{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 9.0,
  "eval_steps": 720,
  "global_step": 32391,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.13892747985551543,
      "grad_norm": 0.595678985118866,
      "learning_rate": 0.00039537834583680656,
      "loss": 0.376,
      "step": 500
    },
    {
      "epoch": 0.2000555709919422,
      "eval_loss": 0.30611157417297363,
      "eval_runtime": 21.4927,
      "eval_samples_per_second": 23.264,
      "eval_steps_per_second": 2.931,
      "step": 720
    },
    {
      "epoch": 0.27785495971103086,
      "grad_norm": 0.7439117431640625,
      "learning_rate": 0.0003907474298416227,
      "loss": 0.2956,
      "step": 1000
    },
    {
      "epoch": 0.4001111419838844,
      "eval_loss": 0.271937757730484,
      "eval_runtime": 21.3424,
      "eval_samples_per_second": 23.428,
      "eval_steps_per_second": 2.952,
      "step": 1440
    },
    {
      "epoch": 0.41678243956654626,
      "grad_norm": 0.5364288091659546,
      "learning_rate": 0.00038611651384643886,
      "loss": 0.2658,
      "step": 1500
    },
    {
      "epoch": 0.5557099194220617,
      "grad_norm": 0.5661336779594421,
      "learning_rate": 0.000381485597851255,
      "loss": 0.2461,
      "step": 2000
    },
    {
      "epoch": 0.6001667129758266,
      "eval_loss": 0.2515988349914551,
      "eval_runtime": 21.3216,
      "eval_samples_per_second": 23.45,
      "eval_steps_per_second": 2.955,
      "step": 2160
    },
    {
      "epoch": 0.6946373992775771,
      "grad_norm": 0.4893731474876404,
      "learning_rate": 0.00037685468185607115,
      "loss": 0.2307,
      "step": 2500
    },
    {
      "epoch": 0.8002222839677688,
      "eval_loss": 0.22781750559806824,
      "eval_runtime": 21.3004,
      "eval_samples_per_second": 23.474,
      "eval_steps_per_second": 2.958,
      "step": 2880
    },
    {
      "epoch": 0.8335648791330925,
      "grad_norm": 0.5245664119720459,
      "learning_rate": 0.0003722237658608873,
      "loss": 0.2205,
      "step": 3000
    },
    {
      "epoch": 0.972492358988608,
      "grad_norm": 0.7262536287307739,
      "learning_rate": 0.00036759284986570345,
      "loss": 0.2154,
      "step": 3500
    },
    {
      "epoch": 1.000277854959711,
      "eval_loss": 0.21747416257858276,
      "eval_runtime": 21.4198,
      "eval_samples_per_second": 23.343,
      "eval_steps_per_second": 2.941,
      "step": 3600
    },
    {
      "epoch": 1.1114198388441234,
      "grad_norm": 0.5650269389152527,
      "learning_rate": 0.0003629619338705196,
      "loss": 0.1815,
      "step": 4000
    },
    {
      "epoch": 1.2003334259516532,
      "eval_loss": 0.2159462422132492,
      "eval_runtime": 21.4142,
      "eval_samples_per_second": 23.349,
      "eval_steps_per_second": 2.942,
      "step": 4320
    },
    {
      "epoch": 1.2503473186996388,
      "grad_norm": 0.6974398493766785,
      "learning_rate": 0.00035833101787533574,
      "loss": 0.1813,
      "step": 4500
    },
    {
      "epoch": 1.3892747985551543,
      "grad_norm": 0.6206701993942261,
      "learning_rate": 0.0003537001018801519,
      "loss": 0.1785,
      "step": 5000
    },
    {
      "epoch": 1.4003889969435954,
      "eval_loss": 0.20631669461727142,
      "eval_runtime": 21.3451,
      "eval_samples_per_second": 23.425,
      "eval_steps_per_second": 2.952,
      "step": 5040
    },
    {
      "epoch": 1.5282022784106695,
      "grad_norm": 0.6988590359687805,
      "learning_rate": 0.00034906918588496804,
