{
  "best_global_step": 460,
  "best_metric": 0.24467714,
  "best_model_checkpoint": "/data/home/scyb089/CODE/scripts/ms-swift/3b-new/v29-20250507-134003/checkpoint-460",
  "epoch": 2.9911123081066524,
  "eval_steps": 20,
  "global_step": 696,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0043091839482897925,
      "grad_norm": 2.422614812850952,
      "learning_rate": 9.99994906450425e-06,
      "loss": 0.3783050775527954,
      "memory(GiB)": 29.06,
      "step": 1,
      "token_acc": 0.8827366746221161,
      "train_speed(iter/s)": 0.066016
    },
    {
      "epoch": 0.02154591974144896,
      "grad_norm": 1.8293063640594482,
      "learning_rate": 9.99872666449397e-06,
      "loss": 0.304108202457428,
      "memory(GiB)": 29.06,
      "step": 5,
      "token_acc": 0.9021085311428756,
      "train_speed(iter/s)": 0.123613
    },
    {
      "epoch": 0.04309183948289792,
      "grad_norm": 1.0181312561035156,
      "learning_rate": 9.994907306529203e-06,
      "loss": 0.30131869316101073,
      "memory(GiB)": 29.06,
      "step": 10,
      "token_acc": 0.9080832657474993,
      "train_speed(iter/s)": 0.141649
    },
    {
      "epoch": 0.06463775922434689,
      "grad_norm": 0.8247085213661194,
      "learning_rate": 9.988543871435342e-06,
      "loss": 0.29039506912231444,
      "memory(GiB)": 29.07,
      "step": 15,
      "token_acc": 0.9026237111961776,
      "train_speed(iter/s)": 0.145623
    },
    {
      "epoch": 0.08618367896579585,
      "grad_norm": 0.8313817977905273,
      "learning_rate": 9.979639600327522e-06,
      "loss": 0.2841599941253662,
      "memory(GiB)": 29.07,
      "step": 20,
      "token_acc": 0.9051915284043123,
      "train_speed(iter/s)": 0.148055
    },
    {
      "epoch": 0.08618367896579585,
      "eval_loss": 0.30637428164482117,
      "eval_runtime": 9.2534,
      "eval_samples_per_second": 16.21,
      "eval_steps_per_second": 4.107,
      "eval_token_acc": 0.8990793096302436,
      "step": 20
    },
    {
      "epoch": 0.10772959870724481,
      "grad_norm": 0.8578795790672302,
      "learning_rate": 9.96819902845557e-06,
      "loss": 0.26584060192108155,
      "memory(GiB)": 29.07,
      "step": 25,
      "token_acc": 0.9061398699976505,
      "train_speed(iter/s)": 0.133467
    },
    {
      "epoch": 0.12927551844869378,
      "grad_norm": 0.8536065220832825,
      "learning_rate": 9.954227982894034e-06,
      "loss": 0.2721074342727661,
      "memory(GiB)": 29.07,
      "step": 30,
      "token_acc": 0.9100939031401429,
      "train_speed(iter/s)": 0.137239
    },
    {
      "epoch": 0.15082143819014274,
      "grad_norm": 0.8567067384719849,
      "learning_rate": 9.937733579574263e-06,
      "loss": 0.26833133697509765,
      "memory(GiB)": 29.07,
      "step": 35,
      "token_acc": 0.9014427903508017,
      "train_speed(iter/s)": 0.139069
    },
    {
      "epoch": 0.1723673579315917,
      "grad_norm": 0.8759805560112,
      "learning_rate": 9.918724219660013e-06,
      "loss": 0.2696810483932495,
      "memory(GiB)": 29.07,
      "step": 40,
      "token_acc": 0.9073056300268096,
      "train_speed(iter/s)": 0.141475
    },
    {
      "epoch": 0.1723673579315917,
      "eval_loss": 0.28540194034576416,
      "eval_runtime": 9.1496,
      "eval_samples_per_second": 16.394,
      "eval_steps_per_second": 4.153,
      "eval_token_acc": 0.9062731282749638,
      "step": 40
    },
    {
      "epoch": 0.19391327767304067,
      "grad_norm": 0.8179587125778198,
      "learning_rate": 9.897209585268459e-06,
      "loss": 0.26938886642456056,
      "memory(GiB)": 29.07,
      "step": 45,
      "token_acc": 0.9068100358422939,
      "train_speed(iter/s)": 0.134153
    },
    {
      "epoch": 0.21545919741448963,
      "grad_norm": 0.7597087621688843,
      "learning_rate": 9.873200634538746e-06,
      "loss": 0.2661460876464844,
      "memory(GiB)": 29.07,
      "step": 50,
      "token_acc": 0.9265829903627394,
      "train_speed(iter/s)": 0.135778
    },
    {
      "epoch": 0.23700511715593858,
      "grad_norm": 0.9001930952072144,
      "learning_rate": 9.846709596050646e-06,
      "loss": 0.2637378692626953,
      "memory(GiB)": 29.07,
      "step": 55,
      "token_acc": 0.9097354466352211,
      "train_speed(iter/s)": 0.137425
    },
    {
      "epoch": 0.25855103689738757,
      "grad_norm": 0.894802451133728,
      "learning_rate": 9.817749962596115e-06,
      "loss": 0.26340594291687014,
      "memory(GiB)": 29.07,
      "step": 60,
      "token_acc": 0.9115004961612785,
      "train_speed(iter/s)": 0.139172
    },
    {
      "epoch": 0.25855103689738757,
      "eval_loss": 0.2748047709465027,
      "eval_runtime": 9.1341,
      "eval_samples_per_second": 16.422,
      "eval_steps_per_second": 4.16,
      "eval_token_acc": 0.9068652121140354,
      "step": 60
    },
    {
      "epoch": 0.28009695663883655,
      "grad_norm": 0.7363941669464111,
      "learning_rate": 9.786336484306966e-06,
      "loss": 0.27098889350891114,
      "memory(GiB)": 29.07,
      "step": 65,
      "token_acc": 0.9083590733590734,
      "train_speed(iter/s)": 0.135246
    },
    {
      "epoch": 0.3016428763802855,
      "grad_norm": 0.8381310105323792,
      "learning_rate": 9.752485161142103e-06,
      "loss": 0.2478638172149658,
      "memory(GiB)": 29.07,
      "step": 70,
      "token_acc": 0.9232435033686237,
      "train_speed(iter/s)": 0.136407
    },
    {
      "epoch": 0.32318879612173446,
      "grad_norm": 0.7714306116104126,
      "learning_rate": 9.716213234738216e-06,
      "loss": 0.2461942672729492,
      "memory(GiB)": 29.07,
      "step": 75,
      "token_acc": 0.9149177216982468,
      "train_speed(iter/s)": 0.137517
    },
    {
      "epoch": 0.3447347158631834,
      "grad_norm": 0.9231055974960327,
      "learning_rate": 9.677539179628005e-06,
      "loss": 0.24781365394592286,
      "memory(GiB)": 29.07,
      "step": 80,
      "token_acc": 0.9261125903385318,
      "train_speed(iter/s)": 0.138742
    },
    {
      "epoch": 0.3447347158631834,
      "eval_loss": 0.26845118403434753,
      "eval_runtime": 9.1426,
      "eval_samples_per_second": 16.407,
      "eval_steps_per_second": 4.156,
      "eval_token_acc": 0.9084194321915984,
      "step": 80
    },
    {
      "epoch": 0.36628063560463237,
