{"event": "train", "step": 20, "loss": -6.127182960510254, "grad_norm": 0.33450567722320557, "lr": 1.5749999999999997e-05, "ex_per_s": 98.44207606719083, "data_frac": 0.8422720855367718, "mem_mb": 63.172096, "nll": -6.127182960510254, "elapsed_s": 13.0} {"event": "train", "step": 40, "loss": -5.890166759490967, "grad_norm": 0.8092024326324463, "lr": 3.0749999999999995e-05, "ex_per_s": 174.58212590277532, "data_frac": 0.7476405950441517, "mem_mb": 63.172096, "nll": -5.890166759490967, "elapsed_s": 14.7} {"event": "train", "step": 60, "loss": -5.977727890014648, "grad_norm": 0.7138806581497192, "lr": 4.5749999999999994e-05, "ex_per_s": 235.33028225588345, "data_frac": 0.6724744004728661, "mem_mb": 63.172096, "nll": -5.977727890014648, "elapsed_s": 16.4} {"event": "train", "step": 80, "loss": -6.0802412033081055, "grad_norm": 0.21126462519168854, "lr": 6.075e-05, "ex_per_s": 284.8776163074289, "data_frac": 0.6111053051792776, "mem_mb": 63.172096, "nll": -6.0802412033081055, "elapsed_s": 18.0} {"event": "train", "step": 100, "loss": -5.948199272155762, "grad_norm": 0.37521979212760925, "lr": 7.575e-05, "ex_per_s": 326.05510565462293, "data_frac": 0.5601211627872854, "mem_mb": 63.172096, "nll": -5.948199272155762, "elapsed_s": 19.7} {"event": "train", "step": 120, "loss": -5.903069972991943, "grad_norm": 1.2512739896774292, "lr": 9.074999999999998e-05, "ex_per_s": 360.8550786231487, "data_frac": 0.517105025047594, "mem_mb": 63.172096, "nll": -5.903069972991943, "elapsed_s": 21.3} {"event": "train", "step": 140, "loss": -5.832436561584473, "grad_norm": 0.16964206099510193, "lr": 0.00010574999999999998, "ex_per_s": 390.63874615787336, "data_frac": 0.48028199111549363, "mem_mb": 63.172096, "nll": -5.832436561584473, "elapsed_s": 23.0} {"event": "train", "step": 160, "loss": -5.735413074493408, "grad_norm": 0.6712958812713623, "lr": 0.00012075, "ex_per_s": 416.39552722454437, "data_frac": 0.4483793394270447, "mem_mb": 63.172096, "nll": -5.735413074493408, "elapsed_s": 24.6} {"event": "train", "step": 180, "loss": -5.9725728034973145, "grad_norm": 1.1546398401260376, "lr": 0.00013575, "ex_per_s": 438.90193776698055, "data_frac": 0.42053162269450794, "mem_mb": 63.172096, "nll": -5.9725728034973145, "elapsed_s": 26.3} {"event": "train", "step": 200, "loss": -5.967144012451172, "grad_norm": 0.9231442213058472, "lr": 0.00015074999999999998, "ex_per_s": 458.71262082059104, "data_frac": 0.39594165406156795, "mem_mb": 63.172096, "nll": -5.967144012451172, "elapsed_s": 27.9} {"event": "train", "step": 220, "loss": -6.027961730957031, "grad_norm": 1.1874021291732788, "lr": 0.00016575, "ex_per_s": 476.39037367456086, "data_frac": 0.374165174594, "mem_mb": 63.172096, "nll": -6.027961730957031, "elapsed_s": 29.6} {"event": "train", "step": 240, "loss": -5.913460731506348, "grad_norm": 0.3065013587474823, "lr": 0.00018075, "ex_per_s": 492.11718844331534, "data_frac": 0.3546592486275402, "mem_mb": 63.172096, "nll": -5.913460731506348, "elapsed_s": 31.3} {"event": "train", "step": 260, "loss": -5.887753963470459, "grad_norm": 0.1304401308298111, "lr": 0.00019574999999999996, "ex_per_s": 506.33592299022763, "data_frac": 0.3371476732115154, "mem_mb": 63.172096, "nll": -5.887753963470459, "elapsed_s": 32.9} {"event": "train", "step": 280, "loss": -5.943431377410889, "grad_norm": 0.33325210213661194, "lr": 0.00021074999999999997, "ex_per_s": 519.1775129457073, "data_frac": 0.3212955807480819, "mem_mb": 63.172096, "nll": -5.943431377410889, "elapsed_s": 34.6} {"event": "train", "step": 300, "loss": -5.9670820236206055, "grad_norm": 0.31373217701911926, "lr": 0.00022574999999999996, "ex_per_s": 530.8560790477146, "data_frac": 0.3069035609502007, "mem_mb": 63.172096, "nll": -5.9670820236206055, "elapsed_s": 36.2} {"event": "train", "step": 320, "loss": -5.965109348297119, "grad_norm": 0.6371338367462158, "lr": 0.00024074999999999997, "ex_per_s": 541.5173815857951, "data_frac": 0.29376747959236177, "mem_mb": 63.172096, "nll": -5.965109348297119, "elapsed_s": 37.9} {"event": "train", "step": 340, "loss": -5.834346294403076, "grad_norm": 0.5424937009811401, "lr": 0.00025575, "ex_per_s": 551.2807919791194, "data_frac": 0.2817307166981207, "mem_mb": 63.172096, "nll": -5.834346294403076, "elapsed_s": 39.5} {"event": "train", "step": 360, "loss": -5.995899677276611, "grad_norm": 0.34699392318725586, "lr": 0.00027074999999999994, "ex_per_s": 560.2277824230845, "data_frac": 0.2706422284543595, "mem_mb": 63.172096, "nll": -5.995899677276611, "elapsed_s": 41.2} {"event": "train", "step": 380, "loss": -6.007436275482178, "grad_norm": 1.0209921598434448, "lr": 0.00028575, "ex_per_s": 568.4919544759127, "data_frac": 0.2604231475451951, "mem_mb": 63.172096, "nll": -6.007436275482178, "elapsed_s": 42.8} {"event": "train", "step": 400, "loss": -6.001394271850586, "grad_norm": 0.8656692504882812, "lr": 0.0003, "ex_per_s": 576.1679547286675, "data_frac": 0.25097119046531924, "mem_mb": 63.172096, "nll": -6.001394271850586, "elapsed_s": 44.5} {"event": "train", "step": 420, "loss": -5.960729122161865, "grad_norm": 1.0123273134231567, "lr": 0.0002999980196294046, "ex_per_s": 583.2266643398896, "data_frac": 0.24219293871331113, "mem_mb": 63.172096, "nll": -5.960729122161865, "elapsed_s": 46.1} {"event": "train", "step": 440, "loss": -5.937890529632568, "grad_norm": 0.7674927115440369, "lr": 0.00029999207857572027, "ex_per_s": 589.8527192528529, "data_frac": 0.2340227351856522, "mem_mb": 63.172096, "nll": -5.937890529632568, "elapsed_s": 47.8} {"event": "train", "step": 460, "loss": -6.101131916046143, "grad_norm": 0.7102874517440796, "lr": 0.00029998217701325043, "ex_per_s": 596.0258246510421, "data_frac": 0.22639651099675748, "mem_mb": 63.172096, "nll": -6.101131916046143, "elapsed_s": 49.4} {"event": "train", "step": 480, "loss": -5.87770938873291, "grad_norm": 0.22956396639347076, "lr": 0.00029996831523249534, "ex_per_s": 601.7982142192087, "data_frac": 0.21926301207171106, "mem_mb": 63.172096, "nll": -5.87770938873291, "elapsed_s": 51.1} {"event": "train", "step": 500, "loss": -5.9230194091796875, "grad_norm": 0.6002429723739624, "lr": 0.0002999504936401433, "ex_per_s": 607.2150182243141, "data_frac": 0.21257728900841413, "mem_mb": 63.172096, "nll": -5.9230194091796875, "elapsed_s": 52.7} {"event": "train", "step": 520, "loss": -6.040375709533691, "grad_norm": 0.28456932306289673, "lr": 0.0002999287127590589, "ex_per_s": 612.2784930174455, "data_frac": 0.20629425460594952, "mem_mb": 63.172096, "nll": -6.040375709533691, "elapsed_s": 54.4} {"event": "train", "step": 540, "loss": -5.8803253173828125, "grad_norm": 0.43941986560821533, "lr": 0.00029990297322826756, "ex_per_s": 616.990459482596, "data_frac": 0.20036575019630623, "mem_mb": 63.172096, "nll": -5.8803253173828125, "elapsed_s": 56.1} {"event": "train", "step": 560, "loss": -5.972140312194824, "grad_norm": 0.32396185398101807, "lr": 0.00029987327580293684, "ex_per_s": 621.4434636961395, "data_frac": 0.19478168306815868, "mem_mb": 63.172096, "nll": -5.972140312194824, "elapsed_s": 57.7} {"event": "train", "step": 580, "loss": -5.924406051635742, "grad_norm": 0.23648443818092346, "lr": 0.00029983962135435424, "ex_per_s": 625.6477395061276, "data_frac": 0.18951497402660306, "mem_mb": 63.172096, "nll": -5.924406051635742, "elapsed_s": 59.4} {"event": "train", "step": 600, "loss": -5.973592758178711, "grad_norm": 0.7366935014724731, "lr": 0.0002998020108699017, "ex_per_s": 629.6335881663848, "data_frac": 0.1845306505540851, "mem_mb": 63.172096, "nll": -5.973592758178711, "elapsed_s": 61.0} {"event": "train", "step": 620, "loss": -5.9527974128723145, "grad_norm": 0.35334017872810364, "lr": 0.0002997604454530266, "ex_per_s": 633.4293278585669, "data_frac": 0.17982774275279492, "mem_mb": 63.172096, "nll": -5.9527974128723145, "elapsed_s": 62.7} {"event": "train", "step": 640, "loss": -6.065530300140381, "grad_norm": 0.4902239739894867, "lr": 0.00029971492632320935, "ex_per_s": 637.0555592917241, "data_frac": 0.17536353752184955, "mem_mb": 63.172096, "nll": -6.065530300140381, "elapsed_s": 64.3} {"event": "train", "step": 660, "loss": -6.006502151489258, "grad_norm": 1.0235997438430786, "lr": 0.0002996654548159277, "ex_per_s": 640.5114390940281, "data_frac": 0.17113933649546795, "mem_mb": 63.172096, "nll": -6.006502151489258, "elapsed_s": 66.0} {"event": "train", "step": 680, "loss": -6.007025718688965, "grad_norm": 0.10780004411935806, "lr": 0.00029961203238261747, "ex_per_s": 643.8478425535486, "data_frac": 0.16712118916906335, "mem_mb": 63.172096, "nll": -6.007025718688965, "elapsed_s": 67.6} {"event": "train", "step": 700, "loss": -5.985134601593018, "grad_norm": 0.9787061214447021, "lr": 0.0002995546605906301, "ex_per_s": 647.0290942134907, "data_frac": 0.16329725007635712, "mem_mb": 63.172096, "nll": -5.985134601593018, "elapsed_s": 69.3} {"event": "train", "step": 720, "loss": -6.06289529800415, "grad_norm": 0.46976229548454285, "lr": 0.00029949334112318624, "ex_per_s": 650.0670187790329, "data_frac": 0.15965182310165238, "mem_mb": 63.172096, "nll": -6.06289529800415, "elapsed_s": 70.9} {"event": "train", "step": 740, "loss": -6.0331926345825195, "grad_norm": 0.2691230773925781, "lr": 0.0002994280757793271, "ex_per_s": 652.9854908547493, "data_frac": 0.15618410924979031, "mem_mb": 63.172096, "nll": -6.0331926345825195, "elapsed_s": 72.6} {"event": "train", "step": 760, "loss": -5.95484733581543, "grad_norm": 1.1793854236602783, "lr": 0.00029935886647386114, "ex_per_s": 655.7564017282845, "data_frac": 0.15285694053408616, "mem_mb": 63.172096, "nll": -5.95484733581543, "elapsed_s": 74.2} {"event": "train", "step": 780, "loss": -5.8956379890441895, "grad_norm": 0.10551509261131287, "lr": 0.00029928571523730785, "ex_per_s": 658.418212955693, "data_frac": 0.1496795075368318, "mem_mb": 63.172096, "nll": -5.8956379890441895, "elapsed_s": 75.9} {"event": "train", "step": 800, "loss": -5.815648555755615, "grad_norm": 0.21799392998218536, "lr": 0.00029920862421583855, "ex_per_s": 660.9735327530854, "data_frac": 0.1466353262384443, "mem_mb": 63.172096, "nll": -5.815648555755615, "elapsed_s": 77.5} {"event": "train", "step": 820, "loss": -5.933268070220947, "grad_norm": 0.29786232113838196, "lr": 0.00029912759567121304, "ex_per_s": 663.4208361343309, "data_frac": 0.14371654612939685, "mem_mb": 63.172096, "nll": -5.933268070220947, "elapsed_s": 79.1} {"event": "train", "step": 840, "loss": -5.806618690490723, "grad_norm": 0.5086231827735901, "lr": 0.0002990426319807135, "ex_per_s": 665.7649634123437, "data_frac": 0.1409169021357129, "mem_mb": 63.172096, "nll": -5.806618690490723, "elapsed_s": 80.8} {"event": "train", "step": 860, "loss": -5.9127631187438965, "grad_norm": 1.5683951377868652, "lr": 0.0002989537356370747, "ex_per_s": 668.0284869468009, "data_frac": 0.13823035998294095, "mem_mb": 63.172096, "nll": -5.9127631187438965, "elapsed_s": 82.4} {"event": "train", "step": 880, "loss": -5.868842124938965, "grad_norm": 0.41443803906440735, "lr": 0.00029886090924841085, "ex_per_s": 670.2000518341164, "data_frac": 0.13565097600524822, "mem_mb": 63.172096, "nll": -5.868842124938965, "elapsed_s": 84.1} {"event": "train", "step": 900, "loss": -5.985984802246094, "grad_norm": 0.4682462811470032, "lr": 0.0002987641555381389, "ex_per_s": 672.2550679009397, "data_frac": 0.13317458195833762, "mem_mb": 63.172096, "nll": -5.985984802246094, "elapsed_s": 85.7} {"event": "train", "step": 920, "loss": -6.024830341339111, "grad_norm": 0.27259010076522827, "lr": 0.0002986634773448989, "ex_per_s": 674.2511453736154, "data_frac": 0.1307875514832568, "mem_mb": 63.172096, "nll": -6.024830341339111, "elapsed_s": 87.4} {"event": "train", "step": 940, "loss": -6.077476501464844, "grad_norm": 1.1316993236541748, "lr": 0.00029855887762247066, "ex_per_s": 676.1801723937366, "data_frac": 0.12848517836076845, "mem_mb": 63.172096, "nll": -6.077476501464844, "elapsed_s": 89.0} {"event": "train", "step": 960, "loss": -5.9673919677734375, "grad_norm": 0.3903808891773224, "lr": 0.00029845035943968696, "ex_per_s": 678.0376518990547, "data_frac": 0.1262659880923631, "mem_mb": 63.172096, "nll": -5.9673919677734375, "elapsed_s": 90.7} {"event": "train", "step": 980, "loss": -5.924966812133789, "grad_norm": 0.4259200096130371, "lr": 0.0002983379259803436, "ex_per_s": 679.809454188512, "data_frac": 0.12412853937296404, "mem_mb": 63.172096, "nll": -5.924966812133789, "elapsed_s": 92.3} {"event": "train", "step": 1000, "loss": -6.07198429107666, "grad_norm": 0.14297768473625183, "lr": 0.00029822158054310604, "ex_per_s": 681.5347739442733, "data_frac": 0.12206511690574481, "mem_mb": 63.172096, "nll": -6.07198429107666, "elapsed_s": 93.9} {"event": "val", "step": 1000, "val_nll": -6.446158300834639, "best": -6.446158300834639, "bad": 0, "mem_mb": 63.172096, "elapsed_s": 156.0} {"event": "train", "step": 1020, "loss": -6.030567646026611, "grad_norm": 0.2499178797006607, "lr": 0.0002981013265414125, "ex_per_s": 414.03253204629584, "data_frac": 0.07276717315492853, "mem_mb": 63.172096, "nll": -6.030567646026611, "elapsed_s": 157.7} {"event": "train", "step": 1040, "loss": -5.972133159637451, "grad_norm": 0.21097296476364136, "lr": 0.000297977167503374, "ex_per_s": 417.7350734172494, "data_frac": 0.07206923609313103, "mem_mb": 63.172096, "nll": -5.972133159637451, "elapsed_s": 159.4} {"event": "train", "step": 1060, "loss": -6.181245803833008, "grad_norm": 0.22506771981716156, "lr": 0.00029784910707167064, "ex_per_s": 421.36885761551207, "data_frac": 0.07138877677187289, "mem_mb": 63.172096, "nll": -6.181245803833008, "elapsed_s": 161.0} {"event": "train", "step": 1080, "loss": -5.813438892364502, "grad_norm": 0.6017510294914246, "lr": 0.0002977171490034448, "ex_per_s": 424.92421766885616, "data_frac": 0.07072241675767744, "mem_mb": 63.172096, "nll": -5.813438892364502, "elapsed_s": 162.7} {"event": "train", "step": 1100, "loss": -6.052811622619629, "grad_norm": 0.1277649700641632, "lr": 0.000297581297170191, "ex_per_s": 428.4126194708756, "data_frac": 0.07006898084567492, "mem_mb": 63.172096, "nll": -6.052811622619629, "elapsed_s": 164.4} {"event": "train", "step": 1120, "loss": -6.052059173583984, "grad_norm": 0.35564756393432617, "lr": 0.0002974415555576423, "ex_per_s": 431.826991424869, "data_frac": 0.0694323008187324, "mem_mb": 63.172096, "nll": -6.052059173583984, "elapsed_s": 166.0} {"event": "train", "step": 1140, "loss": -6.019077777862549, "grad_norm": 0.3295891284942627, "lr": 0.000297297928265653, "ex_per_s": 435.17861336000664, "data_frac": 0.06880734320695316, "mem_mb": 63.172096, "nll": -6.019077777862549, "elapsed_s": 167.7} {"event": "train", "step": 1160, "loss": -5.970592975616455, "grad_norm": 