{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.374257307497089,
  "eval_steps": 1000,
  "global_step": 180000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.9856985041650495e-05,
      "grad_norm": 8.064935684204102,
      "learning_rate": 0.0,
      "loss": 3.4849,
      "step": 1
    },
    {
      "epoch": 0.0014928492520825246,
      "grad_norm": 3.3433420658111572,
      "learning_rate": 0.00019998742849959144,
      "loss": 1.9038,
      "step": 50
    },
    {
      "epoch": 0.0029856985041650493,
      "grad_norm": 2.907883405685425,
      "learning_rate": 0.0001999731427036726,
      "loss": 1.608,
      "step": 100
    },
    {
      "epoch": 0.004478547756247574,
      "grad_norm": 3.2614288330078125,
      "learning_rate": 0.00019995885690775376,
      "loss": 1.5505,
      "step": 150
    },
    {
      "epoch": 0.005971397008330099,
      "grad_norm": 3.8400654792785645,
      "learning_rate": 0.00019994457111183493,
      "loss": 1.5737,
      "step": 200
    },
    {
      "epoch": 0.007464246260412624,
      "grad_norm": 3.3358442783355713,
      "learning_rate": 0.00019993028531591612,
      "loss": 1.567,
      "step": 250
    },
    {
      "epoch": 0.008957095512495149,
      "grad_norm": 2.131911277770996,
      "learning_rate": 0.00019991599951999726,
      "loss": 1.5208,
      "step": 300
    },
    {
      "epoch": 0.010449944764577673,
      "grad_norm": 3.180992364883423,
      "learning_rate": 0.00019990171372407845,
      "loss": 1.5586,
      "step": 350
    },
    {
      "epoch": 0.011942794016660197,
      "grad_norm": 3.024989128112793,
      "learning_rate": 0.0001998874279281596,
      "loss": 1.5267,
      "step": 400
    },
    {
      "epoch": 0.013435643268742723,
      "grad_norm": 3.4935102462768555,
      "learning_rate": 0.00019987314213224078,
      "loss": 1.5609,
      "step": 450
    },
    {
      "epoch": 0.014928492520825247,
      "grad_norm": 3.565504550933838,
      "learning_rate": 0.00019985885633632194,
      "loss": 1.5112,
      "step": 500
    },
    {
      "epoch": 0.01642134177290777,
      "grad_norm": 2.0692882537841797,
      "learning_rate": 0.00019984457054040308,
      "loss": 1.5087,
      "step": 550
    },
    {
      "epoch": 0.017914191024990297,
      "grad_norm": 4.110323905944824,
      "learning_rate": 0.00019983028474448427,
      "loss": 1.5073,
      "step": 600
    },
    {
      "epoch": 0.01940704027707282,
      "grad_norm": 2.8632736206054688,
      "learning_rate": 0.0001998159989485654,
      "loss": 1.472,
      "step": 650
    },
    {
      "epoch": 0.020899889529155346,
      "grad_norm": 2.629347324371338,
      "learning_rate": 0.0001998017131526466,
      "loss": 1.5236,
      "step": 700
    },
    {
      "epoch": 0.022392738781237872,
      "grad_norm": 3.696873188018799,
      "learning_rate": 0.00019978742735672774,
      "loss": 1.5216,
      "step": 750
    },
    {
      "epoch": 0.023885588033320394,
      "grad_norm": 3.253305435180664,
      "learning_rate": 0.00019977314156080893,
      "loss": 1.5915,
      "step": 800
    },
    {
      "epoch": 0.02537843728540292,
      "grad_norm": 2.9587886333465576,
      "learning_rate": 0.0001997588557648901,
      "loss": 1.4862,
      "step": 850
    },
    {
      "epoch": 0.026871286537485446,
      "grad_norm": 3.3311073780059814,
      "learning_rate": 0.00019974456996897126,
      "loss": 1.4703,
      "step": 900
    },
    {
      "epoch": 0.02836413578956797,
      "grad_norm": 2.5960264205932617,
      "learning_rate": 0.00019973028417305243,
      "loss": 1.4818,
      "step": 950
    },
    {
      "epoch": 0.029856985041650495,
      "grad_norm": 3.3142144680023193,
      "learning_rate": 0.0001997159983771336,
      "loss": 1.4746,
      "step": 1000
    },
    {
      "epoch": 0.03134983429373302,
      "grad_norm": 3.5049827098846436,
      "learning_rate": 0.00019970171258121476,
      "loss": 1.442,
      "step": 1050
    },
    {
      "epoch": 0.03284268354581554,
      "grad_norm": 3.0218605995178223,
      "learning_rate": 0.00019968742678529592,
      "loss": 1.5265,
      "step": 1100
    },
    {
      "epoch": 0.03433553279789807,
      "grad_norm": 2.936182975769043,
      "learning_rate": 0.00019967314098937709,
      "loss": 1.5174,
      "step": 1150
    },
    {
      "epoch": 0.035828382049980595,
      "grad_norm": 2.877253293991089,
      "learning_rate": 0.00019965885519345825,
      "loss": 1.4499,
      "step": 1200
    },
    {
      "epoch": 0.03732123130206312,
      "grad_norm": 6.07016658782959,
      "learning_rate": 0.00019964456939753941,
      "loss": 1.4542,
      "step": 1250
    },
    {
      "epoch": 0.03881408055414564,
      "grad_norm": 2.1618189811706543,
      "learning_rate": 0.0001996302836016206,
      "loss": 1.4343,
      "step": 1300
    },
    {
      "epoch": 0.040306929806228166,
      "grad_norm": 2.8267719745635986,
      "learning_rate": 0.00019961599780570174,
      "loss": 1.463,
      "step": 1350
    },
    {
      "epoch": 0.04179977905831069,
      "grad_norm": 2.6036462783813477,
      "learning_rate": 0.00019960171200978294,
      "loss": 1.4557,
      "step": 1400
    },
    {
      "epoch": 0.04329262831039322,
      "grad_norm": 3.0187127590179443,
      "learning_rate": 0.00019958742621386407,
      "loss": 1.4472,
      "step": 1450
    },
    {
      "epoch": 0.044785477562475744,
      "grad_norm": 3.9822633266448975,
      "learning_rate": 0.00019957314041794526,
      "loss": 1.4384,
      "step": 1500
    },
    {
      "epoch": 0.04627832681455826,
      "grad_norm": 2.919654607772827,
      "learning_rate": 0.0001995588546220264,
      "loss": 1.4969,
      "step": 1550
    },
    {
      "epoch": 0.04777117606664079,
      "grad_norm": 2.922963857650757,
      "learning_rate": 0.0001995445688261076,
      "loss": 1.4987,
      "step": 1600
    },
    {
      "epoch": 0.049264025318723315,
      "grad_norm": 2.9638512134552,
      "learning_rate": 0.00019953028303018876,
      "loss": 1.4722,
      "step": 1650
    },
    {
      "epoch": 0.05075687457080584,
      "grad_norm": 3.408391237258911,
      "learning_rate": 0.00019951599723426992,
      "loss": 1.4723,
      "step": 1700
    },
    {
      "epoch": 0.052249723822888366,
      "grad_norm": 3.023597240447998,
      "learning_rate": 0.0001995017114383511,
      "loss": 1.4478,
      "step": 1750
    },
    {
      "epoch": 0.05374257307497089,
      "grad_norm": 2.1655213832855225,
      "learning_rate": 0.00019948742564243225,
      "loss": 1.4071,
      "step": 1800
    },
    {
      "epoch": 0.05523542232705341,
      "grad_norm": 3.796663999557495,
      "learning_rate": 0.00019947313984651342,
      "loss": 1.446,
      "step": 1850
    },
    {
      "epoch": 0.05672827157913594,
      "grad_norm": 3.0415594577789307,
      "learning_rate": 0.00019945885405059458,
      "loss": 1.4324,
      "step": 1900
    },
    {
      "epoch": 0.05822112083121846,
      "grad_norm": 2.524627685546875,
      "learning_rate": 0.00019944456825467575,
      "loss": 1.3963,
      "step": 1950
    },
    {
      "epoch": 0.05971397008330099,
      "grad_norm": 3.2881991863250732,
      "learning_rate": 0.0001994302824587569,
      "loss": 1.4522,
      "step": 2000
    },
    {
      "epoch": 0.061206819335383515,
      "grad_norm": 3.392430067062378,
      "learning_rate": 0.00019941599666283808,
      "loss": 1.4329,
      "step": 2050
    },
    {
      "epoch": 0.06269966858746603,
      "grad_norm": 3.9426393508911133,
      "learning_rate": 0.00019940171086691927,
      "loss": 1.5203,
      "step": 2100
    },
    {
      "epoch": 0.06419251783954856,
      "grad_norm": 3.3737235069274902,
      "learning_rate": 0.0001993874250710004,
      "loss": 1.3674,
      "step": 2150
    },
    {
      "epoch": 0.06568536709163109,
      "grad_norm": 3.783085346221924,
      "learning_rate": 0.0001993731392750816,
      "loss": 1.4339,
      "step": 2200
    },
    {
      "epoch": 0.06717821634371361,
      "grad_norm": 3.4819202423095703,
      "learning_rate": 0.00019935885347916273,
      "loss": 1.4436,
      "step": 2250
    },
    {
      "epoch": 0.06867106559579614,
      "grad_norm": 3.141775608062744,
      "learning_rate": 0.00019934456768324393,
      "loss": 1.4683,
      "step": 2300
    },
    {
      "epoch": 0.07016391484787866,
      "grad_norm": 3.2881035804748535,
      "learning_rate": 0.00019933028188732506,
      "loss": 1.4395,
      "step": 2350
    },
    {
      "epoch": 0.07165676409996119,
      "grad_norm": 3.718122959136963,
      "learning_rate": 0.00019931599609140626,
      "loss": 1.4396,
      "step": 2400
    },
    {
      "epoch": 0.07314961335204372,
      "grad_norm": 4.3829474449157715,
      "learning_rate": 0.00019930171029548742,
      "loss": 1.4477,
      "step": 2450
    },
    {
      "epoch": 0.07464246260412624,
      "grad_norm": 3.3698525428771973,
      "learning_rate": 0.00019928742449956858,
      "loss": 1.3529,
      "step": 2500
    },
    {
      "epoch": 0.07613531185620875,
      "grad_norm": 3.7569565773010254,
      "learning_rate": 0.00019927313870364975,
      "loss": 1.4246,
      "step": 2550
    },
    {
      "epoch": 0.07762816110829128,
      "grad_norm": 3.1486406326293945,
      "learning_rate": 0.00019925885290773091,
      "loss": 1.3813,
      "step": 2600
    },
    {
      "epoch": 0.0791210103603738,
      "grad_norm": 4.0635480880737305,
      "learning_rate": 0.00019924456711181208,
      "loss": 1.4259,
      "step": 2650
    },
    {
      "epoch": 0.08061385961245633,
      "grad_norm": 3.2710611820220947,
      "learning_rate": 0.00019923028131589324,
      "loss": 1.3747,
      "step": 2700
    },
    {
      "epoch": 0.08210670886453886,
      "grad_norm": 3.4968345165252686,
      "learning_rate": 0.0001992159955199744,
      "loss": 1.4721,
      "step": 2750
    },
    {
      "epoch": 0.08359955811662138,
      "grad_norm": 4.274214267730713,
      "learning_rate": 0.00019920170972405557,
      "loss": 1.437,
      "step": 2800
    },
    {
      "epoch": 0.08509240736870391,
      "grad_norm": 2.970602512359619,
      "learning_rate": 0.00019918742392813674,
      "loss": 1.3336,
      "step": 2850
    },
    {
      "epoch": 0.08658525662078644,
      "grad_norm": 4.143342971801758,
      "learning_rate": 0.00019917313813221793,
      "loss": 1.4264,
      "step": 2900
    },
    {
      "epoch": 0.08807810587286896,
      "grad_norm": 3.7546920776367188,
      "learning_rate": 0.00019915885233629907,
      "loss": 1.441,
      "step": 2950
    },
    {
      "epoch": 0.08957095512495149,
      "grad_norm": 3.9160516262054443,
      "learning_rate": 0.00019914456654038026,
      "loss": 1.4261,
      "step": 3000
    },
    {
      "epoch": 0.09106380437703401,
      "grad_norm": 3.842073917388916,
      "learning_rate": 0.0001991302807444614,
      "loss": 1.4076,
      "step": 3050
    },
    {
      "epoch": 0.09255665362911653,
      "grad_norm": 4.392395496368408,
      "learning_rate": 0.0001991159949485426,
      "loss": 1.3789,
      "step": 3100
    },
    {
      "epoch": 0.09404950288119905,
      "grad_norm": 3.822425603866577,
      "learning_rate": 0.00019910170915262373,
      "loss": 1.3877,
      "step": 3150
    },
    {
      "epoch": 0.09554235213328158,
      "grad_norm": 3.1348562240600586,
      "learning_rate": 0.0001990874233567049,
      "loss": 1.4081,
      "step": 3200
    },
    {
      "epoch": 0.0970352013853641,
      "grad_norm": 3.453887939453125,
      "learning_rate": 0.00019907313756078608,
      "loss": 1.4143,
      "step": 3250
    },
    {
      "epoch": 0.09852805063744663,
      "grad_norm": 3.5057384967803955,
      "learning_rate": 0.00019905885176486722,
      "loss": 1.4264,
      "step": 3300
    },
    {
      "epoch": 0.10002089988952916,
      "grad_norm": 3.145796060562134,
      "learning_rate": 0.0001990445659689484,
      "loss": 1.4368,
      "step": 3350
    },
    {
      "epoch": 0.10151374914161168,
      "grad_norm": 3.4077043533325195,
      "learning_rate": 0.00019903028017302955,
      "loss": 1.388,
      "step": 3400
    },
    {
      "epoch": 0.10300659839369421,
      "grad_norm": 3.65567946434021,
      "learning_rate": 0.00019901599437711074,
      "loss": 1.42,
      "step": 3450
    },
    {
      "epoch": 0.10449944764577673,
      "grad_norm": 4.460702419281006,
      "learning_rate": 0.0001990017085811919,
      "loss": 1.3991,
      "step": 3500
    },
    {
      "epoch": 0.10599229689785926,
      "grad_norm": 4.155653476715088,
      "learning_rate": 0.00019898742278527307,
      "loss": 1.371,
      "step": 3550
    },
    {
      "epoch": 0.10748514614994178,
      "grad_norm": 3.8904318809509277,
      "learning_rate": 0.00019897313698935423,
      "loss": 1.4378,
      "step": 3600
    },
    {
      "epoch": 0.1089779954020243,
      "grad_norm": 4.0509233474731445,
      "learning_rate": 0.0001989588511934354,
      "loss": 1.3945,
      "step": 3650
    },
    {
      "epoch": 0.11047084465410682,
      "grad_norm": 3.785123109817505,
      "learning_rate": 0.00019894456539751656,
      "loss": 1.436,
      "step": 3700
    },
    {
      "epoch": 0.11196369390618935,
      "grad_norm": 3.4556167125701904,
      "learning_rate": 0.00019893027960159773,
      "loss": 1.3794,
      "step": 3750
    },
    {
      "epoch": 0.11345654315827187,
      "grad_norm": 4.0479559898376465,
      "learning_rate": 0.0001989159938056789,
      "loss": 1.4734,
      "step": 3800
    },
    {
      "epoch": 0.1149493924103544,
      "grad_norm": 3.890805721282959,
      "learning_rate": 0.00019890170800976006,
      "loss": 1.4341,
      "step": 3850
    },
    {
      "epoch": 0.11644224166243693,
      "grad_norm": 3.8178727626800537,
      "learning_rate": 0.00019888742221384122,
      "loss": 1.4754,
      "step": 3900
    },
    {
      "epoch": 0.11793509091451945,
      "grad_norm": 2.456165075302124,
      "learning_rate": 0.00019887313641792241,
      "loss": 1.3887,
      "step": 3950
    },
    {
      "epoch": 0.11942794016660198,
      "grad_norm": 3.5763051509857178,
      "learning_rate": 0.00019885885062200355,
      "loss": 1.3901,
      "step": 4000
    },
    {
      "epoch": 0.1209207894186845,
      "grad_norm": 3.885662317276001,
      "learning_rate": 0.00019884456482608474,
      "loss": 1.3856,
      "step": 4050
    },
    {
      "epoch": 0.12241363867076703,
      "grad_norm": 3.6095409393310547,
      "learning_rate": 0.00019883027903016588,
      "loss": 1.448,
      "step": 4100
    },
    {
      "epoch": 0.12390648792284956,
      "grad_norm": 3.7112534046173096,
      "learning_rate": 0.00019881599323424707,
      "loss": 1.3537,
      "step": 4150
    },
    {
      "epoch": 0.12539933717493207,
      "grad_norm": 3.3566672801971436,
      "learning_rate": 0.0001988017074383282,
      "loss": 1.4389,
      "step": 4200
    },
    {
      "epoch": 0.1268921864270146,
      "grad_norm": 4.570401191711426,
      "learning_rate": 0.0001987874216424094,
      "loss": 1.4191,
      "step": 4250
    },
    {
      "epoch": 0.12838503567909712,
      "grad_norm": 4.455029010772705,
      "learning_rate": 0.00019877313584649057,
      "loss": 1.3677,
      "step": 4300
    },
    {
      "epoch": 0.12987788493117966,
      "grad_norm": 3.0861828327178955,
      "learning_rate": 0.00019875885005057173,
      "loss": 1.3677,
      "step": 4350
    },
    {
      "epoch": 0.13137073418326217,
      "grad_norm": 4.419896602630615,
      "learning_rate": 0.0001987445642546529,
      "loss": 1.4524,
      "step": 4400
    },
    {
      "epoch": 0.1328635834353447,
      "grad_norm": 5.187576770782471,
      "learning_rate": 0.00019873027845873406,
      "loss": 1.3868,
      "step": 4450
    },
    {
      "epoch": 0.13435643268742722,
      "grad_norm": 5.111696243286133,
      "learning_rate": 0.00019871599266281523,
      "loss": 1.4458,
      "step": 4500
    },
    {
      "epoch": 0.13584928193950974,
      "grad_norm": 3.2652997970581055,
      "learning_rate": 0.0001987017068668964,
      "loss": 1.4529,
      "step": 4550
    },
    {
      "epoch": 0.13734213119159228,
      "grad_norm": 4.190273761749268,
      "learning_rate": 0.00019868742107097755,
      "loss": 1.3991,
      "step": 4600
    },
    {
      "epoch": 0.1388349804436748,
      "grad_norm": 4.85620641708374,
      "learning_rate": 0.00019867313527505872,
      "loss": 1.3916,
      "step": 4650
    },
    {
      "epoch": 0.14032782969575733,
      "grad_norm": 3.030954360961914,
      "learning_rate": 0.00019865884947913988,
      "loss": 1.3805,
      "step": 4700
    },
    {
      "epoch": 0.14182067894783984,
      "grad_norm": 3.264406681060791,
      "learning_rate": 0.00019864456368322108,
      "loss": 1.4048,
      "step": 4750
    },
    {
      "epoch": 0.14331352819992238,
      "grad_norm": 3.2138588428497314,
      "learning_rate": 0.0001986302778873022,
      "loss": 1.4092,
      "step": 4800
    },
    {
      "epoch": 0.1448063774520049,
      "grad_norm": 3.847222328186035,
      "learning_rate": 0.0001986159920913834,
      "loss": 1.3871,
      "step": 4850
    },
    {
      "epoch": 0.14629922670408743,
      "grad_norm": 4.004987716674805,
      "learning_rate": 0.00019860170629546454,
      "loss": 1.3845,
      "step": 4900
    },
    {
      "epoch": 0.14779207595616994,
      "grad_norm": 3.5088725090026855,
      "learning_rate": 0.00019858742049954573,
      "loss": 1.379,
      "step": 4950
    },
    {
      "epoch": 0.14928492520825248,
      "grad_norm": 3.275099277496338,
      "learning_rate": 0.00019857313470362687,
      "loss": 1.3628,
      "step": 5000
    },
    {
      "epoch": 0.150777774460335,
      "grad_norm": 3.7903060913085938,
      "learning_rate": 0.00019855884890770806,
      "loss": 1.3804,
      "step": 5050
    },
    {
      "epoch": 0.1522706237124175,
      "grad_norm": 4.294798374176025,
      "learning_rate": 0.00019854456311178923,
      "loss": 1.3988,
      "step": 5100
    },
    {
      "epoch": 0.15376347296450005,
      "grad_norm": 3.2719295024871826,
      "learning_rate": 0.0001985302773158704,
      "loss": 1.387,
      "step": 5150
    },
    {
      "epoch": 0.15525632221658256,
      "grad_norm": 4.143224239349365,
      "learning_rate": 0.00019851599151995156,
      "loss": 1.393,
      "step": 5200
    },
    {
      "epoch": 0.1567491714686651,
      "grad_norm": 3.404754638671875,
      "learning_rate": 0.00019850170572403272,
      "loss": 1.4205,
      "step": 5250
    },
    {
      "epoch": 0.1582420207207476,
      "grad_norm": 3.607126474380493,
      "learning_rate": 0.0001984874199281139,
      "loss": 1.4239,
      "step": 5300
    },
    {
      "epoch": 0.15973486997283015,
      "grad_norm": 4.140823841094971,
      "learning_rate": 0.00019847313413219505,
      "loss": 1.4204,
      "step": 5350
    },
    {
      "epoch": 0.16122771922491266,
      "grad_norm": 3.893251419067383,
      "learning_rate": 0.00019845884833627622,
      "loss": 1.392,
      "step": 5400
    },
    {
      "epoch": 0.1627205684769952,
      "grad_norm": 4.304211139678955,
      "learning_rate": 0.00019844456254035738,
      "loss": 1.44,
      "step": 5450
    },
    {
      "epoch": 0.16421341772907772,
      "grad_norm": 5.273501873016357,
      "learning_rate": 0.00019843027674443855,
      "loss": 1.445,
      "step": 5500
    },
    {
      "epoch": 0.16570626698116026,
      "grad_norm": 4.787700176239014,
      "learning_rate": 0.00019841599094851974,
      "loss": 1.3668,
      "step": 5550
    },
    {
      "epoch": 0.16719911623324277,
      "grad_norm": 3.7984108924865723,
      "learning_rate": 0.00019840170515260087,
      "loss": 1.3523,
      "step": 5600
    },
    {
      "epoch": 0.16869196548532528,
      "grad_norm": 3.885608673095703,
      "learning_rate": 0.00019838741935668207,
      "loss": 1.3917,
      "step": 5650
    },
    {
      "epoch": 0.17018481473740782,
      "grad_norm": 3.459803342819214,
      "learning_rate": 0.0001983731335607632,
      "loss": 1.3833,
      "step": 5700
    },
    {
      "epoch": 0.17167766398949033,
      "grad_norm": 3.7103006839752197,
      "learning_rate": 0.0001983588477648444,
      "loss": 1.4473,
      "step": 5750
    },
    {
      "epoch": 0.17317051324157287,
      "grad_norm": 6.645928382873535,
      "learning_rate": 0.00019834456196892553,
      "loss": 1.3706,
      "step": 5800
    },
    {
      "epoch": 0.17466336249365538,
      "grad_norm": 3.7201037406921387,
      "learning_rate": 0.0001983302761730067,
      "loss": 1.3733,
      "step": 5850
    },
    {
      "epoch": 0.17615621174573792,
      "grad_norm": 4.050106048583984,
      "learning_rate": 0.0001983159903770879,
      "loss": 1.4096,
      "step": 5900
    },
    {
      "epoch": 0.17764906099782043,
      "grad_norm": 4.190842628479004,
      "learning_rate": 0.00019830170458116903,
      "loss": 1.4404,
      "step": 5950
    },
    {
      "epoch": 0.17914191024990297,
      "grad_norm": 4.393162727355957,
      "learning_rate": 0.00019828741878525022,
      "loss": 1.4443,
      "step": 6000
    },
    {
      "epoch": 0.1806347595019855,
      "grad_norm": 3.597520351409912,
      "learning_rate": 0.00019827313298933136,
      "loss": 1.4063,
      "step": 6050
    },
    {
      "epoch": 0.18212760875406803,
      "grad_norm": 3.608085870742798,
      "learning_rate": 0.00019825884719341255,
      "loss": 1.3857,
      "step": 6100
    },
    {
      "epoch": 0.18362045800615054,
      "grad_norm": 3.7055492401123047,
      "learning_rate": 0.0001982445613974937,
      "loss": 1.3997,
      "step": 6150
    },
    {
      "epoch": 0.18511330725823305,
      "grad_norm": 3.875457763671875,
      "learning_rate": 0.00019823027560157488,
      "loss": 1.4296,
      "step": 6200
    },
    {
      "epoch": 0.1866061565103156,
      "grad_norm": 5.074592590332031,
      "learning_rate": 0.00019821598980565604,
      "loss": 1.3785,
      "step": 6250
    },
    {
      "epoch": 0.1880990057623981,
      "grad_norm": 6.013392448425293,
      "learning_rate": 0.0001982017040097372,
      "loss": 1.4391,
      "step": 6300
    },
    {
      "epoch": 0.18959185501448064,
      "grad_norm": 5.679958820343018,
      "learning_rate": 0.00019818741821381837,
      "loss": 1.367,
      "step": 6350
    },
    {
      "epoch": 0.19108470426656315,
      "grad_norm": 3.6182546615600586,
      "learning_rate": 0.00019817313241789954,
      "loss": 1.4508,
      "step": 6400
    },
    {
      "epoch": 0.1925775535186457,
      "grad_norm": 5.209213733673096,
      "learning_rate": 0.0001981588466219807,
      "loss": 1.3878,
      "step": 6450
    },
    {
      "epoch": 0.1940704027707282,
      "grad_norm": 3.0043230056762695,
      "learning_rate": 0.00019814456082606187,
      "loss": 1.4248,
      "step": 6500
    },
    {
      "epoch": 0.19556325202281075,
      "grad_norm": 3.157851219177246,
      "learning_rate": 0.00019813027503014303,
      "loss": 1.3725,
      "step": 6550
    },
    {
      "epoch": 0.19705610127489326,
      "grad_norm": 3.5292418003082275,
      "learning_rate": 0.0001981159892342242,
      "loss": 1.3932,
      "step": 6600
    },
    {
      "epoch": 0.1985489505269758,
      "grad_norm": 3.2819600105285645,
      "learning_rate": 0.00019810170343830536,
      "loss": 1.3495,
      "step": 6650
    },
    {
      "epoch": 0.2000417997790583,
      "grad_norm": 3.0243399143218994,
      "learning_rate": 0.00019808741764238655,
      "loss": 1.3689,
      "step": 6700
    },
    {
      "epoch": 0.20153464903114082,
      "grad_norm": 3.4495368003845215,
      "learning_rate": 0.0001980731318464677,
      "loss": 1.3725,
      "step": 6750
    },
    {
      "epoch": 0.20302749828322336,
      "grad_norm": 3.538259744644165,
      "learning_rate": 0.00019805884605054888,
      "loss": 1.3905,
      "step": 6800
    },
    {
      "epoch": 0.20452034753530587,
      "grad_norm": 4.162181377410889,
      "learning_rate": 0.00019804456025463002,
      "loss": 1.4129,
      "step": 6850
    },
    {
      "epoch": 0.20601319678738841,
      "grad_norm": 4.592432022094727,
      "learning_rate": 0.0001980302744587112,
      "loss": 1.3634,
      "step": 6900
    },
    {
      "epoch": 0.20750604603947093,
      "grad_norm": 3.45967960357666,
      "learning_rate": 0.00019801598866279237,
      "loss": 1.416,
      "step": 6950
    },
    {
      "epoch": 0.20899889529155347,
      "grad_norm": 4.221930503845215,
      "learning_rate": 0.00019800170286687354,
      "loss": 1.4051,
      "step": 7000
    },
    {
      "epoch": 0.21049174454363598,
      "grad_norm": 4.144239902496338,
      "learning_rate": 0.0001979874170709547,
      "loss": 1.4219,
      "step": 7050
    },
    {
      "epoch": 0.21198459379571852,
      "grad_norm": 4.7492570877075195,
      "learning_rate": 0.00019797313127503587,
      "loss": 1.4028,
      "step": 7100
    },
    {
      "epoch": 0.21347744304780103,
      "grad_norm": 3.5841355323791504,
      "learning_rate": 0.00019795884547911703,
      "loss": 1.4361,
      "step": 7150
    },
    {
      "epoch": 0.21497029229988357,
      "grad_norm": 4.662593364715576,
      "learning_rate": 0.0001979445596831982,
      "loss": 1.3816,
      "step": 7200
    },
    {
      "epoch": 0.21646314155196608,
      "grad_norm": 4.700701713562012,
      "learning_rate": 0.00019793027388727936,
      "loss": 1.4226,
      "step": 7250
    },
    {
      "epoch": 0.2179559908040486,
      "grad_norm": 4.025181293487549,
      "learning_rate": 0.00019791598809136053,
      "loss": 1.4291,
      "step": 7300
    },
    {
      "epoch": 0.21944884005613113,
      "grad_norm": 3.064573049545288,
      "learning_rate": 0.0001979017022954417,
      "loss": 1.4293,
      "step": 7350
    },
    {
      "epoch": 0.22094168930821365,
      "grad_norm": 6.342152118682861,
      "learning_rate": 0.00019788741649952288,
      "loss": 1.4173,
      "step": 7400
    },
    {
      "epoch": 0.22243453856029619,
      "grad_norm": 5.89996337890625,
      "learning_rate": 0.00019787313070360402,
      "loss": 1.396,
      "step": 7450
    },
    {
      "epoch": 0.2239273878123787,
      "grad_norm": 4.462945938110352,
      "learning_rate": 0.0001978588449076852,
      "loss": 1.3868,
      "step": 7500
    },
    {
      "epoch": 0.22542023706446124,
      "grad_norm": 3.6449055671691895,
      "learning_rate": 0.00019784455911176635,
      "loss": 1.396,
      "step": 7550
    },
    {
      "epoch": 0.22691308631654375,
      "grad_norm": 4.674243927001953,
      "learning_rate": 0.00019783027331584754,
      "loss": 1.395,
      "step": 7600
    },
    {
      "epoch": 0.2284059355686263,
      "grad_norm": 3.6160385608673096,
      "learning_rate": 0.00019781598751992868,
      "loss": 1.3918,
      "step": 7650
    },
    {
      "epoch": 0.2298987848207088,
      "grad_norm": 4.326193332672119,
      "learning_rate": 0.00019780170172400987,
      "loss": 1.3947,
      "step": 7700
    },
    {
      "epoch": 0.23139163407279134,
      "grad_norm": 5.4003777503967285,
      "learning_rate": 0.00019778741592809104,
      "loss": 1.344,
      "step": 7750
    },
    {
      "epoch": 0.23288448332487385,
      "grad_norm": 4.711580753326416,
      "learning_rate": 0.0001977731301321722,
      "loss": 1.3959,
      "step": 7800
    },
    {
      "epoch": 0.23437733257695637,
      "grad_norm": 3.4752814769744873,
      "learning_rate": 0.00019775884433625337,
      "loss": 1.3722,
      "step": 7850
    },
    {
      "epoch": 0.2358701818290389,
      "grad_norm": 4.028527736663818,
      "learning_rate": 0.00019774455854033453,
      "loss": 1.3683,
      "step": 7900
    },
    {
      "epoch": 0.23736303108112142,
      "grad_norm": 4.094334602355957,
      "learning_rate": 0.0001977302727444157,
      "loss": 1.3607,
      "step": 7950
    },
    {
      "epoch": 0.23885588033320396,
      "grad_norm": 5.232580661773682,
      "learning_rate": 0.00019771598694849686,
      "loss": 1.4354,
      "step": 8000
    },
    {
      "epoch": 0.24034872958528647,
      "grad_norm": 4.269852161407471,
      "learning_rate": 0.00019770170115257802,
      "loss": 1.4372,
      "step": 8050
    },
    {
      "epoch": 0.241841578837369,
      "grad_norm": 3.312541961669922,
      "learning_rate": 0.0001976874153566592,
      "loss": 1.3826,
      "step": 8100
    },
    {
      "epoch": 0.24333442808945152,
      "grad_norm": 3.8900692462921143,
      "learning_rate": 0.00019767312956074035,
      "loss": 1.4189,
      "step": 8150
    },
    {
      "epoch": 0.24482727734153406,
      "grad_norm": 3.894512414932251,
      "learning_rate": 0.00019765884376482155,
      "loss": 1.3365,
      "step": 8200
    },
    {
      "epoch": 0.24632012659361657,
      "grad_norm": 4.644411563873291,
      "learning_rate": 0.00019764455796890268,
      "loss": 1.4311,
      "step": 8250
    },
    {
      "epoch": 0.2478129758456991,
      "grad_norm": 8.174029350280762,
      "learning_rate": 0.00019763027217298387,
      "loss": 1.361,
      "step": 8300
    },
    {
      "epoch": 0.24930582509778162,
      "grad_norm": 4.615732192993164,
      "learning_rate": 0.000197615986377065,
      "loss": 1.4552,
      "step": 8350
    },
    {
      "epoch": 0.25079867434986414,
      "grad_norm": 4.421249866485596,
      "learning_rate": 0.0001976017005811462,
      "loss": 1.3463,
      "step": 8400
    },
    {
      "epoch": 0.2522915236019467,
      "grad_norm": 2.8386716842651367,
      "learning_rate": 0.00019758741478522734,
      "loss": 1.348,
      "step": 8450
    },
    {
      "epoch": 0.2537843728540292,
      "grad_norm": 4.3141703605651855,
      "learning_rate": 0.0001975731289893085,
      "loss": 1.4306,
      "step": 8500
    },
    {
      "epoch": 0.2552772221061117,
      "grad_norm": 3.947331428527832,
      "learning_rate": 0.0001975588431933897,
      "loss": 1.3823,
      "step": 8550
    },
    {
      "epoch": 0.25677007135819424,
      "grad_norm": 3.2268636226654053,
      "learning_rate": 0.00019754455739747084,
      "loss": 1.4199,
      "step": 8600
    },
    {
      "epoch": 0.2582629206102768,
      "grad_norm": 4.0353102684021,
      "learning_rate": 0.00019753027160155203,
      "loss": 1.3927,
      "step": 8650
    },
    {
      "epoch": 0.2597557698623593,
      "grad_norm": 3.490560293197632,
      "learning_rate": 0.00019751598580563316,
      "loss": 1.401,
      "step": 8700
    },
    {
      "epoch": 0.2612486191144418,
      "grad_norm": 5.577207088470459,
      "learning_rate": 0.00019750170000971436,
      "loss": 1.3586,
      "step": 8750
    },
    {
      "epoch": 0.26274146836652434,
      "grad_norm": 4.168467998504639,
      "learning_rate": 0.0001974874142137955,
      "loss": 1.3303,
      "step": 8800
    },
    {
      "epoch": 0.2642343176186069,
      "grad_norm": 3.812627077102661,
      "learning_rate": 0.00019747312841787669,
      "loss": 1.3717,
      "step": 8850
    },
    {
      "epoch": 0.2657271668706894,
      "grad_norm": 4.875237464904785,
      "learning_rate": 0.00019745884262195785,
      "loss": 1.3873,
      "step": 8900
    },
    {
      "epoch": 0.2672200161227719,
      "grad_norm": 4.048189163208008,
      "learning_rate": 0.00019744455682603902,
      "loss": 1.3775,
      "step": 8950
    },
    {
      "epoch": 0.26871286537485445,
      "grad_norm": 3.9090261459350586,
      "learning_rate": 0.00019743027103012018,
      "loss": 1.4296,
      "step": 9000
    },
    {
      "epoch": 0.270205714626937,
      "grad_norm": 2.8476953506469727,
      "learning_rate": 0.00019741598523420134,
      "loss": 1.4175,
      "step": 9050
    },
    {
      "epoch": 0.2716985638790195,
      "grad_norm": 5.782102584838867,
      "learning_rate": 0.0001974016994382825,
      "loss": 1.3835,
      "step": 9100
    },
    {
      "epoch": 0.273191413131102,
      "grad_norm": 4.640264987945557,
      "learning_rate": 0.00019738741364236367,
      "loss": 1.4524,
      "step": 9150
    },
    {
      "epoch": 0.27468426238318455,
      "grad_norm": 4.81790828704834,
      "learning_rate": 0.00019737312784644484,
      "loss": 1.3183,
      "step": 9200
    },
    {
      "epoch": 0.2761771116352671,
      "grad_norm": 2.685009717941284,
      "learning_rate": 0.000197358842050526,
      "loss": 1.3243,
      "step": 9250
    },
    {
      "epoch": 0.2776699608873496,
      "grad_norm": 5.321321487426758,
      "learning_rate": 0.00019734455625460717,
      "loss": 1.4086,
      "step": 9300
    },
    {
      "epoch": 0.2791628101394321,
      "grad_norm": 3.065791368484497,
      "learning_rate": 0.00019733027045868836,
      "loss": 1.337,
      "step": 9350
    },
    {
      "epoch": 0.28065565939151466,
      "grad_norm": 4.3569817543029785,
      "learning_rate": 0.0001973159846627695,
      "loss": 1.4082,
      "step": 9400
    },
    {
      "epoch": 0.2821485086435972,
      "grad_norm": 4.67582368850708,
      "learning_rate": 0.0001973016988668507,
      "loss": 1.3832,
      "step": 9450
    },
    {
      "epoch": 0.2836413578956797,
      "grad_norm": 4.942144870758057,
      "learning_rate": 0.00019728741307093183,
      "loss": 1.3734,
      "step": 9500
    },
    {
      "epoch": 0.2851342071477622,
      "grad_norm": 4.853246688842773,
      "learning_rate": 0.00019727312727501302,
      "loss": 1.4111,
      "step": 9550
    },
    {
      "epoch": 0.28662705639984476,
      "grad_norm": 3.071237325668335,
      "learning_rate": 0.00019725884147909418,
      "loss": 1.3746,
      "step": 9600
    },
    {
      "epoch": 0.28811990565192724,
      "grad_norm": 4.844615459442139,
      "learning_rate": 0.00019724455568317535,
      "loss": 1.3051,
      "step": 9650
    },
    {
      "epoch": 0.2896127549040098,
      "grad_norm": 5.954223155975342,
      "learning_rate": 0.0001972302698872565,
      "loss": 1.4131,
      "step": 9700
    },
    {
      "epoch": 0.2911056041560923,
      "grad_norm": 3.6717801094055176,
      "learning_rate": 0.00019721598409133768,
      "loss": 1.4166,
      "step": 9750
    },
    {
      "epoch": 0.29259845340817486,
      "grad_norm": 3.6257095336914062,
      "learning_rate": 0.00019720169829541884,
      "loss": 1.3679,
      "step": 9800
    },
    {
      "epoch": 0.29409130266025735,
      "grad_norm": 4.245635032653809,
      "learning_rate": 0.0001971874124995,
      "loss": 1.3171,
      "step": 9850
    },
    {
      "epoch": 0.2955841519123399,
      "grad_norm": 5.362602710723877,
      "learning_rate": 0.00019717312670358117,
      "loss": 1.3932,
      "step": 9900
    },
    {
      "epoch": 0.2970770011644224,
      "grad_norm": 4.6283721923828125,
      "learning_rate": 0.00019715884090766234,
      "loss": 1.3757,
      "step": 9950
    },
    {
      "epoch": 0.29856985041650497,
      "grad_norm": 4.299574851989746,
      "learning_rate": 0.0001971445551117435,
      "loss": 1.4018,
      "step": 10000
    },
    {
      "epoch": 0.30006269966858745,
      "grad_norm": 4.7913641929626465,
      "learning_rate": 0.0001971302693158247,
      "loss": 1.4228,
      "step": 10050
    },
    {
      "epoch": 0.30155554892067,
      "grad_norm": 5.312823295593262,
      "learning_rate": 0.00019711598351990583,
      "loss": 1.3866,
      "step": 10100
    },
    {
      "epoch": 0.30304839817275253,
      "grad_norm": 4.099662780761719,
      "learning_rate": 0.00019710169772398702,
      "loss": 1.454,
      "step": 10150
    },
    {
      "epoch": 0.304541247424835,
      "grad_norm": 4.254878520965576,
      "learning_rate": 0.00019708741192806816,
      "loss": 1.3526,
      "step": 10200
    },
    {
      "epoch": 0.30603409667691756,
      "grad_norm": 4.056606292724609,
      "learning_rate": 0.00019707312613214935,
      "loss": 1.4167,
      "step": 10250
    },
    {
      "epoch": 0.3075269459290001,
      "grad_norm": 3.790809154510498,
      "learning_rate": 0.0001970588403362305,
      "loss": 1.3536,
      "step": 10300
    },
    {
      "epoch": 0.30901979518108263,
      "grad_norm": 4.46298360824585,
      "learning_rate": 0.00019704455454031168,
      "loss": 1.3613,
      "step": 10350
    },
    {
      "epoch": 0.3105126444331651,
      "grad_norm": 4.52452278137207,
      "learning_rate": 0.00019703026874439284,
      "loss": 1.4591,
      "step": 10400
    },
    {
      "epoch": 0.31200549368524766,
      "grad_norm": 4.735177040100098,
      "learning_rate": 0.000197015982948474,
      "loss": 1.4617,
      "step": 10450
    },
    {
      "epoch": 0.3134983429373302,
      "grad_norm": 4.48261833190918,
      "learning_rate": 0.00019700169715255517,
      "loss": 1.4072,
      "step": 10500
    },
    {
      "epoch": 0.31499119218941274,
      "grad_norm": 3.3441503047943115,
      "learning_rate": 0.00019698741135663634,
      "loss": 1.4082,
      "step": 10550
    },
    {
      "epoch": 0.3164840414414952,
      "grad_norm": 3.9771218299865723,
      "learning_rate": 0.0001969731255607175,
      "loss": 1.4184,
      "step": 10600
    },
    {
      "epoch": 0.31797689069357776,
      "grad_norm": 6.366194725036621,
      "learning_rate": 0.00019695883976479867,
      "loss": 1.387,
      "step": 10650
    },
    {
      "epoch": 0.3194697399456603,
      "grad_norm": 5.072678089141846,
      "learning_rate": 0.00019694455396887983,
      "loss": 1.3996,
      "step": 10700
    },
    {
      "epoch": 0.3209625891977428,
      "grad_norm": 3.7204978466033936,
      "learning_rate": 0.000196930268172961,
      "loss": 1.3774,
      "step": 10750
    },
    {
      "epoch": 0.3224554384498253,
      "grad_norm": 4.47731351852417,
      "learning_rate": 0.00019691598237704216,
      "loss": 1.3552,
      "step": 10800
    },
    {
      "epoch": 0.32394828770190787,
      "grad_norm": 3.4569220542907715,
      "learning_rate": 0.00019690169658112335,
      "loss": 1.3794,
      "step": 10850
    },
    {
      "epoch": 0.3254411369539904,
      "grad_norm": 4.344145774841309,
      "learning_rate": 0.0001968874107852045,
      "loss": 1.4419,
      "step": 10900
    },
    {
      "epoch": 0.3269339862060729,
      "grad_norm": 4.089848041534424,
      "learning_rate": 0.00019687312498928568,
      "loss": 1.4116,
      "step": 10950
    },
    {
      "epoch": 0.32842683545815543,
      "grad_norm": 3.995945930480957,
      "learning_rate": 0.00019685883919336682,
      "loss": 1.3532,
      "step": 11000
    },
    {
      "epoch": 0.32991968471023797,
      "grad_norm": 3.8309378623962402,
      "learning_rate": 0.00019684455339744798,
      "loss": 1.3957,
      "step": 11050
    },
    {
      "epoch": 0.3314125339623205,
      "grad_norm": 4.386235237121582,
      "learning_rate": 0.00019683026760152915,
      "loss": 1.3616,
      "step": 11100
    },
    {
      "epoch": 0.332905383214403,
      "grad_norm": 5.133239269256592,
      "learning_rate": 0.00019681598180561031,
      "loss": 1.3959,
      "step": 11150
    },
    {
      "epoch": 0.33439823246648553,
      "grad_norm": 4.216183662414551,
      "learning_rate": 0.0001968016960096915,
      "loss": 1.3446,
      "step": 11200
    },
    {
      "epoch": 0.3358910817185681,
      "grad_norm": 3.631131172180176,
      "learning_rate": 0.00019678741021377264,
      "loss": 1.3779,
      "step": 11250
    },
    {
      "epoch": 0.33738393097065056,
      "grad_norm": 4.603448390960693,
      "learning_rate": 0.00019677312441785384,
      "loss": 1.3962,
      "step": 11300
    },
    {
      "epoch": 0.3388767802227331,
      "grad_norm": 3.6482913494110107,
      "learning_rate": 0.00019675883862193497,
      "loss": 1.3931,
      "step": 11350
    },
    {
      "epoch": 0.34036962947481564,
      "grad_norm": 5.040388107299805,
      "learning_rate": 0.00019674455282601616,
      "loss": 1.3465,
      "step": 11400
    },
    {
      "epoch": 0.3418624787268982,
      "grad_norm": 5.762825012207031,
      "learning_rate": 0.0001967302670300973,
      "loss": 1.3577,
      "step": 11450
    },
    {
      "epoch": 0.34335532797898066,
      "grad_norm": 4.941501617431641,
      "learning_rate": 0.0001967159812341785,
      "loss": 1.3676,
      "step": 11500
    },
    {
      "epoch": 0.3448481772310632,
      "grad_norm": 5.368370532989502,
      "learning_rate": 0.00019670169543825966,
      "loss": 1.4265,
      "step": 11550
    },
    {
      "epoch": 0.34634102648314574,
      "grad_norm": 4.931522369384766,
      "learning_rate": 0.00019668740964234082,
      "loss": 1.3551,
      "step": 11600
    },
    {
      "epoch": 0.3478338757352283,
      "grad_norm": 3.9685990810394287,
      "learning_rate": 0.000196673123846422,
      "loss": 1.402,
      "step": 11650
    },
    {
      "epoch": 0.34932672498731077,
      "grad_norm": 5.771200656890869,
      "learning_rate": 0.00019665883805050315,
      "loss": 1.3596,
      "step": 11700
    },
    {
      "epoch": 0.3508195742393933,
      "grad_norm": 5.142852306365967,
      "learning_rate": 0.00019664455225458432,
      "loss": 1.385,
      "step": 11750
    },
    {
      "epoch": 0.35231242349147585,
      "grad_norm": 3.295628786087036,
      "learning_rate": 0.00019663026645866548,
      "loss": 1.3454,
      "step": 11800
    },
    {
      "epoch": 0.35380527274355833,
      "grad_norm": 4.34658145904541,
      "learning_rate": 0.00019661598066274665,
      "loss": 1.3976,
      "step": 11850
    },
    {
      "epoch": 0.35529812199564087,
      "grad_norm": 4.032591819763184,
      "learning_rate": 0.0001966016948668278,
      "loss": 1.3571,
      "step": 11900
    },
    {
      "epoch": 0.3567909712477234,
      "grad_norm": 3.9286158084869385,
      "learning_rate": 0.00019658740907090898,
      "loss": 1.398,
      "step": 11950
    },
    {
      "epoch": 0.35828382049980595,
      "grad_norm": 5.184597492218018,
      "learning_rate": 0.00019657312327499017,
      "loss": 1.3827,
      "step": 12000
    },
    {
      "epoch": 0.35977666975188843,
      "grad_norm": 4.4749226570129395,
      "learning_rate": 0.0001965588374790713,
      "loss": 1.373,
      "step": 12050
    },
    {
      "epoch": 0.361269519003971,
      "grad_norm": 3.5633764266967773,
      "learning_rate": 0.0001965445516831525,
      "loss": 1.3704,
      "step": 12100
    },
    {
      "epoch": 0.3627623682560535,
      "grad_norm": 7.570897102355957,
      "learning_rate": 0.00019653026588723363,
      "loss": 1.3515,
      "step": 12150
    },
    {
      "epoch": 0.36425521750813605,
      "grad_norm": 4.239411354064941,
      "learning_rate": 0.00019651598009131483,
      "loss": 1.3813,
      "step": 12200
    },
    {
      "epoch": 0.36574806676021854,
      "grad_norm": 3.8941049575805664,
      "learning_rate": 0.00019650169429539596,
      "loss": 1.3954,
      "step": 12250
    },
    {
      "epoch": 0.3672409160123011,
      "grad_norm": 4.8694586753845215,
      "learning_rate": 0.00019648740849947716,
      "loss": 1.3531,
      "step": 12300
    },
    {
      "epoch": 0.3687337652643836,
      "grad_norm": 3.914964437484741,
      "learning_rate": 0.00019647312270355832,
      "loss": 1.3954,
      "step": 12350
    },
    {
      "epoch": 0.3702266145164661,
      "grad_norm": 3.4050538539886475,
      "learning_rate": 0.00019645883690763948,
      "loss": 1.3498,
      "step": 12400
    },
    {
      "epoch": 0.37171946376854864,
      "grad_norm": 4.436797618865967,
      "learning_rate": 0.00019644455111172065,
      "loss": 1.3889,
      "step": 12450
    },
    {
      "epoch": 0.3732123130206312,
      "grad_norm": 2.7660670280456543,
      "learning_rate": 0.00019643026531580181,
      "loss": 1.3392,
      "step": 12500
    },
    {
      "epoch": 0.3747051622727137,
      "grad_norm": 5.364072799682617,
      "learning_rate": 0.00019641597951988298,
      "loss": 1.3157,
      "step": 12550
    },
    {
      "epoch": 0.3761980115247962,
      "grad_norm": 5.123339653015137,
      "learning_rate": 0.00019640169372396414,
      "loss": 1.3523,
      "step": 12600
    },
    {
      "epoch": 0.37769086077687875,
      "grad_norm": 3.4495439529418945,
      "learning_rate": 0.0001963874079280453,
      "loss": 1.3331,
      "step": 12650
    },
    {
      "epoch": 0.3791837100289613,
      "grad_norm": 4.613680362701416,
      "learning_rate": 0.00019637312213212647,
      "loss": 1.3707,
      "step": 12700
    },
    {
      "epoch": 0.3806765592810438,
      "grad_norm": 3.5819404125213623,
      "learning_rate": 0.00019635883633620764,
      "loss": 1.4023,
      "step": 12750
    },
    {
      "epoch": 0.3821694085331263,
      "grad_norm": 3.5075576305389404,
      "learning_rate": 0.00019634455054028883,
      "loss": 1.3815,
      "step": 12800
    },
    {
      "epoch": 0.38366225778520885,
      "grad_norm": 4.425256729125977,
      "learning_rate": 0.00019633026474436997,
      "loss": 1.3801,
      "step": 12850
    },
    {
      "epoch": 0.3851551070372914,
      "grad_norm": 3.5711112022399902,
      "learning_rate": 0.00019631597894845116,
      "loss": 1.3375,
      "step": 12900
    },
    {
      "epoch": 0.3866479562893739,
      "grad_norm": 5.728016376495361,
      "learning_rate": 0.0001963016931525323,
      "loss": 1.3914,
      "step": 12950
    },
    {
      "epoch": 0.3881408055414564,
      "grad_norm": 4.294504642486572,
      "learning_rate": 0.0001962874073566135,
      "loss": 1.4561,
      "step": 13000
    },
    {
      "epoch": 0.38963365479353895,
      "grad_norm": 5.329941749572754,
      "learning_rate": 0.00019627312156069465,
      "loss": 1.4018,
      "step": 13050
    },
    {
      "epoch": 0.3911265040456215,
      "grad_norm": 4.166362762451172,
      "learning_rate": 0.00019625883576477582,
      "loss": 1.4334,
      "step": 13100
    },
    {
      "epoch": 0.392619353297704,
      "grad_norm": 3.3922691345214844,
      "learning_rate": 0.00019624454996885698,
      "loss": 1.3956,
      "step": 13150
    },
    {
      "epoch": 0.3941122025497865,
      "grad_norm": 4.340898036956787,
      "learning_rate": 0.00019623026417293815,
      "loss": 1.3643,
      "step": 13200
    },
    {
      "epoch": 0.39560505180186906,
      "grad_norm": 6.623823165893555,
      "learning_rate": 0.0001962159783770193,
      "loss": 1.362,
      "step": 13250
    },
    {
      "epoch": 0.3970979010539516,
      "grad_norm": 4.490639686584473,
      "learning_rate": 0.00019620169258110048,
      "loss": 1.3742,
      "step": 13300
    },
    {
      "epoch": 0.3985907503060341,
      "grad_norm": 4.179808139801025,
      "learning_rate": 0.00019618740678518164,
      "loss": 1.4191,
      "step": 13350
    },
    {
      "epoch": 0.4000835995581166,
      "grad_norm": 5.623187065124512,
      "learning_rate": 0.0001961731209892628,
      "loss": 1.3793,
      "step": 13400
    },
    {
      "epoch": 0.40157644881019916,
      "grad_norm": 3.9650678634643555,
      "learning_rate": 0.00019615883519334397,
      "loss": 1.3917,
      "step": 13450
    },
    {
      "epoch": 0.40306929806228164,
      "grad_norm": 5.047702312469482,
      "learning_rate": 0.00019614454939742516,
      "loss": 1.4501,
      "step": 13500
    },
    {
      "epoch": 0.4045621473143642,
      "grad_norm": 3.46647310256958,
      "learning_rate": 0.0001961302636015063,
      "loss": 1.4091,
      "step": 13550
    },
    {
      "epoch": 0.4060549965664467,
      "grad_norm": 3.2203481197357178,
      "learning_rate": 0.0001961159778055875,
      "loss": 1.4003,
      "step": 13600
    },
    {
      "epoch": 0.40754784581852926,
      "grad_norm": 3.727679967880249,
      "learning_rate": 0.00019610169200966863,
      "loss": 1.4004,
      "step": 13650
    },
    {
      "epoch": 0.40904069507061175,
      "grad_norm": 4.469257831573486,
      "learning_rate": 0.0001960874062137498,
      "loss": 1.4061,
      "step": 13700
    },
    {
      "epoch": 0.4105335443226943,
      "grad_norm": 4.041538715362549,
      "learning_rate": 0.00019607312041783096,
      "loss": 1.411,
      "step": 13750
    },
    {
      "epoch": 0.41202639357477683,
      "grad_norm": 5.2691779136657715,
      "learning_rate": 0.00019605883462191212,
      "loss": 1.3396,
      "step": 13800
    },
    {
      "epoch": 0.41351924282685937,
      "grad_norm": 6.236726760864258,
      "learning_rate": 0.00019604454882599331,
      "loss": 1.4295,
      "step": 13850
    },
    {
      "epoch": 0.41501209207894185,
      "grad_norm": 6.703745365142822,
      "learning_rate": 0.00019603026303007445,
      "loss": 1.4105,
      "step": 13900
    },
    {
      "epoch": 0.4165049413310244,
      "grad_norm": 3.997664451599121,
      "learning_rate": 0.00019601597723415564,
      "loss": 1.347,
      "step": 13950
    },
    {
      "epoch": 0.41799779058310693,
      "grad_norm": 5.311407566070557,
      "learning_rate": 0.00019600169143823678,
      "loss": 1.4546,
      "step": 14000
    },
    {
      "epoch": 0.4194906398351894,
      "grad_norm": 4.0283098220825195,
      "learning_rate": 0.00019598740564231797,
      "loss": 1.4513,
      "step": 14050
    },
    {
      "epoch": 0.42098348908727196,
      "grad_norm": 7.345764636993408,
      "learning_rate": 0.0001959731198463991,
      "loss": 1.3832,
      "step": 14100
    },
    {
      "epoch": 0.4224763383393545,
      "grad_norm": 4.324542045593262,
      "learning_rate": 0.0001959588340504803,
      "loss": 1.3751,
      "step": 14150
    },
    {
      "epoch": 0.42396918759143704,
      "grad_norm": 3.8322675228118896,
      "learning_rate": 0.00019594454825456147,
      "loss": 1.344,
      "step": 14200
    },
    {
      "epoch": 0.4254620368435195,
      "grad_norm": 4.62548303604126,
      "learning_rate": 0.00019593026245864263,
      "loss": 1.4346,
      "step": 14250
    },
    {
      "epoch": 0.42695488609560206,
      "grad_norm": 4.585489273071289,
      "learning_rate": 0.0001959159766627238,
      "loss": 1.4145,
      "step": 14300
    },
    {
      "epoch": 0.4284477353476846,
      "grad_norm": 3.64227557182312,
      "learning_rate": 0.00019590169086680496,
      "loss": 1.367,
      "step": 14350
    },
    {
      "epoch": 0.42994058459976714,
      "grad_norm": 4.730580806732178,
      "learning_rate": 0.00019588740507088613,
      "loss": 1.3653,
      "step": 14400
    },
    {
      "epoch": 0.4314334338518496,
      "grad_norm": 4.4075398445129395,
      "learning_rate": 0.0001958731192749673,
      "loss": 1.4081,
      "step": 14450
    },
    {
      "epoch": 0.43292628310393216,
      "grad_norm": 3.0498785972595215,
      "learning_rate": 0.00019585883347904845,
      "loss": 1.343,
      "step": 14500
    },
    {
      "epoch": 0.4344191323560147,
      "grad_norm": 4.179199695587158,
      "learning_rate": 0.00019584454768312962,
      "loss": 1.3662,
      "step": 14550
    },
    {
      "epoch": 0.4359119816080972,
      "grad_norm": 4.1148786544799805,
      "learning_rate": 0.00019583026188721078,
      "loss": 1.3785,
      "step": 14600
    },
    {
      "epoch": 0.4374048308601797,
      "grad_norm": 4.012060165405273,
      "learning_rate": 0.00019581597609129198,
      "loss": 1.4056,
      "step": 14650
    },
    {
      "epoch": 0.43889768011226227,
      "grad_norm": 7.186342716217041,
      "learning_rate": 0.0001958016902953731,
      "loss": 1.3943,
      "step": 14700
    },
    {
      "epoch": 0.4403905293643448,
      "grad_norm": 3.951267957687378,
      "learning_rate": 0.0001957874044994543,
      "loss": 1.3976,
      "step": 14750
    },
    {
      "epoch": 0.4418833786164273,
      "grad_norm": 5.276801586151123,
      "learning_rate": 0.00019577311870353544,
      "loss": 1.4137,
      "step": 14800
    },
    {
      "epoch": 0.44337622786850983,
      "grad_norm": 4.107429504394531,
      "learning_rate": 0.00019575883290761663,
      "loss": 1.388,
      "step": 14850
    },
    {
      "epoch": 0.44486907712059237,
      "grad_norm": 4.274941444396973,
      "learning_rate": 0.00019574454711169777,
      "loss": 1.3446,
      "step": 14900
    },
    {
      "epoch": 0.4463619263726749,
      "grad_norm": 4.174200534820557,
      "learning_rate": 0.00019573026131577896,
      "loss": 1.3703,
      "step": 14950
    },
    {
      "epoch": 0.4478547756247574,
      "grad_norm": 3.867125988006592,
      "learning_rate": 0.00019571597551986013,
      "loss": 1.3745,
      "step": 15000
    },
    {
      "epoch": 0.44934762487683994,
      "grad_norm": 6.454402923583984,
      "learning_rate": 0.0001957016897239413,
      "loss": 1.3081,
      "step": 15050
    },
    {
      "epoch": 0.4508404741289225,
      "grad_norm": 3.8726885318756104,
      "learning_rate": 0.00019568740392802246,
      "loss": 1.3527,
      "step": 15100
    },
    {
      "epoch": 0.45233332338100496,
      "grad_norm": 3.6218361854553223,
      "learning_rate": 0.00019567311813210362,
      "loss": 1.4549,
      "step": 15150
    },
    {
      "epoch": 0.4538261726330875,
      "grad_norm": 5.2475361824035645,
      "learning_rate": 0.0001956588323361848,
      "loss": 1.361,
      "step": 15200
    },
    {
      "epoch": 0.45531902188517004,
      "grad_norm": 4.388748645782471,
      "learning_rate": 0.00019564454654026595,
      "loss": 1.3164,
      "step": 15250
    },
    {
      "epoch": 0.4568118711372526,
      "grad_norm": 4.97973108291626,
      "learning_rate": 0.00019563026074434712,
      "loss": 1.3755,
      "step": 15300
    },
    {
      "epoch": 0.45830472038933506,
      "grad_norm": 4.538138389587402,
      "learning_rate": 0.00019561597494842828,
      "loss": 1.4339,
      "step": 15350
    },
    {
      "epoch": 0.4597975696414176,
      "grad_norm": 4.389719009399414,
      "learning_rate": 0.00019560168915250945,
      "loss": 1.4163,
      "step": 15400
    },
    {
      "epoch": 0.46129041889350014,
      "grad_norm": 4.347919464111328,
      "learning_rate": 0.00019558740335659064,
      "loss": 1.3663,
      "step": 15450
    },
    {
      "epoch": 0.4627832681455827,
      "grad_norm": 4.820595741271973,
      "learning_rate": 0.00019557311756067177,
      "loss": 1.3441,
      "step": 15500
    },
    {
      "epoch": 0.46427611739766517,
      "grad_norm": 2.6965413093566895,
      "learning_rate": 0.00019555883176475297,
      "loss": 1.3657,
      "step": 15550
    },
    {
      "epoch": 0.4657689666497477,
      "grad_norm": 4.741116523742676,
      "learning_rate": 0.0001955445459688341,
      "loss": 1.4414,
      "step": 15600
    },
    {
      "epoch": 0.46726181590183025,
      "grad_norm": 3.9512829780578613,
      "learning_rate": 0.0001955302601729153,
      "loss": 1.354,
      "step": 15650
    },
    {
      "epoch": 0.46875466515391273,
      "grad_norm": 7.704863548278809,
      "learning_rate": 0.00019551597437699646,
      "loss": 1.4107,
      "step": 15700
    },
    {
      "epoch": 0.47024751440599527,
      "grad_norm": 3.502988338470459,
      "learning_rate": 0.00019550168858107763,
      "loss": 1.415,
      "step": 15750
    },
    {
      "epoch": 0.4717403636580778,
      "grad_norm": 4.246065139770508,
      "learning_rate": 0.0001954874027851588,
      "loss": 1.3414,
      "step": 15800
    },
    {
      "epoch": 0.47323321291016035,
      "grad_norm": 3.4039735794067383,
      "learning_rate": 0.00019547311698923995,
      "loss": 1.3303,
      "step": 15850
    },
    {
      "epoch": 0.47472606216224283,
      "grad_norm": 3.279521942138672,
      "learning_rate": 0.00019545883119332112,
      "loss": 1.3447,
      "step": 15900
    },
    {
      "epoch": 0.4762189114143254,
      "grad_norm": 2.9335134029388428,
      "learning_rate": 0.00019544454539740228,
      "loss": 1.3998,
      "step": 15950
    },
    {
      "epoch": 0.4777117606664079,
      "grad_norm": 3.317011833190918,
      "learning_rate": 0.00019543025960148345,
      "loss": 1.3506,
      "step": 16000
    },
    {
      "epoch": 0.47920460991849045,
      "grad_norm": 3.5980935096740723,
      "learning_rate": 0.0001954159738055646,
      "loss": 1.3622,
      "step": 16050
    },
    {
      "epoch": 0.48069745917057294,
      "grad_norm": 4.726743698120117,
      "learning_rate": 0.00019540168800964578,
      "loss": 1.3414,
      "step": 16100
    },
    {
      "epoch": 0.4821903084226555,
      "grad_norm": 5.129758358001709,
      "learning_rate": 0.00019538740221372694,
      "loss": 1.3873,
      "step": 16150
    },
    {
      "epoch": 0.483683157674738,
      "grad_norm": 5.122271537780762,
      "learning_rate": 0.0001953731164178081,
      "loss": 1.428,
      "step": 16200
    },
    {
      "epoch": 0.4851760069268205,
      "grad_norm": 3.359868049621582,
      "learning_rate": 0.0001953588306218893,
      "loss": 1.3111,
      "step": 16250
    },
    {
      "epoch": 0.48666885617890304,
      "grad_norm": 5.066514492034912,
      "learning_rate": 0.00019534454482597044,
      "loss": 1.2961,
      "step": 16300
    },
    {
      "epoch": 0.4881617054309856,
      "grad_norm": 4.902595520019531,
      "learning_rate": 0.0001953302590300516,
      "loss": 1.3682,
      "step": 16350
    },
    {
      "epoch": 0.4896545546830681,
      "grad_norm": 5.0537028312683105,
      "learning_rate": 0.00019531597323413277,
      "loss": 1.315,
      "step": 16400
    },
    {
      "epoch": 0.4911474039351506,
      "grad_norm": 3.7002792358398438,
      "learning_rate": 0.00019530168743821393,
      "loss": 1.3441,
      "step": 16450
    },
    {
      "epoch": 0.49264025318723315,
      "grad_norm": 4.845950603485107,
      "learning_rate": 0.00019528740164229512,
      "loss": 1.3887,
      "step": 16500
    },
    {
      "epoch": 0.4941331024393157,
      "grad_norm": 4.933434963226318,
      "learning_rate": 0.00019527311584637626,
      "loss": 1.3865,
      "step": 16550
    },
    {
      "epoch": 0.4956259516913982,
      "grad_norm": 3.8103625774383545,
      "learning_rate": 0.00019525883005045745,
      "loss": 1.3757,
      "step": 16600
    },
    {
      "epoch": 0.4971188009434807,
      "grad_norm": 4.501999855041504,
      "learning_rate": 0.0001952445442545386,
      "loss": 1.4482,
      "step": 16650
    },
    {
      "epoch": 0.49861165019556325,
      "grad_norm": 5.600002765655518,
      "learning_rate": 0.00019523025845861978,
      "loss": 1.4209,
      "step": 16700
    },
    {
      "epoch": 0.5001044994476458,
      "grad_norm": 5.138682842254639,
      "learning_rate": 0.00019521597266270092,
      "loss": 1.4287,
      "step": 16750
    },
    {
      "epoch": 0.5015973486997283,
      "grad_norm": 5.575449466705322,
      "learning_rate": 0.0001952016868667821,
      "loss": 1.3992,
      "step": 16800
    },
    {
      "epoch": 0.5030901979518109,
      "grad_norm": 3.6443893909454346,
      "learning_rate": 0.00019518740107086327,
      "loss": 1.3661,
      "step": 16850
    },
    {
      "epoch": 0.5045830472038934,
      "grad_norm": 3.904905319213867,
      "learning_rate": 0.00019517311527494444,
      "loss": 1.4448,
      "step": 16900
    },
    {
      "epoch": 0.5060758964559758,
      "grad_norm": 4.380904197692871,
      "learning_rate": 0.0001951588294790256,
      "loss": 1.4076,
      "step": 16950
    },
    {
      "epoch": 0.5075687457080584,
      "grad_norm": 3.5924415588378906,
      "learning_rate": 0.00019514454368310677,
      "loss": 1.3833,
      "step": 17000
    },
    {
      "epoch": 0.5090615949601409,
      "grad_norm": 3.7007193565368652,
      "learning_rate": 0.00019513025788718793,
      "loss": 1.391,
      "step": 17050
    },
    {
      "epoch": 0.5105544442122234,
      "grad_norm": 3.9879095554351807,
      "learning_rate": 0.0001951159720912691,
      "loss": 1.3419,
      "step": 17100
    },
    {
      "epoch": 0.512047293464306,
      "grad_norm": 5.663998126983643,
      "learning_rate": 0.00019510168629535026,
      "loss": 1.3481,
      "step": 17150
    },
    {
      "epoch": 0.5135401427163885,
      "grad_norm": 3.9803707599639893,
      "learning_rate": 0.00019508740049943143,
      "loss": 1.4034,
      "step": 17200
    },
    {
      "epoch": 0.5150329919684711,
      "grad_norm": 3.718477725982666,
      "learning_rate": 0.0001950731147035126,
      "loss": 1.406,
      "step": 17250
    },
    {
      "epoch": 0.5165258412205536,
      "grad_norm": 4.864751815795898,
      "learning_rate": 0.00019505882890759378,
      "loss": 1.375,
      "step": 17300
    },
    {
      "epoch": 0.518018690472636,
      "grad_norm": 3.697645664215088,
      "learning_rate": 0.00019504454311167492,
      "loss": 1.4283,
      "step": 17350
    },
    {
      "epoch": 0.5195115397247186,
      "grad_norm": 4.063074111938477,
      "learning_rate": 0.0001950302573157561,
      "loss": 1.378,
      "step": 17400
    },
    {
      "epoch": 0.5210043889768011,
      "grad_norm": 4.223004341125488,
      "learning_rate": 0.00019501597151983725,
      "loss": 1.3789,
      "step": 17450
    },
    {
      "epoch": 0.5224972382288836,
      "grad_norm": 3.329366683959961,
      "learning_rate": 0.00019500168572391844,
      "loss": 1.346,
      "step": 17500
    },
    {
      "epoch": 0.5239900874809662,
      "grad_norm": 4.774710178375244,
      "learning_rate": 0.00019498739992799958,
      "loss": 1.3895,
      "step": 17550
    },
    {
      "epoch": 0.5254829367330487,
      "grad_norm": 6.2145490646362305,
      "learning_rate": 0.00019497311413208077,
      "loss": 1.3715,
      "step": 17600
    },
    {
      "epoch": 0.5269757859851312,
      "grad_norm": 3.9069626331329346,
      "learning_rate": 0.00019495882833616194,
      "loss": 1.4572,
      "step": 17650
    },
    {
      "epoch": 0.5284686352372138,
      "grad_norm": 3.347576141357422,
      "learning_rate": 0.0001949445425402431,
      "loss": 1.359,
      "step": 17700
    },
    {
      "epoch": 0.5299614844892963,
      "grad_norm": 5.305202484130859,
      "learning_rate": 0.00019493025674432427,
      "loss": 1.4038,
      "step": 17750
    },
    {
      "epoch": 0.5314543337413788,
      "grad_norm": 3.865619659423828,
      "learning_rate": 0.00019491597094840543,
      "loss": 1.3913,
      "step": 17800
    },
    {
      "epoch": 0.5329471829934613,
      "grad_norm": 4.791336536407471,
      "learning_rate": 0.0001949016851524866,
      "loss": 1.3852,
      "step": 17850
    },
    {
      "epoch": 0.5344400322455438,
      "grad_norm": 3.7827060222625732,
      "learning_rate": 0.00019488739935656776,
      "loss": 1.3108,
      "step": 17900
    },
    {
      "epoch": 0.5359328814976264,
      "grad_norm": 4.945117473602295,
      "learning_rate": 0.00019487311356064892,
      "loss": 1.346,
      "step": 17950
    },
    {
      "epoch": 0.5374257307497089,
      "grad_norm": 4.561169147491455,
      "learning_rate": 0.0001948588277647301,
      "loss": 1.3904,
      "step": 18000
    },
    {
      "epoch": 0.5389185800017914,
      "grad_norm": 4.608798027038574,
      "learning_rate": 0.00019484454196881125,
      "loss": 1.4133,
      "step": 18050
    },
    {
      "epoch": 0.540411429253874,
      "grad_norm": 4.303143501281738,
      "learning_rate": 0.00019483025617289245,
      "loss": 1.3994,
      "step": 18100
    },
    {
      "epoch": 0.5419042785059565,
      "grad_norm": 5.815835952758789,
      "learning_rate": 0.00019481597037697358,
      "loss": 1.392,
      "step": 18150
    },
    {
      "epoch": 0.543397127758039,
      "grad_norm": 5.349491596221924,
      "learning_rate": 0.00019480168458105477,
      "loss": 1.418,
      "step": 18200
    },
    {
      "epoch": 0.5448899770101215,
      "grad_norm": 3.7407824993133545,
      "learning_rate": 0.0001947873987851359,
      "loss": 1.3747,
      "step": 18250
    },
    {
      "epoch": 0.546382826262204,
      "grad_norm": 5.2810163497924805,
      "learning_rate": 0.0001947731129892171,
      "loss": 1.4023,
      "step": 18300
    },
    {
      "epoch": 0.5478756755142866,
      "grad_norm": 4.417948246002197,
      "learning_rate": 0.00019475882719329824,
      "loss": 1.4005,
      "step": 18350
    },
    {
      "epoch": 0.5493685247663691,
      "grad_norm": 5.287749290466309,
      "learning_rate": 0.00019474454139737943,
      "loss": 1.4211,
      "step": 18400
    },
    {
      "epoch": 0.5508613740184516,
      "grad_norm": 4.0996809005737305,
      "learning_rate": 0.0001947302556014606,
      "loss": 1.3559,
      "step": 18450
    },
    {
      "epoch": 0.5523542232705342,
      "grad_norm": 5.229327201843262,
      "learning_rate": 0.00019471596980554176,
      "loss": 1.4549,
      "step": 18500
    },
    {
      "epoch": 0.5538470725226167,
      "grad_norm": 4.409546852111816,
      "learning_rate": 0.00019470168400962293,
      "loss": 1.3683,
      "step": 18550
    },
    {
      "epoch": 0.5553399217746992,
      "grad_norm": 5.4077229499816895,
      "learning_rate": 0.0001946873982137041,
      "loss": 1.398,
      "step": 18600
    },
    {
      "epoch": 0.5568327710267817,
      "grad_norm": 5.208966255187988,
      "learning_rate": 0.00019467311241778526,
      "loss": 1.3276,
      "step": 18650
    },
    {
      "epoch": 0.5583256202788642,
      "grad_norm": 4.8162617683410645,
      "learning_rate": 0.00019465882662186642,
      "loss": 1.3314,
      "step": 18700
    },
    {
      "epoch": 0.5598184695309467,
      "grad_norm": 5.04697322845459,
      "learning_rate": 0.00019464454082594759,
      "loss": 1.3883,
      "step": 18750
    },
    {
      "epoch": 0.5613113187830293,
      "grad_norm": 4.038108825683594,
      "learning_rate": 0.00019463025503002875,
      "loss": 1.4377,
      "step": 18800
    },
    {
      "epoch": 0.5628041680351118,
      "grad_norm": 4.910576820373535,
      "learning_rate": 0.00019461596923410992,
      "loss": 1.3713,
      "step": 18850
    },
    {
      "epoch": 0.5642970172871944,
      "grad_norm": 5.3433756828308105,
      "learning_rate": 0.00019460168343819108,
      "loss": 1.4359,
      "step": 18900
    },
    {
      "epoch": 0.5657898665392769,
      "grad_norm": 3.9515552520751953,
      "learning_rate": 0.00019458739764227224,
      "loss": 1.3608,
      "step": 18950
    },
    {
      "epoch": 0.5672827157913594,
      "grad_norm": 4.515705108642578,
      "learning_rate": 0.0001945731118463534,
      "loss": 1.3274,
      "step": 19000
    },
    {
      "epoch": 0.568775565043442,
      "grad_norm": 4.434077262878418,
      "learning_rate": 0.00019455882605043457,
      "loss": 1.3681,
      "step": 19050
    },
    {
      "epoch": 0.5702684142955244,
      "grad_norm": 4.534008979797363,
      "learning_rate": 0.00019454454025451574,
      "loss": 1.3863,
      "step": 19100
    },
    {
      "epoch": 0.5717612635476069,
      "grad_norm": 4.200322151184082,
      "learning_rate": 0.00019453025445859693,
      "loss": 1.3743,
      "step": 19150
    },
    {
      "epoch": 0.5732541127996895,
      "grad_norm": 5.686845779418945,
      "learning_rate": 0.00019451596866267807,
      "loss": 1.3225,
      "step": 19200
    },
    {
      "epoch": 0.574746962051772,
      "grad_norm": 7.821211814880371,
      "learning_rate": 0.00019450168286675926,
      "loss": 1.3775,
      "step": 19250
    },
    {
      "epoch": 0.5762398113038545,
      "grad_norm": 5.200834274291992,
      "learning_rate": 0.0001944873970708404,
      "loss": 1.3423,
      "step": 19300
    },
    {
      "epoch": 0.5777326605559371,
      "grad_norm": 5.26302433013916,
      "learning_rate": 0.0001944731112749216,
      "loss": 1.3813,
      "step": 19350
    },
    {
      "epoch": 0.5792255098080196,
      "grad_norm": 3.3207719326019287,
      "learning_rate": 0.00019445882547900273,
      "loss": 1.3922,
      "step": 19400
    },
    {
      "epoch": 0.5807183590601022,
      "grad_norm": 4.619020938873291,
      "learning_rate": 0.00019444453968308392,
      "loss": 1.3533,
      "step": 19450
    },
    {
      "epoch": 0.5822112083121846,
      "grad_norm": 5.780002593994141,
      "learning_rate": 0.00019443025388716508,
      "loss": 1.4035,
      "step": 19500
    },
    {
      "epoch": 0.5837040575642671,
      "grad_norm": 4.961215496063232,
      "learning_rate": 0.00019441596809124625,
      "loss": 1.3687,
      "step": 19550
    },
    {
      "epoch": 0.5851969068163497,
      "grad_norm": 4.50115442276001,
      "learning_rate": 0.0001944016822953274,
      "loss": 1.342,
      "step": 19600
    },
    {
      "epoch": 0.5866897560684322,
      "grad_norm": 3.9477944374084473,
      "learning_rate": 0.00019438739649940858,
      "loss": 1.394,
      "step": 19650
    },
    {
      "epoch": 0.5881826053205147,
      "grad_norm": 3.7466814517974854,
      "learning_rate": 0.00019437311070348974,
      "loss": 1.3414,
      "step": 19700
    },
    {
      "epoch": 0.5896754545725973,
      "grad_norm": 4.382058143615723,
      "learning_rate": 0.0001943588249075709,
      "loss": 1.3669,
      "step": 19750
    },
    {
      "epoch": 0.5911683038246798,
      "grad_norm": 3.7016665935516357,
      "learning_rate": 0.00019434453911165207,
      "loss": 1.4548,
      "step": 19800
    },
    {
      "epoch": 0.5926611530767623,
      "grad_norm": 4.4738030433654785,
      "learning_rate": 0.00019433025331573324,
      "loss": 1.4273,
      "step": 19850
    },
    {
      "epoch": 0.5941540023288449,
      "grad_norm": 5.2445454597473145,
      "learning_rate": 0.0001943159675198144,
      "loss": 1.3746,
      "step": 19900
    },
    {
      "epoch": 0.5956468515809273,
      "grad_norm": 3.766219139099121,
      "learning_rate": 0.0001943016817238956,
      "loss": 1.4391,
      "step": 19950
    },
    {
      "epoch": 0.5971397008330099,
      "grad_norm": 6.310808181762695,
      "learning_rate": 0.00019428739592797673,
      "loss": 1.3316,
      "step": 20000
    },
    {
      "epoch": 0.5986325500850924,
      "grad_norm": 4.055521488189697,
      "learning_rate": 0.00019427311013205792,
      "loss": 1.3784,
      "step": 20050
    },
    {
      "epoch": 0.6001253993371749,
      "grad_norm": 4.933177471160889,
      "learning_rate": 0.00019425882433613906,
      "loss": 1.3352,
      "step": 20100
    },
    {
      "epoch": 0.6016182485892575,
      "grad_norm": 3.8867061138153076,
      "learning_rate": 0.00019424453854022025,
      "loss": 1.4169,
      "step": 20150
    },
    {
      "epoch": 0.60311109784134,
      "grad_norm": 3.364475727081299,
      "learning_rate": 0.0001942302527443014,
      "loss": 1.3767,
      "step": 20200
    },
    {
      "epoch": 0.6046039470934225,
      "grad_norm": 3.48152232170105,
      "learning_rate": 0.00019421596694838258,
      "loss": 1.3659,
      "step": 20250
    },
    {
      "epoch": 0.6060967963455051,
      "grad_norm": 3.3658649921417236,
      "learning_rate": 0.00019420168115246374,
      "loss": 1.3745,
      "step": 20300
    },
    {
      "epoch": 0.6075896455975875,
      "grad_norm": 4.441917896270752,
      "learning_rate": 0.0001941873953565449,
      "loss": 1.3827,
      "step": 20350
    },
    {
      "epoch": 0.60908249484967,
      "grad_norm": 5.014800548553467,
      "learning_rate": 0.00019417310956062607,
      "loss": 1.3792,
      "step": 20400
    },
    {
      "epoch": 0.6105753441017526,
      "grad_norm": 5.472316265106201,
      "learning_rate": 0.00019415882376470724,
      "loss": 1.394,
      "step": 20450
    },
    {
      "epoch": 0.6120681933538351,
      "grad_norm": 5.35073184967041,
      "learning_rate": 0.0001941445379687884,
      "loss": 1.401,
      "step": 20500
    },
    {
      "epoch": 0.6135610426059177,
      "grad_norm": 4.284445762634277,
      "learning_rate": 0.00019413025217286957,
      "loss": 1.4354,
      "step": 20550
    },
    {
      "epoch": 0.6150538918580002,
      "grad_norm": 3.561774492263794,
      "learning_rate": 0.00019411596637695073,
      "loss": 1.3295,
      "step": 20600
    },
    {
      "epoch": 0.6165467411100827,
      "grad_norm": 4.479186534881592,
      "learning_rate": 0.0001941016805810319,
      "loss": 1.3629,
      "step": 20650
    },
    {
      "epoch": 0.6180395903621653,
      "grad_norm": 4.186618804931641,
      "learning_rate": 0.00019408739478511306,
      "loss": 1.3399,
      "step": 20700
    },
    {
      "epoch": 0.6195324396142478,
      "grad_norm": 3.589655637741089,
      "learning_rate": 0.00019407310898919425,
      "loss": 1.3516,
      "step": 20750
    },
    {
      "epoch": 0.6210252888663302,
      "grad_norm": 4.330646514892578,
      "learning_rate": 0.0001940588231932754,
      "loss": 1.3148,
      "step": 20800
    },
    {
      "epoch": 0.6225181381184128,
      "grad_norm": 6.324933052062988,
      "learning_rate": 0.00019404453739735658,
      "loss": 1.331,
      "step": 20850
    },
    {
      "epoch": 0.6240109873704953,
      "grad_norm": 4.652800559997559,
      "learning_rate": 0.00019403025160143772,
      "loss": 1.3604,
      "step": 20900
    },
    {
      "epoch": 0.6255038366225778,
      "grad_norm": 4.831106662750244,
      "learning_rate": 0.0001940159658055189,
      "loss": 1.3322,
      "step": 20950
    },
    {
      "epoch": 0.6269966858746604,
      "grad_norm": 6.0938920974731445,
      "learning_rate": 0.00019400168000960005,
      "loss": 1.4106,
      "step": 21000
    },
    {
      "epoch": 0.6284895351267429,
      "grad_norm": 4.424108028411865,
      "learning_rate": 0.00019398739421368124,
      "loss": 1.4001,
      "step": 21050
    },
    {
      "epoch": 0.6299823843788255,
      "grad_norm": 4.329803466796875,
      "learning_rate": 0.0001939731084177624,
      "loss": 1.4048,
      "step": 21100
    },
    {
      "epoch": 0.631475233630908,
      "grad_norm": 3.915818929672241,
      "learning_rate": 0.00019395882262184357,
      "loss": 1.3619,
      "step": 21150
    },
    {
      "epoch": 0.6329680828829904,
      "grad_norm": 3.9562571048736572,
      "learning_rate": 0.00019394453682592474,
      "loss": 1.3636,
      "step": 21200
    },
    {
      "epoch": 0.634460932135073,
      "grad_norm": 5.229249954223633,
      "learning_rate": 0.0001939302510300059,
      "loss": 1.3502,
      "step": 21250
    },
    {
      "epoch": 0.6359537813871555,
      "grad_norm": 4.149145126342773,
      "learning_rate": 0.00019391596523408706,
      "loss": 1.3603,
      "step": 21300
    },
    {
      "epoch": 0.637446630639238,
      "grad_norm": 4.068868637084961,
      "learning_rate": 0.00019390167943816823,
      "loss": 1.3945,
      "step": 21350
    },
    {
      "epoch": 0.6389394798913206,
      "grad_norm": 3.6808931827545166,
      "learning_rate": 0.0001938873936422494,
      "loss": 1.4015,
      "step": 21400
    },
    {
      "epoch": 0.6404323291434031,
      "grad_norm": 4.391795635223389,
      "learning_rate": 0.00019387310784633056,
      "loss": 1.3428,
      "step": 21450
    },
    {
      "epoch": 0.6419251783954856,
      "grad_norm": 5.109655380249023,
      "learning_rate": 0.00019385882205041172,
      "loss": 1.3874,
      "step": 21500
    },
    {
      "epoch": 0.6434180276475682,
      "grad_norm": 4.560513496398926,
      "learning_rate": 0.0001938445362544929,
      "loss": 1.3833,
      "step": 21550
    },
    {
      "epoch": 0.6449108768996507,
      "grad_norm": 3.6108620166778564,
      "learning_rate": 0.00019383025045857405,
      "loss": 1.3981,
      "step": 21600
    },
    {
      "epoch": 0.6464037261517332,
      "grad_norm": 4.100367546081543,
      "learning_rate": 0.00019381596466265522,
      "loss": 1.3139,
      "step": 21650
    },
    {
      "epoch": 0.6478965754038157,
      "grad_norm": 3.1513540744781494,
      "learning_rate": 0.00019380167886673638,
      "loss": 1.3359,
      "step": 21700
    },
    {
      "epoch": 0.6493894246558982,
      "grad_norm": 4.793807506561279,
      "learning_rate": 0.00019378739307081755,
      "loss": 1.4025,
      "step": 21750
    },
    {
      "epoch": 0.6508822739079808,
      "grad_norm": 4.403114318847656,
      "learning_rate": 0.0001937731072748987,
      "loss": 1.349,
      "step": 21800
    },
    {
      "epoch": 0.6523751231600633,
      "grad_norm": 4.445423603057861,
      "learning_rate": 0.00019375882147897988,
      "loss": 1.4641,
      "step": 21850
    },
    {
      "epoch": 0.6538679724121458,
      "grad_norm": 5.721547603607178,
      "learning_rate": 0.00019374453568306107,
      "loss": 1.3451,
      "step": 21900
    },
    {
      "epoch": 0.6553608216642284,
      "grad_norm": 4.411465644836426,
      "learning_rate": 0.0001937302498871422,
      "loss": 1.3893,
      "step": 21950
    },
    {
      "epoch": 0.6568536709163109,
      "grad_norm": 3.181412696838379,
      "learning_rate": 0.0001937159640912234,
      "loss": 1.385,
      "step": 22000
    },
    {
      "epoch": 0.6583465201683933,
      "grad_norm": 7.374556541442871,
      "learning_rate": 0.00019370167829530453,
      "loss": 1.4051,
      "step": 22050
    },
    {
      "epoch": 0.6598393694204759,
      "grad_norm": 4.452084064483643,
      "learning_rate": 0.00019368739249938573,
      "loss": 1.3389,
      "step": 22100
    },
    {
      "epoch": 0.6613322186725584,
      "grad_norm": 5.60634708404541,
      "learning_rate": 0.0001936731067034669,
      "loss": 1.3924,
      "step": 22150
    },
    {
      "epoch": 0.662825067924641,
      "grad_norm": 4.35741662979126,
      "learning_rate": 0.00019365882090754806,
      "loss": 1.3957,
      "step": 22200
    },
    {
      "epoch": 0.6643179171767235,
      "grad_norm": 4.173916816711426,
      "learning_rate": 0.00019364453511162922,
      "loss": 1.3669,
      "step": 22250
    },
    {
      "epoch": 0.665810766428806,
      "grad_norm": 3.8707377910614014,
      "learning_rate": 0.00019363024931571038,
      "loss": 1.3478,
      "step": 22300
    },
    {
      "epoch": 0.6673036156808886,
      "grad_norm": 4.765937328338623,
      "learning_rate": 0.00019361596351979155,
      "loss": 1.4007,
      "step": 22350
    },
    {
      "epoch": 0.6687964649329711,
      "grad_norm": 4.315809726715088,
      "learning_rate": 0.00019360167772387271,
      "loss": 1.3622,
      "step": 22400
    },
    {
      "epoch": 0.6702893141850536,
      "grad_norm": 5.977219104766846,
