{
  "best_global_step": 500,
  "best_metric": 0.09908278286457062,
  "best_model_checkpoint": "/workspace/anchor/qwen-pretrained/train/checkpoint-500",
  "epoch": 1.0,
  "eval_steps": 50,
  "global_step": 538,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.009306654257794323,
      "grad_norm": 8.3795747756958,
      "learning_rate": 2.962962962962963e-05,
      "loss": 1.0965,
      "step": 5
    },
    {
      "epoch": 0.018613308515588647,
      "grad_norm": 3.7024037837982178,
      "learning_rate": 6.666666666666667e-05,
      "loss": 0.8571,
      "step": 10
    },
    {
      "epoch": 0.02791996277338297,
      "grad_norm": 2.2940096855163574,
      "learning_rate": 0.0001037037037037037,
      "loss": 0.629,
      "step": 15
    },
    {
      "epoch": 0.03722661703117729,
      "grad_norm": 2.2930526733398438,
      "learning_rate": 0.00014074074074074076,
      "loss": 0.527,
      "step": 20
    },
    {
      "epoch": 0.04653327128897161,
      "grad_norm": 4.794444561004639,
      "learning_rate": 0.00017777777777777779,
      "loss": 0.5039,
      "step": 25
    },
    {
      "epoch": 0.05583992554676594,
      "grad_norm": 3.5820388793945312,
      "learning_rate": 0.0001999924406845452,
      "loss": 0.4196,
      "step": 30
    },
    {
      "epoch": 0.06514657980456026,
      "grad_norm": 3.9649713039398193,
      "learning_rate": 0.00019990741151022301,
      "loss": 0.3942,
      "step": 35
    },
    {
      "epoch": 0.07445323406235459,
      "grad_norm": 2.7851450443267822,
      "learning_rate": 0.00019972798462510112,
      "loss": 0.3424,
      "step": 40
    },
    {
      "epoch": 0.08375988832014891,
      "grad_norm": 2.488961935043335,
      "learning_rate": 0.0001994543295610186,
      "loss": 0.2828,
      "step": 45
    },
    {
      "epoch": 0.09306654257794322,
      "grad_norm": 6.255297660827637,
      "learning_rate": 0.00019908670488149148,
      "loss": 0.2322,
      "step": 50
    },
    {
      "epoch": 0.09306654257794322,
      "eval_loss": 0.22483031451702118,
      "eval_runtime": 157.1846,
      "eval_samples_per_second": 1.406,
      "eval_steps_per_second": 0.706,
      "step": 50
    },
    {
      "epoch": 0.10237319683573755,
      "grad_norm": 1.058211326599121,
      "learning_rate": 0.00019862545793740852,
      "loss": 0.2246,
      "step": 55
    },
    {
      "epoch": 0.11167985109353187,
      "grad_norm": 1.2681463956832886,
      "learning_rate": 0.00019807102453883602,
      "loss": 0.2238,
      "step": 60
    },
    {
      "epoch": 0.1209865053513262,
      "grad_norm": 1.2710065841674805,
      "learning_rate": 0.00019742392854324185,
      "loss": 0.2353,
      "step": 65
    },
    {
      "epoch": 0.13029315960912052,
      "grad_norm": 1.0538429021835327,
      "learning_rate": 0.00019668478136052774,
      "loss": 0.21,
      "step": 70
    },
    {
      "epoch": 0.13959981386691483,
      "grad_norm": 1.3220242261886597,
      "learning_rate": 0.00019585428137533744,
      "loss": 0.2222,
      "step": 75
    },
    {
      "epoch": 0.14890646812470917,
      "grad_norm": 1.1070969104766846,
      "learning_rate": 0.00019493321328718652,
      "loss": 0.1745,
      "step": 80
    },
    {
      "epoch": 0.15821312238250348,
      "grad_norm": 1.2554668188095093,
      "learning_rate": 0.00019392244736903772,
      "loss": 0.1952,
      "step": 85
    },
    {
      "epoch": 0.16751977664029782,
      "grad_norm": 1.53813898563385,
      "learning_rate": 0.00019282293864502178,
      "loss": 0.1863,
