{
  "best_global_step": 1200,
  "best_metric": 0.43056953,
  "best_model_checkpoint": "/openpai_config/sft/Long_Cot_data/Stage1-380k-24k-length-Qwen3-8B-Base-resume-iter4600-4p-3e-5/v0-20250826-235423/checkpoint-1200",
  "epoch": 2.2665292804396358,
  "eval_steps": 300,
  "global_step": 3300,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0006869311351537008,
      "grad_norm": 0.07274238765239716,
      "learning_rate": 2.9999997817680842e-05,
      "loss": 0.2985985279083252,
      "memory(GiB)": 56.61,
      "step": 1,
      "token_acc": 0.8812937590030485,
      "train_speed(iter/s)": 0.013692
    },
    {
      "epoch": 0.013738622703074016,
      "grad_norm": 0.10364966094493866,
      "learning_rate": 2.9999127080781484e-05,
      "loss": 0.32278080990439967,
      "memory(GiB)": 70.37,
      "step": 20,
      "token_acc": 0.8852460572314409,
      "train_speed(iter/s)": 0.033578
    },
    {
      "epoch": 0.027477245406148033,
      "grad_norm": 0.0828627347946167,
      "learning_rate": 2.999650842472434e-05,
      "loss": 0.3287534713745117,
      "memory(GiB)": 70.37,
      "step": 40,
      "token_acc": 0.8729899276605303,
      "train_speed(iter/s)": 0.035269
    },
    {
      "epoch": 0.04121586810922205,
      "grad_norm": 0.075782909989357,
      "learning_rate": 2.9992144336611927e-05,
      "loss": 0.32648520469665526,
      "memory(GiB)": 70.37,
      "step": 60,
      "token_acc": 0.871392951813677,
      "train_speed(iter/s)": 0.03605
    },
    {
      "epoch": 0.054954490812296065,
      "grad_norm": 0.09333578497171402,
      "learning_rate": 2.998603532437709e-05,
      "loss": 0.3258840799331665,
      "memory(GiB)": 70.37,
      "step": 80,
      "token_acc": 0.8820731351865331,
      "train_speed(iter/s)": 0.036483
    },
    {
      "epoch": 0.06869311351537008,
      "grad_norm": 0.08618636429309845,
      "learning_rate": 2.9978182099043062e-05,
      "loss": 0.3262542724609375,
      "memory(GiB)": 70.37,
      "step": 100,
      "token_acc": 0.8849732393640833,
      "train_speed(iter/s)": 0.036776
    },
    {
      "epoch": 0.0824317362184441,
      "grad_norm": 0.08525851368904114,
      "learning_rate": 2.9968585574640675e-05,
      "loss": 0.32371375560760496,
      "memory(GiB)": 70.37,
      "step": 120,
      "token_acc": 0.8871000703215756,
      "train_speed(iter/s)": 0.036983
    },
    {
      "epoch": 0.09617035892151812,
      "grad_norm": 0.08820851147174835,
      "learning_rate": 2.995724686810202e-05,
      "loss": 0.3224189281463623,
      "memory(GiB)": 70.37,
      "step": 140,
      "token_acc": 0.8856544898235427,
      "train_speed(iter/s)": 0.03715
    },
    {
      "epoch": 0.10990898162459213,
      "grad_norm": 0.0837075412273407,
      "learning_rate": 2.9944167299130397e-05,
      "loss": 0.32613368034362794,
      "memory(GiB)": 70.4,
      "step": 160,
      "token_acc": 0.8848279812123834,
      "train_speed(iter/s)": 0.037079
    },
    {
      "epoch": 0.12364760432766615,
      "grad_norm": 0.09300094842910767,
      "learning_rate": 2.9929348390046766e-05,
      "loss": 0.32998530864715575,
      "memory(GiB)": 70.4,
      "step": 180,
      "token_acc": 0.8633912245766957,
      "train_speed(iter/s)": 0.037181
    },
    {
      "epoch": 0.13738622703074016,
      "grad_norm": 0.08325231075286865,
      "learning_rate": 2.9912791865612525e-05,
      "loss": 0.3349442958831787,
      "memory(GiB)": 70.4,
      "step": 200,
      "token_acc": 0.8835910835510771,
      "train_speed(iter/s)": 0.037269
    },
    {
      "epoch": 0.15112484973381418,
      "grad_norm": 0.08610066026449203,
      "learning_rate": 2.9894499652828798e-05,
      "loss": 0.33656883239746094,
      "memory(GiB)": 70.4,
      "step": 220,
      "token_acc": 0.8754293623532549,
      "train_speed(iter/s)": 0.037356
    },
    {
      "epoch": 0.1648634724368882,
      "grad_norm": 0.09172473102807999,
      "learning_rate": 2.9874473880712125e-05,
      "loss": 0.3357390403747559,
      "memory(GiB)": 70.4,
      "step": 240,
      "token_acc": 0.8816886144038251,
      "train_speed(iter/s)": 0.037427
    },
    {
      "epoch": 0.17860209513996222,
      "grad_norm": 0.08784262090921402,
      "learning_rate": 2.9852716880046687e-05,
      "loss": 0.33715412616729734,
      "memory(GiB)": 70.4,
      "step": 260,
      "token_acc": 0.8768648698038725,
      "train_speed(iter/s)": 0.037172
    },
    {
      "epoch": 0.19234071784303625,
      "grad_norm": 0.08109364658594131,
      "learning_rate": 2.9829231183113013e-05,
      "loss": 0.3330291509628296,
      "memory(GiB)": 70.4,
      "step": 280,
      "token_acc": 0.884993071444176,
      "train_speed(iter/s)": 0.037251
    },
    {
      "epoch": 0.20607934054611024,
      "grad_norm": 0.08930086344480515,
      "learning_rate": 2.980401952339328e-05,
      "loss": 0.32976431846618653,
      "memory(GiB)": 70.4,
      "step": 300,
      "token_acc": 0.8701908549811445,
      "train_speed(iter/s)": 0.037312
    },
    {
      "epoch": 0.20607934054611024,
      "eval_loss": 0.4335128664970398,
      "eval_runtime": 97.7946,
      "eval_samples_per_second": 38.489,
      "eval_steps_per_second": 0.603,
      "eval_token_acc": 0.8434016162256411,
      "step": 300
    },
    {
      "epoch": 0.21981796324918426,
      "grad_norm": 0.08906491845846176,
      "learning_rate": 2.9777084835253107e-05,
      "loss": 0.3341225624084473,
      "memory(GiB)": 72.41,
      "step": 320,
      "token_acc": 0.8633943248975853,
      "train_speed(iter/s)": 0.036714
    },
    {
      "epoch": 0.23355658595225828,
      "grad_norm": 0.09569748491048813,
      "learning_rate": 2.9748430253600103e-05,
      "loss": 0.3317814826965332,
      "memory(GiB)": 72.41,
      "step": 340,
      "token_acc": 0.8690857078605088,
      "train_speed(iter/s)": 0.036742
    },
    {
      "epoch": 0.2472952086553323,
      "grad_norm": 0.09385869652032852,
      "learning_rate": 2.9718059113518926e-05,
      "loss": 0.33147258758544923,
      "memory(GiB)": 72.41,
      "step": 360,
      "token_acc": 0.8744091464585358,
      "train_speed(iter/s)": 0.036789
    },
    {
      "epoch": 0.2610338313584063,
      "grad_norm": 0.09747012704610825,
      "learning_rate": 2.9685974949883163e-05,
      "loss": 0.3316455841064453,
      "memory(GiB)": 72.41,
      "step": 380,
      "token_acc": 0.8744771091993391,
      "train_speed(iter/s)": 0.036834
    },
    {
      "epoch": 0.2747724540614803,
      "grad_norm": 0.0942377969622612,
