{
  "best_global_step": 5161,
  "best_metric": 0.37820467,
  "best_model_checkpoint": "/home/work/newrag/qwen3/ms-swift-finetuning/output/qwen2.5-bnk-phase2/v1-20250804-040453/checkpoint-5161",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 5161,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00019377028532674515,
      "grad_norm": 0.8763577342033386,
      "learning_rate": 1.1583011583011583e-06,
      "loss": 0.9684686660766602,
      "memory(GiB)": 55.14,
      "step": 1,
      "token_acc": 0.7912246865959498,
      "train_speed(iter/s)": 0.010196
    },
    {
      "epoch": 0.0003875405706534903,
      "grad_norm": 0.848241925239563,
      "learning_rate": 2.3166023166023166e-06,
      "loss": 0.9568785429000854,
      "memory(GiB)": 55.14,
      "step": 2,
      "token_acc": 0.7863654591632134,
      "train_speed(iter/s)": 0.015869
    },
    {
      "epoch": 0.0005813108559802354,
      "grad_norm": 0.8443148136138916,
      "learning_rate": 3.4749034749034742e-06,
      "loss": 0.8965896964073181,
      "memory(GiB)": 66.69,
      "step": 3,
      "token_acc": 0.7995744899005435,
      "train_speed(iter/s)": 0.019428
    },
    {
      "epoch": 0.0007750811413069806,
      "grad_norm": 0.8933870196342468,
      "learning_rate": 4.633204633204633e-06,
      "loss": 0.9800578355789185,
      "memory(GiB)": 78.26,
      "step": 4,
      "token_acc": 0.7830852365415987,
      "train_speed(iter/s)": 0.021886
    },
    {
      "epoch": 0.0009688514266337257,
      "grad_norm": 0.9806432723999023,
      "learning_rate": 5.791505791505791e-06,
      "loss": 1.0483475923538208,
      "memory(GiB)": 78.26,
      "step": 5,
      "token_acc": 0.7753844505170658,
      "train_speed(iter/s)": 0.023671
    },
    {
      "epoch": 0.0011626217119604708,
      "grad_norm": 0.897155225276947,
      "learning_rate": 6.9498069498069484e-06,
      "loss": 1.0103771686553955,
      "memory(GiB)": 78.26,
      "step": 6,
      "token_acc": 0.7809635036496351,
      "train_speed(iter/s)": 0.025066
    },
    {
      "epoch": 0.0013563919972872161,
      "grad_norm": 0.9799181818962097,
      "learning_rate": 8.108108108108107e-06,
      "loss": 1.0178213119506836,
      "memory(GiB)": 78.26,
      "step": 7,
      "token_acc": 0.7772091722595078,
      "train_speed(iter/s)": 0.02613
    },
    {
      "epoch": 0.0015501622826139612,
      "grad_norm": 1.0471524000167847,
      "learning_rate": 9.266409266409266e-06,
      "loss": 1.1097385883331299,
      "memory(GiB)": 78.26,
      "step": 8,
      "token_acc": 0.7599808682440987,
      "train_speed(iter/s)": 0.026976
    },
    {
      "epoch": 0.0017439325679407063,
      "grad_norm": 0.8407416939735413,
      "learning_rate": 1.0424710424710423e-05,
      "loss": 0.9527825117111206,
      "memory(GiB)": 78.26,
      "step": 9,
      "token_acc": 0.7823414284532899,
      "train_speed(iter/s)": 0.027706
    },
    {
      "epoch": 0.0019377028532674514,
      "grad_norm": 0.8665539622306824,
      "learning_rate": 1.1583011583011582e-05,
      "loss": 1.0066951513290405,
      "memory(GiB)": 78.26,
      "step": 10,
      "token_acc": 0.7773281507906927,
      "train_speed(iter/s)": 0.028311
    },
    {
      "epoch": 0.0021314731385941965,
      "grad_norm": 0.8096230030059814,
      "learning_rate": 1.274131274131274e-05,
      "loss": 0.9802660346031189,
      "memory(GiB)": 78.26,
      "step": 11,
      "token_acc": 0.7735733563339918,
      "train_speed(iter/s)": 0.028814
    },
    {
      "epoch": 0.0023252434239209416,
      "grad_norm": 0.6567773818969727,
      "learning_rate": 1.3899613899613897e-05,
      "loss": 0.9128895998001099,
      "memory(GiB)": 78.26,
      "step": 12,
      "token_acc": 0.7855364418288818,
      "train_speed(iter/s)": 0.029273
    },
    {
      "epoch": 0.0025190137092476867,
      "grad_norm": 0.5237937569618225,
      "learning_rate": 1.5057915057915056e-05,
      "loss": 0.9058799743652344,
      "memory(GiB)": 78.26,
      "step": 13,
      "token_acc": 0.7805104547360853,
      "train_speed(iter/s)": 0.029674
    },
    {
      "epoch": 0.0027127839945744322,
      "grad_norm": 0.4156467914581299,
      "learning_rate": 1.6216216216216215e-05,
      "loss": 0.8571181893348694,
      "memory(GiB)": 78.26,
      "step": 14,
      "token_acc": 0.7879932696276072,
      "train_speed(iter/s)": 0.030038
    },
    {
      "epoch": 0.0029065542799011773,
      "grad_norm": 0.2674636244773865,
      "learning_rate": 1.7374517374517374e-05,
      "loss": 0.7994478344917297,
      "memory(GiB)": 78.26,
      "step": 15,
      "token_acc": 0.7901431703158408,
      "train_speed(iter/s)": 0.030338
    },
    {
      "epoch": 0.0031003245652279224,
      "grad_norm": 0.24086351692676544,
      "learning_rate": 1.8532818532818533e-05,
      "loss": 0.8486983180046082,
      "memory(GiB)": 78.26,
      "step": 16,
      "token_acc": 0.7774675145147912,
      "train_speed(iter/s)": 0.0306
    },
    {
      "epoch": 0.0032940948505546675,
      "grad_norm": 0.21650397777557373,
      "learning_rate": 1.9691119691119688e-05,
      "loss": 0.8528725504875183,
      "memory(GiB)": 78.26,
      "step": 17,
      "token_acc": 0.775322540195614,
      "train_speed(iter/s)": 0.030832
    },
    {
      "epoch": 0.0034878651358814126,
      "grad_norm": 0.15441368520259857,
      "learning_rate": 2.0849420849420847e-05,
      "loss": 0.803363025188446,
      "memory(GiB)": 78.26,
      "step": 18,
      "token_acc": 0.7852022395479044,
      "train_speed(iter/s)": 0.031049
    },
    {
      "epoch": 0.0036816354212081577,
      "grad_norm": 0.14429929852485657,
      "learning_rate": 2.200772200772201e-05,
      "loss": 0.8013445734977722,
      "memory(GiB)": 78.26,
      "step": 19,
      "token_acc": 0.7839805825242718,
      "train_speed(iter/s)": 0.031259
    },
    {
      "epoch": 0.003875405706534903,
      "grad_norm": 0.1348273605108261,
      "learning_rate": 2.3166023166023165e-05,
      "loss": 0.7239266633987427,
      "memory(GiB)": 78.26,
      "step": 20,
      "token_acc": 0.8037726258718549,
      "train_speed(iter/s)": 0.031439
    },
    {
      "epoch": 0.004069175991861648,
      "grad_norm": 0.1491738110780716,
      "learning_rate": 2.4324324324324324e-05,
      "loss": 0.8465297222137451,
      "memory(GiB)": 78.26,
      "step": 21,
      "token_acc": 0.7757449962935508,
      "train_speed(iter/s)": 0.031603
    },
    {
      "epoch": 0.004262946277188393,
      "grad_norm": 0.167159304022789,
      "learning_rate": 2.548262548262548e-05,
      "loss": 0.736566424369812,
      "memory(GiB)": 78.26,
      "step": 22,
      "token_acc": 0.8000630083629282,
      "train_speed(iter/s)": 0.03176
    },
    {
      "epoch": 0.004456716562515138,
      "grad_norm": 0.15523645281791687,
      "learning_rate": 2.6640926640926638e-05,
      "loss": 0.6989044547080994,
      "memory(GiB)": 78.26,
      "step": 23,
      "token_acc": 0.8113520593800029,
      "train_speed(iter/s)": 0.031902
    },
    {
      "epoch": 0.004650486847841883,
      "grad_norm": 0.16765183210372925,
      "learning_rate": 2.7799227799227794e-05,
      "loss": 0.6959363222122192,
      "memory(GiB)": 78.26,
      "step": 24,
      "token_acc": 0.8058972412047583,
      "train_speed(iter/s)": 0.03202
    },
    {
      "epoch": 0.004844257133168628,
      "grad_norm": 0.16556301712989807,
      "learning_rate": 2.8957528957528956e-05,
      "loss": 0.7282753586769104,
      "memory(GiB)": 78.26,
      "step": 25,
      "token_acc": 0.8006341844340851,
      "train_speed(iter/s)": 0.03212
    },
    {
      "epoch": 0.005038027418495373,
      "grad_norm": 0.15321362018585205,
      "learning_rate": 3.011583011583011e-05,
      "loss": 0.7563999891281128,
      "memory(GiB)": 78.26,
      "step": 26,
      "token_acc": 0.7961533964690751,
      "train_speed(iter/s)": 0.032232
    },
    {
      "epoch": 0.0052317977038221185,
      "grad_norm": 0.14468924701213837,
      "learning_rate": 3.1274131274131274e-05,
      "loss": 0.7149499654769897,
      "memory(GiB)": 78.26,
      "step": 27,
      "token_acc": 0.8046370478271425,
      "train_speed(iter/s)": 0.032331
    },
    {
      "epoch": 0.0054255679891488644,
      "grad_norm": 0.13254722952842712,
      "learning_rate": 3.243243243243243e-05,
      "loss": 0.7311335802078247,
      "memory(GiB)": 78.26,
      "step": 28,
      "token_acc": 0.8027618690788417,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.0056193382744756095,
      "grad_norm": 0.12081703543663025,
      "learning_rate": 3.3590733590733585e-05,
      "loss": 0.7195248603820801,
      "memory(GiB)": 78.26,
      "step": 29,
      "token_acc": 0.8041594893099605,
      "train_speed(iter/s)": 0.032499
    },
    {
      "epoch": 0.005813108559802355,
      "grad_norm": 0.11528925597667694,
      "learning_rate": 3.474903474903475e-05,
      "loss": 0.6729803681373596,
      "memory(GiB)": 78.26,
      "step": 30,
      "token_acc": 0.8157344552564334,
      "train_speed(iter/s)": 0.03258
    },
    {
      "epoch": 0.0060068788451291,
      "grad_norm": 0.13018707931041718,
      "learning_rate": 3.59073359073359e-05,
      "loss": 0.7402101159095764,
      "memory(GiB)": 78.26,
      "step": 31,
      "token_acc": 0.7976209655764996,
      "train_speed(iter/s)": 0.032663
    },
    {
      "epoch": 0.006200649130455845,
      "grad_norm": 0.12700200080871582,
      "learning_rate": 3.7065637065637065e-05,
      "loss": 0.7580655813217163,
      "memory(GiB)": 78.26,
      "step": 32,
      "token_acc": 0.7951273532668881,
      "train_speed(iter/s)": 0.032732
    },
    {
      "epoch": 0.00639441941578259,
      "grad_norm": 0.12156156450510025,
      "learning_rate": 3.822393822393822e-05,
      "loss": 0.6386986374855042,
      "memory(GiB)": 78.26,
      "step": 33,
      "token_acc": 0.8224443651076996,
      "train_speed(iter/s)": 0.032803
    },
    {
      "epoch": 0.006588189701109335,
      "grad_norm": 0.13328500092029572,
      "learning_rate": 3.9382239382239376e-05,
      "loss": 0.7081706523895264,
      "memory(GiB)": 78.26,
      "step": 34,
      "token_acc": 0.8066472273114339,
      "train_speed(iter/s)": 0.03287
    },
    {
      "epoch": 0.00678195998643608,
      "grad_norm": 0.12701700627803802,
      "learning_rate": 4.054054054054054e-05,
      "loss": 0.6880778074264526,
      "memory(GiB)": 78.26,
      "step": 35,
      "token_acc": 0.8144979985743269,
      "train_speed(iter/s)": 0.032933
    },
    {
      "epoch": 0.006975730271762825,
      "grad_norm": 0.13268497586250305,
      "learning_rate": 4.1698841698841694e-05,
      "loss": 0.7258732914924622,
      "memory(GiB)": 78.26,
      "step": 36,
      "token_acc": 0.8062436274215798,
      "train_speed(iter/s)": 0.033003
    },
    {
      "epoch": 0.00716950055708957,
      "grad_norm": 0.1224246695637703,
      "learning_rate": 4.285714285714285e-05,
      "loss": 0.7171946167945862,
      "memory(GiB)": 78.26,
      "step": 37,
      "token_acc": 0.8046310103255212,
      "train_speed(iter/s)": 0.033068
    },
    {
      "epoch": 0.007363270842416315,
      "grad_norm": 0.12160563468933105,
      "learning_rate": 4.401544401544402e-05,
      "loss": 0.7064481377601624,
      "memory(GiB)": 78.26,
      "step": 38,
      "token_acc": 0.8088185898806071,
      "train_speed(iter/s)": 0.033127
    },
    {
      "epoch": 0.0075570411277430605,
      "grad_norm": 0.11494564265012741,
      "learning_rate": 4.5173745173745174e-05,
      "loss": 0.7347713708877563,
      "memory(GiB)": 78.26,
      "step": 39,
      "token_acc": 0.798022857441745,
      "train_speed(iter/s)": 0.033179
    },
    {
      "epoch": 0.007750811413069806,
      "grad_norm": 0.11666765064001083,
      "learning_rate": 4.633204633204633e-05,
      "loss": 0.759428083896637,
      "memory(GiB)": 78.26,
      "step": 40,
      "token_acc": 0.7921686007998233,
      "train_speed(iter/s)": 0.033222
    },
    {
      "epoch": 0.007944581698396552,
      "grad_norm": 0.11888981610536575,
      "learning_rate": 4.7490347490347485e-05,
      "loss": 0.7375974655151367,
      "memory(GiB)": 78.26,
      "step": 41,
      "token_acc": 0.8020120724346076,
      "train_speed(iter/s)": 0.033273
    },
    {
      "epoch": 0.008138351983723296,
      "grad_norm": 0.11488386988639832,
      "learning_rate": 4.864864864864865e-05,
      "loss": 0.696284294128418,
      "memory(GiB)": 78.26,
      "step": 42,
      "token_acc": 0.8061018470051596,
      "train_speed(iter/s)": 0.033319
    },
    {
      "epoch": 0.008332122269050042,
      "grad_norm": 0.1082736924290657,
      "learning_rate": 4.98069498069498e-05,
      "loss": 0.6931319832801819,
      "memory(GiB)": 78.26,
      "step": 43,
      "token_acc": 0.8120705042391789,
      "train_speed(iter/s)": 0.033355
    },
    {
      "epoch": 0.008525892554376786,
      "grad_norm": 0.12314610928297043,
      "learning_rate": 5.096525096525096e-05,
      "loss": 0.7261099815368652,
      "memory(GiB)": 78.26,
      "step": 44,
      "token_acc": 0.8016441410059914,
      "train_speed(iter/s)": 0.033397
    },
    {
      "epoch": 0.008719662839703532,
      "grad_norm": 0.12343846261501312,
      "learning_rate": 5.212355212355212e-05,
      "loss": 0.6880357265472412,
      "memory(GiB)": 78.26,
      "step": 45,
      "token_acc": 0.8112455396966993,
      "train_speed(iter/s)": 0.033441
    },
    {
      "epoch": 0.008913433125030276,
      "grad_norm": 0.11987273395061493,
      "learning_rate": 5.3281853281853276e-05,
      "loss": 0.6877344250679016,
      "memory(GiB)": 78.26,
      "step": 46,
      "token_acc": 0.8103002813343285,
      "train_speed(iter/s)": 0.033486
    },
    {
      "epoch": 0.009107203410357022,
      "grad_norm": 0.1239103302359581,
      "learning_rate": 5.444015444015443e-05,
      "loss": 0.7504675388336182,
      "memory(GiB)": 78.26,
      "step": 47,
      "token_acc": 0.7963265423078025,
      "train_speed(iter/s)": 0.033526
    },
    {
      "epoch": 0.009300973695683766,
      "grad_norm": 0.11254343390464783,
      "learning_rate": 5.559845559845559e-05,
      "loss": 0.652977466583252,
      "memory(GiB)": 78.26,
      "step": 48,
      "token_acc": 0.8192933347417345,
      "train_speed(iter/s)": 0.033559
    },
    {
      "epoch": 0.009494743981010512,
      "grad_norm": 0.13146735727787018,
      "learning_rate": 5.6756756756756757e-05,
      "loss": 0.7184028029441833,
      "memory(GiB)": 78.26,
      "step": 49,
      "token_acc": 0.8021164329650468,
      "train_speed(iter/s)": 0.033598
    },
    {
      "epoch": 0.009688514266337257,
      "grad_norm": 0.11965050548315048,
      "learning_rate": 5.791505791505791e-05,
      "loss": 0.6229060888290405,
      "memory(GiB)": 78.26,
      "step": 50,
      "token_acc": 0.8271026669398988,
      "train_speed(iter/s)": 0.03363
    },
    {
      "epoch": 0.009882284551664003,
      "grad_norm": 0.12212494015693665,
      "learning_rate": 5.907335907335907e-05,
      "loss": 0.698147177696228,
      "memory(GiB)": 78.26,
      "step": 51,
      "token_acc": 0.8134602899805237,
      "train_speed(iter/s)": 0.033662
    },
    {
      "epoch": 0.010076054836990747,
      "grad_norm": 0.1269523799419403,
      "learning_rate": 6.023166023166022e-05,
      "loss": 0.6502314805984497,
      "memory(GiB)": 78.26,
      "step": 52,
      "token_acc": 0.8226499256566991,
      "train_speed(iter/s)": 0.033693
    },
    {
      "epoch": 0.010269825122317493,
      "grad_norm": 0.13087137043476105,
      "learning_rate": 6.138996138996139e-05,
      "loss": 0.6692061424255371,
      "memory(GiB)": 78.26,
      "step": 53,
      "token_acc": 0.8201304240156289,
      "train_speed(iter/s)": 0.033727
    },
    {
      "epoch": 0.010463595407644237,
      "grad_norm": 0.14609168469905853,
      "learning_rate": 6.254826254826255e-05,
      "loss": 0.7044385671615601,
      "memory(GiB)": 78.26,
      "step": 54,
      "token_acc": 0.8073001694132143,
      "train_speed(iter/s)": 0.033764
    },
    {
      "epoch": 0.010657365692970983,
      "grad_norm": 0.12461866438388824,
      "learning_rate": 6.37065637065637e-05,
      "loss": 0.6916706562042236,
      "memory(GiB)": 78.26,
      "step": 55,
      "token_acc": 0.8121318182938552,
      "train_speed(iter/s)": 0.033788
    },
    {
      "epoch": 0.010851135978297729,
      "grad_norm": 0.12224052846431732,
      "learning_rate": 6.486486486486486e-05,
      "loss": 0.6395901441574097,
      "memory(GiB)": 78.26,
      "step": 56,
      "token_acc": 0.824472191901174,
      "train_speed(iter/s)": 0.033815
    },
    {
      "epoch": 0.011044906263624473,
      "grad_norm": 0.1317582130432129,
      "learning_rate": 6.602316602316601e-05,
      "loss": 0.6729621291160583,
      "memory(GiB)": 78.26,
      "step": 57,
      "token_acc": 0.8167814800579487,
      "train_speed(iter/s)": 0.033843
    },
    {
      "epoch": 0.011238676548951219,
      "grad_norm": 0.12963414192199707,
      "learning_rate": 6.718146718146717e-05,
      "loss": 0.667634129524231,
      "memory(GiB)": 78.26,
      "step": 58,
      "token_acc": 0.8161623128430612,
      "train_speed(iter/s)": 0.033872
    },
    {
      "epoch": 0.011432446834277963,
      "grad_norm": 0.14076951146125793,
      "learning_rate": 6.833976833976833e-05,
      "loss": 0.659672200679779,
      "memory(GiB)": 78.26,
      "step": 59,
      "token_acc": 0.8184602051133901,
      "train_speed(iter/s)": 0.033893
    },
    {
      "epoch": 0.01162621711960471,
      "grad_norm": 0.1304904669523239,
      "learning_rate": 6.94980694980695e-05,
      "loss": 0.6900058388710022,
      "memory(GiB)": 78.26,
      "step": 60,
      "token_acc": 0.8142272123233802,
      "train_speed(iter/s)": 0.033915
    },
    {
      "epoch": 0.011819987404931454,
      "grad_norm": 0.12284844368696213,
      "learning_rate": 7.065637065637065e-05,
      "loss": 0.6068597435951233,
      "memory(GiB)": 78.26,
      "step": 61,
      "token_acc": 0.8333287809674688,
      "train_speed(iter/s)": 0.03394
    },
    {
      "epoch": 0.0120137576902582,
      "grad_norm": 0.16575318574905396,
      "learning_rate": 7.18146718146718e-05,
      "loss": 0.6391609311103821,
      "memory(GiB)": 78.26,
      "step": 62,
      "token_acc": 0.8243709681258777,
      "train_speed(iter/s)": 0.033959
    },
    {
      "epoch": 0.012207527975584944,
      "grad_norm": 0.1536901593208313,
      "learning_rate": 7.297297297297297e-05,
      "loss": 0.6720226407051086,
      "memory(GiB)": 78.26,
      "step": 63,
      "token_acc": 0.8132890573603584,
      "train_speed(iter/s)": 0.03398
    },
    {
      "epoch": 0.01240129826091169,
      "grad_norm": 0.12783144414424896,
      "learning_rate": 7.413127413127413e-05,
      "loss": 0.6299335956573486,
      "memory(GiB)": 78.26,
      "step": 64,
      "token_acc": 0.8247123041659774,
      "train_speed(iter/s)": 0.033999
    },
    {
      "epoch": 0.012595068546238434,
      "grad_norm": 0.12650437653064728,
      "learning_rate": 7.528957528957529e-05,
      "loss": 0.5913993120193481,
      "memory(GiB)": 78.26,
      "step": 65,
      "token_acc": 0.8381598287733052,
      "train_speed(iter/s)": 0.034015
    },
    {
      "epoch": 0.01278883883156518,
      "grad_norm": 0.13705813884735107,
      "learning_rate": 7.644787644787644e-05,
      "loss": 0.6303204298019409,
      "memory(GiB)": 78.26,
      "step": 66,
      "token_acc": 0.8255334138486312,
      "train_speed(iter/s)": 0.034034
    },
    {
      "epoch": 0.012982609116891924,
      "grad_norm": 0.15545400977134705,
      "learning_rate": 7.76061776061776e-05,
      "loss": 0.6292211413383484,
      "memory(GiB)": 78.26,
      "step": 67,
      "token_acc": 0.8286270117314154,
      "train_speed(iter/s)": 0.034055
    },
    {
      "epoch": 0.01317637940221867,
      "grad_norm": 0.14907345175743103,
      "learning_rate": 7.876447876447875e-05,
      "loss": 0.6275659799575806,
      "memory(GiB)": 78.26,
      "step": 68,
      "token_acc": 0.8254116669263935,
      "train_speed(iter/s)": 0.03407
    },
    {
      "epoch": 0.013370149687545414,
      "grad_norm": 0.15840691328048706,
      "learning_rate": 7.992277992277992e-05,
      "loss": 0.6730965375900269,
      "memory(GiB)": 78.26,
      "step": 69,
      "token_acc": 0.8142665820821879,
      "train_speed(iter/s)": 0.034091
    },
    {
      "epoch": 0.01356391997287216,
      "grad_norm": 0.14479568600654602,
      "learning_rate": 8.108108108108108e-05,
      "loss": 0.6207985877990723,
      "memory(GiB)": 78.26,
      "step": 70,
      "token_acc": 0.8265187594377976,
      "train_speed(iter/s)": 0.034111
    },
    {
      "epoch": 0.013757690258198904,
      "grad_norm": 0.13668787479400635,
      "learning_rate": 8.223938223938223e-05,
      "loss": 0.6429966688156128,
      "memory(GiB)": 78.26,
      "step": 71,
      "token_acc": 0.8240099449476115,
      "train_speed(iter/s)": 0.034127
    },
    {
      "epoch": 0.01395146054352565,
      "grad_norm": 0.14420422911643982,
      "learning_rate": 8.339768339768339e-05,
      "loss": 0.6110091805458069,
      "memory(GiB)": 78.26,
      "step": 72,
      "token_acc": 0.8311519082643192,
      "train_speed(iter/s)": 0.034142
    },
    {
      "epoch": 0.014145230828852395,
      "grad_norm": 0.16625623404979706,
      "learning_rate": 8.455598455598454e-05,
      "loss": 0.6166950464248657,
      "memory(GiB)": 78.26,
      "step": 73,
      "token_acc": 0.8307741268199539,
      "train_speed(iter/s)": 0.034155
    },
    {
      "epoch": 0.01433900111417914,
      "grad_norm": 0.14389079809188843,
      "learning_rate": 8.57142857142857e-05,
      "loss": 0.5988722443580627,
      "memory(GiB)": 78.26,
      "step": 74,
      "token_acc": 0.8348723657246945,
      "train_speed(iter/s)": 0.034166
    },
    {
      "epoch": 0.014532771399505887,
      "grad_norm": 0.18358756601810455,
      "learning_rate": 8.687258687258685e-05,
      "loss": 0.7113780379295349,
      "memory(GiB)": 78.26,
      "step": 75,
      "token_acc": 0.8083578854685225,
      "train_speed(iter/s)": 0.034184
    },
    {
      "epoch": 0.01472654168483263,
      "grad_norm": 0.1523071825504303,
      "learning_rate": 8.803088803088804e-05,
      "loss": 0.6291996240615845,
      "memory(GiB)": 78.26,
      "step": 76,
      "token_acc": 0.8278446372484995,
      "train_speed(iter/s)": 0.0342
    },
    {
      "epoch": 0.014920311970159377,
      "grad_norm": 0.16895808279514313,
      "learning_rate": 8.918918918918919e-05,
      "loss": 0.6497290730476379,
      "memory(GiB)": 78.26,
      "step": 77,
      "token_acc": 0.819365872911459,
      "train_speed(iter/s)": 0.034215
    },
    {
      "epoch": 0.015114082255486121,
      "grad_norm": 0.17319287359714508,
      "learning_rate": 9.034749034749035e-05,
      "loss": 0.5960345268249512,
      "memory(GiB)": 78.26,
      "step": 78,
      "token_acc": 0.8366524602867106,
      "train_speed(iter/s)": 0.034224
    },
    {
      "epoch": 0.015307852540812867,
      "grad_norm": 0.14983628690242767,
      "learning_rate": 9.15057915057915e-05,
      "loss": 0.6000748872756958,
      "memory(GiB)": 78.26,
      "step": 79,
      "token_acc": 0.8339163897148855,
      "train_speed(iter/s)": 0.034241
    },
    {
      "epoch": 0.015501622826139611,
      "grad_norm": 0.14565478265285492,
      "learning_rate": 9.266409266409266e-05,
      "loss": 0.6301745772361755,
      "memory(GiB)": 78.26,
      "step": 80,
      "token_acc": 0.8267408675799087,
      "train_speed(iter/s)": 0.034254
    },
    {
      "epoch": 0.015695393111466355,
      "grad_norm": 0.16253505647182465,
      "learning_rate": 9.382239382239381e-05,
      "loss": 0.6245601177215576,
      "memory(GiB)": 78.26,
      "step": 81,
      "token_acc": 0.8304850012464338,
      "train_speed(iter/s)": 0.034267
    },
    {
      "epoch": 0.015889163396793103,
      "grad_norm": 0.15641474723815918,
      "learning_rate": 9.498069498069497e-05,
      "loss": 0.6229044795036316,
      "memory(GiB)": 78.26,
      "step": 82,
      "token_acc": 0.8290063166096224,
      "train_speed(iter/s)": 0.034278
    },
    {
      "epoch": 0.016082933682119847,
      "grad_norm": 0.19741681218147278,
      "learning_rate": 9.613899613899614e-05,
      "loss": 0.5958088636398315,
      "memory(GiB)": 78.26,
      "step": 83,
      "token_acc": 0.8380999073922755,
      "train_speed(iter/s)": 0.034294
    },
    {
      "epoch": 0.01627670396744659,
      "grad_norm": 0.16509681940078735,
      "learning_rate": 9.72972972972973e-05,
      "loss": 0.652334988117218,
      "memory(GiB)": 78.26,
      "step": 84,
      "token_acc": 0.8223709473915191,
      "train_speed(iter/s)": 0.034307
    },
    {
      "epoch": 0.016470474252773336,
      "grad_norm": 0.17223341763019562,
      "learning_rate": 9.845559845559845e-05,
      "loss": 0.655316174030304,
      "memory(GiB)": 78.26,
      "step": 85,
      "token_acc": 0.82221954379727,
      "train_speed(iter/s)": 0.034319
    },
    {
      "epoch": 0.016664244538100084,
      "grad_norm": 0.1611337959766388,
      "learning_rate": 9.96138996138996e-05,
      "loss": 0.6308309435844421,
      "memory(GiB)": 78.26,
      "step": 86,
      "token_acc": 0.8256332148699723,
      "train_speed(iter/s)": 0.034331
    },
    {
      "epoch": 0.016858014823426828,
      "grad_norm": 0.1462010145187378,
      "learning_rate": 0.00010077220077220076,
      "loss": 0.5816379189491272,
      "memory(GiB)": 78.26,
      "step": 87,
      "token_acc": 0.8376213592233009,
      "train_speed(iter/s)": 0.03434
    },
    {
      "epoch": 0.017051785108753572,
      "grad_norm": 0.15553230047225952,
      "learning_rate": 0.00010193050193050192,
      "loss": 0.5978987216949463,
      "memory(GiB)": 78.26,
      "step": 88,
      "token_acc": 0.8393552427369511,
      "train_speed(iter/s)": 0.03435
    },
    {
      "epoch": 0.017245555394080316,
      "grad_norm": 0.172433540225029,
      "learning_rate": 0.00010308880308880307,
      "loss": 0.6419100165367126,
      "memory(GiB)": 78.26,
      "step": 89,
      "token_acc": 0.8243931496649293,
      "train_speed(iter/s)": 0.034363
    },
    {
      "epoch": 0.017439325679407064,
      "grad_norm": 0.17001327872276306,
      "learning_rate": 0.00010424710424710424,
      "loss": 0.6268438100814819,
      "memory(GiB)": 78.26,
      "step": 90,
      "token_acc": 0.8299691153761306,
      "train_speed(iter/s)": 0.034376
    },
    {
      "epoch": 0.017633095964733808,
      "grad_norm": 0.19090093672275543,
      "learning_rate": 0.0001054054054054054,
      "loss": 0.6875048875808716,
      "memory(GiB)": 78.26,
      "step": 91,
      "token_acc": 0.8123634272570442,
      "train_speed(iter/s)": 0.034391
    },
    {
      "epoch": 0.017826866250060552,
      "grad_norm": 0.16166290640830994,
      "learning_rate": 0.00010656370656370655,
      "loss": 0.5984062552452087,
      "memory(GiB)": 78.26,
      "step": 92,
      "token_acc": 0.8348241568976055,
      "train_speed(iter/s)": 0.034401
    },
    {
      "epoch": 0.0180206365353873,
      "grad_norm": 0.14463870227336884,
      "learning_rate": 0.00010772200772200771,
      "loss": 0.5481261014938354,
      "memory(GiB)": 78.26,
      "step": 93,
      "token_acc": 0.8472176412382793,
      "train_speed(iter/s)": 0.034409
    },
    {
      "epoch": 0.018214406820714044,
      "grad_norm": 0.17527909576892853,
      "learning_rate": 0.00010888030888030886,
      "loss": 0.6553927063941956,
      "memory(GiB)": 78.26,
      "step": 94,
      "token_acc": 0.8222487233587132,
      "train_speed(iter/s)": 0.034415
    },
    {
      "epoch": 0.01840817710604079,
      "grad_norm": 0.16232283413410187,
      "learning_rate": 0.00011003861003861002,
      "loss": 0.6176364421844482,
      "memory(GiB)": 78.26,
      "step": 95,
      "token_acc": 0.8296051451559177,
      "train_speed(iter/s)": 0.034422
    },
    {
      "epoch": 0.018601947391367533,
      "grad_norm": 0.1550573855638504,
      "learning_rate": 0.00011119691119691117,
      "loss": 0.5948160886764526,
      "memory(GiB)": 78.26,
      "step": 96,
      "token_acc": 0.8375553097345133,
      "train_speed(iter/s)": 0.034429
    },
    {
      "epoch": 0.01879571767669428,
      "grad_norm": 0.16793379187583923,
      "learning_rate": 0.00011235521235521234,
      "loss": 0.6064925193786621,
      "memory(GiB)": 78.26,
      "step": 97,
      "token_acc": 0.8337825430204662,
      "train_speed(iter/s)": 0.034438
    },
    {
      "epoch": 0.018989487962021025,
      "grad_norm": 0.17881280183792114,
      "learning_rate": 0.00011351351351351351,
      "loss": 0.5967330932617188,
      "memory(GiB)": 78.26,
      "step": 98,
      "token_acc": 0.829712168876606,
      "train_speed(iter/s)": 0.03445
    },
    {
      "epoch": 0.01918325824734777,
      "grad_norm": 0.1565878838300705,
      "learning_rate": 0.00011467181467181467,
      "loss": 0.5600182414054871,
      "memory(GiB)": 78.26,
      "step": 99,
      "token_acc": 0.8435590667538151,
      "train_speed(iter/s)": 0.034461
    },
    {
      "epoch": 0.019377028532674513,
      "grad_norm": 0.15280094742774963,
      "learning_rate": 0.00011583011583011582,
      "loss": 0.5975373983383179,
      "memory(GiB)": 78.26,
      "step": 100,
      "token_acc": 0.8333461057041408,
      "train_speed(iter/s)": 0.034468
    },
    {
      "epoch": 0.01957079881800126,
      "grad_norm": 0.15495073795318604,
      "learning_rate": 0.00011698841698841698,
      "loss": 0.6017282009124756,
      "memory(GiB)": 78.26,
      "step": 101,
      "token_acc": 0.8339026241596184,
      "train_speed(iter/s)": 0.034475
    },
    {
      "epoch": 0.019764569103328005,
      "grad_norm": 0.14694905281066895,
      "learning_rate": 0.00011814671814671814,
      "loss": 0.5602938532829285,
      "memory(GiB)": 78.26,
      "step": 102,
      "token_acc": 0.8411804083454201,
      "train_speed(iter/s)": 0.034482
    },
    {
      "epoch": 0.01995833938865475,
      "grad_norm": 0.1619928628206253,
      "learning_rate": 0.00011930501930501929,
      "loss": 0.5849668979644775,
      "memory(GiB)": 78.26,
      "step": 103,
      "token_acc": 0.8391262944887513,
      "train_speed(iter/s)": 0.03449
    },
    {
      "epoch": 0.020152109673981494,
      "grad_norm": 0.1454261839389801,
      "learning_rate": 0.00012046332046332045,
      "loss": 0.5901338458061218,
      "memory(GiB)": 78.26,
      "step": 104,
      "token_acc": 0.8340299917345614,
      "train_speed(iter/s)": 0.034498
    },
    {
      "epoch": 0.02034587995930824,
      "grad_norm": 0.2236863374710083,
      "learning_rate": 0.00012162162162162162,
      "loss": 0.5838300585746765,
      "memory(GiB)": 78.26,
      "step": 105,
      "token_acc": 0.8396284829721362,
      "train_speed(iter/s)": 0.034506
    },
    {
      "epoch": 0.020539650244634985,
      "grad_norm": 0.15272359549999237,
      "learning_rate": 0.00012277992277992278,
      "loss": 0.5919955372810364,
      "memory(GiB)": 78.26,
      "step": 106,
      "token_acc": 0.8308620948755198,
      "train_speed(iter/s)": 0.034513
    },
    {
      "epoch": 0.02073342052996173,
      "grad_norm": 0.13382184505462646,
      "learning_rate": 0.00012393822393822393,
      "loss": 0.5290847420692444,
      "memory(GiB)": 78.26,
      "step": 107,
      "token_acc": 0.8508332939898966,
      "train_speed(iter/s)": 0.034521
    },
    {
      "epoch": 0.020927190815288474,
      "grad_norm": 0.1610032171010971,
      "learning_rate": 0.0001250965250965251,
      "loss": 0.6414130330085754,
      "memory(GiB)": 78.26,
      "step": 108,
      "token_acc": 0.8247312177217657,
      "train_speed(iter/s)": 0.034529
    },
    {
      "epoch": 0.02112096110061522,
      "grad_norm": 0.1506713628768921,
      "learning_rate": 0.00012625482625482624,
      "loss": 0.6079325675964355,
      "memory(GiB)": 78.26,
      "step": 109,
      "token_acc": 0.8329596412556054,
      "train_speed(iter/s)": 0.034536
    },
    {
      "epoch": 0.021314731385941966,
      "grad_norm": 0.16280919313430786,
      "learning_rate": 0.0001274131274131274,
      "loss": 0.6136084198951721,
      "memory(GiB)": 78.26,
      "step": 110,
      "token_acc": 0.8316319235938349,
      "train_speed(iter/s)": 0.034545
    },
    {
      "epoch": 0.02150850167126871,
      "grad_norm": 0.14740246534347534,
      "learning_rate": 0.00012857142857142855,
      "loss": 0.5931621789932251,
      "memory(GiB)": 78.26,
      "step": 111,
      "token_acc": 0.8302635446262369,
      "train_speed(iter/s)": 0.034549
    },
    {
      "epoch": 0.021702271956595458,
      "grad_norm": 0.15233321487903595,
      "learning_rate": 0.00012972972972972972,
      "loss": 0.5877476930618286,
      "memory(GiB)": 78.26,
      "step": 112,
      "token_acc": 0.8364121451149842,
      "train_speed(iter/s)": 0.034559
    },
    {
      "epoch": 0.021896042241922202,
      "grad_norm": 0.1440098136663437,
      "learning_rate": 0.0001308880308880309,
      "loss": 0.553906261920929,
      "memory(GiB)": 78.26,
      "step": 113,
      "token_acc": 0.8466698357365446,
      "train_speed(iter/s)": 0.034565
    },
    {
      "epoch": 0.022089812527248946,
      "grad_norm": 0.16069145500659943,
      "learning_rate": 0.00013204633204633203,
      "loss": 0.6046357750892639,
      "memory(GiB)": 78.26,
      "step": 114,
      "token_acc": 0.832103537128133,
      "train_speed(iter/s)": 0.034573
    },
    {
      "epoch": 0.02228358281257569,
      "grad_norm": 0.1648361086845398,
      "learning_rate": 0.0001332046332046332,
      "loss": 0.6004931330680847,
      "memory(GiB)": 78.26,
      "step": 115,
      "token_acc": 0.8346290569636716,
      "train_speed(iter/s)": 0.034581
    },
    {
      "epoch": 0.022477353097902438,
      "grad_norm": 0.15239156782627106,
      "learning_rate": 0.00013436293436293434,
      "loss": 0.6015850901603699,
      "memory(GiB)": 78.26,
      "step": 116,
      "token_acc": 0.8332898444838743,
      "train_speed(iter/s)": 0.034589
    },
    {
      "epoch": 0.022671123383229182,
      "grad_norm": 0.15159296989440918,
      "learning_rate": 0.0001355212355212355,
      "loss": 0.5867636203765869,
      "memory(GiB)": 78.26,
      "step": 117,
      "token_acc": 0.8370697910212358,
      "train_speed(iter/s)": 0.034594
    },
    {
      "epoch": 0.022864893668555927,
      "grad_norm": 0.1747017502784729,
      "learning_rate": 0.00013667953667953665,
      "loss": 0.636696457862854,
      "memory(GiB)": 78.26,
      "step": 118,
      "token_acc": 0.8233567399562565,
      "train_speed(iter/s)": 0.034599
    },
    {
      "epoch": 0.02305866395388267,
      "grad_norm": 0.15044157207012177,
      "learning_rate": 0.00013783783783783782,
      "loss": 0.5734996199607849,
      "memory(GiB)": 78.26,
      "step": 119,
      "token_acc": 0.8417291220556745,
      "train_speed(iter/s)": 0.034603
    },
    {
      "epoch": 0.02325243423920942,
      "grad_norm": 0.16326741874217987,
      "learning_rate": 0.000138996138996139,
      "loss": 0.5967447757720947,
      "memory(GiB)": 78.26,
      "step": 120,
      "token_acc": 0.8323221786037829,
      "train_speed(iter/s)": 0.034609
    },
    {
      "epoch": 0.023446204524536163,
      "grad_norm": 0.13857780396938324,
      "learning_rate": 0.00014015444015444016,
      "loss": 0.5407213568687439,
      "memory(GiB)": 78.26,
      "step": 121,
      "token_acc": 0.8461235837180026,
      "train_speed(iter/s)": 0.034612
    },
    {
      "epoch": 0.023639974809862907,
      "grad_norm": 0.13604214787483215,
      "learning_rate": 0.0001413127413127413,
      "loss": 0.5105344653129578,
      "memory(GiB)": 78.26,
      "step": 122,
      "token_acc": 0.8567164915396379,
      "train_speed(iter/s)": 0.034616
    },
    {
      "epoch": 0.02383374509518965,
      "grad_norm": 0.14465397596359253,
      "learning_rate": 0.00014247104247104247,
      "loss": 0.5738057494163513,
      "memory(GiB)": 78.26,
      "step": 123,
      "token_acc": 0.8395190358188781,
      "train_speed(iter/s)": 0.034624
    },
    {
      "epoch": 0.0240275153805164,
      "grad_norm": 0.14596134424209595,
      "learning_rate": 0.0001436293436293436,
      "loss": 0.5468869209289551,
      "memory(GiB)": 78.26,
      "step": 124,
      "token_acc": 0.8466944373600839,
      "train_speed(iter/s)": 0.034632
    },
    {
      "epoch": 0.024221285665843143,
      "grad_norm": 0.14446629583835602,
      "learning_rate": 0.00014478764478764478,
      "loss": 0.5509823560714722,
      "memory(GiB)": 78.26,
      "step": 125,
      "token_acc": 0.8461442816999478,
      "train_speed(iter/s)": 0.034635
    },
    {
      "epoch": 0.024415055951169887,
      "grad_norm": 0.13912013173103333,
      "learning_rate": 0.00014594594594594595,
      "loss": 0.5554008483886719,
      "memory(GiB)": 78.26,
      "step": 126,
      "token_acc": 0.8426597276608645,
      "train_speed(iter/s)": 0.03464
    },
    {
      "epoch": 0.02460882623649663,
      "grad_norm": 0.14441823959350586,
      "learning_rate": 0.0001471042471042471,
      "loss": 0.5533645749092102,
      "memory(GiB)": 78.26,
      "step": 127,
      "token_acc": 0.8452828346917859,
      "train_speed(iter/s)": 0.034642
    },
    {
      "epoch": 0.02480259652182338,
      "grad_norm": 0.1425672024488449,
      "learning_rate": 0.00014826254826254826,
      "loss": 0.5448683500289917,
      "memory(GiB)": 78.26,
      "step": 128,
      "token_acc": 0.8480569849679332,
      "train_speed(iter/s)": 0.034646
    },
    {
      "epoch": 0.024996366807150124,
      "grad_norm": 0.15022799372673035,
      "learning_rate": 0.0001494208494208494,
      "loss": 0.5586391687393188,
      "memory(GiB)": 78.26,
      "step": 129,
      "token_acc": 0.843522056269538,
      "train_speed(iter/s)": 0.034653
    },
    {
      "epoch": 0.025190137092476868,
      "grad_norm": 0.14275088906288147,
      "learning_rate": 0.00015057915057915057,
      "loss": 0.5584322810173035,
      "memory(GiB)": 78.26,
      "step": 130,
      "token_acc": 0.8456992938407218,
      "train_speed(iter/s)": 0.034657
    },
    {
      "epoch": 0.025383907377803615,
      "grad_norm": 0.15598376095294952,
      "learning_rate": 0.0001517374517374517,
      "loss": 0.5684648156166077,
      "memory(GiB)": 78.26,
      "step": 131,
      "token_acc": 0.841918682337542,
      "train_speed(iter/s)": 0.03466
    },
    {
      "epoch": 0.02557767766313036,
      "grad_norm": 0.13116468489170074,
      "learning_rate": 0.00015289575289575288,
      "loss": 0.518610954284668,
      "memory(GiB)": 78.26,
      "step": 132,
      "token_acc": 0.8523458494613052,
      "train_speed(iter/s)": 0.034665
    },
    {
      "epoch": 0.025771447948457104,
      "grad_norm": 0.140156552195549,
      "learning_rate": 0.00015405405405405402,
      "loss": 0.5425719022750854,
      "memory(GiB)": 78.26,
      "step": 133,
      "token_acc": 0.8470962689771895,
      "train_speed(iter/s)": 0.03467
    },
    {
      "epoch": 0.025965218233783848,
      "grad_norm": 0.15615837275981903,
      "learning_rate": 0.0001552123552123552,
      "loss": 0.58672034740448,
      "memory(GiB)": 78.26,
      "step": 134,
      "token_acc": 0.8424782073907695,
      "train_speed(iter/s)": 0.034673
    },
    {
      "epoch": 0.026158988519110596,
      "grad_norm": 0.1514168381690979,
      "learning_rate": 0.00015637065637065634,
      "loss": 0.6196274161338806,
      "memory(GiB)": 78.26,
      "step": 135,
      "token_acc": 0.8292035111586176,
      "train_speed(iter/s)": 0.034681
    },
    {
      "epoch": 0.02635275880443734,
      "grad_norm": 0.15724033117294312,
      "learning_rate": 0.0001575289575289575,
      "loss": 0.5608164668083191,
      "memory(GiB)": 78.26,
      "step": 136,
      "token_acc": 0.8445253505933118,
      "train_speed(iter/s)": 0.034686
    },
    {
      "epoch": 0.026546529089764084,
      "grad_norm": 0.17307862639427185,
      "learning_rate": 0.0001586872586872587,
      "loss": 0.6565619111061096,
      "memory(GiB)": 78.26,
      "step": 137,
      "token_acc": 0.8214262132717068,
      "train_speed(iter/s)": 0.034691
    },
    {
      "epoch": 0.02674029937509083,
      "grad_norm": 0.1405402421951294,
      "learning_rate": 0.00015984555984555984,
      "loss": 0.5477691888809204,
      "memory(GiB)": 78.26,
      "step": 138,
      "token_acc": 0.8483550808604967,
      "train_speed(iter/s)": 0.034696
    },
    {
      "epoch": 0.026934069660417576,
      "grad_norm": 0.15360456705093384,
      "learning_rate": 0.000161003861003861,
      "loss": 0.6261464357376099,
      "memory(GiB)": 78.26,
      "step": 139,
      "token_acc": 0.8274440827470335,
      "train_speed(iter/s)": 0.0347
    },
    {
      "epoch": 0.02712783994574432,
      "grad_norm": 0.12648895382881165,
      "learning_rate": 0.00016216216216216215,
      "loss": 0.4741262197494507,
      "memory(GiB)": 78.26,
      "step": 140,
      "token_acc": 0.8678006997225238,
      "train_speed(iter/s)": 0.034702
    },
    {
      "epoch": 0.027321610231071065,
      "grad_norm": 0.1415882259607315,
      "learning_rate": 0.00016332046332046332,
      "loss": 0.5620299577713013,
      "memory(GiB)": 78.26,
      "step": 141,
      "token_acc": 0.8392420713649041,
      "train_speed(iter/s)": 0.034706
    },
    {
      "epoch": 0.02751538051639781,
      "grad_norm": 0.143955260515213,
      "learning_rate": 0.00016447876447876446,
      "loss": 0.5819345712661743,
      "memory(GiB)": 78.26,
      "step": 142,
      "token_acc": 0.8385089572540735,
      "train_speed(iter/s)": 0.03471
    },
    {
      "epoch": 0.027709150801724557,
      "grad_norm": 0.14960965514183044,
      "learning_rate": 0.00016563706563706563,
      "loss": 0.5373987555503845,
      "memory(GiB)": 78.26,
      "step": 143,
      "token_acc": 0.8491564847739486,
      "train_speed(iter/s)": 0.034715
    },
    {
      "epoch": 0.0279029210870513,
      "grad_norm": 0.13836169242858887,
      "learning_rate": 0.00016679536679536678,
      "loss": 0.5637241005897522,
      "memory(GiB)": 78.26,
      "step": 144,
      "token_acc": 0.8398485764610005,
      "train_speed(iter/s)": 0.034719
    },
    {
      "epoch": 0.028096691372378045,
      "grad_norm": 0.15270289778709412,
      "learning_rate": 0.00016795366795366795,
      "loss": 0.6161123514175415,
      "memory(GiB)": 78.26,
      "step": 145,
      "token_acc": 0.828073843783531,
      "train_speed(iter/s)": 0.034725
    },
    {
      "epoch": 0.02829046165770479,
      "grad_norm": 0.14659975469112396,
      "learning_rate": 0.0001691119691119691,
      "loss": 0.5577420592308044,
      "memory(GiB)": 78.26,
      "step": 146,
      "token_acc": 0.8441303927792653,
      "train_speed(iter/s)": 0.034731
    },
    {
      "epoch": 0.028484231943031537,
      "grad_norm": 0.13934315741062164,
      "learning_rate": 0.00017027027027027026,
      "loss": 0.5665932893753052,
      "memory(GiB)": 78.26,
      "step": 147,
      "token_acc": 0.8396104567059309,
      "train_speed(iter/s)": 0.034731
    },
    {
      "epoch": 0.02867800222835828,
      "grad_norm": 0.15382340550422668,
      "learning_rate": 0.0001714285714285714,
      "loss": 0.6315574645996094,
      "memory(GiB)": 78.26,
      "step": 148,
      "token_acc": 0.8281886687133687,
      "train_speed(iter/s)": 0.034735
    },
    {
      "epoch": 0.028871772513685025,
      "grad_norm": 0.15263354778289795,
      "learning_rate": 0.00017258687258687257,
      "loss": 0.5572026968002319,
      "memory(GiB)": 78.26,
      "step": 149,
      "token_acc": 0.8416983617968031,
      "train_speed(iter/s)": 0.034739
    },
    {
      "epoch": 0.029065542799011773,
      "grad_norm": 0.14366725087165833,
      "learning_rate": 0.0001737451737451737,
      "loss": 0.5455084443092346,
      "memory(GiB)": 78.26,
      "step": 150,
      "token_acc": 0.8478349198978408,
      "train_speed(iter/s)": 0.034744
    },
    {
      "epoch": 0.029259313084338517,
      "grad_norm": 0.14723645150661469,
      "learning_rate": 0.00017490347490347488,
      "loss": 0.5214795470237732,
      "memory(GiB)": 78.26,
      "step": 151,
      "token_acc": 0.8530382721575649,
      "train_speed(iter/s)": 0.034748
    },
    {
      "epoch": 0.02945308336966526,
      "grad_norm": 0.1369503289461136,
      "learning_rate": 0.00017606177606177607,
      "loss": 0.5099643468856812,
      "memory(GiB)": 78.26,
      "step": 152,
      "token_acc": 0.8562671739589939,
      "train_speed(iter/s)": 0.034751
    },
    {
      "epoch": 0.029646853654992006,
      "grad_norm": 0.16232426464557648,
      "learning_rate": 0.00017722007722007722,
      "loss": 0.6399052143096924,
      "memory(GiB)": 78.26,
      "step": 153,
      "token_acc": 0.8246472248353716,
      "train_speed(iter/s)": 0.034755
    },
    {
      "epoch": 0.029840623940318754,
      "grad_norm": 0.15433335304260254,
      "learning_rate": 0.00017837837837837839,
      "loss": 0.5725199580192566,
      "memory(GiB)": 78.26,
      "step": 154,
      "token_acc": 0.8409897602206277,
      "train_speed(iter/s)": 0.034758
    },
    {
      "epoch": 0.030034394225645498,
      "grad_norm": 0.13948160409927368,
      "learning_rate": 0.00017953667953667953,
      "loss": 0.5367989540100098,
      "memory(GiB)": 78.26,
      "step": 155,
      "token_acc": 0.8473491551719958,
      "train_speed(iter/s)": 0.034762
    },
    {
      "epoch": 0.030228164510972242,
      "grad_norm": 0.14237907528877258,
      "learning_rate": 0.0001806949806949807,
      "loss": 0.5689682364463806,
      "memory(GiB)": 78.26,
      "step": 156,
      "token_acc": 0.8421181996941228,
      "train_speed(iter/s)": 0.034765
    },
    {
      "epoch": 0.030421934796298986,
      "grad_norm": 0.13025622069835663,
      "learning_rate": 0.00018185328185328184,
      "loss": 0.5721440315246582,
      "memory(GiB)": 78.26,
      "step": 157,
      "token_acc": 0.8421661117277099,
      "train_speed(iter/s)": 0.034767
    },
    {
      "epoch": 0.030615705081625734,
      "grad_norm": 0.15070052444934845,
      "learning_rate": 0.000183011583011583,
      "loss": 0.5769027471542358,
      "memory(GiB)": 78.26,
      "step": 158,
      "token_acc": 0.8369094922737307,
      "train_speed(iter/s)": 0.034772
    },
    {
      "epoch": 0.030809475366952478,
      "grad_norm": 0.15805287659168243,
      "learning_rate": 0.00018416988416988415,
      "loss": 0.6211342215538025,
      "memory(GiB)": 78.26,
      "step": 159,
      "token_acc": 0.8280907315607631,
      "train_speed(iter/s)": 0.034775
    },
    {
      "epoch": 0.031003245652279222,
      "grad_norm": 0.1476234793663025,
      "learning_rate": 0.00018532818532818532,
      "loss": 0.5813598036766052,
      "memory(GiB)": 78.26,
      "step": 160,
      "token_acc": 0.8377501760032183,
      "train_speed(iter/s)": 0.034778
    },
    {
      "epoch": 0.031197015937605967,
      "grad_norm": 0.1391780972480774,
      "learning_rate": 0.00018648648648648646,
      "loss": 0.5335437059402466,
      "memory(GiB)": 78.26,
      "step": 161,
      "token_acc": 0.848728354978355,
      "train_speed(iter/s)": 0.034779
    },
    {
      "epoch": 0.03139078622293271,
      "grad_norm": 0.1483832597732544,
      "learning_rate": 0.00018764478764478763,
      "loss": 0.5789586305618286,
      "memory(GiB)": 78.26,
      "step": 162,
      "token_acc": 0.8385598141695703,
      "train_speed(iter/s)": 0.034781
    },
    {
      "epoch": 0.031584556508259455,
      "grad_norm": 0.14238472282886505,
      "learning_rate": 0.00018880308880308877,
      "loss": 0.5635126233100891,
      "memory(GiB)": 78.26,
      "step": 163,
      "token_acc": 0.8438286425038327,
      "train_speed(iter/s)": 0.034786
    },
    {
      "epoch": 0.031778326793586206,
      "grad_norm": 0.12377775460481644,
      "learning_rate": 0.00018996138996138994,
      "loss": 0.493597149848938,
      "memory(GiB)": 78.26,
      "step": 164,
      "token_acc": 0.8575653587393326,
      "train_speed(iter/s)": 0.034785
    },
    {
      "epoch": 0.03197209707891295,
      "grad_norm": 0.13860811293125153,
      "learning_rate": 0.00019111969111969108,
      "loss": 0.5463928580284119,
      "memory(GiB)": 78.26,
      "step": 165,
      "token_acc": 0.844146549588483,
      "train_speed(iter/s)": 0.034789
    },
    {
      "epoch": 0.032165867364239695,
      "grad_norm": 0.13424870371818542,
      "learning_rate": 0.00019227799227799228,
      "loss": 0.5533477067947388,
      "memory(GiB)": 78.26,
      "step": 166,
      "token_acc": 0.8481163938685373,
      "train_speed(iter/s)": 0.03479
    },
    {
      "epoch": 0.03235963764956644,
      "grad_norm": 0.12482903897762299,
      "learning_rate": 0.00019343629343629342,
      "loss": 0.5704807043075562,
      "memory(GiB)": 78.26,
      "step": 167,
      "token_acc": 0.8384424192212097,
      "train_speed(iter/s)": 0.034792
    },
    {
      "epoch": 0.03255340793489318,
      "grad_norm": 0.13756157457828522,
      "learning_rate": 0.0001945945945945946,
      "loss": 0.5534645318984985,
      "memory(GiB)": 78.26,
      "step": 168,
      "token_acc": 0.8454033863318016,
      "train_speed(iter/s)": 0.034794
    },
    {
      "epoch": 0.03274717822021993,
      "grad_norm": 0.13826580345630646,
      "learning_rate": 0.00019575289575289573,
      "loss": 0.534266471862793,
      "memory(GiB)": 78.26,
      "step": 169,
      "token_acc": 0.8489137497528039,
      "train_speed(iter/s)": 0.034796
    },
    {
      "epoch": 0.03294094850554667,
      "grad_norm": 0.14315061271190643,
      "learning_rate": 0.0001969111969111969,
      "loss": 0.5491966009140015,
      "memory(GiB)": 78.26,
      "step": 170,
      "token_acc": 0.8464807519310109,
      "train_speed(iter/s)": 0.034798
    },
    {
      "epoch": 0.03313471879087342,
      "grad_norm": 0.12808647751808167,
      "learning_rate": 0.00019806949806949804,
      "loss": 0.5022028684616089,
      "memory(GiB)": 78.26,
      "step": 171,
      "token_acc": 0.8551667255802993,
      "train_speed(iter/s)": 0.034798
    },
    {
      "epoch": 0.03332848907620017,
      "grad_norm": 0.14019440114498138,
      "learning_rate": 0.0001992277992277992,
      "loss": 0.5900301933288574,
      "memory(GiB)": 78.26,
      "step": 172,
      "token_acc": 0.8387483645878762,
      "train_speed(iter/s)": 0.034802
    },
    {
      "epoch": 0.03352225936152691,
      "grad_norm": 0.13789018988609314,
      "learning_rate": 0.00020038610038610038,
      "loss": 0.541344940662384,
      "memory(GiB)": 78.26,
      "step": 173,
      "token_acc": 0.8496686986482905,
      "train_speed(iter/s)": 0.034806
    },
    {
      "epoch": 0.033716029646853655,
      "grad_norm": 0.1350238174200058,
      "learning_rate": 0.00020154440154440152,
      "loss": 0.5568801164627075,
      "memory(GiB)": 78.26,
      "step": 174,
      "token_acc": 0.8458727972794469,
      "train_speed(iter/s)": 0.034809
    },
    {
      "epoch": 0.0339097999321804,
      "grad_norm": 0.14095835387706757,
      "learning_rate": 0.0002027027027027027,
      "loss": 0.5708956122398376,
      "memory(GiB)": 78.26,
      "step": 175,
      "token_acc": 0.845259973423054,
      "train_speed(iter/s)": 0.034811
    },
    {
      "epoch": 0.034103570217507144,
      "grad_norm": 0.13087715208530426,
      "learning_rate": 0.00020386100386100383,
      "loss": 0.5266304016113281,
      "memory(GiB)": 78.26,
      "step": 176,
      "token_acc": 0.8541874507744814,
      "train_speed(iter/s)": 0.034813
    },
    {
      "epoch": 0.03429734050283389,
      "grad_norm": 0.13467884063720703,
      "learning_rate": 0.000205019305019305,
      "loss": 0.4999612867832184,
      "memory(GiB)": 78.26,
      "step": 177,
      "token_acc": 0.859081674116213,
      "train_speed(iter/s)": 0.034816
    },
    {
      "epoch": 0.03449111078816063,
      "grad_norm": 0.1308935433626175,
      "learning_rate": 0.00020617760617760615,
      "loss": 0.5238428115844727,
      "memory(GiB)": 78.26,
      "step": 178,
      "token_acc": 0.8521208527878653,
      "train_speed(iter/s)": 0.034818
    },
    {
      "epoch": 0.034684881073487384,
      "grad_norm": 0.1268441081047058,
      "learning_rate": 0.00020733590733590731,
      "loss": 0.5368191599845886,
      "memory(GiB)": 78.26,
      "step": 179,
      "token_acc": 0.8482607365092318,
      "train_speed(iter/s)": 0.034819
    },
    {
      "epoch": 0.03487865135881413,
      "grad_norm": 0.12708643078804016,
      "learning_rate": 0.00020849420849420848,
      "loss": 0.49307650327682495,
      "memory(GiB)": 78.26,
      "step": 180,
      "token_acc": 0.8607868020304569,
      "train_speed(iter/s)": 0.034821
    },
    {
      "epoch": 0.03507242164414087,
      "grad_norm": 0.12540721893310547,
      "learning_rate": 0.00020965250965250965,
      "loss": 0.507483720779419,
      "memory(GiB)": 78.26,
      "step": 181,
      "token_acc": 0.8567486597287922,
      "train_speed(iter/s)": 0.034824
    },
    {
      "epoch": 0.035266191929467616,
      "grad_norm": 0.1270364373922348,
      "learning_rate": 0.0002108108108108108,
      "loss": 0.5022760629653931,
      "memory(GiB)": 78.26,
      "step": 182,
      "token_acc": 0.8590840060746543,
      "train_speed(iter/s)": 0.034825
    },
    {
      "epoch": 0.03545996221479436,
      "grad_norm": 0.14542265236377716,
      "learning_rate": 0.00021196911196911196,
      "loss": 0.516878604888916,
      "memory(GiB)": 78.26,
      "step": 183,
      "token_acc": 0.8556428363617128,
      "train_speed(iter/s)": 0.034829
    },
    {
      "epoch": 0.035653732500121105,
      "grad_norm": 0.14277629554271698,
      "learning_rate": 0.0002131274131274131,
      "loss": 0.5679965019226074,
      "memory(GiB)": 78.26,
      "step": 184,
      "token_acc": 0.841894944113048,
      "train_speed(iter/s)": 0.034833
    },
    {
      "epoch": 0.03584750278544785,
      "grad_norm": 0.13046088814735413,
      "learning_rate": 0.00021428571428571427,
      "loss": 0.5434874296188354,
      "memory(GiB)": 78.26,
      "step": 185,
      "token_acc": 0.8491733876369872,
      "train_speed(iter/s)": 0.034835
    },
    {
      "epoch": 0.0360412730707746,
      "grad_norm": 0.12494053691625595,
      "learning_rate": 0.00021544401544401542,
      "loss": 0.537643313407898,
      "memory(GiB)": 78.26,
      "step": 186,
      "token_acc": 0.8506308711310349,
      "train_speed(iter/s)": 0.034836
    },
    {
      "epoch": 0.036235043356101344,
      "grad_norm": 0.12971843779087067,
      "learning_rate": 0.00021660231660231659,
      "loss": 0.5533092617988586,
      "memory(GiB)": 78.26,
      "step": 187,
      "token_acc": 0.845800933125972,
      "train_speed(iter/s)": 0.034839
    },
    {
      "epoch": 0.03642881364142809,
      "grad_norm": 0.12924546003341675,
      "learning_rate": 0.00021776061776061773,
      "loss": 0.5168436765670776,
      "memory(GiB)": 78.26,
      "step": 188,
      "token_acc": 0.8556058890147226,
      "train_speed(iter/s)": 0.034839
    },
    {
      "epoch": 0.03662258392675483,
      "grad_norm": 0.1508742719888687,
      "learning_rate": 0.0002189189189189189,
      "loss": 0.6136130690574646,
      "memory(GiB)": 78.26,
      "step": 189,
      "token_acc": 0.8350558426404401,
      "train_speed(iter/s)": 0.034842
    },
    {
      "epoch": 0.03681635421208158,
      "grad_norm": 0.1334015280008316,
      "learning_rate": 0.00022007722007722004,
      "loss": 0.545418918132782,
      "memory(GiB)": 78.26,
      "step": 190,
      "token_acc": 0.8462031558185404,
      "train_speed(iter/s)": 0.034844
    },
    {
      "epoch": 0.03701012449740832,
      "grad_norm": 0.13890019059181213,
      "learning_rate": 0.0002212355212355212,
      "loss": 0.5522346496582031,
      "memory(GiB)": 78.26,
      "step": 191,
      "token_acc": 0.8448294130112312,
      "train_speed(iter/s)": 0.034845
    },
    {
      "epoch": 0.037203894782735066,
      "grad_norm": 0.13158449530601501,
      "learning_rate": 0.00022239382239382235,
      "loss": 0.5597431063652039,
      "memory(GiB)": 78.26,
      "step": 192,
      "token_acc": 0.8428118697781619,
      "train_speed(iter/s)": 0.034848
    },
    {
      "epoch": 0.03739766506806181,
      "grad_norm": 0.13762398064136505,
      "learning_rate": 0.00022355212355212352,
      "loss": 0.5332627892494202,
      "memory(GiB)": 78.26,
      "step": 193,
      "token_acc": 0.8513293253173013,
      "train_speed(iter/s)": 0.034851
    },
    {
      "epoch": 0.03759143535338856,
      "grad_norm": 0.13980168104171753,
      "learning_rate": 0.0002247104247104247,
      "loss": 0.5322573184967041,
      "memory(GiB)": 78.26,
      "step": 194,
      "token_acc": 0.85326330790953,
      "train_speed(iter/s)": 0.034851
    },
    {
      "epoch": 0.037785205638715305,
      "grad_norm": 0.12236251682043076,
      "learning_rate": 0.00022586872586872586,
      "loss": 0.5012400150299072,
      "memory(GiB)": 78.26,
      "step": 195,
      "token_acc": 0.8592677644111584,
      "train_speed(iter/s)": 0.034853
    },
    {
      "epoch": 0.03797897592404205,
      "grad_norm": 0.14247579872608185,
      "learning_rate": 0.00022702702702702703,
      "loss": 0.5467385053634644,
      "memory(GiB)": 78.26,
      "step": 196,
      "token_acc": 0.8513190123595822,
      "train_speed(iter/s)": 0.034856
    },
    {
      "epoch": 0.038172746209368794,
      "grad_norm": 0.12214695662260056,
      "learning_rate": 0.00022818532818532817,
      "loss": 0.4757518768310547,
      "memory(GiB)": 78.26,
      "step": 197,
      "token_acc": 0.8650738047495954,
      "train_speed(iter/s)": 0.034857
    },
    {
      "epoch": 0.03836651649469554,
      "grad_norm": 0.12631294131278992,
      "learning_rate": 0.00022934362934362934,
      "loss": 0.5603289604187012,
      "memory(GiB)": 78.26,
      "step": 198,
      "token_acc": 0.8452558741537236,
      "train_speed(iter/s)": 0.034857
    },
    {
      "epoch": 0.03856028678002228,
      "grad_norm": 0.12949825823307037,
      "learning_rate": 0.00023050193050193048,
      "loss": 0.5440013408660889,
      "memory(GiB)": 78.26,
      "step": 199,
      "token_acc": 0.8481728066281994,
      "train_speed(iter/s)": 0.03486
    },
    {
      "epoch": 0.038754057065349026,
      "grad_norm": 0.13039319217205048,
      "learning_rate": 0.00023166023166023165,
      "loss": 0.5734332799911499,
      "memory(GiB)": 78.26,
      "step": 200,
      "token_acc": 0.8401181070071356,
      "train_speed(iter/s)": 0.034862
    },
    {
      "epoch": 0.03894782735067577,
      "grad_norm": 0.13796895742416382,
      "learning_rate": 0.0002328185328185328,
      "loss": 0.5982975959777832,
      "memory(GiB)": 78.26,
      "step": 201,
      "token_acc": 0.832787772216962,
      "train_speed(iter/s)": 0.03479
    },
    {
      "epoch": 0.03914159763600252,
      "grad_norm": 0.12662553787231445,
      "learning_rate": 0.00023397683397683396,
      "loss": 0.5020790696144104,
      "memory(GiB)": 78.26,
      "step": 202,
      "token_acc": 0.858308341381589,
      "train_speed(iter/s)": 0.034793
    },
    {
      "epoch": 0.039335367921329266,
      "grad_norm": 0.13133689761161804,
      "learning_rate": 0.0002351351351351351,
      "loss": 0.48208650946617126,
      "memory(GiB)": 78.26,
      "step": 203,
      "token_acc": 0.8654059206966738,
      "train_speed(iter/s)": 0.034794
    },
    {
      "epoch": 0.03952913820665601,
      "grad_norm": 0.12844805419445038,
      "learning_rate": 0.00023629343629343627,
      "loss": 0.5480844974517822,
      "memory(GiB)": 78.26,
      "step": 204,
      "token_acc": 0.8480197137937178,
      "train_speed(iter/s)": 0.034797
    },
    {
      "epoch": 0.039722908491982754,
      "grad_norm": 0.13410721719264984,
      "learning_rate": 0.0002374517374517374,
      "loss": 0.524722695350647,
      "memory(GiB)": 78.26,
      "step": 205,
      "token_acc": 0.8535315555041361,
      "train_speed(iter/s)": 0.034801
    },
    {
      "epoch": 0.0399166787773095,
      "grad_norm": 0.15366047620773315,
      "learning_rate": 0.00023861003861003858,
      "loss": 0.593670129776001,
      "memory(GiB)": 78.26,
      "step": 206,
      "token_acc": 0.834993270524899,
      "train_speed(iter/s)": 0.034805
    },
    {
      "epoch": 0.04011044906263624,
      "grad_norm": 0.1396535038948059,
      "learning_rate": 0.00023976833976833972,
      "loss": 0.5373449325561523,
      "memory(GiB)": 78.26,
      "step": 207,
      "token_acc": 0.8462800580988756,
      "train_speed(iter/s)": 0.034806
    },
    {
      "epoch": 0.04030421934796299,
      "grad_norm": 0.14069020748138428,
      "learning_rate": 0.0002409266409266409,
      "loss": 0.583406925201416,
      "memory(GiB)": 78.26,
      "step": 208,
      "token_acc": 0.8375456332490874,
      "train_speed(iter/s)": 0.034809
    },
    {
      "epoch": 0.04049798963328974,
      "grad_norm": 0.13054059445858002,
      "learning_rate": 0.0002420849420849421,
      "loss": 0.5496135950088501,
      "memory(GiB)": 78.26,
      "step": 209,
      "token_acc": 0.843966505507651,
      "train_speed(iter/s)": 0.034811
    },
    {
      "epoch": 0.04069175991861648,
      "grad_norm": 0.14763560891151428,
      "learning_rate": 0.00024324324324324323,
      "loss": 0.5195255279541016,
      "memory(GiB)": 78.26,
      "step": 210,
      "token_acc": 0.8529995042141795,
      "train_speed(iter/s)": 0.034814
    },
    {
      "epoch": 0.04088553020394323,
      "grad_norm": 0.12177236378192902,
      "learning_rate": 0.0002444015444015444,
      "loss": 0.49852877855300903,
      "memory(GiB)": 78.26,
      "step": 211,
      "token_acc": 0.8585345707205675,
      "train_speed(iter/s)": 0.034815
    },
    {
      "epoch": 0.04107930048926997,
      "grad_norm": 0.1217300221323967,
      "learning_rate": 0.00024555984555984557,
      "loss": 0.48390355706214905,
      "memory(GiB)": 78.26,
      "step": 212,
      "token_acc": 0.8644299537231804,
      "train_speed(iter/s)": 0.034818
    },
    {
      "epoch": 0.041273070774596715,
      "grad_norm": 0.14275750517845154,
      "learning_rate": 0.0002467181467181467,
      "loss": 0.5250924825668335,
      "memory(GiB)": 78.26,
      "step": 213,
      "token_acc": 0.8536855100046294,
      "train_speed(iter/s)": 0.03482
    },
    {
      "epoch": 0.04146684105992346,
      "grad_norm": 0.1468067467212677,
      "learning_rate": 0.00024787644787644785,
      "loss": 0.5309884548187256,
      "memory(GiB)": 78.26,
      "step": 214,
      "token_acc": 0.8531648971912378,
      "train_speed(iter/s)": 0.034823
    },
    {
      "epoch": 0.041660611345250204,
      "grad_norm": 0.12400522828102112,
      "learning_rate": 0.000249034749034749,
      "loss": 0.5384760499000549,
      "memory(GiB)": 78.26,
      "step": 215,
      "token_acc": 0.8493547014607857,
      "train_speed(iter/s)": 0.034824
    },
    {
      "epoch": 0.04185438163057695,
      "grad_norm": 0.14342345297336578,
      "learning_rate": 0.0002501930501930502,
      "loss": 0.5606729388237,
      "memory(GiB)": 78.26,
      "step": 216,
      "token_acc": 0.8429284181681875,
      "train_speed(iter/s)": 0.034825
    },
    {
      "epoch": 0.0420481519159037,
      "grad_norm": 0.1313794106245041,
      "learning_rate": 0.0002513513513513513,
      "loss": 0.5184580087661743,
      "memory(GiB)": 78.26,
      "step": 217,
      "token_acc": 0.8552152612420286,
      "train_speed(iter/s)": 0.034827
    },
    {
      "epoch": 0.04224192220123044,
      "grad_norm": 0.13364368677139282,
      "learning_rate": 0.0002525096525096525,
      "loss": 0.5389662981033325,
      "memory(GiB)": 78.26,
      "step": 218,
      "token_acc": 0.8503248639797159,
      "train_speed(iter/s)": 0.034829
    },
    {
      "epoch": 0.04243569248655719,
      "grad_norm": 0.1303595006465912,
      "learning_rate": 0.00025366795366795364,
      "loss": 0.5534095764160156,
      "memory(GiB)": 78.26,
      "step": 219,
      "token_acc": 0.8464527027027027,
      "train_speed(iter/s)": 0.034832
    },
    {
      "epoch": 0.04262946277188393,
      "grad_norm": 0.13916410505771637,
      "learning_rate": 0.0002548262548262548,
      "loss": 0.5153782963752747,
      "memory(GiB)": 78.26,
      "step": 220,
      "token_acc": 0.8532017429948345,
      "train_speed(iter/s)": 0.034834
    },
    {
      "epoch": 0.042823233057210676,
      "grad_norm": 0.12516328692436218,
      "learning_rate": 0.00025598455598455593,
      "loss": 0.5454630851745605,
      "memory(GiB)": 78.26,
      "step": 221,
      "token_acc": 0.8455304060358847,
      "train_speed(iter/s)": 0.034837
    },
    {
      "epoch": 0.04301700334253742,
      "grad_norm": 0.12558779120445251,
      "learning_rate": 0.0002571428571428571,
      "loss": 0.5476839542388916,
      "memory(GiB)": 78.26,
      "step": 222,
      "token_acc": 0.8491302437385303,
      "train_speed(iter/s)": 0.034838
    },
    {
      "epoch": 0.043210773627864164,
      "grad_norm": 0.1301163285970688,
      "learning_rate": 0.00025830115830115827,
      "loss": 0.5624793171882629,
      "memory(GiB)": 78.26,
      "step": 223,
      "token_acc": 0.8417110837775045,
      "train_speed(iter/s)": 0.03484
    },
    {
      "epoch": 0.043404543913190916,
      "grad_norm": 0.1233832985162735,
      "learning_rate": 0.00025945945945945944,
      "loss": 0.5634196996688843,
      "memory(GiB)": 78.26,
      "step": 224,
      "token_acc": 0.8440744270023572,
      "train_speed(iter/s)": 0.034841
    },
    {
      "epoch": 0.04359831419851766,
      "grad_norm": 0.11998777091503143,
      "learning_rate": 0.0002606177606177606,
      "loss": 0.5352566838264465,
      "memory(GiB)": 78.26,
      "step": 225,
      "token_acc": 0.850525063369717,
      "train_speed(iter/s)": 0.034843
    },
    {
      "epoch": 0.043792084483844404,
      "grad_norm": 0.11935053765773773,
      "learning_rate": 0.0002617760617760618,
      "loss": 0.5280268788337708,
      "memory(GiB)": 78.26,
      "step": 226,
      "token_acc": 0.8505288461538462,
      "train_speed(iter/s)": 0.034845
    },
    {
      "epoch": 0.04398585476917115,
      "grad_norm": 0.11708512902259827,
      "learning_rate": 0.00026293436293436294,
      "loss": 0.5135525465011597,
      "memory(GiB)": 78.26,
      "step": 227,
      "token_acc": 0.8566448341432495,
      "train_speed(iter/s)": 0.034846
    },
    {
      "epoch": 0.04417962505449789,
      "grad_norm": 0.1259176880121231,
      "learning_rate": 0.00026409266409266406,
      "loss": 0.5575259327888489,
      "memory(GiB)": 78.26,
      "step": 228,
      "token_acc": 0.8424169123390531,
      "train_speed(iter/s)": 0.034847
    },
    {
      "epoch": 0.04437339533982464,
      "grad_norm": 0.12446990609169006,
      "learning_rate": 0.0002652509652509652,
      "loss": 0.5132482647895813,
      "memory(GiB)": 78.26,
      "step": 229,
      "token_acc": 0.8576641341938295,
      "train_speed(iter/s)": 0.034848
    },
    {
      "epoch": 0.04456716562515138,
      "grad_norm": 0.1376199871301651,
      "learning_rate": 0.0002664092664092664,
      "loss": 0.5118637681007385,
      "memory(GiB)": 78.26,
      "step": 230,
      "token_acc": 0.8575808249721293,
      "train_speed(iter/s)": 0.034849
    },
    {
      "epoch": 0.044760935910478125,
      "grad_norm": 0.13398443162441254,
      "learning_rate": 0.00026756756756756756,
      "loss": 0.543663501739502,
      "memory(GiB)": 78.26,
      "step": 231,
      "token_acc": 0.8458760878036109,
      "train_speed(iter/s)": 0.034849
    },
    {
      "epoch": 0.044954706195804876,
      "grad_norm": 0.13235385715961456,
      "learning_rate": 0.0002687258687258687,
      "loss": 0.5783892869949341,
      "memory(GiB)": 78.26,
      "step": 232,
      "token_acc": 0.8393900968051217,
      "train_speed(iter/s)": 0.034852
    },
    {
      "epoch": 0.04514847648113162,
      "grad_norm": 0.13555991649627686,
      "learning_rate": 0.00026988416988416985,
      "loss": 0.5653671026229858,
      "memory(GiB)": 78.26,
      "step": 233,
      "token_acc": 0.8414533928152148,
      "train_speed(iter/s)": 0.034855
    },
    {
      "epoch": 0.045342246766458365,
      "grad_norm": 0.1324978917837143,
      "learning_rate": 0.000271042471042471,
      "loss": 0.561633825302124,
      "memory(GiB)": 78.26,
      "step": 234,
      "token_acc": 0.845039593124316,
      "train_speed(iter/s)": 0.034858
    },
    {
      "epoch": 0.04553601705178511,
      "grad_norm": 0.1257573366165161,
      "learning_rate": 0.0002722007722007722,
      "loss": 0.5691174864768982,
      "memory(GiB)": 78.26,
      "step": 235,
      "token_acc": 0.8414384744097102,
      "train_speed(iter/s)": 0.03486
    },
    {
      "epoch": 0.04572978733711185,
      "grad_norm": 0.12331625819206238,
      "learning_rate": 0.0002733590733590733,
      "loss": 0.5002002120018005,
      "memory(GiB)": 78.26,
      "step": 236,
      "token_acc": 0.8585492089747352,
      "train_speed(iter/s)": 0.034862
    },
    {
      "epoch": 0.0459235576224386,
      "grad_norm": 0.12218355387449265,
      "learning_rate": 0.00027451737451737447,
      "loss": 0.48562729358673096,
      "memory(GiB)": 78.26,
      "step": 237,
      "token_acc": 0.8611403640740184,
      "train_speed(iter/s)": 0.034864
    },
    {
      "epoch": 0.04611732790776534,
      "grad_norm": 0.12444531172513962,
      "learning_rate": 0.00027567567567567564,
      "loss": 0.5203258991241455,
      "memory(GiB)": 78.26,
      "step": 238,
      "token_acc": 0.8543315991857046,
      "train_speed(iter/s)": 0.034866
    },
    {
      "epoch": 0.046311098193092086,
      "grad_norm": 0.12778066098690033,
      "learning_rate": 0.0002768339768339768,
      "loss": 0.5473844408988953,
      "memory(GiB)": 78.26,
      "step": 239,
      "token_acc": 0.8458585640138409,
      "train_speed(iter/s)": 0.034867
    },
    {
      "epoch": 0.04650486847841884,
      "grad_norm": 0.12521140277385712,
      "learning_rate": 0.000277992277992278,
      "loss": 0.5302430391311646,
      "memory(GiB)": 78.26,
      "step": 240,
      "token_acc": 0.8497068457705965,
      "train_speed(iter/s)": 0.034868
    },
    {
      "epoch": 0.04669863876374558,
      "grad_norm": 0.13087455928325653,
      "learning_rate": 0.00027915057915057915,
      "loss": 0.52344810962677,
      "memory(GiB)": 78.26,
      "step": 241,
      "token_acc": 0.8509464638253397,
      "train_speed(iter/s)": 0.034871
    },
    {
      "epoch": 0.046892409049072326,
      "grad_norm": 0.12300854921340942,
      "learning_rate": 0.0002803088803088803,
      "loss": 0.5085083842277527,
      "memory(GiB)": 78.26,
      "step": 242,
      "token_acc": 0.8563601071279393,
      "train_speed(iter/s)": 0.034873
    },
    {
      "epoch": 0.04708617933439907,
      "grad_norm": 0.1252821385860443,
      "learning_rate": 0.00028146718146718143,
      "loss": 0.5789982080459595,
      "memory(GiB)": 78.26,
      "step": 243,
      "token_acc": 0.8396179117080508,
      "train_speed(iter/s)": 0.034874
    },
    {
      "epoch": 0.047279949619725814,
      "grad_norm": 0.12332039326429367,
      "learning_rate": 0.0002826254826254826,
      "loss": 0.5306107997894287,
      "memory(GiB)": 78.26,
      "step": 244,
      "token_acc": 0.8511578885733525,
      "train_speed(iter/s)": 0.034877
    },
    {
      "epoch": 0.04747371990505256,
      "grad_norm": 0.11339928209781647,
      "learning_rate": 0.00028378378378378377,
      "loss": 0.533706784248352,
      "memory(GiB)": 78.26,
      "step": 245,
      "token_acc": 0.8514316174230886,
      "train_speed(iter/s)": 0.034877
    },
    {
      "epoch": 0.0476674901903793,
      "grad_norm": 0.12364498525857925,
      "learning_rate": 0.00028494208494208494,
      "loss": 0.5375621914863586,
      "memory(GiB)": 78.26,
      "step": 246,
      "token_acc": 0.8459392614747498,
      "train_speed(iter/s)": 0.03488
    },
    {
      "epoch": 0.047861260475706054,
      "grad_norm": 0.12653161585330963,
      "learning_rate": 0.00028610038610038605,
      "loss": 0.4905088245868683,
      "memory(GiB)": 78.26,
      "step": 247,
      "token_acc": 0.8620850743557136,
      "train_speed(iter/s)": 0.034882
    },
    {
      "epoch": 0.0480550307610328,
      "grad_norm": 0.1326380968093872,
      "learning_rate": 0.0002872586872586872,
      "loss": 0.5300474762916565,
      "memory(GiB)": 78.26,
      "step": 248,
      "token_acc": 0.8534217764115839,
      "train_speed(iter/s)": 0.034883
    },
    {
      "epoch": 0.04824880104635954,
      "grad_norm": 0.11917278915643692,
      "learning_rate": 0.0002884169884169884,
      "loss": 0.5086590647697449,
      "memory(GiB)": 78.26,
      "step": 249,
      "token_acc": 0.8568408610460755,
      "train_speed(iter/s)": 0.034884
    },
    {
      "epoch": 0.048442571331686286,
      "grad_norm": 0.12199165672063828,
      "learning_rate": 0.00028957528957528956,
      "loss": 0.5237119197845459,
      "memory(GiB)": 78.26,
      "step": 250,
      "token_acc": 0.8521546095586315,
      "train_speed(iter/s)": 0.034885
    },
    {
      "epoch": 0.04863634161701303,
      "grad_norm": 0.11789362877607346,
      "learning_rate": 0.0002907335907335907,
      "loss": 0.4931395649909973,
      "memory(GiB)": 78.26,
      "step": 251,
      "token_acc": 0.8595119082622757,
      "train_speed(iter/s)": 0.034887
    },
    {
      "epoch": 0.048830111902339775,
      "grad_norm": 0.11933058500289917,
      "learning_rate": 0.0002918918918918919,
      "loss": 0.5037907361984253,
      "memory(GiB)": 78.26,
      "step": 252,
      "token_acc": 0.8580700162252839,
      "train_speed(iter/s)": 0.034889
    },
    {
      "epoch": 0.04902388218766652,
      "grad_norm": 0.11169978976249695,
      "learning_rate": 0.000293050193050193,
      "loss": 0.4690629243850708,
      "memory(GiB)": 78.26,
      "step": 253,
      "token_acc": 0.8666485000123864,
      "train_speed(iter/s)": 0.03489
    },
    {
      "epoch": 0.04921765247299326,
      "grad_norm": 0.11621616035699844,
      "learning_rate": 0.0002942084942084942,
      "loss": 0.49911874532699585,
      "memory(GiB)": 78.26,
      "step": 254,
      "token_acc": 0.8591189560995347,
      "train_speed(iter/s)": 0.034891
    },
    {
      "epoch": 0.049411422758320014,
      "grad_norm": 0.1331455260515213,
      "learning_rate": 0.00029536679536679535,
      "loss": 0.5178690552711487,
      "memory(GiB)": 78.26,
      "step": 255,
      "token_acc": 0.8545419436705194,
      "train_speed(iter/s)": 0.034894
    },
    {
      "epoch": 0.04960519304364676,
      "grad_norm": 0.13746346533298492,
      "learning_rate": 0.0002965250965250965,
      "loss": 0.5907042026519775,
      "memory(GiB)": 78.26,
      "step": 256,
      "token_acc": 0.8399606359750285,
      "train_speed(iter/s)": 0.034896
    },
    {
      "epoch": 0.0497989633289735,
      "grad_norm": 0.13077257573604584,
      "learning_rate": 0.00029768339768339764,
      "loss": 0.5277935266494751,
      "memory(GiB)": 78.26,
      "step": 257,
      "token_acc": 0.8542142655941911,
      "train_speed(iter/s)": 0.034897
    },
    {
      "epoch": 0.04999273361430025,
      "grad_norm": 0.11512145400047302,
      "learning_rate": 0.0002988416988416988,
      "loss": 0.5082737803459167,
      "memory(GiB)": 78.26,
      "step": 258,
      "token_acc": 0.8569431737318288,
      "train_speed(iter/s)": 0.034898
    },
    {
      "epoch": 0.05018650389962699,
      "grad_norm": 0.12492549419403076,
      "learning_rate": 0.0003,
      "loss": 0.5232591032981873,
      "memory(GiB)": 78.26,
      "step": 259,
      "token_acc": 0.8529347048792563,
      "train_speed(iter/s)": 0.034899
    },
    {
      "epoch": 0.050380274184953736,
      "grad_norm": 0.1214490681886673,
      "learning_rate": 0.0002999999691954846,
      "loss": 0.538935124874115,
      "memory(GiB)": 78.26,
      "step": 260,
      "token_acc": 0.848226576457839,
      "train_speed(iter/s)": 0.034901
    },
    {
      "epoch": 0.05057404447028048,
      "grad_norm": 0.12162759155035019,
      "learning_rate": 0.0002999998767819513,
      "loss": 0.5226523876190186,
      "memory(GiB)": 78.26,
      "step": 261,
      "token_acc": 0.8527364343343139,
      "train_speed(iter/s)": 0.034904
    },
    {
      "epoch": 0.05076781475560723,
      "grad_norm": 0.14032401144504547,
      "learning_rate": 0.0002999997227594379,
      "loss": 0.5270801782608032,
      "memory(GiB)": 78.26,
      "step": 262,
      "token_acc": 0.8529345116700657,
      "train_speed(iter/s)": 0.034905
    },
    {
      "epoch": 0.050961585040933975,
      "grad_norm": 0.12755590677261353,
      "learning_rate": 0.00029999950712800773,
      "loss": 0.5621036887168884,
      "memory(GiB)": 78.26,
      "step": 263,
      "token_acc": 0.8452567559094614,
      "train_speed(iter/s)": 0.034908
    },
    {
      "epoch": 0.05115535532626072,
      "grad_norm": 0.12125645577907562,
      "learning_rate": 0.0002999992298877494,
      "loss": 0.48772531747817993,
      "memory(GiB)": 78.26,
      "step": 264,
      "token_acc": 0.860484942704546,
      "train_speed(iter/s)": 0.03491
    },
    {
      "epoch": 0.051349125611587464,
      "grad_norm": 0.14849497377872467,
      "learning_rate": 0.00029999889103877667,
      "loss": 0.5987675189971924,
      "memory(GiB)": 78.26,
      "step": 265,
      "token_acc": 0.8323770133690618,
      "train_speed(iter/s)": 0.034912
    },
    {
      "epoch": 0.05154289589691421,
      "grad_norm": 0.11173395067453384,
      "learning_rate": 0.00029999849058122874,
      "loss": 0.4839743971824646,
      "memory(GiB)": 78.26,
      "step": 266,
      "token_acc": 0.8618188850746991,
      "train_speed(iter/s)": 0.034913
    },
    {
      "epoch": 0.05173666618224095,
      "grad_norm": 0.13394448161125183,
      "learning_rate": 0.0002999980285152701,
      "loss": 0.5341227054595947,
      "memory(GiB)": 78.26,
      "step": 267,
      "token_acc": 0.8524495008126306,
      "train_speed(iter/s)": 0.034913
    },
    {
      "epoch": 0.051930436467567696,
      "grad_norm": 0.13987241685390472,
      "learning_rate": 0.0002999975048410906,
      "loss": 0.5620037913322449,
      "memory(GiB)": 78.26,
      "step": 268,
      "token_acc": 0.8440742478752803,
      "train_speed(iter/s)": 0.034913
    },
    {
      "epoch": 0.05212420675289444,
      "grad_norm": 0.12064801156520844,
      "learning_rate": 0.0002999969195589052,
      "loss": 0.5743050575256348,
      "memory(GiB)": 78.26,
      "step": 269,
      "token_acc": 0.8411526254595733,
      "train_speed(iter/s)": 0.034916
    },
    {
      "epoch": 0.05231797703822119,
      "grad_norm": 0.1192815750837326,
      "learning_rate": 0.00029999627266895444,
      "loss": 0.497215211391449,
      "memory(GiB)": 78.26,
      "step": 270,
      "token_acc": 0.8576876267748479,
      "train_speed(iter/s)": 0.034917
    },
    {
      "epoch": 0.052511747323547936,
      "grad_norm": 0.11295323818922043,
      "learning_rate": 0.0002999955641715039,
      "loss": 0.46060121059417725,
      "memory(GiB)": 78.26,
      "step": 271,
      "token_acc": 0.8695854936493567,
      "train_speed(iter/s)": 0.034918
    },
    {
      "epoch": 0.05270551760887468,
      "grad_norm": 0.11337928473949432,
      "learning_rate": 0.00029999479406684466,
      "loss": 0.47304582595825195,
      "memory(GiB)": 78.26,
      "step": 272,
      "token_acc": 0.8662467580585402,
      "train_speed(iter/s)": 0.034918
    },
    {
      "epoch": 0.052899287894201424,
      "grad_norm": 0.13473428785800934,
      "learning_rate": 0.000299993962355293,
      "loss": 0.5001351833343506,
      "memory(GiB)": 78.26,
      "step": 273,
      "token_acc": 0.8586831727649379,
      "train_speed(iter/s)": 0.03492
    },
    {
      "epoch": 0.05309305817952817,
      "grad_norm": 0.12432650476694107,
      "learning_rate": 0.00029999306903719043,
      "loss": 0.49999624490737915,
      "memory(GiB)": 78.26,
      "step": 274,
      "token_acc": 0.8596558122982282,
      "train_speed(iter/s)": 0.034921
    },
    {
      "epoch": 0.05328682846485491,
      "grad_norm": 0.12385527044534683,
      "learning_rate": 0.0002999921141129039,
      "loss": 0.5260789394378662,
      "memory(GiB)": 78.26,
      "step": 275,
      "token_acc": 0.8528883832638099,
      "train_speed(iter/s)": 0.034922
    },
    {
      "epoch": 0.05348059875018166,
      "grad_norm": 0.1163427084684372,
      "learning_rate": 0.00029999109758282577,
      "loss": 0.5076729655265808,
      "memory(GiB)": 78.26,
      "step": 276,
      "token_acc": 0.8572664593754225,
      "train_speed(iter/s)": 0.034923
    },
    {
      "epoch": 0.0536743690355084,
      "grad_norm": 0.10747300833463669,
      "learning_rate": 0.0002999900194473734,
      "loss": 0.4819478392601013,
      "memory(GiB)": 78.26,
      "step": 277,
      "token_acc": 0.8656782763309278,
      "train_speed(iter/s)": 0.034925
    },
    {
      "epoch": 0.05386813932083515,
      "grad_norm": 0.12642782926559448,
      "learning_rate": 0.00029998887970698966,
      "loss": 0.5688496232032776,
      "memory(GiB)": 78.26,
      "step": 278,
      "token_acc": 0.8424845950704225,
      "train_speed(iter/s)": 0.034926
    },
    {
      "epoch": 0.0540619096061619,
      "grad_norm": 0.11862245947122574,
      "learning_rate": 0.00029998767836214265,
      "loss": 0.5431630611419678,
      "memory(GiB)": 78.26,
      "step": 279,
      "token_acc": 0.8460705036731674,
      "train_speed(iter/s)": 0.034927
    },
    {
      "epoch": 0.05425567989148864,
      "grad_norm": 0.10874070972204208,
      "learning_rate": 0.00029998641541332583,
      "loss": 0.446528822183609,
      "memory(GiB)": 78.26,
      "step": 280,
      "token_acc": 0.875025387583779,
      "train_speed(iter/s)": 0.034929
    },
    {
      "epoch": 0.054449450176815385,
      "grad_norm": 0.1398598551750183,
      "learning_rate": 0.0002999850908610579,
      "loss": 0.5694330334663391,
      "memory(GiB)": 78.26,
      "step": 281,
      "token_acc": 0.841722914998284,
      "train_speed(iter/s)": 0.03493
    },
    {
      "epoch": 0.05464322046214213,
      "grad_norm": 0.11889787018299103,
      "learning_rate": 0.00029998370470588287,
      "loss": 0.4838942885398865,
      "memory(GiB)": 78.26,
      "step": 282,
      "token_acc": 0.8644304682040531,
      "train_speed(iter/s)": 0.034932
    },
    {
      "epoch": 0.054836990747468874,
      "grad_norm": 0.10652502626180649,
      "learning_rate": 0.00029998225694837015,
      "loss": 0.4832991063594818,
      "memory(GiB)": 78.26,
      "step": 283,
      "token_acc": 0.8605131303975627,
      "train_speed(iter/s)": 0.034933
    },
    {
      "epoch": 0.05503076103279562,
      "grad_norm": 0.1357640027999878,
      "learning_rate": 0.0002999807475891143,
      "loss": 0.5402747392654419,
      "memory(GiB)": 78.26,
      "step": 284,
      "token_acc": 0.8523422441967653,
      "train_speed(iter/s)": 0.034934
    },
    {
      "epoch": 0.05522453131812237,
      "grad_norm": 0.12577500939369202,
      "learning_rate": 0.00029997917662873526,
      "loss": 0.5225556492805481,
      "memory(GiB)": 78.26,
      "step": 285,
      "token_acc": 0.8524319637512936,
      "train_speed(iter/s)": 0.034935
    },
    {
      "epoch": 0.05541830160344911,
      "grad_norm": 0.11081529408693314,
      "learning_rate": 0.0002999775440678783,
      "loss": 0.47770658135414124,
      "memory(GiB)": 78.26,
      "step": 286,
      "token_acc": 0.8655869779677737,
      "train_speed(iter/s)": 0.034937
    },
    {
      "epoch": 0.05561207188877586,
      "grad_norm": 0.11565393954515457,
      "learning_rate": 0.00029997584990721396,
      "loss": 0.5191536545753479,
      "memory(GiB)": 78.26,
      "step": 287,
      "token_acc": 0.8520145631067961,
      "train_speed(iter/s)": 0.034937
    },
    {
      "epoch": 0.0558058421741026,
      "grad_norm": 0.1177186667919159,
      "learning_rate": 0.000299974094147438,
      "loss": 0.5068661570549011,
      "memory(GiB)": 78.26,
      "step": 288,
      "token_acc": 0.8555767793372644,
      "train_speed(iter/s)": 0.034939
    },
    {
      "epoch": 0.055999612459429346,
      "grad_norm": 0.11559902131557465,
      "learning_rate": 0.00029997227678927164,
      "loss": 0.5000442862510681,
      "memory(GiB)": 78.26,
      "step": 289,
      "token_acc": 0.8562957392033785,
      "train_speed(iter/s)": 0.03494
    },
    {
      "epoch": 0.05619338274475609,
      "grad_norm": 0.1195712685585022,
      "learning_rate": 0.0002999703978334613,
      "loss": 0.5059924125671387,
      "memory(GiB)": 78.26,
      "step": 290,
      "token_acc": 0.8566373608427013,
      "train_speed(iter/s)": 0.034941
    },
    {
      "epoch": 0.056387153030082834,
      "grad_norm": 0.11978676170110703,
      "learning_rate": 0.00029996845728077874,
      "loss": 0.5218163728713989,
      "memory(GiB)": 78.26,
      "step": 291,
      "token_acc": 0.8552988421821216,
      "train_speed(iter/s)": 0.034943
    },
    {
      "epoch": 0.05658092331540958,
      "grad_norm": 0.11734765022993088,
      "learning_rate": 0.00029996645513202086,
      "loss": 0.5239380598068237,
      "memory(GiB)": 78.26,
      "step": 292,
      "token_acc": 0.85505396631614,
      "train_speed(iter/s)": 0.034945
    },
    {
      "epoch": 0.05677469360073633,
      "grad_norm": 0.11253336071968079,
      "learning_rate": 0.0002999643913880102,
      "loss": 0.512554943561554,
      "memory(GiB)": 78.26,
      "step": 293,
      "token_acc": 0.8555791147627883,
      "train_speed(iter/s)": 0.034947
    },
    {
      "epoch": 0.056968463886063074,
      "grad_norm": 0.1775335669517517,
      "learning_rate": 0.0002999622660495943,
      "loss": 0.48465481400489807,
      "memory(GiB)": 78.26,
      "step": 294,
      "token_acc": 0.8632988755461122,
      "train_speed(iter/s)": 0.034948
    },
    {
      "epoch": 0.05716223417138982,
      "grad_norm": 0.11287941783666611,
      "learning_rate": 0.0002999600791176461,
      "loss": 0.5061824321746826,
      "memory(GiB)": 78.26,
      "step": 295,
      "token_acc": 0.8573063443244403,
      "train_speed(iter/s)": 0.034949
    },
    {
      "epoch": 0.05735600445671656,
      "grad_norm": 0.26354172825813293,
      "learning_rate": 0.00029995783059306373,
      "loss": 0.5662236213684082,
      "memory(GiB)": 78.26,
      "step": 296,
      "token_acc": 0.8466373350094281,
      "train_speed(iter/s)": 0.034951
    },
    {
      "epoch": 0.05754977474204331,
      "grad_norm": 0.4357684552669525,
      "learning_rate": 0.0002999555204767709,
      "loss": 0.5678244233131409,
      "memory(GiB)": 78.26,
      "step": 297,
      "token_acc": 0.8449365109151186,
      "train_speed(iter/s)": 0.034952
    },
    {
      "epoch": 0.05774354502737005,
      "grad_norm": 0.12695299088954926,
      "learning_rate": 0.00029995314876971627,
      "loss": 0.5329843759536743,
      "memory(GiB)": 78.26,
      "step": 298,
      "token_acc": 0.8508813455487176,
      "train_speed(iter/s)": 0.034953
    },
    {
      "epoch": 0.057937315312696795,
      "grad_norm": 0.13667218387126923,
      "learning_rate": 0.00029995071547287414,
      "loss": 0.5417138934135437,
      "memory(GiB)": 78.26,
      "step": 299,
      "token_acc": 0.8479577535288818,
      "train_speed(iter/s)": 0.034954
    },
    {
      "epoch": 0.058131085598023546,
      "grad_norm": 0.10648724436759949,
      "learning_rate": 0.00029994822058724375,
      "loss": 0.4887983202934265,
      "memory(GiB)": 78.26,
      "step": 300,
      "token_acc": 0.8605226687784597,
      "train_speed(iter/s)": 0.034954
    },
    {
      "epoch": 0.05832485588335029,
      "grad_norm": 0.11893422901630402,
      "learning_rate": 0.00029994566411384993,
      "loss": 0.505358099937439,
      "memory(GiB)": 78.26,
      "step": 301,
      "token_acc": 0.8587887578336663,
      "train_speed(iter/s)": 0.034955
    },
    {
      "epoch": 0.058518626168677035,
      "grad_norm": 0.12744036316871643,
      "learning_rate": 0.0002999430460537427,
      "loss": 0.5472061634063721,
      "memory(GiB)": 78.26,
      "step": 302,
      "token_acc": 0.8475548152800153,
      "train_speed(iter/s)": 0.034956
    },
    {
      "epoch": 0.05871239645400378,
      "grad_norm": 0.12299606949090958,
      "learning_rate": 0.00029994036640799726,
      "loss": 0.5060437917709351,
      "memory(GiB)": 78.26,
      "step": 303,
      "token_acc": 0.8581589526852902,
      "train_speed(iter/s)": 0.034956
    },
    {
      "epoch": 0.05890616673933052,
      "grad_norm": 0.11920775473117828,
      "learning_rate": 0.00029993762517771435,
      "loss": 0.5124378204345703,
      "memory(GiB)": 78.26,
      "step": 304,
      "token_acc": 0.853845315310405,
      "train_speed(iter/s)": 0.034957
    },
    {
      "epoch": 0.05909993702465727,
      "grad_norm": 0.11296442151069641,
      "learning_rate": 0.0002999348223640198,
      "loss": 0.5146512985229492,
      "memory(GiB)": 78.26,
      "step": 305,
      "token_acc": 0.856233997982776,
      "train_speed(iter/s)": 0.034958
    },
    {
      "epoch": 0.05929370730998401,
      "grad_norm": 0.11093028634786606,
      "learning_rate": 0.0002999319579680647,
      "loss": 0.485245943069458,
      "memory(GiB)": 78.26,
      "step": 306,
      "token_acc": 0.8640687244248123,
      "train_speed(iter/s)": 0.034959
    },
    {
      "epoch": 0.059487477595310756,
      "grad_norm": 0.11484235525131226,
      "learning_rate": 0.00029992903199102576,
      "loss": 0.4985674023628235,
      "memory(GiB)": 78.26,
      "step": 307,
      "token_acc": 0.8578893469527086,
      "train_speed(iter/s)": 0.03496
    },
    {
      "epoch": 0.05968124788063751,
      "grad_norm": 0.11024551838636398,
      "learning_rate": 0.00029992604443410456,
      "loss": 0.4915925860404968,
      "memory(GiB)": 78.26,
      "step": 308,
      "token_acc": 0.8619262990183482,
      "train_speed(iter/s)": 0.03496
    },
    {
      "epoch": 0.05987501816596425,
      "grad_norm": 0.11654435843229294,
      "learning_rate": 0.00029992299529852827,
      "loss": 0.5016142129898071,
      "memory(GiB)": 78.26,
      "step": 309,
      "token_acc": 0.8607442107550769,
      "train_speed(iter/s)": 0.03496
    },
    {
      "epoch": 0.060068788451290996,
      "grad_norm": 0.10757559537887573,
      "learning_rate": 0.0002999198845855492,
      "loss": 0.4506911039352417,
      "memory(GiB)": 78.26,
      "step": 310,
      "token_acc": 0.8732513679968862,
      "train_speed(iter/s)": 0.03496
    },
    {
      "epoch": 0.06026255873661774,
      "grad_norm": 0.11542216688394547,
      "learning_rate": 0.00029991671229644503,
      "loss": 0.47681623697280884,
      "memory(GiB)": 78.26,
      "step": 311,
      "token_acc": 0.8678860712584043,
      "train_speed(iter/s)": 0.034962
    },
    {
      "epoch": 0.060456329021944484,
      "grad_norm": 0.12022379785776138,
      "learning_rate": 0.0002999134784325187,
      "loss": 0.5177062153816223,
      "memory(GiB)": 78.26,
      "step": 312,
      "token_acc": 0.8545514413146945,
      "train_speed(iter/s)": 0.034964
    },
    {
      "epoch": 0.06065009930727123,
      "grad_norm": 0.11384547501802444,
      "learning_rate": 0.0002999101829950985,
      "loss": 0.5101380348205566,
      "memory(GiB)": 78.26,
      "step": 313,
      "token_acc": 0.8565856442943443,
      "train_speed(iter/s)": 0.034964
    },
    {
      "epoch": 0.06084386959259797,
      "grad_norm": 0.11187402158975601,
      "learning_rate": 0.0002999068259855378,
      "loss": 0.5241718888282776,
      "memory(GiB)": 78.26,
      "step": 314,
      "token_acc": 0.8541580041580041,
      "train_speed(iter/s)": 0.034966
    },
    {
      "epoch": 0.061037639877924724,
      "grad_norm": 0.10513240844011307,
      "learning_rate": 0.0002999034074052156,
      "loss": 0.4202113449573517,
      "memory(GiB)": 78.26,
      "step": 315,
      "token_acc": 0.8803593092589247,
      "train_speed(iter/s)": 0.034966
    },
    {
      "epoch": 0.06123141016325147,
      "grad_norm": 0.11992194503545761,
      "learning_rate": 0.0002998999272555359,
      "loss": 0.506824254989624,
      "memory(GiB)": 78.26,
      "step": 316,
      "token_acc": 0.8580272713024398,
      "train_speed(iter/s)": 0.034967
    },
    {
      "epoch": 0.06142518044857821,
      "grad_norm": 0.11586200445890427,
      "learning_rate": 0.0002998963855379281,
      "loss": 0.5369887351989746,
      "memory(GiB)": 78.26,
      "step": 317,
      "token_acc": 0.8491955856933918,
      "train_speed(iter/s)": 0.034968
    },
    {
      "epoch": 0.061618950733904956,
      "grad_norm": 0.11802522093057632,
      "learning_rate": 0.0002998927822538469,
      "loss": 0.5725922584533691,
      "memory(GiB)": 78.26,
      "step": 318,
      "token_acc": 0.8380715205103811,
      "train_speed(iter/s)": 0.03497
    },
    {
      "epoch": 0.0618127210192317,
      "grad_norm": 0.12824736535549164,
      "learning_rate": 0.0002998891174047722,
      "loss": 0.567888617515564,
      "memory(GiB)": 78.26,
      "step": 319,
      "token_acc": 0.841992673992674,
      "train_speed(iter/s)": 0.034972
    },
    {
      "epoch": 0.062006491304558445,
      "grad_norm": 0.10866405814886093,
      "learning_rate": 0.00029988539099220937,
      "loss": 0.4857517182826996,
      "memory(GiB)": 78.26,
      "step": 320,
      "token_acc": 0.8629794688168165,
      "train_speed(iter/s)": 0.034973
    },
    {
      "epoch": 0.06220026158988519,
      "grad_norm": 0.10862304270267487,
      "learning_rate": 0.00029988160301768884,
      "loss": 0.45832955837249756,
      "memory(GiB)": 78.26,
      "step": 321,
      "token_acc": 0.8700757980143056,
      "train_speed(iter/s)": 0.034974
    },
    {
      "epoch": 0.06239403187521193,
      "grad_norm": 0.11096280068159103,
      "learning_rate": 0.00029987775348276646,
      "loss": 0.5003219842910767,
      "memory(GiB)": 78.26,
      "step": 322,
      "token_acc": 0.8605376800379422,
      "train_speed(iter/s)": 0.034975
    },
    {
      "epoch": 0.06258780216053868,
      "grad_norm": 0.11027340590953827,
      "learning_rate": 0.0002998738423890234,
      "loss": 0.4482397139072418,
      "memory(GiB)": 78.26,
      "step": 323,
      "token_acc": 0.8740220554858348,
      "train_speed(iter/s)": 0.034975
    },
    {
      "epoch": 0.06278157244586542,
      "grad_norm": 0.11176367849111557,
      "learning_rate": 0.000299869869738066,
      "loss": 0.5309886932373047,
      "memory(GiB)": 78.26,
      "step": 324,
      "token_acc": 0.8513593064326234,
      "train_speed(iter/s)": 0.034976
    },
    {
      "epoch": 0.06297534273119217,
      "grad_norm": 0.10626234114170074,
      "learning_rate": 0.000299865835531526,
      "loss": 0.4504719376564026,
      "memory(GiB)": 78.26,
      "step": 325,
      "token_acc": 0.8722543040638998,
      "train_speed(iter/s)": 0.034977
    },
    {
      "epoch": 0.06316911301651891,
      "grad_norm": 0.10870497673749924,
      "learning_rate": 0.00029986173977106017,
      "loss": 0.5229367017745972,
      "memory(GiB)": 78.26,
      "step": 326,
      "token_acc": 0.8503923012467308,
      "train_speed(iter/s)": 0.034978
    },
    {
      "epoch": 0.06336288330184567,
      "grad_norm": 0.11313097178936005,
      "learning_rate": 0.0002998575824583509,
      "loss": 0.4801797866821289,
      "memory(GiB)": 78.26,
      "step": 327,
      "token_acc": 0.8636890035268726,
      "train_speed(iter/s)": 0.034979
    },
    {
      "epoch": 0.06355665358717241,
      "grad_norm": 0.11869515478610992,
      "learning_rate": 0.0002998533635951058,
      "loss": 0.5341205596923828,
      "memory(GiB)": 78.26,
      "step": 328,
      "token_acc": 0.8493568134624201,
      "train_speed(iter/s)": 0.03498
    },
    {
      "epoch": 0.06375042387249916,
      "grad_norm": 0.11751175671815872,
      "learning_rate": 0.00029984908318305743,
      "loss": 0.47566699981689453,
      "memory(GiB)": 78.26,
      "step": 329,
      "token_acc": 0.864566263149548,
      "train_speed(iter/s)": 0.034981
    },
    {
      "epoch": 0.0639441941578259,
      "grad_norm": 0.11967992782592773,
      "learning_rate": 0.000299844741223964,
      "loss": 0.4633430540561676,
      "memory(GiB)": 78.26,
      "step": 330,
      "token_acc": 0.8679728375820995,
      "train_speed(iter/s)": 0.034981
    },
    {
      "epoch": 0.06413796444315265,
      "grad_norm": 0.12265921384096146,
      "learning_rate": 0.00029984033771960895,
      "loss": 0.5029769539833069,
      "memory(GiB)": 78.26,
      "step": 331,
      "token_acc": 0.8578765113276207,
      "train_speed(iter/s)": 0.034982
    },
    {
      "epoch": 0.06433173472847939,
      "grad_norm": 0.11440466344356537,
      "learning_rate": 0.0002998358726718008,
      "loss": 0.5016182661056519,
      "memory(GiB)": 78.26,
      "step": 332,
      "token_acc": 0.8598369870713884,
      "train_speed(iter/s)": 0.034982
    },
    {
      "epoch": 0.06452550501380613,
      "grad_norm": 0.1108684316277504,
      "learning_rate": 0.0002998313460823735,
      "loss": 0.5170926451683044,
      "memory(GiB)": 78.26,
      "step": 333,
      "token_acc": 0.8539208882720333,
      "train_speed(iter/s)": 0.034983
    },
    {
      "epoch": 0.06471927529913288,
      "grad_norm": 0.12245868891477585,
      "learning_rate": 0.00029982675795318616,
      "loss": 0.49607276916503906,
      "memory(GiB)": 78.26,
      "step": 334,
      "token_acc": 0.8618406713164778,
      "train_speed(iter/s)": 0.034983
    },
    {
      "epoch": 0.06491304558445962,
      "grad_norm": 0.11894813925027847,
      "learning_rate": 0.0002998221082861234,
      "loss": 0.49732956290245056,
      "memory(GiB)": 78.26,
      "step": 335,
      "token_acc": 0.8582034755649832,
      "train_speed(iter/s)": 0.034985
    },
    {
      "epoch": 0.06510681586978637,
      "grad_norm": 0.12267972528934479,
      "learning_rate": 0.0002998173970830949,
      "loss": 0.5078924298286438,
      "memory(GiB)": 78.26,
      "step": 336,
      "token_acc": 0.8583969800719866,
      "train_speed(iter/s)": 0.034986
    },
    {
      "epoch": 0.06530058615511311,
      "grad_norm": 0.10648109018802643,
      "learning_rate": 0.0002998126243460357,
      "loss": 0.5070585608482361,
      "memory(GiB)": 78.26,
      "step": 337,
      "token_acc": 0.8565583698958306,
      "train_speed(iter/s)": 0.034987
    },
    {
      "epoch": 0.06549435644043985,
      "grad_norm": 0.11792565882205963,
      "learning_rate": 0.000299807790076906,
      "loss": 0.5286574959754944,
      "memory(GiB)": 78.26,
      "step": 338,
      "token_acc": 0.8529562054765698,
      "train_speed(iter/s)": 0.034989
    },
    {
      "epoch": 0.0656881267257666,
      "grad_norm": 0.12195685505867004,
      "learning_rate": 0.0002998028942776914,
      "loss": 0.4869121313095093,
      "memory(GiB)": 78.26,
      "step": 339,
      "token_acc": 0.8612365934096389,
      "train_speed(iter/s)": 0.03499
    },
    {
      "epoch": 0.06588189701109334,
      "grad_norm": 0.1249171569943428,
      "learning_rate": 0.0002997979369504028,
      "loss": 0.5445664525032043,
      "memory(GiB)": 78.26,
      "step": 340,
      "token_acc": 0.8522178660532028,
      "train_speed(iter/s)": 0.034991
    },
    {
      "epoch": 0.06607566729642009,
      "grad_norm": 0.13457630574703217,
      "learning_rate": 0.0002997929180970763,
      "loss": 0.5670903325080872,
      "memory(GiB)": 78.26,
      "step": 341,
      "token_acc": 0.8372440096177763,
      "train_speed(iter/s)": 0.034991
    },
    {
      "epoch": 0.06626943758174685,
      "grad_norm": 0.11432075500488281,
      "learning_rate": 0.0002997878377197732,
      "loss": 0.5261147022247314,
      "memory(GiB)": 78.26,
      "step": 342,
      "token_acc": 0.8552226935312831,
      "train_speed(iter/s)": 0.034992
    },
    {
      "epoch": 0.06646320786707359,
      "grad_norm": 0.10323546081781387,
      "learning_rate": 0.00029978269582058015,
      "loss": 0.4720154404640198,
      "memory(GiB)": 78.26,
      "step": 343,
      "token_acc": 0.8659866148531952,
      "train_speed(iter/s)": 0.034992
    },
    {
      "epoch": 0.06665697815240033,
      "grad_norm": 0.11725510656833649,
      "learning_rate": 0.0002997774924016092,
      "loss": 0.5160101652145386,
      "memory(GiB)": 78.26,
      "step": 344,
      "token_acc": 0.8551109929549382,
      "train_speed(iter/s)": 0.034992
    },
    {
      "epoch": 0.06685074843772708,
      "grad_norm": 0.12052306532859802,
      "learning_rate": 0.0002997722274649974,
      "loss": 0.537044107913971,
      "memory(GiB)": 78.26,
      "step": 345,
      "token_acc": 0.8476571428571429,
      "train_speed(iter/s)": 0.034993
    },
    {
      "epoch": 0.06704451872305382,
      "grad_norm": 0.11849239468574524,
      "learning_rate": 0.00029976690101290727,
      "loss": 0.5134192705154419,
      "memory(GiB)": 78.26,
      "step": 346,
      "token_acc": 0.8558187985790219,
      "train_speed(iter/s)": 0.034995
    },
    {
      "epoch": 0.06723828900838057,
      "grad_norm": 0.12307219952344894,
      "learning_rate": 0.00029976151304752645,
      "loss": 0.4876058101654053,
      "memory(GiB)": 78.26,
      "step": 347,
      "token_acc": 0.8650527622594661,
      "train_speed(iter/s)": 0.034996
    },
    {
      "epoch": 0.06743205929370731,
      "grad_norm": 0.11402394622564316,
      "learning_rate": 0.00029975606357106804,
      "loss": 0.47068169713020325,
      "memory(GiB)": 78.26,
      "step": 348,
      "token_acc": 0.8668408661682147,
      "train_speed(iter/s)": 0.034998
    },
    {
      "epoch": 0.06762582957903406,
      "grad_norm": 0.12601493299007416,
      "learning_rate": 0.00029975055258577016,
      "loss": 0.5161094665527344,
      "memory(GiB)": 78.26,
      "step": 349,
      "token_acc": 0.8581284381363401,
      "train_speed(iter/s)": 0.034999
    },
    {
      "epoch": 0.0678195998643608,
      "grad_norm": 0.12433503568172455,
      "learning_rate": 0.0002997449800938964,
      "loss": 0.5517579913139343,
      "memory(GiB)": 78.26,
      "step": 350,
      "token_acc": 0.8489635649712355,
      "train_speed(iter/s)": 0.035001
    },
    {
      "epoch": 0.06801337014968754,
      "grad_norm": 0.11672255396842957,
      "learning_rate": 0.0002997393460977355,
      "loss": 0.49452394247055054,
      "memory(GiB)": 78.26,
      "step": 351,
      "token_acc": 0.8589091777061701,
      "train_speed(iter/s)": 0.035002
    },
    {
      "epoch": 0.06820714043501429,
      "grad_norm": 0.11498520523309708,
      "learning_rate": 0.00029973365059960153,
      "loss": 0.5128313302993774,
      "memory(GiB)": 78.26,
      "step": 352,
      "token_acc": 0.8535805294087055,
      "train_speed(iter/s)": 0.035002
    },
    {
      "epoch": 0.06840091072034103,
      "grad_norm": 0.14143286645412445,
      "learning_rate": 0.00029972789360183376,
      "loss": 0.5255135893821716,
      "memory(GiB)": 78.26,
      "step": 353,
      "token_acc": 0.8556857047731469,
      "train_speed(iter/s)": 0.035003
    },
    {
      "epoch": 0.06859468100566778,
      "grad_norm": 0.12254343181848526,
      "learning_rate": 0.00029972207510679675,
      "loss": 0.5137450695037842,
      "memory(GiB)": 78.26,
      "step": 354,
      "token_acc": 0.8559991402933749,
      "train_speed(iter/s)": 0.035004
    },
    {
      "epoch": 0.06878845129099452,
      "grad_norm": 0.11112511903047562,
      "learning_rate": 0.0002997161951168803,
      "loss": 0.46960967779159546,
      "memory(GiB)": 78.26,
      "step": 355,
      "token_acc": 0.8666562937606971,
      "train_speed(iter/s)": 0.035004
    },
    {
      "epoch": 0.06898222157632126,
      "grad_norm": 0.11723387986421585,
      "learning_rate": 0.0002997102536344995,
      "loss": 0.5213668346405029,
      "memory(GiB)": 78.26,
      "step": 356,
      "token_acc": 0.8504984318996416,
      "train_speed(iter/s)": 0.035005
    },
    {
      "epoch": 0.06917599186164802,
      "grad_norm": 0.11225343495607376,
      "learning_rate": 0.0002997042506620946,
      "loss": 0.4935987591743469,
      "memory(GiB)": 78.26,
      "step": 357,
      "token_acc": 0.8581589163069739,
      "train_speed(iter/s)": 0.035007
    },
    {
      "epoch": 0.06936976214697477,
      "grad_norm": 0.11436620354652405,
      "learning_rate": 0.0002996981862021313,
      "loss": 0.47615599632263184,
      "memory(GiB)": 78.26,
      "step": 358,
      "token_acc": 0.8642991737933033,
      "train_speed(iter/s)": 0.035007
    },
    {
      "epoch": 0.06956353243230151,
      "grad_norm": 0.12642012536525726,
      "learning_rate": 0.00029969206025710037,
      "loss": 0.5081407427787781,
      "memory(GiB)": 78.26,
      "step": 359,
      "token_acc": 0.8585974082543831,
      "train_speed(iter/s)": 0.035008
    },
    {
      "epoch": 0.06975730271762826,
      "grad_norm": 0.1168803945183754,
      "learning_rate": 0.0002996858728295179,
      "loss": 0.4714093804359436,
      "memory(GiB)": 78.26,
      "step": 360,
      "token_acc": 0.8668951045236009,
      "train_speed(iter/s)": 0.035008
    },
    {
      "epoch": 0.069951073002955,
      "grad_norm": 0.11019770056009293,
      "learning_rate": 0.00029967962392192526,
      "loss": 0.5050376057624817,
      "memory(GiB)": 78.26,
      "step": 361,
      "token_acc": 0.8604877186782117,
      "train_speed(iter/s)": 0.03501
    },
    {
      "epoch": 0.07014484328828174,
      "grad_norm": 0.12403653562068939,
      "learning_rate": 0.000299673313536889,
      "loss": 0.49756452441215515,
      "memory(GiB)": 78.26,
      "step": 362,
      "token_acc": 0.8619797028974849,
      "train_speed(iter/s)": 0.035011
    },
    {
      "epoch": 0.07033861357360849,
      "grad_norm": 0.10632438957691193,
      "learning_rate": 0.00029966694167700105,
      "loss": 0.50919508934021,
      "memory(GiB)": 78.26,
      "step": 363,
      "token_acc": 0.8557894184337602,
      "train_speed(iter/s)": 0.035011
    },
    {
      "epoch": 0.07053238385893523,
      "grad_norm": 0.12035337090492249,
      "learning_rate": 0.0002996605083448784,
      "loss": 0.5421941876411438,
      "memory(GiB)": 78.26,
      "step": 364,
      "token_acc": 0.8479168897932955,
      "train_speed(iter/s)": 0.035012
    },
    {
      "epoch": 0.07072615414426198,
      "grad_norm": 0.11031734943389893,
      "learning_rate": 0.00029965401354316345,
      "loss": 0.48488667607307434,
      "memory(GiB)": 78.26,
      "step": 365,
      "token_acc": 0.8627623778240666,
      "train_speed(iter/s)": 0.035013
    },
    {
      "epoch": 0.07091992442958872,
      "grad_norm": 0.10722577571868896,
      "learning_rate": 0.00029964745727452375,
      "loss": 0.45194217562675476,
      "memory(GiB)": 78.26,
      "step": 366,
      "token_acc": 0.8705380798689725,
      "train_speed(iter/s)": 0.035013
    },
    {
      "epoch": 0.07111369471491547,
      "grad_norm": 0.10588467866182327,
      "learning_rate": 0.0002996408395416521,
      "loss": 0.45176932215690613,
      "memory(GiB)": 78.26,
      "step": 367,
      "token_acc": 0.8708429432333904,
      "train_speed(iter/s)": 0.035014
    },
    {
      "epoch": 0.07130746500024221,
      "grad_norm": 0.10778294503688812,
      "learning_rate": 0.0002996341603472668,
      "loss": 0.501011312007904,
      "memory(GiB)": 78.26,
      "step": 368,
      "token_acc": 0.8594550505951623,
      "train_speed(iter/s)": 0.035014
    },
    {
      "epoch": 0.07150123528556895,
      "grad_norm": 0.12546683847904205,
      "learning_rate": 0.00029962741969411096,
      "loss": 0.4865407943725586,
      "memory(GiB)": 78.26,
      "step": 369,
      "token_acc": 0.8665692482545868,
      "train_speed(iter/s)": 0.035016
    },
    {
      "epoch": 0.0716950055708957,
      "grad_norm": 0.11863457411527634,
      "learning_rate": 0.0002996206175849532,
      "loss": 0.5059066414833069,
      "memory(GiB)": 78.26,
      "step": 370,
      "token_acc": 0.8579709417580488,
      "train_speed(iter/s)": 0.035017
    },
    {
      "epoch": 0.07188877585622244,
      "grad_norm": 0.1237197294831276,
      "learning_rate": 0.0002996137540225873,
      "loss": 0.5302804708480835,
      "memory(GiB)": 78.26,
      "step": 371,
      "token_acc": 0.8513494809688581,
      "train_speed(iter/s)": 0.035018
    },
    {
      "epoch": 0.0720825461415492,
      "grad_norm": 0.11564111709594727,
      "learning_rate": 0.0002996068290098324,
      "loss": 0.4804614186286926,
      "memory(GiB)": 78.26,
      "step": 372,
      "token_acc": 0.8635471113692303,
      "train_speed(iter/s)": 0.035019
    },
    {
      "epoch": 0.07227631642687594,
      "grad_norm": 0.115287646651268,
      "learning_rate": 0.0002995998425495327,
      "loss": 0.5029768943786621,
      "memory(GiB)": 78.26,
      "step": 373,
      "token_acc": 0.8573006711038692,
      "train_speed(iter/s)": 0.035019
    },
    {
      "epoch": 0.07247008671220269,
      "grad_norm": 0.11475943773984909,
      "learning_rate": 0.0002995927946445578,
      "loss": 0.4446421265602112,
      "memory(GiB)": 78.26,
      "step": 374,
      "token_acc": 0.8731128990987405,
      "train_speed(iter/s)": 0.035021
    },
    {
      "epoch": 0.07266385699752943,
      "grad_norm": 0.11748611181974411,
      "learning_rate": 0.00029958568529780245,
      "loss": 0.4998936653137207,
      "memory(GiB)": 78.26,
      "step": 375,
      "token_acc": 0.8600776778413737,
      "train_speed(iter/s)": 0.035021
    },
    {
      "epoch": 0.07285762728285618,
      "grad_norm": 0.11014379560947418,
      "learning_rate": 0.00029957851451218654,
      "loss": 0.48282113671302795,
      "memory(GiB)": 78.26,
      "step": 376,
      "token_acc": 0.8650632477795992,
      "train_speed(iter/s)": 0.035022
    },
    {
      "epoch": 0.07305139756818292,
      "grad_norm": 0.1131962388753891,
      "learning_rate": 0.0002995712822906554,
      "loss": 0.47199296951293945,
      "memory(GiB)": 78.26,
      "step": 377,
      "token_acc": 0.8649652360874542,
      "train_speed(iter/s)": 0.035023
    },
    {
      "epoch": 0.07324516785350967,
      "grad_norm": 0.11221049726009369,
      "learning_rate": 0.0002995639886361795,
      "loss": 0.5101888179779053,
      "memory(GiB)": 78.26,
      "step": 378,
      "token_acc": 0.8550319599324829,
      "train_speed(iter/s)": 0.035024
    },
    {
      "epoch": 0.07343893813883641,
      "grad_norm": 0.1063636839389801,
      "learning_rate": 0.0002995566335517546,
      "loss": 0.5004944205284119,
      "memory(GiB)": 78.26,
      "step": 379,
      "token_acc": 0.8585596162973695,
      "train_speed(iter/s)": 0.035024
    },
    {
      "epoch": 0.07363270842416315,
      "grad_norm": 0.11803896725177765,
      "learning_rate": 0.00029954921704040147,
      "loss": 0.510295033454895,
      "memory(GiB)": 78.26,
      "step": 380,
      "token_acc": 0.8577513030528667,
      "train_speed(iter/s)": 0.035026
    },
    {
      "epoch": 0.0738264787094899,
      "grad_norm": 0.11037638783454895,
      "learning_rate": 0.00029954173910516635,
      "loss": 0.44948601722717285,
      "memory(GiB)": 78.26,
      "step": 381,
      "token_acc": 0.8719704952581665,
      "train_speed(iter/s)": 0.035026
    },
    {
      "epoch": 0.07402024899481664,
      "grad_norm": 0.1230226382613182,
      "learning_rate": 0.0002995341997491207,
      "loss": 0.5153728723526001,
      "memory(GiB)": 78.26,
      "step": 382,
      "token_acc": 0.8567678516574045,
      "train_speed(iter/s)": 0.035026
    },
    {
      "epoch": 0.07421401928014339,
      "grad_norm": 0.11615514755249023,
      "learning_rate": 0.00029952659897536106,
      "loss": 0.46452564001083374,
      "memory(GiB)": 78.26,
      "step": 383,
      "token_acc": 0.8670107503877628,
      "train_speed(iter/s)": 0.035027
    },
    {
      "epoch": 0.07440778956547013,
      "grad_norm": 0.13006049394607544,
      "learning_rate": 0.00029951893678700927,
      "loss": 0.506874144077301,
      "memory(GiB)": 78.26,
      "step": 384,
      "token_acc": 0.8588143291124011,
      "train_speed(iter/s)": 0.035028
    },
    {
      "epoch": 0.07460155985079688,
      "grad_norm": 0.11657480150461197,
      "learning_rate": 0.00029951121318721243,
      "loss": 0.49863314628601074,
      "memory(GiB)": 78.26,
      "step": 385,
      "token_acc": 0.8593393170109936,
      "train_speed(iter/s)": 0.035029
    },
    {
      "epoch": 0.07479533013612362,
      "grad_norm": 0.11655829846858978,
      "learning_rate": 0.0002995034281791428,
      "loss": 0.5014276504516602,
      "memory(GiB)": 78.26,
      "step": 386,
      "token_acc": 0.861993529502388,
      "train_speed(iter/s)": 0.03503
    },
    {
      "epoch": 0.07498910042145036,
      "grad_norm": 0.10777109861373901,
      "learning_rate": 0.0002994955817659979,
      "loss": 0.4714622497558594,
      "memory(GiB)": 78.26,
      "step": 387,
      "token_acc": 0.8677407562147482,
      "train_speed(iter/s)": 0.035031
    },
    {
      "epoch": 0.07518287070677712,
      "grad_norm": 0.12219083309173584,
      "learning_rate": 0.00029948767395100045,
      "loss": 0.5111258625984192,
      "memory(GiB)": 78.26,
      "step": 388,
      "token_acc": 0.8558166152672241,
      "train_speed(iter/s)": 0.035032
    },
    {
      "epoch": 0.07537664099210387,
      "grad_norm": 0.10944923013448715,
      "learning_rate": 0.00029947970473739844,
      "loss": 0.4479862451553345,
      "memory(GiB)": 78.26,
      "step": 389,
      "token_acc": 0.8729919678714859,
      "train_speed(iter/s)": 0.035033
    },
    {
      "epoch": 0.07557041127743061,
      "grad_norm": 0.10940490663051605,
      "learning_rate": 0.000299471674128465,
      "loss": 0.49675190448760986,
      "memory(GiB)": 78.26,
      "step": 390,
      "token_acc": 0.8615333689812339,
      "train_speed(iter/s)": 0.035032
    },
    {
      "epoch": 0.07576418156275735,
      "grad_norm": 0.11179753392934799,
      "learning_rate": 0.0002994635821274986,
      "loss": 0.4398466646671295,
      "memory(GiB)": 78.26,
      "step": 391,
      "token_acc": 0.8771611786033172,
      "train_speed(iter/s)": 0.035033
    },
    {
      "epoch": 0.0759579518480841,
      "grad_norm": 0.12613913416862488,
      "learning_rate": 0.0002994554287378227,
      "loss": 0.5353314876556396,
      "memory(GiB)": 78.26,
      "step": 392,
      "token_acc": 0.8493086152908249,
      "train_speed(iter/s)": 0.035033
    },
    {
      "epoch": 0.07615172213341084,
      "grad_norm": 0.11761578917503357,
      "learning_rate": 0.00029944721396278623,
      "loss": 0.5115870237350464,
      "memory(GiB)": 78.26,
      "step": 393,
      "token_acc": 0.8544060286040279,
      "train_speed(iter/s)": 0.035034
    },
    {
      "epoch": 0.07634549241873759,
      "grad_norm": 0.11017254739999771,
      "learning_rate": 0.0002994389378057632,
      "loss": 0.48874709010124207,
      "memory(GiB)": 78.26,
      "step": 394,
      "token_acc": 0.8645016225055574,
      "train_speed(iter/s)": 0.035034
    },
    {
      "epoch": 0.07653926270406433,
      "grad_norm": 0.10908011347055435,
      "learning_rate": 0.00029943060027015276,
      "loss": 0.47319239377975464,
      "memory(GiB)": 78.26,
      "step": 395,
      "token_acc": 0.8659232780237955,
      "train_speed(iter/s)": 0.035034
    },
    {
      "epoch": 0.07673303298939108,
      "grad_norm": 0.11207929253578186,
      "learning_rate": 0.0002994222013593795,
      "loss": 0.45064985752105713,
      "memory(GiB)": 78.26,
      "step": 396,
      "token_acc": 0.8735906331309627,
      "train_speed(iter/s)": 0.035035
    },
    {
      "epoch": 0.07692680327471782,
      "grad_norm": 0.11263241618871689,
      "learning_rate": 0.000299413741076893,
      "loss": 0.49612677097320557,
      "memory(GiB)": 78.26,
      "step": 397,
      "token_acc": 0.8607561516527616,
      "train_speed(iter/s)": 0.035036
    },
    {
      "epoch": 0.07712057356004456,
      "grad_norm": 0.12468399852514267,
      "learning_rate": 0.0002994052194261681,
      "loss": 0.530525267124176,
      "memory(GiB)": 78.26,
      "step": 398,
      "token_acc": 0.8516811113159998,
      "train_speed(iter/s)": 0.035037
    },
    {
      "epoch": 0.07731434384537131,
      "grad_norm": 0.10838499665260315,
      "learning_rate": 0.00029939663641070496,
      "loss": 0.46940740942955017,
      "memory(GiB)": 78.26,
      "step": 399,
      "token_acc": 0.8676767140734452,
      "train_speed(iter/s)": 0.035037
    },
    {
      "epoch": 0.07750811413069805,
      "grad_norm": 0.107694610953331,
      "learning_rate": 0.0002993879920340288,
      "loss": 0.4813489019870758,
      "memory(GiB)": 78.26,
      "step": 400,
      "token_acc": 0.8637204826412022,
      "train_speed(iter/s)": 0.035037
    },
    {
      "epoch": 0.0777018844160248,
      "grad_norm": 0.10585100203752518,
      "learning_rate": 0.00029937928629969007,
      "loss": 0.4711493253707886,
      "memory(GiB)": 78.26,
      "step": 401,
      "token_acc": 0.868919624217119,
      "train_speed(iter/s)": 0.034999
    },
    {
      "epoch": 0.07789565470135154,
      "grad_norm": 0.12802836298942566,
      "learning_rate": 0.0002993705192112645,
      "loss": 0.537087619304657,
      "memory(GiB)": 78.26,
      "step": 402,
      "token_acc": 0.8504514311327399,
      "train_speed(iter/s)": 0.035001
    },
    {
      "epoch": 0.0780894249866783,
      "grad_norm": 0.10905808210372925,
      "learning_rate": 0.00029936169077235294,
      "loss": 0.46871036291122437,
      "memory(GiB)": 78.26,
      "step": 403,
      "token_acc": 0.8674802147324147,
      "train_speed(iter/s)": 0.035002
    },
    {
      "epoch": 0.07828319527200504,
      "grad_norm": 0.1310214251279831,
      "learning_rate": 0.0002993528009865815,
      "loss": 0.5204190611839294,
      "memory(GiB)": 78.26,
      "step": 404,
      "token_acc": 0.8552753875213625,
      "train_speed(iter/s)": 0.035003
    },
    {
      "epoch": 0.07847696555733179,
      "grad_norm": 0.10810894519090652,
      "learning_rate": 0.0002993438498576014,
      "loss": 0.45181161165237427,
      "memory(GiB)": 78.26,
      "step": 405,
      "token_acc": 0.8717410764238559,
      "train_speed(iter/s)": 0.035004
    },
    {
      "epoch": 0.07867073584265853,
      "grad_norm": 0.10277310758829117,
      "learning_rate": 0.0002993348373890891,
      "loss": 0.4377118945121765,
      "memory(GiB)": 78.26,
      "step": 406,
      "token_acc": 0.8732325819672131,
      "train_speed(iter/s)": 0.035005
    },
    {
      "epoch": 0.07886450612798528,
      "grad_norm": 0.1221415251493454,
      "learning_rate": 0.0002993257635847464,
      "loss": 0.5307734608650208,
      "memory(GiB)": 78.26,
      "step": 407,
      "token_acc": 0.8525828880078148,
      "train_speed(iter/s)": 0.035005
    },
    {
      "epoch": 0.07905827641331202,
      "grad_norm": 0.11799792945384979,
      "learning_rate": 0.0002993166284483,
      "loss": 0.5113755464553833,
      "memory(GiB)": 78.26,
      "step": 408,
      "token_acc": 0.8569156381218984,
      "train_speed(iter/s)": 0.035006
    },
    {
      "epoch": 0.07925204669863876,
      "grad_norm": 0.11527646332979202,
      "learning_rate": 0.000299307431983502,
      "loss": 0.45992469787597656,
      "memory(GiB)": 78.26,
      "step": 409,
      "token_acc": 0.8694049499736703,
      "train_speed(iter/s)": 0.035007
    },
    {
      "epoch": 0.07944581698396551,
      "grad_norm": 0.11302848160266876,
      "learning_rate": 0.00029929817419412964,
      "loss": 0.492914617061615,
      "memory(GiB)": 78.26,
      "step": 410,
      "token_acc": 0.8602246439421388,
      "train_speed(iter/s)": 0.035007
    },
    {
      "epoch": 0.07963958726929225,
      "grad_norm": 0.11444272100925446,
      "learning_rate": 0.0002992888550839853,
      "loss": 0.5189880728721619,
      "memory(GiB)": 78.26,
      "step": 411,
      "token_acc": 0.8538719731479262,
      "train_speed(iter/s)": 0.035009
    },
    {
      "epoch": 0.079833357554619,
      "grad_norm": 0.1215919554233551,
      "learning_rate": 0.0002992794746568967,
      "loss": 0.5156800150871277,
      "memory(GiB)": 78.26,
      "step": 412,
      "token_acc": 0.8556100806786991,
      "train_speed(iter/s)": 0.03501
    },
    {
      "epoch": 0.08002712783994574,
      "grad_norm": 0.10883322358131409,
      "learning_rate": 0.0002992700329167166,
      "loss": 0.4452913999557495,
      "memory(GiB)": 78.26,
      "step": 413,
      "token_acc": 0.8729308276689324,
      "train_speed(iter/s)": 0.03501
    },
    {
      "epoch": 0.08022089812527249,
      "grad_norm": 0.10672541707754135,
      "learning_rate": 0.00029926052986732285,
      "loss": 0.4543689787387848,
      "memory(GiB)": 78.26,
      "step": 414,
      "token_acc": 0.8705642256902761,
      "train_speed(iter/s)": 0.03501
    },
    {
      "epoch": 0.08041466841059923,
      "grad_norm": 0.10857294499874115,
      "learning_rate": 0.00029925096551261873,
      "loss": 0.48616546392440796,
      "memory(GiB)": 78.26,
      "step": 415,
      "token_acc": 0.8620314125989207,
      "train_speed(iter/s)": 0.035011
    },
    {
      "epoch": 0.08060843869592597,
      "grad_norm": 0.11831134557723999,
      "learning_rate": 0.0002992413398565325,
      "loss": 0.48336830735206604,
      "memory(GiB)": 78.26,
      "step": 416,
      "token_acc": 0.8604266578468662,
      "train_speed(iter/s)": 0.035011
    },
    {
      "epoch": 0.08080220898125272,
      "grad_norm": 0.12000903487205505,
      "learning_rate": 0.0002992316529030178,
      "loss": 0.49910253286361694,
      "memory(GiB)": 78.26,
      "step": 417,
      "token_acc": 0.857624620965228,
      "train_speed(iter/s)": 0.035012
    },
    {
      "epoch": 0.08099597926657948,
      "grad_norm": 0.11360272020101547,
      "learning_rate": 0.0002992219046560532,
      "loss": 0.4726894199848175,
      "memory(GiB)": 78.26,
      "step": 418,
      "token_acc": 0.8649808638600328,
      "train_speed(iter/s)": 0.035013
    },
    {
      "epoch": 0.08118974955190622,
      "grad_norm": 0.11910063773393631,
      "learning_rate": 0.0002992120951196426,
      "loss": 0.49129027128219604,
      "memory(GiB)": 78.26,
      "step": 419,
      "token_acc": 0.8614566125740939,
      "train_speed(iter/s)": 0.035013
    },
    {
      "epoch": 0.08138351983723296,
      "grad_norm": 0.11072386801242828,
      "learning_rate": 0.0002992022242978151,
      "loss": 0.4478093385696411,
      "memory(GiB)": 78.26,
      "step": 420,
      "token_acc": 0.8705150236942144,
      "train_speed(iter/s)": 0.035014
    },
    {
      "epoch": 0.08157729012255971,
      "grad_norm": 0.11215101927518845,
      "learning_rate": 0.0002991922921946248,
      "loss": 0.43505847454071045,
      "memory(GiB)": 78.26,
      "step": 421,
      "token_acc": 0.8739453295629583,
      "train_speed(iter/s)": 0.035015
    },
    {
      "epoch": 0.08177106040788645,
      "grad_norm": 0.12848572432994843,
      "learning_rate": 0.0002991822988141512,
      "loss": 0.4977684020996094,
      "memory(GiB)": 78.26,
      "step": 422,
      "token_acc": 0.8609223368850855,
      "train_speed(iter/s)": 0.035016
    },
    {
      "epoch": 0.0819648306932132,
      "grad_norm": 0.10522928833961487,
      "learning_rate": 0.0002991722441604988,
      "loss": 0.415419340133667,
      "memory(GiB)": 78.26,
      "step": 423,
      "token_acc": 0.880544936757575,
      "train_speed(iter/s)": 0.035017
    },
    {
      "epoch": 0.08215860097853994,
      "grad_norm": 0.12996944785118103,
      "learning_rate": 0.00029916212823779723,
      "loss": 0.4946865737438202,
      "memory(GiB)": 78.26,
      "step": 424,
      "token_acc": 0.8606087064986903,
      "train_speed(iter/s)": 0.035018
    },
    {
      "epoch": 0.08235237126386669,
      "grad_norm": 0.1072884202003479,
      "learning_rate": 0.0002991519510502015,
      "loss": 0.44100552797317505,
      "memory(GiB)": 78.26,
      "step": 425,
      "token_acc": 0.8742977528089888,
      "train_speed(iter/s)": 0.035018
    },
    {
      "epoch": 0.08254614154919343,
      "grad_norm": 0.11559458076953888,
      "learning_rate": 0.0002991417126018916,
      "loss": 0.4932180643081665,
      "memory(GiB)": 78.26,
      "step": 426,
      "token_acc": 0.8621398554887082,
      "train_speed(iter/s)": 0.035018
    },
    {
      "epoch": 0.08273991183452017,
      "grad_norm": 0.10428472608327866,
      "learning_rate": 0.00029913141289707277,
      "loss": 0.44228169322013855,
      "memory(GiB)": 78.26,
      "step": 427,
      "token_acc": 0.8711459857697936,
      "train_speed(iter/s)": 0.035018
    },
    {
      "epoch": 0.08293368211984692,
      "grad_norm": 0.11177484691143036,
      "learning_rate": 0.0002991210519399753,
      "loss": 0.46969670057296753,
      "memory(GiB)": 78.26,
      "step": 428,
      "token_acc": 0.8665406640525826,
      "train_speed(iter/s)": 0.035019
    },
    {
      "epoch": 0.08312745240517366,
      "grad_norm": 0.12332018464803696,
      "learning_rate": 0.00029911062973485476,
      "loss": 0.5035005807876587,
      "memory(GiB)": 78.26,
      "step": 429,
      "token_acc": 0.856846298426235,
      "train_speed(iter/s)": 0.03502
    },
    {
      "epoch": 0.08332122269050041,
      "grad_norm": 0.11600656807422638,
      "learning_rate": 0.00029910014628599184,
      "loss": 0.45422035455703735,
      "memory(GiB)": 78.26,
      "step": 430,
      "token_acc": 0.8697181133128663,
      "train_speed(iter/s)": 0.035021
    },
    {
      "epoch": 0.08351499297582715,
      "grad_norm": 0.12605808675289154,
      "learning_rate": 0.0002990896015976924,
      "loss": 0.48837852478027344,
      "memory(GiB)": 78.26,
      "step": 431,
      "token_acc": 0.8621029303127309,
      "train_speed(iter/s)": 0.035023
    },
    {
      "epoch": 0.0837087632611539,
      "grad_norm": 0.11456278711557388,
      "learning_rate": 0.00029907899567428736,
      "loss": 0.48890623450279236,
      "memory(GiB)": 78.26,
      "step": 432,
      "token_acc": 0.8624089155593656,
      "train_speed(iter/s)": 0.035022
    },
    {
      "epoch": 0.08390253354648065,
      "grad_norm": 0.11591736972332001,
      "learning_rate": 0.0002990683285201329,
      "loss": 0.4833714962005615,
      "memory(GiB)": 78.26,
      "step": 433,
      "token_acc": 0.8619619484549993,
      "train_speed(iter/s)": 0.035023
    },
    {
      "epoch": 0.0840963038318074,
      "grad_norm": 0.11160556972026825,
      "learning_rate": 0.00029905760013961024,
      "loss": 0.5017392635345459,
      "memory(GiB)": 78.26,
      "step": 434,
      "token_acc": 0.857612434705058,
      "train_speed(iter/s)": 0.035023
    },
    {
      "epoch": 0.08429007411713414,
      "grad_norm": 0.11990434676408768,
      "learning_rate": 0.000299046810537126,
      "loss": 0.5158663988113403,
      "memory(GiB)": 78.26,
      "step": 435,
      "token_acc": 0.8545139761525788,
      "train_speed(iter/s)": 0.035025
    },
    {
      "epoch": 0.08448384440246089,
      "grad_norm": 0.1072223111987114,
      "learning_rate": 0.0002990359597171115,
      "loss": 0.4850725531578064,
      "memory(GiB)": 78.26,
      "step": 436,
      "token_acc": 0.8633244854009116,
      "train_speed(iter/s)": 0.035025
    },
    {
      "epoch": 0.08467761468778763,
      "grad_norm": 0.1028514951467514,
      "learning_rate": 0.00029902504768402363,
      "loss": 0.4222199618816376,
      "memory(GiB)": 78.26,
      "step": 437,
      "token_acc": 0.8768839397139292,
      "train_speed(iter/s)": 0.035025
    },
    {
      "epoch": 0.08487138497311437,
      "grad_norm": 0.12057259678840637,
      "learning_rate": 0.0002990140744423443,
      "loss": 0.4803910255432129,
      "memory(GiB)": 78.26,
      "step": 438,
      "token_acc": 0.8642994549535108,
      "train_speed(iter/s)": 0.035027
    },
    {
      "epoch": 0.08506515525844112,
      "grad_norm": 0.11526720970869064,
      "learning_rate": 0.0002990030399965803,
      "loss": 0.505075216293335,
      "memory(GiB)": 78.26,
      "step": 439,
      "token_acc": 0.8569081317921075,
      "train_speed(iter/s)": 0.035027
    },
    {
      "epoch": 0.08525892554376786,
      "grad_norm": 0.115642249584198,
      "learning_rate": 0.000298991944351264,
      "loss": 0.4736385643482208,
      "memory(GiB)": 78.26,
      "step": 440,
      "token_acc": 0.8655344655344656,
      "train_speed(iter/s)": 0.035027
    },
    {
      "epoch": 0.08545269582909461,
      "grad_norm": 0.10979454219341278,
      "learning_rate": 0.0002989807875109525,
      "loss": 0.4606488347053528,
      "memory(GiB)": 78.26,
      "step": 441,
      "token_acc": 0.8691641871787936,
      "train_speed(iter/s)": 0.035028
    },
    {
      "epoch": 0.08564646611442135,
      "grad_norm": 0.1096881628036499,
      "learning_rate": 0.0002989695694802284,
      "loss": 0.46464937925338745,
      "memory(GiB)": 78.26,
      "step": 442,
      "token_acc": 0.8675828444373007,
      "train_speed(iter/s)": 0.035028
    },
    {
      "epoch": 0.0858402363997481,
      "grad_norm": 0.11683948338031769,
      "learning_rate": 0.0002989582902636991,
      "loss": 0.5202597379684448,
      "memory(GiB)": 78.26,
      "step": 443,
      "token_acc": 0.8540174192205939,
      "train_speed(iter/s)": 0.035029
    },
    {
      "epoch": 0.08603400668507484,
      "grad_norm": 0.11013835668563843,
      "learning_rate": 0.00029894694986599735,
      "loss": 0.45811498165130615,
      "memory(GiB)": 78.26,
      "step": 444,
      "token_acc": 0.8691748066748066,
      "train_speed(iter/s)": 0.035029
    },
    {
      "epoch": 0.08622777697040158,
      "grad_norm": 0.12029381096363068,
      "learning_rate": 0.0002989355482917809,
      "loss": 0.5269613862037659,
      "memory(GiB)": 78.26,
      "step": 445,
      "token_acc": 0.8544889657602255,
      "train_speed(iter/s)": 0.03503
    },
    {
      "epoch": 0.08642154725572833,
      "grad_norm": 0.11967018991708755,
      "learning_rate": 0.00029892408554573266,
      "loss": 0.5077260732650757,
      "memory(GiB)": 78.26,
      "step": 446,
      "token_acc": 0.8567639257294429,
      "train_speed(iter/s)": 0.035031
    },
    {
      "epoch": 0.08661531754105507,
      "grad_norm": 0.1161133274435997,
      "learning_rate": 0.00029891256163256085,
      "loss": 0.49124279618263245,
      "memory(GiB)": 78.26,
      "step": 447,
      "token_acc": 0.8613902094308108,
      "train_speed(iter/s)": 0.035031
    },
    {
      "epoch": 0.08680908782638183,
      "grad_norm": 0.11311367154121399,
      "learning_rate": 0.0002989009765569985,
      "loss": 0.4679524898529053,
      "memory(GiB)": 78.26,
      "step": 448,
      "token_acc": 0.8683576233183856,
      "train_speed(iter/s)": 0.035033
    },
    {
      "epoch": 0.08700285811170858,
      "grad_norm": 0.1288428157567978,
      "learning_rate": 0.00029888933032380394,
      "loss": 0.5291860103607178,
      "memory(GiB)": 78.26,
      "step": 449,
      "token_acc": 0.850703275825085,
      "train_speed(iter/s)": 0.035034
    },
    {
      "epoch": 0.08719662839703532,
      "grad_norm": 0.1069454550743103,
      "learning_rate": 0.0002988776229377606,
      "loss": 0.44400641322135925,
      "memory(GiB)": 78.26,
      "step": 450,
      "token_acc": 0.8715773445359871,
      "train_speed(iter/s)": 0.035035
    },
    {
      "epoch": 0.08739039868236206,
      "grad_norm": 0.12414740771055222,
      "learning_rate": 0.00029886585440367703,
      "loss": 0.49137061834335327,
      "memory(GiB)": 78.26,
      "step": 451,
      "token_acc": 0.8623676960997535,
      "train_speed(iter/s)": 0.035036
    },
    {
      "epoch": 0.08758416896768881,
      "grad_norm": 0.11944427341222763,
      "learning_rate": 0.0002988540247263869,
      "loss": 0.4897725582122803,
      "memory(GiB)": 78.26,
      "step": 452,
      "token_acc": 0.8620128910176807,
      "train_speed(iter/s)": 0.035037
    },
    {
      "epoch": 0.08777793925301555,
      "grad_norm": 0.11256846785545349,
      "learning_rate": 0.0002988421339107489,
      "loss": 0.45858433842658997,
      "memory(GiB)": 78.26,
      "step": 453,
      "token_acc": 0.8707719882217386,
      "train_speed(iter/s)": 0.035037
    },
    {
      "epoch": 0.0879717095383423,
      "grad_norm": 0.11249680817127228,
      "learning_rate": 0.000298830181961647,
      "loss": 0.49964413046836853,
      "memory(GiB)": 78.26,
      "step": 454,
      "token_acc": 0.8624331586368352,
      "train_speed(iter/s)": 0.035038
    },
    {
      "epoch": 0.08816547982366904,
      "grad_norm": 0.11055975407361984,
      "learning_rate": 0.00029881816888399014,
      "loss": 0.458358496427536,
      "memory(GiB)": 78.26,
      "step": 455,
      "token_acc": 0.8681250838813582,
      "train_speed(iter/s)": 0.035039
    },
    {
      "epoch": 0.08835925010899578,
      "grad_norm": 0.11990530788898468,
      "learning_rate": 0.0002988060946827124,
      "loss": 0.51871657371521,
      "memory(GiB)": 78.26,
      "step": 456,
      "token_acc": 0.8554156908665106,
      "train_speed(iter/s)": 0.035039
    },
    {
      "epoch": 0.08855302039432253,
      "grad_norm": 0.13425587117671967,
      "learning_rate": 0.00029879395936277303,
      "loss": 0.5187729597091675,
      "memory(GiB)": 78.26,
      "step": 457,
      "token_acc": 0.8552357434987268,
      "train_speed(iter/s)": 0.03504
    },
    {
      "epoch": 0.08874679067964927,
      "grad_norm": 0.12039537727832794,
      "learning_rate": 0.0002987817629291563,
      "loss": 0.5168625116348267,
      "memory(GiB)": 78.26,
      "step": 458,
      "token_acc": 0.852051777849393,
      "train_speed(iter/s)": 0.035041
    },
    {
      "epoch": 0.08894056096497602,
      "grad_norm": 0.10525672137737274,
      "learning_rate": 0.0002987695053868716,
      "loss": 0.47165447473526,
      "memory(GiB)": 78.26,
      "step": 459,
      "token_acc": 0.8652151769798828,
      "train_speed(iter/s)": 0.03504
    },
    {
      "epoch": 0.08913433125030276,
      "grad_norm": 0.12413229048252106,
      "learning_rate": 0.00029875718674095346,
      "loss": 0.4889269471168518,
      "memory(GiB)": 78.26,
      "step": 460,
      "token_acc": 0.8654352837852674,
      "train_speed(iter/s)": 0.03504
    },
    {
      "epoch": 0.0893281015356295,
      "grad_norm": 0.11871378868818283,
      "learning_rate": 0.00029874480699646145,
      "loss": 0.47212713956832886,
      "memory(GiB)": 78.26,
      "step": 461,
      "token_acc": 0.8666805704006882,
      "train_speed(iter/s)": 0.03504
    },
    {
      "epoch": 0.08952187182095625,
      "grad_norm": 0.11364120990037918,
      "learning_rate": 0.0002987323661584803,
      "loss": 0.4883894622325897,
      "memory(GiB)": 78.26,
      "step": 462,
      "token_acc": 0.8616672536642194,
      "train_speed(iter/s)": 0.035041
    },
    {
      "epoch": 0.08971564210628301,
      "grad_norm": 0.11412378400564194,
      "learning_rate": 0.00029871986423211976,
      "loss": 0.48642444610595703,
      "memory(GiB)": 78.26,
      "step": 463,
      "token_acc": 0.8636845270483118,
      "train_speed(iter/s)": 0.035042
    },
    {
      "epoch": 0.08990941239160975,
      "grad_norm": 0.12467852979898453,
      "learning_rate": 0.0002987073012225147,
      "loss": 0.537085235118866,
      "memory(GiB)": 78.26,
      "step": 464,
      "token_acc": 0.851391779396462,
      "train_speed(iter/s)": 0.035042
    },
    {
      "epoch": 0.0901031826769365,
      "grad_norm": 0.11268845200538635,
      "learning_rate": 0.00029869467713482516,
      "loss": 0.5060177445411682,
      "memory(GiB)": 78.26,
      "step": 465,
      "token_acc": 0.8563837385366968,
      "train_speed(iter/s)": 0.035042
    },
    {
      "epoch": 0.09029695296226324,
      "grad_norm": 0.1154741421341896,
      "learning_rate": 0.00029868199197423607,
      "loss": 0.5034860372543335,
      "memory(GiB)": 78.26,
      "step": 466,
      "token_acc": 0.8582380491378634,
      "train_speed(iter/s)": 0.035043
    },
    {
      "epoch": 0.09049072324758999,
      "grad_norm": 0.11442988365888596,
      "learning_rate": 0.0002986692457459577,
      "loss": 0.4907042980194092,
      "memory(GiB)": 78.26,
      "step": 467,
      "token_acc": 0.8586531036135595,
      "train_speed(iter/s)": 0.035045
    },
    {
      "epoch": 0.09068449353291673,
      "grad_norm": 0.1217266321182251,
      "learning_rate": 0.00029865643845522515,
      "loss": 0.5039050579071045,
      "memory(GiB)": 78.26,
      "step": 468,
      "token_acc": 0.859063377160585,
      "train_speed(iter/s)": 0.035045
    },
    {
      "epoch": 0.09087826381824347,
      "grad_norm": 0.11309139430522919,
      "learning_rate": 0.00029864357010729885,
      "loss": 0.459963858127594,
      "memory(GiB)": 78.26,
      "step": 469,
      "token_acc": 0.8677241875771288,
      "train_speed(iter/s)": 0.035045
    },
    {
      "epoch": 0.09107203410357022,
      "grad_norm": 0.10612141340970993,
      "learning_rate": 0.00029863064070746406,
      "loss": 0.4284835755825043,
      "memory(GiB)": 78.26,
      "step": 470,
      "token_acc": 0.8769310129503435,
      "train_speed(iter/s)": 0.035045
    },
    {
      "epoch": 0.09126580438889696,
      "grad_norm": 0.13069190084934235,
      "learning_rate": 0.00029861765026103126,
      "loss": 0.570850670337677,
      "memory(GiB)": 78.26,
      "step": 471,
      "token_acc": 0.8367923900751679,
      "train_speed(iter/s)": 0.035047
    },
    {
      "epoch": 0.0914595746742237,
      "grad_norm": 0.11362801492214203,
      "learning_rate": 0.000298604598773336,
      "loss": 0.47750890254974365,
      "memory(GiB)": 78.26,
      "step": 472,
      "token_acc": 0.8643800970372665,
      "train_speed(iter/s)": 0.035048
    },
    {
      "epoch": 0.09165334495955045,
      "grad_norm": 0.11324970424175262,
      "learning_rate": 0.0002985914862497388,
      "loss": 0.47157877683639526,
      "memory(GiB)": 78.26,
      "step": 473,
      "token_acc": 0.8665956123989698,
      "train_speed(iter/s)": 0.035049
    },
    {
      "epoch": 0.0918471152448772,
      "grad_norm": 0.11150199174880981,
      "learning_rate": 0.00029857831269562544,
      "loss": 0.47126272320747375,
      "memory(GiB)": 78.26,
      "step": 474,
      "token_acc": 0.8654503990877993,
      "train_speed(iter/s)": 0.035049
    },
    {
      "epoch": 0.09204088553020394,
      "grad_norm": 0.10614330321550369,
      "learning_rate": 0.00029856507811640667,
      "loss": 0.4606248736381531,
      "memory(GiB)": 78.26,
      "step": 475,
      "token_acc": 0.8700272670069217,
      "train_speed(iter/s)": 0.03505
    },
    {
      "epoch": 0.09223465581553068,
      "grad_norm": 0.11242176592350006,
      "learning_rate": 0.0002985517825175181,
      "loss": 0.46226733922958374,
      "memory(GiB)": 78.26,
      "step": 476,
      "token_acc": 0.8658786323456982,
      "train_speed(iter/s)": 0.03505
    },
    {
      "epoch": 0.09242842610085743,
      "grad_norm": 0.11099898815155029,
      "learning_rate": 0.0002985384259044208,
      "loss": 0.4800049662590027,
      "memory(GiB)": 78.26,
      "step": 477,
      "token_acc": 0.8640604175921812,
      "train_speed(iter/s)": 0.03505
    },
    {
      "epoch": 0.09262219638618417,
      "grad_norm": 0.11142181605100632,
      "learning_rate": 0.0002985250082826005,
      "loss": 0.48013511300086975,
      "memory(GiB)": 78.26,
      "step": 478,
      "token_acc": 0.8613092239573668,
      "train_speed(iter/s)": 0.03505
    },
    {
      "epoch": 0.09281596667151093,
      "grad_norm": 0.11062056571245193,
      "learning_rate": 0.0002985115296575684,
      "loss": 0.5107330083847046,
      "memory(GiB)": 78.26,
      "step": 479,
      "token_acc": 0.8574547485466223,
      "train_speed(iter/s)": 0.03505
    },
    {
      "epoch": 0.09300973695683767,
      "grad_norm": 0.10769648849964142,
      "learning_rate": 0.00029849799003486035,
      "loss": 0.4654473066329956,
      "memory(GiB)": 78.26,
      "step": 480,
      "token_acc": 0.8700014251104461,
      "train_speed(iter/s)": 0.035052
    },
    {
      "epoch": 0.09320350724216442,
      "grad_norm": 0.11566044390201569,
      "learning_rate": 0.00029848438942003746,
      "loss": 0.5077210664749146,
      "memory(GiB)": 78.26,
      "step": 481,
      "token_acc": 0.857971693966654,
      "train_speed(iter/s)": 0.035053
    },
    {
      "epoch": 0.09339727752749116,
      "grad_norm": 0.10409853607416153,
      "learning_rate": 0.00029847072781868597,
      "loss": 0.4633353352546692,
      "memory(GiB)": 78.26,
      "step": 482,
      "token_acc": 0.8679831912361242,
      "train_speed(iter/s)": 0.035053
    },
    {
      "epoch": 0.0935910478128179,
      "grad_norm": 0.11067858338356018,
      "learning_rate": 0.00029845700523641695,
      "loss": 0.4580146074295044,
      "memory(GiB)": 78.26,
      "step": 483,
      "token_acc": 0.8679867986798679,
      "train_speed(iter/s)": 0.035054
    },
    {
      "epoch": 0.09378481809814465,
      "grad_norm": 0.11449804157018661,
      "learning_rate": 0.0002984432216788667,
      "loss": 0.4559538662433624,
      "memory(GiB)": 78.26,
      "step": 484,
      "token_acc": 0.8718022752208812,
      "train_speed(iter/s)": 0.035054
    },
    {
      "epoch": 0.0939785883834714,
      "grad_norm": 0.11818825453519821,
      "learning_rate": 0.0002984293771516965,
      "loss": 0.5102649331092834,
      "memory(GiB)": 78.26,
      "step": 485,
      "token_acc": 0.8580015707122576,
      "train_speed(iter/s)": 0.035055
    },
    {
      "epoch": 0.09417235866879814,
      "grad_norm": 0.11284741759300232,
      "learning_rate": 0.00029841547166059264,
      "loss": 0.48310020565986633,
      "memory(GiB)": 78.26,
      "step": 486,
      "token_acc": 0.8661896877956481,
      "train_speed(iter/s)": 0.035056
    },
    {
      "epoch": 0.09436612895412488,
      "grad_norm": 0.12485720217227936,
      "learning_rate": 0.0002984015052112665,
      "loss": 0.506294310092926,
      "memory(GiB)": 78.26,
      "step": 487,
      "token_acc": 0.8588789601949635,
      "train_speed(iter/s)": 0.035057
    },
    {
      "epoch": 0.09455989923945163,
      "grad_norm": 0.10859735310077667,
      "learning_rate": 0.0002983874778094545,
      "loss": 0.4783238172531128,
      "memory(GiB)": 78.26,
      "step": 488,
      "token_acc": 0.8641949593834618,
      "train_speed(iter/s)": 0.035057
    },
    {
      "epoch": 0.09475366952477837,
      "grad_norm": 0.10904448479413986,
      "learning_rate": 0.00029837338946091794,
      "loss": 0.4522710144519806,
      "memory(GiB)": 78.26,
      "step": 489,
      "token_acc": 0.8690276365871533,
      "train_speed(iter/s)": 0.035057
    },
    {
      "epoch": 0.09494743981010512,
      "grad_norm": 0.10352976620197296,
      "learning_rate": 0.0002983592401714435,
      "loss": 0.4140471816062927,
      "memory(GiB)": 78.26,
      "step": 490,
      "token_acc": 0.8802987811455075,
      "train_speed(iter/s)": 0.035057
    },
    {
      "epoch": 0.09514121009543186,
      "grad_norm": 0.10894495248794556,
      "learning_rate": 0.00029834502994684247,
      "loss": 0.4381650686264038,
      "memory(GiB)": 78.26,
      "step": 491,
      "token_acc": 0.875124829777576,
      "train_speed(iter/s)": 0.035057
    },
    {
      "epoch": 0.0953349803807586,
      "grad_norm": 0.10618970543146133,
      "learning_rate": 0.00029833075879295146,
      "loss": 0.44262564182281494,
      "memory(GiB)": 78.26,
      "step": 492,
      "token_acc": 0.8727838336831989,
      "train_speed(iter/s)": 0.035057
    },
    {
      "epoch": 0.09552875066608535,
      "grad_norm": 0.11135976016521454,
      "learning_rate": 0.00029831642671563203,
      "loss": 0.504035472869873,
      "memory(GiB)": 78.26,
      "step": 493,
      "token_acc": 0.8569741773191643,
      "train_speed(iter/s)": 0.035057
    },
    {
      "epoch": 0.09572252095141211,
      "grad_norm": 0.11450373381376266,
      "learning_rate": 0.00029830203372077077,
      "loss": 0.46363234519958496,
      "memory(GiB)": 78.26,
      "step": 494,
      "token_acc": 0.8658075691564747,
      "train_speed(iter/s)": 0.035058
    },
    {
      "epoch": 0.09591629123673885,
      "grad_norm": 0.1127939373254776,
      "learning_rate": 0.0002982875798142791,
      "loss": 0.4617582857608795,
      "memory(GiB)": 78.26,
      "step": 495,
      "token_acc": 0.8662511720258121,
      "train_speed(iter/s)": 0.035058
    },
    {
      "epoch": 0.0961100615220656,
      "grad_norm": 0.11518420279026031,
      "learning_rate": 0.00029827306500209387,
      "loss": 0.45998328924179077,
      "memory(GiB)": 78.26,
      "step": 496,
      "token_acc": 0.8671692940370117,
      "train_speed(iter/s)": 0.035059
    },
    {
      "epoch": 0.09630383180739234,
      "grad_norm": 0.10046973824501038,
      "learning_rate": 0.0002982584892901766,
      "loss": 0.4228544235229492,
      "memory(GiB)": 78.26,
      "step": 497,
      "token_acc": 0.8777845109683675,
      "train_speed(iter/s)": 0.035059
    },
    {
      "epoch": 0.09649760209271908,
      "grad_norm": 0.10890112817287445,
      "learning_rate": 0.00029824385268451394,
      "loss": 0.43357470631599426,
      "memory(GiB)": 78.26,
      "step": 498,
      "token_acc": 0.875686566090231,
      "train_speed(iter/s)": 0.035059
    },
    {
      "epoch": 0.09669137237804583,
      "grad_norm": 0.1078747883439064,
      "learning_rate": 0.0002982291551911174,
      "loss": 0.4050006866455078,
      "memory(GiB)": 78.26,
      "step": 499,
      "token_acc": 0.8838355027744645,
      "train_speed(iter/s)": 0.035059
    },
    {
      "epoch": 0.09688514266337257,
      "grad_norm": 0.12394701689481735,
      "learning_rate": 0.0002982143968160238,
      "loss": 0.42825788259506226,
      "memory(GiB)": 78.26,
      "step": 500,
      "token_acc": 0.8775999553895054,
      "train_speed(iter/s)": 0.03506
    },
    {
      "epoch": 0.09688514266337257,
      "eval_loss": 0.5390220284461975,
      "eval_runtime": 1346.167,
      "eval_samples_per_second": 5.013,
      "eval_steps_per_second": 5.013,
      "eval_token_acc": 0.8670573123458688,
      "step": 500
    },
    {
      "epoch": 0.09707891294869932,
      "grad_norm": 0.11574912816286087,
      "learning_rate": 0.0002981995775652948,
      "loss": 0.4860585331916809,
      "memory(GiB)": 78.26,
      "step": 501,
      "token_acc": 0.8617111447871986,
      "train_speed(iter/s)": 0.03204
    },
    {
      "epoch": 0.09727268323402606,
      "grad_norm": 0.10734662413597107,
      "learning_rate": 0.000298184697445017,
      "loss": 0.48478803038597107,
      "memory(GiB)": 78.26,
      "step": 502,
      "token_acc": 0.8619134645052572,
      "train_speed(iter/s)": 0.032046
    },
    {
      "epoch": 0.0974664535193528,
      "grad_norm": 0.10531099885702133,
      "learning_rate": 0.00029816975646130206,
      "loss": 0.40947991609573364,
      "memory(GiB)": 78.26,
      "step": 503,
      "token_acc": 0.8847219204866391,
      "train_speed(iter/s)": 0.032052
    },
    {
      "epoch": 0.09766022380467955,
      "grad_norm": 0.11789926886558533,
      "learning_rate": 0.0002981547546202867,
      "loss": 0.48241478204727173,
      "memory(GiB)": 78.26,
      "step": 504,
      "token_acc": 0.8646450249428649,
      "train_speed(iter/s)": 0.032058
    },
    {
      "epoch": 0.0978539940900063,
      "grad_norm": 0.12707959115505219,
      "learning_rate": 0.0002981396919281325,
      "loss": 0.5034913420677185,
      "memory(GiB)": 78.26,
      "step": 505,
      "token_acc": 0.8592644506741423,
      "train_speed(iter/s)": 0.032064
    },
    {
      "epoch": 0.09804776437533304,
      "grad_norm": 0.11420562863349915,
      "learning_rate": 0.0002981245683910262,
      "loss": 0.4770240783691406,
      "memory(GiB)": 78.26,
      "step": 506,
      "token_acc": 0.8642017671785194,
      "train_speed(iter/s)": 0.03207
    },
    {
      "epoch": 0.09824153466065978,
      "grad_norm": 0.11006432771682739,
      "learning_rate": 0.00029810938401517937,
      "loss": 0.4580528140068054,
      "memory(GiB)": 78.26,
      "step": 507,
      "token_acc": 0.8687396218329851,
      "train_speed(iter/s)": 0.032076
    },
    {
      "epoch": 0.09843530494598653,
      "grad_norm": 0.11123590171337128,
      "learning_rate": 0.00029809413880682866,
      "loss": 0.4647199213504791,
      "memory(GiB)": 78.26,
      "step": 508,
      "token_acc": 0.8687078223879421,
      "train_speed(iter/s)": 0.032082
    },
    {
      "epoch": 0.09862907523131328,
      "grad_norm": 0.11409325152635574,
      "learning_rate": 0.00029807883277223573,
      "loss": 0.44523417949676514,
      "memory(GiB)": 78.26,
      "step": 509,
      "token_acc": 0.8726046297715775,
      "train_speed(iter/s)": 0.032087
    },
    {
      "epoch": 0.09882284551664003,
      "grad_norm": 0.10299310833215714,
      "learning_rate": 0.00029806346591768713,
      "loss": 0.4694310128688812,
      "memory(GiB)": 78.26,
      "step": 510,
      "token_acc": 0.8651645115862245,
      "train_speed(iter/s)": 0.032093
    },
    {
      "epoch": 0.09901661580196677,
      "grad_norm": 0.11479201167821884,
      "learning_rate": 0.0002980480382494945,
      "loss": 0.4353720545768738,
      "memory(GiB)": 78.26,
      "step": 511,
      "token_acc": 0.8747248514197666,
      "train_speed(iter/s)": 0.032098
    },
    {
      "epoch": 0.09921038608729352,
      "grad_norm": 0.12371563911437988,
      "learning_rate": 0.0002980325497739943,
      "loss": 0.4968646466732025,
      "memory(GiB)": 78.26,
      "step": 512,
      "token_acc": 0.8599567167456011,
      "train_speed(iter/s)": 0.032104
    },
    {
      "epoch": 0.09940415637262026,
      "grad_norm": 0.12062571942806244,
      "learning_rate": 0.00029801700049754816,
      "loss": 0.5140507817268372,
      "memory(GiB)": 78.26,
      "step": 513,
      "token_acc": 0.8532771431740046,
      "train_speed(iter/s)": 0.03211
    },
    {
      "epoch": 0.099597926657947,
      "grad_norm": 0.11467831581830978,
      "learning_rate": 0.0002980013904265425,
      "loss": 0.4433932900428772,
      "memory(GiB)": 78.26,
      "step": 514,
      "token_acc": 0.8714406689954483,
      "train_speed(iter/s)": 0.032115
    },
    {
      "epoch": 0.09979169694327375,
      "grad_norm": 0.10191385447978973,
      "learning_rate": 0.00029798571956738887,
      "loss": 0.44032424688339233,
      "memory(GiB)": 78.26,
      "step": 515,
      "token_acc": 0.8739991993594876,
      "train_speed(iter/s)": 0.032121
    },
    {
      "epoch": 0.0999854672286005,
      "grad_norm": 0.1040392592549324,
      "learning_rate": 0.00029796998792652366,
      "loss": 0.43922415375709534,
      "memory(GiB)": 78.26,
      "step": 516,
      "token_acc": 0.8740507096930729,
      "train_speed(iter/s)": 0.032126
    },
    {
      "epoch": 0.10017923751392724,
      "grad_norm": 0.1119905412197113,
      "learning_rate": 0.00029795419551040833,
      "loss": 0.4818880259990692,
      "memory(GiB)": 78.26,
      "step": 517,
      "token_acc": 0.8635854743792163,
      "train_speed(iter/s)": 0.032131
    },
    {
      "epoch": 0.10037300779925398,
      "grad_norm": 0.10882623493671417,
      "learning_rate": 0.00029793834232552923,
      "loss": 0.4754379093647003,
      "memory(GiB)": 78.26,
      "step": 518,
      "token_acc": 0.8650517086330936,
      "train_speed(iter/s)": 0.032137
    },
    {
      "epoch": 0.10056677808458073,
      "grad_norm": 0.11377903819084167,
      "learning_rate": 0.00029792242837839764,
      "loss": 0.4642985165119171,
      "memory(GiB)": 78.26,
      "step": 519,
      "token_acc": 0.8642178631297238,
      "train_speed(iter/s)": 0.032143
    },
    {
      "epoch": 0.10076054836990747,
      "grad_norm": 0.1146177351474762,
      "learning_rate": 0.0002979064536755499,
      "loss": 0.47592130303382874,
      "memory(GiB)": 78.26,
      "step": 520,
      "token_acc": 0.8636443661971831,
      "train_speed(iter/s)": 0.032148
    },
    {
      "epoch": 0.10095431865523422,
      "grad_norm": 0.1103118285536766,
      "learning_rate": 0.00029789041822354725,
      "loss": 0.4815801978111267,
      "memory(GiB)": 78.26,
      "step": 521,
      "token_acc": 0.864503921461087,
      "train_speed(iter/s)": 0.032153
    },
    {
      "epoch": 0.10114808894056096,
      "grad_norm": 0.10768909007310867,
      "learning_rate": 0.00029787432202897586,
      "loss": 0.45158839225769043,
      "memory(GiB)": 78.26,
      "step": 522,
      "token_acc": 0.8713771448091021,
      "train_speed(iter/s)": 0.032159
    },
    {
      "epoch": 0.1013418592258877,
      "grad_norm": 0.11492832750082016,
      "learning_rate": 0.00029785816509844687,
      "loss": 0.510797381401062,
      "memory(GiB)": 78.26,
      "step": 523,
      "token_acc": 0.8555613215706626,
      "train_speed(iter/s)": 0.032164
    },
    {
      "epoch": 0.10153562951121446,
      "grad_norm": 0.11431033164262772,
      "learning_rate": 0.00029784194743859635,
      "loss": 0.4578917324542999,
      "memory(GiB)": 78.26,
      "step": 524,
      "token_acc": 0.8709328500014617,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.1017293997965412,
      "grad_norm": 0.11284809559583664,
      "learning_rate": 0.00029782566905608537,
      "loss": 0.5204124450683594,
      "memory(GiB)": 78.26,
      "step": 525,
      "token_acc": 0.8546162111927511,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.10192317008186795,
      "grad_norm": 0.10947126150131226,
      "learning_rate": 0.00029780932995759993,
      "loss": 0.47740259766578674,
      "memory(GiB)": 78.26,
      "step": 526,
      "token_acc": 0.8662353223100887,
      "train_speed(iter/s)": 0.03218
    },
    {
      "epoch": 0.1021169403671947,
      "grad_norm": 0.10436630249023438,
      "learning_rate": 0.0002977929301498508,
      "loss": 0.41452687978744507,
      "memory(GiB)": 78.26,
      "step": 527,
      "token_acc": 0.8821526069681326,
      "train_speed(iter/s)": 0.032185
    },
    {
      "epoch": 0.10231071065252144,
      "grad_norm": 0.10980913788080215,
      "learning_rate": 0.00029777646963957395,
      "loss": 0.4662986397743225,
      "memory(GiB)": 78.26,
      "step": 528,
      "token_acc": 0.8664124246946333,
      "train_speed(iter/s)": 0.032191
    },
    {
      "epoch": 0.10250448093784818,
      "grad_norm": 0.12164165824651718,
      "learning_rate": 0.00029775994843353015,
      "loss": 0.5069164037704468,
      "memory(GiB)": 78.26,
      "step": 529,
      "token_acc": 0.855151571221513,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.10269825122317493,
      "grad_norm": 0.10582577437162399,
      "learning_rate": 0.000297743366538505,
      "loss": 0.48306143283843994,
      "memory(GiB)": 78.26,
      "step": 530,
      "token_acc": 0.8629473684210527,
      "train_speed(iter/s)": 0.032202
    },
    {
      "epoch": 0.10289202150850167,
      "grad_norm": 0.10868332535028458,
      "learning_rate": 0.00029772672396130914,
      "loss": 0.46425890922546387,
      "memory(GiB)": 78.26,
      "step": 531,
      "token_acc": 0.867730867643638,
      "train_speed(iter/s)": 0.032208
    },
    {
      "epoch": 0.10308579179382842,
      "grad_norm": 0.10470977425575256,
      "learning_rate": 0.0002977100207087783,
      "loss": 0.4376215934753418,
      "memory(GiB)": 78.26,
      "step": 532,
      "token_acc": 0.8725883947271159,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.10327956207915516,
      "grad_norm": 0.10088858008384705,
      "learning_rate": 0.0002976932567877728,
      "loss": 0.44405412673950195,
      "memory(GiB)": 78.26,
      "step": 533,
      "token_acc": 0.8700791271440507,
      "train_speed(iter/s)": 0.032219
    },
    {
      "epoch": 0.1034733323644819,
      "grad_norm": 0.11034450680017471,
      "learning_rate": 0.00029767643220517803,
      "loss": 0.4524500370025635,
      "memory(GiB)": 78.26,
      "step": 534,
      "token_acc": 0.8718685509782921,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.10366710264980865,
      "grad_norm": 0.1017063558101654,
      "learning_rate": 0.0002976595469679044,
      "loss": 0.45182546973228455,
      "memory(GiB)": 78.26,
      "step": 535,
      "token_acc": 0.8708247343881372,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.10386087293513539,
      "grad_norm": 0.12204176932573318,
      "learning_rate": 0.000297642601082887,
      "loss": 0.4296923875808716,
      "memory(GiB)": 78.26,
      "step": 536,
      "token_acc": 0.8769677967208831,
      "train_speed(iter/s)": 0.032234
    },
    {
      "epoch": 0.10405464322046214,
      "grad_norm": 0.11646491289138794,
      "learning_rate": 0.00029762559455708606,
      "loss": 0.5057516694068909,
      "memory(GiB)": 78.26,
      "step": 537,
      "token_acc": 0.8552889453890566,
      "train_speed(iter/s)": 0.03224
    },
    {
      "epoch": 0.10424841350578888,
      "grad_norm": 0.1092115193605423,
      "learning_rate": 0.00029760852739748656,
      "loss": 0.4715476334095001,
      "memory(GiB)": 78.26,
      "step": 538,
      "token_acc": 0.8671120470554063,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.10444218379111564,
      "grad_norm": 0.10526008903980255,
      "learning_rate": 0.00029759139961109843,
      "loss": 0.4452831447124481,
      "memory(GiB)": 78.26,
      "step": 539,
      "token_acc": 0.8757942310248571,
      "train_speed(iter/s)": 0.032249
    },
    {
      "epoch": 0.10463595407644238,
      "grad_norm": 0.10792571306228638,
      "learning_rate": 0.00029757421120495657,
      "loss": 0.4550265669822693,
      "memory(GiB)": 78.26,
      "step": 540,
      "token_acc": 0.8692174444670878,
      "train_speed(iter/s)": 0.032254
    },
    {
      "epoch": 0.10482972436176913,
      "grad_norm": 0.13169942796230316,
      "learning_rate": 0.00029755696218612075,
      "loss": 0.4817619323730469,
      "memory(GiB)": 78.26,
      "step": 541,
      "token_acc": 0.8636959761549925,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 0.10502349464709587,
      "grad_norm": 0.1143367663025856,
      "learning_rate": 0.0002975396525616755,
      "loss": 0.4964498281478882,
      "memory(GiB)": 78.26,
      "step": 542,
      "token_acc": 0.8611530156476469,
      "train_speed(iter/s)": 0.032265
    },
    {
      "epoch": 0.10521726493242262,
      "grad_norm": 0.11364971101284027,
      "learning_rate": 0.0002975222823387304,
      "loss": 0.5033693313598633,
      "memory(GiB)": 78.26,
      "step": 543,
      "token_acc": 0.8565000138033846,
      "train_speed(iter/s)": 0.03227
    },
    {
      "epoch": 0.10541103521774936,
      "grad_norm": 0.11851288378238678,
      "learning_rate": 0.0002975048515244199,
      "loss": 0.4963114261627197,
      "memory(GiB)": 78.26,
      "step": 544,
      "token_acc": 0.8588933277681713,
      "train_speed(iter/s)": 0.032275
    },
    {
      "epoch": 0.1056048055030761,
      "grad_norm": 0.10574258863925934,
      "learning_rate": 0.00029748736012590325,
      "loss": 0.422842800617218,
      "memory(GiB)": 78.26,
      "step": 545,
      "token_acc": 0.8765757477808995,
      "train_speed(iter/s)": 0.03228
    },
    {
      "epoch": 0.10579857578840285,
      "grad_norm": 0.11669930815696716,
      "learning_rate": 0.00029746980815036463,
      "loss": 0.504793107509613,
      "memory(GiB)": 78.26,
      "step": 546,
      "token_acc": 0.8590873623674248,
      "train_speed(iter/s)": 0.032285
    },
    {
      "epoch": 0.10599234607372959,
      "grad_norm": 0.11788391321897507,
      "learning_rate": 0.00029745219560501317,
      "loss": 0.5016002058982849,
      "memory(GiB)": 78.26,
      "step": 547,
      "token_acc": 0.8547953818015547,
      "train_speed(iter/s)": 0.03229
    },
    {
      "epoch": 0.10618611635905634,
      "grad_norm": 0.1270827203989029,
      "learning_rate": 0.0002974345224970828,
      "loss": 0.49133074283599854,
      "memory(GiB)": 78.26,
      "step": 548,
      "token_acc": 0.8635658391797085,
      "train_speed(iter/s)": 0.032296
    },
    {
      "epoch": 0.10637988664438308,
      "grad_norm": 0.11762789636850357,
      "learning_rate": 0.0002974167888338323,
      "loss": 0.4606111943721771,
      "memory(GiB)": 78.26,
      "step": 549,
      "token_acc": 0.8683581749914666,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.10657365692970983,
      "grad_norm": 0.11224767565727234,
      "learning_rate": 0.00029739899462254534,
      "loss": 0.4845007658004761,
      "memory(GiB)": 78.26,
      "step": 550,
      "token_acc": 0.8617370771520969,
      "train_speed(iter/s)": 0.032306
    },
    {
      "epoch": 0.10676742721503657,
      "grad_norm": 0.12319989502429962,
      "learning_rate": 0.00029738113987053057,
      "loss": 0.5313453078269958,
      "memory(GiB)": 78.26,
      "step": 551,
      "token_acc": 0.8511862414550404,
      "train_speed(iter/s)": 0.032312
    },
    {
      "epoch": 0.10696119750036331,
      "grad_norm": 0.11050526797771454,
      "learning_rate": 0.00029736322458512137,
      "loss": 0.4900805354118347,
      "memory(GiB)": 78.26,
      "step": 552,
      "token_acc": 0.8623392457978925,
      "train_speed(iter/s)": 0.032317
    },
    {
      "epoch": 0.10715496778569006,
      "grad_norm": 0.10729413479566574,
      "learning_rate": 0.000297345248773676,
      "loss": 0.42731186747550964,
      "memory(GiB)": 78.26,
      "step": 553,
      "token_acc": 0.8775543301978592,
      "train_speed(iter/s)": 0.032322
    },
    {
      "epoch": 0.1073487380710168,
      "grad_norm": 0.11160185933113098,
      "learning_rate": 0.00029732721244357766,
      "loss": 0.43972256779670715,
      "memory(GiB)": 78.26,
      "step": 554,
      "token_acc": 0.8772162864568294,
      "train_speed(iter/s)": 0.032327
    },
    {
      "epoch": 0.10754250835634356,
      "grad_norm": 0.1065671369433403,
      "learning_rate": 0.0002973091156022343,
      "loss": 0.4213421940803528,
      "memory(GiB)": 78.26,
      "step": 555,
      "token_acc": 0.8795612510860121,
      "train_speed(iter/s)": 0.032332
    },
    {
      "epoch": 0.1077362786416703,
      "grad_norm": 0.12415304034948349,
      "learning_rate": 0.0002972909582570789,
      "loss": 0.5269736051559448,
      "memory(GiB)": 78.26,
      "step": 556,
      "token_acc": 0.8518212183871481,
      "train_speed(iter/s)": 0.032337
    },
    {
      "epoch": 0.10793004892699705,
      "grad_norm": 0.11127987504005432,
      "learning_rate": 0.00029727274041556903,
      "loss": 0.43144795298576355,
      "memory(GiB)": 78.26,
      "step": 557,
      "token_acc": 0.8748079670107538,
      "train_speed(iter/s)": 0.032342
    },
    {
      "epoch": 0.1081238192123238,
      "grad_norm": 0.12338504940271378,
      "learning_rate": 0.0002972544620851873,
      "loss": 0.5014538764953613,
      "memory(GiB)": 78.26,
      "step": 558,
      "token_acc": 0.8602012135858572,
      "train_speed(iter/s)": 0.032347
    },
    {
      "epoch": 0.10831758949765054,
      "grad_norm": 0.11730282008647919,
      "learning_rate": 0.0002972361232734411,
      "loss": 0.5035312175750732,
      "memory(GiB)": 78.26,
      "step": 559,
      "token_acc": 0.857784964507907,
      "train_speed(iter/s)": 0.032351
    },
    {
      "epoch": 0.10851135978297728,
      "grad_norm": 0.11394292116165161,
      "learning_rate": 0.00029721772398786267,
      "loss": 0.5110200047492981,
      "memory(GiB)": 78.26,
      "step": 560,
      "token_acc": 0.8569026399865478,
      "train_speed(iter/s)": 0.032356
    },
    {
      "epoch": 0.10870513006830403,
      "grad_norm": 0.10496672242879868,
      "learning_rate": 0.0002971992642360091,
      "loss": 0.4619162976741791,
      "memory(GiB)": 78.26,
      "step": 561,
      "token_acc": 0.8678721452368306,
      "train_speed(iter/s)": 0.03236
    },
    {
      "epoch": 0.10889890035363077,
      "grad_norm": 0.11340762674808502,
      "learning_rate": 0.0002971807440254623,
      "loss": 0.4587838649749756,
      "memory(GiB)": 78.26,
      "step": 562,
      "token_acc": 0.869656622284513,
      "train_speed(iter/s)": 0.032365
    },
    {
      "epoch": 0.10909267063895751,
      "grad_norm": 0.11113093048334122,
      "learning_rate": 0.0002971621633638291,
      "loss": 0.4310606122016907,
      "memory(GiB)": 78.26,
      "step": 563,
      "token_acc": 0.8775456994698572,
      "train_speed(iter/s)": 0.03237
    },
    {
      "epoch": 0.10928644092428426,
      "grad_norm": 0.11764062941074371,
      "learning_rate": 0.00029714352225874096,
      "loss": 0.4371579587459564,
      "memory(GiB)": 78.26,
      "step": 564,
      "token_acc": 0.8744286816643912,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.109480211209611,
      "grad_norm": 0.12444280833005905,
      "learning_rate": 0.00029712482071785436,
      "loss": 0.4438256621360779,
      "memory(GiB)": 78.26,
      "step": 565,
      "token_acc": 0.8732139048837705,
      "train_speed(iter/s)": 0.032379
    },
    {
      "epoch": 0.10967398149493775,
      "grad_norm": 0.1286301612854004,
      "learning_rate": 0.0002971060587488505,
      "loss": 0.48562532663345337,
      "memory(GiB)": 78.26,
      "step": 566,
      "token_acc": 0.8602812731310141,
      "train_speed(iter/s)": 0.032384
    },
    {
      "epoch": 0.10986775178026449,
      "grad_norm": 0.10392966866493225,
      "learning_rate": 0.00029708723635943536,
      "loss": 0.41354262828826904,
      "memory(GiB)": 78.26,
      "step": 567,
      "token_acc": 0.8819511093547112,
      "train_speed(iter/s)": 0.032389
    },
    {
      "epoch": 0.11006152206559124,
      "grad_norm": 0.12217994779348373,
      "learning_rate": 0.00029706835355733987,
      "loss": 0.4606168568134308,
      "memory(GiB)": 78.26,
      "step": 568,
      "token_acc": 0.8694803015044706,
      "train_speed(iter/s)": 0.032394
    },
    {
      "epoch": 0.11025529235091798,
      "grad_norm": 0.10960107296705246,
      "learning_rate": 0.00029704941035031977,
      "loss": 0.4657290577888489,
      "memory(GiB)": 78.26,
      "step": 569,
      "token_acc": 0.8707625272331154,
      "train_speed(iter/s)": 0.032398
    },
    {
      "epoch": 0.11044906263624474,
      "grad_norm": 0.1072949692606926,
      "learning_rate": 0.0002970304067461554,
      "loss": 0.47186678647994995,
      "memory(GiB)": 78.26,
      "step": 570,
      "token_acc": 0.8641140360004161,
      "train_speed(iter/s)": 0.032403
    },
    {
      "epoch": 0.11064283292157148,
      "grad_norm": 0.1146392896771431,
      "learning_rate": 0.0002970113427526521,
      "loss": 0.460860013961792,
      "memory(GiB)": 78.26,
      "step": 571,
      "token_acc": 0.8702294444126566,
      "train_speed(iter/s)": 0.032408
    },
    {
      "epoch": 0.11083660320689823,
      "grad_norm": 0.11065780371427536,
      "learning_rate": 0.00029699221837764,
      "loss": 0.4562651515007019,
      "memory(GiB)": 78.26,
      "step": 572,
      "token_acc": 0.8716815289879143,
      "train_speed(iter/s)": 0.032412
    },
    {
      "epoch": 0.11103037349222497,
      "grad_norm": 0.10410984605550766,
      "learning_rate": 0.0002969730336289741,
      "loss": 0.410195916891098,
      "memory(GiB)": 78.26,
      "step": 573,
      "token_acc": 0.8815111480133825,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.11122414377755171,
      "grad_norm": 0.1023547351360321,
      "learning_rate": 0.0002969537885145338,
      "loss": 0.44055572152137756,
      "memory(GiB)": 78.26,
      "step": 574,
      "token_acc": 0.8741572450990561,
      "train_speed(iter/s)": 0.032421
    },
    {
      "epoch": 0.11141791406287846,
      "grad_norm": 0.11351530253887177,
      "learning_rate": 0.00029693448304222384,
      "loss": 0.48370641469955444,
      "memory(GiB)": 78.26,
      "step": 575,
      "token_acc": 0.8646675213198818,
      "train_speed(iter/s)": 0.032426
    },
    {
      "epoch": 0.1116116843482052,
      "grad_norm": 0.10755743831396103,
      "learning_rate": 0.0002969151172199734,
      "loss": 0.4608403444290161,
      "memory(GiB)": 78.26,
      "step": 576,
      "token_acc": 0.8711053970511804,
      "train_speed(iter/s)": 0.032431
    },
    {
      "epoch": 0.11180545463353195,
      "grad_norm": 0.11767973005771637,
      "learning_rate": 0.00029689569105573654,
      "loss": 0.4457208216190338,
      "memory(GiB)": 78.26,
      "step": 577,
      "token_acc": 0.8742761124613676,
      "train_speed(iter/s)": 0.032435
    },
    {
      "epoch": 0.11199922491885869,
      "grad_norm": 0.10909876972436905,
      "learning_rate": 0.0002968762045574921,
      "loss": 0.4398786425590515,
      "memory(GiB)": 78.26,
      "step": 578,
      "token_acc": 0.8746623305095859,
      "train_speed(iter/s)": 0.032439
    },
    {
      "epoch": 0.11219299520418544,
      "grad_norm": 0.10908223688602448,
      "learning_rate": 0.0002968566577332438,
      "loss": 0.4776487350463867,
      "memory(GiB)": 78.26,
      "step": 579,
      "token_acc": 0.8675198893116569,
      "train_speed(iter/s)": 0.032444
    },
    {
      "epoch": 0.11238676548951218,
      "grad_norm": 0.11661187559366226,
      "learning_rate": 0.0002968370505910199,
      "loss": 0.4867633581161499,
      "memory(GiB)": 78.26,
      "step": 580,
      "token_acc": 0.8616142945163278,
      "train_speed(iter/s)": 0.032449
    },
    {
      "epoch": 0.11258053577483892,
      "grad_norm": 0.10946566611528397,
      "learning_rate": 0.0002968173831388737,
      "loss": 0.4538075029850006,
      "memory(GiB)": 78.26,
      "step": 581,
      "token_acc": 0.8708697120254691,
      "train_speed(iter/s)": 0.032453
    },
    {
      "epoch": 0.11277430606016567,
      "grad_norm": 0.10774602741003036,
      "learning_rate": 0.00029679765538488315,
      "loss": 0.4607040286064148,
      "memory(GiB)": 78.26,
      "step": 582,
      "token_acc": 0.8665272975798022,
      "train_speed(iter/s)": 0.032457
    },
    {
      "epoch": 0.11296807634549241,
      "grad_norm": 0.11124128848314285,
      "learning_rate": 0.00029677786733715085,
      "loss": 0.48860326409339905,
      "memory(GiB)": 78.26,
      "step": 583,
      "token_acc": 0.8632434789299546,
      "train_speed(iter/s)": 0.032461
    },
    {
      "epoch": 0.11316184663081916,
      "grad_norm": 0.11296314746141434,
      "learning_rate": 0.00029675801900380444,
      "loss": 0.44037532806396484,
      "memory(GiB)": 78.26,
      "step": 584,
      "token_acc": 0.8732820680628273,
      "train_speed(iter/s)": 0.032466
    },
    {
      "epoch": 0.11335561691614592,
      "grad_norm": 0.12023019045591354,
      "learning_rate": 0.000296738110392996,
      "loss": 0.4782663881778717,
      "memory(GiB)": 78.26,
      "step": 585,
      "token_acc": 0.8658891296908863,
      "train_speed(iter/s)": 0.03247
    },
    {
      "epoch": 0.11354938720147266,
      "grad_norm": 0.10668137669563293,
      "learning_rate": 0.0002967181415129027,
      "loss": 0.46215036511421204,
      "memory(GiB)": 78.26,
      "step": 586,
      "token_acc": 0.870391061452514,
      "train_speed(iter/s)": 0.032475
    },
    {
      "epoch": 0.1137431574867994,
      "grad_norm": 0.11411837488412857,
      "learning_rate": 0.00029669811237172615,
      "loss": 0.48195797204971313,
      "memory(GiB)": 78.26,
      "step": 587,
      "token_acc": 0.8628970636393407,
      "train_speed(iter/s)": 0.032479
    },
    {
      "epoch": 0.11393692777212615,
      "grad_norm": 0.11026424914598465,
      "learning_rate": 0.0002966780229776929,
      "loss": 0.4768543541431427,
      "memory(GiB)": 78.26,
      "step": 588,
      "token_acc": 0.8641793084597512,
      "train_speed(iter/s)": 0.032484
    },
    {
      "epoch": 0.11413069805745289,
      "grad_norm": 0.09665640443563461,
      "learning_rate": 0.0002966578733390543,
      "loss": 0.40953487157821655,
      "memory(GiB)": 78.26,
      "step": 589,
      "token_acc": 0.8820130831304016,
      "train_speed(iter/s)": 0.032488
    },
    {
      "epoch": 0.11432446834277964,
      "grad_norm": 0.10579517483711243,
      "learning_rate": 0.00029663766346408623,
      "loss": 0.4386206269264221,
      "memory(GiB)": 78.26,
      "step": 590,
      "token_acc": 0.8759878419452888,
      "train_speed(iter/s)": 0.032492
    },
    {
      "epoch": 0.11451823862810638,
      "grad_norm": 0.1087721437215805,
      "learning_rate": 0.00029661739336108947,
      "loss": 0.456497460603714,
      "memory(GiB)": 78.26,
      "step": 591,
      "token_acc": 0.8705190602070312,
      "train_speed(iter/s)": 0.032497
    },
    {
      "epoch": 0.11471200891343313,
      "grad_norm": 0.11170576512813568,
      "learning_rate": 0.0002965970630383895,
      "loss": 0.5045605897903442,
      "memory(GiB)": 78.26,
      "step": 592,
      "token_acc": 0.8581936467002697,
      "train_speed(iter/s)": 0.032502
    },
    {
      "epoch": 0.11490577919875987,
      "grad_norm": 0.10968808829784393,
      "learning_rate": 0.00029657667250433645,
      "loss": 0.45187485218048096,
      "memory(GiB)": 78.26,
      "step": 593,
      "token_acc": 0.8707526605975126,
      "train_speed(iter/s)": 0.032506
    },
    {
      "epoch": 0.11509954948408661,
      "grad_norm": 0.12064143270254135,
      "learning_rate": 0.00029655622176730543,
      "loss": 0.48971986770629883,
      "memory(GiB)": 78.26,
      "step": 594,
      "token_acc": 0.8627284832344753,
      "train_speed(iter/s)": 0.032511
    },
    {
      "epoch": 0.11529331976941336,
      "grad_norm": 0.10990637540817261,
      "learning_rate": 0.000296535710835696,
      "loss": 0.46801620721817017,
      "memory(GiB)": 78.26,
      "step": 595,
      "token_acc": 0.8661632053702538,
      "train_speed(iter/s)": 0.032515
    },
    {
      "epoch": 0.1154870900547401,
      "grad_norm": 0.11462108045816422,
      "learning_rate": 0.00029651513971793255,
      "loss": 0.4742628037929535,
      "memory(GiB)": 78.26,
      "step": 596,
      "token_acc": 0.8639128007756265,
      "train_speed(iter/s)": 0.03252
    },
    {
      "epoch": 0.11568086034006685,
      "grad_norm": 0.10829971730709076,
      "learning_rate": 0.0002964945084224642,
      "loss": 0.47385069727897644,
      "memory(GiB)": 78.26,
      "step": 597,
      "token_acc": 0.8652491420019095,
      "train_speed(iter/s)": 0.032524
    },
    {
      "epoch": 0.11587463062539359,
      "grad_norm": 0.1086253821849823,
      "learning_rate": 0.00029647381695776474,
      "loss": 0.4455265998840332,
      "memory(GiB)": 78.26,
      "step": 598,
      "token_acc": 0.872541041111267,
      "train_speed(iter/s)": 0.032528
    },
    {
      "epoch": 0.11606840091072033,
      "grad_norm": 0.11207325756549835,
      "learning_rate": 0.0002964530653323328,
      "loss": 0.4860404133796692,
      "memory(GiB)": 78.26,
      "step": 599,
      "token_acc": 0.8606821106821106,
      "train_speed(iter/s)": 0.032532
    },
    {
      "epoch": 0.11626217119604709,
      "grad_norm": 0.11669515818357468,
      "learning_rate": 0.0002964322535546916,
      "loss": 0.5320515632629395,
      "memory(GiB)": 78.26,
      "step": 600,
      "token_acc": 0.8462377317339149,
      "train_speed(iter/s)": 0.032537
    },
    {
      "epoch": 0.11645594148137384,
      "grad_norm": 0.10914132744073868,
      "learning_rate": 0.00029641138163338907,
      "loss": 0.4786490797996521,
      "memory(GiB)": 78.26,
      "step": 601,
      "token_acc": 0.8650941795350415,
      "train_speed(iter/s)": 0.03252
    },
    {
      "epoch": 0.11664971176670058,
      "grad_norm": 0.10866682976484299,
      "learning_rate": 0.0002963904495769978,
      "loss": 0.4337434768676758,
      "memory(GiB)": 78.26,
      "step": 602,
      "token_acc": 0.875823794658342,
      "train_speed(iter/s)": 0.032525
    },
    {
      "epoch": 0.11684348205202733,
      "grad_norm": 0.11499880999326706,
      "learning_rate": 0.0002963694573941153,
      "loss": 0.46066945791244507,
      "memory(GiB)": 78.26,
      "step": 603,
      "token_acc": 0.8695754925999232,
      "train_speed(iter/s)": 0.032529
    },
    {
      "epoch": 0.11703725233735407,
      "grad_norm": 0.09975297749042511,
      "learning_rate": 0.0002963484050933636,
      "loss": 0.4278182089328766,
      "memory(GiB)": 78.26,
      "step": 604,
      "token_acc": 0.8778897736293872,
      "train_speed(iter/s)": 0.032533
    },
    {
      "epoch": 0.11723102262268081,
      "grad_norm": 0.11780435591936111,
      "learning_rate": 0.0002963272926833893,
      "loss": 0.4575609862804413,
      "memory(GiB)": 78.26,
      "step": 605,
      "token_acc": 0.8685906319290465,
      "train_speed(iter/s)": 0.032538
    },
    {
      "epoch": 0.11742479290800756,
      "grad_norm": 0.10950589925050735,
      "learning_rate": 0.00029630612017286393,
      "loss": 0.43958431482315063,
      "memory(GiB)": 78.26,
      "step": 606,
      "token_acc": 0.8741350906095552,
      "train_speed(iter/s)": 0.032542
    },
    {
      "epoch": 0.1176185631933343,
      "grad_norm": 0.11566930264234543,
      "learning_rate": 0.00029628488757048365,
      "loss": 0.4670230448246002,
      "memory(GiB)": 78.26,
      "step": 607,
      "token_acc": 0.8678823105807314,
      "train_speed(iter/s)": 0.032546
    },
    {
      "epoch": 0.11781233347866105,
      "grad_norm": 0.11282102763652802,
      "learning_rate": 0.00029626359488496914,
      "loss": 0.4526304006576538,
      "memory(GiB)": 78.26,
      "step": 608,
      "token_acc": 0.8723770983213429,
      "train_speed(iter/s)": 0.032551
    },
    {
      "epoch": 0.11800610376398779,
      "grad_norm": 0.10305222868919373,
      "learning_rate": 0.0002962422421250661,
      "loss": 0.437641978263855,
      "memory(GiB)": 78.26,
      "step": 609,
      "token_acc": 0.8741574159921818,
      "train_speed(iter/s)": 0.032555
    },
    {
      "epoch": 0.11819987404931454,
      "grad_norm": 0.10737847536802292,
      "learning_rate": 0.0002962208292995444,
      "loss": 0.46929752826690674,
      "memory(GiB)": 78.26,
      "step": 610,
      "token_acc": 0.8661016542219875,
      "train_speed(iter/s)": 0.032559
    },
    {
      "epoch": 0.11839364433464128,
      "grad_norm": 0.11264858394861221,
      "learning_rate": 0.00029619935641719906,
      "loss": 0.4528808295726776,
      "memory(GiB)": 78.26,
      "step": 611,
      "token_acc": 0.8737685426339032,
      "train_speed(iter/s)": 0.032563
    },
    {
      "epoch": 0.11858741461996802,
      "grad_norm": 0.10752102732658386,
      "learning_rate": 0.00029617782348684946,
      "loss": 0.46101704239845276,
      "memory(GiB)": 78.26,
      "step": 612,
      "token_acc": 0.8676157977176102,
      "train_speed(iter/s)": 0.032568
    },
    {
      "epoch": 0.11878118490529477,
      "grad_norm": 0.12059959769248962,
      "learning_rate": 0.00029615623051733986,
      "loss": 0.4816949665546417,
      "memory(GiB)": 78.26,
      "step": 613,
      "token_acc": 0.8623071979434447,
      "train_speed(iter/s)": 0.032572
    },
    {
      "epoch": 0.11897495519062151,
      "grad_norm": 0.11858416348695755,
      "learning_rate": 0.00029613457751753903,
      "loss": 0.4759201407432556,
      "memory(GiB)": 78.26,
      "step": 614,
      "token_acc": 0.8641525737563264,
      "train_speed(iter/s)": 0.032576
    },
    {
      "epoch": 0.11916872547594827,
      "grad_norm": 0.11951098591089249,
      "learning_rate": 0.0002961128644963404,
      "loss": 0.4854854643344879,
      "memory(GiB)": 78.26,
      "step": 615,
      "token_acc": 0.8598115112756648,
      "train_speed(iter/s)": 0.032581
    },
    {
      "epoch": 0.11936249576127501,
      "grad_norm": 0.11944432556629181,
      "learning_rate": 0.0002960910914626621,
      "loss": 0.49400413036346436,
      "memory(GiB)": 78.26,
      "step": 616,
      "token_acc": 0.8579318625243885,
      "train_speed(iter/s)": 0.032585
    },
    {
      "epoch": 0.11955626604660176,
      "grad_norm": 0.10993971675634384,
      "learning_rate": 0.00029606925842544694,
      "loss": 0.421941339969635,
      "memory(GiB)": 78.26,
      "step": 617,
      "token_acc": 0.8773255961925123,
      "train_speed(iter/s)": 0.032588
    },
    {
      "epoch": 0.1197500363319285,
      "grad_norm": 0.10016026347875595,
      "learning_rate": 0.00029604736539366234,
      "loss": 0.4218504726886749,
      "memory(GiB)": 78.26,
      "step": 618,
      "token_acc": 0.8786656708986806,
      "train_speed(iter/s)": 0.032592
    },
    {
      "epoch": 0.11994380661725525,
      "grad_norm": 0.11316211521625519,
      "learning_rate": 0.00029602541237630026,
      "loss": 0.43415647745132446,
      "memory(GiB)": 78.26,
      "step": 619,
      "token_acc": 0.8750876058008518,
      "train_speed(iter/s)": 0.032595
    },
    {
      "epoch": 0.12013757690258199,
      "grad_norm": 0.12216460704803467,
      "learning_rate": 0.00029600339938237746,
      "loss": 0.4619022309780121,
      "memory(GiB)": 78.26,
      "step": 620,
      "token_acc": 0.867951649787651,
      "train_speed(iter/s)": 0.032599
    },
    {
      "epoch": 0.12033134718790874,
      "grad_norm": 0.10427816212177277,
      "learning_rate": 0.0002959813264209353,
      "loss": 0.4056597054004669,
      "memory(GiB)": 78.26,
      "step": 621,
      "token_acc": 0.8820620175323826,
      "train_speed(iter/s)": 0.032603
    },
    {
      "epoch": 0.12052511747323548,
      "grad_norm": 0.113120436668396,
      "learning_rate": 0.0002959591935010397,
      "loss": 0.4769364595413208,
      "memory(GiB)": 78.26,
      "step": 622,
      "token_acc": 0.8646665349552651,
      "train_speed(iter/s)": 0.032607
    },
    {
      "epoch": 0.12071888775856222,
      "grad_norm": 0.10840658843517303,
      "learning_rate": 0.00029593700063178127,
      "loss": 0.4412975609302521,
      "memory(GiB)": 78.26,
      "step": 623,
      "token_acc": 0.8720108105499385,
      "train_speed(iter/s)": 0.032611
    },
    {
      "epoch": 0.12091265804388897,
      "grad_norm": 0.11971483379602432,
      "learning_rate": 0.00029591474782227523,
      "loss": 0.4542793333530426,
      "memory(GiB)": 78.26,
      "step": 624,
      "token_acc": 0.8705333207178225,
      "train_speed(iter/s)": 0.032616
    },
    {
      "epoch": 0.12110642832921571,
      "grad_norm": 0.1130179837346077,
      "learning_rate": 0.00029589243508166136,
      "loss": 0.48266369104385376,
      "memory(GiB)": 78.26,
      "step": 625,
      "token_acc": 0.8646267140680548,
      "train_speed(iter/s)": 0.032619
    },
    {
      "epoch": 0.12130019861454246,
      "grad_norm": 0.11181651800870895,
      "learning_rate": 0.0002958700624191041,
      "loss": 0.41792306303977966,
      "memory(GiB)": 78.26,
      "step": 626,
      "token_acc": 0.8797664711191335,
      "train_speed(iter/s)": 0.032623
    },
    {
      "epoch": 0.1214939688998692,
      "grad_norm": 0.11760863661766052,
      "learning_rate": 0.00029584762984379253,
      "loss": 0.49022918939590454,
      "memory(GiB)": 78.26,
      "step": 627,
      "token_acc": 0.8612506898254378,
      "train_speed(iter/s)": 0.032627
    },
    {
      "epoch": 0.12168773918519595,
      "grad_norm": 0.10427683591842651,
      "learning_rate": 0.00029582513736494027,
      "loss": 0.4219359755516052,
      "memory(GiB)": 78.26,
      "step": 628,
      "token_acc": 0.8782604333868379,
      "train_speed(iter/s)": 0.032631
    },
    {
      "epoch": 0.12188150947052269,
      "grad_norm": 0.10674095898866653,
      "learning_rate": 0.00029580258499178566,
      "loss": 0.4139460325241089,
      "memory(GiB)": 78.26,
      "step": 629,
      "token_acc": 0.8805075685104946,
      "train_speed(iter/s)": 0.032635
    },
    {
      "epoch": 0.12207527975584945,
      "grad_norm": 0.1084229126572609,
      "learning_rate": 0.00029577997273359157,
      "loss": 0.42610836029052734,
      "memory(GiB)": 78.26,
      "step": 630,
      "token_acc": 0.8765509800395612,
      "train_speed(iter/s)": 0.032639
    },
    {
      "epoch": 0.12226905004117619,
      "grad_norm": 0.12075889855623245,
      "learning_rate": 0.00029575730059964534,
      "loss": 0.47887349128723145,
      "memory(GiB)": 78.26,
      "step": 631,
      "token_acc": 0.8641042884486059,
      "train_speed(iter/s)": 0.032643
    },
    {
      "epoch": 0.12246282032650294,
      "grad_norm": 0.13688334822654724,
      "learning_rate": 0.00029573456859925917,
      "loss": 0.5096902847290039,
      "memory(GiB)": 78.26,
      "step": 632,
      "token_acc": 0.8591979075850044,
      "train_speed(iter/s)": 0.032647
    },
    {
      "epoch": 0.12265659061182968,
      "grad_norm": 0.09688639640808105,
      "learning_rate": 0.0002957117767417696,
      "loss": 0.4519810974597931,
      "memory(GiB)": 78.26,
      "step": 633,
      "token_acc": 0.8703838308049426,
      "train_speed(iter/s)": 0.032651
    },
    {
      "epoch": 0.12285036089715642,
      "grad_norm": 0.10635363310575485,
      "learning_rate": 0.0002956889250365379,
      "loss": 0.4540587067604065,
      "memory(GiB)": 78.26,
      "step": 634,
      "token_acc": 0.8705026824989194,
      "train_speed(iter/s)": 0.032654
    },
    {
      "epoch": 0.12304413118248317,
      "grad_norm": 0.1146511361002922,
      "learning_rate": 0.00029566601349294985,
      "loss": 0.44937869906425476,
      "memory(GiB)": 78.26,
      "step": 635,
      "token_acc": 0.8692985300475573,
      "train_speed(iter/s)": 0.032658
    },
    {
      "epoch": 0.12323790146780991,
      "grad_norm": 0.11808592826128006,
      "learning_rate": 0.0002956430421204159,
      "loss": 0.4610764980316162,
      "memory(GiB)": 78.26,
      "step": 636,
      "token_acc": 0.86909560882543,
      "train_speed(iter/s)": 0.032663
    },
    {
      "epoch": 0.12343167175313666,
      "grad_norm": 0.11165952682495117,
      "learning_rate": 0.0002956200109283709,
      "loss": 0.44256392121315,
      "memory(GiB)": 78.26,
      "step": 637,
      "token_acc": 0.872780487804878,
      "train_speed(iter/s)": 0.032667
    },
    {
      "epoch": 0.1236254420384634,
      "grad_norm": 0.11056765913963318,
      "learning_rate": 0.0002955969199262745,
      "loss": 0.44177088141441345,
      "memory(GiB)": 78.26,
      "step": 638,
      "token_acc": 0.8738199880373443,
      "train_speed(iter/s)": 0.032671
    },
    {
      "epoch": 0.12381921232379015,
      "grad_norm": 0.12657155096530914,
      "learning_rate": 0.00029557376912361076,
      "loss": 0.4548972249031067,
      "memory(GiB)": 78.26,
      "step": 639,
      "token_acc": 0.8688163152817969,
      "train_speed(iter/s)": 0.032675
    },
    {
      "epoch": 0.12401298260911689,
      "grad_norm": 0.1116451621055603,
      "learning_rate": 0.00029555055852988836,
      "loss": 0.44486871361732483,
      "memory(GiB)": 78.26,
      "step": 640,
      "token_acc": 0.8735002147061053,
      "train_speed(iter/s)": 0.032679
    },
    {
      "epoch": 0.12420675289444363,
      "grad_norm": 0.11732634902000427,
      "learning_rate": 0.0002955272881546404,
      "loss": 0.4493526816368103,
      "memory(GiB)": 78.26,
      "step": 641,
      "token_acc": 0.8715556519852462,
      "train_speed(iter/s)": 0.032683
    },
    {
      "epoch": 0.12440052317977038,
      "grad_norm": 0.11484182626008987,
      "learning_rate": 0.00029550395800742477,
      "loss": 0.4882141351699829,
      "memory(GiB)": 78.26,
      "step": 642,
      "token_acc": 0.8629650457276468,
      "train_speed(iter/s)": 0.032687
    },
    {
      "epoch": 0.12459429346509712,
      "grad_norm": 0.10542822629213333,
      "learning_rate": 0.0002954805680978237,
      "loss": 0.4319552779197693,
      "memory(GiB)": 78.26,
      "step": 643,
      "token_acc": 0.8768453037188416,
      "train_speed(iter/s)": 0.03269
    },
    {
      "epoch": 0.12478806375042387,
      "grad_norm": 0.10408841073513031,
      "learning_rate": 0.0002954571184354441,
      "loss": 0.402774840593338,
      "memory(GiB)": 78.26,
      "step": 644,
      "token_acc": 0.8839004707464694,
      "train_speed(iter/s)": 0.032694
    },
    {
      "epoch": 0.12498183403575061,
      "grad_norm": 0.10431966185569763,
      "learning_rate": 0.0002954336090299174,
      "loss": 0.4379886984825134,