      "grad_norm": 0.7188398241996765,
      "learning_rate": 9.636482693830488e-06,
      "loss": 0.26428771018981934,
      "memory(GiB)": 29.07,
      "step": 85,
      "token_acc": 0.9042584492261823,
      "train_speed(iter/s)": 0.135179
    },
    {
      "epoch": 0.38782655534608135,
      "grad_norm": 0.8020254969596863,
      "learning_rate": 9.59306468881811e-06,
      "loss": 0.2622120141983032,
      "memory(GiB)": 29.07,
      "step": 90,
      "token_acc": 0.9098044980155814,
      "train_speed(iter/s)": 0.13707
    },
    {
      "epoch": 0.4093724750875303,
      "grad_norm": 0.8326025605201721,
      "learning_rate": 9.547307278865823e-06,
      "loss": 0.2394162893295288,
      "memory(GiB)": 29.07,
      "step": 95,
      "token_acc": 0.9164833305127771,
      "train_speed(iter/s)": 0.138039
    },
    {
      "epoch": 0.43091839482897926,
      "grad_norm": 0.7730870842933655,
      "learning_rate": 9.499233769787534e-06,
      "loss": 0.24491536617279053,
      "memory(GiB)": 29.07,
      "step": 100,
      "token_acc": 0.9174321989744152,
      "train_speed(iter/s)": 0.139304
    },
    {
      "epoch": 0.43091839482897926,
      "eval_loss": 0.26416531205177307,
      "eval_runtime": 9.1324,
      "eval_samples_per_second": 16.425,
      "eval_steps_per_second": 4.161,
      "eval_token_acc": 0.9095295893898576,
      "step": 100
    },
    {
      "epoch": 0.45246431457042824,
      "grad_norm": 0.860744833946228,
      "learning_rate": 9.448868647065644e-06,
      "loss": 0.25905332565307615,
      "memory(GiB)": 29.07,
      "step": 105,
      "token_acc": 0.9042013222435488,
      "train_speed(iter/s)": 0.136744
    },
    {
      "epoch": 0.47401023431187717,
      "grad_norm": 0.7342467904090881,
      "learning_rate": 9.396237563379761e-06,
      "loss": 0.23780291080474852,
      "memory(GiB)": 29.07,
      "step": 110,
      "token_acc": 0.9195750302763405,
      "train_speed(iter/s)": 0.137488
    },
    {
      "epoch": 0.49555615405332615,
      "grad_norm": 0.8625522255897522,
      "learning_rate": 9.341367325540921e-06,
      "loss": 0.23777966499328612,
      "memory(GiB)": 29.07,
      "step": 115,
      "token_acc": 0.9236284378674467,
      "train_speed(iter/s)": 0.138134
    },
    {
      "epoch": 0.5171020737947751,
      "grad_norm": 0.7858663201332092,
      "learning_rate": 9.284285880837947e-06,
      "loss": 0.23805389404296876,
      "memory(GiB)": 29.07,
      "step": 120,
      "token_acc": 0.9197558545092177,
      "train_speed(iter/s)": 0.138839
    },
    {
      "epoch": 0.5171020737947751,
      "eval_loss": 0.26037347316741943,
      "eval_runtime": 9.1916,
      "eval_samples_per_second": 16.319,
      "eval_steps_per_second": 4.134,
      "eval_token_acc": 0.9109209864116758,
      "step": 120
    },
    {
      "epoch": 0.5386479935362241,
      "grad_norm": 0.69041508436203,
      "learning_rate": 9.225022302802951e-06,
      "loss": 0.23021812438964845,
      "memory(GiB)": 29.07,
      "step": 125,
      "token_acc": 0.9175774697625143,
      "train_speed(iter/s)": 0.136057
    },
    {
      "epoch": 0.5601939132776731,
      "grad_norm": 0.8726988434791565,
      "learning_rate": 9.163606776403182e-06,
      "loss": 0.24287738800048828,
      "memory(GiB)": 29.07,
      "step": 130,
      "token_acc": 0.9174245368571058,
      "train_speed(iter/s)": 0.136884
    },
    {
      "epoch": 0.581739833019122,
      "grad_norm": 0.782781720161438,
      "learning_rate": 9.100070582666796e-06,
      "loss": 0.2355494499206543,
      "memory(GiB)": 29.07,
      "step": 135,
      "token_acc": 0.9218754381537254,
      "train_speed(iter/s)": 0.137519
    },
    {
      "epoch": 0.603285752760571,
      "grad_norm": 0.8618416786193848,
      "learning_rate": 9.034446082750352e-06,
      "loss": 0.26097152233123777,
      "memory(GiB)": 29.07,
      "step": 140,
      "token_acc": 0.9200457337339467,
      "train_speed(iter/s)": 0.137992
    },
    {
      "epoch": 0.603285752760571,
      "eval_loss": 0.2570641040802002,
      "eval_runtime": 9.1422,
      "eval_samples_per_second": 16.407,
      "eval_steps_per_second": 4.157,
      "eval_token_acc": 0.9119867373220047,
      "step": 140
    },
    {
      "epoch": 0.6248316725020199,
      "grad_norm": 0.7354313135147095,
      "learning_rate": 8.966766701456177e-06,
      "loss": 0.2450572967529297,
      "memory(GiB)": 29.07,
      "step": 145,
      "token_acc": 0.9104867095521301,
      "train_speed(iter/s)": 0.135924
    },
    {
      "epoch": 0.6463775922434689,
      "grad_norm": 0.7455778121948242,
      "learning_rate": 8.897066910207958e-06,
      "loss": 0.24456512928009033,
      "memory(GiB)": 29.07,
      "step": 150,
      "token_acc": 0.9091944119638131,
      "train_speed(iter/s)": 0.136598
    },
    {
      "epoch": 0.6679235119849178,
      "grad_norm": 0.6894455552101135,
      "learning_rate": 8.825382209493284e-06,
      "loss": 0.22335777282714844,
      "memory(GiB)": 29.07,
      "step": 155,
      "token_acc": 0.9254186825455899,
      "train_speed(iter/s)": 0.137109
    },
    {
      "epoch": 0.6894694317263668,
      "grad_norm": 0.8090242743492126,
      "learning_rate": 8.751749110782013e-06,
      "loss": 0.22939071655273438,
      "memory(GiB)": 29.07,
      "step": 160,
      "token_acc": 0.917912822144448,
      "train_speed(iter/s)": 0.137664
    },
    {
      "epoch": 0.6894694317263668,
      "eval_loss": 0.2559308111667633,
      "eval_runtime": 9.1265,
      "eval_samples_per_second": 16.436,
      "eval_steps_per_second": 4.164,
      "eval_token_acc": 0.9122235708576334,
      "step": 160
    },
    {
      "epoch": 0.7110153514678158,
      "grad_norm": 0.6877977252006531,
      "learning_rate": 8.676205117929752e-06,
      "loss": 0.2350329875946045,
      "memory(GiB)": 29.07,
      "step": 165,
      "token_acc": 0.9184605472599106,
      "train_speed(iter/s)": 0.13599
    },
    {
      "epoch": 0.7325612712092647,
      "grad_norm": 0.7360235452651978,
      "learning_rate": 8.598788708075844e-06,
      "loss": 0.23023662567138672,
      "memory(GiB)": 29.07,
      "step": 170,
      "token_acc": 0.9058568071626164,
      "train_speed(iter/s)": 0.136418
    },
    {
      "epoch": 0.7541071909507137,
