{
  "best_global_step": 320,
  "best_metric": 0.17490149,
  "best_model_checkpoint": "/data/home/scyb089/CODE/scripts/ms-swift/3b/v31-20250504-035000/checkpoint-320",
  "epoch": 2.9977046671767407,
  "eval_steps": 20,
  "global_step": 489,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00612088752869166,
      "grad_norm": 2.4262969493865967,
      "learning_rate": 9.999896814055916e-06,
      "loss": 0.29114457964897156,
      "memory(GiB)": 27.73,
      "step": 1,
      "token_acc": 0.8905354511216637,
      "train_speed(iter/s)": 0.075742
    },
    {
      "epoch": 0.030604437643458302,
      "grad_norm": 1.4515410661697388,
      "learning_rate": 9.99742056433851e-06,
      "loss": 0.2713624835014343,
      "memory(GiB)": 27.73,
      "step": 5,
      "token_acc": 0.9116365441541489,
      "train_speed(iter/s)": 0.152509
    },
    {
      "epoch": 0.061208875286916604,
      "grad_norm": 0.9496105909347534,
      "learning_rate": 9.989684918749365e-06,
      "loss": 0.2552653312683105,
      "memory(GiB)": 27.73,
      "step": 10,
      "token_acc": 0.9153024911032028,
      "train_speed(iter/s)": 0.174854
    },
    {
      "epoch": 0.09181331293037491,
      "grad_norm": 0.7063636183738708,
      "learning_rate": 9.976801044672608e-06,
      "loss": 0.2308896780014038,
      "memory(GiB)": 27.73,
      "step": 15,
      "token_acc": 0.9208086725449933,
      "train_speed(iter/s)": 0.184963
    },
    {
      "epoch": 0.12241775057383321,
      "grad_norm": 0.7325994372367859,
      "learning_rate": 9.958782235357938e-06,
      "loss": 0.23883743286132814,
      "memory(GiB)": 27.73,
      "step": 20,
      "token_acc": 0.9143488349915764,
      "train_speed(iter/s)": 0.1933
    },
    {
      "epoch": 0.12241775057383321,
      "eval_loss": 0.2350648194551468,
      "eval_runtime": 3.0009,
      "eval_samples_per_second": 34.99,
      "eval_steps_per_second": 8.997,
      "eval_token_acc": 0.9214235491373249,
      "step": 20
    },
    {
      "epoch": 0.1530221882172915,
      "grad_norm": 0.6182907223701477,
      "learning_rate": 9.935647082149088e-06,
      "loss": 0.22287518978118898,
      "memory(GiB)": 27.73,
      "step": 25,
      "token_acc": 0.9215506715506715,
      "train_speed(iter/s)": 0.176091
    },
    {
      "epoch": 0.18362662586074982,
      "grad_norm": 0.6147263646125793,
      "learning_rate": 9.90741945530174e-06,
      "loss": 0.2169550657272339,
      "memory(GiB)": 27.73,
      "step": 30,
      "token_acc": 0.9242762338076332,
      "train_speed(iter/s)": 0.182484
    },
    {
      "epoch": 0.21423106350420812,
      "grad_norm": 0.5748075246810913,
      "learning_rate": 9.874128479354833e-06,
      "loss": 0.20252976417541504,
      "memory(GiB)": 27.73,
      "step": 35,
      "token_acc": 0.9266982811494906,
      "train_speed(iter/s)": 0.184908
    },
    {
      "epoch": 0.24483550114766642,
      "grad_norm": 0.5819372534751892,
      "learning_rate": 9.835808503080586e-06,
      "loss": 0.2103517770767212,
      "memory(GiB)": 27.73,
      "step": 40,
      "token_acc": 0.9242556390977443,
      "train_speed(iter/s)": 0.187883
    },
    {
      "epoch": 0.24483550114766642,
      "eval_loss": 0.21685469150543213,
      "eval_runtime": 2.9694,
      "eval_samples_per_second": 35.361,
      "eval_steps_per_second": 9.093,
      "eval_token_acc": 0.9267889015090054,
      "step": 40
    },
    {
      "epoch": 0.2754399387911247,
      "grad_norm": 0.5915679931640625,
      "learning_rate": 9.792499064044343e-06,
      "loss": 0.22011182308197022,
      "memory(GiB)": 27.73,
      "step": 45,
      "token_acc": 0.9259782645267656,
      "train_speed(iter/s)": 0.178276
    },
    {
      "epoch": 0.306044376434583,
      "grad_norm": 0.5741272568702698,
      "learning_rate": 9.744244847810716e-06,
      "loss": 0.20354759693145752,
      "memory(GiB)": 27.73,
      "step": 50,
      "token_acc": 0.9264593625325863,
      "train_speed(iter/s)": 0.180934
    },
    {
      "epoch": 0.3366488140780413,
      "grad_norm": 0.6794053316116333,
      "learning_rate": 9.691095641838168e-06,
      "loss": 0.20430679321289064,
      "memory(GiB)": 27.73,
      "step": 55,
      "token_acc": 0.9300412165115843,
      "train_speed(iter/s)": 0.184099
    },
    {
      "epoch": 0.36725325172149964,
      "grad_norm": 0.638066828250885,
      "learning_rate": 9.633106284109612e-06,
      "loss": 0.20622057914733888,
