{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9997336884154461,
  "eval_steps": 500,
  "global_step": 1877,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0005326231691078562,
      "grad_norm": 0.1129828691482544,
      "learning_rate": 1.0638297872340427e-06,
      "loss": 0.3516,
      "mean_token_accuracy": 0.9021367430686951,
      "step": 1
    },
    {
      "epoch": 0.002663115845539281,
      "grad_norm": 0.07912027090787888,
      "learning_rate": 5.319148936170213e-06,
      "loss": 0.2398,
      "mean_token_accuracy": 0.9323107227683067,
      "step": 5
    },
    {
      "epoch": 0.005326231691078562,
      "grad_norm": 0.07962298393249512,
      "learning_rate": 1.0638297872340426e-05,
      "loss": 0.2164,
      "mean_token_accuracy": 0.9409767866134644,
      "step": 10
    },
    {
      "epoch": 0.007989347536617843,
      "grad_norm": 0.07320517301559448,
      "learning_rate": 1.595744680851064e-05,
      "loss": 0.2285,
      "mean_token_accuracy": 0.9375457763671875,
      "step": 15
    },
    {
      "epoch": 0.010652463382157125,
      "grad_norm": 0.08164672553539276,
      "learning_rate": 2.1276595744680852e-05,
      "loss": 0.2276,
      "mean_token_accuracy": 0.9369352698326111,
      "step": 20
    },
    {
      "epoch": 0.013315579227696404,
      "grad_norm": 0.08709942549467087,
      "learning_rate": 2.6595744680851064e-05,
      "loss": 0.281,
      "mean_token_accuracy": 0.9194993853569031,
      "step": 25
    },
    {
      "epoch": 0.015978695073235686,
      "grad_norm": 0.10631067305803299,
      "learning_rate": 3.191489361702128e-05,
      "loss": 0.2564,
      "mean_token_accuracy": 0.9267643451690674,
      "step": 30
    },
    {
      "epoch": 0.018641810918774968,
      "grad_norm": 0.07647579163312912,
      "learning_rate": 3.723404255319149e-05,
      "loss": 0.2082,
      "mean_token_accuracy": 0.9413186728954315,
      "step": 35
    },
    {
      "epoch": 0.02130492676431425,
      "grad_norm": 0.07922424376010895,
      "learning_rate": 4.2553191489361704e-05,
      "loss": 0.2191,
      "mean_token_accuracy": 0.9362881481647491,
      "step": 40
    },
    {
      "epoch": 0.023968042609853527,
      "grad_norm": 0.0923653244972229,
      "learning_rate": 4.787234042553192e-05,
      "loss": 0.2286,
      "mean_token_accuracy": 0.9335653126239777,
      "step": 45
    },
    {
      "epoch": 0.02663115845539281,
      "grad_norm": 0.07971777766942978,
      "learning_rate": 5.319148936170213e-05,
      "loss": 0.1888,
      "mean_token_accuracy": 0.9448961973190307,
      "step": 50
    },
    {
      "epoch": 0.02929427430093209,
      "grad_norm": 0.09142950922250748,
      "learning_rate": 5.851063829787234e-05,
      "loss": 0.1384,
      "mean_token_accuracy": 0.9579731285572052,
      "step": 55
    },
    {
      "epoch": 0.03195739014647137,
      "grad_norm": 0.13386933505535126,
      "learning_rate": 6.382978723404256e-05,
      "loss": 0.147,
      "mean_token_accuracy": 0.95494504570961,
      "step": 60
    },
    {
      "epoch": 0.03462050599201065,
      "grad_norm": 0.12172385305166245,
      "learning_rate": 6.914893617021277e-05,
      "loss": 0.1365,
      "mean_token_accuracy": 0.9542857050895691,
      "step": 65
    },
    {
      "epoch": 0.037283621837549935,
      "grad_norm": 0.08828116208314896,
      "learning_rate": 7.446808510638298e-05,
      "loss": 0.1243,
      "mean_token_accuracy": 0.9553601861000061,
      "step": 70
    },
    {
      "epoch": 0.03994673768308921,
      "grad_norm": 0.08476319909095764,
      "learning_rate": 7.978723404255319e-05,
      "loss": 0.0951,
      "mean_token_accuracy": 0.9662515103816987,
      "step": 75
    },
    {
      "epoch": 0.0426098535286285,
      "grad_norm": 0.11043168604373932,
      "learning_rate": 8.510638297872341e-05,
      "loss": 0.1085,
      "mean_token_accuracy": 0.9597313702106476,
      "step": 80
    },
    {
      "epoch": 0.045272969374167776,
      "grad_norm": 0.08523018658161163,
      "learning_rate": 9.042553191489363e-05,
      "loss": 0.0903,
      "mean_token_accuracy": 0.965738695859909,
      "step": 85
    },
    {
      "epoch": 0.047936085219707054,
      "grad_norm": 0.09243249148130417,
      "learning_rate": 9.574468085106384e-05,
      "loss": 0.0881,
      "mean_token_accuracy": 0.9664346635341644,
      "step": 90
    },
    {
      "epoch": 0.05059920106524634,
      "grad_norm": 0.12321251630783081,
      "learning_rate": 0.00010106382978723406,
      "loss": 0.0976,
      "mean_token_accuracy": 0.9625030398368836,
      "step": 95
    },
    {
      "epoch": 0.05326231691078562,
      "grad_norm": 0.08379948139190674,
      "learning_rate": 0.00010638297872340425,
      "loss": 0.0707,
      "mean_token_accuracy": 0.9735897183418274,
      "step": 100
    },
    {
      "epoch": 0.0559254327563249,
      "grad_norm": 0.10740843415260315,
      "learning_rate": 0.00011170212765957446,
      "loss": 0.0822,
      "mean_token_accuracy": 0.9676556766033173,
      "step": 105
    },
    {
      "epoch": 0.05858854860186418,
      "grad_norm": 0.12101828306913376,
      "learning_rate": 0.00011702127659574468,
      "loss": 0.0816,
      "mean_token_accuracy": 0.9684371054172516,
      "step": 110
    },
    {
      "epoch": 0.06125166444740346,
      "grad_norm": 0.08741698414087296,
      "learning_rate": 0.0001223404255319149,
      "loss": 0.0633,
      "mean_token_accuracy": 0.9756288051605224,
      "step": 115
    },
    {
      "epoch": 0.06391478029294274,
      "grad_norm": 0.08274981379508972,
      "learning_rate": 0.00012765957446808513,
      "loss": 0.0763,
      "mean_token_accuracy": 0.970195347070694,
      "step": 120
    },
    {
      "epoch": 0.06657789613848203,
      "grad_norm": 0.06632668524980545,
      "learning_rate": 0.00013297872340425532,
      "loss": 0.0744,
      "mean_token_accuracy": 0.970732593536377,
      "step": 125
    },
    {
      "epoch": 0.0692410119840213,
      "grad_norm": 0.08805403858423233,
      "learning_rate": 0.00013829787234042554,
      "loss": 0.0949,
      "mean_token_accuracy": 0.9624908268451691,
      "step": 130
    },
    {
      "epoch": 0.07190412782956059,
      "grad_norm": 0.11924618482589722,
      "learning_rate": 0.00014361702127659576,
      "loss": 0.0682,
      "mean_token_accuracy": 0.9732844889163971,
      "step": 135
    },
    {
      "epoch": 0.07456724367509987,
      "grad_norm": 0.07259126007556915,
      "learning_rate": 0.00014893617021276596,
      "loss": 0.0644,
      "mean_token_accuracy": 0.9742368638515473,
      "step": 140
    },
    {
      "epoch": 0.07723035952063914,
      "grad_norm": 0.07263950258493423,
      "learning_rate": 0.00015425531914893618,
      "loss": 0.0723,
      "mean_token_accuracy": 0.9711477279663085,
      "step": 145
    },
    {
      "epoch": 0.07989347536617843,
      "grad_norm": 0.06513981521129608,
      "learning_rate": 0.00015957446808510637,
      "loss": 0.0671,
      "mean_token_accuracy": 0.9740170657634735,
      "step": 150
    },
    {
      "epoch": 0.08255659121171771,
      "grad_norm": 0.07975617796182632,
      "learning_rate": 0.00016489361702127662,
      "loss": 0.0653,
      "mean_token_accuracy": 0.9741758108139038,
      "step": 155
    },
    {
      "epoch": 0.085219707057257,
      "grad_norm": 0.08921811729669571,
      "learning_rate": 0.00017021276595744682,
      "loss": 0.0695,
      "mean_token_accuracy": 0.9725274622440339,
      "step": 160
    },
    {
      "epoch": 0.08788282290279627,
      "grad_norm": 0.06332514435052872,
      "learning_rate": 0.000175531914893617,
      "loss": 0.0575,
      "mean_token_accuracy": 0.9770817935466767,
      "step": 165
    },
    {
      "epoch": 0.09054593874833555,
      "grad_norm": 0.07502198964357376,
      "learning_rate": 0.00018085106382978726,
      "loss": 0.0725,
      "mean_token_accuracy": 0.9708791077136993,
      "step": 170
    },
    {
      "epoch": 0.09320905459387484,
      "grad_norm": 0.07119857519865036,
      "learning_rate": 0.00018617021276595746,
      "loss": 0.0684,
      "mean_token_accuracy": 0.9726861894130707,
      "step": 175
    },
    {
      "epoch": 0.09587217043941411,
      "grad_norm": 0.09054753929376602,
      "learning_rate": 0.00019148936170212768,
      "loss": 0.0629,
      "mean_token_accuracy": 0.974456650018692,
      "step": 180
    },
    {
      "epoch": 0.0985352862849534,
      "grad_norm": 0.07469576597213745,
      "learning_rate": 0.00019680851063829787,
      "loss": 0.0564,
      "mean_token_accuracy": 0.9774236798286438,
      "step": 185
    },
    {
      "epoch": 0.10119840213049268,
      "grad_norm": 0.06831200420856476,
      "learning_rate": 0.00019999930805760402,
      "loss": 0.0707,
      "mean_token_accuracy": 0.971208781003952,
      "step": 190
    },
    {
      "epoch": 0.10386151797603196,
      "grad_norm": 0.07172739505767822,
      "learning_rate": 0.00019999152381561955,
      "loss": 0.0547,
      "mean_token_accuracy": 0.9777655482292176,
      "step": 195
    },
    {
      "epoch": 0.10652463382157124,
      "grad_norm": 0.06316208839416504,
      "learning_rate": 0.0001999750910791767,
      "loss": 0.0601,
      "mean_token_accuracy": 0.9758241593837738,
      "step": 200
    },
    {
      "epoch": 0.10918774966711052,
      "grad_norm": 0.04626832157373428,
      "learning_rate": 0.00019995001126958025,
      "loss": 0.0466,
      "mean_token_accuracy": 0.9811965644359588,
      "step": 205
    },
    {
      "epoch": 0.1118508655126498,
      "grad_norm": 0.04393989220261574,
      "learning_rate": 0.00019991628655604003,
      "loss": 0.0486,
      "mean_token_accuracy": 0.9803174376487732,
      "step": 210
    },
    {
      "epoch": 0.11451398135818908,
      "grad_norm": 0.08774279803037643,
      "learning_rate": 0.00019987391985548328,
      "loss": 0.0476,
      "mean_token_accuracy": 0.981037837266922,
      "step": 215
    },
    {
      "epoch": 0.11717709720372836,
      "grad_norm": 0.07935914397239685,
      "learning_rate": 0.0001998229148323023,
      "loss": 0.0519,
      "mean_token_accuracy": 0.9785714149475098,
      "step": 220
    },
    {
      "epoch": 0.11984021304926765,
      "grad_norm": 0.07536792755126953,
      "learning_rate": 0.00019976327589803767,
      "loss": 0.0524,
      "mean_token_accuracy": 0.978253960609436,
      "step": 225
    },
    {
      "epoch": 0.12250332889480692,
      "grad_norm": 0.06311788409948349,
      "learning_rate": 0.0001996950082109965,
      "loss": 0.0641,
      "mean_token_accuracy": 0.9737606823444367,