      "loss": 0.1748,
      "step": 5500
    },
    {
      "epoch": 1.6004445679355377,
      "eval_loss": 0.20530453324317932,
      "eval_runtime": 21.3207,
      "eval_samples_per_second": 23.451,
      "eval_steps_per_second": 2.955,
      "step": 5760
    },
    {
      "epoch": 1.667129758266185,
      "grad_norm": 0.5506817698478699,
      "learning_rate": 0.0003444382698897842,
      "loss": 0.1715,
      "step": 6000
    },
    {
      "epoch": 1.8005001389274797,
      "eval_loss": 0.19710968434810638,
      "eval_runtime": 21.2884,
      "eval_samples_per_second": 23.487,
      "eval_steps_per_second": 2.959,
      "step": 6480
    },
    {
      "epoch": 1.8060572381217006,
      "grad_norm": 0.464895099401474,
      "learning_rate": 0.0003398073538946004,
      "loss": 0.1678,
      "step": 6500
    },
    {
      "epoch": 1.9449847179772157,
      "grad_norm": 0.6984584331512451,
      "learning_rate": 0.0003351764378994165,
      "loss": 0.1664,
      "step": 7000
    },
    {
      "epoch": 2.000555709919422,
      "eval_loss": 0.18991900980472565,
      "eval_runtime": 21.3167,
      "eval_samples_per_second": 23.456,
      "eval_steps_per_second": 2.955,
      "step": 7200
    },
    {
      "epoch": 2.0839121978327313,
      "grad_norm": 0.6013078093528748,
      "learning_rate": 0.0003305455219042327,
      "loss": 0.1452,
      "step": 7500
    },
    {
      "epoch": 2.2006112809113643,
      "eval_loss": 0.1970534473657608,
      "eval_runtime": 21.6809,
      "eval_samples_per_second": 23.062,
      "eval_steps_per_second": 2.906,
      "step": 7920
    },
    {
      "epoch": 2.222839677688247,
      "grad_norm": 0.516161322593689,
      "learning_rate": 0.0003259146059090488,
      "loss": 0.1352,
      "step": 8000
    },
    {
      "epoch": 2.361767157543762,
      "grad_norm": 0.6993707418441772,
      "learning_rate": 0.000321283689913865,
      "loss": 0.1363,
      "step": 8500
    },
    {
      "epoch": 2.4006668519033063,
      "eval_loss": 0.18779733777046204,
      "eval_runtime": 21.3417,
      "eval_samples_per_second": 23.428,
      "eval_steps_per_second": 2.952,
      "step": 8640
    },
    {
      "epoch": 2.5006946373992776,
      "grad_norm": 0.542614758014679,
      "learning_rate": 0.00031665277391868115,
      "loss": 0.1345,
      "step": 9000
    },
    {
      "epoch": 2.600722422895249,
      "eval_loss": 0.18519769608974457,
      "eval_runtime": 21.5664,
      "eval_samples_per_second": 23.184,
      "eval_steps_per_second": 2.921,
      "step": 9360
    },
    {
      "epoch": 2.639622117254793,
      "grad_norm": 0.691017746925354,
      "learning_rate": 0.00031202185792349727,
      "loss": 0.1353,
      "step": 9500
    },
    {
      "epoch": 2.7785495971103087,
      "grad_norm": 0.4563291370868683,
      "learning_rate": 0.0003073909419283134,
      "loss": 0.1345,
      "step": 10000
    },
    {
      "epoch": 2.800777993887191,
      "eval_loss": 0.18035078048706055,
      "eval_runtime": 21.2975,
      "eval_samples_per_second": 23.477,
      "eval_steps_per_second": 2.958,
      "step": 10080
    },
    {
      "epoch": 2.917477076965824,
      "grad_norm": 0.6514193415641785,
      "learning_rate": 0.00030276002593312957,
      "loss": 0.1355,
      "step": 10500
    },
    {
      "epoch": 3.000833564879133,
      "eval_loss": 0.17824167013168335,
      "eval_runtime": 21.3358,
      "eval_samples_per_second": 23.435,
      "eval_steps_per_second": 2.953,
      "step": 10800