      "memory(GiB)": 27.73,
      "step": 60,
      "token_acc": 0.9269024685344354,
      "train_speed(iter/s)": 0.187435
    },
    {
      "epoch": 0.36725325172149964,
      "eval_loss": 0.20562399923801422,
      "eval_runtime": 2.9656,
      "eval_samples_per_second": 35.406,
      "eval_steps_per_second": 9.104,
      "eval_token_acc": 0.9300124398290875,
      "step": 60
    },
    {
      "epoch": 0.3978576893649579,
      "grad_norm": 0.6637494564056396,
      "learning_rate": 9.570336606551966e-06,
      "loss": 0.19611797332763672,
      "memory(GiB)": 27.73,
      "step": 65,
      "token_acc": 0.9295485869849984,
      "train_speed(iter/s)": 0.180966
    },
    {
      "epoch": 0.42846212700841624,
      "grad_norm": 0.6041725873947144,
      "learning_rate": 9.502851373303137e-06,
      "loss": 0.19453718662261962,
      "memory(GiB)": 27.73,
      "step": 70,
      "token_acc": 0.935506221719457,
      "train_speed(iter/s)": 0.182759
    },
    {
      "epoch": 0.4590665646518745,
      "grad_norm": 0.6496462225914001,
      "learning_rate": 9.43072021389003e-06,
      "loss": 0.19291472434997559,
      "memory(GiB)": 27.73,
      "step": 75,
      "token_acc": 0.9276501621556994,
      "train_speed(iter/s)": 0.184684
    },
    {
      "epoch": 0.48967100229533284,
      "grad_norm": 0.6091645359992981,
      "learning_rate": 9.354017551386599e-06,
      "loss": 0.19289860725402833,
      "memory(GiB)": 29.73,
      "step": 80,
      "token_acc": 0.9271297939943614,
      "train_speed(iter/s)": 0.186191
    },
    {
      "epoch": 0.48967100229533284,
      "eval_loss": 0.20189911127090454,
      "eval_runtime": 2.9558,
      "eval_samples_per_second": 35.523,
      "eval_steps_per_second": 9.134,
      "eval_token_acc": 0.9315160365622803,
      "step": 80
    },
    {
      "epoch": 0.5202754399387911,
      "grad_norm": 0.6070266962051392,
      "learning_rate": 9.272822525626047e-06,
      "loss": 0.18841828107833863,
      "memory(GiB)": 29.73,
      "step": 85,
      "token_acc": 0.9308790049419735,
      "train_speed(iter/s)": 0.181691
    },
    {
      "epoch": 0.5508798775822494,
      "grad_norm": 0.5664994120597839,
      "learning_rate": 9.187218911546363e-06,
      "loss": 0.18708581924438478,
      "memory(GiB)": 29.73,
      "step": 90,
      "token_acc": 0.9307408366791866,
      "train_speed(iter/s)": 0.183539
    },
    {
      "epoch": 0.5814843152257078,
      "grad_norm": 0.6189126372337341,
      "learning_rate": 9.09729503275351e-06,
      "loss": 0.1919234037399292,
      "memory(GiB)": 29.73,
      "step": 95,
      "token_acc": 0.9302339232778609,
      "train_speed(iter/s)": 0.18417
    },
    {
      "epoch": 0.612088752869166,
      "grad_norm": 0.5816466808319092,
      "learning_rate": 9.003143670391403e-06,
      "loss": 0.17953382730484008,
      "memory(GiB)": 29.73,
      "step": 100,
      "token_acc": 0.9341908338745658,
      "train_speed(iter/s)": 0.185218
    },
    {
      "epoch": 0.612088752869166,
      "eval_loss": 0.19569453597068787,
      "eval_runtime": 2.9576,
      "eval_samples_per_second": 35.502,
      "eval_steps_per_second": 9.129,
      "eval_token_acc": 0.9332143436638001,
      "step": 100
    },
    {
      "epoch": 0.6426931905126243,
      "grad_norm": 0.6038978695869446,
      "learning_rate": 8.904861967412702e-06,
      "loss": 0.18669098615646362,
      "memory(GiB)": 29.73,
      "step": 105,
      "token_acc": 0.9360083136077594,
      "train_speed(iter/s)": 0.182365
    },
    {
      "epoch": 0.6732976281560826,
      "grad_norm": 0.6839295029640198,
      "learning_rate": 8.802551328349222e-06,
      "loss": 0.18886933326721192,
      "memory(GiB)": 29.73,
      "step": 110,
      "token_acc": 0.9355897003717136,
      "train_speed(iter/s)": 0.183447
    },
    {
      "epoch": 0.703902065799541,
      "grad_norm": 0.5784356594085693,
      "learning_rate": 8.696317314685342e-06,
      "loss": 0.1860347270965576,
      "memory(GiB)": 29.73,
      "step": 115,
      "token_acc": 0.934266941773856,
      "train_speed(iter/s)": 0.184623
    },
    {
      "epoch": 0.7345065034429993,
      "grad_norm": 0.6492018103599548,
      "learning_rate": 8.586269535942386e-06,
      "loss": 0.191730272769928,
      "memory(GiB)": 29.73,
      "step": 120,
      "token_acc": 0.9330231994302426,
      "train_speed(iter/s)": 0.185813
    },
    {
      "epoch": 0.7345065034429993,