      "memory(GiB)": 78.26,
      "step": 645,
      "token_acc": 0.8783049896738991,
      "train_speed(iter/s)": 0.032698
    },
    {
      "epoch": 0.12517560432107736,
      "grad_norm": 0.11031137406826019,
      "learning_rate": 0.0002954100398908995,
      "loss": 0.44032973051071167,
      "memory(GiB)": 78.26,
      "step": 646,
      "token_acc": 0.8727701606958573,
      "train_speed(iter/s)": 0.032702
    },
    {
      "epoch": 0.1253693746064041,
      "grad_norm": 0.11465989798307419,
      "learning_rate": 0.000295386411028071,
      "loss": 0.4586394429206848,
      "memory(GiB)": 78.26,
      "step": 647,
      "token_acc": 0.8680211123783254,
      "train_speed(iter/s)": 0.032705
    },
    {
      "epoch": 0.12556314489173084,
      "grad_norm": 0.10031208395957947,
      "learning_rate": 0.0002953627224511367,
      "loss": 0.4253405034542084,
      "memory(GiB)": 78.26,
      "step": 648,
      "token_acc": 0.8783018191380011,
      "train_speed(iter/s)": 0.032708
    },
    {
      "epoch": 0.1257569151770576,
      "grad_norm": 0.11166791617870331,
      "learning_rate": 0.0002953389741698262,
      "loss": 0.45778605341911316,
      "memory(GiB)": 78.26,
      "step": 649,
      "token_acc": 0.8692058831525433,
      "train_speed(iter/s)": 0.032712
    },
    {
      "epoch": 0.12595068546238433,
      "grad_norm": 0.13080237805843353,
      "learning_rate": 0.0002953151661938937,
      "loss": 0.4996272325515747,
      "memory(GiB)": 78.26,
      "step": 650,
      "token_acc": 0.8591764266135061,
      "train_speed(iter/s)": 0.032716
    },
    {
      "epoch": 0.12614445574771108,
      "grad_norm": 0.1140795424580574,
      "learning_rate": 0.00029529129853311765,
      "loss": 0.4572920799255371,
      "memory(GiB)": 78.26,
      "step": 651,
      "token_acc": 0.8702007294029808,
      "train_speed(iter/s)": 0.03272
    },
    {
      "epoch": 0.12633822603303782,
      "grad_norm": 0.10730257630348206,
      "learning_rate": 0.00029526737119730113,
      "loss": 0.4401112198829651,
      "memory(GiB)": 78.26,
      "step": 652,
      "token_acc": 0.8765418681736673,
      "train_speed(iter/s)": 0.032724
    },
    {
      "epoch": 0.1265319963183646,
      "grad_norm": 0.10705429315567017,
      "learning_rate": 0.0002952433841962718,
      "loss": 0.42639651894569397,
      "memory(GiB)": 78.26,
      "step": 653,
      "token_acc": 0.87889592899469,
      "train_speed(iter/s)": 0.032727
    },
    {
      "epoch": 0.12672576660369134,
      "grad_norm": 0.10336251556873322,
      "learning_rate": 0.0002952193375398817,
      "loss": 0.4257362484931946,
      "memory(GiB)": 78.26,
      "step": 654,
      "token_acc": 0.8773996726028077,
      "train_speed(iter/s)": 0.032731
    },
    {
      "epoch": 0.12691953688901808,
      "grad_norm": 0.11652205884456635,
      "learning_rate": 0.0002951952312380075,
      "loss": 0.4846917986869812,
      "memory(GiB)": 78.26,
      "step": 655,
      "token_acc": 0.8638244781206257,
      "train_speed(iter/s)": 0.032734
    },
    {
      "epoch": 0.12711330717434483,
      "grad_norm": 0.1111086755990982,
      "learning_rate": 0.00029517106530055034,
      "loss": 0.4347532093524933,
      "memory(GiB)": 78.26,
      "step": 656,
      "token_acc": 0.8746101794645484,
      "train_speed(iter/s)": 0.032738
    },
    {
      "epoch": 0.12730707745967157,
      "grad_norm": 0.11007381230592728,
      "learning_rate": 0.0002951468397374357,
      "loss": 0.44459158182144165,
      "memory(GiB)": 78.26,
      "step": 657,
      "token_acc": 0.8722839608558633,
      "train_speed(iter/s)": 0.032742
    },
    {
      "epoch": 0.1275008477449983,
      "grad_norm": 0.10285302996635437,
      "learning_rate": 0.00029512255455861375,
      "loss": 0.3928186297416687,
      "memory(GiB)": 78.26,
      "step": 658,
      "token_acc": 0.8879543480957698,
      "train_speed(iter/s)": 0.032745
    },
    {
      "epoch": 0.12769461803032506,
      "grad_norm": 0.1072237491607666,
      "learning_rate": 0.00029509820977405906,
      "loss": 0.4308614730834961,
      "memory(GiB)": 78.26,
      "step": 659,
      "token_acc": 0.876323788978231,
      "train_speed(iter/s)": 0.032749
    },
    {
      "epoch": 0.1278883883156518,
      "grad_norm": 0.12245503067970276,
      "learning_rate": 0.0002950738053937707,
      "loss": 0.49942547082901,
      "memory(GiB)": 78.26,
      "step": 660,
      "token_acc": 0.8578174506458689,
      "train_speed(iter/s)": 0.032753
    },
    {
      "epoch": 0.12808215860097855,
      "grad_norm": 0.10673464089632034,
      "learning_rate": 0.0002950493414277721,
      "loss": 0.4471551477909088,
      "memory(GiB)": 78.26,
      "step": 661,
      "token_acc": 0.8700834952475975,
      "train_speed(iter/s)": 0.032756
    },
    {
      "epoch": 0.1282759288863053,
      "grad_norm": 0.11559055000543594,
      "learning_rate": 0.0002950248178861114,
      "loss": 0.4593106806278229,
      "memory(GiB)": 78.26,
      "step": 662,
      "token_acc": 0.8703804270157487,
      "train_speed(iter/s)": 0.032761
    },
    {
      "epoch": 0.12846969917163203,
      "grad_norm": 0.11425399780273438,
      "learning_rate": 0.000295000234778861,
      "loss": 0.456471711397171,
      "memory(GiB)": 78.26,
      "step": 663,
      "token_acc": 0.868738057655562,
      "train_speed(iter/s)": 0.032764
    },
    {
      "epoch": 0.12866346945695878,
      "grad_norm": 0.11667713522911072,
      "learning_rate": 0.0002949755921161179,
      "loss": 0.4741016924381256,
      "memory(GiB)": 78.26,
      "step": 664,
      "token_acc": 0.864152462756677,
      "train_speed(iter/s)": 0.032768
    },
    {
      "epoch": 0.12885723974228552,
      "grad_norm": 0.1101280227303505,
      "learning_rate": 0.0002949508899080035,
      "loss": 0.4579119384288788,
      "memory(GiB)": 78.26,
      "step": 665,
      "token_acc": 0.8712681302556577,
      "train_speed(iter/s)": 0.032772
    },
    {
      "epoch": 0.12905101002761227,
      "grad_norm": 0.10604141652584076,
      "learning_rate": 0.0002949261281646636,
      "loss": 0.46181899309158325,
      "memory(GiB)": 78.26,
      "step": 666,
      "token_acc": 0.8664379947229551,
      "train_speed(iter/s)": 0.032775
    },
    {
      "epoch": 0.129244780312939,
      "grad_norm": 0.11616591364145279,
      "learning_rate": 0.0002949013068962685,
      "loss": 0.4584234952926636,
      "memory(GiB)": 78.26,
      "step": 667,
      "token_acc": 0.8690121073872609,
      "train_speed(iter/s)": 0.032778
    },
    {
      "epoch": 0.12943855059826576,
      "grad_norm": 0.11570385843515396,
      "learning_rate": 0.00029487642611301305,
      "loss": 0.4578292965888977,
      "memory(GiB)": 78.26,
      "step": 668,
      "token_acc": 0.8677614050350715,
      "train_speed(iter/s)": 0.032782
    },
    {
      "epoch": 0.1296323208835925,
      "grad_norm": 0.09622353315353394,
      "learning_rate": 0.0002948514858251164,
      "loss": 0.41976287961006165,
      "memory(GiB)": 78.26,
      "step": 669,
      "token_acc": 0.8775399043243499,
      "train_speed(iter/s)": 0.032785
    },
    {
      "epoch": 0.12982609116891924,
      "grad_norm": 0.10756651312112808,
      "learning_rate": 0.0002948264860428223,
      "loss": 0.41081976890563965,
      "memory(GiB)": 78.26,
      "step": 670,
      "token_acc": 0.8800941338537481,
      "train_speed(iter/s)": 0.032789
    },
    {
      "epoch": 0.130019861454246,
      "grad_norm": 0.11921778321266174,
      "learning_rate": 0.00029480142677639864,
      "loss": 0.5204399824142456,
      "memory(GiB)": 78.26,
      "step": 671,
      "token_acc": 0.8542876514459217,
      "train_speed(iter/s)": 0.032792
    },
    {
      "epoch": 0.13021363173957273,
      "grad_norm": 0.10356691479682922,
      "learning_rate": 0.00029477630803613806,
      "loss": 0.4329625368118286,
      "memory(GiB)": 78.26,
      "step": 672,
      "token_acc": 0.8725223288914362,
      "train_speed(iter/s)": 0.032795
    },
    {
      "epoch": 0.13040740202489948,
      "grad_norm": 0.11215896904468536,
      "learning_rate": 0.0002947511298323575,
      "loss": 0.4640699625015259,
      "memory(GiB)": 78.26,
      "step": 673,
      "token_acc": 0.8692367949865711,
      "train_speed(iter/s)": 0.032799
    },
    {
      "epoch": 0.13060117231022622,
      "grad_norm": 0.10207220911979675,
      "learning_rate": 0.0002947258921753983,
      "loss": 0.41831666231155396,
      "memory(GiB)": 78.26,
      "step": 674,
      "token_acc": 0.8800374444184413,
      "train_speed(iter/s)": 0.032802
    },
    {
      "epoch": 0.13079494259555297,
      "grad_norm": 0.11593617498874664,
      "learning_rate": 0.0002947005950756262,
      "loss": 0.4830693006515503,
      "memory(GiB)": 78.26,
      "step": 675,
      "token_acc": 0.8652564297671588,
      "train_speed(iter/s)": 0.032806
    },
    {
      "epoch": 0.1309887128808797,
      "grad_norm": 0.10835679620504379,
      "learning_rate": 0.00029467523854343153,
      "loss": 0.41629844903945923,
      "memory(GiB)": 78.26,
      "step": 676,
      "token_acc": 0.8807510973825394,
      "train_speed(iter/s)": 0.032809
    },
    {
      "epoch": 0.13118248316620645,
      "grad_norm": 0.1169678345322609,
      "learning_rate": 0.00029464982258922874,
      "loss": 0.43599221110343933,
      "memory(GiB)": 78.26,
      "step": 677,
      "token_acc": 0.8760780894853193,
      "train_speed(iter/s)": 0.032813
    },
    {
      "epoch": 0.1313762534515332,
      "grad_norm": 0.10601246356964111,
      "learning_rate": 0.00029462434722345697,
      "loss": 0.43275031447410583,
      "memory(GiB)": 78.26,
      "step": 678,
      "token_acc": 0.8755334681042228,
      "train_speed(iter/s)": 0.032817
    },
    {
      "epoch": 0.13157002373685994,
      "grad_norm": 0.10723179578781128,
      "learning_rate": 0.0002945988124565796,
      "loss": 0.432290643453598,
      "memory(GiB)": 78.26,
      "step": 679,
      "token_acc": 0.8754735278964755,
      "train_speed(iter/s)": 0.03282
    },
    {
      "epoch": 0.1317637940221867,
      "grad_norm": 0.1098715141415596,
      "learning_rate": 0.0002945732182990844,
      "loss": 0.46575891971588135,
      "memory(GiB)": 78.26,
      "step": 680,
      "token_acc": 0.8681277056277056,
      "train_speed(iter/s)": 0.032823
    },
    {
      "epoch": 0.13195756430751343,
      "grad_norm": 0.10466992110013962,
      "learning_rate": 0.0002945475647614836,
      "loss": 0.4427857995033264,
      "memory(GiB)": 78.26,
      "step": 681,
      "token_acc": 0.8734357015159308,
      "train_speed(iter/s)": 0.032827
    },
    {
      "epoch": 0.13215133459284018,
      "grad_norm": 0.10467381030321121,
      "learning_rate": 0.0002945218518543138,
      "loss": 0.4318023920059204,
      "memory(GiB)": 78.26,
      "step": 682,
      "token_acc": 0.8751313343570678,
      "train_speed(iter/s)": 0.03283
    },
    {
      "epoch": 0.13234510487816692,
      "grad_norm": 0.1041102334856987,
      "learning_rate": 0.00029449607958813604,
      "loss": 0.43351155519485474,
      "memory(GiB)": 78.26,
      "step": 683,
      "token_acc": 0.873902149545035,
      "train_speed(iter/s)": 0.032833
    },
    {
      "epoch": 0.1325388751634937,
      "grad_norm": 0.11077916622161865,
      "learning_rate": 0.0002944702479735356,
      "loss": 0.43510738015174866,
      "memory(GiB)": 78.26,
      "step": 684,
      "token_acc": 0.8754168786388559,
      "train_speed(iter/s)": 0.032837
    },
    {
      "epoch": 0.13273264544882044,
      "grad_norm": 0.10889974236488342,
      "learning_rate": 0.0002944443570211223,
      "loss": 0.4274645745754242,
      "memory(GiB)": 78.26,
      "step": 685,
      "token_acc": 0.8780943484353106,
      "train_speed(iter/s)": 0.032841
    },
    {
      "epoch": 0.13292641573414718,
      "grad_norm": 0.10870851576328278,
      "learning_rate": 0.00029441840674153017,
      "loss": 0.47596949338912964,
      "memory(GiB)": 78.26,
      "step": 686,
      "token_acc": 0.863644912544668,
      "train_speed(iter/s)": 0.032844
    },
    {
      "epoch": 0.13312018601947392,
      "grad_norm": 0.11301931738853455,
      "learning_rate": 0.0002943923971454177,
      "loss": 0.46415457129478455,
      "memory(GiB)": 78.26,
      "step": 687,
      "token_acc": 0.8705835224996662,
      "train_speed(iter/s)": 0.032847
    },
    {
      "epoch": 0.13331395630480067,
      "grad_norm": 0.10821959376335144,
      "learning_rate": 0.0002943663282434678,
      "loss": 0.43708160519599915,
      "memory(GiB)": 78.26,
      "step": 688,
      "token_acc": 0.8739687603123969,
      "train_speed(iter/s)": 0.03285
    },
    {
      "epoch": 0.1335077265901274,
      "grad_norm": 0.1074993684887886,
      "learning_rate": 0.0002943402000463875,
      "loss": 0.4149300754070282,
      "memory(GiB)": 78.26,
      "step": 689,
      "token_acc": 0.8803831398155441,
      "train_speed(iter/s)": 0.032854
    },
    {
      "epoch": 0.13370149687545416,
      "grad_norm": 0.11187649518251419,
      "learning_rate": 0.0002943140125649086,
      "loss": 0.4131820797920227,
      "memory(GiB)": 78.26,
      "step": 690,
      "token_acc": 0.8804750593824228,
      "train_speed(iter/s)": 0.032857
    },
    {
      "epoch": 0.1338952671607809,
      "grad_norm": 0.10465361177921295,
      "learning_rate": 0.0002942877658097868,
      "loss": 0.45171868801116943,
      "memory(GiB)": 78.26,
      "step": 691,
      "token_acc": 0.8700957481747215,
      "train_speed(iter/s)": 0.032861
    },
    {
      "epoch": 0.13408903744610765,
      "grad_norm": 0.12278591841459274,
      "learning_rate": 0.00029426145979180243,
      "loss": 0.4629440903663635,
      "memory(GiB)": 78.26,
      "step": 692,
      "token_acc": 0.8690040563923638,
      "train_speed(iter/s)": 0.032864
    },
    {
      "epoch": 0.1342828077314344,
      "grad_norm": 0.10109657794237137,
      "learning_rate": 0.00029423509452176005,
      "loss": 0.42186468839645386,
      "memory(GiB)": 78.26,
      "step": 693,
      "token_acc": 0.877427268729604,
      "train_speed(iter/s)": 0.032868
    },
    {
      "epoch": 0.13447657801676113,
      "grad_norm": 0.10865118354558945,
      "learning_rate": 0.00029420867001048867,
      "loss": 0.4467792510986328,
      "memory(GiB)": 78.26,
      "step": 694,
      "token_acc": 0.8724129751527012,
      "train_speed(iter/s)": 0.032871
    },
    {
      "epoch": 0.13467034830208788,
      "grad_norm": 0.1067957952618599,
      "learning_rate": 0.0002941821862688414,
      "loss": 0.43282240629196167,
      "memory(GiB)": 78.26,
      "step": 695,
      "token_acc": 0.8747978812378032,
      "train_speed(iter/s)": 0.032874
    },
    {
      "epoch": 0.13486411858741462,
      "grad_norm": 0.11119311302900314,
      "learning_rate": 0.00029415564330769595,
      "loss": 0.4267749488353729,
      "memory(GiB)": 78.26,
      "step": 696,
      "token_acc": 0.8750176528738879,
      "train_speed(iter/s)": 0.032877
    },
    {
      "epoch": 0.13505788887274137,
      "grad_norm": 0.1177472174167633,
      "learning_rate": 0.00029412904113795417,
      "loss": 0.4727480709552765,
      "memory(GiB)": 78.26,
      "step": 697,
      "token_acc": 0.8640138408304499,
      "train_speed(iter/s)": 0.032881
    },
    {
      "epoch": 0.1352516591580681,
      "grad_norm": 0.1068907380104065,
      "learning_rate": 0.0002941023797705423,
      "loss": 0.4286806583404541,
      "memory(GiB)": 78.26,
      "step": 698,
      "token_acc": 0.8773055332798717,
      "train_speed(iter/s)": 0.032884
    },
    {
      "epoch": 0.13544542944339485,
      "grad_norm": 0.10560762882232666,
      "learning_rate": 0.00029407565921641093,
      "loss": 0.43469613790512085,
      "memory(GiB)": 78.26,
      "step": 699,
      "token_acc": 0.8766754544193146,
      "train_speed(iter/s)": 0.032888
    },
    {
      "epoch": 0.1356391997287216,
      "grad_norm": 0.11222469061613083,
      "learning_rate": 0.0002940488794865348,
      "loss": 0.49300503730773926,
      "memory(GiB)": 78.26,
      "step": 700,
      "token_acc": 0.8590341029442915,
      "train_speed(iter/s)": 0.032891
    },
    {
      "epoch": 0.13583297001404834,
      "grad_norm": 0.10281093418598175,
      "learning_rate": 0.0002940220405919131,
      "loss": 0.40541547536849976,
      "memory(GiB)": 78.26,
      "step": 701,
      "token_acc": 0.8827319887407821,
      "train_speed(iter/s)": 0.032894
    },
    {
      "epoch": 0.1360267402993751,
      "grad_norm": 0.11664767563343048,
      "learning_rate": 0.00029399514254356936,
      "loss": 0.47185465693473816,
      "memory(GiB)": 78.26,
      "step": 702,
      "token_acc": 0.8657896737890655,
      "train_speed(iter/s)": 0.032897
    },
    {
      "epoch": 0.13622051058470183,
      "grad_norm": 0.11212712526321411,
      "learning_rate": 0.0002939681853525512,
      "loss": 0.44248533248901367,
      "memory(GiB)": 78.26,
      "step": 703,
      "token_acc": 0.8741593492114769,
      "train_speed(iter/s)": 0.032901
    },
    {
      "epoch": 0.13641428087002858,
      "grad_norm": 0.10552475601434708,
      "learning_rate": 0.0002939411690299308,
      "loss": 0.41650229692459106,
      "memory(GiB)": 78.26,
      "step": 704,
      "token_acc": 0.8810408921933085,
      "train_speed(iter/s)": 0.032904
    },
    {
      "epoch": 0.13660805115535532,
      "grad_norm": 0.11409718543291092,
      "learning_rate": 0.0002939140935868044,
      "loss": 0.42940688133239746,
      "memory(GiB)": 78.26,
      "step": 705,
      "token_acc": 0.8761153646882736,
      "train_speed(iter/s)": 0.032908
    },
    {
      "epoch": 0.13680182144068206,
      "grad_norm": 0.1144266277551651,
      "learning_rate": 0.0002938869590342927,
      "loss": 0.4688735604286194,
      "memory(GiB)": 78.26,
      "step": 706,
      "token_acc": 0.8650337555096803,
      "train_speed(iter/s)": 0.032911
    },
    {
      "epoch": 0.1369955917260088,
      "grad_norm": 0.11157376319169998,
      "learning_rate": 0.0002938597653835405,
      "loss": 0.45128777623176575,
      "memory(GiB)": 78.26,
      "step": 707,
      "token_acc": 0.8709400486478752,
      "train_speed(iter/s)": 0.032913
    },
    {
      "epoch": 0.13718936201133555,
      "grad_norm": 0.10916193574666977,
      "learning_rate": 0.000293832512645717,
      "loss": 0.43759334087371826,
      "memory(GiB)": 78.26,
      "step": 708,
      "token_acc": 0.8740179495065283,
      "train_speed(iter/s)": 0.032916
    },
    {
      "epoch": 0.1373831322966623,
      "grad_norm": 0.12670759856700897,
      "learning_rate": 0.00029380520083201563,
      "loss": 0.49803271889686584,
      "memory(GiB)": 78.26,
      "step": 709,
      "token_acc": 0.8560126128708614,
      "train_speed(iter/s)": 0.03292
    },
    {
      "epoch": 0.13757690258198904,
      "grad_norm": 0.11388320475816727,
      "learning_rate": 0.00029377782995365404,
      "loss": 0.43962639570236206,
      "memory(GiB)": 78.26,
      "step": 710,
      "token_acc": 0.8743197278911564,
      "train_speed(iter/s)": 0.032923
    },
    {
      "epoch": 0.13777067286731579,
      "grad_norm": 0.116911880671978,
      "learning_rate": 0.0002937504000218743,
      "loss": 0.46620821952819824,
      "memory(GiB)": 78.26,
      "step": 711,
      "token_acc": 0.8658806794868288,
      "train_speed(iter/s)": 0.032926
    },
    {
      "epoch": 0.13796444315264253,
      "grad_norm": 0.10047486424446106,
      "learning_rate": 0.0002937229110479425,
      "loss": 0.411388635635376,
      "memory(GiB)": 78.26,
      "step": 712,
      "token_acc": 0.8812828160820372,
      "train_speed(iter/s)": 0.032929
    },
    {
      "epoch": 0.13815821343796927,
      "grad_norm": 0.10067104548215866,
      "learning_rate": 0.00029369536304314916,
      "loss": 0.4194115698337555,
      "memory(GiB)": 78.26,
      "step": 713,
      "token_acc": 0.8785104275426338,
      "train_speed(iter/s)": 0.032932
    },
    {
      "epoch": 0.13835198372329605,
      "grad_norm": 0.10838618874549866,
      "learning_rate": 0.0002936677560188089,
      "loss": 0.46899646520614624,
      "memory(GiB)": 78.26,
      "step": 714,
      "token_acc": 0.8667539476848265,
      "train_speed(iter/s)": 0.032935
    },
    {
      "epoch": 0.1385457540086228,
      "grad_norm": 0.12074261903762817,
      "learning_rate": 0.00029364008998626086,
      "loss": 0.41880735754966736,
      "memory(GiB)": 78.26,
      "step": 715,
      "token_acc": 0.8823898581649303,
      "train_speed(iter/s)": 0.032938
    },
    {
      "epoch": 0.13873952429394953,
      "grad_norm": 0.11350797116756439,
      "learning_rate": 0.00029361236495686806,
      "loss": 0.4601813554763794,
      "memory(GiB)": 78.26,
      "step": 716,
      "token_acc": 0.8699107434095424,
      "train_speed(iter/s)": 0.032942
    },
    {
      "epoch": 0.13893329457927628,
      "grad_norm": 0.11661586165428162,
      "learning_rate": 0.0002935845809420179,
      "loss": 0.4675025939941406,
      "memory(GiB)": 78.26,
      "step": 717,
      "token_acc": 0.8649476228847703,
      "train_speed(iter/s)": 0.032945
    },
    {
      "epoch": 0.13912706486460302,
      "grad_norm": 0.11661852151155472,
      "learning_rate": 0.0002935567379531222,
      "loss": 0.47289931774139404,
      "memory(GiB)": 78.26,
      "step": 718,
      "token_acc": 0.8682559598494354,
      "train_speed(iter/s)": 0.032949
    },
    {
      "epoch": 0.13932083514992977,
      "grad_norm": 0.11107414215803146,
      "learning_rate": 0.0002935288360016166,
      "loss": 0.4379945397377014,
      "memory(GiB)": 78.26,
      "step": 719,
      "token_acc": 0.8734506763257683,
      "train_speed(iter/s)": 0.032952
    },
    {
      "epoch": 0.1395146054352565,
      "grad_norm": 0.1156369224190712,
      "learning_rate": 0.00029350087509896137,
      "loss": 0.4373580515384674,
      "memory(GiB)": 78.26,
      "step": 720,
      "token_acc": 0.8751508606999936,
      "train_speed(iter/s)": 0.032954
    },
    {
      "epoch": 0.13970837572058326,
      "grad_norm": 0.11756598204374313,
      "learning_rate": 0.00029347285525664065,
      "loss": 0.48482561111450195,
      "memory(GiB)": 78.26,
      "step": 721,
      "token_acc": 0.8630889444773964,
      "train_speed(iter/s)": 0.032958
    },
    {
      "epoch": 0.13990214600591,
      "grad_norm": 0.10152934491634369,
      "learning_rate": 0.00029344477648616304,
      "loss": 0.41498956084251404,
      "memory(GiB)": 78.26,
      "step": 722,
      "token_acc": 0.881353398445286,
      "train_speed(iter/s)": 0.032961
    },
    {
      "epoch": 0.14009591629123674,
      "grad_norm": 0.11098542809486389,
      "learning_rate": 0.0002934166387990612,
      "loss": 0.4146006107330322,
      "memory(GiB)": 78.26,
      "step": 723,
      "token_acc": 0.8806968679173152,
      "train_speed(iter/s)": 0.032963
    },
    {
      "epoch": 0.1402896865765635,
      "grad_norm": 0.11475761979818344,
      "learning_rate": 0.00029338844220689204,
      "loss": 0.4258284270763397,
      "memory(GiB)": 78.26,
      "step": 724,
      "token_acc": 0.8758714403875694,
      "train_speed(iter/s)": 0.032967
    },
    {
      "epoch": 0.14048345686189023,
      "grad_norm": 0.10699902474880219,
      "learning_rate": 0.0002933601867212367,
      "loss": 0.4143221974372864,
      "memory(GiB)": 78.26,
      "step": 725,
      "token_acc": 0.8792861274064282,
      "train_speed(iter/s)": 0.03297
    },
    {
      "epoch": 0.14067722714721698,
      "grad_norm": 0.11193379759788513,
      "learning_rate": 0.0002933318723537004,
      "loss": 0.4646386504173279,
      "memory(GiB)": 78.26,
      "step": 726,
      "token_acc": 0.8705041600375577,
      "train_speed(iter/s)": 0.032973
    },
    {
      "epoch": 0.14087099743254372,
      "grad_norm": 0.10847294330596924,
      "learning_rate": 0.0002933034991159127,
      "loss": 0.431307852268219,
      "memory(GiB)": 78.26,
      "step": 727,
      "token_acc": 0.8776439295122638,
      "train_speed(iter/s)": 0.032976
    },
    {
      "epoch": 0.14106476771787047,
      "grad_norm": 0.11553511023521423,
      "learning_rate": 0.0002932750670195272,
      "loss": 0.47172775864601135,
      "memory(GiB)": 78.26,
      "step": 728,
      "token_acc": 0.8666254148619003,
      "train_speed(iter/s)": 0.032979
    },
    {
      "epoch": 0.1412585380031972,
      "grad_norm": 0.12142791599035263,
      "learning_rate": 0.0002932465760762217,
      "loss": 0.4737962484359741,
      "memory(GiB)": 78.26,
      "step": 729,
      "token_acc": 0.8672086720867209,
      "train_speed(iter/s)": 0.032982
    },
    {
      "epoch": 0.14145230828852395,
      "grad_norm": 0.11232297122478485,
      "learning_rate": 0.0002932180262976982,
      "loss": 0.4628666341304779,
      "memory(GiB)": 78.26,
      "step": 730,
      "token_acc": 0.8675850891410049,
      "train_speed(iter/s)": 0.032985
    },
    {
      "epoch": 0.1416460785738507,
      "grad_norm": 0.11426133662462234,
      "learning_rate": 0.0002931894176956829,
      "loss": 0.48010843992233276,
      "memory(GiB)": 78.26,
      "step": 731,
      "token_acc": 0.8621309527436702,
      "train_speed(iter/s)": 0.032988
    },
    {
      "epoch": 0.14183984885917744,
      "grad_norm": 0.10444355756044388,
      "learning_rate": 0.0002931607502819261,
      "loss": 0.4301010072231293,
      "memory(GiB)": 78.26,
      "step": 732,
      "token_acc": 0.8748947870489872,
      "train_speed(iter/s)": 0.032991
    },
    {
      "epoch": 0.1420336191445042,
      "grad_norm": 0.10665851086378098,
      "learning_rate": 0.0002931320240682023,
      "loss": 0.4206262230873108,
      "memory(GiB)": 78.26,
      "step": 733,
      "token_acc": 0.876729803932444,
      "train_speed(iter/s)": 0.032994
    },
    {
      "epoch": 0.14222738942983093,
      "grad_norm": 0.10180269926786423,
      "learning_rate": 0.00029310323906631006,
      "loss": 0.41888511180877686,
      "memory(GiB)": 78.26,
      "step": 734,
      "token_acc": 0.8802506609223538,
      "train_speed(iter/s)": 0.032997
    },
    {
      "epoch": 0.14242115971515767,
      "grad_norm": 0.109284408390522,
      "learning_rate": 0.00029307439528807223,
      "loss": 0.45739686489105225,
      "memory(GiB)": 78.26,
      "step": 735,
      "token_acc": 0.8659565024411895,
      "train_speed(iter/s)": 0.033
    },
    {
      "epoch": 0.14261493000048442,
      "grad_norm": 0.11013077199459076,
      "learning_rate": 0.0002930454927453357,
      "loss": 0.45072418451309204,
      "memory(GiB)": 78.26,
      "step": 736,
      "token_acc": 0.8717199543645825,
      "train_speed(iter/s)": 0.033003
    },
    {
      "epoch": 0.14280870028581116,
      "grad_norm": 0.11075732111930847,
      "learning_rate": 0.00029301653144997154,
      "loss": 0.45079749822616577,
      "memory(GiB)": 78.26,
      "step": 737,
      "token_acc": 0.8688188438181866,
      "train_speed(iter/s)": 0.033006
    },
    {
      "epoch": 0.1430024705711379,
      "grad_norm": 0.12135940045118332,
      "learning_rate": 0.0002929875114138749,
      "loss": 0.4898369312286377,
      "memory(GiB)": 78.26,
      "step": 738,
      "token_acc": 0.8604469273743017,
      "train_speed(iter/s)": 0.033009
    },
    {
      "epoch": 0.14319624085646465,
      "grad_norm": 0.10148437321186066,
      "learning_rate": 0.00029295843264896506,
      "loss": 0.37936070561408997,
      "memory(GiB)": 78.26,
      "step": 739,
      "token_acc": 0.890900732946855,
      "train_speed(iter/s)": 0.033011
    },
    {
      "epoch": 0.1433900111417914,
      "grad_norm": 0.10243412852287292,
      "learning_rate": 0.0002929292951671855,
      "loss": 0.42304426431655884,
      "memory(GiB)": 78.26,
      "step": 740,
      "token_acc": 0.878807962695576,
      "train_speed(iter/s)": 0.033014
    },
    {
      "epoch": 0.14358378142711814,
      "grad_norm": 0.10791970789432526,
      "learning_rate": 0.0002929000989805038,
      "loss": 0.42843425273895264,
      "memory(GiB)": 78.26,
      "step": 741,
      "token_acc": 0.8764322681978526,
      "train_speed(iter/s)": 0.033017
    },
    {
      "epoch": 0.14377755171244488,
      "grad_norm": 0.11472160369157791,
      "learning_rate": 0.00029287084410091154,
      "loss": 0.4652520716190338,
      "memory(GiB)": 78.26,
      "step": 742,
      "token_acc": 0.8662527909382999,
      "train_speed(iter/s)": 0.03302
    },
    {
      "epoch": 0.14397132199777163,
      "grad_norm": 0.12771129608154297,
      "learning_rate": 0.00029284153054042454,
      "loss": 0.5056222081184387,
      "memory(GiB)": 78.26,
      "step": 743,
      "token_acc": 0.8580256694660118,
      "train_speed(iter/s)": 0.033023
    },
    {
      "epoch": 0.1441650922830984,
      "grad_norm": 0.10892749577760696,
      "learning_rate": 0.0002928121583110826,
      "loss": 0.4871826469898224,
      "memory(GiB)": 78.26,
      "step": 744,
      "token_acc": 0.8604575074297244,
      "train_speed(iter/s)": 0.033026
    },
    {
      "epoch": 0.14435886256842514,
      "grad_norm": 0.10514423251152039,
      "learning_rate": 0.0002927827274249498,
      "loss": 0.4291183352470398,
      "memory(GiB)": 78.26,
      "step": 745,
      "token_acc": 0.8779703110007337,
      "train_speed(iter/s)": 0.033029
    },
    {
      "epoch": 0.1445526328537519,
      "grad_norm": 0.11185097694396973,
      "learning_rate": 0.00029275323789411406,
      "loss": 0.4657382667064667,
      "memory(GiB)": 78.26,
      "step": 746,
      "token_acc": 0.8690815142887272,
      "train_speed(iter/s)": 0.033031
    },
    {
      "epoch": 0.14474640313907863,
      "grad_norm": 0.1169874519109726,
      "learning_rate": 0.00029272368973068765,
      "loss": 0.4796285331249237,
      "memory(GiB)": 78.26,
      "step": 747,
      "token_acc": 0.8630539484445522,
      "train_speed(iter/s)": 0.033034
    },
    {
      "epoch": 0.14494017342440538,
      "grad_norm": 0.10806058347225189,
      "learning_rate": 0.0002926940829468067,
      "loss": 0.44982171058654785,
      "memory(GiB)": 78.26,
      "step": 748,
      "token_acc": 0.8720634405109884,
      "train_speed(iter/s)": 0.033037
    },
    {
      "epoch": 0.14513394370973212,
      "grad_norm": 0.10634516179561615,
      "learning_rate": 0.00029266441755463154,
      "loss": 0.4241579473018646,
      "memory(GiB)": 78.26,
      "step": 749,
      "token_acc": 0.8774991483898016,
      "train_speed(iter/s)": 0.03304
    },
    {
      "epoch": 0.14532771399505887,
      "grad_norm": 0.10630025714635849,
      "learning_rate": 0.00029263469356634656,
      "loss": 0.45012885332107544,
      "memory(GiB)": 78.26,
      "step": 750,
      "token_acc": 0.8703159258521028,
      "train_speed(iter/s)": 0.033043
    },
    {
      "epoch": 0.1455214842803856,
      "grad_norm": 0.11359134316444397,
      "learning_rate": 0.0002926049109941602,
      "loss": 0.47736871242523193,
      "memory(GiB)": 78.26,
      "step": 751,
      "token_acc": 0.8675646281268301,
      "train_speed(iter/s)": 0.033046
    },
    {
      "epoch": 0.14571525456571235,
      "grad_norm": 0.10710503160953522,
      "learning_rate": 0.00029257506985030495,
      "loss": 0.4275802671909332,
      "memory(GiB)": 78.26,
      "step": 752,
      "token_acc": 0.8778828604612576,
      "train_speed(iter/s)": 0.033048
    },
    {
      "epoch": 0.1459090248510391,
      "grad_norm": 0.10511702299118042,
      "learning_rate": 0.00029254517014703737,
      "loss": 0.41856449842453003,
      "memory(GiB)": 78.26,
      "step": 753,
      "token_acc": 0.8796367161589878,
      "train_speed(iter/s)": 0.033051
    },
    {
      "epoch": 0.14610279513636584,
      "grad_norm": 0.10652240365743637,
      "learning_rate": 0.000292515211896638,
      "loss": 0.4258817732334137,
      "memory(GiB)": 78.26,
      "step": 754,
      "token_acc": 0.8783486686751893,
      "train_speed(iter/s)": 0.033054
    },
    {
      "epoch": 0.1462965654216926,
      "grad_norm": 0.10977134853601456,
      "learning_rate": 0.00029248519511141166,
      "loss": 0.45545950531959534,
      "memory(GiB)": 78.26,
      "step": 755,
      "token_acc": 0.8708563737967758,
      "train_speed(iter/s)": 0.033057
    },
    {
      "epoch": 0.14649033570701933,
      "grad_norm": 0.10239318013191223,
      "learning_rate": 0.00029245511980368694,
      "loss": 0.40798917412757874,
      "memory(GiB)": 78.26,
      "step": 756,
      "token_acc": 0.8818212167124625,
      "train_speed(iter/s)": 0.033059
    },
    {
      "epoch": 0.14668410599234608,
      "grad_norm": 0.10805483907461166,
      "learning_rate": 0.0002924249859858166,
      "loss": 0.44618624448776245,
      "memory(GiB)": 78.26,
      "step": 757,
      "token_acc": 0.873049779628478,
      "train_speed(iter/s)": 0.033062
    },
    {
      "epoch": 0.14687787627767282,
      "grad_norm": 0.09752228856086731,
      "learning_rate": 0.0002923947936701774,
      "loss": 0.4019618034362793,
      "memory(GiB)": 78.26,
      "step": 758,
      "token_acc": 0.8834425762325449,
      "train_speed(iter/s)": 0.033065
    },
    {
      "epoch": 0.14707164656299956,
      "grad_norm": 0.10851308703422546,
      "learning_rate": 0.00029236454286917017,
      "loss": 0.4336164891719818,
      "memory(GiB)": 78.26,
      "step": 759,
      "token_acc": 0.8760058721183123,
      "train_speed(iter/s)": 0.033067
    },
    {
      "epoch": 0.1472654168483263,
      "grad_norm": 0.10959117859601974,
      "learning_rate": 0.00029233423359521966,
      "loss": 0.4287548363208771,
      "memory(GiB)": 78.26,
      "step": 760,
      "token_acc": 0.874751174456565,
      "train_speed(iter/s)": 0.033069
    },
    {
      "epoch": 0.14745918713365305,
      "grad_norm": 0.11091278493404388,
      "learning_rate": 0.0002923038658607748,
      "loss": 0.44881102442741394,
      "memory(GiB)": 78.26,
      "step": 761,
      "token_acc": 0.8713587887777806,
      "train_speed(iter/s)": 0.033072
    },
    {
      "epoch": 0.1476529574189798,
      "grad_norm": 0.11895415931940079,
      "learning_rate": 0.0002922734396783083,
      "loss": 0.49504998326301575,
      "memory(GiB)": 78.26,
      "step": 762,
      "token_acc": 0.8594216543375924,
      "train_speed(iter/s)": 0.033075
    },
    {
      "epoch": 0.14784672770430654,
      "grad_norm": 0.09505145251750946,
      "learning_rate": 0.00029224295506031714,
      "loss": 0.38731294870376587,
      "memory(GiB)": 78.26,
      "step": 763,
      "token_acc": 0.8887887484287171,
      "train_speed(iter/s)": 0.033077
    },
    {
      "epoch": 0.14804049798963329,
      "grad_norm": 0.10213866084814072,
      "learning_rate": 0.0002922124120193221,
      "loss": 0.4288552403450012,
      "memory(GiB)": 78.26,
      "step": 764,
      "token_acc": 0.8762533946104032,
      "train_speed(iter/s)": 0.033079
    },
    {
      "epoch": 0.14823426827496003,
      "grad_norm": 0.11133123189210892,
      "learning_rate": 0.00029218181056786806,
      "loss": 0.46136656403541565,
      "memory(GiB)": 78.26,
      "step": 765,
      "token_acc": 0.8676832940918797,
      "train_speed(iter/s)": 0.033082
    },
    {
      "epoch": 0.14842803856028677,
      "grad_norm": 0.1136741042137146,
      "learning_rate": 0.00029215115071852386,
      "loss": 0.44833481311798096,
      "memory(GiB)": 78.26,
      "step": 766,
      "token_acc": 0.8728677891250345,
      "train_speed(iter/s)": 0.033085
    },
    {
      "epoch": 0.14862180884561352,
      "grad_norm": 0.1057792380452156,
      "learning_rate": 0.0002921204324838823,
      "loss": 0.4067574441432953,
      "memory(GiB)": 78.26,
      "step": 767,
      "token_acc": 0.883631027963543,
      "train_speed(iter/s)": 0.033088
    },
    {
      "epoch": 0.14881557913094026,
      "grad_norm": 0.12142736464738846,
      "learning_rate": 0.0002920896558765602,
      "loss": 0.4738287031650543,
      "memory(GiB)": 78.26,
      "step": 768,
      "token_acc": 0.8667360749609578,
      "train_speed(iter/s)": 0.033091
    },
    {
      "epoch": 0.149009349416267,
      "grad_norm": 0.12040547281503677,
      "learning_rate": 0.0002920588209091983,
      "loss": 0.43065720796585083,
      "memory(GiB)": 78.26,
      "step": 769,
      "token_acc": 0.8761208605841793,
      "train_speed(iter/s)": 0.033094
    },
    {
      "epoch": 0.14920311970159375,
      "grad_norm": 0.09921874105930328,
      "learning_rate": 0.0002920279275944614,
      "loss": 0.43212834000587463,
      "memory(GiB)": 78.26,
      "step": 770,
      "token_acc": 0.8759197564892266,
      "train_speed(iter/s)": 0.033096
    },
    {
      "epoch": 0.1493968899869205,
      "grad_norm": 0.141917884349823,
      "learning_rate": 0.0002919969759450382,
      "loss": 0.4761424660682678,
      "memory(GiB)": 78.26,
      "step": 771,
      "token_acc": 0.8653158180933573,
      "train_speed(iter/s)": 0.033099
    },
    {
      "epoch": 0.14959066027224724,
      "grad_norm": 0.1117415577173233,
      "learning_rate": 0.0002919659659736414,
      "loss": 0.4318593144416809,
      "memory(GiB)": 78.26,
      "step": 772,
      "token_acc": 0.8758288466987815,
      "train_speed(iter/s)": 0.033101
    },
    {
      "epoch": 0.14978443055757398,
      "grad_norm": 0.16481183469295502,
      "learning_rate": 0.00029193489769300754,
      "loss": 0.4566386044025421,
      "memory(GiB)": 78.26,
      "step": 773,
      "token_acc": 0.8704999401985408,
      "train_speed(iter/s)": 0.033104
    },
    {
      "epoch": 0.14997820084290073,
      "grad_norm": 0.10933414846658707,
      "learning_rate": 0.0002919037711158973,
      "loss": 0.430907666683197,
      "memory(GiB)": 78.26,
      "step": 774,
      "token_acc": 0.8769761425697039,
      "train_speed(iter/s)": 0.033107
    },
    {
      "epoch": 0.1501719711282275,
      "grad_norm": 0.1175195574760437,
      "learning_rate": 0.00029187258625509513,
      "loss": 0.45036518573760986,
      "memory(GiB)": 78.26,
      "step": 775,
      "token_acc": 0.869928163775235,
      "train_speed(iter/s)": 0.03311
    },
    {
      "epoch": 0.15036574141355424,
      "grad_norm": 0.1023663803935051,
      "learning_rate": 0.0002918413431234096,
      "loss": 0.39650124311447144,
      "memory(GiB)": 78.26,
      "step": 776,
      "token_acc": 0.8873017168666504,
      "train_speed(iter/s)": 0.033112
    },
    {
      "epoch": 0.150559511698881,
      "grad_norm": 0.12010087072849274,
      "learning_rate": 0.000291810041733673,
      "loss": 0.4465680420398712,
      "memory(GiB)": 78.26,
      "step": 777,
      "token_acc": 0.8728560775540641,
      "train_speed(iter/s)": 0.033115
    },
    {
      "epoch": 0.15075328198420773,
      "grad_norm": 0.10836753249168396,
      "learning_rate": 0.0002917786820987416,
      "loss": 0.43890365958213806,
      "memory(GiB)": 78.26,
      "step": 778,
      "token_acc": 0.8746218049034951,
      "train_speed(iter/s)": 0.033118
    },
    {
      "epoch": 0.15094705226953448,
      "grad_norm": 0.11870795488357544,
      "learning_rate": 0.0002917472642314958,
      "loss": 0.46871399879455566,
      "memory(GiB)": 78.26,
      "step": 779,
      "token_acc": 0.867821888084105,
      "train_speed(iter/s)": 0.033121
    },
    {
      "epoch": 0.15114082255486122,
      "grad_norm": 0.1149909496307373,
      "learning_rate": 0.00029171578814483966,
      "loss": 0.44199684262275696,
      "memory(GiB)": 78.26,
      "step": 780,
      "token_acc": 0.8727175418713008,
      "train_speed(iter/s)": 0.033124
    },
    {
      "epoch": 0.15133459284018796,
      "grad_norm": 0.1077098697423935,
      "learning_rate": 0.0002916842538517013,
      "loss": 0.4246280789375305,
      "memory(GiB)": 78.26,
      "step": 781,
      "token_acc": 0.8763708602062753,
      "train_speed(iter/s)": 0.033126
    },
    {
      "epoch": 0.1515283631255147,
      "grad_norm": 0.11270920932292938,
      "learning_rate": 0.0002916526613650326,
      "loss": 0.4364209771156311,
      "memory(GiB)": 78.26,
      "step": 782,
      "token_acc": 0.8724545624382823,
      "train_speed(iter/s)": 0.033129
    },
    {
      "epoch": 0.15172213341084145,
      "grad_norm": 0.10723229497671127,
      "learning_rate": 0.0002916210106978096,
      "loss": 0.47758546471595764,
      "memory(GiB)": 78.26,
      "step": 783,
      "token_acc": 0.8667694833461854,
      "train_speed(iter/s)": 0.033132
    },
    {
      "epoch": 0.1519159036961682,
      "grad_norm": 0.12354382872581482,
      "learning_rate": 0.000291589301863032,
      "loss": 0.4946221709251404,
      "memory(GiB)": 78.26,
      "step": 784,
      "token_acc": 0.8581249637407902,
      "train_speed(iter/s)": 0.033134
    },
    {
      "epoch": 0.15210967398149494,
      "grad_norm": 0.11447127163410187,
      "learning_rate": 0.00029155753487372345,
      "loss": 0.44242554903030396,
      "memory(GiB)": 78.26,
      "step": 785,
      "token_acc": 0.8728641926006219,
      "train_speed(iter/s)": 0.033137
    },
    {
      "epoch": 0.15230344426682169,
      "grad_norm": 0.10308409482240677,
      "learning_rate": 0.0002915257097429315,
      "loss": 0.4122265577316284,
      "memory(GiB)": 78.26,
      "step": 786,
      "token_acc": 0.881477327759794,
      "train_speed(iter/s)": 0.033139
    },
    {
      "epoch": 0.15249721455214843,
      "grad_norm": 0.10119541734457016,
      "learning_rate": 0.00029149382648372763,
      "loss": 0.3953108787536621,
      "memory(GiB)": 78.26,
      "step": 787,
      "token_acc": 0.8823500826124988,
      "train_speed(iter/s)": 0.033141
    },
    {
      "epoch": 0.15269098483747517,
      "grad_norm": 0.10768993198871613,
      "learning_rate": 0.0002914618851092072,
      "loss": 0.4354395270347595,
      "memory(GiB)": 78.26,
      "step": 788,
      "token_acc": 0.8756917373170133,
      "train_speed(iter/s)": 0.033144
    },
    {
      "epoch": 0.15288475512280192,
      "grad_norm": 0.10865043103694916,
      "learning_rate": 0.0002914298856324893,
      "loss": 0.40121009945869446,
      "memory(GiB)": 78.26,
      "step": 789,
      "token_acc": 0.886211232187762,
      "train_speed(iter/s)": 0.033147
    },
    {
      "epoch": 0.15307852540812866,
      "grad_norm": 0.1131611317396164,
      "learning_rate": 0.00029139782806671696,
      "loss": 0.41886383295059204,
      "memory(GiB)": 78.26,
      "step": 790,
      "token_acc": 0.8795520757465404,
      "train_speed(iter/s)": 0.03315
    },
    {
      "epoch": 0.1532722956934554,
      "grad_norm": 0.11826243251562119,
      "learning_rate": 0.0002913657124250571,
      "loss": 0.4526827037334442,
      "memory(GiB)": 78.26,
      "step": 791,
      "token_acc": 0.8706143245597825,
      "train_speed(iter/s)": 0.033152
    },
    {
      "epoch": 0.15346606597878215,
      "grad_norm": 0.12152021378278732,
      "learning_rate": 0.0002913335387207006,
      "loss": 0.4480191767215729,
      "memory(GiB)": 78.26,
      "step": 792,
      "token_acc": 0.8714323428470857,
      "train_speed(iter/s)": 0.033155
    },
    {
      "epoch": 0.1536598362641089,
      "grad_norm": 0.10104276239871979,
      "learning_rate": 0.0002913013069668619,
      "loss": 0.41667938232421875,
      "memory(GiB)": 78.26,
      "step": 793,
      "token_acc": 0.880495535822304,
      "train_speed(iter/s)": 0.033158
    },
    {
      "epoch": 0.15385360654943564,
      "grad_norm": 0.1120065450668335,
      "learning_rate": 0.0002912690171767795,
      "loss": 0.444894015789032,
      "memory(GiB)": 78.26,
      "step": 794,
      "token_acc": 0.8719184281987115,
      "train_speed(iter/s)": 0.033161
    },
    {
      "epoch": 0.15404737683476238,
      "grad_norm": 0.10677407681941986,
      "learning_rate": 0.00029123666936371577,
      "loss": 0.4287039041519165,
      "memory(GiB)": 78.26,
      "step": 795,
      "token_acc": 0.8766376159685128,
      "train_speed(iter/s)": 0.033163
    },
    {
      "epoch": 0.15424114712008913,
      "grad_norm": 0.10139796137809753,
      "learning_rate": 0.0002912042635409568,
      "loss": 0.3969685435295105,
      "memory(GiB)": 78.26,
      "step": 796,
      "token_acc": 0.8835710646968925,
      "train_speed(iter/s)": 0.033166
    },
    {
      "epoch": 0.15443491740541587,
      "grad_norm": 0.11726026982069016,
      "learning_rate": 0.0002911717997218123,
      "loss": 0.45599818229675293,
      "memory(GiB)": 78.26,
      "step": 797,
      "token_acc": 0.8708311582854823,
      "train_speed(iter/s)": 0.033169
    },
    {
      "epoch": 0.15462868769074262,
      "grad_norm": 0.10516592860221863,
      "learning_rate": 0.0002911392779196164,
      "loss": 0.39878836274147034,
      "memory(GiB)": 78.26,
      "step": 798,
      "token_acc": 0.8827302498037836,
      "train_speed(iter/s)": 0.033171
    },
    {
      "epoch": 0.15482245797606936,
      "grad_norm": 0.1172914132475853,
      "learning_rate": 0.00029110669814772644,
      "loss": 0.43762531876564026,
      "memory(GiB)": 78.26,
      "step": 799,
      "token_acc": 0.8754726710209694,
      "train_speed(iter/s)": 0.033174
    },
    {
      "epoch": 0.1550162282613961,
      "grad_norm": 0.1156349629163742,
      "learning_rate": 0.0002910740604195238,
      "loss": 0.46107926964759827,
      "memory(GiB)": 78.26,
      "step": 800,
      "token_acc": 0.869101935557004,
      "train_speed(iter/s)": 0.033176
    },
    {
      "epoch": 0.15520999854672285,
      "grad_norm": 0.1091204434633255,
      "learning_rate": 0.00029104136474841384,
      "loss": 0.4546446204185486,
      "memory(GiB)": 78.26,
      "step": 801,
      "token_acc": 0.8690357722725606,
      "train_speed(iter/s)": 0.033161
    },
    {
      "epoch": 0.1554037688320496,
      "grad_norm": 0.10520661622285843,
      "learning_rate": 0.00029100861114782537,
      "loss": 0.391659677028656,
      "memory(GiB)": 78.26,
      "step": 802,
      "token_acc": 0.883874415497662,
      "train_speed(iter/s)": 0.033164
    },
    {
      "epoch": 0.15559753911737634,
      "grad_norm": 0.10649837553501129,
      "learning_rate": 0.0002909757996312113,
      "loss": 0.4287815988063812,
      "memory(GiB)": 78.26,
      "step": 803,
      "token_acc": 0.8760036358127556,
      "train_speed(iter/s)": 0.033167
    },
    {
      "epoch": 0.15579130940270308,
      "grad_norm": 0.11375483870506287,
      "learning_rate": 0.00029094293021204816,
      "loss": 0.43256044387817383,
      "memory(GiB)": 78.26,
      "step": 804,
      "token_acc": 0.8724521560241453,
      "train_speed(iter/s)": 0.033169
    },
    {
      "epoch": 0.15598507968802985,
      "grad_norm": 0.1098484992980957,
      "learning_rate": 0.00029091000290383626,
      "loss": 0.44678372144699097,
      "memory(GiB)": 78.26,
      "step": 805,
      "token_acc": 0.8710053046965972,
      "train_speed(iter/s)": 0.033171
    },
    {
      "epoch": 0.1561788499733566,
      "grad_norm": 0.12109239399433136,
      "learning_rate": 0.0002908770177200998,
      "loss": 0.43375393748283386,
      "memory(GiB)": 78.26,
      "step": 806,
      "token_acc": 0.8734656356459201,
      "train_speed(iter/s)": 0.033174
    },
    {
      "epoch": 0.15637262025868334,
      "grad_norm": 0.1153409481048584,
      "learning_rate": 0.00029084397467438666,
      "loss": 0.4369393587112427,
      "memory(GiB)": 78.26,
      "step": 807,
      "token_acc": 0.8766715060747307,
      "train_speed(iter/s)": 0.033176
    },
    {
      "epoch": 0.1565663905440101,
      "grad_norm": 0.11903059482574463,
      "learning_rate": 0.0002908108737802685,
      "loss": 0.4392112195491791,
      "memory(GiB)": 78.26,
      "step": 808,
      "token_acc": 0.8715450323167365,
      "train_speed(iter/s)": 0.033178
    },
    {
      "epoch": 0.15676016082933683,
      "grad_norm": 0.10243628919124603,
      "learning_rate": 0.0002907777150513407,
      "loss": 0.40879738330841064,
      "memory(GiB)": 78.26,
      "step": 809,
      "token_acc": 0.8799474375821288,
      "train_speed(iter/s)": 0.033181
    },
    {
      "epoch": 0.15695393111466358,
      "grad_norm": 0.1134597584605217,
      "learning_rate": 0.0002907444985012225,
      "loss": 0.4965320825576782,
      "memory(GiB)": 78.26,
      "step": 810,
      "token_acc": 0.8585183351214547,
      "train_speed(iter/s)": 0.033183
    },
    {
      "epoch": 0.15714770139999032,
      "grad_norm": 0.11033570021390915,
      "learning_rate": 0.0002907112241435568,
      "loss": 0.46451640129089355,
      "memory(GiB)": 78.26,
      "step": 811,
      "token_acc": 0.8680681696762474,
      "train_speed(iter/s)": 0.033186
    },
    {
      "epoch": 0.15734147168531706,
      "grad_norm": 0.10630565881729126,
      "learning_rate": 0.0002906778919920103,
      "loss": 0.4418585002422333,
      "memory(GiB)": 78.26,
      "step": 812,
      "token_acc": 0.8757123405113723,
      "train_speed(iter/s)": 0.033188
    },
    {
      "epoch": 0.1575352419706438,
      "grad_norm": 0.10460490733385086,
      "learning_rate": 0.0002906445020602734,
      "loss": 0.4312971532344818,
      "memory(GiB)": 78.26,
      "step": 813,
      "token_acc": 0.8766326100695175,
      "train_speed(iter/s)": 0.03319
    },
    {
      "epoch": 0.15772901225597055,
      "grad_norm": 0.09658509492874146,
      "learning_rate": 0.0002906110543620603,
      "loss": 0.3654577136039734,
      "memory(GiB)": 78.26,
      "step": 814,
      "token_acc": 0.894104601737202,
      "train_speed(iter/s)": 0.033193
    },
    {
      "epoch": 0.1579227825412973,
      "grad_norm": 0.11015032976865768,
      "learning_rate": 0.0002905775489111087,
      "loss": 0.4334821403026581,
      "memory(GiB)": 78.26,
      "step": 815,
      "token_acc": 0.8754580788450861,
      "train_speed(iter/s)": 0.033195
    },
    {
      "epoch": 0.15811655282662404,
      "grad_norm": 0.10890354961156845,
      "learning_rate": 0.0002905439857211804,
      "loss": 0.4261341392993927,
      "memory(GiB)": 78.26,
      "step": 816,
      "token_acc": 0.8757039216681041,
      "train_speed(iter/s)": 0.033197
    },
    {
      "epoch": 0.15831032311195078,
      "grad_norm": 0.10780946165323257,
      "learning_rate": 0.00029051036480606053,
      "loss": 0.41828641295433044,
      "memory(GiB)": 78.26,
      "step": 817,
      "token_acc": 0.8813302092763193,
      "train_speed(iter/s)": 0.0332
    },
    {
      "epoch": 0.15850409339727753,
      "grad_norm": 0.11241378635168076,
      "learning_rate": 0.0002904766861795582,
      "loss": 0.4504837095737457,
      "memory(GiB)": 78.26,
      "step": 818,
      "token_acc": 0.8700797107404059,
      "train_speed(iter/s)": 0.033202
    },
    {
      "epoch": 0.15869786368260427,
      "grad_norm": 0.10670538991689682,
      "learning_rate": 0.00029044294985550607,
      "loss": 0.4240032434463501,
      "memory(GiB)": 78.26,
      "step": 819,
      "token_acc": 0.8797002130335464,
      "train_speed(iter/s)": 0.033205
    },
    {
      "epoch": 0.15889163396793102,
      "grad_norm": 0.09907601028680801,
      "learning_rate": 0.00029040915584776063,
      "loss": 0.4048020541667938,
      "memory(GiB)": 78.26,
      "step": 820,
      "token_acc": 0.8819687307138449,
      "train_speed(iter/s)": 0.033207
    },
    {
      "epoch": 0.15908540425325776,
      "grad_norm": 0.10909497737884521,
      "learning_rate": 0.00029037530417020194,
      "loss": 0.4121737480163574,
      "memory(GiB)": 78.26,
      "step": 821,
      "token_acc": 0.882557633753806,
      "train_speed(iter/s)": 0.03321
    },
    {
      "epoch": 0.1592791745385845,
      "grad_norm": 0.12102677673101425,
      "learning_rate": 0.00029034139483673373,
      "loss": 0.4818951487541199,
      "memory(GiB)": 78.26,
      "step": 822,
      "token_acc": 0.8623975497008026,
      "train_speed(iter/s)": 0.033212
    },
    {
      "epoch": 0.15947294482391125,
      "grad_norm": 0.11021539568901062,
      "learning_rate": 0.00029030742786128363,
      "loss": 0.42791998386383057,
      "memory(GiB)": 78.26,
      "step": 823,
      "token_acc": 0.8786863654583977,
      "train_speed(iter/s)": 0.033214
    },
    {
      "epoch": 0.159666715109238,
      "grad_norm": 0.12519389390945435,
      "learning_rate": 0.0002902734032578027,
      "loss": 0.4770694077014923,
      "memory(GiB)": 78.26,
      "step": 824,
      "token_acc": 0.8628657500881326,
      "train_speed(iter/s)": 0.033217
    },
    {
      "epoch": 0.15986048539456474,
      "grad_norm": 0.10490487515926361,
      "learning_rate": 0.0002902393210402657,
      "loss": 0.3933722674846649,
      "memory(GiB)": 78.26,
      "step": 825,
      "token_acc": 0.8858756744604317,
      "train_speed(iter/s)": 0.033219
    },
    {
      "epoch": 0.16005425567989148,
      "grad_norm": 0.10252831131219864,
      "learning_rate": 0.0002902051812226712,
      "loss": 0.3960869014263153,
      "memory(GiB)": 78.26,
      "step": 826,
      "token_acc": 0.8850197639086826,
      "train_speed(iter/s)": 0.033221
    },
    {
      "epoch": 0.16024802596521823,
      "grad_norm": 0.11041463166475296,
      "learning_rate": 0.0002901709838190413,
      "loss": 0.4398559033870697,
      "memory(GiB)": 78.26,
      "step": 827,
      "token_acc": 0.8748830196526983,
      "train_speed(iter/s)": 0.033224
    },
    {
      "epoch": 0.16044179625054497,
      "grad_norm": 0.11490801721811295,
      "learning_rate": 0.00029013672884342184,
      "loss": 0.47440311312675476,
      "memory(GiB)": 78.26,
      "step": 828,
      "token_acc": 0.865016464809957,
      "train_speed(iter/s)": 0.033226
    },
    {
      "epoch": 0.16063556653587172,
      "grad_norm": 0.11218751221895218,
      "learning_rate": 0.00029010241630988217,
      "loss": 0.43563181161880493,
      "memory(GiB)": 78.26,
      "step": 829,
      "token_acc": 0.8744653041825095,
      "train_speed(iter/s)": 0.033229
    },
    {
      "epoch": 0.16082933682119846,
      "grad_norm": 0.10691240429878235,
      "learning_rate": 0.00029006804623251547,
      "loss": 0.4449552297592163,
      "memory(GiB)": 78.26,
      "step": 830,
      "token_acc": 0.869553477182905,
      "train_speed(iter/s)": 0.033231
    },
    {
      "epoch": 0.1610231071065252,
      "grad_norm": 0.09648893773555756,
      "learning_rate": 0.00029003361862543834,
      "loss": 0.37794414162635803,
      "memory(GiB)": 78.26,
      "step": 831,
      "token_acc": 0.8909843389926163,
      "train_speed(iter/s)": 0.033233
    },
    {
      "epoch": 0.16121687739185195,
      "grad_norm": 0.10781273245811462,
      "learning_rate": 0.0002899991335027913,
      "loss": 0.4308411180973053,
      "memory(GiB)": 78.26,
      "step": 832,
      "token_acc": 0.877130367419212,
      "train_speed(iter/s)": 0.033235
    },
    {
      "epoch": 0.1614106476771787,
      "grad_norm": 0.10362358391284943,
      "learning_rate": 0.0002899645908787381,
      "loss": 0.40463730692863464,
      "memory(GiB)": 78.26,
      "step": 833,
      "token_acc": 0.8802580076079166,
      "train_speed(iter/s)": 0.033237
    },
    {
      "epoch": 0.16160441796250544,
      "grad_norm": 0.10180053114891052,
      "learning_rate": 0.0002899299907674665,
      "loss": 0.4222317337989807,
      "memory(GiB)": 78.26,
      "step": 834,
      "token_acc": 0.8798938964980744,
      "train_speed(iter/s)": 0.03324
    },
    {
      "epoch": 0.1617981882478322,
      "grad_norm": 0.1112997904419899,
      "learning_rate": 0.0002898953331831876,
      "loss": 0.45217838883399963,
      "memory(GiB)": 78.26,
      "step": 835,
      "token_acc": 0.8695664756612838,
      "train_speed(iter/s)": 0.033243
    },
    {
      "epoch": 0.16199195853315895,
      "grad_norm": 0.10161816328763962,
      "learning_rate": 0.0002898606181401362,
      "loss": 0.42238953709602356,
      "memory(GiB)": 78.26,
      "step": 836,
      "token_acc": 0.8775941837409121,
      "train_speed(iter/s)": 0.033245
    },
    {
      "epoch": 0.1621857288184857,
      "grad_norm": 0.11536618322134018,
      "learning_rate": 0.0002898258456525708,
      "loss": 0.4566255509853363,
      "memory(GiB)": 78.26,
      "step": 837,
      "token_acc": 0.8699845787186317,
      "train_speed(iter/s)": 0.033247
    },
    {
      "epoch": 0.16237949910381244,
      "grad_norm": 0.10816803574562073,
      "learning_rate": 0.0002897910157347733,
      "loss": 0.4683791697025299,
      "memory(GiB)": 78.26,
      "step": 838,
      "token_acc": 0.8670610211706102,
      "train_speed(iter/s)": 0.033249
    },
    {
      "epoch": 0.16257326938913919,
      "grad_norm": 0.11843861639499664,
      "learning_rate": 0.00028975612840104935,
      "loss": 0.45893388986587524,
      "memory(GiB)": 78.26,
      "step": 839,
      "token_acc": 0.8687269042794136,
      "train_speed(iter/s)": 0.033252
    },
    {
      "epoch": 0.16276703967446593,
      "grad_norm": 0.10663704574108124,
      "learning_rate": 0.000289721183665728,
      "loss": 0.4239065945148468,
      "memory(GiB)": 78.26,
      "step": 840,
      "token_acc": 0.8774621714892488,
      "train_speed(iter/s)": 0.033254
    },
    {
      "epoch": 0.16296080995979267,
      "grad_norm": 0.11534672975540161,
      "learning_rate": 0.00028968618154316206,
      "loss": 0.4733141362667084,
      "memory(GiB)": 78.26,
      "step": 841,
      "token_acc": 0.8653584989329564,
      "train_speed(iter/s)": 0.033256
    },
    {
      "epoch": 0.16315458024511942,
      "grad_norm": 0.10526616871356964,
      "learning_rate": 0.00028965112204772786,
      "loss": 0.4100850820541382,
      "memory(GiB)": 78.26,
      "step": 842,
      "token_acc": 0.8815631620593716,
      "train_speed(iter/s)": 0.033258
    },
    {
      "epoch": 0.16334835053044616,
      "grad_norm": 0.11487496644258499,
      "learning_rate": 0.00028961600519382527,
      "loss": 0.44554945826530457,
      "memory(GiB)": 78.26,
      "step": 843,
      "token_acc": 0.87236373596275,
      "train_speed(iter/s)": 0.033261
    },
    {
      "epoch": 0.1635421208157729,
      "grad_norm": 0.12236074358224869,
      "learning_rate": 0.00028958083099587774,
      "loss": 0.45087599754333496,
      "memory(GiB)": 78.26,
      "step": 844,
      "token_acc": 0.8709829274416269,
      "train_speed(iter/s)": 0.033263
    },
    {
      "epoch": 0.16373589110109965,
      "grad_norm": 0.11835192143917084,
      "learning_rate": 0.0002895455994683323,
      "loss": 0.4500479996204376,
      "memory(GiB)": 78.26,
      "step": 845,
      "token_acc": 0.8728866462050668,
      "train_speed(iter/s)": 0.033266
    },
    {
      "epoch": 0.1639296613864264,
      "grad_norm": 0.11468973010778427,
      "learning_rate": 0.0002895103106256593,
      "loss": 0.45714688301086426,
      "memory(GiB)": 78.26,
      "step": 846,
      "token_acc": 0.8693362313994315,
      "train_speed(iter/s)": 0.033268
    },
    {
      "epoch": 0.16412343167175314,
      "grad_norm": 0.11863450706005096,
      "learning_rate": 0.000289474964482353,
      "loss": 0.5020843148231506,
      "memory(GiB)": 78.26,
      "step": 847,
      "token_acc": 0.8581676315127108,
      "train_speed(iter/s)": 0.03327
    },
    {
      "epoch": 0.16431720195707988,
      "grad_norm": 0.09941709041595459,
      "learning_rate": 0.0002894395610529309,
      "loss": 0.3951787054538727,
      "memory(GiB)": 78.26,
      "step": 848,
      "token_acc": 0.8848054109751909,
      "train_speed(iter/s)": 0.033272
    },
    {
      "epoch": 0.16451097224240663,
      "grad_norm": 0.09850473701953888,
      "learning_rate": 0.0002894041003519343,
      "loss": 0.37680599093437195,
      "memory(GiB)": 78.26,
      "step": 849,
      "token_acc": 0.8903767059263413,
      "train_speed(iter/s)": 0.033274
    },
    {
      "epoch": 0.16470474252773337,
      "grad_norm": 0.10843576490879059,
      "learning_rate": 0.0002893685823939276,
      "loss": 0.4342295229434967,
      "memory(GiB)": 78.26,
      "step": 850,
      "token_acc": 0.8752759675488762,
      "train_speed(iter/s)": 0.033276
    },
    {
      "epoch": 0.16489851281306012,
      "grad_norm": 0.09915035963058472,
      "learning_rate": 0.00028933300719349923,
      "loss": 0.41352295875549316,
      "memory(GiB)": 78.26,
      "step": 851,
      "token_acc": 0.8786153965024339,
      "train_speed(iter/s)": 0.033279
    },
    {
      "epoch": 0.16509228309838686,
      "grad_norm": 0.11959869414567947,
      "learning_rate": 0.00028929737476526075,
      "loss": 0.506973147392273,
      "memory(GiB)": 78.26,
      "step": 852,
      "token_acc": 0.8568055984935781,
      "train_speed(iter/s)": 0.033282
    },
    {
      "epoch": 0.1652860533837136,
      "grad_norm": 0.10934106260538101,
      "learning_rate": 0.00028926168512384743,
      "loss": 0.43080487847328186,
      "memory(GiB)": 78.26,
      "step": 853,
      "token_acc": 0.875984682713348,
      "train_speed(iter/s)": 0.033284
    },
    {
      "epoch": 0.16547982366904035,
      "grad_norm": 0.10780926048755646,
      "learning_rate": 0.0002892259382839179,
      "loss": 0.4089958667755127,
      "memory(GiB)": 78.26,
      "step": 854,
      "token_acc": 0.8801118674476281,
      "train_speed(iter/s)": 0.033287
    },
    {
      "epoch": 0.1656735939543671,
      "grad_norm": 0.1189161166548729,
      "learning_rate": 0.0002891901342601543,
      "loss": 0.47321808338165283,
      "memory(GiB)": 78.26,
      "step": 855,
      "token_acc": 0.8673148412925364,
      "train_speed(iter/s)": 0.033289
    },
    {
      "epoch": 0.16586736423969384,
      "grad_norm": 0.11358017474412918,
      "learning_rate": 0.00028915427306726245,
      "loss": 0.4263777732849121,
      "memory(GiB)": 78.26,
      "step": 856,
      "token_acc": 0.8767594964603477,
      "train_speed(iter/s)": 0.033291
    },
    {
      "epoch": 0.16606113452502058,
      "grad_norm": 0.1009831428527832,
      "learning_rate": 0.00028911835471997143,
      "loss": 0.4132530987262726,
      "memory(GiB)": 78.26,
      "step": 857,
      "token_acc": 0.8810461753266181,
      "train_speed(iter/s)": 0.033293
    },
    {
      "epoch": 0.16625490481034733,
      "grad_norm": 0.10224108397960663,
      "learning_rate": 0.0002890823792330339,
      "loss": 0.4362914562225342,
      "memory(GiB)": 78.26,
      "step": 858,
      "token_acc": 0.8741152805026449,
      "train_speed(iter/s)": 0.033295
    },
    {
      "epoch": 0.16644867509567407,
      "grad_norm": 0.1064584031701088,
      "learning_rate": 0.00028904634662122586,
      "loss": 0.44220423698425293,
      "memory(GiB)": 78.26,
      "step": 859,
      "token_acc": 0.8728430508570941,
      "train_speed(iter/s)": 0.033297
    },
    {
      "epoch": 0.16664244538100081,
      "grad_norm": 0.1092744916677475,
      "learning_rate": 0.000289010256899347,
      "loss": 0.422356516122818,
      "memory(GiB)": 78.26,
      "step": 860,
      "token_acc": 0.8780035309255214,
      "train_speed(iter/s)": 0.033299
    },
    {
      "epoch": 0.16683621566632756,
      "grad_norm": 0.1130421906709671,
      "learning_rate": 0.0002889741100822202,
      "loss": 0.45889347791671753,
      "memory(GiB)": 78.26,
      "step": 861,
      "token_acc": 0.867090987216672,
      "train_speed(iter/s)": 0.033302
    },
    {
      "epoch": 0.1670299859516543,
      "grad_norm": 0.11241286247968674,
      "learning_rate": 0.00028893790618469213,
      "loss": 0.47700247168540955,
      "memory(GiB)": 78.26,
      "step": 862,
      "token_acc": 0.862,
      "train_speed(iter/s)": 0.033304
    },
    {
      "epoch": 0.16722375623698105,
      "grad_norm": 0.1090337261557579,
      "learning_rate": 0.0002889016452216325,
      "loss": 0.42925596237182617,
      "memory(GiB)": 78.26,
      "step": 863,
      "token_acc": 0.8770264194853622,
      "train_speed(iter/s)": 0.033306
    },
    {
      "epoch": 0.1674175265223078,
      "grad_norm": 0.12319561094045639,
      "learning_rate": 0.00028886532720793476,
      "loss": 0.4747333526611328,
      "memory(GiB)": 78.26,
      "step": 864,
      "token_acc": 0.8648867565967787,
      "train_speed(iter/s)": 0.033308
    },
    {
      "epoch": 0.16761129680763454,
      "grad_norm": 0.11913838982582092,
      "learning_rate": 0.0002888289521585157,
      "loss": 0.43152928352355957,
      "memory(GiB)": 78.26,
      "step": 865,
      "token_acc": 0.8771402469572392,
      "train_speed(iter/s)": 0.033311
    },
    {
      "epoch": 0.1678050670929613,
      "grad_norm": 0.12217225134372711,
      "learning_rate": 0.0002887925200883155,
      "loss": 0.468718022108078,
      "memory(GiB)": 78.26,
      "step": 866,
      "token_acc": 0.8652350105007839,
      "train_speed(iter/s)": 0.033313
    },
    {
      "epoch": 0.16799883737828805,
      "grad_norm": 0.10935720801353455,
      "learning_rate": 0.0002887560310122978,
      "loss": 0.427009254693985,
      "memory(GiB)": 78.26,
      "step": 867,
      "token_acc": 0.8757216164207825,
      "train_speed(iter/s)": 0.033315
    },
    {
      "epoch": 0.1681926076636148,
      "grad_norm": 0.11316697299480438,
      "learning_rate": 0.0002887194849454496,
      "loss": 0.4437088370323181,
      "memory(GiB)": 78.26,
      "step": 868,
      "token_acc": 0.873464339700426,
      "train_speed(iter/s)": 0.033317
    },
    {
      "epoch": 0.16838637794894154,
      "grad_norm": 0.11642614752054214,
      "learning_rate": 0.00028868288190278145,
      "loss": 0.4325766861438751,
      "memory(GiB)": 78.26,
      "step": 869,
      "token_acc": 0.8753681468056185,
      "train_speed(iter/s)": 0.033319
    },
    {
      "epoch": 0.16858014823426828,
      "grad_norm": 0.10680645704269409,
      "learning_rate": 0.00028864622189932713,
      "loss": 0.443330854177475,
      "memory(GiB)": 78.26,
      "step": 870,
      "token_acc": 0.8720809086981933,
      "train_speed(iter/s)": 0.033321
    },
    {
      "epoch": 0.16877391851959503,
      "grad_norm": 0.11195547878742218,
      "learning_rate": 0.00028860950495014393,
      "loss": 0.44975459575653076,
      "memory(GiB)": 78.26,
      "step": 871,
      "token_acc": 0.8717114605117083,
      "train_speed(iter/s)": 0.033323
    },
    {
      "epoch": 0.16896768880492177,
      "grad_norm": 0.10190173983573914,
      "learning_rate": 0.00028857273107031243,
      "loss": 0.4224821925163269,
      "memory(GiB)": 78.26,
      "step": 872,
      "token_acc": 0.8783074075926575,
      "train_speed(iter/s)": 0.033326
    },
    {
      "epoch": 0.16916145909024852,
      "grad_norm": 0.10820237547159195,
      "learning_rate": 0.0002885359002749367,
      "loss": 0.4363636076450348,
      "memory(GiB)": 78.26,
      "step": 873,
      "token_acc": 0.8747307109231629,
      "train_speed(iter/s)": 0.033328
    },
    {
      "epoch": 0.16935522937557526,
      "grad_norm": 0.10944371670484543,
      "learning_rate": 0.00028849901257914416,
      "loss": 0.4190293550491333,
      "memory(GiB)": 78.26,
      "step": 874,
      "token_acc": 0.87886427298192,
      "train_speed(iter/s)": 0.03333
    },
    {
      "epoch": 0.169548999660902,
      "grad_norm": 0.11039703339338303,
      "learning_rate": 0.0002884620679980855,
      "loss": 0.44972705841064453,
      "memory(GiB)": 78.26,
      "step": 875,
      "token_acc": 0.8701769399209686,
      "train_speed(iter/s)": 0.033332
    },
    {
      "epoch": 0.16974276994622875,
      "grad_norm": 0.10529112815856934,
      "learning_rate": 0.00028842506654693493,
      "loss": 0.45059582591056824,
      "memory(GiB)": 78.26,
      "step": 876,
      "token_acc": 0.8691080109913337,
      "train_speed(iter/s)": 0.033335
    },
    {
      "epoch": 0.1699365402315555,
      "grad_norm": 0.10080941766500473,
      "learning_rate": 0.00028838800824088984,
      "loss": 0.4126817286014557,
      "memory(GiB)": 78.26,
      "step": 877,
      "token_acc": 0.880942601909847,
      "train_speed(iter/s)": 0.033336
    },
    {
      "epoch": 0.17013031051688224,
      "grad_norm": 0.0992860496044159,
      "learning_rate": 0.00028835089309517116,
      "loss": 0.39974620938301086,
      "memory(GiB)": 78.26,
      "step": 878,
      "token_acc": 0.8827877507919747,
      "train_speed(iter/s)": 0.033338
    },
    {
      "epoch": 0.17032408080220898,
      "grad_norm": 0.11538361012935638,
      "learning_rate": 0.0002883137211250231,
      "loss": 0.4553724229335785,
      "memory(GiB)": 78.26,
      "step": 879,
      "token_acc": 0.8721838688599108,
      "train_speed(iter/s)": 0.03334
    },
    {
      "epoch": 0.17051785108753573,
      "grad_norm": 0.10386830568313599,
      "learning_rate": 0.0002882764923457131,
      "loss": 0.40255507826805115,
      "memory(GiB)": 78.26,
      "step": 880,
      "token_acc": 0.8842005576892724,
      "train_speed(iter/s)": 0.033343
    },
    {
      "epoch": 0.17071162137286247,
      "grad_norm": 0.10418614745140076,
      "learning_rate": 0.000288239206772532,
      "loss": 0.41422078013420105,
      "memory(GiB)": 78.26,
      "step": 881,
      "token_acc": 0.8813669826953899,
      "train_speed(iter/s)": 0.033345
    },
    {
      "epoch": 0.17090539165818922,
      "grad_norm": 0.10407593846321106,
      "learning_rate": 0.00028820186442079414,
      "loss": 0.40433794260025024,
      "memory(GiB)": 78.26,
      "step": 882,
      "token_acc": 0.8834683588753167,
      "train_speed(iter/s)": 0.033347
    },
    {
      "epoch": 0.17109916194351596,
      "grad_norm": 0.11592178791761398,
      "learning_rate": 0.0002881644653058369,
      "loss": 0.4496549963951111,
      "memory(GiB)": 78.26,
      "step": 883,
      "token_acc": 0.8705704799275581,
      "train_speed(iter/s)": 0.033349
    },
    {
      "epoch": 0.1712929322288427,
      "grad_norm": 0.10687348991632462,
      "learning_rate": 0.0002881270094430212,
      "loss": 0.3880995512008667,
      "memory(GiB)": 78.26,
      "step": 884,
      "token_acc": 0.889063373346464,
      "train_speed(iter/s)": 0.033351
    },
    {
      "epoch": 0.17148670251416945,
      "grad_norm": 0.11258858442306519,
      "learning_rate": 0.000288089496847731,
      "loss": 0.43804460763931274,
      "memory(GiB)": 78.26,
      "step": 885,
      "token_acc": 0.8751539547598306,
      "train_speed(iter/s)": 0.033353
    },
    {
      "epoch": 0.1716804727994962,
      "grad_norm": 0.10251414030790329,
      "learning_rate": 0.00028805192753537386,
      "loss": 0.39091038703918457,
      "memory(GiB)": 78.26,
      "step": 886,
      "token_acc": 0.884448902027027,
      "train_speed(iter/s)": 0.033355
    },
    {
      "epoch": 0.17187424308482294,
      "grad_norm": 0.10573960095643997,
      "learning_rate": 0.0002880143015213805,
      "loss": 0.43722304701805115,
      "memory(GiB)": 78.26,
      "step": 887,
      "token_acc": 0.8758301120031717,
      "train_speed(iter/s)": 0.033358
    },
    {
      "epoch": 0.17206801337014968,
      "grad_norm": 0.10792998969554901,
      "learning_rate": 0.00028797661882120495,
      "loss": 0.40696701407432556,
      "memory(GiB)": 78.26,
      "step": 888,
      "token_acc": 0.8834754878388297,
      "train_speed(iter/s)": 0.03336
    },
    {
      "epoch": 0.17226178365547642,
      "grad_norm": 0.10105016082525253,
      "learning_rate": 0.0002879388794503245,
      "loss": 0.3817410469055176,
      "memory(GiB)": 78.26,
      "step": 889,
      "token_acc": 0.8874983670803397,
      "train_speed(iter/s)": 0.033362
    },
    {
      "epoch": 0.17245555394080317,
      "grad_norm": 0.1093856692314148,
      "learning_rate": 0.0002879010834242396,
      "loss": 0.4195503294467926,
      "memory(GiB)": 78.26,
      "step": 890,
      "token_acc": 0.8788049002300327,
      "train_speed(iter/s)": 0.033364
    },
    {
      "epoch": 0.1726493242261299,
      "grad_norm": 0.11232632398605347,
      "learning_rate": 0.00028786323075847425,
      "loss": 0.45594799518585205,
      "memory(GiB)": 78.26,
      "step": 891,
      "token_acc": 0.8703043956619689,
      "train_speed(iter/s)": 0.033366
    },
    {
      "epoch": 0.17284309451145666,
      "grad_norm": 0.10968980938196182,
      "learning_rate": 0.00028782532146857546,
      "loss": 0.4293142557144165,
      "memory(GiB)": 78.26,
      "step": 892,
      "token_acc": 0.8755912321954648,
      "train_speed(iter/s)": 0.033368
    },
    {
      "epoch": 0.1730368647967834,
      "grad_norm": 0.12125832587480545,
      "learning_rate": 0.0002877873555701137,
      "loss": 0.49456584453582764,
      "memory(GiB)": 78.26,
      "step": 893,
      "token_acc": 0.8624862149246416,
      "train_speed(iter/s)": 0.03337
    },
    {
      "epoch": 0.17323063508211015,
      "grad_norm": 0.1106322631239891,
      "learning_rate": 0.00028774933307868243,
      "loss": 0.4586373269557953,
      "memory(GiB)": 78.26,
      "step": 894,
      "token_acc": 0.869065453654019,
      "train_speed(iter/s)": 0.033373
    },
    {
      "epoch": 0.1734244053674369,
      "grad_norm": 0.09941691905260086,
      "learning_rate": 0.00028771125400989863,
      "loss": 0.413199782371521,
      "memory(GiB)": 78.26,
      "step": 895,
      "token_acc": 0.8804068187829097,
      "train_speed(iter/s)": 0.033375
    },
    {
      "epoch": 0.17361817565276366,
      "grad_norm": 0.11516234278678894,
      "learning_rate": 0.0002876731183794024,
      "loss": 0.43939366936683655,
      "memory(GiB)": 78.26,
      "step": 896,
      "token_acc": 0.8730681167716084,
      "train_speed(iter/s)": 0.033377
    },
    {
      "epoch": 0.1738119459380904,
      "grad_norm": 0.10177889466285706,
      "learning_rate": 0.000287634926202857,
      "loss": 0.37696516513824463,
      "memory(GiB)": 78.26,
      "step": 897,
      "token_acc": 0.8901657575597098,
      "train_speed(iter/s)": 0.033379
    },
    {
      "epoch": 0.17400571622341715,
      "grad_norm": 0.11825387924909592,
      "learning_rate": 0.00028759667749594903,
      "loss": 0.4378824830055237,
      "memory(GiB)": 78.26,
      "step": 898,
      "token_acc": 0.8729655768399828,
      "train_speed(iter/s)": 0.033381
    },
    {
      "epoch": 0.1741994865087439,
      "grad_norm": 0.1011887937784195,
      "learning_rate": 0.0002875583722743882,
      "loss": 0.4053630232810974,
      "memory(GiB)": 78.26,
      "step": 899,
      "token_acc": 0.8832929164007658,
      "train_speed(iter/s)": 0.033383
    },
    {
      "epoch": 0.17439325679407064,
      "grad_norm": 0.11091629415750504,
      "learning_rate": 0.0002875200105539076,
      "loss": 0.43335989117622375,
      "memory(GiB)": 78.26,
      "step": 900,
      "token_acc": 0.8734870853616364,
      "train_speed(iter/s)": 0.033385
    },
    {
      "epoch": 0.17458702707939738,
      "grad_norm": 0.12012141942977905,
      "learning_rate": 0.00028748159235026337,
      "loss": 0.4563184380531311,
      "memory(GiB)": 78.26,
      "step": 901,
      "token_acc": 0.8695292291774444,
      "train_speed(iter/s)": 0.033387
    },
    {
      "epoch": 0.17478079736472413,
      "grad_norm": 0.1084456518292427,
      "learning_rate": 0.00028744311767923487,
      "loss": 0.42208635807037354,
      "memory(GiB)": 78.26,
      "step": 902,
      "token_acc": 0.8803865467709915,
      "train_speed(iter/s)": 0.033389
    },
    {
      "epoch": 0.17497456765005087,
      "grad_norm": 0.11079172044992447,
      "learning_rate": 0.00028740458655662467,
      "loss": 0.4311895966529846,
      "memory(GiB)": 78.26,
      "step": 903,
      "token_acc": 0.877756061809076,
      "train_speed(iter/s)": 0.033391
    },
    {
      "epoch": 0.17516833793537762,
      "grad_norm": 0.10750513523817062,
      "learning_rate": 0.00028736599899825856,
      "loss": 0.43391120433807373,
      "memory(GiB)": 78.26,
      "step": 904,
      "token_acc": 0.8769369834710744,
      "train_speed(iter/s)": 0.033393
    },
    {
      "epoch": 0.17536210822070436,
      "grad_norm": 0.11065103858709335,
      "learning_rate": 0.00028732735501998556,
      "loss": 0.4614184498786926,
      "memory(GiB)": 78.26,
      "step": 905,
      "token_acc": 0.8664335104252877,
      "train_speed(iter/s)": 0.033395
    },
    {
      "epoch": 0.1755558785060311,
      "grad_norm": 0.11256072670221329,
      "learning_rate": 0.0002872886546376777,
      "loss": 0.4341152012348175,
      "memory(GiB)": 78.26,
      "step": 906,
      "token_acc": 0.8766848053578903,
      "train_speed(iter/s)": 0.033397
    },
    {
      "epoch": 0.17574964879135785,
      "grad_norm": 0.10496652871370316,
      "learning_rate": 0.00028724989786723027,
      "loss": 0.4188949167728424,
      "memory(GiB)": 78.26,
      "step": 907,
      "token_acc": 0.8798962523819606,
      "train_speed(iter/s)": 0.033399
    },
    {
      "epoch": 0.1759434190766846,
      "grad_norm": 0.10520727187395096,
      "learning_rate": 0.00028721108472456173,
      "loss": 0.40807557106018066,
      "memory(GiB)": 78.26,
      "step": 908,
      "token_acc": 0.8816372301654051,
      "train_speed(iter/s)": 0.033401
    },
    {
      "epoch": 0.17613718936201134,
      "grad_norm": 0.10006900131702423,
      "learning_rate": 0.0002871722152256137,
      "loss": 0.41225340962409973,
      "memory(GiB)": 78.26,
      "step": 909,
      "token_acc": 0.8827568245850794,
      "train_speed(iter/s)": 0.033403
    },
    {
      "epoch": 0.17633095964733808,
      "grad_norm": 0.10612796992063522,
      "learning_rate": 0.0002871332893863509,
      "loss": 0.40802818536758423,
      "memory(GiB)": 78.26,
      "step": 910,
      "token_acc": 0.8830624153381265,
      "train_speed(iter/s)": 0.033405
    },
    {
      "epoch": 0.17652472993266483,
      "grad_norm": 0.11015652120113373,
      "learning_rate": 0.0002870943072227613,
      "loss": 0.4355056881904602,
      "memory(GiB)": 78.26,
      "step": 911,
      "token_acc": 0.8744773544633198,
      "train_speed(iter/s)": 0.033407
    },
    {
      "epoch": 0.17671850021799157,
      "grad_norm": 0.11280830204486847,
      "learning_rate": 0.00028705526875085575,
      "loss": 0.39818063378334045,
      "memory(GiB)": 78.26,
      "step": 912,
      "token_acc": 0.8824910029729307,
      "train_speed(iter/s)": 0.033409
    },
    {
      "epoch": 0.17691227050331831,
      "grad_norm": 0.10122046619653702,
      "learning_rate": 0.00028701617398666857,
      "loss": 0.412325382232666,
      "memory(GiB)": 78.26,
      "step": 913,
      "token_acc": 0.8825016812373907,
      "train_speed(iter/s)": 0.033411
    },
    {
      "epoch": 0.17710604078864506,
      "grad_norm": 0.11189986765384674,
      "learning_rate": 0.00028697702294625693,
      "loss": 0.4302537441253662,
      "memory(GiB)": 78.26,
      "step": 914,
      "token_acc": 0.8747316444826105,
      "train_speed(iter/s)": 0.033413
    },
    {
      "epoch": 0.1772998110739718,
      "grad_norm": 0.1030743420124054,
      "learning_rate": 0.0002869378156457013,
      "loss": 0.4212334454059601,
      "memory(GiB)": 78.26,
      "step": 915,
      "token_acc": 0.8769243019428814,
      "train_speed(iter/s)": 0.033415
    },
    {
      "epoch": 0.17749358135929855,
      "grad_norm": 0.09644519537687302,
      "learning_rate": 0.0002868985521011051,
      "loss": 0.3844713568687439,
      "memory(GiB)": 78.26,
      "step": 916,
      "token_acc": 0.8888474264248576,
      "train_speed(iter/s)": 0.033416
    },
    {
      "epoch": 0.1776873516446253,
      "grad_norm": 0.10586938261985779,
      "learning_rate": 0.000286859232328595,
      "loss": 0.4248504638671875,
      "memory(GiB)": 78.26,
      "step": 917,
      "token_acc": 0.8777870642473915,
      "train_speed(iter/s)": 0.033418
    },
    {
      "epoch": 0.17788112192995204,
      "grad_norm": 0.11170489341020584,
      "learning_rate": 0.00028681985634432055,
      "loss": 0.41771912574768066,
      "memory(GiB)": 78.26,
      "step": 918,
      "token_acc": 0.8776203050061971,
      "train_speed(iter/s)": 0.03342
    },
    {
      "epoch": 0.17807489221527878,
      "grad_norm": 0.11335260421037674,
      "learning_rate": 0.00028678042416445463,
      "loss": 0.4187549352645874,
      "memory(GiB)": 78.26,
      "step": 919,
      "token_acc": 0.8776417309627642,
      "train_speed(iter/s)": 0.033422
    },
    {
      "epoch": 0.17826866250060552,
      "grad_norm": 0.10571952164173126,
      "learning_rate": 0.0002867409358051931,
      "loss": 0.4060910642147064,
      "memory(GiB)": 78.26,
      "step": 920,
      "token_acc": 0.8823606660713348,
      "train_speed(iter/s)": 0.033423
    },
    {
      "epoch": 0.17846243278593227,
      "grad_norm": 0.12325315922498703,
      "learning_rate": 0.00028670139128275483,
      "loss": 0.44652315974235535,
      "memory(GiB)": 78.26,
      "step": 921,
      "token_acc": 0.8700105204529983,
      "train_speed(iter/s)": 0.033425
    },
    {
      "epoch": 0.178656203071259,
      "grad_norm": 0.1102391704916954,
      "learning_rate": 0.0002866617906133819,
      "loss": 0.4251522123813629,
      "memory(GiB)": 78.26,
      "step": 922,
      "token_acc": 0.876720608841388,
      "train_speed(iter/s)": 0.033427
    },
    {
      "epoch": 0.17884997335658576,
      "grad_norm": 0.1056000366806984,
      "learning_rate": 0.00028662213381333926,
      "loss": 0.4017042815685272,
      "memory(GiB)": 78.26,
      "step": 923,
      "token_acc": 0.8835219794682811,
      "train_speed(iter/s)": 0.033429
    },
    {
      "epoch": 0.1790437436419125,
      "grad_norm": 0.11068852990865707,
      "learning_rate": 0.00028658242089891513,
      "loss": 0.40017062425613403,
      "memory(GiB)": 78.26,
      "step": 924,
      "token_acc": 0.8838785514205683,
      "train_speed(iter/s)": 0.033431
    },
    {
      "epoch": 0.17923751392723924,
      "grad_norm": 0.10683512687683105,
      "learning_rate": 0.0002865426518864206,
      "loss": 0.4157922565937042,
      "memory(GiB)": 78.26,
      "step": 925,
      "token_acc": 0.8800466330435665,
      "train_speed(iter/s)": 0.033433
    },
    {
      "epoch": 0.17943128421256602,
      "grad_norm": 0.10481857508420944,
      "learning_rate": 0.00028650282679218994,
      "loss": 0.4214838743209839,
      "memory(GiB)": 78.26,
      "step": 926,
      "token_acc": 0.8785436944843359,
      "train_speed(iter/s)": 0.033435
    },
    {
      "epoch": 0.17962505449789276,
      "grad_norm": 0.10210458934307098,
      "learning_rate": 0.0002864629456325803,
      "loss": 0.39417165517807007,
      "memory(GiB)": 78.26,
      "step": 927,
      "token_acc": 0.8880719554934686,
      "train_speed(iter/s)": 0.033437
    },
    {
      "epoch": 0.1798188247832195,
      "grad_norm": 0.10829130560159683,
      "learning_rate": 0.000286423008423972,
      "loss": 0.4183374047279358,
      "memory(GiB)": 78.26,
      "step": 928,
      "token_acc": 0.8781808337845154,
      "train_speed(iter/s)": 0.033438
    },
    {
      "epoch": 0.18001259506854625,
      "grad_norm": 0.10995203256607056,
      "learning_rate": 0.00028638301518276826,
      "loss": 0.447214275598526,
      "memory(GiB)": 78.26,
      "step": 929,
      "token_acc": 0.8710934284538345,
      "train_speed(iter/s)": 0.03344
    },
    {
      "epoch": 0.180206365353873,
      "grad_norm": 0.10909594595432281,
      "learning_rate": 0.00028634296592539547,
      "loss": 0.3997655510902405,
      "memory(GiB)": 78.26,
      "step": 930,
      "token_acc": 0.8848353156450137,
      "train_speed(iter/s)": 0.033442
    },
    {
      "epoch": 0.18040013563919974,
      "grad_norm": 0.1183900386095047,
      "learning_rate": 0.0002863028606683029,
      "loss": 0.4467264413833618,
      "memory(GiB)": 78.26,
      "step": 931,
      "token_acc": 0.8734884573103701,
      "train_speed(iter/s)": 0.033444
    },
    {
      "epoch": 0.18059390592452648,
      "grad_norm": 0.10841910541057587,
      "learning_rate": 0.00028626269942796294,
      "loss": 0.4616568088531494,
      "memory(GiB)": 78.26,
      "step": 932,
      "token_acc": 0.8696779070709253,
      "train_speed(iter/s)": 0.033446
    },
    {
      "epoch": 0.18078767620985323,
      "grad_norm": 0.11348854750394821,
      "learning_rate": 0.0002862224822208707,
      "loss": 0.45820578932762146,
      "memory(GiB)": 78.26,
      "step": 933,
      "token_acc": 0.8705489207844856,
      "train_speed(iter/s)": 0.033448
    },
    {
      "epoch": 0.18098144649517997,
      "grad_norm": 0.10200490057468414,
      "learning_rate": 0.0002861822090635446,
      "loss": 0.40025630593299866,
      "memory(GiB)": 78.26,
      "step": 934,
      "token_acc": 0.8839878318584071,
      "train_speed(iter/s)": 0.03345
    },
    {
      "epoch": 0.18117521678050671,
      "grad_norm": 0.10408560186624527,
      "learning_rate": 0.00028614187997252585,
      "loss": 0.44248849153518677,
      "memory(GiB)": 78.26,
      "step": 935,
      "token_acc": 0.8738729347996039,
      "train_speed(iter/s)": 0.033451
    },
    {
      "epoch": 0.18136898706583346,
      "grad_norm": 0.1146451011300087,
      "learning_rate": 0.0002861014949643787,
      "loss": 0.4483628273010254,
      "memory(GiB)": 78.26,
      "step": 936,
      "token_acc": 0.8693443867125927,
      "train_speed(iter/s)": 0.033453
    },
    {
      "epoch": 0.1815627573511602,
      "grad_norm": 0.11855790764093399,
      "learning_rate": 0.0002860610540556905,
      "loss": 0.41510143876075745,
      "memory(GiB)": 78.26,
      "step": 937,
      "token_acc": 0.8812605670888882,
      "train_speed(iter/s)": 0.033455
    },
    {
      "epoch": 0.18175652763648695,
      "grad_norm": 0.12932439148426056,
      "learning_rate": 0.0002860205572630712,
      "loss": 0.41384294629096985,
      "memory(GiB)": 78.26,
      "step": 938,
      "token_acc": 0.8795511921458625,
      "train_speed(iter/s)": 0.033457
    },
    {
      "epoch": 0.1819502979218137,
      "grad_norm": 0.1010417714715004,
      "learning_rate": 0.00028598000460315404,
      "loss": 0.3942815363407135,
      "memory(GiB)": 78.26,
      "step": 939,
      "token_acc": 0.88588022518277,
      "train_speed(iter/s)": 0.033459
    },
    {
      "epoch": 0.18214406820714044,
      "grad_norm": 0.11028870195150375,
      "learning_rate": 0.00028593939609259506,
      "loss": 0.46208620071411133,
      "memory(GiB)": 78.26,
      "step": 940,
      "token_acc": 0.8662080777943533,
      "train_speed(iter/s)": 0.033461
    },
    {
      "epoch": 0.18233783849246718,
      "grad_norm": 0.10703767091035843,
      "learning_rate": 0.0002858987317480733,
      "loss": 0.4361118674278259,
      "memory(GiB)": 78.26,
      "step": 941,
      "token_acc": 0.8754664583875726,
      "train_speed(iter/s)": 0.033463
    },
    {
      "epoch": 0.18253160877779392,
      "grad_norm": 0.09738662093877792,
      "learning_rate": 0.00028585801158629063,
      "loss": 0.4183694124221802,
      "memory(GiB)": 78.26,
      "step": 942,
      "token_acc": 0.8812399810345217,
      "train_speed(iter/s)": 0.033465
    },
    {
      "epoch": 0.18272537906312067,
      "grad_norm": 0.10861563682556152,
      "learning_rate": 0.000285817235623972,
      "loss": 0.4377916157245636,
      "memory(GiB)": 78.26,
      "step": 943,
      "token_acc": 0.8772842397668782,
      "train_speed(iter/s)": 0.033466
    },
    {
      "epoch": 0.1829191493484474,
      "grad_norm": 0.11255238950252533,
      "learning_rate": 0.0002857764038778651,
      "loss": 0.4062233865261078,
      "memory(GiB)": 78.26,
      "step": 944,
      "token_acc": 0.8827611918896441,
      "train_speed(iter/s)": 0.033468
    },
    {
      "epoch": 0.18311291963377416,
      "grad_norm": 0.1005673035979271,
      "learning_rate": 0.0002857355163647407,
      "loss": 0.3768383264541626,
      "memory(GiB)": 78.26,
      "step": 945,
      "token_acc": 0.8898000897352809,
      "train_speed(iter/s)": 0.03347
    },
    {
      "epoch": 0.1833066899191009,
      "grad_norm": 0.1112765297293663,
      "learning_rate": 0.00028569457310139237,
      "loss": 0.4166128635406494,
      "memory(GiB)": 78.26,
      "step": 946,
      "token_acc": 0.87920341245126,
      "train_speed(iter/s)": 0.033472
    },
    {
      "epoch": 0.18350046020442765,
      "grad_norm": 0.12681414186954498,
      "learning_rate": 0.00028565357410463663,
      "loss": 0.4559955894947052,
      "memory(GiB)": 78.26,
      "step": 947,
      "token_acc": 0.8697389451251998,
      "train_speed(iter/s)": 0.033474
    },
    {
      "epoch": 0.1836942304897544,
      "grad_norm": 0.1174798235297203,
      "learning_rate": 0.0002856125193913128,
      "loss": 0.48667585849761963,
      "memory(GiB)": 78.26,
      "step": 948,
      "token_acc": 0.8624453490125132,
      "train_speed(iter/s)": 0.033476
    },
    {
      "epoch": 0.18388800077508113,
      "grad_norm": 0.11614065617322922,
      "learning_rate": 0.00028557140897828324,
      "loss": 0.43031829595565796,
      "memory(GiB)": 78.26,
      "step": 949,
      "token_acc": 0.8768826126436422,
      "train_speed(iter/s)": 0.033478
    },
    {
      "epoch": 0.18408177106040788,
      "grad_norm": 0.11647061258554459,
      "learning_rate": 0.00028553024288243306,
      "loss": 0.43689054250717163,
      "memory(GiB)": 78.26,
      "step": 950,
      "token_acc": 0.8749047820567076,
      "train_speed(iter/s)": 0.03348
    },
    {
      "epoch": 0.18427554134573462,
      "grad_norm": 0.12167865037918091,
      "learning_rate": 0.0002854890211206703,
      "loss": 0.4729604125022888,
      "memory(GiB)": 78.26,
      "step": 951,
      "token_acc": 0.8637253103406103,
      "train_speed(iter/s)": 0.033483
    },
    {
      "epoch": 0.18446931163106137,
      "grad_norm": 0.11439337581396103,
      "learning_rate": 0.00028544774370992587,
      "loss": 0.4109064042568207,
      "memory(GiB)": 78.26,
      "step": 952,
      "token_acc": 0.8795947901591896,
      "train_speed(iter/s)": 0.033484
    },
    {
      "epoch": 0.1846630819163881,
      "grad_norm": 0.11438382416963577,
      "learning_rate": 0.0002854064106671534,
      "loss": 0.43977493047714233,
      "memory(GiB)": 78.26,
      "step": 953,
      "token_acc": 0.8736031785448225,
      "train_speed(iter/s)": 0.033486
    },
    {
      "epoch": 0.18485685220171486,
      "grad_norm": 0.10898079723119736,
      "learning_rate": 0.0002853650220093296,
      "loss": 0.45643150806427,
      "memory(GiB)": 78.26,
      "step": 954,
      "token_acc": 0.8683694024069004,
      "train_speed(iter/s)": 0.033488
    },
    {
      "epoch": 0.1850506224870416,
      "grad_norm": 0.1164807602763176,
      "learning_rate": 0.0002853235777534539,
      "loss": 0.41862934827804565,
      "memory(GiB)": 78.26,
      "step": 955,
      "token_acc": 0.8782421565296183,
      "train_speed(iter/s)": 0.03349
    },
    {
      "epoch": 0.18524439277236834,
      "grad_norm": 0.11116447299718857,
      "learning_rate": 0.00028528207791654847,
      "loss": 0.4242367148399353,
      "memory(GiB)": 78.26,
      "step": 956,
      "token_acc": 0.8753245376993589,
      "train_speed(iter/s)": 0.033492
    },
    {
      "epoch": 0.18543816305769512,
      "grad_norm": 0.11318708211183548,
      "learning_rate": 0.0002852405225156585,
      "loss": 0.43691354990005493,
      "memory(GiB)": 78.26,
      "step": 957,
      "token_acc": 0.8767160750423891,
      "train_speed(iter/s)": 0.033493
    },
    {
      "epoch": 0.18563193334302186,
      "grad_norm": 0.10934760421514511,
      "learning_rate": 0.00028519891156785187,
      "loss": 0.45028555393218994,
      "memory(GiB)": 78.26,
      "step": 958,
      "token_acc": 0.8715856095936043,
      "train_speed(iter/s)": 0.033495
    },
    {
      "epoch": 0.1858257036283486,
      "grad_norm": 0.11843264102935791,
      "learning_rate": 0.0002851572450902193,
      "loss": 0.45888209342956543,
      "memory(GiB)": 78.26,
      "step": 959,
      "token_acc": 0.8671850891810159,
      "train_speed(iter/s)": 0.033497
    },
    {
      "epoch": 0.18601947391367535,
      "grad_norm": 0.10771480202674866,
      "learning_rate": 0.0002851155230998744,
      "loss": 0.44502004981040955,
      "memory(GiB)": 78.26,
      "step": 960,
      "token_acc": 0.8715418593237414,
      "train_speed(iter/s)": 0.033499
    },
    {
      "epoch": 0.1862132441990021,
      "grad_norm": 0.11727321892976761,
      "learning_rate": 0.00028507374561395345,
      "loss": 0.4620700478553772,
      "memory(GiB)": 78.26,
      "step": 961,
      "token_acc": 0.8692043503148255,
      "train_speed(iter/s)": 0.033501
    },
    {
      "epoch": 0.18640701448432884,
      "grad_norm": 0.11580432206392288,
      "learning_rate": 0.0002850319126496156,
      "loss": 0.43666505813598633,
      "memory(GiB)": 78.26,
      "step": 962,
      "token_acc": 0.8729514358743046,
      "train_speed(iter/s)": 0.033503
    },
    {
      "epoch": 0.18660078476965558,
      "grad_norm": 0.11996878683567047,
      "learning_rate": 0.00028499002422404274,
      "loss": 0.49721530079841614,
      "memory(GiB)": 78.26,
      "step": 963,
      "token_acc": 0.8572844400396432,
      "train_speed(iter/s)": 0.033504
    },
    {
      "epoch": 0.18679455505498233,
      "grad_norm": 0.09926389157772064,
      "learning_rate": 0.00028494808035443966,
      "loss": 0.3834714889526367,
      "memory(GiB)": 78.26,
      "step": 964,
      "token_acc": 0.8917675254643499,
      "train_speed(iter/s)": 0.033506
    },
    {
      "epoch": 0.18698832534030907,
      "grad_norm": 0.11022564768791199,
      "learning_rate": 0.00028490608105803374,
      "loss": 0.43356478214263916,
      "memory(GiB)": 78.26,
      "step": 965,
      "token_acc": 0.8749029825923051,
      "train_speed(iter/s)": 0.033508
    },
    {
      "epoch": 0.1871820956256358,
      "grad_norm": 0.11110399663448334,
      "learning_rate": 0.0002848640263520753,
      "loss": 0.4324399530887604,
      "memory(GiB)": 78.26,
      "step": 966,
      "token_acc": 0.8767509301816591,
      "train_speed(iter/s)": 0.03351
    },
    {
      "epoch": 0.18737586591096256,
      "grad_norm": 0.11281298100948334,
      "learning_rate": 0.00028482191625383733,
      "loss": 0.4322070777416229,
      "memory(GiB)": 78.26,
      "step": 967,
      "token_acc": 0.8765359035058055,
      "train_speed(iter/s)": 0.033512
    },
    {
      "epoch": 0.1875696361962893,
      "grad_norm": 0.1022447943687439,
      "learning_rate": 0.0002847797507806155,
      "loss": 0.39318740367889404,
      "memory(GiB)": 78.26,
      "step": 968,
      "token_acc": 0.8852789546220203,
      "train_speed(iter/s)": 0.033513
    },
    {
      "epoch": 0.18776340648161605,
      "grad_norm": 0.12078309804201126,
      "learning_rate": 0.0002847375299497284,
      "loss": 0.4697941243648529,
      "memory(GiB)": 78.26,
      "step": 969,
      "token_acc": 0.8657820386654392,
      "train_speed(iter/s)": 0.033515
    },
    {
      "epoch": 0.1879571767669428,
      "grad_norm": 0.10926394909620285,
      "learning_rate": 0.00028469525377851715,
      "loss": 0.3991783857345581,
      "memory(GiB)": 78.26,
      "step": 970,
      "token_acc": 0.8852660300136426,
      "train_speed(iter/s)": 0.033517
    },
    {
      "epoch": 0.18815094705226953,
      "grad_norm": 0.11045163869857788,
      "learning_rate": 0.0002846529222843458,
      "loss": 0.4475860893726349,
      "memory(GiB)": 78.26,
      "step": 971,
      "token_acc": 0.8725651014968219,
      "train_speed(iter/s)": 0.033518
    },
    {
      "epoch": 0.18834471733759628,
      "grad_norm": 0.09906210750341415,
      "learning_rate": 0.000284610535484601,
      "loss": 0.34970560669898987,
      "memory(GiB)": 78.26,
      "step": 972,
      "token_acc": 0.8972711247963943,
      "train_speed(iter/s)": 0.03352
    },
    {
      "epoch": 0.18853848762292302,
      "grad_norm": 0.11121566593647003,
      "learning_rate": 0.00028456809339669214,
      "loss": 0.4188978672027588,
      "memory(GiB)": 78.26,
      "step": 973,
      "token_acc": 0.8785746369216949,
      "train_speed(iter/s)": 0.033522
    },
    {
      "epoch": 0.18873225790824977,
      "grad_norm": 0.11328519135713577,
      "learning_rate": 0.00028452559603805137,
      "loss": 0.44297975301742554,
      "memory(GiB)": 78.26,
      "step": 974,
      "token_acc": 0.8741913622081202,
      "train_speed(iter/s)": 0.033524
    },
    {
      "epoch": 0.1889260281935765,
      "grad_norm": 0.10972630977630615,
      "learning_rate": 0.00028448304342613344,
      "loss": 0.4079618453979492,
      "memory(GiB)": 78.26,
      "step": 975,
      "token_acc": 0.8813834977772959,
      "train_speed(iter/s)": 0.033526
    },
    {
      "epoch": 0.18911979847890326,
      "grad_norm": 0.0967109352350235,
      "learning_rate": 0.00028444043557841585,
      "loss": 0.38039693236351013,
      "memory(GiB)": 78.26,
      "step": 976,
      "token_acc": 0.8867412594640688,
      "train_speed(iter/s)": 0.033528
    },
    {
      "epoch": 0.18931356876423,
      "grad_norm": 0.10372382402420044,
      "learning_rate": 0.00028439777251239887,
      "loss": 0.42939719557762146,
      "memory(GiB)": 78.26,
      "step": 977,
      "token_acc": 0.8766995304010031,
      "train_speed(iter/s)": 0.033529
    },
    {
      "epoch": 0.18950733904955674,
      "grad_norm": 0.10110750049352646,
      "learning_rate": 0.00028435505424560527,
      "loss": 0.4236353039741516,
      "memory(GiB)": 78.26,
      "step": 978,
      "token_acc": 0.8794206803637589,
      "train_speed(iter/s)": 0.033531
    },
    {
      "epoch": 0.1897011093348835,
      "grad_norm": 0.11021259427070618,
      "learning_rate": 0.00028431228079558063,
      "loss": 0.4256800711154938,
      "memory(GiB)": 78.26,
      "step": 979,
      "token_acc": 0.8776469925658932,
      "train_speed(iter/s)": 0.033533
    },
    {
      "epoch": 0.18989487962021023,
      "grad_norm": 0.10382870584726334,
      "learning_rate": 0.00028426945217989316,
      "loss": 0.39924290776252747,
      "memory(GiB)": 78.26,
      "step": 980,
      "token_acc": 0.8837884005314389,
      "train_speed(iter/s)": 0.033535
    },
    {
      "epoch": 0.19008864990553698,
      "grad_norm": 0.11175846308469772,
      "learning_rate": 0.00028422656841613377,
      "loss": 0.42587774991989136,
      "memory(GiB)": 78.26,
      "step": 981,
      "token_acc": 0.8782795084309802,
      "train_speed(iter/s)": 0.033537
    },
    {
      "epoch": 0.19028242019086372,
      "grad_norm": 0.10329094529151917,
      "learning_rate": 0.00028418362952191585,
      "loss": 0.40456724166870117,
      "memory(GiB)": 78.26,
      "step": 982,
      "token_acc": 0.881374140565317,
      "train_speed(iter/s)": 0.033539
    },
    {
      "epoch": 0.19047619047619047,
      "grad_norm": 0.11278124898672104,
      "learning_rate": 0.0002841406355148757,
      "loss": 0.4433564245700836,
      "memory(GiB)": 78.26,
      "step": 983,
      "token_acc": 0.871099352158647,
      "train_speed(iter/s)": 0.033541
    },
    {
      "epoch": 0.1906699607615172,
      "grad_norm": 0.11207697540521622,
      "learning_rate": 0.000284097586412672,
      "loss": 0.40271443128585815,
      "memory(GiB)": 78.26,
      "step": 984,
      "token_acc": 0.8841104695500599,
      "train_speed(iter/s)": 0.033542
    },
    {
      "epoch": 0.19086373104684395,
      "grad_norm": 0.11175701022148132,
      "learning_rate": 0.00028405448223298624,
      "loss": 0.451436847448349,
      "memory(GiB)": 78.26,
      "step": 985,
      "token_acc": 0.8711304942705661,
      "train_speed(iter/s)": 0.033544
    },
    {
      "epoch": 0.1910575013321707,
      "grad_norm": 0.10010375082492828,
      "learning_rate": 0.0002840113229935224,
      "loss": 0.3909650444984436,
      "memory(GiB)": 78.26,
      "step": 986,
      "token_acc": 0.8874252112061217,
      "train_speed(iter/s)": 0.033545
    },
    {
      "epoch": 0.19125127161749747,
      "grad_norm": 0.09799065440893173,
      "learning_rate": 0.0002839681087120073,
      "loss": 0.38588500022888184,
      "memory(GiB)": 78.26,
      "step": 987,
      "token_acc": 0.8859629435674921,
      "train_speed(iter/s)": 0.033547
    },
    {
      "epoch": 0.19144504190282421,
      "grad_norm": 0.11740557104349136,
      "learning_rate": 0.0002839248394061899,
      "loss": 0.4740091562271118,
      "memory(GiB)": 78.26,
      "step": 988,
      "token_acc": 0.8610855829982769,
      "train_speed(iter/s)": 0.033549
    },
    {
      "epoch": 0.19163881218815096,
      "grad_norm": 0.11015292257070541,
      "learning_rate": 0.0002838815150938424,
      "loss": 0.41003599762916565,
      "memory(GiB)": 78.26,
      "step": 989,
      "token_acc": 0.882699868938401,
      "train_speed(iter/s)": 0.033551
    },
    {
      "epoch": 0.1918325824734777,
      "grad_norm": 0.11227616667747498,
      "learning_rate": 0.0002838381357927591,
      "loss": 0.41413140296936035,
      "memory(GiB)": 78.26,
      "step": 990,
      "token_acc": 0.880181635836862,
      "train_speed(iter/s)": 0.033552
    },
    {
      "epoch": 0.19202635275880445,
      "grad_norm": 0.11800327897071838,
      "learning_rate": 0.000283794701520757,
      "loss": 0.4505453109741211,
      "memory(GiB)": 78.26,
      "step": 991,
      "token_acc": 0.8734527175620483,
      "train_speed(iter/s)": 0.033554
    },
    {
      "epoch": 0.1922201230441312,
      "grad_norm": 0.11021004617214203,
      "learning_rate": 0.00028375121229567583,
      "loss": 0.42506149411201477,
      "memory(GiB)": 78.26,
      "step": 992,
      "token_acc": 0.8773052126543056,
      "train_speed(iter/s)": 0.033556
    },
    {
      "epoch": 0.19241389332945794,
      "grad_norm": 0.11306871473789215,
      "learning_rate": 0.0002837076681353777,
      "loss": 0.4095050096511841,
      "memory(GiB)": 78.26,
      "step": 993,
      "token_acc": 0.8811114926437571,
      "train_speed(iter/s)": 0.033557
    },
    {
      "epoch": 0.19260766361478468,
      "grad_norm": 0.09795724600553513,
      "learning_rate": 0.00028366406905774746,
      "loss": 0.376755952835083,
      "memory(GiB)": 78.26,
      "step": 994,
      "token_acc": 0.8902818813383374,
      "train_speed(iter/s)": 0.033559
    },
    {
      "epoch": 0.19280143390011142,
      "grad_norm": 0.10513719916343689,
      "learning_rate": 0.0002836204150806923,
      "loss": 0.39423179626464844,
      "memory(GiB)": 78.26,
      "step": 995,
      "token_acc": 0.8858132362606805,
      "train_speed(iter/s)": 0.03356
    },
    {
      "epoch": 0.19299520418543817,
      "grad_norm": 0.10870835930109024,
      "learning_rate": 0.0002835767062221422,
      "loss": 0.43109869956970215,
      "memory(GiB)": 78.26,
      "step": 996,
      "token_acc": 0.8794217244073977,
      "train_speed(iter/s)": 0.033562
    },
    {
      "epoch": 0.1931889744707649,
      "grad_norm": 0.11085118353366852,
      "learning_rate": 0.0002835329425000495,
      "loss": 0.41626110672950745,
      "memory(GiB)": 78.26,
      "step": 997,
      "token_acc": 0.8794915927446048,
      "train_speed(iter/s)": 0.033564
    },
    {
      "epoch": 0.19338274475609166,
      "grad_norm": 0.10917031019926071,
      "learning_rate": 0.00028348912393238914,
      "loss": 0.442068487405777,
      "memory(GiB)": 78.26,
      "step": 998,
      "token_acc": 0.8718288169021695,
      "train_speed(iter/s)": 0.033566
    },
    {
      "epoch": 0.1935765150414184,
      "grad_norm": 0.11420496553182602,
      "learning_rate": 0.00028344525053715857,
      "loss": 0.45237767696380615,
      "memory(GiB)": 78.26,
      "step": 999,
      "token_acc": 0.8713443106414352,
      "train_speed(iter/s)": 0.033568
    },
    {
      "epoch": 0.19377028532674515,
      "grad_norm": 0.10354109853506088,
      "learning_rate": 0.0002834013223323778,
      "loss": 0.40592384338378906,
      "memory(GiB)": 78.26,
      "step": 1000,
      "token_acc": 0.883349086326402,
      "train_speed(iter/s)": 0.033569
    },
    {
      "epoch": 0.19377028532674515,
      "eval_loss": 0.48618289828300476,
      "eval_runtime": 1345.1073,
      "eval_samples_per_second": 5.017,
      "eval_steps_per_second": 5.017,
      "eval_token_acc": 0.8788295367774018,
      "step": 1000
    },
    {
      "epoch": 0.1939640556120719,
      "grad_norm": 0.11008410900831223,
      "learning_rate": 0.00028335733933608937,
      "loss": 0.44370290637016296,
      "memory(GiB)": 78.26,
      "step": 1001,
      "token_acc": 0.8703201902339214,
      "train_speed(iter/s)": 0.032109
    },
    {
      "epoch": 0.19415782589739863,
      "grad_norm": 0.10204144567251205,
      "learning_rate": 0.00028331330156635814,
      "loss": 0.37651312351226807,
      "memory(GiB)": 78.26,
      "step": 1002,
      "token_acc": 0.8898332699723022,
      "train_speed(iter/s)": 0.032111
    },
    {
      "epoch": 0.19435159618272538,
      "grad_norm": 0.10822474956512451,
      "learning_rate": 0.0002832692090412717,
      "loss": 0.4068623185157776,
      "memory(GiB)": 78.26,
      "step": 1003,
      "token_acc": 0.8804925222697431,
      "train_speed(iter/s)": 0.032114
    },
    {
      "epoch": 0.19454536646805212,
      "grad_norm": 0.10033664852380753,
      "learning_rate": 0.0002832250617789401,
      "loss": 0.4149298667907715,
      "memory(GiB)": 78.26,
      "step": 1004,
      "token_acc": 0.8801352320458969,
      "train_speed(iter/s)": 0.032117
    },
    {
      "epoch": 0.19473913675337887,
      "grad_norm": 0.12285728007555008,
      "learning_rate": 0.00028318085979749563,
      "loss": 0.46761053800582886,
      "memory(GiB)": 78.26,
      "step": 1005,
      "token_acc": 0.8641041659816526,
      "train_speed(iter/s)": 0.03212
    },
    {
      "epoch": 0.1949329070387056,
      "grad_norm": 0.09842801839113235,
      "learning_rate": 0.0002831366031150934,
      "loss": 0.39952999353408813,
      "memory(GiB)": 78.26,
      "step": 1006,
      "token_acc": 0.8830670627587659,
      "train_speed(iter/s)": 0.032123
    },
    {
      "epoch": 0.19512667732403235,
      "grad_norm": 0.11463743448257446,
      "learning_rate": 0.0002830922917499108,
      "loss": 0.43408212065696716,
      "memory(GiB)": 78.26,
      "step": 1007,
      "token_acc": 0.8743004464566434,
      "train_speed(iter/s)": 0.032126
    },
    {
      "epoch": 0.1953204476093591,
      "grad_norm": 0.11018949747085571,
      "learning_rate": 0.00028304792572014754,
      "loss": 0.43823209404945374,
      "memory(GiB)": 78.26,
      "step": 1008,
      "token_acc": 0.8741348099243957,
      "train_speed(iter/s)": 0.032128
    },
    {
      "epoch": 0.19551421789468584,
      "grad_norm": 0.11423125863075256,
      "learning_rate": 0.00028300350504402606,
      "loss": 0.45217055082321167,
      "memory(GiB)": 78.26,
      "step": 1009,
      "token_acc": 0.8701353400348177,
      "train_speed(iter/s)": 0.032131
    },
    {
      "epoch": 0.1957079881800126,
      "grad_norm": 0.1115386039018631,
      "learning_rate": 0.0002829590297397912,
      "loss": 0.4279889762401581,
      "memory(GiB)": 78.26,
      "step": 1010,
      "token_acc": 0.8788333144732375,
      "train_speed(iter/s)": 0.032134
    },
    {
      "epoch": 0.19590175846533933,
      "grad_norm": 0.1104380413889885,
      "learning_rate": 0.00028291449982570995,
      "loss": 0.4566521942615509,
      "memory(GiB)": 78.26,
      "step": 1011,
      "token_acc": 0.8691424216142343,
      "train_speed(iter/s)": 0.032137
    },
    {
      "epoch": 0.19609552875066608,
      "grad_norm": 0.11296399682760239,
      "learning_rate": 0.00028286991532007217,
      "loss": 0.43526938557624817,
      "memory(GiB)": 78.26,
      "step": 1012,
      "token_acc": 0.8731874876819551,
      "train_speed(iter/s)": 0.03214
    },
    {
      "epoch": 0.19628929903599282,
      "grad_norm": 0.10835513472557068,
      "learning_rate": 0.0002828252762411898,
      "loss": 0.4160099923610687,
      "memory(GiB)": 78.26,
      "step": 1013,
      "token_acc": 0.880335822929728,
      "train_speed(iter/s)": 0.032143
    },
    {
      "epoch": 0.19648306932131956,
      "grad_norm": 0.09724919497966766,
      "learning_rate": 0.00028278058260739733,
      "loss": 0.3947051763534546,
      "memory(GiB)": 78.26,
      "step": 1014,
      "token_acc": 0.8845330218523385,
      "train_speed(iter/s)": 0.032145
    },
    {
      "epoch": 0.1966768396066463,
      "grad_norm": 0.11479044705629349,
      "learning_rate": 0.0002827358344370516,
      "loss": 0.4259167015552521,
      "memory(GiB)": 78.26,
      "step": 1015,
      "token_acc": 0.8770737733851042,
      "train_speed(iter/s)": 0.032148
    },
    {
      "epoch": 0.19687060989197305,
      "grad_norm": 0.11837134510278702,
      "learning_rate": 0.000282691031748532,
      "loss": 0.4474408030509949,
      "memory(GiB)": 78.26,
      "step": 1016,
      "token_acc": 0.8693976711066432,
      "train_speed(iter/s)": 0.032151
    },
    {
      "epoch": 0.19706438017729982,
      "grad_norm": 0.10522796213626862,
      "learning_rate": 0.00028264617456024,
      "loss": 0.41682255268096924,
      "memory(GiB)": 78.26,
      "step": 1017,
      "token_acc": 0.8796135557400558,
      "train_speed(iter/s)": 0.032154
    },
    {
      "epoch": 0.19725815046262657,
      "grad_norm": 0.10754600167274475,
      "learning_rate": 0.00028260126289059986,
      "loss": 0.4248649477958679,
      "memory(GiB)": 78.26,
      "step": 1018,
      "token_acc": 0.8764376037077302,
      "train_speed(iter/s)": 0.032156
    },
    {
      "epoch": 0.1974519207479533,
      "grad_norm": 0.10504762083292007,
      "learning_rate": 0.00028255629675805785,
      "loss": 0.4158293604850769,
      "memory(GiB)": 78.26,
      "step": 1019,
      "token_acc": 0.8806701766216252,
      "train_speed(iter/s)": 0.032159
    },
    {
      "epoch": 0.19764569103328006,
      "grad_norm": 0.11185994744300842,
      "learning_rate": 0.0002825112761810828,
      "loss": 0.43085163831710815,
      "memory(GiB)": 78.26,
      "step": 1020,
      "token_acc": 0.8744727730563424,
      "train_speed(iter/s)": 0.032162
    },
    {
      "epoch": 0.1978394613186068,
      "grad_norm": 0.09813162684440613,
      "learning_rate": 0.000282466201178166,
      "loss": 0.4018121063709259,
      "memory(GiB)": 78.26,
      "step": 1021,
      "token_acc": 0.8820514020879172,
      "train_speed(iter/s)": 0.032165
    },
    {
      "epoch": 0.19803323160393355,
      "grad_norm": 0.1113823875784874,
      "learning_rate": 0.0002824210717678209,
      "loss": 0.40826714038848877,
      "memory(GiB)": 78.26,
      "step": 1022,
      "token_acc": 0.882307549027434,
      "train_speed(iter/s)": 0.032167
    },
    {
      "epoch": 0.1982270018892603,
      "grad_norm": 0.1155644953250885,
      "learning_rate": 0.00028237588796858323,
      "loss": 0.4487013816833496,
      "memory(GiB)": 78.26,
      "step": 1023,
      "token_acc": 0.8709012113617377,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.19842077217458703,
      "grad_norm": 0.11795882880687714,
      "learning_rate": 0.0002823306497990113,
      "loss": 0.43464547395706177,
      "memory(GiB)": 78.26,
      "step": 1024,
      "token_acc": 0.8770835761743793,
      "train_speed(iter/s)": 0.032173
    },
    {
      "epoch": 0.19861454245991378,
      "grad_norm": 0.1030697301030159,
      "learning_rate": 0.00028228535727768575,
      "loss": 0.4153880476951599,
      "memory(GiB)": 78.26,
      "step": 1025,
      "token_acc": 0.8816005247622172,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.19880831274524052,
      "grad_norm": 0.10959405452013016,
      "learning_rate": 0.00028224001042320923,
      "loss": 0.4293935000896454,
      "memory(GiB)": 78.26,
      "step": 1026,
      "token_acc": 0.8751588677065281,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.19900208303056727,
      "grad_norm": 0.11512494832277298,
      "learning_rate": 0.00028219460925420697,
      "loss": 0.46233439445495605,
      "memory(GiB)": 78.26,
      "step": 1027,
      "token_acc": 0.8720349947631076,
      "train_speed(iter/s)": 0.032181
    },
    {
      "epoch": 0.199195853315894,
      "grad_norm": 0.11418092250823975,
      "learning_rate": 0.00028214915378932653,
      "loss": 0.45269933342933655,
      "memory(GiB)": 78.26,
      "step": 1028,
      "token_acc": 0.8701842783140374,
      "train_speed(iter/s)": 0.032184
    },
    {
      "epoch": 0.19938962360122076,
      "grad_norm": 0.10788623243570328,
      "learning_rate": 0.00028210364404723765,
      "loss": 0.4401698708534241,
      "memory(GiB)": 78.26,
      "step": 1029,
      "token_acc": 0.8730818757377401,
      "train_speed(iter/s)": 0.032187
    },
    {
      "epoch": 0.1995833938865475,
      "grad_norm": 0.10503195226192474,
      "learning_rate": 0.00028205808004663237,
      "loss": 0.39039331674575806,
      "memory(GiB)": 78.26,
      "step": 1030,
      "token_acc": 0.8859542777970212,
      "train_speed(iter/s)": 0.03219
    },
    {
      "epoch": 0.19977716417187424,
      "grad_norm": 0.11032052338123322,
      "learning_rate": 0.0002820124618062251,
      "loss": 0.39245864748954773,
      "memory(GiB)": 78.26,
      "step": 1031,
      "token_acc": 0.8865768832322036,
      "train_speed(iter/s)": 0.032193
    },
    {
      "epoch": 0.199970934457201,
      "grad_norm": 0.10605579614639282,
      "learning_rate": 0.00028196678934475246,
      "loss": 0.42015910148620605,
      "memory(GiB)": 78.26,
      "step": 1032,
      "token_acc": 0.8770413805519122,
      "train_speed(iter/s)": 0.032196
    },
    {
      "epoch": 0.20016470474252773,
      "grad_norm": 0.11556072533130646,
      "learning_rate": 0.00028192106268097334,
      "loss": 0.412725567817688,
      "memory(GiB)": 78.26,
      "step": 1033,
      "token_acc": 0.8829517954994298,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.20035847502785448,
      "grad_norm": 0.11041781306266785,
      "learning_rate": 0.00028187528183366893,
      "loss": 0.4078274369239807,
      "memory(GiB)": 78.26,
      "step": 1034,
      "token_acc": 0.8829038467063116,
      "train_speed(iter/s)": 0.032202
    },
    {
      "epoch": 0.20055224531318122,
      "grad_norm": 0.11429005116224289,
      "learning_rate": 0.0002818294468216426,
      "loss": 0.4424319267272949,
      "memory(GiB)": 78.26,
      "step": 1035,
      "token_acc": 0.8738159769761659,
      "train_speed(iter/s)": 0.032205
    },
    {
      "epoch": 0.20074601559850797,
      "grad_norm": 0.10818706452846527,
      "learning_rate": 0.00028178355766372013,
      "loss": 0.41155439615249634,
      "memory(GiB)": 78.26,
      "step": 1036,
      "token_acc": 0.8821931101407084,
      "train_speed(iter/s)": 0.032207
    },
    {
      "epoch": 0.2009397858838347,
      "grad_norm": 0.10576290637254715,
      "learning_rate": 0.0002817376143787493,
      "loss": 0.425853431224823,
      "memory(GiB)": 78.26,
      "step": 1037,
      "token_acc": 0.8778621912804088,
      "train_speed(iter/s)": 0.03221
    },
    {
      "epoch": 0.20113355616916145,
      "grad_norm": 0.11154097318649292,
      "learning_rate": 0.0002816916169856004,
      "loss": 0.39488485455513,
      "memory(GiB)": 78.26,
      "step": 1038,
      "token_acc": 0.8855062677979039,
      "train_speed(iter/s)": 0.032212
    },
    {
      "epoch": 0.2013273264544882,
      "grad_norm": 0.10686661303043365,
      "learning_rate": 0.00028164556550316563,
      "loss": 0.401602178812027,
      "memory(GiB)": 78.26,
      "step": 1039,
      "token_acc": 0.8835590770383853,
      "train_speed(iter/s)": 0.032215
    },
    {
      "epoch": 0.20152109673981494,
      "grad_norm": 0.11480426788330078,
      "learning_rate": 0.00028159945995035975,
      "loss": 0.4522130489349365,
      "memory(GiB)": 78.26,
      "step": 1040,
      "token_acc": 0.8704494740197641,
      "train_speed(iter/s)": 0.032218
    },
    {
      "epoch": 0.2017148670251417,
      "grad_norm": 0.10808564722537994,
      "learning_rate": 0.0002815533003461193,
      "loss": 0.4125955402851105,
      "memory(GiB)": 78.26,
      "step": 1041,
      "token_acc": 0.8798711669505963,
      "train_speed(iter/s)": 0.032221
    },
    {
      "epoch": 0.20190863731046843,
      "grad_norm": 0.11643750965595245,
      "learning_rate": 0.00028150708670940356,
      "loss": 0.44826722145080566,
      "memory(GiB)": 78.26,
      "step": 1042,
      "token_acc": 0.8713785046728972,
      "train_speed(iter/s)": 0.032223
    },
    {
      "epoch": 0.20210240759579517,
      "grad_norm": 0.1128215417265892,
      "learning_rate": 0.00028146081905919355,
      "loss": 0.4168522357940674,
      "memory(GiB)": 78.26,
      "step": 1043,
      "token_acc": 0.8776015581524763,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.20229617788112192,
      "grad_norm": 0.10721378773450851,
      "learning_rate": 0.00028141449741449264,
      "loss": 0.4148525297641754,
      "memory(GiB)": 78.26,
      "step": 1044,
      "token_acc": 0.880571123565519,
      "train_speed(iter/s)": 0.032228
    },
    {
      "epoch": 0.20248994816644866,
      "grad_norm": 0.11519747227430344,
      "learning_rate": 0.0002813681217943264,
      "loss": 0.44546476006507874,
      "memory(GiB)": 78.26,
      "step": 1045,
      "token_acc": 0.8722711825355682,
      "train_speed(iter/s)": 0.032231
    },
    {
      "epoch": 0.2026837184517754,
      "grad_norm": 0.10372038185596466,
      "learning_rate": 0.00028132169221774256,
      "loss": 0.36725738644599915,
      "memory(GiB)": 78.26,
      "step": 1046,
      "token_acc": 0.8941524609236684,
      "train_speed(iter/s)": 0.032233
    },
    {
      "epoch": 0.20287748873710215,
      "grad_norm": 0.11429732292890549,
      "learning_rate": 0.00028127520870381095,
      "loss": 0.45745980739593506,
      "memory(GiB)": 78.26,
      "step": 1047,
      "token_acc": 0.8703554240277405,
      "train_speed(iter/s)": 0.032236
    },
    {
      "epoch": 0.20307125902242892,
      "grad_norm": 0.11484638601541519,
      "learning_rate": 0.00028122867127162364,
      "loss": 0.4351459741592407,
      "memory(GiB)": 78.26,
      "step": 1048,
      "token_acc": 0.8764367816091954,
      "train_speed(iter/s)": 0.032239
    },
    {
      "epoch": 0.20326502930775567,
      "grad_norm": 0.10716410726308823,
      "learning_rate": 0.0002811820799402948,
      "loss": 0.4354075789451599,
      "memory(GiB)": 78.26,
      "step": 1049,
      "token_acc": 0.8755018944749194,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.2034587995930824,
      "grad_norm": 0.0956411212682724,
      "learning_rate": 0.00028113543472896074,
      "loss": 0.37369605898857117,
      "memory(GiB)": 78.26,
      "step": 1050,
      "token_acc": 0.8899866307131041,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.20365256987840916,
      "grad_norm": 0.10245929658412933,
      "learning_rate": 0.0002810887356567798,
      "loss": 0.3949301242828369,
      "memory(GiB)": 78.26,
      "step": 1051,
      "token_acc": 0.8889433444422662,
      "train_speed(iter/s)": 0.032247
    },
    {
      "epoch": 0.2038463401637359,
      "grad_norm": 0.11300604790449142,
      "learning_rate": 0.0002810419827429327,
      "loss": 0.4184969663619995,
      "memory(GiB)": 78.26,
      "step": 1052,
      "token_acc": 0.8773636026580726,
      "train_speed(iter/s)": 0.032249
    },
    {
      "epoch": 0.20404011044906264,
      "grad_norm": 0.1060163602232933,
      "learning_rate": 0.00028099517600662207,
      "loss": 0.41772571206092834,
      "memory(GiB)": 78.26,
      "step": 1053,
      "token_acc": 0.8804525071341215,
      "train_speed(iter/s)": 0.032252
    },
    {
      "epoch": 0.2042338807343894,
      "grad_norm": 0.11069151014089584,
      "learning_rate": 0.00028094831546707265,
      "loss": 0.4313889741897583,
      "memory(GiB)": 78.26,
      "step": 1054,
      "token_acc": 0.8761178045515395,
      "train_speed(iter/s)": 0.032254
    },
    {
      "epoch": 0.20442765101971613,
      "grad_norm": 0.11553742736577988,
      "learning_rate": 0.00028090140114353133,
      "loss": 0.4176057279109955,
      "memory(GiB)": 78.26,
      "step": 1055,
      "token_acc": 0.8819988121591706,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.20462142130504288,
      "grad_norm": 0.10594391077756882,
      "learning_rate": 0.00028085443305526713,
      "loss": 0.394021213054657,
      "memory(GiB)": 78.26,
      "step": 1056,
      "token_acc": 0.885325837540156,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 0.20481519159036962,
      "grad_norm": 0.11002414673566818,
      "learning_rate": 0.0002808074112215711,
      "loss": 0.41269078850746155,
      "memory(GiB)": 78.26,
      "step": 1057,
      "token_acc": 0.8808142873550017,
      "train_speed(iter/s)": 0.032263
    },
    {
      "epoch": 0.20500896187569637,
      "grad_norm": 0.09443158656358719,
      "learning_rate": 0.0002807603356617563,
      "loss": 0.352535218000412,
      "memory(GiB)": 78.26,
      "step": 1058,
      "token_acc": 0.8964792433000526,
      "train_speed(iter/s)": 0.032265
    },
    {
      "epoch": 0.2052027321610231,
      "grad_norm": 0.11314352601766586,
      "learning_rate": 0.00028071320639515805,
      "loss": 0.4366722106933594,
      "memory(GiB)": 78.26,
      "step": 1059,
      "token_acc": 0.8743527508090615,
      "train_speed(iter/s)": 0.032268
    },
    {
      "epoch": 0.20539650244634985,
      "grad_norm": 0.10947109758853912,
      "learning_rate": 0.00028066602344113353,
      "loss": 0.4176010489463806,
      "memory(GiB)": 78.26,
      "step": 1060,
      "token_acc": 0.8801968582792068,
      "train_speed(iter/s)": 0.03227
    },
    {
      "epoch": 0.2055902727316766,
      "grad_norm": 0.10332785546779633,
      "learning_rate": 0.000280618786819062,
      "loss": 0.39836370944976807,
      "memory(GiB)": 78.26,
      "step": 1061,
      "token_acc": 0.8847858883602182,
      "train_speed(iter/s)": 0.032273
    },
    {
      "epoch": 0.20578404301700334,
      "grad_norm": 0.10330932587385178,
      "learning_rate": 0.0002805714965483449,
      "loss": 0.404694139957428,
      "memory(GiB)": 78.26,
      "step": 1062,
      "token_acc": 0.8835061262959473,
      "train_speed(iter/s)": 0.032275
    },
    {
      "epoch": 0.2059778133023301,
      "grad_norm": 0.12757974863052368,
      "learning_rate": 0.0002805241526484055,
      "loss": 0.4440545439720154,
      "memory(GiB)": 78.26,
      "step": 1063,
      "token_acc": 0.8742229290154692,
      "train_speed(iter/s)": 0.032278
    },
    {
      "epoch": 0.20617158358765683,
      "grad_norm": 0.1015448048710823,
      "learning_rate": 0.00028047675513868936,
      "loss": 0.41053593158721924,
      "memory(GiB)": 78.26,
      "step": 1064,
      "token_acc": 0.8818211780215641,
      "train_speed(iter/s)": 0.03228
    },
    {
      "epoch": 0.20636535387298358,
      "grad_norm": 0.10831795632839203,
      "learning_rate": 0.00028042930403866383,
      "loss": 0.4191955626010895,
      "memory(GiB)": 78.26,
      "step": 1065,
      "token_acc": 0.880605738575983,
      "train_speed(iter/s)": 0.032283
    },
    {
      "epoch": 0.20655912415831032,
      "grad_norm": 0.10870091617107391,
      "learning_rate": 0.0002803817993678183,
      "loss": 0.41718360781669617,
      "memory(GiB)": 78.26,
      "step": 1066,
      "token_acc": 0.8796682921131778,
      "train_speed(iter/s)": 0.032285
    },
    {
      "epoch": 0.20675289444363706,
      "grad_norm": 0.10902893543243408,
      "learning_rate": 0.00028033424114566434,
      "loss": 0.4172331988811493,
      "memory(GiB)": 78.26,
      "step": 1067,
      "token_acc": 0.8794469249603918,
      "train_speed(iter/s)": 0.032288
    },
    {
      "epoch": 0.2069466647289638,
      "grad_norm": 0.11219903081655502,
      "learning_rate": 0.0002802866293917353,
      "loss": 0.42483946681022644,
      "memory(GiB)": 78.26,
      "step": 1068,
      "token_acc": 0.8773620614354346,
      "train_speed(iter/s)": 0.03229
    },
    {
      "epoch": 0.20714043501429055,
      "grad_norm": 0.10504312813282013,
      "learning_rate": 0.00028023896412558664,
      "loss": 0.3886624276638031,
      "memory(GiB)": 78.26,
      "step": 1069,
      "token_acc": 0.8863732842790435,
      "train_speed(iter/s)": 0.032293
    },
    {
      "epoch": 0.2073342052996173,
      "grad_norm": 0.11080943793058395,
      "learning_rate": 0.00028019124536679573,
      "loss": 0.4068402647972107,
      "memory(GiB)": 78.26,
      "step": 1070,
      "token_acc": 0.8815200753561757,
      "train_speed(iter/s)": 0.032296
    },
    {
      "epoch": 0.20752797558494404,
      "grad_norm": 0.11574803292751312,
      "learning_rate": 0.000280143473134962,
      "loss": 0.4340014159679413,
      "memory(GiB)": 78.26,
      "step": 1071,
      "token_acc": 0.8773786767852557,
      "train_speed(iter/s)": 0.032299
    },
    {
      "epoch": 0.20772174587027079,
      "grad_norm": 0.10820218175649643,
      "learning_rate": 0.00028009564744970676,
      "loss": 0.43230772018432617,
      "memory(GiB)": 78.26,
      "step": 1072,
      "token_acc": 0.8758193236979952,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.20791551615559753,
      "grad_norm": 0.11416003853082657,
      "learning_rate": 0.0002800477683306733,
      "loss": 0.4142245352268219,
      "memory(GiB)": 78.26,
      "step": 1073,
      "token_acc": 0.881467683756135,
      "train_speed(iter/s)": 0.032304
    },
    {
      "epoch": 0.20810928644092427,
      "grad_norm": 0.10887296497821808,
      "learning_rate": 0.0002799998357975269,
      "loss": 0.3994034230709076,
      "memory(GiB)": 78.26,
      "step": 1074,
      "token_acc": 0.8853917309454148,
      "train_speed(iter/s)": 0.032306
    },
    {
      "epoch": 0.20830305672625102,
      "grad_norm": 0.10443485528230667,
      "learning_rate": 0.00027995184986995465,
      "loss": 0.41546863317489624,
      "memory(GiB)": 78.26,
      "step": 1075,
      "token_acc": 0.8815385377451952,
      "train_speed(iter/s)": 0.032309
    },
    {
      "epoch": 0.20849682701157776,
      "grad_norm": 0.10981526970863342,
      "learning_rate": 0.0002799038105676658,
      "loss": 0.4088175296783447,
      "memory(GiB)": 78.26,
      "step": 1076,
      "token_acc": 0.8810430263475949,
      "train_speed(iter/s)": 0.032311
    },
    {
      "epoch": 0.2086905972969045,
      "grad_norm": 0.10930271446704865,
      "learning_rate": 0.0002798557179103912,
      "loss": 0.44178467988967896,
      "memory(GiB)": 78.26,
      "step": 1077,
      "token_acc": 0.8711125622398171,
      "train_speed(iter/s)": 0.032314
    },
    {
      "epoch": 0.20888436758223128,
      "grad_norm": 0.10522231459617615,
      "learning_rate": 0.00027980757191788395,
      "loss": 0.4103907644748688,
      "memory(GiB)": 78.26,
      "step": 1078,
      "token_acc": 0.8826355904120181,
      "train_speed(iter/s)": 0.032316
    },
    {
      "epoch": 0.20907813786755802,
      "grad_norm": 0.10563495755195618,
      "learning_rate": 0.00027975937260991886,
      "loss": 0.4192779064178467,
      "memory(GiB)": 78.26,
      "step": 1079,
      "token_acc": 0.8789847870957878,
      "train_speed(iter/s)": 0.032319
    },
    {
      "epoch": 0.20927190815288477,
      "grad_norm": 0.10958468914031982,
      "learning_rate": 0.00027971112000629264,
      "loss": 0.4349307417869568,
      "memory(GiB)": 78.26,
      "step": 1080,
      "token_acc": 0.8740859040094141,
      "train_speed(iter/s)": 0.032321
    },
    {
      "epoch": 0.2094656784382115,
      "grad_norm": 0.09944422543048859,
      "learning_rate": 0.000279662814126824,
      "loss": 0.35245591402053833,
      "memory(GiB)": 78.26,
      "step": 1081,
      "token_acc": 0.896417537322233,
      "train_speed(iter/s)": 0.032324
    },
    {
      "epoch": 0.20965944872353826,
      "grad_norm": 0.10272464156150818,
      "learning_rate": 0.0002796144549913534,
      "loss": 0.4116705358028412,
      "memory(GiB)": 78.26,
      "step": 1082,
      "token_acc": 0.8808783118162493,
      "train_speed(iter/s)": 0.032326
    },
    {
      "epoch": 0.209853219008865,
      "grad_norm": 0.11050140857696533,
      "learning_rate": 0.0002795660426197432,
      "loss": 0.42899516224861145,
      "memory(GiB)": 78.26,
      "step": 1083,
      "token_acc": 0.874714182142658,
      "train_speed(iter/s)": 0.032329
    },
    {
      "epoch": 0.21004698929419174,
      "grad_norm": 0.10963544994592667,
      "learning_rate": 0.0002795175770318778,
      "loss": 0.43252986669540405,
      "memory(GiB)": 78.26,
      "step": 1084,
      "token_acc": 0.8759278350515464,
      "train_speed(iter/s)": 0.032331
    },
    {
      "epoch": 0.2102407595795185,
      "grad_norm": 0.10902027040719986,
      "learning_rate": 0.0002794690582476632,
      "loss": 0.42844176292419434,
      "memory(GiB)": 78.26,
      "step": 1085,
      "token_acc": 0.8774478501489996,
      "train_speed(iter/s)": 0.032334
    },
    {
      "epoch": 0.21043452986484523,
      "grad_norm": 0.10531258583068848,
      "learning_rate": 0.00027942048628702747,
      "loss": 0.39160269498825073,
      "memory(GiB)": 78.26,
      "step": 1086,
      "token_acc": 0.8859537508298296,
      "train_speed(iter/s)": 0.032336
    },
    {
      "epoch": 0.21062830015017198,
      "grad_norm": 0.10602930188179016,
      "learning_rate": 0.0002793718611699203,
      "loss": 0.41201895475387573,
      "memory(GiB)": 78.26,
      "step": 1087,
      "token_acc": 0.8800353219229671,
      "train_speed(iter/s)": 0.032338
    },
    {
      "epoch": 0.21082207043549872,
      "grad_norm": 0.11210530251264572,
      "learning_rate": 0.0002793231829163134,
      "loss": 0.40734055638313293,
      "memory(GiB)": 78.26,
      "step": 1088,
      "token_acc": 0.8804322497961086,
      "train_speed(iter/s)": 0.032341
    },
    {
      "epoch": 0.21101584072082546,
      "grad_norm": 0.12339694052934647,
      "learning_rate": 0.00027927445154620026,
      "loss": 0.4420923888683319,
      "memory(GiB)": 78.26,
      "step": 1089,
      "token_acc": 0.8733834499841585,
      "train_speed(iter/s)": 0.032344
    },
    {
      "epoch": 0.2112096110061522,
      "grad_norm": 0.09618587791919708,
      "learning_rate": 0.00027922566707959607,
      "loss": 0.390455961227417,
      "memory(GiB)": 78.26,
      "step": 1090,
      "token_acc": 0.88821477324435,
      "train_speed(iter/s)": 0.032346
    },
    {
      "epoch": 0.21140338129147895,
      "grad_norm": 0.1139514371752739,
      "learning_rate": 0.00027917682953653805,
      "loss": 0.43686941266059875,
      "memory(GiB)": 78.26,
      "step": 1091,
      "token_acc": 0.8737241340844442,
      "train_speed(iter/s)": 0.032349
    },
    {
      "epoch": 0.2115971515768057,
      "grad_norm": 0.10429059714078903,
      "learning_rate": 0.000279127938937085,
      "loss": 0.41246047616004944,
      "memory(GiB)": 78.26,
      "step": 1092,
      "token_acc": 0.8822545491408587,
      "train_speed(iter/s)": 0.032351
    },
    {
      "epoch": 0.21179092186213244,
      "grad_norm": 0.10676047950983047,
      "learning_rate": 0.0002790789953013176,
      "loss": 0.40249383449554443,
      "memory(GiB)": 78.26,
      "step": 1093,
      "token_acc": 0.8805235195023973,
      "train_speed(iter/s)": 0.032354
    },
    {
      "epoch": 0.21198469214745919,
      "grad_norm": 0.10362397134304047,
      "learning_rate": 0.0002790299986493384,
      "loss": 0.3985450267791748,
      "memory(GiB)": 78.26,
      "step": 1094,
      "token_acc": 0.8856300352798694,
      "train_speed(iter/s)": 0.032356
    },
    {
      "epoch": 0.21217846243278593,
      "grad_norm": 0.12481772899627686,
      "learning_rate": 0.0002789809490012715,
      "loss": 0.48419952392578125,
      "memory(GiB)": 78.26,
      "step": 1095,
      "token_acc": 0.8615846373517498,
      "train_speed(iter/s)": 0.032359
    },
    {
      "epoch": 0.21237223271811267,
      "grad_norm": 0.09449519217014313,
      "learning_rate": 0.00027893184637726304,
      "loss": 0.3890914022922516,
      "memory(GiB)": 78.26,
      "step": 1096,
      "token_acc": 0.8871999258504032,
      "train_speed(iter/s)": 0.032361
    },
    {
      "epoch": 0.21256600300343942,
      "grad_norm": 0.1078861802816391,
      "learning_rate": 0.00027888269079748073,
      "loss": 0.4059605002403259,
      "memory(GiB)": 78.26,
      "step": 1097,
      "token_acc": 0.8837670990726387,
      "train_speed(iter/s)": 0.032363
    },
    {
      "epoch": 0.21275977328876616,
      "grad_norm": 0.1090501993894577,
      "learning_rate": 0.0002788334822821141,
      "loss": 0.4224609136581421,
      "memory(GiB)": 78.26,
      "step": 1098,
      "token_acc": 0.8791557955103201,
      "train_speed(iter/s)": 0.032365
    },
    {
      "epoch": 0.2129535435740929,
      "grad_norm": 0.10729729384183884,
      "learning_rate": 0.00027878422085137437,
      "loss": 0.4196450710296631,
      "memory(GiB)": 78.26,
      "step": 1099,
      "token_acc": 0.8797552019583843,
      "train_speed(iter/s)": 0.032368
    },
    {
      "epoch": 0.21314731385941965,
      "grad_norm": 0.11662878841161728,
      "learning_rate": 0.00027873490652549464,
      "loss": 0.44980597496032715,
      "memory(GiB)": 78.26,
      "step": 1100,
      "token_acc": 0.87146529562982,
      "train_speed(iter/s)": 0.03237
    },
    {
      "epoch": 0.2133410841447464,
      "grad_norm": 0.11625700443983078,
      "learning_rate": 0.00027868553932472955,
      "loss": 0.4741382300853729,
      "memory(GiB)": 78.26,
      "step": 1101,
      "token_acc": 0.8651265929377421,
      "train_speed(iter/s)": 0.032373
    },
    {
      "epoch": 0.21353485443007314,
      "grad_norm": 0.10019952803850174,
      "learning_rate": 0.0002786361192693555,
      "loss": 0.38797110319137573,
      "memory(GiB)": 78.26,
      "step": 1102,
      "token_acc": 0.8859078392670029,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.21372862471539988,
      "grad_norm": 0.10606972128152847,
      "learning_rate": 0.0002785866463796707,
      "loss": 0.41653597354888916,
      "memory(GiB)": 78.26,
      "step": 1103,
      "token_acc": 0.8805738034589087,
      "train_speed(iter/s)": 0.032378
    },
    {
      "epoch": 0.21392239500072663,
      "grad_norm": 0.11103309690952301,
      "learning_rate": 0.000278537120675995,
      "loss": 0.4037442207336426,
      "memory(GiB)": 78.26,
      "step": 1104,
      "token_acc": 0.8811096803090532,
      "train_speed(iter/s)": 0.03238
    },
    {
      "epoch": 0.21411616528605337,
      "grad_norm": 0.10646630078554153,
      "learning_rate": 0.0002784875421786699,
      "loss": 0.38875192403793335,
      "memory(GiB)": 78.26,
      "step": 1105,
      "token_acc": 0.8858713855085408,
      "train_speed(iter/s)": 0.032383
    },
    {
      "epoch": 0.21430993557138012,
      "grad_norm": 0.10254091769456863,
      "learning_rate": 0.0002784379109080586,
      "loss": 0.3985883891582489,
      "memory(GiB)": 78.26,
      "step": 1106,
      "token_acc": 0.8814364477700108,
      "train_speed(iter/s)": 0.032385
    },
    {
      "epoch": 0.21450370585670686,
      "grad_norm": 0.11805382370948792,
      "learning_rate": 0.00027838822688454605,
      "loss": 0.44648706912994385,
      "memory(GiB)": 78.26,
      "step": 1107,
      "token_acc": 0.8738782929855441,
      "train_speed(iter/s)": 0.032388
    },
    {
      "epoch": 0.2146974761420336,
      "grad_norm": 0.10933182388544083,
      "learning_rate": 0.0002783384901285388,
      "loss": 0.43097054958343506,
      "memory(GiB)": 78.26,
      "step": 1108,
      "token_acc": 0.8753957234101638,
      "train_speed(iter/s)": 0.03239
    },
    {
      "epoch": 0.21489124642736038,
      "grad_norm": 0.11015936732292175,
      "learning_rate": 0.00027828870066046505,
      "loss": 0.4059434235095978,
      "memory(GiB)": 78.26,
      "step": 1109,
      "token_acc": 0.8822940702232354,
      "train_speed(iter/s)": 0.032393
    },
    {
      "epoch": 0.21508501671268712,
      "grad_norm": 0.09735988825559616,
      "learning_rate": 0.00027823885850077474,
      "loss": 0.36265531182289124,
      "memory(GiB)": 78.26,
      "step": 1110,
      "token_acc": 0.8935244370144557,
      "train_speed(iter/s)": 0.032395
    },
    {
      "epoch": 0.21527878699801387,
      "grad_norm": 0.10875032097101212,
      "learning_rate": 0.00027818896366993927,
      "loss": 0.41812700033187866,
      "memory(GiB)": 78.26,
      "step": 1111,
      "token_acc": 0.8791208791208791,
      "train_speed(iter/s)": 0.032398
    },
    {
      "epoch": 0.2154725572833406,
      "grad_norm": 0.11396840214729309,
      "learning_rate": 0.0002781390161884519,
      "loss": 0.41851508617401123,
      "memory(GiB)": 78.26,
      "step": 1112,
      "token_acc": 0.8805713529956355,
      "train_speed(iter/s)": 0.0324
    },
    {
      "epoch": 0.21566632756866735,
      "grad_norm": 0.09718259423971176,
      "learning_rate": 0.00027808901607682734,
      "loss": 0.3617076277732849,
      "memory(GiB)": 78.26,
      "step": 1113,
      "token_acc": 0.8916775495666627,
      "train_speed(iter/s)": 0.032402
    },
    {
      "epoch": 0.2158600978539941,
      "grad_norm": 0.11616761237382889,
      "learning_rate": 0.0002780389633556019,
      "loss": 0.40616050362586975,
      "memory(GiB)": 78.26,
      "step": 1114,
      "token_acc": 0.883260254376963,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.21605386813932084,
      "grad_norm": 0.09937416017055511,
      "learning_rate": 0.0002779888580453338,
      "loss": 0.3472048044204712,
      "memory(GiB)": 78.26,
      "step": 1115,
      "token_acc": 0.9003076604876031,
      "train_speed(iter/s)": 0.032407
    },
    {
      "epoch": 0.2162476384246476,
      "grad_norm": 0.10610184073448181,
      "learning_rate": 0.00027793870016660247,
      "loss": 0.43418559432029724,
      "memory(GiB)": 78.26,
      "step": 1116,
      "token_acc": 0.8744638034210666,
      "train_speed(iter/s)": 0.032409
    },
    {
      "epoch": 0.21644140870997433,
      "grad_norm": 0.10146701335906982,
      "learning_rate": 0.0002778884897400091,
      "loss": 0.40924298763275146,
      "memory(GiB)": 78.26,
      "step": 1117,
      "token_acc": 0.8807274502624192,
      "train_speed(iter/s)": 0.032411
    },
    {
      "epoch": 0.21663517899530108,
      "grad_norm": 0.11153874546289444,
      "learning_rate": 0.0002778382267861765,
      "loss": 0.40396177768707275,
      "memory(GiB)": 78.26,
      "step": 1118,
      "token_acc": 0.8842105263157894,
      "train_speed(iter/s)": 0.032414
    },
    {
      "epoch": 0.21682894928062782,
      "grad_norm": 0.1122811809182167,
      "learning_rate": 0.000277787911325749,
      "loss": 0.40342846512794495,
      "memory(GiB)": 78.26,
      "step": 1119,
      "token_acc": 0.882786297835746,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.21702271956595456,
      "grad_norm": 0.1134880930185318,
      "learning_rate": 0.0002777375433793926,
      "loss": 0.4314640164375305,
      "memory(GiB)": 78.26,
      "step": 1120,
      "token_acc": 0.8745123894291522,
      "train_speed(iter/s)": 0.032418
    },
    {
      "epoch": 0.2172164898512813,
      "grad_norm": 0.12474801391363144,
      "learning_rate": 0.0002776871229677946,
      "loss": 0.44416388869285583,
      "memory(GiB)": 78.26,
      "step": 1121,
      "token_acc": 0.8747154810374123,
      "train_speed(iter/s)": 0.032421
    },
    {
      "epoch": 0.21741026013660805,
      "grad_norm": 0.1143270805478096,
      "learning_rate": 0.0002776366501116642,
      "loss": 0.43782278895378113,
      "memory(GiB)": 78.26,
      "step": 1122,
      "token_acc": 0.8753460954479145,
      "train_speed(iter/s)": 0.032423
    },
    {
      "epoch": 0.2176040304219348,
      "grad_norm": 0.10256467759609222,
      "learning_rate": 0.00027758612483173183,
      "loss": 0.3676009774208069,
      "memory(GiB)": 78.26,
      "step": 1123,
      "token_acc": 0.8936442031299904,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.21779780070726154,
      "grad_norm": 0.11031196266412735,
      "learning_rate": 0.00027753554714874957,
      "loss": 0.40533462166786194,
      "memory(GiB)": 78.26,
      "step": 1124,
      "token_acc": 0.8813823163138231,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.21799157099258828,
      "grad_norm": 0.09570778906345367,
      "learning_rate": 0.00027748491708349117,
      "loss": 0.3897078037261963,
      "memory(GiB)": 78.26,
      "step": 1125,
      "token_acc": 0.8868044826994286,
      "train_speed(iter/s)": 0.03243
    },
    {
      "epoch": 0.21818534127791503,
      "grad_norm": 0.10430624336004257,
      "learning_rate": 0.00027743423465675167,
      "loss": 0.3872542083263397,
      "memory(GiB)": 78.26,
      "step": 1126,
      "token_acc": 0.8860606646058733,
      "train_speed(iter/s)": 0.032432
    },
    {
      "epoch": 0.21837911156324177,
      "grad_norm": 0.10934372991323471,
      "learning_rate": 0.0002773834998893476,
      "loss": 0.3712359666824341,
      "memory(GiB)": 78.26,
      "step": 1127,
      "token_acc": 0.8935813061202739,
      "train_speed(iter/s)": 0.032435
    },
    {
      "epoch": 0.21857288184856852,
      "grad_norm": 0.11239821463823318,
      "learning_rate": 0.0002773327128021173,
      "loss": 0.4694061875343323,
      "memory(GiB)": 78.26,
      "step": 1128,
      "token_acc": 0.867498464802099,
      "train_speed(iter/s)": 0.032437
    },
    {
      "epoch": 0.21876665213389526,
      "grad_norm": 0.12323271483182907,
      "learning_rate": 0.00027728187341592025,
      "loss": 0.4466722011566162,
      "memory(GiB)": 78.26,
      "step": 1129,
      "token_acc": 0.8716039707419018,
      "train_speed(iter/s)": 0.032439
    },
    {
      "epoch": 0.218960422419222,
      "grad_norm": 0.1258774995803833,
      "learning_rate": 0.0002772309817516376,
      "loss": 0.4656696021556854,
      "memory(GiB)": 78.26,
      "step": 1130,
      "token_acc": 0.8651681629371566,
      "train_speed(iter/s)": 0.032442
    },
    {
      "epoch": 0.21915419270454875,
      "grad_norm": 0.08737502247095108,
      "learning_rate": 0.0002771800378301719,
      "loss": 0.3400501012802124,
      "memory(GiB)": 78.26,
      "step": 1131,
      "token_acc": 0.9012786337143631,
      "train_speed(iter/s)": 0.032444
    },
    {
      "epoch": 0.2193479629898755,
      "grad_norm": 0.10477188974618912,
      "learning_rate": 0.0002771290416724472,
      "loss": 0.4307701587677002,
      "memory(GiB)": 78.26,
      "step": 1132,
      "token_acc": 0.8758152240789506,
      "train_speed(iter/s)": 0.032447
    },
    {
      "epoch": 0.21954173327520224,
      "grad_norm": 0.10766912251710892,
      "learning_rate": 0.000277077993299409,
      "loss": 0.41765448451042175,
      "memory(GiB)": 78.26,
      "step": 1133,
      "token_acc": 0.8770553935860058,
      "train_speed(iter/s)": 0.032449
    },
    {
      "epoch": 0.21973550356052898,
      "grad_norm": 0.1124100387096405,
      "learning_rate": 0.00027702689273202425,
      "loss": 0.43786174058914185,
      "memory(GiB)": 78.26,
      "step": 1134,
      "token_acc": 0.8744572796884457,
      "train_speed(iter/s)": 0.032452
    },
    {
      "epoch": 0.21992927384585573,
      "grad_norm": 0.0979883000254631,
      "learning_rate": 0.00027697573999128136,
      "loss": 0.3537387549877167,
      "memory(GiB)": 78.26,
      "step": 1135,
      "token_acc": 0.8937490836225014,
      "train_speed(iter/s)": 0.032454
    },
    {
      "epoch": 0.22012304413118247,
      "grad_norm": 0.10655289143323898,
      "learning_rate": 0.00027692453509819,
      "loss": 0.41893625259399414,
      "memory(GiB)": 78.26,
      "step": 1136,
      "token_acc": 0.8789784973184301,
      "train_speed(iter/s)": 0.032456
    },
    {
      "epoch": 0.22031681441650922,
      "grad_norm": 0.10611773282289505,
      "learning_rate": 0.0002768732780737815,
      "loss": 0.39203619956970215,
      "memory(GiB)": 78.26,
      "step": 1137,
      "token_acc": 0.8859876620924084,
      "train_speed(iter/s)": 0.032459
    },
    {
      "epoch": 0.22051058470183596,
      "grad_norm": 0.10541475564241409,
      "learning_rate": 0.0002768219689391085,
      "loss": 0.3815195560455322,
      "memory(GiB)": 78.26,
      "step": 1138,
      "token_acc": 0.8915424198443066,
      "train_speed(iter/s)": 0.032461
    },
    {
      "epoch": 0.22070435498716273,
      "grad_norm": 0.11761334538459778,
      "learning_rate": 0.000276770607715245,
      "loss": 0.4491420090198517,
      "memory(GiB)": 78.26,
      "step": 1139,
      "token_acc": 0.8722042094850799,
      "train_speed(iter/s)": 0.032463
    },
    {
      "epoch": 0.22089812527248948,
      "grad_norm": 0.11273287236690521,
      "learning_rate": 0.0002767191944232865,
      "loss": 0.441051721572876,
      "memory(GiB)": 78.26,
      "step": 1140,
      "token_acc": 0.8757705209438345,
      "train_speed(iter/s)": 0.032466
    },
    {
      "epoch": 0.22109189555781622,
      "grad_norm": 0.10666593909263611,
      "learning_rate": 0.00027666772908434967,
      "loss": 0.41061753034591675,
      "memory(GiB)": 78.26,
      "step": 1141,
      "token_acc": 0.8814865515755317,
      "train_speed(iter/s)": 0.032468
    },
    {
      "epoch": 0.22128566584314296,
      "grad_norm": 0.0970299169421196,
      "learning_rate": 0.0002766162117195729,
      "loss": 0.37377890944480896,
      "memory(GiB)": 78.26,
      "step": 1142,
      "token_acc": 0.8917459199248562,
      "train_speed(iter/s)": 0.03247
    },
    {
      "epoch": 0.2214794361284697,
      "grad_norm": 0.11498536169528961,
      "learning_rate": 0.0002765646423501156,
      "loss": 0.4479539394378662,
      "memory(GiB)": 78.26,
      "step": 1143,
      "token_acc": 0.873067222283145,
      "train_speed(iter/s)": 0.032473
    },
    {
      "epoch": 0.22167320641379645,
      "grad_norm": 0.1132684275507927,
      "learning_rate": 0.00027651302099715886,
      "loss": 0.4535306692123413,
      "memory(GiB)": 78.26,
      "step": 1144,
      "token_acc": 0.873893744313639,
      "train_speed(iter/s)": 0.032475
    },
    {
      "epoch": 0.2218669766991232,
      "grad_norm": 0.11660971492528915,
      "learning_rate": 0.0002764613476819048,
      "loss": 0.42548397183418274,
      "memory(GiB)": 78.26,
      "step": 1145,
      "token_acc": 0.8776913060594131,
      "train_speed(iter/s)": 0.032477
    },
    {
      "epoch": 0.22206074698444994,
      "grad_norm": 0.10114238411188126,
      "learning_rate": 0.0002764096224255771,
      "loss": 0.4015495181083679,
      "memory(GiB)": 78.26,
      "step": 1146,
      "token_acc": 0.8825136612021858,
      "train_speed(iter/s)": 0.032479
    },
    {
      "epoch": 0.22225451726977669,
      "grad_norm": 0.10448514670133591,
      "learning_rate": 0.0002763578452494207,
      "loss": 0.3914845883846283,
      "memory(GiB)": 78.26,
      "step": 1147,
      "token_acc": 0.8844678055190539,
      "train_speed(iter/s)": 0.032482
    },
    {
      "epoch": 0.22244828755510343,
      "grad_norm": 0.11461488157510757,
      "learning_rate": 0.0002763060161747019,
      "loss": 0.4321270287036896,
      "memory(GiB)": 78.26,
      "step": 1148,
      "token_acc": 0.8759273103686238,
      "train_speed(iter/s)": 0.032484
    },
    {
      "epoch": 0.22264205784043017,
      "grad_norm": 0.11510949581861496,
      "learning_rate": 0.00027625413522270833,
      "loss": 0.42202824354171753,
      "memory(GiB)": 78.26,
      "step": 1149,
      "token_acc": 0.8777195685670262,
      "train_speed(iter/s)": 0.032486
    },
    {
      "epoch": 0.22283582812575692,
      "grad_norm": 0.11425234377384186,
      "learning_rate": 0.0002762022024147488,
      "loss": 0.45739829540252686,
      "memory(GiB)": 78.26,
      "step": 1150,
      "token_acc": 0.8678213309024613,
      "train_speed(iter/s)": 0.032488
    },
    {
      "epoch": 0.22302959841108366,
      "grad_norm": 0.11414557695388794,
      "learning_rate": 0.0002761502177721535,
      "loss": 0.4292232096195221,
      "memory(GiB)": 78.26,
      "step": 1151,
      "token_acc": 0.8774583963691377,
      "train_speed(iter/s)": 0.032491
    },
    {
      "epoch": 0.2232233686964104,
      "grad_norm": 0.10301478952169418,
      "learning_rate": 0.00027609818131627407,
      "loss": 0.3989236056804657,
      "memory(GiB)": 78.26,
      "step": 1152,
      "token_acc": 0.8839550828699455,
      "train_speed(iter/s)": 0.032493
    },
    {
      "epoch": 0.22341713898173715,
      "grad_norm": 0.10694558918476105,
      "learning_rate": 0.0002760460930684831,
      "loss": 0.36604946851730347,
      "memory(GiB)": 78.26,
      "step": 1153,
      "token_acc": 0.8924390538257302,
      "train_speed(iter/s)": 0.032496
    },
    {
      "epoch": 0.2236109092670639,
      "grad_norm": 0.13084904849529266,
      "learning_rate": 0.0002759939530501748,
      "loss": 0.42997825145721436,
      "memory(GiB)": 78.26,
      "step": 1154,
      "token_acc": 0.8759590419790543,
      "train_speed(iter/s)": 0.032498
    },
    {
      "epoch": 0.22380467955239064,
      "grad_norm": 0.10484279692173004,
      "learning_rate": 0.00027594176128276435,
      "loss": 0.42240795493125916,
      "memory(GiB)": 78.26,
      "step": 1155,
      "token_acc": 0.8780322748568454,
      "train_speed(iter/s)": 0.032501
    },
    {
      "epoch": 0.22399844983771738,
      "grad_norm": 0.10142414271831512,
      "learning_rate": 0.00027588951778768835,
      "loss": 0.38500741124153137,
      "memory(GiB)": 78.26,
      "step": 1156,
      "token_acc": 0.8893578852952039,
      "train_speed(iter/s)": 0.032503
    },
    {
      "epoch": 0.22419222012304413,
      "grad_norm": 0.11852286010980606,
      "learning_rate": 0.0002758372225864046,
      "loss": 0.41841238737106323,
      "memory(GiB)": 78.26,
      "step": 1157,
      "token_acc": 0.8788539120704378,
      "train_speed(iter/s)": 0.032505
    },
    {
      "epoch": 0.22438599040837087,
      "grad_norm": 0.10002440959215164,
      "learning_rate": 0.0002757848757003922,
      "loss": 0.35776591300964355,
      "memory(GiB)": 78.26,
      "step": 1158,
      "token_acc": 0.8962676092945886,
      "train_speed(iter/s)": 0.032507
    },
    {
      "epoch": 0.22457976069369762,
      "grad_norm": 0.10413894802331924,
      "learning_rate": 0.0002757324771511514,
      "loss": 0.38659295439720154,
      "memory(GiB)": 78.26,
      "step": 1159,
      "token_acc": 0.887719821903838,
      "train_speed(iter/s)": 0.032509
    },
    {
      "epoch": 0.22477353097902436,
      "grad_norm": 0.09892678260803223,
      "learning_rate": 0.0002756800269602036,
      "loss": 0.37748774886131287,
      "memory(GiB)": 78.26,
      "step": 1160,
      "token_acc": 0.8910947249007374,
      "train_speed(iter/s)": 0.032511
    },
    {
      "epoch": 0.2249673012643511,
      "grad_norm": 0.12362544238567352,
      "learning_rate": 0.0002756275251490916,
      "loss": 0.43426865339279175,
      "memory(GiB)": 78.26,
      "step": 1161,
      "token_acc": 0.8750506441941496,
      "train_speed(iter/s)": 0.032513
    },
    {
      "epoch": 0.22516107154967785,
      "grad_norm": 0.10890624672174454,
      "learning_rate": 0.00027557497173937923,
      "loss": 0.3891877830028534,
      "memory(GiB)": 78.26,
      "step": 1162,
      "token_acc": 0.8879472436438945,
      "train_speed(iter/s)": 0.032515
    },
    {
      "epoch": 0.2253548418350046,
      "grad_norm": 0.11398806422948837,
      "learning_rate": 0.00027552236675265174,
      "loss": 0.40979334712028503,
      "memory(GiB)": 78.26,
      "step": 1163,
      "token_acc": 0.881280651429629,
      "train_speed(iter/s)": 0.032518
    },
    {
      "epoch": 0.22554861212033134,
      "grad_norm": 0.10806816816329956,
      "learning_rate": 0.00027546971021051526,
      "loss": 0.39391979575157166,
      "memory(GiB)": 78.26,
      "step": 1164,
      "token_acc": 0.8866101829123446,
      "train_speed(iter/s)": 0.03252
    },
    {
      "epoch": 0.22574238240565808,
      "grad_norm": 0.10718252509832382,
      "learning_rate": 0.00027541700213459726,
      "loss": 0.44121599197387695,
      "memory(GiB)": 78.26,
      "step": 1165,
      "token_acc": 0.8726908749968325,
      "train_speed(iter/s)": 0.032522
    },
    {
      "epoch": 0.22593615269098483,
      "grad_norm": 0.1094525083899498,
      "learning_rate": 0.00027536424254654643,
      "loss": 0.4089512526988983,
      "memory(GiB)": 78.26,
      "step": 1166,
      "token_acc": 0.8825124501342789,
      "train_speed(iter/s)": 0.032524
    },
    {
      "epoch": 0.22612992297631157,
      "grad_norm": 0.12114433944225311,
      "learning_rate": 0.00027531143146803256,
      "loss": 0.4136141538619995,
      "memory(GiB)": 78.26,
      "step": 1167,
      "token_acc": 0.8802897980960939,
      "train_speed(iter/s)": 0.032527
    },
    {
      "epoch": 0.22632369326163831,
      "grad_norm": 0.10031379014253616,
      "learning_rate": 0.00027525856892074646,
      "loss": 0.37666836380958557,
      "memory(GiB)": 78.26,
      "step": 1168,
      "token_acc": 0.8915674529813404,
      "train_speed(iter/s)": 0.032529
    },
    {
      "epoch": 0.2265174635469651,
      "grad_norm": 0.11311759054660797,
      "learning_rate": 0.0002752056549264003,
      "loss": 0.40700918436050415,
      "memory(GiB)": 78.26,
      "step": 1169,
      "token_acc": 0.8810722623836179,
      "train_speed(iter/s)": 0.032531
    },
    {
      "epoch": 0.22671123383229183,
      "grad_norm": 0.11526045948266983,
      "learning_rate": 0.0002751526895067273,
      "loss": 0.4090433120727539,
      "memory(GiB)": 78.26,
      "step": 1170,
      "token_acc": 0.8799791792230562,
      "train_speed(iter/s)": 0.032534
    },
    {
      "epoch": 0.22690500411761858,
      "grad_norm": 0.1078508123755455,
      "learning_rate": 0.0002750996726834817,
      "loss": 0.3891124427318573,
      "memory(GiB)": 78.26,
      "step": 1171,
      "token_acc": 0.887758734679347,
      "train_speed(iter/s)": 0.032536
    },
    {
      "epoch": 0.22709877440294532,
      "grad_norm": 0.12240707129240036,
      "learning_rate": 0.0002750466044784389,
      "loss": 0.4698965549468994,
      "memory(GiB)": 78.26,
      "step": 1172,
      "token_acc": 0.8641207547169811,
      "train_speed(iter/s)": 0.032538
    },
    {
      "epoch": 0.22729254468827206,
      "grad_norm": 0.11151791363954544,
      "learning_rate": 0.00027499348491339564,
      "loss": 0.4254417419433594,
      "memory(GiB)": 78.26,
      "step": 1173,
      "token_acc": 0.8755936161930712,
      "train_speed(iter/s)": 0.03254
    },
    {
      "epoch": 0.2274863149735988,
      "grad_norm": 0.1045253574848175,
      "learning_rate": 0.0002749403140101693,
      "loss": 0.406377375125885,
      "memory(GiB)": 78.26,
      "step": 1174,
      "token_acc": 0.8817293918478969,
      "train_speed(iter/s)": 0.032542
    },
    {
      "epoch": 0.22768008525892555,
      "grad_norm": 0.11521026492118835,
      "learning_rate": 0.00027488709179059886,
      "loss": 0.4493963122367859,
      "memory(GiB)": 78.26,
      "step": 1175,
      "token_acc": 0.8708561714101475,
      "train_speed(iter/s)": 0.032545
    },
    {
      "epoch": 0.2278738555442523,
      "grad_norm": 0.1163213700056076,
      "learning_rate": 0.00027483381827654384,
      "loss": 0.42551738023757935,
      "memory(GiB)": 78.26,
      "step": 1176,
      "token_acc": 0.8768161718256475,
      "train_speed(iter/s)": 0.032547
    },
    {
      "epoch": 0.22806762582957904,
      "grad_norm": 0.11902674287557602,
      "learning_rate": 0.0002747804934898853,
      "loss": 0.4208133816719055,
      "memory(GiB)": 78.26,
      "step": 1177,
      "token_acc": 0.8792419003713664,
      "train_speed(iter/s)": 0.032549
    },
    {
      "epoch": 0.22826139611490578,
      "grad_norm": 0.11104600876569748,
      "learning_rate": 0.00027472711745252514,
      "loss": 0.4281710982322693,
      "memory(GiB)": 78.26,
      "step": 1178,
      "token_acc": 0.8774945561675419,
      "train_speed(iter/s)": 0.032551
    },
    {
      "epoch": 0.22845516640023253,
      "grad_norm": 0.12366992980241776,
      "learning_rate": 0.00027467369018638625,
      "loss": 0.4080888032913208,
      "memory(GiB)": 78.26,
      "step": 1179,
      "token_acc": 0.8825195699186744,
      "train_speed(iter/s)": 0.032554
    },
    {
      "epoch": 0.22864893668555927,
      "grad_norm": 0.1136869415640831,
      "learning_rate": 0.00027462021171341264,
      "loss": 0.401567280292511,
      "memory(GiB)": 78.26,
      "step": 1180,
      "token_acc": 0.8848314606741573,
      "train_speed(iter/s)": 0.032556
    },
    {
      "epoch": 0.22884270697088602,
      "grad_norm": 0.11184633523225784,
      "learning_rate": 0.0002745666820555695,
      "loss": 0.43062543869018555,
      "memory(GiB)": 78.26,
      "step": 1181,
      "token_acc": 0.8777496003087271,
      "train_speed(iter/s)": 0.032558
    },
    {
      "epoch": 0.22903647725621276,
      "grad_norm": 0.1049606204032898,
      "learning_rate": 0.00027451310123484277,
      "loss": 0.41081663966178894,
      "memory(GiB)": 78.26,
      "step": 1182,
      "token_acc": 0.8797746098517307,
      "train_speed(iter/s)": 0.03256
    },
    {
      "epoch": 0.2292302475415395,
      "grad_norm": 0.10288522392511368,
      "learning_rate": 0.0002744594692732395,
      "loss": 0.39744895696640015,
      "memory(GiB)": 78.26,
      "step": 1183,
      "token_acc": 0.8816475626653231,
      "train_speed(iter/s)": 0.032562
    },
    {
      "epoch": 0.22942401782686625,
      "grad_norm": 0.1137736588716507,
      "learning_rate": 0.00027440578619278793,
      "loss": 0.4427635967731476,
      "memory(GiB)": 78.26,
      "step": 1184,
      "token_acc": 0.8732105788154099,
      "train_speed(iter/s)": 0.032564
    },
    {
      "epoch": 0.229617788112193,
      "grad_norm": 0.1281413733959198,
      "learning_rate": 0.000274352052015537,
      "loss": 0.4403104782104492,
      "memory(GiB)": 78.26,
      "step": 1185,
      "token_acc": 0.873717606594319,
      "train_speed(iter/s)": 0.032566
    },
    {
      "epoch": 0.22981155839751974,
      "grad_norm": 0.1134016141295433,
      "learning_rate": 0.00027429826676355685,
      "loss": 0.361285537481308,
      "memory(GiB)": 78.26,
      "step": 1186,
      "token_acc": 0.8938451254627725,
      "train_speed(iter/s)": 0.032569
    },
    {
      "epoch": 0.23000532868284648,
      "grad_norm": 0.1069924458861351,
      "learning_rate": 0.00027424443045893855,
      "loss": 0.42521095275878906,
      "memory(GiB)": 78.26,
      "step": 1187,
      "token_acc": 0.8769493235937237,
      "train_speed(iter/s)": 0.032571
    },
    {
      "epoch": 0.23019909896817323,
      "grad_norm": 0.11626293510198593,
      "learning_rate": 0.0002741905431237941,
      "loss": 0.44659221172332764,
      "memory(GiB)": 78.26,
      "step": 1188,
      "token_acc": 0.8720954699852707,
      "train_speed(iter/s)": 0.032573
    },
    {
      "epoch": 0.23039286925349997,
      "grad_norm": 0.10542602837085724,
      "learning_rate": 0.0002741366047802564,
      "loss": 0.4164351522922516,
      "memory(GiB)": 78.26,
      "step": 1189,
      "token_acc": 0.8788257817485642,
      "train_speed(iter/s)": 0.032575
    },
    {
      "epoch": 0.23058663953882672,
      "grad_norm": 0.12464678287506104,
      "learning_rate": 0.00027408261545047946,
      "loss": 0.47106125950813293,
      "memory(GiB)": 78.26,
      "step": 1190,
      "token_acc": 0.85997171145686,
      "train_speed(iter/s)": 0.032577
    },
    {
      "epoch": 0.23078040982415346,
      "grad_norm": 0.10709256678819656,
      "learning_rate": 0.00027402857515663814,
      "loss": 0.4175183176994324,
      "memory(GiB)": 78.26,
      "step": 1191,
      "token_acc": 0.8793799278364293,
      "train_speed(iter/s)": 0.032579
    },
    {
      "epoch": 0.2309741801094802,
      "grad_norm": 0.10052433609962463,
      "learning_rate": 0.0002739744839209282,
      "loss": 0.38393762707710266,
      "memory(GiB)": 78.26,
      "step": 1192,
      "token_acc": 0.8859593910996107,
      "train_speed(iter/s)": 0.032581
    },
    {
      "epoch": 0.23116795039480695,
      "grad_norm": 0.11317754536867142,
      "learning_rate": 0.0002739203417655664,
      "loss": 0.4203610122203827,
      "memory(GiB)": 78.26,
      "step": 1193,
      "token_acc": 0.878258625139809,
      "train_speed(iter/s)": 0.032583
    },
    {
      "epoch": 0.2313617206801337,
      "grad_norm": 0.11243908107280731,
      "learning_rate": 0.0002738661487127904,
      "loss": 0.3912915885448456,
      "memory(GiB)": 78.26,
      "step": 1194,
      "token_acc": 0.8882206116741347,
      "train_speed(iter/s)": 0.032586
    },
    {
      "epoch": 0.23155549096546044,
      "grad_norm": 0.12153773754835129,
      "learning_rate": 0.00027381190478485863,
      "loss": 0.41862982511520386,
      "memory(GiB)": 78.26,
      "step": 1195,
      "token_acc": 0.8794540140538524,
      "train_speed(iter/s)": 0.032588
    },
    {
      "epoch": 0.23174926125078718,
      "grad_norm": 0.10597037523984909,
      "learning_rate": 0.0002737576100040507,
      "loss": 0.38848644495010376,
      "memory(GiB)": 78.26,
      "step": 1196,
      "token_acc": 0.8870572073398096,
      "train_speed(iter/s)": 0.03259
    },
    {
      "epoch": 0.23194303153611392,
      "grad_norm": 0.10839282721281052,
      "learning_rate": 0.0002737032643926668,
      "loss": 0.3961186110973358,
      "memory(GiB)": 78.26,
      "step": 1197,
      "token_acc": 0.8850163836759011,
      "train_speed(iter/s)": 0.032592
    },
    {
      "epoch": 0.23213680182144067,
      "grad_norm": 0.11460833996534348,
      "learning_rate": 0.0002736488679730282,
      "loss": 0.4457128942012787,
      "memory(GiB)": 78.26,
      "step": 1198,
      "token_acc": 0.8719119445576845,
      "train_speed(iter/s)": 0.032594
    },
    {
      "epoch": 0.2323305721067674,
      "grad_norm": 0.1022963598370552,
      "learning_rate": 0.0002735944207674769,
      "loss": 0.39515194296836853,
      "memory(GiB)": 78.26,
      "step": 1199,
      "token_acc": 0.8834028679578924,
      "train_speed(iter/s)": 0.032596
    },
    {
      "epoch": 0.23252434239209419,
      "grad_norm": 0.11231876909732819,
      "learning_rate": 0.0002735399227983759,
      "loss": 0.4120155870914459,
      "memory(GiB)": 78.26,
      "step": 1200,
      "token_acc": 0.8804409194784245,
      "train_speed(iter/s)": 0.032598
    },
    {
      "epoch": 0.23271811267742093,
      "grad_norm": 0.10544200241565704,
      "learning_rate": 0.00027348537408810903,
      "loss": 0.4136923551559448,
      "memory(GiB)": 78.26,
      "step": 1201,
      "token_acc": 0.8811871376901352,
      "train_speed(iter/s)": 0.03259
    },
    {
      "epoch": 0.23291188296274767,
      "grad_norm": 0.11337179690599442,
      "learning_rate": 0.00027343077465908077,
      "loss": 0.4338820278644562,
      "memory(GiB)": 78.26,
      "step": 1202,
      "token_acc": 0.8770748279047591,
      "train_speed(iter/s)": 0.032592
    },
    {
      "epoch": 0.23310565324807442,
      "grad_norm": 0.11407187581062317,
      "learning_rate": 0.00027337612453371665,
      "loss": 0.4095255136489868,
      "memory(GiB)": 78.26,
      "step": 1203,
      "token_acc": 0.8804728546409807,
      "train_speed(iter/s)": 0.032594
    },
    {
      "epoch": 0.23329942353340116,
      "grad_norm": 0.11431252956390381,
      "learning_rate": 0.00027332142373446297,
      "loss": 0.38796931505203247,
      "memory(GiB)": 78.26,
      "step": 1204,
      "token_acc": 0.8878559697561278,
      "train_speed(iter/s)": 0.032597
    },
    {
      "epoch": 0.2334931938187279,
      "grad_norm": 0.12404537945985794,
      "learning_rate": 0.00027326667228378673,
      "loss": 0.45794570446014404,
      "memory(GiB)": 78.26,
      "step": 1205,
      "token_acc": 0.8690552162153949,
      "train_speed(iter/s)": 0.032599
    },
    {
      "epoch": 0.23368696410405465,
      "grad_norm": 0.10816068947315216,
      "learning_rate": 0.0002732118702041759,
      "loss": 0.42319175601005554,
      "memory(GiB)": 78.26,
      "step": 1206,
      "token_acc": 0.8777571014175469,
      "train_speed(iter/s)": 0.0326
    },
    {
      "epoch": 0.2338807343893814,
      "grad_norm": 0.12030370533466339,
      "learning_rate": 0.0002731570175181392,
      "loss": 0.44653579592704773,
      "memory(GiB)": 78.26,
      "step": 1207,
      "token_acc": 0.8731073377061239,
      "train_speed(iter/s)": 0.032602
    },
    {
      "epoch": 0.23407450467470814,
      "grad_norm": 0.11633247882127762,
      "learning_rate": 0.000273102114248206,
      "loss": 0.44103875756263733,
      "memory(GiB)": 78.26,
      "step": 1208,
      "token_acc": 0.8732270799824535,
      "train_speed(iter/s)": 0.032605
    },
    {
      "epoch": 0.23426827496003488,
      "grad_norm": 0.121209517121315,
      "learning_rate": 0.0002730471604169266,
      "loss": 0.44791626930236816,
      "memory(GiB)": 78.26,
      "step": 1209,
      "token_acc": 0.8704967327820505,
      "train_speed(iter/s)": 0.032607
    },
    {
      "epoch": 0.23446204524536163,
      "grad_norm": 0.10651442408561707,
      "learning_rate": 0.00027299215604687204,
      "loss": 0.3869745433330536,
      "memory(GiB)": 78.26,
      "step": 1210,
      "token_acc": 0.8861974534880438,
      "train_speed(iter/s)": 0.032609
    },
    {
      "epoch": 0.23465581553068837,
      "grad_norm": 0.11880136281251907,
      "learning_rate": 0.000272937101160634,
      "loss": 0.42030277848243713,
      "memory(GiB)": 78.26,
      "step": 1211,
      "token_acc": 0.8782967352155222,
      "train_speed(iter/s)": 0.032611
    },
    {
      "epoch": 0.23484958581601512,
      "grad_norm": 0.1096733883023262,
      "learning_rate": 0.0002728819957808252,
      "loss": 0.40627193450927734,
      "memory(GiB)": 78.26,
      "step": 1212,
      "token_acc": 0.8838417514030116,
      "train_speed(iter/s)": 0.032613
    },
    {
      "epoch": 0.23504335610134186,
      "grad_norm": 0.10282113403081894,
      "learning_rate": 0.0002728268399300786,
      "loss": 0.3663085699081421,
      "memory(GiB)": 78.26,
      "step": 1213,
      "token_acc": 0.8932632961260669,
      "train_speed(iter/s)": 0.032616
    },
    {
      "epoch": 0.2352371263866686,
      "grad_norm": 0.11431027203798294,
      "learning_rate": 0.00027277163363104845,
      "loss": 0.42960235476493835,
      "memory(GiB)": 78.26,
      "step": 1214,
      "token_acc": 0.8752838699642246,
      "train_speed(iter/s)": 0.032618
    },
    {
      "epoch": 0.23543089667199535,
      "grad_norm": 0.11704136431217194,
      "learning_rate": 0.0002727163769064094,
      "loss": 0.42778557538986206,
      "memory(GiB)": 78.26,
      "step": 1215,
      "token_acc": 0.8777720177409135,
      "train_speed(iter/s)": 0.03262
    },
    {
      "epoch": 0.2356246669573221,
      "grad_norm": 0.10846978425979614,
      "learning_rate": 0.00027266106977885674,
      "loss": 0.40238186717033386,
      "memory(GiB)": 78.26,
      "step": 1216,
      "token_acc": 0.88480611332412,
      "train_speed(iter/s)": 0.032622
    },
    {
      "epoch": 0.23581843724264884,
      "grad_norm": 0.1148129478096962,
      "learning_rate": 0.0002726057122711067,
      "loss": 0.4382418096065521,
      "memory(GiB)": 78.26,
      "step": 1217,
      "token_acc": 0.8733262608595923,
      "train_speed(iter/s)": 0.032624
    },
    {
      "epoch": 0.23601220752797558,
      "grad_norm": 0.11113447695970535,
      "learning_rate": 0.00027255030440589614,
      "loss": 0.41040360927581787,
      "memory(GiB)": 78.26,
      "step": 1218,
      "token_acc": 0.8821096468708997,
      "train_speed(iter/s)": 0.032626
    },
    {
      "epoch": 0.23620597781330233,
      "grad_norm": 0.10677137225866318,
      "learning_rate": 0.0002724948462059825,
      "loss": 0.4053942859172821,
      "memory(GiB)": 78.26,
      "step": 1219,
      "token_acc": 0.8811227442521177,
      "train_speed(iter/s)": 0.032628
    },
    {
      "epoch": 0.23639974809862907,
      "grad_norm": 0.11822210252285004,
      "learning_rate": 0.00027243933769414394,
      "loss": 0.4305747449398041,
      "memory(GiB)": 78.26,
      "step": 1220,
      "token_acc": 0.8739931621950513,
      "train_speed(iter/s)": 0.03263
    },
    {
      "epoch": 0.23659351838395581,
      "grad_norm": 0.13333040475845337,
      "learning_rate": 0.00027238377889317935,
      "loss": 0.42261943221092224,
      "memory(GiB)": 78.26,
      "step": 1221,
      "token_acc": 0.8800861924217832,
      "train_speed(iter/s)": 0.032632
    },
    {
      "epoch": 0.23678728866928256,
      "grad_norm": 0.10652173310518265,
      "learning_rate": 0.0002723281698259081,
      "loss": 0.40445369482040405,
      "memory(GiB)": 78.26,
      "step": 1222,
      "token_acc": 0.8841292322492718,
      "train_speed(iter/s)": 0.032634
    },
    {
      "epoch": 0.2369810589546093,
      "grad_norm": 0.10241284221410751,
      "learning_rate": 0.0002722725105151705,
      "loss": 0.3897974491119385,
      "memory(GiB)": 78.26,
      "step": 1223,
      "token_acc": 0.887528428624319,
      "train_speed(iter/s)": 0.032636
    },
    {
      "epoch": 0.23717482923993605,
      "grad_norm": 0.11149775236845016,
      "learning_rate": 0.00027221680098382726,
      "loss": 0.4146060049533844,
      "memory(GiB)": 78.26,
      "step": 1224,
      "token_acc": 0.8800704465401659,
      "train_speed(iter/s)": 0.032638
    },
    {
      "epoch": 0.2373685995252628,
      "grad_norm": 0.11865018308162689,
      "learning_rate": 0.00027216104125475974,
      "loss": 0.390518456697464,
      "memory(GiB)": 78.26,
      "step": 1225,
      "token_acc": 0.8875045692701352,
      "train_speed(iter/s)": 0.032641
    },
    {
      "epoch": 0.23756236981058954,
      "grad_norm": 0.10814682394266129,
      "learning_rate": 0.00027210523135086996,
      "loss": 0.39619240164756775,
      "memory(GiB)": 78.26,
      "step": 1226,
      "token_acc": 0.8837669838825017,
      "train_speed(iter/s)": 0.032643
    },
    {
      "epoch": 0.23775614009591628,
      "grad_norm": 0.10125034302473068,
      "learning_rate": 0.0002720493712950805,
      "loss": 0.3593364953994751,
      "memory(GiB)": 78.26,
      "step": 1227,
      "token_acc": 0.8945650484430858,
      "train_speed(iter/s)": 0.032644
    },
    {
      "epoch": 0.23794991038124302,
      "grad_norm": 0.11461735516786575,
      "learning_rate": 0.0002719934611103348,
      "loss": 0.41169360280036926,
      "memory(GiB)": 78.26,
      "step": 1228,
      "token_acc": 0.8817771660166704,
      "train_speed(iter/s)": 0.032647
    },
    {
      "epoch": 0.23814368066656977,
      "grad_norm": 0.1014741063117981,
      "learning_rate": 0.00027193750081959644,
      "loss": 0.39811670780181885,
      "memory(GiB)": 78.26,
      "step": 1229,
      "token_acc": 0.883717022349185,
      "train_speed(iter/s)": 0.032649
    },
    {
      "epoch": 0.23833745095189654,
      "grad_norm": 0.10858450084924698,
      "learning_rate": 0.00027188149044584997,
      "loss": 0.42627233266830444,
      "memory(GiB)": 78.26,
      "step": 1230,
      "token_acc": 0.8777344179399802,
      "train_speed(iter/s)": 0.032651
    },
    {
      "epoch": 0.23853122123722328,
      "grad_norm": 0.10094203799962997,
      "learning_rate": 0.0002718254300121002,
      "loss": 0.3662244379520416,
      "memory(GiB)": 78.26,
      "step": 1231,
      "token_acc": 0.8925039872408294,
      "train_speed(iter/s)": 0.032653
    },
    {
      "epoch": 0.23872499152255003,
      "grad_norm": 0.10092142224311829,
      "learning_rate": 0.0002717693195413728,
      "loss": 0.3618488907814026,
      "memory(GiB)": 78.26,
      "step": 1232,
      "token_acc": 0.8951329653788259,
      "train_speed(iter/s)": 0.032655
    },
    {
      "epoch": 0.23891876180787677,
      "grad_norm": 0.10050500929355621,
      "learning_rate": 0.0002717131590567138,
      "loss": 0.37352171540260315,
      "memory(GiB)": 78.26,
      "step": 1233,
      "token_acc": 0.8890915338461132,
      "train_speed(iter/s)": 0.032657
    },
    {
      "epoch": 0.23911253209320352,
      "grad_norm": 0.10266852378845215,
      "learning_rate": 0.0002716569485811898,
      "loss": 0.38168448209762573,
      "memory(GiB)": 78.26,
      "step": 1234,
      "token_acc": 0.8879904318660461,
      "train_speed(iter/s)": 0.032659
    },
    {
      "epoch": 0.23930630237853026,
      "grad_norm": 0.10961468517780304,
      "learning_rate": 0.00027160068813788797,
      "loss": 0.423623651266098,
      "memory(GiB)": 78.26,
      "step": 1235,
      "token_acc": 0.8771537798836959,
      "train_speed(iter/s)": 0.032661
    },
    {
      "epoch": 0.239500072663857,
      "grad_norm": 0.10763482749462128,
      "learning_rate": 0.000271544377749916,
      "loss": 0.3990626335144043,
      "memory(GiB)": 78.26,
      "step": 1236,
      "token_acc": 0.8839777513770386,
      "train_speed(iter/s)": 0.032663
    },
    {
      "epoch": 0.23969384294918375,
      "grad_norm": 0.11288584768772125,
      "learning_rate": 0.0002714880174404021,
      "loss": 0.43142029643058777,
      "memory(GiB)": 78.26,
      "step": 1237,
      "token_acc": 0.8738743873247464,
      "train_speed(iter/s)": 0.032665
    },
    {
      "epoch": 0.2398876132345105,
      "grad_norm": 0.1059926375746727,
      "learning_rate": 0.00027143160723249485,
      "loss": 0.39879652857780457,
      "memory(GiB)": 78.26,
      "step": 1238,
      "token_acc": 0.8833674819098997,
      "train_speed(iter/s)": 0.032667
    },
    {
      "epoch": 0.24008138351983724,
      "grad_norm": 0.10575394332408905,
      "learning_rate": 0.00027137514714936357,
      "loss": 0.3970308303833008,
      "memory(GiB)": 78.26,
      "step": 1239,
      "token_acc": 0.8835697867955933,
      "train_speed(iter/s)": 0.032668
    },
    {
      "epoch": 0.24027515380516398,
      "grad_norm": 0.11206567287445068,
      "learning_rate": 0.00027131863721419785,
      "loss": 0.4004877507686615,
      "memory(GiB)": 78.26,
      "step": 1240,
      "token_acc": 0.8835185939591457,
      "train_speed(iter/s)": 0.032671
    },
    {
      "epoch": 0.24046892409049073,
      "grad_norm": 0.11048437654972076,
      "learning_rate": 0.00027126207745020785,
      "loss": 0.41594791412353516,
      "memory(GiB)": 78.26,
      "step": 1241,
      "token_acc": 0.8785890073831009,
      "train_speed(iter/s)": 0.032672
    },
    {
      "epoch": 0.24066269437581747,
      "grad_norm": 0.11345624923706055,
      "learning_rate": 0.0002712054678806242,
      "loss": 0.4346695840358734,
      "memory(GiB)": 78.26,
      "step": 1242,
      "token_acc": 0.8738188213551137,
      "train_speed(iter/s)": 0.032674
    },
    {
      "epoch": 0.24085646466114422,
      "grad_norm": 0.11145740747451782,
      "learning_rate": 0.00027114880852869807,
      "loss": 0.4504337012767792,
      "memory(GiB)": 78.26,
      "step": 1243,
      "token_acc": 0.8700323658937762,
      "train_speed(iter/s)": 0.032676
    },
    {
      "epoch": 0.24105023494647096,
      "grad_norm": 0.11080954968929291,
      "learning_rate": 0.0002710920994177008,
      "loss": 0.43637946248054504,
      "memory(GiB)": 78.26,
      "step": 1244,
      "token_acc": 0.8750756533700138,
      "train_speed(iter/s)": 0.032678
    },
    {
      "epoch": 0.2412440052317977,
      "grad_norm": 0.10086048394441605,
      "learning_rate": 0.00027103534057092447,
      "loss": 0.3988358974456787,
      "memory(GiB)": 78.26,
      "step": 1245,
      "token_acc": 0.8830572217461677,
      "train_speed(iter/s)": 0.03268
    },
    {
      "epoch": 0.24143777551712445,
      "grad_norm": 0.11646781861782074,
      "learning_rate": 0.0002709785320116814,
      "loss": 0.44504404067993164,
      "memory(GiB)": 78.26,
      "step": 1246,
      "token_acc": 0.8725327939193331,
      "train_speed(iter/s)": 0.032682
    },
    {
      "epoch": 0.2416315458024512,
      "grad_norm": 0.11032546311616898,
      "learning_rate": 0.0002709216737633044,
      "loss": 0.44403916597366333,
      "memory(GiB)": 78.26,
      "step": 1247,
      "token_acc": 0.872105901587469,
      "train_speed(iter/s)": 0.032684
    },
    {
      "epoch": 0.24182531608777794,
      "grad_norm": 0.09166669845581055,
      "learning_rate": 0.0002708647658491467,
      "loss": 0.3495952785015106,
      "memory(GiB)": 78.26,
      "step": 1248,
      "token_acc": 0.8977107887579329,
      "train_speed(iter/s)": 0.032685
    },
    {
      "epoch": 0.24201908637310468,
      "grad_norm": 0.10630014538764954,
      "learning_rate": 0.0002708078082925819,
      "loss": 0.4074488878250122,
      "memory(GiB)": 78.26,
      "step": 1249,
      "token_acc": 0.8815855361990353,
      "train_speed(iter/s)": 0.032687
    },
    {
      "epoch": 0.24221285665843142,
      "grad_norm": 0.09900429099798203,
      "learning_rate": 0.000270750801117004,
      "loss": 0.39275825023651123,
      "memory(GiB)": 78.26,
      "step": 1250,
      "token_acc": 0.8847475750400977,
      "train_speed(iter/s)": 0.032689
    },
    {
      "epoch": 0.24240662694375817,
      "grad_norm": 0.10974457859992981,
      "learning_rate": 0.0002706937443458274,
      "loss": 0.4224117398262024,
      "memory(GiB)": 78.26,
      "step": 1251,
      "token_acc": 0.8794166810767191,
      "train_speed(iter/s)": 0.032691
    },
    {
      "epoch": 0.2426003972290849,
      "grad_norm": 0.10894999653100967,
      "learning_rate": 0.0002706366380024868,
      "loss": 0.37922725081443787,
      "memory(GiB)": 78.26,
      "step": 1252,
      "token_acc": 0.8908649728803882,
      "train_speed(iter/s)": 0.032693
    },
    {
      "epoch": 0.24279416751441166,
      "grad_norm": 0.12082328647375107,
      "learning_rate": 0.00027057948211043736,
      "loss": 0.43291229009628296,
      "memory(GiB)": 78.26,
      "step": 1253,
      "token_acc": 0.8723226076593552,
      "train_speed(iter/s)": 0.032695
    },
    {
      "epoch": 0.2429879377997384,
      "grad_norm": 0.10953840613365173,
      "learning_rate": 0.00027052227669315454,
      "loss": 0.38156622648239136,
      "memory(GiB)": 78.26,
      "step": 1254,
      "token_acc": 0.8904028611119321,
      "train_speed(iter/s)": 0.032697
    },
    {
      "epoch": 0.24318170808506515,
      "grad_norm": 0.11070944368839264,
      "learning_rate": 0.00027046502177413415,
      "loss": 0.4315045475959778,
      "memory(GiB)": 78.26,
      "step": 1255,
      "token_acc": 0.8770893778724934,
      "train_speed(iter/s)": 0.032699
    },
    {
      "epoch": 0.2433754783703919,
      "grad_norm": 0.10312207043170929,
      "learning_rate": 0.0002704077173768922,
      "loss": 0.38498249650001526,
      "memory(GiB)": 78.26,
      "step": 1256,
      "token_acc": 0.8877153677921801,
      "train_speed(iter/s)": 0.032701
    },
    {
      "epoch": 0.24356924865571863,
      "grad_norm": 0.11712785810232162,
      "learning_rate": 0.0002703503635249653,
      "loss": 0.4142095446586609,
      "memory(GiB)": 78.26,
      "step": 1257,
      "token_acc": 0.8779902972174204,
      "train_speed(iter/s)": 0.032703
    },
    {
      "epoch": 0.24376301894104538,
      "grad_norm": 0.11441401392221451,
      "learning_rate": 0.0002702929602419102,
      "loss": 0.40991705656051636,
      "memory(GiB)": 78.26,
      "step": 1258,
      "token_acc": 0.8792493116817441,
      "train_speed(iter/s)": 0.032705
    },
    {
      "epoch": 0.24395678922637212,
      "grad_norm": 0.10743826627731323,
      "learning_rate": 0.0002702355075513039,
      "loss": 0.3729992210865021,
      "memory(GiB)": 78.26,
      "step": 1259,
      "token_acc": 0.8912204989885367,
      "train_speed(iter/s)": 0.032707
    },
    {
      "epoch": 0.2441505595116989,
      "grad_norm": 0.12141770869493484,
      "learning_rate": 0.0002701780054767438,
      "loss": 0.456862211227417,
      "memory(GiB)": 78.26,
      "step": 1260,
      "token_acc": 0.868301950047494,
      "train_speed(iter/s)": 0.032708
    },
    {
      "epoch": 0.24434432979702564,
      "grad_norm": 0.11354810744524002,
      "learning_rate": 0.0002701204540418475,
      "loss": 0.43698757886886597,
      "memory(GiB)": 78.26,
      "step": 1261,
      "token_acc": 0.875366250678242,
      "train_speed(iter/s)": 0.03271
    },
    {
      "epoch": 0.24453810008235238,
      "grad_norm": 0.10827391594648361,
      "learning_rate": 0.000270062853270253,
      "loss": 0.39581573009490967,
      "memory(GiB)": 78.26,
      "step": 1262,
      "token_acc": 0.8848191899087549,
      "train_speed(iter/s)": 0.032712
    },
    {
      "epoch": 0.24473187036767913,
      "grad_norm": 0.10404475033283234,
      "learning_rate": 0.0002700052031856184,
      "loss": 0.3672696352005005,
      "memory(GiB)": 78.26,
      "step": 1263,
      "token_acc": 0.8917072443605286,
      "train_speed(iter/s)": 0.032714
    },
    {
      "epoch": 0.24492564065300587,
      "grad_norm": 0.12273856997489929,
      "learning_rate": 0.00026994750381162223,
      "loss": 0.4346576929092407,
      "memory(GiB)": 78.26,
      "step": 1264,
      "token_acc": 0.8755994537428144,
      "train_speed(iter/s)": 0.032716
    },
    {
      "epoch": 0.24511941093833262,
      "grad_norm": 0.11384225636720657,
      "learning_rate": 0.00026988975517196315,
      "loss": 0.4470018148422241,
      "memory(GiB)": 78.26,
      "step": 1265,
      "token_acc": 0.8724487077005063,
      "train_speed(iter/s)": 0.032718
    },
    {
      "epoch": 0.24531318122365936,
      "grad_norm": 0.11222903430461884,
      "learning_rate": 0.00026983195729036004,
      "loss": 0.4389076232910156,
      "memory(GiB)": 78.26,
      "step": 1266,
      "token_acc": 0.8733030464991983,
      "train_speed(iter/s)": 0.032719
    },
    {
      "epoch": 0.2455069515089861,
      "grad_norm": 0.09804686903953552,
      "learning_rate": 0.00026977411019055207,
      "loss": 0.34668290615081787,
      "memory(GiB)": 78.26,
      "step": 1267,
      "token_acc": 0.8978196899835492,
      "train_speed(iter/s)": 0.032721
    },
    {
      "epoch": 0.24570072179431285,
      "grad_norm": 0.10429113358259201,
      "learning_rate": 0.00026971621389629855,
      "loss": 0.4104643762111664,
      "memory(GiB)": 78.26,
      "step": 1268,
      "token_acc": 0.8796510205745658,
      "train_speed(iter/s)": 0.032723
    },
    {
      "epoch": 0.2458944920796396,
      "grad_norm": 0.10256537050008774,
      "learning_rate": 0.0002696582684313791,
      "loss": 0.37179747223854065,
      "memory(GiB)": 78.26,
      "step": 1269,
      "token_acc": 0.8922330825188111,
      "train_speed(iter/s)": 0.032724
    },
    {
      "epoch": 0.24608826236496634,
      "grad_norm": 0.11051318049430847,
      "learning_rate": 0.0002696002738195935,
      "loss": 0.42477959394454956,
      "memory(GiB)": 78.26,
      "step": 1270,
      "token_acc": 0.8768773913540061,
      "train_speed(iter/s)": 0.032726
    },
    {
      "epoch": 0.24628203265029308,
      "grad_norm": 0.10435964167118073,
      "learning_rate": 0.00026954223008476163,
      "loss": 0.3826453983783722,
      "memory(GiB)": 78.26,
      "step": 1271,
      "token_acc": 0.8870310249713184,
      "train_speed(iter/s)": 0.032728
    },
    {
      "epoch": 0.24647580293561983,
      "grad_norm": 0.11433306336402893,
      "learning_rate": 0.0002694841372507236,
      "loss": 0.4138341248035431,
      "memory(GiB)": 78.26,
      "step": 1272,
      "token_acc": 0.8797960325850382,
      "train_speed(iter/s)": 0.03273
    },
    {
      "epoch": 0.24666957322094657,
      "grad_norm": 0.11614301800727844,
      "learning_rate": 0.00026942599534133984,
      "loss": 0.44977250695228577,
      "memory(GiB)": 78.26,
      "step": 1273,
      "token_acc": 0.8709602418042316,
      "train_speed(iter/s)": 0.032732
    },
    {
      "epoch": 0.24686334350627331,
      "grad_norm": 0.18777108192443848,
      "learning_rate": 0.0002693678043804906,
      "loss": 0.3745490312576294,
      "memory(GiB)": 78.26,
      "step": 1274,
      "token_acc": 0.8912466843501327,
      "train_speed(iter/s)": 0.032734
    },
    {
      "epoch": 0.24705711379160006,
      "grad_norm": 0.10248465836048126,
      "learning_rate": 0.0002693095643920766,
      "loss": 0.37728846073150635,
      "memory(GiB)": 78.26,
      "step": 1275,
      "token_acc": 0.8897580936334095,
      "train_speed(iter/s)": 0.032736
    },
    {
      "epoch": 0.2472508840769268,
      "grad_norm": 0.17319722473621368,
      "learning_rate": 0.0002692512754000185,
      "loss": 0.42109817266464233,
      "memory(GiB)": 78.26,
      "step": 1276,
      "token_acc": 0.8798434724156143,
      "train_speed(iter/s)": 0.032737
    },
    {
      "epoch": 0.24744465436225355,
      "grad_norm": 0.1109052449464798,
      "learning_rate": 0.0002691929374282572,
      "loss": 0.3932892084121704,
      "memory(GiB)": 78.26,
      "step": 1277,
      "token_acc": 0.886698000389354,
      "train_speed(iter/s)": 0.032739
    },
    {
      "epoch": 0.2476384246475803,
      "grad_norm": 0.10127067565917969,
      "learning_rate": 0.00026913455050075374,
      "loss": 0.3878341615200043,
      "memory(GiB)": 78.26,
      "step": 1278,
      "token_acc": 0.8887858173572459,
      "train_speed(iter/s)": 0.032741
    },
    {
      "epoch": 0.24783219493290704,
      "grad_norm": 0.11656392365694046,
      "learning_rate": 0.00026907611464148905,
      "loss": 0.4242454171180725,
      "memory(GiB)": 78.26,
      "step": 1279,
      "token_acc": 0.8786354490579843,
      "train_speed(iter/s)": 0.032743
    },
    {
      "epoch": 0.24802596521823378,
      "grad_norm": 0.10642395168542862,
      "learning_rate": 0.00026901762987446436,
      "loss": 0.42408353090286255,
      "memory(GiB)": 78.26,
      "step": 1280,
      "token_acc": 0.8807520778430975,
      "train_speed(iter/s)": 0.032745
    },
    {
      "epoch": 0.24821973550356052,
      "grad_norm": 0.11042629927396774,
      "learning_rate": 0.000268959096223701,
      "loss": 0.3927229344844818,
      "memory(GiB)": 78.26,
      "step": 1281,
      "token_acc": 0.8880031570639305,
      "train_speed(iter/s)": 0.032747
    },
    {
      "epoch": 0.24841350578888727,
      "grad_norm": 0.09721720218658447,
      "learning_rate": 0.0002689005137132402,
      "loss": 0.3534315526485443,
      "memory(GiB)": 78.26,
      "step": 1282,
      "token_acc": 0.8957349486957596,
      "train_speed(iter/s)": 0.032748
    },
    {
      "epoch": 0.248607276074214,
      "grad_norm": 0.10490331053733826,
      "learning_rate": 0.0002688418823671435,
      "loss": 0.39838045835494995,
      "memory(GiB)": 78.26,
      "step": 1283,
      "token_acc": 0.8837405682220588,
      "train_speed(iter/s)": 0.03275
    },
    {
      "epoch": 0.24880104635954076,
      "grad_norm": 0.11232082545757294,
      "learning_rate": 0.0002687832022094923,
      "loss": 0.4016090929508209,
      "memory(GiB)": 78.26,
      "step": 1284,
      "token_acc": 0.8843071140346028,
      "train_speed(iter/s)": 0.032752
    },
    {
      "epoch": 0.2489948166448675,
      "grad_norm": 0.09907688200473785,
      "learning_rate": 0.0002687244732643881,
      "loss": 0.3738039433956146,
      "memory(GiB)": 78.26,
      "step": 1285,
      "token_acc": 0.8901385820445873,
      "train_speed(iter/s)": 0.032754
    },
    {
      "epoch": 0.24918858693019424,
      "grad_norm": 0.11846373975276947,
      "learning_rate": 0.0002686656955559525,
      "loss": 0.45066556334495544,
      "memory(GiB)": 78.26,
      "step": 1286,
      "token_acc": 0.8704005115386851,
      "train_speed(iter/s)": 0.032756
    },
    {
      "epoch": 0.249382357215521,
      "grad_norm": 0.10774929076433182,
      "learning_rate": 0.00026860686910832704,
      "loss": 0.40347960591316223,
      "memory(GiB)": 78.26,
      "step": 1287,
      "token_acc": 0.8835657036827049,
      "train_speed(iter/s)": 0.032758
    },
    {
      "epoch": 0.24957612750084773,
      "grad_norm": 0.10784945636987686,
      "learning_rate": 0.0002685479939456734,
      "loss": 0.41915833950042725,
      "memory(GiB)": 78.26,
      "step": 1288,
      "token_acc": 0.8762773629622079,
      "train_speed(iter/s)": 0.032759
    },
    {
      "epoch": 0.24976989778617448,
      "grad_norm": 0.10360375046730042,
      "learning_rate": 0.000268489070092173,
      "loss": 0.3886149823665619,
      "memory(GiB)": 78.26,
      "step": 1289,
      "token_acc": 0.8865310852948481,
      "train_speed(iter/s)": 0.032761
    },
    {
      "epoch": 0.24996366807150122,
