{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.9676393598144797,
  "eval_steps": 500,
  "global_step": 28000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.007027283427908856,
      "grad_norm": 0.5592142939567566,
      "learning_rate": 0.0001995548787630315,
      "loss": 1.0457,
      "step": 100
    },
    {
      "epoch": 0.014054566855817713,
      "grad_norm": 0.6605722308158875,
      "learning_rate": 0.00019908633009253836,
      "loss": 0.8075,
      "step": 200
    },
    {
      "epoch": 0.02108185028372657,
      "grad_norm": 0.5018069744110107,
      "learning_rate": 0.00019861778142204523,
      "loss": 0.7927,
      "step": 300
    },
    {
      "epoch": 0.028109133711635426,
      "grad_norm": 0.5429800152778625,
      "learning_rate": 0.00019814923275155208,
      "loss": 0.7648,
      "step": 400
    },
    {
      "epoch": 0.03513641713954428,
      "grad_norm": 0.5747570395469666,
      "learning_rate": 0.00019768068408105893,
      "loss": 0.7732,
      "step": 500
    },
    {
      "epoch": 0.03513641713954428,
      "eval_loss": 0.7362164855003357,
      "eval_runtime": 118.3556,
      "eval_samples_per_second": 106.873,
      "eval_steps_per_second": 13.367,
      "step": 500
    },
    {
      "epoch": 0.04216370056745314,
      "grad_norm": 0.6291443109512329,
      "learning_rate": 0.00019721213541056578,
      "loss": 0.7614,
      "step": 600
    },
    {
      "epoch": 0.049190983995361995,
      "grad_norm": 0.6160371899604797,
      "learning_rate": 0.00019674358674007263,
      "loss": 0.7489,
      "step": 700
    },
    {
      "epoch": 0.05621826742327085,
      "grad_norm": 0.5804228782653809,
      "learning_rate": 0.00019627503806957949,
      "loss": 0.7399,
      "step": 800
    },
    {
      "epoch": 0.06324555085117971,
      "grad_norm": 0.6000151634216309,
      "learning_rate": 0.00019580648939908634,
      "loss": 0.7349,
      "step": 900
    },
    {
      "epoch": 0.07027283427908856,
      "grad_norm": 0.5746223330497742,
      "learning_rate": 0.00019533794072859319,
      "loss": 0.7282,
      "step": 1000
    },
    {
      "epoch": 0.07027283427908856,
      "eval_loss": 0.7145671248435974,
      "eval_runtime": 118.2812,
      "eval_samples_per_second": 106.94,
      "eval_steps_per_second": 13.375,
      "step": 1000
    },
    {
      "epoch": 0.07730011770699742,
      "grad_norm": 0.6321994066238403,
      "learning_rate": 0.00019486939205810004,
      "loss": 0.7241,
      "step": 1100
    },
    {
      "epoch": 0.08432740113490628,
      "grad_norm": 0.7580232620239258,
      "learning_rate": 0.00019440084338760689,
      "loss": 0.7254,
      "step": 1200
    },
    {
      "epoch": 0.09135468456281513,
      "grad_norm": 0.6181788444519043,
      "learning_rate": 0.00019393229471711376,
      "loss": 0.7349,
      "step": 1300
    },
    {
      "epoch": 0.09838196799072399,
      "grad_norm": 0.6435703635215759,
      "learning_rate": 0.00019346374604662061,
      "loss": 0.7082,
      "step": 1400
    },
    {
      "epoch": 0.10540925141863285,
      "grad_norm": 0.5934786200523376,
      "learning_rate": 0.00019299519737612746,
      "loss": 0.7145,
      "step": 1500
    },
    {
      "epoch": 0.10540925141863285,
      "eval_loss": 0.6975060701370239,
      "eval_runtime": 118.4288,
      "eval_samples_per_second": 106.807,
      "eval_steps_per_second": 13.358,
      "step": 1500
    },
    {
      "epoch": 0.1124365348465417,
      "grad_norm": 0.4486166536808014,
      "learning_rate": 0.00019252664870563431,
      "loss": 0.7226,
      "step": 1600
    },
    {
      "epoch": 0.11946381827445056,
      "grad_norm": 0.4868922531604767,
      "learning_rate": 0.00019205810003514116,
      "loss": 0.6951,
      "step": 1700
    },
    {
      "epoch": 0.12649110170235942,
      "grad_norm": 0.5752139687538147,
      "learning_rate": 0.00019158955136464801,
      "loss": 0.7117,
      "step": 1800
    },
    {
      "epoch": 0.13351838513026826,
      "grad_norm": 0.5520344972610474,
      "learning_rate": 0.00019112100269415486,
      "loss": 0.7111,
      "step": 1900
    },
    {
      "epoch": 0.14054566855817713,
      "grad_norm": 0.5359894633293152,
      "learning_rate": 0.00019065245402366172,
      "loss": 0.7093,
      "step": 2000
    },
    {
      "epoch": 0.14054566855817713,
      "eval_loss": 0.687991201877594,
      "eval_runtime": 118.7631,
      "eval_samples_per_second": 106.506,
      "eval_steps_per_second": 13.321,
      "step": 2000
    },
    {
      "epoch": 0.14757295198608597,
      "grad_norm": 0.5415408611297607,
      "learning_rate": 0.00019018390535316857,
      "loss": 0.7117,
      "step": 2100
    },
    {
      "epoch": 0.15460023541399484,
      "grad_norm": 0.6338439583778381,
      "learning_rate": 0.00018971535668267542,
      "loss": 0.7253,
      "step": 2200
    },
    {
      "epoch": 0.16162751884190368,
      "grad_norm": 0.6666418313980103,
      "learning_rate": 0.00018924680801218227,
      "loss": 0.7058,
      "step": 2300
    },
    {
      "epoch": 0.16865480226981255,
      "grad_norm": 0.5344674587249756,
      "learning_rate": 0.00018877825934168914,
      "loss": 0.7036,
      "step": 2400
    },
    {
      "epoch": 0.1756820856977214,
      "grad_norm": 0.5522785186767578,
      "learning_rate": 0.000188309710671196,
      "loss": 0.6809,
      "step": 2500
    },
    {
      "epoch": 0.1756820856977214,
      "eval_loss": 0.679720938205719,
      "eval_runtime": 118.6143,
      "eval_samples_per_second": 106.64,
      "eval_steps_per_second": 13.337,
      "step": 2500
    },
    {
      "epoch": 0.18270936912563027,
      "grad_norm": 0.7211841344833374,
      "learning_rate": 0.00018784116200070284,
      "loss": 0.6972,
      "step": 2600
    },
    {
      "epoch": 0.1897366525535391,
      "grad_norm": 0.5469601154327393,
      "learning_rate": 0.00018737261333020967,
      "loss": 0.6996,
      "step": 2700
    },
    {
      "epoch": 0.19676393598144798,
      "grad_norm": 0.5906969904899597,
      "learning_rate": 0.00018690406465971652,
      "loss": 0.6828,
      "step": 2800
    },
    {
      "epoch": 0.20379121940935682,
      "grad_norm": 0.703484833240509,
      "learning_rate": 0.0001864355159892234,
      "loss": 0.6906,
      "step": 2900
    },
    {
      "epoch": 0.2108185028372657,
      "grad_norm": 0.5544711947441101,
      "learning_rate": 0.00018596696731873024,
      "loss": 0.6891,
      "step": 3000
    },
    {
      "epoch": 0.2108185028372657,
      "eval_loss": 0.6728695034980774,
      "eval_runtime": 118.909,
      "eval_samples_per_second": 106.375,
      "eval_steps_per_second": 13.304,
      "step": 3000
    },
    {
      "epoch": 0.21784578626517453,
      "grad_norm": 0.5521793365478516,
      "learning_rate": 0.0001854984186482371,
      "loss": 0.6777,
      "step": 3100
    },
    {
      "epoch": 0.2248730696930834,
      "grad_norm": 0.4634329080581665,
      "learning_rate": 0.00018502986997774394,
      "loss": 0.6744,
      "step": 3200
    },
    {
      "epoch": 0.23190035312099225,
      "grad_norm": 0.7366927266120911,
      "learning_rate": 0.0001845613213072508,
      "loss": 0.6642,
      "step": 3300
    },
    {
      "epoch": 0.23892763654890112,
      "grad_norm": 0.5567039251327515,
      "learning_rate": 0.00018409277263675767,
      "loss": 0.6895,
      "step": 3400
    },
    {
      "epoch": 0.24595491997680996,
      "grad_norm": 0.6940245628356934,
      "learning_rate": 0.0001836242239662645,
      "loss": 0.6708,
      "step": 3500
    },
    {
      "epoch": 0.24595491997680996,
      "eval_loss": 0.6706892848014832,
      "eval_runtime": 118.4648,