0.41435620188713074, "lr": 0.00029715041950807914, "ex_per_s": 438.46595963606427, "data_frac": 0.0681919126036637, "mem_mb": 63.172096, "nll": -5.970592975616455, "elapsed_s": 169.4} {"event": "train", "step": 1180, "loss": -5.9583210945129395, "grad_norm": 0.25305044651031494, "lr": 0.000296999033612654, "ex_per_s": 441.69028001443957, "data_frac": 0.06758850877909278, "mem_mb": 63.172096, "nll": -5.9583210945129395, "elapsed_s": 171.0} {"event": "train", "step": 1200, "loss": -5.946788787841797, "grad_norm": 0.18739229440689087, "lr": 0.00029684377502086165, "ex_per_s": 444.85140639443046, "data_frac": 0.0669965350296369, "mem_mb": 63.172096, "nll": -5.946788787841797, "elapsed_s": 172.7} {"event": "train", "step": 1220, "loss": -6.031352519989014, "grad_norm": 0.7015054821968079, "lr": 0.00029668464828780665, "ex_per_s": 447.95172192951134, "data_frac": 0.06641692561463759, "mem_mb": 63.172096, "nll": -6.031352519989014, "elapsed_s": 174.3} {"event": "train", "step": 1240, "loss": -6.059912204742432, "grad_norm": 0.06966850161552429, "lr": 0.00029652165808208004, "ex_per_s": 450.99099718714285, "data_frac": 0.06584767997388279, "mem_mb": 63.172096, "nll": -6.059912204742432, "elapsed_s": 176.0} {"event": "train", "step": 1260, "loss": -6.0537333488464355, "grad_norm": 0.13834266364574432, "lr": 0.0002963548091856228, "ex_per_s": 453.9731237891592, "data_frac": 0.06528904027595596, "mem_mb": 63.172096, "nll": -6.0537333488464355, "elapsed_s": 177.7} {"event": "train", "step": 1280, "loss": -6.044801235198975, "grad_norm": 0.5171828866004944, "lr": 0.0002961841064935852, "ex_per_s": 456.9043184766815, "data_frac": 0.06474063549749737, "mem_mb": 63.172096, "nll": -6.044801235198975, "elapsed_s": 179.3} {"event": "train", "step": 1300, "loss": -5.868900775909424, "grad_norm": 0.5461729168891907, "lr": 0.0002960095550141835, "ex_per_s": 459.7820110040241, "data_frac": 0.06420180159990015, "mem_mb": 63.172096, "nll": -5.868900775909424, "elapsed_s": 181.0} {"event": "train", "step": 1320, "loss": -5.828946113586426, "grad_norm": 0.2485477179288864, "lr": 0.00029583115986855273, "ex_per_s": 462.6034936544026, "data_frac": 0.06367862490740565, "mem_mb": 63.172096, "nll": -5.828946113586426, "elapsed_s": 182.7} {"event": "train", "step": 1340, "loss": -5.951955318450928, "grad_norm": 0.2226409614086151, "lr": 0.0002956489262905966, "ex_per_s": 465.3849262522626, "data_frac": 0.06316119408036307, "mem_mb": 63.172096, "nll": -5.951955318450928, "elapsed_s": 184.3} {"event": "train", "step": 1360, "loss": -6.0361456871032715, "grad_norm": 0.7093385457992554, "lr": 0.0002954628596268339, "ex_per_s": 468.11066809518434, "data_frac": 0.06265577476391897, "mem_mb": 63.172096, "nll": -6.0361456871032715, "elapsed_s": 186.0} {"event": "train", "step": 1380, "loss": -6.067817687988281, "grad_norm": 0.18961592018604279, "lr": 0.0002952729653362417, "ex_per_s": 470.7954172453888, "data_frac": 0.06216118683836175, "mem_mb": 63.172096, "nll": -6.067817687988281, "elapsed_s": 187.6} {"event": "train", "step": 1400, "loss": -5.796936511993408, "grad_norm": 0.16985869407653809, "lr": 0.00029507924899009505, "ex_per_s": 473.43354884295456, "data_frac": 0.061671624027917704, "mem_mb": 63.172096, "nll": -5.796936511993408, "elapsed_s": 189.3} {"event": "train", "step": 1420, "loss": -5.98995304107666, "grad_norm": 0.5484545826911926, "lr": 0.00029488171627180373, "ex_per_s": 476.0299659053017, "data_frac": 0.06119036570816618, "mem_mb": 63.172096, "nll": -5.98995304107666, "elapsed_s": 191.0} {"event": "train", "step": 1440, "loss": -5.743724822998047, "grad_norm": 0.14672715961933136, "lr": 0.00029468037297674526, "ex_per_s": 478.5817630129601, "data_frac": 0.06071625056132416, "mem_mb": 63.172096, "nll": -5.743724822998047, "elapsed_s": 192.6} {"event": "train", "step": 1460, "loss": -5.802306652069092, "grad_norm": 0.5829465985298157, "lr": 0.00029447522501209515, "ex_per_s": 481.09454207333533, "data_frac": 0.06025186983056, "mem_mb": 63.172096, "nll": -5.802306652069092, "elapsed_s": 194.3} {"event": "train", "step": 1480, "loss": -6.0454816818237305, "grad_norm": 0.5547296404838562, "lr": 0.00029426627839665323, "ex_per_s": 483.5678864908159, "data_frac": 0.05979537552909147, "mem_mb": 63.172096, "nll": -6.0454816818237305, "elapsed_s": 195.9} {"event": "train", "step": 1500, "loss": -5.9572601318359375, "grad_norm": 0.5366147756576538, "lr": 0.00029405353926066754, "ex_per_s": 486.0010141305925, "data_frac": 0.05934592988352912, "mem_mb": 63.172096, "nll": -5.9572601318359375, "elapsed_s": 197.6} {"event": "train", "step": 1520, "loss": -5.95039701461792, "grad_norm": 0.6240704655647278, "lr": 0.000293837013845654, "ex_per_s": 488.3944372383065, "data_frac": 0.05890516865319881, "mem_mb": 63.172096, "nll": -5.95039701461792, "elapsed_s": 199.2} {"event": "train", "step": 1540, "loss": -5.879042625427246, "grad_norm": 0.1304602473974228, "lr": 0.00029361670850421354, "ex_per_s": 490.7452771195747, "data_frac": 0.058470889162323555, "mem_mb": 63.172096, "nll": -5.879042625427246, "elapsed_s": 200.9} {"event": "train", "step": 1560, "loss": -5.98274564743042, "grad_norm": 0.28221723437309265, "lr": 0.0002933926296998457, "ex_per_s": 493.06003943902795, "data_frac": 0.058044007308095266, "mem_mb": 63.172096, "nll": -5.98274564743042, "elapsed_s": 202.5} {"event": "train", "step": 1580, "loss": -5.98096227645874, "grad_norm": 0.08555147051811218, "lr": 0.0002931647840067591, "ex_per_s": 495.3406606517916, "data_frac": 0.05762462740662021, "mem_mb": 63.172096, "nll": -5.98096227645874, "elapsed_s": 204.2} {"event": "train", "step": 1600, "loss": -5.957459926605225, "grad_norm": 0.19032616913318634, "lr": 0.0002929331781096783, "ex_per_s": 497.583462128493, "data_frac": 0.05721238748454551, "mem_mb": 63.172096, "nll": -5.957459926605225, "elapsed_s": 205.8} {"event": "train", "step": 1620, "loss": -5.880980491638184, "grad_norm": 0.1290394365787506, "lr": 0.0002926978188036479, "ex_per_s": 499.7874116712709, "data_frac": 0.056810763346662456, "mem_mb": 63.172096, "nll": -5.880980491638184, "elapsed_s": 207.5} {"event": "train", "step": 1640, "loss": -6.076888561248779, "grad_norm": 0.11601332575082779, "lr": 0.00029245871299383297, "ex_per_s": 501.963245054315, "data_frac": 0.05641150797871458, "mem_mb": 63.172096, "nll": -6.076888561248779, "elapsed_s": 209.1} {"event": "train", "step": 1660, "loss": -6.040598392486572, "grad_norm": 1.0181434154510498, "lr": 0.0002922158676953168, "ex_per_s": 504.1041404418615, "data_frac": 0.05601926214822439, "mem_mb": 63.172096, "nll": -6.040598392486572, "elapsed_s": 210.8} {"event": "train", "step": 1680, "loss": -5.866607666015625, "grad_norm": 0.1540675312280655, "lr": 0.0002919692900328947, "ex_per_s": 506.2090409449146, "data_frac": 0.05563550633092482, "mem_mb": 63.172096, "nll": -5.866607666015625, "elapsed_s": 212.4} {"event": "train", "step": 1700, "loss": -5.927515506744385, "grad_norm": 0.40995556116104126, "lr": 0.0002917189872408653, "ex_per_s": 508.2847804096181, "data_frac": 0.05525558688914628, "mem_mb": 63.172096, "nll": -5.927515506744385, "elapsed_s": 214.1} {"event": "train", "step": 1720, "loss": -5.886694431304932, "grad_norm": 0.9888463020324707, "lr": 0.00029146496666281795, "ex_per_s": 510.32913517366046, "data_frac": 0.054880535257886436, "mem_mb": 63.172096, "nll": -5.886694431304932, "elapsed_s": 215.7} {"event": "train", "step": 1740, "loss": -6.2738423347473145, "grad_norm": 0.3071620762348175, "lr": 0.00029120723575141777, "ex_per_s": 512.3462070248095, "data_frac": 0.05451149899159341, "mem_mb": 63.172096, "nll": -6.2738423347473145, "elapsed_s": 217.4} {"event": "train", "step": 1760, "loss": -5.930585861206055, "grad_norm": 0.22015134990215302, "lr": 0.00029094580206818636, "ex_per_s": 514.3348557064244, "data_frac": 0.05414753270912627, "mem_mb": 63.172096, "nll": -5.930585861206055, "elapsed_s": 219.0} {"event": "train", "step": 1780, "loss": -5.872331619262695, "grad_norm": 0.45501071214675903, "lr": 0.00029068067328328045, "ex_per_s": 516.2958271585276, "data_frac": 0.053789501243533405, "mem_mb": 63.172096, "nll": -5.872331619262695, "elapsed_s": 220.7} {"event": "train", "step": 1800, "loss": -6.050513744354248, "grad_norm": 0.8535847067832947, "lr": 0.0002904118571752668, "ex_per_s": 518.2269511843339, "data_frac": 0.053438342027675195, "mem_mb": 63.172096, "nll": -6.050513744354248, "elapsed_s": 222.3} {"event": "train", "step": 1820, "loss": -5.937288761138916, "grad_norm": 0.5286372900009155, "lr": 0.00029013936163089376, "ex_per_s": 520.1262260552221, "data_frac": 0.053090922643030405, "mem_mb": 63.172096, "nll": -5.937288761138916, "elapsed_s": 224.0} {"event": "train", "step": 1840, "loss": -6.071540832519531, "grad_norm": 0.48307955265045166, "lr": 0.00028986319464486, "ex_per_s": 522.0019431122712, "data_frac": 0.05274813020899129, "mem_mb": 63.172096, "nll": -6.071540832519531, "elapsed_s": 225.6} {"event": "train", "step": 1860, "loss": -6.050912857055664, "grad_norm": 0.2569665014743805, "lr": 0.00028958336431958003, "ex_per_s": 523.8453745316823, "data_frac": 0.05241602146049589, "mem_mb": 63.172096, "nll": -6.050912857055664, "elapsed_s": 227.3} {"event": "train", "step": 1880, "loss": -5.945103645324707, "grad_norm": 0.7374191284179688, "lr": 0.0002892998788649464, "ex_per_s": 525.668862132514, "data_frac": 0.05208409087142406, "mem_mb": 63.172096, "nll": -5.945103645324707, "elapsed_s": 228.9} {"event": "train", "step": 1900, "loss": -5.978987693786621, "grad_norm": 0.3236611783504486, "lr": 0.0002890127465980889, "ex_per_s": 527.466414581861, "data_frac": 0.05175644606524177, "mem_mb": 63.172096, "nll": -5.978987693786621, "elapsed_s": 230.6} {"event": "train", "step": 1920, "loss": -5.965012550354004, "grad_norm": 0.2530168294906616, "lr": 0.0002887219759431305, "ex_per_s": 529.2393571000656, "data_frac": 0.05143276674227902, "mem_mb": 63.172096, "nll": -5.965012550354004, "elapsed_s": 232.2} {"event": "train", "step": 1940, "loss": -6.049437522888184, "grad_norm": 0.41504213213920593, "lr": 0.0002884275754309401, "ex_per_s": 530.9861165841885, "data_frac": 0.051114566896215144, "mem_mb": 63.172096, "nll": -6.049437522888184, "elapsed_s": 233.9} {"event": "train", "step": 1960, "loss": -5.938776969909668, "grad_norm": 0.1256941705942154, "lr": 0.0002881295536988824, "ex_per_s": 532.6959536663604, "data_frac": 0.05080035964863198, "mem_mb": 63.172096, "nll": -5.938776969909668, "elapsed_s": 235.5} {"event": "train", "step": 1980, "loss": -5.965417385101318, "grad_norm": 0.5672934651374817, "lr": 0.00028782791949056447, "ex_per_s": 534.3817472472746, "data_frac": 0.05049045449001353, "mem_mb": 63.172096, "nll": -5.965417385101318, "elapsed_s": 237.2} {"event": "train", "step": 2000, "loss": -5.916933536529541, "grad_norm": 0.11739686876535416, "lr": 0.00028752268165557917, "ex_per_s": 536.034460242721, "data_frac": 0.05018765996713025, "mem_mb": 63.172096, "nll": -5.916933536529541, "elapsed_s": 238.8} {"event": "val", "step": 2000, "val_nll": -6.446376713745543, "best": -6.446376713745543, "bad": 0, "mem_mb": 63.172096, "elapsed_s": 301.1} {"event": "train", "step": 2020, "loss": -5.927313804626465, "grad_norm": 0.3491455912590027, "lr": 0.00028721384914924547, "ex_per_s": 426.9804564906654, "data_frac": 0.03961822323403509, "mem_mb": 63.172096, "nll": -5.927313804626465, "elapsed_s": 302.8} {"event": "train", "step": 2040, "loss": -5.875000953674316, "grad_norm": 0.23502077162265778, "lr": 0.0002869014310323458, "ex_per_s": 428.8477091225559, "data_frac": 0.0394355666244126, "mem_mb": 63.172096, "nll": -5.875000953674316, "elapsed_s": 304.5} {"event": "train", "step": 2060, "loss": -6.1439008712768555, "grad_norm": 0.41506296396255493, "lr": 0.0002865854364708602, "ex_per_s": 430.6956927427364, "data_frac": 0.03925433261250729, "mem_mb": 63.172096, "nll": -6.1439008712768555, "elapsed_s": 306.2} {"event": "train", "step": 2080, "loss": -5.959244251251221, "grad_norm": 0.41924750804901123, "lr": 0.00028626587473569746, "ex_per_s": 432.517232993533, "data_frac": 0.039077424320814104, "mem_mb": 63.172096, "nll": -5.959244251251221, "elapsed_s": 307.8} {"event": "train", "step": 2100, "loss": -5.900227069854736, "grad_norm": 0.6743654608726501, "lr": 0.0002859427552024228, "ex_per_s": 434.3244127243767, "data_frac": 0.03890049378013735, "mem_mb": 63.172096, "nll": -5.900227069854736, "elapsed_s": 309.5} {"event": "train", "step": 2120, "loss": -6.150910377502441, "grad_norm": 0.7002326250076294, "lr": 0.0002856160873509834, "ex_per_s": 436.1133732371835, "data_frac": 0.03872543968800953, "mem_mb": 63.172096, "nll": -6.150910377502441, "elapsed_s": 311.2} {"event": "train", "step": 2140, "loss": -5.977433204650879, "grad_norm": 0.5886155962944031, "lr": 0.00028528588076542964, "ex_per_s": 437.8834652849882, "data_frac": 0.03855101700822096, "mem_mb": 63.172096, "nll": -5.977433204650879, "elapsed_s": 312.8} {"event": "train", "step": 2160, "loss": -5.9281005859375, "grad_norm": 0.32228100299835205, "lr": 0.00028495214513363427, "ex_per_s": 439.6351939320159, "data_frac": 0.03837883306576439, "mem_mb": 63.172096, "nll": -5.9281005859375, "elapsed_s": 314.5} {"event": "train", "step": 2180, "loss": -6.082531452178955, "grad_norm": 0.3972252607345581, "lr": 0.0002846148902470082, "ex_per_s": 441.36700211909135, "data_frac": 0.03820876109962814, "mem_mb": 63.172096, "nll": -6.082531452178955, "elapsed_s": 316.2} {"event": "train", "step": 2200, "loss": -5.961066722869873, "grad_norm": 0.7833836078643799, "lr": 0.0002842741260002131, "ex_per_s": 443.08010024799194, "data_frac": 0.0380408960759021, "mem_mb": 63.172096, "nll": -5.961066722869873, "elapsed_s": 317.8} {"event": "train", "step": 2220, "loss": -5.827590465545654, "grad_norm": 0.5943112969398499, "lr": 0.00028392986239087094, "ex_per_s": 444.77557189191407, "data_frac": 0.037874811439265695, "mem_mb": 63.172096, "nll": -5.827590465545654, "elapsed_s": 319.5} {"event": "train", "step": 2240, "loss": -5.963875770568848, "grad_norm": 0.3622554838657379, "lr": 0.0002835821095192712, "ex_per_s": 446.45300715839574, "data_frac": 0.03771043377101574, "mem_mb": 63.172096, "nll": -5.963875770568848, "elapsed_s": 321.2} {"event": "train", "step": 2260, "loss": -5.738922595977783, "grad_norm": 0.7406855225563049, "lr": 0.0002832308775880741, "ex_per_s": 448.1130800731457, "data_frac": 0.03754856611080323, "mem_mb": 63.172096, "nll": -5.738922595977783, "elapsed_s": 322.8} {"event": "train", "step": 2280, "loss": -6.011047840118408, "grad_norm": 0.31354546546936035, "lr": 0.0002828761769020111, "ex_per_s": 449.7565902657175, "data_frac": 