      "eval_loss": 0.19317017495632172,
      "eval_runtime": 2.9604,
      "eval_samples_per_second": 35.468,
      "eval_steps_per_second": 9.12,
      "eval_token_acc": 0.9342636162042296,
      "step": 120
    },
    {
      "epoch": 0.7651109410864575,
      "grad_norm": 0.5753573775291443,
      "learning_rate": 8.472521536586336e-06,
      "loss": 0.18596272468566893,
      "memory(GiB)": 29.73,
      "step": 125,
      "token_acc": 0.9319849005577779,
      "train_speed(iter/s)": 0.182874
    },
    {
      "epoch": 0.7957153787299158,
      "grad_norm": 0.586589515209198,
      "learning_rate": 8.355190678875577e-06,
      "loss": 0.18235174417495728,
      "memory(GiB)": 29.73,
      "step": 130,
      "token_acc": 0.9364593742793451,
      "train_speed(iter/s)": 0.183919
    },
    {
      "epoch": 0.8263198163733741,
      "grad_norm": 0.6312227845191956,
      "learning_rate": 8.234398021769541e-06,
      "loss": 0.18675460815429687,
      "memory(GiB)": 29.73,
      "step": 135,
      "token_acc": 0.9301861157944705,
      "train_speed(iter/s)": 0.184691
    },
    {
      "epoch": 0.8569242540168325,
      "grad_norm": 0.5985466837882996,
      "learning_rate": 8.110268196023179e-06,
      "loss": 0.18646229505538942,
      "memory(GiB)": 29.73,
      "step": 140,
      "token_acc": 0.9337858101175913,
      "train_speed(iter/s)": 0.185698
    },
    {
      "epoch": 0.8569242540168325,
      "eval_loss": 0.18890073895454407,
      "eval_runtime": 2.9675,
      "eval_samples_per_second": 35.383,
      "eval_steps_per_second": 9.098,
      "eval_token_acc": 0.9360809129752826,
      "step": 140
    },
    {
      "epoch": 0.8875286916602907,
      "grad_norm": 0.6701765060424805,
      "learning_rate": 7.982929275596164e-06,
      "loss": 0.18306300640106202,
      "memory(GiB)": 29.73,
      "step": 145,
      "token_acc": 0.9374918368235448,
      "train_speed(iter/s)": 0.183024
    },
    {
      "epoch": 0.918133129303749,
      "grad_norm": 0.6174297332763672,
      "learning_rate": 7.85251264550948e-06,
      "loss": 0.18363630771636963,
      "memory(GiB)": 29.73,
      "step": 150,
      "token_acc": 0.9375606928038264,
      "train_speed(iter/s)": 0.183651
    },
    {
      "epoch": 0.9487375669472073,
      "grad_norm": 0.6339284181594849,
      "learning_rate": 7.719152866285722e-06,
      "loss": 0.17851275205612183,
      "memory(GiB)": 29.73,
      "step": 155,
      "token_acc": 0.9321391193556705,
      "train_speed(iter/s)": 0.184494
    },
    {
      "epoch": 0.9793420045906657,
      "grad_norm": 0.589568555355072,
      "learning_rate": 7.5829875351130224e-06,
      "loss": 0.1714911699295044,
      "memory(GiB)": 29.73,
      "step": 160,
      "token_acc": 0.936993310933695,
      "train_speed(iter/s)": 0.185314
    },
    {
      "epoch": 0.9793420045906657,
      "eval_loss": 0.18555624783039093,
      "eval_runtime": 2.9746,
      "eval_samples_per_second": 35.299,
      "eval_steps_per_second": 9.077,
      "eval_token_acc": 0.9358970198496404,
      "step": 160
    },
    {
      "epoch": 1.0122417750573833,
      "grad_norm": 0.5571500062942505,
      "learning_rate": 7.44415714387582e-06,
      "loss": 0.1988983631134033,
      "memory(GiB)": 29.73,
      "step": 165,
      "token_acc": 0.9428557149282822,
      "train_speed(iter/s)": 0.183086
    },
    {
      "epoch": 1.0428462127008415,
      "grad_norm": 0.668183445930481,
      "learning_rate": 7.302804934198937e-06,
      "loss": 0.14512217044830322,
      "memory(GiB)": 29.73,
      "step": 170,
      "token_acc": 0.9455400702401195,
      "train_speed(iter/s)": 0.183717
    },
    {
      "epoch": 1.0734506503443,
      "grad_norm": 0.6544790267944336,
      "learning_rate": 7.159076749654559e-06,
      "loss": 0.1422812223434448,
      "memory(GiB)": 29.73,
      "step": 175,
      "token_acc": 0.9467687539877859,
      "train_speed(iter/s)": 0.184527
    },
    {
      "epoch": 1.1040550879877582,
      "grad_norm": 0.6180390119552612,
      "learning_rate": 7.013120885284599e-06,
      "loss": 0.14142969846725464,
      "memory(GiB)": 29.73,
      "step": 180,
      "token_acc": 0.9519427697810598,
      "train_speed(iter/s)": 0.185066
    },
    {
      "epoch": 1.1040550879877582,
      "eval_loss": 0.1868736445903778,
      "eval_runtime": 2.9588,
      "eval_samples_per_second": 35.487,
      "eval_steps_per_second": 9.125,