      "step": 90
    },
    {
      "epoch": 0.17682643089809214,
      "grad_norm": 1.1727522611618042,
      "learning_rate": 0.0001916357259880809,
      "loss": 0.1683,
      "step": 95
    },
    {
      "epoch": 0.18613308515588645,
      "grad_norm": 0.9360535144805908,
      "learning_rate": 0.0001903619311383879,
      "loss": 0.1681,
      "step": 100
    },
    {
      "epoch": 0.18613308515588645,
      "eval_loss": 0.14435869455337524,
      "eval_runtime": 72.7707,
      "eval_samples_per_second": 3.037,
      "eval_steps_per_second": 1.525,
      "step": 100
    },
    {
      "epoch": 0.19543973941368079,
      "grad_norm": 1.1460329294204712,
      "learning_rate": 0.00018900275764346768,
      "loss": 0.1456,
      "step": 105
    },
    {
      "epoch": 0.2047463936714751,
      "grad_norm": 1.0264545679092407,
      "learning_rate": 0.00018755948972102294,
      "loss": 0.1788,
      "step": 110
    },
    {
      "epoch": 0.21405304792926944,
      "grad_norm": 0.9038275480270386,
      "learning_rate": 0.00018603349104553884,
      "loss": 0.1864,
      "step": 115
    },
    {
      "epoch": 0.22335970218706375,
      "grad_norm": 0.8495742082595825,
      "learning_rate": 0.00018442620345981232,
      "loss": 0.1442,
      "step": 120
    },
    {
      "epoch": 0.23266635644485809,
      "grad_norm": 0.7527527213096619,
      "learning_rate": 0.00018273914561262424,
      "loss": 0.1555,
      "step": 125
    },
    {
      "epoch": 0.2419730107026524,
      "grad_norm": 0.6647666692733765,
      "learning_rate": 0.00018097391152384115,
      "loss": 0.1383,
      "step": 130
    },
    {
      "epoch": 0.25127966496044674,
      "grad_norm": 0.8701847195625305,
      "learning_rate": 0.0001791321690783025,
      "loss": 0.1671,
      "step": 135
    },
    {
      "epoch": 0.26058631921824105,
      "grad_norm": 0.6835318803787231,
      "learning_rate": 0.00017721565844991643,
      "loss": 0.1458,
      "step": 140
    },
    {
      "epoch": 0.26989297347603536,
      "grad_norm": 0.8638496994972229,
      "learning_rate": 0.00017522619045745315,
      "loss": 0.163,
      "step": 145
    },
    {
      "epoch": 0.27919962773382967,
      "grad_norm": 1.0149871110916138,
      "learning_rate": 0.00017316564485358923,
      "loss": 0.1518,
      "step": 150
    },
    {
      "epoch": 0.27919962773382967,
      "eval_loss": 0.12961792945861816,
      "eval_runtime": 72.8727,
      "eval_samples_per_second": 3.033,
      "eval_steps_per_second": 1.523,
      "step": 150
    },
    {
      "epoch": 0.28850628199162404,
      "grad_norm": 0.8869348168373108,
      "learning_rate": 0.00017103596854881972,
      "loss": 0.161,
      "step": 155
    },
    {
      "epoch": 0.29781293624941835,
      "grad_norm": 0.6964899897575378,
      "learning_rate": 0.00016883917377191602,
      "loss": 0.1496,
      "step": 160
    },
    {
      "epoch": 0.30711959050721266,
      "grad_norm": 0.7895644307136536,
      "learning_rate": 0.00016657733616866756,
      "loss": 0.178,
      "step": 165
    },
    {
      "epoch": 0.31642624476500697,
      "grad_norm": 0.8652212023735046,
      "learning_rate": 0.00016425259284070398,
      "loss": 0.1572,
      "step": 170
    },
    {
      "epoch": 0.3257328990228013,
      "grad_norm": 0.8598363995552063,
      "learning_rate": 0.00016186714032625035,
      "loss": 0.1329,
      "step": 175
    },
    {
      "epoch": 0.33503955328059565,
      "grad_norm": 0.9525418877601624,
      "learning_rate": 0.000159423232524724,
