{
  "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": 0.45293834805488586,
      "learning_rate": 0.0001995548787630315,
      "loss": 0.8897,
      "step": 100
    },
    {
      "epoch": 0.014054566855817713,
      "grad_norm": 0.5474397540092468,
      "learning_rate": 0.00019908633009253836,
      "loss": 0.7035,
      "step": 200
    },
    {
      "epoch": 0.02108185028372657,
      "grad_norm": 0.38776981830596924,
      "learning_rate": 0.00019861778142204523,
      "loss": 0.6916,
      "step": 300
    },
    {
      "epoch": 0.028109133711635426,
      "grad_norm": 0.39987438917160034,
      "learning_rate": 0.00019814923275155208,
      "loss": 0.6724,
      "step": 400
    },
    {
      "epoch": 0.03513641713954428,
      "grad_norm": 0.4090868830680847,
      "learning_rate": 0.00019768068408105893,
      "loss": 0.6842,
      "step": 500
    },
    {
      "epoch": 0.03513641713954428,
      "eval_loss": 0.6561594605445862,
      "eval_runtime": 192.4196,
      "eval_samples_per_second": 65.737,
      "eval_steps_per_second": 8.222,
      "step": 500
    },
    {
      "epoch": 0.04216370056745314,
      "grad_norm": 0.46690303087234497,
      "learning_rate": 0.00019721213541056578,
      "loss": 0.6738,
      "step": 600
    },
    {
      "epoch": 0.049190983995361995,
      "grad_norm": 0.48969191312789917,
      "learning_rate": 0.00019674358674007263,
      "loss": 0.662,
      "step": 700
    },
    {
      "epoch": 0.05621826742327085,
      "grad_norm": 0.3667367100715637,
      "learning_rate": 0.00019627503806957949,
      "loss": 0.6646,
      "step": 800
    },
    {
      "epoch": 0.06324555085117971,
      "grad_norm": 0.40032121539115906,
      "learning_rate": 0.00019580648939908634,
      "loss": 0.6564,
      "step": 900
    },
    {
      "epoch": 0.07027283427908856,
      "grad_norm": 0.3847985565662384,
      "learning_rate": 0.00019533794072859319,
      "loss": 0.6511,
      "step": 1000
    },
    {
      "epoch": 0.07027283427908856,
      "eval_loss": 0.6426891684532166,
      "eval_runtime": 192.091,
      "eval_samples_per_second": 65.849,
      "eval_steps_per_second": 8.236,
      "step": 1000
    },
    {
      "epoch": 0.07730011770699742,
      "grad_norm": 0.40991073846817017,
      "learning_rate": 0.00019486939205810004,
      "loss": 0.6495,
      "step": 1100
    },
    {
      "epoch": 0.08432740113490628,
      "grad_norm": 0.44681239128112793,
      "learning_rate": 0.00019440084338760689,
      "loss": 0.6516,
      "step": 1200
    },
    {
      "epoch": 0.09135468456281513,
      "grad_norm": 0.4842969477176666,
      "learning_rate": 0.00019393229471711376,
      "loss": 0.6614,
      "step": 1300
    },
    {
      "epoch": 0.09838196799072399,
      "grad_norm": 0.4123075008392334,
      "learning_rate": 0.00019346374604662061,
      "loss": 0.636,
      "step": 1400
    },
    {
      "epoch": 0.10540925141863285,
      "grad_norm": 0.45350509881973267,
      "learning_rate": 0.00019299519737612746,
      "loss": 0.6429,
      "step": 1500
    },
    {
      "epoch": 0.10540925141863285,
      "eval_loss": 0.6345575451850891,
      "eval_runtime": 192.5238,
      "eval_samples_per_second": 65.701,
      "eval_steps_per_second": 8.217,
      "step": 1500
    },
    {
      "epoch": 0.1124365348465417,
      "grad_norm": 0.36109960079193115,
      "learning_rate": 0.00019252664870563431,
      "loss": 0.6542,
      "step": 1600
    },
    {
      "epoch": 0.11946381827445056,
      "grad_norm": 0.3524057865142822,
      "learning_rate": 0.00019205810003514116,
      "loss": 0.6287,
      "step": 1700
    },
    {
      "epoch": 0.12649110170235942,
      "grad_norm": 0.42591190338134766,
      "learning_rate": 0.00019158955136464801,
      "loss": 0.645,
      "step": 1800
    },
    {
      "epoch": 0.13351838513026826,
      "grad_norm": 0.43938347697257996,
      "learning_rate": 0.00019112100269415486,
      "loss": 0.6444,
      "step": 1900
    },
    {
      "epoch": 0.14054566855817713,
      "grad_norm": 0.44198718667030334,
      "learning_rate": 0.00019065245402366172,
      "loss": 0.643,
      "step": 2000
    },
    {
      "epoch": 0.14054566855817713,
      "eval_loss": 0.6290879249572754,
      "eval_runtime": 192.8097,
      "eval_samples_per_second": 65.604,
      "eval_steps_per_second": 8.205,
      "step": 2000
    },
    {
      "epoch": 0.14757295198608597,
      "grad_norm": 0.39245936274528503,
      "learning_rate": 0.00019018390535316857,
      "loss": 0.65,
      "step": 2100
    },
    {
      "epoch": 0.15460023541399484,
      "grad_norm": 0.46238473057746887,
      "learning_rate": 0.00018971535668267542,
      "loss": 0.6557,
      "step": 2200
    },
    {
      "epoch": 0.16162751884190368,
      "grad_norm": 0.5030249357223511,
      "learning_rate": 0.00018924680801218227,
      "loss": 0.6384,
      "step": 2300
    },
    {
      "epoch": 0.16865480226981255,
      "grad_norm": 0.4010857343673706,
      "learning_rate": 0.00018877825934168914,
      "loss": 0.6424,
      "step": 2400
    },
    {
      "epoch": 0.1756820856977214,
      "grad_norm": 0.43222537636756897,
      "learning_rate": 0.000188309710671196,
      "loss": 0.6217,
      "step": 2500
    },
    {
      "epoch": 0.1756820856977214,
      "eval_loss": 0.6259664297103882,
      "eval_runtime": 196.2284,
      "eval_samples_per_second": 64.461,
      "eval_steps_per_second": 8.062,
      "step": 2500
    },
    {
      "epoch": 0.18270936912563027,
      "grad_norm": 0.5106024742126465,
      "learning_rate": 0.00018784116200070284,
      "loss": 0.6367,
      "step": 2600
    },
    {
      "epoch": 0.1897366525535391,
      "grad_norm": 0.463365763425827,
      "learning_rate": 0.00018737261333020967,
      "loss": 0.6392,
      "step": 2700
    },
    {
      "epoch": 0.19676393598144798,
      "grad_norm": 0.46747493743896484,
      "learning_rate": 0.00018690406465971652,
      "loss": 0.6221,
      "step": 2800
    },
    {
      "epoch": 0.20379121940935682,
      "grad_norm": 0.5577239394187927,
      "learning_rate": 0.0001864355159892234,
      "loss": 0.6313,
      "step": 2900
    },
    {
      "epoch": 0.2108185028372657,
      "grad_norm": 0.3764529228210449,
      "learning_rate": 0.00018596696731873024,
      "loss": 0.6309,
      "step": 3000
    },
    {
      "epoch": 0.2108185028372657,
      "eval_loss": 0.61882084608078,
      "eval_runtime": 191.698,
      "eval_samples_per_second": 65.984,
      "eval_steps_per_second": 8.253,
      "step": 3000
    },
    {
      "epoch": 0.21784578626517453,
      "grad_norm": 0.45881420373916626,
      "learning_rate": 0.0001854984186482371,
      "loss": 0.6177,
      "step": 3100
    },
    {
      "epoch": 0.2248730696930834,
      "grad_norm": 0.366277277469635,
      "learning_rate": 0.00018502986997774394,
      "loss": 0.6217,
      "step": 3200
    },
    {
      "epoch": 0.23190035312099225,
      "grad_norm": 0.5218221545219421,
      "learning_rate": 0.0001845613213072508,
      "loss": 0.6085,
      "step": 3300
    },
    {
      "epoch": 0.23892763654890112,
      "grad_norm": 0.42059990763664246,
      "learning_rate": 0.00018409277263675767,
      "loss": 0.624,
      "step": 3400
    },
    {
      "epoch": 0.24595491997680996,
      "grad_norm": 0.4570040702819824,
      "learning_rate": 0.0001836242239662645,
      "loss": 0.6151,
      "step": 3500
    },
    {
      "epoch": 0.24595491997680996,
      "eval_loss": 0.6193254590034485,
      "eval_runtime": 191.8558,
      "eval_samples_per_second": 65.93,
      "eval_steps_per_second": 8.246,
      "step": 3500