      "eval_token_acc": 0.9361025474606522,
      "step": 180
    },
    {
      "epoch": 1.1346595256312164,
      "grad_norm": 0.5758160352706909,
      "learning_rate": 6.86508793459368e-06,
      "loss": 0.13951488733291625,
      "memory(GiB)": 29.73,
      "step": 185,
      "token_acc": 0.9455232473644789,
      "train_speed(iter/s)": 0.182916
    },
    {
      "epoch": 1.1652639632746749,
      "grad_norm": 0.6193047761917114,
      "learning_rate": 6.715130634170636e-06,
      "loss": 0.146294903755188,
      "memory(GiB)": 29.73,
      "step": 190,
      "token_acc": 0.9483162813805095,
      "train_speed(iter/s)": 0.183613
    },
    {
      "epoch": 1.195868400918133,
      "grad_norm": 0.5518524050712585,
      "learning_rate": 6.563403706098833e-06,
      "loss": 0.14221264123916627,
      "memory(GiB)": 29.73,
      "step": 195,
      "token_acc": 0.948411852483479,
      "train_speed(iter/s)": 0.184409
    },
    {
      "epoch": 1.2264728385615915,
      "grad_norm": 0.5622280836105347,
      "learning_rate": 6.410063698317901e-06,
      "loss": 0.14119510650634765,
      "memory(GiB)": 29.73,
      "step": 200,
      "token_acc": 0.949674972130127,
      "train_speed(iter/s)": 0.184756
    },
    {
      "epoch": 1.2264728385615915,
      "eval_loss": 0.18399108946323395,
      "eval_runtime": 2.9642,
      "eval_samples_per_second": 35.423,
      "eval_steps_per_second": 9.109,
      "eval_token_acc": 0.9372816269132997,
      "step": 200
    },
    {
      "epoch": 1.2570772762050497,
      "grad_norm": 0.5925666689872742,
      "learning_rate": 6.255268823101604e-06,
      "loss": 0.13880105018615724,
      "memory(GiB)": 29.73,
      "step": 205,
      "token_acc": 0.9477773473991206,
      "train_speed(iter/s)": 0.183189
    },
    {
      "epoch": 1.287681713848508,
      "grad_norm": 0.6059976816177368,
      "learning_rate": 6.099178793818479e-06,
      "loss": 0.14395406246185302,
      "memory(GiB)": 29.73,
      "step": 210,
      "token_acc": 0.950825693689833,
      "train_speed(iter/s)": 0.183549
    },
    {
      "epoch": 1.3182861514919664,
      "grad_norm": 0.563386857509613,
      "learning_rate": 5.941954660143703e-06,
      "loss": 0.13854990005493165,
      "memory(GiB)": 29.73,
      "step": 215,
      "token_acc": 0.9485802112183096,
      "train_speed(iter/s)": 0.184231
    },
    {
      "epoch": 1.3488905891354246,
      "grad_norm": 0.6268328428268433,
      "learning_rate": 5.783758641892172e-06,
      "loss": 0.14123060703277587,
      "memory(GiB)": 29.73,
      "step": 220,
      "token_acc": 0.9484895429899303,
      "train_speed(iter/s)": 0.184974
    },
    {
      "epoch": 1.3488905891354246,
      "eval_loss": 0.18251946568489075,
      "eval_runtime": 2.9684,
      "eval_samples_per_second": 35.372,
      "eval_steps_per_second": 9.096,
      "eval_token_acc": 0.9383849856671534,
      "step": 220
    },
    {
      "epoch": 1.379495026778883,
      "grad_norm": 0.5578710436820984,
      "learning_rate": 5.624753961644281e-06,
      "loss": 0.14209306240081787,
      "memory(GiB)": 29.73,
      "step": 225,
      "token_acc": 0.9469471634849386,
      "train_speed(iter/s)": 0.183543
    },
    {
      "epoch": 1.4100994644223412,
      "grad_norm": 0.6176152229309082,
      "learning_rate": 5.4651046763370615e-06,
      "loss": 0.1443139672279358,
      "memory(GiB)": 29.73,
      "step": 230,
      "token_acc": 0.9480665287334199,
      "train_speed(iter/s)": 0.184004
    },
    {
      "epoch": 1.4407039020657995,
      "grad_norm": 0.5799533724784851,
      "learning_rate": 5.304975507994453e-06,
      "loss": 0.13565800189971924,
      "memory(GiB)": 29.73,
      "step": 235,
      "token_acc": 0.9535484258411278,
      "train_speed(iter/s)": 0.184791
    },
    {
      "epoch": 1.471308339709258,
      "grad_norm": 0.6195805072784424,
      "learning_rate": 5.144531673771364e-06,
      "loss": 0.13422565460205077,
      "memory(GiB)": 29.73,
      "step": 240,
      "token_acc": 0.9504853626843996,
      "train_speed(iter/s)": 0.185206
    },
    {
      "epoch": 1.471308339709258,
      "eval_loss": 0.1826462596654892,
      "eval_runtime": 2.974,
      "eval_samples_per_second": 35.306,
      "eval_steps_per_second": 9.079,
      "eval_token_acc": 0.9378441235329115,
      "step": 240
    },
    {
      "epoch": 1.501912777352716,
      "grad_norm": 0.5705444812774658,