      "eval_samples_per_second": 106.774,
      "eval_steps_per_second": 13.354,
      "step": 3500
    },
    {
      "epoch": 0.25298220340471883,
      "grad_norm": 0.5577242970466614,
      "learning_rate": 0.00018315567529577135,
      "loss": 0.679,
      "step": 3600
    },
    {
      "epoch": 0.2600094868326277,
      "grad_norm": 0.6587842702865601,
      "learning_rate": 0.0001826871266252782,
      "loss": 0.6918,
      "step": 3700
    },
    {
      "epoch": 0.2670367702605365,
      "grad_norm": 0.6200099587440491,
      "learning_rate": 0.00018221857795478505,
      "loss": 0.7002,
      "step": 3800
    },
    {
      "epoch": 0.2740640536884454,
      "grad_norm": 0.5648295879364014,
      "learning_rate": 0.00018175002928429192,
      "loss": 0.6791,
      "step": 3900
    },
    {
      "epoch": 0.28109133711635426,
      "grad_norm": 0.5908897519111633,
      "learning_rate": 0.00018128148061379877,
      "loss": 0.6839,
      "step": 4000
    },
    {
      "epoch": 0.28109133711635426,
      "eval_loss": 0.664506196975708,
      "eval_runtime": 119.1229,
      "eval_samples_per_second": 106.184,
      "eval_steps_per_second": 13.28,
      "step": 4000
    },
    {
      "epoch": 0.2881186205442631,
      "grad_norm": 0.5778653621673584,
      "learning_rate": 0.00018081293194330562,
      "loss": 0.6654,
      "step": 4100
    },
    {
      "epoch": 0.29514590397217194,
      "grad_norm": 0.700835645198822,
      "learning_rate": 0.00018034438327281247,
      "loss": 0.6897,
      "step": 4200
    },
    {
      "epoch": 0.3021731874000808,
      "grad_norm": 0.5250533819198608,
      "learning_rate": 0.00017987583460231932,
      "loss": 0.6934,
      "step": 4300
    },
    {
      "epoch": 0.3092004708279897,
      "grad_norm": 0.725397527217865,
      "learning_rate": 0.00017940728593182617,
      "loss": 0.6781,
      "step": 4400
    },
    {
      "epoch": 0.31622775425589855,
      "grad_norm": 0.5805392265319824,
      "learning_rate": 0.00017893873726133302,
      "loss": 0.6866,
      "step": 4500
    },
    {
      "epoch": 0.31622775425589855,
      "eval_loss": 0.65822434425354,
      "eval_runtime": 118.7959,
      "eval_samples_per_second": 106.477,
      "eval_steps_per_second": 13.317,
      "step": 4500
    },
    {
      "epoch": 0.32325503768380737,
      "grad_norm": 0.6944029331207275,
      "learning_rate": 0.00017847018859083988,
      "loss": 0.6683,
      "step": 4600
    },
    {
      "epoch": 0.33028232111171624,
      "grad_norm": 0.5891593098640442,
      "learning_rate": 0.00017800163992034673,
      "loss": 0.6817,
      "step": 4700
    },
    {
      "epoch": 0.3373096045396251,
      "grad_norm": 0.6220216751098633,
      "learning_rate": 0.00017753309124985358,
      "loss": 0.6845,
      "step": 4800
    },
    {
      "epoch": 0.344336887967534,
      "grad_norm": 0.7002474665641785,
      "learning_rate": 0.00017706454257936045,
      "loss": 0.6551,
      "step": 4900
    },
    {
      "epoch": 0.3513641713954428,
      "grad_norm": 0.5766635537147522,
      "learning_rate": 0.0001765959939088673,
      "loss": 0.6741,
      "step": 5000
    },
    {
      "epoch": 0.3513641713954428,
      "eval_loss": 0.6550154685974121,
      "eval_runtime": 118.5512,
      "eval_samples_per_second": 106.697,
      "eval_steps_per_second": 13.344,
      "step": 5000
    },
    {
      "epoch": 0.35839145482335166,
      "grad_norm": 0.5734873414039612,
      "learning_rate": 0.00017612744523837415,
      "loss": 0.6583,
      "step": 5100
    },
    {
      "epoch": 0.36541873825126053,
      "grad_norm": 0.4953276515007019,
      "learning_rate": 0.000175658896567881,
      "loss": 0.6621,
      "step": 5200
    },
    {
      "epoch": 0.37244602167916935,
      "grad_norm": 0.7837636470794678,
      "learning_rate": 0.00017519034789738785,
      "loss": 0.6846,
      "step": 5300
    },
    {
      "epoch": 0.3794733051070782,
      "grad_norm": 0.718715488910675,
      "learning_rate": 0.0001747217992268947,
      "loss": 0.6776,
      "step": 5400
    },
    {
      "epoch": 0.3865005885349871,
      "grad_norm": 0.5844186544418335,
      "learning_rate": 0.00017425325055640155,
      "loss": 0.6627,
      "step": 5500
    },
    {
      "epoch": 0.3865005885349871,
      "eval_loss": 0.6532958149909973,
      "eval_runtime": 118.7133,
      "eval_samples_per_second": 106.551,
      "eval_steps_per_second": 13.326,
      "step": 5500
    },
    {
      "epoch": 0.39352787196289596,
      "grad_norm": 0.4426696300506592,
      "learning_rate": 0.0001737847018859084,
      "loss": 0.6546,
      "step": 5600
    },
    {
      "epoch": 0.4005551553908048,
      "grad_norm": 0.5954882502555847,
      "learning_rate": 0.00017331615321541525,
      "loss": 0.6448,
      "step": 5700
    },
    {
      "epoch": 0.40758243881871364,
      "grad_norm": 0.601349413394928,
      "learning_rate": 0.0001728476045449221,
      "loss": 0.6637,
      "step": 5800
    },
    {
      "epoch": 0.4146097222466225,
      "grad_norm": 0.6108406782150269,
      "learning_rate": 0.00017237905587442898,
      "loss": 0.6709,
      "step": 5900
    },
    {
      "epoch": 0.4216370056745314,
      "grad_norm": 0.6442033052444458,
      "learning_rate": 0.00017191050720393583,
      "loss": 0.6588,
      "step": 6000
    },
    {
      "epoch": 0.4216370056745314,
      "eval_loss": 0.6498988270759583,
      "eval_runtime": 118.8665,
      "eval_samples_per_second": 106.413,
      "eval_steps_per_second": 13.309,
      "step": 6000
    },
    {
      "epoch": 0.4286642891024402,
      "grad_norm": 0.4642776846885681,
      "learning_rate": 0.00017144195853344268,
      "loss": 0.6706,
      "step": 6100
    },
    {
      "epoch": 0.43569157253034907,
      "grad_norm": 0.6703388690948486,
      "learning_rate": 0.0001709734098629495,
      "loss": 0.6675,
      "step": 6200
    },
    {
      "epoch": 0.44271885595825794,
      "grad_norm": 0.6001936793327332,
      "learning_rate": 0.00017050486119245636,
      "loss": 0.6545,
      "step": 6300
    },
    {
      "epoch": 0.4497461393861668,
      "grad_norm": 0.7159720063209534,
      "learning_rate": 0.00017003631252196323,
      "loss": 0.6684,
      "step": 6400
    },
    {
      "epoch": 0.4567734228140756,
      "grad_norm": 0.5892972350120544,
      "learning_rate": 0.00016956776385147008,
      "loss": 0.6509,
      "step": 6500
    },
    {
      "epoch": 0.4567734228140756,
      "eval_loss": 0.6456841230392456,
      "eval_runtime": 118.8557,
      "eval_samples_per_second": 106.423,
      "eval_steps_per_second": 13.31,
      "step": 6500
    },
    {
      "epoch": 0.4638007062419845,
      "grad_norm": 0.6351083517074585,
      "learning_rate": 0.00016909921518097693,
      "loss": 0.634,
      "step": 6600
    },
    {
      "epoch": 0.47082798966989337,
      "grad_norm": 0.7939039468765259,
      "learning_rate": 0.00016863066651048378,
      "loss": 0.6595,
      "step": 6700
    },
    {
      "epoch": 0.47785527309780224,
      "grad_norm": 0.5831073522567749,
      "learning_rate": 0.00016816211783999063,
      "loss": 0.6444,
      "step": 6800
    },
    {
      "epoch": 0.48488255652571105,
      "grad_norm": 0.6314815282821655,
      "learning_rate": 0.0001676935691694975,
      "loss": 0.6622,
      "step": 6900
    },
    {
      "epoch": 0.4919098399536199,
      "grad_norm": 0.5781182646751404,
      "learning_rate": 0.00016722502049900433,
      "loss": 0.6448,
      "step": 7000
    },
    {
      "epoch": 0.4919098399536199,
      "eval_loss": 0.6440666317939758,
      "eval_runtime": 118.5556,
      "eval_samples_per_second": 106.693,