    },
    {
      "epoch": 0.25298220340471883,
      "grad_norm": 0.42353853583335876,
      "learning_rate": 0.00018315567529577135,
      "loss": 0.6245,
      "step": 3600
    },
    {
      "epoch": 0.2600094868326277,
      "grad_norm": 0.4249242842197418,
      "learning_rate": 0.0001826871266252782,
      "loss": 0.6323,
      "step": 3700
    },
    {
      "epoch": 0.2670367702605365,
      "grad_norm": 0.48913854360580444,
      "learning_rate": 0.00018221857795478505,
      "loss": 0.6393,
      "step": 3800
    },
    {
      "epoch": 0.2740640536884454,
      "grad_norm": 0.40342265367507935,
      "learning_rate": 0.00018175002928429192,
      "loss": 0.6241,
      "step": 3900
    },
    {
      "epoch": 0.28109133711635426,
      "grad_norm": 0.46278470754623413,
      "learning_rate": 0.00018128148061379877,
      "loss": 0.6235,
      "step": 4000
    },
    {
      "epoch": 0.28109133711635426,
      "eval_loss": 0.6143574714660645,
      "eval_runtime": 192.2015,
      "eval_samples_per_second": 65.811,
      "eval_steps_per_second": 8.231,
      "step": 4000
    },
    {
      "epoch": 0.2881186205442631,
      "grad_norm": 0.5402960777282715,
      "learning_rate": 0.00018081293194330562,
      "loss": 0.6106,
      "step": 4100
    },
    {
      "epoch": 0.29514590397217194,
      "grad_norm": 0.49978354573249817,
      "learning_rate": 0.00018034438327281247,
      "loss": 0.6316,
      "step": 4200
    },
    {
      "epoch": 0.3021731874000808,
      "grad_norm": 0.4171115458011627,
      "learning_rate": 0.00017987583460231932,
      "loss": 0.6329,
      "step": 4300
    },
    {
      "epoch": 0.3092004708279897,
      "grad_norm": 0.48519599437713623,
      "learning_rate": 0.00017940728593182617,
      "loss": 0.6204,
      "step": 4400
    },
    {
      "epoch": 0.31622775425589855,
      "grad_norm": 0.4046633243560791,
      "learning_rate": 0.00017893873726133302,
      "loss": 0.6291,
      "step": 4500
    },
    {
      "epoch": 0.31622775425589855,
      "eval_loss": 0.6102128028869629,
      "eval_runtime": 194.9353,
      "eval_samples_per_second": 64.888,
      "eval_steps_per_second": 8.116,
      "step": 4500
    },
    {
      "epoch": 0.32325503768380737,
      "grad_norm": 0.5373474359512329,
      "learning_rate": 0.00017847018859083988,
      "loss": 0.6137,
      "step": 4600
    },
    {
      "epoch": 0.33028232111171624,
      "grad_norm": 0.45230546593666077,
      "learning_rate": 0.00017800163992034673,
      "loss": 0.6254,
      "step": 4700
    },
    {
      "epoch": 0.3373096045396251,
      "grad_norm": 0.5114957094192505,
      "learning_rate": 0.00017753309124985358,
      "loss": 0.6256,
      "step": 4800
    },
    {
      "epoch": 0.344336887967534,
      "grad_norm": 0.5685564279556274,
      "learning_rate": 0.00017706454257936045,
      "loss": 0.6055,
      "step": 4900
    },
    {
      "epoch": 0.3513641713954428,
      "grad_norm": 0.47926437854766846,
      "learning_rate": 0.0001765959939088673,
      "loss": 0.6151,
      "step": 5000
    },
    {
      "epoch": 0.3513641713954428,
      "eval_loss": 0.6096778512001038,
      "eval_runtime": 191.152,
      "eval_samples_per_second": 66.172,
      "eval_steps_per_second": 8.276,
      "step": 5000
    },
    {
      "epoch": 0.35839145482335166,
      "grad_norm": 0.4746282696723938,
      "learning_rate": 0.00017612744523837415,
      "loss": 0.6122,
      "step": 5100
    },
    {
      "epoch": 0.36541873825126053,
      "grad_norm": 0.43606746196746826,
      "learning_rate": 0.000175658896567881,
      "loss": 0.6087,
      "step": 5200
    },
    {
      "epoch": 0.37244602167916935,
      "grad_norm": 0.6463629603385925,
      "learning_rate": 0.00017519034789738785,
      "loss": 0.6341,
      "step": 5300
    },
    {
      "epoch": 0.3794733051070782,
      "grad_norm": 0.5722381472587585,
      "learning_rate": 0.0001747217992268947,
      "loss": 0.6224,
      "step": 5400
    },
    {
      "epoch": 0.3865005885349871,
      "grad_norm": 0.45890772342681885,
      "learning_rate": 0.00017425325055640155,
      "loss": 0.611,
      "step": 5500
    },
    {
      "epoch": 0.3865005885349871,
      "eval_loss": 0.6061792969703674,
      "eval_runtime": 191.541,
      "eval_samples_per_second": 66.038,
      "eval_steps_per_second": 8.259,
      "step": 5500
    },
    {
      "epoch": 0.39352787196289596,
      "grad_norm": 0.32096725702285767,
      "learning_rate": 0.0001737847018859084,
      "loss": 0.6031,
      "step": 5600
    },
    {
      "epoch": 0.4005551553908048,
      "grad_norm": 0.46782708168029785,
      "learning_rate": 0.00017331615321541525,
      "loss": 0.5948,
      "step": 5700
    },
    {
      "epoch": 0.40758243881871364,
      "grad_norm": 0.5056571364402771,
      "learning_rate": 0.0001728476045449221,
      "loss": 0.6142,
      "step": 5800
    },
    {
      "epoch": 0.4146097222466225,
      "grad_norm": 0.5338990688323975,
      "learning_rate": 0.00017237905587442898,
      "loss": 0.6235,
      "step": 5900
    },
    {
      "epoch": 0.4216370056745314,
      "grad_norm": 0.5611251592636108,
      "learning_rate": 0.00017191050720393583,
      "loss": 0.6064,
      "step": 6000
    },
    {
      "epoch": 0.4216370056745314,
      "eval_loss": 0.6039415001869202,
      "eval_runtime": 192.9342,
      "eval_samples_per_second": 65.561,
      "eval_steps_per_second": 8.2,
      "step": 6000
    },
    {
      "epoch": 0.4286642891024402,
      "grad_norm": 0.34663501381874084,
      "learning_rate": 0.00017144195853344268,
      "loss": 0.619,
      "step": 6100
    },
    {
      "epoch": 0.43569157253034907,
      "grad_norm": 0.5357178449630737,
      "learning_rate": 0.0001709734098629495,
      "loss": 0.6148,
      "step": 6200
    },
    {
      "epoch": 0.44271885595825794,
      "grad_norm": 0.49474236369132996,
      "learning_rate": 0.00017050486119245636,
      "loss": 0.6021,
      "step": 6300
    },
    {
      "epoch": 0.4497461393861668,
      "grad_norm": 0.4276539981365204,
      "learning_rate": 0.00017003631252196323,
      "loss": 0.619,
      "step": 6400
    },
    {
      "epoch": 0.4567734228140756,
      "grad_norm": 0.4469217360019684,
      "learning_rate": 0.00016956776385147008,
      "loss": 0.6044,
      "step": 6500
    },
    {
      "epoch": 0.4567734228140756,
      "eval_loss": 0.6005575656890869,
      "eval_runtime": 191.475,
      "eval_samples_per_second": 66.061,
      "eval_steps_per_second": 8.262,
      "step": 6500
    },
    {
      "epoch": 0.4638007062419845,
      "grad_norm": 0.4944908916950226,
      "learning_rate": 0.00016909921518097693,
      "loss": 0.5871,
      "step": 6600
    },
    {
      "epoch": 0.47082798966989337,
      "grad_norm": 0.6753395795822144,
      "learning_rate": 0.00016863066651048378,
      "loss": 0.6058,
      "step": 6700
    },
    {
      "epoch": 0.47785527309780224,
      "grad_norm": 0.4985552132129669,
      "learning_rate": 0.00016816211783999063,
      "loss": 0.5935,
      "step": 6800
    },
    {
      "epoch": 0.48488255652571105,
      "grad_norm": 0.47422492504119873,
      "learning_rate": 0.0001676935691694975,
      "loss": 0.6138,
      "step": 6900
    },
    {
      "epoch": 0.4919098399536199,
      "grad_norm": 0.4696919620037079,
      "learning_rate": 0.00016722502049900433,
      "loss": 0.595,
      "step": 7000
    },
    {
      "epoch": 0.4919098399536199,
      "eval_loss": 0.5999007225036621,
      "eval_runtime": 192.1728,
      "eval_samples_per_second": 65.821,
      "eval_steps_per_second": 8.232,
      "step": 7000
    },