0.037387274054238674, "mem_mb": 63.172096, "nll": -6.011047840118408, "elapsed_s": 324.5} {"event": "train", "step": 2300, "loss": -6.001675605773926, "grad_norm": 0.14651307463645935, "lr": 0.0002825180178675832, "ex_per_s": 451.38317696677757, "data_frac": 0.037227816014786516, "mem_mb": 63.172096, "nll": -6.001675605773926, "elapsed_s": 326.2} {"event": "train", "step": 2320, "loss": -6.069732189178467, "grad_norm": 0.3306010365486145, "lr": 0.00028215641099275517, "ex_per_s": 452.99035949526836, "data_frac": 0.037073774877562214, "mem_mb": 63.172096, "nll": -6.069732189178467, "elapsed_s": 327.8} {"event": "train", "step": 2340, "loss": -5.851790904998779, "grad_norm": 0.4883081614971161, "lr": 0.0002817913668866473, "ex_per_s": 454.58708022269974, "data_frac": 0.036918028152470604, "mem_mb": 63.172096, "nll": -5.851790904998779, "elapsed_s": 329.5} {"event": "train", "step": 2360, "loss": -5.927648067474365, "grad_norm": 0.5413211584091187, "lr": 0.00028142289625922426, "ex_per_s": 456.16705610152064, "data_frac": 0.036763798721530086, "mem_mb": 63.172096, "nll": -5.927648067474365, "elapsed_s": 331.2} {"event": "train", "step": 2380, "loss": -5.931927680969238, "grad_norm": 0.08088929951190948, "lr": 0.00028105100992098075, "ex_per_s": 457.72859184804565, "data_frac": 0.03661285456820215, "mem_mb": 63.172096, "nll": -5.931927680969238, "elapsed_s": 332.8} {"event": "train", "step": 2400, "loss": -5.957698345184326, "grad_norm": 0.6734814643859863, "lr": 0.00028067571878262454, "ex_per_s": 459.2813777122731, "data_frac": 0.036461935776667134, "mem_mb": 63.172096, "nll": -5.957698345184326, "elapsed_s": 334.5} {"event": "train", "step": 2420, "loss": -5.99725866317749, "grad_norm": 0.15777572989463806, "lr": 0.00028029703385475607, "ex_per_s": 460.8158547794208, "data_frac": 0.03631209235253791, "mem_mb": 63.172096, "nll": -5.99725866317749, "elapsed_s": 336.1} {"event": "train", "step": 2440, "loss": -6.09627628326416, "grad_norm": 0.6938881278038025, "lr": 0.0002799149662475457, "ex_per_s": 462.33794932101495, "data_frac": 0.036163448183922635, "mem_mb": 63.172096, "nll": -6.09627628326416, "elapsed_s": 337.8} {"event": "train", "step": 2460, "loss": -6.132328987121582, "grad_norm": 0.43776780366897583, "lr": 0.0002795295271704076, "ex_per_s": 463.84558909193976, "data_frac": 0.036017280999575386, "mem_mb": 63.172096, "nll": -6.132328987121582, "elapsed_s": 339.5} {"event": "train", "step": 2480, "loss": -5.9446635246276855, "grad_norm": 0.1450570821762085, "lr": 0.00027914072793167077, "ex_per_s": 465.3442468514284, "data_frac": 0.035871576527755265, "mem_mb": 63.172096, "nll": -5.9446635246276855, "elapsed_s": 341.1} {"event": "train", "step": 2500, "loss": -5.754404067993164, "grad_norm": 0.6287614107131958, "lr": 0.0002787485799382476, "ex_per_s": 466.82658645188286, "data_frac": 0.03572970940328002, "mem_mb": 63.172096, "nll": -5.754404067993164, "elapsed_s": 342.8} {"event": "train", "step": 2520, "loss": -5.962614059448242, "grad_norm": 0.22276721894741058, "lr": 0.0002783530946952989, "ex_per_s": 468.298864458377, "data_frac": 0.03558841633996579, "mem_mb": 63.172096, "nll": -5.962614059448242, "elapsed_s": 344.4} {"event": "train", "step": 2540, "loss": -6.032337188720703, "grad_norm": 0.16433095932006836, "lr": 0.00027795428380589644, "ex_per_s": 469.75766718610157, "data_frac": 0.03544809245361104, "mem_mb": 63.172096, "nll": -6.032337188720703, "elapsed_s": 346.1} {"event": "train", "step": 2560, "loss": -5.9483418464660645, "grad_norm": 0.34969472885131836, "lr": 0.00027755215897068257, "ex_per_s": 471.19585820568904, "data_frac": 0.03531251753691441, "mem_mb": 63.172096, "nll": -5.9483418464660645, "elapsed_s": 347.8} {"event": "train", "step": 2580, "loss": -5.963629722595215, "grad_norm": 0.37064749002456665, "lr": 0.0002771467319875269, "ex_per_s": 472.62591898566086, "data_frac": 0.03517472697203905, "mem_mb": 63.172096, "nll": -5.963629722595215, "elapsed_s": 349.4} {"event": "train", "step": 2600, "loss": -5.849702835083008, "grad_norm": 0.17619764804840088, "lr": 0.0002767380147511802, "ex_per_s": 474.04404883002877, "data_frac": 0.0350378088322391, "mem_mb": 63.172096, "nll": -5.849702835083008, "elapsed_s": 351.1} {"event": "train", "step": 2620, "loss": -5.890588283538818, "grad_norm": 0.945696234703064, "lr": 0.0002763260192529254, "ex_per_s": 475.45011146860617, "data_frac": 0.03490256575335866, "mem_mb": 63.172096, "nll": -5.890588283538818, "elapsed_s": 352.7} {"event": "train", "step": 2640, "loss": -5.986608982086182, "grad_norm": 0.4363175630569458, "lr": 0.0002759107575802257, "ex_per_s": 476.84215062660195, "data_frac": 0.03476938401349593, "mem_mb": 63.172096, "nll": -5.986608982086182, "elapsed_s": 354.4} {"event": "train", "step": 2660, "loss": -5.820504188537598, "grad_norm": 0.2584606111049652, "lr": 0.00027549224191637014, "ex_per_s": 478.2170317767965, "data_frac": 0.03463691201355302, "mem_mb": 63.172096, "nll": -5.820504188537598, "elapsed_s": 356.0} {"event": "train", "step": 2680, "loss": -5.9575629234313965, "grad_norm": 0.5263194441795349, "lr": 0.000275070484540116, "ex_per_s": 479.57419901114525, "data_frac": 0.03450596275499934, "mem_mb": 63.172096, "nll": -5.9575629234313965, "elapsed_s": 357.7} {"event": "train", "step": 2700, "loss": -6.134957313537598, "grad_norm": 0.1764765828847885, "lr": 0.00027464549782532844, "ex_per_s": 480.9180552152505, "data_frac": 0.034375857188581625, "mem_mb": 63.172096, "nll": -6.134957313537598, "elapsed_s": 359.4} {"event": "train", "step": 2720, "loss": -5.882750511169434, "grad_norm": 0.17603550851345062, "lr": 0.00027421729424061787, "ex_per_s": 482.2538535677106, "data_frac": 0.034246216854173256, "mem_mb": 63.172096, "nll": -5.882750511169434, "elapsed_s": 361.0} {"event": "train", "step": 2740, "loss": -6.128791809082031, "grad_norm": 0.08595699071884155, "lr": 0.0002737858863489736, "ex_per_s": 483.57448375326345, "data_frac": 0.03411832994266128, "mem_mb": 63.172096, "nll": -6.128791809082031, "elapsed_s": 362.7} {"event": "train", "step": 2760, "loss": -5.961512088775635, "grad_norm": 0.4753054976463318, "lr": 0.00027335128680739583, "ex_per_s": 484.8820361204286, "data_frac": 0.03399338114512934, "mem_mb": 63.172096, "nll": -5.961512088775635, "elapsed_s": 364.3} {"event": "train", "step": 2780, "loss": -6.015606880187988, "grad_norm": 0.3570384979248047, "lr": 0.00027291350836652374, "ex_per_s": 486.18155583285323, "data_frac": 0.033867500640384504, "mem_mb": 63.172096, "nll": -6.015606880187988, "elapsed_s": 366.0} {"event": "train", "step": 2800, "loss": -5.981351852416992, "grad_norm": 0.7248348593711853, "lr": 0.00027247256387026185, "ex_per_s": 487.46535217763443, "data_frac": 0.03374503971164095, "mem_mb": 63.172096, "nll": -5.981351852416992, "elapsed_s": 367.7} {"event": "train", "step": 2820, "loss": -6.007913589477539, "grad_norm": 0.5331605076789856, "lr": 0.000272028466255403, "ex_per_s": 488.7405262650144, "data_frac": 0.0336217097560819, "mem_mb": 63.172096, "nll": -6.007913589477539, "elapsed_s": 369.3} {"event": "train", "step": 2840, "loss": -5.880427837371826, "grad_norm": 0.21495909988880157, "lr": 0.0002715812285512487, "ex_per_s": 490.0081892835462, "data_frac": 0.033498966709166196, "mem_mb": 63.172096, "nll": -5.880427837371826, "elapsed_s": 371.0} {"event": "train", "step": 2860, "loss": -6.05047607421875, "grad_norm": 0.15266244113445282, "lr": 0.0002711308638792272, "ex_per_s": 491.2632858965441, "data_frac": 0.03337801140924267, "mem_mb": 63.172096, "nll": -6.05047607421875, "elapsed_s": 372.6} {"event": "train", "step": 2880, "loss": -5.7739057540893555, "grad_norm": 0.23507851362228394, "lr": 0.000270677385452508, "ex_per_s": 492.5079235546913, "data_frac": 0.033257810502207316, "mem_mb": 63.172096, "nll": -5.7739057540893555, "elapsed_s": 374.3} {"event": "train", "step": 2900, "loss": -5.79514741897583, "grad_norm": 0.39439326524734497, "lr": 0.00027022080657561484, "ex_per_s": 493.74062061682196, "data_frac": 0.03313828569185861, "mem_mb": 63.172096, "nll": -5.79514741897583, "elapsed_s": 376.0} {"event": "train", "step": 2920, "loss": -5.972633361816406, "grad_norm": 0.2030097097158432, "lr": 0.0002697611406440348, "ex_per_s": 494.9631277722764, "data_frac": 0.03302054589109779, "mem_mb": 63.172096, "nll": -5.972633361816406, "elapsed_s": 377.6} {"event": "train", "step": 2940, "loss": -5.9016289710998535, "grad_norm": 0.7790021300315857, "lr": 0.00026929840114382563, "ex_per_s": 496.1761174001425, "data_frac": 0.03290445365315267, "mem_mb": 63.172096, "nll": -5.9016289710998535, "elapsed_s": 379.3} {"event": "train", "step": 2960, "loss": -5.93826150894165, "grad_norm": 0.2902924418449402, "lr": 0.0002688326016512199, "ex_per_s": 497.37627191613063, "data_frac": 0.03278944808994727, "mem_mb": 63.172096, "nll": -5.93826150894165, "elapsed_s": 380.9} {"event": "train", "step": 2980, "loss": -5.894770622253418, "grad_norm": 0.13740569353103638, "lr": 0.0002683637558322267, "ex_per_s": 498.56646351487433, "data_frac": 0.03267609390793888, "mem_mb": 63.172096, "nll": -5.894770622253418, "elapsed_s": 382.6} {"event": "train", "step": 3000, "loss": -5.99525260925293, "grad_norm": 0.12399424612522125, "lr": 0.0002678918774422309, "ex_per_s": 499.7500338732648, "data_frac": 0.03256289805554254, "mem_mb": 63.172096, "nll": -5.99525260925293, "elapsed_s": 384.2} {"event": "val", "step": 3000, "val_nll": -6.448426962465974, "best": -6.448426962465974, "bad": 0, "mem_mb": 63.172096, "elapsed_s": 447.2} {"event": "train", "step": 3020, "loss": -5.960635662078857, "grad_norm": 0.8723112344741821, "lr": 0.0002674169803255893, "ex_per_s": 430.5877987036154, "data_frac": 0.027896652003484165, "mem_mb": 63.172096, "nll": -5.960635662078857, "elapsed_s": 448.9} {"event": "train", "step": 3040, "loss": -5.853154182434082, "grad_norm": 1.4352638721466064, "lr": 0.0002669390784152248, "ex_per_s": 431.8353464190291, "data_frac": 0.02781614810262246, "mem_mb": 63.172096, "nll": -5.853154182434082, "elapsed_s": 450.6} {"event": "train", "step": 3060, "loss": -6.072968006134033, "grad_norm": 0.29385408759117126, "lr": 0.00026645818573221706, "ex_per_s": 433.0732877078356, "data_frac": 0.027736170135678206, "mem_mb": 63.172096, "nll": -6.072968006134033, "elapsed_s": 452.3} {"event": "train", "step": 3080, "loss": -5.969807147979736, "grad_norm": 0.12852391600608826, "lr": 0.00026597431638539183, "ex_per_s": 434.29968866187085, "data_frac": 0.02765840455952802, "mem_mb": 63.172096, "nll": -5.969807147979736, "elapsed_s": 453.9} {"event": "train", "step": 3100, "loss": -6.013445854187012, "grad_norm": 0.40949347615242004, "lr": 0.00026548748457090647, "ex_per_s": 435.52055317558654, "data_frac": 0.027579844980730694, "mem_mb": 63.172096, "nll": -6.013445854187012, "elapsed_s": 455.6} {"event": "train", "step": 3120, "loss": -6.087645053863525, "grad_norm": 1.1411885023117065, "lr": 0.00026499770457183376, "ex_per_s": 436.73224652522094, "data_frac": 0.027502220226973552, "mem_mb": 63.172096, "nll": -6.087645053863525, "elapsed_s": 457.3} {"event": "train", "step": 3140, "loss": -6.00924015045166, "grad_norm": 0.1493375450372696, "lr": 0.0002645049907577427, "ex_per_s": 437.93482932810116, "data_frac": 0.027425257109822423, "mem_mb": 63.172096, "nll": -6.00924015045166, "elapsed_s": 458.9} {"event": "train", "step": 3160, "loss": -6.0221266746521, "grad_norm": 0.18444427847862244, "lr": 0.00026400935758427697, "ex_per_s": 439.1293284715593, "data_frac": 0.027349290763092206, "mem_mb": 63.172096, "nll": -6.0221266746521, "elapsed_s": 460.6} {"event": "train", "step": 3180, "loss": -6.092473030090332, "grad_norm": 0.3708280026912689, "lr": 0.00026351081959273087, "ex_per_s": 440.31587530716234, "data_frac": 0.027273279302234044, "mem_mb": 63.172096, "nll": -6.092473030090332, "elapsed_s": 462.3} {"event": "train", "step": 3200, "loss": -6.002447605133057, "grad_norm": 0.5599419474601746, "lr": 0.0002630093914096226, "ex_per_s": 441.49244102572436, "data_frac": 0.027199240144503457, "mem_mb": 63.172096, "nll": -6.002447605133057, "elapsed_s": 463.9} {"event": "train", "step": 3220, "loss": -5.914905548095703, "grad_norm": 0.23569291830062866, "lr": 0.00026250508774626523, "ex_per_s": 442.6608414293061, "data_frac": 0.027124738080957583, "mem_mb": 63.172096, "nll": -5.914905548095703, "elapsed_s": 465.6} {"event": "train", "step": 3240, "loss": -5.80134391784668, "grad_norm": 0.3219914138317108, "lr": 0.00026199792339833493, "ex_per_s": 443.8209234425412, "data_frac": 0.027049647824194115, "mem_mb": 63.172096, "nll": -5.80134391784668, "elapsed_s": 467.3} {"event": "train", "step": 3260, "loss": -5.96143102645874, "grad_norm": 0.060165196657180786, "lr": 0.00026148791324543704, "ex_per_s": 444.9722718456315, "data_frac": 0.02697855141669005, "mem_mb": 63.172096, "nll": -5.96143102645874, "elapsed_s": 468.9} {"event": "train", "step": 3280, "loss": -6.0138068199157715, "grad_norm": 0.1550036370754242, "lr": 0.0002609750722506695, "ex_per_s": 446.1202422559969, "data_frac": 0.026905501721110267, "mem_mb": 63.172096, "nll": -6.0138068199157715, "elapsed_s": 470.6} {"event": "train", "step": 3300, "loss": -5.803009510040283, "grad_norm": 1.007999062538147, "lr": 0.0002604594154601839, "ex_per_s": 447.2603084563282, "data_frac": 0.02683297217394037, "mem_mb": 63.172096, "nll": -5.803009510040283, "elapsed_s": 472.3} {"event": "train", "step": 3320, "loss": -5.990198135375977, "grad_norm": 0.4588693380355835, "lr": 0.0002599409580027438, "ex_per_s": 448.39271540733097, "data_frac": 0.026760528560112656, "mem_mb": 63.172096, "nll": -5.990198135375977, "elapsed_s": 473.9} {"event": "train", "step": 3340, "loss": -5.93106746673584, "grad_norm": 0.0761118084192276, "lr": 0.0002594197150892811, "ex_per_s": 449.51468148207715, "data_frac": 0.026689630198840245, "mem_mb": 63.172096, "nll": -5.93106746673584, "elapsed_s": 475.6} {"event": "train", "step": 3360, "loss": -5.906250953674316, "grad_norm": 0.21101327240467072, "lr": 0.00025889570201244967, "ex_per_s": 450.62782492201035, "data_frac": 0.026618446316057878, "mem_mb": 63.172096, "nll": -5.906250953674316, "elapsed_s": 477.2} {"event": "train", "step": 3380, "loss": -6.124051570892334, "grad_norm": 0.17848913371562958, "lr": 0.00025836893414617684, "ex_per_s": 451.73554630390083, "data_frac": 0.026547408233217544, "mem_mb": 63.172096, "nll": -6.124051570892334, "elapsed_s": 478.9} {"event": "train", "step": 3400, "loss": -5.882930278778076, "grad_norm": 0.1357809454202652, "lr": 0.00025783942694521214, "ex_per_s": 452.83738219654197, "data_frac": 0.026477206853595475, "mem_mb": 