    },
    {
      "epoch": 3.0564045568213394,
      "grad_norm": 0.7063578367233276,
      "learning_rate": 0.00029812910993794574,
      "loss": 0.1222,
      "step": 11000
    },
    {
      "epoch": 3.1953320366768545,
      "grad_norm": 0.7858085632324219,
      "learning_rate": 0.0002934981939427619,
      "loss": 0.1072,
      "step": 11500
    },
    {
      "epoch": 3.2008891358710754,
      "eval_loss": 0.18112412095069885,
      "eval_runtime": 21.3193,
      "eval_samples_per_second": 23.453,
      "eval_steps_per_second": 2.955,
      "step": 11520
    },
    {
      "epoch": 3.33425951653237,
      "grad_norm": 0.962510883808136,
      "learning_rate": 0.00028886727794757804,
      "loss": 0.1086,
      "step": 12000
    },
    {
      "epoch": 3.4009447068630174,
      "eval_loss": 0.17723555862903595,
      "eval_runtime": 21.2949,
      "eval_samples_per_second": 23.48,
      "eval_steps_per_second": 2.958,
      "step": 12240
    },
    {
      "epoch": 3.4731869963878856,
      "grad_norm": 0.47604429721832275,
      "learning_rate": 0.00028423636195239416,
      "loss": 0.109,
      "step": 12500
    },
    {
      "epoch": 3.6010002778549595,
      "eval_loss": 0.1781689077615738,
      "eval_runtime": 21.2531,
      "eval_samples_per_second": 23.526,
      "eval_steps_per_second": 2.964,
      "step": 12960
    },
    {
      "epoch": 3.612114476243401,
      "grad_norm": 0.49760594964027405,
      "learning_rate": 0.00027960544595721033,
      "loss": 0.1099,
      "step": 13000
    },
    {
      "epoch": 3.7510419560989163,
      "grad_norm": 0.6786069869995117,
      "learning_rate": 0.0002749745299620265,
      "loss": 0.1084,
      "step": 13500
    },
    {
      "epoch": 3.801055848846902,
      "eval_loss": 0.17475004494190216,
      "eval_runtime": 21.2523,
      "eval_samples_per_second": 23.527,
      "eval_steps_per_second": 2.964,
      "step": 13680
    },
    {
      "epoch": 3.889969435954432,
      "grad_norm": 0.8729720115661621,
      "learning_rate": 0.0002703436139668427,
      "loss": 0.1114,
      "step": 14000
    },
    {
      "epoch": 4.001111419838844,
      "eval_loss": 0.16653864085674286,
      "eval_runtime": 21.3451,
      "eval_samples_per_second": 23.425,
      "eval_steps_per_second": 2.951,
      "step": 14400
    },
    {
      "epoch": 4.0288969158099475,
      "grad_norm": 0.7045803070068359,
      "learning_rate": 0.0002657126979716588,
      "loss": 0.1038,
      "step": 14500
    },
    {
      "epoch": 4.167824395665463,
      "grad_norm": 0.9094932079315186,
      "learning_rate": 0.0002610817819764749,
      "loss": 0.0839,
      "step": 15000
    },
    {
      "epoch": 4.201166990830786,
      "eval_loss": 0.1762482225894928,
      "eval_runtime": 21.3199,
      "eval_samples_per_second": 23.452,
      "eval_steps_per_second": 2.955,
      "step": 15120
    },
    {
      "epoch": 4.306751875520978,
      "grad_norm": 0.6102758646011353,
      "learning_rate": 0.0002564508659812911,
      "loss": 0.0878,
      "step": 15500
    },
    {
      "epoch": 4.4012225618227285,
      "eval_loss": 0.17415344715118408,
      "eval_runtime": 21.3432,
      "eval_samples_per_second": 23.427,
      "eval_steps_per_second": 2.952,
      "step": 15840
    },
    {
      "epoch": 4.445679355376494,
      "grad_norm": 0.46317267417907715,
      "learning_rate": 0.00025181994998610727,
      "loss": 0.0883,
      "step": 16000
    },
    {
      "epoch": 4.584606835232009,