      "loss": 0.1509,
      "step": 180
    },
    {
      "epoch": 0.34434620753838996,
      "grad_norm": 0.8453659415245056,
      "learning_rate": 0.0001569231785671332,
      "loss": 0.1619,
      "step": 185
    },
    {
      "epoch": 0.35365286179618427,
      "grad_norm": 0.9399647116661072,
      "learning_rate": 0.00015436934063429013,
      "loss": 0.1637,
      "step": 190
    },
    {
      "epoch": 0.3629595160539786,
      "grad_norm": 0.8372876048088074,
      "learning_rate": 0.00015176413172490004,
      "loss": 0.1622,
      "step": 195
    },
    {
      "epoch": 0.3722661703117729,
      "grad_norm": 0.711920976638794,
      "learning_rate": 0.00014911001337563452,
      "loss": 0.1504,
      "step": 200
    },
    {
      "epoch": 0.3722661703117729,
      "eval_loss": 0.12358199059963226,
      "eval_runtime": 73.0008,
      "eval_samples_per_second": 3.027,
      "eval_steps_per_second": 1.521,
      "step": 200
    },
    {
      "epoch": 0.38157282456956726,
      "grad_norm": 0.8126468658447266,
      "learning_rate": 0.00014640949333534403,
      "loss": 0.1526,
      "step": 205
    },
    {
      "epoch": 0.39087947882736157,
      "grad_norm": 0.635465681552887,
      "learning_rate": 0.0001436651231956064,
      "loss": 0.1358,
      "step": 210
    },
    {
      "epoch": 0.4001861330851559,
      "grad_norm": 0.5986313819885254,
      "learning_rate": 0.00014087949597985065,
      "loss": 0.15,
      "step": 215
    },
    {
      "epoch": 0.4094927873429502,
      "grad_norm": 0.7797956466674805,
      "learning_rate": 0.00013805524369333373,
      "loss": 0.1442,
      "step": 220
    },
    {
      "epoch": 0.41879944160074456,
      "grad_norm": 0.6968953609466553,
      "learning_rate": 0.00013519503483628542,
      "loss": 0.1437,
      "step": 225
    },
    {
      "epoch": 0.42810609585853887,
      "grad_norm": 0.6117879152297974,
      "learning_rate": 0.00013230157188257063,
      "loss": 0.1214,
      "step": 230
    },
    {
      "epoch": 0.4374127501163332,
      "grad_norm": 0.6684589982032776,
      "learning_rate": 0.0001293775887262517,
      "loss": 0.1383,
      "step": 235
    },
    {
      "epoch": 0.4467194043741275,
      "grad_norm": 0.7364048957824707,
      "learning_rate": 0.00012642584809846333,
      "loss": 0.1222,
      "step": 240
    },
    {
      "epoch": 0.4560260586319218,
      "grad_norm": 0.7387944459915161,
      "learning_rate": 0.00012344913895704097,
      "loss": 0.1202,
      "step": 245
    },
    {
      "epoch": 0.46533271288971617,
      "grad_norm": 0.6471113562583923,
      "learning_rate": 0.00012045027385136825,
      "loss": 0.1446,
      "step": 250
    },
    {
      "epoch": 0.46533271288971617,
      "eval_loss": 0.11554089933633804,
      "eval_runtime": 73.039,
      "eval_samples_per_second": 3.026,
      "eval_steps_per_second": 1.52,
      "step": 250
    },
    {
      "epoch": 0.4746393671475105,
      "grad_norm": 0.7449938058853149,
      "learning_rate": 0.0001174320862649344,
      "loss": 0.1425,
      "step": 255
    },
    {
      "epoch": 0.4839460214053048,
      "grad_norm": 0.8666985630989075,
      "learning_rate": 0.00011439742793811205,
      "loss": 0.131,
      "step": 260
    },
    {
      "epoch": 0.4932526756630991,
      "grad_norm": 0.9995303750038147,
      "learning_rate": 0.00011134916617368464,
      "loss": 0.1322,
      "step": 265
    },
    {
      "epoch": 0.5025593299208935,
      "grad_norm": 0.6491286754608154,