      "grad_norm": 0.8410085439682007,
      "learning_rate": 8.51953931204566e-06,
      "loss": 0.23642911911010742,
      "memory(GiB)": 29.07,
      "step": 175,
      "token_acc": 0.9052932094269817,
      "train_speed(iter/s)": 0.137007
    },
    {
      "epoch": 0.7756531106921627,
      "grad_norm": 0.7055257558822632,
      "learning_rate": 8.438497294267117e-06,
      "loss": 0.21671390533447266,
      "memory(GiB)": 29.07,
      "step": 180,
      "token_acc": 0.9202059202059202,
      "train_speed(iter/s)": 0.137438
    },
    {
      "epoch": 0.7756531106921627,
      "eval_loss": 0.251537024974823,
      "eval_runtime": 9.1508,
      "eval_samples_per_second": 16.392,
      "eval_steps_per_second": 4.153,
      "eval_token_acc": 0.9133337280558927,
      "step": 180
    },
    {
      "epoch": 0.7971990304336116,
      "grad_norm": 0.7327454686164856,
      "learning_rate": 8.3557039322117e-06,
      "loss": 0.23624320030212403,
      "memory(GiB)": 29.07,
      "step": 185,
      "token_acc": 0.9212678936605317,
      "train_speed(iter/s)": 0.135745
    },
    {
      "epoch": 0.8187449501750605,
      "grad_norm": 0.7439467906951904,
      "learning_rate": 8.27120139537044e-06,
      "loss": 0.226143741607666,
      "memory(GiB)": 29.07,
      "step": 190,
      "token_acc": 0.9260711777101381,
      "train_speed(iter/s)": 0.136185
    },
    {
      "epoch": 0.8402908699165096,
      "grad_norm": 0.7658076286315918,
      "learning_rate": 8.18503272377554e-06,
      "loss": 0.23765263557434083,
      "memory(GiB)": 29.07,
      "step": 195,
      "token_acc": 0.9223176899233237,
      "train_speed(iter/s)": 0.136807
    },
    {
      "epoch": 0.8618367896579585,
      "grad_norm": 0.739122211933136,
      "learning_rate": 8.097241806078616e-06,
      "loss": 0.2310422420501709,
      "memory(GiB)": 29.07,
      "step": 200,
      "token_acc": 0.9296302733841532,
      "train_speed(iter/s)": 0.13714
    },
    {
      "epoch": 0.8618367896579585,
      "eval_loss": 0.24832946062088013,
      "eval_runtime": 9.1389,
      "eval_samples_per_second": 16.413,
      "eval_steps_per_second": 4.158,
      "eval_token_acc": 0.9134077385357766,
      "step": 200
    },
    {
      "epoch": 0.8833827093994074,
      "grad_norm": 0.7651655077934265,
      "learning_rate": 8.007873357196716e-06,
      "loss": 0.24373788833618165,
      "memory(GiB)": 29.07,
      "step": 205,
      "token_acc": 0.9188066080938974,
      "train_speed(iter/s)": 0.135947
    },
    {
      "epoch": 0.9049286291408565,
      "grad_norm": 0.7528461217880249,
      "learning_rate": 7.916972895537471e-06,
      "loss": 0.23199746608734131,
      "memory(GiB)": 29.07,
      "step": 210,
      "token_acc": 0.922004793261512,
      "train_speed(iter/s)": 0.136394
    },
    {
      "epoch": 0.9264745488823054,
      "grad_norm": 0.8405919075012207,
      "learning_rate": 7.824586719815019e-06,
      "loss": 0.2173825740814209,
      "memory(GiB)": 29.07,
      "step": 215,
      "token_acc": 0.9274689356403538,
      "train_speed(iter/s)": 0.136811
    },
    {
      "epoch": 0.9480204686237543,
      "grad_norm": 0.7239152193069458,
      "learning_rate": 7.730761885468486e-06,
      "loss": 0.22583813667297364,
      "memory(GiB)": 29.07,
      "step": 220,
      "token_acc": 0.9232335730673059,
      "train_speed(iter/s)": 0.137394
    },
    {
      "epoch": 0.9480204686237543,
      "eval_loss": 0.2480185180902481,
      "eval_runtime": 9.1514,
      "eval_samples_per_second": 16.391,
      "eval_steps_per_second": 4.152,
      "eval_token_acc": 0.9135261553035909,
      "step": 220
    },
    {
      "epoch": 0.9695663883652034,
      "grad_norm": 0.7927827835083008,
      "learning_rate": 7.635546180695039e-06,
      "loss": 0.24525394439697265,
      "memory(GiB)": 31.51,
      "step": 225,
      "token_acc": 0.9190034762456547,
      "train_speed(iter/s)": 0.136277
    },
    {
      "epoch": 0.9911123081066523,
      "grad_norm": 0.7584970593452454,
      "learning_rate": 7.538988102109728e-06,
      "loss": 0.24703338146209716,
      "memory(GiB)": 31.51,
      "step": 230,
      "token_acc": 0.9166930084197822,
      "train_speed(iter/s)": 0.136884
    },
    {
      "epoch": 1.0086183678965797,
      "grad_norm": 0.6898382306098938,
      "learning_rate": 7.441136830044495e-06,
      "loss": 0.19301869869232177,
      "memory(GiB)": 31.51,
      "step": 235,
      "token_acc": 0.9391968052558693,
      "train_speed(iter/s)": 0.137591
    },
    {
      "epoch": 1.0301642876380286,
      "grad_norm": 0.8064629435539246,
      "learning_rate": 7.342042203498952e-06,
      "loss": 0.16187149286270142,
      "memory(GiB)": 31.51,
      "step": 240,
      "token_acc": 0.9393010954707055,
      "train_speed(iter/s)": 0.137991
    },
    {
      "epoch": 1.0301642876380286,
      "eval_loss": 0.2533319890499115,
      "eval_runtime": 9.1306,
      "eval_samples_per_second": 16.428,
      "eval_steps_per_second": 4.162,
      "eval_token_acc": 0.914014624470825,
      "step": 240
    },
    {
      "epoch": 1.0517102073794775,
      "grad_norm": 0.8042640089988708,
      "learning_rate": 7.241754694755674e-06,
      "loss": 0.16961312294006348,
      "memory(GiB)": 31.51,
      "step": 245,
      "token_acc": 0.9325969259837942,
      "train_speed(iter/s)": 0.13689
    },
    {
      "epoch": 1.0732561271209264,
      "grad_norm": 0.8370431661605835,
      "learning_rate": 7.140325383672938e-06,
      "loss": 0.1677647829055786,
      "memory(GiB)": 31.51,
      "step": 250,
      "token_acc": 0.9420376456528234,
      "train_speed(iter/s)": 0.13733
    },
    {
      "epoch": 1.0948020468623754,
      "grad_norm": 0.7853599190711975,
      "learning_rate": 7.037805931668006e-06,
      "loss": 0.16614892482757568,
      "memory(GiB)": 31.51,
      "step": 255,
      "token_acc": 0.9391651270517043,
      "train_speed(iter/s)": 0.137609
    },
    {
      "epoch": 1.1163479666038243,
      "grad_norm": 0.6807184815406799,
      "learning_rate": 6.934248555404197e-06,
      "loss": 0.1581436276435852,
      "memory(GiB)": 31.51,
      "step": 260,
      "token_acc": 0.9458589779605179,
      "train_speed(iter/s)": 0.137947
    },
    {
      "epoch": 1.1163479666038243,
      "eval_loss": 0.2524171471595764,