      "eval_steps_per_second": 13.344,
      "step": 7000
    },
    {
      "epoch": 0.4989371233815288,
      "grad_norm": 0.7241762280464172,
      "learning_rate": 0.00016675647182851119,
      "loss": 0.6727,
      "step": 7100
    },
    {
      "epoch": 0.5059644068094377,
      "grad_norm": 0.7668181657791138,
      "learning_rate": 0.00016628792315801804,
      "loss": 0.6564,
      "step": 7200
    },
    {
      "epoch": 0.5129916902373465,
      "grad_norm": 0.5727465152740479,
      "learning_rate": 0.00016581937448752489,
      "loss": 0.6348,
      "step": 7300
    },
    {
      "epoch": 0.5200189736652554,
      "grad_norm": 0.6217190027236938,
      "learning_rate": 0.00016535082581703176,
      "loss": 0.6434,
      "step": 7400
    },
    {
      "epoch": 0.5270462570931642,
      "grad_norm": 0.6559625864028931,
      "learning_rate": 0.0001648822771465386,
      "loss": 0.6551,
      "step": 7500
    },
    {
      "epoch": 0.5270462570931642,
      "eval_loss": 0.6428527235984802,
      "eval_runtime": 118.6739,
      "eval_samples_per_second": 106.586,
      "eval_steps_per_second": 13.331,
      "step": 7500
    },
    {
      "epoch": 0.534073540521073,
      "grad_norm": 0.6792352795600891,
      "learning_rate": 0.00016441372847604546,
      "loss": 0.6622,
      "step": 7600
    },
    {
      "epoch": 0.541100823948982,
      "grad_norm": 0.6426942944526672,
      "learning_rate": 0.00016394517980555231,
      "loss": 0.6529,
      "step": 7700
    },
    {
      "epoch": 0.5481281073768908,
      "grad_norm": 0.8337587118148804,
      "learning_rate": 0.00016347663113505916,
      "loss": 0.6506,
      "step": 7800
    },
    {
      "epoch": 0.5551553908047996,
      "grad_norm": 0.5555398464202881,
      "learning_rate": 0.00016300808246456601,
      "loss": 0.6572,
      "step": 7900
    },
    {
      "epoch": 0.5621826742327085,
      "grad_norm": 0.6205873489379883,
      "learning_rate": 0.00016253953379407286,
      "loss": 0.6545,
      "step": 8000
    },
    {
      "epoch": 0.5621826742327085,
      "eval_loss": 0.639687716960907,
      "eval_runtime": 118.9175,
      "eval_samples_per_second": 106.368,
      "eval_steps_per_second": 13.303,
      "step": 8000
    },
    {
      "epoch": 0.5692099576606173,
      "grad_norm": 0.7055862545967102,
      "learning_rate": 0.00016207098512357971,
      "loss": 0.6428,
      "step": 8100
    },
    {
      "epoch": 0.5762372410885263,
      "grad_norm": 0.5852298140525818,
      "learning_rate": 0.00016160243645308656,
      "loss": 0.6404,
      "step": 8200
    },
    {
      "epoch": 0.5832645245164351,
      "grad_norm": 0.551387369632721,
      "learning_rate": 0.00016113388778259341,
      "loss": 0.6618,
      "step": 8300
    },
    {
      "epoch": 0.5902918079443439,
      "grad_norm": 0.7349231839179993,
      "learning_rate": 0.0001606653391121003,
      "loss": 0.6592,
      "step": 8400
    },
    {
      "epoch": 0.5973190913722528,
      "grad_norm": 0.502613365650177,
      "learning_rate": 0.00016019679044160714,
      "loss": 0.6467,
      "step": 8500
    },
    {
      "epoch": 0.5973190913722528,
      "eval_loss": 0.6377580165863037,
      "eval_runtime": 118.7487,
      "eval_samples_per_second": 106.519,
      "eval_steps_per_second": 13.322,
      "step": 8500
    },
    {
      "epoch": 0.6043463748001616,
      "grad_norm": 0.7003266215324402,
      "learning_rate": 0.000159728241771114,
      "loss": 0.6531,
      "step": 8600
    },
    {
      "epoch": 0.6113736582280704,
      "grad_norm": 0.42152824997901917,
      "learning_rate": 0.00015925969310062084,
      "loss": 0.647,
      "step": 8700
    },
    {
      "epoch": 0.6184009416559794,
      "grad_norm": 0.7964949607849121,
      "learning_rate": 0.0001587911444301277,
      "loss": 0.6623,
      "step": 8800
    },
    {
      "epoch": 0.6254282250838882,
      "grad_norm": 0.6723759770393372,
      "learning_rate": 0.00015832259575963452,
      "loss": 0.6528,
      "step": 8900
    },
    {
      "epoch": 0.6324555085117971,
      "grad_norm": 0.6479921936988831,
      "learning_rate": 0.0001578540470891414,
      "loss": 0.6615,
      "step": 9000
    },
    {
      "epoch": 0.6324555085117971,
      "eval_loss": 0.634124755859375,
      "eval_runtime": 118.6171,
      "eval_samples_per_second": 106.637,
      "eval_steps_per_second": 13.337,
      "step": 9000
    },
    {
      "epoch": 0.6394827919397059,
      "grad_norm": 0.6251150965690613,
      "learning_rate": 0.00015738549841864824,
      "loss": 0.6575,
      "step": 9100
    },
    {
      "epoch": 0.6465100753676147,
      "grad_norm": 0.6354021430015564,
      "learning_rate": 0.0001569169497481551,
      "loss": 0.6371,
      "step": 9200
    },
    {
      "epoch": 0.6535373587955237,
      "grad_norm": 0.6997053027153015,
      "learning_rate": 0.00015644840107766194,
      "loss": 0.6724,
      "step": 9300
    },
    {
      "epoch": 0.6605646422234325,
      "grad_norm": 0.6767123341560364,
      "learning_rate": 0.0001559798524071688,
      "loss": 0.6382,
      "step": 9400
    },
    {
      "epoch": 0.6675919256513413,
      "grad_norm": 0.5579701662063599,
      "learning_rate": 0.00015551130373667567,
      "loss": 0.6362,
      "step": 9500
    },
    {
      "epoch": 0.6675919256513413,
      "eval_loss": 0.6337981224060059,
      "eval_runtime": 118.7227,
      "eval_samples_per_second": 106.542,
      "eval_steps_per_second": 13.325,
      "step": 9500
    },
    {
      "epoch": 0.6746192090792502,
      "grad_norm": 0.6185320019721985,
      "learning_rate": 0.00015504275506618252,
      "loss": 0.638,
      "step": 9600
    },
    {
      "epoch": 0.681646492507159,
      "grad_norm": 0.5869000554084778,
      "learning_rate": 0.00015457420639568935,
      "loss": 0.6557,
      "step": 9700
    },
    {
      "epoch": 0.688673775935068,
      "grad_norm": 0.6124538779258728,
      "learning_rate": 0.0001541056577251962,
      "loss": 0.6394,
      "step": 9800
    },
    {
      "epoch": 0.6957010593629768,
      "grad_norm": 1.008245587348938,
      "learning_rate": 0.00015363710905470305,
      "loss": 0.6516,
      "step": 9900
    },
    {
      "epoch": 0.7027283427908856,
      "grad_norm": 0.5377674698829651,
      "learning_rate": 0.00015316856038420992,
      "loss": 0.6089,
      "step": 10000
    },
    {
      "epoch": 0.7027283427908856,
      "eval_loss": 0.6323862671852112,
      "eval_runtime": 118.7128,
      "eval_samples_per_second": 106.551,
      "eval_steps_per_second": 13.326,
      "step": 10000
    },
    {
      "epoch": 0.7097556262187945,
      "grad_norm": 0.6718229651451111,
      "learning_rate": 0.00015270001171371677,
      "loss": 0.6463,
      "step": 10100
    },
    {
      "epoch": 0.7167829096467033,
      "grad_norm": 0.613488495349884,
      "learning_rate": 0.00015223146304322362,
      "loss": 0.6513,
      "step": 10200
    },
    {
      "epoch": 0.7238101930746121,
      "grad_norm": 0.6889612674713135,
      "learning_rate": 0.00015176291437273047,
      "loss": 0.644,
      "step": 10300
    },
    {
      "epoch": 0.7308374765025211,
      "grad_norm": 0.845743715763092,
      "learning_rate": 0.00015129436570223732,
      "loss": 0.6299,
      "step": 10400
    },
    {
      "epoch": 0.7378647599304299,
      "grad_norm": 0.8277881145477295,
      "learning_rate": 0.00015082581703174417,
      "loss": 0.6592,
      "step": 10500
    },
    {
      "epoch": 0.7378647599304299,
      "eval_loss": 0.629069983959198,
      "eval_runtime": 118.7997,
      "eval_samples_per_second": 106.473,
      "eval_steps_per_second": 13.317,