      "learning_rate": 4.983938715486858e-06,
      "loss": 0.13150216341018678,
      "memory(GiB)": 29.73,
      "step": 245,
      "token_acc": 0.9486215675776085,
      "train_speed(iter/s)": 0.183983
    },
    {
      "epoch": 1.5325172149961745,
      "grad_norm": 0.6076279282569885,
      "learning_rate": 4.82336232882237e-06,
      "loss": 0.1356377363204956,
      "memory(GiB)": 29.73,
      "step": 250,
      "token_acc": 0.9495890326460746,
      "train_speed(iter/s)": 0.184361
    },
    {
      "epoch": 1.5631216526396328,
      "grad_norm": 0.606853187084198,
      "learning_rate": 4.662968192361161e-06,
      "loss": 0.1377635955810547,
      "memory(GiB)": 29.73,
      "step": 255,
      "token_acc": 0.9498974208675264,
      "train_speed(iter/s)": 0.184863
    },
    {
      "epoch": 1.593726090283091,
      "grad_norm": 0.5685162544250488,
      "learning_rate": 4.502921796645424e-06,
      "loss": 0.13207011222839354,
      "memory(GiB)": 29.73,
      "step": 260,
      "token_acc": 0.9555538648813147,
      "train_speed(iter/s)": 0.185408
    },
    {
      "epoch": 1.593726090283091,
      "eval_loss": 0.18062403798103333,
      "eval_runtime": 2.9631,
      "eval_samples_per_second": 35.435,
      "eval_steps_per_second": 9.112,
      "eval_token_acc": 0.9392503650819406,
      "step": 260
    },
    {
      "epoch": 1.6243305279265492,
      "grad_norm": 0.6115860342979431,
      "learning_rate": 4.3433882734274e-06,
      "loss": 0.13854453563690186,
      "memory(GiB)": 29.73,
      "step": 265,
      "token_acc": 0.950067791336045,
      "train_speed(iter/s)": 0.184317
    },
    {
      "epoch": 1.6549349655700076,
      "grad_norm": 0.594256579875946,
      "learning_rate": 4.184532225290687e-06,
      "loss": 0.14273109436035156,
      "memory(GiB)": 29.73,
      "step": 270,
      "token_acc": 0.949110499003211,
      "train_speed(iter/s)": 0.184908
    },
    {
      "epoch": 1.685539403213466,
      "grad_norm": 0.6124013662338257,
      "learning_rate": 4.026517555817527e-06,
      "loss": 0.14174976348876953,
      "memory(GiB)": 29.73,
      "step": 275,
      "token_acc": 0.9440209335390386,
      "train_speed(iter/s)": 0.185319
    },
    {
      "epoch": 1.7161438408569243,
      "grad_norm": 0.5577505230903625,
      "learning_rate": 3.869507300477311e-06,
      "loss": 0.14029152393341066,
      "memory(GiB)": 29.73,
      "step": 280,
      "token_acc": 0.949042973113875,
      "train_speed(iter/s)": 0.185887
    },
    {
      "epoch": 1.7161438408569243,
      "eval_loss": 0.17867250740528107,
      "eval_runtime": 2.9631,
      "eval_samples_per_second": 35.436,
      "eval_steps_per_second": 9.112,
      "eval_token_acc": 0.9399534858564552,
      "step": 280
    },
    {
      "epoch": 1.7467482785003825,
      "grad_norm": 0.5901487469673157,
      "learning_rate": 3.7136634584107787e-06,
      "loss": 0.13561428785324098,
      "memory(GiB)": 32.03,
      "step": 285,
      "token_acc": 0.9532760191419626,
      "train_speed(iter/s)": 0.184838
    },
    {
      "epoch": 1.7773527161438407,
      "grad_norm": 0.5688520073890686,
      "learning_rate": 3.5591468252834654e-06,
      "loss": 0.13801207542419433,
      "memory(GiB)": 32.03,
      "step": 290,
      "token_acc": 0.9521514789965446,
      "train_speed(iter/s)": 0.185292
    },
    {
      "epoch": 1.8079571537872992,
      "grad_norm": 0.5571552515029907,
      "learning_rate": 3.4061168273808896e-06,
      "loss": 0.13579220771789552,
      "memory(GiB)": 32.03,
      "step": 295,
      "token_acc": 0.9481830960882842,
      "train_speed(iter/s)": 0.18574
    },
    {
      "epoch": 1.8385615914307576,
      "grad_norm": 0.5458003282546997,
      "learning_rate": 3.254731357116597e-06,
      "loss": 0.13743789196014405,
      "memory(GiB)": 32.03,
      "step": 300,
      "token_acc": 0.9545541917236455,
      "train_speed(iter/s)": 0.186123
    },
    {
      "epoch": 1.8385615914307576,
      "eval_loss": 0.17722059786319733,
      "eval_runtime": 2.9691,
      "eval_samples_per_second": 35.364,
      "eval_steps_per_second": 9.094,
      "eval_token_acc": 0.9400183893125642,
      "step": 300
    },
    {
      "epoch": 1.8691660290742158,
      "grad_norm": 0.5650319457054138,
      "learning_rate": 3.105146610122839e-06,
      "loss": 0.13116140365600587,
      "memory(GiB)": 32.03,
      "step": 305,
      "token_acc": 0.9474181254005194,