    {
      "epoch": 0.4989371233815288,
      "grad_norm": 0.5600836873054504,
      "learning_rate": 0.00016675647182851119,
      "loss": 0.6183,
      "step": 7100
    },
    {
      "epoch": 0.5059644068094377,
      "grad_norm": 0.600051760673523,
      "learning_rate": 0.00016628792315801804,
      "loss": 0.6035,
      "step": 7200
    },
    {
      "epoch": 0.5129916902373465,
      "grad_norm": 0.4427150785923004,
      "learning_rate": 0.00016581937448752489,
      "loss": 0.5925,
      "step": 7300
    },
    {
      "epoch": 0.5200189736652554,
      "grad_norm": 0.5799112319946289,
      "learning_rate": 0.00016535082581703176,
      "loss": 0.5938,
      "step": 7400
    },
    {
      "epoch": 0.5270462570931642,
      "grad_norm": 0.5303583741188049,
      "learning_rate": 0.0001648822771465386,
      "loss": 0.6041,
      "step": 7500
    },
    {
      "epoch": 0.5270462570931642,
      "eval_loss": 0.5995006561279297,
      "eval_runtime": 192.153,
      "eval_samples_per_second": 65.828,
      "eval_steps_per_second": 8.233,
      "step": 7500
    },
    {
      "epoch": 0.534073540521073,
      "grad_norm": 0.5901041030883789,
      "learning_rate": 0.00016441372847604546,
      "loss": 0.6089,
      "step": 7600
    },
    {
      "epoch": 0.541100823948982,
      "grad_norm": 0.4662562608718872,
      "learning_rate": 0.00016394517980555231,
      "loss": 0.6041,
      "step": 7700
    },
    {
      "epoch": 0.5481281073768908,
      "grad_norm": 0.578330934047699,
      "learning_rate": 0.00016347663113505916,
      "loss": 0.6042,
      "step": 7800
    },
    {
      "epoch": 0.5551553908047996,
      "grad_norm": 0.4467080533504486,
      "learning_rate": 0.00016300808246456601,
      "loss": 0.6106,
      "step": 7900
    },
    {
      "epoch": 0.5621826742327085,
      "grad_norm": 0.5268563628196716,
      "learning_rate": 0.00016253953379407286,
      "loss": 0.6064,
      "step": 8000
    },
    {
      "epoch": 0.5621826742327085,
      "eval_loss": 0.5974312424659729,
      "eval_runtime": 191.7777,
      "eval_samples_per_second": 65.957,
      "eval_steps_per_second": 8.249,
      "step": 8000
    },
    {
      "epoch": 0.5692099576606173,
      "grad_norm": 0.5392386317253113,
      "learning_rate": 0.00016207098512357971,
      "loss": 0.5957,
      "step": 8100
    },
    {
      "epoch": 0.5762372410885263,
      "grad_norm": 0.4071332812309265,
      "learning_rate": 0.00016160243645308656,
      "loss": 0.594,
      "step": 8200
    },
    {
      "epoch": 0.5832645245164351,
      "grad_norm": 0.49810290336608887,
      "learning_rate": 0.00016113388778259341,
      "loss": 0.6087,
      "step": 8300
    },
    {
      "epoch": 0.5902918079443439,
      "grad_norm": 0.5486442446708679,
      "learning_rate": 0.0001606653391121003,
      "loss": 0.6068,
      "step": 8400
    },
    {
      "epoch": 0.5973190913722528,
      "grad_norm": 0.4476531147956848,
      "learning_rate": 0.00016019679044160714,
      "loss": 0.6008,
      "step": 8500
    },
    {
      "epoch": 0.5973190913722528,
      "eval_loss": 0.5947303175926208,
      "eval_runtime": 191.9868,
      "eval_samples_per_second": 65.885,
      "eval_steps_per_second": 8.24,
      "step": 8500
    },
    {
      "epoch": 0.6043463748001616,
      "grad_norm": 0.46748480200767517,
      "learning_rate": 0.000159728241771114,
      "loss": 0.6073,
      "step": 8600
    },
    {
      "epoch": 0.6113736582280704,
      "grad_norm": 0.3768276870250702,
      "learning_rate": 0.00015925969310062084,
      "loss": 0.6025,
      "step": 8700
    },
    {
      "epoch": 0.6184009416559794,
      "grad_norm": 0.4775266647338867,
      "learning_rate": 0.0001587911444301277,
      "loss": 0.6139,
      "step": 8800
    },
    {
      "epoch": 0.6254282250838882,
      "grad_norm": 0.595317542552948,
      "learning_rate": 0.00015832259575963452,
      "loss": 0.606,
      "step": 8900
    },
    {
      "epoch": 0.6324555085117971,
      "grad_norm": 0.5021831393241882,
      "learning_rate": 0.0001578540470891414,
      "loss": 0.6141,
      "step": 9000
    },
    {
      "epoch": 0.6324555085117971,
      "eval_loss": 0.5933237671852112,
      "eval_runtime": 191.7573,
      "eval_samples_per_second": 65.964,
      "eval_steps_per_second": 8.25,
      "step": 9000
    },
    {
      "epoch": 0.6394827919397059,
      "grad_norm": 0.4338054955005646,
      "learning_rate": 0.00015738549841864824,
      "loss": 0.6081,
      "step": 9100
    },
    {
      "epoch": 0.6465100753676147,
      "grad_norm": 0.5167454481124878,
      "learning_rate": 0.0001569169497481551,
      "loss": 0.5881,
      "step": 9200
    },
    {
      "epoch": 0.6535373587955237,
      "grad_norm": 0.4964050054550171,
      "learning_rate": 0.00015644840107766194,
      "loss": 0.6227,
      "step": 9300
    },
    {
      "epoch": 0.6605646422234325,
      "grad_norm": 0.5711501240730286,
      "learning_rate": 0.0001559798524071688,
      "loss": 0.5927,
      "step": 9400
    },
    {
      "epoch": 0.6675919256513413,
      "grad_norm": 0.4459337890148163,
      "learning_rate": 0.00015551130373667567,
      "loss": 0.5921,
      "step": 9500
    },
    {
      "epoch": 0.6675919256513413,
      "eval_loss": 0.593291163444519,
      "eval_runtime": 191.3053,
      "eval_samples_per_second": 66.119,
      "eval_steps_per_second": 8.27,
      "step": 9500
    },
    {
      "epoch": 0.6746192090792502,
      "grad_norm": 0.5358896255493164,
      "learning_rate": 0.00015504275506618252,
      "loss": 0.595,
      "step": 9600
    },
    {
      "epoch": 0.681646492507159,
      "grad_norm": 0.5215060114860535,
      "learning_rate": 0.00015457420639568935,
      "loss": 0.6073,
      "step": 9700
    },
    {
      "epoch": 0.688673775935068,
      "grad_norm": 0.43110188841819763,
      "learning_rate": 0.0001541056577251962,
      "loss": 0.5947,
      "step": 9800
    },
    {
      "epoch": 0.6957010593629768,
      "grad_norm": 0.7894725799560547,
      "learning_rate": 0.00015363710905470305,
      "loss": 0.6036,
      "step": 9900
    },
    {
      "epoch": 0.7027283427908856,
      "grad_norm": 0.40468016266822815,
      "learning_rate": 0.00015316856038420992,
      "loss": 0.5662,
      "step": 10000
    },
    {
      "epoch": 0.7027283427908856,
      "eval_loss": 0.5914621949195862,
      "eval_runtime": 191.9924,
      "eval_samples_per_second": 65.883,
      "eval_steps_per_second": 8.24,
      "step": 10000
    },
    {
      "epoch": 0.7097556262187945,
      "grad_norm": 0.520021915435791,
      "learning_rate": 0.00015270001171371677,
      "loss": 0.6024,
      "step": 10100
    },
    {
      "epoch": 0.7167829096467033,
      "grad_norm": 0.5001341700553894,
      "learning_rate": 0.00015223146304322362,
      "loss": 0.6039,
      "step": 10200
    },
    {
      "epoch": 0.7238101930746121,
      "grad_norm": 0.5081097483634949,
      "learning_rate": 0.00015176291437273047,
      "loss": 0.5995,
      "step": 10300
    },
    {
      "epoch": 0.7308374765025211,
      "grad_norm": 0.7181329131126404,
      "learning_rate": 0.00015129436570223732,
      "loss": 0.5885,
      "step": 10400
    },
    {
      "epoch": 0.7378647599304299,
      "grad_norm": 0.5525663495063782,
      "learning_rate": 0.00015082581703174417,
      "loss": 0.6117,
      "step": 10500
    },
    {
      "epoch": 0.7378647599304299,
      "eval_loss": 0.5893301963806152,
      "eval_runtime": 191.2328,
      "eval_samples_per_second": 66.145,
      "eval_steps_per_second": 8.273,
      "step": 10500
    },
    {
      "epoch": 0.7448920433583387,