      "learning_rate": 0.00010829018112766996,
      "loss": 0.1293,
      "step": 270
    },
    {
      "epoch": 0.5118659841786878,
      "grad_norm": 0.8522428274154663,
      "learning_rate": 0.00010522336308799904,
      "loss": 0.1372,
      "step": 275
    },
    {
      "epoch": 0.5211726384364821,
      "grad_norm": 0.7563990354537964,
      "learning_rate": 0.00010215160974362223,
      "loss": 0.1277,
      "step": 280
    },
    {
      "epoch": 0.5304792926942764,
      "grad_norm": 0.6848676800727844,
      "learning_rate": 9.907782344662195e-05,
      "loss": 0.1294,
      "step": 285
    },
    {
      "epoch": 0.5397859469520707,
      "grad_norm": 0.6876094341278076,
      "learning_rate": 9.600490846991994e-05,
      "loss": 0.1383,
      "step": 290
    },
    {
      "epoch": 0.549092601209865,
      "grad_norm": 0.6399123668670654,
      "learning_rate": 9.29357682631691e-05,
      "loss": 0.1241,
      "step": 295
    },
    {
      "epoch": 0.5583992554676593,
      "grad_norm": 0.6791756749153137,
      "learning_rate": 8.987330270942389e-05,
      "loss": 0.14,
      "step": 300
    },
    {
      "epoch": 0.5583992554676593,
      "eval_loss": 0.1100185289978981,
      "eval_runtime": 72.8167,
      "eval_samples_per_second": 3.035,
      "eval_steps_per_second": 1.524,
      "step": 300
    },
    {
      "epoch": 0.5677059097254537,
      "grad_norm": 0.8283066749572754,
      "learning_rate": 8.682040538517974e-05,
      "loss": 0.1346,
      "step": 305
    },
    {
      "epoch": 0.5770125639832481,
      "grad_norm": 0.7044147253036499,
      "learning_rate": 8.377996082637216e-05,
      "loss": 0.1386,
      "step": 310
    },
    {
      "epoch": 0.5863192182410424,
      "grad_norm": 0.696846067905426,
      "learning_rate": 8.075484180291701e-05,
      "loss": 0.1173,
      "step": 315
    },
    {
      "epoch": 0.5956258724988367,
      "grad_norm": 0.7000551819801331,
      "learning_rate": 7.774790660436858e-05,
      "loss": 0.1319,
      "step": 320
    },
    {
      "epoch": 0.604932526756631,
      "grad_norm": 0.8282796144485474,
      "learning_rate": 7.476199633925895e-05,
      "loss": 0.1373,
      "step": 325
    },
    {
      "epoch": 0.6142391810144253,
      "grad_norm": 0.6185778379440308,
      "learning_rate": 7.179993225067136e-05,
      "loss": 0.1376,
      "step": 330
    },
    {
      "epoch": 0.6235458352722196,
      "grad_norm": 0.6192060112953186,
      "learning_rate": 6.886451305058293e-05,
      "loss": 0.1296,
      "step": 335
    },
    {
      "epoch": 0.6328524895300139,
      "grad_norm": 0.5297410488128662,
      "learning_rate": 6.595851227549657e-05,
      "loss": 0.1114,
      "step": 340
    },
    {
      "epoch": 0.6421591437878083,
      "grad_norm": 0.6853290796279907,
      "learning_rate": 6.308467566585929e-05,
      "loss": 0.1258,
      "step": 345
    },
    {
      "epoch": 0.6514657980456026,
      "grad_norm": 0.5415284633636475,
      "learning_rate": 6.024571857174443e-05,
      "loss": 0.1151,
      "step": 350
    },
    {
      "epoch": 0.6514657980456026,
      "eval_loss": 0.10505883395671844,
      "eval_runtime": 72.7668,
      "eval_samples_per_second": 3.037,
      "eval_steps_per_second": 1.525,
      "step": 350
    },
    {
      "epoch": 0.6607724523033969,
      "grad_norm": 0.6376184225082397,
      "learning_rate": 5.7444323387247544e-05,
      "loss": 0.1196,
      "step": 355
    },
    {
      "epoch": 0.6700791065611913,
      "grad_norm": 0.9076871871948242,