      "grad_norm": 0.10862737894058228,
      "learning_rate": 0.00026843009757202777,
      "loss": 0.4151816666126251,
      "memory(GiB)": 78.26,
      "step": 1290,
      "token_acc": 0.8797664608766098,
      "train_speed(iter/s)": 0.032763
    },
    {
      "epoch": 0.25015743835682797,
      "grad_norm": 0.11150000244379044,
      "learning_rate": 0.00026837107640945905,
      "loss": 0.433391273021698,
      "memory(GiB)": 78.26,
      "step": 1291,
      "token_acc": 0.8773802907537953,
      "train_speed(iter/s)": 0.032765
    },
    {
      "epoch": 0.2503512086421547,
      "grad_norm": 0.10859289765357971,
      "learning_rate": 0.0002683120066287085,
      "loss": 0.41375094652175903,
      "memory(GiB)": 78.26,
      "step": 1292,
      "token_acc": 0.8798327262916498,
      "train_speed(iter/s)": 0.032767
    },
    {
      "epoch": 0.25054497892748145,
      "grad_norm": 0.11398376524448395,
      "learning_rate": 0.0002682528882540376,
      "loss": 0.4081695079803467,
      "memory(GiB)": 78.26,
      "step": 1293,
      "token_acc": 0.8846480067854113,
      "train_speed(iter/s)": 0.032769
    },
    {
      "epoch": 0.2507387492128082,
      "grad_norm": 0.10699091851711273,
      "learning_rate": 0.000268193721309728,
      "loss": 0.40067243576049805,
      "memory(GiB)": 78.26,
      "step": 1294,
      "token_acc": 0.8816360201176189,
      "train_speed(iter/s)": 0.03277
    },
    {
      "epoch": 0.25093251949813494,
      "grad_norm": 0.10412931442260742,
      "learning_rate": 0.00026813450582008103,
      "loss": 0.40070998668670654,
      "memory(GiB)": 78.26,
      "step": 1295,
      "token_acc": 0.8844565031409646,
      "train_speed(iter/s)": 0.032772
    },
    {
      "epoch": 0.2511262897834617,
      "grad_norm": 0.11724124103784561,
      "learning_rate": 0.00026807524180941814,
      "loss": 0.462046355009079,
      "memory(GiB)": 78.26,
      "step": 1296,
      "token_acc": 0.866198113456813,
      "train_speed(iter/s)": 0.032774
    },
    {
      "epoch": 0.25132006006878843,
      "grad_norm": 0.10965701937675476,
      "learning_rate": 0.0002680159293020806,
      "loss": 0.4059637784957886,
      "memory(GiB)": 78.26,
      "step": 1297,
      "token_acc": 0.8829141864372988,
      "train_speed(iter/s)": 0.032776
    },
    {
      "epoch": 0.2515138303541152,
      "grad_norm": 0.10186024010181427,
      "learning_rate": 0.0002679565683224297,
      "loss": 0.3853279948234558,
      "memory(GiB)": 78.26,
      "step": 1298,
      "token_acc": 0.8870169740948418,
      "train_speed(iter/s)": 0.032778
    },
    {
      "epoch": 0.2517076006394419,
      "grad_norm": 0.11461776494979858,
      "learning_rate": 0.00026789715889484657,
      "loss": 0.39657965302467346,
      "memory(GiB)": 78.26,
      "step": 1299,
      "token_acc": 0.8847986900967808,
      "train_speed(iter/s)": 0.03278
    },
    {
      "epoch": 0.25190137092476866,
      "grad_norm": 0.12204183638095856,
      "learning_rate": 0.0002678377010437323,
      "loss": 0.4581944942474365,
      "memory(GiB)": 78.26,
      "step": 1300,
      "token_acc": 0.8694807389051589,
      "train_speed(iter/s)": 0.032781
    },
    {
      "epoch": 0.2520951412100954,
      "grad_norm": 0.10922391712665558,
      "learning_rate": 0.00026777819479350775,
      "loss": 0.39435988664627075,
      "memory(GiB)": 78.26,
      "step": 1301,
      "token_acc": 0.8852810715217581,
      "train_speed(iter/s)": 0.032783
    },
    {
      "epoch": 0.25228891149542215,
      "grad_norm": 0.10757733881473541,
      "learning_rate": 0.00026771864016861377,
      "loss": 0.38533589243888855,
      "memory(GiB)": 78.26,
      "step": 1302,
      "token_acc": 0.8885692617484767,
      "train_speed(iter/s)": 0.032785
    },
    {
      "epoch": 0.2524826817807489,
      "grad_norm": 0.11346130073070526,
      "learning_rate": 0.0002676590371935111,
      "loss": 0.4119528830051422,
      "memory(GiB)": 78.26,
      "step": 1303,
      "token_acc": 0.8788996980878899,
      "train_speed(iter/s)": 0.032787
    },
    {
      "epoch": 0.25267645206607564,
      "grad_norm": 0.11655943840742111,
      "learning_rate": 0.0002675993858926802,
      "loss": 0.4352225661277771,
      "memory(GiB)": 78.26,
      "step": 1304,
      "token_acc": 0.8748795761078998,
      "train_speed(iter/s)": 0.032788
    },
    {
      "epoch": 0.2528702223514024,
      "grad_norm": 0.10136115550994873,
      "learning_rate": 0.00026753968629062146,
      "loss": 0.39623382687568665,
      "memory(GiB)": 78.26,
      "step": 1305,
      "token_acc": 0.882677549344216,
      "train_speed(iter/s)": 0.03279
    },
    {
      "epoch": 0.2530639926367292,
      "grad_norm": 0.10475181043148041,
      "learning_rate": 0.0002674799384118552,
      "loss": 0.3807075023651123,
      "memory(GiB)": 78.26,
      "step": 1306,
      "token_acc": 0.8890237979601748,
      "train_speed(iter/s)": 0.032792
    },
    {
      "epoch": 0.25325776292205593,
      "grad_norm": 0.11299892514944077,
      "learning_rate": 0.0002674201422809214,
      "loss": 0.4387453496456146,
      "memory(GiB)": 78.26,
      "step": 1307,
      "token_acc": 0.8713827248539375,
      "train_speed(iter/s)": 0.032794
    },
    {
      "epoch": 0.2534515332073827,
      "grad_norm": 0.11837562918663025,
      "learning_rate": 0.00026736029792238003,
      "loss": 0.46790987253189087,
      "memory(GiB)": 78.26,
      "step": 1308,
      "token_acc": 0.8648062202398891,
      "train_speed(iter/s)": 0.032796
    },
    {
      "epoch": 0.2536453034927094,
      "grad_norm": 0.10069328546524048,
      "learning_rate": 0.0002673004053608106,
      "loss": 0.3791685700416565,
      "memory(GiB)": 78.26,
      "step": 1309,
      "token_acc": 0.8905825121616565,
      "train_speed(iter/s)": 0.032797
    },
    {
      "epoch": 0.25383907377803616,
      "grad_norm": 0.12146264314651489,
      "learning_rate": 0.0002672404646208128,
      "loss": 0.4696671664714813,
      "memory(GiB)": 78.26,
      "step": 1310,
      "token_acc": 0.8655859144344838,
      "train_speed(iter/s)": 0.032799
    },
    {
      "epoch": 0.2540328440633629,
      "grad_norm": 0.1213066577911377,
      "learning_rate": 0.00026718047572700575,
      "loss": 0.40812620520591736,
      "memory(GiB)": 78.26,
      "step": 1311,
      "token_acc": 0.8800956738768719,
      "train_speed(iter/s)": 0.032801
    },
    {
      "epoch": 0.25422661434868965,
      "grad_norm": 0.1058630719780922,
      "learning_rate": 0.0002671204387040286,
      "loss": 0.4007977247238159,
      "memory(GiB)": 78.26,
      "step": 1312,
      "token_acc": 0.8834997096479827,
      "train_speed(iter/s)": 0.032802
    },
    {
      "epoch": 0.2544203846340164,
      "grad_norm": 0.09740731865167618,
      "learning_rate": 0.00026706035357654007,
      "loss": 0.34724316000938416,
      "memory(GiB)": 78.26,
      "step": 1313,
      "token_acc": 0.8995111614130565,
      "train_speed(iter/s)": 0.032804
    },
    {
      "epoch": 0.25461415491934314,
      "grad_norm": 0.12260702252388,
      "learning_rate": 0.00026700022036921884,
      "loss": 0.47982773184776306,
      "memory(GiB)": 78.26,
      "step": 1314,
      "token_acc": 0.8605628010809541,
      "train_speed(iter/s)": 0.032806
    },
    {
      "epoch": 0.2548079252046699,
      "grad_norm": 0.09923295676708221,
      "learning_rate": 0.00026694003910676315,
      "loss": 0.3682483732700348,
      "memory(GiB)": 78.26,
      "step": 1315,
      "token_acc": 0.8914272901666708,
      "train_speed(iter/s)": 0.032808
    },
    {
      "epoch": 0.2550016954899966,
      "grad_norm": 0.11524137109518051,
      "learning_rate": 0.0002668798098138911,
      "loss": 0.4038434326648712,
      "memory(GiB)": 78.26,
      "step": 1316,
      "token_acc": 0.8844845437065523,
      "train_speed(iter/s)": 0.03281
    },
    {
      "epoch": 0.25519546577532337,
      "grad_norm": 0.11129481345415115,
      "learning_rate": 0.00026681953251534053,
      "loss": 0.4236079454421997,
      "memory(GiB)": 78.26,
      "step": 1317,
      "token_acc": 0.8786745197919983,
      "train_speed(iter/s)": 0.032811
    },
    {
      "epoch": 0.2553892360606501,
      "grad_norm": 0.10136806964874268,
      "learning_rate": 0.00026675920723586886,
      "loss": 0.3650326430797577,
      "memory(GiB)": 78.26,
      "step": 1318,
      "token_acc": 0.8948705179282869,
      "train_speed(iter/s)": 0.032813
    },
    {
      "epoch": 0.25558300634597686,
      "grad_norm": 0.10744194686412811,
      "learning_rate": 0.0002666988340002533,
      "loss": 0.4144882559776306,
      "memory(GiB)": 78.26,
      "step": 1319,
      "token_acc": 0.8820147315987494,
      "train_speed(iter/s)": 0.032815
    },
    {
      "epoch": 0.2557767766313036,
      "grad_norm": 0.10830198973417282,
      "learning_rate": 0.00026663841283329086,
      "loss": 0.4078242778778076,
      "memory(GiB)": 78.26,
      "step": 1320,
      "token_acc": 0.8824059014869888,
      "train_speed(iter/s)": 0.032817
    },
    {
      "epoch": 0.25597054691663035,
      "grad_norm": 0.09699589014053345,
      "learning_rate": 0.000266577943759798,
      "loss": 0.34918874502182007,
      "memory(GiB)": 78.26,
      "step": 1321,
      "token_acc": 0.897431914673294,
      "train_speed(iter/s)": 0.032818
    },
    {
      "epoch": 0.2561643172019571,
      "grad_norm": 0.11712899804115295,
      "learning_rate": 0.00026651742680461115,
      "loss": 0.42532727122306824,
      "memory(GiB)": 78.26,
      "step": 1322,
      "token_acc": 0.8809470377019749,
      "train_speed(iter/s)": 0.03282
    },
    {
      "epoch": 0.25635808748728384,
      "grad_norm": 0.12428336590528488,
      "learning_rate": 0.0002664568619925862,
      "loss": 0.44392460584640503,
      "memory(GiB)": 78.26,
      "step": 1323,
      "token_acc": 0.8718965574699237,
      "train_speed(iter/s)": 0.032822
    },
    {
      "epoch": 0.2565518577726106,
      "grad_norm": 0.1330966353416443,
      "learning_rate": 0.00026639624934859853,
      "loss": 0.46010035276412964,
      "memory(GiB)": 78.26,
      "step": 1324,
      "token_acc": 0.86755315416462,
      "train_speed(iter/s)": 0.032824
    },
    {
      "epoch": 0.2567456280579373,
      "grad_norm": 0.1076708659529686,
      "learning_rate": 0.0002663355888975437,
      "loss": 0.3726995587348938,
      "memory(GiB)": 78.26,
      "step": 1325,
      "token_acc": 0.8889115964031298,
      "train_speed(iter/s)": 0.032826
    },
    {
      "epoch": 0.25693939834326407,
      "grad_norm": 0.12054485827684402,
      "learning_rate": 0.0002662748806643364,
      "loss": 0.45460277795791626,
      "memory(GiB)": 78.26,
      "step": 1326,
      "token_acc": 0.8696869355809753,
      "train_speed(iter/s)": 0.032828
    },
    {
      "epoch": 0.2571331686285908,
      "grad_norm": 0.1159156784415245,
      "learning_rate": 0.00026621412467391125,
      "loss": 0.413094162940979,
      "memory(GiB)": 78.26,
      "step": 1327,
      "token_acc": 0.8797178700263824,
      "train_speed(iter/s)": 0.03283
    },
    {
      "epoch": 0.25732693891391756,
      "grad_norm": 0.10506505519151688,
      "learning_rate": 0.00026615332095122223,
      "loss": 0.3767690360546112,
      "memory(GiB)": 78.26,
      "step": 1328,
      "token_acc": 0.8906741666898722,
      "train_speed(iter/s)": 0.032831
    },
    {
      "epoch": 0.2575207091992443,
      "grad_norm": 0.10869252681732178,
      "learning_rate": 0.00026609246952124323,
      "loss": 0.4055717885494232,
      "memory(GiB)": 78.26,
      "step": 1329,
      "token_acc": 0.8826902784786897,
      "train_speed(iter/s)": 0.032833
    },
    {
      "epoch": 0.25771447948457105,
      "grad_norm": 0.10867463052272797,
      "learning_rate": 0.00026603157040896736,
      "loss": 0.41014426946640015,
      "memory(GiB)": 78.26,
      "step": 1330,
      "token_acc": 0.8821821358824135,
      "train_speed(iter/s)": 0.032835
    },
    {
      "epoch": 0.2579082497698978,
      "grad_norm": 0.11635489761829376,
      "learning_rate": 0.0002659706236394077,
      "loss": 0.41088828444480896,
      "memory(GiB)": 78.26,
      "step": 1331,
      "token_acc": 0.8814593374412788,
      "train_speed(iter/s)": 0.032837
    },
    {
      "epoch": 0.25810202005522453,
      "grad_norm": 0.11617472767829895,
      "learning_rate": 0.00026590962923759664,
      "loss": 0.4540368616580963,
      "memory(GiB)": 78.26,
      "step": 1332,
      "token_acc": 0.8705997580226997,
      "train_speed(iter/s)": 0.032838
    },
    {
      "epoch": 0.2582957903405513,
      "grad_norm": 0.10581759363412857,
      "learning_rate": 0.0002658485872285863,
      "loss": 0.40073099732398987,
      "memory(GiB)": 78.26,
      "step": 1333,
      "token_acc": 0.8806860367272187,
      "train_speed(iter/s)": 0.03284
    },
    {
      "epoch": 0.258489560625878,
      "grad_norm": 0.10285675525665283,
      "learning_rate": 0.0002657874976374481,
      "loss": 0.39908578991889954,
      "memory(GiB)": 78.26,
      "step": 1334,
      "token_acc": 0.8866406011983345,
      "train_speed(iter/s)": 0.032842
    },
    {
      "epoch": 0.25868333091120477,
      "grad_norm": 0.1177554577589035,
      "learning_rate": 0.00026572636048927334,
      "loss": 0.4447701573371887,
      "memory(GiB)": 78.26,
      "step": 1335,
      "token_acc": 0.8721481689350358,
      "train_speed(iter/s)": 0.032844
    },
    {
      "epoch": 0.2588771011965315,
      "grad_norm": 0.10632134228944778,
      "learning_rate": 0.00026566517580917267,
      "loss": 0.405487984418869,
      "memory(GiB)": 78.26,
      "step": 1336,
      "token_acc": 0.881686990206509,
      "train_speed(iter/s)": 0.032845
    },
    {
      "epoch": 0.25907087148185826,
      "grad_norm": 0.10160496830940247,
      "learning_rate": 0.00026560394362227624,
      "loss": 0.3819845914840698,
      "memory(GiB)": 78.26,
      "step": 1337,
      "token_acc": 0.887969021231139,
      "train_speed(iter/s)": 0.032847
    },
    {
      "epoch": 0.259264641767185,
      "grad_norm": 0.10578127205371857,
      "learning_rate": 0.0002655426639537337,
      "loss": 0.40268826484680176,
      "memory(GiB)": 78.26,
      "step": 1338,
      "token_acc": 0.8844663787785317,
      "train_speed(iter/s)": 0.032849
    },
    {
      "epoch": 0.25945841205251174,
      "grad_norm": 0.11469519138336182,
      "learning_rate": 0.0002654813368287144,
      "loss": 0.4174302816390991,
      "memory(GiB)": 78.26,
      "step": 1339,
      "token_acc": 0.8773725181492171,
      "train_speed(iter/s)": 0.03285
    },
    {
      "epoch": 0.2596521823378385,
      "grad_norm": 0.1754118949174881,
      "learning_rate": 0.0002654199622724069,
      "loss": 0.4385494291782379,
      "memory(GiB)": 78.26,
      "step": 1340,
      "token_acc": 0.8754544961198242,
      "train_speed(iter/s)": 0.032852
    },
    {
      "epoch": 0.25984595262316523,
      "grad_norm": 0.12439852207899094,
      "learning_rate": 0.00026535854031001953,
      "loss": 0.46192917227745056,
      "memory(GiB)": 78.26,
      "step": 1341,
      "token_acc": 0.869569399467761,
      "train_speed(iter/s)": 0.032854
    },
    {
      "epoch": 0.260039722908492,
      "grad_norm": 0.10508039593696594,
      "learning_rate": 0.00026529707096677977,
      "loss": 0.3968399167060852,
      "memory(GiB)": 78.26,
      "step": 1342,
      "token_acc": 0.8842156942967024,
      "train_speed(iter/s)": 0.032856
    },
    {
      "epoch": 0.2602334931938187,
      "grad_norm": 0.10867384821176529,
      "learning_rate": 0.0002652355542679349,
      "loss": 0.41717660427093506,
      "memory(GiB)": 78.26,
      "step": 1343,
      "token_acc": 0.8766409013494039,
      "train_speed(iter/s)": 0.032858
    },
    {
      "epoch": 0.26042726347914547,
      "grad_norm": 0.1032068282365799,
      "learning_rate": 0.0002651739902387513,
      "loss": 0.4130229353904724,
      "memory(GiB)": 78.26,
      "step": 1344,
      "token_acc": 0.8800225915765693,
      "train_speed(iter/s)": 0.032859
    },
    {
      "epoch": 0.2606210337644722,
      "grad_norm": 0.11675221472978592,
      "learning_rate": 0.00026511237890451504,
      "loss": 0.45590946078300476,
      "memory(GiB)": 78.26,
      "step": 1345,
      "token_acc": 0.8688484994867459,
      "train_speed(iter/s)": 0.032861
    },
    {
      "epoch": 0.26081480404979895,
      "grad_norm": 0.10524659603834152,
      "learning_rate": 0.00026505072029053167,
      "loss": 0.39327630400657654,
      "memory(GiB)": 78.26,
      "step": 1346,
      "token_acc": 0.8866813620279599,
      "train_speed(iter/s)": 0.032863
    },
    {
      "epoch": 0.2610085743351257,
      "grad_norm": 0.11350347101688385,
      "learning_rate": 0.0002649890144221259,
      "loss": 0.40386056900024414,
      "memory(GiB)": 78.26,
      "step": 1347,
      "token_acc": 0.8825792875394559,
      "train_speed(iter/s)": 0.032865
    },
    {
      "epoch": 0.26120234462045244,
      "grad_norm": 0.10424939543008804,
      "learning_rate": 0.000264927261324642,
      "loss": 0.37450239062309265,
      "memory(GiB)": 78.26,
      "step": 1348,
      "token_acc": 0.8905393229501867,
      "train_speed(iter/s)": 0.032866
    },
    {
      "epoch": 0.2613961149057792,
      "grad_norm": 0.11633609235286713,
      "learning_rate": 0.00026486546102344374,
      "loss": 0.40495017170906067,
      "memory(GiB)": 78.26,
      "step": 1349,
      "token_acc": 0.8839027845893956,
      "train_speed(iter/s)": 0.032868
    },
    {
      "epoch": 0.26158988519110593,
      "grad_norm": 0.12242694199085236,
      "learning_rate": 0.000264803613543914,
      "loss": 0.42412567138671875,
      "memory(GiB)": 78.26,
      "step": 1350,
      "token_acc": 0.8782494429526367,
      "train_speed(iter/s)": 0.03287
    },
    {
      "epoch": 0.2617836554764327,
      "grad_norm": 0.10657955706119537,
      "learning_rate": 0.00026474171891145536,
      "loss": 0.39834100008010864,
      "memory(GiB)": 78.26,
      "step": 1351,
      "token_acc": 0.8844101811062164,
      "train_speed(iter/s)": 0.032872
    },
    {
      "epoch": 0.2619774257617594,
      "grad_norm": 0.12258057296276093,
      "learning_rate": 0.0002646797771514895,
      "loss": 0.44253265857696533,
      "memory(GiB)": 78.26,
      "step": 1352,
      "token_acc": 0.8755840383886855,
      "train_speed(iter/s)": 0.032873
    },
    {
      "epoch": 0.26217119604708616,
      "grad_norm": 0.10096532851457596,
      "learning_rate": 0.0002646177882894576,
      "loss": 0.34528648853302,
      "memory(GiB)": 78.26,
      "step": 1353,
      "token_acc": 0.8981727315320943,
      "train_speed(iter/s)": 0.032875
    },
    {
      "epoch": 0.2623649663324129,
      "grad_norm": 0.10675135254859924,
      "learning_rate": 0.0002645557523508202,
      "loss": 0.3865097761154175,
      "memory(GiB)": 78.26,
      "step": 1354,
      "token_acc": 0.8894203275279439,
      "train_speed(iter/s)": 0.032877
    },
    {
      "epoch": 0.26255873661773965,
      "grad_norm": 0.10857495665550232,
      "learning_rate": 0.00026449366936105696,
      "loss": 0.4091225862503052,
      "memory(GiB)": 78.26,
      "step": 1355,
      "token_acc": 0.8810944625407167,
      "train_speed(iter/s)": 0.032878
    },
    {
      "epoch": 0.2627525069030664,
      "grad_norm": 0.11212481558322906,
      "learning_rate": 0.0002644315393456672,
      "loss": 0.4342299699783325,
      "memory(GiB)": 78.26,
      "step": 1356,
      "token_acc": 0.8745918208676722,
      "train_speed(iter/s)": 0.03288
    },
    {
      "epoch": 0.26294627718839314,
      "grad_norm": 0.10644607990980148,
      "learning_rate": 0.00026436936233016937,
      "loss": 0.3997873067855835,
      "memory(GiB)": 78.26,
      "step": 1357,
      "token_acc": 0.8845487023221603,
      "train_speed(iter/s)": 0.032882
    },
    {
      "epoch": 0.2631400474737199,
      "grad_norm": 0.1112193912267685,
      "learning_rate": 0.0002643071383401012,
      "loss": 0.3876175284385681,
      "memory(GiB)": 78.26,
      "step": 1358,
      "token_acc": 0.8877623875481936,
      "train_speed(iter/s)": 0.032883
    },
    {
      "epoch": 0.26333381775904663,
      "grad_norm": 0.10909376293420792,
      "learning_rate": 0.00026424486740101973,
      "loss": 0.3942829370498657,
      "memory(GiB)": 78.26,
      "step": 1359,
      "token_acc": 0.886093114358897,
      "train_speed(iter/s)": 0.032885
    },
    {
      "epoch": 0.2635275880443734,
      "grad_norm": 0.10721178352832794,
      "learning_rate": 0.00026418254953850136,
      "loss": 0.4055170714855194,
      "memory(GiB)": 78.26,
      "step": 1360,
      "token_acc": 0.8812773628637651,
      "train_speed(iter/s)": 0.032887
    },
    {
      "epoch": 0.2637213583297001,
      "grad_norm": 0.11268987506628036,
      "learning_rate": 0.00026412018477814164,
      "loss": 0.4035116136074066,
      "memory(GiB)": 78.26,
      "step": 1361,
      "token_acc": 0.884084150849012,
      "train_speed(iter/s)": 0.032888
    },
    {
      "epoch": 0.26391512861502686,
      "grad_norm": 0.0998779758810997,
      "learning_rate": 0.0002640577731455556,
      "loss": 0.37671032547950745,
      "memory(GiB)": 78.26,
      "step": 1362,
      "token_acc": 0.8892367801327339,
      "train_speed(iter/s)": 0.03289
    },
    {
      "epoch": 0.2641088989003536,
      "grad_norm": 0.109964519739151,
      "learning_rate": 0.0002639953146663772,
      "loss": 0.38640278577804565,
      "memory(GiB)": 78.26,
      "step": 1363,
      "token_acc": 0.8855807365439093,
      "train_speed(iter/s)": 0.032892
    },
    {
      "epoch": 0.26430266918568035,
      "grad_norm": 0.10154106467962265,
      "learning_rate": 0.0002639328093662599,
      "loss": 0.3642141819000244,
      "memory(GiB)": 78.26,
      "step": 1364,
      "token_acc": 0.894092631464166,
      "train_speed(iter/s)": 0.032893
    },
    {
      "epoch": 0.2644964394710071,
      "grad_norm": 0.10464335232973099,
      "learning_rate": 0.00026387025727087635,
      "loss": 0.3907454311847687,
      "memory(GiB)": 78.26,
      "step": 1365,
      "token_acc": 0.8862689656913484,
      "train_speed(iter/s)": 0.032895
    },
    {
      "epoch": 0.26469020975633384,
      "grad_norm": 0.10311836749315262,
      "learning_rate": 0.00026380765840591834,
      "loss": 0.3700196146965027,
      "memory(GiB)": 78.26,
      "step": 1366,
      "token_acc": 0.8920064072753578,
      "train_speed(iter/s)": 0.032896
    },
    {
      "epoch": 0.26488398004166064,
      "grad_norm": 0.11652354896068573,
      "learning_rate": 0.00026374501279709684,
      "loss": 0.4107242822647095,
      "memory(GiB)": 78.26,
      "step": 1367,
      "token_acc": 0.8811259676284307,
      "train_speed(iter/s)": 0.032898
    },
    {
      "epoch": 0.2650777503269874,
      "grad_norm": 0.11389324814081192,
      "learning_rate": 0.0002636823204701421,
      "loss": 0.4210531711578369,
      "memory(GiB)": 78.26,
      "step": 1368,
      "token_acc": 0.8800777400635059,
      "train_speed(iter/s)": 0.032899
    },
    {
      "epoch": 0.2652715206123141,
      "grad_norm": 0.100397489964962,
      "learning_rate": 0.00026361958145080367,
      "loss": 0.36525481939315796,
      "memory(GiB)": 78.26,
      "step": 1369,
      "token_acc": 0.8945630471509752,
      "train_speed(iter/s)": 0.032901
    },
    {
      "epoch": 0.26546529089764087,
      "grad_norm": 0.11270000785589218,
      "learning_rate": 0.00026355679576485003,
      "loss": 0.44845372438430786,
      "memory(GiB)": 78.26,
      "step": 1370,
      "token_acc": 0.8699417372881356,
      "train_speed(iter/s)": 0.032902
    },
    {
      "epoch": 0.2656590611829676,
      "grad_norm": 0.10077559947967529,
      "learning_rate": 0.00026349396343806897,
      "loss": 0.3875451385974884,
      "memory(GiB)": 78.26,
      "step": 1371,
      "token_acc": 0.8859615802040374,
      "train_speed(iter/s)": 0.032904
    },
    {
      "epoch": 0.26585283146829436,
      "grad_norm": 0.10320444405078888,
      "learning_rate": 0.0002634310844962674,
      "loss": 0.3602311909198761,
      "memory(GiB)": 78.26,
      "step": 1372,
      "token_acc": 0.8927622360728608,
      "train_speed(iter/s)": 0.032906
    },
    {
      "epoch": 0.2660466017536211,
      "grad_norm": 0.10504072159528732,
      "learning_rate": 0.0002633681589652715,
      "loss": 0.38470515608787537,
      "memory(GiB)": 78.26,
      "step": 1373,
      "token_acc": 0.8860445526020218,
      "train_speed(iter/s)": 0.032907
    },
    {
      "epoch": 0.26624037203894785,
      "grad_norm": 0.10138935595750809,
      "learning_rate": 0.00026330518687092626,
      "loss": 0.3703993260860443,
      "memory(GiB)": 78.26,
      "step": 1374,
      "token_acc": 0.8889246858813908,
      "train_speed(iter/s)": 0.032909
    },
    {
      "epoch": 0.2664341423242746,
      "grad_norm": 0.09620644897222519,
      "learning_rate": 0.0002632421682390962,
      "loss": 0.37472549080848694,
      "memory(GiB)": 78.26,
      "step": 1375,
      "token_acc": 0.8922912631003824,
      "train_speed(iter/s)": 0.03291
    },
    {
      "epoch": 0.26662791260960134,
      "grad_norm": 0.10792719572782516,
      "learning_rate": 0.00026317910309566476,
      "loss": 0.3973167836666107,
      "memory(GiB)": 78.26,
      "step": 1376,
      "token_acc": 0.8853990475942648,
      "train_speed(iter/s)": 0.032912
    },
    {
      "epoch": 0.2668216828949281,
      "grad_norm": 0.10119035840034485,
      "learning_rate": 0.00026311599146653443,
      "loss": 0.3766027092933655,
      "memory(GiB)": 78.26,
      "step": 1377,
      "token_acc": 0.8914306730415594,
      "train_speed(iter/s)": 0.032914
    },
    {
      "epoch": 0.2670154531802548,
      "grad_norm": 0.10936938971281052,
      "learning_rate": 0.00026305283337762684,
      "loss": 0.385044664144516,
      "memory(GiB)": 78.26,
      "step": 1378,
      "token_acc": 0.8883062401678028,
      "train_speed(iter/s)": 0.032915
    },
    {
      "epoch": 0.26720922346558157,
      "grad_norm": 0.10729166865348816,
      "learning_rate": 0.0002629896288548827,
      "loss": 0.3800238370895386,
      "memory(GiB)": 78.26,
      "step": 1379,
      "token_acc": 0.8888697301491508,
      "train_speed(iter/s)": 0.032917
    },
    {
      "epoch": 0.2674029937509083,
      "grad_norm": 0.11573578417301178,
      "learning_rate": 0.0002629263779242619,
      "loss": 0.4426755905151367,
      "memory(GiB)": 78.26,
      "step": 1380,
      "token_acc": 0.8741497664796414,
      "train_speed(iter/s)": 0.032919
    },
    {
      "epoch": 0.26759676403623506,
      "grad_norm": 0.09716112166643143,
      "learning_rate": 0.00026286308061174315,
      "loss": 0.36622488498687744,
      "memory(GiB)": 78.26,
      "step": 1381,
      "token_acc": 0.8954976071882019,
      "train_speed(iter/s)": 0.032921
    },
    {
      "epoch": 0.2677905343215618,
      "grad_norm": 0.11388426274061203,
      "learning_rate": 0.0002627997369433246,
      "loss": 0.40867888927459717,
      "memory(GiB)": 78.26,
      "step": 1382,
      "token_acc": 0.8815565203408908,
      "train_speed(iter/s)": 0.032922
    },
    {
      "epoch": 0.26798430460688855,
      "grad_norm": 0.10194810479879379,
      "learning_rate": 0.0002627363469450229,
      "loss": 0.3862283229827881,
      "memory(GiB)": 78.26,
      "step": 1383,
      "token_acc": 0.8868069456729203,
      "train_speed(iter/s)": 0.032924
    },
    {
      "epoch": 0.2681780748922153,
      "grad_norm": 0.11335241794586182,
      "learning_rate": 0.0002626729106428742,
      "loss": 0.3988998532295227,
      "memory(GiB)": 78.26,
      "step": 1384,
      "token_acc": 0.8836738942922578,
      "train_speed(iter/s)": 0.032926
    },
    {
      "epoch": 0.26837184517754203,
      "grad_norm": 0.10034345090389252,
      "learning_rate": 0.0002626094280629335,
      "loss": 0.3520752489566803,
      "memory(GiB)": 78.26,
      "step": 1385,
      "token_acc": 0.8951549706812225,
      "train_speed(iter/s)": 0.032928
    },
    {
      "epoch": 0.2685656154628688,
      "grad_norm": 0.10252499580383301,
      "learning_rate": 0.0002625458992312747,
      "loss": 0.37120020389556885,
      "memory(GiB)": 78.26,
      "step": 1386,
      "token_acc": 0.892410527781646,
      "train_speed(iter/s)": 0.032929
    },
    {
      "epoch": 0.2687593857481955,
      "grad_norm": 0.11847876012325287,
      "learning_rate": 0.0002624823241739909,
      "loss": 0.4186937212944031,
      "memory(GiB)": 78.26,
      "step": 1387,
      "token_acc": 0.8775456306121251,
      "train_speed(iter/s)": 0.032931
    },
    {
      "epoch": 0.26895315603352227,
      "grad_norm": 0.11340092122554779,
      "learning_rate": 0.000262418702917194,
      "loss": 0.41182029247283936,
      "memory(GiB)": 78.26,
      "step": 1388,
      "token_acc": 0.8774313059586292,
      "train_speed(iter/s)": 0.032932
    },
    {
      "epoch": 0.269146926318849,
      "grad_norm": 0.11647920310497284,
      "learning_rate": 0.0002623550354870151,
      "loss": 0.41684606671333313,
      "memory(GiB)": 78.26,
      "step": 1389,
      "token_acc": 0.8813554314750806,
      "train_speed(iter/s)": 0.032934
    },
    {
      "epoch": 0.26934069660417576,
      "grad_norm": 0.10434102267026901,
      "learning_rate": 0.00026229132190960395,
      "loss": 0.35181209444999695,
      "memory(GiB)": 78.26,
      "step": 1390,
      "token_acc": 0.8950941151016876,
      "train_speed(iter/s)": 0.032935
    },
    {
      "epoch": 0.2695344668895025,
      "grad_norm": 0.11331314593553543,
      "learning_rate": 0.0002622275622111295,
      "loss": 0.3922780156135559,
      "memory(GiB)": 78.26,
      "step": 1391,
      "token_acc": 0.8852905742642418,
      "train_speed(iter/s)": 0.032937
    },
    {
      "epoch": 0.26972823717482924,
      "grad_norm": 0.10961979627609253,
      "learning_rate": 0.00026216375641777964,
      "loss": 0.40969350934028625,
      "memory(GiB)": 78.26,
      "step": 1392,
      "token_acc": 0.8799139553643452,
      "train_speed(iter/s)": 0.032939
    },
    {
      "epoch": 0.269922007460156,
      "grad_norm": 0.10636827349662781,
      "learning_rate": 0.000262099904555761,
      "loss": 0.39670512080192566,
      "memory(GiB)": 78.26,
      "step": 1393,
      "token_acc": 0.8852959931980019,
      "train_speed(iter/s)": 0.03294
    },
    {
      "epoch": 0.27011577774548273,
      "grad_norm": 0.11483940482139587,
      "learning_rate": 0.00026203600665129935,
      "loss": 0.40846797823905945,
      "memory(GiB)": 78.26,
      "step": 1394,
      "token_acc": 0.8796793825144723,
      "train_speed(iter/s)": 0.032942
    },
    {
      "epoch": 0.2703095480308095,
      "grad_norm": 0.10763109475374222,
      "learning_rate": 0.0002619720627306393,
      "loss": 0.4060547351837158,
      "memory(GiB)": 78.26,
      "step": 1395,
      "token_acc": 0.8830100404773169,
      "train_speed(iter/s)": 0.032943
    },
    {
      "epoch": 0.2705033183161362,
      "grad_norm": 0.11100554466247559,
      "learning_rate": 0.00026190807282004414,
      "loss": 0.4114495813846588,
      "memory(GiB)": 78.26,
      "step": 1396,
      "token_acc": 0.8800386130606068,
      "train_speed(iter/s)": 0.032945
    },
    {
      "epoch": 0.27069708860146297,
      "grad_norm": 0.1212446540594101,
      "learning_rate": 0.0002618440369457965,
      "loss": 0.45782333612442017,
      "memory(GiB)": 78.26,
      "step": 1397,
      "token_acc": 0.868626656635222,
      "train_speed(iter/s)": 0.032946
    },
    {
      "epoch": 0.2708908588867897,
      "grad_norm": 0.10512025654315948,
      "learning_rate": 0.0002617799551341975,
      "loss": 0.40602394938468933,
      "memory(GiB)": 78.26,
      "step": 1398,
      "token_acc": 0.8847831850941259,
      "train_speed(iter/s)": 0.032948
    },
    {
      "epoch": 0.27108462917211645,
      "grad_norm": 0.1056622564792633,
      "learning_rate": 0.00026171582741156725,
      "loss": 0.3937338590621948,
      "memory(GiB)": 78.26,
      "step": 1399,
      "token_acc": 0.8871733032741762,
      "train_speed(iter/s)": 0.032949
    },
    {
      "epoch": 0.2712783994574432,
      "grad_norm": 0.10223392397165298,
      "learning_rate": 0.0002616516538042448,
      "loss": 0.4050544500350952,
      "memory(GiB)": 78.26,
      "step": 1400,
      "token_acc": 0.883415804468436,
      "train_speed(iter/s)": 0.032951
    },
    {
      "epoch": 0.27147216974276994,
      "grad_norm": 0.10160177946090698,
      "learning_rate": 0.0002615874343385879,
      "loss": 0.38988304138183594,
      "memory(GiB)": 78.26,
      "step": 1401,
      "token_acc": 0.8859642291617994,
      "train_speed(iter/s)": 0.032943
    },
    {
      "epoch": 0.2716659400280967,
      "grad_norm": 0.12478124350309372,
      "learning_rate": 0.00026152316904097327,
      "loss": 0.4281242787837982,
      "memory(GiB)": 78.26,
      "step": 1402,
      "token_acc": 0.875607687959299,
      "train_speed(iter/s)": 0.032945
    },
    {
      "epoch": 0.27185971031342343,
      "grad_norm": 0.10745180398225784,
      "learning_rate": 0.00026145885793779633,
      "loss": 0.3870825171470642,
      "memory(GiB)": 78.26,
      "step": 1403,
      "token_acc": 0.8886628235560077,
      "train_speed(iter/s)": 0.032946
    },
    {
      "epoch": 0.2720534805987502,
      "grad_norm": 0.10077864676713943,
      "learning_rate": 0.0002613945010554715,
      "loss": 0.37229087948799133,
      "memory(GiB)": 78.26,
      "step": 1404,
      "token_acc": 0.8895456714202407,
      "train_speed(iter/s)": 0.032948
    },
    {
      "epoch": 0.2722472508840769,
      "grad_norm": 0.1082373782992363,
      "learning_rate": 0.00026133009842043174,
      "loss": 0.418215274810791,
      "memory(GiB)": 78.26,
      "step": 1405,
      "token_acc": 0.8777365051766379,
      "train_speed(iter/s)": 0.03295
    },
    {
      "epoch": 0.27244102116940366,
      "grad_norm": 0.1040397435426712,
      "learning_rate": 0.00026126565005912903,
      "loss": 0.397979736328125,
      "memory(GiB)": 78.26,
      "step": 1406,
      "token_acc": 0.8846828451006571,
      "train_speed(iter/s)": 0.032951
    },
    {
      "epoch": 0.2726347914547304,
      "grad_norm": 0.10922178626060486,
      "learning_rate": 0.000261201155998034,
      "loss": 0.41443806886672974,
      "memory(GiB)": 78.26,
      "step": 1407,
      "token_acc": 0.8808320476671809,
      "train_speed(iter/s)": 0.032953
    },
    {
      "epoch": 0.27282856174005715,
      "grad_norm": 0.11000542342662811,
      "learning_rate": 0.0002611366162636361,
      "loss": 0.4330993890762329,
      "memory(GiB)": 78.26,
      "step": 1408,
      "token_acc": 0.8738227795410791,
      "train_speed(iter/s)": 0.032954
    },
    {
      "epoch": 0.2730223320253839,
      "grad_norm": 0.1046384945511818,
      "learning_rate": 0.00026107203088244357,
      "loss": 0.37256211042404175,
      "memory(GiB)": 78.26,
      "step": 1409,
      "token_acc": 0.8898919399606564,
      "train_speed(iter/s)": 0.032956
    },
    {
      "epoch": 0.27321610231071064,
      "grad_norm": 0.10583927482366562,
      "learning_rate": 0.0002610073998809833,
      "loss": 0.38221731781959534,
      "memory(GiB)": 78.26,
      "step": 1410,
      "token_acc": 0.8893005757282106,
      "train_speed(iter/s)": 0.032958
    },
    {
      "epoch": 0.2734098725960374,
      "grad_norm": 0.11508861929178238,
      "learning_rate": 0.0002609427232858011,
      "loss": 0.4040870666503906,
      "memory(GiB)": 78.26,
      "step": 1411,
      "token_acc": 0.8844921965991148,
      "train_speed(iter/s)": 0.032959
    },
    {
      "epoch": 0.27360364288136413,
      "grad_norm": 0.10771467536687851,
      "learning_rate": 0.0002608780011234612,
      "loss": 0.3735302686691284,
      "memory(GiB)": 78.26,
      "step": 1412,
      "token_acc": 0.8907560765756076,
      "train_speed(iter/s)": 0.032961
    },
    {
      "epoch": 0.2737974131666909,
      "grad_norm": 0.11977894604206085,
      "learning_rate": 0.0002608132334205469,
      "loss": 0.43817439675331116,
      "memory(GiB)": 78.26,
      "step": 1413,
      "token_acc": 0.8712272477356252,
      "train_speed(iter/s)": 0.032963
    },
    {
      "epoch": 0.2739911834520176,
      "grad_norm": 0.11429792642593384,
      "learning_rate": 0.00026074842020365994,
      "loss": 0.41938716173171997,
      "memory(GiB)": 78.26,
      "step": 1414,
      "token_acc": 0.8812332907125421,
      "train_speed(iter/s)": 0.032964
    },
    {
      "epoch": 0.27418495373734436,
      "grad_norm": 0.12319884449243546,
      "learning_rate": 0.00026068356149942085,
      "loss": 0.4341338574886322,
      "memory(GiB)": 78.26,
      "step": 1415,
      "token_acc": 0.8744184650478735,
      "train_speed(iter/s)": 0.032966
    },
    {
      "epoch": 0.2743787240226711,
      "grad_norm": 0.10541116446256638,
      "learning_rate": 0.00026061865733446887,
      "loss": 0.39338019490242004,
      "memory(GiB)": 78.26,
      "step": 1416,
      "token_acc": 0.8870288248337029,
      "train_speed(iter/s)": 0.032967
    },
    {
      "epoch": 0.27457249430799785,
      "grad_norm": 0.10803982615470886,
      "learning_rate": 0.00026055370773546193,
      "loss": 0.3971567749977112,
      "memory(GiB)": 78.26,
      "step": 1417,
      "token_acc": 0.8830211291457155,
      "train_speed(iter/s)": 0.032969
    },
    {
      "epoch": 0.2747662645933246,
      "grad_norm": 0.11194406449794769,
      "learning_rate": 0.00026048871272907657,
      "loss": 0.386283278465271,
      "memory(GiB)": 78.26,
      "step": 1418,
      "token_acc": 0.8865788499180024,
      "train_speed(iter/s)": 0.032971
    },
    {
      "epoch": 0.27496003487865134,
      "grad_norm": 0.09532847255468369,
      "learning_rate": 0.00026042367234200783,
      "loss": 0.365975558757782,
      "memory(GiB)": 78.26,
      "step": 1419,
      "token_acc": 0.8936005233952846,
      "train_speed(iter/s)": 0.032972
    },
    {
      "epoch": 0.2751538051639781,
      "grad_norm": 0.10494247823953629,
      "learning_rate": 0.0002603585866009697,
      "loss": 0.4001488983631134,
      "memory(GiB)": 78.26,
      "step": 1420,
      "token_acc": 0.8836253515033528,
      "train_speed(iter/s)": 0.032974
    },
    {
      "epoch": 0.2753475754493048,
      "grad_norm": 0.11904102563858032,
      "learning_rate": 0.00026029345553269466,
      "loss": 0.4139590263366699,
      "memory(GiB)": 78.26,
      "step": 1421,
      "token_acc": 0.881683852450303,
      "train_speed(iter/s)": 0.032975
    },
    {
      "epoch": 0.27554134573463157,
      "grad_norm": 0.09683738648891449,
      "learning_rate": 0.00026022827916393366,
      "loss": 0.3371363878250122,
      "memory(GiB)": 78.26,
      "step": 1422,
      "token_acc": 0.8991735972207601,
      "train_speed(iter/s)": 0.032977
    },
    {
      "epoch": 0.2757351160199583,
      "grad_norm": 0.09993797540664673,
      "learning_rate": 0.0002601630575214565,
      "loss": 0.3939957916736603,
      "memory(GiB)": 78.26,
      "step": 1423,
      "token_acc": 0.8834102564102564,
      "train_speed(iter/s)": 0.032978
    },
    {
      "epoch": 0.27592888630528506,
      "grad_norm": 0.11129175126552582,
      "learning_rate": 0.0002600977906320514,
      "loss": 0.41668498516082764,
      "memory(GiB)": 78.26,
      "step": 1424,
      "token_acc": 0.8775945117501095,
      "train_speed(iter/s)": 0.03298
    },
    {
      "epoch": 0.2761226565906118,
      "grad_norm": 0.10565146058797836,
      "learning_rate": 0.00026003247852252525,
      "loss": 0.37214991450309753,
      "memory(GiB)": 78.26,
      "step": 1425,
      "token_acc": 0.8892408742152987,
      "train_speed(iter/s)": 0.032981
    },
    {
      "epoch": 0.27631642687593855,
      "grad_norm": 0.11600112915039062,
      "learning_rate": 0.0002599671212197035,
      "loss": 0.42271527647972107,
      "memory(GiB)": 78.26,
      "step": 1426,
      "token_acc": 0.8791572967215056,
      "train_speed(iter/s)": 0.032983
    },
    {
      "epoch": 0.27651019716126535,
      "grad_norm": 0.1128147765994072,
      "learning_rate": 0.0002599017187504301,
      "loss": 0.4270755648612976,
      "memory(GiB)": 78.26,
      "step": 1427,
      "token_acc": 0.875413155465563,
      "train_speed(iter/s)": 0.032985
    },
    {
      "epoch": 0.2767039674465921,
      "grad_norm": 0.09815947711467743,
      "learning_rate": 0.0002598362711415677,
      "loss": 0.35846811532974243,
      "memory(GiB)": 78.26,
      "step": 1428,
      "token_acc": 0.8941823402029246,
      "train_speed(iter/s)": 0.032986
    },
    {
      "epoch": 0.27689773773191884,
      "grad_norm": 0.11663207411766052,
      "learning_rate": 0.0002597707784199973,
      "loss": 0.41208210587501526,
      "memory(GiB)": 78.26,
      "step": 1429,
      "token_acc": 0.8810604466620577,
      "train_speed(iter/s)": 0.032988
    },
    {
      "epoch": 0.2770915080172456,
      "grad_norm": 0.10307564586400986,
      "learning_rate": 0.0002597052406126185,
      "loss": 0.37274155020713806,
      "memory(GiB)": 78.26,
      "step": 1430,
      "token_acc": 0.8910128388017119,
      "train_speed(iter/s)": 0.032989
    },
    {
      "epoch": 0.2772852783025723,
      "grad_norm": 0.11798793822526932,
      "learning_rate": 0.0002596396577463495,
      "loss": 0.4077189266681671,
      "memory(GiB)": 78.26,
      "step": 1431,
      "token_acc": 0.8813992194674013,
      "train_speed(iter/s)": 0.032991
    },
    {
      "epoch": 0.27747904858789907,
      "grad_norm": 0.11641109734773636,
      "learning_rate": 0.00025957402984812695,
      "loss": 0.40781551599502563,
      "memory(GiB)": 78.26,
      "step": 1432,
      "token_acc": 0.8810986109516703,
      "train_speed(iter/s)": 0.032992
    },
    {
      "epoch": 0.2776728188732258,
      "grad_norm": 0.11191576719284058,
      "learning_rate": 0.000259508356944906,
      "loss": 0.4027900695800781,
      "memory(GiB)": 78.26,
      "step": 1433,
      "token_acc": 0.8829726161088467,
      "train_speed(iter/s)": 0.032994
    },
    {
      "epoch": 0.27786658915855256,
      "grad_norm": 0.11352302134037018,
      "learning_rate": 0.0002594426390636602,
      "loss": 0.423457533121109,
      "memory(GiB)": 78.26,
      "step": 1434,
      "token_acc": 0.8779633306084316,
      "train_speed(iter/s)": 0.032996
    },
    {
      "epoch": 0.2780603594438793,
      "grad_norm": 0.10580118000507355,
      "learning_rate": 0.00025937687623138174,
      "loss": 0.41249287128448486,
      "memory(GiB)": 78.26,
      "step": 1435,
      "token_acc": 0.8809974249829208,
      "train_speed(iter/s)": 0.032997
    },
    {
      "epoch": 0.27825412972920605,
      "grad_norm": 0.10057955980300903,
      "learning_rate": 0.00025931106847508115,
      "loss": 0.37016963958740234,
      "memory(GiB)": 78.26,
      "step": 1436,
      "token_acc": 0.8921429696110849,
      "train_speed(iter/s)": 0.032999
    },
    {
      "epoch": 0.2784479000145328,
      "grad_norm": 0.11049242317676544,
      "learning_rate": 0.0002592452158217873,
      "loss": 0.39204567670822144,
      "memory(GiB)": 78.26,
      "step": 1437,
      "token_acc": 0.8855197058366565,
      "train_speed(iter/s)": 0.033
    },
    {
      "epoch": 0.27864167029985953,
      "grad_norm": 0.1143956407904625,
      "learning_rate": 0.00025917931829854795,
      "loss": 0.41014882922172546,
      "memory(GiB)": 78.26,
      "step": 1438,
      "token_acc": 0.8808566548416765,
      "train_speed(iter/s)": 0.033002
    },
    {
      "epoch": 0.2788354405851863,
      "grad_norm": 0.10722561925649643,
      "learning_rate": 0.00025911337593242874,
      "loss": 0.4017128646373749,
      "memory(GiB)": 78.26,
      "step": 1439,
      "token_acc": 0.8847192571825407,
      "train_speed(iter/s)": 0.033004
    },
    {
      "epoch": 0.279029210870513,
      "grad_norm": 0.10540410131216049,
      "learning_rate": 0.0002590473887505141,
      "loss": 0.3711482286453247,
      "memory(GiB)": 78.26,
      "step": 1440,
      "token_acc": 0.8929875288413797,
      "train_speed(iter/s)": 0.033005
    },
    {
      "epoch": 0.27922298115583977,
      "grad_norm": 0.10699062794446945,
      "learning_rate": 0.0002589813567799066,
      "loss": 0.37242835760116577,
      "memory(GiB)": 78.26,
      "step": 1441,
      "token_acc": 0.8892943052514933,
      "train_speed(iter/s)": 0.033007
    },
    {
      "epoch": 0.2794167514411665,
      "grad_norm": 0.11198758333921432,
      "learning_rate": 0.0002589152800477275,
      "loss": 0.38732579350471497,
      "memory(GiB)": 78.26,
      "step": 1442,
      "token_acc": 0.8864687236455637,
      "train_speed(iter/s)": 0.033008
    },
    {
      "epoch": 0.27961052172649326,
      "grad_norm": 0.1112254187464714,
      "learning_rate": 0.00025884915858111614,
      "loss": 0.41368624567985535,
      "memory(GiB)": 78.26,
      "step": 1443,
      "token_acc": 0.8818168540966846,
      "train_speed(iter/s)": 0.033009
    },
    {
      "epoch": 0.27980429201182,
      "grad_norm": 0.09880448877811432,
      "learning_rate": 0.00025878299240723055,
      "loss": 0.3723769783973694,
      "memory(GiB)": 78.26,
      "step": 1444,
      "token_acc": 0.8905043044032355,
      "train_speed(iter/s)": 0.033011
    },
    {
      "epoch": 0.27999806229714674,
      "grad_norm": 0.10194198787212372,
      "learning_rate": 0.0002587167815532468,
      "loss": 0.3568089008331299,
      "memory(GiB)": 78.26,
      "step": 1445,
      "token_acc": 0.8941702819956616,
      "train_speed(iter/s)": 0.033012
    },
    {
      "epoch": 0.2801918325824735,
      "grad_norm": 0.10794655978679657,
      "learning_rate": 0.00025865052604635955,
      "loss": 0.3879980146884918,
      "memory(GiB)": 78.26,
      "step": 1446,
      "token_acc": 0.8870534199744252,
      "train_speed(iter/s)": 0.033014
    },
    {
      "epoch": 0.28038560286780023,
      "grad_norm": 0.10190257430076599,
      "learning_rate": 0.0002585842259137817,
      "loss": 0.385148823261261,
      "memory(GiB)": 78.26,
      "step": 1447,
      "token_acc": 0.8866731047802994,
      "train_speed(iter/s)": 0.033015
    },
    {
      "epoch": 0.280579373153127,
      "grad_norm": 0.11218508332967758,
      "learning_rate": 0.0002585178811827445,
      "loss": 0.41854822635650635,
      "memory(GiB)": 78.26,
      "step": 1448,
      "token_acc": 0.8806162104733414,
      "train_speed(iter/s)": 0.033017
    },
    {
      "epoch": 0.2807731434384537,
      "grad_norm": 0.10860633105039597,
      "learning_rate": 0.00025845149188049747,
      "loss": 0.3808492124080658,
      "memory(GiB)": 78.26,
      "step": 1449,
      "token_acc": 0.8890830143110403,
      "train_speed(iter/s)": 0.033018
    },
    {
      "epoch": 0.28096691372378046,
      "grad_norm": 0.10633924603462219,
      "learning_rate": 0.0002583850580343086,
      "loss": 0.3995111584663391,
      "memory(GiB)": 78.26,
      "step": 1450,
      "token_acc": 0.8848618846379894,
      "train_speed(iter/s)": 0.03302
    },
    {
      "epoch": 0.2811606840091072,
      "grad_norm": 0.1051037460565567,
      "learning_rate": 0.00025831857967146394,
      "loss": 0.3739997148513794,
      "memory(GiB)": 78.26,
      "step": 1451,
      "token_acc": 0.8915456874466268,
      "train_speed(iter/s)": 0.033021
    },
    {
      "epoch": 0.28135445429443395,
      "grad_norm": 0.1118805930018425,
      "learning_rate": 0.0002582520568192679,
      "loss": 0.4328954517841339,
      "memory(GiB)": 78.26,
      "step": 1452,
      "token_acc": 0.874195172358973,
      "train_speed(iter/s)": 0.033023
    },
    {
      "epoch": 0.2815482245797607,
      "grad_norm": 0.10707706958055496,
      "learning_rate": 0.0002581854895050434,
      "loss": 0.4252171814441681,
      "memory(GiB)": 78.26,
      "step": 1453,
      "token_acc": 0.8767298393819819,
      "train_speed(iter/s)": 0.033024
    },
    {
      "epoch": 0.28174199486508744,
      "grad_norm": 0.1086982861161232,
      "learning_rate": 0.0002581188777561313,
      "loss": 0.41112756729125977,
      "memory(GiB)": 78.26,
      "step": 1454,
      "token_acc": 0.8803484635202252,
      "train_speed(iter/s)": 0.033026
    },
    {
      "epoch": 0.2819357651504142,
      "grad_norm": 0.10656489431858063,
      "learning_rate": 0.00025805222159989077,
      "loss": 0.3699858486652374,
      "memory(GiB)": 78.26,
      "step": 1455,
      "token_acc": 0.8922368486911256,
      "train_speed(iter/s)": 0.033027
    },
    {
      "epoch": 0.28212953543574093,
      "grad_norm": 0.10233739018440247,
      "learning_rate": 0.00025798552106369937,
      "loss": 0.3772289752960205,
      "memory(GiB)": 78.26,
      "step": 1456,
      "token_acc": 0.8884009801305224,
      "train_speed(iter/s)": 0.033028
    },
    {
      "epoch": 0.2823233057210677,
      "grad_norm": 0.10542133450508118,
      "learning_rate": 0.00025791877617495275,
      "loss": 0.37354397773742676,
      "memory(GiB)": 78.26,
      "step": 1457,
      "token_acc": 0.8912464826222526,
      "train_speed(iter/s)": 0.03303
    },
    {
      "epoch": 0.2825170760063944,
      "grad_norm": 0.11344679445028305,
      "learning_rate": 0.0002578519869610649,
      "loss": 0.4273817241191864,
      "memory(GiB)": 78.26,
      "step": 1458,
      "token_acc": 0.8761668213330422,
      "train_speed(iter/s)": 0.033031
    },
    {
      "epoch": 0.28271084629172116,
      "grad_norm": 0.10357240587472916,
      "learning_rate": 0.000257785153449468,
      "loss": 0.3900695741176605,
      "memory(GiB)": 78.26,
      "step": 1459,
      "token_acc": 0.884189494968947,
      "train_speed(iter/s)": 0.033032
    },
    {
      "epoch": 0.2829046165770479,
      "grad_norm": 0.11411286890506744,
      "learning_rate": 0.00025771827566761215,
      "loss": 0.4180591106414795,
      "memory(GiB)": 78.26,
      "step": 1460,
      "token_acc": 0.8776129660386003,
      "train_speed(iter/s)": 0.033034
    },
    {
      "epoch": 0.28309838686237465,
      "grad_norm": 0.10684863477945328,
      "learning_rate": 0.00025765135364296606,
      "loss": 0.42027878761291504,
      "memory(GiB)": 78.26,
      "step": 1461,
      "token_acc": 0.8778430934841382,
      "train_speed(iter/s)": 0.033035
    },
    {
      "epoch": 0.2832921571477014,
      "grad_norm": 0.10738670825958252,
      "learning_rate": 0.0002575843874030163,
      "loss": 0.36749184131622314,
      "memory(GiB)": 78.26,
      "step": 1462,
      "token_acc": 0.889355581127733,
      "train_speed(iter/s)": 0.033037
    },
    {
      "epoch": 0.28348592743302814,
      "grad_norm": 0.10112845152616501,
      "learning_rate": 0.0002575173769752677,
      "loss": 0.3824685513973236,
      "memory(GiB)": 78.26,
      "step": 1463,
      "token_acc": 0.8880091942280679,
      "train_speed(iter/s)": 0.033038
    },
    {
      "epoch": 0.2836796977183549,
      "grad_norm": 0.11604847013950348,
      "learning_rate": 0.00025745032238724325,
      "loss": 0.4152447283267975,
      "memory(GiB)": 78.26,
      "step": 1464,
      "token_acc": 0.8827828574982894,
      "train_speed(iter/s)": 0.033039
    },
    {
      "epoch": 0.28387346800368163,
      "grad_norm": 0.10715403407812119,
      "learning_rate": 0.0002573832236664842,
      "loss": 0.3738415837287903,
      "memory(GiB)": 78.26,
      "step": 1465,
      "token_acc": 0.8922089917371876,
      "train_speed(iter/s)": 0.033041
    },
    {
      "epoch": 0.2840672382890084,
      "grad_norm": 0.10670117288827896,
      "learning_rate": 0.0002573160808405496,
      "loss": 0.40999048948287964,
      "memory(GiB)": 78.26,
      "step": 1466,
      "token_acc": 0.8812486533074768,
      "train_speed(iter/s)": 0.033042
    },
    {
      "epoch": 0.2842610085743351,
      "grad_norm": 0.11364596337080002,
      "learning_rate": 0.00025724889393701687,
      "loss": 0.38645139336586,
      "memory(GiB)": 78.26,
      "step": 1467,
      "token_acc": 0.8858692377222055,
      "train_speed(iter/s)": 0.033044
    },
    {
      "epoch": 0.28445477885966186,
      "grad_norm": 0.10818950086832047,
      "learning_rate": 0.00025718166298348163,
      "loss": 0.43197697401046753,
      "memory(GiB)": 78.26,
      "step": 1468,
      "token_acc": 0.8740517566040781,
      "train_speed(iter/s)": 0.033045
    },
    {
      "epoch": 0.2846485491449886,
      "grad_norm": 0.10857094824314117,
      "learning_rate": 0.00025711438800755725,
      "loss": 0.3983537256717682,
      "memory(GiB)": 78.26,
      "step": 1469,
      "token_acc": 0.8840075154730327,
      "train_speed(iter/s)": 0.033047
    },
    {
      "epoch": 0.28484231943031535,
      "grad_norm": 0.10339067131280899,
      "learning_rate": 0.00025704706903687544,
      "loss": 0.3655926585197449,
      "memory(GiB)": 78.26,
      "step": 1470,
      "token_acc": 0.8920360215324245,
      "train_speed(iter/s)": 0.033048
    },
    {
      "epoch": 0.2850360897156421,
      "grad_norm": 0.09809229522943497,
      "learning_rate": 0.0002569797060990859,
      "loss": 0.36431875824928284,
      "memory(GiB)": 78.26,
      "step": 1471,
      "token_acc": 0.8942954390742002,
      "train_speed(iter/s)": 0.033049
    },
    {
      "epoch": 0.28522986000096884,
      "grad_norm": 0.12293403595685959,
      "learning_rate": 0.0002569122992218564,
      "loss": 0.4425593316555023,
      "memory(GiB)": 78.26,
      "step": 1472,
      "token_acc": 0.872041270483512,
      "train_speed(iter/s)": 0.033051
    },
    {
      "epoch": 0.2854236302862956,
      "grad_norm": 0.11606067419052124,
      "learning_rate": 0.00025684484843287284,
      "loss": 0.4095402956008911,
      "memory(GiB)": 78.26,
      "step": 1473,
      "token_acc": 0.8805144353865623,
      "train_speed(iter/s)": 0.033052
    },
    {
      "epoch": 0.2856174005716223,
      "grad_norm": 0.09995657950639725,
      "learning_rate": 0.00025677735375983894,
      "loss": 0.38141509890556335,
      "memory(GiB)": 78.26,
      "step": 1474,
      "token_acc": 0.886579869804707,
      "train_speed(iter/s)": 0.033053
    },
    {
      "epoch": 0.28581117085694907,
      "grad_norm": 0.09991477429866791,
      "learning_rate": 0.00025670981523047664,
      "loss": 0.3756733238697052,
      "memory(GiB)": 78.26,
      "step": 1475,
      "token_acc": 0.8895817295355402,
      "train_speed(iter/s)": 0.033055
    },
    {
      "epoch": 0.2860049411422758,
      "grad_norm": 0.10868589580059052,
      "learning_rate": 0.00025664223287252586,
      "loss": 0.40122172236442566,
      "memory(GiB)": 78.26,
      "step": 1476,
      "token_acc": 0.8827901370963692,
      "train_speed(iter/s)": 0.033056
    },
    {
      "epoch": 0.28619871142760256,
      "grad_norm": 0.10444658994674683,
      "learning_rate": 0.0002565746067137444,
      "loss": 0.37232351303100586,
      "memory(GiB)": 78.26,
      "step": 1477,
      "token_acc": 0.8921252470612712,
      "train_speed(iter/s)": 0.033058
    },
    {
      "epoch": 0.2863924817129293,
      "grad_norm": 0.10679329186677933,
      "learning_rate": 0.0002565069367819082,
      "loss": 0.36880356073379517,
      "memory(GiB)": 78.26,
      "step": 1478,
      "token_acc": 0.8917536644521776,
      "train_speed(iter/s)": 0.033059
    },
    {
      "epoch": 0.28658625199825605,
      "grad_norm": 0.09651319682598114,
      "learning_rate": 0.0002564392231048111,
      "loss": 0.36866456270217896,
      "memory(GiB)": 78.26,
      "step": 1479,
      "token_acc": 0.8893619056730125,
      "train_speed(iter/s)": 0.033061
    },
    {
      "epoch": 0.2867800222835828,
      "grad_norm": 0.10870502889156342,
      "learning_rate": 0.000256371465710265,
      "loss": 0.3834042549133301,
      "memory(GiB)": 78.26,
      "step": 1480,
      "token_acc": 0.8883792468895053,
      "train_speed(iter/s)": 0.033062
    },
    {
      "epoch": 0.28697379256890954,
      "grad_norm": 0.11162568628787994,
      "learning_rate": 0.0002563036646260996,
      "loss": 0.4060906171798706,
      "memory(GiB)": 78.26,
      "step": 1481,
      "token_acc": 0.8793504766127245,
      "train_speed(iter/s)": 0.033064
    },
    {
      "epoch": 0.2871675628542363,
      "grad_norm": 0.10246509313583374,
      "learning_rate": 0.00025623581988016257,
      "loss": 0.37621062994003296,
      "memory(GiB)": 78.26,
      "step": 1482,
      "token_acc": 0.8903611537529288,
      "train_speed(iter/s)": 0.033065
    },
    {
      "epoch": 0.287361333139563,
      "grad_norm": 0.10581538081169128,
      "learning_rate": 0.0002561679315003197,
      "loss": 0.40835872292518616,
      "memory(GiB)": 78.26,
      "step": 1483,
      "token_acc": 0.8827440663756468,
      "train_speed(iter/s)": 0.033066
    },
    {
      "epoch": 0.28755510342488977,
      "grad_norm": 0.11528300493955612,
      "learning_rate": 0.0002560999995144545,
      "loss": 0.43208250403404236,
      "memory(GiB)": 78.26,
      "step": 1484,
      "token_acc": 0.8761382138334439,
      "train_speed(iter/s)": 0.033068
    },
    {
      "epoch": 0.2877488737102165,
      "grad_norm": 0.11191680282354355,
      "learning_rate": 0.00025603202395046857,
      "loss": 0.40369465947151184,
      "memory(GiB)": 78.26,
      "step": 1485,
      "token_acc": 0.8817523721627779,
      "train_speed(iter/s)": 0.033069
    },
    {
      "epoch": 0.28794264399554326,
      "grad_norm": 0.0994502380490303,
      "learning_rate": 0.00025596400483628113,
      "loss": 0.3518386781215668,
      "memory(GiB)": 78.26,
      "step": 1486,
      "token_acc": 0.8963413044045166,
      "train_speed(iter/s)": 0.033071
    },
    {
      "epoch": 0.28813641428087,
      "grad_norm": 0.10107074677944183,
      "learning_rate": 0.00025589594219982957,
      "loss": 0.36090126633644104,
      "memory(GiB)": 78.26,
      "step": 1487,
      "token_acc": 0.8948038903451824,
      "train_speed(iter/s)": 0.033072
    },
    {
      "epoch": 0.2883301845661968,
      "grad_norm": 0.10665128380060196,
      "learning_rate": 0.000255827836069069,
      "loss": 0.381521075963974,
      "memory(GiB)": 78.26,
      "step": 1488,
      "token_acc": 0.8875178437492439,
      "train_speed(iter/s)": 0.033073
    },
    {
      "epoch": 0.28852395485152355,
      "grad_norm": 0.11443497240543365,
      "learning_rate": 0.00025575968647197246,
      "loss": 0.4305599629878998,
      "memory(GiB)": 78.26,
      "step": 1489,
      "token_acc": 0.8745530313214142,
      "train_speed(iter/s)": 0.033075
    },
    {
      "epoch": 0.2887177251368503,
      "grad_norm": 0.10215198248624802,
      "learning_rate": 0.0002556914934365308,
      "loss": 0.38611581921577454,
      "memory(GiB)": 78.26,
      "step": 1490,
      "token_acc": 0.8878939990051101,
      "train_speed(iter/s)": 0.033076
    },
    {
      "epoch": 0.28891149542217703,
      "grad_norm": 0.11031733453273773,
      "learning_rate": 0.00025562325699075275,
      "loss": 0.3934246003627777,
      "memory(GiB)": 78.26,
      "step": 1491,
      "token_acc": 0.8847335423197492,
      "train_speed(iter/s)": 0.033078
    },
    {
      "epoch": 0.2891052657075038,
      "grad_norm": 0.107778400182724,
      "learning_rate": 0.00025555497716266487,
      "loss": 0.4159546196460724,
      "memory(GiB)": 78.26,
      "step": 1492,
      "token_acc": 0.8804967649961618,
      "train_speed(iter/s)": 0.033079
    },
    {
      "epoch": 0.2892990359928305,
      "grad_norm": 0.11043369024991989,
      "learning_rate": 0.00025548665398031145,
      "loss": 0.42264997959136963,
      "memory(GiB)": 78.26,
      "step": 1493,
      "token_acc": 0.8791478474270945,
      "train_speed(iter/s)": 0.033081
    },
    {
      "epoch": 0.28949280627815727,
      "grad_norm": 0.11627716571092606,
      "learning_rate": 0.0002554182874717547,
      "loss": 0.40543925762176514,
      "memory(GiB)": 78.26,
      "step": 1494,
      "token_acc": 0.8830718414533444,
      "train_speed(iter/s)": 0.033082
    },
    {
      "epoch": 0.289686576563484,
      "grad_norm": 0.11875671148300171,
      "learning_rate": 0.00025534987766507466,
      "loss": 0.42699331045150757,
      "memory(GiB)": 78.26,
      "step": 1495,
      "token_acc": 0.8727887840821125,
      "train_speed(iter/s)": 0.033084
    },
    {
      "epoch": 0.28988034684881075,
      "grad_norm": 0.1101786196231842,
      "learning_rate": 0.00025528142458836896,
      "loss": 0.36886993050575256,
      "memory(GiB)": 78.26,
      "step": 1496,
      "token_acc": 0.8915269892151803,
      "train_speed(iter/s)": 0.033085
    },
    {
      "epoch": 0.2900741171341375,
      "grad_norm": 0.1435171663761139,
      "learning_rate": 0.0002552129282697532,
      "loss": 0.38396596908569336,
      "memory(GiB)": 78.26,
      "step": 1497,
      "token_acc": 0.8871555969652603,
      "train_speed(iter/s)": 0.033087
    },
    {
      "epoch": 0.29026788741946424,
      "grad_norm": 0.10523149371147156,
      "learning_rate": 0.0002551443887373605,
      "loss": 0.36478835344314575,
      "memory(GiB)": 78.26,
      "step": 1498,
      "token_acc": 0.8914113351325608,
      "train_speed(iter/s)": 0.033088
    },
    {
      "epoch": 0.290461657704791,
      "grad_norm": 0.1243639886379242,
      "learning_rate": 0.00025507580601934215,
      "loss": 0.4351167678833008,
      "memory(GiB)": 78.26,
      "step": 1499,
      "token_acc": 0.874244422464007,
      "train_speed(iter/s)": 0.033089
    },
    {
      "epoch": 0.29065542799011773,
      "grad_norm": 0.12400247901678085,
      "learning_rate": 0.0002550071801438667,
      "loss": 0.4268084466457367,
      "memory(GiB)": 78.26,
      "step": 1500,
      "token_acc": 0.8754112107334064,
      "train_speed(iter/s)": 0.033091
    },
    {
      "epoch": 0.29065542799011773,
      "eval_loss": 0.46029698848724365,
      "eval_runtime": 1344.1819,
      "eval_samples_per_second": 5.021,
      "eval_steps_per_second": 5.021,
      "eval_token_acc": 0.8845504800698284,
      "step": 1500
    },
    {
      "epoch": 0.2908491982754445,
      "grad_norm": 0.11321935057640076,
      "learning_rate": 0.0002549385111391207,
      "loss": 0.4121960699558258,
      "memory(GiB)": 78.26,
      "step": 1501,
      "token_acc": 0.881211333653493,
      "train_speed(iter/s)": 0.03214
    },
    {
      "epoch": 0.2910429685607712,
      "grad_norm": 0.11633153259754181,
      "learning_rate": 0.0002548697990333084,
      "loss": 0.4356468617916107,
      "memory(GiB)": 78.26,
      "step": 1502,
      "token_acc": 0.8736777331731473,
      "train_speed(iter/s)": 0.032142
    },
    {
      "epoch": 0.29123673884609796,
      "grad_norm": 0.10829450935125351,
      "learning_rate": 0.00025480104385465166,
      "loss": 0.4199885129928589,
      "memory(GiB)": 78.26,
      "step": 1503,
      "token_acc": 0.8805436914067747,
      "train_speed(iter/s)": 0.032144
    },
    {
      "epoch": 0.2914305091314247,
      "grad_norm": 0.11650849878787994,
      "learning_rate": 0.0002547322456313901,
      "loss": 0.4100668430328369,
      "memory(GiB)": 78.26,
      "step": 1504,
      "token_acc": 0.8807525592548188,
      "train_speed(iter/s)": 0.032146
    },
    {
      "epoch": 0.29162427941675145,
      "grad_norm": 0.10901083797216415,
      "learning_rate": 0.000254663404391781,
      "loss": 0.4044414460659027,
      "memory(GiB)": 78.26,
      "step": 1505,
      "token_acc": 0.883741382734262,
      "train_speed(iter/s)": 0.032148
    },
    {
      "epoch": 0.2918180497020782,
      "grad_norm": 0.10262013971805573,
      "learning_rate": 0.00025459452016409926,
      "loss": 0.40257662534713745,
      "memory(GiB)": 78.26,
      "step": 1506,
      "token_acc": 0.8821285962936418,
      "train_speed(iter/s)": 0.03215
    },
    {
      "epoch": 0.29201181998740494,
      "grad_norm": 0.10731082409620285,
      "learning_rate": 0.0002545255929766376,
      "loss": 0.4021913409233093,
      "memory(GiB)": 78.26,
      "step": 1507,
      "token_acc": 0.8832824363320539,
      "train_speed(iter/s)": 0.032152
    },
    {
      "epoch": 0.2922055902727317,
      "grad_norm": 0.11936552822589874,
      "learning_rate": 0.00025445662285770613,
      "loss": 0.4097455143928528,
      "memory(GiB)": 78.26,
      "step": 1508,
      "token_acc": 0.8790452832446013,
      "train_speed(iter/s)": 0.032154
    },
    {
      "epoch": 0.29239936055805843,
      "grad_norm": 0.11192671209573746,
      "learning_rate": 0.00025438760983563285,
      "loss": 0.40953508019447327,
      "memory(GiB)": 78.26,
      "step": 1509,
      "token_acc": 0.8821702104868513,
      "train_speed(iter/s)": 0.032156
    },
    {
      "epoch": 0.2925931308433852,
      "grad_norm": 0.10715434700250626,
      "learning_rate": 0.0002543185539387632,
      "loss": 0.3857203722000122,
      "memory(GiB)": 78.26,
      "step": 1510,
      "token_acc": 0.8867008985879332,
      "train_speed(iter/s)": 0.032158
    },
    {
      "epoch": 0.2927869011287119,
      "grad_norm": 0.1083342507481575,
      "learning_rate": 0.0002542494551954602,
      "loss": 0.3879879117012024,
      "memory(GiB)": 78.26,
      "step": 1511,
      "token_acc": 0.886165023879687,
      "train_speed(iter/s)": 0.032159
    },
    {
      "epoch": 0.29298067141403866,
      "grad_norm": 0.12331650406122208,
      "learning_rate": 0.0002541803136341048,
      "loss": 0.4193812608718872,
      "memory(GiB)": 78.26,
      "step": 1512,
      "token_acc": 0.8781190284970506,
      "train_speed(iter/s)": 0.032161
    },
    {
      "epoch": 0.2931744416993654,
      "grad_norm": 0.10431553423404694,
      "learning_rate": 0.0002541111292830951,
      "loss": 0.3808317184448242,
      "memory(GiB)": 78.26,
      "step": 1513,
      "token_acc": 0.888267724649629,
      "train_speed(iter/s)": 0.032163
    },
    {
      "epoch": 0.29336821198469215,
      "grad_norm": 0.09962674230337143,
      "learning_rate": 0.00025404190217084697,
      "loss": 0.3735466003417969,
      "memory(GiB)": 78.26,
      "step": 1514,
      "token_acc": 0.8892632578004508,
      "train_speed(iter/s)": 0.032165
    },
    {
      "epoch": 0.2935619822700189,
      "grad_norm": 0.1109800636768341,
      "learning_rate": 0.000253972632325794,
      "loss": 0.42662137746810913,
      "memory(GiB)": 78.26,
      "step": 1515,
      "token_acc": 0.8756379486330658,
      "train_speed(iter/s)": 0.032167
    },
    {
      "epoch": 0.29375575255534564,
      "grad_norm": 0.10075878351926804,
      "learning_rate": 0.000253903319776387,
      "loss": 0.3623042702674866,
      "memory(GiB)": 78.26,
      "step": 1516,
      "token_acc": 0.8929071827001877,
      "train_speed(iter/s)": 0.032169
    },
    {
      "epoch": 0.2939495228406724,
      "grad_norm": 0.10288105905056,
      "learning_rate": 0.0002538339645510946,
      "loss": 0.38466522097587585,
      "memory(GiB)": 78.26,
      "step": 1517,
      "token_acc": 0.8891691358327007,
      "train_speed(iter/s)": 0.032171
    },
    {
      "epoch": 0.29414329312599913,
      "grad_norm": 0.11288794130086899,
      "learning_rate": 0.00025376456667840284,
      "loss": 0.39476266503334045,
      "memory(GiB)": 78.26,
      "step": 1518,
      "token_acc": 0.8875907973596265,
      "train_speed(iter/s)": 0.032173
    },
    {
      "epoch": 0.29433706341132587,
      "grad_norm": 0.12219920009374619,
      "learning_rate": 0.0002536951261868153,
      "loss": 0.40511706471443176,
      "memory(GiB)": 78.26,
      "step": 1519,
      "token_acc": 0.8838210765731614,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.2945308336966526,
      "grad_norm": 0.1097761020064354,
      "learning_rate": 0.000253625643104853,
      "loss": 0.4082501530647278,
      "memory(GiB)": 78.26,
      "step": 1520,
      "token_acc": 0.8827916295636687,
      "train_speed(iter/s)": 0.032177
    },
    {
      "epoch": 0.29472460398197936,
      "grad_norm": 0.21266750991344452,
      "learning_rate": 0.0002535561174610546,
      "loss": 0.4297519028186798,
      "memory(GiB)": 78.26,
      "step": 1521,
      "token_acc": 0.8791513393088441,
      "train_speed(iter/s)": 0.032179
    },
    {
      "epoch": 0.2949183742673061,
      "grad_norm": 0.10368236899375916,
      "learning_rate": 0.00025348654928397614,
      "loss": 0.3687269389629364,
      "memory(GiB)": 78.26,
      "step": 1522,
      "token_acc": 0.8922836811893236,
      "train_speed(iter/s)": 0.03218
    },
    {
      "epoch": 0.29511214455263285,
      "grad_norm": 0.11547058075666428,
      "learning_rate": 0.000253416938602191,
      "loss": 0.38897940516471863,
      "memory(GiB)": 78.26,
      "step": 1523,
      "token_acc": 0.8873057637889094,
      "train_speed(iter/s)": 0.032182
    },
    {
      "epoch": 0.2953059148379596,
      "grad_norm": 0.40387094020843506,
      "learning_rate": 0.0002533472854442903,
      "loss": 0.4147917926311493,
      "memory(GiB)": 78.26,
      "step": 1524,
      "token_acc": 0.8777598930968507,
      "train_speed(iter/s)": 0.032184
    },
    {
      "epoch": 0.29549968512328634,
      "grad_norm": 0.10571064054965973,
      "learning_rate": 0.0002532775898388824,
      "loss": 0.38508111238479614,
      "memory(GiB)": 78.26,
      "step": 1525,
      "token_acc": 0.8872821274090757,
      "train_speed(iter/s)": 0.032186
    },
    {
      "epoch": 0.2956934554086131,
      "grad_norm": 0.1151953935623169,
      "learning_rate": 0.0002532078518145931,
      "loss": 0.4073837697505951,
      "memory(GiB)": 78.26,
      "step": 1526,
      "token_acc": 0.8840981728275945,
      "train_speed(iter/s)": 0.032188
    },
    {
      "epoch": 0.2958872256939398,
      "grad_norm": 0.11016645282506943,
      "learning_rate": 0.0002531380714000659,
      "loss": 0.39761802554130554,
      "memory(GiB)": 78.26,
      "step": 1527,
      "token_acc": 0.8864857603439011,
      "train_speed(iter/s)": 0.03219
    },
    {
      "epoch": 0.29608099597926657,
      "grad_norm": 0.10478914529085159,
      "learning_rate": 0.00025306824862396127,
      "loss": 0.3668254613876343,
      "memory(GiB)": 78.26,
      "step": 1528,
      "token_acc": 0.8923076923076924,
      "train_speed(iter/s)": 0.032192
    },
    {
      "epoch": 0.2962747662645933,
      "grad_norm": 0.3135347068309784,
      "learning_rate": 0.0002529983835149574,
      "loss": 0.3972938656806946,
      "memory(GiB)": 78.26,
      "step": 1529,
      "token_acc": 0.8850560993077107,
      "train_speed(iter/s)": 0.032193
    },
    {
      "epoch": 0.29646853654992006,
      "grad_norm": 0.11302391439676285,
      "learning_rate": 0.00025292847610174974,
      "loss": 0.4002307057380676,
      "memory(GiB)": 78.26,
      "step": 1530,
      "token_acc": 0.8843288710519805,
      "train_speed(iter/s)": 0.032195
    },
    {
      "epoch": 0.2966623068352468,
      "grad_norm": 0.1242542490363121,
      "learning_rate": 0.0002528585264130511,
      "loss": 0.40723949670791626,
      "memory(GiB)": 78.26,
      "step": 1531,
      "token_acc": 0.8828371991110276,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.29685607712057355,
      "grad_norm": 0.11407161504030228,
      "learning_rate": 0.00025278853447759184,
      "loss": 0.38927173614501953,
      "memory(GiB)": 78.26,
      "step": 1532,
      "token_acc": 0.888841623906628,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.2970498474059003,
      "grad_norm": 0.10127895325422287,
      "learning_rate": 0.0002527185003241194,
      "loss": 0.38897159695625305,
      "memory(GiB)": 78.26,
      "step": 1533,
      "token_acc": 0.884356180097505,
      "train_speed(iter/s)": 0.032201
    },
    {
      "epoch": 0.29724361769122704,
      "grad_norm": 0.11514552682638168,
      "learning_rate": 0.0002526484239813987,
      "loss": 0.41849443316459656,
      "memory(GiB)": 78.26,
      "step": 1534,
      "token_acc": 0.8769607696674535,
      "train_speed(iter/s)": 0.032203
    },
    {
      "epoch": 0.2974373879765538,
      "grad_norm": 0.10013052076101303,
      "learning_rate": 0.00025257830547821205,
      "loss": 0.35412314534187317,
      "memory(GiB)": 78.26,
      "step": 1535,
      "token_acc": 0.8960811384876806,
      "train_speed(iter/s)": 0.032205
    },
    {
      "epoch": 0.2976311582618805,
      "grad_norm": 0.107215017080307,
      "learning_rate": 0.0002525081448433589,
      "loss": 0.4127185642719269,
      "memory(GiB)": 78.26,
      "step": 1536,
      "token_acc": 0.8802447552447552,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.29782492854720727,
      "grad_norm": 0.10350769758224487,
      "learning_rate": 0.00025243794210565623,
      "loss": 0.35154351592063904,
      "memory(GiB)": 78.26,
      "step": 1537,
      "token_acc": 0.8971347925653725,
      "train_speed(iter/s)": 0.032208
    },
    {
      "epoch": 0.298018698832534,
      "grad_norm": 0.11700846254825592,
      "learning_rate": 0.00025236769729393806,
      "loss": 0.42805176973342896,
      "memory(GiB)": 78.26,
      "step": 1538,
      "token_acc": 0.8778269617706237,
      "train_speed(iter/s)": 0.03221
    },
    {
      "epoch": 0.29821246911786076,
      "grad_norm": 0.11825554072856903,
      "learning_rate": 0.0002522974104370559,
      "loss": 0.42821887135505676,
      "memory(GiB)": 78.26,
      "step": 1539,
      "token_acc": 0.8769199262748311,
      "train_speed(iter/s)": 0.032211
    },
    {
      "epoch": 0.2984062394031875,
      "grad_norm": 0.11428213119506836,
      "learning_rate": 0.0002522270815638784,
      "loss": 0.4251292645931244,
      "memory(GiB)": 78.26,
      "step": 1540,
      "token_acc": 0.8747427227286092,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.29860000968851425,
      "grad_norm": 0.11575201153755188,
      "learning_rate": 0.00025215671070329164,
      "loss": 0.4403133988380432,
      "memory(GiB)": 78.26,
      "step": 1541,
      "token_acc": 0.872397366589948,
      "train_speed(iter/s)": 0.032215
    },
    {
      "epoch": 0.298793779973841,
      "grad_norm": 0.10235166549682617,
      "learning_rate": 0.0002520862978841987,
      "loss": 0.3714492619037628,
      "memory(GiB)": 78.26,
      "step": 1542,
      "token_acc": 0.8913319238900634,
      "train_speed(iter/s)": 0.032217
    },
    {
      "epoch": 0.29898755025916773,
      "grad_norm": 0.11813291907310486,
      "learning_rate": 0.00025201584313552,
      "loss": 0.4298678934574127,
      "memory(GiB)": 78.26,
      "step": 1543,
      "token_acc": 0.875,
      "train_speed(iter/s)": 0.032219
    },
    {
      "epoch": 0.2991813205444945,
      "grad_norm": 0.1107359528541565,
      "learning_rate": 0.0002519453464861933,
      "loss": 0.3796873390674591,
      "memory(GiB)": 78.26,
      "step": 1544,
      "token_acc": 0.8897960202731381,
      "train_speed(iter/s)": 0.03222
    },
    {
      "epoch": 0.2993750908298212,
      "grad_norm": 0.12202285975217819,
      "learning_rate": 0.0002518748079651734,
      "loss": 0.3539576232433319,
      "memory(GiB)": 78.26,
      "step": 1545,
      "token_acc": 0.8958181376124483,
      "train_speed(iter/s)": 0.032222
    },
    {
      "epoch": 0.29956886111514797,
      "grad_norm": 0.10455404222011566,
      "learning_rate": 0.00025180422760143244,
      "loss": 0.3648805618286133,
      "memory(GiB)": 78.26,
      "step": 1546,
      "token_acc": 0.8921280583557064,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.2997626314004747,
      "grad_norm": 0.10837128758430481,
      "learning_rate": 0.0002517336054239596,
      "loss": 0.4049074351787567,
      "memory(GiB)": 78.26,
      "step": 1547,
      "token_acc": 0.8826463706255946,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.29995640168580145,
      "grad_norm": 0.11739413440227509,
      "learning_rate": 0.00025166294146176124,
      "loss": 0.38583841919898987,
      "memory(GiB)": 78.26,
      "step": 1548,
      "token_acc": 0.8861551627309925,
      "train_speed(iter/s)": 0.032227
    },
    {
      "epoch": 0.30015017197112825,
      "grad_norm": 0.10927794128656387,
      "learning_rate": 0.00025159223574386114,
      "loss": 0.40299391746520996,
      "memory(GiB)": 78.26,
      "step": 1549,
      "token_acc": 0.8808586662200263,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.300343942256455,
      "grad_norm": 0.10551624745130539,
      "learning_rate": 0.0002515214882992999,
      "loss": 0.3709675073623657,
      "memory(GiB)": 78.26,
      "step": 1550,
      "token_acc": 0.8917359439947495,
      "train_speed(iter/s)": 0.032231
    },
    {
      "epoch": 0.30053771254178174,
      "grad_norm": 0.11021958291530609,
      "learning_rate": 0.00025145069915713536,
      "loss": 0.3905988335609436,
      "memory(GiB)": 78.26,
      "step": 1551,
      "token_acc": 0.8831299218774511,
      "train_speed(iter/s)": 0.032233
    },
    {
      "epoch": 0.3007314828271085,
      "grad_norm": 0.09675493836402893,
      "learning_rate": 0.0002513798683464427,
      "loss": 0.3553582429885864,
      "memory(GiB)": 78.26,
      "step": 1552,
      "token_acc": 0.895078622611465,
      "train_speed(iter/s)": 0.032234
    },
    {
      "epoch": 0.30092525311243523,
      "grad_norm": 0.11715512722730637,
      "learning_rate": 0.0002513089958963139,
      "loss": 0.38959836959838867,
      "memory(GiB)": 78.26,
      "step": 1553,
      "token_acc": 0.8870248193528119,
      "train_speed(iter/s)": 0.032236
    },
    {
      "epoch": 0.301119023397762,
      "grad_norm": 0.10786168277263641,
      "learning_rate": 0.00025123808183585817,
      "loss": 0.3897908329963684,
      "memory(GiB)": 78.26,
      "step": 1554,
      "token_acc": 0.8833446083074009,
      "train_speed(iter/s)": 0.032238
    },
    {
      "epoch": 0.3013127936830887,
      "grad_norm": 0.11172870546579361,
      "learning_rate": 0.00025116712619420185,
      "loss": 0.37313905358314514,
      "memory(GiB)": 78.26,
      "step": 1555,
      "token_acc": 0.8884811242923387,
      "train_speed(iter/s)": 0.03224
    },
    {
      "epoch": 0.30150656396841546,
      "grad_norm": 0.11459864675998688,
      "learning_rate": 0.0002510961290004884,
      "loss": 0.41675081849098206,
      "memory(GiB)": 78.26,
      "step": 1556,
      "token_acc": 0.8796063237082496,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.3017003342537422,
      "grad_norm": 0.12527529895305634,
      "learning_rate": 0.00025102509028387813,
      "loss": 0.3714950680732727,
      "memory(GiB)": 78.26,
      "step": 1557,
      "token_acc": 0.8910757252156046,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.30189410453906895,
      "grad_norm": 0.11035227030515671,
      "learning_rate": 0.00025095401007354867,
      "loss": 0.39393147826194763,
      "memory(GiB)": 78.26,
      "step": 1558,
      "token_acc": 0.8863604634929498,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.3020878748243957,
      "grad_norm": 0.10444723814725876,
      "learning_rate": 0.0002508828883986945,
      "loss": 0.3709162771701813,
      "memory(GiB)": 78.26,
      "step": 1559,
      "token_acc": 0.8904844563673663,
      "train_speed(iter/s)": 0.032248
    },
    {
      "epoch": 0.30228164510972244,
      "grad_norm": 0.10122967511415482,
      "learning_rate": 0.0002508117252885273,
      "loss": 0.34817183017730713,
      "memory(GiB)": 78.26,
      "step": 1560,
      "token_acc": 0.8977317721590587,
      "train_speed(iter/s)": 0.03225
    },
    {
      "epoch": 0.3024754153950492,
      "grad_norm": 0.10534863919019699,
      "learning_rate": 0.00025074052077227556,
      "loss": 0.4088999927043915,
      "memory(GiB)": 78.26,
      "step": 1561,
      "token_acc": 0.8811859443631039,
      "train_speed(iter/s)": 0.032251
    },
    {
      "epoch": 0.30266918568037593,
      "grad_norm": 0.0965069830417633,
      "learning_rate": 0.000250669274879185,
      "loss": 0.3669951558113098,
      "memory(GiB)": 78.26,
      "step": 1562,
      "token_acc": 0.8907134896627016,
      "train_speed(iter/s)": 0.032253
    },
    {
      "epoch": 0.3028629559657027,
      "grad_norm": 0.10454553365707397,
      "learning_rate": 0.0002505979876385181,
      "loss": 0.40625712275505066,
      "memory(GiB)": 78.26,
      "step": 1563,
      "token_acc": 0.8799307565696124,
      "train_speed(iter/s)": 0.032255
    },
    {
      "epoch": 0.3030567262510294,
      "grad_norm": 0.10287556052207947,
      "learning_rate": 0.0002505266590795545,
      "loss": 0.415050208568573,
      "memory(GiB)": 78.26,
      "step": 1564,
      "token_acc": 0.8815440689198144,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.30325049653635616,
      "grad_norm": 0.11195293813943863,
      "learning_rate": 0.00025045528923159073,
      "loss": 0.4277609586715698,
      "memory(GiB)": 78.26,
      "step": 1565,
      "token_acc": 0.8775790135451792,
      "train_speed(iter/s)": 0.032258
    },
    {
      "epoch": 0.3034442668216829,
      "grad_norm": 0.10438180714845657,
      "learning_rate": 0.0002503838781239404,
      "loss": 0.3552014231681824,
      "memory(GiB)": 78.26,
      "step": 1566,
      "token_acc": 0.896136847274513,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 0.30363803710700965,
      "grad_norm": 0.12397732585668564,
      "learning_rate": 0.0002503124257859339,
      "loss": 0.46164411306381226,
      "memory(GiB)": 78.26,
      "step": 1567,
      "token_acc": 0.8641721234798877,
      "train_speed(iter/s)": 0.032262
    },
    {
      "epoch": 0.3038318073923364,
      "grad_norm": 0.11469985544681549,
      "learning_rate": 0.0002502409322469186,
      "loss": 0.39702335000038147,
      "memory(GiB)": 78.26,
      "step": 1568,
      "token_acc": 0.883688332266813,
      "train_speed(iter/s)": 0.032264
    },
    {
      "epoch": 0.30402557767766314,
      "grad_norm": 0.11066870391368866,
      "learning_rate": 0.00025016939753625886,
      "loss": 0.40544670820236206,
      "memory(GiB)": 78.26,
      "step": 1569,
      "token_acc": 0.8816060016671298,
      "train_speed(iter/s)": 0.032266
    },
    {
      "epoch": 0.3042193479629899,
      "grad_norm": 0.09953310340642929,
      "learning_rate": 0.0002500978216833359,
      "loss": 0.39714691042900085,
      "memory(GiB)": 78.26,
      "step": 1570,
      "token_acc": 0.8836974458757173,
      "train_speed(iter/s)": 0.032267
    },
    {
      "epoch": 0.3044131182483166,
      "grad_norm": 0.11722380667924881,
      "learning_rate": 0.00025002620471754785,
      "loss": 0.41713088750839233,
      "memory(GiB)": 78.26,
      "step": 1571,
      "token_acc": 0.8797300061362242,
      "train_speed(iter/s)": 0.032269
    },
    {
      "epoch": 0.30460688853364337,
      "grad_norm": 0.11423590779304504,
      "learning_rate": 0.00024995454666830967,
      "loss": 0.40586379170417786,
      "memory(GiB)": 78.26,
      "step": 1572,
      "token_acc": 0.8844951044863364,
      "train_speed(iter/s)": 0.032271
    },
    {
      "epoch": 0.3048006588189701,
      "grad_norm": 0.11461616307497025,
      "learning_rate": 0.00024988284756505334,
      "loss": 0.3949568271636963,
      "memory(GiB)": 78.26,
      "step": 1573,
      "token_acc": 0.8844877454711524,
      "train_speed(iter/s)": 0.032273
    },
    {
      "epoch": 0.30499442910429686,
      "grad_norm": 0.09611064195632935,
      "learning_rate": 0.0002498111074372276,
      "loss": 0.3817123770713806,
      "memory(GiB)": 78.26,
      "step": 1574,
      "token_acc": 0.8891472188103176,
      "train_speed(iter/s)": 0.032274
    },
    {
      "epoch": 0.3051881993896236,
      "grad_norm": 0.12749864161014557,
      "learning_rate": 0.0002497393263142979,
      "loss": 0.43550729751586914,
      "memory(GiB)": 78.26,
      "step": 1575,
      "token_acc": 0.876248012718601,
      "train_speed(iter/s)": 0.032276
    },
    {
      "epoch": 0.30538196967495035,
      "grad_norm": 0.10742782801389694,
      "learning_rate": 0.00024966750422574684,
      "loss": 0.3822196125984192,
      "memory(GiB)": 78.26,
      "step": 1576,
      "token_acc": 0.8897888795986622,
      "train_speed(iter/s)": 0.032278
    },
    {
      "epoch": 0.3055757399602771,
      "grad_norm": 0.1094922348856926,
      "learning_rate": 0.0002495956412010736,
      "loss": 0.3977155089378357,
      "memory(GiB)": 78.26,
      "step": 1577,
      "token_acc": 0.8843427182499494,
      "train_speed(iter/s)": 0.03228
    },
    {
      "epoch": 0.30576951024560384,
      "grad_norm": 0.11330767720937729,
      "learning_rate": 0.0002495237372697943,
      "loss": 0.3944382071495056,
      "memory(GiB)": 78.26,
      "step": 1578,
      "token_acc": 0.8850218853421586,
      "train_speed(iter/s)": 0.032281
    },
    {
      "epoch": 0.3059632805309306,
      "grad_norm": 0.10812770575284958,
      "learning_rate": 0.0002494517924614418,
      "loss": 0.39382147789001465,
      "memory(GiB)": 78.26,
      "step": 1579,
      "token_acc": 0.8852722927933593,
      "train_speed(iter/s)": 0.032283
    },
    {
      "epoch": 0.3061570508162573,
      "grad_norm": 0.12108810991048813,
      "learning_rate": 0.00024937980680556576,
      "loss": 0.4625246226787567,
      "memory(GiB)": 78.26,
      "step": 1580,
      "token_acc": 0.8664573098687892,
      "train_speed(iter/s)": 0.032285
    },
    {
      "epoch": 0.30635082110158407,
      "grad_norm": 0.11010700464248657,
      "learning_rate": 0.00024930778033173265,
      "loss": 0.40388697385787964,
      "memory(GiB)": 78.26,
      "step": 1581,
      "token_acc": 0.8823420361501623,
      "train_speed(iter/s)": 0.032287
    },
    {
      "epoch": 0.3065445913869108,
      "grad_norm": 0.1091642677783966,
      "learning_rate": 0.0002492357130695256,
      "loss": 0.4263598322868347,
      "memory(GiB)": 78.26,
      "step": 1582,
      "token_acc": 0.8781599433828733,
      "train_speed(iter/s)": 0.032289
    },
    {
      "epoch": 0.30673836167223756,
      "grad_norm": 0.10464166849851608,
      "learning_rate": 0.0002491636050485447,
      "loss": 0.37317129969596863,
      "memory(GiB)": 78.26,
      "step": 1583,
      "token_acc": 0.890936937421071,
      "train_speed(iter/s)": 0.03229
    },
    {
      "epoch": 0.3069321319575643,
      "grad_norm": 0.10919937491416931,
      "learning_rate": 0.00024909145629840645,
      "loss": 0.39577916264533997,
      "memory(GiB)": 78.26,
      "step": 1584,
      "token_acc": 0.8861203036791901,
      "train_speed(iter/s)": 0.032292
    },
    {
      "epoch": 0.30712590224289105,
      "grad_norm": 0.10316384583711624,
      "learning_rate": 0.0002490192668487445,
      "loss": 0.3726585805416107,
      "memory(GiB)": 78.26,
      "step": 1585,
      "token_acc": 0.8914358661264361,
      "train_speed(iter/s)": 0.032294
    },
    {
      "epoch": 0.3073196725282178,
      "grad_norm": 0.11321194469928741,
      "learning_rate": 0.00024894703672920894,
      "loss": 0.4326629340648651,
      "memory(GiB)": 78.26,
      "step": 1586,
      "token_acc": 0.8764373258299288,
      "train_speed(iter/s)": 0.032296
    },
    {
      "epoch": 0.30751344281354454,
      "grad_norm": 0.10429013520479202,
      "learning_rate": 0.0002488747659694665,
      "loss": 0.36875998973846436,
      "memory(GiB)": 78.26,
      "step": 1587,
      "token_acc": 0.8898948094499052,
      "train_speed(iter/s)": 0.032297
    },
    {
      "epoch": 0.3077072130988713,
      "grad_norm": 0.10509074479341507,
      "learning_rate": 0.0002488024545992009,
      "loss": 0.36970990896224976,
      "memory(GiB)": 78.26,
      "step": 1588,
      "token_acc": 0.8925745257452574,
      "train_speed(iter/s)": 0.032299
    },
    {
      "epoch": 0.307900983384198,
      "grad_norm": 0.09186790883541107,
      "learning_rate": 0.0002487301026481122,
      "loss": 0.3301558494567871,
      "memory(GiB)": 78.26,
      "step": 1589,
      "token_acc": 0.9027383654937571,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.30809475366952477,
      "grad_norm": 0.11293933540582657,
      "learning_rate": 0.00024865771014591733,
      "loss": 0.38940760493278503,
      "memory(GiB)": 78.26,
      "step": 1590,
      "token_acc": 0.8861825562725317,
      "train_speed(iter/s)": 0.032303
    },
    {
      "epoch": 0.3082885239548515,
      "grad_norm": 0.11680306494235992,
      "learning_rate": 0.0002485852771223499,
      "loss": 0.41928166151046753,
      "memory(GiB)": 78.26,
      "step": 1591,
      "token_acc": 0.8788892413276046,
      "train_speed(iter/s)": 0.032304
    },
    {
      "epoch": 0.30848229424017826,
      "grad_norm": 0.10015156865119934,
      "learning_rate": 0.00024851280360716014,
      "loss": 0.3776377737522125,
      "memory(GiB)": 78.26,
      "step": 1592,
      "token_acc": 0.8896725440806046,
      "train_speed(iter/s)": 0.032306
    },
    {
      "epoch": 0.308676064525505,
      "grad_norm": 0.11279778182506561,
      "learning_rate": 0.00024844028963011476,
      "loss": 0.39112144708633423,
      "memory(GiB)": 78.26,
      "step": 1593,
      "token_acc": 0.8870161362751848,
      "train_speed(iter/s)": 0.032308
    },
    {
      "epoch": 0.30886983481083174,
      "grad_norm": 0.09899583458900452,
      "learning_rate": 0.0002483677352209972,
      "loss": 0.35154473781585693,
      "memory(GiB)": 78.26,
      "step": 1594,
      "token_acc": 0.897304444024716,
      "train_speed(iter/s)": 0.032309
    },
    {
      "epoch": 0.3090636050961585,
      "grad_norm": 0.11332813650369644,
      "learning_rate": 0.0002482951404096076,
      "loss": 0.40011727809906006,
      "memory(GiB)": 78.26,
      "step": 1595,
      "token_acc": 0.8834134615384616,
      "train_speed(iter/s)": 0.032311
    },
    {
      "epoch": 0.30925737538148523,
      "grad_norm": 0.10168785601854324,
      "learning_rate": 0.00024822250522576247,
      "loss": 0.4107932150363922,
      "memory(GiB)": 78.26,
      "step": 1596,
      "token_acc": 0.8791540056914098,
      "train_speed(iter/s)": 0.032313
    },
    {
      "epoch": 0.309451145666812,
      "grad_norm": 0.10856655985116959,
      "learning_rate": 0.0002481498296992951,
      "loss": 0.40053847432136536,
      "memory(GiB)": 78.26,
      "step": 1597,
      "token_acc": 0.8820065490768003,
      "train_speed(iter/s)": 0.032314
    },
    {
      "epoch": 0.3096449159521387,
      "grad_norm": 0.10789895802736282,
      "learning_rate": 0.0002480771138600553,
      "loss": 0.3848547041416168,
      "memory(GiB)": 78.26,
      "step": 1598,
      "token_acc": 0.8867153284671533,
      "train_speed(iter/s)": 0.032316
    },
    {
      "epoch": 0.30983868623746547,
      "grad_norm": 0.11236843466758728,
      "learning_rate": 0.00024800435773790946,
      "loss": 0.40924182534217834,
      "memory(GiB)": 78.26,
      "step": 1599,
      "token_acc": 0.8796616904126696,
      "train_speed(iter/s)": 0.032318
    },
    {
      "epoch": 0.3100324565227922,
      "grad_norm": 0.10669440031051636,
      "learning_rate": 0.00024793156136274037,
      "loss": 0.38883164525032043,
      "memory(GiB)": 78.26,
      "step": 1600,
      "token_acc": 0.8854082720253724,
      "train_speed(iter/s)": 0.032319
    },
    {
      "epoch": 0.31022622680811895,
      "grad_norm": 0.10899162292480469,
      "learning_rate": 0.0002478587247644475,
      "loss": 0.3585772216320038,
      "memory(GiB)": 78.26,
      "step": 1601,
      "token_acc": 0.8958569414886266,
      "train_speed(iter/s)": 0.032313
    },
    {
      "epoch": 0.3104199970934457,
      "grad_norm": 0.11266548186540604,
      "learning_rate": 0.00024778584797294684,
      "loss": 0.3955519199371338,
      "memory(GiB)": 78.26,
      "step": 1602,
      "token_acc": 0.8861248676036985,
      "train_speed(iter/s)": 0.032315
    },
    {
      "epoch": 0.31061376737877244,
      "grad_norm": 0.1098904013633728,
      "learning_rate": 0.0002477129310181708,
      "loss": 0.42345693707466125,
      "memory(GiB)": 78.26,
      "step": 1603,
      "token_acc": 0.8796804389928986,
      "train_speed(iter/s)": 0.032316
    },
    {
      "epoch": 0.3108075376640992,
      "grad_norm": 0.10730142146348953,
      "learning_rate": 0.0002476399739300683,
      "loss": 0.40332454442977905,
      "memory(GiB)": 78.26,
      "step": 1604,
      "token_acc": 0.8845008085794536,
      "train_speed(iter/s)": 0.032318
    },
    {
      "epoch": 0.31100130794942593,
      "grad_norm": 0.10026847571134567,
      "learning_rate": 0.0002475669767386049,
      "loss": 0.3628842830657959,
      "memory(GiB)": 78.26,
      "step": 1605,
      "token_acc": 0.8923505698501966,
      "train_speed(iter/s)": 0.032319
    },
    {
      "epoch": 0.3111950782347527,
      "grad_norm": 0.12212540209293365,
      "learning_rate": 0.00024749393947376234,
      "loss": 0.38043132424354553,
      "memory(GiB)": 78.26,
      "step": 1606,
      "token_acc": 0.8881142442463197,
      "train_speed(iter/s)": 0.032321
    },
    {
      "epoch": 0.3113888485200794,
      "grad_norm": 0.10906349122524261,
      "learning_rate": 0.00024742086216553914,
      "loss": 0.4101215898990631,
      "memory(GiB)": 78.26,
      "step": 1607,
      "token_acc": 0.8820161592380544,
      "train_speed(iter/s)": 0.032323
    },
    {
      "epoch": 0.31158261880540616,
      "grad_norm": 0.10822609812021255,
      "learning_rate": 0.00024734774484395,
      "loss": 0.40345677733421326,
      "memory(GiB)": 78.26,
      "step": 1608,
      "token_acc": 0.8825836216839678,
      "train_speed(iter/s)": 0.032324
    },
    {
      "epoch": 0.31177638909073296,
      "grad_norm": 0.13473811745643616,
      "learning_rate": 0.00024727458753902624,
      "loss": 0.492877721786499,
      "memory(GiB)": 78.26,
      "step": 1609,
      "token_acc": 0.8599673445153629,
      "train_speed(iter/s)": 0.032326
    },
    {
      "epoch": 0.3119701593760597,
      "grad_norm": 0.10517790168523788,
      "learning_rate": 0.0002472013902808155,
      "loss": 0.375487744808197,
      "memory(GiB)": 78.26,
      "step": 1610,
      "token_acc": 0.8903864596563075,
      "train_speed(iter/s)": 0.032328
    },
    {
      "epoch": 0.31216392966138645,
      "grad_norm": 0.11642046272754669,
      "learning_rate": 0.00024712815309938186,
      "loss": 0.39327579736709595,
      "memory(GiB)": 78.26,
      "step": 1611,
      "token_acc": 0.8876873924797247,
      "train_speed(iter/s)": 0.03233
    },
    {
      "epoch": 0.3123576999467132,
      "grad_norm": 0.10896303504705429,
      "learning_rate": 0.00024705487602480583,
      "loss": 0.38862502574920654,
      "memory(GiB)": 78.26,
      "step": 1612,
      "token_acc": 0.88512,
      "train_speed(iter/s)": 0.032331
    },
    {
      "epoch": 0.31255147023203994,
      "grad_norm": 0.10455754399299622,
      "learning_rate": 0.0002469815590871842,
      "loss": 0.3765313923358917,
      "memory(GiB)": 78.26,
      "step": 1613,
      "token_acc": 0.890468422279189,
      "train_speed(iter/s)": 0.032333
    },
    {
      "epoch": 0.3127452405173667,
      "grad_norm": 0.10820908099412918,
      "learning_rate": 0.00024690820231663036,
      "loss": 0.40285009145736694,
      "memory(GiB)": 78.26,
      "step": 1614,
      "token_acc": 0.8830803366074086,
      "train_speed(iter/s)": 0.032335
    },
    {
      "epoch": 0.31293901080269343,
      "grad_norm": 0.10197239369153976,
      "learning_rate": 0.0002468348057432737,
      "loss": 0.3643296957015991,
      "memory(GiB)": 78.26,
      "step": 1615,
      "token_acc": 0.8937665096266442,
      "train_speed(iter/s)": 0.032336
    },
    {
      "epoch": 0.3131327810880202,
      "grad_norm": 0.12070825695991516,
      "learning_rate": 0.00024676136939726036,
      "loss": 0.4440290331840515,
      "memory(GiB)": 78.26,
      "step": 1616,
      "token_acc": 0.8726066239878395,
      "train_speed(iter/s)": 0.032338
    },
    {
      "epoch": 0.3133265513733469,
      "grad_norm": 0.10100740194320679,
      "learning_rate": 0.0002466878933087525,
      "loss": 0.38823482394218445,
      "memory(GiB)": 78.26,
      "step": 1617,
      "token_acc": 0.8885969521807672,
      "train_speed(iter/s)": 0.03234
    },
    {
      "epoch": 0.31352032165867366,
      "grad_norm": 0.10464068502187729,
      "learning_rate": 0.00024661437750792865,
      "loss": 0.3649863004684448,
      "memory(GiB)": 78.26,
      "step": 1618,
      "token_acc": 0.8917689623982427,
      "train_speed(iter/s)": 0.032341
    },
    {
      "epoch": 0.3137140919440004,
      "grad_norm": 0.09842365980148315,
      "learning_rate": 0.00024654082202498395,
      "loss": 0.37648287415504456,
      "memory(GiB)": 78.26,
      "step": 1619,
      "token_acc": 0.8893629913904108,
      "train_speed(iter/s)": 0.032343
    },
    {
      "epoch": 0.31390786222932715,
      "grad_norm": 0.11136578768491745,
      "learning_rate": 0.00024646722689012946,
      "loss": 0.3890235722064972,
      "memory(GiB)": 78.26,
      "step": 1620,
      "token_acc": 0.8880463144161774,
      "train_speed(iter/s)": 0.032345
    },
    {
      "epoch": 0.3141016325146539,
      "grad_norm": 0.12551820278167725,
      "learning_rate": 0.0002463935921335927,
      "loss": 0.39017459750175476,
      "memory(GiB)": 78.26,
      "step": 1621,
      "token_acc": 0.8850291533218363,
      "train_speed(iter/s)": 0.032347
    },
    {
      "epoch": 0.31429540279998064,
      "grad_norm": 0.12295407056808472,
      "learning_rate": 0.00024631991778561747,
      "loss": 0.4268419146537781,
      "memory(GiB)": 78.26,
      "step": 1622,
      "token_acc": 0.8766729419818586,
      "train_speed(iter/s)": 0.032348
    },
    {
      "epoch": 0.3144891730853074,
      "grad_norm": 0.10500724613666534,
      "learning_rate": 0.00024624620387646377,
      "loss": 0.3578689992427826,
      "memory(GiB)": 78.26,
      "step": 1623,
      "token_acc": 0.8940465211599549,
      "train_speed(iter/s)": 0.03235
    },
    {
      "epoch": 0.3146829433706341,
      "grad_norm": 0.1127677783370018,
      "learning_rate": 0.0002461724504364079,
      "loss": 0.41033974289894104,
      "memory(GiB)": 78.26,
      "step": 1624,
      "token_acc": 0.8811105837683911,
      "train_speed(iter/s)": 0.032352
    },
    {
      "epoch": 0.31487671365596087,
      "grad_norm": 0.10852906107902527,
      "learning_rate": 0.0002460986574957424,
      "loss": 0.4147697687149048,
      "memory(GiB)": 78.26,
      "step": 1625,
      "token_acc": 0.8803296877476621,
      "train_speed(iter/s)": 0.032353
    },
    {
      "epoch": 0.3150704839412876,
      "grad_norm": 0.10529500991106033,
      "learning_rate": 0.000246024825084776,
      "loss": 0.3946833312511444,
      "memory(GiB)": 78.26,
      "step": 1626,
      "token_acc": 0.8831530219494435,
      "train_speed(iter/s)": 0.032355
    },
    {
      "epoch": 0.31526425422661436,
      "grad_norm": 0.10391051322221756,
      "learning_rate": 0.00024595095323383365,
      "loss": 0.3614901900291443,
      "memory(GiB)": 78.26,
      "step": 1627,
      "token_acc": 0.8932863813825308,
      "train_speed(iter/s)": 0.032357
    },
    {
      "epoch": 0.3154580245119411,
      "grad_norm": 0.11637852340936661,
      "learning_rate": 0.00024587704197325655,
      "loss": 0.43050843477249146,
      "memory(GiB)": 78.26,
      "step": 1628,
      "token_acc": 0.8762489252571971,
      "train_speed(iter/s)": 0.032359
    },
    {
      "epoch": 0.31565179479726785,
      "grad_norm": 0.11949366331100464,
      "learning_rate": 0.0002458030913334019,
      "loss": 0.4221411347389221,
      "memory(GiB)": 78.26,
      "step": 1629,
      "token_acc": 0.8803364945335391,
      "train_speed(iter/s)": 0.03236
    },
    {
      "epoch": 0.3158455650825946,
      "grad_norm": 0.10639101266860962,
      "learning_rate": 0.0002457291013446434,
      "loss": 0.387704074382782,
      "memory(GiB)": 78.26,
      "step": 1630,
      "token_acc": 0.8859933917540583,
      "train_speed(iter/s)": 0.032362
    },
    {
      "epoch": 0.31603933536792134,
      "grad_norm": 0.12497800588607788,
      "learning_rate": 0.00024565507203737054,
      "loss": 0.4306849539279938,
      "memory(GiB)": 78.26,
      "step": 1631,
      "token_acc": 0.8744147105336599,
      "train_speed(iter/s)": 0.032364
    },
    {
      "epoch": 0.3162331056532481,
      "grad_norm": 0.10723750293254852,
      "learning_rate": 0.0002455810034419893,
      "loss": 0.41962188482284546,
      "memory(GiB)": 78.26,
      "step": 1632,
      "token_acc": 0.8786517987789495,
      "train_speed(iter/s)": 0.032366
    },
    {
      "epoch": 0.3164268759385748,
      "grad_norm": 0.10330141335725784,
      "learning_rate": 0.0002455068955889216,
      "loss": 0.39232727885246277,
      "memory(GiB)": 78.26,
      "step": 1633,
      "token_acc": 0.8855892466395748,
      "train_speed(iter/s)": 0.032367
    },
    {
      "epoch": 0.31662064622390157,
      "grad_norm": 0.09860360622406006,
      "learning_rate": 0.0002454327485086055,
      "loss": 0.37398630380630493,
      "memory(GiB)": 78.26,
      "step": 1634,
      "token_acc": 0.8935310637039957,
      "train_speed(iter/s)": 0.032369
    },
    {
      "epoch": 0.3168144165092283,
      "grad_norm": 0.10215026140213013,
      "learning_rate": 0.00024535856223149524,
      "loss": 0.3739304840564728,
      "memory(GiB)": 78.26,
      "step": 1635,
      "token_acc": 0.889224391616981,
      "train_speed(iter/s)": 0.03237
    },
    {
      "epoch": 0.31700818679455506,
      "grad_norm": 0.10608868300914764,
      "learning_rate": 0.000245284336788061,
      "loss": 0.40002018213272095,
      "memory(GiB)": 78.26,
      "step": 1636,
      "token_acc": 0.8821046707934721,
      "train_speed(iter/s)": 0.032372
    },
    {
      "epoch": 0.3172019570798818,
      "grad_norm": 0.10536207258701324,
      "learning_rate": 0.0002452100722087893,
      "loss": 0.38114506006240845,
      "memory(GiB)": 78.26,
      "step": 1637,
      "token_acc": 0.8877112648882257,
      "train_speed(iter/s)": 0.032373
    },
    {
      "epoch": 0.31739572736520855,
      "grad_norm": 0.10534202307462692,
      "learning_rate": 0.00024513576852418256,
      "loss": 0.3625592887401581,
      "memory(GiB)": 78.26,
      "step": 1638,
      "token_acc": 0.8946749986647439,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.3175894976505353,
      "grad_norm": 0.10395082831382751,
      "learning_rate": 0.0002450614257647593,
      "loss": 0.40679264068603516,
      "memory(GiB)": 78.26,
      "step": 1639,
      "token_acc": 0.882569104812343,
      "train_speed(iter/s)": 0.032377
    },
    {
      "epoch": 0.31778326793586203,
      "grad_norm": 0.11221817880868912,
      "learning_rate": 0.00024498704396105404,
      "loss": 0.4055024981498718,
      "memory(GiB)": 78.26,
      "step": 1640,
      "token_acc": 0.882479675964494,
      "train_speed(iter/s)": 0.032378
    },
    {
      "epoch": 0.3179770382211888,
      "grad_norm": 0.09974303096532822,
      "learning_rate": 0.00024491262314361745,
      "loss": 0.38338178396224976,
      "memory(GiB)": 78.26,
      "step": 1641,
      "token_acc": 0.8880624860078108,
      "train_speed(iter/s)": 0.03238
    },
    {
      "epoch": 0.3181708085065155,
      "grad_norm": 0.10050232708454132,
      "learning_rate": 0.0002448381633430161,
      "loss": 0.35944464802742004,
      "memory(GiB)": 78.26,
      "step": 1642,
      "token_acc": 0.8940914437298231,
      "train_speed(iter/s)": 0.032382
    },
    {
      "epoch": 0.31836457879184227,
      "grad_norm": 0.11217102408409119,
      "learning_rate": 0.0002447636645898327,
      "loss": 0.39601394534111023,
      "memory(GiB)": 78.26,
      "step": 1643,
      "token_acc": 0.8860522531810413,
      "train_speed(iter/s)": 0.032383
    },
    {
      "epoch": 0.318558349077169,
      "grad_norm": 0.11227677762508392,
      "learning_rate": 0.00024468912691466587,
      "loss": 0.39023369550704956,
      "memory(GiB)": 78.26,
      "step": 1644,
      "token_acc": 0.8874575239640919,
      "train_speed(iter/s)": 0.032385
    },
    {
      "epoch": 0.31875211936249576,
      "grad_norm": 0.10062558948993683,
      "learning_rate": 0.00024461455034813017,
      "loss": 0.37629178166389465,
      "memory(GiB)": 78.26,
      "step": 1645,
      "token_acc": 0.8884143616228832,
      "train_speed(iter/s)": 0.032386
    },
    {
      "epoch": 0.3189458896478225,
      "grad_norm": 0.12126388400793076,
      "learning_rate": 0.0002445399349208563,
      "loss": 0.4465317726135254,
      "memory(GiB)": 78.26,
      "step": 1646,
      "token_acc": 0.8706858890345129,
      "train_speed(iter/s)": 0.032388
    },
    {
      "epoch": 0.31913965993314924,
      "grad_norm": 0.10243180394172668,
      "learning_rate": 0.00024446528066349074,
      "loss": 0.3463560938835144,
      "memory(GiB)": 78.26,
      "step": 1647,
      "token_acc": 0.8977993900931345,
      "train_speed(iter/s)": 0.03239
    },
    {
      "epoch": 0.319333430218476,
      "grad_norm": 0.09814903885126114,
      "learning_rate": 0.00024439058760669603,
      "loss": 0.3848741054534912,
      "memory(GiB)": 78.26,
      "step": 1648,
      "token_acc": 0.8868541204061775,
      "train_speed(iter/s)": 0.032391
    },
    {
      "epoch": 0.31952720050380273,
      "grad_norm": 0.10367216914892197,
      "learning_rate": 0.00024431585578115064,
      "loss": 0.38090386986732483,
      "memory(GiB)": 78.26,
      "step": 1649,
      "token_acc": 0.8876089324618737,
      "train_speed(iter/s)": 0.032393
    },
    {
      "epoch": 0.3197209707891295,
      "grad_norm": 0.11322217434644699,
      "learning_rate": 0.00024424108521754886,
      "loss": 0.3961893916130066,
      "memory(GiB)": 78.26,
      "step": 1650,
      "token_acc": 0.8858755383929897,
      "train_speed(iter/s)": 0.032394
    },
    {
      "epoch": 0.3199147410744562,
      "grad_norm": 0.11258938163518906,
      "learning_rate": 0.00024416627594660105,
      "loss": 0.40742677450180054,
      "memory(GiB)": 78.26,
      "step": 1651,
      "token_acc": 0.8804218539370391,
      "train_speed(iter/s)": 0.032396
    },
    {
      "epoch": 0.32010851135978297,
      "grad_norm": 0.10289296507835388,
      "learning_rate": 0.00024409142799903342,
      "loss": 0.37639862298965454,
      "memory(GiB)": 78.26,
      "step": 1652,
      "token_acc": 0.8907713884992987,
      "train_speed(iter/s)": 0.032398
    },
    {
      "epoch": 0.3203022816451097,
      "grad_norm": 0.10312007367610931,
      "learning_rate": 0.00024401654140558795,
      "loss": 0.39015206694602966,
      "memory(GiB)": 78.26,
      "step": 1653,
      "token_acc": 0.8851129761136217,
      "train_speed(iter/s)": 0.032399
    },
    {
      "epoch": 0.32049605193043645,
      "grad_norm": 0.09907987713813782,
      "learning_rate": 0.00024394161619702257,
      "loss": 0.35583794116973877,
      "memory(GiB)": 78.26,
      "step": 1654,
      "token_acc": 0.8962283436398017,
      "train_speed(iter/s)": 0.032401
    },
    {
      "epoch": 0.3206898222157632,
      "grad_norm": 0.11065588891506195,
      "learning_rate": 0.00024386665240411115,
      "loss": 0.3847392499446869,
      "memory(GiB)": 78.26,
      "step": 1655,
      "token_acc": 0.886949811533553,
      "train_speed(iter/s)": 0.032402
    },
    {
      "epoch": 0.32088359250108994,
      "grad_norm": 0.10860442370176315,
      "learning_rate": 0.0002437916500576433,
      "loss": 0.4008861184120178,
      "memory(GiB)": 78.33,
      "step": 1656,
      "token_acc": 0.8821946594228897,
      "train_speed(iter/s)": 0.032404
    },
    {
      "epoch": 0.3210773627864167,
      "grad_norm": 0.11284346878528595,
      "learning_rate": 0.0002437166091884244,
      "loss": 0.3918432593345642,
      "memory(GiB)": 78.33,
      "step": 1657,
      "token_acc": 0.8856465073739568,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.32127113307174343,
      "grad_norm": 0.11324102431535721,
      "learning_rate": 0.00024364152982727592,
      "loss": 0.4203750789165497,
      "memory(GiB)": 78.33,
      "step": 1658,
      "token_acc": 0.8781059802334023,
      "train_speed(iter/s)": 0.032407
    },
    {
      "epoch": 0.3214649033570702,
      "grad_norm": 0.1033095270395279,
      "learning_rate": 0.0002435664120050349,
      "loss": 0.3659226894378662,
      "memory(GiB)": 78.33,
      "step": 1659,
      "token_acc": 0.8937110992198721,
      "train_speed(iter/s)": 0.032409
    },
    {
      "epoch": 0.3216586736423969,
      "grad_norm": 0.11475743353366852,
      "learning_rate": 0.0002434912557525542,
      "loss": 0.4008634686470032,
      "memory(GiB)": 78.33,
      "step": 1660,
      "token_acc": 0.8813870157718026,
      "train_speed(iter/s)": 0.03241
    },
    {
      "epoch": 0.32185244392772366,
      "grad_norm": 0.11086868494749069,
      "learning_rate": 0.0002434160611007026,
      "loss": 0.4097922742366791,
      "memory(GiB)": 78.33,
      "step": 1661,
      "token_acc": 0.8791088493742418,
      "train_speed(iter/s)": 0.032412
    },
    {
      "epoch": 0.3220462142130504,
      "grad_norm": 0.09878856688737869,
      "learning_rate": 0.0002433408280803645,
      "loss": 0.3229445517063141,
      "memory(GiB)": 78.33,
      "step": 1662,
      "token_acc": 0.9044224128327949,
      "train_speed(iter/s)": 0.032414
    },
    {
      "epoch": 0.32223998449837715,
      "grad_norm": 0.1650267392396927,
      "learning_rate": 0.00024326555672244012,
      "loss": 0.3968316912651062,
      "memory(GiB)": 78.33,
      "step": 1663,
      "token_acc": 0.8846646732165742,
      "train_speed(iter/s)": 0.032415
    },
    {
      "epoch": 0.3224337547837039,
      "grad_norm": 0.11585035920143127,
      "learning_rate": 0.0002431902470578455,
      "loss": 0.4106610417366028,
      "memory(GiB)": 78.33,
      "step": 1664,
      "token_acc": 0.8812659238112779,
      "train_speed(iter/s)": 0.032417
    },
    {
      "epoch": 0.32262752506903064,
      "grad_norm": 0.10727370530366898,
      "learning_rate": 0.00024311489911751224,
      "loss": 0.3810875117778778,
      "memory(GiB)": 78.33,
      "step": 1665,
      "token_acc": 0.8862208987547374,
      "train_speed(iter/s)": 0.032419
    },
    {
      "epoch": 0.3228212953543574,
      "grad_norm": 0.10232355445623398,
      "learning_rate": 0.00024303951293238785,
      "loss": 0.40213391184806824,
      "memory(GiB)": 78.33,
      "step": 1666,
      "token_acc": 0.8846539618856569,
      "train_speed(iter/s)": 0.03242
    },
    {
      "epoch": 0.32301506563968413,
      "grad_norm": 0.10427332669496536,
      "learning_rate": 0.00024296408853343544,
      "loss": 0.381198525428772,
      "memory(GiB)": 78.33,
      "step": 1667,
      "token_acc": 0.886510858088634,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.3232088359250109,
      "grad_norm": 0.11577033996582031,
      "learning_rate": 0.0002428886259516338,
      "loss": 0.38930296897888184,
      "memory(GiB)": 78.33,
      "step": 1668,
      "token_acc": 0.8849813571961223,
      "train_speed(iter/s)": 0.032424
    },
    {
      "epoch": 0.3234026062103376,
      "grad_norm": 0.10479265451431274,
      "learning_rate": 0.0002428131252179775,
      "loss": 0.38501253724098206,
      "memory(GiB)": 78.33,
      "step": 1669,
      "token_acc": 0.8847304574878965,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.3235963764956644,
      "grad_norm": 0.10768333077430725,
      "learning_rate": 0.00024273758636347663,
      "loss": 0.4145703613758087,
      "memory(GiB)": 78.33,
      "step": 1670,
      "token_acc": 0.8764765532929537,
      "train_speed(iter/s)": 0.032427
    },
    {
      "epoch": 0.32379014678099116,
      "grad_norm": 0.1042892262339592,
      "learning_rate": 0.00024266200941915712,
      "loss": 0.3692252039909363,
      "memory(GiB)": 78.33,
      "step": 1671,
      "token_acc": 0.8922046134447549,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.3239839170663179,
      "grad_norm": 0.11078062653541565,
      "learning_rate": 0.00024258639441606042,
      "loss": 0.4250616133213043,
      "memory(GiB)": 78.33,
      "step": 1672,
      "token_acc": 0.876742678050362,
      "train_speed(iter/s)": 0.03243
    },
    {
      "epoch": 0.32417768735164465,
      "grad_norm": 0.10978075116872787,
      "learning_rate": 0.00024251074138524365,
      "loss": 0.40720629692077637,
      "memory(GiB)": 78.33,
      "step": 1673,
      "token_acc": 0.8835114016903205,
      "train_speed(iter/s)": 0.032432
    },
    {
      "epoch": 0.3243714576369714,
      "grad_norm": 0.09793038666248322,
      "learning_rate": 0.00024243505035777954,
      "loss": 0.34907984733581543,
      "memory(GiB)": 78.33,
      "step": 1674,
      "token_acc": 0.8983231559561888,
      "train_speed(iter/s)": 0.032433
    },
    {
      "epoch": 0.32456522792229814,
      "grad_norm": 0.11205164343118668,
      "learning_rate": 0.0002423593213647564,
      "loss": 0.3938961625099182,
      "memory(GiB)": 78.33,
      "step": 1675,
      "token_acc": 0.8874208948595537,
      "train_speed(iter/s)": 0.032435
    },
    {
      "epoch": 0.3247589982076249,
      "grad_norm": 0.12075439840555191,
      "learning_rate": 0.0002422835544372782,
      "loss": 0.42155468463897705,
      "memory(GiB)": 78.33,
      "step": 1676,
      "token_acc": 0.8779719830355995,
      "train_speed(iter/s)": 0.032436
    },
    {
      "epoch": 0.3249527684929516,
      "grad_norm": 0.10721995681524277,
      "learning_rate": 0.0002422077496064644,
      "loss": 0.4283701479434967,
      "memory(GiB)": 78.33,
      "step": 1677,
      "token_acc": 0.874123831775701,
      "train_speed(iter/s)": 0.032438
    },
    {
      "epoch": 0.32514653877827837,
      "grad_norm": 0.12848171591758728,
      "learning_rate": 0.00024213190690345018,
      "loss": 0.4088367521762848,
      "memory(GiB)": 78.33,
      "step": 1678,
      "token_acc": 0.8799862555836692,
      "train_speed(iter/s)": 0.03244
    },
    {
      "epoch": 0.3253403090636051,
      "grad_norm": 0.10981763154268265,
      "learning_rate": 0.00024205602635938604,
      "loss": 0.41480377316474915,
      "memory(GiB)": 78.33,
      "step": 1679,
      "token_acc": 0.878927367125194,
      "train_speed(iter/s)": 0.032441
    },
    {
      "epoch": 0.32553407934893186,
      "grad_norm": 0.10942694544792175,
      "learning_rate": 0.0002419801080054383,
      "loss": 0.39324793219566345,
      "memory(GiB)": 78.33,
      "step": 1680,
      "token_acc": 0.8848377444002208,
      "train_speed(iter/s)": 0.032443
    },
    {
      "epoch": 0.3257278496342586,
      "grad_norm": 0.10424201935529709,
      "learning_rate": 0.00024190415187278855,
      "loss": 0.38931453227996826,
      "memory(GiB)": 78.33,
      "step": 1681,
      "token_acc": 0.8873810462300805,
      "train_speed(iter/s)": 0.032445
    },
    {
      "epoch": 0.32592161991958535,
      "grad_norm": 0.10016170144081116,
      "learning_rate": 0.0002418281579926341,
      "loss": 0.3769393861293793,
      "memory(GiB)": 78.33,
      "step": 1682,
      "token_acc": 0.888585472419443,
      "train_speed(iter/s)": 0.032446
    },
    {
      "epoch": 0.3261153902049121,
      "grad_norm": 0.10119081288576126,
      "learning_rate": 0.0002417521263961876,
      "loss": 0.34021657705307007,
      "memory(GiB)": 78.33,
      "step": 1683,
      "token_acc": 0.8990787809976002,
      "train_speed(iter/s)": 0.032448
    },
    {
      "epoch": 0.32630916049023884,
      "grad_norm": 0.10256388038396835,
      "learning_rate": 0.00024167605711467738,
      "loss": 0.3520572781562805,
      "memory(GiB)": 78.33,
      "step": 1684,
      "token_acc": 0.8965411006861784,
      "train_speed(iter/s)": 0.03245
    },
    {
      "epoch": 0.3265029307755656,
      "grad_norm": 0.10342499613761902,
      "learning_rate": 0.00024159995017934702,
      "loss": 0.37822040915489197,
      "memory(GiB)": 78.33,
      "step": 1685,
      "token_acc": 0.8892546779308175,
      "train_speed(iter/s)": 0.032451
    },
    {
      "epoch": 0.3266967010608923,
      "grad_norm": 0.10906314849853516,
      "learning_rate": 0.00024152380562145575,
      "loss": 0.3807571232318878,
      "memory(GiB)": 78.33,
      "step": 1686,
      "token_acc": 0.8904570071765867,
      "train_speed(iter/s)": 0.032453
    },
    {
      "epoch": 0.32689047134621907,
      "grad_norm": 0.10030427575111389,
      "learning_rate": 0.00024144762347227822,
      "loss": 0.35491418838500977,
      "memory(GiB)": 78.33,
      "step": 1687,
      "token_acc": 0.8963004898220396,
      "train_speed(iter/s)": 0.032454
    },
    {
      "epoch": 0.3270842416315458,
      "grad_norm": 0.10785649716854095,
      "learning_rate": 0.0002413714037631044,
      "loss": 0.4118516743183136,
      "memory(GiB)": 78.33,
      "step": 1688,
      "token_acc": 0.8789999213774667,
      "train_speed(iter/s)": 0.032456
    },
    {
      "epoch": 0.32727801191687256,
      "grad_norm": 0.10498232394456863,
      "learning_rate": 0.00024129514652523976,
      "loss": 0.35566800832748413,
      "memory(GiB)": 78.33,
      "step": 1689,
      "token_acc": 0.8963449018301123,
      "train_speed(iter/s)": 0.032457
    },
    {
      "epoch": 0.3274717822021993,
      "grad_norm": 0.09992900490760803,
      "learning_rate": 0.0002412188517900053,
      "loss": 0.3673107326030731,
      "memory(GiB)": 78.33,
      "step": 1690,
      "token_acc": 0.891633487007544,
      "train_speed(iter/s)": 0.032459
    },
    {
      "epoch": 0.32766555248752605,
      "grad_norm": 0.10761499404907227,
      "learning_rate": 0.00024114251958873726,
      "loss": 0.39024174213409424,
      "memory(GiB)": 78.33,
      "step": 1691,
      "token_acc": 0.886815954076521,
      "train_speed(iter/s)": 0.03246
    },
    {
      "epoch": 0.3278593227728528,
      "grad_norm": 0.10518249124288559,
      "learning_rate": 0.00024106614995278731,
      "loss": 0.38077983260154724,
      "memory(GiB)": 78.33,
      "step": 1692,
      "token_acc": 0.8875674808915496,
      "train_speed(iter/s)": 0.032462
    },
    {
      "epoch": 0.32805309305817953,
      "grad_norm": 0.10248047858476639,
      "learning_rate": 0.00024098974291352255,
      "loss": 0.36225420236587524,
      "memory(GiB)": 78.33,
      "step": 1693,
      "token_acc": 0.8952485416562797,
      "train_speed(iter/s)": 0.032464
    },
    {
      "epoch": 0.3282468633435063,
      "grad_norm": 0.10516565293073654,
      "learning_rate": 0.00024091329850232536,
      "loss": 0.391397625207901,
      "memory(GiB)": 78.33,
      "step": 1694,
      "token_acc": 0.8858848043629254,
      "train_speed(iter/s)": 0.032465
    },
    {
      "epoch": 0.328440633628833,
      "grad_norm": 0.1099853664636612,
      "learning_rate": 0.00024083681675059356,
      "loss": 0.38383805751800537,
      "memory(GiB)": 78.33,
      "step": 1695,
      "token_acc": 0.8891934663426214,
      "train_speed(iter/s)": 0.032467
    },
    {
      "epoch": 0.32863440391415977,
      "grad_norm": 0.11349144577980042,
      "learning_rate": 0.00024076029768974025,
      "loss": 0.4004535675048828,
      "memory(GiB)": 78.33,
      "step": 1696,
      "token_acc": 0.883281280229955,
      "train_speed(iter/s)": 0.032468
    },
    {
      "epoch": 0.3288281741994865,
      "grad_norm": 0.11258336901664734,
      "learning_rate": 0.00024068374135119384,
      "loss": 0.4041019678115845,
      "memory(GiB)": 78.33,
      "step": 1697,
      "token_acc": 0.8841508871328463,
      "train_speed(iter/s)": 0.03247
    },
    {
      "epoch": 0.32902194448481326,
      "grad_norm": 0.10647819191217422,
      "learning_rate": 0.00024060714776639813,
      "loss": 0.36608096957206726,
      "memory(GiB)": 78.33,
      "step": 1698,
      "token_acc": 0.8930285460648598,
      "train_speed(iter/s)": 0.032471
    },
    {
      "epoch": 0.32921571477014,
      "grad_norm": 0.10232997685670853,
      "learning_rate": 0.00024053051696681208,
      "loss": 0.37514373660087585,
      "memory(GiB)": 78.33,
      "step": 1699,
      "token_acc": 0.8901580589433514,
      "train_speed(iter/s)": 0.032473
    },
    {
      "epoch": 0.32940948505546674,
      "grad_norm": 0.10426058620214462,
      "learning_rate": 0.00024045384898391007,
      "loss": 0.3706841468811035,
      "memory(GiB)": 78.33,
      "step": 1700,
      "token_acc": 0.8918139787870547,
      "train_speed(iter/s)": 0.032474
    },
    {
      "epoch": 0.3296032553407935,
      "grad_norm": 0.11431006342172623,
      "learning_rate": 0.0002403771438491817,
      "loss": 0.40189433097839355,
      "memory(GiB)": 78.33,
      "step": 1701,
      "token_acc": 0.8844837336666964,
      "train_speed(iter/s)": 0.032476
    },
    {
      "epoch": 0.32979702562612023,
      "grad_norm": 0.11498620361089706,
      "learning_rate": 0.00024030040159413185,
      "loss": 0.4026452898979187,
      "memory(GiB)": 78.33,
      "step": 1702,
      "token_acc": 0.8820736746475984,
      "train_speed(iter/s)": 0.032478
    },
    {
      "epoch": 0.329990795911447,
      "grad_norm": 0.10887457430362701,
      "learning_rate": 0.0002402236222502805,
      "loss": 0.3983825445175171,
      "memory(GiB)": 78.33,
      "step": 1703,
      "token_acc": 0.8844022811427336,
      "train_speed(iter/s)": 0.032479
    },
    {
      "epoch": 0.3301845661967737,
      "grad_norm": 0.10820619016885757,
      "learning_rate": 0.00024014680584916322,
      "loss": 0.37078577280044556,
      "memory(GiB)": 78.33,
      "step": 1704,
      "token_acc": 0.8899442436468578,
      "train_speed(iter/s)": 0.032481
    },
    {
      "epoch": 0.33037833648210047,
      "grad_norm": 0.12457577884197235,
      "learning_rate": 0.00024006995242233038,
      "loss": 0.4524819552898407,
      "memory(GiB)": 78.33,
      "step": 1705,
      "token_acc": 0.867928674577761,
      "train_speed(iter/s)": 0.032482
    },
    {
      "epoch": 0.3305721067674272,
      "grad_norm": 0.11018163710832596,
      "learning_rate": 0.0002399930620013478,
      "loss": 0.39005085825920105,
      "memory(GiB)": 78.33,
      "step": 1706,
      "token_acc": 0.8860519838801287,
      "train_speed(iter/s)": 0.032484
    },
    {
      "epoch": 0.33076587705275395,
      "grad_norm": 0.10681940615177155,
      "learning_rate": 0.00023991613461779644,
      "loss": 0.38101130723953247,
      "memory(GiB)": 78.33,
      "step": 1707,
      "token_acc": 0.890979347101932,
      "train_speed(iter/s)": 0.032485
    },
    {
      "epoch": 0.3309596473380807,
      "grad_norm": 0.10825340449810028,
      "learning_rate": 0.00023983917030327248,
      "loss": 0.36492252349853516,
      "memory(GiB)": 78.33,
      "step": 1708,
      "token_acc": 0.8953067555147058,
      "train_speed(iter/s)": 0.032487
    },
    {
      "epoch": 0.33115341762340744,
      "grad_norm": 0.0985020250082016,
      "learning_rate": 0.00023976216908938719,
      "loss": 0.36632204055786133,
      "memory(GiB)": 78.33,
      "step": 1709,
      "token_acc": 0.8902666632577404,
      "train_speed(iter/s)": 0.032488
    },
    {
      "epoch": 0.3313471879087342,
      "grad_norm": 0.10576551407575607,
      "learning_rate": 0.00023968513100776703,
      "loss": 0.3914228081703186,
      "memory(GiB)": 78.33,
      "step": 1710,
      "token_acc": 0.8861784230125749,
      "train_speed(iter/s)": 0.03249
    },
    {
      "epoch": 0.33154095819406093,
      "grad_norm": 0.10542413592338562,
      "learning_rate": 0.00023960805609005365,
      "loss": 0.40728819370269775,
      "memory(GiB)": 78.33,
      "step": 1711,
      "token_acc": 0.8833345786943635,
      "train_speed(iter/s)": 0.032491
    },
    {
      "epoch": 0.3317347284793877,
      "grad_norm": 0.10389473289251328,
      "learning_rate": 0.0002395309443679038,
      "loss": 0.3726048171520233,
      "memory(GiB)": 78.33,
      "step": 1712,
      "token_acc": 0.8894928520954504,
      "train_speed(iter/s)": 0.032493
    },
    {
      "epoch": 0.3319284987647144,
      "grad_norm": 0.11621604114770889,
      "learning_rate": 0.0002394537958729893,
      "loss": 0.39150145649909973,
      "memory(GiB)": 78.33,
      "step": 1713,
      "token_acc": 0.8860767814882987,
      "train_speed(iter/s)": 0.032494
    },
    {
      "epoch": 0.33212226905004116,
      "grad_norm": 0.13983365893363953,
      "learning_rate": 0.00023937661063699707,
      "loss": 0.4037204384803772,
      "memory(GiB)": 78.33,
      "step": 1714,
      "token_acc": 0.882203057624461,
      "train_speed(iter/s)": 0.032496
    },
    {
      "epoch": 0.3323160393353679,
      "grad_norm": 0.11795634776353836,
      "learning_rate": 0.00023929938869162928,
      "loss": 0.43276962637901306,
      "memory(GiB)": 78.33,
      "step": 1715,
      "token_acc": 0.875131748448296,
      "train_speed(iter/s)": 0.032498
    },
    {
      "epoch": 0.33250980962069465,
      "grad_norm": 0.10593996942043304,
      "learning_rate": 0.00023922213006860292,
      "loss": 0.3847067654132843,
      "memory(GiB)": 78.33,
      "step": 1716,
      "token_acc": 0.888629105839416,
      "train_speed(iter/s)": 0.032499
    },
    {
      "epoch": 0.3327035799060214,
      "grad_norm": 0.11202581226825714,
      "learning_rate": 0.00023914483479965025,
      "loss": 0.4195789396762848,
      "memory(GiB)": 78.33,
      "step": 1717,
      "token_acc": 0.8793484873624247,
      "train_speed(iter/s)": 0.0325
    },
    {
      "epoch": 0.33289735019134814,
      "grad_norm": 0.11114699393510818,
      "learning_rate": 0.00023906750291651858,
      "loss": 0.3933086395263672,
      "memory(GiB)": 78.33,
      "step": 1718,
      "token_acc": 0.8864894491641545,
      "train_speed(iter/s)": 0.032502
    },
    {
      "epoch": 0.3330911204766749,
      "grad_norm": 0.11813364177942276,
      "learning_rate": 0.00023899013445097007,
      "loss": 0.39278751611709595,
      "memory(GiB)": 78.33,
      "step": 1719,
      "token_acc": 0.8872432671919998,
      "train_speed(iter/s)": 0.032504
    },
    {
      "epoch": 0.33328489076200163,
      "grad_norm": 0.1009705662727356,
      "learning_rate": 0.0002389127294347821,
      "loss": 0.38593360781669617,
      "memory(GiB)": 78.33,
      "step": 1720,
      "token_acc": 0.8872761447292852,
      "train_speed(iter/s)": 0.032505
    },
    {
      "epoch": 0.3334786610473284,
      "grad_norm": 0.10921531915664673,
      "learning_rate": 0.000238835287899747,
      "loss": 0.38367462158203125,
      "memory(GiB)": 78.33,
      "step": 1721,
      "token_acc": 0.8886438230274386,
      "train_speed(iter/s)": 0.032507
    },
    {
      "epoch": 0.3336724313326551,
      "grad_norm": 0.1057095155119896,
      "learning_rate": 0.00023875780987767204,
      "loss": 0.3982135057449341,
      "memory(GiB)": 78.33,
      "step": 1722,
      "token_acc": 0.8859919028340081,
      "train_speed(iter/s)": 0.032508
    },
    {
      "epoch": 0.33386620161798186,
      "grad_norm": 0.10099222511053085,
      "learning_rate": 0.0002386802954003795,
      "loss": 0.37414735555648804,
      "memory(GiB)": 78.33,
      "step": 1723,
      "token_acc": 0.8888457355911139,
      "train_speed(iter/s)": 0.032509
    },
    {
      "epoch": 0.3340599719033086,
      "grad_norm": 0.10116644203662872,
      "learning_rate": 0.0002386027444997068,
      "loss": 0.4056088328361511,
      "memory(GiB)": 78.33,
      "step": 1724,
      "token_acc": 0.8838143450628663,
      "train_speed(iter/s)": 0.032511
    },
    {
      "epoch": 0.33425374218863535,
      "grad_norm": 0.0998440757393837,
      "learning_rate": 0.000238525157207506,
      "loss": 0.3904368281364441,
      "memory(GiB)": 78.33,
      "step": 1725,
      "token_acc": 0.8871362252283732,
      "train_speed(iter/s)": 0.032512
    },
    {
      "epoch": 0.3344475124739621,
      "grad_norm": 0.1074642464518547,
      "learning_rate": 0.0002384475335556444,
      "loss": 0.3967469334602356,
      "memory(GiB)": 78.33,
      "step": 1726,
      "token_acc": 0.8839466666666667,
      "train_speed(iter/s)": 0.032514
    },
    {
      "epoch": 0.33464128275928884,
      "grad_norm": 0.11831879615783691,
      "learning_rate": 0.00023836987357600414,
      "loss": 0.45349621772766113,
      "memory(GiB)": 78.33,
      "step": 1727,
      "token_acc": 0.8706680645349176,
      "train_speed(iter/s)": 0.032515
    },
    {
      "epoch": 0.3348350530446156,
      "grad_norm": 0.11764407902956009,
      "learning_rate": 0.00023829217730048219,
      "loss": 0.4561113119125366,
      "memory(GiB)": 78.33,
      "step": 1728,
      "token_acc": 0.8694065038177228,
      "train_speed(iter/s)": 0.032517
    },
    {
      "epoch": 0.3350288233299423,
      "grad_norm": 0.1027785986661911,
      "learning_rate": 0.00023821444476099048,
      "loss": 0.38838091492652893,
      "memory(GiB)": 78.33,
      "step": 1729,
      "token_acc": 0.8851640132492787,
      "train_speed(iter/s)": 0.032518
    },
    {
      "epoch": 0.33522259361526907,
      "grad_norm": 0.1072501540184021,
      "learning_rate": 0.0002381366759894559,
      "loss": 0.3992714285850525,
      "memory(GiB)": 78.33,
      "step": 1730,
      "token_acc": 0.8841534590542398,
      "train_speed(iter/s)": 0.03252
    },
    {
      "epoch": 0.33541636390059587,
      "grad_norm": 0.09161798655986786,
      "learning_rate": 0.00023805887101782018,
      "loss": 0.3421012759208679,
      "memory(GiB)": 78.33,
      "step": 1731,
      "token_acc": 0.8976631008677287,
      "train_speed(iter/s)": 0.032521
    },
    {
      "epoch": 0.3356101341859226,
      "grad_norm": 0.12624742090702057,
      "learning_rate": 0.00023798102987803994,
      "loss": 0.47877341508865356,
      "memory(GiB)": 78.33,
      "step": 1732,
      "token_acc": 0.8629468940234756,
      "train_speed(iter/s)": 0.032523
    },
    {
      "epoch": 0.33580390447124936,
      "grad_norm": 0.11039040237665176,
      "learning_rate": 0.00023790315260208654,
      "loss": 0.373019278049469,
      "memory(GiB)": 78.33,
      "step": 1733,
      "token_acc": 0.8893133633557658,
      "train_speed(iter/s)": 0.032525
    },
    {
      "epoch": 0.3359976747565761,
      "grad_norm": 0.11160407960414886,
      "learning_rate": 0.0002378252392219463,
      "loss": 0.3813738524913788,
      "memory(GiB)": 78.33,
      "step": 1734,
      "token_acc": 0.887247745813654,
      "train_speed(iter/s)": 0.032526
    },
    {
      "epoch": 0.33619144504190285,
      "grad_norm": 0.11238150298595428,
      "learning_rate": 0.0002377472897696204,
      "loss": 0.41692203283309937,
      "memory(GiB)": 78.33,
      "step": 1735,
      "token_acc": 0.8769641013877738,
      "train_speed(iter/s)": 0.032528
    },
    {
      "epoch": 0.3363852153272296,
      "grad_norm": 0.1071707084774971,
      "learning_rate": 0.00023766930427712471,
      "loss": 0.39206287264823914,
      "memory(GiB)": 78.33,
      "step": 1736,
      "token_acc": 0.8842453737762436,
      "train_speed(iter/s)": 0.032529
    },
    {
      "epoch": 0.33657898561255634,
      "grad_norm": 0.10300777852535248,
      "learning_rate": 0.00023759128277649,
      "loss": 0.3877497613430023,
      "memory(GiB)": 78.33,
      "step": 1737,
      "token_acc": 0.8867694590155248,
      "train_speed(iter/s)": 0.032531
    },
    {
      "epoch": 0.3367727558978831,
      "grad_norm": 0.1066984236240387,
      "learning_rate": 0.0002375132252997618,
      "loss": 0.3994975686073303,
      "memory(GiB)": 78.33,
      "step": 1738,
      "token_acc": 0.8828811405205282,
      "train_speed(iter/s)": 0.032532
    },
    {
      "epoch": 0.3369665261832098,
      "grad_norm": 0.100070521235466,
      "learning_rate": 0.00023743513187900037,
      "loss": 0.36340388655662537,
      "memory(GiB)": 78.33,
      "step": 1739,
      "token_acc": 0.8943034295679132,
      "train_speed(iter/s)": 0.032534
    },
    {
      "epoch": 0.33716029646853657,
      "grad_norm": 0.10749375075101852,
      "learning_rate": 0.00023735700254628078,
      "loss": 0.3817083537578583,
      "memory(GiB)": 78.33,
      "step": 1740,
      "token_acc": 0.8865700752298691,
      "train_speed(iter/s)": 0.032535
    },
    {
      "epoch": 0.3373540667538633,
      "grad_norm": 0.1144825890660286,
      "learning_rate": 0.00023727883733369292,
      "loss": 0.3933083415031433,
      "memory(GiB)": 78.33,
      "step": 1741,
      "token_acc": 0.8860413411849537,
      "train_speed(iter/s)": 0.032537
    },
    {
      "epoch": 0.33754783703919006,
      "grad_norm": 0.11627618223428726,
      "learning_rate": 0.00023720063627334124,
      "loss": 0.38191571831703186,
      "memory(GiB)": 78.33,
      "step": 1742,
      "token_acc": 0.8882073834046985,
      "train_speed(iter/s)": 0.032538
    },
    {
      "epoch": 0.3377416073245168,
      "grad_norm": 0.10807470977306366,
      "learning_rate": 0.00023712239939734512,
      "loss": 0.3824315667152405,
      "memory(GiB)": 78.33,
      "step": 1743,
      "token_acc": 0.8895115896074527,
      "train_speed(iter/s)": 0.032539
    },
    {
      "epoch": 0.33793537760984355,
      "grad_norm": 0.11028466373682022,
      "learning_rate": 0.00023704412673783852,
      "loss": 0.4183183014392853,
      "memory(GiB)": 78.33,
      "step": 1744,
      "token_acc": 0.8795684778318642,
      "train_speed(iter/s)": 0.032541
    },
    {
      "epoch": 0.3381291478951703,
      "grad_norm": 0.10906746983528137,
      "learning_rate": 0.00023696581832697002,
      "loss": 0.38075122237205505,
      "memory(GiB)": 78.33,
      "step": 1745,
      "token_acc": 0.8904931905328574,
      "train_speed(iter/s)": 0.032542
    },
    {
      "epoch": 0.33832291818049703,
      "grad_norm": 0.10776349902153015,
      "learning_rate": 0.00023688747419690312,
      "loss": 0.39934486150741577,
      "memory(GiB)": 78.33,
      "step": 1746,
      "token_acc": 0.8850526197309921,
      "train_speed(iter/s)": 0.032544
    },
    {
      "epoch": 0.3385166884658238,
      "grad_norm": 0.12053883075714111,
      "learning_rate": 0.00023680909437981583,
      "loss": 0.4140859544277191,
      "memory(GiB)": 78.33,
      "step": 1747,
      "token_acc": 0.8816291216714198,
      "train_speed(iter/s)": 0.032546
    },
    {
      "epoch": 0.3387104587511505,
      "grad_norm": 0.12190855294466019,
      "learning_rate": 0.00023673067890790078,
      "loss": 0.46971094608306885,
      "memory(GiB)": 78.33,
      "step": 1748,
      "token_acc": 0.8653503654117791,
      "train_speed(iter/s)": 0.032547
    },
    {
      "epoch": 0.33890422903647727,
      "grad_norm": 0.10787336528301239,
      "learning_rate": 0.00023665222781336538,
      "loss": 0.3823373317718506,
      "memory(GiB)": 78.33,
      "step": 1749,
      "token_acc": 0.8879386276459724,
      "train_speed(iter/s)": 0.032549
    },
    {
      "epoch": 0.339097999321804,
      "grad_norm": 0.10566065460443497,
      "learning_rate": 0.0002365737411284316,
      "loss": 0.36630237102508545,
      "memory(GiB)": 78.33,
      "step": 1750,
      "token_acc": 0.8929974516882565,
      "train_speed(iter/s)": 0.03255
    },
    {
      "epoch": 0.33929176960713076,
      "grad_norm": 0.11304379254579544,
      "learning_rate": 0.000236495218885336,
      "loss": 0.41917577385902405,
      "memory(GiB)": 78.33,
      "step": 1751,
      "token_acc": 0.8782786241558034,
      "train_speed(iter/s)": 0.032551
    },
    {
      "epoch": 0.3394855398924575,
      "grad_norm": 0.1184999868273735,
      "learning_rate": 0.00023641666111632977,
      "loss": 0.436492919921875,
      "memory(GiB)": 78.33,
      "step": 1752,
      "token_acc": 0.8726875365530827,
      "train_speed(iter/s)": 0.032553
    },
    {
      "epoch": 0.33967931017778424,
      "grad_norm": 0.11077001690864563,
      "learning_rate": 0.00023633806785367873,
      "loss": 0.35720402002334595,
      "memory(GiB)": 78.33,
      "step": 1753,
      "token_acc": 0.8934273097826086,
      "train_speed(iter/s)": 0.032554
    },
    {
      "epoch": 0.339873080463111,
      "grad_norm": 0.10461857914924622,
      "learning_rate": 0.00023625943912966322,
      "loss": 0.366897851228714,
      "memory(GiB)": 78.33,
      "step": 1754,
      "token_acc": 0.8917701004031939,
      "train_speed(iter/s)": 0.032556
    },
    {
      "epoch": 0.34006685074843773,
      "grad_norm": 0.1258544921875,
      "learning_rate": 0.0002361807749765782,
      "loss": 0.43176522850990295,
      "memory(GiB)": 78.33,
      "step": 1755,
      "token_acc": 0.875682894257618,
      "train_speed(iter/s)": 0.032557
    },
    {
      "epoch": 0.3402606210337645,
      "grad_norm": 0.10822499543428421,
      "learning_rate": 0.0002361020754267331,
      "loss": 0.33935534954071045,
      "memory(GiB)": 78.33,
      "step": 1756,
      "token_acc": 0.9004844907278499,
      "train_speed(iter/s)": 0.032559
    },
    {
      "epoch": 0.3404543913190912,
      "grad_norm": 0.1091342493891716,
      "learning_rate": 0.00023602334051245195,
      "loss": 0.38166943192481995,
      "memory(GiB)": 78.33,
      "step": 1757,
      "token_acc": 0.8881519274376417,
      "train_speed(iter/s)": 0.03256
    },
    {
      "epoch": 0.34064816160441796,
      "grad_norm": 0.0942879170179367,
      "learning_rate": 0.00023594457026607335,
      "loss": 0.3533993363380432,
      "memory(GiB)": 78.33,
      "step": 1758,
      "token_acc": 0.8947904639000204,
      "train_speed(iter/s)": 0.032562
    },
    {
      "epoch": 0.3408419318897447,
      "grad_norm": 0.10075850039720535,
      "learning_rate": 0.00023586576471995035,
      "loss": 0.35853201150894165,
      "memory(GiB)": 78.33,
      "step": 1759,
      "token_acc": 0.8928562603474264,
      "train_speed(iter/s)": 0.032563
    },
    {
      "epoch": 0.34103570217507145,
      "grad_norm": 0.10983631759881973,
      "learning_rate": 0.00023578692390645043,
      "loss": 0.38611915707588196,
      "memory(GiB)": 78.33,
      "step": 1760,
      "token_acc": 0.8869917407878017,
      "train_speed(iter/s)": 0.032564
    },
    {
      "epoch": 0.3412294724603982,
      "grad_norm": 0.10747679322957993,
      "learning_rate": 0.00023570804785795572,
      "loss": 0.3835248649120331,
      "memory(GiB)": 78.33,
      "step": 1761,
      "token_acc": 0.8901472855851544,
      "train_speed(iter/s)": 0.032566
    },
    {
      "epoch": 0.34142324274572494,
      "grad_norm": 0.11863560974597931,
      "learning_rate": 0.00023562913660686263,
      "loss": 0.4406958222389221,
      "memory(GiB)": 78.33,
      "step": 1762,
      "token_acc": 0.8729787360066339,
      "train_speed(iter/s)": 0.032567
    },
    {
      "epoch": 0.3416170130310517,
      "grad_norm": 0.11419124901294708,
      "learning_rate": 0.00023555019018558224,
      "loss": 0.39518317580223083,
      "memory(GiB)": 78.33,
      "step": 1763,
      "token_acc": 0.88392173041787,
      "train_speed(iter/s)": 0.032569
    },
    {
      "epoch": 0.34181078331637843,
      "grad_norm": 0.11686435341835022,
      "learning_rate": 0.0002354712086265399,
      "loss": 0.43183597922325134,
      "memory(GiB)": 78.33,
      "step": 1764,
      "token_acc": 0.8757511860832894,
      "train_speed(iter/s)": 0.03257
    },
    {
      "epoch": 0.3420045536017052,
      "grad_norm": 0.10107649117708206,
      "learning_rate": 0.0002353921919621755,
      "loss": 0.39429807662963867,
      "memory(GiB)": 78.33,
      "step": 1765,
      "token_acc": 0.8837560239497639,
      "train_speed(iter/s)": 0.032572
    },
    {
      "epoch": 0.3421983238870319,
      "grad_norm": 0.10357075929641724,
      "learning_rate": 0.00023531314022494324,
      "loss": 0.3501340448856354,
      "memory(GiB)": 78.33,
      "step": 1766,
      "token_acc": 0.8969428540563897,
      "train_speed(iter/s)": 0.032573
    },
    {
      "epoch": 0.34239209417235866,
      "grad_norm": 0.0919223427772522,
      "learning_rate": 0.0002352340534473119,
      "loss": 0.3332497477531433,
      "memory(GiB)": 78.33,
      "step": 1767,
      "token_acc": 0.9006399880469158,
      "train_speed(iter/s)": 0.032574
    },
    {
      "epoch": 0.3425858644576854,
      "grad_norm": 0.10394643247127533,
      "learning_rate": 0.00023515493166176442,
      "loss": 0.36848586797714233,
      "memory(GiB)": 78.33,
      "step": 1768,
      "token_acc": 0.8926898258640502,
      "train_speed(iter/s)": 0.032576
    },
    {
      "epoch": 0.34277963474301215,
      "grad_norm": 0.10993330925703049,
      "learning_rate": 0.00023507577490079832,
      "loss": 0.41358712315559387,
      "memory(GiB)": 78.33,
      "step": 1769,
      "token_acc": 0.8809283276450512,
      "train_speed(iter/s)": 0.032577
    },
    {
      "epoch": 0.3429734050283389,
      "grad_norm": 0.10172808170318604,
      "learning_rate": 0.00023499658319692542,
      "loss": 0.3734714686870575,
      "memory(GiB)": 78.33,
      "step": 1770,
      "token_acc": 0.8903681089061836,
      "train_speed(iter/s)": 0.032579
    },
    {
      "epoch": 0.34316717531366564,
      "grad_norm": 0.10930664837360382,
      "learning_rate": 0.00023491735658267182,
      "loss": 0.37374699115753174,
      "memory(GiB)": 78.33,
      "step": 1771,
      "token_acc": 0.8907253336311352,
      "train_speed(iter/s)": 0.03258
    },
    {
      "epoch": 0.3433609455989924,
      "grad_norm": 0.1160641685128212,
      "learning_rate": 0.0002348380950905781,
      "loss": 0.3713539242744446,
      "memory(GiB)": 78.33,
      "step": 1772,
      "token_acc": 0.8907073847185395,
      "train_speed(iter/s)": 0.032582
    },
    {
      "epoch": 0.34355471588431913,
      "grad_norm": 0.11093118786811829,
      "learning_rate": 0.000234758798753199,
      "loss": 0.40102115273475647,
      "memory(GiB)": 78.33,
      "step": 1773,
      "token_acc": 0.8830998248686515,
      "train_speed(iter/s)": 0.032583
    },
    {
      "epoch": 0.3437484861696459,
      "grad_norm": 0.10876762866973877,
      "learning_rate": 0.00023467946760310368,
      "loss": 0.3644455671310425,
      "memory(GiB)": 78.33,
      "step": 1774,
      "token_acc": 0.8928700486448923,
      "train_speed(iter/s)": 0.032585
    },
    {
      "epoch": 0.3439422564549726,
      "grad_norm": 0.1010998785495758,
      "learning_rate": 0.00023460010167287564,
      "loss": 0.37988942861557007,
      "memory(GiB)": 78.33,
      "step": 1775,
      "token_acc": 0.8903876473385219,
      "train_speed(iter/s)": 0.032586
    },
    {
      "epoch": 0.34413602674029936,
      "grad_norm": 0.10760042816400528,
      "learning_rate": 0.00023452070099511249,
      "loss": 0.39640262722969055,
      "memory(GiB)": 78.33,
      "step": 1776,
      "token_acc": 0.8831872960402909,
      "train_speed(iter/s)": 0.032587
    },
    {
      "epoch": 0.3443297970256261,
      "grad_norm": 0.10678626596927643,
      "learning_rate": 0.00023444126560242634,
      "loss": 0.39018261432647705,
      "memory(GiB)": 78.33,
      "step": 1777,
      "token_acc": 0.8859104540432472,
      "train_speed(iter/s)": 0.032589
    },
    {
      "epoch": 0.34452356731095285,
      "grad_norm": 0.1159418597817421,
      "learning_rate": 0.00023436179552744333,
      "loss": 0.39666786789894104,
      "memory(GiB)": 78.33,
      "step": 1778,
      "token_acc": 0.8834541062801933,
      "train_speed(iter/s)": 0.03259
    },
    {
      "epoch": 0.3447173375962796,
      "grad_norm": 0.10813209414482117,
      "learning_rate": 0.00023428229080280403,
      "loss": 0.4084324836730957,
      "memory(GiB)": 78.33,
      "step": 1779,
      "token_acc": 0.8810328586468665,
      "train_speed(iter/s)": 0.032592
    },
    {
      "epoch": 0.34491110788160634,
      "grad_norm": 0.1118760034441948,
      "learning_rate": 0.00023420275146116318,
      "loss": 0.4295273721218109,
      "memory(GiB)": 78.33,
      "step": 1780,
      "token_acc": 0.876175413371675,
      "train_speed(iter/s)": 0.032593
    },
    {
      "epoch": 0.3451048781669331,
      "grad_norm": 0.10425037890672684,
      "learning_rate": 0.00023412317753518968,
      "loss": 0.38702112436294556,
      "memory(GiB)": 78.33,
      "step": 1781,
      "token_acc": 0.88680838382835,
      "train_speed(iter/s)": 0.032595
    },
    {
      "epoch": 0.3452986484522598,
      "grad_norm": 0.11203489452600479,
      "learning_rate": 0.0002340435690575666,
      "loss": 0.36375969648361206,
      "memory(GiB)": 78.33,
      "step": 1782,
      "token_acc": 0.8949717774449948,
      "train_speed(iter/s)": 0.032596
    },
    {
      "epoch": 0.34549241873758657,
      "grad_norm": 0.10534288734197617,
      "learning_rate": 0.00023396392606099144,
      "loss": 0.37565791606903076,
      "memory(GiB)": 78.33,
      "step": 1783,
      "token_acc": 0.8892402497364794,
      "train_speed(iter/s)": 0.032597
    },
    {
      "epoch": 0.3456861890229133,
      "grad_norm": 0.10718753933906555,
      "learning_rate": 0.00023388424857817566,
      "loss": 0.37750717997550964,
      "memory(GiB)": 78.33,
      "step": 1784,
      "token_acc": 0.8892705765022024,
      "train_speed(iter/s)": 0.032599
    },
    {
      "epoch": 0.34587995930824006,
      "grad_norm": 0.11478148400783539,
      "learning_rate": 0.00023380453664184492,
      "loss": 0.42205026745796204,
      "memory(GiB)": 78.33,
      "step": 1785,
      "token_acc": 0.8783005541671194,
      "train_speed(iter/s)": 0.0326
    },
    {
      "epoch": 0.3460737295935668,
      "grad_norm": 0.10276792198419571,
      "learning_rate": 0.00023372479028473908,
      "loss": 0.3618147671222687,
      "memory(GiB)": 78.33,
      "step": 1786,
      "token_acc": 0.8970774091627172,
      "train_speed(iter/s)": 0.032601
    },
    {
      "epoch": 0.34626749987889355,
      "grad_norm": 0.11334466934204102,
      "learning_rate": 0.0002336450095396121,
      "loss": 0.3954538404941559,
      "memory(GiB)": 78.33,
      "step": 1787,
      "token_acc": 0.8844070906454381,
      "train_speed(iter/s)": 0.032603
    },
    {
      "epoch": 0.3464612701642203,
      "grad_norm": 0.12592172622680664,
      "learning_rate": 0.00023356519443923205,
      "loss": 0.45259174704551697,
      "memory(GiB)": 78.33,
      "step": 1788,
      "token_acc": 0.8711917422583673,
      "train_speed(iter/s)": 0.032604
    },
    {
      "epoch": 0.34665504044954704,
      "grad_norm": 0.09646034240722656,
      "learning_rate": 0.00023348534501638115,
      "loss": 0.35588333010673523,
      "memory(GiB)": 78.33,
      "step": 1789,
      "token_acc": 0.8962561097320675,
      "train_speed(iter/s)": 0.032606
    },
    {
      "epoch": 0.3468488107348738,
      "grad_norm": 0.10078699886798859,
      "learning_rate": 0.00023340546130385574,
      "loss": 0.3622387647628784,
      "memory(GiB)": 78.33,
      "step": 1790,
      "token_acc": 0.8921391399616544,
      "train_speed(iter/s)": 0.032607
    },
    {
      "epoch": 0.3470425810202005,
      "grad_norm": 0.1054060235619545,
      "learning_rate": 0.00023332554333446617,
      "loss": 0.39420005679130554,
      "memory(GiB)": 78.33,
      "step": 1791,
      "token_acc": 0.8858397062477569,
      "train_speed(iter/s)": 0.032609
    },
    {
      "epoch": 0.3472363513055273,
      "grad_norm": 0.10570289194583893,
      "learning_rate": 0.0002332455911410369,
      "loss": 0.37943026423454285,
      "memory(GiB)": 78.33,
      "step": 1792,
      "token_acc": 0.8889375085464242,
      "train_speed(iter/s)": 0.03261
    },
    {
      "epoch": 0.34743012159085407,
      "grad_norm": 0.11169130355119705,
      "learning_rate": 0.00023316560475640646,
      "loss": 0.4090282917022705,
      "memory(GiB)": 78.33,
      "step": 1793,
      "token_acc": 0.8791112570459148,
      "train_speed(iter/s)": 0.032612
    },
    {
      "epoch": 0.3476238918761808,
      "grad_norm": 0.11017318069934845,
      "learning_rate": 0.00023308558421342743,
      "loss": 0.38704198598861694,
      "memory(GiB)": 78.33,
      "step": 1794,
      "token_acc": 0.8879233394591757,
      "train_speed(iter/s)": 0.032613
    },
    {
      "epoch": 0.34781766216150756,
      "grad_norm": 0.1121833398938179,
      "learning_rate": 0.0002330055295449663,
      "loss": 0.3916083872318268,
      "memory(GiB)": 78.33,
      "step": 1795,
      "token_acc": 0.8842932088924959,
      "train_speed(iter/s)": 0.032615
    },
    {
      "epoch": 0.3480114324468343,
      "grad_norm": 0.10421619564294815,
      "learning_rate": 0.00023292544078390377,
      "loss": 0.3865489065647125,
      "memory(GiB)": 78.33,
      "step": 1796,
      "token_acc": 0.8867510896758053,
      "train_speed(iter/s)": 0.032616
    },
    {
      "epoch": 0.34820520273216105,
      "grad_norm": 0.10179581493139267,
      "learning_rate": 0.00023284531796313444,
      "loss": 0.36844325065612793,
      "memory(GiB)": 78.33,
      "step": 1797,
      "token_acc": 0.8937036571572201,
      "train_speed(iter/s)": 0.032617
    },
    {
      "epoch": 0.3483989730174878,
      "grad_norm": 0.12043815106153488,
      "learning_rate": 0.0002327651611155669,
      "loss": 0.3939439058303833,
      "memory(GiB)": 78.33,
      "step": 1798,
      "token_acc": 0.8850653819683414,
      "train_speed(iter/s)": 0.032619
    },
    {
      "epoch": 0.34859274330281453,
      "grad_norm": 0.10927697271108627,
      "learning_rate": 0.00023268497027412364,
      "loss": 0.4002224802970886,
      "memory(GiB)": 78.33,
      "step": 1799,
      "token_acc": 0.8828930991564099,
      "train_speed(iter/s)": 0.03262
    },
    {
      "epoch": 0.3487865135881413,
      "grad_norm": 0.09782009571790695,
      "learning_rate": 0.0002326047454717413,
      "loss": 0.3829875886440277,
      "memory(GiB)": 78.33,
      "step": 1800,
      "token_acc": 0.8896661604720058,
      "train_speed(iter/s)": 0.032621
    },
    {
      "epoch": 0.348980283873468,
      "grad_norm": 0.09550854563713074,
      "learning_rate": 0.0002325244867413703,
      "loss": 0.34558382630348206,
      "memory(GiB)": 78.33,
      "step": 1801,
      "token_acc": 0.897237290273748,
      "train_speed(iter/s)": 0.032615
    },
    {
      "epoch": 0.34917405415879477,
      "grad_norm": 0.10129474103450775,
      "learning_rate": 0.00023244419411597508,
      "loss": 0.3574253022670746,
      "memory(GiB)": 78.33,
      "step": 1802,
      "token_acc": 0.8967041694242224,
      "train_speed(iter/s)": 0.032617
    },
    {
      "epoch": 0.3493678244441215,
      "grad_norm": 0.11037204414606094,
      "learning_rate": 0.00023236386762853398,
      "loss": 0.37563031911849976,
      "memory(GiB)": 78.33,
      "step": 1803,
      "token_acc": 0.889018691588785,
      "train_speed(iter/s)": 0.032618
    },
    {
      "epoch": 0.34956159472944826,
      "grad_norm": 0.10526623576879501,
      "learning_rate": 0.00023228350731203923,
      "loss": 0.37744981050491333,
      "memory(GiB)": 78.33,
      "step": 1804,
      "token_acc": 0.8889050705114202,
      "train_speed(iter/s)": 0.032619
    },
    {
      "epoch": 0.349755365014775,
      "grad_norm": 0.11370985209941864,
      "learning_rate": 0.000232203113199497,
      "loss": 0.38519221544265747,
      "memory(GiB)": 78.33,
      "step": 1805,
      "token_acc": 0.8884900839518189,
      "train_speed(iter/s)": 0.032621
    },
    {
      "epoch": 0.34994913530010174,
      "grad_norm": 0.11202775686979294,
      "learning_rate": 0.00023212268532392733,
      "loss": 0.40250468254089355,
      "memory(GiB)": 78.33,
      "step": 1806,
      "token_acc": 0.882124123298167,
      "train_speed(iter/s)": 0.032622
    },
    {
      "epoch": 0.3501429055854285,
      "grad_norm": 0.10065510869026184,
      "learning_rate": 0.00023204222371836405,
      "loss": 0.3592451810836792,
      "memory(GiB)": 78.33,
      "step": 1807,
      "token_acc": 0.8944679954711814,
      "train_speed(iter/s)": 0.032623
    },
    {
      "epoch": 0.35033667587075523,
      "grad_norm": 0.11026205122470856,
      "learning_rate": 0.00023196172841585488,
      "loss": 0.41680392622947693,
      "memory(GiB)": 78.33,
      "step": 1808,
      "token_acc": 0.8787629092507141,
      "train_speed(iter/s)": 0.032625
    },
    {
      "epoch": 0.350530446156082,
      "grad_norm": 0.10042080283164978,
      "learning_rate": 0.00023188119944946147,
      "loss": 0.36419713497161865,
      "memory(GiB)": 78.33,
      "step": 1809,
      "token_acc": 0.8931367912134112,
      "train_speed(iter/s)": 0.032626
    },
    {
      "epoch": 0.3507242164414087,
      "grad_norm": 0.12159177660942078,
      "learning_rate": 0.00023180063685225924,
      "loss": 0.3895268142223358,
      "memory(GiB)": 78.33,
      "step": 1810,
      "token_acc": 0.884988553363131,
      "train_speed(iter/s)": 0.032628
    },
    {
      "epoch": 0.35091798672673546,
      "grad_norm": 0.09737487882375717,
      "learning_rate": 0.0002317200406573374,
      "loss": 0.3568708002567291,
      "memory(GiB)": 78.33,
      "step": 1811,
      "token_acc": 0.8948162336884952,
      "train_speed(iter/s)": 0.032629
    },
    {
      "epoch": 0.3511117570120622,
      "grad_norm": 0.105777308344841,
      "learning_rate": 0.00023163941089779892,
      "loss": 0.3754398226737976,
      "memory(GiB)": 78.33,
      "step": 1812,
      "token_acc": 0.8915336122076238,
      "train_speed(iter/s)": 0.03263
    },
    {
      "epoch": 0.35130552729738895,
      "grad_norm": 0.10280515253543854,
      "learning_rate": 0.00023155874760676069,
      "loss": 0.3664350211620331,
      "memory(GiB)": 78.33,
      "step": 1813,
      "token_acc": 0.8921140983953261,
      "train_speed(iter/s)": 0.032632
    },
    {
      "epoch": 0.3514992975827157,
      "grad_norm": 0.10777980089187622,
      "learning_rate": 0.00023147805081735325,
      "loss": 0.3817192614078522,
      "memory(GiB)": 78.33,
      "step": 1814,
      "token_acc": 0.889120756816917,
      "train_speed(iter/s)": 0.032633
    },
    {
      "epoch": 0.35169306786804244,
      "grad_norm": 0.10182831436395645,
      "learning_rate": 0.0002313973205627209,
      "loss": 0.37246525287628174,
      "memory(GiB)": 78.33,
      "step": 1815,
      "token_acc": 0.8915448019143845,
      "train_speed(iter/s)": 0.032635
    },
    {
      "epoch": 0.3518868381533692,
      "grad_norm": 0.10677202045917511,
      "learning_rate": 0.00023131655687602174,
      "loss": 0.3982056677341461,
      "memory(GiB)": 78.33,
      "step": 1816,
      "token_acc": 0.8849781395220943,
      "train_speed(iter/s)": 0.032636
    },
    {
      "epoch": 0.35208060843869593,
      "grad_norm": 0.10674361884593964,
      "learning_rate": 0.00023123575979042767,
      "loss": 0.38551852107048035,
      "memory(GiB)": 78.33,
      "step": 1817,
      "token_acc": 0.8879332677063364,
      "train_speed(iter/s)": 0.032637
    },
    {
      "epoch": 0.3522743787240227,
      "grad_norm": 0.10658242553472519,
      "learning_rate": 0.00023115492933912412,
      "loss": 0.416787713766098,
      "memory(GiB)": 78.33,
      "step": 1818,
      "token_acc": 0.8809510899034811,
      "train_speed(iter/s)": 0.032639
    },
    {
      "epoch": 0.3524681490093494,
      "grad_norm": 0.10621768981218338,
      "learning_rate": 0.00023107406555531042,
      "loss": 0.37745460867881775,
      "memory(GiB)": 78.33,
      "step": 1819,
      "token_acc": 0.8909349857578331,
      "train_speed(iter/s)": 0.03264
    },
    {
      "epoch": 0.35266191929467616,
      "grad_norm": 0.11966609209775925,
      "learning_rate": 0.00023099316847219944,
      "loss": 0.43290236592292786,
      "memory(GiB)": 78.33,
      "step": 1820,
      "token_acc": 0.8743744060817231,
      "train_speed(iter/s)": 0.032642
    },
    {
      "epoch": 0.3528556895800029,
      "grad_norm": 0.11253046244382858,
      "learning_rate": 0.00023091223812301778,
      "loss": 0.4067782759666443,
      "memory(GiB)": 78.33,
      "step": 1821,
      "token_acc": 0.8815392109108622,
      "train_speed(iter/s)": 0.032643
    },
    {
      "epoch": 0.35304945986532965,
      "grad_norm": 0.09745576977729797,
      "learning_rate": 0.00023083127454100573,
      "loss": 0.36597940325737,
      "memory(GiB)": 78.33,
      "step": 1822,
      "token_acc": 0.8929222319310948,
      "train_speed(iter/s)": 0.032645
    },
    {
      "epoch": 0.3532432301506564,
      "grad_norm": 0.10735499858856201,
      "learning_rate": 0.00023075027775941722,
      "loss": 0.3935870826244354,
      "memory(GiB)": 78.33,
      "step": 1823,
      "token_acc": 0.8852148867679449,
      "train_speed(iter/s)": 0.032646
    },
    {
      "epoch": 0.35343700043598314,
      "grad_norm": 0.10999494791030884,
      "learning_rate": 0.00023066924781151976,
      "loss": 0.3992076814174652,
      "memory(GiB)": 78.33,
      "step": 1824,
      "token_acc": 0.8823376247649356,
      "train_speed(iter/s)": 0.032648
    },
    {
      "epoch": 0.3536307707213099,
      "grad_norm": 0.10191714018583298,
      "learning_rate": 0.00023058818473059456,
      "loss": 0.38821935653686523,
      "memory(GiB)": 78.33,
      "step": 1825,
      "token_acc": 0.8841636697432286,
      "train_speed(iter/s)": 0.032649
    },
    {
      "epoch": 0.35382454100663663,
      "grad_norm": 0.10946159809827805,
      "learning_rate": 0.00023050708854993645,
      "loss": 0.3874565660953522,
      "memory(GiB)": 78.33,
      "step": 1826,
      "token_acc": 0.8877764842840512,
      "train_speed(iter/s)": 0.03265
    },
    {
      "epoch": 0.3540183112919634,
      "grad_norm": 0.10883501917123795,
      "learning_rate": 0.00023042595930285374,
      "loss": 0.36587873101234436,
      "memory(GiB)": 78.33,
      "step": 1827,
      "token_acc": 0.8928685827436592,
      "train_speed(iter/s)": 0.032652
    },
    {
      "epoch": 0.3542120815772901,
      "grad_norm": 0.1008007600903511,
      "learning_rate": 0.0002303447970226684,
      "loss": 0.3552745282649994,
      "memory(GiB)": 78.33,
      "step": 1828,
      "token_acc": 0.8956754006614093,
      "train_speed(iter/s)": 0.032653
    },
    {
      "epoch": 0.35440585186261686,
      "grad_norm": 0.1125805526971817,
      "learning_rate": 0.00023026360174271593,
      "loss": 0.4129788279533386,
      "memory(GiB)": 78.33,
      "step": 1829,
      "token_acc": 0.8825399481301766,
      "train_speed(iter/s)": 0.032655
    },
    {
      "epoch": 0.3545996221479436,
      "grad_norm": 0.11501467972993851,
      "learning_rate": 0.00023018237349634553,
      "loss": 0.41495800018310547,
      "memory(GiB)": 78.33,
      "step": 1830,
      "token_acc": 0.8798725839542166,
      "train_speed(iter/s)": 0.032656
    },
    {
      "epoch": 0.35479339243327035,
      "grad_norm": 0.10721118748188019,
      "learning_rate": 0.00023010111231691973,
      "loss": 0.36478832364082336,
      "memory(GiB)": 78.33,
      "step": 1831,
      "token_acc": 0.891846109675567,
      "train_speed(iter/s)": 0.032657
    },
    {
      "epoch": 0.3549871627185971,
      "grad_norm": 0.11395079642534256,
      "learning_rate": 0.00023001981823781472,
      "loss": 0.3958333730697632,
      "memory(GiB)": 78.33,
      "step": 1832,
      "token_acc": 0.8860460306674152,
      "train_speed(iter/s)": 0.032659
    },
    {
      "epoch": 0.35518093300392384,
      "grad_norm": 0.11087855696678162,
      "learning_rate": 0.00022993849129242014,
      "loss": 0.4031826853752136,
      "memory(GiB)": 78.33,
      "step": 1833,
      "token_acc": 0.882696344865269,
      "train_speed(iter/s)": 0.03266
    },
    {
      "epoch": 0.3553747032892506,
      "grad_norm": 0.1058662161231041,
      "learning_rate": 0.00022985713151413913,
      "loss": 0.37713879346847534,
      "memory(GiB)": 78.33,
      "step": 1834,
      "token_acc": 0.8884106402887314,
      "train_speed(iter/s)": 0.032661
    },
    {
      "epoch": 0.3555684735745773,
      "grad_norm": 0.10403905063867569,
      "learning_rate": 0.00022977573893638836,
      "loss": 0.3826426863670349,
      "memory(GiB)": 78.33,
      "step": 1835,
      "token_acc": 0.8869446715726356,
      "train_speed(iter/s)": 0.032663
    },
    {
      "epoch": 0.35576224385990407,
      "grad_norm": 0.09630994498729706,
      "learning_rate": 0.00022969431359259797,
      "loss": 0.35036543011665344,
      "memory(GiB)": 78.33,
      "step": 1836,
      "token_acc": 0.8947007008903202,
      "train_speed(iter/s)": 0.032664
    },
    {
      "epoch": 0.3559560141452308,
      "grad_norm": 0.12186194956302643,
      "learning_rate": 0.0002296128555162115,
      "loss": 0.4338938593864441,
      "memory(GiB)": 78.33,
      "step": 1837,
      "token_acc": 0.8740203761755486,
      "train_speed(iter/s)": 0.032665
    },
    {
      "epoch": 0.35614978443055756,
      "grad_norm": 0.11005796492099762,
      "learning_rate": 0.000229531364740686,
      "loss": 0.37460899353027344,
      "memory(GiB)": 78.33,
      "step": 1838,
      "token_acc": 0.8911450261367517,
      "train_speed(iter/s)": 0.032667
    },
    {
      "epoch": 0.3563435547158843,
      "grad_norm": 0.10339885205030441,
      "learning_rate": 0.00022944984129949196,
      "loss": 0.3677971661090851,
      "memory(GiB)": 78.33,
      "step": 1839,
      "token_acc": 0.8935775621157461,
      "train_speed(iter/s)": 0.032668
    },
    {
      "epoch": 0.35653732500121105,
      "grad_norm": 0.10451404750347137,
      "learning_rate": 0.00022936828522611316,
      "loss": 0.3537764847278595,
      "memory(GiB)": 78.33,
      "step": 1840,
      "token_acc": 0.8961053422788898,
      "train_speed(iter/s)": 0.032669
    },
    {
      "epoch": 0.3567310952865378,
      "grad_norm": 0.09972133487462997,
      "learning_rate": 0.00022928669655404688,
      "loss": 0.39150598645210266,
      "memory(GiB)": 78.33,
      "step": 1841,
      "token_acc": 0.885431667414245,
      "train_speed(iter/s)": 0.032671
    },
    {
      "epoch": 0.35692486557186454,
      "grad_norm": 0.10803169012069702,
      "learning_rate": 0.0002292050753168038,
      "loss": 0.3714596629142761,
      "memory(GiB)": 78.33,
      "step": 1842,
      "token_acc": 0.8923362209736792,
      "train_speed(iter/s)": 0.032672
    },
    {
      "epoch": 0.3571186358571913,
      "grad_norm": 0.1168799102306366,
      "learning_rate": 0.00022912342154790804,
      "loss": 0.421195387840271,
      "memory(GiB)": 78.33,
      "step": 1843,
      "token_acc": 0.8766641213771124,
      "train_speed(iter/s)": 0.032674
    },
    {
      "epoch": 0.357312406142518,
      "grad_norm": 0.10394692420959473,
      "learning_rate": 0.00022904173528089686,
      "loss": 0.4081256091594696,
      "memory(GiB)": 78.33,
      "step": 1844,
      "token_acc": 0.8812434789587736,
      "train_speed(iter/s)": 0.032675
    },
    {
      "epoch": 0.35750617642784477,
      "grad_norm": 0.0959138348698616,
      "learning_rate": 0.00022896001654932105,
      "loss": 0.3497539162635803,
      "memory(GiB)": 78.33,
      "step": 1845,
      "token_acc": 0.8946740342957947,
      "train_speed(iter/s)": 0.032677
    },
    {
      "epoch": 0.3576999467131715,
      "grad_norm": 0.10984820127487183,
      "learning_rate": 0.0002288782653867448,
      "loss": 0.4027239680290222,
      "memory(GiB)": 78.33,
      "step": 1846,
      "token_acc": 0.8789472180150316,
      "train_speed(iter/s)": 0.032678
    },
    {
      "epoch": 0.35789371699849826,
      "grad_norm": 0.11082874238491058,
      "learning_rate": 0.0002287964818267453,
      "loss": 0.3869192600250244,
      "memory(GiB)": 78.33,
      "step": 1847,
      "token_acc": 0.8861776989530293,
      "train_speed(iter/s)": 0.032679
    },
    {
      "epoch": 0.358087487283825,
      "grad_norm": 0.10880383849143982,
      "learning_rate": 0.0002287146659029134,
      "loss": 0.39860600233078003,
      "memory(GiB)": 78.33,
      "step": 1848,
      "token_acc": 0.8812581533555588,
      "train_speed(iter/s)": 0.032681
    },
    {
      "epoch": 0.35828125756915175,
      "grad_norm": 0.11083029955625534,
      "learning_rate": 0.00022863281764885315,
      "loss": 0.3671538233757019,
      "memory(GiB)": 78.33,
      "step": 1849,
      "token_acc": 0.8913858792713418,
      "train_speed(iter/s)": 0.032682
    },
    {
      "epoch": 0.3584750278544785,
      "grad_norm": 0.1035129502415657,
      "learning_rate": 0.00022855093709818168,
      "loss": 0.35645681619644165,
      "memory(GiB)": 78.33,
      "step": 1850,
      "token_acc": 0.8927785347447856,
      "train_speed(iter/s)": 0.032683
    },
    {
      "epoch": 0.35866879813980523,
      "grad_norm": 0.11510751396417618,
      "learning_rate": 0.00022846902428452957,
      "loss": 0.4069887101650238,
      "memory(GiB)": 78.33,
      "step": 1851,
      "token_acc": 0.8827348409542743,
      "train_speed(iter/s)": 0.032685
    },
    {
      "epoch": 0.35886256842513203,
      "grad_norm": 0.11083865165710449,
      "learning_rate": 0.00022838707924154072,
      "loss": 0.3828306794166565,
      "memory(GiB)": 78.33,
      "step": 1852,
      "token_acc": 0.8873923603444469,
      "train_speed(iter/s)": 0.032686
    },
    {
      "epoch": 0.3590563387104588,
      "grad_norm": 0.11098314076662064,
      "learning_rate": 0.00022830510200287204,
      "loss": 0.4004696309566498,
      "memory(GiB)": 78.33,
      "step": 1853,
      "token_acc": 0.8826161182080414,
      "train_speed(iter/s)": 0.032688
    },
    {
      "epoch": 0.3592501089957855,
      "grad_norm": 0.1142255887389183,
      "learning_rate": 0.00022822309260219382,
      "loss": 0.39147108793258667,
      "memory(GiB)": 78.33,
      "step": 1854,
      "token_acc": 0.8852366362257191,
      "train_speed(iter/s)": 0.032689
    },
    {
      "epoch": 0.35944387928111227,
      "grad_norm": 0.10014226287603378,
      "learning_rate": 0.00022814105107318952,
      "loss": 0.35710304975509644,
      "memory(GiB)": 78.33,
      "step": 1855,
      "token_acc": 0.8953069555241322,
      "train_speed(iter/s)": 0.03269
    },
    {
      "epoch": 0.359637649566439,
      "grad_norm": 0.10325758904218674,
      "learning_rate": 0.00022805897744955587,
      "loss": 0.3418915569782257,
      "memory(GiB)": 78.33,
      "step": 1856,
      "token_acc": 0.899773435287454,
      "train_speed(iter/s)": 0.032691
    },
    {
      "epoch": 0.35983141985176575,
      "grad_norm": 0.10661393404006958,
      "learning_rate": 0.00022797687176500257,
      "loss": 0.36435335874557495,
      "memory(GiB)": 78.33,
      "step": 1857,
      "token_acc": 0.8942655906926346,
      "train_speed(iter/s)": 0.032693
    },
    {
      "epoch": 0.3600251901370925,
      "grad_norm": 0.11469036340713501,
      "learning_rate": 0.0002278947340532528,
      "loss": 0.415936678647995,
      "memory(GiB)": 78.33,
      "step": 1858,
      "token_acc": 0.8799646174259177,
      "train_speed(iter/s)": 0.032694
    },
    {
      "epoch": 0.36021896042241924,
      "grad_norm": 0.11351092904806137,
      "learning_rate": 0.0002278125643480426,
      "loss": 0.4251920282840729,
      "memory(GiB)": 78.33,
      "step": 1859,
      "token_acc": 0.8783027287232604,
      "train_speed(iter/s)": 0.032696
    },
    {
      "epoch": 0.360412730707746,
      "grad_norm": 0.12002420425415039,
      "learning_rate": 0.00022773036268312135,
      "loss": 0.40197598934173584,
      "memory(GiB)": 78.33,
      "step": 1860,
      "token_acc": 0.884971045690204,
      "train_speed(iter/s)": 0.032697
    },
    {
      "epoch": 0.36060650099307273,
      "grad_norm": 0.11968737095594406,
      "learning_rate": 0.00022764812909225143,
      "loss": 0.444561243057251,
      "memory(GiB)": 78.33,
      "step": 1861,
      "token_acc": 0.8720333353463373,
      "train_speed(iter/s)": 0.032698
    },
    {
      "epoch": 0.3608002712783995,
      "grad_norm": 0.12199822068214417,
      "learning_rate": 0.00022756586360920834,
      "loss": 0.4252670705318451,
      "memory(GiB)": 78.33,
      "step": 1862,
      "token_acc": 0.8757837872600405,
      "train_speed(iter/s)": 0.0327
    },
    {
      "epoch": 0.3609940415637262,
      "grad_norm": 0.11111042648553848,
      "learning_rate": 0.00022748356626778085,
      "loss": 0.38725805282592773,
      "memory(GiB)": 78.33,
      "step": 1863,
      "token_acc": 0.88624693011833,
      "train_speed(iter/s)": 0.032701
    },
    {
      "epoch": 0.36118781184905296,
      "grad_norm": 0.09851629287004471,
      "learning_rate": 0.00022740123710177063,
      "loss": 0.37134605646133423,
      "memory(GiB)": 78.33,
      "step": 1864,
      "token_acc": 0.8892687066811259,
      "train_speed(iter/s)": 0.032702
    },
    {
      "epoch": 0.3613815821343797,
      "grad_norm": 0.10534118860960007,
      "learning_rate": 0.0002273188761449925,
      "loss": 0.36640486121177673,
      "memory(GiB)": 78.33,
      "step": 1865,
      "token_acc": 0.8921657754010696,
      "train_speed(iter/s)": 0.032703
    },
    {
      "epoch": 0.36157535241970645,
      "grad_norm": 0.10792740434408188,
      "learning_rate": 0.00022723648343127428,
      "loss": 0.3896613121032715,
      "memory(GiB)": 78.33,
      "step": 1866,
      "token_acc": 0.8864207221350079,
      "train_speed(iter/s)": 0.032705
    },
    {
      "epoch": 0.3617691227050332,
      "grad_norm": 0.11643590033054352,
      "learning_rate": 0.0002271540589944569,
      "loss": 0.3730244040489197,
      "memory(GiB)": 78.33,
      "step": 1867,
      "token_acc": 0.8894352346164653,
      "train_speed(iter/s)": 0.032706
    },
    {
      "epoch": 0.36196289299035994,
      "grad_norm": 0.11337530612945557,
      "learning_rate": 0.00022707160286839425,
      "loss": 0.40268588066101074,
      "memory(GiB)": 78.33,
      "step": 1868,
      "token_acc": 0.8809530219410419,
      "train_speed(iter/s)": 0.032707
    },
    {
      "epoch": 0.3621566632756867,
      "grad_norm": 0.11132289469242096,
      "learning_rate": 0.00022698911508695335,
      "loss": 0.4088488221168518,
      "memory(GiB)": 78.33,
      "step": 1869,
      "token_acc": 0.8813299380751201,
      "train_speed(iter/s)": 0.032709
    },
    {
      "epoch": 0.36235043356101343,
      "grad_norm": 0.10401839762926102,
      "learning_rate": 0.00022690659568401405,
      "loss": 0.3792577087879181,
      "memory(GiB)": 78.33,
      "step": 1870,
      "token_acc": 0.8880251617357506,
      "train_speed(iter/s)": 0.03271
    },
    {
      "epoch": 0.3625442038463402,
      "grad_norm": 0.10719176381826401,
      "learning_rate": 0.0002268240446934694,
      "loss": 0.3922528028488159,
      "memory(GiB)": 78.33,
      "step": 1871,
      "token_acc": 0.8880327346157934,
      "train_speed(iter/s)": 0.032711
    },
    {
      "epoch": 0.3627379741316669,
      "grad_norm": 0.10341445356607437,
      "learning_rate": 0.00022674146214922522,
      "loss": 0.3721281886100769,
      "memory(GiB)": 78.33,
      "step": 1872,
      "token_acc": 0.8919034280378818,
      "train_speed(iter/s)": 0.032713
    },
    {
      "epoch": 0.36293174441699366,
      "grad_norm": 0.09941703081130981,
      "learning_rate": 0.00022665884808520045,
      "loss": 0.35231590270996094,
      "memory(GiB)": 78.33,
      "step": 1873,
      "token_acc": 0.895152484700744,
      "train_speed(iter/s)": 0.032714
    },
    {
      "epoch": 0.3631255147023204,
      "grad_norm": 0.12776227295398712,
      "learning_rate": 0.00022657620253532681,
      "loss": 0.42608171701431274,
      "memory(GiB)": 78.33,
      "step": 1874,
      "token_acc": 0.8764504054897068,
      "train_speed(iter/s)": 0.032715
    },
    {
      "epoch": 0.36331928498764715,
      "grad_norm": 0.10132501274347305,
      "learning_rate": 0.00022649352553354913,
      "loss": 0.3703892230987549,
      "memory(GiB)": 78.33,
      "step": 1875,
      "token_acc": 0.8915755231679516,
      "train_speed(iter/s)": 0.032717
    },
    {
      "epoch": 0.3635130552729739,
      "grad_norm": 0.11589387059211731,
      "learning_rate": 0.00022641081711382508,
      "loss": 0.4376241862773895,
      "memory(GiB)": 78.33,
      "step": 1876,
      "token_acc": 0.8734218545929473,
      "train_speed(iter/s)": 0.032718
    },
    {
      "epoch": 0.36370682555830064,
      "grad_norm": 0.10825785249471664,
      "learning_rate": 0.00022632807731012519,
      "loss": 0.3789633810520172,
      "memory(GiB)": 78.33,
      "step": 1877,
      "token_acc": 0.8895960154952961,
      "train_speed(iter/s)": 0.032719
    },
    {
      "epoch": 0.3639005958436274,
      "grad_norm": 0.10502450913190842,
      "learning_rate": 0.00022624530615643291,
      "loss": 0.3538724482059479,
      "memory(GiB)": 78.33,
      "step": 1878,
      "token_acc": 0.8984881209503239,
      "train_speed(iter/s)": 0.03272
    },
    {
      "epoch": 0.36409436612895413,
      "grad_norm": 0.10411059856414795,
      "learning_rate": 0.00022616250368674465,
      "loss": 0.35804876685142517,
      "memory(GiB)": 78.33,
      "step": 1879,
      "token_acc": 0.8947292874777516,
      "train_speed(iter/s)": 0.032722
    },
    {
      "epoch": 0.36428813641428087,
      "grad_norm": 0.09905564039945602,
      "learning_rate": 0.00022607966993506954,
      "loss": 0.36274391412734985,
      "memory(GiB)": 78.33,
      "step": 1880,
      "token_acc": 0.8931178818261462,
      "train_speed(iter/s)": 0.032723
    },
    {
      "epoch": 0.3644819066996076,
      "grad_norm": 0.11155527085065842,
      "learning_rate": 0.0002259968049354296,
      "loss": 0.4275517165660858,
      "memory(GiB)": 78.33,
      "step": 1881,
      "token_acc": 0.8751891922192948,
      "train_speed(iter/s)": 0.032724
    },
    {
      "epoch": 0.36467567698493436,
      "grad_norm": 0.12305615842342377,
      "learning_rate": 0.00022591390872185978,
      "loss": 0.4187135696411133,
      "memory(GiB)": 78.33,
      "step": 1882,
      "token_acc": 0.8782666539960088,
      "train_speed(iter/s)": 0.032726
    },
    {
      "epoch": 0.3648694472702611,
      "grad_norm": 0.2306802123785019,
      "learning_rate": 0.00022583098132840783,
      "loss": 0.40811386704444885,
      "memory(GiB)": 78.33,
      "step": 1883,
      "token_acc": 0.8822494609011463,
      "train_speed(iter/s)": 0.032727
    },
    {
      "epoch": 0.36506321755558785,
      "grad_norm": 0.11205735057592392,
      "learning_rate": 0.00022574802278913409,
      "loss": 0.3602101504802704,
      "memory(GiB)": 78.33,
      "step": 1884,
      "token_acc": 0.8919279519679787,
      "train_speed(iter/s)": 0.032728
    },
    {
      "epoch": 0.3652569878409146,
      "grad_norm": 0.10754162818193436,
      "learning_rate": 0.00022566503313811202,
      "loss": 0.41668108105659485,
      "memory(GiB)": 78.33,
      "step": 1885,
      "token_acc": 0.8790157211209843,
      "train_speed(iter/s)": 0.032729
    },
    {
      "epoch": 0.36545075812624134,
      "grad_norm": 0.13130834698677063,
      "learning_rate": 0.00022558201240942765,
      "loss": 0.3803432881832123,
      "memory(GiB)": 78.33,
      "step": 1886,
      "token_acc": 0.8907192443382269,
      "train_speed(iter/s)": 0.03273
    },
    {
      "epoch": 0.3656445284115681,
      "grad_norm": 0.12778477370738983,
      "learning_rate": 0.00022549896063717978,
      "loss": 0.45700541138648987,
      "memory(GiB)": 78.33,
      "step": 1887,
      "token_acc": 0.8673952641165756,
      "train_speed(iter/s)": 0.032732
    },
    {
      "epoch": 0.3658382986968948,
      "grad_norm": 0.10660150647163391,
      "learning_rate": 0.00022541587785548006,
      "loss": 0.388899028301239,
      "memory(GiB)": 78.33,
      "step": 1888,
      "token_acc": 0.8842618950793005,
      "train_speed(iter/s)": 0.032733
    },
    {
      "epoch": 0.36603206898222157,
      "grad_norm": 0.10151253640651703,
      "learning_rate": 0.0002253327640984528,
      "loss": 0.3622681200504303,
      "memory(GiB)": 78.33,
      "step": 1889,
      "token_acc": 0.8935134049603374,
      "train_speed(iter/s)": 0.032734
    },
    {
      "epoch": 0.3662258392675483,
      "grad_norm": 0.12292792648077011,
      "learning_rate": 0.00022524961940023505,
      "loss": 0.42959490418434143,
      "memory(GiB)": 78.33,
      "step": 1890,
      "token_acc": 0.8763984746777599,
      "train_speed(iter/s)": 0.032736
    },
    {
      "epoch": 0.36641960955287506,
      "grad_norm": 0.11249762028455734,
      "learning_rate": 0.00022516644379497658,
      "loss": 0.3739752471446991,
      "memory(GiB)": 78.33,
      "step": 1891,
      "token_acc": 0.8892815758980301,
      "train_speed(iter/s)": 0.032737
    },
    {
      "epoch": 0.3666133798382018,
      "grad_norm": 0.14134949445724487,
      "learning_rate": 0.00022508323731683984,
      "loss": 0.37695708870887756,
      "memory(GiB)": 78.33,
      "step": 1892,
      "token_acc": 0.8897990726429675,
      "train_speed(iter/s)": 0.032738
    },
    {
      "epoch": 0.36680715012352855,
      "grad_norm": 0.10767961293458939,
      "learning_rate": 0.000225,
      "loss": 0.3695901930332184,
      "memory(GiB)": 78.33,
      "step": 1893,
      "token_acc": 0.8925884180704907,
      "train_speed(iter/s)": 0.032739
    },
    {
      "epoch": 0.3670009204088553,
      "grad_norm": 0.10858670622110367,
      "learning_rate": 0.00022491673187864482,
      "loss": 0.39885541796684265,
      "memory(GiB)": 78.33,
      "step": 1894,
      "token_acc": 0.8855477140227064,
      "train_speed(iter/s)": 0.032741
    },
    {
      "epoch": 0.36719469069418204,
      "grad_norm": 0.12462179362773895,
      "learning_rate": 0.00022483343298697472,
      "loss": 0.38520297408103943,
      "memory(GiB)": 78.33,
      "step": 1895,
      "token_acc": 0.8870938651413259,
      "train_speed(iter/s)": 0.032742
    },
    {
      "epoch": 0.3673884609795088,
      "grad_norm": 0.09840144217014313,
      "learning_rate": 0.00022475010335920288,
      "loss": 0.35938745737075806,
      "memory(GiB)": 78.33,
      "step": 1896,
      "token_acc": 0.893057469002165,
      "train_speed(iter/s)": 0.032743
    },
    {
      "epoch": 0.3675822312648355,
      "grad_norm": 0.11222008615732193,
      "learning_rate": 0.00022466674302955495,
      "loss": 0.4017634689807892,
      "memory(GiB)": 78.33,
      "step": 1897,
      "token_acc": 0.8824197671116161,
      "train_speed(iter/s)": 0.032745
    },
    {
      "epoch": 0.36777600155016227,
      "grad_norm": 0.0982629731297493,
      "learning_rate": 0.00022458335203226932,
      "loss": 0.3599035441875458,
      "memory(GiB)": 78.33,
      "step": 1898,
      "token_acc": 0.8942642891179495,
      "train_speed(iter/s)": 0.032746
    },
    {
      "epoch": 0.367969771835489,
      "grad_norm": 0.09826286882162094,
      "learning_rate": 0.00022449993040159685,
      "loss": 0.35435935854911804,
      "memory(GiB)": 78.33,
      "step": 1899,
      "token_acc": 0.8933514973118933,
      "train_speed(iter/s)": 0.032747
    },
    {
      "epoch": 0.36816354212081576,
      "grad_norm": 0.11281174421310425,
      "learning_rate": 0.0002244164781718011,
      "loss": 0.3880941569805145,
      "memory(GiB)": 78.33,
      "step": 1900,
      "token_acc": 0.889101803692551,
      "train_speed(iter/s)": 0.032748
    },
    {
      "epoch": 0.3683573124061425,
      "grad_norm": 0.1215512827038765,
      "learning_rate": 0.0002243329953771581,
      "loss": 0.43437880277633667,
      "memory(GiB)": 78.33,
      "step": 1901,
      "token_acc": 0.875211292666291,
      "train_speed(iter/s)": 0.03275
    },
    {
      "epoch": 0.36855108269146925,
      "grad_norm": 0.11155978590250015,
      "learning_rate": 0.0002242494820519565,
      "loss": 0.37402305006980896,
      "memory(GiB)": 78.33,
      "step": 1902,
      "token_acc": 0.8897708186595872,
      "train_speed(iter/s)": 0.032751
    },
    {
      "epoch": 0.368744852976796,
      "grad_norm": 0.10142076760530472,
      "learning_rate": 0.00022416593823049746,
      "loss": 0.3670305609703064,
      "memory(GiB)": 78.33,
      "step": 1903,
      "token_acc": 0.8930082952429321,
      "train_speed(iter/s)": 0.032752
    },
    {
      "epoch": 0.36893862326212273,
      "grad_norm": 0.12013176828622818,
      "learning_rate": 0.00022408236394709464,
      "loss": 0.42799264192581177,
      "memory(GiB)": 78.33,
      "step": 1904,
      "token_acc": 0.8759844248616682,
      "train_speed(iter/s)": 0.032753
    },
    {
      "epoch": 0.3691323935474495,
      "grad_norm": 0.11704428493976593,
      "learning_rate": 0.0002239987592360743,
      "loss": 0.3853393793106079,
      "memory(GiB)": 78.33,
      "step": 1905,
      "token_acc": 0.8859077310659739,
      "train_speed(iter/s)": 0.032755
    },
    {
      "epoch": 0.3693261638327762,
      "grad_norm": 0.09620506316423416,
      "learning_rate": 0.00022391512413177516,
      "loss": 0.34337371587753296,
      "memory(GiB)": 78.33,
      "step": 1906,
      "token_acc": 0.8966349706853229,
      "train_speed(iter/s)": 0.032756
    },
    {
      "epoch": 0.36951993411810297,
      "grad_norm": 0.1018129214644432,
      "learning_rate": 0.00022383145866854834,
      "loss": 0.3677298128604889,
      "memory(GiB)": 78.33,
      "step": 1907,
      "token_acc": 0.8905968102763946,
      "train_speed(iter/s)": 0.032757
    },
    {
      "epoch": 0.3697137044034297,
      "grad_norm": 0.10136358439922333,
      "learning_rate": 0.00022374776288075745,
      "loss": 0.3641367554664612,
      "memory(GiB)": 78.33,
      "step": 1908,
      "token_acc": 0.8920720537349888,
      "train_speed(iter/s)": 0.032758
    },
    {
      "epoch": 0.36990747468875645,
      "grad_norm": 0.10752905905246735,
      "learning_rate": 0.00022366403680277875,
      "loss": 0.373902291059494,
      "memory(GiB)": 78.33,
      "step": 1909,
      "token_acc": 0.888663967611336,
      "train_speed(iter/s)": 0.032759
    },
    {
      "epoch": 0.3701012449740832,
      "grad_norm": 0.09833654016256332,
      "learning_rate": 0.00022358028046900067,
      "loss": 0.3543311059474945,
      "memory(GiB)": 78.33,
      "step": 1910,
      "token_acc": 0.89600593545196,
      "train_speed(iter/s)": 0.03276
    },
    {
      "epoch": 0.37029501525940994,
      "grad_norm": 0.10382269322872162,
      "learning_rate": 0.00022349649391382423,
      "loss": 0.39178794622421265,
      "memory(GiB)": 78.33,
      "step": 1911,
      "token_acc": 0.8857472274074888,
      "train_speed(iter/s)": 0.032762
    },
    {
      "epoch": 0.3704887855447367,
      "grad_norm": 0.11607307940721512,
      "learning_rate": 0.0002234126771716628,
      "loss": 0.39549145102500916,
      "memory(GiB)": 78.33,
      "step": 1912,
      "token_acc": 0.8873361227336123,
      "train_speed(iter/s)": 0.032763
    },
    {
      "epoch": 0.3706825558300635,
      "grad_norm": 0.098874531686306,
      "learning_rate": 0.0002233288302769422,
      "loss": 0.37269628047943115,
      "memory(GiB)": 78.33,
      "step": 1913,
      "token_acc": 0.8919546544147023,
      "train_speed(iter/s)": 0.032764
    },
    {
      "epoch": 0.37087632611539023,
      "grad_norm": 0.10173005610704422,
      "learning_rate": 0.00022324495326410057,
      "loss": 0.37610965967178345,
      "memory(GiB)": 78.33,
      "step": 1914,
      "token_acc": 0.8908233494774407,
      "train_speed(iter/s)": 0.032765
    },
    {
      "epoch": 0.371070096400717,
      "grad_norm": 0.12058980017900467,
      "learning_rate": 0.00022316104616758848,
      "loss": 0.43350180983543396,
      "memory(GiB)": 78.33,
      "step": 1915,
      "token_acc": 0.8748074872645422,
      "train_speed(iter/s)": 0.032767
    },
    {
      "epoch": 0.3712638666860437,
      "grad_norm": 0.1139712780714035,
      "learning_rate": 0.0002230771090218688,
      "loss": 0.3701034486293793,
      "memory(GiB)": 78.33,
      "step": 1916,
      "token_acc": 0.8913546669752908,
      "train_speed(iter/s)": 0.032768
    },
    {
      "epoch": 0.37145763697137046,
      "grad_norm": 0.11131946742534637,
      "learning_rate": 0.00022299314186141676,
      "loss": 0.37333056330680847,
      "memory(GiB)": 78.33,
      "step": 1917,
      "token_acc": 0.8892822835185947,
      "train_speed(iter/s)": 0.032769
    },
    {
      "epoch": 0.3716514072566972,
      "grad_norm": 0.12218397855758667,
      "learning_rate": 0.00022290914472072,
      "loss": 0.4149776101112366,
      "memory(GiB)": 78.33,
      "step": 1918,
      "token_acc": 0.8801127922136988,
      "train_speed(iter/s)": 0.03277
    },
    {
      "epoch": 0.37184517754202395,
      "grad_norm": 0.10744242370128632,
      "learning_rate": 0.00022282511763427838,
      "loss": 0.3630349338054657,
      "memory(GiB)": 78.33,
      "step": 1919,
      "token_acc": 0.8921786701935794,
      "train_speed(iter/s)": 0.032772
    },
    {
      "epoch": 0.3720389478273507,
      "grad_norm": 0.10889404267072678,
      "learning_rate": 0.00022274106063660404,
      "loss": 0.3649406433105469,
      "memory(GiB)": 78.33,
      "step": 1920,
      "token_acc": 0.8912664400768435,
      "train_speed(iter/s)": 0.032773
    },
    {
      "epoch": 0.37223271811267744,
      "grad_norm": 0.11465780436992645,
      "learning_rate": 0.00022265697376222141,
      "loss": 0.3944970965385437,
      "memory(GiB)": 78.33,
      "step": 1921,
      "token_acc": 0.8845351473922902,
      "train_speed(iter/s)": 0.032774
    },
    {
      "epoch": 0.3724264883980042,
      "grad_norm": 0.11644507199525833,
      "learning_rate": 0.00022257285704566735,
      "loss": 0.41171303391456604,
      "memory(GiB)": 78.33,
      "step": 1922,
      "token_acc": 0.8799286944436051,
      "train_speed(iter/s)": 0.032775
    },
    {
      "epoch": 0.37262025868333093,
      "grad_norm": 0.10672671347856522,
      "learning_rate": 0.00022248871052149078,
      "loss": 0.37590137124061584,
      "memory(GiB)": 78.33,
      "step": 1923,
      "token_acc": 0.8882163998831992,
      "train_speed(iter/s)": 0.032777
    },
    {
      "epoch": 0.3728140289686577,
      "grad_norm": 0.11380830407142639,
      "learning_rate": 0.00022240453422425294,
      "loss": 0.41342946887016296,
      "memory(GiB)": 78.33,
      "step": 1924,
      "token_acc": 0.8813163889880054,
      "train_speed(iter/s)": 0.032778
    },
    {
      "epoch": 0.3730077992539844,
      "grad_norm": 0.10751762241125107,
      "learning_rate": 0.00022232032818852732,
      "loss": 0.37841910123825073,
      "memory(GiB)": 78.33,
      "step": 1925,
      "token_acc": 0.887600209372159,
      "train_speed(iter/s)": 0.032779
    },
    {
      "epoch": 0.37320156953931116,
      "grad_norm": 0.11165986955165863,
      "learning_rate": 0.0002222360924488996,
      "loss": 0.39904457330703735,
      "memory(GiB)": 78.33,
      "step": 1926,
      "token_acc": 0.8864683010230814,
      "train_speed(iter/s)": 0.03278
    },
    {
      "epoch": 0.3733953398246379,
      "grad_norm": 0.10432970523834229,
      "learning_rate": 0.00022215182703996765,
      "loss": 0.35338613390922546,
      "memory(GiB)": 78.33,
      "step": 1927,
      "token_acc": 0.8975270862625093,
      "train_speed(iter/s)": 0.032782
    },
    {
      "epoch": 0.37358911010996465,
      "grad_norm": 0.09983038157224655,
      "learning_rate": 0.00022206753199634148,
      "loss": 0.36258718371391296,
      "memory(GiB)": 78.33,
      "step": 1928,
      "token_acc": 0.8939695669934641,
      "train_speed(iter/s)": 0.032783
    },
    {
      "epoch": 0.3737828803952914,
      "grad_norm": 0.10431524366140366,
      "learning_rate": 0.00022198320735264344,
      "loss": 0.3931850492954254,
      "memory(GiB)": 78.33,
      "step": 1929,
      "token_acc": 0.8851347908067064,
      "train_speed(iter/s)": 0.032784
    },
    {
      "epoch": 0.37397665068061814,
      "grad_norm": 0.11230375617742538,
      "learning_rate": 0.00022189885314350787,
      "loss": 0.391851007938385,
      "memory(GiB)": 78.33,
      "step": 1930,
      "token_acc": 0.8859056476850633,
      "train_speed(iter/s)": 0.032785
    },
    {
      "epoch": 0.3741704209659449,
      "grad_norm": 0.1120762825012207,
      "learning_rate": 0.00022181446940358135,
      "loss": 0.41800612211227417,
      "memory(GiB)": 78.33,
      "step": 1931,
      "token_acc": 0.877407731234931,
      "train_speed(iter/s)": 0.032787
    },
    {
      "epoch": 0.3743641912512716,
      "grad_norm": 0.1048278734087944,
      "learning_rate": 0.00022173005616752252,
      "loss": 0.3735467791557312,
      "memory(GiB)": 78.33,
      "step": 1932,
      "token_acc": 0.8902271252433485,
      "train_speed(iter/s)": 0.032788
    },
    {
      "epoch": 0.37455796153659837,
      "grad_norm": 0.12189824134111404,
      "learning_rate": 0.00022164561347000212,
      "loss": 0.3999292254447937,
      "memory(GiB)": 78.33,
      "step": 1933,
      "token_acc": 0.8858239307268434,
      "train_speed(iter/s)": 0.032789
    },
    {
      "epoch": 0.3747517318219251,
      "grad_norm": 0.10540164262056351,