      "grad_norm": 0.41122502088546753,
      "learning_rate": 0.00015035726836125102,
      "loss": 0.6135,
      "step": 10600
    },
    {
      "epoch": 0.7519193267862476,
      "grad_norm": 0.501287043094635,
      "learning_rate": 0.00014988871969075787,
      "loss": 0.6045,
      "step": 10700
    },
    {
      "epoch": 0.7589466102141564,
      "grad_norm": 0.4683501720428467,
      "learning_rate": 0.00014942017102026472,
      "loss": 0.6044,
      "step": 10800
    },
    {
      "epoch": 0.7659738936420654,
      "grad_norm": 0.5879314541816711,
      "learning_rate": 0.00014895162234977158,
      "loss": 0.5979,
      "step": 10900
    },
    {
      "epoch": 0.7730011770699742,
      "grad_norm": 0.4201727509498596,
      "learning_rate": 0.00014848307367927845,
      "loss": 0.6133,
      "step": 11000
    },
    {
      "epoch": 0.7730011770699742,
      "eval_loss": 0.5890333652496338,
      "eval_runtime": 191.871,
      "eval_samples_per_second": 65.925,
      "eval_steps_per_second": 8.245,
      "step": 11000
    },
    {
      "epoch": 0.780028460497883,
      "grad_norm": 0.4720204770565033,
      "learning_rate": 0.0001480145250087853,
      "loss": 0.5926,
      "step": 11100
    },
    {
      "epoch": 0.7870557439257919,
      "grad_norm": 0.4714803695678711,
      "learning_rate": 0.00014754597633829215,
      "loss": 0.5936,
      "step": 11200
    },
    {
      "epoch": 0.7940830273537007,
      "grad_norm": 0.44766145944595337,
      "learning_rate": 0.000147077427667799,
      "loss": 0.5898,
      "step": 11300
    },
    {
      "epoch": 0.8011103107816095,
      "grad_norm": 0.5793395638465881,
      "learning_rate": 0.00014660887899730585,
      "loss": 0.5792,
      "step": 11400
    },
    {
      "epoch": 0.8081375942095185,
      "grad_norm": 0.5724961757659912,
      "learning_rate": 0.0001461403303268127,
      "loss": 0.5976,
      "step": 11500
    },
    {
      "epoch": 0.8081375942095185,
      "eval_loss": 0.5884692072868347,
      "eval_runtime": 192.9015,
      "eval_samples_per_second": 65.572,
      "eval_steps_per_second": 8.201,
      "step": 11500
    },
    {
      "epoch": 0.8151648776374273,
      "grad_norm": 0.4588712751865387,
      "learning_rate": 0.00014567178165631955,
      "loss": 0.5876,
      "step": 11600
    },
    {
      "epoch": 0.8221921610653362,
      "grad_norm": 0.5474816560745239,
      "learning_rate": 0.0001452032329858264,
      "loss": 0.6004,
      "step": 11700
    },
    {
      "epoch": 0.829219444493245,
      "grad_norm": 0.480760395526886,
      "learning_rate": 0.00014473468431533325,
      "loss": 0.6027,
      "step": 11800
    },
    {
      "epoch": 0.8362467279211538,
      "grad_norm": 0.44635719060897827,
      "learning_rate": 0.0001442661356448401,
      "loss": 0.5955,
      "step": 11900
    },
    {
      "epoch": 0.8432740113490628,
      "grad_norm": 0.370127409696579,
      "learning_rate": 0.00014379758697434698,
      "loss": 0.5873,
      "step": 12000
    },
    {
      "epoch": 0.8432740113490628,
      "eval_loss": 0.5867401361465454,
      "eval_runtime": 191.0689,
      "eval_samples_per_second": 66.201,
      "eval_steps_per_second": 8.28,
      "step": 12000
    },
    {
      "epoch": 0.8503012947769716,
      "grad_norm": 0.5330467224121094,
      "learning_rate": 0.00014332903830385383,
      "loss": 0.6036,
      "step": 12100
    },
    {
      "epoch": 0.8573285782048804,
      "grad_norm": 0.516147255897522,
      "learning_rate": 0.00014286048963336068,
      "loss": 0.5808,
      "step": 12200
    },
    {
      "epoch": 0.8643558616327893,
      "grad_norm": 0.5745691061019897,
      "learning_rate": 0.00014239194096286753,
      "loss": 0.6026,
      "step": 12300
    },
    {
      "epoch": 0.8713831450606981,
      "grad_norm": 0.5167753100395203,
      "learning_rate": 0.00014192339229237436,
      "loss": 0.5858,
      "step": 12400
    },
    {
      "epoch": 0.8784104284886071,
      "grad_norm": 0.644183874130249,
      "learning_rate": 0.00014145484362188123,
      "loss": 0.6009,
      "step": 12500
    },
    {
      "epoch": 0.8784104284886071,
      "eval_loss": 0.585637629032135,
      "eval_runtime": 192.0709,
      "eval_samples_per_second": 65.856,
      "eval_steps_per_second": 8.237,
      "step": 12500
    },
    {
      "epoch": 0.8854377119165159,
      "grad_norm": 0.5942461490631104,
      "learning_rate": 0.00014098629495138808,
      "loss": 0.5872,
      "step": 12600
    },
    {
      "epoch": 0.8924649953444247,
      "grad_norm": 0.5303371548652649,
      "learning_rate": 0.00014051774628089493,
      "loss": 0.603,
      "step": 12700
    },
    {
      "epoch": 0.8994922787723336,
      "grad_norm": 0.6340320706367493,
      "learning_rate": 0.00014004919761040178,
      "loss": 0.5967,
      "step": 12800
    },
    {
      "epoch": 0.9065195622002424,
      "grad_norm": 0.3721257150173187,
      "learning_rate": 0.00013958064893990863,
      "loss": 0.5951,
      "step": 12900
    },
    {
      "epoch": 0.9135468456281512,
      "grad_norm": 0.5798757672309875,
      "learning_rate": 0.0001391121002694155,
      "loss": 0.5872,
      "step": 13000
    },
    {
      "epoch": 0.9135468456281512,
      "eval_loss": 0.5851631164550781,
      "eval_runtime": 191.7607,
      "eval_samples_per_second": 65.962,
      "eval_steps_per_second": 8.25,
      "step": 13000
    },
    {
      "epoch": 0.9205741290560602,
      "grad_norm": 0.49079081416130066,
      "learning_rate": 0.00013864355159892236,
      "loss": 0.5945,
      "step": 13100
    },
    {
      "epoch": 0.927601412483969,
      "grad_norm": 0.6511490941047668,
      "learning_rate": 0.00013817500292842918,
      "loss": 0.5967,
      "step": 13200
    },
    {
      "epoch": 0.9346286959118779,
      "grad_norm": 0.6914525032043457,
      "learning_rate": 0.00013770645425793603,
      "loss": 0.595,
      "step": 13300
    },
    {
      "epoch": 0.9416559793397867,
      "grad_norm": 0.5525299310684204,
      "learning_rate": 0.00013723790558744289,
      "loss": 0.5994,
      "step": 13400
    },
    {
      "epoch": 0.9486832627676955,
      "grad_norm": 0.5471107363700867,
      "learning_rate": 0.00013676935691694976,
      "loss": 0.5778,
      "step": 13500
    },
    {
      "epoch": 0.9486832627676955,
      "eval_loss": 0.5845098495483398,
      "eval_runtime": 192.0133,
      "eval_samples_per_second": 65.876,
      "eval_steps_per_second": 8.239,
      "step": 13500
    },
    {
      "epoch": 0.9557105461956045,
      "grad_norm": 0.6058303117752075,
      "learning_rate": 0.0001363008082464566,
      "loss": 0.6011,
      "step": 13600
    },
    {
      "epoch": 0.9627378296235133,
      "grad_norm": 0.5132064819335938,
      "learning_rate": 0.00013583225957596346,
      "loss": 0.5814,
      "step": 13700
    },
    {
      "epoch": 0.9697651130514221,
      "grad_norm": 0.42012500762939453,
      "learning_rate": 0.0001353637109054703,
      "loss": 0.5834,
      "step": 13800
    },
    {
      "epoch": 0.976792396479331,
      "grad_norm": 0.5156047344207764,
      "learning_rate": 0.00013489516223497716,
      "loss": 0.5912,
      "step": 13900
    },
    {
      "epoch": 0.9838196799072398,
      "grad_norm": 0.578301727771759,
      "learning_rate": 0.00013442661356448404,
      "loss": 0.5888,
      "step": 14000
    },
    {
      "epoch": 0.9838196799072398,
      "eval_loss": 0.5840687155723572,
      "eval_runtime": 191.6947,
      "eval_samples_per_second": 65.985,
      "eval_steps_per_second": 8.253,
      "step": 14000
    },
    {
      "epoch": 0.9908469633351488,