      "learning_rate": 2.9652181496943888e-05,
      "loss": 0.33130803108215334,
      "memory(GiB)": 72.41,
      "step": 400,
      "token_acc": 0.8743932496768547,
      "train_speed(iter/s)": 0.036889
    },
    {
      "epoch": 0.28851107676455434,
      "grad_norm": 0.08894450962543488,
      "learning_rate": 2.9616682687895038e-05,
      "loss": 0.3286163806915283,
      "memory(GiB)": 72.41,
      "step": 420,
      "token_acc": 0.8871290646641795,
      "train_speed(iter/s)": 0.036939
    },
    {
      "epoch": 0.30224969946762836,
      "grad_norm": 0.09808226674795151,
      "learning_rate": 2.9579482654415627e-05,
      "loss": 0.3300768375396729,
      "memory(GiB)": 72.41,
      "step": 440,
      "token_acc": 0.8833857380634584,
      "train_speed(iter/s)": 0.036991
    },
    {
      "epoch": 0.3159883221707024,
      "grad_norm": 0.09648197889328003,
      "learning_rate": 2.9540585726188883e-05,
      "loss": 0.3316764831542969,
      "memory(GiB)": 72.41,
      "step": 460,
      "token_acc": 0.8862556346113715,
      "train_speed(iter/s)": 0.037032
    },
    {
      "epoch": 0.3297269448737764,
      "grad_norm": 0.08601760119199753,
      "learning_rate": 2.9499996430398296e-05,
      "loss": 0.3263013124465942,
      "memory(GiB)": 72.41,
      "step": 480,
      "token_acc": 0.8804854791092939,
      "train_speed(iter/s)": 0.037074
    },
    {
      "epoch": 0.34346556757685043,
      "grad_norm": 0.08579235523939133,
      "learning_rate": 2.945771949120071e-05,
      "loss": 0.33027398586273193,
      "memory(GiB)": 72.41,
      "step": 500,
      "token_acc": 0.8894077316408038,
      "train_speed(iter/s)": 0.037103
    },
    {
      "epoch": 0.35720419027992445,
      "grad_norm": 0.08523233234882355,
      "learning_rate": 2.9413759829176497e-05,
      "loss": 0.32302305698394773,
      "memory(GiB)": 72.41,
      "step": 520,
      "token_acc": 0.8851045135756845,
      "train_speed(iter/s)": 0.037131
    },
    {
      "epoch": 0.37094281298299847,
      "grad_norm": 0.08190900087356567,
      "learning_rate": 2.9368122560756822e-05,
      "loss": 0.3292397975921631,
      "memory(GiB)": 72.41,
      "step": 540,
      "token_acc": 0.8731282020961862,
      "train_speed(iter/s)": 0.037164
    },
    {
      "epoch": 0.3846814356860725,
      "grad_norm": 0.0842556357383728,
      "learning_rate": 2.9320812997628184e-05,
      "loss": 0.3262279748916626,
      "memory(GiB)": 72.41,
      "step": 560,
      "token_acc": 0.8791384953802229,
      "train_speed(iter/s)": 0.037189
    },
    {
      "epoch": 0.3984200583891465,
      "grad_norm": 0.09352608770132065,
      "learning_rate": 2.9271836646114166e-05,
      "loss": 0.3300283908843994,
      "memory(GiB)": 72.41,
      "step": 580,
      "token_acc": 0.8848605133294101,
      "train_speed(iter/s)": 0.037218
    },
    {
      "epoch": 0.4121586810922205,
      "grad_norm": 0.08545698970556259,
      "learning_rate": 2.922119920653457e-05,
      "loss": 0.33294997215270994,
      "memory(GiB)": 72.41,
      "step": 600,
      "token_acc": 0.8852764804825887,
      "train_speed(iter/s)": 0.037244
    },
    {
      "epoch": 0.4121586810922205,
      "eval_loss": 0.43278947472572327,
      "eval_runtime": 98.2411,
      "eval_samples_per_second": 38.314,
      "eval_steps_per_second": 0.601,
      "eval_token_acc": 0.8435347324817475,
      "step": 600
    },
    {
      "epoch": 0.4258973037952945,
      "grad_norm": 0.09264256060123444,
      "learning_rate": 2.916890657254194e-05,
      "loss": 0.32485041618347166,
      "memory(GiB)": 72.41,
      "step": 620,
      "token_acc": 0.8645305113582833,
      "train_speed(iter/s)": 0.036936
    },
    {
      "epoch": 0.4396359264983685,
      "grad_norm": 0.08746380358934402,
      "learning_rate": 2.9114964830435648e-05,
      "loss": 0.32976808547973635,
      "memory(GiB)": 72.41,
      "step": 640,
      "token_acc": 0.8707312036033282,
      "train_speed(iter/s)": 0.036935
    },
    {
      "epoch": 0.45337454920144254,
      "grad_norm": 0.08795125037431717,
      "learning_rate": 2.9059380258453473e-05,
      "loss": 0.3318798303604126,
      "memory(GiB)": 72.41,
      "step": 660,
      "token_acc": 0.8750859732168144,
      "train_speed(iter/s)": 0.03695
    },
    {
      "epoch": 0.46711317190451657,
      "grad_norm": 0.09344589710235596,
      "learning_rate": 2.9002159326040897e-05,
      "loss": 0.3324501752853394,
      "memory(GiB)": 72.41,
      "step": 680,
      "token_acc": 0.8820933572199229,
      "train_speed(iter/s)": 0.036974
    },
    {
      "epoch": 0.4808517946075906,
      "grad_norm": 0.08175963908433914,
      "learning_rate": 2.894330869309814e-05,
      "loss": 0.3342601776123047,
      "memory(GiB)": 72.41,
      "step": 700,
      "token_acc": 0.888566898590454,
      "train_speed(iter/s)": 0.037001
    },
    {
      "epoch": 0.4945904173106646,
      "grad_norm": 0.08012371510267258,
      "learning_rate": 2.8882835209205e-05,
      "loss": 0.3325735807418823,
      "memory(GiB)": 72.41,
      "step": 720,
      "token_acc": 0.8726016398435306,
      "train_speed(iter/s)": 0.03703
    },
    {
      "epoch": 0.5083290400137386,
      "grad_norm": 0.08838852494955063,
      "learning_rate": 2.8820745912823653e-05,
      "loss": 0.3366635799407959,
      "memory(GiB)": 72.41,
      "step": 740,
      "token_acc": 0.8669464276591397,
      "train_speed(iter/s)": 0.037058
    },
    {
      "epoch": 0.5220676627168126,
      "grad_norm": 0.08403000980615616,
      "learning_rate": 2.8757048030479438e-05,
      "loss": 0.33883938789367674,
      "memory(GiB)": 72.41,
      "step": 760,
      "token_acc": 0.8731299584725523,
      "train_speed(iter/s)": 0.037083
    },
    {
      "epoch": 0.5358062854198866,
      "grad_norm": 0.09687156975269318,
      "learning_rate": 2.8691748975919784e-05,
      "loss": 0.3338437557220459,
      "memory(GiB)": 72.41,
      "step": 780,
      "token_acc": 0.8848287485135103,
      "train_speed(iter/s)": 0.037115
    },
    {
      "epoch": 0.5495449081229606,
      "grad_norm": 0.08311958611011505,
      "learning_rate": 2.862485634925131e-05,
      "loss": 0.3358239889144897,
      "memory(GiB)": 72.41,
      "step": 800,
      "token_acc": 0.8782527593369446,
      "train_speed(iter/s)": 0.037145
    },
    {
      "epoch": 0.5632835308260347,
      "grad_norm": 0.08971302956342697,
      "learning_rate": 2.855637793605527e-05,
      "loss": 0.3317149877548218,