      "step": 230
    },
    {
      "epoch": 0.12516644474034622,
      "grad_norm": 0.05745382234454155,
      "learning_rate": 0.00019961811767580648,
      "loss": 0.0556,
      "mean_token_accuracy": 0.9775213539600373,
      "step": 235
    },
    {
      "epoch": 0.1278295605858855,
      "grad_norm": 0.0590946190059185,
      "learning_rate": 0.0001995326109429049,
      "loss": 0.0601,
      "mean_token_accuracy": 0.9754822731018067,
      "step": 240
    },
    {
      "epoch": 0.13049267643142476,
      "grad_norm": 0.07194741815328598,
      "learning_rate": 0.00019943849540796375,
      "loss": 0.0538,
      "mean_token_accuracy": 0.978327214717865,
      "step": 245
    },
    {
      "epoch": 0.13315579227696406,
      "grad_norm": 0.07567504793405533,
      "learning_rate": 0.0001993357792112498,
      "loss": 0.0614,
      "mean_token_accuracy": 0.974957263469696,
      "step": 250
    },
    {
      "epoch": 0.13581890812250333,
      "grad_norm": 0.054747238755226135,
      "learning_rate": 0.0001992244712369207,
      "loss": 0.0479,
      "mean_token_accuracy": 0.9809889972209931,
      "step": 255
    },
    {
      "epoch": 0.1384820239680426,
      "grad_norm": 0.07008703052997589,
      "learning_rate": 0.00019910458111225646,
      "loss": 0.053,
      "mean_token_accuracy": 0.978302800655365,
      "step": 260
    },
    {
      "epoch": 0.1411451398135819,
      "grad_norm": 0.06375347077846527,
      "learning_rate": 0.00019897611920682677,
      "loss": 0.0677,
      "mean_token_accuracy": 0.9719413816928864,
      "step": 265
    },
    {
      "epoch": 0.14380825565912117,
      "grad_norm": 0.06617304682731628,
      "learning_rate": 0.00019883909663159424,
      "loss": 0.0534,
      "mean_token_accuracy": 0.9783882677555085,
      "step": 270
    },
    {
      "epoch": 0.14647137150466044,
      "grad_norm": 0.06903678178787231,
      "learning_rate": 0.0001986935252379532,
      "loss": 0.0553,
      "mean_token_accuracy": 0.9774358928203583,
      "step": 275
    },
    {
      "epoch": 0.14913448735019974,
      "grad_norm": 0.05433755740523338,
      "learning_rate": 0.00019853941761670483,
      "loss": 0.0541,
      "mean_token_accuracy": 0.9783272087574005,
      "step": 280
    },
    {
      "epoch": 0.151797603195739,
      "grad_norm": 0.06586781144142151,
      "learning_rate": 0.00019837678709696798,
      "loss": 0.041,
      "mean_token_accuracy": 0.9836629867553711,
      "step": 285
    },
    {
      "epoch": 0.15446071904127828,
      "grad_norm": 0.07344241440296173,
      "learning_rate": 0.00019820564774502644,
      "loss": 0.0699,
      "mean_token_accuracy": 0.9718070745468139,
      "step": 290
    },
    {
      "epoch": 0.15712383488681758,
      "grad_norm": 0.06562484800815582,
      "learning_rate": 0.0001980260143631122,
      "loss": 0.0503,
      "mean_token_accuracy": 0.9793650686740876,
      "step": 295
    },
    {
      "epoch": 0.15978695073235685,
      "grad_norm": 0.05266420543193817,
      "learning_rate": 0.00019783790248812533,
      "loss": 0.0524,
      "mean_token_accuracy": 0.978253960609436,
      "step": 300
    },
    {
      "epoch": 0.16245006657789615,
      "grad_norm": 0.09536128491163254,
      "learning_rate": 0.00019764132839029,
      "loss": 0.0655,
      "mean_token_accuracy": 0.9735408842563629,
      "step": 305
    },
    {
      "epoch": 0.16511318242343542,
      "grad_norm": 0.06160730496048927,
      "learning_rate": 0.00019743630907174725,
      "loss": 0.0568,
      "mean_token_accuracy": 0.9770451605319976,
      "step": 310
    },
    {
      "epoch": 0.1677762982689747,
      "grad_norm": 0.07126890867948532,
      "learning_rate": 0.0001972228622650846,
      "loss": 0.0649,
      "mean_token_accuracy": 0.973968243598938,
      "step": 315
    },
    {
      "epoch": 0.170439414114514,
      "grad_norm": 0.06424631178379059,
      "learning_rate": 0.0001970010064318021,
      "loss": 0.0578,
      "mean_token_accuracy": 0.9766178131103516,
      "step": 320
    },
    {
      "epoch": 0.17310252996005326,
      "grad_norm": 0.0797090232372284,
      "learning_rate": 0.00019677076076071566,
      "loss": 0.0511,
      "mean_token_accuracy": 0.9792185425758362,
      "step": 325
    },
    {
      "epoch": 0.17576564580559254,
      "grad_norm": 0.06460501998662949,
      "learning_rate": 0.00019653214516629735,
      "loss": 0.0583,
      "mean_token_accuracy": 0.976691073179245,
      "step": 330
    },
    {
      "epoch": 0.17842876165113183,
      "grad_norm": 0.06340578198432922,
      "learning_rate": 0.00019628518028695307,
      "loss": 0.0521,
      "mean_token_accuracy": 0.9784615218639374,
      "step": 335
    },
    {
      "epoch": 0.1810918774966711,
      "grad_norm": 0.048455528914928436,
      "learning_rate": 0.00019602988748323717,
      "loss": 0.0515,
      "mean_token_accuracy": 0.9792307555675507,
      "step": 340
    },
    {
      "epoch": 0.18375499334221038,
      "grad_norm": 0.052041687071323395,
      "learning_rate": 0.00019576628883600535,
      "loss": 0.0467,
      "mean_token_accuracy": 0.9806593298912049,
      "step": 345
    },
    {
      "epoch": 0.18641810918774968,
      "grad_norm": 0.07374859601259232,
      "learning_rate": 0.00019549440714450444,
      "loss": 0.0607,
      "mean_token_accuracy": 0.9754334390163422,
      "step": 350
    },
    {
      "epoch": 0.18908122503328895,
      "grad_norm": 0.05594540387392044,
      "learning_rate": 0.00019521426592440072,
      "loss": 0.066,
      "mean_token_accuracy": 0.9731135427951813,
      "step": 355
    },
    {
      "epoch": 0.19174434087882822,
      "grad_norm": 0.0629001334309578,
      "learning_rate": 0.00019492588940574586,
      "loss": 0.0555,
      "mean_token_accuracy": 0.9774358808994293,
      "step": 360
    },
    {
      "epoch": 0.19440745672436752,
      "grad_norm": 0.05497188866138458,
      "learning_rate": 0.0001946293025308813,
      "loss": 0.0481,
      "mean_token_accuracy": 0.9801220834255219,
      "step": 365
    },
    {
      "epoch": 0.1970705725699068,
      "grad_norm": 0.06228744983673096,
      "learning_rate": 0.00019432453095228076,
      "loss": 0.0663,
      "mean_token_accuracy": 0.9721977889537812,
      "step": 370
    },
    {
      "epoch": 0.19973368841544606,
      "grad_norm": 0.04110695794224739,
      "learning_rate": 0.00019401160103033174,
      "loss": 0.0512,
      "mean_token_accuracy": 0.9791330754756927,
      "step": 375
    },
    {
      "epoch": 0.20239680426098536,
      "grad_norm": 0.05923238396644592,
      "learning_rate": 0.00019369053983105532,
      "loss": 0.0462,
      "mean_token_accuracy": 0.9814529776573181,
      "step": 380
    },
    {
      "epoch": 0.20505992010652463,
      "grad_norm": 0.06135060265660286,
      "learning_rate": 0.00019336137512376532,
      "loss": 0.0506,
      "mean_token_accuracy": 0.9793650686740876,
      "step": 385
    },
    {
      "epoch": 0.20772303595206393,
      "grad_norm": 0.05882129445672035,
      "learning_rate": 0.00019302413537866642,
      "loss": 0.061,
      "mean_token_accuracy": 0.9756898581981659,
      "step": 390
    },
    {
      "epoch": 0.2103861517976032,
      "grad_norm": 0.0644676685333252,
      "learning_rate": 0.0001926788497643916,
      "loss": 0.0578,
      "mean_token_accuracy": 0.9761660397052765,
      "step": 395
    },
    {
      "epoch": 0.21304926764314247,
      "grad_norm": 0.048850301653146744,
      "learning_rate": 0.00019232554814547953,
      "loss": 0.0427,
      "mean_token_accuracy": 0.9819169521331788,
      "step": 400
    },
    {
      "epoch": 0.21571238348868177,
      "grad_norm": 0.05742983892560005,
      "learning_rate": 0.00019196426107979128,
      "loss": 0.0581,
      "mean_token_accuracy": 0.9765811860561371,
      "step": 405
    },
    {
      "epoch": 0.21837549933422104,
      "grad_norm": 0.05089128762483597,
      "learning_rate": 0.00019159501981586737,
      "loss": 0.0475,
      "mean_token_accuracy": 0.9808180510997773,
      "step": 410
    },
    {
      "epoch": 0.2210386151797603,
      "grad_norm": 0.06413483619689941,
      "learning_rate": 0.00019121785629022501,
      "loss": 0.0583,
      "mean_token_accuracy": 0.9765567600727081,
      "step": 415
    },
    {
      "epoch": 0.2237017310252996,
      "grad_norm": 0.057210683822631836,
      "learning_rate": 0.00019083280312459593,
      "loss": 0.066,
      "mean_token_accuracy": 0.9729426026344299,
      "step": 420
    },
    {
      "epoch": 0.22636484687083888,
      "grad_norm": 0.07737241685390472,
      "learning_rate": 0.0001904398936231047,
      "loss": 0.0612,
      "mean_token_accuracy": 0.9754578590393066,
      "step": 425
    },
    {
      "epoch": 0.22902796271637815,
      "grad_norm": 0.06051091104745865,
      "learning_rate": 0.00019003916176938836,
      "loss": 0.0512,
      "mean_token_accuracy": 0.9791941285133362,
      "step": 430
    },
    {
      "epoch": 0.23169107856191745,
      "grad_norm": 0.06462576240301132,
      "learning_rate": 0.00018963064222365694,
      "loss": 0.047,
      "mean_token_accuracy": 0.9807447969913483,
      "step": 435
    },
    {
      "epoch": 0.23435419440745672,
      "grad_norm": 0.065574049949646,
      "learning_rate": 0.00018921437031969558,
      "loss": 0.0476,
      "mean_token_accuracy": 0.9804151117801666,
      "step": 440
    },
    {
      "epoch": 0.237017310252996,
      "grad_norm": 0.05336959287524223,
      "learning_rate": 0.0001887903820618087,
      "loss": 0.0388,
      "mean_token_accuracy": 0.9844322204589844,
      "step": 445
    },
    {
      "epoch": 0.2396804260985353,
      "grad_norm": 0.06592069566249847,
      "learning_rate": 0.00018835871412170563,
      "loss": 0.0502,
      "mean_token_accuracy": 0.9795970559120178,
      "step": 450
    },
    {
      "epoch": 0.24234354194407456,
      "grad_norm": 0.04266435280442238,
      "learning_rate": 0.0001879194038353289,
      "loss": 0.0561,
      "mean_token_accuracy": 0.9772527337074279,
      "step": 455
    },
    {
      "epoch": 0.24500665778961384,
      "grad_norm": 0.04808111861348152,
      "learning_rate": 0.00018747248919962498,
      "loss": 0.0389,
      "mean_token_accuracy": 0.9841758131980896,
      "step": 460
    },
    {
      "epoch": 0.24766977363515313,
      "grad_norm": 0.05507846549153328,
      "learning_rate": 0.00018701800886925782,
      "loss": 0.0543,
      "mean_token_accuracy": 0.9776800870895386,
      "step": 465
    },
    {
      "epoch": 0.25033288948069243,