      "grad_norm": 0.5532825589179993,
      "learning_rate": 0.00013395806489399086,
      "loss": 0.5951,
      "step": 14100
    },
    {
      "epoch": 0.9978742467630576,
      "grad_norm": 0.7117305994033813,
      "learning_rate": 0.00013348951622349771,
      "loss": 0.6017,
      "step": 14200
    },
    {
      "epoch": 1.0049015301909665,
      "grad_norm": 0.4331432580947876,
      "learning_rate": 0.00013302096755300456,
      "loss": 0.535,
      "step": 14300
    },
    {
      "epoch": 1.0119288136188753,
      "grad_norm": 0.5378111004829407,
      "learning_rate": 0.00013255241888251141,
      "loss": 0.5451,
      "step": 14400
    },
    {
      "epoch": 1.0189560970467841,
      "grad_norm": 0.5767349600791931,
      "learning_rate": 0.0001320838702120183,
      "loss": 0.5469,
      "step": 14500
    },
    {
      "epoch": 1.0189560970467841,
      "eval_loss": 0.5859206318855286,
      "eval_runtime": 192.2055,
      "eval_samples_per_second": 65.81,
      "eval_steps_per_second": 8.231,
      "step": 14500
    },
    {
      "epoch": 1.025983380474693,
      "grad_norm": 0.5575023889541626,
      "learning_rate": 0.00013161532154152514,
      "loss": 0.5449,
      "step": 14600
    },
    {
      "epoch": 1.0330106639026018,
      "grad_norm": 0.5510827302932739,
      "learning_rate": 0.000131146772871032,
      "loss": 0.5339,
      "step": 14700
    },
    {
      "epoch": 1.0400379473305108,
      "grad_norm": 0.5780773758888245,
      "learning_rate": 0.00013067822420053884,
      "loss": 0.5334,
      "step": 14800
    },
    {
      "epoch": 1.0470652307584196,
      "grad_norm": 0.6217513084411621,
      "learning_rate": 0.0001302096755300457,
      "loss": 0.5404,
      "step": 14900
    },
    {
      "epoch": 1.0540925141863284,
      "grad_norm": 0.5290889739990234,
      "learning_rate": 0.00012974112685955254,
      "loss": 0.5368,
      "step": 15000
    },
    {
      "epoch": 1.0540925141863284,
      "eval_loss": 0.5861626863479614,
      "eval_runtime": 192.0995,
      "eval_samples_per_second": 65.846,
      "eval_steps_per_second": 8.235,
      "step": 15000
    },
    {
      "epoch": 1.0611197976142372,
      "grad_norm": 0.5673081278800964,
      "learning_rate": 0.0001292725781890594,
      "loss": 0.5381,
      "step": 15100
    },
    {
      "epoch": 1.068147081042146,
      "grad_norm": 0.4198685884475708,
      "learning_rate": 0.00012880402951856624,
      "loss": 0.5379,
      "step": 15200
    },
    {
      "epoch": 1.075174364470055,
      "grad_norm": 0.45610734820365906,
      "learning_rate": 0.0001283354808480731,
      "loss": 0.5349,
      "step": 15300
    },
    {
      "epoch": 1.082201647897964,
      "grad_norm": 0.5499436259269714,
      "learning_rate": 0.00012786693217757994,
      "loss": 0.5411,
      "step": 15400
    },
    {
      "epoch": 1.0892289313258727,
      "grad_norm": 0.6300696730613708,
      "learning_rate": 0.0001273983835070868,
      "loss": 0.546,
      "step": 15500
    },
    {
      "epoch": 1.0892289313258727,
      "eval_loss": 0.5862005352973938,
      "eval_runtime": 191.9358,
      "eval_samples_per_second": 65.902,
      "eval_steps_per_second": 8.242,
      "step": 15500
    },
    {
      "epoch": 1.0962562147537815,
      "grad_norm": 0.6075917482376099,
      "learning_rate": 0.00012692983483659367,
      "loss": 0.5417,
      "step": 15600
    },
    {
      "epoch": 1.1032834981816904,
      "grad_norm": 0.46201905608177185,
      "learning_rate": 0.00012646128616610052,
      "loss": 0.5283,
      "step": 15700
    },
    {
      "epoch": 1.1103107816095992,
      "grad_norm": 0.4050121605396271,
      "learning_rate": 0.00012599273749560737,
      "loss": 0.5479,
      "step": 15800
    },
    {
      "epoch": 1.1173380650375082,
      "grad_norm": 0.4577913284301758,
      "learning_rate": 0.0001255241888251142,
      "loss": 0.5452,
      "step": 15900
    },
    {
      "epoch": 1.124365348465417,
      "grad_norm": 0.4985063672065735,
      "learning_rate": 0.00012505564015462105,
      "loss": 0.5444,
      "step": 16000
    },
    {
      "epoch": 1.124365348465417,
      "eval_loss": 0.5855303406715393,
      "eval_runtime": 191.4632,
      "eval_samples_per_second": 66.065,
      "eval_steps_per_second": 8.263,
      "step": 16000
    },
    {
      "epoch": 1.1313926318933258,
      "grad_norm": 0.4636366665363312,
      "learning_rate": 0.00012458709148412792,
      "loss": 0.5352,
      "step": 16100
    },
    {
      "epoch": 1.1384199153212347,
      "grad_norm": 0.5733815431594849,
      "learning_rate": 0.00012411854281363477,
      "loss": 0.5339,
      "step": 16200
    },
    {
      "epoch": 1.1454471987491435,
      "grad_norm": 0.5225148797035217,
      "learning_rate": 0.00012364999414314162,
      "loss": 0.5356,
      "step": 16300
    },
    {
      "epoch": 1.1524744821770523,
      "grad_norm": 0.5425903797149658,
      "learning_rate": 0.00012318144547264847,
      "loss": 0.544,
      "step": 16400
    },
    {
      "epoch": 1.1595017656049613,
      "grad_norm": 0.49336591362953186,
      "learning_rate": 0.00012271289680215532,
      "loss": 0.5386,
      "step": 16500
    },
    {
      "epoch": 1.1595017656049613,
      "eval_loss": 0.5844515562057495,
      "eval_runtime": 195.2478,
      "eval_samples_per_second": 64.784,
      "eval_steps_per_second": 8.103,
      "step": 16500
    },
    {
      "epoch": 1.1665290490328701,
      "grad_norm": 0.45879772305488586,
      "learning_rate": 0.0001222443481316622,
      "loss": 0.5334,
      "step": 16600
    },
    {
      "epoch": 1.173556332460779,
      "grad_norm": 0.4931933879852295,
      "learning_rate": 0.00012177579946116904,
      "loss": 0.5362,
      "step": 16700
    },
    {
      "epoch": 1.1805836158886878,
      "grad_norm": 0.7223409414291382,
      "learning_rate": 0.00012130725079067589,
      "loss": 0.5331,
      "step": 16800
    },
    {
      "epoch": 1.1876108993165966,
      "grad_norm": 0.4814956784248352,
      "learning_rate": 0.00012083870212018274,
      "loss": 0.5279,
      "step": 16900
    },
    {
      "epoch": 1.1946381827445056,
      "grad_norm": 0.5712910294532776,
      "learning_rate": 0.00012037015344968959,
      "loss": 0.5269,
      "step": 17000
    },
    {
      "epoch": 1.1946381827445056,
      "eval_loss": 0.5852189660072327,
      "eval_runtime": 191.4648,
      "eval_samples_per_second": 66.064,
      "eval_steps_per_second": 8.263,
      "step": 17000
    },
    {
      "epoch": 1.2016654661724144,
      "grad_norm": 0.5194661021232605,
      "learning_rate": 0.00011990160477919645,
      "loss": 0.5456,
      "step": 17100
    },
    {
      "epoch": 1.2086927496003232,
      "grad_norm": 0.42699483036994934,
      "learning_rate": 0.0001194330561087033,
      "loss": 0.5519,
      "step": 17200
    },
    {
      "epoch": 1.215720033028232,
      "grad_norm": 0.49879905581474304,
      "learning_rate": 0.00011896450743821015,
      "loss": 0.5327,
      "step": 17300
    },
    {
      "epoch": 1.2227473164561409,
      "grad_norm": 0.7226296663284302,
      "learning_rate": 0.000118495958767717,
      "loss": 0.545,
      "step": 17400
    },
    {
      "epoch": 1.22977459988405,
      "grad_norm": 0.5727146863937378,
      "learning_rate": 0.00011802741009722384,
      "loss": 0.5428,
      "step": 17500
    },
    {
      "epoch": 1.22977459988405,
      "eval_loss": 0.5834037065505981,
      "eval_runtime": 191.7102,
      "eval_samples_per_second": 65.98,
      "eval_steps_per_second": 8.252,
      "step": 17500
    },
    {
      "epoch": 1.2368018833119587,