      "eval_runtime": 9.156,
      "eval_samples_per_second": 16.383,
      "eval_steps_per_second": 4.15,
      "eval_token_acc": 0.9150063649012701,
      "step": 260
    },
    {
      "epoch": 1.1378938863452734,
      "grad_norm": 0.7507938146591187,
      "learning_rate": 6.8297060001951545e-06,
      "loss": 0.16150083541870117,
      "memory(GiB)": 31.51,
      "step": 265,
      "token_acc": 0.9309385421629347,
      "train_speed(iter/s)": 0.137099
    },
    {
      "epoch": 1.1594398060867224,
      "grad_norm": 0.7291717529296875,
      "learning_rate": 6.724231513139853e-06,
      "loss": 0.16564717292785644,
      "memory(GiB)": 31.51,
      "step": 270,
      "token_acc": 0.9417148494231771,
      "train_speed(iter/s)": 0.137339
    },
    {
      "epoch": 1.1809857258281713,
      "grad_norm": 0.7378965616226196,
      "learning_rate": 6.617878816002032e-06,
      "loss": 0.1618717670440674,
      "memory(GiB)": 31.51,
      "step": 275,
      "token_acc": 0.9485524256651017,
      "train_speed(iter/s)": 0.137622
    },
    {
      "epoch": 1.2025316455696202,
      "grad_norm": 0.8035087585449219,
      "learning_rate": 6.510702077847864e-06,
      "loss": 0.1574448823928833,
      "memory(GiB)": 31.51,
      "step": 280,
      "token_acc": 0.9411747078213965,
      "train_speed(iter/s)": 0.138001
    },
    {
      "epoch": 1.2025316455696202,
      "eval_loss": 0.24995924532413483,
      "eval_runtime": 9.1411,
      "eval_samples_per_second": 16.409,
      "eval_steps_per_second": 4.157,
      "eval_token_acc": 0.9145474999259895,
      "step": 280
    },
    {
      "epoch": 1.2240775653110691,
      "grad_norm": 0.7014583349227905,
      "learning_rate": 6.402755887455792e-06,
      "loss": 0.1643718123435974,
      "memory(GiB)": 31.51,
      "step": 285,
      "token_acc": 0.9339765241569784,
      "train_speed(iter/s)": 0.137005
    },
    {
      "epoch": 1.2456234850525183,
      "grad_norm": 0.7766486406326294,
      "learning_rate": 6.294095225512604e-06,
      "loss": 0.16814930438995362,
      "memory(GiB)": 31.51,
      "step": 290,
      "token_acc": 0.9360814298463542,
      "train_speed(iter/s)": 0.137309
    },
    {
      "epoch": 1.2671694047939672,
      "grad_norm": 0.7775722146034241,
      "learning_rate": 6.184775436609885e-06,
      "loss": 0.1682277202606201,
      "memory(GiB)": 31.51,
      "step": 295,
      "token_acc": 0.9411593528110813,
      "train_speed(iter/s)": 0.137689
    },
    {
      "epoch": 1.2887153245354162,
      "grad_norm": 0.7489521503448486,
      "learning_rate": 6.074852201055121e-06,
      "loss": 0.16042615175247193,
      "memory(GiB)": 31.51,
      "step": 300,
      "token_acc": 0.9415187229598687,
      "train_speed(iter/s)": 0.138077
    },
    {
      "epoch": 1.2887153245354162,
      "eval_loss": 0.25172922015190125,
      "eval_runtime": 9.1377,
      "eval_samples_per_second": 16.415,
      "eval_steps_per_second": 4.159,
      "eval_token_acc": 0.9150211669972468,
      "step": 300
    },
    {
      "epoch": 1.310261244276865,
      "grad_norm": 0.792445182800293,
      "learning_rate": 5.964381506511823e-06,
      "loss": 0.16529514789581298,
      "memory(GiB)": 31.51,
      "step": 305,
      "token_acc": 0.9390098282355103,
      "train_speed(iter/s)": 0.137303
    },
    {
      "epoch": 1.331807164018314,
      "grad_norm": 0.7657850980758667,
      "learning_rate": 5.853419619483083e-06,
      "loss": 0.16101518869400025,
      "memory(GiB)": 31.51,
      "step": 310,
      "token_acc": 0.9423482091553342,
      "train_speed(iter/s)": 0.137569
    },
    {
      "epoch": 1.353353083759763,
      "grad_norm": 0.7221185564994812,
      "learning_rate": 5.742023056653131e-06,
      "loss": 0.16527080535888672,
      "memory(GiB)": 31.51,
      "step": 315,
      "token_acc": 0.9436092441929018,
      "train_speed(iter/s)": 0.13783
    },
    {
      "epoch": 1.3748990035012119,
      "grad_norm": 0.7651124596595764,
      "learning_rate": 5.630248556101448e-06,
      "loss": 0.16076445579528809,
      "memory(GiB)": 31.51,
      "step": 320,
      "token_acc": 0.941539852342926,
      "train_speed(iter/s)": 0.138138
    },
    {
      "epoch": 1.3748990035012119,
      "eval_loss": 0.2519991993904114,
      "eval_runtime": 9.1625,
      "eval_samples_per_second": 16.371,
      "eval_steps_per_second": 4.147,
      "eval_token_acc": 0.9146955208857575,
      "step": 320
    },
    {
      "epoch": 1.3964449232426608,
      "grad_norm": 0.6687130331993103,
      "learning_rate": 5.51815304840412e-06,
      "loss": 0.16071187257766723,
      "memory(GiB)": 31.51,
      "step": 325,
      "token_acc": 0.93119978263401,
      "train_speed(iter/s)": 0.137342
    },
    {
      "epoch": 1.41799084298411,
      "grad_norm": 0.7091411352157593,
      "learning_rate": 5.405793627637157e-06,
      "loss": 0.15800976753234863,
      "memory(GiB)": 31.51,
      "step": 330,
      "token_acc": 0.9493767600253226,
      "train_speed(iter/s)": 0.137567
    },
    {
      "epoch": 1.4395367627255589,
      "grad_norm": 0.7872418761253357,
      "learning_rate": 5.293227522296517e-06,
      "loss": 0.16303786039352416,
      "memory(GiB)": 31.51,
      "step": 335,
      "token_acc": 0.9474813818783616,
      "train_speed(iter/s)": 0.137773
    },
    {
      "epoch": 1.4610826824670078,
      "grad_norm": 0.696894109249115,
      "learning_rate": 5.180512066149682e-06,
      "loss": 0.1651884913444519,
      "memory(GiB)": 31.51,
      "step": 340,
      "token_acc": 0.9437182487584908,
      "train_speed(iter/s)": 0.138053
    },
    {
      "epoch": 1.4610826824670078,
      "eval_loss": 0.2488545924425125,
      "eval_runtime": 9.1671,
      "eval_samples_per_second": 16.363,
      "eval_steps_per_second": 4.145,
      "eval_token_acc": 0.9151395837650611,
      "step": 340
    },
    {
      "epoch": 1.4826286022084567,
      "grad_norm": 0.7297951579093933,
      "learning_rate": 5.06770466903361e-06,
      "loss": 0.15690993070602416,
      "memory(GiB)": 31.51,
      "step": 345,
      "token_acc": 0.9340376019152534,
      "train_speed(iter/s)": 0.137212
    },
    {
      "epoch": 1.5041745219499059,
      "grad_norm": 0.7707265019416809,
      "learning_rate": 4.954862787613937e-06,