      "grad_norm": 0.049451231956481934,
      "learning_rate": 0.00018655600215326546,
      "loss": 0.0466,
      "mean_token_accuracy": 0.9812942504882812,
      "step": 470
    },
    {
      "epoch": 0.2529960053262317,
      "grad_norm": 0.0632142499089241,
      "learning_rate": 0.00018608650901166032,
      "loss": 0.0438,
      "mean_token_accuracy": 0.9818925321102142,
      "step": 475
    },
    {
      "epoch": 0.255659121171771,
      "grad_norm": 0.06887644529342651,
      "learning_rate": 0.0001856095700519726,
      "loss": 0.0568,
      "mean_token_accuracy": 0.9767521262168884,
      "step": 480
    },
    {
      "epoch": 0.2583222370173103,
      "grad_norm": 0.06655330210924149,
      "learning_rate": 0.0001851252265257384,
      "loss": 0.0546,
      "mean_token_accuracy": 0.9772893607616424,
      "step": 485
    },
    {
      "epoch": 0.2609853528628495,
      "grad_norm": 0.051667094230651855,
      "learning_rate": 0.0001846335203249316,
      "loss": 0.0598,
      "mean_token_accuracy": 0.975567752122879,
      "step": 490
    },
    {
      "epoch": 0.2636484687083888,
      "grad_norm": 0.062494028359651566,
      "learning_rate": 0.00018413449397834051,
      "loss": 0.0524,
      "mean_token_accuracy": 0.9785714209079742,
      "step": 495
    },
    {
      "epoch": 0.2663115845539281,
      "grad_norm": 0.07585468888282776,
      "learning_rate": 0.00018362819064788956,
      "loss": 0.055,
      "mean_token_accuracy": 0.9775335609912872,
      "step": 500
    },
    {
      "epoch": 0.26897470039946736,
      "grad_norm": 0.06810704618692398,
      "learning_rate": 0.00018311465412490608,
      "loss": 0.0574,
      "mean_token_accuracy": 0.9759340524673462,
      "step": 505
    },
    {
      "epoch": 0.27163781624500666,
      "grad_norm": 0.05042650178074837,
      "learning_rate": 0.00018259392882633265,
      "loss": 0.056,
      "mean_token_accuracy": 0.9768742263317108,
      "step": 510
    },
    {
      "epoch": 0.27430093209054596,
      "grad_norm": 0.07071951776742935,
      "learning_rate": 0.00018206605979088542,
      "loss": 0.0658,
      "mean_token_accuracy": 0.972893762588501,
      "step": 515
    },
    {
      "epoch": 0.2769640479360852,
      "grad_norm": 0.05426500737667084,
      "learning_rate": 0.0001815310926751586,
      "loss": 0.0489,
      "mean_token_accuracy": 0.979426109790802,
      "step": 520
    },
    {
      "epoch": 0.2796271637816245,
      "grad_norm": 0.047436077147722244,
      "learning_rate": 0.00018098907374967555,
      "loss": 0.047,
      "mean_token_accuracy": 0.980610477924347,
      "step": 525
    },
    {
      "epoch": 0.2822902796271638,
      "grad_norm": 0.07206933200359344,
      "learning_rate": 0.00018044004989488664,
      "loss": 0.0441,
      "mean_token_accuracy": 0.9823198795318604,
      "step": 530
    },
    {
      "epoch": 0.28495339547270304,
      "grad_norm": 0.05135834217071533,
      "learning_rate": 0.00017988406859711456,
      "loss": 0.042,
      "mean_token_accuracy": 0.9827960848808288,
      "step": 535
    },
    {
      "epoch": 0.28761651131824234,
      "grad_norm": 0.055974993854761124,
      "learning_rate": 0.00017932117794444713,
      "loss": 0.0471,
      "mean_token_accuracy": 0.9802075505256653,
      "step": 540
    },
    {
      "epoch": 0.29027962716378164,
      "grad_norm": 0.05277843400835991,
      "learning_rate": 0.00017875142662257786,
      "loss": 0.0514,
      "mean_token_accuracy": 0.9788400411605835,
      "step": 545
    },
    {
      "epoch": 0.2929427430093209,
      "grad_norm": 0.06408589333295822,
      "learning_rate": 0.00017817486391059532,
      "loss": 0.0487,
      "mean_token_accuracy": 0.9800854623317719,
      "step": 550
    },
    {
      "epoch": 0.2956058588548602,
      "grad_norm": 0.043870504945516586,
      "learning_rate": 0.0001775915396767205,
      "loss": 0.0489,
      "mean_token_accuracy": 0.979902309179306,
      "step": 555
    },
    {
      "epoch": 0.2982689747003995,
      "grad_norm": 0.04837007820606232,
      "learning_rate": 0.00017700150437399405,
      "loss": 0.0661,
      "mean_token_accuracy": 0.9734798431396484,
      "step": 560
    },
    {
      "epoch": 0.3009320905459387,
      "grad_norm": 0.04017208516597748,
      "learning_rate": 0.0001764048090359121,
      "loss": 0.05,
      "mean_token_accuracy": 0.9792307496070862,
      "step": 565
    },
    {
      "epoch": 0.303595206391478,
      "grad_norm": 0.0461135059595108,
      "learning_rate": 0.00017580150527201241,
      "loss": 0.0428,
      "mean_token_accuracy": 0.9829426050186157,
      "step": 570
    },
    {
      "epoch": 0.3062583222370173,
      "grad_norm": 0.06603188067674637,
      "learning_rate": 0.0001751916452634105,
      "loss": 0.0625,
      "mean_token_accuracy": 0.9739193975925445,
      "step": 575
    },
    {
      "epoch": 0.30892143808255657,
      "grad_norm": 0.05519532039761543,
      "learning_rate": 0.0001745752817582865,
      "loss": 0.0559,
      "mean_token_accuracy": 0.9767765462398529,
      "step": 580
    },
    {
      "epoch": 0.31158455392809586,
      "grad_norm": 0.06411441415548325,
      "learning_rate": 0.00017395246806732267,
      "loss": 0.055,
      "mean_token_accuracy": 0.9773748338222503,
      "step": 585
    },
    {
      "epoch": 0.31424766977363516,
      "grad_norm": 0.05509982258081436,
      "learning_rate": 0.00017332325805909256,
      "loss": 0.061,
      "mean_token_accuracy": 0.9751526117324829,
      "step": 590
    },
    {
      "epoch": 0.3169107856191744,
      "grad_norm": 0.0655801072716713,
      "learning_rate": 0.00017268770615540177,
      "loss": 0.0466,
      "mean_token_accuracy": 0.9806593298912049,
      "step": 595
    },
    {
      "epoch": 0.3195739014647137,
      "grad_norm": 0.06518279016017914,
      "learning_rate": 0.00017204586732658087,
      "loss": 0.0412,
      "mean_token_accuracy": 0.9829548120498657,
      "step": 600
    },
    {
      "epoch": 0.322237017310253,
      "grad_norm": 0.06401462852954865,
      "learning_rate": 0.00017139779708673085,
      "loss": 0.0552,
      "mean_token_accuracy": 0.9774725079536438,
      "step": 605
    },
    {
      "epoch": 0.3249001331557923,
      "grad_norm": 0.047077760100364685,
      "learning_rate": 0.00017074355148892167,
      "loss": 0.0446,
      "mean_token_accuracy": 0.9820268452167511,
      "step": 610
    },
    {
      "epoch": 0.32756324900133155,
      "grad_norm": 0.0499393492937088,
      "learning_rate": 0.00017008318712034403,
      "loss": 0.0503,
      "mean_token_accuracy": 0.9791696965694427,
      "step": 615
    },
    {
      "epoch": 0.33022636484687085,
      "grad_norm": 0.05421934276819229,
      "learning_rate": 0.00016941676109741508,
      "loss": 0.0541,
      "mean_token_accuracy": 0.9778876543045044,
      "step": 620
    },
    {
      "epoch": 0.33288948069241014,
      "grad_norm": 0.045506466180086136,
      "learning_rate": 0.00016874433106083814,
      "loss": 0.0379,
      "mean_token_accuracy": 0.9842490673065185,
      "step": 625
    },
    {
      "epoch": 0.3355525965379494,
      "grad_norm": 0.05099815875291824,
      "learning_rate": 0.00016806595517061744,
      "loss": 0.0546,
      "mean_token_accuracy": 0.976752120256424,
      "step": 630
    },
    {
      "epoch": 0.3382157123834887,
      "grad_norm": 0.0524461530148983,
      "learning_rate": 0.00016738169210102764,
      "loss": 0.0514,
      "mean_token_accuracy": 0.9787912011146546,
      "step": 635
    },
    {
      "epoch": 0.340878828229028,
      "grad_norm": 0.04635464772582054,
      "learning_rate": 0.00016669160103553884,
      "loss": 0.0503,
      "mean_token_accuracy": 0.978864461183548,
      "step": 640
    },
    {
      "epoch": 0.34354194407456723,
      "grad_norm": 0.0492749884724617,
      "learning_rate": 0.00016599574166169782,
      "loss": 0.0431,
      "mean_token_accuracy": 0.9821611523628235,
      "step": 645
    },
    {
      "epoch": 0.34620505992010653,
      "grad_norm": 0.04817197471857071,
      "learning_rate": 0.0001652941741659655,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9868620157241821,
      "step": 650
    },
    {
      "epoch": 0.3488681757656458,
      "grad_norm": 0.057936232537031174,
      "learning_rate": 0.00016458695922851125,
      "loss": 0.0423,
      "mean_token_accuracy": 0.98202685713768,
      "step": 655
    },
    {
      "epoch": 0.35153129161118507,
      "grad_norm": 0.04186012223362923,
      "learning_rate": 0.0001638741580179645,
      "loss": 0.0556,
      "mean_token_accuracy": 0.9770085275173187,
      "step": 660
    },
    {
      "epoch": 0.35419440745672437,
      "grad_norm": 0.06112098693847656,
      "learning_rate": 0.0001631558321861241,
      "loss": 0.0513,
      "mean_token_accuracy": 0.9787911951541901,
      "step": 665
    },
    {
      "epoch": 0.35685752330226367,
      "grad_norm": 0.055144574493169785,
      "learning_rate": 0.00016243204386262616,
      "loss": 0.0503,
      "mean_token_accuracy": 0.9787545680999756,
      "step": 670
    },
    {
      "epoch": 0.3595206391478029,
      "grad_norm": 0.06168635934591293,
      "learning_rate": 0.0001617028556495699,
      "loss": 0.0568,
      "mean_token_accuracy": 0.9764590799808502,
      "step": 675
    },
    {
      "epoch": 0.3621837549933422,
      "grad_norm": 0.05613715201616287,
      "learning_rate": 0.00016096833061610336,
      "loss": 0.0518,
      "mean_token_accuracy": 0.9790279984474182,
      "step": 680
    },
    {
      "epoch": 0.3648468708388815,
      "grad_norm": 0.048708170652389526,
      "learning_rate": 0.0001602285322929684,
      "loss": 0.0595,
      "mean_token_accuracy": 0.9754334390163422,
      "step": 685
    },
    {
      "epoch": 0.36750998668442075,
      "grad_norm": 0.06413900852203369,
      "learning_rate": 0.00015948352466700562,
      "loss": 0.0549,
      "mean_token_accuracy": 0.9773382127285004,
      "step": 690
    },
    {
      "epoch": 0.37017310252996005,
      "grad_norm": 0.05412254482507706,
      "learning_rate": 0.00015873337217562012,
      "loss": 0.0576,
      "mean_token_accuracy": 0.9758485853672028,
      "step": 695
    },
    {
      "epoch": 0.37283621837549935,
      "grad_norm": 0.07209640741348267,
      "learning_rate": 0.00015797813970120806,
      "loss": 0.0504,
      "mean_token_accuracy": 0.9789987623691558,
      "step": 700
    },
    {
      "epoch": 0.3754993342210386,
      "grad_norm": 0.0462648831307888,
      "learning_rate": 0.00015721789256554493,
      "loss": 0.0432,
      "mean_token_accuracy": 0.9820268452167511,