      "grad_norm": 0.6738116145133972,
      "learning_rate": 0.00011755886142673072,
      "loss": 0.5494,
      "step": 17600
    },
    {
      "epoch": 1.2438291667398675,
      "grad_norm": 0.5708216428756714,
      "learning_rate": 0.00011709031275623757,
      "loss": 0.5321,
      "step": 17700
    },
    {
      "epoch": 1.2508564501677764,
      "grad_norm": 0.7585426568984985,
      "learning_rate": 0.00011662176408574442,
      "loss": 0.5405,
      "step": 17800
    },
    {
      "epoch": 1.2578837335956852,
      "grad_norm": 0.48215654492378235,
      "learning_rate": 0.00011615321541525125,
      "loss": 0.5317,
      "step": 17900
    },
    {
      "epoch": 1.2649110170235942,
      "grad_norm": 0.7234981060028076,
      "learning_rate": 0.0001156846667447581,
      "loss": 0.5198,
      "step": 18000
    },
    {
      "epoch": 1.2649110170235942,
      "eval_loss": 0.583666980266571,
      "eval_runtime": 191.0164,
      "eval_samples_per_second": 66.219,
      "eval_steps_per_second": 8.282,
      "step": 18000
    },
    {
      "epoch": 1.271938300451503,
      "grad_norm": 0.6166390180587769,
      "learning_rate": 0.00011521611807426498,
      "loss": 0.5391,
      "step": 18100
    },
    {
      "epoch": 1.2789655838794118,
      "grad_norm": 0.5026915669441223,
      "learning_rate": 0.00011474756940377183,
      "loss": 0.5453,
      "step": 18200
    },
    {
      "epoch": 1.2859928673073207,
      "grad_norm": 0.44040101766586304,
      "learning_rate": 0.00011427902073327867,
      "loss": 0.5459,
      "step": 18300
    },
    {
      "epoch": 1.2930201507352295,
      "grad_norm": 0.4429463744163513,
      "learning_rate": 0.00011381047206278552,
      "loss": 0.5302,
      "step": 18400
    },
    {
      "epoch": 1.3000474341631385,
      "grad_norm": 0.5252630710601807,
      "learning_rate": 0.00011334192339229237,
      "loss": 0.5373,
      "step": 18500
    },
    {
      "epoch": 1.3000474341631385,
      "eval_loss": 0.5821723341941833,
      "eval_runtime": 191.389,
      "eval_samples_per_second": 66.091,
      "eval_steps_per_second": 8.266,
      "step": 18500
    },
    {
      "epoch": 1.3070747175910473,
      "grad_norm": 0.590726912021637,
      "learning_rate": 0.00011287337472179925,
      "loss": 0.5382,
      "step": 18600
    },
    {
      "epoch": 1.3141020010189561,
      "grad_norm": 0.6306003928184509,
      "learning_rate": 0.00011240482605130608,
      "loss": 0.5199,
      "step": 18700
    },
    {
      "epoch": 1.321129284446865,
      "grad_norm": 0.5457239151000977,
      "learning_rate": 0.00011193627738081293,
      "loss": 0.5318,
      "step": 18800
    },
    {
      "epoch": 1.3281565678747738,
      "grad_norm": 0.5287586450576782,
      "learning_rate": 0.00011146772871031978,
      "loss": 0.545,
      "step": 18900
    },
    {
      "epoch": 1.3351838513026828,
      "grad_norm": 0.5725194811820984,
      "learning_rate": 0.00011099918003982663,
      "loss": 0.5439,
      "step": 19000
    },
    {
      "epoch": 1.3351838513026828,
      "eval_loss": 0.5815557241439819,
      "eval_runtime": 191.8001,
      "eval_samples_per_second": 65.949,
      "eval_steps_per_second": 8.248,
      "step": 19000
    },
    {
      "epoch": 1.3422111347305914,
      "grad_norm": 0.6155562400817871,
      "learning_rate": 0.0001105306313693335,
      "loss": 0.5337,
      "step": 19100
    },
    {
      "epoch": 1.3492384181585004,
      "grad_norm": 0.45534420013427734,
      "learning_rate": 0.00011006208269884035,
      "loss": 0.5437,
      "step": 19200
    },
    {
      "epoch": 1.3562657015864092,
      "grad_norm": 0.5221378803253174,
      "learning_rate": 0.0001095935340283472,
      "loss": 0.5433,
      "step": 19300
    },
    {
      "epoch": 1.363292985014318,
      "grad_norm": 0.6740151047706604,
      "learning_rate": 0.00010912498535785405,
      "loss": 0.5315,
      "step": 19400
    },
    {
      "epoch": 1.3703202684422269,
      "grad_norm": 0.5568748116493225,
      "learning_rate": 0.0001086564366873609,
      "loss": 0.5442,
      "step": 19500
    },
    {
      "epoch": 1.3703202684422269,
      "eval_loss": 0.5810661911964417,
      "eval_runtime": 193.8958,
      "eval_samples_per_second": 65.236,
      "eval_steps_per_second": 8.159,
      "step": 19500
    },
    {
      "epoch": 1.3773475518701357,
      "grad_norm": 0.6538209915161133,
      "learning_rate": 0.00010818788801686776,
      "loss": 0.5485,
      "step": 19600
    },
    {
      "epoch": 1.3843748352980447,
      "grad_norm": 0.3977202773094177,
      "learning_rate": 0.00010771933934637461,
      "loss": 0.551,
      "step": 19700
    },
    {
      "epoch": 1.3914021187259535,
      "grad_norm": 0.6462295651435852,
      "learning_rate": 0.00010725079067588146,
      "loss": 0.535,
      "step": 19800
    },
    {
      "epoch": 1.3984294021538624,
      "grad_norm": 0.6669672131538391,
      "learning_rate": 0.00010678224200538831,
      "loss": 0.5372,
      "step": 19900
    },
    {
      "epoch": 1.4054566855817712,
      "grad_norm": 0.672690749168396,
      "learning_rate": 0.00010631369333489516,
      "loss": 0.5562,
      "step": 20000
    },
    {
      "epoch": 1.4054566855817712,
      "eval_loss": 0.5803663730621338,
      "eval_runtime": 191.2217,
      "eval_samples_per_second": 66.148,
      "eval_steps_per_second": 8.273,
      "step": 20000
    },
    {
      "epoch": 1.41248396900968,
      "grad_norm": 0.5990594029426575,
      "learning_rate": 0.00010584514466440203,
      "loss": 0.5402,
      "step": 20100
    },
    {
      "epoch": 1.419511252437589,
      "grad_norm": 0.7736939787864685,
      "learning_rate": 0.00010537659599390888,
      "loss": 0.5514,
      "step": 20200
    },
    {
      "epoch": 1.4265385358654978,
      "grad_norm": 0.7517827749252319,
      "learning_rate": 0.00010490804732341573,
      "loss": 0.5199,
      "step": 20300
    },
    {
      "epoch": 1.4335658192934067,
      "grad_norm": 0.40587854385375977,
      "learning_rate": 0.00010443949865292258,
      "loss": 0.5453,
      "step": 20400
    },
    {
      "epoch": 1.4405931027213155,
      "grad_norm": 0.5752481818199158,
      "learning_rate": 0.00010397094998242943,
      "loss": 0.5327,
      "step": 20500
    },
    {
      "epoch": 1.4405931027213155,
      "eval_loss": 0.5793671607971191,
      "eval_runtime": 191.4019,
      "eval_samples_per_second": 66.086,
      "eval_steps_per_second": 8.265,
      "step": 20500
    },
    {
      "epoch": 1.4476203861492243,
      "grad_norm": 0.5109778642654419,
      "learning_rate": 0.00010350240131193629,
      "loss": 0.534,
      "step": 20600
    },
    {
      "epoch": 1.4546476695771333,
      "grad_norm": 0.5415691137313843,
      "learning_rate": 0.00010303385264144314,
      "loss": 0.5346,
      "step": 20700
    },
    {
      "epoch": 1.4616749530050421,
      "grad_norm": 0.6798147559165955,
      "learning_rate": 0.00010256530397094999,
      "loss": 0.5389,
      "step": 20800
    },
    {
      "epoch": 1.468702236432951,
      "grad_norm": 0.5930722951889038,
      "learning_rate": 0.00010209675530045684,
      "loss": 0.5553,
      "step": 20900
    },
    {
      "epoch": 1.4757295198608598,
      "grad_norm": 0.46035194396972656,
      "learning_rate": 0.00010162820662996368,
      "loss": 0.5488,
      "step": 21000
    },
    {
      "epoch": 1.4757295198608598,
      "eval_loss": 0.5797436833381653,
      "eval_runtime": 191.1985,
      "eval_samples_per_second": 66.156,
      "eval_steps_per_second": 8.274,
      "step": 21000
    },
    {
      "epoch": 1.4827568032887686,