      "memory(GiB)": 72.41,
      "step": 820,
      "token_acc": 0.8807387938176727,
      "train_speed(iter/s)": 0.037168
    },
    {
      "epoch": 0.5770221535291087,
      "grad_norm": 0.08680903911590576,
      "learning_rate": 2.848632170648139e-05,
      "loss": 0.3325679779052734,
      "memory(GiB)": 72.41,
      "step": 840,
      "token_acc": 0.8763491204567425,
      "train_speed(iter/s)": 0.037194
    },
    {
      "epoch": 0.5907607762321827,
      "grad_norm": 0.08324452489614487,
      "learning_rate": 2.8414695814320224e-05,
      "loss": 0.3364755868911743,
      "memory(GiB)": 72.41,
      "step": 860,
      "token_acc": 0.8762890432412627,
      "train_speed(iter/s)": 0.037217
    },
    {
      "epoch": 0.6044993989352567,
      "grad_norm": 0.08400746434926987,
      "learning_rate": 2.834150859605415e-05,
      "loss": 0.3361694812774658,
      "memory(GiB)": 72.41,
      "step": 880,
      "token_acc": 0.8813631606405008,
      "train_speed(iter/s)": 0.037241
    },
    {
      "epoch": 0.6182380216383307,
      "grad_norm": 0.08885052800178528,
      "learning_rate": 2.8266768569887078e-05,
      "loss": 0.33670692443847655,
      "memory(GiB)": 72.41,
      "step": 900,
      "token_acc": 0.8707416462917685,
      "train_speed(iter/s)": 0.037262
    },
    {
      "epoch": 0.6182380216383307,
      "eval_loss": 0.43130558729171753,
      "eval_runtime": 98.6067,
      "eval_samples_per_second": 38.172,
      "eval_steps_per_second": 0.598,
      "eval_token_acc": 0.8438052359165757,
      "step": 900
    },
    {
      "epoch": 0.6319766443414048,
      "grad_norm": 0.08866509050130844,
      "learning_rate": 2.8190484434753047e-05,
      "loss": 0.3301401615142822,
      "memory(GiB)": 72.41,
      "step": 920,
      "token_acc": 0.8694062120826557,
      "train_speed(iter/s)": 0.037051
    },
    {
      "epoch": 0.6457152670444788,
      "grad_norm": 0.0904315784573555,
      "learning_rate": 2.811266506930373e-05,
      "loss": 0.3325372219085693,
      "memory(GiB)": 72.41,
      "step": 940,
      "token_acc": 0.8789977772514003,
      "train_speed(iter/s)": 0.037059
    },
    {
      "epoch": 0.6594538897475528,
      "grad_norm": 0.08624011278152466,
      "learning_rate": 2.80333195308751e-05,
      "loss": 0.33223259449005127,
      "memory(GiB)": 72.41,
      "step": 960,
      "token_acc": 0.8850906146562588,
      "train_speed(iter/s)": 0.037068
    },
    {
      "epoch": 0.6731925124506268,
      "grad_norm": 0.09679839015007019,
      "learning_rate": 2.7952457054433193e-05,
      "loss": 0.3346273183822632,
      "memory(GiB)": 72.41,
      "step": 980,
      "token_acc": 0.8790110005974765,
      "train_speed(iter/s)": 0.037086
    },
    {
      "epoch": 0.6869311351537009,
      "grad_norm": 0.08788934350013733,
      "learning_rate": 2.787008705149932e-05,
      "loss": 0.331668758392334,
      "memory(GiB)": 72.41,
      "step": 1000,
      "token_acc": 0.8727144921802968,
      "train_speed(iter/s)": 0.037101
    },
    {
      "epoch": 0.7006697578567749,
      "grad_norm": 0.07837922871112823,
      "learning_rate": 2.7786219109054618e-05,
      "loss": 0.33264620304107667,
      "memory(GiB)": 72.41,
      "step": 1020,
      "token_acc": 0.8763268597016804,
      "train_speed(iter/s)": 0.037116
    },
    {
      "epoch": 0.7144083805598489,
      "grad_norm": 0.08805106580257416,
      "learning_rate": 2.770086298842426e-05,
      "loss": 0.33027656078338624,
      "memory(GiB)": 72.41,
      "step": 1040,
      "token_acc": 0.8825189704888438,
      "train_speed(iter/s)": 0.037133
    },
    {
      "epoch": 0.7281470032629229,
      "grad_norm": 0.07906992733478546,
      "learning_rate": 2.7614028624141333e-05,
      "loss": 0.3281256914138794,
      "memory(GiB)": 72.41,
      "step": 1060,
      "token_acc": 0.8848472346715264,
      "train_speed(iter/s)": 0.037145
    },
    {
      "epoch": 0.7418856259659969,
      "grad_norm": 0.08966130018234253,
      "learning_rate": 2.7525726122790556e-05,
      "loss": 0.33036127090454104,
      "memory(GiB)": 72.41,
      "step": 1080,
      "token_acc": 0.8711485117219462,
      "train_speed(iter/s)": 0.037158
    },
    {
      "epoch": 0.755624248669071,
      "grad_norm": 0.0860779657959938,
      "learning_rate": 2.7435965761831987e-05,
      "loss": 0.32908782958984373,
      "memory(GiB)": 72.41,
      "step": 1100,
      "token_acc": 0.868969553437273,
      "train_speed(iter/s)": 0.037175
    },
    {
      "epoch": 0.769362871372145,
      "grad_norm": 0.08935214579105377,
      "learning_rate": 2.7344757988404845e-05,
      "loss": 0.33276095390319826,
      "memory(GiB)": 72.41,
      "step": 1120,
      "token_acc": 0.8682377205407336,
      "train_speed(iter/s)": 0.037189
    },
    {
      "epoch": 0.783101494075219,
      "grad_norm": 0.08778548985719681,
      "learning_rate": 2.725211341811158e-05,
      "loss": 0.33044397830963135,
      "memory(GiB)": 72.41,
      "step": 1140,
      "token_acc": 0.8793117868359244,
      "train_speed(iter/s)": 0.037201
    },
    {
      "epoch": 0.796840116778293,
      "grad_norm": 0.09090530127286911,
      "learning_rate": 2.71580428337823e-05,
      "loss": 0.32858192920684814,
      "memory(GiB)": 72.41,
      "step": 1160,
      "token_acc": 0.8733004247503157,
      "train_speed(iter/s)": 0.037216
    },
    {
      "epoch": 0.8105787394813669,
      "grad_norm": 0.08553975820541382,
      "learning_rate": 2.7062557184219806e-05,
      "loss": 0.3291203498840332,
      "memory(GiB)": 72.41,
      "step": 1180,
      "token_acc": 0.8764492301755195,
      "train_speed(iter/s)": 0.037231
    },
    {
      "epoch": 0.824317362184441,
      "grad_norm": 0.08339999616146088,
      "learning_rate": 2.6965667582925247e-05,
      "loss": 0.3333151817321777,
      "memory(GiB)": 72.41,
      "step": 1200,
      "token_acc": 0.8695492444572112,
      "train_speed(iter/s)": 0.037244
    },
    {
      "epoch": 0.824317362184441,
      "eval_loss": 0.43056952953338623,
      "eval_runtime": 99.0376,
      "eval_samples_per_second": 38.006,
      "eval_steps_per_second": 0.596,
      "eval_token_acc": 0.8441900740502716,
      "step": 1200
    },
    {
      "epoch": 0.838055984887515,
      "grad_norm": 0.08720073848962784,
      "learning_rate": 2.686738530680462e-05,
      "loss": 0.33159494400024414,
      "memory(GiB)": 72.41,
      "step": 1220,
      "token_acc": 0.8612746275278335,
      "train_speed(iter/s)": 0.037076
    },