      "learning_rate": 0.00022156114134570305,
      "loss": 0.39689502120018005,
      "memory(GiB)": 78.33,
      "step": 1934,
      "token_acc": 0.8854811041222644,
      "train_speed(iter/s)": 0.03279
    },
    {
      "epoch": 0.37494550210725186,
      "grad_norm": 0.10333568602800369,
      "learning_rate": 0.00022147663982932038,
      "loss": 0.3748435080051422,
      "memory(GiB)": 78.33,
      "step": 1935,
      "token_acc": 0.8905851118145615,
      "train_speed(iter/s)": 0.032791
    },
    {
      "epoch": 0.3751392723925786,
      "grad_norm": 0.1070760041475296,
      "learning_rate": 0.00022139210895556104,
      "loss": 0.37235966324806213,
      "memory(GiB)": 78.33,
      "step": 1936,
      "token_acc": 0.8913205876656614,
      "train_speed(iter/s)": 0.032792
    },
    {
      "epoch": 0.37533304267790535,
      "grad_norm": 0.10744938999414444,
      "learning_rate": 0.00022130754875914415,
      "loss": 0.36817583441734314,
      "memory(GiB)": 78.33,
      "step": 1937,
      "token_acc": 0.8937919693024531,
      "train_speed(iter/s)": 0.032794
    },
    {
      "epoch": 0.3755268129632321,
      "grad_norm": 0.09824454039335251,
      "learning_rate": 0.0002212229592748009,
      "loss": 0.34159597754478455,
      "memory(GiB)": 78.33,
      "step": 1938,
      "token_acc": 0.8967411946420885,
      "train_speed(iter/s)": 0.032795
    },
    {
      "epoch": 0.37572058324855884,
      "grad_norm": 0.10819046944379807,
      "learning_rate": 0.00022113834053727444,
      "loss": 0.3992760181427002,
      "memory(GiB)": 78.33,
      "step": 1939,
      "token_acc": 0.8845097429519071,
      "train_speed(iter/s)": 0.032796
    },
    {
      "epoch": 0.3759143535338856,
      "grad_norm": 0.09654027968645096,
      "learning_rate": 0.00022105369258131998,
      "loss": 0.340084046125412,
      "memory(GiB)": 78.33,
      "step": 1940,
      "token_acc": 0.9001615206348463,
      "train_speed(iter/s)": 0.032797
    },
    {
      "epoch": 0.3761081238192123,
      "grad_norm": 0.09880183637142181,
      "learning_rate": 0.00022096901544170467,
      "loss": 0.36598512530326843,
      "memory(GiB)": 78.33,
      "step": 1941,
      "token_acc": 0.8919111291880625,
      "train_speed(iter/s)": 0.032798
    },
    {
      "epoch": 0.37630189410453907,
      "grad_norm": 0.10682184994220734,
      "learning_rate": 0.0002208843091532077,
      "loss": 0.39188694953918457,
      "memory(GiB)": 78.33,
      "step": 1942,
      "token_acc": 0.8864231527226587,
      "train_speed(iter/s)": 0.0328
    },
    {
      "epoch": 0.3764956643898658,
      "grad_norm": 0.11017350852489471,
      "learning_rate": 0.00022079957375062021,
      "loss": 0.40813326835632324,
      "memory(GiB)": 78.33,
      "step": 1943,
      "token_acc": 0.8828356812531744,
      "train_speed(iter/s)": 0.032801
    },
    {
      "epoch": 0.37668943467519256,
      "grad_norm": 0.11248282343149185,
      "learning_rate": 0.00022071480926874536,
      "loss": 0.41364431381225586,
      "memory(GiB)": 78.33,
      "step": 1944,
      "token_acc": 0.8817496143077539,
      "train_speed(iter/s)": 0.032802
    },
    {
      "epoch": 0.3768832049605193,
      "grad_norm": 0.09806734323501587,
      "learning_rate": 0.00022063001574239814,
      "loss": 0.36278462409973145,
      "memory(GiB)": 78.33,
      "step": 1945,
      "token_acc": 0.8927522524764797,
      "train_speed(iter/s)": 0.032803
    },
    {
      "epoch": 0.37707697524584605,
      "grad_norm": 0.10731010138988495,
      "learning_rate": 0.00022054519320640557,
      "loss": 0.39958709478378296,
      "memory(GiB)": 78.33,
      "step": 1946,
      "token_acc": 0.882800608828006,
      "train_speed(iter/s)": 0.032804
    },
    {
      "epoch": 0.3772707455311728,
      "grad_norm": 0.1078319326043129,
      "learning_rate": 0.0002204603416956065,
      "loss": 0.39028337597846985,
      "memory(GiB)": 78.33,
      "step": 1947,
      "token_acc": 0.8870954252738982,
      "train_speed(iter/s)": 0.032805
    },
    {
      "epoch": 0.37746451581649954,
      "grad_norm": 0.10992579162120819,
      "learning_rate": 0.00022037546124485178,
      "loss": 0.3915446102619171,
      "memory(GiB)": 78.33,
      "step": 1948,
      "token_acc": 0.8861552284559077,
      "train_speed(iter/s)": 0.032807
    },
    {
      "epoch": 0.3776582861018263,
      "grad_norm": 0.10269248485565186,
      "learning_rate": 0.00022029055188900405,
      "loss": 0.35439544916152954,
      "memory(GiB)": 78.33,
      "step": 1949,
      "token_acc": 0.894891822424076,
      "train_speed(iter/s)": 0.032808
    },
    {
      "epoch": 0.377852056387153,
      "grad_norm": 0.11859910935163498,
      "learning_rate": 0.00022020561366293789,
      "loss": 0.36357784271240234,
      "memory(GiB)": 78.33,
      "step": 1950,
      "token_acc": 0.8931177855959062,
      "train_speed(iter/s)": 0.032809
    },
    {
      "epoch": 0.37804582667247977,
      "grad_norm": 0.11494617909193039,
      "learning_rate": 0.0002201206466015397,
      "loss": 0.4254799485206604,
      "memory(GiB)": 78.33,
      "step": 1951,
      "token_acc": 0.8750034286968209,
      "train_speed(iter/s)": 0.03281
    },
    {
      "epoch": 0.3782395969578065,
      "grad_norm": 0.09963639825582504,
      "learning_rate": 0.00022003565073970774,
      "loss": 0.3571300506591797,
      "memory(GiB)": 78.33,
      "step": 1952,
      "token_acc": 0.8945331269019273,
      "train_speed(iter/s)": 0.032812
    },
    {
      "epoch": 0.37843336724313326,
      "grad_norm": 0.10854914784431458,
      "learning_rate": 0.0002199506261123521,
      "loss": 0.3844650089740753,
      "memory(GiB)": 78.33,
      "step": 1953,
      "token_acc": 0.8875416461849364,
      "train_speed(iter/s)": 0.032813
    },
    {
      "epoch": 0.37862713752846,
      "grad_norm": 0.11779830604791641,
      "learning_rate": 0.00021986557275439464,
      "loss": 0.40272995829582214,
      "memory(GiB)": 78.33,
      "step": 1954,
      "token_acc": 0.8847069242264903,
      "train_speed(iter/s)": 0.032814
    },
    {
      "epoch": 0.37882090781378674,
      "grad_norm": 0.09850434213876724,
      "learning_rate": 0.00021978049070076912,
      "loss": 0.3412163257598877,
      "memory(GiB)": 78.33,
      "step": 1955,
      "token_acc": 0.8973165531228873,
      "train_speed(iter/s)": 0.032816
    },
    {
      "epoch": 0.3790146780991135,
      "grad_norm": 0.10487658530473709,
      "learning_rate": 0.00021969537998642097,
      "loss": 0.37174132466316223,
      "memory(GiB)": 78.33,
      "step": 1956,
      "token_acc": 0.8908394592093765,
      "train_speed(iter/s)": 0.032817
    },
    {
      "epoch": 0.37920844838444023,
      "grad_norm": 0.10610850155353546,
      "learning_rate": 0.00021961024064630745,
      "loss": 0.3750998079776764,
      "memory(GiB)": 78.33,
      "step": 1957,
      "token_acc": 0.8877445652173913,
      "train_speed(iter/s)": 0.032818
    },
    {
      "epoch": 0.379402218669767,
      "grad_norm": 0.11280324310064316,
      "learning_rate": 0.00021952507271539762,
      "loss": 0.40227746963500977,
      "memory(GiB)": 78.33,
      "step": 1958,
      "token_acc": 0.8827386807356404,
      "train_speed(iter/s)": 0.032819
    },
    {
      "epoch": 0.3795959889550937,
      "grad_norm": 0.11210876703262329,
      "learning_rate": 0.00021943987622867223,
      "loss": 0.3853500187397003,
      "memory(GiB)": 78.33,
      "step": 1959,
      "token_acc": 0.8886800763470856,
      "train_speed(iter/s)": 0.03282
    },
    {
      "epoch": 0.37978975924042047,
      "grad_norm": 0.10147024691104889,
      "learning_rate": 0.00021935465122112377,
      "loss": 0.34372827410697937,
      "memory(GiB)": 78.33,
      "step": 1960,
      "token_acc": 0.8980276961812841,
      "train_speed(iter/s)": 0.032821
    },
    {
      "epoch": 0.3799835295257472,
      "grad_norm": 0.1177581325173378,
      "learning_rate": 0.00021926939772775637,
      "loss": 0.4162188172340393,
      "memory(GiB)": 78.33,
      "step": 1961,
      "token_acc": 0.8789121688698469,
      "train_speed(iter/s)": 0.032823
    },
    {
      "epoch": 0.38017729981107395,
      "grad_norm": 0.10330435633659363,
      "learning_rate": 0.00021918411578358601,
      "loss": 0.3811953067779541,
      "memory(GiB)": 78.33,
      "step": 1962,
      "token_acc": 0.8878933276780744,
      "train_speed(iter/s)": 0.032824
    },
    {
      "epoch": 0.3803710700964007,
      "grad_norm": 0.1050238385796547,
      "learning_rate": 0.0002190988054236402,
      "loss": 0.39693543314933777,
      "memory(GiB)": 78.33,
      "step": 1963,
      "token_acc": 0.8836019246925838,
      "train_speed(iter/s)": 0.032825
    },
    {
      "epoch": 0.38056484038172744,
      "grad_norm": 0.1056143268942833,
      "learning_rate": 0.0002190134666829583,
      "loss": 0.3753167390823364,
      "memory(GiB)": 78.33,
      "step": 1964,
      "token_acc": 0.8911448414921026,
      "train_speed(iter/s)": 0.032826
    },
    {
      "epoch": 0.3807586106670542,
      "grad_norm": 0.1035081148147583,
      "learning_rate": 0.0002189280995965912,
      "loss": 0.3400823771953583,
      "memory(GiB)": 78.33,
      "step": 1965,
      "token_acc": 0.8973672105419812,
      "train_speed(iter/s)": 0.032828
    },
    {
      "epoch": 0.38095238095238093,
      "grad_norm": 0.11154508590698242,
      "learning_rate": 0.00021884270419960137,
      "loss": 0.37742310762405396,
      "memory(GiB)": 78.33,
      "step": 1966,
      "token_acc": 0.8898611151892416,
      "train_speed(iter/s)": 0.032829
    },
    {
      "epoch": 0.3811461512377077,
      "grad_norm": 0.10220933705568314,
      "learning_rate": 0.00021875728052706304,
      "loss": 0.3489319086074829,
      "memory(GiB)": 78.33,
      "step": 1967,
      "token_acc": 0.8957410562180579,
      "train_speed(iter/s)": 0.03283
    },
    {
      "epoch": 0.3813399215230344,
      "grad_norm": 0.12063480913639069,
      "learning_rate": 0.00021867182861406206,
      "loss": 0.40946558117866516,
      "memory(GiB)": 78.33,
      "step": 1968,
      "token_acc": 0.8786677692548975,
      "train_speed(iter/s)": 0.032831
    },
    {
      "epoch": 0.38153369180836116,
      "grad_norm": 0.1056455671787262,
      "learning_rate": 0.00021858634849569576,
      "loss": 0.38398033380508423,
      "memory(GiB)": 78.33,
      "step": 1969,
      "token_acc": 0.8870220312216843,
      "train_speed(iter/s)": 0.032832
    },
    {
      "epoch": 0.3817274620936879,
      "grad_norm": 0.10487000644207001,
      "learning_rate": 0.00021850084020707316,
      "loss": 0.39127257466316223,
      "memory(GiB)": 78.33,
      "step": 1970,
      "token_acc": 0.8856672733165466,
      "train_speed(iter/s)": 0.032833
    },
    {
      "epoch": 0.38192123237901465,
      "grad_norm": 0.10356573760509491,
      "learning_rate": 0.0002184153037833148,
      "loss": 0.39030513167381287,
      "memory(GiB)": 78.33,
      "step": 1971,
      "token_acc": 0.8840334008097166,
      "train_speed(iter/s)": 0.032835
    },
    {
      "epoch": 0.3821150026643414,
      "grad_norm": 0.10742707550525665,
      "learning_rate": 0.0002183297392595528,
      "loss": 0.3961334824562073,
      "memory(GiB)": 78.33,
      "step": 1972,
      "token_acc": 0.8832247032291641,
      "train_speed(iter/s)": 0.032836
    },
    {
      "epoch": 0.38230877294966814,
      "grad_norm": 0.10838883370161057,
      "learning_rate": 0.00021824414667093075,
      "loss": 0.3667401075363159,
      "memory(GiB)": 78.33,
      "step": 1973,
      "token_acc": 0.8917944849551588,
      "train_speed(iter/s)": 0.032837
    },
    {
      "epoch": 0.38250254323499494,
      "grad_norm": 0.1165713295340538,
      "learning_rate": 0.00021815852605260386,
      "loss": 0.3943071663379669,
      "memory(GiB)": 78.33,
      "step": 1974,
      "token_acc": 0.8838169344539811,
      "train_speed(iter/s)": 0.032838
    },
    {
      "epoch": 0.3826963135203217,
      "grad_norm": 0.10990098863840103,
      "learning_rate": 0.0002180728774397389,
      "loss": 0.4022403061389923,
      "memory(GiB)": 78.33,
      "step": 1975,
      "token_acc": 0.8827039398397908,
      "train_speed(iter/s)": 0.032839
    },
    {
      "epoch": 0.38289008380564843,
      "grad_norm": 0.12068414688110352,
      "learning_rate": 0.00021798720086751395,
      "loss": 0.431316614151001,
      "memory(GiB)": 78.33,
      "step": 1976,
      "token_acc": 0.8732160741111668,
      "train_speed(iter/s)": 0.032841
    },
    {
      "epoch": 0.3830838540909752,
      "grad_norm": 0.09716933965682983,
      "learning_rate": 0.0002179014963711187,
      "loss": 0.35583171248435974,
      "memory(GiB)": 78.33,
      "step": 1977,
      "token_acc": 0.8953317742009447,
      "train_speed(iter/s)": 0.032842
    },
    {
      "epoch": 0.3832776243763019,
      "grad_norm": 0.11288797110319138,
      "learning_rate": 0.00021781576398575433,
      "loss": 0.392235666513443,
      "memory(GiB)": 78.33,
      "step": 1978,
      "token_acc": 0.885842526497945,
      "train_speed(iter/s)": 0.032843
    },
    {
      "epoch": 0.38347139466162866,
      "grad_norm": 0.10276081413030624,
      "learning_rate": 0.0002177300037466334,
      "loss": 0.37976735830307007,
      "memory(GiB)": 78.33,
      "step": 1979,
      "token_acc": 0.8875089992800576,
      "train_speed(iter/s)": 0.032844
    },
    {
      "epoch": 0.3836651649469554,
      "grad_norm": 0.11369860172271729,
      "learning_rate": 0.00021764421568897993,
      "loss": 0.3621234893798828,
      "memory(GiB)": 78.33,
      "step": 1980,
      "token_acc": 0.890562048175558,
      "train_speed(iter/s)": 0.032845
    },
    {
      "epoch": 0.38385893523228215,
      "grad_norm": 0.10219123214483261,
      "learning_rate": 0.00021755839984802944,
      "loss": 0.3366636335849762,
      "memory(GiB)": 78.33,
      "step": 1981,
      "token_acc": 0.900472891235016,
      "train_speed(iter/s)": 0.032847
    },
    {
      "epoch": 0.3840527055176089,
      "grad_norm": 0.10489702969789505,
      "learning_rate": 0.0002174725562590288,
      "loss": 0.37687474489212036,
      "memory(GiB)": 78.33,
      "step": 1982,
      "token_acc": 0.8895873939660143,
      "train_speed(iter/s)": 0.032848
    },
    {
      "epoch": 0.38424647580293564,
      "grad_norm": 0.09669921547174454,
      "learning_rate": 0.00021738668495723616,
      "loss": 0.3477326035499573,
      "memory(GiB)": 78.33,
      "step": 1983,
      "token_acc": 0.8987814906182253,
      "train_speed(iter/s)": 0.032849
    },
    {
      "epoch": 0.3844402460882624,
      "grad_norm": 0.11556685715913773,
      "learning_rate": 0.0002173007859779213,
      "loss": 0.4000082015991211,
      "memory(GiB)": 78.33,
      "step": 1984,
      "token_acc": 0.8822343958445688,
      "train_speed(iter/s)": 0.03285
    },
    {
      "epoch": 0.3846340163735891,
      "grad_norm": 0.10442518442869186,
      "learning_rate": 0.00021721485935636523,
      "loss": 0.3522685170173645,
      "memory(GiB)": 78.33,
      "step": 1985,
      "token_acc": 0.8955117718187637,
      "train_speed(iter/s)": 0.032851
    },
    {
      "epoch": 0.38482778665891587,
      "grad_norm": 0.11232437193393707,
      "learning_rate": 0.00021712890512786027,
      "loss": 0.39102703332901,
      "memory(GiB)": 78.33,
      "step": 1986,
      "token_acc": 0.8859773174722014,
      "train_speed(iter/s)": 0.032852
    },
    {
      "epoch": 0.3850215569442426,
      "grad_norm": 0.1146976500749588,
      "learning_rate": 0.00021704292332771013,
      "loss": 0.4080618619918823,
      "memory(GiB)": 78.33,
      "step": 1987,
      "token_acc": 0.8815463453091564,
      "train_speed(iter/s)": 0.032853
    },
    {
      "epoch": 0.38521532722956936,
      "grad_norm": 0.10926152765750885,
      "learning_rate": 0.00021695691399122987,
      "loss": 0.3856127858161926,
      "memory(GiB)": 78.33,
      "step": 1988,
      "token_acc": 0.8882630577068502,
      "train_speed(iter/s)": 0.032855
    },
    {
      "epoch": 0.3854090975148961,
      "grad_norm": 0.11085885018110275,
      "learning_rate": 0.00021687087715374585,
      "loss": 0.37670475244522095,
      "memory(GiB)": 78.33,
      "step": 1989,
      "token_acc": 0.8898128898128899,
      "train_speed(iter/s)": 0.032856
    },
    {
      "epoch": 0.38560286780022285,
      "grad_norm": 0.12100964039564133,
      "learning_rate": 0.00021678481285059567,
      "loss": 0.43839654326438904,
      "memory(GiB)": 78.33,
      "step": 1990,
      "token_acc": 0.8723092599668832,
      "train_speed(iter/s)": 0.032857
    },
    {
      "epoch": 0.3857966380855496,
      "grad_norm": 0.12123683094978333,
      "learning_rate": 0.00021669872111712828,
      "loss": 0.44955089688301086,
      "memory(GiB)": 78.33,
      "step": 1991,
      "token_acc": 0.871977240398293,
      "train_speed(iter/s)": 0.032858
    },
    {
      "epoch": 0.38599040837087634,
      "grad_norm": 0.11107171326875687,
      "learning_rate": 0.0002166126019887039,
      "loss": 0.4144379794597626,
      "memory(GiB)": 78.33,
      "step": 1992,
      "token_acc": 0.8786039878072653,
      "train_speed(iter/s)": 0.032859
    },
    {
      "epoch": 0.3861841786562031,
      "grad_norm": 0.10761483013629913,
      "learning_rate": 0.00021652645550069392,
      "loss": 0.3946702480316162,
      "memory(GiB)": 78.33,
      "step": 1993,
      "token_acc": 0.8829077659918314,
      "train_speed(iter/s)": 0.03286
    },
    {
      "epoch": 0.3863779489415298,
      "grad_norm": 0.10240019857883453,
      "learning_rate": 0.000216440281688481,
      "loss": 0.36723431944847107,
      "memory(GiB)": 78.33,
      "step": 1994,
      "token_acc": 0.8909840895698291,
      "train_speed(iter/s)": 0.032862
    },
    {
      "epoch": 0.38657171922685657,
      "grad_norm": 0.10443782061338425,
      "learning_rate": 0.00021635408058745908,
      "loss": 0.3627747893333435,
      "memory(GiB)": 78.33,
      "step": 1995,
      "token_acc": 0.8937532210377844,
      "train_speed(iter/s)": 0.032863
    },
    {
      "epoch": 0.3867654895121833,
      "grad_norm": 0.10117712616920471,
      "learning_rate": 0.00021626785223303327,
      "loss": 0.36782556772232056,
      "memory(GiB)": 78.33,
      "step": 1996,
      "token_acc": 0.8936070197430273,
      "train_speed(iter/s)": 0.032864
    },
    {
      "epoch": 0.38695925979751006,
      "grad_norm": 0.10845784842967987,
      "learning_rate": 0.00021618159666061983,
      "loss": 0.4153798818588257,
      "memory(GiB)": 78.33,
      "step": 1997,
      "token_acc": 0.8783169270276054,
      "train_speed(iter/s)": 0.032865
    },
    {
      "epoch": 0.3871530300828368,
      "grad_norm": 0.11158988624811172,
      "learning_rate": 0.00021609531390564635,
      "loss": 0.3699086904525757,
      "memory(GiB)": 78.33,
      "step": 1998,
      "token_acc": 0.8905605642615397,
      "train_speed(iter/s)": 0.032866
    },
    {
      "epoch": 0.38734680036816355,
      "grad_norm": 0.1030752882361412,
      "learning_rate": 0.0002160090040035513,
      "loss": 0.3778277039527893,
      "memory(GiB)": 78.33,
      "step": 1999,
      "token_acc": 0.8885486675143369,
      "train_speed(iter/s)": 0.032867
    },
    {
      "epoch": 0.3875405706534903,
      "grad_norm": 0.10194329917430878,
      "learning_rate": 0.00021592266698978462,
      "loss": 0.3767800033092499,
      "memory(GiB)": 78.33,
      "step": 2000,
      "token_acc": 0.8890578277836907,
      "train_speed(iter/s)": 0.032868
    },
    {
      "epoch": 0.3875405706534903,
      "eval_loss": 0.44034290313720703,
      "eval_runtime": 1345.0564,
      "eval_samples_per_second": 5.018,
      "eval_steps_per_second": 5.018,
      "eval_token_acc": 0.8890313859070251,
      "step": 2000
    },
    {
      "epoch": 0.38773434093881703,
      "grad_norm": 0.12286636233329773,
      "learning_rate": 0.00021583630289980724,
      "loss": 0.41180068254470825,
      "memory(GiB)": 78.33,
      "step": 2001,
      "token_acc": 0.8807242798353909,
      "train_speed(iter/s)": 0.032152
    },
    {
      "epoch": 0.3879281112241438,
      "grad_norm": 0.10262475907802582,
      "learning_rate": 0.00021574991176909113,
      "loss": 0.37820303440093994,
      "memory(GiB)": 78.33,
      "step": 2002,
      "token_acc": 0.8894734153940447,
      "train_speed(iter/s)": 0.032154
    },
    {
      "epoch": 0.3881218815094705,
      "grad_norm": 0.10098009556531906,
      "learning_rate": 0.00021566349363311949,
      "loss": 0.3635365962982178,
      "memory(GiB)": 78.33,
      "step": 2003,
      "token_acc": 0.8932263226068625,
      "train_speed(iter/s)": 0.032155
    },
    {
      "epoch": 0.38831565179479727,
      "grad_norm": 0.11192551255226135,
      "learning_rate": 0.00021557704852738654,
      "loss": 0.39867013692855835,
      "memory(GiB)": 78.33,
      "step": 2004,
      "token_acc": 0.8858497030607584,
      "train_speed(iter/s)": 0.032157
    },
    {
      "epoch": 0.388509422080124,
      "grad_norm": 0.0996481254696846,
      "learning_rate": 0.00021549057648739768,
      "loss": 0.3613511025905609,
      "memory(GiB)": 78.33,
      "step": 2005,
      "token_acc": 0.8933584214808787,
      "train_speed(iter/s)": 0.032158
    },
    {
      "epoch": 0.38870319236545076,
      "grad_norm": 0.12326376140117645,
      "learning_rate": 0.00021540407754866924,
      "loss": 0.4421766996383667,
      "memory(GiB)": 78.33,
      "step": 2006,
      "token_acc": 0.8737238044062332,
      "train_speed(iter/s)": 0.03216
    },
    {
      "epoch": 0.3888969626507775,
      "grad_norm": 0.09973074495792389,
      "learning_rate": 0.00021531755174672868,
      "loss": 0.34252458810806274,
      "memory(GiB)": 78.33,
      "step": 2007,
      "token_acc": 0.8975569907844689,
      "train_speed(iter/s)": 0.032161
    },
    {
      "epoch": 0.38909073293610424,
      "grad_norm": 0.09278019517660141,
      "learning_rate": 0.00021523099911711447,
      "loss": 0.3389296233654022,
      "memory(GiB)": 78.33,
      "step": 2008,
      "token_acc": 0.89790143300206,
      "train_speed(iter/s)": 0.032162
    },
    {
      "epoch": 0.389284503221431,
      "grad_norm": 0.11313237994909286,
      "learning_rate": 0.00021514441969537607,
      "loss": 0.43270501494407654,
      "memory(GiB)": 78.33,
      "step": 2009,
      "token_acc": 0.8762101038587737,
      "train_speed(iter/s)": 0.032164
    },
    {
      "epoch": 0.38947827350675773,
      "grad_norm": 0.10002917051315308,
      "learning_rate": 0.00021505781351707402,
      "loss": 0.3512948751449585,
      "memory(GiB)": 78.33,
      "step": 2010,
      "token_acc": 0.8962549078828148,
      "train_speed(iter/s)": 0.032165
    },
    {
      "epoch": 0.3896720437920845,
      "grad_norm": 0.11693061888217926,
      "learning_rate": 0.0002149711806177798,
      "loss": 0.40658366680145264,
      "memory(GiB)": 78.33,
      "step": 2011,
      "token_acc": 0.8811087609929867,
      "train_speed(iter/s)": 0.032167
    },
    {
      "epoch": 0.3898658140774112,
      "grad_norm": 0.10790550708770752,
      "learning_rate": 0.00021488452103307585,
      "loss": 0.41037797927856445,
      "memory(GiB)": 78.33,
      "step": 2012,
      "token_acc": 0.88296488946684,
      "train_speed(iter/s)": 0.032168
    },
    {
      "epoch": 0.39005958436273797,
      "grad_norm": 0.11255151033401489,
      "learning_rate": 0.0002147978347985556,
      "loss": 0.37893742322921753,
      "memory(GiB)": 78.33,
      "step": 2013,
      "token_acc": 0.889498343046509,
      "train_speed(iter/s)": 0.032169
    },
    {
      "epoch": 0.3902533546480647,
      "grad_norm": 0.10197418183088303,
      "learning_rate": 0.0002147111219498234,
      "loss": 0.3498515188694,
      "memory(GiB)": 78.33,
      "step": 2014,
      "token_acc": 0.8973286219081272,
      "train_speed(iter/s)": 0.032171
    },
    {
      "epoch": 0.39044712493339145,
      "grad_norm": 0.09706109017133713,
      "learning_rate": 0.00021462438252249457,
      "loss": 0.340999037027359,
      "memory(GiB)": 78.33,
      "step": 2015,
      "token_acc": 0.896882369710006,
      "train_speed(iter/s)": 0.032172
    },
    {
      "epoch": 0.3906408952187182,
      "grad_norm": 0.10438670963048935,
      "learning_rate": 0.00021453761655219528,
      "loss": 0.389445424079895,
      "memory(GiB)": 78.33,
      "step": 2016,
      "token_acc": 0.8829340326399674,
      "train_speed(iter/s)": 0.032174
    },
    {
      "epoch": 0.39083466550404494,
      "grad_norm": 0.11134763062000275,
      "learning_rate": 0.00021445082407456272,
      "loss": 0.43515315651893616,
      "memory(GiB)": 78.33,
      "step": 2017,
      "token_acc": 0.8730318643677499,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.3910284357893717,
      "grad_norm": 0.09228318929672241,
      "learning_rate": 0.00021436400512524483,
      "loss": 0.33381155133247375,
      "memory(GiB)": 78.33,
      "step": 2018,
      "token_acc": 0.902964766542538,
      "train_speed(iter/s)": 0.032177
    },
    {
      "epoch": 0.39122220607469843,
      "grad_norm": 0.10420046001672745,
      "learning_rate": 0.00021427715973990056,
      "loss": 0.3501606285572052,
      "memory(GiB)": 78.33,
      "step": 2019,
      "token_acc": 0.8966936364399899,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.3914159763600252,
      "grad_norm": 0.09859726577997208,
      "learning_rate": 0.00021419028795419953,
      "loss": 0.3734872341156006,
      "memory(GiB)": 78.33,
      "step": 2020,
      "token_acc": 0.8910265718019457,
      "train_speed(iter/s)": 0.032179
    },
    {
      "epoch": 0.3916097466453519,
      "grad_norm": 0.11447544395923615,
      "learning_rate": 0.00021410338980382238,
      "loss": 0.38576456904411316,
      "memory(GiB)": 78.33,
      "step": 2021,
      "token_acc": 0.887767163988371,
      "train_speed(iter/s)": 0.032181
    },
    {
      "epoch": 0.39180351693067866,
      "grad_norm": 0.10994785279035568,
      "learning_rate": 0.00021401646532446053,
      "loss": 0.40534737706184387,
      "memory(GiB)": 78.33,
      "step": 2022,
      "token_acc": 0.8793884919720366,
      "train_speed(iter/s)": 0.032182
    },
    {
      "epoch": 0.3919972872160054,
      "grad_norm": 0.10715804994106293,
      "learning_rate": 0.00021392951455181619,
      "loss": 0.3926943242549896,
      "memory(GiB)": 78.33,
      "step": 2023,
      "token_acc": 0.8848688300385429,
      "train_speed(iter/s)": 0.032183
    },
    {
      "epoch": 0.39219105750133215,
      "grad_norm": 0.10480938851833344,
      "learning_rate": 0.00021384253752160235,
      "loss": 0.3569415807723999,
      "memory(GiB)": 78.33,
      "step": 2024,
      "token_acc": 0.8943907463159563,
      "train_speed(iter/s)": 0.032185
    },
    {
      "epoch": 0.3923848277866589,
      "grad_norm": 0.09940195083618164,
      "learning_rate": 0.00021375553426954285,
      "loss": 0.3529174029827118,
      "memory(GiB)": 78.33,
      "step": 2025,
      "token_acc": 0.8969288159844653,
      "train_speed(iter/s)": 0.032186
    },
    {
      "epoch": 0.39257859807198564,
      "grad_norm": 0.11303528398275375,
      "learning_rate": 0.00021366850483137226,
      "loss": 0.40097349882125854,
      "memory(GiB)": 78.33,
      "step": 2026,
      "token_acc": 0.8815832710978342,
      "train_speed(iter/s)": 0.032188
    },
    {
      "epoch": 0.3927723683573124,
      "grad_norm": 0.0959119200706482,
      "learning_rate": 0.00021358144924283584,
      "loss": 0.35317641496658325,
      "memory(GiB)": 78.33,
      "step": 2027,
      "token_acc": 0.8957997557997558,
      "train_speed(iter/s)": 0.032189
    },
    {
      "epoch": 0.39296613864263913,
      "grad_norm": 0.11829999089241028,
      "learning_rate": 0.0002134943675396898,
      "loss": 0.38520050048828125,
      "memory(GiB)": 78.33,
      "step": 2028,
      "token_acc": 0.8890619591554171,
      "train_speed(iter/s)": 0.032191
    },
    {
      "epoch": 0.3931599089279659,
      "grad_norm": 0.11327025294303894,
      "learning_rate": 0.0002134072597577008,
      "loss": 0.39436087012290955,
      "memory(GiB)": 78.33,
      "step": 2029,
      "token_acc": 0.8814638027048528,
      "train_speed(iter/s)": 0.032192
    },
    {
      "epoch": 0.3933536792132926,
      "grad_norm": 0.09940643608570099,
      "learning_rate": 0.0002133201259326464,
      "loss": 0.3501843810081482,
      "memory(GiB)": 78.33,
      "step": 2030,
      "token_acc": 0.8952846160250895,
      "train_speed(iter/s)": 0.032194
    },
    {
      "epoch": 0.39354744949861936,
      "grad_norm": 0.09761599451303482,
      "learning_rate": 0.0002132329661003148,
      "loss": 0.3546067178249359,
      "memory(GiB)": 78.33,
      "step": 2031,
      "token_acc": 0.8960130106943965,
      "train_speed(iter/s)": 0.032195
    },
    {
      "epoch": 0.3937412197839461,
      "grad_norm": 0.10584171116352081,
      "learning_rate": 0.00021314578029650493,
      "loss": 0.39139533042907715,
      "memory(GiB)": 78.33,
      "step": 2032,
      "token_acc": 0.8860886829913964,
      "train_speed(iter/s)": 0.032196
    },
    {
      "epoch": 0.39393499006927285,
      "grad_norm": 0.1090681329369545,
      "learning_rate": 0.00021305856855702624,
      "loss": 0.37363752722740173,
      "memory(GiB)": 78.33,
      "step": 2033,
      "token_acc": 0.8886917688801884,
      "train_speed(iter/s)": 0.032198
    },
    {
      "epoch": 0.39412876035459965,
      "grad_norm": 0.10559553653001785,
      "learning_rate": 0.00021297133091769904,
      "loss": 0.36771360039711,
      "memory(GiB)": 78.33,
      "step": 2034,
      "token_acc": 0.8921826625386997,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.3943225306399264,
      "grad_norm": 0.10535501688718796,
      "learning_rate": 0.00021288406741435412,
      "loss": 0.36465299129486084,
      "memory(GiB)": 78.33,
      "step": 2035,
      "token_acc": 0.8928280358598207,
      "train_speed(iter/s)": 0.0322
    },
    {
      "epoch": 0.39451630092525314,
      "grad_norm": 0.1058509349822998,
      "learning_rate": 0.0002127967780828329,
      "loss": 0.3727305829524994,
      "memory(GiB)": 78.33,
      "step": 2036,
      "token_acc": 0.8917116094237866,
      "train_speed(iter/s)": 0.032202
    },
    {
      "epoch": 0.3947100712105799,
      "grad_norm": 0.10819867253303528,
      "learning_rate": 0.00021270946295898755,
      "loss": 0.36687490344047546,
      "memory(GiB)": 78.33,
      "step": 2037,
      "token_acc": 0.8913552022967728,
      "train_speed(iter/s)": 0.032203
    },
    {
      "epoch": 0.3949038414959066,
      "grad_norm": 0.10262859612703323,
      "learning_rate": 0.0002126221220786807,
      "loss": 0.3651004731655121,
      "memory(GiB)": 78.33,
      "step": 2038,
      "token_acc": 0.892148337595908,
      "train_speed(iter/s)": 0.032204
    },
    {
      "epoch": 0.39509761178123337,
      "grad_norm": 0.10185900330543518,
      "learning_rate": 0.0002125347554777856,
      "loss": 0.3557824492454529,
      "memory(GiB)": 78.33,
      "step": 2039,
      "token_acc": 0.8953365688963917,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.3952913820665601,
      "grad_norm": 0.09542527794837952,
      "learning_rate": 0.000212447363192186,
      "loss": 0.36669260263442993,
      "memory(GiB)": 78.33,
      "step": 2040,
      "token_acc": 0.8931179156718667,
      "train_speed(iter/s)": 0.032207
    },
    {
      "epoch": 0.39548515235188686,
      "grad_norm": 0.10032487660646439,
      "learning_rate": 0.00021235994525777637,
      "loss": 0.3561350107192993,
      "memory(GiB)": 78.33,
      "step": 2041,
      "token_acc": 0.8934586347077768,
      "train_speed(iter/s)": 0.032208
    },
    {
      "epoch": 0.3956789226372136,
      "grad_norm": 0.1082734614610672,
      "learning_rate": 0.0002122725017104615,
      "loss": 0.34290170669555664,
      "memory(GiB)": 78.33,
      "step": 2042,
      "token_acc": 0.8980645927333175,
      "train_speed(iter/s)": 0.03221
    },
    {
      "epoch": 0.39587269292254035,
      "grad_norm": 0.10429167002439499,
      "learning_rate": 0.00021218503258615688,
      "loss": 0.38119545578956604,
      "memory(GiB)": 78.33,
      "step": 2043,
      "token_acc": 0.8884871605078859,
      "train_speed(iter/s)": 0.032211
    },
    {
      "epoch": 0.3960664632078671,
      "grad_norm": 0.10330045223236084,
      "learning_rate": 0.00021209753792078836,
      "loss": 0.36754027009010315,
      "memory(GiB)": 78.33,
      "step": 2044,
      "token_acc": 0.8926996316089867,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.39626023349319384,
      "grad_norm": 0.10568582266569138,
      "learning_rate": 0.00021201001775029244,
      "loss": 0.37445148825645447,
      "memory(GiB)": 78.33,
      "step": 2045,
      "token_acc": 0.890682963949396,
      "train_speed(iter/s)": 0.032214
    },
    {
      "epoch": 0.3964540037785206,
      "grad_norm": 0.11252961307764053,
      "learning_rate": 0.00021192247211061595,
      "loss": 0.40704840421676636,
      "memory(GiB)": 78.33,
      "step": 2046,
      "token_acc": 0.8822551520536559,
      "train_speed(iter/s)": 0.032215
    },
    {
      "epoch": 0.3966477740638473,
      "grad_norm": 0.10527341812849045,
      "learning_rate": 0.0002118349010377162,
      "loss": 0.4128043055534363,
      "memory(GiB)": 78.33,
      "step": 2047,
      "token_acc": 0.8777996488801404,
      "train_speed(iter/s)": 0.032217
    },
    {
      "epoch": 0.39684154434917407,
      "grad_norm": 0.1078762486577034,
      "learning_rate": 0.00021174730456756106,
      "loss": 0.3727009892463684,
      "memory(GiB)": 78.33,
      "step": 2048,
      "token_acc": 0.8923324669454176,
      "train_speed(iter/s)": 0.032218
    },
    {
      "epoch": 0.3970353146345008,
      "grad_norm": 0.10819050669670105,
      "learning_rate": 0.00021165968273612875,
      "loss": 0.37524309754371643,
      "memory(GiB)": 78.33,
      "step": 2049,
      "token_acc": 0.8889318457969734,
      "train_speed(iter/s)": 0.03222
    },
    {
      "epoch": 0.39722908491982756,
      "grad_norm": 0.10360529273748398,
      "learning_rate": 0.0002115720355794078,
      "loss": 0.37673884630203247,
      "memory(GiB)": 78.33,
      "step": 2050,
      "token_acc": 0.8891292318527089,
      "train_speed(iter/s)": 0.032221
    },
    {
      "epoch": 0.3974228552051543,
      "grad_norm": 0.10623595118522644,
      "learning_rate": 0.00021148436313339739,
      "loss": 0.3622683882713318,
      "memory(GiB)": 78.33,
      "step": 2051,
      "token_acc": 0.8929465428046949,
      "train_speed(iter/s)": 0.032222
    },
    {
      "epoch": 0.39761662549048105,
      "grad_norm": 0.1145068109035492,
      "learning_rate": 0.0002113966654341069,
      "loss": 0.37567758560180664,
      "memory(GiB)": 78.33,
      "step": 2052,
      "token_acc": 0.8903638151425762,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.3978103957758078,
      "grad_norm": 0.11284542083740234,
      "learning_rate": 0.00021130894251755608,
      "loss": 0.3825456202030182,
      "memory(GiB)": 78.33,
      "step": 2053,
      "token_acc": 0.8858395490519843,
      "train_speed(iter/s)": 0.032225
    },
    {
      "epoch": 0.39800416606113453,
      "grad_norm": 0.11392559856176376,
      "learning_rate": 0.00021122119441977516,
      "loss": 0.38240864872932434,
      "memory(GiB)": 78.33,
      "step": 2054,
      "token_acc": 0.8883792048929664,
      "train_speed(iter/s)": 0.032227
    },
    {
      "epoch": 0.3981979363464613,
      "grad_norm": 0.11121159791946411,
      "learning_rate": 0.00021113342117680463,
      "loss": 0.4033408761024475,
      "memory(GiB)": 78.33,
      "step": 2055,
      "token_acc": 0.8815422034050986,
      "train_speed(iter/s)": 0.032228
    },
    {
      "epoch": 0.398391706631788,
      "grad_norm": 0.11232481151819229,
      "learning_rate": 0.00021104562282469523,
      "loss": 0.3962811529636383,
      "memory(GiB)": 78.33,
      "step": 2056,
      "token_acc": 0.8822988914577035,
      "train_speed(iter/s)": 0.03223
    },
    {
      "epoch": 0.39858547691711477,
      "grad_norm": 0.1073499396443367,
      "learning_rate": 0.00021095779939950827,
      "loss": 0.36865997314453125,
      "memory(GiB)": 78.33,
      "step": 2057,
      "token_acc": 0.890293265087614,
      "train_speed(iter/s)": 0.032231
    },
    {
      "epoch": 0.3987792472024415,
      "grad_norm": 0.10009787231683731,
      "learning_rate": 0.00021086995093731506,
      "loss": 0.36048176884651184,
      "memory(GiB)": 78.33,
      "step": 2058,
      "token_acc": 0.8924960221731766,
      "train_speed(iter/s)": 0.032232
    },
    {
      "epoch": 0.39897301748776826,
      "grad_norm": 0.10518264770507812,
      "learning_rate": 0.00021078207747419737,
      "loss": 0.375562846660614,
      "memory(GiB)": 78.33,
      "step": 2059,
      "token_acc": 0.8906748031710573,
      "train_speed(iter/s)": 0.032234
    },
    {
      "epoch": 0.399166787773095,
      "grad_norm": 0.11405156552791595,
      "learning_rate": 0.00021069417904624713,
      "loss": 0.39197370409965515,
      "memory(GiB)": 78.33,
      "step": 2060,
      "token_acc": 0.8833886430678466,
      "train_speed(iter/s)": 0.032235
    },
    {
      "epoch": 0.39936055805842174,
      "grad_norm": 0.11229176819324493,
      "learning_rate": 0.00021060625568956672,
      "loss": 0.38142168521881104,
      "memory(GiB)": 78.33,
      "step": 2061,
      "token_acc": 0.8872789019385878,
      "train_speed(iter/s)": 0.032236
    },
    {
      "epoch": 0.3995543283437485,
      "grad_norm": 0.10720710456371307,
      "learning_rate": 0.0002105183074402685,
      "loss": 0.3908834755420685,
      "memory(GiB)": 78.33,
      "step": 2062,
      "token_acc": 0.8867632578421719,
      "train_speed(iter/s)": 0.032238
    },
    {
      "epoch": 0.39974809862907523,
      "grad_norm": 0.11731129139661789,
      "learning_rate": 0.00021043033433447523,
      "loss": 0.3812180757522583,
      "memory(GiB)": 78.33,
      "step": 2063,
      "token_acc": 0.8876133286235723,
      "train_speed(iter/s)": 0.032239
    },
    {
      "epoch": 0.399941868914402,
      "grad_norm": 0.1061178669333458,
      "learning_rate": 0.00021034233640831985,
      "loss": 0.35581403970718384,
      "memory(GiB)": 78.33,
      "step": 2064,
      "token_acc": 0.8961584794538884,
      "train_speed(iter/s)": 0.032241
    },
    {
      "epoch": 0.4001356391997287,
      "grad_norm": 0.12066232413053513,
      "learning_rate": 0.0002102543136979454,
      "loss": 0.407576322555542,
      "memory(GiB)": 78.33,
      "step": 2065,
      "token_acc": 0.8834347797630735,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.40032940948505547,
      "grad_norm": 0.10802853107452393,
      "learning_rate": 0.00021016626623950523,
      "loss": 0.3750949800014496,
      "memory(GiB)": 78.33,
      "step": 2066,
      "token_acc": 0.8900487525093204,
      "train_speed(iter/s)": 0.032243
    },
    {
      "epoch": 0.4005231797703822,
      "grad_norm": 0.10871551930904388,
      "learning_rate": 0.00021007819406916283,
      "loss": 0.3510313928127289,
      "memory(GiB)": 78.33,
      "step": 2067,
      "token_acc": 0.8955925167910985,
      "train_speed(iter/s)": 0.032245
    },
    {
      "epoch": 0.40071695005570895,
      "grad_norm": 0.10944053530693054,
      "learning_rate": 0.0002099900972230917,
      "loss": 0.3768633306026459,
      "memory(GiB)": 78.33,
      "step": 2068,
      "token_acc": 0.8907095472471034,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.4009107203410357,
      "grad_norm": 0.11056249588727951,
      "learning_rate": 0.0002099019757374757,
      "loss": 0.38695505261421204,
      "memory(GiB)": 78.33,
      "step": 2069,
      "token_acc": 0.8846898656182149,
      "train_speed(iter/s)": 0.032247
    },
    {
      "epoch": 0.40110449062636244,
      "grad_norm": 0.10126353055238724,
      "learning_rate": 0.00020981382964850858,
      "loss": 0.3784712851047516,
      "memory(GiB)": 78.33,
      "step": 2070,
      "token_acc": 0.8864102038725541,
      "train_speed(iter/s)": 0.032249
    },
    {
      "epoch": 0.4012982609116892,
      "grad_norm": 0.11109844595193863,
      "learning_rate": 0.00020972565899239441,
      "loss": 0.39917057752609253,
      "memory(GiB)": 78.33,
      "step": 2071,
      "token_acc": 0.884891448825875,
      "train_speed(iter/s)": 0.03225
    },
    {
      "epoch": 0.40149203119701593,
      "grad_norm": 0.10675584524869919,
      "learning_rate": 0.0002096374638053472,
      "loss": 0.3992624580860138,
      "memory(GiB)": 78.33,
      "step": 2072,
      "token_acc": 0.8833933488667656,
      "train_speed(iter/s)": 0.032251
    },
    {
      "epoch": 0.4016858014823427,
      "grad_norm": 0.09866248071193695,
      "learning_rate": 0.000209549244123591,
      "loss": 0.3732259273529053,
      "memory(GiB)": 78.33,
      "step": 2073,
      "token_acc": 0.891376350844958,
      "train_speed(iter/s)": 0.032253
    },
    {
      "epoch": 0.4018795717676694,
      "grad_norm": 0.11382555216550827,
      "learning_rate": 0.00020946099998336019,
      "loss": 0.40181201696395874,
      "memory(GiB)": 78.33,
      "step": 2074,
      "token_acc": 0.8830451706345614,
      "train_speed(iter/s)": 0.032254
    },
    {
      "epoch": 0.40207334205299616,
      "grad_norm": 0.09255445003509521,
      "learning_rate": 0.0002093727314208989,
      "loss": 0.317599356174469,
      "memory(GiB)": 78.33,
      "step": 2075,
      "token_acc": 0.9049172687019342,
      "train_speed(iter/s)": 0.032255
    },
    {
      "epoch": 0.4022671123383229,
      "grad_norm": 0.12287932634353638,
      "learning_rate": 0.00020928443847246134,
      "loss": 0.41253185272216797,
      "memory(GiB)": 78.33,
      "step": 2076,
      "token_acc": 0.8792145844223732,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.40246088262364965,
      "grad_norm": 0.1004108339548111,
      "learning_rate": 0.0002091961211743119,
      "loss": 0.32616835832595825,
      "memory(GiB)": 78.33,
      "step": 2077,
      "token_acc": 0.9020430729022693,
      "train_speed(iter/s)": 0.032258
    },
    {
      "epoch": 0.4026546529089764,
      "grad_norm": 0.10672204196453094,
      "learning_rate": 0.00020910777956272485,
      "loss": 0.37214604020118713,
      "memory(GiB)": 78.33,
      "step": 2078,
      "token_acc": 0.8916562179414111,
      "train_speed(iter/s)": 0.032259
    },
    {
      "epoch": 0.40284842319430314,
      "grad_norm": 0.10667048394680023,
      "learning_rate": 0.00020901941367398446,
      "loss": 0.3868388533592224,
      "memory(GiB)": 78.33,
      "step": 2079,
      "token_acc": 0.886404833836858,
      "train_speed(iter/s)": 0.032261
    },
    {
      "epoch": 0.4030421934796299,
      "grad_norm": 0.1087040826678276,
      "learning_rate": 0.0002089310235443849,
      "loss": 0.3963964879512787,
      "memory(GiB)": 78.33,
      "step": 2080,
      "token_acc": 0.8864254703328509,
      "train_speed(iter/s)": 0.032262
    },
    {
      "epoch": 0.40323596376495663,
      "grad_norm": 0.10210565477609634,
      "learning_rate": 0.0002088426092102305,
      "loss": 0.38553428649902344,
      "memory(GiB)": 78.33,
      "step": 2081,
      "token_acc": 0.8849096532638776,
      "train_speed(iter/s)": 0.032263
    },
    {
      "epoch": 0.4034297340502834,
      "grad_norm": 0.10996957868337631,
      "learning_rate": 0.0002087541707078353,
      "loss": 0.40852048993110657,
      "memory(GiB)": 78.33,
      "step": 2082,
      "token_acc": 0.8819868995633188,
      "train_speed(iter/s)": 0.032264
    },
    {
      "epoch": 0.4036235043356101,
      "grad_norm": 0.1148100420832634,
      "learning_rate": 0.00020866570807352337,
      "loss": 0.40155094861984253,
      "memory(GiB)": 78.33,
      "step": 2083,
      "token_acc": 0.8844750224349387,
      "train_speed(iter/s)": 0.032266
    },
    {
      "epoch": 0.40381727462093686,
      "grad_norm": 0.09997189044952393,
      "learning_rate": 0.0002085772213436288,
      "loss": 0.35030126571655273,
      "memory(GiB)": 78.33,
      "step": 2084,
      "token_acc": 0.8974117047048323,
      "train_speed(iter/s)": 0.032267
    },
    {
      "epoch": 0.4040110449062636,
      "grad_norm": 0.10798052698373795,
      "learning_rate": 0.00020848871055449537,
      "loss": 0.3880666196346283,
      "memory(GiB)": 78.33,
      "step": 2085,
      "token_acc": 0.8854523021703774,
      "train_speed(iter/s)": 0.032269
    },
    {
      "epoch": 0.40420481519159035,
      "grad_norm": 0.11830330640077591,
      "learning_rate": 0.00020840017574247683,
      "loss": 0.446191668510437,
      "memory(GiB)": 78.33,
      "step": 2086,
      "token_acc": 0.8699704224566103,
      "train_speed(iter/s)": 0.03227
    },
    {
      "epoch": 0.4043985854769171,
      "grad_norm": 0.0969855859875679,
      "learning_rate": 0.00020831161694393683,
      "loss": 0.36765626072883606,
      "memory(GiB)": 78.33,
      "step": 2087,
      "token_acc": 0.892250186892599,
      "train_speed(iter/s)": 0.032271
    },
    {
      "epoch": 0.40459235576224384,
      "grad_norm": 0.10226907581090927,
      "learning_rate": 0.00020822303419524893,
      "loss": 0.36062684655189514,
      "memory(GiB)": 78.33,
      "step": 2088,
      "token_acc": 0.893473640557773,
      "train_speed(iter/s)": 0.032273
    },
    {
      "epoch": 0.4047861260475706,
      "grad_norm": 0.12443238496780396,
      "learning_rate": 0.0002081344275327963,
      "loss": 0.4231499135494232,
      "memory(GiB)": 78.33,
      "step": 2089,
      "token_acc": 0.8769953656024717,
      "train_speed(iter/s)": 0.032274
    },
    {
      "epoch": 0.4049798963328973,
      "grad_norm": 0.11125290393829346,
      "learning_rate": 0.00020804579699297218,
      "loss": 0.38650333881378174,
      "memory(GiB)": 78.33,
      "step": 2090,
      "token_acc": 0.8851398261545724,
      "train_speed(iter/s)": 0.032275
    },
    {
      "epoch": 0.40517366661822407,
      "grad_norm": 0.0958833247423172,
      "learning_rate": 0.00020795714261217949,
      "loss": 0.34758636355400085,
      "memory(GiB)": 78.33,
      "step": 2091,
      "token_acc": 0.8987275662330263,
      "train_speed(iter/s)": 0.032277
    },
    {
      "epoch": 0.4053674369035508,
      "grad_norm": 0.10977054387331009,
      "learning_rate": 0.00020786846442683095,
      "loss": 0.3909391760826111,
      "memory(GiB)": 78.33,
      "step": 2092,
      "token_acc": 0.886636506003132,
      "train_speed(iter/s)": 0.032278
    },
    {
      "epoch": 0.40556120718887756,
      "grad_norm": 0.10875298827886581,
      "learning_rate": 0.00020777976247334906,
      "loss": 0.3958970308303833,
      "memory(GiB)": 78.33,
      "step": 2093,
      "token_acc": 0.8842546456415968,
      "train_speed(iter/s)": 0.032279
    },
    {
      "epoch": 0.4057549774742043,
      "grad_norm": 0.11530545353889465,
      "learning_rate": 0.00020769103678816616,
      "loss": 0.3912769556045532,
      "memory(GiB)": 78.33,
      "step": 2094,
      "token_acc": 0.8857514450867052,
      "train_speed(iter/s)": 0.032281
    },
    {
      "epoch": 0.4059487477595311,
      "grad_norm": 0.11220169067382812,
      "learning_rate": 0.00020760228740772423,
      "loss": 0.39823243021965027,
      "memory(GiB)": 78.33,
      "step": 2095,
      "token_acc": 0.8826204060355064,
      "train_speed(iter/s)": 0.032282
    },
    {
      "epoch": 0.40614251804485785,
      "grad_norm": 0.10903418809175491,
      "learning_rate": 0.00020751351436847497,
      "loss": 0.4057612121105194,
      "memory(GiB)": 78.33,
      "step": 2096,
      "token_acc": 0.883694474539545,
      "train_speed(iter/s)": 0.032283
    },
    {
      "epoch": 0.4063362883301846,
      "grad_norm": 0.11058089882135391,
      "learning_rate": 0.00020742471770687998,
      "loss": 0.40834489464759827,
      "memory(GiB)": 78.33,
      "step": 2097,
      "token_acc": 0.8802559666034989,
      "train_speed(iter/s)": 0.032285
    },
    {
      "epoch": 0.40653005861551134,
      "grad_norm": 0.10095040500164032,
      "learning_rate": 0.00020733589745941034,
      "loss": 0.3721862733364105,
      "memory(GiB)": 78.33,
      "step": 2098,
      "token_acc": 0.8911490351522632,
      "train_speed(iter/s)": 0.032286
    },
    {
      "epoch": 0.4067238289008381,
      "grad_norm": 0.12033317983150482,
      "learning_rate": 0.00020724705366254693,
      "loss": 0.43703311681747437,
      "memory(GiB)": 78.33,
      "step": 2099,
      "token_acc": 0.874896265560166,
      "train_speed(iter/s)": 0.032287
    },
    {
      "epoch": 0.4069175991861648,
      "grad_norm": 0.10075613856315613,
      "learning_rate": 0.0002071581863527803,
      "loss": 0.33554336428642273,
      "memory(GiB)": 78.33,
      "step": 2100,
      "token_acc": 0.900580013797685,
      "train_speed(iter/s)": 0.032289
    },
    {
      "epoch": 0.40711136947149157,
      "grad_norm": 0.10525397956371307,
      "learning_rate": 0.00020706929556661068,
      "loss": 0.37453317642211914,
      "memory(GiB)": 78.33,
      "step": 2101,
      "token_acc": 0.8898639629661186,
      "train_speed(iter/s)": 0.03229
    },
    {
      "epoch": 0.4073051397568183,
      "grad_norm": 0.10155075043439865,
      "learning_rate": 0.00020698038134054782,
      "loss": 0.38334396481513977,
      "memory(GiB)": 78.33,
      "step": 2102,
      "token_acc": 0.885774579297657,
      "train_speed(iter/s)": 0.032291
    },
    {
      "epoch": 0.40749891004214506,
      "grad_norm": 0.11583290249109268,
      "learning_rate": 0.00020689144371111118,
      "loss": 0.40256598591804504,
      "memory(GiB)": 78.33,
      "step": 2103,
      "token_acc": 0.8832213494842434,
      "train_speed(iter/s)": 0.032293
    },
    {
      "epoch": 0.4076926803274718,
      "grad_norm": 0.10212710499763489,
      "learning_rate": 0.00020680248271482993,
      "loss": 0.3900299668312073,
      "memory(GiB)": 78.33,
      "step": 2104,
      "token_acc": 0.8837178537411366,
      "train_speed(iter/s)": 0.032294
    },
    {
      "epoch": 0.40788645061279855,
      "grad_norm": 0.1136767715215683,
      "learning_rate": 0.0002067134983882427,
      "loss": 0.40916168689727783,
      "memory(GiB)": 78.33,
      "step": 2105,
      "token_acc": 0.8802716981132076,
      "train_speed(iter/s)": 0.032296
    },
    {
      "epoch": 0.4080802208981253,
      "grad_norm": 0.1174011304974556,
      "learning_rate": 0.00020662449076789768,
      "loss": 0.3446325957775116,
      "memory(GiB)": 78.33,
      "step": 2106,
      "token_acc": 0.8981224992305324,
      "train_speed(iter/s)": 0.032297
    },
    {
      "epoch": 0.40827399118345203,
      "grad_norm": 0.10661471635103226,
      "learning_rate": 0.00020653545989035278,
      "loss": 0.3691784143447876,
      "memory(GiB)": 78.33,
      "step": 2107,
      "token_acc": 0.8892190390760778,
      "train_speed(iter/s)": 0.032299
    },
    {
      "epoch": 0.4084677614687788,
      "grad_norm": 0.10399375855922699,
      "learning_rate": 0.00020644640579217533,
      "loss": 0.3859068751335144,
      "memory(GiB)": 78.33,
      "step": 2108,
      "token_acc": 0.8871030059500812,
      "train_speed(iter/s)": 0.0323
    },
    {
      "epoch": 0.4086615317541055,
      "grad_norm": 0.10391335189342499,
      "learning_rate": 0.0002063573285099422,
      "loss": 0.3771337866783142,
      "memory(GiB)": 78.33,
      "step": 2109,
      "token_acc": 0.8897867506583363,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.40885530203943227,
      "grad_norm": 0.11487588286399841,
      "learning_rate": 0.00020626822808023993,
      "loss": 0.3902420699596405,
      "memory(GiB)": 78.33,
      "step": 2110,
      "token_acc": 0.8852734179451692,
      "train_speed(iter/s)": 0.032302
    },
    {
      "epoch": 0.409049072324759,
      "grad_norm": 0.11222491413354874,
      "learning_rate": 0.00020617910453966438,
      "loss": 0.3725607693195343,
      "memory(GiB)": 78.33,
      "step": 2111,
      "token_acc": 0.8908452798762715,
      "train_speed(iter/s)": 0.032304
    },
    {
      "epoch": 0.40924284261008576,
      "grad_norm": 0.10845697671175003,
      "learning_rate": 0.00020608995792482102,
      "loss": 0.36426377296447754,
      "memory(GiB)": 78.33,
      "step": 2112,
      "token_acc": 0.8920298507462686,
      "train_speed(iter/s)": 0.032305
    },
    {
      "epoch": 0.4094366128954125,
      "grad_norm": 0.11745316535234451,
      "learning_rate": 0.00020600078827232469,
      "loss": 0.44262993335723877,
      "memory(GiB)": 78.33,
      "step": 2113,
      "token_acc": 0.8721979182380449,
      "train_speed(iter/s)": 0.032307
    },
    {
      "epoch": 0.40963038318073924,
      "grad_norm": 0.10488930344581604,
      "learning_rate": 0.00020591159561879991,
      "loss": 0.3662855923175812,
      "memory(GiB)": 78.33,
      "step": 2114,
      "token_acc": 0.8902525438621355,
      "train_speed(iter/s)": 0.032308
    },
    {
      "epoch": 0.409824153466066,
      "grad_norm": 0.10289178043603897,
      "learning_rate": 0.00020582238000088033,
      "loss": 0.37403732538223267,
      "memory(GiB)": 78.33,
      "step": 2115,
      "token_acc": 0.8886182232818493,
      "train_speed(iter/s)": 0.032309
    },
    {
      "epoch": 0.41001792375139273,
      "grad_norm": 0.10993564128875732,
      "learning_rate": 0.0002057331414552093,
      "loss": 0.400392085313797,
      "memory(GiB)": 78.33,
      "step": 2116,
      "token_acc": 0.882896215297761,
      "train_speed(iter/s)": 0.032311
    },
    {
      "epoch": 0.4102116940367195,
      "grad_norm": 0.10764973610639572,
      "learning_rate": 0.00020564388001843945,
      "loss": 0.3797903060913086,
      "memory(GiB)": 78.33,
      "step": 2117,
      "token_acc": 0.8884940778341793,
      "train_speed(iter/s)": 0.032312
    },
    {
      "epoch": 0.4104054643220462,
      "grad_norm": 0.11092360317707062,
      "learning_rate": 0.00020555459572723294,
      "loss": 0.3895651698112488,
      "memory(GiB)": 78.33,
      "step": 2118,
      "token_acc": 0.8856181915590213,
      "train_speed(iter/s)": 0.032313
    },
    {
      "epoch": 0.41059923460737296,
      "grad_norm": 0.11945401877164841,
      "learning_rate": 0.00020546528861826107,
      "loss": 0.40205317735671997,
      "memory(GiB)": 78.33,
      "step": 2119,
      "token_acc": 0.8840717802127998,
      "train_speed(iter/s)": 0.032315
    },
    {
      "epoch": 0.4107930048926997,
      "grad_norm": 0.1141401007771492,
      "learning_rate": 0.0002053759587282048,
      "loss": 0.42232364416122437,
      "memory(GiB)": 78.33,
      "step": 2120,
      "token_acc": 0.8797556077277742,
      "train_speed(iter/s)": 0.032316
    },
    {
      "epoch": 0.41098677517802645,
      "grad_norm": 0.10188201814889908,
      "learning_rate": 0.00020528660609375426,
      "loss": 0.36200839281082153,
      "memory(GiB)": 78.33,
      "step": 2121,
      "token_acc": 0.8925009859710407,
      "train_speed(iter/s)": 0.032317
    },
    {
      "epoch": 0.4111805454633532,
      "grad_norm": 0.09827426820993423,
      "learning_rate": 0.000205197230751609,
      "loss": 0.3723769187927246,
      "memory(GiB)": 78.33,
      "step": 2122,
      "token_acc": 0.8900832517140059,
      "train_speed(iter/s)": 0.032318
    },
    {
      "epoch": 0.41137431574867994,
      "grad_norm": 0.1049884557723999,
      "learning_rate": 0.00020510783273847778,
      "loss": 0.37664487957954407,
      "memory(GiB)": 78.33,
      "step": 2123,
      "token_acc": 0.8893548130258234,
      "train_speed(iter/s)": 0.032319
    },
    {
      "epoch": 0.4115680860340067,
      "grad_norm": 0.10434847325086594,
      "learning_rate": 0.00020501841209107896,
      "loss": 0.3857555091381073,
      "memory(GiB)": 78.33,
      "step": 2124,
      "token_acc": 0.8891336270190896,
      "train_speed(iter/s)": 0.032321
    },
    {
      "epoch": 0.41176185631933343,
      "grad_norm": 0.10164911299943924,
      "learning_rate": 0.00020492896884613987,
      "loss": 0.3675006628036499,
      "memory(GiB)": 78.33,
      "step": 2125,
      "token_acc": 0.8910501783714814,
      "train_speed(iter/s)": 0.032322
    },
    {
      "epoch": 0.4119556266046602,
      "grad_norm": 0.10134454071521759,
      "learning_rate": 0.00020483950304039724,
      "loss": 0.3476465940475464,
      "memory(GiB)": 78.33,
      "step": 2126,
      "token_acc": 0.896970502258836,
      "train_speed(iter/s)": 0.032323
    },
    {
      "epoch": 0.4121493968899869,
      "grad_norm": 0.10433205217123032,
      "learning_rate": 0.00020475001471059712,
      "loss": 0.3762246072292328,
      "memory(GiB)": 78.33,
      "step": 2127,
      "token_acc": 0.8882155392268263,
      "train_speed(iter/s)": 0.032324
    },
    {
      "epoch": 0.41234316717531366,
      "grad_norm": 0.10764237493276596,
      "learning_rate": 0.0002046605038934948,
      "loss": 0.3889080286026001,
      "memory(GiB)": 78.33,
      "step": 2128,
      "token_acc": 0.8841791697411931,
      "train_speed(iter/s)": 0.032326
    },
    {
      "epoch": 0.4125369374606404,
      "grad_norm": 0.10777831077575684,
      "learning_rate": 0.00020457097062585473,
      "loss": 0.38341259956359863,
      "memory(GiB)": 78.33,
      "step": 2129,
      "token_acc": 0.8884803921568627,
      "train_speed(iter/s)": 0.032327
    },
    {
      "epoch": 0.41273070774596715,
      "grad_norm": 0.15417581796646118,
      "learning_rate": 0.00020448141494445066,
      "loss": 0.37628644704818726,
      "memory(GiB)": 78.33,
      "step": 2130,
      "token_acc": 0.8896055119537908,
      "train_speed(iter/s)": 0.032328
    },
    {
      "epoch": 0.4129244780312939,
      "grad_norm": 0.1092265397310257,
      "learning_rate": 0.00020439183688606547,
      "loss": 0.367519348859787,
      "memory(GiB)": 78.33,
      "step": 2131,
      "token_acc": 0.8922348315610759,
      "train_speed(iter/s)": 0.03233
    },
    {
      "epoch": 0.41311824831662064,
      "grad_norm": 0.10159304738044739,
      "learning_rate": 0.0002043022364874913,
      "loss": 0.33829817175865173,
      "memory(GiB)": 78.33,
      "step": 2132,
      "token_acc": 0.8995331695331695,
      "train_speed(iter/s)": 0.032331
    },
    {
      "epoch": 0.4133120186019474,
      "grad_norm": 0.1057777926325798,
      "learning_rate": 0.00020421261378552948,
      "loss": 0.36473432183265686,
      "memory(GiB)": 78.33,
      "step": 2133,
      "token_acc": 0.8916222307734594,
      "train_speed(iter/s)": 0.032332
    },
    {
      "epoch": 0.41350578888727413,
      "grad_norm": 0.10533451288938522,
      "learning_rate": 0.00020412296881699039,
      "loss": 0.3586021065711975,
      "memory(GiB)": 78.33,
      "step": 2134,
      "token_acc": 0.8953077991205739,
      "train_speed(iter/s)": 0.032333
    },
    {
      "epoch": 0.4136995591726009,
      "grad_norm": 0.10619760304689407,
      "learning_rate": 0.00020403330161869373,
      "loss": 0.35481902956962585,
      "memory(GiB)": 78.33,
      "step": 2135,
      "token_acc": 0.8953304521977443,
      "train_speed(iter/s)": 0.032335
    },
    {
      "epoch": 0.4138933294579276,
      "grad_norm": 0.11591517180204391,
      "learning_rate": 0.0002039436122274681,
      "loss": 0.40828073024749756,
      "memory(GiB)": 78.33,
      "step": 2136,
      "token_acc": 0.8834233806329039,
      "train_speed(iter/s)": 0.032336
    },
    {
      "epoch": 0.41408709974325436,
      "grad_norm": 0.11356504261493683,
      "learning_rate": 0.00020385390068015146,
      "loss": 0.3991680443286896,
      "memory(GiB)": 78.33,
      "step": 2137,
      "token_acc": 0.8845186434119241,
      "train_speed(iter/s)": 0.032337
    },
    {
      "epoch": 0.4142808700285811,
      "grad_norm": 0.11383282393217087,
      "learning_rate": 0.00020376416701359067,
      "loss": 0.41105103492736816,
      "memory(GiB)": 78.33,
      "step": 2138,
      "token_acc": 0.8808600337268128,
      "train_speed(iter/s)": 0.032338
    },
    {
      "epoch": 0.41447464031390785,
      "grad_norm": 0.09987689554691315,
      "learning_rate": 0.00020367441126464177,
      "loss": 0.3558513820171356,
      "memory(GiB)": 78.33,
      "step": 2139,
      "token_acc": 0.8945576407506702,
      "train_speed(iter/s)": 0.03234
    },
    {
      "epoch": 0.4146684105992346,
      "grad_norm": 0.11728887259960175,
      "learning_rate": 0.00020358463347016988,
      "loss": 0.41319242119789124,
      "memory(GiB)": 78.33,
      "step": 2140,
      "token_acc": 0.880601774015885,
      "train_speed(iter/s)": 0.032341
    },
    {
      "epoch": 0.41486218088456134,
      "grad_norm": 0.10223786532878876,
      "learning_rate": 0.0002034948336670492,
      "loss": 0.3577488362789154,
      "memory(GiB)": 78.33,
      "step": 2141,
      "token_acc": 0.892794648592376,
      "train_speed(iter/s)": 0.032342
    },
    {
      "epoch": 0.4150559511698881,
      "grad_norm": 0.10343615710735321,
      "learning_rate": 0.00020340501189216285,
      "loss": 0.36109933257102966,
      "memory(GiB)": 78.33,
      "step": 2142,
      "token_acc": 0.8946854716704895,
      "train_speed(iter/s)": 0.032343
    },
    {
      "epoch": 0.4152497214552148,
      "grad_norm": 0.10884758830070496,
      "learning_rate": 0.000203315168182403,
      "loss": 0.38284680247306824,
      "memory(GiB)": 78.33,
      "step": 2143,
      "token_acc": 0.8858148616687751,
      "train_speed(iter/s)": 0.032345
    },
    {
      "epoch": 0.41544349174054157,
      "grad_norm": 0.10763585567474365,
      "learning_rate": 0.00020322530257467104,
      "loss": 0.3719366788864136,
      "memory(GiB)": 78.33,
      "step": 2144,
      "token_acc": 0.8914604948124502,
      "train_speed(iter/s)": 0.032346
    },
    {
      "epoch": 0.4156372620258683,
      "grad_norm": 0.10268665105104446,
      "learning_rate": 0.00020313541510587707,
      "loss": 0.3508341312408447,
      "memory(GiB)": 78.33,
      "step": 2145,
      "token_acc": 0.8950760245695297,
      "train_speed(iter/s)": 0.032347
    },
    {
      "epoch": 0.41583103231119506,
      "grad_norm": 0.10579564422369003,
      "learning_rate": 0.00020304550581294026,
      "loss": 0.39762815833091736,
      "memory(GiB)": 78.33,
      "step": 2146,
      "token_acc": 0.8849706763321503,
      "train_speed(iter/s)": 0.032348
    },
    {
      "epoch": 0.4160248025965218,
      "grad_norm": 0.09803339838981628,
      "learning_rate": 0.00020295557473278886,
      "loss": 0.33986738324165344,
      "memory(GiB)": 78.33,
      "step": 2147,
      "token_acc": 0.8999627421758569,
      "train_speed(iter/s)": 0.03235
    },
    {
      "epoch": 0.41621857288184855,
      "grad_norm": 0.1106083020567894,
      "learning_rate": 0.00020286562190235998,
      "loss": 0.3605062961578369,
      "memory(GiB)": 78.33,
      "step": 2148,
      "token_acc": 0.8942120900468518,
      "train_speed(iter/s)": 0.032351
    },
    {
      "epoch": 0.4164123431671753,
      "grad_norm": 0.11620527505874634,
      "learning_rate": 0.00020277564735859957,
      "loss": 0.42198172211647034,
      "memory(GiB)": 78.33,
      "step": 2149,
      "token_acc": 0.8788444418918531,
      "train_speed(iter/s)": 0.032352
    },
    {
      "epoch": 0.41660611345250204,
      "grad_norm": 0.0997370034456253,
      "learning_rate": 0.0002026856511384627,
      "loss": 0.33660730719566345,
      "memory(GiB)": 78.33,
      "step": 2150,
      "token_acc": 0.9009054193173719,
      "train_speed(iter/s)": 0.032354
    },
    {
      "epoch": 0.4167998837378288,
      "grad_norm": 0.10104697942733765,
      "learning_rate": 0.00020259563327891316,
      "loss": 0.3682084381580353,
      "memory(GiB)": 78.33,
      "step": 2151,
      "token_acc": 0.8905543542362886,
      "train_speed(iter/s)": 0.032355
    },
    {
      "epoch": 0.4169936540231555,
      "grad_norm": 0.10348115861415863,
      "learning_rate": 0.00020250559381692373,
      "loss": 0.35385698080062866,
      "memory(GiB)": 78.33,
      "step": 2152,
      "token_acc": 0.8960956832554303,
      "train_speed(iter/s)": 0.032356
    },
    {
      "epoch": 0.41718742430848227,
      "grad_norm": 0.10488869249820709,
      "learning_rate": 0.00020241553278947604,
      "loss": 0.36309656500816345,
      "memory(GiB)": 78.33,
      "step": 2153,
      "token_acc": 0.8921550221268607,
      "train_speed(iter/s)": 0.032357
    },
    {
      "epoch": 0.417381194593809,
      "grad_norm": 0.11151447147130966,
      "learning_rate": 0.00020232545023356058,
      "loss": 0.37561148405075073,
      "memory(GiB)": 78.33,
      "step": 2154,
      "token_acc": 0.889549997111663,
      "train_speed(iter/s)": 0.032359
    },
    {
      "epoch": 0.41757496487913576,
      "grad_norm": 0.11074910312891006,
      "learning_rate": 0.0002022353461861767,
      "loss": 0.3915635049343109,
      "memory(GiB)": 78.33,
      "step": 2155,
      "token_acc": 0.8837005402042029,
      "train_speed(iter/s)": 0.03236
    },
    {
      "epoch": 0.41776873516446256,
      "grad_norm": 0.10543368011713028,
      "learning_rate": 0.00020214522068433247,
      "loss": 0.34513017535209656,
      "memory(GiB)": 78.33,
      "step": 2156,
      "token_acc": 0.8980891719745223,
      "train_speed(iter/s)": 0.032361
    },
    {
      "epoch": 0.4179625054497893,
      "grad_norm": 0.0990806296467781,
      "learning_rate": 0.00020205507376504494,
      "loss": 0.34623822569847107,
      "memory(GiB)": 78.33,
      "step": 2157,
      "token_acc": 0.8984733024327651,
      "train_speed(iter/s)": 0.032362
    },
    {
      "epoch": 0.41815627573511605,
      "grad_norm": 0.0954173356294632,
      "learning_rate": 0.00020196490546533987,
      "loss": 0.3334534168243408,
      "memory(GiB)": 78.33,
      "step": 2158,
      "token_acc": 0.9018397113381105,
      "train_speed(iter/s)": 0.032363
    },
    {
      "epoch": 0.4183500460204428,
      "grad_norm": 0.10484931617975235,
      "learning_rate": 0.0002018747158222517,
      "loss": 0.33561134338378906,
      "memory(GiB)": 78.33,
      "step": 2159,
      "token_acc": 0.9003498385360603,
      "train_speed(iter/s)": 0.032365
    },
    {
      "epoch": 0.41854381630576953,
      "grad_norm": 0.10665947943925858,
      "learning_rate": 0.00020178450487282385,
      "loss": 0.3735751509666443,
      "memory(GiB)": 78.33,
      "step": 2160,
      "token_acc": 0.892504140620826,
      "train_speed(iter/s)": 0.032366
    },
    {
      "epoch": 0.4187375865910963,
      "grad_norm": 0.10096178948879242,
      "learning_rate": 0.00020169427265410837,
      "loss": 0.3531975746154785,
      "memory(GiB)": 78.33,
      "step": 2161,
      "token_acc": 0.8952592753308743,
      "train_speed(iter/s)": 0.032367
    },
    {
      "epoch": 0.418931356876423,
      "grad_norm": 0.11227719485759735,
      "learning_rate": 0.00020160401920316597,
      "loss": 0.3715410828590393,
      "memory(GiB)": 78.33,
      "step": 2162,
      "token_acc": 0.8928775907477384,
      "train_speed(iter/s)": 0.032368
    },
    {
      "epoch": 0.41912512716174977,
      "grad_norm": 0.10699176788330078,
      "learning_rate": 0.0002015137445570663,
      "loss": 0.3732505738735199,
      "memory(GiB)": 78.33,
      "step": 2163,
      "token_acc": 0.8906586310804994,
      "train_speed(iter/s)": 0.03237
    },
    {
      "epoch": 0.4193188974470765,
      "grad_norm": 0.0976463332772255,
      "learning_rate": 0.0002014234487528875,
      "loss": 0.37260702252388,
      "memory(GiB)": 78.33,
      "step": 2164,
      "token_acc": 0.8918560748130923,
      "train_speed(iter/s)": 0.032371
    },
    {
      "epoch": 0.41951266773240325,
      "grad_norm": 0.10294061154127121,
      "learning_rate": 0.00020133313182771646,
      "loss": 0.37537047266960144,
      "memory(GiB)": 78.33,
      "step": 2165,
      "token_acc": 0.8896890512075307,
      "train_speed(iter/s)": 0.032372
    },
    {
      "epoch": 0.41970643801773,
      "grad_norm": 0.10132710635662079,
      "learning_rate": 0.00020124279381864883,
      "loss": 0.35512280464172363,
      "memory(GiB)": 78.33,
      "step": 2166,
      "token_acc": 0.8946123521681998,
      "train_speed(iter/s)": 0.032373
    },
    {
      "epoch": 0.41990020830305674,
      "grad_norm": 0.1205829605460167,
      "learning_rate": 0.00020115243476278883,
      "loss": 0.4342115521430969,
      "memory(GiB)": 78.33,
      "step": 2167,
      "token_acc": 0.8729766390354182,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.4200939785883835,
      "grad_norm": 0.10426265001296997,
      "learning_rate": 0.00020106205469724937,
      "loss": 0.36752408742904663,
      "memory(GiB)": 78.33,
      "step": 2168,
      "token_acc": 0.8942200862982198,
      "train_speed(iter/s)": 0.032376
    },
    {
      "epoch": 0.42028774887371023,
      "grad_norm": 0.10518850386142731,
      "learning_rate": 0.00020097165365915188,
      "loss": 0.36646583676338196,
      "memory(GiB)": 78.33,
      "step": 2169,
      "token_acc": 0.8950598184818482,
      "train_speed(iter/s)": 0.032377
    },
    {
      "epoch": 0.420481519159037,
      "grad_norm": 0.097692109644413,
      "learning_rate": 0.00020088123168562663,
      "loss": 0.31673404574394226,
      "memory(GiB)": 78.33,
      "step": 2170,
      "token_acc": 0.9040827884466869,
      "train_speed(iter/s)": 0.032378
    },
    {
      "epoch": 0.4206752894443637,
      "grad_norm": 0.10072191804647446,
      "learning_rate": 0.00020079078881381232,
      "loss": 0.3458814322948456,
      "memory(GiB)": 78.33,
      "step": 2171,
      "token_acc": 0.8972006834012354,
      "train_speed(iter/s)": 0.03238
    },
    {
      "epoch": 0.42086905972969046,
      "grad_norm": 0.10148821771144867,
      "learning_rate": 0.00020070032508085617,
      "loss": 0.362411230802536,
      "memory(GiB)": 78.33,
      "step": 2172,
      "token_acc": 0.893191234333064,
      "train_speed(iter/s)": 0.032381
    },
    {
      "epoch": 0.4210628300150172,
      "grad_norm": 0.09655621647834778,
      "learning_rate": 0.0002006098405239142,
      "loss": 0.36690768599510193,
      "memory(GiB)": 78.33,
      "step": 2173,
      "token_acc": 0.8898027945694182,
      "train_speed(iter/s)": 0.032382
    },
    {
      "epoch": 0.42125660030034395,
      "grad_norm": 0.10610742121934891,
      "learning_rate": 0.00020051933518015077,
      "loss": 0.388561487197876,
      "memory(GiB)": 78.33,
      "step": 2174,
      "token_acc": 0.8857004153406333,
      "train_speed(iter/s)": 0.032383
    },
    {
      "epoch": 0.4214503705856707,
      "grad_norm": 0.11180911213159561,
      "learning_rate": 0.00020042880908673888,
      "loss": 0.38997843861579895,
      "memory(GiB)": 78.33,
      "step": 2175,
      "token_acc": 0.8855060034305318,
      "train_speed(iter/s)": 0.032384
    },
    {
      "epoch": 0.42164414087099744,
      "grad_norm": 0.11040447652339935,
      "learning_rate": 0.00020033826228085997,
      "loss": 0.3600209355354309,
      "memory(GiB)": 78.33,
      "step": 2176,
      "token_acc": 0.8954944743553415,
      "train_speed(iter/s)": 0.032386
    },
    {
      "epoch": 0.4218379111563242,
      "grad_norm": 0.11115849018096924,
      "learning_rate": 0.0002002476947997042,
      "loss": 0.3961622714996338,
      "memory(GiB)": 78.33,
      "step": 2177,
      "token_acc": 0.8841036617111124,
      "train_speed(iter/s)": 0.032387
    },
    {
      "epoch": 0.42203168144165093,
      "grad_norm": 0.10502826422452927,
      "learning_rate": 0.0002001571066804699,
      "loss": 0.38119229674339294,
      "memory(GiB)": 78.33,
      "step": 2178,
      "token_acc": 0.8862857463902466,
      "train_speed(iter/s)": 0.032388
    },
    {
      "epoch": 0.4222254517269777,
      "grad_norm": 0.11460091918706894,
      "learning_rate": 0.00020006649796036412,
      "loss": 0.4165613353252411,
      "memory(GiB)": 78.33,
      "step": 2179,
      "token_acc": 0.8790522573127372,
      "train_speed(iter/s)": 0.032389
    },
    {
      "epoch": 0.4224192220123044,
      "grad_norm": 0.09722710400819778,
      "learning_rate": 0.0001999758686766023,
      "loss": 0.3486678898334503,
      "memory(GiB)": 78.33,
      "step": 2180,
      "token_acc": 0.8953782314223028,
      "train_speed(iter/s)": 0.03239
    },
    {
      "epoch": 0.42261299229763116,
      "grad_norm": 0.10160472989082336,
      "learning_rate": 0.0001998852188664083,
      "loss": 0.3687483072280884,
      "memory(GiB)": 78.33,
      "step": 2181,
      "token_acc": 0.8916645094486151,
      "train_speed(iter/s)": 0.032392
    },
    {
      "epoch": 0.4228067625829579,
      "grad_norm": 0.11015883833169937,
      "learning_rate": 0.00019979454856701442,
      "loss": 0.36141785979270935,
      "memory(GiB)": 78.33,
      "step": 2182,
      "token_acc": 0.895195164429874,
      "train_speed(iter/s)": 0.032393
    },
    {
      "epoch": 0.42300053286828465,
      "grad_norm": 0.10014893114566803,
      "learning_rate": 0.00019970385781566146,
      "loss": 0.3551277220249176,
      "memory(GiB)": 78.33,
      "step": 2183,
      "token_acc": 0.8938043448633031,
      "train_speed(iter/s)": 0.032395
    },
    {
      "epoch": 0.4231943031536114,
      "grad_norm": 0.08740631490945816,
      "learning_rate": 0.00019961314664959849,
      "loss": 0.3233758211135864,
      "memory(GiB)": 78.33,
      "step": 2184,
      "token_acc": 0.9035294117647059,
      "train_speed(iter/s)": 0.032396
    },
    {
      "epoch": 0.42338807343893814,
      "grad_norm": 0.10040253400802612,
      "learning_rate": 0.00019952241510608302,
      "loss": 0.34599989652633667,
      "memory(GiB)": 78.33,
      "step": 2185,
      "token_acc": 0.8977024740700782,
      "train_speed(iter/s)": 0.032397
    },
    {
      "epoch": 0.4235818437242649,
      "grad_norm": 0.10159041732549667,
      "learning_rate": 0.00019943166322238095,
      "loss": 0.34493720531463623,
      "memory(GiB)": 78.33,
      "step": 2186,
      "token_acc": 0.8977762408895037,
      "train_speed(iter/s)": 0.032398
    },
    {
      "epoch": 0.42377561400959163,
      "grad_norm": 0.11545497179031372,
      "learning_rate": 0.00019934089103576652,
      "loss": 0.3942737877368927,
      "memory(GiB)": 78.33,
      "step": 2187,
      "token_acc": 0.8845215505557061,
      "train_speed(iter/s)": 0.0324
    },
    {
      "epoch": 0.42396938429491837,
      "grad_norm": 0.10246077924966812,
      "learning_rate": 0.00019925009858352233,
      "loss": 0.37554022669792175,
      "memory(GiB)": 78.33,
      "step": 2188,
      "token_acc": 0.8895408954606556,
      "train_speed(iter/s)": 0.032401
    },
    {
      "epoch": 0.4241631545802451,
      "grad_norm": 0.12075801938772202,
      "learning_rate": 0.00019915928590293918,
      "loss": 0.40932536125183105,
      "memory(GiB)": 78.33,
      "step": 2189,
      "token_acc": 0.8816431763492609,
      "train_speed(iter/s)": 0.032402
    },
    {
      "epoch": 0.42435692486557186,
      "grad_norm": 0.12044385075569153,
      "learning_rate": 0.00019906845303131643,
      "loss": 0.4146193861961365,
      "memory(GiB)": 78.33,
      "step": 2190,
      "token_acc": 0.8788535074955163,
      "train_speed(iter/s)": 0.032403
    },
    {
      "epoch": 0.4245506951508986,
      "grad_norm": 0.10225894302129745,
      "learning_rate": 0.00019897760000596145,
      "loss": 0.3517759144306183,
      "memory(GiB)": 78.33,
      "step": 2191,
      "token_acc": 0.8963014924651073,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.42474446543622535,
      "grad_norm": 0.10850612074136734,
      "learning_rate": 0.00019888672686419005,
      "loss": 0.34386181831359863,
      "memory(GiB)": 78.33,
      "step": 2192,
      "token_acc": 0.898326711713018,
      "train_speed(iter/s)": 0.032406
    },
    {
      "epoch": 0.4249382357215521,
      "grad_norm": 0.10484492033720016,
      "learning_rate": 0.0001987958336433263,
      "loss": 0.36306411027908325,
      "memory(GiB)": 78.33,
      "step": 2193,
      "token_acc": 0.8922604656216891,
      "train_speed(iter/s)": 0.032407
    },
    {
      "epoch": 0.42513200600687884,
      "grad_norm": 0.09952452033758163,
      "learning_rate": 0.00019870492038070252,
      "loss": 0.3434096574783325,
      "memory(GiB)": 78.33,
      "step": 2194,
      "token_acc": 0.8980537301407457,
      "train_speed(iter/s)": 0.032408
    },
    {
      "epoch": 0.4253257762922056,
      "grad_norm": 0.11556005477905273,
      "learning_rate": 0.00019861398711365917,
      "loss": 0.3726232945919037,
      "memory(GiB)": 78.33,
      "step": 2195,
      "token_acc": 0.890572481838881,
      "train_speed(iter/s)": 0.032409
    },
    {
      "epoch": 0.4255195465775323,
      "grad_norm": 0.0965765044093132,
      "learning_rate": 0.00019852303387954496,
      "loss": 0.34968358278274536,
      "memory(GiB)": 78.33,
      "step": 2196,
      "token_acc": 0.8960782916458824,
      "train_speed(iter/s)": 0.03241
    },
    {
      "epoch": 0.42571331686285907,
      "grad_norm": 0.11270653456449509,
      "learning_rate": 0.00019843206071571692,
      "loss": 0.37671494483947754,
      "memory(GiB)": 78.33,
      "step": 2197,
      "token_acc": 0.8884828552286984,
      "train_speed(iter/s)": 0.032412
    },
    {
      "epoch": 0.4259070871481858,
      "grad_norm": 0.10049509257078171,
      "learning_rate": 0.0001983410676595401,
      "loss": 0.36629050970077515,
      "memory(GiB)": 78.33,
      "step": 2198,
      "token_acc": 0.893262368452885,
      "train_speed(iter/s)": 0.032413
    },
    {
      "epoch": 0.42610085743351256,
      "grad_norm": 0.10621780902147293,
      "learning_rate": 0.0001982500547483878,
      "loss": 0.3855476677417755,
      "memory(GiB)": 78.33,
      "step": 2199,
      "token_acc": 0.8878023674729799,
      "train_speed(iter/s)": 0.032414
    },
    {
      "epoch": 0.4262946277188393,
      "grad_norm": 0.10796981304883957,
      "learning_rate": 0.00019815902201964153,
      "loss": 0.34390783309936523,
      "memory(GiB)": 78.33,
      "step": 2200,
      "token_acc": 0.8990232532326988,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.42648839800416605,
      "grad_norm": 0.11975818127393723,
      "learning_rate": 0.00019806796951069087,
      "loss": 0.3976423442363739,
      "memory(GiB)": 78.33,
      "step": 2201,
      "token_acc": 0.885062679353572,
      "train_speed(iter/s)": 0.032411
    },
    {
      "epoch": 0.4266821682894928,
      "grad_norm": 0.10373754054307938,
      "learning_rate": 0.00019797689725893337,
      "loss": 0.3727421462535858,
      "memory(GiB)": 78.33,
      "step": 2202,
      "token_acc": 0.8891960318280459,
      "train_speed(iter/s)": 0.032412
    },
    {
      "epoch": 0.42687593857481954,
      "grad_norm": 0.10288981348276138,
      "learning_rate": 0.00019788580530177507,
      "loss": 0.35971593856811523,
      "memory(GiB)": 78.33,
      "step": 2203,
      "token_acc": 0.8920195439739413,
      "train_speed(iter/s)": 0.032413
    },
    {
      "epoch": 0.4270697088601463,
      "grad_norm": 0.10778756439685822,
      "learning_rate": 0.0001977946936766298,
      "loss": 0.38950392603874207,
      "memory(GiB)": 78.33,
      "step": 2204,
      "token_acc": 0.8854846286340623,
      "train_speed(iter/s)": 0.032415
    },
    {
      "epoch": 0.427263479145473,
      "grad_norm": 0.1137860119342804,
      "learning_rate": 0.0001977035624209195,
      "loss": 0.3468484580516815,
      "memory(GiB)": 78.33,
      "step": 2205,
      "token_acc": 0.8975911374468634,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.42745724943079977,
      "grad_norm": 0.11108643561601639,
      "learning_rate": 0.00019761241157207428,
      "loss": 0.3628512918949127,
      "memory(GiB)": 78.33,
      "step": 2206,
      "token_acc": 0.8934348239771646,
      "train_speed(iter/s)": 0.032417
    },
    {
      "epoch": 0.4276510197161265,
      "grad_norm": 0.09750779718160629,
      "learning_rate": 0.00019752124116753224,
      "loss": 0.3502158522605896,
      "memory(GiB)": 78.33,
      "step": 2207,
      "token_acc": 0.8962943407094514,
      "train_speed(iter/s)": 0.032418
    },
    {
      "epoch": 0.42784479000145326,
      "grad_norm": 0.09157220274209976,
      "learning_rate": 0.0001974300512447395,
      "loss": 0.33457863330841064,
      "memory(GiB)": 78.33,
      "step": 2208,
      "token_acc": 0.900385138196647,
      "train_speed(iter/s)": 0.032419
    },
    {
      "epoch": 0.42803856028678,
      "grad_norm": 0.11325064301490784,
      "learning_rate": 0.0001973388418411502,
      "loss": 0.37056490778923035,
      "memory(GiB)": 78.33,
      "step": 2209,
      "token_acc": 0.8906734699794161,
      "train_speed(iter/s)": 0.03242
    },
    {
      "epoch": 0.42823233057210675,
      "grad_norm": 0.11709318310022354,
      "learning_rate": 0.00019724761299422654,
      "loss": 0.40491345524787903,
      "memory(GiB)": 78.33,
      "step": 2210,
      "token_acc": 0.8828931933278018,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.4284261008574335,
      "grad_norm": 0.10216681659221649,
      "learning_rate": 0.00019715636474143864,
      "loss": 0.38466590642929077,
      "memory(GiB)": 78.33,
      "step": 2211,
      "token_acc": 0.8875497597803707,
      "train_speed(iter/s)": 0.032423
    },
    {
      "epoch": 0.42861987114276023,
      "grad_norm": 0.1109510064125061,
      "learning_rate": 0.00019706509712026456,
      "loss": 0.38734912872314453,
      "memory(GiB)": 78.33,
      "step": 2212,
      "token_acc": 0.8852111060119321,
      "train_speed(iter/s)": 0.032424
    },
    {
      "epoch": 0.428813641428087,
      "grad_norm": 0.10640700161457062,
      "learning_rate": 0.00019697381016819043,
      "loss": 0.3618239462375641,
      "memory(GiB)": 78.33,
      "step": 2213,
      "token_acc": 0.8929901199097012,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.4290074117134137,
      "grad_norm": 0.13030792772769928,
      "learning_rate": 0.00019688250392271026,
      "loss": 0.3900720775127411,
      "memory(GiB)": 78.33,
      "step": 2214,
      "token_acc": 0.8874647759294609,
      "train_speed(iter/s)": 0.032427
    },
    {
      "epoch": 0.42920118199874047,
      "grad_norm": 0.09482412040233612,
      "learning_rate": 0.00019679117842132592,
      "loss": 0.35719773173332214,
      "memory(GiB)": 78.33,
      "step": 2215,
      "token_acc": 0.8954372137062723,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.4293949522840672,
      "grad_norm": 0.10285639762878418,
      "learning_rate": 0.00019669983370154722,
      "loss": 0.3917454481124878,
      "memory(GiB)": 78.33,
      "step": 2216,
      "token_acc": 0.8865480649188514,
      "train_speed(iter/s)": 0.032429
    },
    {
      "epoch": 0.429588722569394,
      "grad_norm": 0.11368583887815475,
      "learning_rate": 0.000196608469800892,
      "loss": 0.42137694358825684,
      "memory(GiB)": 78.33,
      "step": 2217,
      "token_acc": 0.8776351492381549,
      "train_speed(iter/s)": 0.03243
    },
    {
      "epoch": 0.42978249285472075,
      "grad_norm": 0.10096923261880875,
      "learning_rate": 0.0001965170867568858,
      "loss": 0.3742446005344391,
      "memory(GiB)": 78.33,
      "step": 2218,
      "token_acc": 0.8895205325753609,
      "train_speed(iter/s)": 0.032431
    },
    {
      "epoch": 0.4299762631400475,
      "grad_norm": 0.10487914085388184,
      "learning_rate": 0.00019642568460706214,
      "loss": 0.36482223868370056,
      "memory(GiB)": 78.33,
      "step": 2219,
      "token_acc": 0.8928659320226409,
      "train_speed(iter/s)": 0.032432
    },
    {
      "epoch": 0.43017003342537424,
      "grad_norm": 0.2587502598762512,
      "learning_rate": 0.00019633426338896227,
      "loss": 0.36105775833129883,
      "memory(GiB)": 78.33,
      "step": 2220,
      "token_acc": 0.8932460356156114,
      "train_speed(iter/s)": 0.032433
    },
    {
      "epoch": 0.430363803710701,
      "grad_norm": 0.10027684271335602,
      "learning_rate": 0.0001962428231401354,
      "loss": 0.3383048176765442,
      "memory(GiB)": 78.33,
      "step": 2221,
      "token_acc": 0.8997930908043268,
      "train_speed(iter/s)": 0.032435
    },
    {
      "epoch": 0.43055757399602773,
      "grad_norm": 0.11608999222517014,
      "learning_rate": 0.00019615136389813847,
      "loss": 0.38931986689567566,
      "memory(GiB)": 78.33,
      "step": 2222,
      "token_acc": 0.8857191859135605,
      "train_speed(iter/s)": 0.032436
    },
    {
      "epoch": 0.4307513442813545,
      "grad_norm": 0.10588336735963821,
      "learning_rate": 0.00019605988570053622,
      "loss": 0.354093462228775,
      "memory(GiB)": 78.33,
      "step": 2223,
      "token_acc": 0.8957208040448884,
      "train_speed(iter/s)": 0.032437
    },
    {
      "epoch": 0.4309451145666812,
      "grad_norm": 0.16060945391654968,
      "learning_rate": 0.0001959683885849013,
      "loss": 0.3510138690471649,
      "memory(GiB)": 78.33,
      "step": 2224,
      "token_acc": 0.8974735682241519,
      "train_speed(iter/s)": 0.032438
    },
    {
      "epoch": 0.43113888485200796,
      "grad_norm": 0.1098427101969719,
      "learning_rate": 0.00019587687258881391,
      "loss": 0.3766689896583557,
      "memory(GiB)": 78.33,
      "step": 2225,
      "token_acc": 0.8884950490521164,
      "train_speed(iter/s)": 0.032439
    },
    {
      "epoch": 0.4313326551373347,
      "grad_norm": 0.10366590321063995,
      "learning_rate": 0.00019578533774986217,
      "loss": 0.3672398626804352,
      "memory(GiB)": 78.33,
      "step": 2226,
      "token_acc": 0.8924895345973898,
      "train_speed(iter/s)": 0.032441
    },
    {
      "epoch": 0.43152642542266145,
      "grad_norm": 0.11767855286598206,
      "learning_rate": 0.00019569378410564197,
      "loss": 0.3826836347579956,
      "memory(GiB)": 78.33,
      "step": 2227,
      "token_acc": 0.8886111359104841,
      "train_speed(iter/s)": 0.032442
    },
    {
      "epoch": 0.4317201957079882,
      "grad_norm": 0.0977967232465744,
      "learning_rate": 0.0001956022116937568,
      "loss": 0.3510884940624237,
      "memory(GiB)": 78.33,
      "step": 2228,
      "token_acc": 0.8963877167735799,
      "train_speed(iter/s)": 0.032443
    },
    {
      "epoch": 0.43191396599331494,
      "grad_norm": 0.09899823367595673,
      "learning_rate": 0.00019551062055181786,
      "loss": 0.3398961126804352,
      "memory(GiB)": 78.33,
      "step": 2229,
      "token_acc": 0.8998167084699981,
      "train_speed(iter/s)": 0.032444
    },
    {
      "epoch": 0.4321077362786417,
      "grad_norm": 0.1056177020072937,
      "learning_rate": 0.0001954190107174442,
      "loss": 0.38561806082725525,
      "memory(GiB)": 78.33,
      "step": 2230,
      "token_acc": 0.8863317429132753,
      "train_speed(iter/s)": 0.032445
    },
    {
      "epoch": 0.43230150656396843,
      "grad_norm": 0.09486104547977448,
      "learning_rate": 0.00019532738222826233,
      "loss": 0.3347775936126709,
      "memory(GiB)": 78.33,
      "step": 2231,
      "token_acc": 0.9010725841379837,
      "train_speed(iter/s)": 0.032447
    },
    {
      "epoch": 0.4324952768492952,
      "grad_norm": 0.1015876904129982,
      "learning_rate": 0.0001952357351219066,
      "loss": 0.3641873002052307,
      "memory(GiB)": 78.33,
      "step": 2232,
      "token_acc": 0.8932900972811175,
      "train_speed(iter/s)": 0.032448
    },
    {
      "epoch": 0.4326890471346219,
      "grad_norm": 0.10654337704181671,
      "learning_rate": 0.00019514406943601896,
      "loss": 0.37854257225990295,
      "memory(GiB)": 78.33,
      "step": 2233,
      "token_acc": 0.8883228206372311,
      "train_speed(iter/s)": 0.032449
    },
    {
      "epoch": 0.43288281741994866,
      "grad_norm": 0.09729617834091187,
      "learning_rate": 0.00019505238520824893,
      "loss": 0.342031866312027,
      "memory(GiB)": 78.33,
      "step": 2234,
      "token_acc": 0.8982541675900618,
      "train_speed(iter/s)": 0.03245
    },
    {
      "epoch": 0.4330765877052754,
      "grad_norm": 0.11875928193330765,
      "learning_rate": 0.00019496068247625361,
      "loss": 0.41712290048599243,
      "memory(GiB)": 78.33,
      "step": 2235,
      "token_acc": 0.8765791742219315,
      "train_speed(iter/s)": 0.032451
    },
    {
      "epoch": 0.43327035799060215,
      "grad_norm": 0.1124706119298935,
      "learning_rate": 0.00019486896127769794,
      "loss": 0.3886880576610565,
      "memory(GiB)": 78.33,
      "step": 2236,
      "token_acc": 0.8862263618402484,
      "train_speed(iter/s)": 0.032453
    },
    {
      "epoch": 0.4334641282759289,
      "grad_norm": 0.10713918507099152,
      "learning_rate": 0.00019477722165025418,
      "loss": 0.3783111572265625,
      "memory(GiB)": 78.33,
      "step": 2237,
      "token_acc": 0.8903876189123953,
      "train_speed(iter/s)": 0.032454
    },
    {
      "epoch": 0.43365789856125564,
      "grad_norm": 0.09784865379333496,
      "learning_rate": 0.00019468546363160224,
      "loss": 0.3712804913520813,
      "memory(GiB)": 78.33,
      "step": 2238,
      "token_acc": 0.8909272229695255,
      "train_speed(iter/s)": 0.032455
    },
    {
      "epoch": 0.4338516688465824,
      "grad_norm": 0.11485416442155838,
      "learning_rate": 0.0001945936872594297,
      "loss": 0.40341106057167053,
      "memory(GiB)": 78.33,
      "step": 2239,
      "token_acc": 0.8822438849351115,
      "train_speed(iter/s)": 0.032456
    },
    {
      "epoch": 0.4340454391319091,
      "grad_norm": 0.10090679675340652,
      "learning_rate": 0.00019450189257143148,
      "loss": 0.3727151155471802,
      "memory(GiB)": 78.33,
      "step": 2240,
      "token_acc": 0.8921047498293182,
      "train_speed(iter/s)": 0.032457
    },
    {
      "epoch": 0.43423920941723587,
      "grad_norm": 0.11711379885673523,
      "learning_rate": 0.0001944100796053102,
      "loss": 0.38696539402008057,
      "memory(GiB)": 78.33,
      "step": 2241,
      "token_acc": 0.8871316037592384,
      "train_speed(iter/s)": 0.032458
    },
    {
      "epoch": 0.4344329797025626,
      "grad_norm": 0.12041328847408295,
      "learning_rate": 0.00019431824839877582,
      "loss": 0.41463562846183777,
      "memory(GiB)": 78.33,
      "step": 2242,
      "token_acc": 0.881887840766859,
      "train_speed(iter/s)": 0.03246
    },
    {
      "epoch": 0.43462674998788936,
      "grad_norm": 0.1023503914475441,
      "learning_rate": 0.00019422639898954603,
      "loss": 0.3793856203556061,
      "memory(GiB)": 78.33,
      "step": 2243,
      "token_acc": 0.887026578933734,
      "train_speed(iter/s)": 0.032461
    },
    {
      "epoch": 0.4348205202732161,
      "grad_norm": 0.12375747412443161,
      "learning_rate": 0.00019413453141534575,
      "loss": 0.41560643911361694,
      "memory(GiB)": 78.33,
      "step": 2244,
      "token_acc": 0.878722458568798,
      "train_speed(iter/s)": 0.032462
    },
    {
      "epoch": 0.43501429055854285,
      "grad_norm": 0.09916850179433823,
      "learning_rate": 0.00019404264571390743,
      "loss": 0.360500693321228,
      "memory(GiB)": 78.33,
      "step": 2245,
      "token_acc": 0.8931444381820977,
      "train_speed(iter/s)": 0.032463
    },
    {
      "epoch": 0.4352080608438696,
      "grad_norm": 0.10854557156562805,
      "learning_rate": 0.00019395074192297106,
      "loss": 0.39239490032196045,
      "memory(GiB)": 78.33,
      "step": 2246,
      "token_acc": 0.8853895766150005,
      "train_speed(iter/s)": 0.032464
    },
    {
      "epoch": 0.43540183112919634,
      "grad_norm": 0.10064675658941269,
      "learning_rate": 0.000193858820080284,
      "loss": 0.3592490553855896,
      "memory(GiB)": 78.33,
      "step": 2247,
      "token_acc": 0.8954478346456692,
      "train_speed(iter/s)": 0.032465
    },
    {
      "epoch": 0.4355956014145231,
      "grad_norm": 0.11303189396858215,
      "learning_rate": 0.00019376688022360099,
      "loss": 0.35404136776924133,
      "memory(GiB)": 78.33,
      "step": 2248,
      "token_acc": 0.8956532831584654,
      "train_speed(iter/s)": 0.032467
    },
    {
      "epoch": 0.4357893716998498,
      "grad_norm": 0.1282435804605484,
      "learning_rate": 0.00019367492239068417,
      "loss": 0.37942934036254883,
      "memory(GiB)": 78.33,
      "step": 2249,
      "token_acc": 0.889295596034125,
      "train_speed(iter/s)": 0.032468
    },
    {
      "epoch": 0.43598314198517657,
      "grad_norm": 0.1114896610379219,
      "learning_rate": 0.0001935829466193032,
      "loss": 0.3755391538143158,
      "memory(GiB)": 78.33,
      "step": 2250,
      "token_acc": 0.8880601168939605,
      "train_speed(iter/s)": 0.032469
    },
    {
      "epoch": 0.4361769122705033,
      "grad_norm": 0.10226267576217651,
      "learning_rate": 0.00019349095294723487,
      "loss": 0.37153515219688416,
      "memory(GiB)": 78.33,
      "step": 2251,
      "token_acc": 0.8909354937328035,
      "train_speed(iter/s)": 0.03247
    },
    {
      "epoch": 0.43637068255583006,
      "grad_norm": 0.10146588832139969,
      "learning_rate": 0.00019339894141226355,
      "loss": 0.36199501156806946,
      "memory(GiB)": 78.33,
      "step": 2252,
      "token_acc": 0.8924786237263836,
      "train_speed(iter/s)": 0.032471
    },
    {
      "epoch": 0.4365644528411568,
      "grad_norm": 0.11089852452278137,
      "learning_rate": 0.00019330691205218082,
      "loss": 0.38684314489364624,
      "memory(GiB)": 78.33,
      "step": 2253,
      "token_acc": 0.8873638079006635,
      "train_speed(iter/s)": 0.032472
    },
    {
      "epoch": 0.43675822312648355,
      "grad_norm": 0.10390076041221619,
      "learning_rate": 0.00019321486490478563,
      "loss": 0.37561658024787903,
      "memory(GiB)": 78.33,
      "step": 2254,
      "token_acc": 0.8915121559575482,
      "train_speed(iter/s)": 0.032474
    },
    {
      "epoch": 0.4369519934118103,
      "grad_norm": 0.09438452124595642,
      "learning_rate": 0.00019312280000788416,
      "loss": 0.34584975242614746,
      "memory(GiB)": 78.33,
      "step": 2255,
      "token_acc": 0.8961015827777495,
      "train_speed(iter/s)": 0.032475
    },
    {
      "epoch": 0.43714576369713704,
      "grad_norm": 0.09714444726705551,
      "learning_rate": 0.00019303071739928997,
      "loss": 0.3477434515953064,
      "memory(GiB)": 78.33,
      "step": 2256,
      "token_acc": 0.8965367189699747,
      "train_speed(iter/s)": 0.032476
    },
    {
      "epoch": 0.4373395339824638,
      "grad_norm": 0.10147438943386078,
      "learning_rate": 0.00019293861711682393,
      "loss": 0.352212131023407,
      "memory(GiB)": 78.33,
      "step": 2257,
      "token_acc": 0.895200889829359,
      "train_speed(iter/s)": 0.032477
    },
    {
      "epoch": 0.4375333042677905,
      "grad_norm": 0.10559714585542679,
      "learning_rate": 0.00019284649919831394,
      "loss": 0.37872010469436646,
      "memory(GiB)": 78.33,
      "step": 2258,
      "token_acc": 0.8881790060998526,
      "train_speed(iter/s)": 0.032478
    },
    {
      "epoch": 0.43772707455311727,
      "grad_norm": 0.10056591033935547,
      "learning_rate": 0.00019275436368159548,
      "loss": 0.3722417652606964,
      "memory(GiB)": 78.33,
      "step": 2259,
      "token_acc": 0.8918299445471349,
      "train_speed(iter/s)": 0.032479
    },
    {
      "epoch": 0.437920844838444,
      "grad_norm": 0.10316530615091324,
      "learning_rate": 0.00019266221060451096,
      "loss": 0.3741195797920227,
      "memory(GiB)": 78.33,
      "step": 2260,
      "token_acc": 0.8905550295555158,
      "train_speed(iter/s)": 0.03248
    },
    {
      "epoch": 0.43811461512377076,
      "grad_norm": 0.1018475815653801,
      "learning_rate": 0.00019257004000491017,
      "loss": 0.36181291937828064,
      "memory(GiB)": 78.33,
      "step": 2261,
      "token_acc": 0.8936356535939978,
      "train_speed(iter/s)": 0.032481
    },
    {
      "epoch": 0.4383083854090975,
      "grad_norm": 0.10764992982149124,
      "learning_rate": 0.00019247785192065003,
      "loss": 0.39701730012893677,
      "memory(GiB)": 78.33,
      "step": 2262,
      "token_acc": 0.8833886405397252,
      "train_speed(iter/s)": 0.032483
    },
    {
      "epoch": 0.43850215569442424,
      "grad_norm": 0.09933136403560638,
      "learning_rate": 0.00019238564638959473,
      "loss": 0.331234335899353,
      "memory(GiB)": 78.33,
      "step": 2263,
      "token_acc": 0.901470844428959,
      "train_speed(iter/s)": 0.032484
    },
    {
      "epoch": 0.438695925979751,
      "grad_norm": 0.10954437404870987,
      "learning_rate": 0.00019229342344961547,
      "loss": 0.38815680146217346,
      "memory(GiB)": 78.33,
      "step": 2264,
      "token_acc": 0.8864250798765668,
      "train_speed(iter/s)": 0.032485
    },
    {
      "epoch": 0.43888969626507773,
      "grad_norm": 0.09880734980106354,
      "learning_rate": 0.00019220118313859074,
      "loss": 0.3450104892253876,
      "memory(GiB)": 78.33,
      "step": 2265,
      "token_acc": 0.8972709063595393,
      "train_speed(iter/s)": 0.032486
    },
    {
      "epoch": 0.4390834665504045,
      "grad_norm": 0.09824536740779877,
      "learning_rate": 0.0001921089254944061,
      "loss": 0.3457093834877014,
      "memory(GiB)": 78.33,
      "step": 2266,
      "token_acc": 0.898,
      "train_speed(iter/s)": 0.032487
    },
    {
      "epoch": 0.4392772368357312,
      "grad_norm": 0.11076337844133377,
      "learning_rate": 0.00019201665055495427,
      "loss": 0.4086833894252777,
      "memory(GiB)": 78.33,
      "step": 2267,
      "token_acc": 0.8818113491168733,
      "train_speed(iter/s)": 0.032488
    },
    {
      "epoch": 0.43947100712105797,
      "grad_norm": 0.10491563379764557,
      "learning_rate": 0.00019192435835813502,
      "loss": 0.36850589513778687,
      "memory(GiB)": 78.33,
      "step": 2268,
      "token_acc": 0.8915903890160183,
      "train_speed(iter/s)": 0.032489
    },
    {
      "epoch": 0.4396647774063847,
      "grad_norm": 0.10932449251413345,
      "learning_rate": 0.00019183204894185522,
      "loss": 0.3876577317714691,
      "memory(GiB)": 78.33,
      "step": 2269,
      "token_acc": 0.8863226177709189,
      "train_speed(iter/s)": 0.03249
    },
    {
      "epoch": 0.43985854769171145,
      "grad_norm": 0.11134737730026245,
      "learning_rate": 0.00019173972234402887,
      "loss": 0.40485280752182007,
      "memory(GiB)": 78.33,
      "step": 2270,
      "token_acc": 0.8811312687068596,
      "train_speed(iter/s)": 0.032491
    },
    {
      "epoch": 0.4400523179770382,
      "grad_norm": 0.1066315695643425,
      "learning_rate": 0.00019164737860257692,
      "loss": 0.38198620080947876,
      "memory(GiB)": 78.33,
      "step": 2271,
      "token_acc": 0.8884394005212859,
      "train_speed(iter/s)": 0.032493
    },
    {
      "epoch": 0.44024608826236494,
      "grad_norm": 0.10852167755365372,
      "learning_rate": 0.00019155501775542752,
      "loss": 0.3655702471733093,
      "memory(GiB)": 78.33,
      "step": 2272,
      "token_acc": 0.8904522077437578,
      "train_speed(iter/s)": 0.032494
    },
    {
      "epoch": 0.4404398585476917,
      "grad_norm": 0.10043393820524216,
      "learning_rate": 0.00019146263984051574,
      "loss": 0.3396553099155426,
      "memory(GiB)": 78.33,
      "step": 2273,
      "token_acc": 0.8992605233219567,
      "train_speed(iter/s)": 0.032495
    },
    {
      "epoch": 0.44063362883301843,
      "grad_norm": 0.09665820002555847,
      "learning_rate": 0.00019137024489578354,
      "loss": 0.3522893190383911,
      "memory(GiB)": 78.33,
      "step": 2274,
      "token_acc": 0.8970611596505162,
      "train_speed(iter/s)": 0.032496
    },
    {
      "epoch": 0.4408273991183452,
      "grad_norm": 0.10279867798089981,
      "learning_rate": 0.00019127783295918015,
      "loss": 0.38115227222442627,
      "memory(GiB)": 78.33,
      "step": 2275,
      "token_acc": 0.8877062684693637,
      "train_speed(iter/s)": 0.032497
    },
    {
      "epoch": 0.4410211694036719,
      "grad_norm": 0.10720198601484299,
      "learning_rate": 0.00019118540406866158,
      "loss": 0.3606134355068207,
      "memory(GiB)": 78.33,
      "step": 2276,
      "token_acc": 0.8946840521564694,
      "train_speed(iter/s)": 0.032499
    },
    {
      "epoch": 0.4412149396889987,
      "grad_norm": 0.10232347249984741,
      "learning_rate": 0.00019109295826219086,
      "loss": 0.3502247929573059,
      "memory(GiB)": 78.33,
      "step": 2277,
      "token_acc": 0.8963676036452317,
      "train_speed(iter/s)": 0.0325
    },
    {
      "epoch": 0.44140870997432546,
      "grad_norm": 0.10972965508699417,
      "learning_rate": 0.00019100049557773798,
      "loss": 0.3618561327457428,
      "memory(GiB)": 78.33,
      "step": 2278,
      "token_acc": 0.8917501192179303,
      "train_speed(iter/s)": 0.032501
    },
    {
      "epoch": 0.4416024802596522,
      "grad_norm": 0.10983320325613022,
      "learning_rate": 0.00019090801605327982,
      "loss": 0.37896856665611267,
      "memory(GiB)": 78.33,
      "step": 2279,
      "token_acc": 0.8861007817761251,
      "train_speed(iter/s)": 0.032502
    },
    {
      "epoch": 0.44179625054497895,
      "grad_norm": 0.10652820765972137,
      "learning_rate": 0.00019081551972680025,
      "loss": 0.3781920075416565,
      "memory(GiB)": 78.33,
      "step": 2280,
      "token_acc": 0.8892166502785972,
      "train_speed(iter/s)": 0.032503
    },
    {
      "epoch": 0.4419900208303057,
      "grad_norm": 0.10808536410331726,
      "learning_rate": 0.00019072300663628997,
      "loss": 0.3892320990562439,
      "memory(GiB)": 78.33,
      "step": 2281,
      "token_acc": 0.8866169049621531,
      "train_speed(iter/s)": 0.032504
    },
    {
      "epoch": 0.44218379111563244,
      "grad_norm": 0.09705408662557602,
      "learning_rate": 0.00019063047681974656,
      "loss": 0.33386996388435364,
      "memory(GiB)": 78.33,
      "step": 2282,
      "token_acc": 0.9006132756132756,
      "train_speed(iter/s)": 0.032505
    },
    {
      "epoch": 0.4423775614009592,
      "grad_norm": 0.10078884661197662,
      "learning_rate": 0.0001905379303151746,
      "loss": 0.33445224165916443,
      "memory(GiB)": 78.33,
      "step": 2283,
      "token_acc": 0.8993649371524941,
      "train_speed(iter/s)": 0.032506
    },
    {
      "epoch": 0.44257133168628593,
      "grad_norm": 0.11122056096792221,
      "learning_rate": 0.0001904453671605853,
      "loss": 0.37664756178855896,
      "memory(GiB)": 78.33,
      "step": 2284,
      "token_acc": 0.888987135970792,
      "train_speed(iter/s)": 0.032507
    },
    {
      "epoch": 0.4427651019716127,
      "grad_norm": 0.1016409620642662,
      "learning_rate": 0.00019035278739399692,
      "loss": 0.35757166147232056,
      "memory(GiB)": 78.33,
      "step": 2285,
      "token_acc": 0.8957771559399644,
      "train_speed(iter/s)": 0.032508
    },
    {
      "epoch": 0.4429588722569394,
      "grad_norm": 0.09880708903074265,
      "learning_rate": 0.00019026019105343445,
      "loss": 0.3487710952758789,
      "memory(GiB)": 78.33,
      "step": 2286,
      "token_acc": 0.8957860078103066,
      "train_speed(iter/s)": 0.03251
    },
    {
      "epoch": 0.44315264254226616,
      "grad_norm": 0.1038837805390358,
      "learning_rate": 0.00019016757817692966,
      "loss": 0.35564103722572327,
      "memory(GiB)": 78.33,
      "step": 2287,
      "token_acc": 0.893324717649042,
      "train_speed(iter/s)": 0.032511
    },
    {
      "epoch": 0.4433464128275929,
      "grad_norm": 0.11149082332849503,
      "learning_rate": 0.0001900749488025212,
      "loss": 0.3836996853351593,
      "memory(GiB)": 78.33,
      "step": 2288,
      "token_acc": 0.8873978009585566,
      "train_speed(iter/s)": 0.032512
    },
    {
      "epoch": 0.44354018311291965,
      "grad_norm": 0.10695455968379974,
      "learning_rate": 0.00018998230296825438,
      "loss": 0.3890923857688904,
      "memory(GiB)": 78.33,
      "step": 2289,
      "token_acc": 0.8848062223414241,
      "train_speed(iter/s)": 0.032513
    },
    {
      "epoch": 0.4437339533982464,
      "grad_norm": 0.10037211328744888,
      "learning_rate": 0.00018988964071218136,
      "loss": 0.34572556614875793,
      "memory(GiB)": 78.33,
      "step": 2290,
      "token_acc": 0.8985781478077526,
      "train_speed(iter/s)": 0.032514
    },
    {
      "epoch": 0.44392772368357314,
      "grad_norm": 0.09722217172384262,
      "learning_rate": 0.000189796962072361,
      "loss": 0.3216269314289093,
      "memory(GiB)": 78.33,
      "step": 2291,
      "token_acc": 0.906508290027221,
      "train_speed(iter/s)": 0.032515
    },
    {
      "epoch": 0.4441214939688999,
      "grad_norm": 0.1133374273777008,
      "learning_rate": 0.000189704267086859,
      "loss": 0.3567155599594116,
      "memory(GiB)": 78.33,
      "step": 2292,
      "token_acc": 0.8946020433729573,
      "train_speed(iter/s)": 0.032516
    },
    {
      "epoch": 0.4443152642542266,
      "grad_norm": 0.10639848560094833,
      "learning_rate": 0.00018961155579374757,
      "loss": 0.3713446855545044,
      "memory(GiB)": 78.33,
      "step": 2293,
      "token_acc": 0.8898174505758534,
      "train_speed(iter/s)": 0.032518
    },
    {
      "epoch": 0.44450903453955337,
      "grad_norm": 0.1109897717833519,
      "learning_rate": 0.0001895188282311058,
      "loss": 0.3658873438835144,
      "memory(GiB)": 78.33,
      "step": 2294,
      "token_acc": 0.8911519867265497,
      "train_speed(iter/s)": 0.032519
    },
    {
      "epoch": 0.4447028048248801,
      "grad_norm": 0.1037331074476242,
      "learning_rate": 0.00018942608443701936,
      "loss": 0.34084218740463257,
      "memory(GiB)": 78.33,
      "step": 2295,
      "token_acc": 0.898319518564778,
      "train_speed(iter/s)": 0.03252
    },
    {
      "epoch": 0.44489657511020686,
      "grad_norm": 0.11509410291910172,
      "learning_rate": 0.00018933332444958062,
      "loss": 0.38112330436706543,
      "memory(GiB)": 78.33,
      "step": 2296,
      "token_acc": 0.8888923143323982,
      "train_speed(iter/s)": 0.032521
    },
    {
      "epoch": 0.4450903453955336,
      "grad_norm": 0.09891083091497421,
      "learning_rate": 0.00018924054830688858,
      "loss": 0.33499619364738464,
      "memory(GiB)": 78.33,
      "step": 2297,
      "token_acc": 0.9014303407076423,
      "train_speed(iter/s)": 0.032522
    },
    {
      "epoch": 0.44528411568086035,
      "grad_norm": 0.11420368403196335,
      "learning_rate": 0.0001891477560470489,
      "loss": 0.3879055976867676,
      "memory(GiB)": 78.33,
      "step": 2298,
      "token_acc": 0.8863246853022871,
      "train_speed(iter/s)": 0.032523
    },
    {
      "epoch": 0.4454778859661871,
      "grad_norm": 0.10501628369092941,
      "learning_rate": 0.0001890549477081739,
      "loss": 0.3551484942436218,
      "memory(GiB)": 78.33,
      "step": 2299,
      "token_acc": 0.8968862632448985,
      "train_speed(iter/s)": 0.032524
    },
    {
      "epoch": 0.44567165625151384,
      "grad_norm": 0.10620047897100449,
      "learning_rate": 0.00018896212332838243,
      "loss": 0.36860981583595276,
      "memory(GiB)": 78.33,
      "step": 2300,
      "token_acc": 0.8929374124117507,
      "train_speed(iter/s)": 0.032525
    },
    {
      "epoch": 0.4458654265368406,
      "grad_norm": 0.10044432431459427,
      "learning_rate": 0.0001888692829458,
      "loss": 0.3472827672958374,
      "memory(GiB)": 78.33,
      "step": 2301,
      "token_acc": 0.8963743001866169,
      "train_speed(iter/s)": 0.032526
    },
    {
      "epoch": 0.4460591968221673,
      "grad_norm": 0.13254094123840332,
      "learning_rate": 0.00018877642659855852,
      "loss": 0.4520479440689087,
      "memory(GiB)": 78.33,
      "step": 2302,
      "token_acc": 0.8687972508591065,
      "train_speed(iter/s)": 0.032527
    },
    {
      "epoch": 0.44625296710749407,
      "grad_norm": 0.10516183078289032,
      "learning_rate": 0.00018868355432479674,
      "loss": 0.37986326217651367,
      "memory(GiB)": 78.33,
      "step": 2303,
      "token_acc": 0.8892758400680562,
      "train_speed(iter/s)": 0.032529
    },
    {
      "epoch": 0.4464467373928208,
      "grad_norm": 0.12673194706439972,
      "learning_rate": 0.00018859066616265966,
      "loss": 0.4348051846027374,
      "memory(GiB)": 78.33,
      "step": 2304,
      "token_acc": 0.8738275643379285,
      "train_speed(iter/s)": 0.03253
    },
    {
      "epoch": 0.44664050767814756,
      "grad_norm": 0.09944983571767807,
      "learning_rate": 0.00018849776215029907,
      "loss": 0.3587363064289093,
      "memory(GiB)": 78.33,
      "step": 2305,
      "token_acc": 0.8943836362735872,
      "train_speed(iter/s)": 0.032531
    },
    {
      "epoch": 0.4468342779634743,
      "grad_norm": 0.09945741295814514,
      "learning_rate": 0.0001884048423258731,
      "loss": 0.35953009128570557,
      "memory(GiB)": 78.33,
      "step": 2306,
      "token_acc": 0.8941004794663331,
      "train_speed(iter/s)": 0.032532
    },
    {
      "epoch": 0.44702804824880105,
      "grad_norm": 0.10500568151473999,
      "learning_rate": 0.00018831190672754638,
      "loss": 0.38044261932373047,
      "memory(GiB)": 78.33,
      "step": 2307,
      "token_acc": 0.8867819141508644,
      "train_speed(iter/s)": 0.032533
    },
    {
      "epoch": 0.4472218185341278,
      "grad_norm": 0.09857631474733353,
      "learning_rate": 0.00018821895539349008,
      "loss": 0.36684414744377136,
      "memory(GiB)": 78.33,
      "step": 2308,
      "token_acc": 0.8926224492510708,
      "train_speed(iter/s)": 0.032534
    },
    {
      "epoch": 0.44741558881945453,
      "grad_norm": 0.10511371493339539,
      "learning_rate": 0.00018812598836188182,
      "loss": 0.36754751205444336,
      "memory(GiB)": 78.33,
      "step": 2309,
      "token_acc": 0.8913974514883942,
      "train_speed(iter/s)": 0.032535
    },
    {
      "epoch": 0.4476093591047813,
      "grad_norm": 0.10627889633178711,
      "learning_rate": 0.0001880330056709057,
      "loss": 0.3721667230129242,
      "memory(GiB)": 78.33,
      "step": 2310,
      "token_acc": 0.8893509074742403,
      "train_speed(iter/s)": 0.032537
    },
    {
      "epoch": 0.447803129390108,
      "grad_norm": 0.09434963017702103,
      "learning_rate": 0.00018794000735875208,
      "loss": 0.3414641320705414,
      "memory(GiB)": 78.33,
      "step": 2311,
      "token_acc": 0.9010579609772407,
      "train_speed(iter/s)": 0.032538
    },
    {
      "epoch": 0.44799689967543477,
      "grad_norm": 0.10588974505662918,
      "learning_rate": 0.00018784699346361802,
      "loss": 0.36833858489990234,
      "memory(GiB)": 78.33,
      "step": 2312,
      "token_acc": 0.8927507447864945,
      "train_speed(iter/s)": 0.032539
    },
    {
      "epoch": 0.4481906699607615,
      "grad_norm": 0.10945326089859009,
      "learning_rate": 0.00018775396402370673,
      "loss": 0.3979440927505493,
      "memory(GiB)": 78.33,
      "step": 2313,
      "token_acc": 0.8817913102618431,
      "train_speed(iter/s)": 0.03254
    },
    {
      "epoch": 0.44838444024608826,
      "grad_norm": 0.10148966312408447,
      "learning_rate": 0.00018766091907722795,
      "loss": 0.35080429911613464,
      "memory(GiB)": 78.33,
      "step": 2314,
      "token_acc": 0.8943991907576,
      "train_speed(iter/s)": 0.032541
    },
    {
      "epoch": 0.448578210531415,
      "grad_norm": 0.10452762246131897,
      "learning_rate": 0.00018756785866239767,
      "loss": 0.36064577102661133,
      "memory(GiB)": 78.33,
      "step": 2315,
      "token_acc": 0.8935257756351483,
      "train_speed(iter/s)": 0.032542
    },
    {
      "epoch": 0.44877198081674174,
      "grad_norm": 0.10478675365447998,
      "learning_rate": 0.00018747478281743842,
      "loss": 0.36270397901535034,
      "memory(GiB)": 78.33,
      "step": 2316,
      "token_acc": 0.8914383656318249,
      "train_speed(iter/s)": 0.032543
    },
    {
      "epoch": 0.4489657511020685,
      "grad_norm": 0.10872726887464523,
      "learning_rate": 0.0001873816915805788,
      "loss": 0.3819129467010498,
      "memory(GiB)": 78.33,
      "step": 2317,
      "token_acc": 0.8880903215425461,
      "train_speed(iter/s)": 0.032544
    },
    {
      "epoch": 0.44915952138739523,
      "grad_norm": 0.09838545322418213,
      "learning_rate": 0.00018728858499005398,
      "loss": 0.34659987688064575,
      "memory(GiB)": 78.33,
      "step": 2318,
      "token_acc": 0.89670946219167,
      "train_speed(iter/s)": 0.032545
    },
    {
      "epoch": 0.449353291672722,
      "grad_norm": 0.11515597254037857,
      "learning_rate": 0.00018719546308410538,
      "loss": 0.389648973941803,
      "memory(GiB)": 78.33,
      "step": 2319,
      "token_acc": 0.8872989067638433,
      "train_speed(iter/s)": 0.032546
    },
    {
      "epoch": 0.4495470619580487,
      "grad_norm": 0.10723573714494705,
      "learning_rate": 0.00018710232590098057,
      "loss": 0.3719337582588196,
      "memory(GiB)": 78.33,
      "step": 2320,
      "token_acc": 0.8903108357787197,
      "train_speed(iter/s)": 0.032548
    },
    {
      "epoch": 0.44974083224337547,
      "grad_norm": 0.10834118723869324,
      "learning_rate": 0.00018700917347893358,
      "loss": 0.3716052770614624,
      "memory(GiB)": 78.33,
      "step": 2321,
      "token_acc": 0.889080622347949,
      "train_speed(iter/s)": 0.032549
    },
    {
      "epoch": 0.4499346025287022,
      "grad_norm": 0.10417478531599045,
      "learning_rate": 0.0001869160058562245,
      "loss": 0.35477250814437866,
      "memory(GiB)": 78.33,
      "step": 2322,
      "token_acc": 0.8943120009258188,
      "train_speed(iter/s)": 0.03255
    },
    {
      "epoch": 0.45012837281402895,
      "grad_norm": 0.10129362344741821,
      "learning_rate": 0.00018682282307111987,
      "loss": 0.3582378327846527,
      "memory(GiB)": 78.33,
      "step": 2323,
      "token_acc": 0.8946028116311099,
      "train_speed(iter/s)": 0.032551
    },
    {
      "epoch": 0.4503221430993557,
      "grad_norm": 0.09624646604061127,
      "learning_rate": 0.0001867296251618923,
      "loss": 0.3303111493587494,
      "memory(GiB)": 78.33,
      "step": 2324,
      "token_acc": 0.9006094476429071,
      "train_speed(iter/s)": 0.032552
    },
    {
      "epoch": 0.45051591338468244,
      "grad_norm": 0.10416833311319351,
      "learning_rate": 0.00018663641216682075,
      "loss": 0.3703954517841339,
      "memory(GiB)": 78.33,
      "step": 2325,
      "token_acc": 0.8913434299411637,
      "train_speed(iter/s)": 0.032553
    },
    {
      "epoch": 0.4507096836700092,
      "grad_norm": 0.09914438426494598,
      "learning_rate": 0.0001865431841241903,
      "loss": 0.3545830547809601,
      "memory(GiB)": 78.33,
      "step": 2326,
      "token_acc": 0.8953699331461094,
      "train_speed(iter/s)": 0.032554
    },
    {
      "epoch": 0.45090345395533593,
      "grad_norm": 0.10621084272861481,
      "learning_rate": 0.00018644994107229216,
      "loss": 0.37236636877059937,
      "memory(GiB)": 78.33,
      "step": 2327,
      "token_acc": 0.8887367838522704,
      "train_speed(iter/s)": 0.032555
    },
    {
      "epoch": 0.4510972242406627,
      "grad_norm": 0.09897017478942871,
      "learning_rate": 0.0001863566830494237,
      "loss": 0.33905407786369324,
      "memory(GiB)": 78.33,
      "step": 2328,
      "token_acc": 0.8993225346373268,
      "train_speed(iter/s)": 0.032557
    },
    {
      "epoch": 0.4512909945259894,
      "grad_norm": 0.09863407909870148,
      "learning_rate": 0.00018626341009388866,
      "loss": 0.3825795352458954,
      "memory(GiB)": 78.33,
      "step": 2329,
      "token_acc": 0.8874984754238322,
      "train_speed(iter/s)": 0.032558
    },
    {
      "epoch": 0.45148476481131616,
      "grad_norm": 0.09864397346973419,
      "learning_rate": 0.00018617012224399662,
      "loss": 0.3443533778190613,
      "memory(GiB)": 78.33,
      "step": 2330,
      "token_acc": 0.9003264309386758,
      "train_speed(iter/s)": 0.032559
    },
    {
      "epoch": 0.4516785350966429,
      "grad_norm": 0.09809895604848862,
      "learning_rate": 0.00018607681953806341,
      "loss": 0.34165775775909424,
      "memory(GiB)": 78.33,
      "step": 2331,
      "token_acc": 0.8970808010361662,
      "train_speed(iter/s)": 0.03256
    },
    {
      "epoch": 0.45187230538196965,
      "grad_norm": 0.10424879193305969,
      "learning_rate": 0.00018598350201441108,
      "loss": 0.33717483282089233,
      "memory(GiB)": 78.33,
      "step": 2332,
      "token_acc": 0.8975676890237052,
      "train_speed(iter/s)": 0.032561
    },
    {
      "epoch": 0.4520660756672964,
      "grad_norm": 0.09718424081802368,
      "learning_rate": 0.00018589016971136752,
      "loss": 0.32958295941352844,
      "memory(GiB)": 78.33,
      "step": 2333,
      "token_acc": 0.9014591294853342,
      "train_speed(iter/s)": 0.032562
    },
    {
      "epoch": 0.45225984595262314,
      "grad_norm": 0.1105201467871666,
      "learning_rate": 0.00018579682266726686,
      "loss": 0.38120901584625244,
      "memory(GiB)": 78.33,
      "step": 2334,
      "token_acc": 0.8889513793200416,
      "train_speed(iter/s)": 0.032563
    },
    {
      "epoch": 0.4524536162379499,
      "grad_norm": 0.1020718514919281,
      "learning_rate": 0.00018570346092044917,
      "loss": 0.3501797616481781,
      "memory(GiB)": 78.33,
      "step": 2335,
      "token_acc": 0.8964525407478428,
      "train_speed(iter/s)": 0.032564
    },
    {
      "epoch": 0.45264738652327663,
      "grad_norm": 0.10678357630968094,
      "learning_rate": 0.00018561008450926076,
      "loss": 0.3577033579349518,
      "memory(GiB)": 78.33,
      "step": 2336,
      "token_acc": 0.8953428424127857,
      "train_speed(iter/s)": 0.032565
    },
    {
      "epoch": 0.4528411568086034,
      "grad_norm": 0.09743531793355942,
      "learning_rate": 0.0001855166934720537,
      "loss": 0.33748358488082886,
      "memory(GiB)": 78.33,
      "step": 2337,
      "token_acc": 0.8984052721601771,
      "train_speed(iter/s)": 0.032566
    },
    {
      "epoch": 0.4530349270939302,
      "grad_norm": 0.10229934751987457,
      "learning_rate": 0.00018542328784718632,
      "loss": 0.3752382695674896,
      "memory(GiB)": 78.33,
      "step": 2338,
      "token_acc": 0.891382828441375,
      "train_speed(iter/s)": 0.032567
    },
    {
      "epoch": 0.4532286973792569,
      "grad_norm": 0.11242741346359253,
      "learning_rate": 0.00018532986767302276,
      "loss": 0.4054182767868042,
      "memory(GiB)": 78.33,
      "step": 2339,
      "token_acc": 0.8786372906672795,
      "train_speed(iter/s)": 0.032568
    },
    {
      "epoch": 0.45342246766458366,
      "grad_norm": 0.09873122721910477,
      "learning_rate": 0.0001852364329879332,
      "loss": 0.3532402813434601,
      "memory(GiB)": 78.33,
      "step": 2340,
      "token_acc": 0.8958922821738706,
      "train_speed(iter/s)": 0.032569
    },
    {
      "epoch": 0.4536162379499104,
      "grad_norm": 0.10772307217121124,
      "learning_rate": 0.00018514298383029372,
      "loss": 0.37972134351730347,
      "memory(GiB)": 78.33,
      "step": 2341,
      "token_acc": 0.8879170528266914,
      "train_speed(iter/s)": 0.03257
    },
    {
      "epoch": 0.45381000823523715,
      "grad_norm": 0.11007421463727951,
      "learning_rate": 0.00018504952023848647,
      "loss": 0.37062400579452515,
      "memory(GiB)": 78.33,
      "step": 2342,
      "token_acc": 0.8900929211930756,
      "train_speed(iter/s)": 0.032571
    },
    {
      "epoch": 0.4540037785205639,
      "grad_norm": 0.09810297191143036,
      "learning_rate": 0.00018495604225089946,
      "loss": 0.33862578868865967,
      "memory(GiB)": 78.33,
      "step": 2343,
      "token_acc": 0.9004052456801936,
      "train_speed(iter/s)": 0.032572
    },
    {
      "epoch": 0.45419754880589064,
      "grad_norm": 0.09834323078393936,
      "learning_rate": 0.00018486254990592656,
      "loss": 0.34658533334732056,
      "memory(GiB)": 78.33,
      "step": 2344,
      "token_acc": 0.8964155452144384,
      "train_speed(iter/s)": 0.032573
    },
    {
      "epoch": 0.4543913190912174,
      "grad_norm": 0.11022671312093735,
      "learning_rate": 0.00018476904324196764,
      "loss": 0.4045500159263611,
      "memory(GiB)": 78.33,
      "step": 2345,
      "token_acc": 0.8820878509132359,
      "train_speed(iter/s)": 0.032574
    },
    {
      "epoch": 0.4545850893765441,
      "grad_norm": 0.10049665719270706,
      "learning_rate": 0.0001846755222974284,
      "loss": 0.3371380865573883,
      "memory(GiB)": 78.33,
      "step": 2346,
      "token_acc": 0.9003629250212349,
      "train_speed(iter/s)": 0.032576
    },
    {
      "epoch": 0.45477885966187087,
      "grad_norm": 0.09852764755487442,
      "learning_rate": 0.0001845819871107204,
      "loss": 0.36543235182762146,
      "memory(GiB)": 78.33,
      "step": 2347,
      "token_acc": 0.8905981201226657,
      "train_speed(iter/s)": 0.032577
    },
    {
      "epoch": 0.4549726299471976,
      "grad_norm": 0.11310271173715591,
      "learning_rate": 0.00018448843772026098,
      "loss": 0.3698401153087616,
      "memory(GiB)": 78.33,
      "step": 2348,
      "token_acc": 0.8884804666140106,
      "train_speed(iter/s)": 0.032578
    },
    {
      "epoch": 0.45516640023252436,
      "grad_norm": 0.1051260381937027,
      "learning_rate": 0.0001843948741644735,
      "loss": 0.3575577437877655,
      "memory(GiB)": 78.33,
      "step": 2349,
      "token_acc": 0.8939034259509037,
      "train_speed(iter/s)": 0.032579
    },
    {
      "epoch": 0.4553601705178511,
      "grad_norm": 0.10609371960163116,
      "learning_rate": 0.00018430129648178693,
      "loss": 0.3754083216190338,
      "memory(GiB)": 78.33,
      "step": 2350,
      "token_acc": 0.8901177784276713,
      "train_speed(iter/s)": 0.03258
    },
    {
      "epoch": 0.45555394080317785,
      "grad_norm": 0.10359574854373932,
      "learning_rate": 0.0001842077047106362,
      "loss": 0.37402597069740295,
      "memory(GiB)": 78.33,
      "step": 2351,
      "token_acc": 0.8886889138857643,
      "train_speed(iter/s)": 0.032581
    },
    {
      "epoch": 0.4557477110885046,
      "grad_norm": 0.10701092332601547,
      "learning_rate": 0.00018411409888946197,
      "loss": 0.3984612822532654,
      "memory(GiB)": 78.33,
      "step": 2352,
      "token_acc": 0.8840575810574465,
      "train_speed(iter/s)": 0.032582
    },
    {
      "epoch": 0.45594148137383134,
      "grad_norm": 0.09848210960626602,
      "learning_rate": 0.00018402047905671063,
      "loss": 0.3315700590610504,
      "memory(GiB)": 78.33,
      "step": 2353,
      "token_acc": 0.9021433150310767,
      "train_speed(iter/s)": 0.032583
    },
    {
      "epoch": 0.4561352516591581,
      "grad_norm": 0.10119795799255371,
      "learning_rate": 0.0001839268452508344,
      "loss": 0.3340100347995758,
      "memory(GiB)": 78.33,
      "step": 2354,
      "token_acc": 0.9002782696686061,
      "train_speed(iter/s)": 0.032584
    },
    {
      "epoch": 0.4563290219444848,
      "grad_norm": 0.10151761025190353,
      "learning_rate": 0.00018383319751029114,
      "loss": 0.3325027823448181,
      "memory(GiB)": 78.33,
      "step": 2355,
      "token_acc": 0.902809093079392,
      "train_speed(iter/s)": 0.032585
    },
    {
      "epoch": 0.45652279222981157,
      "grad_norm": 0.11158560961484909,
      "learning_rate": 0.00018373953587354452,
      "loss": 0.38792747259140015,
      "memory(GiB)": 78.33,
      "step": 2356,
      "token_acc": 0.8858068315665489,
      "train_speed(iter/s)": 0.032586
    },
    {
      "epoch": 0.4567165625151383,
      "grad_norm": 0.0991615429520607,
      "learning_rate": 0.00018364586037906391,
      "loss": 0.35615038871765137,
      "memory(GiB)": 78.33,
      "step": 2357,
      "token_acc": 0.8954778266800505,
      "train_speed(iter/s)": 0.032587
    },
    {
      "epoch": 0.45691033280046506,
      "grad_norm": 0.10101715475320816,
      "learning_rate": 0.00018355217106532436,
      "loss": 0.36160457134246826,
      "memory(GiB)": 78.33,
      "step": 2358,
      "token_acc": 0.8935985113941181,
      "train_speed(iter/s)": 0.032588
    },
    {
      "epoch": 0.4571041030857918,
      "grad_norm": 0.10320735722780228,
      "learning_rate": 0.0001834584679708066,
      "loss": 0.3799397945404053,
      "memory(GiB)": 78.33,
      "step": 2359,
      "token_acc": 0.8880186336993524,
      "train_speed(iter/s)": 0.032589
    },
    {
      "epoch": 0.45729787337111855,
      "grad_norm": 0.1062033548951149,
      "learning_rate": 0.00018336475113399692,
      "loss": 0.3934144079685211,
      "memory(GiB)": 78.33,
      "step": 2360,
      "token_acc": 0.8855768736176123,
      "train_speed(iter/s)": 0.03259
    },
    {
      "epoch": 0.4574916436564453,
      "grad_norm": 0.10630171746015549,
      "learning_rate": 0.00018327102059338744,
      "loss": 0.3696288466453552,
      "memory(GiB)": 78.33,
      "step": 2361,
      "token_acc": 0.8901493818220828,
      "train_speed(iter/s)": 0.032592
    },
    {
      "epoch": 0.45768541394177203,
      "grad_norm": 0.10573367774486542,
      "learning_rate": 0.00018317727638747576,
      "loss": 0.3717585504055023,
      "memory(GiB)": 78.33,
      "step": 2362,
      "token_acc": 0.8903601270722656,
      "train_speed(iter/s)": 0.032593
    },
    {
      "epoch": 0.4578791842270988,
      "grad_norm": 0.10207115113735199,
      "learning_rate": 0.0001830835185547652,
      "loss": 0.3541256785392761,
      "memory(GiB)": 78.33,
      "step": 2363,
      "token_acc": 0.896280064694527,
      "train_speed(iter/s)": 0.032594
    },
    {
      "epoch": 0.4580729545124255,
      "grad_norm": 0.10319995135068893,
      "learning_rate": 0.0001829897471337645,
      "loss": 0.3481341004371643,
      "memory(GiB)": 78.33,
      "step": 2364,
      "token_acc": 0.8952120676258607,
      "train_speed(iter/s)": 0.032595
    },
    {
      "epoch": 0.45826672479775227,
      "grad_norm": 0.10591937601566315,
      "learning_rate": 0.00018289596216298823,
      "loss": 0.36794811487197876,
      "memory(GiB)": 78.33,
      "step": 2365,
      "token_acc": 0.8902500987974934,
      "train_speed(iter/s)": 0.032596
    },
    {
      "epoch": 0.458460495083079,
      "grad_norm": 0.1113501638174057,
      "learning_rate": 0.00018280216368095638,
      "loss": 0.37933623790740967,
      "memory(GiB)": 78.33,
      "step": 2366,
      "token_acc": 0.8892398270522216,
      "train_speed(iter/s)": 0.032597
    },
    {
      "epoch": 0.45865426536840576,
      "grad_norm": 0.10290928184986115,
      "learning_rate": 0.00018270835172619443,
      "loss": 0.3481866717338562,
      "memory(GiB)": 78.33,
      "step": 2367,
      "token_acc": 0.8950770760815515,
      "train_speed(iter/s)": 0.032598
    },
    {
      "epoch": 0.4588480356537325,
      "grad_norm": 0.11763869225978851,
      "learning_rate": 0.00018261452633723356,
      "loss": 0.41069701313972473,
      "memory(GiB)": 78.33,
      "step": 2368,
      "token_acc": 0.8784230338208672,
      "train_speed(iter/s)": 0.032599
    },
    {
      "epoch": 0.45904180593905924,
      "grad_norm": 0.1066950112581253,
      "learning_rate": 0.00018252068755261029,
      "loss": 0.3848106861114502,
      "memory(GiB)": 78.33,
      "step": 2369,
      "token_acc": 0.8873462694725335,
      "train_speed(iter/s)": 0.0326
    },
    {
      "epoch": 0.459235576224386,
      "grad_norm": 0.10410353541374207,
      "learning_rate": 0.00018242683541086678,
      "loss": 0.36915817856788635,
      "memory(GiB)": 78.33,
      "step": 2370,
      "token_acc": 0.8926594658498576,
      "train_speed(iter/s)": 0.032601
    },
    {
      "epoch": 0.45942934650971273,
      "grad_norm": 0.10082308948040009,
      "learning_rate": 0.00018233296995055065,
      "loss": 0.370003342628479,
      "memory(GiB)": 78.33,
      "step": 2371,
      "token_acc": 0.8912730226099714,
      "train_speed(iter/s)": 0.032602
    },
    {
      "epoch": 0.4596231167950395,
      "grad_norm": 0.10870834439992905,
      "learning_rate": 0.00018223909121021495,
      "loss": 0.3969360589981079,
      "memory(GiB)": 78.33,
      "step": 2372,
      "token_acc": 0.8850561134081512,
      "train_speed(iter/s)": 0.032603
    },
    {
      "epoch": 0.4598168870803662,
      "grad_norm": 0.09935883432626724,
      "learning_rate": 0.00018214519922841817,
      "loss": 0.3541335165500641,
      "memory(GiB)": 78.33,
      "step": 2373,
      "token_acc": 0.8943371776597123,
      "train_speed(iter/s)": 0.032604
    },
    {
      "epoch": 0.46001065736569297,
      "grad_norm": 0.12180610001087189,
      "learning_rate": 0.00018205129404372431,
      "loss": 0.4192396402359009,
      "memory(GiB)": 78.33,
      "step": 2374,
      "token_acc": 0.8759435110786462,
      "train_speed(iter/s)": 0.032605
    },
    {
      "epoch": 0.4602044276510197,
      "grad_norm": 0.11153632402420044,
      "learning_rate": 0.00018195737569470273,
      "loss": 0.37499576807022095,
      "memory(GiB)": 78.33,
      "step": 2375,
      "token_acc": 0.889001271529303,
      "train_speed(iter/s)": 0.032607
    },
    {
      "epoch": 0.46039819793634645,
      "grad_norm": 0.10651109367609024,
      "learning_rate": 0.0001818634442199282,
      "loss": 0.3690948486328125,
      "memory(GiB)": 78.33,
      "step": 2376,
      "token_acc": 0.8935253398571071,
      "train_speed(iter/s)": 0.032608
    },
    {
      "epoch": 0.4605919682216732,
      "grad_norm": 0.09841850399971008,
      "learning_rate": 0.00018176949965798093,
      "loss": 0.3518884778022766,
      "memory(GiB)": 78.33,
      "step": 2377,
      "token_acc": 0.8960413453472955,
      "train_speed(iter/s)": 0.032609
    },
    {
      "epoch": 0.46078573850699994,
      "grad_norm": 0.1050775870680809,
      "learning_rate": 0.0001816755420474465,
      "loss": 0.3738871216773987,
      "memory(GiB)": 78.33,
      "step": 2378,
      "token_acc": 0.8914807588995391,
      "train_speed(iter/s)": 0.03261
    },
    {
      "epoch": 0.4609795087923267,
      "grad_norm": 0.10667918622493744,
      "learning_rate": 0.0001815815714269158,
      "loss": 0.38406267762184143,
      "memory(GiB)": 78.33,
      "step": 2379,
      "token_acc": 0.8884490352348994,
      "train_speed(iter/s)": 0.032611
    },
    {
      "epoch": 0.46117327907765343,
      "grad_norm": 0.11324939876794815,
      "learning_rate": 0.00018148758783498504,
      "loss": 0.38455072045326233,
      "memory(GiB)": 78.33,
      "step": 2380,
      "token_acc": 0.8882405081991431,
      "train_speed(iter/s)": 0.032612
    },
    {
      "epoch": 0.4613670493629802,
      "grad_norm": 0.10635250061750412,
      "learning_rate": 0.00018139359131025588,
      "loss": 0.3805847764015198,
      "memory(GiB)": 78.33,
      "step": 2381,
      "token_acc": 0.8862944162436548,
      "train_speed(iter/s)": 0.032613
    },
    {
      "epoch": 0.4615608196483069,
      "grad_norm": 0.10868912190198898,
      "learning_rate": 0.00018129958189133522,
      "loss": 0.3691996932029724,
      "memory(GiB)": 78.33,
      "step": 2382,
      "token_acc": 0.8913490258405722,
      "train_speed(iter/s)": 0.032614
    },
    {
      "epoch": 0.46175458993363366,
      "grad_norm": 0.09865929931402206,
      "learning_rate": 0.00018120555961683514,
      "loss": 0.36495789885520935,
      "memory(GiB)": 78.33,
      "step": 2383,
      "token_acc": 0.8911807418711695,
      "train_speed(iter/s)": 0.032615
    },
    {
      "epoch": 0.4619483602189604,
      "grad_norm": 0.1034301221370697,
      "learning_rate": 0.00018111152452537327,
      "loss": 0.3717435598373413,
      "memory(GiB)": 78.33,
      "step": 2384,
      "token_acc": 0.8909826100745283,
      "train_speed(iter/s)": 0.032616
    },
    {
      "epoch": 0.46214213050428715,
      "grad_norm": 0.10854472219944,
      "learning_rate": 0.00018101747665557225,
      "loss": 0.3835892081260681,
      "memory(GiB)": 78.33,
      "step": 2385,
      "token_acc": 0.8888919333625602,
      "train_speed(iter/s)": 0.032617
    },
    {
      "epoch": 0.4623359007896139,
      "grad_norm": 0.11397820711135864,
      "learning_rate": 0.00018092341604606014,
      "loss": 0.39861786365509033,
      "memory(GiB)": 78.33,
      "step": 2386,
      "token_acc": 0.8822021941317055,
      "train_speed(iter/s)": 0.032618
    },
    {
      "epoch": 0.46252967107494064,
      "grad_norm": 0.1112833321094513,
      "learning_rate": 0.00018082934273547008,
      "loss": 0.40602025389671326,
      "memory(GiB)": 78.33,
      "step": 2387,
      "token_acc": 0.8811936155447606,
      "train_speed(iter/s)": 0.032619
    },
    {
      "epoch": 0.4627234413602674,
      "grad_norm": 0.11048243939876556,
      "learning_rate": 0.00018073525676244053,
      "loss": 0.3989701271057129,
      "memory(GiB)": 78.33,
      "step": 2388,
      "token_acc": 0.8846206164812145,
      "train_speed(iter/s)": 0.03262
    },
    {
      "epoch": 0.46291721164559413,
      "grad_norm": 0.12440559267997742,
      "learning_rate": 0.00018064115816561515,
      "loss": 0.397320419549942,
      "memory(GiB)": 78.33,
      "step": 2389,
      "token_acc": 0.8848505094371137,
      "train_speed(iter/s)": 0.032621
    },
    {
      "epoch": 0.4631109819309209,
      "grad_norm": 0.11104737967252731,
      "learning_rate": 0.00018054704698364273,
      "loss": 0.39772453904151917,
      "memory(GiB)": 78.33,
      "step": 2390,
      "token_acc": 0.8835264012326995,
      "train_speed(iter/s)": 0.032622
    },
    {
      "epoch": 0.4633047522162476,
      "grad_norm": 0.1087903156876564,
      "learning_rate": 0.00018045292325517736,
      "loss": 0.39890480041503906,
      "memory(GiB)": 78.33,
      "step": 2391,
      "token_acc": 0.8840932546508243,
      "train_speed(iter/s)": 0.032623
    },
    {
      "epoch": 0.46349852250157436,
      "grad_norm": 0.10867461562156677,
      "learning_rate": 0.00018035878701887803,
      "loss": 0.3494797348976135,
      "memory(GiB)": 78.33,
      "step": 2392,
      "token_acc": 0.897205366586873,
      "train_speed(iter/s)": 0.032624
    },
    {
      "epoch": 0.4636922927869011,
      "grad_norm": 0.11319278180599213,
      "learning_rate": 0.00018026463831340915,
      "loss": 0.36147987842559814,
      "memory(GiB)": 78.33,
      "step": 2393,
      "token_acc": 0.8932517509038962,
      "train_speed(iter/s)": 0.032625
    },
    {
      "epoch": 0.46388606307222785,
      "grad_norm": 0.1191987693309784,
      "learning_rate": 0.00018017047717744006,
      "loss": 0.4088186025619507,
      "memory(GiB)": 78.33,
      "step": 2394,
      "token_acc": 0.88241711618886,
      "train_speed(iter/s)": 0.032626
    },
    {
      "epoch": 0.4640798333575546,
      "grad_norm": 0.10697056353092194,
      "learning_rate": 0.00018007630364964524,
      "loss": 0.3530442714691162,
      "memory(GiB)": 78.33,
      "step": 2395,
      "token_acc": 0.8948714966856611,
      "train_speed(iter/s)": 0.032628
    },
    {
      "epoch": 0.46427360364288134,
      "grad_norm": 0.1151200458407402,
      "learning_rate": 0.00017998211776870435,
      "loss": 0.4034122824668884,
      "memory(GiB)": 78.33,
      "step": 2396,
      "token_acc": 0.8807363035709694,
      "train_speed(iter/s)": 0.032629
    },
    {
      "epoch": 0.4644673739282081,
      "grad_norm": 0.11261675506830215,
      "learning_rate": 0.00017988791957330205,
      "loss": 0.3917164206504822,
      "memory(GiB)": 78.33,
      "step": 2397,
      "token_acc": 0.8845557080623023,
      "train_speed(iter/s)": 0.03263
    },
    {
      "epoch": 0.4646611442135348,
      "grad_norm": 0.10412374138832092,
      "learning_rate": 0.00017979370910212807,
      "loss": 0.34287336468696594,
      "memory(GiB)": 78.33,
      "step": 2398,
      "token_acc": 0.8996082907468417,
      "train_speed(iter/s)": 0.032631
    },
    {
      "epoch": 0.4648549144988616,
      "grad_norm": 0.1128680557012558,
      "learning_rate": 0.00017969948639387715,
      "loss": 0.35632070899009705,
      "memory(GiB)": 78.33,
      "step": 2399,
      "token_acc": 0.8946635868277659,
      "train_speed(iter/s)": 0.032632
    },
    {
      "epoch": 0.46504868478418837,
      "grad_norm": 0.10247381776571274,
      "learning_rate": 0.00017960525148724916,
      "loss": 0.3603074848651886,
      "memory(GiB)": 78.33,
      "step": 2400,
      "token_acc": 0.891370611730082,
      "train_speed(iter/s)": 0.032633
    },
    {
      "epoch": 0.4652424550695151,
      "grad_norm": 0.10234736651182175,
      "learning_rate": 0.00017951100442094878,
      "loss": 0.3643084764480591,
      "memory(GiB)": 78.33,
      "step": 2401,
      "token_acc": 0.892455605758117,
      "train_speed(iter/s)": 0.032628
    },
    {
      "epoch": 0.46543622535484186,
      "grad_norm": 0.11007906496524811,
      "learning_rate": 0.00017941674523368594,
      "loss": 0.36536821722984314,
      "memory(GiB)": 78.33,
      "step": 2402,
      "token_acc": 0.8915501381254806,
      "train_speed(iter/s)": 0.032629
    },
    {
      "epoch": 0.4656299956401686,
      "grad_norm": 0.0991489514708519,
      "learning_rate": 0.00017932247396417538,
      "loss": 0.32952260971069336,
      "memory(GiB)": 78.33,
      "step": 2403,
      "token_acc": 0.9006100723689086,
      "train_speed(iter/s)": 0.032631
    },
    {
      "epoch": 0.46582376592549535,
      "grad_norm": 0.09857763350009918,
      "learning_rate": 0.00017922819065113683,
      "loss": 0.36439892649650574,
      "memory(GiB)": 78.33,
      "step": 2404,
      "token_acc": 0.8917956246540396,
      "train_speed(iter/s)": 0.032631
    },
    {
      "epoch": 0.4660175362108221,
      "grad_norm": 0.11544130742549896,
      "learning_rate": 0.000179133895333295,
      "loss": 0.39752769470214844,
      "memory(GiB)": 78.33,
      "step": 2405,
      "token_acc": 0.8802836879432624,
      "train_speed(iter/s)": 0.032633
    },
    {
      "epoch": 0.46621130649614884,
      "grad_norm": 0.11335242539644241,
      "learning_rate": 0.0001790395880493795,
      "loss": 0.3888709843158722,
      "memory(GiB)": 78.33,
      "step": 2406,
      "token_acc": 0.8840124504810413,
      "train_speed(iter/s)": 0.032634
    },
    {
      "epoch": 0.4664050767814756,
      "grad_norm": 0.1116030365228653,
      "learning_rate": 0.00017894526883812485,
      "loss": 0.3878939747810364,
      "memory(GiB)": 78.33,
      "step": 2407,
      "token_acc": 0.885872988604478,
      "train_speed(iter/s)": 0.032635
    },
    {
      "epoch": 0.4665988470668023,
      "grad_norm": 0.10574699193239212,
      "learning_rate": 0.00017885093773827048,
      "loss": 0.36204928159713745,
      "memory(GiB)": 78.33,
      "step": 2408,
      "token_acc": 0.8934056241568953,
      "train_speed(iter/s)": 0.032636
    },
    {
      "epoch": 0.46679261735212907,
      "grad_norm": 0.10200771689414978,
      "learning_rate": 0.00017875659478856076,
      "loss": 0.3507916033267975,
      "memory(GiB)": 78.33,
      "step": 2409,
      "token_acc": 0.8951724847051565,
      "train_speed(iter/s)": 0.032636
    },
    {
      "epoch": 0.4669863876374558,
      "grad_norm": 0.10833270847797394,
      "learning_rate": 0.00017866224002774478,
      "loss": 0.38488560914993286,
      "memory(GiB)": 78.33,
      "step": 2410,
      "token_acc": 0.8863565847742081,
      "train_speed(iter/s)": 0.032637
    },
    {
      "epoch": 0.46718015792278256,
      "grad_norm": 0.1008668914437294,
      "learning_rate": 0.00017856787349457672,
      "loss": 0.37106162309646606,
      "memory(GiB)": 78.33,
      "step": 2411,
      "token_acc": 0.8896542726679713,
      "train_speed(iter/s)": 0.032638
    },
    {
      "epoch": 0.4673739282081093,
      "grad_norm": 0.11065692454576492,
      "learning_rate": 0.0001784734952278153,
      "loss": 0.3777768313884735,
      "memory(GiB)": 78.33,
      "step": 2412,
      "token_acc": 0.8877977553825012,
      "train_speed(iter/s)": 0.03264
    },
    {
      "epoch": 0.46756769849343605,
      "grad_norm": 0.10646090656518936,
      "learning_rate": 0.00017837910526622436,
      "loss": 0.3846604824066162,
      "memory(GiB)": 78.33,
      "step": 2413,
      "token_acc": 0.8871676430824695,
      "train_speed(iter/s)": 0.032641
    },
    {
      "epoch": 0.4677614687787628,
      "grad_norm": 0.09328921884298325,
      "learning_rate": 0.00017828470364857226,
      "loss": 0.32804617285728455,
      "memory(GiB)": 78.33,
      "step": 2414,
      "token_acc": 0.9045698603985205,
      "train_speed(iter/s)": 0.032641
    },
    {
      "epoch": 0.46795523906408953,
      "grad_norm": 0.09627443552017212,
      "learning_rate": 0.00017819029041363232,
      "loss": 0.33626145124435425,
      "memory(GiB)": 78.33,
      "step": 2415,
      "token_acc": 0.8980459016393443,
      "train_speed(iter/s)": 0.032642
    },
    {
      "epoch": 0.4681490093494163,
      "grad_norm": 0.10673625767230988,
      "learning_rate": 0.00017809586560018262,
      "loss": 0.38183170557022095,
      "memory(GiB)": 78.33,
      "step": 2416,
      "token_acc": 0.8858918947761593,
      "train_speed(iter/s)": 0.032643
    },
    {
      "epoch": 0.468342779634743,
      "grad_norm": 0.1077309101819992,
      "learning_rate": 0.00017800142924700592,
      "loss": 0.3792920410633087,
      "memory(GiB)": 78.33,
      "step": 2417,
      "token_acc": 0.8886526780784098,
      "train_speed(iter/s)": 0.032644
    },
    {
      "epoch": 0.46853654992006977,
      "grad_norm": 0.11391132324934006,
      "learning_rate": 0.00017790698139288983,
      "loss": 0.3999425768852234,
      "memory(GiB)": 78.33,
      "step": 2418,
      "token_acc": 0.8836907644413697,
      "train_speed(iter/s)": 0.032645
    },
    {
      "epoch": 0.4687303202053965,
      "grad_norm": 0.10699284076690674,
      "learning_rate": 0.0001778125220766266,
      "loss": 0.3917970061302185,
      "memory(GiB)": 78.33,
      "step": 2419,
      "token_acc": 0.8875225537820958,
      "train_speed(iter/s)": 0.032646
    },
    {
      "epoch": 0.46892409049072326,
      "grad_norm": 0.10588902235031128,
      "learning_rate": 0.00017771805133701322,
      "loss": 0.3656570315361023,
      "memory(GiB)": 78.33,
      "step": 2420,
      "token_acc": 0.8920440225153323,
      "train_speed(iter/s)": 0.032648
    },
    {
      "epoch": 0.46911786077605,
      "grad_norm": 0.09696738421916962,
      "learning_rate": 0.00017762356921285127,
      "loss": 0.35896116495132446,
      "memory(GiB)": 78.33,
      "step": 2421,
      "token_acc": 0.8952547723573225,
      "train_speed(iter/s)": 0.032648
    },
    {
      "epoch": 0.46931163106137674,
      "grad_norm": 0.09918226301670074,
      "learning_rate": 0.00017752907574294726,
      "loss": 0.36196303367614746,
      "memory(GiB)": 78.33,
      "step": 2422,
      "token_acc": 0.8924877517691889,
      "train_speed(iter/s)": 0.032649
    },
    {
      "epoch": 0.4695054013467035,
      "grad_norm": 0.09119024872779846,
      "learning_rate": 0.000177434570966112,
      "loss": 0.34092551469802856,
      "memory(GiB)": 78.33,
      "step": 2423,
      "token_acc": 0.898368601754819,
      "train_speed(iter/s)": 0.03265
    },
    {
      "epoch": 0.46969917163203023,
      "grad_norm": 0.10667548328638077,
      "learning_rate": 0.00017734005492116135,
      "loss": 0.4041289687156677,
      "memory(GiB)": 78.33,
      "step": 2424,
      "token_acc": 0.8804930885083557,
      "train_speed(iter/s)": 0.032651
    },
    {
      "epoch": 0.469892941917357,
      "grad_norm": 0.10172632336616516,
      "learning_rate": 0.00017724552764691545,
      "loss": 0.3615379333496094,
      "memory(GiB)": 78.33,
      "step": 2425,
      "token_acc": 0.8944812914528845,
      "train_speed(iter/s)": 0.032652
    },
    {
      "epoch": 0.4700867122026837,
      "grad_norm": 0.10548295080661774,
      "learning_rate": 0.00017715098918219926,
      "loss": 0.36227190494537354,
      "memory(GiB)": 78.33,
      "step": 2426,
      "token_acc": 0.8919847828938738,
      "train_speed(iter/s)": 0.032653
    },
    {
      "epoch": 0.47028048248801047,
      "grad_norm": 0.10230764001607895,
      "learning_rate": 0.0001770564395658422,
      "loss": 0.3501635789871216,
      "memory(GiB)": 78.33,
      "step": 2427,
      "token_acc": 0.8956358685880147,
      "train_speed(iter/s)": 0.032654
    },
    {
      "epoch": 0.4704742527733372,
      "grad_norm": 0.10815402865409851,
      "learning_rate": 0.00017696187883667837,
      "loss": 0.3826008439064026,
      "memory(GiB)": 78.33,
      "step": 2428,
      "token_acc": 0.8863275727763192,
      "train_speed(iter/s)": 0.032655
    },
    {
      "epoch": 0.47066802305866395,
      "grad_norm": 0.12040967494249344,
      "learning_rate": 0.00017686730703354641,
      "loss": 0.43001601099967957,
      "memory(GiB)": 78.33,
      "step": 2429,
      "token_acc": 0.8732980332829047,
      "train_speed(iter/s)": 0.032656
    },
    {
      "epoch": 0.4708617933439907,
      "grad_norm": 0.10938671231269836,
      "learning_rate": 0.00017677272419528952,
      "loss": 0.39348623156547546,
      "memory(GiB)": 78.33,
      "step": 2430,
      "token_acc": 0.8847729835272117,
      "train_speed(iter/s)": 0.032657
    },
    {
      "epoch": 0.47105556362931744,
      "grad_norm": 0.09924609214067459,
      "learning_rate": 0.00017667813036075538,
      "loss": 0.35848769545555115,
      "memory(GiB)": 78.33,
      "step": 2431,
      "token_acc": 0.8955514503969244,
      "train_speed(iter/s)": 0.032658
    },
    {
      "epoch": 0.4712493339146442,
      "grad_norm": 0.09913709759712219,
      "learning_rate": 0.00017658352556879623,
      "loss": 0.3611469864845276,
      "memory(GiB)": 78.33,
      "step": 2432,
      "token_acc": 0.8925030260887481,
      "train_speed(iter/s)": 0.032659
    },
    {
      "epoch": 0.47144310419997093,
      "grad_norm": 0.10867290198802948,
      "learning_rate": 0.00017648890985826881,
      "loss": 0.34665271639823914,
      "memory(GiB)": 78.33,
      "step": 2433,
      "token_acc": 0.8978396543446951,
      "train_speed(iter/s)": 0.03266
    },
    {
      "epoch": 0.4716368744852977,
      "grad_norm": 0.10434415936470032,
      "learning_rate": 0.00017639428326803432,
      "loss": 0.3893589973449707,
      "memory(GiB)": 78.33,
      "step": 2434,
      "token_acc": 0.8848601637972565,
      "train_speed(iter/s)": 0.032661
    },
    {
      "epoch": 0.4718306447706244,
      "grad_norm": 0.10091494768857956,
      "learning_rate": 0.00017629964583695847,
      "loss": 0.3660696744918823,
      "memory(GiB)": 78.33,
      "step": 2435,
      "token_acc": 0.893289756957954,
      "train_speed(iter/s)": 0.032662
    },
    {
      "epoch": 0.47202441505595116,
      "grad_norm": 0.11700880527496338,
      "learning_rate": 0.00017620499760391133,
      "loss": 0.3919360041618347,
      "memory(GiB)": 78.33,
      "step": 2436,
      "token_acc": 0.8846436236304077,
      "train_speed(iter/s)": 0.032663
    },
    {
      "epoch": 0.4722181853412779,
      "grad_norm": 0.11263095587491989,
      "learning_rate": 0.00017611033860776752,
      "loss": 0.36513423919677734,
      "memory(GiB)": 78.33,
      "step": 2437,
      "token_acc": 0.8920323843097422,
      "train_speed(iter/s)": 0.032664
    },
    {
      "epoch": 0.47241195562660465,
      "grad_norm": 0.09740516543388367,
      "learning_rate": 0.0001760156688874061,
      "loss": 0.3375285565853119,
      "memory(GiB)": 78.33,
      "step": 2438,
      "token_acc": 0.899163848916149,
      "train_speed(iter/s)": 0.032665
    },
    {
      "epoch": 0.4726057259119314,
      "grad_norm": 0.10824833065271378,
      "learning_rate": 0.00017592098848171037,
      "loss": 0.36911553144454956,
      "memory(GiB)": 78.33,
      "step": 2439,
      "token_acc": 0.8915403549682601,
      "train_speed(iter/s)": 0.032666
    },
    {
      "epoch": 0.47279949619725814,
      "grad_norm": 0.1003868579864502,
      "learning_rate": 0.00017582629742956816,
      "loss": 0.3386095464229584,
      "memory(GiB)": 78.33,
      "step": 2440,
      "token_acc": 0.8985514852524404,
      "train_speed(iter/s)": 0.032667
    },
    {
      "epoch": 0.4729932664825849,
      "grad_norm": 0.13465555012226105,
      "learning_rate": 0.00017573159576987155,
      "loss": 0.4634130001068115,
      "memory(GiB)": 78.33,
      "step": 2441,
      "token_acc": 0.8691293051762041,
      "train_speed(iter/s)": 0.032668
    },
    {
      "epoch": 0.47318703676791163,
      "grad_norm": 0.11581727117300034,
      "learning_rate": 0.0001756368835415172,
      "loss": 0.3709140717983246,
      "memory(GiB)": 78.33,
      "step": 2442,
      "token_acc": 0.8906245185938318,
      "train_speed(iter/s)": 0.032669
    },
    {
      "epoch": 0.4733808070532384,
      "grad_norm": 0.09655480831861496,
      "learning_rate": 0.00017554216078340582,
      "loss": 0.3356662094593048,
      "memory(GiB)": 78.33,
      "step": 2443,
      "token_acc": 0.901089996601929,
      "train_speed(iter/s)": 0.03267
    },
    {
      "epoch": 0.4735745773385651,
      "grad_norm": 0.10045115649700165,
      "learning_rate": 0.00017544742753444268,
      "loss": 0.33761128783226013,
      "memory(GiB)": 78.33,
      "step": 2444,
      "token_acc": 0.8995756718528995,
      "train_speed(iter/s)": 0.032671
    },
    {
      "epoch": 0.47376834762389186,
      "grad_norm": 0.10537243634462357,
      "learning_rate": 0.0001753526838335373,
      "loss": 0.3550539016723633,
      "memory(GiB)": 78.33,
      "step": 2445,
      "token_acc": 0.8979015012132363,
      "train_speed(iter/s)": 0.032672
    },
    {
      "epoch": 0.4739621179092186,
      "grad_norm": 0.09614920616149902,
      "learning_rate": 0.0001752579297196034,
      "loss": 0.3335365056991577,
      "memory(GiB)": 78.33,
      "step": 2446,
      "token_acc": 0.9002796136248093,
      "train_speed(iter/s)": 0.032673
    },
    {
      "epoch": 0.47415588819454535,
      "grad_norm": 0.10471241921186447,
      "learning_rate": 0.00017516316523155903,
      "loss": 0.3631875216960907,
      "memory(GiB)": 78.33,
      "step": 2447,
      "token_acc": 0.8914798436857261,
      "train_speed(iter/s)": 0.032674
    },
    {
      "epoch": 0.4743496584798721,
      "grad_norm": 0.10309155285358429,
      "learning_rate": 0.00017506839040832653,
      "loss": 0.36359232664108276,
      "memory(GiB)": 78.33,
      "step": 2448,
      "token_acc": 0.8915206063477025,
      "train_speed(iter/s)": 0.032675
    },
    {
      "epoch": 0.47454342876519884,
      "grad_norm": 0.09676847606897354,
      "learning_rate": 0.00017497360528883252,
      "loss": 0.33657437562942505,
      "memory(GiB)": 78.33,
      "step": 2449,
      "token_acc": 0.8991928296794235,
      "train_speed(iter/s)": 0.032676
    },
    {
      "epoch": 0.4747371990505256,
      "grad_norm": 0.1178392842411995,
      "learning_rate": 0.0001748788099120077,
      "loss": 0.3946092426776886,
      "memory(GiB)": 78.33,
      "step": 2450,
      "token_acc": 0.8864816204051013,
      "train_speed(iter/s)": 0.032677
    },
    {
      "epoch": 0.4749309693358523,
      "grad_norm": 0.10294267535209656,
      "learning_rate": 0.00017478400431678715,
      "loss": 0.35818547010421753,
      "memory(GiB)": 78.33,
      "step": 2451,
      "token_acc": 0.8954294409377818,
      "train_speed(iter/s)": 0.032678
    },
    {
      "epoch": 0.47512473962117907,
      "grad_norm": 0.09975744038820267,
      "learning_rate": 0.00017468918854211007,
      "loss": 0.3438222408294678,
      "memory(GiB)": 78.33,
      "step": 2452,
      "token_acc": 0.899183906851024,
      "train_speed(iter/s)": 0.032679
    },
    {
      "epoch": 0.4753185099065058,
      "grad_norm": 0.12452986091375351,
      "learning_rate": 0.00017459436262691987,
      "loss": 0.41207653284072876,
      "memory(GiB)": 78.33,
      "step": 2453,
      "token_acc": 0.8813101862650361,
      "train_speed(iter/s)": 0.03268
    },
    {
      "epoch": 0.47551228019183256,
      "grad_norm": 0.09743805229663849,
      "learning_rate": 0.00017449952661016395,
      "loss": 0.3469730019569397,
      "memory(GiB)": 78.33,
      "step": 2454,
      "token_acc": 0.8969273247713859,
      "train_speed(iter/s)": 0.032681
    },
    {
      "epoch": 0.4757060504771593,
      "grad_norm": 0.10268343985080719,
      "learning_rate": 0.0001744046805307942,
      "loss": 0.33460330963134766,
      "memory(GiB)": 78.33,
      "step": 2455,
      "token_acc": 0.900157210232957,
      "train_speed(iter/s)": 0.032682
    },
    {
      "epoch": 0.47589982076248605,
      "grad_norm": 0.1134437769651413,
      "learning_rate": 0.00017430982442776636,
      "loss": 0.3615866005420685,
      "memory(GiB)": 78.33,
      "step": 2456,
      "token_acc": 0.8943463421872682,
      "train_speed(iter/s)": 0.032683
    },
    {
      "epoch": 0.4760935910478128,
      "grad_norm": 0.10063590854406357,
      "learning_rate": 0.0001742149583400404,
      "loss": 0.35712823271751404,
      "memory(GiB)": 78.33,
      "step": 2457,
      "token_acc": 0.8925671118305679,
      "train_speed(iter/s)": 0.032684
    },
    {
      "epoch": 0.47628736133313954,
      "grad_norm": 0.11840621381998062,
      "learning_rate": 0.0001741200823065804,
      "loss": 0.3843688666820526,
      "memory(GiB)": 78.33,
      "step": 2458,
      "token_acc": 0.8865816184002235,
      "train_speed(iter/s)": 0.032685
    },
    {
      "epoch": 0.47648113161846634,
      "grad_norm": 0.11133860051631927,
      "learning_rate": 0.00017402519636635445,
      "loss": 0.36577725410461426,
      "memory(GiB)": 78.33,
      "step": 2459,
      "token_acc": 0.8931690359777295,
      "train_speed(iter/s)": 0.032686
    },
    {
      "epoch": 0.4766749019037931,
      "grad_norm": 0.10411540418863297,
      "learning_rate": 0.00017393030055833477,
      "loss": 0.40943437814712524,
      "memory(GiB)": 78.33,
      "step": 2460,
      "token_acc": 0.8801027690550957,
      "train_speed(iter/s)": 0.032688
    },
    {
      "epoch": 0.4768686721891198,
      "grad_norm": 0.11033840477466583,
      "learning_rate": 0.00017383539492149755,
      "loss": 0.39954739809036255,
      "memory(GiB)": 78.33,
      "step": 2461,
      "token_acc": 0.8828463277143869,
      "train_speed(iter/s)": 0.032688
    },
    {
      "epoch": 0.47706244247444657,
      "grad_norm": 0.10192213207483292,
      "learning_rate": 0.00017374047949482324,
      "loss": 0.36716240644454956,
      "memory(GiB)": 78.33,
      "step": 2462,
      "token_acc": 0.8911974494758457,
      "train_speed(iter/s)": 0.032689
    },
    {
      "epoch": 0.4772562127597733,
      "grad_norm": 0.10164597630500793,
      "learning_rate": 0.000173645554317296,
      "loss": 0.37092214822769165,
      "memory(GiB)": 78.33,
      "step": 2463,
      "token_acc": 0.8932100329364074,
      "train_speed(iter/s)": 0.03269
    },
    {
      "epoch": 0.47744998304510006,
      "grad_norm": 0.10854342579841614,
      "learning_rate": 0.0001735506194279043,
      "loss": 0.37675389647483826,
      "memory(GiB)": 78.33,
      "step": 2464,
      "token_acc": 0.8900420837124658,
      "train_speed(iter/s)": 0.032691
    },
    {
      "epoch": 0.4776437533304268,
      "grad_norm": 0.1078154519200325,
      "learning_rate": 0.00017345567486564033,
      "loss": 0.3637319505214691,
      "memory(GiB)": 78.33,
      "step": 2465,
      "token_acc": 0.8951967543337022,
      "train_speed(iter/s)": 0.032692
    },
    {
      "epoch": 0.47783752361575355,
      "grad_norm": 0.11544948071241379,
      "learning_rate": 0.00017336072066950043,
      "loss": 0.42091140151023865,
      "memory(GiB)": 78.33,
      "step": 2466,
      "token_acc": 0.8770483917519161,
      "train_speed(iter/s)": 0.032693
    },
    {
      "epoch": 0.4780312939010803,
      "grad_norm": 0.10689322650432587,
      "learning_rate": 0.00017326575687848483,
      "loss": 0.36379310488700867,
      "memory(GiB)": 78.33,
      "step": 2467,
      "token_acc": 0.8929775576862291,
      "train_speed(iter/s)": 0.032694
    },
    {
      "epoch": 0.47822506418640703,
      "grad_norm": 0.1052493080496788,
      "learning_rate": 0.00017317078353159767,
      "loss": 0.36684519052505493,
      "memory(GiB)": 78.33,
      "step": 2468,
      "token_acc": 0.8944831280128549,
      "train_speed(iter/s)": 0.032695
    },
    {
      "epoch": 0.4784188344717338,
      "grad_norm": 0.11039218306541443,
      "learning_rate": 0.00017307580066784706,
      "loss": 0.36911848187446594,
      "memory(GiB)": 78.33,
      "step": 2469,
      "token_acc": 0.891165015793186,
      "train_speed(iter/s)": 0.032696
    },
    {
      "epoch": 0.4786126047570605,
      "grad_norm": 0.09582363814115524,
      "learning_rate": 0.00017298080832624512,
      "loss": 0.32416924834251404,
      "memory(GiB)": 78.33,
      "step": 2470,
      "token_acc": 0.9032258064516129,
      "train_speed(iter/s)": 0.032697
    },
    {
      "epoch": 0.47880637504238727,
      "grad_norm": 0.09783096611499786,
      "learning_rate": 0.00017288580654580766,
      "loss": 0.3526180386543274,
      "memory(GiB)": 78.33,
      "step": 2471,
      "token_acc": 0.8959444254195,
      "train_speed(iter/s)": 0.032698
    },
    {
      "epoch": 0.479000145327714,
      "grad_norm": 0.11712806671857834,
      "learning_rate": 0.00017279079536555448,
      "loss": 0.416570246219635,
      "memory(GiB)": 78.33,
      "step": 2472,
      "token_acc": 0.8787166033411736,
      "train_speed(iter/s)": 0.032699
    },
    {
      "epoch": 0.47919391561304076,
      "grad_norm": 0.10478947311639786,
      "learning_rate": 0.00017269577482450927,
      "loss": 0.3899994194507599,
      "memory(GiB)": 78.33,
      "step": 2473,
      "token_acc": 0.8852699947849998,
      "train_speed(iter/s)": 0.0327
    },
    {
      "epoch": 0.4793876858983675,
      "grad_norm": 0.10049055516719818,
      "learning_rate": 0.0001726007449616994,
      "loss": 0.32086169719696045,
      "memory(GiB)": 78.33,
      "step": 2474,
      "token_acc": 0.9046260601387818,
      "train_speed(iter/s)": 0.032701
    },
    {
      "epoch": 0.47958145618369424,
      "grad_norm": 0.09648868441581726,
      "learning_rate": 0.00017250570581615632,
      "loss": 0.3393422067165375,
      "memory(GiB)": 78.33,
      "step": 2475,
      "token_acc": 0.897726723095526,
      "train_speed(iter/s)": 0.032702
    },
    {
      "epoch": 0.479775226469021,
      "grad_norm": 0.1213340163230896,
      "learning_rate": 0.00017241065742691508,
      "loss": 0.3405624032020569,
      "memory(GiB)": 78.33,
      "step": 2476,
      "token_acc": 0.8992950060470193,
      "train_speed(iter/s)": 0.032703
    },
    {
      "epoch": 0.47996899675434773,
      "grad_norm": 0.10845254361629486,
      "learning_rate": 0.00017231559983301467,
      "loss": 0.3818724751472473,
      "memory(GiB)": 78.33,
      "step": 2477,
      "token_acc": 0.8866860090264346,
      "train_speed(iter/s)": 0.032704
    },
    {
      "epoch": 0.4801627670396745,
      "grad_norm": 0.1041831448674202,
      "learning_rate": 0.00017222053307349775,
      "loss": 0.36897408962249756,
      "memory(GiB)": 78.33,
      "step": 2478,
      "token_acc": 0.8889274835675733,
      "train_speed(iter/s)": 0.032705
    },
    {
      "epoch": 0.4803565373250012,
      "grad_norm": 0.11223351210355759,
      "learning_rate": 0.00017212545718741084,
      "loss": 0.39438849687576294,
      "memory(GiB)": 78.33,
      "step": 2479,
      "token_acc": 0.8818676281855405,
      "train_speed(iter/s)": 0.032706
    },
    {
      "epoch": 0.48055030761032796,
      "grad_norm": 0.11003533750772476,
      "learning_rate": 0.0001720303722138041,
      "loss": 0.3478088080883026,
      "memory(GiB)": 78.33,
      "step": 2480,
      "token_acc": 0.8965669378188336,
      "train_speed(iter/s)": 0.032707
    },
    {
      "epoch": 0.4807440778956547,
      "grad_norm": 0.11993599683046341,
      "learning_rate": 0.0001719352781917315,
      "loss": 0.35039135813713074,
      "memory(GiB)": 78.33,
      "step": 2481,
      "token_acc": 0.896633438940834,
      "train_speed(iter/s)": 0.032708
    },
    {
      "epoch": 0.48093784818098145,
      "grad_norm": 0.10672726482152939,
      "learning_rate": 0.00017184017516025076,
      "loss": 0.3407592177391052,
      "memory(GiB)": 78.33,
      "step": 2482,
      "token_acc": 0.8988511857298958,
      "train_speed(iter/s)": 0.032709
    },
    {
      "epoch": 0.4811316184663082,
      "grad_norm": 0.1108117550611496,
      "learning_rate": 0.00017174506315842316,
      "loss": 0.3430328965187073,
      "memory(GiB)": 78.33,
      "step": 2483,
      "token_acc": 0.9010159105568695,
      "train_speed(iter/s)": 0.03271
    },
    {
      "epoch": 0.48132538875163494,
      "grad_norm": 0.10908929258584976,
      "learning_rate": 0.00017164994222531384,
      "loss": 0.365764319896698,
      "memory(GiB)": 78.33,
      "step": 2484,
      "token_acc": 0.8944871455075537,
      "train_speed(iter/s)": 0.032711
    },
    {
      "epoch": 0.4815191590369617,
      "grad_norm": 0.10668904334306717,
      "learning_rate": 0.0001715548123999915,
      "loss": 0.36919450759887695,
      "memory(GiB)": 78.33,
      "step": 2485,
      "token_acc": 0.8912290460418194,
      "train_speed(iter/s)": 0.032712
    },
    {
      "epoch": 0.48171292932228843,
      "grad_norm": 0.10080403089523315,
      "learning_rate": 0.0001714596737215285,
      "loss": 0.33419883251190186,
      "memory(GiB)": 78.33,
      "step": 2486,
      "token_acc": 0.899380222317015,
      "train_speed(iter/s)": 0.032713
    },
    {
      "epoch": 0.4819066996076152,
      "grad_norm": 0.10038257390260696,
      "learning_rate": 0.00017136452622900083,
      "loss": 0.3448052406311035,
      "memory(GiB)": 78.33,
      "step": 2487,
      "token_acc": 0.898629288314102,
      "train_speed(iter/s)": 0.032714
    },
    {
      "epoch": 0.4821004698929419,
      "grad_norm": 0.10354321449995041,
      "learning_rate": 0.0001712693699614882,
      "loss": 0.35112708806991577,
      "memory(GiB)": 78.33,
      "step": 2488,
      "token_acc": 0.8958072674031679,
      "train_speed(iter/s)": 0.032715
    },
    {
      "epoch": 0.48229424017826866,
      "grad_norm": 0.10565353184938431,
      "learning_rate": 0.00017117420495807372,
      "loss": 0.34346991777420044,
      "memory(GiB)": 78.33,
      "step": 2489,
      "token_acc": 0.8978623685413809,
      "train_speed(iter/s)": 0.032716
    },
    {
      "epoch": 0.4824880104635954,
      "grad_norm": 0.10667932033538818,
      "learning_rate": 0.00017107903125784433,
      "loss": 0.39348533749580383,
      "memory(GiB)": 78.33,
      "step": 2490,
      "token_acc": 0.8839731259234479,
      "train_speed(iter/s)": 0.032717
    },
    {
      "epoch": 0.48268178074892215,
      "grad_norm": 0.10690039396286011,
      "learning_rate": 0.00017098384889989044,
      "loss": 0.36393576860427856,
      "memory(GiB)": 78.33,
      "step": 2491,
      "token_acc": 0.891016713091922,
      "train_speed(iter/s)": 0.032718
    },
    {
      "epoch": 0.4828755510342489,
      "grad_norm": 0.10752350091934204,
      "learning_rate": 0.0001708886579233059,
      "loss": 0.3595428764820099,
      "memory(GiB)": 78.33,
      "step": 2492,
      "token_acc": 0.8925832633641197,
      "train_speed(iter/s)": 0.032719
    },
    {
      "epoch": 0.48306932131957564,
      "grad_norm": 0.10408618301153183,
      "learning_rate": 0.00017079345836718828,
      "loss": 0.35838645696640015,
      "memory(GiB)": 78.33,
      "step": 2493,
      "token_acc": 0.893934211276438,
      "train_speed(iter/s)": 0.03272
    },
    {
      "epoch": 0.4832630916049024,
      "grad_norm": 0.10144350677728653,
      "learning_rate": 0.0001706982502706385,
      "loss": 0.3558150827884674,
      "memory(GiB)": 78.33,
      "step": 2494,
      "token_acc": 0.8957832584387083,
      "train_speed(iter/s)": 0.032721
    },
    {
      "epoch": 0.48345686189022913,
      "grad_norm": 0.10692603886127472,
      "learning_rate": 0.00017060303367276121,
      "loss": 0.4040507376194,
      "memory(GiB)": 78.33,
      "step": 2495,
      "token_acc": 0.8803178194638112,
      "train_speed(iter/s)": 0.032722
    },
    {
      "epoch": 0.4836506321755559,
      "grad_norm": 0.10149496793746948,
      "learning_rate": 0.00017050780861266432,
      "loss": 0.36347493529319763,
      "memory(GiB)": 78.33,
      "step": 2496,
      "token_acc": 0.8917304707027652,
      "train_speed(iter/s)": 0.032723
    },
    {
      "epoch": 0.4838444024608826,
      "grad_norm": 0.11008831858634949,
      "learning_rate": 0.00017041257512945943,
      "loss": 0.36779606342315674,
      "memory(GiB)": 78.33,
      "step": 2497,
      "token_acc": 0.8886516076638493,
      "train_speed(iter/s)": 0.032724
    },
    {
      "epoch": 0.48403817274620936,
      "grad_norm": 0.09817710518836975,
      "learning_rate": 0.00017031733326226142,
      "loss": 0.3289712071418762,
      "memory(GiB)": 78.33,
      "step": 2498,
      "token_acc": 0.9038888272495285,
      "train_speed(iter/s)": 0.032725
    },
    {
      "epoch": 0.4842319430315361,
      "grad_norm": 0.11836884170770645,
      "learning_rate": 0.00017022208305018867,
      "loss": 0.4116940498352051,
      "memory(GiB)": 78.33,
      "step": 2499,
      "token_acc": 0.88121387283237,
      "train_speed(iter/s)": 0.032726
    },
    {
      "epoch": 0.48442571331686285,
      "grad_norm": 0.11396266520023346,
      "learning_rate": 0.00017012682453236303,
      "loss": 0.3659469485282898,
      "memory(GiB)": 78.33,
      "step": 2500,
      "token_acc": 0.8918359785092324,
      "train_speed(iter/s)": 0.032727
    },
    {
      "epoch": 0.48442571331686285,
      "eval_loss": 0.4225236177444458,
      "eval_runtime": 1345.6225,
      "eval_samples_per_second": 5.016,
      "eval_steps_per_second": 5.016,
      "eval_token_acc": 0.8928025735859316,
      "step": 2500
    },
    {
      "epoch": 0.4846194836021896,
      "grad_norm": 0.10215523093938828,
      "learning_rate": 0.00017003155774790966,
      "loss": 0.37822431325912476,
      "memory(GiB)": 78.33,
      "step": 2501,
      "token_acc": 0.8875518780506904,
      "train_speed(iter/s)": 0.032161
    },
    {
      "epoch": 0.48481325388751634,
      "grad_norm": 0.09499707072973251,
      "learning_rate": 0.00016993628273595732,
      "loss": 0.3357214629650116,
      "memory(GiB)": 78.33,
      "step": 2502,
      "token_acc": 0.8995496714816547,
      "train_speed(iter/s)": 0.032162
    },
    {
      "epoch": 0.4850070241728431,
      "grad_norm": 0.1082739308476448,
      "learning_rate": 0.00016984099953563792,
      "loss": 0.3810504078865051,
      "memory(GiB)": 78.33,
      "step": 2503,
      "token_acc": 0.8882280116266552,
      "train_speed(iter/s)": 0.032163
    },
    {
      "epoch": 0.4852007944581698,
      "grad_norm": 0.1032496765255928,
      "learning_rate": 0.0001697457081860869,
      "loss": 0.36440324783325195,
      "memory(GiB)": 78.33,
      "step": 2504,
      "token_acc": 0.8939489607997896,
      "train_speed(iter/s)": 0.032164
    },
    {
      "epoch": 0.48539456474349657,
      "grad_norm": 0.08897780627012253,
      "learning_rate": 0.00016965040872644294,
      "loss": 0.3292064964771271,
      "memory(GiB)": 78.33,
      "step": 2505,
      "token_acc": 0.9021104037673323,
      "train_speed(iter/s)": 0.032165
    },
    {
      "epoch": 0.4855883350288233,
      "grad_norm": 0.10932856798171997,
      "learning_rate": 0.0001695551011958481,
      "loss": 0.39690276980400085,
      "memory(GiB)": 78.33,
      "step": 2506,
      "token_acc": 0.8848341960945107,
      "train_speed(iter/s)": 0.032167
    },
    {
      "epoch": 0.48578210531415006,
      "grad_norm": 0.11414424329996109,
      "learning_rate": 0.0001694597856334477,
      "loss": 0.39790499210357666,
      "memory(GiB)": 78.33,
      "step": 2507,
      "token_acc": 0.8816454951776053,
      "train_speed(iter/s)": 0.032168
    },
    {
      "epoch": 0.4859758755994768,
      "grad_norm": 0.09871453046798706,
      "learning_rate": 0.00016936446207839042,
      "loss": 0.3503097593784332,
      "memory(GiB)": 78.33,
      "step": 2508,
      "token_acc": 0.8961086541229425,
      "train_speed(iter/s)": 0.032169
    },
    {
      "epoch": 0.48616964588480355,
      "grad_norm": 0.09798671305179596,
      "learning_rate": 0.0001692691305698282,
      "loss": 0.35509318113327026,
      "memory(GiB)": 78.33,
      "step": 2509,
      "token_acc": 0.8953029405960135,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.4863634161701303,
      "grad_norm": 0.10052474588155746,
      "learning_rate": 0.00016917379114691635,
      "loss": 0.34173664450645447,
      "memory(GiB)": 78.33,
      "step": 2510,
      "token_acc": 0.8989416623644811,
      "train_speed(iter/s)": 0.032171
    },
    {
      "epoch": 0.48655718645545704,
      "grad_norm": 0.11249368637800217,
      "learning_rate": 0.00016907844384881325,
      "loss": 0.41095811128616333,
      "memory(GiB)": 78.33,
      "step": 2511,
      "token_acc": 0.8812950699043415,
      "train_speed(iter/s)": 0.032172
    },
    {
      "epoch": 0.4867509567407838,
      "grad_norm": 0.10713987797498703,
      "learning_rate": 0.00016898308871468059,
      "loss": 0.3698621988296509,
      "memory(GiB)": 78.33,
      "step": 2512,
      "token_acc": 0.8919433163888336,
      "train_speed(iter/s)": 0.032174
    },
    {
      "epoch": 0.4869447270261105,
      "grad_norm": 0.09921532869338989,
      "learning_rate": 0.00016888772578368326,
      "loss": 0.36754289269447327,
      "memory(GiB)": 78.33,
      "step": 2513,
      "token_acc": 0.8912722283784497,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.48713849731143727,
      "grad_norm": 0.10043738037347794,
      "learning_rate": 0.00016879235509498943,
      "loss": 0.3569088280200958,
      "memory(GiB)": 78.33,
      "step": 2514,
      "token_acc": 0.8948838737949167,
      "train_speed(iter/s)": 0.032176
    },
    {
      "epoch": 0.487332267596764,
      "grad_norm": 0.1124173253774643,
      "learning_rate": 0.00016869697668777043,
      "loss": 0.37260574102401733,
      "memory(GiB)": 78.33,
      "step": 2515,
      "token_acc": 0.8900623953736113,
      "train_speed(iter/s)": 0.032177
    },
    {
      "epoch": 0.48752603788209076,
      "grad_norm": 0.10620686411857605,
      "learning_rate": 0.00016860159060120062,
      "loss": 0.36633095145225525,
      "memory(GiB)": 78.33,
      "step": 2516,
      "token_acc": 0.8924985397602425,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.4877198081674175,
      "grad_norm": 0.10621728748083115,
      "learning_rate": 0.00016850619687445778,
      "loss": 0.36821886897087097,
      "memory(GiB)": 78.33,
      "step": 2517,
      "token_acc": 0.8912531612453126,
      "train_speed(iter/s)": 0.032179
    },
    {
      "epoch": 0.48791357845274425,
      "grad_norm": 0.1054278165102005,
      "learning_rate": 0.0001684107955467226,
      "loss": 0.3753064274787903,
      "memory(GiB)": 78.33,
      "step": 2518,
      "token_acc": 0.891217046851523,
      "train_speed(iter/s)": 0.03218
    },
    {
      "epoch": 0.488107348738071,
      "grad_norm": 0.09800657629966736,
      "learning_rate": 0.00016831538665717895,
      "loss": 0.35065239667892456,
      "memory(GiB)": 78.33,
      "step": 2519,
      "token_acc": 0.8958839022878686,
      "train_speed(iter/s)": 0.032181
    },
    {
      "epoch": 0.4883011190233978,
      "grad_norm": 0.12331446260213852,
      "learning_rate": 0.00016821997024501386,
      "loss": 0.4164249897003174,
      "memory(GiB)": 78.33,
      "step": 2520,
      "token_acc": 0.8785413567155202,
      "train_speed(iter/s)": 0.032183
    },
    {
      "epoch": 0.48849488930872453,
      "grad_norm": 0.10607205331325531,
      "learning_rate": 0.00016812454634941739,
      "loss": 0.34385666251182556,
      "memory(GiB)": 78.33,
      "step": 2521,
      "token_acc": 0.8990350297422339,
      "train_speed(iter/s)": 0.032184
    },
    {
      "epoch": 0.4886886595940513,
      "grad_norm": 0.10245993733406067,
      "learning_rate": 0.00016802911500958268,
      "loss": 0.38447538018226624,
      "memory(GiB)": 78.33,
      "step": 2522,
      "token_acc": 0.8873564827779333,
      "train_speed(iter/s)": 0.032185
    },
    {
      "epoch": 0.488882429879378,
      "grad_norm": 0.10764322429895401,
      "learning_rate": 0.00016793367626470598,
      "loss": 0.36548304557800293,
      "memory(GiB)": 78.33,
      "step": 2523,
      "token_acc": 0.8924651924651925,
      "train_speed(iter/s)": 0.032186
    },
    {
      "epoch": 0.48907620016470477,
      "grad_norm": 0.11456014215946198,
      "learning_rate": 0.0001678382301539866,
      "loss": 0.3953639566898346,
      "memory(GiB)": 78.33,
      "step": 2524,
      "token_acc": 0.8825286212045794,
      "train_speed(iter/s)": 0.032187
    },
    {
      "epoch": 0.4892699704500315,
      "grad_norm": 0.10977276414632797,
      "learning_rate": 0.00016774277671662672,
      "loss": 0.38997790217399597,
      "memory(GiB)": 78.33,
      "step": 2525,
      "token_acc": 0.8871661125759487,
      "train_speed(iter/s)": 0.032188
    },
    {
      "epoch": 0.48946374073535825,
      "grad_norm": 0.10588563233613968,
      "learning_rate": 0.00016764731599183173,
      "loss": 0.3698723614215851,
      "memory(GiB)": 78.33,
      "step": 2526,
      "token_acc": 0.8911557154232681,
      "train_speed(iter/s)": 0.03219
    },
    {
      "epoch": 0.489657511020685,
      "grad_norm": 0.1090141013264656,
      "learning_rate": 0.00016755184801880976,
      "loss": 0.3647971749305725,
      "memory(GiB)": 78.33,
      "step": 2527,
      "token_acc": 0.8910625354128768,
      "train_speed(iter/s)": 0.032191
    },
    {
      "epoch": 0.48985128130601174,
      "grad_norm": 0.11149538308382034,
      "learning_rate": 0.00016745637283677227,
      "loss": 0.4022625982761383,
      "memory(GiB)": 78.33,
      "step": 2528,
      "token_acc": 0.8828381516777448,
      "train_speed(iter/s)": 0.032192
    },
    {
      "epoch": 0.4900450515913385,
      "grad_norm": 0.102491594851017,
      "learning_rate": 0.0001673608904849333,
      "loss": 0.3276543915271759,
      "memory(GiB)": 78.33,
      "step": 2529,
      "token_acc": 0.9035930180828514,
      "train_speed(iter/s)": 0.032193
    },
    {
      "epoch": 0.49023882187666523,
      "grad_norm": 0.09906461089849472,
      "learning_rate": 0.00016726540100251013,
      "loss": 0.33269554376602173,
      "memory(GiB)": 78.33,
      "step": 2530,
      "token_acc": 0.9001250312578144,
      "train_speed(iter/s)": 0.032194
    },
    {
      "epoch": 0.490432592161992,
      "grad_norm": 0.10517586022615433,
      "learning_rate": 0.00016716990442872286,
      "loss": 0.3707316517829895,
      "memory(GiB)": 78.33,
      "step": 2531,
      "token_acc": 0.8911390265885534,
      "train_speed(iter/s)": 0.032195
    },
    {
      "epoch": 0.4906263624473187,
      "grad_norm": 0.10482536256313324,
      "learning_rate": 0.00016707440080279448,
      "loss": 0.37238699197769165,
      "memory(GiB)": 78.33,
      "step": 2532,
      "token_acc": 0.8896209510682288,
      "train_speed(iter/s)": 0.032196
    },
    {
      "epoch": 0.49082013273264546,
      "grad_norm": 0.10107911378145218,
      "learning_rate": 0.00016697889016395085,
      "loss": 0.34942498803138733,
      "memory(GiB)": 78.33,
      "step": 2533,
      "token_acc": 0.8957952468007313,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.4910139030179722,
      "grad_norm": 0.11440771073102951,
      "learning_rate": 0.00016688337255142078,
      "loss": 0.39366450905799866,
      "memory(GiB)": 78.33,
      "step": 2534,
      "token_acc": 0.8840879915976193,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.49120767330329895,
      "grad_norm": 0.11480151861906052,
      "learning_rate": 0.00016678784800443593,
      "loss": 0.3709234595298767,
      "memory(GiB)": 78.33,
      "step": 2535,
      "token_acc": 0.8920728858433266,
      "train_speed(iter/s)": 0.0322
    },
    {
      "epoch": 0.4914014435886257,
      "grad_norm": 0.10538162291049957,
      "learning_rate": 0.00016669231656223082,
      "loss": 0.3718525171279907,
      "memory(GiB)": 78.33,
      "step": 2536,
      "token_acc": 0.8928541556305238,
      "train_speed(iter/s)": 0.032201
    },
    {
      "epoch": 0.49159521387395244,
      "grad_norm": 0.10045890510082245,
      "learning_rate": 0.00016659677826404273,
      "loss": 0.3458371162414551,
      "memory(GiB)": 78.33,
      "step": 2537,
      "token_acc": 0.8953016402557687,
      "train_speed(iter/s)": 0.032202
    },
    {
      "epoch": 0.4917889841592792,
      "grad_norm": 0.09948378801345825,
      "learning_rate": 0.00016650123314911188,
      "loss": 0.3648657500743866,
      "memory(GiB)": 78.33,
      "step": 2538,
      "token_acc": 0.8927988345515284,
      "train_speed(iter/s)": 0.032203
    },
    {
      "epoch": 0.49198275444460593,
      "grad_norm": 0.11414239555597305,
      "learning_rate": 0.00016640568125668117,
      "loss": 0.39501774311065674,
      "memory(GiB)": 78.33,
      "step": 2539,
      "token_acc": 0.8854694665701183,
      "train_speed(iter/s)": 0.032204
    },
    {
      "epoch": 0.4921765247299327,
      "grad_norm": 0.11848455667495728,
      "learning_rate": 0.00016631012262599632,
      "loss": 0.4026211202144623,
      "memory(GiB)": 78.33,
      "step": 2540,
      "token_acc": 0.8802395209580839,
      "train_speed(iter/s)": 0.032205
    },
    {
      "epoch": 0.4923702950152594,
      "grad_norm": 0.10458105802536011,
      "learning_rate": 0.0001662145572963058,
      "loss": 0.37393152713775635,
      "memory(GiB)": 78.33,
      "step": 2541,
      "token_acc": 0.8897102626590847,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.49256406530058616,
      "grad_norm": 0.11251001805067062,
      "learning_rate": 0.0001661189853068609,
      "loss": 0.39233535528182983,
      "memory(GiB)": 78.33,
      "step": 2542,
      "token_acc": 0.8865367607200936,
      "train_speed(iter/s)": 0.032208
    },
    {
      "epoch": 0.4927578355859129,
      "grad_norm": 0.11127537488937378,
      "learning_rate": 0.00016602340669691563,
      "loss": 0.38980281352996826,
      "memory(GiB)": 78.33,
      "step": 2543,
      "token_acc": 0.8849459159053324,
      "train_speed(iter/s)": 0.032209
    },
    {
      "epoch": 0.49295160587123965,
      "grad_norm": 0.11086184531450272,
      "learning_rate": 0.00016592782150572666,
      "loss": 0.3862883448600769,
      "memory(GiB)": 78.33,
      "step": 2544,
      "token_acc": 0.8859040848435406,
      "train_speed(iter/s)": 0.03221
    },
    {
      "epoch": 0.4931453761565664,
      "grad_norm": 0.10376356542110443,
      "learning_rate": 0.00016583222977255337,
      "loss": 0.36207854747772217,
      "memory(GiB)": 78.33,
      "step": 2545,
      "token_acc": 0.8943563260789377,
      "train_speed(iter/s)": 0.032211
    },
    {
      "epoch": 0.49333914644189314,
      "grad_norm": 0.10361557453870773,
      "learning_rate": 0.00016573663153665792,
      "loss": 0.36922216415405273,
      "memory(GiB)": 78.33,
      "step": 2546,
      "token_acc": 0.8913299533978594,
      "train_speed(iter/s)": 0.032212
    },
    {
      "epoch": 0.4935329167272199,
      "grad_norm": 0.09568388015031815,
      "learning_rate": 0.000165641026837305,
      "loss": 0.32921454310417175,
      "memory(GiB)": 78.33,
      "step": 2547,
      "token_acc": 0.9026607486174076,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.49372668701254663,
      "grad_norm": 0.1072535291314125,
      "learning_rate": 0.00016554541571376212,
      "loss": 0.34089717268943787,
      "memory(GiB)": 78.33,
      "step": 2548,
      "token_acc": 0.899007279947055,
      "train_speed(iter/s)": 0.032214
    },
    {
      "epoch": 0.49392045729787337,
      "grad_norm": 0.11042487621307373,
      "learning_rate": 0.00016544979820529924,
      "loss": 0.381551057100296,
      "memory(GiB)": 78.33,
      "step": 2549,
      "token_acc": 0.8886517557338797,
      "train_speed(iter/s)": 0.032215
    },
    {
      "epoch": 0.4941142275832001,
      "grad_norm": 0.10860154032707214,
      "learning_rate": 0.0001653541743511891,
      "loss": 0.3577283024787903,
      "memory(GiB)": 78.33,
      "step": 2550,
      "token_acc": 0.8940223463687151,
      "train_speed(iter/s)": 0.032216
    },
    {
      "epoch": 0.49430799786852686,
      "grad_norm": 0.10476289689540863,
      "learning_rate": 0.00016525854419070698,
      "loss": 0.34515032172203064,
      "memory(GiB)": 78.33,
      "step": 2551,
      "token_acc": 0.896808724928779,
      "train_speed(iter/s)": 0.032218
    },
    {
      "epoch": 0.4945017681538536,
      "grad_norm": 0.11404412239789963,
      "learning_rate": 0.00016516290776313075,
      "loss": 0.3582231402397156,
      "memory(GiB)": 78.33,
      "step": 2552,
      "token_acc": 0.8928893430305944,
      "train_speed(iter/s)": 0.032219
    },
    {
      "epoch": 0.49469553843918035,
      "grad_norm": 0.11500284075737,
      "learning_rate": 0.00016506726510774085,
      "loss": 0.3722653388977051,
      "memory(GiB)": 78.33,
      "step": 2553,
      "token_acc": 0.8909366240293843,
      "train_speed(iter/s)": 0.03222
    },
    {
      "epoch": 0.4948893087245071,
      "grad_norm": 0.10306849330663681,
      "learning_rate": 0.00016497161626382028,
      "loss": 0.361613392829895,
      "memory(GiB)": 78.33,
      "step": 2554,
      "token_acc": 0.8950341710758377,
      "train_speed(iter/s)": 0.032221
    },
    {
      "epoch": 0.49508307900983384,
      "grad_norm": 0.10094691812992096,
      "learning_rate": 0.0001648759612706546,
      "loss": 0.364359587430954,
      "memory(GiB)": 78.33,
      "step": 2555,
      "token_acc": 0.8946505999798327,
      "train_speed(iter/s)": 0.032222
    },
    {
      "epoch": 0.4952768492951606,
      "grad_norm": 0.11370456963777542,
      "learning_rate": 0.00016478030016753195,
      "loss": 0.3835192620754242,
      "memory(GiB)": 78.33,
      "step": 2556,
      "token_acc": 0.88659125721692,
      "train_speed(iter/s)": 0.032223
    },
    {
      "epoch": 0.4954706195804873,
      "grad_norm": 0.09802607446908951,
      "learning_rate": 0.00016468463299374283,
      "loss": 0.3576071858406067,
      "memory(GiB)": 78.33,
      "step": 2557,
      "token_acc": 0.8939546925566343,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.49566438986581407,
      "grad_norm": 0.10389211028814316,
      "learning_rate": 0.00016458895978858034,
      "loss": 0.33778145909309387,
      "memory(GiB)": 78.33,
      "step": 2558,
      "token_acc": 0.901243754650792,
      "train_speed(iter/s)": 0.032225
    },
    {
      "epoch": 0.4958581601511408,
      "grad_norm": 0.10826588422060013,
      "learning_rate": 0.00016449328059134008,
      "loss": 0.39494338631629944,
      "memory(GiB)": 78.33,
      "step": 2559,
      "token_acc": 0.8841801579743431,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.49605193043646756,
      "grad_norm": 0.1128348559141159,
      "learning_rate": 0.00016439759544132,
      "loss": 0.35508641600608826,
      "memory(GiB)": 78.33,
      "step": 2560,
      "token_acc": 0.8955709517264593,
      "train_speed(iter/s)": 0.032227
    },
    {
      "epoch": 0.4962457007217943,
      "grad_norm": 0.10237224400043488,
      "learning_rate": 0.00016430190437782057,
      "loss": 0.3304956555366516,
      "memory(GiB)": 78.33,
      "step": 2561,
      "token_acc": 0.9007643970512228,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.49643947100712105,
      "grad_norm": 0.10696551948785782,
      "learning_rate": 0.00016420620744014473,
      "loss": 0.34542012214660645,
      "memory(GiB)": 78.33,
      "step": 2562,
      "token_acc": 0.8984604105571847,
      "train_speed(iter/s)": 0.03223
    },
    {
      "epoch": 0.4966332412924478,
      "grad_norm": 0.10525023192167282,
      "learning_rate": 0.00016411050466759775,
      "loss": 0.37965142726898193,
      "memory(GiB)": 78.33,
      "step": 2563,
      "token_acc": 0.8895881006864989,
      "train_speed(iter/s)": 0.032231
    },
    {
      "epoch": 0.49682701157777454,
      "grad_norm": 0.10651940107345581,
      "learning_rate": 0.00016401479609948736,
      "loss": 0.36009615659713745,
      "memory(GiB)": 78.33,
      "step": 2564,
      "token_acc": 0.8930840596007872,
      "train_speed(iter/s)": 0.032232
    },
    {
      "epoch": 0.4970207818631013,
      "grad_norm": 0.10184381902217865,
      "learning_rate": 0.00016391908177512362,
      "loss": 0.3449605405330658,
      "memory(GiB)": 78.33,
      "step": 2565,
      "token_acc": 0.8950802590120822,
      "train_speed(iter/s)": 0.032233
    },
    {
      "epoch": 0.497214552148428,
      "grad_norm": 0.10290346294641495,
      "learning_rate": 0.00016382336173381899,
      "loss": 0.3701861798763275,
      "memory(GiB)": 78.33,
      "step": 2566,
      "token_acc": 0.8891216519527159,
      "train_speed(iter/s)": 0.032234
    },
    {
      "epoch": 0.49740832243375477,
      "grad_norm": 0.10077276080846786,
      "learning_rate": 0.00016372763601488818,
      "loss": 0.36543571949005127,
      "memory(GiB)": 78.33,
      "step": 2567,
      "token_acc": 0.891035628960216,
      "train_speed(iter/s)": 0.032235
    },
    {
      "epoch": 0.4976020927190815,
      "grad_norm": 0.10975835472345352,
      "learning_rate": 0.00016363190465764837,
      "loss": 0.3935272693634033,
      "memory(GiB)": 78.33,
      "step": 2568,
      "token_acc": 0.882976987106251,
      "train_speed(iter/s)": 0.032236
    },
    {
      "epoch": 0.49779586300440826,
      "grad_norm": 0.10267551988363266,
      "learning_rate": 0.0001635361677014191,
      "loss": 0.3506276607513428,
      "memory(GiB)": 78.33,
      "step": 2569,
      "token_acc": 0.8953292213611682,
      "train_speed(iter/s)": 0.032237
    },
    {
      "epoch": 0.497989633289735,
      "grad_norm": 0.10422598570585251,
      "learning_rate": 0.00016344042518552198,
      "loss": 0.3643769323825836,
      "memory(GiB)": 78.33,
      "step": 2570,
      "token_acc": 0.8916066426570628,
      "train_speed(iter/s)": 0.032238
    },
    {
      "epoch": 0.49818340357506175,
      "grad_norm": 0.110415019094944,
      "learning_rate": 0.00016334467714928112,
      "loss": 0.38099807500839233,
      "memory(GiB)": 78.33,
      "step": 2571,
      "token_acc": 0.8894837676823347,
      "train_speed(iter/s)": 0.032239
    },
    {
      "epoch": 0.4983771738603885,
      "grad_norm": 0.0991300493478775,
      "learning_rate": 0.00016324892363202273,
      "loss": 0.33743083477020264,
      "memory(GiB)": 78.33,
      "step": 2572,
      "token_acc": 0.8997278841710735,
      "train_speed(iter/s)": 0.03224
    },
    {
      "epoch": 0.49857094414571523,
      "grad_norm": 0.11681834608316422,
      "learning_rate": 0.00016315316467307544,
      "loss": 0.41430431604385376,
      "memory(GiB)": 78.33,
      "step": 2573,
      "token_acc": 0.8792227548003788,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.498764714431042,
      "grad_norm": 0.10748863965272903,
      "learning_rate": 0.0001630574003117699,
      "loss": 0.393510639667511,
      "memory(GiB)": 78.33,
      "step": 2574,
      "token_acc": 0.8857627401373405,
      "train_speed(iter/s)": 0.032243
    },
    {
      "epoch": 0.4989584847163687,
      "grad_norm": 0.10948104411363602,
      "learning_rate": 0.00016296163058743919,
      "loss": 0.35196927189826965,
      "memory(GiB)": 78.33,
      "step": 2575,
      "token_acc": 0.8924734374072535,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.49915225500169547,
      "grad_norm": 0.1006598174571991,
      "learning_rate": 0.00016286585553941857,
      "loss": 0.32736673951148987,
      "memory(GiB)": 78.33,
      "step": 2576,
      "token_acc": 0.9026124709933379,
      "train_speed(iter/s)": 0.032245
    },
    {
      "epoch": 0.4993460252870222,
      "grad_norm": 0.09987809509038925,
      "learning_rate": 0.00016277007520704533,
      "loss": 0.3588752746582031,