      "grad_norm": 0.6532848477363586,
      "learning_rate": 0.00024718903399092345,
      "loss": 0.0887,
      "step": 16500
    },
    {
      "epoch": 4.601278132814671,
      "eval_loss": 0.18139471113681793,
      "eval_runtime": 21.5324,
      "eval_samples_per_second": 23.221,
      "eval_steps_per_second": 2.926,
      "step": 16560
    },
    {
      "epoch": 4.723534315087524,
      "grad_norm": 0.8044198155403137,
      "learning_rate": 0.00024255811799573954,
      "loss": 0.0928,
      "step": 17000
    },
    {
      "epoch": 4.801333703806613,
      "eval_loss": 0.16956336796283722,
      "eval_runtime": 22.0239,
      "eval_samples_per_second": 22.703,
      "eval_steps_per_second": 2.861,
      "step": 17280
    },
    {
      "epoch": 4.86246179494304,
      "grad_norm": 0.6072468161582947,
      "learning_rate": 0.00023792720200055571,
      "loss": 0.0915,
      "step": 17500
    },
    {
      "epoch": 5.001389274798555,
      "grad_norm": 0.7845910787582397,
      "learning_rate": 0.0002332962860053719,
      "loss": 0.0888,
      "step": 18000
    },
    {
      "epoch": 5.001389274798555,
      "eval_loss": 0.17226029932498932,
      "eval_runtime": 21.2768,
      "eval_samples_per_second": 23.5,
      "eval_steps_per_second": 2.961,
      "step": 18000
    },
    {
      "epoch": 5.14031675465407,
      "grad_norm": 0.4631412923336029,
      "learning_rate": 0.00022866537001018804,
      "loss": 0.067,
      "step": 18500
    },
    {
      "epoch": 5.201444845790498,
      "eval_loss": 0.18070575594902039,
      "eval_runtime": 21.433,
      "eval_samples_per_second": 23.328,
      "eval_steps_per_second": 2.939,
      "step": 18720
    },
    {
      "epoch": 5.279244234509586,
      "grad_norm": 0.5497326850891113,
      "learning_rate": 0.00022403445401500416,
      "loss": 0.0696,
      "step": 19000
    },
    {
      "epoch": 5.401500416782439,
      "eval_loss": 0.17943890392780304,
      "eval_runtime": 21.3084,
      "eval_samples_per_second": 23.465,
      "eval_steps_per_second": 2.957,
      "step": 19440
    },
    {
      "epoch": 5.418171714365101,
      "grad_norm": 0.455014705657959,
      "learning_rate": 0.00021940353801982033,
      "loss": 0.0717,
      "step": 19500
    },
    {
      "epoch": 5.5570991942206165,
      "grad_norm": 0.4856937527656555,
      "learning_rate": 0.00021477262202463648,
      "loss": 0.0724,
      "step": 20000
    },
    {
      "epoch": 5.601555987774382,
      "eval_loss": 0.1766962856054306,
      "eval_runtime": 21.2925,
      "eval_samples_per_second": 23.482,
      "eval_steps_per_second": 2.959,
      "step": 20160
    },
    {
      "epoch": 5.6960266740761325,
      "grad_norm": 0.9877897500991821,
      "learning_rate": 0.00021014170602945265,
      "loss": 0.0748,
      "step": 20500
    },
    {
      "epoch": 5.801611558766324,
      "eval_loss": 0.17774970829486847,
      "eval_runtime": 21.463,
      "eval_samples_per_second": 23.296,
      "eval_steps_per_second": 2.935,
      "step": 20880
    },
    {
      "epoch": 5.834954153931648,
      "grad_norm": 0.5059243440628052,
      "learning_rate": 0.0002055107900342688,
      "loss": 0.0732,
      "step": 21000
    },
    {
      "epoch": 5.973881633787163,
      "grad_norm": 0.6043376326560974,
      "learning_rate": 0.00020087987403908492,
      "loss": 0.0731,
      "step": 21500
    },
    {
      "epoch": 6.001667129758266,
      "eval_loss": 0.1752805858850479,
      "eval_runtime": 21.3362,