      "loss": 0.15354007482528687,
      "memory(GiB)": 31.51,
      "step": 350,
      "token_acc": 0.9396699063799393,
      "train_speed(iter/s)": 0.13744
    },
    {
      "epoch": 1.5257204416913548,
      "grad_norm": 0.7526496052742004,
      "learning_rate": 4.842043896120332e-06,
      "loss": 0.16020708084106444,
      "memory(GiB)": 31.51,
      "step": 355,
      "token_acc": 0.9479154768703598,
      "train_speed(iter/s)": 0.137715
    },
    {
      "epoch": 1.5472663614328037,
      "grad_norm": 0.7758511900901794,
      "learning_rate": 4.729305457072913e-06,
      "loss": 0.16963763236999513,
      "memory(GiB)": 31.51,
      "step": 360,
      "token_acc": 0.9411747908278363,
      "train_speed(iter/s)": 0.138029
    },
    {
      "epoch": 1.5472663614328037,
      "eval_loss": 0.24962776899337769,
      "eval_runtime": 9.1377,
      "eval_samples_per_second": 16.415,
      "eval_steps_per_second": 4.159,
      "eval_token_acc": 0.9158204801799935,
      "step": 360
    },
    {
      "epoch": 1.5688122811742526,
      "grad_norm": 0.7562235593795776,
      "learning_rate": 4.616704892014613e-06,
      "loss": 0.1591555118560791,
      "memory(GiB)": 31.51,
      "step": 365,
      "token_acc": 0.9352809509107749,
      "train_speed(iter/s)": 0.137302
    },
    {
      "epoch": 1.5903582009157016,
      "grad_norm": 0.7587376236915588,
      "learning_rate": 4.504299552264428e-06,
      "loss": 0.15684648752212524,
      "memory(GiB)": 31.51,
      "step": 370,
      "token_acc": 0.9416271651313239,
      "train_speed(iter/s)": 0.137471
    },
    {
      "epoch": 1.6119041206571505,
      "grad_norm": 0.8137562870979309,
      "learning_rate": 4.392146689706426e-06,
      "loss": 0.1647357702255249,
      "memory(GiB)": 31.51,
      "step": 375,
      "token_acc": 0.9458877409154104,
      "train_speed(iter/s)": 0.137806
    },
    {
      "epoch": 1.6334500403985994,
      "grad_norm": 0.7551019191741943,
      "learning_rate": 4.280303427629404e-06,
      "loss": 0.15853278636932372,
      "memory(GiB)": 31.51,
      "step": 380,
      "token_acc": 0.9448746907604604,
      "train_speed(iter/s)": 0.138095
    },
    {
      "epoch": 1.6334500403985994,
      "eval_loss": 0.25000789761543274,
      "eval_runtime": 9.1322,
      "eval_samples_per_second": 16.425,
      "eval_steps_per_second": 4.161,
      "eval_token_acc": 0.9155096361644809,
      "step": 380
    },
    {
      "epoch": 1.6549959601400483,
      "grad_norm": 0.7977014780044556,
      "learning_rate": 4.168826731632052e-06,
      "loss": 0.15717003345489503,
      "memory(GiB)": 31.51,
      "step": 385,
      "token_acc": 0.9312666413084824,
      "train_speed(iter/s)": 0.137411
    },
    {
      "epoch": 1.6765418798814973,
      "grad_norm": 0.7832633852958679,
      "learning_rate": 4.057773380608411e-06,
      "loss": 0.1634294271469116,
      "memory(GiB)": 31.51,
      "step": 390,
      "token_acc": 0.9465973781793492,
      "train_speed(iter/s)": 0.137768
    },
    {
      "epoch": 1.6980877996229464,
      "grad_norm": 0.8293562531471252,
      "learning_rate": 3.947199937828447e-06,
      "loss": 0.16505708694458007,
      "memory(GiB)": 31.51,
      "step": 395,
      "token_acc": 0.9415725074644342,
      "train_speed(iter/s)": 0.138088
    },
    {
      "epoch": 1.7196337193643954,
      "grad_norm": 0.7886548042297363,
      "learning_rate": 3.8371627221284495e-06,
      "loss": 0.1561971426010132,
      "memory(GiB)": 31.51,
      "step": 400,
      "token_acc": 0.9470925236321971,
      "train_speed(iter/s)": 0.138285
    },
    {
      "epoch": 1.7196337193643954,
      "eval_loss": 0.24771690368652344,
      "eval_runtime": 9.1277,
      "eval_samples_per_second": 16.434,
      "eval_steps_per_second": 4.163,
      "eval_token_acc": 0.915642855028272,
      "step": 400
    },
    {
      "epoch": 1.7411796391058443,
      "grad_norm": 0.7254658937454224,
      "learning_rate": 3.727717779225912e-06,
      "loss": 0.1556318521499634,
      "memory(GiB)": 31.51,
      "step": 405,
      "token_acc": 0.9359159282917783,
      "train_speed(iter/s)": 0.137626
    },
    {
      "epoch": 1.7627255588472934,
      "grad_norm": 0.7896953225135803,
      "learning_rate": 3.6189208531735354e-06,
      "loss": 0.16613179445266724,
      "memory(GiB)": 31.51,
      "step": 410,
      "token_acc": 0.9352341759749168,
      "train_speed(iter/s)": 0.137957
    },
    {
      "epoch": 1.7842714785887424,
      "grad_norm": 0.6848239898681641,
      "learning_rate": 3.510827357966876e-06,
      "loss": 0.1551806092262268,
      "memory(GiB)": 31.51,
      "step": 415,
      "token_acc": 0.9506668360218469,
      "train_speed(iter/s)": 0.138211
    },
    {
      "epoch": 1.8058173983301913,
      "grad_norm": 0.7046887874603271,
      "learning_rate": 3.403492349320101e-06,
      "loss": 0.15121963024139404,
      "memory(GiB)": 31.51,
      "step": 420,
      "token_acc": 0.9443215339233039,
      "train_speed(iter/s)": 0.138347
    },
    {
      "epoch": 1.8058173983301913,
      "eval_loss": 0.24637845158576965,
      "eval_runtime": 9.144,
      "eval_samples_per_second": 16.404,
      "eval_steps_per_second": 4.156,
      "eval_token_acc": 0.9157020634121792,
      "step": 420
    },
    {
      "epoch": 1.8273633180716402,
      "grad_norm": 0.7264747619628906,
      "learning_rate": 3.29697049662423e-06,
      "loss": 0.1486160159111023,
      "memory(GiB)": 31.51,
      "step": 425,
      "token_acc": 0.9378997513712539,
      "train_speed(iter/s)": 0.137609
    },
    {
      "epoch": 1.8489092378130891,
      "grad_norm": 0.6881827116012573,
      "learning_rate": 3.191316055102146e-06,
      "loss": 0.14999903440475465,
      "memory(GiB)": 31.51,
      "step": 430,
      "token_acc": 0.9458710676835081,
      "train_speed(iter/s)": 0.13782
    },
    {
      "epoch": 1.870455157554538,
      "grad_norm": 0.7096033096313477,
      "learning_rate": 3.0865828381745515e-06,
      "loss": 0.15066919326782227,
      "memory(GiB)": 31.51,
      "step": 435,
      "token_acc": 0.9486315094650982,
      "train_speed(iter/s)": 0.137979
    },
    {
      "epoch": 1.892001077295987,
      "grad_norm": 0.7479064464569092,
      "learning_rate": 2.982824190050958e-06,