      "learning_rate": 0.00019358739192795388,
      "loss": 1.3836,
      "step": 22450
    },
    {
      "epoch": 0.6717821634371361,
      "grad_norm": 4.420149803161621,
      "learning_rate": 0.00019357310613203504,
      "loss": 1.3488,
      "step": 22500
    },
    {
      "epoch": 0.6732750126892186,
      "grad_norm": 3.9537293910980225,
      "learning_rate": 0.0001935588203361162,
      "loss": 1.3892,
      "step": 22550
    },
    {
      "epoch": 0.6747678619413011,
      "grad_norm": 5.468355655670166,
      "learning_rate": 0.0001935445345401974,
      "loss": 1.4475,
      "step": 22600
    },
    {
      "epoch": 0.6762607111933837,
      "grad_norm": 4.3148674964904785,
      "learning_rate": 0.00019353024874427854,
      "loss": 1.3866,
      "step": 22650
    },
    {
      "epoch": 0.6777535604454662,
      "grad_norm": 4.004809379577637,
      "learning_rate": 0.00019351596294835973,
      "loss": 1.3811,
      "step": 22700
    },
    {
      "epoch": 0.6792464096975488,
      "grad_norm": 4.404988765716553,
      "learning_rate": 0.00019350167715244087,
      "loss": 1.3648,
      "step": 22750
    },
    {
      "epoch": 0.6807392589496313,
      "grad_norm": 5.115052223205566,
      "learning_rate": 0.00019348739135652206,
      "loss": 1.3523,
      "step": 22800
    },
    {
      "epoch": 0.6822321082017138,
      "grad_norm": 5.083719730377197,
      "learning_rate": 0.0001934731055606032,
      "loss": 1.3978,
      "step": 22850
    },
    {
      "epoch": 0.6837249574537964,
      "grad_norm": 4.038282871246338,
      "learning_rate": 0.0001934588197646844,
      "loss": 1.2788,
      "step": 22900
    },
    {
      "epoch": 0.6852178067058788,
      "grad_norm": 5.438407897949219,
      "learning_rate": 0.00019344453396876555,
      "loss": 1.4181,
      "step": 22950
    },
    {
      "epoch": 0.6867106559579613,
      "grad_norm": 5.381191730499268,
      "learning_rate": 0.00019343024817284672,
      "loss": 1.4145,
      "step": 23000
    },
    {
      "epoch": 0.6882035052100439,
      "grad_norm": 3.183706283569336,
      "learning_rate": 0.00019341596237692788,
      "loss": 1.3271,
      "step": 23050
    },
    {
      "epoch": 0.6896963544621264,
      "grad_norm": 4.063404083251953,
      "learning_rate": 0.00019340167658100905,
      "loss": 1.3717,
      "step": 23100
    },
    {
      "epoch": 0.6911892037142089,
      "grad_norm": 5.677481651306152,
      "learning_rate": 0.0001933873907850902,
      "loss": 1.344,
      "step": 23150
    },
    {
      "epoch": 0.6926820529662915,
      "grad_norm": 5.376470565795898,
      "learning_rate": 0.00019337310498917138,
      "loss": 1.295,
      "step": 23200
    },
    {
      "epoch": 0.694174902218374,
      "grad_norm": 3.8844990730285645,
      "learning_rate": 0.00019335881919325254,
      "loss": 1.381,
      "step": 23250
    },
    {
      "epoch": 0.6956677514704566,
      "grad_norm": 5.285017013549805,
      "learning_rate": 0.0001933445333973337,
      "loss": 1.3505,
      "step": 23300
    },
    {
      "epoch": 0.697160600722539,
      "grad_norm": 4.133642196655273,
      "learning_rate": 0.00019333024760141487,
      "loss": 1.3604,
      "step": 23350
    },
    {
      "epoch": 0.6986534499746215,
      "grad_norm": 5.037447929382324,
      "learning_rate": 0.00019331596180549606,
      "loss": 1.3209,
      "step": 23400
    },
    {
      "epoch": 0.7001462992267041,
      "grad_norm": 4.916257858276367,
      "learning_rate": 0.0001933016760095772,
      "loss": 1.4084,
      "step": 23450
    },
    {
      "epoch": 0.7016391484787866,
      "grad_norm": 3.469505786895752,
      "learning_rate": 0.0001932873902136584,
      "loss": 1.3705,
      "step": 23500
    },
    {
      "epoch": 0.7031319977308691,
      "grad_norm": 3.624896287918091,
      "learning_rate": 0.00019327310441773953,
      "loss": 1.3507,
      "step": 23550
    },
    {
      "epoch": 0.7046248469829517,
      "grad_norm": 4.352174758911133,
      "learning_rate": 0.00019325881862182072,
      "loss": 1.3184,
      "step": 23600
    },
    {
      "epoch": 0.7061176962350342,
      "grad_norm": 6.4549150466918945,
      "learning_rate": 0.00019324453282590186,
      "loss": 1.4161,
      "step": 23650
    },
    {
      "epoch": 0.7076105454871167,
      "grad_norm": 6.221735000610352,
      "learning_rate": 0.00019323024702998305,
      "loss": 1.3425,
      "step": 23700
    },
    {
      "epoch": 0.7091033947391993,
      "grad_norm": 4.285811901092529,
      "learning_rate": 0.00019321596123406421,
      "loss": 1.3418,
      "step": 23750
    },
    {
      "epoch": 0.7105962439912817,
      "grad_norm": 4.533527374267578,
      "learning_rate": 0.00019320167543814538,
      "loss": 1.3723,
      "step": 23800
    },
    {
      "epoch": 0.7120890932433643,
      "grad_norm": 3.838109016418457,
      "learning_rate": 0.00019318738964222654,
      "loss": 1.3624,
      "step": 23850
    },
    {
      "epoch": 0.7135819424954468,
      "grad_norm": 3.8401126861572266,
      "learning_rate": 0.0001931731038463077,
      "loss": 1.4239,
      "step": 23900
    },
    {
      "epoch": 0.7150747917475293,
      "grad_norm": 5.9049787521362305,
      "learning_rate": 0.00019315881805038887,
      "loss": 1.3642,
      "step": 23950
    },
    {
      "epoch": 0.7165676409996119,
      "grad_norm": 5.397033214569092,
      "learning_rate": 0.00019314453225447004,
      "loss": 1.3298,
      "step": 24000
    },
    {
      "epoch": 0.7180604902516944,
      "grad_norm": 5.649387359619141,
      "learning_rate": 0.0001931302464585512,
      "loss": 1.4102,
      "step": 24050
    },
    {
      "epoch": 0.7195533395037769,
      "grad_norm": 5.697938919067383,
      "learning_rate": 0.00019311596066263237,
      "loss": 1.3765,
      "step": 24100
    },
    {
      "epoch": 0.7210461887558595,
      "grad_norm": 4.244998455047607,
      "learning_rate": 0.00019310167486671353,
      "loss": 1.3471,
      "step": 24150
    },
    {
      "epoch": 0.722539038007942,
      "grad_norm": 4.579226493835449,
      "learning_rate": 0.0001930873890707947,
      "loss": 1.3897,
      "step": 24200
    },
    {
      "epoch": 0.7240318872600244,
      "grad_norm": 8.80657958984375,
      "learning_rate": 0.00019307310327487586,
      "loss": 1.3628,
      "step": 24250
    },
    {
      "epoch": 0.725524736512107,
      "grad_norm": 5.42709493637085,
      "learning_rate": 0.00019305881747895703,
      "loss": 1.3587,
      "step": 24300
    },
    {
      "epoch": 0.7270175857641895,
      "grad_norm": 4.2680840492248535,
      "learning_rate": 0.0001930445316830382,
      "loss": 1.3414,
      "step": 24350
    },
    {
      "epoch": 0.7285104350162721,
      "grad_norm": 5.641107559204102,
      "learning_rate": 0.00019303024588711935,
      "loss": 1.4095,
      "step": 24400
    },
    {
      "epoch": 0.7300032842683546,
      "grad_norm": 5.29530143737793,
      "learning_rate": 0.00019301596009120052,
      "loss": 1.389,
      "step": 24450
    },
    {
      "epoch": 0.7314961335204371,
      "grad_norm": 4.110136032104492,
      "learning_rate": 0.00019300167429528168,
      "loss": 1.4641,
      "step": 24500
    },
    {
      "epoch": 0.7329889827725197,
      "grad_norm": 4.390302658081055,
      "learning_rate": 0.00019298738849936288,
      "loss": 1.4028,
      "step": 24550
    },
    {
      "epoch": 0.7344818320246022,
      "grad_norm": 3.717409372329712,
      "learning_rate": 0.000192973102703444,
      "loss": 1.436,
      "step": 24600
    },
    {
      "epoch": 0.7359746812766846,
      "grad_norm": 5.399808406829834,
      "learning_rate": 0.0001929588169075252,
      "loss": 1.2887,
      "step": 24650
    },
    {
      "epoch": 0.7374675305287672,
      "grad_norm": 5.027936935424805,
      "learning_rate": 0.00019294453111160634,
      "loss": 1.3591,
      "step": 24700
    },
    {
      "epoch": 0.7389603797808497,
      "grad_norm": 5.12235689163208,
      "learning_rate": 0.00019293024531568753,
      "loss": 1.4305,
      "step": 24750
    },
    {
      "epoch": 0.7404532290329322,
      "grad_norm": 4.445746898651123,
      "learning_rate": 0.0001929159595197687,
      "loss": 1.3609,
      "step": 24800
    },
    {
      "epoch": 0.7419460782850148,
      "grad_norm": 4.590288162231445,
      "learning_rate": 0.00019290167372384986,
      "loss": 1.3811,
      "step": 24850
    },
    {
      "epoch": 0.7434389275370973,
      "grad_norm": 3.841573476791382,
      "learning_rate": 0.00019288738792793103,
      "loss": 1.3632,
      "step": 24900
    },
    {
      "epoch": 0.7449317767891799,
      "grad_norm": 5.2155327796936035,
      "learning_rate": 0.0001928731021320122,
      "loss": 1.4422,
      "step": 24950
    },
    {
      "epoch": 0.7464246260412624,
      "grad_norm": 4.942720413208008,
      "learning_rate": 0.00019285881633609336,
      "loss": 1.3916,
      "step": 25000
    },
    {
      "epoch": 0.7479174752933448,
      "grad_norm": 5.292357921600342,
      "learning_rate": 0.00019284453054017452,
      "loss": 1.3048,
      "step": 25050
    },
    {
      "epoch": 0.7494103245454274,
      "grad_norm": 6.690375804901123,
      "learning_rate": 0.0001928302447442557,
      "loss": 1.3967,
      "step": 25100
    },
    {
      "epoch": 0.7509031737975099,
      "grad_norm": 4.590394496917725,
      "learning_rate": 0.00019281595894833685,
      "loss": 1.3925,
      "step": 25150
    },
    {
      "epoch": 0.7523960230495924,
      "grad_norm": 4.220889091491699,
      "learning_rate": 0.00019280167315241802,
      "loss": 1.3263,
      "step": 25200
    },
    {
      "epoch": 0.753888872301675,
      "grad_norm": 4.847371578216553,
      "learning_rate": 0.00019278738735649918,
      "loss": 1.4141,
      "step": 25250
    },
    {
      "epoch": 0.7553817215537575,
      "grad_norm": 5.392793655395508,
      "learning_rate": 0.00019277310156058035,
      "loss": 1.3147,
      "step": 25300
    },
    {
      "epoch": 0.75687457080584,
      "grad_norm": 4.261468887329102,
      "learning_rate": 0.00019275881576466154,
      "loss": 1.3102,
      "step": 25350
    },
    {
      "epoch": 0.7583674200579226,
      "grad_norm": 4.498802661895752,
      "learning_rate": 0.00019274452996874267,
      "loss": 1.3912,
      "step": 25400
    },
    {
      "epoch": 0.759860269310005,
      "grad_norm": 5.2235283851623535,
      "learning_rate": 0.00019273024417282387,
      "loss": 1.3375,
      "step": 25450
    },
    {
      "epoch": 0.7613531185620876,
      "grad_norm": 6.409016132354736,
      "learning_rate": 0.000192715958376905,
      "loss": 1.3339,
      "step": 25500
    },
    {
      "epoch": 0.7628459678141701,
      "grad_norm": 4.4392805099487305,
      "learning_rate": 0.0001927016725809862,
      "loss": 1.3599,
      "step": 25550
    },
    {
      "epoch": 0.7643388170662526,
      "grad_norm": 5.580776691436768,
      "learning_rate": 0.00019268738678506736,
      "loss": 1.3812,
      "step": 25600
    },
    {
      "epoch": 0.7658316663183352,
      "grad_norm": 5.473046779632568,
      "learning_rate": 0.00019267310098914852,
      "loss": 1.3582,
      "step": 25650
    },
    {
      "epoch": 0.7673245155704177,
      "grad_norm": 6.353076934814453,
      "learning_rate": 0.0001926588151932297,
      "loss": 1.3388,
      "step": 25700
    },
    {
      "epoch": 0.7688173648225002,
      "grad_norm": 4.247453689575195,
      "learning_rate": 0.00019264452939731085,
      "loss": 1.3745,
      "step": 25750
    },
    {
      "epoch": 0.7703102140745828,
      "grad_norm": 5.048892498016357,
      "learning_rate": 0.00019263024360139202,
      "loss": 1.405,
      "step": 25800
    },
    {
      "epoch": 0.7718030633266653,
      "grad_norm": 4.883440017700195,
      "learning_rate": 0.00019261595780547318,
      "loss": 1.4082,
      "step": 25850
    },
    {
      "epoch": 0.7732959125787477,
      "grad_norm": 4.221151828765869,
      "learning_rate": 0.00019260167200955435,
      "loss": 1.3766,
      "step": 25900
    },
    {
      "epoch": 0.7747887618308303,
      "grad_norm": 8.68738079071045,
      "learning_rate": 0.0001925873862136355,
      "loss": 1.4046,
      "step": 25950
    },
    {
      "epoch": 0.7762816110829128,
      "grad_norm": 4.479017734527588,
      "learning_rate": 0.00019257310041771668,
      "loss": 1.3452,
      "step": 26000
    },
    {
      "epoch": 0.7777744603349954,
      "grad_norm": 4.299473285675049,
      "learning_rate": 0.00019255881462179787,
      "loss": 1.3702,
      "step": 26050
    },
    {
      "epoch": 0.7792673095870779,
      "grad_norm": 5.272589683532715,
      "learning_rate": 0.000192544528825879,
      "loss": 1.3725,
      "step": 26100
    },
    {
      "epoch": 0.7807601588391604,
      "grad_norm": 4.644061088562012,
      "learning_rate": 0.0001925302430299602,
      "loss": 1.4143,
      "step": 26150
    },
    {
      "epoch": 0.782253008091243,
      "grad_norm": 4.7033185958862305,
      "learning_rate": 0.00019251595723404134,
      "loss": 1.3666,
      "step": 26200
    },
    {
      "epoch": 0.7837458573433255,
      "grad_norm": 3.5262560844421387,
      "learning_rate": 0.00019250167143812253,
      "loss": 1.4135,
      "step": 26250
    },
    {
      "epoch": 0.785238706595408,
      "grad_norm": 3.8599159717559814,
      "learning_rate": 0.00019248738564220367,
      "loss": 1.3258,
      "step": 26300
    },
    {
      "epoch": 0.7867315558474905,
      "grad_norm": 5.743364334106445,
      "learning_rate": 0.00019247309984628486,
      "loss": 1.3946,
      "step": 26350
    },
    {
      "epoch": 0.788224405099573,
      "grad_norm": 5.478078365325928,
      "learning_rate": 0.00019245881405036602,
      "loss": 1.3881,
      "step": 26400
    },
    {
      "epoch": 0.7897172543516555,
      "grad_norm": 5.912649631500244,
      "learning_rate": 0.0001924445282544472,
      "loss": 1.4139,
      "step": 26450
    },
    {
      "epoch": 0.7912101036037381,
      "grad_norm": 3.753570079803467,
      "learning_rate": 0.00019243024245852835,
      "loss": 1.3267,
      "step": 26500
    },
    {
      "epoch": 0.7927029528558206,
      "grad_norm": 4.155190944671631,
      "learning_rate": 0.00019241595666260952,
      "loss": 1.3539,
      "step": 26550
    },
    {
      "epoch": 0.7941958021079032,
      "grad_norm": 4.314638614654541,
      "learning_rate": 0.00019240167086669068,
      "loss": 1.3635,
      "step": 26600
    },
    {
      "epoch": 0.7956886513599857,
      "grad_norm": 4.177329063415527,
      "learning_rate": 0.00019238738507077185,
      "loss": 1.3642,
      "step": 26650
    },
    {
      "epoch": 0.7971815006120682,
      "grad_norm": 4.3052144050598145,
      "learning_rate": 0.000192373099274853,
      "loss": 1.3011,
      "step": 26700
    },
    {
      "epoch": 0.7986743498641508,
      "grad_norm": 5.800657272338867,
      "learning_rate": 0.00019235881347893417,
      "loss": 1.3247,
      "step": 26750
    },
    {
      "epoch": 0.8001671991162332,
      "grad_norm": 4.211179733276367,
      "learning_rate": 0.00019234452768301534,
      "loss": 1.3399,
      "step": 26800
    },
    {
      "epoch": 0.8016600483683157,
      "grad_norm": 4.8080315589904785,
      "learning_rate": 0.0001923302418870965,
      "loss": 1.3439,
      "step": 26850
    },
    {
      "epoch": 0.8031528976203983,
      "grad_norm": 5.03045129776001,
      "learning_rate": 0.00019231595609117767,
      "loss": 1.4009,
      "step": 26900
    },
    {
      "epoch": 0.8046457468724808,
      "grad_norm": 8.932162284851074,
      "learning_rate": 0.00019230167029525883,
      "loss": 1.3298,
      "step": 26950
    },
    {
      "epoch": 0.8061385961245633,
      "grad_norm": 5.651429176330566,
      "learning_rate": 0.00019228738449934,
      "loss": 1.4044,
      "step": 27000
    },
    {
      "epoch": 0.8076314453766459,
      "grad_norm": 4.428074836730957,
      "learning_rate": 0.00019227309870342116,
      "loss": 1.3988,
      "step": 27050
    },
    {
      "epoch": 0.8091242946287284,
      "grad_norm": 6.512781620025635,
      "learning_rate": 0.00019225881290750233,
      "loss": 1.3529,
      "step": 27100
    },
    {
      "epoch": 0.810617143880811,
      "grad_norm": 5.393406867980957,
      "learning_rate": 0.0001922445271115835,
      "loss": 1.3266,
      "step": 27150
    },
    {
      "epoch": 0.8121099931328934,
      "grad_norm": 3.74702525138855,
      "learning_rate": 0.00019223024131566468,
      "loss": 1.416,
      "step": 27200
    },
    {
      "epoch": 0.8136028423849759,
      "grad_norm": 3.856290817260742,
      "learning_rate": 0.00019221595551974582,
      "loss": 1.4082,
      "step": 27250
    },
    {
      "epoch": 0.8150956916370585,
      "grad_norm": 4.597848415374756,
      "learning_rate": 0.000192201669723827,
      "loss": 1.3751,
      "step": 27300
    },
    {
      "epoch": 0.816588540889141,
      "grad_norm": 3.875593662261963,
      "learning_rate": 0.00019218738392790815,
      "loss": 1.4024,
      "step": 27350
    },
    {
      "epoch": 0.8180813901412235,
      "grad_norm": 5.470495700836182,
      "learning_rate": 0.00019217309813198934,
      "loss": 1.394,
      "step": 27400
    },
    {
      "epoch": 0.8195742393933061,
      "grad_norm": 4.562834739685059,
      "learning_rate": 0.00019215881233607048,
      "loss": 1.3161,
      "step": 27450
    },
    {
      "epoch": 0.8210670886453886,
      "grad_norm": 4.183352470397949,
      "learning_rate": 0.00019214452654015167,
      "loss": 1.3427,
      "step": 27500
    },
    {
      "epoch": 0.8225599378974711,
      "grad_norm": 5.345236301422119,
      "learning_rate": 0.00019213024074423284,
      "loss": 1.3551,
      "step": 27550
    },
    {
      "epoch": 0.8240527871495537,
      "grad_norm": 5.217405319213867,
      "learning_rate": 0.000192115954948314,
      "loss": 1.3174,
      "step": 27600
    },
    {
      "epoch": 0.8255456364016361,
      "grad_norm": 4.486965179443359,
      "learning_rate": 0.00019210166915239517,
      "loss": 1.3931,
      "step": 27650
    },
    {
      "epoch": 0.8270384856537187,
      "grad_norm": 4.183075904846191,
      "learning_rate": 0.00019208738335647633,
      "loss": 1.3672,
      "step": 27700
    },
    {
      "epoch": 0.8285313349058012,
      "grad_norm": 4.916398525238037,
      "learning_rate": 0.0001920730975605575,
      "loss": 1.3993,
      "step": 27750
    },
    {
      "epoch": 0.8300241841578837,
      "grad_norm": 3.99765682220459,
      "learning_rate": 0.00019205881176463866,
      "loss": 1.362,
      "step": 27800
    },
    {
      "epoch": 0.8315170334099663,
      "grad_norm": 3.948981523513794,
      "learning_rate": 0.00019204452596871982,
      "loss": 1.3912,
      "step": 27850
    },
    {
      "epoch": 0.8330098826620488,
      "grad_norm": 6.524683475494385,
      "learning_rate": 0.000192030240172801,
      "loss": 1.3845,
      "step": 27900
    },
    {
      "epoch": 0.8345027319141313,
      "grad_norm": 4.598319053649902,
      "learning_rate": 0.00019201595437688215,
      "loss": 1.38,
      "step": 27950
    },
    {
      "epoch": 0.8359955811662139,
      "grad_norm": 4.287365436553955,
      "learning_rate": 0.00019200166858096335,
      "loss": 1.3687,
      "step": 28000
    },
    {
      "epoch": 0.8374884304182963,
      "grad_norm": 5.497170448303223,
      "learning_rate": 0.00019198738278504448,
      "loss": 1.4324,
      "step": 28050
    },
    {
      "epoch": 0.8389812796703788,
      "grad_norm": 3.9373042583465576,
      "learning_rate": 0.00019197309698912567,
      "loss": 1.3892,
      "step": 28100
    },
    {
      "epoch": 0.8404741289224614,
      "grad_norm": 5.157922267913818,
      "learning_rate": 0.0001919588111932068,
      "loss": 1.3764,
      "step": 28150
    },
    {
      "epoch": 0.8419669781745439,
      "grad_norm": 4.0024895668029785,
      "learning_rate": 0.000191944525397288,
      "loss": 1.3571,
      "step": 28200
    },
    {
      "epoch": 0.8434598274266265,
      "grad_norm": 4.894957542419434,
      "learning_rate": 0.00019193023960136917,
      "loss": 1.3997,
      "step": 28250
    },
    {
      "epoch": 0.844952676678709,
      "grad_norm": 4.829247951507568,
      "learning_rate": 0.00019191595380545033,
      "loss": 1.3751,
      "step": 28300
    },
    {
      "epoch": 0.8464455259307915,
      "grad_norm": 4.123593330383301,
      "learning_rate": 0.0001919016680095315,
      "loss": 1.4183,
      "step": 28350
    },
    {
      "epoch": 0.8479383751828741,
      "grad_norm": 4.012912273406982,
      "learning_rate": 0.00019188738221361266,
      "loss": 1.3757,
      "step": 28400
    },
    {
      "epoch": 0.8494312244349566,
      "grad_norm": 5.17979097366333,
      "learning_rate": 0.00019187309641769383,
      "loss": 1.3975,
      "step": 28450
    },
    {
      "epoch": 0.850924073687039,
      "grad_norm": 4.665642261505127,
      "learning_rate": 0.000191858810621775,
      "loss": 1.4539,
      "step": 28500
    },
    {
      "epoch": 0.8524169229391216,
      "grad_norm": 4.333798408508301,
      "learning_rate": 0.00019184452482585616,
      "loss": 1.3711,
      "step": 28550
    },
    {
      "epoch": 0.8539097721912041,
      "grad_norm": 4.485865116119385,
      "learning_rate": 0.00019183023902993732,
      "loss": 1.4299,
      "step": 28600
    },
    {
      "epoch": 0.8554026214432866,
      "grad_norm": 6.3318328857421875,
      "learning_rate": 0.00019181595323401849,
      "loss": 1.3621,
      "step": 28650
    },
    {
      "epoch": 0.8568954706953692,
      "grad_norm": 3.915356159210205,
      "learning_rate": 0.00019180166743809965,
      "loss": 1.4561,
      "step": 28700
    },
    {
      "epoch": 0.8583883199474517,
      "grad_norm": 4.613978862762451,
      "learning_rate": 0.00019178738164218081,
      "loss": 1.3267,
      "step": 28750
    },
    {
      "epoch": 0.8598811691995343,
      "grad_norm": 4.935427188873291,
      "learning_rate": 0.000191773095846262,
      "loss": 1.3962,
      "step": 28800
    },
    {
      "epoch": 0.8613740184516168,
      "grad_norm": 4.204285144805908,
      "learning_rate": 0.00019175881005034314,
      "loss": 1.3644,
      "step": 28850
    },
    {
      "epoch": 0.8628668677036992,
      "grad_norm": 3.901266098022461,
      "learning_rate": 0.00019174452425442434,
      "loss": 1.344,
      "step": 28900
    },
    {
      "epoch": 0.8643597169557818,
      "grad_norm": 5.1740922927856445,
      "learning_rate": 0.00019173023845850547,
      "loss": 1.3852,
      "step": 28950
    },
    {
      "epoch": 0.8658525662078643,
      "grad_norm": 4.984529972076416,
      "learning_rate": 0.00019171595266258667,
      "loss": 1.3853,
      "step": 29000
    },
    {
      "epoch": 0.8673454154599468,
      "grad_norm": 4.719232082366943,
      "learning_rate": 0.00019170166686666783,
      "loss": 1.3962,
      "step": 29050
    },
    {
      "epoch": 0.8688382647120294,
      "grad_norm": 6.084499359130859,
      "learning_rate": 0.000191687381070749,
      "loss": 1.3377,
      "step": 29100
    },
    {
      "epoch": 0.8703311139641119,
      "grad_norm": 5.261173248291016,
      "learning_rate": 0.00019167309527483016,
      "loss": 1.3673,
      "step": 29150
    },
    {
      "epoch": 0.8718239632161944,
      "grad_norm": 5.18539571762085,
      "learning_rate": 0.00019165880947891132,
      "loss": 1.358,
      "step": 29200
    },
    {
      "epoch": 0.873316812468277,
      "grad_norm": 4.789880275726318,
      "learning_rate": 0.0001916445236829925,
      "loss": 1.3426,
      "step": 29250
    },
    {
      "epoch": 0.8748096617203595,
      "grad_norm": 3.2945287227630615,
      "learning_rate": 0.00019163023788707365,
      "loss": 1.3308,
      "step": 29300
    },
    {
      "epoch": 0.876302510972442,
      "grad_norm": 5.060371398925781,
      "learning_rate": 0.00019161595209115482,
      "loss": 1.3497,
      "step": 29350
    },
    {
      "epoch": 0.8777953602245245,
      "grad_norm": 4.771635055541992,
      "learning_rate": 0.00019160166629523598,
      "loss": 1.3517,
      "step": 29400
    },
    {
      "epoch": 0.879288209476607,
      "grad_norm": 4.408042907714844,
      "learning_rate": 0.00019158738049931715,
      "loss": 1.3636,
      "step": 29450
    },
    {
      "epoch": 0.8807810587286896,
      "grad_norm": 5.183899879455566,
      "learning_rate": 0.0001915730947033983,
      "loss": 1.3569,
      "step": 29500
    },
    {
      "epoch": 0.8822739079807721,
      "grad_norm": 4.176016807556152,
      "learning_rate": 0.00019155880890747948,
      "loss": 1.3659,
      "step": 29550
    },
    {
      "epoch": 0.8837667572328546,
      "grad_norm": 4.783830165863037,
      "learning_rate": 0.00019154452311156064,
      "loss": 1.368,
      "step": 29600
    },
    {
      "epoch": 0.8852596064849372,
      "grad_norm": 3.7985739707946777,
      "learning_rate": 0.0001915302373156418,
      "loss": 1.4091,
      "step": 29650
    },
    {
      "epoch": 0.8867524557370197,
      "grad_norm": 4.401581287384033,
      "learning_rate": 0.00019151595151972297,
      "loss": 1.3613,
      "step": 29700
    },
    {
      "epoch": 0.8882453049891021,
      "grad_norm": 6.214754581451416,
      "learning_rate": 0.00019150166572380414,
      "loss": 1.4077,
      "step": 29750
    },
    {
      "epoch": 0.8897381542411847,
      "grad_norm": 4.197826862335205,
      "learning_rate": 0.0001914873799278853,
      "loss": 1.347,
      "step": 29800
    },
    {
      "epoch": 0.8912310034932672,
      "grad_norm": 3.1535027027130127,
      "learning_rate": 0.0001914730941319665,
      "loss": 1.3452,
      "step": 29850
    },
    {
      "epoch": 0.8927238527453498,
      "grad_norm": 4.618538856506348,
      "learning_rate": 0.00019145880833604763,
      "loss": 1.3417,
      "step": 29900
    },
    {
      "epoch": 0.8942167019974323,
      "grad_norm": 3.913506031036377,
      "learning_rate": 0.00019144452254012882,
      "loss": 1.3503,
      "step": 29950
    },
    {
      "epoch": 0.8957095512495148,
      "grad_norm": 4.013761043548584,
      "learning_rate": 0.00019143023674420996,
      "loss": 1.4109,
      "step": 30000
    },
    {
      "epoch": 0.8972024005015974,
      "grad_norm": 4.6257758140563965,
      "learning_rate": 0.00019141595094829115,
      "loss": 1.3679,
      "step": 30050
    },
    {
      "epoch": 0.8986952497536799,
      "grad_norm": 4.422520637512207,
      "learning_rate": 0.0001914016651523723,
      "loss": 1.3893,
      "step": 30100
    },
    {
      "epoch": 0.9001880990057624,
      "grad_norm": 4.911538124084473,
      "learning_rate": 0.00019138737935645348,
      "loss": 1.3346,
      "step": 30150
    },
    {
      "epoch": 0.901680948257845,
      "grad_norm": 6.410045623779297,
      "learning_rate": 0.00019137309356053464,
      "loss": 1.3651,
      "step": 30200
    },
    {
      "epoch": 0.9031737975099274,
      "grad_norm": 3.8565444946289062,
      "learning_rate": 0.0001913588077646158,
      "loss": 1.3436,
      "step": 30250
    },
    {
      "epoch": 0.9046666467620099,
      "grad_norm": 4.632288455963135,
      "learning_rate": 0.00019134452196869697,
      "loss": 1.3858,
      "step": 30300
    },
    {
      "epoch": 0.9061594960140925,
      "grad_norm": 5.382070064544678,
      "learning_rate": 0.00019133023617277814,
      "loss": 1.3872,
      "step": 30350
    },
    {
      "epoch": 0.907652345266175,
      "grad_norm": 4.425014019012451,
      "learning_rate": 0.0001913159503768593,
      "loss": 1.4488,
      "step": 30400
    },
    {
      "epoch": 0.9091451945182576,
      "grad_norm": 3.7266223430633545,
      "learning_rate": 0.00019130166458094047,
      "loss": 1.376,
      "step": 30450
    },
    {
      "epoch": 0.9106380437703401,
      "grad_norm": 7.25547981262207,
      "learning_rate": 0.00019128737878502163,
      "loss": 1.3488,
      "step": 30500
    },
    {
      "epoch": 0.9121308930224226,
      "grad_norm": 5.824578285217285,
      "learning_rate": 0.0001912730929891028,
      "loss": 1.4001,
      "step": 30550
    },
    {
      "epoch": 0.9136237422745052,
      "grad_norm": 5.808544635772705,
      "learning_rate": 0.00019125880719318396,
      "loss": 1.3132,
      "step": 30600
    },
    {
      "epoch": 0.9151165915265876,
      "grad_norm": 4.343785285949707,
      "learning_rate": 0.00019124452139726515,
      "loss": 1.3592,
      "step": 30650
    },
    {
      "epoch": 0.9166094407786701,
      "grad_norm": 5.0626325607299805,
      "learning_rate": 0.0001912302356013463,
      "loss": 1.4418,
      "step": 30700
    },
    {
      "epoch": 0.9181022900307527,
      "grad_norm": 4.336055755615234,
      "learning_rate": 0.00019121594980542748,
      "loss": 1.3624,
      "step": 30750
    },
    {
      "epoch": 0.9195951392828352,
      "grad_norm": 6.215260982513428,
      "learning_rate": 0.00019120166400950862,
      "loss": 1.4053,
      "step": 30800
    },
    {
      "epoch": 0.9210879885349177,
      "grad_norm": 4.496364116668701,
      "learning_rate": 0.0001911873782135898,
      "loss": 1.4117,
      "step": 30850
    },
    {
      "epoch": 0.9225808377870003,
      "grad_norm": 4.023138046264648,
      "learning_rate": 0.00019117309241767095,
      "loss": 1.3783,
      "step": 30900
    },
    {
      "epoch": 0.9240736870390828,
      "grad_norm": 3.8177783489227295,
      "learning_rate": 0.00019115880662175214,
      "loss": 1.3127,
      "step": 30950
    },
    {
      "epoch": 0.9255665362911654,
      "grad_norm": 3.893087387084961,
      "learning_rate": 0.0001911445208258333,
      "loss": 1.3359,
      "step": 31000
    },
    {
      "epoch": 0.9270593855432478,
      "grad_norm": 3.9967384338378906,
      "learning_rate": 0.00019113023502991447,
      "loss": 1.3897,
      "step": 31050
    },
    {
      "epoch": 0.9285522347953303,
      "grad_norm": 3.2351789474487305,
      "learning_rate": 0.00019111594923399563,
      "loss": 1.3406,
      "step": 31100
    },
    {
      "epoch": 0.9300450840474129,
      "grad_norm": 4.571500778198242,
      "learning_rate": 0.0001911016634380768,
      "loss": 1.3783,
      "step": 31150
    },
    {
      "epoch": 0.9315379332994954,
      "grad_norm": 6.241518020629883,
      "learning_rate": 0.00019108737764215796,
      "loss": 1.3659,
      "step": 31200
    },
    {
      "epoch": 0.9330307825515779,
      "grad_norm": 3.2547826766967773,
      "learning_rate": 0.00019107309184623913,
      "loss": 1.3137,
      "step": 31250
    },
    {
      "epoch": 0.9345236318036605,
      "grad_norm": 5.386812686920166,
      "learning_rate": 0.0001910588060503203,
      "loss": 1.3974,
      "step": 31300
    },
    {
      "epoch": 0.936016481055743,
      "grad_norm": 4.239005088806152,
      "learning_rate": 0.00019104452025440146,
      "loss": 1.3789,
      "step": 31350
    },
    {
      "epoch": 0.9375093303078255,
      "grad_norm": 3.7252326011657715,
      "learning_rate": 0.00019103023445848262,
      "loss": 1.3622,
      "step": 31400
    },
    {
      "epoch": 0.9390021795599081,
      "grad_norm": 5.890969276428223,
      "learning_rate": 0.00019101594866256381,
      "loss": 1.3905,
      "step": 31450
    },
    {
      "epoch": 0.9404950288119905,
      "grad_norm": 5.796853542327881,
      "learning_rate": 0.00019100166286664495,
      "loss": 1.3797,
      "step": 31500
    },
    {
      "epoch": 0.9419878780640731,
      "grad_norm": 4.130026817321777,
      "learning_rate": 0.00019098737707072614,
      "loss": 1.3365,
      "step": 31550
    },
    {
      "epoch": 0.9434807273161556,
      "grad_norm": 3.65081524848938,
      "learning_rate": 0.00019097309127480728,
      "loss": 1.4347,
      "step": 31600
    },
    {
      "epoch": 0.9449735765682381,
      "grad_norm": 4.91404390335083,
      "learning_rate": 0.00019095880547888847,
      "loss": 1.4056,
      "step": 31650
    },
    {
      "epoch": 0.9464664258203207,
      "grad_norm": 4.7272114753723145,
      "learning_rate": 0.00019094451968296964,
      "loss": 1.2828,
      "step": 31700
    },
    {
      "epoch": 0.9479592750724032,
      "grad_norm": 5.111660957336426,
      "learning_rate": 0.0001909302338870508,
      "loss": 1.3486,
      "step": 31750
    },
    {
      "epoch": 0.9494521243244857,
      "grad_norm": 3.1206741333007812,
      "learning_rate": 0.00019091594809113197,
      "loss": 1.3416,
      "step": 31800
    },
    {
      "epoch": 0.9509449735765683,
      "grad_norm": 4.359163761138916,
      "learning_rate": 0.00019090166229521313,
      "loss": 1.3471,
      "step": 31850
    },
    {
      "epoch": 0.9524378228286507,
      "grad_norm": 4.667807102203369,
      "learning_rate": 0.0001908873764992943,
      "loss": 1.4042,
      "step": 31900
    },
    {
      "epoch": 0.9539306720807332,
      "grad_norm": 4.410369873046875,
      "learning_rate": 0.00019087309070337546,
      "loss": 1.3909,
      "step": 31950
    },
    {
      "epoch": 0.9554235213328158,
      "grad_norm": 4.987631320953369,
      "learning_rate": 0.00019085880490745663,
      "loss": 1.3548,
      "step": 32000
    },
    {
      "epoch": 0.9569163705848983,
      "grad_norm": 4.66121768951416,
      "learning_rate": 0.0001908445191115378,
      "loss": 1.3753,
      "step": 32050
    },
    {
      "epoch": 0.9584092198369809,
      "grad_norm": 6.010061740875244,
      "learning_rate": 0.00019083023331561896,
      "loss": 1.4128,
      "step": 32100
    },
    {
      "epoch": 0.9599020690890634,
      "grad_norm": 3.118246078491211,
      "learning_rate": 0.00019081594751970012,
      "loss": 1.3625,
      "step": 32150
    },
    {
      "epoch": 0.9613949183411459,
      "grad_norm": 3.821732759475708,
      "learning_rate": 0.00019080166172378128,
      "loss": 1.3691,
      "step": 32200
    },
    {
      "epoch": 0.9628877675932285,
      "grad_norm": 5.321105480194092,
      "learning_rate": 0.00019078737592786245,
      "loss": 1.3499,
      "step": 32250
    },
    {
      "epoch": 0.964380616845311,
      "grad_norm": 5.963992595672607,
      "learning_rate": 0.00019077309013194361,
      "loss": 1.3145,
      "step": 32300
    },
    {
      "epoch": 0.9658734660973934,
      "grad_norm": 4.071386337280273,
      "learning_rate": 0.00019075880433602478,
      "loss": 1.3215,
      "step": 32350
    },
    {
      "epoch": 0.967366315349476,
      "grad_norm": 3.7641048431396484,
      "learning_rate": 0.00019074451854010594,
      "loss": 1.3846,
      "step": 32400
    },
    {
      "epoch": 0.9688591646015585,
      "grad_norm": 5.985085487365723,
      "learning_rate": 0.0001907302327441871,
      "loss": 1.3383,
      "step": 32450
    },
    {
      "epoch": 0.970352013853641,
      "grad_norm": 6.602528095245361,
      "learning_rate": 0.0001907159469482683,
      "loss": 1.4195,
      "step": 32500
    },
    {
      "epoch": 0.9718448631057236,
      "grad_norm": 5.370469570159912,
      "learning_rate": 0.00019070166115234944,
      "loss": 1.3485,
      "step": 32550
    },
    {
      "epoch": 0.9733377123578061,
      "grad_norm": 4.588457107543945,
      "learning_rate": 0.00019068737535643063,
      "loss": 1.3896,
      "step": 32600
    },
    {
      "epoch": 0.9748305616098887,
      "grad_norm": 6.389341354370117,
      "learning_rate": 0.00019067308956051177,
      "loss": 1.3118,
      "step": 32650
    },
    {
      "epoch": 0.9763234108619712,
      "grad_norm": 4.051975250244141,
      "learning_rate": 0.00019065880376459296,
      "loss": 1.4065,
      "step": 32700
    },
    {
      "epoch": 0.9778162601140536,
      "grad_norm": 5.474813938140869,
      "learning_rate": 0.0001906445179686741,
      "loss": 1.3967,
      "step": 32750
    },
    {
      "epoch": 0.9793091093661362,
      "grad_norm": 4.033454895019531,
      "learning_rate": 0.0001906302321727553,
      "loss": 1.3212,
      "step": 32800
    },
    {
      "epoch": 0.9808019586182187,
      "grad_norm": 4.46946382522583,
      "learning_rate": 0.00019061594637683645,
      "loss": 1.3917,
      "step": 32850
    },
    {
      "epoch": 0.9822948078703012,
      "grad_norm": 4.98955774307251,
      "learning_rate": 0.00019060166058091762,
      "loss": 1.3814,
      "step": 32900
    },
    {
      "epoch": 0.9837876571223838,
      "grad_norm": 4.879435062408447,
      "learning_rate": 0.00019058737478499878,
      "loss": 1.3486,
      "step": 32950
    },
    {
      "epoch": 0.9852805063744663,
      "grad_norm": 4.0941925048828125,
      "learning_rate": 0.00019057308898907995,
      "loss": 1.3289,
      "step": 33000
    },
    {
      "epoch": 0.9867733556265488,
      "grad_norm": 5.2992706298828125,
      "learning_rate": 0.0001905588031931611,
      "loss": 1.3458,
      "step": 33050
    },
    {
      "epoch": 0.9882662048786314,
      "grad_norm": 4.284411430358887,
      "learning_rate": 0.00019054451739724228,
      "loss": 1.3932,
      "step": 33100
    },
    {
      "epoch": 0.9897590541307139,
      "grad_norm": 4.969634532928467,
      "learning_rate": 0.00019053023160132344,
      "loss": 1.4044,
      "step": 33150
    },
    {
      "epoch": 0.9912519033827965,
      "grad_norm": 5.532953262329102,
      "learning_rate": 0.0001905159458054046,
      "loss": 1.3658,
      "step": 33200
    },
    {
      "epoch": 0.9927447526348789,
      "grad_norm": 4.402670383453369,
      "learning_rate": 0.00019050166000948577,
      "loss": 1.343,
      "step": 33250
    },
    {
      "epoch": 0.9942376018869614,
      "grad_norm": 4.458900451660156,
      "learning_rate": 0.00019048737421356696,
      "loss": 1.4002,
      "step": 33300
    },
    {
      "epoch": 0.995730451139044,
      "grad_norm": 3.934340238571167,
      "learning_rate": 0.0001904730884176481,
      "loss": 1.413,
      "step": 33350
    },
    {
      "epoch": 0.9972233003911265,
      "grad_norm": 4.274466514587402,
      "learning_rate": 0.0001904588026217293,
      "loss": 1.3552,
      "step": 33400
    },
    {
      "epoch": 0.998716149643209,
      "grad_norm": 4.774006366729736,
      "learning_rate": 0.00019044451682581043,
      "loss": 1.3825,
      "step": 33450
    },
    {
      "epoch": 1.0002089988952916,
      "grad_norm": 3.870762825012207,
      "learning_rate": 0.00019043023102989162,
      "loss": 1.3715,
      "step": 33500
    },
    {
      "epoch": 1.001701848147374,
      "grad_norm": 3.595686912536621,
      "learning_rate": 0.00019041594523397276,
      "loss": 1.3352,
      "step": 33550
    },
    {
      "epoch": 1.0031946973994565,
      "grad_norm": 5.492071151733398,
      "learning_rate": 0.00019040165943805395,
      "loss": 1.2927,
      "step": 33600
    },
    {
      "epoch": 1.004687546651539,
      "grad_norm": 5.381584167480469,
      "learning_rate": 0.00019038737364213511,
      "loss": 1.2545,
      "step": 33650
    },
    {
      "epoch": 1.0061803959036217,
      "grad_norm": 5.983656883239746,
      "learning_rate": 0.00019037308784621628,
      "loss": 1.2393,
      "step": 33700
    },
    {
      "epoch": 1.0076732451557042,
      "grad_norm": 4.956115245819092,
      "learning_rate": 0.00019035880205029744,
      "loss": 1.2931,
      "step": 33750
    },
    {
      "epoch": 1.0091660944077867,
      "grad_norm": 4.604572296142578,
      "learning_rate": 0.0001903445162543786,
      "loss": 1.2786,
      "step": 33800
    },
    {
      "epoch": 1.0106589436598692,
      "grad_norm": 5.043067932128906,
      "learning_rate": 0.00019033023045845977,
      "loss": 1.2933,
      "step": 33850
    },
    {
      "epoch": 1.0121517929119517,
      "grad_norm": 4.53336238861084,
      "learning_rate": 0.00019031594466254094,
      "loss": 1.3114,
      "step": 33900
    },
    {
      "epoch": 1.0136446421640344,
      "grad_norm": 5.106605052947998,
      "learning_rate": 0.0001903016588666221,
      "loss": 1.2994,
      "step": 33950
    },
    {
      "epoch": 1.0151374914161169,
      "grad_norm": 5.796351909637451,
      "learning_rate": 0.00019028737307070327,
      "loss": 1.2555,
      "step": 34000
    },
    {
      "epoch": 1.0166303406681994,
      "grad_norm": 4.970977306365967,
      "learning_rate": 0.00019027308727478443,
      "loss": 1.3343,
      "step": 34050
    },
    {
      "epoch": 1.0181231899202818,
      "grad_norm": 4.233397960662842,
      "learning_rate": 0.00019025880147886562,
      "loss": 1.2485,
      "step": 34100
    },
    {
      "epoch": 1.0196160391723643,
      "grad_norm": 4.012045383453369,
      "learning_rate": 0.00019024451568294676,
      "loss": 1.2816,
      "step": 34150
    },
    {
      "epoch": 1.0211088884244468,
      "grad_norm": 4.715073585510254,
      "learning_rate": 0.00019023022988702795,
      "loss": 1.2541,
      "step": 34200
    },
    {
      "epoch": 1.0226017376765295,
      "grad_norm": 3.7285947799682617,
      "learning_rate": 0.0001902159440911091,
      "loss": 1.3094,
      "step": 34250
    },
    {
      "epoch": 1.024094586928612,
      "grad_norm": 4.012038707733154,
      "learning_rate": 0.00019020165829519028,
      "loss": 1.2791,
      "step": 34300
    },
    {
      "epoch": 1.0255874361806945,
      "grad_norm": 4.04899263381958,
      "learning_rate": 0.00019018737249927142,
      "loss": 1.3085,
      "step": 34350
    },
    {
      "epoch": 1.027080285432777,
      "grad_norm": 4.380834102630615,
      "learning_rate": 0.0001901730867033526,
      "loss": 1.3212,
      "step": 34400
    },
    {
      "epoch": 1.0285731346848594,
      "grad_norm": 5.275148868560791,
      "learning_rate": 0.00019015880090743378,
      "loss": 1.2749,
      "step": 34450
    },
    {
      "epoch": 1.0300659839369422,
      "grad_norm": 4.162272930145264,
      "learning_rate": 0.00019014451511151494,
      "loss": 1.296,
      "step": 34500
    },
    {
      "epoch": 1.0315588331890246,
      "grad_norm": 4.128006935119629,
      "learning_rate": 0.0001901302293155961,
      "loss": 1.2547,
      "step": 34550
    },
    {
      "epoch": 1.0330516824411071,
      "grad_norm": 3.930121421813965,
      "learning_rate": 0.00019011594351967727,
      "loss": 1.2831,
      "step": 34600
    },
    {
      "epoch": 1.0345445316931896,
      "grad_norm": 4.515873908996582,
      "learning_rate": 0.00019010165772375843,
      "loss": 1.2747,
      "step": 34650
    },
    {
      "epoch": 1.036037380945272,
      "grad_norm": 4.484467506408691,
      "learning_rate": 0.0001900873719278396,
      "loss": 1.289,
      "step": 34700
    },
    {
      "epoch": 1.0375302301973546,
      "grad_norm": 3.8937742710113525,
      "learning_rate": 0.00019007308613192076,
      "loss": 1.2939,
      "step": 34750
    },
    {
      "epoch": 1.0390230794494373,
      "grad_norm": 4.367059230804443,
      "learning_rate": 0.00019005880033600193,
      "loss": 1.2522,
      "step": 34800
    },
    {
      "epoch": 1.0405159287015198,
      "grad_norm": 5.143396854400635,
      "learning_rate": 0.0001900445145400831,
      "loss": 1.2996,
      "step": 34850
    },
    {
      "epoch": 1.0420087779536022,
      "grad_norm": 5.060800552368164,
      "learning_rate": 0.00019003022874416426,
      "loss": 1.3122,
      "step": 34900
    },
    {
      "epoch": 1.0435016272056847,
      "grad_norm": 4.775914192199707,
      "learning_rate": 0.00019001594294824542,
      "loss": 1.2595,
      "step": 34950
    },
    {
      "epoch": 1.0449944764577672,
      "grad_norm": 5.256932258605957,
      "learning_rate": 0.0001900016571523266,
      "loss": 1.3148,
      "step": 35000
    },
    {
      "epoch": 1.04648732570985,
      "grad_norm": 4.253088474273682,
      "learning_rate": 0.00018998737135640775,
      "loss": 1.2575,
      "step": 35050
    },
    {
      "epoch": 1.0479801749619324,
      "grad_norm": 4.784812927246094,
      "learning_rate": 0.00018997308556048892,
      "loss": 1.2911,
      "step": 35100
    },
    {
      "epoch": 1.049473024214015,
      "grad_norm": 3.9015183448791504,
      "learning_rate": 0.0001899587997645701,
      "loss": 1.2647,
      "step": 35150
    },
    {
      "epoch": 1.0509658734660974,
      "grad_norm": 4.729675769805908,
      "learning_rate": 0.00018994451396865125,
      "loss": 1.2876,
      "step": 35200
    },
    {
      "epoch": 1.0524587227181799,
      "grad_norm": 4.921074867248535,
      "learning_rate": 0.00018993022817273244,
      "loss": 1.2956,
      "step": 35250
    },
    {
      "epoch": 1.0539515719702623,
      "grad_norm": 4.466576099395752,
      "learning_rate": 0.00018991594237681357,
      "loss": 1.2938,
      "step": 35300
    },
    {
      "epoch": 1.055444421222345,
      "grad_norm": 4.142183780670166,
      "learning_rate": 0.00018990165658089477,
      "loss": 1.282,
      "step": 35350
    },
    {
      "epoch": 1.0569372704744275,
      "grad_norm": 4.372234344482422,
      "learning_rate": 0.0001898873707849759,
      "loss": 1.2793,
      "step": 35400
    },
    {
      "epoch": 1.05843011972651,
      "grad_norm": 5.349823474884033,
      "learning_rate": 0.0001898730849890571,
      "loss": 1.3678,
      "step": 35450
    },
    {
      "epoch": 1.0599229689785925,
      "grad_norm": 4.57612419128418,
      "learning_rate": 0.00018985879919313826,
      "loss": 1.3256,
      "step": 35500
    },
    {
      "epoch": 1.061415818230675,
      "grad_norm": 5.3342156410217285,
      "learning_rate": 0.00018984451339721942,
      "loss": 1.2897,
      "step": 35550
    },
    {
      "epoch": 1.0629086674827577,
      "grad_norm": 5.361871242523193,
      "learning_rate": 0.0001898302276013006,
      "loss": 1.2933,
      "step": 35600
    },
    {
      "epoch": 1.0644015167348402,
      "grad_norm": 3.6657068729400635,
      "learning_rate": 0.00018981594180538175,
      "loss": 1.3359,
      "step": 35650
    },
    {
      "epoch": 1.0658943659869227,
      "grad_norm": 4.428374290466309,
      "learning_rate": 0.00018980165600946292,
      "loss": 1.3278,
      "step": 35700
    },
    {
      "epoch": 1.0673872152390051,
      "grad_norm": 5.079537391662598,
      "learning_rate": 0.00018978737021354408,
      "loss": 1.2699,
      "step": 35750
    },
    {
      "epoch": 1.0688800644910876,
      "grad_norm": 5.303153038024902,
      "learning_rate": 0.00018977308441762525,
      "loss": 1.2985,
      "step": 35800
    },
    {
      "epoch": 1.0703729137431701,
      "grad_norm": 4.319901943206787,
      "learning_rate": 0.0001897587986217064,
      "loss": 1.2661,
      "step": 35850
    },
    {
      "epoch": 1.0718657629952528,
      "grad_norm": 6.0582451820373535,
      "learning_rate": 0.00018974451282578758,
      "loss": 1.2837,
      "step": 35900
    },
    {
      "epoch": 1.0733586122473353,
      "grad_norm": 5.738245964050293,
      "learning_rate": 0.00018973022702986877,
      "loss": 1.3075,
      "step": 35950
    },
    {
      "epoch": 1.0748514614994178,
      "grad_norm": 4.602933883666992,
      "learning_rate": 0.0001897159412339499,
      "loss": 1.2585,
      "step": 36000
    },
    {
      "epoch": 1.0763443107515003,
      "grad_norm": 4.273153305053711,
      "learning_rate": 0.0001897016554380311,
      "loss": 1.3382,
      "step": 36050
    },
    {
      "epoch": 1.0778371600035828,
      "grad_norm": 4.225873947143555,
      "learning_rate": 0.00018968736964211224,
      "loss": 1.2964,
      "step": 36100
    },
    {
      "epoch": 1.0793300092556652,
      "grad_norm": 4.828727722167969,
      "learning_rate": 0.00018967308384619343,
      "loss": 1.3734,
      "step": 36150
    },
    {
      "epoch": 1.080822858507748,
      "grad_norm": 3.7251532077789307,
      "learning_rate": 0.00018965879805027457,
      "loss": 1.2706,
      "step": 36200
    },
    {
      "epoch": 1.0823157077598304,
      "grad_norm": 4.92055606842041,
      "learning_rate": 0.00018964451225435576,
      "loss": 1.3187,
      "step": 36250
    },
    {
      "epoch": 1.083808557011913,
      "grad_norm": 4.827970027923584,
      "learning_rate": 0.00018963022645843692,
      "loss": 1.2574,
      "step": 36300
    },
    {
      "epoch": 1.0853014062639954,
      "grad_norm": 4.742151737213135,
      "learning_rate": 0.00018961594066251809,
      "loss": 1.3076,
      "step": 36350
    },
    {
      "epoch": 1.086794255516078,
      "grad_norm": 3.70794939994812,
      "learning_rate": 0.00018960165486659925,
      "loss": 1.3075,
      "step": 36400
    },
    {
      "epoch": 1.0882871047681606,
      "grad_norm": 4.062809944152832,
      "learning_rate": 0.00018958736907068042,
      "loss": 1.2628,
      "step": 36450
    },
    {
      "epoch": 1.089779954020243,
      "grad_norm": 5.072466850280762,
      "learning_rate": 0.00018957308327476158,
      "loss": 1.3278,
      "step": 36500
    },
    {
      "epoch": 1.0912728032723256,
      "grad_norm": 4.530171871185303,
      "learning_rate": 0.00018955879747884274,
      "loss": 1.3233,
      "step": 36550
    },
    {
      "epoch": 1.092765652524408,
      "grad_norm": 3.9391825199127197,
      "learning_rate": 0.0001895445116829239,
      "loss": 1.2973,
      "step": 36600
    },
    {
      "epoch": 1.0942585017764905,
      "grad_norm": 6.774008274078369,
      "learning_rate": 0.00018953022588700507,
      "loss": 1.2801,
      "step": 36650
    },
    {
      "epoch": 1.0957513510285732,
      "grad_norm": 4.565464496612549,
      "learning_rate": 0.00018951594009108624,
      "loss": 1.3718,
      "step": 36700
    },
    {
      "epoch": 1.0972442002806557,
      "grad_norm": 5.68241548538208,
      "learning_rate": 0.00018950165429516743,
      "loss": 1.2928,
      "step": 36750
    },
    {
      "epoch": 1.0987370495327382,
      "grad_norm": 4.235448360443115,
      "learning_rate": 0.00018948736849924857,
      "loss": 1.3097,
      "step": 36800
    },
    {
      "epoch": 1.1002298987848207,
      "grad_norm": 4.672369003295898,
      "learning_rate": 0.00018947308270332976,
      "loss": 1.2875,
      "step": 36850
    },
    {
      "epoch": 1.1017227480369032,
      "grad_norm": 6.449750900268555,
      "learning_rate": 0.0001894587969074109,
      "loss": 1.2823,
      "step": 36900
    },
    {
      "epoch": 1.1032155972889857,
      "grad_norm": 5.093149662017822,
      "learning_rate": 0.0001894445111114921,
      "loss": 1.315,
      "step": 36950
    },
    {
      "epoch": 1.1047084465410684,
      "grad_norm": 5.433828353881836,
      "learning_rate": 0.00018943022531557323,
      "loss": 1.2781,
      "step": 37000
    },
    {
      "epoch": 1.1062012957931509,
      "grad_norm": 5.176681995391846,
      "learning_rate": 0.00018941593951965442,
      "loss": 1.2652,
      "step": 37050
    },
    {
      "epoch": 1.1076941450452333,
      "grad_norm": 4.09026575088501,
      "learning_rate": 0.00018940165372373558,
      "loss": 1.3045,
      "step": 37100
    },
    {
      "epoch": 1.1091869942973158,
      "grad_norm": 5.223085880279541,
      "learning_rate": 0.00018938736792781675,
      "loss": 1.3009,
      "step": 37150
    },
    {
      "epoch": 1.1106798435493983,
      "grad_norm": 5.105752944946289,
      "learning_rate": 0.0001893730821318979,
      "loss": 1.2915,
      "step": 37200
    },
    {
      "epoch": 1.1121726928014808,
      "grad_norm": 3.945962905883789,
      "learning_rate": 0.00018935879633597908,
      "loss": 1.341,
      "step": 37250
    },
    {
      "epoch": 1.1136655420535635,
      "grad_norm": 4.856802463531494,
      "learning_rate": 0.00018934451054006024,
      "loss": 1.3318,
      "step": 37300
    },
    {
      "epoch": 1.115158391305646,
      "grad_norm": 4.858597755432129,
      "learning_rate": 0.0001893302247441414,
      "loss": 1.2756,
      "step": 37350
    },
    {
      "epoch": 1.1166512405577285,
      "grad_norm": 4.594838619232178,
      "learning_rate": 0.00018931593894822257,
      "loss": 1.2712,
      "step": 37400
    },
    {
      "epoch": 1.118144089809811,
      "grad_norm": 4.745405673980713,
      "learning_rate": 0.00018930165315230374,
      "loss": 1.2878,
      "step": 37450
    },
    {
      "epoch": 1.1196369390618934,
      "grad_norm": 4.112718105316162,
      "learning_rate": 0.0001892873673563849,
      "loss": 1.3337,
      "step": 37500
    },
    {
      "epoch": 1.1211297883139761,
      "grad_norm": 4.297769069671631,
      "learning_rate": 0.00018927308156046607,
      "loss": 1.3324,
      "step": 37550
    },
    {
      "epoch": 1.1226226375660586,
      "grad_norm": 5.46763801574707,
      "learning_rate": 0.00018925879576454723,
      "loss": 1.276,
      "step": 37600
    },
    {
      "epoch": 1.124115486818141,
      "grad_norm": 5.648976802825928,
      "learning_rate": 0.0001892445099686284,
      "loss": 1.3238,
      "step": 37650
    },
    {
      "epoch": 1.1256083360702236,
      "grad_norm": 4.185100078582764,
      "learning_rate": 0.00018923022417270956,
      "loss": 1.2848,
      "step": 37700
    },
    {
      "epoch": 1.127101185322306,
      "grad_norm": 4.676313877105713,
      "learning_rate": 0.00018921593837679072,
      "loss": 1.3038,
      "step": 37750
    },
    {
      "epoch": 1.1285940345743888,
      "grad_norm": 4.426568508148193,
      "learning_rate": 0.00018920165258087192,
      "loss": 1.3343,
      "step": 37800
    },
    {
      "epoch": 1.1300868838264713,
      "grad_norm": 4.887205600738525,
      "learning_rate": 0.00018918736678495305,
      "loss": 1.3456,
      "step": 37850
    },
    {
      "epoch": 1.1315797330785538,
      "grad_norm": 5.455615043640137,
      "learning_rate": 0.00018917308098903424,
      "loss": 1.3258,
      "step": 37900
    },
    {
      "epoch": 1.1330725823306362,
      "grad_norm": 5.676678657531738,
      "learning_rate": 0.00018915879519311538,
      "loss": 1.3702,
      "step": 37950
    },
    {
      "epoch": 1.1345654315827187,
      "grad_norm": 3.9528415203094482,
      "learning_rate": 0.00018914450939719657,
      "loss": 1.3097,
      "step": 38000
    },
    {
      "epoch": 1.1360582808348014,
      "grad_norm": 4.538026332855225,
      "learning_rate": 0.0001891302236012777,
      "loss": 1.316,
      "step": 38050
    },
    {
      "epoch": 1.137551130086884,
      "grad_norm": 4.46422815322876,
      "learning_rate": 0.0001891159378053589,
      "loss": 1.2479,
      "step": 38100
    },
    {
      "epoch": 1.1390439793389664,
      "grad_norm": 4.823958396911621,
      "learning_rate": 0.00018910165200944007,
      "loss": 1.2616,
      "step": 38150
    },
    {
      "epoch": 1.1405368285910489,
      "grad_norm": 5.7128167152404785,
      "learning_rate": 0.00018908736621352123,
      "loss": 1.2853,
      "step": 38200
    },
    {
      "epoch": 1.1420296778431314,
      "grad_norm": 3.5420877933502197,
      "learning_rate": 0.0001890730804176024,
      "loss": 1.3008,
      "step": 38250
    },
    {
      "epoch": 1.1435225270952138,
      "grad_norm": 4.617981910705566,
      "learning_rate": 0.00018905879462168356,
      "loss": 1.2888,
      "step": 38300
    },
    {
      "epoch": 1.1450153763472963,
      "grad_norm": 4.542544364929199,
      "learning_rate": 0.00018904450882576473,
      "loss": 1.3009,
      "step": 38350
    },
    {
      "epoch": 1.146508225599379,
      "grad_norm": 4.301428318023682,
      "learning_rate": 0.0001890302230298459,
      "loss": 1.2834,
      "step": 38400
    },
    {
      "epoch": 1.1480010748514615,
      "grad_norm": 4.136764049530029,
      "learning_rate": 0.00018901593723392706,
      "loss": 1.3895,
      "step": 38450
    },
    {
      "epoch": 1.149493924103544,
      "grad_norm": 5.59256649017334,
      "learning_rate": 0.00018900165143800822,
      "loss": 1.295,
      "step": 38500
    },
    {
      "epoch": 1.1509867733556265,
      "grad_norm": 3.4925365447998047,
      "learning_rate": 0.00018898736564208939,
      "loss": 1.3385,
      "step": 38550
    },
    {
      "epoch": 1.152479622607709,
      "grad_norm": 4.884555816650391,
      "learning_rate": 0.00018897307984617058,
      "loss": 1.2776,
      "step": 38600
    },
    {
      "epoch": 1.1539724718597917,
      "grad_norm": 3.866908550262451,
      "learning_rate": 0.00018895879405025171,
      "loss": 1.3173,
      "step": 38650
    },
    {
      "epoch": 1.1554653211118742,
      "grad_norm": 4.8104939460754395,
      "learning_rate": 0.0001889445082543329,
      "loss": 1.2988,
      "step": 38700
    },
    {
      "epoch": 1.1569581703639567,
      "grad_norm": 3.814675807952881,
      "learning_rate": 0.00018893022245841404,
      "loss": 1.348,
      "step": 38750
    },
    {
      "epoch": 1.1584510196160391,
      "grad_norm": 5.720306873321533,
      "learning_rate": 0.00018891593666249524,
      "loss": 1.2844,
      "step": 38800
    },
    {
      "epoch": 1.1599438688681216,
      "grad_norm": 4.06850528717041,
      "learning_rate": 0.00018890165086657637,
      "loss": 1.3514,
      "step": 38850
    },
    {
      "epoch": 1.1614367181202043,
      "grad_norm": 6.193358898162842,
      "learning_rate": 0.00018888736507065757,
      "loss": 1.3587,
      "step": 38900
    },
    {
      "epoch": 1.1629295673722868,
      "grad_norm": 4.8998212814331055,
      "learning_rate": 0.00018887307927473873,
      "loss": 1.317,
      "step": 38950
    },
    {
      "epoch": 1.1644224166243693,
      "grad_norm": 4.4293107986450195,
      "learning_rate": 0.0001888587934788199,
      "loss": 1.3473,
      "step": 39000
    },
    {
      "epoch": 1.1659152658764518,
      "grad_norm": 6.795536994934082,
      "learning_rate": 0.00018884450768290106,
      "loss": 1.2793,
      "step": 39050
    },
    {
      "epoch": 1.1674081151285343,
      "grad_norm": 4.158294200897217,
      "learning_rate": 0.00018883022188698222,
      "loss": 1.3175,
      "step": 39100
    },
    {
      "epoch": 1.168900964380617,
      "grad_norm": 5.839204788208008,
      "learning_rate": 0.0001888159360910634,
      "loss": 1.2931,
      "step": 39150
    },
    {
      "epoch": 1.1703938136326995,
      "grad_norm": 6.633917331695557,
      "learning_rate": 0.00018880165029514455,
      "loss": 1.3053,
      "step": 39200
    },
    {
      "epoch": 1.171886662884782,
      "grad_norm": 4.409125328063965,
      "learning_rate": 0.00018878736449922572,
      "loss": 1.3074,
      "step": 39250
    },
    {
      "epoch": 1.1733795121368644,
      "grad_norm": 4.820318698883057,
      "learning_rate": 0.00018877307870330688,
      "loss": 1.329,
      "step": 39300
    },
    {
      "epoch": 1.174872361388947,
      "grad_norm": 5.104337215423584,
      "learning_rate": 0.00018875879290738805,
      "loss": 1.2628,
      "step": 39350
    },
    {
      "epoch": 1.1763652106410294,
      "grad_norm": 5.449405670166016,
      "learning_rate": 0.00018874450711146924,
      "loss": 1.2672,
      "step": 39400
    },
    {
      "epoch": 1.1778580598931119,
      "grad_norm": 5.3521504402160645,
      "learning_rate": 0.00018873022131555038,
      "loss": 1.3236,
      "step": 39450
    },
    {
      "epoch": 1.1793509091451946,
      "grad_norm": 5.748175144195557,
      "learning_rate": 0.00018871593551963157,
      "loss": 1.284,
      "step": 39500
    },
    {
      "epoch": 1.180843758397277,
      "grad_norm": 4.418118953704834,
      "learning_rate": 0.0001887016497237127,
      "loss": 1.299,
      "step": 39550
    },
    {
      "epoch": 1.1823366076493596,
      "grad_norm": 4.267107009887695,
      "learning_rate": 0.0001886873639277939,
      "loss": 1.3287,
      "step": 39600
    },
    {
      "epoch": 1.183829456901442,
      "grad_norm": 4.607259273529053,
      "learning_rate": 0.00018867307813187503,
      "loss": 1.3087,
      "step": 39650
    },
    {
      "epoch": 1.1853223061535245,
      "grad_norm": 4.254667282104492,
      "learning_rate": 0.00018865879233595623,
      "loss": 1.3015,
      "step": 39700
    },
    {
      "epoch": 1.1868151554056072,
      "grad_norm": 5.307114124298096,
      "learning_rate": 0.0001886445065400374,
      "loss": 1.2802,
      "step": 39750
    },
    {
      "epoch": 1.1883080046576897,
      "grad_norm": 3.9305639266967773,
      "learning_rate": 0.00018863022074411856,
      "loss": 1.2975,
      "step": 39800
    },
    {
      "epoch": 1.1898008539097722,
      "grad_norm": 4.650544166564941,
      "learning_rate": 0.00018861593494819972,
      "loss": 1.3239,
      "step": 39850
    },
    {
      "epoch": 1.1912937031618547,
      "grad_norm": 4.182717800140381,
      "learning_rate": 0.00018860164915228086,
      "loss": 1.3948,
      "step": 39900
    },
    {
      "epoch": 1.1927865524139372,
      "grad_norm": 5.322524547576904,
      "learning_rate": 0.00018858736335636205,
      "loss": 1.3411,
      "step": 39950
    },
    {
      "epoch": 1.1942794016660199,
      "grad_norm": 5.221969127655029,
      "learning_rate": 0.0001885730775604432,
      "loss": 1.3047,
      "step": 40000
    },
    {
      "epoch": 1.1957722509181024,
      "grad_norm": 4.373313903808594,
      "learning_rate": 0.00018855879176452438,
      "loss": 1.366,
      "step": 40050
    },
    {
      "epoch": 1.1972651001701848,
      "grad_norm": 4.842396259307861,
      "learning_rate": 0.00018854450596860554,
      "loss": 1.3294,
      "step": 40100
    },
    {
      "epoch": 1.1987579494222673,
      "grad_norm": 5.024688720703125,
      "learning_rate": 0.0001885302201726867,
      "loss": 1.3531,
      "step": 40150
    },
    {
      "epoch": 1.2002507986743498,
      "grad_norm": 5.733693599700928,
      "learning_rate": 0.00018851593437676787,
      "loss": 1.3177,
      "step": 40200
    },
    {
      "epoch": 1.2017436479264325,
      "grad_norm": 4.634098529815674,
      "learning_rate": 0.00018850164858084904,
      "loss": 1.3534,
      "step": 40250
    },
    {
      "epoch": 1.203236497178515,
      "grad_norm": 5.662175178527832,
      "learning_rate": 0.0001884873627849302,
      "loss": 1.2813,
      "step": 40300
    },
    {
      "epoch": 1.2047293464305975,
      "grad_norm": 4.25814151763916,
      "learning_rate": 0.00018847307698901137,
      "loss": 1.3704,
      "step": 40350
    },
    {
      "epoch": 1.20622219568268,
      "grad_norm": 4.733590126037598,
      "learning_rate": 0.00018845879119309253,
      "loss": 1.2999,
      "step": 40400
    },
    {
      "epoch": 1.2077150449347624,
      "grad_norm": 5.873198509216309,
      "learning_rate": 0.0001884445053971737,
      "loss": 1.3697,
      "step": 40450
    },
    {
      "epoch": 1.209207894186845,
      "grad_norm": 4.847556114196777,
      "learning_rate": 0.00018843021960125486,
      "loss": 1.2706,
      "step": 40500
    },
    {
      "epoch": 1.2107007434389274,
      "grad_norm": 4.76710319519043,
      "learning_rate": 0.00018841593380533605,
      "loss": 1.3873,
      "step": 40550
    },
    {
      "epoch": 1.2121935926910101,
      "grad_norm": 4.693939685821533,
      "learning_rate": 0.0001884016480094172,
      "loss": 1.2752,
      "step": 40600
    },
    {
      "epoch": 1.2136864419430926,
      "grad_norm": 5.377971649169922,
      "learning_rate": 0.00018838736221349838,
      "loss": 1.3083,
      "step": 40650
    },
    {
      "epoch": 1.215179291195175,
      "grad_norm": 4.257247447967529,
      "learning_rate": 0.00018837307641757952,
      "loss": 1.3318,
      "step": 40700
    },
    {
      "epoch": 1.2166721404472576,
      "grad_norm": 4.820394039154053,
      "learning_rate": 0.0001883587906216607,
      "loss": 1.3063,
      "step": 40750
    },
    {
      "epoch": 1.21816498969934,
      "grad_norm": 3.7783172130584717,
      "learning_rate": 0.00018834450482574188,
      "loss": 1.2765,
      "step": 40800
    },
    {
      "epoch": 1.2196578389514228,
      "grad_norm": 5.447601795196533,
      "learning_rate": 0.00018833021902982304,
      "loss": 1.3067,
      "step": 40850
    },
    {
      "epoch": 1.2211506882035053,
      "grad_norm": 4.7725510597229,
      "learning_rate": 0.0001883159332339042,
      "loss": 1.3366,
      "step": 40900
    },
    {
      "epoch": 1.2226435374555877,
      "grad_norm": 5.6868672370910645,
      "learning_rate": 0.00018830164743798537,
      "loss": 1.3441,
      "step": 40950
    },
    {
      "epoch": 1.2241363867076702,
      "grad_norm": 8.460538864135742,
      "learning_rate": 0.00018828736164206653,
      "loss": 1.3305,
      "step": 41000
    },
    {
      "epoch": 1.2256292359597527,
      "grad_norm": 5.594083309173584,
      "learning_rate": 0.0001882730758461477,
      "loss": 1.2825,
      "step": 41050
    },
    {
      "epoch": 1.2271220852118354,
      "grad_norm": 6.125982284545898,
      "learning_rate": 0.00018825879005022886,
      "loss": 1.3253,
      "step": 41100
    },
    {
      "epoch": 1.228614934463918,
      "grad_norm": 4.173381328582764,
      "learning_rate": 0.00018824450425431003,
      "loss": 1.3363,
      "step": 41150
    },
    {
      "epoch": 1.2301077837160004,
      "grad_norm": 4.884373664855957,
      "learning_rate": 0.0001882302184583912,
      "loss": 1.3124,
      "step": 41200
    },
    {
      "epoch": 1.2316006329680829,
      "grad_norm": 6.418700695037842,
      "learning_rate": 0.00018821593266247239,
      "loss": 1.3123,
      "step": 41250
    },
    {
      "epoch": 1.2330934822201653,
      "grad_norm": 4.717264175415039,
      "learning_rate": 0.00018820164686655352,
      "loss": 1.287,
      "step": 41300
    },
    {
      "epoch": 1.234586331472248,
      "grad_norm": 5.55648136138916,
      "learning_rate": 0.00018818736107063471,
      "loss": 1.3426,
      "step": 41350
    },
    {
      "epoch": 1.2360791807243305,
      "grad_norm": 5.190313339233398,
      "learning_rate": 0.00018817307527471585,
      "loss": 1.2945,
      "step": 41400
    },
    {
      "epoch": 1.237572029976413,
      "grad_norm": 6.027787208557129,
      "learning_rate": 0.00018815878947879704,
      "loss": 1.3313,
      "step": 41450
    },
    {
      "epoch": 1.2390648792284955,
      "grad_norm": 4.4704413414001465,
      "learning_rate": 0.00018814450368287818,
      "loss": 1.3031,
      "step": 41500
    },
    {
      "epoch": 1.240557728480578,
      "grad_norm": 4.408693790435791,
      "learning_rate": 0.00018813021788695937,
      "loss": 1.287,
      "step": 41550
    },
    {
      "epoch": 1.2420505777326605,
      "grad_norm": 5.230643272399902,
      "learning_rate": 0.00018811593209104054,
      "loss": 1.2846,
      "step": 41600
    },
    {
      "epoch": 1.243543426984743,
      "grad_norm": 4.442893981933594,
      "learning_rate": 0.0001881016462951217,
      "loss": 1.3199,
      "step": 41650
    },
    {
      "epoch": 1.2450362762368257,
      "grad_norm": 3.9362857341766357,
      "learning_rate": 0.00018808736049920287,
      "loss": 1.3301,
      "step": 41700
    },
    {
      "epoch": 1.2465291254889082,
      "grad_norm": 4.765340805053711,
      "learning_rate": 0.00018807307470328403,
      "loss": 1.3092,
      "step": 41750
    },
    {
      "epoch": 1.2480219747409906,
      "grad_norm": 4.7868146896362305,
      "learning_rate": 0.0001880587889073652,
      "loss": 1.3077,
      "step": 41800
    },
    {
      "epoch": 1.2495148239930731,
      "grad_norm": 5.724763870239258,
      "learning_rate": 0.00018804450311144636,
      "loss": 1.3035,
      "step": 41850
    },
    {
      "epoch": 1.2510076732451556,
      "grad_norm": 5.832367897033691,
      "learning_rate": 0.00018803021731552753,
      "loss": 1.3703,
      "step": 41900
    },
    {
      "epoch": 1.2525005224972383,
      "grad_norm": 5.079206466674805,
      "learning_rate": 0.0001880159315196087,
      "loss": 1.3328,
      "step": 41950
    },
    {
      "epoch": 1.2539933717493208,
      "grad_norm": 5.704753398895264,
      "learning_rate": 0.00018800164572368986,
      "loss": 1.354,
      "step": 42000
    },
    {
      "epoch": 1.2554862210014033,
      "grad_norm": 5.396346569061279,
      "learning_rate": 0.00018798735992777105,
      "loss": 1.2794,
      "step": 42050
    },
    {
      "epoch": 1.2569790702534858,
      "grad_norm": 5.461711406707764,
      "learning_rate": 0.00018797307413185218,
      "loss": 1.3335,
      "step": 42100
    },
    {
      "epoch": 1.2584719195055682,
      "grad_norm": 3.3528285026550293,
      "learning_rate": 0.00018795878833593338,
      "loss": 1.296,
      "step": 42150
    },
    {
      "epoch": 1.259964768757651,
      "grad_norm": 4.413696765899658,
      "learning_rate": 0.00018794450254001451,
      "loss": 1.2828,
      "step": 42200
    },
    {
      "epoch": 1.2614576180097334,
      "grad_norm": 5.430131435394287,
      "learning_rate": 0.0001879302167440957,
      "loss": 1.3107,
      "step": 42250
    },
    {
      "epoch": 1.262950467261816,
      "grad_norm": 4.623260974884033,
      "learning_rate": 0.00018791593094817684,
      "loss": 1.3252,
      "step": 42300
    },
    {
      "epoch": 1.2644433165138984,
      "grad_norm": 5.377962112426758,
      "learning_rate": 0.00018790164515225803,
      "loss": 1.3073,
      "step": 42350
    },
    {
      "epoch": 1.265936165765981,
      "grad_norm": 5.34269905090332,
      "learning_rate": 0.0001878873593563392,
      "loss": 1.3705,
      "step": 42400
    },
    {
      "epoch": 1.2674290150180636,
      "grad_norm": 5.574904918670654,
      "learning_rate": 0.00018787307356042036,
      "loss": 1.3848,
      "step": 42450
    },
    {
      "epoch": 1.268921864270146,
      "grad_norm": 4.49995231628418,
      "learning_rate": 0.00018785878776450153,
      "loss": 1.3169,
      "step": 42500
    },
    {
      "epoch": 1.2704147135222286,
      "grad_norm": 3.9160735607147217,
      "learning_rate": 0.00018784450196858267,
      "loss": 1.2994,
      "step": 42550
    },
    {
      "epoch": 1.271907562774311,
      "grad_norm": 6.018070697784424,
      "learning_rate": 0.00018783021617266386,
      "loss": 1.3213,
      "step": 42600
    },
    {
      "epoch": 1.2734004120263935,
      "grad_norm": 5.017271041870117,
      "learning_rate": 0.000187815930376745,
      "loss": 1.2986,
      "step": 42650
    },
    {
      "epoch": 1.274893261278476,
      "grad_norm": 4.147556304931641,
      "learning_rate": 0.0001878016445808262,
      "loss": 1.2738,
      "step": 42700
    },
    {
      "epoch": 1.2763861105305585,
      "grad_norm": 4.864907741546631,
      "learning_rate": 0.00018778735878490735,
      "loss": 1.3414,
      "step": 42750
    },
    {
      "epoch": 1.2778789597826412,
      "grad_norm": 4.564859390258789,
      "learning_rate": 0.00018777307298898852,
      "loss": 1.3028,
      "step": 42800
    },
    {
      "epoch": 1.2793718090347237,
      "grad_norm": 4.96591854095459,
      "learning_rate": 0.00018775878719306968,
      "loss": 1.3197,
      "step": 42850
    },
    {
      "epoch": 1.2808646582868062,
      "grad_norm": 4.384543418884277,
      "learning_rate": 0.00018774450139715085,
      "loss": 1.2802,
      "step": 42900
    },
    {
      "epoch": 1.2823575075388887,
      "grad_norm": 3.956608295440674,
      "learning_rate": 0.000187730215601232,
      "loss": 1.2764,
      "step": 42950
    },
    {
      "epoch": 1.2838503567909711,
      "grad_norm": 4.815617561340332,
      "learning_rate": 0.00018771592980531318,
      "loss": 1.2384,
      "step": 43000
    },
    {
      "epoch": 1.2853432060430539,
      "grad_norm": 4.381842136383057,
      "learning_rate": 0.00018770164400939434,
      "loss": 1.3016,
      "step": 43050
    },
    {
      "epoch": 1.2868360552951363,
      "grad_norm": 6.777477741241455,
      "learning_rate": 0.0001876873582134755,
      "loss": 1.2956,
      "step": 43100
    },
    {
      "epoch": 1.2883289045472188,
      "grad_norm": 4.342850208282471,
      "learning_rate": 0.00018767307241755667,
      "loss": 1.338,
      "step": 43150
    },
    {
      "epoch": 1.2898217537993013,
      "grad_norm": 5.081860065460205,
      "learning_rate": 0.00018765878662163786,
      "loss": 1.292,
      "step": 43200
    },
    {
      "epoch": 1.2913146030513838,
      "grad_norm": 4.7942986488342285,
      "learning_rate": 0.000187644500825719,
      "loss": 1.2969,
      "step": 43250
    },
    {
      "epoch": 1.2928074523034665,
      "grad_norm": 3.766878604888916,
      "learning_rate": 0.0001876302150298002,
      "loss": 1.3561,
      "step": 43300
    },
    {
      "epoch": 1.294300301555549,
      "grad_norm": 5.662395000457764,
      "learning_rate": 0.00018761592923388133,
      "loss": 1.2993,
      "step": 43350
    },
    {
      "epoch": 1.2957931508076315,
      "grad_norm": 4.3839545249938965,
      "learning_rate": 0.00018760164343796252,
      "loss": 1.3283,
      "step": 43400
    },
    {
      "epoch": 1.297286000059714,
      "grad_norm": 5.862764835357666,
      "learning_rate": 0.00018758735764204368,
      "loss": 1.3876,
      "step": 43450
    },
    {
      "epoch": 1.2987788493117964,
      "grad_norm": 4.883128643035889,
      "learning_rate": 0.00018757307184612485,
      "loss": 1.3456,
      "step": 43500
    },
    {
      "epoch": 1.3002716985638791,
      "grad_norm": 5.6903533935546875,
      "learning_rate": 0.000187558786050206,
      "loss": 1.351,
      "step": 43550
    },
    {
      "epoch": 1.3017645478159616,
      "grad_norm": 3.8554911613464355,
      "learning_rate": 0.00018754450025428718,
      "loss": 1.3385,
      "step": 43600
    },
    {
      "epoch": 1.303257397068044,
      "grad_norm": 4.693499565124512,
      "learning_rate": 0.00018753021445836834,
      "loss": 1.2923,
      "step": 43650
    },
    {
      "epoch": 1.3047502463201266,
      "grad_norm": 4.351410388946533,
      "learning_rate": 0.0001875159286624495,
      "loss": 1.37,
      "step": 43700
    },
    {
      "epoch": 1.306243095572209,
      "grad_norm": 3.915044069290161,
      "learning_rate": 0.00018750164286653067,
      "loss": 1.3378,
      "step": 43750
    },
    {
      "epoch": 1.3077359448242916,
      "grad_norm": 5.484678745269775,
      "learning_rate": 0.00018748735707061184,
      "loss": 1.2753,
      "step": 43800
    },
    {
      "epoch": 1.309228794076374,
      "grad_norm": 4.64120626449585,
      "learning_rate": 0.000187473071274693,
      "loss": 1.346,
      "step": 43850
    },
    {
      "epoch": 1.3107216433284568,
      "grad_norm": 6.287013053894043,
      "learning_rate": 0.00018745878547877417,
      "loss": 1.3233,
      "step": 43900
    },
    {
      "epoch": 1.3122144925805392,
      "grad_norm": 4.840941905975342,
      "learning_rate": 0.00018744449968285533,
      "loss": 1.3651,
      "step": 43950
    },
    {
      "epoch": 1.3137073418326217,
      "grad_norm": 4.027961730957031,
      "learning_rate": 0.00018743021388693652,
      "loss": 1.3241,
      "step": 44000
    },
    {
      "epoch": 1.3152001910847042,
      "grad_norm": 5.4013190269470215,
      "learning_rate": 0.00018741592809101766,
      "loss": 1.3048,
      "step": 44050
    },
    {
      "epoch": 1.3166930403367867,
      "grad_norm": 5.291625022888184,
      "learning_rate": 0.00018740164229509885,
      "loss": 1.3171,
      "step": 44100
    },
    {
      "epoch": 1.3181858895888694,
      "grad_norm": 3.7480239868164062,
      "learning_rate": 0.00018738735649918,
      "loss": 1.3527,
      "step": 44150
    },
    {
      "epoch": 1.3196787388409519,
      "grad_norm": 3.080198287963867,
      "learning_rate": 0.00018737307070326118,
      "loss": 1.3782,
      "step": 44200
    },
    {
      "epoch": 1.3211715880930344,
      "grad_norm": 5.115420818328857,
      "learning_rate": 0.00018735878490734235,
      "loss": 1.3359,
      "step": 44250
    },
    {
      "epoch": 1.3226644373451169,
      "grad_norm": 4.522632598876953,
      "learning_rate": 0.0001873444991114235,
      "loss": 1.3557,
      "step": 44300
    },
    {
      "epoch": 1.3241572865971993,
      "grad_norm": 5.980693817138672,
      "learning_rate": 0.00018733021331550468,
      "loss": 1.3356,
      "step": 44350
    },
    {
      "epoch": 1.325650135849282,
      "grad_norm": 4.4641947746276855,
      "learning_rate": 0.00018731592751958584,
      "loss": 1.3179,
      "step": 44400
    },
    {
      "epoch": 1.3271429851013645,
      "grad_norm": 5.057164669036865,
      "learning_rate": 0.000187301641723667,
      "loss": 1.2966,
      "step": 44450
    },
    {
      "epoch": 1.328635834353447,
      "grad_norm": 4.383616924285889,
      "learning_rate": 0.00018728735592774817,
      "loss": 1.317,
      "step": 44500
    },
    {
      "epoch": 1.3301286836055295,
      "grad_norm": 5.158847332000732,
      "learning_rate": 0.00018727307013182933,
      "loss": 1.3278,
      "step": 44550
    },
    {
      "epoch": 1.331621532857612,
      "grad_norm": 4.217287063598633,
      "learning_rate": 0.0001872587843359105,
      "loss": 1.2695,
      "step": 44600
    },
    {
      "epoch": 1.3331143821096947,
      "grad_norm": 3.9984827041625977,
      "learning_rate": 0.00018724449853999166,
      "loss": 1.287,
      "step": 44650
    },
    {
      "epoch": 1.3346072313617772,
      "grad_norm": 3.536656379699707,
      "learning_rate": 0.00018723021274407285,
      "loss": 1.3528,
      "step": 44700
    },
    {
      "epoch": 1.3361000806138597,
      "grad_norm": 8.800326347351074,
      "learning_rate": 0.000187215926948154,
      "loss": 1.3632,
      "step": 44750
    },
    {
      "epoch": 1.3375929298659421,
      "grad_norm": 5.808169364929199,
      "learning_rate": 0.00018720164115223518,
      "loss": 1.3802,
      "step": 44800
    },
    {
      "epoch": 1.3390857791180246,
      "grad_norm": 3.631208896636963,
      "learning_rate": 0.00018718735535631632,
      "loss": 1.3953,
      "step": 44850
    },
    {
      "epoch": 1.340578628370107,
      "grad_norm": 3.4300222396850586,
      "learning_rate": 0.0001871730695603975,
      "loss": 1.3154,
      "step": 44900
    },
    {
      "epoch": 1.3420714776221896,
      "grad_norm": 4.281446933746338,
      "learning_rate": 0.00018715878376447865,
      "loss": 1.3453,
      "step": 44950
    },
    {
      "epoch": 1.3435643268742723,