    {
      "epoch": 0.851794607590589,
      "grad_norm": 0.08212270587682724,
      "learning_rate": 2.676772179485629e-05,
      "loss": 0.3343451976776123,
      "memory(GiB)": 72.41,
      "step": 1240,
      "token_acc": 0.874983845473253,
      "train_speed(iter/s)": 0.037077
    },
    {
      "epoch": 0.865533230293663,
      "grad_norm": 0.08813036233186722,
      "learning_rate": 2.6666688646839574e-05,
      "loss": 0.3311768531799316,
      "memory(GiB)": 72.41,
      "step": 1260,
      "token_acc": 0.8867381088510741,
      "train_speed(iter/s)": 0.037081
    },
    {
      "epoch": 0.879271852996737,
      "grad_norm": 0.08220379054546356,
      "learning_rate": 2.6564297621924696e-05,
      "loss": 0.33231358528137206,
      "memory(GiB)": 72.41,
      "step": 1280,
      "token_acc": 0.8829307086452494,
      "train_speed(iter/s)": 0.037092
    },
    {
      "epoch": 0.8930104756998111,
      "grad_norm": 0.08680058270692825,
      "learning_rate": 2.6460560637324113e-05,
      "loss": 0.3345161199569702,
      "memory(GiB)": 72.41,
      "step": 1300,
      "token_acc": 0.8796637788480545,
      "train_speed(iter/s)": 0.037101
    },
    {
      "epoch": 0.9067490984028851,
      "grad_norm": 0.09523913264274597,
      "learning_rate": 2.6355489766905496e-05,
      "loss": 0.33291900157928467,
      "memory(GiB)": 72.41,
      "step": 1320,
      "token_acc": 0.8843804465166379,
      "train_speed(iter/s)": 0.037112
    },
    {
      "epoch": 0.9204877211059591,
      "grad_norm": 0.08730066567659378,
      "learning_rate": 2.6249097239786456e-05,
      "loss": 0.33270628452301027,
      "memory(GiB)": 72.41,
      "step": 1340,
      "token_acc": 0.8798414921651669,
      "train_speed(iter/s)": 0.037122
    },
    {
      "epoch": 0.9342263438090331,
      "grad_norm": 0.0898577868938446,
      "learning_rate": 2.6141395438911216e-05,
      "loss": 0.3346142530441284,
      "memory(GiB)": 72.41,
      "step": 1360,
      "token_acc": 0.8822741759710485,
      "train_speed(iter/s)": 0.037134
    },
    {
      "epoch": 0.9479649665121072,
      "grad_norm": 0.08267758786678314,
      "learning_rate": 2.603239689960935e-05,
      "loss": 0.33362205028533937,
      "memory(GiB)": 72.41,
      "step": 1380,
      "token_acc": 0.875556680651538,
      "train_speed(iter/s)": 0.037147
    },
    {
      "epoch": 0.9617035892151812,
      "grad_norm": 0.08812654763460159,
      "learning_rate": 2.5922114308136826e-05,
      "loss": 0.3352126359939575,
      "memory(GiB)": 72.41,
      "step": 1400,
      "token_acc": 0.8624814158268795,
      "train_speed(iter/s)": 0.037157
    },
    {
      "epoch": 0.9754422119182552,
      "grad_norm": 0.08206140995025635,
      "learning_rate": 2.5810560500199454e-05,
      "loss": 0.32973828315734866,
      "memory(GiB)": 72.41,
      "step": 1420,
      "token_acc": 0.8715281106649746,
      "train_speed(iter/s)": 0.037167
    },
    {
      "epoch": 0.9891808346213292,
      "grad_norm": 0.08127112686634064,
      "learning_rate": 2.5697748459458945e-05,
      "loss": 0.33533248901367185,
      "memory(GiB)": 72.41,
      "step": 1440,
      "token_acc": 0.8709140359593514,
      "train_speed(iter/s)": 0.037179
    },
    {
      "epoch": 1.002747724540615,
      "grad_norm": 0.13326233625411987,
      "learning_rate": 2.5583691316021758e-05,
      "loss": 0.32816076278686523,
      "memory(GiB)": 72.41,
      "step": 1460,
      "token_acc": 0.8743706550630328,
      "train_speed(iter/s)": 0.037189
    },
    {
      "epoch": 1.016486347243689,
      "grad_norm": 0.09180466085672379,
      "learning_rate": 2.5468402344910895e-05,
      "loss": 0.30609779357910155,
      "memory(GiB)": 72.41,
      "step": 1480,
      "token_acc": 0.890758443681086,
      "train_speed(iter/s)": 0.037196
    },
    {
      "epoch": 1.030224969946763,
      "grad_norm": 0.09560491889715195,
      "learning_rate": 2.5351894964520832e-05,
      "loss": 0.3120020627975464,
      "memory(GiB)": 72.41,
      "step": 1500,
      "token_acc": 0.8784849657881351,
      "train_speed(iter/s)": 0.037203
    },
    {
      "epoch": 1.030224969946763,
      "eval_loss": 0.43459564447402954,
      "eval_runtime": 99.9824,
      "eval_samples_per_second": 37.647,
      "eval_steps_per_second": 0.59,
      "eval_token_acc": 0.8433591170448618,
      "step": 1500
    },
    {
      "epoch": 1.043963592649837,
      "grad_norm": 0.08771445602178574,
      "learning_rate": 2.523418273505576e-05,
      "loss": 0.31691765785217285,
      "memory(GiB)": 72.41,
      "step": 1520,
      "token_acc": 0.8715976442080544,
      "train_speed(iter/s)": 0.037064
    },
    {
      "epoch": 1.057702215352911,
      "grad_norm": 0.07920734584331512,
      "learning_rate": 2.511527935695133e-05,
      "loss": 0.31117587089538573,
      "memory(GiB)": 72.41,
      "step": 1540,
      "token_acc": 0.891813564000213,
      "train_speed(iter/s)": 0.037065
    },
    {
      "epoch": 1.071440838055985,
      "grad_norm": 0.08091707527637482,
      "learning_rate": 2.499519866928006e-05,
      "loss": 0.3078420639038086,
      "memory(GiB)": 72.41,
      "step": 1560,
      "token_acc": 0.8856428771072653,
      "train_speed(iter/s)": 0.037064
    },
    {
      "epoch": 1.0851794607590588,
      "grad_norm": 0.08118196576833725,
      "learning_rate": 2.487395464814062e-05,
      "loss": 0.30695157051086425,
      "memory(GiB)": 72.41,
      "step": 1580,
      "token_acc": 0.896800843691485,
      "train_speed(iter/s)": 0.037065
    },
    {
      "epoch": 1.0989180834621328,
      "grad_norm": 0.08192740380764008,
      "learning_rate": 2.475156140503116e-05,
      "loss": 0.30917532444000245,
      "memory(GiB)": 72.41,
      "step": 1600,
      "token_acc": 0.8825165654283423,
      "train_speed(iter/s)": 0.037072
    },
    {
      "epoch": 1.1126567061652068,
      "grad_norm": 0.08665605634450912,
      "learning_rate": 2.4628033185206914e-05,
      "loss": 0.3106253147125244,
      "memory(GiB)": 72.41,
      "step": 1620,
      "token_acc": 0.885648658540145,
      "train_speed(iter/s)": 0.037078
    },
    {
      "epoch": 1.1263953288682809,
      "grad_norm": 0.08092272281646729,
      "learning_rate": 2.4503384366022153e-05,
      "loss": 0.3136306285858154,
      "memory(GiB)": 72.41,
      "step": 1640,
      "token_acc": 0.8739959227894275,
      "train_speed(iter/s)": 0.037085
    },
    {
      "epoch": 1.1401339515713549,
      "grad_norm": 0.08443531394004822,