      "loss": 0.165749990940094,
      "memory(GiB)": 31.51,
      "step": 440,
      "token_acc": 0.9481958622195534,
      "train_speed(iter/s)": 0.138187
    },
    {
      "epoch": 1.892001077295987,
      "eval_loss": 0.24523746967315674,
      "eval_runtime": 9.1328,
      "eval_samples_per_second": 16.424,
      "eval_steps_per_second": 4.161,
      "eval_token_acc": 0.9161757304834365,
      "step": 440
    },
    {
      "epoch": 1.913546997037436,
      "grad_norm": 0.7331624031066895,
      "learning_rate": 2.8800929585596506e-06,
      "loss": 0.15496289730072021,
      "memory(GiB)": 31.51,
      "step": 445,
      "token_acc": 0.9352829677768751,
      "train_speed(iter/s)": 0.137542
    },
    {
      "epoch": 1.9350929167788848,
      "grad_norm": 0.6734929084777832,
      "learning_rate": 2.778441468230483e-06,
      "loss": 0.1523799180984497,
      "memory(GiB)": 31.51,
      "step": 450,
      "token_acc": 0.9479633806554332,
      "train_speed(iter/s)": 0.1377
    },
    {
      "epoch": 1.956638836520334,
      "grad_norm": 0.7542054057121277,
      "learning_rate": 2.6779214936442056e-06,
      "loss": 0.16172744035720826,
      "memory(GiB)": 31.51,
      "step": 455,
      "token_acc": 0.935499950154521,
      "train_speed(iter/s)": 0.137884
    },
    {
      "epoch": 1.978184756261783,
      "grad_norm": 0.7129687070846558,
      "learning_rate": 2.5785842330619038e-06,
      "loss": 0.15356701612472534,
      "memory(GiB)": 31.51,
      "step": 460,
      "token_acc": 0.941601546088564,
      "train_speed(iter/s)": 0.13804
    },
    {
      "epoch": 1.978184756261783,
      "eval_loss": 0.24467714130878448,
      "eval_runtime": 9.1316,
      "eval_samples_per_second": 16.426,
      "eval_steps_per_second": 4.161,
      "eval_token_acc": 0.9170934604339974,
      "step": 460
    },
    {
      "epoch": 1.9997306760032318,
      "grad_norm": 0.7861402630805969,
      "learning_rate": 2.480480282347961e-06,
      "loss": 0.15792056322097778,
      "memory(GiB)": 31.51,
      "step": 465,
      "token_acc": 0.9337420552337027,
      "train_speed(iter/s)": 0.137533
    },
    {
      "epoch": 2.0172367357931593,
      "grad_norm": 0.6826748847961426,
      "learning_rate": 2.383659609199873e-06,
      "loss": 0.14240689277648927,
      "memory(GiB)": 31.51,
      "step": 470,
      "token_acc": 0.9578195371952166,
      "train_speed(iter/s)": 0.137988
    },
    {
      "epoch": 2.0387826555346082,
      "grad_norm": 0.6501537561416626,
      "learning_rate": 2.2881715276979705e-06,
      "loss": 0.10814023017883301,
      "memory(GiB)": 31.51,
      "step": 475,
      "token_acc": 0.9586633663366336,
      "train_speed(iter/s)": 0.138103
    },
    {
      "epoch": 2.060328575276057,
      "grad_norm": 0.6575304269790649,
      "learning_rate": 2.1940646731880887e-06,
      "loss": 0.1118842363357544,
      "memory(GiB)": 31.51,
      "step": 480,
      "token_acc": 0.9698543524895563,
      "train_speed(iter/s)": 0.138265
    },
    {
      "epoch": 2.060328575276057,
      "eval_loss": 0.26591211557388306,
      "eval_runtime": 9.1643,
      "eval_samples_per_second": 16.368,
      "eval_steps_per_second": 4.147,
      "eval_token_acc": 0.9162201367713668,
      "step": 480
    },
    {
      "epoch": 2.081874495017506,
      "grad_norm": 0.765779972076416,
      "learning_rate": 2.101386977509907e-06,
      "loss": 0.12155743837356567,
      "memory(GiB)": 31.51,
      "step": 485,
      "token_acc": 0.946289860026969,
      "train_speed(iter/s)": 0.137773
    },
    {
      "epoch": 2.103420414758955,
      "grad_norm": 0.7146125435829163,
      "learning_rate": 2.010185644583641e-06,
      "loss": 0.11463183164596558,
      "memory(GiB)": 31.51,
      "step": 490,
      "token_acc": 0.9624759934997784,
      "train_speed(iter/s)": 0.137939
    },
    {
      "epoch": 2.124966334500404,
      "grad_norm": 0.6777431964874268,
      "learning_rate": 1.920507126367448e-06,
      "loss": 0.10685477256774903,
      "memory(GiB)": 31.51,
      "step": 495,
      "token_acc": 0.9612281857095818,
      "train_speed(iter/s)": 0.138102
    },
    {
      "epoch": 2.146512254241853,
      "grad_norm": 0.7272450923919678,
      "learning_rate": 1.8323970991978823e-06,
      "loss": 0.10419889688491821,
      "memory(GiB)": 31.51,
      "step": 500,
      "token_acc": 0.9610325296357052,
      "train_speed(iter/s)": 0.138212
    },
    {
      "epoch": 2.146512254241853,
      "eval_loss": 0.26627564430236816,
      "eval_runtime": 9.1679,
      "eval_samples_per_second": 16.361,
      "eval_steps_per_second": 4.145,
      "eval_token_acc": 0.9159833032357382,
      "step": 500
    },
    {
      "epoch": 2.168058173983302,
      "grad_norm": 0.7035279273986816,
      "learning_rate": 1.7459004405253544e-06,
      "loss": 0.1082868218421936,
      "memory(GiB)": 31.51,
      "step": 505,
      "token_acc": 0.9475257941268758,
      "train_speed(iter/s)": 0.137652
    },
    {
      "epoch": 2.1896040937247507,
      "grad_norm": 0.601047158241272,
      "learning_rate": 1.6610612060565235e-06,
      "loss": 0.09674398303031921,
      "memory(GiB)": 31.51,
      "step": 510,
      "token_acc": 0.9646692233940556,
      "train_speed(iter/s)": 0.137835
    },
    {
      "epoch": 2.2111500134661997,
      "grad_norm": 0.7340168356895447,
      "learning_rate": 1.5779226073152071e-06,
      "loss": 0.1145021677017212,
      "memory(GiB)": 31.51,
      "step": 515,
      "token_acc": 0.9572068592615479,
      "train_speed(iter/s)": 0.138109
    },
    {
      "epoch": 2.2326959332076486,
      "grad_norm": 0.7059099078178406,
      "learning_rate": 1.4965269896332884e-06,
      "loss": 0.1138340711593628,
      "memory(GiB)": 31.51,
      "step": 520,
      "token_acc": 0.9643348939686037,
      "train_speed(iter/s)": 0.138301
    },
    {
      "epoch": 2.2326959332076486,
      "eval_loss": 0.26528117060661316,
      "eval_runtime": 9.1705,
      "eval_samples_per_second": 16.357,
      "eval_steps_per_second": 4.144,
      "eval_token_acc": 0.9159536990437847,
      "step": 520
    },
    {
      "epoch": 2.254241852949098,
      "grad_norm": 0.7410480976104736,
      "learning_rate": 1.4169158105827768e-06,
      "loss": 0.11105086803436279,