      "eval_samples_per_second": 23.434,
      "eval_steps_per_second": 2.953,
      "step": 21600
    },
    {
      "epoch": 6.112809113642679,
      "grad_norm": 0.49161791801452637,
      "learning_rate": 0.0001962489580439011,
      "loss": 0.0579,
      "step": 22000
    },
    {
      "epoch": 6.201722700750208,
      "eval_loss": 0.18562264740467072,
      "eval_runtime": 21.3241,
      "eval_samples_per_second": 23.448,
      "eval_steps_per_second": 2.954,
      "step": 22320
    },
    {
      "epoch": 6.251736593498194,
      "grad_norm": 0.5732259154319763,
      "learning_rate": 0.00019161804204871724,
      "loss": 0.0548,
      "step": 22500
    },
    {
      "epoch": 6.390664073353709,
      "grad_norm": 0.807499885559082,
      "learning_rate": 0.0001869871260535334,
      "loss": 0.0582,
      "step": 23000
    },
    {
      "epoch": 6.401778271742151,
      "eval_loss": 0.17600271105766296,
      "eval_runtime": 21.3547,
      "eval_samples_per_second": 23.414,
      "eval_steps_per_second": 2.95,
      "step": 23040
    },
    {
      "epoch": 6.529591553209225,
      "grad_norm": 0.5904980897903442,
      "learning_rate": 0.00018235621005834954,
      "loss": 0.0597,
      "step": 23500
    },
    {
      "epoch": 6.601833842734093,
      "eval_loss": 0.17754560708999634,
      "eval_runtime": 21.3447,
      "eval_samples_per_second": 23.425,
      "eval_steps_per_second": 2.952,
      "step": 23760
    },
    {
      "epoch": 6.66851903306474,
      "grad_norm": 0.721977174282074,
      "learning_rate": 0.0001777252940631657,
      "loss": 0.0591,
      "step": 24000
    },
    {
      "epoch": 6.801889413726035,
      "eval_loss": 0.18270494043827057,
      "eval_runtime": 21.3444,
      "eval_samples_per_second": 23.425,
      "eval_steps_per_second": 2.952,
      "step": 24480
    },
    {
      "epoch": 6.807446512920255,
      "grad_norm": 0.3316296339035034,
      "learning_rate": 0.00017309437806798186,
      "loss": 0.06,
      "step": 24500
    },
    {
      "epoch": 6.946373992775771,
      "grad_norm": 0.64435213804245,
      "learning_rate": 0.000168463462072798,
      "loss": 0.0591,
      "step": 25000
    },
    {
      "epoch": 7.001944984717977,
      "eval_loss": 0.17571093142032623,
      "eval_runtime": 21.3307,
      "eval_samples_per_second": 23.44,
      "eval_steps_per_second": 2.953,
      "step": 25200
    },
    {
      "epoch": 7.085301472631286,
      "grad_norm": 0.743349552154541,
      "learning_rate": 0.00016383254607761416,
      "loss": 0.0479,
      "step": 25500
    },
    {
      "epoch": 7.20200055570992,
      "eval_loss": 0.19375747442245483,
      "eval_runtime": 21.2935,
      "eval_samples_per_second": 23.481,
      "eval_steps_per_second": 2.959,
      "step": 25920
    },
    {
      "epoch": 7.2242289524868015,
      "grad_norm": 0.6206533908843994,
      "learning_rate": 0.00015920163008243033,
      "loss": 0.0431,
      "step": 26000
    },
    {
      "epoch": 7.3631564323423175,
      "grad_norm": 0.5083501935005188,
      "learning_rate": 0.00015457071408724645,
      "loss": 0.0457,
      "step": 26500
    },
    {
      "epoch": 7.402056126701861,
      "eval_loss": 0.18405260145664215,
      "eval_runtime": 21.2846,
      "eval_samples_per_second": 23.491,
      "eval_steps_per_second": 2.96,
      "step": 26640
    },
    {
      "epoch": 7.502083912197833,
      "grad_norm": 0.5129767060279846,
      "learning_rate": 0.00014993979809206263,
      "loss": 0.0463,
      "step": 27000