      "grad_norm": 4.406055450439453,
      "learning_rate": 0.00018714449796855984,
      "loss": 1.3131,
      "step": 45000
    },
    {
      "epoch": 1.3450571761263548,
      "grad_norm": 4.559666156768799,
      "learning_rate": 0.000187130212172641,
      "loss": 1.3005,
      "step": 45050
    },
    {
      "epoch": 1.3465500253784373,
      "grad_norm": 4.3830342292785645,
      "learning_rate": 0.00018711592637672217,
      "loss": 1.2796,
      "step": 45100
    },
    {
      "epoch": 1.3480428746305197,
      "grad_norm": 3.840522289276123,
      "learning_rate": 0.00018710164058080334,
      "loss": 1.3683,
      "step": 45150
    },
    {
      "epoch": 1.3495357238826022,
      "grad_norm": 6.221634387969971,
      "learning_rate": 0.00018708735478488447,
      "loss": 1.3393,
      "step": 45200
    },
    {
      "epoch": 1.351028573134685,
      "grad_norm": 5.086376190185547,
      "learning_rate": 0.00018707306898896567,
      "loss": 1.3007,
      "step": 45250
    },
    {
      "epoch": 1.3525214223867674,
      "grad_norm": 5.3263373374938965,
      "learning_rate": 0.0001870587831930468,
      "loss": 1.3334,
      "step": 45300
    },
    {
      "epoch": 1.35401427163885,
      "grad_norm": 3.790532350540161,
      "learning_rate": 0.000187044497397128,
      "loss": 1.3282,
      "step": 45350
    },
    {
      "epoch": 1.3555071208909324,
      "grad_norm": 5.929101943969727,
      "learning_rate": 0.00018703021160120916,
      "loss": 1.3048,
      "step": 45400
    },
    {
      "epoch": 1.3569999701430149,
      "grad_norm": 4.900948524475098,
      "learning_rate": 0.00018701592580529032,
      "loss": 1.3282,
      "step": 45450
    },
    {
      "epoch": 1.3584928193950976,
      "grad_norm": 4.735415935516357,
      "learning_rate": 0.0001870016400093715,
      "loss": 1.3076,
      "step": 45500
    },
    {
      "epoch": 1.35998566864718,
      "grad_norm": 3.394000768661499,
      "learning_rate": 0.00018698735421345265,
      "loss": 1.3283,
      "step": 45550
    },
    {
      "epoch": 1.3614785178992626,
      "grad_norm": 4.902571678161621,
      "learning_rate": 0.00018697306841753382,
      "loss": 1.3456,
      "step": 45600
    },
    {
      "epoch": 1.362971367151345,
      "grad_norm": 3.773273229598999,
      "learning_rate": 0.00018695878262161498,
      "loss": 1.3095,
      "step": 45650
    },
    {
      "epoch": 1.3644642164034275,
      "grad_norm": 3.6063497066497803,
      "learning_rate": 0.00018694449682569615,
      "loss": 1.3088,
      "step": 45700
    },
    {
      "epoch": 1.3659570656555102,
      "grad_norm": 5.380884170532227,
      "learning_rate": 0.0001869302110297773,
      "loss": 1.2992,
      "step": 45750
    },
    {
      "epoch": 1.3674499149075927,
      "grad_norm": 3.108400821685791,
      "learning_rate": 0.00018691592523385848,
      "loss": 1.3638,
      "step": 45800
    },
    {
      "epoch": 1.3689427641596752,
      "grad_norm": 5.948930740356445,
      "learning_rate": 0.00018690163943793967,
      "loss": 1.4066,
      "step": 45850
    },
    {
      "epoch": 1.3704356134117577,
      "grad_norm": 3.5984768867492676,
      "learning_rate": 0.0001868873536420208,
      "loss": 1.3167,
      "step": 45900
    },
    {
      "epoch": 1.3719284626638402,
      "grad_norm": 4.595841884613037,
      "learning_rate": 0.000186873067846102,
      "loss": 1.2317,
      "step": 45950
    },
    {
      "epoch": 1.3734213119159226,
      "grad_norm": 4.499626159667969,
      "learning_rate": 0.00018685878205018314,
      "loss": 1.3094,
      "step": 46000
    },
    {
      "epoch": 1.3749141611680051,
      "grad_norm": 4.876091480255127,
      "learning_rate": 0.00018684449625426433,
      "loss": 1.3723,
      "step": 46050
    },
    {
      "epoch": 1.3764070104200878,
      "grad_norm": 5.354704856872559,
      "learning_rate": 0.00018683021045834547,
      "loss": 1.3672,
      "step": 46100
    },
    {
      "epoch": 1.3778998596721703,
      "grad_norm": 4.629252910614014,
      "learning_rate": 0.00018681592466242666,
      "loss": 1.3295,
      "step": 46150
    },
    {
      "epoch": 1.3793927089242528,
      "grad_norm": 3.6931405067443848,
      "learning_rate": 0.00018680163886650782,
      "loss": 1.2989,
      "step": 46200
    },
    {
      "epoch": 1.3808855581763353,
      "grad_norm": 6.378023624420166,
      "learning_rate": 0.00018678735307058899,
      "loss": 1.3022,
      "step": 46250
    },
    {
      "epoch": 1.3823784074284178,
      "grad_norm": 4.780524253845215,
      "learning_rate": 0.00018677306727467015,
      "loss": 1.298,
      "step": 46300
    },
    {
      "epoch": 1.3838712566805005,
      "grad_norm": 4.4194560050964355,
      "learning_rate": 0.00018675878147875132,
      "loss": 1.3396,
      "step": 46350
    },
    {
      "epoch": 1.385364105932583,
      "grad_norm": 5.250397205352783,
      "learning_rate": 0.00018674449568283248,
      "loss": 1.347,
      "step": 46400
    },
    {
      "epoch": 1.3868569551846655,
      "grad_norm": 5.2740607261657715,
      "learning_rate": 0.00018673020988691364,
      "loss": 1.3962,
      "step": 46450
    },
    {
      "epoch": 1.388349804436748,
      "grad_norm": 4.3260884284973145,
      "learning_rate": 0.0001867159240909948,
      "loss": 1.3236,
      "step": 46500
    },
    {
      "epoch": 1.3898426536888304,
      "grad_norm": 4.773256778717041,
      "learning_rate": 0.00018670163829507597,
      "loss": 1.324,
      "step": 46550
    },
    {
      "epoch": 1.3913355029409131,
      "grad_norm": 6.666245460510254,
      "learning_rate": 0.00018668735249915714,
      "loss": 1.3242,
      "step": 46600
    },
    {
      "epoch": 1.3928283521929956,
      "grad_norm": 7.051351547241211,
      "learning_rate": 0.00018667306670323833,
      "loss": 1.3228,
      "step": 46650
    },
    {
      "epoch": 1.394321201445078,
      "grad_norm": 7.103764057159424,
      "learning_rate": 0.00018665878090731947,
      "loss": 1.3815,
      "step": 46700
    },
    {
      "epoch": 1.3958140506971606,
      "grad_norm": 4.093582630157471,
      "learning_rate": 0.00018664449511140066,
      "loss": 1.3352,
      "step": 46750
    },
    {
      "epoch": 1.397306899949243,
      "grad_norm": 4.276589870452881,
      "learning_rate": 0.0001866302093154818,
      "loss": 1.2578,
      "step": 46800
    },
    {
      "epoch": 1.3987997492013258,
      "grad_norm": 4.325264930725098,
      "learning_rate": 0.000186615923519563,
      "loss": 1.297,
      "step": 46850
    },
    {
      "epoch": 1.4002925984534083,
      "grad_norm": 4.978733062744141,
      "learning_rate": 0.00018660163772364415,
      "loss": 1.3142,
      "step": 46900
    },
    {
      "epoch": 1.4017854477054907,
      "grad_norm": 4.2968854904174805,
      "learning_rate": 0.00018658735192772532,
      "loss": 1.3085,
      "step": 46950
    },
    {
      "epoch": 1.4032782969575732,
      "grad_norm": 3.587477922439575,
      "learning_rate": 0.00018657306613180648,
      "loss": 1.3348,
      "step": 47000
    },
    {
      "epoch": 1.4047711462096557,
      "grad_norm": 5.918453693389893,
      "learning_rate": 0.00018655878033588765,
      "loss": 1.3163,
      "step": 47050
    },
    {
      "epoch": 1.4062639954617382,
      "grad_norm": 4.217932224273682,
      "learning_rate": 0.0001865444945399688,
      "loss": 1.358,
      "step": 47100
    },
    {
      "epoch": 1.4077568447138207,
      "grad_norm": 4.987464904785156,
      "learning_rate": 0.00018653020874404998,
      "loss": 1.2865,
      "step": 47150
    },
    {
      "epoch": 1.4092496939659034,
      "grad_norm": 4.03493070602417,
      "learning_rate": 0.00018651592294813114,
      "loss": 1.2961,
      "step": 47200
    },
    {
      "epoch": 1.4107425432179859,
      "grad_norm": 4.768476486206055,
      "learning_rate": 0.0001865016371522123,
      "loss": 1.3301,
      "step": 47250
    },
    {
      "epoch": 1.4122353924700684,
      "grad_norm": 4.223608016967773,
      "learning_rate": 0.00018648735135629347,
      "loss": 1.358,
      "step": 47300
    },
    {
      "epoch": 1.4137282417221508,
      "grad_norm": 5.97542142868042,
      "learning_rate": 0.00018647306556037464,
      "loss": 1.3001,
      "step": 47350
    },
    {
      "epoch": 1.4152210909742333,
      "grad_norm": 3.4876928329467773,
      "learning_rate": 0.0001864587797644558,
      "loss": 1.358,
      "step": 47400
    },
    {
      "epoch": 1.416713940226316,
      "grad_norm": 3.8894033432006836,
      "learning_rate": 0.000186444493968537,
      "loss": 1.3079,
      "step": 47450
    },
    {
      "epoch": 1.4182067894783985,
      "grad_norm": 5.399030685424805,
      "learning_rate": 0.00018643020817261813,
      "loss": 1.3059,
      "step": 47500
    },
    {
      "epoch": 1.419699638730481,
      "grad_norm": 3.9021189212799072,
      "learning_rate": 0.00018641592237669932,
      "loss": 1.2862,
      "step": 47550
    },
    {
      "epoch": 1.4211924879825635,
      "grad_norm": 4.103232383728027,
      "learning_rate": 0.00018640163658078046,
      "loss": 1.3533,
      "step": 47600
    },
    {
      "epoch": 1.422685337234646,
      "grad_norm": 4.994762897491455,
      "learning_rate": 0.00018638735078486165,
      "loss": 1.3515,
      "step": 47650
    },
    {
      "epoch": 1.4241781864867287,
      "grad_norm": 4.363476276397705,
      "learning_rate": 0.00018637306498894282,
      "loss": 1.3524,
      "step": 47700
    },
    {
      "epoch": 1.4256710357388112,
      "grad_norm": 4.2915873527526855,
      "learning_rate": 0.00018635877919302398,
      "loss": 1.3245,
      "step": 47750
    },
    {
      "epoch": 1.4271638849908936,
      "grad_norm": 4.654321670532227,
      "learning_rate": 0.00018634449339710514,
      "loss": 1.369,
      "step": 47800
    },
    {
      "epoch": 1.4286567342429761,
      "grad_norm": 4.786460876464844,
      "learning_rate": 0.00018633020760118628,
      "loss": 1.3485,
      "step": 47850
    },
    {
      "epoch": 1.4301495834950586,
      "grad_norm": 4.434101104736328,
      "learning_rate": 0.00018631592180526747,
      "loss": 1.3131,
      "step": 47900
    },
    {
      "epoch": 1.4316424327471413,
      "grad_norm": 4.622028827667236,
      "learning_rate": 0.0001863016360093486,
      "loss": 1.3383,
      "step": 47950
    },
    {
      "epoch": 1.4331352819992238,
      "grad_norm": 5.579216480255127,
      "learning_rate": 0.0001862873502134298,
      "loss": 1.3214,
      "step": 48000
    },
    {
      "epoch": 1.4346281312513063,
      "grad_norm": 5.444978713989258,
      "learning_rate": 0.00018627306441751097,
      "loss": 1.3243,
      "step": 48050
    },
    {
      "epoch": 1.4361209805033888,
      "grad_norm": 4.104014873504639,
      "learning_rate": 0.00018625877862159213,
      "loss": 1.241,
      "step": 48100
    },
    {
      "epoch": 1.4376138297554713,
      "grad_norm": 4.517495632171631,
      "learning_rate": 0.0001862444928256733,
      "loss": 1.4197,
      "step": 48150
    },
    {
      "epoch": 1.4391066790075537,
      "grad_norm": 5.38925838470459,
      "learning_rate": 0.00018623020702975446,
      "loss": 1.3125,
      "step": 48200
    },
    {
      "epoch": 1.4405995282596362,
      "grad_norm": 6.211074352264404,
      "learning_rate": 0.00018621592123383563,
      "loss": 1.2659,
      "step": 48250
    },
    {
      "epoch": 1.442092377511719,
      "grad_norm": 5.461385726928711,
      "learning_rate": 0.0001862016354379168,
      "loss": 1.3043,
      "step": 48300
    },
    {
      "epoch": 1.4435852267638014,
      "grad_norm": 8.296045303344727,
      "learning_rate": 0.00018618734964199796,
      "loss": 1.3422,
      "step": 48350
    },
    {
      "epoch": 1.445078076015884,
      "grad_norm": 4.640946388244629,
      "learning_rate": 0.00018617306384607912,
      "loss": 1.3317,
      "step": 48400
    },
    {
      "epoch": 1.4465709252679664,
      "grad_norm": 4.698127269744873,
      "learning_rate": 0.00018615877805016029,
      "loss": 1.3241,
      "step": 48450
    },
    {
      "epoch": 1.4480637745200489,
      "grad_norm": 4.810885429382324,
      "learning_rate": 0.00018614449225424148,
      "loss": 1.3204,
      "step": 48500
    },
    {
      "epoch": 1.4495566237721316,
      "grad_norm": 3.57252836227417,
      "learning_rate": 0.00018613020645832261,
      "loss": 1.3412,
      "step": 48550
    },
    {
      "epoch": 1.451049473024214,
      "grad_norm": 4.79307222366333,
      "learning_rate": 0.0001861159206624038,
      "loss": 1.325,
      "step": 48600
    },
    {
      "epoch": 1.4525423222762965,
      "grad_norm": 5.495840072631836,
      "learning_rate": 0.00018610163486648494,
      "loss": 1.3191,
      "step": 48650
    },
    {
      "epoch": 1.454035171528379,
      "grad_norm": 4.352155685424805,
      "learning_rate": 0.00018608734907056614,
      "loss": 1.3129,
      "step": 48700
    },
    {
      "epoch": 1.4555280207804615,
      "grad_norm": 4.606874942779541,
      "learning_rate": 0.00018607306327464727,
      "loss": 1.3212,
      "step": 48750
    },
    {
      "epoch": 1.4570208700325442,
      "grad_norm": 5.2679219245910645,
      "learning_rate": 0.00018605877747872846,
      "loss": 1.3531,
      "step": 48800
    },
    {
      "epoch": 1.4585137192846267,
      "grad_norm": 4.821314811706543,
      "learning_rate": 0.00018604449168280963,
      "loss": 1.3195,
      "step": 48850
    },
    {
      "epoch": 1.4600065685367092,
      "grad_norm": 4.137834072113037,
      "learning_rate": 0.0001860302058868908,
      "loss": 1.2751,
      "step": 48900
    },
    {
      "epoch": 1.4614994177887917,
      "grad_norm": 6.577256202697754,
      "learning_rate": 0.00018601592009097196,
      "loss": 1.3971,
      "step": 48950
    },
    {
      "epoch": 1.4629922670408742,
      "grad_norm": 4.749688148498535,
      "learning_rate": 0.00018600163429505312,
      "loss": 1.3557,
      "step": 49000
    },
    {
      "epoch": 1.4644851162929569,
      "grad_norm": 5.410492897033691,
      "learning_rate": 0.0001859873484991343,
      "loss": 1.2467,
      "step": 49050
    },
    {
      "epoch": 1.4659779655450393,
      "grad_norm": 4.303316593170166,
      "learning_rate": 0.00018597306270321545,
      "loss": 1.3166,
      "step": 49100
    },
    {
      "epoch": 1.4674708147971218,
      "grad_norm": 5.683919906616211,
      "learning_rate": 0.00018595877690729662,
      "loss": 1.2932,
      "step": 49150
    },
    {
      "epoch": 1.4689636640492043,
      "grad_norm": 5.251864433288574,
      "learning_rate": 0.00018594449111137778,
      "loss": 1.3179,
      "step": 49200
    },
    {
      "epoch": 1.4704565133012868,
      "grad_norm": 4.565097808837891,
      "learning_rate": 0.00018593020531545895,
      "loss": 1.327,
      "step": 49250
    },
    {
      "epoch": 1.4719493625533693,
      "grad_norm": 5.532310485839844,
      "learning_rate": 0.00018591591951954014,
      "loss": 1.3868,
      "step": 49300
    },
    {
      "epoch": 1.4734422118054518,
      "grad_norm": 4.5699782371521,
      "learning_rate": 0.00018590163372362128,
      "loss": 1.2583,
      "step": 49350
    },
    {
      "epoch": 1.4749350610575345,
      "grad_norm": 4.792230606079102,
      "learning_rate": 0.00018588734792770247,
      "loss": 1.3805,
      "step": 49400
    },
    {
      "epoch": 1.476427910309617,
      "grad_norm": 4.52764368057251,
      "learning_rate": 0.0001858730621317836,
      "loss": 1.3465,
      "step": 49450
    },
    {
      "epoch": 1.4779207595616994,
      "grad_norm": 5.21061372756958,
      "learning_rate": 0.0001858587763358648,
      "loss": 1.3778,
      "step": 49500
    },
    {
      "epoch": 1.479413608813782,
      "grad_norm": 4.855043888092041,
      "learning_rate": 0.00018584449053994593,
      "loss": 1.3313,
      "step": 49550
    },
    {
      "epoch": 1.4809064580658644,
      "grad_norm": 4.446802616119385,
      "learning_rate": 0.00018583020474402713,
      "loss": 1.3224,
      "step": 49600
    },
    {
      "epoch": 1.4823993073179471,
      "grad_norm": 5.054227828979492,
      "learning_rate": 0.0001858159189481083,
      "loss": 1.3782,
      "step": 49650
    },
    {
      "epoch": 1.4838921565700296,
      "grad_norm": 6.877427101135254,
      "learning_rate": 0.00018580163315218946,
      "loss": 1.3399,
      "step": 49700
    },
    {
      "epoch": 1.485385005822112,
      "grad_norm": 6.227908611297607,
      "learning_rate": 0.00018578734735627062,
      "loss": 1.2996,
      "step": 49750
    },
    {
      "epoch": 1.4868778550741946,
      "grad_norm": 6.2445759773254395,
      "learning_rate": 0.00018577306156035179,
      "loss": 1.3005,
      "step": 49800
    },
    {
      "epoch": 1.488370704326277,
      "grad_norm": 6.488719463348389,
      "learning_rate": 0.00018575877576443295,
      "loss": 1.3755,
      "step": 49850
    },
    {
      "epoch": 1.4898635535783598,
      "grad_norm": 5.3022589683532715,
      "learning_rate": 0.00018574448996851411,
      "loss": 1.3402,
      "step": 49900
    },
    {
      "epoch": 1.4913564028304422,
      "grad_norm": 4.986827373504639,
      "learning_rate": 0.00018573020417259528,
      "loss": 1.3695,
      "step": 49950
    },
    {
      "epoch": 1.4928492520825247,
      "grad_norm": 5.669780731201172,
      "learning_rate": 0.00018571591837667644,
      "loss": 1.314,
      "step": 50000
    },
    {
      "epoch": 1.4943421013346072,
      "grad_norm": 5.21511697769165,
      "learning_rate": 0.0001857016325807576,
      "loss": 1.363,
      "step": 50050
    },
    {
      "epoch": 1.4958349505866897,
      "grad_norm": 4.303940296173096,
      "learning_rate": 0.0001856873467848388,
      "loss": 1.2926,
      "step": 50100
    },
    {
      "epoch": 1.4973277998387724,
      "grad_norm": 4.7364020347595215,
      "learning_rate": 0.00018567306098891994,
      "loss": 1.3325,
      "step": 50150
    },
    {
      "epoch": 1.4988206490908549,
      "grad_norm": 4.325379371643066,
      "learning_rate": 0.00018565877519300113,
      "loss": 1.31,
      "step": 50200
    },
    {
      "epoch": 1.5003134983429374,
      "grad_norm": 5.442165374755859,
      "learning_rate": 0.00018564448939708227,
      "loss": 1.3445,
      "step": 50250
    },
    {
      "epoch": 1.5018063475950199,
      "grad_norm": 5.861702919006348,
      "learning_rate": 0.00018563020360116346,
      "loss": 1.3392,
      "step": 50300
    },
    {
      "epoch": 1.5032991968471023,
      "grad_norm": 4.277650356292725,
      "learning_rate": 0.00018561591780524462,
      "loss": 1.3127,
      "step": 50350
    },
    {
      "epoch": 1.504792046099185,
      "grad_norm": 5.670056343078613,
      "learning_rate": 0.00018560163200932576,
      "loss": 1.346,
      "step": 50400
    },
    {
      "epoch": 1.5062848953512673,
      "grad_norm": 4.479572772979736,
      "learning_rate": 0.00018558734621340695,
      "loss": 1.3254,
      "step": 50450
    },
    {
      "epoch": 1.50777774460335,
      "grad_norm": 4.820215225219727,
      "learning_rate": 0.0001855730604174881,
      "loss": 1.3043,
      "step": 50500
    },
    {
      "epoch": 1.5092705938554325,
      "grad_norm": 5.21779203414917,
      "learning_rate": 0.00018555877462156928,
      "loss": 1.2953,
      "step": 50550
    },
    {
      "epoch": 1.510763443107515,
      "grad_norm": 4.987244129180908,
      "learning_rate": 0.00018554448882565042,
      "loss": 1.3008,
      "step": 50600
    },
    {
      "epoch": 1.5122562923595977,
      "grad_norm": 4.658651828765869,
      "learning_rate": 0.0001855302030297316,
      "loss": 1.3146,
      "step": 50650
    },
    {
      "epoch": 1.51374914161168,
      "grad_norm": 4.724532604217529,
      "learning_rate": 0.00018551591723381278,
      "loss": 1.3467,
      "step": 50700
    },
    {
      "epoch": 1.5152419908637627,
      "grad_norm": 6.199383735656738,
      "learning_rate": 0.00018550163143789394,
      "loss": 1.3244,
      "step": 50750
    },
    {
      "epoch": 1.5167348401158451,
      "grad_norm": 4.255542278289795,
      "learning_rate": 0.0001854873456419751,
      "loss": 1.3373,
      "step": 50800
    },
    {
      "epoch": 1.5182276893679276,
      "grad_norm": 5.011137962341309,
      "learning_rate": 0.00018547305984605627,
      "loss": 1.2945,
      "step": 50850
    },
    {
      "epoch": 1.51972053862001,
      "grad_norm": 4.276299476623535,
      "learning_rate": 0.00018545877405013743,
      "loss": 1.3503,
      "step": 50900
    },
    {
      "epoch": 1.5212133878720926,
      "grad_norm": 5.867751598358154,
      "learning_rate": 0.0001854444882542186,
      "loss": 1.345,
      "step": 50950
    },
    {
      "epoch": 1.5227062371241753,
      "grad_norm": 3.6653764247894287,
      "learning_rate": 0.00018543020245829976,
      "loss": 1.3353,
      "step": 51000
    },
    {
      "epoch": 1.5241990863762576,
      "grad_norm": 4.758911609649658,
      "learning_rate": 0.00018541591666238093,
      "loss": 1.3488,
      "step": 51050
    },
    {
      "epoch": 1.5256919356283403,
      "grad_norm": 4.028106689453125,
      "learning_rate": 0.0001854016308664621,
      "loss": 1.3215,
      "step": 51100
    },
    {
      "epoch": 1.5271847848804228,
      "grad_norm": 4.353278636932373,
      "learning_rate": 0.00018538734507054329,
      "loss": 1.2984,
      "step": 51150
    },
    {
      "epoch": 1.5286776341325052,
      "grad_norm": 7.112466812133789,
      "learning_rate": 0.00018537305927462442,
      "loss": 1.3057,
      "step": 51200
    },
    {
      "epoch": 1.530170483384588,
      "grad_norm": 4.703035354614258,
      "learning_rate": 0.00018535877347870561,
      "loss": 1.3289,
      "step": 51250
    },
    {
      "epoch": 1.5316633326366702,
      "grad_norm": 4.110894203186035,
      "learning_rate": 0.00018534448768278675,
      "loss": 1.3587,
      "step": 51300
    },
    {
      "epoch": 1.533156181888753,
      "grad_norm": 5.426900863647461,
      "learning_rate": 0.00018533020188686794,
      "loss": 1.316,
      "step": 51350
    },
    {
      "epoch": 1.5346490311408354,
      "grad_norm": 4.927971839904785,
      "learning_rate": 0.00018531591609094908,
      "loss": 1.3775,
      "step": 51400
    },
    {
      "epoch": 1.5361418803929179,
      "grad_norm": 4.867948055267334,
      "learning_rate": 0.00018530163029503027,
      "loss": 1.3038,
      "step": 51450
    },
    {
      "epoch": 1.5376347296450006,
      "grad_norm": 4.8490471839904785,
      "learning_rate": 0.00018528734449911144,
      "loss": 1.3851,
      "step": 51500
    },
    {
      "epoch": 1.5391275788970828,
      "grad_norm": 4.634105682373047,
      "learning_rate": 0.0001852730587031926,
      "loss": 1.3215,
      "step": 51550
    },
    {
      "epoch": 1.5406204281491656,
      "grad_norm": 3.7799713611602783,
      "learning_rate": 0.00018525877290727377,
      "loss": 1.3371,
      "step": 51600
    },
    {
      "epoch": 1.542113277401248,
      "grad_norm": 3.9134316444396973,
      "learning_rate": 0.00018524448711135493,
      "loss": 1.3073,
      "step": 51650
    },
    {
      "epoch": 1.5436061266533305,
      "grad_norm": 3.5114309787750244,
      "learning_rate": 0.0001852302013154361,
      "loss": 1.3859,
      "step": 51700
    },
    {
      "epoch": 1.5450989759054132,
      "grad_norm": 3.7832348346710205,
      "learning_rate": 0.00018521591551951726,
      "loss": 1.369,
      "step": 51750
    },
    {
      "epoch": 1.5465918251574955,
      "grad_norm": 3.666222333908081,
      "learning_rate": 0.00018520162972359843,
      "loss": 1.3552,
      "step": 51800
    },
    {
      "epoch": 1.5480846744095782,
      "grad_norm": 4.288083076477051,
      "learning_rate": 0.0001851873439276796,
      "loss": 1.3139,
      "step": 51850
    },
    {
      "epoch": 1.5495775236616607,
      "grad_norm": 6.1794939041137695,
      "learning_rate": 0.00018517305813176075,
      "loss": 1.2984,
      "step": 51900
    },
    {
      "epoch": 1.5510703729137432,
      "grad_norm": 4.89127779006958,
      "learning_rate": 0.00018515877233584195,
      "loss": 1.3251,
      "step": 51950
    },
    {
      "epoch": 1.5525632221658257,
      "grad_norm": 3.1737000942230225,
      "learning_rate": 0.00018514448653992308,
      "loss": 1.342,
      "step": 52000
    },
    {
      "epoch": 1.5540560714179081,
      "grad_norm": 5.618998050689697,
      "learning_rate": 0.00018513020074400428,
      "loss": 1.3234,
      "step": 52050
    },
    {
      "epoch": 1.5555489206699908,
      "grad_norm": 4.3047685623168945,
      "learning_rate": 0.0001851159149480854,
      "loss": 1.3305,
      "step": 52100
    },
    {
      "epoch": 1.557041769922073,
      "grad_norm": 5.245848178863525,
      "learning_rate": 0.0001851016291521666,
      "loss": 1.3052,
      "step": 52150
    },
    {
      "epoch": 1.5585346191741558,
      "grad_norm": 5.337231636047363,
      "learning_rate": 0.00018508734335624774,
      "loss": 1.3353,
      "step": 52200
    },
    {
      "epoch": 1.5600274684262383,
      "grad_norm": 4.924198627471924,
      "learning_rate": 0.00018507305756032893,
      "loss": 1.3022,
      "step": 52250
    },
    {
      "epoch": 1.5615203176783208,
      "grad_norm": 4.464673042297363,
      "learning_rate": 0.0001850587717644101,
      "loss": 1.3168,
      "step": 52300
    },
    {
      "epoch": 1.5630131669304035,
      "grad_norm": 3.7909131050109863,
      "learning_rate": 0.00018504448596849126,
      "loss": 1.3307,
      "step": 52350
    },
    {
      "epoch": 1.5645060161824857,
      "grad_norm": 4.331950664520264,
      "learning_rate": 0.00018503020017257243,
      "loss": 1.3574,
      "step": 52400
    },
    {
      "epoch": 1.5659988654345685,
      "grad_norm": 4.650835990905762,
      "learning_rate": 0.0001850159143766536,
      "loss": 1.3356,
      "step": 52450
    },
    {
      "epoch": 1.567491714686651,
      "grad_norm": 5.968230247497559,
      "learning_rate": 0.00018500162858073476,
      "loss": 1.3515,
      "step": 52500
    },
    {
      "epoch": 1.5689845639387334,
      "grad_norm": 3.9560632705688477,
      "learning_rate": 0.00018498734278481592,
      "loss": 1.3262,
      "step": 52550
    },
    {
      "epoch": 1.5704774131908161,
      "grad_norm": 3.6332058906555176,
      "learning_rate": 0.0001849730569888971,
      "loss": 1.3708,
      "step": 52600
    },
    {
      "epoch": 1.5719702624428984,
      "grad_norm": 5.171941757202148,
      "learning_rate": 0.00018495877119297825,
      "loss": 1.3201,
      "step": 52650
    },
    {
      "epoch": 1.573463111694981,
      "grad_norm": 4.531052589416504,
      "learning_rate": 0.00018494448539705942,
      "loss": 1.364,
      "step": 52700
    },
    {
      "epoch": 1.5749559609470636,
      "grad_norm": 3.860273838043213,
      "learning_rate": 0.0001849301996011406,
      "loss": 1.3729,
      "step": 52750
    },
    {
      "epoch": 1.576448810199146,
      "grad_norm": 4.562480926513672,
      "learning_rate": 0.00018491591380522175,
      "loss": 1.2745,
      "step": 52800
    },
    {
      "epoch": 1.5779416594512288,
      "grad_norm": 4.759164333343506,
      "learning_rate": 0.00018490162800930294,
      "loss": 1.3399,
      "step": 52850
    },
    {
      "epoch": 1.579434508703311,
      "grad_norm": 6.885429859161377,
      "learning_rate": 0.00018488734221338408,
      "loss": 1.2915,
      "step": 52900
    },
    {
      "epoch": 1.5809273579553937,
      "grad_norm": 4.555217742919922,
      "learning_rate": 0.00018487305641746527,
      "loss": 1.3897,
      "step": 52950
    },
    {
      "epoch": 1.5824202072074762,
      "grad_norm": 5.686909198760986,
      "learning_rate": 0.0001848587706215464,
      "loss": 1.3234,
      "step": 53000
    },
    {
      "epoch": 1.5839130564595587,
      "grad_norm": 3.994685411453247,
      "learning_rate": 0.00018484448482562757,
      "loss": 1.3265,
      "step": 53050
    },
    {
      "epoch": 1.5854059057116412,
      "grad_norm": 4.623023986816406,
      "learning_rate": 0.00018483019902970876,
      "loss": 1.3479,
      "step": 53100
    },
    {
      "epoch": 1.5868987549637237,
      "grad_norm": 6.000095844268799,
      "learning_rate": 0.0001848159132337899,
      "loss": 1.3333,
      "step": 53150
    },
    {
      "epoch": 1.5883916042158064,
      "grad_norm": 7.388947010040283,
      "learning_rate": 0.0001848016274378711,
      "loss": 1.306,
      "step": 53200
    },
    {
      "epoch": 1.5898844534678886,
      "grad_norm": 4.04311990737915,
      "learning_rate": 0.00018478734164195223,
      "loss": 1.3263,
      "step": 53250
    },
    {
      "epoch": 1.5913773027199714,
      "grad_norm": 5.370228290557861,
      "learning_rate": 0.00018477305584603342,
      "loss": 1.3279,
      "step": 53300
    },
    {
      "epoch": 1.5928701519720538,
      "grad_norm": 5.453476905822754,
      "learning_rate": 0.00018475877005011458,
      "loss": 1.3477,
      "step": 53350
    },
    {
      "epoch": 1.5943630012241363,
      "grad_norm": 3.9221408367156982,
      "learning_rate": 0.00018474448425419575,
      "loss": 1.3391,
      "step": 53400
    },
    {
      "epoch": 1.595855850476219,
      "grad_norm": 4.4245781898498535,
      "learning_rate": 0.0001847301984582769,
      "loss": 1.3263,
      "step": 53450
    },
    {
      "epoch": 1.5973486997283013,
      "grad_norm": 4.352360725402832,
      "learning_rate": 0.00018471591266235808,
      "loss": 1.3059,
      "step": 53500
    },
    {
      "epoch": 1.598841548980384,
      "grad_norm": 5.2914509773254395,
      "learning_rate": 0.00018470162686643924,
      "loss": 1.3353,
      "step": 53550
    },
    {
      "epoch": 1.6003343982324665,
      "grad_norm": 5.199918746948242,
      "learning_rate": 0.0001846873410705204,
      "loss": 1.3667,
      "step": 53600
    },
    {
      "epoch": 1.601827247484549,
      "grad_norm": 4.404350280761719,
      "learning_rate": 0.00018467305527460157,
      "loss": 1.3245,
      "step": 53650
    },
    {
      "epoch": 1.6033200967366317,
      "grad_norm": 4.452194690704346,
      "learning_rate": 0.00018465876947868274,
      "loss": 1.3194,
      "step": 53700
    },
    {
      "epoch": 1.604812945988714,
      "grad_norm": 4.387017726898193,
      "learning_rate": 0.0001846444836827639,
      "loss": 1.2448,
      "step": 53750
    },
    {
      "epoch": 1.6063057952407966,
      "grad_norm": 5.375933647155762,
      "learning_rate": 0.0001846301978868451,
      "loss": 1.3515,
      "step": 53800
    },
    {
      "epoch": 1.6077986444928791,
      "grad_norm": 4.564199447631836,
      "learning_rate": 0.00018461591209092623,
      "loss": 1.3182,
      "step": 53850
    },
    {
      "epoch": 1.6092914937449616,
      "grad_norm": 5.435900688171387,
      "learning_rate": 0.00018460162629500742,
      "loss": 1.3071,
      "step": 53900
    },
    {
      "epoch": 1.6107843429970443,
      "grad_norm": 3.600592613220215,
      "learning_rate": 0.00018458734049908856,
      "loss": 1.4009,
      "step": 53950
    },
    {
      "epoch": 1.6122771922491266,
      "grad_norm": 5.12654447555542,
      "learning_rate": 0.00018457305470316975,
      "loss": 1.3285,
      "step": 54000
    },
    {
      "epoch": 1.6137700415012093,
      "grad_norm": 3.6472108364105225,
      "learning_rate": 0.0001845587689072509,
      "loss": 1.32,
      "step": 54050
    },
    {
      "epoch": 1.6152628907532918,
      "grad_norm": 5.0967559814453125,
      "learning_rate": 0.00018454448311133208,
      "loss": 1.3018,
      "step": 54100
    },
    {
      "epoch": 1.6167557400053743,
      "grad_norm": 5.078010082244873,
      "learning_rate": 0.00018453019731541325,
      "loss": 1.3539,
      "step": 54150
    },
    {
      "epoch": 1.6182485892574567,
      "grad_norm": 4.601762294769287,
      "learning_rate": 0.0001845159115194944,
      "loss": 1.3153,
      "step": 54200
    },
    {
      "epoch": 1.6197414385095392,
      "grad_norm": 3.8307271003723145,
      "learning_rate": 0.00018450162572357557,
      "loss": 1.34,
      "step": 54250
    },
    {
      "epoch": 1.621234287761622,
      "grad_norm": 6.145284175872803,
      "learning_rate": 0.00018448733992765674,
      "loss": 1.3315,
      "step": 54300
    },
    {
      "epoch": 1.6227271370137042,
      "grad_norm": 9.96362590789795,
      "learning_rate": 0.0001844730541317379,
      "loss": 1.3269,
      "step": 54350
    },
    {
      "epoch": 1.624219986265787,
      "grad_norm": 5.126537322998047,
      "learning_rate": 0.00018445876833581907,
      "loss": 1.3604,
      "step": 54400
    },
    {
      "epoch": 1.6257128355178694,
      "grad_norm": 3.604433059692383,
      "learning_rate": 0.00018444448253990023,
      "loss": 1.3651,
      "step": 54450
    },
    {
      "epoch": 1.6272056847699519,
      "grad_norm": 4.835113048553467,
      "learning_rate": 0.0001844301967439814,
      "loss": 1.2787,
      "step": 54500
    },
    {
      "epoch": 1.6286985340220346,
      "grad_norm": 4.534168243408203,
      "learning_rate": 0.00018441591094806256,
      "loss": 1.3354,
      "step": 54550
    },
    {
      "epoch": 1.6301913832741168,
      "grad_norm": 7.929208755493164,
      "learning_rate": 0.00018440162515214375,
      "loss": 1.3361,
      "step": 54600
    },
    {
      "epoch": 1.6316842325261995,
      "grad_norm": 4.582381248474121,
      "learning_rate": 0.0001843873393562249,
      "loss": 1.354,
      "step": 54650
    },
    {
      "epoch": 1.633177081778282,
      "grad_norm": 4.781309604644775,
      "learning_rate": 0.00018437305356030608,
      "loss": 1.3035,
      "step": 54700
    },
    {
      "epoch": 1.6346699310303645,
      "grad_norm": 6.0308308601379395,
      "learning_rate": 0.00018435876776438722,
      "loss": 1.3392,
      "step": 54750
    },
    {
      "epoch": 1.6361627802824472,
      "grad_norm": 5.485566139221191,
      "learning_rate": 0.0001843444819684684,
      "loss": 1.3408,
      "step": 54800
    },
    {
      "epoch": 1.6376556295345295,
      "grad_norm": 6.658644676208496,
      "learning_rate": 0.00018433019617254955,
      "loss": 1.3772,
      "step": 54850
    },
    {
      "epoch": 1.6391484787866122,
      "grad_norm": 4.54780912399292,
      "learning_rate": 0.00018431591037663074,
      "loss": 1.3511,
      "step": 54900
    },
    {
      "epoch": 1.6406413280386947,
      "grad_norm": 5.600758075714111,
      "learning_rate": 0.0001843016245807119,
      "loss": 1.3301,
      "step": 54950
    },
    {
      "epoch": 1.6421341772907772,
      "grad_norm": 4.194091796875,
      "learning_rate": 0.00018428733878479307,
      "loss": 1.3159,
      "step": 55000
    },
    {
      "epoch": 1.6436270265428599,
      "grad_norm": 6.306137561798096,
      "learning_rate": 0.00018427305298887424,
      "loss": 1.3401,
      "step": 55050
    },
    {
      "epoch": 1.6451198757949421,
      "grad_norm": 4.377103328704834,
      "learning_rate": 0.0001842587671929554,
      "loss": 1.34,
      "step": 55100
    },
    {
      "epoch": 1.6466127250470248,
      "grad_norm": 3.450533390045166,
      "learning_rate": 0.00018424448139703657,
      "loss": 1.3367,
      "step": 55150
    },
    {
      "epoch": 1.6481055742991073,
      "grad_norm": 3.5845348834991455,
      "learning_rate": 0.00018423019560111773,
      "loss": 1.3755,
      "step": 55200
    },
    {
      "epoch": 1.6495984235511898,
      "grad_norm": 4.741848468780518,
      "learning_rate": 0.0001842159098051989,
      "loss": 1.3116,
      "step": 55250
    },
    {
      "epoch": 1.6510912728032723,
      "grad_norm": 4.1155619621276855,
      "learning_rate": 0.00018420162400928006,
      "loss": 1.3682,
      "step": 55300
    },
    {
      "epoch": 1.6525841220553548,
      "grad_norm": 4.3722028732299805,
      "learning_rate": 0.00018418733821336122,
      "loss": 1.2916,
      "step": 55350
    },
    {
      "epoch": 1.6540769713074375,
      "grad_norm": 4.6453118324279785,
      "learning_rate": 0.00018417305241744242,
      "loss": 1.3743,
      "step": 55400
    },
    {
      "epoch": 1.6555698205595197,
      "grad_norm": 4.749643802642822,
      "learning_rate": 0.00018415876662152355,
      "loss": 1.3649,
      "step": 55450
    },
    {
      "epoch": 1.6570626698116024,
      "grad_norm": 3.9023261070251465,
      "learning_rate": 0.00018414448082560475,
      "loss": 1.4002,
      "step": 55500
    },
    {
      "epoch": 1.658555519063685,
      "grad_norm": 4.34132719039917,
      "learning_rate": 0.00018413019502968588,
      "loss": 1.3344,
      "step": 55550
    },
    {
      "epoch": 1.6600483683157674,
      "grad_norm": 5.479841709136963,
      "learning_rate": 0.00018411590923376707,
      "loss": 1.3511,
      "step": 55600
    },
    {
      "epoch": 1.6615412175678501,
      "grad_norm": 3.340822696685791,
      "learning_rate": 0.0001841016234378482,
      "loss": 1.3789,
      "step": 55650
    },
    {
      "epoch": 1.6630340668199324,
      "grad_norm": 5.259148597717285,
      "learning_rate": 0.00018408733764192938,
      "loss": 1.3092,
      "step": 55700
    },
    {
      "epoch": 1.664526916072015,
      "grad_norm": 3.590052843093872,
      "learning_rate": 0.00018407305184601057,
      "loss": 1.3596,
      "step": 55750
    },
    {
      "epoch": 1.6660197653240976,
      "grad_norm": 5.688297748565674,
      "learning_rate": 0.0001840587660500917,
      "loss": 1.3629,
      "step": 55800
    },
    {
      "epoch": 1.66751261457618,
      "grad_norm": 5.826660633087158,
      "learning_rate": 0.0001840444802541729,
      "loss": 1.3757,
      "step": 55850
    },
    {
      "epoch": 1.6690054638282628,
      "grad_norm": 3.0926566123962402,
      "learning_rate": 0.00018403019445825404,
      "loss": 1.3507,
      "step": 55900
    },
    {
      "epoch": 1.670498313080345,
      "grad_norm": 7.726663589477539,
      "learning_rate": 0.00018401590866233523,
      "loss": 1.3162,
      "step": 55950
    },
    {
      "epoch": 1.6719911623324277,
      "grad_norm": 4.687432765960693,
      "learning_rate": 0.0001840016228664164,
      "loss": 1.289,
      "step": 56000
    },
    {
      "epoch": 1.6734840115845102,
      "grad_norm": 4.279799461364746,
      "learning_rate": 0.00018398733707049756,
      "loss": 1.3438,
      "step": 56050
    },
    {
      "epoch": 1.6749768608365927,
      "grad_norm": 6.4994797706604,
      "learning_rate": 0.00018397305127457872,
      "loss": 1.3373,
      "step": 56100
    },
    {
      "epoch": 1.6764697100886754,
      "grad_norm": 4.511581897735596,
      "learning_rate": 0.00018395876547865989,
      "loss": 1.3549,
      "step": 56150
    },
    {
      "epoch": 1.6779625593407577,
      "grad_norm": 5.1127190589904785,
      "learning_rate": 0.00018394447968274105,
      "loss": 1.2734,
      "step": 56200
    },
    {
      "epoch": 1.6794554085928404,
      "grad_norm": 4.7369561195373535,
      "learning_rate": 0.00018393019388682222,
      "loss": 1.3193,
      "step": 56250
    },
    {
      "epoch": 1.6809482578449229,
      "grad_norm": 5.101438522338867,
      "learning_rate": 0.00018391590809090338,
      "loss": 1.3311,
      "step": 56300
    },
    {
      "epoch": 1.6824411070970053,
      "grad_norm": 5.68798303604126,
      "learning_rate": 0.00018390162229498454,
      "loss": 1.3723,
      "step": 56350
    },
    {
      "epoch": 1.6839339563490878,
      "grad_norm": 5.620144367218018,
      "learning_rate": 0.0001838873364990657,
      "loss": 1.3051,
      "step": 56400
    },
    {
      "epoch": 1.6854268056011703,
      "grad_norm": 4.335910797119141,
      "learning_rate": 0.00018387305070314687,
      "loss": 1.2625,
      "step": 56450
    },
    {
      "epoch": 1.686919654853253,
      "grad_norm": 5.112996578216553,
      "learning_rate": 0.00018385876490722804,
      "loss": 1.3288,
      "step": 56500
    },
    {
      "epoch": 1.6884125041053353,
      "grad_norm": 5.485559463500977,
      "learning_rate": 0.00018384447911130923,
      "loss": 1.3171,
      "step": 56550
    },
    {
      "epoch": 1.689905353357418,
      "grad_norm": 5.58554744720459,
      "learning_rate": 0.00018383019331539037,
      "loss": 1.3002,
      "step": 56600
    },
    {
      "epoch": 1.6913982026095005,
      "grad_norm": 4.202605247497559,
      "learning_rate": 0.00018381590751947156,
      "loss": 1.3018,
      "step": 56650
    },
    {
      "epoch": 1.692891051861583,
      "grad_norm": 5.206381797790527,
      "learning_rate": 0.0001838016217235527,
      "loss": 1.3696,
      "step": 56700
    },
    {
      "epoch": 1.6943839011136657,
      "grad_norm": 5.87765645980835,
      "learning_rate": 0.0001837873359276339,
      "loss": 1.3616,
      "step": 56750
    },
    {
      "epoch": 1.695876750365748,
      "grad_norm": 3.4311394691467285,
      "learning_rate": 0.00018377305013171505,
      "loss": 1.3249,
      "step": 56800
    },
    {
      "epoch": 1.6973695996178306,
      "grad_norm": 2.7731411457061768,
      "learning_rate": 0.00018375876433579622,
      "loss": 1.3715,
      "step": 56850
    },
    {
      "epoch": 1.698862448869913,
      "grad_norm": 6.399103164672852,
      "learning_rate": 0.00018374447853987738,
      "loss": 1.3654,
      "step": 56900
    },
    {
      "epoch": 1.7003552981219956,
      "grad_norm": 3.8202717304229736,
      "learning_rate": 0.00018373019274395855,
      "loss": 1.3565,
      "step": 56950
    },
    {
      "epoch": 1.7018481473740783,
      "grad_norm": 4.040468215942383,
      "learning_rate": 0.0001837159069480397,
      "loss": 1.3373,
      "step": 57000
    },
    {
      "epoch": 1.7033409966261606,
      "grad_norm": 4.522890090942383,
      "learning_rate": 0.00018370162115212088,
      "loss": 1.3393,
      "step": 57050
    },
    {
      "epoch": 1.7048338458782433,
      "grad_norm": 4.771418571472168,
      "learning_rate": 0.00018368733535620204,
      "loss": 1.2437,
      "step": 57100
    },
    {
      "epoch": 1.7063266951303258,
      "grad_norm": 4.333567142486572,
      "learning_rate": 0.0001836730495602832,
      "loss": 1.2876,
      "step": 57150
    },
    {
      "epoch": 1.7078195443824082,
      "grad_norm": 5.1498703956604,
      "learning_rate": 0.00018365876376436437,
      "loss": 1.4321,
      "step": 57200
    },
    {
      "epoch": 1.709312393634491,
      "grad_norm": 3.8978734016418457,
      "learning_rate": 0.00018364447796844556,
      "loss": 1.2815,
      "step": 57250
    },
    {
      "epoch": 1.7108052428865732,
      "grad_norm": 5.365299224853516,
      "learning_rate": 0.0001836301921725267,
      "loss": 1.2443,
      "step": 57300
    },
    {
      "epoch": 1.712298092138656,
      "grad_norm": 4.474836826324463,
      "learning_rate": 0.0001836159063766079,
      "loss": 1.3634,
      "step": 57350
    },
    {
      "epoch": 1.7137909413907384,
      "grad_norm": 3.8736836910247803,
      "learning_rate": 0.00018360162058068903,
      "loss": 1.3889,
      "step": 57400
    },
    {
      "epoch": 1.7152837906428209,
      "grad_norm": 4.919806957244873,
      "learning_rate": 0.00018358733478477022,
      "loss": 1.3216,
      "step": 57450
    },
    {
      "epoch": 1.7167766398949034,
      "grad_norm": 6.2346954345703125,
      "learning_rate": 0.00018357304898885136,
      "loss": 1.3294,
      "step": 57500
    },
    {
      "epoch": 1.7182694891469859,
      "grad_norm": 4.782268524169922,
      "learning_rate": 0.00018355876319293255,
      "loss": 1.363,
      "step": 57550
    },
    {
      "epoch": 1.7197623383990686,
      "grad_norm": 5.770125389099121,
      "learning_rate": 0.00018354447739701372,
      "loss": 1.3441,
      "step": 57600
    },
    {
      "epoch": 1.721255187651151,
      "grad_norm": 5.742440223693848,
      "learning_rate": 0.00018353019160109488,
      "loss": 1.302,
      "step": 57650
    },
    {
      "epoch": 1.7227480369032335,
      "grad_norm": 5.430489540100098,
      "learning_rate": 0.00018351590580517604,
      "loss": 1.3183,
      "step": 57700
    },
    {
      "epoch": 1.724240886155316,
      "grad_norm": 6.17573356628418,
      "learning_rate": 0.0001835016200092572,
      "loss": 1.3421,
      "step": 57750
    },
    {
      "epoch": 1.7257337354073985,
      "grad_norm": 5.362539291381836,
      "learning_rate": 0.00018348733421333837,
      "loss": 1.3333,
      "step": 57800
    },
    {
      "epoch": 1.7272265846594812,
      "grad_norm": 4.4574294090271,
      "learning_rate": 0.00018347304841741954,
      "loss": 1.2792,
      "step": 57850
    },
    {
      "epoch": 1.7287194339115635,
      "grad_norm": 4.277164936065674,
      "learning_rate": 0.0001834587626215007,
      "loss": 1.342,
      "step": 57900
    },
    {
      "epoch": 1.7302122831636462,
      "grad_norm": 4.775753498077393,
      "learning_rate": 0.00018344447682558187,
      "loss": 1.3618,
      "step": 57950
    },
    {
      "epoch": 1.7317051324157287,
      "grad_norm": 4.7877702713012695,
      "learning_rate": 0.00018343019102966303,
      "loss": 1.3631,
      "step": 58000
    },
    {
      "epoch": 1.7331979816678111,
      "grad_norm": 5.616472244262695,
      "learning_rate": 0.00018341590523374422,
      "loss": 1.3558,
      "step": 58050
    },
    {
      "epoch": 1.7346908309198938,
      "grad_norm": 5.436590671539307,
      "learning_rate": 0.00018340161943782536,
      "loss": 1.3999,
      "step": 58100
    },
    {
      "epoch": 1.736183680171976,
      "grad_norm": 4.245835304260254,
      "learning_rate": 0.00018338733364190655,
      "loss": 1.3592,
      "step": 58150
    },
    {
      "epoch": 1.7376765294240588,
      "grad_norm": 4.440322399139404,
      "learning_rate": 0.0001833730478459877,
      "loss": 1.312,
      "step": 58200
    },
    {
      "epoch": 1.7391693786761413,
      "grad_norm": 4.439347743988037,
      "learning_rate": 0.00018335876205006888,
      "loss": 1.3769,
      "step": 58250
    },
    {
      "epoch": 1.7406622279282238,
      "grad_norm": 4.374370098114014,
      "learning_rate": 0.00018334447625415002,
      "loss": 1.3328,
      "step": 58300
    },
    {
      "epoch": 1.7421550771803065,
      "grad_norm": 3.9568932056427,
      "learning_rate": 0.00018333019045823119,
      "loss": 1.2992,
      "step": 58350
    },
    {
      "epoch": 1.7436479264323888,
      "grad_norm": 4.133091926574707,
      "learning_rate": 0.00018331590466231238,
      "loss": 1.3685,
      "step": 58400
    },
    {
      "epoch": 1.7451407756844715,
      "grad_norm": 4.492006778717041,
      "learning_rate": 0.00018330161886639351,
      "loss": 1.3543,
      "step": 58450
    },
    {
      "epoch": 1.746633624936554,
      "grad_norm": 2.6073009967803955,
      "learning_rate": 0.0001832873330704747,
      "loss": 1.36,
      "step": 58500
    },
    {
      "epoch": 1.7481264741886364,
      "grad_norm": 4.137593746185303,
      "learning_rate": 0.00018327304727455584,
      "loss": 1.3599,
      "step": 58550
    },
    {
      "epoch": 1.749619323440719,
      "grad_norm": 4.932790756225586,
      "learning_rate": 0.00018325876147863704,
      "loss": 1.3265,
      "step": 58600
    },
    {
      "epoch": 1.7511121726928014,
      "grad_norm": 4.623594760894775,
      "learning_rate": 0.00018324447568271817,
      "loss": 1.3186,
      "step": 58650
    },
    {
      "epoch": 1.752605021944884,
      "grad_norm": 4.066229343414307,
      "learning_rate": 0.00018323018988679936,
      "loss": 1.3196,
      "step": 58700
    },
    {
      "epoch": 1.7540978711969666,
      "grad_norm": 4.105048656463623,
      "learning_rate": 0.00018321590409088053,
      "loss": 1.3663,
      "step": 58750
    },
    {
      "epoch": 1.755590720449049,
      "grad_norm": 4.454251766204834,
      "learning_rate": 0.0001832016182949617,
      "loss": 1.2823,
      "step": 58800
    },
    {
      "epoch": 1.7570835697011316,
      "grad_norm": 4.3030242919921875,
      "learning_rate": 0.00018318733249904286,
      "loss": 1.3267,
      "step": 58850
    },
    {
      "epoch": 1.758576418953214,
      "grad_norm": 5.350499629974365,
      "learning_rate": 0.00018317304670312402,
      "loss": 1.2992,
      "step": 58900
    },
    {
      "epoch": 1.7600692682052967,
      "grad_norm": 4.497105598449707,
      "learning_rate": 0.0001831587609072052,
      "loss": 1.2759,
      "step": 58950
    },
    {
      "epoch": 1.761562117457379,
      "grad_norm": 4.683063983917236,
      "learning_rate": 0.00018314447511128635,
      "loss": 1.3342,
      "step": 59000
    },
    {
      "epoch": 1.7630549667094617,
      "grad_norm": 3.4937946796417236,
      "learning_rate": 0.00018313018931536752,
      "loss": 1.2969,
      "step": 59050
    },
    {
      "epoch": 1.7645478159615442,
      "grad_norm": 5.422043800354004,
      "learning_rate": 0.00018311590351944868,
      "loss": 1.3116,
      "step": 59100
    },
    {
      "epoch": 1.7660406652136267,
      "grad_norm": 3.8882923126220703,
      "learning_rate": 0.00018310161772352985,
      "loss": 1.3073,
      "step": 59150
    },
    {
      "epoch": 1.7675335144657094,
      "grad_norm": 5.60578727722168,
      "learning_rate": 0.00018308733192761104,
      "loss": 1.31,
      "step": 59200
    },
    {
      "epoch": 1.7690263637177917,
      "grad_norm": 5.257917881011963,
      "learning_rate": 0.00018307304613169218,
      "loss": 1.3617,
      "step": 59250
    },
    {
      "epoch": 1.7705192129698744,
      "grad_norm": 4.838101863861084,
      "learning_rate": 0.00018305876033577337,
      "loss": 1.367,
      "step": 59300
    },
    {
      "epoch": 1.7720120622219568,
      "grad_norm": 4.63686990737915,
      "learning_rate": 0.0001830444745398545,
      "loss": 1.3653,
      "step": 59350
    },
    {
      "epoch": 1.7735049114740393,
      "grad_norm": 3.8766531944274902,
      "learning_rate": 0.0001830301887439357,
      "loss": 1.298,
      "step": 59400
    },
    {
      "epoch": 1.774997760726122,
      "grad_norm": 5.356032848358154,
      "learning_rate": 0.00018301590294801686,
      "loss": 1.3132,
      "step": 59450
    },
    {
      "epoch": 1.7764906099782043,
      "grad_norm": 4.647774696350098,
      "learning_rate": 0.00018300161715209803,
      "loss": 1.3594,
      "step": 59500
    },
    {
      "epoch": 1.777983459230287,
      "grad_norm": 3.6505117416381836,
      "learning_rate": 0.0001829873313561792,
      "loss": 1.3178,
      "step": 59550
    },
    {
      "epoch": 1.7794763084823695,
      "grad_norm": 4.109731674194336,
      "learning_rate": 0.00018297304556026036,
      "loss": 1.2996,
      "step": 59600
    },
    {
      "epoch": 1.780969157734452,
      "grad_norm": 3.942424774169922,
      "learning_rate": 0.00018295875976434152,
      "loss": 1.3066,
      "step": 59650
    },
    {
      "epoch": 1.7824620069865345,
      "grad_norm": 5.511881351470947,
      "learning_rate": 0.00018294447396842268,
      "loss": 1.3009,
      "step": 59700
    },
    {
      "epoch": 1.783954856238617,
      "grad_norm": 4.0874552726745605,
      "learning_rate": 0.00018293018817250385,
      "loss": 1.3426,
      "step": 59750
    },
    {
      "epoch": 1.7854477054906996,
      "grad_norm": 4.787434101104736,
      "learning_rate": 0.00018291590237658501,
      "loss": 1.3765,
      "step": 59800
    },
    {
      "epoch": 1.7869405547427821,
      "grad_norm": 4.477612018585205,
      "learning_rate": 0.00018290161658066618,
      "loss": 1.3215,
      "step": 59850
    },
    {
      "epoch": 1.7884334039948646,
      "grad_norm": 6.102437973022461,
      "learning_rate": 0.00018288733078474737,
      "loss": 1.2529,
      "step": 59900
    },
    {
      "epoch": 1.789926253246947,
      "grad_norm": 4.351520538330078,
      "learning_rate": 0.0001828730449888285,
      "loss": 1.2889,
      "step": 59950
    },
    {
      "epoch": 1.7914191024990296,
      "grad_norm": 5.043869972229004,
      "learning_rate": 0.0001828587591929097,
      "loss": 1.3149,
      "step": 60000
    },
    {
      "epoch": 1.7929119517511123,
      "grad_norm": 4.172428607940674,
      "learning_rate": 0.00018284447339699084,
      "loss": 1.2703,
      "step": 60050
    },
    {
      "epoch": 1.7944048010031945,
      "grad_norm": 5.1870903968811035,
      "learning_rate": 0.00018283018760107203,
      "loss": 1.3107,
      "step": 60100
    },
    {
      "epoch": 1.7958976502552773,
      "grad_norm": 4.141382694244385,
      "learning_rate": 0.00018281590180515317,
      "loss": 1.3252,
      "step": 60150
    },
    {
      "epoch": 1.7973904995073597,
      "grad_norm": 3.9119277000427246,
      "learning_rate": 0.00018280161600923436,
      "loss": 1.3875,
      "step": 60200
    },
    {
      "epoch": 1.7988833487594422,
      "grad_norm": 4.159456729888916,
      "learning_rate": 0.00018278733021331552,
      "loss": 1.3871,
      "step": 60250
    },
    {
      "epoch": 1.800376198011525,
      "grad_norm": 4.536618709564209,
      "learning_rate": 0.0001827730444173967,
      "loss": 1.2892,
      "step": 60300
    },
    {
      "epoch": 1.8018690472636072,
      "grad_norm": 7.839147090911865,
      "learning_rate": 0.00018275875862147785,
      "loss": 1.2953,
      "step": 60350
    },
    {
      "epoch": 1.80336189651569,
      "grad_norm": 3.6673262119293213,
      "learning_rate": 0.00018274447282555902,
      "loss": 1.3245,
      "step": 60400
    },
    {
      "epoch": 1.8048547457677724,
      "grad_norm": 4.275418281555176,
      "learning_rate": 0.00018273018702964018,
      "loss": 1.3061,
      "step": 60450
    },
    {
      "epoch": 1.8063475950198549,
      "grad_norm": 4.154748916625977,
      "learning_rate": 0.00018271590123372135,
      "loss": 1.3127,
      "step": 60500
    },
    {
      "epoch": 1.8078404442719376,
      "grad_norm": 4.65784215927124,
      "learning_rate": 0.0001827016154378025,
      "loss": 1.3717,
      "step": 60550
    },
    {
      "epoch": 1.8093332935240198,
      "grad_norm": 6.874159812927246,
      "learning_rate": 0.00018268732964188368,
      "loss": 1.3043,
      "step": 60600
    },
    {
      "epoch": 1.8108261427761025,
      "grad_norm": 5.431716442108154,
      "learning_rate": 0.00018267304384596484,
      "loss": 1.3138,
      "step": 60650
    },
    {
      "epoch": 1.812318992028185,
      "grad_norm": 4.631954669952393,
      "learning_rate": 0.00018265875805004603,
      "loss": 1.3241,
      "step": 60700
    },
    {
      "epoch": 1.8138118412802675,
      "grad_norm": 4.336903095245361,
      "learning_rate": 0.00018264447225412717,
      "loss": 1.2709,
      "step": 60750
    },
    {
      "epoch": 1.81530469053235,
      "grad_norm": 4.103146553039551,
      "learning_rate": 0.00018263018645820836,
      "loss": 1.2912,
      "step": 60800
    },
    {
      "epoch": 1.8167975397844325,
      "grad_norm": 4.709009170532227,
      "learning_rate": 0.0001826159006622895,
      "loss": 1.304,
      "step": 60850
    },
    {
      "epoch": 1.8182903890365152,
      "grad_norm": 5.0871148109436035,
      "learning_rate": 0.00018260161486637066,
      "loss": 1.301,
      "step": 60900
    },
    {
      "epoch": 1.8197832382885977,
      "grad_norm": 4.265575408935547,
      "learning_rate": 0.00018258732907045183,
      "loss": 1.3914,
      "step": 60950
    },
    {
      "epoch": 1.8212760875406802,
      "grad_norm": 4.705594539642334,
      "learning_rate": 0.000182573043274533,
      "loss": 1.3056,
      "step": 61000
    },
    {
      "epoch": 1.8227689367927626,
      "grad_norm": 3.86602783203125,
      "learning_rate": 0.00018255875747861418,
      "loss": 1.3394,
      "step": 61050
    },
    {
      "epoch": 1.8242617860448451,
      "grad_norm": 3.9397497177124023,
      "learning_rate": 0.00018254447168269532,
      "loss": 1.3625,
      "step": 61100
    },
    {
      "epoch": 1.8257546352969278,
      "grad_norm": 4.313930988311768,
      "learning_rate": 0.00018253018588677651,
      "loss": 1.3445,
      "step": 61150
    },
    {
      "epoch": 1.82724748454901,
      "grad_norm": 5.848763942718506,
      "learning_rate": 0.00018251590009085765,
      "loss": 1.3896,
      "step": 61200
    },
    {
      "epoch": 1.8287403338010928,
      "grad_norm": 4.323721885681152,
      "learning_rate": 0.00018250161429493884,
      "loss": 1.3616,
      "step": 61250
    },
    {
      "epoch": 1.8302331830531753,
      "grad_norm": 5.140015125274658,
      "learning_rate": 0.00018248732849901998,
      "loss": 1.2641,
      "step": 61300
    },
    {
      "epoch": 1.8317260323052578,
      "grad_norm": 5.685877799987793,
      "learning_rate": 0.00018247304270310117,
      "loss": 1.3396,
      "step": 61350
    },
    {
      "epoch": 1.8332188815573405,
      "grad_norm": 4.268862724304199,
      "learning_rate": 0.00018245875690718234,
      "loss": 1.3201,
      "step": 61400
    },
    {
      "epoch": 1.8347117308094227,
      "grad_norm": 5.0206732749938965,
      "learning_rate": 0.0001824444711112635,
      "loss": 1.372,
      "step": 61450
    },
    {
      "epoch": 1.8362045800615054,
      "grad_norm": 4.32529354095459,
      "learning_rate": 0.00018243018531534467,
      "loss": 1.3,
      "step": 61500
    },
    {
      "epoch": 1.837697429313588,
      "grad_norm": 5.349672317504883,
      "learning_rate": 0.00018241589951942583,
      "loss": 1.2957,
      "step": 61550
    },
    {
      "epoch": 1.8391902785656704,
      "grad_norm": 4.735100746154785,
      "learning_rate": 0.000182401613723507,
      "loss": 1.3546,
      "step": 61600
    },
    {
      "epoch": 1.8406831278177531,
      "grad_norm": 4.489027500152588,
      "learning_rate": 0.00018238732792758816,
      "loss": 1.2695,
      "step": 61650
    },
    {
      "epoch": 1.8421759770698354,
      "grad_norm": 3.2674996852874756,
      "learning_rate": 0.00018237304213166933,
      "loss": 1.3425,
      "step": 61700
    },
    {
      "epoch": 1.843668826321918,
      "grad_norm": 3.9815094470977783,
      "learning_rate": 0.0001823587563357505,
      "loss": 1.3396,
      "step": 61750
    },
    {
      "epoch": 1.8451616755740006,
      "grad_norm": 3.9253129959106445,
      "learning_rate": 0.00018234447053983165,
      "loss": 1.3641,
      "step": 61800
    },
    {
      "epoch": 1.846654524826083,
      "grad_norm": 6.536600112915039,
      "learning_rate": 0.00018233018474391285,
      "loss": 1.2578,
      "step": 61850
    },
    {
      "epoch": 1.8481473740781655,
      "grad_norm": 4.420536518096924,
      "learning_rate": 0.00018231589894799398,
      "loss": 1.3635,
      "step": 61900
    },
    {
      "epoch": 1.849640223330248,
      "grad_norm": 3.770279884338379,
      "learning_rate": 0.00018230161315207518,
      "loss": 1.2869,
      "step": 61950
    },
    {
      "epoch": 1.8511330725823307,
      "grad_norm": 4.154125213623047,
      "learning_rate": 0.0001822873273561563,
      "loss": 1.3246,
      "step": 62000
    },
    {
      "epoch": 1.8526259218344132,
      "grad_norm": 3.884587049484253,
      "learning_rate": 0.0001822730415602375,
      "loss": 1.329,
      "step": 62050
    },
    {
      "epoch": 1.8541187710864957,
      "grad_norm": 5.329457759857178,
      "learning_rate": 0.00018225875576431864,
      "loss": 1.3787,
      "step": 62100
    },
    {
      "epoch": 1.8556116203385782,
      "grad_norm": 5.966410160064697,
      "learning_rate": 0.00018224446996839983,
      "loss": 1.3574,
      "step": 62150
    },
    {
      "epoch": 1.8571044695906607,
      "grad_norm": 5.356328010559082,
      "learning_rate": 0.000182230184172481,
      "loss": 1.3113,
      "step": 62200
    },
    {
      "epoch": 1.8585973188427434,
      "grad_norm": 3.6018612384796143,
      "learning_rate": 0.00018221589837656216,
      "loss": 1.3296,
      "step": 62250
    },
    {
      "epoch": 1.8600901680948256,
      "grad_norm": 5.247616767883301,
      "learning_rate": 0.00018220161258064333,
      "loss": 1.3518,
      "step": 62300
    },
    {
      "epoch": 1.8615830173469083,
      "grad_norm": 5.828862190246582,
      "learning_rate": 0.0001821873267847245,
      "loss": 1.3027,
      "step": 62350
    },
    {
      "epoch": 1.8630758665989908,
      "grad_norm": 5.451898574829102,
      "learning_rate": 0.00018217304098880566,
      "loss": 1.3647,
      "step": 62400
    },
    {
      "epoch": 1.8645687158510733,
      "grad_norm": 3.905299663543701,
      "learning_rate": 0.00018215875519288682,
      "loss": 1.3528,
      "step": 62450
    },
    {
      "epoch": 1.866061565103156,
      "grad_norm": 4.312819004058838,
      "learning_rate": 0.000182144469396968,
      "loss": 1.3487,
      "step": 62500
    },
    {
      "epoch": 1.8675544143552383,
      "grad_norm": 4.3358073234558105,
      "learning_rate": 0.00018213018360104915,
      "loss": 1.2925,
      "step": 62550
    },
    {
      "epoch": 1.869047263607321,
      "grad_norm": 5.664048194885254,
      "learning_rate": 0.00018211589780513032,
      "loss": 1.3822,
      "step": 62600
    },
    {
      "epoch": 1.8705401128594035,
      "grad_norm": 4.992645263671875,
      "learning_rate": 0.0001821016120092115,
      "loss": 1.3819,
      "step": 62650
    },
    {
      "epoch": 1.872032962111486,
      "grad_norm": 4.036514759063721,
      "learning_rate": 0.00018208732621329265,
      "loss": 1.3193,
      "step": 62700
    },
    {
      "epoch": 1.8735258113635687,
      "grad_norm": 4.094419956207275,
      "learning_rate": 0.00018207304041737384,
      "loss": 1.3649,
      "step": 62750
    },
    {
      "epoch": 1.875018660615651,
      "grad_norm": 5.437168121337891,
      "learning_rate": 0.00018205875462145497,
      "loss": 1.3445,
      "step": 62800
    },
    {
      "epoch": 1.8765115098677336,
      "grad_norm": 4.937995433807373,
      "learning_rate": 0.00018204446882553617,
      "loss": 1.3588,
      "step": 62850
    },
    {
      "epoch": 1.8780043591198161,
      "grad_norm": 3.918588876724243,
      "learning_rate": 0.00018203018302961733,
      "loss": 1.3195,
      "step": 62900
    },
    {
      "epoch": 1.8794972083718986,
      "grad_norm": 4.650214672088623,
      "learning_rate": 0.0001820158972336985,
      "loss": 1.3812,
      "step": 62950
    },
    {
      "epoch": 1.880990057623981,
      "grad_norm": 3.844249725341797,
      "learning_rate": 0.00018200161143777966,
      "loss": 1.2877,
      "step": 63000
    },
    {
      "epoch": 1.8824829068760636,
      "grad_norm": 4.988340854644775,
      "learning_rate": 0.00018198732564186083,
      "loss": 1.2705,
      "step": 63050
    },
    {
      "epoch": 1.8839757561281463,
      "grad_norm": 4.229214668273926,
      "learning_rate": 0.000181973039845942,
      "loss": 1.3314,
      "step": 63100
    },
    {
      "epoch": 1.8854686053802288,
      "grad_norm": 4.095324993133545,
      "learning_rate": 0.00018195875405002315,
      "loss": 1.3779,
      "step": 63150
    },
    {
      "epoch": 1.8869614546323112,
      "grad_norm": 5.325993537902832,
      "learning_rate": 0.00018194446825410432,
      "loss": 1.3163,
      "step": 63200
    },
    {
      "epoch": 1.8884543038843937,
      "grad_norm": 5.301263809204102,
      "learning_rate": 0.00018193018245818548,
      "loss": 1.4235,
      "step": 63250
    },
    {
      "epoch": 1.8899471531364762,
      "grad_norm": 5.347028732299805,
      "learning_rate": 0.00018191589666226665,
      "loss": 1.356,
      "step": 63300
    },
    {
      "epoch": 1.891440002388559,
      "grad_norm": 4.957263469696045,
      "learning_rate": 0.00018190161086634784,
      "loss": 1.2915,
      "step": 63350
    },
    {
      "epoch": 1.8929328516406412,
      "grad_norm": 3.9531173706054688,
      "learning_rate": 0.00018188732507042898,
      "loss": 1.2905,
      "step": 63400
    },
    {
      "epoch": 1.8944257008927239,
      "grad_norm": 5.983827590942383,
      "learning_rate": 0.00018187303927451017,
      "loss": 1.3129,
      "step": 63450
    },
    {
      "epoch": 1.8959185501448064,
      "grad_norm": 4.4479522705078125,
      "learning_rate": 0.0001818587534785913,
      "loss": 1.3528,
      "step": 63500
    },
    {
      "epoch": 1.8974113993968889,
      "grad_norm": 4.956698417663574,
      "learning_rate": 0.00018184446768267247,
      "loss": 1.2675,
      "step": 63550
    },
    {
      "epoch": 1.8989042486489716,
      "grad_norm": 3.7053678035736084,
      "learning_rate": 0.00018183018188675364,
      "loss": 1.3316,
      "step": 63600
    },
    {
      "epoch": 1.9003970979010538,
      "grad_norm": 5.076517105102539,
      "learning_rate": 0.0001818158960908348,
      "loss": 1.3603,
      "step": 63650
    },
    {
      "epoch": 1.9018899471531365,
      "grad_norm": 4.377649307250977,
      "learning_rate": 0.000181801610294916,
      "loss": 1.3491,
      "step": 63700
    },
    {
      "epoch": 1.903382796405219,
      "grad_norm": 5.296698570251465,
      "learning_rate": 0.00018178732449899713,
      "loss": 1.3768,
      "step": 63750
    },
    {
      "epoch": 1.9048756456573015,
      "grad_norm": 4.674185752868652,
      "learning_rate": 0.00018177303870307832,
      "loss": 1.3599,
      "step": 63800
    },
    {
      "epoch": 1.9063684949093842,
      "grad_norm": 5.758728981018066,
      "learning_rate": 0.00018175875290715946,
      "loss": 1.3239,
      "step": 63850
    },
    {
      "epoch": 1.9078613441614665,
      "grad_norm": 4.195470333099365,
      "learning_rate": 0.00018174446711124065,
      "loss": 1.3132,
      "step": 63900
    },
    {
      "epoch": 1.9093541934135492,
      "grad_norm": 4.118239402770996,
      "learning_rate": 0.0001817301813153218,
      "loss": 1.2887,
      "step": 63950
    },
    {
      "epoch": 1.9108470426656317,
      "grad_norm": 3.598353147506714,
      "learning_rate": 0.00018171589551940298,
      "loss": 1.3341,
      "step": 64000
    },
    {
      "epoch": 1.9123398919177141,
      "grad_norm": 3.8190245628356934,
      "learning_rate": 0.00018170160972348415,
      "loss": 1.3011,
      "step": 64050
    },
    {
      "epoch": 1.9138327411697966,
      "grad_norm": 6.505800724029541,
      "learning_rate": 0.0001816873239275653,
      "loss": 1.3583,
      "step": 64100
    },
    {
      "epoch": 1.915325590421879,
      "grad_norm": 4.134885311126709,
      "learning_rate": 0.00018167303813164647,
      "loss": 1.3363,
      "step": 64150
    },
    {
      "epoch": 1.9168184396739618,
      "grad_norm": 3.753544807434082,
      "learning_rate": 0.00018165875233572764,
      "loss": 1.3415,
      "step": 64200
    },
    {
      "epoch": 1.9183112889260443,
      "grad_norm": 3.716143846511841,
      "learning_rate": 0.0001816444665398088,
      "loss": 1.3276,
      "step": 64250
    },
    {
      "epoch": 1.9198041381781268,
      "grad_norm": 4.619287490844727,
      "learning_rate": 0.00018163018074388997,
      "loss": 1.3624,
      "step": 64300
    },
    {
      "epoch": 1.9212969874302093,
      "grad_norm": 3.8095598220825195,
      "learning_rate": 0.00018161589494797113,
      "loss": 1.3488,
      "step": 64350
    },
    {
      "epoch": 1.9227898366822918,
      "grad_norm": 6.268500328063965,
      "learning_rate": 0.0001816016091520523,
      "loss": 1.3481,
      "step": 64400
    },
    {
      "epoch": 1.9242826859343745,
      "grad_norm": 6.642594814300537,
      "learning_rate": 0.00018158732335613346,
      "loss": 1.3392,
      "step": 64450
    },
    {
      "epoch": 1.9257755351864567,
      "grad_norm": 5.921909332275391,
      "learning_rate": 0.00018157303756021465,
      "loss": 1.3218,
      "step": 64500
    },
    {
      "epoch": 1.9272683844385394,
      "grad_norm": 4.467897415161133,
      "learning_rate": 0.0001815587517642958,
      "loss": 1.3087,
      "step": 64550
    },
    {
      "epoch": 1.928761233690622,
      "grad_norm": 6.775788307189941,
      "learning_rate": 0.00018154446596837698,
      "loss": 1.3712,
      "step": 64600
    },
    {
      "epoch": 1.9302540829427044,
      "grad_norm": 4.561954975128174,
      "learning_rate": 0.00018153018017245812,
      "loss": 1.3214,
      "step": 64650
    },
    {
      "epoch": 1.931746932194787,
      "grad_norm": 5.288097381591797,
      "learning_rate": 0.0001815158943765393,
      "loss": 1.3926,
      "step": 64700
    },
    {
      "epoch": 1.9332397814468694,
      "grad_norm": 4.883842945098877,
      "learning_rate": 0.00018150160858062045,
      "loss": 1.2545,
      "step": 64750
    },
    {
      "epoch": 1.934732630698952,
      "grad_norm": 4.528720378875732,
      "learning_rate": 0.00018148732278470164,
      "loss": 1.3733,
      "step": 64800
    },
    {
      "epoch": 1.9362254799510346,
      "grad_norm": 3.9458446502685547,
      "learning_rate": 0.0001814730369887828,
      "loss": 1.3685,
      "step": 64850
    },
    {
      "epoch": 1.937718329203117,
      "grad_norm": 5.117423057556152,
      "learning_rate": 0.00018145875119286397,
      "loss": 1.3712,
      "step": 64900
    },
    {
      "epoch": 1.9392111784551997,
      "grad_norm": 4.276913166046143,
      "learning_rate": 0.00018144446539694514,
      "loss": 1.2996,
      "step": 64950
    },
    {
      "epoch": 1.940704027707282,
      "grad_norm": 3.9713099002838135,
      "learning_rate": 0.0001814301796010263,
      "loss": 1.301,
      "step": 65000
    },
    {
      "epoch": 1.9421968769593647,
      "grad_norm": 4.240658760070801,
      "learning_rate": 0.00018141589380510747,
      "loss": 1.4013,
      "step": 65050
    },
    {
      "epoch": 1.9436897262114472,
      "grad_norm": 4.406320571899414,
      "learning_rate": 0.00018140160800918863,
      "loss": 1.3481,
      "step": 65100
    },
    {
      "epoch": 1.9451825754635297,
      "grad_norm": 4.139687538146973,
      "learning_rate": 0.0001813873222132698,
      "loss": 1.3179,
      "step": 65150
    },
    {
      "epoch": 1.9466754247156122,
      "grad_norm": 3.8167126178741455,
      "learning_rate": 0.00018137303641735096,
      "loss": 1.2714,
      "step": 65200
    },
    {
      "epoch": 1.9481682739676947,
      "grad_norm": 4.680548191070557,
      "learning_rate": 0.00018135875062143212,
      "loss": 1.3571,
      "step": 65250
    },
    {
      "epoch": 1.9496611232197774,
      "grad_norm": 8.530779838562012,
      "learning_rate": 0.00018134446482551332,
      "loss": 1.323,
      "step": 65300
    },
    {
      "epoch": 1.9511539724718598,
      "grad_norm": 4.576430320739746,
      "learning_rate": 0.00018133017902959445,
      "loss": 1.3709,
      "step": 65350
    },
    {
      "epoch": 1.9526468217239423,
      "grad_norm": 5.320089817047119,
      "learning_rate": 0.00018131589323367565,
      "loss": 1.3857,
      "step": 65400
    },
    {
      "epoch": 1.9541396709760248,
      "grad_norm": 4.247587203979492,
      "learning_rate": 0.00018130160743775678,
      "loss": 1.3637,
      "step": 65450
    },
    {
      "epoch": 1.9556325202281073,
      "grad_norm": 4.2958903312683105,
      "learning_rate": 0.00018128732164183797,
      "loss": 1.2861,
      "step": 65500
    },
    {
      "epoch": 1.95712536948019,
      "grad_norm": 4.818902969360352,
      "learning_rate": 0.00018127303584591914,
      "loss": 1.3264,
      "step": 65550
    },
    {
      "epoch": 1.9586182187322723,
      "grad_norm": 4.755589962005615,
      "learning_rate": 0.0001812587500500003,
      "loss": 1.3821,
      "step": 65600
    },
    {
      "epoch": 1.960111067984355,
      "grad_norm": 3.589580774307251,
      "learning_rate": 0.00018124446425408147,
      "loss": 1.2799,
      "step": 65650
    },
    {
      "epoch": 1.9616039172364375,
      "grad_norm": 5.7323431968688965,
      "learning_rate": 0.00018123017845816263,
      "loss": 1.2616,
      "step": 65700
    },
    {
      "epoch": 1.96309676648852,
      "grad_norm": 6.535925388336182,
      "learning_rate": 0.0001812158926622438,
      "loss": 1.3391,
      "step": 65750
    },
    {
      "epoch": 1.9645896157406026,
      "grad_norm": 3.783618688583374,
      "learning_rate": 0.00018120160686632496,
      "loss": 1.2743,
      "step": 65800
    },
    {
      "epoch": 1.966082464992685,
      "grad_norm": 4.055028438568115,
      "learning_rate": 0.00018118732107040613,
      "loss": 1.3319,
      "step": 65850
    },
    {
      "epoch": 1.9675753142447676,
      "grad_norm": 4.827310562133789,
      "learning_rate": 0.0001811730352744873,
      "loss": 1.3518,
      "step": 65900
    },
    {
      "epoch": 1.96906816349685,
      "grad_norm": 4.298334121704102,
      "learning_rate": 0.00018115874947856846,
      "loss": 1.3755,
      "step": 65950
    },
    {
      "epoch": 1.9705610127489326,
      "grad_norm": 4.718268871307373,
      "learning_rate": 0.00018114446368264962,
      "loss": 1.2615,
      "step": 66000
    },
    {
      "epoch": 1.9720538620010153,
      "grad_norm": 4.387974262237549,
      "learning_rate": 0.00018113017788673079,
      "loss": 1.3475,
      "step": 66050
    },
    {
      "epoch": 1.9735467112530976,
      "grad_norm": 4.896945476531982,
      "learning_rate": 0.00018111589209081198,
      "loss": 1.3145,
      "step": 66100
    },
    {
      "epoch": 1.9750395605051803,
      "grad_norm": 5.453362941741943,
      "learning_rate": 0.00018110160629489312,
      "loss": 1.3377,
      "step": 66150
    },
    {
      "epoch": 1.9765324097572627,
      "grad_norm": 6.6892499923706055,
      "learning_rate": 0.00018108732049897428,
      "loss": 1.2734,
      "step": 66200
    },
    {
      "epoch": 1.9780252590093452,
      "grad_norm": 4.2804765701293945,
      "learning_rate": 0.00018107303470305544,
      "loss": 1.3742,
      "step": 66250
    },
    {
      "epoch": 1.9795181082614277,
      "grad_norm": 4.235034465789795,
      "learning_rate": 0.0001810587489071366,
      "loss": 1.2921,
      "step": 66300
    },
    {
      "epoch": 1.9810109575135102,
      "grad_norm": 4.113801956176758,
      "learning_rate": 0.0001810444631112178,
      "loss": 1.3475,
      "step": 66350
    },
    {
      "epoch": 1.982503806765593,
      "grad_norm": 4.276202201843262,
      "learning_rate": 0.00018103017731529894,
      "loss": 1.2955,
      "step": 66400
    },
    {
      "epoch": 1.9839966560176754,
      "grad_norm": 4.72022008895874,
      "learning_rate": 0.00018101589151938013,
      "loss": 1.3676,
      "step": 66450
    },
    {
      "epoch": 1.9854895052697579,
      "grad_norm": 4.182514190673828,
      "learning_rate": 0.00018100160572346127,
      "loss": 1.3706,
      "step": 66500
    },
    {
      "epoch": 1.9869823545218404,
      "grad_norm": 3.8737592697143555,
      "learning_rate": 0.00018098731992754246,
      "loss": 1.3103,
      "step": 66550
    },
    {
      "epoch": 1.9884752037739228,
      "grad_norm": 6.266764163970947,
      "learning_rate": 0.0001809730341316236,
      "loss": 1.3175,
      "step": 66600
    },
    {
      "epoch": 1.9899680530260055,
      "grad_norm": 5.243070125579834,
      "learning_rate": 0.0001809587483357048,
      "loss": 1.3484,
      "step": 66650
    },
    {
      "epoch": 1.9914609022780878,
      "grad_norm": 4.289467811584473,
      "learning_rate": 0.00018094446253978595,
      "loss": 1.3806,
      "step": 66700
    },
    {
      "epoch": 1.9929537515301705,
      "grad_norm": 5.450198173522949,
      "learning_rate": 0.00018093017674386712,
      "loss": 1.3155,
      "step": 66750
    },
    {
      "epoch": 1.994446600782253,
      "grad_norm": 4.389066219329834,
      "learning_rate": 0.00018091589094794828,
      "loss": 1.293,
      "step": 66800
    },
    {
      "epoch": 1.9959394500343355,
      "grad_norm": 4.606175422668457,
      "learning_rate": 0.00018090160515202945,
      "loss": 1.3558,
      "step": 66850
    },
    {
      "epoch": 1.9974322992864182,
      "grad_norm": 4.8083319664001465,
      "learning_rate": 0.0001808873193561106,
      "loss": 1.2621,
      "step": 66900
    },
    {
      "epoch": 1.9989251485385005,
      "grad_norm": 3.760852813720703,
      "learning_rate": 0.00018087303356019178,
      "loss": 1.3539,
      "step": 66950
    },
    {
      "epoch": 2.000417997790583,
      "grad_norm": 3.6912026405334473,
      "learning_rate": 0.00018085874776427294,
      "loss": 1.2619,
      "step": 67000
    },
    {
      "epoch": 2.0019108470426654,
      "grad_norm": 4.681028842926025,
      "learning_rate": 0.0001808444619683541,
      "loss": 1.2457,
      "step": 67050
    },
    {
      "epoch": 2.003403696294748,
      "grad_norm": 4.194242477416992,
      "learning_rate": 0.00018083017617243527,
      "loss": 1.297,
      "step": 67100
    },
    {
      "epoch": 2.004896545546831,
      "grad_norm": 3.831331729888916,
      "learning_rate": 0.00018081589037651646,
      "loss": 1.2989,
      "step": 67150
    },
    {
      "epoch": 2.006389394798913,
      "grad_norm": 4.249265193939209,
      "learning_rate": 0.0001808016045805976,
      "loss": 1.2078,
      "step": 67200
    },
    {
      "epoch": 2.007882244050996,
      "grad_norm": 7.441315650939941,
      "learning_rate": 0.0001807873187846788,
      "loss": 1.2945,
      "step": 67250
    },
    {
      "epoch": 2.009375093303078,
      "grad_norm": 4.2568745613098145,
      "learning_rate": 0.00018077303298875993,
      "loss": 1.2295,
      "step": 67300
    },
    {
      "epoch": 2.0108679425551608,
      "grad_norm": 5.394452095031738,
      "learning_rate": 0.00018075874719284112,
      "loss": 1.2135,
      "step": 67350
    },
    {
      "epoch": 2.0123607918072435,
      "grad_norm": 3.199411630630493,
      "learning_rate": 0.00018074446139692226,
      "loss": 1.2314,
      "step": 67400
    },
    {
      "epoch": 2.0138536410593257,
      "grad_norm": 3.7905383110046387,
      "learning_rate": 0.00018073017560100345,
      "loss": 1.2227,
      "step": 67450
    },
    {
      "epoch": 2.0153464903114084,
      "grad_norm": 7.900835990905762,
      "learning_rate": 0.00018071588980508462,
      "loss": 1.2067,
      "step": 67500
    },
    {
      "epoch": 2.0168393395634907,