      "memory(GiB)": 31.51,
      "step": 525,
      "token_acc": 0.9514605435256503,
      "train_speed(iter/s)": 0.137827
    },
    {
      "epoch": 2.275787772690547,
      "grad_norm": 0.6698100566864014,
      "learning_rate": 1.3391296188600594e-06,
      "loss": 0.10843292474746705,
      "memory(GiB)": 31.51,
      "step": 530,
      "token_acc": 0.9629225092250923,
      "train_speed(iter/s)": 0.137941
    },
    {
      "epoch": 2.297333692431996,
      "grad_norm": 0.6693587303161621,
      "learning_rate": 1.2632080336330532e-06,
      "loss": 0.11362366676330567,
      "memory(GiB)": 31.51,
      "step": 535,
      "token_acc": 0.9621976353183642,
      "train_speed(iter/s)": 0.138116
    },
    {
      "epoch": 2.3188796121734447,
      "grad_norm": 0.6858277320861816,
      "learning_rate": 1.1891897243618184e-06,
      "loss": 0.10754673480987549,
      "memory(GiB)": 31.51,
      "step": 540,
      "token_acc": 0.9638930030070464,
      "train_speed(iter/s)": 0.138279
    },
    {
      "epoch": 2.3188796121734447,
      "eval_loss": 0.2651301622390747,
      "eval_runtime": 9.1447,
      "eval_samples_per_second": 16.403,
      "eval_steps_per_second": 4.155,
      "eval_token_acc": 0.91620533467539,
      "step": 540
    },
    {
      "epoch": 2.3404255319148937,
      "grad_norm": 0.7329075932502747,
      "learning_rate": 1.1171123911028692e-06,
      "loss": 0.10752699375152588,
      "memory(GiB)": 31.51,
      "step": 545,
      "token_acc": 0.946654961925566,
      "train_speed(iter/s)": 0.137811
    },
    {
      "epoch": 2.3619714516563426,
      "grad_norm": 0.7588092684745789,
      "learning_rate": 1.047012745307255e-06,
      "loss": 0.10413261651992797,
      "memory(GiB)": 31.51,
      "step": 550,
      "token_acc": 0.9638513608403786,
      "train_speed(iter/s)": 0.137959
    },
    {
      "epoch": 2.3835173713977915,
      "grad_norm": 0.6776463389396667,
      "learning_rate": 9.789264911221546e-07,
      "loss": 0.11203373670578003,
      "memory(GiB)": 31.51,
      "step": 555,
      "token_acc": 0.9539418840061927,
      "train_speed(iter/s)": 0.138182
    },
    {
      "epoch": 2.4050632911392404,
      "grad_norm": 0.6233177185058594,
      "learning_rate": 9.128883072055411e-07,
      "loss": 0.10640518665313721,
      "memory(GiB)": 31.51,
      "step": 560,
      "token_acc": 0.9605239362389232,
      "train_speed(iter/s)": 0.138441
    },
    {
      "epoch": 2.4050632911392404,
      "eval_loss": 0.26431551575660706,
      "eval_runtime": 9.1559,
      "eval_samples_per_second": 16.383,
      "eval_steps_per_second": 4.15,
      "eval_token_acc": 0.9158056780840167,
      "step": 560
    },
    {
      "epoch": 2.4266092108806894,
      "grad_norm": 0.6891351938247681,
      "learning_rate": 8.489318290631454e-07,
      "loss": 0.11017493009567261,
      "memory(GiB)": 31.51,
      "step": 565,
      "token_acc": 0.9501483222252186,
      "train_speed(iter/s)": 0.137996
    },
    {
      "epoch": 2.4481551306221383,
      "grad_norm": 0.685417890548706,
      "learning_rate": 7.870896319167548e-07,
      "loss": 0.10502817630767822,
      "memory(GiB)": 31.51,
      "step": 570,
      "token_acc": 0.9675666865866247,
      "train_speed(iter/s)": 0.138123
    },
    {
      "epoch": 2.4697010503635872,
      "grad_norm": 0.8273110389709473,
      "learning_rate": 7.273932141125256e-07,
      "loss": 0.11376097202301025,
      "memory(GiB)": 31.51,
      "step": 575,
      "token_acc": 0.9588286984389538,
      "train_speed(iter/s)": 0.138286
    },
    {
      "epoch": 2.4912469701050366,
      "grad_norm": 0.7995973825454712,
      "learning_rate": 6.698729810778065e-07,
      "loss": 0.1191399335861206,
      "memory(GiB)": 31.51,
      "step": 580,
      "token_acc": 0.9582757592998997,
      "train_speed(iter/s)": 0.138394
    },
    {
      "epoch": 2.4912469701050366,
      "eval_loss": 0.26525548100471497,
      "eval_runtime": 9.1367,
      "eval_samples_per_second": 16.417,
      "eval_steps_per_second": 4.159,
      "eval_token_acc": 0.9161461262914828,
      "step": 580
    },
    {
      "epoch": 2.5127928898464855,
      "grad_norm": 0.6976614594459534,
      "learning_rate": 6.145582298346153e-07,
      "loss": 0.10850718021392822,
      "memory(GiB)": 31.51,
      "step": 585,
      "token_acc": 0.9468696569536905,
      "train_speed(iter/s)": 0.137863
    },
    {
      "epoch": 2.5343388095879344,
      "grad_norm": 0.7127689123153687,
      "learning_rate": 5.614771340776559e-07,
      "loss": 0.1049992561340332,
      "memory(GiB)": 31.51,
      "step": 590,
      "token_acc": 0.9636775106082037,
      "train_speed(iter/s)": 0.137998
    },
    {
      "epoch": 2.5558847293293834,
      "grad_norm": 0.7365370392799377,
      "learning_rate": 5.106567298245008e-07,
      "loss": 0.11682652235031128,
      "memory(GiB)": 31.51,
      "step": 595,
      "token_acc": 0.9585528403681371,
      "train_speed(iter/s)": 0.138108
    },
    {
      "epoch": 2.5774306490708323,
      "grad_norm": 0.6995398998260498,
      "learning_rate": 4.6212290164521554e-07,
      "loss": 0.10941903591156006,
      "memory(GiB)": 31.51,
      "step": 600,
      "token_acc": 0.9588730068630993,
      "train_speed(iter/s)": 0.138228
    },
    {
      "epoch": 2.5774306490708323,
      "eval_loss": 0.2648448944091797,
      "eval_runtime": 9.1555,
      "eval_samples_per_second": 16.384,
      "eval_steps_per_second": 4.15,
      "eval_token_acc": 0.91620533467539,
      "step": 600
    },
    {
      "epoch": 2.5989765688122812,
      "grad_norm": 0.6558308601379395,
      "learning_rate": 4.159003694784647e-07,
      "loss": 0.09994454979896546,
      "memory(GiB)": 31.51,
      "step": 605,
      "token_acc": 0.9488222044057573,
      "train_speed(iter/s)": 0.137823
    },
    {
      "epoch": 2.62052248855373,
      "grad_norm": 0.5623044371604919,
      "learning_rate": 3.7201267604080436e-07,
      "loss": 0.10503623485565186,
      "memory(GiB)": 31.51,
      "step": 610,
      "token_acc": 0.9605492530908896,
      "train_speed(iter/s)": 0.137935
    },
    {
      "epoch": 2.642068408295179,
      "grad_norm": 0.7411386966705322,
      "learning_rate": 3.3048217483556743e-07,
      "loss": 0.10335917472839355,
      "memory(GiB)": 31.51,