      "train_speed(iter/s)": 0.185096
    },
    {
      "epoch": 1.899770466717674,
      "grad_norm": 0.5377411842346191,
      "learning_rate": 2.95751692409194e-06,
      "loss": 0.141234827041626,
      "memory(GiB)": 32.03,
      "step": 310,
      "token_acc": 0.9441545925827463,
      "train_speed(iter/s)": 0.185517
    },
    {
      "epoch": 1.9303749043611322,
      "grad_norm": 0.5164922475814819,
      "learning_rate": 2.8119946195346375e-06,
      "loss": 0.13225994110107422,
      "memory(GiB)": 32.03,
      "step": 315,
      "token_acc": 0.9511941848390446,
      "train_speed(iter/s)": 0.185993
    },
    {
      "epoch": 1.9609793420045907,
      "grad_norm": 0.5725815892219543,
      "learning_rate": 2.6687298426196974e-06,
      "loss": 0.1376440167427063,
      "memory(GiB)": 32.03,
      "step": 320,
      "token_acc": 0.9511019926015729,
      "train_speed(iter/s)": 0.186367
    },
    {
      "epoch": 1.9609793420045907,
      "eval_loss": 0.17490148544311523,
      "eval_runtime": 2.9672,
      "eval_samples_per_second": 35.386,
      "eval_steps_per_second": 9.099,
      "eval_token_acc": 0.9408945859700363,
      "step": 320
    },
    {
      "epoch": 1.9915837796480491,
      "grad_norm": 0.5404472947120667,
      "learning_rate": 2.527870410256966e-06,
      "loss": 0.1282888650894165,
      "memory(GiB)": 32.03,
      "step": 325,
      "token_acc": 0.9526283186802439,
      "train_speed(iter/s)": 0.185372
    },
    {
      "epoch": 2.0244835501147667,
      "grad_norm": 0.5364894270896912,
      "learning_rate": 2.389561657583681e-06,
      "loss": 0.14493446350097655,
      "memory(GiB)": 32.03,
      "step": 330,
      "token_acc": 0.9557499000952937,
      "train_speed(iter/s)": 0.185654
    },
    {
      "epoch": 2.055087987758225,
      "grad_norm": 0.5487475395202637,
      "learning_rate": 2.253946288011419e-06,
      "loss": 0.10785077810287476,
      "memory(GiB)": 32.03,
      "step": 335,
      "token_acc": 0.9601302750594264,
      "train_speed(iter/s)": 0.186057
    },
    {
      "epoch": 2.085692425401683,
      "grad_norm": 0.5607486963272095,
      "learning_rate": 2.121164225988387e-06,
      "loss": 0.10882875919342042,
      "memory(GiB)": 32.03,
      "step": 340,
      "token_acc": 0.9595018647214418,
      "train_speed(iter/s)": 0.186452
    },
    {
      "epoch": 2.085692425401683,
      "eval_loss": 0.18441390991210938,
      "eval_runtime": 2.9649,
      "eval_samples_per_second": 35.415,
      "eval_steps_per_second": 9.107,
      "eval_token_acc": 0.9401265617394127,
      "step": 340
    },
    {
      "epoch": 2.1162968630451418,
      "grad_norm": 0.539606511592865,
      "learning_rate": 1.9913524726289784e-06,
      "loss": 0.1085669994354248,
      "memory(GiB)": 32.03,
      "step": 345,
      "token_acc": 0.9560936203659229,
      "train_speed(iter/s)": 0.185455
    },
    {
      "epoch": 2.1469013006886,
      "grad_norm": 0.5162772536277771,
      "learning_rate": 1.8646449643595565e-06,
      "loss": 0.10514116287231445,
      "memory(GiB)": 32.03,
      "step": 350,
      "token_acc": 0.9644231329796537,
      "train_speed(iter/s)": 0.185643
    },
    {
      "epoch": 2.177505738332058,
      "grad_norm": 0.5563685894012451,
      "learning_rate": 1.7411724347262826e-06,
      "loss": 0.11478805541992188,
      "memory(GiB)": 32.03,
      "step": 355,
      "token_acc": 0.9576824543864034,
      "train_speed(iter/s)": 0.186082
    },
    {
      "epoch": 2.2081101759755164,
      "grad_norm": 0.5719069242477417,
      "learning_rate": 1.621062279507617e-06,
      "loss": 0.10535905361175538,
      "memory(GiB)": 32.03,
      "step": 360,
      "token_acc": 0.9599832227797368,
      "train_speed(iter/s)": 0.1864
    },
    {
      "epoch": 2.2081101759755164,
      "eval_loss": 0.1818896383047104,
      "eval_runtime": 2.968,
      "eval_samples_per_second": 35.378,
      "eval_steps_per_second": 9.097,
      "eval_token_acc": 0.9408621342419817,
      "step": 360
    },
    {
      "epoch": 2.2387146136189746,
      "grad_norm": 0.5362414121627808,
      "learning_rate": 1.5044384252706312e-06,
      "loss": 0.1083768367767334,
      "memory(GiB)": 32.03,
      "step": 365,
      "token_acc": 0.9576676318628286,
      "train_speed(iter/s)": 0.185562
    },
    {
      "epoch": 2.269319051262433,
      "grad_norm": 0.5509895086288452,
      "learning_rate": 1.3914212015067653e-06,