      "memory(GiB)": 78.33,
      "step": 2577,
      "token_acc": 0.8945117224048975,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.49953979557234895,
      "grad_norm": 0.10553018003702164,
      "learning_rate": 0.00016267428962965906,
      "loss": 0.3588329255580902,
      "memory(GiB)": 78.33,
      "step": 2578,
      "token_acc": 0.8936772386426634,
      "train_speed(iter/s)": 0.032247
    },
    {
      "epoch": 0.4997335658576757,
      "grad_norm": 0.10368996113538742,
      "learning_rate": 0.00016257849884660148,
      "loss": 0.36948418617248535,
      "memory(GiB)": 78.33,
      "step": 2579,
      "token_acc": 0.8906377490590371,
      "train_speed(iter/s)": 0.032248
    },
    {
      "epoch": 0.49992733614300244,
      "grad_norm": 0.1034737378358841,
      "learning_rate": 0.00016248270289721646,
      "loss": 0.3439171612262726,
      "memory(GiB)": 78.33,
      "step": 2580,
      "token_acc": 0.8986508719973676,
      "train_speed(iter/s)": 0.032249
    },
    {
      "epoch": 0.5001211064283292,
      "grad_norm": 0.10310948640108109,
      "learning_rate": 0.00016238690182084986,
      "loss": 0.336532324552536,
      "memory(GiB)": 78.33,
      "step": 2581,
      "token_acc": 0.9005147686293942,
      "train_speed(iter/s)": 0.03225
    },
    {
      "epoch": 0.5003148767136559,
      "grad_norm": 0.1009165570139885,
      "learning_rate": 0.0001622910956568498,
      "loss": 0.33198171854019165,
      "memory(GiB)": 78.33,
      "step": 2582,
      "token_acc": 0.9027183088253284,
      "train_speed(iter/s)": 0.032251
    },
    {
      "epoch": 0.5005086469989827,
      "grad_norm": 0.11531993001699448,
      "learning_rate": 0.00016219528444456658,
      "loss": 0.38337087631225586,
      "memory(GiB)": 78.33,
      "step": 2583,
      "token_acc": 0.8874007793332371,
      "train_speed(iter/s)": 0.032253
    },
    {
      "epoch": 0.5007024172843094,
      "grad_norm": 0.11452654004096985,
      "learning_rate": 0.0001620994682233523,
      "loss": 0.3833463191986084,
      "memory(GiB)": 78.33,
      "step": 2584,
      "token_acc": 0.887452540227807,
      "train_speed(iter/s)": 0.032254
    },
    {
      "epoch": 0.5008961875696362,
      "grad_norm": 0.09829486906528473,
      "learning_rate": 0.00016200364703256132,
      "loss": 0.3433375358581543,
      "memory(GiB)": 78.33,
      "step": 2585,
      "token_acc": 0.898637268412188,
      "train_speed(iter/s)": 0.032255
    },
    {
      "epoch": 0.5010899578549629,
      "grad_norm": 0.11422152817249298,
      "learning_rate": 0.00016190782091154993,
      "loss": 0.3638113737106323,
      "memory(GiB)": 78.33,
      "step": 2586,
      "token_acc": 0.8920692223941321,
      "train_speed(iter/s)": 0.032256
    },
    {
      "epoch": 0.5012837281402897,
      "grad_norm": 0.1095975786447525,
      "learning_rate": 0.00016181198989967648,
      "loss": 0.375558078289032,
      "memory(GiB)": 78.33,
      "step": 2587,
      "token_acc": 0.8875113896788801,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.5014774984256164,
      "grad_norm": 0.12371022254228592,
      "learning_rate": 0.0001617161540363014,
      "loss": 0.3656739592552185,
      "memory(GiB)": 78.33,
      "step": 2588,
      "token_acc": 0.890730205842746,
      "train_speed(iter/s)": 0.032258
    },
    {
      "epoch": 0.5016712687109431,
      "grad_norm": 0.1015966534614563,
      "learning_rate": 0.00016162031336078707,
      "loss": 0.35260072350502014,
      "memory(GiB)": 78.33,
      "step": 2589,
      "token_acc": 0.8963511941792216,
      "train_speed(iter/s)": 0.032259
    },
    {
      "epoch": 0.5018650389962699,
      "grad_norm": 0.09963703155517578,
      "learning_rate": 0.00016152446791249775,
      "loss": 0.3328312933444977,
      "memory(GiB)": 78.33,
      "step": 2590,
      "token_acc": 0.9016321007428681,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 0.5020588092815966,
      "grad_norm": 0.11252304911613464,
      "learning_rate": 0.00016142861773079983,
      "loss": 0.3886357545852661,
      "memory(GiB)": 78.33,
      "step": 2591,
      "token_acc": 0.8855965345045425,
      "train_speed(iter/s)": 0.032261
    },
    {
      "epoch": 0.5022525795669234,
      "grad_norm": 0.10422030091285706,
      "learning_rate": 0.00016133276285506152,
      "loss": 0.3719256520271301,
      "memory(GiB)": 78.33,
      "step": 2592,
      "token_acc": 0.8892891253922772,
      "train_speed(iter/s)": 0.032262
    },
    {
      "epoch": 0.5024463498522501,
      "grad_norm": 0.11819025874137878,
      "learning_rate": 0.00016123690332465294,
      "loss": 0.3973425626754761,
      "memory(GiB)": 78.33,
      "step": 2593,
      "token_acc": 0.8866967825205125,
      "train_speed(iter/s)": 0.032264
    },
    {
      "epoch": 0.5026401201375769,
      "grad_norm": 0.11547715216875076,
      "learning_rate": 0.00016114103917894617,
      "loss": 0.3757579028606415,
      "memory(GiB)": 78.33,
      "step": 2594,
      "token_acc": 0.8914766343788905,
      "train_speed(iter/s)": 0.032265
    },
    {
      "epoch": 0.5028338904229036,
      "grad_norm": 0.1119546890258789,
      "learning_rate": 0.0001610451704573153,
      "loss": 0.365306556224823,
      "memory(GiB)": 78.33,
      "step": 2595,
      "token_acc": 0.8929053225410841,
      "train_speed(iter/s)": 0.032266
    },
    {
      "epoch": 0.5030276607082304,
      "grad_norm": 0.10307694226503372,
      "learning_rate": 0.00016094929719913612,
      "loss": 0.34190264344215393,
      "memory(GiB)": 78.33,
      "step": 2596,
      "token_acc": 0.8982385908726982,
      "train_speed(iter/s)": 0.032267
    },
    {
      "epoch": 0.5032214309935571,
      "grad_norm": 0.10450293868780136,
      "learning_rate": 0.00016085341944378634,
      "loss": 0.37576237320899963,
      "memory(GiB)": 78.33,
      "step": 2597,
      "token_acc": 0.8903388098419935,
      "train_speed(iter/s)": 0.032268
    },
    {
      "epoch": 0.5034152012788838,
      "grad_norm": 0.11997717618942261,
      "learning_rate": 0.00016075753723064558,
      "loss": 0.3967727720737457,
      "memory(GiB)": 78.33,
      "step": 2598,
      "token_acc": 0.8815336184366833,
      "train_speed(iter/s)": 0.032269
    },
    {
      "epoch": 0.5036089715642106,
      "grad_norm": 0.10347646474838257,
      "learning_rate": 0.00016066165059909523,
      "loss": 0.3663371801376343,
      "memory(GiB)": 78.33,
      "step": 2599,
      "token_acc": 0.8920181267202313,
      "train_speed(iter/s)": 0.03227
    },
    {
      "epoch": 0.5038027418495373,
      "grad_norm": 0.10675547271966934,
      "learning_rate": 0.00016056575958851843,
      "loss": 0.3652355670928955,
      "memory(GiB)": 78.33,
      "step": 2600,
      "token_acc": 0.890871055842873,
      "train_speed(iter/s)": 0.032271
    },
    {
      "epoch": 0.5039965121348641,
      "grad_norm": 0.11141776293516159,
      "learning_rate": 0.0001604698642383003,
      "loss": 0.37232065200805664,
      "memory(GiB)": 78.33,
      "step": 2601,
      "token_acc": 0.8899046440325167,
      "train_speed(iter/s)": 0.032267
    },
    {
      "epoch": 0.5041902824201908,
      "grad_norm": 0.08926242589950562,
      "learning_rate": 0.00016037396458782759,
      "loss": 0.3233528435230255,
      "memory(GiB)": 78.33,
      "step": 2602,
      "token_acc": 0.9036984098624263,
      "train_speed(iter/s)": 0.032268
    },
    {
      "epoch": 0.5043840527055176,
      "grad_norm": 0.09639449417591095,
      "learning_rate": 0.00016027806067648884,
      "loss": 0.31952396035194397,
      "memory(GiB)": 78.33,
      "step": 2603,
      "token_acc": 0.9034606910634639,
      "train_speed(iter/s)": 0.032269
    },
    {
      "epoch": 0.5045778229908443,
      "grad_norm": 0.09628106653690338,
      "learning_rate": 0.0001601821525436744,
      "loss": 0.3303004503250122,
      "memory(GiB)": 78.33,
      "step": 2604,
      "token_acc": 0.9007667907185853,
      "train_speed(iter/s)": 0.03227
    },
    {
      "epoch": 0.504771593276171,
      "grad_norm": 0.10307995975017548,
      "learning_rate": 0.0001600862402287763,
      "loss": 0.370724618434906,
      "memory(GiB)": 78.33,
      "step": 2605,
      "token_acc": 0.8890643615105422,
      "train_speed(iter/s)": 0.032271
    },
    {
      "epoch": 0.5049653635614978,
      "grad_norm": 0.11356647312641144,
      "learning_rate": 0.00015999032377118834,
      "loss": 0.3665652275085449,
      "memory(GiB)": 78.33,
      "step": 2606,
      "token_acc": 0.8920088410173396,
      "train_speed(iter/s)": 0.032272
    },
    {
      "epoch": 0.5051591338468245,
      "grad_norm": 0.11644966900348663,
      "learning_rate": 0.0001598944032103059,
      "loss": 0.38838085532188416,
      "memory(GiB)": 78.33,
      "step": 2607,
      "token_acc": 0.8869162270049968,
      "train_speed(iter/s)": 0.032273
    },
    {
      "epoch": 0.5053529041321513,
      "grad_norm": 0.10509419441223145,
      "learning_rate": 0.0001597984785855262,
      "loss": 0.37085598707199097,
      "memory(GiB)": 78.33,
      "step": 2608,
      "token_acc": 0.8929580496730619,
      "train_speed(iter/s)": 0.032274
    },
    {
      "epoch": 0.505546674417478,
      "grad_norm": 0.10611365735530853,
      "learning_rate": 0.0001597025499362481,
      "loss": 0.3378239870071411,
      "memory(GiB)": 78.33,
      "step": 2609,
      "token_acc": 0.9013872771389387,
      "train_speed(iter/s)": 0.032275
    },
    {
      "epoch": 0.5057404447028048,
      "grad_norm": 0.11018887907266617,
      "learning_rate": 0.000159606617301872,
      "loss": 0.3307594954967499,
      "memory(GiB)": 78.33,
      "step": 2610,
      "token_acc": 0.9017131424284106,
      "train_speed(iter/s)": 0.032276
    },
    {
      "epoch": 0.5059342149881316,
      "grad_norm": 0.1073288694024086,
      "learning_rate": 0.00015951068072180002,
      "loss": 0.38456863164901733,
      "memory(GiB)": 78.33,
      "step": 2611,
      "token_acc": 0.8856738792250035,
      "train_speed(iter/s)": 0.032277
    },
    {
      "epoch": 0.5061279852734584,
      "grad_norm": 0.11159303784370422,
      "learning_rate": 0.0001594147402354359,
      "loss": 0.3925580680370331,
      "memory(GiB)": 78.33,
      "step": 2612,
      "token_acc": 0.8856811904125457,
      "train_speed(iter/s)": 0.032279
    },
    {
      "epoch": 0.5063217555587851,
      "grad_norm": 0.09800157696008682,
      "learning_rate": 0.00015931879588218503,
      "loss": 0.3453458845615387,
      "memory(GiB)": 78.33,
      "step": 2613,
      "token_acc": 0.8966416230180362,
      "train_speed(iter/s)": 0.03228
    },
    {
      "epoch": 0.5065155258441119,
      "grad_norm": 0.09460814297199249,
      "learning_rate": 0.00015922284770145424,
      "loss": 0.32773369550704956,
      "memory(GiB)": 78.33,
      "step": 2614,
      "token_acc": 0.9022055463357145,
      "train_speed(iter/s)": 0.032281
    },
    {
      "epoch": 0.5067092961294386,
      "grad_norm": 0.10572027415037155,
      "learning_rate": 0.00015912689573265208,
      "loss": 0.37495627999305725,
      "memory(GiB)": 78.33,
      "step": 2615,
      "token_acc": 0.8893242702918832,
      "train_speed(iter/s)": 0.032282
    },
    {
      "epoch": 0.5069030664147653,
      "grad_norm": 0.10272035747766495,
      "learning_rate": 0.00015903094001518857,
      "loss": 0.3710392713546753,
      "memory(GiB)": 78.33,
      "step": 2616,
      "token_acc": 0.8908351579605881,
      "train_speed(iter/s)": 0.032283
    },
    {
      "epoch": 0.5070968367000921,
      "grad_norm": 0.11143457889556885,
      "learning_rate": 0.0001589349805884754,
      "loss": 0.38444697856903076,
      "memory(GiB)": 78.33,
      "step": 2617,
      "token_acc": 0.8872893621420449,
      "train_speed(iter/s)": 0.032284
    },
    {
      "epoch": 0.5072906069854188,
      "grad_norm": 0.10727585107088089,
      "learning_rate": 0.00015883901749192555,
      "loss": 0.3834106922149658,
      "memory(GiB)": 78.33,
      "step": 2618,
      "token_acc": 0.888004011444415,
      "train_speed(iter/s)": 0.032285
    },
    {
      "epoch": 0.5074843772707456,
      "grad_norm": 0.10124249756336212,
      "learning_rate": 0.00015874305076495372,
      "loss": 0.37049931287765503,
      "memory(GiB)": 78.33,
      "step": 2619,
      "token_acc": 0.8904576168763391,
      "train_speed(iter/s)": 0.032286
    },
    {
      "epoch": 0.5076781475560723,
      "grad_norm": 0.1076463907957077,
      "learning_rate": 0.00015864708044697597,
      "loss": 0.35899150371551514,
      "memory(GiB)": 78.33,
      "step": 2620,
      "token_acc": 0.8948525469168901,
      "train_speed(iter/s)": 0.032287
    },
    {
      "epoch": 0.5078719178413991,
      "grad_norm": 0.09248903393745422,
      "learning_rate": 0.00015855110657740998,
      "loss": 0.3522607386112213,
      "memory(GiB)": 78.33,
      "step": 2621,
      "token_acc": 0.8958730306269461,
      "train_speed(iter/s)": 0.032288
    },
    {
      "epoch": 0.5080656881267258,
      "grad_norm": 0.10546861588954926,
      "learning_rate": 0.00015845512919567467,
      "loss": 0.3607703447341919,
      "memory(GiB)": 78.33,
      "step": 2622,
      "token_acc": 0.8937468225724453,
      "train_speed(iter/s)": 0.032289
    },
    {
      "epoch": 0.5082594584120526,
      "grad_norm": 0.08962032198905945,
      "learning_rate": 0.00015835914834119066,
      "loss": 0.3159025311470032,
      "memory(GiB)": 78.33,
      "step": 2623,
      "token_acc": 0.9060619623054353,
      "train_speed(iter/s)": 0.03229
    },
    {
      "epoch": 0.5084532286973793,
      "grad_norm": 0.10019563883543015,
      "learning_rate": 0.0001582631640533798,
      "loss": 0.34771329164505005,
      "memory(GiB)": 78.33,
      "step": 2624,
      "token_acc": 0.8975572054242686,
      "train_speed(iter/s)": 0.032291
    },
    {
      "epoch": 0.508646998982706,
      "grad_norm": 0.11215566098690033,
      "learning_rate": 0.00015816717637166545,
      "loss": 0.38069719076156616,
      "memory(GiB)": 78.33,
      "step": 2625,
      "token_acc": 0.8880416751549722,
      "train_speed(iter/s)": 0.032292
    },
    {
      "epoch": 0.5088407692680328,
      "grad_norm": 0.10857607424259186,
      "learning_rate": 0.00015807118533547228,
      "loss": 0.36432355642318726,
      "memory(GiB)": 78.33,
      "step": 2626,
      "token_acc": 0.8919505825043285,
      "train_speed(iter/s)": 0.032293
    },
    {
      "epoch": 0.5090345395533595,
      "grad_norm": 0.10065297037363052,
      "learning_rate": 0.00015797519098422638,
      "loss": 0.3655664920806885,
      "memory(GiB)": 78.33,
      "step": 2627,
      "token_acc": 0.8914216996258685,
      "train_speed(iter/s)": 0.032294
    },
    {
      "epoch": 0.5092283098386863,
      "grad_norm": 0.10752709954977036,
      "learning_rate": 0.00015787919335735523,
      "loss": 0.35718318819999695,
      "memory(GiB)": 78.33,
      "step": 2628,
      "token_acc": 0.8937131050767414,
      "train_speed(iter/s)": 0.032295
    },
    {
      "epoch": 0.509422080124013,
      "grad_norm": 0.10011646896600723,
      "learning_rate": 0.0001577831924942877,
      "loss": 0.3524722158908844,
      "memory(GiB)": 78.33,
      "step": 2629,
      "token_acc": 0.8952115870315963,
      "train_speed(iter/s)": 0.032296
    },
    {
      "epoch": 0.5096158504093398,
      "grad_norm": 0.09924966841936111,
      "learning_rate": 0.00015768718843445386,
      "loss": 0.35439997911453247,
      "memory(GiB)": 78.33,
      "step": 2630,
      "token_acc": 0.8962517059855722,
      "train_speed(iter/s)": 0.032297
    },
    {
      "epoch": 0.5098096206946665,
      "grad_norm": 0.10966690629720688,
      "learning_rate": 0.00015759118121728516,
      "loss": 0.4074428677558899,
      "memory(GiB)": 78.33,
      "step": 2631,
      "token_acc": 0.8822711142654365,
      "train_speed(iter/s)": 0.032298
    },
    {
      "epoch": 0.5100033909799933,
      "grad_norm": 0.11511880904436111,
      "learning_rate": 0.00015749517088221434,
      "loss": 0.4096870720386505,
      "memory(GiB)": 78.33,
      "step": 2632,
      "token_acc": 0.8811336465830012,
      "train_speed(iter/s)": 0.032299
    },
    {
      "epoch": 0.51019716126532,
      "grad_norm": 0.10656805336475372,
      "learning_rate": 0.00015739915746867546,
      "loss": 0.3948022127151489,
      "memory(GiB)": 78.33,
      "step": 2633,
      "token_acc": 0.8854893199651264,
      "train_speed(iter/s)": 0.0323
    },
    {
      "epoch": 0.5103909315506467,
      "grad_norm": 0.09918279200792313,
      "learning_rate": 0.00015730314101610376,
      "loss": 0.32525914907455444,
      "memory(GiB)": 78.33,
      "step": 2634,
      "token_acc": 0.9005807402270954,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.5105847018359735,
      "grad_norm": 0.09818840026855469,
      "learning_rate": 0.00015720712156393579,
      "loss": 0.33824896812438965,
      "memory(GiB)": 78.33,
      "step": 2635,
      "token_acc": 0.9008060904612629,
      "train_speed(iter/s)": 0.032302
    },
    {
      "epoch": 0.5107784721213002,
      "grad_norm": 0.09556617587804794,
      "learning_rate": 0.00015711109915160932,
      "loss": 0.3609238564968109,
      "memory(GiB)": 78.33,
      "step": 2636,
      "token_acc": 0.8941985496374093,
      "train_speed(iter/s)": 0.032303
    },
    {
      "epoch": 0.510972242406627,
      "grad_norm": 0.10253620892763138,
      "learning_rate": 0.00015701507381856342,
      "loss": 0.35230863094329834,
      "memory(GiB)": 78.33,
      "step": 2637,
      "token_acc": 0.8980842250604073,
      "train_speed(iter/s)": 0.032304
    },
    {
      "epoch": 0.5111660126919537,
      "grad_norm": 0.10715737193822861,
      "learning_rate": 0.00015691904560423818,
      "loss": 0.37759995460510254,
      "memory(GiB)": 78.33,
      "step": 2638,
      "token_acc": 0.8882215743440234,
      "train_speed(iter/s)": 0.032305
    },
    {
      "epoch": 0.5113597829772805,
      "grad_norm": 0.1301286518573761,
      "learning_rate": 0.00015682301454807496,
      "loss": 0.3717860281467438,
      "memory(GiB)": 78.33,
      "step": 2639,
      "token_acc": 0.8927836337553795,
      "train_speed(iter/s)": 0.032306
    },
    {
      "epoch": 0.5115535532626072,
      "grad_norm": 0.09518956393003464,
      "learning_rate": 0.00015672698068951632,
      "loss": 0.3172317445278168,
      "memory(GiB)": 78.33,
      "step": 2640,
      "token_acc": 0.9052939066263517,
      "train_speed(iter/s)": 0.032307
    },
    {
      "epoch": 0.511747323547934,
      "grad_norm": 0.09442702680826187,
      "learning_rate": 0.00015663094406800592,
      "loss": 0.3069186210632324,
      "memory(GiB)": 78.33,
      "step": 2641,
      "token_acc": 0.9095811612439332,
      "train_speed(iter/s)": 0.032308
    },
    {
      "epoch": 0.5119410938332607,
      "grad_norm": 0.11501803249120712,
      "learning_rate": 0.00015653490472298864,
      "loss": 0.3687857389450073,
      "memory(GiB)": 78.33,
      "step": 2642,
      "token_acc": 0.8907549189100189,
      "train_speed(iter/s)": 0.032309
    },
    {
      "epoch": 0.5121348641185874,
      "grad_norm": 0.0969555526971817,
      "learning_rate": 0.00015643886269391043,
      "loss": 0.3279804587364197,
      "memory(GiB)": 78.33,
      "step": 2643,
      "token_acc": 0.9027313266443702,
      "train_speed(iter/s)": 0.03231
    },
    {
      "epoch": 0.5123286344039142,
      "grad_norm": 0.10562612861394882,
      "learning_rate": 0.00015634281802021826,
      "loss": 0.345773845911026,
      "memory(GiB)": 78.33,
      "step": 2644,
      "token_acc": 0.8951558127530089,
      "train_speed(iter/s)": 0.032311
    },
    {
      "epoch": 0.5125224046892409,
      "grad_norm": 0.10318602621555328,
      "learning_rate": 0.0001562467707413603,
      "loss": 0.38389700651168823,
      "memory(GiB)": 78.33,
      "step": 2645,
      "token_acc": 0.8879924999383219,
      "train_speed(iter/s)": 0.032312
    },
    {
      "epoch": 0.5127161749745677,
      "grad_norm": 0.0975809097290039,
      "learning_rate": 0.00015615072089678574,
      "loss": 0.3352702260017395,
      "memory(GiB)": 78.33,
      "step": 2646,
      "token_acc": 0.8997019155590305,
      "train_speed(iter/s)": 0.032313
    },
    {
      "epoch": 0.5129099452598944,
      "grad_norm": 0.102164626121521,
      "learning_rate": 0.00015605466852594481,
      "loss": 0.36365604400634766,
      "memory(GiB)": 78.33,
      "step": 2647,
      "token_acc": 0.8929846070017708,
      "train_speed(iter/s)": 0.032314
    },
    {
      "epoch": 0.5131037155452212,
      "grad_norm": 0.09489741176366806,
      "learning_rate": 0.00015595861366828883,
      "loss": 0.32492002844810486,
      "memory(GiB)": 78.33,
      "step": 2648,
      "token_acc": 0.9037146465248159,
      "train_speed(iter/s)": 0.032315
    },
    {
      "epoch": 0.5132974858305479,
      "grad_norm": 0.0985584408044815,
      "learning_rate": 0.00015586255636327012,
      "loss": 0.3407394289970398,
      "memory(GiB)": 78.33,
      "step": 2649,
      "token_acc": 0.9006668376506797,
      "train_speed(iter/s)": 0.032316
    },
    {
      "epoch": 0.5134912561158747,
      "grad_norm": 0.0985412672162056,
      "learning_rate": 0.00015576649665034197,
      "loss": 0.3170143961906433,
      "memory(GiB)": 78.33,
      "step": 2650,
      "token_acc": 0.907279489904357,
      "train_speed(iter/s)": 0.032317
    },
    {
      "epoch": 0.5136850264012014,
      "grad_norm": 0.10823136568069458,
      "learning_rate": 0.00015567043456895868,
      "loss": 0.37427809834480286,
      "memory(GiB)": 78.33,
      "step": 2651,
      "token_acc": 0.8882963136611381,
      "train_speed(iter/s)": 0.032318
    },
    {
      "epoch": 0.5138787966865281,
      "grad_norm": 0.09839842468500137,
      "learning_rate": 0.0001555743701585756,
      "loss": 0.3554091155529022,
      "memory(GiB)": 78.33,
      "step": 2652,
      "token_acc": 0.8943349139006738,
      "train_speed(iter/s)": 0.032319
    },
    {
      "epoch": 0.5140725669718549,
      "grad_norm": 0.10772431641817093,
      "learning_rate": 0.00015547830345864885,
      "loss": 0.3907697796821594,
      "memory(GiB)": 78.33,
      "step": 2653,
      "token_acc": 0.8854761904761905,
      "train_speed(iter/s)": 0.03232
    },
    {
      "epoch": 0.5142663372571816,
      "grad_norm": 0.10258005559444427,
      "learning_rate": 0.00015538223450863565,
      "loss": 0.3574678897857666,
      "memory(GiB)": 78.33,
      "step": 2654,
      "token_acc": 0.893127167301461,
      "train_speed(iter/s)": 0.032322
    },
    {
      "epoch": 0.5144601075425084,
      "grad_norm": 0.10088451206684113,
      "learning_rate": 0.0001552861633479941,
      "loss": 0.37425774335861206,
      "memory(GiB)": 78.33,
      "step": 2655,
      "token_acc": 0.8893967324057473,
      "train_speed(iter/s)": 0.032322
    },
    {
      "epoch": 0.5146538778278351,
      "grad_norm": 0.10441171377897263,
      "learning_rate": 0.00015519009001618327,
      "loss": 0.3345174193382263,
      "memory(GiB)": 78.33,
      "step": 2656,
      "token_acc": 0.8989620545132345,
      "train_speed(iter/s)": 0.032323
    },
    {
      "epoch": 0.5148476481131619,
      "grad_norm": 0.10342314839363098,
      "learning_rate": 0.000155094014552663,
      "loss": 0.3678382635116577,
      "memory(GiB)": 78.33,
      "step": 2657,
      "token_acc": 0.8913637207329038,
      "train_speed(iter/s)": 0.032324
    },
    {
      "epoch": 0.5150414183984886,
      "grad_norm": 0.10229925811290741,
      "learning_rate": 0.00015499793699689406,
      "loss": 0.3740781247615814,
      "memory(GiB)": 78.33,
      "step": 2658,
      "token_acc": 0.8912131626660894,
      "train_speed(iter/s)": 0.032325
    },
    {
      "epoch": 0.5152351886838153,
      "grad_norm": 0.10311522334814072,
      "learning_rate": 0.0001549018573883381,
      "loss": 0.3588406443595886,
      "memory(GiB)": 78.33,
      "step": 2659,
      "token_acc": 0.8927260150055872,
      "train_speed(iter/s)": 0.032326
    },
    {
      "epoch": 0.5154289589691421,
      "grad_norm": 0.1111614927649498,
      "learning_rate": 0.00015480577576645758,
      "loss": 0.3687216639518738,
      "memory(GiB)": 78.33,
      "step": 2660,
      "token_acc": 0.8915793809579852,
      "train_speed(iter/s)": 0.032327
    },
    {
      "epoch": 0.5156227292544688,
      "grad_norm": 0.10557236522436142,
      "learning_rate": 0.00015470969217071582,
      "loss": 0.39404910802841187,
      "memory(GiB)": 78.33,
      "step": 2661,
      "token_acc": 0.884908754848619,
      "train_speed(iter/s)": 0.032328
    },
    {
      "epoch": 0.5158164995397956,
      "grad_norm": 0.11680735647678375,
      "learning_rate": 0.00015461360664057692,
      "loss": 0.391072154045105,
      "memory(GiB)": 78.33,
      "step": 2662,
      "token_acc": 0.8852517451358978,
      "train_speed(iter/s)": 0.03233
    },
    {
      "epoch": 0.5160102698251223,
      "grad_norm": 0.1172829270362854,
      "learning_rate": 0.00015451751921550583,
      "loss": 0.38999128341674805,
      "memory(GiB)": 78.33,
      "step": 2663,
      "token_acc": 0.8861199122593224,
      "train_speed(iter/s)": 0.032331
    },
    {
      "epoch": 0.5162040401104491,
      "grad_norm": 0.09996732324361801,
      "learning_rate": 0.0001544214299349682,
      "loss": 0.3332677483558655,
      "memory(GiB)": 78.33,
      "step": 2664,
      "token_acc": 0.900718860279702,
      "train_speed(iter/s)": 0.032332
    },
    {
      "epoch": 0.5163978103957758,
      "grad_norm": 0.11197924613952637,
      "learning_rate": 0.00015432533883843048,
      "loss": 0.35336822271347046,
      "memory(GiB)": 78.33,
      "step": 2665,
      "token_acc": 0.8947896063077465,
      "train_speed(iter/s)": 0.032333
    },
    {
      "epoch": 0.5165915806811026,
      "grad_norm": 0.10348919034004211,
      "learning_rate": 0.0001542292459653599,
      "loss": 0.3489812910556793,
      "memory(GiB)": 78.33,
      "step": 2666,
      "token_acc": 0.8955829903978052,
      "train_speed(iter/s)": 0.032334
    },
    {
      "epoch": 0.5167853509664293,
      "grad_norm": 0.10220210254192352,
      "learning_rate": 0.00015413315135522432,
      "loss": 0.3711531162261963,
      "memory(GiB)": 78.33,
      "step": 2667,
      "token_acc": 0.8892863670783613,
      "train_speed(iter/s)": 0.032335
    },
    {
      "epoch": 0.516979121251756,
      "grad_norm": 0.1041957437992096,
      "learning_rate": 0.00015403705504749238,
      "loss": 0.3776172697544098,
      "memory(GiB)": 78.33,
      "step": 2668,
      "token_acc": 0.8862548629964455,
      "train_speed(iter/s)": 0.032336
    },
    {
      "epoch": 0.5171728915370828,
      "grad_norm": 0.10291523486375809,
      "learning_rate": 0.0001539409570816335,
      "loss": 0.36820337176322937,
      "memory(GiB)": 78.33,
      "step": 2669,
      "token_acc": 0.8891734392557556,
      "train_speed(iter/s)": 0.032337
    },
    {
      "epoch": 0.5173666618224095,
      "grad_norm": 0.10148876905441284,
      "learning_rate": 0.00015384485749711768,
      "loss": 0.34953948855400085,
      "memory(GiB)": 78.33,
      "step": 2670,
      "token_acc": 0.8953482075165169,
      "train_speed(iter/s)": 0.032338
    },
    {
      "epoch": 0.5175604321077363,
      "grad_norm": 0.09922674298286438,
      "learning_rate": 0.0001537487563334155,
      "loss": 0.306792676448822,
      "memory(GiB)": 78.33,
      "step": 2671,
      "token_acc": 0.9090061848682538,
      "train_speed(iter/s)": 0.032339
    },
    {
      "epoch": 0.517754202393063,
      "grad_norm": 0.10890819132328033,
      "learning_rate": 0.00015365265362999846,
      "loss": 0.3788120746612549,
      "memory(GiB)": 78.33,
      "step": 2672,
      "token_acc": 0.8892348255357343,
      "train_speed(iter/s)": 0.03234
    },
    {
      "epoch": 0.5179479726783898,
      "grad_norm": 0.09601421654224396,
      "learning_rate": 0.00015355654942633833,
      "loss": 0.3305688500404358,
      "memory(GiB)": 78.33,
      "step": 2673,
      "token_acc": 0.9015699037955296,
      "train_speed(iter/s)": 0.032341
    },
    {
      "epoch": 0.5181417429637165,
      "grad_norm": 0.09569491446018219,
      "learning_rate": 0.00015346044376190782,
      "loss": 0.31505027413368225,
      "memory(GiB)": 78.33,
      "step": 2674,
      "token_acc": 0.9047790339157246,
      "train_speed(iter/s)": 0.032342
    },
    {
      "epoch": 0.5183355132490433,
      "grad_norm": 0.11235766857862473,
      "learning_rate": 0.00015336433667618004,
      "loss": 0.368362158536911,
      "memory(GiB)": 78.33,
      "step": 2675,
      "token_acc": 0.8934327846364883,
      "train_speed(iter/s)": 0.032343
    },
    {
      "epoch": 0.51852928353437,
      "grad_norm": 0.09810005128383636,
      "learning_rate": 0.00015326822820862883,
      "loss": 0.3768079876899719,
      "memory(GiB)": 78.33,
      "step": 2676,
      "token_acc": 0.8896894012167788,
      "train_speed(iter/s)": 0.032344
    },
    {
      "epoch": 0.5187230538196967,
      "grad_norm": 0.1080244854092598,
      "learning_rate": 0.00015317211839872846,
      "loss": 0.3397291898727417,
      "memory(GiB)": 78.33,
      "step": 2677,
      "token_acc": 0.90013633265167,
      "train_speed(iter/s)": 0.032345
    },
    {
      "epoch": 0.5189168241050235,
      "grad_norm": 0.10235600918531418,
      "learning_rate": 0.00015307600728595383,
      "loss": 0.3567368686199188,
      "memory(GiB)": 78.33,
      "step": 2678,
      "token_acc": 0.89522800645682,
      "train_speed(iter/s)": 0.032346
    },
    {
      "epoch": 0.5191105943903502,
      "grad_norm": 0.10997829586267471,
      "learning_rate": 0.00015297989490978037,
      "loss": 0.3763918876647949,
      "memory(GiB)": 78.33,
      "step": 2679,
      "token_acc": 0.8877150980098758,
      "train_speed(iter/s)": 0.032347
    },
    {
      "epoch": 0.519304364675677,
      "grad_norm": 0.11425752192735672,
      "learning_rate": 0.00015288378130968395,
      "loss": 0.350836843252182,
      "memory(GiB)": 78.33,
      "step": 2680,
      "token_acc": 0.8964464422016792,
      "train_speed(iter/s)": 0.032348
    },
    {
      "epoch": 0.5194981349610037,
      "grad_norm": 0.10013525933027267,
      "learning_rate": 0.00015278766652514103,
      "loss": 0.332501083612442,
      "memory(GiB)": 78.33,
      "step": 2681,
      "token_acc": 0.8996237243171544,
      "train_speed(iter/s)": 0.032349
    },
    {
      "epoch": 0.5196919052463305,
      "grad_norm": 0.09917152673006058,
      "learning_rate": 0.00015269155059562863,
      "loss": 0.35868263244628906,
      "memory(GiB)": 78.33,
      "step": 2682,
      "token_acc": 0.8944104296421408,
      "train_speed(iter/s)": 0.03235
    },
    {
      "epoch": 0.5198856755316572,
      "grad_norm": 0.10261884331703186,
      "learning_rate": 0.00015259543356062406,
      "loss": 0.3666459321975708,
      "memory(GiB)": 78.33,
      "step": 2683,
      "token_acc": 0.8901406178581991,
      "train_speed(iter/s)": 0.032351
    },
    {
      "epoch": 0.520079445816984,
      "grad_norm": 0.11286422610282898,
      "learning_rate": 0.00015249931545960517,
      "loss": 0.38164597749710083,
      "memory(GiB)": 78.33,
      "step": 2684,
      "token_acc": 0.8868490627692125,
      "train_speed(iter/s)": 0.032352
    },
    {
      "epoch": 0.5202732161023107,
      "grad_norm": 0.10428116470575333,
      "learning_rate": 0.0001524031963320503,
      "loss": 0.34657108783721924,
      "memory(GiB)": 78.33,
      "step": 2685,
      "token_acc": 0.8960804399396162,
      "train_speed(iter/s)": 0.032353
    },
    {
      "epoch": 0.5204669863876374,
      "grad_norm": 0.10740246623754501,
      "learning_rate": 0.00015230707621743809,
      "loss": 0.3531382381916046,
      "memory(GiB)": 78.33,
      "step": 2686,
      "token_acc": 0.8949211908931699,
      "train_speed(iter/s)": 0.032354
    },
    {
      "epoch": 0.5206607566729642,
      "grad_norm": 0.10802320390939713,
      "learning_rate": 0.00015221095515524768,
      "loss": 0.38364243507385254,
      "memory(GiB)": 78.33,
      "step": 2687,
      "token_acc": 0.8869632473001232,
      "train_speed(iter/s)": 0.032355
    },
    {
      "epoch": 0.5208545269582909,
      "grad_norm": 0.10873539000749588,
      "learning_rate": 0.00015211483318495854,
      "loss": 0.37823835015296936,
      "memory(GiB)": 78.33,
      "step": 2688,
      "token_acc": 0.8876921812052947,
      "train_speed(iter/s)": 0.032356
    },
    {
      "epoch": 0.5210482972436177,
      "grad_norm": 0.09489670395851135,
      "learning_rate": 0.00015201871034605064,
      "loss": 0.32838284969329834,
      "memory(GiB)": 78.33,
      "step": 2689,
      "token_acc": 0.903539469079638,
      "train_speed(iter/s)": 0.032357
    },
    {
      "epoch": 0.5212420675289444,
      "grad_norm": 0.09505011886358261,
      "learning_rate": 0.00015192258667800414,
      "loss": 0.3134308457374573,
      "memory(GiB)": 78.33,
      "step": 2690,
      "token_acc": 0.9066839378238342,
      "train_speed(iter/s)": 0.032358
    },
    {
      "epoch": 0.5214358378142712,
      "grad_norm": 0.10631779581308365,
      "learning_rate": 0.00015182646222029964,
      "loss": 0.39203402400016785,
      "memory(GiB)": 78.33,
      "step": 2691,
      "token_acc": 0.88480466768138,
      "train_speed(iter/s)": 0.032359
    },
    {
      "epoch": 0.5216296080995979,
      "grad_norm": 0.11067686975002289,
      "learning_rate": 0.00015173033701241804,
      "loss": 0.38351666927337646,
      "memory(GiB)": 78.33,
      "step": 2692,
      "token_acc": 0.8888694485755111,
      "train_speed(iter/s)": 0.03236
    },
    {
      "epoch": 0.5218233783849247,
      "grad_norm": 0.11114225536584854,
      "learning_rate": 0.00015163421109384048,
      "loss": 0.3793320655822754,
      "memory(GiB)": 78.33,
      "step": 2693,
      "token_acc": 0.8876760563380282,
      "train_speed(iter/s)": 0.032361
    },
    {
      "epoch": 0.5220171486702514,
      "grad_norm": 0.10015545785427094,
      "learning_rate": 0.0001515380845040485,
      "loss": 0.31924664974212646,
      "memory(GiB)": 78.33,
      "step": 2694,
      "token_acc": 0.9028615097223401,
      "train_speed(iter/s)": 0.032362
    },
    {
      "epoch": 0.5222109189555781,
      "grad_norm": 0.11381295323371887,
      "learning_rate": 0.00015144195728252396,
      "loss": 0.36840251088142395,
      "memory(GiB)": 78.33,
      "step": 2695,
      "token_acc": 0.889958889452047,
      "train_speed(iter/s)": 0.032363
    },
    {
      "epoch": 0.5224046892409049,
      "grad_norm": 0.09790168702602386,
      "learning_rate": 0.00015134582946874875,
      "loss": 0.3553347587585449,
      "memory(GiB)": 78.33,
      "step": 2696,
      "token_acc": 0.8931691146224018,
      "train_speed(iter/s)": 0.032364
    },
    {
      "epoch": 0.5225984595262316,
      "grad_norm": 0.10989172011613846,
      "learning_rate": 0.00015124970110220526,
      "loss": 0.37042251229286194,
      "memory(GiB)": 78.33,
      "step": 2697,
      "token_acc": 0.8891240509008175,
      "train_speed(iter/s)": 0.032365
    },
    {
      "epoch": 0.5227922298115584,
      "grad_norm": 0.11018986999988556,
      "learning_rate": 0.00015115357222237596,
      "loss": 0.34254854917526245,
      "memory(GiB)": 78.33,
      "step": 2698,
      "token_acc": 0.8972499929196522,
      "train_speed(iter/s)": 0.032366
    },
    {
      "epoch": 0.5229860000968851,
      "grad_norm": 0.10905101150274277,
      "learning_rate": 0.00015105744286874354,
      "loss": 0.37283697724342346,
      "memory(GiB)": 78.33,
      "step": 2699,
      "token_acc": 0.8891810881122045,
      "train_speed(iter/s)": 0.032367
    },
    {
      "epoch": 0.5231797703822119,
      "grad_norm": 0.10779301077127457,
      "learning_rate": 0.00015096131308079086,
      "loss": 0.35086601972579956,
      "memory(GiB)": 78.33,
      "step": 2700,
      "token_acc": 0.894660338178582,
      "train_speed(iter/s)": 0.032368
    },
    {
      "epoch": 0.5233735406675386,
      "grad_norm": 0.10483751446008682,
      "learning_rate": 0.00015086518289800108,
      "loss": 0.3449743688106537,
      "memory(GiB)": 78.33,
      "step": 2701,
      "token_acc": 0.8989076300800705,
      "train_speed(iter/s)": 0.032369
    },
    {
      "epoch": 0.5235673109528654,
      "grad_norm": 0.11302126944065094,
      "learning_rate": 0.00015076905235985748,
      "loss": 0.4011099338531494,
      "memory(GiB)": 78.33,
      "step": 2702,
      "token_acc": 0.8809768411165576,
      "train_speed(iter/s)": 0.03237
    },
    {
      "epoch": 0.5237610812381921,
      "grad_norm": 0.09251462668180466,
      "learning_rate": 0.0001506729215058434,
      "loss": 0.2971772253513336,
      "memory(GiB)": 78.33,
      "step": 2703,
      "token_acc": 0.9121730788649904,
      "train_speed(iter/s)": 0.032371
    },
    {
      "epoch": 0.5239548515235188,
      "grad_norm": 0.1044599711894989,
      "learning_rate": 0.0001505767903754424,
      "loss": 0.3731986880302429,
      "memory(GiB)": 78.33,
      "step": 2704,
      "token_acc": 0.8906576333350313,
      "train_speed(iter/s)": 0.032372
    },
    {
      "epoch": 0.5241486218088456,
      "grad_norm": 0.11340048164129257,
      "learning_rate": 0.00015048065900813805,
      "loss": 0.4006243944168091,
      "memory(GiB)": 78.33,
      "step": 2705,
      "token_acc": 0.8819369331891173,
      "train_speed(iter/s)": 0.032373
    },
    {
      "epoch": 0.5243423920941723,
      "grad_norm": 0.1074516549706459,
      "learning_rate": 0.00015038452744341416,
      "loss": 0.3777906894683838,
      "memory(GiB)": 78.33,
      "step": 2706,
      "token_acc": 0.8883070102521474,
      "train_speed(iter/s)": 0.032374
    },
    {
      "epoch": 0.5245361623794991,
      "grad_norm": 0.10201266407966614,
      "learning_rate": 0.00015028839572075447,
      "loss": 0.36712780594825745,
      "memory(GiB)": 78.33,
      "step": 2707,
      "token_acc": 0.8929159371815307,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.5247299326648258,
      "grad_norm": 0.1035100668668747,
      "learning_rate": 0.0001501922638796429,
      "loss": 0.36527976393699646,
      "memory(GiB)": 78.33,
      "step": 2708,
      "token_acc": 0.8913647477217159,
      "train_speed(iter/s)": 0.032376
    },
    {
      "epoch": 0.5249237029501526,
      "grad_norm": 0.1076916977763176,
      "learning_rate": 0.00015009613195956343,
      "loss": 0.38062483072280884,
      "memory(GiB)": 78.33,
      "step": 2709,
      "token_acc": 0.8895733585435028,
      "train_speed(iter/s)": 0.032377
    },
    {
      "epoch": 0.5251174732354793,
      "grad_norm": 0.09896742552518845,
      "learning_rate": 0.00015,
      "loss": 0.3540569841861725,
      "memory(GiB)": 78.33,
      "step": 2710,
      "token_acc": 0.8951074201119336,
      "train_speed(iter/s)": 0.032378
    },
    {
      "epoch": 0.525311243520806,
      "grad_norm": 0.09472720324993134,
      "learning_rate": 0.00014990386804043652,
      "loss": 0.3247393071651459,
      "memory(GiB)": 78.33,
      "step": 2711,
      "token_acc": 0.901641010008637,
      "train_speed(iter/s)": 0.032379
    },
    {
      "epoch": 0.5255050138061328,
      "grad_norm": 0.09038470685482025,
      "learning_rate": 0.0001498077361203571,
      "loss": 0.3078896105289459,
      "memory(GiB)": 78.33,
      "step": 2712,
      "token_acc": 0.9055924034928663,
      "train_speed(iter/s)": 0.03238
    },
    {
      "epoch": 0.5256987840914595,
      "grad_norm": 0.10283378511667252,
      "learning_rate": 0.00014971160427924553,
      "loss": 0.3739337623119354,
      "memory(GiB)": 78.33,
      "step": 2713,
      "token_acc": 0.8916592017517667,
      "train_speed(iter/s)": 0.032381
    },
    {
      "epoch": 0.5258925543767863,
      "grad_norm": 0.10884794592857361,
      "learning_rate": 0.00014961547255658587,
      "loss": 0.3748420178890228,
      "memory(GiB)": 78.33,
      "step": 2714,
      "token_acc": 0.890541823727448,
      "train_speed(iter/s)": 0.032382
    },
    {
      "epoch": 0.526086324662113,
      "grad_norm": 0.10610167682170868,
      "learning_rate": 0.00014951934099186195,
      "loss": 0.39279741048812866,
      "memory(GiB)": 78.33,
      "step": 2715,
      "token_acc": 0.8841656649546498,
      "train_speed(iter/s)": 0.032383
    },
    {
      "epoch": 0.5262800949474398,
      "grad_norm": 0.10813671350479126,
      "learning_rate": 0.00014942320962455766,
      "loss": 0.36780744791030884,
      "memory(GiB)": 78.33,
      "step": 2716,
      "token_acc": 0.8913100322496296,
      "train_speed(iter/s)": 0.032384
    },
    {
      "epoch": 0.5264738652327665,
      "grad_norm": 0.103311687707901,
      "learning_rate": 0.0001493270784941566,
      "loss": 0.37279364466667175,
      "memory(GiB)": 78.33,
      "step": 2717,
      "token_acc": 0.8909463356782836,
      "train_speed(iter/s)": 0.032385
    },
    {
      "epoch": 0.5266676355180933,
      "grad_norm": 0.10846543312072754,
      "learning_rate": 0.00014923094764014247,
      "loss": 0.36826059222221375,
      "memory(GiB)": 78.33,
      "step": 2718,
      "token_acc": 0.8907717095394189,
      "train_speed(iter/s)": 0.032386
    },
    {
      "epoch": 0.52686140580342,
      "grad_norm": 0.09876382350921631,
      "learning_rate": 0.0001491348171019989,
      "loss": 0.358868807554245,
      "memory(GiB)": 78.33,
      "step": 2719,
      "token_acc": 0.8912382783074005,
      "train_speed(iter/s)": 0.032387
    },
    {
      "epoch": 0.5270551760887467,
      "grad_norm": 0.09218183159828186,
      "learning_rate": 0.00014903868691920911,
      "loss": 0.31816765666007996,
      "memory(GiB)": 78.33,
      "step": 2720,
      "token_acc": 0.9072353603603603,
      "train_speed(iter/s)": 0.032388
    },
    {
      "epoch": 0.5272489463740735,
      "grad_norm": 0.10111497342586517,
      "learning_rate": 0.0001489425571312565,
      "loss": 0.3605380952358246,
      "memory(GiB)": 78.33,
      "step": 2721,
      "token_acc": 0.8904896257271978,
      "train_speed(iter/s)": 0.032389
    },
    {
      "epoch": 0.5274427166594002,
      "grad_norm": 0.10805071890354156,
      "learning_rate": 0.00014884642777762404,
      "loss": 0.35685086250305176,
      "memory(GiB)": 78.33,
      "step": 2722,
      "token_acc": 0.8938120928038679,
      "train_speed(iter/s)": 0.03239
    },
    {
      "epoch": 0.527636486944727,
      "grad_norm": 0.10611846297979355,
      "learning_rate": 0.00014875029889779476,
      "loss": 0.366268128156662,
      "memory(GiB)": 78.33,
      "step": 2723,
      "token_acc": 0.8927133114965365,
      "train_speed(iter/s)": 0.032391
    },
    {
      "epoch": 0.5278302572300537,
      "grad_norm": 0.10951778292655945,
      "learning_rate": 0.00014865417053125122,
      "loss": 0.35670673847198486,
      "memory(GiB)": 78.33,
      "step": 2724,
      "token_acc": 0.8942021014849336,
      "train_speed(iter/s)": 0.032392
    },
    {
      "epoch": 0.5280240275153805,
      "grad_norm": 0.10806987434625626,
      "learning_rate": 0.000148558042717476,
      "loss": 0.34922337532043457,
      "memory(GiB)": 78.33,
      "step": 2725,
      "token_acc": 0.8949587478767289,
      "train_speed(iter/s)": 0.032393
    },
    {
      "epoch": 0.5282177978007072,
      "grad_norm": 0.09839235991239548,
      "learning_rate": 0.0001484619154959515,
      "loss": 0.34449273347854614,
      "memory(GiB)": 78.33,
      "step": 2726,
      "token_acc": 0.8982513498786348,
      "train_speed(iter/s)": 0.032394
    },
    {
      "epoch": 0.528411568086034,
      "grad_norm": 0.10562780499458313,
      "learning_rate": 0.00014836578890615952,
      "loss": 0.34781414270401,
      "memory(GiB)": 78.33,
      "step": 2727,
      "token_acc": 0.8980476002925787,
      "train_speed(iter/s)": 0.032395
    },
    {
      "epoch": 0.5286053383713607,
      "grad_norm": 0.11507020145654678,
      "learning_rate": 0.00014826966298758202,
      "loss": 0.3725394010543823,
      "memory(GiB)": 78.33,
      "step": 2728,
      "token_acc": 0.8897241588360109,
      "train_speed(iter/s)": 0.032396
    },
    {
      "epoch": 0.5287991086566874,
      "grad_norm": 0.1054215356707573,
      "learning_rate": 0.00014817353777970036,
      "loss": 0.3395271897315979,
      "memory(GiB)": 78.33,
      "step": 2729,
      "token_acc": 0.8995790406558103,
      "train_speed(iter/s)": 0.032397
    },
    {
      "epoch": 0.5289928789420142,
      "grad_norm": 0.11207542568445206,
      "learning_rate": 0.00014807741332199584,
      "loss": 0.3861173093318939,
      "memory(GiB)": 78.33,
      "step": 2730,
      "token_acc": 0.8875055334218681,
      "train_speed(iter/s)": 0.032398
    },
    {
      "epoch": 0.5291866492273409,
      "grad_norm": 0.11059516668319702,
      "learning_rate": 0.00014798128965394936,
      "loss": 0.3895573019981384,
      "memory(GiB)": 78.33,
      "step": 2731,
      "token_acc": 0.8851554949115925,
      "train_speed(iter/s)": 0.032399
    },
    {
      "epoch": 0.5293804195126677,
      "grad_norm": 0.1085188090801239,
      "learning_rate": 0.0001478851668150414,
      "loss": 0.37232252955436707,
      "memory(GiB)": 78.33,
      "step": 2732,
      "token_acc": 0.8892735392827885,
      "train_speed(iter/s)": 0.0324
    },
    {
      "epoch": 0.5295741897979945,
      "grad_norm": 0.09967105835676193,
      "learning_rate": 0.00014778904484475235,
      "loss": 0.3428942561149597,
      "memory(GiB)": 78.33,
      "step": 2733,
      "token_acc": 0.8976353126642144,
      "train_speed(iter/s)": 0.032401
    },
    {
      "epoch": 0.5297679600833213,
      "grad_norm": 0.11610176414251328,
      "learning_rate": 0.0001476929237825619,
      "loss": 0.40362459421157837,
      "memory(GiB)": 78.33,
      "step": 2734,
      "token_acc": 0.8841030058763807,
      "train_speed(iter/s)": 0.032402
    },
    {
      "epoch": 0.529961730368648,
      "grad_norm": 0.10144146531820297,
      "learning_rate": 0.00014759680366794974,
      "loss": 0.3451736271381378,
      "memory(GiB)": 78.33,
      "step": 2735,
      "token_acc": 0.8984730993684548,
      "train_speed(iter/s)": 0.032403
    },
    {
      "epoch": 0.5301555006539748,
      "grad_norm": 0.097495436668396,
      "learning_rate": 0.0001475006845403948,
      "loss": 0.3148418366909027,
      "memory(GiB)": 78.33,
      "step": 2736,
      "token_acc": 0.9057284734309409,
      "train_speed(iter/s)": 0.032404
    },
    {
      "epoch": 0.5303492709393015,
      "grad_norm": 0.10190415382385254,
      "learning_rate": 0.00014740456643937591,
      "loss": 0.35173338651657104,
      "memory(GiB)": 78.33,
      "step": 2737,
      "token_acc": 0.895157419172667,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.5305430412246283,
      "grad_norm": 0.118934765458107,
      "learning_rate": 0.00014730844940437138,
      "loss": 0.3918536901473999,
      "memory(GiB)": 78.33,
      "step": 2738,
      "token_acc": 0.8861999064400436,
      "train_speed(iter/s)": 0.032406
    },
    {
      "epoch": 0.530736811509955,
      "grad_norm": 0.10452469438314438,
      "learning_rate": 0.00014721233347485892,
      "loss": 0.36255908012390137,
      "memory(GiB)": 78.33,
      "step": 2739,
      "token_acc": 0.8927415372798595,
      "train_speed(iter/s)": 0.032407
    },
    {
      "epoch": 0.5309305817952817,
      "grad_norm": 0.10265407711267471,
      "learning_rate": 0.00014711621869031608,
      "loss": 0.3520941436290741,
      "memory(GiB)": 78.33,
      "step": 2740,
      "token_acc": 0.8986634006070611,
      "train_speed(iter/s)": 0.032408
    },
    {
      "epoch": 0.5311243520806085,
      "grad_norm": 0.10262811183929443,
      "learning_rate": 0.00014702010509021963,
      "loss": 0.33090177178382874,
      "memory(GiB)": 78.33,
      "step": 2741,
      "token_acc": 0.9019104647847163,
      "train_speed(iter/s)": 0.032409
    },
    {
      "epoch": 0.5313181223659352,
      "grad_norm": 0.10777027159929276,
      "learning_rate": 0.0001469239927140462,
      "loss": 0.3813689649105072,
      "memory(GiB)": 78.33,
      "step": 2742,
      "token_acc": 0.8873789612676056,
      "train_speed(iter/s)": 0.03241
    },
    {
      "epoch": 0.531511892651262,
      "grad_norm": 0.10505778342485428,
      "learning_rate": 0.00014682788160127154,
      "loss": 0.37121495604515076,
      "memory(GiB)": 78.33,
      "step": 2743,
      "token_acc": 0.8890379187973286,
      "train_speed(iter/s)": 0.032411
    },
    {
      "epoch": 0.5317056629365887,
      "grad_norm": 0.11017142981290817,
      "learning_rate": 0.00014673177179137114,
      "loss": 0.3495400547981262,
      "memory(GiB)": 78.33,
      "step": 2744,
      "token_acc": 0.8959156433529097,
      "train_speed(iter/s)": 0.032412
    },
    {
      "epoch": 0.5318994332219155,
      "grad_norm": 0.11223854124546051,
      "learning_rate": 0.00014663566332381994,
      "loss": 0.3930763900279999,
      "memory(GiB)": 78.33,
      "step": 2745,
      "token_acc": 0.8840929833038063,
      "train_speed(iter/s)": 0.032413
    },
    {
      "epoch": 0.5320932035072422,
      "grad_norm": 0.11058208346366882,
      "learning_rate": 0.00014653955623809215,
      "loss": 0.3463561534881592,
      "memory(GiB)": 78.33,
      "step": 2746,
      "token_acc": 0.8972069472423368,
      "train_speed(iter/s)": 0.032413
    },
    {
      "epoch": 0.532286973792569,
      "grad_norm": 0.10111848264932632,
      "learning_rate": 0.00014644345057366167,
      "loss": 0.33341148495674133,
      "memory(GiB)": 78.33,
      "step": 2747,
      "token_acc": 0.9006800088456435,
      "train_speed(iter/s)": 0.032414
    },
    {
      "epoch": 0.5324807440778957,
      "grad_norm": 0.10276893526315689,
      "learning_rate": 0.00014634734637000154,
      "loss": 0.35466858744621277,
      "memory(GiB)": 78.33,
      "step": 2748,
      "token_acc": 0.8964071405257475,
      "train_speed(iter/s)": 0.032415
    },
    {
      "epoch": 0.5326745143632224,
      "grad_norm": 0.09825938940048218,
      "learning_rate": 0.0001462512436665845,
      "loss": 0.35709887742996216,
      "memory(GiB)": 78.33,
      "step": 2749,
      "token_acc": 0.8944281524926686,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.5328682846485492,
      "grad_norm": 0.1046825423836708,
      "learning_rate": 0.00014615514250288232,
      "loss": 0.3530852496623993,
      "memory(GiB)": 78.33,
      "step": 2750,
      "token_acc": 0.8973124966203428,
      "train_speed(iter/s)": 0.032417
    },
    {
      "epoch": 0.5330620549338759,
      "grad_norm": 0.09290501475334167,
      "learning_rate": 0.00014605904291836643,
      "loss": 0.3255302309989929,
      "memory(GiB)": 78.33,
      "step": 2751,
      "token_acc": 0.9017647611001596,
      "train_speed(iter/s)": 0.032418
    },
    {
      "epoch": 0.5332558252192027,
      "grad_norm": 0.09813160449266434,
      "learning_rate": 0.0001459629449525076,
      "loss": 0.35900381207466125,
      "memory(GiB)": 78.33,
      "step": 2752,
      "token_acc": 0.8953540595838115,
      "train_speed(iter/s)": 0.032419
    },
    {
      "epoch": 0.5334495955045294,
      "grad_norm": 0.10881511121988297,
      "learning_rate": 0.0001458668486447757,
      "loss": 0.3696974217891693,
      "memory(GiB)": 78.33,
      "step": 2753,
      "token_acc": 0.8900663072442213,
      "train_speed(iter/s)": 0.03242
    },
    {
      "epoch": 0.5336433657898562,
      "grad_norm": 0.09696649760007858,
      "learning_rate": 0.00014577075403464013,
      "loss": 0.3419326841831207,
      "memory(GiB)": 78.33,
      "step": 2754,
      "token_acc": 0.8965491278704982,
      "train_speed(iter/s)": 0.032421
    },
    {
      "epoch": 0.5338371360751829,
      "grad_norm": 0.10864214599132538,
      "learning_rate": 0.0001456746611615695,
      "loss": 0.3853573799133301,
      "memory(GiB)": 78.33,
      "step": 2755,
      "token_acc": 0.8867099400407064,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.5340309063605096,
      "grad_norm": 0.105714350938797,
      "learning_rate": 0.00014557857006503182,
      "loss": 0.3608066737651825,
      "memory(GiB)": 78.33,
      "step": 2756,
      "token_acc": 0.8907904278462654,
      "train_speed(iter/s)": 0.032423
    },
    {
      "epoch": 0.5342246766458364,
      "grad_norm": 0.10220605880022049,
      "learning_rate": 0.00014548248078449417,
      "loss": 0.35340073704719543,
      "memory(GiB)": 78.33,
      "step": 2757,
      "token_acc": 0.8937097622685939,
      "train_speed(iter/s)": 0.032424
    },
    {
      "epoch": 0.5344184469311631,
      "grad_norm": 0.12155355513095856,
      "learning_rate": 0.00014538639335942303,
      "loss": 0.3872877061367035,
      "memory(GiB)": 78.33,
      "step": 2758,
      "token_acc": 0.8883419182369433,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.5346122172164899,
      "grad_norm": 0.11298363655805588,
      "learning_rate": 0.0001452903078292842,
      "loss": 0.36656370759010315,
      "memory(GiB)": 78.33,
      "step": 2759,
      "token_acc": 0.8930003250623036,
      "train_speed(iter/s)": 0.032426
    },
    {
      "epoch": 0.5348059875018166,
      "grad_norm": 0.1053992435336113,
      "learning_rate": 0.00014519422423354243,
      "loss": 0.3747759163379669,
      "memory(GiB)": 78.33,
      "step": 2760,
      "token_acc": 0.8891952102478419,
      "train_speed(iter/s)": 0.032427
    },
    {
      "epoch": 0.5349997577871434,
      "grad_norm": 0.09907464683055878,
      "learning_rate": 0.00014509814261166193,
      "loss": 0.33399906754493713,
      "memory(GiB)": 78.33,
      "step": 2761,
      "token_acc": 0.8976413830072366,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.5351935280724701,
      "grad_norm": 0.0989702120423317,
      "learning_rate": 0.00014500206300310594,
      "loss": 0.3469410538673401,
      "memory(GiB)": 78.33,
      "step": 2762,
      "token_acc": 0.8955082378865913,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.5353872983577969,
      "grad_norm": 0.09886979311704636,
      "learning_rate": 0.00014490598544733695,
      "loss": 0.3301083445549011,
      "memory(GiB)": 78.33,
      "step": 2763,
      "token_acc": 0.903747359636999,
      "train_speed(iter/s)": 0.032429
    },
    {
      "epoch": 0.5355810686431236,
      "grad_norm": 0.11633176356554031,
      "learning_rate": 0.00014480990998381674,
      "loss": 0.36927592754364014,
      "memory(GiB)": 78.33,
      "step": 2764,
      "token_acc": 0.8909205359551572,
      "train_speed(iter/s)": 0.03243
    },
    {
      "epoch": 0.5357748389284503,
      "grad_norm": 0.10796020925045013,
      "learning_rate": 0.00014471383665200585,
      "loss": 0.3573460876941681,
      "memory(GiB)": 78.33,
      "step": 2765,
      "token_acc": 0.8945837063563116,
      "train_speed(iter/s)": 0.032431
    },
    {
      "epoch": 0.5359686092137771,
      "grad_norm": 0.10915590077638626,
      "learning_rate": 0.00014461776549136435,
      "loss": 0.3306170701980591,
      "memory(GiB)": 78.33,
      "step": 2766,
      "token_acc": 0.9005574136008918,
      "train_speed(iter/s)": 0.032432
    },
    {
      "epoch": 0.5361623794991038,
      "grad_norm": 0.09537974745035172,
      "learning_rate": 0.00014452169654135115,
      "loss": 0.3284014165401459,
      "memory(GiB)": 78.33,
      "step": 2767,
      "token_acc": 0.9025765702541481,
      "train_speed(iter/s)": 0.032433
    },
    {
      "epoch": 0.5363561497844306,
      "grad_norm": 0.09599710255861282,
      "learning_rate": 0.00014442562984142446,
      "loss": 0.34253832697868347,
      "memory(GiB)": 78.33,
      "step": 2768,
      "token_acc": 0.8958434143036872,
      "train_speed(iter/s)": 0.032434
    },
    {
      "epoch": 0.5365499200697573,
      "grad_norm": 0.10483718663454056,
      "learning_rate": 0.0001443295654310413,
      "loss": 0.3458617329597473,
      "memory(GiB)": 78.33,
      "step": 2769,
      "token_acc": 0.8977225020990764,
      "train_speed(iter/s)": 0.032435
    },
    {
      "epoch": 0.5367436903550841,
      "grad_norm": 0.10079243779182434,
      "learning_rate": 0.000144233503349658,
      "loss": 0.3476545810699463,
      "memory(GiB)": 78.33,
      "step": 2770,
      "token_acc": 0.8959448042804844,
      "train_speed(iter/s)": 0.032436
    },
    {
      "epoch": 0.5369374606404108,
      "grad_norm": 0.10120805352926254,
      "learning_rate": 0.00014413744363672988,
      "loss": 0.35417628288269043,
      "memory(GiB)": 78.33,
      "step": 2771,
      "token_acc": 0.8963836812952706,
      "train_speed(iter/s)": 0.032437
    },
    {
      "epoch": 0.5371312309257376,
      "grad_norm": 0.10923577845096588,
      "learning_rate": 0.00014404138633171114,
      "loss": 0.388058602809906,
      "memory(GiB)": 78.33,
      "step": 2772,
      "token_acc": 0.8873768455115426,
      "train_speed(iter/s)": 0.032438
    },
    {
      "epoch": 0.5373250012110643,
      "grad_norm": 0.10782677680253983,
      "learning_rate": 0.00014394533147405519,
      "loss": 0.3875938057899475,
      "memory(GiB)": 78.33,
      "step": 2773,
      "token_acc": 0.8839462734913853,
      "train_speed(iter/s)": 0.032439
    },
    {
      "epoch": 0.537518771496391,
      "grad_norm": 0.10002797842025757,
      "learning_rate": 0.00014384927910321424,
      "loss": 0.35020288825035095,
      "memory(GiB)": 78.33,
      "step": 2774,
      "token_acc": 0.8941337890860073,
      "train_speed(iter/s)": 0.03244
    },
    {
      "epoch": 0.5377125417817178,
      "grad_norm": 0.09954270720481873,
      "learning_rate": 0.0001437532292586397,
      "loss": 0.3441551625728607,
      "memory(GiB)": 78.33,
      "step": 2775,
      "token_acc": 0.9002784888463811,
      "train_speed(iter/s)": 0.032441
    },
    {
      "epoch": 0.5379063120670445,
      "grad_norm": 0.10037088394165039,
      "learning_rate": 0.00014365718197978172,
      "loss": 0.34260034561157227,
      "memory(GiB)": 78.33,
      "step": 2776,
      "token_acc": 0.89792182924082,
      "train_speed(iter/s)": 0.032442
    },
    {
      "epoch": 0.5381000823523713,
      "grad_norm": 0.1107875406742096,
      "learning_rate": 0.00014356113730608954,
      "loss": 0.3863231837749481,
      "memory(GiB)": 78.33,
      "step": 2777,
      "token_acc": 0.8852114077738917,
      "train_speed(iter/s)": 0.032443
    },
    {
      "epoch": 0.538293852637698,
      "grad_norm": 0.10534848272800446,
      "learning_rate": 0.00014346509527701133,
      "loss": 0.38511624932289124,
      "memory(GiB)": 78.33,
      "step": 2778,
      "token_acc": 0.8850161787561467,
      "train_speed(iter/s)": 0.032444
    },
    {
      "epoch": 0.5384876229230248,
      "grad_norm": 0.10484083741903305,
      "learning_rate": 0.00014336905593199405,
      "loss": 0.3411652445793152,
      "memory(GiB)": 78.33,
      "step": 2779,
      "token_acc": 0.8983436180578109,
      "train_speed(iter/s)": 0.032445
    },
    {
      "epoch": 0.5386813932083515,
      "grad_norm": 0.09589672088623047,
      "learning_rate": 0.00014327301931048368,
      "loss": 0.32667702436447144,
      "memory(GiB)": 78.33,
      "step": 2780,
      "token_acc": 0.902627403397424,
      "train_speed(iter/s)": 0.032445
    },
    {
      "epoch": 0.5388751634936783,
      "grad_norm": 0.11629176884889603,
      "learning_rate": 0.00014317698545192504,
      "loss": 0.39405229687690735,
      "memory(GiB)": 78.33,
      "step": 2781,
      "token_acc": 0.8830865833285327,
      "train_speed(iter/s)": 0.032446
    },
    {
      "epoch": 0.539068933779005,
      "grad_norm": 0.11103569716215134,
      "learning_rate": 0.00014308095439576188,
      "loss": 0.37723052501678467,
      "memory(GiB)": 78.33,
      "step": 2782,
      "token_acc": 0.8878711653086893,
      "train_speed(iter/s)": 0.032447
    },
    {
      "epoch": 0.5392627040643317,
      "grad_norm": 0.10809724777936935,
      "learning_rate": 0.00014298492618143658,
      "loss": 0.3773004710674286,
      "memory(GiB)": 78.33,
      "step": 2783,
      "token_acc": 0.8881794280195311,
      "train_speed(iter/s)": 0.032448
    },
    {
      "epoch": 0.5394564743496585,
      "grad_norm": 0.10347483307123184,
      "learning_rate": 0.0001428889008483906,
      "loss": 0.3722970187664032,
      "memory(GiB)": 78.33,
      "step": 2784,
      "token_acc": 0.8895179639134346,
      "train_speed(iter/s)": 0.032449
    },
    {
      "epoch": 0.5396502446349852,
      "grad_norm": 0.10195406526327133,
      "learning_rate": 0.00014279287843606422,
      "loss": 0.3863418996334076,
      "memory(GiB)": 78.33,
      "step": 2785,
      "token_acc": 0.8859166207897503,
      "train_speed(iter/s)": 0.03245
    },
    {
      "epoch": 0.539844014920312,
      "grad_norm": 0.10171383619308472,
      "learning_rate": 0.00014269685898389624,
      "loss": 0.3722241222858429,
      "memory(GiB)": 78.33,
      "step": 2786,
      "token_acc": 0.889989379306938,
      "train_speed(iter/s)": 0.032451
    },
    {
      "epoch": 0.5400377852056387,
      "grad_norm": 0.10024231672286987,
      "learning_rate": 0.00014260084253132457,
      "loss": 0.3747716546058655,
      "memory(GiB)": 78.33,
      "step": 2787,
      "token_acc": 0.8894973834732871,
      "train_speed(iter/s)": 0.032452
    },
    {
      "epoch": 0.5402315554909655,
      "grad_norm": 0.09642113000154495,
      "learning_rate": 0.00014250482911778563,
      "loss": 0.34401756525039673,
      "memory(GiB)": 78.33,
      "step": 2788,
      "token_acc": 0.8965368793647757,
      "train_speed(iter/s)": 0.032453
    },
    {
      "epoch": 0.5404253257762922,
      "grad_norm": 0.09063483029603958,
      "learning_rate": 0.00014240881878271487,
      "loss": 0.31945639848709106,
      "memory(GiB)": 78.33,
      "step": 2789,
      "token_acc": 0.9047063862187114,
      "train_speed(iter/s)": 0.032454
    },
    {
      "epoch": 0.540619096061619,
      "grad_norm": 0.1115802749991417,
      "learning_rate": 0.00014231281156554615,
      "loss": 0.37289944291114807,
      "memory(GiB)": 78.33,
      "step": 2790,
      "token_acc": 0.8908885411232829,
      "train_speed(iter/s)": 0.032455
    },
    {
      "epoch": 0.5408128663469457,
      "grad_norm": 0.1019621267914772,
      "learning_rate": 0.00014221680750571228,
      "loss": 0.3510439693927765,
      "memory(GiB)": 78.33,
      "step": 2791,
      "token_acc": 0.8948518823287389,
      "train_speed(iter/s)": 0.032456
    },
    {
      "epoch": 0.5410066366322724,
      "grad_norm": 0.09841447323560715,
      "learning_rate": 0.00014212080664264477,
      "loss": 0.3427790403366089,
      "memory(GiB)": 78.33,
      "step": 2792,
      "token_acc": 0.8993469501181047,
      "train_speed(iter/s)": 0.032456
    },
    {
      "epoch": 0.5412004069175992,
      "grad_norm": 0.10309161990880966,
      "learning_rate": 0.00014202480901577362,
      "loss": 0.34060966968536377,
      "memory(GiB)": 78.33,
      "step": 2793,
      "token_acc": 0.900046502721777,
      "train_speed(iter/s)": 0.032457
    },
    {
      "epoch": 0.5413941772029259,
      "grad_norm": 0.10474463552236557,
      "learning_rate": 0.00014192881466452775,
      "loss": 0.3322206437587738,
      "memory(GiB)": 78.33,
      "step": 2794,
      "token_acc": 0.8990165235382479,
      "train_speed(iter/s)": 0.032458
    },
    {
      "epoch": 0.5415879474882527,
      "grad_norm": 0.11492858082056046,
      "learning_rate": 0.00014183282362833455,
      "loss": 0.37814396619796753,
      "memory(GiB)": 78.33,
      "step": 2795,
      "token_acc": 0.8911760892116183,
      "train_speed(iter/s)": 0.032459
    },
    {
      "epoch": 0.5417817177735794,
      "grad_norm": 0.11693891882896423,
      "learning_rate": 0.00014173683594662014,
      "loss": 0.3709845244884491,
      "memory(GiB)": 78.33,
      "step": 2796,
      "token_acc": 0.8902887474818534,
      "train_speed(iter/s)": 0.03246
    },
    {
      "epoch": 0.5419754880589062,
      "grad_norm": 0.1105792298913002,
      "learning_rate": 0.00014164085165880932,
      "loss": 0.3588966131210327,
      "memory(GiB)": 78.33,
      "step": 2797,
      "token_acc": 0.8919163847269213,
      "train_speed(iter/s)": 0.032461
    },
    {
      "epoch": 0.5421692583442329,
      "grad_norm": 0.10854232311248779,
      "learning_rate": 0.00014154487080432528,
      "loss": 0.3848278820514679,
      "memory(GiB)": 78.33,
      "step": 2798,
      "token_acc": 0.8873765907664213,
      "train_speed(iter/s)": 0.032462
    },
    {
      "epoch": 0.5423630286295597,
      "grad_norm": 0.09642491489648819,
      "learning_rate": 0.00014144889342259002,
      "loss": 0.3380579352378845,
      "memory(GiB)": 78.33,
      "step": 2799,
      "token_acc": 0.8996538924558587,
      "train_speed(iter/s)": 0.032463
    },
    {
      "epoch": 0.5425567989148864,
      "grad_norm": 0.10592179000377655,
      "learning_rate": 0.000141352919553024,
      "loss": 0.37832123041152954,
      "memory(GiB)": 78.33,
      "step": 2800,
      "token_acc": 0.8886095633018012,
      "train_speed(iter/s)": 0.032464
    },
    {
      "epoch": 0.5427505692002131,
      "grad_norm": 0.10104614496231079,
      "learning_rate": 0.0001412569492350463,
      "loss": 0.36123543977737427,
      "memory(GiB)": 78.33,
      "step": 2801,
      "token_acc": 0.8938950807577793,
      "train_speed(iter/s)": 0.03246
    },
    {
      "epoch": 0.5429443394855399,
      "grad_norm": 0.10246588289737701,
      "learning_rate": 0.00014116098250807445,
      "loss": 0.35974156856536865,
      "memory(GiB)": 78.33,
      "step": 2802,
      "token_acc": 0.893833531441162,
      "train_speed(iter/s)": 0.032461
    },
    {
      "epoch": 0.5431381097708666,
      "grad_norm": 0.10312814265489578,
      "learning_rate": 0.00014106501941152459,
      "loss": 0.357723206281662,
      "memory(GiB)": 78.33,
      "step": 2803,
      "token_acc": 0.8940801971556195,
      "train_speed(iter/s)": 0.032462
    },
    {
      "epoch": 0.5433318800561934,
      "grad_norm": 0.10023822635412216,
      "learning_rate": 0.0001409690599848114,
      "loss": 0.33914467692375183,
      "memory(GiB)": 78.33,
      "step": 2804,
      "token_acc": 0.8970815359216062,
      "train_speed(iter/s)": 0.032463
    },
    {
      "epoch": 0.5435256503415201,
      "grad_norm": 0.10362284630537033,
      "learning_rate": 0.0001408731042673479,
      "loss": 0.3780994415283203,
      "memory(GiB)": 78.33,
      "step": 2805,
      "token_acc": 0.8886399282993502,
      "train_speed(iter/s)": 0.032464
    },
    {
      "epoch": 0.5437194206268469,
      "grad_norm": 0.10058917850255966,
      "learning_rate": 0.00014077715229854576,
      "loss": 0.34822142124176025,
      "memory(GiB)": 78.33,
      "step": 2806,
      "token_acc": 0.8962744785862123,
      "train_speed(iter/s)": 0.032465
    },
    {
      "epoch": 0.5439131909121736,
      "grad_norm": 0.10602036118507385,
      "learning_rate": 0.00014068120411781497,
      "loss": 0.3826134502887726,
      "memory(GiB)": 78.33,
      "step": 2807,
      "token_acc": 0.8868417471764866,
      "train_speed(iter/s)": 0.032466
    },
    {
      "epoch": 0.5441069611975003,
      "grad_norm": 0.09917270392179489,
      "learning_rate": 0.0001405852597645641,
      "loss": 0.35776597261428833,
      "memory(GiB)": 78.33,
      "step": 2808,
      "token_acc": 0.8961601781909743,
      "train_speed(iter/s)": 0.032467
    },
    {
      "epoch": 0.5443007314828271,
      "grad_norm": 0.20605452358722687,
      "learning_rate": 0.00014048931927819995,
      "loss": 0.3961770832538605,
      "memory(GiB)": 78.33,
      "step": 2809,
      "token_acc": 0.8819182185180362,
      "train_speed(iter/s)": 0.032467
    },
    {
      "epoch": 0.5444945017681538,
      "grad_norm": 0.11399701237678528,
      "learning_rate": 0.00014039338269812796,
      "loss": 0.39141345024108887,
      "memory(GiB)": 78.33,
      "step": 2810,
      "token_acc": 0.8836162513606542,
      "train_speed(iter/s)": 0.032468
    },
    {
      "epoch": 0.5446882720534806,
      "grad_norm": 0.10171796381473541,
      "learning_rate": 0.0001402974500637519,
      "loss": 0.3601566255092621,
      "memory(GiB)": 78.33,
      "step": 2811,
      "token_acc": 0.8961741354505248,
      "train_speed(iter/s)": 0.032469
    },
    {
      "epoch": 0.5448820423388073,
      "grad_norm": 0.09617502242326736,
      "learning_rate": 0.00014020152141447375,
      "loss": 0.34712108969688416,
      "memory(GiB)": 78.33,
      "step": 2812,
      "token_acc": 0.8963927855711423,
      "train_speed(iter/s)": 0.03247
    },
    {
      "epoch": 0.5450758126241341,
      "grad_norm": 0.10345807671546936,
      "learning_rate": 0.00014010559678969407,
      "loss": 0.36251404881477356,
      "memory(GiB)": 78.33,
      "step": 2813,
      "token_acc": 0.8929795640683466,
      "train_speed(iter/s)": 0.032471
    },
    {
      "epoch": 0.5452695829094608,
      "grad_norm": 0.09582247585058212,
      "learning_rate": 0.00014000967622881166,
      "loss": 0.3179134130477905,
      "memory(GiB)": 78.33,
      "step": 2814,
      "token_acc": 0.9051344993351744,
      "train_speed(iter/s)": 0.032472
    },
    {
      "epoch": 0.5454633531947876,
      "grad_norm": 0.125186488032341,
      "learning_rate": 0.0001399137597712237,
      "loss": 0.36605405807495117,
      "memory(GiB)": 78.33,
      "step": 2815,
      "token_acc": 0.8923891831614162,
      "train_speed(iter/s)": 0.032473
    },
    {
      "epoch": 0.5456571234801143,
      "grad_norm": 0.10813166946172714,
      "learning_rate": 0.00013981784745632558,
      "loss": 0.36997631192207336,
      "memory(GiB)": 78.33,
      "step": 2816,
      "token_acc": 0.8912589239073655,
      "train_speed(iter/s)": 0.032474
    },
    {
      "epoch": 0.545850893765441,
      "grad_norm": 0.10874950140714645,
      "learning_rate": 0.00013972193932351113,
      "loss": 0.36081942915916443,
      "memory(GiB)": 78.33,
      "step": 2817,
      "token_acc": 0.89447387926538,
      "train_speed(iter/s)": 0.032475
    },
    {
      "epoch": 0.5460446640507678,
      "grad_norm": 0.09806004166603088,
      "learning_rate": 0.00013962603541217244,
      "loss": 0.34919509291648865,
      "memory(GiB)": 78.33,
      "step": 2818,
      "token_acc": 0.8967576382560315,
      "train_speed(iter/s)": 0.032476
    },
    {
      "epoch": 0.5462384343360945,
      "grad_norm": 0.1022680252790451,
      "learning_rate": 0.0001395301357616997,
      "loss": 0.3764371871948242,
      "memory(GiB)": 78.33,
      "step": 2819,
      "token_acc": 0.8876494023904382,
      "train_speed(iter/s)": 0.032477
    },
    {
      "epoch": 0.5464322046214213,
      "grad_norm": 0.10649038106203079,
      "learning_rate": 0.00013943424041148154,
      "loss": 0.38199445605278015,
      "memory(GiB)": 78.33,
      "step": 2820,
      "token_acc": 0.8861415110903026,
      "train_speed(iter/s)": 0.032478
    },
    {
      "epoch": 0.546625974906748,
      "grad_norm": 0.10255075246095657,
      "learning_rate": 0.00013933834940090475,
      "loss": 0.36467188596725464,
      "memory(GiB)": 78.33,
      "step": 2821,
      "token_acc": 0.8910845821691643,
      "train_speed(iter/s)": 0.032479
    },
    {
      "epoch": 0.5468197451920748,
      "grad_norm": 0.10376548022031784,
      "learning_rate": 0.00013924246276935442,
      "loss": 0.32367387413978577,
      "memory(GiB)": 78.33,
      "step": 2822,
      "token_acc": 0.9025662838508436,
      "train_speed(iter/s)": 0.03248
    },
    {
      "epoch": 0.5470135154774015,
      "grad_norm": 0.10719739645719528,
      "learning_rate": 0.00013914658055621363,
      "loss": 0.3771483302116394,
      "memory(GiB)": 78.33,
      "step": 2823,
      "token_acc": 0.8899454274847032,
      "train_speed(iter/s)": 0.03248
    },
    {
      "epoch": 0.5472072857627283,
      "grad_norm": 0.11074833571910858,
      "learning_rate": 0.00013905070280086386,
      "loss": 0.36725014448165894,
      "memory(GiB)": 78.33,
      "step": 2824,
      "token_acc": 0.8935861944063543,
      "train_speed(iter/s)": 0.032481
    },
    {
      "epoch": 0.547401056048055,
      "grad_norm": 0.10274602472782135,
      "learning_rate": 0.0001389548295426847,
      "loss": 0.3463117480278015,
      "memory(GiB)": 78.33,
      "step": 2825,
      "token_acc": 0.8964165043036476,
      "train_speed(iter/s)": 0.032482
    },
    {
      "epoch": 0.5475948263333817,
      "grad_norm": 0.10460248589515686,
      "learning_rate": 0.0001388589608210538,
      "loss": 0.35207241773605347,
      "memory(GiB)": 78.33,
      "step": 2826,
      "token_acc": 0.8935823860501049,
      "train_speed(iter/s)": 0.032483
    },
    {
      "epoch": 0.5477885966187085,
      "grad_norm": 0.11349425464868546,
      "learning_rate": 0.0001387630966753471,
      "loss": 0.377665638923645,
      "memory(GiB)": 78.33,
      "step": 2827,
      "token_acc": 0.8892741984547059,
      "train_speed(iter/s)": 0.032484
    },
    {
      "epoch": 0.5479823669040352,
      "grad_norm": 0.11117535084486008,
      "learning_rate": 0.0001386672371449385,
      "loss": 0.38577699661254883,
      "memory(GiB)": 78.33,
      "step": 2828,
      "token_acc": 0.887010551652719,
      "train_speed(iter/s)": 0.032485
    },
    {
      "epoch": 0.548176137189362,
      "grad_norm": 0.09501176327466965,
      "learning_rate": 0.0001385713822692001,
      "loss": 0.3158635199069977,
      "memory(GiB)": 78.33,
      "step": 2829,
      "token_acc": 0.9051193667815651,
      "train_speed(iter/s)": 0.032486
    },
    {
      "epoch": 0.5483699074746887,
      "grad_norm": 0.0978478193283081,
      "learning_rate": 0.00013847553208750222,
      "loss": 0.33727461099624634,
      "memory(GiB)": 78.33,
      "step": 2830,
      "token_acc": 0.9000374953130859,
      "train_speed(iter/s)": 0.032487
    },
    {
      "epoch": 0.5485636777600155,
      "grad_norm": 0.11235343664884567,
      "learning_rate": 0.0001383796866392129,
      "loss": 0.3979712128639221,
      "memory(GiB)": 78.33,
      "step": 2831,
      "token_acc": 0.8826383993927235,
      "train_speed(iter/s)": 0.032488
    },
    {
      "epoch": 0.5487574480453422,
      "grad_norm": 0.11587338894605637,
      "learning_rate": 0.0001382838459636986,
      "loss": 0.4109783470630646,
      "memory(GiB)": 78.33,
      "step": 2832,
      "token_acc": 0.8784734879286655,
      "train_speed(iter/s)": 0.032489
    },
    {
      "epoch": 0.548951218330669,
      "grad_norm": 0.09932563453912735,
      "learning_rate": 0.0001381880101003235,
      "loss": 0.3608021140098572,
      "memory(GiB)": 78.33,
      "step": 2833,
      "token_acc": 0.892873446614253,
      "train_speed(iter/s)": 0.03249
    },
    {
      "epoch": 0.5491449886159957,
      "grad_norm": 0.1027938574552536,
      "learning_rate": 0.00013809217908845008,
      "loss": 0.3469353914260864,
      "memory(GiB)": 78.33,
      "step": 2834,
      "token_acc": 0.8969847914789889,
      "train_speed(iter/s)": 0.032491
    },
    {
      "epoch": 0.5493387589013224,
      "grad_norm": 0.1079033836722374,
      "learning_rate": 0.00013799635296743868,
      "loss": 0.3682827055454254,
      "memory(GiB)": 78.33,
      "step": 2835,
      "token_acc": 0.8886550135749367,
      "train_speed(iter/s)": 0.032492
    },
    {
      "epoch": 0.5495325291866492,
      "grad_norm": 0.10848791897296906,
      "learning_rate": 0.00013790053177664766,
      "loss": 0.3711949288845062,
      "memory(GiB)": 78.33,
      "step": 2836,
      "token_acc": 0.8927511279981002,
      "train_speed(iter/s)": 0.032492
    },
    {
      "epoch": 0.5497262994719759,
      "grad_norm": 0.10043694078922272,
      "learning_rate": 0.00013780471555543343,
      "loss": 0.3552500009536743,
      "memory(GiB)": 78.33,
      "step": 2837,
      "token_acc": 0.8943481138318994,
      "train_speed(iter/s)": 0.032493
    },
    {
      "epoch": 0.5499200697573027,
      "grad_norm": 0.10964163392782211,
      "learning_rate": 0.00013770890434315012,
      "loss": 0.3622366786003113,
      "memory(GiB)": 78.33,
      "step": 2838,
      "token_acc": 0.8923598464684952,
      "train_speed(iter/s)": 0.032494
    },
    {
      "epoch": 0.5501138400426294,
      "grad_norm": 0.1045740619301796,
      "learning_rate": 0.00013761309817915014,
      "loss": 0.3492569923400879,
      "memory(GiB)": 78.33,
      "step": 2839,
      "token_acc": 0.8954375260019415,
      "train_speed(iter/s)": 0.032495
    },
    {
      "epoch": 0.5503076103279562,
      "grad_norm": 0.11920612305402756,
      "learning_rate": 0.00013751729710278354,
      "loss": 0.32998332381248474,
      "memory(GiB)": 78.33,
      "step": 2840,
      "token_acc": 0.9013660163219938,
      "train_speed(iter/s)": 0.032496
    },
    {
      "epoch": 0.5505013806132829,
      "grad_norm": 0.09879645705223083,
      "learning_rate": 0.00013742150115339852,
      "loss": 0.31799447536468506,
      "memory(GiB)": 78.33,
      "step": 2841,
      "token_acc": 0.9049382374886009,
      "train_speed(iter/s)": 0.032497
    },
    {
      "epoch": 0.5506951508986097,
      "grad_norm": 0.10987204313278198,
      "learning_rate": 0.0001373257103703409,
      "loss": 0.3859724998474121,
      "memory(GiB)": 78.33,
      "step": 2842,
      "token_acc": 0.885055264529223,
      "train_speed(iter/s)": 0.032498
    },
    {
      "epoch": 0.5508889211839364,
      "grad_norm": 0.1112445592880249,
      "learning_rate": 0.00013722992479295461,
      "loss": 0.3733959496021271,
      "memory(GiB)": 78.33,
      "step": 2843,
      "token_acc": 0.8924345643599857,
      "train_speed(iter/s)": 0.032499
    },
    {
      "epoch": 0.5510826914692631,
      "grad_norm": 0.10867438465356827,
      "learning_rate": 0.00013713414446058143,
      "loss": 0.3374147415161133,
      "memory(GiB)": 78.33,
      "step": 2844,
      "token_acc": 0.8994350282485876,
      "train_speed(iter/s)": 0.0325
    },
    {
      "epoch": 0.5512764617545899,
      "grad_norm": 0.10243318974971771,
      "learning_rate": 0.00013703836941256073,
      "loss": 0.36659640073776245,
      "memory(GiB)": 78.33,
      "step": 2845,
      "token_acc": 0.8927052150622341,
      "train_speed(iter/s)": 0.032501
    },
    {
      "epoch": 0.5514702320399166,
      "grad_norm": 0.10131695866584778,
      "learning_rate": 0.00013694259968823007,
      "loss": 0.3265146017074585,
      "memory(GiB)": 78.33,
      "step": 2846,
      "token_acc": 0.9022528879828551,
      "train_speed(iter/s)": 0.032501
    },
    {
      "epoch": 0.5516640023252434,
      "grad_norm": 0.1028154119849205,
      "learning_rate": 0.00013684683532692456,
      "loss": 0.34832775592803955,
      "memory(GiB)": 78.33,
      "step": 2847,
      "token_acc": 0.8964531768308241,
      "train_speed(iter/s)": 0.032502
    },
    {
      "epoch": 0.5518577726105701,
      "grad_norm": 0.10362169146537781,
      "learning_rate": 0.00013675107636797727,
      "loss": 0.35685330629348755,
      "memory(GiB)": 78.33,
      "step": 2848,
      "token_acc": 0.8953478775207021,
      "train_speed(iter/s)": 0.032503
    },
    {
      "epoch": 0.5520515428958969,
      "grad_norm": 0.10931023210287094,
      "learning_rate": 0.00013665532285071885,
      "loss": 0.3599531650543213,
      "memory(GiB)": 78.33,
      "step": 2849,
      "token_acc": 0.894084200338737,
      "train_speed(iter/s)": 0.032504
    },
    {
      "epoch": 0.5522453131812236,
      "grad_norm": 0.10313741117715836,
      "learning_rate": 0.00013655957481447796,
      "loss": 0.3336341977119446,
      "memory(GiB)": 78.33,
      "step": 2850,
      "token_acc": 0.8995257854179016,
      "train_speed(iter/s)": 0.032505
    },
    {
      "epoch": 0.5524390834665504,
      "grad_norm": 0.10181530565023422,
      "learning_rate": 0.00013646383229858088,
      "loss": 0.3731972575187683,
      "memory(GiB)": 78.33,
      "step": 2851,
      "token_acc": 0.8911425098754178,
      "train_speed(iter/s)": 0.032506
    },
    {
      "epoch": 0.5526328537518771,
      "grad_norm": 0.09568587690591812,
      "learning_rate": 0.00013636809534235155,
      "loss": 0.3068424463272095,
      "memory(GiB)": 78.33,
      "step": 2852,
      "token_acc": 0.9089695605517314,
      "train_speed(iter/s)": 0.032507
    },
    {
      "epoch": 0.5528266240372038,
      "grad_norm": 0.10480746626853943,
      "learning_rate": 0.00013627236398511183,
      "loss": 0.3546777069568634,
      "memory(GiB)": 78.33,
      "step": 2853,
      "token_acc": 0.8944366727644464,
      "train_speed(iter/s)": 0.032508
    },
    {
      "epoch": 0.5530203943225307,
      "grad_norm": 0.1180429756641388,
      "learning_rate": 0.00013617663826618102,
      "loss": 0.40899646282196045,
      "memory(GiB)": 78.33,
      "step": 2854,
      "token_acc": 0.8810652619256658,
      "train_speed(iter/s)": 0.032509
    },
    {
      "epoch": 0.5532141646078574,
      "grad_norm": 0.09670621156692505,
      "learning_rate": 0.0001360809182248764,
      "loss": 0.33177006244659424,
      "memory(GiB)": 78.33,
      "step": 2855,
      "token_acc": 0.9025397480900268,
      "train_speed(iter/s)": 0.03251
    },
    {
      "epoch": 0.5534079348931842,
      "grad_norm": 0.10695379227399826,
      "learning_rate": 0.00013598520390051264,
      "loss": 0.34610581398010254,
      "memory(GiB)": 78.33,
      "step": 2856,
      "token_acc": 0.898416321871879,
      "train_speed(iter/s)": 0.032511
    },
    {
      "epoch": 0.5536017051785109,
      "grad_norm": 0.10109010338783264,
      "learning_rate": 0.00013588949533240222,
      "loss": 0.3390766978263855,
      "memory(GiB)": 78.33,
      "step": 2857,
      "token_acc": 0.89864415955117,
      "train_speed(iter/s)": 0.032511
    },
    {
      "epoch": 0.5537954754638377,
      "grad_norm": 0.101463183760643,
      "learning_rate": 0.00013579379255985528,
      "loss": 0.3392132818698883,
      "memory(GiB)": 78.33,
      "step": 2858,
      "token_acc": 0.898758677643294,
      "train_speed(iter/s)": 0.032512
    },
    {
      "epoch": 0.5539892457491644,
      "grad_norm": 0.09950132668018341,
      "learning_rate": 0.00013569809562217943,
      "loss": 0.32112976908683777,
      "memory(GiB)": 78.33,
      "step": 2859,
      "token_acc": 0.902511359616902,
      "train_speed(iter/s)": 0.032513
    },
    {
      "epoch": 0.5541830160344912,
      "grad_norm": 0.10303032398223877,
      "learning_rate": 0.00013560240455868003,
      "loss": 0.3673378825187683,
      "memory(GiB)": 78.33,
      "step": 2860,
      "token_acc": 0.8928259417236241,
      "train_speed(iter/s)": 0.032514
    },
    {
      "epoch": 0.5543767863198179,
      "grad_norm": 0.11683917790651321,
      "learning_rate": 0.00013550671940865992,
      "loss": 0.3810235261917114,
      "memory(GiB)": 78.33,
      "step": 2861,
      "token_acc": 0.8879626045791527,
      "train_speed(iter/s)": 0.032515
    },
    {
      "epoch": 0.5545705566051446,
      "grad_norm": 0.11016968637704849,
      "learning_rate": 0.0001354110402114196,
      "loss": 0.3756512999534607,
      "memory(GiB)": 78.33,
      "step": 2862,
      "token_acc": 0.8894170776635569,
      "train_speed(iter/s)": 0.032516
    },
    {
      "epoch": 0.5547643268904714,
      "grad_norm": 0.09833884984254837,
      "learning_rate": 0.00013531536700625715,
      "loss": 0.3510299026966095,
      "memory(GiB)": 78.33,
      "step": 2863,
      "token_acc": 0.8975372743295384,
      "train_speed(iter/s)": 0.032517
    },
    {
      "epoch": 0.5549580971757981,
      "grad_norm": 0.09793855249881744,
      "learning_rate": 0.00013521969983246803,
      "loss": 0.36319395899772644,
      "memory(GiB)": 78.33,
      "step": 2864,
      "token_acc": 0.892717529189427,
      "train_speed(iter/s)": 0.032517
    },
    {
      "epoch": 0.5551518674611249,
      "grad_norm": 0.11692684143781662,
      "learning_rate": 0.0001351240387293454,
      "loss": 0.407484233379364,
      "memory(GiB)": 78.33,
      "step": 2865,
      "token_acc": 0.8815019139833371,
      "train_speed(iter/s)": 0.032518
    },
    {
      "epoch": 0.5553456377464516,
      "grad_norm": 0.10228422284126282,
      "learning_rate": 0.0001350283837361797,
      "loss": 0.340445339679718,
      "memory(GiB)": 78.33,
      "step": 2866,
      "token_acc": 0.8981877995519129,
      "train_speed(iter/s)": 0.032519
    },
    {
      "epoch": 0.5555394080317784,
      "grad_norm": 0.09636260569095612,
      "learning_rate": 0.00013493273489225915,
      "loss": 0.32771408557891846,
      "memory(GiB)": 78.33,
      "step": 2867,
      "token_acc": 0.9005267118133935,
      "train_speed(iter/s)": 0.03252
    },
    {
      "epoch": 0.5557331783171051,
      "grad_norm": 0.10914786905050278,
      "learning_rate": 0.00013483709223686922,
      "loss": 0.35096076130867004,
      "memory(GiB)": 78.33,
      "step": 2868,
      "token_acc": 0.8962214601291438,
      "train_speed(iter/s)": 0.032521
    },
    {
      "epoch": 0.5559269486024319,
      "grad_norm": 0.10578683018684387,
      "learning_rate": 0.00013474145580929297,
      "loss": 0.36036306619644165,
      "memory(GiB)": 78.33,
      "step": 2869,
      "token_acc": 0.8910474829632886,
      "train_speed(iter/s)": 0.032522
    },
    {
      "epoch": 0.5561207188877586,
      "grad_norm": 0.10106171667575836,
      "learning_rate": 0.00013464582564881087,
      "loss": 0.34388551115989685,
      "memory(GiB)": 78.33,
      "step": 2870,
      "token_acc": 0.8960813087667808,
      "train_speed(iter/s)": 0.032523
    },
    {
      "epoch": 0.5563144891730853,
      "grad_norm": 0.11406324058771133,
      "learning_rate": 0.00013455020179470073,
      "loss": 0.3458818197250366,
      "memory(GiB)": 78.33,
      "step": 2871,
      "token_acc": 0.8983081299595748,
      "train_speed(iter/s)": 0.032524
    },
    {
      "epoch": 0.5565082594584121,
      "grad_norm": 0.09773491322994232,
      "learning_rate": 0.00013445458428623788,
      "loss": 0.33812999725341797,
      "memory(GiB)": 78.33,
      "step": 2872,
      "token_acc": 0.899073884568154,
      "train_speed(iter/s)": 0.032525
    },
    {
      "epoch": 0.5567020297437388,
      "grad_norm": 0.10276000201702118,
      "learning_rate": 0.000134358973162695,
      "loss": 0.3428000807762146,
      "memory(GiB)": 78.33,
      "step": 2873,
      "token_acc": 0.8970639124763755,
      "train_speed(iter/s)": 0.032525
    },
    {
      "epoch": 0.5568958000290656,
      "grad_norm": 0.10354767739772797,
      "learning_rate": 0.00013426336846334208,
      "loss": 0.359587162733078,
      "memory(GiB)": 78.33,
      "step": 2874,
      "token_acc": 0.8922890471140736,
      "train_speed(iter/s)": 0.032526
    },
    {
      "epoch": 0.5570895703143923,
      "grad_norm": 0.12495241314172745,
      "learning_rate": 0.0001341677702274466,
      "loss": 0.3915575444698334,
      "memory(GiB)": 78.33,
      "step": 2875,
      "token_acc": 0.8831632312720966,
      "train_speed(iter/s)": 0.032527
    },
    {
      "epoch": 0.5572833405997191,
      "grad_norm": 0.09519120305776596,
      "learning_rate": 0.00013407217849427332,
      "loss": 0.3396569788455963,
      "memory(GiB)": 78.33,
      "step": 2876,
      "token_acc": 0.8975312241823395,
      "train_speed(iter/s)": 0.032528
    },
    {
      "epoch": 0.5574771108850458,
      "grad_norm": 0.10415952652692795,
      "learning_rate": 0.0001339765933030844,
      "loss": 0.36318159103393555,
      "memory(GiB)": 78.33,
      "step": 2877,
      "token_acc": 0.8913055970578094,
      "train_speed(iter/s)": 0.032529
    },
    {
      "epoch": 0.5576708811703726,
      "grad_norm": 0.09890494495630264,
      "learning_rate": 0.00013388101469313907,
      "loss": 0.34472253918647766,
      "memory(GiB)": 78.33,
      "step": 2878,
      "token_acc": 0.8975010936132983,
      "train_speed(iter/s)": 0.03253
    },
    {
      "epoch": 0.5578646514556993,
      "grad_norm": 0.10054649412631989,
      "learning_rate": 0.0001337854427036942,
      "loss": 0.3560715615749359,
      "memory(GiB)": 78.33,
      "step": 2879,
      "token_acc": 0.8977050131798505,
      "train_speed(iter/s)": 0.032531
    },
    {
      "epoch": 0.558058421741026,
      "grad_norm": 0.10295595228672028,
      "learning_rate": 0.00013368987737400368,
      "loss": 0.3403087556362152,
      "memory(GiB)": 78.33,
      "step": 2880,
      "token_acc": 0.8993411629905471,
      "train_speed(iter/s)": 0.032532
    },
    {
      "epoch": 0.5582521920263528,
      "grad_norm": 0.11277662217617035,
      "learning_rate": 0.00013359431874331886,
      "loss": 0.38854244351387024,
      "memory(GiB)": 78.33,
      "step": 2881,
      "token_acc": 0.8866336122301415,
      "train_speed(iter/s)": 0.032533
    },
    {
      "epoch": 0.5584459623116795,
      "grad_norm": 0.09544237703084946,
      "learning_rate": 0.0001334987668508881,
      "loss": 0.35088953375816345,
      "memory(GiB)": 78.33,
      "step": 2882,
      "token_acc": 0.8954308027790456,
      "train_speed(iter/s)": 0.032533
    },
    {
      "epoch": 0.5586397325970063,
      "grad_norm": 0.09790968149900436,
      "learning_rate": 0.0001334032217359572,
      "loss": 0.3613382875919342,
      "memory(GiB)": 78.33,
      "step": 2883,
      "token_acc": 0.8920747907163267,
      "train_speed(iter/s)": 0.032534
    },
    {
      "epoch": 0.558833502882333,
      "grad_norm": 0.10290265083312988,
      "learning_rate": 0.00013330768343776918,
      "loss": 0.35918739438056946,
      "memory(GiB)": 78.33,
      "step": 2884,
      "token_acc": 0.8945180422371265,
      "train_speed(iter/s)": 0.032535
    },
    {
      "epoch": 0.5590272731676598,
      "grad_norm": 0.11126357316970825,
      "learning_rate": 0.00013321215199556404,
      "loss": 0.34764158725738525,
      "memory(GiB)": 78.33,
      "step": 2885,
      "token_acc": 0.8981644381890688,
      "train_speed(iter/s)": 0.032536
    },
    {
      "epoch": 0.5592210434529865,
      "grad_norm": 0.10342656821012497,
      "learning_rate": 0.0001331166274485792,
      "loss": 0.3714340329170227,
      "memory(GiB)": 78.33,
      "step": 2886,
      "token_acc": 0.8889918774544479,
      "train_speed(iter/s)": 0.032537
    },
    {
      "epoch": 0.5594148137383133,
      "grad_norm": 0.09674359858036041,
      "learning_rate": 0.00013302110983604912,
      "loss": 0.3426961302757263,
      "memory(GiB)": 78.33,
      "step": 2887,
      "token_acc": 0.8992525579932754,
      "train_speed(iter/s)": 0.032538
    },
    {
      "epoch": 0.55960858402364,
      "grad_norm": 0.1152225211262703,
      "learning_rate": 0.00013292559919720554,
      "loss": 0.3851836621761322,
      "memory(GiB)": 78.33,
      "step": 2888,
      "token_acc": 0.8864522728725784,
      "train_speed(iter/s)": 0.032539
    },
    {
      "epoch": 0.5598023543089667,
      "grad_norm": 0.2307002991437912,
      "learning_rate": 0.00013283009557127712,
      "loss": 0.3549976050853729,
      "memory(GiB)": 78.33,
      "step": 2889,
      "token_acc": 0.8931888544891641,
      "train_speed(iter/s)": 0.03254
    },
    {
      "epoch": 0.5599961245942935,
      "grad_norm": 0.10933719575405121,
      "learning_rate": 0.0001327345989974898,
      "loss": 0.40044257044792175,
      "memory(GiB)": 78.33,
      "step": 2890,
      "token_acc": 0.8821779976652399,
      "train_speed(iter/s)": 0.03254
    },
    {
      "epoch": 0.5601898948796202,
      "grad_norm": 0.08904779702425003,
      "learning_rate": 0.00013263910951506668,
      "loss": 0.31354546546936035,
      "memory(GiB)": 78.33,
      "step": 2891,
      "token_acc": 0.9041093058828393,
      "train_speed(iter/s)": 0.032541
    },
    {
      "epoch": 0.560383665164947,
      "grad_norm": 0.10197526961565018,
      "learning_rate": 0.00013254362716322776,
      "loss": 0.3503738045692444,
      "memory(GiB)": 78.33,
      "step": 2892,
      "token_acc": 0.8949555782295996,
      "train_speed(iter/s)": 0.032542
    },
    {
      "epoch": 0.5605774354502737,
      "grad_norm": 0.11662586033344269,
      "learning_rate": 0.00013244815198119024,
      "loss": 0.3930249810218811,
      "memory(GiB)": 78.33,
      "step": 2893,
      "token_acc": 0.8859180035650623,
      "train_speed(iter/s)": 0.032543
    },
    {
      "epoch": 0.5607712057356005,
      "grad_norm": 0.11080929636955261,
      "learning_rate": 0.0001323526840081683,
      "loss": 0.3820610046386719,
      "memory(GiB)": 78.33,
      "step": 2894,
      "token_acc": 0.8872899535216303,
      "train_speed(iter/s)": 0.032544
    },
    {
      "epoch": 0.5609649760209272,
      "grad_norm": 0.10336665064096451,
      "learning_rate": 0.00013225722328337323,
      "loss": 0.34890395402908325,
      "memory(GiB)": 78.33,
      "step": 2895,
      "token_acc": 0.8951938666153719,
      "train_speed(iter/s)": 0.032545
    },
    {
      "epoch": 0.561158746306254,
      "grad_norm": 0.09736684709787369,
      "learning_rate": 0.0001321617698460134,
      "loss": 0.33721378445625305,
      "memory(GiB)": 78.33,
      "step": 2896,
      "token_acc": 0.9000335025642348,
      "train_speed(iter/s)": 0.032546
    },
    {
      "epoch": 0.5613525165915807,
      "grad_norm": 0.10685203224420547,
      "learning_rate": 0.00013206632373529396,
      "loss": 0.35752072930336,
      "memory(GiB)": 78.33,
      "step": 2897,
      "token_acc": 0.8925185941373779,
      "train_speed(iter/s)": 0.032546
    },
    {
      "epoch": 0.5615462868769074,
      "grad_norm": 0.10666251182556152,
      "learning_rate": 0.00013197088499041732,
      "loss": 0.3803097903728485,
      "memory(GiB)": 78.33,
      "step": 2898,
      "token_acc": 0.8869661896345288,
      "train_speed(iter/s)": 0.032547
    },
    {
      "epoch": 0.5617400571622342,
      "grad_norm": 0.09380005300045013,
      "learning_rate": 0.00013187545365058261,
      "loss": 0.3471141755580902,
      "memory(GiB)": 78.33,
      "step": 2899,
      "token_acc": 0.8968915295993044,
      "train_speed(iter/s)": 0.032548
    },
    {
      "epoch": 0.5619338274475609,
      "grad_norm": 0.10400458425283432,
      "learning_rate": 0.00013178002975498614,
      "loss": 0.3575308322906494,
      "memory(GiB)": 78.33,
      "step": 2900,
      "token_acc": 0.8943859748659613,
      "train_speed(iter/s)": 0.032549
    },
    {
      "epoch": 0.5621275977328877,
      "grad_norm": 0.09898320585489273,
      "learning_rate": 0.00013168461334282103,
      "loss": 0.34250903129577637,
      "memory(GiB)": 78.33,
      "step": 2901,
      "token_acc": 0.895111494593793,
      "train_speed(iter/s)": 0.03255
    },
    {
      "epoch": 0.5623213680182144,
      "grad_norm": 0.09259682148694992,
      "learning_rate": 0.00013158920445327738,
      "loss": 0.30667221546173096,
      "memory(GiB)": 78.33,
      "step": 2902,
      "token_acc": 0.9075316927665921,
      "train_speed(iter/s)": 0.032551
    },
    {
      "epoch": 0.5625151383035412,
      "grad_norm": 0.10204022377729416,
      "learning_rate": 0.0001314938031255422,
      "loss": 0.35288378596305847,
      "memory(GiB)": 78.33,
      "step": 2903,
      "token_acc": 0.8955367449018854,
      "train_speed(iter/s)": 0.032552
    },
    {
      "epoch": 0.5627089085888679,
      "grad_norm": 0.0911250039935112,
      "learning_rate": 0.00013139840939879933,
      "loss": 0.3128979206085205,
      "memory(GiB)": 78.33,
      "step": 2904,
      "token_acc": 0.9082222013523666,
      "train_speed(iter/s)": 0.032553
    },
    {
      "epoch": 0.5629026788741947,
      "grad_norm": 0.10293769836425781,
      "learning_rate": 0.00013130302331222963,
      "loss": 0.34590229392051697,
      "memory(GiB)": 78.33,
      "step": 2905,
      "token_acc": 0.8969505783385909,
      "train_speed(iter/s)": 0.032553
    },
    {
      "epoch": 0.5630964491595214,
      "grad_norm": 0.09894620627164841,
      "learning_rate": 0.00013120764490501057,
      "loss": 0.33227023482322693,
      "memory(GiB)": 78.33,
      "step": 2906,
      "token_acc": 0.9007358424899319,
      "train_speed(iter/s)": 0.032554
    },
    {
      "epoch": 0.5632902194448481,
      "grad_norm": 0.10376887768507004,
      "learning_rate": 0.00013111227421631674,
      "loss": 0.36455827951431274,
      "memory(GiB)": 78.33,
      "step": 2907,
      "token_acc": 0.8945593638331069,
      "train_speed(iter/s)": 0.032555
    },
    {
      "epoch": 0.5634839897301749,
      "grad_norm": 0.10269268602132797,
      "learning_rate": 0.00013101691128531942,
      "loss": 0.3566417694091797,
      "memory(GiB)": 78.33,
      "step": 2908,
      "token_acc": 0.8924794993526112,
      "train_speed(iter/s)": 0.032556
    },
    {
      "epoch": 0.5636777600155016,
      "grad_norm": 0.11318562924861908,
      "learning_rate": 0.00013092155615118672,
      "loss": 0.3609810471534729,
      "memory(GiB)": 78.33,
      "step": 2909,
      "token_acc": 0.8935661992803078,
      "train_speed(iter/s)": 0.032557
    },
    {
      "epoch": 0.5638715303008284,
      "grad_norm": 0.09679584205150604,
      "learning_rate": 0.00013082620885308363,
      "loss": 0.33762603998184204,
      "memory(GiB)": 78.33,
      "step": 2910,
      "token_acc": 0.8995445957075723,
      "train_speed(iter/s)": 0.032558
    },
    {
      "epoch": 0.5640653005861551,
      "grad_norm": 0.11616901308298111,
      "learning_rate": 0.00013073086943017173,
      "loss": 0.351492315530777,
      "memory(GiB)": 78.33,
      "step": 2911,
      "token_acc": 0.8948040342142218,
      "train_speed(iter/s)": 0.032559
    },
    {
      "epoch": 0.5642590708714819,
      "grad_norm": 0.10330884903669357,
      "learning_rate": 0.00013063553792160958,
      "loss": 0.3321197032928467,
      "memory(GiB)": 78.33,
      "step": 2912,
      "token_acc": 0.8990494397535977,
      "train_speed(iter/s)": 0.03256
    },
    {
      "epoch": 0.5644528411568086,
      "grad_norm": 0.10223756730556488,
      "learning_rate": 0.0001305402143665523,
      "loss": 0.32255592942237854,
      "memory(GiB)": 78.33,
      "step": 2913,
      "token_acc": 0.9044650149741356,
      "train_speed(iter/s)": 0.03256
    },
    {
      "epoch": 0.5646466114421353,
      "grad_norm": 0.09594756364822388,
      "learning_rate": 0.00013044489880415194,
      "loss": 0.3452211320400238,
      "memory(GiB)": 78.33,
      "step": 2914,
      "token_acc": 0.8991926235148401,
      "train_speed(iter/s)": 0.032561
    },
    {
      "epoch": 0.5648403817274621,
      "grad_norm": 0.09111212939023972,
      "learning_rate": 0.00013034959127355703,
      "loss": 0.3073510229587555,
      "memory(GiB)": 78.33,
      "step": 2915,
      "token_acc": 0.9079986434765758,
      "train_speed(iter/s)": 0.032562
    },
    {
      "epoch": 0.5650341520127888,
      "grad_norm": 0.10147152841091156,
      "learning_rate": 0.00013025429181391304,
      "loss": 0.3540181815624237,
      "memory(GiB)": 78.33,
      "step": 2916,
      "token_acc": 0.8969390771546049,
      "train_speed(iter/s)": 0.032563
    },
    {
      "epoch": 0.5652279222981156,
      "grad_norm": 0.10001610219478607,
      "learning_rate": 0.00013015900046436205,
      "loss": 0.35491663217544556,
      "memory(GiB)": 78.33,
      "step": 2917,
      "token_acc": 0.8945139415897574,
      "train_speed(iter/s)": 0.032564
    },
    {
      "epoch": 0.5654216925834423,
      "grad_norm": 0.09321217238903046,
      "learning_rate": 0.00013006371726404265,
      "loss": 0.3310143053531647,
      "memory(GiB)": 78.33,
      "step": 2918,
      "token_acc": 0.9035609732687047,
      "train_speed(iter/s)": 0.032564
    },
    {
      "epoch": 0.5656154628687691,
      "grad_norm": 0.10879889130592346,
      "learning_rate": 0.0001299684422520903,
      "loss": 0.3868524730205536,
      "memory(GiB)": 78.33,
      "step": 2919,
      "token_acc": 0.8842917053674463,
      "train_speed(iter/s)": 0.032565
    },
    {
      "epoch": 0.5658092331540958,
      "grad_norm": 0.09358043968677521,
      "learning_rate": 0.00012987317546763697,
      "loss": 0.3301950991153717,
      "memory(GiB)": 78.33,
      "step": 2920,
      "token_acc": 0.9018064033885635,
      "train_speed(iter/s)": 0.032566
    },
    {
      "epoch": 0.5660030034394226,
      "grad_norm": 0.09514044970273972,
      "learning_rate": 0.00012977791694981136,
      "loss": 0.34315773844718933,
      "memory(GiB)": 78.33,
      "step": 2921,
      "token_acc": 0.8958858102434929,
      "train_speed(iter/s)": 0.032567
    },
    {
      "epoch": 0.5661967737247493,
      "grad_norm": 0.10747494548559189,
      "learning_rate": 0.00012968266673773858,
      "loss": 0.3809034824371338,
      "memory(GiB)": 78.33,
      "step": 2922,
      "token_acc": 0.8869167528719987,
      "train_speed(iter/s)": 0.032568
    },
    {
      "epoch": 0.566390544010076,
      "grad_norm": 0.09282265603542328,
      "learning_rate": 0.00012958742487054054,
      "loss": 0.3388338088989258,
      "memory(GiB)": 78.33,
      "step": 2923,
      "token_acc": 0.9004572462527813,
      "train_speed(iter/s)": 0.032569
    },
    {
      "epoch": 0.5665843142954028,
      "grad_norm": 0.11763904243707657,
      "learning_rate": 0.00012949219138733565,
      "loss": 0.403276652097702,
      "memory(GiB)": 78.33,
      "step": 2924,
      "token_acc": 0.8811169562025936,
      "train_speed(iter/s)": 0.03257
    },
    {
      "epoch": 0.5667780845807295,
      "grad_norm": 0.09826403111219406,
      "learning_rate": 0.00012939696632723876,
      "loss": 0.34657537937164307,
      "memory(GiB)": 78.33,
      "step": 2925,
      "token_acc": 0.8960861857357153,
      "train_speed(iter/s)": 0.03257
    },
    {
      "epoch": 0.5669718548660563,
      "grad_norm": 0.09996719658374786,
      "learning_rate": 0.00012930174972936148,
      "loss": 0.34664636850357056,
      "memory(GiB)": 78.33,
      "step": 2926,
      "token_acc": 0.8960221082560519,
      "train_speed(iter/s)": 0.032571
    },
    {
      "epoch": 0.567165625151383,
      "grad_norm": 0.09884995222091675,
      "learning_rate": 0.00012920654163281172,
      "loss": 0.3174511790275574,
      "memory(GiB)": 78.33,
      "step": 2927,
      "token_acc": 0.9049272486772487,
      "train_speed(iter/s)": 0.032572
    },
    {
      "epoch": 0.5673593954367098,
      "grad_norm": 0.10128315538167953,
      "learning_rate": 0.00012911134207669412,
      "loss": 0.33141183853149414,
      "memory(GiB)": 78.33,
      "step": 2928,
      "token_acc": 0.9011783988470342,
      "train_speed(iter/s)": 0.032573
    },
    {
      "epoch": 0.5675531657220365,
      "grad_norm": 0.11448942869901657,
      "learning_rate": 0.00012901615110010956,
      "loss": 0.3866661489009857,
      "memory(GiB)": 78.33,
      "step": 2929,
      "token_acc": 0.8871186120469788,
      "train_speed(iter/s)": 0.032574
    },
    {
      "epoch": 0.5677469360073633,
      "grad_norm": 0.09554385393857956,
      "learning_rate": 0.00012892096874215562,
      "loss": 0.31373119354248047,
      "memory(GiB)": 78.33,
      "step": 2930,
      "token_acc": 0.9060102797220859,
      "train_speed(iter/s)": 0.032575
    },
    {
      "epoch": 0.56794070629269,
      "grad_norm": 0.10202641785144806,
      "learning_rate": 0.00012882579504192628,
      "loss": 0.3446533977985382,
      "memory(GiB)": 78.33,
      "step": 2931,
      "token_acc": 0.8982907151647751,
      "train_speed(iter/s)": 0.032576
    },
    {
      "epoch": 0.5681344765780167,
      "grad_norm": 0.10661419481039047,
      "learning_rate": 0.00012873063003851184,
      "loss": 0.35394954681396484,
      "memory(GiB)": 78.33,
      "step": 2932,
      "token_acc": 0.8950952106174265,
      "train_speed(iter/s)": 0.032577
    },
    {
      "epoch": 0.5683282468633435,
      "grad_norm": 0.10493983328342438,
      "learning_rate": 0.00012863547377099918,
      "loss": 0.34970593452453613,
      "memory(GiB)": 78.33,
      "step": 2933,
      "token_acc": 0.8950512907225792,
      "train_speed(iter/s)": 0.032578
    },
    {
      "epoch": 0.5685220171486702,
      "grad_norm": 0.11756689846515656,
      "learning_rate": 0.0001285403262784715,
      "loss": 0.3975431025028229,
      "memory(GiB)": 78.33,
      "step": 2934,
      "token_acc": 0.8840085287846482,
      "train_speed(iter/s)": 0.032578
    },
    {
      "epoch": 0.568715787433997,
      "grad_norm": 0.10590403527021408,
      "learning_rate": 0.00012844518760000848,
      "loss": 0.37834692001342773,
      "memory(GiB)": 78.33,
      "step": 2935,
      "token_acc": 0.8883588043448963,
      "train_speed(iter/s)": 0.032579
    },
    {
      "epoch": 0.5689095577193237,
      "grad_norm": 0.0980941578745842,
      "learning_rate": 0.0001283500577746862,
      "loss": 0.32845622301101685,
      "memory(GiB)": 78.33,
      "step": 2936,
      "token_acc": 0.9010894350139347,
      "train_speed(iter/s)": 0.03258
    },
    {
      "epoch": 0.5691033280046505,
      "grad_norm": 0.09987200796604156,
      "learning_rate": 0.00012825493684157682,
      "loss": 0.334673672914505,
      "memory(GiB)": 78.33,
      "step": 2937,
      "token_acc": 0.8996542412235318,
      "train_speed(iter/s)": 0.032581
    },
    {
      "epoch": 0.5692970982899772,
      "grad_norm": 0.1141529381275177,
      "learning_rate": 0.0001281598248397493,
      "loss": 0.39329400658607483,
      "memory(GiB)": 78.33,
      "step": 2938,
      "token_acc": 0.886091163126108,
      "train_speed(iter/s)": 0.032582
    },
    {
      "epoch": 0.569490868575304,
      "grad_norm": 0.09881948679685593,
      "learning_rate": 0.0001280647218082685,
      "loss": 0.3378741145133972,
      "memory(GiB)": 78.33,
      "step": 2939,
      "token_acc": 0.8991566664118831,
      "train_speed(iter/s)": 0.032583
    },
    {
      "epoch": 0.5696846388606307,
      "grad_norm": 0.10220920294523239,
      "learning_rate": 0.00012796962778619593,
      "loss": 0.35935983061790466,
      "memory(GiB)": 78.33,
      "step": 2940,
      "token_acc": 0.8916394328504156,
      "train_speed(iter/s)": 0.032583
    },
    {
      "epoch": 0.5698784091459574,
      "grad_norm": 0.10239671915769577,
      "learning_rate": 0.00012787454281258916,
      "loss": 0.34669652581214905,
      "memory(GiB)": 78.33,
      "step": 2941,
      "token_acc": 0.8974839774985307,
      "train_speed(iter/s)": 0.032584
    },
    {
      "epoch": 0.5700721794312842,
      "grad_norm": 0.10328203439712524,
      "learning_rate": 0.0001277794669265022,
      "loss": 0.33819228410720825,
      "memory(GiB)": 78.33,
      "step": 2942,
      "token_acc": 0.8988887064104669,
      "train_speed(iter/s)": 0.032585
    },
    {
      "epoch": 0.5702659497166109,
      "grad_norm": 0.10673517733812332,
      "learning_rate": 0.00012768440016698533,
      "loss": 0.36250045895576477,
      "memory(GiB)": 78.33,
      "step": 2943,
      "token_acc": 0.8937628879783616,
      "train_speed(iter/s)": 0.032586
    },
    {
      "epoch": 0.5704597200019377,
      "grad_norm": 0.10135802626609802,
      "learning_rate": 0.0001275893425730849,
      "loss": 0.3382996618747711,
      "memory(GiB)": 78.33,
      "step": 2944,
      "token_acc": 0.8998493193390834,
      "train_speed(iter/s)": 0.032587
    },
    {
      "epoch": 0.5706534902872644,
      "grad_norm": 0.09103412926197052,
      "learning_rate": 0.00012749429418384368,
      "loss": 0.3121355175971985,
      "memory(GiB)": 78.33,
      "step": 2945,
      "token_acc": 0.907722643769968,
      "train_speed(iter/s)": 0.032588
    },
    {
      "epoch": 0.5708472605725912,
      "grad_norm": 0.10437119007110596,
      "learning_rate": 0.00012739925503830058,
      "loss": 0.3552001118659973,
      "memory(GiB)": 78.33,
      "step": 2946,
      "token_acc": 0.8937973264809601,
      "train_speed(iter/s)": 0.032589
    },
    {
      "epoch": 0.5710410308579179,
      "grad_norm": 0.09624162316322327,
      "learning_rate": 0.00012730422517549076,
      "loss": 0.34696635603904724,
      "memory(GiB)": 78.33,
      "step": 2947,
      "token_acc": 0.89641196508939,
      "train_speed(iter/s)": 0.032589
    },
    {
      "epoch": 0.5712348011432447,
      "grad_norm": 0.0997808501124382,
      "learning_rate": 0.0001272092046344455,
      "loss": 0.33023878931999207,
      "memory(GiB)": 78.33,
      "step": 2948,
      "token_acc": 0.9004624871531346,
      "train_speed(iter/s)": 0.03259
    },
    {
      "epoch": 0.5714285714285714,
      "grad_norm": 0.10305418819189072,
      "learning_rate": 0.0001271141934541923,
      "loss": 0.3459644019603729,
      "memory(GiB)": 78.33,
      "step": 2949,
      "token_acc": 0.8978806469604016,
      "train_speed(iter/s)": 0.032591
    },
    {
      "epoch": 0.5716223417138981,
      "grad_norm": 0.09685331583023071,
      "learning_rate": 0.00012701919167375488,
      "loss": 0.34739017486572266,
      "memory(GiB)": 78.33,
      "step": 2950,
      "token_acc": 0.8971735019041944,
      "train_speed(iter/s)": 0.032592
    },
    {
      "epoch": 0.5718161119992249,
      "grad_norm": 0.09449600428342819,
      "learning_rate": 0.00012692419933215288,
      "loss": 0.3246384859085083,
      "memory(GiB)": 78.33,
      "step": 2951,
      "token_acc": 0.9017080904162831,
      "train_speed(iter/s)": 0.032593
    },
    {
      "epoch": 0.5720098822845516,
      "grad_norm": 0.10447492450475693,
      "learning_rate": 0.00012682921646840233,
      "loss": 0.37390848994255066,
      "memory(GiB)": 78.33,
      "step": 2952,
      "token_acc": 0.8906833284795214,
      "train_speed(iter/s)": 0.032594
    },
    {
      "epoch": 0.5722036525698784,
      "grad_norm": 0.11614307761192322,
      "learning_rate": 0.00012673424312151517,
      "loss": 0.40665432810783386,
      "memory(GiB)": 78.33,
      "step": 2953,
      "token_acc": 0.8813434035031336,
      "train_speed(iter/s)": 0.032594
    },
    {
      "epoch": 0.5723974228552051,
      "grad_norm": 0.09071079641580582,
      "learning_rate": 0.0001266392793304996,
      "loss": 0.33611175417900085,
      "memory(GiB)": 78.33,
      "step": 2954,
      "token_acc": 0.8994031782555547,
      "train_speed(iter/s)": 0.032595
    },
    {
      "epoch": 0.5725911931405319,
      "grad_norm": 0.10135752707719803,
      "learning_rate": 0.00012654432513435965,
      "loss": 0.36446985602378845,
      "memory(GiB)": 78.33,
      "step": 2955,
      "token_acc": 0.8912203779100115,
      "train_speed(iter/s)": 0.032596
    },
    {
      "epoch": 0.5727849634258586,
      "grad_norm": 0.10277879238128662,
      "learning_rate": 0.00012644938057209567,
      "loss": 0.3852466940879822,
      "memory(GiB)": 78.33,
      "step": 2956,
      "token_acc": 0.8869683281547821,
      "train_speed(iter/s)": 0.032597
    },
    {
      "epoch": 0.5729787337111854,
      "grad_norm": 0.10757278650999069,
      "learning_rate": 0.00012635444568270398,
      "loss": 0.35662007331848145,
      "memory(GiB)": 78.33,
      "step": 2957,
      "token_acc": 0.893989887382211,
      "train_speed(iter/s)": 0.032598
    },
    {
      "epoch": 0.5731725039965121,
      "grad_norm": 0.11290912330150604,
      "learning_rate": 0.00012625952050517673,
      "loss": 0.365123450756073,
      "memory(GiB)": 78.33,
      "step": 2958,
      "token_acc": 0.890861820997898,
      "train_speed(iter/s)": 0.032599
    },
    {
      "epoch": 0.5733662742818388,
      "grad_norm": 0.11363532394170761,
      "learning_rate": 0.00012616460507850242,
      "loss": 0.3688386082649231,
      "memory(GiB)": 78.33,
      "step": 2959,
      "token_acc": 0.8900687757909216,
      "train_speed(iter/s)": 0.032599
    },
    {
      "epoch": 0.5735600445671656,
      "grad_norm": 0.10175785422325134,
      "learning_rate": 0.00012606969944166523,
      "loss": 0.3480740785598755,
      "memory(GiB)": 78.33,
      "step": 2960,
      "token_acc": 0.8961689531507941,
      "train_speed(iter/s)": 0.0326
    },
    {
      "epoch": 0.5737538148524923,
      "grad_norm": 0.10226535052061081,
      "learning_rate": 0.00012597480363364558,
      "loss": 0.36537787318229675,
      "memory(GiB)": 78.33,
      "step": 2961,
      "token_acc": 0.8930911178973194,
      "train_speed(iter/s)": 0.032601
    },
    {
      "epoch": 0.5739475851378191,
      "grad_norm": 0.10284057259559631,
      "learning_rate": 0.0001258799176934196,
      "loss": 0.3630596399307251,
      "memory(GiB)": 78.33,
      "step": 2962,
      "token_acc": 0.8929182958930716,
      "train_speed(iter/s)": 0.032602
    },
    {
      "epoch": 0.5741413554231458,
      "grad_norm": 0.10394224524497986,
      "learning_rate": 0.00012578504165995953,
      "loss": 0.3422529995441437,
      "memory(GiB)": 78.33,
      "step": 2963,
      "token_acc": 0.8979251265505924,
      "train_speed(iter/s)": 0.032603
    },
    {
      "epoch": 0.5743351257084726,
      "grad_norm": 0.10181540995836258,
      "learning_rate": 0.00012569017557223362,
      "loss": 0.35082393884658813,
      "memory(GiB)": 78.33,
      "step": 2964,
      "token_acc": 0.8960905890389365,
      "train_speed(iter/s)": 0.032604
    },
    {
      "epoch": 0.5745288959937993,
      "grad_norm": 0.10336862504482269,
      "learning_rate": 0.00012559531946920578,
      "loss": 0.36020544171333313,
      "memory(GiB)": 78.33,
      "step": 2965,
      "token_acc": 0.8921231485743684,
      "train_speed(iter/s)": 0.032604
    },
    {
      "epoch": 0.574722666279126,
      "grad_norm": 0.10156133770942688,
      "learning_rate": 0.00012550047338983603,
      "loss": 0.36286041140556335,
      "memory(GiB)": 78.33,
      "step": 2966,
      "token_acc": 0.8922610015174507,
      "train_speed(iter/s)": 0.032605
    },
    {
      "epoch": 0.5749164365644528,
      "grad_norm": 0.09440826624631882,
      "learning_rate": 0.00012540563737308016,
      "loss": 0.3350200057029724,
      "memory(GiB)": 78.33,
      "step": 2967,
      "token_acc": 0.8994114241605558,
      "train_speed(iter/s)": 0.032606
    },
    {
      "epoch": 0.5751102068497795,
      "grad_norm": 0.117142453789711,
      "learning_rate": 0.00012531081145788987,
      "loss": 0.4059712290763855,
      "memory(GiB)": 78.33,
      "step": 2968,
      "token_acc": 0.8823188237682337,
      "train_speed(iter/s)": 0.032607
    },
    {
      "epoch": 0.5753039771351063,
      "grad_norm": 0.10292242467403412,
      "learning_rate": 0.00012521599568321283,
      "loss": 0.3511313199996948,
      "memory(GiB)": 78.33,
      "step": 2969,
      "token_acc": 0.8962899832210296,
      "train_speed(iter/s)": 0.032608
    },
    {
      "epoch": 0.575497747420433,
      "grad_norm": 0.10700348764657974,
      "learning_rate": 0.00012512119008799226,
      "loss": 0.3738468885421753,
      "memory(GiB)": 78.33,
      "step": 2970,
      "token_acc": 0.8902907594626264,
      "train_speed(iter/s)": 0.032609
    },
    {
      "epoch": 0.5756915177057598,
      "grad_norm": 0.10004039853811264,
      "learning_rate": 0.0001250263947111675,
      "loss": 0.3366740942001343,
      "memory(GiB)": 78.33,
      "step": 2971,
      "token_acc": 0.9001073537305422,
      "train_speed(iter/s)": 0.032609
    },
    {
      "epoch": 0.5758852879910865,
      "grad_norm": 0.10136931389570236,
      "learning_rate": 0.00012493160959167347,
      "loss": 0.35105177760124207,
      "memory(GiB)": 78.33,
      "step": 2972,
      "token_acc": 0.8961308889362609,
      "train_speed(iter/s)": 0.03261
    },
    {
      "epoch": 0.5760790582764133,
      "grad_norm": 0.10934333503246307,
      "learning_rate": 0.000124836834768441,
      "loss": 0.37874335050582886,
      "memory(GiB)": 78.33,
      "step": 2973,
      "token_acc": 0.8891520475888849,
      "train_speed(iter/s)": 0.032611
    },
    {
      "epoch": 0.57627282856174,
      "grad_norm": 0.09537496417760849,
      "learning_rate": 0.0001247420702803966,
      "loss": 0.3463536500930786,
      "memory(GiB)": 78.33,
      "step": 2974,
      "token_acc": 0.8957062728096603,
      "train_speed(iter/s)": 0.032612
    },
    {
      "epoch": 0.5764665988470667,
      "grad_norm": 0.11320365220308304,
      "learning_rate": 0.00012464731616646267,
      "loss": 0.3580982983112335,
      "memory(GiB)": 78.33,
      "step": 2975,
      "token_acc": 0.8927235580700927,
      "train_speed(iter/s)": 0.032613
    },
    {
      "epoch": 0.5766603691323936,
      "grad_norm": 0.1089281439781189,
      "learning_rate": 0.0001245525724655573,
      "loss": 0.3534199297428131,
      "memory(GiB)": 78.33,
      "step": 2976,
      "token_acc": 0.8954800657857427,
      "train_speed(iter/s)": 0.032614
    },
    {
      "epoch": 0.5768541394177203,
      "grad_norm": 0.11043395847082138,
      "learning_rate": 0.00012445783921659416,
      "loss": 0.3937772214412689,
      "memory(GiB)": 78.33,
      "step": 2977,
      "token_acc": 0.8828118422898835,
      "train_speed(iter/s)": 0.032614
    },
    {
      "epoch": 0.5770479097030471,
      "grad_norm": 0.09332658350467682,
      "learning_rate": 0.00012436311645848286,
      "loss": 0.3024745285511017,
      "memory(GiB)": 78.33,
      "step": 2978,
      "token_acc": 0.9083586811047573,
      "train_speed(iter/s)": 0.032615
    },
    {
      "epoch": 0.5772416799883738,
      "grad_norm": 0.10384024679660797,
      "learning_rate": 0.00012426840423012845,
      "loss": 0.3482080101966858,
      "memory(GiB)": 78.33,
      "step": 2979,
      "token_acc": 0.8957580870528065,
      "train_speed(iter/s)": 0.032616
    },
    {
      "epoch": 0.5774354502737006,
      "grad_norm": 0.10794834792613983,
      "learning_rate": 0.0001241737025704319,
      "loss": 0.35951748490333557,
      "memory(GiB)": 78.33,
      "step": 2980,
      "token_acc": 0.894366402850384,
      "train_speed(iter/s)": 0.032617
    },
    {
      "epoch": 0.5776292205590273,
      "grad_norm": 0.13303066790103912,
      "learning_rate": 0.00012407901151828963,
      "loss": 0.3524818420410156,
      "memory(GiB)": 78.33,
      "step": 2981,
      "token_acc": 0.8969385499557914,
      "train_speed(iter/s)": 0.032618
    },
    {
      "epoch": 0.5778229908443541,
      "grad_norm": 0.11768164485692978,
      "learning_rate": 0.00012398433111259386,
      "loss": 0.3718627691268921,
      "memory(GiB)": 78.33,
      "step": 2982,
      "token_acc": 0.8906458313955291,
      "train_speed(iter/s)": 0.032619
    },
    {
      "epoch": 0.5780167611296808,
      "grad_norm": 0.10189583152532578,
      "learning_rate": 0.00012388966139223245,
      "loss": 0.3596772253513336,
      "memory(GiB)": 78.33,
      "step": 2983,
      "token_acc": 0.8941813261163735,
      "train_speed(iter/s)": 0.03262
    },
    {
      "epoch": 0.5782105314150076,
      "grad_norm": 0.10249454528093338,
      "learning_rate": 0.00012379500239608865,
      "loss": 0.3092750906944275,
      "memory(GiB)": 78.33,
      "step": 2984,
      "token_acc": 0.9063748245708734,
      "train_speed(iter/s)": 0.03262
    },
    {
      "epoch": 0.5784043017003343,
      "grad_norm": 0.12180888652801514,
      "learning_rate": 0.00012370035416304153,
      "loss": 0.3937700092792511,
      "memory(GiB)": 78.33,
      "step": 2985,
      "token_acc": 0.8859492919528534,
      "train_speed(iter/s)": 0.032621
    },
    {
      "epoch": 0.578598071985661,
      "grad_norm": 0.0898217260837555,
      "learning_rate": 0.00012360571673196565,
      "loss": 0.3232322931289673,
      "memory(GiB)": 78.33,
      "step": 2986,
      "token_acc": 0.9051626763526832,
      "train_speed(iter/s)": 0.032622
    },
    {
      "epoch": 0.5787918422709878,
      "grad_norm": 0.09598486870527267,
      "learning_rate": 0.0001235110901417312,
      "loss": 0.32468363642692566,
      "memory(GiB)": 78.33,
      "step": 2987,
      "token_acc": 0.901153603034134,
      "train_speed(iter/s)": 0.032623
    },
    {
      "epoch": 0.5789856125563145,
      "grad_norm": 0.10317398607730865,
      "learning_rate": 0.00012341647443120374,
      "loss": 0.33831608295440674,
      "memory(GiB)": 78.33,
      "step": 2988,
      "token_acc": 0.900511402902557,
      "train_speed(iter/s)": 0.032624
    },
    {
      "epoch": 0.5791793828416413,
      "grad_norm": 0.10351257771253586,
      "learning_rate": 0.0001233218696392446,
      "loss": 0.3458552956581116,
      "memory(GiB)": 78.33,
      "step": 2989,
      "token_acc": 0.8962962962962963,
      "train_speed(iter/s)": 0.032624
    },
    {
      "epoch": 0.579373153126968,
      "grad_norm": 0.09751173853874207,
      "learning_rate": 0.00012322727580471048,
      "loss": 0.338079035282135,
      "memory(GiB)": 78.33,
      "step": 2990,
      "token_acc": 0.8986410108266975,
      "train_speed(iter/s)": 0.032625
    },
    {
      "epoch": 0.5795669234122948,
      "grad_norm": 0.10420162230730057,
      "learning_rate": 0.00012313269296645356,
      "loss": 0.3703789710998535,
      "memory(GiB)": 78.33,
      "step": 2991,
      "token_acc": 0.8920427978106005,
      "train_speed(iter/s)": 0.032626
    },
    {
      "epoch": 0.5797606936976215,
      "grad_norm": 0.10459905862808228,
      "learning_rate": 0.00012303812116332163,
      "loss": 0.35847145318984985,
      "memory(GiB)": 78.33,
      "step": 2992,
      "token_acc": 0.893757727364214,
      "train_speed(iter/s)": 0.032627
    },
    {
      "epoch": 0.5799544639829483,
      "grad_norm": 0.09815093129873276,
      "learning_rate": 0.0001229435604341578,
      "loss": 0.3507353663444519,
      "memory(GiB)": 78.33,
      "step": 2993,
      "token_acc": 0.8970371099517669,
      "train_speed(iter/s)": 0.032628
    },
    {
      "epoch": 0.580148234268275,
      "grad_norm": 0.10171345621347427,
      "learning_rate": 0.00012284901081780077,
      "loss": 0.34216856956481934,
      "memory(GiB)": 78.33,
      "step": 2994,
      "token_acc": 0.8977114312267658,
      "train_speed(iter/s)": 0.032629
    },
    {
      "epoch": 0.5803420045536017,
      "grad_norm": 0.09991803765296936,
      "learning_rate": 0.00012275447235308453,
      "loss": 0.33179599046707153,
      "memory(GiB)": 78.33,
      "step": 2995,
      "token_acc": 0.9008316831683169,
      "train_speed(iter/s)": 0.032629
    },
    {
      "epoch": 0.5805357748389285,
      "grad_norm": 0.12053931504487991,
      "learning_rate": 0.00012265994507883863,
      "loss": 0.3269277811050415,
      "memory(GiB)": 78.33,
      "step": 2996,
      "token_acc": 0.9013863084178934,
      "train_speed(iter/s)": 0.03263
    },
    {
      "epoch": 0.5807295451242552,
      "grad_norm": 0.10479523986577988,
      "learning_rate": 0.00012256542903388797,
      "loss": 0.3800506591796875,
      "memory(GiB)": 78.33,
      "step": 2997,
      "token_acc": 0.8882348086815394,
      "train_speed(iter/s)": 0.032631
    },
    {
      "epoch": 0.580923315409582,
      "grad_norm": 0.10834317654371262,
      "learning_rate": 0.00012247092425705274,
      "loss": 0.3906557559967041,
      "memory(GiB)": 78.33,
      "step": 2998,
      "token_acc": 0.8865690162121335,
      "train_speed(iter/s)": 0.032632
    },
    {
      "epoch": 0.5811170856949087,
      "grad_norm": 0.09533150494098663,
      "learning_rate": 0.0001223764307871487,
      "loss": 0.32510894536972046,
      "memory(GiB)": 78.33,
      "step": 2999,
      "token_acc": 0.9044620191368873,
      "train_speed(iter/s)": 0.032633
    },
    {
      "epoch": 0.5813108559802355,
      "grad_norm": 0.11682204157114029,
      "learning_rate": 0.00012228194866298678,
      "loss": 0.3705749809741974,
      "memory(GiB)": 78.33,
      "step": 3000,
      "token_acc": 0.8889822334235664,
      "train_speed(iter/s)": 0.032633
    },
    {
      "epoch": 0.5813108559802355,
      "eval_loss": 0.40707629919052124,
      "eval_runtime": 1344.136,
      "eval_samples_per_second": 5.021,
      "eval_steps_per_second": 5.021,
      "eval_token_acc": 0.8961095095372348,
      "step": 3000
    },
    {
      "epoch": 0.5815046262655622,
      "grad_norm": 0.1150721088051796,
      "learning_rate": 0.00012218747792337335,
      "loss": 0.3748778998851776,
      "memory(GiB)": 78.33,
      "step": 3001,
      "token_acc": 0.8879601402956698,
      "train_speed(iter/s)": 0.03216
    },
    {
      "epoch": 0.581698396550889,
      "grad_norm": 0.10332752019166946,
      "learning_rate": 0.00012209301860711017,
      "loss": 0.34594491124153137,
      "memory(GiB)": 78.33,
      "step": 3002,
      "token_acc": 0.8974166376049575,
      "train_speed(iter/s)": 0.032161
    },
    {
      "epoch": 0.5818921668362157,
      "grad_norm": 0.10606271028518677,
      "learning_rate": 0.00012199857075299403,
      "loss": 0.33615994453430176,
      "memory(GiB)": 78.33,
      "step": 3003,
      "token_acc": 0.900671290493301,
      "train_speed(iter/s)": 0.032162
    },
    {
      "epoch": 0.5820859371215424,
      "grad_norm": 0.11355333030223846,
      "learning_rate": 0.00012190413439981741,
      "loss": 0.3375508785247803,
      "memory(GiB)": 78.33,
      "step": 3004,
      "token_acc": 0.8986658671863289,
      "train_speed(iter/s)": 0.032163
    },
    {
      "epoch": 0.5822797074068692,
      "grad_norm": 0.0899207592010498,
      "learning_rate": 0.00012180970958636769,
      "loss": 0.32201769948005676,
      "memory(GiB)": 78.33,
      "step": 3005,
      "token_acc": 0.9033656775887123,
      "train_speed(iter/s)": 0.032164
    },
    {
      "epoch": 0.5824734776921959,
      "grad_norm": 0.10488248616456985,
      "learning_rate": 0.00012171529635142777,
      "loss": 0.3354138433933258,
      "memory(GiB)": 78.33,
      "step": 3006,
      "token_acc": 0.8999153020892151,
      "train_speed(iter/s)": 0.032165
    },
    {
      "epoch": 0.5826672479775227,
      "grad_norm": 0.11078818142414093,
      "learning_rate": 0.00012162089473377564,
      "loss": 0.39330416917800903,
      "memory(GiB)": 78.33,
      "step": 3007,
      "token_acc": 0.8855033101341029,
      "train_speed(iter/s)": 0.032166
    },
    {
      "epoch": 0.5828610182628494,
      "grad_norm": 0.09363757818937302,
      "learning_rate": 0.00012152650477218462,
      "loss": 0.34882262349128723,
      "memory(GiB)": 78.33,
      "step": 3008,
      "token_acc": 0.8959220498015157,
      "train_speed(iter/s)": 0.032167
    },
    {
      "epoch": 0.5830547885481762,
      "grad_norm": 0.11309264600276947,
      "learning_rate": 0.00012143212650542327,
      "loss": 0.38344091176986694,
      "memory(GiB)": 78.33,
      "step": 3009,
      "token_acc": 0.8874630723781388,
      "train_speed(iter/s)": 0.032168
    },
    {
      "epoch": 0.5832485588335029,
      "grad_norm": 0.1011514663696289,
      "learning_rate": 0.00012133775997225515,
      "loss": 0.3377688527107239,
      "memory(GiB)": 78.33,
      "step": 3010,
      "token_acc": 0.8985852981969487,
      "train_speed(iter/s)": 0.032169
    },
    {
      "epoch": 0.5834423291188297,
      "grad_norm": 0.09630289673805237,
      "learning_rate": 0.00012124340521143926,
      "loss": 0.3347204327583313,
      "memory(GiB)": 78.33,
      "step": 3011,
      "token_acc": 0.9005485624724494,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.5836360994041564,
      "grad_norm": 0.09772694110870361,
      "learning_rate": 0.0001211490622617295,
      "loss": 0.33570656180381775,
      "memory(GiB)": 78.33,
      "step": 3012,
      "token_acc": 0.9012391298992437,
      "train_speed(iter/s)": 0.032171
    },
    {
      "epoch": 0.5838298696894831,
      "grad_norm": 0.11302408576011658,
      "learning_rate": 0.00012105473116187517,
      "loss": 0.36947059631347656,
      "memory(GiB)": 78.33,
      "step": 3013,
      "token_acc": 0.8906702025072324,
      "train_speed(iter/s)": 0.032172
    },
    {
      "epoch": 0.5840236399748099,
      "grad_norm": 0.09968356788158417,
      "learning_rate": 0.0001209604119506205,
      "loss": 0.3836635947227478,
      "memory(GiB)": 78.33,
      "step": 3014,
      "token_acc": 0.8870716703312993,
      "train_speed(iter/s)": 0.032172
    },
    {
      "epoch": 0.5842174102601366,
      "grad_norm": 0.1118093952536583,
      "learning_rate": 0.00012086610466670495,
      "loss": 0.3891493082046509,
      "memory(GiB)": 78.33,
      "step": 3015,
      "token_acc": 0.8857653605512787,
      "train_speed(iter/s)": 0.032173
    },
    {
      "epoch": 0.5844111805454634,
      "grad_norm": 0.10725325345993042,
      "learning_rate": 0.00012077180934886317,
      "loss": 0.3714507520198822,
      "memory(GiB)": 78.33,
      "step": 3016,
      "token_acc": 0.8898680855780956,
      "train_speed(iter/s)": 0.032174
    },
    {
      "epoch": 0.5846049508307901,
      "grad_norm": 0.10335146635770798,
      "learning_rate": 0.00012067752603582458,
      "loss": 0.32297301292419434,
      "memory(GiB)": 78.33,
      "step": 3017,
      "token_acc": 0.9023583923365265,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.5847987211161169,
      "grad_norm": 0.09637671709060669,
      "learning_rate": 0.00012058325476631404,
      "loss": 0.33302274346351624,
      "memory(GiB)": 78.33,
      "step": 3018,
      "token_acc": 0.9011041882171811,
      "train_speed(iter/s)": 0.032176
    },
    {
      "epoch": 0.5849924914014436,
      "grad_norm": 0.10183451324701309,
      "learning_rate": 0.0001204889955790512,
      "loss": 0.35752299427986145,
      "memory(GiB)": 78.33,
      "step": 3019,
      "token_acc": 0.8940680214357722,
      "train_speed(iter/s)": 0.032177
    },
    {
      "epoch": 0.5851862616867703,
      "grad_norm": 0.09284891933202744,
      "learning_rate": 0.00012039474851275087,
      "loss": 0.31136441230773926,
      "memory(GiB)": 78.33,
      "step": 3020,
      "token_acc": 0.906412598570867,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.5853800319720971,
      "grad_norm": 0.12801574170589447,
      "learning_rate": 0.00012030051360612282,
      "loss": 0.36925309896469116,
      "memory(GiB)": 78.33,
      "step": 3021,
      "token_acc": 0.890402707664066,
      "train_speed(iter/s)": 0.032179
    },
    {
      "epoch": 0.5855738022574238,
      "grad_norm": 0.10945330560207367,
      "learning_rate": 0.0001202062908978719,
      "loss": 0.37747853994369507,
      "memory(GiB)": 78.33,
      "step": 3022,
      "token_acc": 0.8870110767568217,
      "train_speed(iter/s)": 0.03218
    },
    {
      "epoch": 0.5857675725427506,
      "grad_norm": 0.10129844397306442,
      "learning_rate": 0.00012011208042669797,
      "loss": 0.3541242778301239,
      "memory(GiB)": 78.33,
      "step": 3023,
      "token_acc": 0.893968527045489,
      "train_speed(iter/s)": 0.032181
    },
    {
      "epoch": 0.5859613428280773,
      "grad_norm": 0.10958780348300934,
      "learning_rate": 0.00012001788223129563,
      "loss": 0.37422338128089905,
      "memory(GiB)": 78.33,
      "step": 3024,
      "token_acc": 0.8888448160126929,
      "train_speed(iter/s)": 0.032182
    },
    {
      "epoch": 0.5861551131134041,
      "grad_norm": 0.09564699977636337,
      "learning_rate": 0.00011992369635035475,
      "loss": 0.3554079532623291,
      "memory(GiB)": 78.33,
      "step": 3025,
      "token_acc": 0.8962656629191224,
      "train_speed(iter/s)": 0.032183
    },
    {
      "epoch": 0.5863488833987308,
      "grad_norm": 0.11680517345666885,
      "learning_rate": 0.00011982952282255994,
      "loss": 0.37712615728378296,
      "memory(GiB)": 78.33,
      "step": 3026,
      "token_acc": 0.8882045642012648,
      "train_speed(iter/s)": 0.032184
    },
    {
      "epoch": 0.5865426536840576,
      "grad_norm": 0.1049395278096199,
      "learning_rate": 0.00011973536168659089,
      "loss": 0.36447393894195557,
      "memory(GiB)": 78.33,
      "step": 3027,
      "token_acc": 0.8916041342205862,
      "train_speed(iter/s)": 0.032185
    },
    {
      "epoch": 0.5867364239693843,
      "grad_norm": 0.1006232500076294,
      "learning_rate": 0.00011964121298112194,
      "loss": 0.34536200761795044,
      "memory(GiB)": 78.33,
      "step": 3028,
      "token_acc": 0.897648835202761,
      "train_speed(iter/s)": 0.032185
    },
    {
      "epoch": 0.586930194254711,
      "grad_norm": 0.12608642876148224,
      "learning_rate": 0.00011954707674482263,
      "loss": 0.3553128242492676,
      "memory(GiB)": 78.33,
      "step": 3029,
      "token_acc": 0.8939438229119608,
      "train_speed(iter/s)": 0.032186
    },
    {
      "epoch": 0.5871239645400378,
      "grad_norm": 0.10169053822755814,
      "learning_rate": 0.00011945295301635724,
      "loss": 0.3616080582141876,
      "memory(GiB)": 78.33,
      "step": 3030,
      "token_acc": 0.892506928160307,
      "train_speed(iter/s)": 0.032187
    },
    {
      "epoch": 0.5873177348253645,
      "grad_norm": 0.09629808366298676,
      "learning_rate": 0.00011935884183438483,
      "loss": 0.33636754751205444,
      "memory(GiB)": 78.33,
      "step": 3031,
      "token_acc": 0.9002651087436792,
      "train_speed(iter/s)": 0.032188
    },
    {
      "epoch": 0.5875115051106913,
      "grad_norm": 0.10230912268161774,
      "learning_rate": 0.00011926474323755947,
      "loss": 0.352095365524292,
      "memory(GiB)": 78.33,
      "step": 3032,
      "token_acc": 0.8956463675213675,
      "train_speed(iter/s)": 0.032189
    },
    {
      "epoch": 0.587705275396018,
      "grad_norm": 0.09654974937438965,
      "learning_rate": 0.00011917065726452991,
      "loss": 0.33314049243927,
      "memory(GiB)": 78.33,
      "step": 3033,
      "token_acc": 0.9005795098091339,
      "train_speed(iter/s)": 0.03219
    },
    {
      "epoch": 0.5878990456813448,
      "grad_norm": 0.10360792279243469,
      "learning_rate": 0.00011907658395393982,
      "loss": 0.35966095328330994,
      "memory(GiB)": 78.33,
      "step": 3034,
      "token_acc": 0.8929588157145568,
      "train_speed(iter/s)": 0.032191
    },
    {
      "epoch": 0.5880928159666715,
      "grad_norm": 0.09588982164859772,
      "learning_rate": 0.00011898252334442771,
      "loss": 0.3092660903930664,
      "memory(GiB)": 78.33,
      "step": 3035,
      "token_acc": 0.9049737273763518,
      "train_speed(iter/s)": 0.032192
    },
    {
      "epoch": 0.5882865862519983,
      "grad_norm": 0.11765430122613907,
      "learning_rate": 0.00011888847547462669,
      "loss": 0.3882688879966736,
      "memory(GiB)": 78.33,
      "step": 3036,
      "token_acc": 0.8858762254901961,
      "train_speed(iter/s)": 0.032193
    },
    {
      "epoch": 0.588480356537325,
      "grad_norm": 0.1067943125963211,
      "learning_rate": 0.00011879444038316485,
      "loss": 0.34796375036239624,
      "memory(GiB)": 78.33,
      "step": 3037,
      "token_acc": 0.8972947233713282,
      "train_speed(iter/s)": 0.032194
    },
    {
      "epoch": 0.5886741268226517,
      "grad_norm": 0.12499988824129105,
      "learning_rate": 0.0001187004181086648,
      "loss": 0.34971049427986145,
      "memory(GiB)": 78.33,
      "step": 3038,
      "token_acc": 0.8963723439392272,
      "train_speed(iter/s)": 0.032195
    },
    {
      "epoch": 0.5888678971079785,
      "grad_norm": 0.10188789665699005,
      "learning_rate": 0.0001186064086897441,
      "loss": 0.355500727891922,
      "memory(GiB)": 78.33,
      "step": 3039,
      "token_acc": 0.8935444389469024,
      "train_speed(iter/s)": 0.032196
    },
    {
      "epoch": 0.5890616673933052,
      "grad_norm": 0.10140416026115417,
      "learning_rate": 0.00011851241216501492,
      "loss": 0.3354438245296478,
      "memory(GiB)": 78.33,
      "step": 3040,
      "token_acc": 0.8993545502808126,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.589255437678632,
      "grad_norm": 0.10051760822534561,
      "learning_rate": 0.00011841842857308416,
      "loss": 0.35278022289276123,
      "memory(GiB)": 78.33,
      "step": 3041,
      "token_acc": 0.8940331066341589,
      "train_speed(iter/s)": 0.032198
    },
    {
      "epoch": 0.5894492079639587,
      "grad_norm": 0.11390508711338043,
      "learning_rate": 0.00011832445795255348,
      "loss": 0.36132004857063293,
      "memory(GiB)": 78.33,
      "step": 3042,
      "token_acc": 0.891473121085595,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.5896429782492855,
      "grad_norm": 0.1137322410941124,
      "learning_rate": 0.00011823050034201902,
      "loss": 0.38379916548728943,
      "memory(GiB)": 78.33,
      "step": 3043,
      "token_acc": 0.8879169618381741,
      "train_speed(iter/s)": 0.0322
    },
    {
      "epoch": 0.5898367485346122,
      "grad_norm": 0.09486231952905655,
      "learning_rate": 0.00011813655578007181,
      "loss": 0.3457268476486206,
      "memory(GiB)": 78.33,
      "step": 3044,
      "token_acc": 0.8960932211584953,
      "train_speed(iter/s)": 0.0322
    },
    {
      "epoch": 0.590030518819939,
      "grad_norm": 0.09855726361274719,
      "learning_rate": 0.00011804262430529727,
      "loss": 0.33907079696655273,
      "memory(GiB)": 78.33,
      "step": 3045,
      "token_acc": 0.8978774011737636,
      "train_speed(iter/s)": 0.032201
    },
    {
      "epoch": 0.5902242891052657,
      "grad_norm": 0.10438596457242966,
      "learning_rate": 0.0001179487059562757,
      "loss": 0.32949021458625793,
      "memory(GiB)": 78.33,
      "step": 3046,
      "token_acc": 0.9006081617063784,
      "train_speed(iter/s)": 0.032202
    },
    {
      "epoch": 0.5904180593905924,
      "grad_norm": 0.10170579701662064,
      "learning_rate": 0.0001178548007715818,
      "loss": 0.3779696226119995,
      "memory(GiB)": 78.33,
      "step": 3047,
      "token_acc": 0.887333483856401,
      "train_speed(iter/s)": 0.032203
    },
    {
      "epoch": 0.5906118296759192,
      "grad_norm": 0.10463224351406097,
      "learning_rate": 0.000117760908789785,
      "loss": 0.37798070907592773,
      "memory(GiB)": 78.33,
      "step": 3048,
      "token_acc": 0.8907643133516514,
      "train_speed(iter/s)": 0.032204
    },
    {
      "epoch": 0.5908055999612459,
      "grad_norm": 0.09821716696023941,
      "learning_rate": 0.00011766703004944934,
      "loss": 0.3446789085865021,
      "memory(GiB)": 78.33,
      "step": 3049,
      "token_acc": 0.8958164505672609,
      "train_speed(iter/s)": 0.032205
    },
    {
      "epoch": 0.5909993702465727,
      "grad_norm": 0.12338458746671677,
      "learning_rate": 0.00011757316458913317,
      "loss": 0.3765670657157898,
      "memory(GiB)": 78.33,
      "step": 3050,
      "token_acc": 0.8903645271623579,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.5911931405318994,
      "grad_norm": 0.10490836948156357,
      "learning_rate": 0.00011747931244738973,
      "loss": 0.3376193344593048,
      "memory(GiB)": 78.33,
      "step": 3051,
      "token_acc": 0.8987225491603272,
      "train_speed(iter/s)": 0.032207
    },
    {
      "epoch": 0.5913869108172262,
      "grad_norm": 0.10726680606603622,
      "learning_rate": 0.00011738547366276645,
      "loss": 0.36437898874282837,
      "memory(GiB)": 78.33,
      "step": 3052,
      "token_acc": 0.8919469627617164,
      "train_speed(iter/s)": 0.032208
    },
    {
      "epoch": 0.5915806811025529,
      "grad_norm": 0.10234256833791733,
      "learning_rate": 0.00011729164827380557,
      "loss": 0.335938960313797,
      "memory(GiB)": 78.33,
      "step": 3053,
      "token_acc": 0.8999602227525855,
      "train_speed(iter/s)": 0.032209
    },
    {
      "epoch": 0.5917744513878797,
      "grad_norm": 0.09306668490171432,
      "learning_rate": 0.00011719783631904362,
      "loss": 0.3333013653755188,
      "memory(GiB)": 78.33,
      "step": 3054,
      "token_acc": 0.8992164035860263,
      "train_speed(iter/s)": 0.03221
    },
    {
      "epoch": 0.5919682216732064,
      "grad_norm": 0.10699167847633362,
      "learning_rate": 0.00011710403783701172,
      "loss": 0.36375150084495544,
      "memory(GiB)": 78.33,
      "step": 3055,
      "token_acc": 0.8928638271703965,
      "train_speed(iter/s)": 0.032211
    },
    {
      "epoch": 0.5921619919585331,
      "grad_norm": 0.1067671999335289,
      "learning_rate": 0.0001170102528662355,
      "loss": 0.37255626916885376,
      "memory(GiB)": 78.33,
      "step": 3056,
      "token_acc": 0.8900267627942542,
      "train_speed(iter/s)": 0.032212
    },
    {
      "epoch": 0.5923557622438599,
      "grad_norm": 0.11016660928726196,
      "learning_rate": 0.00011691648144523482,
      "loss": 0.3919016420841217,
      "memory(GiB)": 78.33,
      "step": 3057,
      "token_acc": 0.8844034568760513,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.5925495325291866,
      "grad_norm": 0.10009429603815079,
      "learning_rate": 0.00011682272361252423,
      "loss": 0.32291102409362793,
      "memory(GiB)": 78.33,
      "step": 3058,
      "token_acc": 0.9015109432285302,
      "train_speed(iter/s)": 0.032214
    },
    {
      "epoch": 0.5927433028145134,
      "grad_norm": 0.1145942285656929,
      "learning_rate": 0.00011672897940661254,
      "loss": 0.3807049095630646,
      "memory(GiB)": 78.33,
      "step": 3059,
      "token_acc": 0.8866095219750233,
      "train_speed(iter/s)": 0.032215
    },
    {
      "epoch": 0.5929370730998401,
      "grad_norm": 0.12313009053468704,
      "learning_rate": 0.00011663524886600309,
      "loss": 0.40620920062065125,
      "memory(GiB)": 78.33,
      "step": 3060,
      "token_acc": 0.8780752287360217,
      "train_speed(iter/s)": 0.032216
    },
    {
      "epoch": 0.5931308433851669,
      "grad_norm": 0.09408815950155258,
      "learning_rate": 0.00011654153202919341,
      "loss": 0.3425576388835907,
      "memory(GiB)": 78.33,
      "step": 3061,
      "token_acc": 0.8985101178563487,
      "train_speed(iter/s)": 0.032216
    },
    {
      "epoch": 0.5933246136704936,
      "grad_norm": 0.09209151566028595,
      "learning_rate": 0.00011644782893467559,
      "loss": 0.2937813997268677,
      "memory(GiB)": 78.33,
      "step": 3062,
      "token_acc": 0.9103508585965656,
      "train_speed(iter/s)": 0.032217
    },
    {
      "epoch": 0.5935183839558204,
      "grad_norm": 0.10529588907957077,
      "learning_rate": 0.00011635413962093607,
      "loss": 0.33743685483932495,
      "memory(GiB)": 78.33,
      "step": 3063,
      "token_acc": 0.9011610773855131,
      "train_speed(iter/s)": 0.032218
    },
    {
      "epoch": 0.5937121542411471,
      "grad_norm": 0.10583112388849258,
      "learning_rate": 0.00011626046412645546,
      "loss": 0.384732186794281,
      "memory(GiB)": 78.33,
      "step": 3064,
      "token_acc": 0.886413967142195,
      "train_speed(iter/s)": 0.032219
    },
    {
      "epoch": 0.5939059245264738,
      "grad_norm": 0.10110019147396088,
      "learning_rate": 0.00011616680248970887,
      "loss": 0.35833442211151123,
      "memory(GiB)": 78.33,
      "step": 3065,
      "token_acc": 0.8952516405918607,
      "train_speed(iter/s)": 0.03222
    },
    {
      "epoch": 0.5940996948118006,
      "grad_norm": 0.10245722532272339,
      "learning_rate": 0.0001160731547491656,
      "loss": 0.39178815484046936,
      "memory(GiB)": 78.33,
      "step": 3066,
      "token_acc": 0.8862267136788917,
      "train_speed(iter/s)": 0.032221
    },
    {
      "epoch": 0.5942934650971273,
      "grad_norm": 0.11041680723428726,
      "learning_rate": 0.00011597952094328933,
      "loss": 0.36065673828125,
      "memory(GiB)": 78.33,
      "step": 3067,
      "token_acc": 0.8902277926345974,
      "train_speed(iter/s)": 0.032222
    },
    {
      "epoch": 0.5944872353824541,
      "grad_norm": 0.1108374372124672,
      "learning_rate": 0.00011588590111053803,
      "loss": 0.3488714396953583,
      "memory(GiB)": 78.33,
      "step": 3068,
      "token_acc": 0.8970215739365883,
      "train_speed(iter/s)": 0.032223
    },
    {
      "epoch": 0.5946810056677808,
      "grad_norm": 0.09806838631629944,
      "learning_rate": 0.00011579229528936375,
      "loss": 0.3352605104446411,
      "memory(GiB)": 78.33,
      "step": 3069,
      "token_acc": 0.9009188114555908,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.5948747759531076,
      "grad_norm": 0.0958922728896141,
      "learning_rate": 0.00011569870351821308,
      "loss": 0.3334873616695404,
      "memory(GiB)": 78.33,
      "step": 3070,
      "token_acc": 0.8991558523371551,
      "train_speed(iter/s)": 0.032225
    },
    {
      "epoch": 0.5950685462384343,
      "grad_norm": 0.08912528306245804,
      "learning_rate": 0.00011560512583552649,
      "loss": 0.2977232336997986,
      "memory(GiB)": 78.33,
      "step": 3071,
      "token_acc": 0.9110933022487736,
      "train_speed(iter/s)": 0.032225
    },
    {
      "epoch": 0.595262316523761,
      "grad_norm": 0.11086632311344147,
      "learning_rate": 0.000115511562279739,
      "loss": 0.385816752910614,
      "memory(GiB)": 78.33,
      "step": 3072,
      "token_acc": 0.8883094428021641,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.5954560868090878,
      "grad_norm": 0.09204145520925522,
      "learning_rate": 0.0001154180128892796,
      "loss": 0.31663045287132263,
      "memory(GiB)": 78.33,
      "step": 3073,
      "token_acc": 0.9048406026202965,
      "train_speed(iter/s)": 0.032227
    },
    {
      "epoch": 0.5956498570944145,
      "grad_norm": 0.11149836331605911,
      "learning_rate": 0.00011532447770257153,
      "loss": 0.4004219174385071,
      "memory(GiB)": 78.33,
      "step": 3074,
      "token_acc": 0.883167884804189,
      "train_speed(iter/s)": 0.032228
    },
    {
      "epoch": 0.5958436273797413,
      "grad_norm": 0.10164281725883484,
      "learning_rate": 0.00011523095675803232,
      "loss": 0.3479783535003662,
      "memory(GiB)": 78.33,
      "step": 3075,
      "token_acc": 0.8950982001203567,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.596037397665068,
      "grad_norm": 0.09977372735738754,
      "learning_rate": 0.00011513745009407339,
      "loss": 0.34006303548812866,
      "memory(GiB)": 78.33,
      "step": 3076,
      "token_acc": 0.8972231350640815,
      "train_speed(iter/s)": 0.03223
    },
    {
      "epoch": 0.5962311679503948,
      "grad_norm": 0.09573015570640564,
      "learning_rate": 0.00011504395774910056,
      "loss": 0.32083189487457275,
      "memory(GiB)": 78.33,
      "step": 3077,
      "token_acc": 0.9044013992586017,
      "train_speed(iter/s)": 0.032231
    },
    {
      "epoch": 0.5964249382357215,
      "grad_norm": 0.10882751643657684,
      "learning_rate": 0.00011495047976151352,
      "loss": 0.35326480865478516,
      "memory(GiB)": 78.33,
      "step": 3078,
      "token_acc": 0.8945277761309065,
      "train_speed(iter/s)": 0.032232
    },
    {
      "epoch": 0.5966187085210483,
      "grad_norm": 0.10612433403730392,
      "learning_rate": 0.00011485701616970628,
      "loss": 0.37883704900741577,
      "memory(GiB)": 78.33,
      "step": 3079,
      "token_acc": 0.8884327457135958,
      "train_speed(iter/s)": 0.032233
    },
    {
      "epoch": 0.596812478806375,
      "grad_norm": 0.114121213555336,
      "learning_rate": 0.00011476356701206683,
      "loss": 0.3798381984233856,
      "memory(GiB)": 78.33,
      "step": 3080,
      "token_acc": 0.887470047620492,
      "train_speed(iter/s)": 0.032234
    },
    {
      "epoch": 0.5970062490917017,
      "grad_norm": 0.09286189824342728,
      "learning_rate": 0.00011467013232697721,
      "loss": 0.30042147636413574,
      "memory(GiB)": 78.33,
      "step": 3081,
      "token_acc": 0.9096791788565478,
      "train_speed(iter/s)": 0.032235
    },
    {
      "epoch": 0.5972000193770285,
      "grad_norm": 0.103480763733387,
      "learning_rate": 0.00011457671215281367,
      "loss": 0.35093623399734497,
      "memory(GiB)": 78.33,
      "step": 3082,
      "token_acc": 0.8954995525533328,
      "train_speed(iter/s)": 0.032235
    },
    {
      "epoch": 0.5973937896623552,
      "grad_norm": 0.0931040421128273,
      "learning_rate": 0.00011448330652794625,
      "loss": 0.3155452013015747,
      "memory(GiB)": 78.33,
      "step": 3083,
      "token_acc": 0.9038670039754246,
      "train_speed(iter/s)": 0.032236
    },
    {
      "epoch": 0.597587559947682,
      "grad_norm": 0.10366553068161011,
      "learning_rate": 0.00011438991549073928,
      "loss": 0.3622359037399292,
      "memory(GiB)": 78.33,
      "step": 3084,
      "token_acc": 0.8940353881278539,
      "train_speed(iter/s)": 0.032237
    },
    {
      "epoch": 0.5977813302330087,
      "grad_norm": 0.09329306334257126,
      "learning_rate": 0.00011429653907955083,
      "loss": 0.3334829807281494,
      "memory(GiB)": 78.33,
      "step": 3085,
      "token_acc": 0.9010285939268665,
      "train_speed(iter/s)": 0.032238
    },
    {
      "epoch": 0.5979751005183355,
      "grad_norm": 0.09546185284852982,
      "learning_rate": 0.00011420317733273319,
      "loss": 0.32923221588134766,
      "memory(GiB)": 78.33,
      "step": 3086,
      "token_acc": 0.9020269413227124,
      "train_speed(iter/s)": 0.032239
    },
    {
      "epoch": 0.5981688708036622,
      "grad_norm": 0.09393859654664993,
      "learning_rate": 0.00011410983028863249,
      "loss": 0.315865159034729,
      "memory(GiB)": 78.33,
      "step": 3087,
      "token_acc": 0.9042561262423184,
      "train_speed(iter/s)": 0.03224
    },
    {
      "epoch": 0.598362641088989,
      "grad_norm": 0.10992894321680069,
      "learning_rate": 0.0001140164979855889,
      "loss": 0.39187929034233093,
      "memory(GiB)": 78.33,
      "step": 3088,
      "token_acc": 0.8844834915686552,
      "train_speed(iter/s)": 0.032241
    },
    {
      "epoch": 0.5985564113743157,
      "grad_norm": 0.10544363409280777,
      "learning_rate": 0.00011392318046193656,
      "loss": 0.3612055778503418,
      "memory(GiB)": 78.33,
      "step": 3089,
      "token_acc": 0.8903434157031227,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.5987501816596424,
      "grad_norm": 0.10547579824924469,
      "learning_rate": 0.00011382987775600336,
      "loss": 0.37269455194473267,
      "memory(GiB)": 78.33,
      "step": 3090,
      "token_acc": 0.8905209503713244,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.5989439519449692,
      "grad_norm": 0.10508310049772263,
      "learning_rate": 0.00011373658990611134,
      "loss": 0.36013132333755493,
      "memory(GiB)": 78.33,
      "step": 3091,
      "token_acc": 0.8923727863707689,
      "train_speed(iter/s)": 0.032243
    },
    {
      "epoch": 0.5991377222302959,
      "grad_norm": 0.10769172757863998,
      "learning_rate": 0.00011364331695057627,
      "loss": 0.36448052525520325,
      "memory(GiB)": 78.33,
      "step": 3092,
      "token_acc": 0.8934169278996865,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.5993314925156227,
      "grad_norm": 0.10621456056833267,
      "learning_rate": 0.00011355005892770788,
      "loss": 0.35971012711524963,
      "memory(GiB)": 78.33,
      "step": 3093,
      "token_acc": 0.8954783003862758,
      "train_speed(iter/s)": 0.032245
    },
    {
      "epoch": 0.5995252628009494,
      "grad_norm": 0.10547253489494324,
      "learning_rate": 0.00011345681587580971,
      "loss": 0.38230541348457336,
      "memory(GiB)": 78.33,
      "step": 3094,
      "token_acc": 0.8890573803785643,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.5997190330862762,
      "grad_norm": 0.11314172297716141,
      "learning_rate": 0.00011336358783317918,
      "loss": 0.3828127086162567,
      "memory(GiB)": 78.33,
      "step": 3095,
      "token_acc": 0.8876047830374754,
      "train_speed(iter/s)": 0.032247
    },
    {
      "epoch": 0.5999128033716029,
      "grad_norm": 0.1083948016166687,
      "learning_rate": 0.00011327037483810767,
      "loss": 0.34444913268089294,
      "memory(GiB)": 78.33,
      "step": 3096,
      "token_acc": 0.8984524686809138,
      "train_speed(iter/s)": 0.032248
    },
    {
      "epoch": 0.6001065736569298,
      "grad_norm": 0.1084941178560257,
      "learning_rate": 0.00011317717692888012,
      "loss": 0.35603395104408264,
      "memory(GiB)": 78.33,
      "step": 3097,
      "token_acc": 0.8943460416891673,
      "train_speed(iter/s)": 0.032249
    },
    {
      "epoch": 0.6003003439422565,
      "grad_norm": 0.10329198837280273,
      "learning_rate": 0.0001130839941437755,
      "loss": 0.3441343903541565,
      "memory(GiB)": 78.33,
      "step": 3098,
      "token_acc": 0.8977043317759704,
      "train_speed(iter/s)": 0.03225
    },
    {
      "epoch": 0.6004941142275833,
      "grad_norm": 0.10437928140163422,
      "learning_rate": 0.00011299082652106642,
      "loss": 0.32754242420196533,
      "memory(GiB)": 78.33,
      "step": 3099,
      "token_acc": 0.9024354801209021,
      "train_speed(iter/s)": 0.032251
    },
    {
      "epoch": 0.60068788451291,
      "grad_norm": 0.10441198199987411,
      "learning_rate": 0.00011289767409901936,
      "loss": 0.34080564975738525,
      "memory(GiB)": 78.33,
      "step": 3100,
      "token_acc": 0.9004198514371837,
      "train_speed(iter/s)": 0.032252
    },
    {
      "epoch": 0.6008816547982367,
      "grad_norm": 0.09256397187709808,
      "learning_rate": 0.00011280453691589461,
      "loss": 0.34055426716804504,
      "memory(GiB)": 78.33,
      "step": 3101,
      "token_acc": 0.8970888269749687,
      "train_speed(iter/s)": 0.032252
    },
    {
      "epoch": 0.6010754250835635,
      "grad_norm": 0.10573244839906693,
      "learning_rate": 0.00011271141500994595,
      "loss": 0.37066394090652466,
      "memory(GiB)": 78.33,
      "step": 3102,
      "token_acc": 0.890913167778463,
      "train_speed(iter/s)": 0.032253
    },
    {
      "epoch": 0.6012691953688902,
      "grad_norm": 0.1301407516002655,
      "learning_rate": 0.0001126183084194212,
      "loss": 0.34675294160842896,
      "memory(GiB)": 78.33,
      "step": 3103,
      "token_acc": 0.8968794248992484,
      "train_speed(iter/s)": 0.032254
    },
    {
      "epoch": 0.601462965654217,
      "grad_norm": 0.09512253850698471,
      "learning_rate": 0.00011252521718256159,
      "loss": 0.34905996918678284,
      "memory(GiB)": 78.33,
      "step": 3104,
      "token_acc": 0.8967058594888738,
      "train_speed(iter/s)": 0.032255
    },
    {
      "epoch": 0.6016567359395437,
      "grad_norm": 0.11211273819208145,
      "learning_rate": 0.0001124321413376023,
      "loss": 0.3654071092605591,
      "memory(GiB)": 78.33,
      "step": 3105,
      "token_acc": 0.8934378629500581,
      "train_speed(iter/s)": 0.032256
    },
    {
      "epoch": 0.6018505062248705,
      "grad_norm": 0.10533998906612396,
      "learning_rate": 0.00011233908092277203,
      "loss": 0.3517376780509949,
      "memory(GiB)": 78.33,
      "step": 3106,
      "token_acc": 0.8942503737936659,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.6020442765101972,
      "grad_norm": 0.11167991906404495,
      "learning_rate": 0.00011224603597629322,
      "loss": 0.3245575428009033,
      "memory(GiB)": 78.33,
      "step": 3107,
      "token_acc": 0.9026755252289459,
      "train_speed(iter/s)": 0.032258
    },
    {
      "epoch": 0.602238046795524,
      "grad_norm": 0.11581658571958542,
      "learning_rate": 0.00011215300653638199,
      "loss": 0.36745405197143555,
      "memory(GiB)": 78.33,
      "step": 3108,
      "token_acc": 0.892201612120035,
      "train_speed(iter/s)": 0.032259
    },
    {
      "epoch": 0.6024318170808507,
      "grad_norm": 0.10512839257717133,
      "learning_rate": 0.00011205999264124786,
      "loss": 0.3423347771167755,
      "memory(GiB)": 78.33,
      "step": 3109,
      "token_acc": 0.8962487920583326,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 0.6026255873661774,
      "grad_norm": 0.10966593772172928,
      "learning_rate": 0.00011196699432909435,
      "loss": 0.3704321086406708,
      "memory(GiB)": 78.33,
      "step": 3110,
      "token_acc": 0.8911972524843059,
      "train_speed(iter/s)": 0.032261
    },
    {
      "epoch": 0.6028193576515042,
      "grad_norm": 0.09262005239725113,
      "learning_rate": 0.00011187401163811816,
      "loss": 0.31054481863975525,
      "memory(GiB)": 78.33,
      "step": 3111,
      "token_acc": 0.9069608679751311,
      "train_speed(iter/s)": 0.032261
    },
    {
      "epoch": 0.6030131279368309,
      "grad_norm": 0.11650776118040085,
      "learning_rate": 0.00011178104460650993,
      "loss": 0.40273237228393555,
      "memory(GiB)": 78.33,
      "step": 3112,
      "token_acc": 0.8809349247778527,
      "train_speed(iter/s)": 0.032262
    },
    {
      "epoch": 0.6032068982221577,
      "grad_norm": 0.10926227271556854,
      "learning_rate": 0.00011168809327245361,
      "loss": 0.3614426851272583,
      "memory(GiB)": 78.33,
      "step": 3113,
      "token_acc": 0.8933805533961015,
      "train_speed(iter/s)": 0.032263
    },
    {
      "epoch": 0.6034006685074844,
      "grad_norm": 0.09592621773481369,
      "learning_rate": 0.00011159515767412688,
      "loss": 0.3322632312774658,
      "memory(GiB)": 78.33,
      "step": 3114,
      "token_acc": 0.9001508923461158,
      "train_speed(iter/s)": 0.032264
    },
    {
      "epoch": 0.6035944387928112,
      "grad_norm": 0.09553615748882294,
      "learning_rate": 0.00011150223784970092,
      "loss": 0.31129273772239685,
      "memory(GiB)": 78.33,
      "step": 3115,
      "token_acc": 0.9059551739629611,
      "train_speed(iter/s)": 0.032265
    },
    {
      "epoch": 0.6037882090781379,
      "grad_norm": 0.113133504986763,
      "learning_rate": 0.0001114093338373403,
      "loss": 0.38860899209976196,
      "memory(GiB)": 78.33,
      "step": 3116,
      "token_acc": 0.885463585843814,
      "train_speed(iter/s)": 0.032266
    },
    {
      "epoch": 0.6039819793634646,
      "grad_norm": 0.09849842637777328,
      "learning_rate": 0.0001113164456752033,
      "loss": 0.3431742489337921,
      "memory(GiB)": 78.33,
      "step": 3117,
      "token_acc": 0.8975163928200713,
      "train_speed(iter/s)": 0.032267
    },
    {
      "epoch": 0.6041757496487914,
      "grad_norm": 0.10167428106069565,
      "learning_rate": 0.00011122357340144148,
      "loss": 0.37376028299331665,
      "memory(GiB)": 78.33,
      "step": 3118,
      "token_acc": 0.8888616757830571,
      "train_speed(iter/s)": 0.032268
    },
    {
      "epoch": 0.6043695199341181,
      "grad_norm": 0.11035677045583725,
      "learning_rate": 0.00011113071705420004,
      "loss": 0.36835113167762756,
      "memory(GiB)": 78.33,
      "step": 3119,
      "token_acc": 0.8914904111178809,
      "train_speed(iter/s)": 0.032269
    },
    {
      "epoch": 0.6045632902194449,
      "grad_norm": 0.10013226419687271,
      "learning_rate": 0.00011103787667161753,
      "loss": 0.3682772219181061,
      "memory(GiB)": 78.33,
      "step": 3120,
      "token_acc": 0.8906798812862472,
      "train_speed(iter/s)": 0.03227
    },
    {
      "epoch": 0.6047570605047716,
      "grad_norm": 0.10240922123193741,
      "learning_rate": 0.00011094505229182605,
      "loss": 0.33320745825767517,
      "memory(GiB)": 78.33,
      "step": 3121,
      "token_acc": 0.8995880980499467,
      "train_speed(iter/s)": 0.03227
    },
    {
      "epoch": 0.6049508307900984,
      "grad_norm": 0.10636462271213531,
      "learning_rate": 0.00011085224395295109,
      "loss": 0.3461098074913025,
      "memory(GiB)": 78.33,
      "step": 3122,
      "token_acc": 0.8970420099088271,
      "train_speed(iter/s)": 0.032271
    },
    {
      "epoch": 0.6051446010754251,
      "grad_norm": 0.15055952966213226,
      "learning_rate": 0.00011075945169311141,
      "loss": 0.4095023572444916,
      "memory(GiB)": 78.33,
      "step": 3123,
      "token_acc": 0.8795589223532286,
      "train_speed(iter/s)": 0.032272
    },
    {
      "epoch": 0.6053383713607519,
      "grad_norm": 0.10108061134815216,
      "learning_rate": 0.00011066667555041942,
      "loss": 0.33705589175224304,
      "memory(GiB)": 78.33,
      "step": 3124,
      "token_acc": 0.8990451697627971,
      "train_speed(iter/s)": 0.032273
    },
    {
      "epoch": 0.6055321416460786,
      "grad_norm": 0.10497518628835678,
      "learning_rate": 0.00011057391556298065,
      "loss": 0.3509282171726227,
      "memory(GiB)": 78.33,
      "step": 3125,
      "token_acc": 0.8944882780464354,
      "train_speed(iter/s)": 0.032274
    },
    {
      "epoch": 0.6057259119314053,
      "grad_norm": 0.0899055078625679,
      "learning_rate": 0.0001104811717688942,
      "loss": 0.3112187385559082,
      "memory(GiB)": 78.33,
      "step": 3126,
      "token_acc": 0.9061983165066279,
      "train_speed(iter/s)": 0.032275
    },
    {
      "epoch": 0.6059196822167321,
      "grad_norm": 0.10332870483398438,
      "learning_rate": 0.00011038844420625239,
      "loss": 0.40198662877082825,
      "memory(GiB)": 78.33,
      "step": 3127,
      "token_acc": 0.8823683727275112,
      "train_speed(iter/s)": 0.032276
    },
    {
      "epoch": 0.6061134525020588,
      "grad_norm": 0.1093793734908104,
      "learning_rate": 0.00011029573291314094,
      "loss": 0.36452385783195496,
      "memory(GiB)": 78.33,
      "step": 3128,
      "token_acc": 0.8916964258920266,
      "train_speed(iter/s)": 0.032277
    },
    {
      "epoch": 0.6063072227873856,
      "grad_norm": 0.10296124219894409,
      "learning_rate": 0.00011020303792763896,
      "loss": 0.31651031970977783,
      "memory(GiB)": 78.33,
      "step": 3129,
      "token_acc": 0.9070627097315436,
      "train_speed(iter/s)": 0.032277
    },
    {
      "epoch": 0.6065009930727123,
      "grad_norm": 0.10293302685022354,
      "learning_rate": 0.00011011035928781861,
      "loss": 0.34327608346939087,
      "memory(GiB)": 78.33,
      "step": 3130,
      "token_acc": 0.896756412974348,
      "train_speed(iter/s)": 0.032278
    },
    {
      "epoch": 0.6066947633580391,
      "grad_norm": 0.1089029610157013,
      "learning_rate": 0.00011001769703174564,
      "loss": 0.3363596796989441,
      "memory(GiB)": 78.33,
      "step": 3131,
      "token_acc": 0.9004472523691865,
      "train_speed(iter/s)": 0.032279
    },
    {
      "epoch": 0.6068885336433658,
      "grad_norm": 0.10041207820177078,
      "learning_rate": 0.0001099250511974788,
      "loss": 0.3575053811073303,
      "memory(GiB)": 78.33,
      "step": 3132,
      "token_acc": 0.8955265421790395,
      "train_speed(iter/s)": 0.03228
    },
    {
      "epoch": 0.6070823039286926,
      "grad_norm": 0.11202222853899002,
      "learning_rate": 0.00010983242182307032,
      "loss": 0.363272488117218,
      "memory(GiB)": 78.33,
      "step": 3133,
      "token_acc": 0.8899773926149209,
      "train_speed(iter/s)": 0.032281
    },
    {
      "epoch": 0.6072760742140193,
      "grad_norm": 0.10849933326244354,
      "learning_rate": 0.00010973980894656555,
      "loss": 0.3696746826171875,
      "memory(GiB)": 78.33,
      "step": 3134,
      "token_acc": 0.8896282494898412,
      "train_speed(iter/s)": 0.032282
    },
    {
      "epoch": 0.607469844499346,
      "grad_norm": 0.10315241664648056,
      "learning_rate": 0.00010964721260600305,
      "loss": 0.34507814049720764,
      "memory(GiB)": 78.33,
      "step": 3135,
      "token_acc": 0.897193022987178,
      "train_speed(iter/s)": 0.032283
    },
    {
      "epoch": 0.6076636147846728,
      "grad_norm": 0.11026407033205032,
      "learning_rate": 0.00010955463283941472,
      "loss": 0.3476986289024353,
      "memory(GiB)": 78.33,
      "step": 3136,
      "token_acc": 0.8966703526231724,
      "train_speed(iter/s)": 0.032284
    },
    {
      "epoch": 0.6078573850699995,
      "grad_norm": 0.09877464920282364,
      "learning_rate": 0.00010946206968482542,
      "loss": 0.33894383907318115,
      "memory(GiB)": 78.33,
      "step": 3137,
      "token_acc": 0.8988343465809174,
      "train_speed(iter/s)": 0.032284
    },
    {
      "epoch": 0.6080511553553263,
      "grad_norm": 0.09642918407917023,
      "learning_rate": 0.00010936952318025344,
      "loss": 0.3351293206214905,
      "memory(GiB)": 78.33,
      "step": 3138,
      "token_acc": 0.8997663427327279,
      "train_speed(iter/s)": 0.032285
    },
    {
      "epoch": 0.608244925640653,
      "grad_norm": 0.10105688869953156,
      "learning_rate": 0.00010927699336371003,
      "loss": 0.32609879970550537,
      "memory(GiB)": 78.33,
      "step": 3139,
      "token_acc": 0.9029436501261564,
      "train_speed(iter/s)": 0.032286
    },
    {
      "epoch": 0.6084386959259798,
      "grad_norm": 0.10417565703392029,
      "learning_rate": 0.00010918448027319971,
      "loss": 0.35680675506591797,
      "memory(GiB)": 78.33,
      "step": 3140,
      "token_acc": 0.8950784207679827,
      "train_speed(iter/s)": 0.032287
    },
    {
      "epoch": 0.6086324662113065,
      "grad_norm": 0.10269538313150406,
      "learning_rate": 0.00010909198394672018,
      "loss": 0.36641865968704224,
      "memory(GiB)": 78.33,
      "step": 3141,
      "token_acc": 0.8919290565190549,
      "train_speed(iter/s)": 0.032288
    },
    {
      "epoch": 0.6088262364966333,
      "grad_norm": 0.10089493542909622,
      "learning_rate": 0.000108999504422262,
      "loss": 0.37600627541542053,
      "memory(GiB)": 78.33,
      "step": 3142,
      "token_acc": 0.8877864583333334,
      "train_speed(iter/s)": 0.032289
    },
    {
      "epoch": 0.60902000678196,
      "grad_norm": 0.10720504820346832,
      "learning_rate": 0.00010890704173780916,
      "loss": 0.35292261838912964,
      "memory(GiB)": 78.33,
      "step": 3143,
      "token_acc": 0.8938151494093121,
      "train_speed(iter/s)": 0.03229
    },
    {
      "epoch": 0.6092137770672867,
      "grad_norm": 0.09673202037811279,
      "learning_rate": 0.00010881459593133842,
      "loss": 0.32286348938941956,
      "memory(GiB)": 78.33,
      "step": 3144,
      "token_acc": 0.9035548686244204,
      "train_speed(iter/s)": 0.032291
    },
    {
      "epoch": 0.6094075473526135,
      "grad_norm": 0.10169202834367752,
      "learning_rate": 0.00010872216704081986,
      "loss": 0.345810204744339,
      "memory(GiB)": 78.33,
      "step": 3145,
      "token_acc": 0.8962894744659993,
      "train_speed(iter/s)": 0.032291
    },
    {
      "epoch": 0.6096013176379402,
      "grad_norm": 0.09147030860185623,
      "learning_rate": 0.00010862975510421642,
      "loss": 0.30978208780288696,
      "memory(GiB)": 78.33,
      "step": 3146,
      "token_acc": 0.9076807434741014,
      "train_speed(iter/s)": 0.032292
    },
    {
      "epoch": 0.609795087923267,
      "grad_norm": 0.09909648448228836,
      "learning_rate": 0.00010853736015948425,
      "loss": 0.339855819940567,
      "memory(GiB)": 78.33,
      "step": 3147,
      "token_acc": 0.8987607001405391,
      "train_speed(iter/s)": 0.032293
    },
    {
      "epoch": 0.6099888582085937,
      "grad_norm": 0.21334044635295868,
      "learning_rate": 0.00010844498224457246,
      "loss": 0.34740781784057617,
      "memory(GiB)": 78.33,
      "step": 3148,
      "token_acc": 0.8966986427186447,
      "train_speed(iter/s)": 0.032294
    },
    {
      "epoch": 0.6101826284939205,
      "grad_norm": 0.11788555234670639,
      "learning_rate": 0.00010835262139742303,
      "loss": 0.36575233936309814,
      "memory(GiB)": 78.33,
      "step": 3149,
      "token_acc": 0.8924587929111414,
      "train_speed(iter/s)": 0.032295
    },
    {
      "epoch": 0.6103763987792472,
      "grad_norm": 0.09829849749803543,
      "learning_rate": 0.00010826027765597116,
      "loss": 0.3394019901752472,
      "memory(GiB)": 78.33,
      "step": 3150,
      "token_acc": 0.8985713198162762,
      "train_speed(iter/s)": 0.032296
    },
    {
      "epoch": 0.610570169064574,
      "grad_norm": 0.1058797761797905,
      "learning_rate": 0.00010816795105814479,
      "loss": 0.3209468424320221,
      "memory(GiB)": 78.33,
      "step": 3151,
      "token_acc": 0.9035917319430982,
      "train_speed(iter/s)": 0.032296
    },
    {
      "epoch": 0.6107639393499007,
      "grad_norm": 0.09663695096969604,
      "learning_rate": 0.000108075641641865,
      "loss": 0.3185003399848938,
      "memory(GiB)": 78.33,
      "step": 3152,
      "token_acc": 0.904547132985177,
      "train_speed(iter/s)": 0.032297
    },
    {
      "epoch": 0.6109577096352274,
      "grad_norm": 0.09900177270174026,
      "learning_rate": 0.00010798334944504572,
      "loss": 0.3523489832878113,
      "memory(GiB)": 78.33,
      "step": 3153,
      "token_acc": 0.8945492208011177,
      "train_speed(iter/s)": 0.032298
    },
    {
      "epoch": 0.6111514799205542,
      "grad_norm": 0.11104562878608704,
      "learning_rate": 0.00010789107450559386,
      "loss": 0.37791165709495544,
      "memory(GiB)": 78.33,
      "step": 3154,
      "token_acc": 0.8904507301368287,
      "train_speed(iter/s)": 0.032299
    },
    {
      "epoch": 0.6113452502058809,
      "grad_norm": 0.11416266113519669,
      "learning_rate": 0.00010779881686140927,
      "loss": 0.34979140758514404,
      "memory(GiB)": 78.33,
      "step": 3155,
      "token_acc": 0.895737220889463,
      "train_speed(iter/s)": 0.0323
    },
    {
      "epoch": 0.6115390204912077,
      "grad_norm": 0.10298382490873337,
      "learning_rate": 0.00010770657655038453,
      "loss": 0.339542031288147,
      "memory(GiB)": 78.33,
      "step": 3156,
      "token_acc": 0.8982814065053508,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.6117327907765344,
      "grad_norm": 0.09547599405050278,
      "learning_rate": 0.00010761435361040531,
      "loss": 0.36547231674194336,
      "memory(GiB)": 78.33,
      "step": 3157,
      "token_acc": 0.8895893839578759,
      "train_speed(iter/s)": 0.032302
    },
    {
      "epoch": 0.6119265610618612,
      "grad_norm": 0.10469914972782135,
      "learning_rate": 0.00010752214807934996,
      "loss": 0.31865012645721436,
      "memory(GiB)": 78.33,
      "step": 3158,
      "token_acc": 0.9054883979435876,
      "train_speed(iter/s)": 0.032302
    },
    {
      "epoch": 0.6121203313471879,
      "grad_norm": 0.10441134870052338,
      "learning_rate": 0.00010742995999508987,
      "loss": 0.36503875255584717,
      "memory(GiB)": 78.33,
      "step": 3159,
      "token_acc": 0.8934720034758051,
      "train_speed(iter/s)": 0.032303
    },
    {
      "epoch": 0.6123141016325147,
      "grad_norm": 0.10514453798532486,
      "learning_rate": 0.00010733778939548905,
      "loss": 0.3756176829338074,
      "memory(GiB)": 78.33,
      "step": 3160,
      "token_acc": 0.8889959795120339,
      "train_speed(iter/s)": 0.032304
    },
    {
      "epoch": 0.6125078719178414,
      "grad_norm": 0.10831741243600845,
      "learning_rate": 0.00010724563631840451,
      "loss": 0.39807701110839844,
      "memory(GiB)": 78.33,
      "step": 3161,
      "token_acc": 0.8838846894490271,
      "train_speed(iter/s)": 0.032305
    },
    {
      "epoch": 0.6127016422031681,
      "grad_norm": 0.10723893344402313,
      "learning_rate": 0.00010715350080168606,
      "loss": 0.38960257172584534,
      "memory(GiB)": 78.33,
      "step": 3162,
      "token_acc": 0.8852941176470588,
      "train_speed(iter/s)": 0.032306
    },
    {
      "epoch": 0.6128954124884949,
      "grad_norm": 0.10730850696563721,
      "learning_rate": 0.00010706138288317609,
      "loss": 0.3630613386631012,
      "memory(GiB)": 78.33,
      "step": 3163,
      "token_acc": 0.8937855052504734,
      "train_speed(iter/s)": 0.032307
    },
    {
      "epoch": 0.6130891827738216,
      "grad_norm": 0.09366155415773392,
      "learning_rate": 0.00010696928260070999,
      "loss": 0.3480615019798279,
      "memory(GiB)": 78.33,
      "step": 3164,
      "token_acc": 0.8973462201951934,
      "train_speed(iter/s)": 0.032308
    },
    {
      "epoch": 0.6132829530591484,
      "grad_norm": 0.10169912874698639,
      "learning_rate": 0.00010687719999211583,
      "loss": 0.36037591099739075,
      "memory(GiB)": 78.33,
      "step": 3165,
      "token_acc": 0.892155500597438,
      "train_speed(iter/s)": 0.032309
    },
    {
      "epoch": 0.6134767233444751,
      "grad_norm": 0.10275442153215408,
      "learning_rate": 0.00010678513509521435,
      "loss": 0.3725361227989197,
      "memory(GiB)": 78.33,
      "step": 3166,
      "token_acc": 0.8887706641060908,
      "train_speed(iter/s)": 0.032309
    },
    {
      "epoch": 0.6136704936298019,
      "grad_norm": 0.10723954439163208,
      "learning_rate": 0.00010669308794781914,
      "loss": 0.3526693880558014,
      "memory(GiB)": 78.33,
      "step": 3167,
      "token_acc": 0.8959228785459904,
      "train_speed(iter/s)": 0.03231
    },
    {
      "epoch": 0.6138642639151286,
      "grad_norm": 0.1202649399638176,
      "learning_rate": 0.0001066010585877364,
      "loss": 0.4182667136192322,
      "memory(GiB)": 78.33,
      "step": 3168,
      "token_acc": 0.8773098963363596,
      "train_speed(iter/s)": 0.032311
    },
    {
      "epoch": 0.6140580342004553,
      "grad_norm": 0.1005726233124733,
      "learning_rate": 0.00010650904705276513,
      "loss": 0.33930516242980957,
      "memory(GiB)": 78.33,
      "step": 3169,
      "token_acc": 0.8973083132625784,
      "train_speed(iter/s)": 0.032312
    },
    {
      "epoch": 0.6142518044857821,
      "grad_norm": 0.10162294656038284,
      "learning_rate": 0.0001064170533806968,
      "loss": 0.3408072292804718,
      "memory(GiB)": 78.33,
      "step": 3170,
      "token_acc": 0.8978023358030893,
      "train_speed(iter/s)": 0.032313
    },
    {
      "epoch": 0.6144455747711088,
      "grad_norm": 0.0949673056602478,
      "learning_rate": 0.00010632507760931581,
      "loss": 0.3130902349948883,
      "memory(GiB)": 78.33,
      "step": 3171,
      "token_acc": 0.9057638586329737,
      "train_speed(iter/s)": 0.032314
    },
    {
      "epoch": 0.6146393450564356,
      "grad_norm": 0.10987738519906998,
      "learning_rate": 0.000106233119776399,
      "loss": 0.35283100605010986,
      "memory(GiB)": 78.33,
      "step": 3172,
      "token_acc": 0.8905293376688052,
      "train_speed(iter/s)": 0.032315
    },
    {
      "epoch": 0.6148331153417623,
      "grad_norm": 0.10509679466485977,
      "learning_rate": 0.00010614117991971598,
      "loss": 0.36292973160743713,
      "memory(GiB)": 78.33,
      "step": 3173,
      "token_acc": 0.8947354138398914,
      "train_speed(iter/s)": 0.032316
    },
    {
      "epoch": 0.6150268856270891,
      "grad_norm": 0.1026267558336258,
      "learning_rate": 0.00010604925807702895,
      "loss": 0.3576149046421051,
      "memory(GiB)": 78.33,
      "step": 3174,
      "token_acc": 0.8960543506663182,
      "train_speed(iter/s)": 0.032317
    },
    {
      "epoch": 0.6152206559124158,
      "grad_norm": 0.1011863574385643,
      "learning_rate": 0.00010595735428609256,
      "loss": 0.3429381251335144,
      "memory(GiB)": 78.33,
      "step": 3175,
      "token_acc": 0.8972618182729163,
      "train_speed(iter/s)": 0.032317
    },
    {
      "epoch": 0.6154144261977426,
      "grad_norm": 0.09697016328573227,
      "learning_rate": 0.0001058654685846543,
      "loss": 0.3507676422595978,
      "memory(GiB)": 78.33,
      "step": 3176,
      "token_acc": 0.8970009395870896,
      "train_speed(iter/s)": 0.032318
    },
    {
      "epoch": 0.6156081964830693,
      "grad_norm": 0.0992831438779831,
      "learning_rate": 0.00010577360101045396,
      "loss": 0.3351018726825714,
      "memory(GiB)": 78.33,
      "step": 3177,
      "token_acc": 0.8992732671719934,
      "train_speed(iter/s)": 0.032319
    },
    {
      "epoch": 0.615801966768396,
      "grad_norm": 0.1062702089548111,
      "learning_rate": 0.00010568175160122414,
      "loss": 0.38858115673065186,
      "memory(GiB)": 78.33,
      "step": 3178,
      "token_acc": 0.8845210155749946,
      "train_speed(iter/s)": 0.03232
    },
    {
      "epoch": 0.6159957370537228,
      "grad_norm": 0.09542369842529297,
      "learning_rate": 0.00010558992039468979,
      "loss": 0.32729557156562805,
      "memory(GiB)": 78.33,
      "step": 3179,
      "token_acc": 0.9012937964120754,
      "train_speed(iter/s)": 0.032321
    },
    {
      "epoch": 0.6161895073390495,
      "grad_norm": 0.09666220843791962,
      "learning_rate": 0.00010549810742856847,
      "loss": 0.3226274847984314,
      "memory(GiB)": 78.33,
      "step": 3180,
      "token_acc": 0.9038066656136471,
      "train_speed(iter/s)": 0.032322
    },
    {
      "epoch": 0.6163832776243763,
      "grad_norm": 0.12249480187892914,
      "learning_rate": 0.0001054063127405703,
      "loss": 0.4036695063114166,
      "memory(GiB)": 78.33,
      "step": 3181,
      "token_acc": 0.8821568293927539,
      "train_speed(iter/s)": 0.032323
    },
    {
      "epoch": 0.616577047909703,
      "grad_norm": 0.10205438733100891,
      "learning_rate": 0.00010531453636839771,
      "loss": 0.3525814116001129,
      "memory(GiB)": 78.33,
      "step": 3182,
      "token_acc": 0.8946187098204389,
      "train_speed(iter/s)": 0.032323
    },
    {
      "epoch": 0.6167708181950298,
      "grad_norm": 0.10525421798229218,
      "learning_rate": 0.00010522277834974585,
      "loss": 0.3661832809448242,
      "memory(GiB)": 78.33,
      "step": 3183,
      "token_acc": 0.8905501755879536,
      "train_speed(iter/s)": 0.032324
    },
    {
      "epoch": 0.6169645884803565,
      "grad_norm": 0.09824883937835693,
      "learning_rate": 0.00010513103872230206,
      "loss": 0.3395775854587555,
      "memory(GiB)": 78.33,
      "step": 3184,
      "token_acc": 0.8979004582756048,
      "train_speed(iter/s)": 0.032325
    },
    {
      "epoch": 0.6171583587656833,
      "grad_norm": 0.10357562452554703,
      "learning_rate": 0.00010503931752374637,
      "loss": 0.330016553401947,
      "memory(GiB)": 78.33,
      "step": 3185,
      "token_acc": 0.903437815975733,
      "train_speed(iter/s)": 0.032326
    },
    {
      "epoch": 0.61735212905101,
      "grad_norm": 0.09775417298078537,
      "learning_rate": 0.00010494761479175107,
      "loss": 0.3158339262008667,
      "memory(GiB)": 78.33,
      "step": 3186,
      "token_acc": 0.9043734015345268,
      "train_speed(iter/s)": 0.032327
    },
    {
      "epoch": 0.6175458993363367,
      "grad_norm": 0.09946906566619873,
      "learning_rate": 0.000104855930563981,
      "loss": 0.3278699517250061,
      "memory(GiB)": 78.33,
      "step": 3187,
      "token_acc": 0.9031324520176881,
      "train_speed(iter/s)": 0.032328
    },
    {
      "epoch": 0.6177396696216635,
      "grad_norm": 0.09458330273628235,
      "learning_rate": 0.00010476426487809338,
      "loss": 0.3177909255027771,
      "memory(GiB)": 78.33,
      "step": 3188,
      "token_acc": 0.9047277202338222,
      "train_speed(iter/s)": 0.032329
    },
    {
      "epoch": 0.6179334399069902,
      "grad_norm": 0.09574276953935623,
      "learning_rate": 0.00010467261777173763,
      "loss": 0.3308386206626892,
      "memory(GiB)": 78.33,
      "step": 3189,
      "token_acc": 0.9015635322616163,
      "train_speed(iter/s)": 0.032329
    },
    {
      "epoch": 0.618127210192317,
      "grad_norm": 0.12292765825986862,
      "learning_rate": 0.00010458098928255584,
      "loss": 0.3934580385684967,
      "memory(GiB)": 78.33,
      "step": 3190,
      "token_acc": 0.8843327802796871,
      "train_speed(iter/s)": 0.03233
    },
    {
      "epoch": 0.6183209804776437,
      "grad_norm": 0.10484007745981216,
      "learning_rate": 0.00010448937944818211,
      "loss": 0.33495286107063293,
      "memory(GiB)": 78.33,
      "step": 3191,
      "token_acc": 0.9003335885376791,
      "train_speed(iter/s)": 0.032331
    },
    {
      "epoch": 0.6185147507629705,
      "grad_norm": 0.09170569479465485,
      "learning_rate": 0.00010439778830624321,
      "loss": 0.32464128732681274,
      "memory(GiB)": 78.33,
      "step": 3192,
      "token_acc": 0.9023982419262373,
      "train_speed(iter/s)": 0.032332
    },
    {
      "epoch": 0.6187085210482972,
      "grad_norm": 0.10959605127573013,
      "learning_rate": 0.00010430621589435801,
      "loss": 0.36965322494506836,
      "memory(GiB)": 78.33,
      "step": 3193,
      "token_acc": 0.8880881729546418,
      "train_speed(iter/s)": 0.032333
    },
    {
      "epoch": 0.618902291333624,
      "grad_norm": 0.10908178985118866,
      "learning_rate": 0.00010421466225013776,
      "loss": 0.39024633169174194,
      "memory(GiB)": 78.33,
      "step": 3194,
      "token_acc": 0.8856543263021954,
      "train_speed(iter/s)": 0.032334
    },
    {
      "epoch": 0.6190960616189507,
      "grad_norm": 0.10703379660844803,
      "learning_rate": 0.0001041231274111861,
      "loss": 0.3596927523612976,
      "memory(GiB)": 78.33,
      "step": 3195,
      "token_acc": 0.8922619047619048,
      "train_speed(iter/s)": 0.032334
    },
    {
      "epoch": 0.6192898319042774,
      "grad_norm": 0.10117658227682114,
      "learning_rate": 0.00010403161141509872,
      "loss": 0.34861043095588684,
      "memory(GiB)": 78.33,
      "step": 3196,
      "token_acc": 0.89493396959272,
      "train_speed(iter/s)": 0.032335
    },
    {
      "epoch": 0.6194836021896042,
      "grad_norm": 0.09983016550540924,
      "learning_rate": 0.0001039401142994638,
      "loss": 0.34936952590942383,
      "memory(GiB)": 78.33,
      "step": 3197,
      "token_acc": 0.8953481619141646,
      "train_speed(iter/s)": 0.032336
    },
    {
      "epoch": 0.6196773724749309,
      "grad_norm": 0.08768882602453232,
      "learning_rate": 0.00010384863610186155,
      "loss": 0.3130777180194855,
      "memory(GiB)": 78.33,
      "step": 3198,
      "token_acc": 0.9038072289156627,
      "train_speed(iter/s)": 0.032337
    },
    {
      "epoch": 0.6198711427602577,
      "grad_norm": 0.09570645540952682,
      "learning_rate": 0.00010375717685986459,
      "loss": 0.31712067127227783,
      "memory(GiB)": 78.33,
      "step": 3199,
      "token_acc": 0.9068876881917851,
      "train_speed(iter/s)": 0.032338
    },
    {
      "epoch": 0.6200649130455844,
      "grad_norm": 0.09237085282802582,
      "learning_rate": 0.0001036657366110377,
      "loss": 0.33153122663497925,
      "memory(GiB)": 78.33,
      "step": 3200,
      "token_acc": 0.9015142538922724,
      "train_speed(iter/s)": 0.032339
    },
    {
      "epoch": 0.6202586833309112,
      "grad_norm": 0.11354105174541473,
      "learning_rate": 0.00010357431539293784,
      "loss": 0.3915308713912964,
      "memory(GiB)": 78.33,
      "step": 3201,
      "token_acc": 0.8843476653936183,
      "train_speed(iter/s)": 0.032336
    },
    {
      "epoch": 0.6204524536162379,
      "grad_norm": 0.09536038339138031,
      "learning_rate": 0.00010348291324311418,
      "loss": 0.3190258741378784,
      "memory(GiB)": 78.33,
      "step": 3202,
      "token_acc": 0.9054259501965924,
      "train_speed(iter/s)": 0.032337
    },
    {
      "epoch": 0.6206462239015647,
      "grad_norm": 0.10331059992313385,
      "learning_rate": 0.00010339153019910797,
      "loss": 0.34562522172927856,
      "memory(GiB)": 78.33,
      "step": 3203,
      "token_acc": 0.8966327964544374,
      "train_speed(iter/s)": 0.032337
    },
    {
      "epoch": 0.6208399941868914,
      "grad_norm": 0.1046903133392334,
      "learning_rate": 0.00010330016629845274,
      "loss": 0.3687852919101715,
      "memory(GiB)": 78.33,
      "step": 3204,
      "token_acc": 0.8896271888020478,
      "train_speed(iter/s)": 0.032338
    },
    {
      "epoch": 0.6210337644722181,
      "grad_norm": 0.12125846743583679,
      "learning_rate": 0.00010320882157867408,
      "loss": 0.32853639125823975,
      "memory(GiB)": 78.33,
      "step": 3205,
      "token_acc": 0.9024209282514908,
      "train_speed(iter/s)": 0.032339
    },
    {
      "epoch": 0.6212275347575449,
      "grad_norm": 0.09862146526575089,
      "learning_rate": 0.00010311749607728976,
      "loss": 0.3248175084590912,
      "memory(GiB)": 78.33,
      "step": 3206,
      "token_acc": 0.9051060112761058,
      "train_speed(iter/s)": 0.03234
    },
    {
      "epoch": 0.6214213050428716,
      "grad_norm": 0.10406461358070374,
      "learning_rate": 0.00010302618983180955,
      "loss": 0.3480234742164612,
      "memory(GiB)": 78.33,
      "step": 3207,
      "token_acc": 0.8964719652212685,
      "train_speed(iter/s)": 0.032341
    },
    {
      "epoch": 0.6216150753281984,
      "grad_norm": 0.0927383154630661,
      "learning_rate": 0.00010293490287973539,
      "loss": 0.31758391857147217,
      "memory(GiB)": 78.33,
      "step": 3208,
      "token_acc": 0.9062631046646439,
      "train_speed(iter/s)": 0.032341
    },
    {
      "epoch": 0.6218088456135251,
      "grad_norm": 0.10273189842700958,
      "learning_rate": 0.00010284363525856138,
      "loss": 0.34117552638053894,
      "memory(GiB)": 78.33,
      "step": 3209,
      "token_acc": 0.897176918619091,
      "train_speed(iter/s)": 0.032342
    },
    {
      "epoch": 0.6220026158988519,
      "grad_norm": 0.10961279273033142,
      "learning_rate": 0.00010275238700577344,
      "loss": 0.3611040711402893,
      "memory(GiB)": 78.33,
      "step": 3210,
      "token_acc": 0.8908705991531813,
      "train_speed(iter/s)": 0.032343
    },
    {
      "epoch": 0.6221963861841786,
      "grad_norm": 0.11966651678085327,
      "learning_rate": 0.00010266115815884978,
      "loss": 0.3134154975414276,
      "memory(GiB)": 78.33,
      "step": 3211,
      "token_acc": 0.9054380748603662,
      "train_speed(iter/s)": 0.032344
    },
    {
      "epoch": 0.6223901564695054,
      "grad_norm": 0.10242405533790588,
      "learning_rate": 0.00010256994875526048,
      "loss": 0.3505587577819824,
      "memory(GiB)": 78.33,
      "step": 3212,
      "token_acc": 0.8964926844453668,
      "train_speed(iter/s)": 0.032345
    },
    {
      "epoch": 0.6225839267548321,
      "grad_norm": 0.09929542243480682,
      "learning_rate": 0.0001024787588324677,
      "loss": 0.3694930672645569,
      "memory(GiB)": 78.33,
      "step": 3213,
      "token_acc": 0.8909321660752674,
      "train_speed(iter/s)": 0.032346
    },
    {
      "epoch": 0.6227776970401588,
      "grad_norm": 0.11470238864421844,
      "learning_rate": 0.00010238758842792571,
      "loss": 0.34870725870132446,
      "memory(GiB)": 78.33,
      "step": 3214,
      "token_acc": 0.8959076482318906,
      "train_speed(iter/s)": 0.032347
    },
    {
      "epoch": 0.6229714673254856,
      "grad_norm": 0.10671926289796829,
      "learning_rate": 0.00010229643757908047,
      "loss": 0.34584519267082214,
      "memory(GiB)": 78.33,
      "step": 3215,
      "token_acc": 0.8956783512316554,
      "train_speed(iter/s)": 0.032347
    },
    {
      "epoch": 0.6231652376108123,
      "grad_norm": 0.10204023122787476,
      "learning_rate": 0.00010220530632337022,
      "loss": 0.33685633540153503,
      "memory(GiB)": 78.33,
      "step": 3216,
      "token_acc": 0.8971209005303604,
      "train_speed(iter/s)": 0.032348
    },
    {
      "epoch": 0.6233590078961391,
      "grad_norm": 0.09859669208526611,
      "learning_rate": 0.0001021141946982249,
      "loss": 0.33895328640937805,
      "memory(GiB)": 78.33,
      "step": 3217,
      "token_acc": 0.8996550051210177,
      "train_speed(iter/s)": 0.032349
    },
    {
      "epoch": 0.6235527781814659,
      "grad_norm": 0.09662448614835739,
      "learning_rate": 0.00010202310274106659,
      "loss": 0.3390568494796753,
      "memory(GiB)": 78.33,
      "step": 3218,
      "token_acc": 0.8996795034847637,
      "train_speed(iter/s)": 0.03235
    },
    {
      "epoch": 0.6237465484667927,
      "grad_norm": 0.12932774424552917,
      "learning_rate": 0.00010193203048930914,
      "loss": 0.3782866597175598,
      "memory(GiB)": 78.33,
      "step": 3219,
      "token_acc": 0.8888921265808031,
      "train_speed(iter/s)": 0.032351
    },
    {
      "epoch": 0.6239403187521194,
      "grad_norm": 0.09401915222406387,
      "learning_rate": 0.0001018409779803584,
      "loss": 0.29952648282051086,
      "memory(GiB)": 78.33,
      "step": 3220,
      "token_acc": 0.9101730793176441,
      "train_speed(iter/s)": 0.032352
    },
    {
      "epoch": 0.6241340890374462,
      "grad_norm": 0.09950324892997742,
      "learning_rate": 0.00010174994525161215,
      "loss": 0.32064443826675415,
      "memory(GiB)": 78.33,
      "step": 3221,
      "token_acc": 0.9026459405633293,
      "train_speed(iter/s)": 0.032352
    },
    {
      "epoch": 0.6243278593227729,
      "grad_norm": 0.11111953109502792,
      "learning_rate": 0.00010165893234045988,
      "loss": 0.3619105815887451,
      "memory(GiB)": 78.33,