      "step": 615,
      "token_acc": 0.9596229517824632,
      "train_speed(iter/s)": 0.138049
    },
    {
      "epoch": 2.663614328036628,
      "grad_norm": 0.7703331112861633,
      "learning_rate": 2.9133001876746004e-07,
      "loss": 0.11330341100692749,
      "memory(GiB)": 31.51,
      "step": 620,
      "token_acc": 0.9627358888545153,
      "train_speed(iter/s)": 0.138195
    },
    {
      "epoch": 2.663614328036628,
      "eval_loss": 0.2645653188228607,
      "eval_runtime": 9.1574,
      "eval_samples_per_second": 16.38,
      "eval_steps_per_second": 4.15,
      "eval_token_acc": 0.9161461262914828,
      "step": 620
    },
    {
      "epoch": 2.685160247778077,
      "grad_norm": 0.7208820581436157,
      "learning_rate": 2.545761493686666e-07,
      "loss": 0.10512195825576783,
      "memory(GiB)": 31.51,
      "step": 625,
      "token_acc": 0.9453959214438257,
      "train_speed(iter/s)": 0.137691
    },
    {
      "epoch": 2.706706167519526,
      "grad_norm": 0.792917013168335,
      "learning_rate": 2.2023928664194229e-07,
      "loss": 0.10448248386383056,
      "memory(GiB)": 31.51,
      "step": 630,
      "token_acc": 0.9634454263743831,
      "train_speed(iter/s)": 0.137865
    },
    {
      "epoch": 2.728252087260975,
      "grad_norm": 0.7074964046478271,
      "learning_rate": 1.8833691952587829e-07,
      "loss": 0.10274065732955932,
      "memory(GiB)": 31.51,
      "step": 635,
      "token_acc": 0.9614498168320434,
      "train_speed(iter/s)": 0.137973
    },
    {
      "epoch": 2.7497980070024237,
      "grad_norm": 0.695501446723938,
      "learning_rate": 1.5888529698718347e-07,
      "loss": 0.111275053024292,
      "memory(GiB)": 31.51,
      "step": 640,
      "token_acc": 0.9627026215729437,
      "train_speed(iter/s)": 0.13808
    },
    {
      "epoch": 2.7497980070024237,
      "eval_loss": 0.26474642753601074,
      "eval_runtime": 9.1444,
      "eval_samples_per_second": 16.403,
      "eval_steps_per_second": 4.156,
      "eval_token_acc": 0.9162349388673436,
      "step": 640
    },
    {
      "epoch": 2.7713439267438726,
      "grad_norm": 0.7010114192962646,
      "learning_rate": 1.3189941974453502e-07,
      "loss": 0.11862779855728149,
      "memory(GiB)": 31.51,
      "step": 645,
      "token_acc": 0.9463942439720986,
      "train_speed(iter/s)": 0.13766
    },
    {
      "epoch": 2.7928898464853216,
      "grad_norm": 0.7053817510604858,
      "learning_rate": 1.0739303262819301e-07,
      "loss": 0.10773177146911621,
      "memory(GiB)": 31.51,
      "step": 650,
      "token_acc": 0.9672267425750056,
      "train_speed(iter/s)": 0.137765
    },
    {
      "epoch": 2.814435766226771,
      "grad_norm": 0.7381494641304016,
      "learning_rate": 8.537861757929422e-08,
      "loss": 0.10787509679794312,
      "memory(GiB)": 31.51,
      "step": 655,
      "token_acc": 0.9632690990902866,
      "train_speed(iter/s)": 0.137917
    },
    {
      "epoch": 2.83598168596822,
      "grad_norm": 0.7262890934944153,
      "learning_rate": 6.58673872923693e-08,
      "loss": 0.11206209659576416,
      "memory(GiB)": 31.51,
      "step": 660,
      "token_acc": 0.965990990990991,
      "train_speed(iter/s)": 0.138079
    },
    {
      "epoch": 2.83598168596822,
      "eval_loss": 0.2648203372955322,
      "eval_runtime": 9.1399,
      "eval_samples_per_second": 16.412,
      "eval_steps_per_second": 4.158,
      "eval_token_acc": 0.916279345155274,
      "step": 660
    },
    {
      "epoch": 2.857527605709669,
      "grad_norm": 0.7241911888122559,
      "learning_rate": 4.88692795043344e-08,
      "loss": 0.10918653011322021,
      "memory(GiB)": 31.51,
      "step": 665,
      "token_acc": 0.9504457917261055,
      "train_speed(iter/s)": 0.137668
    },
    {
      "epoch": 2.8790735254511177,
      "grad_norm": 0.7945267558097839,
      "learning_rate": 3.439295193286174e-08,
      "loss": 0.11153676509857177,
      "memory(GiB)": 31.51,
      "step": 670,
      "token_acc": 0.9583095218657305,
      "train_speed(iter/s)": 0.13784
    },
    {
      "epoch": 2.9006194451925666,
      "grad_norm": 0.7374799847602844,
      "learning_rate": 2.2445777866709208e-08,
      "loss": 0.10855717658996582,
      "memory(GiB)": 31.51,
      "step": 675,
      "token_acc": 0.9629715143294179,
      "train_speed(iter/s)": 0.13794
    },
    {
      "epoch": 2.9221653649340156,
      "grad_norm": 0.6748504042625427,
      "learning_rate": 1.3033842410251074e-08,
      "loss": 0.11381592750549316,
      "memory(GiB)": 31.51,
      "step": 680,
      "token_acc": 0.9596367864459332,
      "train_speed(iter/s)": 0.138055
    },
    {
      "epoch": 2.9221653649340156,
      "eval_loss": 0.26483333110809326,
      "eval_runtime": 9.1492,
      "eval_samples_per_second": 16.395,
      "eval_steps_per_second": 4.153,
      "eval_token_acc": 0.9160425116196453,
      "step": 680
    },
    {
      "epoch": 2.9437112846754645,
      "grad_norm": 0.7286980748176575,
      "learning_rate": 6.16193938412557e-09,
      "loss": 0.10510704517364503,
      "memory(GiB)": 31.51,
      "step": 685,
      "token_acc": 0.9509096674461929,
      "train_speed(iter/s)": 0.137639
    },
    {
      "epoch": 2.9652572044169134,
      "grad_norm": 0.6892649531364441,
      "learning_rate": 1.8335688835802169e-09,
      "loss": 0.105083167552948,
      "memory(GiB)": 31.51,
      "step": 690,
      "token_acc": 0.9589703497799398,
      "train_speed(iter/s)": 0.137777
    },
    {
      "epoch": 2.9868031241583624,
      "grad_norm": 0.7908564209938049,
      "learning_rate": 5.093549575119205e-11,
      "loss": 0.10409483909606934,
      "memory(GiB)": 31.51,
      "step": 695,
      "token_acc": 0.9638513775207209,
      "train_speed(iter/s)": 0.137894
    },
    {
      "epoch": 2.9911123081066524,
      "eval_loss": 0.26468953490257263,
      "eval_runtime": 9.1426,
      "eval_samples_per_second": 16.407,
      "eval_steps_per_second": 4.156,
      "eval_token_acc": 0.9161757304834365,
      "step": 696
    }
  ],
  "logging_steps": 5,
  "max_steps": 696,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 20,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 8.392166013990339e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}