      "grad_norm": 4.803965091705322,
      "learning_rate": 0.00018070160400916578,
      "loss": 1.2894,
      "step": 67550
    },
    {
      "epoch": 2.0183321888155734,
      "grad_norm": 4.046397686004639,
      "learning_rate": 0.00018068731821324694,
      "loss": 1.2627,
      "step": 67600
    },
    {
      "epoch": 2.019825038067656,
      "grad_norm": 4.9588518142700195,
      "learning_rate": 0.0001806730324173281,
      "loss": 1.1931,
      "step": 67650
    },
    {
      "epoch": 2.0213178873197384,
      "grad_norm": 3.5710065364837646,
      "learning_rate": 0.00018065874662140927,
      "loss": 1.1937,
      "step": 67700
    },
    {
      "epoch": 2.022810736571821,
      "grad_norm": 4.107451915740967,
      "learning_rate": 0.00018064446082549044,
      "loss": 1.2236,
      "step": 67750
    },
    {
      "epoch": 2.0243035858239034,
      "grad_norm": 4.8303422927856445,
      "learning_rate": 0.0001806301750295716,
      "loss": 1.2223,
      "step": 67800
    },
    {
      "epoch": 2.025796435075986,
      "grad_norm": 6.539272308349609,
      "learning_rate": 0.00018061588923365277,
      "loss": 1.3199,
      "step": 67850
    },
    {
      "epoch": 2.0272892843280688,
      "grad_norm": 4.415231227874756,
      "learning_rate": 0.00018060160343773393,
      "loss": 1.2988,
      "step": 67900
    },
    {
      "epoch": 2.028782133580151,
      "grad_norm": 4.396175861358643,
      "learning_rate": 0.00018058731764181512,
      "loss": 1.1952,
      "step": 67950
    },
    {
      "epoch": 2.0302749828322337,
      "grad_norm": 3.2673914432525635,
      "learning_rate": 0.00018057303184589626,
      "loss": 1.2285,
      "step": 68000
    },
    {
      "epoch": 2.031767832084316,
      "grad_norm": 3.6252634525299072,
      "learning_rate": 0.00018055874604997745,
      "loss": 1.314,
      "step": 68050
    },
    {
      "epoch": 2.0332606813363987,
      "grad_norm": 4.550121307373047,
      "learning_rate": 0.0001805444602540586,
      "loss": 1.2846,
      "step": 68100
    },
    {
      "epoch": 2.034753530588481,
      "grad_norm": 4.809107303619385,
      "learning_rate": 0.00018053017445813978,
      "loss": 1.3165,
      "step": 68150
    },
    {
      "epoch": 2.0362463798405637,
      "grad_norm": 4.466585159301758,
      "learning_rate": 0.00018051588866222092,
      "loss": 1.2543,
      "step": 68200
    },
    {
      "epoch": 2.0377392290926464,
      "grad_norm": 4.243835926055908,
      "learning_rate": 0.0001805016028663021,
      "loss": 1.2908,
      "step": 68250
    },
    {
      "epoch": 2.0392320783447286,
      "grad_norm": 4.369659423828125,
      "learning_rate": 0.00018048731707038328,
      "loss": 1.2167,
      "step": 68300
    },
    {
      "epoch": 2.0407249275968113,
      "grad_norm": 4.6665496826171875,
      "learning_rate": 0.00018047303127446444,
      "loss": 1.2453,
      "step": 68350
    },
    {
      "epoch": 2.0422177768488936,
      "grad_norm": 4.3825507164001465,
      "learning_rate": 0.0001804587454785456,
      "loss": 1.2272,
      "step": 68400
    },
    {
      "epoch": 2.0437106261009763,
      "grad_norm": 4.299841403961182,
      "learning_rate": 0.00018044445968262677,
      "loss": 1.2198,
      "step": 68450
    },
    {
      "epoch": 2.045203475353059,
      "grad_norm": 3.937131881713867,
      "learning_rate": 0.00018043017388670794,
      "loss": 1.2763,
      "step": 68500
    },
    {
      "epoch": 2.0466963246051413,
      "grad_norm": 6.704258918762207,
      "learning_rate": 0.0001804158880907891,
      "loss": 1.3289,
      "step": 68550
    },
    {
      "epoch": 2.048189173857224,
      "grad_norm": 3.8973612785339355,
      "learning_rate": 0.00018040160229487026,
      "loss": 1.2907,
      "step": 68600
    },
    {
      "epoch": 2.0496820231093063,
      "grad_norm": 8.648911476135254,
      "learning_rate": 0.00018038731649895143,
      "loss": 1.2488,
      "step": 68650
    },
    {
      "epoch": 2.051174872361389,
      "grad_norm": 4.553196430206299,
      "learning_rate": 0.0001803730307030326,
      "loss": 1.2489,
      "step": 68700
    },
    {
      "epoch": 2.0526677216134717,
      "grad_norm": 4.379563808441162,
      "learning_rate": 0.00018035874490711376,
      "loss": 1.253,
      "step": 68750
    },
    {
      "epoch": 2.054160570865554,
      "grad_norm": 6.393334865570068,
      "learning_rate": 0.00018034445911119492,
      "loss": 1.1783,
      "step": 68800
    },
    {
      "epoch": 2.0556534201176366,
      "grad_norm": 4.846323013305664,
      "learning_rate": 0.0001803301733152761,
      "loss": 1.2583,
      "step": 68850
    },
    {
      "epoch": 2.057146269369719,
      "grad_norm": 5.24686861038208,
      "learning_rate": 0.00018031588751935725,
      "loss": 1.333,
      "step": 68900
    },
    {
      "epoch": 2.0586391186218016,
      "grad_norm": 4.824219226837158,
      "learning_rate": 0.00018030160172343842,
      "loss": 1.2931,
      "step": 68950
    },
    {
      "epoch": 2.0601319678738843,
      "grad_norm": 5.001037120819092,
      "learning_rate": 0.0001802873159275196,
      "loss": 1.2625,
      "step": 69000
    },
    {
      "epoch": 2.0616248171259666,
      "grad_norm": 4.036411762237549,
      "learning_rate": 0.00018027303013160075,
      "loss": 1.2585,
      "step": 69050
    },
    {
      "epoch": 2.0631176663780493,
      "grad_norm": 4.597596645355225,
      "learning_rate": 0.00018025874433568194,
      "loss": 1.2812,
      "step": 69100
    },
    {
      "epoch": 2.0646105156301315,
      "grad_norm": 7.785471439361572,
      "learning_rate": 0.00018024445853976308,
      "loss": 1.2749,
      "step": 69150
    },
    {
      "epoch": 2.0661033648822142,
      "grad_norm": 4.905806541442871,
      "learning_rate": 0.00018023017274384427,
      "loss": 1.2619,
      "step": 69200
    },
    {
      "epoch": 2.0675962141342965,
      "grad_norm": 5.271987438201904,
      "learning_rate": 0.0001802158869479254,
      "loss": 1.2769,
      "step": 69250
    },
    {
      "epoch": 2.069089063386379,
      "grad_norm": 4.129353046417236,
      "learning_rate": 0.0001802016011520066,
      "loss": 1.2552,
      "step": 69300
    },
    {
      "epoch": 2.070581912638462,
      "grad_norm": 4.855686187744141,
      "learning_rate": 0.00018018731535608776,
      "loss": 1.2452,
      "step": 69350
    },
    {
      "epoch": 2.072074761890544,
      "grad_norm": 4.6511383056640625,
      "learning_rate": 0.00018017302956016893,
      "loss": 1.2529,
      "step": 69400
    },
    {
      "epoch": 2.073567611142627,
      "grad_norm": 4.4711995124816895,
      "learning_rate": 0.0001801587437642501,
      "loss": 1.2721,
      "step": 69450
    },
    {
      "epoch": 2.075060460394709,
      "grad_norm": 4.466591835021973,
      "learning_rate": 0.00018014445796833126,
      "loss": 1.2654,
      "step": 69500
    },
    {
      "epoch": 2.076553309646792,
      "grad_norm": 5.165157318115234,
      "learning_rate": 0.00018013017217241242,
      "loss": 1.2522,
      "step": 69550
    },
    {
      "epoch": 2.0780461588988746,
      "grad_norm": 4.762765407562256,
      "learning_rate": 0.00018011588637649358,
      "loss": 1.2318,
      "step": 69600
    },
    {
      "epoch": 2.079539008150957,
      "grad_norm": 5.024138927459717,
      "learning_rate": 0.00018010160058057475,
      "loss": 1.2713,
      "step": 69650
    },
    {
      "epoch": 2.0810318574030395,
      "grad_norm": 5.054821014404297,
      "learning_rate": 0.00018008731478465591,
      "loss": 1.2714,
      "step": 69700
    },
    {
      "epoch": 2.082524706655122,
      "grad_norm": 5.052605152130127,
      "learning_rate": 0.00018007302898873708,
      "loss": 1.2001,
      "step": 69750
    },
    {
      "epoch": 2.0840175559072045,
      "grad_norm": 6.121365547180176,
      "learning_rate": 0.00018005874319281827,
      "loss": 1.2372,
      "step": 69800
    },
    {
      "epoch": 2.085510405159287,
      "grad_norm": 4.6751627922058105,
      "learning_rate": 0.0001800444573968994,
      "loss": 1.2366,
      "step": 69850
    },
    {
      "epoch": 2.0870032544113695,
      "grad_norm": 4.396819591522217,
      "learning_rate": 0.0001800301716009806,
      "loss": 1.2878,
      "step": 69900
    },
    {
      "epoch": 2.088496103663452,
      "grad_norm": 5.096658229827881,
      "learning_rate": 0.00018001588580506174,
      "loss": 1.2265,
      "step": 69950
    },
    {
      "epoch": 2.0899889529155344,
      "grad_norm": 4.45212459564209,
      "learning_rate": 0.00018000160000914293,
      "loss": 1.2261,
      "step": 70000
    },
    {
      "epoch": 2.091481802167617,
      "grad_norm": 4.837035179138184,
      "learning_rate": 0.00017998731421322407,
      "loss": 1.2857,
      "step": 70050
    },
    {
      "epoch": 2.0929746514197,
      "grad_norm": 4.402126789093018,
      "learning_rate": 0.00017997302841730526,
      "loss": 1.2614,
      "step": 70100
    },
    {
      "epoch": 2.094467500671782,
      "grad_norm": 3.9073615074157715,
      "learning_rate": 0.00017995874262138642,
      "loss": 1.2485,
      "step": 70150
    },
    {
      "epoch": 2.095960349923865,
      "grad_norm": 5.109444618225098,
      "learning_rate": 0.0001799444568254676,
      "loss": 1.2674,
      "step": 70200
    },
    {
      "epoch": 2.097453199175947,
      "grad_norm": 6.542872428894043,
      "learning_rate": 0.00017993017102954875,
      "loss": 1.2715,
      "step": 70250
    },
    {
      "epoch": 2.09894604842803,
      "grad_norm": 5.090799808502197,
      "learning_rate": 0.00017991588523362992,
      "loss": 1.2821,
      "step": 70300
    },
    {
      "epoch": 2.100438897680112,
      "grad_norm": 5.429018497467041,
      "learning_rate": 0.00017990159943771108,
      "loss": 1.2994,
      "step": 70350
    },
    {
      "epoch": 2.1019317469321948,
      "grad_norm": 8.873945236206055,
      "learning_rate": 0.00017988731364179225,
      "loss": 1.2912,
      "step": 70400
    },
    {
      "epoch": 2.1034245961842775,
      "grad_norm": 5.144947528839111,
      "learning_rate": 0.0001798730278458734,
      "loss": 1.2744,
      "step": 70450
    },
    {
      "epoch": 2.1049174454363597,
      "grad_norm": 5.346638202667236,
      "learning_rate": 0.00017985874204995458,
      "loss": 1.311,
      "step": 70500
    },
    {
      "epoch": 2.1064102946884424,
      "grad_norm": 5.384441375732422,
      "learning_rate": 0.00017984445625403574,
      "loss": 1.2662,
      "step": 70550
    },
    {
      "epoch": 2.1079031439405247,
      "grad_norm": 4.645060062408447,
      "learning_rate": 0.00017983017045811693,
      "loss": 1.2366,
      "step": 70600
    },
    {
      "epoch": 2.1093959931926074,
      "grad_norm": 4.9864349365234375,
      "learning_rate": 0.00017981588466219807,
      "loss": 1.2531,
      "step": 70650
    },
    {
      "epoch": 2.11088884244469,
      "grad_norm": 4.057093143463135,
      "learning_rate": 0.00017980159886627926,
      "loss": 1.2477,
      "step": 70700
    },
    {
      "epoch": 2.1123816916967724,
      "grad_norm": 4.121668815612793,
      "learning_rate": 0.0001797873130703604,
      "loss": 1.2533,
      "step": 70750
    },
    {
      "epoch": 2.113874540948855,
      "grad_norm": 5.140614032745361,
      "learning_rate": 0.0001797730272744416,
      "loss": 1.1833,
      "step": 70800
    },
    {
      "epoch": 2.1153673902009373,
      "grad_norm": 5.469926834106445,
      "learning_rate": 0.00017975874147852273,
      "loss": 1.2746,
      "step": 70850
    },
    {
      "epoch": 2.11686023945302,
      "grad_norm": 3.0967893600463867,
      "learning_rate": 0.00017974445568260392,
      "loss": 1.2449,
      "step": 70900
    },
    {
      "epoch": 2.1183530887051027,
      "grad_norm": 4.644883155822754,
      "learning_rate": 0.00017973016988668508,
      "loss": 1.2304,
      "step": 70950
    },
    {
      "epoch": 2.119845937957185,
      "grad_norm": 4.095651626586914,
      "learning_rate": 0.00017971588409076625,
      "loss": 1.3378,
      "step": 71000
    },
    {
      "epoch": 2.1213387872092677,
      "grad_norm": 5.656352519989014,
      "learning_rate": 0.00017970159829484741,
      "loss": 1.268,
      "step": 71050
    },
    {
      "epoch": 2.12283163646135,
      "grad_norm": 5.86411190032959,
      "learning_rate": 0.00017968731249892858,
      "loss": 1.3146,
      "step": 71100
    },
    {
      "epoch": 2.1243244857134327,
      "grad_norm": 4.751001358032227,
      "learning_rate": 0.00017967302670300974,
      "loss": 1.2861,
      "step": 71150
    },
    {
      "epoch": 2.1258173349655154,
      "grad_norm": 4.90724515914917,
      "learning_rate": 0.0001796587409070909,
      "loss": 1.2307,
      "step": 71200
    },
    {
      "epoch": 2.1273101842175977,
      "grad_norm": 4.715629577636719,
      "learning_rate": 0.00017964445511117207,
      "loss": 1.2602,
      "step": 71250
    },
    {
      "epoch": 2.1288030334696804,
      "grad_norm": 3.9297828674316406,
      "learning_rate": 0.00017963016931525324,
      "loss": 1.2705,
      "step": 71300
    },
    {
      "epoch": 2.1302958827217626,
      "grad_norm": 4.373486518859863,
      "learning_rate": 0.0001796158835193344,
      "loss": 1.264,
      "step": 71350
    },
    {
      "epoch": 2.1317887319738453,
      "grad_norm": 5.277180194854736,
      "learning_rate": 0.00017960159772341557,
      "loss": 1.2703,
      "step": 71400
    },
    {
      "epoch": 2.133281581225928,
      "grad_norm": 3.8668055534362793,
      "learning_rate": 0.00017958731192749673,
      "loss": 1.2655,
      "step": 71450
    },
    {
      "epoch": 2.1347744304780103,
      "grad_norm": 4.491888046264648,
      "learning_rate": 0.0001795730261315779,
      "loss": 1.2659,
      "step": 71500
    },
    {
      "epoch": 2.136267279730093,
      "grad_norm": 4.461415767669678,
      "learning_rate": 0.00017955874033565906,
      "loss": 1.3075,
      "step": 71550
    },
    {
      "epoch": 2.1377601289821753,
      "grad_norm": 5.062393665313721,
      "learning_rate": 0.00017954445453974023,
      "loss": 1.2328,
      "step": 71600
    },
    {
      "epoch": 2.139252978234258,
      "grad_norm": 5.79408073425293,
      "learning_rate": 0.0001795301687438214,
      "loss": 1.2528,
      "step": 71650
    },
    {
      "epoch": 2.1407458274863402,
      "grad_norm": 5.496243476867676,
      "learning_rate": 0.00017951588294790255,
      "loss": 1.2409,
      "step": 71700
    },
    {
      "epoch": 2.142238676738423,
      "grad_norm": 5.3767805099487305,
      "learning_rate": 0.00017950159715198375,
      "loss": 1.2383,
      "step": 71750
    },
    {
      "epoch": 2.1437315259905056,
      "grad_norm": 6.0737199783325195,
      "learning_rate": 0.00017948731135606488,
      "loss": 1.2936,
      "step": 71800
    },
    {
      "epoch": 2.145224375242588,
      "grad_norm": 5.363690376281738,
      "learning_rate": 0.00017947302556014608,
      "loss": 1.2774,
      "step": 71850
    },
    {
      "epoch": 2.1467172244946706,
      "grad_norm": 4.2048773765563965,
      "learning_rate": 0.0001794587397642272,
      "loss": 1.2476,
      "step": 71900
    },
    {
      "epoch": 2.148210073746753,
      "grad_norm": 5.775049686431885,
      "learning_rate": 0.0001794444539683084,
      "loss": 1.2646,
      "step": 71950
    },
    {
      "epoch": 2.1497029229988356,
      "grad_norm": 4.7285475730896,
      "learning_rate": 0.00017943016817238957,
      "loss": 1.3038,
      "step": 72000
    },
    {
      "epoch": 2.1511957722509183,
      "grad_norm": 4.310266971588135,
      "learning_rate": 0.00017941588237647073,
      "loss": 1.2478,
      "step": 72050
    },
    {
      "epoch": 2.1526886215030006,
      "grad_norm": 5.618819713592529,
      "learning_rate": 0.0001794015965805519,
      "loss": 1.2788,
      "step": 72100
    },
    {
      "epoch": 2.1541814707550833,
      "grad_norm": 5.624669075012207,
      "learning_rate": 0.00017938731078463306,
      "loss": 1.3053,
      "step": 72150
    },
    {
      "epoch": 2.1556743200071655,
      "grad_norm": 3.442650079727173,
      "learning_rate": 0.00017937302498871423,
      "loss": 1.2774,
      "step": 72200
    },
    {
      "epoch": 2.1571671692592482,
      "grad_norm": 5.232537746429443,
      "learning_rate": 0.0001793587391927954,
      "loss": 1.3006,
      "step": 72250
    },
    {
      "epoch": 2.1586600185113305,
      "grad_norm": 4.750761985778809,
      "learning_rate": 0.00017934445339687656,
      "loss": 1.2696,
      "step": 72300
    },
    {
      "epoch": 2.160152867763413,
      "grad_norm": 5.479654788970947,
      "learning_rate": 0.00017933016760095772,
      "loss": 1.2896,
      "step": 72350
    },
    {
      "epoch": 2.161645717015496,
      "grad_norm": 4.499126434326172,
      "learning_rate": 0.0001793158818050389,
      "loss": 1.2962,
      "step": 72400
    },
    {
      "epoch": 2.163138566267578,
      "grad_norm": 5.156326770782471,
      "learning_rate": 0.00017930159600912008,
      "loss": 1.2852,
      "step": 72450
    },
    {
      "epoch": 2.164631415519661,
      "grad_norm": 4.520088195800781,
      "learning_rate": 0.00017928731021320122,
      "loss": 1.2527,
      "step": 72500
    },
    {
      "epoch": 2.166124264771743,
      "grad_norm": 4.284939289093018,
      "learning_rate": 0.0001792730244172824,
      "loss": 1.3213,
      "step": 72550
    },
    {
      "epoch": 2.167617114023826,
      "grad_norm": 4.373524188995361,
      "learning_rate": 0.00017925873862136355,
      "loss": 1.2853,
      "step": 72600
    },
    {
      "epoch": 2.1691099632759085,
      "grad_norm": 3.9682133197784424,
      "learning_rate": 0.00017924445282544474,
      "loss": 1.2707,
      "step": 72650
    },
    {
      "epoch": 2.170602812527991,
      "grad_norm": 5.218033313751221,
      "learning_rate": 0.00017923016702952587,
      "loss": 1.2492,
      "step": 72700
    },
    {
      "epoch": 2.1720956617800735,
      "grad_norm": 4.900534152984619,
      "learning_rate": 0.00017921588123360707,
      "loss": 1.256,
      "step": 72750
    },
    {
      "epoch": 2.173588511032156,
      "grad_norm": 4.519747257232666,
      "learning_rate": 0.00017920159543768823,
      "loss": 1.2367,
      "step": 72800
    },
    {
      "epoch": 2.1750813602842385,
      "grad_norm": 4.465813159942627,
      "learning_rate": 0.0001791873096417694,
      "loss": 1.286,
      "step": 72850
    },
    {
      "epoch": 2.176574209536321,
      "grad_norm": 6.142160892486572,
      "learning_rate": 0.00017917302384585056,
      "loss": 1.2959,
      "step": 72900
    },
    {
      "epoch": 2.1780670587884035,
      "grad_norm": 3.851306200027466,
      "learning_rate": 0.00017915873804993173,
      "loss": 1.2511,
      "step": 72950
    },
    {
      "epoch": 2.179559908040486,
      "grad_norm": 5.0324788093566895,
      "learning_rate": 0.0001791444522540129,
      "loss": 1.2652,
      "step": 73000
    },
    {
      "epoch": 2.1810527572925684,
      "grad_norm": 4.664618492126465,
      "learning_rate": 0.00017913016645809405,
      "loss": 1.2639,
      "step": 73050
    },
    {
      "epoch": 2.182545606544651,
      "grad_norm": 4.418179988861084,
      "learning_rate": 0.00017911588066217522,
      "loss": 1.2948,
      "step": 73100
    },
    {
      "epoch": 2.184038455796734,
      "grad_norm": 4.895113945007324,
      "learning_rate": 0.00017910159486625638,
      "loss": 1.2283,
      "step": 73150
    },
    {
      "epoch": 2.185531305048816,
      "grad_norm": 4.2365946769714355,
      "learning_rate": 0.00017908730907033755,
      "loss": 1.2772,
      "step": 73200
    },
    {
      "epoch": 2.187024154300899,
      "grad_norm": 6.705420017242432,
      "learning_rate": 0.00017907302327441874,
      "loss": 1.273,
      "step": 73250
    },
    {
      "epoch": 2.188517003552981,
      "grad_norm": 5.094226360321045,
      "learning_rate": 0.00017905873747849988,
      "loss": 1.2391,
      "step": 73300
    },
    {
      "epoch": 2.1900098528050638,
      "grad_norm": 4.424404621124268,
      "learning_rate": 0.00017904445168258107,
      "loss": 1.2198,
      "step": 73350
    },
    {
      "epoch": 2.1915027020571465,
      "grad_norm": 5.0152201652526855,
      "learning_rate": 0.0001790301658866622,
      "loss": 1.3167,
      "step": 73400
    },
    {
      "epoch": 2.1929955513092287,
      "grad_norm": 4.916745662689209,
      "learning_rate": 0.0001790158800907434,
      "loss": 1.2482,
      "step": 73450
    },
    {
      "epoch": 2.1944884005613114,
      "grad_norm": 4.396121978759766,
      "learning_rate": 0.00017900159429482454,
      "loss": 1.2425,
      "step": 73500
    },
    {
      "epoch": 2.1959812498133937,
      "grad_norm": 4.289492607116699,
      "learning_rate": 0.00017898730849890573,
      "loss": 1.2647,
      "step": 73550
    },
    {
      "epoch": 2.1974740990654764,
      "grad_norm": 6.336133003234863,
      "learning_rate": 0.0001789730227029869,
      "loss": 1.2513,
      "step": 73600
    },
    {
      "epoch": 2.198966948317559,
      "grad_norm": 4.202422618865967,
      "learning_rate": 0.00017895873690706806,
      "loss": 1.2657,
      "step": 73650
    },
    {
      "epoch": 2.2004597975696414,
      "grad_norm": 4.372437953948975,
      "learning_rate": 0.00017894445111114922,
      "loss": 1.2432,
      "step": 73700
    },
    {
      "epoch": 2.201952646821724,
      "grad_norm": 5.008980751037598,
      "learning_rate": 0.0001789301653152304,
      "loss": 1.2848,
      "step": 73750
    },
    {
      "epoch": 2.2034454960738064,
      "grad_norm": 5.251125812530518,
      "learning_rate": 0.00017891587951931155,
      "loss": 1.3296,
      "step": 73800
    },
    {
      "epoch": 2.204938345325889,
      "grad_norm": 3.898746967315674,
      "learning_rate": 0.00017890159372339272,
      "loss": 1.2639,
      "step": 73850
    },
    {
      "epoch": 2.2064311945779713,
      "grad_norm": 5.178562164306641,
      "learning_rate": 0.00017888730792747388,
      "loss": 1.3065,
      "step": 73900
    },
    {
      "epoch": 2.207924043830054,
      "grad_norm": 5.667030334472656,
      "learning_rate": 0.00017887302213155505,
      "loss": 1.2954,
      "step": 73950
    },
    {
      "epoch": 2.2094168930821367,
      "grad_norm": 7.237538814544678,
      "learning_rate": 0.0001788587363356362,
      "loss": 1.2558,
      "step": 74000
    },
    {
      "epoch": 2.210909742334219,
      "grad_norm": 4.029843330383301,
      "learning_rate": 0.00017884445053971737,
      "loss": 1.3004,
      "step": 74050
    },
    {
      "epoch": 2.2124025915863017,
      "grad_norm": 5.686726093292236,
      "learning_rate": 0.00017883016474379854,
      "loss": 1.278,
      "step": 74100
    },
    {
      "epoch": 2.213895440838384,
      "grad_norm": 5.373737812042236,
      "learning_rate": 0.0001788158789478797,
      "loss": 1.3089,
      "step": 74150
    },
    {
      "epoch": 2.2153882900904667,
      "grad_norm": 4.493228912353516,
      "learning_rate": 0.00017880159315196087,
      "loss": 1.2856,
      "step": 74200
    },
    {
      "epoch": 2.2168811393425494,
      "grad_norm": 5.083597183227539,
      "learning_rate": 0.00017878730735604203,
      "loss": 1.2792,
      "step": 74250
    },
    {
      "epoch": 2.2183739885946316,
      "grad_norm": 5.1655144691467285,
      "learning_rate": 0.0001787730215601232,
      "loss": 1.2852,
      "step": 74300
    },
    {
      "epoch": 2.2198668378467143,
      "grad_norm": 3.2413179874420166,
      "learning_rate": 0.00017875873576420436,
      "loss": 1.2617,
      "step": 74350
    },
    {
      "epoch": 2.2213596870987966,
      "grad_norm": 4.30172061920166,
      "learning_rate": 0.00017874444996828555,
      "loss": 1.2949,
      "step": 74400
    },
    {
      "epoch": 2.2228525363508793,
      "grad_norm": 4.27219295501709,
      "learning_rate": 0.0001787301641723667,
      "loss": 1.2293,
      "step": 74450
    },
    {
      "epoch": 2.2243453856029616,
      "grad_norm": 5.430578708648682,
      "learning_rate": 0.00017871587837644788,
      "loss": 1.2662,
      "step": 74500
    },
    {
      "epoch": 2.2258382348550443,
      "grad_norm": 5.0630292892456055,
      "learning_rate": 0.00017870159258052902,
      "loss": 1.308,
      "step": 74550
    },
    {
      "epoch": 2.227331084107127,
      "grad_norm": 4.326575756072998,
      "learning_rate": 0.0001786873067846102,
      "loss": 1.3244,
      "step": 74600
    },
    {
      "epoch": 2.2288239333592093,
      "grad_norm": 4.2859206199646,
      "learning_rate": 0.00017867302098869138,
      "loss": 1.2879,
      "step": 74650
    },
    {
      "epoch": 2.230316782611292,
      "grad_norm": 3.931136131286621,
      "learning_rate": 0.00017865873519277254,
      "loss": 1.3046,
      "step": 74700
    },
    {
      "epoch": 2.231809631863374,
      "grad_norm": 3.9645185470581055,
      "learning_rate": 0.0001786444493968537,
      "loss": 1.2712,
      "step": 74750
    },
    {
      "epoch": 2.233302481115457,
      "grad_norm": 5.313785076141357,
      "learning_rate": 0.00017863016360093487,
      "loss": 1.2514,
      "step": 74800
    },
    {
      "epoch": 2.2347953303675396,
      "grad_norm": 4.994843482971191,
      "learning_rate": 0.00017861587780501604,
      "loss": 1.2803,
      "step": 74850
    },
    {
      "epoch": 2.236288179619622,
      "grad_norm": 6.394142150878906,
      "learning_rate": 0.0001786015920090972,
      "loss": 1.2424,
      "step": 74900
    },
    {
      "epoch": 2.2377810288717046,
      "grad_norm": 4.67982816696167,
      "learning_rate": 0.00017858730621317837,
      "loss": 1.2775,
      "step": 74950
    },
    {
      "epoch": 2.239273878123787,
      "grad_norm": 5.697408199310303,
      "learning_rate": 0.00017857302041725953,
      "loss": 1.2665,
      "step": 75000
    },
    {
      "epoch": 2.2407667273758696,
      "grad_norm": 6.522719383239746,
      "learning_rate": 0.0001785587346213407,
      "loss": 1.3017,
      "step": 75050
    },
    {
      "epoch": 2.2422595766279523,
      "grad_norm": 3.7727861404418945,
      "learning_rate": 0.00017854444882542186,
      "loss": 1.3516,
      "step": 75100
    },
    {
      "epoch": 2.2437524258800345,
      "grad_norm": 4.810925483703613,
      "learning_rate": 0.00017853016302950302,
      "loss": 1.277,
      "step": 75150
    },
    {
      "epoch": 2.2452452751321172,
      "grad_norm": 3.6915650367736816,
      "learning_rate": 0.00017851587723358422,
      "loss": 1.3159,
      "step": 75200
    },
    {
      "epoch": 2.2467381243841995,
      "grad_norm": 4.427685260772705,
      "learning_rate": 0.00017850159143766535,
      "loss": 1.2725,
      "step": 75250
    },
    {
      "epoch": 2.248230973636282,
      "grad_norm": 7.222316265106201,
      "learning_rate": 0.00017848730564174655,
      "loss": 1.3272,
      "step": 75300
    },
    {
      "epoch": 2.249723822888365,
      "grad_norm": 4.113325595855713,
      "learning_rate": 0.00017847301984582768,
      "loss": 1.2861,
      "step": 75350
    },
    {
      "epoch": 2.251216672140447,
      "grad_norm": 3.8266818523406982,
      "learning_rate": 0.00017845873404990887,
      "loss": 1.2688,
      "step": 75400
    },
    {
      "epoch": 2.25270952139253,
      "grad_norm": 5.281492710113525,
      "learning_rate": 0.00017844444825399004,
      "loss": 1.2743,
      "step": 75450
    },
    {
      "epoch": 2.254202370644612,
      "grad_norm": 5.222156047821045,
      "learning_rate": 0.0001784301624580712,
      "loss": 1.2324,
      "step": 75500
    },
    {
      "epoch": 2.255695219896695,
      "grad_norm": 4.961234092712402,
      "learning_rate": 0.00017841587666215237,
      "loss": 1.3248,
      "step": 75550
    },
    {
      "epoch": 2.2571880691487776,
      "grad_norm": 5.501556396484375,
      "learning_rate": 0.00017840159086623353,
      "loss": 1.2868,
      "step": 75600
    },
    {
      "epoch": 2.25868091840086,
      "grad_norm": 4.440338611602783,
      "learning_rate": 0.0001783873050703147,
      "loss": 1.237,
      "step": 75650
    },
    {
      "epoch": 2.2601737676529425,
      "grad_norm": 7.373712539672852,
      "learning_rate": 0.00017837301927439586,
      "loss": 1.2931,
      "step": 75700
    },
    {
      "epoch": 2.261666616905025,
      "grad_norm": 4.324766635894775,
      "learning_rate": 0.00017835873347847703,
      "loss": 1.2762,
      "step": 75750
    },
    {
      "epoch": 2.2631594661571075,
      "grad_norm": 7.707190990447998,
      "learning_rate": 0.0001783444476825582,
      "loss": 1.2565,
      "step": 75800
    },
    {
      "epoch": 2.26465231540919,
      "grad_norm": 3.9111223220825195,
      "learning_rate": 0.00017833016188663936,
      "loss": 1.2134,
      "step": 75850
    },
    {
      "epoch": 2.2661451646612725,
      "grad_norm": 4.735013961791992,
      "learning_rate": 0.00017831587609072055,
      "loss": 1.1908,
      "step": 75900
    },
    {
      "epoch": 2.267638013913355,
      "grad_norm": 5.743906021118164,
      "learning_rate": 0.00017830159029480169,
      "loss": 1.294,
      "step": 75950
    },
    {
      "epoch": 2.2691308631654374,
      "grad_norm": 3.792663097381592,
      "learning_rate": 0.00017828730449888288,
      "loss": 1.3725,
      "step": 76000
    },
    {
      "epoch": 2.27062371241752,
      "grad_norm": 4.77357292175293,
      "learning_rate": 0.00017827301870296402,
      "loss": 1.2511,
      "step": 76050
    },
    {
      "epoch": 2.272116561669603,
      "grad_norm": 5.1196675300598145,
      "learning_rate": 0.0001782587329070452,
      "loss": 1.3099,
      "step": 76100
    },
    {
      "epoch": 2.273609410921685,
      "grad_norm": 5.5518574714660645,
      "learning_rate": 0.00017824444711112634,
      "loss": 1.2841,
      "step": 76150
    },
    {
      "epoch": 2.275102260173768,
      "grad_norm": 4.075642108917236,
      "learning_rate": 0.00017823016131520754,
      "loss": 1.2857,
      "step": 76200
    },
    {
      "epoch": 2.27659510942585,
      "grad_norm": 6.330162048339844,
      "learning_rate": 0.0001782158755192887,
      "loss": 1.253,
      "step": 76250
    },
    {
      "epoch": 2.278087958677933,
      "grad_norm": 3.8504278659820557,
      "learning_rate": 0.00017820158972336987,
      "loss": 1.3167,
      "step": 76300
    },
    {
      "epoch": 2.279580807930015,
      "grad_norm": 5.074968338012695,
      "learning_rate": 0.00017818730392745103,
      "loss": 1.3109,
      "step": 76350
    },
    {
      "epoch": 2.2810736571820978,
      "grad_norm": 4.519435405731201,
      "learning_rate": 0.0001781730181315322,
      "loss": 1.2721,
      "step": 76400
    },
    {
      "epoch": 2.2825665064341805,
      "grad_norm": 5.285332202911377,
      "learning_rate": 0.00017815873233561336,
      "loss": 1.3091,
      "step": 76450
    },
    {
      "epoch": 2.2840593556862627,
      "grad_norm": 6.193671226501465,
      "learning_rate": 0.00017814444653969452,
      "loss": 1.2676,
      "step": 76500
    },
    {
      "epoch": 2.2855522049383454,
      "grad_norm": 4.4518961906433105,
      "learning_rate": 0.0001781301607437757,
      "loss": 1.2637,
      "step": 76550
    },
    {
      "epoch": 2.2870450541904277,
      "grad_norm": 5.584630966186523,
      "learning_rate": 0.00017811587494785685,
      "loss": 1.2906,
      "step": 76600
    },
    {
      "epoch": 2.2885379034425104,
      "grad_norm": 5.528042793273926,
      "learning_rate": 0.00017810158915193802,
      "loss": 1.2701,
      "step": 76650
    },
    {
      "epoch": 2.2900307526945927,
      "grad_norm": 5.087693214416504,
      "learning_rate": 0.00017808730335601918,
      "loss": 1.33,
      "step": 76700
    },
    {
      "epoch": 2.2915236019466754,
      "grad_norm": 4.278517723083496,
      "learning_rate": 0.00017807301756010035,
      "loss": 1.2854,
      "step": 76750
    },
    {
      "epoch": 2.293016451198758,
      "grad_norm": 4.385780334472656,
      "learning_rate": 0.0001780587317641815,
      "loss": 1.2567,
      "step": 76800
    },
    {
      "epoch": 2.2945093004508403,
      "grad_norm": 4.630320072174072,
      "learning_rate": 0.00017804444596826268,
      "loss": 1.2706,
      "step": 76850
    },
    {
      "epoch": 2.296002149702923,
      "grad_norm": 6.519657135009766,
      "learning_rate": 0.00017803016017234384,
      "loss": 1.2892,
      "step": 76900
    },
    {
      "epoch": 2.2974949989550053,
      "grad_norm": 5.163790702819824,
      "learning_rate": 0.000178015874376425,
      "loss": 1.3251,
      "step": 76950
    },
    {
      "epoch": 2.298987848207088,
      "grad_norm": 4.371760368347168,
      "learning_rate": 0.00017800158858050617,
      "loss": 1.3014,
      "step": 77000
    },
    {
      "epoch": 2.3004806974591707,
      "grad_norm": 4.700384616851807,
      "learning_rate": 0.00017798730278458736,
      "loss": 1.2648,
      "step": 77050
    },
    {
      "epoch": 2.301973546711253,
      "grad_norm": 4.602656364440918,
      "learning_rate": 0.0001779730169886685,
      "loss": 1.3043,
      "step": 77100
    },
    {
      "epoch": 2.3034663959633357,
      "grad_norm": 4.436239719390869,
      "learning_rate": 0.0001779587311927497,
      "loss": 1.2702,
      "step": 77150
    },
    {
      "epoch": 2.304959245215418,
      "grad_norm": 5.805758476257324,
      "learning_rate": 0.00017794444539683083,
      "loss": 1.2556,
      "step": 77200
    },
    {
      "epoch": 2.3064520944675007,
      "grad_norm": 5.682462215423584,
      "learning_rate": 0.00017793015960091202,
      "loss": 1.2422,
      "step": 77250
    },
    {
      "epoch": 2.3079449437195834,
      "grad_norm": 5.270015716552734,
      "learning_rate": 0.00017791587380499316,
      "loss": 1.2666,
      "step": 77300
    },
    {
      "epoch": 2.3094377929716656,
      "grad_norm": 5.771697998046875,
      "learning_rate": 0.00017790158800907435,
      "loss": 1.2735,
      "step": 77350
    },
    {
      "epoch": 2.3109306422237483,
      "grad_norm": 4.319996356964111,
      "learning_rate": 0.00017788730221315551,
      "loss": 1.2362,
      "step": 77400
    },
    {
      "epoch": 2.3124234914758306,
      "grad_norm": 7.646838188171387,
      "learning_rate": 0.00017787301641723668,
      "loss": 1.2998,
      "step": 77450
    },
    {
      "epoch": 2.3139163407279133,
      "grad_norm": 4.850773334503174,
      "learning_rate": 0.00017785873062131784,
      "loss": 1.2505,
      "step": 77500
    },
    {
      "epoch": 2.315409189979996,
      "grad_norm": 8.903724670410156,
      "learning_rate": 0.000177844444825399,
      "loss": 1.2483,
      "step": 77550
    },
    {
      "epoch": 2.3169020392320783,
      "grad_norm": 4.112529754638672,
      "learning_rate": 0.00017783015902948017,
      "loss": 1.2622,
      "step": 77600
    },
    {
      "epoch": 2.318394888484161,
      "grad_norm": 5.173572063446045,
      "learning_rate": 0.00017781587323356134,
      "loss": 1.2783,
      "step": 77650
    },
    {
      "epoch": 2.3198877377362432,
      "grad_norm": 3.9499335289001465,
      "learning_rate": 0.0001778015874376425,
      "loss": 1.3203,
      "step": 77700
    },
    {
      "epoch": 2.321380586988326,
      "grad_norm": 4.4116926193237305,
      "learning_rate": 0.00017778730164172367,
      "loss": 1.2819,
      "step": 77750
    },
    {
      "epoch": 2.3228734362404087,
      "grad_norm": 5.434197902679443,
      "learning_rate": 0.00017777301584580483,
      "loss": 1.2591,
      "step": 77800
    },
    {
      "epoch": 2.324366285492491,
      "grad_norm": 3.9472904205322266,
      "learning_rate": 0.00017775873004988602,
      "loss": 1.2609,
      "step": 77850
    },
    {
      "epoch": 2.3258591347445736,
      "grad_norm": 4.981082439422607,
      "learning_rate": 0.00017774444425396716,
      "loss": 1.3558,
      "step": 77900
    },
    {
      "epoch": 2.327351983996656,
      "grad_norm": 4.96250057220459,
      "learning_rate": 0.00017773015845804835,
      "loss": 1.2478,
      "step": 77950
    },
    {
      "epoch": 2.3288448332487386,
      "grad_norm": 7.093445301055908,
      "learning_rate": 0.0001777158726621295,
      "loss": 1.2659,
      "step": 78000
    },
    {
      "epoch": 2.3303376825008213,
      "grad_norm": 4.415938854217529,
      "learning_rate": 0.00017770158686621068,
      "loss": 1.2994,
      "step": 78050
    },
    {
      "epoch": 2.3318305317529036,
      "grad_norm": 6.983063697814941,
      "learning_rate": 0.00017768730107029185,
      "loss": 1.2994,
      "step": 78100
    },
    {
      "epoch": 2.3333233810049863,
      "grad_norm": 6.34420919418335,
      "learning_rate": 0.000177673015274373,
      "loss": 1.3172,
      "step": 78150
    },
    {
      "epoch": 2.3348162302570685,
      "grad_norm": 4.156967639923096,
      "learning_rate": 0.00017765872947845418,
      "loss": 1.2425,
      "step": 78200
    },
    {
      "epoch": 2.3363090795091512,
      "grad_norm": 4.001344680786133,
      "learning_rate": 0.00017764444368253534,
      "loss": 1.3034,
      "step": 78250
    },
    {
      "epoch": 2.337801928761234,
      "grad_norm": 5.008638381958008,
      "learning_rate": 0.0001776301578866165,
      "loss": 1.2912,
      "step": 78300
    },
    {
      "epoch": 2.339294778013316,
      "grad_norm": 3.917297840118408,
      "learning_rate": 0.00017761587209069767,
      "loss": 1.2807,
      "step": 78350
    },
    {
      "epoch": 2.340787627265399,
      "grad_norm": 4.353655815124512,
      "learning_rate": 0.00017760158629477884,
      "loss": 1.3237,
      "step": 78400
    },
    {
      "epoch": 2.342280476517481,
      "grad_norm": 5.5648698806762695,
      "learning_rate": 0.00017758730049886,
      "loss": 1.3145,
      "step": 78450
    },
    {
      "epoch": 2.343773325769564,
      "grad_norm": 3.928065538406372,
      "learning_rate": 0.00017757301470294116,
      "loss": 1.297,
      "step": 78500
    },
    {
      "epoch": 2.345266175021646,
      "grad_norm": 4.029613494873047,
      "learning_rate": 0.00017755872890702233,
      "loss": 1.2477,
      "step": 78550
    },
    {
      "epoch": 2.346759024273729,
      "grad_norm": 3.9108121395111084,
      "learning_rate": 0.0001775444431111035,
      "loss": 1.2537,
      "step": 78600
    },
    {
      "epoch": 2.3482518735258116,
      "grad_norm": 4.727762699127197,
      "learning_rate": 0.00017753015731518469,
      "loss": 1.2956,
      "step": 78650
    },
    {
      "epoch": 2.349744722777894,
      "grad_norm": 4.287507057189941,
      "learning_rate": 0.00017751587151926582,
      "loss": 1.2689,
      "step": 78700
    },
    {
      "epoch": 2.3512375720299765,
      "grad_norm": 4.162285804748535,
      "learning_rate": 0.00017750158572334701,
      "loss": 1.2803,
      "step": 78750
    },
    {
      "epoch": 2.352730421282059,
      "grad_norm": 5.458852291107178,
      "learning_rate": 0.00017748729992742815,
      "loss": 1.2789,
      "step": 78800
    },
    {
      "epoch": 2.3542232705341415,
      "grad_norm": 4.103569984436035,
      "learning_rate": 0.00017747301413150934,
      "loss": 1.3299,
      "step": 78850
    },
    {
      "epoch": 2.3557161197862238,
      "grad_norm": 4.286417007446289,
      "learning_rate": 0.0001774587283355905,
      "loss": 1.3333,
      "step": 78900
    },
    {
      "epoch": 2.3572089690383065,
      "grad_norm": 4.811729431152344,
      "learning_rate": 0.00017744444253967167,
      "loss": 1.3414,
      "step": 78950
    },
    {
      "epoch": 2.358701818290389,
      "grad_norm": 3.960171699523926,
      "learning_rate": 0.00017743015674375284,
      "loss": 1.2661,
      "step": 79000
    },
    {
      "epoch": 2.3601946675424714,
      "grad_norm": 5.161068439483643,
      "learning_rate": 0.000177415870947834,
      "loss": 1.2622,
      "step": 79050
    },
    {
      "epoch": 2.361687516794554,
      "grad_norm": 5.81963586807251,
      "learning_rate": 0.00017740158515191517,
      "loss": 1.3336,
      "step": 79100
    },
    {
      "epoch": 2.3631803660466364,
      "grad_norm": 5.815190315246582,
      "learning_rate": 0.00017738729935599633,
      "loss": 1.3078,
      "step": 79150
    },
    {
      "epoch": 2.364673215298719,
      "grad_norm": 3.994666814804077,
      "learning_rate": 0.0001773730135600775,
      "loss": 1.2237,
      "step": 79200
    },
    {
      "epoch": 2.366166064550802,
      "grad_norm": 5.232024669647217,
      "learning_rate": 0.00017735872776415866,
      "loss": 1.3267,
      "step": 79250
    },
    {
      "epoch": 2.367658913802884,
      "grad_norm": 4.517995357513428,
      "learning_rate": 0.00017734444196823983,
      "loss": 1.3168,
      "step": 79300
    },
    {
      "epoch": 2.3691517630549668,
      "grad_norm": 4.49736213684082,
      "learning_rate": 0.000177330156172321,
      "loss": 1.3403,
      "step": 79350
    },
    {
      "epoch": 2.370644612307049,
      "grad_norm": 5.295841693878174,
      "learning_rate": 0.00017731587037640216,
      "loss": 1.2639,
      "step": 79400
    },
    {
      "epoch": 2.3721374615591317,
      "grad_norm": 6.010612487792969,
      "learning_rate": 0.00017730158458048332,
      "loss": 1.273,
      "step": 79450
    },
    {
      "epoch": 2.3736303108112144,
      "grad_norm": 4.014612674713135,
      "learning_rate": 0.00017728729878456448,
      "loss": 1.3027,
      "step": 79500
    },
    {
      "epoch": 2.3751231600632967,
      "grad_norm": 5.472930908203125,
      "learning_rate": 0.00017727301298864565,
      "loss": 1.2927,
      "step": 79550
    },
    {
      "epoch": 2.3766160093153794,
      "grad_norm": 5.090315341949463,
      "learning_rate": 0.00017725872719272681,
      "loss": 1.2994,
      "step": 79600
    },
    {
      "epoch": 2.3781088585674617,
      "grad_norm": 4.583268165588379,
      "learning_rate": 0.00017724444139680798,
      "loss": 1.3249,
      "step": 79650
    },
    {
      "epoch": 2.3796017078195444,
      "grad_norm": 5.765147686004639,
      "learning_rate": 0.00017723015560088917,
      "loss": 1.2582,
      "step": 79700
    },
    {
      "epoch": 2.381094557071627,
      "grad_norm": 4.751801490783691,
      "learning_rate": 0.0001772158698049703,
      "loss": 1.2497,
      "step": 79750
    },
    {
      "epoch": 2.3825874063237094,
      "grad_norm": 6.316953182220459,
      "learning_rate": 0.0001772015840090515,
      "loss": 1.2949,
      "step": 79800
    },
    {
      "epoch": 2.384080255575792,
      "grad_norm": 4.6090779304504395,
      "learning_rate": 0.00017718729821313264,
      "loss": 1.2922,
      "step": 79850
    },
    {
      "epoch": 2.3855731048278743,
      "grad_norm": 5.004429340362549,
      "learning_rate": 0.00017717301241721383,
      "loss": 1.336,
      "step": 79900
    },
    {
      "epoch": 2.387065954079957,
      "grad_norm": 5.7317352294921875,
      "learning_rate": 0.00017715872662129497,
      "loss": 1.2296,
      "step": 79950
    },
    {
      "epoch": 2.3885588033320397,
      "grad_norm": 4.938337326049805,
      "learning_rate": 0.00017714444082537616,
      "loss": 1.3118,
      "step": 80000
    },
    {
      "epoch": 2.390051652584122,
      "grad_norm": 7.3811869621276855,
      "learning_rate": 0.00017713015502945732,
      "loss": 1.2989,
      "step": 80050
    },
    {
      "epoch": 2.3915445018362047,
      "grad_norm": 4.3925604820251465,
      "learning_rate": 0.0001771158692335385,
      "loss": 1.3183,
      "step": 80100
    },
    {
      "epoch": 2.393037351088287,
      "grad_norm": 3.382279396057129,
      "learning_rate": 0.00017710158343761965,
      "loss": 1.2749,
      "step": 80150
    },
    {
      "epoch": 2.3945302003403697,
      "grad_norm": 6.085289478302002,
      "learning_rate": 0.00017708729764170082,
      "loss": 1.2881,
      "step": 80200
    },
    {
      "epoch": 2.3960230495924524,
      "grad_norm": 3.96476149559021,
      "learning_rate": 0.00017707301184578198,
      "loss": 1.2656,
      "step": 80250
    },
    {
      "epoch": 2.3975158988445346,
      "grad_norm": 4.456584930419922,
      "learning_rate": 0.00017705872604986315,
      "loss": 1.2969,
      "step": 80300
    },
    {
      "epoch": 2.3990087480966173,
      "grad_norm": 5.09492301940918,
      "learning_rate": 0.0001770444402539443,
      "loss": 1.2679,
      "step": 80350
    },
    {
      "epoch": 2.4005015973486996,
      "grad_norm": 4.211208820343018,
      "learning_rate": 0.00017703015445802548,
      "loss": 1.2235,
      "step": 80400
    },
    {
      "epoch": 2.4019944466007823,
      "grad_norm": 4.683793067932129,
      "learning_rate": 0.00017701586866210664,
      "loss": 1.2791,
      "step": 80450
    },
    {
      "epoch": 2.403487295852865,
      "grad_norm": 4.003592014312744,
      "learning_rate": 0.00017700158286618783,
      "loss": 1.2498,
      "step": 80500
    },
    {
      "epoch": 2.4049801451049473,
      "grad_norm": 5.065978527069092,
      "learning_rate": 0.00017698729707026897,
      "loss": 1.2681,
      "step": 80550
    },
    {
      "epoch": 2.40647299435703,
      "grad_norm": 4.453160762786865,
      "learning_rate": 0.00017697301127435016,
      "loss": 1.3567,
      "step": 80600
    },
    {
      "epoch": 2.4079658436091123,
      "grad_norm": 4.852217197418213,
      "learning_rate": 0.0001769587254784313,
      "loss": 1.2808,
      "step": 80650
    },
    {
      "epoch": 2.409458692861195,
      "grad_norm": 4.629857063293457,
      "learning_rate": 0.0001769444396825125,
      "loss": 1.2994,
      "step": 80700
    },
    {
      "epoch": 2.4109515421132772,
      "grad_norm": 4.611933708190918,
      "learning_rate": 0.00017693015388659363,
      "loss": 1.2477,
      "step": 80750
    },
    {
      "epoch": 2.41244439136536,
      "grad_norm": 4.3425822257995605,
      "learning_rate": 0.00017691586809067482,
      "loss": 1.2861,
      "step": 80800
    },
    {
      "epoch": 2.4139372406174426,
      "grad_norm": 5.061252593994141,
      "learning_rate": 0.00017690158229475598,
      "loss": 1.345,
      "step": 80850
    },
    {
      "epoch": 2.415430089869525,
      "grad_norm": 5.230067253112793,
      "learning_rate": 0.00017688729649883715,
      "loss": 1.3097,
      "step": 80900
    },
    {
      "epoch": 2.4169229391216076,
      "grad_norm": 4.092568874359131,
      "learning_rate": 0.00017687301070291831,
      "loss": 1.2609,
      "step": 80950
    },
    {
      "epoch": 2.41841578837369,
      "grad_norm": 7.115472316741943,
      "learning_rate": 0.00017685872490699948,
      "loss": 1.3306,
      "step": 81000
    },
    {
      "epoch": 2.4199086376257726,
      "grad_norm": 4.142023086547852,
      "learning_rate": 0.00017684443911108064,
      "loss": 1.2715,
      "step": 81050
    },
    {
      "epoch": 2.421401486877855,
      "grad_norm": 4.5259575843811035,
      "learning_rate": 0.0001768301533151618,
      "loss": 1.3097,
      "step": 81100
    },
    {
      "epoch": 2.4228943361299375,
      "grad_norm": 4.749493598937988,
      "learning_rate": 0.00017681586751924297,
      "loss": 1.2413,
      "step": 81150
    },
    {
      "epoch": 2.4243871853820202,
      "grad_norm": 3.844097137451172,
      "learning_rate": 0.00017680158172332414,
      "loss": 1.3365,
      "step": 81200
    },
    {
      "epoch": 2.4258800346341025,
      "grad_norm": 4.504172325134277,
      "learning_rate": 0.0001767872959274053,
      "loss": 1.2889,
      "step": 81250
    },
    {
      "epoch": 2.427372883886185,
      "grad_norm": 4.036706447601318,
      "learning_rate": 0.0001767730101314865,
      "loss": 1.3041,
      "step": 81300
    },
    {
      "epoch": 2.4288657331382675,
      "grad_norm": 4.100889205932617,
      "learning_rate": 0.00017675872433556763,
      "loss": 1.2886,
      "step": 81350
    },
    {
      "epoch": 2.43035858239035,
      "grad_norm": 5.082368850708008,
      "learning_rate": 0.00017674443853964882,
      "loss": 1.333,
      "step": 81400
    },
    {
      "epoch": 2.431851431642433,
      "grad_norm": 4.644407749176025,
      "learning_rate": 0.00017673015274372996,
      "loss": 1.3191,
      "step": 81450
    },
    {
      "epoch": 2.433344280894515,
      "grad_norm": 4.445605754852295,
      "learning_rate": 0.00017671586694781115,
      "loss": 1.2961,
      "step": 81500
    },
    {
      "epoch": 2.434837130146598,
      "grad_norm": 4.809011936187744,
      "learning_rate": 0.00017670158115189232,
      "loss": 1.2496,
      "step": 81550
    },
    {
      "epoch": 2.43632997939868,
      "grad_norm": 5.294357776641846,
      "learning_rate": 0.00017668729535597348,
      "loss": 1.2885,
      "step": 81600
    },
    {
      "epoch": 2.437822828650763,
      "grad_norm": 4.103408336639404,
      "learning_rate": 0.00017667300956005465,
      "loss": 1.2612,
      "step": 81650
    },
    {
      "epoch": 2.4393156779028455,
      "grad_norm": 4.427671909332275,
      "learning_rate": 0.0001766587237641358,
      "loss": 1.2849,
      "step": 81700
    },
    {
      "epoch": 2.440808527154928,
      "grad_norm": 5.336637496948242,
      "learning_rate": 0.00017664443796821698,
      "loss": 1.2923,
      "step": 81750
    },
    {
      "epoch": 2.4423013764070105,
      "grad_norm": 5.235950946807861,
      "learning_rate": 0.00017663015217229814,
      "loss": 1.3086,
      "step": 81800
    },
    {
      "epoch": 2.4437942256590928,
      "grad_norm": 5.3591766357421875,
      "learning_rate": 0.0001766158663763793,
      "loss": 1.3326,
      "step": 81850
    },
    {
      "epoch": 2.4452870749111755,
      "grad_norm": 5.432180881500244,
      "learning_rate": 0.00017660158058046047,
      "loss": 1.2918,
      "step": 81900
    },
    {
      "epoch": 2.446779924163258,
      "grad_norm": 5.184196949005127,
      "learning_rate": 0.00017658729478454163,
      "loss": 1.2652,
      "step": 81950
    },
    {
      "epoch": 2.4482727734153404,
      "grad_norm": 4.962531566619873,
      "learning_rate": 0.0001765730089886228,
      "loss": 1.2579,
      "step": 82000
    },
    {
      "epoch": 2.449765622667423,
      "grad_norm": 3.7370457649230957,
      "learning_rate": 0.00017655872319270396,
      "loss": 1.3204,
      "step": 82050
    },
    {
      "epoch": 2.4512584719195054,
      "grad_norm": 4.21932315826416,
      "learning_rate": 0.00017654443739678513,
      "loss": 1.333,
      "step": 82100
    },
    {
      "epoch": 2.452751321171588,
      "grad_norm": 4.366473197937012,
      "learning_rate": 0.0001765301516008663,
      "loss": 1.2594,
      "step": 82150
    },
    {
      "epoch": 2.454244170423671,
      "grad_norm": 5.760409832000732,
      "learning_rate": 0.00017651586580494746,
      "loss": 1.3004,
      "step": 82200
    },
    {
      "epoch": 2.455737019675753,
      "grad_norm": 4.652368545532227,
      "learning_rate": 0.00017650158000902862,
      "loss": 1.3092,
      "step": 82250
    },
    {
      "epoch": 2.457229868927836,
      "grad_norm": 3.5619237422943115,
      "learning_rate": 0.0001764872942131098,
      "loss": 1.3315,
      "step": 82300
    },
    {
      "epoch": 2.458722718179918,
      "grad_norm": 3.285752534866333,
      "learning_rate": 0.00017647300841719098,
      "loss": 1.2906,
      "step": 82350
    },
    {
      "epoch": 2.4602155674320008,
      "grad_norm": 4.153928279876709,
      "learning_rate": 0.00017645872262127212,
      "loss": 1.3055,
      "step": 82400
    },
    {
      "epoch": 2.4617084166840835,
      "grad_norm": 3.767779588699341,
      "learning_rate": 0.0001764444368253533,
      "loss": 1.264,
      "step": 82450
    },
    {
      "epoch": 2.4632012659361657,
      "grad_norm": 4.4513983726501465,
      "learning_rate": 0.00017643015102943445,
      "loss": 1.2749,
      "step": 82500
    },
    {
      "epoch": 2.4646941151882484,
      "grad_norm": 4.600104808807373,
      "learning_rate": 0.00017641586523351564,
      "loss": 1.3194,
      "step": 82550
    },
    {
      "epoch": 2.4661869644403307,
      "grad_norm": 5.654660701751709,
      "learning_rate": 0.00017640157943759677,
      "loss": 1.2525,
      "step": 82600
    },
    {
      "epoch": 2.4676798136924134,
      "grad_norm": 5.852449893951416,
      "learning_rate": 0.00017638729364167797,
      "loss": 1.2609,
      "step": 82650
    },
    {
      "epoch": 2.469172662944496,
      "grad_norm": 4.8410139083862305,
      "learning_rate": 0.00017637300784575913,
      "loss": 1.2892,
      "step": 82700
    },
    {
      "epoch": 2.4706655121965784,
      "grad_norm": 4.150029182434082,
      "learning_rate": 0.0001763587220498403,
      "loss": 1.2893,
      "step": 82750
    },
    {
      "epoch": 2.472158361448661,
      "grad_norm": 4.355432033538818,
      "learning_rate": 0.00017634443625392146,
      "loss": 1.294,
      "step": 82800
    },
    {
      "epoch": 2.4736512107007433,
      "grad_norm": 4.689866542816162,
      "learning_rate": 0.00017633015045800262,
      "loss": 1.3327,
      "step": 82850
    },
    {
      "epoch": 2.475144059952826,
      "grad_norm": 5.386442184448242,
      "learning_rate": 0.0001763158646620838,
      "loss": 1.3027,
      "step": 82900
    },
    {
      "epoch": 2.4766369092049083,
      "grad_norm": 4.756774425506592,
      "learning_rate": 0.00017630157886616495,
      "loss": 1.2994,
      "step": 82950
    },
    {
      "epoch": 2.478129758456991,
      "grad_norm": 5.8331451416015625,
      "learning_rate": 0.00017628729307024612,
      "loss": 1.298,
      "step": 83000
    },
    {
      "epoch": 2.4796226077090737,
      "grad_norm": 4.216109275817871,
      "learning_rate": 0.00017627300727432728,
      "loss": 1.3288,
      "step": 83050
    },
    {
      "epoch": 2.481115456961156,
      "grad_norm": 5.123735427856445,
      "learning_rate": 0.00017625872147840845,
      "loss": 1.3103,
      "step": 83100
    },
    {
      "epoch": 2.4826083062132387,
      "grad_norm": 3.7272541522979736,
      "learning_rate": 0.00017624443568248964,
      "loss": 1.2753,
      "step": 83150
    },
    {
      "epoch": 2.484101155465321,
      "grad_norm": 5.656236171722412,
      "learning_rate": 0.00017623014988657078,
      "loss": 1.3013,
      "step": 83200
    },
    {
      "epoch": 2.4855940047174037,
      "grad_norm": 4.702991962432861,
      "learning_rate": 0.00017621586409065197,
      "loss": 1.3073,
      "step": 83250
    },
    {
      "epoch": 2.487086853969486,
      "grad_norm": 4.512374401092529,
      "learning_rate": 0.0001762015782947331,
      "loss": 1.2951,
      "step": 83300
    },
    {
      "epoch": 2.4885797032215686,
      "grad_norm": 3.560082197189331,
      "learning_rate": 0.0001761872924988143,
      "loss": 1.3083,
      "step": 83350
    },
    {
      "epoch": 2.4900725524736513,
      "grad_norm": 5.189292907714844,
      "learning_rate": 0.00017617300670289544,
      "loss": 1.3153,
      "step": 83400
    },
    {
      "epoch": 2.4915654017257336,
      "grad_norm": 4.624311447143555,
      "learning_rate": 0.00017615872090697663,
      "loss": 1.2606,
      "step": 83450
    },
    {
      "epoch": 2.4930582509778163,
      "grad_norm": 4.030190944671631,
      "learning_rate": 0.0001761444351110578,
      "loss": 1.2552,
      "step": 83500
    },
    {
      "epoch": 2.4945511002298986,
      "grad_norm": 5.9866042137146,
      "learning_rate": 0.00017613014931513896,
      "loss": 1.2818,
      "step": 83550
    },
    {
      "epoch": 2.4960439494819813,
      "grad_norm": 4.025645732879639,
      "learning_rate": 0.00017611586351922012,
      "loss": 1.325,
      "step": 83600
    },
    {
      "epoch": 2.497536798734064,
      "grad_norm": 5.12217903137207,
      "learning_rate": 0.0001761015777233013,
      "loss": 1.2656,
      "step": 83650
    },
    {
      "epoch": 2.4990296479861462,
      "grad_norm": 3.5725717544555664,
      "learning_rate": 0.00017608729192738245,
      "loss": 1.2605,
      "step": 83700
    },
    {
      "epoch": 2.500522497238229,
      "grad_norm": 6.810322284698486,
      "learning_rate": 0.00017607300613146362,
      "loss": 1.3275,
      "step": 83750
    },
    {
      "epoch": 2.502015346490311,
      "grad_norm": 6.723217487335205,
      "learning_rate": 0.00017605872033554478,
      "loss": 1.3038,
      "step": 83800
    },
    {
      "epoch": 2.503508195742394,
      "grad_norm": 5.044365882873535,
      "learning_rate": 0.00017604443453962595,
      "loss": 1.2821,
      "step": 83850
    },
    {
      "epoch": 2.5050010449944766,
      "grad_norm": 4.397944450378418,
      "learning_rate": 0.0001760301487437071,
      "loss": 1.3281,
      "step": 83900
    },
    {
      "epoch": 2.506493894246559,
      "grad_norm": 3.9087088108062744,
      "learning_rate": 0.0001760158629477883,
      "loss": 1.2674,
      "step": 83950
    },
    {
      "epoch": 2.5079867434986416,
      "grad_norm": 4.3125128746032715,
      "learning_rate": 0.00017600157715186944,
      "loss": 1.2654,
      "step": 84000
    },
    {
      "epoch": 2.509479592750724,
      "grad_norm": 5.792150020599365,
      "learning_rate": 0.00017598729135595063,
      "loss": 1.3846,
      "step": 84050
    },
    {
      "epoch": 2.5109724420028066,
      "grad_norm": 4.449563026428223,
      "learning_rate": 0.00017597300556003177,
      "loss": 1.2402,
      "step": 84100
    },
    {
      "epoch": 2.5124652912548893,
      "grad_norm": 3.286771774291992,
      "learning_rate": 0.00017595871976411296,
      "loss": 1.2547,
      "step": 84150
    },
    {
      "epoch": 2.5139581405069715,
      "grad_norm": 4.84338903427124,
      "learning_rate": 0.0001759444339681941,
      "loss": 1.2127,
      "step": 84200
    },
    {
      "epoch": 2.5154509897590542,
      "grad_norm": 4.913907051086426,
      "learning_rate": 0.0001759301481722753,
      "loss": 1.3453,
      "step": 84250
    },
    {
      "epoch": 2.5169438390111365,
      "grad_norm": 3.828268527984619,
      "learning_rate": 0.00017591586237635645,
      "loss": 1.3056,
      "step": 84300
    },
    {
      "epoch": 2.518436688263219,
      "grad_norm": 4.86277961730957,
      "learning_rate": 0.00017590157658043762,
      "loss": 1.2613,
      "step": 84350
    },
    {
      "epoch": 2.519929537515302,
      "grad_norm": 3.3477306365966797,
      "learning_rate": 0.00017588729078451878,
      "loss": 1.2784,
      "step": 84400
    },
    {
      "epoch": 2.521422386767384,
      "grad_norm": 3.982935905456543,
      "learning_rate": 0.00017587300498859995,
      "loss": 1.279,
      "step": 84450
    },
    {
      "epoch": 2.522915236019467,
      "grad_norm": 4.321762561798096,
      "learning_rate": 0.0001758587191926811,
      "loss": 1.2706,
      "step": 84500
    },
    {
      "epoch": 2.524408085271549,
      "grad_norm": 3.989497423171997,
      "learning_rate": 0.00017584443339676228,
      "loss": 1.2431,
      "step": 84550
    },
    {
      "epoch": 2.525900934523632,
      "grad_norm": 4.457726955413818,
      "learning_rate": 0.00017583014760084344,
      "loss": 1.2686,
      "step": 84600
    },
    {
      "epoch": 2.5273937837757146,
      "grad_norm": 4.100541114807129,
      "learning_rate": 0.0001758158618049246,
      "loss": 1.2379,
      "step": 84650
    },
    {
      "epoch": 2.528886633027797,
      "grad_norm": 5.386853218078613,
      "learning_rate": 0.00017580157600900577,
      "loss": 1.2767,
      "step": 84700
    },
    {
      "epoch": 2.5303794822798795,
      "grad_norm": 4.51826286315918,
      "learning_rate": 0.00017578729021308694,
      "loss": 1.2072,
      "step": 84750
    },
    {
      "epoch": 2.531872331531962,
      "grad_norm": 7.255303382873535,
      "learning_rate": 0.0001757730044171681,
      "loss": 1.2324,
      "step": 84800
    },
    {
      "epoch": 2.5333651807840445,
      "grad_norm": 4.7532572746276855,
      "learning_rate": 0.00017575871862124927,
      "loss": 1.2534,
      "step": 84850
    },
    {
      "epoch": 2.534858030036127,
      "grad_norm": 4.200450897216797,
      "learning_rate": 0.00017574443282533043,
      "loss": 1.2732,
      "step": 84900
    },
    {
      "epoch": 2.5363508792882095,
      "grad_norm": 4.062733173370361,
      "learning_rate": 0.0001757301470294116,
      "loss": 1.2987,
      "step": 84950
    },
    {
      "epoch": 2.537843728540292,
      "grad_norm": 4.786027908325195,
      "learning_rate": 0.00017571586123349279,
      "loss": 1.2593,
      "step": 85000
    },
    {
      "epoch": 2.5393365777923744,
      "grad_norm": 6.346192359924316,
      "learning_rate": 0.00017570157543757392,
      "loss": 1.2854,
      "step": 85050
    },
    {
      "epoch": 2.540829427044457,
      "grad_norm": 5.030992031097412,
      "learning_rate": 0.00017568728964165512,
      "loss": 1.2498,
      "step": 85100
    },
    {
      "epoch": 2.54232227629654,
      "grad_norm": 3.7949612140655518,
      "learning_rate": 0.00017567300384573625,
      "loss": 1.275,
      "step": 85150
    },
    {
      "epoch": 2.543815125548622,
      "grad_norm": 4.442857265472412,
      "learning_rate": 0.00017565871804981745,
      "loss": 1.2156,
      "step": 85200
    },
    {
      "epoch": 2.5453079748007044,
      "grad_norm": 5.143223285675049,
      "learning_rate": 0.00017564443225389858,
      "loss": 1.3554,
      "step": 85250
    },
    {
      "epoch": 2.546800824052787,
      "grad_norm": 4.240771770477295,
      "learning_rate": 0.00017563014645797977,
      "loss": 1.3288,
      "step": 85300
    },
    {
      "epoch": 2.5482936733048698,
      "grad_norm": 5.060192108154297,
      "learning_rate": 0.00017561586066206094,
      "loss": 1.3284,
      "step": 85350
    },
    {
      "epoch": 2.549786522556952,
      "grad_norm": 4.155733585357666,
      "learning_rate": 0.0001756015748661421,
      "loss": 1.2422,
      "step": 85400
    },
    {
      "epoch": 2.5512793718090347,
      "grad_norm": 5.4249749183654785,
      "learning_rate": 0.00017558728907022327,
      "loss": 1.262,
      "step": 85450
    },
    {
      "epoch": 2.552772221061117,
      "grad_norm": 5.704741954803467,
      "learning_rate": 0.00017557300327430443,
      "loss": 1.308,
      "step": 85500
    },
    {
      "epoch": 2.5542650703131997,
      "grad_norm": 4.013110637664795,
      "learning_rate": 0.0001755587174783856,
      "loss": 1.2667,
      "step": 85550
    },
    {
      "epoch": 2.5557579195652824,
      "grad_norm": 4.5984673500061035,
      "learning_rate": 0.00017554443168246676,
      "loss": 1.2756,
      "step": 85600
    },
    {
      "epoch": 2.5572507688173647,
      "grad_norm": 4.716318130493164,
      "learning_rate": 0.00017553014588654793,
      "loss": 1.2896,
      "step": 85650
    },
    {
      "epoch": 2.5587436180694474,
      "grad_norm": 3.9019134044647217,
      "learning_rate": 0.0001755158600906291,
      "loss": 1.3419,
      "step": 85700
    },
    {
      "epoch": 2.5602364673215297,
      "grad_norm": 4.762269020080566,
      "learning_rate": 0.00017550157429471026,
      "loss": 1.2856,
      "step": 85750
    },
    {
      "epoch": 2.5617293165736124,
      "grad_norm": 4.8936567306518555,
      "learning_rate": 0.00017548728849879145,
      "loss": 1.3033,
      "step": 85800
    },
    {
      "epoch": 2.563222165825695,
      "grad_norm": 5.605986595153809,
      "learning_rate": 0.00017547300270287259,
      "loss": 1.2928,
      "step": 85850
    },
    {
      "epoch": 2.5647150150777773,
      "grad_norm": 5.456994533538818,
      "learning_rate": 0.00017545871690695378,
      "loss": 1.2646,
      "step": 85900
    },
    {
      "epoch": 2.56620786432986,
      "grad_norm": 5.125797748565674,
      "learning_rate": 0.00017544443111103491,
      "loss": 1.3337,
      "step": 85950
    },
    {
      "epoch": 2.5677007135819423,
      "grad_norm": 6.698281288146973,
      "learning_rate": 0.0001754301453151161,
      "loss": 1.231,
      "step": 86000
    },
    {
      "epoch": 2.569193562834025,
      "grad_norm": 4.8110175132751465,
      "learning_rate": 0.00017541585951919724,
      "loss": 1.2829,
      "step": 86050
    },
    {
      "epoch": 2.5706864120861077,
      "grad_norm": 4.389436721801758,
      "learning_rate": 0.00017540157372327844,
      "loss": 1.3394,
      "step": 86100
    },
    {
      "epoch": 2.57217926133819,
      "grad_norm": 4.567478179931641,
      "learning_rate": 0.0001753872879273596,
      "loss": 1.3179,
      "step": 86150
    },
    {
      "epoch": 2.5736721105902727,
      "grad_norm": 6.0036444664001465,
      "learning_rate": 0.00017537300213144077,
      "loss": 1.2837,
      "step": 86200
    },
    {
      "epoch": 2.575164959842355,
      "grad_norm": 4.954610824584961,
      "learning_rate": 0.00017535871633552193,
      "loss": 1.3057,
      "step": 86250
    },
    {
      "epoch": 2.5766578090944376,
      "grad_norm": 3.623244524002075,
      "learning_rate": 0.0001753444305396031,
      "loss": 1.3203,
      "step": 86300
    },
    {
      "epoch": 2.5781506583465204,
      "grad_norm": 3.969407081604004,
      "learning_rate": 0.00017533014474368426,
      "loss": 1.2465,
      "step": 86350
    },
    {
      "epoch": 2.5796435075986026,
      "grad_norm": 4.700910568237305,
      "learning_rate": 0.00017531585894776542,
      "loss": 1.2673,
      "step": 86400
    },
    {
      "epoch": 2.5811363568506853,
      "grad_norm": 4.05939245223999,
      "learning_rate": 0.0001753015731518466,
      "loss": 1.2177,
      "step": 86450
    },
    {
      "epoch": 2.5826292061027676,
      "grad_norm": 4.144320011138916,
      "learning_rate": 0.00017528728735592775,
      "loss": 1.2198,
      "step": 86500
    },
    {
      "epoch": 2.5841220553548503,
      "grad_norm": 4.483496189117432,
      "learning_rate": 0.00017527300156000892,
      "loss": 1.3403,
      "step": 86550
    },
    {
      "epoch": 2.585614904606933,
      "grad_norm": 6.164941787719727,
      "learning_rate": 0.0001752587157640901,
      "loss": 1.3121,
      "step": 86600
    },
    {
      "epoch": 2.5871077538590153,
      "grad_norm": 5.265034198760986,
      "learning_rate": 0.00017524442996817125,
      "loss": 1.2484,
      "step": 86650
    },
    {
      "epoch": 2.588600603111098,
      "grad_norm": 4.718167304992676,
      "learning_rate": 0.00017523014417225244,
      "loss": 1.3256,
      "step": 86700
    },
    {
      "epoch": 2.5900934523631802,
      "grad_norm": 3.9518258571624756,
      "learning_rate": 0.00017521585837633358,
      "loss": 1.271,
      "step": 86750
    },
    {
      "epoch": 2.591586301615263,
      "grad_norm": 4.604781627655029,
      "learning_rate": 0.00017520157258041477,
      "loss": 1.2821,
      "step": 86800
    },
    {
      "epoch": 2.5930791508673456,
      "grad_norm": 3.5907840728759766,
      "learning_rate": 0.0001751872867844959,
      "loss": 1.2958,
      "step": 86850
    },
    {
      "epoch": 2.594572000119428,
      "grad_norm": 5.262489318847656,
      "learning_rate": 0.0001751730009885771,
      "loss": 1.3186,
      "step": 86900
    },
    {
      "epoch": 2.5960648493715106,
      "grad_norm": 4.812371253967285,
      "learning_rate": 0.00017515871519265826,
      "loss": 1.2586,
      "step": 86950
    },
    {
      "epoch": 2.597557698623593,
      "grad_norm": 4.645204544067383,
      "learning_rate": 0.00017514442939673943,
      "loss": 1.2656,
      "step": 87000
    },
    {
      "epoch": 2.5990505478756756,
      "grad_norm": 5.977488040924072,
      "learning_rate": 0.0001751301436008206,
      "loss": 1.3167,
      "step": 87050
    },
    {
      "epoch": 2.6005433971277583,
      "grad_norm": 4.534546375274658,
      "learning_rate": 0.00017511585780490176,
      "loss": 1.3263,
      "step": 87100
    },
    {
      "epoch": 2.6020362463798405,
      "grad_norm": 4.562127113342285,
      "learning_rate": 0.00017510157200898292,
      "loss": 1.3122,
      "step": 87150
    },
    {
      "epoch": 2.6035290956319233,
      "grad_norm": 3.588327646255493,
      "learning_rate": 0.00017508728621306409,
      "loss": 1.2953,
      "step": 87200
    },
    {
      "epoch": 2.6050219448840055,
      "grad_norm": 5.136357307434082,
      "learning_rate": 0.00017507300041714525,
      "loss": 1.2831,
      "step": 87250
    },
    {
      "epoch": 2.606514794136088,
      "grad_norm": 4.8900299072265625,
      "learning_rate": 0.00017505871462122641,
      "loss": 1.2847,
      "step": 87300
    },
    {
      "epoch": 2.608007643388171,
      "grad_norm": 3.7218592166900635,
      "learning_rate": 0.00017504442882530758,
      "loss": 1.2828,
      "step": 87350
    },
    {
      "epoch": 2.609500492640253,
      "grad_norm": 3.4648168087005615,
      "learning_rate": 0.00017503014302938874,
      "loss": 1.2576,
      "step": 87400
    },
    {
      "epoch": 2.6109933418923355,
      "grad_norm": 3.6955080032348633,
      "learning_rate": 0.0001750158572334699,
      "loss": 1.3208,
      "step": 87450
    },
    {
      "epoch": 2.612486191144418,
      "grad_norm": 4.296437740325928,
      "learning_rate": 0.00017500157143755107,
      "loss": 1.2455,
      "step": 87500
    },
    {
      "epoch": 2.613979040396501,
      "grad_norm": 4.3054518699646,
      "learning_rate": 0.00017498728564163224,
      "loss": 1.2722,
      "step": 87550
    },
    {
      "epoch": 2.615471889648583,
      "grad_norm": 6.258134841918945,
      "learning_rate": 0.0001749729998457134,
      "loss": 1.2777,
      "step": 87600
    },
    {
      "epoch": 2.616964738900666,
      "grad_norm": 5.092291355133057,
      "learning_rate": 0.0001749587140497946,
      "loss": 1.2906,
      "step": 87650
    },
    {
      "epoch": 2.618457588152748,
      "grad_norm": 3.986844778060913,
      "learning_rate": 0.00017494442825387573,
      "loss": 1.3326,
      "step": 87700
    },
    {
      "epoch": 2.619950437404831,
      "grad_norm": 5.269211292266846,
      "learning_rate": 0.00017493014245795692,
      "loss": 1.3253,
      "step": 87750
    },
    {
      "epoch": 2.6214432866569135,
      "grad_norm": 3.6939330101013184,
      "learning_rate": 0.00017491585666203806,
      "loss": 1.318,
      "step": 87800
    },
    {
      "epoch": 2.6229361359089958,
      "grad_norm": 5.028196334838867,
      "learning_rate": 0.00017490157086611925,
      "loss": 1.3089,
      "step": 87850
    },
    {
      "epoch": 2.6244289851610785,
      "grad_norm": 5.179739475250244,
      "learning_rate": 0.0001748872850702004,
      "loss": 1.2471,
      "step": 87900
    },
    {
      "epoch": 2.6259218344131607,
      "grad_norm": 4.581385135650635,
      "learning_rate": 0.00017487299927428158,
      "loss": 1.2703,
      "step": 87950
    },
    {
      "epoch": 2.6274146836652434,
      "grad_norm": 4.185626983642578,
      "learning_rate": 0.00017485871347836275,
      "loss": 1.2967,
      "step": 88000
    },
    {
      "epoch": 2.628907532917326,
      "grad_norm": 4.166527271270752,
      "learning_rate": 0.0001748444276824439,
      "loss": 1.282,
      "step": 88050
    },
    {
      "epoch": 2.6304003821694084,
      "grad_norm": 5.894084930419922,
      "learning_rate": 0.00017483014188652508,
      "loss": 1.3337,
      "step": 88100
    },
    {
      "epoch": 2.631893231421491,
      "grad_norm": 5.775843620300293,
      "learning_rate": 0.00017481585609060624,
      "loss": 1.3186,
      "step": 88150
    },
    {
      "epoch": 2.6333860806735734,
      "grad_norm": 4.601424217224121,
      "learning_rate": 0.0001748015702946874,
      "loss": 1.2961,
      "step": 88200
    },
    {
      "epoch": 2.634878929925656,
      "grad_norm": 4.345365524291992,
      "learning_rate": 0.00017478728449876857,
      "loss": 1.2776,
      "step": 88250
    },
    {
      "epoch": 2.636371779177739,
      "grad_norm": 4.751996040344238,
      "learning_rate": 0.00017477299870284973,
      "loss": 1.2847,
      "step": 88300
    },
    {
      "epoch": 2.637864628429821,
      "grad_norm": 4.37894344329834,
      "learning_rate": 0.0001747587129069309,
      "loss": 1.2631,
      "step": 88350
    },
    {
      "epoch": 2.6393574776819038,
      "grad_norm": 4.729623794555664,
      "learning_rate": 0.00017474442711101206,
      "loss": 1.3288,
      "step": 88400
    },
    {
      "epoch": 2.640850326933986,
      "grad_norm": 4.7951788902282715,
      "learning_rate": 0.00017473014131509326,
      "loss": 1.3207,
      "step": 88450
    },
    {
      "epoch": 2.6423431761860687,
      "grad_norm": 3.692270040512085,
      "learning_rate": 0.0001747158555191744,
      "loss": 1.3038,
      "step": 88500
    },
    {
      "epoch": 2.6438360254381514,
      "grad_norm": 4.1009297370910645,
      "learning_rate": 0.00017470156972325559,
      "loss": 1.2874,
      "step": 88550
    },
    {
      "epoch": 2.6453288746902337,
      "grad_norm": 4.919840335845947,
      "learning_rate": 0.00017468728392733672,
      "loss": 1.294,
      "step": 88600
    },
    {
      "epoch": 2.6468217239423164,
      "grad_norm": 3.7501938343048096,
      "learning_rate": 0.00017467299813141791,
      "loss": 1.3196,
      "step": 88650
    },
    {
      "epoch": 2.6483145731943987,
      "grad_norm": 4.218129634857178,
      "learning_rate": 0.00017465871233549905,
      "loss": 1.2823,
      "step": 88700
    },
    {
      "epoch": 2.6498074224464814,
      "grad_norm": 5.627344131469727,
      "learning_rate": 0.00017464442653958024,
      "loss": 1.2796,
      "step": 88750
    },
    {
      "epoch": 2.651300271698564,
      "grad_norm": 4.543562412261963,
      "learning_rate": 0.0001746301407436614,
      "loss": 1.2605,
      "step": 88800
    },
    {
      "epoch": 2.6527931209506463,
      "grad_norm": 4.705561637878418,
      "learning_rate": 0.00017461585494774257,
      "loss": 1.2822,
      "step": 88850
    },
    {
      "epoch": 2.654285970202729,
      "grad_norm": 4.356922149658203,
      "learning_rate": 0.00017460156915182374,
      "loss": 1.3247,
      "step": 88900
    },
    {
      "epoch": 2.6557788194548113,
      "grad_norm": 4.276957988739014,
      "learning_rate": 0.0001745872833559049,
      "loss": 1.3103,
      "step": 88950
    },
    {
      "epoch": 2.657271668706894,
      "grad_norm": 4.122916221618652,
      "learning_rate": 0.00017457299755998607,
      "loss": 1.302,
      "step": 89000
    },
    {
      "epoch": 2.6587645179589767,
      "grad_norm": 4.473240852355957,
      "learning_rate": 0.00017455871176406723,
      "loss": 1.274,
      "step": 89050
    },
    {
      "epoch": 2.660257367211059,
      "grad_norm": 5.299698829650879,
      "learning_rate": 0.0001745444259681484,
      "loss": 1.2879,
      "step": 89100
    },
    {
      "epoch": 2.6617502164631417,
      "grad_norm": 5.3960347175598145,
      "learning_rate": 0.00017453014017222956,
      "loss": 1.3083,
      "step": 89150
    },
    {
      "epoch": 2.663243065715224,
      "grad_norm": 9.475055694580078,
      "learning_rate": 0.00017451585437631073,
      "loss": 1.3123,
      "step": 89200
    },
    {
      "epoch": 2.6647359149673067,
      "grad_norm": 4.01664924621582,
      "learning_rate": 0.00017450156858039192,
      "loss": 1.2673,
      "step": 89250
    },
    {
      "epoch": 2.6662287642193894,
      "grad_norm": 4.74378776550293,
      "learning_rate": 0.00017448728278447306,
      "loss": 1.2704,
      "step": 89300
    },
    {
      "epoch": 2.6677216134714716,
      "grad_norm": 3.8527019023895264,
      "learning_rate": 0.00017447299698855425,
      "loss": 1.2638,
      "step": 89350
    },
    {
      "epoch": 2.6692144627235543,
      "grad_norm": 5.14741849899292,
      "learning_rate": 0.00017445871119263538,
      "loss": 1.2843,
      "step": 89400
    },
    {
      "epoch": 2.6707073119756366,
      "grad_norm": 4.814597129821777,
      "learning_rate": 0.00017444442539671658,
      "loss": 1.2955,
      "step": 89450
    },
    {
      "epoch": 2.6722001612277193,
      "grad_norm": 3.8238747119903564,
      "learning_rate": 0.00017443013960079771,
      "loss": 1.2921,
      "step": 89500
    },
    {
      "epoch": 2.673693010479802,
      "grad_norm": 5.118923187255859,
      "learning_rate": 0.0001744158538048789,
      "loss": 1.3114,
      "step": 89550
    },
    {
      "epoch": 2.6751858597318843,
      "grad_norm": 4.3184380531311035,
      "learning_rate": 0.00017440156800896007,
      "loss": 1.2387,
      "step": 89600
    },
    {
      "epoch": 2.6766787089839665,
      "grad_norm": 4.6351728439331055,
      "learning_rate": 0.00017438728221304123,
      "loss": 1.2524,
      "step": 89650
    },
    {
      "epoch": 2.6781715582360492,
      "grad_norm": 4.391300678253174,
      "learning_rate": 0.0001743729964171224,
      "loss": 1.2591,
      "step": 89700
    },
    {
      "epoch": 2.679664407488132,
      "grad_norm": 3.749375104904175,
      "learning_rate": 0.00017435871062120354,
      "loss": 1.247,
      "step": 89750
    },
    {
      "epoch": 2.681157256740214,
      "grad_norm": 4.260756492614746,
      "learning_rate": 0.00017434442482528473,
      "loss": 1.2672,
      "step": 89800
    },
    {
      "epoch": 2.682650105992297,
      "grad_norm": 3.6358819007873535,
      "learning_rate": 0.00017433013902936587,
      "loss": 1.2941,
      "step": 89850
    },
    {
      "epoch": 2.684142955244379,
      "grad_norm": 4.653028964996338,
      "learning_rate": 0.00017431585323344706,
      "loss": 1.3237,
      "step": 89900
    },
    {
      "epoch": 2.685635804496462,
      "grad_norm": 4.692080497741699,
      "learning_rate": 0.00017430156743752822,
      "loss": 1.325,
      "step": 89950
    },
    {
      "epoch": 2.6871286537485446,
      "grad_norm": 4.317993640899658,
      "learning_rate": 0.0001742872816416094,
      "loss": 1.2491,
      "step": 90000
    },
    {
      "epoch": 2.688621503000627,
      "grad_norm": 3.7262911796569824,
      "learning_rate": 0.00017427299584569055,
      "loss": 1.3218,
      "step": 90050
    },
    {
      "epoch": 2.6901143522527096,
      "grad_norm": 4.5458855628967285,
      "learning_rate": 0.00017425871004977172,