      "step": 705
    },
    {
      "epoch": 0.3781624500665779,
      "grad_norm": 0.054127827286720276,
      "learning_rate": 0.00015645269652413572,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9852625012397767,
      "step": 710
    },
    {
      "epoch": 0.3808255659121172,
      "grad_norm": 0.04338300973176956,
      "learning_rate": 0.00015568261776052747,
      "loss": 0.0457,
      "mean_token_accuracy": 0.9812576115131378,
      "step": 715
    },
    {
      "epoch": 0.38348868175765644,
      "grad_norm": 0.05176251381635666,
      "learning_rate": 0.0001549077228805851,
      "loss": 0.0402,
      "mean_token_accuracy": 0.9832844853401184,
      "step": 720
    },
    {
      "epoch": 0.38615179760319573,
      "grad_norm": 0.06750281155109406,
      "learning_rate": 0.00015412807890673012,
      "loss": 0.0519,
      "mean_token_accuracy": 0.9786324739456177,
      "step": 725
    },
    {
      "epoch": 0.38881491344873503,
      "grad_norm": 0.05919007956981659,
      "learning_rate": 0.00015334375327214435,
      "loss": 0.0386,
      "mean_token_accuracy": 0.9842979073524475,
      "step": 730
    },
    {
      "epoch": 0.3914780292942743,
      "grad_norm": 0.0641694962978363,
      "learning_rate": 0.00015255481381493686,
      "loss": 0.045,
      "mean_token_accuracy": 0.9814407587051391,
      "step": 735
    },
    {
      "epoch": 0.3941411451398136,
      "grad_norm": 0.07224282622337341,
      "learning_rate": 0.00015176132877227672,
      "loss": 0.0474,
      "mean_token_accuracy": 0.9796214699745178,
      "step": 740
    },
    {
      "epoch": 0.3968042609853529,
      "grad_norm": 0.032872989773750305,
      "learning_rate": 0.00015096336677449123,
      "loss": 0.0506,
      "mean_token_accuracy": 0.9786691009998322,
      "step": 745
    },
    {
      "epoch": 0.3994673768308921,
      "grad_norm": 0.04536876827478409,
      "learning_rate": 0.0001501609968391295,
      "loss": 0.0458,
      "mean_token_accuracy": 0.9807936370372772,
      "step": 750
    },
    {
      "epoch": 0.4021304926764314,
      "grad_norm": 0.03958405926823616,
      "learning_rate": 0.00014935428836499332,
      "loss": 0.0466,
      "mean_token_accuracy": 0.9809889972209931,
      "step": 755
    },
    {
      "epoch": 0.4047936085219707,
      "grad_norm": 0.06668851524591446,
      "learning_rate": 0.0001485433111261346,
      "loss": 0.0502,
      "mean_token_accuracy": 0.9785347878932953,
      "step": 760
    },
    {
      "epoch": 0.40745672436750996,
      "grad_norm": 0.042506150901317596,
      "learning_rate": 0.0001477281352658203,
      "loss": 0.0476,
      "mean_token_accuracy": 0.9801465034484863,
      "step": 765
    },
    {
      "epoch": 0.41011984021304926,
      "grad_norm": 0.04714113101363182,
      "learning_rate": 0.00014690883129046584,
      "loss": 0.0416,
      "mean_token_accuracy": 0.9826251447200776,
      "step": 770
    },
    {
      "epoch": 0.41278295605858856,
      "grad_norm": 0.08716364204883575,
      "learning_rate": 0.0001460854700635366,
      "loss": 0.0647,
      "mean_token_accuracy": 0.9730402827262878,
      "step": 775
    },
    {
      "epoch": 0.41544607190412786,
      "grad_norm": 0.06549763679504395,
      "learning_rate": 0.00014525812279941896,
      "loss": 0.0482,
      "mean_token_accuracy": 0.9799389362335205,
      "step": 780
    },
    {
      "epoch": 0.4181091877496671,
      "grad_norm": 0.05666585639119148,
      "learning_rate": 0.00014442686105726067,
      "loss": 0.0472,
      "mean_token_accuracy": 0.9804151237010956,
      "step": 785
    },
    {
      "epoch": 0.4207723035952064,
      "grad_norm": 0.059463534504175186,
      "learning_rate": 0.00014359175673478162,
      "loss": 0.0479,
      "mean_token_accuracy": 0.9803662836551666,
      "step": 790
    },
    {
      "epoch": 0.4234354194407457,
      "grad_norm": 0.04722663387656212,
      "learning_rate": 0.00014275288206205524,
      "loss": 0.0524,
      "mean_token_accuracy": 0.9780830085277558,
      "step": 795
    },
    {
      "epoch": 0.42609853528628494,
      "grad_norm": 0.045738983899354935,
      "learning_rate": 0.00014191030959526105,
      "loss": 0.0456,
      "mean_token_accuracy": 0.9813186585903168,
      "step": 800
    },
    {
      "epoch": 0.42876165113182424,
      "grad_norm": 0.039609119296073914,
      "learning_rate": 0.00014106411221040933,
      "loss": 0.0436,
      "mean_token_accuracy": 0.9822588443756104,
      "step": 805
    },
    {
      "epoch": 0.43142476697736354,
      "grad_norm": 0.05956882983446121,
      "learning_rate": 0.00014021436309703765,
      "loss": 0.0415,
      "mean_token_accuracy": 0.9826861917972565,
      "step": 810
    },
    {
      "epoch": 0.4340878828229028,
      "grad_norm": 0.062378112226724625,
      "learning_rate": 0.00013936113575188075,
      "loss": 0.0576,
      "mean_token_accuracy": 0.9756043791770935,
      "step": 815
    },
    {
      "epoch": 0.4367509986684421,
      "grad_norm": 0.06362305581569672,
      "learning_rate": 0.00013850450397251345,
      "loss": 0.0455,
      "mean_token_accuracy": 0.980976790189743,
      "step": 820
    },
    {
      "epoch": 0.4394141145139814,
      "grad_norm": 0.04958374425768852,
      "learning_rate": 0.0001376445418509679,
      "loss": 0.054,
      "mean_token_accuracy": 0.9774480938911438,
      "step": 825
    },
    {
      "epoch": 0.4420772303595206,
      "grad_norm": 0.045249808579683304,
      "learning_rate": 0.00013678132376732517,
      "loss": 0.0436,
      "mean_token_accuracy": 0.9819169580936432,
      "step": 830
    },
    {
      "epoch": 0.4447403462050599,
      "grad_norm": 0.059343721717596054,
      "learning_rate": 0.00013591492438328183,
      "loss": 0.0505,
      "mean_token_accuracy": 0.9791208624839782,
      "step": 835
    },
    {
      "epoch": 0.4474034620505992,
      "grad_norm": 0.04187900200486183,
      "learning_rate": 0.0001350454186356924,
      "loss": 0.0505,
      "mean_token_accuracy": 0.9787179350852966,
      "step": 840
    },
    {
      "epoch": 0.45006657789613846,
      "grad_norm": 0.052214790135622025,
      "learning_rate": 0.00013417288173008776,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9863491952419281,
      "step": 845
    },
    {
      "epoch": 0.45272969374167776,
      "grad_norm": 0.04296575114130974,
      "learning_rate": 0.00013329738913417068,
      "loss": 0.0501,
      "mean_token_accuracy": 0.9794993758201599,
      "step": 850
    },
    {
      "epoch": 0.45539280958721706,
      "grad_norm": 0.049184828996658325,
      "learning_rate": 0.00013241901657128825,
      "loss": 0.0408,
      "mean_token_accuracy": 0.9823565125465393,
      "step": 855
    },
    {
      "epoch": 0.4580559254327563,
      "grad_norm": 0.06455284357070923,
      "learning_rate": 0.00013153784001388247,
      "loss": 0.0448,
      "mean_token_accuracy": 0.9815140306949616,
      "step": 860
    },
    {
      "epoch": 0.4607190412782956,
      "grad_norm": 0.07309287786483765,
      "learning_rate": 0.00013065393567691913,
      "loss": 0.0534,
      "mean_token_accuracy": 0.9781806886196136,
      "step": 865
    },
    {
      "epoch": 0.4633821571238349,
      "grad_norm": 0.04752277955412865,
      "learning_rate": 0.00012976738001129606,
      "loss": 0.0458,
      "mean_token_accuracy": 0.9805549561977387,
      "step": 870
    },
    {
      "epoch": 0.46604527296937415,
      "grad_norm": 0.04298139736056328,
      "learning_rate": 0.00012887824969723034,
      "loss": 0.0413,
      "mean_token_accuracy": 0.9831013262271882,
      "step": 875
    },
    {
      "epoch": 0.46870838881491345,
      "grad_norm": 0.04882394149899483,
      "learning_rate": 0.00012798662163762635,
      "loss": 0.0428,
      "mean_token_accuracy": 0.9822100043296814,
      "step": 880
    },
    {
      "epoch": 0.47137150466045274,
      "grad_norm": 0.05154480040073395,
      "learning_rate": 0.00012709257295142422,
      "loss": 0.0372,
      "mean_token_accuracy": 0.9849450349807739,
      "step": 885
    },
    {
      "epoch": 0.474034620505992,
      "grad_norm": 0.04906482994556427,
      "learning_rate": 0.00012619618096692943,
      "loss": 0.0436,
      "mean_token_accuracy": 0.9814041376113891,
      "step": 890
    },
    {
      "epoch": 0.4766977363515313,
      "grad_norm": 0.04769090563058853,
      "learning_rate": 0.0001252975232151248,
      "loss": 0.0499,
      "mean_token_accuracy": 0.9794871747493744,
      "step": 895
    },
    {
      "epoch": 0.4793608521970706,
      "grad_norm": 0.06067895516753197,
      "learning_rate": 0.0001243966774229645,
      "loss": 0.0438,
      "mean_token_accuracy": 0.9816849708557129,
      "step": 900
    },
    {
      "epoch": 0.48202396804260983,
      "grad_norm": 0.042605381458997726,
      "learning_rate": 0.00012349372150665118,
      "loss": 0.0373,
      "mean_token_accuracy": 0.9843467473983765,
      "step": 905
    },
    {
      "epoch": 0.48468708388814913,
      "grad_norm": 0.062114305794239044,
      "learning_rate": 0.00012258873356489714,
      "loss": 0.0459,
      "mean_token_accuracy": 0.9810500562191009,
      "step": 910
    },
    {
      "epoch": 0.4873501997336884,
      "grad_norm": 0.05644230544567108,
      "learning_rate": 0.00012168179187216893,
      "loss": 0.0371,
      "mean_token_accuracy": 0.9850549340248108,
      "step": 915
    },
    {
      "epoch": 0.49001331557922767,
      "grad_norm": 0.0462392121553421,
      "learning_rate": 0.0001207729748719177,
      "loss": 0.041,
      "mean_token_accuracy": 0.9826495587825775,
      "step": 920
    },
    {
      "epoch": 0.49267643142476697,
      "grad_norm": 0.04839859530329704,
      "learning_rate": 0.00011986236116979406,
      "loss": 0.0411,
      "mean_token_accuracy": 0.9826251327991485,
      "step": 925
    },
    {
      "epoch": 0.49533954727030627,
      "grad_norm": 0.05229158326983452,
      "learning_rate": 0.0001189500295268495,
      "loss": 0.0413,
      "mean_token_accuracy": 0.9825274586677551,
      "step": 930
    },
    {
      "epoch": 0.4980026631158455,
      "grad_norm": 0.07273437082767487,
      "learning_rate": 0.0001180360588527242,
      "loss": 0.0417,
      "mean_token_accuracy": 0.9828571259975434,
      "step": 935
    },
    {
      "epoch": 0.5006657789613849,
      "grad_norm": 0.05428444594144821,
      "learning_rate": 0.00011712052819882171,
      "loss": 0.0488,
      "mean_token_accuracy": 0.9799145221710205,
      "step": 940
    },
    {
      "epoch": 0.5033288948069241,