63.172096, "nll": -5.882930278778076, "elapsed_s": 480.6} {"event": "train", "step": 3420, "loss": -5.836485862731934, "grad_norm": 0.08457709103822708, "lr": 0.00025730719594467387, "ex_per_s": 453.9326046793671, "data_frac": 0.026407846383766313, "mem_mb": 63.172096, "nll": -5.836485862731934, "elapsed_s": 482.2} {"event": "train", "step": 3440, "loss": -5.916750907897949, "grad_norm": 0.1257239431142807, "lr": 0.00025677225675959357, "ex_per_s": 455.0164140700459, "data_frac": 0.026340400458785985, "mem_mb": 63.172096, "nll": -5.916750907897949, "elapsed_s": 483.9} {"event": "train", "step": 3460, "loss": -6.056111812591553, "grad_norm": 0.16061939299106598, "lr": 0.00025623462508445754, "ex_per_s": 456.09778403755973, "data_frac": 0.02627306471403625, "mem_mb": 63.172096, "nll": -6.056111812591553, "elapsed_s": 485.6} {"event": "train", "step": 3480, "loss": -5.91658353805542, "grad_norm": 0.0613369457423687, "lr": 0.00025569431669274684, "ex_per_s": 457.1709751060341, "data_frac": 0.026205384447405805, "mem_mb": 63.172096, "nll": -5.91658353805542, "elapsed_s": 487.2} {"event": "train", "step": 3500, "loss": -6.033098220825195, "grad_norm": 0.9037821888923645, "lr": 0.00025515134743647386, "ex_per_s": 458.2359453287765, "data_frac": 0.02613944810327155, "mem_mb": 63.172096, "nll": -6.033098220825195, "elapsed_s": 488.9} {"event": "train", "step": 3520, "loss": -5.936584949493408, "grad_norm": 0.632296085357666, "lr": 0.0002546057332457181, "ex_per_s": 459.29611431862634, "data_frac": 0.026072503370399513, "mem_mb": 63.172096, "nll": -5.936584949493408, "elapsed_s": 490.5} {"event": "train", "step": 3540, "loss": -6.138355255126953, "grad_norm": 0.14636145532131195, "lr": 0.00025405749012815786, "ex_per_s": 460.345192508744, "data_frac": 0.026008208682993972, "mem_mb": 63.172096, "nll": -6.138355255126953, "elapsed_s": 492.2} {"event": "train", "step": 3560, "loss": -6.138952255249023, "grad_norm": 0.07016687095165253, "lr": 0.0002535066341686014, "ex_per_s": 461.39342371790025, "data_frac": 0.025944317599626204, "mem_mb": 63.172096, "nll": -6.138952255249023, "elapsed_s": 493.9} {"event": "train", "step": 3580, "loss": -5.901885986328125, "grad_norm": 0.531997561454773, "lr": 0.0002529531815285146, "ex_per_s": 462.433211098931, "data_frac": 0.02588030647451052, "mem_mb": 63.172096, "nll": -5.901885986328125, "elapsed_s": 495.5} {"event": "train", "step": 3600, "loss": -6.097924709320068, "grad_norm": 0.15150265395641327, "lr": 0.00025239714844554667, "ex_per_s": 463.46915124760625, "data_frac": 0.025815899021020765, "mem_mb": 63.172096, "nll": -6.097924709320068, "elapsed_s": 497.2} {"event": "train", "step": 3620, "loss": -5.9058942794799805, "grad_norm": 1.4655643701553345, "lr": 0.0002518385512330544, "ex_per_s": 464.49791481650817, "data_frac": 0.02575105824364544, "mem_mb": 63.172096, "nll": -5.9058942794799805, "elapsed_s": 498.8} {"event": "train", "step": 3640, "loss": -6.051377296447754, "grad_norm": 0.38966429233551025, "lr": 0.00025127740627962274, "ex_per_s": 465.5197741316041, "data_frac": 0.025687466797317076, "mem_mb": 63.172096, "nll": -6.051377296447754, "elapsed_s": 500.5} {"event": "train", "step": 3660, "loss": -5.830473899841309, "grad_norm": 0.4287618100643158, "lr": 0.00025071373004858453, "ex_per_s": 466.5332695195651, "data_frac": 0.025625831938876832, "mem_mb": 63.172096, "nll": -5.830473899841309, "elapsed_s": 502.1} {"event": "train", "step": 3680, "loss": -5.990777492523193, "grad_norm": 0.13760708272457123, "lr": 0.00025014753907753737, "ex_per_s": 467.53953064763084, "data_frac": 0.02556424382577638, "mem_mb": 63.172096, "nll": -5.990777492523193, "elapsed_s": 503.8} {"event": "train", "step": 3700, "loss": -6.07009744644165, "grad_norm": 0.14207005500793457, "lr": 0.0002495788499778582, "ex_per_s": 468.5424990222505, "data_frac": 0.025502344632763685, "mem_mb": 63.172096, "nll": -6.07009744644165, "elapsed_s": 505.4} {"event": "train", "step": 3720, "loss": -5.944260120391846, "grad_norm": 0.9022865891456604, "lr": 0.0002490076794342163, "ex_per_s": 469.53923576043843, "data_frac": 0.025440136666147292, "mem_mb": 63.172096, "nll": -5.944260120391846, "elapsed_s": 507.1} {"event": "train", "step": 3740, "loss": -6.191204071044922, "grad_norm": 0.2780141532421112, "lr": 0.00024843404420408344, "ex_per_s": 470.5260032658707, "data_frac": 0.025379605154576292, "mem_mb": 63.172096, "nll": -6.191204071044922, "elapsed_s": 508.7} {"event": "train", "step": 3760, "loss": -6.145015716552734, "grad_norm": 0.4090437889099121, "lr": 0.0002478579611172425, "ex_per_s": 471.5150560492497, "data_frac": 0.025319020186920282, "mem_mb": 63.172096, "nll": -6.145015716552734, "elapsed_s": 510.4} {"event": "train", "step": 3780, "loss": -5.9640116691589355, "grad_norm": 0.35761308670043945, "lr": 0.00024727944707529366, "ex_per_s": 472.4967801767024, "data_frac": 0.02525836619039493, "mem_mb": 63.172096, "nll": -5.9640116691589355, "elapsed_s": 512.0} {"event": "train", "step": 3800, "loss": -6.098962306976318, "grad_norm": 0.10082433372735977, "lr": 0.0002466985190511583, "ex_per_s": 473.4688460933677, "data_frac": 0.025198772803594238, "mem_mb": 63.172096, "nll": -6.098962306976318, "elapsed_s": 513.7} {"event": "train", "step": 3820, "loss": -5.969749450683594, "grad_norm": 0.5901807546615601, "lr": 0.0002461151940885812, "ex_per_s": 474.4374222545825, "data_frac": 0.025138909316231306, "mem_mb": 63.172096, "nll": -5.969749450683594, "elapsed_s": 515.3} {"event": "train", "step": 3840, "loss": -5.876527309417725, "grad_norm": 0.4000459313392639, "lr": 0.0002455294893016306, "ex_per_s": 475.40116412850966, "data_frac": 0.025079401500095983, "mem_mb": 63.172096, "nll": -5.876527309417725, "elapsed_s": 517.0} {"event": "train", "step": 3860, "loss": -6.013072490692139, "grad_norm": 0.34891465306282043, "lr": 0.000244941421874196, "ex_per_s": 476.3581459888232, "data_frac": 0.025020085377950498, "mem_mb": 63.172096, "nll": -6.013072490692139, "elapsed_s": 518.6} {"event": "train", "step": 3880, "loss": -5.889781475067139, "grad_norm": 0.3309165835380554, "lr": 0.00024435100905948387, "ex_per_s": 477.3101191387411, "data_frac": 0.024961543186666966, "mem_mb": 63.172096, "nll": -5.889781475067139, "elapsed_s": 520.3} {"event": "train", "step": 3900, "loss": -6.184322834014893, "grad_norm": 0.23490779101848602, "lr": 0.0002437582681795117, "ex_per_s": 478.25635994475783, "data_frac": 0.024903072548304393, "mem_mb": 63.172096, "nll": -6.184322834014893, "elapsed_s": 521.9} {"event": "train", "step": 3920, "loss": -6.035952568054199, "grad_norm": 0.9003032445907593, "lr": 0.00024316321662459957, "ex_per_s": 479.1949706190986, "data_frac": 0.024846045509361394, "mem_mb": 63.172096, "nll": -6.035952568054199, "elapsed_s": 523.6} {"event": "train", "step": 3940, "loss": -5.921543121337891, "grad_norm": 0.2802773416042328, "lr": 0.00024256587185286015, "ex_per_s": 480.1261110719133, "data_frac": 0.024788631012311976, "mem_mb": 63.172096, "nll": -5.921543121337891, "elapsed_s": 525.2} {"event": "train", "step": 3960, "loss": -6.066773891448975, "grad_norm": 0.6582695245742798, "lr": 0.0002419662513896863, "ex_per_s": 481.05214949747653, "data_frac": 0.024731920308130626, "mem_mb": 63.172096, "nll": -6.066773891448975, "elapsed_s": 526.9} {"event": "train", "step": 3980, "loss": -5.804901123046875, "grad_norm": 0.08341794461011887, "lr": 0.00024136437282723714, "ex_per_s": 481.97011887538605, "data_frac": 0.024676518442450477, "mem_mb": 63.172096, "nll": -5.804901123046875, "elapsed_s": 528.5} {"event": "train", "step": 4000, "loss": -6.0546875, "grad_norm": 1.0237009525299072, "lr": 0.0002407602538239216, "ex_per_s": 482.88813446394505, "data_frac": 0.024620759898405283, "mem_mb": 63.172096, "nll": -6.0546875, "elapsed_s": 530.2} {"event": "val", "step": 4000, "val_nll": -6.449017662456785, "best": -6.449017662456785, "bad": 0, "mem_mb": 63.172096, "elapsed_s": 592.8} {"event": "train", "step": 4020, "loss": -5.988595008850098, "grad_norm": 0.5850511193275452, "lr": 0.00024015391210388054, "ex_per_s": 432.76469956496106, "data_frac": 0.021973381254870004, "mem_mb": 63.172096, "nll": -5.988595008850098, "elapsed_s": 594.5} {"event": "train", "step": 4040, "loss": -6.13120174407959, "grad_norm": 0.5277165174484253, "lr": 0.00023954536545646676, "ex_per_s": 433.70269347013254, "data_frac": 0.02192944849121131, "mem_mb": 63.172096, "nll": -6.13120174407959, "elapsed_s": 596.2} {"event": "train", "step": 4060, "loss": -5.86958122253418, "grad_norm": 0.15309298038482666, "lr": 0.0002389346317357231, "ex_per_s": 434.63519931951674, "data_frac": 0.021885683541267263, "mem_mb": 63.172096, "nll": -5.86958122253418, "elapsed_s": 597.9} {"event": "train", "step": 4080, "loss": -5.929775714874268, "grad_norm": 0.2759925127029419, "lr": 0.0002383217288598584, "ex_per_s": 435.56319363616285, "data_frac": 0.021842169625427777, "mem_mb": 63.172096, "nll": -5.929775714874268, "elapsed_s": 599.5} {"event": "train", "step": 4100, "loss": -5.942184925079346, "grad_norm": 0.10391180217266083, "lr": 0.00023770667481072205, "ex_per_s": 436.48591548069123, "data_frac": 0.02179923994378526, "mem_mb": 63.172096, "nll": -5.942184925079346, "elapsed_s": 601.2} {"event": "train", "step": 4120, "loss": -6.106082916259766, "grad_norm": 0.3815624415874481, "lr": 0.0002370894876332764, "ex_per_s": 437.4035725716959, "data_frac": 0.021756049944611135, "mem_mb": 63.172096, "nll": -6.106082916259766, "elapsed_s": 602.9} {"event": "train", "step": 4140, "loss": -5.917718410491943, "grad_norm": 0.49873754382133484, "lr": 0.00023647018543506706, "ex_per_s": 438.3157230899145, "data_frac": 0.02171432345404453, "mem_mb": 63.172096, "nll": -5.917718410491943, "elapsed_s": 604.5} {"event": "train", "step": 4160, "loss": -6.135861873626709, "grad_norm": 0.08190706372261047, "lr": 0.00023584878638569206, "ex_per_s": 439.22396789820795, "data_frac": 0.021671625056085508, "mem_mb": 63.172096, "nll": -6.135861873626709, "elapsed_s": 606.2} {"event": "train", "step": 4180, "loss": -5.97361421585083, "grad_norm": 0.12099456787109375, "lr": 0.00023522530871626857, "ex_per_s": 440.1271989101843, "data_frac": 0.021629015720143044, "mem_mb": 63.172096, "nll": -5.97361421585083, "elapsed_s": 607.9} {"event": "train", "step": 4200, "loss": -6.030475616455078, "grad_norm": 0.4139261245727539, "lr": 0.0002345997707188979, "ex_per_s": 441.02427416088386, "data_frac": 0.02158865872068281, "mem_mb": 63.172096, "nll": -6.030475616455078, "elapsed_s": 609.5} {"event": "train", "step": 4220, "loss": -5.9651570320129395, "grad_norm": 0.48177316784858704, "lr": 0.00023397219074612914, "ex_per_s": 441.91946027551563, "data_frac": 0.02154728164328536, "mem_mb": 63.172096, "nll": -5.9651570320129395, "elapsed_s": 611.2} {"event": "train", "step": 4240, "loss": -5.995877265930176, "grad_norm": 0.5227839946746826, "lr": 0.00023334258721042025, "ex_per_s": 442.8086785319816, "data_frac": 0.02150616830834748, "mem_mb": 63.172096, "nll": -5.995877265930176, "elapsed_s": 612.9} {"event": "train", "step": 4260, "loss": -5.888874053955078, "grad_norm": 0.44130757451057434, "lr": 0.00023271097858359853, "ex_per_s": 443.69100080309374, "data_frac": 0.021466288070846447, "mem_mb": 63.172096, "nll": -5.888874053955078, "elapsed_s": 614.5} {"event": "train", "step": 4280, "loss": -5.85167121887207, "grad_norm": 0.08139557391405106, "lr": 0.00023207738339631794, "ex_per_s": 444.56607464932114, "data_frac": 0.021426107945876433, "mem_mb": 63.172096, "nll": -5.85167121887207, "elapsed_s": 616.2} {"event": "train", "step": 4300, "loss": -6.264202117919922, "grad_norm": 0.23190699517726898, "lr": 0.00023144182023751594, "ex_per_s": 445.43727453500367, "data_frac": 0.02138693435452946, "mem_mb": 63.172096, "nll": -6.264202117919922, "elapsed_s": 617.9} {"event": "train", "step": 4320, "loss": -6.130671501159668, "grad_norm": 0.9142364859580994, "lr": 0.00023080430775386808, "ex_per_s": 446.31214825784303, "data_frac": 0.021347744257896477, "mem_mb": 63.172096, "nll": -6.130671501159668, "elapsed_s": 619.5} {"event": "train", "step": 4340, "loss": -6.082424163818359, "grad_norm": 0.6023508310317993, "lr": 0.00023016486464924057, "ex_per_s": 447.1758031332629, "data_frac": 0.021309575323638967, "mem_mb": 63.172096, "nll": -6.082424163818359, "elapsed_s": 621.2} {"event": "train", "step": 4360, "loss": -5.84864616394043, "grad_norm": 0.6140193939208984, "lr": 0.0002295235096841419, "ex_per_s": 448.0432945020356, "data_frac": 0.021270624002705484, "mem_mb": 63.172096, "nll": -5.84864616394043, "elapsed_s": 622.8} {"event": "train", "step": 4380, "loss": -6.008236885070801, "grad_norm": 0.5911669731140137, "lr": 0.00022888026167517227, "ex_per_s": 448.9010298070629, "data_frac": 0.02123407127254171, "mem_mb": 63.172096, "nll": -6.008236885070801, "elapsed_s": 624.5} {"event": "train", "step": 4400, "loss": -6.048521041870117, "grad_norm": 0.2583250403404236, "lr": 0.00022823513949447166, "ex_per_s": 449.75906636988947, "data_frac": 0.021196354129307045, "mem_mb": 63.172096, "nll": -6.048521041870117, "elapsed_s": 626.2} {"event": "train", "step": 4420, "loss": -5.898808479309082, "grad_norm": 0.41499820351600647, "lr": 0.00022758816206916584, "ex_per_s": 450.61585553492955, "data_frac": 0.021157932671056395, "mem_mb": 63.172096, "nll": -5.898808479309082, "elapsed_s": 627.8} {"event": "train", "step": 4440, "loss": -6.091491222381592, "grad_norm": 0.3114559054374695, "lr": 0.00022693934838081143, "ex_per_s": 451.4630339610992, "data_frac": 0.02112105543607776, "mem_mb": 63.172096, "nll": -6.091491222381592, "elapsed_s": 629.5} {"event": "train", "step": 4460, "loss": -5.935984134674072, "grad_norm": 0.29296553134918213, "lr": 0.0002262887174648388, "ex_per_s": 452.31251076682787, "data_frac": 0.0210846248496339, "mem_mb": 63.172096, "nll": -5.935984134674072, "elapsed_s": 631.1} {"event": "train", "step": 4480, "loss": -6.062878608703613, "grad_norm": 0.20007658004760742, "lr": 0.00022563628840999367, "ex_per_s": 453.15599042737495, "data_frac": 0.021047356125102095, "mem_mb": 63.172096, "nll": -6.062878608703613, "elapsed_s": 632.8} {"event": "train", "step": 4500, "loss": -6.151100158691406, "grad_norm": 1.0532455444335938, "lr": 0.000224982080357777, "ex_per_s": 453.98665120977273, "data_frac": 0.021012550462921355, "mem_mb": 63.172096, "nll": -6.151100158691406, "elapsed_s": 634.4} {"event": "train", "step": 4520, "loss": -6.017053127288818, "grad_norm": 0.5221448540687561, "lr": 0.00022432611250188336, "ex_per_s": 454.82076858961835, "data_frac": 0.020976437156003957, "mem_mb": 63.172096, "nll": -6.017053127288818, "elapsed_s": 