    },
    {
      "epoch": 7.602111697693804,
      "eval_loss": 0.18968620896339417,
      "eval_runtime": 21.3052,
      "eval_samples_per_second": 23.468,
      "eval_steps_per_second": 2.957,
      "step": 27360
    },
    {
      "epoch": 7.641011392053348,
      "grad_norm": 0.5119397044181824,
      "learning_rate": 0.00014530888209687877,
      "loss": 0.0463,
      "step": 27500
    },
    {
      "epoch": 7.779938871908864,
      "grad_norm": 0.6329432725906372,
      "learning_rate": 0.00014067796610169492,
      "loss": 0.0482,
      "step": 28000
    },
    {
      "epoch": 7.802167268685746,
      "eval_loss": 0.18267524242401123,
      "eval_runtime": 21.2984,
      "eval_samples_per_second": 23.476,
      "eval_steps_per_second": 2.958,
      "step": 28080
    },
    {
      "epoch": 7.918866351764379,
      "grad_norm": 0.9381836652755737,
      "learning_rate": 0.00013604705010651107,
      "loss": 0.0483,
      "step": 28500
    },
    {
      "epoch": 8.002222839677689,
      "eval_loss": 0.18136519193649292,
      "eval_runtime": 21.3111,
      "eval_samples_per_second": 23.462,
      "eval_steps_per_second": 2.956,
      "step": 28800
    },
    {
      "epoch": 8.057793831619895,
      "grad_norm": 0.8777905106544495,
      "learning_rate": 0.00013141613411132722,
      "loss": 0.0408,
      "step": 29000
    },
    {
      "epoch": 8.19672131147541,
      "grad_norm": 0.6021227240562439,
      "learning_rate": 0.0001267852181161434,
      "loss": 0.0353,
      "step": 29500
    },
    {
      "epoch": 8.20227841066963,
      "eval_loss": 0.19863204658031464,
      "eval_runtime": 21.3525,
      "eval_samples_per_second": 23.416,
      "eval_steps_per_second": 2.95,
      "step": 29520
    },
    {
      "epoch": 8.335648791330925,
      "grad_norm": 0.37129077315330505,
      "learning_rate": 0.00012215430212095954,
      "loss": 0.0346,
      "step": 30000
    },
    {
      "epoch": 8.402333981661572,
      "eval_loss": 0.2027880847454071,
      "eval_runtime": 21.3394,
      "eval_samples_per_second": 23.431,
      "eval_steps_per_second": 2.952,
      "step": 30240
    },
    {
      "epoch": 8.474576271186441,
      "grad_norm": 0.5072668194770813,
      "learning_rate": 0.00011752338612577569,
      "loss": 0.0358,
      "step": 30500
    },
    {
      "epoch": 8.602389552653515,
      "eval_loss": 0.1992003470659256,
      "eval_runtime": 21.3327,
      "eval_samples_per_second": 23.438,
      "eval_steps_per_second": 2.953,
      "step": 30960
    },
    {
      "epoch": 8.613503751041955,
      "grad_norm": 0.7495487332344055,
      "learning_rate": 0.00011289247013059183,
      "loss": 0.0369,
      "step": 31000
    },
    {
      "epoch": 8.752431230897471,
      "grad_norm": 0.618772566318512,
      "learning_rate": 0.000108261554135408,
      "loss": 0.0374,
      "step": 31500
    },
    {
      "epoch": 8.802445123645457,
      "eval_loss": 0.20048412680625916,
      "eval_runtime": 21.4267,
      "eval_samples_per_second": 23.335,
      "eval_steps_per_second": 2.94,
      "step": 31680
    },
    {
      "epoch": 8.891358710752987,
      "grad_norm": 0.839154064655304,
      "learning_rate": 0.00010363063814022413,
      "loss": 0.0373,
      "step": 32000
    }
  ],
  "logging_steps": 500,
  "max_steps": 43188,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 12,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.2282623257814761e+19,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}