      "grad_norm": 0.03856281563639641,
      "learning_rate": 0.00011620351675147195,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9858730018138886,
      "step": 945
    },
    {
      "epoch": 0.5059920106524634,
      "grad_norm": 0.05268105864524841,
      "learning_rate": 0.0001152851038250819,
      "loss": 0.0427,
      "mean_token_accuracy": 0.9826739728450775,
      "step": 950
    },
    {
      "epoch": 0.5086551264980027,
      "grad_norm": 0.051574211567640305,
      "learning_rate": 0.00011436536885527576,
      "loss": 0.0467,
      "mean_token_accuracy": 0.9806593298912049,
      "step": 955
    },
    {
      "epoch": 0.511318242343542,
      "grad_norm": 0.058736130595207214,
      "learning_rate": 0.00011344439139202421,
      "loss": 0.0433,
      "mean_token_accuracy": 0.9818681299686431,
      "step": 960
    },
    {
      "epoch": 0.5139813581890812,
      "grad_norm": 0.04859258979558945,
      "learning_rate": 0.00011252225109276404,
      "loss": 0.0619,
      "mean_token_accuracy": 0.9739438176155091,
      "step": 965
    },
    {
      "epoch": 0.5166444740346205,
      "grad_norm": 0.04141055420041084,
      "learning_rate": 0.00011159902771550837,
      "loss": 0.0382,
      "mean_token_accuracy": 0.9844200074672699,
      "step": 970
    },
    {
      "epoch": 0.5193075898801598,
      "grad_norm": 0.06450913846492767,
      "learning_rate": 0.00011067480111194817,
      "loss": 0.047,
      "mean_token_accuracy": 0.9803784966468811,
      "step": 975
    },
    {
      "epoch": 0.521970705725699,
      "grad_norm": 0.036628931760787964,
      "learning_rate": 0.00010974965122054579,
      "loss": 0.0444,
      "mean_token_accuracy": 0.9814529776573181,
      "step": 980
    },
    {
      "epoch": 0.5246338215712384,
      "grad_norm": 0.0419347807765007,
      "learning_rate": 0.00010882365805962083,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9871550559997558,
      "step": 985
    },
    {
      "epoch": 0.5272969374167776,
      "grad_norm": 0.061081454157829285,
      "learning_rate": 0.00010789690172042912,
      "loss": 0.0485,
      "mean_token_accuracy": 0.9795482158660889,
      "step": 990
    },
    {
      "epoch": 0.5299600532623169,
      "grad_norm": 0.03295666724443436,
      "learning_rate": 0.00010696946236023567,
      "loss": 0.0442,
      "mean_token_accuracy": 0.9816483318805694,
      "step": 995
    },
    {
      "epoch": 0.5326231691078562,
      "grad_norm": 0.05213489755988121,
      "learning_rate": 0.00010604142019538135,
      "loss": 0.0424,
      "mean_token_accuracy": 0.982710599899292,
      "step": 1000
    },
    {
      "epoch": 0.5352862849533955,
      "grad_norm": 0.042385201901197433,
      "learning_rate": 0.00010511285549434509,
      "loss": 0.0369,
      "mean_token_accuracy": 0.9848351597785949,
      "step": 1005
    },
    {
      "epoch": 0.5379494007989347,
      "grad_norm": 0.05076534301042557,
      "learning_rate": 0.00010418384857080117,
      "loss": 0.0429,
      "mean_token_accuracy": 0.9821611642837524,
      "step": 1010
    },
    {
      "epoch": 0.5406125166444741,
      "grad_norm": 0.06207333505153656,
      "learning_rate": 0.00010325447977667263,
      "loss": 0.0547,
      "mean_token_accuracy": 0.976752120256424,
      "step": 1015
    },
    {
      "epoch": 0.5432756324900133,
      "grad_norm": 0.056006304919719696,
      "learning_rate": 0.00010232482949518156,
      "loss": 0.0409,
      "mean_token_accuracy": 0.9827472329139709,
      "step": 1020
    },
    {
      "epoch": 0.5459387483355526,
      "grad_norm": 0.04523596540093422,
      "learning_rate": 0.00010139497813389654,
      "loss": 0.0468,
      "mean_token_accuracy": 0.9804395496845245,
      "step": 1025
    },
    {
      "epoch": 0.5486018641810919,
      "grad_norm": 0.056653622537851334,
      "learning_rate": 0.00010046500611777798,
      "loss": 0.0468,
      "mean_token_accuracy": 0.9804639697074891,
      "step": 1030
    },
    {
      "epoch": 0.5512649800266312,
      "grad_norm": 0.05866164341568947,
      "learning_rate": 9.953499388222202e-05,
      "loss": 0.045,
      "mean_token_accuracy": 0.9809889972209931,
      "step": 1035
    },
    {
      "epoch": 0.5539280958721704,
      "grad_norm": 0.053372107446193695,
      "learning_rate": 9.860502186610349e-05,
      "loss": 0.0517,
      "mean_token_accuracy": 0.9779609084129334,
      "step": 1040
    },
    {
      "epoch": 0.5565912117177098,
      "grad_norm": 0.05768711119890213,
      "learning_rate": 9.767517050481846e-05,
      "loss": 0.0423,
      "mean_token_accuracy": 0.9827716588973999,
      "step": 1045
    },
    {
      "epoch": 0.559254327563249,
      "grad_norm": 0.045836519449949265,
      "learning_rate": 9.67455202233274e-05,
      "loss": 0.0363,
      "mean_token_accuracy": 0.9848595678806304,
      "step": 1050
    },
    {
      "epoch": 0.5619174434087882,
      "grad_norm": 0.05728695169091225,
      "learning_rate": 9.581615142919887e-05,
      "loss": 0.0396,
      "mean_token_accuracy": 0.9834065854549408,
      "step": 1055
    },
    {
      "epoch": 0.5645805592543276,
      "grad_norm": 0.03806497901678085,
      "learning_rate": 9.488714450565491e-05,
      "loss": 0.0516,
      "mean_token_accuracy": 0.9779364943504334,
      "step": 1060
    },
    {
      "epoch": 0.5672436750998668,
      "grad_norm": 0.04686364531517029,
      "learning_rate": 9.395857980461867e-05,
      "loss": 0.0597,
      "mean_token_accuracy": 0.9751526176929474,
      "step": 1065
    },
    {
      "epoch": 0.5699067909454061,
      "grad_norm": 0.06154019758105278,
      "learning_rate": 9.303053763976434e-05,
      "loss": 0.0417,
      "mean_token_accuracy": 0.9825518727302551,
      "step": 1070
    },
    {
      "epoch": 0.5725699067909454,
      "grad_norm": 0.06537701189517975,
      "learning_rate": 9.210309827957089e-05,
      "loss": 0.0446,
      "mean_token_accuracy": 0.9810866832733154,
      "step": 1075
    },
    {
      "epoch": 0.5752330226364847,
      "grad_norm": 0.04378455877304077,
      "learning_rate": 9.117634194037922e-05,
      "loss": 0.0445,
      "mean_token_accuracy": 0.9811599373817443,
      "step": 1080
    },
    {
      "epoch": 0.5778961384820239,
      "grad_norm": 0.03224113956093788,
      "learning_rate": 9.025034877945422e-05,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9851502895355224,
      "step": 1085
    },
    {
      "epoch": 0.5805592543275633,
      "grad_norm": 0.050813645124435425,
      "learning_rate": 8.932519888805185e-05,
      "loss": 0.0453,
      "mean_token_accuracy": 0.9811965763568878,
      "step": 1090
    },
    {
      "epoch": 0.5832223701731025,
      "grad_norm": 0.06234807148575783,
      "learning_rate": 8.840097228449165e-05,
      "loss": 0.0446,
      "mean_token_accuracy": 0.9812820434570313,
      "step": 1095
    },
    {
      "epoch": 0.5858854860186418,
      "grad_norm": 0.046289362013339996,
      "learning_rate": 8.747774890723599e-05,
      "loss": 0.0361,
      "mean_token_accuracy": 0.9846764206886292,
      "step": 1100
    },
    {
      "epoch": 0.5885486018641811,
      "grad_norm": 0.04741502180695534,
      "learning_rate": 8.655560860797582e-05,
      "loss": 0.0392,
      "mean_token_accuracy": 0.9833821594715119,
      "step": 1105
    },
    {
      "epoch": 0.5912117177097204,
      "grad_norm": 0.035679835826158524,
      "learning_rate": 8.563463114472425e-05,
      "loss": 0.0386,
      "mean_token_accuracy": 0.9833333194255829,
      "step": 1110
    },
    {
      "epoch": 0.5938748335552596,
      "grad_norm": 0.03642835095524788,
      "learning_rate": 8.471489617491812e-05,
      "loss": 0.0424,
      "mean_token_accuracy": 0.9821611642837524,
      "step": 1115
    },
    {
      "epoch": 0.596537949400799,
      "grad_norm": 0.05134493485093117,
      "learning_rate": 8.379648324852808e-05,
      "loss": 0.0416,
      "mean_token_accuracy": 0.9830402791500091,
      "step": 1120
    },
    {
      "epoch": 0.5992010652463382,
      "grad_norm": 0.04159806668758392,
      "learning_rate": 8.287947180117832e-05,
      "loss": 0.0464,
      "mean_token_accuracy": 0.980549442768097,
      "step": 1125
    },
    {
      "epoch": 0.6018641810918774,
      "grad_norm": 0.04069337248802185,
      "learning_rate": 8.196394114727585e-05,
      "loss": 0.0387,
      "mean_token_accuracy": 0.9839316129684448,
      "step": 1130
    },
    {
      "epoch": 0.6045272969374168,
      "grad_norm": 0.055958956480026245,
      "learning_rate": 8.104997047315048e-05,
      "loss": 0.0469,
      "mean_token_accuracy": 0.9801464974880219,
      "step": 1135
    },
    {
      "epoch": 0.607190412782956,
      "grad_norm": 0.05381285026669502,
      "learning_rate": 8.013763883020596e-05,
      "loss": 0.0518,
      "mean_token_accuracy": 0.9784493148326874,
      "step": 1140
    },
    {
      "epoch": 0.6098535286284953,
      "grad_norm": 0.06098148599267006,
      "learning_rate": 7.92270251280823e-05,
      "loss": 0.0495,
      "mean_token_accuracy": 0.9790842354297637,
      "step": 1145
    },
    {
      "epoch": 0.6125166444740346,
      "grad_norm": 0.04997801408171654,
      "learning_rate": 7.831820812783108e-05,
      "loss": 0.0471,
      "mean_token_accuracy": 0.9798290371894837,
      "step": 1150
    },
    {
      "epoch": 0.6151797603195739,
      "grad_norm": 0.045936260372400284,
      "learning_rate": 7.741126643510292e-05,
      "loss": 0.0369,
      "mean_token_accuracy": 0.9849816679954528,
      "step": 1155
    },
    {
      "epoch": 0.6178428761651131,
      "grad_norm": 0.10932360589504242,
      "learning_rate": 7.650627849334881e-05,
      "loss": 0.038,
      "mean_token_accuracy": 0.984285694360733,
      "step": 1160
    },
    {
      "epoch": 0.6205059920106525,
      "grad_norm": 0.06736107170581818,
      "learning_rate": 7.560332257703555e-05,
      "loss": 0.0499,
      "mean_token_accuracy": 0.9795238018035889,
      "step": 1165
    },
    {
      "epoch": 0.6231691078561917,
      "grad_norm": 0.13129392266273499,
      "learning_rate": 7.470247678487522e-05,
      "loss": 0.0493,
      "mean_token_accuracy": 0.978888875246048,
      "step": 1170
    },
    {
      "epoch": 0.625832223701731,
      "grad_norm": 0.04084044694900513,
      "learning_rate": 7.380381903307061e-05,
      "loss": 0.0411,
      "mean_token_accuracy": 0.9828205049037934,
      "step": 1175
    },
    {
      "epoch": 0.6284953395472703,
      "grad_norm": 0.041778337210416794,
      "learning_rate": 7.290742704857585e-05,
      "loss": 0.0487,