      "learning_rate": 5.468313701602141e-05,
      "loss": 0.1199,
      "step": 360
    },
    {
      "epoch": 0.6793857608189856,
      "grad_norm": 0.8146594166755676,
      "learning_rate": 5.196476837034381e-05,
      "loss": 0.1278,
      "step": 365
    },
    {
      "epoch": 0.6886924150767799,
      "grad_norm": 0.7426274418830872,
      "learning_rate": 4.9291785906081925e-05,
      "loss": 0.1356,
      "step": 370
    },
    {
      "epoch": 0.6979990693345742,
      "grad_norm": 0.8531163334846497,
      "learning_rate": 4.6666715195881586e-05,
      "loss": 0.1322,
      "step": 375
    },
    {
      "epoch": 0.7073057235923685,
      "grad_norm": 0.6669689416885376,
      "learning_rate": 4.409203654287538e-05,
      "loss": 0.1228,
      "step": 380
    },
    {
      "epoch": 0.7166123778501629,
      "grad_norm": 0.7575223445892334,
      "learning_rate": 4.1570182637163155e-05,
      "loss": 0.1384,
      "step": 385
    },
    {
      "epoch": 0.7259190321079572,
      "grad_norm": 0.677855908870697,
      "learning_rate": 3.910353625728035e-05,
      "loss": 0.1195,
      "step": 390
    },
    {
      "epoch": 0.7352256863657515,
      "grad_norm": 0.6284500360488892,
      "learning_rate": 3.66944280188248e-05,
      "loss": 0.119,
      "step": 395
    },
    {
      "epoch": 0.7445323406235458,
      "grad_norm": 0.5671315789222717,
      "learning_rate": 3.434513417237041e-05,
      "loss": 0.1159,
      "step": 400
    },
    {
      "epoch": 0.7445323406235458,
      "eval_loss": 0.10208839178085327,
      "eval_runtime": 72.8321,
      "eval_samples_per_second": 3.034,
      "eval_steps_per_second": 1.524,
      "step": 400
    },
    {
      "epoch": 0.7538389948813402,
      "grad_norm": 0.6890803575515747,
      "learning_rate": 3.205787445274707e-05,
      "loss": 0.1254,
      "step": 405
    },
    {
      "epoch": 0.7631456491391345,
      "grad_norm": 0.6065118312835693,
      "learning_rate": 2.9834809981720224e-05,
      "loss": 0.1076,
      "step": 410
    },
    {
      "epoch": 0.7724523033969288,
      "grad_norm": 0.734563946723938,
      "learning_rate": 2.7678041226050533e-05,
      "loss": 0.1283,
      "step": 415
    },
    {
      "epoch": 0.7817589576547231,
      "grad_norm": 0.6849725842475891,
      "learning_rate": 2.5589606012863963e-05,
      "loss": 0.1235,
      "step": 420
    },
    {
      "epoch": 0.7910656119125175,
      "grad_norm": 0.7363895773887634,
      "learning_rate": 2.3571477604206792e-05,
      "loss": 0.1329,
      "step": 425
    },
    {
      "epoch": 0.8003722661703118,
      "grad_norm": 0.6717341542243958,
      "learning_rate": 2.1625562832605284e-05,
      "loss": 0.1339,
      "step": 430
    },
    {
      "epoch": 0.8096789204281061,
      "grad_norm": 0.6806132197380066,
      "learning_rate": 1.975370029939111e-05,
      "loss": 0.1198,
      "step": 435
    },
    {
      "epoch": 0.8189855746859004,
      "grad_norm": 0.7354431748390198,
      "learning_rate": 1.795765863749549e-05,
      "loss": 0.1436,
      "step": 440
    },
    {
      "epoch": 0.8282922289436947,
      "grad_norm": 0.5758335590362549,
      "learning_rate": 1.623913484035282e-05,
      "loss": 0.108,
      "step": 445
    },
    {
      "epoch": 0.8375988832014891,
      "grad_norm": 0.5895151495933533,
      "learning_rate": 1.4599752658493304e-05,
      "loss": 0.1384,
      "step": 450
    },
    {
      "epoch": 0.8375988832014891,
      "eval_loss": 0.10044225305318832,
      "eval_runtime": 72.8277,