      "step": 10500
    },
    {
      "epoch": 0.7448920433583387,
      "grad_norm": 0.5254293084144592,
      "learning_rate": 0.00015035726836125102,
      "loss": 0.6634,
      "step": 10600
    },
    {
      "epoch": 0.7519193267862476,
      "grad_norm": 0.7291231155395508,
      "learning_rate": 0.00014988871969075787,
      "loss": 0.653,
      "step": 10700
    },
    {
      "epoch": 0.7589466102141564,
      "grad_norm": 0.5473717451095581,
      "learning_rate": 0.00014942017102026472,
      "loss": 0.6477,
      "step": 10800
    },
    {
      "epoch": 0.7659738936420654,
      "grad_norm": 0.717761218547821,
      "learning_rate": 0.00014895162234977158,
      "loss": 0.6481,
      "step": 10900
    },
    {
      "epoch": 0.7730011770699742,
      "grad_norm": 0.5327322483062744,
      "learning_rate": 0.00014848307367927845,
      "loss": 0.6581,
      "step": 11000
    },
    {
      "epoch": 0.7730011770699742,
      "eval_loss": 0.6284623146057129,
      "eval_runtime": 119.1069,
      "eval_samples_per_second": 106.199,
      "eval_steps_per_second": 13.282,
      "step": 11000
    },
    {
      "epoch": 0.780028460497883,
      "grad_norm": 0.5596719980239868,
      "learning_rate": 0.0001480145250087853,
      "loss": 0.6372,
      "step": 11100
    },
    {
      "epoch": 0.7870557439257919,
      "grad_norm": 0.56830894947052,
      "learning_rate": 0.00014754597633829215,
      "loss": 0.6353,
      "step": 11200
    },
    {
      "epoch": 0.7940830273537007,
      "grad_norm": 0.6329615712165833,
      "learning_rate": 0.000147077427667799,
      "loss": 0.6365,
      "step": 11300
    },
    {
      "epoch": 0.8011103107816095,
      "grad_norm": 0.8399169445037842,
      "learning_rate": 0.00014660887899730585,
      "loss": 0.6212,
      "step": 11400
    },
    {
      "epoch": 0.8081375942095185,
      "grad_norm": 0.7220659255981445,
      "learning_rate": 0.0001461403303268127,
      "loss": 0.6438,
      "step": 11500
    },
    {
      "epoch": 0.8081375942095185,
      "eval_loss": 0.6267364621162415,
      "eval_runtime": 118.3852,
      "eval_samples_per_second": 106.846,
      "eval_steps_per_second": 13.363,
      "step": 11500
    },
    {
      "epoch": 0.8151648776374273,
      "grad_norm": 0.5614886283874512,
      "learning_rate": 0.00014567178165631955,
      "loss": 0.6262,
      "step": 11600
    },
    {
      "epoch": 0.8221921610653362,
      "grad_norm": 0.7679696679115295,
      "learning_rate": 0.0001452032329858264,
      "loss": 0.6473,
      "step": 11700
    },
    {
      "epoch": 0.829219444493245,
      "grad_norm": 0.6253560185432434,
      "learning_rate": 0.00014473468431533325,
      "loss": 0.6483,
      "step": 11800
    },
    {
      "epoch": 0.8362467279211538,
      "grad_norm": 0.5834682583808899,
      "learning_rate": 0.0001442661356448401,
      "loss": 0.6411,
      "step": 11900
    },
    {
      "epoch": 0.8432740113490628,
      "grad_norm": 0.4512103497982025,
      "learning_rate": 0.00014379758697434698,
      "loss": 0.6321,
      "step": 12000
    },
    {
      "epoch": 0.8432740113490628,
      "eval_loss": 0.6246777772903442,
      "eval_runtime": 118.7689,
      "eval_samples_per_second": 106.501,
      "eval_steps_per_second": 13.32,
      "step": 12000
    },
    {
      "epoch": 0.8503012947769716,
      "grad_norm": 0.6617989540100098,
      "learning_rate": 0.00014332903830385383,
      "loss": 0.6485,
      "step": 12100
    },
    {
      "epoch": 0.8573285782048804,
      "grad_norm": 0.6235445737838745,
      "learning_rate": 0.00014286048963336068,
      "loss": 0.6254,
      "step": 12200
    },
    {
      "epoch": 0.8643558616327893,
      "grad_norm": 0.612450122833252,
      "learning_rate": 0.00014239194096286753,
      "loss": 0.649,
      "step": 12300
    },
    {
      "epoch": 0.8713831450606981,
      "grad_norm": 0.7379807829856873,
      "learning_rate": 0.00014192339229237436,
      "loss": 0.6284,
      "step": 12400
    },
    {
      "epoch": 0.8784104284886071,
      "grad_norm": 0.8035106658935547,
      "learning_rate": 0.00014145484362188123,
      "loss": 0.6466,
      "step": 12500
    },
    {
      "epoch": 0.8784104284886071,
      "eval_loss": 0.6244432330131531,
      "eval_runtime": 118.6995,
      "eval_samples_per_second": 106.563,
      "eval_steps_per_second": 13.328,
      "step": 12500
    },
    {
      "epoch": 0.8854377119165159,
      "grad_norm": 0.6433550715446472,
      "learning_rate": 0.00014098629495138808,
      "loss": 0.6324,
      "step": 12600
    },
    {
      "epoch": 0.8924649953444247,
      "grad_norm": 0.7102698087692261,
      "learning_rate": 0.00014051774628089493,
      "loss": 0.6536,
      "step": 12700
    },
    {
      "epoch": 0.8994922787723336,
      "grad_norm": 0.7628334164619446,
      "learning_rate": 0.00014004919761040178,
      "loss": 0.6434,
      "step": 12800
    },
    {
      "epoch": 0.9065195622002424,
      "grad_norm": 0.5142523050308228,
      "learning_rate": 0.00013958064893990863,
      "loss": 0.6406,
      "step": 12900
    },
    {
      "epoch": 0.9135468456281512,
      "grad_norm": 0.8540221452713013,
      "learning_rate": 0.0001391121002694155,
      "loss": 0.6317,
      "step": 13000
    },
    {
      "epoch": 0.9135468456281512,
      "eval_loss": 0.6225576996803284,
      "eval_runtime": 118.6623,
      "eval_samples_per_second": 106.597,
      "eval_steps_per_second": 13.332,
      "step": 13000
    },
    {
      "epoch": 0.9205741290560602,
      "grad_norm": 0.6770111918449402,
      "learning_rate": 0.00013864355159892236,
      "loss": 0.6381,
      "step": 13100
    },
    {
      "epoch": 0.927601412483969,
      "grad_norm": 0.7313960194587708,
      "learning_rate": 0.00013817500292842918,
      "loss": 0.643,
      "step": 13200
    },
    {
      "epoch": 0.9346286959118779,
      "grad_norm": 0.8158569931983948,
      "learning_rate": 0.00013770645425793603,
      "loss": 0.6381,
      "step": 13300
    },
    {
      "epoch": 0.9416559793397867,
      "grad_norm": 0.6563596725463867,
      "learning_rate": 0.00013723790558744289,
      "loss": 0.6445,
      "step": 13400
    },
    {
      "epoch": 0.9486832627676955,
      "grad_norm": 0.6007642149925232,
      "learning_rate": 0.00013676935691694976,
      "loss": 0.6165,
      "step": 13500
    },
    {
      "epoch": 0.9486832627676955,
      "eval_loss": 0.6213079690933228,
      "eval_runtime": 118.8856,
      "eval_samples_per_second": 106.396,
      "eval_steps_per_second": 13.307,
      "step": 13500
    },
    {
      "epoch": 0.9557105461956045,
      "grad_norm": 0.7704166173934937,
      "learning_rate": 0.0001363008082464566,
      "loss": 0.6475,
      "step": 13600
    },
    {
      "epoch": 0.9627378296235133,
      "grad_norm": 0.6467058658599854,
      "learning_rate": 0.00013583225957596346,
      "loss": 0.626,
      "step": 13700
    },
    {
      "epoch": 0.9697651130514221,
      "grad_norm": 0.5320102572441101,
      "learning_rate": 0.0001353637109054703,
      "loss": 0.6283,
      "step": 13800
    },
    {
      "epoch": 0.976792396479331,
      "grad_norm": 0.6444761157035828,
      "learning_rate": 0.00013489516223497716,
      "loss": 0.6381,
      "step": 13900
    },
    {
      "epoch": 0.9838196799072398,
      "grad_norm": 0.7598044872283936,
      "learning_rate": 0.00013442661356448404,
      "loss": 0.6305,
      "step": 14000
    },
    {
      "epoch": 0.9838196799072398,
      "eval_loss": 0.6203290820121765,
      "eval_runtime": 118.8928,
      "eval_samples_per_second": 106.39,
      "eval_steps_per_second": 13.306,
      "step": 14000
    },
    {