      "learning_rate": 2.437762945525686e-05,
      "loss": 0.317700719833374,
      "memory(GiB)": 72.41,
      "step": 1660,
      "token_acc": 0.8844603109257694,
      "train_speed(iter/s)": 0.037093
    },
    {
      "epoch": 1.153872574274429,
      "grad_norm": 0.0871417298913002,
      "learning_rate": 2.425078308942815e-05,
      "loss": 0.3168033123016357,
      "memory(GiB)": 72.41,
      "step": 1680,
      "token_acc": 0.8770761500087126,
      "train_speed(iter/s)": 0.0371
    },
    {
      "epoch": 1.167611196977503,
      "grad_norm": 0.0831030011177063,
      "learning_rate": 2.4122860032086763e-05,
      "loss": 0.31378917694091796,
      "memory(GiB)": 72.41,
      "step": 1700,
      "token_acc": 0.8724602675981399,
      "train_speed(iter/s)": 0.037109
    },
    {
      "epoch": 1.181349819680577,
      "grad_norm": 0.08853127807378769,
      "learning_rate": 2.3993875172098737e-05,
      "loss": 0.31873183250427245,
      "memory(GiB)": 72.41,
      "step": 1720,
      "token_acc": 0.887010406125234,
      "train_speed(iter/s)": 0.037115
    },
    {
      "epoch": 1.195088442383651,
      "grad_norm": 0.08741605281829834,
      "learning_rate": 2.3863843521912497e-05,
      "loss": 0.31804475784301756,
      "memory(GiB)": 72.41,
      "step": 1740,
      "token_acc": 0.8780459946059642,
      "train_speed(iter/s)": 0.037121
    },
    {
      "epoch": 1.208827065086725,
      "grad_norm": 0.0871034637093544,
      "learning_rate": 2.3732780215811563e-05,
      "loss": 0.31754317283630373,
      "memory(GiB)": 72.41,
      "step": 1760,
      "token_acc": 0.8814304365386241,
      "train_speed(iter/s)": 0.03713
    },
    {
      "epoch": 1.222565687789799,
      "grad_norm": 0.08191618323326111,
      "learning_rate": 2.3600700508153103e-05,
      "loss": 0.31642465591430663,
      "memory(GiB)": 72.41,
      "step": 1780,
      "token_acc": 0.8742795205201467,
      "train_speed(iter/s)": 0.037136
    },
    {
      "epoch": 1.236304310492873,
      "grad_norm": 0.0853760614991188,
      "learning_rate": 2.346761977159248e-05,
      "loss": 0.31632657051086427,
      "memory(GiB)": 72.41,
      "step": 1800,
      "token_acc": 0.8847685541277756,
      "train_speed(iter/s)": 0.037145
    },
    {
      "epoch": 1.236304310492873,
      "eval_loss": 0.4353775978088379,
      "eval_runtime": 98.8783,
      "eval_samples_per_second": 38.067,
      "eval_steps_per_second": 0.597,
      "eval_token_acc": 0.8432962434627534,
      "step": 1800
    },
    {
      "epoch": 1.250042933195947,
      "grad_norm": 0.08532160520553589,
      "learning_rate": 2.3333553495294033e-05,
      "loss": 0.3085492610931396,
      "memory(GiB)": 72.41,
      "step": 1820,
      "token_acc": 0.8650615076497046,
      "train_speed(iter/s)": 0.037034
    },
    {
      "epoch": 1.263781555899021,
      "grad_norm": 0.09372863918542862,
      "learning_rate": 2.3198517283128316e-05,
      "loss": 0.314247727394104,
      "memory(GiB)": 72.41,
      "step": 1840,
      "token_acc": 0.8870913422078638,
      "train_speed(iter/s)": 0.037033
    },
    {
      "epoch": 1.277520178602095,
      "grad_norm": 0.0852976143360138,
      "learning_rate": 2.3062526851855962e-05,
      "loss": 0.31310009956359863,
      "memory(GiB)": 72.41,
      "step": 1860,
      "token_acc": 0.8906191502723332,
      "train_speed(iter/s)": 0.037035
    },
    {
      "epoch": 1.2912588013051691,
      "grad_norm": 0.07643554359674454,
      "learning_rate": 2.2925598029298437e-05,
      "loss": 0.3103055715560913,
      "memory(GiB)": 72.41,
      "step": 1880,
      "token_acc": 0.887677412229967,
      "train_speed(iter/s)": 0.037042
    },
    {
      "epoch": 1.3049974240082431,
      "grad_norm": 0.08572836965322495,
      "learning_rate": 2.278774675249585e-05,
      "loss": 0.31417174339294435,
      "memory(GiB)": 72.41,
      "step": 1900,
      "token_acc": 0.8894350828946791,
      "train_speed(iter/s)": 0.037046
    },
    {
      "epoch": 1.3187360467113172,
      "grad_norm": 0.08305912464857101,
      "learning_rate": 2.264898906585204e-05,
      "loss": 0.31093263626098633,
      "memory(GiB)": 72.41,
      "step": 1920,
      "token_acc": 0.8820933517164594,
      "train_speed(iter/s)": 0.037052
    },
    {
      "epoch": 1.3324746694143912,
      "grad_norm": 0.08301204442977905,
      "learning_rate": 2.2509341119267193e-05,
      "loss": 0.3095247268676758,
      "memory(GiB)": 72.41,
      "step": 1940,
      "token_acc": 0.8671473791714781,
      "train_speed(iter/s)": 0.037058
    },
    {
      "epoch": 1.3462132921174652,
      "grad_norm": 0.0815897062420845,
      "learning_rate": 2.236881916625816e-05,
      "loss": 0.3098980188369751,
      "memory(GiB)": 72.41,
      "step": 1960,
      "token_acc": 0.8935473891956769,
      "train_speed(iter/s)": 0.037064
    },
    {
      "epoch": 1.3599519148205392,
      "grad_norm": 0.08821182698011398,
      "learning_rate": 2.2227439562066734e-05,
      "loss": 0.30906736850738525,
      "memory(GiB)": 72.41,
      "step": 1980,
      "token_acc": 0.8789100589878682,
      "train_speed(iter/s)": 0.037069
    },
    {
      "epoch": 1.3736905375236133,
      "grad_norm": 0.0886450707912445,
      "learning_rate": 2.2085218761756058e-05,
      "loss": 0.3117701768875122,
      "memory(GiB)": 72.41,
      "step": 2000,
      "token_acc": 0.8915136412607484,
      "train_speed(iter/s)": 0.037071
    },
    {
      "epoch": 1.3874291602266873,
      "grad_norm": 0.08608590811491013,
      "learning_rate": 2.1942173318295443e-05,
      "loss": 0.3138264179229736,
      "memory(GiB)": 72.41,
      "step": 2020,
      "token_acc": 0.8859320703790349,
      "train_speed(iter/s)": 0.037077
    },
    {
      "epoch": 1.4011677829297613,
      "grad_norm": 0.07869933545589447,
      "learning_rate": 2.1798319880633795e-05,
      "loss": 0.3135652542114258,
      "memory(GiB)": 72.41,
      "step": 2040,
      "token_acc": 0.8910194771797223,
      "train_speed(iter/s)": 0.037082
    },
    {
      "epoch": 1.4149064056328353,
      "grad_norm": 0.07948032766580582,
      "learning_rate": 2.165367519176183e-05,
      "loss": 0.3114771842956543,
      "memory(GiB)": 72.41,
      "step": 2060,
      "token_acc": 0.888334672346102,
      "train_speed(iter/s)": 0.037086
    },
    {
      "epoch": 1.4286450283359093,
      "grad_norm": 0.08114151656627655,
      "learning_rate": 2.1508256086763372e-05,
      "loss": 0.3094203948974609,
      "memory(GiB)": 72.41,