      "mean_token_accuracy": 0.9800610363483429,
      "step": 1180
    },
    {
      "epoch": 0.6311584553928096,
      "grad_norm": 0.043811947107315063,
      "learning_rate": 7.201337836237365e-05,
      "loss": 0.0424,
      "mean_token_accuracy": 0.9816605448722839,
      "step": 1185
    },
    {
      "epoch": 0.6338215712383488,
      "grad_norm": 0.060183800756931305,
      "learning_rate": 7.112175030276969e-05,
      "loss": 0.0498,
      "mean_token_accuracy": 0.9791941285133362,
      "step": 1190
    },
    {
      "epoch": 0.6364846870838882,
      "grad_norm": 0.042345814406871796,
      "learning_rate": 7.023261998870395e-05,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9873259902000427,
      "step": 1195
    },
    {
      "epoch": 0.6391478029294274,
      "grad_norm": 0.06444382667541504,
      "learning_rate": 6.934606432308086e-05,
      "loss": 0.0484,
      "mean_token_accuracy": 0.9798046231269837,
      "step": 1200
    },
    {
      "epoch": 0.6418109187749668,
      "grad_norm": 0.046876195818185806,
      "learning_rate": 6.846215998611757e-05,
      "loss": 0.0486,
      "mean_token_accuracy": 0.9789499282836914,
      "step": 1205
    },
    {
      "epoch": 0.644474034620506,
      "grad_norm": 0.057887110859155655,
      "learning_rate": 6.758098342871174e-05,
      "loss": 0.057,
      "mean_token_accuracy": 0.9761416196823121,
      "step": 1210
    },
    {
      "epoch": 0.6471371504660453,
      "grad_norm": 0.03972464054822922,
      "learning_rate": 6.670261086582933e-05,
      "loss": 0.0473,
      "mean_token_accuracy": 0.9801709234714509,
      "step": 1215
    },
    {
      "epoch": 0.6498002663115846,
      "grad_norm": 0.043679483234882355,
      "learning_rate": 6.582711826991226e-05,
      "loss": 0.0397,
      "mean_token_accuracy": 0.9834798455238343,
      "step": 1220
    },
    {
      "epoch": 0.6524633821571239,
      "grad_norm": 0.05936583876609802,
      "learning_rate": 6.495458136430765e-05,
      "loss": 0.0376,
      "mean_token_accuracy": 0.9842490732669831,
      "step": 1225
    },
    {
      "epoch": 0.6551264980026631,
      "grad_norm": 0.04087141901254654,
      "learning_rate": 6.408507561671819e-05,
      "loss": 0.0462,
      "mean_token_accuracy": 0.9808546841144562,
      "step": 1230
    },
    {
      "epoch": 0.6577896138482024,
      "grad_norm": 0.05433586984872818,
      "learning_rate": 6.321867623267481e-05,
      "loss": 0.049,
      "mean_token_accuracy": 0.9793772757053375,
      "step": 1235
    },
    {
      "epoch": 0.6604527296937417,
      "grad_norm": 0.06001468747854233,
      "learning_rate": 6.23554581490321e-05,
      "loss": 0.0492,
      "mean_token_accuracy": 0.9789987683296204,
      "step": 1240
    },
    {
      "epoch": 0.6631158455392809,
      "grad_norm": 0.04763103276491165,
      "learning_rate": 6.149549602748656e-05,
      "loss": 0.0576,
      "mean_token_accuracy": 0.9754822731018067,
      "step": 1245
    },
    {
      "epoch": 0.6657789613848203,
      "grad_norm": 0.06517472863197327,
      "learning_rate": 6.063886424811929e-05,
      "loss": 0.0438,
      "mean_token_accuracy": 0.9815750777721405,
      "step": 1250
    },
    {
      "epoch": 0.6684420772303595,
      "grad_norm": 0.0524207204580307,
      "learning_rate": 5.9785636902962374e-05,
      "loss": 0.0444,
      "mean_token_accuracy": 0.9813186764717102,
      "step": 1255
    },
    {
      "epoch": 0.6711051930758988,
      "grad_norm": 0.05194844305515289,
      "learning_rate": 5.893588778959067e-05,
      "loss": 0.0406,
      "mean_token_accuracy": 0.9824053585529328,
      "step": 1260
    },
    {
      "epoch": 0.6737683089214381,
      "grad_norm": 0.04531911760568619,
      "learning_rate": 5.8089690404738925e-05,
      "loss": 0.0394,
      "mean_token_accuracy": 0.9833089053630829,
      "step": 1265
    },
    {
      "epoch": 0.6764314247669774,
      "grad_norm": 0.05924287810921669,
      "learning_rate": 5.7247117937944786e-05,
      "loss": 0.0421,
      "mean_token_accuracy": 0.9821977853775025,
      "step": 1270
    },
    {
      "epoch": 0.6790945406125166,
      "grad_norm": 0.04600124806165695,
      "learning_rate": 5.640824326521841e-05,
      "loss": 0.0482,
      "mean_token_accuracy": 0.9791574954986573,
      "step": 1275
    },
    {
      "epoch": 0.681757656458056,
      "grad_norm": 0.060442570596933365,
      "learning_rate": 5.5573138942739365e-05,
      "loss": 0.0464,
      "mean_token_accuracy": 0.9802686095237731,
      "step": 1280
    },
    {
      "epoch": 0.6844207723035952,
      "grad_norm": 0.05491984635591507,
      "learning_rate": 5.4741877200581057e-05,
      "loss": 0.0571,
      "mean_token_accuracy": 0.9753601789474488,
      "step": 1285
    },
    {
      "epoch": 0.6870838881491345,
      "grad_norm": 0.048727113753557205,
      "learning_rate": 5.391452993646342e-05,
      "loss": 0.0524,
      "mean_token_accuracy": 0.9780952274799347,
      "step": 1290
    },
    {
      "epoch": 0.6897470039946738,
      "grad_norm": 0.04571819305419922,
      "learning_rate": 5.30911687095342e-05,
      "loss": 0.0419,
      "mean_token_accuracy": 0.982209998369217,
      "step": 1295
    },
    {
      "epoch": 0.6924101198402131,
      "grad_norm": 0.06429927051067352,
      "learning_rate": 5.227186473417971e-05,
      "loss": 0.0446,
      "mean_token_accuracy": 0.9807325899600983,
      "step": 1300
    },
    {
      "epoch": 0.6950732356857523,
      "grad_norm": 0.05295795202255249,
      "learning_rate": 5.145668887386543e-05,
      "loss": 0.0389,
      "mean_token_accuracy": 0.9835286796092987,
      "step": 1305
    },
    {
      "epoch": 0.6977363515312917,
      "grad_norm": 0.06455942988395691,
      "learning_rate": 5.064571163500667e-05,
      "loss": 0.0541,
      "mean_token_accuracy": 0.97710622549057,
      "step": 1310
    },
    {
      "epoch": 0.7003994673768309,
      "grad_norm": 0.05818232521414757,
      "learning_rate": 4.983900316087051e-05,
      "loss": 0.0468,
      "mean_token_accuracy": 0.9796092748641968,
      "step": 1315
    },
    {
      "epoch": 0.7030625832223701,
      "grad_norm": 0.04409760236740112,
      "learning_rate": 4.90366332255088e-05,
      "loss": 0.0534,
      "mean_token_accuracy": 0.9773870468139648,
      "step": 1320
    },
    {
      "epoch": 0.7057256990679095,
      "grad_norm": 0.06393323838710785,
      "learning_rate": 4.823867122772329e-05,
      "loss": 0.0428,
      "mean_token_accuracy": 0.9825396656990051,
      "step": 1325
    },
    {
      "epoch": 0.7083888149134487,
      "grad_norm": 0.04404666647315025,
      "learning_rate": 4.744518618506319e-05,
      "loss": 0.0452,
      "mean_token_accuracy": 0.9808302700519562,
      "step": 1330
    },
    {
      "epoch": 0.711051930758988,
      "grad_norm": 0.05861349031329155,
      "learning_rate": 4.665624672785566e-05,
      "loss": 0.0408,
      "mean_token_accuracy": 0.982637345790863,
      "step": 1335
    },
    {
      "epoch": 0.7137150466045273,
      "grad_norm": 0.04552765563130379,
      "learning_rate": 4.5871921093269875e-05,
      "loss": 0.0461,
      "mean_token_accuracy": 0.9806715369224548,
      "step": 1340
    },
    {
      "epoch": 0.7163781624500666,
      "grad_norm": 0.06690757721662521,
      "learning_rate": 4.5092277119414975e-05,
      "loss": 0.0432,
      "mean_token_accuracy": 0.9817826449871063,
      "step": 1345
    },
    {
      "epoch": 0.7190412782956058,
      "grad_norm": 0.04385219141840935,
      "learning_rate": 4.431738223947252e-05,
      "loss": 0.033,
      "mean_token_accuracy": 0.9860195100307465,
      "step": 1350
    },
    {
      "epoch": 0.7217043941411452,
      "grad_norm": 0.042622532695531845,
      "learning_rate": 4.35473034758643e-05,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9848229348659515,
      "step": 1355
    },
    {
      "epoch": 0.7243675099866844,
      "grad_norm": 0.04397781938314438,
      "learning_rate": 4.2782107434455054e-05,
      "loss": 0.0427,
      "mean_token_accuracy": 0.9819902241230011,
      "step": 1360
    },
    {
      "epoch": 0.7270306258322237,
      "grad_norm": 0.049341488629579544,
      "learning_rate": 4.202186029879195e-05,
      "loss": 0.042,
      "mean_token_accuracy": 0.9824175596237182,
      "step": 1365
    },
    {
      "epoch": 0.729693741677763,
      "grad_norm": 0.04645610973238945,
      "learning_rate": 4.12666278243799e-05,
      "loss": 0.032,
      "mean_token_accuracy": 0.9869352757930756,
      "step": 1370
    },
    {
      "epoch": 0.7323568575233023,
      "grad_norm": 0.039163220673799515,
      "learning_rate": 4.0516475332994383e-05,
      "loss": 0.0394,
      "mean_token_accuracy": 0.983870553970337,
      "step": 1375
    },
    {
      "epoch": 0.7350199733688415,
      "grad_norm": 0.054667674005031586,
      "learning_rate": 3.9771467707031615e-05,
      "loss": 0.0418,
      "mean_token_accuracy": 0.9822100043296814,
      "step": 1380
    },
    {
      "epoch": 0.7376830892143809,
      "grad_norm": 0.04490191861987114,
      "learning_rate": 3.903166938389664e-05,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9861904561519623,
      "step": 1385
    },
    {
      "epoch": 0.7403462050599201,
      "grad_norm": 0.0483897365629673,
      "learning_rate": 3.8297144350430144e-05,
      "loss": 0.0506,
      "mean_token_accuracy": 0.9783150017261505,
      "step": 1390
    },
    {
      "epoch": 0.7430093209054593,
      "grad_norm": 0.050037335604429245,
      "learning_rate": 3.756795613737388e-05,
      "loss": 0.0416,
      "mean_token_accuracy": 0.9822344243526459,
      "step": 1395
    },
    {
      "epoch": 0.7456724367509987,
      "grad_norm": 0.05803540349006653,
      "learning_rate": 3.684416781387589e-05,
      "loss": 0.0395,
      "mean_token_accuracy": 0.9835408866405487,
      "step": 1400
    },
    {
      "epoch": 0.748335552596538,
      "grad_norm": 0.04605701565742493,
      "learning_rate": 3.6125841982035536e-05,
      "loss": 0.0367,
      "mean_token_accuracy": 0.9844322204589844,
      "step": 1405
    },
    {
      "epoch": 0.7509986684420772,
      "grad_norm": 0.05056578665971756,
      "learning_rate": 3.5413040771488746e-05,
      "loss": 0.0442,
      "mean_token_accuracy": 0.9809784710407257,
      "step": 1410
    },
    {
      "epoch": 0.7536617842876165,
      "grad_norm": 0.04270164296030998,
      "learning_rate": 3.47058258340345e-05,
      "loss": 0.0467,
      "mean_token_accuracy": 0.9802930176258087,
      "step": 1415
    },
    {
      "epoch": 0.7563249001331558,