      "epoch": 0.9908469633351488,
      "grad_norm": 0.735137403011322,
      "learning_rate": 0.00013395806489399086,
      "loss": 0.6378,
      "step": 14100
    },
    {
      "epoch": 0.9978742467630576,
      "grad_norm": 0.758840799331665,
      "learning_rate": 0.00013348951622349771,
      "loss": 0.6469,
      "step": 14200
    },
    {
      "epoch": 1.0049015301909665,
      "grad_norm": 0.5422857403755188,
      "learning_rate": 0.00013302096755300456,
      "loss": 0.586,
      "step": 14300
    },
    {
      "epoch": 1.0119288136188753,
      "grad_norm": 0.5595451593399048,
      "learning_rate": 0.00013255241888251141,
      "loss": 0.5967,
      "step": 14400
    },
    {
      "epoch": 1.0189560970467841,
      "grad_norm": 0.6194477081298828,
      "learning_rate": 0.0001320838702120183,
      "loss": 0.602,
      "step": 14500
    },
    {
      "epoch": 1.0189560970467841,
      "eval_loss": 0.6205016374588013,
      "eval_runtime": 118.9575,
      "eval_samples_per_second": 106.332,
      "eval_steps_per_second": 13.299,
      "step": 14500
    },
    {
      "epoch": 1.025983380474693,
      "grad_norm": 0.710670530796051,
      "learning_rate": 0.00013161532154152514,
      "loss": 0.5954,
      "step": 14600
    },
    {
      "epoch": 1.0330106639026018,
      "grad_norm": 0.613923192024231,
      "learning_rate": 0.000131146772871032,
      "loss": 0.5855,
      "step": 14700
    },
    {
      "epoch": 1.0400379473305108,
      "grad_norm": 0.7411431670188904,
      "learning_rate": 0.00013067822420053884,
      "loss": 0.5881,
      "step": 14800
    },
    {
      "epoch": 1.0470652307584196,
      "grad_norm": 0.7143989205360413,
      "learning_rate": 0.0001302096755300457,
      "loss": 0.5891,
      "step": 14900
    },
    {
      "epoch": 1.0540925141863284,
      "grad_norm": 0.6597478985786438,
      "learning_rate": 0.00012974112685955254,
      "loss": 0.5915,
      "step": 15000
    },
    {
      "epoch": 1.0540925141863284,
      "eval_loss": 0.6205988526344299,
      "eval_runtime": 118.9337,
      "eval_samples_per_second": 106.353,
      "eval_steps_per_second": 13.302,
      "step": 15000
    },
    {
      "epoch": 1.0611197976142372,
      "grad_norm": 0.5813501477241516,
      "learning_rate": 0.0001292725781890594,
      "loss": 0.5877,
      "step": 15100
    },
    {
      "epoch": 1.068147081042146,
      "grad_norm": 0.5610823631286621,
      "learning_rate": 0.00012880402951856624,
      "loss": 0.5911,
      "step": 15200
    },
    {
      "epoch": 1.075174364470055,
      "grad_norm": 0.5839988589286804,
      "learning_rate": 0.0001283354808480731,
      "loss": 0.5903,
      "step": 15300
    },
    {
      "epoch": 1.082201647897964,
      "grad_norm": 0.6421522498130798,
      "learning_rate": 0.00012786693217757994,
      "loss": 0.591,
      "step": 15400
    },
    {
      "epoch": 1.0892289313258727,
      "grad_norm": 0.6854695081710815,
      "learning_rate": 0.0001273983835070868,
      "loss": 0.5998,
      "step": 15500
    },
    {
      "epoch": 1.0892289313258727,
      "eval_loss": 0.6208451986312866,
      "eval_runtime": 118.6829,
      "eval_samples_per_second": 106.578,
      "eval_steps_per_second": 13.33,
      "step": 15500
    },
    {
      "epoch": 1.0962562147537815,
      "grad_norm": 0.7333750128746033,
      "learning_rate": 0.00012692983483659367,
      "loss": 0.5916,
      "step": 15600
    },
    {
      "epoch": 1.1032834981816904,
      "grad_norm": 0.5983602404594421,
      "learning_rate": 0.00012646128616610052,
      "loss": 0.581,
      "step": 15700
    },
    {
      "epoch": 1.1103107816095992,
      "grad_norm": 0.5112642645835876,
      "learning_rate": 0.00012599273749560737,
      "loss": 0.5984,
      "step": 15800
    },
    {
      "epoch": 1.1173380650375082,
      "grad_norm": 0.5569522976875305,
      "learning_rate": 0.0001255241888251142,
      "loss": 0.5983,
      "step": 15900
    },
    {
      "epoch": 1.124365348465417,
      "grad_norm": 0.6089062690734863,
      "learning_rate": 0.00012505564015462105,
      "loss": 0.6019,
      "step": 16000
    },
    {
      "epoch": 1.124365348465417,
      "eval_loss": 0.6191478967666626,
      "eval_runtime": 118.6608,
      "eval_samples_per_second": 106.598,
      "eval_steps_per_second": 13.332,
      "step": 16000
    },
    {
      "epoch": 1.1313926318933258,
      "grad_norm": 0.5710394978523254,
      "learning_rate": 0.00012458709148412792,
      "loss": 0.5907,
      "step": 16100
    },
    {
      "epoch": 1.1384199153212347,
      "grad_norm": 0.7836496233940125,
      "learning_rate": 0.00012411854281363477,
      "loss": 0.5818,
      "step": 16200
    },
    {
      "epoch": 1.1454471987491435,
      "grad_norm": 0.600236177444458,
      "learning_rate": 0.00012364999414314162,
      "loss": 0.5834,
      "step": 16300
    },
    {
      "epoch": 1.1524744821770523,
      "grad_norm": 0.7090241312980652,
      "learning_rate": 0.00012318144547264847,
      "loss": 0.5963,
      "step": 16400
    },
    {
      "epoch": 1.1595017656049613,
      "grad_norm": 0.5439143180847168,
      "learning_rate": 0.00012271289680215532,
      "loss": 0.5939,
      "step": 16500
    },
    {
      "epoch": 1.1595017656049613,
      "eval_loss": 0.6187065243721008,
      "eval_runtime": 118.7823,
      "eval_samples_per_second": 106.489,
      "eval_steps_per_second": 13.318,
      "step": 16500
    },
    {
      "epoch": 1.1665290490328701,
      "grad_norm": 0.6133089661598206,
      "learning_rate": 0.0001222443481316622,
      "loss": 0.583,
      "step": 16600
    },
    {
      "epoch": 1.173556332460779,
      "grad_norm": 0.6947652101516724,
      "learning_rate": 0.00012177579946116904,
      "loss": 0.5851,
      "step": 16700
    },
    {
      "epoch": 1.1805836158886878,
      "grad_norm": 0.799213707447052,
      "learning_rate": 0.00012130725079067589,
      "loss": 0.5815,
      "step": 16800
    },
    {
      "epoch": 1.1876108993165966,
      "grad_norm": 0.6443912982940674,
      "learning_rate": 0.00012083870212018274,
      "loss": 0.5794,
      "step": 16900
    },
    {
      "epoch": 1.1946381827445056,
      "grad_norm": 0.8435219526290894,
      "learning_rate": 0.00012037015344968959,
      "loss": 0.5747,
      "step": 17000
    },
    {
      "epoch": 1.1946381827445056,
      "eval_loss": 0.6187562942504883,
      "eval_runtime": 118.6879,
      "eval_samples_per_second": 106.574,
      "eval_steps_per_second": 13.329,
      "step": 17000
    },
    {
      "epoch": 1.2016654661724144,
      "grad_norm": 0.7162328958511353,
      "learning_rate": 0.00011990160477919645,
      "loss": 0.596,
      "step": 17100
    },
    {
      "epoch": 1.2086927496003232,
      "grad_norm": 0.6594322919845581,
      "learning_rate": 0.0001194330561087033,
      "loss": 0.6054,
      "step": 17200
    },
    {
      "epoch": 1.215720033028232,
      "grad_norm": 0.5395209193229675,
      "learning_rate": 0.00011896450743821015,
      "loss": 0.5835,
      "step": 17300
    },
    {
      "epoch": 1.2227473164561409,
      "grad_norm": 0.7208767533302307,
      "learning_rate": 0.000118495958767717,
      "loss": 0.595,
      "step": 17400
    },
    {
      "epoch": 1.22977459988405,
      "grad_norm": 0.6752803921699524,
      "learning_rate": 0.00011802741009722384,
      "loss": 0.5937,
      "step": 17500
    },
    {
      "epoch": 1.22977459988405,
      "eval_loss": 0.6167559623718262,
      "eval_runtime": 118.8202,
      "eval_samples_per_second": 106.455,
      "eval_steps_per_second": 13.314,
      "step": 17500
    },
    {
      "epoch": 1.2368018833119587,