      "loss": 0.10523394346237183,
      "memory(GiB)": 32.03,
      "step": 370,
      "token_acc": 0.9611730489578284,
      "train_speed(iter/s)": 0.185837
    },
    {
      "epoch": 2.2999234889058915,
      "grad_norm": 0.5473169684410095,
      "learning_rate": 1.2821272164789544e-06,
      "loss": 0.10544384717941284,
      "memory(GiB)": 32.03,
      "step": 375,
      "token_acc": 0.9595612848087743,
      "train_speed(iter/s)": 0.186197
    },
    {
      "epoch": 2.3305279265493497,
      "grad_norm": 0.5506060719490051,
      "learning_rate": 1.1766692369082255e-06,
      "loss": 0.10901916027069092,
      "memory(GiB)": 32.03,
      "step": 380,
      "token_acc": 0.9595015576323987,
      "train_speed(iter/s)": 0.18649
    },
    {
      "epoch": 2.3305279265493497,
      "eval_loss": 0.18156129121780396,
      "eval_runtime": 2.9586,
      "eval_samples_per_second": 35.49,
      "eval_steps_per_second": 9.126,
      "eval_token_acc": 0.9409054032127211,
      "step": 380
    },
    {
      "epoch": 2.361132364192808,
      "grad_norm": 0.5408657193183899,
      "learning_rate": 1.0751560716238968e-06,
      "loss": 0.10397273302078247,
      "memory(GiB)": 32.03,
      "step": 385,
      "token_acc": 0.9595956316091394,
      "train_speed(iter/s)": 0.185677
    },
    {
      "epoch": 2.391736801836266,
      "grad_norm": 0.565687358379364,
      "learning_rate": 9.776924592974257e-07,
      "loss": 0.10786858797073365,
      "memory(GiB)": 32.03,
      "step": 390,
      "token_acc": 0.9580994261477046,
      "train_speed(iter/s)": 0.185992
    },
    {
      "epoch": 2.4223412394797244,
      "grad_norm": 0.5598291754722595,
      "learning_rate": 8.843789603757446e-07,
      "loss": 0.10539779663085938,
      "memory(GiB)": 32.03,
      "step": 395,
      "token_acc": 0.9650639142961036,
      "train_speed(iter/s)": 0.186263
    },
    {
      "epoch": 2.452945677123183,
      "grad_norm": 0.5249931216239929,
      "learning_rate": 7.953118533255821e-07,
      "loss": 0.10175876617431641,
      "memory(GiB)": 32.03,
      "step": 400,
      "token_acc": 0.9630585035539542,
      "train_speed(iter/s)": 0.186474
    },
    {
      "epoch": 2.452945677123183,
      "eval_loss": 0.18106642365455627,
      "eval_runtime": 2.9642,
      "eval_samples_per_second": 35.423,
      "eval_steps_per_second": 9.109,
      "eval_token_acc": 0.9406998756017091,
      "step": 400
    },
    {
      "epoch": 2.4835501147666412,
      "grad_norm": 0.5907252430915833,
      "learning_rate": 7.105830352958143e-07,
      "loss": 0.10489962100982667,
      "memory(GiB)": 32.03,
      "step": 405,
      "token_acc": 0.9566693129911521,
      "train_speed(iter/s)": 0.185755
    },
    {
      "epoch": 2.5141545524100994,
      "grad_norm": 0.5359929203987122,
      "learning_rate": 6.302799273003546e-07,
      "loss": 0.10674171447753907,
      "memory(GiB)": 32.03,
      "step": 410,
      "token_acc": 0.9589793334758233,
      "train_speed(iter/s)": 0.186023
    },
    {
      "epoch": 2.5447589900535577,
      "grad_norm": 0.5575762987136841,
      "learning_rate": 5.544853840193981e-07,
      "loss": 0.10684401988983154,
      "memory(GiB)": 32.03,
      "step": 415,
      "token_acc": 0.961750384418247,
      "train_speed(iter/s)": 0.186382
    },
    {
      "epoch": 2.575363427697016,
      "grad_norm": 0.5757756233215332,
      "learning_rate": 4.832776083120983e-07,
      "loss": 0.10789649486541748,
      "memory(GiB)": 32.03,
      "step": 420,
      "token_acc": 0.9595656339827531,
      "train_speed(iter/s)": 0.186545
    },
    {
      "epoch": 2.575363427697016,
      "eval_loss": 0.18043935298919678,
      "eval_runtime": 2.9657,
      "eval_samples_per_second": 35.404,
      "eval_steps_per_second": 9.104,
      "eval_token_acc": 0.9409162204554059,
      "step": 420
    },
    {
      "epoch": 2.6059678653404745,
      "grad_norm": 0.5543663501739502,
      "learning_rate": 4.167300705288718e-07,
      "loss": 0.10247746706008912,
      "memory(GiB)": 32.03,
      "step": 425,
      "token_acc": 0.9594139385812865,
      "train_speed(iter/s)": 0.185739
    },
    {
      "epoch": 2.6365723029839327,
      "grad_norm": 0.545873761177063,
      "learning_rate": 3.5491143270657445e-07,
      "loss": 0.1048995852470398,
      "memory(GiB)": 32.03,
      "step": 430,
      "token_acc": 0.9606733925249195,
      "train_speed(iter/s)": 0.185979