      "loss": 1.3188,
      "step": 90100
    },
    {
      "epoch": 2.691607201504792,
      "grad_norm": 4.8145599365234375,
      "learning_rate": 0.00017424442425385288,
      "loss": 1.2746,
      "step": 90150
    },
    {
      "epoch": 2.6931000507568745,
      "grad_norm": 5.353755474090576,
      "learning_rate": 0.00017423013845793405,
      "loss": 1.2153,
      "step": 90200
    },
    {
      "epoch": 2.6945929000089572,
      "grad_norm": 5.111592769622803,
      "learning_rate": 0.0001742158526620152,
      "loss": 1.2482,
      "step": 90250
    },
    {
      "epoch": 2.6960857492610395,
      "grad_norm": 5.903563022613525,
      "learning_rate": 0.00017420156686609638,
      "loss": 1.2749,
      "step": 90300
    },
    {
      "epoch": 2.697578598513122,
      "grad_norm": 5.421234130859375,
      "learning_rate": 0.00017418728107017754,
      "loss": 1.3023,
      "step": 90350
    },
    {
      "epoch": 2.6990714477652045,
      "grad_norm": 4.8063154220581055,
      "learning_rate": 0.00017417299527425873,
      "loss": 1.3413,
      "step": 90400
    },
    {
      "epoch": 2.700564297017287,
      "grad_norm": 4.831643104553223,
      "learning_rate": 0.00017415870947833987,
      "loss": 1.3733,
      "step": 90450
    },
    {
      "epoch": 2.70205714626937,
      "grad_norm": 5.314348220825195,
      "learning_rate": 0.00017414442368242106,
      "loss": 1.2819,
      "step": 90500
    },
    {
      "epoch": 2.703549995521452,
      "grad_norm": 5.592355728149414,
      "learning_rate": 0.0001741301378865022,
      "loss": 1.2959,
      "step": 90550
    },
    {
      "epoch": 2.705042844773535,
      "grad_norm": 4.042795181274414,
      "learning_rate": 0.0001741158520905834,
      "loss": 1.28,
      "step": 90600
    },
    {
      "epoch": 2.706535694025617,
      "grad_norm": 4.347578525543213,
      "learning_rate": 0.00017410156629466456,
      "loss": 1.3491,
      "step": 90650
    },
    {
      "epoch": 2.7080285432777,
      "grad_norm": 5.4727678298950195,
      "learning_rate": 0.00017408728049874572,
      "loss": 1.3294,
      "step": 90700
    },
    {
      "epoch": 2.7095213925297825,
      "grad_norm": 4.353364944458008,
      "learning_rate": 0.00017407299470282688,
      "loss": 1.27,
      "step": 90750
    },
    {
      "epoch": 2.711014241781865,
      "grad_norm": 3.916985034942627,
      "learning_rate": 0.00017405870890690805,
      "loss": 1.297,
      "step": 90800
    },
    {
      "epoch": 2.7125070910339475,
      "grad_norm": 4.451666831970215,
      "learning_rate": 0.00017404442311098921,
      "loss": 1.3051,
      "step": 90850
    },
    {
      "epoch": 2.7139999402860298,
      "grad_norm": 5.726661205291748,
      "learning_rate": 0.00017403013731507038,
      "loss": 1.3353,
      "step": 90900
    },
    {
      "epoch": 2.7154927895381125,
      "grad_norm": 4.238119125366211,
      "learning_rate": 0.00017401585151915154,
      "loss": 1.3279,
      "step": 90950
    },
    {
      "epoch": 2.716985638790195,
      "grad_norm": 4.4539642333984375,
      "learning_rate": 0.0001740015657232327,
      "loss": 1.2492,
      "step": 91000
    },
    {
      "epoch": 2.7184784880422774,
      "grad_norm": 5.241584300994873,
      "learning_rate": 0.00017398727992731387,
      "loss": 1.3039,
      "step": 91050
    },
    {
      "epoch": 2.71997133729436,
      "grad_norm": 7.063384056091309,
      "learning_rate": 0.00017397299413139506,
      "loss": 1.2849,
      "step": 91100
    },
    {
      "epoch": 2.7214641865464424,
      "grad_norm": 3.5526702404022217,
      "learning_rate": 0.0001739587083354762,
      "loss": 1.3074,
      "step": 91150
    },
    {
      "epoch": 2.722957035798525,
      "grad_norm": 4.808590888977051,
      "learning_rate": 0.0001739444225395574,
      "loss": 1.3124,
      "step": 91200
    },
    {
      "epoch": 2.724449885050608,
      "grad_norm": 6.006200790405273,
      "learning_rate": 0.00017393013674363853,
      "loss": 1.2375,
      "step": 91250
    },
    {
      "epoch": 2.72594273430269,
      "grad_norm": 3.5086710453033447,
      "learning_rate": 0.00017391585094771972,
      "loss": 1.2535,
      "step": 91300
    },
    {
      "epoch": 2.727435583554773,
      "grad_norm": 4.1073832511901855,
      "learning_rate": 0.00017390156515180086,
      "loss": 1.3369,
      "step": 91350
    },
    {
      "epoch": 2.728928432806855,
      "grad_norm": 4.759340286254883,
      "learning_rate": 0.00017388727935588205,
      "loss": 1.2975,
      "step": 91400
    },
    {
      "epoch": 2.7304212820589377,
      "grad_norm": 4.313277244567871,
      "learning_rate": 0.00017387299355996322,
      "loss": 1.2607,
      "step": 91450
    },
    {
      "epoch": 2.7319141313110205,
      "grad_norm": 4.681273937225342,
      "learning_rate": 0.00017385870776404438,
      "loss": 1.2994,
      "step": 91500
    },
    {
      "epoch": 2.7334069805631027,
      "grad_norm": 5.569629192352295,
      "learning_rate": 0.00017384442196812555,
      "loss": 1.29,
      "step": 91550
    },
    {
      "epoch": 2.7348998298151854,
      "grad_norm": 5.809197425842285,
      "learning_rate": 0.0001738301361722067,
      "loss": 1.3245,
      "step": 91600
    },
    {
      "epoch": 2.7363926790672677,
      "grad_norm": 3.6838326454162598,
      "learning_rate": 0.00017381585037628788,
      "loss": 1.2904,
      "step": 91650
    },
    {
      "epoch": 2.7378855283193504,
      "grad_norm": 5.378135681152344,
      "learning_rate": 0.00017380156458036904,
      "loss": 1.2837,
      "step": 91700
    },
    {
      "epoch": 2.739378377571433,
      "grad_norm": 5.717269420623779,
      "learning_rate": 0.0001737872787844502,
      "loss": 1.301,
      "step": 91750
    },
    {
      "epoch": 2.7408712268235154,
      "grad_norm": 5.068012237548828,
      "learning_rate": 0.00017377299298853137,
      "loss": 1.2899,
      "step": 91800
    },
    {
      "epoch": 2.7423640760755976,
      "grad_norm": 5.394426345825195,
      "learning_rate": 0.00017375870719261253,
      "loss": 1.2719,
      "step": 91850
    },
    {
      "epoch": 2.7438569253276803,
      "grad_norm": 4.564161777496338,
      "learning_rate": 0.00017374442139669373,
      "loss": 1.2951,
      "step": 91900
    },
    {
      "epoch": 2.745349774579763,
      "grad_norm": 5.253499507904053,
      "learning_rate": 0.00017373013560077486,
      "loss": 1.2543,
      "step": 91950
    },
    {
      "epoch": 2.7468426238318453,
      "grad_norm": 3.995630979537964,
      "learning_rate": 0.00017371584980485605,
      "loss": 1.2854,
      "step": 92000
    },
    {
      "epoch": 2.748335473083928,
      "grad_norm": 3.954859733581543,
      "learning_rate": 0.0001737015640089372,
      "loss": 1.287,
      "step": 92050
    },
    {
      "epoch": 2.7498283223360103,
      "grad_norm": 4.409485340118408,
      "learning_rate": 0.00017368727821301838,
      "loss": 1.2889,
      "step": 92100
    },
    {
      "epoch": 2.751321171588093,
      "grad_norm": 5.021519660949707,
      "learning_rate": 0.00017367299241709952,
      "loss": 1.3849,
      "step": 92150
    },
    {
      "epoch": 2.7528140208401757,
      "grad_norm": 4.707982063293457,
      "learning_rate": 0.00017365870662118071,
      "loss": 1.3099,
      "step": 92200
    },
    {
      "epoch": 2.754306870092258,
      "grad_norm": 3.8836755752563477,
      "learning_rate": 0.00017364442082526188,
      "loss": 1.2707,
      "step": 92250
    },
    {
      "epoch": 2.7557997193443406,
      "grad_norm": 4.692137718200684,
      "learning_rate": 0.00017363013502934304,
      "loss": 1.3051,
      "step": 92300
    },
    {
      "epoch": 2.757292568596423,
      "grad_norm": 4.50584602355957,
      "learning_rate": 0.0001736158492334242,
      "loss": 1.2969,
      "step": 92350
    },
    {
      "epoch": 2.7587854178485056,
      "grad_norm": 5.027088642120361,
      "learning_rate": 0.00017360156343750535,
      "loss": 1.2846,
      "step": 92400
    },
    {
      "epoch": 2.7602782671005883,
      "grad_norm": 5.339383125305176,
      "learning_rate": 0.00017358727764158654,
      "loss": 1.3153,
      "step": 92450
    },
    {
      "epoch": 2.7617711163526706,
      "grad_norm": 3.3670814037323,
      "learning_rate": 0.00017357299184566767,
      "loss": 1.2268,
      "step": 92500
    },
    {
      "epoch": 2.7632639656047533,
      "grad_norm": 4.426408290863037,
      "learning_rate": 0.00017355870604974887,
      "loss": 1.2737,
      "step": 92550
    },
    {
      "epoch": 2.7647568148568356,
      "grad_norm": 4.831822395324707,
      "learning_rate": 0.00017354442025383003,
      "loss": 1.3036,
      "step": 92600
    },
    {
      "epoch": 2.7662496641089183,
      "grad_norm": 5.008368015289307,
      "learning_rate": 0.0001735301344579112,
      "loss": 1.2846,
      "step": 92650
    },
    {
      "epoch": 2.767742513361001,
      "grad_norm": 4.761803150177002,
      "learning_rate": 0.00017351584866199236,
      "loss": 1.2715,
      "step": 92700
    },
    {
      "epoch": 2.7692353626130832,
      "grad_norm": 4.651097774505615,
      "learning_rate": 0.00017350156286607352,
      "loss": 1.2337,
      "step": 92750
    },
    {
      "epoch": 2.770728211865166,
      "grad_norm": 4.635573387145996,
      "learning_rate": 0.0001734872770701547,
      "loss": 1.3133,
      "step": 92800
    },
    {
      "epoch": 2.772221061117248,
      "grad_norm": 4.60062313079834,
      "learning_rate": 0.00017347299127423585,
      "loss": 1.2319,
      "step": 92850
    },
    {
      "epoch": 2.773713910369331,
      "grad_norm": 5.222136497497559,
      "learning_rate": 0.00017345870547831702,
      "loss": 1.3058,
      "step": 92900
    },
    {
      "epoch": 2.7752067596214136,
      "grad_norm": 5.110498428344727,
      "learning_rate": 0.00017344441968239818,
      "loss": 1.2217,
      "step": 92950
    },
    {
      "epoch": 2.776699608873496,
      "grad_norm": 4.382105350494385,
      "learning_rate": 0.00017343013388647935,
      "loss": 1.2964,
      "step": 93000
    },
    {
      "epoch": 2.7781924581255786,
      "grad_norm": 4.167068958282471,
      "learning_rate": 0.00017341584809056054,
      "loss": 1.3275,
      "step": 93050
    },
    {
      "epoch": 2.779685307377661,
      "grad_norm": 2.9642536640167236,
      "learning_rate": 0.00017340156229464168,
      "loss": 1.3143,
      "step": 93100
    },
    {
      "epoch": 2.7811781566297435,
      "grad_norm": 3.9589407444000244,
      "learning_rate": 0.00017338727649872287,
      "loss": 1.2388,
      "step": 93150
    },
    {
      "epoch": 2.7826710058818263,
      "grad_norm": 4.6660051345825195,
      "learning_rate": 0.000173372990702804,
      "loss": 1.3026,
      "step": 93200
    },
    {
      "epoch": 2.7841638551339085,
      "grad_norm": 4.213698387145996,
      "learning_rate": 0.0001733587049068852,
      "loss": 1.3192,
      "step": 93250
    },
    {
      "epoch": 2.7856567043859912,
      "grad_norm": 4.1265645027160645,
      "learning_rate": 0.00017334441911096636,
      "loss": 1.2899,
      "step": 93300
    },
    {
      "epoch": 2.7871495536380735,
      "grad_norm": 5.5079545974731445,
      "learning_rate": 0.00017333013331504753,
      "loss": 1.2968,
      "step": 93350
    },
    {
      "epoch": 2.788642402890156,
      "grad_norm": 4.825951099395752,
      "learning_rate": 0.0001733158475191287,
      "loss": 1.2432,
      "step": 93400
    },
    {
      "epoch": 2.790135252142239,
      "grad_norm": 3.7230324745178223,
      "learning_rate": 0.00017330156172320986,
      "loss": 1.385,
      "step": 93450
    },
    {
      "epoch": 2.791628101394321,
      "grad_norm": 4.764435291290283,
      "learning_rate": 0.00017328727592729102,
      "loss": 1.2362,
      "step": 93500
    },
    {
      "epoch": 2.793120950646404,
      "grad_norm": 5.750154972076416,
      "learning_rate": 0.00017327299013137219,
      "loss": 1.2868,
      "step": 93550
    },
    {
      "epoch": 2.794613799898486,
      "grad_norm": 3.704388380050659,
      "learning_rate": 0.00017325870433545335,
      "loss": 1.2839,
      "step": 93600
    },
    {
      "epoch": 2.796106649150569,
      "grad_norm": 5.4798407554626465,
      "learning_rate": 0.00017324441853953452,
      "loss": 1.3227,
      "step": 93650
    },
    {
      "epoch": 2.7975994984026515,
      "grad_norm": 4.121763229370117,
      "learning_rate": 0.00017323013274361568,
      "loss": 1.2249,
      "step": 93700
    },
    {
      "epoch": 2.799092347654734,
      "grad_norm": 4.939391136169434,
      "learning_rate": 0.00017321584694769684,
      "loss": 1.3017,
      "step": 93750
    },
    {
      "epoch": 2.8005851969068165,
      "grad_norm": 4.476954936981201,
      "learning_rate": 0.000173201561151778,
      "loss": 1.2851,
      "step": 93800
    },
    {
      "epoch": 2.8020780461588988,
      "grad_norm": 4.639068603515625,
      "learning_rate": 0.0001731872753558592,
      "loss": 1.3026,
      "step": 93850
    },
    {
      "epoch": 2.8035708954109815,
      "grad_norm": 4.282186031341553,
      "learning_rate": 0.00017317298955994034,
      "loss": 1.2374,
      "step": 93900
    },
    {
      "epoch": 2.805063744663064,
      "grad_norm": 5.077556610107422,
      "learning_rate": 0.00017315870376402153,
      "loss": 1.3172,
      "step": 93950
    },
    {
      "epoch": 2.8065565939151464,
      "grad_norm": 5.154215335845947,
      "learning_rate": 0.00017314441796810267,
      "loss": 1.2998,
      "step": 94000
    },
    {
      "epoch": 2.8080494431672287,
      "grad_norm": 6.019194602966309,
      "learning_rate": 0.00017313013217218386,
      "loss": 1.3121,
      "step": 94050
    },
    {
      "epoch": 2.8095422924193114,
      "grad_norm": 4.445097923278809,
      "learning_rate": 0.00017311584637626502,
      "loss": 1.2887,
      "step": 94100
    },
    {
      "epoch": 2.811035141671394,
      "grad_norm": 3.711930990219116,
      "learning_rate": 0.0001731015605803462,
      "loss": 1.3031,
      "step": 94150
    },
    {
      "epoch": 2.8125279909234764,
      "grad_norm": 7.101813316345215,
      "learning_rate": 0.00017308727478442735,
      "loss": 1.2877,
      "step": 94200
    },
    {
      "epoch": 2.814020840175559,
      "grad_norm": 5.068753719329834,
      "learning_rate": 0.00017307298898850852,
      "loss": 1.2459,
      "step": 94250
    },
    {
      "epoch": 2.8155136894276414,
      "grad_norm": 5.185253143310547,
      "learning_rate": 0.00017305870319258968,
      "loss": 1.2292,
      "step": 94300
    },
    {
      "epoch": 2.817006538679724,
      "grad_norm": 4.654400825500488,
      "learning_rate": 0.00017304441739667085,
      "loss": 1.325,
      "step": 94350
    },
    {
      "epoch": 2.8184993879318068,
      "grad_norm": 4.575232028961182,
      "learning_rate": 0.000173030131600752,
      "loss": 1.3077,
      "step": 94400
    },
    {
      "epoch": 2.819992237183889,
      "grad_norm": 5.062366485595703,
      "learning_rate": 0.00017301584580483318,
      "loss": 1.319,
      "step": 94450
    },
    {
      "epoch": 2.8214850864359717,
      "grad_norm": 3.9217934608459473,
      "learning_rate": 0.00017300156000891434,
      "loss": 1.3042,
      "step": 94500
    },
    {
      "epoch": 2.822977935688054,
      "grad_norm": 7.488821506500244,
      "learning_rate": 0.00017298727421299553,
      "loss": 1.302,
      "step": 94550
    },
    {
      "epoch": 2.8244707849401367,
      "grad_norm": 4.662975311279297,
      "learning_rate": 0.00017297298841707667,
      "loss": 1.2965,
      "step": 94600
    },
    {
      "epoch": 2.8259636341922194,
      "grad_norm": 4.872460842132568,
      "learning_rate": 0.00017295870262115786,
      "loss": 1.2729,
      "step": 94650
    },
    {
      "epoch": 2.8274564834443017,
      "grad_norm": 4.52310848236084,
      "learning_rate": 0.000172944416825239,
      "loss": 1.2872,
      "step": 94700
    },
    {
      "epoch": 2.8289493326963844,
      "grad_norm": 4.012574672698975,
      "learning_rate": 0.0001729301310293202,
      "loss": 1.2537,
      "step": 94750
    },
    {
      "epoch": 2.8304421819484666,
      "grad_norm": 5.984830379486084,
      "learning_rate": 0.00017291584523340133,
      "loss": 1.3413,
      "step": 94800
    },
    {
      "epoch": 2.8319350312005493,
      "grad_norm": 4.736281394958496,
      "learning_rate": 0.00017290155943748252,
      "loss": 1.3116,
      "step": 94850
    },
    {
      "epoch": 2.833427880452632,
      "grad_norm": 4.227638244628906,
      "learning_rate": 0.00017288727364156369,
      "loss": 1.2607,
      "step": 94900
    },
    {
      "epoch": 2.8349207297047143,
      "grad_norm": 4.373624324798584,
      "learning_rate": 0.00017287298784564485,
      "loss": 1.3173,
      "step": 94950
    },
    {
      "epoch": 2.836413578956797,
      "grad_norm": 3.9513514041900635,
      "learning_rate": 0.00017285870204972602,
      "loss": 1.2344,
      "step": 95000
    },
    {
      "epoch": 2.8379064282088793,
      "grad_norm": 4.8411664962768555,
      "learning_rate": 0.00017284441625380715,
      "loss": 1.2618,
      "step": 95050
    },
    {
      "epoch": 2.839399277460962,
      "grad_norm": 6.098468780517578,
      "learning_rate": 0.00017283013045788834,
      "loss": 1.2512,
      "step": 95100
    },
    {
      "epoch": 2.8408921267130447,
      "grad_norm": 5.991581439971924,
      "learning_rate": 0.00017281584466196948,
      "loss": 1.306,
      "step": 95150
    },
    {
      "epoch": 2.842384975965127,
      "grad_norm": 4.810219764709473,
      "learning_rate": 0.00017280155886605067,
      "loss": 1.3146,
      "step": 95200
    },
    {
      "epoch": 2.8438778252172097,
      "grad_norm": 4.411154747009277,
      "learning_rate": 0.00017278727307013184,
      "loss": 1.3104,
      "step": 95250
    },
    {
      "epoch": 2.845370674469292,
      "grad_norm": 5.2504472732543945,
      "learning_rate": 0.000172772987274213,
      "loss": 1.3055,
      "step": 95300
    },
    {
      "epoch": 2.8468635237213746,
      "grad_norm": 3.6604950428009033,
      "learning_rate": 0.00017275870147829417,
      "loss": 1.2861,
      "step": 95350
    },
    {
      "epoch": 2.8483563729734573,
      "grad_norm": 4.993444442749023,
      "learning_rate": 0.00017274441568237533,
      "loss": 1.3199,
      "step": 95400
    },
    {
      "epoch": 2.8498492222255396,
      "grad_norm": 4.28807258605957,
      "learning_rate": 0.0001727301298864565,
      "loss": 1.2676,
      "step": 95450
    },
    {
      "epoch": 2.8513420714776223,
      "grad_norm": 4.000975608825684,
      "learning_rate": 0.00017271584409053766,
      "loss": 1.2916,
      "step": 95500
    },
    {
      "epoch": 2.8528349207297046,
      "grad_norm": 3.3593053817749023,
      "learning_rate": 0.00017270155829461883,
      "loss": 1.2802,
      "step": 95550
    },
    {
      "epoch": 2.8543277699817873,
      "grad_norm": 4.179107666015625,
      "learning_rate": 0.0001726872724987,
      "loss": 1.3147,
      "step": 95600
    },
    {
      "epoch": 2.85582061923387,
      "grad_norm": 3.5861895084381104,
      "learning_rate": 0.00017267298670278116,
      "loss": 1.2653,
      "step": 95650
    },
    {
      "epoch": 2.8573134684859522,
      "grad_norm": 4.41680383682251,
      "learning_rate": 0.00017265870090686235,
      "loss": 1.2961,
      "step": 95700
    },
    {
      "epoch": 2.858806317738035,
      "grad_norm": 6.727049827575684,
      "learning_rate": 0.00017264441511094349,
      "loss": 1.2936,
      "step": 95750
    },
    {
      "epoch": 2.860299166990117,
      "grad_norm": 4.895215034484863,
      "learning_rate": 0.00017263012931502468,
      "loss": 1.2528,
      "step": 95800
    },
    {
      "epoch": 2.8617920162422,
      "grad_norm": 5.4271955490112305,
      "learning_rate": 0.00017261584351910581,
      "loss": 1.3365,
      "step": 95850
    },
    {
      "epoch": 2.8632848654942826,
      "grad_norm": 4.273880481719971,
      "learning_rate": 0.000172601557723187,
      "loss": 1.2779,
      "step": 95900
    },
    {
      "epoch": 2.864777714746365,
      "grad_norm": 4.767480850219727,
      "learning_rate": 0.00017258727192726814,
      "loss": 1.2716,
      "step": 95950
    },
    {
      "epoch": 2.8662705639984476,
      "grad_norm": 4.384129524230957,
      "learning_rate": 0.00017257298613134934,
      "loss": 1.2969,
      "step": 96000
    },
    {
      "epoch": 2.86776341325053,
      "grad_norm": 4.527589797973633,
      "learning_rate": 0.0001725587003354305,
      "loss": 1.266,
      "step": 96050
    },
    {
      "epoch": 2.8692562625026126,
      "grad_norm": 4.149820804595947,
      "learning_rate": 0.00017254441453951167,
      "loss": 1.31,
      "step": 96100
    },
    {
      "epoch": 2.8707491117546953,
      "grad_norm": 4.574423313140869,
      "learning_rate": 0.00017253012874359283,
      "loss": 1.3,
      "step": 96150
    },
    {
      "epoch": 2.8722419610067775,
      "grad_norm": 4.80661153793335,
      "learning_rate": 0.000172515842947674,
      "loss": 1.304,
      "step": 96200
    },
    {
      "epoch": 2.87373481025886,
      "grad_norm": 4.065107822418213,
      "learning_rate": 0.00017250155715175516,
      "loss": 1.2581,
      "step": 96250
    },
    {
      "epoch": 2.8752276595109425,
      "grad_norm": 4.727302551269531,
      "learning_rate": 0.00017248727135583632,
      "loss": 1.321,
      "step": 96300
    },
    {
      "epoch": 2.876720508763025,
      "grad_norm": 4.85087251663208,
      "learning_rate": 0.0001724729855599175,
      "loss": 1.3013,
      "step": 96350
    },
    {
      "epoch": 2.8782133580151075,
      "grad_norm": 5.125801086425781,
      "learning_rate": 0.00017245869976399865,
      "loss": 1.2956,
      "step": 96400
    },
    {
      "epoch": 2.87970620726719,
      "grad_norm": 4.144227504730225,
      "learning_rate": 0.00017244441396807982,
      "loss": 1.2929,
      "step": 96450
    },
    {
      "epoch": 2.8811990565192724,
      "grad_norm": 4.702892303466797,
      "learning_rate": 0.000172430128172161,
      "loss": 1.361,
      "step": 96500
    },
    {
      "epoch": 2.882691905771355,
      "grad_norm": 5.188230991363525,
      "learning_rate": 0.00017241584237624215,
      "loss": 1.2474,
      "step": 96550
    },
    {
      "epoch": 2.884184755023438,
      "grad_norm": 5.263503551483154,
      "learning_rate": 0.00017240155658032334,
      "loss": 1.2728,
      "step": 96600
    },
    {
      "epoch": 2.88567760427552,
      "grad_norm": 3.8955039978027344,
      "learning_rate": 0.00017238727078440448,
      "loss": 1.2638,
      "step": 96650
    },
    {
      "epoch": 2.887170453527603,
      "grad_norm": 4.1187238693237305,
      "learning_rate": 0.00017237298498848567,
      "loss": 1.2934,
      "step": 96700
    },
    {
      "epoch": 2.888663302779685,
      "grad_norm": 6.000136375427246,
      "learning_rate": 0.00017235869919256683,
      "loss": 1.293,
      "step": 96750
    },
    {
      "epoch": 2.890156152031768,
      "grad_norm": 5.494427680969238,
      "learning_rate": 0.000172344413396648,
      "loss": 1.331,
      "step": 96800
    },
    {
      "epoch": 2.8916490012838505,
      "grad_norm": 5.483157157897949,
      "learning_rate": 0.00017233012760072916,
      "loss": 1.3299,
      "step": 96850
    },
    {
      "epoch": 2.8931418505359328,
      "grad_norm": 4.264294147491455,
      "learning_rate": 0.00017231584180481033,
      "loss": 1.2966,
      "step": 96900
    },
    {
      "epoch": 2.8946346997880155,
      "grad_norm": 3.7987403869628906,
      "learning_rate": 0.0001723015560088915,
      "loss": 1.3028,
      "step": 96950
    },
    {
      "epoch": 2.8961275490400977,
      "grad_norm": 4.712928771972656,
      "learning_rate": 0.00017228727021297266,
      "loss": 1.2154,
      "step": 97000
    },
    {
      "epoch": 2.8976203982921804,
      "grad_norm": 5.0749640464782715,
      "learning_rate": 0.00017227298441705382,
      "loss": 1.269,
      "step": 97050
    },
    {
      "epoch": 2.899113247544263,
      "grad_norm": 4.288541316986084,
      "learning_rate": 0.00017225869862113499,
      "loss": 1.2575,
      "step": 97100
    },
    {
      "epoch": 2.9006060967963454,
      "grad_norm": 4.452602386474609,
      "learning_rate": 0.00017224441282521615,
      "loss": 1.2613,
      "step": 97150
    },
    {
      "epoch": 2.902098946048428,
      "grad_norm": 4.485718250274658,
      "learning_rate": 0.00017223012702929731,
      "loss": 1.2678,
      "step": 97200
    },
    {
      "epoch": 2.9035917953005104,
      "grad_norm": 3.692293643951416,
      "learning_rate": 0.00017221584123337848,
      "loss": 1.2929,
      "step": 97250
    },
    {
      "epoch": 2.905084644552593,
      "grad_norm": 4.699619770050049,
      "learning_rate": 0.00017220155543745967,
      "loss": 1.2601,
      "step": 97300
    },
    {
      "epoch": 2.906577493804676,
      "grad_norm": 4.477214336395264,
      "learning_rate": 0.0001721872696415408,
      "loss": 1.3272,
      "step": 97350
    },
    {
      "epoch": 2.908070343056758,
      "grad_norm": 4.274310111999512,
      "learning_rate": 0.000172172983845622,
      "loss": 1.3221,
      "step": 97400
    },
    {
      "epoch": 2.9095631923088408,
      "grad_norm": 4.423600196838379,
      "learning_rate": 0.00017215869804970314,
      "loss": 1.3161,
      "step": 97450
    },
    {
      "epoch": 2.911056041560923,
      "grad_norm": 4.693662166595459,
      "learning_rate": 0.00017214441225378433,
      "loss": 1.3374,
      "step": 97500
    },
    {
      "epoch": 2.9125488908130057,
      "grad_norm": 5.746738910675049,
      "learning_rate": 0.0001721301264578655,
      "loss": 1.3171,
      "step": 97550
    },
    {
      "epoch": 2.9140417400650884,
      "grad_norm": 5.049290657043457,
      "learning_rate": 0.00017211584066194666,
      "loss": 1.2713,
      "step": 97600
    },
    {
      "epoch": 2.9155345893171707,
      "grad_norm": 4.0549468994140625,
      "learning_rate": 0.00017210155486602782,
      "loss": 1.3214,
      "step": 97650
    },
    {
      "epoch": 2.9170274385692534,
      "grad_norm": 5.128767013549805,
      "learning_rate": 0.00017208726907010896,
      "loss": 1.2717,
      "step": 97700
    },
    {
      "epoch": 2.9185202878213357,
      "grad_norm": 5.148744106292725,
      "learning_rate": 0.00017207298327419015,
      "loss": 1.2602,
      "step": 97750
    },
    {
      "epoch": 2.9200131370734184,
      "grad_norm": 3.682936429977417,
      "learning_rate": 0.0001720586974782713,
      "loss": 1.2941,
      "step": 97800
    },
    {
      "epoch": 2.921505986325501,
      "grad_norm": 4.665682315826416,
      "learning_rate": 0.00017204441168235248,
      "loss": 1.2959,
      "step": 97850
    },
    {
      "epoch": 2.9229988355775833,
      "grad_norm": 4.523177146911621,
      "learning_rate": 0.00017203012588643365,
      "loss": 1.2755,
      "step": 97900
    },
    {
      "epoch": 2.924491684829666,
      "grad_norm": 4.719842433929443,
      "learning_rate": 0.0001720158400905148,
      "loss": 1.2523,
      "step": 97950
    },
    {
      "epoch": 2.9259845340817483,
      "grad_norm": 3.7076125144958496,
      "learning_rate": 0.00017200155429459598,
      "loss": 1.2909,
      "step": 98000
    },
    {
      "epoch": 2.927477383333831,
      "grad_norm": 8.15734577178955,
      "learning_rate": 0.00017198726849867714,
      "loss": 1.2369,
      "step": 98050
    },
    {
      "epoch": 2.9289702325859137,
      "grad_norm": 4.653754711151123,
      "learning_rate": 0.0001719729827027583,
      "loss": 1.3744,
      "step": 98100
    },
    {
      "epoch": 2.930463081837996,
      "grad_norm": 4.599735736846924,
      "learning_rate": 0.00017195869690683947,
      "loss": 1.333,
      "step": 98150
    },
    {
      "epoch": 2.9319559310900787,
      "grad_norm": 5.794803619384766,
      "learning_rate": 0.00017194441111092063,
      "loss": 1.3269,
      "step": 98200
    },
    {
      "epoch": 2.933448780342161,
      "grad_norm": 4.360086441040039,
      "learning_rate": 0.0001719301253150018,
      "loss": 1.3025,
      "step": 98250
    },
    {
      "epoch": 2.9349416295942437,
      "grad_norm": 4.956981182098389,
      "learning_rate": 0.00017191583951908296,
      "loss": 1.2484,
      "step": 98300
    },
    {
      "epoch": 2.9364344788463264,
      "grad_norm": 3.210866928100586,
      "learning_rate": 0.00017190155372316416,
      "loss": 1.2796,
      "step": 98350
    },
    {
      "epoch": 2.9379273280984086,
      "grad_norm": 5.565187931060791,
      "learning_rate": 0.0001718872679272453,
      "loss": 1.3243,
      "step": 98400
    },
    {
      "epoch": 2.939420177350491,
      "grad_norm": 4.15312385559082,
      "learning_rate": 0.00017187298213132649,
      "loss": 1.2922,
      "step": 98450
    },
    {
      "epoch": 2.9409130266025736,
      "grad_norm": 4.836455345153809,
      "learning_rate": 0.00017185869633540762,
      "loss": 1.2495,
      "step": 98500
    },
    {
      "epoch": 2.9424058758546563,
      "grad_norm": 5.803861618041992,
      "learning_rate": 0.00017184441053948881,
      "loss": 1.3052,
      "step": 98550
    },
    {
      "epoch": 2.9438987251067386,
      "grad_norm": 5.279322147369385,
      "learning_rate": 0.00017183012474356995,
      "loss": 1.299,
      "step": 98600
    },
    {
      "epoch": 2.9453915743588213,
      "grad_norm": 3.5537776947021484,
      "learning_rate": 0.00017181583894765114,
      "loss": 1.3192,
      "step": 98650
    },
    {
      "epoch": 2.9468844236109035,
      "grad_norm": 4.724547386169434,
      "learning_rate": 0.0001718015531517323,
      "loss": 1.2877,
      "step": 98700
    },
    {
      "epoch": 2.9483772728629862,
      "grad_norm": 3.863722801208496,
      "learning_rate": 0.00017178726735581347,
      "loss": 1.2713,
      "step": 98750
    },
    {
      "epoch": 2.949870122115069,
      "grad_norm": 3.8161838054656982,
      "learning_rate": 0.00017177298155989464,
      "loss": 1.2484,
      "step": 98800
    },
    {
      "epoch": 2.951362971367151,
      "grad_norm": 4.72680139541626,
      "learning_rate": 0.0001717586957639758,
      "loss": 1.3184,
      "step": 98850
    },
    {
      "epoch": 2.952855820619234,
      "grad_norm": 4.270802021026611,
      "learning_rate": 0.00017174440996805697,
      "loss": 1.2655,
      "step": 98900
    },
    {
      "epoch": 2.954348669871316,
      "grad_norm": 4.669827461242676,
      "learning_rate": 0.00017173012417213813,
      "loss": 1.3366,
      "step": 98950
    },
    {
      "epoch": 2.955841519123399,
      "grad_norm": 4.791524887084961,
      "learning_rate": 0.0001717158383762193,
      "loss": 1.2887,
      "step": 99000
    },
    {
      "epoch": 2.9573343683754816,
      "grad_norm": 5.6565423011779785,
      "learning_rate": 0.00017170155258030046,
      "loss": 1.2693,
      "step": 99050
    },
    {
      "epoch": 2.958827217627564,
      "grad_norm": 3.3838722705841064,
      "learning_rate": 0.00017168726678438163,
      "loss": 1.3281,
      "step": 99100
    },
    {
      "epoch": 2.9603200668796466,
      "grad_norm": 5.198723793029785,
      "learning_rate": 0.00017167298098846282,
      "loss": 1.2931,
      "step": 99150
    },
    {
      "epoch": 2.961812916131729,
      "grad_norm": 3.6596312522888184,
      "learning_rate": 0.00017165869519254395,
      "loss": 1.235,
      "step": 99200
    },
    {
      "epoch": 2.9633057653838115,
      "grad_norm": 4.816101551055908,
      "learning_rate": 0.00017164440939662515,
      "loss": 1.3069,
      "step": 99250
    },
    {
      "epoch": 2.9647986146358942,
      "grad_norm": 4.188026428222656,
      "learning_rate": 0.00017163012360070628,
      "loss": 1.3104,
      "step": 99300
    },
    {
      "epoch": 2.9662914638879765,
      "grad_norm": 4.76021671295166,
      "learning_rate": 0.00017161583780478748,
      "loss": 1.3125,
      "step": 99350
    },
    {
      "epoch": 2.967784313140059,
      "grad_norm": 4.8385138511657715,
      "learning_rate": 0.00017160155200886861,
      "loss": 1.2871,
      "step": 99400
    },
    {
      "epoch": 2.9692771623921415,
      "grad_norm": 10.232873916625977,
      "learning_rate": 0.0001715872662129498,
      "loss": 1.2901,
      "step": 99450
    },
    {
      "epoch": 2.970770011644224,
      "grad_norm": 4.217905521392822,
      "learning_rate": 0.00017157298041703097,
      "loss": 1.3533,
      "step": 99500
    },
    {
      "epoch": 2.972262860896307,
      "grad_norm": 4.916757583618164,
      "learning_rate": 0.00017155869462111213,
      "loss": 1.3062,
      "step": 99550
    },
    {
      "epoch": 2.973755710148389,
      "grad_norm": 4.588007926940918,
      "learning_rate": 0.0001715444088251933,
      "loss": 1.2755,
      "step": 99600
    },
    {
      "epoch": 2.975248559400472,
      "grad_norm": 7.835936546325684,
      "learning_rate": 0.00017153012302927446,
      "loss": 1.2472,
      "step": 99650
    },
    {
      "epoch": 2.976741408652554,
      "grad_norm": 3.825573444366455,
      "learning_rate": 0.00017151583723335563,
      "loss": 1.252,
      "step": 99700
    },
    {
      "epoch": 2.978234257904637,
      "grad_norm": 4.329678535461426,
      "learning_rate": 0.0001715015514374368,
      "loss": 1.2689,
      "step": 99750
    },
    {
      "epoch": 2.9797271071567195,
      "grad_norm": 5.514796733856201,
      "learning_rate": 0.00017148726564151796,
      "loss": 1.2776,
      "step": 99800
    },
    {
      "epoch": 2.9812199564088018,
      "grad_norm": 5.769738674163818,
      "learning_rate": 0.00017147297984559912,
      "loss": 1.2717,
      "step": 99850
    },
    {
      "epoch": 2.9827128056608845,
      "grad_norm": 5.01526403427124,
      "learning_rate": 0.0001714586940496803,
      "loss": 1.2571,
      "step": 99900
    },
    {
      "epoch": 2.9842056549129667,
      "grad_norm": 4.526731491088867,
      "learning_rate": 0.00017144440825376148,
      "loss": 1.3401,
      "step": 99950
    },
    {
      "epoch": 2.9856985041650494,
      "grad_norm": 3.969651222229004,
      "learning_rate": 0.00017143012245784262,
      "loss": 1.2814,
      "step": 100000
    },
    {
      "epoch": 2.987191353417132,
      "grad_norm": 4.796939849853516,
      "learning_rate": 0.0001714158366619238,
      "loss": 1.286,
      "step": 100050
    },
    {
      "epoch": 2.9886842026692144,
      "grad_norm": 5.400031089782715,
      "learning_rate": 0.00017140155086600495,
      "loss": 1.2939,
      "step": 100100
    },
    {
      "epoch": 2.990177051921297,
      "grad_norm": 4.253983497619629,
      "learning_rate": 0.00017138726507008614,
      "loss": 1.2753,
      "step": 100150
    },
    {
      "epoch": 2.9916699011733794,
      "grad_norm": 4.808609485626221,
      "learning_rate": 0.0001713729792741673,
      "loss": 1.3019,
      "step": 100200
    },
    {
      "epoch": 2.993162750425462,
      "grad_norm": 4.482404708862305,
      "learning_rate": 0.00017135869347824844,
      "loss": 1.2642,
      "step": 100250
    },
    {
      "epoch": 2.994655599677545,
      "grad_norm": 5.035706043243408,
      "learning_rate": 0.00017134440768232963,
      "loss": 1.2927,
      "step": 100300
    },
    {
      "epoch": 2.996148448929627,
      "grad_norm": 4.565020561218262,
      "learning_rate": 0.00017133012188641077,
      "loss": 1.3068,
      "step": 100350
    },
    {
      "epoch": 2.9976412981817098,
      "grad_norm": 5.071930408477783,
      "learning_rate": 0.00017131583609049196,
      "loss": 1.3127,
      "step": 100400
    },
    {
      "epoch": 2.999134147433792,
      "grad_norm": 5.127132415771484,
      "learning_rate": 0.0001713015502945731,
      "loss": 1.3057,
      "step": 100450
    },
    {
      "epoch": 3.0006269966858747,
      "grad_norm": 5.550358772277832,
      "learning_rate": 0.0001712872644986543,
      "loss": 1.2978,
      "step": 100500
    },
    {
      "epoch": 3.002119845937957,
      "grad_norm": 5.580805778503418,
      "learning_rate": 0.00017127297870273545,
      "loss": 1.2057,
      "step": 100550
    },
    {
      "epoch": 3.0036126951900397,
      "grad_norm": 3.326228618621826,
      "learning_rate": 0.00017125869290681662,
      "loss": 1.1558,
      "step": 100600
    },
    {
      "epoch": 3.0051055444421224,
      "grad_norm": 4.807900905609131,
      "learning_rate": 0.00017124440711089778,
      "loss": 1.1667,
      "step": 100650
    },
    {
      "epoch": 3.0065983936942047,
      "grad_norm": 4.419902801513672,
      "learning_rate": 0.00017123012131497895,
      "loss": 1.1765,
      "step": 100700
    },
    {
      "epoch": 3.0080912429462874,
      "grad_norm": 4.567293643951416,
      "learning_rate": 0.0001712158355190601,
      "loss": 1.2552,
      "step": 100750
    },
    {
      "epoch": 3.0095840921983696,
      "grad_norm": 3.129995346069336,
      "learning_rate": 0.00017120154972314128,
      "loss": 1.2222,
      "step": 100800
    },
    {
      "epoch": 3.0110769414504523,
      "grad_norm": 4.515715599060059,
      "learning_rate": 0.00017118726392722244,
      "loss": 1.2794,
      "step": 100850
    },
    {
      "epoch": 3.012569790702535,
      "grad_norm": 4.2073283195495605,
      "learning_rate": 0.0001711729781313036,
      "loss": 1.2108,
      "step": 100900
    },
    {
      "epoch": 3.0140626399546173,
      "grad_norm": 4.649296760559082,
      "learning_rate": 0.00017115869233538477,
      "loss": 1.2171,
      "step": 100950
    },
    {
      "epoch": 3.0155554892067,
      "grad_norm": 5.268732070922852,
      "learning_rate": 0.00017114440653946596,
      "loss": 1.2566,
      "step": 101000
    },
    {
      "epoch": 3.0170483384587823,
      "grad_norm": 4.506618976593018,
      "learning_rate": 0.0001711301207435471,
      "loss": 1.1528,
      "step": 101050
    },
    {
      "epoch": 3.018541187710865,
      "grad_norm": 5.6522016525268555,
      "learning_rate": 0.0001711158349476283,
      "loss": 1.2399,
      "step": 101100
    },
    {
      "epoch": 3.0200340369629477,
      "grad_norm": 4.122422695159912,
      "learning_rate": 0.00017110154915170943,
      "loss": 1.2239,
      "step": 101150
    },
    {
      "epoch": 3.02152688621503,
      "grad_norm": 3.759425401687622,
      "learning_rate": 0.00017108726335579062,
      "loss": 1.2463,
      "step": 101200
    },
    {
      "epoch": 3.0230197354671127,
      "grad_norm": 3.6985554695129395,
      "learning_rate": 0.00017107297755987176,
      "loss": 1.1758,
      "step": 101250
    },
    {
      "epoch": 3.024512584719195,
      "grad_norm": 4.267302513122559,
      "learning_rate": 0.00017105869176395295,
      "loss": 1.1626,
      "step": 101300
    },
    {
      "epoch": 3.0260054339712776,
      "grad_norm": 4.210064888000488,
      "learning_rate": 0.00017104440596803412,
      "loss": 1.1989,
      "step": 101350
    },
    {
      "epoch": 3.02749828322336,
      "grad_norm": 4.68282413482666,
      "learning_rate": 0.00017103012017211528,
      "loss": 1.1906,
      "step": 101400
    },
    {
      "epoch": 3.0289911324754426,
      "grad_norm": 4.420328617095947,
      "learning_rate": 0.00017101583437619645,
      "loss": 1.2236,
      "step": 101450
    },
    {
      "epoch": 3.0304839817275253,
      "grad_norm": 4.2961931228637695,
      "learning_rate": 0.0001710015485802776,
      "loss": 1.159,
      "step": 101500
    },
    {
      "epoch": 3.0319768309796076,
      "grad_norm": 5.494664669036865,
      "learning_rate": 0.00017098726278435878,
      "loss": 1.2286,
      "step": 101550
    },
    {
      "epoch": 3.0334696802316903,
      "grad_norm": 4.834953308105469,
      "learning_rate": 0.00017097297698843994,
      "loss": 1.1724,
      "step": 101600
    },
    {
      "epoch": 3.0349625294837725,
      "grad_norm": 4.731471538543701,
      "learning_rate": 0.0001709586911925211,
      "loss": 1.2089,
      "step": 101650
    },
    {
      "epoch": 3.0364553787358552,
      "grad_norm": 3.613417863845825,
      "learning_rate": 0.00017094440539660227,
      "loss": 1.2864,
      "step": 101700
    },
    {
      "epoch": 3.037948227987938,
      "grad_norm": 4.939467430114746,
      "learning_rate": 0.00017093011960068343,
      "loss": 1.2509,
      "step": 101750
    },
    {
      "epoch": 3.03944107724002,
      "grad_norm": 4.377845287322998,
      "learning_rate": 0.00017091583380476463,
      "loss": 1.2377,
      "step": 101800
    },
    {
      "epoch": 3.040933926492103,
      "grad_norm": 5.249380111694336,
      "learning_rate": 0.00017090154800884576,
      "loss": 1.1936,
      "step": 101850
    },
    {
      "epoch": 3.042426775744185,
      "grad_norm": 5.340421676635742,
      "learning_rate": 0.00017088726221292695,
      "loss": 1.2575,
      "step": 101900
    },
    {
      "epoch": 3.043919624996268,
      "grad_norm": 4.105010986328125,
      "learning_rate": 0.0001708729764170081,
      "loss": 1.1913,
      "step": 101950
    },
    {
      "epoch": 3.0454124742483506,
      "grad_norm": 4.164525032043457,
      "learning_rate": 0.00017085869062108928,
      "loss": 1.2782,
      "step": 102000
    },
    {
      "epoch": 3.046905323500433,
      "grad_norm": 5.213939666748047,
      "learning_rate": 0.00017084440482517042,
      "loss": 1.2888,
      "step": 102050
    },
    {
      "epoch": 3.0483981727525156,
      "grad_norm": 6.272274494171143,
      "learning_rate": 0.0001708301190292516,
      "loss": 1.1914,
      "step": 102100
    },
    {
      "epoch": 3.049891022004598,
      "grad_norm": 4.293304443359375,
      "learning_rate": 0.00017081583323333278,
      "loss": 1.2037,
      "step": 102150
    },
    {
      "epoch": 3.0513838712566805,
      "grad_norm": 4.839235305786133,
      "learning_rate": 0.00017080154743741394,
      "loss": 1.2241,
      "step": 102200
    },
    {
      "epoch": 3.0528767205087632,
      "grad_norm": 5.0241498947143555,
      "learning_rate": 0.0001707872616414951,
      "loss": 1.2341,
      "step": 102250
    },
    {
      "epoch": 3.0543695697608455,
      "grad_norm": 6.890449047088623,
      "learning_rate": 0.00017077297584557627,
      "loss": 1.2411,
      "step": 102300
    },
    {
      "epoch": 3.055862419012928,
      "grad_norm": 5.459249496459961,
      "learning_rate": 0.00017075869004965744,
      "loss": 1.2553,
      "step": 102350
    },
    {
      "epoch": 3.0573552682650105,
      "grad_norm": 4.772298336029053,
      "learning_rate": 0.0001707444042537386,
      "loss": 1.1836,
      "step": 102400
    },
    {
      "epoch": 3.058848117517093,
      "grad_norm": 4.730257511138916,
      "learning_rate": 0.00017073011845781977,
      "loss": 1.2091,
      "step": 102450
    },
    {
      "epoch": 3.0603409667691754,
      "grad_norm": 4.203242301940918,
      "learning_rate": 0.00017071583266190093,
      "loss": 1.2106,
      "step": 102500
    },
    {
      "epoch": 3.061833816021258,
      "grad_norm": 5.499764442443848,
      "learning_rate": 0.0001707015468659821,
      "loss": 1.2492,
      "step": 102550
    },
    {
      "epoch": 3.063326665273341,
      "grad_norm": 4.593760967254639,
      "learning_rate": 0.0001706872610700633,
      "loss": 1.2849,
      "step": 102600
    },
    {
      "epoch": 3.064819514525423,
      "grad_norm": 4.561386585235596,
      "learning_rate": 0.00017067297527414442,
      "loss": 1.2404,
      "step": 102650
    },
    {
      "epoch": 3.066312363777506,
      "grad_norm": 5.402337551116943,
      "learning_rate": 0.00017065868947822562,
      "loss": 1.1952,
      "step": 102700
    },
    {
      "epoch": 3.067805213029588,
      "grad_norm": 5.212750434875488,
      "learning_rate": 0.00017064440368230675,
      "loss": 1.2505,
      "step": 102750
    },
    {
      "epoch": 3.069298062281671,
      "grad_norm": 5.780417442321777,
      "learning_rate": 0.00017063011788638795,
      "loss": 1.2653,
      "step": 102800
    },
    {
      "epoch": 3.0707909115337535,
      "grad_norm": 5.937560081481934,
      "learning_rate": 0.00017061583209046908,
      "loss": 1.2068,
      "step": 102850
    },
    {
      "epoch": 3.0722837607858358,
      "grad_norm": 4.921445369720459,
      "learning_rate": 0.00017060154629455025,
      "loss": 1.1634,
      "step": 102900
    },
    {
      "epoch": 3.0737766100379185,
      "grad_norm": 5.346931457519531,
      "learning_rate": 0.00017058726049863144,
      "loss": 1.2136,
      "step": 102950
    },
    {
      "epoch": 3.0752694592900007,
      "grad_norm": 4.446613788604736,
      "learning_rate": 0.00017057297470271258,
      "loss": 1.1925,
      "step": 103000
    },
    {
      "epoch": 3.0767623085420834,
      "grad_norm": 4.870328903198242,
      "learning_rate": 0.00017055868890679377,
      "loss": 1.1855,
      "step": 103050
    },
    {
      "epoch": 3.078255157794166,
      "grad_norm": 5.902019500732422,
      "learning_rate": 0.0001705444031108749,
      "loss": 1.2081,
      "step": 103100
    },
    {
      "epoch": 3.0797480070462484,
      "grad_norm": 4.221609592437744,
      "learning_rate": 0.0001705301173149561,
      "loss": 1.2125,
      "step": 103150
    },
    {
      "epoch": 3.081240856298331,
      "grad_norm": 4.263118267059326,
      "learning_rate": 0.00017051583151903726,
      "loss": 1.2164,
      "step": 103200
    },
    {
      "epoch": 3.0827337055504134,
      "grad_norm": 7.199416160583496,
      "learning_rate": 0.00017050154572311843,
      "loss": 1.2191,
      "step": 103250
    },
    {
      "epoch": 3.084226554802496,
      "grad_norm": 5.3002119064331055,
      "learning_rate": 0.0001704872599271996,
      "loss": 1.2828,
      "step": 103300
    },
    {
      "epoch": 3.0857194040545783,
      "grad_norm": 4.3020405769348145,
      "learning_rate": 0.00017047297413128076,
      "loss": 1.2341,
      "step": 103350
    },
    {
      "epoch": 3.087212253306661,
      "grad_norm": 5.1998443603515625,
      "learning_rate": 0.00017045868833536192,
      "loss": 1.2271,
      "step": 103400
    },
    {
      "epoch": 3.0887051025587438,
      "grad_norm": 4.382910251617432,
      "learning_rate": 0.00017044440253944309,
      "loss": 1.1994,
      "step": 103450
    },
    {
      "epoch": 3.090197951810826,
      "grad_norm": 5.094580173492432,
      "learning_rate": 0.00017043011674352425,
      "loss": 1.2557,
      "step": 103500
    },
    {
      "epoch": 3.0916908010629087,
      "grad_norm": 3.9512152671813965,
      "learning_rate": 0.00017041583094760542,
      "loss": 1.283,
      "step": 103550
    },
    {
      "epoch": 3.093183650314991,
      "grad_norm": 6.366419792175293,
      "learning_rate": 0.00017040154515168658,
      "loss": 1.2327,
      "step": 103600
    },
    {
      "epoch": 3.0946764995670737,
      "grad_norm": 3.0555920600891113,
      "learning_rate": 0.00017038725935576777,
      "loss": 1.1865,
      "step": 103650
    },
    {
      "epoch": 3.0961693488191564,
      "grad_norm": 5.861559867858887,
      "learning_rate": 0.0001703729735598489,
      "loss": 1.2394,
      "step": 103700
    },
    {
      "epoch": 3.0976621980712387,
      "grad_norm": 4.790325164794922,
      "learning_rate": 0.0001703586877639301,
      "loss": 1.2256,
      "step": 103750
    },
    {
      "epoch": 3.0991550473233214,
      "grad_norm": 4.111683368682861,
      "learning_rate": 0.00017034440196801124,
      "loss": 1.2228,
      "step": 103800
    },
    {
      "epoch": 3.1006478965754036,
      "grad_norm": 4.152289390563965,
      "learning_rate": 0.00017033011617209243,
      "loss": 1.2547,
      "step": 103850
    },
    {
      "epoch": 3.1021407458274863,
      "grad_norm": 5.741196155548096,
      "learning_rate": 0.00017031583037617357,
      "loss": 1.2049,
      "step": 103900
    },
    {
      "epoch": 3.103633595079569,
      "grad_norm": 6.526316165924072,
      "learning_rate": 0.00017030154458025476,
      "loss": 1.2845,
      "step": 103950
    },
    {
      "epoch": 3.1051264443316513,
      "grad_norm": 4.736927509307861,
      "learning_rate": 0.00017028725878433592,
      "loss": 1.2777,
      "step": 104000
    },
    {
      "epoch": 3.106619293583734,
      "grad_norm": 4.231018543243408,
      "learning_rate": 0.0001702729729884171,
      "loss": 1.2387,
      "step": 104050
    },
    {
      "epoch": 3.1081121428358163,
      "grad_norm": 4.244640350341797,
      "learning_rate": 0.00017025868719249825,
      "loss": 1.1534,
      "step": 104100
    },
    {
      "epoch": 3.109604992087899,
      "grad_norm": 4.443108558654785,
      "learning_rate": 0.00017024440139657942,
      "loss": 1.198,
      "step": 104150
    },
    {
      "epoch": 3.1110978413399817,
      "grad_norm": 4.030060768127441,
      "learning_rate": 0.00017023011560066058,
      "loss": 1.2005,
      "step": 104200
    },
    {
      "epoch": 3.112590690592064,
      "grad_norm": 4.423096179962158,
      "learning_rate": 0.00017021582980474175,
      "loss": 1.2502,
      "step": 104250
    },
    {
      "epoch": 3.1140835398441467,
      "grad_norm": 4.914687156677246,
      "learning_rate": 0.0001702015440088229,
      "loss": 1.2291,
      "step": 104300
    },
    {
      "epoch": 3.115576389096229,
      "grad_norm": 6.267758369445801,
      "learning_rate": 0.00017018725821290408,
      "loss": 1.2694,
      "step": 104350
    },
    {
      "epoch": 3.1170692383483116,
      "grad_norm": 4.870142936706543,
      "learning_rate": 0.00017017297241698524,
      "loss": 1.1742,
      "step": 104400
    },
    {
      "epoch": 3.1185620876003943,
      "grad_norm": 4.136995792388916,
      "learning_rate": 0.00017015868662106643,
      "loss": 1.2683,
      "step": 104450
    },
    {
      "epoch": 3.1200549368524766,
      "grad_norm": 3.877889633178711,
      "learning_rate": 0.00017014440082514757,
      "loss": 1.2226,
      "step": 104500
    },
    {
      "epoch": 3.1215477861045593,
      "grad_norm": 5.544692039489746,
      "learning_rate": 0.00017013011502922876,
      "loss": 1.2163,
      "step": 104550
    },
    {
      "epoch": 3.1230406353566416,
      "grad_norm": 4.616161346435547,
      "learning_rate": 0.0001701158292333099,
      "loss": 1.2569,
      "step": 104600
    },
    {
      "epoch": 3.1245334846087243,
      "grad_norm": 4.393960475921631,
      "learning_rate": 0.0001701015434373911,
      "loss": 1.277,
      "step": 104650
    },
    {
      "epoch": 3.126026333860807,
      "grad_norm": 5.330219268798828,
      "learning_rate": 0.00017008725764147223,
      "loss": 1.1929,
      "step": 104700
    },
    {
      "epoch": 3.1275191831128892,
      "grad_norm": 5.1908979415893555,
      "learning_rate": 0.00017007297184555342,
      "loss": 1.2356,
      "step": 104750
    },
    {
      "epoch": 3.129012032364972,
      "grad_norm": 3.941892623901367,
      "learning_rate": 0.00017005868604963459,
      "loss": 1.1774,
      "step": 104800
    },
    {
      "epoch": 3.130504881617054,
      "grad_norm": 5.030502796173096,
      "learning_rate": 0.00017004440025371575,
      "loss": 1.2153,
      "step": 104850
    },
    {
      "epoch": 3.131997730869137,
      "grad_norm": 3.437596082687378,
      "learning_rate": 0.00017003011445779692,
      "loss": 1.2438,
      "step": 104900
    },
    {
      "epoch": 3.133490580121219,
      "grad_norm": 4.600546836853027,
      "learning_rate": 0.00017001582866187808,
      "loss": 1.2477,
      "step": 104950
    },
    {
      "epoch": 3.134983429373302,
      "grad_norm": 7.366837501525879,
      "learning_rate": 0.00017000154286595924,
      "loss": 1.2411,
      "step": 105000
    },
    {
      "epoch": 3.1364762786253846,
      "grad_norm": 4.779984474182129,
      "learning_rate": 0.0001699872570700404,
      "loss": 1.1974,
      "step": 105050
    },
    {
      "epoch": 3.137969127877467,
      "grad_norm": 3.6350603103637695,
      "learning_rate": 0.00016997297127412157,
      "loss": 1.2811,
      "step": 105100
    },
    {
      "epoch": 3.1394619771295496,
      "grad_norm": 3.3783535957336426,
      "learning_rate": 0.00016995868547820274,
      "loss": 1.2377,
      "step": 105150
    },
    {
      "epoch": 3.140954826381632,
      "grad_norm": 5.765346527099609,
      "learning_rate": 0.0001699443996822839,
      "loss": 1.2648,
      "step": 105200
    },
    {
      "epoch": 3.1424476756337145,
      "grad_norm": 6.171380996704102,
      "learning_rate": 0.0001699301138863651,
      "loss": 1.2653,
      "step": 105250
    },
    {
      "epoch": 3.1439405248857972,
      "grad_norm": 4.658510684967041,
      "learning_rate": 0.00016991582809044623,
      "loss": 1.2838,
      "step": 105300
    },
    {
      "epoch": 3.1454333741378795,
      "grad_norm": 4.044935703277588,
      "learning_rate": 0.00016990154229452742,
      "loss": 1.2807,
      "step": 105350
    },
    {
      "epoch": 3.146926223389962,
      "grad_norm": 3.6364893913269043,
      "learning_rate": 0.00016988725649860856,
      "loss": 1.2631,
      "step": 105400
    },
    {
      "epoch": 3.1484190726420445,
      "grad_norm": 4.018857955932617,
      "learning_rate": 0.00016987297070268975,
      "loss": 1.2301,
      "step": 105450
    },
    {
      "epoch": 3.149911921894127,
      "grad_norm": 4.4695725440979,
      "learning_rate": 0.0001698586849067709,
      "loss": 1.2299,
      "step": 105500
    },
    {
      "epoch": 3.1514047711462094,
      "grad_norm": 3.9617760181427,
      "learning_rate": 0.00016984439911085206,
      "loss": 1.2364,
      "step": 105550
    },
    {
      "epoch": 3.152897620398292,
      "grad_norm": 4.522684574127197,
      "learning_rate": 0.00016983011331493325,
      "loss": 1.2737,
      "step": 105600
    },
    {
      "epoch": 3.154390469650375,
      "grad_norm": 4.221574783325195,
      "learning_rate": 0.00016981582751901439,
      "loss": 1.2214,
      "step": 105650
    },
    {
      "epoch": 3.155883318902457,
      "grad_norm": 4.848055839538574,
      "learning_rate": 0.00016980154172309558,
      "loss": 1.1756,
      "step": 105700
    },
    {
      "epoch": 3.15737616815454,
      "grad_norm": 4.364899635314941,
      "learning_rate": 0.00016978725592717671,
      "loss": 1.2188,
      "step": 105750
    },
    {
      "epoch": 3.158869017406622,
      "grad_norm": 3.952366352081299,
      "learning_rate": 0.0001697729701312579,
      "loss": 1.2422,
      "step": 105800
    },
    {
      "epoch": 3.1603618666587048,
      "grad_norm": 4.575440883636475,
      "learning_rate": 0.00016975868433533907,
      "loss": 1.2711,
      "step": 105850
    },
    {
      "epoch": 3.1618547159107875,
      "grad_norm": 5.710911273956299,
      "learning_rate": 0.00016974439853942024,
      "loss": 1.1996,
      "step": 105900
    },
    {
      "epoch": 3.1633475651628697,
      "grad_norm": 3.885944128036499,
      "learning_rate": 0.0001697301127435014,
      "loss": 1.2633,
      "step": 105950
    },
    {
      "epoch": 3.1648404144149525,
      "grad_norm": 4.504218578338623,
      "learning_rate": 0.00016971582694758256,
      "loss": 1.2028,
      "step": 106000
    },
    {
      "epoch": 3.1663332636670347,
      "grad_norm": 5.381597995758057,
      "learning_rate": 0.00016970154115166373,
      "loss": 1.2445,
      "step": 106050
    },
    {
      "epoch": 3.1678261129191174,
      "grad_norm": 3.7343199253082275,
      "learning_rate": 0.0001696872553557449,
      "loss": 1.2553,
      "step": 106100
    },
    {
      "epoch": 3.1693189621712,
      "grad_norm": 4.142834186553955,
      "learning_rate": 0.00016967296955982606,
      "loss": 1.2319,
      "step": 106150
    },
    {
      "epoch": 3.1708118114232824,
      "grad_norm": 3.856889486312866,
      "learning_rate": 0.00016965868376390722,
      "loss": 1.2911,
      "step": 106200
    },
    {
      "epoch": 3.172304660675365,
      "grad_norm": 4.8363938331604,
      "learning_rate": 0.0001696443979679884,
      "loss": 1.2712,
      "step": 106250
    },
    {
      "epoch": 3.1737975099274474,
      "grad_norm": 5.556827068328857,
      "learning_rate": 0.00016963011217206955,
      "loss": 1.2295,
      "step": 106300
    },
    {
      "epoch": 3.17529035917953,
      "grad_norm": 4.371210098266602,
      "learning_rate": 0.00016961582637615072,
      "loss": 1.2745,
      "step": 106350
    },
    {
      "epoch": 3.1767832084316128,
      "grad_norm": 4.789484024047852,
      "learning_rate": 0.0001696015405802319,
      "loss": 1.232,
      "step": 106400
    },
    {
      "epoch": 3.178276057683695,
      "grad_norm": 3.9675486087799072,
      "learning_rate": 0.00016958725478431305,
      "loss": 1.2131,
      "step": 106450
    },
    {
      "epoch": 3.1797689069357777,
      "grad_norm": 6.18342924118042,
      "learning_rate": 0.00016957296898839424,
      "loss": 1.2337,
      "step": 106500
    },
    {
      "epoch": 3.18126175618786,
      "grad_norm": 3.3635101318359375,
      "learning_rate": 0.00016955868319247538,
      "loss": 1.2632,
      "step": 106550
    },
    {
      "epoch": 3.1827546054399427,
      "grad_norm": 3.658357620239258,
      "learning_rate": 0.00016954439739655657,
      "loss": 1.2306,
      "step": 106600
    },
    {
      "epoch": 3.1842474546920254,
      "grad_norm": 5.325497627258301,
      "learning_rate": 0.00016953011160063773,
      "loss": 1.2381,
      "step": 106650
    },
    {
      "epoch": 3.1857403039441077,
      "grad_norm": 5.2231268882751465,
      "learning_rate": 0.0001695158258047189,
      "loss": 1.2236,
      "step": 106700
    },
    {
      "epoch": 3.1872331531961904,
      "grad_norm": 5.6110334396362305,
      "learning_rate": 0.00016950154000880006,
      "loss": 1.2479,
      "step": 106750
    },
    {
      "epoch": 3.1887260024482726,
      "grad_norm": 4.795280933380127,
      "learning_rate": 0.00016948725421288123,
      "loss": 1.1846,
      "step": 106800
    },
    {
      "epoch": 3.1902188517003554,
      "grad_norm": 5.203749179840088,
      "learning_rate": 0.0001694729684169624,
      "loss": 1.2531,
      "step": 106850
    },
    {
      "epoch": 3.191711700952438,
      "grad_norm": 6.236663818359375,
      "learning_rate": 0.00016945868262104356,
      "loss": 1.1977,
      "step": 106900
    },
    {
      "epoch": 3.1932045502045203,
      "grad_norm": 4.13303279876709,
      "learning_rate": 0.00016944439682512472,
      "loss": 1.2418,
      "step": 106950
    },
    {
      "epoch": 3.194697399456603,
      "grad_norm": 6.106944561004639,
      "learning_rate": 0.00016943011102920589,
      "loss": 1.1926,
      "step": 107000
    },
    {
      "epoch": 3.1961902487086853,
      "grad_norm": 4.1147332191467285,
      "learning_rate": 0.00016941582523328705,
      "loss": 1.2549,
      "step": 107050
    },
    {
      "epoch": 3.197683097960768,
      "grad_norm": 4.951442718505859,
      "learning_rate": 0.00016940153943736824,
      "loss": 1.2964,
      "step": 107100
    },
    {
      "epoch": 3.1991759472128503,
      "grad_norm": 6.196277141571045,
      "learning_rate": 0.00016938725364144938,
      "loss": 1.2541,
      "step": 107150
    },
    {
      "epoch": 3.200668796464933,
      "grad_norm": 5.174891948699951,
      "learning_rate": 0.00016937296784553057,
      "loss": 1.238,
      "step": 107200
    },
    {
      "epoch": 3.2021616457170157,
      "grad_norm": 6.037936687469482,
      "learning_rate": 0.0001693586820496117,
      "loss": 1.2919,
      "step": 107250
    },
    {
      "epoch": 3.203654494969098,
      "grad_norm": 4.104307651519775,
      "learning_rate": 0.0001693443962536929,
      "loss": 1.178,
      "step": 107300
    },
    {
      "epoch": 3.2051473442211806,
      "grad_norm": 4.1462273597717285,
      "learning_rate": 0.00016933011045777404,
      "loss": 1.2468,
      "step": 107350
    },
    {
      "epoch": 3.206640193473263,
      "grad_norm": 3.701998233795166,
      "learning_rate": 0.00016931582466185523,
      "loss": 1.2262,
      "step": 107400
    },
    {
      "epoch": 3.2081330427253456,
      "grad_norm": 4.66901159286499,
      "learning_rate": 0.0001693015388659364,
      "loss": 1.2413,
      "step": 107450
    },
    {
      "epoch": 3.2096258919774283,
      "grad_norm": 5.96376371383667,
      "learning_rate": 0.00016928725307001756,
      "loss": 1.1554,
      "step": 107500
    },
    {
      "epoch": 3.2111187412295106,
      "grad_norm": 3.801471710205078,
      "learning_rate": 0.00016927296727409872,
      "loss": 1.2821,
      "step": 107550
    },
    {
      "epoch": 3.2126115904815933,
      "grad_norm": 5.026028633117676,
      "learning_rate": 0.0001692586814781799,
      "loss": 1.2307,
      "step": 107600
    },
    {
      "epoch": 3.2141044397336755,
      "grad_norm": 5.475889682769775,
      "learning_rate": 0.00016924439568226105,
      "loss": 1.2,
      "step": 107650
    },
    {
      "epoch": 3.2155972889857583,
      "grad_norm": 6.249619007110596,
      "learning_rate": 0.00016923010988634222,
      "loss": 1.2448,
      "step": 107700
    },
    {
      "epoch": 3.2170901382378405,
      "grad_norm": 4.803418159484863,
      "learning_rate": 0.00016921582409042338,
      "loss": 1.2669,
      "step": 107750
    },
    {
      "epoch": 3.218582987489923,
      "grad_norm": 3.9626855850219727,
      "learning_rate": 0.00016920153829450455,
      "loss": 1.2167,
      "step": 107800
    },
    {
      "epoch": 3.220075836742006,
      "grad_norm": 3.5219650268554688,
      "learning_rate": 0.0001691872524985857,
      "loss": 1.2554,
      "step": 107850
    },
    {
      "epoch": 3.221568685994088,
      "grad_norm": 4.510613441467285,
      "learning_rate": 0.0001691729667026669,
      "loss": 1.1923,
      "step": 107900
    },
    {
      "epoch": 3.223061535246171,
      "grad_norm": 5.354753494262695,
      "learning_rate": 0.00016915868090674804,
      "loss": 1.1883,
      "step": 107950
    },
    {
      "epoch": 3.224554384498253,
      "grad_norm": 4.148642063140869,
      "learning_rate": 0.00016914439511082923,
      "loss": 1.2497,
      "step": 108000
    },
    {
      "epoch": 3.226047233750336,
      "grad_norm": 4.1666646003723145,
      "learning_rate": 0.00016913010931491037,
      "loss": 1.1872,
      "step": 108050
    },
    {
      "epoch": 3.2275400830024186,
      "grad_norm": 4.137986183166504,
      "learning_rate": 0.00016911582351899156,
      "loss": 1.2788,
      "step": 108100
    },
    {
      "epoch": 3.229032932254501,
      "grad_norm": 3.901366710662842,
      "learning_rate": 0.0001691015377230727,
      "loss": 1.2651,
      "step": 108150
    },
    {
      "epoch": 3.2305257815065835,
      "grad_norm": 3.96573543548584,
      "learning_rate": 0.00016908725192715386,
      "loss": 1.2277,
      "step": 108200
    },
    {
      "epoch": 3.232018630758666,
      "grad_norm": 4.226304054260254,
      "learning_rate": 0.00016907296613123506,
      "loss": 1.2792,
      "step": 108250
    },
    {
      "epoch": 3.2335114800107485,
      "grad_norm": 5.907839775085449,
      "learning_rate": 0.0001690586803353162,
      "loss": 1.2123,
      "step": 108300
    },
    {
      "epoch": 3.235004329262831,
      "grad_norm": 3.7594401836395264,
      "learning_rate": 0.00016904439453939738,
      "loss": 1.1956,
      "step": 108350
    },
    {
      "epoch": 3.2364971785149135,
      "grad_norm": 3.6791820526123047,
      "learning_rate": 0.00016903010874347852,
      "loss": 1.2103,
      "step": 108400
    },
    {
      "epoch": 3.237990027766996,
      "grad_norm": 4.249289512634277,
      "learning_rate": 0.00016901582294755971,
      "loss": 1.2348,
      "step": 108450
    },
    {
      "epoch": 3.2394828770190784,
      "grad_norm": 5.203214645385742,
      "learning_rate": 0.00016900153715164085,
      "loss": 1.2122,
      "step": 108500
    },
    {
      "epoch": 3.240975726271161,
      "grad_norm": 5.2223591804504395,
      "learning_rate": 0.00016898725135572204,
      "loss": 1.1984,
      "step": 108550
    },
    {
      "epoch": 3.242468575523244,
      "grad_norm": 3.9512999057769775,
      "learning_rate": 0.0001689729655598032,
      "loss": 1.2791,
      "step": 108600
    },
    {
      "epoch": 3.243961424775326,
      "grad_norm": 5.953644752502441,
      "learning_rate": 0.00016895867976388437,
      "loss": 1.2582,
      "step": 108650
    },
    {
      "epoch": 3.245454274027409,
      "grad_norm": 4.990479469299316,
      "learning_rate": 0.00016894439396796554,
      "loss": 1.2594,
      "step": 108700
    },
    {
      "epoch": 3.246947123279491,
      "grad_norm": 4.3011651039123535,
      "learning_rate": 0.0001689301081720467,
      "loss": 1.2351,
      "step": 108750
    },
    {
      "epoch": 3.248439972531574,
      "grad_norm": 4.554538249969482,
      "learning_rate": 0.00016891582237612787,
      "loss": 1.2482,
      "step": 108800
    },
    {
      "epoch": 3.2499328217836565,
      "grad_norm": 4.715415000915527,
      "learning_rate": 0.00016890153658020903,
      "loss": 1.2691,
      "step": 108850
    },
    {
      "epoch": 3.2514256710357388,
      "grad_norm": 5.759798049926758,
      "learning_rate": 0.0001688872507842902,
      "loss": 1.2224,
      "step": 108900
    },
    {
      "epoch": 3.2529185202878215,
      "grad_norm": 3.7031774520874023,
      "learning_rate": 0.00016887296498837136,
      "loss": 1.2166,
      "step": 108950
    },
    {
      "epoch": 3.2544113695399037,
      "grad_norm": 5.22156286239624,
      "learning_rate": 0.00016885867919245253,
      "loss": 1.2465,
      "step": 109000
    },
    {
      "epoch": 3.2559042187919864,
      "grad_norm": 5.56991720199585,
      "learning_rate": 0.00016884439339653372,
      "loss": 1.2274,
      "step": 109050
    },
    {
      "epoch": 3.257397068044069,
      "grad_norm": 4.750823020935059,
      "learning_rate": 0.00016883010760061485,
      "loss": 1.1885,
      "step": 109100
    },
    {
      "epoch": 3.2588899172961514,
      "grad_norm": 4.139292240142822,
      "learning_rate": 0.00016881582180469605,
      "loss": 1.2308,
      "step": 109150
    },
    {
      "epoch": 3.260382766548234,
      "grad_norm": 3.652226209640503,
      "learning_rate": 0.00016880153600877718,
      "loss": 1.212,
      "step": 109200
    },
    {
      "epoch": 3.2618756158003164,
      "grad_norm": 5.819267272949219,
      "learning_rate": 0.00016878725021285838,
      "loss": 1.1852,
      "step": 109250
    },
    {
      "epoch": 3.263368465052399,
      "grad_norm": 4.4434027671813965,
      "learning_rate": 0.00016877296441693954,
      "loss": 1.2124,
      "step": 109300
    },
    {
      "epoch": 3.2648613143044813,
      "grad_norm": 3.6519103050231934,
      "learning_rate": 0.0001687586786210207,
      "loss": 1.2033,
      "step": 109350
    },
    {
      "epoch": 3.266354163556564,
      "grad_norm": 5.874778747558594,
      "learning_rate": 0.00016874439282510187,
      "loss": 1.2745,
      "step": 109400
    },
    {
      "epoch": 3.2678470128086468,
      "grad_norm": 4.622878551483154,
      "learning_rate": 0.00016873010702918303,
      "loss": 1.2599,
      "step": 109450
    },
    {
      "epoch": 3.269339862060729,
      "grad_norm": 4.694761276245117,
      "learning_rate": 0.0001687158212332642,
      "loss": 1.2233,
      "step": 109500
    },
    {
      "epoch": 3.2708327113128117,
      "grad_norm": 5.996031284332275,
      "learning_rate": 0.00016870153543734536,
      "loss": 1.2452,
      "step": 109550
    },
    {
      "epoch": 3.272325560564894,
      "grad_norm": 3.6771366596221924,
      "learning_rate": 0.00016868724964142653,
      "loss": 1.253,
      "step": 109600
    },
    {
      "epoch": 3.2738184098169767,
      "grad_norm": 3.622293710708618,
      "learning_rate": 0.0001686729638455077,
      "loss": 1.2456,
      "step": 109650
    },
    {
      "epoch": 3.2753112590690594,
      "grad_norm": 6.092465400695801,
      "learning_rate": 0.00016865867804958886,
      "loss": 1.2742,
      "step": 109700
    },
    {
      "epoch": 3.2768041083211417,
      "grad_norm": 4.962725639343262,
      "learning_rate": 0.00016864439225367005,
      "loss": 1.2426,
      "step": 109750
    },
    {
      "epoch": 3.2782969575732244,
      "grad_norm": 4.702395915985107,
      "learning_rate": 0.0001686301064577512,
      "loss": 1.2324,
      "step": 109800
    },
    {
      "epoch": 3.2797898068253066,
      "grad_norm": 4.226596832275391,
      "learning_rate": 0.00016861582066183238,
      "loss": 1.2791,
      "step": 109850
    },
    {
      "epoch": 3.2812826560773893,
      "grad_norm": 4.105664253234863,
      "learning_rate": 0.00016860153486591352,
      "loss": 1.2606,
      "step": 109900
    },
    {
      "epoch": 3.2827755053294716,
      "grad_norm": 4.2797369956970215,
      "learning_rate": 0.0001685872490699947,
      "loss": 1.3038,
      "step": 109950
    },
    {
      "epoch": 3.2842683545815543,
      "grad_norm": 5.301023483276367,
      "learning_rate": 0.00016857296327407585,
      "loss": 1.1694,
      "step": 110000
    },
    {
      "epoch": 3.285761203833637,
      "grad_norm": 4.061399459838867,
      "learning_rate": 0.00016855867747815704,
      "loss": 1.2366,
      "step": 110050
    },
    {
      "epoch": 3.2872540530857193,
      "grad_norm": 4.011737823486328,
      "learning_rate": 0.0001685443916822382,
      "loss": 1.2354,
      "step": 110100
    },
    {
      "epoch": 3.288746902337802,
      "grad_norm": 4.765646457672119,
      "learning_rate": 0.00016853010588631937,
      "loss": 1.2935,
      "step": 110150
    },
    {
      "epoch": 3.2902397515898842,
      "grad_norm": 4.911751747131348,
      "learning_rate": 0.00016851582009040053,
      "loss": 1.2079,
      "step": 110200
    },
    {
      "epoch": 3.291732600841967,
      "grad_norm": 5.475641250610352,
      "learning_rate": 0.0001685015342944817,
      "loss": 1.2861,
      "step": 110250
    },
    {
      "epoch": 3.2932254500940497,
      "grad_norm": 4.465792655944824,
      "learning_rate": 0.00016848724849856286,
      "loss": 1.239,
      "step": 110300
    },
    {
      "epoch": 3.294718299346132,
      "grad_norm": 6.820103645324707,
      "learning_rate": 0.00016847296270264403,
      "loss": 1.1682,
      "step": 110350
    },
    {
      "epoch": 3.2962111485982146,
      "grad_norm": 6.0468010902404785,
      "learning_rate": 0.0001684586769067252,
      "loss": 1.2195,
      "step": 110400
    },
    {
      "epoch": 3.297703997850297,
      "grad_norm": 6.527349472045898,
      "learning_rate": 0.00016844439111080635,
      "loss": 1.2655,
      "step": 110450
    },
    {
      "epoch": 3.2991968471023796,
      "grad_norm": 4.763921737670898,
      "learning_rate": 0.00016843010531488752,
      "loss": 1.2427,
      "step": 110500
    },
    {
      "epoch": 3.3006896963544623,
      "grad_norm": 5.6205949783325195,
      "learning_rate": 0.0001684158195189687,
      "loss": 1.2629,
      "step": 110550
    },
    {
      "epoch": 3.3021825456065446,
      "grad_norm": 4.788918972015381,
      "learning_rate": 0.00016840153372304985,
      "loss": 1.246,
      "step": 110600
    },
    {
      "epoch": 3.3036753948586273,
      "grad_norm": 4.326768398284912,
      "learning_rate": 0.00016838724792713104,
      "loss": 1.23,
      "step": 110650
    },
    {
      "epoch": 3.3051682441107095,
      "grad_norm": 4.26240873336792,
      "learning_rate": 0.00016837296213121218,
      "loss": 1.2487,
      "step": 110700
    },
    {
      "epoch": 3.3066610933627922,
      "grad_norm": 5.802476406097412,
      "learning_rate": 0.00016835867633529334,
      "loss": 1.2401,
      "step": 110750
    },
    {
      "epoch": 3.308153942614875,
      "grad_norm": 3.884272575378418,
      "learning_rate": 0.0001683443905393745,
      "loss": 1.2199,
      "step": 110800
    },
    {
      "epoch": 3.309646791866957,
      "grad_norm": 5.074997425079346,
      "learning_rate": 0.00016833010474345567,
      "loss": 1.2285,
      "step": 110850
    },
    {
      "epoch": 3.31113964111904,
      "grad_norm": 3.9242136478424072,
      "learning_rate": 0.00016831581894753686,
      "loss": 1.2276,
      "step": 110900
    },
    {
      "epoch": 3.312632490371122,
      "grad_norm": 5.645691394805908,
      "learning_rate": 0.000168301533151618,
      "loss": 1.2031,
      "step": 110950
    },
    {
      "epoch": 3.314125339623205,
      "grad_norm": 6.003003120422363,
      "learning_rate": 0.0001682872473556992,
      "loss": 1.2442,
      "step": 111000
    },
    {
      "epoch": 3.3156181888752876,
      "grad_norm": 4.822206020355225,
      "learning_rate": 0.00016827296155978033,
      "loss": 1.2398,
      "step": 111050
    },
    {
      "epoch": 3.31711103812737,
      "grad_norm": 3.79050350189209,
      "learning_rate": 0.00016825867576386152,
      "loss": 1.2108,
      "step": 111100
    },
    {
      "epoch": 3.3186038873794526,
      "grad_norm": 5.795800685882568,
      "learning_rate": 0.00016824438996794266,
      "loss": 1.2537,
      "step": 111150
    },
    {
      "epoch": 3.320096736631535,
      "grad_norm": 4.554315090179443,
      "learning_rate": 0.00016823010417202385,
      "loss": 1.2485,
      "step": 111200
    },
    {
      "epoch": 3.3215895858836175,
      "grad_norm": 3.562445640563965,
      "learning_rate": 0.00016821581837610502,
      "loss": 1.2802,
      "step": 111250
    },
    {
      "epoch": 3.3230824351357002,
      "grad_norm": 4.7466278076171875,
      "learning_rate": 0.00016820153258018618,
      "loss": 1.2108,
      "step": 111300
    },
    {
      "epoch": 3.3245752843877825,
      "grad_norm": 4.278189659118652,
      "learning_rate": 0.00016818724678426735,
      "loss": 1.2035,
      "step": 111350
    },
    {
      "epoch": 3.326068133639865,
      "grad_norm": 3.9288856983184814,
      "learning_rate": 0.0001681729609883485,
      "loss": 1.2155,
      "step": 111400
    },
    {
      "epoch": 3.3275609828919475,
      "grad_norm": 5.247483253479004,
      "learning_rate": 0.00016815867519242967,
      "loss": 1.2523,
      "step": 111450
    },
    {
      "epoch": 3.32905383214403,
      "grad_norm": 5.946502685546875,
      "learning_rate": 0.00016814438939651084,
      "loss": 1.2254,
      "step": 111500
    },
    {
      "epoch": 3.3305466813961124,
      "grad_norm": 4.208354473114014,
      "learning_rate": 0.000168130103600592,
      "loss": 1.2168,
      "step": 111550
    },
    {
      "epoch": 3.332039530648195,
      "grad_norm": 6.033560752868652,
      "learning_rate": 0.00016811581780467317,
      "loss": 1.2876,
      "step": 111600
    },
    {
      "epoch": 3.333532379900278,
      "grad_norm": 4.848031520843506,
      "learning_rate": 0.00016810153200875433,
      "loss": 1.2404,
      "step": 111650
    },
    {
      "epoch": 3.33502522915236,
      "grad_norm": 3.873622179031372,
      "learning_rate": 0.00016808724621283553,
      "loss": 1.2489,
      "step": 111700
    },
    {
      "epoch": 3.336518078404443,
      "grad_norm": 4.604708671569824,
      "learning_rate": 0.00016807296041691666,
      "loss": 1.2255,
      "step": 111750
    },
    {
      "epoch": 3.338010927656525,
      "grad_norm": 5.416150093078613,
      "learning_rate": 0.00016805867462099785,
      "loss": 1.2108,
      "step": 111800
    },
    {
      "epoch": 3.339503776908608,
      "grad_norm": 4.228431224822998,
      "learning_rate": 0.000168044388825079,
      "loss": 1.2447,
      "step": 111850
    },
    {
      "epoch": 3.3409966261606905,
      "grad_norm": 4.664206027984619,
      "learning_rate": 0.00016803010302916018,
      "loss": 1.2296,
      "step": 111900
    },
    {
      "epoch": 3.3424894754127727,
      "grad_norm": 4.114374160766602,
      "learning_rate": 0.00016801581723324132,
      "loss": 1.1755,
      "step": 111950
    },
    {
      "epoch": 3.3439823246648555,
      "grad_norm": 5.543557643890381,
      "learning_rate": 0.0001680015314373225,
      "loss": 1.2252,
      "step": 112000
    },
    {
      "epoch": 3.3454751739169377,
      "grad_norm": 4.992297649383545,
      "learning_rate": 0.00016798724564140368,
      "loss": 1.2215,
      "step": 112050
    },
    {
      "epoch": 3.3469680231690204,
      "grad_norm": 5.890657901763916,
      "learning_rate": 0.00016797295984548484,
      "loss": 1.258,
      "step": 112100
    },
    {
      "epoch": 3.3484608724211027,
      "grad_norm": 8.049378395080566,
      "learning_rate": 0.000167958674049566,
      "loss": 1.2354,
      "step": 112150
    },
    {
      "epoch": 3.3499537216731854,
      "grad_norm": 4.3732075691223145,
      "learning_rate": 0.00016794438825364717,
      "loss": 1.1945,
      "step": 112200
    },
    {
      "epoch": 3.351446570925268,
      "grad_norm": 6.537196159362793,
      "learning_rate": 0.00016793010245772834,
      "loss": 1.2857,
      "step": 112250
    },
    {
      "epoch": 3.3529394201773504,
      "grad_norm": 4.264357089996338,
      "learning_rate": 0.0001679158166618095,
      "loss": 1.2425,
      "step": 112300
    },
    {
      "epoch": 3.354432269429433,
      "grad_norm": 5.0003662109375,
      "learning_rate": 0.00016790153086589067,
      "loss": 1.2781,
      "step": 112350
    },
    {
      "epoch": 3.3559251186815153,
      "grad_norm": 3.599029064178467,
      "learning_rate": 0.00016788724506997183,
      "loss": 1.1866,
      "step": 112400
    },
    {
      "epoch": 3.357417967933598,
      "grad_norm": 5.207790374755859,
      "learning_rate": 0.000167872959274053,
      "loss": 1.248,
      "step": 112450
    },
    {
      "epoch": 3.3589108171856807,
      "grad_norm": 4.075089931488037,
      "learning_rate": 0.0001678586734781342,
      "loss": 1.195,
      "step": 112500
    },
    {
      "epoch": 3.360403666437763,
      "grad_norm": 5.74088716506958,
      "learning_rate": 0.00016784438768221532,
      "loss": 1.2343,
      "step": 112550
    },
    {
      "epoch": 3.3618965156898457,
      "grad_norm": 4.6949262619018555,
      "learning_rate": 0.00016783010188629652,
      "loss": 1.2295,