636.1} {"event": "train", "step": 4540, "loss": -6.008523464202881, "grad_norm": 0.7809440493583679, "lr": 0.00022366840408763816, "ex_per_s": 455.6545108254643, "data_frac": 0.02093917110594982, "mem_mb": 63.172096, "nll": -6.008523464202881, "elapsed_s": 637.7} {"event": "train", "step": 4560, "loss": -6.045727729797363, "grad_norm": 0.3165197968482971, "lr": 0.00022300897441143247, "ex_per_s": 456.4783422700639, "data_frac": 0.020902529466712925, "mem_mb": 63.172096, "nll": -6.045727729797363, "elapsed_s": 639.4} {"event": "train", "step": 4580, "loss": -6.019979000091553, "grad_norm": 0.9400734901428223, "lr": 0.0002223478428201573, "ex_per_s": 457.30104260426907, "data_frac": 0.020865551095896166, "mem_mb": 63.172096, "nll": -6.019979000091553, "elapsed_s": 641.0} {"event": "train", "step": 4600, "loss": -5.857340335845947, "grad_norm": 0.9673236012458801, "lr": 0.0002216850287106357, "ex_per_s": 458.1178849120131, "data_frac": 0.020829667914049877, "mem_mb": 63.172096, "nll": -5.857340335845947, "elapsed_s": 642.7} {"event": "train", "step": 4620, "loss": -6.031040668487549, "grad_norm": 0.16937778890132904, "lr": 0.00022102055152905406, "ex_per_s": 458.93108066740274, "data_frac": 0.020793526122604405, "mem_mb": 63.172096, "nll": -6.031040668487549, "elapsed_s": 644.3} {"event": "train", "step": 4640, "loss": -5.836943626403809, "grad_norm": 0.24049833416938782, "lr": 0.00022035443077039114, "ex_per_s": 459.7394088528926, "data_frac": 0.020757194708007905, "mem_mb": 63.172096, "nll": -5.836943626403809, "elapsed_s": 646.0} {"event": "train", "step": 4660, "loss": -5.98054838180542, "grad_norm": 0.4861999750137329, "lr": 0.00021968668597784646, "ex_per_s": 460.54703458027825, "data_frac": 0.02072167262480686, "mem_mb": 63.172096, "nll": -5.98054838180542, "elapsed_s": 647.6} {"event": "train", "step": 4680, "loss": -6.033840179443359, "grad_norm": 0.10847526788711548, "lr": 0.00021901733674226665, "ex_per_s": 461.3475541848972, "data_frac": 0.020686758020416644, "mem_mb": 63.172096, "nll": -6.033840179443359, "elapsed_s": 649.3} {"event": "train", "step": 4700, "loss": -5.9106621742248535, "grad_norm": 0.39642345905303955, "lr": 0.00021834640270157097, "ex_per_s": 462.14617943172846, "data_frac": 0.020651254319259117, "mem_mb": 63.172096, "nll": -5.9106621742248535, "elapsed_s": 650.9} {"event": "train", "step": 4720, "loss": -6.068989276885986, "grad_norm": 0.11651307344436646, "lr": 0.00021767390354017485, "ex_per_s": 462.9401235968654, "data_frac": 0.02061541878557859, "mem_mb": 63.172096, "nll": -6.068989276885986, "elapsed_s": 652.6} {"event": "train", "step": 4740, "loss": -6.0415191650390625, "grad_norm": 0.3979460597038269, "lr": 0.0002169998589884126, "ex_per_s": 463.730681929204, "data_frac": 0.02057978538872914, "mem_mb": 63.172096, "nll": -6.0415191650390625, "elapsed_s": 654.2} {"event": "train", "step": 4760, "loss": -5.995904445648193, "grad_norm": 0.7380417585372925, "lr": 0.00021632428882195843, "ex_per_s": 464.51527994352267, "data_frac": 0.02054465758639486, "mem_mb": 63.172096, "nll": -5.995904445648193, "elapsed_s": 655.9} {"event": "train", "step": 4780, "loss": -6.063714981079102, "grad_norm": 0.9543397426605225, "lr": 0.00021564721286124628, "ex_per_s": 465.2923073925672, "data_frac": 0.020509670492865423, "mem_mb": 63.172096, "nll": -6.063714981079102, "elapsed_s": 657.5} {"event": "train", "step": 4800, "loss": -6.11692476272583, "grad_norm": 0.2982669472694397, "lr": 0.00021496865097088845, "ex_per_s": 466.0654676335814, "data_frac": 0.02047530894596501, "mem_mb": 63.172096, "nll": -6.11692476272583, "elapsed_s": 659.2} {"event": "train", "step": 4820, "loss": -6.034458637237549, "grad_norm": 0.2549459934234619, "lr": 0.00021428862305909247, "ex_per_s": 466.8341057242957, "data_frac": 0.020441129842097384, "mem_mb": 63.172096, "nll": -6.034458637237549, "elapsed_s": 660.8} {"event": "train", "step": 4840, "loss": -5.937367916107178, "grad_norm": 0.12078606337308884, "lr": 0.00021360714907707733, "ex_per_s": 467.6426613374139, "data_frac": 0.020408808033552703, "mem_mb": 63.172096, "nll": -5.937367916107178, "elapsed_s": 662.4} {"event": "train", "step": 4860, "loss": -5.925081729888916, "grad_norm": 0.5514074563980103, "lr": 0.00021292424901848804, "ex_per_s": 468.4041506560077, "data_frac": 0.02037500797112407, "mem_mb": 63.172096, "nll": -5.925081729888916, "elapsed_s": 664.1} {"event": "train", "step": 4880, "loss": -6.0286054611206055, "grad_norm": 0.5464956164360046, "lr": 0.00021223994291880895, "ex_per_s": 469.1619209948904, "data_frac": 0.020341149125285434, "mem_mb": 63.172096, "nll": -6.0286054611206055, "elapsed_s": 665.7} {"event": "train", "step": 4900, "loss": -5.959802627563477, "grad_norm": 0.15207037329673767, "lr": 0.00021155425085477597, "ex_per_s": 469.9167325047674, "data_frac": 0.020307008116320257, "mem_mb": 63.172096, "nll": -5.959802627563477, "elapsed_s": 667.4} {"event": "train", "step": 4920, "loss": -5.951364994049072, "grad_norm": 0.19883348047733307, "lr": 0.00021086719294378776, "ex_per_s": 470.6678997926624, "data_frac": 0.02027317739592139, "mem_mb": 63.172096, "nll": -5.951364994049072, "elapsed_s": 669.0} {"event": "train", "step": 4940, "loss": -6.045966625213623, "grad_norm": 0.18041402101516724, "lr": 0.00021017878934331512, "ex_per_s": 471.4158288354038, "data_frac": 0.020240822005324653, "mem_mb": 63.172096, "nll": -6.045966625213623, "elapsed_s": 670.7} {"event": "train", "step": 4960, "loss": -6.062763214111328, "grad_norm": 0.43319758772850037, "lr": 0.00020948906025030994, "ex_per_s": 472.1615301302819, "data_frac": 0.020206947437123216, "mem_mb": 63.172096, "nll": -6.062763214111328, "elapsed_s": 672.4} {"event": "train", "step": 4980, "loss": -6.074322700500488, "grad_norm": 0.08253835886716843, "lr": 0.00020879802590061233, "ex_per_s": 472.9036485928949, "data_frac": 0.020173439385885995, "mem_mb": 63.172096, "nll": -6.074322700500488, "elapsed_s": 674.0} {"event": "train", "step": 5000, "loss": -5.884579181671143, "grad_norm": 0.4432583451271057, "lr": 0.00020810570656835728, "ex_per_s": 473.6400851103788, "data_frac": 0.02014042460776673, "mem_mb": 63.172096, "nll": -5.884579181671143, "elapsed_s": 675.7} {"event": "val", "step": 5000, "val_nll": -6.451341391408747, "best": -6.451341391408747, "bad": 0, "mem_mb": 63.172096, "elapsed_s": 738.4} {"event": "train", "step": 5020, "loss": -5.963519096374512, "grad_norm": 0.11028175055980682, "lr": 0.0002074121225653795, "ex_per_s": 434.14387999755394, "data_frac": 0.01840337406100457, "mem_mb": 63.172096, "nll": -5.963519096374512, "elapsed_s": 740.1} {"event": "train", "step": 5040, "loss": -6.095760822296143, "grad_norm": 0.33837297558784485, "lr": 0.00020671729424061788, "ex_per_s": 434.8918445762101, "data_frac": 0.01837603679585418, "mem_mb": 63.172096, "nll": -6.095760822296143, "elapsed_s": 741.7} {"event": "train", "step": 5060, "loss": -6.009047508239746, "grad_norm": 0.15719111263751984, "lr": 0.00020602124197951796, "ex_per_s": 435.6356452631175, "data_frac": 0.01834913787263624, "mem_mb": 63.172096, "nll": -6.009047508239746, "elapsed_s": 743.4} {"event": "train", "step": 5080, "loss": -6.130947113037109, "grad_norm": 0.4316101670265198, "lr": 0.00020532398620343433, "ex_per_s": 436.37723697019896, "data_frac": 0.018322123555049495, "mem_mb": 63.172096, "nll": -6.130947113037109, "elapsed_s": 745.1} {"event": "train", "step": 5100, "loss": -6.0651655197143555, "grad_norm": 0.13152849674224854, "lr": 0.0002046255473690312, "ex_per_s": 437.1138108259951, "data_frac": 0.01829602383072739, "mem_mb": 63.172096, "nll": -6.0651655197143555, "elapsed_s": 746.8} {"event": "train", "step": 5120, "loss": -6.004025459289551, "grad_norm": 0.30758246779441833, "lr": 0.00020392594596768237, "ex_per_s": 437.848563005452, "data_frac": 0.01826915345157043, "mem_mb": 63.172096, "nll": -6.004025459289551, "elapsed_s": 748.4} {"event": "train", "step": 5140, "loss": -5.945947647094727, "grad_norm": 1.1781293153762817, "lr": 0.00020322520252486997, "ex_per_s": 438.57842764190684, "data_frac": 0.018243955735686667, "mem_mb": 63.172096, "nll": -5.945947647094727, "elapsed_s": 750.1} {"event": "train", "step": 5160, "loss": -6.163727283477783, "grad_norm": 0.5714467167854309, "lr": 0.0002025233375995822, "ex_per_s": 439.308546497034, "data_frac": 0.018218019256924584, "mem_mb": 63.172096, "nll": -6.163727283477783, "elapsed_s": 751.8} {"event": "train", "step": 5180, "loss": -5.9410505294799805, "grad_norm": 0.28637781739234924, "lr": 0.0002018203717837105, "ex_per_s": 440.03524089154723, "data_frac": 0.018191759185883696, "mem_mb": 63.172096, "nll": -5.9410505294799805, "elapsed_s": 753.4} {"event": "train", "step": 5200, "loss": -5.825963973999023, "grad_norm": 0.07314343005418777, "lr": 0.00020111632570144484, "ex_per_s": 440.75611275117996, "data_frac": 0.01816670191858294, "mem_mb": 63.172096, "nll": -5.825963973999023, "elapsed_s": 755.1} {"event": "train", "step": 5220, "loss": -5.926499843597412, "grad_norm": 0.23671551048755646, "lr": 0.00020041122000866903, "ex_per_s": 441.47546491894417, "data_frac": 0.018140512529685602, "mem_mb": 63.172096, "nll": -5.926499843597412, "elapsed_s": 756.8} {"event": "train", "step": 5240, "loss": -5.778651237487793, "grad_norm": 0.10853724926710129, "lr": 0.00019970507539235469, "ex_per_s": 442.1927750732476, "data_frac": 0.018114613356323105, "mem_mb": 63.172096, "nll": -5.778651237487793, "elapsed_s": 758.4} {"event": "train", "step": 5260, "loss": -5.950465679168701, "grad_norm": 0.17484615743160248, "lr": 0.0001989979125699542, "ex_per_s": 442.9076764354519, "data_frac": 0.018089591316341806, "mem_mb": 63.172096, "nll": -5.950465679168701, "elapsed_s": 760.1} {"event": "train", "step": 5280, "loss": -6.006011009216309, "grad_norm": 0.6513213515281677, "lr": 0.00019828975228879288, "ex_per_s": 443.61893816396565, "data_frac": 0.018064001489041068, "mem_mb": 63.172096, "nll": -6.006011009216309, "elapsed_s": 761.8} {"event": "train", "step": 5300, "loss": -5.8948445320129395, "grad_norm": 0.1329512894153595, "lr": 0.00019758061532546025, "ex_per_s": 444.32771939736404, "data_frac": 0.01803874242006059, "mem_mb": 63.172096, "nll": -5.8948445320129395, "elapsed_s": 763.4} {"event": "train", "step": 5320, "loss": -6.074074745178223, "grad_norm": 0.8051087856292725, "lr": 0.0001968705224852007, "ex_per_s": 445.03294203584625, "data_frac": 0.018013821027906676, "mem_mb": 63.172096, "nll": -6.074074745178223, "elapsed_s": 765.1} {"event": "train", "step": 5340, "loss": -5.973247051239014, "grad_norm": 0.836872935295105, "lr": 0.00019615949460130276, "ex_per_s": 445.7358769645221, "data_frac": 0.017989141663322616, "mem_mb": 63.172096, "nll": -5.973247051239014, "elapsed_s": 766.8} {"event": "train", "step": 5360, "loss": -6.0123066902160645, "grad_norm": 0.5429688096046448, "lr": 0.00019544755253448823, "ex_per_s": 446.43787544000395, "data_frac": 0.017964230324843084, "mem_mb": 63.172096, "nll": -6.0123066902160645, "elapsed_s": 768.4} {"event": "train", "step": 5380, "loss": -5.862411022186279, "grad_norm": 0.09952083230018616, "lr": 0.0001947347171722997, "ex_per_s": 447.1359511196163, "data_frac": 0.017939984473633253, "mem_mb": 63.172096, "nll": -5.862411022186279, "elapsed_s": 770.1} {"event": "train", "step": 5400, "loss": -5.9537153244018555, "grad_norm": 0.28390973806381226, "lr": 0.00019402100942848827, "ex_per_s": 447.8318286139858, "data_frac": 0.017915805414347553, "mem_mb": 63.172096, "nll": -5.9537153244018555, "elapsed_s": 771.8} {"event": "train", "step": 5420, "loss": -5.837810516357422, "grad_norm": 0.2740914523601532, "lr": 0.00019330645024239946, "ex_per_s": 448.5259907338343, "data_frac": 0.017891121077094327, "mem_mb": 63.172096, "nll": -5.837810516357422, "elapsed_s": 773.4} {"event": "train", "step": 5440, "loss": -6.021592140197754, "grad_norm": 0.6354823112487793, "lr": 0.00019259106057835926, "ex_per_s": 449.21427709955935, "data_frac": 0.017866760128557386, "mem_mb": 63.172096, "nll": -6.021592140197754, "elapsed_s": 775.1} {"event": "train", "step": 5460, "loss": -6.100337982177734, "grad_norm": 0.2544364631175995, "lr": 0.00019187486142505883, "ex_per_s": 449.8953288276771, "data_frac": 0.01784335890247568, "mem_mb": 63.172096, "nll": -6.100337982177734, "elapsed_s": 776.8} {"event": "train", "step": 5480, "loss": -6.102818489074707, "grad_norm": 0.1958138346672058, "lr": 0.00019115787379493887, "ex_per_s": 450.5774527506075, "data_frac": 0.01781942727887436, "mem_mb": 63.172096, "nll": -6.102818489074707, "elapsed_s": 778.4} {"event": "train", "step": 5500, "loss": -6.082809925079346, "grad_norm": 0.7104038000106812, "lr": 0.00019044011872357296, "ex_per_s": 451.25393816288664, "data_frac": 0.017795991242438386, "mem_mb": 63.172096, "nll": -6.082809925079346, "elapsed_s": 780.1} {"event": "train", "step": 5520, "loss": -5.929188251495361, "grad_norm": 0.08591355383396149, "lr": 0.00018972161726905055, "ex_per_s": 451.9274268359975, "data_frac": 0.017772854237934316, "mem_mb": 63.172096, "nll": -5.929188251495361, "elapsed_s": 781.8} {"event": "train", "step": 5540, "loss": -5.922160625457764, "grad_norm": 0.6538212895393372, "lr": 0.00018900239051135913, "ex_per_s": 452.600201304547, "data_frac": 0.017749643289138323, "mem_mb": 63.172096, "nll": -5.922160625457764, "elapsed_s": 783.4} {"event": "train", "step": 5560, "loss": -5.980275630950928, "grad_norm": 0.5744074583053589, "lr": 0.00018828245955176568, "ex_per_s": 453.26916203362237, "data_frac": 0.017726443585478625, "mem_mb": 63.172096, "nll": -5.980275630950928, "elapsed_s": 785.1} {"event": "train", "step": 5580, "loss": -6.0292768478393555, "grad_norm": 0.5735440254211426, "lr": 0.00018756184551219765, "ex_per_s": 453.93451572186615, "data_frac": 0.01770282123687129, "mem_mb": 63.172096, "nll": -6.0292768478393555, "elapsed_s": 786.8} {"event": "train", "step": 5600, "loss": -6.086175918579102, "grad_norm": 0.326840877532959, "lr": 0.00018684056953462326, "ex_per_s": 454.60015143935794, "data_frac": 0.017678864208861705, "mem_mb": 63.172096, "nll": -6.086175918579102, "elapsed_s": 788.4} {"event": "train", "step": 5620, "loss": -5.822642803192139, "grad_norm": 0.09933295845985413, "lr": 0.00018611865278043117, "ex_per_s": 455.2621909692198, "data_frac": 0.017655815975823108, "mem_mb": 63.172096, "nll": -5.822642803192139, "elapsed_s": 790.1} {"event": "train", "step": 5640, "loss": -6.117214202880859, "grad_norm": 0.9376333355903625, "lr": 0.00018539611642980975, "ex_per_s": 455.92185215223384, "data_frac": 0.01763321557960828, "mem_mb": 63.172096, "nll": -6.117214202880859, "elapsed_s": 791.8} {"event": "train", "step": 