      "grad_norm": 0.7853017449378967,
      "learning_rate": 0.00011755886142673072,
      "loss": 0.6035,
      "step": 17600
    },
    {
      "epoch": 1.2438291667398675,
      "grad_norm": 0.6420643329620361,
      "learning_rate": 0.00011709031275623757,
      "loss": 0.5834,
      "step": 17700
    },
    {
      "epoch": 1.2508564501677764,
      "grad_norm": 0.8449912667274475,
      "learning_rate": 0.00011662176408574442,
      "loss": 0.5964,
      "step": 17800
    },
    {
      "epoch": 1.2578837335956852,
      "grad_norm": 0.6199436783790588,
      "learning_rate": 0.00011615321541525125,
      "loss": 0.5771,
      "step": 17900
    },
    {
      "epoch": 1.2649110170235942,
      "grad_norm": 1.0329114198684692,
      "learning_rate": 0.0001156846667447581,
      "loss": 0.5678,
      "step": 18000
    },
    {
      "epoch": 1.2649110170235942,
      "eval_loss": 0.6171479821205139,
      "eval_runtime": 118.9379,
      "eval_samples_per_second": 106.35,
      "eval_steps_per_second": 13.301,
      "step": 18000
    },
    {
      "epoch": 1.271938300451503,
      "grad_norm": 0.7668006420135498,
      "learning_rate": 0.00011521611807426498,
      "loss": 0.5895,
      "step": 18100
    },
    {
      "epoch": 1.2789655838794118,
      "grad_norm": 0.661259651184082,
      "learning_rate": 0.00011474756940377183,
      "loss": 0.5951,
      "step": 18200
    },
    {
      "epoch": 1.2859928673073207,
      "grad_norm": 0.5448057055473328,
      "learning_rate": 0.00011427902073327867,
      "loss": 0.5958,
      "step": 18300
    },
    {
      "epoch": 1.2930201507352295,
      "grad_norm": 0.5419151782989502,
      "learning_rate": 0.00011381047206278552,
      "loss": 0.5788,
      "step": 18400
    },
    {
      "epoch": 1.3000474341631385,
      "grad_norm": 0.595245361328125,
      "learning_rate": 0.00011334192339229237,
      "loss": 0.5849,
      "step": 18500
    },
    {
      "epoch": 1.3000474341631385,
      "eval_loss": 0.6158913969993591,
      "eval_runtime": 118.7534,
      "eval_samples_per_second": 106.515,
      "eval_steps_per_second": 13.322,
      "step": 18500
    },
    {
      "epoch": 1.3070747175910473,
      "grad_norm": 0.6495450139045715,
      "learning_rate": 0.00011287337472179925,
      "loss": 0.5849,
      "step": 18600
    },
    {
      "epoch": 1.3141020010189561,
      "grad_norm": 0.686590313911438,
      "learning_rate": 0.00011240482605130608,
      "loss": 0.5679,
      "step": 18700
    },
    {
      "epoch": 1.321129284446865,
      "grad_norm": 0.60063636302948,
      "learning_rate": 0.00011193627738081293,
      "loss": 0.5805,
      "step": 18800
    },
    {
      "epoch": 1.3281565678747738,
      "grad_norm": 0.6396400332450867,
      "learning_rate": 0.00011146772871031978,
      "loss": 0.5996,
      "step": 18900
    },
    {
      "epoch": 1.3351838513026828,
      "grad_norm": 0.6631867289543152,
      "learning_rate": 0.00011099918003982663,
      "loss": 0.5926,
      "step": 19000
    },
    {
      "epoch": 1.3351838513026828,
      "eval_loss": 0.6149775981903076,
      "eval_runtime": 118.2284,
      "eval_samples_per_second": 106.988,
      "eval_steps_per_second": 13.381,
      "step": 19000
    },
    {
      "epoch": 1.3422111347305914,
      "grad_norm": 0.6868234276771545,
      "learning_rate": 0.0001105306313693335,
      "loss": 0.5826,
      "step": 19100
    },
    {
      "epoch": 1.3492384181585004,
      "grad_norm": 0.6282151937484741,
      "learning_rate": 0.00011006208269884035,
      "loss": 0.5927,
      "step": 19200
    },
    {
      "epoch": 1.3562657015864092,
      "grad_norm": 0.6498789191246033,
      "learning_rate": 0.0001095935340283472,
      "loss": 0.598,
      "step": 19300
    },
    {
      "epoch": 1.363292985014318,
      "grad_norm": 0.7654560804367065,
      "learning_rate": 0.00010912498535785405,
      "loss": 0.5819,
      "step": 19400
    },
    {
      "epoch": 1.3703202684422269,
      "grad_norm": 0.6079320907592773,
      "learning_rate": 0.0001086564366873609,
      "loss": 0.5941,
      "step": 19500
    },
    {
      "epoch": 1.3703202684422269,
      "eval_loss": 0.6134491562843323,
      "eval_runtime": 118.1186,
      "eval_samples_per_second": 107.087,
      "eval_steps_per_second": 13.393,
      "step": 19500
    },
    {
      "epoch": 1.3773475518701357,
      "grad_norm": 0.7858242392539978,
      "learning_rate": 0.00010818788801686776,
      "loss": 0.6035,
      "step": 19600
    },
    {
      "epoch": 1.3843748352980447,
      "grad_norm": 0.46175357699394226,
      "learning_rate": 0.00010771933934637461,
      "loss": 0.6054,
      "step": 19700
    },
    {
      "epoch": 1.3914021187259535,
      "grad_norm": 0.817308783531189,
      "learning_rate": 0.00010725079067588146,
      "loss": 0.5844,
      "step": 19800
    },
    {
      "epoch": 1.3984294021538624,
      "grad_norm": 0.7891727685928345,
      "learning_rate": 0.00010678224200538831,
      "loss": 0.5885,
      "step": 19900
    },
    {
      "epoch": 1.4054566855817712,
      "grad_norm": 0.7089536786079407,
      "learning_rate": 0.00010631369333489516,
      "loss": 0.6108,
      "step": 20000
    },
    {
      "epoch": 1.4054566855817712,
      "eval_loss": 0.6131945252418518,
      "eval_runtime": 119.3154,
      "eval_samples_per_second": 106.013,
      "eval_steps_per_second": 13.259,
      "step": 20000
    },
    {
      "epoch": 1.41248396900968,
      "grad_norm": 0.6867943406105042,
      "learning_rate": 0.00010584514466440203,
      "loss": 0.5863,
      "step": 20100
    },
    {
      "epoch": 1.419511252437589,
      "grad_norm": 0.9261388182640076,
      "learning_rate": 0.00010537659599390888,
      "loss": 0.6027,
      "step": 20200
    },
    {
      "epoch": 1.4265385358654978,
      "grad_norm": 0.7854331135749817,
      "learning_rate": 0.00010490804732341573,
      "loss": 0.5637,
      "step": 20300
    },
    {
      "epoch": 1.4335658192934067,
      "grad_norm": 0.4610428214073181,
      "learning_rate": 0.00010443949865292258,
      "loss": 0.5961,
      "step": 20400
    },
    {
      "epoch": 1.4405931027213155,
      "grad_norm": 0.651196300983429,
      "learning_rate": 0.00010397094998242943,
      "loss": 0.5829,
      "step": 20500
    },
    {
      "epoch": 1.4405931027213155,
      "eval_loss": 0.6108871698379517,
      "eval_runtime": 119.1208,
      "eval_samples_per_second": 106.186,
      "eval_steps_per_second": 13.281,
      "step": 20500
    },
    {
      "epoch": 1.4476203861492243,
      "grad_norm": 0.7416488528251648,
      "learning_rate": 0.00010350240131193629,
      "loss": 0.5869,
      "step": 20600
    },
    {
      "epoch": 1.4546476695771333,
      "grad_norm": 0.6142196655273438,
      "learning_rate": 0.00010303385264144314,
      "loss": 0.5858,
      "step": 20700
    },
    {
      "epoch": 1.4616749530050421,
      "grad_norm": 0.644241213798523,
      "learning_rate": 0.00010256530397094999,
      "loss": 0.5861,
      "step": 20800
    },
    {
      "epoch": 1.468702236432951,
      "grad_norm": 0.8656560182571411,
      "learning_rate": 0.00010209675530045684,
      "loss": 0.6099,
      "step": 20900
    },
    {
      "epoch": 1.4757295198608598,
      "grad_norm": 0.5667104721069336,
      "learning_rate": 0.00010162820662996368,
      "loss": 0.601,
      "step": 21000
    },
    {
      "epoch": 1.4757295198608598,
      "eval_loss": 0.6112544536590576,
      "eval_runtime": 119.0171,
      "eval_samples_per_second": 106.279,
      "eval_steps_per_second": 13.292,
      "step": 21000
    },
    {
      "epoch": 1.4827568032887686,
      "grad_norm": 0.6884378790855408,