      "step": 3222,
      "token_acc": 0.8926043878273178,
      "train_speed(iter/s)": 0.032353
    },
    {
      "epoch": 0.6245216296080996,
      "grad_norm": 0.10427284240722656,
      "learning_rate": 0.0001015679392842831,
      "loss": 0.3563496768474579,
      "memory(GiB)": 78.33,
      "step": 3223,
      "token_acc": 0.8932768896563249,
      "train_speed(iter/s)": 0.032354
    },
    {
      "epoch": 0.6247153998934264,
      "grad_norm": 0.10940185189247131,
      "learning_rate": 0.00010147696612045502,
      "loss": 0.35510045289993286,
      "memory(GiB)": 78.33,
      "step": 3224,
      "token_acc": 0.892183207930252,
      "train_speed(iter/s)": 0.032355
    },
    {
      "epoch": 0.6249091701787531,
      "grad_norm": 0.10383673757314682,
      "learning_rate": 0.00010138601288634085,
      "loss": 0.34664323925971985,
      "memory(GiB)": 78.33,
      "step": 3225,
      "token_acc": 0.897200460505995,
      "train_speed(iter/s)": 0.032356
    },
    {
      "epoch": 0.6251029404640799,
      "grad_norm": 0.1071397215127945,
      "learning_rate": 0.00010129507961929748,
      "loss": 0.3815678358078003,
      "memory(GiB)": 78.33,
      "step": 3226,
      "token_acc": 0.8863624317591562,
      "train_speed(iter/s)": 0.032357
    },
    {
      "epoch": 0.6252967107494066,
      "grad_norm": 0.09263280034065247,
      "learning_rate": 0.00010120416635667364,
      "loss": 0.30259522795677185,
      "memory(GiB)": 78.33,
      "step": 3227,
      "token_acc": 0.9077235976497765,
      "train_speed(iter/s)": 0.032358
    },
    {
      "epoch": 0.6254904810347334,
      "grad_norm": 0.12750935554504395,
      "learning_rate": 0.00010111327313580994,
      "loss": 0.3879462480545044,
      "memory(GiB)": 78.33,
      "step": 3228,
      "token_acc": 0.8852969484232577,
      "train_speed(iter/s)": 0.032359
    },
    {
      "epoch": 0.6256842513200601,
      "grad_norm": 0.09431038796901703,
      "learning_rate": 0.00010102239999403857,
      "loss": 0.3204158842563629,
      "memory(GiB)": 78.33,
      "step": 3229,
      "token_acc": 0.9041399892865342,
      "train_speed(iter/s)": 0.032359
    },
    {
      "epoch": 0.6258780216053869,
      "grad_norm": 0.1006726399064064,
      "learning_rate": 0.00010093154696868362,
      "loss": 0.34341204166412354,
      "memory(GiB)": 78.33,
      "step": 3230,
      "token_acc": 0.8983400146801422,
      "train_speed(iter/s)": 0.03236
    },
    {
      "epoch": 0.6260717918907136,
      "grad_norm": 0.09969964623451233,
      "learning_rate": 0.0001008407140970608,
      "loss": 0.318872332572937,
      "memory(GiB)": 78.33,
      "step": 3231,
      "token_acc": 0.9055216821471309,
      "train_speed(iter/s)": 0.032361
    },
    {
      "epoch": 0.6262655621760403,
      "grad_norm": 0.1142285168170929,
      "learning_rate": 0.00010074990141647767,
      "loss": 0.33352869749069214,
      "memory(GiB)": 78.33,
      "step": 3232,
      "token_acc": 0.9006948304613674,
      "train_speed(iter/s)": 0.032362
    },
    {
      "epoch": 0.6264593324613671,
      "grad_norm": 0.10268845409154892,
      "learning_rate": 0.00010065910896423346,
      "loss": 0.33974435925483704,
      "memory(GiB)": 78.33,
      "step": 3233,
      "token_acc": 0.8981594097137761,
      "train_speed(iter/s)": 0.032363
    },
    {
      "epoch": 0.6266531027466938,
      "grad_norm": 0.11217531561851501,
      "learning_rate": 0.000100568336777619,
      "loss": 0.40757495164871216,
      "memory(GiB)": 78.33,
      "step": 3234,
      "token_acc": 0.8796624013068336,
      "train_speed(iter/s)": 0.032364
    },
    {
      "epoch": 0.6268468730320206,
      "grad_norm": 0.11023017019033432,
      "learning_rate": 0.00010047758489391698,
      "loss": 0.37037602066993713,
      "memory(GiB)": 78.33,
      "step": 3235,
      "token_acc": 0.8889917278880083,
      "train_speed(iter/s)": 0.032365
    },
    {
      "epoch": 0.6270406433173473,
      "grad_norm": 0.09839659184217453,
      "learning_rate": 0.00010038685335040149,
      "loss": 0.330047607421875,
      "memory(GiB)": 78.33,
      "step": 3236,
      "token_acc": 0.9030212027856163,
      "train_speed(iter/s)": 0.032365
    },
    {
      "epoch": 0.6272344136026741,
      "grad_norm": 0.10475650429725647,
      "learning_rate": 0.00010029614218433851,
      "loss": 0.3539055585861206,
      "memory(GiB)": 78.33,
      "step": 3237,
      "token_acc": 0.8929911208269236,
      "train_speed(iter/s)": 0.032366
    },
    {
      "epoch": 0.6274281838880008,
      "grad_norm": 0.11595990508794785,
      "learning_rate": 0.00010020545143298555,
      "loss": 0.3532840311527252,
      "memory(GiB)": 78.33,
      "step": 3238,
      "token_acc": 0.8959765685216587,
      "train_speed(iter/s)": 0.032367
    },
    {
      "epoch": 0.6276219541733276,
      "grad_norm": 0.0995573177933693,
      "learning_rate": 0.0001001147811335917,
      "loss": 0.3350241482257843,
      "memory(GiB)": 78.33,
      "step": 3239,
      "token_acc": 0.900135550084373,
      "train_speed(iter/s)": 0.032368
    },
    {
      "epoch": 0.6278157244586543,
      "grad_norm": 0.10098245739936829,
      "learning_rate": 0.0001000241313233977,
      "loss": 0.3346031606197357,
      "memory(GiB)": 78.33,
      "step": 3240,
      "token_acc": 0.9002930289944479,
      "train_speed(iter/s)": 0.032369
    },
    {
      "epoch": 0.628009494743981,
      "grad_norm": 0.09921301156282425,
      "learning_rate": 9.993350203963586e-05,
      "loss": 0.329167902469635,
      "memory(GiB)": 78.33,
      "step": 3241,
      "token_acc": 0.9022172464638741,
      "train_speed(iter/s)": 0.03237
    },
    {
      "epoch": 0.6282032650293078,
      "grad_norm": 0.09442011266946793,
      "learning_rate": 9.984289331953012e-05,
      "loss": 0.32895606756210327,
      "memory(GiB)": 78.33,
      "step": 3242,
      "token_acc": 0.9014620960805996,
      "train_speed(iter/s)": 0.03237
    },
    {
      "epoch": 0.6283970353146345,
      "grad_norm": 0.09743113070726395,
      "learning_rate": 9.975230520029581e-05,
      "loss": 0.33065280318260193,
      "memory(GiB)": 78.33,
      "step": 3243,
      "token_acc": 0.9017754207977865,
      "train_speed(iter/s)": 0.032371
    },
    {
      "epoch": 0.6285908055999613,
      "grad_norm": 0.10759622603654861,
      "learning_rate": 9.966173771913999e-05,
      "loss": 0.3607182502746582,
      "memory(GiB)": 78.33,
      "step": 3244,
      "token_acc": 0.8951261723983921,
      "train_speed(iter/s)": 0.032372
    },
    {
      "epoch": 0.628784575885288,
      "grad_norm": 0.10882527381181717,
      "learning_rate": 9.957119091326111e-05,
      "loss": 0.3616001307964325,
      "memory(GiB)": 78.33,
      "step": 3245,
      "token_acc": 0.8927234113899291,
      "train_speed(iter/s)": 0.032373
    },
    {
      "epoch": 0.6289783461706148,
      "grad_norm": 0.10181569308042526,
      "learning_rate": 9.948066481984919e-05,
      "loss": 0.32742297649383545,
      "memory(GiB)": 78.33,
      "step": 3246,
      "token_acc": 0.9032007815193656,
      "train_speed(iter/s)": 0.032374
    },
    {
      "epoch": 0.6291721164559415,
      "grad_norm": 0.09965640306472778,
      "learning_rate": 9.939015947608579e-05,
      "loss": 0.32857993245124817,
      "memory(GiB)": 78.33,
      "step": 3247,
      "token_acc": 0.901941986907338,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.6293658867412683,
      "grad_norm": 0.11179827153682709,
      "learning_rate": 9.929967491914378e-05,
      "loss": 0.3481506407260895,
      "memory(GiB)": 78.33,
      "step": 3248,
      "token_acc": 0.8953111091968875,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.629559657026595,
      "grad_norm": 0.10431916266679764,
      "learning_rate": 9.920921118618772e-05,
      "loss": 0.36535415053367615,
      "memory(GiB)": 78.33,
      "step": 3249,
      "token_acc": 0.8926919518963923,
      "train_speed(iter/s)": 0.032376
    },
    {
      "epoch": 0.6297534273119217,
      "grad_norm": 0.11817894130945206,
      "learning_rate": 9.911876831437334e-05,
      "loss": 0.4042690098285675,
      "memory(GiB)": 78.33,
      "step": 3250,
      "token_acc": 0.8816427447580516,
      "train_speed(iter/s)": 0.032377
    },
    {
      "epoch": 0.6299471975972485,
      "grad_norm": 0.10380962491035461,
      "learning_rate": 9.902834634084809e-05,
      "loss": 0.33863088488578796,
      "memory(GiB)": 78.33,
      "step": 3251,
      "token_acc": 0.8970420766782492,
      "train_speed(iter/s)": 0.032378
    },
    {
      "epoch": 0.6301409678825752,
      "grad_norm": 0.10345399379730225,
      "learning_rate": 9.893794530275065e-05,
      "loss": 0.33550775051116943,
      "memory(GiB)": 78.33,
      "step": 3252,
      "token_acc": 0.8967549151027703,
      "train_speed(iter/s)": 0.032379
    },
    {
      "epoch": 0.630334738167902,
      "grad_norm": 0.10701797902584076,
      "learning_rate": 9.884756523721115e-05,
      "loss": 0.33787888288497925,
      "memory(GiB)": 78.33,
      "step": 3253,
      "token_acc": 0.8981328701693443,
      "train_speed(iter/s)": 0.03238
    },
    {
      "epoch": 0.6305285084532287,
      "grad_norm": 0.11159101128578186,
      "learning_rate": 9.875720618135118e-05,
      "loss": 0.36240842938423157,
      "memory(GiB)": 78.33,
      "step": 3254,
      "token_acc": 0.8929006465315301,
      "train_speed(iter/s)": 0.032381
    },
    {
      "epoch": 0.6307222787385555,
      "grad_norm": 0.10191599279642105,
      "learning_rate": 9.866686817228351e-05,
      "loss": 0.35280781984329224,
      "memory(GiB)": 78.33,
      "step": 3255,
      "token_acc": 0.8953659778617565,
      "train_speed(iter/s)": 0.032381
    },
    {
      "epoch": 0.6309160490238822,
      "grad_norm": 0.10553169250488281,
      "learning_rate": 9.857655124711252e-05,
      "loss": 0.3603316843509674,
      "memory(GiB)": 78.33,
      "step": 3256,
      "token_acc": 0.8932977749043693,
      "train_speed(iter/s)": 0.032382
    },
    {
      "epoch": 0.631109819309209,
      "grad_norm": 0.12384941428899765,
      "learning_rate": 9.84862554429337e-05,
      "loss": 0.33838897943496704,
      "memory(GiB)": 78.33,
      "step": 3257,
      "token_acc": 0.8999357294705097,
      "train_speed(iter/s)": 0.032383
    },
    {
      "epoch": 0.6313035895945357,
      "grad_norm": 0.09418811649084091,
      "learning_rate": 9.839598079683399e-05,
      "loss": 0.3201935887336731,
      "memory(GiB)": 78.33,
      "step": 3258,
      "token_acc": 0.9044349822983446,
      "train_speed(iter/s)": 0.032384
    },
    {
      "epoch": 0.6314973598798624,
      "grad_norm": 0.08840276300907135,
      "learning_rate": 9.830572734589162e-05,
      "loss": 0.295772910118103,
      "memory(GiB)": 78.33,
      "step": 3259,
      "token_acc": 0.9113684822640047,
      "train_speed(iter/s)": 0.032385
    },
    {
      "epoch": 0.6316911301651892,
      "grad_norm": 0.10140841454267502,
      "learning_rate": 9.82154951271761e-05,
      "loss": 0.3301541805267334,
      "memory(GiB)": 78.33,
      "step": 3260,
      "token_acc": 0.9020520414639306,
      "train_speed(iter/s)": 0.032385
    },
    {
      "epoch": 0.6318849004505159,
      "grad_norm": 0.11607446521520615,
      "learning_rate": 9.81252841777483e-05,
      "loss": 0.376188725233078,
      "memory(GiB)": 78.33,
      "step": 3261,
      "token_acc": 0.8894073426006934,
      "train_speed(iter/s)": 0.032386
    },
    {
      "epoch": 0.6320786707358427,
      "grad_norm": 0.09963465481996536,
      "learning_rate": 9.803509453466015e-05,
      "loss": 0.3333686590194702,
      "memory(GiB)": 78.33,
      "step": 3262,
      "token_acc": 0.8974518502488639,
      "train_speed(iter/s)": 0.032387
    },
    {
      "epoch": 0.6322724410211694,
      "grad_norm": 0.0981069952249527,
      "learning_rate": 9.794492623495509e-05,
      "loss": 0.3355132043361664,
      "memory(GiB)": 78.33,
      "step": 3263,
      "token_acc": 0.8998677114073471,
      "train_speed(iter/s)": 0.032388
    },
    {
      "epoch": 0.6324662113064962,
      "grad_norm": 0.11496897041797638,
      "learning_rate": 9.785477931566753e-05,
      "loss": 0.3889937400817871,
      "memory(GiB)": 78.33,
      "step": 3264,
      "token_acc": 0.884316427783903,
      "train_speed(iter/s)": 0.032389
    },
    {
      "epoch": 0.6326599815918229,
      "grad_norm": 0.10167025029659271,
      "learning_rate": 9.77646538138233e-05,
      "loss": 0.33570176362991333,
      "memory(GiB)": 78.33,
      "step": 3265,
      "token_acc": 0.8996981339187706,
      "train_speed(iter/s)": 0.03239
    },
    {
      "epoch": 0.6328537518771497,
      "grad_norm": 0.1109309270977974,
      "learning_rate": 9.767454976643939e-05,
      "loss": 0.3936588168144226,
      "memory(GiB)": 78.33,
      "step": 3266,
      "token_acc": 0.8842383328441444,
      "train_speed(iter/s)": 0.032391
    },
    {
      "epoch": 0.6330475221624764,
      "grad_norm": 0.1110692024230957,
      "learning_rate": 9.758446721052394e-05,
      "loss": 0.35908806324005127,
      "memory(GiB)": 78.33,
      "step": 3267,
      "token_acc": 0.8943084600760456,
      "train_speed(iter/s)": 0.032391
    },
    {
      "epoch": 0.6332412924478031,
      "grad_norm": 0.10611660033464432,
      "learning_rate": 9.749440618307628e-05,
      "loss": 0.3543277680873871,
      "memory(GiB)": 78.33,
      "step": 3268,
      "token_acc": 0.8957095521023766,
      "train_speed(iter/s)": 0.032392
    },
    {
      "epoch": 0.6334350627331299,
      "grad_norm": 0.0958552435040474,
      "learning_rate": 9.740436672108685e-05,
      "loss": 0.3312506675720215,
      "memory(GiB)": 78.33,
      "step": 3269,
      "token_acc": 0.901442661563374,
      "train_speed(iter/s)": 0.032393
    },
    {
      "epoch": 0.6336288330184566,
      "grad_norm": 0.1073455959558487,
      "learning_rate": 9.731434886153735e-05,
      "loss": 0.37120985984802246,
      "memory(GiB)": 78.33,
      "step": 3270,
      "token_acc": 0.8908422837761447,
      "train_speed(iter/s)": 0.032394
    },
    {
      "epoch": 0.6338226033037834,
      "grad_norm": 0.10486573725938797,
      "learning_rate": 9.722435264140043e-05,
      "loss": 0.34987497329711914,
      "memory(GiB)": 78.33,
      "step": 3271,
      "token_acc": 0.8957631675770527,
      "train_speed(iter/s)": 0.032395
    },
    {
      "epoch": 0.6340163735891101,
      "grad_norm": 0.0972420871257782,
      "learning_rate": 9.713437809764002e-05,
      "loss": 0.3343609869480133,
      "memory(GiB)": 78.33,
      "step": 3272,
      "token_acc": 0.9014839885446498,
      "train_speed(iter/s)": 0.032395
    },
    {
      "epoch": 0.6342101438744369,
      "grad_norm": 0.09882809221744537,
      "learning_rate": 9.704442526721112e-05,
      "loss": 0.31326884031295776,
      "memory(GiB)": 78.33,
      "step": 3273,
      "token_acc": 0.9058581477465455,
      "train_speed(iter/s)": 0.032396
    },
    {
      "epoch": 0.6344039141597636,
      "grad_norm": 0.15750029683113098,
      "learning_rate": 9.69544941870597e-05,
      "loss": 0.32955771684646606,
      "memory(GiB)": 78.33,
      "step": 3274,
      "token_acc": 0.9013154082401345,
      "train_speed(iter/s)": 0.032397
    },
    {
      "epoch": 0.6345976844450903,
      "grad_norm": 0.10274738818407059,
      "learning_rate": 9.686458489412296e-05,
      "loss": 0.37514811754226685,
      "memory(GiB)": 78.33,
      "step": 3275,
      "token_acc": 0.8886818439601394,
      "train_speed(iter/s)": 0.032398
    },
    {
      "epoch": 0.6347914547304171,
      "grad_norm": 0.10508158802986145,
      "learning_rate": 9.677469742532896e-05,
      "loss": 0.34448888897895813,
      "memory(GiB)": 78.33,
      "step": 3276,
      "token_acc": 0.8970773975620502,
      "train_speed(iter/s)": 0.032399
    },
    {
      "epoch": 0.6349852250157438,
      "grad_norm": 0.1056324765086174,
      "learning_rate": 9.668483181759696e-05,
      "loss": 0.35540270805358887,
      "memory(GiB)": 78.33,
      "step": 3277,
      "token_acc": 0.8941543700340522,
      "train_speed(iter/s)": 0.032399
    },
    {
      "epoch": 0.6351789953010706,
      "grad_norm": 0.1182514950633049,
      "learning_rate": 9.659498810783716e-05,
      "loss": 0.38734346628189087,
      "memory(GiB)": 78.33,
      "step": 3278,
      "token_acc": 0.8861613506065693,
      "train_speed(iter/s)": 0.0324
    },
    {
      "epoch": 0.6353727655863973,
      "grad_norm": 0.09534314274787903,
      "learning_rate": 9.65051663329508e-05,
      "loss": 0.3013858199119568,
      "memory(GiB)": 78.33,
      "step": 3279,
      "token_acc": 0.9102011699243829,
      "train_speed(iter/s)": 0.032401
    },
    {
      "epoch": 0.6355665358717241,
      "grad_norm": 0.11075269430875778,
      "learning_rate": 9.641536652983008e-05,
      "loss": 0.3610069453716278,
      "memory(GiB)": 78.33,
      "step": 3280,
      "token_acc": 0.8919886003799873,
      "train_speed(iter/s)": 0.032402
    },
    {
      "epoch": 0.6357603061570508,
      "grad_norm": 0.1012001633644104,
      "learning_rate": 9.63255887353582e-05,
      "loss": 0.32832014560699463,
      "memory(GiB)": 78.33,
      "step": 3281,
      "token_acc": 0.9021164021164021,
      "train_speed(iter/s)": 0.032403
    },
    {
      "epoch": 0.6359540764423776,
      "grad_norm": 0.09684737026691437,
      "learning_rate": 9.623583298640937e-05,
      "loss": 0.35295921564102173,
      "memory(GiB)": 78.33,
      "step": 3282,
      "token_acc": 0.8938139656941098,
      "train_speed(iter/s)": 0.032404
    },
    {
      "epoch": 0.6361478467277043,
      "grad_norm": 0.12095388025045395,
      "learning_rate": 9.614609931984854e-05,
      "loss": 0.3737364709377289,
      "memory(GiB)": 78.33,
      "step": 3283,
      "token_acc": 0.8893832689984383,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.636341617013031,
      "grad_norm": 0.10381918400526047,
      "learning_rate": 9.60563877725319e-05,
      "loss": 0.3494938313961029,
      "memory(GiB)": 78.33,
      "step": 3284,
      "token_acc": 0.8952139037433156,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.6365353872983578,
      "grad_norm": 0.10965927690267563,
      "learning_rate": 9.596669838130627e-05,
      "loss": 0.36877620220184326,
      "memory(GiB)": 78.33,
      "step": 3285,
      "token_acc": 0.8932712103153104,
      "train_speed(iter/s)": 0.032406
    },
    {
      "epoch": 0.6367291575836845,
      "grad_norm": 0.09375525265932083,
      "learning_rate": 9.587703118300955e-05,
      "loss": 0.33250147104263306,
      "memory(GiB)": 78.33,
      "step": 3286,
      "token_acc": 0.9001272429821601,
      "train_speed(iter/s)": 0.032407
    },
    {
      "epoch": 0.6369229278690113,
      "grad_norm": 0.11146512627601624,
      "learning_rate": 9.578738621447052e-05,
      "loss": 0.3969725966453552,
      "memory(GiB)": 78.33,
      "step": 3287,
      "token_acc": 0.8843091971199565,
      "train_speed(iter/s)": 0.032408
    },
    {
      "epoch": 0.637116698154338,
      "grad_norm": 0.10300832241773605,
      "learning_rate": 9.569776351250867e-05,
      "loss": 0.3498913645744324,
      "memory(GiB)": 78.33,
      "step": 3288,
      "token_acc": 0.896644596566042,
      "train_speed(iter/s)": 0.032409
    },
    {
      "epoch": 0.6373104684396648,
      "grad_norm": 0.10484420508146286,
      "learning_rate": 9.560816311393456e-05,
      "loss": 0.3171748220920563,
      "memory(GiB)": 78.33,
      "step": 3289,
      "token_acc": 0.9038370440549502,
      "train_speed(iter/s)": 0.032409
    },
    {
      "epoch": 0.6375042387249915,
      "grad_norm": 0.11446377635002136,
      "learning_rate": 9.551858505554935e-05,
      "loss": 0.3828757703304291,
      "memory(GiB)": 78.33,
      "step": 3290,
      "token_acc": 0.8870528771384136,
      "train_speed(iter/s)": 0.03241
    },
    {
      "epoch": 0.6376980090103183,
      "grad_norm": 0.10568659007549286,
      "learning_rate": 9.542902937414528e-05,
      "loss": 0.3619765043258667,
      "memory(GiB)": 78.33,
      "step": 3291,
      "token_acc": 0.8938068279293189,
      "train_speed(iter/s)": 0.032411
    },
    {
      "epoch": 0.637891779295645,
      "grad_norm": 0.09850284457206726,
      "learning_rate": 9.533949610650519e-05,
      "loss": 0.3398081660270691,
      "memory(GiB)": 78.33,
      "step": 3292,
      "token_acc": 0.8959392848759878,
      "train_speed(iter/s)": 0.032412
    },
    {
      "epoch": 0.6380855495809717,
      "grad_norm": 0.10148289799690247,
      "learning_rate": 9.524998528940282e-05,
      "loss": 0.3508596420288086,
      "memory(GiB)": 78.33,
      "step": 3293,
      "token_acc": 0.8957725382847403,
      "train_speed(iter/s)": 0.032413
    },
    {
      "epoch": 0.6382793198662985,
      "grad_norm": 0.09729644656181335,
      "learning_rate": 9.516049695960277e-05,
      "loss": 0.33748623728752136,
      "memory(GiB)": 78.33,
      "step": 3294,
      "token_acc": 0.8995384056843525,
      "train_speed(iter/s)": 0.032414
    },
    {
      "epoch": 0.6384730901516252,
      "grad_norm": 0.11033184826374054,
      "learning_rate": 9.507103115386013e-05,
      "loss": 0.36528030037879944,
      "memory(GiB)": 78.33,
      "step": 3295,
      "token_acc": 0.8912418790604698,
      "train_speed(iter/s)": 0.032415
    },
    {
      "epoch": 0.638666860436952,
      "grad_norm": 0.09952208399772644,
      "learning_rate": 9.498158790892105e-05,
      "loss": 0.3313583433628082,
      "memory(GiB)": 78.33,
      "step": 3296,
      "token_acc": 0.9018691588785047,
      "train_speed(iter/s)": 0.032415
    },
    {
      "epoch": 0.6388606307222787,
      "grad_norm": 0.11369100958108902,
      "learning_rate": 9.489216726152218e-05,
      "loss": 0.3779648542404175,
      "memory(GiB)": 78.33,
      "step": 3297,
      "token_acc": 0.8898235347901619,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.6390544010076055,
      "grad_norm": 0.10706917941570282,
      "learning_rate": 9.480276924839101e-05,
      "loss": 0.3606652319431305,
      "memory(GiB)": 78.33,
      "step": 3298,
      "token_acc": 0.8924252794889346,
      "train_speed(iter/s)": 0.032417
    },
    {
      "epoch": 0.6392481712929322,
      "grad_norm": 0.107958123087883,
      "learning_rate": 9.471339390624573e-05,
      "loss": 0.36460253596305847,
      "memory(GiB)": 78.33,
      "step": 3299,
      "token_acc": 0.8916521320743369,
      "train_speed(iter/s)": 0.032418
    },
    {
      "epoch": 0.639441941578259,
      "grad_norm": 0.10166801512241364,
      "learning_rate": 9.462404127179517e-05,
      "loss": 0.34821823239326477,
      "memory(GiB)": 78.33,
      "step": 3300,
      "token_acc": 0.8970175808932179,
      "train_speed(iter/s)": 0.032418
    },
    {
      "epoch": 0.6396357118635857,
      "grad_norm": 0.11107532680034637,
      "learning_rate": 9.453471138173893e-05,
      "loss": 0.3458663523197174,
      "memory(GiB)": 78.33,
      "step": 3301,
      "token_acc": 0.8959231235833094,
      "train_speed(iter/s)": 0.032419
    },
    {
      "epoch": 0.6398294821489124,
      "grad_norm": 0.11017818003892899,
      "learning_rate": 9.444540427276707e-05,
      "loss": 0.36541783809661865,
      "memory(GiB)": 78.33,
      "step": 3302,
      "token_acc": 0.8919458960959115,
      "train_speed(iter/s)": 0.03242
    },
    {
      "epoch": 0.6400232524342392,
      "grad_norm": 0.09530369192361832,
      "learning_rate": 9.435611998156055e-05,
      "loss": 0.31924229860305786,
      "memory(GiB)": 78.33,
      "step": 3303,
      "token_acc": 0.9019935658738,
      "train_speed(iter/s)": 0.032421
    },
    {
      "epoch": 0.6402170227195659,
      "grad_norm": 0.11861889809370041,
      "learning_rate": 9.42668585447907e-05,
      "loss": 0.3566751480102539,
      "memory(GiB)": 78.33,
      "step": 3304,
      "token_acc": 0.8942750741652464,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.6404107930048927,
      "grad_norm": 0.09427324682474136,
      "learning_rate": 9.417761999911965e-05,
      "loss": 0.35689249634742737,
      "memory(GiB)": 78.33,
      "step": 3305,
      "token_acc": 0.8942868072275602,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.6406045632902194,
      "grad_norm": 0.09987188875675201,
      "learning_rate": 9.40884043812001e-05,
      "loss": 0.3596974313259125,
      "memory(GiB)": 78.33,
      "step": 3306,
      "token_acc": 0.8958737616178123,
      "train_speed(iter/s)": 0.032423
    },
    {
      "epoch": 0.6407983335755462,
      "grad_norm": 0.0936884880065918,
      "learning_rate": 9.399921172767525e-05,
      "loss": 0.32182028889656067,
      "memory(GiB)": 78.33,
      "step": 3307,
      "token_acc": 0.9035014440174715,
      "train_speed(iter/s)": 0.032424
    },
    {
      "epoch": 0.6409921038608729,
      "grad_norm": 0.10136570781469345,
      "learning_rate": 9.3910042075179e-05,
      "loss": 0.34020885825157166,
      "memory(GiB)": 78.33,
      "step": 3308,
      "token_acc": 0.8990686593025774,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.6411858741461997,
      "grad_norm": 0.09547661989927292,
      "learning_rate": 9.38208954603356e-05,
      "loss": 0.320523738861084,
      "memory(GiB)": 78.33,
      "step": 3309,
      "token_acc": 0.9048575160868935,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.6413796444315264,
      "grad_norm": 0.09538991749286652,
      "learning_rate": 9.373177191976007e-05,
      "loss": 0.3119708299636841,
      "memory(GiB)": 78.33,
      "step": 3310,
      "token_acc": 0.906541066892464,
      "train_speed(iter/s)": 0.032426
    },
    {
      "epoch": 0.6415734147168531,
      "grad_norm": 0.09219719469547272,
      "learning_rate": 9.36426714900578e-05,
      "loss": 0.3281041383743286,
      "memory(GiB)": 78.33,
      "step": 3311,
      "token_acc": 0.9008394285994796,
      "train_speed(iter/s)": 0.032427
    },
    {
      "epoch": 0.6417671850021799,
      "grad_norm": 0.09361285716295242,
      "learning_rate": 9.355359420782467e-05,
      "loss": 0.3183348476886749,
      "memory(GiB)": 78.33,
      "step": 3312,
      "token_acc": 0.9031101372046116,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.6419609552875066,
      "grad_norm": 0.10026978701353073,
      "learning_rate": 9.346454010964722e-05,
      "loss": 0.3270958364009857,
      "memory(GiB)": 78.33,
      "step": 3313,
      "token_acc": 0.90237444958475,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.6421547255728334,
      "grad_norm": 0.10751143842935562,
      "learning_rate": 9.337550923210228e-05,
      "loss": 0.3549875020980835,
      "memory(GiB)": 78.33,
      "step": 3314,
      "token_acc": 0.8923901567449037,
      "train_speed(iter/s)": 0.032429
    },
    {
      "epoch": 0.6423484958581601,
      "grad_norm": 0.10847034305334091,
      "learning_rate": 9.328650161175735e-05,
      "loss": 0.3592216372489929,
      "memory(GiB)": 78.33,
      "step": 3315,
      "token_acc": 0.8951582706875199,
      "train_speed(iter/s)": 0.03243
    },
    {
      "epoch": 0.6425422661434869,
      "grad_norm": 0.09726880490779877,
      "learning_rate": 9.319751728517007e-05,
      "loss": 0.31801801919937134,
      "memory(GiB)": 78.33,
      "step": 3316,
      "token_acc": 0.9054611953118507,
      "train_speed(iter/s)": 0.032431
    },
    {
      "epoch": 0.6427360364288136,
      "grad_norm": 0.10297710448503494,
      "learning_rate": 9.31085562888888e-05,
      "loss": 0.3436722159385681,
      "memory(GiB)": 78.33,
      "step": 3317,
      "token_acc": 0.897666406481307,
      "train_speed(iter/s)": 0.032432
    },
    {
      "epoch": 0.6429298067141404,
      "grad_norm": 0.1005515605211258,
      "learning_rate": 9.301961865945217e-05,
      "loss": 0.3575522005558014,
      "memory(GiB)": 78.33,
      "step": 3318,
      "token_acc": 0.8908450704225352,
      "train_speed(iter/s)": 0.032433
    },
    {
      "epoch": 0.6431235769994671,
      "grad_norm": 0.11259230971336365,
      "learning_rate": 9.29307044333893e-05,
      "loss": 0.3851853311061859,
      "memory(GiB)": 78.33,
      "step": 3319,
      "token_acc": 0.8852057033875216,
      "train_speed(iter/s)": 0.032433
    },
    {
      "epoch": 0.6433173472847938,
      "grad_norm": 0.1026308611035347,
      "learning_rate": 9.284181364721967e-05,
      "loss": 0.33121898770332336,
      "memory(GiB)": 78.33,
      "step": 3320,
      "token_acc": 0.9000698616738857,
      "train_speed(iter/s)": 0.032434
    },
    {
      "epoch": 0.6435111175701206,
      "grad_norm": 0.10448215901851654,
      "learning_rate": 9.275294633745302e-05,
      "loss": 0.35694530606269836,
      "memory(GiB)": 78.33,
      "step": 3321,
      "token_acc": 0.8939416251877595,
      "train_speed(iter/s)": 0.032435
    },
    {
      "epoch": 0.6437048878554473,
      "grad_norm": 0.09308988600969315,
      "learning_rate": 9.266410254058966e-05,
      "loss": 0.3261651396751404,
      "memory(GiB)": 78.33,
      "step": 3322,
      "token_acc": 0.9015043921939929,
      "train_speed(iter/s)": 0.032436
    },
    {
      "epoch": 0.6438986581407741,
      "grad_norm": 0.10853360593318939,
      "learning_rate": 9.257528229312e-05,
      "loss": 0.36176615953445435,
      "memory(GiB)": 78.33,
      "step": 3323,
      "token_acc": 0.8933630754077465,
      "train_speed(iter/s)": 0.032437
    },
    {
      "epoch": 0.6440924284261008,
      "grad_norm": 0.10240910202264786,
      "learning_rate": 9.248648563152499e-05,
      "loss": 0.3467825651168823,
      "memory(GiB)": 78.33,
      "step": 3324,
      "token_acc": 0.8980282133305365,
      "train_speed(iter/s)": 0.032437
    },
    {
      "epoch": 0.6442861987114276,
      "grad_norm": 0.10371696203947067,
      "learning_rate": 9.239771259227577e-05,
      "loss": 0.35027819871902466,
      "memory(GiB)": 78.33,
      "step": 3325,
      "token_acc": 0.8956741618858588,
      "train_speed(iter/s)": 0.032438
    },
    {
      "epoch": 0.6444799689967543,
      "grad_norm": 0.11719199270009995,
      "learning_rate": 9.230896321183379e-05,
      "loss": 0.3709411919116974,
      "memory(GiB)": 78.33,
      "step": 3326,
      "token_acc": 0.8905724445013935,
      "train_speed(iter/s)": 0.032439
    },
    {
      "epoch": 0.644673739282081,
      "grad_norm": 0.09655088186264038,
      "learning_rate": 9.222023752665094e-05,
      "loss": 0.35260531306266785,
      "memory(GiB)": 78.33,
      "step": 3327,
      "token_acc": 0.8940183178216111,
      "train_speed(iter/s)": 0.03244
    },
    {
      "epoch": 0.6448675095674078,
      "grad_norm": 0.10662583261728287,
      "learning_rate": 9.213153557316904e-05,
      "loss": 0.3466799259185791,
      "memory(GiB)": 78.33,
      "step": 3328,
      "token_acc": 0.8973979206598129,
      "train_speed(iter/s)": 0.03244
    },
    {
      "epoch": 0.6450612798527345,
      "grad_norm": 0.1059870645403862,
      "learning_rate": 9.204285738782053e-05,
      "loss": 0.35201188921928406,
      "memory(GiB)": 78.33,
      "step": 3329,
      "token_acc": 0.8970674486803519,
      "train_speed(iter/s)": 0.032441
    },
    {
      "epoch": 0.6452550501380613,
      "grad_norm": 0.09518367052078247,
      "learning_rate": 9.195420300702782e-05,
      "loss": 0.3428179919719696,
      "memory(GiB)": 78.33,
      "step": 3330,
      "token_acc": 0.894964203373377,
      "train_speed(iter/s)": 0.032442
    },
    {
      "epoch": 0.645448820423388,
      "grad_norm": 0.10190161317586899,
      "learning_rate": 9.18655724672037e-05,
      "loss": 0.34315353631973267,
      "memory(GiB)": 78.33,
      "step": 3331,
      "token_acc": 0.8971641880907941,
      "train_speed(iter/s)": 0.032443
    },
    {
      "epoch": 0.6456425907087148,
      "grad_norm": 0.09689725935459137,
      "learning_rate": 9.177696580475109e-05,
      "loss": 0.32769232988357544,
      "memory(GiB)": 78.33,
      "step": 3332,
      "token_acc": 0.9002272727272728,
      "train_speed(iter/s)": 0.032444
    },
    {
      "epoch": 0.6458363609940415,
      "grad_norm": 0.09515693038702011,
      "learning_rate": 9.168838305606311e-05,
      "loss": 0.2983517050743103,
      "memory(GiB)": 78.33,
      "step": 3333,
      "token_acc": 0.9090839107005388,
      "train_speed(iter/s)": 0.032444
    },
    {
      "epoch": 0.6460301312793683,
      "grad_norm": 0.11088063567876816,
      "learning_rate": 9.159982425752319e-05,
      "loss": 0.38608160614967346,
      "memory(GiB)": 78.33,
      "step": 3334,
      "token_acc": 0.8878626217586326,
      "train_speed(iter/s)": 0.032445
    },
    {
      "epoch": 0.646223901564695,
      "grad_norm": 0.09727376699447632,
      "learning_rate": 9.151128944550465e-05,
      "loss": 0.3358725905418396,
      "memory(GiB)": 78.33,
      "step": 3335,
      "token_acc": 0.9006044153050249,
      "train_speed(iter/s)": 0.032446
    },
    {
      "epoch": 0.6464176718500217,
      "grad_norm": 0.1008610799908638,
      "learning_rate": 9.142277865637124e-05,
      "loss": 0.350691556930542,
      "memory(GiB)": 78.33,
      "step": 3336,
      "token_acc": 0.8959597901617732,
      "train_speed(iter/s)": 0.032447
    },
    {
      "epoch": 0.6466114421353485,
      "grad_norm": 0.10170338302850723,
      "learning_rate": 9.133429192647661e-05,
      "loss": 0.3640455901622772,
      "memory(GiB)": 78.33,
      "step": 3337,
      "token_acc": 0.8921849481941582,
      "train_speed(iter/s)": 0.032447
    },
    {
      "epoch": 0.6468052124206752,
      "grad_norm": 0.10477067530155182,
      "learning_rate": 9.124582929216471e-05,
      "loss": 0.3344863951206207,
      "memory(GiB)": 78.33,
      "step": 3338,
      "token_acc": 0.8980883180952892,
      "train_speed(iter/s)": 0.032448
    },
    {
      "epoch": 0.646998982706002,
      "grad_norm": 0.10310588032007217,
      "learning_rate": 9.11573907897695e-05,
      "loss": 0.35701289772987366,
      "memory(GiB)": 78.33,
      "step": 3339,
      "token_acc": 0.895986649522611,
      "train_speed(iter/s)": 0.032449
    },
    {
      "epoch": 0.6471927529913288,
      "grad_norm": 0.0928465947508812,
      "learning_rate": 9.106897645561506e-05,
      "loss": 0.3364104628562927,
      "memory(GiB)": 78.33,
      "step": 3340,
      "token_acc": 0.8999741468459153,
      "train_speed(iter/s)": 0.03245
    },
    {
      "epoch": 0.6473865232766556,
      "grad_norm": 0.09853307157754898,
      "learning_rate": 9.098058632601557e-05,
      "loss": 0.33070456981658936,
      "memory(GiB)": 78.33,
      "step": 3341,
      "token_acc": 0.900615836532763,
      "train_speed(iter/s)": 0.03245
    },
    {
      "epoch": 0.6475802935619823,
      "grad_norm": 0.10872960090637207,
      "learning_rate": 9.089222043727512e-05,
      "loss": 0.38177114725112915,
      "memory(GiB)": 78.33,
      "step": 3342,
      "token_acc": 0.8859876130382949,
      "train_speed(iter/s)": 0.032451
    },
    {
      "epoch": 0.6477740638473091,
      "grad_norm": 0.10483535379171371,
      "learning_rate": 9.08038788256881e-05,
      "loss": 0.3313630223274231,
      "memory(GiB)": 78.33,
      "step": 3343,
      "token_acc": 0.9004806641905178,
      "train_speed(iter/s)": 0.032452
    },
    {
      "epoch": 0.6479678341326358,
      "grad_norm": 0.09030815213918686,
      "learning_rate": 9.071556152753866e-05,
      "loss": 0.3115072548389435,
      "memory(GiB)": 78.33,
      "step": 3344,
      "token_acc": 0.9051830718021874,
      "train_speed(iter/s)": 0.032453
    },
    {
      "epoch": 0.6481616044179626,
      "grad_norm": 0.1190243512392044,
      "learning_rate": 9.062726857910111e-05,
      "loss": 0.3789428472518921,
      "memory(GiB)": 78.33,
      "step": 3345,
      "token_acc": 0.8881057268722466,
      "train_speed(iter/s)": 0.032454
    },
    {
      "epoch": 0.6483553747032893,
      "grad_norm": 0.10349465906620026,
      "learning_rate": 9.05390000166398e-05,
      "loss": 0.33017659187316895,
      "memory(GiB)": 78.33,
      "step": 3346,
      "token_acc": 0.9012410514386799,
      "train_speed(iter/s)": 0.032455
    },
    {
      "epoch": 0.648549144988616,
      "grad_norm": 0.10818509757518768,
      "learning_rate": 9.045075587640894e-05,
      "loss": 0.3425545394420624,
      "memory(GiB)": 78.33,
      "step": 3347,
      "token_acc": 0.8974606721217429,
      "train_speed(iter/s)": 0.032455
    },
    {
      "epoch": 0.6487429152739428,
      "grad_norm": 0.09324289113283157,
      "learning_rate": 9.036253619465285e-05,
      "loss": 0.3058326244354248,
      "memory(GiB)": 78.33,
      "step": 3348,
      "token_acc": 0.9082648317882186,
      "train_speed(iter/s)": 0.032456
    },
    {
      "epoch": 0.6489366855592695,
      "grad_norm": 0.10658305138349533,
      "learning_rate": 9.027434100760559e-05,
      "loss": 0.3650735318660736,
      "memory(GiB)": 78.33,
      "step": 3349,
      "token_acc": 0.8911853628023353,
      "train_speed(iter/s)": 0.032457
    },
    {
      "epoch": 0.6491304558445963,
      "grad_norm": 0.10458887368440628,
      "learning_rate": 9.018617035149141e-05,
      "loss": 0.34233736991882324,
      "memory(GiB)": 78.33,
      "step": 3350,
      "token_acc": 0.89612028460196,
      "train_speed(iter/s)": 0.032458
    },
    {
      "epoch": 0.649324226129923,
      "grad_norm": 0.10068278759717941,
      "learning_rate": 9.00980242625243e-05,
      "loss": 0.3308500647544861,
      "memory(GiB)": 78.33,
      "step": 3351,
      "token_acc": 0.8995980595980596,
      "train_speed(iter/s)": 0.032458
    },
    {
      "epoch": 0.6495179964152498,
      "grad_norm": 0.10845746099948883,
      "learning_rate": 9.000990277690828e-05,
      "loss": 0.349687397480011,
      "memory(GiB)": 78.33,
      "step": 3352,
      "token_acc": 0.8954418825203011,
      "train_speed(iter/s)": 0.032459
    },
    {
      "epoch": 0.6497117667005765,
      "grad_norm": 0.10593696683645248,
      "learning_rate": 8.992180593083718e-05,
      "loss": 0.35012173652648926,
      "memory(GiB)": 78.33,
      "step": 3353,
      "token_acc": 0.8978386677689855,
      "train_speed(iter/s)": 0.03246
    },
    {
      "epoch": 0.6499055369859033,
      "grad_norm": 0.10230911523103714,
      "learning_rate": 8.983373376049473e-05,
      "loss": 0.32801395654678345,
      "memory(GiB)": 78.33,
      "step": 3354,
      "token_acc": 0.9028890794570777,
      "train_speed(iter/s)": 0.032461
    },
    {
      "epoch": 0.65009930727123,
      "grad_norm": 0.09752920269966125,
      "learning_rate": 8.97456863020546e-05,
      "loss": 0.33961349725723267,
      "memory(GiB)": 78.33,
      "step": 3355,
      "token_acc": 0.8986005964670796,
      "train_speed(iter/s)": 0.032462
    },
    {
      "epoch": 0.6502930775565567,
      "grad_norm": 0.09952106326818466,
      "learning_rate": 8.965766359168017e-05,
      "loss": 0.3362662196159363,
      "memory(GiB)": 78.33,
      "step": 3356,
      "token_acc": 0.8996824741203501,
      "train_speed(iter/s)": 0.032462
    },
    {
      "epoch": 0.6504868478418835,
      "grad_norm": 0.09671690315008163,
      "learning_rate": 8.956966566552476e-05,
      "loss": 0.3496435880661011,
      "memory(GiB)": 78.33,
      "step": 3357,
      "token_acc": 0.8963326523493946,
      "train_speed(iter/s)": 0.032463
    },
    {
      "epoch": 0.6506806181272102,
      "grad_norm": 0.10482124984264374,
      "learning_rate": 8.948169255973147e-05,
      "loss": 0.3635352551937103,
      "memory(GiB)": 78.33,
      "step": 3358,
      "token_acc": 0.8922592685440517,
      "train_speed(iter/s)": 0.032464
    },
    {
      "epoch": 0.650874388412537,
      "grad_norm": 0.104715995490551,
      "learning_rate": 8.939374431043325e-05,
      "loss": 0.33896604180336,
      "memory(GiB)": 78.33,
      "step": 3359,
      "token_acc": 0.8968829586978186,
      "train_speed(iter/s)": 0.032465
    },
    {
      "epoch": 0.6510681586978637,
      "grad_norm": 0.10908566415309906,
      "learning_rate": 8.930582095375283e-05,
      "loss": 0.354704350233078,
      "memory(GiB)": 78.33,
      "step": 3360,
      "token_acc": 0.8975314829897876,
      "train_speed(iter/s)": 0.032465
    },
    {
      "epoch": 0.6512619289831905,
      "grad_norm": 0.10956370085477829,
      "learning_rate": 8.921792252580263e-05,
      "loss": 0.3756663203239441,
      "memory(GiB)": 78.33,
      "step": 3361,
      "token_acc": 0.8876317598533455,
      "train_speed(iter/s)": 0.032466
    },
    {
      "epoch": 0.6514556992685172,
      "grad_norm": 0.08828612416982651,
      "learning_rate": 8.913004906268495e-05,
      "loss": 0.3033355474472046,
      "memory(GiB)": 78.33,
      "step": 3362,
      "token_acc": 0.9071128935601346,
      "train_speed(iter/s)": 0.032467
    },
    {
      "epoch": 0.651649469553844,
      "grad_norm": 0.10207577049732208,
      "learning_rate": 8.904220060049172e-05,
      "loss": 0.3187861144542694,
      "memory(GiB)": 78.33,
      "step": 3363,
      "token_acc": 0.9046379128103328,
      "train_speed(iter/s)": 0.032468
    },
    {
      "epoch": 0.6518432398391707,
      "grad_norm": 0.10097214579582214,
      "learning_rate": 8.895437717530473e-05,
      "loss": 0.31431207060813904,
      "memory(GiB)": 78.33,
      "step": 3364,
      "token_acc": 0.9061064038885495,
      "train_speed(iter/s)": 0.032468
    },
    {
      "epoch": 0.6520370101244974,
      "grad_norm": 0.10009994357824326,
      "learning_rate": 8.886657882319537e-05,
      "loss": 0.32470712065696716,
      "memory(GiB)": 78.33,
      "step": 3365,
      "token_acc": 0.9029475799698543,
      "train_speed(iter/s)": 0.032469
    },
    {
      "epoch": 0.6522307804098242,
      "grad_norm": 0.1002596840262413,
      "learning_rate": 8.877880558022478e-05,
      "loss": 0.340162992477417,
      "memory(GiB)": 78.33,
      "step": 3366,
      "token_acc": 0.9000396877893901,
      "train_speed(iter/s)": 0.03247
    },
    {
      "epoch": 0.6524245506951509,
      "grad_norm": 0.09453920274972916,
      "learning_rate": 8.869105748244392e-05,
      "loss": 0.3194851279258728,
      "memory(GiB)": 78.33,
      "step": 3367,
      "token_acc": 0.9037842692144585,
      "train_speed(iter/s)": 0.032471
    },
    {
      "epoch": 0.6526183209804777,
      "grad_norm": 0.09737507253885269,
      "learning_rate": 8.86033345658931e-05,
      "loss": 0.33026716113090515,
      "memory(GiB)": 78.33,
      "step": 3368,
      "token_acc": 0.901304817031296,
      "train_speed(iter/s)": 0.032471
    },
    {
      "epoch": 0.6528120912658044,
      "grad_norm": 0.27890968322753906,
      "learning_rate": 8.851563686660263e-05,
      "loss": 0.3396485149860382,
      "memory(GiB)": 78.33,
      "step": 3369,
      "token_acc": 0.8994968003062954,
      "train_speed(iter/s)": 0.032472
    },
    {
      "epoch": 0.6530058615511312,
      "grad_norm": 0.09766080230474472,
      "learning_rate": 8.842796442059217e-05,
      "loss": 0.33352982997894287,
      "memory(GiB)": 78.33,
      "step": 3370,
      "token_acc": 0.9007515752714021,
      "train_speed(iter/s)": 0.032473
    },
    {
      "epoch": 0.6531996318364579,
      "grad_norm": 0.0987740233540535,
      "learning_rate": 8.834031726387126e-05,
      "loss": 0.33398640155792236,
      "memory(GiB)": 78.33,
      "step": 3371,
      "token_acc": 0.9004823690252018,
      "train_speed(iter/s)": 0.032474
    },
    {
      "epoch": 0.6533934021217846,
      "grad_norm": 0.09064597636461258,
      "learning_rate": 8.825269543243891e-05,
      "loss": 0.3236205577850342,
      "memory(GiB)": 78.33,
      "step": 3372,
      "token_acc": 0.9059911385089578,
      "train_speed(iter/s)": 0.032474
    },
    {
      "epoch": 0.6535871724071114,
      "grad_norm": 0.0983690693974495,
      "learning_rate": 8.816509896228374e-05,
      "loss": 0.32531166076660156,
      "memory(GiB)": 78.33,
      "step": 3373,
      "token_acc": 0.9019883621521745,
      "train_speed(iter/s)": 0.032475
    },
    {
      "epoch": 0.6537809426924381,
      "grad_norm": 0.10600029677152634,
      "learning_rate": 8.807752788938406e-05,
      "loss": 0.33703696727752686,
      "memory(GiB)": 78.33,
      "step": 3374,
      "token_acc": 0.898842851810377,
      "train_speed(iter/s)": 0.032476
    },
    {
      "epoch": 0.6539747129777649,
      "grad_norm": 0.10279665142297745,
      "learning_rate": 8.798998224970756e-05,
      "loss": 0.36671680212020874,
      "memory(GiB)": 78.33,
      "step": 3375,
      "token_acc": 0.8890161118835365,
      "train_speed(iter/s)": 0.032477
    },
    {
      "epoch": 0.6541684832630916,
      "grad_norm": 0.10394495725631714,
      "learning_rate": 8.790246207921164e-05,
      "loss": 0.344325453042984,
      "memory(GiB)": 78.33,
      "step": 3376,
      "token_acc": 0.8953350437253441,
      "train_speed(iter/s)": 0.032478
    },
    {
      "epoch": 0.6543622535484184,
      "grad_norm": 0.10396216064691544,
      "learning_rate": 8.781496741384314e-05,
      "loss": 0.3480278551578522,
      "memory(GiB)": 78.33,
      "step": 3377,
      "token_acc": 0.8966755763448045,
      "train_speed(iter/s)": 0.032478
    },
    {
      "epoch": 0.6545560238337451,
      "grad_norm": 0.10462364554405212,
      "learning_rate": 8.772749828953848e-05,
      "loss": 0.3206149935722351,
      "memory(GiB)": 78.33,
      "step": 3378,
      "token_acc": 0.9034575662325999,
      "train_speed(iter/s)": 0.032479
    },
    {
      "epoch": 0.6547497941190719,
      "grad_norm": 0.10066306591033936,
      "learning_rate": 8.764005474222365e-05,
      "loss": 0.32818686962127686,
      "memory(GiB)": 78.33,
      "step": 3379,
      "token_acc": 0.9018111463427743,
      "train_speed(iter/s)": 0.03248
    },
    {
      "epoch": 0.6549435644043986,
      "grad_norm": 0.10638129711151123,
      "learning_rate": 8.7552636807814e-05,
      "loss": 0.3548268675804138,
      "memory(GiB)": 78.33,
      "step": 3380,
      "token_acc": 0.8971172104263767,
      "train_speed(iter/s)": 0.032481
    },
    {
      "epoch": 0.6551373346897253,
      "grad_norm": 0.08810947835445404,
      "learning_rate": 8.746524452221442e-05,
      "loss": 0.31981220841407776,
      "memory(GiB)": 78.33,
      "step": 3381,
      "token_acc": 0.9033225494131816,
      "train_speed(iter/s)": 0.032481
    },
    {
      "epoch": 0.6553311049750521,
      "grad_norm": 0.0968799740076065,
      "learning_rate": 8.737787792131926e-05,
      "loss": 0.29892611503601074,
      "memory(GiB)": 78.33,
      "step": 3382,
      "token_acc": 0.9107401514971624,
      "train_speed(iter/s)": 0.032482
    },
    {
      "epoch": 0.6555248752603788,
      "grad_norm": 0.10424499958753586,
      "learning_rate": 8.729053704101246e-05,
      "loss": 0.3598312735557556,
      "memory(GiB)": 78.33,
      "step": 3383,
      "token_acc": 0.8916409545107142,
      "train_speed(iter/s)": 0.032483
    },
    {
      "epoch": 0.6557186455457056,
      "grad_norm": 0.1125241369009018,
      "learning_rate": 8.720322191716708e-05,
      "loss": 0.3586179316043854,
      "memory(GiB)": 78.33,
      "step": 3384,
      "token_acc": 0.8950719017611892,
      "train_speed(iter/s)": 0.032484
    },
    {
      "epoch": 0.6559124158310323,
      "grad_norm": 0.1094302088022232,
      "learning_rate": 8.71159325856459e-05,
      "loss": 0.35498055815696716,
      "memory(GiB)": 78.33,
      "step": 3385,
      "token_acc": 0.8951116185440015,
      "train_speed(iter/s)": 0.032485
    },
    {
      "epoch": 0.6561061861163591,
      "grad_norm": 0.09570290893316269,
      "learning_rate": 8.702866908230096e-05,
      "loss": 0.30832698941230774,
      "memory(GiB)": 78.33,
      "step": 3386,
      "token_acc": 0.9092654575646714,
      "train_speed(iter/s)": 0.032485
    },
    {
      "epoch": 0.6562999564016858,
      "grad_norm": 0.09967100620269775,
      "learning_rate": 8.694143144297376e-05,
      "loss": 0.32282644510269165,
      "memory(GiB)": 78.33,
      "step": 3387,
      "token_acc": 0.9025936599423631,
      "train_speed(iter/s)": 0.032486
    },
    {
      "epoch": 0.6564937266870126,
      "grad_norm": 0.09364413470029831,
      "learning_rate": 8.685421970349511e-05,
      "loss": 0.3126744031906128,
      "memory(GiB)": 78.33,
      "step": 3388,
      "token_acc": 0.9047869806163645,
      "train_speed(iter/s)": 0.032487
    },
    {
      "epoch": 0.6566874969723393,
      "grad_norm": 0.11045239120721817,
      "learning_rate": 8.676703389968515e-05,
      "loss": 0.3734998106956482,
      "memory(GiB)": 78.33,
      "step": 3389,
      "token_acc": 0.8892863324294698,
      "train_speed(iter/s)": 0.032488
    },
    {
      "epoch": 0.656881267257666,
      "grad_norm": 0.0982174277305603,
      "learning_rate": 8.667987406735363e-05,
      "loss": 0.3432157337665558,
      "memory(GiB)": 78.33,
      "step": 3390,
      "token_acc": 0.8978511530398323,
      "train_speed(iter/s)": 0.032488
    },
    {
      "epoch": 0.6570750375429928,
      "grad_norm": 0.09781992435455322,
      "learning_rate": 8.659274024229918e-05,
      "loss": 0.34358012676239014,
      "memory(GiB)": 78.33,
      "step": 3391,
      "token_acc": 0.8971234051399319,
      "train_speed(iter/s)": 0.032489
    },
    {
      "epoch": 0.6572688078283195,
      "grad_norm": 0.09683690965175629,
      "learning_rate": 8.65056324603102e-05,
      "loss": 0.3415444493293762,
      "memory(GiB)": 78.33,
      "step": 3392,
      "token_acc": 0.8975164282953227,
      "train_speed(iter/s)": 0.03249
    },
    {
      "epoch": 0.6574625781136463,
      "grad_norm": 0.09927819669246674,
      "learning_rate": 8.641855075716413e-05,
      "loss": 0.33681708574295044,
      "memory(GiB)": 78.33,
      "step": 3393,
      "token_acc": 0.9005583061202992,
      "train_speed(iter/s)": 0.032491
    },
    {
      "epoch": 0.657656348398973,
      "grad_norm": 0.10389687120914459,
      "learning_rate": 8.633149516862775e-05,
      "loss": 0.3464776277542114,
      "memory(GiB)": 78.33,
      "step": 3394,
      "token_acc": 0.8982288677996028,
      "train_speed(iter/s)": 0.032491
    },
    {
      "epoch": 0.6578501186842998,
      "grad_norm": 0.09941697865724564,
      "learning_rate": 8.624446573045717e-05,
      "loss": 0.33485954999923706,
      "memory(GiB)": 78.33,
      "step": 3395,
      "token_acc": 0.8993164654649689,
      "train_speed(iter/s)": 0.032492
    },
    {
      "epoch": 0.6580438889696265,
      "grad_norm": 0.10718297958374023,
      "learning_rate": 8.61574624783976e-05,
      "loss": 0.3694123923778534,
      "memory(GiB)": 78.33,
      "step": 3396,
      "token_acc": 0.8908513223063265,
      "train_speed(iter/s)": 0.032493
    },
    {
      "epoch": 0.6582376592549533,
      "grad_norm": 0.09040777385234833,
      "learning_rate": 8.607048544818386e-05,
      "loss": 0.29911673069000244,
      "memory(GiB)": 78.33,
      "step": 3397,
      "token_acc": 0.9073398896617894,
      "train_speed(iter/s)": 0.032494
    },
    {
      "epoch": 0.65843142954028,
      "grad_norm": 0.09987498074769974,
      "learning_rate": 8.598353467553945e-05,
      "loss": 0.3411368727684021,
      "memory(GiB)": 78.33,
      "step": 3398,
      "token_acc": 0.898384837029792,
      "train_speed(iter/s)": 0.032495
    },
    {
      "epoch": 0.6586251998256067,
      "grad_norm": 0.10136095434427261,
      "learning_rate": 8.58966101961776e-05,
      "loss": 0.3573547601699829,
      "memory(GiB)": 78.33,
      "step": 3399,
      "token_acc": 0.8953447859995176,
      "train_speed(iter/s)": 0.032495
    },
    {
      "epoch": 0.6588189701109335,
      "grad_norm": 0.10064728558063507,
      "learning_rate": 8.580971204580049e-05,
      "loss": 0.3405356705188751,
      "memory(GiB)": 78.33,
      "step": 3400,
      "token_acc": 0.8976995468804462,
      "train_speed(iter/s)": 0.032496
    },
    {
      "epoch": 0.6590127403962602,
      "grad_norm": 0.11121074110269547,
      "learning_rate": 8.572284026009947e-05,
      "loss": 0.3688167333602905,
      "memory(GiB)": 78.33,
      "step": 3401,
      "token_acc": 0.8922276364073873,
      "train_speed(iter/s)": 0.032493
    },
    {
      "epoch": 0.659206510681587,
      "grad_norm": 0.09966862946748734,
      "learning_rate": 8.563599487475517e-05,
      "loss": 0.32531312108039856,
      "memory(GiB)": 78.33,
      "step": 3402,
      "token_acc": 0.9004441763544637,
      "train_speed(iter/s)": 0.032494
    },
    {
      "epoch": 0.6594002809669137,
      "grad_norm": 0.09778161346912384,
      "learning_rate": 8.554917592543724e-05,
      "loss": 0.3067454695701599,
      "memory(GiB)": 78.33,
      "step": 3403,
      "token_acc": 0.9069717187542206,
      "train_speed(iter/s)": 0.032495
    },
    {
      "epoch": 0.6595940512522405,
      "grad_norm": 0.10241376608610153,
      "learning_rate": 8.546238344780468e-05,
      "loss": 0.3466450572013855,
      "memory(GiB)": 78.33,
      "step": 3404,
      "token_acc": 0.8972174435433221,
      "train_speed(iter/s)": 0.032496
    },
    {
      "epoch": 0.6597878215375672,
      "grad_norm": 0.10399215668439865,
      "learning_rate": 8.537561747750542e-05,
      "loss": 0.35718533396720886,
      "memory(GiB)": 78.33,
      "step": 3405,
      "token_acc": 0.894998862084661,
      "train_speed(iter/s)": 0.032496
    },
    {
      "epoch": 0.659981591822894,
      "grad_norm": 0.12148671597242355,
      "learning_rate": 8.528887805017661e-05,
      "loss": 0.38953131437301636,
      "memory(GiB)": 78.33,
      "step": 3406,
      "token_acc": 0.8871596564074931,
      "train_speed(iter/s)": 0.032497
    },
    {
      "epoch": 0.6601753621082207,
      "grad_norm": 0.10579821467399597,
      "learning_rate": 8.520216520144442e-05,
      "loss": 0.33564966917037964,
      "memory(GiB)": 78.33,
      "step": 3407,
      "token_acc": 0.8999078135039249,
      "train_speed(iter/s)": 0.032498
    },
    {
      "epoch": 0.6603691323935474,
      "grad_norm": 0.11077598482370377,
      "learning_rate": 8.51154789669241e-05,
      "loss": 0.3519205152988434,
      "memory(GiB)": 78.33,
      "step": 3408,
      "token_acc": 0.89746127867707,
      "train_speed(iter/s)": 0.032499
    },
    {
      "epoch": 0.6605629026788742,
      "grad_norm": 0.09937475621700287,
      "learning_rate": 8.502881938222021e-05,
      "loss": 0.36071473360061646,
      "memory(GiB)": 78.33,
      "step": 3409,
      "token_acc": 0.8904598045082784,
      "train_speed(iter/s)": 0.032499
    },
    {
      "epoch": 0.6607566729642009,
      "grad_norm": 0.09916140139102936,
      "learning_rate": 8.494218648292594e-05,
      "loss": 0.3261912763118744,
      "memory(GiB)": 78.33,
      "step": 3410,
      "token_acc": 0.9021432978352424,
      "train_speed(iter/s)": 0.0325
    },
    {
      "epoch": 0.6609504432495277,
      "grad_norm": 0.10734451562166214,
      "learning_rate": 8.485558030462389e-05,
      "loss": 0.32132577896118164,
      "memory(GiB)": 78.33,
      "step": 3411,
      "token_acc": 0.9042697182232066,
      "train_speed(iter/s)": 0.032501
    },
    {
      "epoch": 0.6611442135348544,
      "grad_norm": 0.09797913581132889,
      "learning_rate": 8.476900088288554e-05,
      "loss": 0.33835476636886597,
      "memory(GiB)": 78.33,
      "step": 3412,
      "token_acc": 0.9016252123424863,
      "train_speed(iter/s)": 0.032502
    },
    {
      "epoch": 0.6613379838201812,
      "grad_norm": 0.10399965941905975,
      "learning_rate": 8.468244825327132e-05,
      "loss": 0.3561461269855499,
      "memory(GiB)": 78.33,
      "step": 3413,
      "token_acc": 0.8911394426552394,
      "train_speed(iter/s)": 0.032502
    },
    {
      "epoch": 0.6615317541055079,
      "grad_norm": 0.14848816394805908,
      "learning_rate": 8.459592245133076e-05,
      "loss": 0.3276277780532837,
      "memory(GiB)": 78.33,
      "step": 3414,
      "token_acc": 0.9025031223980017,
      "train_speed(iter/s)": 0.032503
    },
    {
      "epoch": 0.6617255243908347,
      "grad_norm": 0.10193125158548355,
      "learning_rate": 8.450942351260228e-05,
      "loss": 0.3469090163707733,
      "memory(GiB)": 78.33,
      "step": 3415,
      "token_acc": 0.89754166435967,
      "train_speed(iter/s)": 0.032504
    },
    {
      "epoch": 0.6619192946761614,
      "grad_norm": 0.0984477773308754,
      "learning_rate": 8.442295147261347e-05,
      "loss": 0.33425408601760864,
      "memory(GiB)": 78.33,
      "step": 3416,
      "token_acc": 0.8991217336922533,
      "train_speed(iter/s)": 0.032505
    },
    {
      "epoch": 0.6621130649614881,
      "grad_norm": 0.09679713100194931,
      "learning_rate": 8.43365063668805e-05,
      "loss": 0.3265281617641449,
      "memory(GiB)": 78.33,
      "step": 3417,
      "token_acc": 0.9010822632198733,
      "train_speed(iter/s)": 0.032505
    },
    {
      "epoch": 0.6623068352468149,
      "grad_norm": 0.10674044489860535,
      "learning_rate": 8.425008823090885e-05,
      "loss": 0.32397177815437317,
      "memory(GiB)": 78.33,
      "step": 3418,
      "token_acc": 0.9022934871751365,
      "train_speed(iter/s)": 0.032506
    },
    {
      "epoch": 0.6625006055321416,
      "grad_norm": 0.09785594791173935,
      "learning_rate": 8.416369710019276e-05,
      "loss": 0.36127400398254395,
      "memory(GiB)": 78.33,
      "step": 3419,
      "token_acc": 0.8940439348587084,
      "train_speed(iter/s)": 0.032507
    },
    {
      "epoch": 0.6626943758174684,
      "grad_norm": 0.0925799012184143,
      "learning_rate": 8.407733301021534e-05,
      "loss": 0.3258381485939026,
      "memory(GiB)": 78.33,
      "step": 3420,
      "token_acc": 0.9027204212265125,
      "train_speed(iter/s)": 0.032508
    },
    {
      "epoch": 0.6628881461027951,
      "grad_norm": 0.09700040519237518,
      "learning_rate": 8.399099599644869e-05,
      "loss": 0.3296799957752228,
      "memory(GiB)": 78.33,
      "step": 3421,
      "token_acc": 0.9014329474066767,
      "train_speed(iter/s)": 0.032509
    },
    {
      "epoch": 0.6630819163881219,
      "grad_norm": 0.09748223423957825,
      "learning_rate": 8.390468609435364e-05,
      "loss": 0.3318982422351837,
      "memory(GiB)": 78.33,
      "step": 3422,
      "token_acc": 0.8991830021301265,
      "train_speed(iter/s)": 0.032509
    },
    {
      "epoch": 0.6632756866734486,
      "grad_norm": 0.1090392917394638,
      "learning_rate": 8.381840333938017e-05,
      "loss": 0.36316922307014465,
      "memory(GiB)": 78.33,
      "step": 3423,
      "token_acc": 0.8909560870306198,
      "train_speed(iter/s)": 0.03251
    },
    {
      "epoch": 0.6634694569587753,
      "grad_norm": 0.10358504951000214,
      "learning_rate": 8.37321477669667e-05,
      "loss": 0.3546435236930847,
      "memory(GiB)": 78.33,
      "step": 3424,
      "token_acc": 0.8954622502570373,
      "train_speed(iter/s)": 0.032511
    },
    {
      "epoch": 0.6636632272441021,
      "grad_norm": 0.1012730523943901,
      "learning_rate": 8.364591941254091e-05,
      "loss": 0.3311161994934082,
      "memory(GiB)": 78.33,
      "step": 3425,
      "token_acc": 0.9006039512744396,
      "train_speed(iter/s)": 0.032512
    },
    {
      "epoch": 0.6638569975294288,
      "grad_norm": 0.10826534032821655,
      "learning_rate": 8.355971831151901e-05,
      "loss": 0.36814382672309875,
      "memory(GiB)": 78.33,
      "step": 3426,
      "token_acc": 0.8914054979142154,
      "train_speed(iter/s)": 0.032512
    },
    {
      "epoch": 0.6640507678147556,
      "grad_norm": 0.10928687453269958,
      "learning_rate": 8.347354449930611e-05,
      "loss": 0.35287556052207947,
      "memory(GiB)": 78.33,
      "step": 3427,
      "token_acc": 0.8944273647136077,
      "train_speed(iter/s)": 0.032513
    },
    {
      "epoch": 0.6642445381000823,
      "grad_norm": 0.10338309407234192,
      "learning_rate": 8.338739801129611e-05,
      "loss": 0.3381505608558655,
      "memory(GiB)": 78.33,
      "step": 3428,
      "token_acc": 0.8976415225684292,
      "train_speed(iter/s)": 0.032514
    },
    {
      "epoch": 0.6644383083854091,
      "grad_norm": 0.09947198629379272,
      "learning_rate": 8.330127888287165e-05,
      "loss": 0.3421975374221802,
      "memory(GiB)": 78.33,
      "step": 3429,
      "token_acc": 0.8979885810067572,
      "train_speed(iter/s)": 0.032515
    },
    {
      "epoch": 0.6646320786707358,
      "grad_norm": 0.11111465096473694,
      "learning_rate": 8.321518714940434e-05,
      "loss": 0.35146766901016235,
      "memory(GiB)": 78.33,
      "step": 3430,
      "token_acc": 0.8957091043216328,
      "train_speed(iter/s)": 0.032516
    },
    {
      "epoch": 0.6648258489560626,
      "grad_norm": 0.09252513200044632,
      "learning_rate": 8.312912284625412e-05,
      "loss": 0.3317872881889343,
      "memory(GiB)": 78.33,
      "step": 3431,
      "token_acc": 0.8990752267472879,
      "train_speed(iter/s)": 0.032516
    },
    {
      "epoch": 0.6650196192413893,
      "grad_norm": 0.09670988470315933,
      "learning_rate": 8.30430860087701e-05,
      "loss": 0.3309570848941803,
      "memory(GiB)": 78.33,
      "step": 3432,
      "token_acc": 0.9016609407692506,
      "train_speed(iter/s)": 0.032517
    },
    {
      "epoch": 0.665213389526716,
      "grad_norm": 0.10486641526222229,
      "learning_rate": 8.295707667228987e-05,
      "loss": 0.33131322264671326,
      "memory(GiB)": 78.33,
      "step": 3433,
      "token_acc": 0.9014130083054687,
      "train_speed(iter/s)": 0.032518
    },
    {
      "epoch": 0.6654071598120428,
      "grad_norm": 0.09246989339590073,
      "learning_rate": 8.287109487213974e-05,
      "loss": 0.3162573575973511,
      "memory(GiB)": 78.33,
      "step": 3434,
      "token_acc": 0.9054944021074419,
      "train_speed(iter/s)": 0.032519
    },
    {
      "epoch": 0.6656009300973695,
      "grad_norm": 0.08992563188076019,
      "learning_rate": 8.278514064363477e-05,
      "loss": 0.31209853291511536,
      "memory(GiB)": 78.33,
      "step": 3435,
      "token_acc": 0.9090142626291013,
      "train_speed(iter/s)": 0.032519
    },
    {
      "epoch": 0.6657947003826963,
      "grad_norm": 0.102995365858078,
      "learning_rate": 8.269921402207863e-05,
      "loss": 0.3250240385532379,
      "memory(GiB)": 78.33,
      "step": 3436,
      "token_acc": 0.9044458079991075,
      "train_speed(iter/s)": 0.03252
    },
    {
      "epoch": 0.665988470668023,
      "grad_norm": 0.09776205569505692,
      "learning_rate": 8.261331504276378e-05,
      "loss": 0.31458067893981934,
      "memory(GiB)": 78.33,
      "step": 3437,
      "token_acc": 0.9068092088396656,
      "train_speed(iter/s)": 0.032521
    },
    {
      "epoch": 0.6661822409533498,
      "grad_norm": 0.12002206593751907,
      "learning_rate": 8.25274437409712e-05,
      "loss": 0.4206904470920563,
      "memory(GiB)": 78.33,
      "step": 3438,
      "token_acc": 0.8826992485651737,
      "train_speed(iter/s)": 0.032522
    },
    {
      "epoch": 0.6663760112386765,
      "grad_norm": 0.09399055689573288,
      "learning_rate": 8.244160015197054e-05,
      "loss": 0.3172173798084259,
      "memory(GiB)": 78.33,
      "step": 3439,
      "token_acc": 0.9048228875854827,
      "train_speed(iter/s)": 0.032522
    },
    {
      "epoch": 0.6665697815240033,
      "grad_norm": 0.10963206738233566,
      "learning_rate": 8.235578431102004e-05,
      "loss": 0.357374906539917,
      "memory(GiB)": 78.33,
      "step": 3440,
      "token_acc": 0.8945577178923919,
      "train_speed(iter/s)": 0.032523
    },
    {
      "epoch": 0.66676355180933,
      "grad_norm": 0.09495268017053604,
      "learning_rate": 8.226999625336662e-05,
      "loss": 0.3170054852962494,
      "memory(GiB)": 78.33,
      "step": 3441,
      "token_acc": 0.9031417624521073,
      "train_speed(iter/s)": 0.032524
    },
    {
      "epoch": 0.6669573220946567,
      "grad_norm": 0.10406588762998581,
      "learning_rate": 8.21842360142457e-05,
      "loss": 0.3348080813884735,
      "memory(GiB)": 78.33,
      "step": 3442,
      "token_acc": 0.8991230567735322,
      "train_speed(iter/s)": 0.032524
    },
    {
      "epoch": 0.6671510923799835,
      "grad_norm": 0.0943104475736618,
      "learning_rate": 8.209850362888126e-05,
      "loss": 0.3240373432636261,
      "memory(GiB)": 78.33,
      "step": 3443,
      "token_acc": 0.9041046505639388,
      "train_speed(iter/s)": 0.032525
    },
    {
      "epoch": 0.6673448626653102,
      "grad_norm": 0.10466547310352325,
      "learning_rate": 8.201279913248606e-05,
      "loss": 0.35648593306541443,
      "memory(GiB)": 78.33,
      "step": 3444,
      "token_acc": 0.8950629795859273,
      "train_speed(iter/s)": 0.032526
    },
    {
      "epoch": 0.667538632950637,
      "grad_norm": 0.1078016608953476,
      "learning_rate": 8.192712256026111e-05,
      "loss": 0.3920597434043884,
      "memory(GiB)": 78.33,
      "step": 3445,
      "token_acc": 0.8864499320972257,
      "train_speed(iter/s)": 0.032527
    },
    {
      "epoch": 0.6677324032359637,
      "grad_norm": 0.0912395566701889,
      "learning_rate": 8.18414739473961e-05,
      "loss": 0.3125567138195038,
      "memory(GiB)": 78.33,
      "step": 3446,
      "token_acc": 0.9068691030238295,
      "train_speed(iter/s)": 0.032528
    },
    {
      "epoch": 0.6679261735212905,
      "grad_norm": 0.10371241718530655,
      "learning_rate": 8.175585332906928e-05,
      "loss": 0.35416972637176514,
      "memory(GiB)": 78.33,
      "step": 3447,
      "token_acc": 0.8940065267570507,
      "train_speed(iter/s)": 0.032528
    },
    {
      "epoch": 0.6681199438066172,
      "grad_norm": 0.1040647029876709,
      "learning_rate": 8.167026074044719e-05,
      "loss": 0.34062814712524414,
      "memory(GiB)": 78.33,
      "step": 3448,
      "token_acc": 0.8997950416935733,
      "train_speed(iter/s)": 0.032529
    },
    {
      "epoch": 0.668313714091944,
      "grad_norm": 0.1095561683177948,
      "learning_rate": 8.158469621668522e-05,
      "loss": 0.35409224033355713,
      "memory(GiB)": 78.33,
      "step": 3449,
      "token_acc": 0.8959530893440019,
      "train_speed(iter/s)": 0.03253
    },
    {
      "epoch": 0.6685074843772707,
      "grad_norm": 0.09262026101350784,
      "learning_rate": 8.149915979292683e-05,
      "loss": 0.3020830750465393,
      "memory(GiB)": 78.33,
      "step": 3450,
      "token_acc": 0.9076835711816701,
      "train_speed(iter/s)": 0.032531
    },
    {
      "epoch": 0.6687012546625974,
      "grad_norm": 0.109112448990345,
      "learning_rate": 8.141365150430421e-05,
      "loss": 0.3594379723072052,
      "memory(GiB)": 78.33,
      "step": 3451,
      "token_acc": 0.893763065114973,
      "train_speed(iter/s)": 0.032531
    },
    {
      "epoch": 0.6688950249479242,
      "grad_norm": 0.10545790940523148,
      "learning_rate": 8.132817138593792e-05,
      "loss": 0.339529812335968,
      "memory(GiB)": 78.33,
      "step": 3452,
      "token_acc": 0.8981471052038748,
      "train_speed(iter/s)": 0.032532
    },
    {
      "epoch": 0.6690887952332509,
      "grad_norm": 0.10486025363206863,
      "learning_rate": 8.124271947293695e-05,
      "loss": 0.3370751142501831,
      "memory(GiB)": 78.33,
      "step": 3453,
      "token_acc": 0.9002499368137269,
      "train_speed(iter/s)": 0.032533
    },
    {
      "epoch": 0.6692825655185777,
      "grad_norm": 0.10516992211341858,
      "learning_rate": 8.115729580039863e-05,
      "loss": 0.31734785437583923,
      "memory(GiB)": 78.33,
      "step": 3454,
      "token_acc": 0.9048377947651333,
      "train_speed(iter/s)": 0.032534
    },
    {
      "epoch": 0.6694763358039044,
      "grad_norm": 0.09757381677627563,
      "learning_rate": 8.107190040340878e-05,
      "loss": 0.3384143114089966,
      "memory(GiB)": 78.33,
      "step": 3455,
      "token_acc": 0.8986423042761071,
      "train_speed(iter/s)": 0.032534
    },
    {
      "epoch": 0.6696701060892312,
      "grad_norm": 0.10361029207706451,
      "learning_rate": 8.09865333170417e-05,
      "loss": 0.36140722036361694,
      "memory(GiB)": 78.33,
      "step": 3456,
      "token_acc": 0.8917902428540726,
      "train_speed(iter/s)": 0.032535
    },
    {
      "epoch": 0.6698638763745579,
      "grad_norm": 0.10952255129814148,
      "learning_rate": 8.090119457635973e-05,
      "loss": 0.3548586666584015,
      "memory(GiB)": 78.33,
      "step": 3457,
      "token_acc": 0.8944674205396651,
      "train_speed(iter/s)": 0.032536
    },
    {
      "epoch": 0.6700576466598847,
      "grad_norm": 0.09695498645305634,
      "learning_rate": 8.081588421641399e-05,
      "loss": 0.34991052746772766,
      "memory(GiB)": 78.33,
      "step": 3458,
      "token_acc": 0.8968775374621687,
      "train_speed(iter/s)": 0.032537
    },
    {
      "epoch": 0.6702514169452114,
      "grad_norm": 0.096768319606781,
      "learning_rate": 8.073060227224364e-05,
      "loss": 0.3430905044078827,
      "memory(GiB)": 78.33,
      "step": 3459,
      "token_acc": 0.8984231274638633,
      "train_speed(iter/s)": 0.032537
    },
    {
      "epoch": 0.6704451872305381,
      "grad_norm": 0.10539654642343521,
      "learning_rate": 8.064534877887625e-05,
      "loss": 0.3595712184906006,
      "memory(GiB)": 78.33,
      "step": 3460,
      "token_acc": 0.8929441752331867,
      "train_speed(iter/s)": 0.032538
    },
    {
      "epoch": 0.670638957515865,
      "grad_norm": 0.109794020652771,
      "learning_rate": 8.056012377132778e-05,
      "loss": 0.38728106021881104,
      "memory(GiB)": 78.33,
      "step": 3461,
      "token_acc": 0.8862599615278923,
      "train_speed(iter/s)": 0.032539
    },
    {
      "epoch": 0.6708327278011917,
      "grad_norm": 0.10039215534925461,
      "learning_rate": 8.047492728460232e-05,
      "loss": 0.3494102358818054,
      "memory(GiB)": 78.33,
      "step": 3462,
      "token_acc": 0.8965799655454828,
      "train_speed(iter/s)": 0.03254
    },
    {
      "epoch": 0.6710264980865185,
      "grad_norm": 0.11146795004606247,
      "learning_rate": 8.038975935369256e-05,
      "loss": 0.34334322810173035,
      "memory(GiB)": 78.33,
      "step": 3463,
      "token_acc": 0.8990417830644919,
      "train_speed(iter/s)": 0.032541
    },
    {
      "epoch": 0.6712202683718452,
      "grad_norm": 0.10396383702754974,
      "learning_rate": 8.030462001357903e-05,
      "loss": 0.3301684558391571,
      "memory(GiB)": 78.33,
      "step": 3464,
      "token_acc": 0.9044740024183797,
      "train_speed(iter/s)": 0.032541
    },
    {
      "epoch": 0.671414038657172,
      "grad_norm": 0.10532711446285248,
      "learning_rate": 8.02195092992309e-05,
      "loss": 0.3416173756122589,
      "memory(GiB)": 78.33,
      "step": 3465,
      "token_acc": 0.8965047796620201,
      "train_speed(iter/s)": 0.032542
    },
    {
      "epoch": 0.6716078089424987,
      "grad_norm": 0.10811363905668259,
      "learning_rate": 8.013442724560537e-05,
      "loss": 0.3821423649787903,
      "memory(GiB)": 78.33,
      "step": 3466,
      "token_acc": 0.8860308779126075,
      "train_speed(iter/s)": 0.032543
    },
    {
      "epoch": 0.6718015792278255,
      "grad_norm": 0.10519671440124512,
      "learning_rate": 8.004937388764793e-05,
      "loss": 0.36607247591018677,
      "memory(GiB)": 78.33,
      "step": 3467,
      "token_acc": 0.892551655187736,
      "train_speed(iter/s)": 0.032543
    },
    {
      "epoch": 0.6719953495131522,
      "grad_norm": 0.10614033788442612,
      "learning_rate": 7.996434926029227e-05,
      "loss": 0.36225974559783936,
      "memory(GiB)": 78.33,
      "step": 3468,
      "token_acc": 0.8928445644517492,
      "train_speed(iter/s)": 0.032544
    },
    {
      "epoch": 0.672189119798479,
      "grad_norm": 0.11390534043312073,
      "learning_rate": 7.987935339846025e-05,
      "loss": 0.3544899821281433,
      "memory(GiB)": 78.33,
      "step": 3469,
      "token_acc": 0.895393061480967,
      "train_speed(iter/s)": 0.032545
    },
    {
      "epoch": 0.6723828900838057,
      "grad_norm": 0.09529854357242584,
      "learning_rate": 7.979438633706206e-05,
      "loss": 0.3296326994895935,
      "memory(GiB)": 78.33,
      "step": 3470,
      "token_acc": 0.8999336332128899,
      "train_speed(iter/s)": 0.032546
    },
    {
      "epoch": 0.6725766603691324,
      "grad_norm": 0.09383497387170792,
      "learning_rate": 7.97094481109959e-05,
      "loss": 0.31189560890197754,
      "memory(GiB)": 78.33,
      "step": 3471,
      "token_acc": 0.9070854413313131,
      "train_speed(iter/s)": 0.032547
    },
    {
      "epoch": 0.6727704306544592,
      "grad_norm": 0.10056065768003464,
      "learning_rate": 7.962453875514821e-05,
      "loss": 0.3454468250274658,
      "memory(GiB)": 78.33,
      "step": 3472,
      "token_acc": 0.8972385445115727,
      "train_speed(iter/s)": 0.032547
    },
    {
      "epoch": 0.6729642009397859,
      "grad_norm": 0.09275511652231216,
      "learning_rate": 7.953965830439349e-05,
      "loss": 0.3189311623573303,
      "memory(GiB)": 78.33,
      "step": 3473,
      "token_acc": 0.9043669819878191,
      "train_speed(iter/s)": 0.032548
    },
    {
      "epoch": 0.6731579712251127,
      "grad_norm": 0.09983082860708237,
      "learning_rate": 7.945480679359443e-05,
      "loss": 0.33363524079322815,
      "memory(GiB)": 78.33,
      "step": 3474,
      "token_acc": 0.901058171118153,
      "train_speed(iter/s)": 0.032549
    },
    {
      "epoch": 0.6733517415104394,
      "grad_norm": 0.3244771361351013,
      "learning_rate": 7.936998425760186e-05,
      "loss": 0.3636043071746826,
      "memory(GiB)": 78.33,
      "step": 3475,
      "token_acc": 0.8894176781990268,
      "train_speed(iter/s)": 0.03255
    },
    {
      "epoch": 0.6735455117957662,
      "grad_norm": 0.10669998079538345,
      "learning_rate": 7.928519073125461e-05,
      "loss": 0.35683298110961914,
      "memory(GiB)": 78.33,
      "step": 3476,
      "token_acc": 0.8926932475117707,
      "train_speed(iter/s)": 0.03255
    },
    {
      "epoch": 0.6737392820810929,
      "grad_norm": 0.09464793652296066,
      "learning_rate": 7.920042624937976e-05,
      "loss": 0.3355555236339569,
      "memory(GiB)": 78.33,
      "step": 3477,
      "token_acc": 0.9,
      "train_speed(iter/s)": 0.032551
    },
    {
      "epoch": 0.6739330523664196,
      "grad_norm": 0.10269086062908173,
      "learning_rate": 7.911569084679229e-05,
      "loss": 0.357306569814682,
      "memory(GiB)": 78.33,
      "step": 3478,
      "token_acc": 0.8936895083236547,
      "train_speed(iter/s)": 0.032552
    },
    {
      "epoch": 0.6741268226517464,
      "grad_norm": 0.09656988829374313,
      "learning_rate": 7.903098455829535e-05,
      "loss": 0.32650524377822876,
      "memory(GiB)": 78.33,
      "step": 3479,
      "token_acc": 0.9009720369210662,
      "train_speed(iter/s)": 0.032552
    },
    {
      "epoch": 0.6743205929370731,
      "grad_norm": 0.10245799273252487,
      "learning_rate": 7.894630741868004e-05,
      "loss": 0.3412097096443176,
      "memory(GiB)": 78.33,
      "step": 3480,
      "token_acc": 0.8984286408553378,
      "train_speed(iter/s)": 0.032553
    },
    {
      "epoch": 0.6745143632223999,
      "grad_norm": 0.0951901376247406,
      "learning_rate": 7.88616594627255e-05,
      "loss": 0.31834906339645386,
      "memory(GiB)": 78.33,
      "step": 3481,
      "token_acc": 0.9045248313917842,
      "train_speed(iter/s)": 0.032554
    },
    {
      "epoch": 0.6747081335077266,
      "grad_norm": 0.09806448966264725,
      "learning_rate": 7.877704072519911e-05,
      "loss": 0.3509221076965332,
      "memory(GiB)": 78.33,
      "step": 3482,
      "token_acc": 0.8939818589482267,
      "train_speed(iter/s)": 0.032555
    },
    {
      "epoch": 0.6749019037930534,
      "grad_norm": 0.10363534837961197,
      "learning_rate": 7.869245124085581e-05,
      "loss": 0.3548724353313446,
      "memory(GiB)": 78.33,
      "step": 3483,
      "token_acc": 0.8933054967997143,
      "train_speed(iter/s)": 0.032555
    },
    {
      "epoch": 0.6750956740783801,
      "grad_norm": 0.10435101389884949,
      "learning_rate": 7.860789104443896e-05,
      "loss": 0.36100128293037415,
      "memory(GiB)": 78.33,
      "step": 3484,
      "token_acc": 0.8941173294457493,
      "train_speed(iter/s)": 0.032556
    },
    {
      "epoch": 0.6752894443637069,
      "grad_norm": 0.09948544204235077,
      "learning_rate": 7.852336017067964e-05,
      "loss": 0.32601821422576904,
      "memory(GiB)": 78.33,
      "step": 3485,
      "token_acc": 0.901149788196911,
      "train_speed(iter/s)": 0.032557
    },
    {
      "epoch": 0.6754832146490336,
      "grad_norm": 0.10575611889362335,
      "learning_rate": 7.843885865429693e-05,
      "loss": 0.36941444873809814,
      "memory(GiB)": 78.33,
      "step": 3486,
      "token_acc": 0.8913294131211806,
      "train_speed(iter/s)": 0.032557
    },
    {
      "epoch": 0.6756769849343603,
      "grad_norm": 0.10058741271495819,
      "learning_rate": 7.835438652999791e-05,
      "loss": 0.3273415267467499,
      "memory(GiB)": 78.33,
      "step": 3487,
      "token_acc": 0.9005714607437435,
      "train_speed(iter/s)": 0.032558
    },
    {
      "epoch": 0.6758707552196871,
      "grad_norm": 0.09746947884559631,
      "learning_rate": 7.826994383247747e-05,
      "loss": 0.31902948021888733,
      "memory(GiB)": 78.33,
      "step": 3488,
      "token_acc": 0.9043132861435046,
      "train_speed(iter/s)": 0.032559
    },
    {
      "epoch": 0.6760645255050138,
      "grad_norm": 0.0960841104388237,
      "learning_rate": 7.818553059641867e-05,
      "loss": 0.32989510893821716,
      "memory(GiB)": 78.33,
      "step": 3489,
      "token_acc": 0.9024306199803854,
      "train_speed(iter/s)": 0.03256
    },
    {
      "epoch": 0.6762582957903406,
      "grad_norm": 0.10085577517747879,
      "learning_rate": 7.810114685649207e-05,
      "loss": 0.3315393626689911,
      "memory(GiB)": 78.33,
      "step": 3490,
      "token_acc": 0.8987733391228832,
      "train_speed(iter/s)": 0.03256
    },
    {
      "epoch": 0.6764520660756673,
      "grad_norm": 0.1091570183634758,
      "learning_rate": 7.801679264735652e-05,
      "loss": 0.3644520044326782,
      "memory(GiB)": 78.33,
      "step": 3491,
      "token_acc": 0.8908445452104651,
      "train_speed(iter/s)": 0.032561
    },
    {
      "epoch": 0.6766458363609941,
      "grad_norm": 0.11124227195978165,
      "learning_rate": 7.793246800365848e-05,
      "loss": 0.36782822012901306,
      "memory(GiB)": 78.33,
      "step": 3492,
      "token_acc": 0.8893680254071851,
      "train_speed(iter/s)": 0.032562
    },
    {
      "epoch": 0.6768396066463208,
      "grad_norm": 0.11075468361377716,
      "learning_rate": 7.784817296003237e-05,
      "loss": 0.3444291353225708,
      "memory(GiB)": 78.33,
      "step": 3493,
      "token_acc": 0.8969756577329727,
      "train_speed(iter/s)": 0.032563
    },
    {
      "epoch": 0.6770333769316476,
      "grad_norm": 0.10527750104665756,
      "learning_rate": 7.776390755110041e-05,
      "loss": 0.36560899019241333,
      "memory(GiB)": 78.33,
      "step": 3494,
      "token_acc": 0.8897428155792825,
      "train_speed(iter/s)": 0.032564
    },
    {
      "epoch": 0.6772271472169743,
      "grad_norm": 0.10618400573730469,
      "learning_rate": 7.767967181147265e-05,
      "loss": 0.35203975439071655,
      "memory(GiB)": 78.33,
      "step": 3495,
      "token_acc": 0.8978139451084018,
      "train_speed(iter/s)": 0.032564
    },
    {
      "epoch": 0.677420917502301,
      "grad_norm": 0.09277181327342987,
      "learning_rate": 7.759546577574708e-05,
      "loss": 0.3013547956943512,
      "memory(GiB)": 78.33,
      "step": 3496,
      "token_acc": 0.9089381973172872,
      "train_speed(iter/s)": 0.032565
    },
    {
      "epoch": 0.6776146877876278,
      "grad_norm": 0.09669741243124008,
      "learning_rate": 7.751128947850921e-05,
      "loss": 0.3188782334327698,
      "memory(GiB)": 78.33,
      "step": 3497,
      "token_acc": 0.9045035268583831,
      "train_speed(iter/s)": 0.032566
    },
    {
      "epoch": 0.6778084580729545,
      "grad_norm": 0.09220936894416809,
      "learning_rate": 7.742714295433265e-05,
      "loss": 0.3000409007072449,
      "memory(GiB)": 78.33,
      "step": 3498,
      "token_acc": 0.9094853531300161,
      "train_speed(iter/s)": 0.032566
    },
    {
      "epoch": 0.6780022283582813,
      "grad_norm": 0.1013851910829544,
      "learning_rate": 7.734302623777857e-05,
      "loss": 0.35849493741989136,
      "memory(GiB)": 78.33,
      "step": 3499,
      "token_acc": 0.8939226809538817,
      "train_speed(iter/s)": 0.032567
    },
    {
      "epoch": 0.678195998643608,
      "grad_norm": 0.10033900290727615,
      "learning_rate": 7.7258939363396e-05,
      "loss": 0.3273484408855438,
      "memory(GiB)": 78.33,
      "step": 3500,
      "token_acc": 0.902882797731569,
      "train_speed(iter/s)": 0.032568
    },
    {
      "epoch": 0.678195998643608,
      "eval_loss": 0.3950374126434326,
      "eval_runtime": 1346.3478,
      "eval_samples_per_second": 5.013,
      "eval_steps_per_second": 5.013,
      "eval_token_acc": 0.8989412672013904,
      "step": 3500
    },
    {
      "epoch": 0.6783897689289348,
      "grad_norm": 0.09484906494617462,
      "learning_rate": 7.717488236572166e-05,
      "loss": 0.3013468384742737,
      "memory(GiB)": 78.33,
      "step": 3501,
      "token_acc": 0.9093957031354772,
      "train_speed(iter/s)": 0.032166
    },
    {
      "epoch": 0.6785835392142615,
      "grad_norm": 0.1061227023601532,
      "learning_rate": 7.709085527927994e-05,
      "loss": 0.3452511131763458,
      "memory(GiB)": 78.33,
      "step": 3502,
      "token_acc": 0.8964009033475315,
      "train_speed(iter/s)": 0.032166
    },
    {
      "epoch": 0.6787773094995883,
      "grad_norm": 0.10096945613622665,
      "learning_rate": 7.700685813858317e-05,
      "loss": 0.330334335565567,
      "memory(GiB)": 78.33,
      "step": 3503,
      "token_acc": 0.9003261023677868,
      "train_speed(iter/s)": 0.032167
    },
    {
      "epoch": 0.678971079784915,
      "grad_norm": 0.09609825909137726,
      "learning_rate": 7.692289097813119e-05,
      "loss": 0.32408520579338074,
      "memory(GiB)": 78.33,
      "step": 3504,
      "token_acc": 0.9045769211614124,
      "train_speed(iter/s)": 0.032168
    },
    {
      "epoch": 0.6791648500702417,
      "grad_norm": 0.08847975730895996,
      "learning_rate": 7.683895383241152e-05,
      "loss": 0.3026714622974396,
      "memory(GiB)": 78.33,
      "step": 3505,
      "token_acc": 0.9074179743223966,
      "train_speed(iter/s)": 0.032169
    },
    {
      "epoch": 0.6793586203555685,
      "grad_norm": 0.10310973227024078,
      "learning_rate": 7.675504673589942e-05,
      "loss": 0.3564712107181549,
      "memory(GiB)": 78.33,
      "step": 3506,
      "token_acc": 0.8931326434619002,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.6795523906408952,
      "grad_norm": 0.09145442396402359,
      "learning_rate": 7.66711697230578e-05,
      "loss": 0.29558998346328735,
      "memory(GiB)": 78.33,
      "step": 3507,
      "token_acc": 0.9112166695232732,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.679746160926222,
      "grad_norm": 0.11162301152944565,
      "learning_rate": 7.65873228283372e-05,
      "loss": 0.37096238136291504,
      "memory(GiB)": 78.33,
      "step": 3508,
      "token_acc": 0.8897501419647927,
      "train_speed(iter/s)": 0.032171
    },
    {
      "epoch": 0.6799399312115487,
      "grad_norm": 0.09270161390304565,
      "learning_rate": 7.650350608617573e-05,
      "loss": 0.30103829503059387,
      "memory(GiB)": 78.33,
      "step": 3509,
      "token_acc": 0.9094411328318193,
      "train_speed(iter/s)": 0.032172
    },
    {
      "epoch": 0.6801337014968755,
      "grad_norm": 0.11339203268289566,
      "learning_rate": 7.641971953099932e-05,
      "loss": 0.36674901843070984,
      "memory(GiB)": 78.33,
      "step": 3510,
      "token_acc": 0.8918313357088867,
      "train_speed(iter/s)": 0.032173
    },
    {
      "epoch": 0.6803274717822022,
      "grad_norm": 0.11586851626634598,
      "learning_rate": 7.633596319722123e-05,
      "loss": 0.3334874212741852,
      "memory(GiB)": 78.33,
      "step": 3511,
      "token_acc": 0.8987653673752917,
      "train_speed(iter/s)": 0.032174
    },
    {
      "epoch": 0.680521242067529,
      "grad_norm": 0.09603973478078842,
      "learning_rate": 7.625223711924251e-05,
      "loss": 0.3322643041610718,
      "memory(GiB)": 78.33,
      "step": 3512,
      "token_acc": 0.9002838519764508,
      "train_speed(iter/s)": 0.032174
    },
    {
      "epoch": 0.6807150123528557,
      "grad_norm": 0.09271606057882309,
      "learning_rate": 7.616854133145168e-05,
      "loss": 0.3315150737762451,
      "memory(GiB)": 78.33,
      "step": 3513,
      "token_acc": 0.9017941454202077,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.6809087826381824,
      "grad_norm": 0.10619889944791794,
      "learning_rate": 7.608487586822484e-05,
      "loss": 0.3216935396194458,
      "memory(GiB)": 78.33,
      "step": 3514,
      "token_acc": 0.9023508686672336,
      "train_speed(iter/s)": 0.032176
    },
    {
      "epoch": 0.6811025529235092,
      "grad_norm": 0.10402272641658783,
      "learning_rate": 7.600124076392569e-05,
      "loss": 0.3527657687664032,
      "memory(GiB)": 78.33,
      "step": 3515,
      "token_acc": 0.8967574223559994,
      "train_speed(iter/s)": 0.032177
    },
    {
      "epoch": 0.6812963232088359,
      "grad_norm": 0.09466518461704254,
      "learning_rate": 7.591763605290532e-05,
      "loss": 0.33050382137298584,
      "memory(GiB)": 78.33,
      "step": 3516,
      "token_acc": 0.9007306626354246,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.6814900934941627,
      "grad_norm": 0.10368410497903824,
      "learning_rate": 7.583406176950252e-05,
      "loss": 0.35686179995536804,
      "memory(GiB)": 78.33,
      "step": 3517,
      "token_acc": 0.8919963619827194,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.6816838637794894,
      "grad_norm": 0.1045493334531784,
      "learning_rate": 7.57505179480435e-05,
      "loss": 0.34543779492378235,
      "memory(GiB)": 78.33,
      "step": 3518,
      "token_acc": 0.89681269102023,
      "train_speed(iter/s)": 0.032179
    },
    {
      "epoch": 0.6818776340648162,
      "grad_norm": 0.10913045704364777,
      "learning_rate": 7.56670046228419e-05,
      "loss": 0.3246757388114929,
      "memory(GiB)": 78.33,
      "step": 3519,
      "token_acc": 0.901501614470673,
      "train_speed(iter/s)": 0.03218
    },
    {
      "epoch": 0.6820714043501429,
      "grad_norm": 0.10240820050239563,
      "learning_rate": 7.55835218281989e-05,
      "loss": 0.35042083263397217,
      "memory(GiB)": 78.33,
      "step": 3520,
      "token_acc": 0.8953667078178932,
      "train_speed(iter/s)": 0.032181
    },
    {
      "epoch": 0.6822651746354697,
      "grad_norm": 0.08281584084033966,
      "learning_rate": 7.55000695984031e-05,
      "loss": 0.28523337841033936,
      "memory(GiB)": 78.33,
      "step": 3521,
      "token_acc": 0.9129096006250287,
      "train_speed(iter/s)": 0.032182
    },
    {
      "epoch": 0.6824589449207964,
      "grad_norm": 0.10515667498111725,
      "learning_rate": 7.54166479677307e-05,
      "loss": 0.3321034014225006,
      "memory(GiB)": 78.33,
      "step": 3522,
      "token_acc": 0.8979174190441797,
      "train_speed(iter/s)": 0.032182
    },
    {
      "epoch": 0.6826527152061231,
      "grad_norm": 0.0955941304564476,
      "learning_rate": 7.5333256970445e-05,
      "loss": 0.3286648988723755,
      "memory(GiB)": 78.33,
      "step": 3523,
      "token_acc": 0.9015049247537623,
      "train_speed(iter/s)": 0.032183
    },
    {
      "epoch": 0.6828464854914499,
      "grad_norm": 0.09993550926446915,
      "learning_rate": 7.52498966407971e-05,
      "loss": 0.3626023232936859,
      "memory(GiB)": 78.33,
      "step": 3524,
      "token_acc": 0.8902110817941953,
      "train_speed(iter/s)": 0.032184
    },
    {
      "epoch": 0.6830402557767766,
      "grad_norm": 0.0978488177061081,
      "learning_rate": 7.516656701302527e-05,
      "loss": 0.3279617726802826,
      "memory(GiB)": 78.33,
      "step": 3525,
      "token_acc": 0.9018437274393292,
      "train_speed(iter/s)": 0.032185
    },
    {
      "epoch": 0.6832340260621034,
      "grad_norm": 0.11416416615247726,
      "learning_rate": 7.508326812135521e-05,
      "loss": 0.36074215173721313,
      "memory(GiB)": 78.33,
      "step": 3526,
      "token_acc": 0.891769873711183,
      "train_speed(iter/s)": 0.032186
    },
    {
      "epoch": 0.6834277963474301,
      "grad_norm": 0.10453812032938004,
      "learning_rate": 7.500000000000002e-05,
      "loss": 0.3612178862094879,
      "memory(GiB)": 78.33,
      "step": 3527,
      "token_acc": 0.8947847093932073,
      "train_speed(iter/s)": 0.032187
    },
    {
      "epoch": 0.6836215666327569,
      "grad_norm": 0.09948738664388657,
      "learning_rate": 7.49167626831601e-05,
      "loss": 0.34740930795669556,
      "memory(GiB)": 78.33,
      "step": 3528,
      "token_acc": 0.8971372214521804,
      "train_speed(iter/s)": 0.032187
    },
    {
      "epoch": 0.6838153369180836,
      "grad_norm": 0.09419828653335571,
      "learning_rate": 7.483355620502344e-05,
      "loss": 0.3287207782268524,
      "memory(GiB)": 78.33,
      "step": 3529,
      "token_acc": 0.9001759014951627,
      "train_speed(iter/s)": 0.032188
    },
    {
      "epoch": 0.6840091072034103,
      "grad_norm": 0.0952015146613121,
      "learning_rate": 7.475038059976492e-05,
      "loss": 0.3348086476325989,
      "memory(GiB)": 78.33,
      "step": 3530,
      "token_acc": 0.8985031033223805,
      "train_speed(iter/s)": 0.032189
    },
    {
      "epoch": 0.6842028774887371,
      "grad_norm": 0.09707789123058319,
      "learning_rate": 7.466723590154719e-05,
      "loss": 0.32328489422798157,
      "memory(GiB)": 78.33,
      "step": 3531,
      "token_acc": 0.9013299458170223,
      "train_speed(iter/s)": 0.03219
    },
    {
      "epoch": 0.6843966477740638,
      "grad_norm": 0.09463740885257721,
      "learning_rate": 7.458412214451992e-05,
      "loss": 0.3230316638946533,
      "memory(GiB)": 78.33,
      "step": 3532,
      "token_acc": 0.9013585495772858,
      "train_speed(iter/s)": 0.03219
    },
    {
      "epoch": 0.6845904180593906,
      "grad_norm": 0.09360900521278381,
      "learning_rate": 7.450103936282022e-05,
      "loss": 0.3111358880996704,
      "memory(GiB)": 78.33,
      "step": 3533,
      "token_acc": 0.9069836922693888,
      "train_speed(iter/s)": 0.032191
    },
    {
      "epoch": 0.6847841883447173,
      "grad_norm": 0.10580860823392868,
      "learning_rate": 7.441798759057238e-05,
      "loss": 0.3405246138572693,
      "memory(GiB)": 78.33,
      "step": 3534,
      "token_acc": 0.897554945054945,
      "train_speed(iter/s)": 0.032192
    },
    {
      "epoch": 0.6849779586300441,
      "grad_norm": 0.10613741725683212,
      "learning_rate": 7.433496686188794e-05,
      "loss": 0.3461315929889679,
      "memory(GiB)": 78.33,
      "step": 3535,
      "token_acc": 0.8966271542167963,
      "train_speed(iter/s)": 0.032193
    },
    {
      "epoch": 0.6851717289153708,
      "grad_norm": 0.1079440712928772,
      "learning_rate": 7.425197721086587e-05,
      "loss": 0.3699726462364197,
      "memory(GiB)": 78.33,
      "step": 3536,
      "token_acc": 0.8911746065279391,
      "train_speed(iter/s)": 0.032194
    },
    {
      "epoch": 0.6853654992006976,
      "grad_norm": 0.09260550886392593,
      "learning_rate": 7.416901867159219e-05,
      "loss": 0.29420557618141174,
      "memory(GiB)": 78.33,
      "step": 3537,
      "token_acc": 0.9136413222721161,
      "train_speed(iter/s)": 0.032194
    },
    {
      "epoch": 0.6855592694860243,
      "grad_norm": 0.10380339622497559,
      "learning_rate": 7.408609127814019e-05,
      "loss": 0.346148818731308,
      "memory(GiB)": 78.33,
      "step": 3538,
      "token_acc": 0.8991937710530675,
      "train_speed(iter/s)": 0.032195
    },
    {
      "epoch": 0.685753039771351,
      "grad_norm": 0.09581930190324783,
      "learning_rate": 7.400319506457039e-05,
      "loss": 0.3182103633880615,
      "memory(GiB)": 78.33,
      "step": 3539,
      "token_acc": 0.9065670467951239,
      "train_speed(iter/s)": 0.032196
    },
    {
      "epoch": 0.6859468100566778,
      "grad_norm": 0.11510851234197617,
      "learning_rate": 7.39203300649305e-05,
      "loss": 0.3609999716281891,
      "memory(GiB)": 78.33,
      "step": 3540,
      "token_acc": 0.892073988583549,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.6861405803420045,
      "grad_norm": 0.11120989918708801,
      "learning_rate": 7.383749631325538e-05,
      "loss": 0.3553025722503662,
      "memory(GiB)": 78.33,
      "step": 3541,
      "token_acc": 0.8961922030825022,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.6863343506273313,
      "grad_norm": 0.1011735275387764,
      "learning_rate": 7.375469384356705e-05,
      "loss": 0.33759036660194397,
      "memory(GiB)": 78.33,
      "step": 3542,
      "token_acc": 0.8999340949033392,
      "train_speed(iter/s)": 0.032198
    },
    {
      "epoch": 0.686528120912658,
      "grad_norm": 0.11077761650085449,
      "learning_rate": 7.367192268987479e-05,
      "loss": 0.38028085231781006,
      "memory(GiB)": 78.33,
      "step": 3543,
      "token_acc": 0.8891689052590767,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.6867218911979848,
      "grad_norm": 0.10262443870306015,
      "learning_rate": 7.35891828861749e-05,
      "loss": 0.35626545548439026,
      "memory(GiB)": 78.33,
      "step": 3544,
      "token_acc": 0.8935800392486684,
      "train_speed(iter/s)": 0.0322
    },
    {
      "epoch": 0.6869156614833115,
      "grad_norm": 0.09497423470020294,
      "learning_rate": 7.350647446645084e-05,
      "loss": 0.3267231583595276,
      "memory(GiB)": 78.33,
      "step": 3545,
      "token_acc": 0.9022584504935687,
      "train_speed(iter/s)": 0.032201
    },
    {
      "epoch": 0.6871094317686383,
      "grad_norm": 0.09908232092857361,
      "learning_rate": 7.342379746467317e-05,
      "loss": 0.3486153185367584,
      "memory(GiB)": 78.33,
      "step": 3546,
      "token_acc": 0.8962997272399773,
      "train_speed(iter/s)": 0.032201
    },
    {
      "epoch": 0.687303202053965,
      "grad_norm": 0.10710590332746506,
      "learning_rate": 7.334115191479958e-05,
      "loss": 0.35125356912612915,
      "memory(GiB)": 78.33,
      "step": 3547,
      "token_acc": 0.8947753459732976,
      "train_speed(iter/s)": 0.032202
    },
    {
      "epoch": 0.6874969723392917,
      "grad_norm": 0.09003735333681107,
      "learning_rate": 7.325853785077478e-05,
      "loss": 0.3048159182071686,
      "memory(GiB)": 78.33,
      "step": 3548,
      "token_acc": 0.907744732028764,
      "train_speed(iter/s)": 0.032203
    },
    {
      "epoch": 0.6876907426246185,
      "grad_norm": 0.11373089253902435,
      "learning_rate": 7.317595530653055e-05,
      "loss": 0.3735848367214203,
      "memory(GiB)": 78.33,
      "step": 3549,
      "token_acc": 0.8891771564011403,
      "train_speed(iter/s)": 0.032204
    },
    {
      "epoch": 0.6878845129099452,
      "grad_norm": 0.10183597356081009,
      "learning_rate": 7.30934043159859e-05,
      "loss": 0.3513960838317871,
      "memory(GiB)": 78.33,
      "step": 3550,
      "token_acc": 0.8948632459312839,
      "train_speed(iter/s)": 0.032205
    },
    {
      "epoch": 0.688078283195272,
      "grad_norm": 0.10609740763902664,
      "learning_rate": 7.301088491304664e-05,
      "loss": 0.32522451877593994,
      "memory(GiB)": 78.33,
      "step": 3551,
      "token_acc": 0.9038668263767314,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.6882720534805987,
      "grad_norm": 0.10257891565561295,
      "learning_rate": 7.292839713160572e-05,
      "loss": 0.34625405073165894,
      "memory(GiB)": 78.33,
      "step": 3552,
      "token_acc": 0.8976674404121611,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.6884658237659255,
      "grad_norm": 0.09951046854257584,
      "learning_rate": 7.28459410055431e-05,
      "loss": 0.32491254806518555,
      "memory(GiB)": 78.33,
      "step": 3553,
      "token_acc": 0.9030543047794716,
      "train_speed(iter/s)": 0.032207
    },
    {
      "epoch": 0.6886595940512522,
      "grad_norm": 0.10017047822475433,
      "learning_rate": 7.276351656872567e-05,
      "loss": 0.3391422927379608,
      "memory(GiB)": 78.33,
      "step": 3554,
      "token_acc": 0.8989633042619714,
      "train_speed(iter/s)": 0.032208
    },
    {
      "epoch": 0.688853364336579,
      "grad_norm": 0.09386296570301056,
      "learning_rate": 7.268112385500751e-05,
      "loss": 0.3156377077102661,
      "memory(GiB)": 78.33,
      "step": 3555,
      "token_acc": 0.9045845939027228,
      "train_speed(iter/s)": 0.032209
    },
    {
      "epoch": 0.6890471346219057,
      "grad_norm": 0.0977504625916481,
      "learning_rate": 7.259876289822932e-05,
      "loss": 0.2982153594493866,
      "memory(GiB)": 78.33,
      "step": 3556,
      "token_acc": 0.9096926576077448,
      "train_speed(iter/s)": 0.03221
    },
    {
      "epoch": 0.6892409049072324,
      "grad_norm": 0.10796993225812912,
      "learning_rate": 7.25164337322191e-05,
      "loss": 0.33437347412109375,
      "memory(GiB)": 78.33,
      "step": 3557,
      "token_acc": 0.9011005762019451,
      "train_speed(iter/s)": 0.03221
    },
    {
      "epoch": 0.6894346751925592,
      "grad_norm": 0.09524484723806381,
      "learning_rate": 7.243413639079164e-05,
      "loss": 0.3119055926799774,
      "memory(GiB)": 78.33,
      "step": 3558,
      "token_acc": 0.9061108966936833,
      "train_speed(iter/s)": 0.032211
    },
    {
      "epoch": 0.6896284454778859,
      "grad_norm": 0.09212891757488251,
      "learning_rate": 7.235187090774861e-05,
      "loss": 0.31174278259277344,
      "memory(GiB)": 78.33,
      "step": 3559,
      "token_acc": 0.9064973268274461,
      "train_speed(iter/s)": 0.032212
    },
    {
      "epoch": 0.6898222157632127,
      "grad_norm": 0.09447701275348663,
      "learning_rate": 7.22696373168787e-05,
      "loss": 0.3395736515522003,
      "memory(GiB)": 78.33,
      "step": 3560,
      "token_acc": 0.899402390438247,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.6900159860485394,
      "grad_norm": 0.09689807891845703,
      "learning_rate": 7.218743565195736e-05,
      "loss": 0.33678123354911804,
      "memory(GiB)": 78.33,
      "step": 3561,
      "token_acc": 0.8996187046549806,
      "train_speed(iter/s)": 0.032214
    },
    {
      "epoch": 0.6902097563338662,
      "grad_norm": 0.10127212852239609,
      "learning_rate": 7.210526594674724e-05,
      "loss": 0.35242077708244324,
      "memory(GiB)": 78.33,
      "step": 3562,
      "token_acc": 0.8943310420532657,
      "train_speed(iter/s)": 0.032214
    },
    {
      "epoch": 0.6904035266191929,
      "grad_norm": 0.0959462970495224,
      "learning_rate": 7.202312823499738e-05,
      "loss": 0.30481332540512085,
      "memory(GiB)": 78.33,
      "step": 3563,
      "token_acc": 0.9080480651942592,
      "train_speed(iter/s)": 0.032215
    },
    {
      "epoch": 0.6905972969045197,
      "grad_norm": 0.09734071046113968,
      "learning_rate": 7.194102255044415e-05,
      "loss": 0.31086277961730957,
      "memory(GiB)": 78.33,
      "step": 3564,
      "token_acc": 0.9077373688064437,
      "train_speed(iter/s)": 0.032216
    },
    {
      "epoch": 0.6907910671898464,
      "grad_norm": 0.10370344668626785,
      "learning_rate": 7.185894892681048e-05,
      "loss": 0.3449724316596985,
      "memory(GiB)": 78.33,
      "step": 3565,
      "token_acc": 0.8976901609168102,
      "train_speed(iter/s)": 0.032217
    },
    {
      "epoch": 0.6909848374751731,
      "grad_norm": 0.10260160267353058,
      "learning_rate": 7.17769073978062e-05,
      "loss": 0.3641132712364197,
      "memory(GiB)": 78.33,
      "step": 3566,
      "token_acc": 0.8912745666262711,
      "train_speed(iter/s)": 0.032218
    },
    {
      "epoch": 0.6911786077604999,
      "grad_norm": 0.10054480284452438,
      "learning_rate": 7.169489799712799e-05,
      "loss": 0.3622581660747528,
      "memory(GiB)": 78.33,
      "step": 3567,
      "token_acc": 0.8936806148590948,
      "train_speed(iter/s)": 0.032218
    },
    {
      "epoch": 0.6913723780458266,
      "grad_norm": 0.0974610224366188,
      "learning_rate": 7.161292075845926e-05,
      "loss": 0.3504323363304138,
      "memory(GiB)": 78.33,
      "step": 3568,
      "token_acc": 0.8961522654885441,
      "train_speed(iter/s)": 0.032219
    },
    {
      "epoch": 0.6915661483311534,
      "grad_norm": 0.10981731861829758,
      "learning_rate": 7.153097571547038e-05,
      "loss": 0.3665009140968323,
      "memory(GiB)": 78.33,
      "step": 3569,
      "token_acc": 0.8888823146559375,
      "train_speed(iter/s)": 0.03222
    },
    {
      "epoch": 0.6917599186164801,
      "grad_norm": 0.10058823972940445,
      "learning_rate": 7.144906290181832e-05,
      "loss": 0.3594769239425659,
      "memory(GiB)": 78.33,
      "step": 3570,
      "token_acc": 0.8944776275804255,
      "train_speed(iter/s)": 0.032221
    },
    {
      "epoch": 0.6919536889018069,
      "grad_norm": 0.11253203451633453,
      "learning_rate": 7.136718235114686e-05,
      "loss": 0.36248651146888733,
      "memory(GiB)": 78.33,
      "step": 3571,
      "token_acc": 0.8918202731428905,
      "train_speed(iter/s)": 0.032222
    },
    {
      "epoch": 0.6921474591871336,
      "grad_norm": 0.11419709771871567,
      "learning_rate": 7.128533409708656e-05,
      "loss": 0.3548899292945862,
      "memory(GiB)": 78.33,
      "step": 3572,
      "token_acc": 0.8954283004658169,
      "train_speed(iter/s)": 0.032222
    },
    {
      "epoch": 0.6923412294724604,
      "grad_norm": 0.10111390799283981,
      "learning_rate": 7.120351817325469e-05,
      "loss": 0.33687660098075867,
      "memory(GiB)": 78.33,
      "step": 3573,
      "token_acc": 0.8996950958880688,
      "train_speed(iter/s)": 0.032223
    },
    {
      "epoch": 0.6925349997577871,
      "grad_norm": 0.10509663820266724,
      "learning_rate": 7.112173461325525e-05,
      "loss": 0.36578071117401123,
      "memory(GiB)": 78.33,
      "step": 3574,
      "token_acc": 0.889935521532624,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.6927287700431138,
      "grad_norm": 0.11243990808725357,
      "learning_rate": 7.10399834506789e-05,
      "loss": 0.3173210620880127,
      "memory(GiB)": 78.33,
      "step": 3575,
      "token_acc": 0.9059147392523605,
      "train_speed(iter/s)": 0.032225
    },
    {
      "epoch": 0.6929225403284406,
      "grad_norm": 0.10139571875333786,
      "learning_rate": 7.095826471910313e-05,
      "loss": 0.3282420039176941,
      "memory(GiB)": 78.33,
      "step": 3576,
      "token_acc": 0.902722545120832,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.6931163106137673,
      "grad_norm": 0.0943324863910675,
      "learning_rate": 7.087657845209196e-05,
      "loss": 0.32712027430534363,
      "memory(GiB)": 78.33,
      "step": 3577,
      "token_acc": 0.9013381123058543,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.6933100808990941,
      "grad_norm": 0.13730654120445251,
      "learning_rate": 7.079492468319618e-05,
      "loss": 0.3298265337944031,
      "memory(GiB)": 78.33,
      "step": 3578,
      "token_acc": 0.901703923714241,
      "train_speed(iter/s)": 0.032227
    },
    {
      "epoch": 0.6935038511844208,
      "grad_norm": 0.10465666651725769,
      "learning_rate": 7.071330344595314e-05,
      "loss": 0.34704989194869995,
      "memory(GiB)": 78.33,
      "step": 3579,
      "token_acc": 0.8968518114360541,
      "train_speed(iter/s)": 0.032228
    },
    {
      "epoch": 0.6936976214697476,
      "grad_norm": 0.10772595554590225,
      "learning_rate": 7.063171477388688e-05,
      "loss": 0.35015714168548584,
      "memory(GiB)": 78.33,
      "step": 3580,
      "token_acc": 0.894221193443329,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.6938913917550743,
      "grad_norm": 0.10779722779989243,
      "learning_rate": 7.055015870050809e-05,
      "loss": 0.3421753942966461,
      "memory(GiB)": 78.33,
      "step": 3581,
      "token_acc": 0.8975127263354598,
      "train_speed(iter/s)": 0.03223
    },
    {
      "epoch": 0.694085162040401,
      "grad_norm": 0.09361686557531357,
      "learning_rate": 7.046863525931395e-05,
      "loss": 0.3193609416484833,
      "memory(GiB)": 78.33,
      "step": 3582,
      "token_acc": 0.9052085463095482,
      "train_speed(iter/s)": 0.03223
    },
    {
      "epoch": 0.6942789323257279,
      "grad_norm": 0.10432402044534683,
      "learning_rate": 7.038714448378846e-05,
      "loss": 0.3353957235813141,
      "memory(GiB)": 78.33,
      "step": 3583,
      "token_acc": 0.8987451070688465,
      "train_speed(iter/s)": 0.032231
    },
    {
      "epoch": 0.6944727026110546,
      "grad_norm": 0.09659415483474731,
      "learning_rate": 7.030568640740201e-05,
      "loss": 0.3202442526817322,
      "memory(GiB)": 78.33,
      "step": 3584,
      "token_acc": 0.9025401069518717,
      "train_speed(iter/s)": 0.032232
    },
    {
      "epoch": 0.6946664728963814,
      "grad_norm": 0.09830014407634735,
      "learning_rate": 7.022426106361163e-05,
      "loss": 0.31335949897766113,
      "memory(GiB)": 78.33,
      "step": 3585,
      "token_acc": 0.906624622120296,
      "train_speed(iter/s)": 0.032233
    },
    {
      "epoch": 0.6948602431817081,
      "grad_norm": 0.09905708581209183,
      "learning_rate": 7.014286848586088e-05,
      "loss": 0.3123447895050049,
      "memory(GiB)": 78.33,
      "step": 3586,
      "token_acc": 0.9041099546303709,
      "train_speed(iter/s)": 0.032233
    },
    {
      "epoch": 0.6950540134670349,
      "grad_norm": 0.09144581854343414,
      "learning_rate": 7.006150870757989e-05,
      "loss": 0.3135398030281067,
      "memory(GiB)": 78.33,
      "step": 3587,
      "token_acc": 0.9057776084938225,
      "train_speed(iter/s)": 0.032234
    },
    {
      "epoch": 0.6952477837523616,
      "grad_norm": 0.10966409742832184,
      "learning_rate": 6.99801817621853e-05,
      "loss": 0.3710702657699585,
      "memory(GiB)": 78.33,
      "step": 3588,
      "token_acc": 0.8925160936022211,
      "train_speed(iter/s)": 0.032235
    },
    {
      "epoch": 0.6954415540376884,
      "grad_norm": 0.09838887304067612,
      "learning_rate": 6.989888768308024e-05,
      "loss": 0.3230520486831665,
      "memory(GiB)": 78.33,
      "step": 3589,
      "token_acc": 0.901048865761444,
      "train_speed(iter/s)": 0.032236
    },
    {
      "epoch": 0.6956353243230151,
      "grad_norm": 0.0968567356467247,
      "learning_rate": 6.981762650365443e-05,
      "loss": 0.34151309728622437,
      "memory(GiB)": 78.33,
      "step": 3590,
      "token_acc": 0.8988749172733289,
      "train_speed(iter/s)": 0.032237
    },
    {
      "epoch": 0.6958290946083419,
      "grad_norm": 0.09684547781944275,
      "learning_rate": 6.973639825728401e-05,
      "loss": 0.3080379366874695,
      "memory(GiB)": 78.33,
      "step": 3591,
      "token_acc": 0.9065901444686325,
      "train_speed(iter/s)": 0.032237
    },
    {
      "epoch": 0.6960228648936686,
      "grad_norm": 0.09826657176017761,
      "learning_rate": 6.965520297733161e-05,
      "loss": 0.3160586953163147,
      "memory(GiB)": 78.33,
      "step": 3592,
      "token_acc": 0.9051656151419558,
      "train_speed(iter/s)": 0.032238
    },
    {
      "epoch": 0.6962166351789953,
      "grad_norm": 0.10795781761407852,
      "learning_rate": 6.957404069714629e-05,
      "loss": 0.3557586073875427,
      "memory(GiB)": 78.33,
      "step": 3593,
      "token_acc": 0.8951461742901105,
      "train_speed(iter/s)": 0.032239
    },
    {
      "epoch": 0.6964104054643221,
      "grad_norm": 0.10063203424215317,
      "learning_rate": 6.949291145006353e-05,
      "loss": 0.31674492359161377,
      "memory(GiB)": 78.33,
      "step": 3594,
      "token_acc": 0.9046947410840218,
      "train_speed(iter/s)": 0.03224
    },
    {
      "epoch": 0.6966041757496488,
      "grad_norm": 0.10855990648269653,
      "learning_rate": 6.941181526940546e-05,
      "loss": 0.3493427634239197,
      "memory(GiB)": 78.33,
      "step": 3595,
      "token_acc": 0.895584936843348,
      "train_speed(iter/s)": 0.032241
    },
    {
      "epoch": 0.6967979460349756,
      "grad_norm": 0.09571418911218643,
      "learning_rate": 6.933075218848022e-05,
      "loss": 0.3080763518810272,
      "memory(GiB)": 78.33,
      "step": 3596,
      "token_acc": 0.9050366723085603,
      "train_speed(iter/s)": 0.032241
    },
    {
      "epoch": 0.6969917163203023,
      "grad_norm": 0.0985865592956543,
      "learning_rate": 6.924972224058278e-05,
      "loss": 0.3179894983768463,
      "memory(GiB)": 78.33,
      "step": 3597,
      "token_acc": 0.9029965847747053,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.6971854866056291,
      "grad_norm": 0.10527476668357849,
      "learning_rate": 6.916872545899427e-05,
      "loss": 0.34273964166641235,
      "memory(GiB)": 78.33,
      "step": 3598,
      "token_acc": 0.8985260287978188,
      "train_speed(iter/s)": 0.032243
    },
    {
      "epoch": 0.6973792568909558,
      "grad_norm": 0.10250242054462433,
      "learning_rate": 6.908776187698222e-05,
      "loss": 0.34843844175338745,
      "memory(GiB)": 78.33,
      "step": 3599,
      "token_acc": 0.8955212090892335,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.6975730271762826,
      "grad_norm": 0.10623464733362198,
      "learning_rate": 6.900683152780059e-05,
      "loss": 0.34149301052093506,
      "memory(GiB)": 78.33,
      "step": 3600,
      "token_acc": 0.8972149929701756,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.6977667974616093,
      "grad_norm": 0.10202284902334213,
      "learning_rate": 6.892593444468954e-05,
      "loss": 0.34131869673728943,
      "memory(GiB)": 78.33,
      "step": 3601,
      "token_acc": 0.8982538616521155,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.697960567746936,
      "grad_norm": 0.24693740904331207,
      "learning_rate": 6.884507066087584e-05,
      "loss": 0.341753214597702,
      "memory(GiB)": 78.33,
      "step": 3602,
      "token_acc": 0.8990424814565071,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.6981543380322628,
      "grad_norm": 0.11156189441680908,
      "learning_rate": 6.87642402095723e-05,
      "loss": 0.3889864385128021,
      "memory(GiB)": 78.33,
      "step": 3603,
      "token_acc": 0.885625468967735,
      "train_speed(iter/s)": 0.032243
    },
    {
      "epoch": 0.6983481083175895,
      "grad_norm": 0.10035014897584915,
      "learning_rate": 6.868344312397823e-05,
      "loss": 0.3296211063861847,
      "memory(GiB)": 78.33,
      "step": 3604,
      "token_acc": 0.9014160142449337,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.6985418786029163,
      "grad_norm": 0.09430637955665588,
      "learning_rate": 6.860267943727912e-05,
      "loss": 0.31993281841278076,
      "memory(GiB)": 78.33,
      "step": 3605,
      "token_acc": 0.9017905151091515,
      "train_speed(iter/s)": 0.032245
    },
    {
      "epoch": 0.698735648888243,
      "grad_norm": 0.09148237109184265,
      "learning_rate": 6.852194918264679e-05,
      "loss": 0.3246019780635834,
      "memory(GiB)": 78.33,
      "step": 3606,
      "token_acc": 0.9024793779101373,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.6989294191735698,
      "grad_norm": 0.10248222947120667,
      "learning_rate": 6.844125239323933e-05,
      "loss": 0.33933860063552856,
      "memory(GiB)": 78.33,
      "step": 3607,
      "token_acc": 0.8989493117584775,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.6991231894588965,
      "grad_norm": 0.09796436131000519,
      "learning_rate": 6.836058910220102e-05,
      "loss": 0.3447071611881256,
      "memory(GiB)": 78.33,
      "step": 3608,
      "token_acc": 0.8983826213034023,
      "train_speed(iter/s)": 0.032247
    },
    {
      "epoch": 0.6993169597442233,
      "grad_norm": 0.10065195709466934,
      "learning_rate": 6.827995934266259e-05,
      "loss": 0.35696181654930115,
      "memory(GiB)": 78.33,
      "step": 3609,
      "token_acc": 0.8937146709536897,
      "train_speed(iter/s)": 0.032248
    },
    {
      "epoch": 0.69951073002955,
      "grad_norm": 0.08694633841514587,
      "learning_rate": 6.819936314774074e-05,
      "loss": 0.3089888095855713,
      "memory(GiB)": 78.33,
      "step": 3610,
      "token_acc": 0.9059485912481589,
      "train_speed(iter/s)": 0.032249
    },
    {
      "epoch": 0.6997045003148767,
      "grad_norm": 0.11966582387685776,
      "learning_rate": 6.81188005505385e-05,
      "loss": 0.3933137357234955,
      "memory(GiB)": 78.33,
      "step": 3611,
      "token_acc": 0.8836271567891973,
      "train_speed(iter/s)": 0.032249
    },
    {
      "epoch": 0.6998982706002035,
      "grad_norm": 0.11115437746047974,
      "learning_rate": 6.803827158414512e-05,
      "loss": 0.36456286907196045,
      "memory(GiB)": 78.33,
      "step": 3612,
      "token_acc": 0.8943506748392162,
      "train_speed(iter/s)": 0.03225
    },
    {
      "epoch": 0.7000920408855302,
      "grad_norm": 0.09643102437257767,
      "learning_rate": 6.795777628163599e-05,
      "loss": 0.3220784366130829,
      "memory(GiB)": 78.33,
      "step": 3613,
      "token_acc": 0.9028170546008266,
      "train_speed(iter/s)": 0.032251
    },
    {
      "epoch": 0.700285811170857,
      "grad_norm": 0.09812808781862259,
      "learning_rate": 6.78773146760727e-05,
      "loss": 0.3644194006919861,
      "memory(GiB)": 78.33,
      "step": 3614,
      "token_acc": 0.8943334712050802,
      "train_speed(iter/s)": 0.032252
    },
    {
      "epoch": 0.7004795814561837,
      "grad_norm": 0.08493607491254807,
      "learning_rate": 6.779688680050296e-05,
      "loss": 0.2795659899711609,
      "memory(GiB)": 78.33,
      "step": 3615,
      "token_acc": 0.9142221903464104,
      "train_speed(iter/s)": 0.032253
    },
    {
      "epoch": 0.7006733517415105,
      "grad_norm": 0.10760082304477692,
      "learning_rate": 6.771649268796073e-05,
      "loss": 0.3571077883243561,
      "memory(GiB)": 78.33,
      "step": 3616,
      "token_acc": 0.8955498133792708,
      "train_speed(iter/s)": 0.032253
    },
    {
      "epoch": 0.7008671220268372,
      "grad_norm": 0.10084082186222076,
      "learning_rate": 6.7636132371466e-05,
      "loss": 0.3274276554584503,
      "memory(GiB)": 78.33,
      "step": 3617,
      "token_acc": 0.9002256477176956,
      "train_speed(iter/s)": 0.032254
    },
    {
      "epoch": 0.701060892312164,
      "grad_norm": 0.10685470700263977,
      "learning_rate": 6.755580588402492e-05,
      "loss": 0.34894925355911255,
      "memory(GiB)": 78.33,
      "step": 3618,
      "token_acc": 0.8940826727066817,
      "train_speed(iter/s)": 0.032255
    },
    {
      "epoch": 0.7012546625974907,
      "grad_norm": 0.11531514674425125,
      "learning_rate": 6.74755132586297e-05,
      "loss": 0.40001946687698364,
      "memory(GiB)": 78.33,
      "step": 3619,
      "token_acc": 0.885813818463325,
      "train_speed(iter/s)": 0.032256
    },
    {
      "epoch": 0.7014484328828174,
      "grad_norm": 0.1036171242594719,
      "learning_rate": 6.739525452825871e-05,
      "loss": 0.35469454526901245,
      "memory(GiB)": 78.33,
      "step": 3620,
      "token_acc": 0.8936814976351638,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.7016422031681442,
      "grad_norm": 0.10101023316383362,
      "learning_rate": 6.731502972587637e-05,
      "loss": 0.31730735301971436,
      "memory(GiB)": 78.33,
      "step": 3621,
      "token_acc": 0.9038396016249508,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.7018359734534709,
      "grad_norm": 0.10922136157751083,
      "learning_rate": 6.72348388844331e-05,
      "loss": 0.3624739944934845,
      "memory(GiB)": 78.33,
      "step": 3622,
      "token_acc": 0.8922535011220635,
      "train_speed(iter/s)": 0.032258
    },
    {
      "epoch": 0.7020297437387977,
      "grad_norm": 0.09917715191841125,
      "learning_rate": 6.715468203686553e-05,
      "loss": 0.3212898373603821,
      "memory(GiB)": 78.33,
      "step": 3623,
      "token_acc": 0.9037098911227606,
      "train_speed(iter/s)": 0.032259
    },
    {
      "epoch": 0.7022235140241244,
      "grad_norm": 0.1009821966290474,
      "learning_rate": 6.70745592160962e-05,
      "loss": 0.3310352563858032,
      "memory(GiB)": 78.33,
      "step": 3624,
      "token_acc": 0.9000106797671811,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 0.7024172843094512,
      "grad_norm": 0.1044231653213501,
      "learning_rate": 6.699447045503368e-05,
      "loss": 0.3453122079372406,
      "memory(GiB)": 78.33,
      "step": 3625,
      "token_acc": 0.8964209722297998,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 0.7026110545947779,
      "grad_norm": 0.10527423024177551,
      "learning_rate": 6.69144157865726e-05,
      "loss": 0.34135711193084717,
      "memory(GiB)": 78.33,
      "step": 3626,
      "token_acc": 0.9015418633837328,
      "train_speed(iter/s)": 0.032261
    },
    {
      "epoch": 0.7028048248801047,
      "grad_norm": 0.10108046233654022,
      "learning_rate": 6.683439524359351e-05,
      "loss": 0.3644823431968689,
      "memory(GiB)": 78.33,
      "step": 3627,
      "token_acc": 0.8922698922698923,
      "train_speed(iter/s)": 0.032262
    },
    {
      "epoch": 0.7029985951654314,
      "grad_norm": 0.10416973382234573,
      "learning_rate": 6.675440885896313e-05,
      "loss": 0.3607743978500366,
      "memory(GiB)": 78.33,
      "step": 3628,
      "token_acc": 0.8915753781950965,
      "train_speed(iter/s)": 0.032263
    },
    {
      "epoch": 0.7031923654507581,
      "grad_norm": 0.09729216992855072,
      "learning_rate": 6.66744566655338e-05,
      "loss": 0.3052827715873718,
      "memory(GiB)": 78.33,
      "step": 3629,
      "token_acc": 0.907312711319753,
      "train_speed(iter/s)": 0.032264
    },
    {
      "epoch": 0.7033861357360849,
      "grad_norm": 0.11203377693891525,
      "learning_rate": 6.659453869614426e-05,
      "loss": 0.35420122742652893,
      "memory(GiB)": 78.33,
      "step": 3630,
      "token_acc": 0.8935774567300815,
      "train_speed(iter/s)": 0.032265
    },
    {
      "epoch": 0.7035799060214116,
      "grad_norm": 0.09870520234107971,
      "learning_rate": 6.651465498361885e-05,
      "loss": 0.34778672456741333,
      "memory(GiB)": 78.33,
      "step": 3631,
      "token_acc": 0.8956560099647083,
      "train_speed(iter/s)": 0.032265
    },
    {
      "epoch": 0.7037736763067384,
      "grad_norm": 0.10209493339061737,
      "learning_rate": 6.643480556076796e-05,
      "loss": 0.3381112515926361,
      "memory(GiB)": 78.33,
      "step": 3632,
      "token_acc": 0.9000748406410488,
      "train_speed(iter/s)": 0.032266
    },
    {
      "epoch": 0.7039674465920651,
      "grad_norm": 0.10792220383882523,
      "learning_rate": 6.635499046038794e-05,
      "loss": 0.3817068338394165,
      "memory(GiB)": 78.33,
      "step": 3633,
      "token_acc": 0.8868650435828074,
      "train_speed(iter/s)": 0.032267
    },
    {
      "epoch": 0.7041612168773919,
      "grad_norm": 0.1084747463464737,
      "learning_rate": 6.627520971526088e-05,
      "loss": 0.36603009700775146,
      "memory(GiB)": 78.33,
      "step": 3634,
      "token_acc": 0.8899790989317232,
      "train_speed(iter/s)": 0.032268
    },
    {
      "epoch": 0.7043549871627186,
      "grad_norm": 0.10230504721403122,
      "learning_rate": 6.619546335815503e-05,
      "loss": 0.3367150127887726,
      "memory(GiB)": 78.33,
      "step": 3635,
      "token_acc": 0.898678290999525,
      "train_speed(iter/s)": 0.032268
    },
    {
      "epoch": 0.7045487574480453,
      "grad_norm": 0.0944862887263298,
      "learning_rate": 6.61157514218243e-05,
      "loss": 0.33380186557769775,
      "memory(GiB)": 78.33,
      "step": 3636,
      "token_acc": 0.9002516989680343,
      "train_speed(iter/s)": 0.032269
    },
    {
      "epoch": 0.7047425277333721,
      "grad_norm": 0.1018281877040863,
      "learning_rate": 6.603607393900852e-05,
      "loss": 0.3347662687301636,
      "memory(GiB)": 78.33,
      "step": 3637,
      "token_acc": 0.8994067307187483,
      "train_speed(iter/s)": 0.03227
    },
    {
      "epoch": 0.7049362980186988,
      "grad_norm": 0.10647895187139511,
      "learning_rate": 6.595643094243335e-05,
      "loss": 0.3350790739059448,
      "memory(GiB)": 78.33,
      "step": 3638,
      "token_acc": 0.9001211906239839,
      "train_speed(iter/s)": 0.032271
    },
    {
      "epoch": 0.7051300683040256,
      "grad_norm": 0.10115791112184525,
      "learning_rate": 6.587682246481036e-05,
      "loss": 0.3523610234260559,
      "memory(GiB)": 78.33,
      "step": 3639,
      "token_acc": 0.8971207273951843,
      "train_speed(iter/s)": 0.032272
    },
    {
      "epoch": 0.7053238385893523,
      "grad_norm": 0.10161978751420975,
      "learning_rate": 6.579724853883684e-05,
      "loss": 0.3423236012458801,
      "memory(GiB)": 78.33,
      "step": 3640,
      "token_acc": 0.90029210201101,
      "train_speed(iter/s)": 0.032272
    },
    {
      "epoch": 0.7055176088746791,
      "grad_norm": 0.09845045208930969,
      "learning_rate": 6.571770919719592e-05,
      "loss": 0.32784488797187805,
      "memory(GiB)": 78.33,
      "step": 3641,
      "token_acc": 0.9032830751813925,
      "train_speed(iter/s)": 0.032273
    },
    {
      "epoch": 0.7057113791600058,
      "grad_norm": 0.09614937752485275,
      "learning_rate": 6.563820447255663e-05,
      "loss": 0.33316075801849365,
      "memory(GiB)": 78.33,
      "step": 3642,
      "token_acc": 0.8991082090500266,
      "train_speed(iter/s)": 0.032274
    },
    {
      "epoch": 0.7059051494453326,
      "grad_norm": 0.10282139480113983,
      "learning_rate": 6.555873439757366e-05,
      "loss": 0.32991480827331543,
      "memory(GiB)": 78.33,
      "step": 3643,
      "token_acc": 0.9022564374834086,
      "train_speed(iter/s)": 0.032274
    },
    {
      "epoch": 0.7060989197306593,
      "grad_norm": 0.11456986516714096,
      "learning_rate": 6.547929900488749e-05,
      "loss": 0.36666494607925415,
      "memory(GiB)": 78.33,
      "step": 3644,
      "token_acc": 0.889168765743073,
      "train_speed(iter/s)": 0.032275
    },
    {
      "epoch": 0.706292690015986,
      "grad_norm": 0.10417237877845764,
      "learning_rate": 6.539989832712439e-05,
      "loss": 0.36223533749580383,
      "memory(GiB)": 78.33,
      "step": 3645,
      "token_acc": 0.8901788429112419,
      "train_speed(iter/s)": 0.032276
    },
    {
      "epoch": 0.7064864603013128,
      "grad_norm": 0.09668152034282684,
      "learning_rate": 6.532053239689631e-05,
      "loss": 0.32515013217926025,
      "memory(GiB)": 78.33,
      "step": 3646,
      "token_acc": 0.901520462181827,
      "train_speed(iter/s)": 0.032277
    },
    {
      "epoch": 0.7066802305866395,
      "grad_norm": 0.10676920413970947,
      "learning_rate": 6.524120124680104e-05,
      "loss": 0.35744109749794006,
      "memory(GiB)": 78.33,
      "step": 3647,
      "token_acc": 0.8961123110151188,
      "train_speed(iter/s)": 0.032278
    },
    {
      "epoch": 0.7068740008719663,
      "grad_norm": 0.10558585822582245,
      "learning_rate": 6.51619049094219e-05,
      "loss": 0.3541412353515625,
      "memory(GiB)": 78.33,
      "step": 3648,
      "token_acc": 0.8950543561476038,
      "train_speed(iter/s)": 0.032278
    },
    {
      "epoch": 0.707067771157293,
      "grad_norm": 0.10470271855592728,
      "learning_rate": 6.508264341732815e-05,
      "loss": 0.36191800236701965,
      "memory(GiB)": 78.33,
      "step": 3649,
      "token_acc": 0.8917385076885658,
      "train_speed(iter/s)": 0.032279
    },
    {
      "epoch": 0.7072615414426198,
      "grad_norm": 0.10651751607656479,
      "learning_rate": 6.500341680307457e-05,
      "loss": 0.34071311354637146,
      "memory(GiB)": 78.33,
      "step": 3650,
      "token_acc": 0.8990374215877136,
      "train_speed(iter/s)": 0.03228
    },
    {
      "epoch": 0.7074553117279465,
      "grad_norm": 0.10043511539697647,
      "learning_rate": 6.492422509920167e-05,
      "loss": 0.32280540466308594,
      "memory(GiB)": 78.33,
      "step": 3651,
      "token_acc": 0.9044051318031472,
      "train_speed(iter/s)": 0.03228
    },
    {
      "epoch": 0.7076490820132733,
      "grad_norm": 0.11648175865411758,
      "learning_rate": 6.484506833823559e-05,
      "loss": 0.36728546023368835,
      "memory(GiB)": 78.33,
      "step": 3652,
      "token_acc": 0.890401807018102,
      "train_speed(iter/s)": 0.032281
    },
    {
      "epoch": 0.7078428522986,
      "grad_norm": 0.10505972057580948,
      "learning_rate": 6.476594655268814e-05,
      "loss": 0.3796302080154419,
      "memory(GiB)": 78.33,
      "step": 3653,
      "token_acc": 0.8889887273321353,
      "train_speed(iter/s)": 0.032282
    },
    {
      "epoch": 0.7080366225839267,
      "grad_norm": 0.1008792519569397,
      "learning_rate": 6.468685977505676e-05,
      "loss": 0.31763315200805664,
      "memory(GiB)": 78.33,
      "step": 3654,
      "token_acc": 0.905323467447716,
      "train_speed(iter/s)": 0.032283
    },
    {
      "epoch": 0.7082303928692535,
      "grad_norm": 0.09355062991380692,
      "learning_rate": 6.460780803782448e-05,
      "loss": 0.31916338205337524,
      "memory(GiB)": 78.33,
      "step": 3655,
      "token_acc": 0.9061786058790678,
      "train_speed(iter/s)": 0.032283
    },
    {
      "epoch": 0.7084241631545802,
      "grad_norm": 0.10963544994592667,
      "learning_rate": 6.452879137346007e-05,
      "loss": 0.3544541895389557,
      "memory(GiB)": 78.33,
      "step": 3656,
      "token_acc": 0.8935794884272636,
      "train_speed(iter/s)": 0.032284
    },
    {
      "epoch": 0.708617933439907,
      "grad_norm": 0.09566653519868851,
      "learning_rate": 6.444980981441775e-05,
      "loss": 0.3308408558368683,
      "memory(GiB)": 78.33,
      "step": 3657,
      "token_acc": 0.9017979944484733,
      "train_speed(iter/s)": 0.032285
    },
    {
      "epoch": 0.7088117037252337,
      "grad_norm": 0.10379898548126221,
      "learning_rate": 6.437086339313735e-05,
      "loss": 0.32990795373916626,
      "memory(GiB)": 78.33,
      "step": 3658,
      "token_acc": 0.8986880466472303,
      "train_speed(iter/s)": 0.032286
    },
    {
      "epoch": 0.7090054740105605,
      "grad_norm": 0.11237087845802307,
      "learning_rate": 6.429195214204428e-05,
      "loss": 0.3397267162799835,
      "memory(GiB)": 78.33,
      "step": 3659,
      "token_acc": 0.8990950108233576,
      "train_speed(iter/s)": 0.032286
    },
    {
      "epoch": 0.7091992442958872,
      "grad_norm": 0.10298223793506622,
      "learning_rate": 6.421307609354957e-05,
      "loss": 0.32215797901153564,
      "memory(GiB)": 78.33,
      "step": 3660,
      "token_acc": 0.9034365070024715,
      "train_speed(iter/s)": 0.032287
    },
    {
      "epoch": 0.709393014581214,
      "grad_norm": 0.11267991364002228,
      "learning_rate": 6.413423528004968e-05,
      "loss": 0.3442709445953369,
      "memory(GiB)": 78.33,
      "step": 3661,
      "token_acc": 0.8958262848706671,
      "train_speed(iter/s)": 0.032288
    },
    {
      "epoch": 0.7095867848665407,
      "grad_norm": 0.09616924822330475,
      "learning_rate": 6.40554297339266e-05,
      "loss": 0.3290918469429016,
      "memory(GiB)": 78.33,
      "step": 3662,
      "token_acc": 0.9001304801670146,
      "train_speed(iter/s)": 0.032289
    },
    {
      "epoch": 0.7097805551518674,
      "grad_norm": 0.09400229156017303,
      "learning_rate": 6.3976659487548e-05,
      "loss": 0.3193153440952301,
      "memory(GiB)": 78.33,
      "step": 3663,
      "token_acc": 0.9033483612785573,
      "train_speed(iter/s)": 0.032289
    },
    {
      "epoch": 0.7099743254371942,
      "grad_norm": 0.0955566018819809,
      "learning_rate": 6.38979245732669e-05,
      "loss": 0.3172321915626526,
      "memory(GiB)": 78.33,
      "step": 3664,
      "token_acc": 0.905316643945966,
      "train_speed(iter/s)": 0.03229
    },
    {
      "epoch": 0.7101680957225209,
      "grad_norm": 0.1006002202630043,
      "learning_rate": 6.381922502342182e-05,
      "loss": 0.31889939308166504,
      "memory(GiB)": 78.33,
      "step": 3665,
      "token_acc": 0.9029685900544152,
      "train_speed(iter/s)": 0.032291
    },
    {
      "epoch": 0.7103618660078477,
      "grad_norm": 0.10536913573741913,
      "learning_rate": 6.37405608703368e-05,
      "loss": 0.3331918716430664,
      "memory(GiB)": 78.33,
      "step": 3666,
      "token_acc": 0.9007023825919295,
      "train_speed(iter/s)": 0.032292
    },
    {
      "epoch": 0.7105556362931744,
      "grad_norm": 0.09231302887201309,
      "learning_rate": 6.366193214632123e-05,
      "loss": 0.32499587535858154,
      "memory(GiB)": 78.33,
      "step": 3667,
      "token_acc": 0.903846596533229,
      "train_speed(iter/s)": 0.032292
    },
    {
      "epoch": 0.7107494065785012,
      "grad_norm": 0.1009543314576149,
      "learning_rate": 6.35833388836702e-05,
      "loss": 0.3330029845237732,
      "memory(GiB)": 78.33,
      "step": 3668,
      "token_acc": 0.9015039132436442,
      "train_speed(iter/s)": 0.032293
    },
    {
      "epoch": 0.7109431768638279,
      "grad_norm": 0.10333767533302307,
      "learning_rate": 6.350478111466399e-05,
      "loss": 0.3277633786201477,
      "memory(GiB)": 78.33,
      "step": 3669,
      "token_acc": 0.9013294198895028,
      "train_speed(iter/s)": 0.032294
    },
    {
      "epoch": 0.7111369471491547,
      "grad_norm": 0.12599217891693115,
      "learning_rate": 6.342625887156839e-05,
      "loss": 0.38569512963294983,
      "memory(GiB)": 78.33,
      "step": 3670,
      "token_acc": 0.8859112900889416,
      "train_speed(iter/s)": 0.032295
    },
    {
      "epoch": 0.7113307174344814,
      "grad_norm": 0.10265874862670898,
      "learning_rate": 6.334777218663461e-05,
      "loss": 0.36148056387901306,
      "memory(GiB)": 78.33,
      "step": 3671,
      "token_acc": 0.8917856569870046,
      "train_speed(iter/s)": 0.032295
    },
    {
      "epoch": 0.7115244877198081,
      "grad_norm": 0.1083201915025711,
      "learning_rate": 6.326932109209922e-05,
      "loss": 0.34301403164863586,
      "memory(GiB)": 78.33,
      "step": 3672,
      "token_acc": 0.8956853082741233,
      "train_speed(iter/s)": 0.032296
    },
    {
      "epoch": 0.7117182580051349,
      "grad_norm": 0.11269722878932953,
      "learning_rate": 6.319090562018419e-05,
      "loss": 0.34771838784217834,
      "memory(GiB)": 78.33,
      "step": 3673,
      "token_acc": 0.8975308282074455,
      "train_speed(iter/s)": 0.032297
    },
    {
      "epoch": 0.7119120282904616,
      "grad_norm": 0.1110716462135315,
      "learning_rate": 6.311252580309682e-05,
      "loss": 0.361664742231369,
      "memory(GiB)": 78.33,
      "step": 3674,
      "token_acc": 0.8944661718966526,
      "train_speed(iter/s)": 0.032298
    },
    {
      "epoch": 0.7121057985757884,
      "grad_norm": 0.11600884050130844,
      "learning_rate": 6.303418167302994e-05,
      "loss": 0.38586270809173584,
      "memory(GiB)": 78.33,
      "step": 3675,
      "token_acc": 0.8842967627548681,
      "train_speed(iter/s)": 0.032298
    },
    {
      "epoch": 0.7122995688611151,
      "grad_norm": 0.1069970428943634,
      "learning_rate": 6.295587326216149e-05,
      "loss": 0.34455546736717224,
      "memory(GiB)": 78.33,
      "step": 3676,
      "token_acc": 0.898117085260072,
      "train_speed(iter/s)": 0.032299
    },
    {
      "epoch": 0.7124933391464419,
      "grad_norm": 0.10031075775623322,
      "learning_rate": 6.287760060265485e-05,
      "loss": 0.33782869577407837,
      "memory(GiB)": 78.33,
      "step": 3677,
      "token_acc": 0.8992902208201893,
      "train_speed(iter/s)": 0.0323
    },
    {
      "epoch": 0.7126871094317686,
      "grad_norm": 0.10248647630214691,
      "learning_rate": 6.279936372665874e-05,
      "loss": 0.33945292234420776,
      "memory(GiB)": 78.33,
      "step": 3678,
      "token_acc": 0.8977860913237329,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.7128808797170954,
      "grad_norm": 0.10595156252384186,
      "learning_rate": 6.27211626663071e-05,
      "loss": 0.36181285977363586,
      "memory(GiB)": 78.33,
      "step": 3679,
      "token_acc": 0.8937350415317471,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.7130746500024221,
      "grad_norm": 0.09816795587539673,
      "learning_rate": 6.264299745371922e-05,
      "loss": 0.33072012662887573,
      "memory(GiB)": 78.33,
      "step": 3680,
      "token_acc": 0.9027671022290545,
      "train_speed(iter/s)": 0.032302
    },
    {
      "epoch": 0.7132684202877488,
      "grad_norm": 0.09826884418725967,
      "learning_rate": 6.256486812099961e-05,
      "loss": 0.32538434863090515,
      "memory(GiB)": 78.33,
      "step": 3681,
      "token_acc": 0.901966256501332,
      "train_speed(iter/s)": 0.032303
    },
    {
      "epoch": 0.7134621905730756,
      "grad_norm": 0.09893721342086792,
      "learning_rate": 6.248677470023819e-05,
      "loss": 0.31855225563049316,
      "memory(GiB)": 78.33,
      "step": 3682,
      "token_acc": 0.9033751256414326,
      "train_speed(iter/s)": 0.032304
    },
    {
      "epoch": 0.7136559608584023,
      "grad_norm": 0.10306079685688019,
      "learning_rate": 6.240871722350998e-05,
      "loss": 0.3382839262485504,
      "memory(GiB)": 78.33,
      "step": 3683,
      "token_acc": 0.8988074461896451,
      "train_speed(iter/s)": 0.032304
    },
    {
      "epoch": 0.7138497311437291,
      "grad_norm": 0.10051840543746948,
      "learning_rate": 6.233069572287527e-05,
      "loss": 0.31791266798973083,
      "memory(GiB)": 78.33,
      "step": 3684,
      "token_acc": 0.90387971539804,
      "train_speed(iter/s)": 0.032305
    },
    {
      "epoch": 0.7140435014290558,
      "grad_norm": 0.11294718831777573,
      "learning_rate": 6.22527102303796e-05,
      "loss": 0.3825894892215729,
      "memory(GiB)": 78.33,
      "step": 3685,
      "token_acc": 0.8891299885640559,
      "train_speed(iter/s)": 0.032306
    },
    {
      "epoch": 0.7142372717143826,
      "grad_norm": 0.0918063074350357,
      "learning_rate": 6.217476077805369e-05,
      "loss": 0.31590160727500916,
      "memory(GiB)": 78.33,
      "step": 3686,
      "token_acc": 0.9046886822386566,
      "train_speed(iter/s)": 0.032307
    },
    {
      "epoch": 0.7144310419997093,
      "grad_norm": 0.09278496354818344,
      "learning_rate": 6.209684739791347e-05,
      "loss": 0.3343659043312073,
      "memory(GiB)": 78.33,
      "step": 3687,
      "token_acc": 0.8985444692228377,
      "train_speed(iter/s)": 0.032307
    },
    {
      "epoch": 0.714624812285036,
      "grad_norm": 0.11030339449644089,
      "learning_rate": 6.201897012196005e-05,
      "loss": 0.33175480365753174,
      "memory(GiB)": 78.33,
      "step": 3688,
      "token_acc": 0.9004147113390811,
      "train_speed(iter/s)": 0.032308
    },
    {
      "epoch": 0.7148185825703628,
      "grad_norm": 0.09746501594781876,
      "learning_rate": 6.194112898217978e-05,
      "loss": 0.3286242187023163,
      "memory(GiB)": 78.33,
      "step": 3689,
      "token_acc": 0.9018423513356415,
      "train_speed(iter/s)": 0.032309
    },
    {
      "epoch": 0.7150123528556895,
      "grad_norm": 0.10385642945766449,
      "learning_rate": 6.186332401054406e-05,
      "loss": 0.3613872528076172,
      "memory(GiB)": 78.33,
      "step": 3690,
      "token_acc": 0.891545143039286,
      "train_speed(iter/s)": 0.03231
    },
    {
      "epoch": 0.7152061231410163,
      "grad_norm": 0.10428830236196518,
      "learning_rate": 6.17855552390095e-05,
      "loss": 0.3543146252632141,
      "memory(GiB)": 78.33,
      "step": 3691,
      "token_acc": 0.8936885245901639,
      "train_speed(iter/s)": 0.03231
    },
    {
      "epoch": 0.715399893426343,
      "grad_norm": 0.09830185770988464,
      "learning_rate": 6.170782269951783e-05,
      "loss": 0.2993859648704529,
      "memory(GiB)": 78.33,
      "step": 3692,
      "token_acc": 0.9107328114684523,
      "train_speed(iter/s)": 0.032311
    },
    {
      "epoch": 0.7155936637116698,
      "grad_norm": 0.10561109334230423,
      "learning_rate": 6.163012642399587e-05,
      "loss": 0.3516858220100403,
      "memory(GiB)": 78.33,
      "step": 3693,
      "token_acc": 0.8974372357638708,
      "train_speed(iter/s)": 0.032312
    },
    {
      "epoch": 0.7157874339969965,
      "grad_norm": 0.0973869115114212,
      "learning_rate": 6.155246644435558e-05,
      "loss": 0.3680126965045929,
      "memory(GiB)": 78.33,
      "step": 3694,
      "token_acc": 0.8894600330110823,
      "train_speed(iter/s)": 0.032312
    },
    {
      "epoch": 0.7159812042823233,
      "grad_norm": 0.09466208517551422,
      "learning_rate": 6.147484279249396e-05,
      "loss": 0.32024601101875305,
      "memory(GiB)": 78.33,
      "step": 3695,
      "token_acc": 0.9047858675212543,
      "train_speed(iter/s)": 0.032313
    },
    {
      "epoch": 0.71617497456765,
      "grad_norm": 0.10687924921512604,
      "learning_rate": 6.13972555002932e-05,
      "loss": 0.33103132247924805,
      "memory(GiB)": 78.33,
      "step": 3696,
      "token_acc": 0.9001981178801387,
      "train_speed(iter/s)": 0.032314
    },
    {
      "epoch": 0.7163687448529767,
      "grad_norm": 0.09712281078100204,
      "learning_rate": 6.131970459962046e-05,
      "loss": 0.3220095634460449,
      "memory(GiB)": 78.33,
      "step": 3697,
      "token_acc": 0.9038578331704998,
      "train_speed(iter/s)": 0.032315
    },
    {
      "epoch": 0.7165625151383035,
      "grad_norm": 0.11033158749341965,
      "learning_rate": 6.124219012232798e-05,
      "loss": 0.36744678020477295,
      "memory(GiB)": 78.33,
      "step": 3698,
      "token_acc": 0.8917987391739822,
      "train_speed(iter/s)": 0.032315
    },
    {
      "epoch": 0.7167562854236302,
      "grad_norm": 0.09989314526319504,
      "learning_rate": 6.1164712100253e-05,
      "loss": 0.33150678873062134,
      "memory(GiB)": 78.33,
      "step": 3699,
      "token_acc": 0.8993894415495552,
      "train_speed(iter/s)": 0.032316
    },
    {
      "epoch": 0.716950055708957,
      "grad_norm": 0.09984250366687775,
      "learning_rate": 6.108727056521783e-05,
      "loss": 0.30265843868255615,
      "memory(GiB)": 78.33,
      "step": 3700,
      "token_acc": 0.9069431182649025,
      "train_speed(iter/s)": 0.032317
    },
    {
      "epoch": 0.7171438259942837,
      "grad_norm": 0.09416959434747696,
      "learning_rate": 6.100986554902988e-05,
      "loss": 0.3297763168811798,
      "memory(GiB)": 78.33,
      "step": 3701,
      "token_acc": 0.9030869331246025,
      "train_speed(iter/s)": 0.032318
    },
    {
      "epoch": 0.7173375962796105,
      "grad_norm": 0.10124395787715912,
      "learning_rate": 6.0932497083481404e-05,
      "loss": 0.3468609154224396,
      "memory(GiB)": 78.33,
      "step": 3702,
      "token_acc": 0.8960674907226953,
      "train_speed(iter/s)": 0.032318
    },
    {
      "epoch": 0.7175313665649372,
      "grad_norm": 0.10942904651165009,
      "learning_rate": 6.08551652003497e-05,
      "loss": 0.3687775135040283,
      "memory(GiB)": 78.33,
      "step": 3703,
      "token_acc": 0.890260192821466,
      "train_speed(iter/s)": 0.032319
    },
    {
      "epoch": 0.7177251368502641,
      "grad_norm": 0.0992734357714653,
      "learning_rate": 6.077786993139706e-05,
      "loss": 0.34562569856643677,
      "memory(GiB)": 78.33,
      "step": 3704,
      "token_acc": 0.8980179107439656,
      "train_speed(iter/s)": 0.03232
    },
    {
      "epoch": 0.7179189071355908,
      "grad_norm": 0.09762708097696304,
      "learning_rate": 6.070061130837074e-05,
      "loss": 0.3264696002006531,
      "memory(GiB)": 78.33,
      "step": 3705,
      "token_acc": 0.9032848106089475,
      "train_speed(iter/s)": 0.032321
    },
    {
      "epoch": 0.7181126774209176,
      "grad_norm": 0.1083015501499176,
      "learning_rate": 6.0623389363002925e-05,
      "loss": 0.3502708375453949,
      "memory(GiB)": 78.33,
      "step": 3706,
      "token_acc": 0.8949533057084544,
      "train_speed(iter/s)": 0.032321
    },
    {
      "epoch": 0.7183064477062443,
      "grad_norm": 0.1071147471666336,
      "learning_rate": 6.054620412701069e-05,
      "loss": 0.33522728085517883,
      "memory(GiB)": 78.33,
      "step": 3707,
      "token_acc": 0.89790950744559,
      "train_speed(iter/s)": 0.032322
    },
    {
      "epoch": 0.718500217991571,
      "grad_norm": 0.08930321782827377,
      "learning_rate": 6.0469055632096186e-05,
      "loss": 0.2909185290336609,
      "memory(GiB)": 78.33,
      "step": 3708,
      "token_acc": 0.9108066749253079,
      "train_speed(iter/s)": 0.032323
    },
    {
      "epoch": 0.7186939882768978,
      "grad_norm": 0.09592770040035248,
      "learning_rate": 6.039194390994632e-05,
      "loss": 0.3235243260860443,
      "memory(GiB)": 78.33,
      "step": 3709,
      "token_acc": 0.9030587138380254,
      "train_speed(iter/s)": 0.032323
    },
    {
      "epoch": 0.7188877585622245,
      "grad_norm": 0.11995202302932739,
      "learning_rate": 6.031486899223295e-05,
      "loss": 0.39347177743911743,
      "memory(GiB)": 78.33,
      "step": 3710,
      "token_acc": 0.8830674846625767,
      "train_speed(iter/s)": 0.032324
    },
    {
      "epoch": 0.7190815288475513,
      "grad_norm": 0.11326012015342712,
      "learning_rate": 6.0237830910612816e-05,
      "loss": 0.36819055676460266,
      "memory(GiB)": 78.33,
      "step": 3711,
      "token_acc": 0.8888051139157516,
      "train_speed(iter/s)": 0.032325
    },
    {
      "epoch": 0.719275299132878,
      "grad_norm": 0.099938303232193,
      "learning_rate": 6.0160829696727535e-05,
      "loss": 0.3334549367427826,
      "memory(GiB)": 78.33,
      "step": 3712,
      "token_acc": 0.9019944353433705,
      "train_speed(iter/s)": 0.032326
    },
    {
      "epoch": 0.7194690694182048,
      "grad_norm": 0.10359780490398407,
      "learning_rate": 6.008386538220357e-05,
      "loss": 0.3458460569381714,
      "memory(GiB)": 78.33,
      "step": 3713,
      "token_acc": 0.8952391251453135,
      "train_speed(iter/s)": 0.032326
    },
    {
      "epoch": 0.7196628397035315,
      "grad_norm": 0.09753096848726273,
      "learning_rate": 6.0006937998652174e-05,
      "loss": 0.3544035255908966,
      "memory(GiB)": 78.33,
      "step": 3714,
      "token_acc": 0.8957159256940096,
      "train_speed(iter/s)": 0.032327
    },
    {
      "epoch": 0.7198566099888583,
      "grad_norm": 0.10259803384542465,
      "learning_rate": 5.993004757766961e-05,
      "loss": 0.3598324954509735,
      "memory(GiB)": 78.33,
      "step": 3715,
      "token_acc": 0.8931450536543698,
      "train_speed(iter/s)": 0.032328
    },
    {
      "epoch": 0.720050380274185,
      "grad_norm": 0.10409785062074661,
      "learning_rate": 5.9853194150836776e-05,
      "loss": 0.34401389956474304,
      "memory(GiB)": 78.33,
      "step": 3716,
      "token_acc": 0.8969972797656414,
      "train_speed(iter/s)": 0.032328
    },
    {
      "epoch": 0.7202441505595117,
      "grad_norm": 0.10885158181190491,
      "learning_rate": 5.977637774971945e-05,
      "loss": 0.34587976336479187,
      "memory(GiB)": 78.33,
      "step": 3717,
      "token_acc": 0.8954475087437647,
      "train_speed(iter/s)": 0.032329
    },
    {
      "epoch": 0.7204379208448385,
      "grad_norm": 0.0933261513710022,
      "learning_rate": 5.9699598405868184e-05,
      "loss": 0.30791398882865906,
      "memory(GiB)": 78.33,
      "step": 3718,
      "token_acc": 0.9086631971823582,
      "train_speed(iter/s)": 0.03233
    },
    {
      "epoch": 0.7206316911301652,
      "grad_norm": 0.10228514671325684,
      "learning_rate": 5.962285615081831e-05,
      "loss": 0.32020220160484314,
      "memory(GiB)": 78.33,
      "step": 3719,
      "token_acc": 0.9048829981248876,
      "train_speed(iter/s)": 0.032331
    },
    {
      "epoch": 0.720825461415492,
      "grad_norm": 0.10021097958087921,
      "learning_rate": 5.9546151016089935e-05,
      "loss": 0.3199518024921417,
      "memory(GiB)": 78.33,
      "step": 3720,
      "token_acc": 0.9031611499255526,
      "train_speed(iter/s)": 0.032331
    },
    {
      "epoch": 0.7210192317008187,
      "grad_norm": 0.0932881161570549,
      "learning_rate": 5.946948303318788e-05,
      "loss": 0.30388596653938293,
      "memory(GiB)": 78.33,
      "step": 3721,
      "token_acc": 0.9094827586206896,
      "train_speed(iter/s)": 0.032332
    },
    {
      "epoch": 0.7212130019861455,
      "grad_norm": 0.09003807604312897,
      "learning_rate": 5.939285223360185e-05,
      "loss": 0.3002552092075348,
      "memory(GiB)": 78.33,
      "step": 3722,
      "token_acc": 0.9080838912957857,
      "train_speed(iter/s)": 0.032333
    },
    {
      "epoch": 0.7214067722714722,
      "grad_norm": 0.08356068283319473,
      "learning_rate": 5.931625864880612e-05,
      "loss": 0.2760453224182129,
      "memory(GiB)": 78.33,
      "step": 3723,
      "token_acc": 0.9180010863661053,
      "train_speed(iter/s)": 0.032333
    },
    {
      "epoch": 0.721600542556799,
      "grad_norm": 0.09654924273490906,
      "learning_rate": 5.9239702310259726e-05,
      "loss": 0.3264361023902893,
      "memory(GiB)": 78.33,
      "step": 3724,
      "token_acc": 0.9010648901012432,
      "train_speed(iter/s)": 0.032334
    },
    {
      "epoch": 0.7217943128421257,
      "grad_norm": 0.1332063376903534,
      "learning_rate": 5.916318324940643e-05,
      "loss": 0.3769557476043701,
      "memory(GiB)": 78.33,
      "step": 3725,
      "token_acc": 0.8894034827713968,
      "train_speed(iter/s)": 0.032335
    },
    {
      "epoch": 0.7219880831274524,
      "grad_norm": 0.10913225263357162,
      "learning_rate": 5.9086701497674636e-05,
      "loss": 0.3281722962856293,
      "memory(GiB)": 78.33,
      "step": 3726,
      "token_acc": 0.9037930929929353,
      "train_speed(iter/s)": 0.032335
    },
    {
      "epoch": 0.7221818534127792,
      "grad_norm": 0.0964721292257309,
      "learning_rate": 5.9010257086477465e-05,
      "loss": 0.31369680166244507,
      "memory(GiB)": 78.33,
      "step": 3727,
      "token_acc": 0.9058114664301737,
      "train_speed(iter/s)": 0.032336
    },
    {
      "epoch": 0.7223756236981059,
      "grad_norm": 0.09561733901500702,
      "learning_rate": 5.893385004721265e-05,
      "loss": 0.31413954496383667,
      "memory(GiB)": 78.33,
      "step": 3728,
      "token_acc": 0.9042112486339492,
      "train_speed(iter/s)": 0.032337
    },
    {
      "epoch": 0.7225693939834327,
      "grad_norm": 0.09478065371513367,
      "learning_rate": 5.885748041126273e-05,
      "loss": 0.32629698514938354,
      "memory(GiB)": 78.33,
      "step": 3729,
      "token_acc": 0.9058655593044381,
      "train_speed(iter/s)": 0.032338
    },
    {
      "epoch": 0.7227631642687594,
      "grad_norm": 0.09354487806558609,
      "learning_rate": 5.8781148209994684e-05,
      "loss": 0.31036120653152466,
      "memory(GiB)": 78.33,
      "step": 3730,
      "token_acc": 0.9080553710987384,
      "train_speed(iter/s)": 0.032338
    },
    {
      "epoch": 0.7229569345540862,
      "grad_norm": 0.10508932918310165,
      "learning_rate": 5.870485347476023e-05,
      "loss": 0.3312036693096161,
      "memory(GiB)": 78.33,
      "step": 3731,
      "token_acc": 0.9002311682453766,
      "train_speed(iter/s)": 0.032339
    },
    {
      "epoch": 0.7231507048394129,
      "grad_norm": 0.09594292938709259,
      "learning_rate": 5.862859623689564e-05,
      "loss": 0.33407771587371826,
      "memory(GiB)": 78.33,
      "step": 3732,
      "token_acc": 0.8988324898531843,
      "train_speed(iter/s)": 0.03234
    },
    {
      "epoch": 0.7233444751247396,
      "grad_norm": 0.09291541576385498,
      "learning_rate": 5.855237652772182e-05,
      "loss": 0.30694496631622314,
      "memory(GiB)": 78.33,
      "step": 3733,
      "token_acc": 0.9065133287402618,
      "train_speed(iter/s)": 0.03234
    },
    {
      "epoch": 0.7235382454100664,
      "grad_norm": 0.10045888274908066,
      "learning_rate": 5.847619437854425e-05,
      "loss": 0.3402080535888672,
      "memory(GiB)": 78.33,
      "step": 3734,
      "token_acc": 0.8967975233486823,
      "train_speed(iter/s)": 0.032341
    },
    {
      "epoch": 0.7237320156953931,
      "grad_norm": 0.10660912096500397,
      "learning_rate": 5.8400049820652944e-05,
      "loss": 0.36076852679252625,
      "memory(GiB)": 78.33,
      "step": 3735,
      "token_acc": 0.8916170432607337,
      "train_speed(iter/s)": 0.032342
    },
    {
      "epoch": 0.7239257859807199,
      "grad_norm": 0.10676801949739456,
      "learning_rate": 5.8323942885322605e-05,
      "loss": 0.3120606243610382,
      "memory(GiB)": 78.33,
      "step": 3736,
      "token_acc": 0.907629155894226,
      "train_speed(iter/s)": 0.032343
    },
    {
      "epoch": 0.7241195562660466,
      "grad_norm": 0.10707957297563553,
      "learning_rate": 5.8247873603812364e-05,
      "loss": 0.35654643177986145,
      "memory(GiB)": 78.33,
      "step": 3737,
      "token_acc": 0.8950555588218002,
      "train_speed(iter/s)": 0.032343
    },
    {
      "epoch": 0.7243133265513734,
      "grad_norm": 0.11162468791007996,
      "learning_rate": 5.8171842007365906e-05,
      "loss": 0.3537403345108032,
      "memory(GiB)": 78.33,
      "step": 3738,
      "token_acc": 0.895508172545164,
      "train_speed(iter/s)": 0.032344
    },
    {
      "epoch": 0.7245070968367001,
      "grad_norm": 0.10323680192232132,
      "learning_rate": 5.809584812721145e-05,
      "loss": 0.34856048226356506,
      "memory(GiB)": 78.33,
      "step": 3739,
      "token_acc": 0.8983748949285514,
      "train_speed(iter/s)": 0.032345
    },
    {
      "epoch": 0.7247008671220269,
      "grad_norm": 0.1021660715341568,
      "learning_rate": 5.801989199456167e-05,
      "loss": 0.3216470181941986,
      "memory(GiB)": 78.33,
      "step": 3740,
      "token_acc": 0.9034610943704413,
      "train_speed(iter/s)": 0.032346
    },
    {
      "epoch": 0.7248946374073536,
      "grad_norm": 0.09654397517442703,
      "learning_rate": 5.794397364061391e-05,
      "loss": 0.33012211322784424,
      "memory(GiB)": 78.33,
      "step": 3741,
      "token_acc": 0.9008311162385515,
      "train_speed(iter/s)": 0.032346
    },
    {
      "epoch": 0.7250884076926803,
      "grad_norm": 0.11007391661405563,
      "learning_rate": 5.786809309654982e-05,
      "loss": 0.3086721897125244,
      "memory(GiB)": 78.33,
      "step": 3742,
      "token_acc": 0.9068431837791199,
      "train_speed(iter/s)": 0.032347
    },
    {
      "epoch": 0.7252821779780071,
      "grad_norm": 0.10440944880247116,
      "learning_rate": 5.7792250393535575e-05,
      "loss": 0.3401702344417572,
      "memory(GiB)": 78.33,
      "step": 3743,
      "token_acc": 0.8993664717348928,
      "train_speed(iter/s)": 0.032348
    },
    {
      "epoch": 0.7254759482633338,
      "grad_norm": 0.11093514412641525,
      "learning_rate": 5.771644556272181e-05,
      "loss": 0.3583501875400543,
      "memory(GiB)": 78.33,
      "step": 3744,
      "token_acc": 0.8940592570311325,
      "train_speed(iter/s)": 0.032349
    },
    {
      "epoch": 0.7256697185486606,
      "grad_norm": 0.10289919376373291,
      "learning_rate": 5.7640678635243606e-05,
      "loss": 0.3268412947654724,
      "memory(GiB)": 78.33,
      "step": 3745,
      "token_acc": 0.9035391502660698,
      "train_speed(iter/s)": 0.032349
    },
    {
      "epoch": 0.7258634888339873,
      "grad_norm": 0.095411017537117,
      "learning_rate": 5.756494964222047e-05,
      "loss": 0.32108423113822937,
      "memory(GiB)": 78.33,
      "step": 3746,
      "token_acc": 0.9025717749632419,
      "train_speed(iter/s)": 0.03235
    },
    {
      "epoch": 0.7260572591193141,
      "grad_norm": 0.10212237387895584,
      "learning_rate": 5.748925861475631e-05,
      "loss": 0.35688862204551697,
      "memory(GiB)": 78.33,
      "step": 3747,
      "token_acc": 0.8916946508368809,
      "train_speed(iter/s)": 0.032351
    },
    {
      "epoch": 0.7262510294046408,
      "grad_norm": 0.11504284292459488,
      "learning_rate": 5.741360558393953e-05,
      "loss": 0.3791845738887787,
      "memory(GiB)": 78.33,
      "step": 3748,
      "token_acc": 0.8880055524007824,
      "train_speed(iter/s)": 0.032352
    },
    {
      "epoch": 0.7264447996899676,
      "grad_norm": 0.09927377104759216,
      "learning_rate": 5.733799058084284e-05,
      "loss": 0.33466947078704834,
      "memory(GiB)": 78.33,
      "step": 3749,
      "token_acc": 0.899494293530826,
      "train_speed(iter/s)": 0.032352
    },
    {
      "epoch": 0.7266385699752943,
      "grad_norm": 0.09528510272502899,
      "learning_rate": 5.7262413636523343e-05,
      "loss": 0.3107728362083435,
      "memory(GiB)": 78.33,
      "step": 3750,
      "token_acc": 0.9060210094798873,
      "train_speed(iter/s)": 0.032353
    },
    {
      "epoch": 0.726832340260621,
      "grad_norm": 0.09576904028654099,
      "learning_rate": 5.718687478202252e-05,
      "loss": 0.3216266930103302,
      "memory(GiB)": 78.33,
      "step": 3751,
      "token_acc": 0.9044459023011707,
      "train_speed(iter/s)": 0.032354
    },
    {
      "epoch": 0.7270261105459478,
      "grad_norm": 0.10115578025579453,
      "learning_rate": 5.7111374048366204e-05,
      "loss": 0.3389231264591217,
      "memory(GiB)": 78.33,
      "step": 3752,
      "token_acc": 0.8973095364944059,
      "train_speed(iter/s)": 0.032354
    },
    {
      "epoch": 0.7272198808312745,
      "grad_norm": 0.10029512643814087,
      "learning_rate": 5.703591146656458e-05,
      "loss": 0.31098711490631104,
      "memory(GiB)": 78.33,
      "step": 3753,
      "token_acc": 0.906193576626361,
      "train_speed(iter/s)": 0.032355
    },
    {
      "epoch": 0.7274136511166013,
      "grad_norm": 0.10533425211906433,
      "learning_rate": 5.696048706761211e-05,
      "loss": 0.33651649951934814,
      "memory(GiB)": 78.33,
      "step": 3754,
      "token_acc": 0.8975294455616202,
      "train_speed(iter/s)": 0.032356
    },
    {
      "epoch": 0.727607421401928,
      "grad_norm": 0.09996731579303741,
      "learning_rate": 5.688510088248772e-05,
      "loss": 0.31849730014801025,
      "memory(GiB)": 78.33,
      "step": 3755,
      "token_acc": 0.9020276110740658,
      "train_speed(iter/s)": 0.032357
    },
    {
      "epoch": 0.7278011916872548,
      "grad_norm": 0.10525427013635635,
      "learning_rate": 5.6809752942154505e-05,
      "loss": 0.3365975618362427,
      "memory(GiB)": 78.33,
      "step": 3756,
      "token_acc": 0.8993686078063035,
      "train_speed(iter/s)": 0.032357
    },
    {
      "epoch": 0.7279949619725815,
      "grad_norm": 0.11113554239273071,
      "learning_rate": 5.673444327755986e-05,
      "loss": 0.36573588848114014,
      "memory(GiB)": 78.33,
      "step": 3757,
      "token_acc": 0.8919042258663178,
      "train_speed(iter/s)": 0.032358
    },
    {
      "epoch": 0.7281887322579083,
      "grad_norm": 0.11570943146944046,
      "learning_rate": 5.6659171919635504e-05,
      "loss": 0.3126344382762909,
      "memory(GiB)": 78.33,
      "step": 3758,
      "token_acc": 0.9062464954581138,
      "train_speed(iter/s)": 0.032359
    },
    {
      "epoch": 0.728382502543235,
      "grad_norm": 0.11019326746463776,
      "learning_rate": 5.6583938899297404e-05,
      "loss": 0.3683561086654663,
      "memory(GiB)": 78.33,
      "step": 3759,
      "token_acc": 0.8913200775664959,
      "train_speed(iter/s)": 0.032359
    },
    {
      "epoch": 0.7285762728285617,
      "grad_norm": 0.11105161160230637,
      "learning_rate": 5.650874424744579e-05,
      "loss": 0.34568020701408386,
      "memory(GiB)": 78.33,
      "step": 3760,
      "token_acc": 0.898310245090577,
      "train_speed(iter/s)": 0.03236
    },
    {
      "epoch": 0.7287700431138885,
      "grad_norm": 0.10734663158655167,
      "learning_rate": 5.643358799496508e-05,
      "loss": 0.3574303090572357,
      "memory(GiB)": 78.33,
      "step": 3761,
      "token_acc": 0.8934692112234416,
      "train_speed(iter/s)": 0.032361
    },
    {
      "epoch": 0.7289638133992152,
      "grad_norm": 0.1054970771074295,
      "learning_rate": 5.635847017272404e-05,
      "loss": 0.33780547976493835,
      "memory(GiB)": 78.33,
      "step": 3762,
      "token_acc": 0.898160262743318,
      "train_speed(iter/s)": 0.032362
    },
    {
      "epoch": 0.729157583684542,
      "grad_norm": 0.10268279910087585,
      "learning_rate": 5.628339081157556e-05,
      "loss": 0.3292725682258606,
      "memory(GiB)": 78.33,
      "step": 3763,
      "token_acc": 0.9022685641820487,
      "train_speed(iter/s)": 0.032362
    },
    {
      "epoch": 0.7293513539698687,
      "grad_norm": 0.10835665464401245,
      "learning_rate": 5.620834994235673e-05,
      "loss": 0.3525397479534149,
      "memory(GiB)": 78.33,
      "step": 3764,
      "token_acc": 0.8940980485483103,
      "train_speed(iter/s)": 0.032363
    },
    {
      "epoch": 0.7295451242551955,
      "grad_norm": 0.09963172674179077,
      "learning_rate": 5.613334759588885e-05,
      "loss": 0.3277450501918793,
      "memory(GiB)": 78.33,
      "step": 3765,
      "token_acc": 0.9008741844812914,
      "train_speed(iter/s)": 0.032364
    },
    {
      "epoch": 0.7297388945405222,
      "grad_norm": 0.10072033107280731,
      "learning_rate": 5.605838380297742e-05,
      "loss": 0.3499065041542053,
      "memory(GiB)": 78.33,
      "step": 3766,
      "token_acc": 0.8946452543131308,
      "train_speed(iter/s)": 0.032364
    },
    {
      "epoch": 0.729932664825849,
      "grad_norm": 0.09813085943460464,
      "learning_rate": 5.5983458594412075e-05,
      "loss": 0.3212866187095642,
      "memory(GiB)": 78.33,
      "step": 3767,
      "token_acc": 0.9034384826047267,
      "train_speed(iter/s)": 0.032365
    },
    {
      "epoch": 0.7301264351111757,
      "grad_norm": 0.1046958863735199,
      "learning_rate": 5.5908572000966545e-05,
      "loss": 0.3336329162120819,
      "memory(GiB)": 78.33,
      "step": 3768,
      "token_acc": 0.9022138594066974,
      "train_speed(iter/s)": 0.032366
    },
    {
      "epoch": 0.7303202053965024,
      "grad_norm": 0.10364958643913269,
      "learning_rate": 5.583372405339888e-05,
      "loss": 0.3599414527416229,
      "memory(GiB)": 78.33,
      "step": 3769,
      "token_acc": 0.8923859525040124,
      "train_speed(iter/s)": 0.032367
    },
    {
      "epoch": 0.7305139756818292,
      "grad_norm": 0.11301209777593613,
      "learning_rate": 5.5758914782451094e-05,
      "loss": 0.3293749690055847,
      "memory(GiB)": 78.33,
      "step": 3770,
      "token_acc": 0.9023539911864785,
      "train_speed(iter/s)": 0.032367
    },
    {
      "epoch": 0.7307077459671559,
      "grad_norm": 0.10696696490049362,
      "learning_rate": 5.5684144218849364e-05,
      "loss": 0.3482256531715393,
      "memory(GiB)": 78.33,
      "step": 3771,
      "token_acc": 0.8958611481975968,
      "train_speed(iter/s)": 0.032368
    },
    {
      "epoch": 0.7309015162524827,
      "grad_norm": 0.1035601943731308,
      "learning_rate": 5.5609412393303983e-05,
      "loss": 0.35395896434783936,
      "memory(GiB)": 78.33,
      "step": 3772,
      "token_acc": 0.8959904183792906,
      "train_speed(iter/s)": 0.032369
    },
    {
      "epoch": 0.7310952865378094,
      "grad_norm": 0.10243549197912216,
      "learning_rate": 5.553471933650922e-05,
      "loss": 0.37264150381088257,
      "memory(GiB)": 78.33,
      "step": 3773,
      "token_acc": 0.8881558441558441,
      "train_speed(iter/s)": 0.032369
    },
    {
      "epoch": 0.7312890568231362,
      "grad_norm": 0.10226655006408691,
      "learning_rate": 5.546006507914369e-05,
      "loss": 0.3363930284976959,
      "memory(GiB)": 78.33,
      "step": 3774,
      "token_acc": 0.8988527079466847,
      "train_speed(iter/s)": 0.03237
    },
    {
      "epoch": 0.7314828271084629,
      "grad_norm": 0.11039669811725616,
      "learning_rate": 5.5385449651869815e-05,
      "loss": 0.34301939606666565,
      "memory(GiB)": 78.33,
      "step": 3775,
      "token_acc": 0.8971273781025576,
      "train_speed(iter/s)": 0.032371
    },
    {
      "epoch": 0.7316765973937897,
      "grad_norm": 0.10423394292593002,
      "learning_rate": 5.531087308533414e-05,
      "loss": 0.3568193316459656,
      "memory(GiB)": 78.33,
      "step": 3776,
      "token_acc": 0.8942443919716647,
      "train_speed(iter/s)": 0.032372
    },
    {
      "epoch": 0.7318703676791164,
      "grad_norm": 0.11283637583255768,
      "learning_rate": 5.52363354101673e-05,
      "loss": 0.36575639247894287,
      "memory(GiB)": 78.33,
      "step": 3777,
      "token_acc": 0.8924577682607662,
      "train_speed(iter/s)": 0.032372
    },
    {
      "epoch": 0.7320641379644431,
      "grad_norm": 0.09643404930830002,
      "learning_rate": 5.51618366569839e-05,
      "loss": 0.30832773447036743,
      "memory(GiB)": 78.33,
      "step": 3778,
      "token_acc": 0.9075776332077857,
      "train_speed(iter/s)": 0.032373
    },
    {
      "epoch": 0.7322579082497699,
      "grad_norm": 0.10145960003137589,
      "learning_rate": 5.508737685638259e-05,
      "loss": 0.32317882776260376,
      "memory(GiB)": 78.33,
      "step": 3779,
      "token_acc": 0.9013355214758644,
      "train_speed(iter/s)": 0.032374
    },
    {
      "epoch": 0.7324516785350966,
      "grad_norm": 0.10353747755289078,
      "learning_rate": 5.501295603894594e-05,
      "loss": 0.34720805287361145,
      "memory(GiB)": 78.33,
      "step": 3780,
      "token_acc": 0.8948657321398412,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.7326454488204234,
      "grad_norm": 0.09833744168281555,
      "learning_rate": 5.49385742352407e-05,
      "loss": 0.33571141958236694,
      "memory(GiB)": 78.33,
      "step": 3781,
      "token_acc": 0.9020605793832017,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.7328392191057501,
      "grad_norm": 0.09903203696012497,
      "learning_rate": 5.486423147581744e-05,
      "loss": 0.3094678521156311,
      "memory(GiB)": 78.33,
      "step": 3782,
      "token_acc": 0.9053752361325778,
      "train_speed(iter/s)": 0.032376
    },
    {
      "epoch": 0.7330329893910769,
      "grad_norm": 0.1019536554813385,
      "learning_rate": 5.4789927791210694e-05,
      "loss": 0.34034597873687744,
      "memory(GiB)": 78.33,
      "step": 3783,
      "token_acc": 0.8967783128042779,
      "train_speed(iter/s)": 0.032377
    },
    {
      "epoch": 0.7332267596764036,
      "grad_norm": 0.10213874280452728,
      "learning_rate": 5.4715663211939e-05,
      "loss": 0.34146547317504883,
      "memory(GiB)": 78.33,
      "step": 3784,
      "token_acc": 0.8989790738338316,
      "train_speed(iter/s)": 0.032377
    },
    {
      "epoch": 0.7334205299617303,
      "grad_norm": 0.09893448650836945,
      "learning_rate": 5.4641437768504824e-05,
      "loss": 0.3442443907260895,
      "memory(GiB)": 78.33,
      "step": 3785,
      "token_acc": 0.8979535398230089,
      "train_speed(iter/s)": 0.032378
    },
    {
      "epoch": 0.7336143002470571,
      "grad_norm": 0.10614677518606186,
      "learning_rate": 5.456725149139454e-05,
      "loss": 0.34340110421180725,
      "memory(GiB)": 78.33,
      "step": 3786,
      "token_acc": 0.895698897732694,
      "train_speed(iter/s)": 0.032379
    },
    {
      "epoch": 0.7338080705323838,
      "grad_norm": 0.0985707938671112,
      "learning_rate": 5.449310441107838e-05,
      "loss": 0.33455923199653625,
      "memory(GiB)": 78.33,
      "step": 3787,
      "token_acc": 0.8980871664348876,
      "train_speed(iter/s)": 0.032379
    },
    {
      "epoch": 0.7340018408177106,
      "grad_norm": 0.1156158521771431,
      "learning_rate": 5.4418996558010667e-05,
      "loss": 0.3170754015445709,
      "memory(GiB)": 78.33,
      "step": 3788,
      "token_acc": 0.9032800672834315,
      "train_speed(iter/s)": 0.03238
    },
    {
      "epoch": 0.7341956111030373,
      "grad_norm": 0.11739075183868408,
      "learning_rate": 5.434492796262942e-05,
      "loss": 0.38899776339530945,
      "memory(GiB)": 78.33,
      "step": 3789,
      "token_acc": 0.8846523748952937,
      "train_speed(iter/s)": 0.032381
    },
    {
      "epoch": 0.7343893813883641,
      "grad_norm": 0.1062496230006218,
      "learning_rate": 5.4270898655356625e-05,
      "loss": 0.335146427154541,
      "memory(GiB)": 78.33,
      "step": 3790,
      "token_acc": 0.8996969696969697,
      "train_speed(iter/s)": 0.032382
    },
    {
      "epoch": 0.7345831516736908,
      "grad_norm": 0.11981339007616043,
      "learning_rate": 5.4196908666598075e-05,
      "loss": 0.41245871782302856,
      "memory(GiB)": 78.33,
      "step": 3791,
      "token_acc": 0.8796483136350534,
      "train_speed(iter/s)": 0.032382
    },
    {
      "epoch": 0.7347769219590176,
      "grad_norm": 0.10465206950902939,
      "learning_rate": 5.412295802674348e-05,
      "loss": 0.3469778895378113,
      "memory(GiB)": 78.33,
      "step": 3792,
      "token_acc": 0.8962869653889989,
      "train_speed(iter/s)": 0.032383
    },
    {
      "epoch": 0.7349706922443443,
      "grad_norm": 0.0999392494559288,
      "learning_rate": 5.4049046766166335e-05,
      "loss": 0.3266690969467163,
      "memory(GiB)": 78.33,
      "step": 3793,
      "token_acc": 0.9025028714907711,
      "train_speed(iter/s)": 0.032384
    },
    {
      "epoch": 0.735164462529671,
      "grad_norm": 0.10005468130111694,
      "learning_rate": 5.397517491522393e-05,
      "loss": 0.34978044033050537,
      "memory(GiB)": 78.33,
      "step": 3794,
      "token_acc": 0.8954549081141459,
      "train_speed(iter/s)": 0.032384
    },
    {
      "epoch": 0.7353582328149978,
      "grad_norm": 0.11191490292549133,
      "learning_rate": 5.390134250425753e-05,
      "loss": 0.3621137738227844,
      "memory(GiB)": 78.33,
      "step": 3795,
      "token_acc": 0.8916416576208416,
      "train_speed(iter/s)": 0.032385
    },
    {
      "epoch": 0.7355520031003245,
      "grad_norm": 0.0992647185921669,
      "learning_rate": 5.382754956359204e-05,
      "loss": 0.33923065662384033,
      "memory(GiB)": 78.33,
      "step": 3796,
      "token_acc": 0.8980301274623407,
      "train_speed(iter/s)": 0.032386
    },
    {
      "epoch": 0.7357457733856513,
      "grad_norm": 0.10284343361854553,
      "learning_rate": 5.37537961235362e-05,
      "loss": 0.3275540769100189,
      "memory(GiB)": 78.33,
      "step": 3797,
      "token_acc": 0.9016500897717233,
      "train_speed(iter/s)": 0.032387
    },
    {
      "epoch": 0.735939543670978,
      "grad_norm": 0.09948836266994476,
      "learning_rate": 5.368008221438251e-05,
      "loss": 0.3380359709262848,
      "memory(GiB)": 78.33,
      "step": 3798,
      "token_acc": 0.8977125965043193,
      "train_speed(iter/s)": 0.032387
    },
    {
      "epoch": 0.7361333139563048,
      "grad_norm": 0.09476862102746964,
      "learning_rate": 5.360640786640729e-05,
      "loss": 0.3325924873352051,
      "memory(GiB)": 78.33,
      "step": 3799,
      "token_acc": 0.9025462002294754,
      "train_speed(iter/s)": 0.032388
    },
    {
      "epoch": 0.7363270842416315,
      "grad_norm": 0.10117123275995255,
      "learning_rate": 5.3532773109870544e-05,
      "loss": 0.3492782711982727,
      "memory(GiB)": 78.33,
      "step": 3800,
      "token_acc": 0.8976592977893368,
      "train_speed(iter/s)": 0.032389
    },
    {
      "epoch": 0.7365208545269583,
      "grad_norm": 0.11379625648260117,
      "learning_rate": 5.3459177975016e-05,
      "loss": 0.34902098774909973,
      "memory(GiB)": 78.33,
      "step": 3801,
      "token_acc": 0.8949467686309791,
      "train_speed(iter/s)": 0.032386
    },
    {
      "epoch": 0.736714624812285,
      "grad_norm": 0.10552668571472168,
      "learning_rate": 5.338562249207128e-05,
      "loss": 0.35885700583457947,
      "memory(GiB)": 78.33,
      "step": 3802,
      "token_acc": 0.8908405111749141,
      "train_speed(iter/s)": 0.032387
    },
    {
      "epoch": 0.7369083950976117,
      "grad_norm": 0.10673682391643524,
      "learning_rate": 5.331210669124752e-05,
      "loss": 0.34370338916778564,
      "memory(GiB)": 78.33,
      "step": 3803,
      "token_acc": 0.897491594013152,
      "train_speed(iter/s)": 0.032387
    },
    {
      "epoch": 0.7371021653829385,
      "grad_norm": 0.09526897221803665,
      "learning_rate": 5.323863060273966e-05,
      "loss": 0.303422749042511,
      "memory(GiB)": 78.33,
      "step": 3804,
      "token_acc": 0.909460807476243,
      "train_speed(iter/s)": 0.032388
    },
    {
      "epoch": 0.7372959356682652,
      "grad_norm": 0.08512071520090103,
      "learning_rate": 5.3165194256726275e-05,
      "loss": 0.2895103394985199,
      "memory(GiB)": 78.33,
      "step": 3805,
      "token_acc": 0.9112764240057593,
      "train_speed(iter/s)": 0.032389
    },
    {
      "epoch": 0.737489705953592,
      "grad_norm": 0.103308767080307,
      "learning_rate": 5.309179768336967e-05,
      "loss": 0.34746530652046204,
      "memory(GiB)": 78.33,
      "step": 3806,
      "token_acc": 0.8955654981027354,
      "train_speed(iter/s)": 0.032389
    },
    {
      "epoch": 0.7376834762389187,
      "grad_norm": 0.10745726525783539,
      "learning_rate": 5.301844091281573e-05,
      "loss": 0.33433717489242554,
      "memory(GiB)": 78.33,
      "step": 3807,
      "token_acc": 0.9007903055848261,
      "train_speed(iter/s)": 0.03239
    },
    {
      "epoch": 0.7378772465242455,
      "grad_norm": 0.09775236248970032,
      "learning_rate": 5.294512397519414e-05,
      "loss": 0.3289939761161804,
      "memory(GiB)": 78.33,
      "step": 3808,
      "token_acc": 0.9027567020738493,
      "train_speed(iter/s)": 0.032391
    },
    {
      "epoch": 0.7380710168095722,
      "grad_norm": 0.10481875389814377,
      "learning_rate": 5.287184690061811e-05,
      "loss": 0.32689011096954346,
      "memory(GiB)": 78.33,
      "step": 3809,
      "token_acc": 0.9029928528886242,
      "train_speed(iter/s)": 0.032391
    },
    {
      "epoch": 0.738264787094899,
      "grad_norm": 0.1059269830584526,
      "learning_rate": 5.279860971918449e-05,
      "loss": 0.37165161967277527,
      "memory(GiB)": 78.33,
      "step": 3810,
      "token_acc": 0.8881279218652295,
      "train_speed(iter/s)": 0.032392
    },
    {
      "epoch": 0.7384585573802257,
      "grad_norm": 0.09881359338760376,
      "learning_rate": 5.272541246097376e-05,
      "loss": 0.309497207403183,
      "memory(GiB)": 78.33,
      "step": 3811,
      "token_acc": 0.9069961956668375,
      "train_speed(iter/s)": 0.032393
    },
    {
      "epoch": 0.7386523276655524,
      "grad_norm": 0.10563083738088608,
      "learning_rate": 5.265225515605001e-05,
      "loss": 0.3451386094093323,
      "memory(GiB)": 78.33,
      "step": 3812,
      "token_acc": 0.8983088749126485,
      "train_speed(iter/s)": 0.032394
    },
    {
      "epoch": 0.7388460979508792,
      "grad_norm": 0.10384704172611237,
      "learning_rate": 5.257913783446086e-05,
      "loss": 0.35985156893730164,
      "memory(GiB)": 78.33,
      "step": 3813,
      "token_acc": 0.8926537745954783,
      "train_speed(iter/s)": 0.032394
    },
    {
      "epoch": 0.7390398682362059,
      "grad_norm": 0.10183234512805939,
      "learning_rate": 5.250606052623762e-05,
      "loss": 0.315267413854599,
      "memory(GiB)": 78.33,
      "step": 3814,
      "token_acc": 0.9050984876045943,
      "train_speed(iter/s)": 0.032395
    },
    {
      "epoch": 0.7392336385215327,
      "grad_norm": 0.10309257358312607,
      "learning_rate": 5.2433023261395113e-05,
      "loss": 0.34021520614624023,
      "memory(GiB)": 78.33,
      "step": 3815,
      "token_acc": 0.8987062341073574,
      "train_speed(iter/s)": 0.032396
    },
    {
      "epoch": 0.7394274088068594,
      "grad_norm": 0.11030527949333191,
      "learning_rate": 5.236002606993167e-05,
      "loss": 0.35504722595214844,
      "memory(GiB)": 78.33,
      "step": 3816,
      "token_acc": 0.8941146380971327,
      "train_speed(iter/s)": 0.032396
    },
    {
      "epoch": 0.7396211790921862,
      "grad_norm": 0.09362789988517761,
      "learning_rate": 5.228706898182921e-05,
      "loss": 0.3009899854660034,
      "memory(GiB)": 78.33,
      "step": 3817,
      "token_acc": 0.909036220077704,
      "train_speed(iter/s)": 0.032397
    },
    {
      "epoch": 0.7398149493775129,
      "grad_norm": 0.1043652817606926,
      "learning_rate": 5.221415202705316e-05,
      "loss": 0.3409806489944458,
      "memory(GiB)": 78.33,
      "step": 3818,
      "token_acc": 0.8996104618809126,
      "train_speed(iter/s)": 0.032398
    },
    {
      "epoch": 0.7400087196628397,
      "grad_norm": 0.10487966984510422,
      "learning_rate": 5.214127523555249e-05,
      "loss": 0.3638575077056885,
      "memory(GiB)": 78.33,
      "step": 3819,
      "token_acc": 0.8908792978947954,
      "train_speed(iter/s)": 0.032398
    },
    {
      "epoch": 0.7402024899481664,
      "grad_norm": 0.10344404727220535,
      "learning_rate": 5.206843863725959e-05,
      "loss": 0.3182569742202759,
      "memory(GiB)": 78.33,
      "step": 3820,
      "token_acc": 0.9035023713973003,
      "train_speed(iter/s)": 0.032399
    },
    {
      "epoch": 0.7403962602334931,
      "grad_norm": 0.09493482857942581,
      "learning_rate": 5.199564226209051e-05,
      "loss": 0.31004172563552856,
      "memory(GiB)": 78.33,
      "step": 3821,
      "token_acc": 0.905801490532747,
      "train_speed(iter/s)": 0.0324
    },
    {
      "epoch": 0.7405900305188199,
      "grad_norm": 0.10709976404905319,
      "learning_rate": 5.192288613994464e-05,
      "loss": 0.35697808861732483,
      "memory(GiB)": 78.33,
      "step": 3822,
      "token_acc": 0.8928056450375824,
      "train_speed(iter/s)": 0.032401
    },
    {
      "epoch": 0.7407838008041466,
      "grad_norm": 0.09773747622966766,
      "learning_rate": 5.185017030070487e-05,
      "loss": 0.34137359261512756,
      "memory(GiB)": 78.33,
      "step": 3823,
      "token_acc": 0.8971325975603316,
      "train_speed(iter/s)": 0.032401
    },
    {
      "epoch": 0.7409775710894734,
      "grad_norm": 0.10251982510089874,
      "learning_rate": 5.1777494774237534e-05,
      "loss": 0.3589448034763336,
      "memory(GiB)": 78.33,
      "step": 3824,
      "token_acc": 0.8923671206762486,
      "train_speed(iter/s)": 0.032402
    },
    {
      "epoch": 0.7411713413748001,
      "grad_norm": 0.09967760741710663,
      "learning_rate": 5.170485959039244e-05,
      "loss": 0.33076420426368713,
      "memory(GiB)": 78.33,
      "step": 3825,
      "token_acc": 0.9000113526707157,
      "train_speed(iter/s)": 0.032403
    },
    {
      "epoch": 0.741365111660127,
      "grad_norm": 0.09976795315742493,
      "learning_rate": 5.163226477900281e-05,
      "loss": 0.33364495635032654,
      "memory(GiB)": 78.33,
      "step": 3826,
      "token_acc": 0.9020033344801122,
      "train_speed(iter/s)": 0.032403
    },
    {
      "epoch": 0.7415588819454537,
      "grad_norm": 0.09653392434120178,
      "learning_rate": 5.155971036988522e-05,
      "loss": 0.32705751061439514,
      "memory(GiB)": 78.33,
      "step": 3827,
      "token_acc": 0.8996669334641502,
      "train_speed(iter/s)": 0.032404
    },
    {
      "epoch": 0.7417526522307805,
      "grad_norm": 0.10461395233869553,
      "learning_rate": 5.148719639283984e-05,
      "loss": 0.35183149576187134,
      "memory(GiB)": 78.33,
      "step": 3828,
      "token_acc": 0.8943066752210661,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.7419464225161072,
      "grad_norm": 0.09616634994745255,
      "learning_rate": 5.1414722877650025e-05,
      "loss": 0.34106606245040894,
      "memory(GiB)": 78.33,
      "step": 3829,
      "token_acc": 0.8989504013171434,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.742140192801434,
      "grad_norm": 0.0909215584397316,
      "learning_rate": 5.134228985408262e-05,
      "loss": 0.3328753709793091,
      "memory(GiB)": 78.33,
      "step": 3830,
      "token_acc": 0.8994169096209913,
      "train_speed(iter/s)": 0.032406
    },
    {
      "epoch": 0.7423339630867607,
      "grad_norm": 0.107168048620224,
      "learning_rate": 5.126989735188782e-05,
      "loss": 0.3591649532318115,
      "memory(GiB)": 78.33,
      "step": 3831,
      "token_acc": 0.8922912205567451,
      "train_speed(iter/s)": 0.032407
    },
    {
      "epoch": 0.7425277333720874,
      "grad_norm": 0.10682724416255951,
      "learning_rate": 5.119754540079914e-05,
      "loss": 0.35992226004600525,
      "memory(GiB)": 78.33,
      "step": 3832,
      "token_acc": 0.893421090615167,
      "train_speed(iter/s)": 0.032408
    },
    {
      "epoch": 0.7427215036574142,
      "grad_norm": 0.10819026827812195,
      "learning_rate": 5.1125234030533494e-05,
      "loss": 0.3347112536430359,
      "memory(GiB)": 78.33,
      "step": 3833,
      "token_acc": 0.8996055354371444,
      "train_speed(iter/s)": 0.032408
    },
    {
      "epoch": 0.7429152739427409,
      "grad_norm": 0.09705478698015213,
      "learning_rate": 5.1052963270791045e-05,
      "loss": 0.33945244550704956,
      "memory(GiB)": 78.33,
      "step": 3834,
      "token_acc": 0.89829610196889,
      "train_speed(iter/s)": 0.032409
    },
    {
      "epoch": 0.7431090442280677,
      "grad_norm": 0.09978172183036804,
      "learning_rate": 5.098073315125545e-05,
      "loss": 0.314132958650589,
      "memory(GiB)": 78.33,
      "step": 3835,
      "token_acc": 0.9039825686962838,
      "train_speed(iter/s)": 0.03241
    },
    {
      "epoch": 0.7433028145133944,
      "grad_norm": 0.1035098284482956,
      "learning_rate": 5.09085437015935e-05,
      "loss": 0.3336668908596039,
      "memory(GiB)": 78.33,
      "step": 3836,
      "token_acc": 0.9000965741587883,
      "train_speed(iter/s)": 0.03241
    },
    {
      "epoch": 0.7434965847987212,
      "grad_norm": 0.10066191852092743,
      "learning_rate": 5.083639495145534e-05,
      "loss": 0.3203166127204895,
      "memory(GiB)": 78.33,
      "step": 3837,
      "token_acc": 0.9036123032904149,
      "train_speed(iter/s)": 0.032411
    },
    {
      "epoch": 0.7436903550840479,
      "grad_norm": 0.11868193745613098,
      "learning_rate": 5.076428693047439e-05,
      "loss": 0.3131275177001953,
      "memory(GiB)": 78.33,
      "step": 3838,
      "token_acc": 0.9023334587881069,
      "train_speed(iter/s)": 0.032412
    },
    {
      "epoch": 0.7438841253693746,
      "grad_norm": 0.108455590903759,
      "learning_rate": 5.069221966826738e-05,
      "loss": 0.3502568006515503,
      "memory(GiB)": 78.33,
      "step": 3839,
      "token_acc": 0.8950985915492958,
      "train_speed(iter/s)": 0.032412
    },
    {
      "epoch": 0.7440778956547014,
      "grad_norm": 0.10078644752502441,
      "learning_rate": 5.06201931944342e-05,
      "loss": 0.33233729004859924,
      "memory(GiB)": 78.33,
      "step": 3840,
      "token_acc": 0.8990275860042892,
      "train_speed(iter/s)": 0.032413
    },
    {
      "epoch": 0.7442716659400281,
      "grad_norm": 0.10141783952713013,
      "learning_rate": 5.054820753855817e-05,
      "loss": 0.3332747220993042,
      "memory(GiB)": 78.33,
      "step": 3841,
      "token_acc": 0.902451309603761,
      "train_speed(iter/s)": 0.032414
    },
    {
      "epoch": 0.7444654362253549,
      "grad_norm": 0.0925682932138443,
      "learning_rate": 5.047626273020568e-05,
      "loss": 0.30119389295578003,
      "memory(GiB)": 78.33,
      "step": 3842,
      "token_acc": 0.9076820241130888,
      "train_speed(iter/s)": 0.032414
    },
    {
      "epoch": 0.7446592065106816,
      "grad_norm": 0.09351833164691925,
      "learning_rate": 5.040435879892639e-05,
      "loss": 0.3276900053024292,
      "memory(GiB)": 78.33,
      "step": 3843,
      "token_acc": 0.9007225727661307,
      "train_speed(iter/s)": 0.032415
    },
    {
      "epoch": 0.7448529767960084,
      "grad_norm": 0.1048940047621727,
      "learning_rate": 5.0332495774253165e-05,
      "loss": 0.33195894956588745,
      "memory(GiB)": 78.33,
      "step": 3844,
      "token_acc": 0.8993898230210663,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.7450467470813351,
      "grad_norm": 0.09521888941526413,
      "learning_rate": 5.026067368570211e-05,
      "loss": 0.28278443217277527,
      "memory(GiB)": 78.33,
      "step": 3845,
      "token_acc": 0.9150288711129947,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.7452405173666619,
      "grad_norm": 0.12543639540672302,
      "learning_rate": 5.018889256277241e-05,
      "loss": 0.32016804814338684,
      "memory(GiB)": 78.33,
      "step": 3846,
      "token_acc": 0.9048030793261185,
      "train_speed(iter/s)": 0.032417
    },
    {
      "epoch": 0.7454342876519886,
      "grad_norm": 0.09576548635959625,
      "learning_rate": 5.011715243494663e-05,
      "loss": 0.3216573894023895,
      "memory(GiB)": 78.33,
      "step": 3847,
      "token_acc": 0.9055984653440355,
      "train_speed(iter/s)": 0.032418
    },
    {
      "epoch": 0.7456280579373153,
      "grad_norm": 0.09915050864219666,
      "learning_rate": 5.004545333169028e-05,
      "loss": 0.3430394232273102,
      "memory(GiB)": 78.33,
      "step": 3848,
      "token_acc": 0.8962812609632637,
      "train_speed(iter/s)": 0.032419
    },
    {
      "epoch": 0.7458218282226421,
      "grad_norm": 0.09718713909387589,
      "learning_rate": 4.997379528245215e-05,
      "loss": 0.3420059084892273,
      "memory(GiB)": 78.33,
      "step": 3849,
      "token_acc": 0.8961742910262117,
      "train_speed(iter/s)": 0.032419
    },
    {
      "epoch": 0.7460155985079688,
      "grad_norm": 0.10262469947338104,
      "learning_rate": 4.990217831666409e-05,
      "loss": 0.3337242007255554,
      "memory(GiB)": 78.33,
      "step": 3850,
      "token_acc": 0.9016568394653292,
      "train_speed(iter/s)": 0.03242
    },
    {
      "epoch": 0.7462093687932956,
      "grad_norm": 0.09261801093816757,
      "learning_rate": 4.983060246374115e-05,
      "loss": 0.2943536341190338,
      "memory(GiB)": 78.33,
      "step": 3851,
      "token_acc": 0.9127831715210356,
      "train_speed(iter/s)": 0.032421
    },
    {
      "epoch": 0.7464031390786223,
      "grad_norm": 0.11604174971580505,
      "learning_rate": 4.9759067753081414e-05,
      "loss": 0.3443485200405121,
      "memory(GiB)": 78.33,
      "step": 3852,
      "token_acc": 0.8987715146996856,
      "train_speed(iter/s)": 0.032421
    },
    {
      "epoch": 0.7465969093639491,
      "grad_norm": 0.10266567766666412,
      "learning_rate": 4.968757421406608e-05,
      "loss": 0.3430730998516083,
      "memory(GiB)": 78.33,
      "step": 3853,
      "token_acc": 0.897044232074782,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.7467906796492758,
      "grad_norm": 0.10636512190103531,
      "learning_rate": 4.961612187605958e-05,
      "loss": 0.3752942383289337,
      "memory(GiB)": 78.33,
      "step": 3854,
      "token_acc": 0.886795650331874,
      "train_speed(iter/s)": 0.032423
    },
    {
      "epoch": 0.7469844499346026,
      "grad_norm": 0.11186108738183975,
      "learning_rate": 4.954471076840922e-05,
      "loss": 0.37107086181640625,
      "memory(GiB)": 78.33,
      "step": 3855,
      "token_acc": 0.8894728014759322,
      "train_speed(iter/s)": 0.032423
    },
    {
      "epoch": 0.7471782202199293,
      "grad_norm": 0.10150814801454544,
      "learning_rate": 4.947334092044552e-05,
      "loss": 0.3377690315246582,
      "memory(GiB)": 78.33,
      "step": 3856,
      "token_acc": 0.8990276738967838,
      "train_speed(iter/s)": 0.032424
    },
    {
      "epoch": 0.747371990505256,
      "grad_norm": 0.10091419517993927,
      "learning_rate": 4.9402012361481934e-05,
      "loss": 0.30891650915145874,
      "memory(GiB)": 78.33,
      "step": 3857,
      "token_acc": 0.9063477562771641,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.7475657607905828,
      "grad_norm": 0.09730595350265503,
      "learning_rate": 4.9330725120815054e-05,
      "loss": 0.3267110586166382,
      "memory(GiB)": 78.33,
      "step": 3858,
      "token_acc": 0.900251940973829,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.7477595310759095,
      "grad_norm": 0.10822945088148117,
      "learning_rate": 4.925947922772445e-05,
      "loss": 0.4060097932815552,
      "memory(GiB)": 78.33,
      "step": 3859,
      "token_acc": 0.8794587622081238,
      "train_speed(iter/s)": 0.032426
    },
    {
      "epoch": 0.7479533013612363,
      "grad_norm": 0.10585605353116989,
      "learning_rate": 4.918827471147268e-05,
      "loss": 0.338094025850296,
      "memory(GiB)": 78.33,
      "step": 3860,
      "token_acc": 0.8971483116573201,
      "train_speed(iter/s)": 0.032427
    },
    {
      "epoch": 0.748147071646563,
      "grad_norm": 0.12144403904676437,
      "learning_rate": 4.911711160130546e-05,
      "loss": 0.36428892612457275,
      "memory(GiB)": 78.33,
      "step": 3861,
      "token_acc": 0.8935286935286936,
      "train_speed(iter/s)": 0.032427
    },
    {
      "epoch": 0.7483408419318898,
      "grad_norm": 0.09023165702819824,
      "learning_rate": 4.904598992645132e-05,
      "loss": 0.3032684028148651,
      "memory(GiB)": 78.33,
      "step": 3862,
      "token_acc": 0.9088954518606025,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.7485346122172165,
      "grad_norm": 0.10547629743814468,
      "learning_rate": 4.897490971612187e-05,
      "loss": 0.33661413192749023,
      "memory(GiB)": 78.33,
      "step": 3863,
      "token_acc": 0.8987709009233841,
      "train_speed(iter/s)": 0.032429
    },
    {
      "epoch": 0.7487283825025433,
      "grad_norm": 0.10550445318222046,
      "learning_rate": 4.890387099951164e-05,
      "loss": 0.34583577513694763,
      "memory(GiB)": 78.33,
      "step": 3864,
      "token_acc": 0.8972740894421393,
      "train_speed(iter/s)": 0.03243
    },
    {
      "epoch": 0.74892215278787,
      "grad_norm": 0.09834988415241241,
      "learning_rate": 4.883287380579816e-05,
      "loss": 0.3329737186431885,
      "memory(GiB)": 78.33,
      "step": 3865,
      "token_acc": 0.900711867118138,
      "train_speed(iter/s)": 0.03243
    },
    {
      "epoch": 0.7491159230731967,
      "grad_norm": 0.0946660041809082,
      "learning_rate": 4.876191816414186e-05,
      "loss": 0.310972660779953,
      "memory(GiB)": 78.33,
      "step": 3866,
      "token_acc": 0.9068650482902636,
      "train_speed(iter/s)": 0.032431
    },
    {
      "epoch": 0.7493096933585235,
      "grad_norm": 0.09578848630189896,
      "learning_rate": 4.869100410368609e-05,
      "loss": 0.31944969296455383,
      "memory(GiB)": 78.33,
      "step": 3867,
      "token_acc": 0.904564120614807,
      "train_speed(iter/s)": 0.032432
    },
    {
      "epoch": 0.7495034636438502,
      "grad_norm": 0.09447766840457916,
      "learning_rate": 4.862013165355728e-05,
      "loss": 0.30866914987564087,
      "memory(GiB)": 78.33,
      "step": 3868,
      "token_acc": 0.9051225521981585,
      "train_speed(iter/s)": 0.032432
    },
    {
      "epoch": 0.749697233929177,
      "grad_norm": 0.09313666820526123,
      "learning_rate": 4.8549300842864576e-05,
      "loss": 0.32266178727149963,
      "memory(GiB)": 78.33,
      "step": 3869,
      "token_acc": 0.90316239526255,
      "train_speed(iter/s)": 0.032433
    },
    {
      "epoch": 0.7498910042145037,
      "grad_norm": 0.10308795422315598,
      "learning_rate": 4.84785117007001e-05,
      "loss": 0.3472236096858978,
      "memory(GiB)": 78.33,
      "step": 3870,
      "token_acc": 0.894967925873129,