      "step": 2080,
      "token_acc": 0.8877183536236418,
      "train_speed(iter/s)": 0.03709
    },
    {
      "epoch": 1.4423836510389834,
      "grad_norm": 0.0871092826128006,
      "learning_rate": 2.1362079490855968e-05,
      "loss": 0.3111464738845825,
      "memory(GiB)": 72.41,
      "step": 2100,
      "token_acc": 0.881144622390869,
      "train_speed(iter/s)": 0.037093
    },
    {
      "epoch": 1.4423836510389834,
      "eval_loss": 0.4344118535518646,
      "eval_runtime": 99.5043,
      "eval_samples_per_second": 37.828,
      "eval_steps_per_second": 0.593,
      "eval_token_acc": 0.8434381641208087,
      "step": 2100
    },
    {
      "epoch": 1.4561222737420574,
      "grad_norm": 0.08234430849552155,
      "learning_rate": 2.1215162417420926e-05,
      "loss": 0.3089058637619019,
      "memory(GiB)": 72.41,
      "step": 2120,
      "token_acc": 0.8675401686436827,
      "train_speed(iter/s)": 0.036998
    },
    {
      "epoch": 1.4698608964451314,
      "grad_norm": 0.08042703568935394,
      "learning_rate": 2.1067521966023165e-05,
      "loss": 0.31057741641998293,
      "memory(GiB)": 72.41,
      "step": 2140,
      "token_acc": 0.8962609916378795,
      "train_speed(iter/s)": 0.036997
    },
    {
      "epoch": 1.4835995191482054,
      "grad_norm": 0.08385493606328964,
      "learning_rate": 2.0919175320421023e-05,
      "loss": 0.3134245634078979,
      "memory(GiB)": 72.41,
      "step": 2160,
      "token_acc": 0.8868531518893562,
      "train_speed(iter/s)": 0.036998
    },
    {
      "epoch": 1.4973381418512794,
      "grad_norm": 0.09416038542985916,
      "learning_rate": 2.0770139746566223e-05,
      "loss": 0.31356468200683596,
      "memory(GiB)": 72.41,
      "step": 2180,
      "token_acc": 0.8753653697079176,
      "train_speed(iter/s)": 0.037001
    },
    {
      "epoch": 1.5110767645543535,
      "grad_norm": 0.08807655423879623,
      "learning_rate": 2.062043259059432e-05,
      "loss": 0.31597309112548827,
      "memory(GiB)": 72.41,
      "step": 2200,
      "token_acc": 0.8919064810265528,
      "train_speed(iter/s)": 0.037007
    },
    {
      "epoch": 1.5248153872574275,
      "grad_norm": 0.08815551549196243,
      "learning_rate": 2.047007127680579e-05,
      "loss": 0.3196309804916382,
      "memory(GiB)": 72.41,
      "step": 2220,
      "token_acc": 0.8772186268233043,
      "train_speed(iter/s)": 0.037012
    },
    {
      "epoch": 1.5385540099605015,
      "grad_norm": 0.08227042853832245,
      "learning_rate": 2.0319073305638035e-05,
      "loss": 0.31729488372802733,
      "memory(GiB)": 72.41,
      "step": 2240,
      "token_acc": 0.8858860714860183,
      "train_speed(iter/s)": 0.037014
    },
    {
      "epoch": 1.5522926326635753,
      "grad_norm": 0.08253244310617447,
      "learning_rate": 2.0167456251628524e-05,
      "loss": 0.31553847789764405,
      "memory(GiB)": 72.41,
      "step": 2260,
      "token_acc": 0.8908496364853541,
      "train_speed(iter/s)": 0.037016
    },
    {
      "epoch": 1.5660312553666493,
      "grad_norm": 0.08127789944410324,
      "learning_rate": 2.00152377613693e-05,
      "loss": 0.3174169063568115,
      "memory(GiB)": 72.41,
      "step": 2280,
      "token_acc": 0.8759036896828214,
      "train_speed(iter/s)": 0.037021
    },
    {
      "epoch": 1.5797698780697234,
      "grad_norm": 0.08351726084947586,
      "learning_rate": 1.9862435551453103e-05,
      "loss": 0.31812009811401365,
      "memory(GiB)": 72.41,
      "step": 2300,
      "token_acc": 0.8801280981073656,
      "train_speed(iter/s)": 0.037028
    },
    {
      "epoch": 1.5935085007727974,
      "grad_norm": 0.08042768388986588,
      "learning_rate": 1.9709067406411352e-05,
      "loss": 0.3188045024871826,
      "memory(GiB)": 72.41,
      "step": 2320,
      "token_acc": 0.8883485418399553,
      "train_speed(iter/s)": 0.037034
    },
    {
      "epoch": 1.6072471234758714,
      "grad_norm": 0.0847587063908577,
      "learning_rate": 1.9555151176644223e-05,
      "loss": 0.31552605628967284,
      "memory(GiB)": 72.41,
      "step": 2340,
      "token_acc": 0.8933710959011879,
      "train_speed(iter/s)": 0.03704
    },
    {
      "epoch": 1.6209857461789454,
      "grad_norm": 0.0849500447511673,
      "learning_rate": 1.9400704776343047e-05,
      "loss": 0.3190001010894775,
      "memory(GiB)": 72.41,
      "step": 2360,
      "token_acc": 0.8655127619672538,
      "train_speed(iter/s)": 0.037046
    },
    {
      "epoch": 1.6347243688820194,
      "grad_norm": 0.0820319652557373,
      "learning_rate": 1.9245746181405306e-05,
      "loss": 0.3157363414764404,
      "memory(GiB)": 72.41,
      "step": 2380,
      "token_acc": 0.8931401676158139,
      "train_speed(iter/s)": 0.037052
    },
    {
      "epoch": 1.6484629915850935,
      "grad_norm": 0.07777854800224304,
      "learning_rate": 1.9090293427342406e-05,
      "loss": 0.30912251472473146,
      "memory(GiB)": 72.41,
      "step": 2400,
      "token_acc": 0.8933260366449716,
      "train_speed(iter/s)": 0.037059
    },
    {
      "epoch": 1.6484629915850935,
      "eval_loss": 0.43367844820022583,
      "eval_runtime": 99.7879,
      "eval_samples_per_second": 37.72,
      "eval_steps_per_second": 0.591,
      "eval_token_acc": 0.843613657031226,
      "step": 2400
    },
    {
      "epoch": 1.6622016142881675,
      "grad_norm": 0.08155303448438644,
      "learning_rate": 1.893436460718056e-05,
      "loss": 0.3163402795791626,
      "memory(GiB)": 72.41,
      "step": 2420,
      "token_acc": 0.8672117073299662,
      "train_speed(iter/s)": 0.036975
    },
    {
      "epoch": 1.6759402369912415,
      "grad_norm": 0.08382421731948853,
      "learning_rate": 1.877797786935495e-05,
      "loss": 0.3165715217590332,
      "memory(GiB)": 72.41,
      "step": 2440,
      "token_acc": 0.8854491510650321,
      "train_speed(iter/s)": 0.03697
    },
    {
      "epoch": 1.6896788596943155,
      "grad_norm": 0.08178658783435822,
      "learning_rate": 1.862115141559744e-05,
      "loss": 0.3171123504638672,
      "memory(GiB)": 72.41,
      "step": 2460,
      "token_acc": 0.8836779780841286,
      "train_speed(iter/s)": 0.036973
    },
    {
      "epoch": 1.7034174823973895,
      "grad_norm": 0.07710675150156021,
      "learning_rate": 1.8463903498818088e-05,
      "loss": 0.31471326351165774,
      "memory(GiB)": 72.41,
      "step": 2480,
      "token_acc": 0.8908230830682876,
      "train_speed(iter/s)": 0.036976
    },
    {