      "grad_norm": 0.04214705154299736,
      "learning_rate": 3.4004258338302195e-05,
      "loss": 0.0362,
      "mean_token_accuracy": 0.9850060999393463,
      "step": 1420
    },
    {
      "epoch": 0.758988015978695,
      "grad_norm": 0.04405741021037102,
      "learning_rate": 3.3308398964461206e-05,
      "loss": 0.0452,
      "mean_token_accuracy": 0.9807936370372772,
      "step": 1425
    },
    {
      "epoch": 0.7616511318242344,
      "grad_norm": 0.06789631396532059,
      "learning_rate": 3.261830789897241e-05,
      "loss": 0.0472,
      "mean_token_accuracy": 0.9795726358890533,
      "step": 1430
    },
    {
      "epoch": 0.7643142476697736,
      "grad_norm": 0.05779273435473442,
      "learning_rate": 3.193404482938256e-05,
      "loss": 0.0547,
      "mean_token_accuracy": 0.9761782586574554,
      "step": 1435
    },
    {
      "epoch": 0.7669773635153129,
      "grad_norm": 0.0402994304895401,
      "learning_rate": 3.1255668939161894e-05,
      "loss": 0.0396,
      "mean_token_accuracy": 0.983247846364975,
      "step": 1440
    },
    {
      "epoch": 0.7696404793608522,
      "grad_norm": 0.0458141565322876,
      "learning_rate": 3.058323890258498e-05,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9849328339099884,
      "step": 1445
    },
    {
      "epoch": 0.7723035952063915,
      "grad_norm": 0.038610294461250305,
      "learning_rate": 2.9916812879655975e-05,
      "loss": 0.0501,
      "mean_token_accuracy": 0.9783516347408294,
      "step": 1450
    },
    {
      "epoch": 0.7749667110519307,
      "grad_norm": 0.06346781551837921,
      "learning_rate": 2.925644851107835e-05,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9874236702919006,
      "step": 1455
    },
    {
      "epoch": 0.7776298268974701,
      "grad_norm": 0.037486497312784195,
      "learning_rate": 2.860220291326915e-05,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9851404011249543,
      "step": 1460
    },
    {
      "epoch": 0.7802929427430093,
      "grad_norm": 0.06529121845960617,
      "learning_rate": 2.7954132673419143e-05,
      "loss": 0.0422,
      "mean_token_accuracy": 0.9822344183921814,
      "step": 1465
    },
    {
      "epoch": 0.7829560585885486,
      "grad_norm": 0.045860741287469864,
      "learning_rate": 2.7312293844598246e-05,
      "loss": 0.0435,
      "mean_token_accuracy": 0.9818559169769288,
      "step": 1470
    },
    {
      "epoch": 0.7856191744340879,
      "grad_norm": 0.05481775850057602,
      "learning_rate": 2.6676741940907478e-05,
      "loss": 0.0604,
      "mean_token_accuracy": 0.9741025567054749,
      "step": 1475
    },
    {
      "epoch": 0.7882822902796272,
      "grad_norm": 0.08080944418907166,
      "learning_rate": 2.6047531932677383e-05,
      "loss": 0.0592,
      "mean_token_accuracy": 0.975018298625946,
      "step": 1480
    },
    {
      "epoch": 0.7909454061251664,
      "grad_norm": 0.06755299866199493,
      "learning_rate": 2.542471824171353e-05,
      "loss": 0.0526,
      "mean_token_accuracy": 0.977777761220932,
      "step": 1485
    },
    {
      "epoch": 0.7936085219707057,
      "grad_norm": 0.05000187084078789,
      "learning_rate": 2.4808354736589523e-05,
      "loss": 0.0437,
      "mean_token_accuracy": 0.9818559110164642,
      "step": 1490
    },
    {
      "epoch": 0.796271637816245,
      "grad_norm": 0.05813051760196686,
      "learning_rate": 2.419849472798761e-05,
      "loss": 0.0451,
      "mean_token_accuracy": 0.9809035301208496,
      "step": 1495
    },
    {
      "epoch": 0.7989347536617842,
      "grad_norm": 0.04271751642227173,
      "learning_rate": 2.359519096408791e-05,
      "loss": 0.045,
      "mean_token_accuracy": 0.9813552975654602,
      "step": 1500
    },
    {
      "epoch": 0.8015978695073236,
      "grad_norm": 0.04828868433833122,
      "learning_rate": 2.2998495626005957e-05,
      "loss": 0.0339,
      "mean_token_accuracy": 0.985848581790924,
      "step": 1505
    },
    {
      "epoch": 0.8042609853528628,
      "grad_norm": 0.03289498761296272,
      "learning_rate": 2.240846032327949e-05,
      "loss": 0.032,
      "mean_token_accuracy": 0.9865933954715729,
      "step": 1510
    },
    {
      "epoch": 0.8069241011984021,
      "grad_norm": 0.05242108181118965,
      "learning_rate": 2.1825136089404718e-05,
      "loss": 0.0463,
      "mean_token_accuracy": 0.9802197635173797,
      "step": 1515
    },
    {
      "epoch": 0.8095872170439414,
      "grad_norm": 0.050270598381757736,
      "learning_rate": 2.1248573377422155e-05,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9858363628387451,
      "step": 1520
    },
    {
      "epoch": 0.8122503328894807,
      "grad_norm": 0.05596785992383957,
      "learning_rate": 2.0678822055552906e-05,
      "loss": 0.0463,
      "mean_token_accuracy": 0.9803540706634521,
      "step": 1525
    },
    {
      "epoch": 0.8149134487350199,
      "grad_norm": 0.053749240934848785,
      "learning_rate": 2.0115931402885458e-05,
      "loss": 0.0472,
      "mean_token_accuracy": 0.9804517567157746,
      "step": 1530
    },
    {
      "epoch": 0.8175765645805593,
      "grad_norm": 0.04804141819477081,
      "learning_rate": 1.955995010511338e-05,
      "loss": 0.0518,
      "mean_token_accuracy": 0.9780585944652558,
      "step": 1535
    },
    {
      "epoch": 0.8202396804260985,
      "grad_norm": 0.045841652899980545,
      "learning_rate": 1.901092625032448e-05,
      "loss": 0.0485,
      "mean_token_accuracy": 0.979365062713623,
      "step": 1540
    },
    {
      "epoch": 0.8229027962716379,
      "grad_norm": 0.0423155315220356,
      "learning_rate": 1.84689073248414e-05,
      "loss": 0.0407,
      "mean_token_accuracy": 0.9830280661582946,
      "step": 1545
    },
    {
      "epoch": 0.8255659121171771,
      "grad_norm": 0.04051283746957779,
      "learning_rate": 1.7933940209114597e-05,
      "loss": 0.0396,
      "mean_token_accuracy": 0.9830036520957947,
      "step": 1550
    },
    {
      "epoch": 0.8282290279627164,
      "grad_norm": 0.04280205816030502,
      "learning_rate": 1.7406071173667372e-05,
      "loss": 0.0417,
      "mean_token_accuracy": 0.9825030326843261,
      "step": 1555
    },
    {
      "epoch": 0.8308921438082557,
      "grad_norm": 0.045416004955768585,
      "learning_rate": 1.6885345875093918e-05,
      "loss": 0.0449,
      "mean_token_accuracy": 0.9808058440685272,
      "step": 1560
    },
    {
      "epoch": 0.833555259653795,
      "grad_norm": 0.044623058289289474,
      "learning_rate": 1.6371809352110447e-05,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9845054864883422,
      "step": 1565
    },
    {
      "epoch": 0.8362183754993342,
      "grad_norm": 0.03994055837392807,
      "learning_rate": 1.5865506021659516e-05,
      "loss": 0.0441,
      "mean_token_accuracy": 0.981098884344101,
      "step": 1570
    },
    {
      "epoch": 0.8388814913448736,
      "grad_norm": 0.06405629962682724,
      "learning_rate": 1.5366479675068435e-05,
      "loss": 0.0528,
      "mean_token_accuracy": 0.9776190400123597,
      "step": 1575
    },
    {
      "epoch": 0.8415446071904128,
      "grad_norm": 0.06854992359876633,
      "learning_rate": 1.4874773474261638e-05,
      "loss": 0.0372,
      "mean_token_accuracy": 0.9850305020809174,
      "step": 1580
    },
    {
      "epoch": 0.844207723035952,
      "grad_norm": 0.05664476752281189,
      "learning_rate": 1.4390429948027428e-05,
      "loss": 0.0432,
      "mean_token_accuracy": 0.9814163386821747,
      "step": 1585
    },
    {
      "epoch": 0.8468708388814914,
      "grad_norm": 0.04452156275510788,
      "learning_rate": 1.3913490988339718e-05,
      "loss": 0.0415,
      "mean_token_accuracy": 0.9819780170917511,
      "step": 1590
    },
    {
      "epoch": 0.8495339547270306,
      "grad_norm": 0.055095456540584564,
      "learning_rate": 1.3443997846734535e-05,
      "loss": 0.0508,
      "mean_token_accuracy": 0.9781196355819702,
      "step": 1595
    },
    {
      "epoch": 0.8521970705725699,
      "grad_norm": 0.049195777624845505,
      "learning_rate": 1.2981991130742211e-05,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9868620216846467,
      "step": 1600
    },
    {
      "epoch": 0.8548601864181092,
      "grad_norm": 0.050247907638549805,
      "learning_rate": 1.2527510800375043e-05,
      "loss": 0.0358,
      "mean_token_accuracy": 0.9846397936344147,
      "step": 1605
    },
    {
      "epoch": 0.8575233022636485,
      "grad_norm": 0.059565331786870956,
      "learning_rate": 1.20805961646711e-05,
      "loss": 0.0451,
      "mean_token_accuracy": 0.9804273426532746,
      "step": 1610
    },
    {
      "epoch": 0.8601864181091877,
      "grad_norm": 0.042766787111759186,
      "learning_rate": 1.1641285878294372e-05,
      "loss": 0.0411,
      "mean_token_accuracy": 0.982661771774292,
      "step": 1615
    },
    {
      "epoch": 0.8628495339547271,
      "grad_norm": 0.037687476724386215,
      "learning_rate": 1.1209617938191307e-05,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9880708038806916,
      "step": 1620
    },
    {
      "epoch": 0.8655126498002663,
      "grad_norm": 0.06686782836914062,
      "learning_rate": 1.0785629680304432e-05,
      "loss": 0.042,
      "mean_token_accuracy": 0.9822710454463959,
      "step": 1625
    },
    {
      "epoch": 0.8681757656458056,
      "grad_norm": 0.04809387028217316,
      "learning_rate": 1.0369357776343103e-05,
      "loss": 0.0401,
      "mean_token_accuracy": 0.9831379592418671,
      "step": 1630
    },
    {
      "epoch": 0.8708388814913449,
      "grad_norm": 0.041997481137514114,
      "learning_rate": 9.960838230611635e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9850183010101319,
      "step": 1635
    },
    {
      "epoch": 0.8735019973368842,
      "grad_norm": 0.0419871024787426,
      "learning_rate": 9.560106376895306e-06,
      "loss": 0.039,
      "mean_token_accuracy": 0.983711838722229,
      "step": 1640
    },
    {
      "epoch": 0.8761651131824234,
      "grad_norm": 0.05675376579165459,
      "learning_rate": 9.167196875404094e-06,
      "loss": 0.0457,
      "mean_token_accuracy": 0.9801098823547363,
      "step": 1645
    },
    {
      "epoch": 0.8788282290279628,
      "grad_norm": 0.04745902493596077,
      "learning_rate": 8.782143709775015e-06,
      "loss": 0.0487,
      "mean_token_accuracy": 0.9794505357742309,
      "step": 1650
    },
    {
      "epoch": 0.881491344873502,
      "grad_norm": 0.04839067533612251,
      "learning_rate": 8.40498018413266e-06,