      "learning_rate": 0.00010115965795947056,
      "loss": 0.5882,
      "step": 21100
    },
    {
      "epoch": 1.4897840867166776,
      "grad_norm": 0.7136459946632385,
      "learning_rate": 0.0001006911092889774,
      "loss": 0.5905,
      "step": 21200
    },
    {
      "epoch": 1.4968113701445864,
      "grad_norm": 0.8048639297485352,
      "learning_rate": 0.00010022256061848426,
      "loss": 0.5875,
      "step": 21300
    },
    {
      "epoch": 1.5038386535724952,
      "grad_norm": 0.7300230860710144,
      "learning_rate": 9.975401194799109e-05,
      "loss": 0.5916,
      "step": 21400
    },
    {
      "epoch": 1.510865937000404,
      "grad_norm": 0.7058496475219727,
      "learning_rate": 9.928546327749796e-05,
      "loss": 0.5768,
      "step": 21500
    },
    {
      "epoch": 1.510865937000404,
      "eval_loss": 0.6108001470565796,
      "eval_runtime": 118.8504,
      "eval_samples_per_second": 106.428,
      "eval_steps_per_second": 13.311,
      "step": 21500
    },
    {
      "epoch": 1.5178932204283129,
      "grad_norm": 0.8157733678817749,
      "learning_rate": 9.881691460700481e-05,
      "loss": 0.5872,
      "step": 21600
    },
    {
      "epoch": 1.524920503856222,
      "grad_norm": 0.8618035316467285,
      "learning_rate": 9.834836593651166e-05,
      "loss": 0.5861,
      "step": 21700
    },
    {
      "epoch": 1.5319477872841305,
      "grad_norm": 0.7457069158554077,
      "learning_rate": 9.787981726601851e-05,
      "loss": 0.5948,
      "step": 21800
    },
    {
      "epoch": 1.5389750707120395,
      "grad_norm": 0.7307142615318298,
      "learning_rate": 9.741126859552536e-05,
      "loss": 0.6068,
      "step": 21900
    },
    {
      "epoch": 1.5460023541399484,
      "grad_norm": 0.7692698836326599,
      "learning_rate": 9.694271992503222e-05,
      "loss": 0.5732,
      "step": 22000
    },
    {
      "epoch": 1.5460023541399484,
      "eval_loss": 0.6103039979934692,
      "eval_runtime": 118.5221,
      "eval_samples_per_second": 106.723,
      "eval_steps_per_second": 13.348,
      "step": 22000
    },
    {
      "epoch": 1.5530296375678572,
      "grad_norm": 0.7781071662902832,
      "learning_rate": 9.647417125453907e-05,
      "loss": 0.5915,
      "step": 22100
    },
    {
      "epoch": 1.5600569209957662,
      "grad_norm": 0.7720737457275391,
      "learning_rate": 9.600562258404592e-05,
      "loss": 0.6007,
      "step": 22200
    },
    {
      "epoch": 1.5670842044236748,
      "grad_norm": 0.632757306098938,
      "learning_rate": 9.553707391355277e-05,
      "loss": 0.5838,
      "step": 22300
    },
    {
      "epoch": 1.5741114878515838,
      "grad_norm": 0.7514855265617371,
      "learning_rate": 9.506852524305962e-05,
      "loss": 0.5894,
      "step": 22400
    },
    {
      "epoch": 1.5811387712794926,
      "grad_norm": 0.5634511113166809,
      "learning_rate": 9.459997657256649e-05,
      "loss": 0.5774,
      "step": 22500
    },
    {
      "epoch": 1.5811387712794926,
      "eval_loss": 0.6085862517356873,
      "eval_runtime": 119.1435,
      "eval_samples_per_second": 106.166,
      "eval_steps_per_second": 13.278,
      "step": 22500
    },
    {
      "epoch": 1.5881660547074015,
      "grad_norm": 0.8155964612960815,
      "learning_rate": 9.413142790207334e-05,
      "loss": 0.582,
      "step": 22600
    },
    {
      "epoch": 1.5951933381353105,
      "grad_norm": 0.7442721128463745,
      "learning_rate": 9.366287923158019e-05,
      "loss": 0.595,
      "step": 22700
    },
    {
      "epoch": 1.602220621563219,
      "grad_norm": 0.6977400183677673,
      "learning_rate": 9.319433056108704e-05,
      "loss": 0.5906,
      "step": 22800
    },
    {
      "epoch": 1.6092479049911281,
      "grad_norm": 0.7066090703010559,
      "learning_rate": 9.272578189059389e-05,
      "loss": 0.5917,
      "step": 22900
    },
    {
      "epoch": 1.616275188419037,
      "grad_norm": 0.9004433751106262,
      "learning_rate": 9.225723322010075e-05,
      "loss": 0.5821,
      "step": 23000
    },
    {
      "epoch": 1.616275188419037,
      "eval_loss": 0.608232319355011,
      "eval_runtime": 118.5994,
      "eval_samples_per_second": 106.653,
      "eval_steps_per_second": 13.339,
      "step": 23000
    },
    {
      "epoch": 1.6233024718469458,
      "grad_norm": 0.6980244517326355,
      "learning_rate": 9.17886845496076e-05,
      "loss": 0.5986,
      "step": 23100
    },
    {
      "epoch": 1.6303297552748546,
      "grad_norm": 0.6959982514381409,
      "learning_rate": 9.132013587911444e-05,
      "loss": 0.5751,
      "step": 23200
    },
    {
      "epoch": 1.6373570387027634,
      "grad_norm": 0.6452066898345947,
      "learning_rate": 9.08515872086213e-05,
      "loss": 0.5834,
      "step": 23300
    },
    {
      "epoch": 1.6443843221306724,
      "grad_norm": 0.563113272190094,
      "learning_rate": 9.038303853812815e-05,
      "loss": 0.5943,
      "step": 23400
    },
    {
      "epoch": 1.6514116055585812,
      "grad_norm": 0.6849614977836609,
      "learning_rate": 8.991448986763502e-05,
      "loss": 0.581,
      "step": 23500
    },
    {
      "epoch": 1.6514116055585812,
      "eval_loss": 0.607568621635437,
      "eval_runtime": 119.0065,
      "eval_samples_per_second": 106.288,
      "eval_steps_per_second": 13.293,
      "step": 23500
    },
    {
      "epoch": 1.65843888898649,
      "grad_norm": 0.8570700287818909,
      "learning_rate": 8.944594119714185e-05,
      "loss": 0.592,
      "step": 23600
    },
    {
      "epoch": 1.6654661724143989,
      "grad_norm": 0.4864564538002014,
      "learning_rate": 8.89773925266487e-05,
      "loss": 0.5722,
      "step": 23700
    },
    {
      "epoch": 1.6724934558423077,
      "grad_norm": 0.6979348063468933,
      "learning_rate": 8.850884385615557e-05,
      "loss": 0.578,
      "step": 23800
    },
    {
      "epoch": 1.6795207392702167,
      "grad_norm": 0.7699964046478271,
      "learning_rate": 8.804029518566242e-05,
      "loss": 0.5897,
      "step": 23900
    },
    {
      "epoch": 1.6865480226981253,
      "grad_norm": 0.8256754875183105,
      "learning_rate": 8.757174651516927e-05,
      "loss": 0.5845,
      "step": 24000
    },
    {
      "epoch": 1.6865480226981253,
      "eval_loss": 0.6071833372116089,
      "eval_runtime": 119.1197,
      "eval_samples_per_second": 106.187,
      "eval_steps_per_second": 13.281,
      "step": 24000
    },
    {
      "epoch": 1.6935753061260344,
      "grad_norm": 0.6591055393218994,
      "learning_rate": 8.710319784467612e-05,
      "loss": 0.5867,
      "step": 24100
    },
    {
      "epoch": 1.7006025895539432,
      "grad_norm": 0.7720032334327698,
      "learning_rate": 8.663464917418297e-05,
      "loss": 0.5905,
      "step": 24200
    },
    {
      "epoch": 1.707629872981852,
      "grad_norm": 0.5807234048843384,
      "learning_rate": 8.616610050368983e-05,
      "loss": 0.5849,
      "step": 24300
    },
    {
      "epoch": 1.714657156409761,
      "grad_norm": 0.6583465337753296,
      "learning_rate": 8.569755183319668e-05,
      "loss": 0.5706,
      "step": 24400
    },
    {
      "epoch": 1.7216844398376696,
      "grad_norm": 0.7280032634735107,
      "learning_rate": 8.522900316270352e-05,
      "loss": 0.5843,
      "step": 24500
    },
    {
      "epoch": 1.7216844398376696,
      "eval_loss": 0.6051200032234192,
      "eval_runtime": 119.1844,
      "eval_samples_per_second": 106.13,
      "eval_steps_per_second": 13.274,
      "step": 24500
    },
    {
      "epoch": 1.7287117232655786,
      "grad_norm": 0.7264565825462341,
      "learning_rate": 8.476045449221038e-05,