      "epoch": 1.7171561051004636,
      "grad_norm": 0.08034602552652359,
      "learning_rate": 1.8306252420980704e-05,
      "loss": 0.31853632926940917,
      "memory(GiB)": 72.41,
      "step": 2500,
      "token_acc": 0.8883582169845952,
      "train_speed(iter/s)": 0.036978
    },
    {
      "epoch": 1.7308947278035376,
      "grad_norm": 0.07948100566864014,
      "learning_rate": 1.8148216530972714e-05,
      "loss": 0.3109827995300293,
      "memory(GiB)": 72.41,
      "step": 2520,
      "token_acc": 0.8949681174869483,
      "train_speed(iter/s)": 0.036981
    },
    {
      "epoch": 1.7446333505066116,
      "grad_norm": 0.07899336516857147,
      "learning_rate": 1.7989814222469538e-05,
      "loss": 0.3090771436691284,
      "memory(GiB)": 72.41,
      "step": 2540,
      "token_acc": 0.8906750005261931,
      "train_speed(iter/s)": 0.036983
    },
    {
      "epoch": 1.7583719732096856,
      "grad_norm": 0.07443471997976303,
      "learning_rate": 1.783106393179375e-05,
      "loss": 0.31173481941223147,
      "memory(GiB)": 72.41,
      "step": 2560,
      "token_acc": 0.8891939493597887,
      "train_speed(iter/s)": 0.036988
    },
    {
      "epoch": 1.7721105959127597,
      "grad_norm": 0.07555528730154037,
      "learning_rate": 1.767198413576931e-05,
      "loss": 0.30927410125732424,
      "memory(GiB)": 72.41,
      "step": 2580,
      "token_acc": 0.8850328545945815,
      "train_speed(iter/s)": 0.036994
    },
    {
      "epoch": 1.7858492186158337,
      "grad_norm": 0.08017897605895996,
      "learning_rate": 1.7512593349571046e-05,
      "loss": 0.31209754943847656,
      "memory(GiB)": 72.41,
      "step": 2600,
      "token_acc": 0.8816632260591382,
      "train_speed(iter/s)": 0.036998
    },
    {
      "epoch": 1.7995878413189077,
      "grad_norm": 0.07677578181028366,
      "learning_rate": 1.7352910124569695e-05,
      "loss": 0.30882983207702636,
      "memory(GiB)": 72.41,
      "step": 2620,
      "token_acc": 0.8925267013383078,
      "train_speed(iter/s)": 0.037
    },
    {
      "epoch": 1.8133264640219817,
      "grad_norm": 0.07692938297986984,
      "learning_rate": 1.7192953046172726e-05,
      "loss": 0.3074300289154053,
      "memory(GiB)": 72.41,
      "step": 2640,
      "token_acc": 0.8861350676140611,
      "train_speed(iter/s)": 0.037005
    },
    {
      "epoch": 1.8270650867250557,
      "grad_norm": 0.07619909197092056,
      "learning_rate": 1.7032740731661178e-05,
      "loss": 0.30927472114562987,
      "memory(GiB)": 72.41,
      "step": 2660,
      "token_acc": 0.8921508449028042,
      "train_speed(iter/s)": 0.037009
    },
    {
      "epoch": 1.8408037094281298,
      "grad_norm": 0.08186180889606476,
      "learning_rate": 1.687229182802284e-05,
      "loss": 0.3076324939727783,
      "memory(GiB)": 72.41,
      "step": 2680,
      "token_acc": 0.874112111934862,
      "train_speed(iter/s)": 0.037013
    },
    {
      "epoch": 1.8545423321312038,
      "grad_norm": 0.0749615728855133,
      "learning_rate": 1.6711625009781926e-05,
      "loss": 0.3025542736053467,
      "memory(GiB)": 72.41,
      "step": 2700,
      "token_acc": 0.9005740784776456,
      "train_speed(iter/s)": 0.037016
    },
    {
      "epoch": 1.8545423321312038,
      "eval_loss": 0.43260329961776733,
      "eval_runtime": 99.7086,
      "eval_samples_per_second": 37.75,
      "eval_steps_per_second": 0.592,
      "eval_token_acc": 0.8437799954640701,
      "step": 2700
    },
    {
      "epoch": 1.8682809548342778,
      "grad_norm": 0.07678617537021637,
      "learning_rate": 1.655075897682555e-05,
      "loss": 0.3069960117340088,
      "memory(GiB)": 72.41,
      "step": 2720,
      "token_acc": 0.8656336346071796,
      "train_speed(iter/s)": 0.036945
    },
    {
      "epoch": 1.8820195775373518,
      "grad_norm": 0.08224895596504211,
      "learning_rate": 1.6389712452227295e-05,
      "loss": 0.31150364875793457,
      "memory(GiB)": 72.41,
      "step": 2740,
      "token_acc": 0.8871026948734946,
      "train_speed(iter/s)": 0.036944
    },
    {
      "epoch": 1.8957582002404259,
      "grad_norm": 0.07674538344144821,
      "learning_rate": 1.6228504180068003e-05,
      "loss": 0.31361680030822753,
      "memory(GiB)": 72.41,
      "step": 2760,
      "token_acc": 0.8885069679173144,
      "train_speed(iter/s)": 0.036944
    },
    {
      "epoch": 1.9094968229434999,
      "grad_norm": 0.07724355906248093,
      "learning_rate": 1.60671529232542e-05,
      "loss": 0.31092076301574706,
      "memory(GiB)": 72.41,
      "step": 2780,
      "token_acc": 0.8775532573683428,
      "train_speed(iter/s)": 0.036948
    },
    {
      "epoch": 1.923235445646574,
      "grad_norm": 0.0748470202088356,
      "learning_rate": 1.5905677461334292e-05,
      "loss": 0.3125690698623657,
      "memory(GiB)": 72.41,
      "step": 2800,
      "token_acc": 0.8846377126342211,
      "train_speed(iter/s)": 0.036916
    },
    {
      "epoch": 1.936974068349648,
      "grad_norm": 0.08404634892940521,
      "learning_rate": 1.574409658831281e-05,
      "loss": 0.3153404235839844,
      "memory(GiB)": 72.41,
      "step": 2820,
      "token_acc": 0.8762131944710342,
      "train_speed(iter/s)": 0.036918
    },
    {
      "epoch": 1.950712691052722,
      "grad_norm": 0.07959295809268951,
      "learning_rate": 1.558242911046302e-05,
      "loss": 0.31249830722808836,
      "memory(GiB)": 72.41,
      "step": 2840,
      "token_acc": 0.8892035392544179,
      "train_speed(iter/s)": 0.036921
    },
    {
      "epoch": 1.964451313755796,
      "grad_norm": 0.08044803887605667,
      "learning_rate": 1.5420693844138036e-05,
      "loss": 0.3130341053009033,
      "memory(GiB)": 72.41,
      "step": 2860,
      "token_acc": 0.8932265094341124,
      "train_speed(iter/s)": 0.036924
    },
    {
      "epoch": 1.97818993645887,
      "grad_norm": 0.07583785802125931,
      "learning_rate": 1.525890961358083e-05,
      "loss": 0.3141756772994995,
      "memory(GiB)": 72.41,
      "step": 2880,
      "token_acc": 0.8839660044002189,
      "train_speed(iter/s)": 0.036928
    },
    {
      "epoch": 1.991928559161944,
      "grad_norm": 0.07464556396007538,
      "learning_rate": 1.5097095248733284e-05,
      "loss": 0.31082568168640134,
      "memory(GiB)": 72.41,
      "step": 2900,
      "token_acc": 0.8775183645838733,
      "train_speed(iter/s)": 0.036932
    },
    {
      "epoch": 2.00549544908123,
      "grad_norm": 0.12503303587436676,