      "step": 112600
    },
    {
      "epoch": 3.363389364941928,
      "grad_norm": 5.432741165161133,
      "learning_rate": 0.00016781581609037765,
      "loss": 1.3118,
      "step": 112650
    },
    {
      "epoch": 3.3648822141940107,
      "grad_norm": 4.652493000030518,
      "learning_rate": 0.00016780153029445885,
      "loss": 1.2686,
      "step": 112700
    },
    {
      "epoch": 3.3663750634460934,
      "grad_norm": 4.528727054595947,
      "learning_rate": 0.00016778724449854,
      "loss": 1.2235,
      "step": 112750
    },
    {
      "epoch": 3.3678679126981756,
      "grad_norm": 4.215224266052246,
      "learning_rate": 0.00016777295870262117,
      "loss": 1.2993,
      "step": 112800
    },
    {
      "epoch": 3.3693607619502584,
      "grad_norm": 4.306272506713867,
      "learning_rate": 0.00016775867290670234,
      "loss": 1.2298,
      "step": 112850
    },
    {
      "epoch": 3.3708536112023406,
      "grad_norm": 3.477867841720581,
      "learning_rate": 0.0001677443871107835,
      "loss": 1.2557,
      "step": 112900
    },
    {
      "epoch": 3.3723464604544233,
      "grad_norm": 4.306337833404541,
      "learning_rate": 0.00016773010131486467,
      "loss": 1.1917,
      "step": 112950
    },
    {
      "epoch": 3.373839309706506,
      "grad_norm": 4.982789993286133,
      "learning_rate": 0.00016771581551894583,
      "loss": 1.2572,
      "step": 113000
    },
    {
      "epoch": 3.3753321589585883,
      "grad_norm": 4.850037097930908,
      "learning_rate": 0.000167701529723027,
      "loss": 1.2686,
      "step": 113050
    },
    {
      "epoch": 3.376825008210671,
      "grad_norm": 4.561743259429932,
      "learning_rate": 0.00016768724392710816,
      "loss": 1.2696,
      "step": 113100
    },
    {
      "epoch": 3.3783178574627533,
      "grad_norm": 4.380636692047119,
      "learning_rate": 0.00016767295813118933,
      "loss": 1.2335,
      "step": 113150
    },
    {
      "epoch": 3.379810706714836,
      "grad_norm": 5.846193313598633,
      "learning_rate": 0.00016765867233527052,
      "loss": 1.2353,
      "step": 113200
    },
    {
      "epoch": 3.3813035559669187,
      "grad_norm": 3.9056570529937744,
      "learning_rate": 0.00016764438653935166,
      "loss": 1.3195,
      "step": 113250
    },
    {
      "epoch": 3.382796405219001,
      "grad_norm": 5.165102958679199,
      "learning_rate": 0.00016763010074343285,
      "loss": 1.2193,
      "step": 113300
    },
    {
      "epoch": 3.3842892544710836,
      "grad_norm": 4.945791244506836,
      "learning_rate": 0.00016761581494751399,
      "loss": 1.22,
      "step": 113350
    },
    {
      "epoch": 3.385782103723166,
      "grad_norm": 6.68800687789917,
      "learning_rate": 0.00016760152915159515,
      "loss": 1.2546,
      "step": 113400
    },
    {
      "epoch": 3.3872749529752486,
      "grad_norm": 8.491449356079102,
      "learning_rate": 0.00016758724335567632,
      "loss": 1.2185,
      "step": 113450
    },
    {
      "epoch": 3.3887678022273313,
      "grad_norm": 5.642709255218506,
      "learning_rate": 0.00016757295755975748,
      "loss": 1.2071,
      "step": 113500
    },
    {
      "epoch": 3.3902606514794136,
      "grad_norm": 4.33923864364624,
      "learning_rate": 0.00016755867176383867,
      "loss": 1.2902,
      "step": 113550
    },
    {
      "epoch": 3.3917535007314963,
      "grad_norm": 9.632731437683105,
      "learning_rate": 0.0001675443859679198,
      "loss": 1.2163,
      "step": 113600
    },
    {
      "epoch": 3.3932463499835785,
      "grad_norm": 4.608974456787109,
      "learning_rate": 0.000167530100172001,
      "loss": 1.2247,
      "step": 113650
    },
    {
      "epoch": 3.3947391992356613,
      "grad_norm": 3.9156320095062256,
      "learning_rate": 0.00016751581437608214,
      "loss": 1.2184,
      "step": 113700
    },
    {
      "epoch": 3.3962320484877435,
      "grad_norm": 3.8621175289154053,
      "learning_rate": 0.00016750152858016333,
      "loss": 1.1579,
      "step": 113750
    },
    {
      "epoch": 3.397724897739826,
      "grad_norm": 4.917006969451904,
      "learning_rate": 0.00016748724278424447,
      "loss": 1.2958,
      "step": 113800
    },
    {
      "epoch": 3.399217746991909,
      "grad_norm": 4.837984561920166,
      "learning_rate": 0.00016747295698832566,
      "loss": 1.25,
      "step": 113850
    },
    {
      "epoch": 3.400710596243991,
      "grad_norm": 5.688441753387451,
      "learning_rate": 0.00016745867119240682,
      "loss": 1.2573,
      "step": 113900
    },
    {
      "epoch": 3.402203445496074,
      "grad_norm": 5.892030239105225,
      "learning_rate": 0.000167444385396488,
      "loss": 1.2146,
      "step": 113950
    },
    {
      "epoch": 3.403696294748156,
      "grad_norm": 5.0287652015686035,
      "learning_rate": 0.00016743009960056915,
      "loss": 1.2544,
      "step": 114000
    },
    {
      "epoch": 3.405189144000239,
      "grad_norm": 3.9118645191192627,
      "learning_rate": 0.00016741581380465032,
      "loss": 1.2881,
      "step": 114050
    },
    {
      "epoch": 3.4066819932523216,
      "grad_norm": 4.481536865234375,
      "learning_rate": 0.00016740152800873148,
      "loss": 1.2773,
      "step": 114100
    },
    {
      "epoch": 3.408174842504404,
      "grad_norm": 4.5398125648498535,
      "learning_rate": 0.00016738724221281265,
      "loss": 1.2502,
      "step": 114150
    },
    {
      "epoch": 3.4096676917564865,
      "grad_norm": 4.823166847229004,
      "learning_rate": 0.0001673729564168938,
      "loss": 1.2654,
      "step": 114200
    },
    {
      "epoch": 3.411160541008569,
      "grad_norm": 6.366767406463623,
      "learning_rate": 0.00016735867062097498,
      "loss": 1.2552,
      "step": 114250
    },
    {
      "epoch": 3.4126533902606515,
      "grad_norm": 4.806399822235107,
      "learning_rate": 0.00016734438482505614,
      "loss": 1.1956,
      "step": 114300
    },
    {
      "epoch": 3.4141462395127338,
      "grad_norm": 7.901057720184326,
      "learning_rate": 0.00016733009902913733,
      "loss": 1.2192,
      "step": 114350
    },
    {
      "epoch": 3.4156390887648165,
      "grad_norm": 4.8132500648498535,
      "learning_rate": 0.00016731581323321847,
      "loss": 1.2719,
      "step": 114400
    },
    {
      "epoch": 3.417131938016899,
      "grad_norm": 4.0474629402160645,
      "learning_rate": 0.00016730152743729966,
      "loss": 1.2212,
      "step": 114450
    },
    {
      "epoch": 3.4186247872689814,
      "grad_norm": 5.218587875366211,
      "learning_rate": 0.0001672872416413808,
      "loss": 1.2162,
      "step": 114500
    },
    {
      "epoch": 3.420117636521064,
      "grad_norm": 4.62558650970459,
      "learning_rate": 0.000167272955845462,
      "loss": 1.2715,
      "step": 114550
    },
    {
      "epoch": 3.4216104857731464,
      "grad_norm": 5.178062915802002,
      "learning_rate": 0.00016725867004954313,
      "loss": 1.2799,
      "step": 114600
    },
    {
      "epoch": 3.423103335025229,
      "grad_norm": 4.345101833343506,
      "learning_rate": 0.00016724438425362432,
      "loss": 1.2453,
      "step": 114650
    },
    {
      "epoch": 3.424596184277312,
      "grad_norm": 4.486167907714844,
      "learning_rate": 0.00016723009845770549,
      "loss": 1.2591,
      "step": 114700
    },
    {
      "epoch": 3.426089033529394,
      "grad_norm": 4.727272033691406,
      "learning_rate": 0.00016721581266178665,
      "loss": 1.3124,
      "step": 114750
    },
    {
      "epoch": 3.427581882781477,
      "grad_norm": 3.8477623462677,
      "learning_rate": 0.00016720152686586782,
      "loss": 1.2389,
      "step": 114800
    },
    {
      "epoch": 3.429074732033559,
      "grad_norm": 3.8839187622070312,
      "learning_rate": 0.00016718724106994898,
      "loss": 1.2635,
      "step": 114850
    },
    {
      "epoch": 3.4305675812856418,
      "grad_norm": 4.389588356018066,
      "learning_rate": 0.00016717295527403014,
      "loss": 1.3258,
      "step": 114900
    },
    {
      "epoch": 3.4320604305377245,
      "grad_norm": 4.100683689117432,
      "learning_rate": 0.0001671586694781113,
      "loss": 1.1715,
      "step": 114950
    },
    {
      "epoch": 3.4335532797898067,
      "grad_norm": 4.422749042510986,
      "learning_rate": 0.00016714438368219247,
      "loss": 1.2415,
      "step": 115000
    },
    {
      "epoch": 3.4350461290418894,
      "grad_norm": 5.368492126464844,
      "learning_rate": 0.00016713009788627364,
      "loss": 1.2644,
      "step": 115050
    },
    {
      "epoch": 3.4365389782939717,
      "grad_norm": 4.524289608001709,
      "learning_rate": 0.0001671158120903548,
      "loss": 1.2413,
      "step": 115100
    },
    {
      "epoch": 3.4380318275460544,
      "grad_norm": 6.099424362182617,
      "learning_rate": 0.000167101526294436,
      "loss": 1.2915,
      "step": 115150
    },
    {
      "epoch": 3.439524676798137,
      "grad_norm": 4.018702507019043,
      "learning_rate": 0.00016708724049851713,
      "loss": 1.2868,
      "step": 115200
    },
    {
      "epoch": 3.4410175260502194,
      "grad_norm": 4.0083112716674805,
      "learning_rate": 0.00016707295470259832,
      "loss": 1.1582,
      "step": 115250
    },
    {
      "epoch": 3.442510375302302,
      "grad_norm": 4.636148452758789,
      "learning_rate": 0.00016705866890667946,
      "loss": 1.2637,
      "step": 115300
    },
    {
      "epoch": 3.4440032245543843,
      "grad_norm": 4.421489238739014,
      "learning_rate": 0.00016704438311076065,
      "loss": 1.2649,
      "step": 115350
    },
    {
      "epoch": 3.445496073806467,
      "grad_norm": 5.654540538787842,
      "learning_rate": 0.00016703009731484182,
      "loss": 1.3091,
      "step": 115400
    },
    {
      "epoch": 3.4469889230585498,
      "grad_norm": 4.134968280792236,
      "learning_rate": 0.00016701581151892298,
      "loss": 1.2496,
      "step": 115450
    },
    {
      "epoch": 3.448481772310632,
      "grad_norm": 4.880650043487549,
      "learning_rate": 0.00016700152572300415,
      "loss": 1.2744,
      "step": 115500
    },
    {
      "epoch": 3.4499746215627147,
      "grad_norm": 6.422858238220215,
      "learning_rate": 0.0001669872399270853,
      "loss": 1.1798,
      "step": 115550
    },
    {
      "epoch": 3.451467470814797,
      "grad_norm": 5.5306315422058105,
      "learning_rate": 0.00016697295413116648,
      "loss": 1.2321,
      "step": 115600
    },
    {
      "epoch": 3.4529603200668797,
      "grad_norm": 3.598203659057617,
      "learning_rate": 0.00016695866833524764,
      "loss": 1.2494,
      "step": 115650
    },
    {
      "epoch": 3.4544531693189624,
      "grad_norm": 4.61868143081665,
      "learning_rate": 0.0001669443825393288,
      "loss": 1.2525,
      "step": 115700
    },
    {
      "epoch": 3.4559460185710447,
      "grad_norm": 3.731189727783203,
      "learning_rate": 0.00016693009674340997,
      "loss": 1.2407,
      "step": 115750
    },
    {
      "epoch": 3.4574388678231274,
      "grad_norm": 5.61367654800415,
      "learning_rate": 0.00016691581094749114,
      "loss": 1.176,
      "step": 115800
    },
    {
      "epoch": 3.4589317170752096,
      "grad_norm": 4.095551490783691,
      "learning_rate": 0.0001669015251515723,
      "loss": 1.2203,
      "step": 115850
    },
    {
      "epoch": 3.4604245663272923,
      "grad_norm": 5.298232555389404,
      "learning_rate": 0.00016688723935565346,
      "loss": 1.2516,
      "step": 115900
    },
    {
      "epoch": 3.4619174155793746,
      "grad_norm": 5.313197135925293,
      "learning_rate": 0.00016687295355973466,
      "loss": 1.2375,
      "step": 115950
    },
    {
      "epoch": 3.4634102648314573,
      "grad_norm": 5.016303062438965,
      "learning_rate": 0.0001668586677638158,
      "loss": 1.2515,
      "step": 116000
    },
    {
      "epoch": 3.46490311408354,
      "grad_norm": 4.23463249206543,
      "learning_rate": 0.00016684438196789696,
      "loss": 1.2218,
      "step": 116050
    },
    {
      "epoch": 3.4663959633356223,
      "grad_norm": 4.841723918914795,
      "learning_rate": 0.00016683009617197812,
      "loss": 1.2503,
      "step": 116100
    },
    {
      "epoch": 3.467888812587705,
      "grad_norm": 4.741985321044922,
      "learning_rate": 0.0001668158103760593,
      "loss": 1.257,
      "step": 116150
    },
    {
      "epoch": 3.4693816618397872,
      "grad_norm": 5.977380275726318,
      "learning_rate": 0.00016680152458014048,
      "loss": 1.2774,
      "step": 116200
    },
    {
      "epoch": 3.47087451109187,
      "grad_norm": 4.758930683135986,
      "learning_rate": 0.00016678723878422162,
      "loss": 1.2718,
      "step": 116250
    },
    {
      "epoch": 3.4723673603439527,
      "grad_norm": 3.923983097076416,
      "learning_rate": 0.0001667729529883028,
      "loss": 1.2274,
      "step": 116300
    },
    {
      "epoch": 3.473860209596035,
      "grad_norm": 4.535523891448975,
      "learning_rate": 0.00016675866719238395,
      "loss": 1.2749,
      "step": 116350
    },
    {
      "epoch": 3.4753530588481176,
      "grad_norm": 3.575572967529297,
      "learning_rate": 0.00016674438139646514,
      "loss": 1.2378,
      "step": 116400
    },
    {
      "epoch": 3.4768459081002,
      "grad_norm": 4.385488510131836,
      "learning_rate": 0.00016673009560054628,
      "loss": 1.278,
      "step": 116450
    },
    {
      "epoch": 3.4783387573522826,
      "grad_norm": 3.994697332382202,
      "learning_rate": 0.00016671580980462747,
      "loss": 1.2324,
      "step": 116500
    },
    {
      "epoch": 3.479831606604365,
      "grad_norm": 4.184852600097656,
      "learning_rate": 0.00016670152400870863,
      "loss": 1.2528,
      "step": 116550
    },
    {
      "epoch": 3.4813244558564476,
      "grad_norm": 4.509354591369629,
      "learning_rate": 0.0001666872382127898,
      "loss": 1.2764,
      "step": 116600
    },
    {
      "epoch": 3.4828173051085303,
      "grad_norm": 4.86820125579834,
      "learning_rate": 0.00016667295241687096,
      "loss": 1.2332,
      "step": 116650
    },
    {
      "epoch": 3.4843101543606125,
      "grad_norm": 5.312245845794678,
      "learning_rate": 0.00016665866662095213,
      "loss": 1.2443,
      "step": 116700
    },
    {
      "epoch": 3.4858030036126952,
      "grad_norm": 4.096804618835449,
      "learning_rate": 0.0001666443808250333,
      "loss": 1.2352,
      "step": 116750
    },
    {
      "epoch": 3.4872958528647775,
      "grad_norm": 4.2911200523376465,
      "learning_rate": 0.00016663009502911446,
      "loss": 1.2366,
      "step": 116800
    },
    {
      "epoch": 3.48878870211686,
      "grad_norm": 5.922074794769287,
      "learning_rate": 0.00016661580923319562,
      "loss": 1.28,
      "step": 116850
    },
    {
      "epoch": 3.490281551368943,
      "grad_norm": 4.1473612785339355,
      "learning_rate": 0.00016660152343727678,
      "loss": 1.1911,
      "step": 116900
    },
    {
      "epoch": 3.491774400621025,
      "grad_norm": 4.428061485290527,
      "learning_rate": 0.00016658723764135795,
      "loss": 1.2649,
      "step": 116950
    },
    {
      "epoch": 3.493267249873108,
      "grad_norm": 4.338399887084961,
      "learning_rate": 0.00016657295184543914,
      "loss": 1.2596,
      "step": 117000
    },
    {
      "epoch": 3.49476009912519,
      "grad_norm": 5.209356307983398,
      "learning_rate": 0.00016655866604952028,
      "loss": 1.2898,
      "step": 117050
    },
    {
      "epoch": 3.496252948377273,
      "grad_norm": 4.742622375488281,
      "learning_rate": 0.00016654438025360147,
      "loss": 1.2243,
      "step": 117100
    },
    {
      "epoch": 3.4977457976293556,
      "grad_norm": 5.037703037261963,
      "learning_rate": 0.0001665300944576826,
      "loss": 1.2425,
      "step": 117150
    },
    {
      "epoch": 3.499238646881438,
      "grad_norm": 4.114877700805664,
      "learning_rate": 0.0001665158086617638,
      "loss": 1.2498,
      "step": 117200
    },
    {
      "epoch": 3.5007314961335205,
      "grad_norm": 3.778136968612671,
      "learning_rate": 0.00016650152286584494,
      "loss": 1.2793,
      "step": 117250
    },
    {
      "epoch": 3.502224345385603,
      "grad_norm": 3.3782308101654053,
      "learning_rate": 0.00016648723706992613,
      "loss": 1.234,
      "step": 117300
    },
    {
      "epoch": 3.5037171946376855,
      "grad_norm": 3.4435694217681885,
      "learning_rate": 0.0001664729512740073,
      "loss": 1.2218,
      "step": 117350
    },
    {
      "epoch": 3.505210043889768,
      "grad_norm": 6.202013969421387,
      "learning_rate": 0.00016645866547808846,
      "loss": 1.2541,
      "step": 117400
    },
    {
      "epoch": 3.5067028931418505,
      "grad_norm": 4.274670600891113,
      "learning_rate": 0.00016644437968216962,
      "loss": 1.257,
      "step": 117450
    },
    {
      "epoch": 3.508195742393933,
      "grad_norm": 4.702452182769775,
      "learning_rate": 0.0001664300938862508,
      "loss": 1.2419,
      "step": 117500
    },
    {
      "epoch": 3.5096885916460154,
      "grad_norm": 2.897954225540161,
      "learning_rate": 0.00016641580809033195,
      "loss": 1.2716,
      "step": 117550
    },
    {
      "epoch": 3.511181440898098,
      "grad_norm": 4.723432540893555,
      "learning_rate": 0.00016640152229441312,
      "loss": 1.2357,
      "step": 117600
    },
    {
      "epoch": 3.512674290150181,
      "grad_norm": 4.234017848968506,
      "learning_rate": 0.00016638723649849428,
      "loss": 1.239,
      "step": 117650
    },
    {
      "epoch": 3.514167139402263,
      "grad_norm": 6.463681221008301,
      "learning_rate": 0.00016637295070257545,
      "loss": 1.2579,
      "step": 117700
    },
    {
      "epoch": 3.515659988654346,
      "grad_norm": 5.568453311920166,
      "learning_rate": 0.0001663586649066566,
      "loss": 1.2486,
      "step": 117750
    },
    {
      "epoch": 3.517152837906428,
      "grad_norm": 4.636147975921631,
      "learning_rate": 0.0001663443791107378,
      "loss": 1.2411,
      "step": 117800
    },
    {
      "epoch": 3.518645687158511,
      "grad_norm": 5.019303798675537,
      "learning_rate": 0.00016633009331481894,
      "loss": 1.2608,
      "step": 117850
    },
    {
      "epoch": 3.5201385364105935,
      "grad_norm": 5.414350986480713,
      "learning_rate": 0.00016631580751890013,
      "loss": 1.2642,
      "step": 117900
    },
    {
      "epoch": 3.5216313856626758,
      "grad_norm": 3.891533136367798,
      "learning_rate": 0.00016630152172298127,
      "loss": 1.2433,
      "step": 117950
    },
    {
      "epoch": 3.5231242349147585,
      "grad_norm": 3.9345791339874268,
      "learning_rate": 0.00016628723592706246,
      "loss": 1.2654,
      "step": 118000
    },
    {
      "epoch": 3.5246170841668407,
      "grad_norm": 3.8173060417175293,
      "learning_rate": 0.0001662729501311436,
      "loss": 1.2531,
      "step": 118050
    },
    {
      "epoch": 3.5261099334189234,
      "grad_norm": 4.534266471862793,
      "learning_rate": 0.0001662586643352248,
      "loss": 1.2924,
      "step": 118100
    },
    {
      "epoch": 3.527602782671006,
      "grad_norm": 5.261129856109619,
      "learning_rate": 0.00016624437853930596,
      "loss": 1.2433,
      "step": 118150
    },
    {
      "epoch": 3.5290956319230884,
      "grad_norm": 4.014307022094727,
      "learning_rate": 0.00016623009274338712,
      "loss": 1.2501,
      "step": 118200
    },
    {
      "epoch": 3.5305884811751707,
      "grad_norm": 4.939244270324707,
      "learning_rate": 0.00016621580694746828,
      "loss": 1.2156,
      "step": 118250
    },
    {
      "epoch": 3.5320813304272534,
      "grad_norm": 7.338710308074951,
      "learning_rate": 0.00016620152115154945,
      "loss": 1.2571,
      "step": 118300
    },
    {
      "epoch": 3.533574179679336,
      "grad_norm": 4.462497711181641,
      "learning_rate": 0.00016618723535563061,
      "loss": 1.2532,
      "step": 118350
    },
    {
      "epoch": 3.5350670289314188,
      "grad_norm": 3.8068175315856934,
      "learning_rate": 0.00016617294955971178,
      "loss": 1.2707,
      "step": 118400
    },
    {
      "epoch": 3.536559878183501,
      "grad_norm": 4.011793613433838,
      "learning_rate": 0.00016615866376379294,
      "loss": 1.2012,
      "step": 118450
    },
    {
      "epoch": 3.5380527274355833,
      "grad_norm": 4.955504894256592,
      "learning_rate": 0.0001661443779678741,
      "loss": 1.2312,
      "step": 118500
    },
    {
      "epoch": 3.539545576687666,
      "grad_norm": 5.355716228485107,
      "learning_rate": 0.00016613009217195527,
      "loss": 1.2677,
      "step": 118550
    },
    {
      "epoch": 3.5410384259397487,
      "grad_norm": 3.9078450202941895,
      "learning_rate": 0.00016611580637603644,
      "loss": 1.2216,
      "step": 118600
    },
    {
      "epoch": 3.542531275191831,
      "grad_norm": 4.30654764175415,
      "learning_rate": 0.0001661015205801176,
      "loss": 1.2864,
      "step": 118650
    },
    {
      "epoch": 3.5440241244439137,
      "grad_norm": 5.134391784667969,
      "learning_rate": 0.00016608723478419877,
      "loss": 1.2577,
      "step": 118700
    },
    {
      "epoch": 3.545516973695996,
      "grad_norm": 3.4655020236968994,
      "learning_rate": 0.00016607294898827993,
      "loss": 1.2427,
      "step": 118750
    },
    {
      "epoch": 3.5470098229480787,
      "grad_norm": 4.221212863922119,
      "learning_rate": 0.0001660586631923611,
      "loss": 1.243,
      "step": 118800
    },
    {
      "epoch": 3.5485026722001614,
      "grad_norm": 5.412088394165039,
      "learning_rate": 0.0001660443773964423,
      "loss": 1.2377,
      "step": 118850
    },
    {
      "epoch": 3.5499955214522436,
      "grad_norm": 5.42820930480957,
      "learning_rate": 0.00016603009160052343,
      "loss": 1.245,
      "step": 118900
    },
    {
      "epoch": 3.5514883707043263,
      "grad_norm": 4.585496425628662,
      "learning_rate": 0.00016601580580460462,
      "loss": 1.2885,
      "step": 118950
    },
    {
      "epoch": 3.5529812199564086,
      "grad_norm": 4.376836776733398,
      "learning_rate": 0.00016600152000868575,
      "loss": 1.2915,
      "step": 119000
    },
    {
      "epoch": 3.5544740692084913,
      "grad_norm": 5.338039875030518,
      "learning_rate": 0.00016598723421276695,
      "loss": 1.1785,
      "step": 119050
    },
    {
      "epoch": 3.555966918460574,
      "grad_norm": 4.422656059265137,
      "learning_rate": 0.00016597294841684808,
      "loss": 1.235,
      "step": 119100
    },
    {
      "epoch": 3.5574597677126563,
      "grad_norm": 5.377553939819336,
      "learning_rate": 0.00016595866262092928,
      "loss": 1.2796,
      "step": 119150
    },
    {
      "epoch": 3.558952616964739,
      "grad_norm": 4.3863959312438965,
      "learning_rate": 0.00016594437682501044,
      "loss": 1.2497,
      "step": 119200
    },
    {
      "epoch": 3.5604454662168212,
      "grad_norm": 5.522170066833496,
      "learning_rate": 0.0001659300910290916,
      "loss": 1.2899,
      "step": 119250
    },
    {
      "epoch": 3.561938315468904,
      "grad_norm": 3.6835429668426514,
      "learning_rate": 0.00016591580523317277,
      "loss": 1.2272,
      "step": 119300
    },
    {
      "epoch": 3.5634311647209866,
      "grad_norm": 3.4317328929901123,
      "learning_rate": 0.00016590151943725393,
      "loss": 1.2315,
      "step": 119350
    },
    {
      "epoch": 3.564924013973069,
      "grad_norm": 4.688202381134033,
      "learning_rate": 0.0001658872336413351,
      "loss": 1.2938,
      "step": 119400
    },
    {
      "epoch": 3.5664168632251516,
      "grad_norm": 3.5351858139038086,
      "learning_rate": 0.00016587294784541626,
      "loss": 1.2247,
      "step": 119450
    },
    {
      "epoch": 3.567909712477234,
      "grad_norm": 4.518115043640137,
      "learning_rate": 0.00016585866204949743,
      "loss": 1.3105,
      "step": 119500
    },
    {
      "epoch": 3.5694025617293166,
      "grad_norm": 3.226865530014038,
      "learning_rate": 0.0001658443762535786,
      "loss": 1.2498,
      "step": 119550
    },
    {
      "epoch": 3.5708954109813993,
      "grad_norm": 5.850762367248535,
      "learning_rate": 0.00016583009045765976,
      "loss": 1.2802,
      "step": 119600
    },
    {
      "epoch": 3.5723882602334815,
      "grad_norm": 6.21754789352417,
      "learning_rate": 0.00016581580466174095,
      "loss": 1.23,
      "step": 119650
    },
    {
      "epoch": 3.5738811094855643,
      "grad_norm": 5.319993019104004,
      "learning_rate": 0.0001658015188658221,
      "loss": 1.3102,
      "step": 119700
    },
    {
      "epoch": 3.5753739587376465,
      "grad_norm": 4.379292964935303,
      "learning_rate": 0.00016578723306990328,
      "loss": 1.203,
      "step": 119750
    },
    {
      "epoch": 3.5768668079897292,
      "grad_norm": 3.9974188804626465,
      "learning_rate": 0.00016577294727398442,
      "loss": 1.29,
      "step": 119800
    },
    {
      "epoch": 3.578359657241812,
      "grad_norm": 4.0016188621521,
      "learning_rate": 0.0001657586614780656,
      "loss": 1.2131,
      "step": 119850
    },
    {
      "epoch": 3.579852506493894,
      "grad_norm": 5.10989236831665,
      "learning_rate": 0.00016574437568214675,
      "loss": 1.2519,
      "step": 119900
    },
    {
      "epoch": 3.581345355745977,
      "grad_norm": 3.3042585849761963,
      "learning_rate": 0.00016573008988622794,
      "loss": 1.2462,
      "step": 119950
    },
    {
      "epoch": 3.582838204998059,
      "grad_norm": 6.814448356628418,
      "learning_rate": 0.0001657158040903091,
      "loss": 1.2539,
      "step": 120000
    },
    {
      "epoch": 3.584331054250142,
      "grad_norm": 6.447416305541992,
      "learning_rate": 0.00016570151829439027,
      "loss": 1.2407,
      "step": 120050
    },
    {
      "epoch": 3.5858239035022246,
      "grad_norm": 4.890771389007568,
      "learning_rate": 0.00016568723249847143,
      "loss": 1.282,
      "step": 120100
    },
    {
      "epoch": 3.587316752754307,
      "grad_norm": 3.6665565967559814,
      "learning_rate": 0.0001656729467025526,
      "loss": 1.269,
      "step": 120150
    },
    {
      "epoch": 3.5888096020063895,
      "grad_norm": 3.8311381340026855,
      "learning_rate": 0.00016565866090663376,
      "loss": 1.2546,
      "step": 120200
    },
    {
      "epoch": 3.590302451258472,
      "grad_norm": 3.9717676639556885,
      "learning_rate": 0.00016564437511071493,
      "loss": 1.2402,
      "step": 120250
    },
    {
      "epoch": 3.5917953005105545,
      "grad_norm": 5.8756585121154785,
      "learning_rate": 0.0001656300893147961,
      "loss": 1.2934,
      "step": 120300
    },
    {
      "epoch": 3.593288149762637,
      "grad_norm": 4.234692573547363,
      "learning_rate": 0.00016561580351887725,
      "loss": 1.2773,
      "step": 120350
    },
    {
      "epoch": 3.5947809990147195,
      "grad_norm": 5.237219333648682,
      "learning_rate": 0.00016560151772295842,
      "loss": 1.2638,
      "step": 120400
    },
    {
      "epoch": 3.5962738482668017,
      "grad_norm": 3.977701187133789,
      "learning_rate": 0.0001655872319270396,
      "loss": 1.2452,
      "step": 120450
    },
    {
      "epoch": 3.5977666975188844,
      "grad_norm": 4.238033294677734,
      "learning_rate": 0.00016557294613112075,
      "loss": 1.2813,
      "step": 120500
    },
    {
      "epoch": 3.599259546770967,
      "grad_norm": 5.174859523773193,
      "learning_rate": 0.00016555866033520194,
      "loss": 1.2477,
      "step": 120550
    },
    {
      "epoch": 3.60075239602305,
      "grad_norm": 4.832043647766113,
      "learning_rate": 0.00016554437453928308,
      "loss": 1.2081,
      "step": 120600
    },
    {
      "epoch": 3.602245245275132,
      "grad_norm": 5.991222858428955,
      "learning_rate": 0.00016553008874336427,
      "loss": 1.2275,
      "step": 120650
    },
    {
      "epoch": 3.6037380945272144,
      "grad_norm": 5.4362101554870605,
      "learning_rate": 0.0001655158029474454,
      "loss": 1.262,
      "step": 120700
    },
    {
      "epoch": 3.605230943779297,
      "grad_norm": 5.659782409667969,
      "learning_rate": 0.0001655015171515266,
      "loss": 1.3115,
      "step": 120750
    },
    {
      "epoch": 3.60672379303138,
      "grad_norm": 4.6319146156311035,
      "learning_rate": 0.00016548723135560776,
      "loss": 1.3261,
      "step": 120800
    },
    {
      "epoch": 3.608216642283462,
      "grad_norm": 4.129281520843506,
      "learning_rate": 0.00016547294555968893,
      "loss": 1.3001,
      "step": 120850
    },
    {
      "epoch": 3.6097094915355448,
      "grad_norm": 4.629218101501465,
      "learning_rate": 0.0001654586597637701,
      "loss": 1.2856,
      "step": 120900
    },
    {
      "epoch": 3.611202340787627,
      "grad_norm": 4.741670608520508,
      "learning_rate": 0.00016544437396785126,
      "loss": 1.2671,
      "step": 120950
    },
    {
      "epoch": 3.6126951900397097,
      "grad_norm": 3.8376450538635254,
      "learning_rate": 0.00016543008817193242,
      "loss": 1.1981,
      "step": 121000
    },
    {
      "epoch": 3.6141880392917924,
      "grad_norm": 5.130064010620117,
      "learning_rate": 0.0001654158023760136,
      "loss": 1.2174,
      "step": 121050
    },
    {
      "epoch": 3.6156808885438747,
      "grad_norm": 4.116199970245361,
      "learning_rate": 0.00016540151658009475,
      "loss": 1.2786,
      "step": 121100
    },
    {
      "epoch": 3.6171737377959574,
      "grad_norm": 3.8612489700317383,
      "learning_rate": 0.00016538723078417592,
      "loss": 1.2788,
      "step": 121150
    },
    {
      "epoch": 3.6186665870480397,
      "grad_norm": 4.064988136291504,
      "learning_rate": 0.00016537294498825708,
      "loss": 1.2465,
      "step": 121200
    },
    {
      "epoch": 3.6201594363001224,
      "grad_norm": 4.097548007965088,
      "learning_rate": 0.00016535865919233825,
      "loss": 1.2659,
      "step": 121250
    },
    {
      "epoch": 3.621652285552205,
      "grad_norm": 4.937625885009766,
      "learning_rate": 0.0001653443733964194,
      "loss": 1.2071,
      "step": 121300
    },
    {
      "epoch": 3.6231451348042873,
      "grad_norm": 4.893852710723877,
      "learning_rate": 0.00016533008760050057,
      "loss": 1.2962,
      "step": 121350
    },
    {
      "epoch": 3.62463798405637,
      "grad_norm": 5.306116580963135,
      "learning_rate": 0.00016531580180458174,
      "loss": 1.27,
      "step": 121400
    },
    {
      "epoch": 3.6261308333084523,
      "grad_norm": 4.8264360427856445,
      "learning_rate": 0.0001653015160086629,
      "loss": 1.2717,
      "step": 121450
    },
    {
      "epoch": 3.627623682560535,
      "grad_norm": 4.300417423248291,
      "learning_rate": 0.00016528723021274407,
      "loss": 1.3015,
      "step": 121500
    },
    {
      "epoch": 3.6291165318126177,
      "grad_norm": 4.886301517486572,
      "learning_rate": 0.00016527294441682523,
      "loss": 1.2295,
      "step": 121550
    },
    {
      "epoch": 3.6306093810647,
      "grad_norm": 5.038539409637451,
      "learning_rate": 0.00016525865862090643,
      "loss": 1.2542,
      "step": 121600
    },
    {
      "epoch": 3.6321022303167827,
      "grad_norm": 4.360227584838867,
      "learning_rate": 0.00016524437282498756,
      "loss": 1.2712,
      "step": 121650
    },
    {
      "epoch": 3.633595079568865,
      "grad_norm": 4.984081745147705,
      "learning_rate": 0.00016523008702906875,
      "loss": 1.2949,
      "step": 121700
    },
    {
      "epoch": 3.6350879288209477,
      "grad_norm": 5.2939229011535645,
      "learning_rate": 0.0001652158012331499,
      "loss": 1.2393,
      "step": 121750
    },
    {
      "epoch": 3.6365807780730304,
      "grad_norm": 5.242954730987549,
      "learning_rate": 0.00016520151543723108,
      "loss": 1.1846,
      "step": 121800
    },
    {
      "epoch": 3.6380736273251126,
      "grad_norm": 4.06920862197876,
      "learning_rate": 0.00016518722964131225,
      "loss": 1.2599,
      "step": 121850
    },
    {
      "epoch": 3.6395664765771953,
      "grad_norm": 5.329120635986328,
      "learning_rate": 0.0001651729438453934,
      "loss": 1.2407,
      "step": 121900
    },
    {
      "epoch": 3.6410593258292776,
      "grad_norm": 5.278786659240723,
      "learning_rate": 0.00016515865804947458,
      "loss": 1.2645,
      "step": 121950
    },
    {
      "epoch": 3.6425521750813603,
      "grad_norm": 3.5512640476226807,
      "learning_rate": 0.00016514437225355574,
      "loss": 1.2492,
      "step": 122000
    },
    {
      "epoch": 3.644045024333443,
      "grad_norm": 3.9851057529449463,
      "learning_rate": 0.0001651300864576369,
      "loss": 1.27,
      "step": 122050
    },
    {
      "epoch": 3.6455378735855253,
      "grad_norm": 2.720550298690796,
      "learning_rate": 0.00016511580066171807,
      "loss": 1.2387,
      "step": 122100
    },
    {
      "epoch": 3.647030722837608,
      "grad_norm": 4.045743465423584,
      "learning_rate": 0.00016510151486579924,
      "loss": 1.2768,
      "step": 122150
    },
    {
      "epoch": 3.6485235720896902,
      "grad_norm": 4.347272872924805,
      "learning_rate": 0.0001650872290698804,
      "loss": 1.2114,
      "step": 122200
    },
    {
      "epoch": 3.650016421341773,
      "grad_norm": 4.8486151695251465,
      "learning_rate": 0.00016507294327396157,
      "loss": 1.2761,
      "step": 122250
    },
    {
      "epoch": 3.6515092705938557,
      "grad_norm": 4.846822261810303,
      "learning_rate": 0.00016505865747804276,
      "loss": 1.2563,
      "step": 122300
    },
    {
      "epoch": 3.653002119845938,
      "grad_norm": 5.770145893096924,
      "learning_rate": 0.0001650443716821239,
      "loss": 1.2244,
      "step": 122350
    },
    {
      "epoch": 3.6544949690980206,
      "grad_norm": 4.940767765045166,
      "learning_rate": 0.0001650300858862051,
      "loss": 1.2658,
      "step": 122400
    },
    {
      "epoch": 3.655987818350103,
      "grad_norm": 4.317561626434326,
      "learning_rate": 0.00016501580009028622,
      "loss": 1.2454,
      "step": 122450
    },
    {
      "epoch": 3.6574806676021856,
      "grad_norm": 4.162029266357422,
      "learning_rate": 0.00016500151429436742,
      "loss": 1.2757,
      "step": 122500
    },
    {
      "epoch": 3.6589735168542683,
      "grad_norm": 3.620096445083618,
      "learning_rate": 0.00016498722849844855,
      "loss": 1.2615,
      "step": 122550
    },
    {
      "epoch": 3.6604663661063506,
      "grad_norm": 4.369839191436768,
      "learning_rate": 0.00016497294270252975,
      "loss": 1.2435,
      "step": 122600
    },
    {
      "epoch": 3.661959215358433,
      "grad_norm": 5.848557472229004,
      "learning_rate": 0.0001649586569066109,
      "loss": 1.2422,
      "step": 122650
    },
    {
      "epoch": 3.6634520646105155,
      "grad_norm": 4.413859844207764,
      "learning_rate": 0.00016494437111069207,
      "loss": 1.2505,
      "step": 122700
    },
    {
      "epoch": 3.6649449138625982,
      "grad_norm": 4.773179531097412,
      "learning_rate": 0.00016493008531477324,
      "loss": 1.2856,
      "step": 122750
    },
    {
      "epoch": 3.666437763114681,
      "grad_norm": 6.004321098327637,
      "learning_rate": 0.0001649157995188544,
      "loss": 1.2183,
      "step": 122800
    },
    {
      "epoch": 3.667930612366763,
      "grad_norm": 4.601461410522461,
      "learning_rate": 0.00016490151372293557,
      "loss": 1.2328,
      "step": 122850
    },
    {
      "epoch": 3.6694234616188455,
      "grad_norm": 4.456672668457031,
      "learning_rate": 0.00016488722792701673,
      "loss": 1.2269,
      "step": 122900
    },
    {
      "epoch": 3.670916310870928,
      "grad_norm": 4.381960391998291,
      "learning_rate": 0.0001648729421310979,
      "loss": 1.2971,
      "step": 122950
    },
    {
      "epoch": 3.672409160123011,
      "grad_norm": 4.233649730682373,
      "learning_rate": 0.00016485865633517906,
      "loss": 1.2597,
      "step": 123000
    },
    {
      "epoch": 3.673902009375093,
      "grad_norm": 3.898301601409912,
      "learning_rate": 0.00016484437053926023,
      "loss": 1.2237,
      "step": 123050
    },
    {
      "epoch": 3.675394858627176,
      "grad_norm": 4.5442585945129395,
      "learning_rate": 0.00016483008474334142,
      "loss": 1.2503,
      "step": 123100
    },
    {
      "epoch": 3.676887707879258,
      "grad_norm": 6.81040620803833,
      "learning_rate": 0.00016481579894742256,
      "loss": 1.2685,
      "step": 123150
    },
    {
      "epoch": 3.678380557131341,
      "grad_norm": 3.5733816623687744,
      "learning_rate": 0.00016480151315150375,
      "loss": 1.2621,
      "step": 123200
    },
    {
      "epoch": 3.6798734063834235,
      "grad_norm": 5.7314252853393555,
      "learning_rate": 0.00016478722735558489,
      "loss": 1.2356,
      "step": 123250
    },
    {
      "epoch": 3.681366255635506,
      "grad_norm": 3.6596696376800537,
      "learning_rate": 0.00016477294155966608,
      "loss": 1.2131,
      "step": 123300
    },
    {
      "epoch": 3.6828591048875885,
      "grad_norm": 4.004091739654541,
      "learning_rate": 0.00016475865576374722,
      "loss": 1.2796,
      "step": 123350
    },
    {
      "epoch": 3.6843519541396708,
      "grad_norm": 4.247511863708496,
      "learning_rate": 0.0001647443699678284,
      "loss": 1.2252,
      "step": 123400
    },
    {
      "epoch": 3.6858448033917535,
      "grad_norm": 4.852141380310059,
      "learning_rate": 0.00016473008417190957,
      "loss": 1.2485,
      "step": 123450
    },
    {
      "epoch": 3.687337652643836,
      "grad_norm": 4.641091346740723,
      "learning_rate": 0.00016471579837599074,
      "loss": 1.2414,
      "step": 123500
    },
    {
      "epoch": 3.6888305018959184,
      "grad_norm": 3.5827832221984863,
      "learning_rate": 0.0001647015125800719,
      "loss": 1.2304,
      "step": 123550
    },
    {
      "epoch": 3.690323351148001,
      "grad_norm": 3.877167224884033,
      "learning_rate": 0.00016468722678415307,
      "loss": 1.2609,
      "step": 123600
    },
    {
      "epoch": 3.6918162004000834,
      "grad_norm": 4.917414665222168,
      "learning_rate": 0.00016467294098823423,
      "loss": 1.3357,
      "step": 123650
    },
    {
      "epoch": 3.693309049652166,
      "grad_norm": 5.0784406661987305,
      "learning_rate": 0.0001646586551923154,
      "loss": 1.2776,
      "step": 123700
    },
    {
      "epoch": 3.694801898904249,
      "grad_norm": 5.322165489196777,
      "learning_rate": 0.00016464436939639656,
      "loss": 1.265,
      "step": 123750
    },
    {
      "epoch": 3.696294748156331,
      "grad_norm": 4.516918182373047,
      "learning_rate": 0.00016463008360047772,
      "loss": 1.2559,
      "step": 123800
    },
    {
      "epoch": 3.697787597408414,
      "grad_norm": 4.489822864532471,
      "learning_rate": 0.0001646157978045589,
      "loss": 1.2206,
      "step": 123850
    },
    {
      "epoch": 3.699280446660496,
      "grad_norm": 5.039775848388672,
      "learning_rate": 0.00016460151200864005,
      "loss": 1.2285,
      "step": 123900
    },
    {
      "epoch": 3.7007732959125788,
      "grad_norm": 3.899019241333008,
      "learning_rate": 0.00016458722621272122,
      "loss": 1.2484,
      "step": 123950
    },
    {
      "epoch": 3.7022661451646615,
      "grad_norm": 5.370189189910889,
      "learning_rate": 0.00016457294041680238,
      "loss": 1.2901,
      "step": 124000
    },
    {
      "epoch": 3.7037589944167437,
      "grad_norm": 4.206950664520264,
      "learning_rate": 0.00016455865462088355,
      "loss": 1.2384,
      "step": 124050
    },
    {
      "epoch": 3.7052518436688264,
      "grad_norm": 6.357316970825195,
      "learning_rate": 0.0001645443688249647,
      "loss": 1.2222,
      "step": 124100
    },
    {
      "epoch": 3.7067446929209087,
      "grad_norm": 5.480586051940918,
      "learning_rate": 0.00016453008302904588,
      "loss": 1.2402,
      "step": 124150
    },
    {
      "epoch": 3.7082375421729914,
      "grad_norm": 4.471677303314209,
      "learning_rate": 0.00016451579723312704,
      "loss": 1.2962,
      "step": 124200
    },
    {
      "epoch": 3.709730391425074,
      "grad_norm": 4.873519420623779,
      "learning_rate": 0.00016450151143720823,
      "loss": 1.2438,
      "step": 124250
    },
    {
      "epoch": 3.7112232406771564,
      "grad_norm": 4.237666130065918,
      "learning_rate": 0.00016448722564128937,
      "loss": 1.3002,
      "step": 124300
    },
    {
      "epoch": 3.712716089929239,
      "grad_norm": 5.313760280609131,
      "learning_rate": 0.00016447293984537056,
      "loss": 1.217,
      "step": 124350
    },
    {
      "epoch": 3.7142089391813213,
      "grad_norm": 6.441144943237305,
      "learning_rate": 0.0001644586540494517,
      "loss": 1.3176,
      "step": 124400
    },
    {
      "epoch": 3.715701788433404,
      "grad_norm": 5.12811803817749,
      "learning_rate": 0.0001644443682535329,
      "loss": 1.2937,
      "step": 124450
    },
    {
      "epoch": 3.7171946376854867,
      "grad_norm": 4.984979152679443,
      "learning_rate": 0.00016443008245761406,
      "loss": 1.1999,
      "step": 124500
    },
    {
      "epoch": 3.718687486937569,
      "grad_norm": 4.063045501708984,
      "learning_rate": 0.00016441579666169522,
      "loss": 1.235,
      "step": 124550
    },
    {
      "epoch": 3.7201803361896517,
      "grad_norm": 3.6467928886413574,
      "learning_rate": 0.00016440151086577639,
      "loss": 1.1857,
      "step": 124600
    },
    {
      "epoch": 3.721673185441734,
      "grad_norm": 4.508729934692383,
      "learning_rate": 0.00016438722506985755,
      "loss": 1.3138,
      "step": 124650
    },
    {
      "epoch": 3.7231660346938167,
      "grad_norm": 5.27504825592041,
      "learning_rate": 0.00016437293927393872,
      "loss": 1.2026,
      "step": 124700
    },
    {
      "epoch": 3.7246588839458994,
      "grad_norm": 4.412220478057861,
      "learning_rate": 0.00016435865347801988,
      "loss": 1.2699,
      "step": 124750
    },
    {
      "epoch": 3.7261517331979817,
      "grad_norm": 5.649445533752441,
      "learning_rate": 0.00016434436768210104,
      "loss": 1.2803,
      "step": 124800
    },
    {
      "epoch": 3.7276445824500644,
      "grad_norm": 3.868227005004883,
      "learning_rate": 0.0001643300818861822,
      "loss": 1.2614,
      "step": 124850
    },
    {
      "epoch": 3.7291374317021466,
      "grad_norm": 4.377996444702148,
      "learning_rate": 0.00016431579609026337,
      "loss": 1.2548,
      "step": 124900
    },
    {
      "epoch": 3.7306302809542293,
      "grad_norm": 5.023475646972656,
      "learning_rate": 0.00016430151029434454,
      "loss": 1.2775,
      "step": 124950
    },
    {
      "epoch": 3.732123130206312,
      "grad_norm": 8.013720512390137,
      "learning_rate": 0.0001642872244984257,
      "loss": 1.2716,
      "step": 125000
    },
    {
      "epoch": 3.7336159794583943,
      "grad_norm": 4.947465419769287,
      "learning_rate": 0.0001642729387025069,
      "loss": 1.2681,
      "step": 125050
    },
    {
      "epoch": 3.7351088287104766,
      "grad_norm": 3.9547464847564697,
      "learning_rate": 0.00016425865290658803,
      "loss": 1.244,
      "step": 125100
    },
    {
      "epoch": 3.7366016779625593,
      "grad_norm": 5.331160545349121,
      "learning_rate": 0.00016424436711066922,
      "loss": 1.3086,
      "step": 125150
    },
    {
      "epoch": 3.738094527214642,
      "grad_norm": 3.980027914047241,
      "learning_rate": 0.00016423008131475036,
      "loss": 1.2461,
      "step": 125200
    },
    {
      "epoch": 3.7395873764667242,
      "grad_norm": 3.4833130836486816,
      "learning_rate": 0.00016421579551883155,
      "loss": 1.2107,
      "step": 125250
    },
    {
      "epoch": 3.741080225718807,
      "grad_norm": 4.746307373046875,
      "learning_rate": 0.00016420150972291272,
      "loss": 1.2875,
      "step": 125300
    },
    {
      "epoch": 3.742573074970889,
      "grad_norm": 3.9247684478759766,
      "learning_rate": 0.00016418722392699388,
      "loss": 1.2185,
      "step": 125350
    },
    {
      "epoch": 3.744065924222972,
      "grad_norm": 5.135631084442139,
      "learning_rate": 0.00016417293813107505,
      "loss": 1.2512,
      "step": 125400
    },
    {
      "epoch": 3.7455587734750546,
      "grad_norm": 4.875301361083984,
      "learning_rate": 0.0001641586523351562,
      "loss": 1.2794,
      "step": 125450
    },
    {
      "epoch": 3.747051622727137,
      "grad_norm": 5.335410118103027,
      "learning_rate": 0.00016414436653923738,
      "loss": 1.2617,
      "step": 125500
    },
    {
      "epoch": 3.7485444719792196,
      "grad_norm": 4.500457286834717,
      "learning_rate": 0.00016413008074331854,
      "loss": 1.2482,
      "step": 125550
    },
    {
      "epoch": 3.750037321231302,
      "grad_norm": 4.4383087158203125,
      "learning_rate": 0.0001641157949473997,
      "loss": 1.2534,
      "step": 125600
    },
    {
      "epoch": 3.7515301704833846,
      "grad_norm": 4.272806644439697,
      "learning_rate": 0.00016410150915148087,
      "loss": 1.2564,
      "step": 125650
    },
    {
      "epoch": 3.7530230197354673,
      "grad_norm": 6.498635292053223,
      "learning_rate": 0.00016408722335556204,
      "loss": 1.2422,
      "step": 125700
    },
    {
      "epoch": 3.7545158689875495,
      "grad_norm": 4.720774173736572,
      "learning_rate": 0.00016407293755964323,
      "loss": 1.2543,
      "step": 125750
    },
    {
      "epoch": 3.7560087182396322,
      "grad_norm": 4.762580871582031,
      "learning_rate": 0.00016405865176372436,
      "loss": 1.2453,
      "step": 125800
    },
    {
      "epoch": 3.7575015674917145,
      "grad_norm": 3.98309326171875,
      "learning_rate": 0.00016404436596780556,
      "loss": 1.2295,
      "step": 125850
    },
    {
      "epoch": 3.758994416743797,
      "grad_norm": 5.4489336013793945,
      "learning_rate": 0.0001640300801718867,
      "loss": 1.2466,
      "step": 125900
    },
    {
      "epoch": 3.76048726599588,
      "grad_norm": 5.064475059509277,
      "learning_rate": 0.00016401579437596789,
      "loss": 1.2451,
      "step": 125950
    },
    {
      "epoch": 3.761980115247962,
      "grad_norm": 4.22551155090332,
      "learning_rate": 0.00016400150858004902,
      "loss": 1.2346,
      "step": 126000
    },
    {
      "epoch": 3.763472964500045,
      "grad_norm": 4.364709854125977,
      "learning_rate": 0.00016398722278413021,
      "loss": 1.2894,
      "step": 126050
    },
    {
      "epoch": 3.764965813752127,
      "grad_norm": 4.05735969543457,
      "learning_rate": 0.00016397293698821138,
      "loss": 1.1943,
      "step": 126100
    },
    {
      "epoch": 3.76645866300421,
      "grad_norm": 4.884429931640625,
      "learning_rate": 0.00016395865119229254,
      "loss": 1.2587,
      "step": 126150
    },
    {
      "epoch": 3.7679515122562925,
      "grad_norm": 4.451826572418213,
      "learning_rate": 0.0001639443653963737,
      "loss": 1.2856,
      "step": 126200
    },
    {
      "epoch": 3.769444361508375,
      "grad_norm": 5.6031646728515625,
      "learning_rate": 0.00016393007960045487,
      "loss": 1.2705,
      "step": 126250
    },
    {
      "epoch": 3.7709372107604575,
      "grad_norm": 4.248857498168945,
      "learning_rate": 0.00016391579380453604,
      "loss": 1.3043,
      "step": 126300
    },
    {
      "epoch": 3.7724300600125398,
      "grad_norm": 4.71324348449707,
      "learning_rate": 0.0001639015080086172,
      "loss": 1.2658,
      "step": 126350
    },
    {
      "epoch": 3.7739229092646225,
      "grad_norm": 5.27174186706543,
      "learning_rate": 0.00016388722221269837,
      "loss": 1.2325,
      "step": 126400
    },
    {
      "epoch": 3.775415758516705,
      "grad_norm": 4.696889877319336,
      "learning_rate": 0.00016387293641677953,
      "loss": 1.2819,
      "step": 126450
    },
    {
      "epoch": 3.7769086077687875,
      "grad_norm": 5.210821151733398,
      "learning_rate": 0.0001638586506208607,
      "loss": 1.2733,
      "step": 126500
    },
    {
      "epoch": 3.77840145702087,
      "grad_norm": 4.370712757110596,
      "learning_rate": 0.00016384436482494186,
      "loss": 1.2121,
      "step": 126550
    },
    {
      "epoch": 3.7798943062729524,
      "grad_norm": 3.9094676971435547,
      "learning_rate": 0.00016383007902902303,
      "loss": 1.2744,
      "step": 126600
    },
    {
      "epoch": 3.781387155525035,
      "grad_norm": 4.375615119934082,
      "learning_rate": 0.0001638157932331042,
      "loss": 1.2808,
      "step": 126650
    },
    {
      "epoch": 3.782880004777118,
      "grad_norm": 4.995912075042725,
      "learning_rate": 0.00016380150743718536,
      "loss": 1.3116,
      "step": 126700
    },
    {
      "epoch": 3.7843728540292,
      "grad_norm": 3.9363934993743896,
      "learning_rate": 0.00016378722164126652,
      "loss": 1.2554,
      "step": 126750
    },
    {
      "epoch": 3.785865703281283,
      "grad_norm": 3.885279893875122,
      "learning_rate": 0.00016377293584534768,
      "loss": 1.2099,
      "step": 126800
    },
    {
      "epoch": 3.787358552533365,
      "grad_norm": 4.745761871337891,
      "learning_rate": 0.00016375865004942885,
      "loss": 1.2336,
      "step": 126850
    },
    {
      "epoch": 3.7888514017854478,
      "grad_norm": 5.675398826599121,
      "learning_rate": 0.00016374436425351004,
      "loss": 1.2736,
      "step": 126900
    },
    {
      "epoch": 3.7903442510375305,
      "grad_norm": 5.933438301086426,
      "learning_rate": 0.00016373007845759118,
      "loss": 1.2245,
      "step": 126950
    },
    {
      "epoch": 3.7918371002896127,
      "grad_norm": 4.410390853881836,
      "learning_rate": 0.00016371579266167237,
      "loss": 1.2786,
      "step": 127000
    },
    {
      "epoch": 3.7933299495416954,
      "grad_norm": 3.837341785430908,
      "learning_rate": 0.0001637015068657535,
      "loss": 1.2154,
      "step": 127050
    },
    {
      "epoch": 3.7948227987937777,
      "grad_norm": 4.170938014984131,
      "learning_rate": 0.0001636872210698347,
      "loss": 1.1933,
      "step": 127100
    },
    {
      "epoch": 3.7963156480458604,
      "grad_norm": 5.0940985679626465,
      "learning_rate": 0.00016367293527391584,
      "loss": 1.2399,
      "step": 127150
    },
    {
      "epoch": 3.797808497297943,
      "grad_norm": 3.5157582759857178,
      "learning_rate": 0.00016365864947799703,
      "loss": 1.2217,
      "step": 127200
    },
    {
      "epoch": 3.7993013465500254,
      "grad_norm": 4.230498790740967,
      "learning_rate": 0.0001636443636820782,
      "loss": 1.2665,
      "step": 127250
    },
    {
      "epoch": 3.8007941958021076,
      "grad_norm": 4.24791955947876,
      "learning_rate": 0.00016363007788615936,
      "loss": 1.28,
      "step": 127300
    },
    {
      "epoch": 3.8022870450541904,
      "grad_norm": 4.877384185791016,
      "learning_rate": 0.00016361579209024052,
      "loss": 1.3076,
      "step": 127350
    },
    {
      "epoch": 3.803779894306273,
      "grad_norm": 6.252450942993164,
      "learning_rate": 0.0001636015062943217,
      "loss": 1.2648,
      "step": 127400
    },
    {
      "epoch": 3.8052727435583553,
      "grad_norm": 4.372878551483154,
      "learning_rate": 0.00016358722049840285,
      "loss": 1.2749,
      "step": 127450
    },
    {
      "epoch": 3.806765592810438,
      "grad_norm": 5.042996883392334,
      "learning_rate": 0.00016357293470248402,
      "loss": 1.2379,
      "step": 127500
    },
    {
      "epoch": 3.8082584420625203,
      "grad_norm": 4.833359241485596,
      "learning_rate": 0.00016355864890656518,
      "loss": 1.2456,
      "step": 127550
    },
    {
      "epoch": 3.809751291314603,
      "grad_norm": 3.7556586265563965,
      "learning_rate": 0.00016354436311064635,
      "loss": 1.2998,
      "step": 127600
    },
    {
      "epoch": 3.8112441405666857,
      "grad_norm": 4.22800874710083,
      "learning_rate": 0.0001635300773147275,
      "loss": 1.2307,
      "step": 127650
    },
    {
      "epoch": 3.812736989818768,
      "grad_norm": 3.8682031631469727,
      "learning_rate": 0.0001635157915188087,
      "loss": 1.2429,
      "step": 127700
    },
    {
      "epoch": 3.8142298390708507,
      "grad_norm": 5.100236415863037,
      "learning_rate": 0.00016350150572288984,
      "loss": 1.2725,
      "step": 127750
    },
    {
      "epoch": 3.815722688322933,
      "grad_norm": 5.069801330566406,
      "learning_rate": 0.00016348721992697103,
      "loss": 1.3199,
      "step": 127800
    },
    {
      "epoch": 3.8172155375750156,
      "grad_norm": 4.382961273193359,
      "learning_rate": 0.00016347293413105217,
      "loss": 1.2727,
      "step": 127850
    },
    {
      "epoch": 3.8187083868270983,
      "grad_norm": 4.622864246368408,
      "learning_rate": 0.00016345864833513336,
      "loss": 1.3309,
      "step": 127900
    },
    {
      "epoch": 3.8202012360791806,
      "grad_norm": 5.895843982696533,
      "learning_rate": 0.00016344436253921453,
      "loss": 1.2453,
      "step": 127950
    },
    {
      "epoch": 3.8216940853312633,
      "grad_norm": 4.68233060836792,
      "learning_rate": 0.0001634300767432957,
      "loss": 1.2948,
      "step": 128000
    },
    {
      "epoch": 3.8231869345833456,
      "grad_norm": 4.286251068115234,
      "learning_rate": 0.00016341579094737686,
      "loss": 1.2758,
      "step": 128050
    },
    {
      "epoch": 3.8246797838354283,
      "grad_norm": 3.9140446186065674,
      "learning_rate": 0.00016340150515145802,
      "loss": 1.2791,
      "step": 128100
    },
    {
      "epoch": 3.826172633087511,
      "grad_norm": 5.249776840209961,
      "learning_rate": 0.00016338721935553918,
      "loss": 1.3379,
      "step": 128150
    },
    {
      "epoch": 3.8276654823395933,
      "grad_norm": 4.369449138641357,
      "learning_rate": 0.00016337293355962035,
      "loss": 1.179,
      "step": 128200
    },
    {
      "epoch": 3.829158331591676,
      "grad_norm": 5.655234336853027,
      "learning_rate": 0.00016335864776370151,
      "loss": 1.2535,
      "step": 128250
    },
    {
      "epoch": 3.830651180843758,
      "grad_norm": 5.233351230621338,
      "learning_rate": 0.00016334436196778268,
      "loss": 1.2601,
      "step": 128300
    },
    {
      "epoch": 3.832144030095841,
      "grad_norm": 4.451437950134277,
      "learning_rate": 0.00016333007617186384,
      "loss": 1.2733,
      "step": 128350
    },
    {
      "epoch": 3.8336368793479236,
      "grad_norm": 3.807561159133911,
      "learning_rate": 0.000163315790375945,
      "loss": 1.2721,
      "step": 128400
    },
    {
      "epoch": 3.835129728600006,
      "grad_norm": 4.507883071899414,
      "learning_rate": 0.00016330150458002617,
      "loss": 1.2498,
      "step": 128450
    },
    {
      "epoch": 3.8366225778520886,
      "grad_norm": 4.246575832366943,
      "learning_rate": 0.00016328721878410736,
      "loss": 1.2299,
      "step": 128500
    },
    {
      "epoch": 3.838115427104171,
      "grad_norm": 5.045656204223633,
      "learning_rate": 0.0001632729329881885,
      "loss": 1.2536,
      "step": 128550
    },
    {
      "epoch": 3.8396082763562536,
      "grad_norm": 3.3326852321624756,
      "learning_rate": 0.0001632586471922697,
      "loss": 1.2401,
      "step": 128600
    },
    {
      "epoch": 3.8411011256083363,
      "grad_norm": 4.201671600341797,
      "learning_rate": 0.00016324436139635083,
      "loss": 1.2681,
      "step": 128650
    },
    {
      "epoch": 3.8425939748604185,
      "grad_norm": 3.979217052459717,
      "learning_rate": 0.00016323007560043202,
      "loss": 1.2901,
      "step": 128700
    },
    {
      "epoch": 3.8440868241125012,
      "grad_norm": 4.908186435699463,
      "learning_rate": 0.0001632157898045132,
      "loss": 1.2443,
      "step": 128750
    },
    {
      "epoch": 3.8455796733645835,
      "grad_norm": 7.0551252365112305,
      "learning_rate": 0.00016320150400859435,
      "loss": 1.3312,
      "step": 128800
    },
    {
      "epoch": 3.847072522616666,
      "grad_norm": 2.9734935760498047,
      "learning_rate": 0.00016318721821267552,
      "loss": 1.1726,
      "step": 128850
    },
    {
      "epoch": 3.848565371868749,
      "grad_norm": 8.724651336669922,
      "learning_rate": 0.00016317293241675668,
      "loss": 1.3131,
      "step": 128900
    },
    {
      "epoch": 3.850058221120831,
      "grad_norm": 4.633723735809326,
      "learning_rate": 0.00016315864662083785,
      "loss": 1.2309,
      "step": 128950
    },
    {
      "epoch": 3.851551070372914,
      "grad_norm": 5.132857799530029,
      "learning_rate": 0.000163144360824919,
      "loss": 1.1988,
      "step": 129000
    },
    {
      "epoch": 3.853043919624996,
      "grad_norm": 3.8106892108917236,
      "learning_rate": 0.00016313007502900018,
      "loss": 1.2284,
      "step": 129050
    },
    {
      "epoch": 3.854536768877079,
      "grad_norm": 3.9938535690307617,
      "learning_rate": 0.00016311578923308134,
      "loss": 1.2252,
      "step": 129100
    },
    {
      "epoch": 3.8560296181291616,
      "grad_norm": 4.499520301818848,
      "learning_rate": 0.0001631015034371625,
      "loss": 1.2834,
      "step": 129150
    },
    {
      "epoch": 3.857522467381244,
      "grad_norm": 3.780081033706665,
      "learning_rate": 0.00016308721764124367,
      "loss": 1.1696,
      "step": 129200
    },
    {
      "epoch": 3.8590153166333265,
      "grad_norm": 5.509545803070068,
      "learning_rate": 0.00016307293184532483,
      "loss": 1.2129,
      "step": 129250
    },
    {
      "epoch": 3.860508165885409,
      "grad_norm": 3.87497615814209,
      "learning_rate": 0.000163058646049406,
      "loss": 1.2971,
      "step": 129300
    },
    {
      "epoch": 3.8620010151374915,
      "grad_norm": 4.839779853820801,
      "learning_rate": 0.00016304436025348716,
      "loss": 1.2385,
      "step": 129350
    },
    {
      "epoch": 3.863493864389574,
      "grad_norm": 7.801334857940674,
      "learning_rate": 0.00016303007445756833,
      "loss": 1.2482,
      "step": 129400
    },
    {
      "epoch": 3.8649867136416565,
      "grad_norm": 3.884004831314087,
      "learning_rate": 0.0001630157886616495,
      "loss": 1.264,
      "step": 129450
    },
    {
      "epoch": 3.8664795628937387,
      "grad_norm": 5.332859516143799,
      "learning_rate": 0.00016300150286573066,
      "loss": 1.2263,
      "step": 129500
    },
    {
      "epoch": 3.8679724121458214,
      "grad_norm": 4.195265293121338,
      "learning_rate": 0.00016298721706981185,
      "loss": 1.2579,
      "step": 129550
    },
    {
      "epoch": 3.869465261397904,
      "grad_norm": 4.393342018127441,
      "learning_rate": 0.000162972931273893,
      "loss": 1.2513,
      "step": 129600
    },
    {
      "epoch": 3.8709581106499864,
      "grad_norm": 5.231433391571045,
      "learning_rate": 0.00016295864547797418,
      "loss": 1.2508,
      "step": 129650
    },
    {
      "epoch": 3.872450959902069,
      "grad_norm": 4.832424640655518,
      "learning_rate": 0.00016294435968205532,
      "loss": 1.2902,
      "step": 129700
    },
    {
      "epoch": 3.8739438091541514,
      "grad_norm": 3.9749419689178467,
      "learning_rate": 0.0001629300738861365,
      "loss": 1.2803,
      "step": 129750
    },
    {
      "epoch": 3.875436658406234,
      "grad_norm": 4.696380138397217,
      "learning_rate": 0.00016291578809021765,
      "loss": 1.2226,
      "step": 129800
    },
    {
      "epoch": 3.876929507658317,
      "grad_norm": 3.757120132446289,
      "learning_rate": 0.00016290150229429884,
      "loss": 1.2499,
      "step": 129850
    },
    {
      "epoch": 3.878422356910399,
      "grad_norm": 4.010384559631348,
      "learning_rate": 0.00016288721649838,
      "loss": 1.3455,
      "step": 129900
    },
    {
      "epoch": 3.8799152061624818,
      "grad_norm": 3.7654688358306885,
      "learning_rate": 0.00016287293070246117,
      "loss": 1.2689,
      "step": 129950
    },
    {
      "epoch": 3.881408055414564,
      "grad_norm": 4.330887317657471,
      "learning_rate": 0.00016285864490654233,
      "loss": 1.2988,
      "step": 130000
    },
    {
      "epoch": 3.8829009046666467,
      "grad_norm": 3.8731918334960938,
      "learning_rate": 0.0001628443591106235,
      "loss": 1.2158,
      "step": 130050
    },
    {
      "epoch": 3.8843937539187294,
      "grad_norm": 4.860089302062988,
      "learning_rate": 0.00016283007331470466,
      "loss": 1.3152,
      "step": 130100
    },
    {
      "epoch": 3.8858866031708117,
      "grad_norm": 3.8552567958831787,
      "learning_rate": 0.00016281578751878583,
      "loss": 1.2038,
      "step": 130150
    },
    {
      "epoch": 3.8873794524228944,
      "grad_norm": 3.8789255619049072,
      "learning_rate": 0.000162801501722867,
      "loss": 1.1898,
      "step": 130200
    },
    {
      "epoch": 3.8888723016749767,
      "grad_norm": 5.833735466003418,
      "learning_rate": 0.00016278721592694815,
      "loss": 1.327,
      "step": 130250
    },
    {
      "epoch": 3.8903651509270594,
      "grad_norm": 5.1781463623046875,
      "learning_rate": 0.00016277293013102932,
      "loss": 1.3002,
      "step": 130300
    },
    {
      "epoch": 3.891858000179142,
      "grad_norm": 5.102818012237549,
      "learning_rate": 0.0001627586443351105,
      "loss": 1.257,
      "step": 130350
    },
    {
      "epoch": 3.8933508494312243,
      "grad_norm": 4.642506122589111,
      "learning_rate": 0.00016274435853919165,
      "loss": 1.2387,
      "step": 130400
    },
    {
      "epoch": 3.894843698683307,
      "grad_norm": 4.327921390533447,
      "learning_rate": 0.00016273007274327284,
      "loss": 1.2563,
      "step": 130450
    },
    {
      "epoch": 3.8963365479353893,
      "grad_norm": 4.99731969833374,
      "learning_rate": 0.00016271578694735398,
      "loss": 1.2621,
      "step": 130500
    },
    {
      "epoch": 3.897829397187472,
      "grad_norm": 3.6538639068603516,
      "learning_rate": 0.00016270150115143517,
      "loss": 1.2284,
      "step": 130550
    },
    {
      "epoch": 3.8993222464395547,
      "grad_norm": 4.081072807312012,
      "learning_rate": 0.0001626872153555163,
      "loss": 1.2901,
      "step": 130600
    },
    {
      "epoch": 3.900815095691637,
      "grad_norm": 5.384579658508301,
      "learning_rate": 0.0001626729295595975,
      "loss": 1.2898,
      "step": 130650
    },
    {
      "epoch": 3.9023079449437197,
      "grad_norm": 4.1314520835876465,
      "learning_rate": 0.00016265864376367866,
      "loss": 1.2692,
      "step": 130700
    },
    {
      "epoch": 3.903800794195802,
      "grad_norm": 5.35874080657959,
      "learning_rate": 0.00016264435796775983,
      "loss": 1.2377,
      "step": 130750
    },
    {
      "epoch": 3.9052936434478847,
      "grad_norm": 4.041543006896973,
      "learning_rate": 0.000162630072171841,
      "loss": 1.2757,
      "step": 130800
    },
    {
      "epoch": 3.9067864926999674,
      "grad_norm": 4.784633636474609,
      "learning_rate": 0.00016261578637592216,
      "loss": 1.234,
      "step": 130850
    },
    {
      "epoch": 3.9082793419520496,
      "grad_norm": 4.274779319763184,
      "learning_rate": 0.00016260150058000332,
      "loss": 1.2506,
      "step": 130900
    },
    {
      "epoch": 3.9097721912041323,
      "grad_norm": 4.858269214630127,
      "learning_rate": 0.0001625872147840845,
      "loss": 1.2345,
      "step": 130950
    },
    {
      "epoch": 3.9112650404562146,
      "grad_norm": 3.8560070991516113,
      "learning_rate": 0.00016257292898816565,
      "loss": 1.2689,
      "step": 131000
    },
    {
      "epoch": 3.9127578897082973,
      "grad_norm": 4.95020866394043,
      "learning_rate": 0.00016255864319224682,
      "loss": 1.2156,
      "step": 131050
    },
    {
      "epoch": 3.91425073896038,
      "grad_norm": 4.7120795249938965,
      "learning_rate": 0.00016254435739632798,
      "loss": 1.3119,
      "step": 131100
    },
    {
      "epoch": 3.9157435882124623,
      "grad_norm": 5.613494873046875,
      "learning_rate": 0.00016253007160040917,
      "loss": 1.2806,
      "step": 131150
    },
    {
      "epoch": 3.917236437464545,
      "grad_norm": 5.0516767501831055,
      "learning_rate": 0.0001625157858044903,
      "loss": 1.2254,
      "step": 131200
    },
    {
      "epoch": 3.9187292867166272,
      "grad_norm": 4.780202388763428,
      "learning_rate": 0.0001625015000085715,
      "loss": 1.2915,
      "step": 131250
    },
    {
      "epoch": 3.92022213596871,
      "grad_norm": 5.579587459564209,
      "learning_rate": 0.00016248721421265264,
      "loss": 1.2509,
      "step": 131300
    },
    {
      "epoch": 3.9217149852207926,
      "grad_norm": 4.543426990509033,
      "learning_rate": 0.00016247292841673383,
      "loss": 1.2974,
      "step": 131350
    },
    {
      "epoch": 3.923207834472875,
      "grad_norm": 5.310088634490967,
      "learning_rate": 0.000162458642620815,
      "loss": 1.2247,
      "step": 131400
    },
    {
      "epoch": 3.9247006837249576,
      "grad_norm": 4.900899887084961,
      "learning_rate": 0.00016244435682489616,
      "loss": 1.3038,
      "step": 131450
    },
    {
      "epoch": 3.92619353297704,
      "grad_norm": 4.816722869873047,
      "learning_rate": 0.00016243007102897732,
      "loss": 1.2484,
      "step": 131500
    },
    {
      "epoch": 3.9276863822291226,
      "grad_norm": 3.618678092956543,
      "learning_rate": 0.0001624157852330585,
      "loss": 1.3475,
      "step": 131550
    },
    {
      "epoch": 3.9291792314812053,
      "grad_norm": 5.137662410736084,
      "learning_rate": 0.00016240149943713965,
      "loss": 1.2443,
      "step": 131600
    },
    {
      "epoch": 3.9306720807332876,
      "grad_norm": 4.798673152923584,
      "learning_rate": 0.00016238721364122082,
      "loss": 1.2654,
      "step": 131650
    },
    {
      "epoch": 3.93216492998537,
      "grad_norm": 4.472255229949951,
      "learning_rate": 0.00016237292784530198,
      "loss": 1.2686,
      "step": 131700
    },
    {
      "epoch": 3.9336577792374525,
      "grad_norm": 4.61024284362793,
      "learning_rate": 0.00016235864204938315,
      "loss": 1.2491,
      "step": 131750
    },
    {
      "epoch": 3.9351506284895352,
      "grad_norm": 4.1179280281066895,
      "learning_rate": 0.0001623443562534643,
      "loss": 1.2044,
      "step": 131800
    },
    {
      "epoch": 3.9366434777416175,
      "grad_norm": 6.594708442687988,
      "learning_rate": 0.00016233007045754548,
      "loss": 1.2111,
      "step": 131850
    },
    {
      "epoch": 3.9381363269937,
      "grad_norm": 8.489596366882324,
      "learning_rate": 0.00016231578466162664,
      "loss": 1.2011,
      "step": 131900
    },
    {
      "epoch": 3.9396291762457825,
      "grad_norm": 5.426854133605957,
      "learning_rate": 0.0001623014988657078,
      "loss": 1.268,
      "step": 131950
    },
    {
      "epoch": 3.941122025497865,
      "grad_norm": 4.155928611755371,
      "learning_rate": 0.00016228721306978897,
      "loss": 1.2722,
      "step": 132000
    },
    {
      "epoch": 3.942614874749948,
      "grad_norm": 4.766868591308594,
      "learning_rate": 0.00016227292727387014,
      "loss": 1.288,
      "step": 132050
    },
    {
      "epoch": 3.94410772400203,
      "grad_norm": 4.16912317276001,
      "learning_rate": 0.0001622586414779513,
      "loss": 1.2261,
      "step": 132100
    },
    {
      "epoch": 3.945600573254113,
      "grad_norm": 6.41418981552124,
      "learning_rate": 0.00016224435568203247,
      "loss": 1.2414,
      "step": 132150
    },
    {
      "epoch": 3.947093422506195,
      "grad_norm": 4.736821174621582,
      "learning_rate": 0.00016223006988611366,
      "loss": 1.2481,
      "step": 132200
    },
    {
      "epoch": 3.948586271758278,
      "grad_norm": 4.794166564941406,
      "learning_rate": 0.0001622157840901948,
      "loss": 1.2831,
      "step": 132250
    },
    {
      "epoch": 3.9500791210103605,
      "grad_norm": 6.948697090148926,
      "learning_rate": 0.000162201498294276,
      "loss": 1.2545,
      "step": 132300
    },
    {
      "epoch": 3.951571970262443,
      "grad_norm": 5.114023208618164,
      "learning_rate": 0.00016218721249835712,
      "loss": 1.2809,
      "step": 132350
    },
    {
      "epoch": 3.9530648195145255,
      "grad_norm": 4.972701072692871,
      "learning_rate": 0.00016217292670243832,
      "loss": 1.2143,
      "step": 132400
    },
    {
      "epoch": 3.9545576687666077,
      "grad_norm": 3.860616445541382,
      "learning_rate": 0.00016215864090651945,
      "loss": 1.2209,
      "step": 132450
    },
    {
      "epoch": 3.9560505180186905,
      "grad_norm": 3.6322524547576904,
      "learning_rate": 0.00016214435511060065,
      "loss": 1.2252,
      "step": 132500
    },
    {
      "epoch": 3.957543367270773,
      "grad_norm": 5.153745174407959,
      "learning_rate": 0.0001621300693146818,
      "loss": 1.2748,
      "step": 132550
    },
    {
      "epoch": 3.9590362165228554,
      "grad_norm": 4.3787617683410645,
      "learning_rate": 0.00016211578351876297,
      "loss": 1.2634,
      "step": 132600
    },
    {
      "epoch": 3.960529065774938,
      "grad_norm": 4.335618019104004,
      "learning_rate": 0.00016210149772284414,
      "loss": 1.2661,
      "step": 132650
    },
    {
      "epoch": 3.9620219150270204,
      "grad_norm": 4.789446830749512,
      "learning_rate": 0.0001620872119269253,
      "loss": 1.2877,
      "step": 132700
    },
    {
      "epoch": 3.963514764279103,
      "grad_norm": 5.508213996887207,
      "learning_rate": 0.00016207292613100647,
      "loss": 1.2759,
      "step": 132750
    },
    {
      "epoch": 3.965007613531186,
      "grad_norm": 3.3553807735443115,
      "learning_rate": 0.00016205864033508763,
      "loss": 1.2907,
      "step": 132800
    },
    {
      "epoch": 3.966500462783268,
      "grad_norm": 4.773813247680664,
      "learning_rate": 0.0001620443545391688,
      "loss": 1.3115,
      "step": 132850
    },
    {
      "epoch": 3.9679933120353508,
      "grad_norm": 4.2718939781188965,
      "learning_rate": 0.00016203006874324996,
      "loss": 1.2467,
      "step": 132900
    },
    {
      "epoch": 3.969486161287433,
      "grad_norm": 4.36405086517334,
      "learning_rate": 0.00016201578294733113,
      "loss": 1.2536,
      "step": 132950
    },
    {
      "epoch": 3.9709790105395157,
      "grad_norm": 5.21968936920166,
      "learning_rate": 0.00016200149715141232,
      "loss": 1.3317,
      "step": 133000
    },
    {
      "epoch": 3.9724718597915984,
      "grad_norm": 3.792954683303833,
      "learning_rate": 0.00016198721135549346,
      "loss": 1.2041,
      "step": 133050
    },
    {
      "epoch": 3.9739647090436807,
      "grad_norm": 4.445356369018555,
      "learning_rate": 0.00016197292555957465,
      "loss": 1.2283,
      "step": 133100
    },
    {
      "epoch": 3.9754575582957634,
      "grad_norm": 4.6043548583984375,
      "learning_rate": 0.00016195863976365579,
      "loss": 1.328,
      "step": 133150
    },
    {
      "epoch": 3.9769504075478457,
      "grad_norm": 4.0124053955078125,
      "learning_rate": 0.00016194435396773698,
      "loss": 1.1772,
      "step": 133200
    },
    {
      "epoch": 3.9784432567999284,
      "grad_norm": 4.683640480041504,
      "learning_rate": 0.00016193006817181811,
      "loss": 1.2748,
      "step": 133250
    },
    {
      "epoch": 3.979936106052011,
      "grad_norm": 5.253026008605957,
      "learning_rate": 0.0001619157823758993,
      "loss": 1.2193,
      "step": 133300
    },
    {
      "epoch": 3.9814289553040934,
      "grad_norm": 4.60040283203125,
      "learning_rate": 0.00016190149657998047,
      "loss": 1.2518,
      "step": 133350
    },
    {
      "epoch": 3.982921804556176,
      "grad_norm": 5.973727226257324,
      "learning_rate": 0.00016188721078406164,
      "loss": 1.242,
      "step": 133400
    },
    {
      "epoch": 3.9844146538082583,
      "grad_norm": 5.09537410736084,
      "learning_rate": 0.0001618729249881428,
      "loss": 1.2484,
      "step": 133450
    },
    {
      "epoch": 3.985907503060341,
      "grad_norm": 4.428155899047852,
      "learning_rate": 0.00016185863919222397,
      "loss": 1.2112,
      "step": 133500
    },
    {
      "epoch": 3.9874003523124237,
      "grad_norm": 4.2854838371276855,
      "learning_rate": 0.00016184435339630513,
      "loss": 1.2402,
      "step": 133550
    },
    {
      "epoch": 3.988893201564506,
      "grad_norm": 4.689759254455566,
      "learning_rate": 0.0001618300676003863,
      "loss": 1.2705,
      "step": 133600
    },
    {
      "epoch": 3.9903860508165887,
      "grad_norm": 4.272946357727051,
      "learning_rate": 0.00016181578180446746,
      "loss": 1.2105,
      "step": 133650
    },
    {
      "epoch": 3.991878900068671,
      "grad_norm": 4.82036018371582,
      "learning_rate": 0.00016180149600854862,
      "loss": 1.2379,
      "step": 133700
    },
    {
      "epoch": 3.9933717493207537,
      "grad_norm": 5.448759078979492,
      "learning_rate": 0.0001617872102126298,
      "loss": 1.2077,
      "step": 133750
    },
    {
      "epoch": 3.9948645985728364,
      "grad_norm": 3.519653797149658,
      "learning_rate": 0.00016177292441671098,
      "loss": 1.2057,
      "step": 133800
    },
    {
      "epoch": 3.9963574478249186,
      "grad_norm": 4.229268550872803,
      "learning_rate": 0.00016175863862079212,
      "loss": 1.3109,
      "step": 133850
    },
    {
      "epoch": 3.997850297077001,
      "grad_norm": 3.772096633911133,
      "learning_rate": 0.0001617443528248733,
      "loss": 1.3218,
      "step": 133900
    },
    {
      "epoch": 3.9993431463290836,
      "grad_norm": 3.7670326232910156,
      "learning_rate": 0.00016173006702895445,
      "loss": 1.2882,
      "step": 133950
    },
    {
      "epoch": 4.000835995581166,
      "grad_norm": 4.392240524291992,
      "learning_rate": 0.00016171578123303564,
      "loss": 1.2225,
      "step": 134000
    },
    {
      "epoch": 4.002328844833249,
      "grad_norm": 4.192978382110596,
      "learning_rate": 0.00016170149543711678,
      "loss": 1.1469,
      "step": 134050
    },
    {
      "epoch": 4.003821694085331,
      "grad_norm": 4.702342987060547,
      "learning_rate": 0.00016168720964119797,
      "loss": 1.1452,
      "step": 134100
    },
    {
      "epoch": 4.0053145433374135,
      "grad_norm": 4.7361555099487305,
      "learning_rate": 0.00016167292384527913,
      "loss": 1.1871,
      "step": 134150
    },
    {
      "epoch": 4.006807392589496,
      "grad_norm": 3.4626450538635254,
      "learning_rate": 0.0001616586380493603,
      "loss": 1.159,
      "step": 134200
    },
    {
      "epoch": 4.008300241841579,
      "grad_norm": 4.5900654792785645,
      "learning_rate": 0.00016164435225344146,
      "loss": 1.2012,
      "step": 134250
    },
    {
      "epoch": 4.009793091093662,
      "grad_norm": 6.448458194732666,
      "learning_rate": 0.00016163006645752263,
      "loss": 1.1656,
      "step": 134300
    },
    {
      "epoch": 4.0112859403457435,
      "grad_norm": 4.073904991149902,
      "learning_rate": 0.0001616157806616038,
      "loss": 1.131,
      "step": 134350
    },
    {
      "epoch": 4.012778789597826,
      "grad_norm": 5.392414569854736,
      "learning_rate": 0.00016160149486568496,
      "loss": 1.1462,
      "step": 134400
    },
    {
      "epoch": 4.014271638849909,
      "grad_norm": 4.959211349487305,
      "learning_rate": 0.00016158720906976612,
      "loss": 1.1275,
      "step": 134450
    },
    {
      "epoch": 4.015764488101992,
      "grad_norm": 4.862436771392822,
      "learning_rate": 0.00016157292327384729,
      "loss": 1.1395,
      "step": 134500
    },
    {
      "epoch": 4.017257337354074,
      "grad_norm": 4.025382995605469,
      "learning_rate": 0.00016155863747792845,
      "loss": 1.1529,
      "step": 134550
    },
    {
      "epoch": 4.018750186606156,
      "grad_norm": 5.5243449211120605,
      "learning_rate": 0.00016154435168200961,
      "loss": 1.1552,
      "step": 134600
    },
    {
      "epoch": 4.020243035858239,
      "grad_norm": 5.48097038269043,
      "learning_rate": 0.00016153006588609078,
      "loss": 1.1542,
      "step": 134650
    },
    {
      "epoch": 4.0217358851103215,
      "grad_norm": 5.017364978790283,
      "learning_rate": 0.00016151578009017194,
      "loss": 1.1294,
      "step": 134700
    },
    {
      "epoch": 4.023228734362404,
      "grad_norm": 3.903317451477051,
      "learning_rate": 0.0001615014942942531,
      "loss": 1.1743,
      "step": 134750
    },
    {
      "epoch": 4.024721583614487,
      "grad_norm": 5.144746780395508,
      "learning_rate": 0.00016148720849833427,
      "loss": 1.184,
      "step": 134800
    },
    {
      "epoch": 4.026214432866569,
      "grad_norm": 3.201718330383301,
      "learning_rate": 0.00016147292270241547,
      "loss": 1.1294,
      "step": 134850
    },
    {
      "epoch": 4.0277072821186515,
      "grad_norm": 5.051429748535156,
      "learning_rate": 0.0001614586369064966,
      "loss": 1.1625,
      "step": 134900
    },
    {
      "epoch": 4.029200131370734,
      "grad_norm": 4.083528995513916,
      "learning_rate": 0.0001614443511105778,
      "loss": 1.1865,
      "step": 134950
    },
    {
      "epoch": 4.030692980622817,
      "grad_norm": 4.694931983947754,
      "learning_rate": 0.00016143006531465893,
      "loss": 1.1607,
      "step": 135000
    },
    {
      "epoch": 4.0321858298749,
      "grad_norm": 3.7577016353607178,
      "learning_rate": 0.00016141577951874012,
      "loss": 1.2009,
      "step": 135050