      "loss": 0.5866,
      "step": 24600
    },
    {
      "epoch": 1.7357390066934875,
      "grad_norm": 0.7795102596282959,
      "learning_rate": 8.429190582171723e-05,
      "loss": 0.5807,
      "step": 24700
    },
    {
      "epoch": 1.7427662901213963,
      "grad_norm": 0.702314019203186,
      "learning_rate": 8.38233571512241e-05,
      "loss": 0.5856,
      "step": 24800
    },
    {
      "epoch": 1.7497935735493053,
      "grad_norm": 0.7014954090118408,
      "learning_rate": 8.335480848073095e-05,
      "loss": 0.5812,
      "step": 24900
    },
    {
      "epoch": 1.756820856977214,
      "grad_norm": 0.6724287867546082,
      "learning_rate": 8.288625981023778e-05,
      "loss": 0.5803,
      "step": 25000
    },
    {
      "epoch": 1.756820856977214,
      "eval_loss": 0.6055319309234619,
      "eval_runtime": 119.2213,
      "eval_samples_per_second": 106.097,
      "eval_steps_per_second": 13.269,
      "step": 25000
    },
    {
      "epoch": 1.763848140405123,
      "grad_norm": 0.9690128564834595,
      "learning_rate": 8.241771113974465e-05,
      "loss": 0.5674,
      "step": 25100
    },
    {
      "epoch": 1.7708754238330318,
      "grad_norm": 1.0121440887451172,
      "learning_rate": 8.19491624692515e-05,
      "loss": 0.5913,
      "step": 25200
    },
    {
      "epoch": 1.7779027072609406,
      "grad_norm": 0.7207921743392944,
      "learning_rate": 8.148061379875836e-05,
      "loss": 0.592,
      "step": 25300
    },
    {
      "epoch": 1.7849299906888496,
      "grad_norm": 0.6326346397399902,
      "learning_rate": 8.10120651282652e-05,
      "loss": 0.592,
      "step": 25400
    },
    {
      "epoch": 1.7919572741167582,
      "grad_norm": 0.7215606570243835,
      "learning_rate": 8.054351645777205e-05,
      "loss": 0.578,
      "step": 25500
    },
    {
      "epoch": 1.7919572741167582,
      "eval_loss": 0.6045902371406555,
      "eval_runtime": 118.5573,
      "eval_samples_per_second": 106.691,
      "eval_steps_per_second": 13.344,
      "step": 25500
    },
    {
      "epoch": 1.7989845575446672,
      "grad_norm": 0.6932191848754883,
      "learning_rate": 8.007496778727891e-05,
      "loss": 0.5881,
      "step": 25600
    },
    {
      "epoch": 1.806011840972576,
      "grad_norm": 0.7887512445449829,
      "learning_rate": 7.960641911678576e-05,
      "loss": 0.5875,
      "step": 25700
    },
    {
      "epoch": 1.8130391244004849,
      "grad_norm": 0.5214329957962036,
      "learning_rate": 7.913787044629261e-05,
      "loss": 0.5757,
      "step": 25800
    },
    {
      "epoch": 1.8200664078283937,
      "grad_norm": 0.6298120021820068,
      "learning_rate": 7.866932177579946e-05,
      "loss": 0.5634,
      "step": 25900
    },
    {
      "epoch": 1.8270936912563025,
      "grad_norm": 0.6642977595329285,
      "learning_rate": 7.820077310530631e-05,
      "loss": 0.5708,
      "step": 26000
    },
    {
      "epoch": 1.8270936912563025,
      "eval_loss": 0.6030368804931641,
      "eval_runtime": 118.8611,
      "eval_samples_per_second": 106.418,
      "eval_steps_per_second": 13.31,
      "step": 26000
    },
    {
      "epoch": 1.8341209746842115,
      "grad_norm": 0.5501639246940613,
      "learning_rate": 7.773222443481318e-05,
      "loss": 0.5692,
      "step": 26100
    },
    {
      "epoch": 1.8411482581121204,
      "grad_norm": 0.6441388726234436,
      "learning_rate": 7.726367576432003e-05,
      "loss": 0.5881,
      "step": 26200
    },
    {
      "epoch": 1.8481755415400292,
      "grad_norm": 0.6169604063034058,
      "learning_rate": 7.679512709382688e-05,
      "loss": 0.5762,
      "step": 26300
    },
    {
      "epoch": 1.855202824967938,
      "grad_norm": 0.6980007290840149,
      "learning_rate": 7.632657842333373e-05,
      "loss": 0.5739,
      "step": 26400
    },
    {
      "epoch": 1.8622301083958468,
      "grad_norm": 0.7939792275428772,
      "learning_rate": 7.585802975284058e-05,
      "loss": 0.5746,
      "step": 26500
    },
    {
      "epoch": 1.8622301083958468,
      "eval_loss": 0.6025614142417908,
      "eval_runtime": 118.9608,
      "eval_samples_per_second": 106.329,
      "eval_steps_per_second": 13.298,
      "step": 26500
    },
    {
      "epoch": 1.8692573918237558,
      "grad_norm": 0.6231071352958679,
      "learning_rate": 7.538948108234744e-05,
      "loss": 0.5634,
      "step": 26600
    },
    {
      "epoch": 1.8762846752516644,
      "grad_norm": 0.6939712762832642,
      "learning_rate": 7.492093241185428e-05,
      "loss": 0.5792,
      "step": 26700
    },
    {
      "epoch": 1.8833119586795735,
      "grad_norm": 0.6055401563644409,
      "learning_rate": 7.445238374136114e-05,
      "loss": 0.5929,
      "step": 26800
    },
    {
      "epoch": 1.8903392421074823,
      "grad_norm": 0.7465933561325073,
      "learning_rate": 7.398383507086799e-05,
      "loss": 0.5817,
      "step": 26900
    },
    {
      "epoch": 1.897366525535391,
      "grad_norm": 0.6460291743278503,
      "learning_rate": 7.351528640037484e-05,
      "loss": 0.5831,
      "step": 27000
    },
    {
      "epoch": 1.897366525535391,
      "eval_loss": 0.6021212935447693,
      "eval_runtime": 118.8793,
      "eval_samples_per_second": 106.402,
      "eval_steps_per_second": 13.308,
      "step": 27000
    },
    {
      "epoch": 1.9043938089633001,
      "grad_norm": 0.5837533473968506,
      "learning_rate": 7.304673772988169e-05,
      "loss": 0.5807,
      "step": 27100
    },
    {
      "epoch": 1.9114210923912087,
      "grad_norm": 0.7371869087219238,
      "learning_rate": 7.257818905938854e-05,
      "loss": 0.5967,
      "step": 27200
    },
    {
      "epoch": 1.9184483758191178,
      "grad_norm": 0.8853654861450195,
      "learning_rate": 7.21096403888954e-05,
      "loss": 0.5967,
      "step": 27300
    },
    {
      "epoch": 1.9254756592470266,
      "grad_norm": 0.7515887022018433,
      "learning_rate": 7.164109171840226e-05,
      "loss": 0.5708,
      "step": 27400
    },
    {
      "epoch": 1.9325029426749354,
      "grad_norm": 0.6723042130470276,
      "learning_rate": 7.11725430479091e-05,
      "loss": 0.5687,
      "step": 27500
    },
    {
      "epoch": 1.9325029426749354,
      "eval_loss": 0.6017782092094421,
      "eval_runtime": 118.9577,
      "eval_samples_per_second": 106.332,
      "eval_steps_per_second": 13.299,
      "step": 27500
    },
    {
      "epoch": 1.9395302261028444,
      "grad_norm": 0.5766080617904663,
      "learning_rate": 7.070399437741596e-05,
      "loss": 0.5876,
      "step": 27600
    },
    {
      "epoch": 1.946557509530753,
      "grad_norm": 0.6501230597496033,
      "learning_rate": 7.02354457069228e-05,
      "loss": 0.5726,
      "step": 27700
    },
    {
      "epoch": 1.953584792958662,
      "grad_norm": 0.7194878458976746,
      "learning_rate": 6.976689703642966e-05,
      "loss": 0.5741,
      "step": 27800
    },
    {
      "epoch": 1.9606120763865709,
      "grad_norm": 0.8090994954109192,
      "learning_rate": 6.929834836593652e-05,
      "loss": 0.5764,
      "step": 27900
    },
    {
      "epoch": 1.9676393598144797,
      "grad_norm": 0.6500638127326965,
      "learning_rate": 6.882979969544337e-05,
      "loss": 0.5752,
      "step": 28000
    },
    {
      "epoch": 1.9676393598144797,
      "eval_loss": 0.6010117530822754,
      "eval_runtime": 119.0388,
      "eval_samples_per_second": 106.259,
      "eval_steps_per_second": 13.29,
      "step": 28000
    }
  ],
  "logging_steps": 100,
  "max_steps": 42690,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 9.547282008374477e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}