5660, "loss": -6.010015964508057, "grad_norm": 0.5045461058616638, "lr": 0.00018467298168112556, "ex_per_s": 456.5778685620361, "data_frac": 0.017610170723156935, "mem_mb": 63.172096, "nll": -6.010015964508057, "elapsed_s": 793.4} {"event": "train", "step": 5680, "loss": -5.97000789642334, "grad_norm": 0.2903425991535187, "lr": 0.00018394926975030145, "ex_per_s": 457.2336462149869, "data_frac": 0.017586981338358097, "mem_mb": 63.172096, "nll": -5.97000789642334, "elapsed_s": 795.1} {"event": "train", "step": 5700, "loss": -6.035181045532227, "grad_norm": 0.6137940287590027, "lr": 0.00018322500187019415, "ex_per_s": 457.88729353930944, "data_frac": 0.017563944502899633, "mem_mb": 63.172096, "nll": -6.035181045532227, "elapsed_s": 796.7} {"event": "train", "step": 5720, "loss": -5.836230754852295, "grad_norm": 0.14391456544399261, "lr": 0.00018250019928997124, "ex_per_s": 458.5379997343351, "data_frac": 0.01754071876215441, "mem_mb": 63.172096, "nll": -5.836230754852295, "elapsed_s": 798.4} {"event": "train", "step": 5740, "loss": -5.999297142028809, "grad_norm": 0.4474968910217285, "lr": 0.00018177488327448776, "ex_per_s": 459.1843532558089, "data_frac": 0.0175184354170113, "mem_mb": 63.172096, "nll": -5.999297142028809, "elapsed_s": 800.1} {"event": "train", "step": 5760, "loss": -5.998048305511475, "grad_norm": 0.42669227719306946, "lr": 0.00018104907510366235, "ex_per_s": 459.8316107135414, "data_frac": 0.017495752040438906, "mem_mb": 63.172096, "nll": -5.998048305511475, "elapsed_s": 801.7} {"event": "train", "step": 5780, "loss": -6.012313365936279, "grad_norm": 0.15818586945533752, "lr": 0.00018032279607185304, "ex_per_s": 460.4758521672175, "data_frac": 0.01747302897658259, "mem_mb": 63.172096, "nll": -6.012313365936279, "elapsed_s": 803.4} {"event": "train", "step": 5800, "loss": -5.985251426696777, "grad_norm": 0.7574462890625, "lr": 0.00017959606748723213, "ex_per_s": 461.11746600518865, "data_frac": 0.01745039252197654, "mem_mb": 63.172096, "nll": -5.985251426696777, "elapsed_s": 805.0} {"event": "train", "step": 5820, "loss": -5.907803058624268, "grad_norm": 0.7600855827331543, "lr": 0.00017886891067116147, "ex_per_s": 461.75578396689247, "data_frac": 0.017428152651410008, "mem_mb": 63.172096, "nll": -5.907803058624268, "elapsed_s": 806.7} {"event": "train", "step": 5840, "loss": -5.911012649536133, "grad_norm": 0.42235249280929565, "lr": 0.00017814134695756664, "ex_per_s": 462.393076028716, "data_frac": 0.017406215423222852, "mem_mb": 63.172096, "nll": -5.911012649536133, "elapsed_s": 808.4} {"event": "train", "step": 5860, "loss": -6.003688812255859, "grad_norm": 0.14280028641223907, "lr": 0.00017741339769231104, "ex_per_s": 463.0280132453667, "data_frac": 0.017383855776519644, "mem_mb": 63.172096, "nll": -6.003688812255859, "elapsed_s": 810.0} {"event": "train", "step": 5880, "loss": -6.015819549560547, "grad_norm": 0.09298168867826462, "lr": 0.00017668508423256986, "ex_per_s": 463.66147744628984, "data_frac": 0.017362913525563684, "mem_mb": 63.172096, "nll": -6.015819549560547, "elapsed_s": 811.7} {"event": "train", "step": 5900, "loss": -6.138203144073486, "grad_norm": 0.5889818072319031, "lr": 0.0001759564279462032, "ex_per_s": 464.2919755836876, "data_frac": 0.01734066642374814, "mem_mb": 63.172096, "nll": -6.138203144073486, "elapsed_s": 813.3} {"event": "train", "step": 5920, "loss": -6.000471115112305, "grad_norm": 0.0827980637550354, "lr": 0.00017522745021112944, "ex_per_s": 464.92014478950256, "data_frac": 0.01731955132400747, "mem_mb": 63.172096, "nll": -6.000471115112305, "elapsed_s": 815.0} {"event": "train", "step": 5940, "loss": -5.944952964782715, "grad_norm": 0.17220646142959595, "lr": 0.0001744981724146976, "ex_per_s": 465.5476218796696, "data_frac": 0.017297669176425782, "mem_mb": 63.172096, "nll": -5.944952964782715, "elapsed_s": 816.6} {"event": "train", "step": 5960, "loss": -5.905746936798096, "grad_norm": 0.300721138715744, "lr": 0.00017376861595306043, "ex_per_s": 466.1735089742926, "data_frac": 0.017276063815224688, "mem_mb": 63.172096, "nll": -5.905746936798096, "elapsed_s": 818.3} {"event": "train", "step": 5980, "loss": -6.032399654388428, "grad_norm": 0.31951671838760376, "lr": 0.00017303880223054632, "ex_per_s": 466.79710789482476, "data_frac": 0.017254475082462754, "mem_mb": 63.172096, "nll": -6.032399654388428, "elapsed_s": 819.9} {"event": "train", "step": 6000, "loss": -5.991265296936035, "grad_norm": 0.25476381182670593, "lr": 0.00017230875265903138, "ex_per_s": 467.4182533323611, "data_frac": 0.017232847454053447, "mem_mb": 63.172096, "nll": -5.991265296936035, "elapsed_s": 821.6} {"event": "val", "step": 6000, "val_nll": -6.4510178656746975, "best": -6.451341391408747, "bad": 1, "mem_mb": 63.172096, "elapsed_s": 884.3} {"event": "train", "step": 6020, "loss": -5.845580577850342, "grad_norm": 0.10442115366458893, "lr": 0.0001715784886573111, "ex_per_s": 434.8740936896804, "data_frac": 0.01599186490852533, "mem_mb": 63.172096, "nll": -5.845580577850342, "elapsed_s": 886.0} {"event": "train", "step": 6040, "loss": -5.865780830383301, "grad_norm": 0.8336315751075745, "lr": 0.0001708480316504724, "ex_per_s": 435.49824473393966, "data_frac": 0.015973590344626185, "mem_mb": 63.172096, "nll": -5.865780830383301, "elapsed_s": 887.7} {"event": "train", "step": 6060, "loss": -6.025774955749512, "grad_norm": 0.6253706216812134, "lr": 0.00017011740306926445, "ex_per_s": 436.1206904714395, "data_frac": 0.015955482534398387, "mem_mb": 63.172096, "nll": -6.025774955749512, "elapsed_s": 889.3} {"event": "train", "step": 6080, "loss": -6.006911277770996, "grad_norm": 0.2337626814842224, "lr": 0.00016938662434947037, "ex_per_s": 436.7381030705624, "data_frac": 0.01593892432962375, "mem_mb": 63.172096, "nll": -6.006911277770996, "elapsed_s": 891.0} {"event": "train", "step": 6100, "loss": -6.139987468719482, "grad_norm": 0.6494742035865784, "lr": 0.00016865571693127818, "ex_per_s": 437.3567286372113, "data_frac": 0.0159210761540566, "mem_mb": 63.172096, "nll": -6.139987468719482, "elapsed_s": 892.7} {"event": "train", "step": 6120, "loss": -5.9965620040893555, "grad_norm": 0.1604577749967575, "lr": 0.00016792470225865166, "ex_per_s": 437.972111827, "data_frac": 0.015903231148994047, "mem_mb": 63.172096, "nll": -5.9965620040893555, "elapsed_s": 894.3} {"event": "train", "step": 6140, "loss": -6.004493236541748, "grad_norm": 0.205854132771492, "lr": 0.00016719360177870138, "ex_per_s": 438.5855675165876, "data_frac": 0.01588531890568884, "mem_mb": 63.172096, "nll": -6.004493236541748, "elapsed_s": 896.0} {"event": "train", "step": 6160, "loss": -6.0662713050842285, "grad_norm": 0.43451347947120667, "lr": 0.0001664624369410554, "ex_per_s": 439.1975003973213, "data_frac": 0.015867442644511825, "mem_mb": 63.172096, "nll": -6.0662713050842285, "elapsed_s": 897.7} {"event": "train", "step": 6180, "loss": -6.043064117431641, "grad_norm": 0.5282727479934692, "lr": 0.00016573122919722992, "ex_per_s": 439.8076396844412, "data_frac": 0.01584986920647434, "mem_mb": 63.172096, "nll": -6.043064117431641, "elapsed_s": 899.3} {"event": "train", "step": 6200, "loss": -5.989233016967773, "grad_norm": 0.13463576138019562, "lr": 0.000165, "ex_per_s": 440.41580347555083, "data_frac": 0.015832137114613672, "mem_mb": 63.172096, "nll": -5.989233016967773, "elapsed_s": 901.0} {"event": "train", "step": 6220, "loss": -5.934233665466309, "grad_norm": 0.12305151671171188, "lr": 0.00016426877080277007, "ex_per_s": 441.02099330391155, "data_frac": 0.015814469302190555, "mem_mb": 63.172096, "nll": -5.934233665466309, "elapsed_s": 902.7} {"event": "train", "step": 6240, "loss": -5.9045634269714355, "grad_norm": 0.4926239848136902, "lr": 0.0001635375630589446, "ex_per_s": 441.6231197916958, "data_frac": 0.01579748430108748, "mem_mb": 63.172096, "nll": -5.9045634269714355, "elapsed_s": 904.3} {"event": "train", "step": 6260, "loss": -5.910966396331787, "grad_norm": 0.08821208029985428, "lr": 0.00016280639822129864, "ex_per_s": 442.221509714068, "data_frac": 0.015781914735166, "mem_mb": 63.172096, "nll": -5.910966396331787, "elapsed_s": 906.0} {"event": "train", "step": 6280, "loss": -5.90130090713501, "grad_norm": 0.7586851119995117, "lr": 0.00016207529774134834, "ex_per_s": 442.8225232588008, "data_frac": 0.015764911043174054, "mem_mb": 63.172096, "nll": -5.90130090713501, "elapsed_s": 907.7} {"event": "train", "step": 6300, "loss": -5.986000061035156, "grad_norm": 0.22178561985492706, "lr": 0.00016134428306872187, "ex_per_s": 443.4203697553097, "data_frac": 0.015748249350431644, "mem_mb": 63.172096, "nll": -5.986000061035156, "elapsed_s": 909.3} {"event": "train", "step": 6320, "loss": -5.82438850402832, "grad_norm": 0.5996934771537781, "lr": 0.00016061337565052966, "ex_per_s": 444.0171608989584, "data_frac": 0.01573208991476889, "mem_mb": 63.172096, "nll": -5.82438850402832, "elapsed_s": 911.0} {"event": "train", "step": 6340, "loss": -5.885747909545898, "grad_norm": 0.15591992437839508, "lr": 0.00015988259693073558, "ex_per_s": 444.61125202906396, "data_frac": 0.015716130240149583, "mem_mb": 63.172096, "nll": -5.885747909545898, "elapsed_s": 912.7} {"event": "train", "step": 6360, "loss": -6.166389465332031, "grad_norm": 0.25792139768600464, "lr": 0.0001591519683495276, "ex_per_s": 445.20437050351916, "data_frac": 0.015699135550130198, "mem_mb": 63.172096, "nll": -6.166389465332031, "elapsed_s": 914.3} {"event": "train", "step": 6380, "loss": -5.900102138519287, "grad_norm": 0.5128691792488098, "lr": 0.0001584215113426889, "ex_per_s": 445.7948756611308, "data_frac": 0.015682444054440983, "mem_mb": 63.172096, "nll": -5.900102138519287, "elapsed_s": 916.0} {"event": "train", "step": 6400, "loss": -5.893473148345947, "grad_norm": 0.4209344685077667, "lr": 0.00015769124734096862, "ex_per_s": 446.38239652058706, "data_frac": 0.015666454613666822, "mem_mb": 63.172096, "nll": -5.893473148345947, "elapsed_s": 917.6} {"event": "train", "step": 6420, "loss": -5.984138011932373, "grad_norm": 0.33714768290519714, "lr": 0.00015696119776945368, "ex_per_s": 446.9702193169224, "data_frac": 0.015649580100569348, "mem_mb": 63.172096, "nll": -5.984138011932373, "elapsed_s": 919.3} {"event": "train", "step": 6440, "loss": -5.978874683380127, "grad_norm": 0.22228765487670898, "lr": 0.00015623138404693952, "ex_per_s": 447.55593793910293, "data_frac": 0.01563298054568029, "mem_mb": 63.172096, "nll": -5.978874683380127, "elapsed_s": 921.0} {"event": "train", "step": 6460, "loss": -5.928905963897705, "grad_norm": 0.5250478982925415, "lr": 0.0001555018275853024, "ex_per_s": 448.1399588037066, "data_frac": 0.01561628860016076, "mem_mb": 63.172096, "nll": -5.928905963897705, "elapsed_s": 922.6} {"event": "train", "step": 6480, "loss": -5.935487747192383, "grad_norm": 0.3554326295852661, "lr": 0.0001547725497888706, "ex_per_s": 448.7199688493219, "data_frac": 0.015600318295177171, "mem_mb": 63.172096, "nll": -5.935487747192383, "elapsed_s": 924.3} {"event": "train", "step": 6500, "loss": -5.958847999572754, "grad_norm": 0.13338904082775116, "lr": 0.0001540435720537968, "ex_per_s": 449.29793064137186, "data_frac": 0.015584983415871962, "mem_mb": 63.172096, "nll": -5.958847999572754, "elapsed_s": 925.9} {"event": "train", "step": 6520, "loss": -6.03798246383667, "grad_norm": 0.09675446152687073, "lr": 0.0001533149157674301, "ex_per_s": 449.8752307960371, "data_frac": 0.01556935165013719, "mem_mb": 63.172096, "nll": -6.03798246383667, "elapsed_s": 927.6} {"event": "train", "step": 6540, "loss": -6.160595893859863, "grad_norm": 0.13690438866615295, "lr": 0.00015258660230768893, "ex_per_s": 450.45061868319027, "data_frac": 0.015553542596535025, "mem_mb": 63.172096, "nll": -6.160595893859863, "elapsed_s": 929.2} {"event": "train", "step": 6560, "loss": -6.051356315612793, "grad_norm": 0.6654970049858093, "lr": 0.0001518586530424334, "ex_per_s": 451.0234938614236, "data_frac": 0.015537994597689894, "mem_mb": 63.172096, "nll": -6.051356315612793, "elapsed_s": 930.9} {"event": "train", "step": 6580, "loss": -6.044302463531494, "grad_norm": 0.636034369468689, "lr": 0.00015113108932883853, "ex_per_s": 451.5962681527937, "data_frac": 0.015522526235371058, "mem_mb": 63.172096, "nll": -6.044302463531494, "elapsed_s": 932.6} {"event": "train", "step": 6600, "loss": -6.01459264755249, "grad_norm": 0.35435131192207336, "lr": 0.00015040393251276787, "ex_per_s": 452.1682391410482, "data_frac": 0.01550641220197719, "mem_mb": 63.172096, "nll": -6.01459264755249, "elapsed_s": 934.2} {"event": "train", "step": 6620, "loss": -6.203455924987793, "grad_norm": 0.08980391174554825, "lr": 0.00014967720392814694, "ex_per_s": 452.73743467603305, "data_frac": 0.015490663391672519, "mem_mb": 63.172096, "nll": -6.203455924987793, "elapsed_s": 935.9} {"event": "train", "step": 6640, "loss": -6.102722644805908, "grad_norm": 0.22767707705497742, "lr": 0.00014895092489633762, "ex_per_s": 453.3046535204575, "data_frac": 0.015475455120345774, "mem_mb": 63.172096, "nll": -6.102722644805908, "elapsed_s": 937.5} {"event": "train", "step": 6660, "loss": -5.9721999168396, "grad_norm": 0.11184004694223404, "lr": 0.00014822511672551224, "ex_per_s": 453.8679946407155, "data_frac": 0.015460619642449016, "mem_mb": 63.172096, "nll": -5.9721999168396, "elapsed_s": 939.2} {"event": "train", "step": 6680, "loss": -5.798809051513672, "grad_norm": 0.21747368574142456, "lr": 0.0001474998007100288, "ex_per_s": 454.43150652909156, "data_frac": 0.015445535592289346, "mem_mb": 63.172096, "nll": -5.798809051513672, "elapsed_s": 940.8} {"event": "train", "step": 6700, "loss": -5.860888957977295, "grad_norm": 0.8249330520629883, "lr": 0.00014677499812980582, "ex_per_s": 454.99519156996456, "data_frac": 0.015430098155843962, "mem_mb": 63.172096, "nll": -5.860888957977295, "elapsed_s": 942.5} {"event": "train", "step": 6720, "loss": -6.2116265296936035, "grad_norm": 0.5215477347373962, "lr": 0.0001460507302496985, "ex_per_s": 455.5550612370481, "data_frac": 0.015414832901844602, "mem_mb": 63.172096, "nll": -6.2116265296936035, "elapsed_s": 944.1} {"event": "train", "step": 6740, "loss": -5.9734296798706055, "grad_norm": 0.30069291591644287, "lr": 0.0001453270183188744, "ex_per_s": 456.11494062084125, "data_frac": 0.015399261512600695, "mem_mb": 63.172096, "nll": -5.9734296798706055, "elapsed_s": 945.8} {"event": "train", "step": 6760, "loss": -5.904926300048828, "grad_norm": 0.26177921891212463, "lr": 0.00014460388357019025, "ex_per_s": 456.67241697381587, "data_frac": 0.015384045516572405, "mem_mb": 63.172096, "nll": -5.904926300048828, "elapsed_s": 947.4} {"event": "train", "step": 6780, "loss": -5.997712135314941, "grad_norm": 