    },
    {
      "epoch": 4.033678679126981,
      "grad_norm": 3.913961887359619,
      "learning_rate": 0.00016140149372282126,
      "loss": 1.148,
      "step": 135100
    },
    {
      "epoch": 4.035171528379064,
      "grad_norm": 4.2388129234313965,
      "learning_rate": 0.00016138720792690245,
      "loss": 1.1673,
      "step": 135150
    },
    {
      "epoch": 4.036664377631147,
      "grad_norm": 4.703933238983154,
      "learning_rate": 0.00016137292213098362,
      "loss": 1.1249,
      "step": 135200
    },
    {
      "epoch": 4.0381572268832295,
      "grad_norm": 5.607854843139648,
      "learning_rate": 0.00016135863633506478,
      "loss": 1.1609,
      "step": 135250
    },
    {
      "epoch": 4.039650076135312,
      "grad_norm": 4.3828840255737305,
      "learning_rate": 0.00016134435053914595,
      "loss": 1.1788,
      "step": 135300
    },
    {
      "epoch": 4.041142925387394,
      "grad_norm": 5.778615474700928,
      "learning_rate": 0.0001613300647432271,
      "loss": 1.2101,
      "step": 135350
    },
    {
      "epoch": 4.042635774639477,
      "grad_norm": 4.011424541473389,
      "learning_rate": 0.00016131577894730828,
      "loss": 1.1991,
      "step": 135400
    },
    {
      "epoch": 4.0441286238915595,
      "grad_norm": 5.844552993774414,
      "learning_rate": 0.00016130149315138944,
      "loss": 1.2172,
      "step": 135450
    },
    {
      "epoch": 4.045621473143642,
      "grad_norm": 4.851437091827393,
      "learning_rate": 0.0001612872073554706,
      "loss": 1.2379,
      "step": 135500
    },
    {
      "epoch": 4.047114322395725,
      "grad_norm": 6.846874713897705,
      "learning_rate": 0.00016127292155955177,
      "loss": 1.1666,
      "step": 135550
    },
    {
      "epoch": 4.048607171647807,
      "grad_norm": 3.9584429264068604,
      "learning_rate": 0.00016125863576363294,
      "loss": 1.191,
      "step": 135600
    },
    {
      "epoch": 4.050100020899889,
      "grad_norm": 3.547799587249756,
      "learning_rate": 0.00016124434996771413,
      "loss": 1.234,
      "step": 135650
    },
    {
      "epoch": 4.051592870151972,
      "grad_norm": 4.4900641441345215,
      "learning_rate": 0.00016123006417179526,
      "loss": 1.1759,
      "step": 135700
    },
    {
      "epoch": 4.053085719404055,
      "grad_norm": 4.560009479522705,
      "learning_rate": 0.00016121577837587646,
      "loss": 1.1759,
      "step": 135750
    },
    {
      "epoch": 4.0545785686561375,
      "grad_norm": 4.02714204788208,
      "learning_rate": 0.0001612014925799576,
      "loss": 1.1788,
      "step": 135800
    },
    {
      "epoch": 4.056071417908219,
      "grad_norm": 4.854091644287109,
      "learning_rate": 0.00016118720678403879,
      "loss": 1.1621,
      "step": 135850
    },
    {
      "epoch": 4.057564267160302,
      "grad_norm": 4.28786563873291,
      "learning_rate": 0.00016117292098811992,
      "loss": 1.1694,
      "step": 135900
    },
    {
      "epoch": 4.059057116412385,
      "grad_norm": 5.054505825042725,
      "learning_rate": 0.00016115863519220111,
      "loss": 1.2124,
      "step": 135950
    },
    {
      "epoch": 4.0605499656644675,
      "grad_norm": 4.706223011016846,
      "learning_rate": 0.00016114434939628228,
      "loss": 1.1867,
      "step": 136000
    },
    {
      "epoch": 4.06204281491655,
      "grad_norm": 3.443118095397949,
      "learning_rate": 0.00016113006360036344,
      "loss": 1.1398,
      "step": 136050
    },
    {
      "epoch": 4.063535664168632,
      "grad_norm": 5.108642578125,
      "learning_rate": 0.0001611157778044446,
      "loss": 1.1902,
      "step": 136100
    },
    {
      "epoch": 4.065028513420715,
      "grad_norm": 5.258908271789551,
      "learning_rate": 0.00016110149200852577,
      "loss": 1.2093,
      "step": 136150
    },
    {
      "epoch": 4.066521362672797,
      "grad_norm": 3.8509016036987305,
      "learning_rate": 0.00016108720621260694,
      "loss": 1.1354,
      "step": 136200
    },
    {
      "epoch": 4.06801421192488,
      "grad_norm": 5.995665073394775,
      "learning_rate": 0.0001610729204166881,
      "loss": 1.1516,
      "step": 136250
    },
    {
      "epoch": 4.069507061176962,
      "grad_norm": 4.4564738273620605,
      "learning_rate": 0.00016105863462076927,
      "loss": 1.1405,
      "step": 136300
    },
    {
      "epoch": 4.070999910429045,
      "grad_norm": 7.378493309020996,
      "learning_rate": 0.00016104434882485043,
      "loss": 1.1817,
      "step": 136350
    },
    {
      "epoch": 4.072492759681127,
      "grad_norm": 4.1566009521484375,
      "learning_rate": 0.0001610300630289316,
      "loss": 1.1988,
      "step": 136400
    },
    {
      "epoch": 4.07398560893321,
      "grad_norm": 4.735561370849609,
      "learning_rate": 0.0001610157772330128,
      "loss": 1.1847,
      "step": 136450
    },
    {
      "epoch": 4.075478458185293,
      "grad_norm": 6.040157318115234,
      "learning_rate": 0.00016100149143709393,
      "loss": 1.1762,
      "step": 136500
    },
    {
      "epoch": 4.076971307437375,
      "grad_norm": 4.695454120635986,
      "learning_rate": 0.00016098720564117512,
      "loss": 1.2151,
      "step": 136550
    },
    {
      "epoch": 4.078464156689457,
      "grad_norm": 5.111282825469971,
      "learning_rate": 0.00016097291984525626,
      "loss": 1.2581,
      "step": 136600
    },
    {
      "epoch": 4.07995700594154,
      "grad_norm": 6.601406574249268,
      "learning_rate": 0.00016095863404933745,
      "loss": 1.1902,
      "step": 136650
    },
    {
      "epoch": 4.081449855193623,
      "grad_norm": 4.173478126525879,
      "learning_rate": 0.00016094434825341858,
      "loss": 1.1809,
      "step": 136700
    },
    {
      "epoch": 4.082942704445705,
      "grad_norm": 4.189050674438477,
      "learning_rate": 0.00016093006245749978,
      "loss": 1.1837,
      "step": 136750
    },
    {
      "epoch": 4.084435553697787,
      "grad_norm": 4.181232929229736,
      "learning_rate": 0.00016091577666158094,
      "loss": 1.124,
      "step": 136800
    },
    {
      "epoch": 4.08592840294987,
      "grad_norm": 6.015085697174072,
      "learning_rate": 0.0001609014908656621,
      "loss": 1.2253,
      "step": 136850
    },
    {
      "epoch": 4.087421252201953,
      "grad_norm": 4.773750305175781,
      "learning_rate": 0.00016088720506974327,
      "loss": 1.1498,
      "step": 136900
    },
    {
      "epoch": 4.088914101454035,
      "grad_norm": 4.635052680969238,
      "learning_rate": 0.00016087291927382443,
      "loss": 1.1946,
      "step": 136950
    },
    {
      "epoch": 4.090406950706118,
      "grad_norm": 3.5755093097686768,
      "learning_rate": 0.0001608586334779056,
      "loss": 1.178,
      "step": 137000
    },
    {
      "epoch": 4.0918997999582,
      "grad_norm": 5.496055603027344,
      "learning_rate": 0.00016084434768198676,
      "loss": 1.1409,
      "step": 137050
    },
    {
      "epoch": 4.093392649210283,
      "grad_norm": 4.117304801940918,
      "learning_rate": 0.00016083006188606793,
      "loss": 1.2,
      "step": 137100
    },
    {
      "epoch": 4.094885498462365,
      "grad_norm": 5.538839817047119,
      "learning_rate": 0.0001608157760901491,
      "loss": 1.1778,
      "step": 137150
    },
    {
      "epoch": 4.096378347714448,
      "grad_norm": 4.6166558265686035,
      "learning_rate": 0.00016080149029423026,
      "loss": 1.1797,
      "step": 137200
    },
    {
      "epoch": 4.097871196966531,
      "grad_norm": 4.904277324676514,
      "learning_rate": 0.00016078720449831142,
      "loss": 1.1951,
      "step": 137250
    },
    {
      "epoch": 4.0993640462186125,
      "grad_norm": 4.139499664306641,
      "learning_rate": 0.0001607729187023926,
      "loss": 1.1909,
      "step": 137300
    },
    {
      "epoch": 4.100856895470695,
      "grad_norm": 4.454082489013672,
      "learning_rate": 0.00016075863290647375,
      "loss": 1.1944,
      "step": 137350
    },
    {
      "epoch": 4.102349744722778,
      "grad_norm": 6.1854963302612305,
      "learning_rate": 0.00016074434711055492,
      "loss": 1.1934,
      "step": 137400
    },
    {
      "epoch": 4.103842593974861,
      "grad_norm": 3.4759559631347656,
      "learning_rate": 0.00016073006131463608,
      "loss": 1.1489,
      "step": 137450
    },
    {
      "epoch": 4.105335443226943,
      "grad_norm": 5.522613525390625,
      "learning_rate": 0.00016071577551871727,
      "loss": 1.1765,
      "step": 137500
    },
    {
      "epoch": 4.106828292479025,
      "grad_norm": 4.545987606048584,
      "learning_rate": 0.0001607014897227984,
      "loss": 1.2346,
      "step": 137550
    },
    {
      "epoch": 4.108321141731108,
      "grad_norm": 3.9791605472564697,
      "learning_rate": 0.0001606872039268796,
      "loss": 1.1124,
      "step": 137600
    },
    {
      "epoch": 4.109813990983191,
      "grad_norm": 5.259634971618652,
      "learning_rate": 0.00016067291813096074,
      "loss": 1.2323,
      "step": 137650
    },
    {
      "epoch": 4.111306840235273,
      "grad_norm": 3.549207925796509,
      "learning_rate": 0.00016065863233504193,
      "loss": 1.2269,
      "step": 137700
    },
    {
      "epoch": 4.112799689487356,
      "grad_norm": 4.890096664428711,
      "learning_rate": 0.00016064434653912307,
      "loss": 1.142,
      "step": 137750
    },
    {
      "epoch": 4.114292538739438,
      "grad_norm": 3.7870566844940186,
      "learning_rate": 0.00016063006074320426,
      "loss": 1.1446,
      "step": 137800
    },
    {
      "epoch": 4.1157853879915205,
      "grad_norm": 5.0174689292907715,
      "learning_rate": 0.00016061577494728543,
      "loss": 1.2022,
      "step": 137850
    },
    {
      "epoch": 4.117278237243603,
      "grad_norm": 4.542319297790527,
      "learning_rate": 0.0001606014891513666,
      "loss": 1.1809,
      "step": 137900
    },
    {
      "epoch": 4.118771086495686,
      "grad_norm": 5.590012073516846,
      "learning_rate": 0.00016058720335544776,
      "loss": 1.1811,
      "step": 137950
    },
    {
      "epoch": 4.120263935747769,
      "grad_norm": 4.569715976715088,
      "learning_rate": 0.00016057291755952892,
      "loss": 1.1715,
      "step": 138000
    },
    {
      "epoch": 4.12175678499985,
      "grad_norm": 3.725158214569092,
      "learning_rate": 0.00016055863176361008,
      "loss": 1.1825,
      "step": 138050
    },
    {
      "epoch": 4.123249634251933,
      "grad_norm": 3.941608428955078,
      "learning_rate": 0.00016054434596769125,
      "loss": 1.1482,
      "step": 138100
    },
    {
      "epoch": 4.124742483504016,
      "grad_norm": 4.545751094818115,
      "learning_rate": 0.00016053006017177241,
      "loss": 1.2089,
      "step": 138150
    },
    {
      "epoch": 4.1262353327560986,
      "grad_norm": 6.809751033782959,
      "learning_rate": 0.00016051577437585358,
      "loss": 1.1664,
      "step": 138200
    },
    {
      "epoch": 4.12772818200818,
      "grad_norm": 5.444116592407227,
      "learning_rate": 0.00016050148857993474,
      "loss": 1.2127,
      "step": 138250
    },
    {
      "epoch": 4.129221031260263,
      "grad_norm": 4.276831150054932,
      "learning_rate": 0.00016048720278401593,
      "loss": 1.2138,
      "step": 138300
    },
    {
      "epoch": 4.130713880512346,
      "grad_norm": 3.9716556072235107,
      "learning_rate": 0.00016047291698809707,
      "loss": 1.2237,
      "step": 138350
    },
    {
      "epoch": 4.1322067297644285,
      "grad_norm": 3.9793570041656494,
      "learning_rate": 0.00016045863119217826,
      "loss": 1.1997,
      "step": 138400
    },
    {
      "epoch": 4.133699579016511,
      "grad_norm": 4.590352535247803,
      "learning_rate": 0.0001604443453962594,
      "loss": 1.1631,
      "step": 138450
    },
    {
      "epoch": 4.135192428268593,
      "grad_norm": 3.9549121856689453,
      "learning_rate": 0.0001604300596003406,
      "loss": 1.1518,
      "step": 138500
    },
    {
      "epoch": 4.136685277520676,
      "grad_norm": 5.076075077056885,
      "learning_rate": 0.00016041577380442173,
      "loss": 1.1806,
      "step": 138550
    },
    {
      "epoch": 4.138178126772758,
      "grad_norm": 6.228057384490967,
      "learning_rate": 0.00016040148800850292,
      "loss": 1.1547,
      "step": 138600
    },
    {
      "epoch": 4.139670976024841,
      "grad_norm": 4.276813507080078,
      "learning_rate": 0.0001603872022125841,
      "loss": 1.1351,
      "step": 138650
    },
    {
      "epoch": 4.141163825276924,
      "grad_norm": 5.215205669403076,
      "learning_rate": 0.00016037291641666525,
      "loss": 1.2021,
      "step": 138700
    },
    {
      "epoch": 4.142656674529006,
      "grad_norm": 4.386962413787842,
      "learning_rate": 0.00016035863062074642,
      "loss": 1.2346,
      "step": 138750
    },
    {
      "epoch": 4.144149523781088,
      "grad_norm": 4.739329814910889,
      "learning_rate": 0.00016034434482482758,
      "loss": 1.1971,
      "step": 138800
    },
    {
      "epoch": 4.145642373033171,
      "grad_norm": 3.994267225265503,
      "learning_rate": 0.00016033005902890875,
      "loss": 1.2259,
      "step": 138850
    },
    {
      "epoch": 4.147135222285254,
      "grad_norm": 4.481820106506348,
      "learning_rate": 0.0001603157732329899,
      "loss": 1.2428,
      "step": 138900
    },
    {
      "epoch": 4.1486280715373365,
      "grad_norm": 5.454226016998291,
      "learning_rate": 0.00016030148743707108,
      "loss": 1.1596,
      "step": 138950
    },
    {
      "epoch": 4.150120920789418,
      "grad_norm": 5.072449684143066,
      "learning_rate": 0.00016028720164115224,
      "loss": 1.1755,
      "step": 139000
    },
    {
      "epoch": 4.151613770041501,
      "grad_norm": 4.940500259399414,
      "learning_rate": 0.0001602729158452334,
      "loss": 1.1765,
      "step": 139050
    },
    {
      "epoch": 4.153106619293584,
      "grad_norm": 4.294040203094482,
      "learning_rate": 0.0001602586300493146,
      "loss": 1.1638,
      "step": 139100
    },
    {
      "epoch": 4.154599468545666,
      "grad_norm": 3.98103928565979,
      "learning_rate": 0.00016024434425339573,
      "loss": 1.192,
      "step": 139150
    },
    {
      "epoch": 4.156092317797749,
      "grad_norm": 5.18140983581543,
      "learning_rate": 0.00016023005845747693,
      "loss": 1.2121,
      "step": 139200
    },
    {
      "epoch": 4.157585167049831,
      "grad_norm": 4.342568874359131,
      "learning_rate": 0.00016021577266155806,
      "loss": 1.2056,
      "step": 139250
    },
    {
      "epoch": 4.159078016301914,
      "grad_norm": 4.284645080566406,
      "learning_rate": 0.00016020148686563926,
      "loss": 1.1864,
      "step": 139300
    },
    {
      "epoch": 4.160570865553996,
      "grad_norm": 5.542572498321533,
      "learning_rate": 0.0001601872010697204,
      "loss": 1.1903,
      "step": 139350
    },
    {
      "epoch": 4.162063714806079,
      "grad_norm": 4.752786159515381,
      "learning_rate": 0.00016017291527380158,
      "loss": 1.2012,
      "step": 139400
    },
    {
      "epoch": 4.163556564058162,
      "grad_norm": 4.298015117645264,
      "learning_rate": 0.00016015862947788275,
      "loss": 1.2108,
      "step": 139450
    },
    {
      "epoch": 4.165049413310244,
      "grad_norm": 4.048568248748779,
      "learning_rate": 0.00016014434368196391,
      "loss": 1.2191,
      "step": 139500
    },
    {
      "epoch": 4.166542262562326,
      "grad_norm": 4.178681373596191,
      "learning_rate": 0.00016013005788604508,
      "loss": 1.1892,
      "step": 139550
    },
    {
      "epoch": 4.168035111814409,
      "grad_norm": 6.201132297515869,
      "learning_rate": 0.00016011577209012622,
      "loss": 1.1917,
      "step": 139600
    },
    {
      "epoch": 4.169527961066492,
      "grad_norm": 4.938151836395264,
      "learning_rate": 0.0001601014862942074,
      "loss": 1.1954,
      "step": 139650
    },
    {
      "epoch": 4.171020810318574,
      "grad_norm": 6.060943126678467,
      "learning_rate": 0.00016008720049828855,
      "loss": 1.1471,
      "step": 139700
    },
    {
      "epoch": 4.172513659570656,
      "grad_norm": 5.0269880294799805,
      "learning_rate": 0.00016007291470236974,
      "loss": 1.1868,
      "step": 139750
    },
    {
      "epoch": 4.174006508822739,
      "grad_norm": 4.509561538696289,
      "learning_rate": 0.0001600586289064509,
      "loss": 1.1906,
      "step": 139800
    },
    {
      "epoch": 4.175499358074822,
      "grad_norm": 4.10978364944458,
      "learning_rate": 0.00016004434311053207,
      "loss": 1.1515,
      "step": 139850
    },
    {
      "epoch": 4.176992207326904,
      "grad_norm": 3.7452332973480225,
      "learning_rate": 0.00016003005731461323,
      "loss": 1.1807,
      "step": 139900
    },
    {
      "epoch": 4.178485056578987,
      "grad_norm": 4.911968231201172,
      "learning_rate": 0.0001600157715186944,
      "loss": 1.1568,
      "step": 139950
    },
    {
      "epoch": 4.179977905831069,
      "grad_norm": 4.606587886810303,
      "learning_rate": 0.00016000148572277556,
      "loss": 1.2099,
      "step": 140000
    },
    {
      "epoch": 4.181470755083152,
      "grad_norm": 4.736166954040527,
      "learning_rate": 0.00015998719992685672,
      "loss": 1.2101,
      "step": 140050
    },
    {
      "epoch": 4.182963604335234,
      "grad_norm": 4.431340217590332,
      "learning_rate": 0.0001599729141309379,
      "loss": 1.1885,
      "step": 140100
    },
    {
      "epoch": 4.184456453587317,
      "grad_norm": 4.86411190032959,
      "learning_rate": 0.00015995862833501905,
      "loss": 1.2,
      "step": 140150
    },
    {
      "epoch": 4.1859493028394,
      "grad_norm": 4.4166951179504395,
      "learning_rate": 0.00015994434253910022,
      "loss": 1.1717,
      "step": 140200
    },
    {
      "epoch": 4.1874421520914815,
      "grad_norm": 4.2804856300354,
      "learning_rate": 0.0001599300567431814,
      "loss": 1.1695,
      "step": 140250
    },
    {
      "epoch": 4.188935001343564,
      "grad_norm": 3.3184804916381836,
      "learning_rate": 0.00015991577094726255,
      "loss": 1.1864,
      "step": 140300
    },
    {
      "epoch": 4.190427850595647,
      "grad_norm": 3.712420701980591,
      "learning_rate": 0.00015990148515134374,
      "loss": 1.2254,
      "step": 140350
    },
    {
      "epoch": 4.19192069984773,
      "grad_norm": 4.40815544128418,
      "learning_rate": 0.00015988719935542488,
      "loss": 1.1645,
      "step": 140400
    },
    {
      "epoch": 4.193413549099812,
      "grad_norm": 4.534137725830078,
      "learning_rate": 0.00015987291355950607,
      "loss": 1.2137,
      "step": 140450
    },
    {
      "epoch": 4.194906398351894,
      "grad_norm": 3.4060471057891846,
      "learning_rate": 0.00015985862776358723,
      "loss": 1.1804,
      "step": 140500
    },
    {
      "epoch": 4.196399247603977,
      "grad_norm": 3.3433594703674316,
      "learning_rate": 0.0001598443419676684,
      "loss": 1.1801,
      "step": 140550
    },
    {
      "epoch": 4.19789209685606,
      "grad_norm": 4.448024272918701,
      "learning_rate": 0.00015983005617174956,
      "loss": 1.2327,
      "step": 140600
    },
    {
      "epoch": 4.199384946108142,
      "grad_norm": 4.073503494262695,
      "learning_rate": 0.00015981577037583073,
      "loss": 1.2058,
      "step": 140650
    },
    {
      "epoch": 4.200877795360224,
      "grad_norm": 4.200987339019775,
      "learning_rate": 0.0001598014845799119,
      "loss": 1.166,
      "step": 140700
    },
    {
      "epoch": 4.202370644612307,
      "grad_norm": 6.871157169342041,
      "learning_rate": 0.00015978719878399306,
      "loss": 1.1599,
      "step": 140750
    },
    {
      "epoch": 4.2038634938643895,
      "grad_norm": 6.4246745109558105,
      "learning_rate": 0.00015977291298807422,
      "loss": 1.2119,
      "step": 140800
    },
    {
      "epoch": 4.205356343116472,
      "grad_norm": 4.552119731903076,
      "learning_rate": 0.0001597586271921554,
      "loss": 1.2171,
      "step": 140850
    },
    {
      "epoch": 4.206849192368555,
      "grad_norm": 4.876185894012451,
      "learning_rate": 0.00015974434139623655,
      "loss": 1.2087,
      "step": 140900
    },
    {
      "epoch": 4.208342041620637,
      "grad_norm": 5.435263633728027,
      "learning_rate": 0.00015973005560031774,
      "loss": 1.1805,
      "step": 140950
    },
    {
      "epoch": 4.2098348908727194,
      "grad_norm": 6.356777191162109,
      "learning_rate": 0.00015971576980439888,
      "loss": 1.1849,
      "step": 141000
    },
    {
      "epoch": 4.211327740124802,
      "grad_norm": 5.153776168823242,
      "learning_rate": 0.00015970148400848007,
      "loss": 1.1769,
      "step": 141050
    },
    {
      "epoch": 4.212820589376885,
      "grad_norm": 3.6175692081451416,
      "learning_rate": 0.0001596871982125612,
      "loss": 1.1857,
      "step": 141100
    },
    {
      "epoch": 4.214313438628968,
      "grad_norm": 3.5825202465057373,
      "learning_rate": 0.0001596729124166424,
      "loss": 1.2036,
      "step": 141150
    },
    {
      "epoch": 4.215806287881049,
      "grad_norm": 4.836036205291748,
      "learning_rate": 0.00015965862662072354,
      "loss": 1.2118,
      "step": 141200
    },
    {
      "epoch": 4.217299137133132,
      "grad_norm": 4.852596759796143,
      "learning_rate": 0.00015964434082480473,
      "loss": 1.1825,
      "step": 141250
    },
    {
      "epoch": 4.218791986385215,
      "grad_norm": 6.009531497955322,
      "learning_rate": 0.0001596300550288859,
      "loss": 1.2429,
      "step": 141300
    },
    {
      "epoch": 4.2202848356372975,
      "grad_norm": 4.5213727951049805,
      "learning_rate": 0.00015961576923296706,
      "loss": 1.1961,
      "step": 141350
    },
    {
      "epoch": 4.22177768488938,
      "grad_norm": 4.069146633148193,
      "learning_rate": 0.00015960148343704822,
      "loss": 1.214,
      "step": 141400
    },
    {
      "epoch": 4.223270534141462,
      "grad_norm": 4.735863208770752,
      "learning_rate": 0.0001595871976411294,
      "loss": 1.2156,
      "step": 141450
    },
    {
      "epoch": 4.224763383393545,
      "grad_norm": 4.679515361785889,
      "learning_rate": 0.00015957291184521055,
      "loss": 1.1919,
      "step": 141500
    },
    {
      "epoch": 4.226256232645627,
      "grad_norm": 4.299509048461914,
      "learning_rate": 0.00015955862604929172,
      "loss": 1.198,
      "step": 141550
    },
    {
      "epoch": 4.22774908189771,
      "grad_norm": 4.518818378448486,
      "learning_rate": 0.00015954434025337288,
      "loss": 1.1918,
      "step": 141600
    },
    {
      "epoch": 4.229241931149793,
      "grad_norm": 2.8362114429473877,
      "learning_rate": 0.00015953005445745405,
      "loss": 1.1891,
      "step": 141650
    },
    {
      "epoch": 4.230734780401875,
      "grad_norm": 5.10905647277832,
      "learning_rate": 0.0001595157686615352,
      "loss": 1.2033,
      "step": 141700
    },
    {
      "epoch": 4.232227629653957,
      "grad_norm": 5.38252592086792,
      "learning_rate": 0.0001595014828656164,
      "loss": 1.2235,
      "step": 141750
    },
    {
      "epoch": 4.23372047890604,
      "grad_norm": 4.693549156188965,
      "learning_rate": 0.00015948719706969754,
      "loss": 1.2876,
      "step": 141800
    },
    {
      "epoch": 4.235213328158123,
      "grad_norm": 5.683078289031982,
      "learning_rate": 0.00015947291127377873,
      "loss": 1.175,
      "step": 141850
    },
    {
      "epoch": 4.2367061774102055,
      "grad_norm": 4.098508358001709,
      "learning_rate": 0.00015945862547785987,
      "loss": 1.2489,
      "step": 141900
    },
    {
      "epoch": 4.238199026662287,
      "grad_norm": 4.055385589599609,
      "learning_rate": 0.00015944433968194106,
      "loss": 1.1783,
      "step": 141950
    },
    {
      "epoch": 4.23969187591437,
      "grad_norm": 4.988245010375977,
      "learning_rate": 0.0001594300538860222,
      "loss": 1.1549,
      "step": 142000
    },
    {
      "epoch": 4.241184725166453,
      "grad_norm": 4.256673812866211,
      "learning_rate": 0.0001594157680901034,
      "loss": 1.2084,
      "step": 142050
    },
    {
      "epoch": 4.242677574418535,
      "grad_norm": 5.259324550628662,
      "learning_rate": 0.00015940148229418456,
      "loss": 1.1815,
      "step": 142100
    },
    {
      "epoch": 4.244170423670618,
      "grad_norm": 5.063663005828857,
      "learning_rate": 0.00015938719649826572,
      "loss": 1.1878,
      "step": 142150
    },
    {
      "epoch": 4.2456632729227,
      "grad_norm": 3.863372325897217,
      "learning_rate": 0.00015937291070234689,
      "loss": 1.1603,
      "step": 142200
    },
    {
      "epoch": 4.247156122174783,
      "grad_norm": 4.990203380584717,
      "learning_rate": 0.00015935862490642802,
      "loss": 1.2208,
      "step": 142250
    },
    {
      "epoch": 4.248648971426865,
      "grad_norm": 3.7190535068511963,
      "learning_rate": 0.00015934433911050922,
      "loss": 1.2135,
      "step": 142300
    },
    {
      "epoch": 4.250141820678948,
      "grad_norm": 6.787582874298096,
      "learning_rate": 0.00015933005331459035,
      "loss": 1.1978,
      "step": 142350
    },
    {
      "epoch": 4.251634669931031,
      "grad_norm": 4.888009548187256,
      "learning_rate": 0.00015931576751867155,
      "loss": 1.2557,
      "step": 142400
    },
    {
      "epoch": 4.253127519183113,
      "grad_norm": 4.288732528686523,
      "learning_rate": 0.0001593014817227527,
      "loss": 1.2057,
      "step": 142450
    },
    {
      "epoch": 4.254620368435195,
      "grad_norm": 4.282674789428711,
      "learning_rate": 0.00015928719592683387,
      "loss": 1.1783,
      "step": 142500
    },
    {
      "epoch": 4.256113217687278,
      "grad_norm": 4.2092719078063965,
      "learning_rate": 0.00015927291013091504,
      "loss": 1.214,
      "step": 142550
    },
    {
      "epoch": 4.257606066939361,
      "grad_norm": 3.5486562252044678,
      "learning_rate": 0.0001592586243349962,
      "loss": 1.1579,
      "step": 142600
    },
    {
      "epoch": 4.2590989161914425,
      "grad_norm": 4.74271821975708,
      "learning_rate": 0.00015924433853907737,
      "loss": 1.145,
      "step": 142650
    },
    {
      "epoch": 4.260591765443525,
      "grad_norm": 5.327475070953369,
      "learning_rate": 0.00015923005274315853,
      "loss": 1.1631,
      "step": 142700
    },
    {
      "epoch": 4.262084614695608,
      "grad_norm": 4.441630840301514,
      "learning_rate": 0.0001592157669472397,
      "loss": 1.1797,
      "step": 142750
    },
    {
      "epoch": 4.263577463947691,
      "grad_norm": 4.510640621185303,
      "learning_rate": 0.00015920148115132086,
      "loss": 1.1697,
      "step": 142800
    },
    {
      "epoch": 4.265070313199773,
      "grad_norm": 4.31403112411499,
      "learning_rate": 0.00015918719535540203,
      "loss": 1.2071,
      "step": 142850
    },
    {
      "epoch": 4.266563162451856,
      "grad_norm": 4.649517059326172,
      "learning_rate": 0.00015917290955948322,
      "loss": 1.1932,
      "step": 142900
    },
    {
      "epoch": 4.268056011703938,
      "grad_norm": 3.873218059539795,
      "learning_rate": 0.00015915862376356436,
      "loss": 1.1632,
      "step": 142950
    },
    {
      "epoch": 4.269548860956021,
      "grad_norm": 4.308788776397705,
      "learning_rate": 0.00015914433796764555,
      "loss": 1.2257,
      "step": 143000
    },
    {
      "epoch": 4.271041710208103,
      "grad_norm": 5.109141826629639,
      "learning_rate": 0.00015913005217172669,
      "loss": 1.2442,
      "step": 143050
    },
    {
      "epoch": 4.272534559460186,
      "grad_norm": 3.967876434326172,
      "learning_rate": 0.00015911576637580788,
      "loss": 1.2319,
      "step": 143100
    },
    {
      "epoch": 4.274027408712268,
      "grad_norm": 5.368978500366211,
      "learning_rate": 0.00015910148057988904,
      "loss": 1.1588,
      "step": 143150
    },
    {
      "epoch": 4.2755202579643505,
      "grad_norm": 5.190979480743408,
      "learning_rate": 0.0001590871947839702,
      "loss": 1.2443,
      "step": 143200
    },
    {
      "epoch": 4.277013107216433,
      "grad_norm": 3.6689977645874023,
      "learning_rate": 0.00015907290898805137,
      "loss": 1.2022,
      "step": 143250
    },
    {
      "epoch": 4.278505956468516,
      "grad_norm": 4.946050643920898,
      "learning_rate": 0.00015905862319213254,
      "loss": 1.2022,
      "step": 143300
    },
    {
      "epoch": 4.279998805720599,
      "grad_norm": 3.930406093597412,
      "learning_rate": 0.0001590443373962137,
      "loss": 1.2069,
      "step": 143350
    },
    {
      "epoch": 4.2814916549726805,
      "grad_norm": 4.970086574554443,
      "learning_rate": 0.00015903005160029487,
      "loss": 1.1877,
      "step": 143400
    },
    {
      "epoch": 4.282984504224763,
      "grad_norm": 4.933574199676514,
      "learning_rate": 0.00015901576580437603,
      "loss": 1.2037,
      "step": 143450
    },
    {
      "epoch": 4.284477353476846,
      "grad_norm": 5.583881378173828,
      "learning_rate": 0.0001590014800084572,
      "loss": 1.2082,
      "step": 143500
    },
    {
      "epoch": 4.285970202728929,
      "grad_norm": 3.963679552078247,
      "learning_rate": 0.00015898719421253836,
      "loss": 1.176,
      "step": 143550
    },
    {
      "epoch": 4.287463051981011,
      "grad_norm": 3.982883930206299,
      "learning_rate": 0.00015897290841661952,
      "loss": 1.2513,
      "step": 143600
    },
    {
      "epoch": 4.288955901233093,
      "grad_norm": 3.9288153648376465,
      "learning_rate": 0.0001589586226207007,
      "loss": 1.187,
      "step": 143650
    },
    {
      "epoch": 4.290448750485176,
      "grad_norm": 5.109147071838379,
      "learning_rate": 0.00015894433682478188,
      "loss": 1.2322,
      "step": 143700
    },
    {
      "epoch": 4.2919415997372585,
      "grad_norm": 7.246605396270752,
      "learning_rate": 0.00015893005102886302,
      "loss": 1.1822,
      "step": 143750
    },
    {
      "epoch": 4.293434448989341,
      "grad_norm": 4.197552680969238,
      "learning_rate": 0.0001589157652329442,
      "loss": 1.2281,
      "step": 143800
    },
    {
      "epoch": 4.294927298241424,
      "grad_norm": 6.113961696624756,
      "learning_rate": 0.00015890147943702535,
      "loss": 1.1814,
      "step": 143850
    },
    {
      "epoch": 4.296420147493506,
      "grad_norm": 6.120628356933594,
      "learning_rate": 0.00015888719364110654,
      "loss": 1.1497,
      "step": 143900
    },
    {
      "epoch": 4.2979129967455885,
      "grad_norm": 5.152296543121338,
      "learning_rate": 0.0001588729078451877,
      "loss": 1.2277,
      "step": 143950
    },
    {
      "epoch": 4.299405845997671,
      "grad_norm": 4.7058000564575195,
      "learning_rate": 0.00015885862204926887,
      "loss": 1.1586,
      "step": 144000
    },
    {
      "epoch": 4.300898695249754,
      "grad_norm": 4.757035732269287,
      "learning_rate": 0.00015884433625335003,
      "loss": 1.1701,
      "step": 144050
    },
    {
      "epoch": 4.302391544501837,
      "grad_norm": 5.381356239318848,
      "learning_rate": 0.0001588300504574312,
      "loss": 1.2242,
      "step": 144100
    },
    {
      "epoch": 4.303884393753918,
      "grad_norm": 4.9179511070251465,
      "learning_rate": 0.00015881576466151236,
      "loss": 1.2313,
      "step": 144150
    },
    {
      "epoch": 4.305377243006001,
      "grad_norm": 4.601654529571533,
      "learning_rate": 0.00015880147886559353,
      "loss": 1.2054,
      "step": 144200
    },
    {
      "epoch": 4.306870092258084,
      "grad_norm": 4.032548904418945,
      "learning_rate": 0.0001587871930696747,
      "loss": 1.2266,
      "step": 144250
    },
    {
      "epoch": 4.3083629415101665,
      "grad_norm": 4.067842960357666,
      "learning_rate": 0.00015877290727375586,
      "loss": 1.1795,
      "step": 144300
    },
    {
      "epoch": 4.309855790762249,
      "grad_norm": 4.99582576751709,
      "learning_rate": 0.00015875862147783702,
      "loss": 1.2216,
      "step": 144350
    },
    {
      "epoch": 4.311348640014331,
      "grad_norm": 6.543945789337158,
      "learning_rate": 0.0001587443356819182,
      "loss": 1.1862,
      "step": 144400
    },
    {
      "epoch": 4.312841489266414,
      "grad_norm": 4.526428699493408,
      "learning_rate": 0.00015873004988599935,
      "loss": 1.1575,
      "step": 144450
    },
    {
      "epoch": 4.3143343385184965,
      "grad_norm": 5.83212947845459,
      "learning_rate": 0.00015871576409008054,
      "loss": 1.2166,
      "step": 144500
    },
    {
      "epoch": 4.315827187770579,
      "grad_norm": 3.717582941055298,
      "learning_rate": 0.00015870147829416168,
      "loss": 1.2081,
      "step": 144550
    },
    {
      "epoch": 4.317320037022661,
      "grad_norm": 4.536131381988525,
      "learning_rate": 0.00015868719249824287,
      "loss": 1.2142,
      "step": 144600
    },
    {
      "epoch": 4.318812886274744,
      "grad_norm": 4.8904290199279785,
      "learning_rate": 0.000158672906702324,
      "loss": 1.2321,
      "step": 144650
    },
    {
      "epoch": 4.320305735526826,
      "grad_norm": 3.6146597862243652,
      "learning_rate": 0.0001586586209064052,
      "loss": 1.1443,
      "step": 144700
    },
    {
      "epoch": 4.321798584778909,
      "grad_norm": 4.939590930938721,
      "learning_rate": 0.00015864433511048637,
      "loss": 1.2164,
      "step": 144750
    },
    {
      "epoch": 4.323291434030992,
      "grad_norm": 3.37874698638916,
      "learning_rate": 0.00015863004931456753,
      "loss": 1.2104,
      "step": 144800
    },
    {
      "epoch": 4.3247842832830745,
      "grad_norm": 4.682936191558838,
      "learning_rate": 0.0001586157635186487,
      "loss": 1.2236,
      "step": 144850
    },
    {
      "epoch": 4.326277132535156,
      "grad_norm": 3.446786880493164,
      "learning_rate": 0.00015860147772272983,
      "loss": 1.1452,
      "step": 144900
    },
    {
      "epoch": 4.327769981787239,
      "grad_norm": 3.864423990249634,
      "learning_rate": 0.00015858719192681102,
      "loss": 1.1497,
      "step": 144950
    },
    {
      "epoch": 4.329262831039322,
      "grad_norm": 5.804245471954346,
      "learning_rate": 0.00015857290613089216,
      "loss": 1.2297,
      "step": 145000
    },
    {
      "epoch": 4.3307556802914045,
      "grad_norm": 6.271811485290527,
      "learning_rate": 0.00015855862033497335,
      "loss": 1.1358,
      "step": 145050
    },
    {
      "epoch": 4.332248529543486,
      "grad_norm": 5.019343852996826,
      "learning_rate": 0.00015854433453905452,
      "loss": 1.2191,
      "step": 145100
    },
    {
      "epoch": 4.333741378795569,
      "grad_norm": 7.10085916519165,
      "learning_rate": 0.00015853004874313568,
      "loss": 1.1678,
      "step": 145150
    },
    {
      "epoch": 4.335234228047652,
      "grad_norm": 4.680956840515137,
      "learning_rate": 0.00015851576294721685,
      "loss": 1.155,
      "step": 145200
    },
    {
      "epoch": 4.336727077299734,
      "grad_norm": 4.227858066558838,
      "learning_rate": 0.000158501477151298,
      "loss": 1.1563,
      "step": 145250
    },
    {
      "epoch": 4.338219926551817,
      "grad_norm": 6.304920196533203,
      "learning_rate": 0.00015848719135537918,
      "loss": 1.2083,
      "step": 145300
    },
    {
      "epoch": 4.339712775803899,
      "grad_norm": 3.7268805503845215,
      "learning_rate": 0.00015847290555946034,
      "loss": 1.2171,
      "step": 145350
    },
    {
      "epoch": 4.341205625055982,
      "grad_norm": 3.7859396934509277,
      "learning_rate": 0.0001584586197635415,
      "loss": 1.198,
      "step": 145400
    },
    {
      "epoch": 4.342698474308064,
      "grad_norm": 4.7007246017456055,
      "learning_rate": 0.00015844433396762267,
      "loss": 1.2418,
      "step": 145450
    },
    {
      "epoch": 4.344191323560147,
      "grad_norm": 4.652912616729736,
      "learning_rate": 0.00015843004817170383,
      "loss": 1.2408,
      "step": 145500
    },
    {
      "epoch": 4.34568417281223,
      "grad_norm": 5.060214519500732,
      "learning_rate": 0.00015841576237578503,
      "loss": 1.1922,
      "step": 145550
    },
    {
      "epoch": 4.347177022064312,
      "grad_norm": 4.346338272094727,
      "learning_rate": 0.00015840147657986616,
      "loss": 1.223,
      "step": 145600
    },
    {
      "epoch": 4.348669871316394,
      "grad_norm": 3.7870049476623535,
      "learning_rate": 0.00015838719078394736,
      "loss": 1.205,
      "step": 145650
    },
    {
      "epoch": 4.350162720568477,
      "grad_norm": 4.397316932678223,
      "learning_rate": 0.0001583729049880285,
      "loss": 1.1928,
      "step": 145700
    },
    {
      "epoch": 4.35165556982056,
      "grad_norm": 4.058841228485107,
      "learning_rate": 0.00015835861919210969,
      "loss": 1.2002,
      "step": 145750
    },
    {
      "epoch": 4.353148419072642,
      "grad_norm": 4.240140914916992,
      "learning_rate": 0.00015834433339619082,
      "loss": 1.1746,
      "step": 145800
    },
    {
      "epoch": 4.354641268324724,
      "grad_norm": 4.605095863342285,
      "learning_rate": 0.00015833004760027201,
      "loss": 1.1907,
      "step": 145850
    },
    {
      "epoch": 4.356134117576807,
      "grad_norm": 4.193150997161865,
      "learning_rate": 0.00015831576180435318,
      "loss": 1.2754,
      "step": 145900
    },
    {
      "epoch": 4.35762696682889,
      "grad_norm": 4.516674995422363,
      "learning_rate": 0.00015830147600843434,
      "loss": 1.2,
      "step": 145950
    },
    {
      "epoch": 4.359119816080972,
      "grad_norm": 4.528346538543701,
      "learning_rate": 0.0001582871902125155,
      "loss": 1.2143,
      "step": 146000
    },
    {
      "epoch": 4.360612665333055,
      "grad_norm": 4.620771408081055,
      "learning_rate": 0.00015827290441659667,
      "loss": 1.2008,
      "step": 146050
    },
    {
      "epoch": 4.362105514585137,
      "grad_norm": 5.53831148147583,
      "learning_rate": 0.00015825861862067784,
      "loss": 1.1736,
      "step": 146100
    },
    {
      "epoch": 4.3635983638372196,
      "grad_norm": 5.173369884490967,
      "learning_rate": 0.000158244332824759,
      "loss": 1.2242,
      "step": 146150
    },
    {
      "epoch": 4.365091213089302,
      "grad_norm": 4.372010231018066,
      "learning_rate": 0.00015823004702884017,
      "loss": 1.1915,
      "step": 146200
    },
    {
      "epoch": 4.366584062341385,
      "grad_norm": 6.224375247955322,
      "learning_rate": 0.00015821576123292133,
      "loss": 1.2133,
      "step": 146250
    },
    {
      "epoch": 4.368076911593468,
      "grad_norm": 8.175760269165039,
      "learning_rate": 0.0001582014754370025,
      "loss": 1.208,
      "step": 146300
    },
    {
      "epoch": 4.3695697608455495,
      "grad_norm": 4.286147594451904,
      "learning_rate": 0.0001581871896410837,
      "loss": 1.2283,
      "step": 146350
    },
    {
      "epoch": 4.371062610097632,
      "grad_norm": 5.579278945922852,
      "learning_rate": 0.00015817290384516483,
      "loss": 1.1817,
      "step": 146400
    },
    {
      "epoch": 4.372555459349715,
      "grad_norm": 3.485745668411255,
      "learning_rate": 0.00015815861804924602,
      "loss": 1.2079,
      "step": 146450
    },
    {
      "epoch": 4.374048308601798,
      "grad_norm": 5.126974582672119,
      "learning_rate": 0.00015814433225332716,
      "loss": 1.1943,
      "step": 146500
    },
    {
      "epoch": 4.37554115785388,
      "grad_norm": 5.43192720413208,
      "learning_rate": 0.00015813004645740835,
      "loss": 1.1759,
      "step": 146550
    },
    {
      "epoch": 4.377034007105962,
      "grad_norm": 5.335154056549072,
      "learning_rate": 0.0001581157606614895,
      "loss": 1.2406,
      "step": 146600
    },
    {
      "epoch": 4.378526856358045,
      "grad_norm": 4.4820756912231445,
      "learning_rate": 0.00015810147486557068,
      "loss": 1.2259,
      "step": 146650
    },
    {
      "epoch": 4.3800197056101275,
      "grad_norm": 4.541141986846924,
      "learning_rate": 0.00015808718906965184,
      "loss": 1.2041,
      "step": 146700
    },
    {
      "epoch": 4.38151255486221,
      "grad_norm": 5.367811679840088,
      "learning_rate": 0.000158072903273733,
      "loss": 1.1894,
      "step": 146750
    },
    {
      "epoch": 4.383005404114293,
      "grad_norm": 3.965756893157959,
      "learning_rate": 0.00015805861747781417,
      "loss": 1.2322,
      "step": 146800
    },
    {
      "epoch": 4.384498253366375,
      "grad_norm": 5.187005519866943,
      "learning_rate": 0.00015804433168189533,
      "loss": 1.2183,
      "step": 146850
    },
    {
      "epoch": 4.3859911026184575,
      "grad_norm": 4.8554205894470215,
      "learning_rate": 0.0001580300458859765,
      "loss": 1.2169,
      "step": 146900
    },
    {
      "epoch": 4.38748395187054,
      "grad_norm": 5.468644618988037,
      "learning_rate": 0.00015801576009005766,
      "loss": 1.2101,
      "step": 146950
    },
    {
      "epoch": 4.388976801122623,
      "grad_norm": 3.7732479572296143,
      "learning_rate": 0.00015800147429413883,
      "loss": 1.226,
      "step": 147000
    },
    {
      "epoch": 4.390469650374705,
      "grad_norm": 4.265769958496094,
      "learning_rate": 0.00015798718849822,
      "loss": 1.1803,
      "step": 147050
    },
    {
      "epoch": 4.391962499626787,
      "grad_norm": 5.331263065338135,
      "learning_rate": 0.00015797290270230116,
      "loss": 1.1799,
      "step": 147100
    },
    {
      "epoch": 4.39345534887887,
      "grad_norm": 4.761427879333496,
      "learning_rate": 0.00015795861690638235,
      "loss": 1.2091,
      "step": 147150
    },
    {
      "epoch": 4.394948198130953,
      "grad_norm": 4.461631774902344,
      "learning_rate": 0.0001579443311104635,
      "loss": 1.2328,
      "step": 147200
    },
    {
      "epoch": 4.3964410473830355,
      "grad_norm": 4.383672714233398,
      "learning_rate": 0.00015793004531454468,
      "loss": 1.2279,
      "step": 147250
    },
    {
      "epoch": 4.397933896635118,
      "grad_norm": 6.641529560089111,
      "learning_rate": 0.00015791575951862582,
      "loss": 1.2231,
      "step": 147300
    },
    {
      "epoch": 4.3994267458872,
      "grad_norm": 3.944716215133667,
      "learning_rate": 0.000157901473722707,
      "loss": 1.2308,
      "step": 147350
    },
    {
      "epoch": 4.400919595139283,
      "grad_norm": 3.688462495803833,
      "learning_rate": 0.00015788718792678817,
      "loss": 1.252,
      "step": 147400
    },
    {
      "epoch": 4.4024124443913655,
      "grad_norm": 5.539252281188965,
      "learning_rate": 0.00015787290213086934,
      "loss": 1.2468,
      "step": 147450
    },
    {
      "epoch": 4.403905293643448,
      "grad_norm": 5.409310340881348,
      "learning_rate": 0.0001578586163349505,
      "loss": 1.2176,
      "step": 147500
    },
    {
      "epoch": 4.40539814289553,
      "grad_norm": 4.958510875701904,
      "learning_rate": 0.00015784433053903164,
      "loss": 1.2498,
      "step": 147550
    },
    {
      "epoch": 4.406890992147613,
      "grad_norm": 5.825056076049805,
      "learning_rate": 0.00015783004474311283,
      "loss": 1.2347,
      "step": 147600
    },
    {
      "epoch": 4.408383841399695,
      "grad_norm": 6.798030853271484,
      "learning_rate": 0.00015781575894719397,
      "loss": 1.2459,
      "step": 147650
    },
    {
      "epoch": 4.409876690651778,
      "grad_norm": 4.517495632171631,
      "learning_rate": 0.00015780147315127516,
      "loss": 1.2076,
      "step": 147700
    },
    {
      "epoch": 4.411369539903861,
      "grad_norm": 4.233908653259277,
      "learning_rate": 0.00015778718735535633,
      "loss": 1.1905,
      "step": 147750
    },
    {
      "epoch": 4.412862389155943,
      "grad_norm": 4.206421375274658,
      "learning_rate": 0.0001577729015594375,
      "loss": 1.2357,
      "step": 147800
    },
    {
      "epoch": 4.414355238408025,
      "grad_norm": 4.130222797393799,
      "learning_rate": 0.00015775861576351866,
      "loss": 1.201,
      "step": 147850
    },
    {
      "epoch": 4.415848087660108,
      "grad_norm": 3.8784375190734863,
      "learning_rate": 0.00015774432996759982,
      "loss": 1.2536,
      "step": 147900
    },
    {
      "epoch": 4.417340936912191,
      "grad_norm": 5.533395290374756,
      "learning_rate": 0.00015773004417168098,
      "loss": 1.208,
      "step": 147950
    },
    {
      "epoch": 4.4188337861642735,
      "grad_norm": 5.0397419929504395,
      "learning_rate": 0.00015771575837576215,
      "loss": 1.1951,
      "step": 148000
    },
    {
      "epoch": 4.420326635416355,
      "grad_norm": 4.333613872528076,
      "learning_rate": 0.00015770147257984331,
      "loss": 1.196,
      "step": 148050
    },
    {
      "epoch": 4.421819484668438,
      "grad_norm": 5.430739879608154,
      "learning_rate": 0.00015768718678392448,
      "loss": 1.1636,
      "step": 148100
    },
    {
      "epoch": 4.423312333920521,
      "grad_norm": 4.669544219970703,
      "learning_rate": 0.00015767290098800564,
      "loss": 1.2302,
      "step": 148150
    },
    {
      "epoch": 4.424805183172603,
      "grad_norm": 4.452166557312012,
      "learning_rate": 0.00015765861519208683,
      "loss": 1.1997,
      "step": 148200
    },
    {
      "epoch": 4.426298032424686,
      "grad_norm": 3.695939779281616,
      "learning_rate": 0.00015764432939616797,
      "loss": 1.2327,
      "step": 148250
    },
    {
      "epoch": 4.427790881676768,
      "grad_norm": 5.5830535888671875,
      "learning_rate": 0.00015763004360024916,
      "loss": 1.2306,
      "step": 148300
    },
    {
      "epoch": 4.429283730928851,
      "grad_norm": 3.978583335876465,
      "learning_rate": 0.0001576157578043303,
      "loss": 1.1639,
      "step": 148350
    },
    {
      "epoch": 4.430776580180933,
      "grad_norm": 3.8745334148406982,
      "learning_rate": 0.0001576014720084115,
      "loss": 1.22,
      "step": 148400
    },
    {
      "epoch": 4.432269429433016,
      "grad_norm": 4.7659711837768555,
      "learning_rate": 0.00015758718621249263,
      "loss": 1.1871,
      "step": 148450
    },
    {
      "epoch": 4.433762278685099,
      "grad_norm": 4.240323066711426,
      "learning_rate": 0.00015757290041657382,
      "loss": 1.1782,
      "step": 148500
    },
    {
      "epoch": 4.435255127937181,
      "grad_norm": 4.188791751861572,
      "learning_rate": 0.000157558614620655,
      "loss": 1.2258,
      "step": 148550
    },
    {
      "epoch": 4.436747977189263,
      "grad_norm": 4.706462860107422,
      "learning_rate": 0.00015754432882473615,
      "loss": 1.2095,
      "step": 148600
    },
    {
      "epoch": 4.438240826441346,
      "grad_norm": 3.878901958465576,
      "learning_rate": 0.00015753004302881732,
      "loss": 1.2249,
      "step": 148650
    },
    {
      "epoch": 4.439733675693429,
      "grad_norm": 3.91372013092041,
      "learning_rate": 0.00015751575723289848,
      "loss": 1.2382,
      "step": 148700
    },
    {
      "epoch": 4.441226524945511,
      "grad_norm": 3.9671475887298584,
      "learning_rate": 0.00015750147143697965,
      "loss": 1.2729,
      "step": 148750
    },
    {
      "epoch": 4.442719374197593,
      "grad_norm": 4.4688591957092285,
      "learning_rate": 0.0001574871856410608,
      "loss": 1.2032,
      "step": 148800
    },
    {
      "epoch": 4.444212223449676,
      "grad_norm": 4.55964469909668,
      "learning_rate": 0.00015747289984514198,
      "loss": 1.2129,
      "step": 148850
    },
    {
      "epoch": 4.445705072701759,
      "grad_norm": 5.789206027984619,
      "learning_rate": 0.00015745861404922314,
      "loss": 1.1947,
      "step": 148900
    },
    {
      "epoch": 4.447197921953841,
      "grad_norm": 4.169732570648193,
      "learning_rate": 0.0001574443282533043,
      "loss": 1.2135,
      "step": 148950
    },
    {
      "epoch": 4.448690771205923,
      "grad_norm": 4.305688858032227,
      "learning_rate": 0.0001574300424573855,
      "loss": 1.1931,
      "step": 149000
    },
    {
      "epoch": 4.450183620458006,
      "grad_norm": 5.780201435089111,
      "learning_rate": 0.00015741575666146663,
      "loss": 1.2308,
      "step": 149050
    },
    {
      "epoch": 4.451676469710089,
      "grad_norm": 5.42836332321167,
      "learning_rate": 0.00015740147086554783,
      "loss": 1.1898,
      "step": 149100
    },
    {
      "epoch": 4.453169318962171,
      "grad_norm": 4.967898845672607,
      "learning_rate": 0.00015738718506962896,
      "loss": 1.1961,
      "step": 149150
    },
    {
      "epoch": 4.454662168214254,
      "grad_norm": 4.566812038421631,
      "learning_rate": 0.00015737289927371015,
      "loss": 1.2633,
      "step": 149200
    },
    {
      "epoch": 4.456155017466337,
      "grad_norm": 4.523815155029297,
      "learning_rate": 0.0001573586134777913,
      "loss": 1.2104,
      "step": 149250
    },
    {
      "epoch": 4.4576478667184185,
      "grad_norm": 4.3855180740356445,
      "learning_rate": 0.00015734432768187248,
      "loss": 1.1904,
      "step": 149300
    },
    {
      "epoch": 4.459140715970501,
      "grad_norm": 4.116260051727295,
      "learning_rate": 0.00015733004188595365,
      "loss": 1.2162,
      "step": 149350
    },
    {
      "epoch": 4.460633565222584,
      "grad_norm": 4.964864253997803,
      "learning_rate": 0.0001573157560900348,
      "loss": 1.233,
      "step": 149400
    },
    {
      "epoch": 4.462126414474667,
      "grad_norm": 4.709635257720947,
      "learning_rate": 0.00015730147029411598,
      "loss": 1.2226,
      "step": 149450
    },
    {
      "epoch": 4.463619263726748,
      "grad_norm": 5.310390949249268,
      "learning_rate": 0.00015728718449819714,
      "loss": 1.1655,
      "step": 149500
    },
    {
      "epoch": 4.465112112978831,
      "grad_norm": 5.156966209411621,
      "learning_rate": 0.0001572728987022783,
      "loss": 1.1738,
      "step": 149550
    },
    {
      "epoch": 4.466604962230914,
      "grad_norm": 4.476001739501953,
      "learning_rate": 0.00015725861290635947,
      "loss": 1.188,
      "step": 149600
    },
    {
      "epoch": 4.468097811482997,
      "grad_norm": 4.519705295562744,
      "learning_rate": 0.00015724432711044064,
      "loss": 1.1512,
      "step": 149650
    },
    {
      "epoch": 4.469590660735079,
      "grad_norm": 5.426435470581055,
      "learning_rate": 0.0001572300413145218,
      "loss": 1.2137,
      "step": 149700
    },
    {
      "epoch": 4.471083509987161,
      "grad_norm": 3.799715757369995,
      "learning_rate": 0.00015721575551860297,
      "loss": 1.2298,
      "step": 149750
    },
    {
      "epoch": 4.472576359239244,
      "grad_norm": 3.640909194946289,
      "learning_rate": 0.00015720146972268416,
      "loss": 1.1837,
      "step": 149800
    },
    {
      "epoch": 4.4740692084913265,
      "grad_norm": 3.8437955379486084,
      "learning_rate": 0.0001571871839267653,
      "loss": 1.2025,
      "step": 149850
    },
    {
      "epoch": 4.475562057743409,
      "grad_norm": 3.9769036769866943,
      "learning_rate": 0.0001571728981308465,
      "loss": 1.1918,
      "step": 149900
    },
    {
      "epoch": 4.477054906995492,
      "grad_norm": 4.485401630401611,
      "learning_rate": 0.00015715861233492762,
      "loss": 1.1588,
      "step": 149950
    },
    {
      "epoch": 4.478547756247574,
      "grad_norm": 4.401375770568848,
      "learning_rate": 0.00015714432653900882,
      "loss": 1.2089,
      "step": 150000
    },
    {
      "epoch": 4.480040605499656,
      "grad_norm": 4.181282043457031,
      "learning_rate": 0.00015713004074308998,
      "loss": 1.2139,
      "step": 150050
    },
    {
      "epoch": 4.481533454751739,
      "grad_norm": 4.993724822998047,
      "learning_rate": 0.00015711575494717112,
      "loss": 1.2247,
      "step": 150100
    },
    {
      "epoch": 4.483026304003822,
      "grad_norm": 4.870810031890869,
      "learning_rate": 0.0001571014691512523,
      "loss": 1.281,
      "step": 150150
    },
    {
      "epoch": 4.4845191532559046,
      "grad_norm": 5.145561218261719,
      "learning_rate": 0.00015708718335533345,
      "loss": 1.2263,
      "step": 150200
    },
    {
      "epoch": 4.486012002507986,
      "grad_norm": 4.650093078613281,
      "learning_rate": 0.00015707289755941464,
      "loss": 1.2627,
      "step": 150250
    },
    {
      "epoch": 4.487504851760069,
      "grad_norm": 4.286979675292969,
      "learning_rate": 0.00015705861176349578,
      "loss": 1.1671,
      "step": 150300
    },
    {
      "epoch": 4.488997701012152,
      "grad_norm": 4.116031646728516,
      "learning_rate": 0.00015704432596757697,
      "loss": 1.1746,
      "step": 150350
    },
    {
      "epoch": 4.4904905502642345,
      "grad_norm": 4.999283313751221,
      "learning_rate": 0.00015703004017165813,
      "loss": 1.1993,
      "step": 150400
    },
    {
      "epoch": 4.491983399516317,
      "grad_norm": 5.618175983428955,
      "learning_rate": 0.0001570157543757393,
      "loss": 1.2608,
      "step": 150450
    },
    {
      "epoch": 4.493476248768399,
      "grad_norm": 4.471452236175537,
      "learning_rate": 0.00015700146857982046,
      "loss": 1.1811,
      "step": 150500
    },
    {
      "epoch": 4.494969098020482,
      "grad_norm": 4.4648518562316895,
      "learning_rate": 0.00015698718278390163,
      "loss": 1.2314,
      "step": 150550
    },
    {
      "epoch": 4.496461947272564,
      "grad_norm": 5.73147439956665,
      "learning_rate": 0.0001569728969879828,
      "loss": 1.243,
      "step": 150600
    },
    {
      "epoch": 4.497954796524647,
      "grad_norm": 6.473092555999756,
      "learning_rate": 0.00015695861119206396,
      "loss": 1.2208,
      "step": 150650
    },
    {
      "epoch": 4.49944764577673,
      "grad_norm": 6.348034381866455,
      "learning_rate": 0.00015694432539614512,
      "loss": 1.214,
      "step": 150700
    },
    {
      "epoch": 4.500940495028812,
      "grad_norm": 4.37158727645874,
      "learning_rate": 0.00015693003960022629,
      "loss": 1.2777,
      "step": 150750
    },
    {
      "epoch": 4.502433344280894,
      "grad_norm": 4.451486110687256,
      "learning_rate": 0.00015691575380430745,
      "loss": 1.1995,
      "step": 150800
    },
    {
      "epoch": 4.503926193532977,
      "grad_norm": 4.328061580657959,
      "learning_rate": 0.00015690146800838864,
      "loss": 1.225,
      "step": 150850
    },
    {
      "epoch": 4.50541904278506,
      "grad_norm": 5.565524578094482,
      "learning_rate": 0.00015688718221246978,
      "loss": 1.2471,
      "step": 150900
    },
    {
      "epoch": 4.506911892037142,
      "grad_norm": 4.532073497772217,
      "learning_rate": 0.00015687289641655097,
      "loss": 1.2396,
      "step": 150950
    },
    {
      "epoch": 4.508404741289224,
      "grad_norm": 5.4925618171691895,
      "learning_rate": 0.0001568586106206321,
      "loss": 1.2135,
      "step": 151000
    },
    {
      "epoch": 4.509897590541307,
      "grad_norm": 5.885026931762695,
      "learning_rate": 0.0001568443248247133,
      "loss": 1.2067,
      "step": 151050
    },
    {
      "epoch": 4.51139043979339,
      "grad_norm": 4.559401988983154,
      "learning_rate": 0.00015683003902879444,
      "loss": 1.1973,
      "step": 151100
    },
    {
      "epoch": 4.512883289045472,
      "grad_norm": 2.85530686378479,
      "learning_rate": 0.00015681575323287563,
      "loss": 1.2574,
      "step": 151150
    },
    {
      "epoch": 4.514376138297555,
      "grad_norm": 3.6753365993499756,
      "learning_rate": 0.0001568014674369568,
      "loss": 1.2358,
      "step": 151200
    },
    {
      "epoch": 4.515868987549637,
      "grad_norm": 4.216865539550781,
      "learning_rate": 0.00015678718164103796,
      "loss": 1.2624,
      "step": 151250
    },
    {
      "epoch": 4.51736183680172,
      "grad_norm": 4.118535995483398,
      "learning_rate": 0.00015677289584511912,
      "loss": 1.2322,
      "step": 151300
    },
    {
      "epoch": 4.518854686053802,
      "grad_norm": 3.65169095993042,
      "learning_rate": 0.0001567586100492003,
      "loss": 1.1924,
      "step": 151350
    },
    {
      "epoch": 4.520347535305885,
      "grad_norm": 4.938554763793945,
      "learning_rate": 0.00015674432425328145,
      "loss": 1.2083,
      "step": 151400
    },
    {
      "epoch": 4.521840384557967,
      "grad_norm": 5.959898471832275,
      "learning_rate": 0.00015673003845736262,
      "loss": 1.3091,
      "step": 151450
    },
    {
      "epoch": 4.52333323381005,
      "grad_norm": 4.388880252838135,
      "learning_rate": 0.00015671575266144378,
      "loss": 1.2481,
      "step": 151500
    },
    {
      "epoch": 4.524826083062132,
      "grad_norm": 4.406062602996826,
      "learning_rate": 0.00015670146686552495,
      "loss": 1.2337,
      "step": 151550
    },
    {
      "epoch": 4.526318932314215,
      "grad_norm": 3.515559196472168,
      "learning_rate": 0.0001566871810696061,
      "loss": 1.1663,
      "step": 151600
    },
    {
      "epoch": 4.527811781566298,
      "grad_norm": 5.7760701179504395,
      "learning_rate": 0.0001566728952736873,
      "loss": 1.1952,
      "step": 151650
    },
    {
      "epoch": 4.52930463081838,
      "grad_norm": 3.783639907836914,
      "learning_rate": 0.00015665860947776844,
      "loss": 1.2086,
      "step": 151700
    },
    {
      "epoch": 4.530797480070462,
      "grad_norm": 4.640422344207764,
      "learning_rate": 0.00015664432368184963,
      "loss": 1.2152,
      "step": 151750
    },
    {
      "epoch": 4.532290329322545,
      "grad_norm": 5.015472412109375,
      "learning_rate": 0.00015663003788593077,
      "loss": 1.2133,
      "step": 151800
    },
    {
      "epoch": 4.533783178574628,
      "grad_norm": 5.7063117027282715,
      "learning_rate": 0.00015661575209001196,
      "loss": 1.2345,
      "step": 151850
    },
    {
      "epoch": 4.53527602782671,
      "grad_norm": 4.530743598937988,
      "learning_rate": 0.0001566014662940931,
      "loss": 1.2408,
      "step": 151900
    },
    {
      "epoch": 4.536768877078792,
      "grad_norm": 3.597856044769287,
      "learning_rate": 0.0001565871804981743,
      "loss": 1.2222,
      "step": 151950
    },
    {
      "epoch": 4.538261726330875,
      "grad_norm": 5.361748218536377,
      "learning_rate": 0.00015657289470225546,
      "loss": 1.1898,
      "step": 152000
    },
    {
      "epoch": 4.539754575582958,
      "grad_norm": 6.137198448181152,
      "learning_rate": 0.00015655860890633662,
      "loss": 1.1988,
      "step": 152050
    },
    {
      "epoch": 4.54124742483504,
      "grad_norm": 4.32033634185791,
      "learning_rate": 0.00015654432311041779,
      "loss": 1.2221,
      "step": 152100
    },
    {
      "epoch": 4.542740274087123,
      "grad_norm": 4.112492084503174,
      "learning_rate": 0.00015653003731449895,
      "loss": 1.229,
      "step": 152150
    },
    {
      "epoch": 4.544233123339206,
      "grad_norm": 4.592637062072754,
      "learning_rate": 0.00015651575151858012,
      "loss": 1.2376,
      "step": 152200
    },
    {
      "epoch": 4.5457259725912875,
      "grad_norm": 4.471153736114502,
      "learning_rate": 0.00015650146572266128,
      "loss": 1.1976,
      "step": 152250
    },
    {
      "epoch": 4.54721882184337,
      "grad_norm": 6.161034107208252,
      "learning_rate": 0.00015648717992674244,
      "loss": 1.1973,
      "step": 152300
    },
    {
      "epoch": 4.548711671095453,
      "grad_norm": 4.57336950302124,
      "learning_rate": 0.0001564728941308236,
      "loss": 1.1773,
      "step": 152350
    },
    {
      "epoch": 4.550204520347536,
      "grad_norm": 3.936795949935913,
      "learning_rate": 0.00015645860833490477,
      "loss": 1.2448,
      "step": 152400
    },
    {
      "epoch": 4.5516973695996175,
      "grad_norm": 3.8005549907684326,
      "learning_rate": 0.00015644432253898597,
      "loss": 1.2391,
      "step": 152450
    },
    {
      "epoch": 4.5531902188517,
      "grad_norm": 4.863176345825195,
      "learning_rate": 0.0001564300367430671,
      "loss": 1.2185,
      "step": 152500
    },
    {
      "epoch": 4.554683068103783,
      "grad_norm": 4.168069362640381,
      "learning_rate": 0.0001564157509471483,
      "loss": 1.1793,
      "step": 152550
    },
    {
      "epoch": 4.556175917355866,
      "grad_norm": 4.241150856018066,
      "learning_rate": 0.00015640146515122943,
      "loss": 1.2087,
      "step": 152600
    },
    {
      "epoch": 4.557668766607948,
      "grad_norm": 5.077085971832275,
      "learning_rate": 0.00015638717935531062,
      "loss": 1.1999,
      "step": 152650
    },
    {
      "epoch": 4.55916161586003,
      "grad_norm": 4.880026817321777,
      "learning_rate": 0.00015637289355939176,
      "loss": 1.1691,
      "step": 152700
    },
    {
      "epoch": 4.560654465112113,
      "grad_norm": 4.0811662673950195,
      "learning_rate": 0.00015635860776347293,
      "loss": 1.2294,
      "step": 152750
    },
    {
      "epoch": 4.5621473143641955,
      "grad_norm": 3.9246151447296143,
      "learning_rate": 0.00015634432196755412,
      "loss": 1.2704,
      "step": 152800
    },
    {
      "epoch": 4.563640163616278,
      "grad_norm": 4.361538887023926,
      "learning_rate": 0.00015633003617163526,
      "loss": 1.2227,
      "step": 152850
    },
    {
      "epoch": 4.565133012868361,
      "grad_norm": 3.964625358581543,
      "learning_rate": 0.00015631575037571645,
      "loss": 1.2269,
      "step": 152900
    },
    {
      "epoch": 4.566625862120443,
      "grad_norm": 4.382391452789307,
      "learning_rate": 0.00015630146457979759,
      "loss": 1.2066,
      "step": 152950
    },
    {
      "epoch": 4.5681187113725255,
      "grad_norm": 4.823386192321777,
      "learning_rate": 0.00015628717878387878,
      "loss": 1.2191,
      "step": 153000
    },
    {
      "epoch": 4.569611560624608,
      "grad_norm": 4.2271223068237305,
      "learning_rate": 0.00015627289298795994,
      "loss": 1.2611,
      "step": 153050
    },
    {
      "epoch": 4.571104409876691,
      "grad_norm": 5.857795238494873,
      "learning_rate": 0.0001562586071920411,
      "loss": 1.2222,
      "step": 153100
    },
    {
      "epoch": 4.572597259128774,
      "grad_norm": 3.7787981033325195,
      "learning_rate": 0.00015624432139612227,
      "loss": 1.2648,
      "step": 153150
    },
    {
      "epoch": 4.574090108380855,
      "grad_norm": 4.284102916717529,
      "learning_rate": 0.00015623003560020344,
      "loss": 1.2814,
      "step": 153200
    },
    {
      "epoch": 4.575582957632938,
      "grad_norm": 6.498587131500244,
      "learning_rate": 0.0001562157498042846,
      "loss": 1.209,
      "step": 153250
    },
    {
      "epoch": 4.577075806885021,
      "grad_norm": 3.2543442249298096,
      "learning_rate": 0.00015620146400836577,
      "loss": 1.1867,
      "step": 153300
    },
    {
      "epoch": 4.5785686561371035,
      "grad_norm": 3.8081936836242676,
      "learning_rate": 0.00015618717821244693,
      "loss": 1.2242,
      "step": 153350
    },
    {
      "epoch": 4.580061505389185,
      "grad_norm": 3.8605005741119385,
      "learning_rate": 0.0001561728924165281,
      "loss": 1.1957,
      "step": 153400
    },
    {
      "epoch": 4.581554354641268,
      "grad_norm": 6.710892677307129,
      "learning_rate": 0.00015615860662060926,
      "loss": 1.1945,
      "step": 153450
    },
    {
      "epoch": 4.583047203893351,
      "grad_norm": 4.212560176849365,
      "learning_rate": 0.00015614432082469045,
      "loss": 1.2124,
      "step": 153500
    },
    {
      "epoch": 4.5845400531454334,
      "grad_norm": 4.561640739440918,
      "learning_rate": 0.0001561300350287716,
      "loss": 1.2355,
      "step": 153550
    },
    {
      "epoch": 4.586032902397516,
      "grad_norm": 3.4632205963134766,
      "learning_rate": 0.00015611574923285278,
      "loss": 1.2606,
      "step": 153600
    },
    {
      "epoch": 4.587525751649599,
      "grad_norm": 3.846022844314575,
      "learning_rate": 0.00015610146343693392,
      "loss": 1.2386,
      "step": 153650
    },
    {
      "epoch": 4.589018600901681,
      "grad_norm": 4.297937870025635,
      "learning_rate": 0.0001560871776410151,
      "loss": 1.2084,
      "step": 153700
    },
    {
      "epoch": 4.590511450153763,
      "grad_norm": 3.9638116359710693,
      "learning_rate": 0.00015607289184509625,
      "loss": 1.2334,
      "step": 153750
    },
    {
      "epoch": 4.592004299405846,
      "grad_norm": 5.708649158477783,
      "learning_rate": 0.00015605860604917744,
      "loss": 1.1872,
      "step": 153800
    },
    {
      "epoch": 4.593497148657929,
      "grad_norm": 4.930756092071533,
      "learning_rate": 0.0001560443202532586,
      "loss": 1.2585,
      "step": 153850
    },
    {
      "epoch": 4.594989997910011,
      "grad_norm": 4.887782573699951,
      "learning_rate": 0.00015603003445733977,
      "loss": 1.2325,
      "step": 153900
    },
    {
      "epoch": 4.596482847162093,
      "grad_norm": 5.5042290687561035,
      "learning_rate": 0.00015601574866142093,
      "loss": 1.261,
      "step": 153950
    },
    {
      "epoch": 4.597975696414176,
      "grad_norm": 5.776236534118652,
      "learning_rate": 0.0001560014628655021,
      "loss": 1.205,
      "step": 154000
    },
    {
      "epoch": 4.599468545666259,
      "grad_norm": 4.465933799743652,
      "learning_rate": 0.00015598717706958326,
      "loss": 1.2886,
      "step": 154050
    },
    {
      "epoch": 4.600961394918341,
      "grad_norm": 4.536434173583984,
      "learning_rate": 0.00015597289127366443,
      "loss": 1.2107,
      "step": 154100
    },
    {
      "epoch": 4.602454244170424,
      "grad_norm": 6.132566452026367,
      "learning_rate": 0.0001559586054777456,
      "loss": 1.2365,
      "step": 154150
    },
    {
      "epoch": 4.603947093422506,
      "grad_norm": 5.581535816192627,
      "learning_rate": 0.00015594431968182676,
      "loss": 1.2163,
      "step": 154200
    },
    {
      "epoch": 4.605439942674589,
      "grad_norm": 4.754735946655273,
      "learning_rate": 0.00015593003388590792,
      "loss": 1.1939,
      "step": 154250
    },
    {
      "epoch": 4.606932791926671,
      "grad_norm": 4.529079914093018,
      "learning_rate": 0.0001559157480899891,
      "loss": 1.2434,
      "step": 154300
    },
    {
      "epoch": 4.608425641178754,
      "grad_norm": 3.8123152256011963,
      "learning_rate": 0.00015590146229407025,
      "loss": 1.1727,
      "step": 154350
    },
    {
      "epoch": 4.609918490430836,
      "grad_norm": 7.249644756317139,
      "learning_rate": 0.00015588717649815144,
      "loss": 1.2392,
      "step": 154400
    },
    {
      "epoch": 4.611411339682919,
      "grad_norm": 5.355637073516846,
      "learning_rate": 0.00015587289070223258,
      "loss": 1.1899,
      "step": 154450
    },
    {
      "epoch": 4.612904188935001,
      "grad_norm": 3.952632427215576,
      "learning_rate": 0.00015585860490631377,
      "loss": 1.248,
      "step": 154500
    },
    {
      "epoch": 4.614397038187084,
      "grad_norm": 5.123134136199951,
      "learning_rate": 0.0001558443191103949,
      "loss": 1.1783,
      "step": 154550
    },
    {
      "epoch": 4.615889887439167,
      "grad_norm": 5.032320022583008,
      "learning_rate": 0.0001558300333144761,
      "loss": 1.2189,
      "step": 154600
    },
    {
      "epoch": 4.6173827366912485,
      "grad_norm": 4.798953533172607,
      "learning_rate": 0.00015581574751855726,
      "loss": 1.2064,
      "step": 154650
    },
    {
      "epoch": 4.618875585943331,
      "grad_norm": 4.720959663391113,
      "learning_rate": 0.00015580146172263843,
      "loss": 1.2685,
      "step": 154700
    },
    {
      "epoch": 4.620368435195414,
      "grad_norm": 5.178957939147949,
      "learning_rate": 0.0001557871759267196,
      "loss": 1.2969,
      "step": 154750
    },
    {
      "epoch": 4.621861284447497,
      "grad_norm": 2.8498313426971436,
      "learning_rate": 0.00015577289013080076,
      "loss": 1.2293,
      "step": 154800
    },
    {
      "epoch": 4.623354133699579,
      "grad_norm": 4.07916784286499,
      "learning_rate": 0.00015575860433488192,
      "loss": 1.2489,
      "step": 154850
    },
    {
      "epoch": 4.624846982951661,
      "grad_norm": 4.014176845550537,
      "learning_rate": 0.0001557443185389631,
      "loss": 1.2244,
      "step": 154900
    },
    {
      "epoch": 4.626339832203744,
      "grad_norm": 6.280478477478027,
      "learning_rate": 0.00015573003274304425,
      "loss": 1.2029,
      "step": 154950
    },
    {
      "epoch": 4.627832681455827,
      "grad_norm": 4.045797348022461,
      "learning_rate": 0.00015571574694712542,
      "loss": 1.1694,
      "step": 155000
    },
    {
      "epoch": 4.629325530707909,
      "grad_norm": 4.864063262939453,
      "learning_rate": 0.00015570146115120658,
      "loss": 1.1799,
      "step": 155050
    },
    {
      "epoch": 4.630818379959992,
      "grad_norm": 3.5234014987945557,
      "learning_rate": 0.00015568717535528777,
      "loss": 1.2153,
      "step": 155100
    },
    {
      "epoch": 4.632311229212074,
      "grad_norm": 4.5075273513793945,
      "learning_rate": 0.0001556728895593689,
      "loss": 1.1483,
      "step": 155150
    },
    {
      "epoch": 4.6338040784641565,
      "grad_norm": 5.258499622344971,
      "learning_rate": 0.0001556586037634501,
      "loss": 1.2606,
      "step": 155200
    },
    {
      "epoch": 4.635296927716239,
      "grad_norm": 5.108059883117676,
      "learning_rate": 0.00015564431796753124,
      "loss": 1.2517,
      "step": 155250
    },
    {
      "epoch": 4.636789776968322,
      "grad_norm": 4.033377647399902,
      "learning_rate": 0.00015563003217161243,
      "loss": 1.1887,
      "step": 155300
    },
    {
      "epoch": 4.638282626220404,
      "grad_norm": 5.546184062957764,
      "learning_rate": 0.00015561574637569357,
      "loss": 1.1573,
      "step": 155350
    },
    {
      "epoch": 4.6397754754724865,
      "grad_norm": 5.48836088180542,
      "learning_rate": 0.00015560146057977473,
      "loss": 1.2637,
      "step": 155400
    },
    {
      "epoch": 4.641268324724569,
      "grad_norm": 5.246400833129883,
      "learning_rate": 0.00015558717478385593,
      "loss": 1.2008,
      "step": 155450
    },
    {
      "epoch": 4.642761173976652,
      "grad_norm": 4.412383556365967,
      "learning_rate": 0.00015557288898793706,
      "loss": 1.168,
      "step": 155500
    },
    {
      "epoch": 4.644254023228735,
      "grad_norm": 3.9869327545166016,
      "learning_rate": 0.00015555860319201826,
      "loss": 1.1371,
      "step": 155550
    },
    {
      "epoch": 4.645746872480817,
      "grad_norm": 3.883467197418213,
      "learning_rate": 0.0001555443173960994,
      "loss": 1.216,
      "step": 155600
    },
    {
      "epoch": 4.647239721732899,
      "grad_norm": 4.335739612579346,
      "learning_rate": 0.00015553003160018059,
      "loss": 1.1792,
      "step": 155650
    },
    {
      "epoch": 4.648732570984982,
      "grad_norm": 4.266781806945801,
      "learning_rate": 0.00015551574580426175,
      "loss": 1.2357,
      "step": 155700
    },
    {
      "epoch": 4.6502254202370645,
      "grad_norm": 4.26521635055542,
      "learning_rate": 0.00015550146000834291,
      "loss": 1.2613,
      "step": 155750
    },
    {
      "epoch": 4.651718269489147,
      "grad_norm": 4.667625427246094,
      "learning_rate": 0.00015548717421242408,
      "loss": 1.213,
      "step": 155800
    },
    {
      "epoch": 4.653211118741229,
      "grad_norm": 4.781721591949463,
      "learning_rate": 0.00015547288841650524,
      "loss": 1.2303,
      "step": 155850
    },
    {
      "epoch": 4.654703967993312,
      "grad_norm": 4.377721309661865,
      "learning_rate": 0.0001554586026205864,
      "loss": 1.2062,
      "step": 155900
    },
    {
      "epoch": 4.6561968172453945,
      "grad_norm": 3.987110137939453,
      "learning_rate": 0.00015544431682466757,
      "loss": 1.1393,
      "step": 155950
    },
    {
      "epoch": 4.657689666497477,
      "grad_norm": 5.804355144500732,
      "learning_rate": 0.00015543003102874874,
      "loss": 1.2368,
      "step": 156000
    },
    {
      "epoch": 4.65918251574956,
      "grad_norm": 4.505835056304932,
      "learning_rate": 0.0001554157452328299,
      "loss": 1.2318,
      "step": 156050
    },
    {
      "epoch": 4.660675365001643,
      "grad_norm": 5.454282760620117,
      "learning_rate": 0.00015540145943691107,
      "loss": 1.2058,
      "step": 156100
    },
    {
      "epoch": 4.662168214253724,
      "grad_norm": 4.576279163360596,
      "learning_rate": 0.00015538717364099223,
      "loss": 1.1932,
      "step": 156150
    },
    {
      "epoch": 4.663661063505807,
      "grad_norm": 4.093862533569336,
      "learning_rate": 0.0001553728878450734,
      "loss": 1.2505,
      "step": 156200
    },
    {
      "epoch": 4.66515391275789,
      "grad_norm": 4.815280437469482,
      "learning_rate": 0.0001553586020491546,
      "loss": 1.3025,
      "step": 156250
    },
    {
      "epoch": 4.6666467620099725,
      "grad_norm": 5.615932464599609,
      "learning_rate": 0.00015534431625323573,
      "loss": 1.2428,
      "step": 156300
    },
    {
      "epoch": 4.668139611262054,
      "grad_norm": 4.396317005157471,
      "learning_rate": 0.00015533003045731692,
      "loss": 1.2452,
      "step": 156350
    },
    {
      "epoch": 4.669632460514137,
      "grad_norm": 5.061831951141357,
      "learning_rate": 0.00015531574466139805,
      "loss": 1.2262,
      "step": 156400
    },
    {
      "epoch": 4.67112530976622,
      "grad_norm": 4.902180194854736,
      "learning_rate": 0.00015530145886547925,
      "loss": 1.2485,
      "step": 156450
    },
    {
      "epoch": 4.6726181590183025,
      "grad_norm": 5.020074367523193,
      "learning_rate": 0.0001552871730695604,
      "loss": 1.2692,
      "step": 156500
    },
    {
      "epoch": 4.674111008270385,
      "grad_norm": 4.696333885192871,
      "learning_rate": 0.00015527288727364158,
      "loss": 1.2462,
      "step": 156550
    },
    {
      "epoch": 4.675603857522468,
      "grad_norm": 4.284012794494629,
      "learning_rate": 0.00015525860147772274,
      "loss": 1.1821,
      "step": 156600
    },
    {
      "epoch": 4.67709670677455,
      "grad_norm": 5.056790828704834,
      "learning_rate": 0.0001552443156818039,
      "loss": 1.2571,
      "step": 156650
    },
    {
      "epoch": 4.678589556026632,
      "grad_norm": 5.424617767333984,
      "learning_rate": 0.00015523002988588507,
      "loss": 1.2369,
      "step": 156700
    },
    {
      "epoch": 4.680082405278715,
      "grad_norm": 4.2647624015808105,
      "learning_rate": 0.00015521574408996623,
      "loss": 1.1531,
      "step": 156750
    },
    {
      "epoch": 4.681575254530798,
      "grad_norm": 5.538541316986084,
      "learning_rate": 0.0001552014582940474,
      "loss": 1.1512,
      "step": 156800
    },
    {
      "epoch": 4.68306810378288,
      "grad_norm": 5.385500907897949,
      "learning_rate": 0.00015518717249812856,
      "loss": 1.1912,
      "step": 156850
    },
    {
      "epoch": 4.684560953034962,
      "grad_norm": 4.456097602844238,
      "learning_rate": 0.00015517288670220973,
      "loss": 1.266,
      "step": 156900
    },
    {
      "epoch": 4.686053802287045,
      "grad_norm": 4.176798343658447,
      "learning_rate": 0.00015515860090629092,
      "loss": 1.1689,
      "step": 156950
    },
    {
      "epoch": 4.687546651539128,
      "grad_norm": 4.8275980949401855,
      "learning_rate": 0.00015514431511037206,
      "loss": 1.2168,
      "step": 157000
    },
    {
      "epoch": 4.6890395007912105,
      "grad_norm": 4.710529327392578,
      "learning_rate": 0.00015513002931445325,
      "loss": 1.2356,
      "step": 157050
    },
    {
      "epoch": 4.690532350043292,
      "grad_norm": 3.7281603813171387,
      "learning_rate": 0.0001551157435185344,
      "loss": 1.1486,
      "step": 157100
    },
    {
      "epoch": 4.692025199295375,
      "grad_norm": 5.062285900115967,
      "learning_rate": 0.00015510145772261558,
      "loss": 1.215,
      "step": 157150
    },
    {
      "epoch": 4.693518048547458,
      "grad_norm": 3.41444993019104,
      "learning_rate": 0.00015508717192669672,
      "loss": 1.2052,
      "step": 157200
    },
    {
      "epoch": 4.69501089779954,
      "grad_norm": 3.7359328269958496,
      "learning_rate": 0.0001550728861307779,
      "loss": 1.2664,
      "step": 157250
    },
    {
      "epoch": 4.696503747051623,
      "grad_norm": 4.192359924316406,
      "learning_rate": 0.00015505860033485907,
      "loss": 1.2171,
      "step": 157300
    },
    {
      "epoch": 4.697996596303705,
      "grad_norm": 5.198591232299805,
      "learning_rate": 0.00015504431453894024,
      "loss": 1.2094,
      "step": 157350
    },
    {
      "epoch": 4.699489445555788,
      "grad_norm": 4.99566650390625,
      "learning_rate": 0.0001550300287430214,
      "loss": 1.1983,
      "step": 157400
    },
    {
      "epoch": 4.70098229480787,
      "grad_norm": 4.944712162017822,
      "learning_rate": 0.00015501574294710257,
      "loss": 1.2355,
      "step": 157450
    },
    {
      "epoch": 4.702475144059953,
      "grad_norm": 4.957249164581299,
      "learning_rate": 0.00015500145715118373,
      "loss": 1.1986,
      "step": 157500
    },
    {
      "epoch": 4.703967993312036,
      "grad_norm": 3.442450523376465,
      "learning_rate": 0.0001549871713552649,
      "loss": 1.2072,
      "step": 157550
    },
    {
      "epoch": 4.705460842564118,