      "eval_samples_per_second": 3.035,
      "eval_steps_per_second": 1.524,
      "step": 450
    },
    {
      "epoch": 0.8469055374592834,
      "grad_norm": 0.5299936532974243,
      "learning_rate": 1.30410610653389e-05,
      "loss": 0.1082,
      "step": 455
    },
    {
      "epoch": 0.8562121917170777,
      "grad_norm": 0.6784066557884216,
      "learning_rate": 1.1564532793653072e-05,
      "loss": 0.1279,
      "step": 460
    },
    {
      "epoch": 0.865518845974872,
      "grad_norm": 0.7089002728462219,
      "learning_rate": 1.0171562944026103e-05,
      "loss": 0.121,
      "step": 465
    },
    {
      "epoch": 0.8748255002326664,
      "grad_norm": 0.7150669097900391,
      "learning_rate": 8.863467666711866e-06,
      "loss": 0.1,
      "step": 470
    },
    {
      "epoch": 0.8841321544904607,
      "grad_norm": 0.7433251142501831,
      "learning_rate": 7.641482918060506e-06,
      "loss": 0.1111,
      "step": 475
    },
    {
      "epoch": 0.893438808748255,
      "grad_norm": 0.6310339570045471,
      "learning_rate": 6.506763292722851e-06,
      "loss": 0.0885,
      "step": 480
    },
    {
      "epoch": 0.9027454630060493,
      "grad_norm": 0.5972334742546082,
      "learning_rate": 5.460380932729303e-06,
      "loss": 0.1061,
      "step": 485
    },
    {
      "epoch": 0.9120521172638436,
      "grad_norm": 0.6366889476776123,
      "learning_rate": 4.503324514474483e-06,
      "loss": 0.1347,
      "step": 490
    },
    {
      "epoch": 0.9213587715216379,
      "grad_norm": 0.7572676539421082,
      "learning_rate": 3.636498314564307e-06,
      "loss": 0.1174,
      "step": 495
    },
    {
      "epoch": 0.9306654257794323,
      "grad_norm": 0.6025388836860657,
      "learning_rate": 2.860721355408602e-06,
      "loss": 0.1112,
      "step": 500
    },
    {
      "epoch": 0.9306654257794323,
      "eval_loss": 0.09908278286457062,
      "eval_runtime": 72.8525,
      "eval_samples_per_second": 3.034,
      "eval_steps_per_second": 1.524,
      "step": 500
    },
    {
      "epoch": 0.9399720800372267,
      "grad_norm": 0.6565231680870056,
      "learning_rate": 2.1767266313661105e-06,
      "loss": 0.1077,
      "step": 505
    },
    {
      "epoch": 0.949278734295021,
      "grad_norm": 0.5436176657676697,
      "learning_rate": 1.5851604161734256e-06,
      "loss": 0.1193,
      "step": 510
    },
    {
      "epoch": 0.9585853885528153,
      "grad_norm": 0.8013174533843994,
      "learning_rate": 1.0865816523119466e-06,
      "loss": 0.1193,
      "step": 515
    },
    {
      "epoch": 0.9678920428106096,
      "grad_norm": 0.6958430409431458,
      "learning_rate": 6.814614228900507e-07,
      "loss": 0.1152,
      "step": 520
    },
    {
      "epoch": 0.9771986970684039,
      "grad_norm": 0.5581568479537964,
      "learning_rate": 3.701825065392184e-07,
      "loss": 0.1249,
      "step": 525
    },
    {
      "epoch": 0.9865053513261982,
      "grad_norm": 0.7202214002609253,
      "learning_rate": 1.5303901574502678e-07,
      "loss": 0.1164,
      "step": 530
    },
    {
      "epoch": 0.9958120055839925,
      "grad_norm": 0.5646489262580872,
      "learning_rate": 3.023611895428546e-08,
      "loss": 0.1061,
      "step": 535
    }
  ],
  "logging_steps": 5,
  "max_steps": 538,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 2,
        "early_stopping_threshold": 0.001
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.3722943184871096e+18,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}