      "loss": 0.0417,
      "mean_token_accuracy": 0.9824664056301117,
      "step": 1655
    },
    {
      "epoch": 0.8841544607190412,
      "grad_norm": 0.04664967581629753,
      "learning_rate": 8.035738920208714e-06,
      "loss": 0.0425,
      "mean_token_accuracy": 0.9820756852626801,
      "step": 1660
    },
    {
      "epoch": 0.8868175765645806,
      "grad_norm": 0.06269760429859161,
      "learning_rate": 7.67445185452046e-06,
      "loss": 0.0389,
      "mean_token_accuracy": 0.9834432125091552,
      "step": 1665
    },
    {
      "epoch": 0.8894806924101198,
      "grad_norm": 0.037270188331604004,
      "learning_rate": 7.321150235608399e-06,
      "loss": 0.043,
      "mean_token_accuracy": 0.9818803250789643,
      "step": 1670
    },
    {
      "epoch": 0.8921438082556591,
      "grad_norm": 0.05021100491285324,
      "learning_rate": 6.9758646213336165e-06,
      "loss": 0.039,
      "mean_token_accuracy": 0.9833943724632264,
      "step": 1675
    },
    {
      "epoch": 0.8948069241011984,
      "grad_norm": 0.04479273036122322,
      "learning_rate": 6.6386248762347004e-06,
      "loss": 0.0441,
      "mean_token_accuracy": 0.981575071811676,
      "step": 1680
    },
    {
      "epoch": 0.8974700399467377,
      "grad_norm": 0.04378955066204071,
      "learning_rate": 6.309460168944692e-06,
      "loss": 0.0449,
      "mean_token_accuracy": 0.9806104898452759,
      "step": 1685
    },
    {
      "epoch": 0.9001331557922769,
      "grad_norm": 0.06034992262721062,
      "learning_rate": 5.988398969668285e-06,
      "loss": 0.0469,
      "mean_token_accuracy": 0.9805616497993469,
      "step": 1690
    },
    {
      "epoch": 0.9027962716378163,
      "grad_norm": 0.046302057802677155,
      "learning_rate": 5.6754690477192396e-06,
      "loss": 0.0438,
      "mean_token_accuracy": 0.980976790189743,
      "step": 1695
    },
    {
      "epoch": 0.9054593874833555,
      "grad_norm": 0.06407734751701355,
      "learning_rate": 5.370697469118713e-06,
      "loss": 0.0456,
      "mean_token_accuracy": 0.9805005967617035,
      "step": 1700
    },
    {
      "epoch": 0.9081225033288948,
      "grad_norm": 0.05884678661823273,
      "learning_rate": 5.074110594254133e-06,
      "loss": 0.0475,
      "mean_token_accuracy": 0.9797313630580902,
      "step": 1705
    },
    {
      "epoch": 0.9107856191744341,
      "grad_norm": 0.06042284518480301,
      "learning_rate": 4.78573407559928e-06,
      "loss": 0.0472,
      "mean_token_accuracy": 0.979413902759552,
      "step": 1710
    },
    {
      "epoch": 0.9134487350199734,
      "grad_norm": 0.05121051147580147,
      "learning_rate": 4.5055928554955665e-06,
      "loss": 0.0379,
      "mean_token_accuracy": 0.9844688534736633,
      "step": 1715
    },
    {
      "epoch": 0.9161118508655126,
      "grad_norm": 0.0401000939309597,
      "learning_rate": 4.233711163994669e-06,
      "loss": 0.0368,
      "mean_token_accuracy": 0.9846520006656647,
      "step": 1720
    },
    {
      "epoch": 0.918774966711052,
      "grad_norm": 0.046601567417383194,
      "learning_rate": 3.970112516762825e-06,
      "loss": 0.0496,
      "mean_token_accuracy": 0.9794139087200164,
      "step": 1725
    },
    {
      "epoch": 0.9214380825565912,
      "grad_norm": 0.046900924295186996,
      "learning_rate": 3.7148197130469576e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9858485758304596,
      "step": 1730
    },
    {
      "epoch": 0.9241011984021305,
      "grad_norm": 0.04627379775047302,
      "learning_rate": 3.467854833702644e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.9845298945903778,
      "step": 1735
    },
    {
      "epoch": 0.9267643142476698,
      "grad_norm": 0.035060226917266846,
      "learning_rate": 3.229239239284354e-06,
      "loss": 0.0419,
      "mean_token_accuracy": 0.9822954714298249,
      "step": 1740
    },
    {
      "epoch": 0.929427430093209,
      "grad_norm": 0.055220045149326324,
      "learning_rate": 2.9989935681979164e-06,
      "loss": 0.0441,
      "mean_token_accuracy": 0.9810866773128509,
      "step": 1745
    },
    {
      "epoch": 0.9320905459387483,
      "grad_norm": 0.075019471347332,
      "learning_rate": 2.777137734915403e-06,
      "loss": 0.0456,
      "mean_token_accuracy": 0.9808424830436706,
      "step": 1750
    },
    {
      "epoch": 0.9347536617842876,
      "grad_norm": 0.05054498836398125,
      "learning_rate": 2.563690928252749e-06,
      "loss": 0.039,
      "mean_token_accuracy": 0.9836629867553711,
      "step": 1755
    },
    {
      "epoch": 0.9374167776298269,
      "grad_norm": 0.06797026097774506,
      "learning_rate": 2.358671609710017e-06,
      "loss": 0.0401,
      "mean_token_accuracy": 0.9828693389892578,
      "step": 1760
    },
    {
      "epoch": 0.9400798934753661,
      "grad_norm": 0.0382985956966877,
      "learning_rate": 2.1620975118746835e-06,
      "loss": 0.0432,
      "mean_token_accuracy": 0.981636130809784,
      "step": 1765
    },
    {
      "epoch": 0.9427430093209055,
      "grad_norm": 0.04903360456228256,
      "learning_rate": 1.9739856368878096e-06,
      "loss": 0.0385,
      "mean_token_accuracy": 0.9839194059371948,
      "step": 1770
    },
    {
      "epoch": 0.9454061251664447,
      "grad_norm": 0.052963729947805405,
      "learning_rate": 1.794352254973597e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.985384601354599,
      "step": 1775
    },
    {
      "epoch": 0.948069241011984,
      "grad_norm": 0.06124390289187431,
      "learning_rate": 1.6232129030320453e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.9850793480873108,
      "step": 1780
    },
    {
      "epoch": 0.9507323568575233,
      "grad_norm": 0.039376676082611084,
      "learning_rate": 1.4605823832951948e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9882173299789428,
      "step": 1785
    },
    {
      "epoch": 0.9533954727030626,
      "grad_norm": 0.06673986464738846,
      "learning_rate": 1.3064747620468054e-06,
      "loss": 0.0416,
      "mean_token_accuracy": 0.9819657981395722,
      "step": 1790
    },
    {
      "epoch": 0.9560585885486018,
      "grad_norm": 0.05445706844329834,
      "learning_rate": 1.1609033684057857e-06,
      "loss": 0.0486,
      "mean_token_accuracy": 0.9793284356594085,
      "step": 1795
    },
    {
      "epoch": 0.9587217043941412,
      "grad_norm": 0.04978228360414505,
      "learning_rate": 1.0238807931732487e-06,
      "loss": 0.0459,
      "mean_token_accuracy": 0.9800854563713074,
      "step": 1800
    },
    {
      "epoch": 0.9613848202396804,
      "grad_norm": 0.056231539696455,
      "learning_rate": 8.95418887743571e-07,
      "loss": 0.0406,
      "mean_token_accuracy": 0.9825762987136841,
      "step": 1805
    },
    {
      "epoch": 0.9640479360852197,
      "grad_norm": 0.0571223720908165,
      "learning_rate": 7.75528763079314e-07,
      "loss": 0.0451,
      "mean_token_accuracy": 0.980451762676239,
      "step": 1810
    },
    {
      "epoch": 0.966711051930759,
      "grad_norm": 0.06353303045034409,
      "learning_rate": 6.642207887502027e-07,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9854822874069213,
      "step": 1815
    },
    {
      "epoch": 0.9693741677762983,
      "grad_norm": 0.04823755845427513,
      "learning_rate": 5.615045920362549e-07,
      "loss": 0.03,
      "mean_token_accuracy": 0.987533563375473,
      "step": 1820
    },
    {
      "epoch": 0.9720372836218375,
      "grad_norm": 0.052398230880498886,
      "learning_rate": 4.673890570951023e-07,
      "loss": 0.0423,
      "mean_token_accuracy": 0.981550657749176,
      "step": 1825
    },
    {
      "epoch": 0.9747003994673769,
      "grad_norm": 0.056029096245765686,
      "learning_rate": 3.8188232419352764e-07,
      "loss": 0.0403,
      "mean_token_accuracy": 0.9830036520957947,
      "step": 1830
    },
    {
      "epoch": 0.9773635153129161,
      "grad_norm": 0.04497731104493141,
      "learning_rate": 3.049917890034837e-07,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9856898486614227,
      "step": 1835
    },
    {
      "epoch": 0.9800266311584553,
      "grad_norm": 0.03512905538082123,
      "learning_rate": 2.3672410196232675e-07,
      "loss": 0.0407,
      "mean_token_accuracy": 0.9827838599681854,
      "step": 1840
    },
    {
      "epoch": 0.9826897470039947,
      "grad_norm": 0.05528529733419418,
      "learning_rate": 1.7708516769769924e-07,
      "loss": 0.0404,
      "mean_token_accuracy": 0.9827594459056854,
      "step": 1845
    },
    {
      "epoch": 0.9853528628495339,
      "grad_norm": 0.05035589635372162,
      "learning_rate": 1.2608014451672702e-07,
      "loss": 0.0419,
      "mean_token_accuracy": 0.982686185836792,
      "step": 1850
    },
    {
      "epoch": 0.9880159786950732,
      "grad_norm": 0.04821354150772095,
      "learning_rate": 8.371344395996516e-08,
      "loss": 0.0402,
      "mean_token_accuracy": 0.9832600593566895,
      "step": 1855
    },
    {
      "epoch": 0.9906790945406125,
      "grad_norm": 0.06377455592155457,
      "learning_rate": 4.998873041975882e-08,
      "loss": 0.045,
      "mean_token_accuracy": 0.9806715369224548,
      "step": 1860
    },
    {
      "epoch": 0.9933422103861518,
      "grad_norm": 0.05035337433218956,
      "learning_rate": 2.490892082331886e-08,
      "loss": 0.0428,
      "mean_token_accuracy": 0.9816605448722839,
      "step": 1865
    },
    {
      "epoch": 0.996005326231691,
      "grad_norm": 0.04119347408413887,
      "learning_rate": 8.476184380468155e-09,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9854578614234925,
      "step": 1870
    },
    {
      "epoch": 0.9986684420772304,
      "grad_norm": 0.04923999309539795,
      "learning_rate": 6.919423959805826e-10,
      "loss": 0.0373,
      "mean_token_accuracy": 0.984322327375412,
      "step": 1875
    },
    {
      "epoch": 0.9997336884154461,
      "eval_loss": 0.04217471182346344,
      "eval_mean_token_accuracy": 0.982079415995975,
      "eval_runtime": 981.8954,
      "eval_samples_per_second": 3.158,
      "eval_steps_per_second": 1.58,
      "step": 1877
    },
    {
      "epoch": 0.9997336884154461,
      "step": 1877,
      "total_flos": 2.4518665439700582e+17,
      "train_loss": 0.05413007871165959,
      "train_runtime": 8945.5939,
      "train_samples_per_second": 0.839,
      "train_steps_per_second": 0.21
    }
  ],
  "logging_steps": 5,
  "max_steps": 1877,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.4518665439700582e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}