      "grad_norm": 0.6289058327674866,
      "learning_rate": 0.00010115965795947056,
      "loss": 0.5373,
      "step": 21100
    },
    {
      "epoch": 1.4897840867166776,
      "grad_norm": 0.5587871670722961,
      "learning_rate": 0.0001006911092889774,
      "loss": 0.5433,
      "step": 21200
    },
    {
      "epoch": 1.4968113701445864,
      "grad_norm": 0.6402022838592529,
      "learning_rate": 0.00010022256061848426,
      "loss": 0.5369,
      "step": 21300
    },
    {
      "epoch": 1.5038386535724952,
      "grad_norm": 0.7074068784713745,
      "learning_rate": 9.975401194799109e-05,
      "loss": 0.5422,
      "step": 21400
    },
    {
      "epoch": 1.510865937000404,
      "grad_norm": 0.6283564567565918,
      "learning_rate": 9.928546327749796e-05,
      "loss": 0.53,
      "step": 21500
    },
    {
      "epoch": 1.510865937000404,
      "eval_loss": 0.5776032209396362,
      "eval_runtime": 195.2716,
      "eval_samples_per_second": 64.776,
      "eval_steps_per_second": 8.102,
      "step": 21500
    },
    {
      "epoch": 1.5178932204283129,
      "grad_norm": 0.5513471961021423,
      "learning_rate": 9.881691460700481e-05,
      "loss": 0.5363,
      "step": 21600
    },
    {
      "epoch": 1.524920503856222,
      "grad_norm": 0.6834294199943542,
      "learning_rate": 9.834836593651166e-05,
      "loss": 0.5357,
      "step": 21700
    },
    {
      "epoch": 1.5319477872841305,
      "grad_norm": 0.5920292139053345,
      "learning_rate": 9.787981726601851e-05,
      "loss": 0.5425,
      "step": 21800
    },
    {
      "epoch": 1.5389750707120395,
      "grad_norm": 0.7391265630722046,
      "learning_rate": 9.741126859552536e-05,
      "loss": 0.5536,
      "step": 21900
    },
    {
      "epoch": 1.5460023541399484,
      "grad_norm": 0.59616619348526,
      "learning_rate": 9.694271992503222e-05,
      "loss": 0.5266,
      "step": 22000
    },
    {
      "epoch": 1.5460023541399484,
      "eval_loss": 0.5772854089736938,
      "eval_runtime": 191.6654,
      "eval_samples_per_second": 65.995,
      "eval_steps_per_second": 8.254,
      "step": 22000
    },
    {
      "epoch": 1.5530296375678572,
      "grad_norm": 0.7118886709213257,
      "learning_rate": 9.647417125453907e-05,
      "loss": 0.5402,
      "step": 22100
    },
    {
      "epoch": 1.5600569209957662,
      "grad_norm": 0.5848612189292908,
      "learning_rate": 9.600562258404592e-05,
      "loss": 0.5471,
      "step": 22200
    },
    {
      "epoch": 1.5670842044236748,
      "grad_norm": 0.6379753947257996,
      "learning_rate": 9.553707391355277e-05,
      "loss": 0.5319,
      "step": 22300
    },
    {
      "epoch": 1.5741114878515838,
      "grad_norm": 0.5326395630836487,
      "learning_rate": 9.506852524305962e-05,
      "loss": 0.5362,
      "step": 22400
    },
    {
      "epoch": 1.5811387712794926,
      "grad_norm": 0.44863948225975037,
      "learning_rate": 9.459997657256649e-05,
      "loss": 0.5339,
      "step": 22500
    },
    {
      "epoch": 1.5811387712794926,
      "eval_loss": 0.5767233371734619,
      "eval_runtime": 191.4103,
      "eval_samples_per_second": 66.083,
      "eval_steps_per_second": 8.265,
      "step": 22500
    },
    {
      "epoch": 1.5881660547074015,
      "grad_norm": 0.6882546544075012,
      "learning_rate": 9.413142790207334e-05,
      "loss": 0.5338,
      "step": 22600
    },
    {
      "epoch": 1.5951933381353105,
      "grad_norm": 0.6498861312866211,
      "learning_rate": 9.366287923158019e-05,
      "loss": 0.5449,
      "step": 22700
    },
    {
      "epoch": 1.602220621563219,
      "grad_norm": 0.5820225477218628,
      "learning_rate": 9.319433056108704e-05,
      "loss": 0.541,
      "step": 22800
    },
    {
      "epoch": 1.6092479049911281,
      "grad_norm": 0.6746723651885986,
      "learning_rate": 9.272578189059389e-05,
      "loss": 0.5437,
      "step": 22900
    },
    {
      "epoch": 1.616275188419037,
      "grad_norm": 0.6235536932945251,
      "learning_rate": 9.225723322010075e-05,
      "loss": 0.5332,
      "step": 23000
    },
    {
      "epoch": 1.616275188419037,
      "eval_loss": 0.575995683670044,
      "eval_runtime": 191.6609,
      "eval_samples_per_second": 65.997,
      "eval_steps_per_second": 8.254,
      "step": 23000
    },
    {
      "epoch": 1.6233024718469458,
      "grad_norm": 0.7106947898864746,
      "learning_rate": 9.17886845496076e-05,
      "loss": 0.5466,
      "step": 23100
    },
    {
      "epoch": 1.6303297552748546,
      "grad_norm": 0.5090802907943726,
      "learning_rate": 9.132013587911444e-05,
      "loss": 0.5244,
      "step": 23200
    },
    {
      "epoch": 1.6373570387027634,
      "grad_norm": 0.5423375368118286,
      "learning_rate": 9.08515872086213e-05,
      "loss": 0.5331,
      "step": 23300
    },
    {
      "epoch": 1.6443843221306724,
      "grad_norm": 0.40430954098701477,
      "learning_rate": 9.038303853812815e-05,
      "loss": 0.543,
      "step": 23400
    },
    {
      "epoch": 1.6514116055585812,
      "grad_norm": 0.5849967002868652,
      "learning_rate": 8.991448986763502e-05,
      "loss": 0.5313,
      "step": 23500
    },
    {
      "epoch": 1.6514116055585812,
      "eval_loss": 0.5759484171867371,
      "eval_runtime": 198.3535,
      "eval_samples_per_second": 63.77,
      "eval_steps_per_second": 7.976,
      "step": 23500
    },
    {
      "epoch": 1.65843888898649,
      "grad_norm": 0.642866849899292,
      "learning_rate": 8.944594119714185e-05,
      "loss": 0.5375,
      "step": 23600
    },
    {
      "epoch": 1.6654661724143989,
      "grad_norm": 0.4610786437988281,
      "learning_rate": 8.89773925266487e-05,
      "loss": 0.5262,
      "step": 23700
    },
    {
      "epoch": 1.6724934558423077,
      "grad_norm": 0.5731543302536011,
      "learning_rate": 8.850884385615557e-05,
      "loss": 0.5316,
      "step": 23800
    },
    {
      "epoch": 1.6795207392702167,
      "grad_norm": 0.6324673295021057,
      "learning_rate": 8.804029518566242e-05,
      "loss": 0.5391,
      "step": 23900
    },
    {
      "epoch": 1.6865480226981253,
      "grad_norm": 0.7439347505569458,
      "learning_rate": 8.757174651516927e-05,
      "loss": 0.5328,
      "step": 24000
    },
    {
      "epoch": 1.6865480226981253,
      "eval_loss": 0.5753322839736938,
      "eval_runtime": 191.1076,
      "eval_samples_per_second": 66.188,
      "eval_steps_per_second": 8.278,
      "step": 24000
    },
    {
      "epoch": 1.6935753061260344,
      "grad_norm": 0.5916799306869507,
      "learning_rate": 8.710319784467612e-05,
      "loss": 0.5357,
      "step": 24100
    },
    {
      "epoch": 1.7006025895539432,
      "grad_norm": 0.641130268573761,
      "learning_rate": 8.663464917418297e-05,
      "loss": 0.5413,
      "step": 24200
    },
    {
      "epoch": 1.707629872981852,
      "grad_norm": 0.5414934754371643,
      "learning_rate": 8.616610050368983e-05,
      "loss": 0.5372,
      "step": 24300
    },
    {
      "epoch": 1.714657156409761,
      "grad_norm": 0.5349559783935547,
      "learning_rate": 8.569755183319668e-05,
      "loss": 0.5248,
      "step": 24400
    },
    {
      "epoch": 1.7216844398376696,
      "grad_norm": 0.657092273235321,
      "learning_rate": 8.522900316270352e-05,
      "loss": 0.5323,
      "step": 24500
    },
    {
      "epoch": 1.7216844398376696,
      "eval_loss": 0.5741798281669617,
      "eval_runtime": 191.815,
      "eval_samples_per_second": 65.944,
      "eval_steps_per_second": 8.248,
      "step": 24500
    },
    {
      "epoch": 1.7287117232655786,
      "grad_norm": 0.5300177335739136,