      "grad_norm": 4.920107364654541,
      "learning_rate": 0.00015497288555934606,
      "loss": 1.2014,
      "step": 157600
    },
    {
      "epoch": 4.7069536918162,
      "grad_norm": 3.697744131088257,
      "learning_rate": 0.00015495859976342723,
      "loss": 1.1869,
      "step": 157650
    },
    {
      "epoch": 4.708446541068283,
      "grad_norm": 6.023653984069824,
      "learning_rate": 0.0001549443139675084,
      "loss": 1.2255,
      "step": 157700
    },
    {
      "epoch": 4.709939390320366,
      "grad_norm": 4.989515781402588,
      "learning_rate": 0.00015493002817158958,
      "loss": 1.1685,
      "step": 157750
    },
    {
      "epoch": 4.7114322395724475,
      "grad_norm": 3.7004759311676025,
      "learning_rate": 0.00015491574237567072,
      "loss": 1.1584,
      "step": 157800
    },
    {
      "epoch": 4.71292508882453,
      "grad_norm": 4.318152904510498,
      "learning_rate": 0.0001549014565797519,
      "loss": 1.2163,
      "step": 157850
    },
    {
      "epoch": 4.714417938076613,
      "grad_norm": 4.550259590148926,
      "learning_rate": 0.00015488717078383305,
      "loss": 1.2507,
      "step": 157900
    },
    {
      "epoch": 4.715910787328696,
      "grad_norm": 4.413512706756592,
      "learning_rate": 0.00015487288498791424,
      "loss": 1.2037,
      "step": 157950
    },
    {
      "epoch": 4.717403636580778,
      "grad_norm": 3.8001725673675537,
      "learning_rate": 0.00015485859919199538,
      "loss": 1.2543,
      "step": 158000
    },
    {
      "epoch": 4.718896485832861,
      "grad_norm": 3.4299263954162598,
      "learning_rate": 0.00015484431339607654,
      "loss": 1.1916,
      "step": 158050
    },
    {
      "epoch": 4.720389335084943,
      "grad_norm": 4.620875835418701,
      "learning_rate": 0.00015483002760015773,
      "loss": 1.2697,
      "step": 158100
    },
    {
      "epoch": 4.721882184337026,
      "grad_norm": 4.62977409362793,
      "learning_rate": 0.00015481574180423887,
      "loss": 1.1801,
      "step": 158150
    },
    {
      "epoch": 4.723375033589108,
      "grad_norm": 3.916433572769165,
      "learning_rate": 0.00015480145600832006,
      "loss": 1.1801,
      "step": 158200
    },
    {
      "epoch": 4.724867882841191,
      "grad_norm": 4.217404842376709,
      "learning_rate": 0.0001547871702124012,
      "loss": 1.2328,
      "step": 158250
    },
    {
      "epoch": 4.726360732093273,
      "grad_norm": 6.347287654876709,
      "learning_rate": 0.0001547728844164824,
      "loss": 1.207,
      "step": 158300
    },
    {
      "epoch": 4.7278535813453555,
      "grad_norm": 4.39561653137207,
      "learning_rate": 0.00015475859862056353,
      "loss": 1.1908,
      "step": 158350
    },
    {
      "epoch": 4.729346430597438,
      "grad_norm": 4.472975730895996,
      "learning_rate": 0.00015474431282464472,
      "loss": 1.2007,
      "step": 158400
    },
    {
      "epoch": 4.730839279849521,
      "grad_norm": 4.133266448974609,
      "learning_rate": 0.0001547300270287259,
      "loss": 1.2425,
      "step": 158450
    },
    {
      "epoch": 4.732332129101604,
      "grad_norm": 3.9597208499908447,
      "learning_rate": 0.00015471574123280705,
      "loss": 1.2095,
      "step": 158500
    },
    {
      "epoch": 4.733824978353686,
      "grad_norm": 4.040492057800293,
      "learning_rate": 0.00015470145543688822,
      "loss": 1.1701,
      "step": 158550
    },
    {
      "epoch": 4.735317827605768,
      "grad_norm": 4.925788402557373,
      "learning_rate": 0.00015468716964096938,
      "loss": 1.2441,
      "step": 158600
    },
    {
      "epoch": 4.736810676857851,
      "grad_norm": 3.6005380153656006,
      "learning_rate": 0.00015467288384505055,
      "loss": 1.2654,
      "step": 158650
    },
    {
      "epoch": 4.7383035261099336,
      "grad_norm": 4.446198463439941,
      "learning_rate": 0.0001546585980491317,
      "loss": 1.2385,
      "step": 158700
    },
    {
      "epoch": 4.739796375362016,
      "grad_norm": 4.482244491577148,
      "learning_rate": 0.00015464431225321288,
      "loss": 1.2554,
      "step": 158750
    },
    {
      "epoch": 4.741289224614098,
      "grad_norm": 3.8426311016082764,
      "learning_rate": 0.00015463002645729404,
      "loss": 1.1721,
      "step": 158800
    },
    {
      "epoch": 4.742782073866181,
      "grad_norm": 5.686476707458496,
      "learning_rate": 0.0001546157406613752,
      "loss": 1.2192,
      "step": 158850
    },
    {
      "epoch": 4.7442749231182635,
      "grad_norm": 5.65896463394165,
      "learning_rate": 0.0001546014548654564,
      "loss": 1.1794,
      "step": 158900
    },
    {
      "epoch": 4.745767772370346,
      "grad_norm": 7.737217426300049,
      "learning_rate": 0.00015458716906953753,
      "loss": 1.259,
      "step": 158950
    },
    {
      "epoch": 4.747260621622429,
      "grad_norm": 5.426517963409424,
      "learning_rate": 0.00015457288327361873,
      "loss": 1.2547,
      "step": 159000
    },
    {
      "epoch": 4.748753470874511,
      "grad_norm": 4.5140509605407715,
      "learning_rate": 0.00015455859747769986,
      "loss": 1.2253,
      "step": 159050
    },
    {
      "epoch": 4.750246320126593,
      "grad_norm": 4.854976177215576,
      "learning_rate": 0.00015454431168178105,
      "loss": 1.2273,
      "step": 159100
    },
    {
      "epoch": 4.751739169378676,
      "grad_norm": 5.603672504425049,
      "learning_rate": 0.00015453002588586222,
      "loss": 1.2483,
      "step": 159150
    },
    {
      "epoch": 4.753232018630759,
      "grad_norm": 5.602535247802734,
      "learning_rate": 0.00015451574008994338,
      "loss": 1.2304,
      "step": 159200
    },
    {
      "epoch": 4.7547248678828415,
      "grad_norm": 5.0422773361206055,
      "learning_rate": 0.00015450145429402455,
      "loss": 1.2187,
      "step": 159250
    },
    {
      "epoch": 4.756217717134923,
      "grad_norm": 4.779214859008789,
      "learning_rate": 0.0001544871684981057,
      "loss": 1.1922,
      "step": 159300
    },
    {
      "epoch": 4.757710566387006,
      "grad_norm": 3.4088544845581055,
      "learning_rate": 0.00015447288270218688,
      "loss": 1.2072,
      "step": 159350
    },
    {
      "epoch": 4.759203415639089,
      "grad_norm": 5.209439277648926,
      "learning_rate": 0.00015445859690626804,
      "loss": 1.2119,
      "step": 159400
    },
    {
      "epoch": 4.7606962648911715,
      "grad_norm": 4.001313209533691,
      "learning_rate": 0.0001544443111103492,
      "loss": 1.2626,
      "step": 159450
    },
    {
      "epoch": 4.762189114143254,
      "grad_norm": 3.787076234817505,
      "learning_rate": 0.00015443002531443037,
      "loss": 1.2197,
      "step": 159500
    },
    {
      "epoch": 4.763681963395336,
      "grad_norm": 4.906338214874268,
      "learning_rate": 0.00015441573951851154,
      "loss": 1.2964,
      "step": 159550
    },
    {
      "epoch": 4.765174812647419,
      "grad_norm": 5.104478359222412,
      "learning_rate": 0.00015440145372259273,
      "loss": 1.23,
      "step": 159600
    },
    {
      "epoch": 4.766667661899501,
      "grad_norm": 4.3545918464660645,
      "learning_rate": 0.00015438716792667387,
      "loss": 1.2015,
      "step": 159650
    },
    {
      "epoch": 4.768160511151584,
      "grad_norm": 4.524704456329346,
      "learning_rate": 0.00015437288213075506,
      "loss": 1.2358,
      "step": 159700
    },
    {
      "epoch": 4.769653360403666,
      "grad_norm": 4.5480637550354,
      "learning_rate": 0.0001543585963348362,
      "loss": 1.1828,
      "step": 159750
    },
    {
      "epoch": 4.771146209655749,
      "grad_norm": 4.825510501861572,
      "learning_rate": 0.0001543443105389174,
      "loss": 1.1657,
      "step": 159800
    },
    {
      "epoch": 4.772639058907831,
      "grad_norm": 3.6792662143707275,
      "learning_rate": 0.00015433002474299852,
      "loss": 1.2789,
      "step": 159850
    },
    {
      "epoch": 4.774131908159914,
      "grad_norm": 4.740036964416504,
      "learning_rate": 0.00015431573894707972,
      "loss": 1.2292,
      "step": 159900
    },
    {
      "epoch": 4.775624757411997,
      "grad_norm": 5.538552761077881,
      "learning_rate": 0.00015430145315116088,
      "loss": 1.1691,
      "step": 159950
    },
    {
      "epoch": 4.7771176066640795,
      "grad_norm": 4.624987602233887,
      "learning_rate": 0.00015428716735524205,
      "loss": 1.2157,
      "step": 160000
    },
    {
      "epoch": 4.778610455916161,
      "grad_norm": 6.003055572509766,
      "learning_rate": 0.0001542728815593232,
      "loss": 1.2021,
      "step": 160050
    },
    {
      "epoch": 4.780103305168244,
      "grad_norm": 3.8452155590057373,
      "learning_rate": 0.00015425859576340437,
      "loss": 1.254,
      "step": 160100
    },
    {
      "epoch": 4.781596154420327,
      "grad_norm": 4.149327278137207,
      "learning_rate": 0.00015424430996748554,
      "loss": 1.235,
      "step": 160150
    },
    {
      "epoch": 4.783089003672409,
      "grad_norm": 4.331029415130615,
      "learning_rate": 0.0001542300241715667,
      "loss": 1.2077,
      "step": 160200
    },
    {
      "epoch": 4.784581852924491,
      "grad_norm": 4.8423662185668945,
      "learning_rate": 0.00015421573837564787,
      "loss": 1.2223,
      "step": 160250
    },
    {
      "epoch": 4.786074702176574,
      "grad_norm": 4.183012008666992,
      "learning_rate": 0.00015420145257972903,
      "loss": 1.2438,
      "step": 160300
    },
    {
      "epoch": 4.787567551428657,
      "grad_norm": 3.7776644229888916,
      "learning_rate": 0.0001541871667838102,
      "loss": 1.2093,
      "step": 160350
    },
    {
      "epoch": 4.789060400680739,
      "grad_norm": 4.586690902709961,
      "learning_rate": 0.0001541728809878914,
      "loss": 1.2353,
      "step": 160400
    },
    {
      "epoch": 4.790553249932822,
      "grad_norm": 4.448139190673828,
      "learning_rate": 0.00015415859519197253,
      "loss": 1.2149,
      "step": 160450
    },
    {
      "epoch": 4.792046099184905,
      "grad_norm": 4.895113945007324,
      "learning_rate": 0.00015414430939605372,
      "loss": 1.2682,
      "step": 160500
    },
    {
      "epoch": 4.793538948436987,
      "grad_norm": 6.31041955947876,
      "learning_rate": 0.00015413002360013486,
      "loss": 1.2296,
      "step": 160550
    },
    {
      "epoch": 4.795031797689069,
      "grad_norm": 5.237212181091309,
      "learning_rate": 0.00015411573780421602,
      "loss": 1.2121,
      "step": 160600
    },
    {
      "epoch": 4.796524646941152,
      "grad_norm": 5.017521381378174,
      "learning_rate": 0.00015410145200829719,
      "loss": 1.2413,
      "step": 160650
    },
    {
      "epoch": 4.798017496193235,
      "grad_norm": 4.569025039672852,
      "learning_rate": 0.00015408716621237835,
      "loss": 1.2321,
      "step": 160700
    },
    {
      "epoch": 4.7995103454453165,
      "grad_norm": 4.4759650230407715,
      "learning_rate": 0.00015407288041645954,
      "loss": 1.2579,
      "step": 160750
    },
    {
      "epoch": 4.801003194697399,
      "grad_norm": 4.48403263092041,
      "learning_rate": 0.00015405859462054068,
      "loss": 1.2332,
      "step": 160800
    },
    {
      "epoch": 4.802496043949482,
      "grad_norm": 4.603891849517822,
      "learning_rate": 0.00015404430882462187,
      "loss": 1.2004,
      "step": 160850
    },
    {
      "epoch": 4.803988893201565,
      "grad_norm": 3.970634937286377,
      "learning_rate": 0.000154030023028703,
      "loss": 1.2121,
      "step": 160900
    },
    {
      "epoch": 4.805481742453647,
      "grad_norm": 5.254501819610596,
      "learning_rate": 0.0001540157372327842,
      "loss": 1.3071,
      "step": 160950
    },
    {
      "epoch": 4.80697459170573,
      "grad_norm": 3.6827309131622314,
      "learning_rate": 0.00015400145143686534,
      "loss": 1.2399,
      "step": 161000
    },
    {
      "epoch": 4.808467440957812,
      "grad_norm": 4.880918502807617,
      "learning_rate": 0.00015398716564094653,
      "loss": 1.1842,
      "step": 161050
    },
    {
      "epoch": 4.809960290209895,
      "grad_norm": 3.4792227745056152,
      "learning_rate": 0.0001539728798450277,
      "loss": 1.22,
      "step": 161100
    },
    {
      "epoch": 4.811453139461977,
      "grad_norm": 4.746030807495117,
      "learning_rate": 0.00015395859404910886,
      "loss": 1.2068,
      "step": 161150
    },
    {
      "epoch": 4.81294598871406,
      "grad_norm": 4.76997709274292,
      "learning_rate": 0.00015394430825319002,
      "loss": 1.2713,
      "step": 161200
    },
    {
      "epoch": 4.814438837966142,
      "grad_norm": 3.699706792831421,
      "learning_rate": 0.0001539300224572712,
      "loss": 1.2059,
      "step": 161250
    },
    {
      "epoch": 4.8159316872182245,
      "grad_norm": 5.2487664222717285,
      "learning_rate": 0.00015391573666135235,
      "loss": 1.2295,
      "step": 161300
    },
    {
      "epoch": 4.817424536470307,
      "grad_norm": 4.549964904785156,
      "learning_rate": 0.00015390145086543352,
      "loss": 1.2362,
      "step": 161350
    },
    {
      "epoch": 4.81891738572239,
      "grad_norm": 6.63731575012207,
      "learning_rate": 0.00015388716506951468,
      "loss": 1.2018,
      "step": 161400
    },
    {
      "epoch": 4.820410234974473,
      "grad_norm": 4.054471969604492,
      "learning_rate": 0.00015387287927359585,
      "loss": 1.2234,
      "step": 161450
    },
    {
      "epoch": 4.8219030842265544,
      "grad_norm": 4.350498199462891,
      "learning_rate": 0.000153858593477677,
      "loss": 1.2006,
      "step": 161500
    },
    {
      "epoch": 4.823395933478637,
      "grad_norm": 4.074748992919922,
      "learning_rate": 0.0001538443076817582,
      "loss": 1.2353,
      "step": 161550
    },
    {
      "epoch": 4.82488878273072,
      "grad_norm": 4.194019317626953,
      "learning_rate": 0.00015383002188583934,
      "loss": 1.2294,
      "step": 161600
    },
    {
      "epoch": 4.826381631982803,
      "grad_norm": 4.11326789855957,
      "learning_rate": 0.00015381573608992053,
      "loss": 1.2195,
      "step": 161650
    },
    {
      "epoch": 4.827874481234885,
      "grad_norm": 5.007305145263672,
      "learning_rate": 0.00015380145029400167,
      "loss": 1.221,
      "step": 161700
    },
    {
      "epoch": 4.829367330486967,
      "grad_norm": 5.089259624481201,
      "learning_rate": 0.00015378716449808286,
      "loss": 1.1968,
      "step": 161750
    },
    {
      "epoch": 4.83086017973905,
      "grad_norm": 5.502153396606445,
      "learning_rate": 0.000153772878702164,
      "loss": 1.2386,
      "step": 161800
    },
    {
      "epoch": 4.8323530289911325,
      "grad_norm": 4.014642238616943,
      "learning_rate": 0.0001537585929062452,
      "loss": 1.2358,
      "step": 161850
    },
    {
      "epoch": 4.833845878243215,
      "grad_norm": 4.9020915031433105,
      "learning_rate": 0.00015374430711032636,
      "loss": 1.1838,
      "step": 161900
    },
    {
      "epoch": 4.835338727495298,
      "grad_norm": 3.0818114280700684,
      "learning_rate": 0.00015373002131440752,
      "loss": 1.3071,
      "step": 161950
    },
    {
      "epoch": 4.83683157674738,
      "grad_norm": 4.447099208831787,
      "learning_rate": 0.00015371573551848869,
      "loss": 1.2656,
      "step": 162000
    },
    {
      "epoch": 4.838324425999462,
      "grad_norm": 3.805255174636841,
      "learning_rate": 0.00015370144972256985,
      "loss": 1.233,
      "step": 162050
    },
    {
      "epoch": 4.839817275251545,
      "grad_norm": 4.1598968505859375,
      "learning_rate": 0.00015368716392665102,
      "loss": 1.256,
      "step": 162100
    },
    {
      "epoch": 4.841310124503628,
      "grad_norm": 4.9868364334106445,
      "learning_rate": 0.00015367287813073218,
      "loss": 1.1727,
      "step": 162150
    },
    {
      "epoch": 4.84280297375571,
      "grad_norm": 5.927983283996582,
      "learning_rate": 0.00015365859233481334,
      "loss": 1.1832,
      "step": 162200
    },
    {
      "epoch": 4.844295823007792,
      "grad_norm": 4.542764663696289,
      "learning_rate": 0.0001536443065388945,
      "loss": 1.2064,
      "step": 162250
    },
    {
      "epoch": 4.845788672259875,
      "grad_norm": 5.132274627685547,
      "learning_rate": 0.00015363002074297567,
      "loss": 1.2295,
      "step": 162300
    },
    {
      "epoch": 4.847281521511958,
      "grad_norm": 4.797247409820557,
      "learning_rate": 0.00015361573494705687,
      "loss": 1.1985,
      "step": 162350
    },
    {
      "epoch": 4.8487743707640405,
      "grad_norm": 3.799029588699341,
      "learning_rate": 0.000153601449151138,
      "loss": 1.2029,
      "step": 162400
    },
    {
      "epoch": 4.850267220016123,
      "grad_norm": 3.758556842803955,
      "learning_rate": 0.0001535871633552192,
      "loss": 1.2401,
      "step": 162450
    },
    {
      "epoch": 4.851760069268205,
      "grad_norm": 4.282034397125244,
      "learning_rate": 0.00015357287755930033,
      "loss": 1.1842,
      "step": 162500
    },
    {
      "epoch": 4.853252918520288,
      "grad_norm": 4.154748916625977,
      "learning_rate": 0.00015355859176338152,
      "loss": 1.2201,
      "step": 162550
    },
    {
      "epoch": 4.85474576777237,
      "grad_norm": 4.305574893951416,
      "learning_rate": 0.0001535443059674627,
      "loss": 1.206,
      "step": 162600
    },
    {
      "epoch": 4.856238617024453,
      "grad_norm": 4.58249568939209,
      "learning_rate": 0.00015353002017154385,
      "loss": 1.1851,
      "step": 162650
    },
    {
      "epoch": 4.857731466276535,
      "grad_norm": 3.7399301528930664,
      "learning_rate": 0.00015351573437562502,
      "loss": 1.2155,
      "step": 162700
    },
    {
      "epoch": 4.859224315528618,
      "grad_norm": 4.139153480529785,
      "learning_rate": 0.00015350144857970618,
      "loss": 1.2233,
      "step": 162750
    },
    {
      "epoch": 4.8607171647807,
      "grad_norm": 4.756731986999512,
      "learning_rate": 0.00015348716278378735,
      "loss": 1.2671,
      "step": 162800
    },
    {
      "epoch": 4.862210014032783,
      "grad_norm": 4.127963542938232,
      "learning_rate": 0.0001534728769878685,
      "loss": 1.2087,
      "step": 162850
    },
    {
      "epoch": 4.863702863284866,
      "grad_norm": 5.56052303314209,
      "learning_rate": 0.00015345859119194968,
      "loss": 1.2257,
      "step": 162900
    },
    {
      "epoch": 4.8651957125369485,
      "grad_norm": 6.622554302215576,
      "learning_rate": 0.00015344430539603084,
      "loss": 1.2274,
      "step": 162950
    },
    {
      "epoch": 4.86668856178903,
      "grad_norm": 3.089730739593506,
      "learning_rate": 0.000153430019600112,
      "loss": 1.2615,
      "step": 163000
    },
    {
      "epoch": 4.868181411041113,
      "grad_norm": 5.314934253692627,
      "learning_rate": 0.0001534157338041932,
      "loss": 1.2035,
      "step": 163050
    },
    {
      "epoch": 4.869674260293196,
      "grad_norm": 4.352352619171143,
      "learning_rate": 0.00015340144800827434,
      "loss": 1.2113,
      "step": 163100
    },
    {
      "epoch": 4.871167109545278,
      "grad_norm": 4.171778202056885,
      "learning_rate": 0.00015338716221235553,
      "loss": 1.1977,
      "step": 163150
    },
    {
      "epoch": 4.87265995879736,
      "grad_norm": 3.6624767780303955,
      "learning_rate": 0.00015337287641643666,
      "loss": 1.1419,
      "step": 163200
    },
    {
      "epoch": 4.874152808049443,
      "grad_norm": 3.778160810470581,
      "learning_rate": 0.00015335859062051783,
      "loss": 1.257,
      "step": 163250
    },
    {
      "epoch": 4.875645657301526,
      "grad_norm": 4.410187244415283,
      "learning_rate": 0.000153344304824599,
      "loss": 1.1846,
      "step": 163300
    },
    {
      "epoch": 4.877138506553608,
      "grad_norm": 4.407317161560059,
      "learning_rate": 0.00015333001902868016,
      "loss": 1.2594,
      "step": 163350
    },
    {
      "epoch": 4.878631355805691,
      "grad_norm": 3.622882843017578,
      "learning_rate": 0.00015331573323276135,
      "loss": 1.1928,
      "step": 163400
    },
    {
      "epoch": 4.880124205057773,
      "grad_norm": 4.268970012664795,
      "learning_rate": 0.0001533014474368425,
      "loss": 1.2575,
      "step": 163450
    },
    {
      "epoch": 4.881617054309856,
      "grad_norm": 5.875527381896973,
      "learning_rate": 0.00015328716164092368,
      "loss": 1.1402,
      "step": 163500
    },
    {
      "epoch": 4.883109903561938,
      "grad_norm": 4.78771448135376,
      "learning_rate": 0.00015327287584500482,
      "loss": 1.2003,
      "step": 163550
    },
    {
      "epoch": 4.884602752814021,
      "grad_norm": 4.128746509552002,
      "learning_rate": 0.000153258590049086,
      "loss": 1.1964,
      "step": 163600
    },
    {
      "epoch": 4.886095602066104,
      "grad_norm": 2.846057176589966,
      "learning_rate": 0.00015324430425316715,
      "loss": 1.2154,
      "step": 163650
    },
    {
      "epoch": 4.8875884513181855,
      "grad_norm": 3.893822193145752,
      "learning_rate": 0.00015323001845724834,
      "loss": 1.1727,
      "step": 163700
    },
    {
      "epoch": 4.889081300570268,
      "grad_norm": 3.671567916870117,
      "learning_rate": 0.0001532157326613295,
      "loss": 1.247,
      "step": 163750
    },
    {
      "epoch": 4.890574149822351,
      "grad_norm": 4.6587066650390625,
      "learning_rate": 0.00015320144686541067,
      "loss": 1.2278,
      "step": 163800
    },
    {
      "epoch": 4.892066999074434,
      "grad_norm": 4.010086536407471,
      "learning_rate": 0.00015318716106949183,
      "loss": 1.2153,
      "step": 163850
    },
    {
      "epoch": 4.893559848326516,
      "grad_norm": 3.343904495239258,
      "learning_rate": 0.000153172875273573,
      "loss": 1.1807,
      "step": 163900
    },
    {
      "epoch": 4.895052697578598,
      "grad_norm": 4.4615702629089355,
      "learning_rate": 0.00015315858947765416,
      "loss": 1.1961,
      "step": 163950
    },
    {
      "epoch": 4.896545546830681,
      "grad_norm": 4.361425399780273,
      "learning_rate": 0.00015314430368173533,
      "loss": 1.2436,
      "step": 164000
    },
    {
      "epoch": 4.898038396082764,
      "grad_norm": 4.161865234375,
      "learning_rate": 0.0001531300178858165,
      "loss": 1.2104,
      "step": 164050
    },
    {
      "epoch": 4.899531245334846,
      "grad_norm": 5.719996929168701,
      "learning_rate": 0.00015311573208989766,
      "loss": 1.2363,
      "step": 164100
    },
    {
      "epoch": 4.901024094586928,
      "grad_norm": 4.253158092498779,
      "learning_rate": 0.00015310144629397882,
      "loss": 1.2513,
      "step": 164150
    },
    {
      "epoch": 4.902516943839011,
      "grad_norm": 2.575803518295288,
      "learning_rate": 0.00015308716049806,
      "loss": 1.1488,
      "step": 164200
    },
    {
      "epoch": 4.9040097930910935,
      "grad_norm": 6.142227649688721,
      "learning_rate": 0.00015307287470214115,
      "loss": 1.2197,
      "step": 164250
    },
    {
      "epoch": 4.905502642343176,
      "grad_norm": 5.662699222564697,
      "learning_rate": 0.00015305858890622234,
      "loss": 1.2282,
      "step": 164300
    },
    {
      "epoch": 4.906995491595259,
      "grad_norm": 4.345313549041748,
      "learning_rate": 0.00015304430311030348,
      "loss": 1.263,
      "step": 164350
    },
    {
      "epoch": 4.908488340847342,
      "grad_norm": 4.106235980987549,
      "learning_rate": 0.00015303001731438467,
      "loss": 1.2568,
      "step": 164400
    },
    {
      "epoch": 4.9099811900994235,
      "grad_norm": 5.795431613922119,
      "learning_rate": 0.0001530157315184658,
      "loss": 1.1668,
      "step": 164450
    },
    {
      "epoch": 4.911474039351506,
      "grad_norm": 4.605996608734131,
      "learning_rate": 0.000153001445722547,
      "loss": 1.2513,
      "step": 164500
    },
    {
      "epoch": 4.912966888603589,
      "grad_norm": 6.11320686340332,
      "learning_rate": 0.00015298715992662816,
      "loss": 1.1922,
      "step": 164550
    },
    {
      "epoch": 4.914459737855672,
      "grad_norm": 6.641072750091553,
      "learning_rate": 0.00015297287413070933,
      "loss": 1.2272,
      "step": 164600
    },
    {
      "epoch": 4.915952587107753,
      "grad_norm": 3.955686092376709,
      "learning_rate": 0.0001529585883347905,
      "loss": 1.2089,
      "step": 164650
    },
    {
      "epoch": 4.917445436359836,
      "grad_norm": 3.445519208908081,
      "learning_rate": 0.00015294430253887166,
      "loss": 1.2013,
      "step": 164700
    },
    {
      "epoch": 4.918938285611919,
      "grad_norm": 5.487844944000244,
      "learning_rate": 0.00015293001674295282,
      "loss": 1.2818,
      "step": 164750
    },
    {
      "epoch": 4.9204311348640015,
      "grad_norm": 5.098935127258301,
      "learning_rate": 0.000152915730947034,
      "loss": 1.2195,
      "step": 164800
    },
    {
      "epoch": 4.921923984116084,
      "grad_norm": 3.2200589179992676,
      "learning_rate": 0.00015290144515111515,
      "loss": 1.2425,
      "step": 164850
    },
    {
      "epoch": 4.923416833368167,
      "grad_norm": 5.7381205558776855,
      "learning_rate": 0.00015288715935519632,
      "loss": 1.2012,
      "step": 164900
    },
    {
      "epoch": 4.924909682620249,
      "grad_norm": 4.34859037399292,
      "learning_rate": 0.00015287287355927748,
      "loss": 1.2595,
      "step": 164950
    },
    {
      "epoch": 4.9264025318723315,
      "grad_norm": 4.303558349609375,
      "learning_rate": 0.00015285858776335867,
      "loss": 1.2802,
      "step": 165000
    },
    {
      "epoch": 4.927895381124414,
      "grad_norm": 3.9082958698272705,
      "learning_rate": 0.0001528443019674398,
      "loss": 1.2252,
      "step": 165050
    },
    {
      "epoch": 4.929388230376497,
      "grad_norm": 3.7715795040130615,
      "learning_rate": 0.000152830016171521,
      "loss": 1.2849,
      "step": 165100
    },
    {
      "epoch": 4.930881079628579,
      "grad_norm": 5.810252666473389,
      "learning_rate": 0.00015281573037560214,
      "loss": 1.1903,
      "step": 165150
    },
    {
      "epoch": 4.932373928880661,
      "grad_norm": 4.941216468811035,
      "learning_rate": 0.00015280144457968333,
      "loss": 1.199,
      "step": 165200
    },
    {
      "epoch": 4.933866778132744,
      "grad_norm": 4.764664173126221,
      "learning_rate": 0.0001527871587837645,
      "loss": 1.2358,
      "step": 165250
    },
    {
      "epoch": 4.935359627384827,
      "grad_norm": 4.884801387786865,
      "learning_rate": 0.00015277287298784566,
      "loss": 1.168,
      "step": 165300
    },
    {
      "epoch": 4.9368524766369095,
      "grad_norm": 4.915900230407715,
      "learning_rate": 0.00015275858719192683,
      "loss": 1.2395,
      "step": 165350
    },
    {
      "epoch": 4.938345325888992,
      "grad_norm": 4.347561836242676,
      "learning_rate": 0.000152744301396008,
      "loss": 1.2134,
      "step": 165400
    },
    {
      "epoch": 4.939838175141074,
      "grad_norm": 5.786562919616699,
      "learning_rate": 0.00015273001560008916,
      "loss": 1.2413,
      "step": 165450
    },
    {
      "epoch": 4.941331024393157,
      "grad_norm": 4.780857086181641,
      "learning_rate": 0.00015271572980417032,
      "loss": 1.2336,
      "step": 165500
    },
    {
      "epoch": 4.9428238736452395,
      "grad_norm": 4.805835723876953,
      "learning_rate": 0.00015270144400825148,
      "loss": 1.2464,
      "step": 165550
    },
    {
      "epoch": 4.944316722897322,
      "grad_norm": 4.181761264801025,
      "learning_rate": 0.00015268715821233265,
      "loss": 1.2486,
      "step": 165600
    },
    {
      "epoch": 4.945809572149404,
      "grad_norm": 6.043191432952881,
      "learning_rate": 0.00015267287241641381,
      "loss": 1.2485,
      "step": 165650
    },
    {
      "epoch": 4.947302421401487,
      "grad_norm": 3.7125275135040283,
      "learning_rate": 0.00015265858662049498,
      "loss": 1.2216,
      "step": 165700
    },
    {
      "epoch": 4.948795270653569,
      "grad_norm": 4.299423694610596,
      "learning_rate": 0.00015264430082457614,
      "loss": 1.1784,
      "step": 165750
    },
    {
      "epoch": 4.950288119905652,
      "grad_norm": 3.839174270629883,
      "learning_rate": 0.00015263001502865734,
      "loss": 1.2377,
      "step": 165800
    },
    {
      "epoch": 4.951780969157735,
      "grad_norm": 4.8078837394714355,
      "learning_rate": 0.00015261572923273847,
      "loss": 1.2435,
      "step": 165850
    },
    {
      "epoch": 4.953273818409817,
      "grad_norm": 4.367580413818359,
      "learning_rate": 0.00015260144343681964,
      "loss": 1.1683,
      "step": 165900
    },
    {
      "epoch": 4.954766667661899,
      "grad_norm": 4.1319756507873535,
      "learning_rate": 0.0001525871576409008,
      "loss": 1.259,
      "step": 165950
    },
    {
      "epoch": 4.956259516913982,
      "grad_norm": 4.9268293380737305,
      "learning_rate": 0.00015257287184498197,
      "loss": 1.1774,
      "step": 166000
    },
    {
      "epoch": 4.957752366166065,
      "grad_norm": 3.851490020751953,
      "learning_rate": 0.00015255858604906316,
      "loss": 1.2051,
      "step": 166050
    },
    {
      "epoch": 4.9592452154181474,
      "grad_norm": 6.583017826080322,
      "learning_rate": 0.0001525443002531443,
      "loss": 1.2744,
      "step": 166100
    },
    {
      "epoch": 4.960738064670229,
      "grad_norm": 3.7507100105285645,
      "learning_rate": 0.0001525300144572255,
      "loss": 1.1927,
      "step": 166150
    },
    {
      "epoch": 4.962230913922312,
      "grad_norm": 5.387413501739502,
      "learning_rate": 0.00015251572866130663,
      "loss": 1.203,
      "step": 166200
    },
    {
      "epoch": 4.963723763174395,
      "grad_norm": 3.7446823120117188,
      "learning_rate": 0.00015250144286538782,
      "loss": 1.2062,
      "step": 166250
    },
    {
      "epoch": 4.965216612426477,
      "grad_norm": 4.162785053253174,
      "learning_rate": 0.00015248715706946895,
      "loss": 1.1969,
      "step": 166300
    },
    {
      "epoch": 4.96670946167856,
      "grad_norm": 4.15479040145874,
      "learning_rate": 0.00015247287127355015,
      "loss": 1.2164,
      "step": 166350
    },
    {
      "epoch": 4.968202310930642,
      "grad_norm": 4.103176593780518,
      "learning_rate": 0.0001524585854776313,
      "loss": 1.2318,
      "step": 166400
    },
    {
      "epoch": 4.969695160182725,
      "grad_norm": 4.903412818908691,
      "learning_rate": 0.00015244429968171248,
      "loss": 1.1984,
      "step": 166450
    },
    {
      "epoch": 4.971188009434807,
      "grad_norm": 5.026261806488037,
      "learning_rate": 0.00015243001388579364,
      "loss": 1.2472,
      "step": 166500
    },
    {
      "epoch": 4.97268085868689,
      "grad_norm": 4.777016639709473,
      "learning_rate": 0.0001524157280898748,
      "loss": 1.2363,
      "step": 166550
    },
    {
      "epoch": 4.974173707938972,
      "grad_norm": 3.636580467224121,
      "learning_rate": 0.00015240144229395597,
      "loss": 1.2543,
      "step": 166600
    },
    {
      "epoch": 4.9756665571910546,
      "grad_norm": 3.8117880821228027,
      "learning_rate": 0.00015238715649803713,
      "loss": 1.1817,
      "step": 166650
    },
    {
      "epoch": 4.977159406443137,
      "grad_norm": 5.417634963989258,
      "learning_rate": 0.0001523728707021183,
      "loss": 1.2251,
      "step": 166700
    },
    {
      "epoch": 4.97865225569522,
      "grad_norm": 4.955143451690674,
      "learning_rate": 0.00015235858490619946,
      "loss": 1.2164,
      "step": 166750
    },
    {
      "epoch": 4.980145104947303,
      "grad_norm": 4.069868564605713,
      "learning_rate": 0.00015234429911028063,
      "loss": 1.2496,
      "step": 166800
    },
    {
      "epoch": 4.981637954199385,
      "grad_norm": 3.862034320831299,
      "learning_rate": 0.00015233001331436182,
      "loss": 1.19,
      "step": 166850
    },
    {
      "epoch": 4.983130803451467,
      "grad_norm": 5.259671211242676,
      "learning_rate": 0.00015231572751844296,
      "loss": 1.216,
      "step": 166900
    },
    {
      "epoch": 4.98462365270355,
      "grad_norm": 4.1818742752075195,
      "learning_rate": 0.00015230144172252415,
      "loss": 1.2089,
      "step": 166950
    },
    {
      "epoch": 4.986116501955633,
      "grad_norm": 4.551473617553711,
      "learning_rate": 0.0001522871559266053,
      "loss": 1.2218,
      "step": 167000
    },
    {
      "epoch": 4.987609351207715,
      "grad_norm": 4.714193820953369,
      "learning_rate": 0.00015227287013068648,
      "loss": 1.2117,
      "step": 167050
    },
    {
      "epoch": 4.989102200459797,
      "grad_norm": 5.234292030334473,
      "learning_rate": 0.00015225858433476762,
      "loss": 1.1861,
      "step": 167100
    },
    {
      "epoch": 4.99059504971188,
      "grad_norm": 3.052459239959717,
      "learning_rate": 0.0001522442985388488,
      "loss": 1.1598,
      "step": 167150
    },
    {
      "epoch": 4.9920878989639625,
      "grad_norm": 4.330234527587891,
      "learning_rate": 0.00015223001274292997,
      "loss": 1.2177,
      "step": 167200
    },
    {
      "epoch": 4.993580748216045,
      "grad_norm": 4.3281636238098145,
      "learning_rate": 0.00015221572694701114,
      "loss": 1.291,
      "step": 167250
    },
    {
      "epoch": 4.995073597468128,
      "grad_norm": 4.501579761505127,
      "learning_rate": 0.0001522014411510923,
      "loss": 1.215,
      "step": 167300
    },
    {
      "epoch": 4.996566446720211,
      "grad_norm": 3.957705497741699,
      "learning_rate": 0.00015218715535517347,
      "loss": 1.2056,
      "step": 167350
    },
    {
      "epoch": 4.9980592959722925,
      "grad_norm": 3.6151764392852783,
      "learning_rate": 0.00015217286955925463,
      "loss": 1.2531,
      "step": 167400
    },
    {
      "epoch": 4.999552145224375,
      "grad_norm": 5.759849548339844,
      "learning_rate": 0.0001521585837633358,
      "loss": 1.233,
      "step": 167450
    },
    {
      "epoch": 5.001044994476458,
      "grad_norm": 4.063636779785156,
      "learning_rate": 0.00015214429796741696,
      "loss": 1.0927,
      "step": 167500
    },
    {
      "epoch": 5.002537843728541,
      "grad_norm": 5.337904930114746,
      "learning_rate": 0.00015213001217149813,
      "loss": 1.0732,
      "step": 167550
    },
    {
      "epoch": 5.004030692980622,
      "grad_norm": 3.7725415229797363,
      "learning_rate": 0.0001521157263755793,
      "loss": 1.0745,
      "step": 167600
    },
    {
      "epoch": 5.005523542232705,
      "grad_norm": 5.607946395874023,
      "learning_rate": 0.00015210144057966048,
      "loss": 1.0961,
      "step": 167650
    },
    {
      "epoch": 5.007016391484788,
      "grad_norm": 5.750009059906006,
      "learning_rate": 0.00015208715478374162,
      "loss": 1.1483,
      "step": 167700
    },
    {
      "epoch": 5.0085092407368705,
      "grad_norm": 4.2717180252075195,
      "learning_rate": 0.0001520728689878228,
      "loss": 1.1728,
      "step": 167750
    },
    {
      "epoch": 5.010002089988953,
      "grad_norm": 4.048334121704102,
      "learning_rate": 0.00015205858319190395,
      "loss": 1.1091,
      "step": 167800
    },
    {
      "epoch": 5.011494939241035,
      "grad_norm": 4.57917594909668,
      "learning_rate": 0.00015204429739598514,
      "loss": 1.1404,
      "step": 167850
    },
    {
      "epoch": 5.012987788493118,
      "grad_norm": 5.619372367858887,
      "learning_rate": 0.00015203001160006628,
      "loss": 1.1803,
      "step": 167900
    },
    {
      "epoch": 5.0144806377452005,
      "grad_norm": 4.788344383239746,
      "learning_rate": 0.00015201572580414747,
      "loss": 1.1131,
      "step": 167950
    },
    {
      "epoch": 5.015973486997283,
      "grad_norm": 4.1209588050842285,
      "learning_rate": 0.00015200144000822863,
      "loss": 1.108,
      "step": 168000
    },
    {
      "epoch": 5.017466336249366,
      "grad_norm": 3.88391375541687,
      "learning_rate": 0.0001519871542123098,
      "loss": 1.1692,
      "step": 168050
    },
    {
      "epoch": 5.018959185501448,
      "grad_norm": 5.6544904708862305,
      "learning_rate": 0.00015197286841639096,
      "loss": 1.1457,
      "step": 168100
    },
    {
      "epoch": 5.02045203475353,
      "grad_norm": 5.817676544189453,
      "learning_rate": 0.00015195858262047213,
      "loss": 1.124,
      "step": 168150
    },
    {
      "epoch": 5.021944884005613,
      "grad_norm": 6.914028644561768,
      "learning_rate": 0.0001519442968245533,
      "loss": 1.0979,
      "step": 168200
    },
    {
      "epoch": 5.023437733257696,
      "grad_norm": 4.256136417388916,
      "learning_rate": 0.00015193001102863446,
      "loss": 1.1524,
      "step": 168250
    },
    {
      "epoch": 5.0249305825097785,
      "grad_norm": 4.776482582092285,
      "learning_rate": 0.00015191572523271562,
      "loss": 1.1005,
      "step": 168300
    },
    {
      "epoch": 5.02642343176186,
      "grad_norm": 4.402430057525635,
      "learning_rate": 0.0001519014394367968,
      "loss": 1.1295,
      "step": 168350
    },
    {
      "epoch": 5.027916281013943,
      "grad_norm": 4.203427314758301,
      "learning_rate": 0.00015188715364087795,
      "loss": 1.1377,
      "step": 168400
    },
    {
      "epoch": 5.029409130266026,
      "grad_norm": 3.8063766956329346,
      "learning_rate": 0.00015187286784495912,
      "loss": 1.0963,
      "step": 168450
    },
    {
      "epoch": 5.0309019795181085,
      "grad_norm": 4.944723129272461,
      "learning_rate": 0.00015185858204904028,
      "loss": 1.1298,
      "step": 168500
    },
    {
      "epoch": 5.032394828770191,
      "grad_norm": 4.621240615844727,
      "learning_rate": 0.00015184429625312145,
      "loss": 1.1597,
      "step": 168550
    },
    {
      "epoch": 5.033887678022273,
      "grad_norm": 5.244666576385498,
      "learning_rate": 0.0001518300104572026,
      "loss": 1.1266,
      "step": 168600
    },
    {
      "epoch": 5.035380527274356,
      "grad_norm": 3.7527828216552734,
      "learning_rate": 0.00015181572466128377,
      "loss": 1.1254,
      "step": 168650
    },
    {
      "epoch": 5.036873376526438,
      "grad_norm": 5.852657794952393,
      "learning_rate": 0.00015180143886536497,
      "loss": 1.1909,
      "step": 168700
    },
    {
      "epoch": 5.038366225778521,
      "grad_norm": 4.197393417358398,
      "learning_rate": 0.0001517871530694461,
      "loss": 1.1742,
      "step": 168750
    },
    {
      "epoch": 5.039859075030604,
      "grad_norm": 4.88790225982666,
      "learning_rate": 0.0001517728672735273,
      "loss": 1.1273,
      "step": 168800
    },
    {
      "epoch": 5.041351924282686,
      "grad_norm": 5.1781134605407715,
      "learning_rate": 0.00015175858147760843,
      "loss": 1.1011,
      "step": 168850
    },
    {
      "epoch": 5.042844773534768,
      "grad_norm": 6.221542835235596,
      "learning_rate": 0.00015174429568168963,
      "loss": 1.0775,
      "step": 168900
    },
    {
      "epoch": 5.044337622786851,
      "grad_norm": 3.557209014892578,
      "learning_rate": 0.00015173000988577076,
      "loss": 1.1295,
      "step": 168950
    },
    {
      "epoch": 5.045830472038934,
      "grad_norm": 4.908199310302734,
      "learning_rate": 0.00015171572408985195,
      "loss": 1.1428,
      "step": 169000
    },
    {
      "epoch": 5.0473233212910165,
      "grad_norm": 3.2522714138031006,
      "learning_rate": 0.00015170143829393312,
      "loss": 1.1454,
      "step": 169050
    },
    {
      "epoch": 5.048816170543098,
      "grad_norm": 4.945195198059082,
      "learning_rate": 0.00015168715249801428,
      "loss": 1.1541,
      "step": 169100
    },
    {
      "epoch": 5.050309019795181,
      "grad_norm": 3.7441771030426025,
      "learning_rate": 0.00015167286670209545,
      "loss": 1.1223,
      "step": 169150
    },
    {
      "epoch": 5.051801869047264,
      "grad_norm": 4.811500549316406,
      "learning_rate": 0.0001516585809061766,
      "loss": 1.1861,
      "step": 169200
    },
    {
      "epoch": 5.053294718299346,
      "grad_norm": 4.175551414489746,
      "learning_rate": 0.00015164429511025778,
      "loss": 1.1655,
      "step": 169250
    },
    {
      "epoch": 5.054787567551428,
      "grad_norm": 4.7304205894470215,
      "learning_rate": 0.00015163000931433894,
      "loss": 1.1238,
      "step": 169300
    },
    {
      "epoch": 5.056280416803511,
      "grad_norm": 4.252381801605225,
      "learning_rate": 0.0001516157235184201,
      "loss": 1.1683,
      "step": 169350
    },
    {
      "epoch": 5.057773266055594,
      "grad_norm": 3.368849039077759,
      "learning_rate": 0.00015160143772250127,
      "loss": 1.1584,
      "step": 169400
    },
    {
      "epoch": 5.059266115307676,
      "grad_norm": 4.560303688049316,
      "learning_rate": 0.00015158715192658244,
      "loss": 1.1668,
      "step": 169450
    },
    {
      "epoch": 5.060758964559759,
      "grad_norm": 4.084552764892578,
      "learning_rate": 0.00015157286613066363,
      "loss": 1.1385,
      "step": 169500
    },
    {
      "epoch": 5.062251813811841,
      "grad_norm": 4.2204060554504395,
      "learning_rate": 0.00015155858033474477,
      "loss": 1.1486,
      "step": 169550
    },
    {
      "epoch": 5.063744663063924,
      "grad_norm": 3.872606039047241,
      "learning_rate": 0.00015154429453882596,
      "loss": 1.1732,
      "step": 169600
    },
    {
      "epoch": 5.065237512316006,
      "grad_norm": 6.434457302093506,
      "learning_rate": 0.0001515300087429071,
      "loss": 1.1658,
      "step": 169650
    },
    {
      "epoch": 5.066730361568089,
      "grad_norm": 4.881868362426758,
      "learning_rate": 0.0001515157229469883,
      "loss": 1.1769,
      "step": 169700
    },
    {
      "epoch": 5.068223210820172,
      "grad_norm": 4.622376918792725,
      "learning_rate": 0.00015150143715106942,
      "loss": 1.0902,
      "step": 169750
    },
    {
      "epoch": 5.0697160600722535,
      "grad_norm": 3.6144649982452393,
      "learning_rate": 0.00015148715135515062,
      "loss": 1.1098,
      "step": 169800
    },
    {
      "epoch": 5.071208909324336,
      "grad_norm": 4.620079517364502,
      "learning_rate": 0.00015147286555923178,
      "loss": 1.1716,
      "step": 169850
    },
    {
      "epoch": 5.072701758576419,
      "grad_norm": 4.734360694885254,
      "learning_rate": 0.00015145857976331295,
      "loss": 1.1539,
      "step": 169900
    },
    {
      "epoch": 5.074194607828502,
      "grad_norm": 4.301705360412598,
      "learning_rate": 0.0001514442939673941,
      "loss": 1.1533,
      "step": 169950
    },
    {
      "epoch": 5.075687457080584,
      "grad_norm": 5.320747375488281,
      "learning_rate": 0.00015143000817147527,
      "loss": 1.1233,
      "step": 170000
    },
    {
      "epoch": 5.077180306332666,
      "grad_norm": 4.841482639312744,
      "learning_rate": 0.00015141572237555644,
      "loss": 1.1201,
      "step": 170050
    },
    {
      "epoch": 5.078673155584749,
      "grad_norm": 5.517151832580566,
      "learning_rate": 0.0001514014365796376,
      "loss": 1.0986,
      "step": 170100
    },
    {
      "epoch": 5.080166004836832,
      "grad_norm": 4.323915004730225,
      "learning_rate": 0.00015138715078371877,
      "loss": 1.1362,
      "step": 170150
    },
    {
      "epoch": 5.081658854088914,
      "grad_norm": 4.015235900878906,
      "learning_rate": 0.00015137286498779993,
      "loss": 1.1571,
      "step": 170200
    },
    {
      "epoch": 5.083151703340997,
      "grad_norm": 5.928144454956055,
      "learning_rate": 0.0001513585791918811,
      "loss": 1.1672,
      "step": 170250
    },
    {
      "epoch": 5.084644552593079,
      "grad_norm": 4.04873514175415,
      "learning_rate": 0.0001513442933959623,
      "loss": 1.1028,
      "step": 170300
    },
    {
      "epoch": 5.0861374018451615,
      "grad_norm": 5.569055557250977,
      "learning_rate": 0.00015133000760004343,
      "loss": 1.1479,
      "step": 170350
    },
    {
      "epoch": 5.087630251097244,
      "grad_norm": 5.128480911254883,
      "learning_rate": 0.00015131572180412462,
      "loss": 1.1567,
      "step": 170400
    },
    {
      "epoch": 5.089123100349327,
      "grad_norm": 5.323390007019043,
      "learning_rate": 0.00015130143600820576,
      "loss": 1.1072,
      "step": 170450
    },
    {
      "epoch": 5.09061594960141,
      "grad_norm": 3.7644312381744385,
      "learning_rate": 0.00015128715021228695,
      "loss": 1.1435,
      "step": 170500
    },
    {
      "epoch": 5.092108798853491,
      "grad_norm": 4.050746917724609,
      "learning_rate": 0.00015127286441636809,
      "loss": 1.2082,
      "step": 170550
    },
    {
      "epoch": 5.093601648105574,
      "grad_norm": 6.779710292816162,
      "learning_rate": 0.00015125857862044928,
      "loss": 1.1314,
      "step": 170600
    },
    {
      "epoch": 5.095094497357657,
      "grad_norm": 4.560274124145508,
      "learning_rate": 0.00015124429282453044,
      "loss": 1.2157,
      "step": 170650
    },
    {
      "epoch": 5.0965873466097396,
      "grad_norm": 5.227869987487793,
      "learning_rate": 0.0001512300070286116,
      "loss": 1.187,
      "step": 170700
    },
    {
      "epoch": 5.098080195861822,
      "grad_norm": 5.666630744934082,
      "learning_rate": 0.00015121572123269277,
      "loss": 1.2005,
      "step": 170750
    },
    {
      "epoch": 5.099573045113904,
      "grad_norm": 4.06814432144165,
      "learning_rate": 0.00015120143543677394,
      "loss": 1.1351,
      "step": 170800
    },
    {
      "epoch": 5.101065894365987,
      "grad_norm": 5.10957145690918,
      "learning_rate": 0.0001511871496408551,
      "loss": 1.193,
      "step": 170850
    },
    {
      "epoch": 5.1025587436180695,
      "grad_norm": 5.230775833129883,
      "learning_rate": 0.00015117286384493627,
      "loss": 1.2186,
      "step": 170900
    },
    {
      "epoch": 5.104051592870152,
      "grad_norm": 4.77035665512085,
      "learning_rate": 0.00015115857804901743,
      "loss": 1.165,
      "step": 170950
    },
    {
      "epoch": 5.105544442122235,
      "grad_norm": 3.762669801712036,
      "learning_rate": 0.0001511442922530986,
      "loss": 1.0899,
      "step": 171000
    },
    {
      "epoch": 5.107037291374317,
      "grad_norm": 4.551312446594238,
      "learning_rate": 0.00015113000645717976,
      "loss": 1.1306,
      "step": 171050
    },
    {
      "epoch": 5.108530140626399,
      "grad_norm": 3.9182794094085693,
      "learning_rate": 0.00015111572066126092,
      "loss": 1.1365,
      "step": 171100
    },
    {
      "epoch": 5.110022989878482,
      "grad_norm": 4.611422538757324,
      "learning_rate": 0.0001511014348653421,
      "loss": 1.2169,
      "step": 171150
    },
    {
      "epoch": 5.111515839130565,
      "grad_norm": 4.5237321853637695,
      "learning_rate": 0.00015108714906942325,
      "loss": 1.1373,
      "step": 171200
    },
    {
      "epoch": 5.1130086883826475,
      "grad_norm": 4.893795967102051,
      "learning_rate": 0.00015107286327350442,
      "loss": 1.1307,
      "step": 171250
    },
    {
      "epoch": 5.114501537634729,
      "grad_norm": 4.26010799407959,
      "learning_rate": 0.00015105857747758558,
      "loss": 1.1461,
      "step": 171300
    },
    {
      "epoch": 5.115994386886812,
      "grad_norm": 5.365528583526611,
      "learning_rate": 0.00015104429168166675,
      "loss": 1.1156,
      "step": 171350
    },
    {
      "epoch": 5.117487236138895,
      "grad_norm": 3.8632538318634033,
      "learning_rate": 0.0001510300058857479,
      "loss": 1.1354,
      "step": 171400
    },
    {
      "epoch": 5.1189800853909775,
      "grad_norm": 5.570279598236084,
      "learning_rate": 0.0001510157200898291,
      "loss": 1.1758,
      "step": 171450
    },
    {
      "epoch": 5.120472934643059,
      "grad_norm": 4.220273017883301,
      "learning_rate": 0.00015100143429391024,
      "loss": 1.149,
      "step": 171500
    },
    {
      "epoch": 5.121965783895142,
      "grad_norm": 5.220869541168213,
      "learning_rate": 0.00015098714849799143,
      "loss": 1.208,
      "step": 171550
    },
    {
      "epoch": 5.123458633147225,
      "grad_norm": 3.9469594955444336,
      "learning_rate": 0.00015097286270207257,
      "loss": 1.1431,
      "step": 171600
    },
    {
      "epoch": 5.124951482399307,
      "grad_norm": 3.932490587234497,
      "learning_rate": 0.00015095857690615376,
      "loss": 1.1621,
      "step": 171650
    },
    {
      "epoch": 5.12644433165139,
      "grad_norm": 5.017124176025391,
      "learning_rate": 0.00015094429111023493,
      "loss": 1.1341,
      "step": 171700
    },
    {
      "epoch": 5.127937180903472,
      "grad_norm": 5.0581488609313965,
      "learning_rate": 0.0001509300053143161,
      "loss": 1.1756,
      "step": 171750
    },
    {
      "epoch": 5.129430030155555,
      "grad_norm": 5.208496570587158,
      "learning_rate": 0.00015091571951839726,
      "loss": 1.1226,
      "step": 171800
    },
    {
      "epoch": 5.130922879407637,
      "grad_norm": 3.9350638389587402,
      "learning_rate": 0.00015090143372247842,
      "loss": 1.1443,
      "step": 171850
    },
    {
      "epoch": 5.13241572865972,
      "grad_norm": 3.7098276615142822,
      "learning_rate": 0.00015088714792655959,
      "loss": 1.1346,
      "step": 171900
    },
    {
      "epoch": 5.133908577911803,
      "grad_norm": 4.124345779418945,
      "learning_rate": 0.00015087286213064075,
      "loss": 1.1955,
      "step": 171950
    },
    {
      "epoch": 5.135401427163885,
      "grad_norm": 3.5282959938049316,
      "learning_rate": 0.00015085857633472192,
      "loss": 1.1457,
      "step": 172000
    },
    {
      "epoch": 5.136894276415967,
      "grad_norm": 4.17992639541626,
      "learning_rate": 0.00015084429053880308,
      "loss": 1.161,
      "step": 172050
    },
    {
      "epoch": 5.13838712566805,
      "grad_norm": 3.838498830795288,
      "learning_rate": 0.00015083000474288424,
      "loss": 1.1381,
      "step": 172100
    },
    {
      "epoch": 5.139879974920133,
      "grad_norm": 5.275324821472168,
      "learning_rate": 0.00015081571894696544,
      "loss": 1.0803,
      "step": 172150
    },
    {
      "epoch": 5.141372824172215,
      "grad_norm": 4.769461154937744,
      "learning_rate": 0.00015080143315104657,
      "loss": 1.1494,
      "step": 172200
    },
    {
      "epoch": 5.142865673424297,
      "grad_norm": 4.681296348571777,
      "learning_rate": 0.00015078714735512777,
      "loss": 1.1726,
      "step": 172250
    },
    {
      "epoch": 5.14435852267638,
      "grad_norm": 4.308079719543457,
      "learning_rate": 0.0001507728615592089,
      "loss": 1.1331,
      "step": 172300
    },
    {
      "epoch": 5.145851371928463,
      "grad_norm": 4.144331455230713,
      "learning_rate": 0.0001507585757632901,
      "loss": 1.1579,
      "step": 172350
    },
    {
      "epoch": 5.147344221180545,
      "grad_norm": 4.138561725616455,
      "learning_rate": 0.00015074428996737123,
      "loss": 1.1462,
      "step": 172400
    },
    {
      "epoch": 5.148837070432628,
      "grad_norm": 4.038719177246094,
      "learning_rate": 0.00015073000417145242,
      "loss": 1.1604,
      "step": 172450
    },
    {
      "epoch": 5.15032991968471,
      "grad_norm": 3.903709888458252,
      "learning_rate": 0.0001507157183755336,
      "loss": 1.1504,
      "step": 172500
    },
    {
      "epoch": 5.151822768936793,
      "grad_norm": 5.332316875457764,
      "learning_rate": 0.00015070143257961475,
      "loss": 1.1616,
      "step": 172550
    },
    {
      "epoch": 5.153315618188875,
      "grad_norm": 4.094432353973389,
      "learning_rate": 0.00015068714678369592,
      "loss": 1.1771,
      "step": 172600
    },
    {
      "epoch": 5.154808467440958,
      "grad_norm": 4.575977325439453,
      "learning_rate": 0.00015067286098777708,
      "loss": 1.1587,
      "step": 172650
    },
    {
      "epoch": 5.156301316693041,
      "grad_norm": 4.0832414627075195,
      "learning_rate": 0.00015065857519185825,
      "loss": 1.1432,
      "step": 172700
    },
    {
      "epoch": 5.1577941659451225,
      "grad_norm": 3.969329595565796,
      "learning_rate": 0.0001506442893959394,
      "loss": 1.1735,
      "step": 172750
    },
    {
      "epoch": 5.159287015197205,
      "grad_norm": 4.063047885894775,
      "learning_rate": 0.00015063000360002058,
      "loss": 1.1071,
      "step": 172800
    },
    {
      "epoch": 5.160779864449288,
      "grad_norm": 4.360136985778809,
      "learning_rate": 0.00015061571780410174,
      "loss": 1.154,
      "step": 172850
    },
    {
      "epoch": 5.162272713701371,
      "grad_norm": 3.850008010864258,
      "learning_rate": 0.0001506014320081829,
      "loss": 1.2022,
      "step": 172900
    },
    {
      "epoch": 5.163765562953453,
      "grad_norm": 4.514410972595215,
      "learning_rate": 0.0001505871462122641,
      "loss": 1.1265,
      "step": 172950
    },
    {
      "epoch": 5.165258412205535,
      "grad_norm": 4.641385555267334,
      "learning_rate": 0.00015057286041634524,
      "loss": 1.1373,
      "step": 173000
    },
    {
      "epoch": 5.166751261457618,
      "grad_norm": 4.439742088317871,
      "learning_rate": 0.00015055857462042643,
      "loss": 1.1792,
      "step": 173050
    },
    {
      "epoch": 5.168244110709701,
      "grad_norm": 4.819581985473633,
      "learning_rate": 0.00015054428882450756,
      "loss": 1.1055,
      "step": 173100
    },
    {
      "epoch": 5.169736959961783,
      "grad_norm": 4.719875335693359,
      "learning_rate": 0.00015053000302858876,
      "loss": 1.1097,
      "step": 173150
    },
    {
      "epoch": 5.171229809213866,
      "grad_norm": 3.7248473167419434,
      "learning_rate": 0.0001505157172326699,
      "loss": 1.1809,
      "step": 173200
    },
    {
      "epoch": 5.172722658465948,
      "grad_norm": 4.468180179595947,
      "learning_rate": 0.00015050143143675109,
      "loss": 1.1218,
      "step": 173250
    },
    {
      "epoch": 5.1742155077180305,
      "grad_norm": 5.018815994262695,
      "learning_rate": 0.00015048714564083225,
      "loss": 1.1562,
      "step": 173300
    },
    {
      "epoch": 5.175708356970113,
      "grad_norm": 5.088767051696777,
      "learning_rate": 0.00015047285984491342,
      "loss": 1.0996,
      "step": 173350
    },
    {
      "epoch": 5.177201206222196,
      "grad_norm": 5.453111171722412,
      "learning_rate": 0.00015045857404899458,
      "loss": 1.2058,
      "step": 173400
    },
    {
      "epoch": 5.178694055474279,
      "grad_norm": 5.479541778564453,
      "learning_rate": 0.00015044428825307574,
      "loss": 1.1409,
      "step": 173450
    },
    {
      "epoch": 5.1801869047263605,
      "grad_norm": 5.977447509765625,
      "learning_rate": 0.0001504300024571569,
      "loss": 1.1197,
      "step": 173500
    },
    {
      "epoch": 5.181679753978443,
      "grad_norm": 5.65791130065918,
      "learning_rate": 0.00015041571666123807,
      "loss": 1.1486,
      "step": 173550
    },
    {
      "epoch": 5.183172603230526,
      "grad_norm": 6.653776168823242,
      "learning_rate": 0.00015040143086531924,
      "loss": 1.1734,
      "step": 173600
    },
    {
      "epoch": 5.184665452482609,
      "grad_norm": 4.445573806762695,
      "learning_rate": 0.0001503871450694004,
      "loss": 1.1776,
      "step": 173650
    },
    {
      "epoch": 5.186158301734691,
      "grad_norm": 4.370992183685303,
      "learning_rate": 0.00015037285927348157,
      "loss": 1.1274,
      "step": 173700
    },
    {
      "epoch": 5.187651150986773,
      "grad_norm": 4.479329586029053,
      "learning_rate": 0.00015035857347756273,
      "loss": 1.1387,
      "step": 173750
    },
    {
      "epoch": 5.189144000238856,
      "grad_norm": 4.528141498565674,
      "learning_rate": 0.0001503442876816439,
      "loss": 1.1586,
      "step": 173800
    },
    {
      "epoch": 5.1906368494909385,
      "grad_norm": 4.396657943725586,
      "learning_rate": 0.00015033000188572506,
      "loss": 1.1918,
      "step": 173850
    },
    {
      "epoch": 5.192129698743021,
      "grad_norm": 4.700211048126221,
      "learning_rate": 0.00015031571608980623,
      "loss": 1.144,
      "step": 173900
    },
    {
      "epoch": 5.193622547995103,
      "grad_norm": 3.748830795288086,
      "learning_rate": 0.0001503014302938874,
      "loss": 1.1584,
      "step": 173950
    },
    {
      "epoch": 5.195115397247186,
      "grad_norm": 4.841403484344482,
      "learning_rate": 0.00015028714449796856,
      "loss": 1.1419,
      "step": 174000
    },
    {
      "epoch": 5.1966082464992684,
      "grad_norm": 5.811513423919678,
      "learning_rate": 0.00015027285870204972,
      "loss": 1.1155,
      "step": 174050
    },
    {
      "epoch": 5.198101095751351,
      "grad_norm": 4.072244167327881,
      "learning_rate": 0.0001502585729061309,
      "loss": 1.0956,
      "step": 174100
    },
    {
      "epoch": 5.199593945003434,
      "grad_norm": 4.929732799530029,
      "learning_rate": 0.00015024428711021205,
      "loss": 1.1761,
      "step": 174150
    },
    {
      "epoch": 5.201086794255516,
      "grad_norm": 6.669888019561768,
      "learning_rate": 0.00015023000131429324,
      "loss": 1.1291,
      "step": 174200
    },
    {
      "epoch": 5.202579643507598,
      "grad_norm": 4.264486789703369,
      "learning_rate": 0.00015021571551837438,
      "loss": 1.1596,
      "step": 174250
    },
    {
      "epoch": 5.204072492759681,
      "grad_norm": 4.881564140319824,
      "learning_rate": 0.00015020142972245557,
      "loss": 1.1884,
      "step": 174300
    },
    {
      "epoch": 5.205565342011764,
      "grad_norm": 5.406867504119873,
      "learning_rate": 0.00015018714392653674,
      "loss": 1.1987,
      "step": 174350
    },
    {
      "epoch": 5.2070581912638465,
      "grad_norm": 5.110748767852783,
      "learning_rate": 0.0001501728581306179,
      "loss": 1.1758,
      "step": 174400
    },
    {
      "epoch": 5.208551040515928,
      "grad_norm": 4.510517120361328,
      "learning_rate": 0.00015015857233469906,
      "loss": 1.1746,
      "step": 174450
    },
    {
      "epoch": 5.210043889768011,
      "grad_norm": 4.028598308563232,
      "learning_rate": 0.00015014428653878023,
      "loss": 1.1727,
      "step": 174500
    },
    {
      "epoch": 5.211536739020094,
      "grad_norm": 5.600837707519531,
      "learning_rate": 0.0001501300007428614,
      "loss": 1.2241,
      "step": 174550
    },
    {
      "epoch": 5.213029588272176,
      "grad_norm": 5.198084831237793,
      "learning_rate": 0.00015011571494694256,
      "loss": 1.1599,
      "step": 174600
    },
    {
      "epoch": 5.214522437524259,
      "grad_norm": 4.601158142089844,
      "learning_rate": 0.00015010142915102372,
      "loss": 1.1764,
      "step": 174650
    },
    {
      "epoch": 5.216015286776341,
      "grad_norm": 5.581248760223389,
      "learning_rate": 0.0001500871433551049,
      "loss": 1.174,
      "step": 174700
    },
    {
      "epoch": 5.217508136028424,
      "grad_norm": 6.802553176879883,
      "learning_rate": 0.00015007285755918605,
      "loss": 1.1922,
      "step": 174750
    },
    {
      "epoch": 5.219000985280506,
      "grad_norm": 4.5715651512146,
      "learning_rate": 0.00015005857176326722,
      "loss": 1.1835,
      "step": 174800
    },
    {
      "epoch": 5.220493834532589,
      "grad_norm": 3.920050621032715,
      "learning_rate": 0.00015004428596734838,
      "loss": 1.1529,
      "step": 174850
    },
    {
      "epoch": 5.221986683784672,
      "grad_norm": 5.379642009735107,
      "learning_rate": 0.00015003000017142957,
      "loss": 1.1016,
      "step": 174900
    },
    {
      "epoch": 5.223479533036754,
      "grad_norm": 5.268077373504639,
      "learning_rate": 0.0001500157143755107,
      "loss": 1.1792,
      "step": 174950
    },
    {
      "epoch": 5.224972382288836,
      "grad_norm": 4.864146709442139,
      "learning_rate": 0.0001500014285795919,
      "loss": 1.1511,
      "step": 175000
    },
    {
      "epoch": 5.226465231540919,
      "grad_norm": 3.050143241882324,
      "learning_rate": 0.00014998714278367304,
      "loss": 1.2307,
      "step": 175050
    },
    {
      "epoch": 5.227958080793002,
      "grad_norm": 4.5110764503479,
      "learning_rate": 0.00014997285698775423,
      "loss": 1.1558,
      "step": 175100
    },
    {
      "epoch": 5.229450930045084,
      "grad_norm": 4.925455093383789,
      "learning_rate": 0.0001499585711918354,
      "loss": 1.194,
      "step": 175150
    },
    {
      "epoch": 5.230943779297166,
      "grad_norm": 5.321854591369629,
      "learning_rate": 0.00014994428539591656,
      "loss": 1.1927,
      "step": 175200
    },
    {
      "epoch": 5.232436628549249,
      "grad_norm": 3.1668155193328857,
      "learning_rate": 0.00014992999959999773,
      "loss": 1.1243,
      "step": 175250
    },
    {
      "epoch": 5.233929477801332,
      "grad_norm": 5.259143829345703,
      "learning_rate": 0.0001499157138040789,
      "loss": 1.181,
      "step": 175300
    },
    {
      "epoch": 5.235422327053414,
      "grad_norm": 5.783114910125732,
      "learning_rate": 0.00014990142800816006,
      "loss": 1.1156,
      "step": 175350
    },
    {
      "epoch": 5.236915176305497,
      "grad_norm": 5.6330952644348145,
      "learning_rate": 0.00014988714221224122,
      "loss": 1.1513,
      "step": 175400
    },
    {
      "epoch": 5.238408025557579,
      "grad_norm": 4.543177127838135,
      "learning_rate": 0.00014987285641632238,
      "loss": 1.1898,
      "step": 175450
    },
    {
      "epoch": 5.239900874809662,
      "grad_norm": 3.058159828186035,
      "learning_rate": 0.00014985857062040355,
      "loss": 1.2119,
      "step": 175500
    },
    {
      "epoch": 5.241393724061744,
      "grad_norm": 5.873176097869873,
      "learning_rate": 0.00014984428482448471,
      "loss": 1.1351,
      "step": 175550
    },
    {
      "epoch": 5.242886573313827,
      "grad_norm": 4.126943588256836,
      "learning_rate": 0.0001498299990285659,
      "loss": 1.2393,
      "step": 175600
    },
    {
      "epoch": 5.24437942256591,
      "grad_norm": 5.249665260314941,
      "learning_rate": 0.00014981571323264704,
      "loss": 1.2033,
      "step": 175650
    },
    {
      "epoch": 5.2458722718179915,
      "grad_norm": 6.529717445373535,
      "learning_rate": 0.00014980142743672824,
      "loss": 1.1383,
      "step": 175700
    },
    {
      "epoch": 5.247365121070074,
      "grad_norm": 4.902076244354248,
      "learning_rate": 0.00014978714164080937,
      "loss": 1.1252,
      "step": 175750
    },
    {
      "epoch": 5.248857970322157,
      "grad_norm": 5.3138017654418945,
      "learning_rate": 0.00014977285584489056,
      "loss": 1.1629,
      "step": 175800
    },
    {
      "epoch": 5.25035081957424,
      "grad_norm": 4.891197204589844,
      "learning_rate": 0.0001497585700489717,
      "loss": 1.144,
      "step": 175850
    },
    {
      "epoch": 5.2518436688263215,
      "grad_norm": 5.521958827972412,
      "learning_rate": 0.0001497442842530529,
      "loss": 1.2136,
      "step": 175900
    },
    {
      "epoch": 5.253336518078404,
      "grad_norm": 3.749346971511841,
      "learning_rate": 0.00014972999845713406,
      "loss": 1.1887,
      "step": 175950
    },
    {
      "epoch": 5.254829367330487,
      "grad_norm": 3.947131633758545,
      "learning_rate": 0.00014971571266121522,
      "loss": 1.1959,
      "step": 176000
    },
    {
      "epoch": 5.25632221658257,
      "grad_norm": 3.47645902633667,
      "learning_rate": 0.0001497014268652964,
      "loss": 1.1597,
      "step": 176050
    },
    {
      "epoch": 5.257815065834652,
      "grad_norm": 3.3379621505737305,
      "learning_rate": 0.00014968714106937755,
      "loss": 1.1643,
      "step": 176100
    },
    {
      "epoch": 5.259307915086735,
      "grad_norm": 3.426164388656616,
      "learning_rate": 0.00014967285527345872,
      "loss": 1.131,
      "step": 176150
    },
    {
      "epoch": 5.260800764338817,
      "grad_norm": 4.857641696929932,
      "learning_rate": 0.00014965856947753988,
      "loss": 1.1171,
      "step": 176200
    },
    {
      "epoch": 5.2622936135908995,
      "grad_norm": 3.9440646171569824,
      "learning_rate": 0.00014964428368162105,
      "loss": 1.1529,
      "step": 176250
    },
    {
      "epoch": 5.263786462842982,
      "grad_norm": 4.392827987670898,
      "learning_rate": 0.0001496299978857022,
      "loss": 1.1946,
      "step": 176300
    },
    {
      "epoch": 5.265279312095065,
      "grad_norm": 4.131694316864014,
      "learning_rate": 0.00014961571208978338,
      "loss": 1.1702,
      "step": 176350
    },
    {
      "epoch": 5.266772161347147,
      "grad_norm": 4.955690383911133,
      "learning_rate": 0.00014960142629386454,
      "loss": 1.1059,
      "step": 176400
    },
    {
      "epoch": 5.2682650105992295,
      "grad_norm": 3.52262544631958,
      "learning_rate": 0.0001495871404979457,
      "loss": 1.1873,
      "step": 176450
    },
    {
      "epoch": 5.269757859851312,
      "grad_norm": 5.748971939086914,
      "learning_rate": 0.00014957285470202687,
      "loss": 1.1285,
      "step": 176500
    },
    {
      "epoch": 5.271250709103395,
      "grad_norm": 5.502897262573242,
      "learning_rate": 0.00014955856890610803,
      "loss": 1.1292,
      "step": 176550
    },
    {
      "epoch": 5.272743558355478,
      "grad_norm": 3.8092453479766846,
      "learning_rate": 0.0001495442831101892,
      "loss": 1.1545,
      "step": 176600
    },
    {
      "epoch": 5.274236407607559,
      "grad_norm": 4.742307186126709,
      "learning_rate": 0.00014952999731427036,
      "loss": 1.1732,
      "step": 176650
    },
    {
      "epoch": 5.275729256859642,
      "grad_norm": 5.108506202697754,
      "learning_rate": 0.00014951571151835153,
      "loss": 1.2033,
      "step": 176700
    },
    {
      "epoch": 5.277222106111725,
      "grad_norm": 3.655372381210327,
      "learning_rate": 0.00014950142572243272,
      "loss": 1.1837,
      "step": 176750
    },
    {
      "epoch": 5.2787149553638075,
      "grad_norm": 4.246975898742676,
      "learning_rate": 0.00014948713992651386,
      "loss": 1.1572,
      "step": 176800
    },
    {
      "epoch": 5.28020780461589,
      "grad_norm": 3.185184955596924,
      "learning_rate": 0.00014947285413059505,
      "loss": 1.1773,
      "step": 176850
    },
    {
      "epoch": 5.281700653867972,
      "grad_norm": 4.376567363739014,
      "learning_rate": 0.0001494585683346762,
      "loss": 1.1662,
      "step": 176900
    },
    {
      "epoch": 5.283193503120055,
      "grad_norm": 4.7851457595825195,
      "learning_rate": 0.00014944428253875738,
      "loss": 1.2018,
      "step": 176950
    },
    {
      "epoch": 5.2846863523721375,
      "grad_norm": 4.13614559173584,
      "learning_rate": 0.00014942999674283852,
      "loss": 1.1137,
      "step": 177000
    },
    {
      "epoch": 5.28617920162422,
      "grad_norm": 5.2566609382629395,
      "learning_rate": 0.0001494157109469197,
      "loss": 1.1698,
      "step": 177050
    },
    {
      "epoch": 5.287672050876303,
      "grad_norm": 4.4063825607299805,
      "learning_rate": 0.00014940142515100087,
      "loss": 1.1507,
      "step": 177100
    },
    {
      "epoch": 5.289164900128385,
      "grad_norm": 6.457488059997559,
      "learning_rate": 0.00014938713935508204,
      "loss": 1.2139,
      "step": 177150
    },
    {
      "epoch": 5.290657749380467,
      "grad_norm": 3.639181137084961,
      "learning_rate": 0.0001493728535591632,
      "loss": 1.146,
      "step": 177200
    },
    {
      "epoch": 5.29215059863255,
      "grad_norm": 4.316925525665283,
      "learning_rate": 0.00014935856776324437,
      "loss": 1.1696,
      "step": 177250
    },
    {
      "epoch": 5.293643447884633,
      "grad_norm": 3.8231780529022217,
      "learning_rate": 0.00014934428196732553,
      "loss": 1.2053,
      "step": 177300
    },
    {
      "epoch": 5.2951362971367155,
      "grad_norm": 5.597898483276367,
      "learning_rate": 0.0001493299961714067,
      "loss": 1.2097,
      "step": 177350
    },
    {
      "epoch": 5.296629146388797,
      "grad_norm": 5.123006820678711,
      "learning_rate": 0.00014931571037548786,
      "loss": 1.1213,
      "step": 177400
    },
    {
      "epoch": 5.29812199564088,
      "grad_norm": 4.827045917510986,
      "learning_rate": 0.00014930142457956903,
      "loss": 1.1784,
      "step": 177450
    },
    {
      "epoch": 5.299614844892963,
      "grad_norm": 4.859231472015381,
      "learning_rate": 0.0001492871387836502,
      "loss": 1.1418,
      "step": 177500
    },
    {
      "epoch": 5.3011076941450455,
      "grad_norm": 5.134034633636475,
      "learning_rate": 0.00014927285298773138,
      "loss": 1.1989,
      "step": 177550
    },
    {
      "epoch": 5.302600543397128,
      "grad_norm": 6.3303327560424805,
      "learning_rate": 0.00014925856719181252,
      "loss": 1.1866,
      "step": 177600
    },
    {
      "epoch": 5.30409339264921,
      "grad_norm": 4.182743549346924,
      "learning_rate": 0.0001492442813958937,
      "loss": 1.1781,
      "step": 177650
    },
    {
      "epoch": 5.305586241901293,
      "grad_norm": 4.055047512054443,
      "learning_rate": 0.00014922999559997485,
      "loss": 1.1735,
      "step": 177700
    },
    {
      "epoch": 5.307079091153375,
      "grad_norm": 5.255928993225098,
      "learning_rate": 0.00014921570980405604,
      "loss": 1.1627,
      "step": 177750
    },
    {
      "epoch": 5.308571940405458,
      "grad_norm": 4.9416584968566895,
      "learning_rate": 0.0001492014240081372,
      "loss": 1.0982,
      "step": 177800
    },
    {
      "epoch": 5.31006478965754,
      "grad_norm": 6.426959037780762,
      "learning_rate": 0.00014918713821221837,
      "loss": 1.1585,
      "step": 177850
    },
    {
      "epoch": 5.311557638909623,
      "grad_norm": 3.433701753616333,
      "learning_rate": 0.00014917285241629953,
      "loss": 1.1728,
      "step": 177900
    },
    {
      "epoch": 5.313050488161705,
      "grad_norm": 4.7097320556640625,
      "learning_rate": 0.0001491585666203807,
      "loss": 1.1183,
      "step": 177950
    },
    {
      "epoch": 5.314543337413788,
      "grad_norm": 5.154112815856934,
      "learning_rate": 0.00014914428082446186,
      "loss": 1.2097,
      "step": 178000
    },
    {
      "epoch": 5.316036186665871,
      "grad_norm": 4.5182929039001465,
      "learning_rate": 0.00014912999502854303,
      "loss": 1.2144,
      "step": 178050
    },
    {
      "epoch": 5.3175290359179535,
      "grad_norm": 4.899813175201416,
      "learning_rate": 0.0001491157092326242,
      "loss": 1.1513,
      "step": 178100
    },
    {
      "epoch": 5.319021885170035,
      "grad_norm": 4.790624141693115,
      "learning_rate": 0.00014910142343670536,
      "loss": 1.1579,
      "step": 178150
    },
    {
      "epoch": 5.320514734422118,
      "grad_norm": 4.103714466094971,
      "learning_rate": 0.00014908713764078652,
      "loss": 1.1407,
      "step": 178200
    },
    {
      "epoch": 5.322007583674201,
      "grad_norm": 3.641630172729492,
      "learning_rate": 0.0001490728518448677,
      "loss": 1.1401,
      "step": 178250
    },
    {
      "epoch": 5.323500432926283,
      "grad_norm": 4.197120666503906,
      "learning_rate": 0.00014905856604894885,
      "loss": 1.151,
      "step": 178300
    },
    {
      "epoch": 5.324993282178365,
      "grad_norm": 6.3722405433654785,
      "learning_rate": 0.00014904428025303004,
      "loss": 1.2082,
      "step": 178350
    },
    {
      "epoch": 5.326486131430448,
      "grad_norm": 6.299606800079346,
      "learning_rate": 0.00014902999445711118,
      "loss": 1.2194,
      "step": 178400
    },
    {
      "epoch": 5.327978980682531,
      "grad_norm": 4.542901992797852,
      "learning_rate": 0.00014901570866119237,
      "loss": 1.199,
      "step": 178450
    },
    {
      "epoch": 5.329471829934613,
      "grad_norm": 5.075666904449463,
      "learning_rate": 0.0001490014228652735,
      "loss": 1.1677,
      "step": 178500
    },
    {
      "epoch": 5.330964679186696,
      "grad_norm": 4.868073463439941,
      "learning_rate": 0.0001489871370693547,
      "loss": 1.1416,
      "step": 178550
    },
    {
      "epoch": 5.332457528438778,
      "grad_norm": 4.414193630218506,
      "learning_rate": 0.00014897285127343587,
      "loss": 1.1757,
      "step": 178600
    },
    {
      "epoch": 5.3339503776908606,
      "grad_norm": 4.902929306030273,
      "learning_rate": 0.00014895856547751703,
      "loss": 1.1555,
      "step": 178650
    },
    {
      "epoch": 5.335443226942943,
      "grad_norm": 3.7560372352600098,
      "learning_rate": 0.0001489442796815982,
      "loss": 1.1768,
      "step": 178700
    },
    {
      "epoch": 5.336936076195026,
      "grad_norm": 4.610017776489258,
      "learning_rate": 0.00014892999388567936,
      "loss": 1.1467,
      "step": 178750
    },
    {
      "epoch": 5.338428925447109,
      "grad_norm": 3.9438891410827637,
      "learning_rate": 0.00014891570808976053,
      "loss": 1.1995,
      "step": 178800
    },
    {
      "epoch": 5.3399217746991905,
      "grad_norm": 4.497828483581543,
      "learning_rate": 0.0001489014222938417,
      "loss": 1.1962,
      "step": 178850
    },
    {
      "epoch": 5.341414623951273,
      "grad_norm": 6.109000205993652,
      "learning_rate": 0.00014888713649792285,
      "loss": 1.1826,
      "step": 178900
    },
    {
      "epoch": 5.342907473203356,
      "grad_norm": 4.150059223175049,
      "learning_rate": 0.00014887285070200402,
      "loss": 1.1613,
      "step": 178950
    },
    {
      "epoch": 5.344400322455439,
      "grad_norm": 4.3024492263793945,
      "learning_rate": 0.00014885856490608518,
      "loss": 1.1438,
      "step": 179000
    },
    {
      "epoch": 5.345893171707521,
      "grad_norm": 4.783211708068848,
      "learning_rate": 0.00014884427911016635,
      "loss": 1.1864,
      "step": 179050
    },
    {
      "epoch": 5.347386020959603,
      "grad_norm": 5.038503646850586,
      "learning_rate": 0.0001488299933142475,
      "loss": 1.2297,
      "step": 179100
    },
    {
      "epoch": 5.348878870211686,
      "grad_norm": 5.009586334228516,
      "learning_rate": 0.00014881570751832868,
      "loss": 1.1509,
      "step": 179150
    },
    {
      "epoch": 5.3503717194637685,
      "grad_norm": 5.057339191436768,
      "learning_rate": 0.00014880142172240984,
      "loss": 1.162,
      "step": 179200
    },
    {
      "epoch": 5.351864568715851,
      "grad_norm": 7.249011993408203,
      "learning_rate": 0.000148787135926491,
      "loss": 1.1634,
      "step": 179250
    },
    {
      "epoch": 5.353357417967934,
      "grad_norm": 4.474646091461182,
      "learning_rate": 0.00014877285013057217,
      "loss": 1.1652,
      "step": 179300
    },
    {
      "epoch": 5.354850267220016,
      "grad_norm": 4.699128150939941,
      "learning_rate": 0.00014875856433465334,
      "loss": 1.1875,
      "step": 179350
    },
    {
      "epoch": 5.3563431164720985,
      "grad_norm": 3.6825220584869385,
      "learning_rate": 0.00014874427853873453,
      "loss": 1.1896,
      "step": 179400
    },
    {
      "epoch": 5.357835965724181,
      "grad_norm": 5.787204742431641,
      "learning_rate": 0.00014872999274281567,
      "loss": 1.1793,
      "step": 179450
    },
    {
      "epoch": 5.359328814976264,
      "grad_norm": 6.388000011444092,
      "learning_rate": 0.00014871570694689686,
      "loss": 1.2219,
      "step": 179500
    },
    {
      "epoch": 5.360821664228347,
      "grad_norm": 4.790463924407959,
      "learning_rate": 0.000148701421150978,
      "loss": 1.1739,
      "step": 179550
    },
    {
      "epoch": 5.362314513480428,
      "grad_norm": 2.8773367404937744,
      "learning_rate": 0.0001486871353550592,
      "loss": 1.1894,
      "step": 179600
    },
    {
      "epoch": 5.363807362732511,
      "grad_norm": 3.305040121078491,
      "learning_rate": 0.00014867284955914032,
      "loss": 1.2278,
      "step": 179650
    },
    {
      "epoch": 5.365300211984594,
      "grad_norm": 4.368061065673828,
      "learning_rate": 0.00014865856376322152,
      "loss": 1.214,
      "step": 179700
    },
    {
      "epoch": 5.3667930612366765,
      "grad_norm": 5.629176139831543,
      "learning_rate": 0.00014864427796730268,
      "loss": 1.1611,
      "step": 179750
    },
    {
      "epoch": 5.368285910488759,
      "grad_norm": 5.811650276184082,
      "learning_rate": 0.00014862999217138385,
      "loss": 1.1861,
      "step": 179800
    },
    {
      "epoch": 5.369778759740841,
      "grad_norm": 3.6000754833221436,
      "learning_rate": 0.000148615706375465,
      "loss": 1.1636,
      "step": 179850
    },
    {
      "epoch": 5.371271608992924,
      "grad_norm": 4.922330379486084,
      "learning_rate": 0.00014860142057954617,
      "loss": 1.1534,
      "step": 179900
    },
    {
      "epoch": 5.3727644582450065,
      "grad_norm": 5.780492305755615,
      "learning_rate": 0.00014858713478362734,
      "loss": 1.2066,
      "step": 179950
    },
    {
      "epoch": 5.374257307497089,
      "grad_norm": 4.670080184936523,
      "learning_rate": 0.0001485728489877085,
      "loss": 1.1812,
      "step": 180000
    }
  ],
  "logging_steps": 50,
  "max_steps": 700001,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 21,
  "save_steps": 10000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 4.5513784407443374e+18,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}