    },
    {
      "epoch": 2.667176740627391,
      "grad_norm": 0.5157749056816101,
      "learning_rate": 2.9788547772478416e-07,
      "loss": 0.10134179592132568,
      "memory(GiB)": 32.03,
      "step": 435,
      "token_acc": 0.9625057155921354,
      "train_speed(iter/s)": 0.186374
    },
    {
      "epoch": 2.697781178270849,
      "grad_norm": 0.5268109440803528,
      "learning_rate": 2.457110434962645e-07,
      "loss": 0.10287988185882568,
      "memory(GiB)": 32.03,
      "step": 440,
      "token_acc": 0.9619926437374976,
      "train_speed(iter/s)": 0.186687
    },
    {
      "epoch": 2.697781178270849,
      "eval_loss": 0.18037647008895874,
      "eval_runtime": 2.9654,
      "eval_samples_per_second": 35.409,
      "eval_steps_per_second": 9.105,
      "eval_token_acc": 0.941046027367624,
      "step": 440
    },
    {
      "epoch": 2.7283856159143074,
      "grad_norm": 0.5612174272537231,
      "learning_rate": 1.984419622595224e-07,
      "loss": 0.10738935470581054,
      "memory(GiB)": 32.03,
      "step": 445,
      "token_acc": 0.9575748052706009,
      "train_speed(iter/s)": 0.185904
    },
    {
      "epoch": 2.758990053557766,
      "grad_norm": 0.5592919588088989,
      "learning_rate": 1.561270050360897e-07,
      "loss": 0.1064642071723938,
      "memory(GiB)": 32.03,
      "step": 450,
      "token_acc": 0.9620171040990858,
      "train_speed(iter/s)": 0.186133
    },
    {
      "epoch": 2.7895944912012243,
      "grad_norm": 0.5563650131225586,
      "learning_rate": 1.1880983130983626e-07,
      "loss": 0.11139054298400879,
      "memory(GiB)": 32.03,
      "step": 455,
      "token_acc": 0.9608461825416563,
      "train_speed(iter/s)": 0.186386
    },
    {
      "epoch": 2.8201989288446825,
      "grad_norm": 0.5721088647842407,
      "learning_rate": 8.652894398024137e-08,
      "loss": 0.1031692385673523,
      "memory(GiB)": 32.03,
      "step": 460,
      "token_acc": 0.9663138518177634,
      "train_speed(iter/s)": 0.186591
    },
    {
      "epoch": 2.8201989288446825,
      "eval_loss": 0.1801983118057251,
      "eval_runtime": 2.9604,
      "eval_samples_per_second": 35.468,
      "eval_steps_per_second": 9.12,
      "eval_token_acc": 0.941110930823733,
      "step": 460
    },
    {
      "epoch": 2.8508033664881407,
      "grad_norm": 0.5719510316848755,
      "learning_rate": 5.9317649636088656e-08,
      "loss": 0.10809749364852905,
      "memory(GiB)": 32.03,
      "step": 465,
      "token_acc": 0.9583455558814902,
      "train_speed(iter/s)": 0.185906
    },
    {
      "epoch": 2.881407804131599,
      "grad_norm": 0.5535134673118591,
      "learning_rate": 3.720402419058966e-08,
      "loss": 0.10542930364608764,
      "memory(GiB)": 32.03,
      "step": 470,
      "token_acc": 0.9625262091383702,
      "train_speed(iter/s)": 0.18623
    },
    {
      "epoch": 2.9120122417750576,
      "grad_norm": 0.5565011501312256,
      "learning_rate": 2.0210883913376334e-08,
      "loss": 0.10625803470611572,
      "memory(GiB)": 32.03,
      "step": 475,
      "token_acc": 0.9594989624827711,
      "train_speed(iter/s)": 0.186434
    },
    {
      "epoch": 2.942616679418516,
      "grad_norm": 0.5512571334838867,
      "learning_rate": 8.35576188926046e-09,
      "loss": 0.098574298620224,
      "memory(GiB)": 32.03,
      "step": 480,
      "token_acc": 0.9657953849295488,
      "train_speed(iter/s)": 0.186661
    },
    {
      "epoch": 2.942616679418516,
      "eval_loss": 0.18013475835323334,
      "eval_runtime": 2.956,
      "eval_samples_per_second": 35.52,
      "eval_steps_per_second": 9.134,
      "eval_token_acc": 0.9411433825517875,
      "step": 480
    },
    {
      "epoch": 2.973221117061974,
      "grad_norm": 0.5630972385406494,
      "learning_rate": 1.6508899280515134e-09,
      "loss": 0.10670101642608643,
      "memory(GiB)": 32.03,
      "step": 485,
      "token_acc": 0.9557970508071842,
      "train_speed(iter/s)": 0.186012
    },
    {
      "epoch": 2.9977046671767407,
      "eval_loss": 0.18032938241958618,
      "eval_runtime": 2.9668,
      "eval_samples_per_second": 35.392,
      "eval_steps_per_second": 9.101,
      "eval_token_acc": 0.9413380929201146,
      "step": 489
    }
  ],
  "logging_steps": 5,
  "max_steps": 489,
  "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": 7.018693783621468e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}