      "learning_rate": 8.476045449221038e-05,
      "loss": 0.5396,
      "step": 24600
    },
    {
      "epoch": 1.7357390066934875,
      "grad_norm": 0.5802739262580872,
      "learning_rate": 8.429190582171723e-05,
      "loss": 0.5343,
      "step": 24700
    },
    {
      "epoch": 1.7427662901213963,
      "grad_norm": 0.5847595930099487,
      "learning_rate": 8.38233571512241e-05,
      "loss": 0.5337,
      "step": 24800
    },
    {
      "epoch": 1.7497935735493053,
      "grad_norm": 0.5918022394180298,
      "learning_rate": 8.335480848073095e-05,
      "loss": 0.5332,
      "step": 24900
    },
    {
      "epoch": 1.756820856977214,
      "grad_norm": 0.6495380997657776,
      "learning_rate": 8.288625981023778e-05,
      "loss": 0.5337,
      "step": 25000
    },
    {
      "epoch": 1.756820856977214,
      "eval_loss": 0.5734357833862305,
      "eval_runtime": 191.5001,
      "eval_samples_per_second": 66.052,
      "eval_steps_per_second": 8.261,
      "step": 25000
    },
    {
      "epoch": 1.763848140405123,
      "grad_norm": 0.773017406463623,
      "learning_rate": 8.241771113974465e-05,
      "loss": 0.5195,
      "step": 25100
    },
    {
      "epoch": 1.7708754238330318,
      "grad_norm": 0.6224300861358643,
      "learning_rate": 8.19491624692515e-05,
      "loss": 0.5409,
      "step": 25200
    },
    {
      "epoch": 1.7779027072609406,
      "grad_norm": 0.561305582523346,
      "learning_rate": 8.148061379875836e-05,
      "loss": 0.5411,
      "step": 25300
    },
    {
      "epoch": 1.7849299906888496,
      "grad_norm": 0.5424569249153137,
      "learning_rate": 8.10120651282652e-05,
      "loss": 0.5385,
      "step": 25400
    },
    {
      "epoch": 1.7919572741167582,
      "grad_norm": 0.6570726037025452,
      "learning_rate": 8.054351645777205e-05,
      "loss": 0.5304,
      "step": 25500
    },
    {
      "epoch": 1.7919572741167582,
      "eval_loss": 0.5725218057632446,
      "eval_runtime": 191.5153,
      "eval_samples_per_second": 66.047,
      "eval_steps_per_second": 8.26,
      "step": 25500
    },
    {
      "epoch": 1.7989845575446672,
      "grad_norm": 0.619318962097168,
      "learning_rate": 8.007496778727891e-05,
      "loss": 0.5404,
      "step": 25600
    },
    {
      "epoch": 1.806011840972576,
      "grad_norm": 0.666038990020752,
      "learning_rate": 7.960641911678576e-05,
      "loss": 0.5367,
      "step": 25700
    },
    {
      "epoch": 1.8130391244004849,
      "grad_norm": 0.4837888777256012,
      "learning_rate": 7.913787044629261e-05,
      "loss": 0.5275,
      "step": 25800
    },
    {
      "epoch": 1.8200664078283937,
      "grad_norm": 0.5173748135566711,
      "learning_rate": 7.866932177579946e-05,
      "loss": 0.5159,
      "step": 25900
    },
    {
      "epoch": 1.8270936912563025,
      "grad_norm": 0.5783088207244873,
      "learning_rate": 7.820077310530631e-05,
      "loss": 0.526,
      "step": 26000
    },
    {
      "epoch": 1.8270936912563025,
      "eval_loss": 0.571327805519104,
      "eval_runtime": 191.7206,
      "eval_samples_per_second": 65.976,
      "eval_steps_per_second": 8.252,
      "step": 26000
    },
    {
      "epoch": 1.8341209746842115,
      "grad_norm": 0.4488949477672577,
      "learning_rate": 7.773222443481318e-05,
      "loss": 0.5264,
      "step": 26100
    },
    {
      "epoch": 1.8411482581121204,
      "grad_norm": 0.5412874817848206,
      "learning_rate": 7.726367576432003e-05,
      "loss": 0.5366,
      "step": 26200
    },
    {
      "epoch": 1.8481755415400292,
      "grad_norm": 0.5360791087150574,
      "learning_rate": 7.679512709382688e-05,
      "loss": 0.5277,
      "step": 26300
    },
    {
      "epoch": 1.855202824967938,
      "grad_norm": 0.5853073596954346,
      "learning_rate": 7.632657842333373e-05,
      "loss": 0.5221,
      "step": 26400
    },
    {
      "epoch": 1.8622301083958468,
      "grad_norm": 0.733930230140686,
      "learning_rate": 7.585802975284058e-05,
      "loss": 0.53,
      "step": 26500
    },
    {
      "epoch": 1.8622301083958468,
      "eval_loss": 0.5721610188484192,
      "eval_runtime": 191.5535,
      "eval_samples_per_second": 66.034,
      "eval_steps_per_second": 8.259,
      "step": 26500
    },
    {
      "epoch": 1.8692573918237558,
      "grad_norm": 0.5376147031784058,
      "learning_rate": 7.538948108234744e-05,
      "loss": 0.5203,
      "step": 26600
    },
    {
      "epoch": 1.8762846752516644,
      "grad_norm": 0.6613253951072693,
      "learning_rate": 7.492093241185428e-05,
      "loss": 0.5316,
      "step": 26700
    },
    {
      "epoch": 1.8833119586795735,
      "grad_norm": 0.6095169186592102,
      "learning_rate": 7.445238374136114e-05,
      "loss": 0.5396,
      "step": 26800
    },
    {
      "epoch": 1.8903392421074823,
      "grad_norm": 0.5879153609275818,
      "learning_rate": 7.398383507086799e-05,
      "loss": 0.5328,
      "step": 26900
    },
    {
      "epoch": 1.897366525535391,
      "grad_norm": 0.4561489522457123,
      "learning_rate": 7.351528640037484e-05,
      "loss": 0.5334,
      "step": 27000
    },
    {
      "epoch": 1.897366525535391,
      "eval_loss": 0.5708492398262024,
      "eval_runtime": 191.398,
      "eval_samples_per_second": 66.087,
      "eval_steps_per_second": 8.265,
      "step": 27000
    },
    {
      "epoch": 1.9043938089633001,
      "grad_norm": 1.0811113119125366,
      "learning_rate": 7.304673772988169e-05,
      "loss": 0.5299,
      "step": 27100
    },
    {
      "epoch": 1.9114210923912087,
      "grad_norm": 0.5830855965614319,
      "learning_rate": 7.257818905938854e-05,
      "loss": 0.5477,
      "step": 27200
    },
    {
      "epoch": 1.9184483758191178,
      "grad_norm": 0.6300702095031738,
      "learning_rate": 7.21096403888954e-05,
      "loss": 0.5481,
      "step": 27300
    },
    {
      "epoch": 1.9254756592470266,
      "grad_norm": 0.6268986463546753,
      "learning_rate": 7.164109171840226e-05,
      "loss": 0.5253,
      "step": 27400
    },
    {
      "epoch": 1.9325029426749354,
      "grad_norm": 0.6296914219856262,
      "learning_rate": 7.11725430479091e-05,
      "loss": 0.5233,
      "step": 27500
    },
    {
      "epoch": 1.9325029426749354,
      "eval_loss": 0.5700781941413879,
      "eval_runtime": 191.7014,
      "eval_samples_per_second": 65.983,
      "eval_steps_per_second": 8.252,
      "step": 27500
    },
    {
      "epoch": 1.9395302261028444,
      "grad_norm": 0.47624173760414124,
      "learning_rate": 7.070399437741596e-05,
      "loss": 0.54,
      "step": 27600
    },
    {
      "epoch": 1.946557509530753,
      "grad_norm": 0.6563848853111267,
      "learning_rate": 7.02354457069228e-05,
      "loss": 0.5273,
      "step": 27700
    },
    {
      "epoch": 1.953584792958662,
      "grad_norm": 0.6281931400299072,
      "learning_rate": 6.976689703642966e-05,
      "loss": 0.5265,
      "step": 27800
    },
    {
      "epoch": 1.9606120763865709,
      "grad_norm": 0.7378135323524475,
      "learning_rate": 6.929834836593652e-05,
      "loss": 0.5241,
      "step": 27900
    },
    {
      "epoch": 1.9676393598144797,
      "grad_norm": 0.5289490222930908,
      "learning_rate": 6.882979969544337e-05,
      "loss": 0.5327,
      "step": 28000
    },
    {
      "epoch": 1.9676393598144797,
      "eval_loss": 0.5690564513206482,
      "eval_runtime": 191.2308,
      "eval_samples_per_second": 66.145,
      "eval_steps_per_second": 8.273,
      "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": 5.0643079170298675e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}