      "learning_rate": 1.4935269583044581e-05,
      "loss": 0.2993995904922485,
      "memory(GiB)": 72.41,
      "step": 2920,
      "token_acc": 0.8823204490957476,
      "train_speed(iter/s)": 0.036937
    },
    {
      "epoch": 2.019234071784304,
      "grad_norm": 0.08332613110542297,
      "learning_rate": 1.4773451451279213e-05,
      "loss": 0.29198360443115234,
      "memory(GiB)": 72.41,
      "step": 2940,
      "token_acc": 0.8980348203187491,
      "train_speed(iter/s)": 0.036941
    },
    {
      "epoch": 2.032972694487378,
      "grad_norm": 0.08240070939064026,
      "learning_rate": 1.461165968732479e-05,
      "loss": 0.2935274362564087,
      "memory(GiB)": 72.41,
      "step": 2960,
      "token_acc": 0.8904662128095143,
      "train_speed(iter/s)": 0.036946
    },
    {
      "epoch": 2.046711317190452,
      "grad_norm": 0.08113058656454086,
      "learning_rate": 1.4449913122000005e-05,
      "loss": 0.29198508262634276,
      "memory(GiB)": 72.41,
      "step": 2980,
      "token_acc": 0.8908448858293387,
      "train_speed(iter/s)": 0.036953
    },
    {
      "epoch": 2.060449939893526,
      "grad_norm": 0.08105536550283432,
      "learning_rate": 1.4288230580862905e-05,
      "loss": 0.290987491607666,
      "memory(GiB)": 72.41,
      "step": 3000,
      "token_acc": 0.8764201959142056,
      "train_speed(iter/s)": 0.036958
    },
    {
      "epoch": 2.060449939893526,
      "eval_loss": 0.4393101930618286,
      "eval_runtime": 100.5668,
      "eval_samples_per_second": 37.428,
      "eval_steps_per_second": 0.587,
      "eval_token_acc": 0.84265782805066,
      "step": 3000
    },
    {
      "epoch": 2.0741885625966,
      "grad_norm": 0.08123844116926193,
      "learning_rate": 1.412663088201982e-05,
      "loss": 0.29090156555175783,
      "memory(GiB)": 72.41,
      "step": 3020,
      "token_acc": 0.8735638063478551,
      "train_speed(iter/s)": 0.036892
    },
    {
      "epoch": 2.087927185299674,
      "grad_norm": 0.0814339891076088,
      "learning_rate": 1.3965132833935126e-05,
      "loss": 0.2902204990386963,
      "memory(GiB)": 72.41,
      "step": 3040,
      "token_acc": 0.8839492383548759,
      "train_speed(iter/s)": 0.036893
    },
    {
      "epoch": 2.101665808002748,
      "grad_norm": 0.0760771632194519,
      "learning_rate": 1.380375523324215e-05,
      "loss": 0.29666552543640134,
      "memory(GiB)": 72.41,
      "step": 3060,
      "token_acc": 0.8921821581883145,
      "train_speed(iter/s)": 0.036894
    },
    {
      "epoch": 2.115404430705822,
      "grad_norm": 0.0808984562754631,
      "learning_rate": 1.3642516862555433e-05,
      "loss": 0.28961887359619143,
      "memory(GiB)": 72.41,
      "step": 3080,
      "token_acc": 0.8996385382943057,
      "train_speed(iter/s)": 0.036898
    },
    {
      "epoch": 2.129143053408896,
      "grad_norm": 0.08168598264455795,
      "learning_rate": 1.3481436488284648e-05,
      "loss": 0.2952747821807861,
      "memory(GiB)": 72.41,
      "step": 3100,
      "token_acc": 0.8932001882680204,
      "train_speed(iter/s)": 0.036901
    },
    {
      "epoch": 2.14288167611197,
      "grad_norm": 0.08384265005588531,
      "learning_rate": 1.3320532858450382e-05,
      "loss": 0.29767014980316164,
      "memory(GiB)": 72.41,
      "step": 3120,
      "token_acc": 0.8847414688023099,
      "train_speed(iter/s)": 0.036904
    },
    {
      "epoch": 2.156620298815044,
      "grad_norm": 0.08613137155771255,
      "learning_rate": 1.3159824700502083e-05,
      "loss": 0.2987870693206787,
      "memory(GiB)": 72.41,
      "step": 3140,
      "token_acc": 0.8912630847005318,
      "train_speed(iter/s)": 0.036909
    },
    {
      "epoch": 2.1703589215181176,
      "grad_norm": 0.08044470101594925,
      "learning_rate": 1.2999330719138363e-05,
      "loss": 0.29793477058410645,
      "memory(GiB)": 72.41,
      "step": 3160,
      "token_acc": 0.8911335210006078,
      "train_speed(iter/s)": 0.036914
    },
    {
      "epoch": 2.1840975442211916,
      "grad_norm": 0.0807594358921051,
      "learning_rate": 1.283906959413e-05,
      "loss": 0.2947986125946045,
      "memory(GiB)": 72.41,
      "step": 3180,
      "token_acc": 0.8855782459322568,
      "train_speed(iter/s)": 0.036917
    },
    {
      "epoch": 2.1978361669242656,
      "grad_norm": 0.08066173642873764,
      "learning_rate": 1.267905997814578e-05,
      "loss": 0.2977961778640747,
      "memory(GiB)": 72.41,
      "step": 3200,
      "token_acc": 0.8803478438446615,
      "train_speed(iter/s)": 0.036922
    },
    {
      "epoch": 2.2115747896273397,
      "grad_norm": 0.08045271784067154,
      "learning_rate": 1.2519320494581581e-05,
      "loss": 0.29424285888671875,
      "memory(GiB)": 72.41,
      "step": 3220,
      "token_acc": 0.8828892872837293,
      "train_speed(iter/s)": 0.036927
    },
    {
      "epoch": 2.2253134123304137,
      "grad_norm": 0.08237405866384506,
      "learning_rate": 1.2359869735392746e-05,
      "loss": 0.29676170349121095,
      "memory(GiB)": 72.41,
      "step": 3240,
      "token_acc": 0.8954581030873394,
      "train_speed(iter/s)": 0.036931
    },
    {
      "epoch": 2.2390520350334877,
      "grad_norm": 0.08597618341445923,
      "learning_rate": 1.220072625893023e-05,
      "loss": 0.296732759475708,
      "memory(GiB)": 72.41,
      "step": 3260,
      "token_acc": 0.8882645330425789,
      "train_speed(iter/s)": 0.036937
    },
    {
      "epoch": 2.2527906577365617,
      "grad_norm": 0.07742371410131454,
      "learning_rate": 1.2041908587780571e-05,
      "loss": 0.293271803855896,
      "memory(GiB)": 72.41,
      "step": 3280,
      "token_acc": 0.8859188183637006,
      "train_speed(iter/s)": 0.036942
    },
    {
      "epoch": 2.2665292804396358,
      "grad_norm": 0.07885393500328064,
      "learning_rate": 1.1883435206610095e-05,
      "loss": 0.29781594276428225,
      "memory(GiB)": 72.41,
      "step": 3300,
      "token_acc": 0.8873998820001778,
      "train_speed(iter/s)": 0.036945
    },
    {
      "epoch": 2.2665292804396358,
      "eval_loss": 0.4392697215080261,
      "eval_runtime": 99.0852,
      "eval_samples_per_second": 37.988,
      "eval_steps_per_second": 0.595,
      "eval_token_acc": 0.8427081339178593,
      "step": 3300
    }
  ],
  "logging_steps": 20,
  "max_steps": 5824,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 300,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.596186893058048e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}