0.5915903449058533, "lr": 0.0001438813472195688, "ex_per_s": 457.2252457498495, "data_frac": 0.015369325859214843, "mem_mb": 63.172096, "nll": -5.997712135314941, "elapsed_s": 949.1} {"event": "train", "step": 6800, "loss": -6.091069221496582, "grad_norm": 0.25981003046035767, "lr": 0.00014315943046537674, "ex_per_s": 457.7799016636418, "data_frac": 0.015353839261386127, "mem_mb": 63.172096, "nll": -6.091069221496582, "elapsed_s": 950.7} {"event": "train", "step": 6820, "loss": -6.110374450683594, "grad_norm": 0.20165389776229858, "lr": 0.00014243815448780235, "ex_per_s": 458.32998073303213, "data_frac": 0.015339378212033428, "mem_mb": 63.172096, "nll": -6.110374450683594, "elapsed_s": 952.4} {"event": "train", "step": 6840, "loss": -5.831211090087891, "grad_norm": 0.47765493392944336, "lr": 0.00014171754044823431, "ex_per_s": 458.8800617762177, "data_frac": 0.01532406847501627, "mem_mb": 63.172096, "nll": -5.831211090087891, "elapsed_s": 954.0} {"event": "train", "step": 6860, "loss": -5.952399253845215, "grad_norm": 0.10546395182609558, "lr": 0.0001409976094886409, "ex_per_s": 459.42518239463794, "data_frac": 0.01530907021981324, "mem_mb": 63.172096, "nll": -5.952399253845215, "elapsed_s": 955.7} {"event": "train", "step": 6880, "loss": -6.026334285736084, "grad_norm": 0.09351973980665207, "lr": 0.00014027838273094942, "ex_per_s": 459.96847522395655, "data_frac": 0.015293907241170569, "mem_mb": 63.172096, "nll": -6.026334285736084, "elapsed_s": 957.3} {"event": "train", "step": 6900, "loss": -6.007990837097168, "grad_norm": 0.36722251772880554, "lr": 0.00013955988127642706, "ex_per_s": 460.50954414861593, "data_frac": 0.015278596261491806, "mem_mb": 63.172096, "nll": -6.007990837097168, "elapsed_s": 959.0} {"event": "train", "step": 6920, "loss": -6.066556930541992, "grad_norm": 0.6775072813034058, "lr": 0.0001388421262050611, "ex_per_s": 461.0490173644032, "data_frac": 0.015263825697914913, "mem_mb": 63.172096, "nll": -6.066556930541992, "elapsed_s": 960.6} {"event": "train", "step": 6940, "loss": -6.063632488250732, "grad_norm": 0.42312923073768616, "lr": 0.00013812513857494116, "ex_per_s": 461.5869297747408, "data_frac": 0.015248664686345123, "mem_mb": 63.172096, "nll": -6.063632488250732, "elapsed_s": 962.3} {"event": "train", "step": 6960, "loss": -5.74311637878418, "grad_norm": 0.1734769195318222, "lr": 0.00013740893942164073, "ex_per_s": 462.12320505266604, "data_frac": 0.015233435676769843, "mem_mb": 63.172096, "nll": -5.74311637878418, "elapsed_s": 963.9} {"event": "train", "step": 6980, "loss": -6.198734283447266, "grad_norm": 0.22104515135288239, "lr": 0.0001366935497576006, "ex_per_s": 462.65757209720203, "data_frac": 0.015218856973134007, "mem_mb": 63.172096, "nll": -6.198734283447266, "elapsed_s": 965.6} {"event": "train", "step": 7000, "loss": -6.127958297729492, "grad_norm": 0.46031367778778076, "lr": 0.00013597899057151173, "ex_per_s": 463.18940013900493, "data_frac": 0.015204333329347558, "mem_mb": 63.172096, "nll": -6.127958297729492, "elapsed_s": 967.3} {"event": "val", "step": 7000, "val_nll": -6.452620423016065, "best": -6.452620423016065, "bad": 0, "mem_mb": 63.172096, "elapsed_s": 1030.0} {"event": "train", "step": 7020, "loss": -6.143035888671875, "grad_norm": 0.18793991208076477, "lr": 0.0001352652828277003, "ex_per_s": 435.499166219203, "data_frac": 0.014266525213905068, "mem_mb": 63.172096, "nll": -6.143035888671875, "elapsed_s": 1031.7} {"event": "train", "step": 7040, "loss": -6.059739589691162, "grad_norm": 0.0796942263841629, "lr": 0.00013455244746551177, "ex_per_s": 436.0337657530127, "data_frac": 0.014254495805678676, "mem_mb": 63.172096, "nll": -6.059739589691162, "elapsed_s": 1033.4} {"event": "train", "step": 7060, "loss": -5.981436729431152, "grad_norm": 0.9037418365478516, "lr": 0.00013384050539869724, "ex_per_s": 436.5668464209024, "data_frac": 0.014241905913391724, "mem_mb": 63.172096, "nll": -5.981436729431152, "elapsed_s": 1035.0} {"event": "train", "step": 7080, "loss": -6.0684494972229, "grad_norm": 1.0192948579788208, "lr": 0.00013312947751479927, "ex_per_s": 437.09831850824924, "data_frac": 0.014229407521342129, "mem_mb": 63.172096, "nll": -6.0684494972229, "elapsed_s": 1036.7} {"event": "train", "step": 7100, "loss": -5.966437339782715, "grad_norm": 0.26642298698425293, "lr": 0.00013241938467453975, "ex_per_s": 437.6282808137786, "data_frac": 0.014216881195762908, "mem_mb": 63.172096, "nll": -5.966437339782715, "elapsed_s": 1038.4} {"event": "train", "step": 7120, "loss": -6.044669151306152, "grad_norm": 0.11367853730916977, "lr": 0.00013171024771120715, "ex_per_s": 438.1570401644622, "data_frac": 0.014204099467615638, "mem_mb": 63.172096, "nll": -6.044669151306152, "elapsed_s": 1040.0} {"event": "train", "step": 7140, "loss": -5.94468879699707, "grad_norm": 0.17536327242851257, "lr": 0.0001310020874300458, "ex_per_s": 438.682336045985, "data_frac": 0.014191540153530201, "mem_mb": 63.172096, "nll": -5.94468879699707, "elapsed_s": 1041.7} {"event": "train", "step": 7160, "loss": -6.030123233795166, "grad_norm": 0.4111756980419159, "lr": 0.00013029492460764526, "ex_per_s": 439.2074813853688, "data_frac": 0.014179117231880667, "mem_mb": 63.172096, "nll": -6.030123233795166, "elapsed_s": 1043.4} {"event": "train", "step": 7180, "loss": -5.949474334716797, "grad_norm": 0.5399172902107239, "lr": 0.00012958877999133094, "ex_per_s": 439.7299404551598, "data_frac": 0.0141666337222829, "mem_mb": 63.172096, "nll": -5.949474334716797, "elapsed_s": 1045.0} {"event": "train", "step": 7200, "loss": -5.964656829833984, "grad_norm": 0.12445804476737976, "lr": 0.0001288836742985552, "ex_per_s": 440.2518838347297, "data_frac": 0.01415406358861941, "mem_mb": 63.172096, "nll": -5.964656829833984, "elapsed_s": 1046.7} {"event": "train", "step": 7220, "loss": -6.0633697509765625, "grad_norm": 0.09721891582012177, "lr": 0.00012817962821628953, "ex_per_s": 440.771863660748, "data_frac": 0.014141807372200096, "mem_mb": 63.172096, "nll": -6.0633697509765625, "elapsed_s": 1048.4} {"event": "train", "step": 7240, "loss": -5.958276748657227, "grad_norm": 0.21005748212337494, "lr": 0.00012747666240041779, "ex_per_s": 441.28977911196074, "data_frac": 0.014129624870399534, "mem_mb": 63.172096, "nll": -5.958276748657227, "elapsed_s": 1050.1} {"event": "train", "step": 7260, "loss": -6.075646877288818, "grad_norm": 0.5768629908561707, "lr": 0.00012677479747513008, "ex_per_s": 441.8061290533091, "data_frac": 0.014117706225896275, "mem_mb": 63.172096, "nll": -6.075646877288818, "elapsed_s": 1051.7} {"event": "train", "step": 7280, "loss": -5.869856357574463, "grad_norm": 0.5724324584007263, "lr": 0.00012607405403231763, "ex_per_s": 442.32118801865465, "data_frac": 0.014106203092735333, "mem_mb": 63.172096, "nll": -5.869856357574463, "elapsed_s": 1053.4} {"event": "train", "step": 7300, "loss": -5.925808906555176, "grad_norm": 0.8045802712440491, "lr": 0.0001253744526309688, "ex_per_s": 442.835510448364, "data_frac": 0.014094126382167785, "mem_mb": 63.172096, "nll": -5.925808906555176, "elapsed_s": 1055.1} {"event": "train", "step": 7320, "loss": -6.073298454284668, "grad_norm": 0.15662191808223724, "lr": 0.00012467601379656567, "ex_per_s": 443.3474451050441, "data_frac": 0.01408244276092795, "mem_mb": 63.172096, "nll": -6.073298454284668, "elapsed_s": 1056.7} {"event": "train", "step": 7340, "loss": -6.013782024383545, "grad_norm": 0.23902083933353424, "lr": 0.000123978758020482, "ex_per_s": 443.8590893090002, "data_frac": 0.014070374799582508, "mem_mb": 63.172096, "nll": -6.013782024383545, "elapsed_s": 1058.4} {"event": "train", "step": 7360, "loss": -5.995161533355713, "grad_norm": 0.32357728481292725, "lr": 0.00012328270575938212, "ex_per_s": 444.3693443822925, "data_frac": 0.014058446156679519, "mem_mb": 63.172096, "nll": -5.995161533355713, "elapsed_s": 1060.1} {"event": "train", "step": 7380, "loss": -6.078542709350586, "grad_norm": 0.18394799530506134, "lr": 0.00012258787743462041, "ex_per_s": 444.8797380694157, "data_frac": 0.014046596823032727, "mem_mb": 63.172096, "nll": -6.078542709350586, "elapsed_s": 1061.7} {"event": "train", "step": 7400, "loss": -6.002654075622559, "grad_norm": 0.4082891643047333, "lr": 0.00012189429343164274, "ex_per_s": 445.3886616981001, "data_frac": 0.014034831483114702, "mem_mb": 63.172096, "nll": -6.002654075622559, "elapsed_s": 1063.4} {"event": "train", "step": 7420, "loss": -6.04995584487915, "grad_norm": 0.10554604977369308, "lr": 0.00012120197409938763, "ex_per_s": 445.8959650537012, "data_frac": 0.014022788608154384, "mem_mb": 63.172096, "nll": -6.04995584487915, "elapsed_s": 1065.0} {"event": "train", "step": 7440, "loss": -5.782918453216553, "grad_norm": 0.40111109614372253, "lr": 0.00012051093974969008, "ex_per_s": 446.4019065362579, "data_frac": 0.014010957714179833, "mem_mb": 63.172096, "nll": -5.782918453216553, "elapsed_s": 1066.7} {"event": "train", "step": 7460, "loss": -6.212433338165283, "grad_norm": 0.19006435573101044, "lr": 0.00011982121065668482, "ex_per_s": 446.9052117231076, "data_frac": 0.013999526848268513, "mem_mb": 63.172096, "nll": -6.212433338165283, "elapsed_s": 1068.4} {"event": "train", "step": 7480, "loss": -5.966171741485596, "grad_norm": 0.2248275727033615, "lr": 0.00011913280705621224, "ex_per_s": 447.4084803369072, "data_frac": 0.01398765675496729, "mem_mb": 63.172096, "nll": -5.966171741485596, "elapsed_s": 1070.0} {"event": "train", "step": 7500, "loss": -5.777978897094727, "grad_norm": 0.09725591540336609, "lr": 0.00011844574914522403, "ex_per_s": 447.90888442257904, "data_frac": 0.01397628791363518, "mem_mb": 63.172096, "nll": -5.777978897094727, "elapsed_s": 1071.7} {"event": "train", "step": 7520, "loss": -6.059506416320801, "grad_norm": 0.633493185043335, "lr": 0.00011776005708119105, "ex_per_s": 448.4085239596869, "data_frac": 0.013965612501412698, "mem_mb": 63.172096, "nll": -6.059506416320801, "elapsed_s": 1073.4} {"event": "train", "step": 7540, "loss": -5.845093727111816, "grad_norm": 0.49088048934936523, "lr": 0.00011707575098151199, "ex_per_s": 448.90474368344593, "data_frac": 0.013954379444889026, "mem_mb": 63.172096, "nll": -5.845093727111816, "elapsed_s": 1075.0} {"event": "train", "step": 7560, "loss": -5.903280258178711, "grad_norm": 0.38983631134033203, "lr": 0.00011639285092292267, "ex_per_s": 449.3992040503426, "data_frac": 0.013942780871253432, "mem_mb": 63.172096, "nll": -5.903280258178711, "elapsed_s": 1076.7} {"event": "train", "step": 7580, "loss": -5.869455814361572, "grad_norm": 0.334137499332428, "lr": 0.00011571137694090757, "ex_per_s": 449.8916592130112, "data_frac": 0.013931168618652612, "mem_mb": 63.172096, "nll": -5.869455814361572, "elapsed_s": 1078.3} {"event": "train", "step": 7600, "loss": -6.059781074523926, "grad_norm": 0.06554337590932846, "lr": 0.00011503134902911152, "ex_per_s": 450.3829848966832, "data_frac": 0.01391974269315809, "mem_mb": 63.172096, "nll": -6.059781074523926, "elapsed_s": 1080.0} {"event": "train", "step": 7620, "loss": -5.949527740478516, "grad_norm": 0.3119913339614868, "lr": 0.00011435278713875375, "ex_per_s": 450.8715388303437, "data_frac": 0.013908285364373297, "mem_mb": 63.172096, "nll": -5.949527740478516, "elapsed_s": 1081.7} {"event": "train", "step": 7640, "loss": -5.9106526374816895, "grad_norm": 0.9023550748825073, "lr": 0.00011367571117804159, "ex_per_s": 451.36028501080386, "data_frac": 0.013897279393726019, "mem_mb": 63.172096, "nll": -5.9106526374816895, "elapsed_s": 1083.3} {"event": "train", "step": 7660, "loss": -6.055073261260986, "grad_norm": 0.6704477071762085, "lr": 0.00011300014101158744, "ex_per_s": 451.8477230803373, "data_frac": 0.013885820631278615, "mem_mb": 63.172096, "nll": -6.055073261260986, "elapsed_s": 1085.0} {"event": "train", "step": 7680, "loss": -5.966360092163086, "grad_norm": 0.1272115707397461, "lr": 0.00011232609645982515, "ex_per_s": 452.3307329152857, "data_frac": 0.013874734564464152, "mem_mb": 63.172096, "nll": -5.966360092163086, "elapsed_s": 1086.7} {"event": "train", "step": 7700, "loss": -5.911745071411133, "grad_norm": 0.22376087307929993, "lr": 0.00011165359729842902, "ex_per_s": 452.8140531247324, "data_frac": 0.013863680288563947, "mem_mb": 63.172096, "nll": -5.911745071411133, "elapsed_s": 1088.3} {"event": "train", "step": 7720, "loss": -6.039160251617432, "grad_norm": 0.1337025761604309, "lr": 0.00011098266325773331, "ex_per_s": 453.29657987859093, "data_frac": 0.013852676815500243, "mem_mb": 63.172096, "nll": -6.039160251617432, "elapsed_s": 1090.0} {"event": "train", "step": 7740, "loss": -6.050570487976074, "grad_norm": 0.6986706852912903, "lr": 0.00011031331402215353, "ex_per_s": 453.7775551713585, "data_frac": 0.01384147298615776, "mem_mb": 63.172096, "nll": -6.050570487976074, "elapsed_s": 1091.7} {"event": "train", "step": 7760, "loss": -5.964038848876953, "grad_norm": 0.7634562849998474, "lr": 0.00010964556922960883, "ex_per_s": 454.2565596825296, "data_frac": 0.013830998599737318, "mem_mb": 63.172096, "nll": -5.964038848876953, "elapsed_s": 1093.3} {"event": "train", "step": 7780, "loss": -6.050996780395508, "grad_norm": 0.2514401376247406, "lr": 0.00010897944847094595, "ex_per_s": 454.7353133597909, "data_frac": 0.013820297756969697, "mem_mb": 63.172096, "nll": -6.050996780395508, "elapsed_s": 1095.0} {"event": "train", "step": 7800, "loss": -6.058783054351807, "grad_norm": 0.09801717847585678, "lr": 0.00010831497128936428, "ex_per_s": 455.2123848291423, "data_frac": 0.01380976852129098, "mem_mb": 63.172096, "nll": -6.058783054351807, "elapsed_s": 1096.7} {"event": "train", "step": 7820, "loss": -5.90444278717041, "grad_norm": 0.10992229729890823, "lr": 0.00010765215717984271, "ex_per_s": 455.6887728998113, "data_frac": 0.013799023586899556, "mem_mb": 63.172096, "nll": -5.90444278717041, "elapsed_s": 1098.3} {"event": "train", "step": 7840, "loss": -5.9600396156311035, "grad_norm": 0.0855255201458931, "lr": 0.00010699102558856754, "ex_per_s": 456.16235201288805, "data_frac": 0.013788387668001521, "mem_mb": 63.172096, "nll": -5.9600396156311035, "elapsed_s": 1100.0} {"event": "train", "step": 7860, "loss": -6.119227409362793, "grad_norm": 0.3351425230503082, "lr": 0.0001063315959123618, "ex_per_s": 456.6380140085002, "data_frac": 0.01377753438584892, "mem_mb": 63.172096, "nll": -6.119227409362793, "elapsed_s": 1101.7} {"event": "train", "step": 7880, "loss": -6.06404972076416, "grad_norm": 0.16217826306819916, "lr": 0.0001056738874981166, "ex_per_s": 457.1112616949043, "data_frac": 0.01376695935024472, "mem_mb": 63.172096, "nll": -6.06404972076416, "elapsed_s": 1103.3} {"event": "train", "step": 7900, "loss": -6.058303356170654, "grad_norm": 