      "train_speed(iter/s)": 0.032434
    },
    {
      "epoch": 0.7500847744998305,
      "grad_norm": 0.10628514736890793,
      "learning_rate": 4.840776425613886e-05,
      "loss": 0.3184339702129364,
      "memory(GiB)": 78.33,
      "step": 3871,
      "token_acc": 0.9032356532356532,
      "train_speed(iter/s)": 0.032434
    },
    {
      "epoch": 0.7502785447851572,
      "grad_norm": 0.09150487929582596,
      "learning_rate": 4.833705853823872e-05,
      "loss": 0.3118763864040375,
      "memory(GiB)": 78.33,
      "step": 3872,
      "token_acc": 0.9086181652950976,
      "train_speed(iter/s)": 0.032435
    },
    {
      "epoch": 0.750472315070484,
      "grad_norm": 0.10045349597930908,
      "learning_rate": 4.826639457604039e-05,
      "loss": 0.317714124917984,
      "memory(GiB)": 78.33,
      "step": 3873,
      "token_acc": 0.9055200119029906,
      "train_speed(iter/s)": 0.032436
    },
    {
      "epoch": 0.7506660853558107,
      "grad_norm": 0.09495769441127777,
      "learning_rate": 4.819577239856754e-05,
      "loss": 0.3022938072681427,
      "memory(GiB)": 78.33,
      "step": 3874,
      "token_acc": 0.9077078844206755,
      "train_speed(iter/s)": 0.032436
    },
    {
      "epoch": 0.7508598556411374,
      "grad_norm": 0.09479624778032303,
      "learning_rate": 4.812519203482655e-05,
      "loss": 0.3185397982597351,
      "memory(GiB)": 78.33,
      "step": 3875,
      "token_acc": 0.9046973580441641,
      "train_speed(iter/s)": 0.032437
    },
    {
      "epoch": 0.7510536259264642,
      "grad_norm": 0.10061098635196686,
      "learning_rate": 4.805465351380666e-05,
      "loss": 0.3666335940361023,
      "memory(GiB)": 78.33,
      "step": 3876,
      "token_acc": 0.8898552194528259,
      "train_speed(iter/s)": 0.032438
    },
    {
      "epoch": 0.7512473962117909,
      "grad_norm": 0.1007319763302803,
      "learning_rate": 4.798415686447997e-05,
      "loss": 0.3402915298938751,
      "memory(GiB)": 78.33,
      "step": 3877,
      "token_acc": 0.8979099678456591,
      "train_speed(iter/s)": 0.032438
    },
    {
      "epoch": 0.7514411664971177,
      "grad_norm": 0.09538646787405014,
      "learning_rate": 4.791370211580132e-05,
      "loss": 0.3278900384902954,
      "memory(GiB)": 78.33,
      "step": 3878,
      "token_acc": 0.9009969149180286,
      "train_speed(iter/s)": 0.032439
    },
    {
      "epoch": 0.7516349367824444,
      "grad_norm": 0.10622713714838028,
      "learning_rate": 4.7843289296708384e-05,
      "loss": 0.33736902475357056,
      "memory(GiB)": 78.33,
      "step": 3879,
      "token_acc": 0.8988551338682051,
      "train_speed(iter/s)": 0.03244
    },
    {
      "epoch": 0.7518287070677712,
      "grad_norm": 0.10398625582456589,
      "learning_rate": 4.777291843612153e-05,
      "loss": 0.31338000297546387,
      "memory(GiB)": 78.33,
      "step": 3880,
      "token_acc": 0.9074961626413558,
      "train_speed(iter/s)": 0.03244
    },
    {
      "epoch": 0.7520224773530979,
      "grad_norm": 0.1060367301106453,
      "learning_rate": 4.770258956294408e-05,
      "loss": 0.3464204668998718,
      "memory(GiB)": 78.33,
      "step": 3881,
      "token_acc": 0.8942851343906753,
      "train_speed(iter/s)": 0.032441
    },
    {
      "epoch": 0.7522162476384247,
      "grad_norm": 0.11519785970449448,
      "learning_rate": 4.7632302706061925e-05,
      "loss": 0.3710397779941559,
      "memory(GiB)": 78.33,
      "step": 3882,
      "token_acc": 0.8889779944825816,
      "train_speed(iter/s)": 0.032442
    },
    {
      "epoch": 0.7524100179237514,
      "grad_norm": 0.0994989275932312,
      "learning_rate": 4.756205789434379e-05,
      "loss": 0.3080540597438812,
      "memory(GiB)": 78.33,
      "step": 3883,
      "token_acc": 0.9073027181909703,
      "train_speed(iter/s)": 0.032442
    },
    {
      "epoch": 0.7526037882090781,
      "grad_norm": 0.11023864895105362,
      "learning_rate": 4.749185515664109e-05,
      "loss": 0.3553957939147949,
      "memory(GiB)": 78.33,
      "step": 3884,
      "token_acc": 0.8958528037383178,
      "train_speed(iter/s)": 0.032443
    },
    {
      "epoch": 0.7527975584944049,
      "grad_norm": 0.10181353241205215,
      "learning_rate": 4.742169452178796e-05,
      "loss": 0.3252992630004883,
      "memory(GiB)": 78.33,
      "step": 3885,
      "token_acc": 0.9014353801817778,
      "train_speed(iter/s)": 0.032444
    },
    {
      "epoch": 0.7529913287797316,
      "grad_norm": 0.0931096225976944,
      "learning_rate": 4.735157601860123e-05,
      "loss": 0.3285648226737976,
      "memory(GiB)": 78.33,
      "step": 3886,
      "token_acc": 0.9007917114651549,
      "train_speed(iter/s)": 0.032444
    },
    {
      "epoch": 0.7531850990650584,
      "grad_norm": 0.09871333092451096,
      "learning_rate": 4.7281499675880564e-05,
      "loss": 0.31424829363822937,
      "memory(GiB)": 78.33,
      "step": 3887,
      "token_acc": 0.9050426360669545,
      "train_speed(iter/s)": 0.032445
    },
    {
      "epoch": 0.7533788693503851,
      "grad_norm": 0.0981900617480278,
      "learning_rate": 4.7211465522408124e-05,
      "loss": 0.2891225814819336,
      "memory(GiB)": 78.33,
      "step": 3888,
      "token_acc": 0.9127433168959252,
      "train_speed(iter/s)": 0.032446
    },
    {
      "epoch": 0.7535726396357119,
      "grad_norm": 0.10310321301221848,
      "learning_rate": 4.714147358694883e-05,
      "loss": 0.341362327337265,
      "memory(GiB)": 78.33,
      "step": 3889,
      "token_acc": 0.899036050593414,
      "train_speed(iter/s)": 0.032446
    },
    {
      "epoch": 0.7537664099210386,
      "grad_norm": 0.08967840671539307,
      "learning_rate": 4.7071523898250246e-05,
      "loss": 0.3140985369682312,
      "memory(GiB)": 78.33,
      "step": 3890,
      "token_acc": 0.902944590528984,
      "train_speed(iter/s)": 0.032447
    },
    {
      "epoch": 0.7539601802063653,
      "grad_norm": 0.10281458497047424,
      "learning_rate": 4.700161648504261e-05,
      "loss": 0.34009885787963867,
      "memory(GiB)": 78.33,
      "step": 3891,
      "token_acc": 0.8981371841155235,
      "train_speed(iter/s)": 0.032448
    },
    {
      "epoch": 0.7541539504916921,
      "grad_norm": 0.0905076265335083,
      "learning_rate": 4.6931751376038735e-05,
      "loss": 0.2962040305137634,
      "memory(GiB)": 78.33,
      "step": 3892,
      "token_acc": 0.9114481536497017,
      "train_speed(iter/s)": 0.032448
    },
    {
      "epoch": 0.7543477207770188,
      "grad_norm": 0.11229556798934937,
      "learning_rate": 4.6861928599934086e-05,
      "loss": 0.38348227739334106,
      "memory(GiB)": 78.33,
      "step": 3893,
      "token_acc": 0.8880430168388284,
      "train_speed(iter/s)": 0.032449
    },
    {
      "epoch": 0.7545414910623456,
      "grad_norm": 0.10800474882125854,
      "learning_rate": 4.679214818540683e-05,
      "loss": 0.3647887110710144,
      "memory(GiB)": 78.33,
      "step": 3894,
      "token_acc": 0.8906179648806843,
      "train_speed(iter/s)": 0.03245
    },
    {
      "epoch": 0.7547352613476723,
      "grad_norm": 0.09147538244724274,
      "learning_rate": 4.672241016111761e-05,
      "loss": 0.32499605417251587,
      "memory(GiB)": 78.33,
      "step": 3895,
      "token_acc": 0.9020622260544657,
      "train_speed(iter/s)": 0.03245
    },
    {
      "epoch": 0.7549290316329991,
      "grad_norm": 0.09846985340118408,
      "learning_rate": 4.6652714555709734e-05,
      "loss": 0.32619956135749817,
      "memory(GiB)": 78.33,
      "step": 3896,
      "token_acc": 0.9018954111099443,
      "train_speed(iter/s)": 0.032451
    },
    {
      "epoch": 0.7551228019183258,
      "grad_norm": 0.09857229888439178,
      "learning_rate": 4.658306139780902e-05,
      "loss": 0.31706997752189636,
      "memory(GiB)": 78.33,
      "step": 3897,
      "token_acc": 0.9044509455647268,
      "train_speed(iter/s)": 0.032451
    },
    {
      "epoch": 0.7553165722036526,
      "grad_norm": 0.09921880811452866,
      "learning_rate": 4.6513450716023924e-05,
      "loss": 0.32637161016464233,
      "memory(GiB)": 78.33,
      "step": 3898,
      "token_acc": 0.9035741835947544,
      "train_speed(iter/s)": 0.032452
    },
    {
      "epoch": 0.7555103424889793,
      "grad_norm": 0.10863330215215683,
      "learning_rate": 4.64438825389454e-05,
      "loss": 0.33484184741973877,
      "memory(GiB)": 78.33,
      "step": 3899,
      "token_acc": 0.8991859737006888,
      "train_speed(iter/s)": 0.032453
    },
    {
      "epoch": 0.755704112774306,
      "grad_norm": 0.10528206825256348,
      "learning_rate": 4.637435689514693e-05,
      "loss": 0.3199610710144043,
      "memory(GiB)": 78.33,
      "step": 3900,
      "token_acc": 0.9030449124587668,
      "train_speed(iter/s)": 0.032453
    },
    {
      "epoch": 0.7558978830596328,
      "grad_norm": 0.09049303084611893,
      "learning_rate": 4.630487381318466e-05,
      "loss": 0.3117978572845459,
      "memory(GiB)": 78.33,
      "step": 3901,
      "token_acc": 0.906941374967325,
      "train_speed(iter/s)": 0.032454
    },
    {
      "epoch": 0.7560916533449595,
      "grad_norm": 0.09448360651731491,
      "learning_rate": 4.6235433321597124e-05,
      "loss": 0.2956331968307495,
      "memory(GiB)": 78.33,
      "step": 3902,
      "token_acc": 0.9114258416231102,
      "train_speed(iter/s)": 0.032455
    },
    {
      "epoch": 0.7562854236302863,
      "grad_norm": 0.09764683246612549,
      "learning_rate": 4.616603544890537e-05,
      "loss": 0.32111823558807373,
      "memory(GiB)": 78.33,
      "step": 3903,
      "token_acc": 0.9027792510873024,
      "train_speed(iter/s)": 0.032455
    },
    {
      "epoch": 0.756479193915613,
      "grad_norm": 0.10489743947982788,
      "learning_rate": 4.609668022361299e-05,
      "loss": 0.32993102073669434,
      "memory(GiB)": 78.33,
      "step": 3904,
      "token_acc": 0.9007670182166826,
      "train_speed(iter/s)": 0.032456
    },
    {
      "epoch": 0.7566729642009398,
      "grad_norm": 0.09750455617904663,
      "learning_rate": 4.6027367674206034e-05,
      "loss": 0.32114407420158386,
      "memory(GiB)": 78.33,
      "step": 3905,
      "token_acc": 0.9016605685336335,
      "train_speed(iter/s)": 0.032457
    },
    {
      "epoch": 0.7568667344862665,
      "grad_norm": 0.10215258598327637,
      "learning_rate": 4.595809782915298e-05,
      "loss": 0.3243396580219269,
      "memory(GiB)": 78.33,
      "step": 3906,
      "token_acc": 0.9038705677423164,
      "train_speed(iter/s)": 0.032457
    },
    {
      "epoch": 0.7570605047715933,
      "grad_norm": 0.10810741782188416,
      "learning_rate": 4.58888707169049e-05,
      "loss": 0.33883118629455566,
      "memory(GiB)": 78.33,
      "step": 3907,
      "token_acc": 0.8982128982128982,
      "train_speed(iter/s)": 0.032458
    },
    {
      "epoch": 0.75725427505692,
      "grad_norm": 0.09572840481996536,
      "learning_rate": 4.581968636589521e-05,
      "loss": 0.3282477557659149,
      "memory(GiB)": 78.33,
      "step": 3908,
      "token_acc": 0.9006182113741041,
      "train_speed(iter/s)": 0.032459
    },
    {
      "epoch": 0.7574480453422467,
      "grad_norm": 0.11090946942567825,
      "learning_rate": 4.575054480453975e-05,
      "loss": 0.3538365960121155,
      "memory(GiB)": 78.33,
      "step": 3909,
      "token_acc": 0.8962534285800404,
      "train_speed(iter/s)": 0.032459
    },
    {
      "epoch": 0.7576418156275735,
      "grad_norm": 0.10753259807825089,
      "learning_rate": 4.568144606123683e-05,
      "loss": 0.3474193513393402,
      "memory(GiB)": 78.33,
      "step": 3910,
      "token_acc": 0.8973380854262961,
      "train_speed(iter/s)": 0.03246
    },
    {
      "epoch": 0.7578355859129002,
      "grad_norm": 0.10958699882030487,
      "learning_rate": 4.561239016436716e-05,
      "loss": 0.3798098564147949,
      "memory(GiB)": 78.33,
      "step": 3911,
      "token_acc": 0.8861535899641975,
      "train_speed(iter/s)": 0.032461
    },
    {
      "epoch": 0.758029356198227,
      "grad_norm": 0.10337840020656586,
      "learning_rate": 4.5543377142293856e-05,
      "loss": 0.3272544741630554,
      "memory(GiB)": 78.33,
      "step": 3912,
      "token_acc": 0.9024021501763817,
      "train_speed(iter/s)": 0.032461
    },
    {
      "epoch": 0.7582231264835537,
      "grad_norm": 0.10478585213422775,
      "learning_rate": 4.5474407023362374e-05,
      "loss": 0.334266722202301,
      "memory(GiB)": 78.33,
      "step": 3913,
      "token_acc": 0.8995371061669404,
      "train_speed(iter/s)": 0.032462
    },
    {
      "epoch": 0.7584168967688805,
      "grad_norm": 0.10417360812425613,
      "learning_rate": 4.5405479835900685e-05,
      "loss": 0.29992321133613586,
      "memory(GiB)": 78.33,
      "step": 3914,
      "token_acc": 0.9098251531019792,
      "train_speed(iter/s)": 0.032463
    },
    {
      "epoch": 0.7586106670542072,
      "grad_norm": 0.09397785365581512,
      "learning_rate": 4.533659560821898e-05,
      "loss": 0.31968510150909424,
      "memory(GiB)": 78.33,
      "step": 3915,
      "token_acc": 0.9051098506427563,
      "train_speed(iter/s)": 0.032463
    },
    {
      "epoch": 0.758804437339534,
      "grad_norm": 0.09900712221860886,
      "learning_rate": 4.526775436860988e-05,
      "loss": 0.316898375749588,
      "memory(GiB)": 78.33,
      "step": 3916,
      "token_acc": 0.9043134151167392,
      "train_speed(iter/s)": 0.032464
    },
    {
      "epoch": 0.7589982076248607,
      "grad_norm": 0.08824368566274643,
      "learning_rate": 4.519895614534833e-05,
      "loss": 0.3089524209499359,
      "memory(GiB)": 78.33,
      "step": 3917,
      "token_acc": 0.9057074210072986,
      "train_speed(iter/s)": 0.032465
    },
    {
      "epoch": 0.7591919779101874,
      "grad_norm": 0.09777352958917618,
      "learning_rate": 4.513020096669161e-05,
      "loss": 0.31515491008758545,
      "memory(GiB)": 78.33,
      "step": 3918,
      "token_acc": 0.9042631075068986,
      "train_speed(iter/s)": 0.032465
    },
    {
      "epoch": 0.7593857481955142,
      "grad_norm": 0.10553678125143051,
      "learning_rate": 4.506148886087925e-05,
      "loss": 0.3489447236061096,
      "memory(GiB)": 78.33,
      "step": 3919,
      "token_acc": 0.8933819207711297,
      "train_speed(iter/s)": 0.032466
    },
    {
      "epoch": 0.7595795184808409,
      "grad_norm": 0.10306905955076218,
      "learning_rate": 4.4992819856133285e-05,
      "loss": 0.35067906975746155,
      "memory(GiB)": 78.33,
      "step": 3920,
      "token_acc": 0.8936492337113867,
      "train_speed(iter/s)": 0.032467
    },
    {
      "epoch": 0.7597732887661677,
      "grad_norm": 0.12322760373353958,
      "learning_rate": 4.492419398065784e-05,
      "loss": 0.42081135511398315,
      "memory(GiB)": 78.33,
      "step": 3921,
      "token_acc": 0.8778385546162771,
      "train_speed(iter/s)": 0.032468
    },
    {
      "epoch": 0.7599670590514944,
      "grad_norm": 0.0956936627626419,
      "learning_rate": 4.485561126263944e-05,
      "loss": 0.30873122811317444,
      "memory(GiB)": 78.33,
      "step": 3922,
      "token_acc": 0.9064496298078208,
      "train_speed(iter/s)": 0.032468
    },
    {
      "epoch": 0.7601608293368212,
      "grad_norm": 0.10779135674238205,
      "learning_rate": 4.4787071730246834e-05,
      "loss": 0.33779773116111755,
      "memory(GiB)": 78.33,
      "step": 3923,
      "token_acc": 0.8981919669942968,
      "train_speed(iter/s)": 0.032469
    },
    {
      "epoch": 0.7603545996221479,
      "grad_norm": 0.094157375395298,
      "learning_rate": 4.471857541163103e-05,
      "loss": 0.3073067367076874,
      "memory(GiB)": 78.33,
      "step": 3924,
      "token_acc": 0.9095217049695541,
      "train_speed(iter/s)": 0.03247
    },
    {
      "epoch": 0.7605483699074747,
      "grad_norm": 0.09603159129619598,
      "learning_rate": 4.465012233492535e-05,
      "loss": 0.31469446420669556,
      "memory(GiB)": 78.33,
      "step": 3925,
      "token_acc": 0.9049290982825726,
      "train_speed(iter/s)": 0.03247
    },
    {
      "epoch": 0.7607421401928014,
      "grad_norm": 0.09002597630023956,
      "learning_rate": 4.4581712528245226e-05,
      "loss": 0.30716589093208313,
      "memory(GiB)": 78.33,
      "step": 3926,
      "token_acc": 0.9074883267996704,
      "train_speed(iter/s)": 0.032471
    },
    {
      "epoch": 0.7609359104781281,
      "grad_norm": 0.09838228672742844,
      "learning_rate": 4.4513346019688514e-05,
      "loss": 0.3335438072681427,
      "memory(GiB)": 78.33,
      "step": 3927,
      "token_acc": 0.8979319720777014,
      "train_speed(iter/s)": 0.032471
    },
    {
      "epoch": 0.7611296807634549,
      "grad_norm": 0.10846085846424103,
      "learning_rate": 4.444502283733512e-05,
      "loss": 0.3647676110267639,
      "memory(GiB)": 78.33,
      "step": 3928,
      "token_acc": 0.8943367089359427,
      "train_speed(iter/s)": 0.032472
    },
    {
      "epoch": 0.7613234510487816,
      "grad_norm": 0.09758076071739197,
      "learning_rate": 4.437674300924724e-05,
      "loss": 0.32494428753852844,
      "memory(GiB)": 78.33,
      "step": 3929,
      "token_acc": 0.9025488420235477,
      "train_speed(iter/s)": 0.032473
    },
    {
      "epoch": 0.7615172213341084,
      "grad_norm": 0.09471366554498672,
      "learning_rate": 4.430850656346919e-05,
      "loss": 0.31988587975502014,
      "memory(GiB)": 78.33,
      "step": 3930,
      "token_acc": 0.9036783713916415,
      "train_speed(iter/s)": 0.032473
    },
    {
      "epoch": 0.7617109916194351,
      "grad_norm": 0.11201035976409912,
      "learning_rate": 4.4240313528027545e-05,
      "loss": 0.35305842757225037,
      "memory(GiB)": 78.33,
      "step": 3931,
      "token_acc": 0.8956776695356096,
      "train_speed(iter/s)": 0.032474
    },
    {
      "epoch": 0.7619047619047619,
      "grad_norm": 0.09610579162836075,
      "learning_rate": 4.417216393093102e-05,
      "loss": 0.3121870160102844,
      "memory(GiB)": 78.33,
      "step": 3932,
      "token_acc": 0.9052815243581577,
      "train_speed(iter/s)": 0.032475
    },
    {
      "epoch": 0.7620985321900886,
      "grad_norm": 0.09443981200456619,
      "learning_rate": 4.410405780017041e-05,
      "loss": 0.2963784635066986,
      "memory(GiB)": 78.33,
      "step": 3933,
      "token_acc": 0.9106874456134799,
      "train_speed(iter/s)": 0.032476
    },
    {
      "epoch": 0.7622923024754154,
      "grad_norm": 0.0956021174788475,
      "learning_rate": 4.403599516371884e-05,
      "loss": 0.28646087646484375,
      "memory(GiB)": 78.33,
      "step": 3934,
      "token_acc": 0.9147495102155052,
      "train_speed(iter/s)": 0.032476
    },
    {
      "epoch": 0.7624860727607421,
      "grad_norm": 0.08942447602748871,
      "learning_rate": 4.396797604953143e-05,
      "loss": 0.31425604224205017,
      "memory(GiB)": 78.33,
      "step": 3935,
      "token_acc": 0.9043874975732867,
      "train_speed(iter/s)": 0.032477
    },
    {
      "epoch": 0.7626798430460688,
      "grad_norm": 0.10449796169996262,
      "learning_rate": 4.3900000485545445e-05,
      "loss": 0.3526667654514313,
      "memory(GiB)": 78.33,
      "step": 3936,
      "token_acc": 0.8961538461538462,
      "train_speed(iter/s)": 0.032478
    },
    {
      "epoch": 0.7628736133313956,
      "grad_norm": 0.10736709088087082,
      "learning_rate": 4.3832068499680276e-05,
      "loss": 0.3539313077926636,
      "memory(GiB)": 78.33,
      "step": 3937,
      "token_acc": 0.8955620144206584,
      "train_speed(iter/s)": 0.032478
    },
    {
      "epoch": 0.7630673836167223,
      "grad_norm": 0.10252580791711807,
      "learning_rate": 4.376418011983741e-05,
      "loss": 0.3328987956047058,
      "memory(GiB)": 78.33,
      "step": 3938,
      "token_acc": 0.9002593667342903,
      "train_speed(iter/s)": 0.032479
    },
    {
      "epoch": 0.7632611539020491,
      "grad_norm": 0.09203853458166122,
      "learning_rate": 4.369633537390041e-05,
      "loss": 0.30216261744499207,
      "memory(GiB)": 78.33,
      "step": 3939,
      "token_acc": 0.9080275516593613,
      "train_speed(iter/s)": 0.03248
    },
    {
      "epoch": 0.7634549241873758,
      "grad_norm": 0.10732089728116989,
      "learning_rate": 4.3628534289734996e-05,
      "loss": 0.3707207441329956,
      "memory(GiB)": 78.33,
      "step": 3940,
      "token_acc": 0.8905606159184558,
      "train_speed(iter/s)": 0.03248
    },
    {
      "epoch": 0.7636486944727026,
      "grad_norm": 0.10623017698526382,
      "learning_rate": 4.3560776895188856e-05,
      "loss": 0.3556334972381592,
      "memory(GiB)": 78.33,
      "step": 3941,
      "token_acc": 0.8951635685631957,
      "train_speed(iter/s)": 0.032481
    },
    {
      "epoch": 0.7638424647580293,
      "grad_norm": 0.10812726616859436,
      "learning_rate": 4.3493063218091784e-05,
      "loss": 0.3475589156150818,
      "memory(GiB)": 78.33,
      "step": 3942,
      "token_acc": 0.8958073637165181,
      "train_speed(iter/s)": 0.032482
    },
    {
      "epoch": 0.764036235043356,
      "grad_norm": 0.10460913926362991,
      "learning_rate": 4.342539328625559e-05,
      "loss": 0.36832839250564575,
      "memory(GiB)": 78.33,
      "step": 3943,
      "token_acc": 0.8916668762733607,
      "train_speed(iter/s)": 0.032482
    },
    {
      "epoch": 0.7642300053286828,
      "grad_norm": 0.09330563992261887,
      "learning_rate": 4.335776712747416e-05,
      "loss": 0.2897856533527374,
      "memory(GiB)": 78.33,
      "step": 3944,
      "token_acc": 0.9121778725053284,
      "train_speed(iter/s)": 0.032483
    },
    {
      "epoch": 0.7644237756140095,
      "grad_norm": 0.1091540977358818,
      "learning_rate": 4.329018476952336e-05,
      "loss": 0.35247859358787537,
      "memory(GiB)": 78.33,
      "step": 3945,
      "token_acc": 0.8945392125308443,
      "train_speed(iter/s)": 0.032484
    },
    {
      "epoch": 0.7646175458993363,
      "grad_norm": 0.1135362833738327,
      "learning_rate": 4.3222646240161014e-05,
      "loss": 0.37347984313964844,
      "memory(GiB)": 78.33,
      "step": 3946,
      "token_acc": 0.8913823511524661,
      "train_speed(iter/s)": 0.032484
    },
    {
      "epoch": 0.7648113161846631,
      "grad_norm": 0.09469881653785706,
      "learning_rate": 4.315515156712714e-05,
      "loss": 0.2968969941139221,
      "memory(GiB)": 78.33,
      "step": 3947,
      "token_acc": 0.9084439266538344,
      "train_speed(iter/s)": 0.032485
    },
    {
      "epoch": 0.7650050864699899,
      "grad_norm": 0.09792491793632507,
      "learning_rate": 4.308770077814354e-05,
      "loss": 0.33856552839279175,
      "memory(GiB)": 78.33,
      "step": 3948,
      "token_acc": 0.8967915156396832,
      "train_speed(iter/s)": 0.032486
    },
    {
      "epoch": 0.7651988567553166,
      "grad_norm": 0.10359305143356323,
      "learning_rate": 4.3020293900914075e-05,
      "loss": 0.31331631541252136,
      "memory(GiB)": 78.33,
      "step": 3949,
      "token_acc": 0.9054741250373916,
      "train_speed(iter/s)": 0.032486
    },
    {
      "epoch": 0.7653926270406434,
      "grad_norm": 0.10234098881483078,
      "learning_rate": 4.295293096312457e-05,
      "loss": 0.3446482717990875,
      "memory(GiB)": 78.33,
      "step": 3950,
      "token_acc": 0.8954361313351653,
      "train_speed(iter/s)": 0.032487
    },
    {
      "epoch": 0.7655863973259701,
      "grad_norm": 0.10561185330152512,
      "learning_rate": 4.288561199244277e-05,
      "loss": 0.33969467878341675,
      "memory(GiB)": 78.33,
      "step": 3951,
      "token_acc": 0.8973197015750207,
      "train_speed(iter/s)": 0.032487
    },
    {
      "epoch": 0.7657801676112969,
      "grad_norm": 0.10193517059087753,
      "learning_rate": 4.281833701651841e-05,
      "loss": 0.34807661175727844,
      "memory(GiB)": 78.33,
      "step": 3952,
      "token_acc": 0.8959937746531784,
      "train_speed(iter/s)": 0.032488
    },
    {
      "epoch": 0.7659739378966236,
      "grad_norm": 0.10420801490545273,
      "learning_rate": 4.275110606298307e-05,
      "loss": 0.34150460362434387,
      "memory(GiB)": 78.33,
      "step": 3953,
      "token_acc": 0.8981601881311385,
      "train_speed(iter/s)": 0.032489
    },
    {
      "epoch": 0.7661677081819503,
      "grad_norm": 0.1057552695274353,
      "learning_rate": 4.26839191594504e-05,
      "loss": 0.34214961528778076,
      "memory(GiB)": 78.33,
      "step": 3954,
      "token_acc": 0.8992475734944477,
      "train_speed(iter/s)": 0.032489
    },
    {
      "epoch": 0.7663614784672771,
      "grad_norm": 0.10361424833536148,
      "learning_rate": 4.2616776333515844e-05,
      "loss": 0.3399190306663513,
      "memory(GiB)": 78.33,
      "step": 3955,
      "token_acc": 0.895477586158072,
      "train_speed(iter/s)": 0.03249
    },
    {
      "epoch": 0.7665552487526038,
      "grad_norm": 0.10096311569213867,
      "learning_rate": 4.254967761275672e-05,
      "loss": 0.3292155861854553,
      "memory(GiB)": 78.33,
      "step": 3956,
      "token_acc": 0.8995320229977269,
      "train_speed(iter/s)": 0.032491
    },
    {
      "epoch": 0.7667490190379306,
      "grad_norm": 0.10260067880153656,
      "learning_rate": 4.2482623024732334e-05,
      "loss": 0.3111547827720642,
      "memory(GiB)": 78.33,
      "step": 3957,
      "token_acc": 0.9036062791684345,
      "train_speed(iter/s)": 0.032491
    },
    {
      "epoch": 0.7669427893232573,
      "grad_norm": 0.10281860083341599,
      "learning_rate": 4.241561259698376e-05,
      "loss": 0.3256504237651825,
      "memory(GiB)": 78.33,
      "step": 3958,
      "token_acc": 0.900353960698157,
      "train_speed(iter/s)": 0.032492
    },
    {
      "epoch": 0.7671365596085841,
      "grad_norm": 0.1012672632932663,
      "learning_rate": 4.2348646357033944e-05,
      "loss": 0.3229145407676697,
      "memory(GiB)": 78.33,
      "step": 3959,
      "token_acc": 0.902934926196083,
      "train_speed(iter/s)": 0.032493
    },
    {
      "epoch": 0.7673303298939108,
      "grad_norm": 0.11049380153417587,
      "learning_rate": 4.228172433238783e-05,
      "loss": 0.3576306700706482,
      "memory(GiB)": 78.33,
      "step": 3960,
      "token_acc": 0.892226424530366,
      "train_speed(iter/s)": 0.032493
    },
    {
      "epoch": 0.7675241001792376,
      "grad_norm": 0.10522151738405228,
      "learning_rate": 4.2214846550532026e-05,
      "loss": 0.34801146388053894,
      "memory(GiB)": 78.33,
      "step": 3961,
      "token_acc": 0.8971957437375305,
      "train_speed(iter/s)": 0.032494
    },
    {
      "epoch": 0.7677178704645643,
      "grad_norm": 0.10627347230911255,
      "learning_rate": 4.2148013038935054e-05,
      "loss": 0.34363991022109985,
      "memory(GiB)": 78.33,
      "step": 3962,
      "token_acc": 0.8970234186911797,
      "train_speed(iter/s)": 0.032495
    },
    {
      "epoch": 0.767911640749891,
      "grad_norm": 0.09957989305257797,
      "learning_rate": 4.2081223825047214e-05,
      "loss": 0.3352537453174591,
      "memory(GiB)": 78.33,
      "step": 3963,
      "token_acc": 0.8995469112609128,
      "train_speed(iter/s)": 0.032495
    },
    {
      "epoch": 0.7681054110352178,
      "grad_norm": 0.10130537301301956,
      "learning_rate": 4.201447893630065e-05,
      "loss": 0.331424355506897,
      "memory(GiB)": 78.33,
      "step": 3964,
      "token_acc": 0.9031016042780748,
      "train_speed(iter/s)": 0.032496
    },
    {
      "epoch": 0.7682991813205445,
      "grad_norm": 0.08747418969869614,
      "learning_rate": 4.194777840010926e-05,
      "loss": 0.29280680418014526,
      "memory(GiB)": 78.33,
      "step": 3965,
      "token_acc": 0.9117611623816294,
      "train_speed(iter/s)": 0.032497
    },
    {
      "epoch": 0.7684929516058713,
      "grad_norm": 0.09481216967105865,
      "learning_rate": 4.1881122243868715e-05,
      "loss": 0.34195125102996826,
      "memory(GiB)": 78.33,
      "step": 3966,
      "token_acc": 0.8987415491352898,
      "train_speed(iter/s)": 0.032497
    },
    {
      "epoch": 0.768686721891198,
      "grad_norm": 0.09969401359558105,
      "learning_rate": 4.181451049495657e-05,
      "loss": 0.3363822400569916,
      "memory(GiB)": 78.33,
      "step": 3967,
      "token_acc": 0.9003247963367232,
      "train_speed(iter/s)": 0.032498
    },
    {
      "epoch": 0.7688804921765248,
      "grad_norm": 0.10244782269001007,
      "learning_rate": 4.174794318073202e-05,
      "loss": 0.3243018090724945,
      "memory(GiB)": 78.33,
      "step": 3968,
      "token_acc": 0.900588806951352,
      "train_speed(iter/s)": 0.032499
    },
    {
      "epoch": 0.7690742624618515,
      "grad_norm": 0.09450022131204605,
      "learning_rate": 4.168142032853605e-05,
      "loss": 0.3295527398586273,
      "memory(GiB)": 78.33,
      "step": 3969,
      "token_acc": 0.9003620957213587,
      "train_speed(iter/s)": 0.032499
    },
    {
      "epoch": 0.7692680327471783,
      "grad_norm": 0.09936364740133286,
      "learning_rate": 4.16149419656914e-05,
      "loss": 0.31701233983039856,
      "memory(GiB)": 78.33,
      "step": 3970,
      "token_acc": 0.9036584700943242,
      "train_speed(iter/s)": 0.0325
    },
    {
      "epoch": 0.769461803032505,
      "grad_norm": 0.11445298790931702,
      "learning_rate": 4.15485081195025e-05,
      "loss": 0.3477141261100769,
      "memory(GiB)": 78.33,
      "step": 3971,
      "token_acc": 0.8983733184957288,
      "train_speed(iter/s)": 0.032501
    },
    {
      "epoch": 0.7696555733178317,
      "grad_norm": 0.09958308935165405,
      "learning_rate": 4.148211881725547e-05,
      "loss": 0.31291618943214417,
      "memory(GiB)": 78.33,
      "step": 3972,
      "token_acc": 0.9050311502015601,
      "train_speed(iter/s)": 0.032501
    },
    {
      "epoch": 0.7698493436031585,
      "grad_norm": 0.10363567620515823,
      "learning_rate": 4.141577408621827e-05,
      "loss": 0.33989396691322327,
      "memory(GiB)": 78.33,
      "step": 3973,
      "token_acc": 0.8975318087789396,
      "train_speed(iter/s)": 0.032502
    },
    {
      "epoch": 0.7700431138884852,
      "grad_norm": 0.10342077165842056,
      "learning_rate": 4.134947395364043e-05,
      "loss": 0.33245235681533813,
      "memory(GiB)": 78.33,
      "step": 3974,
      "token_acc": 0.8993934211581113,
      "train_speed(iter/s)": 0.032502
    },
    {
      "epoch": 0.770236884173812,
      "grad_norm": 0.10198888927698135,
      "learning_rate": 4.128321844675318e-05,
      "loss": 0.32281598448753357,
      "memory(GiB)": 78.33,
      "step": 3975,
      "token_acc": 0.9067534827377347,
      "train_speed(iter/s)": 0.032503
    },
    {
      "epoch": 0.7704306544591387,
      "grad_norm": 0.09553690254688263,
      "learning_rate": 4.121700759276946e-05,
      "loss": 0.31830698251724243,
      "memory(GiB)": 78.33,
      "step": 3976,
      "token_acc": 0.905037092785722,
      "train_speed(iter/s)": 0.032504
    },
    {
      "epoch": 0.7706244247444655,
      "grad_norm": 0.09600663185119629,
      "learning_rate": 4.1150841418883845e-05,
      "loss": 0.3234773874282837,
      "memory(GiB)": 78.33,
      "step": 3977,
      "token_acc": 0.904327039295211,
      "train_speed(iter/s)": 0.032504
    },
    {
      "epoch": 0.7708181950297922,
      "grad_norm": 0.10290994495153427,
      "learning_rate": 4.1084719952272524e-05,
      "loss": 0.32649847865104675,
      "memory(GiB)": 78.33,
      "step": 3978,
      "token_acc": 0.9009803039841713,
      "train_speed(iter/s)": 0.032505
    },
    {
      "epoch": 0.771011965315119,
      "grad_norm": 0.09713493287563324,
      "learning_rate": 4.101864322009335e-05,
      "loss": 0.326259970664978,
      "memory(GiB)": 78.33,
      "step": 3979,
      "token_acc": 0.9027269351159213,
      "train_speed(iter/s)": 0.032506
    },
    {
      "epoch": 0.7712057356004457,
      "grad_norm": 0.1079777181148529,
      "learning_rate": 4.0952611249485906e-05,
      "loss": 0.33296194672584534,
      "memory(GiB)": 78.33,
      "step": 3980,
      "token_acc": 0.9001692971437933,
      "train_speed(iter/s)": 0.032506
    },
    {
      "epoch": 0.7713995058857724,
      "grad_norm": 0.10135926306247711,
      "learning_rate": 4.0886624067571215e-05,
      "loss": 0.3391050100326538,
      "memory(GiB)": 78.33,
      "step": 3981,
      "token_acc": 0.9001635300313202,
      "train_speed(iter/s)": 0.032507
    },
    {
      "epoch": 0.7715932761710992,
      "grad_norm": 0.09656066447496414,
      "learning_rate": 4.0820681701452034e-05,
      "loss": 0.3262328803539276,
      "memory(GiB)": 78.33,
      "step": 3982,
      "token_acc": 0.9005832056541294,
      "train_speed(iter/s)": 0.032508
    },
    {
      "epoch": 0.7717870464564259,
      "grad_norm": 0.0958666205406189,
      "learning_rate": 4.0754784178212616e-05,
      "loss": 0.30787983536720276,
      "memory(GiB)": 78.33,
      "step": 3983,
      "token_acc": 0.9089610766847405,
      "train_speed(iter/s)": 0.032508
    },
    {
      "epoch": 0.7719808167417527,
      "grad_norm": 0.11135300248861313,
      "learning_rate": 4.068893152491888e-05,
      "loss": 0.3504694402217865,
      "memory(GiB)": 78.33,
      "step": 3984,
      "token_acc": 0.8946419448868588,
      "train_speed(iter/s)": 0.032509
    },
    {
      "epoch": 0.7721745870270794,
      "grad_norm": 0.09396642446517944,
      "learning_rate": 4.062312376861828e-05,
      "loss": 0.2973524332046509,
      "memory(GiB)": 78.33,
      "step": 3985,
      "token_acc": 0.9072842658845541,
      "train_speed(iter/s)": 0.032509
    },
    {
      "epoch": 0.7723683573124062,
      "grad_norm": 0.10528262704610825,
      "learning_rate": 4.0557360936339754e-05,
      "loss": 0.3348216414451599,
      "memory(GiB)": 78.33,
      "step": 3986,
      "token_acc": 0.8983733985893191,
      "train_speed(iter/s)": 0.03251
    },
    {
      "epoch": 0.7725621275977329,
      "grad_norm": 0.10733997821807861,
      "learning_rate": 4.049164305509398e-05,
      "loss": 0.36273688077926636,
      "memory(GiB)": 78.33,
      "step": 3987,
      "token_acc": 0.8929391901242769,
      "train_speed(iter/s)": 0.032511
    },
    {
      "epoch": 0.7727558978830597,
      "grad_norm": 0.09470584243535995,
      "learning_rate": 4.042597015187301e-05,
      "loss": 0.3208099603652954,
      "memory(GiB)": 78.33,
      "step": 3988,
      "token_acc": 0.9038011841385144,
      "train_speed(iter/s)": 0.032511
    },
    {
      "epoch": 0.7729496681683864,
      "grad_norm": 0.10173063725233078,
      "learning_rate": 4.036034225365047e-05,
      "loss": 0.33486083149909973,
      "memory(GiB)": 78.33,
      "step": 3989,
      "token_acc": 0.9000950410912953,
      "train_speed(iter/s)": 0.032512
    },
    {
      "epoch": 0.7731434384537131,
      "grad_norm": 0.09967856109142303,
      "learning_rate": 4.029475938738149e-05,
      "loss": 0.31119605898857117,
      "memory(GiB)": 78.33,
      "step": 3990,
      "token_acc": 0.9062775816416593,
      "train_speed(iter/s)": 0.032513
    },
    {
      "epoch": 0.7733372087390399,
      "grad_norm": 0.10067766904830933,
      "learning_rate": 4.0229221580002736e-05,
      "loss": 0.31823796033859253,
      "memory(GiB)": 78.33,
      "step": 3991,
      "token_acc": 0.9016983016983017,
      "train_speed(iter/s)": 0.032513
    },
    {
      "epoch": 0.7735309790243666,
      "grad_norm": 0.10268165916204453,
      "learning_rate": 4.016372885843228e-05,
      "loss": 0.3410794138908386,
      "memory(GiB)": 78.33,
      "step": 3992,
      "token_acc": 0.8990299823633157,
      "train_speed(iter/s)": 0.032514
    },
    {
      "epoch": 0.7737247493096934,
      "grad_norm": 0.1047457829117775,
      "learning_rate": 4.0098281249569845e-05,
      "loss": 0.3615070581436157,
      "memory(GiB)": 78.33,
      "step": 3993,
      "token_acc": 0.8931236597569693,
      "train_speed(iter/s)": 0.032515
    },
    {
      "epoch": 0.7739185195950201,
      "grad_norm": 0.10737626999616623,
      "learning_rate": 4.0032878780296476e-05,
      "loss": 0.3728790879249573,
      "memory(GiB)": 78.33,
      "step": 3994,
      "token_acc": 0.8893093379417198,
      "train_speed(iter/s)": 0.032515
    },
    {
      "epoch": 0.7741122898803469,
      "grad_norm": 0.10214757919311523,
      "learning_rate": 3.9967521477474726e-05,
      "loss": 0.34999844431877136,
      "memory(GiB)": 78.33,
      "step": 3995,
      "token_acc": 0.8950407514850117,
      "train_speed(iter/s)": 0.032516
    },
    {
      "epoch": 0.7743060601656736,
      "grad_norm": 0.09131192415952682,
      "learning_rate": 3.990220936794859e-05,
      "loss": 0.302643358707428,
      "memory(GiB)": 78.33,
      "step": 3996,
      "token_acc": 0.9078601502709995,
      "train_speed(iter/s)": 0.032517
    },
    {
      "epoch": 0.7744998304510003,
      "grad_norm": 0.10554607212543488,
      "learning_rate": 3.98369424785435e-05,
      "loss": 0.33304092288017273,
      "memory(GiB)": 78.33,
      "step": 3997,
      "token_acc": 0.9006529325678262,
      "train_speed(iter/s)": 0.032517
    },
    {
      "epoch": 0.7746936007363271,
      "grad_norm": 0.0956023558974266,
      "learning_rate": 3.977172083606634e-05,
      "loss": 0.32441043853759766,
      "memory(GiB)": 78.33,
      "step": 3998,
      "token_acc": 0.9003093260588937,
      "train_speed(iter/s)": 0.032518
    },
    {
      "epoch": 0.7748873710216538,
      "grad_norm": 0.09569355845451355,
      "learning_rate": 3.9706544467305316e-05,
      "loss": 0.2999122738838196,
      "memory(GiB)": 78.33,
      "step": 3999,
      "token_acc": 0.9101693571824883,
      "train_speed(iter/s)": 0.032519
    },
    {
      "epoch": 0.7750811413069806,
      "grad_norm": 0.1037716194987297,
      "learning_rate": 3.964141339903026e-05,
      "loss": 0.33561694622039795,
      "memory(GiB)": 78.33,
      "step": 4000,
      "token_acc": 0.9002222414639074,
      "train_speed(iter/s)": 0.03252
    },
    {
      "epoch": 0.7750811413069806,
      "eval_loss": 0.38567423820495605,
      "eval_runtime": 1344.6983,
      "eval_samples_per_second": 5.019,
      "eval_steps_per_second": 5.019,
      "eval_token_acc": 0.9009321879927275,
      "step": 4000
    },
    {
      "epoch": 0.7752749115923073,
      "grad_norm": 0.09706872701644897,
      "learning_rate": 3.9576327657992144e-05,
      "loss": 0.31833964586257935,
      "memory(GiB)": 78.33,
      "step": 4001,
      "token_acc": 0.9042262932234534,
      "train_speed(iter/s)": 0.032165
    },
    {
      "epoch": 0.7754686818776341,
      "grad_norm": 0.10196474194526672,
      "learning_rate": 3.951128727092346e-05,
      "loss": 0.3550868630409241,
      "memory(GiB)": 78.33,
      "step": 4002,
      "token_acc": 0.8949007501286321,
      "train_speed(iter/s)": 0.032166
    },
    {
      "epoch": 0.7756624521629608,
      "grad_norm": 0.09914345294237137,
      "learning_rate": 3.9446292264538046e-05,
      "loss": 0.33100420236587524,
      "memory(GiB)": 78.33,
      "step": 4003,
      "token_acc": 0.900188852592895,
      "train_speed(iter/s)": 0.032167
    },
    {
      "epoch": 0.7758562224482876,
      "grad_norm": 0.11188572645187378,
      "learning_rate": 3.93813426655311e-05,
      "loss": 0.35677969455718994,
      "memory(GiB)": 78.33,
      "step": 4004,
      "token_acc": 0.8938544116760148,
      "train_speed(iter/s)": 0.032168
    },
    {
      "epoch": 0.7760499927336143,
      "grad_norm": 0.10255058854818344,
      "learning_rate": 3.9316438500579103e-05,
      "loss": 0.32014256715774536,
      "memory(GiB)": 78.33,
      "step": 4005,
      "token_acc": 0.9037872184890489,
      "train_speed(iter/s)": 0.032168
    },
    {
      "epoch": 0.776243763018941,
      "grad_norm": 0.10134012997150421,
      "learning_rate": 3.925157979634005e-05,
      "loss": 0.32840850949287415,
      "memory(GiB)": 78.33,
      "step": 4006,
      "token_acc": 0.9019783197831979,
      "train_speed(iter/s)": 0.032169
    },
    {
      "epoch": 0.7764375333042678,
      "grad_norm": 0.10212652385234833,
      "learning_rate": 3.918676657945308e-05,
      "loss": 0.3236843943595886,
      "memory(GiB)": 78.33,
      "step": 4007,
      "token_acc": 0.9032230290158711,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.7766313035895945,
      "grad_norm": 0.09699150919914246,
      "learning_rate": 3.9121998876538775e-05,
      "loss": 0.3039090037345886,
      "memory(GiB)": 78.33,
      "step": 4008,
      "token_acc": 0.9065003465003465,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.7768250738749213,
      "grad_norm": 0.10348644107580185,
      "learning_rate": 3.905727671419891e-05,
      "loss": 0.36719679832458496,
      "memory(GiB)": 78.33,
      "step": 4009,
      "token_acc": 0.8896081591671731,
      "train_speed(iter/s)": 0.032171
    },
    {
      "epoch": 0.777018844160248,
      "grad_norm": 0.09221908450126648,
      "learning_rate": 3.899260011901666e-05,
      "loss": 0.2919962406158447,
      "memory(GiB)": 78.33,
      "step": 4010,
      "token_acc": 0.9112894709050586,
      "train_speed(iter/s)": 0.032172
    },
    {
      "epoch": 0.7772126144455748,
      "grad_norm": 0.10690614581108093,
      "learning_rate": 3.892796911755642e-05,
      "loss": 0.3488962650299072,
      "memory(GiB)": 78.33,
      "step": 4011,
      "token_acc": 0.8949435444280806,
      "train_speed(iter/s)": 0.032173
    },
    {
      "epoch": 0.7774063847309015,
      "grad_norm": 0.11350135505199432,
      "learning_rate": 3.886338373636385e-05,
      "loss": 0.36426064372062683,
      "memory(GiB)": 78.33,
      "step": 4012,
      "token_acc": 0.892725139337049,
      "train_speed(iter/s)": 0.032173
    },
    {
      "epoch": 0.7776001550162283,
      "grad_norm": 0.10248113423585892,
      "learning_rate": 3.8798844001965976e-05,
      "loss": 0.3667606711387634,
      "memory(GiB)": 78.33,
      "step": 4013,
      "token_acc": 0.8912181013079423,
      "train_speed(iter/s)": 0.032174
    },
    {
      "epoch": 0.777793925301555,
      "grad_norm": 0.09691166877746582,
      "learning_rate": 3.873434994087095e-05,
      "loss": 0.3389664888381958,
      "memory(GiB)": 78.33,
      "step": 4014,
      "token_acc": 0.8992298909214586,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.7779876955868817,
      "grad_norm": 0.09862842410802841,
      "learning_rate": 3.866990157956823e-05,
      "loss": 0.35453882813453674,
      "memory(GiB)": 78.33,
      "step": 4015,
      "token_acc": 0.8935185185185185,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.7781814658722085,
      "grad_norm": 0.11863162368535995,
      "learning_rate": 3.86054989445285e-05,
      "loss": 0.3620249032974243,
      "memory(GiB)": 78.33,
      "step": 4016,
      "token_acc": 0.8918080939947781,
      "train_speed(iter/s)": 0.032176
    },
    {
      "epoch": 0.7783752361575352,
      "grad_norm": 0.10086744278669357,
      "learning_rate": 3.854114206220364e-05,
      "loss": 0.3244988024234772,
      "memory(GiB)": 78.33,
      "step": 4017,
      "token_acc": 0.9012764208850194,
      "train_speed(iter/s)": 0.032177
    },
    {
      "epoch": 0.778569006442862,
      "grad_norm": 0.12051405757665634,
      "learning_rate": 3.8476830959026735e-05,
      "loss": 0.38493812084198,
      "memory(GiB)": 78.33,
      "step": 4018,
      "token_acc": 0.8884372177055104,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.7787627767281887,
      "grad_norm": 0.10893973708152771,
      "learning_rate": 3.8412565661412056e-05,
      "loss": 0.35063496232032776,
      "memory(GiB)": 78.33,
      "step": 4019,
      "token_acc": 0.8946015424164524,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.7789565470135155,
      "grad_norm": 0.09621429443359375,
      "learning_rate": 3.834834619575519e-05,
      "loss": 0.3024739623069763,
      "memory(GiB)": 78.33,
      "step": 4020,
      "token_acc": 0.9086043745798645,
      "train_speed(iter/s)": 0.032179
    },
    {
      "epoch": 0.7791503172988422,
      "grad_norm": 0.09711124747991562,
      "learning_rate": 3.8284172588432716e-05,
      "loss": 0.32116764783859253,
      "memory(GiB)": 78.33,
      "step": 4021,
      "token_acc": 0.9038179359553471,
      "train_speed(iter/s)": 0.03218
    },
    {
      "epoch": 0.779344087584169,
      "grad_norm": 0.10393381118774414,
      "learning_rate": 3.822004486580251e-05,
      "loss": 0.32372620701789856,
      "memory(GiB)": 78.33,
      "step": 4022,
      "token_acc": 0.9042860117188661,
      "train_speed(iter/s)": 0.032181
    },
    {
      "epoch": 0.7795378578694957,
      "grad_norm": 0.1017252653837204,
      "learning_rate": 3.815596305420349e-05,
      "loss": 0.353562593460083,
      "memory(GiB)": 78.33,
      "step": 4023,
      "token_acc": 0.8964561961709364,
      "train_speed(iter/s)": 0.032181
    },
    {
      "epoch": 0.7797316281548224,
      "grad_norm": 0.09764081239700317,
      "learning_rate": 3.809192717995584e-05,
      "loss": 0.3372447192668915,
      "memory(GiB)": 78.33,
      "step": 4024,
      "token_acc": 0.9003608608738545,
      "train_speed(iter/s)": 0.032182
    },
    {
      "epoch": 0.7799253984401492,
      "grad_norm": 0.1030493900179863,
      "learning_rate": 3.8027937269360757e-05,
      "loss": 0.32105734944343567,
      "memory(GiB)": 78.33,
      "step": 4025,
      "token_acc": 0.9018271999087826,
      "train_speed(iter/s)": 0.032183
    },
    {
      "epoch": 0.7801191687254759,
      "grad_norm": 0.10457610338926315,
      "learning_rate": 3.796399334870061e-05,
      "loss": 0.3225201964378357,
      "memory(GiB)": 78.33,
      "step": 4026,
      "token_acc": 0.9034423897581793,
      "train_speed(iter/s)": 0.032184
    },
    {
      "epoch": 0.7803129390108027,
      "grad_norm": 0.10154607892036438,
      "learning_rate": 3.7900095444238965e-05,
      "loss": 0.3304111063480377,
      "memory(GiB)": 78.33,
      "step": 4027,
      "token_acc": 0.9023520164459762,
      "train_speed(iter/s)": 0.032184
    },
    {
      "epoch": 0.7805067092961294,
      "grad_norm": 0.09185813367366791,
      "learning_rate": 3.783624358222036e-05,
      "loss": 0.29117369651794434,
      "memory(GiB)": 78.33,
      "step": 4028,
      "token_acc": 0.9135486512770641,
      "train_speed(iter/s)": 0.032185
    },
    {
      "epoch": 0.7807004795814562,
      "grad_norm": 0.10176742821931839,
      "learning_rate": 3.777243778887047e-05,
      "loss": 0.33322426676750183,
      "memory(GiB)": 78.33,
      "step": 4029,
      "token_acc": 0.8997748686733928,
      "train_speed(iter/s)": 0.032186
    },
    {
      "epoch": 0.7808942498667829,
      "grad_norm": 0.10486084967851639,
      "learning_rate": 3.770867809039604e-05,
      "loss": 0.3594956398010254,
      "memory(GiB)": 78.33,
      "step": 4030,
      "token_acc": 0.8941469133331427,
      "train_speed(iter/s)": 0.032186
    },
    {
      "epoch": 0.7810880201521097,
      "grad_norm": 0.09931764006614685,
      "learning_rate": 3.764496451298492e-05,
      "loss": 0.30685773491859436,
      "memory(GiB)": 78.33,
      "step": 4031,
      "token_acc": 0.9082772375966093,
      "train_speed(iter/s)": 0.032187
    },
    {
      "epoch": 0.7812817904374364,
      "grad_norm": 0.1024843156337738,
      "learning_rate": 3.758129708280593e-05,
      "loss": 0.3311740458011627,
      "memory(GiB)": 78.33,
      "step": 4032,
      "token_acc": 0.8998150183641189,
      "train_speed(iter/s)": 0.032188
    },
    {
      "epoch": 0.7814755607227631,
      "grad_norm": 0.10810218751430511,
      "learning_rate": 3.751767582600908e-05,
      "loss": 0.3451961576938629,
      "memory(GiB)": 78.33,
      "step": 4033,
      "token_acc": 0.8986039894484128,
      "train_speed(iter/s)": 0.032188
    },
    {
      "epoch": 0.7816693310080899,
      "grad_norm": 0.11089378595352173,
      "learning_rate": 3.745410076872528e-05,
      "loss": 0.3630661964416504,
      "memory(GiB)": 78.33,
      "step": 4034,
      "token_acc": 0.8932108218478816,
      "train_speed(iter/s)": 0.032189
    },
    {
      "epoch": 0.7818631012934166,
      "grad_norm": 0.09974395483732224,
      "learning_rate": 3.739057193706651e-05,
      "loss": 0.3139222264289856,
      "memory(GiB)": 78.33,
      "step": 4035,
      "token_acc": 0.9046637100646875,
      "train_speed(iter/s)": 0.03219
    },
    {
      "epoch": 0.7820568715787434,
      "grad_norm": 0.10395882278680801,
      "learning_rate": 3.7327089357125794e-05,
      "loss": 0.3373940885066986,
      "memory(GiB)": 78.33,
      "step": 4036,
      "token_acc": 0.9000392222782541,
      "train_speed(iter/s)": 0.032191
    },
    {
      "epoch": 0.7822506418640701,
      "grad_norm": 0.09552538394927979,
      "learning_rate": 3.7263653054977106e-05,
      "loss": 0.3279712200164795,
      "memory(GiB)": 78.33,
      "step": 4037,
      "token_acc": 0.9012946233238103,
      "train_speed(iter/s)": 0.032191
    },
    {
      "epoch": 0.7824444121493969,
      "grad_norm": 0.09712161868810654,
      "learning_rate": 3.7200263056675424e-05,
      "loss": 0.3257930278778076,
      "memory(GiB)": 78.33,
      "step": 4038,
      "token_acc": 0.9013559672911707,
      "train_speed(iter/s)": 0.032192
    },
    {
      "epoch": 0.7826381824347236,
      "grad_norm": 0.09889403730630875,
      "learning_rate": 3.713691938825677e-05,
      "loss": 0.30284392833709717,
      "memory(GiB)": 78.33,
      "step": 4039,
      "token_acc": 0.9077128339058109,
      "train_speed(iter/s)": 0.032193
    },
    {
      "epoch": 0.7828319527200504,
      "grad_norm": 0.11424127966165543,
      "learning_rate": 3.7073622075738085e-05,
      "loss": 0.340787798166275,
      "memory(GiB)": 78.33,
      "step": 4040,
      "token_acc": 0.8969184444871549,
      "train_speed(iter/s)": 0.032193
    },
    {
      "epoch": 0.7830257230053771,
      "grad_norm": 0.1108805313706398,
      "learning_rate": 3.701037114511727e-05,
      "loss": 0.3701401948928833,
      "memory(GiB)": 78.33,
      "step": 4041,
      "token_acc": 0.8902578623930594,
      "train_speed(iter/s)": 0.032194
    },
    {
      "epoch": 0.7832194932907038,
      "grad_norm": 0.10004852712154388,
      "learning_rate": 3.694716662237317e-05,
      "loss": 0.3417167067527771,
      "memory(GiB)": 78.33,
      "step": 4042,
      "token_acc": 0.8959614939126406,
      "train_speed(iter/s)": 0.032195
    },
    {
      "epoch": 0.7834132635760306,
      "grad_norm": 0.09257663786411285,
      "learning_rate": 3.6884008533465575e-05,
      "loss": 0.2933918833732605,
      "memory(GiB)": 78.33,
      "step": 4043,
      "token_acc": 0.9114590016681637,
      "train_speed(iter/s)": 0.032195
    },
    {
      "epoch": 0.7836070338613573,
      "grad_norm": 0.10794571042060852,
      "learning_rate": 3.682089690433522e-05,
      "loss": 0.34912464022636414,
      "memory(GiB)": 78.33,
      "step": 4044,
      "token_acc": 0.8977625199506128,
      "train_speed(iter/s)": 0.032196
    },
    {
      "epoch": 0.7838008041466841,
      "grad_norm": 0.09921461343765259,
      "learning_rate": 3.675783176090373e-05,
      "loss": 0.3479025363922119,
      "memory(GiB)": 78.33,
      "step": 4045,
      "token_acc": 0.8996169970498421,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.7839945744320108,
      "grad_norm": 0.091604083776474,
      "learning_rate": 3.669481312907369e-05,
      "loss": 0.3121592104434967,
      "memory(GiB)": 78.33,
      "step": 4046,
      "token_acc": 0.9041123022455497,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.7841883447173376,
      "grad_norm": 0.09730125963687897,
      "learning_rate": 3.663184103472852e-05,
      "loss": 0.35410815477371216,
      "memory(GiB)": 78.33,
      "step": 4047,
      "token_acc": 0.8955283408408409,
      "train_speed(iter/s)": 0.032198
    },
    {
      "epoch": 0.7843821150026643,
      "grad_norm": 0.09932620078325272,
      "learning_rate": 3.6568915503732577e-05,
      "loss": 0.3147258162498474,
      "memory(GiB)": 78.33,
      "step": 4048,
      "token_acc": 0.9058458354888774,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.784575885287991,
      "grad_norm": 0.09612176567316055,
      "learning_rate": 3.650603656193105e-05,
      "loss": 0.3254881203174591,
      "memory(GiB)": 78.33,
      "step": 4049,
      "token_acc": 0.902174750301692,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.7847696555733178,
      "grad_norm": 0.08921913802623749,
      "learning_rate": 3.6443204235149995e-05,
      "loss": 0.30229154229164124,
      "memory(GiB)": 78.33,
      "step": 4050,
      "token_acc": 0.9090634861704102,
      "train_speed(iter/s)": 0.0322
    },
    {
      "epoch": 0.7849634258586445,
      "grad_norm": 0.09413807839155197,
      "learning_rate": 3.638041854919634e-05,
      "loss": 0.32846495509147644,
      "memory(GiB)": 78.33,
      "step": 4051,
      "token_acc": 0.901771947119638,
      "train_speed(iter/s)": 0.032201
    },
    {
      "epoch": 0.7851571961439713,
      "grad_norm": 0.09915482252836227,
      "learning_rate": 3.6317679529857844e-05,
      "loss": 0.32399147748947144,
      "memory(GiB)": 78.33,
      "step": 4052,
      "token_acc": 0.903708087800108,
      "train_speed(iter/s)": 0.032202
    },
    {
      "epoch": 0.785350966429298,
      "grad_norm": 0.10495869070291519,
      "learning_rate": 3.625498720290315e-05,
      "loss": 0.3514251112937927,
      "memory(GiB)": 78.33,
      "step": 4053,
      "token_acc": 0.8951106706309878,
      "train_speed(iter/s)": 0.032202
    },
    {
      "epoch": 0.7855447367146248,
      "grad_norm": 0.10493289679288864,
      "learning_rate": 3.619234159408168e-05,
      "loss": 0.34921297430992126,
      "memory(GiB)": 78.33,
      "step": 4054,
      "token_acc": 0.8932814420319541,
      "train_speed(iter/s)": 0.032203
    },
    {
      "epoch": 0.7857385069999515,
      "grad_norm": 0.09735289216041565,
      "learning_rate": 3.6129742729123625e-05,
      "loss": 0.30153387784957886,
      "memory(GiB)": 78.33,
      "step": 4055,
      "token_acc": 0.9098055563826937,
      "train_speed(iter/s)": 0.032204
    },
    {
      "epoch": 0.7859322772852783,
      "grad_norm": 0.09159845113754272,
      "learning_rate": 3.606719063374006e-05,
      "loss": 0.30557680130004883,
      "memory(GiB)": 78.33,
      "step": 4056,
      "token_acc": 0.907328966162636,
      "train_speed(iter/s)": 0.032204
    },
    {
      "epoch": 0.786126047570605,
      "grad_norm": 0.0924990177154541,
      "learning_rate": 3.600468533362279e-05,
      "loss": 0.2820108234882355,
      "memory(GiB)": 78.33,
      "step": 4057,
      "token_acc": 0.9154507656948067,
      "train_speed(iter/s)": 0.032205
    },
    {
      "epoch": 0.7863198178559317,
      "grad_norm": 0.1056913435459137,
      "learning_rate": 3.594222685444441e-05,
      "loss": 0.33108171820640564,
      "memory(GiB)": 78.33,
      "step": 4058,
      "token_acc": 0.8996641845524894,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.7865135881412585,
      "grad_norm": 0.11276807636022568,
      "learning_rate": 3.587981522185829e-05,
      "loss": 0.31702518463134766,
      "memory(GiB)": 78.33,
      "step": 4059,
      "token_acc": 0.9057384014048604,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.7867073584265852,
      "grad_norm": 0.09739526361227036,
      "learning_rate": 3.5817450461498634e-05,
      "loss": 0.3203843832015991,
      "memory(GiB)": 78.33,
      "step": 4060,
      "token_acc": 0.9029685310230603,
      "train_speed(iter/s)": 0.032207
    },
    {
      "epoch": 0.786901128711912,
      "grad_norm": 0.09344157576560974,
      "learning_rate": 3.575513259898027e-05,
      "loss": 0.31456148624420166,
      "memory(GiB)": 78.33,
      "step": 4061,
      "token_acc": 0.9028077753779697,
      "train_speed(iter/s)": 0.032208
    },
    {
      "epoch": 0.7870948989972387,
      "grad_norm": 0.09507586807012558,
      "learning_rate": 3.569286165989881e-05,
      "loss": 0.326376736164093,
      "memory(GiB)": 78.33,
      "step": 4062,
      "token_acc": 0.900634411067798,
      "train_speed(iter/s)": 0.032208
    },
    {
      "epoch": 0.7872886692825655,
      "grad_norm": 0.10197113454341888,
      "learning_rate": 3.5630637669830645e-05,
      "loss": 0.3019747734069824,
      "memory(GiB)": 78.33,
      "step": 4063,
      "token_acc": 0.9101308469795503,
      "train_speed(iter/s)": 0.032209
    },
    {
      "epoch": 0.7874824395678922,
      "grad_norm": 0.10532142966985703,
      "learning_rate": 3.556846065433279e-05,
      "loss": 0.33741429448127747,
      "memory(GiB)": 78.33,
      "step": 4064,
      "token_acc": 0.900261802212651,
      "train_speed(iter/s)": 0.03221
    },
    {
      "epoch": 0.787676209853219,
      "grad_norm": 0.10591472685337067,
      "learning_rate": 3.550633063894301e-05,
      "loss": 0.3443160951137543,
      "memory(GiB)": 78.33,
      "step": 4065,
      "token_acc": 0.8973382917044889,
      "train_speed(iter/s)": 0.032211
    },
    {
      "epoch": 0.7878699801385457,
      "grad_norm": 0.10201407968997955,
      "learning_rate": 3.544424764917983e-05,
      "loss": 0.33403322100639343,
      "memory(GiB)": 78.33,
      "step": 4066,
      "token_acc": 0.8988485568234625,
      "train_speed(iter/s)": 0.032211
    },
    {
      "epoch": 0.7880637504238724,
      "grad_norm": 0.09689588844776154,
      "learning_rate": 3.538221171054239e-05,
      "loss": 0.3143858015537262,
      "memory(GiB)": 78.33,
      "step": 4067,
      "token_acc": 0.9059065757302792,
      "train_speed(iter/s)": 0.032212
    },
    {
      "epoch": 0.7882575207091993,
      "grad_norm": 0.09144670516252518,
      "learning_rate": 3.532022284851048e-05,
      "loss": 0.3066175878047943,
      "memory(GiB)": 78.33,
      "step": 4068,
      "token_acc": 0.9066947038975011,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.788451290994526,
      "grad_norm": 0.09848953783512115,
      "learning_rate": 3.525828108854464e-05,
      "loss": 0.3331592381000519,
      "memory(GiB)": 78.33,
      "step": 4069,
      "token_acc": 0.9009377322888673,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.7886450612798528,
      "grad_norm": 0.11075198650360107,
      "learning_rate": 3.519638645608596e-05,
      "loss": 0.37963664531707764,
      "memory(GiB)": 78.33,
      "step": 4070,
      "token_acc": 0.8859161767085604,
      "train_speed(iter/s)": 0.032214
    },
    {
      "epoch": 0.7888388315651795,
      "grad_norm": 0.09600254893302917,
      "learning_rate": 3.513453897655622e-05,
      "loss": 0.30960047245025635,
      "memory(GiB)": 78.33,
      "step": 4071,
      "token_acc": 0.9076073849781613,
      "train_speed(iter/s)": 0.032215
    },
    {
      "epoch": 0.7890326018505063,
      "grad_norm": 0.10882295668125153,
      "learning_rate": 3.507273867535793e-05,
      "loss": 0.35110461711883545,
      "memory(GiB)": 78.33,
      "step": 4072,
      "token_acc": 0.8942715048811936,
      "train_speed(iter/s)": 0.032215
    },
    {
      "epoch": 0.789226372135833,
      "grad_norm": 0.10820236057043076,
      "learning_rate": 3.5010985577874066e-05,
      "loss": 0.34106817841529846,
      "memory(GiB)": 78.33,
      "step": 4073,
      "token_acc": 0.8956751190629425,
      "train_speed(iter/s)": 0.032216
    },
    {
      "epoch": 0.7894201424211598,
      "grad_norm": 0.11212671548128128,
      "learning_rate": 3.494927970946831e-05,
      "loss": 0.3203110694885254,
      "memory(GiB)": 78.33,
      "step": 4074,
      "token_acc": 0.9028558826772521,
      "train_speed(iter/s)": 0.032217
    },
    {
      "epoch": 0.7896139127064865,
      "grad_norm": 0.09849805384874344,
      "learning_rate": 3.4887621095484905e-05,
      "loss": 0.3194783627986908,
      "memory(GiB)": 78.33,
      "step": 4075,
      "token_acc": 0.9021505376344086,
      "train_speed(iter/s)": 0.032217
    },
    {
      "epoch": 0.7898076829918133,
      "grad_norm": 0.10881268978118896,
      "learning_rate": 3.482600976124871e-05,
      "loss": 0.33694878220558167,
      "memory(GiB)": 78.33,
      "step": 4076,
      "token_acc": 0.8999273783587509,
      "train_speed(iter/s)": 0.032218
    },
    {
      "epoch": 0.79000145327714,
      "grad_norm": 0.08902069926261902,
      "learning_rate": 3.476444573206515e-05,
      "loss": 0.28375521302223206,
      "memory(GiB)": 78.33,
      "step": 4077,
      "token_acc": 0.9137976536071591,
      "train_speed(iter/s)": 0.032219
    },
    {
      "epoch": 0.7901952235624667,
      "grad_norm": 0.11011825501918793,
      "learning_rate": 3.4702929033220174e-05,
      "loss": 0.36214667558670044,
      "memory(GiB)": 78.33,
      "step": 4078,
      "token_acc": 0.8909107047754306,
      "train_speed(iter/s)": 0.032219
    },
    {
      "epoch": 0.7903889938477935,
      "grad_norm": 0.09892592579126358,
      "learning_rate": 3.464145968998045e-05,
      "loss": 0.3139350414276123,
      "memory(GiB)": 78.33,
      "step": 4079,
      "token_acc": 0.904188324670132,
      "train_speed(iter/s)": 0.03222
    },
    {
      "epoch": 0.7905827641331202,
      "grad_norm": 0.09487756341695786,
      "learning_rate": 3.4580037727593033e-05,
      "loss": 0.32229354977607727,
      "memory(GiB)": 78.33,
      "step": 4080,
      "token_acc": 0.9033888688296281,
      "train_speed(iter/s)": 0.032221
    },
    {
      "epoch": 0.790776534418447,
      "grad_norm": 0.09870006889104843,
      "learning_rate": 3.4518663171285563e-05,
      "loss": 0.3076024651527405,
      "memory(GiB)": 78.33,
      "step": 4081,
      "token_acc": 0.9079937641113859,
      "train_speed(iter/s)": 0.032222
    },
    {
      "epoch": 0.7909703047037737,
      "grad_norm": 0.08845412731170654,
      "learning_rate": 3.445733604626626e-05,
      "loss": 0.3001633286476135,
      "memory(GiB)": 78.33,
      "step": 4082,
      "token_acc": 0.9098186843296562,
      "train_speed(iter/s)": 0.032222
    },
    {
      "epoch": 0.7911640749891005,
      "grad_norm": 0.10530709475278854,
      "learning_rate": 3.4396056377723766e-05,
      "loss": 0.32672393321990967,
      "memory(GiB)": 78.33,
      "step": 4083,
      "token_acc": 0.9015249886758266,
      "train_speed(iter/s)": 0.032223
    },
    {
      "epoch": 0.7913578452744272,
      "grad_norm": 0.09617147594690323,
      "learning_rate": 3.433482419082734e-05,
      "loss": 0.31130674481391907,
      "memory(GiB)": 78.33,
      "step": 4084,
      "token_acc": 0.9056026236676687,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.791551615559754,
      "grad_norm": 0.09556283056735992,
      "learning_rate": 3.4273639510726617e-05,
      "loss": 0.3225637972354889,
      "memory(GiB)": 78.33,
      "step": 4085,
      "token_acc": 0.9032013734599071,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.7917453858450807,
      "grad_norm": 0.10910790413618088,
      "learning_rate": 3.4212502362551864e-05,
      "loss": 0.3229255974292755,
      "memory(GiB)": 78.33,
      "step": 4086,
      "token_acc": 0.9025904897090135,
      "train_speed(iter/s)": 0.032225
    },
    {
      "epoch": 0.7919391561304074,
      "grad_norm": 0.10885477811098099,
      "learning_rate": 3.415141277141372e-05,
      "loss": 0.36971014738082886,
      "memory(GiB)": 78.33,
      "step": 4087,
      "token_acc": 0.8883316991175234,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.7921329264157342,
      "grad_norm": 0.10303046554327011,
      "learning_rate": 3.409037076240334e-05,
      "loss": 0.3518933653831482,
      "memory(GiB)": 78.33,
      "step": 4088,
      "token_acc": 0.8961959684672769,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.7923266967010609,
      "grad_norm": 0.09691599756479263,
      "learning_rate": 3.4029376360592284e-05,
      "loss": 0.3284243941307068,
      "memory(GiB)": 78.33,
      "step": 4089,
      "token_acc": 0.9007073741985853,
      "train_speed(iter/s)": 0.032227
    },
    {
      "epoch": 0.7925204669863877,
      "grad_norm": 0.10007129609584808,
      "learning_rate": 3.396842959103262e-05,
      "loss": 0.30757999420166016,
      "memory(GiB)": 78.33,
      "step": 4090,
      "token_acc": 0.9069416756121249,
      "train_speed(iter/s)": 0.032228
    },
    {
      "epoch": 0.7927142372717144,
      "grad_norm": 0.12611159682273865,
      "learning_rate": 3.3907530478756793e-05,
      "loss": 0.3721444308757782,
      "memory(GiB)": 78.33,
      "step": 4091,
      "token_acc": 0.8881480744352672,
      "train_speed(iter/s)": 0.032228
    },
    {
      "epoch": 0.7929080075570412,
      "grad_norm": 0.10851044952869415,
      "learning_rate": 3.38466790487777e-05,
      "loss": 0.3504268229007721,
      "memory(GiB)": 78.33,
      "step": 4092,
      "token_acc": 0.8952597994530538,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.7931017778423679,
      "grad_norm": 0.09891311824321747,
      "learning_rate": 3.378587532608872e-05,
      "loss": 0.3422066867351532,
      "memory(GiB)": 78.33,
      "step": 4093,
      "token_acc": 0.8977621295777654,
      "train_speed(iter/s)": 0.03223
    },
    {
      "epoch": 0.7932955481276946,
      "grad_norm": 0.10003877431154251,
      "learning_rate": 3.372511933566355e-05,
      "loss": 0.3378649652004242,
      "memory(GiB)": 78.33,
      "step": 4094,
      "token_acc": 0.8983621269912497,
      "train_speed(iter/s)": 0.032231
    },
    {
      "epoch": 0.7934893184130214,
      "grad_norm": 0.10703529417514801,
      "learning_rate": 3.366441110245627e-05,
      "loss": 0.34462597966194153,
      "memory(GiB)": 78.33,
      "step": 4095,
      "token_acc": 0.896659132256464,
      "train_speed(iter/s)": 0.032231
    },
    {
      "epoch": 0.7936830886983481,
      "grad_norm": 0.10170795023441315,
      "learning_rate": 3.360375065140142e-05,
      "loss": 0.33412742614746094,
      "memory(GiB)": 78.33,
      "step": 4096,
      "token_acc": 0.9000777389083237,
      "train_speed(iter/s)": 0.032232
    },
    {
      "epoch": 0.7938768589836749,
      "grad_norm": 0.09521227329969406,
      "learning_rate": 3.354313800741387e-05,
      "loss": 0.32723385095596313,
      "memory(GiB)": 78.33,
      "step": 4097,
      "token_acc": 0.9021288088982738,
      "train_speed(iter/s)": 0.032233
    },
    {
      "epoch": 0.7940706292690016,
      "grad_norm": 0.10300496965646744,
      "learning_rate": 3.3482573195388854e-05,
      "loss": 0.35852038860321045,
      "memory(GiB)": 78.33,
      "step": 4098,
      "token_acc": 0.8925953869368398,
      "train_speed(iter/s)": 0.032233
    },
    {
      "epoch": 0.7942643995543284,
      "grad_norm": 0.09719771891832352,
      "learning_rate": 3.342205624020194e-05,
      "loss": 0.3195003569126129,
      "memory(GiB)": 78.33,
      "step": 4099,
      "token_acc": 0.9037184087959075,
      "train_speed(iter/s)": 0.032234
    },
    {
      "epoch": 0.7944581698396551,
      "grad_norm": 0.09721704572439194,
      "learning_rate": 3.336158716670913e-05,
      "loss": 0.3213385343551636,
      "memory(GiB)": 78.33,
      "step": 4100,
      "token_acc": 0.9026950799122532,
      "train_speed(iter/s)": 0.032235
    },
    {
      "epoch": 0.7946519401249819,
      "grad_norm": 0.1064034178853035,
      "learning_rate": 3.330116599974666e-05,
      "loss": 0.34772008657455444,
      "memory(GiB)": 78.33,
      "step": 4101,
      "token_acc": 0.8958247453035746,
      "train_speed(iter/s)": 0.032235
    },
    {
      "epoch": 0.7948457104103086,
      "grad_norm": 0.09646876156330109,
      "learning_rate": 3.324079276413114e-05,
      "loss": 0.3242574632167816,
      "memory(GiB)": 78.33,
      "step": 4102,
      "token_acc": 0.9017802644964394,
      "train_speed(iter/s)": 0.032236
    },
    {
      "epoch": 0.7950394806956353,
      "grad_norm": 0.09698193520307541,
      "learning_rate": 3.318046748465949e-05,
      "loss": 0.3176778554916382,
      "memory(GiB)": 78.33,
      "step": 4103,
      "token_acc": 0.9066922523386903,
      "train_speed(iter/s)": 0.032237
    },
    {
      "epoch": 0.7952332509809621,
      "grad_norm": 0.09748407453298569,
      "learning_rate": 3.312019018610884e-05,
      "loss": 0.31994813680648804,
      "memory(GiB)": 78.33,
      "step": 4104,
      "token_acc": 0.9049721189591078,
      "train_speed(iter/s)": 0.032237
    },
    {
      "epoch": 0.7954270212662888,
      "grad_norm": 0.09335286915302277,
      "learning_rate": 3.305996089323681e-05,
      "loss": 0.3147365152835846,
      "memory(GiB)": 78.33,
      "step": 4105,
      "token_acc": 0.9052823315118397,
      "train_speed(iter/s)": 0.032238
    },
    {
      "epoch": 0.7956207915516156,
      "grad_norm": 0.09451750665903091,
      "learning_rate": 3.299977963078115e-05,
      "loss": 0.31313663721084595,
      "memory(GiB)": 78.33,
      "step": 4106,
      "token_acc": 0.9066659892287369,
      "train_speed(iter/s)": 0.032239
    },
    {
      "epoch": 0.7958145618369423,
      "grad_norm": 0.10167912393808365,
      "learning_rate": 3.29396464234599e-05,
      "loss": 0.3258751928806305,
      "memory(GiB)": 78.33,
      "step": 4107,
      "token_acc": 0.9005933489719884,
      "train_speed(iter/s)": 0.032239
    },
    {
      "epoch": 0.7960083321222691,
      "grad_norm": 0.11654622852802277,
      "learning_rate": 3.287956129597142e-05,
      "loss": 0.39499345421791077,
      "memory(GiB)": 78.33,
      "step": 4108,
      "token_acc": 0.883764646907701,
      "train_speed(iter/s)": 0.03224
    },
    {
      "epoch": 0.7962021024075958,
      "grad_norm": 0.09777145832777023,
      "learning_rate": 3.281952427299424e-05,
      "loss": 0.3374737501144409,
      "memory(GiB)": 78.33,
      "step": 4109,
      "token_acc": 0.8980137309365337,
      "train_speed(iter/s)": 0.032241
    },
    {
      "epoch": 0.7963958726929226,
      "grad_norm": 0.09370385110378265,
      "learning_rate": 3.2759535379187214e-05,
      "loss": 0.3268592059612274,
      "memory(GiB)": 78.33,
      "step": 4110,
      "token_acc": 0.9006880405179416,
      "train_speed(iter/s)": 0.032241
    },
    {
      "epoch": 0.7965896429782493,
      "grad_norm": 0.09882562607526779,
      "learning_rate": 3.269959463918934e-05,
      "loss": 0.32429930567741394,
      "memory(GiB)": 78.33,
      "step": 4111,
      "token_acc": 0.9020517759682456,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.796783413263576,
      "grad_norm": 0.09826599806547165,
      "learning_rate": 3.263970207761997e-05,
      "loss": 0.32509344816207886,
      "memory(GiB)": 78.33,
      "step": 4112,
      "token_acc": 0.9012852928782132,
      "train_speed(iter/s)": 0.032243
    },
    {
      "epoch": 0.7969771835489028,
      "grad_norm": 0.0987883061170578,
      "learning_rate": 3.257985771907856e-05,
      "loss": 0.3253635764122009,
      "memory(GiB)": 78.33,
      "step": 4113,
      "token_acc": 0.9022095821483264,
      "train_speed(iter/s)": 0.032243
    },
    {
      "epoch": 0.7971709538342295,
      "grad_norm": 0.10667404532432556,
      "learning_rate": 3.252006158814478e-05,
      "loss": 0.333732932806015,
      "memory(GiB)": 78.33,
      "step": 4114,
      "token_acc": 0.8992663220134353,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.7973647241195563,
      "grad_norm": 0.10525521636009216,
      "learning_rate": 3.246031370937851e-05,
      "loss": 0.3293308913707733,
      "memory(GiB)": 78.33,
      "step": 4115,
      "token_acc": 0.9034758242413402,
      "train_speed(iter/s)": 0.032245
    },
    {
      "epoch": 0.797558494404883,
      "grad_norm": 0.09787974506616592,
      "learning_rate": 3.240061410731981e-05,
      "loss": 0.31691908836364746,
      "memory(GiB)": 78.33,
      "step": 4116,
      "token_acc": 0.9044979630296816,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.7977522646902098,
      "grad_norm": 0.10533692687749863,
      "learning_rate": 3.234096280648892e-05,
      "loss": 0.35985928773880005,
      "memory(GiB)": 78.33,
      "step": 4117,
      "token_acc": 0.8940023752969121,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.7979460349755365,
      "grad_norm": 0.09776397794485092,
      "learning_rate": 3.228135983138618e-05,
      "loss": 0.320921391248703,
      "memory(GiB)": 78.33,
      "step": 4118,
      "token_acc": 0.9044611133043822,
      "train_speed(iter/s)": 0.032247
    },
    {
      "epoch": 0.7981398052608633,
      "grad_norm": 0.10431236773729324,
      "learning_rate": 3.222180520649224e-05,
      "loss": 0.3434464633464813,
      "memory(GiB)": 78.33,
      "step": 4119,
      "token_acc": 0.8964882943143813,
      "train_speed(iter/s)": 0.032248
    },
    {
      "epoch": 0.79833357554619,
      "grad_norm": 0.09289419651031494,
      "learning_rate": 3.216229895626769e-05,
      "loss": 0.2984315752983093,
      "memory(GiB)": 78.33,
      "step": 4120,
      "token_acc": 0.9101831539207184,
      "train_speed(iter/s)": 0.032248
    },
    {
      "epoch": 0.7985273458315167,
      "grad_norm": 0.09537661075592041,
      "learning_rate": 3.2102841105153414e-05,
      "loss": 0.3033888041973114,
      "memory(GiB)": 78.33,
      "step": 4121,
      "token_acc": 0.9094319549819424,
      "train_speed(iter/s)": 0.032249
    },
    {
      "epoch": 0.7987211161168435,
      "grad_norm": 0.09890120476484299,
      "learning_rate": 3.2043431677570295e-05,
      "loss": 0.33709731698036194,
      "memory(GiB)": 78.33,
      "step": 4122,
      "token_acc": 0.899075500770416,
      "train_speed(iter/s)": 0.03225
    },
    {
      "epoch": 0.7989148864021702,
      "grad_norm": 0.10694926232099533,
      "learning_rate": 3.19840706979194e-05,
      "loss": 0.3586742877960205,
      "memory(GiB)": 78.33,
      "step": 4123,
      "token_acc": 0.8929906273359783,
      "train_speed(iter/s)": 0.03225
    },
    {
      "epoch": 0.799108656687497,
      "grad_norm": 0.09874139726161957,
      "learning_rate": 3.1924758190581886e-05,
      "loss": 0.3332917392253876,
      "memory(GiB)": 78.33,
      "step": 4124,
      "token_acc": 0.9008399249775748,
      "train_speed(iter/s)": 0.032251
    },
    {
      "epoch": 0.7993024269728237,
      "grad_norm": 0.10014494508504868,
      "learning_rate": 3.186549417991895e-05,
      "loss": 0.3542085289955139,
      "memory(GiB)": 78.33,
      "step": 4125,
      "token_acc": 0.892789227527368,
      "train_speed(iter/s)": 0.032252
    },
    {
      "epoch": 0.7994961972581505,
      "grad_norm": 0.10449165850877762,
      "learning_rate": 3.1806278690272005e-05,
      "loss": 0.3253108263015747,
      "memory(GiB)": 78.33,
      "step": 4126,
      "token_acc": 0.9014500959391023,
      "train_speed(iter/s)": 0.032253
    },
    {
      "epoch": 0.7996899675434772,
      "grad_norm": 0.1035015657544136,
      "learning_rate": 3.174711174596238e-05,
      "loss": 0.3380778133869171,
      "memory(GiB)": 78.33,
      "step": 4127,
      "token_acc": 0.9001489459211732,
      "train_speed(iter/s)": 0.032253
    },
    {
      "epoch": 0.799883737828804,
      "grad_norm": 0.10077136009931564,
      "learning_rate": 3.1687993371291525e-05,
      "loss": 0.32532066106796265,
      "memory(GiB)": 78.33,
      "step": 4128,
      "token_acc": 0.9027537506026488,
      "train_speed(iter/s)": 0.032254
    },
    {
      "epoch": 0.8000775081141307,
      "grad_norm": 0.10059913992881775,
      "learning_rate": 3.162892359054098e-05,
      "loss": 0.3570432662963867,
      "memory(GiB)": 78.33,
      "step": 4129,
      "token_acc": 0.8933640100302458,
      "train_speed(iter/s)": 0.032255
    },
    {
      "epoch": 0.8002712783994574,
      "grad_norm": 0.1088043823838234,
      "learning_rate": 3.156990242797226e-05,
      "loss": 0.3387261927127838,
      "memory(GiB)": 78.33,
      "step": 4130,
      "token_acc": 0.8997582550409318,
      "train_speed(iter/s)": 0.032255
    },
    {
      "epoch": 0.8004650486847842,
      "grad_norm": 0.10076975077390671,
      "learning_rate": 3.151092990782695e-05,
      "loss": 0.32494884729385376,
      "memory(GiB)": 78.33,
      "step": 4131,
      "token_acc": 0.9004781745476916,
      "train_speed(iter/s)": 0.032256
    },
    {
      "epoch": 0.8006588189701109,
      "grad_norm": 0.1097509041428566,
      "learning_rate": 3.145200605432662e-05,
      "loss": 0.3796813189983368,
      "memory(GiB)": 78.33,
      "step": 4132,
      "token_acc": 0.8863678979936246,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.8008525892554377,
      "grad_norm": 0.10167836397886276,
      "learning_rate": 3.1393130891672944e-05,
      "loss": 0.3483812212944031,
      "memory(GiB)": 78.33,
      "step": 4133,
      "token_acc": 0.896118628960376,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.8010463595407644,
      "grad_norm": 0.10103829205036163,
      "learning_rate": 3.1334304444047495e-05,
      "loss": 0.3420543372631073,
      "memory(GiB)": 78.33,
      "step": 4134,
      "token_acc": 0.8974770039421813,
      "train_speed(iter/s)": 0.032258
    },
    {
      "epoch": 0.8012401298260912,
      "grad_norm": 0.10454893857240677,
      "learning_rate": 3.1275526735611896e-05,
      "loss": 0.31737667322158813,
      "memory(GiB)": 78.33,
      "step": 4135,
      "token_acc": 0.9039817974971559,
      "train_speed(iter/s)": 0.032259
    },
    {
      "epoch": 0.8014339001114179,
      "grad_norm": 0.14402469992637634,
      "learning_rate": 3.12167977905077e-05,
      "loss": 0.36191561818122864,
      "memory(GiB)": 78.33,
      "step": 4136,
      "token_acc": 0.8928331193053263,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 0.8016276703967447,
      "grad_norm": 0.09334685653448105,
      "learning_rate": 3.1158117632856454e-05,
      "loss": 0.3147181570529938,
      "memory(GiB)": 78.33,
      "step": 4137,
      "token_acc": 0.905297142173861,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 0.8018214406820714,
      "grad_norm": 0.09789223968982697,
      "learning_rate": 3.109948628675974e-05,
      "loss": 0.3045772612094879,
      "memory(GiB)": 78.33,
      "step": 4138,
      "token_acc": 0.9074807619528195,
      "train_speed(iter/s)": 0.032261
    },
    {
      "epoch": 0.8020152109673981,
      "grad_norm": 0.11379068344831467,
      "learning_rate": 3.104090377629899e-05,
      "loss": 0.3640574514865875,
      "memory(GiB)": 78.33,
      "step": 4139,
      "token_acc": 0.8914368285480933,
      "train_speed(iter/s)": 0.032262
    },
    {
      "epoch": 0.8022089812527249,
      "grad_norm": 0.11807750910520554,
      "learning_rate": 3.098237012553562e-05,
      "loss": 0.35527053475379944,
      "memory(GiB)": 78.33,
      "step": 4140,
      "token_acc": 0.8920514040932889,
      "train_speed(iter/s)": 0.032262
    },
    {
      "epoch": 0.8024027515380516,
      "grad_norm": 0.09791693091392517,
      "learning_rate": 3.0923885358510946e-05,
      "loss": 0.3340778946876526,
      "memory(GiB)": 78.33,
      "step": 4141,
      "token_acc": 0.9000325538494927,
      "train_speed(iter/s)": 0.032263
    },
    {
      "epoch": 0.8025965218233784,
      "grad_norm": 0.10278962552547455,
      "learning_rate": 3.086544949924627e-05,
      "loss": 0.34071099758148193,
      "memory(GiB)": 78.33,
      "step": 4142,
      "token_acc": 0.9005618602581219,
      "train_speed(iter/s)": 0.032264
    },
    {
      "epoch": 0.8027902921087051,
      "grad_norm": 0.100987508893013,
      "learning_rate": 3.0807062571742755e-05,
      "loss": 0.32279205322265625,
      "memory(GiB)": 78.33,
      "step": 4143,
      "token_acc": 0.9030330758672659,
      "train_speed(iter/s)": 0.032264
    },
    {
      "epoch": 0.8029840623940319,
      "grad_norm": 0.10697422176599503,
      "learning_rate": 3.074872459998143e-05,
      "loss": 0.3390357792377472,
      "memory(GiB)": 78.33,
      "step": 4144,
      "token_acc": 0.9002963590177815,
      "train_speed(iter/s)": 0.032265
    },
    {
      "epoch": 0.8031778326793586,
      "grad_norm": 0.09046490490436554,
      "learning_rate": 3.069043560792336e-05,
      "loss": 0.3146419823169708,
      "memory(GiB)": 78.33,
      "step": 4145,
      "token_acc": 0.9062027231467473,
      "train_speed(iter/s)": 0.032266
    },
    {
      "epoch": 0.8033716029646853,
      "grad_norm": 0.09514185786247253,
      "learning_rate": 3.063219561950936e-05,
      "loss": 0.3226296007633209,
      "memory(GiB)": 78.33,
      "step": 4146,
      "token_acc": 0.9025114098311594,
      "train_speed(iter/s)": 0.032266
    },
    {
      "epoch": 0.8035653732500121,
      "grad_norm": 0.10075315088033676,
      "learning_rate": 3.057400465866016e-05,
      "loss": 0.31182047724723816,
      "memory(GiB)": 78.33,
      "step": 4147,
      "token_acc": 0.9075695604585517,
      "train_speed(iter/s)": 0.032267
    },
    {
      "epoch": 0.8037591435353388,
      "grad_norm": 0.10024786740541458,
      "learning_rate": 3.0515862749276353e-05,
      "loss": 0.3048425018787384,
      "memory(GiB)": 78.33,
      "step": 4148,
      "token_acc": 0.9080320590439946,
      "train_speed(iter/s)": 0.032268
    },
    {
      "epoch": 0.8039529138206656,
      "grad_norm": 0.09824330359697342,
      "learning_rate": 3.0457769915238368e-05,
      "loss": 0.3147960603237152,
      "memory(GiB)": 78.33,
      "step": 4149,
      "token_acc": 0.9054583112323508,
      "train_speed(iter/s)": 0.032268
    },
    {
      "epoch": 0.8041466841059923,
      "grad_norm": 0.10100057721138,
      "learning_rate": 3.03997261804065e-05,
      "loss": 0.35047298669815063,
      "memory(GiB)": 78.33,
      "step": 4150,
      "token_acc": 0.8985602958658038,
      "train_speed(iter/s)": 0.032269
    },
    {
      "epoch": 0.8043404543913191,
      "grad_norm": 0.10795464366674423,
      "learning_rate": 3.034173156862084e-05,
      "loss": 0.3506011664867401,
      "memory(GiB)": 78.33,
      "step": 4151,
      "token_acc": 0.8946405657028967,
      "train_speed(iter/s)": 0.03227
    },
    {
      "epoch": 0.8045342246766458,
      "grad_norm": 0.09722840040922165,
      "learning_rate": 3.028378610370141e-05,
      "loss": 0.30743473768234253,
      "memory(GiB)": 78.33,
      "step": 4152,
      "token_acc": 0.9074273940345369,
      "train_speed(iter/s)": 0.03227
    },
    {
      "epoch": 0.8047279949619726,
      "grad_norm": 0.10855165868997574,
      "learning_rate": 3.022588980944792e-05,
      "loss": 0.37705856561660767,
      "memory(GiB)": 78.33,
      "step": 4153,
      "token_acc": 0.889402452187872,
      "train_speed(iter/s)": 0.032271
    },
    {
      "epoch": 0.8049217652472993,
      "grad_norm": 0.09210513532161713,
      "learning_rate": 3.0168042709639932e-05,
      "loss": 0.30485376715660095,
      "memory(GiB)": 78.33,
      "step": 4154,
      "token_acc": 0.9075583735909822,
      "train_speed(iter/s)": 0.032271
    },
    {
      "epoch": 0.805115535532626,
      "grad_norm": 0.10625799745321274,
      "learning_rate": 3.011024482803684e-05,
      "loss": 0.3092432916164398,
      "memory(GiB)": 78.33,
      "step": 4155,
      "token_acc": 0.9070884146341464,
      "train_speed(iter/s)": 0.032272
    },
    {
      "epoch": 0.8053093058179528,
      "grad_norm": 0.10007067769765854,
      "learning_rate": 3.0052496188377735e-05,
      "loss": 0.3402232229709625,
      "memory(GiB)": 78.33,
      "step": 4156,
      "token_acc": 0.8997760214195608,
      "train_speed(iter/s)": 0.032273
    },
    {
      "epoch": 0.8055030761032795,
      "grad_norm": 0.10713425278663635,
      "learning_rate": 2.999479681438156e-05,
      "loss": 0.3407253921031952,
      "memory(GiB)": 78.33,
      "step": 4157,
      "token_acc": 0.8987352889513438,
      "train_speed(iter/s)": 0.032273
    },
    {
      "epoch": 0.8056968463886063,
      "grad_norm": 0.09479415416717529,
      "learning_rate": 2.993714672974698e-05,
      "loss": 0.3190673887729645,
      "memory(GiB)": 78.33,
      "step": 4158,
      "token_acc": 0.9022564050323435,
      "train_speed(iter/s)": 0.032274
    },
    {
      "epoch": 0.805890616673933,
      "grad_norm": 0.10293829441070557,
      "learning_rate": 2.987954595815247e-05,
      "loss": 0.3430511951446533,
      "memory(GiB)": 78.33,
      "step": 4159,
      "token_acc": 0.8964779192630723,
      "train_speed(iter/s)": 0.032275
    },
    {
      "epoch": 0.8060843869592598,
      "grad_norm": 0.09652558714151382,
      "learning_rate": 2.98219945232562e-05,
      "loss": 0.3083397150039673,
      "memory(GiB)": 78.33,
      "step": 4160,
      "token_acc": 0.9067033176387339,
      "train_speed(iter/s)": 0.032275
    },
    {
      "epoch": 0.8062781572445865,
      "grad_norm": 0.11536381393671036,
      "learning_rate": 2.9764492448696098e-05,
      "loss": 0.3753798305988312,
      "memory(GiB)": 78.33,
      "step": 4161,
      "token_acc": 0.8890807651434643,
      "train_speed(iter/s)": 0.032276
    },
    {
      "epoch": 0.8064719275299133,
      "grad_norm": 0.09727875888347626,
      "learning_rate": 2.970703975808979e-05,
      "loss": 0.32429182529449463,
      "memory(GiB)": 78.33,
      "step": 4162,
      "token_acc": 0.9036051879533964,
      "train_speed(iter/s)": 0.032277
    },
    {
      "epoch": 0.80666569781524,
      "grad_norm": 0.10008195042610168,
      "learning_rate": 2.964963647503465e-05,
      "loss": 0.326107919216156,
      "memory(GiB)": 78.33,
      "step": 4163,
      "token_acc": 0.9037652447142781,
      "train_speed(iter/s)": 0.032278
    },
    {
      "epoch": 0.8068594681005667,
      "grad_norm": 0.0974980965256691,
      "learning_rate": 2.9592282623107765e-05,
      "loss": 0.31634485721588135,
      "memory(GiB)": 78.33,
      "step": 4164,
      "token_acc": 0.9050801412659604,
      "train_speed(iter/s)": 0.032278
    },
    {
      "epoch": 0.8070532383858935,
      "grad_norm": 0.10140351206064224,
      "learning_rate": 2.953497822586583e-05,
      "loss": 0.32958346605300903,
      "memory(GiB)": 78.33,
      "step": 4165,
      "token_acc": 0.900670556920557,
      "train_speed(iter/s)": 0.032279
    },
    {
      "epoch": 0.8072470086712202,
      "grad_norm": 0.10700841248035431,
      "learning_rate": 2.9477723306845414e-05,
      "loss": 0.34634923934936523,
      "memory(GiB)": 78.33,
      "step": 4166,
      "token_acc": 0.8981357595218982,
      "train_speed(iter/s)": 0.032279
    },
    {
      "epoch": 0.807440778956547,
      "grad_norm": 0.09883598238229752,
      "learning_rate": 2.9420517889562574e-05,
      "loss": 0.3215301036834717,
      "memory(GiB)": 78.33,
      "step": 4167,
      "token_acc": 0.9049081311541975,
      "train_speed(iter/s)": 0.03228
    },
    {
      "epoch": 0.8076345492418737,
      "grad_norm": 0.10146182030439377,
      "learning_rate": 2.9363361997513145e-05,
      "loss": 0.3307911157608032,
      "memory(GiB)": 78.33,
      "step": 4168,
      "token_acc": 0.9003170028818444,
      "train_speed(iter/s)": 0.032281
    },
    {
      "epoch": 0.8078283195272005,
      "grad_norm": 0.10559407621622086,
      "learning_rate": 2.9306255654172572e-05,
      "loss": 0.33857250213623047,
      "memory(GiB)": 78.33,
      "step": 4169,
      "token_acc": 0.8989095106466997,
      "train_speed(iter/s)": 0.032281
    },
    {
      "epoch": 0.8080220898125272,
      "grad_norm": 0.10372511297464371,
      "learning_rate": 2.9249198882995973e-05,
      "loss": 0.31755688786506653,
      "memory(GiB)": 78.33,
      "step": 4170,
      "token_acc": 0.9028887891034276,
      "train_speed(iter/s)": 0.032282
    },
    {
      "epoch": 0.808215860097854,
      "grad_norm": 0.10989030450582504,
      "learning_rate": 2.91921917074181e-05,
      "loss": 0.3677654266357422,
      "memory(GiB)": 78.33,
      "step": 4171,
      "token_acc": 0.8904222209172588,
      "train_speed(iter/s)": 0.032283
    },
    {
      "epoch": 0.8084096303831807,
      "grad_norm": 0.10606292635202408,
      "learning_rate": 2.9135234150853276e-05,
      "loss": 0.35702183842658997,
      "memory(GiB)": 78.33,
      "step": 4172,
      "token_acc": 0.8924521259797307,
      "train_speed(iter/s)": 0.032283
    },
    {
      "epoch": 0.8086034006685074,
      "grad_norm": 0.09554623067378998,
      "learning_rate": 2.907832623669559e-05,
      "loss": 0.32034456729888916,
      "memory(GiB)": 78.33,
      "step": 4173,
      "token_acc": 0.9017811962351988,
      "train_speed(iter/s)": 0.032284
    },
    {
      "epoch": 0.8087971709538342,
      "grad_norm": 0.09658126533031464,
      "learning_rate": 2.90214679883186e-05,
      "loss": 0.3179253339767456,
      "memory(GiB)": 78.33,
      "step": 4174,
      "token_acc": 0.9031473726867486,
      "train_speed(iter/s)": 0.032285
    },
    {
      "epoch": 0.8089909412391609,
      "grad_norm": 0.09535779058933258,
      "learning_rate": 2.8964659429075543e-05,
      "loss": 0.3256949186325073,
      "memory(GiB)": 78.33,
      "step": 4175,
      "token_acc": 0.9016154247003648,
      "train_speed(iter/s)": 0.032285
    },
    {
      "epoch": 0.8091847115244877,
      "grad_norm": 0.09355632960796356,
      "learning_rate": 2.890790058229919e-05,
      "loss": 0.3242107629776001,
      "memory(GiB)": 78.33,
      "step": 4176,
      "token_acc": 0.9035344140313579,
      "train_speed(iter/s)": 0.032286
    },
    {
      "epoch": 0.8093784818098144,
      "grad_norm": 0.10124616324901581,
      "learning_rate": 2.8851191471301903e-05,
      "loss": 0.31790977716445923,
      "memory(GiB)": 78.33,
      "step": 4177,
      "token_acc": 0.9053812224322622,
      "train_speed(iter/s)": 0.032287
    },
    {
      "epoch": 0.8095722520951412,
      "grad_norm": 0.10209498554468155,
      "learning_rate": 2.8794532119375712e-05,
      "loss": 0.3283519148826599,
      "memory(GiB)": 78.33,
      "step": 4178,
      "token_acc": 0.9008973858759266,
      "train_speed(iter/s)": 0.032287
    },
    {
      "epoch": 0.8097660223804679,
      "grad_norm": 0.11376667767763138,
      "learning_rate": 2.8737922549792103e-05,
      "loss": 0.3559627830982208,
      "memory(GiB)": 78.33,
      "step": 4179,
      "token_acc": 0.8943670846197467,
      "train_speed(iter/s)": 0.032288
    },
    {
      "epoch": 0.8099597926657947,
      "grad_norm": 0.0886598601937294,
      "learning_rate": 2.868136278580214e-05,
      "loss": 0.3123997449874878,
      "memory(GiB)": 78.33,
      "step": 4180,
      "token_acc": 0.9062552047395845,
      "train_speed(iter/s)": 0.032289
    },
    {
      "epoch": 0.8101535629511214,
      "grad_norm": 0.09667520970106125,
      "learning_rate": 2.8624852850636432e-05,
      "loss": 0.3109186589717865,
      "memory(GiB)": 78.33,
      "step": 4181,
      "token_acc": 0.9074359598582596,
      "train_speed(iter/s)": 0.032289
    },
    {
      "epoch": 0.8103473332364481,
      "grad_norm": 0.09185932576656342,
      "learning_rate": 2.856839276750514e-05,
      "loss": 0.3074433505535126,
      "memory(GiB)": 78.33,
      "step": 4182,
      "token_acc": 0.9070412315028378,
      "train_speed(iter/s)": 0.03229
    },
    {
      "epoch": 0.8105411035217749,
      "grad_norm": 0.09254490584135056,
      "learning_rate": 2.851198255959793e-05,
      "loss": 0.3106043040752411,
      "memory(GiB)": 78.33,
      "step": 4183,
      "token_acc": 0.9067268041237113,
      "train_speed(iter/s)": 0.032291
    },
    {
      "epoch": 0.8107348738071016,
      "grad_norm": 0.10701245069503784,
      "learning_rate": 2.8455622250083953e-05,
      "loss": 0.32825660705566406,
      "memory(GiB)": 78.33,
      "step": 4184,
      "token_acc": 0.9028698224852071,
      "train_speed(iter/s)": 0.032291
    },
    {
      "epoch": 0.8109286440924284,
      "grad_norm": 0.09500328451395035,
      "learning_rate": 2.8399311862111978e-05,
      "loss": 0.28734296560287476,
      "memory(GiB)": 78.33,
      "step": 4185,
      "token_acc": 0.9113909591935806,
      "train_speed(iter/s)": 0.032292
    },
    {
      "epoch": 0.8111224143777551,
      "grad_norm": 0.11596144735813141,
      "learning_rate": 2.834305141881017e-05,
      "loss": 0.3888537287712097,
      "memory(GiB)": 78.33,
      "step": 4186,
      "token_acc": 0.8838215903227655,
      "train_speed(iter/s)": 0.032293
    },
    {
      "epoch": 0.8113161846630819,
      "grad_norm": 0.10639077425003052,
      "learning_rate": 2.8286840943286178e-05,
      "loss": 0.34377238154411316,
      "memory(GiB)": 78.33,
      "step": 4187,
      "token_acc": 0.9001137980085349,
      "train_speed(iter/s)": 0.032293
    },
    {
      "epoch": 0.8115099549484086,
      "grad_norm": 0.10659915953874588,
      "learning_rate": 2.823068045862718e-05,
      "loss": 0.36640581488609314,
      "memory(GiB)": 78.33,
      "step": 4188,
      "token_acc": 0.8934043246767358,
      "train_speed(iter/s)": 0.032294
    },
    {
      "epoch": 0.8117037252337354,
      "grad_norm": 0.10051631182432175,
      "learning_rate": 2.817456998789978e-05,
      "loss": 0.3765765428543091,
      "memory(GiB)": 78.33,
      "step": 4189,
      "token_acc": 0.8884732824427481,
      "train_speed(iter/s)": 0.032294
    },
    {
      "epoch": 0.8118974955190622,
      "grad_norm": 0.10304338485002518,
      "learning_rate": 2.8118509554150076e-05,
      "loss": 0.34053856134414673,
      "memory(GiB)": 78.33,
      "step": 4190,
      "token_acc": 0.8995534483709136,
      "train_speed(iter/s)": 0.032295
    },
    {
      "epoch": 0.812091265804389,
      "grad_norm": 0.09566718339920044,
      "learning_rate": 2.8062499180403532e-05,
      "loss": 0.3288200795650482,
      "memory(GiB)": 78.33,
      "step": 4191,
      "token_acc": 0.900259556989139,
      "train_speed(iter/s)": 0.032296
    },
    {
      "epoch": 0.8122850360897157,
      "grad_norm": 0.10790709406137466,
      "learning_rate": 2.800653888966519e-05,
      "loss": 0.32764533162117004,
      "memory(GiB)": 78.33,
      "step": 4192,
      "token_acc": 0.9018782286721013,
      "train_speed(iter/s)": 0.032296
    },
    {
      "epoch": 0.8124788063750424,
      "grad_norm": 0.09332863241434097,
      "learning_rate": 2.7950628704919426e-05,
      "loss": 0.3051561415195465,
      "memory(GiB)": 78.33,
      "step": 4193,
      "token_acc": 0.9071980963712076,
      "train_speed(iter/s)": 0.032297
    },
    {
      "epoch": 0.8126725766603692,
      "grad_norm": 0.10486903041601181,
      "learning_rate": 2.7894768649130044e-05,
      "loss": 0.358365923166275,
      "memory(GiB)": 78.33,
      "step": 4194,
      "token_acc": 0.8943505007578234,
      "train_speed(iter/s)": 0.032298
    },
    {
      "epoch": 0.8128663469456959,
      "grad_norm": 0.08717334270477295,
      "learning_rate": 2.783895874524028e-05,
      "loss": 0.30397284030914307,
      "memory(GiB)": 78.33,
      "step": 4195,
      "token_acc": 0.9074197783971805,
      "train_speed(iter/s)": 0.032298
    },
    {
      "epoch": 0.8130601172310227,
      "grad_norm": 0.09161286801099777,
      "learning_rate": 2.7783199016172765e-05,
      "loss": 0.3110318183898926,
      "memory(GiB)": 78.33,
      "step": 4196,
      "token_acc": 0.907268415519644,
      "train_speed(iter/s)": 0.032299
    },
    {
      "epoch": 0.8132538875163494,
      "grad_norm": 0.10017764568328857,
      "learning_rate": 2.772748948482949e-05,
      "loss": 0.33178234100341797,
      "memory(GiB)": 78.33,
      "step": 4197,
      "token_acc": 0.8992754418706773,
      "train_speed(iter/s)": 0.0323
    },
    {
      "epoch": 0.8134476578016762,
      "grad_norm": 0.09984217584133148,
      "learning_rate": 2.7671830174091824e-05,
      "loss": 0.33245033025741577,
      "memory(GiB)": 78.33,
      "step": 4198,
      "token_acc": 0.9006675354378872,
      "train_speed(iter/s)": 0.0323
    },
    {
      "epoch": 0.8136414280870029,
      "grad_norm": 0.09028760343790054,
      "learning_rate": 2.7616221106820645e-05,
      "loss": 0.29679808020591736,
      "memory(GiB)": 78.33,
      "step": 4199,
      "token_acc": 0.908898198152728,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.8138351983723296,
      "grad_norm": 0.10453902184963226,
      "learning_rate": 2.7560662305856036e-05,
      "loss": 0.3445345163345337,
      "memory(GiB)": 78.33,
      "step": 4200,
      "token_acc": 0.8972627996028932,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.8140289686576564,
      "grad_norm": 0.09469582140445709,
      "learning_rate": 2.7505153794017487e-05,
      "loss": 0.3133549690246582,
      "memory(GiB)": 78.33,
      "step": 4201,
      "token_acc": 0.905493996966347,
      "train_speed(iter/s)": 0.032299
    },
    {
      "epoch": 0.8142227389429831,
      "grad_norm": 0.10252521187067032,
      "learning_rate": 2.744969559410385e-05,
      "loss": 0.32331135869026184,
      "memory(GiB)": 78.33,
      "step": 4202,
      "token_acc": 0.9024316857682103,
      "train_speed(iter/s)": 0.0323
    },
    {
      "epoch": 0.8144165092283099,
      "grad_norm": 0.09786242991685867,
      "learning_rate": 2.7394287728893265e-05,
      "loss": 0.3233289122581482,
      "memory(GiB)": 78.33,
      "step": 4203,
      "token_acc": 0.9021999564365062,
      "train_speed(iter/s)": 0.0323
    },
    {
      "epoch": 0.8146102795136366,
      "grad_norm": 0.09682147949934006,
      "learning_rate": 2.733893022114327e-05,
      "loss": 0.3290054202079773,
      "memory(GiB)": 78.33,
      "step": 4204,
      "token_acc": 0.9020085944667086,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.8148040497989634,
      "grad_norm": 0.09070180356502533,
      "learning_rate": 2.728362309359062e-05,
      "loss": 0.3250106871128082,
      "memory(GiB)": 78.33,
      "step": 4205,
      "token_acc": 0.9020958593997225,
      "train_speed(iter/s)": 0.032302
    },
    {
      "epoch": 0.8149978200842901,
      "grad_norm": 0.10017143189907074,
      "learning_rate": 2.7228366368951525e-05,
      "loss": 0.36836662888526917,
      "memory(GiB)": 78.33,
      "step": 4206,
      "token_acc": 0.8892802334649929,
      "train_speed(iter/s)": 0.032302
    },
    {
      "epoch": 0.8151915903696169,
      "grad_norm": 0.10824721306562424,
      "learning_rate": 2.7173160069921357e-05,
      "loss": 0.31380200386047363,
      "memory(GiB)": 78.33,
      "step": 4207,
      "token_acc": 0.9043143002803976,
      "train_speed(iter/s)": 0.032303
    },
    {
      "epoch": 0.8153853606549436,
      "grad_norm": 0.09697490185499191,
      "learning_rate": 2.7118004219174838e-05,
      "loss": 0.3323417901992798,
      "memory(GiB)": 78.33,
      "step": 4208,
      "token_acc": 0.8995884773662551,
      "train_speed(iter/s)": 0.032304
    },
    {
      "epoch": 0.8155791309402703,
      "grad_norm": 0.0982755497097969,
      "learning_rate": 2.706289883936595e-05,
      "loss": 0.3338177800178528,
      "memory(GiB)": 78.33,
      "step": 4209,
      "token_acc": 0.8988449691991787,
      "train_speed(iter/s)": 0.032304
    },
    {
      "epoch": 0.8157729012255971,
      "grad_norm": 0.09161195158958435,
      "learning_rate": 2.7007843953127917e-05,
      "loss": 0.3077443540096283,
      "memory(GiB)": 78.33,
      "step": 4210,
      "token_acc": 0.9083794426997657,
      "train_speed(iter/s)": 0.032305
    },
    {
      "epoch": 0.8159666715109238,
      "grad_norm": 0.10864771902561188,
      "learning_rate": 2.6952839583073355e-05,
      "loss": 0.36505550146102905,
      "memory(GiB)": 78.33,
      "step": 4211,
      "token_acc": 0.8941932795618918,
      "train_speed(iter/s)": 0.032306
    },
    {
      "epoch": 0.8161604417962506,
      "grad_norm": 0.10426463186740875,
      "learning_rate": 2.6897885751793956e-05,
      "loss": 0.3455352485179901,
      "memory(GiB)": 78.33,
      "step": 4212,
      "token_acc": 0.8963470566017374,
      "train_speed(iter/s)": 0.032306
    },
    {
      "epoch": 0.8163542120815773,
      "grad_norm": 0.1076427772641182,
      "learning_rate": 2.6842982481860768e-05,
      "loss": 0.3745838403701782,
      "memory(GiB)": 78.33,
      "step": 4213,
      "token_acc": 0.8894979479395064,
      "train_speed(iter/s)": 0.032307
    },
    {
      "epoch": 0.8165479823669041,
      "grad_norm": 0.09672832489013672,
      "learning_rate": 2.6788129795824054e-05,
      "loss": 0.34184157848358154,
      "memory(GiB)": 78.33,
      "step": 4214,
      "token_acc": 0.8963274582336017,
      "train_speed(iter/s)": 0.032308
    },
    {
      "epoch": 0.8167417526522308,
      "grad_norm": 0.10097566246986389,
      "learning_rate": 2.6733327716213236e-05,
      "loss": 0.30746835470199585,
      "memory(GiB)": 78.33,
      "step": 4215,
      "token_acc": 0.9063103281853282,
      "train_speed(iter/s)": 0.032308
    },
    {
      "epoch": 0.8169355229375576,
      "grad_norm": 0.10790715366601944,
      "learning_rate": 2.667857626553705e-05,
      "loss": 0.34957319498062134,
      "memory(GiB)": 78.33,
      "step": 4216,
      "token_acc": 0.8963655436258176,
      "train_speed(iter/s)": 0.032309
    },
    {
      "epoch": 0.8171292932228843,
      "grad_norm": 0.10515126585960388,
      "learning_rate": 2.662387546628332e-05,
      "loss": 0.3403055965900421,
      "memory(GiB)": 78.33,
      "step": 4217,
      "token_acc": 0.8979972597148937,
      "train_speed(iter/s)": 0.032309
    },
    {
      "epoch": 0.817323063508211,
      "grad_norm": 0.10571117699146271,
      "learning_rate": 2.6569225340919202e-05,
      "loss": 0.3421645164489746,
      "memory(GiB)": 78.33,
      "step": 4218,
      "token_acc": 0.8976865691740056,
      "train_speed(iter/s)": 0.03231
    },
    {
      "epoch": 0.8175168337935378,
      "grad_norm": 0.10033620893955231,
      "learning_rate": 2.651462591189097e-05,
      "loss": 0.32853877544403076,
      "memory(GiB)": 78.33,
      "step": 4219,
      "token_acc": 0.9037694765715578,
      "train_speed(iter/s)": 0.032311
    },
    {
      "epoch": 0.8177106040788645,
      "grad_norm": 0.10828068852424622,
      "learning_rate": 2.6460077201624058e-05,
      "loss": 0.3596659302711487,
      "memory(GiB)": 78.33,
      "step": 4220,
      "token_acc": 0.8916654015485046,
      "train_speed(iter/s)": 0.032311
    },
    {
      "epoch": 0.8179043743641913,
      "grad_norm": 0.09712978452444077,
      "learning_rate": 2.6405579232523066e-05,
      "loss": 0.31051602959632874,
      "memory(GiB)": 78.33,
      "step": 4221,
      "token_acc": 0.9060798808091227,
      "train_speed(iter/s)": 0.032312
    },
    {
      "epoch": 0.818098144649518,
      "grad_norm": 0.10436290502548218,
      "learning_rate": 2.6351132026971823e-05,
      "loss": 0.3625693619251251,
      "memory(GiB)": 78.33,
      "step": 4222,
      "token_acc": 0.8920728157908663,
      "train_speed(iter/s)": 0.032313
    },
    {
      "epoch": 0.8182919149348448,
      "grad_norm": 0.10246887058019638,
      "learning_rate": 2.6296735607333202e-05,
      "loss": 0.34993377327919006,
      "memory(GiB)": 78.33,
      "step": 4223,
      "token_acc": 0.8966694249069094,
      "train_speed(iter/s)": 0.032313
    },
    {
      "epoch": 0.8184856852201715,
      "grad_norm": 0.09719887375831604,
      "learning_rate": 2.6242389995949286e-05,
      "loss": 0.31395548582077026,
      "memory(GiB)": 78.33,
      "step": 4224,
      "token_acc": 0.9058810845776508,
      "train_speed(iter/s)": 0.032314
    },
    {
      "epoch": 0.8186794555054983,
      "grad_norm": 0.09809587150812149,
      "learning_rate": 2.618809521514132e-05,
      "loss": 0.3162482976913452,
      "memory(GiB)": 78.33,
      "step": 4225,
      "token_acc": 0.9043280182232346,
      "train_speed(iter/s)": 0.032315
    },
    {
      "epoch": 0.818873225790825,
      "grad_norm": 0.09690183401107788,
      "learning_rate": 2.613385128720961e-05,
      "loss": 0.308881551027298,
      "memory(GiB)": 78.33,
      "step": 4226,
      "token_acc": 0.9085266774992675,
      "train_speed(iter/s)": 0.032315
    },
    {
      "epoch": 0.8190669960761517,
      "grad_norm": 0.10261222720146179,
      "learning_rate": 2.6079658234433575e-05,
      "loss": 0.3542589843273163,
      "memory(GiB)": 78.33,
      "step": 4227,
      "token_acc": 0.8961422008770098,
      "train_speed(iter/s)": 0.032316
    },
    {
      "epoch": 0.8192607663614785,
      "grad_norm": 0.09491068869829178,
      "learning_rate": 2.602551607907179e-05,
      "loss": 0.31204402446746826,
      "memory(GiB)": 78.33,
      "step": 4228,
      "token_acc": 0.9075091359188966,
      "train_speed(iter/s)": 0.032316
    },
    {
      "epoch": 0.8194545366468052,
      "grad_norm": 0.10251081734895706,
      "learning_rate": 2.5971424843361865e-05,
      "loss": 0.3522550165653229,
      "memory(GiB)": 78.33,
      "step": 4229,
      "token_acc": 0.8963797611666712,
      "train_speed(iter/s)": 0.032317
    },
    {
      "epoch": 0.819648306932132,
      "grad_norm": 0.11336258798837662,
      "learning_rate": 2.591738454952055e-05,
      "loss": 0.33237776160240173,
      "memory(GiB)": 78.33,
      "step": 4230,
      "token_acc": 0.8990595976358925,
      "train_speed(iter/s)": 0.032318
    },
    {
      "epoch": 0.8198420772174587,
      "grad_norm": 0.0893642008304596,
      "learning_rate": 2.5863395219743565e-05,
      "loss": 0.3061623275279999,
      "memory(GiB)": 78.33,
      "step": 4231,
      "token_acc": 0.9074907292954264,
      "train_speed(iter/s)": 0.032318
    },
    {
      "epoch": 0.8200358475027855,
      "grad_norm": 0.10706465691328049,
      "learning_rate": 2.5809456876205897e-05,
      "loss": 0.35384026169776917,
      "memory(GiB)": 78.33,
      "step": 4232,
      "token_acc": 0.8951451380857006,
      "train_speed(iter/s)": 0.032319
    },
    {
      "epoch": 0.8202296177881122,
      "grad_norm": 0.11483049392700195,
      "learning_rate": 2.575556954106142e-05,
      "loss": 0.345024049282074,
      "memory(GiB)": 78.33,
      "step": 4233,
      "token_acc": 0.8970334598137288,
      "train_speed(iter/s)": 0.03232
    },
    {
      "epoch": 0.820423388073439,
      "grad_norm": 0.08666758239269257,
      "learning_rate": 2.57017332364431e-05,
      "loss": 0.2846605181694031,
      "memory(GiB)": 78.33,
      "step": 4234,
      "token_acc": 0.9113689315771956,
      "train_speed(iter/s)": 0.03232
    },
    {
      "epoch": 0.8206171583587657,
      "grad_norm": 0.10516150295734406,
      "learning_rate": 2.564794798446298e-05,
      "loss": 0.353068470954895,
      "memory(GiB)": 78.33,
      "step": 4235,
      "token_acc": 0.8956983240223464,
      "train_speed(iter/s)": 0.032321
    },
    {
      "epoch": 0.8208109286440924,
      "grad_norm": 0.10344908386468887,
      "learning_rate": 2.559421380721207e-05,
      "loss": 0.3344540596008301,
      "memory(GiB)": 78.33,
      "step": 4236,
      "token_acc": 0.8966224100413132,
      "train_speed(iter/s)": 0.032321
    },
    {
      "epoch": 0.8210046989294192,
      "grad_norm": 0.09849216789007187,
      "learning_rate": 2.554053072676049e-05,
      "loss": 0.3291811943054199,
      "memory(GiB)": 78.33,
      "step": 4237,
      "token_acc": 0.9034900284900285,
      "train_speed(iter/s)": 0.032322
    },
    {
      "epoch": 0.8211984692147459,
      "grad_norm": 0.10524845868349075,
      "learning_rate": 2.5486898765157227e-05,
      "loss": 0.36191216111183167,
      "memory(GiB)": 78.33,
      "step": 4238,
      "token_acc": 0.8912652571926766,
      "train_speed(iter/s)": 0.032323
    },
    {
      "epoch": 0.8213922395000727,
      "grad_norm": 0.10183076560497284,
      "learning_rate": 2.5433317944430497e-05,
      "loss": 0.31616219878196716,
      "memory(GiB)": 78.33,
      "step": 4239,
      "token_acc": 0.9054447035789353,
      "train_speed(iter/s)": 0.032323
    },
    {
      "epoch": 0.8215860097853994,
      "grad_norm": 0.10111679881811142,
      "learning_rate": 2.5379788286587317e-05,
      "loss": 0.3214479386806488,
      "memory(GiB)": 78.33,
      "step": 4240,
      "token_acc": 0.9007918449076181,
      "train_speed(iter/s)": 0.032324
    },
    {
      "epoch": 0.8217797800707262,
      "grad_norm": 0.10107513517141342,
      "learning_rate": 2.532630981361376e-05,
      "loss": 0.3232714831829071,
      "memory(GiB)": 78.33,
      "step": 4241,
      "token_acc": 0.9016899178246744,
      "train_speed(iter/s)": 0.032325
    },
    {
      "epoch": 0.8219735503560529,
      "grad_norm": 0.10409308224916458,
      "learning_rate": 2.5272882547474877e-05,
      "loss": 0.3329227566719055,
      "memory(GiB)": 78.33,
      "step": 4242,
      "token_acc": 0.8991776806795668,
      "train_speed(iter/s)": 0.032325
    },
    {
      "epoch": 0.8221673206413797,
      "grad_norm": 0.09440822154283524,
      "learning_rate": 2.5219506510114647e-05,
      "loss": 0.33597031235694885,
      "memory(GiB)": 78.33,
      "step": 4243,
      "token_acc": 0.9017857142857143,
      "train_speed(iter/s)": 0.032326
    },
    {
      "epoch": 0.8223610909267064,
      "grad_norm": 0.09774498641490936,
      "learning_rate": 2.5166181723456147e-05,
      "loss": 0.3483470678329468,
      "memory(GiB)": 78.33,
      "step": 4244,
      "token_acc": 0.8975424647632815,
      "train_speed(iter/s)": 0.032327
    },
    {
      "epoch": 0.8225548612120331,
      "grad_norm": 0.09778609871864319,
      "learning_rate": 2.5112908209401144e-05,
      "loss": 0.29115307331085205,
      "memory(GiB)": 78.33,
      "step": 4245,
      "token_acc": 0.9137274419007227,
      "train_speed(iter/s)": 0.032327
    },
    {
      "epoch": 0.8227486314973599,
      "grad_norm": 0.09965585917234421,
      "learning_rate": 2.5059685989830636e-05,
      "loss": 0.32971906661987305,
      "memory(GiB)": 78.33,
      "step": 4246,
      "token_acc": 0.9021879021879022,
      "train_speed(iter/s)": 0.032328
    },
    {
      "epoch": 0.8229424017826866,
      "grad_norm": 0.10306161642074585,
      "learning_rate": 2.5006515086604368e-05,
      "loss": 0.3302563726902008,
      "memory(GiB)": 78.33,
      "step": 4247,
      "token_acc": 0.9004143366267783,
      "train_speed(iter/s)": 0.032328
    },
    {
      "epoch": 0.8231361720680134,
      "grad_norm": 0.08937028795480728,
      "learning_rate": 2.4953395521561053e-05,
      "loss": 0.30359315872192383,
      "memory(GiB)": 78.33,
      "step": 4248,
      "token_acc": 0.9071417392665241,
      "train_speed(iter/s)": 0.032329
    },
    {
      "epoch": 0.8233299423533401,
      "grad_norm": 0.10544510930776596,
      "learning_rate": 2.4900327316518326e-05,
      "loss": 0.3514764606952667,
      "memory(GiB)": 78.33,
      "step": 4249,
      "token_acc": 0.8938008836033232,
      "train_speed(iter/s)": 0.03233
    },
    {
      "epoch": 0.8235237126386669,
      "grad_norm": 0.10431473702192307,
      "learning_rate": 2.48473104932727e-05,
      "loss": 0.3315693438053131,
      "memory(GiB)": 78.33,
      "step": 4250,
      "token_acc": 0.8996884388561255,
      "train_speed(iter/s)": 0.03233
    },
    {
      "epoch": 0.8237174829239936,
      "grad_norm": 0.10015621036291122,
      "learning_rate": 2.479434507359967e-05,
      "loss": 0.321992427110672,
      "memory(GiB)": 78.33,
      "step": 4251,
      "token_acc": 0.9025872260867175,
      "train_speed(iter/s)": 0.032331
    },
    {
      "epoch": 0.8239112532093203,
      "grad_norm": 0.10041700303554535,
      "learning_rate": 2.474143107925352e-05,
      "loss": 0.3204496204853058,
      "memory(GiB)": 78.33,
      "step": 4252,
      "token_acc": 0.9054709141274239,
      "train_speed(iter/s)": 0.032332
    },
    {
      "epoch": 0.8241050234946471,
      "grad_norm": 0.09637311100959778,
      "learning_rate": 2.4688568531967467e-05,
      "loss": 0.31289243698120117,
      "memory(GiB)": 78.33,
      "step": 4253,
      "token_acc": 0.9049771363624349,
      "train_speed(iter/s)": 0.032332
    },
    {
      "epoch": 0.8242987937799738,
      "grad_norm": 0.10459338128566742,
      "learning_rate": 2.463575745345356e-05,
      "loss": 0.34315067529678345,
      "memory(GiB)": 78.33,
      "step": 4254,
      "token_acc": 0.8960445130013222,
      "train_speed(iter/s)": 0.032333
    },
    {
      "epoch": 0.8244925640653006,
      "grad_norm": 0.09525644779205322,
      "learning_rate": 2.4582997865402727e-05,
      "loss": 0.32378891110420227,
      "memory(GiB)": 78.33,
      "step": 4255,
      "token_acc": 0.9033018867924528,
      "train_speed(iter/s)": 0.032333
    },
    {
      "epoch": 0.8246863343506273,
      "grad_norm": 0.0998750701546669,
      "learning_rate": 2.453028978948477e-05,
      "loss": 0.31236085295677185,
      "memory(GiB)": 78.33,
      "step": 4256,
      "token_acc": 0.9072579542034853,
      "train_speed(iter/s)": 0.032334
    },
    {
      "epoch": 0.8248801046359541,
      "grad_norm": 0.09134574234485626,
      "learning_rate": 2.4477633247348238e-05,
      "loss": 0.3001454472541809,
      "memory(GiB)": 78.33,
      "step": 4257,
      "token_acc": 0.9097239492663517,
      "train_speed(iter/s)": 0.032335
    },
    {
      "epoch": 0.8250738749212808,
      "grad_norm": 0.09457145631313324,
      "learning_rate": 2.4425028260620715e-05,
      "loss": 0.32455000281333923,
      "memory(GiB)": 78.33,
      "step": 4258,
      "token_acc": 0.902986820556812,
      "train_speed(iter/s)": 0.032335
    },
    {
      "epoch": 0.8252676452066076,
      "grad_norm": 0.1064552590250969,
      "learning_rate": 2.4372474850908404e-05,
      "loss": 0.3257259130477905,
      "memory(GiB)": 78.33,
      "step": 4259,
      "token_acc": 0.9031152183633925,
      "train_speed(iter/s)": 0.032336
    },
    {
      "epoch": 0.8254614154919343,
      "grad_norm": 0.10509302467107773,
      "learning_rate": 2.4319973039796397e-05,
      "loss": 0.3298068344593048,
      "memory(GiB)": 78.33,
      "step": 4260,
      "token_acc": 0.8993223921422355,
      "train_speed(iter/s)": 0.032336
    },
    {
      "epoch": 0.825655185777261,
      "grad_norm": 0.09444321691989899,
      "learning_rate": 2.4267522848848635e-05,
      "loss": 0.29824700951576233,
      "memory(GiB)": 78.33,
      "step": 4261,
      "token_acc": 0.908746618575293,
      "train_speed(iter/s)": 0.032337
    },
    {
      "epoch": 0.8258489560625878,
      "grad_norm": 0.09932407736778259,
      "learning_rate": 2.4215124299607802e-05,
      "loss": 0.33362001180648804,
      "memory(GiB)": 78.33,
      "step": 4262,
      "token_acc": 0.8990427838595377,
      "train_speed(iter/s)": 0.032338
    },
    {
      "epoch": 0.8260427263479145,
      "grad_norm": 0.09288123995065689,
      "learning_rate": 2.416277741359538e-05,
      "loss": 0.31670305132865906,
      "memory(GiB)": 78.33,
      "step": 4263,
      "token_acc": 0.9042440573573285,
      "train_speed(iter/s)": 0.032338
    },
    {
      "epoch": 0.8262364966332413,
      "grad_norm": 0.11048420518636703,
      "learning_rate": 2.411048221231162e-05,
      "loss": 0.3662915825843811,
      "memory(GiB)": 78.33,
      "step": 4264,
      "token_acc": 0.8945862079354081,
      "train_speed(iter/s)": 0.032339
    },
    {
      "epoch": 0.826430266918568,
      "grad_norm": 0.10002768039703369,
      "learning_rate": 2.4058238717235628e-05,
      "loss": 0.32539236545562744,
      "memory(GiB)": 78.33,
      "step": 4265,
      "token_acc": 0.9030578297219892,
      "train_speed(iter/s)": 0.03234
    },
    {
      "epoch": 0.8266240372038948,
      "grad_norm": 0.10625031590461731,
      "learning_rate": 2.4006046949825186e-05,
      "loss": 0.31954246759414673,
      "memory(GiB)": 78.33,
      "step": 4266,
      "token_acc": 0.9040422214225741,
      "train_speed(iter/s)": 0.03234
    },
    {
      "epoch": 0.8268178074892215,
      "grad_norm": 0.09548264741897583,
      "learning_rate": 2.3953906931516848e-05,
      "loss": 0.31130972504615784,
      "memory(GiB)": 78.33,
      "step": 4267,
      "token_acc": 0.9068991470145509,
      "train_speed(iter/s)": 0.032341
    },
    {
      "epoch": 0.8270115777745483,
      "grad_norm": 0.1018620952963829,
      "learning_rate": 2.390181868372593e-05,
      "loss": 0.37592288851737976,
      "memory(GiB)": 78.33,
      "step": 4268,
      "token_acc": 0.8873175527489979,
      "train_speed(iter/s)": 0.032341
    },
    {
      "epoch": 0.827205348059875,
      "grad_norm": 0.10313120484352112,
      "learning_rate": 2.384978222784646e-05,
      "loss": 0.3224887251853943,
      "memory(GiB)": 78.33,
      "step": 4269,
      "token_acc": 0.9022740524781341,
      "train_speed(iter/s)": 0.032342
    },
    {
      "epoch": 0.8273991183452017,
      "grad_norm": 0.1032845601439476,
      "learning_rate": 2.379779758525123e-05,
      "loss": 0.33453071117401123,
      "memory(GiB)": 78.33,
      "step": 4270,
      "token_acc": 0.8983879998821207,
      "train_speed(iter/s)": 0.032343
    },
    {
      "epoch": 0.8275928886305285,
      "grad_norm": 0.09773199260234833,
      "learning_rate": 2.3745864777291674e-05,
      "loss": 0.3257528841495514,
      "memory(GiB)": 78.33,
      "step": 4271,
      "token_acc": 0.9016068290233492,
      "train_speed(iter/s)": 0.032343
    },
    {
      "epoch": 0.8277866589158552,
      "grad_norm": 0.09065309911966324,
      "learning_rate": 2.369398382529807e-05,
      "loss": 0.29769429564476013,
      "memory(GiB)": 78.33,
      "step": 4272,
      "token_acc": 0.9092235329627487,
      "train_speed(iter/s)": 0.032344
    },
    {
      "epoch": 0.827980429201182,
      "grad_norm": 0.08960139006376266,
      "learning_rate": 2.3642154750579272e-05,
      "loss": 0.31141579151153564,
      "memory(GiB)": 78.33,
      "step": 4273,
      "token_acc": 0.9071020707684546,
      "train_speed(iter/s)": 0.032345
    },
    {
      "epoch": 0.8281741994865087,
      "grad_norm": 0.10664994269609451,
      "learning_rate": 2.3590377574422892e-05,
      "loss": 0.37568601965904236,
      "memory(GiB)": 78.33,
      "step": 4274,
      "token_acc": 0.8949000498366465,
      "train_speed(iter/s)": 0.032345
    },
    {
      "epoch": 0.8283679697718355,
      "grad_norm": 0.11605322360992432,
      "learning_rate": 2.3538652318095198e-05,
      "loss": 0.3583415746688843,
      "memory(GiB)": 78.33,
      "step": 4275,
      "token_acc": 0.8928342520189719,
      "train_speed(iter/s)": 0.032346
    },
    {
      "epoch": 0.8285617400571622,
      "grad_norm": 0.10450875759124756,
      "learning_rate": 2.348697900284111e-05,
      "loss": 0.34997832775115967,
      "memory(GiB)": 78.33,
      "step": 4276,
      "token_acc": 0.8950859618248274,
      "train_speed(iter/s)": 0.032347
    },
    {
      "epoch": 0.828755510342489,
      "grad_norm": 0.09127886593341827,
      "learning_rate": 2.3435357649884357e-05,
      "loss": 0.31003209948539734,
      "memory(GiB)": 78.33,
      "step": 4277,
      "token_acc": 0.9049302739590545,
      "train_speed(iter/s)": 0.032347
    },
    {
      "epoch": 0.8289492806278157,
      "grad_norm": 0.10241317003965378,
      "learning_rate": 2.3383788280427074e-05,
      "loss": 0.347523957490921,
      "memory(GiB)": 78.33,
      "step": 4278,
      "token_acc": 0.8969976905311778,
      "train_speed(iter/s)": 0.032348
    },
    {
      "epoch": 0.8291430509131424,
      "grad_norm": 0.09595993906259537,
      "learning_rate": 2.3332270915650285e-05,
      "loss": 0.3249264061450958,
      "memory(GiB)": 78.33,
      "step": 4279,
      "token_acc": 0.9030240097254584,
      "train_speed(iter/s)": 0.032348
    },
    {
      "epoch": 0.8293368211984692,
      "grad_norm": 0.0956515520811081,
      "learning_rate": 2.328080557671352e-05,
      "loss": 0.3111531734466553,
      "memory(GiB)": 78.33,
      "step": 4280,
      "token_acc": 0.9046177726038028,
      "train_speed(iter/s)": 0.032349
    },
    {
      "epoch": 0.8295305914837959,
      "grad_norm": 0.08913633227348328,
      "learning_rate": 2.3229392284754994e-05,
      "loss": 0.30270490050315857,
      "memory(GiB)": 78.33,
      "step": 4281,
      "token_acc": 0.9090954225002482,
      "train_speed(iter/s)": 0.03235
    },
    {
      "epoch": 0.8297243617691227,
      "grad_norm": 0.10194297879934311,
      "learning_rate": 2.3178031060891507e-05,
      "loss": 0.35240015387535095,
      "memory(GiB)": 78.33,
      "step": 4282,
      "token_acc": 0.8955933833943508,
      "train_speed(iter/s)": 0.03235
    },
    {
      "epoch": 0.8299181320544494,
      "grad_norm": 0.10122035443782806,
      "learning_rate": 2.312672192621846e-05,
      "loss": 0.33214977383613586,
      "memory(GiB)": 78.33,
      "step": 4283,
      "token_acc": 0.9003510196977169,
      "train_speed(iter/s)": 0.032351
    },
    {
      "epoch": 0.8301119023397762,
      "grad_norm": 0.09273627400398254,
      "learning_rate": 2.307546490180997e-05,
      "loss": 0.3047739863395691,
      "memory(GiB)": 78.33,
      "step": 4284,
      "token_acc": 0.909302266165913,
      "train_speed(iter/s)": 0.032351
    },
    {
      "epoch": 0.8303056726251029,
      "grad_norm": 0.1077071949839592,
      "learning_rate": 2.3024260008718642e-05,
      "loss": 0.34293609857559204,
      "memory(GiB)": 78.33,
      "step": 4285,
      "token_acc": 0.896573135034446,
      "train_speed(iter/s)": 0.032352
    },
    {
      "epoch": 0.8304994429104297,
      "grad_norm": 0.0920027419924736,
      "learning_rate": 2.2973107267975703e-05,
      "loss": 0.3134732246398926,
      "memory(GiB)": 78.33,
      "step": 4286,
      "token_acc": 0.9056259577262666,
      "train_speed(iter/s)": 0.032353
    },
    {
      "epoch": 0.8306932131957564,
      "grad_norm": 0.09819093346595764,
      "learning_rate": 2.292200670059095e-05,
      "loss": 0.32229092717170715,
      "memory(GiB)": 78.33,
      "step": 4287,
      "token_acc": 0.9032614234754544,
      "train_speed(iter/s)": 0.032353
    },
    {
      "epoch": 0.8308869834810831,
      "grad_norm": 0.10240425914525986,
      "learning_rate": 2.2870958327552774e-05,
      "loss": 0.35109835863113403,
      "memory(GiB)": 78.33,
      "step": 4288,
      "token_acc": 0.8959884028881269,
      "train_speed(iter/s)": 0.032354
    },
    {
      "epoch": 0.8310807537664099,
      "grad_norm": 0.10077088326215744,
      "learning_rate": 2.2819962169828088e-05,
      "loss": 0.32499995827674866,
      "memory(GiB)": 78.33,
      "step": 4289,
      "token_acc": 0.9027956087719758,
      "train_speed(iter/s)": 0.032355
    },
    {
      "epoch": 0.8312745240517366,
      "grad_norm": 0.09702709317207336,
      "learning_rate": 2.276901824836237e-05,
      "loss": 0.3244347870349884,
      "memory(GiB)": 78.33,
      "step": 4290,
      "token_acc": 0.9018396633577902,
      "train_speed(iter/s)": 0.032355
    },
    {
      "epoch": 0.8314682943370634,
      "grad_norm": 0.10198424011468887,
      "learning_rate": 2.2718126584079734e-05,
      "loss": 0.3321649730205536,
      "memory(GiB)": 78.33,
      "step": 4291,
      "token_acc": 0.9009141765588757,
      "train_speed(iter/s)": 0.032356
    },
    {
      "epoch": 0.8316620646223901,
      "grad_norm": 0.1072782501578331,
      "learning_rate": 2.266728719788269e-05,
      "loss": 0.3425625264644623,
      "memory(GiB)": 78.33,
      "step": 4292,
      "token_acc": 0.8971817900278725,
      "train_speed(iter/s)": 0.032356
    },
    {
      "epoch": 0.8318558349077169,
      "grad_norm": 0.10138484835624695,
      "learning_rate": 2.2616500110652352e-05,
      "loss": 0.32518693804740906,
      "memory(GiB)": 78.33,
      "step": 4293,
      "token_acc": 0.9017971942055487,
      "train_speed(iter/s)": 0.032357
    },
    {
      "epoch": 0.8320496051930436,
      "grad_norm": 0.09211868792772293,
      "learning_rate": 2.2565765343248353e-05,
      "loss": 0.3022071123123169,
      "memory(GiB)": 78.33,
      "step": 4294,
      "token_acc": 0.9080617941061924,
      "train_speed(iter/s)": 0.032358
    },
    {
      "epoch": 0.8322433754783704,
      "grad_norm": 0.09903164952993393,
      "learning_rate": 2.2515082916508824e-05,
      "loss": 0.3520633280277252,
      "memory(GiB)": 78.33,
      "step": 4295,
      "token_acc": 0.8949784239892258,
      "train_speed(iter/s)": 0.032358
    },
    {
      "epoch": 0.8324371457636971,
      "grad_norm": 0.10028848052024841,
      "learning_rate": 2.24644528512504e-05,
      "loss": 0.34844347834587097,
      "memory(GiB)": 78.33,
      "step": 4296,
      "token_acc": 0.8961258624855694,
      "train_speed(iter/s)": 0.032359
    },
    {
      "epoch": 0.8326309160490238,
      "grad_norm": 0.0993753969669342,
      "learning_rate": 2.2413875168268154e-05,
      "loss": 0.35692083835601807,
      "memory(GiB)": 78.33,
      "step": 4297,
      "token_acc": 0.8944187141847869,
      "train_speed(iter/s)": 0.03236
    },
    {
      "epoch": 0.8328246863343506,
      "grad_norm": 0.10167410969734192,
      "learning_rate": 2.2363349888335775e-05,
      "loss": 0.349956214427948,
      "memory(GiB)": 78.33,
      "step": 4298,
      "token_acc": 0.895724891202321,
      "train_speed(iter/s)": 0.03236
    },
    {
      "epoch": 0.8330184566196773,
      "grad_norm": 0.0991780236363411,
      "learning_rate": 2.2312877032205346e-05,
      "loss": 0.3107830584049225,
      "memory(GiB)": 78.33,
      "step": 4299,
      "token_acc": 0.9071795583946675,
      "train_speed(iter/s)": 0.032361
    },
    {
      "epoch": 0.8332122269050041,
      "grad_norm": 0.09722703695297241,
      "learning_rate": 2.22624566206074e-05,
      "loss": 0.3044697344303131,
      "memory(GiB)": 78.33,
      "step": 4300,
      "token_acc": 0.9069954220544535,
      "train_speed(iter/s)": 0.032362
    },
    {
      "epoch": 0.8334059971903308,
      "grad_norm": 0.09022455662488937,
      "learning_rate": 2.2212088674250956e-05,
      "loss": 0.2870331108570099,
      "memory(GiB)": 78.33,
      "step": 4301,
      "token_acc": 0.91363100759646,
      "train_speed(iter/s)": 0.032362
    },
    {
      "epoch": 0.8335997674756576,
      "grad_norm": 0.1041305810213089,
      "learning_rate": 2.216177321382348e-05,
      "loss": 0.3358173966407776,
      "memory(GiB)": 78.33,
      "step": 4302,
      "token_acc": 0.8962132245849112,
      "train_speed(iter/s)": 0.032363
    },
    {
      "epoch": 0.8337935377609843,
      "grad_norm": 0.10265124589204788,
      "learning_rate": 2.2111510259990913e-05,
      "loss": 0.3599543869495392,
      "memory(GiB)": 78.33,
      "step": 4303,
      "token_acc": 0.8918628427620197,
      "train_speed(iter/s)": 0.032363
    },
    {
      "epoch": 0.833987308046311,
      "grad_norm": 0.0928591713309288,
      "learning_rate": 2.2061299833397532e-05,
      "loss": 0.29637521505355835,
      "memory(GiB)": 78.33,
      "step": 4304,
      "token_acc": 0.910638866613205,
      "train_speed(iter/s)": 0.032364
    },
    {
      "epoch": 0.8341810783316378,
      "grad_norm": 0.0997898206114769,
      "learning_rate": 2.2011141954666185e-05,
      "loss": 0.33638525009155273,
      "memory(GiB)": 78.33,
      "step": 4305,
      "token_acc": 0.8974762101779065,
      "train_speed(iter/s)": 0.032365
    },
    {
      "epoch": 0.8343748486169645,
      "grad_norm": 0.10566650331020355,
      "learning_rate": 2.1961036644398035e-05,
      "loss": 0.3401532769203186,
      "memory(GiB)": 78.33,
      "step": 4306,
      "token_acc": 0.8974769961412882,
      "train_speed(iter/s)": 0.032365
    },
    {
      "epoch": 0.8345686189022913,
      "grad_norm": 0.11155321449041367,
      "learning_rate": 2.1910983923172686e-05,
      "loss": 0.29791954159736633,
      "memory(GiB)": 78.33,
      "step": 4307,
      "token_acc": 0.910283068563032,
      "train_speed(iter/s)": 0.032366
    },
    {
      "epoch": 0.834762389187618,
      "grad_norm": 0.10973121225833893,
      "learning_rate": 2.1860983811548118e-05,
      "loss": 0.3326932191848755,
      "memory(GiB)": 78.33,
      "step": 4308,
      "token_acc": 0.8998096353785328,
      "train_speed(iter/s)": 0.032366
    },
    {
      "epoch": 0.8349561594729448,
      "grad_norm": 0.08969270437955856,
      "learning_rate": 2.1811036330060676e-05,
      "loss": 0.30071404576301575,
      "memory(GiB)": 78.33,
      "step": 4309,
      "token_acc": 0.9097683498797705,
      "train_speed(iter/s)": 0.032367
    },
    {
      "epoch": 0.8351499297582715,
      "grad_norm": 0.09654086083173752,
      "learning_rate": 2.1761141499225278e-05,
      "loss": 0.3438301384449005,
      "memory(GiB)": 78.33,
      "step": 4310,
      "token_acc": 0.8972290781405053,
      "train_speed(iter/s)": 0.032368
    },
    {
      "epoch": 0.8353437000435984,
      "grad_norm": 0.09115693718194962,
      "learning_rate": 2.171129933953489e-05,
      "loss": 0.2993430197238922,
      "memory(GiB)": 78.33,
      "step": 4311,
      "token_acc": 0.9102314363897779,
      "train_speed(iter/s)": 0.032368
    },
    {
      "epoch": 0.8355374703289251,
      "grad_norm": 0.1047009527683258,
      "learning_rate": 2.1661509871461168e-05,
      "loss": 0.33541637659072876,
      "memory(GiB)": 78.33,
      "step": 4312,
      "token_acc": 0.8988593808067237,
      "train_speed(iter/s)": 0.032369
    },
    {
      "epoch": 0.8357312406142519,
      "grad_norm": 0.09559044986963272,
      "learning_rate": 2.1611773115453913e-05,
      "loss": 0.3060503900051117,
      "memory(GiB)": 78.33,
      "step": 4313,
      "token_acc": 0.9065379777703622,
      "train_speed(iter/s)": 0.03237
    },
    {
      "epoch": 0.8359250108995786,
      "grad_norm": 0.10747367143630981,
      "learning_rate": 2.1562089091941376e-05,
      "loss": 0.3518577218055725,
      "memory(GiB)": 78.33,
      "step": 4314,
      "token_acc": 0.8972739541160594,
      "train_speed(iter/s)": 0.03237
    },
    {
      "epoch": 0.8361187811849053,
      "grad_norm": 0.10569079220294952,
      "learning_rate": 2.1512457821330102e-05,
      "loss": 0.33122116327285767,
      "memory(GiB)": 78.33,
      "step": 4315,
      "token_acc": 0.9022820362785254,
      "train_speed(iter/s)": 0.032371
    },
    {
      "epoch": 0.8363125514702321,
      "grad_norm": 0.10672451555728912,
      "learning_rate": 2.1462879324004973e-05,
      "loss": 0.356486439704895,
      "memory(GiB)": 78.33,
      "step": 4316,
      "token_acc": 0.892833182626958,
      "train_speed(iter/s)": 0.032371
    },
    {
      "epoch": 0.8365063217555588,
      "grad_norm": 0.10485262423753738,
      "learning_rate": 2.1413353620329294e-05,
      "loss": 0.33778226375579834,
      "memory(GiB)": 78.33,
      "step": 4317,
      "token_acc": 0.8977509922093194,
      "train_speed(iter/s)": 0.032372
    },
    {
      "epoch": 0.8367000920408856,
      "grad_norm": 0.1112034022808075,
      "learning_rate": 2.136388073064446e-05,
      "loss": 0.31891724467277527,
      "memory(GiB)": 78.33,
      "step": 4318,
      "token_acc": 0.9021532012195121,
      "train_speed(iter/s)": 0.032373
    },
    {
      "epoch": 0.8368938623262123,
      "grad_norm": 0.09961036592721939,
      "learning_rate": 2.131446067527044e-05,
      "loss": 0.3315172493457794,
      "memory(GiB)": 78.33,
      "step": 4319,
      "token_acc": 0.9015923147782603,
      "train_speed(iter/s)": 0.032373
    },
    {
      "epoch": 0.8370876326115391,
      "grad_norm": 0.10558105260133743,
      "learning_rate": 2.126509347450534e-05,
      "loss": 0.36373376846313477,
      "memory(GiB)": 78.33,
      "step": 4320,
      "token_acc": 0.8906225980015373,
      "train_speed(iter/s)": 0.032374
    },
    {
      "epoch": 0.8372814028968658,
      "grad_norm": 0.09915536642074585,
      "learning_rate": 2.1215779148625578e-05,
      "loss": 0.3262394964694977,
      "memory(GiB)": 78.33,
      "step": 4321,
      "token_acc": 0.8996407847471677,
      "train_speed(iter/s)": 0.032374
    },
    {
      "epoch": 0.8374751731821926,
      "grad_norm": 0.11209924519062042,
      "learning_rate": 2.11665177178859e-05,
      "loss": 0.3615911900997162,
      "memory(GiB)": 78.33,
      "step": 4322,
      "token_acc": 0.8943596998400787,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.8376689434675193,
      "grad_norm": 0.09516075998544693,
      "learning_rate": 2.111730920251924e-05,
      "loss": 0.3138526678085327,
      "memory(GiB)": 78.33,
      "step": 4323,
      "token_acc": 0.9044929966462814,
      "train_speed(iter/s)": 0.032376
    },
    {
      "epoch": 0.837862713752846,
      "grad_norm": 0.10355053097009659,
      "learning_rate": 2.1068153622736943e-05,
      "loss": 0.3259844183921814,
      "memory(GiB)": 78.33,
      "step": 4324,
      "token_acc": 0.902735473289597,
      "train_speed(iter/s)": 0.032376
    },
    {
      "epoch": 0.8380564840381728,
      "grad_norm": 0.11187773942947388,
      "learning_rate": 2.101905099872848e-05,
      "loss": 0.3713458776473999,
      "memory(GiB)": 78.33,
      "step": 4325,
      "token_acc": 0.888154201235107,
      "train_speed(iter/s)": 0.032377
    },
    {
      "epoch": 0.8382502543234995,
      "grad_norm": 0.10848337411880493,
      "learning_rate": 2.0970001350661635e-05,
      "loss": 0.3444962799549103,
      "memory(GiB)": 78.33,
      "step": 4326,
      "token_acc": 0.8968367889420521,
      "train_speed(iter/s)": 0.032378
    },
    {
      "epoch": 0.8384440246088263,
      "grad_norm": 0.08781024068593979,
      "learning_rate": 2.0921004698682407e-05,
      "loss": 0.328144371509552,
      "memory(GiB)": 78.33,
      "step": 4327,
      "token_acc": 0.9009247631008106,
      "train_speed(iter/s)": 0.032378
    },
    {
      "epoch": 0.838637794894153,
      "grad_norm": 0.09519962966442108,
      "learning_rate": 2.087206106291502e-05,
      "loss": 0.31252622604370117,
      "memory(GiB)": 78.33,
      "step": 4328,
      "token_acc": 0.9068911656474201,
      "train_speed(iter/s)": 0.032379
    },
    {
      "epoch": 0.8388315651794798,
      "grad_norm": 0.09435832500457764,
      "learning_rate": 2.082317046346197e-05,
      "loss": 0.32768622040748596,
      "memory(GiB)": 78.33,
      "step": 4329,
      "token_acc": 0.900981393912031,
      "train_speed(iter/s)": 0.032379
    },
    {
      "epoch": 0.8390253354648065,
      "grad_norm": 0.10018561035394669,
      "learning_rate": 2.077433292040388e-05,
      "loss": 0.3233901560306549,
      "memory(GiB)": 78.33,
      "step": 4330,
      "token_acc": 0.9020335985853227,
      "train_speed(iter/s)": 0.03238
    },
    {
      "epoch": 0.8392191057501333,
      "grad_norm": 0.09643685072660446,
      "learning_rate": 2.072554845379974e-05,
      "loss": 0.316950261592865,
      "memory(GiB)": 78.33,
      "step": 4331,
      "token_acc": 0.9074275988617929,
      "train_speed(iter/s)": 0.032381
    },
    {
      "epoch": 0.83941287603546,
      "grad_norm": 0.0906091183423996,
      "learning_rate": 2.067681708368657e-05,
      "loss": 0.31494781374931335,
      "memory(GiB)": 78.33,
      "step": 4332,
      "token_acc": 0.9049370490875654,
      "train_speed(iter/s)": 0.032381
    },
    {
      "epoch": 0.8396066463207867,
      "grad_norm": 0.10366849601268768,
      "learning_rate": 2.0628138830079695e-05,
      "loss": 0.3571404218673706,
      "memory(GiB)": 78.33,
      "step": 4333,
      "token_acc": 0.8948438109026116,
      "train_speed(iter/s)": 0.032382
    },
    {
      "epoch": 0.8398004166061135,
      "grad_norm": 0.10383996367454529,
      "learning_rate": 2.0579513712972535e-05,
      "loss": 0.3651527166366577,
      "memory(GiB)": 78.33,
      "step": 4334,
      "token_acc": 0.8928912539227726,
      "train_speed(iter/s)": 0.032383
    },
    {
      "epoch": 0.8399941868914402,
      "grad_norm": 0.1002940759062767,
      "learning_rate": 2.0530941752336767e-05,
      "loss": 0.33056554198265076,
      "memory(GiB)": 78.33,
      "step": 4335,
      "token_acc": 0.9013376077824453,
      "train_speed(iter/s)": 0.032383
    },
    {
      "epoch": 0.840187957176767,
      "grad_norm": 0.09768969565629959,
      "learning_rate": 2.0482422968122198e-05,
      "loss": 0.327396422624588,
      "memory(GiB)": 78.33,
      "step": 4336,
      "token_acc": 0.9023116889636305,
      "train_speed(iter/s)": 0.032384
    },
    {
      "epoch": 0.8403817274620937,
      "grad_norm": 0.0857066735625267,
      "learning_rate": 2.043395738025674e-05,
      "loss": 0.30105772614479065,
      "memory(GiB)": 78.33,
      "step": 4337,
      "token_acc": 0.9095752105336621,
      "train_speed(iter/s)": 0.032385
    },
    {
      "epoch": 0.8405754977474205,
      "grad_norm": 0.09931961447000504,
      "learning_rate": 2.0385545008646597e-05,
      "loss": 0.3311472535133362,
      "memory(GiB)": 78.33,
      "step": 4338,
      "token_acc": 0.9012172284644194,
      "train_speed(iter/s)": 0.032385
    },
    {
      "epoch": 0.8407692680327472,
      "grad_norm": 0.10935309529304504,
      "learning_rate": 2.0337185873176004e-05,
      "loss": 0.34328311681747437,
      "memory(GiB)": 78.33,
      "step": 4339,
      "token_acc": 0.8972643180476284,
      "train_speed(iter/s)": 0.032386
    },
    {
      "epoch": 0.840963038318074,
      "grad_norm": 0.08845790475606918,
      "learning_rate": 2.0288879993707335e-05,
      "loss": 0.2846664786338806,
      "memory(GiB)": 78.33,
      "step": 4340,
      "token_acc": 0.9113034491667743,
      "train_speed(iter/s)": 0.032386
    },
    {
      "epoch": 0.8411568086034007,
      "grad_norm": 0.10597343742847443,
      "learning_rate": 2.0240627390081137e-05,
      "loss": 0.32449671626091003,
      "memory(GiB)": 78.33,
      "step": 4341,
      "token_acc": 0.9027229011608319,
      "train_speed(iter/s)": 0.032387
    },
    {
      "epoch": 0.8413505788887274,
      "grad_norm": 0.1017051637172699,
      "learning_rate": 2.0192428082115992e-05,
      "loss": 0.32706570625305176,
      "memory(GiB)": 78.33,
      "step": 4342,
      "token_acc": 0.9033635878270155,
      "train_speed(iter/s)": 0.032388
    },
    {
      "epoch": 0.8415443491740542,
      "grad_norm": 0.10161115974187851,
      "learning_rate": 2.0144282089608778e-05,
      "loss": 0.34076470136642456,
      "memory(GiB)": 78.33,
      "step": 4343,
      "token_acc": 0.8972409669428547,
      "train_speed(iter/s)": 0.032388
    },
    {
      "epoch": 0.8417381194593809,
      "grad_norm": 0.10345038771629333,
      "learning_rate": 2.009618943233419e-05,
      "loss": 0.3462626338005066,
      "memory(GiB)": 78.33,
      "step": 4344,
      "token_acc": 0.8954670063412068,
      "train_speed(iter/s)": 0.032389
    },
    {
      "epoch": 0.8419318897447077,
      "grad_norm": 0.10364454239606857,
      "learning_rate": 2.0048150130045303e-05,
      "loss": 0.35689064860343933,
      "memory(GiB)": 78.33,
      "step": 4345,
      "token_acc": 0.893178860336336,
      "train_speed(iter/s)": 0.03239
    },
    {
      "epoch": 0.8421256600300344,
      "grad_norm": 0.08873011916875839,
      "learning_rate": 2.000016420247308e-05,
      "loss": 0.28760528564453125,
      "memory(GiB)": 78.33,
      "step": 4346,
      "token_acc": 0.9117839149094981,
      "train_speed(iter/s)": 0.03239
    },
    {
      "epoch": 0.8423194303153612,
      "grad_norm": 0.10340781509876251,
      "learning_rate": 1.9952231669326668e-05,
      "loss": 0.3450697958469391,
      "memory(GiB)": 78.33,
      "step": 4347,
      "token_acc": 0.896965636602183,
      "train_speed(iter/s)": 0.032391
    },
    {
      "epoch": 0.8425132006006879,
      "grad_norm": 0.08939331769943237,
      "learning_rate": 1.9904352550293224e-05,
      "loss": 0.28838780522346497,
      "memory(GiB)": 78.33,
      "step": 4348,
      "token_acc": 0.9093310888218653,
      "train_speed(iter/s)": 0.032391
    },
    {
      "epoch": 0.8427069708860147,
      "grad_norm": 0.09946542233228683,
      "learning_rate": 1.9856526865037947e-05,
      "loss": 0.322831392288208,
      "memory(GiB)": 78.33,
      "step": 4349,
      "token_acc": 0.90141040206157,
      "train_speed(iter/s)": 0.032392
    },
    {
      "epoch": 0.8429007411713414,
      "grad_norm": 0.10646877437829971,
      "learning_rate": 1.980875463320426e-05,
      "loss": 0.3633573651313782,
      "memory(GiB)": 78.33,
      "step": 4350,
      "token_acc": 0.8931559868080351,
      "train_speed(iter/s)": 0.032393
    },
    {
      "epoch": 0.8430945114566681,
      "grad_norm": 0.09976527839899063,
      "learning_rate": 1.9761035874413333e-05,
      "loss": 0.3256649374961853,
      "memory(GiB)": 78.33,
      "step": 4351,
      "token_acc": 0.9018026445528784,
      "train_speed(iter/s)": 0.032393
    },
    {
      "epoch": 0.8432882817419949,
      "grad_norm": 0.10646556317806244,
      "learning_rate": 1.9713370608264674e-05,
      "loss": 0.32194364070892334,
      "memory(GiB)": 78.33,
      "step": 4352,
      "token_acc": 0.9037880343886623,
      "train_speed(iter/s)": 0.032394
    },
    {
      "epoch": 0.8434820520273216,
      "grad_norm": 0.10743288695812225,
      "learning_rate": 1.966575885433565e-05,
      "loss": 0.34353500604629517,
      "memory(GiB)": 78.33,
      "step": 4353,
      "token_acc": 0.8954018292107992,
      "train_speed(iter/s)": 0.032394
    },
    {
      "epoch": 0.8436758223126484,
      "grad_norm": 0.1006690040230751,
      "learning_rate": 1.9618200632181673e-05,
      "loss": 0.3364703059196472,
      "memory(GiB)": 78.33,
      "step": 4354,
      "token_acc": 0.8980853059647961,
      "train_speed(iter/s)": 0.032395
    },
    {
      "epoch": 0.8438695925979751,
      "grad_norm": 0.09362746775150299,
      "learning_rate": 1.9570695961336203e-05,
      "loss": 0.28917554020881653,
      "memory(GiB)": 78.33,
      "step": 4355,
      "token_acc": 0.9127130038690557,
      "train_speed(iter/s)": 0.032396
    },
    {
      "epoch": 0.8440633628833019,
      "grad_norm": 0.11022236943244934,
      "learning_rate": 1.9523244861310626e-05,
      "loss": 0.31461644172668457,
      "memory(GiB)": 78.33,
      "step": 4356,
      "token_acc": 0.905908453993461,
      "train_speed(iter/s)": 0.032396
    },
    {
      "epoch": 0.8442571331686286,
      "grad_norm": 0.09002205729484558,
      "learning_rate": 1.9475847351594458e-05,
      "loss": 0.28434687852859497,
      "memory(GiB)": 78.33,
      "step": 4357,
      "token_acc": 0.9123356771629471,
      "train_speed(iter/s)": 0.032397
    },
    {
      "epoch": 0.8444509034539553,
      "grad_norm": 0.09616533666849136,
      "learning_rate": 1.9428503451655125e-05,
      "loss": 0.33912795782089233,
      "memory(GiB)": 78.33,
      "step": 4358,
      "token_acc": 0.8976422492479367,
      "train_speed(iter/s)": 0.032397
    },
    {
      "epoch": 0.8446446737392821,
      "grad_norm": 0.09272696822881699,
      "learning_rate": 1.9381213180938003e-05,
      "loss": 0.30497756600379944,
      "memory(GiB)": 78.33,
      "step": 4359,
      "token_acc": 0.9095019342359768,
      "train_speed(iter/s)": 0.032398
    },
    {
      "epoch": 0.8448384440246088,
      "grad_norm": 0.09096968919038773,
      "learning_rate": 1.9333976558866476e-05,
      "loss": 0.295366108417511,
      "memory(GiB)": 78.33,
      "step": 4360,
      "token_acc": 0.9114807813484562,
      "train_speed(iter/s)": 0.032399
    },
    {
      "epoch": 0.8450322143099356,
      "grad_norm": 0.1177552342414856,
      "learning_rate": 1.928679360484194e-05,
      "loss": 0.3394724428653717,
      "memory(GiB)": 78.33,
      "step": 4361,
      "token_acc": 0.8994290351563632,
      "train_speed(iter/s)": 0.032399
    },
    {
      "epoch": 0.8452259845952623,
      "grad_norm": 0.10825799405574799,
      "learning_rate": 1.9239664338243637e-05,
      "loss": 0.32095205783843994,
      "memory(GiB)": 78.33,
      "step": 4362,
      "token_acc": 0.90408615136876,
      "train_speed(iter/s)": 0.0324
    },
    {
      "epoch": 0.8454197548805891,
      "grad_norm": 0.09334609657526016,
      "learning_rate": 1.9192588778428842e-05,
      "loss": 0.30651578307151794,
      "memory(GiB)": 78.33,
      "step": 4363,
      "token_acc": 0.907121790842721,
      "train_speed(iter/s)": 0.0324
    },
    {
      "epoch": 0.8456135251659158,
      "grad_norm": 0.10030210018157959,
      "learning_rate": 1.91455669447328e-05,
      "loss": 0.3436662554740906,
      "memory(GiB)": 78.33,
      "step": 4364,
      "token_acc": 0.9008844323772145,
      "train_speed(iter/s)": 0.032401
    },
    {
      "epoch": 0.8458072954512426,
      "grad_norm": 0.09695718437433243,
      "learning_rate": 1.909859885646861e-05,
      "loss": 0.33061671257019043,
      "memory(GiB)": 78.33,
      "step": 4365,
      "token_acc": 0.9015153412648717,
      "train_speed(iter/s)": 0.032402
    },
    {
      "epoch": 0.8460010657365693,
      "grad_norm": 0.1165538877248764,
      "learning_rate": 1.9051684532927332e-05,
      "loss": 0.3574296534061432,
      "memory(GiB)": 78.33,
      "step": 4366,
      "token_acc": 0.8923136197264496,
      "train_speed(iter/s)": 0.032402
    },
    {
      "epoch": 0.846194836021896,
      "grad_norm": 0.09451232105493546,
      "learning_rate": 1.9004823993377927e-05,
      "loss": 0.3224382996559143,
      "memory(GiB)": 78.33,
      "step": 4367,
      "token_acc": 0.9037022153011504,
      "train_speed(iter/s)": 0.032403
    },
    {
      "epoch": 0.8463886063072228,
      "grad_norm": 0.08888131380081177,
      "learning_rate": 1.895801725706727e-05,
      "loss": 0.2944706082344055,
      "memory(GiB)": 78.33,
      "step": 4368,
      "token_acc": 0.9106517341753416,
      "train_speed(iter/s)": 0.032403
    },
    {
      "epoch": 0.8465823765925495,
      "grad_norm": 0.10048877447843552,
      "learning_rate": 1.8911264343220184e-05,
      "loss": 0.3280665874481201,
      "memory(GiB)": 78.33,
      "step": 4369,
      "token_acc": 0.9011113631208891,
      "train_speed(iter/s)": 0.032404
    },
    {
      "epoch": 0.8467761468778763,
      "grad_norm": 0.10502710193395615,
      "learning_rate": 1.8864565271039274e-05,
      "loss": 0.3245999217033386,
      "memory(GiB)": 78.33,
      "step": 4370,
      "token_acc": 0.9021815622800844,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.846969917163203,
      "grad_norm": 0.1014811173081398,
      "learning_rate": 1.8817920059705194e-05,
      "loss": 0.3340781629085541,
      "memory(GiB)": 78.33,
      "step": 4371,
      "token_acc": 0.8999215070643642,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.8471636874485298,
      "grad_norm": 0.09874725341796875,
      "learning_rate": 1.8771328728376338e-05,
      "loss": 0.3380013704299927,
      "memory(GiB)": 78.33,
      "step": 4372,
      "token_acc": 0.8971284634760706,
      "train_speed(iter/s)": 0.032406
    },
    {
      "epoch": 0.8473574577338565,
      "grad_norm": 0.14172475039958954,
      "learning_rate": 1.8724791296189034e-05,
      "loss": 0.32453781366348267,
      "memory(GiB)": 78.33,
      "step": 4373,
      "token_acc": 0.9030038451530833,
      "train_speed(iter/s)": 0.032406
    },
    {
      "epoch": 0.8475512280191833,
      "grad_norm": 0.11031965911388397,
      "learning_rate": 1.867830778225744e-05,
      "loss": 0.3707287907600403,
      "memory(GiB)": 78.33,
      "step": 4374,
      "token_acc": 0.892270504861836,
      "train_speed(iter/s)": 0.032407
    },
    {
      "epoch": 0.84774499830451,
      "grad_norm": 0.10298004001379013,
      "learning_rate": 1.8631878205673552e-05,
      "loss": 0.34424299001693726,
      "memory(GiB)": 78.33,
      "step": 4375,
      "token_acc": 0.8972931413876296,
      "train_speed(iter/s)": 0.032408
    },
    {
      "epoch": 0.8479387685898367,
      "grad_norm": 0.11782942712306976,
      "learning_rate": 1.858550258550736e-05,
      "loss": 0.32887881994247437,
      "memory(GiB)": 78.33,
      "step": 4376,
      "token_acc": 0.9019160805238904,
      "train_speed(iter/s)": 0.032408
    },
    {
      "epoch": 0.8481325388751635,
      "grad_norm": 0.09481247514486313,
      "learning_rate": 1.8539180940806436e-05,
      "loss": 0.32029709219932556,
      "memory(GiB)": 78.33,
      "step": 4377,
      "token_acc": 0.9027707297269867,
      "train_speed(iter/s)": 0.032409
    },
    {
      "epoch": 0.8483263091604902,
      "grad_norm": 0.11217660456895828,
      "learning_rate": 1.8492913290596407e-05,
      "loss": 0.3478485345840454,
      "memory(GiB)": 78.33,
      "step": 4378,
      "token_acc": 0.8978001381079215,
      "train_speed(iter/s)": 0.032409
    },
    {
      "epoch": 0.848520079445817,
      "grad_norm": 0.10071595013141632,
      "learning_rate": 1.8446699653880638e-05,
      "loss": 0.32309237122535706,
      "memory(GiB)": 78.33,
      "step": 4379,
      "token_acc": 0.9025203747208446,
      "train_speed(iter/s)": 0.03241
    },
    {
      "epoch": 0.8487138497311437,
      "grad_norm": 0.10318909585475922,
      "learning_rate": 1.8400540049640278e-05,
      "loss": 0.3701336681842804,
      "memory(GiB)": 78.33,
      "step": 4380,
      "token_acc": 0.8895177546906498,
      "train_speed(iter/s)": 0.032411
    },
    {
      "epoch": 0.8489076200164705,
      "grad_norm": 0.1191054955124855,
      "learning_rate": 1.8354434496834346e-05,
      "loss": 0.37648195028305054,
      "memory(GiB)": 78.33,
      "step": 4381,
      "token_acc": 0.8879827766179541,
      "train_speed(iter/s)": 0.032411
    },
    {
      "epoch": 0.8491013903017972,
      "grad_norm": 0.09198344498872757,
      "learning_rate": 1.830838301439958e-05,
      "loss": 0.32460615038871765,
      "memory(GiB)": 78.33,
      "step": 4382,
      "token_acc": 0.9025374855824683,
      "train_speed(iter/s)": 0.032412
    },
    {
      "epoch": 0.849295160587124,
      "grad_norm": 0.09382152557373047,
      "learning_rate": 1.826238562125068e-05,
      "loss": 0.31965401768684387,
      "memory(GiB)": 78.33,
      "step": 4383,
      "token_acc": 0.9034426770175306,
      "train_speed(iter/s)": 0.032413
    },
    {
      "epoch": 0.8494889308724507,
      "grad_norm": 0.10211214423179626,
      "learning_rate": 1.821644233627985e-05,
      "loss": 0.35185420513153076,
      "memory(GiB)": 78.33,
      "step": 4384,
      "token_acc": 0.8940360240160107,
      "train_speed(iter/s)": 0.032413
    },
    {
      "epoch": 0.8496827011577774,
      "grad_norm": 0.09837459772825241,
      "learning_rate": 1.8170553178357366e-05,
      "loss": 0.3056308627128601,
      "memory(GiB)": 78.33,
      "step": 4385,
      "token_acc": 0.9083345412378605,
      "train_speed(iter/s)": 0.032414
    },
    {
      "epoch": 0.8498764714431042,
      "grad_norm": 0.09455987811088562,
      "learning_rate": 1.8124718166331066e-05,
      "loss": 0.2900204658508301,
      "memory(GiB)": 78.33,
      "step": 4386,
      "token_acc": 0.9121223003129328,
      "train_speed(iter/s)": 0.032414
    },
    {
      "epoch": 0.8500702417284309,
      "grad_norm": 0.09906947612762451,
      "learning_rate": 1.8078937319026654e-05,
      "loss": 0.33766764402389526,
      "memory(GiB)": 78.33,
      "step": 4387,
      "token_acc": 0.8973069206205668,
      "train_speed(iter/s)": 0.032415
    },
    {
      "epoch": 0.8502640120137577,
      "grad_norm": 0.09509405493736267,
      "learning_rate": 1.8033210655247527e-05,
      "loss": 0.3071034550666809,
      "memory(GiB)": 78.33,
      "step": 4388,
      "token_acc": 0.9057726952298896,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.8504577822990844,
      "grad_norm": 0.10163518041372299,
      "learning_rate": 1.7987538193774857e-05,
      "loss": 0.34781065583229065,
      "memory(GiB)": 78.33,
      "step": 4389,
      "token_acc": 0.8967510195262917,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.8506515525844112,
      "grad_norm": 0.09907959401607513,
      "learning_rate": 1.794191995336761e-05,
      "loss": 0.3162297308444977,
      "memory(GiB)": 78.33,
      "step": 4390,
      "token_acc": 0.9030727923627685,
      "train_speed(iter/s)": 0.032417
    },
    {
      "epoch": 0.8508453228697379,
      "grad_norm": 0.10068470239639282,
      "learning_rate": 1.7896355952762314e-05,
      "loss": 0.32280969619750977,
      "memory(GiB)": 78.33,
      "step": 4391,
      "token_acc": 0.9020573146084884,
      "train_speed(iter/s)": 0.032417
    },
    {
      "epoch": 0.8510390931550647,
      "grad_norm": 0.1039271429181099,
      "learning_rate": 1.785084621067343e-05,
      "loss": 0.3339514434337616,
      "memory(GiB)": 78.33,
      "step": 4392,
      "token_acc": 0.9003369695585127,
      "train_speed(iter/s)": 0.032418
    },
    {
      "epoch": 0.8512328634403914,
      "grad_norm": 0.1004716232419014,
      "learning_rate": 1.780539074579299e-05,
      "loss": 0.3353673815727234,
      "memory(GiB)": 78.33,
      "step": 4393,
      "token_acc": 0.8975802074107934,
      "train_speed(iter/s)": 0.032418
    },
    {
      "epoch": 0.8514266337257181,
      "grad_norm": 0.10031532496213913,
      "learning_rate": 1.7759989576790778e-05,
      "loss": 0.32534271478652954,
      "memory(GiB)": 78.33,
      "step": 4394,
      "token_acc": 0.9057053186123409,
      "train_speed(iter/s)": 0.032419
    },
    {
      "epoch": 0.8516204040110449,
      "grad_norm": 0.10088721662759781,
      "learning_rate": 1.7714642722314278e-05,
      "loss": 0.3248327970504761,
      "memory(GiB)": 78.33,
      "step": 4395,
      "token_acc": 0.9025890722491108,
      "train_speed(iter/s)": 0.03242
    },
    {
      "epoch": 0.8518141742963716,
      "grad_norm": 0.10467974096536636,
      "learning_rate": 1.766935020098862e-05,
      "loss": 0.33937883377075195,
      "memory(GiB)": 78.33,
      "step": 4396,
      "token_acc": 0.8982584532423584,
      "train_speed(iter/s)": 0.03242
    },
    {
      "epoch": 0.8520079445816984,
      "grad_norm": 0.0988173633813858,
      "learning_rate": 1.7624112031416725e-05,
      "loss": 0.33925455808639526,
      "memory(GiB)": 78.33,
      "step": 4397,
      "token_acc": 0.8979691440036875,
      "train_speed(iter/s)": 0.032421
    },
    {
      "epoch": 0.8522017148670251,
      "grad_norm": 0.10114793479442596,
      "learning_rate": 1.7578928232179102e-05,
      "loss": 0.29743602871894836,
      "memory(GiB)": 78.33,
      "step": 4398,
      "token_acc": 0.9106254871918932,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.8523954851523519,
      "grad_norm": 0.1041836142539978,
      "learning_rate": 1.753379882183395e-05,
      "loss": 0.3369101881980896,
      "memory(GiB)": 78.33,
      "step": 4399,
      "token_acc": 0.8994292404507537,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.8525892554376786,
      "grad_norm": 0.10185372084379196,
      "learning_rate": 1.748872381891713e-05,
      "loss": 0.3253825902938843,
      "memory(GiB)": 78.33,
      "step": 4400,
      "token_acc": 0.9039803494668677,
      "train_speed(iter/s)": 0.032423
    },
    {
      "epoch": 0.8527830257230054,
      "grad_norm": 0.09283772110939026,
      "learning_rate": 1.7443703241942143e-05,
      "loss": 0.2999337911605835,
      "memory(GiB)": 78.33,
      "step": 4401,
      "token_acc": 0.9094778884305121,
      "train_speed(iter/s)": 0.03242
    },
    {
      "epoch": 0.8529767960083321,
      "grad_norm": 0.10008923709392548,
      "learning_rate": 1.739873710940015e-05,
      "loss": 0.32751649618148804,
      "memory(GiB)": 78.33,
      "step": 4402,
      "token_acc": 0.9021663945992309,
      "train_speed(iter/s)": 0.032421
    },
    {
      "epoch": 0.8531705662936588,
      "grad_norm": 0.11449175328016281,
      "learning_rate": 1.7353825439759948e-05,
      "loss": 0.3371378481388092,
      "memory(GiB)": 78.33,
      "step": 4403,
      "token_acc": 0.8989161168708766,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.8533643365789856,
      "grad_norm": 0.10284436494112015,
      "learning_rate": 1.7308968251467997e-05,
      "loss": 0.34078723192214966,
      "memory(GiB)": 78.33,
      "step": 4404,
      "token_acc": 0.8978489916236173,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.8535581068643123,
      "grad_norm": 0.09477044641971588,
      "learning_rate": 1.726416556294834e-05,
      "loss": 0.31810081005096436,
      "memory(GiB)": 78.33,
      "step": 4405,
      "token_acc": 0.9051635577183023,
      "train_speed(iter/s)": 0.032423
    },
    {
      "epoch": 0.8537518771496391,
      "grad_norm": 0.09910459071397781,
      "learning_rate": 1.721941739260264e-05,
      "loss": 0.340251088142395,
      "memory(GiB)": 78.33,
      "step": 4406,
      "token_acc": 0.8982416457979913,
      "train_speed(iter/s)": 0.032423
    },
    {
      "epoch": 0.8539456474349658,
      "grad_norm": 0.09987306594848633,
      "learning_rate": 1.7174723758810166e-05,
      "loss": 0.3514068126678467,
      "memory(GiB)": 78.33,
      "step": 4407,
      "token_acc": 0.8941308390766874,
      "train_speed(iter/s)": 0.032424
    },
    {
      "epoch": 0.8541394177202926,
      "grad_norm": 0.10393253713846207,
      "learning_rate": 1.7130084679927763e-05,
      "loss": 0.34633970260620117,
      "memory(GiB)": 78.33,
      "step": 4408,
      "token_acc": 0.896145340536909,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.8543331880056193,
      "grad_norm": 0.10434923321008682,
      "learning_rate": 1.708550017429e-05,
      "loss": 0.34398695826530457,
      "memory(GiB)": 78.33,
      "step": 4409,
      "token_acc": 0.8962004803657725,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.854526958290946,
      "grad_norm": 0.09756119549274445,
      "learning_rate": 1.70409702602088e-05,
      "loss": 0.31827184557914734,
      "memory(GiB)": 78.33,
      "step": 4410,
      "token_acc": 0.9032197224841708,
      "train_speed(iter/s)": 0.032426
    },
    {
      "epoch": 0.8547207285762728,
      "grad_norm": 0.09933782368898392,
      "learning_rate": 1.699649495597389e-05,
      "loss": 0.30512192845344543,
      "memory(GiB)": 78.33,
      "step": 4411,
      "token_acc": 0.9053738951249932,
      "train_speed(iter/s)": 0.032427
    },
    {
      "epoch": 0.8549144988615995,
      "grad_norm": 0.0985947772860527,
      "learning_rate": 1.695207427985246e-05,
      "loss": 0.31603682041168213,
      "memory(GiB)": 78.33,
      "step": 4412,
      "token_acc": 0.9057937693987814,
      "train_speed(iter/s)": 0.032427
    },
    {
      "epoch": 0.8551082691469263,
      "grad_norm": 0.09728779643774033,
      "learning_rate": 1.690770825008924e-05,
      "loss": 0.3370038568973541,
      "memory(GiB)": 78.33,
      "step": 4413,
      "token_acc": 0.898977370525996,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.855302039432253,
      "grad_norm": 0.09609750658273697,
      "learning_rate": 1.6863396884906583e-05,
      "loss": 0.30580419301986694,
      "memory(GiB)": 78.33,
      "step": 4414,
      "token_acc": 0.9062045771938325,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.8554958097175798,
      "grad_norm": 0.08634244650602341,
      "learning_rate": 1.681914020250431e-05,
      "loss": 0.3033207356929779,
      "memory(GiB)": 78.33,
      "step": 4415,
      "token_acc": 0.9060820975714593,
      "train_speed(iter/s)": 0.032429
    },
    {
      "epoch": 0.8556895800029065,
      "grad_norm": 0.10316123068332672,
      "learning_rate": 1.677493822105992e-05,
      "loss": 0.3632691502571106,
      "memory(GiB)": 78.33,
      "step": 4416,
      "token_acc": 0.8925554382259767,
      "train_speed(iter/s)": 0.032429
    },
    {
      "epoch": 0.8558833502882333,
      "grad_norm": 0.08851286768913269,
      "learning_rate": 1.6730790958728253e-05,
      "loss": 0.2938583493232727,
      "memory(GiB)": 78.33,
      "step": 4417,
      "token_acc": 0.9093023255813953,
      "train_speed(iter/s)": 0.03243
    },
    {
      "epoch": 0.85607712057356,
      "grad_norm": 0.10702818632125854,
      "learning_rate": 1.6686698433641836e-05,
      "loss": 0.33791351318359375,
      "memory(GiB)": 78.33,
      "step": 4418,
      "token_acc": 0.8968233799237612,
      "train_speed(iter/s)": 0.032431
    },
    {
      "epoch": 0.8562708908588867,
      "grad_norm": 0.10163046419620514,
      "learning_rate": 1.6642660663910658e-05,
      "loss": 0.34928232431411743,
      "memory(GiB)": 78.33,
      "step": 4419,
      "token_acc": 0.8963705698818635,
      "train_speed(iter/s)": 0.032431
    },
    {
      "epoch": 0.8564646611442135,
      "grad_norm": 0.09223022311925888,
      "learning_rate": 1.6598677667622175e-05,
      "loss": 0.29846811294555664,
      "memory(GiB)": 78.33,
      "step": 4420,
      "token_acc": 0.9110857397454819,
      "train_speed(iter/s)": 0.032432
    },
    {
      "epoch": 0.8566584314295402,
      "grad_norm": 0.0997534915804863,
      "learning_rate": 1.655474946284142e-05,
      "loss": 0.3577570915222168,
      "memory(GiB)": 78.33,
      "step": 4421,
      "token_acc": 0.8929023263019287,
      "train_speed(iter/s)": 0.032432
    },
    {
      "epoch": 0.856852201714867,
      "grad_norm": 0.10170114785432816,
      "learning_rate": 1.6510876067610833e-05,
      "loss": 0.3271295130252838,
      "memory(GiB)": 78.33,
      "step": 4422,
      "token_acc": 0.9011473179963333,
      "train_speed(iter/s)": 0.032433
    },
    {
      "epoch": 0.8570459720001937,
      "grad_norm": 0.10372772812843323,
      "learning_rate": 1.6467057499950497e-05,
      "loss": 0.3079608380794525,
      "memory(GiB)": 78.33,
      "step": 4423,
      "token_acc": 0.9084609878310665,
      "train_speed(iter/s)": 0.032434
    },
    {
      "epoch": 0.8572397422855205,
      "grad_norm": 0.09529042989015579,
      "learning_rate": 1.6423293777857765e-05,
      "loss": 0.31907567381858826,
      "memory(GiB)": 78.33,
      "step": 4424,
      "token_acc": 0.9041177847722602,
      "train_speed(iter/s)": 0.032434
    },
    {
      "epoch": 0.8574335125708472,
      "grad_norm": 0.10136647522449493,
      "learning_rate": 1.6379584919307644e-05,
      "loss": 0.3494371175765991,
      "memory(GiB)": 78.33,
      "step": 4425,
      "token_acc": 0.8960502692998205,
      "train_speed(iter/s)": 0.032435
    },
    {
      "epoch": 0.857627282856174,
      "grad_norm": 0.091824971139431,
      "learning_rate": 1.6335930942252535e-05,
      "loss": 0.30003249645233154,
      "memory(GiB)": 78.33,
      "step": 4426,
      "token_acc": 0.9094136715634568,
      "train_speed(iter/s)": 0.032435
    },
    {
      "epoch": 0.8578210531415007,
      "grad_norm": 0.10738231241703033,
      "learning_rate": 1.6292331864622265e-05,
      "loss": 0.338476300239563,
      "memory(GiB)": 78.33,
      "step": 4427,
      "token_acc": 0.8967210902712185,
      "train_speed(iter/s)": 0.032436
    },
    {
      "epoch": 0.8580148234268274,
      "grad_norm": 0.09496748447418213,
      "learning_rate": 1.6248787704324163e-05,
      "loss": 0.3244837522506714,
      "memory(GiB)": 78.33,
      "step": 4428,
      "token_acc": 0.9002851103573711,
      "train_speed(iter/s)": 0.032437
    },
    {
      "epoch": 0.8582085937121542,
      "grad_norm": 0.10821764171123505,
      "learning_rate": 1.620529847924295e-05,
      "loss": 0.3445894420146942,
      "memory(GiB)": 78.33,
      "step": 4429,
      "token_acc": 0.8967734961898071,
      "train_speed(iter/s)": 0.032437
    },
    {
      "epoch": 0.8584023639974809,
      "grad_norm": 0.11172088235616684,
      "learning_rate": 1.616186420724089e-05,
      "loss": 0.3518655598163605,
      "memory(GiB)": 78.33,
      "step": 4430,
      "token_acc": 0.8950802436686354,
      "train_speed(iter/s)": 0.032438
    },
    {
      "epoch": 0.8585961342828077,
      "grad_norm": 0.09562534093856812,
      "learning_rate": 1.611848490615757e-05,
      "loss": 0.31965699791908264,
      "memory(GiB)": 78.33,
      "step": 4431,
      "token_acc": 0.9038684271484827,
      "train_speed(iter/s)": 0.032438
    },
    {
      "epoch": 0.8587899045681344,
      "grad_norm": 0.10791201889514923,
      "learning_rate": 1.6075160593810044e-05,
      "loss": 0.3649354577064514,
      "memory(GiB)": 78.33,
      "step": 4432,
      "token_acc": 0.8911983213293257,
      "train_speed(iter/s)": 0.032439
    },
    {
      "epoch": 0.8589836748534613,
      "grad_norm": 0.10779910534620285,
      "learning_rate": 1.6031891287992747e-05,
      "loss": 0.3215060234069824,
      "memory(GiB)": 78.33,
      "step": 4433,
      "token_acc": 0.9031464957595514,
      "train_speed(iter/s)": 0.03244
    },
    {
      "epoch": 0.859177445138788,
      "grad_norm": 0.09658445417881012,
      "learning_rate": 1.5988677006477568e-05,
      "loss": 0.33394524455070496,
      "memory(GiB)": 78.33,
      "step": 4434,
      "token_acc": 0.8990924985430022,
      "train_speed(iter/s)": 0.03244
    },
    {
      "epoch": 0.8593712154241148,
      "grad_norm": 0.10684984922409058,
      "learning_rate": 1.594551776701377e-05,
      "loss": 0.33571135997772217,
      "memory(GiB)": 78.33,
      "step": 4435,
      "token_acc": 0.9019298688193743,
      "train_speed(iter/s)": 0.032441
    },
    {
      "epoch": 0.8595649857094415,
      "grad_norm": 0.09724808484315872,
      "learning_rate": 1.5902413587327978e-05,
      "loss": 0.32527902722358704,
      "memory(GiB)": 78.33,
      "step": 4436,
      "token_acc": 0.9014134468456669,
      "train_speed(iter/s)": 0.032441
    },
    {
      "epoch": 0.8597587559947683,
      "grad_norm": 0.1140824407339096,
      "learning_rate": 1.5859364485124294e-05,
      "loss": 0.3552241027355194,
      "memory(GiB)": 78.33,
      "step": 4437,
      "token_acc": 0.8934269717466584,
      "train_speed(iter/s)": 0.032442
    },
    {
      "epoch": 0.859952526280095,
      "grad_norm": 0.11029309034347534,
      "learning_rate": 1.5816370478084106e-05,
      "loss": 0.3606140613555908,
      "memory(GiB)": 78.33,
      "step": 4438,
      "token_acc": 0.8945596964690044,
      "train_speed(iter/s)": 0.032443
    },
    {
      "epoch": 0.8601462965654217,
      "grad_norm": 0.10048742592334747,
      "learning_rate": 1.5773431583866226e-05,
      "loss": 0.3234768509864807,
      "memory(GiB)": 78.33,
      "step": 4439,
      "token_acc": 0.9028618152085037,
      "train_speed(iter/s)": 0.032443
    },
    {
      "epoch": 0.8603400668507485,
      "grad_norm": 0.10467264801263809,
      "learning_rate": 1.573054782010681e-05,
      "loss": 0.33086928725242615,
      "memory(GiB)": 78.33,
      "step": 4440,
      "token_acc": 0.9003659258462036,
      "train_speed(iter/s)": 0.032444
    },
    {
      "epoch": 0.8605338371360752,
      "grad_norm": 0.11301259696483612,
      "learning_rate": 1.568771920441932e-05,
      "loss": 0.36950597167015076,
      "memory(GiB)": 78.33,
      "step": 4441,
      "token_acc": 0.8924123686288804,
      "train_speed(iter/s)": 0.032445
    },
    {
      "epoch": 0.860727607421402,
      "grad_norm": 0.1058172732591629,
      "learning_rate": 1.5644945754394732e-05,
      "loss": 0.36208900809288025,
      "memory(GiB)": 78.33,
      "step": 4442,
      "token_acc": 0.8900826684333324,
      "train_speed(iter/s)": 0.032445
    },
    {
      "epoch": 0.8609213777067287,
      "grad_norm": 0.10789795219898224,
      "learning_rate": 1.5602227487601114e-05,
      "loss": 0.35568827390670776,
      "memory(GiB)": 78.33,
      "step": 4443,
      "token_acc": 0.8932357022646078,
      "train_speed(iter/s)": 0.032446
    },
    {
      "epoch": 0.8611151479920555,
      "grad_norm": 0.10026465356349945,
      "learning_rate": 1.5559564421584114e-05,
      "loss": 0.32759472727775574,
      "memory(GiB)": 78.33,
      "step": 4444,
      "token_acc": 0.8990959539379413,
      "train_speed(iter/s)": 0.032446
    },
    {
      "epoch": 0.8613089182773822,
      "grad_norm": 0.09213128685951233,
      "learning_rate": 1.5516956573866564e-05,
      "loss": 0.3294154107570648,
      "memory(GiB)": 78.33,
      "step": 4445,
      "token_acc": 0.9012261580381471,
      "train_speed(iter/s)": 0.032447
    },
    {
      "epoch": 0.861502688562709,
      "grad_norm": 0.09709835052490234,
      "learning_rate": 1.5474403961948627e-05,
      "loss": 0.31836092472076416,
      "memory(GiB)": 78.33,
      "step": 4446,
      "token_acc": 0.9042838018741634,
      "train_speed(iter/s)": 0.032447
    },
    {
      "epoch": 0.8616964588480357,
      "grad_norm": 0.10115919262170792,
      "learning_rate": 1.5431906603307846e-05,
      "loss": 0.3285466432571411,
      "memory(GiB)": 78.33,
      "step": 4447,
      "token_acc": 0.8986334681362831,
      "train_speed(iter/s)": 0.032448
    },
    {
      "epoch": 0.8618902291333624,
      "grad_norm": 0.10198438912630081,
      "learning_rate": 1.5389464515398976e-05,
      "loss": 0.326136976480484,
      "memory(GiB)": 78.33,
      "step": 4448,
      "token_acc": 0.9012800087897599,
      "train_speed(iter/s)": 0.032449
    },
    {
      "epoch": 0.8620839994186892,
      "grad_norm": 0.10833071917295456,
      "learning_rate": 1.5347077715654198e-05,
      "loss": 0.3257608115673065,
      "memory(GiB)": 78.33,
      "step": 4449,
      "token_acc": 0.9010049449672994,
      "train_speed(iter/s)": 0.032449
    },
    {
      "epoch": 0.8622777697040159,
      "grad_norm": 0.09669304639101028,
      "learning_rate": 1.5304746221482827e-05,
      "loss": 0.30072930455207825,
      "memory(GiB)": 78.33,
      "step": 4450,
      "token_acc": 0.9068726155150487,
      "train_speed(iter/s)": 0.03245
    },
    {
      "epoch": 0.8624715399893427,
      "grad_norm": 0.11190321296453476,
      "learning_rate": 1.52624700502716e-05,
      "loss": 0.3294280469417572,
      "memory(GiB)": 78.33,
      "step": 4451,
      "token_acc": 0.9005425479444106,
      "train_speed(iter/s)": 0.03245
    },
    {
      "epoch": 0.8626653102746694,
      "grad_norm": 0.09287480264902115,
      "learning_rate": 1.5220249219384484e-05,
      "loss": 0.3197273015975952,
      "memory(GiB)": 78.33,
      "step": 4452,
      "token_acc": 0.9034961140224298,
      "train_speed(iter/s)": 0.032451
    },
    {
      "epoch": 0.8628590805599962,
      "grad_norm": 0.09530606120824814,
      "learning_rate": 1.5178083746162666e-05,
      "loss": 0.3117508590221405,
      "memory(GiB)": 78.33,
      "step": 4453,
      "token_acc": 0.9065753927662811,
      "train_speed(iter/s)": 0.032452
    },
    {
      "epoch": 0.8630528508453229,
      "grad_norm": 0.10420526564121246,
      "learning_rate": 1.5135973647924665e-05,
      "loss": 0.30883246660232544,
      "memory(GiB)": 78.33,
      "step": 4454,
      "token_acc": 0.9062847265094456,
      "train_speed(iter/s)": 0.032452
    },
    {
      "epoch": 0.8632466211306496,
      "grad_norm": 0.08543951064348221,
      "learning_rate": 1.5093918941966193e-05,
      "loss": 0.27868953347206116,
      "memory(GiB)": 78.33,
      "step": 4455,
      "token_acc": 0.9159319449077825,
      "train_speed(iter/s)": 0.032453
    },
    {
      "epoch": 0.8634403914159764,
      "grad_norm": 0.09152313321828842,
      "learning_rate": 1.5051919645560334e-05,
      "loss": 0.31067970395088196,
      "memory(GiB)": 78.33,
      "step": 4456,
      "token_acc": 0.9047191907111177,
      "train_speed(iter/s)": 0.032453
    },
    {
      "epoch": 0.8636341617013031,
      "grad_norm": 0.11151348054409027,
      "learning_rate": 1.5009975775957207e-05,
      "loss": 0.3601279556751251,
      "memory(GiB)": 78.33,
      "step": 4457,
      "token_acc": 0.8918564063150479,
      "train_speed(iter/s)": 0.032454
    },
    {
      "epoch": 0.8638279319866299,
      "grad_norm": 0.10080854594707489,
      "learning_rate": 1.4968087350384395e-05,
      "loss": 0.34047406911849976,
      "memory(GiB)": 78.33,
      "step": 4458,
      "token_acc": 0.8975244985557326,
      "train_speed(iter/s)": 0.032455
    },
    {
      "epoch": 0.8640217022719566,
      "grad_norm": 0.0961776152253151,
      "learning_rate": 1.4926254386046554e-05,
      "loss": 0.3052217662334442,
      "memory(GiB)": 78.33,
      "step": 4459,
      "token_acc": 0.908130910695084,
      "train_speed(iter/s)": 0.032455
    },
    {
      "epoch": 0.8642154725572834,
      "grad_norm": 0.10806053876876831,
      "learning_rate": 1.4884476900125591e-05,
      "loss": 0.33401399850845337,
      "memory(GiB)": 78.33,
      "step": 4460,
      "token_acc": 0.9004724605246799,
      "train_speed(iter/s)": 0.032456
    },
    {
      "epoch": 0.8644092428426101,
      "grad_norm": 0.09109724313020706,
      "learning_rate": 1.484275490978068e-05,
      "loss": 0.32588818669319153,
      "memory(GiB)": 78.33,
      "step": 4461,
      "token_acc": 0.90133139220889,
      "train_speed(iter/s)": 0.032456
    },
    {
      "epoch": 0.8646030131279369,
      "grad_norm": 0.08890355378389359,
      "learning_rate": 1.4801088432148112e-05,
      "loss": 0.2869184613227844,
      "memory(GiB)": 78.33,
      "step": 4462,
      "token_acc": 0.911701588985605,
      "train_speed(iter/s)": 0.032457
    },
    {
      "epoch": 0.8647967834132636,
      "grad_norm": 0.09064006060361862,
      "learning_rate": 1.4759477484341513e-05,
      "loss": 0.29877418279647827,
      "memory(GiB)": 78.33,
      "step": 4463,
      "token_acc": 0.9077449822904369,
      "train_speed(iter/s)": 0.032458
    },
    {
      "epoch": 0.8649905536985903,
      "grad_norm": 0.10824266821146011,
      "learning_rate": 1.47179220834515e-05,
      "loss": 0.3239610493183136,
      "memory(GiB)": 78.33,
      "step": 4464,
      "token_acc": 0.9016067776803973,
      "train_speed(iter/s)": 0.032458
    },
    {
      "epoch": 0.8651843239839171,
      "grad_norm": 0.1040419191122055,
      "learning_rate": 1.46764222465461e-05,
      "loss": 0.35535627603530884,
      "memory(GiB)": 78.33,
      "step": 4465,
      "token_acc": 0.8935964513193813,
      "train_speed(iter/s)": 0.032459
    },
    {
      "epoch": 0.8653780942692438,
      "grad_norm": 0.13735035061836243,
      "learning_rate": 1.463497799067036e-05,
      "loss": 0.3214113116264343,
      "memory(GiB)": 78.33,
      "step": 4466,
      "token_acc": 0.9032424242424243,
      "train_speed(iter/s)": 0.032459
    },
    {
      "epoch": 0.8655718645545706,
      "grad_norm": 0.11260360479354858,
      "learning_rate": 1.4593589332846567e-05,
      "loss": 0.3748854696750641,
      "memory(GiB)": 78.33,
      "step": 4467,
      "token_acc": 0.8932340525328331,
      "train_speed(iter/s)": 0.03246
    },
    {
      "epoch": 0.8657656348398973,
      "grad_norm": 0.09724284708499908,
      "learning_rate": 1.4552256290074138e-05,
      "loss": 0.32235243916511536,
      "memory(GiB)": 78.33,
      "step": 4468,
      "token_acc": 0.9010934523343511,
      "train_speed(iter/s)": 0.032461
    },
    {
      "epoch": 0.8659594051252241,
      "grad_norm": 0.1042628064751625,
      "learning_rate": 1.451097887932966e-05,
      "loss": 0.3547298014163971,
      "memory(GiB)": 78.33,
      "step": 4469,
      "token_acc": 0.8934734091775081,
      "train_speed(iter/s)": 0.032461
    },
    {
      "epoch": 0.8661531754105508,
      "grad_norm": 0.11357161402702332,
      "learning_rate": 1.4469757117566888e-05,
      "loss": 0.3611631393432617,
      "memory(GiB)": 78.33,
      "step": 4470,
      "token_acc": 0.8912460223373058,
      "train_speed(iter/s)": 0.032462
    },
    {
      "epoch": 0.8663469456958776,
      "grad_norm": 0.11139743030071259,
      "learning_rate": 1.4428591021716729e-05,
      "loss": 0.35104861855506897,
      "memory(GiB)": 78.33,
      "step": 4471,
      "token_acc": 0.8951944796856144,
      "train_speed(iter/s)": 0.032462
    },
    {
      "epoch": 0.8665407159812043,
      "grad_norm": 0.11307670921087265,
      "learning_rate": 1.4387480608687174e-05,
      "loss": 0.36476895213127136,
      "memory(GiB)": 78.33,
      "step": 4472,
      "token_acc": 0.8911314038074601,
      "train_speed(iter/s)": 0.032463
    },
    {
      "epoch": 0.866734486266531,
      "grad_norm": 0.10000620037317276,
      "learning_rate": 1.4346425895363384e-05,
      "loss": 0.33012938499450684,
      "memory(GiB)": 78.33,
      "step": 4473,
      "token_acc": 0.9017330185642236,
      "train_speed(iter/s)": 0.032464
    },
    {
      "epoch": 0.8669282565518578,
      "grad_norm": 0.09574344009160995,
      "learning_rate": 1.4305426898607602e-05,
      "loss": 0.31818804144859314,
      "memory(GiB)": 78.33,
      "step": 4474,
      "token_acc": 0.9036768787502313,
      "train_speed(iter/s)": 0.032464
    },
    {
      "epoch": 0.8671220268371845,
      "grad_norm": 0.09933432936668396,
      "learning_rate": 1.426448363525931e-05,
      "loss": 0.3508920669555664,
      "memory(GiB)": 78.33,
      "step": 4475,
      "token_acc": 0.8937370802838147,
      "train_speed(iter/s)": 0.032465
    },
    {
      "epoch": 0.8673157971225113,
      "grad_norm": 0.10301670432090759,
      "learning_rate": 1.4223596122134873e-05,
      "loss": 0.3236181437969208,
      "memory(GiB)": 78.33,
      "step": 4476,
      "token_acc": 0.9032188377619357,
      "train_speed(iter/s)": 0.032465
    },
    {
      "epoch": 0.867509567407838,
      "grad_norm": 0.11474580317735672,
      "learning_rate": 1.4182764376028006e-05,
      "loss": 0.3661247491836548,
      "memory(GiB)": 78.33,
      "step": 4477,
      "token_acc": 0.8923494649417026,
      "train_speed(iter/s)": 0.032466
    },
    {
      "epoch": 0.8677033376931648,
      "grad_norm": 0.10032869875431061,
      "learning_rate": 1.414198841370936e-05,
      "loss": 0.3041604459285736,
      "memory(GiB)": 78.33,
      "step": 4478,
      "token_acc": 0.9077770323157777,
      "train_speed(iter/s)": 0.032467
    },
    {
      "epoch": 0.8678971079784915,
      "grad_norm": 0.10304979979991913,
      "learning_rate": 1.4101268251926707e-05,
      "loss": 0.3436087667942047,
      "memory(GiB)": 78.33,
      "step": 4479,
      "token_acc": 0.8969530197494768,
      "train_speed(iter/s)": 0.032467
    },
    {
      "epoch": 0.8680908782638183,
      "grad_norm": 0.10889364033937454,
      "learning_rate": 1.4060603907404933e-05,
      "loss": 0.35096222162246704,
      "memory(GiB)": 78.33,
      "step": 4480,
      "token_acc": 0.8916971772764852,
      "train_speed(iter/s)": 0.032468
    },
    {
      "epoch": 0.868284648549145,
      "grad_norm": 0.09136100858449936,
      "learning_rate": 1.401999539684593e-05,
      "loss": 0.3105822205543518,
      "memory(GiB)": 78.33,
      "step": 4481,
      "token_acc": 0.9048854192898514,
      "train_speed(iter/s)": 0.032468
    },
    {
      "epoch": 0.8684784188344717,
      "grad_norm": 0.09702181816101074,
      "learning_rate": 1.3979442736928803e-05,
      "loss": 0.31895220279693604,
      "memory(GiB)": 78.33,
      "step": 4482,
      "token_acc": 0.9040166789823709,
      "train_speed(iter/s)": 0.032469
    },
    {
      "epoch": 0.8686721891197985,
      "grad_norm": 0.09936001151800156,
      "learning_rate": 1.3938945944309499e-05,
      "loss": 0.3287636339664459,
      "memory(GiB)": 78.33,
      "step": 4483,
      "token_acc": 0.9009811937857727,
      "train_speed(iter/s)": 0.032469
    },
    {
      "epoch": 0.8688659594051252,
      "grad_norm": 0.09482026845216751,
      "learning_rate": 1.3898505035621226e-05,
      "loss": 0.30222681164741516,
      "memory(GiB)": 78.33,
      "step": 4484,
      "token_acc": 0.9081456725285801,
      "train_speed(iter/s)": 0.03247
    },
    {
      "epoch": 0.869059729690452,
      "grad_norm": 0.09608285129070282,
      "learning_rate": 1.3858120027474134e-05,
      "loss": 0.2800081968307495,
      "memory(GiB)": 78.33,
      "step": 4485,
      "token_acc": 0.9137849484969519,
      "train_speed(iter/s)": 0.032471
    },
    {
      "epoch": 0.8692534999757787,
      "grad_norm": 0.09473740309476852,
      "learning_rate": 1.3817790936455402e-05,
      "loss": 0.3255942165851593,
      "memory(GiB)": 78.33,
      "step": 4486,
      "token_acc": 0.9004664035672185,
      "train_speed(iter/s)": 0.032471
    },
    {
      "epoch": 0.8694472702611055,
      "grad_norm": 0.09641645103693008,
      "learning_rate": 1.3777517779129316e-05,
      "loss": 0.31738942861557007,
      "memory(GiB)": 78.33,
      "step": 4487,
      "token_acc": 0.9032964135021097,
      "train_speed(iter/s)": 0.032472
    },
    {
      "epoch": 0.8696410405464322,
      "grad_norm": 0.12395960092544556,
      "learning_rate": 1.3737300572037075e-05,
      "loss": 0.30116546154022217,
      "memory(GiB)": 78.33,
      "step": 4488,
      "token_acc": 0.9070750061682704,
      "train_speed(iter/s)": 0.032472
    },
    {
      "epoch": 0.869834810831759,
      "grad_norm": 0.09995721280574799,
      "learning_rate": 1.3697139331697065e-05,
      "loss": 0.34251344203948975,
      "memory(GiB)": 78.33,
      "step": 4489,
      "token_acc": 0.896267974399468,
      "train_speed(iter/s)": 0.032473
    },
    {
      "epoch": 0.8700285811170857,
      "grad_norm": 0.10301598161458969,
      "learning_rate": 1.3657034074604478e-05,
      "loss": 0.3315487205982208,
      "memory(GiB)": 78.33,
      "step": 4490,
      "token_acc": 0.9004393623612911,
      "train_speed(iter/s)": 0.032473
    },
    {
      "epoch": 0.8702223514024124,
      "grad_norm": 0.099857397377491,
      "learning_rate": 1.3616984817231685e-05,
      "loss": 0.3251858353614807,
      "memory(GiB)": 78.33,
      "step": 4491,
      "token_acc": 0.9022668466644167,
      "train_speed(iter/s)": 0.032474
    },
    {
      "epoch": 0.8704161216877392,
      "grad_norm": 0.09973510354757309,
      "learning_rate": 1.3576991576028013e-05,
      "loss": 0.3382074534893036,
      "memory(GiB)": 78.33,
      "step": 4492,
      "token_acc": 0.9003876364378252,
      "train_speed(iter/s)": 0.032475
    },
    {
      "epoch": 0.8706098919730659,
      "grad_norm": 0.08957747370004654,
      "learning_rate": 1.3537054367419703e-05,
      "loss": 0.29314741492271423,
      "memory(GiB)": 78.33,
      "step": 4493,
      "token_acc": 0.9126814476038974,
      "train_speed(iter/s)": 0.032475
    },
    {
      "epoch": 0.8708036622583927,
      "grad_norm": 0.10090366750955582,
      "learning_rate": 1.3497173207810068e-05,
      "loss": 0.32285887002944946,
      "memory(GiB)": 78.33,
      "step": 4494,
      "token_acc": 0.9043573045641298,
      "train_speed(iter/s)": 0.032476
    },
    {
      "epoch": 0.8709974325437194,
      "grad_norm": 0.09529086202383041,
      "learning_rate": 1.3457348113579358e-05,
      "loss": 0.30941274762153625,
      "memory(GiB)": 78.33,
      "step": 4495,
      "token_acc": 0.9068646306126112,
      "train_speed(iter/s)": 0.032476
    },
    {
      "epoch": 0.8711912028290462,
      "grad_norm": 0.09300097078084946,
      "learning_rate": 1.3417579101084869e-05,
      "loss": 0.3176164925098419,
      "memory(GiB)": 78.33,
      "step": 4496,
      "token_acc": 0.9025844421699079,
      "train_speed(iter/s)": 0.032477
    },
    {
      "epoch": 0.8713849731143729,
      "grad_norm": 0.10441552102565765,
      "learning_rate": 1.3377866186660701e-05,
      "loss": 0.33653199672698975,
      "memory(GiB)": 78.33,
      "step": 4497,
      "token_acc": 0.8984428473648186,
      "train_speed(iter/s)": 0.032477
    },
    {
      "epoch": 0.8715787433996997,
      "grad_norm": 0.10322162508964539,
      "learning_rate": 1.3338209386618092e-05,
      "loss": 0.32316023111343384,
      "memory(GiB)": 78.33,
      "step": 4498,
      "token_acc": 0.9009378129837021,
      "train_speed(iter/s)": 0.032478
    },
    {
      "epoch": 0.8717725136850264,
      "grad_norm": 0.09797031432390213,
      "learning_rate": 1.329860871724513e-05,
      "loss": 0.3057693541049957,
      "memory(GiB)": 78.33,
      "step": 4499,
      "token_acc": 0.9071016563475685,
      "train_speed(iter/s)": 0.032479
    },
    {
      "epoch": 0.8719662839703531,
      "grad_norm": 0.10819875448942184,
      "learning_rate": 1.3259064194806885e-05,
      "loss": 0.3662481904029846,
      "memory(GiB)": 78.33,
      "step": 4500,
      "token_acc": 0.8935516888433982,
      "train_speed(iter/s)": 0.032479
    },
    {
      "epoch": 0.8719662839703531,
      "eval_loss": 0.3799753785133362,
      "eval_runtime": 1344.7448,
      "eval_samples_per_second": 5.019,
      "eval_steps_per_second": 5.019,
      "eval_token_acc": 0.9022423587138677,
      "step": 4500
    },
    {
      "epoch": 0.8721600542556799,
      "grad_norm": 0.09284574538469315,
      "learning_rate": 1.3219575835545332e-05,
      "loss": 0.30319368839263916,
      "memory(GiB)": 78.33,
      "step": 4501,
      "token_acc": 0.9082558951495945,
      "train_speed(iter/s)": 0.032168
    },
    {
      "epoch": 0.8723538245410066,
      "grad_norm": 0.098308265209198,
      "learning_rate": 1.3180143655679397e-05,
      "loss": 0.3219000995159149,
      "memory(GiB)": 78.33,
      "step": 4502,
      "token_acc": 0.9041865907009868,
      "train_speed(iter/s)": 0.032168
    },
    {
      "epoch": 0.8725475948263334,
      "grad_norm": 0.10174136608839035,
      "learning_rate": 1.3140767671404995e-05,
      "loss": 0.33587872982025146,
      "memory(GiB)": 78.33,
      "step": 4503,
      "token_acc": 0.8989667675363268,
      "train_speed(iter/s)": 0.032169
    },
    {
      "epoch": 0.8727413651116601,
      "grad_norm": 0.09019061923027039,
      "learning_rate": 1.3101447898894852e-05,
      "loss": 0.2792420983314514,
      "memory(GiB)": 78.33,
      "step": 4504,
      "token_acc": 0.9144691759092409,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.8729351353969869,
      "grad_norm": 0.1039118766784668,
      "learning_rate": 1.306218435429865e-05,
      "loss": 0.35971078276634216,
      "memory(GiB)": 78.33,
      "step": 4505,
      "token_acc": 0.8914640875954808,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.8731289056823136,
      "grad_norm": 0.10616055876016617,
      "learning_rate": 1.3022977053743005e-05,
      "loss": 0.3402102291584015,
      "memory(GiB)": 78.33,
      "step": 4506,
      "token_acc": 0.8962266862085811,
      "train_speed(iter/s)": 0.032171
    },
    {
      "epoch": 0.8733226759676403,
      "grad_norm": 0.09928639978170395,
      "learning_rate": 1.298382601333139e-05,
      "loss": 0.3082139194011688,
      "memory(GiB)": 78.33,
      "step": 4507,
      "token_acc": 0.9083105179101072,
      "train_speed(iter/s)": 0.032171
    },
    {
      "epoch": 0.8735164462529671,
      "grad_norm": 0.09501931816339493,
      "learning_rate": 1.294473124914422e-05,
      "loss": 0.3152191638946533,
      "memory(GiB)": 78.33,
      "step": 4508,
      "token_acc": 0.9062588283637671,
      "train_speed(iter/s)": 0.032172
    },
    {
      "epoch": 0.8737102165382938,
      "grad_norm": 0.09296244382858276,
      "learning_rate": 1.2905692777238719e-05,
      "loss": 0.27433520555496216,
      "memory(GiB)": 78.33,
      "step": 4509,
      "token_acc": 0.9149881928045562,
      "train_speed(iter/s)": 0.032173
    },
    {
      "epoch": 0.8739039868236206,
      "grad_norm": 0.09587821364402771,
      "learning_rate": 1.2866710613649062e-05,
      "loss": 0.2978772521018982,
      "memory(GiB)": 78.33,
      "step": 4510,
      "token_acc": 0.9090977199510526,
      "train_speed(iter/s)": 0.032173
    },
    {
      "epoch": 0.8740977571089473,
      "grad_norm": 0.10060916841030121,
      "learning_rate": 1.282778477438629e-05,
      "loss": 0.321035236120224,
      "memory(GiB)": 78.33,
      "step": 4511,
      "token_acc": 0.9027070865260921,
      "train_speed(iter/s)": 0.032174
    },
    {
      "epoch": 0.8742915273942741,
      "grad_norm": 0.097939632833004,
      "learning_rate": 1.2788915275438267e-05,
      "loss": 0.2997584044933319,
      "memory(GiB)": 78.33,
      "step": 4512,
      "token_acc": 0.9089030803906837,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.8744852976796008,
      "grad_norm": 0.08998732268810272,
      "learning_rate": 1.2750102132769735e-05,
      "loss": 0.2946266531944275,
      "memory(GiB)": 78.33,
      "step": 4513,
      "token_acc": 0.9110179997024842,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.8746790679649276,
      "grad_norm": 0.10302318632602692,
      "learning_rate": 1.2711345362322295e-05,
      "loss": 0.3208523988723755,
      "memory(GiB)": 78.33,
      "step": 4514,
      "token_acc": 0.9040156111839498,
      "train_speed(iter/s)": 0.032176
    },
    {
      "epoch": 0.8748728382502543,
      "grad_norm": 0.10049934685230255,
      "learning_rate": 1.2672644980014445e-05,
      "loss": 0.35380083322525024,
      "memory(GiB)": 78.33,
      "step": 4515,
      "token_acc": 0.8943812514845214,
      "train_speed(iter/s)": 0.032177
    },
    {
      "epoch": 0.875066608535581,
      "grad_norm": 0.09158279001712799,
      "learning_rate": 1.2634001001741373e-05,
      "loss": 0.3017105460166931,
      "memory(GiB)": 78.33,
      "step": 4516,
      "token_acc": 0.908971506687753,
      "train_speed(iter/s)": 0.032177
    },
    {
      "epoch": 0.8752603788209078,
      "grad_norm": 0.10271737724542618,
      "learning_rate": 1.2595413443375297e-05,
      "loss": 0.3498397767543793,
      "memory(GiB)": 78.33,
      "step": 4517,
      "token_acc": 0.896111417539989,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.8754541491062345,
      "grad_norm": 0.09800441563129425,
      "learning_rate": 1.2556882320765122e-05,
      "loss": 0.3219001591205597,
      "memory(GiB)": 78.33,
      "step": 4518,
      "token_acc": 0.9034795042897998,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.8756479193915613,
      "grad_norm": 0.10516981035470963,
      "learning_rate": 1.2518407649736607e-05,
      "loss": 0.3301500082015991,
      "memory(GiB)": 78.33,
      "step": 4519,
      "token_acc": 0.8994242640075973,
      "train_speed(iter/s)": 0.032179
    },
    {
      "epoch": 0.875841689676888,
      "grad_norm": 0.10065259039402008,
      "learning_rate": 1.2479989446092359e-05,
      "loss": 0.3331465423107147,
      "memory(GiB)": 78.33,
      "step": 4520,
      "token_acc": 0.8990757119011435,
      "train_speed(iter/s)": 0.03218
    },
    {
      "epoch": 0.8760354599622148,
      "grad_norm": 0.11081155389547348,
      "learning_rate": 1.2441627725611708e-05,
      "loss": 0.34419184923171997,
      "memory(GiB)": 78.33,
      "step": 4521,
      "token_acc": 0.8977344241661422,
      "train_speed(iter/s)": 0.03218
    },
    {
      "epoch": 0.8762292302475415,
      "grad_norm": 0.0998803898692131,
      "learning_rate": 1.240332250405095e-05,
      "loss": 0.31978899240493774,
      "memory(GiB)": 78.33,
      "step": 4522,
      "token_acc": 0.9027906976744186,
      "train_speed(iter/s)": 0.032181
    },
    {
      "epoch": 0.8764230005328683,
      "grad_norm": 0.10122760385274887,
      "learning_rate": 1.236507379714295e-05,
      "loss": 0.31909558176994324,
      "memory(GiB)": 78.33,
      "step": 4523,
      "token_acc": 0.9035904628330996,
      "train_speed(iter/s)": 0.032182
    },
    {
      "epoch": 0.876616770818195,
      "grad_norm": 0.10056735575199127,
      "learning_rate": 1.2326881620597556e-05,
      "loss": 0.3240450620651245,
      "memory(GiB)": 78.33,
      "step": 4524,
      "token_acc": 0.9028395198933096,
      "train_speed(iter/s)": 0.032182
    },
    {
      "epoch": 0.8768105411035217,
      "grad_norm": 0.08745722472667694,
      "learning_rate": 1.2288745990101323e-05,
      "loss": 0.3004869520664215,
      "memory(GiB)": 78.33,
      "step": 4525,
      "token_acc": 0.907592263761578,
      "train_speed(iter/s)": 0.032183
    },
    {
      "epoch": 0.8770043113888485,
      "grad_norm": 0.10847844928503036,
      "learning_rate": 1.2250666921317537e-05,
      "loss": 0.34397241473197937,
      "memory(GiB)": 78.33,
      "step": 4526,
      "token_acc": 0.8972039724980901,
      "train_speed(iter/s)": 0.032184
    },
    {
      "epoch": 0.8771980816741752,
      "grad_norm": 0.09692453593015671,
      "learning_rate": 1.221264442988632e-05,
      "loss": 0.33984488248825073,
      "memory(GiB)": 78.33,
      "step": 4527,
      "token_acc": 0.9006118417883123,
      "train_speed(iter/s)": 0.032184
    },
    {
      "epoch": 0.877391851959502,
      "grad_norm": 0.09908100217580795,
      "learning_rate": 1.2174678531424497e-05,
      "loss": 0.31903359293937683,
      "memory(GiB)": 78.33,
      "step": 4528,
      "token_acc": 0.9019696566409369,
      "train_speed(iter/s)": 0.032185
    },
    {
      "epoch": 0.8775856222448287,
      "grad_norm": 0.10452962666749954,
      "learning_rate": 1.2136769241525762e-05,
      "loss": 0.33924567699432373,
      "memory(GiB)": 78.33,
      "step": 4529,
      "token_acc": 0.897684942468475,
      "train_speed(iter/s)": 0.032185
    },
    {
      "epoch": 0.8777793925301555,
      "grad_norm": 0.09859445691108704,
      "learning_rate": 1.2098916575760376e-05,
      "loss": 0.3426741659641266,
      "memory(GiB)": 78.33,
      "step": 4530,
      "token_acc": 0.896549017555794,
      "train_speed(iter/s)": 0.032186
    },
    {
      "epoch": 0.8779731628154822,
      "grad_norm": 0.1066657230257988,
      "learning_rate": 1.2061120549675518e-05,
      "loss": 0.3669888377189636,
      "memory(GiB)": 78.33,
      "step": 4531,
      "token_acc": 0.8908226988867771,
      "train_speed(iter/s)": 0.032187
    },
    {
      "epoch": 0.878166933100809,
      "grad_norm": 0.1010642871260643,
      "learning_rate": 1.2023381178795022e-05,
      "loss": 0.31678637862205505,
      "memory(GiB)": 78.33,
      "step": 4532,
      "token_acc": 0.9021789270887037,
      "train_speed(iter/s)": 0.032187
    },
    {
      "epoch": 0.8783607033861357,
      "grad_norm": 0.10661202669143677,
      "learning_rate": 1.1985698478619454e-05,
      "loss": 0.3384348452091217,
      "memory(GiB)": 78.33,
      "step": 4533,
      "token_acc": 0.899502755811167,
      "train_speed(iter/s)": 0.032188
    },
    {
      "epoch": 0.8785544736714624,
      "grad_norm": 0.09901930391788483,
      "learning_rate": 1.1948072464626101e-05,
      "loss": 0.3115423619747162,
      "memory(GiB)": 78.33,
      "step": 4534,
      "token_acc": 0.9044517138387143,
      "train_speed(iter/s)": 0.032189
    },
    {
      "epoch": 0.8787482439567892,
      "grad_norm": 0.09338109940290451,
      "learning_rate": 1.1910503152268952e-05,
      "loss": 0.2875381410121918,
      "memory(GiB)": 78.33,
      "step": 4535,
      "token_acc": 0.9132529457108248,
      "train_speed(iter/s)": 0.032189
    },
    {
      "epoch": 0.8789420142421159,
      "grad_norm": 0.10732939094305038,
      "learning_rate": 1.187299055697883e-05,
      "loss": 0.3420470058917999,
      "memory(GiB)": 78.33,
      "step": 4536,
      "token_acc": 0.8977093546629971,
      "train_speed(iter/s)": 0.03219
    },
    {
      "epoch": 0.8791357845274427,
      "grad_norm": 0.10017619282007217,
      "learning_rate": 1.1835534694163057e-05,
      "loss": 0.3340654969215393,
      "memory(GiB)": 78.33,
      "step": 4537,
      "token_acc": 0.8995184135977337,
      "train_speed(iter/s)": 0.03219
    },
    {
      "epoch": 0.8793295548127694,
      "grad_norm": 0.09547953307628632,
      "learning_rate": 1.1798135579205831e-05,
      "loss": 0.3285997807979584,
      "memory(GiB)": 78.33,
      "step": 4538,
      "token_acc": 0.9024012202922796,
      "train_speed(iter/s)": 0.032191
    },
    {
      "epoch": 0.8795233250980962,
      "grad_norm": 0.09887305647134781,
      "learning_rate": 1.1760793227467947e-05,
      "loss": 0.34439530968666077,
      "memory(GiB)": 78.33,
      "step": 4539,
      "token_acc": 0.8960512218764348,
      "train_speed(iter/s)": 0.032192
    },
    {
      "epoch": 0.8797170953834229,
      "grad_norm": 0.10936824232339859,
      "learning_rate": 1.1723507654286885e-05,
      "loss": 0.3436868488788605,
      "memory(GiB)": 78.33,
      "step": 4540,
      "token_acc": 0.8984078161751402,
      "train_speed(iter/s)": 0.032192
    },
    {
      "epoch": 0.8799108656687497,
      "grad_norm": 0.09683515876531601,
      "learning_rate": 1.1686278874976912e-05,
      "loss": 0.32740306854248047,
      "memory(GiB)": 78.33,
      "step": 4541,
      "token_acc": 0.9006859993763642,
      "train_speed(iter/s)": 0.032193
    },
    {
      "epoch": 0.8801046359540764,
      "grad_norm": 0.10903044044971466,
      "learning_rate": 1.1649106904828798e-05,
      "loss": 0.3218260705471039,
      "memory(GiB)": 78.33,
      "step": 4542,
      "token_acc": 0.9046106833637916,
      "train_speed(iter/s)": 0.032194
    },
    {
      "epoch": 0.8802984062394031,
      "grad_norm": 0.09545977413654327,
      "learning_rate": 1.1611991759110128e-05,
      "loss": 0.31498417258262634,
      "memory(GiB)": 78.33,
      "step": 4543,
      "token_acc": 0.904716222873292,
      "train_speed(iter/s)": 0.032194
    },
    {
      "epoch": 0.8804921765247299,
      "grad_norm": 0.09254451096057892,
      "learning_rate": 1.1574933453065078e-05,
      "loss": 0.2897379696369171,
      "memory(GiB)": 78.33,
      "step": 4544,
      "token_acc": 0.9108022299442514,
      "train_speed(iter/s)": 0.032195
    },
    {
      "epoch": 0.8806859468100566,
      "grad_norm": 0.10882263630628586,
      "learning_rate": 1.1537932001914485e-05,
      "loss": 0.3277333974838257,
      "memory(GiB)": 78.33,
      "step": 4545,
      "token_acc": 0.9019514213860503,
      "train_speed(iter/s)": 0.032196
    },
    {
      "epoch": 0.8808797170953834,
      "grad_norm": 0.0869324579834938,
      "learning_rate": 1.1500987420855845e-05,
      "loss": 0.2924439609050751,
      "memory(GiB)": 78.33,
      "step": 4546,
      "token_acc": 0.9105908898603864,
      "train_speed(iter/s)": 0.032196
    },
    {
      "epoch": 0.8810734873807101,
      "grad_norm": 0.09957166761159897,
      "learning_rate": 1.1464099725063237e-05,
      "loss": 0.3209323287010193,
      "memory(GiB)": 78.33,
      "step": 4547,
      "token_acc": 0.9023074369189907,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.8812672576660369,
      "grad_norm": 0.11669695377349854,
      "learning_rate": 1.1427268929687555e-05,
      "loss": 0.3712838888168335,
      "memory(GiB)": 78.33,
      "step": 4548,
      "token_acc": 0.8882153971045866,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.8814610279513636,
      "grad_norm": 0.0909273773431778,
      "learning_rate": 1.1390495049856036e-05,
      "loss": 0.31092870235443115,
      "memory(GiB)": 78.33,
      "step": 4549,
      "token_acc": 0.9068354688855991,
      "train_speed(iter/s)": 0.032198
    },
    {
      "epoch": 0.8816547982366904,
      "grad_norm": 0.08977729827165604,
      "learning_rate": 1.1353778100672828e-05,
      "loss": 0.29384180903434753,
      "memory(GiB)": 78.33,
      "step": 4550,
      "token_acc": 0.9103132495094121,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.8818485685220171,
      "grad_norm": 0.09765194356441498,
      "learning_rate": 1.131711809721852e-05,
      "loss": 0.31177225708961487,
      "memory(GiB)": 78.33,
      "step": 4551,
      "token_acc": 0.9054578904333606,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.8820423388073438,
      "grad_norm": 0.11460500210523605,
      "learning_rate": 1.1280515054550366e-05,
      "loss": 0.3225688934326172,
      "memory(GiB)": 78.33,
      "step": 4552,
      "token_acc": 0.9011132164850781,
      "train_speed(iter/s)": 0.0322
    },
    {
      "epoch": 0.8822361090926706,
      "grad_norm": 0.09452386945486069,
      "learning_rate": 1.1243968987702206e-05,
      "loss": 0.29615986347198486,
      "memory(GiB)": 78.33,
      "step": 4553,
      "token_acc": 0.9087365694797036,
      "train_speed(iter/s)": 0.032201
    },
    {
      "epoch": 0.8824298793779974,
      "grad_norm": 0.09222650527954102,
      "learning_rate": 1.1207479911684487e-05,
      "loss": 0.2946526110172272,
      "memory(GiB)": 78.33,
      "step": 4554,
      "token_acc": 0.9109519196451333,
      "train_speed(iter/s)": 0.032201
    },
    {
      "epoch": 0.8826236496633242,
      "grad_norm": 0.10203875601291656,
      "learning_rate": 1.11710478414843e-05,
      "loss": 0.34442439675331116,
      "memory(GiB)": 78.33,
      "step": 4555,
      "token_acc": 0.8982755761896866,
      "train_speed(iter/s)": 0.032202
    },
    {
      "epoch": 0.8828174199486509,
      "grad_norm": 0.10598360747098923,
      "learning_rate": 1.1134672792065209e-05,
      "loss": 0.329555869102478,
      "memory(GiB)": 78.33,
      "step": 4556,
      "token_acc": 0.9016222021252543,
      "train_speed(iter/s)": 0.032202
    },
    {
      "epoch": 0.8830111902339777,
      "grad_norm": 0.10227588564157486,
      "learning_rate": 1.109835477836748e-05,
      "loss": 0.33669596910476685,
      "memory(GiB)": 78.33,
      "step": 4557,
      "token_acc": 0.8991185669604776,
      "train_speed(iter/s)": 0.032203
    },
    {
      "epoch": 0.8832049605193044,
      "grad_norm": 0.09594012796878815,
      "learning_rate": 1.1062093815307865e-05,
      "loss": 0.3027840554714203,
      "memory(GiB)": 78.33,
      "step": 4558,
      "token_acc": 0.9077541203159593,
      "train_speed(iter/s)": 0.032204
    },
    {
      "epoch": 0.8833987308046312,
      "grad_norm": 0.11179719865322113,
      "learning_rate": 1.1025889917779735e-05,
      "loss": 0.3367825746536255,
      "memory(GiB)": 78.33,
      "step": 4559,
      "token_acc": 0.9006670902160102,
      "train_speed(iter/s)": 0.032204
    },
    {
      "epoch": 0.8835925010899579,
      "grad_norm": 0.09386321157217026,
      "learning_rate": 1.0989743100653008e-05,
      "loss": 0.3008558750152588,
      "memory(GiB)": 78.33,
      "step": 4560,
      "token_acc": 0.908220500378284,
      "train_speed(iter/s)": 0.032205
    },
    {
      "epoch": 0.8837862713752846,
      "grad_norm": 0.08674740046262741,
      "learning_rate": 1.0953653378774097e-05,
      "loss": 0.27751424908638,
      "memory(GiB)": 78.33,
      "step": 4561,
      "token_acc": 0.9156325156325157,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.8839800416606114,
      "grad_norm": 0.10045383870601654,
      "learning_rate": 1.0917620766966123e-05,
      "loss": 0.30626630783081055,
      "memory(GiB)": 78.33,
      "step": 4562,
      "token_acc": 0.9081538638719336,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.8841738119459381,
      "grad_norm": 0.10451143980026245,
      "learning_rate": 1.0881645280028534e-05,
      "loss": 0.3387284576892853,
      "memory(GiB)": 78.33,
      "step": 4563,
      "token_acc": 0.8987345454545455,
      "train_speed(iter/s)": 0.032207
    },
    {
      "epoch": 0.8843675822312649,
      "grad_norm": 0.09835859388113022,
      "learning_rate": 1.0845726932737509e-05,
      "loss": 0.33612358570098877,
      "memory(GiB)": 78.33,
      "step": 4564,
      "token_acc": 0.8987608426270136,
      "train_speed(iter/s)": 0.032207
    },
    {
      "epoch": 0.8845613525165916,
      "grad_norm": 0.10052934288978577,
      "learning_rate": 1.0809865739845646e-05,
      "loss": 0.31725549697875977,
      "memory(GiB)": 78.33,
      "step": 4565,
      "token_acc": 0.9065557583555249,
      "train_speed(iter/s)": 0.032208
    },
    {
      "epoch": 0.8847551228019184,
      "grad_norm": 0.10602091997861862,
      "learning_rate": 1.0774061716082117e-05,
      "loss": 0.34389835596084595,
      "memory(GiB)": 78.33,
      "step": 4566,
      "token_acc": 0.8980959097320169,
      "train_speed(iter/s)": 0.032209
    },
    {
      "epoch": 0.8849488930872451,
      "grad_norm": 0.09629754722118378,
      "learning_rate": 1.0738314876152587e-05,
      "loss": 0.29018452763557434,
      "memory(GiB)": 78.33,
      "step": 4567,
      "token_acc": 0.9128355904015205,
      "train_speed(iter/s)": 0.032209
    },
    {
      "epoch": 0.8851426633725719,
      "grad_norm": 0.11927231401205063,
      "learning_rate": 1.0702625234739215e-05,
      "loss": 0.3924431800842285,
      "memory(GiB)": 78.33,
      "step": 4568,
      "token_acc": 0.8830037082818294,
      "train_speed(iter/s)": 0.03221
    },
    {
      "epoch": 0.8853364336578986,
      "grad_norm": 0.10137398540973663,
      "learning_rate": 1.0666992806500774e-05,
      "loss": 0.3232523202896118,
      "memory(GiB)": 78.33,
      "step": 4569,
      "token_acc": 0.9024434010228191,
      "train_speed(iter/s)": 0.032211
    },
    {
      "epoch": 0.8855302039432253,
      "grad_norm": 0.09592381864786148,
      "learning_rate": 1.0631417606072356e-05,
      "loss": 0.31686073541641235,
      "memory(GiB)": 78.33,
      "step": 4570,
      "token_acc": 0.9042706275456697,
      "train_speed(iter/s)": 0.032211
    },
    {
      "epoch": 0.8857239742285521,
      "grad_norm": 0.10683749616146088,
      "learning_rate": 1.0595899648065742e-05,
      "loss": 0.33387985825538635,
      "memory(GiB)": 78.33,
      "step": 4571,
      "token_acc": 0.8986985376504726,
      "train_speed(iter/s)": 0.032212
    },
    {
      "epoch": 0.8859177445138788,
      "grad_norm": 0.10224435478448868,
      "learning_rate": 1.0560438947069077e-05,
      "loss": 0.32763242721557617,
      "memory(GiB)": 78.33,
      "step": 4572,
      "token_acc": 0.9004544285346824,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.8861115147992056,
      "grad_norm": 0.1058693677186966,
      "learning_rate": 1.0525035517647012e-05,
      "loss": 0.33532196283340454,
      "memory(GiB)": 78.33,
      "step": 4573,
      "token_acc": 0.899757553151809,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.8863052850845323,
      "grad_norm": 0.10156890749931335,
      "learning_rate": 1.0489689374340699e-05,
      "loss": 0.32516196370124817,
      "memory(GiB)": 78.33,
      "step": 4574,
      "token_acc": 0.9034920634920635,
      "train_speed(iter/s)": 0.032214
    },
    {
      "epoch": 0.8864990553698591,
      "grad_norm": 0.09546992182731628,
      "learning_rate": 1.0454400531667723e-05,
      "loss": 0.32520344853401184,
      "memory(GiB)": 78.33,
      "step": 4575,
      "token_acc": 0.9030310559006212,
      "train_speed(iter/s)": 0.032214
    },
    {
      "epoch": 0.8866928256551858,
      "grad_norm": 0.0944749042391777,
      "learning_rate": 1.0419169004122208e-05,
      "loss": 0.308064341545105,
      "memory(GiB)": 78.33,
      "step": 4576,
      "token_acc": 0.9058090431558348,
      "train_speed(iter/s)": 0.032215
    },
    {
      "epoch": 0.8868865959405126,
      "grad_norm": 0.10865960270166397,
      "learning_rate": 1.0383994806174678e-05,
      "loss": 0.34614354372024536,
      "memory(GiB)": 78.33,
      "step": 4577,
      "token_acc": 0.8971537001897533,
      "train_speed(iter/s)": 0.032216
    },
    {
      "epoch": 0.8870803662258393,
      "grad_norm": 0.11621136218309402,
      "learning_rate": 1.0348877952272094e-05,
      "loss": 0.3003866672515869,
      "memory(GiB)": 78.33,
      "step": 4578,
      "token_acc": 0.908298606977021,
      "train_speed(iter/s)": 0.032216
    },
    {
      "epoch": 0.887274136511166,
      "grad_norm": 0.09108857810497284,
      "learning_rate": 1.0313818456837918e-05,
      "loss": 0.29279422760009766,
      "memory(GiB)": 78.33,
      "step": 4579,
      "token_acc": 0.9107714628777266,
      "train_speed(iter/s)": 0.032217
    },
    {
      "epoch": 0.8874679067964928,
      "grad_norm": 0.1540272831916809,
      "learning_rate": 1.0278816334271984e-05,
      "loss": 0.3414173722267151,
      "memory(GiB)": 78.33,
      "step": 4580,
      "token_acc": 0.8987891249714416,
      "train_speed(iter/s)": 0.032217
    },
    {
      "epoch": 0.8876616770818195,
      "grad_norm": 0.10262391716241837,
      "learning_rate": 1.024387159895067e-05,
      "loss": 0.35209545493125916,
      "memory(GiB)": 78.33,
      "step": 4581,
      "token_acc": 0.894268269722155,
      "train_speed(iter/s)": 0.032218
    },
    {
      "epoch": 0.8878554473671463,
      "grad_norm": 0.10410673916339874,
      "learning_rate": 1.0208984265226649e-05,
      "loss": 0.3540557324886322,
      "memory(GiB)": 78.33,
      "step": 4582,
      "token_acc": 0.896634891466636,
      "train_speed(iter/s)": 0.032219
    },
    {
      "epoch": 0.888049217652473,
      "grad_norm": 0.09851502627134323,
      "learning_rate": 1.0174154347429141e-05,
      "loss": 0.3153938055038452,
      "memory(GiB)": 78.33,
      "step": 4583,
      "token_acc": 0.904395755821965,
      "train_speed(iter/s)": 0.032219
    },
    {
      "epoch": 0.8882429879377998,
      "grad_norm": 0.09433472901582718,
      "learning_rate": 1.013938185986372e-05,
      "loss": 0.30990996956825256,
      "memory(GiB)": 78.33,
      "step": 4584,
      "token_acc": 0.9047421073145381,
      "train_speed(iter/s)": 0.03222
    },
    {
      "epoch": 0.8884367582231265,
      "grad_norm": 0.10146637260913849,
      "learning_rate": 1.0104666816812362e-05,
      "loss": 0.33719319105148315,
      "memory(GiB)": 78.33,
      "step": 4585,
      "token_acc": 0.89920724801812,
      "train_speed(iter/s)": 0.032221
    },
    {
      "epoch": 0.8886305285084533,
      "grad_norm": 0.11201319843530655,
      "learning_rate": 1.0070009232533476e-05,
      "loss": 0.3517708480358124,
      "memory(GiB)": 78.33,
      "step": 4586,
      "token_acc": 0.8955111278762731,
      "train_speed(iter/s)": 0.032221
    },
    {
      "epoch": 0.88882429879378,
      "grad_norm": 0.09309983253479004,
      "learning_rate": 1.0035409121261828e-05,
      "loss": 0.2867148518562317,
      "memory(GiB)": 78.33,
      "step": 4587,
      "token_acc": 0.912418334951329,
      "train_speed(iter/s)": 0.032222
    },
    {
      "epoch": 0.8890180690791067,
      "grad_norm": 0.09345608949661255,
      "learning_rate": 1.0000866497208714e-05,
      "loss": 0.310814768075943,
      "memory(GiB)": 78.33,
      "step": 4588,
      "token_acc": 0.9047558526727841,
      "train_speed(iter/s)": 0.032222
    },
    {
      "epoch": 0.8892118393644335,
      "grad_norm": 0.1012081727385521,
      "learning_rate": 9.96638137456159e-06,
      "loss": 0.3262358605861664,
      "memory(GiB)": 78.33,
      "step": 4589,
      "token_acc": 0.9022921972678861,
      "train_speed(iter/s)": 0.032223
    },
    {
      "epoch": 0.8894056096497602,
      "grad_norm": 0.10403067618608475,
      "learning_rate": 9.931953767484518e-06,
      "loss": 0.3487986922264099,
      "memory(GiB)": 78.33,
      "step": 4590,
      "token_acc": 0.8952626667398583,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.889599379935087,
      "grad_norm": 0.11049600690603256,
      "learning_rate": 9.89758369011781e-06,
      "loss": 0.3473038971424103,
      "memory(GiB)": 78.33,
      "step": 4591,
      "token_acc": 0.8961088918479065,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.8897931502204137,
      "grad_norm": 0.08729095757007599,
      "learning_rate": 9.863271156578174e-06,
      "loss": 0.28431570529937744,
      "memory(GiB)": 78.33,
      "step": 4592,
      "token_acc": 0.9126174143741486,
      "train_speed(iter/s)": 0.032225
    },
    {
      "epoch": 0.8899869205057405,
      "grad_norm": 0.09780146926641464,
      "learning_rate": 9.829016180958681e-06,
      "loss": 0.3170766234397888,
      "memory(GiB)": 78.33,
      "step": 4593,
      "token_acc": 0.9040569020021075,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.8901806907910672,
      "grad_norm": 0.11150769889354706,
      "learning_rate": 9.794818777328767e-06,
      "loss": 0.29845190048217773,
      "memory(GiB)": 78.33,
      "step": 4594,
      "token_acc": 0.909273330479452,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.890374461076394,
      "grad_norm": 0.10244564712047577,
      "learning_rate": 9.760678959734292e-06,
      "loss": 0.34681934118270874,
      "memory(GiB)": 78.33,
      "step": 4595,
      "token_acc": 0.8985476177715848,
      "train_speed(iter/s)": 0.032227
    },
    {
      "epoch": 0.8905682313617207,
      "grad_norm": 0.10472705215215683,
      "learning_rate": 9.726596742197307e-06,
      "loss": 0.3317742645740509,
      "memory(GiB)": 78.33,
      "step": 4596,
      "token_acc": 0.8990034710558727,
      "train_speed(iter/s)": 0.032227
    },
    {
      "epoch": 0.8907620016470474,
      "grad_norm": 0.0967646911740303,
      "learning_rate": 9.692572138716347e-06,
      "loss": 0.34411588311195374,
      "memory(GiB)": 78.33,
      "step": 4597,
      "token_acc": 0.8969721095713183,
      "train_speed(iter/s)": 0.032228
    },
    {
      "epoch": 0.8909557719323742,
      "grad_norm": 0.10965237766504288,
      "learning_rate": 9.658605163266203e-06,
      "loss": 0.3249712884426117,
      "memory(GiB)": 78.33,
      "step": 4598,
      "token_acc": 0.9018522218099831,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.8911495422177009,
      "grad_norm": 0.10542988777160645,
      "learning_rate": 9.624695829798045e-06,
      "loss": 0.33859795331954956,
      "memory(GiB)": 78.33,
      "step": 4599,
      "token_acc": 0.898973819608289,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.8913433125030277,
      "grad_norm": 0.09840527176856995,
      "learning_rate": 9.590844152239353e-06,
      "loss": 0.3379114270210266,
      "memory(GiB)": 78.33,
      "step": 4600,
      "token_acc": 0.8994015604878418,
      "train_speed(iter/s)": 0.03223
    },
    {
      "epoch": 0.8915370827883544,
      "grad_norm": 0.0945395678281784,
      "learning_rate": 9.557050144493884e-06,
      "loss": 0.30370837450027466,
      "memory(GiB)": 78.33,
      "step": 4601,
      "token_acc": 0.9077565971057858,
      "train_speed(iter/s)": 0.032227
    },
    {
      "epoch": 0.8917308530736812,
      "grad_norm": 0.10549890249967575,
      "learning_rate": 9.523313820441803e-06,
      "loss": 0.3490893840789795,
      "memory(GiB)": 78.33,
      "step": 4602,
      "token_acc": 0.897390101976148,
      "train_speed(iter/s)": 0.032228
    },
    {
      "epoch": 0.8919246233590079,
      "grad_norm": 0.10837141424417496,
      "learning_rate": 9.489635193939444e-06,
      "loss": 0.36473971605300903,
      "memory(GiB)": 78.33,
      "step": 4603,
      "token_acc": 0.8911379629370041,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.8921183936443347,
      "grad_norm": 0.09465762972831726,
      "learning_rate": 9.456014278819606e-06,
      "loss": 0.3148060142993927,
      "memory(GiB)": 78.33,
      "step": 4604,
      "token_acc": 0.9028953344343518,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.8923121639296614,
      "grad_norm": 0.09781907498836517,
      "learning_rate": 9.422451088891264e-06,
      "loss": 0.3031347990036011,
      "memory(GiB)": 78.33,
      "step": 4605,
      "token_acc": 0.908634026587614,
      "train_speed(iter/s)": 0.03223
    },
    {
      "epoch": 0.8925059342149881,
      "grad_norm": 0.09446154534816742,
      "learning_rate": 9.38894563793972e-06,
      "loss": 0.3180665373802185,
      "memory(GiB)": 78.33,
      "step": 4606,
      "token_acc": 0.905185446482564,
      "train_speed(iter/s)": 0.03223
    },
    {
      "epoch": 0.8926997045003149,
      "grad_norm": 0.09713321179151535,
      "learning_rate": 9.355497939726569e-06,
      "loss": 0.31975382566452026,
      "memory(GiB)": 78.33,
      "step": 4607,
      "token_acc": 0.904540804281105,
      "train_speed(iter/s)": 0.032231
    },
    {
      "epoch": 0.8928934747856416,
      "grad_norm": 0.0889185443520546,
      "learning_rate": 9.322108007989654e-06,
      "loss": 0.28848981857299805,
      "memory(GiB)": 78.33,
      "step": 4608,
      "token_acc": 0.9133419560242761,
      "train_speed(iter/s)": 0.032232
    },
    {
      "epoch": 0.8930872450709684,
      "grad_norm": 0.10388769209384918,
      "learning_rate": 9.288775856443187e-06,
      "loss": 0.3318116068840027,
      "memory(GiB)": 78.33,
      "step": 4609,
      "token_acc": 0.9007829299222584,
      "train_speed(iter/s)": 0.032232
    },
    {
      "epoch": 0.8932810153562951,
      "grad_norm": 0.10328590124845505,
      "learning_rate": 9.255501498777485e-06,
      "loss": 0.35095512866973877,
      "memory(GiB)": 78.33,
      "step": 4610,
      "token_acc": 0.8963482109922538,
      "train_speed(iter/s)": 0.032233
    },
    {
      "epoch": 0.8934747856416219,
      "grad_norm": 0.09728133678436279,
      "learning_rate": 9.222284948659297e-06,
      "loss": 0.32896798849105835,
      "memory(GiB)": 78.33,
      "step": 4611,
      "token_acc": 0.9007275166400083,
      "train_speed(iter/s)": 0.032233
    },
    {
      "epoch": 0.8936685559269486,
      "grad_norm": 0.09340497106313705,
      "learning_rate": 9.189126219731513e-06,
      "loss": 0.3199424147605896,
      "memory(GiB)": 78.33,
      "step": 4612,
      "token_acc": 0.9017920064431693,
      "train_speed(iter/s)": 0.032234
    },
    {
      "epoch": 0.8938623262122753,
      "grad_norm": 0.12456195801496506,
      "learning_rate": 9.156025325613319e-06,
      "loss": 0.38881370425224304,
      "memory(GiB)": 78.33,
      "step": 4613,
      "token_acc": 0.8865035516969219,
      "train_speed(iter/s)": 0.032235
    },
    {
      "epoch": 0.8940560964976021,
      "grad_norm": 0.10831289738416672,
      "learning_rate": 9.122982279900192e-06,
      "loss": 0.31916946172714233,
      "memory(GiB)": 78.33,
      "step": 4614,
      "token_acc": 0.9046099809746817,
      "train_speed(iter/s)": 0.032235
    },
    {
      "epoch": 0.8942498667829288,
      "grad_norm": 0.09680543094873428,
      "learning_rate": 9.089997096163692e-06,
      "loss": 0.31975850462913513,
      "memory(GiB)": 78.33,
      "step": 4615,
      "token_acc": 0.9020113186492016,
      "train_speed(iter/s)": 0.032236
    },
    {
      "epoch": 0.8944436370682556,
      "grad_norm": 0.09476674348115921,
      "learning_rate": 9.057069787951832e-06,
      "loss": 0.3149741291999817,
      "memory(GiB)": 78.33,
      "step": 4616,
      "token_acc": 0.9046471784987686,
      "train_speed(iter/s)": 0.032236
    },
    {
      "epoch": 0.8946374073535823,
      "grad_norm": 0.09538931399583817,
      "learning_rate": 9.024200368788676e-06,
      "loss": 0.298635333776474,
      "memory(GiB)": 78.33,
      "step": 4617,
      "token_acc": 0.9104565695584527,
      "train_speed(iter/s)": 0.032237
    },
    {
      "epoch": 0.8948311776389091,
      "grad_norm": 0.09865786135196686,
      "learning_rate": 8.991388852174592e-06,
      "loss": 0.3136424124240875,
      "memory(GiB)": 78.33,
      "step": 4618,
      "token_acc": 0.9061650538218985,
      "train_speed(iter/s)": 0.032238
    },
    {
      "epoch": 0.8950249479242358,
      "grad_norm": 0.09958308935165405,
      "learning_rate": 8.958635251586166e-06,
      "loss": 0.32274237275123596,
      "memory(GiB)": 78.33,
      "step": 4619,
      "token_acc": 0.9019686621132985,
      "train_speed(iter/s)": 0.032238
    },
    {
      "epoch": 0.8952187182095626,
      "grad_norm": 0.1041649580001831,
      "learning_rate": 8.925939580476138e-06,
      "loss": 0.33783018589019775,
      "memory(GiB)": 78.33,
      "step": 4620,
      "token_acc": 0.8964275424949583,
      "train_speed(iter/s)": 0.032239
    },
    {
      "epoch": 0.8954124884948893,
      "grad_norm": 0.09779240190982819,
      "learning_rate": 8.893301852273582e-06,
      "loss": 0.3440534174442291,
      "memory(GiB)": 78.33,
      "step": 4621,
      "token_acc": 0.8966321509124462,
      "train_speed(iter/s)": 0.03224
    },
    {
      "epoch": 0.895606258780216,
      "grad_norm": 0.11845773458480835,
      "learning_rate": 8.860722080383593e-06,
      "loss": 0.35379868745803833,
      "memory(GiB)": 78.33,
      "step": 4622,
      "token_acc": 0.8932370241687239,
      "train_speed(iter/s)": 0.03224
    },
    {
      "epoch": 0.8958000290655428,
      "grad_norm": 0.0978410616517067,
      "learning_rate": 8.82820027818763e-06,
      "loss": 0.3412013053894043,
      "memory(GiB)": 78.33,
      "step": 4623,
      "token_acc": 0.9002301375594958,
      "train_speed(iter/s)": 0.032241
    },
    {
      "epoch": 0.8959937993508695,
      "grad_norm": 0.09728839993476868,
      "learning_rate": 8.795736459043246e-06,
      "loss": 0.3061705529689789,
      "memory(GiB)": 78.33,
      "step": 4624,
      "token_acc": 0.9070262390670554,
      "train_speed(iter/s)": 0.032241
    },
    {
      "epoch": 0.8961875696361963,
      "grad_norm": 0.09861348569393158,
      "learning_rate": 8.763330636284204e-06,
      "loss": 0.3142586350440979,
      "memory(GiB)": 78.33,
      "step": 4625,
      "token_acc": 0.9079194559839123,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.896381339921523,
      "grad_norm": 0.09855727106332779,
      "learning_rate": 8.730982823220445e-06,
      "loss": 0.32500770688056946,
      "memory(GiB)": 78.33,
      "step": 4626,
      "token_acc": 0.9023219335979276,
      "train_speed(iter/s)": 0.032243
    },
    {
      "epoch": 0.8965751102068498,
      "grad_norm": 0.10426346212625504,
      "learning_rate": 8.698693033138054e-06,
      "loss": 0.32699429988861084,
      "memory(GiB)": 78.33,
      "step": 4627,
      "token_acc": 0.9018576936558009,
      "train_speed(iter/s)": 0.032243
    },
    {
      "epoch": 0.8967688804921765,
      "grad_norm": 0.09685856848955154,
      "learning_rate": 8.666461279299408e-06,
      "loss": 0.3376917243003845,
      "memory(GiB)": 78.33,
      "step": 4628,
      "token_acc": 0.8998025786638719,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.8969626507775033,
      "grad_norm": 0.10653463006019592,
      "learning_rate": 8.634287574942833e-06,
      "loss": 0.3381001055240631,
      "memory(GiB)": 78.33,
      "step": 4629,
      "token_acc": 0.9000566335807326,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.89715642106283,
      "grad_norm": 0.1089102178812027,
      "learning_rate": 8.602171933283025e-06,
      "loss": 0.3288399279117584,
      "memory(GiB)": 78.33,
      "step": 4630,
      "token_acc": 0.9028040685495066,
      "train_speed(iter/s)": 0.032245
    },
    {
      "epoch": 0.8973501913481567,
      "grad_norm": 0.10240910202264786,
      "learning_rate": 8.570114367510717e-06,
      "loss": 0.33557647466659546,
      "memory(GiB)": 78.33,
      "step": 4631,
      "token_acc": 0.897394733136178,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.8975439616334835,
      "grad_norm": 0.10204704850912094,
      "learning_rate": 8.53811489079279e-06,
      "loss": 0.3257525861263275,
      "memory(GiB)": 78.33,
      "step": 4632,
      "token_acc": 0.9036119817088836,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.8977377319188102,
      "grad_norm": 0.10609758645296097,
      "learning_rate": 8.506173516272319e-06,
      "loss": 0.362753689289093,
      "memory(GiB)": 78.33,
      "step": 4633,
      "token_acc": 0.8909109816971714,
      "train_speed(iter/s)": 0.032247
    },
    {
      "epoch": 0.897931502204137,
      "grad_norm": 0.09235696494579315,
      "learning_rate": 8.474290257068456e-06,
      "loss": 0.3118291199207306,
      "memory(GiB)": 78.33,
      "step": 4634,
      "token_acc": 0.9042984381050729,
      "train_speed(iter/s)": 0.032247
    },
    {
      "epoch": 0.8981252724894637,
      "grad_norm": 0.10148902982473373,
      "learning_rate": 8.44246512627656e-06,
      "loss": 0.3247118592262268,
      "memory(GiB)": 78.33,
      "step": 4635,
      "token_acc": 0.9009333957432749,
      "train_speed(iter/s)": 0.032248
    },
    {
      "epoch": 0.8983190427747905,
      "grad_norm": 0.10464513301849365,
      "learning_rate": 8.41069813696799e-06,
      "loss": 0.3417477309703827,
      "memory(GiB)": 78.33,
      "step": 4636,
      "token_acc": 0.898022001552421,
      "train_speed(iter/s)": 0.032249
    },
    {
      "epoch": 0.8985128130601172,
      "grad_norm": 0.0930318832397461,
      "learning_rate": 8.37898930219038e-06,
      "loss": 0.3098987638950348,
      "memory(GiB)": 78.33,
      "step": 4637,
      "token_acc": 0.9070895522388059,
      "train_speed(iter/s)": 0.032249
    },
    {
      "epoch": 0.898706583345444,
      "grad_norm": 0.09498867392539978,
      "learning_rate": 8.34733863496736e-06,
      "loss": 0.3298591375350952,
      "memory(GiB)": 78.33,
      "step": 4638,
      "token_acc": 0.9012791647079199,
      "train_speed(iter/s)": 0.03225
    },
    {
      "epoch": 0.8989003536307707,
      "grad_norm": 0.09717161953449249,
      "learning_rate": 8.315746148298713e-06,
      "loss": 0.3042880892753601,
      "memory(GiB)": 78.33,
      "step": 4639,
      "token_acc": 0.9077266387726639,
      "train_speed(iter/s)": 0.03225
    },
    {
      "epoch": 0.8990941239160974,
      "grad_norm": 0.09742318093776703,
      "learning_rate": 8.284211855160328e-06,
      "loss": 0.32740819454193115,
      "memory(GiB)": 78.33,
      "step": 4640,
      "token_acc": 0.9031217784356744,
      "train_speed(iter/s)": 0.032251
    },
    {
      "epoch": 0.8992878942014242,
      "grad_norm": 0.10585065186023712,
      "learning_rate": 8.252735768504176e-06,
      "loss": 0.32093313336372375,
      "memory(GiB)": 78.33,
      "step": 4641,
      "token_acc": 0.9033283470456245,
      "train_speed(iter/s)": 0.032252
    },
    {
      "epoch": 0.8994816644867509,
      "grad_norm": 0.10464109480381012,
      "learning_rate": 8.221317901258367e-06,
      "loss": 0.3451445698738098,
      "memory(GiB)": 78.33,
      "step": 4642,
      "token_acc": 0.8969477704080168,
      "train_speed(iter/s)": 0.032252
    },
    {
      "epoch": 0.8996754347720777,
      "grad_norm": 0.09843490272760391,
      "learning_rate": 8.189958266326996e-06,
      "loss": 0.3220871090888977,
      "memory(GiB)": 78.33,
      "step": 4643,
      "token_acc": 0.9026759937819232,
      "train_speed(iter/s)": 0.032253
    },
    {
      "epoch": 0.8998692050574044,
      "grad_norm": 0.10574258118867874,
      "learning_rate": 8.158656876590375e-06,
      "loss": 0.36457881331443787,
      "memory(GiB)": 78.33,
      "step": 4644,
      "token_acc": 0.8903435022820082,
      "train_speed(iter/s)": 0.032254
    },
    {
      "epoch": 0.9000629753427312,
      "grad_norm": 0.0989030972123146,
      "learning_rate": 8.127413744904804e-06,
      "loss": 0.3353177309036255,
      "memory(GiB)": 78.33,
      "step": 4645,
      "token_acc": 0.8995719135954338,
      "train_speed(iter/s)": 0.032254
    },
    {
      "epoch": 0.9002567456280579,
      "grad_norm": 0.10432767868041992,
      "learning_rate": 8.096228884102652e-06,
      "loss": 0.32852283120155334,
      "memory(GiB)": 78.33,
      "step": 4646,
      "token_acc": 0.9020051081818983,
      "train_speed(iter/s)": 0.032255
    },
    {
      "epoch": 0.9004505159133847,
      "grad_norm": 0.09884372353553772,
      "learning_rate": 8.065102306992439e-06,
      "loss": 0.34128913283348083,
      "memory(GiB)": 78.33,
      "step": 4647,
      "token_acc": 0.8975473369505038,
      "train_speed(iter/s)": 0.032255
    },
    {
      "epoch": 0.9006442861987114,
      "grad_norm": 0.09996260702610016,
      "learning_rate": 8.034034026358587e-06,
      "loss": 0.32066965103149414,
      "memory(GiB)": 78.33,
      "step": 4648,
      "token_acc": 0.9021558379855089,
      "train_speed(iter/s)": 0.032256
    },
    {
      "epoch": 0.9008380564840381,
      "grad_norm": 0.10087880492210388,
      "learning_rate": 8.003024054961776e-06,
      "loss": 0.34059831500053406,
      "memory(GiB)": 78.33,
      "step": 4649,
      "token_acc": 0.89795299887043,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.9010318267693649,
      "grad_norm": 0.10195588320493698,
      "learning_rate": 7.972072405538582e-06,
      "loss": 0.3243820369243622,
      "memory(GiB)": 78.33,
      "step": 4650,
      "token_acc": 0.9031757307864341,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.9012255970546916,
      "grad_norm": 0.11210127919912338,
      "learning_rate": 7.941179090801687e-06,
      "loss": 0.33561965823173523,
      "memory(GiB)": 78.33,
      "step": 4651,
      "token_acc": 0.8998974967630556,
      "train_speed(iter/s)": 0.032258
    },
    {
      "epoch": 0.9014193673400184,
      "grad_norm": 0.09735889732837677,
      "learning_rate": 7.91034412343982e-06,
      "loss": 0.3177420496940613,
      "memory(GiB)": 78.33,
      "step": 4652,
      "token_acc": 0.9025480310669028,
      "train_speed(iter/s)": 0.032258
    },
    {
      "epoch": 0.9016131376253451,
      "grad_norm": 0.1058276817202568,
      "learning_rate": 7.879567516117691e-06,
      "loss": 0.34389030933380127,
      "memory(GiB)": 78.33,
      "step": 4653,
      "token_acc": 0.9005196304849884,
      "train_speed(iter/s)": 0.032259
    },
    {
      "epoch": 0.9018069079106719,
      "grad_norm": 0.10890252888202667,
      "learning_rate": 7.848849281476149e-06,
      "loss": 0.3455137014389038,
      "memory(GiB)": 78.33,
      "step": 4654,
      "token_acc": 0.89581208468552,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 0.9020006781959986,
      "grad_norm": 0.09760237485170364,
      "learning_rate": 7.818189432131921e-06,
      "loss": 0.3121528923511505,
      "memory(GiB)": 78.33,
      "step": 4655,
      "token_acc": 0.9046669988374024,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 0.9021944484813254,
      "grad_norm": 0.0973569005727768,
      "learning_rate": 7.787587980677868e-06,
      "loss": 0.3131091594696045,
      "memory(GiB)": 78.33,
      "step": 4656,
      "token_acc": 0.9063661202185792,
      "train_speed(iter/s)": 0.032261
    },
    {
      "epoch": 0.9023882187666521,
      "grad_norm": 0.09090343117713928,
      "learning_rate": 7.75704493968285e-06,
      "loss": 0.29967737197875977,
      "memory(GiB)": 78.33,
      "step": 4657,
      "token_acc": 0.9070998676422846,
      "train_speed(iter/s)": 0.032262
    },
    {
      "epoch": 0.9025819890519788,
      "grad_norm": 0.10325030982494354,
      "learning_rate": 7.726560321691682e-06,
      "loss": 0.3381047248840332,
      "memory(GiB)": 78.33,
      "step": 4658,
      "token_acc": 0.8999558238845531,
      "train_speed(iter/s)": 0.032262
    },
    {
      "epoch": 0.9027757593373056,
      "grad_norm": 0.10501549392938614,
      "learning_rate": 7.696134139225219e-06,
      "loss": 0.33340954780578613,
      "memory(GiB)": 78.33,
      "step": 4659,
      "token_acc": 0.8982804760006833,
      "train_speed(iter/s)": 0.032263
    },
    {
      "epoch": 0.9029695296226323,
      "grad_norm": 0.098429374396801,
      "learning_rate": 7.66576640478031e-06,
      "loss": 0.30632802844047546,
      "memory(GiB)": 78.33,
      "step": 4660,
      "token_acc": 0.9067490984028851,
      "train_speed(iter/s)": 0.032263
    },
    {
      "epoch": 0.9031632999079591,
      "grad_norm": 0.11389485746622086,
      "learning_rate": 7.635457130829832e-06,
      "loss": 0.34137189388275146,
      "memory(GiB)": 78.33,
      "step": 4661,
      "token_acc": 0.8986180210060807,
      "train_speed(iter/s)": 0.032264
    },
    {
      "epoch": 0.9033570701932858,
      "grad_norm": 0.09606049209833145,
      "learning_rate": 7.6052063298225715e-06,
      "loss": 0.29577910900115967,
      "memory(GiB)": 78.33,
      "step": 4662,
      "token_acc": 0.9094975613979395,
      "train_speed(iter/s)": 0.032265
    },
    {
      "epoch": 0.9035508404786126,
      "grad_norm": 0.10640320926904678,
      "learning_rate": 7.575014014183378e-06,
      "loss": 0.30323442816734314,
      "memory(GiB)": 78.33,
      "step": 4663,
      "token_acc": 0.9066974069962713,
      "train_speed(iter/s)": 0.032265
    },
    {
      "epoch": 0.9037446107639393,
      "grad_norm": 0.10014378279447556,
      "learning_rate": 7.5448801963130305e-06,
      "loss": 0.337171733379364,
      "memory(GiB)": 78.33,
      "step": 4664,
      "token_acc": 0.8995919717688575,
      "train_speed(iter/s)": 0.032266
    },
    {
      "epoch": 0.903938381049266,
      "grad_norm": 0.10196644067764282,
      "learning_rate": 7.5148048885883105e-06,
      "loss": 0.31075039505958557,
      "memory(GiB)": 78.33,
      "step": 4665,
      "token_acc": 0.9083940206975852,
      "train_speed(iter/s)": 0.032266
    },
    {
      "epoch": 0.9041321513345928,
      "grad_norm": 0.09376231580972672,
      "learning_rate": 7.484788103361955e-06,
      "loss": 0.3261268436908722,
      "memory(GiB)": 78.33,
      "step": 4666,
      "token_acc": 0.9021167228378149,
      "train_speed(iter/s)": 0.032267
    },
    {
      "epoch": 0.9043259216199195,
      "grad_norm": 0.09824056923389435,
      "learning_rate": 7.454829852962635e-06,
      "loss": 0.32310348749160767,
      "memory(GiB)": 78.33,
      "step": 4667,
      "token_acc": 0.9040854978354979,
      "train_speed(iter/s)": 0.032268
    },
    {
      "epoch": 0.9045196919052463,
      "grad_norm": 0.09825246036052704,
      "learning_rate": 7.424930149695074e-06,
      "loss": 0.32509177923202515,
      "memory(GiB)": 78.33,
      "step": 4668,
      "token_acc": 0.9037440744840898,
      "train_speed(iter/s)": 0.032268
    },
    {
      "epoch": 0.904713462190573,
      "grad_norm": 0.11145921051502228,
      "learning_rate": 7.395089005839783e-06,
      "loss": 0.32568415999412537,
      "memory(GiB)": 78.33,
      "step": 4669,
      "token_acc": 0.9009174311926605,
      "train_speed(iter/s)": 0.032269
    },
    {
      "epoch": 0.9049072324758998,
      "grad_norm": 0.10170602053403854,
      "learning_rate": 7.365306433653423e-06,
      "loss": 0.3407144844532013,
      "memory(GiB)": 78.33,
      "step": 4670,
      "token_acc": 0.8984231756508985,
      "train_speed(iter/s)": 0.032269
    },
    {
      "epoch": 0.9051010027612265,
      "grad_norm": 0.1278442144393921,
      "learning_rate": 7.335582445368443e-06,
      "loss": 0.3306085467338562,
      "memory(GiB)": 78.33,
      "step": 4671,
      "token_acc": 0.8980235373613379,
      "train_speed(iter/s)": 0.03227
    },
    {
      "epoch": 0.9052947730465533,
      "grad_norm": 0.0913781076669693,
      "learning_rate": 7.305917053193294e-06,
      "loss": 0.28779953718185425,
      "memory(GiB)": 78.33,
      "step": 4672,
      "token_acc": 0.9126665944305992,
      "train_speed(iter/s)": 0.032271
    },
    {
      "epoch": 0.90548854333188,
      "grad_norm": 0.08933483064174652,
      "learning_rate": 7.276310269312347e-06,
      "loss": 0.3098442852497101,
      "memory(GiB)": 78.33,
      "step": 4673,
      "token_acc": 0.9050055824339412,
      "train_speed(iter/s)": 0.032271
    },
    {
      "epoch": 0.9056823136172067,
      "grad_norm": 0.09485527127981186,
      "learning_rate": 7.24676210588589e-06,
      "loss": 0.3187693655490875,
      "memory(GiB)": 78.33,
      "step": 4674,
      "token_acc": 0.9054063940302876,
      "train_speed(iter/s)": 0.032272
    },
    {
      "epoch": 0.9058760839025335,
      "grad_norm": 0.10101988166570663,
      "learning_rate": 7.217272575050198e-06,
      "loss": 0.3332398235797882,
      "memory(GiB)": 78.33,
      "step": 4675,
      "token_acc": 0.9006496386548044,
      "train_speed(iter/s)": 0.032272
    },
    {
      "epoch": 0.9060698541878603,
      "grad_norm": 0.1058175191283226,
      "learning_rate": 7.187841688917351e-06,
      "loss": 0.30771443247795105,
      "memory(GiB)": 78.33,
      "step": 4676,
      "token_acc": 0.9090541632983024,
      "train_speed(iter/s)": 0.032273
    },
    {
      "epoch": 0.9062636244731871,
      "grad_norm": 0.10288235545158386,
      "learning_rate": 7.158469459575444e-06,
      "loss": 0.31234169006347656,
      "memory(GiB)": 78.33,
      "step": 4677,
      "token_acc": 0.9043364814657251,
      "train_speed(iter/s)": 0.032274
    },
    {
      "epoch": 0.9064573947585138,
      "grad_norm": 0.11373648047447205,
      "learning_rate": 7.129155899088429e-06,
      "loss": 0.36738601326942444,
      "memory(GiB)": 78.33,
      "step": 4678,
      "token_acc": 0.8934572225850501,
      "train_speed(iter/s)": 0.032274
    },
    {
      "epoch": 0.9066511650438406,
      "grad_norm": 0.09895353019237518,
      "learning_rate": 7.099901019496157e-06,
      "loss": 0.32815736532211304,
      "memory(GiB)": 78.33,
      "step": 4679,
      "token_acc": 0.9006158244900203,
      "train_speed(iter/s)": 0.032275
    },
    {
      "epoch": 0.9068449353291673,
      "grad_norm": 0.10247547179460526,
      "learning_rate": 7.070704832814467e-06,
      "loss": 0.32758232951164246,
      "memory(GiB)": 78.33,
      "step": 4680,
      "token_acc": 0.9017008208823009,
      "train_speed(iter/s)": 0.032275
    },
    {
      "epoch": 0.9070387056144941,
      "grad_norm": 0.09913128614425659,
      "learning_rate": 7.041567351034899e-06,
      "loss": 0.31721755862236023,
      "memory(GiB)": 78.33,
      "step": 4681,
      "token_acc": 0.9058820262163964,
      "train_speed(iter/s)": 0.032276
    },
    {
      "epoch": 0.9072324758998208,
      "grad_norm": 0.1092270091176033,
      "learning_rate": 7.0124885861251145e-06,
      "loss": 0.31192123889923096,
      "memory(GiB)": 78.33,
      "step": 4682,
      "token_acc": 0.9066539550794942,
      "train_speed(iter/s)": 0.032277
    },
    {
      "epoch": 0.9074262461851476,
      "grad_norm": 0.09576813131570816,
      "learning_rate": 6.983468550028442e-06,
      "loss": 0.3214895725250244,
      "memory(GiB)": 78.33,
      "step": 4683,
      "token_acc": 0.9031646569427624,
      "train_speed(iter/s)": 0.032277
    },
    {
      "epoch": 0.9076200164704743,
      "grad_norm": 0.09397298097610474,
      "learning_rate": 6.954507254664266e-06,
      "loss": 0.3123416006565094,
      "memory(GiB)": 78.33,
      "step": 4684,
      "token_acc": 0.9053452877968654,
      "train_speed(iter/s)": 0.032278
    },
    {
      "epoch": 0.907813786755801,
      "grad_norm": 0.11117564886808395,
      "learning_rate": 6.925604711927751e-06,
      "loss": 0.36098116636276245,
      "memory(GiB)": 78.33,
      "step": 4685,
      "token_acc": 0.8910591965540965,
      "train_speed(iter/s)": 0.032278
    },
    {
      "epoch": 0.9080075570411278,
      "grad_norm": 0.0936315655708313,
      "learning_rate": 6.896760933689904e-06,
      "loss": 0.3093935251235962,
      "memory(GiB)": 78.33,
      "step": 4686,
      "token_acc": 0.9063284652831155,
      "train_speed(iter/s)": 0.032279
    },
    {
      "epoch": 0.9082013273264545,
      "grad_norm": 0.09810250252485275,
      "learning_rate": 6.867975931797715e-06,
      "loss": 0.2804984450340271,
      "memory(GiB)": 78.33,
      "step": 4687,
      "token_acc": 0.9146265467818037,
      "train_speed(iter/s)": 0.03228
    },
    {
      "epoch": 0.9083950976117813,
      "grad_norm": 0.09873296320438385,
      "learning_rate": 6.839249718073875e-06,
      "loss": 0.3098883032798767,
      "memory(GiB)": 78.33,
      "step": 4688,
      "token_acc": 0.9055900621118013,
      "train_speed(iter/s)": 0.03228
    },
    {
      "epoch": 0.908588867897108,
      "grad_norm": 0.10540574043989182,
      "learning_rate": 6.810582304317081e-06,
      "loss": 0.31421953439712524,
      "memory(GiB)": 78.33,
      "step": 4689,
      "token_acc": 0.9060614765871876,
      "train_speed(iter/s)": 0.032281
    },
    {
      "epoch": 0.9087826381824348,
      "grad_norm": 0.12720443308353424,
      "learning_rate": 6.781973702301796e-06,
      "loss": 0.3550341725349426,
      "memory(GiB)": 78.33,
      "step": 4690,
      "token_acc": 0.8944157949020186,
      "train_speed(iter/s)": 0.032282
    },
    {
      "epoch": 0.9089764084677615,
      "grad_norm": 0.1082942858338356,
      "learning_rate": 6.7534239237783065e-06,
      "loss": 0.3174242675304413,
      "memory(GiB)": 78.33,
      "step": 4691,
      "token_acc": 0.9045280815373182,
      "train_speed(iter/s)": 0.032282
    },
    {
      "epoch": 0.9091701787530883,
      "grad_norm": 0.0936630368232727,
      "learning_rate": 6.724932980472813e-06,
      "loss": 0.3189206123352051,
      "memory(GiB)": 78.33,
      "step": 4692,
      "token_acc": 0.9040681037771843,
      "train_speed(iter/s)": 0.032283
    },
    {
      "epoch": 0.909363949038415,
      "grad_norm": 0.1038210541009903,
      "learning_rate": 6.696500884087258e-06,
      "loss": 0.3283487558364868,
      "memory(GiB)": 78.33,
      "step": 4693,
      "token_acc": 0.901529364943999,
      "train_speed(iter/s)": 0.032283
    },
    {
      "epoch": 0.9095577193237417,
      "grad_norm": 0.09301093220710754,
      "learning_rate": 6.668127646299548e-06,
      "loss": 0.3160119652748108,
      "memory(GiB)": 78.33,
      "step": 4694,
      "token_acc": 0.906036029207408,
      "train_speed(iter/s)": 0.032284
    },
    {
      "epoch": 0.9097514896090685,
      "grad_norm": 0.10908481478691101,
      "learning_rate": 6.639813278763262e-06,
      "loss": 0.35018086433410645,
      "memory(GiB)": 78.33,
      "step": 4695,
      "token_acc": 0.8955263546434372,
      "train_speed(iter/s)": 0.032285
    },
    {
      "epoch": 0.9099452598943952,
      "grad_norm": 0.09090767800807953,
      "learning_rate": 6.611557793107914e-06,
      "loss": 0.30631187558174133,
      "memory(GiB)": 78.33,
      "step": 4696,
      "token_acc": 0.9081879446779534,
      "train_speed(iter/s)": 0.032285
    },
    {
      "epoch": 0.910139030179722,
      "grad_norm": 0.10230287909507751,
      "learning_rate": 6.583361200938769e-06,
      "loss": 0.3390125036239624,
      "memory(GiB)": 78.33,
      "step": 4697,
      "token_acc": 0.8979241998125084,
      "train_speed(iter/s)": 0.032286
    },
    {
      "epoch": 0.9103328004650487,
      "grad_norm": 0.10016355663537979,
      "learning_rate": 6.5552235138369494e-06,
      "loss": 0.32282742857933044,
      "memory(GiB)": 78.33,
      "step": 4698,
      "token_acc": 0.904220375941925,
      "train_speed(iter/s)": 0.032286
    },
    {
      "epoch": 0.9105265707503755,
      "grad_norm": 0.10659368336200714,
      "learning_rate": 6.527144743359342e-06,
      "loss": 0.35568249225616455,
      "memory(GiB)": 78.33,
      "step": 4699,
      "token_acc": 0.8939569935121965,
      "train_speed(iter/s)": 0.032287
    },
    {
      "epoch": 0.9107203410357022,
      "grad_norm": 0.10241632908582687,
      "learning_rate": 6.499124901038621e-06,
      "loss": 0.33742234110832214,
      "memory(GiB)": 78.33,
      "step": 4700,
      "token_acc": 0.8987963512711178,
      "train_speed(iter/s)": 0.032287
    },
    {
      "epoch": 0.910914111321029,
      "grad_norm": 0.09770441055297852,
      "learning_rate": 6.471163998383366e-06,
      "loss": 0.3216272294521332,
      "memory(GiB)": 78.33,
      "step": 4701,
      "token_acc": 0.903465820499494,
      "train_speed(iter/s)": 0.032288
    },
    {
      "epoch": 0.9111078816063557,
      "grad_norm": 0.09685520827770233,
      "learning_rate": 6.44326204687779e-06,
      "loss": 0.2898436188697815,
      "memory(GiB)": 78.33,
      "step": 4702,
      "token_acc": 0.9125583603703411,
      "train_speed(iter/s)": 0.032289
    },
    {
      "epoch": 0.9113016518916824,
      "grad_norm": 0.10244952887296677,
      "learning_rate": 6.415419057982024e-06,
      "loss": 0.3338771164417267,
      "memory(GiB)": 78.33,
      "step": 4703,
      "token_acc": 0.8992232679442114,
      "train_speed(iter/s)": 0.032289
    },
    {
      "epoch": 0.9114954221770092,
      "grad_norm": 0.09772542864084244,
      "learning_rate": 6.387635043131923e-06,
      "loss": 0.31243184208869934,
      "memory(GiB)": 78.33,
      "step": 4704,
      "token_acc": 0.9057063523545078,
      "train_speed(iter/s)": 0.03229
    },
    {
      "epoch": 0.9116891924623359,
      "grad_norm": 0.10280278325080872,
      "learning_rate": 6.359910013739122e-06,
      "loss": 0.3368726372718811,
      "memory(GiB)": 78.33,
      "step": 4705,
      "token_acc": 0.8969188578803032,
      "train_speed(iter/s)": 0.03229
    },
    {
      "epoch": 0.9118829627476627,
      "grad_norm": 0.09057314693927765,
      "learning_rate": 6.332243981191032e-06,
      "loss": 0.30282315611839294,
      "memory(GiB)": 78.33,
      "step": 4706,
      "token_acc": 0.9087763447625039,
      "train_speed(iter/s)": 0.032291
    },
    {
      "epoch": 0.9120767330329894,
      "grad_norm": 0.10076764971017838,
      "learning_rate": 6.304636956850828e-06,
      "loss": 0.32998421788215637,
      "memory(GiB)": 78.33,
      "step": 4707,
      "token_acc": 0.900506177543134,
      "train_speed(iter/s)": 0.032291
    },
    {
      "epoch": 0.9122705033183162,
      "grad_norm": 0.10298432409763336,
      "learning_rate": 6.277088952057508e-06,
      "loss": 0.3607329726219177,
      "memory(GiB)": 78.33,
      "step": 4708,
      "token_acc": 0.8920216927933519,
      "train_speed(iter/s)": 0.032292
    },
    {
      "epoch": 0.9124642736036429,
      "grad_norm": 0.09257783740758896,
      "learning_rate": 6.249599978125685e-06,
      "loss": 0.31179288029670715,
      "memory(GiB)": 78.33,
      "step": 4709,
      "token_acc": 0.9052415912583234,
      "train_speed(iter/s)": 0.032293
    },
    {
      "epoch": 0.9126580438889697,
      "grad_norm": 0.10241623967885971,
      "learning_rate": 6.222170046345914e-06,
      "loss": 0.3377891480922699,
      "memory(GiB)": 78.33,
      "step": 4710,
      "token_acc": 0.8971870153484144,
      "train_speed(iter/s)": 0.032293
    },
    {
      "epoch": 0.9128518141742964,
      "grad_norm": 0.09825358539819717,
      "learning_rate": 6.194799167984365e-06,
      "loss": 0.2925589084625244,
      "memory(GiB)": 78.33,
      "step": 4711,
      "token_acc": 0.9106943825234062,
      "train_speed(iter/s)": 0.032294
    },
    {
      "epoch": 0.9130455844596231,
      "grad_norm": 0.09656988829374313,
      "learning_rate": 6.167487354282963e-06,
      "loss": 0.32839635014533997,
      "memory(GiB)": 78.33,
      "step": 4712,
      "token_acc": 0.9005813656994237,
      "train_speed(iter/s)": 0.032294
    },
    {
      "epoch": 0.9132393547449499,
      "grad_norm": 0.10096679627895355,
      "learning_rate": 6.140234616459483e-06,
      "loss": 0.33763110637664795,
      "memory(GiB)": 78.33,
      "step": 4713,
      "token_acc": 0.895675275944236,
      "train_speed(iter/s)": 0.032295
    },
    {
      "epoch": 0.9134331250302766,
      "grad_norm": 0.10386110842227936,
      "learning_rate": 6.113040965707256e-06,
      "loss": 0.30816999077796936,
      "memory(GiB)": 78.33,
      "step": 4714,
      "token_acc": 0.904618748291883,
      "train_speed(iter/s)": 0.032295
    },
    {
      "epoch": 0.9136268953156034,
      "grad_norm": 0.09449145197868347,
      "learning_rate": 6.085906413195546e-06,
      "loss": 0.30260464549064636,
      "memory(GiB)": 78.33,
      "step": 4715,
      "token_acc": 0.9089724446857608,
      "train_speed(iter/s)": 0.032296
    },
    {
      "epoch": 0.9138206656009301,
      "grad_norm": 0.1062229797244072,
      "learning_rate": 6.058830970069156e-06,
      "loss": 0.3432777523994446,
      "memory(GiB)": 78.33,
      "step": 4716,
      "token_acc": 0.8961035016855469,
      "train_speed(iter/s)": 0.032297
    },
    {
      "epoch": 0.9140144358862569,
      "grad_norm": 0.10210466384887695,
      "learning_rate": 6.03181464744874e-06,
      "loss": 0.31807243824005127,
      "memory(GiB)": 78.33,
      "step": 4717,
      "token_acc": 0.9045282715354299,
      "train_speed(iter/s)": 0.032297
    },
    {
      "epoch": 0.9142082061715836,
      "grad_norm": 0.10354772955179214,
      "learning_rate": 6.00485745643064e-06,
      "loss": 0.3176378011703491,
      "memory(GiB)": 78.33,
      "step": 4718,
      "token_acc": 0.9056022642137684,
      "train_speed(iter/s)": 0.032298
    },
    {
      "epoch": 0.9144019764569103,
      "grad_norm": 0.10732652246952057,
      "learning_rate": 5.977959408086863e-06,
      "loss": 0.3399035930633545,
      "memory(GiB)": 78.33,
      "step": 4719,
      "token_acc": 0.8993071737036692,
      "train_speed(iter/s)": 0.032298
    },
    {
      "epoch": 0.9145957467422371,
      "grad_norm": 0.08825431764125824,
      "learning_rate": 5.951120513465207e-06,
      "loss": 0.2766910195350647,
      "memory(GiB)": 78.33,
      "step": 4720,
      "token_acc": 0.9151678797839868,
      "train_speed(iter/s)": 0.032299
    },
    {
      "epoch": 0.9147895170275638,
      "grad_norm": 0.08927903324365616,
      "learning_rate": 5.924340783589071e-06,
      "loss": 0.2686108350753784,
      "memory(GiB)": 78.33,
      "step": 4721,
      "token_acc": 0.917844274653106,
      "train_speed(iter/s)": 0.0323
    },
    {
      "epoch": 0.9149832873128906,
      "grad_norm": 0.09068787842988968,
      "learning_rate": 5.897620229457639e-06,
      "loss": 0.2903996407985687,
      "memory(GiB)": 78.33,
      "step": 4722,
      "token_acc": 0.9110736173485845,
      "train_speed(iter/s)": 0.0323
    },
    {
      "epoch": 0.9151770575982173,
      "grad_norm": 0.10634169727563858,
      "learning_rate": 5.870958862045782e-06,
      "loss": 0.3310278356075287,
      "memory(GiB)": 78.33,
      "step": 4723,
      "token_acc": 0.9006753122572163,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.9153708278835441,
      "grad_norm": 0.09822243452072144,
      "learning_rate": 5.844356692304009e-06,
      "loss": 0.316002756357193,
      "memory(GiB)": 78.33,
      "step": 4724,
      "token_acc": 0.9054254372961755,
      "train_speed(iter/s)": 0.032301
    },
    {
      "epoch": 0.9155645981688708,
      "grad_norm": 0.10107910633087158,
      "learning_rate": 5.817813731158544e-06,
      "loss": 0.3441876769065857,
      "memory(GiB)": 78.33,
      "step": 4725,
      "token_acc": 0.8951170707621944,
      "train_speed(iter/s)": 0.032302
    },
    {
      "epoch": 0.9157583684541976,
      "grad_norm": 0.10910195857286453,
      "learning_rate": 5.791329989511301e-06,
      "loss": 0.34992867708206177,
      "memory(GiB)": 78.33,
      "step": 4726,
      "token_acc": 0.8951081299872788,
      "train_speed(iter/s)": 0.032302
    },
    {
      "epoch": 0.9159521387395243,
      "grad_norm": 0.10190389305353165,
      "learning_rate": 5.764905478239895e-06,
      "loss": 0.3404694199562073,
      "memory(GiB)": 78.33,
      "step": 4727,
      "token_acc": 0.8978785857238158,
      "train_speed(iter/s)": 0.032303
    },
    {
      "epoch": 0.916145909024851,
      "grad_norm": 0.09621675312519073,
      "learning_rate": 5.7385402081975284e-06,
      "loss": 0.3171209990978241,
      "memory(GiB)": 78.33,
      "step": 4728,
      "token_acc": 0.9050346044239381,
      "train_speed(iter/s)": 0.032304
    },
    {
      "epoch": 0.9163396793101778,
      "grad_norm": 0.09636913985013962,
      "learning_rate": 5.712234190213172e-06,
      "loss": 0.317484587430954,
      "memory(GiB)": 78.33,
      "step": 4729,
      "token_acc": 0.9054396568531877,
      "train_speed(iter/s)": 0.032304
    },
    {
      "epoch": 0.9165334495955045,
      "grad_norm": 0.10614298284053802,
      "learning_rate": 5.685987435091399e-06,
      "loss": 0.3546099066734314,
      "memory(GiB)": 78.33,
      "step": 4730,
      "token_acc": 0.8933270676691729,
      "train_speed(iter/s)": 0.032305
    },
    {
      "epoch": 0.9167272198808313,
      "grad_norm": 0.1044028177857399,
      "learning_rate": 5.659799953612438e-06,
      "loss": 0.349134236574173,
      "memory(GiB)": 78.33,
      "step": 4731,
      "token_acc": 0.8962899543378996,
      "train_speed(iter/s)": 0.032305
    },
    {
      "epoch": 0.916920990166158,
      "grad_norm": 0.09492190182209015,
      "learning_rate": 5.633671756532232e-06,
      "loss": 0.2976999878883362,
      "memory(GiB)": 78.33,
      "step": 4732,
      "token_acc": 0.9110198151256294,
      "train_speed(iter/s)": 0.032306
    },
    {
      "epoch": 0.9171147604514848,
      "grad_norm": 0.10704229772090912,
      "learning_rate": 5.607602854582266e-06,
      "loss": 0.38176852464675903,
      "memory(GiB)": 78.33,
      "step": 4733,
      "token_acc": 0.8874598960061953,
      "train_speed(iter/s)": 0.032306
    },
    {
      "epoch": 0.9173085307368115,
      "grad_norm": 0.09060212224721909,
      "learning_rate": 5.581593258469841e-06,
      "loss": 0.30857348442077637,
      "memory(GiB)": 78.33,
      "step": 4734,
      "token_acc": 0.9065131425913003,
      "train_speed(iter/s)": 0.032307
    },
    {
      "epoch": 0.9175023010221383,
      "grad_norm": 0.10411375015974045,
      "learning_rate": 5.555642978877678e-06,
      "loss": 0.3482334315776825,
      "memory(GiB)": 78.33,
      "step": 4735,
      "token_acc": 0.8967308452374231,
      "train_speed(iter/s)": 0.032307
    },
    {
      "epoch": 0.917696071307465,
      "grad_norm": 0.11206993460655212,
      "learning_rate": 5.529752026464351e-06,
      "loss": 0.35386189818382263,
      "memory(GiB)": 78.33,
      "step": 4736,
      "token_acc": 0.8936384358071106,
      "train_speed(iter/s)": 0.032308
    },
    {
      "epoch": 0.9178898415927917,
      "grad_norm": 0.0973845049738884,
      "learning_rate": 5.5039204118639215e-06,
      "loss": 0.34648823738098145,
      "memory(GiB)": 78.33,
      "step": 4737,
      "token_acc": 0.8961357595583725,
      "train_speed(iter/s)": 0.032309
    },
    {
      "epoch": 0.9180836118781185,
      "grad_norm": 0.09547542035579681,
      "learning_rate": 5.478148145686151e-06,
      "loss": 0.32375043630599976,
      "memory(GiB)": 78.33,
      "step": 4738,
      "token_acc": 0.9037339556592765,
      "train_speed(iter/s)": 0.032309
    },
    {
      "epoch": 0.9182773821634452,
      "grad_norm": 0.10374420881271362,
      "learning_rate": 5.452435238516373e-06,
      "loss": 0.34971579909324646,
      "memory(GiB)": 78.33,
      "step": 4739,
      "token_acc": 0.8960960267766973,
      "train_speed(iter/s)": 0.03231
    },
    {
      "epoch": 0.918471152448772,
      "grad_norm": 0.098647341132164,
      "learning_rate": 5.426781700915573e-06,
      "loss": 0.33870449662208557,
      "memory(GiB)": 78.33,
      "step": 4740,
      "token_acc": 0.8994294051744353,
      "train_speed(iter/s)": 0.03231
    },
    {
      "epoch": 0.9186649227340987,
      "grad_norm": 0.1018560454249382,
      "learning_rate": 5.401187543420405e-06,
      "loss": 0.3089270293712616,
      "memory(GiB)": 78.33,
      "step": 4741,
      "token_acc": 0.9068556551923633,
      "train_speed(iter/s)": 0.032311
    },
    {
      "epoch": 0.9188586930194255,
      "grad_norm": 0.10648074001073837,
      "learning_rate": 5.375652776542994e-06,
      "loss": 0.33178946375846863,
      "memory(GiB)": 78.33,
      "step": 4742,
      "token_acc": 0.8994858134060215,
      "train_speed(iter/s)": 0.032312
    },
    {
      "epoch": 0.9190524633047522,
      "grad_norm": 0.09511461853981018,
      "learning_rate": 5.350177410771217e-06,
      "loss": 0.31413573026657104,
      "memory(GiB)": 78.33,
      "step": 4743,
      "token_acc": 0.9058918735479156,
      "train_speed(iter/s)": 0.032312
    },
    {
      "epoch": 0.919246233590079,
      "grad_norm": 0.10511302202939987,
      "learning_rate": 5.324761456568455e-06,
      "loss": 0.3475594222545624,
      "memory(GiB)": 78.33,
      "step": 4744,
      "token_acc": 0.8972395423451783,
      "train_speed(iter/s)": 0.032313
    },
    {
      "epoch": 0.9194400038754057,
      "grad_norm": 0.0897306576371193,
      "learning_rate": 5.29940492437374e-06,
      "loss": 0.30537813901901245,
      "memory(GiB)": 78.33,
      "step": 4745,
      "token_acc": 0.9067023393016396,
      "train_speed(iter/s)": 0.032313
    },
    {
      "epoch": 0.9196337741607324,
      "grad_norm": 0.10397239774465561,
      "learning_rate": 5.274107824601692e-06,
      "loss": 0.3287217617034912,
      "memory(GiB)": 78.33,
      "step": 4746,
      "token_acc": 0.9034060827107477,
      "train_speed(iter/s)": 0.032314
    },
    {
      "epoch": 0.9198275444460592,
      "grad_norm": 0.10913705080747604,
      "learning_rate": 5.248870167642466e-06,
      "loss": 0.37385323643684387,
      "memory(GiB)": 78.33,
      "step": 4747,
      "token_acc": 0.8894808050059192,
      "train_speed(iter/s)": 0.032315
    },
    {
      "epoch": 0.9200213147313859,
      "grad_norm": 0.09678813070058823,
      "learning_rate": 5.22369196386192e-06,
      "loss": 0.29931795597076416,
      "memory(GiB)": 78.33,
      "step": 4748,
      "token_acc": 0.9094277587871513,
      "train_speed(iter/s)": 0.032315
    },
    {
      "epoch": 0.9202150850167127,
      "grad_norm": 0.1072593703866005,
      "learning_rate": 5.198573223601332e-06,
      "loss": 0.33650124073028564,
      "memory(GiB)": 78.33,
      "step": 4749,
      "token_acc": 0.9004353969032035,
      "train_speed(iter/s)": 0.032316
    },
    {
      "epoch": 0.9204088553020394,
      "grad_norm": 0.10376523435115814,
      "learning_rate": 5.173513957177716e-06,
      "loss": 0.35175377130508423,
      "memory(GiB)": 78.33,
      "step": 4750,
      "token_acc": 0.8930210695128021,
      "train_speed(iter/s)": 0.032316
    },
    {
      "epoch": 0.9206026255873662,
      "grad_norm": 0.09530597925186157,
      "learning_rate": 5.148514174883539e-06,
      "loss": 0.3145235478878021,
      "memory(GiB)": 78.33,
      "step": 4751,
      "token_acc": 0.9033812729498164,
      "train_speed(iter/s)": 0.032317
    },
    {
      "epoch": 0.9207963958726929,
      "grad_norm": 0.09804235398769379,
      "learning_rate": 5.123573886986887e-06,
      "loss": 0.32991063594818115,
      "memory(GiB)": 78.33,
      "step": 4752,
      "token_acc": 0.90162206627079,
      "train_speed(iter/s)": 0.032318
    },
    {
      "epoch": 0.9209901661580197,
      "grad_norm": 0.09455542266368866,
      "learning_rate": 5.098693103731466e-06,
      "loss": 0.3050658106803894,
      "memory(GiB)": 78.33,
      "step": 4753,
      "token_acc": 0.9097025408482637,
      "train_speed(iter/s)": 0.032318
    },
    {
      "epoch": 0.9211839364433464,
      "grad_norm": 0.09994572401046753,
      "learning_rate": 5.073871835336402e-06,
      "loss": 0.3111793100833893,
      "memory(GiB)": 78.33,
      "step": 4754,
      "token_acc": 0.9068566536879021,
      "train_speed(iter/s)": 0.032319
    },
    {
      "epoch": 0.9213777067286731,
      "grad_norm": 0.0913093090057373,
      "learning_rate": 5.049110091996505e-06,
      "loss": 0.2801298499107361,
      "memory(GiB)": 78.33,
      "step": 4755,
      "token_acc": 0.9132923800810352,
      "train_speed(iter/s)": 0.032319
    },
    {
      "epoch": 0.9215714770139999,
      "grad_norm": 0.1138007864356041,
      "learning_rate": 5.024407883882059e-06,
      "loss": 0.32883748412132263,
      "memory(GiB)": 78.33,
      "step": 4756,
      "token_acc": 0.8995948792740237,
      "train_speed(iter/s)": 0.03232
    },
    {
      "epoch": 0.9217652472993266,
      "grad_norm": 0.10155371576547623,
      "learning_rate": 4.999765221138946e-06,
      "loss": 0.338005006313324,
      "memory(GiB)": 78.33,
      "step": 4757,
      "token_acc": 0.8987489089322083,
      "train_speed(iter/s)": 0.03232
    },
    {
      "epoch": 0.9219590175846534,
      "grad_norm": 0.09977617859840393,
      "learning_rate": 4.975182113888571e-06,
      "loss": 0.3259299397468567,
      "memory(GiB)": 78.33,
      "step": 4758,
      "token_acc": 0.9017150933652252,
      "train_speed(iter/s)": 0.032321
    },
    {
      "epoch": 0.9221527878699801,
      "grad_norm": 0.10653609782457352,
      "learning_rate": 4.950658572227856e-06,
      "loss": 0.3374466300010681,
      "memory(GiB)": 78.33,
      "step": 4759,
      "token_acc": 0.8985345429169574,
      "train_speed(iter/s)": 0.032322
    },
    {
      "epoch": 0.9223465581553069,
      "grad_norm": 0.11073584109544754,
      "learning_rate": 4.926194606229311e-06,
      "loss": 0.3533399701118469,
      "memory(GiB)": 78.33,
      "step": 4760,
      "token_acc": 0.8946103673189152,
      "train_speed(iter/s)": 0.032322
    },
    {
      "epoch": 0.9225403284406336,
      "grad_norm": 0.10663000494241714,
      "learning_rate": 4.901790225940916e-06,
      "loss": 0.3574818968772888,
      "memory(GiB)": 78.33,
      "step": 4761,
      "token_acc": 0.8953725220185329,
      "train_speed(iter/s)": 0.032323
    },
    {
      "epoch": 0.9227340987259603,
      "grad_norm": 0.08861353248357773,
      "learning_rate": 4.877445441386218e-06,
      "loss": 0.2784247100353241,
      "memory(GiB)": 78.33,
      "step": 4762,
      "token_acc": 0.9163010066405425,
      "train_speed(iter/s)": 0.032323
    },
    {
      "epoch": 0.9229278690112871,
      "grad_norm": 0.10696634650230408,
      "learning_rate": 4.853160262564271e-06,
      "loss": 0.34982746839523315,
      "memory(GiB)": 78.33,
      "step": 4763,
      "token_acc": 0.8978552430535907,
      "train_speed(iter/s)": 0.032324
    },
    {
      "epoch": 0.9231216392966138,
      "grad_norm": 0.09375255554914474,
      "learning_rate": 4.8289346994496434e-06,
      "loss": 0.3086707890033722,
      "memory(GiB)": 78.33,
      "step": 4764,
      "token_acc": 0.9071370640713706,
      "train_speed(iter/s)": 0.032324
    },
    {
      "epoch": 0.9233154095819406,
      "grad_norm": 0.10051163285970688,
      "learning_rate": 4.804768761992445e-06,
      "loss": 0.33212947845458984,
      "memory(GiB)": 78.33,
      "step": 4765,
      "token_acc": 0.9005619215513442,
      "train_speed(iter/s)": 0.032325
    },
    {
      "epoch": 0.9235091798672673,
      "grad_norm": 0.0955612063407898,
      "learning_rate": 4.780662460118234e-06,
      "loss": 0.31439507007598877,
      "memory(GiB)": 78.33,
      "step": 4766,
      "token_acc": 0.9068176100628931,
      "train_speed(iter/s)": 0.032326
    },
    {
      "epoch": 0.9237029501525941,
      "grad_norm": 0.09498281031847,
      "learning_rate": 4.756615803728192e-06,
      "loss": 0.31122851371765137,
      "memory(GiB)": 78.33,
      "step": 4767,
      "token_acc": 0.9053871107818846,
      "train_speed(iter/s)": 0.032326
    },
    {
      "epoch": 0.9238967204379208,
      "grad_norm": 0.10263197124004364,
      "learning_rate": 4.732628802698835e-06,
      "loss": 0.31166380643844604,
      "memory(GiB)": 78.33,
      "step": 4768,
      "token_acc": 0.9051054675202015,
      "train_speed(iter/s)": 0.032327
    },
    {
      "epoch": 0.9240904907232476,
      "grad_norm": 0.08937201648950577,
      "learning_rate": 4.708701466882348e-06,
      "loss": 0.2815870940685272,
      "memory(GiB)": 78.33,
      "step": 4769,
      "token_acc": 0.9132003898738179,
      "train_speed(iter/s)": 0.032327
    },
    {
      "epoch": 0.9242842610085743,
      "grad_norm": 0.09476013481616974,
      "learning_rate": 4.684833806106286e-06,
      "loss": 0.32919424772262573,
      "memory(GiB)": 78.33,
      "step": 4770,
      "token_acc": 0.9013624509493285,
      "train_speed(iter/s)": 0.032328
    },
    {
      "epoch": 0.924478031293901,
      "grad_norm": 0.09523847699165344,
      "learning_rate": 4.661025830173742e-06,
      "loss": 0.3461284935474396,
      "memory(GiB)": 78.33,
      "step": 4771,
      "token_acc": 0.8952430099455808,
      "train_speed(iter/s)": 0.032328
    },
    {
      "epoch": 0.9246718015792278,
      "grad_norm": 0.09518402069807053,
      "learning_rate": 4.63727754886331e-06,
      "loss": 0.30630695819854736,
      "memory(GiB)": 78.33,
      "step": 4772,
      "token_acc": 0.9087172109035788,
      "train_speed(iter/s)": 0.032329
    },
    {
      "epoch": 0.9248655718645545,
      "grad_norm": 0.09036926180124283,
      "learning_rate": 4.613588971929022e-06,
      "loss": 0.31284967064857483,
      "memory(GiB)": 78.33,
      "step": 4773,
      "token_acc": 0.9051664182811724,
      "train_speed(iter/s)": 0.03233
    },
    {
      "epoch": 0.9250593421498813,
      "grad_norm": 0.09823027998209,
      "learning_rate": 4.589960109100444e-06,
      "loss": 0.30616480112075806,
      "memory(GiB)": 78.33,
      "step": 4774,
      "token_acc": 0.904891304347826,
      "train_speed(iter/s)": 0.03233
    },
    {
      "epoch": 0.925253112435208,
      "grad_norm": 0.10480865091085434,
      "learning_rate": 4.566390970082562e-06,
      "loss": 0.32189154624938965,
      "memory(GiB)": 78.33,
      "step": 4775,
      "token_acc": 0.9040938430251407,
      "train_speed(iter/s)": 0.032331
    },
    {
      "epoch": 0.9254468827205348,
      "grad_norm": 0.09449231624603271,
      "learning_rate": 4.54288156455585e-06,
      "loss": 0.3326607942581177,
      "memory(GiB)": 78.33,
      "step": 4776,
      "token_acc": 0.9016650796510282,
      "train_speed(iter/s)": 0.032331
    },
    {
      "epoch": 0.9256406530058615,
      "grad_norm": 0.09553948044776917,
      "learning_rate": 4.519431902176285e-06,
      "loss": 0.318975567817688,
      "memory(GiB)": 78.33,
      "step": 4777,
      "token_acc": 0.9035005686844433,
      "train_speed(iter/s)": 0.032332
    },
    {
      "epoch": 0.9258344232911883,
      "grad_norm": 0.09830733388662338,
      "learning_rate": 4.496041992575227e-06,
      "loss": 0.3239940404891968,
      "memory(GiB)": 78.33,
      "step": 4778,
      "token_acc": 0.900301950436604,
      "train_speed(iter/s)": 0.032332
    },
    {
      "epoch": 0.926028193576515,
      "grad_norm": 0.09941928833723068,
      "learning_rate": 4.472711845359594e-06,
      "loss": 0.3222883641719818,
      "memory(GiB)": 78.33,
      "step": 4779,
      "token_acc": 0.9041993341209322,
      "train_speed(iter/s)": 0.032333
    },
    {
      "epoch": 0.9262219638618417,
      "grad_norm": 0.10502801835536957,
      "learning_rate": 4.449441470111653e-06,
      "loss": 0.35170549154281616,
      "memory(GiB)": 78.33,
      "step": 4780,
      "token_acc": 0.895190294701809,
      "train_speed(iter/s)": 0.032333
    },
    {
      "epoch": 0.9264157341471685,
      "grad_norm": 0.08964619785547256,
      "learning_rate": 4.426230876389208e-06,
      "loss": 0.2930634617805481,
      "memory(GiB)": 78.33,
      "step": 4781,
      "token_acc": 0.9114340692728378,
      "train_speed(iter/s)": 0.032334
    },
    {
      "epoch": 0.9266095044324952,
      "grad_norm": 0.09876509755849838,
      "learning_rate": 4.403080073725451e-06,
      "loss": 0.31545448303222656,
      "memory(GiB)": 78.33,
      "step": 4782,
      "token_acc": 0.9063012667079041,
      "train_speed(iter/s)": 0.032334
    },
    {
      "epoch": 0.926803274717822,
      "grad_norm": 0.10196632146835327,
      "learning_rate": 4.379989071629059e-06,
      "loss": 0.3357621133327484,
      "memory(GiB)": 78.33,
      "step": 4783,
      "token_acc": 0.9003710320186891,
      "train_speed(iter/s)": 0.032335
    },
    {
      "epoch": 0.9269970450031487,
      "grad_norm": 0.10657868534326553,
      "learning_rate": 4.356957879584111e-06,
      "loss": 0.3701134920120239,
      "memory(GiB)": 78.33,
      "step": 4784,
      "token_acc": 0.8883001847625258,
      "train_speed(iter/s)": 0.032336
    },
    {
      "epoch": 0.9271908152884755,
      "grad_norm": 0.09407640993595123,
      "learning_rate": 4.333986507050125e-06,
      "loss": 0.3258771002292633,
      "memory(GiB)": 78.33,
      "step": 4785,
      "token_acc": 0.9031313034290654,
      "train_speed(iter/s)": 0.032336
    },
    {
      "epoch": 0.9273845855738022,
      "grad_norm": 0.09928041696548462,
      "learning_rate": 4.311074963462119e-06,
      "loss": 0.34973931312561035,
      "memory(GiB)": 78.33,
      "step": 4786,
      "token_acc": 0.8954454830273685,
      "train_speed(iter/s)": 0.032337
    },
    {
      "epoch": 0.927578355859129,
      "grad_norm": 0.10799898207187653,
      "learning_rate": 4.2882232582304e-06,
      "loss": 0.34097224473953247,
      "memory(GiB)": 78.33,
      "step": 4787,
      "token_acc": 0.9003566184182924,
      "train_speed(iter/s)": 0.032337
    },
    {
      "epoch": 0.9277721261444557,
      "grad_norm": 0.09870768338441849,
      "learning_rate": 4.265431400740843e-06,
      "loss": 0.3039855659008026,
      "memory(GiB)": 78.33,
      "step": 4788,
      "token_acc": 0.907001012860256,
      "train_speed(iter/s)": 0.032338
    },
    {
      "epoch": 0.9279658964297824,
      "grad_norm": 0.0973040908575058,
      "learning_rate": 4.242699400354627e-06,
      "loss": 0.34012576937675476,
      "memory(GiB)": 78.33,
      "step": 4789,
      "token_acc": 0.8992445949466007,
      "train_speed(iter/s)": 0.032339
    },
    {
      "epoch": 0.9281596667151092,
      "grad_norm": 0.09325951337814331,
      "learning_rate": 4.220027266408432e-06,
      "loss": 0.30864644050598145,
      "memory(GiB)": 78.33,
      "step": 4790,
      "token_acc": 0.9057424223046425,
      "train_speed(iter/s)": 0.032339
    },
    {
      "epoch": 0.9283534370004359,
      "grad_norm": 0.2708342671394348,
      "learning_rate": 4.197415008214294e-06,
      "loss": 0.3377147316932678,
      "memory(GiB)": 78.33,
      "step": 4791,
      "token_acc": 0.897419232539876,
      "train_speed(iter/s)": 0.03234
    },
    {
      "epoch": 0.9285472072857627,
      "grad_norm": 0.10447484999895096,
      "learning_rate": 4.174862635059667e-06,
      "loss": 0.36220014095306396,
      "memory(GiB)": 78.33,
      "step": 4792,
      "token_acc": 0.8931025706286587,
      "train_speed(iter/s)": 0.03234
    },
    {
      "epoch": 0.9287409775710894,
      "grad_norm": 0.10496804118156433,
      "learning_rate": 4.152370156207457e-06,
      "loss": 0.3289092481136322,
      "memory(GiB)": 78.33,
      "step": 4793,
      "token_acc": 0.9012323899291816,
      "train_speed(iter/s)": 0.032341
    },
    {
      "epoch": 0.9289347478564162,
      "grad_norm": 0.09275829046964645,
      "learning_rate": 4.129937580895876e-06,
      "loss": 0.30164456367492676,
      "memory(GiB)": 78.33,
      "step": 4794,
      "token_acc": 0.9096051284813877,
      "train_speed(iter/s)": 0.032341
    },
    {
      "epoch": 0.9291285181417429,
      "grad_norm": 0.09018470346927643,
      "learning_rate": 4.107564918338635e-06,
      "loss": 0.2863234579563141,
      "memory(GiB)": 78.33,
      "step": 4795,
      "token_acc": 0.9150141643059491,
      "train_speed(iter/s)": 0.032342
    },
    {
      "epoch": 0.9293222884270697,
      "grad_norm": 0.09511598944664001,
      "learning_rate": 4.085252177724751e-06,
      "loss": 0.29719942808151245,
      "memory(GiB)": 78.33,
      "step": 4796,
      "token_acc": 0.9094084830157199,
      "train_speed(iter/s)": 0.032342
    },
    {
      "epoch": 0.9295160587123965,
      "grad_norm": 0.11475857347249985,
      "learning_rate": 4.062999368218678e-06,
      "loss": 0.38321453332901,
      "memory(GiB)": 78.33,
      "step": 4797,
      "token_acc": 0.8886180329842301,
      "train_speed(iter/s)": 0.032343
    },
    {
      "epoch": 0.9297098289977233,
      "grad_norm": 0.10678976774215698,
      "learning_rate": 4.040806498960236e-06,
      "loss": 0.3360079824924469,
      "memory(GiB)": 78.33,
      "step": 4798,
      "token_acc": 0.8982757019471875,
      "train_speed(iter/s)": 0.032344
    },
    {
      "epoch": 0.92990359928305,
      "grad_norm": 0.09599542617797852,
      "learning_rate": 4.0186735790646355e-06,
      "loss": 0.30816134810447693,
      "memory(GiB)": 78.33,
      "step": 4799,
      "token_acc": 0.9078661001140015,
      "train_speed(iter/s)": 0.032344
    },
    {
      "epoch": 0.9300973695683767,
      "grad_norm": 0.10428732633590698,
      "learning_rate": 3.996600617622503e-06,
      "loss": 0.3343293070793152,
      "memory(GiB)": 78.33,
      "step": 4800,
      "token_acc": 0.8990470605483067,
      "train_speed(iter/s)": 0.032345
    },
    {
      "epoch": 0.9302911398537035,
      "grad_norm": 0.10038130730390549,
      "learning_rate": 3.974587623699721e-06,
      "loss": 0.2978077828884125,
      "memory(GiB)": 78.33,
      "step": 4801,
      "token_acc": 0.9088284412239237,
      "train_speed(iter/s)": 0.032343
    },
    {
      "epoch": 0.9304849101390302,
      "grad_norm": 0.10180116444826126,
      "learning_rate": 3.9526346063376735e-06,
      "loss": 0.31616196036338806,
      "memory(GiB)": 78.33,
      "step": 4802,
      "token_acc": 0.9038762241790799,
      "train_speed(iter/s)": 0.032343
    },
    {
      "epoch": 0.930678680424357,
      "grad_norm": 0.10563566535711288,
      "learning_rate": 3.930741574553048e-06,
      "loss": 0.35403674840927124,
      "memory(GiB)": 78.33,
      "step": 4803,
      "token_acc": 0.8924158321943599,
      "train_speed(iter/s)": 0.032344
    },
    {
      "epoch": 0.9308724507096837,
      "grad_norm": 0.10199618339538574,
      "learning_rate": 3.908908537337868e-06,
      "loss": 0.35334300994873047,
      "memory(GiB)": 78.33,
      "step": 4804,
      "token_acc": 0.8947171598813293,
      "train_speed(iter/s)": 0.032344
    },
    {
      "epoch": 0.9310662209950105,
      "grad_norm": 0.08841365575790405,
      "learning_rate": 3.887135503659594e-06,
      "loss": 0.27694904804229736,
      "memory(GiB)": 78.33,
      "step": 4805,
      "token_acc": 0.9153520015775981,
      "train_speed(iter/s)": 0.032345
    },
    {
      "epoch": 0.9312599912803372,
      "grad_norm": 0.10381640493869781,
      "learning_rate": 3.8654224824609396e-06,
      "loss": 0.2975887358188629,
      "memory(GiB)": 78.33,
      "step": 4806,
      "token_acc": 0.9085034107542324,
      "train_speed(iter/s)": 0.032346
    },
    {
      "epoch": 0.931453761565664,
      "grad_norm": 0.10239759087562561,
      "learning_rate": 3.8437694826601025e-06,
      "loss": 0.34857773780822754,
      "memory(GiB)": 78.33,
      "step": 4807,
      "token_acc": 0.8943446895898057,
      "train_speed(iter/s)": 0.032346
    },
    {
      "epoch": 0.9316475318509907,
      "grad_norm": 0.12824027240276337,
      "learning_rate": 3.8221765131504714e-06,
      "loss": 0.32302480936050415,
      "memory(GiB)": 78.33,
      "step": 4808,
      "token_acc": 0.904497486048953,
      "train_speed(iter/s)": 0.032347
    },
    {
      "epoch": 0.9318413021363174,
      "grad_norm": 0.09815438836812973,
      "learning_rate": 3.8006435828009162e-06,
      "loss": 0.30667853355407715,
      "memory(GiB)": 78.33,
      "step": 4809,
      "token_acc": 0.907749177788046,
      "train_speed(iter/s)": 0.032347
    },
    {
      "epoch": 0.9320350724216442,
      "grad_norm": 0.09541095048189163,
      "learning_rate": 3.7791707004555802e-06,
      "loss": 0.3058740794658661,
      "memory(GiB)": 78.33,
      "step": 4810,
      "token_acc": 0.9084823790877341,
      "train_speed(iter/s)": 0.032348
    },
    {
      "epoch": 0.9322288427069709,
      "grad_norm": 0.10464680939912796,
      "learning_rate": 3.7577578749339255e-06,
      "loss": 0.3602639138698578,
      "memory(GiB)": 78.33,
      "step": 4811,
      "token_acc": 0.8935190262090141,
      "train_speed(iter/s)": 0.032348
    },
    {
      "epoch": 0.9324226129922977,
      "grad_norm": 0.10377727448940277,
      "learning_rate": 3.7364051150307993e-06,
      "loss": 0.33127960562705994,
      "memory(GiB)": 78.33,
      "step": 4812,
      "token_acc": 0.9014416334773097,
      "train_speed(iter/s)": 0.032349
    },
    {
      "epoch": 0.9326163832776244,
      "grad_norm": 0.09491308033466339,
      "learning_rate": 3.715112429516337e-06,
      "loss": 0.32317915558815,
      "memory(GiB)": 78.33,
      "step": 4813,
      "token_acc": 0.9034436015745158,
      "train_speed(iter/s)": 0.032349
    },
    {
      "epoch": 0.9328101535629512,
      "grad_norm": 0.10311052948236465,
      "learning_rate": 3.6938798271360594e-06,
      "loss": 0.3292618691921234,
      "memory(GiB)": 78.33,
      "step": 4814,
      "token_acc": 0.8985615698012804,
      "train_speed(iter/s)": 0.03235
    },
    {
      "epoch": 0.9330039238482779,
      "grad_norm": 0.0882132351398468,
      "learning_rate": 3.672707316610707e-06,
      "loss": 0.30853813886642456,
      "memory(GiB)": 78.33,
      "step": 4815,
      "token_acc": 0.9053889463403424,
      "train_speed(iter/s)": 0.032351
    },
    {
      "epoch": 0.9331976941336046,
      "grad_norm": 0.09400478005409241,
      "learning_rate": 3.6515949066364236e-06,
      "loss": 0.3131456971168518,
      "memory(GiB)": 78.33,
      "step": 4816,
      "token_acc": 0.9059575833888848,
      "train_speed(iter/s)": 0.032351
    },
    {
      "epoch": 0.9333914644189314,
      "grad_norm": 0.09559616446495056,
      "learning_rate": 3.6305426058846565e-06,
      "loss": 0.3201538324356079,
      "memory(GiB)": 78.33,
      "step": 4817,
      "token_acc": 0.9050924765675906,
      "train_speed(iter/s)": 0.032352
    },
    {
      "epoch": 0.9335852347042581,
      "grad_norm": 0.1044517233967781,
      "learning_rate": 3.6095504230021387e-06,
      "loss": 0.34499791264533997,
      "memory(GiB)": 78.33,
      "step": 4818,
      "token_acc": 0.8969500515691764,
      "train_speed(iter/s)": 0.032352
    },
    {
      "epoch": 0.9337790049895849,
      "grad_norm": 0.0850808322429657,
      "learning_rate": 3.5886183666109405e-06,
      "loss": 0.27842819690704346,
      "memory(GiB)": 78.33,
      "step": 4819,
      "token_acc": 0.9156751866093644,
      "train_speed(iter/s)": 0.032353
    },
    {
      "epoch": 0.9339727752749116,
      "grad_norm": 0.09250815957784653,
      "learning_rate": 3.567746445308367e-06,
      "loss": 0.30809280276298523,
      "memory(GiB)": 78.33,
      "step": 4820,
      "token_acc": 0.9050716781110709,
      "train_speed(iter/s)": 0.032353
    },
    {
      "epoch": 0.9341665455602384,
      "grad_norm": 0.09440483152866364,
      "learning_rate": 3.5469346676671616e-06,
      "loss": 0.2813994288444519,
      "memory(GiB)": 78.33,
      "step": 4821,
      "token_acc": 0.9127701474449607,
      "train_speed(iter/s)": 0.032354
    },
    {
      "epoch": 0.9343603158455651,
      "grad_norm": 0.09050828963518143,
      "learning_rate": 3.526183042235203e-06,
      "loss": 0.2954739034175873,
      "memory(GiB)": 78.33,
      "step": 4822,
      "token_acc": 0.9097762259958737,
      "train_speed(iter/s)": 0.032354
    },
    {
      "epoch": 0.9345540861308919,
      "grad_norm": 0.10041210055351257,
      "learning_rate": 3.5054915775357907e-06,
      "loss": 0.3556936979293823,
      "memory(GiB)": 78.33,
      "step": 4823,
      "token_acc": 0.8935934907970614,
      "train_speed(iter/s)": 0.032355
    },
    {
      "epoch": 0.9347478564162186,
      "grad_norm": 0.10176945477724075,
      "learning_rate": 3.4848602820674255e-06,
      "loss": 0.3064349889755249,
      "memory(GiB)": 78.33,
      "step": 4824,
      "token_acc": 0.9054652880354506,
      "train_speed(iter/s)": 0.032355
    },
    {
      "epoch": 0.9349416267015453,
      "grad_norm": 0.10641314834356308,
      "learning_rate": 3.464289164303963e-06,
      "loss": 0.32750222086906433,
      "memory(GiB)": 78.33,
      "step": 4825,
      "token_acc": 0.9015335861015222,
      "train_speed(iter/s)": 0.032356
    },
    {
      "epoch": 0.9351353969868721,
      "grad_norm": 0.09517652541399002,
      "learning_rate": 3.4437782326945274e-06,
      "loss": 0.31871315836906433,
      "memory(GiB)": 78.33,
      "step": 4826,
      "token_acc": 0.9053494708163051,
      "train_speed(iter/s)": 0.032357
    },
    {
      "epoch": 0.9353291672721988,
      "grad_norm": 0.09240631759166718,
      "learning_rate": 3.4233274956634803e-06,
      "loss": 0.2959226667881012,
      "memory(GiB)": 78.33,
      "step": 4827,
      "token_acc": 0.9104759299781182,
      "train_speed(iter/s)": 0.032357
    },
    {
      "epoch": 0.9355229375575256,
      "grad_norm": 0.10840350389480591,
      "learning_rate": 3.402936961610503e-06,
      "loss": 0.33262190222740173,
      "memory(GiB)": 78.33,
      "step": 4828,
      "token_acc": 0.9010374853331686,
      "train_speed(iter/s)": 0.032358
    },
    {
      "epoch": 0.9357167078428523,
      "grad_norm": 0.09562506526708603,
      "learning_rate": 3.3826066389105123e-06,
      "loss": 0.3282950818538666,
      "memory(GiB)": 78.33,
      "step": 4829,
      "token_acc": 0.9013872354899055,
      "train_speed(iter/s)": 0.032358
    },
    {
      "epoch": 0.9359104781281791,
      "grad_norm": 0.1005750298500061,
      "learning_rate": 3.3623365359137453e-06,
      "loss": 0.3179299831390381,
      "memory(GiB)": 78.33,
      "step": 4830,
      "token_acc": 0.9042720884875364,
      "train_speed(iter/s)": 0.032359
    },
    {
      "epoch": 0.9361042484135058,
      "grad_norm": 0.09987672418355942,
      "learning_rate": 3.3421266609456766e-06,
      "loss": 0.27341562509536743,
      "memory(GiB)": 78.33,
      "step": 4831,
      "token_acc": 0.9170077307867213,
      "train_speed(iter/s)": 0.032359
    },
    {
      "epoch": 0.9362980186988326,
      "grad_norm": 0.10075198113918304,
      "learning_rate": 3.321977022307032e-06,
      "loss": 0.3398180305957794,
      "memory(GiB)": 78.33,
      "step": 4832,
      "token_acc": 0.8986838767860665,
      "train_speed(iter/s)": 0.03236
    },
    {
      "epoch": 0.9364917889841593,
      "grad_norm": 0.10808294266462326,
      "learning_rate": 3.301887628273825e-06,
      "loss": 0.35527899861335754,
      "memory(GiB)": 78.33,
      "step": 4833,
      "token_acc": 0.8941093763730631,
      "train_speed(iter/s)": 0.032361
    },
    {
      "epoch": 0.936685559269486,
      "grad_norm": 0.09991144388914108,
      "learning_rate": 3.2818584870972887e-06,
      "loss": 0.319998562335968,
      "memory(GiB)": 78.33,
      "step": 4834,
      "token_acc": 0.902095910695979,
      "train_speed(iter/s)": 0.032361
    },
    {
      "epoch": 0.9368793295548128,
      "grad_norm": 0.10339634865522385,
      "learning_rate": 3.2618896070039422e-06,
      "loss": 0.3507855236530304,
      "memory(GiB)": 78.33,
      "step": 4835,
      "token_acc": 0.8963165437078712,
      "train_speed(iter/s)": 0.032362
    },
    {
      "epoch": 0.9370730998401395,
      "grad_norm": 0.10384315997362137,
      "learning_rate": 3.241980996195559e-06,
      "loss": 0.34698525071144104,
      "memory(GiB)": 78.33,
      "step": 4836,
      "token_acc": 0.8970609075192019,
      "train_speed(iter/s)": 0.032362
    },
    {
      "epoch": 0.9372668701254663,
      "grad_norm": 0.0977269858121872,
      "learning_rate": 3.2221326628490973e-06,
      "loss": 0.33199065923690796,
      "memory(GiB)": 78.33,
      "step": 4837,
      "token_acc": 0.900308274874384,
      "train_speed(iter/s)": 0.032363
    },
    {
      "epoch": 0.937460640410793,
      "grad_norm": 0.11115771532058716,
      "learning_rate": 3.2023446151168363e-06,
      "loss": 0.35356926918029785,
      "memory(GiB)": 78.33,
      "step": 4838,
      "token_acc": 0.8959869212498843,
      "train_speed(iter/s)": 0.032363
    },
    {
      "epoch": 0.9376544106961198,
      "grad_norm": 0.09477823227643967,
      "learning_rate": 3.1826168611262417e-06,
      "loss": 0.30472081899642944,
      "memory(GiB)": 78.33,
      "step": 4839,
      "token_acc": 0.9091226468297661,
      "train_speed(iter/s)": 0.032364
    },
    {
      "epoch": 0.9378481809814465,
      "grad_norm": 0.09087901562452316,
      "learning_rate": 3.162949408980048e-06,
      "loss": 0.295918732881546,
      "memory(GiB)": 78.33,
      "step": 4840,
      "token_acc": 0.9117901101274023,
      "train_speed(iter/s)": 0.032364
    },
    {
      "epoch": 0.9380419512667733,
      "grad_norm": 0.10526914894580841,
      "learning_rate": 3.143342266756177e-06,
      "loss": 0.3462876081466675,
      "memory(GiB)": 78.33,
      "step": 4841,
      "token_acc": 0.8977767847167915,
      "train_speed(iter/s)": 0.032365
    },
    {
      "epoch": 0.9382357215521,
      "grad_norm": 0.08861321955919266,
      "learning_rate": 3.1237954425078537e-06,
      "loss": 0.30201205611228943,
      "memory(GiB)": 78.33,
      "step": 4842,
      "token_acc": 0.9064061563062162,
      "train_speed(iter/s)": 0.032366
    },
    {
      "epoch": 0.9384294918374267,
      "grad_norm": 0.09867441654205322,
      "learning_rate": 3.1043089442634394e-06,
      "loss": 0.3132460415363312,
      "memory(GiB)": 78.33,
      "step": 4843,
      "token_acc": 0.9052049214169717,
      "train_speed(iter/s)": 0.032366
    },
    {
      "epoch": 0.9386232621227535,
      "grad_norm": 0.09859292209148407,
      "learning_rate": 3.0848827800265817e-06,
      "loss": 0.3280273675918579,
      "memory(GiB)": 78.33,
      "step": 4844,
      "token_acc": 0.900389837343729,
      "train_speed(iter/s)": 0.032367
    },
    {
      "epoch": 0.9388170324080802,
      "grad_norm": 0.10763488709926605,
      "learning_rate": 3.0655169577761483e-06,
      "loss": 0.3330923914909363,
      "memory(GiB)": 78.33,
      "step": 4845,
      "token_acc": 0.8993531614620408,
      "train_speed(iter/s)": 0.032367
    },
    {
      "epoch": 0.939010802693407,
      "grad_norm": 0.10462969541549683,
      "learning_rate": 3.04621148546616e-06,
      "loss": 0.34643369913101196,
      "memory(GiB)": 78.33,
      "step": 4846,
      "token_acc": 0.8975769701011448,
      "train_speed(iter/s)": 0.032368
    },
    {
      "epoch": 0.9392045729787337,
      "grad_norm": 0.10669691115617752,
      "learning_rate": 3.0269663710259405e-06,
      "loss": 0.3267556130886078,
      "memory(GiB)": 78.33,
      "step": 4847,
      "token_acc": 0.9032104437471986,
      "train_speed(iter/s)": 0.032368
    },
    {
      "epoch": 0.9393983432640605,
      "grad_norm": 0.09605014324188232,
      "learning_rate": 3.007781622359934e-06,
      "loss": 0.31911730766296387,
      "memory(GiB)": 78.33,
      "step": 4848,
      "token_acc": 0.9046072137571867,
      "train_speed(iter/s)": 0.032369
    },
    {
      "epoch": 0.9395921135493872,
      "grad_norm": 0.09630642086267471,
      "learning_rate": 2.988657247347853e-06,
      "loss": 0.31961339712142944,
      "memory(GiB)": 78.33,
      "step": 4849,
      "token_acc": 0.9028399069553437,
      "train_speed(iter/s)": 0.032369
    },
    {
      "epoch": 0.939785883834714,
      "grad_norm": 0.09434531629085541,
      "learning_rate": 2.969593253844582e-06,
      "loss": 0.29251518845558167,
      "memory(GiB)": 78.33,
      "step": 4850,
      "token_acc": 0.9108287448821121,
      "train_speed(iter/s)": 0.03237
    },
    {
      "epoch": 0.9399796541200407,
      "grad_norm": 0.09503661096096039,
      "learning_rate": 2.950589649680224e-06,
      "loss": 0.327890545129776,
      "memory(GiB)": 78.33,
      "step": 4851,
      "token_acc": 0.9029296643340055,
      "train_speed(iter/s)": 0.032371
    },
    {
      "epoch": 0.9401734244053674,
      "grad_norm": 0.09823209792375565,
      "learning_rate": 2.931646442660085e-06,
      "loss": 0.3063351511955261,
      "memory(GiB)": 78.33,
      "step": 4852,
      "token_acc": 0.9096577564030869,
      "train_speed(iter/s)": 0.032371
    },
    {
      "epoch": 0.9403671946906942,
      "grad_norm": 0.10142441838979721,
      "learning_rate": 2.912763640564608e-06,
      "loss": 0.3455069363117218,
      "memory(GiB)": 78.33,
      "step": 4853,
      "token_acc": 0.8964419326593966,
      "train_speed(iter/s)": 0.032372
    },
    {
      "epoch": 0.9405609649760209,
      "grad_norm": 0.09713559597730637,
      "learning_rate": 2.8939412511495066e-06,
      "loss": 0.3110349178314209,
      "memory(GiB)": 78.33,
      "step": 4854,
      "token_acc": 0.9043400246719748,
      "train_speed(iter/s)": 0.032372
    },
    {
      "epoch": 0.9407547352613477,
      "grad_norm": 0.09745633602142334,
      "learning_rate": 2.875179282145612e-06,
      "loss": 0.33706194162368774,
      "memory(GiB)": 78.33,
      "step": 4855,
      "token_acc": 0.9004770564824333,
      "train_speed(iter/s)": 0.032373
    },
    {
      "epoch": 0.9409485055466744,
      "grad_norm": 0.10170414298772812,
      "learning_rate": 2.8564777412589944e-06,
      "loss": 0.33610811829566956,
      "memory(GiB)": 78.33,
      "step": 4856,
      "token_acc": 0.9004509018036072,
      "train_speed(iter/s)": 0.032373
    },
    {
      "epoch": 0.9411422758320012,
      "grad_norm": 0.09948156028985977,
      "learning_rate": 2.8378366361708593e-06,
      "loss": 0.3124832808971405,
      "memory(GiB)": 78.33,
      "step": 4857,
      "token_acc": 0.9069445621169759,
      "train_speed(iter/s)": 0.032374
    },
    {
      "epoch": 0.9413360461173279,
      "grad_norm": 0.10482876747846603,
      "learning_rate": 2.8192559745376152e-06,
      "loss": 0.35402458906173706,
      "memory(GiB)": 78.33,
      "step": 4858,
      "token_acc": 0.8951886335504009,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.9415298164026547,
      "grad_norm": 0.12244974076747894,
      "learning_rate": 2.8007357639908743e-06,
      "loss": 0.34525981545448303,
      "memory(GiB)": 78.33,
      "step": 4859,
      "token_acc": 0.8989004930936265,
      "train_speed(iter/s)": 0.032375
    },
    {
      "epoch": 0.9417235866879814,
      "grad_norm": 0.10524637997150421,
      "learning_rate": 2.7822760121373187e-06,
      "loss": 0.3427123427391052,
      "memory(GiB)": 78.33,
      "step": 4860,
      "token_acc": 0.8977062523122457,
      "train_speed(iter/s)": 0.032376
    },
    {
      "epoch": 0.9419173569733081,
      "grad_norm": 0.10324079543352127,
      "learning_rate": 2.7638767265589168e-06,
      "loss": 0.3470621109008789,
      "memory(GiB)": 78.33,
      "step": 4861,
      "token_acc": 0.8988676079966448,
      "train_speed(iter/s)": 0.032376
    },
    {
      "epoch": 0.9421111272586349,
      "grad_norm": 0.10326438397169113,
      "learning_rate": 2.7455379148127064e-06,
      "loss": 0.3370799422264099,
      "memory(GiB)": 78.33,
      "step": 4862,
      "token_acc": 0.8991185653891851,
      "train_speed(iter/s)": 0.032377
    },
    {
      "epoch": 0.9423048975439616,
      "grad_norm": 0.10224196314811707,
      "learning_rate": 2.7272595844309797e-06,
      "loss": 0.3339768052101135,
      "memory(GiB)": 78.33,
      "step": 4863,
      "token_acc": 0.9005996573386637,
      "train_speed(iter/s)": 0.032377
    },
    {
      "epoch": 0.9424986678292884,
      "grad_norm": 0.09687656909227371,
      "learning_rate": 2.7090417429211143e-06,
      "loss": 0.316772997379303,
      "memory(GiB)": 78.33,
      "step": 4864,
      "token_acc": 0.9048740035311112,
      "train_speed(iter/s)": 0.032378
    },
    {
      "epoch": 0.9426924381146151,
      "grad_norm": 0.10042627900838852,
      "learning_rate": 2.6908843977656415e-06,
      "loss": 0.3032311499118805,
      "memory(GiB)": 78.33,
      "step": 4865,
      "token_acc": 0.9067732763020391,
      "train_speed(iter/s)": 0.032378
    },
    {
      "epoch": 0.9428862083999419,
      "grad_norm": 0.10065774619579315,
      "learning_rate": 2.6727875564223287e-06,
      "loss": 0.32765907049179077,
      "memory(GiB)": 78.33,
      "step": 4866,
      "token_acc": 0.9011883622455948,
      "train_speed(iter/s)": 0.032379
    },
    {
      "epoch": 0.9430799786852686,
      "grad_norm": 0.10036752372980118,
      "learning_rate": 2.654751226323981e-06,
      "loss": 0.3179859519004822,
      "memory(GiB)": 78.33,
      "step": 4867,
      "token_acc": 0.9036099036099036,
      "train_speed(iter/s)": 0.03238
    },
    {
      "epoch": 0.9432737489705953,
      "grad_norm": 0.09917581081390381,
      "learning_rate": 2.6367754148786225e-06,
      "loss": 0.3333044648170471,
      "memory(GiB)": 78.33,
      "step": 4868,
      "token_acc": 0.8983144059174526,
      "train_speed(iter/s)": 0.03238
    },
    {
      "epoch": 0.9434675192559221,
      "grad_norm": 0.09930333495140076,
      "learning_rate": 2.6188601294694135e-06,
      "loss": 0.3436489999294281,
      "memory(GiB)": 78.33,
      "step": 4869,
      "token_acc": 0.8961160505381376,
      "train_speed(iter/s)": 0.032381
    },
    {
      "epoch": 0.9436612895412488,
      "grad_norm": 0.10882118344306946,
      "learning_rate": 2.601005377454635e-06,
      "loss": 0.3218170702457428,
      "memory(GiB)": 78.33,
      "step": 4870,
      "token_acc": 0.9025652352254441,
      "train_speed(iter/s)": 0.032381
    },
    {
      "epoch": 0.9438550598265756,
      "grad_norm": 0.10568146407604218,
      "learning_rate": 2.5832111661677203e-06,
      "loss": 0.3271356225013733,
      "memory(GiB)": 78.33,
      "step": 4871,
      "token_acc": 0.9018029976218129,
      "train_speed(iter/s)": 0.032382
    },
    {
      "epoch": 0.9440488301119023,
      "grad_norm": 0.11422929167747498,
      "learning_rate": 2.5654775029171903e-06,
      "loss": 0.3454614281654358,
      "memory(GiB)": 78.33,
      "step": 4872,
      "token_acc": 0.8961740297778863,
      "train_speed(iter/s)": 0.032382
    },
    {
      "epoch": 0.9442426003972291,
      "grad_norm": 0.10132316499948502,
      "learning_rate": 2.547804394986819e-06,
      "loss": 0.3422181010246277,
      "memory(GiB)": 78.33,
      "step": 4873,
      "token_acc": 0.8978968880151073,
      "train_speed(iter/s)": 0.032383
    },
    {
      "epoch": 0.9444363706825558,
      "grad_norm": 0.09934362024068832,
      "learning_rate": 2.5301918496353322e-06,
      "loss": 0.31417742371559143,
      "memory(GiB)": 78.33,
      "step": 4874,
      "token_acc": 0.9052588444201611,
      "train_speed(iter/s)": 0.032383
    },
    {
      "epoch": 0.9446301409678826,
      "grad_norm": 0.10042758285999298,
      "learning_rate": 2.5126398740967446e-06,
      "loss": 0.3282407522201538,
      "memory(GiB)": 78.33,
      "step": 4875,
      "token_acc": 0.901354957914186,
      "train_speed(iter/s)": 0.032384
    },
    {
      "epoch": 0.9448239112532093,
      "grad_norm": 0.09526004642248154,
      "learning_rate": 2.4951484755800886e-06,
      "loss": 0.29993996024131775,
      "memory(GiB)": 78.33,
      "step": 4876,
      "token_acc": 0.9110852110852111,
      "train_speed(iter/s)": 0.032384
    },
    {
      "epoch": 0.945017681538536,
      "grad_norm": 0.21196365356445312,
      "learning_rate": 2.4777176612695513e-06,
      "loss": 0.3271010220050812,
      "memory(GiB)": 78.33,
      "step": 4877,
      "token_acc": 0.9029842588543944,
      "train_speed(iter/s)": 0.032385
    },
    {
      "epoch": 0.9452114518238628,
      "grad_norm": 0.10081423819065094,
      "learning_rate": 2.4603474383244724e-06,
      "loss": 0.32736343145370483,
      "memory(GiB)": 78.33,
      "step": 4878,
      "token_acc": 0.904177119187912,
      "train_speed(iter/s)": 0.032386
    },
    {
      "epoch": 0.9454052221091895,
      "grad_norm": 0.10198832303285599,
      "learning_rate": 2.4430378138792282e-06,
      "loss": 0.3308171331882477,
      "memory(GiB)": 78.33,
      "step": 4879,
      "token_acc": 0.899286116294914,
      "train_speed(iter/s)": 0.032386
    },
    {
      "epoch": 0.9455989923945163,
      "grad_norm": 0.10578960925340652,
      "learning_rate": 2.425788795043382e-06,
      "loss": 0.3238624930381775,
      "memory(GiB)": 78.33,
      "step": 4880,
      "token_acc": 0.904650030083374,
      "train_speed(iter/s)": 0.032387
    },
    {
      "epoch": 0.945792762679843,
      "grad_norm": 0.09681179374456406,
      "learning_rate": 2.4086003889015326e-06,
      "loss": 0.2874050438404083,
      "memory(GiB)": 78.33,
      "step": 4881,
      "token_acc": 0.9121931027345872,
      "train_speed(iter/s)": 0.032387
    },
    {
      "epoch": 0.9459865329651698,
      "grad_norm": 0.10076259076595306,
      "learning_rate": 2.3914726025134335e-06,
      "loss": 0.3133833706378937,
      "memory(GiB)": 78.33,
      "step": 4882,
      "token_acc": 0.9046173554639635,
      "train_speed(iter/s)": 0.032388
    },
    {
      "epoch": 0.9461803032504965,
      "grad_norm": 0.09847967326641083,
      "learning_rate": 2.3744054429139402e-06,
      "loss": 0.31684496998786926,
      "memory(GiB)": 78.33,
      "step": 4883,
      "token_acc": 0.903607284929505,
      "train_speed(iter/s)": 0.032389
    },
    {
      "epoch": 0.9463740735358233,
      "grad_norm": 0.09512414038181305,
      "learning_rate": 2.3573989171129792e-06,
      "loss": 0.3404044806957245,
      "memory(GiB)": 78.33,
      "step": 4884,
      "token_acc": 0.8973556187081547,
      "train_speed(iter/s)": 0.032389
    },
    {
      "epoch": 0.94656784382115,
      "grad_norm": 0.111565001308918,
      "learning_rate": 2.340453032095613e-06,
      "loss": 0.36841219663619995,
      "memory(GiB)": 78.33,
      "step": 4885,
      "token_acc": 0.8914011037920598,
      "train_speed(iter/s)": 0.03239
    },
    {
      "epoch": 0.9467616141064767,
      "grad_norm": 0.09721893817186356,
      "learning_rate": 2.3235677948219234e-06,
      "loss": 0.3191075921058655,
      "memory(GiB)": 78.33,
      "step": 4886,
      "token_acc": 0.9030001088968747,
      "train_speed(iter/s)": 0.03239
    },
    {
      "epoch": 0.9469553843918035,
      "grad_norm": 0.09971962869167328,
      "learning_rate": 2.3067432122271966e-06,
      "loss": 0.33436092734336853,
      "memory(GiB)": 78.33,
      "step": 4887,
      "token_acc": 0.8994865769240593,
      "train_speed(iter/s)": 0.032391
    },
    {
      "epoch": 0.9471491546771302,
      "grad_norm": 0.103705994784832,
      "learning_rate": 2.289979291221672e-06,
      "loss": 0.3510754108428955,
      "memory(GiB)": 78.33,
      "step": 4888,
      "token_acc": 0.8964129530759531,
      "train_speed(iter/s)": 0.032391
    },
    {
      "epoch": 0.947342924962457,
      "grad_norm": 0.10477303713560104,
      "learning_rate": 2.273276038690791e-06,
      "loss": 0.3268052637577057,
      "memory(GiB)": 78.33,
      "step": 4889,
      "token_acc": 0.9006867252517022,
      "train_speed(iter/s)": 0.032392
    },
    {
      "epoch": 0.9475366952477837,
      "grad_norm": 0.10907679796218872,
      "learning_rate": 2.256633461495e-06,
      "loss": 0.35595622658729553,
      "memory(GiB)": 78.33,
      "step": 4890,
      "token_acc": 0.8929480987090542,
      "train_speed(iter/s)": 0.032392
    },
    {
      "epoch": 0.9477304655331105,
      "grad_norm": 0.09888530522584915,
      "learning_rate": 2.240051566469864e-06,
      "loss": 0.3075358271598816,
      "memory(GiB)": 78.33,
      "step": 4891,
      "token_acc": 0.9061078252957234,
      "train_speed(iter/s)": 0.032393
    },
    {
      "epoch": 0.9479242358184372,
      "grad_norm": 0.09880520403385162,
      "learning_rate": 2.2235303604260347e-06,
      "loss": 0.34165963530540466,
      "memory(GiB)": 78.33,
      "step": 4892,
      "token_acc": 0.8993901612684646,
      "train_speed(iter/s)": 0.032393
    },
    {
      "epoch": 0.948118006103764,
      "grad_norm": 0.09279303252696991,
      "learning_rate": 2.207069850149168e-06,
      "loss": 0.31139740347862244,
      "memory(GiB)": 78.33,
      "step": 4893,
      "token_acc": 0.906871677108554,
      "train_speed(iter/s)": 0.032394
    },
    {
      "epoch": 0.9483117763890907,
      "grad_norm": 0.10540986061096191,
      "learning_rate": 2.190670042400089e-06,
      "loss": 0.3575151264667511,
      "memory(GiB)": 78.33,
      "step": 4894,
      "token_acc": 0.8956390443200994,
      "train_speed(iter/s)": 0.032394
    },
    {
      "epoch": 0.9485055466744174,
      "grad_norm": 0.10701259970664978,
      "learning_rate": 2.174330943914593e-06,
      "loss": 0.3582616150379181,
      "memory(GiB)": 78.33,
      "step": 4895,
      "token_acc": 0.8956076759061834,
      "train_speed(iter/s)": 0.032395
    },
    {
      "epoch": 0.9486993169597442,
      "grad_norm": 0.09107998758554459,
      "learning_rate": 2.1580525614036115e-06,
      "loss": 0.325961172580719,
      "memory(GiB)": 78.33,
      "step": 4896,
      "token_acc": 0.9035096153846154,
      "train_speed(iter/s)": 0.032396
    },
    {
      "epoch": 0.9488930872450709,
      "grad_norm": 0.09532765299081802,
      "learning_rate": 2.141834901553113e-06,
      "loss": 0.304240345954895,
      "memory(GiB)": 78.33,
      "step": 4897,
      "token_acc": 0.9070364131139044,
      "train_speed(iter/s)": 0.032396
    },
    {
      "epoch": 0.9490868575303977,
      "grad_norm": 0.0961633175611496,
      "learning_rate": 2.12567797102412e-06,
      "loss": 0.317226380109787,
      "memory(GiB)": 78.33,
      "step": 4898,
      "token_acc": 0.9035757011830386,
      "train_speed(iter/s)": 0.032397
    },
    {
      "epoch": 0.9492806278157244,
      "grad_norm": 0.092696912586689,
      "learning_rate": 2.10958177645274e-06,
      "loss": 0.31177228689193726,
      "memory(GiB)": 78.33,
      "step": 4899,
      "token_acc": 0.9065967318942909,
      "train_speed(iter/s)": 0.032397
    },
    {
      "epoch": 0.9494743981010512,
      "grad_norm": 0.09380333125591278,
      "learning_rate": 2.0935463244500683e-06,
      "loss": 0.3175070583820343,
      "memory(GiB)": 78.33,
      "step": 4900,
      "token_acc": 0.9043746832999542,
      "train_speed(iter/s)": 0.032398
    },
    {
      "epoch": 0.9496681683863779,
      "grad_norm": 0.09854871779680252,
      "learning_rate": 2.077571621602353e-06,
      "loss": 0.3263704180717468,
      "memory(GiB)": 78.33,
      "step": 4901,
      "token_acc": 0.9024825089818742,
      "train_speed(iter/s)": 0.032398
    },
    {
      "epoch": 0.9498619386717047,
      "grad_norm": 0.11277089267969131,
      "learning_rate": 2.0616576744707624e-06,
      "loss": 0.35801947116851807,
      "memory(GiB)": 78.33,
      "step": 4902,
      "token_acc": 0.895363334124911,
      "train_speed(iter/s)": 0.032399
    },
    {
      "epoch": 0.9500557089570314,
      "grad_norm": 0.09970259666442871,
      "learning_rate": 2.0458044895916513e-06,
      "loss": 0.31966596841812134,
      "memory(GiB)": 78.33,
      "step": 4903,
      "token_acc": 0.9028877820418606,
      "train_speed(iter/s)": 0.032399
    },
    {
      "epoch": 0.9502494792423581,
      "grad_norm": 0.10125202685594559,
      "learning_rate": 2.0300120734763113e-06,
      "loss": 0.3093150556087494,
      "memory(GiB)": 78.33,
      "step": 4904,
      "token_acc": 0.905799933852491,
      "train_speed(iter/s)": 0.0324
    },
    {
      "epoch": 0.9504432495276849,
      "grad_norm": 0.09156003594398499,
      "learning_rate": 2.014280432611104e-06,
      "loss": 0.29941630363464355,
      "memory(GiB)": 78.33,
      "step": 4905,
      "token_acc": 0.9120291488636662,
      "train_speed(iter/s)": 0.032401
    },
    {
      "epoch": 0.9506370198130116,
      "grad_norm": 0.10564015060663223,
      "learning_rate": 1.998609573457477e-06,
      "loss": 0.3411896824836731,
      "memory(GiB)": 78.33,
      "step": 4906,
      "token_acc": 0.8986840243384746,
      "train_speed(iter/s)": 0.032401
    },
    {
      "epoch": 0.9508307900983384,
      "grad_norm": 0.10158051550388336,
      "learning_rate": 1.982999502451832e-06,
      "loss": 0.33457452058792114,
      "memory(GiB)": 78.33,
      "step": 4907,
      "token_acc": 0.8990842883153047,
      "train_speed(iter/s)": 0.032402
    },
    {
      "epoch": 0.9510245603836651,
      "grad_norm": 0.09677699208259583,
      "learning_rate": 1.9674502260056733e-06,
      "loss": 0.3045703172683716,
      "memory(GiB)": 78.33,
      "step": 4908,
      "token_acc": 0.9080135296660817,
      "train_speed(iter/s)": 0.032402
    },
    {
      "epoch": 0.9512183306689919,
      "grad_norm": 0.09688906371593475,
      "learning_rate": 1.9519617505055098e-06,
      "loss": 0.3192395865917206,
      "memory(GiB)": 78.33,
      "step": 4909,
      "token_acc": 0.9055065341226405,
      "train_speed(iter/s)": 0.032403
    },
    {
      "epoch": 0.9514121009543186,
      "grad_norm": 0.1135900616645813,
      "learning_rate": 1.936534082312835e-06,
      "loss": 0.35068824887275696,
      "memory(GiB)": 78.33,
      "step": 4910,
      "token_acc": 0.898571652801758,
      "train_speed(iter/s)": 0.032403
    },
    {
      "epoch": 0.9516058712396454,
      "grad_norm": 0.09443049877882004,
      "learning_rate": 1.9211672277642475e-06,
      "loss": 0.30319830775260925,
      "memory(GiB)": 78.33,
      "step": 4911,
      "token_acc": 0.906641655112416,
      "train_speed(iter/s)": 0.032404
    },
    {
      "epoch": 0.9517996415249721,
      "grad_norm": 0.10316184908151627,
      "learning_rate": 1.9058611931712986e-06,
      "loss": 0.31675225496292114,
      "memory(GiB)": 78.33,
      "step": 4912,
      "token_acc": 0.9031579571039222,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.9519934118102988,
      "grad_norm": 0.10155156254768372,
      "learning_rate": 1.8906159848206092e-06,
      "loss": 0.30963414907455444,
      "memory(GiB)": 78.33,
      "step": 4913,
      "token_acc": 0.905279359704479,
      "train_speed(iter/s)": 0.032405
    },
    {
      "epoch": 0.9521871820956256,
      "grad_norm": 0.11472687125205994,
      "learning_rate": 1.8754316089737876e-06,
      "loss": 0.3777286112308502,
      "memory(GiB)": 78.33,
      "step": 4914,
      "token_acc": 0.8904687163389038,
      "train_speed(iter/s)": 0.032406
    },
    {
      "epoch": 0.9523809523809523,
      "grad_norm": 0.09569194912910461,
      "learning_rate": 1.8603080718674612e-06,
      "loss": 0.3033876419067383,
      "memory(GiB)": 78.33,
      "step": 4915,
      "token_acc": 0.9085435990308607,
      "train_speed(iter/s)": 0.032406
    },
    {
      "epoch": 0.9525747226662791,
      "grad_norm": 0.09913970530033112,
      "learning_rate": 1.8452453797132948e-06,
      "loss": 0.3156486749649048,
      "memory(GiB)": 78.33,
      "step": 4916,
      "token_acc": 0.9022938238862399,
      "train_speed(iter/s)": 0.032407
    },
    {
      "epoch": 0.9527684929516058,
      "grad_norm": 0.09308433532714844,
      "learning_rate": 1.8302435386978897e-06,
      "loss": 0.30741560459136963,
      "memory(GiB)": 78.33,
      "step": 4917,
      "token_acc": 0.907486671172186,
      "train_speed(iter/s)": 0.032407
    },
    {
      "epoch": 0.9529622632369327,
      "grad_norm": 0.09139852225780487,
      "learning_rate": 1.8153025549829836e-06,
      "loss": 0.32832008600234985,
      "memory(GiB)": 78.33,
      "step": 4918,
      "token_acc": 0.9023957409050577,
      "train_speed(iter/s)": 0.032408
    },
    {
      "epoch": 0.9531560335222594,
      "grad_norm": 0.11089830100536346,
      "learning_rate": 1.800422434705151e-06,
      "loss": 0.3461150527000427,
      "memory(GiB)": 78.33,
      "step": 4919,
      "token_acc": 0.8961630109366407,
      "train_speed(iter/s)": 0.032408
    },
    {
      "epoch": 0.9533498038075862,
      "grad_norm": 0.10732929408550262,
      "learning_rate": 1.7856031839761363e-06,
      "loss": 0.3544648289680481,
      "memory(GiB)": 78.33,
      "step": 4920,
      "token_acc": 0.8927505075839006,
      "train_speed(iter/s)": 0.032409
    },
    {
      "epoch": 0.9535435740929129,
      "grad_norm": 0.10486234724521637,
      "learning_rate": 1.7708448088825545e-06,
      "loss": 0.341609388589859,
      "memory(GiB)": 78.33,
      "step": 4921,
      "token_acc": 0.8979067310163635,
      "train_speed(iter/s)": 0.032409
    },
    {
      "epoch": 0.9537373443782396,
      "grad_norm": 0.10882783681154251,
      "learning_rate": 1.7561473154860728e-06,
      "loss": 0.34817302227020264,
      "memory(GiB)": 78.33,
      "step": 4922,
      "token_acc": 0.8951537080828513,
      "train_speed(iter/s)": 0.03241
    },
    {
      "epoch": 0.9539311146635664,
      "grad_norm": 0.09959416091442108,
      "learning_rate": 1.7415107098233628e-06,
      "loss": 0.3298611044883728,
      "memory(GiB)": 78.33,
      "step": 4923,
      "token_acc": 0.9004360465116279,
      "train_speed(iter/s)": 0.03241
    },
    {
      "epoch": 0.9541248849488931,
      "grad_norm": 0.08811165392398834,
      "learning_rate": 1.7269349979060654e-06,
      "loss": 0.28269749879837036,
      "memory(GiB)": 78.33,
      "step": 4924,
      "token_acc": 0.9127474034881442,
      "train_speed(iter/s)": 0.032411
    },
    {
      "epoch": 0.9543186552342199,
      "grad_norm": 0.10352582484483719,
      "learning_rate": 1.7124201857208252e-06,
      "loss": 0.3340131640434265,
      "memory(GiB)": 78.33,
      "step": 4925,
      "token_acc": 0.8997261368174816,
      "train_speed(iter/s)": 0.032412
    },
    {
      "epoch": 0.9545124255195466,
      "grad_norm": 0.10261107236146927,
      "learning_rate": 1.6979662792292404e-06,
      "loss": 0.32005396485328674,
      "memory(GiB)": 78.33,
      "step": 4926,
      "token_acc": 0.9025667147748818,
      "train_speed(iter/s)": 0.032412
    },
    {
      "epoch": 0.9547061958048734,
      "grad_norm": 0.09662662446498871,
      "learning_rate": 1.6835732843679451e-06,
      "loss": 0.32958412170410156,
      "memory(GiB)": 78.33,
      "step": 4927,
      "token_acc": 0.9002728335715213,
      "train_speed(iter/s)": 0.032413
    },
    {
      "epoch": 0.9548999660902001,
      "grad_norm": 0.09711389243602753,
      "learning_rate": 1.6692412070485106e-06,
      "loss": 0.3142922818660736,
      "memory(GiB)": 78.33,
      "step": 4928,
      "token_acc": 0.904241246186658,
      "train_speed(iter/s)": 0.032413
    },
    {
      "epoch": 0.9550937363755269,
      "grad_norm": 0.10424994677305222,
      "learning_rate": 1.6549700531575284e-06,
      "loss": 0.335157573223114,
      "memory(GiB)": 78.33,
      "step": 4929,
      "token_acc": 0.899737302977233,
      "train_speed(iter/s)": 0.032414
    },
    {
      "epoch": 0.9552875066608536,
      "grad_norm": 0.09868663549423218,
      "learning_rate": 1.6407598285565093e-06,
      "loss": 0.32267120480537415,
      "memory(GiB)": 78.33,
      "step": 4930,
      "token_acc": 0.9046480108427011,
      "train_speed(iter/s)": 0.032414
    },
    {
      "epoch": 0.9554812769461803,
      "grad_norm": 0.09154357016086578,
      "learning_rate": 1.6266105390820017e-06,
      "loss": 0.31675025820732117,
      "memory(GiB)": 78.33,
      "step": 4931,
      "token_acc": 0.903805316214699,
      "train_speed(iter/s)": 0.032415
    },
    {
      "epoch": 0.9556750472315071,
      "grad_norm": 0.09340998530387878,
      "learning_rate": 1.6125221905455231e-06,
      "loss": 0.30397865176200867,
      "memory(GiB)": 78.33,
      "step": 4932,
      "token_acc": 0.9084064579411999,
      "train_speed(iter/s)": 0.032415
    },
    {
      "epoch": 0.9558688175168338,
      "grad_norm": 0.10670307278633118,
      "learning_rate": 1.598494788733462e-06,
      "loss": 0.32941100001335144,
      "memory(GiB)": 78.33,
      "step": 4933,
      "token_acc": 0.9001266174608566,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.9560625878021606,
      "grad_norm": 0.09027762711048126,
      "learning_rate": 1.584528339407326e-06,
      "loss": 0.2942469120025635,
      "memory(GiB)": 78.33,
      "step": 4934,
      "token_acc": 0.9111908000494621,
      "train_speed(iter/s)": 0.032416
    },
    {
      "epoch": 0.9562563580874873,
      "grad_norm": 0.09194032102823257,
      "learning_rate": 1.57062284830346e-06,
      "loss": 0.303377628326416,
      "memory(GiB)": 78.33,
      "step": 4935,
      "token_acc": 0.9077935247705968,
      "train_speed(iter/s)": 0.032417
    },
    {
      "epoch": 0.9564501283728141,
      "grad_norm": 0.13369497656822205,
      "learning_rate": 1.5567783211332619e-06,
      "loss": 0.3248521685600281,
      "memory(GiB)": 78.33,
      "step": 4936,
      "token_acc": 0.9012696041822256,
      "train_speed(iter/s)": 0.032417
    },
    {
      "epoch": 0.9566438986581408,
      "grad_norm": 0.0936022698879242,
      "learning_rate": 1.5429947635830164e-06,
      "loss": 0.3220188021659851,
      "memory(GiB)": 78.33,
      "step": 4937,
      "token_acc": 0.9041003304295584,
      "train_speed(iter/s)": 0.032418
    },
    {
      "epoch": 0.9568376689434676,
      "grad_norm": 0.09517528116703033,
      "learning_rate": 1.529272181314012e-06,
      "loss": 0.3031097650527954,
      "memory(GiB)": 78.33,
      "step": 4938,
      "token_acc": 0.9060737583491436,
      "train_speed(iter/s)": 0.032419
    },
    {
      "epoch": 0.9570314392287943,
      "grad_norm": 0.09666939824819565,
      "learning_rate": 1.5156105799625063e-06,
      "loss": 0.3083738684654236,
      "memory(GiB)": 78.33,
      "step": 4939,
      "token_acc": 0.9067667594099428,
      "train_speed(iter/s)": 0.032419
    },
    {
      "epoch": 0.957225209514121,
      "grad_norm": 0.09546789526939392,
      "learning_rate": 1.5020099651396444e-06,
      "loss": 0.3306988775730133,
      "memory(GiB)": 78.33,
      "step": 4940,
      "token_acc": 0.9035588697408784,
      "train_speed(iter/s)": 0.03242
    },
    {
      "epoch": 0.9574189797994478,
      "grad_norm": 0.10004635155200958,
      "learning_rate": 1.4884703424315915e-06,
      "loss": 0.3262169063091278,
      "memory(GiB)": 78.33,
      "step": 4941,
      "token_acc": 0.9031863057731064,
      "train_speed(iter/s)": 0.03242
    },
    {
      "epoch": 0.9576127500847745,
      "grad_norm": 0.09645616263151169,
      "learning_rate": 1.474991717399432e-06,
      "loss": 0.2961626946926117,
      "memory(GiB)": 78.33,
      "step": 4942,
      "token_acc": 0.9100462809072081,
      "train_speed(iter/s)": 0.032421
    },
    {
      "epoch": 0.9578065203701013,
      "grad_norm": 0.09985015541315079,
      "learning_rate": 1.4615740955792044e-06,
      "loss": 0.3318878412246704,
      "memory(GiB)": 78.33,
      "step": 4943,
      "token_acc": 0.9009440305532396,
      "train_speed(iter/s)": 0.032421
    },
    {
      "epoch": 0.958000290655428,
      "grad_norm": 0.10011833906173706,
      "learning_rate": 1.4482174824818671e-06,
      "loss": 0.337127149105072,
      "memory(GiB)": 78.33,
      "step": 4944,
      "token_acc": 0.8997083603370059,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.9581940609407548,
      "grad_norm": 0.10610119998455048,
      "learning_rate": 1.4349218835933486e-06,
      "loss": 0.35346800088882446,
      "memory(GiB)": 78.33,
      "step": 4945,
      "token_acc": 0.8942115189322501,
      "train_speed(iter/s)": 0.032422
    },
    {
      "epoch": 0.9583878312260815,
      "grad_norm": 0.09986083954572678,
      "learning_rate": 1.4216873043745137e-06,
      "loss": 0.3105732202529907,
      "memory(GiB)": 78.33,
      "step": 4946,
      "token_acc": 0.9050974597942831,
      "train_speed(iter/s)": 0.032423
    },
    {
      "epoch": 0.9585816015114083,
      "grad_norm": 0.10908082872629166,
      "learning_rate": 1.4085137502611477e-06,
      "loss": 0.3552319407463074,
      "memory(GiB)": 78.33,
      "step": 4947,
      "token_acc": 0.8946395037842144,
      "train_speed(iter/s)": 0.032423
    },
    {
      "epoch": 0.958775371796735,
      "grad_norm": 0.09129244834184647,
      "learning_rate": 1.3954012266640059e-06,
      "loss": 0.27050110697746277,
      "memory(GiB)": 78.33,
      "step": 4948,
      "token_acc": 0.9176398959428793,
      "train_speed(iter/s)": 0.032424
    },
    {
      "epoch": 0.9589691420820617,
      "grad_norm": 0.10130368173122406,
      "learning_rate": 1.3823497389687466e-06,
      "loss": 0.3376561403274536,
      "memory(GiB)": 78.33,
      "step": 4949,
      "token_acc": 0.8982817684028808,
      "train_speed(iter/s)": 0.032424
    },
    {
      "epoch": 0.9591629123673885,
      "grad_norm": 0.10016176849603653,
      "learning_rate": 1.369359292535932e-06,
      "loss": 0.3218695819377899,
      "memory(GiB)": 78.33,
      "step": 4950,
      "token_acc": 0.9014069086330542,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.9593566826527152,
      "grad_norm": 0.09046775102615356,
      "learning_rate": 1.356429892701144e-06,
      "loss": 0.30442506074905396,
      "memory(GiB)": 78.33,
      "step": 4951,
      "token_acc": 0.9069761971963496,
      "train_speed(iter/s)": 0.032425
    },
    {
      "epoch": 0.959550452938042,
      "grad_norm": 0.09234574437141418,
      "learning_rate": 1.343561544774785e-06,
      "loss": 0.30501365661621094,
      "memory(GiB)": 78.33,
      "step": 4952,
      "token_acc": 0.9067395915863726,
      "train_speed(iter/s)": 0.032426
    },
    {
      "epoch": 0.9597442232233687,
      "grad_norm": 0.09642866253852844,
      "learning_rate": 1.3307542540422766e-06,
      "loss": 0.29419055581092834,
      "memory(GiB)": 78.33,
      "step": 4953,
      "token_acc": 0.9106110240706452,
      "train_speed(iter/s)": 0.032426
    },
    {
      "epoch": 0.9599379935086955,
      "grad_norm": 0.10034112632274628,
      "learning_rate": 1.3180080257638782e-06,
      "loss": 0.339307963848114,
      "memory(GiB)": 78.33,
      "step": 4954,
      "token_acc": 0.8978796816739941,
      "train_speed(iter/s)": 0.032427
    },
    {
      "epoch": 0.9601317637940222,
      "grad_norm": 0.10334701836109161,
      "learning_rate": 1.3053228651748349e-06,
      "loss": 0.3403671681880951,
      "memory(GiB)": 78.33,
      "step": 4955,
      "token_acc": 0.8992078746578135,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.960325534079349,
      "grad_norm": 0.10150493681430817,
      "learning_rate": 1.2926987774852627e-06,
      "loss": 0.3322233259677887,
      "memory(GiB)": 78.33,
      "step": 4956,
      "token_acc": 0.9011913901447297,
      "train_speed(iter/s)": 0.032428
    },
    {
      "epoch": 0.9605193043646757,
      "grad_norm": 0.09975534677505493,
      "learning_rate": 1.2801357678802138e-06,
      "loss": 0.3389909267425537,
      "memory(GiB)": 78.33,
      "step": 4957,
      "token_acc": 0.8986236035148656,
      "train_speed(iter/s)": 0.032429
    },
    {
      "epoch": 0.9607130746500024,
      "grad_norm": 0.10334469377994537,
      "learning_rate": 1.2676338415196774e-06,
      "loss": 0.34487882256507874,
      "memory(GiB)": 78.33,
      "step": 4958,
      "token_acc": 0.897090561398716,
      "train_speed(iter/s)": 0.032429
    },
    {
      "epoch": 0.9609068449353292,
      "grad_norm": 0.10029633343219757,
      "learning_rate": 1.2551930035385126e-06,
      "loss": 0.32758665084838867,
      "memory(GiB)": 78.33,
      "step": 4959,
      "token_acc": 0.9024535719127016,
      "train_speed(iter/s)": 0.03243
    },
    {
      "epoch": 0.9611006152206559,
      "grad_norm": 0.1052929162979126,
      "learning_rate": 1.2428132590465156e-06,
      "loss": 0.3499680757522583,
      "memory(GiB)": 78.33,
      "step": 4960,
      "token_acc": 0.8937179730499146,
      "train_speed(iter/s)": 0.03243
    },
    {
      "epoch": 0.9612943855059827,
      "grad_norm": 0.10789167135953903,
      "learning_rate": 1.2304946131283521e-06,
      "loss": 0.347482293844223,
      "memory(GiB)": 78.33,
      "step": 4961,
      "token_acc": 0.8981575675370183,
      "train_speed(iter/s)": 0.032431
    },
    {
      "epoch": 0.9614881557913094,
      "grad_norm": 0.09253862500190735,
      "learning_rate": 1.2182370708436584e-06,
      "loss": 0.2795042097568512,
      "memory(GiB)": 78.33,
      "step": 4962,
      "token_acc": 0.9132865314612585,
      "train_speed(iter/s)": 0.032431
    },
    {
      "epoch": 0.9616819260766362,
      "grad_norm": 0.11232099682092667,
      "learning_rate": 1.2060406372269238e-06,
      "loss": 0.3673211932182312,
      "memory(GiB)": 78.33,
      "step": 4963,
      "token_acc": 0.8896090319882919,
      "train_speed(iter/s)": 0.032432
    },
    {
      "epoch": 0.9618756963619629,
      "grad_norm": 0.09660880267620087,
      "learning_rate": 1.1939053172875245e-06,
      "loss": 0.31402865052223206,
      "memory(GiB)": 78.33,
      "step": 4964,
      "token_acc": 0.9040212859337283,
      "train_speed(iter/s)": 0.032433
    },
    {
      "epoch": 0.9620694666472897,
      "grad_norm": 0.10611861944198608,
      "learning_rate": 1.1818311160098237e-06,
      "loss": 0.3377360999584198,
      "memory(GiB)": 78.33,
      "step": 4965,
      "token_acc": 0.9000355008579374,
      "train_speed(iter/s)": 0.032433
    },
    {
      "epoch": 0.9622632369326164,
      "grad_norm": 0.1003570705652237,
      "learning_rate": 1.1698180383529542e-06,
      "loss": 0.32752174139022827,
      "memory(GiB)": 78.33,
      "step": 4966,
      "token_acc": 0.9013712208822072,
      "train_speed(iter/s)": 0.032434
    },
    {
      "epoch": 0.9624570072179431,
      "grad_norm": 0.08952160179615021,
      "learning_rate": 1.1578660892510528e-06,
      "loss": 0.2788778245449066,
      "memory(GiB)": 78.33,
      "step": 4967,
      "token_acc": 0.9154586305821423,
      "train_speed(iter/s)": 0.032434
    },
    {
      "epoch": 0.9626507775032699,
      "grad_norm": 0.10649926215410233,
      "learning_rate": 1.1459752736130756e-06,
      "loss": 0.3794569969177246,
      "memory(GiB)": 78.33,
      "step": 4968,
      "token_acc": 0.888167308750688,
      "train_speed(iter/s)": 0.032435
    },
    {
      "epoch": 0.9628445477885966,
      "grad_norm": 0.08885262906551361,
      "learning_rate": 1.1341455963229329e-06,
      "loss": 0.29271164536476135,
      "memory(GiB)": 78.33,
      "step": 4969,
      "token_acc": 0.9096743030637593,
      "train_speed(iter/s)": 0.032435
    },
    {
      "epoch": 0.9630383180739234,
      "grad_norm": 0.10036831349134445,
      "learning_rate": 1.1223770622393714e-06,
      "loss": 0.3443622887134552,
      "memory(GiB)": 78.33,
      "step": 4970,
      "token_acc": 0.8970222654561176,
      "train_speed(iter/s)": 0.032436
    },
    {
      "epoch": 0.9632320883592501,
      "grad_norm": 0.09779267013072968,
      "learning_rate": 1.110669676196041e-06,
      "loss": 0.3445608615875244,
      "memory(GiB)": 78.33,
      "step": 4971,
      "token_acc": 0.8951095773995004,
      "train_speed(iter/s)": 0.032436
    },
    {
      "epoch": 0.9634258586445769,
      "grad_norm": 0.10155529528856277,
      "learning_rate": 1.0990234430014954e-06,
      "loss": 0.3322911262512207,
      "memory(GiB)": 78.33,
      "step": 4972,
      "token_acc": 0.9017332921313593,
      "train_speed(iter/s)": 0.032437
    },
    {
      "epoch": 0.9636196289299036,
      "grad_norm": 0.09281053394079208,
      "learning_rate": 1.087438367439125e-06,
      "loss": 0.29993924498558044,
      "memory(GiB)": 78.33,
      "step": 4973,
      "token_acc": 0.9089661368003393,
      "train_speed(iter/s)": 0.032437
    },
    {
      "epoch": 0.9638133992152303,
      "grad_norm": 0.09526029974222183,
      "learning_rate": 1.0759144542672737e-06,
      "loss": 0.29557523131370544,
      "memory(GiB)": 78.33,
      "step": 4974,
      "token_acc": 0.9109068897204491,
      "train_speed(iter/s)": 0.032438
    },
    {
      "epoch": 0.9640071695005571,
      "grad_norm": 0.10219217091798782,
      "learning_rate": 1.0644517082190883e-06,
      "loss": 0.3419698476791382,
      "memory(GiB)": 78.33,
      "step": 4975,
      "token_acc": 0.8975590462199119,
      "train_speed(iter/s)": 0.032439
    },
    {
      "epoch": 0.9642009397858838,
      "grad_norm": 0.09392113983631134,
      "learning_rate": 1.0530501340026532e-06,
      "loss": 0.3047352135181427,
      "memory(GiB)": 78.33,
      "step": 4976,
      "token_acc": 0.9068194872330426,
      "train_speed(iter/s)": 0.032439
    },
    {
      "epoch": 0.9643947100712106,
      "grad_norm": 0.10133232921361923,
      "learning_rate": 1.0417097363008886e-06,
      "loss": 0.32953399419784546,
      "memory(GiB)": 78.33,
      "step": 4977,
      "token_acc": 0.9026956897794436,
      "train_speed(iter/s)": 0.03244
    },
    {
      "epoch": 0.9645884803565373,
      "grad_norm": 0.09701909869909286,
      "learning_rate": 1.030430519771569e-06,
      "loss": 0.3382679224014282,
      "memory(GiB)": 78.33,
      "step": 4978,
      "token_acc": 0.8983569375214284,
      "train_speed(iter/s)": 0.03244
    },
    {
      "epoch": 0.9647822506418641,
      "grad_norm": 0.09171733260154724,
      "learning_rate": 1.0192124890474385e-06,
      "loss": 0.3097023665904999,
      "memory(GiB)": 78.33,
      "step": 4979,
      "token_acc": 0.9057717083225972,
      "train_speed(iter/s)": 0.032441
    },
    {
      "epoch": 0.9649760209271908,
      "grad_norm": 0.0920836329460144,
      "learning_rate": 1.0080556487359947e-06,
      "loss": 0.30298176407814026,
      "memory(GiB)": 78.33,
      "step": 4980,
      "token_acc": 0.9079930043486482,
      "train_speed(iter/s)": 0.032441
    },
    {
      "epoch": 0.9651697912125176,
      "grad_norm": 0.09740208089351654,
      "learning_rate": 9.969600034196557e-07,
      "loss": 0.3513997197151184,
      "memory(GiB)": 78.33,
      "step": 4981,
      "token_acc": 0.8963819470346885,
      "train_speed(iter/s)": 0.032442
    },
    {
      "epoch": 0.9653635614978443,
      "grad_norm": 0.09901650995016098,
      "learning_rate": 9.859255576557257e-07,
      "loss": 0.32587093114852905,
      "memory(GiB)": 78.33,
      "step": 4982,
      "token_acc": 0.9019118199881854,
      "train_speed(iter/s)": 0.032442
    },
    {
      "epoch": 0.965557331783171,
      "grad_norm": 0.10267394781112671,
      "learning_rate": 9.749523159763295e-07,
      "loss": 0.32968056201934814,
      "memory(GiB)": 78.33,
      "step": 4983,
      "token_acc": 0.9018928833455613,
      "train_speed(iter/s)": 0.032443
    },
    {
      "epoch": 0.9657511020684978,
      "grad_norm": 0.0950947031378746,
      "learning_rate": 9.64040282888462e-07,
      "loss": 0.32611650228500366,
      "memory(GiB)": 78.33,
      "step": 4984,
      "token_acc": 0.9009006650259203,
      "train_speed(iter/s)": 0.032443
    },
    {
      "epoch": 0.9659448723538245,
      "grad_norm": 0.10337654501199722,
      "learning_rate": 9.531894628740044e-07,
      "loss": 0.3508460521697998,
      "memory(GiB)": 78.33,
      "step": 4985,
      "token_acc": 0.8948303758520995,
      "train_speed(iter/s)": 0.032444
    },
    {
      "epoch": 0.9661386426391513,
      "grad_norm": 0.08894824236631393,
      "learning_rate": 9.423998603896921e-07,
      "loss": 0.29168346524238586,
      "memory(GiB)": 78.33,
      "step": 4986,
      "token_acc": 0.9120740535223346,
      "train_speed(iter/s)": 0.032444
    },
    {
      "epoch": 0.966332412924478,
      "grad_norm": 0.0989081859588623,
      "learning_rate": 9.316714798670799e-07,
      "loss": 0.32772505283355713,
      "memory(GiB)": 78.33,
      "step": 4987,
      "token_acc": 0.899692881430295,
      "train_speed(iter/s)": 0.032445
    },
    {
      "epoch": 0.9665261832098048,
      "grad_norm": 0.08811601996421814,
      "learning_rate": 9.210043257126098e-07,
      "loss": 0.289480984210968,
      "memory(GiB)": 78.33,
      "step": 4988,
      "token_acc": 0.9128572101125182,
      "train_speed(iter/s)": 0.032445
    },
    {
      "epoch": 0.9667199534951315,
      "grad_norm": 0.10313121974468231,
      "learning_rate": 9.103984023075772e-07,
      "loss": 0.35241398215293884,
      "memory(GiB)": 78.33,
      "step": 4989,
      "token_acc": 0.8971772553485724,
      "train_speed(iter/s)": 0.032446
    },
    {
      "epoch": 0.9669137237804583,
      "grad_norm": 0.09983363002538681,
      "learning_rate": 8.998537140081141e-07,
      "loss": 0.31030166149139404,
      "memory(GiB)": 78.33,
      "step": 4990,
      "token_acc": 0.9058147247402166,
      "train_speed(iter/s)": 0.032447
    },
    {
      "epoch": 0.967107494065785,
      "grad_norm": 0.10593412816524506,
      "learning_rate": 8.893702651452062e-07,
      "loss": 0.3430192768573761,
      "memory(GiB)": 78.33,
      "step": 4991,
      "token_acc": 0.8970623145400594,
      "train_speed(iter/s)": 0.032447
    },
    {
      "epoch": 0.9673012643511117,
      "grad_norm": 0.09164869040250778,
      "learning_rate": 8.789480600246757e-07,
      "loss": 0.3059519827365875,
      "memory(GiB)": 78.33,
      "step": 4992,
      "token_acc": 0.9062849909936225,
      "train_speed(iter/s)": 0.032448
    },
    {
      "epoch": 0.9674950346364385,
      "grad_norm": 0.11753620952367783,
      "learning_rate": 8.685871029272318e-07,
      "loss": 0.35421380400657654,
      "memory(GiB)": 78.33,
      "step": 4993,
      "token_acc": 0.8949581180397317,
      "train_speed(iter/s)": 0.032448
    },
    {
      "epoch": 0.9676888049217652,
      "grad_norm": 0.09262137115001678,
      "learning_rate": 8.582873981083705e-07,
      "loss": 0.29231324791908264,
      "memory(GiB)": 78.33,
      "step": 4994,
      "token_acc": 0.9102221280876863,
      "train_speed(iter/s)": 0.032449
    },
    {
      "epoch": 0.967882575207092,
      "grad_norm": 0.09335288405418396,
      "learning_rate": 8.480489497984744e-07,
      "loss": 0.3362676203250885,
      "memory(GiB)": 78.33,
      "step": 4995,
      "token_acc": 0.8979033950843279,
      "train_speed(iter/s)": 0.032449
    },
    {
      "epoch": 0.9680763454924187,
      "grad_norm": 0.12949591875076294,
      "learning_rate": 8.378717622027465e-07,
      "loss": 0.32554298639297485,
      "memory(GiB)": 78.33,
      "step": 4996,
      "token_acc": 0.9030780971762911,
      "train_speed(iter/s)": 0.03245
    },
    {
      "epoch": 0.9682701157777455,
      "grad_norm": 0.10350757837295532,
      "learning_rate": 8.277558395012096e-07,
      "loss": 0.335059255361557,
      "memory(GiB)": 78.33,
      "step": 4997,
      "token_acc": 0.8990599887317261,
      "train_speed(iter/s)": 0.03245
    },
    {
      "epoch": 0.9684638860630722,
      "grad_norm": 0.11247258633375168,
      "learning_rate": 8.177011858487903e-07,
      "loss": 0.3429482579231262,
      "memory(GiB)": 78.33,
      "step": 4998,
      "token_acc": 0.895786360575093,
      "train_speed(iter/s)": 0.032451
    },
    {
      "epoch": 0.968657656348399,
      "grad_norm": 0.10709885507822037,
      "learning_rate": 8.077078053751518e-07,
      "loss": 0.3340757191181183,
      "memory(GiB)": 78.33,
      "step": 4999,
      "token_acc": 0.8992043255199533,
      "train_speed(iter/s)": 0.032451
    },
    {
      "epoch": 0.9688514266337257,
      "grad_norm": 0.09967630356550217,
      "learning_rate": 7.97775702184894e-07,
      "loss": 0.30826178193092346,
      "memory(GiB)": 78.33,
      "step": 5000,
      "token_acc": 0.9074506820281506,
      "train_speed(iter/s)": 0.032452
    },
    {
      "epoch": 0.9688514266337257,
      "eval_loss": 0.3782345950603485,
      "eval_runtime": 1344.6451,
      "eval_samples_per_second": 5.019,
      "eval_steps_per_second": 5.019,
      "eval_token_acc": 0.9026852655344548,
      "step": 5000
    },
    {
      "epoch": 0.9690451969190524,
      "grad_norm": 0.10758214443922043,
      "learning_rate": 7.87904880357354e-07,
      "loss": 0.3177351951599121,
      "memory(GiB)": 78.33,
      "step": 5001,
      "token_acc": 0.9042991375981465,
      "train_speed(iter/s)": 0.032169
    },
    {
      "epoch": 0.9692389672043792,
      "grad_norm": 0.0970822274684906,
      "learning_rate": 7.780953439467719e-07,
      "loss": 0.3285868167877197,
      "memory(GiB)": 78.33,
      "step": 5002,
      "token_acc": 0.9016248076571994,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.9694327374897059,
      "grad_norm": 0.10737350583076477,
      "learning_rate": 7.683470969821748e-07,
      "loss": 0.3421315550804138,
      "memory(GiB)": 78.33,
      "step": 5003,
      "token_acc": 0.894730186830209,
      "train_speed(iter/s)": 0.03217
    },
    {
      "epoch": 0.9696265077750327,
      "grad_norm": 0.09953954815864563,
      "learning_rate": 7.586601434674266e-07,
      "loss": 0.31448549032211304,
      "memory(GiB)": 78.33,
      "step": 5004,
      "token_acc": 0.9061734010562758,
      "train_speed(iter/s)": 0.032171
    },
    {
      "epoch": 0.9698202780603594,
      "grad_norm": 0.10338281095027924,
      "learning_rate": 7.490344873812615e-07,
      "loss": 0.33236753940582275,
      "memory(GiB)": 78.33,
      "step": 5005,
      "token_acc": 0.8993050377307842,
      "train_speed(iter/s)": 0.032171
    },
    {
      "epoch": 0.9700140483456862,
      "grad_norm": 0.09365373104810715,
      "learning_rate": 7.394701326771335e-07,
      "loss": 0.3056509792804718,
      "memory(GiB)": 78.33,
      "step": 5006,
      "token_acc": 0.906607994493338,
      "train_speed(iter/s)": 0.032172
    },
    {
      "epoch": 0.9702078186310129,
      "grad_norm": 0.10147716104984283,
      "learning_rate": 7.29967083283417e-07,
      "loss": 0.31711748242378235,
      "memory(GiB)": 78.33,
      "step": 5007,
      "token_acc": 0.902543880455408,
      "train_speed(iter/s)": 0.032173
    },
    {
      "epoch": 0.9704015889163397,
      "grad_norm": 0.09770286083221436,
      "learning_rate": 7.205253431032564e-07,
      "loss": 0.324236124753952,
      "memory(GiB)": 78.33,
      "step": 5008,
      "token_acc": 0.9028327266972622,
      "train_speed(iter/s)": 0.032173
    },
    {
      "epoch": 0.9705953592016664,
      "grad_norm": 0.10440776497125626,
      "learning_rate": 7.111449160146332e-07,
      "loss": 0.3207136392593384,
      "memory(GiB)": 78.33,
      "step": 5009,
      "token_acc": 0.9042985518859825,
      "train_speed(iter/s)": 0.032174
    },
    {
      "epoch": 0.9707891294869931,
      "grad_norm": 0.09528158605098724,
      "learning_rate": 7.018258058703319e-07,
      "loss": 0.2945985496044159,
      "memory(GiB)": 78.33,
      "step": 5010,
      "token_acc": 0.9116842726151536,
      "train_speed(iter/s)": 0.032174
    },
    {
      "epoch": 0.9709828997723199,
      "grad_norm": 0.09875106066465378,
      "learning_rate": 6.925680164979741e-07,
      "loss": 0.33711299300193787,
      "memory(GiB)": 78.33,
      "step": 5011,
      "token_acc": 0.8979212309573547,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.9711766700576466,
      "grad_norm": 0.09752146899700165,
      "learning_rate": 6.833715516999849e-07,
      "loss": 0.33156195282936096,
      "memory(GiB)": 78.33,
      "step": 5012,
      "token_acc": 0.9005984838409363,
      "train_speed(iter/s)": 0.032175
    },
    {
      "epoch": 0.9713704403429734,
      "grad_norm": 0.11073119193315506,
      "learning_rate": 6.742364152535929e-07,
      "loss": 0.3404943645000458,
      "memory(GiB)": 78.33,
      "step": 5013,
      "token_acc": 0.8999028182701652,
      "train_speed(iter/s)": 0.032176
    },
    {
      "epoch": 0.9715642106283001,
      "grad_norm": 0.08712846785783768,
      "learning_rate": 6.651626109108465e-07,
      "loss": 0.30490002036094666,
      "memory(GiB)": 78.33,
      "step": 5014,
      "token_acc": 0.907512204600102,
      "train_speed(iter/s)": 0.032176
    },
    {
      "epoch": 0.9717579809136269,
      "grad_norm": 0.0969480574131012,
      "learning_rate": 6.561501423985816e-07,
      "loss": 0.32621052861213684,
      "memory(GiB)": 78.33,
      "step": 5015,
      "token_acc": 0.9022739990842956,
      "train_speed(iter/s)": 0.032177
    },
    {
      "epoch": 0.9719517511989536,
      "grad_norm": 0.11033257842063904,
      "learning_rate": 6.471990134185035e-07,
      "loss": 0.34958821535110474,
      "memory(GiB)": 78.33,
      "step": 5016,
      "token_acc": 0.8971919453168299,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.9721455214842804,
      "grad_norm": 0.11200974881649017,
      "learning_rate": 6.383092276470381e-07,
      "loss": 0.37351301312446594,
      "memory(GiB)": 78.33,
      "step": 5017,
      "token_acc": 0.890080579498554,
      "train_speed(iter/s)": 0.032178
    },
    {
      "epoch": 0.9723392917696071,
      "grad_norm": 0.10688678920269012,
      "learning_rate": 6.294807887354647e-07,
      "loss": 0.32566171884536743,
      "memory(GiB)": 78.33,
      "step": 5018,
      "token_acc": 0.9027785449925426,
      "train_speed(iter/s)": 0.032179
    },
    {
      "epoch": 0.9725330620549338,
      "grad_norm": 0.11659563332796097,
      "learning_rate": 6.207137003098994e-07,
      "loss": 0.3676939904689789,
      "memory(GiB)": 78.33,
      "step": 5019,
      "token_acc": 0.8909811380567443,
      "train_speed(iter/s)": 0.032179
    },
    {
      "epoch": 0.9727268323402606,
      "grad_norm": 0.10240488499403,
      "learning_rate": 6.120079659711786e-07,
      "loss": 0.3170427680015564,
      "memory(GiB)": 78.33,
      "step": 5020,
      "token_acc": 0.9030933713471133,
      "train_speed(iter/s)": 0.03218
    },
    {
      "epoch": 0.9729206026255873,
      "grad_norm": 0.1029011458158493,
      "learning_rate": 6.033635892950084e-07,
      "loss": 0.3273800313472748,
      "memory(GiB)": 78.33,
      "step": 5021,
      "token_acc": 0.9038344491783323,
      "train_speed(iter/s)": 0.03218
    },
    {
      "epoch": 0.9731143729109141,
      "grad_norm": 0.09669654816389084,
      "learning_rate": 5.94780573831849e-07,
      "loss": 0.3153877854347229,
      "memory(GiB)": 78.33,
      "step": 5022,
      "token_acc": 0.9058483637541851,
      "train_speed(iter/s)": 0.032181
    },
    {
      "epoch": 0.9733081431962408,
      "grad_norm": 0.09650097042322159,
      "learning_rate": 5.862589231069803e-07,
      "loss": 0.3036450445652008,
      "memory(GiB)": 78.33,
      "step": 5023,
      "token_acc": 0.9092272045795284,
      "train_speed(iter/s)": 0.032181
    },
    {
      "epoch": 0.9735019134815676,
      "grad_norm": 0.11860737949609756,
      "learning_rate": 5.777986406204694e-07,
      "loss": 0.3921282887458801,
      "memory(GiB)": 78.33,
      "step": 5024,
      "token_acc": 0.8855097849722156,
      "train_speed(iter/s)": 0.032182
    },
    {
      "epoch": 0.9736956837668943,
      "grad_norm": 0.10267064720392227,
      "learning_rate": 5.693997298472031e-07,
      "loss": 0.31443023681640625,
      "memory(GiB)": 78.33,
      "step": 5025,
      "token_acc": 0.9051502501866032,
      "train_speed(iter/s)": 0.032183
    },
    {
      "epoch": 0.973889454052221,
      "grad_norm": 0.09616965055465698,
      "learning_rate": 5.610621942368054e-07,
      "loss": 0.3315185606479645,
      "memory(GiB)": 78.33,
      "step": 5026,
      "token_acc": 0.902337848564771,
      "train_speed(iter/s)": 0.032183
    },
    {
      "epoch": 0.9740832243375478,
      "grad_norm": 0.09976300597190857,
      "learning_rate": 5.527860372137538e-07,
      "loss": 0.32565930485725403,
      "memory(GiB)": 78.33,
      "step": 5027,
      "token_acc": 0.9027216527952472,
      "train_speed(iter/s)": 0.032184
    },
    {
      "epoch": 0.9742769946228745,
      "grad_norm": 0.09962465614080429,
      "learning_rate": 5.445712621772791e-07,
      "loss": 0.327309787273407,
      "memory(GiB)": 78.33,
      "step": 5028,
      "token_acc": 0.9021066306645948,
      "train_speed(iter/s)": 0.032184
    },
    {
      "epoch": 0.9744707649082013,
      "grad_norm": 0.10362134128808975,
      "learning_rate": 5.364178725014157e-07,
      "loss": 0.3133904039859772,
      "memory(GiB)": 78.33,
      "step": 5029,
      "token_acc": 0.9041495198902606,
      "train_speed(iter/s)": 0.032185
    },
    {
      "epoch": 0.974664535193528,
      "grad_norm": 0.10740216076374054,
      "learning_rate": 5.283258715349514e-07,
      "loss": 0.33964803814888,
      "memory(GiB)": 78.33,
      "step": 5030,
      "token_acc": 0.8985633557311141,
      "train_speed(iter/s)": 0.032186
    },
    {
      "epoch": 0.9748583054788548,
      "grad_norm": 0.0993296429514885,
      "learning_rate": 5.202952626015445e-07,
      "loss": 0.3082018196582794,
      "memory(GiB)": 78.33,
      "step": 5031,
      "token_acc": 0.9063476667744369,
      "train_speed(iter/s)": 0.032186
    },
    {
      "epoch": 0.9750520757641815,
      "grad_norm": 0.09285027533769608,
      "learning_rate": 5.123260489995229e-07,
      "loss": 0.2862505316734314,
      "memory(GiB)": 78.33,
      "step": 5032,
      "token_acc": 0.9113854235062376,
      "train_speed(iter/s)": 0.032187
    },
    {
      "epoch": 0.9752458460495083,
      "grad_norm": 0.11154574900865555,
      "learning_rate": 5.044182340021019e-07,
      "loss": 0.36379557847976685,
      "memory(GiB)": 78.33,
      "step": 5033,
      "token_acc": 0.8903923823000501,
      "train_speed(iter/s)": 0.032187
    },
    {
      "epoch": 0.975439616334835,
      "grad_norm": 0.09221196174621582,
      "learning_rate": 4.965718208572001e-07,
      "loss": 0.3152211308479309,
      "memory(GiB)": 78.33,
      "step": 5034,
      "token_acc": 0.9047773077880069,
      "train_speed(iter/s)": 0.032188
    },
    {
      "epoch": 0.9756333866201617,
      "grad_norm": 0.09758051484823227,
      "learning_rate": 4.887868127875561e-07,
      "loss": 0.3170757293701172,
      "memory(GiB)": 78.33,
      "step": 5035,
      "token_acc": 0.9045536265328575,
      "train_speed(iter/s)": 0.032188
    },
    {
      "epoch": 0.9758271569054885,
      "grad_norm": 0.09600334614515305,
      "learning_rate": 4.810632129907122e-07,
      "loss": 0.31009405851364136,
      "memory(GiB)": 78.33,
      "step": 5036,
      "token_acc": 0.9058032803330491,
      "train_speed(iter/s)": 0.032189
    },
    {
      "epoch": 0.9760209271908152,
      "grad_norm": 0.10319238901138306,
      "learning_rate": 4.7340102463891415e-07,
      "loss": 0.33604225516319275,
      "memory(GiB)": 78.33,
      "step": 5037,
      "token_acc": 0.8996856559863233,
      "train_speed(iter/s)": 0.032189
    },
    {
      "epoch": 0.976214697476142,
      "grad_norm": 0.10404365509748459,
      "learning_rate": 4.6580025087926134e-07,
      "loss": 0.3151894509792328,
      "memory(GiB)": 78.33,
      "step": 5038,
      "token_acc": 0.9064724919093851,
      "train_speed(iter/s)": 0.03219
    },
    {
      "epoch": 0.9764084677614687,
      "grad_norm": 0.10096472501754761,
      "learning_rate": 4.5826089483358973e-07,
      "loss": 0.3288641571998596,
      "memory(GiB)": 78.33,
      "step": 5039,
      "token_acc": 0.9003461989642643,
      "train_speed(iter/s)": 0.032191
    },
    {
      "epoch": 0.9766022380467956,
      "grad_norm": 0.08929524570703506,
      "learning_rate": 4.5078295959850576e-07,
      "loss": 0.2969154715538025,
      "memory(GiB)": 78.33,
      "step": 5040,
      "token_acc": 0.9102983397827986,
      "train_speed(iter/s)": 0.032191
    },
    {
      "epoch": 0.9767960083321223,
      "grad_norm": 0.09447763115167618,
      "learning_rate": 4.4336644824540245e-07,
      "loss": 0.3048000633716583,
      "memory(GiB)": 78.33,
      "step": 5041,
      "token_acc": 0.9062621145943711,
      "train_speed(iter/s)": 0.032192
    },
    {
      "epoch": 0.9769897786174491,
      "grad_norm": 0.10533607751131058,
      "learning_rate": 4.360113638204432e-07,
      "loss": 0.3521908223628998,
      "memory(GiB)": 78.33,
      "step": 5042,
      "token_acc": 0.8957643566617194,
      "train_speed(iter/s)": 0.032192
    },
    {
      "epoch": 0.9771835489027758,
      "grad_norm": 0.09984603524208069,
      "learning_rate": 4.287177093445615e-07,
      "loss": 0.3105589747428894,
      "memory(GiB)": 78.33,
      "step": 5043,
      "token_acc": 0.9076200993926008,
      "train_speed(iter/s)": 0.032193
    },
    {
      "epoch": 0.9773773191881026,
      "grad_norm": 0.09986462444067001,
      "learning_rate": 4.2148548781344437e-07,
      "loss": 0.2763623893260956,
      "memory(GiB)": 78.33,
      "step": 5044,
      "token_acc": 0.9144782780290841,
      "train_speed(iter/s)": 0.032193
    },
    {
      "epoch": 0.9775710894734293,
      "grad_norm": 0.09815599024295807,
      "learning_rate": 4.143147021975823e-07,
      "loss": 0.33960387110710144,
      "memory(GiB)": 78.33,
      "step": 5045,
      "token_acc": 0.9006058664958497,
      "train_speed(iter/s)": 0.032194
    },
    {
      "epoch": 0.977764859758756,
      "grad_norm": 0.10120173543691635,
      "learning_rate": 4.0720535544216945e-07,
      "loss": 0.3160873055458069,
      "memory(GiB)": 78.33,
      "step": 5046,
      "token_acc": 0.9033082947099249,
      "train_speed(iter/s)": 0.032195
    },
    {
      "epoch": 0.9779586300440828,
      "grad_norm": 0.09940861910581589,
      "learning_rate": 4.0015745046725336e-07,
      "loss": 0.31816571950912476,
      "memory(GiB)": 78.33,
      "step": 5047,
      "token_acc": 0.9058804471083752,
      "train_speed(iter/s)": 0.032195
    },
    {
      "epoch": 0.9781524003294095,
      "grad_norm": 0.11098282784223557,
      "learning_rate": 3.931709901675684e-07,
      "loss": 0.353601336479187,
      "memory(GiB)": 78.33,
      "step": 5048,
      "token_acc": 0.8934923500340327,
      "train_speed(iter/s)": 0.032196
    },
    {
      "epoch": 0.9783461706147363,
      "grad_norm": 0.09338078647851944,
      "learning_rate": 3.862459774126525e-07,
      "loss": 0.3041617274284363,
      "memory(GiB)": 78.33,
      "step": 5049,
      "token_acc": 0.9068331108843003,
      "train_speed(iter/s)": 0.032196
    },
    {
      "epoch": 0.978539940900063,
      "grad_norm": 0.10786343365907669,
      "learning_rate": 3.793824150467806e-07,
      "loss": 0.3523224890232086,
      "memory(GiB)": 78.33,
      "step": 5050,
      "token_acc": 0.8945492180312787,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.9787337111853898,
      "grad_norm": 0.09506326168775558,
      "learning_rate": 3.7258030588901424e-07,
      "loss": 0.3265533447265625,
      "memory(GiB)": 78.33,
      "step": 5051,
      "token_acc": 0.9006531536959823,
      "train_speed(iter/s)": 0.032197
    },
    {
      "epoch": 0.9789274814707165,
      "grad_norm": 0.0921454057097435,
      "learning_rate": 3.6583965273316864e-07,
      "loss": 0.29769742488861084,
      "memory(GiB)": 78.33,
      "step": 5052,
      "token_acc": 0.9098529003608105,
      "train_speed(iter/s)": 0.032198
    },
    {
      "epoch": 0.9791212517560433,
      "grad_norm": 0.10511661320924759,
      "learning_rate": 3.591604583478125e-07,
      "loss": 0.33176693320274353,
      "memory(GiB)": 78.33,
      "step": 5053,
      "token_acc": 0.9007189710979348,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.97931502204137,
      "grad_norm": 0.09943993389606476,
      "learning_rate": 3.5254272547623474e-07,
      "loss": 0.32589173316955566,
      "memory(GiB)": 78.33,
      "step": 5054,
      "token_acc": 0.9014144342263095,
      "train_speed(iter/s)": 0.032199
    },
    {
      "epoch": 0.9795087923266967,
      "grad_norm": 0.10941484570503235,
      "learning_rate": 3.4598645683656113e-07,
      "loss": 0.3675105571746826,
      "memory(GiB)": 78.33,
      "step": 5055,
      "token_acc": 0.8914224336351082,
      "train_speed(iter/s)": 0.0322
    },
    {
      "epoch": 0.9797025626120235,
      "grad_norm": 0.10779840499162674,
      "learning_rate": 3.3949165512160423e-07,
      "loss": 0.35344696044921875,
      "memory(GiB)": 78.33,
      "step": 5056,
      "token_acc": 0.8957955624622291,
      "train_speed(iter/s)": 0.0322
    },
    {
      "epoch": 0.9798963328973502,
      "grad_norm": 0.09050939232110977,
      "learning_rate": 3.330583229989636e-07,
      "loss": 0.30384790897369385,
      "memory(GiB)": 78.33,
      "step": 5057,
      "token_acc": 0.9091451737259681,
      "train_speed(iter/s)": 0.032201
    },
    {
      "epoch": 0.980090103182677,
      "grad_norm": 0.10062376409769058,
      "learning_rate": 3.2668646311097556e-07,
      "loss": 0.30701377987861633,
      "memory(GiB)": 78.33,
      "step": 5058,
      "token_acc": 0.9093784940958303,
      "train_speed(iter/s)": 0.032201
    },
    {
      "epoch": 0.9802838734680037,
      "grad_norm": 0.09237212687730789,
      "learning_rate": 3.2037607807473e-07,
      "loss": 0.30352169275283813,
      "memory(GiB)": 78.33,
      "step": 5059,
      "token_acc": 0.9058777531604327,
      "train_speed(iter/s)": 0.032202
    },
    {
      "epoch": 0.9804776437533305,
      "grad_norm": 0.0979812890291214,
      "learning_rate": 3.1412717048207025e-07,
      "loss": 0.31397053599357605,
      "memory(GiB)": 78.33,
      "step": 5060,
      "token_acc": 0.9068339778781405,
      "train_speed(iter/s)": 0.032203
    },
    {
      "epoch": 0.9806714140386572,
      "grad_norm": 0.09479079395532608,
      "learning_rate": 3.0793974289961e-07,
      "loss": 0.31937700510025024,
      "memory(GiB)": 78.33,
      "step": 5061,
      "token_acc": 0.9035309120858683,
      "train_speed(iter/s)": 0.032203
    },
    {
      "epoch": 0.980865184323984,
      "grad_norm": 0.12520615756511688,
      "learning_rate": 3.01813797868683e-07,
      "loss": 0.35134872794151306,
      "memory(GiB)": 78.33,
      "step": 5062,
      "token_acc": 0.8932484641205903,
      "train_speed(iter/s)": 0.032204
    },
    {
      "epoch": 0.9810589546093107,
      "grad_norm": 0.10819295048713684,
      "learning_rate": 2.957493379053599e-07,
      "loss": 0.3610383868217468,
      "memory(GiB)": 78.33,
      "step": 5063,
      "token_acc": 0.8935895511184507,
      "train_speed(iter/s)": 0.032204
    },
    {
      "epoch": 0.9812527248946374,
      "grad_norm": 0.09377988427877426,
      "learning_rate": 2.8974636550049833e-07,
      "loss": 0.3174287676811218,
      "memory(GiB)": 78.33,
      "step": 5064,
      "token_acc": 0.9046624721817732,
      "train_speed(iter/s)": 0.032205
    },
    {
      "epoch": 0.9814464951799642,
      "grad_norm": 0.09167510271072388,
      "learning_rate": 2.83804883119676e-07,
      "loss": 0.297576367855072,
      "memory(GiB)": 78.33,
      "step": 5065,
      "token_acc": 0.9104416645391882,
      "train_speed(iter/s)": 0.032205
    },
    {
      "epoch": 0.9816402654652909,
      "grad_norm": 0.09928658604621887,
      "learning_rate": 2.7792489320322407e-07,
      "loss": 0.3319474458694458,
      "memory(GiB)": 78.33,
      "step": 5066,
      "token_acc": 0.9005888179616993,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.9818340357506177,
      "grad_norm": 0.09505660086870193,
      "learning_rate": 2.721063981661942e-07,
      "loss": 0.32442140579223633,
      "memory(GiB)": 78.33,
      "step": 5067,
      "token_acc": 0.9013683579704355,
      "train_speed(iter/s)": 0.032206
    },
    {
      "epoch": 0.9820278060359444,
      "grad_norm": 0.09601055830717087,
      "learning_rate": 2.663494003984079e-07,
      "loss": 0.3260546922683716,
      "memory(GiB)": 78.33,
      "step": 5068,
      "token_acc": 0.9013065431263338,
      "train_speed(iter/s)": 0.032207
    },
    {
      "epoch": 0.9822215763212712,
      "grad_norm": 0.10125764459371567,
      "learning_rate": 2.6065390226444047e-07,
      "loss": 0.3628811836242676,
      "memory(GiB)": 78.33,
      "step": 5069,
      "token_acc": 0.8924914675767918,
      "train_speed(iter/s)": 0.032208
    },
    {
      "epoch": 0.9824153466065979,
      "grad_norm": 0.10014388710260391,
      "learning_rate": 2.5501990610355406e-07,
      "loss": 0.32086047530174255,
      "memory(GiB)": 78.33,
      "step": 5070,
      "token_acc": 0.9021128125605737,
      "train_speed(iter/s)": 0.032208
    },
    {
      "epoch": 0.9826091168919246,
      "grad_norm": 0.09349211305379868,
      "learning_rate": 2.4944741422979754e-07,
      "loss": 0.31767165660858154,
      "memory(GiB)": 78.33,
      "step": 5071,
      "token_acc": 0.9040567600306544,
      "train_speed(iter/s)": 0.032209
    },
    {
      "epoch": 0.9828028871772514,
      "grad_norm": 0.09731042385101318,
      "learning_rate": 2.4393642893194007e-07,
      "loss": 0.3093053102493286,
      "memory(GiB)": 78.33,
      "step": 5072,
      "token_acc": 0.9075150674702651,
      "train_speed(iter/s)": 0.032209
    },
    {
      "epoch": 0.9829966574625781,
      "grad_norm": 0.10741175711154938,
      "learning_rate": 2.3848695247350446e-07,
      "loss": 0.31964316964149475,
      "memory(GiB)": 78.33,
      "step": 5073,
      "token_acc": 0.9033793824646376,
      "train_speed(iter/s)": 0.03221
    },
    {
      "epoch": 0.9831904277479049,
      "grad_norm": 0.09962072223424911,
      "learning_rate": 2.330989870927169e-07,
      "loss": 0.33166712522506714,
      "memory(GiB)": 78.33,
      "step": 5074,
      "token_acc": 0.8991797207209679,
      "train_speed(iter/s)": 0.03221
    },
    {
      "epoch": 0.9833841980332316,
      "grad_norm": 0.11339244991540909,
      "learning_rate": 2.2777253500257386e-07,
      "loss": 0.3660873770713806,
      "memory(GiB)": 78.33,
      "step": 5075,
      "token_acc": 0.8902705205370726,
      "train_speed(iter/s)": 0.032211
    },
    {
      "epoch": 0.9835779683185584,
      "grad_norm": 0.10337600857019424,
      "learning_rate": 2.2250759839077536e-07,
      "loss": 0.3016257882118225,
      "memory(GiB)": 78.33,
      "step": 5076,
      "token_acc": 0.9081414405155412,
      "train_speed(iter/s)": 0.032212
    },
    {
      "epoch": 0.9837717386038851,
      "grad_norm": 0.09644091874361038,
      "learning_rate": 2.173041794197916e-07,
      "loss": 0.3238067924976349,
      "memory(GiB)": 78.33,
      "step": 5077,
      "token_acc": 0.9044397813242034,
      "train_speed(iter/s)": 0.032212
    },
    {
      "epoch": 0.9839655088892119,
      "grad_norm": 0.10839894413948059,
      "learning_rate": 2.1216228022679638e-07,
      "loss": 0.3413659334182739,
      "memory(GiB)": 78.33,
      "step": 5078,
      "token_acc": 0.8964040304440267,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.9841592791745386,
      "grad_norm": 0.08885491639375687,
      "learning_rate": 2.070819029237003e-07,
      "loss": 0.29514098167419434,
      "memory(GiB)": 78.33,
      "step": 5079,
      "token_acc": 0.9091107924858441,
      "train_speed(iter/s)": 0.032213
    },
    {
      "epoch": 0.9843530494598653,
      "grad_norm": 0.09718841314315796,
      "learning_rate": 2.0206304959716756e-07,
      "loss": 0.3231916129589081,
      "memory(GiB)": 78.33,
      "step": 5080,
      "token_acc": 0.9021267809209168,
      "train_speed(iter/s)": 0.032214
    },
    {
      "epoch": 0.9845468197451921,
      "grad_norm": 0.11555752903223038,
      "learning_rate": 1.971057223085659e-07,
      "loss": 0.3575592637062073,
      "memory(GiB)": 78.33,
      "step": 5081,
      "token_acc": 0.893569844789357,
      "train_speed(iter/s)": 0.032214
    },
    {
      "epoch": 0.9847405900305188,
      "grad_norm": 0.11206049472093582,
      "learning_rate": 1.9220992309399997e-07,
      "loss": 0.37476587295532227,
      "memory(GiB)": 78.33,
      "step": 5082,
      "token_acc": 0.8887654848355404,
      "train_speed(iter/s)": 0.032215
    },
    {
      "epoch": 0.9849343603158456,
      "grad_norm": 0.08786389976739883,
      "learning_rate": 1.873756539643112e-07,
      "loss": 0.2945099174976349,
      "memory(GiB)": 78.33,
      "step": 5083,
      "token_acc": 0.9125009193655153,
      "train_speed(iter/s)": 0.032215
    },
    {
      "epoch": 0.9851281306011723,
      "grad_norm": 0.09714756906032562,
      "learning_rate": 1.8260291690506135e-07,
      "loss": 0.32386794686317444,
      "memory(GiB)": 78.33,
      "step": 5084,
      "token_acc": 0.9020605635215081,
      "train_speed(iter/s)": 0.032216
    },
    {
      "epoch": 0.9853219008864991,
      "grad_norm": 0.09232458472251892,
      "learning_rate": 1.7789171387654898e-07,
      "loss": 0.3251858949661255,
      "memory(GiB)": 78.33,
      "step": 5085,
      "token_acc": 0.901725535610886,
      "train_speed(iter/s)": 0.032216
    },
    {
      "epoch": 0.9855156711718258,
      "grad_norm": 0.11348365992307663,
      "learning_rate": 1.7324204681377628e-07,
      "loss": 0.35106420516967773,
      "memory(GiB)": 78.33,
      "step": 5086,
      "token_acc": 0.895291405992756,
      "train_speed(iter/s)": 0.032217
    },
    {
      "epoch": 0.9857094414571526,
      "grad_norm": 0.09008847177028656,
      "learning_rate": 1.6865391762649893e-07,
      "loss": 0.2719075083732605,
      "memory(GiB)": 78.33,
      "step": 5087,
      "token_acc": 0.9164248403946604,
      "train_speed(iter/s)": 0.032218
    },
    {
      "epoch": 0.9859032117424793,
      "grad_norm": 0.1007675901055336,
      "learning_rate": 1.6412732819919284e-07,
      "loss": 0.29824161529541016,
      "memory(GiB)": 78.33,
      "step": 5088,
      "token_acc": 0.9085992132867133,
      "train_speed(iter/s)": 0.032218
    },
    {
      "epoch": 0.986096982027806,
      "grad_norm": 0.12334459275007248,
      "learning_rate": 1.596622803910208e-07,
      "loss": 0.3122141361236572,
      "memory(GiB)": 78.33,
      "step": 5089,
      "token_acc": 0.90404706917409,
      "train_speed(iter/s)": 0.032219
    },
    {
      "epoch": 0.9862907523131328,
      "grad_norm": 0.1014440655708313,
      "learning_rate": 1.552587760359325e-07,
      "loss": 0.31267303228378296,
      "memory(GiB)": 78.33,
      "step": 5090,
      "token_acc": 0.9055354659248956,
      "train_speed(iter/s)": 0.032219
    },
    {
      "epoch": 0.9864845225984595,
      "grad_norm": 0.11133712530136108,
      "learning_rate": 1.5091681694253122e-07,
      "loss": 0.35740742087364197,
      "memory(GiB)": 78.33,
      "step": 5091,
      "token_acc": 0.8951927600808125,
      "train_speed(iter/s)": 0.03222
    },
    {
      "epoch": 0.9866782928837863,
      "grad_norm": 0.10128425806760788,
      "learning_rate": 1.4663640489420702e-07,
      "loss": 0.3182716965675354,
      "memory(GiB)": 78.33,
      "step": 5092,
      "token_acc": 0.9042219609160648,
      "train_speed(iter/s)": 0.03222
    },
    {
      "epoch": 0.986872063169113,
      "grad_norm": 0.09385867416858673,
      "learning_rate": 1.4241754164903696e-07,
      "loss": 0.2903617322444916,
      "memory(GiB)": 78.33,
      "step": 5093,
      "token_acc": 0.9133545725178879,
      "train_speed(iter/s)": 0.032221
    },
    {
      "epoch": 0.9870658334544398,
      "grad_norm": 0.09520326554775238,
      "learning_rate": 1.3826022893980159e-07,
      "loss": 0.32819217443466187,
      "memory(GiB)": 78.33,
      "step": 5094,
      "token_acc": 0.901990578939371,
      "train_speed(iter/s)": 0.032221
    },
    {
      "epoch": 0.9872596037397665,
      "grad_norm": 0.09930814802646637,
      "learning_rate": 1.3416446847401842e-07,
      "loss": 0.30430757999420166,
      "memory(GiB)": 78.33,
      "step": 5095,
      "token_acc": 0.9086877119749543,
      "train_speed(iter/s)": 0.032222
    },
    {
      "epoch": 0.9874533740250933,
      "grad_norm": 0.10373541712760925,
      "learning_rate": 1.3013026193395836e-07,
      "loss": 0.3304864168167114,
      "memory(GiB)": 78.33,
      "step": 5096,
      "token_acc": 0.9003137958085845,
      "train_speed(iter/s)": 0.032223
    },
    {
      "epoch": 0.98764714431042,
      "grad_norm": 0.09495776146650314,
      "learning_rate": 1.2615761097654608e-07,
      "loss": 0.31185808777809143,
      "memory(GiB)": 78.33,
      "step": 5097,
      "token_acc": 0.9045740484060134,
      "train_speed(iter/s)": 0.032223
    },
    {
      "epoch": 0.9878409145957467,
      "grad_norm": 0.10554935038089752,
      "learning_rate": 1.2224651723347634e-07,
      "loss": 0.3295019567012787,
      "memory(GiB)": 78.33,
      "step": 5098,
      "token_acc": 0.9011663040850858,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.9880346848810735,
      "grad_norm": 0.09783808141946793,
      "learning_rate": 1.1839698231113082e-07,
      "loss": 0.32202041149139404,
      "memory(GiB)": 78.33,
      "step": 5099,
      "token_acc": 0.9030985169491526,
      "train_speed(iter/s)": 0.032224
    },
    {
      "epoch": 0.9882284551664002,
      "grad_norm": 0.1201200857758522,
      "learning_rate": 1.1460900779061144e-07,
      "loss": 0.31718727946281433,
      "memory(GiB)": 78.33,
      "step": 5100,
      "token_acc": 0.9047082558230932,
      "train_speed(iter/s)": 0.032225
    },
    {
      "epoch": 0.988422225451727,
      "grad_norm": 0.10204530507326126,
      "learning_rate": 1.1088259522777365e-07,
      "loss": 0.33351755142211914,
      "memory(GiB)": 78.33,
      "step": 5101,
      "token_acc": 0.90131747431921,
      "train_speed(iter/s)": 0.032225
    },
    {
      "epoch": 0.9886159957370537,
      "grad_norm": 0.1120908334851265,
      "learning_rate": 1.0721774615310985e-07,
      "loss": 0.35009273886680603,
      "memory(GiB)": 78.33,
      "step": 5102,
      "token_acc": 0.895470053070508,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.9888097660223805,
      "grad_norm": 0.10349807143211365,
      "learning_rate": 1.0361446207189928e-07,
      "loss": 0.3232646584510803,
      "memory(GiB)": 78.33,
      "step": 5103,
      "token_acc": 0.9045182551383227,
      "train_speed(iter/s)": 0.032226
    },
    {
      "epoch": 0.9890035363077072,
      "grad_norm": 0.10206515341997147,
      "learning_rate": 1.0007274446409141e-07,
      "loss": 0.32843074202537537,
      "memory(GiB)": 78.33,
      "step": 5104,
      "token_acc": 0.9033802574615097,
      "train_speed(iter/s)": 0.032227
    },
    {
      "epoch": 0.989197306593034,
      "grad_norm": 0.10101883858442307,
      "learning_rate": 9.65925947843893e-08,
      "loss": 0.3663754165172577,
      "memory(GiB)": 78.33,
      "step": 5105,
      "token_acc": 0.8890428585568643,
      "train_speed(iter/s)": 0.032227
    },
    {
      "epoch": 0.9893910768783607,
      "grad_norm": 0.10111556947231293,
      "learning_rate": 9.317401446216621e-08,
      "loss": 0.3458866775035858,
      "memory(GiB)": 78.33,
      "step": 5106,
      "token_acc": 0.898968688533305,
      "train_speed(iter/s)": 0.032228
    },
    {
      "epoch": 0.9895848471636874,
      "grad_norm": 0.10001836717128754,
      "learning_rate": 8.981700490151567e-08,
      "loss": 0.34546831250190735,
      "memory(GiB)": 78.33,
      "step": 5107,
      "token_acc": 0.8980206216602694,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.9897786174490142,
      "grad_norm": 0.10675647109746933,
      "learning_rate": 8.652156748126804e-08,
      "loss": 0.34267866611480713,
      "memory(GiB)": 78.33,
      "step": 5108,
      "token_acc": 0.8978237122930847,
      "train_speed(iter/s)": 0.032229
    },
    {
      "epoch": 0.9899723877343409,
      "grad_norm": 0.10720111429691315,
      "learning_rate": 8.328770355495729e-08,
      "loss": 0.3194417953491211,
      "memory(GiB)": 78.33,
      "step": 5109,
      "token_acc": 0.9027690371302706,
      "train_speed(iter/s)": 0.03223
    },
    {
      "epoch": 0.9901661580196677,
      "grad_norm": 0.08818720281124115,
      "learning_rate": 8.011541445078762e-08,
      "loss": 0.2857103645801544,
      "memory(GiB)": 78.33,
      "step": 5110,
      "token_acc": 0.91203895313451,
      "train_speed(iter/s)": 0.03223
    },
    {
      "epoch": 0.9903599283049944,
      "grad_norm": 0.09907843917608261,
      "learning_rate": 7.700470147173343e-08,
      "loss": 0.3331596553325653,
      "memory(GiB)": 78.33,
      "step": 5111,
      "token_acc": 0.8999431495167709,
      "train_speed(iter/s)": 0.032231
    },
    {
      "epoch": 0.9905536985903212,
      "grad_norm": 0.13084760308265686,
      "learning_rate": 7.395556589542274e-08,
      "loss": 0.3520573377609253,
      "memory(GiB)": 78.33,
      "step": 5112,
      "token_acc": 0.8953529427741111,
      "train_speed(iter/s)": 0.032231
    },
    {
      "epoch": 0.9907474688756479,
      "grad_norm": 0.09977789223194122,
      "learning_rate": 7.09680089742537e-08,
      "loss": 0.34126073122024536,
      "memory(GiB)": 78.33,
      "step": 5113,
      "token_acc": 0.8968176914778857,
      "train_speed(iter/s)": 0.032232
    },
    {
      "epoch": 0.9909412391609747,
      "grad_norm": 0.0944044291973114,
      "learning_rate": 6.804203193524483e-08,
      "loss": 0.32394513487815857,
      "memory(GiB)": 78.33,
      "step": 5114,
      "token_acc": 0.9009689518649718,
      "train_speed(iter/s)": 0.032232
    },
    {
      "epoch": 0.9911350094463014,
      "grad_norm": 0.09428620338439941,
      "learning_rate": 6.517763598021808e-08,
      "loss": 0.2987945079803467,
      "memory(GiB)": 78.33,
      "step": 5115,
      "token_acc": 0.9094635777663906,
      "train_speed(iter/s)": 0.032233
    },
    {
      "epoch": 0.9913287797316281,
      "grad_norm": 0.10901875793933868,
      "learning_rate": 6.237482228563239e-08,
      "loss": 0.3245546817779541,
      "memory(GiB)": 78.33,
      "step": 5116,
      "token_acc": 0.9043344214726151,
      "train_speed(iter/s)": 0.032234
    },
    {
      "epoch": 0.9915225500169549,
      "grad_norm": 0.10262294858694077,
      "learning_rate": 5.963359200270024e-08,
      "loss": 0.3375054895877838,
      "memory(GiB)": 78.33,
      "step": 5117,
      "token_acc": 0.8971592035573682,
      "train_speed(iter/s)": 0.032234
    },
    {
      "epoch": 0.9917163203022816,
      "grad_norm": 0.0967864915728569,
      "learning_rate": 5.6953946257287665e-08,
      "loss": 0.3116954267024994,
      "memory(GiB)": 78.33,
      "step": 5118,
      "token_acc": 0.9069773955911599,
      "train_speed(iter/s)": 0.032235
    },
    {
      "epoch": 0.9919100905876084,
      "grad_norm": 0.09405119717121124,
      "learning_rate": 5.433588615003093e-08,
      "loss": 0.30092549324035645,
      "memory(GiB)": 78.33,
      "step": 5119,
      "token_acc": 0.9092479884464617,
      "train_speed(iter/s)": 0.032235
    },
    {
      "epoch": 0.9921038608729351,
      "grad_norm": 0.0960017740726471,
      "learning_rate": 5.177941275620323e-08,
      "loss": 0.29913923144340515,
      "memory(GiB)": 78.33,
      "step": 5120,
      "token_acc": 0.9077315436241611,
      "train_speed(iter/s)": 0.032236
    },
    {
      "epoch": 0.9922976311582619,
      "grad_norm": 0.10961098968982697,
      "learning_rate": 4.928452712584796e-08,
      "loss": 0.3457680344581604,
      "memory(GiB)": 78.33,
      "step": 5121,
      "token_acc": 0.8968839910971175,
      "train_speed(iter/s)": 0.032236
    },
    {
      "epoch": 0.9924914014435886,
      "grad_norm": 0.10494101792573929,
      "learning_rate": 4.6851230283678766e-08,
      "loss": 0.32835787534713745,
      "memory(GiB)": 78.33,
      "step": 5122,
      "token_acc": 0.9018704634282524,
      "train_speed(iter/s)": 0.032237
    },
    {
      "epoch": 0.9926851717289153,
      "grad_norm": 0.09790334105491638,
      "learning_rate": 4.44795232290962e-08,
      "loss": 0.30867090821266174,
      "memory(GiB)": 78.33,
      "step": 5123,
      "token_acc": 0.9083395542284313,
      "train_speed(iter/s)": 0.032237
    },
    {
      "epoch": 0.9928789420142421,
      "grad_norm": 0.09199753403663635,
      "learning_rate": 4.216940693622106e-08,
      "loss": 0.3020906448364258,
      "memory(GiB)": 78.33,
      "step": 5124,
      "token_acc": 0.9092012383900929,
      "train_speed(iter/s)": 0.032238
    },
    {
      "epoch": 0.9930727122995688,
      "grad_norm": 0.10397046059370041,
      "learning_rate": 3.9920882353911e-08,
      "loss": 0.3302524983882904,
      "memory(GiB)": 78.33,
      "step": 5125,
      "token_acc": 0.9000474552141416,
      "train_speed(iter/s)": 0.032238
    },
    {
      "epoch": 0.9932664825848956,
      "grad_norm": 0.1026199460029602,
      "learning_rate": 3.773395040567728e-08,
      "loss": 0.3645437955856323,
      "memory(GiB)": 78.33,
      "step": 5126,
      "token_acc": 0.8931143232588699,
      "train_speed(iter/s)": 0.032239
    },
    {
      "epoch": 0.9934602528702223,
      "grad_norm": 0.1036728173494339,
      "learning_rate": 3.56086119897514e-08,
      "loss": 0.34227946400642395,
      "memory(GiB)": 78.33,
      "step": 5127,
      "token_acc": 0.8958297382801266,
      "train_speed(iter/s)": 0.03224
    },
    {
      "epoch": 0.9936540231555491,
      "grad_norm": 0.10324281454086304,
      "learning_rate": 3.354486797906841e-08,
      "loss": 0.3703176975250244,
      "memory(GiB)": 78.33,
      "step": 5128,
      "token_acc": 0.8903992961943412,
      "train_speed(iter/s)": 0.03224
    },
    {
      "epoch": 0.9938477934408758,
      "grad_norm": 0.10356750339269638,
      "learning_rate": 3.154271922125029e-08,
      "loss": 0.3422203063964844,
      "memory(GiB)": 78.33,
      "step": 5129,
      "token_acc": 0.9005716619028175,
      "train_speed(iter/s)": 0.032241
    },
    {
      "epoch": 0.9940415637262026,
      "grad_norm": 0.10634325444698334,
      "learning_rate": 2.960216653865588e-08,
      "loss": 0.33742478489875793,
      "memory(GiB)": 78.33,
      "step": 5130,
      "token_acc": 0.8997738043946575,
      "train_speed(iter/s)": 0.032241
    },
    {
      "epoch": 0.9942353340115293,
      "grad_norm": 0.09013240784406662,
      "learning_rate": 2.7723210728314292e-08,
      "loss": 0.29299721121788025,
      "memory(GiB)": 78.33,
      "step": 5131,
      "token_acc": 0.910865125192264,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.994429104296856,
      "grad_norm": 0.09943090379238129,
      "learning_rate": 2.5905852561958208e-08,
      "loss": 0.3339230716228485,
      "memory(GiB)": 78.33,
      "step": 5132,
      "token_acc": 0.8997598211081357,
      "train_speed(iter/s)": 0.032242
    },
    {
      "epoch": 0.9946228745821828,
      "grad_norm": 0.10124190896749496,
      "learning_rate": 2.415009278604052e-08,
      "loss": 0.3518436551094055,
      "memory(GiB)": 78.33,
      "step": 5133,
      "token_acc": 0.8950833333333333,
      "train_speed(iter/s)": 0.032243
    },
    {
      "epoch": 0.9948166448675095,
      "grad_norm": 0.09643140435218811,
      "learning_rate": 2.245593212166774e-08,
      "loss": 0.3150833249092102,
      "memory(GiB)": 78.33,
      "step": 5134,
      "token_acc": 0.9049085264157237,
      "train_speed(iter/s)": 0.032243
    },
    {
      "epoch": 0.9950104151528363,
      "grad_norm": 0.10642070323228836,
      "learning_rate": 2.0823371264699907e-08,
      "loss": 0.33650028705596924,
      "memory(GiB)": 78.33,
      "step": 5135,
      "token_acc": 0.9001945581787031,
      "train_speed(iter/s)": 0.032244
    },
    {
      "epoch": 0.995204185438163,
      "grad_norm": 0.10311167687177658,
      "learning_rate": 1.9252410885683965e-08,
      "loss": 0.3397340178489685,
      "memory(GiB)": 78.33,
      "step": 5136,
      "token_acc": 0.8996154508408426,
      "train_speed(iter/s)": 0.032245
    },
    {
      "epoch": 0.9953979557234898,
      "grad_norm": 0.10818962752819061,
      "learning_rate": 1.7743051629837135e-08,
      "loss": 0.3583690822124481,
      "memory(GiB)": 78.33,
      "step": 5137,
      "token_acc": 0.8913712208308734,
      "train_speed(iter/s)": 0.032245
    },
    {
      "epoch": 0.9955917260088165,
      "grad_norm": 0.11495489627122879,
      "learning_rate": 1.6295294117080192e-08,
      "loss": 0.3481709063053131,
      "memory(GiB)": 78.33,
      "step": 5138,
      "token_acc": 0.8962670979044539,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.9957854962941433,
      "grad_norm": 0.09410503506660461,
      "learning_rate": 1.490913894208745e-08,
      "loss": 0.30858170986175537,
      "memory(GiB)": 78.33,
      "step": 5139,
      "token_acc": 0.9072154599071401,
      "train_speed(iter/s)": 0.032246
    },
    {
      "epoch": 0.99597926657947,
      "grad_norm": 0.10258731245994568,
      "learning_rate": 1.3584586674153519e-08,
      "loss": 0.3513551652431488,
      "memory(GiB)": 78.33,
      "step": 5140,
      "token_acc": 0.8945655624933856,
      "train_speed(iter/s)": 0.032247
    },
    {
      "epoch": 0.9961730368647967,
      "grad_norm": 0.10465652495622635,
      "learning_rate": 1.2321637857326538e-08,
      "loss": 0.32509326934814453,
      "memory(GiB)": 78.33,
      "step": 5141,
      "token_acc": 0.9022335312411648,
      "train_speed(iter/s)": 0.032247
    },
    {
      "epoch": 0.9963668071501235,
      "grad_norm": 0.17421969771385193,
      "learning_rate": 1.112029301032491e-08,
      "loss": 0.3402435779571533,
      "memory(GiB)": 78.33,
      "step": 5142,
      "token_acc": 0.8987543069175722,
      "train_speed(iter/s)": 0.032248
    },
    {
      "epoch": 0.9965605774354502,
      "grad_norm": 0.09737160056829453,
      "learning_rate": 9.980552626587257e-09,
      "loss": 0.31006062030792236,
      "memory(GiB)": 78.33,
      "step": 5143,
      "token_acc": 0.9062544199598314,
      "train_speed(iter/s)": 0.032248
    },
    {
      "epoch": 0.996754347720777,
      "grad_norm": 0.0954391285777092,
      "learning_rate": 8.902417174205812e-09,
      "loss": 0.3203745186328888,
      "memory(GiB)": 78.33,
      "step": 5144,
      "token_acc": 0.903609002530297,
      "train_speed(iter/s)": 0.032249
    },
    {
      "epoch": 0.9969481180061037,
      "grad_norm": 0.10394702851772308,
      "learning_rate": 7.885887096026333e-09,
      "loss": 0.33040565252304077,
      "memory(GiB)": 78.33,
      "step": 5145,
      "token_acc": 0.9019830523281209,
      "train_speed(iter/s)": 0.03225
    },
    {
      "epoch": 0.9971418882914305,
      "grad_norm": 0.1102806031703949,
      "learning_rate": 6.930962809564844e-09,
      "loss": 0.38149887323379517,
      "memory(GiB)": 78.33,
      "step": 5146,
      "token_acc": 0.8889834487615917,
      "train_speed(iter/s)": 0.03225
    },
    {
      "epoch": 0.9973356585767572,
      "grad_norm": 0.09441733360290527,
      "learning_rate": 6.0376447070242805e-09,
      "loss": 0.2967412769794464,
      "memory(GiB)": 78.33,
      "step": 5147,
      "token_acc": 0.910029761147712,
      "train_speed(iter/s)": 0.032251
    },
    {
      "epoch": 0.997529428862084,
      "grad_norm": 0.09879467636346817,
      "learning_rate": 5.205933155311149e-09,
      "loss": 0.31042206287384033,
      "memory(GiB)": 78.33,
      "step": 5148,
      "token_acc": 0.9073388532511939,
      "train_speed(iter/s)": 0.032251
    },
    {
      "epoch": 0.9977231991474107,
      "grad_norm": 0.09496015310287476,
      "learning_rate": 4.435828496035521e-09,
      "loss": 0.30281057953834534,
      "memory(GiB)": 78.33,
      "step": 5149,
      "token_acc": 0.9083017847485128,
      "train_speed(iter/s)": 0.032252
    },
    {
      "epoch": 0.9979169694327374,
      "grad_norm": 0.09966279566287994,
      "learning_rate": 3.727331045511039e-09,
      "loss": 0.3183283805847168,
      "memory(GiB)": 78.33,
      "step": 5150,
      "token_acc": 0.9036394691893312,
      "train_speed(iter/s)": 0.032252
    },
    {
      "epoch": 0.9981107397180642,
      "grad_norm": 0.0968567430973053,
      "learning_rate": 3.0804410947216084e-09,
      "loss": 0.34515106678009033,
      "memory(GiB)": 78.33,
      "step": 5151,
      "token_acc": 0.8957895251601545,
      "train_speed(iter/s)": 0.032253
    },
    {
      "epoch": 0.9983045100033909,
      "grad_norm": 0.09464351087808609,
      "learning_rate": 2.4951589093713533e-09,
      "loss": 0.32018736004829407,
      "memory(GiB)": 78.33,
      "step": 5152,
      "token_acc": 0.90508582795118,
      "train_speed(iter/s)": 0.032253
    },
    {
      "epoch": 0.9984982802887177,
      "grad_norm": 0.0978541150689125,
      "learning_rate": 1.9714847298513135e-09,
      "loss": 0.3221741318702698,
      "memory(GiB)": 78.33,
      "step": 5153,
      "token_acc": 0.90467557008248,
      "train_speed(iter/s)": 0.032254
    },
    {
      "epoch": 0.9986920505740444,
      "grad_norm": 0.09578622877597809,
      "learning_rate": 1.5094187712394456e-09,
      "loss": 0.3303835988044739,
      "memory(GiB)": 78.33,
      "step": 5154,
      "token_acc": 0.9011171856429759,
      "train_speed(iter/s)": 0.032255
    },
    {
      "epoch": 0.9988858208593712,
      "grad_norm": 0.09848620742559433,
      "learning_rate": 1.1089612233339261e-09,
      "loss": 0.30692458152770996,
      "memory(GiB)": 78.33,
      "step": 5155,
      "token_acc": 0.9065349757288363,
      "train_speed(iter/s)": 0.032255
    },
    {
      "epoch": 0.9990795911446979,
      "grad_norm": 0.09920880943536758,
      "learning_rate": 7.701122505865409e-10,
      "loss": 0.3253172039985657,
      "memory(GiB)": 78.33,
      "step": 5156,
      "token_acc": 0.9011366073343341,
      "train_speed(iter/s)": 0.032256
    },
    {
      "epoch": 0.9992733614300247,
      "grad_norm": 0.09585346281528473,
      "learning_rate": 4.928719922026037e-10,
      "loss": 0.30521735548973083,
      "memory(GiB)": 78.33,
      "step": 5157,
      "token_acc": 0.906159781992823,
      "train_speed(iter/s)": 0.032256
    },
    {
      "epoch": 0.9994671317153514,
      "grad_norm": 0.10266774892807007,
      "learning_rate": 2.772405620410367e-10,
      "loss": 0.341879278421402,
      "memory(GiB)": 78.33,
      "step": 5158,
      "token_acc": 0.8973592287271203,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.9996609020006781,
      "grad_norm": 0.10256272554397583,
      "learning_rate": 1.2321804866433082e-10,
      "loss": 0.3247736990451813,
      "memory(GiB)": 78.33,
      "step": 5159,
      "token_acc": 0.9010874626783905,
      "train_speed(iter/s)": 0.032257
    },
    {
      "epoch": 0.9998546722860049,
      "grad_norm": 0.09452533721923828,
      "learning_rate": 3.0804515321891657e-11,
      "loss": 0.3123696446418762,
      "memory(GiB)": 78.33,
      "step": 5160,
      "token_acc": 0.9040158570691574,
      "train_speed(iter/s)": 0.032258
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.1213284358382225,
      "learning_rate": 0.0,
      "loss": 0.31489574909210205,
      "memory(GiB)": 78.33,
      "step": 5161,
      "token_acc": 0.9069757440220196,
      "train_speed(iter/s)": 0.03226
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.37820467352867126,
      "eval_runtime": 1344.7251,
      "eval_samples_per_second": 5.019,
      "eval_steps_per_second": 5.019,
      "eval_token_acc": 0.9026974626241785,
      "step": 5161
    }
  ],
  "logging_steps": 1,
  "max_steps": 5161,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 200,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 5.701470853092039e+19,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}