0.36136093735694885, "lr": 0.00010501791964222304, "ex_per_s": 457.58305786277805, "data_frac": 0.013756549399677603, "mem_mb": 63.172096, "nll": -6.058303356170654, "elapsed_s": 1105.0} {"event": "train", "step": 7920, "loss": -6.087601661682129, "grad_norm": 0.1115318164229393, "lr": 0.00010436371159000636, "ex_per_s": 458.05332743453295, "data_frac": 0.013745720703270885, "mem_mb": 63.172096, "nll": -6.087601661682129, "elapsed_s": 1106.6} {"event": "train", "step": 7940, "loss": -6.047421455383301, "grad_norm": 0.14144843816757202, "lr": 0.00010371128253516118, "ex_per_s": 458.52243756174965, "data_frac": 0.013735388851968643, "mem_mb": 63.172096, "nll": -6.047421455383301, "elapsed_s": 1108.3} {"event": "train", "step": 7960, "loss": -5.993412971496582, "grad_norm": 0.15868671238422394, "lr": 0.00010306065161918857, "ex_per_s": 458.9903248969247, "data_frac": 0.013724650341187212, "mem_mb": 63.172096, "nll": -5.993412971496582, "elapsed_s": 1110.0} {"event": "train", "step": 7980, "loss": -5.841272830963135, "grad_norm": 0.12058326601982117, "lr": 0.00010241183793083414, "ex_per_s": 459.4578246472062, "data_frac": 0.013713939469465149, "mem_mb": 63.172096, "nll": -5.841272830963135, "elapsed_s": 1111.6} {"event": "train", "step": 8000, "loss": -6.048815727233887, "grad_norm": 0.10572365671396255, "lr": 0.00010176486050552833, "ex_per_s": 459.9234417367169, "data_frac": 0.013704269989759677, "mem_mb": 63.172096, "nll": -6.048815727233887, "elapsed_s": 1113.3} {"event": "val", "step": 8000, "val_nll": -6.4525101503040005, "best": -6.452620423016065, "bad": 1, "mem_mb": 63.172096, "elapsed_s": 1176.0} {"event": "train", "step": 8020, "loss": -6.045403957366943, "grad_norm": 0.23722118139266968, "lr": 0.00010111973832482765, "ex_per_s": 435.8547830072498, "data_frac": 0.01296456806697792, "mem_mb": 63.172096, "nll": -6.045403957366943, "elapsed_s": 1177.7} {"event": "train", "step": 8040, "loss": -6.0338358879089355, "grad_norm": 0.19862113893032074, "lr": 0.00010047649031585807, "ex_per_s": 436.3230896925882, "data_frac": 0.012955086251741792, "mem_mb": 63.172096, "nll": -6.0338358879089355, "elapsed_s": 1179.4} {"event": "train", "step": 8060, "loss": -5.833107948303223, "grad_norm": 0.7448070049285889, "lr": 9.983513535075943e-05, "ex_per_s": 436.7913377794145, "data_frac": 0.012945721850947629, "mem_mb": 63.172096, "nll": -5.833107948303223, "elapsed_s": 1181.0} {"event": "train", "step": 8080, "loss": -5.996209144592285, "grad_norm": 0.21905829012393951, "lr": 9.919569224613192e-05, "ex_per_s": 437.2577389480707, "data_frac": 0.012936420646612655, "mem_mb": 63.172096, "nll": -5.996209144592285, "elapsed_s": 1182.7} {"event": "train", "step": 8100, "loss": -5.941924571990967, "grad_norm": 0.1042368933558464, "lr": 9.855817976248399e-05, "ex_per_s": 437.7227267835294, "data_frac": 0.012927140876253792, "mem_mb": 63.172096, "nll": -5.941924571990967, "elapsed_s": 1184.4} {"event": "train", "step": 8120, "loss": -5.902727127075195, "grad_norm": 0.5438287854194641, "lr": 9.792261660368207e-05, "ex_per_s": 438.1855883902168, "data_frac": 0.012917914747091749, "mem_mb": 63.172096, "nll": -5.902727127075195, "elapsed_s": 1186.0} {"event": "train", "step": 8140, "loss": -5.896433353424072, "grad_norm": 0.10775536298751831, "lr": 9.72890214164015e-05, "ex_per_s": 438.64749268231395, "data_frac": 0.012909033454447218, "mem_mb": 63.172096, "nll": -5.896433353424072, "elapsed_s": 1187.7} {"event": "train", "step": 8160, "loss": -6.042757034301758, "grad_norm": 0.2162817120552063, "lr": 9.66574127895797e-05, "ex_per_s": 439.108325325323, "data_frac": 0.0128996749150138, "mem_mb": 63.172096, "nll": -6.042757034301758, "elapsed_s": 1189.4} {"event": "train", "step": 8180, "loss": -5.94603157043457, "grad_norm": 0.6871351599693298, "lr": 9.602780925387089e-05, "ex_per_s": 439.56723120285227, "data_frac": 0.012890354590748348, "mem_mb": 63.172096, "nll": -5.94603157043457, "elapsed_s": 1191.0} {"event": "train", "step": 8200, "loss": -6.174124717712402, "grad_norm": 0.22373701632022858, "lr": 9.540022928110205e-05, "ex_per_s": 440.02410163104787, "data_frac": 0.012881501994412565, "mem_mb": 63.172096, "nll": -6.174124717712402, "elapsed_s": 1192.7} {"event": "train", "step": 8220, "loss": -5.907902717590332, "grad_norm": 0.20531129837036133, "lr": 9.477469128373143e-05, "ex_per_s": 440.4804084544424, "data_frac": 0.01287300033604471, "mem_mb": 63.172096, "nll": -5.907902717590332, "elapsed_s": 1194.4} {"event": "train", "step": 8240, "loss": -5.841577053070068, "grad_norm": 0.2810773551464081, "lr": 9.415121361430792e-05, "ex_per_s": 440.9366450634489, "data_frac": 0.012864105844562478, "mem_mb": 63.172096, "nll": -5.841577053070068, "elapsed_s": 1196.0} {"event": "train", "step": 8260, "loss": -6.188292026519775, "grad_norm": 0.53098464012146, "lr": 9.352981456493298e-05, "ex_per_s": 441.3920815247135, "data_frac": 0.012855046553762961, "mem_mb": 63.172096, "nll": -6.188292026519775, "elapsed_s": 1197.7} {"event": "train", "step": 8280, "loss": -6.154677391052246, "grad_norm": 0.08171036839485168, "lr": 9.291051236672361e-05, "ex_per_s": 441.84659974262036, "data_frac": 0.012846295857153507, "mem_mb": 63.172096, "nll": -6.154677391052246, "elapsed_s": 1199.4} {"event": "train", "step": 8300, "loss": -6.048665523529053, "grad_norm": 0.1065034568309784, "lr": 9.229332518927795e-05, "ex_per_s": 442.2995576633166, "data_frac": 0.01283731316056153, "mem_mb": 63.172096, "nll": -6.048665523529053, "elapsed_s": 1201.0} {"event": "train", "step": 8320, "loss": -5.893360614776611, "grad_norm": 0.19964249432086945, "lr": 9.167827114014155e-05, "ex_per_s": 442.751015706744, "data_frac": 0.012828690608300029, "mem_mb": 63.172096, "nll": -5.893360614776611, "elapsed_s": 1202.7} {"event": "train", "step": 8340, "loss": -6.060128688812256, "grad_norm": 0.5994720458984375, "lr": 9.106536826427686e-05, "ex_per_s": 443.20311555563273, "data_frac": 0.012819827693446114, "mem_mb": 63.172096, "nll": -6.060128688812256, "elapsed_s": 1204.4} {"event": "train", "step": 8360, "loss": -6.014883518218994, "grad_norm": 0.162236750125885, "lr": 9.045463454353313e-05, "ex_per_s": 443.6535170166261, "data_frac": 0.012811295992001152, "mem_mb": 63.172096, "nll": -6.014883518218994, "elapsed_s": 1206.0} {"event": "train", "step": 8380, "loss": -6.0564045906066895, "grad_norm": 0.1596960574388504, "lr": 8.984608789611947e-05, "ex_per_s": 444.1028729536987, "data_frac": 0.01280306998020508, "mem_mb": 63.172096, "nll": -6.0564045906066895, "elapsed_s": 1207.7} {"event": "train", "step": 8400, "loss": -6.036610126495361, "grad_norm": 0.2125064879655838, "lr": 8.92397461760784e-05, "ex_per_s": 444.5521329010374, "data_frac": 0.012794474696391717, "mem_mb": 63.172096, "nll": -6.036610126495361, "elapsed_s": 1209.4} {"event": "train", "step": 8420, "loss": -6.003643989562988, "grad_norm": 0.15908116102218628, "lr": 8.863562717276283e-05, "ex_per_s": 444.99881827740035, "data_frac": 0.012786114000167229, "mem_mb": 63.172096, "nll": -6.003643989562988, "elapsed_s": 1211.0} {"event": "train", "step": 8440, "loss": -6.0317254066467285, "grad_norm": 0.3935907185077667, "lr": 8.803374861031363e-05, "ex_per_s": 445.44570701002067, "data_frac": 0.01277841411411241, "mem_mb": 63.172096, "nll": -6.0317254066467285, "elapsed_s": 1212.7} {"event": "train", "step": 8460, "loss": -5.912224769592285, "grad_norm": 0.39181002974510193, "lr": 8.743412814713985e-05, "ex_per_s": 445.8917871683247, "data_frac": 0.012770239716006743, "mem_mb": 63.172096, "nll": -5.912224769592285, "elapsed_s": 1214.3} {"event": "train", "step": 8480, "loss": -5.880106449127197, "grad_norm": 0.2561168074607849, "lr": 8.683678337540044e-05, "ex_per_s": 446.3375573260238, "data_frac": 0.012761624495935502, "mem_mb": 63.172096, "nll": -5.880106449127197, "elapsed_s": 1216.0} {"event": "train", "step": 8500, "loss": -5.96785306930542, "grad_norm": 0.7290065884590149, "lr": 8.62417318204883e-05, "ex_per_s": 446.7811415847918, "data_frac": 0.01275415175536077, "mem_mb": 63.172096, "nll": -5.96785306930542, "elapsed_s": 1217.6} {"event": "train", "step": 8520, "loss": -5.934292316436768, "grad_norm": 0.5016697645187378, "lr": 8.564899094051614e-05, "ex_per_s": 447.2246619527779, "data_frac": 0.012745738329191087, "mem_mb": 63.172096, "nll": -5.934292316436768, "elapsed_s": 1219.3} {"event": "train", "step": 8540, "loss": -5.9071455001831055, "grad_norm": 0.3838145434856415, "lr": 8.505857812580402e-05, "ex_per_s": 447.6655568235119, "data_frac": 0.012737375751037361, "mem_mb": 63.172096, "nll": -5.9071455001831055, "elapsed_s": 1221.0} {"event": "train", "step": 8560, "loss": -5.969155788421631, "grad_norm": 0.19728022813796997, "lr": 8.447051069836941e-05, "ex_per_s": 448.1067468613782, "data_frac": 0.012729258223815262, "mem_mb": 63.172096, "nll": -5.969155788421631, "elapsed_s": 1222.6} {"event": "train", "step": 8580, "loss": -5.959328651428223, "grad_norm": 0.0832996517419815, "lr": 8.38848059114188e-05, "ex_per_s": 448.5474746641875, "data_frac": 0.012721114598250812, "mem_mb": 63.172096, "nll": -5.959328651428223, "elapsed_s": 1224.3} {"event": "train", "step": 8600, "loss": -6.023653507232666, "grad_norm": 0.12594954669475555, "lr": 8.330148094884171e-05, "ex_per_s": 448.98660642963034, "data_frac": 0.012712924392463408, "mem_mb": 63.172096, "nll": -6.023653507232666, "elapsed_s": 1225.9} {"event": "train", "step": 8620, "loss": -5.995419979095459, "grad_norm": 0.4333614706993103, "lr": 8.27205529247063e-05, "ex_per_s": 449.4245949497393, "data_frac": 0.012704530713900564, "mem_mb": 63.172096, "nll": -5.995419979095459, "elapsed_s": 1227.6} {"event": "train", "step": 8640, "loss": -5.9659552574157715, "grad_norm": 0.7941544651985168, "lr": 8.214203888275743e-05, "ex_per_s": 449.8611701797482, "data_frac": 0.012696212532186892, "mem_mb": 63.172096, "nll": -5.9659552574157715, "elapsed_s": 1229.2} {"event": "train", "step": 8660, "loss": -5.950146198272705, "grad_norm": 0.11132169514894485, "lr": 8.156595579591652e-05, "ex_per_s": 450.2956548642715, "data_frac": 0.0126880673292909, "mem_mb": 63.172096, "nll": -5.950146198272705, "elapsed_s": 1230.9} {"event": "train", "step": 8680, "loss": -5.904598712921143, "grad_norm": 0.3401385545730591, "lr": 8.09923205657837e-05, "ex_per_s": 450.73041582765376, "data_frac": 0.012679671850813168, "mem_mb": 63.172096, "nll": -5.904598712921143, "elapsed_s": 1232.5} {"event": "train", "step": 8700, "loss": -5.932109355926514, "grad_norm": 0.14903250336647034, "lr": 8.042115002214181e-05, "ex_per_s": 451.16289227723917, "data_frac": 0.01267201908044638, "mem_mb": 63.172096, "nll": -5.932109355926514, "elapsed_s": 1234.2} {"event": "train", "step": 8720, "loss": -6.164109230041504, "grad_norm": 0.2782130539417267, "lr": 7.985246092246267e-05, "ex_per_s": 451.5956099077527, "data_frac": 0.012663954177439, "mem_mb": 63.172096, "nll": -6.164109230041504, "elapsed_s": 1235.8} {"event": "train", "step": 8740, "loss": -5.901725769042969, "grad_norm": 0.05761490389704704, "lr": 7.928626995141546e-05, "ex_per_s": 452.0265520206857, "data_frac": 0.01265626336456975, "mem_mb": 63.172096, "nll": -5.901725769042969, "elapsed_s": 1237.5} {"event": "train", "step": 8760, "loss": -6.071680545806885, "grad_norm": 0.5674680471420288, "lr": 7.872259372037726e-05, "ex_per_s": 452.4565557922564, "data_frac": 0.012647986743199861, "mem_mb": 63.172096, "nll": -6.071680545806885, "elapsed_s": 1239.1} {"event": "train", "step": 8780, "loss": -5.981593608856201, "grad_norm": 0.13353751599788666, "lr": 7.816144876694563e-05, "ex_per_s": 452.88619262964096, "data_frac": 0.012639876313338934, "mem_mb": 63.172096, "nll": -5.981593608856201, "elapsed_s": 1240.8} {"event": "train", "step": 8800, "loss": -5.953164577484131, "grad_norm": 0.11745991557836533, "lr": 7.760285155445326e-05, "ex_per_s": 453.314286863673, "data_frac": 0.012631755733282199, "mem_mb": 63.172096, "nll": -5.953164577484131, "elapsed_s": 1242.4} {"event": "train", "step": 8820, "loss": -5.991349220275879, "grad_norm": 0.12780264019966125, "lr": 7.70468184714854e-05, "ex_per_s": 453.74232117966244, "data_frac": 0.012623617880776192, "mem_mb": 63.172096, "nll": -5.991349220275879, "elapsed_s": 1244.1} {"event": "train", "step": 8840, "loss": -5.961200714111328, "grad_norm": 0.3145582377910614, "lr": 7.649336583139853e-05, "ex_per_s": 454.16913849889335, "data_frac": 0.01261544527251631, "mem_mb": 63.172096, "nll": -5.961200714111328, "elapsed_s": 1245.7} {"event": "train", "step": 8860, "loss": -5.993384838104248, "grad_norm": 0.12342141568660736, "lr": 7.594250987184208e-05, "ex_per_s": 454.5946401860425, "data_frac": 0.01260755974317293, "mem_mb": 63.172096, "nll": -5.993384838104248, "elapsed_s": 1247.4} {"event": "train", "step": 8880, "loss": -5.972760200500488, "grad_norm": 0.3834097683429718, "lr": 7.53942667542819e-05, "ex_per_s": 455.01944188435795, "data_frac": 0.012599318521609695, "mem_mb": 63.172096, "nll": -5.972760200500488, "elapsed_s": 1249.0} {"event": "train", "step": 8900, "loss": -5.9486403465271, "grad_norm": 0.192343071103096, "lr": 7.484865256352612e-05, "ex_per_s": 455.44268373878344, "data_frac": 0.012591281884223836, "mem_mb": 63.172096, "nll": -5.9486403465271, "elapsed_s": 1250.7} {"event": "train", "step": 8920, "loss": -6.098045349121094, "grad_norm": 0.17420361936092377, "lr": 7.430568330725319e-05, "ex_per_s": 455.8651231543002, "data_frac": 0.012583339212613568, "mem_mb": 63.172096, "nll": -6.098045349121094, "elapsed_s": 1252.3} {"event": "train", "step": 8940, "loss": -6.102135181427002, "grad_norm": 0.2864559292793274, "lr": 7.376537491554246e-05, "ex_per_s": 456.28548261206447, "data_frac": 0.012576124640807129, "mem_mb": 63.172096, "nll": -6.102135181427002, "elapsed_s": 1254.0} {"event": "train", "step": 8960, "loss": -6.060303688049316, "grad_norm": 0.23316338658332825, "lr": 7.322774324040643e-05, "ex_per_s": 456.7035345408622, "data_frac": 0.012568289110288788, "mem_mb": 63.172096, "nll": -6.060303688049316, "elapsed_s": 1255.7} {"event": "train", "step": 8980, "loss": -5.915591239929199, "grad_norm": 0.8363661766052246, "lr": 7.26928040553261e-05, "ex_per_s": 457.11987100684854, "data_frac": 0.012560245868906826, "mem_mb": 63.172096, "nll": -5.915591239929199, "elapsed_s": 1257.3} {"event": "train", "step": 9000, "loss": -6.094305992126465, "grad_norm": 0.2917490303516388, "lr": 7.216057305478784e-05, "ex_per_s": 457.53610506150386, "data_frac": 0.0125522387347816, "mem_mb": 63.172096, "nll": -6.094305992126465, "elapsed_s": 1259.0} {"event": "val", "step": 9000, "val_nll": -6.452362914860797, "best": -6.452620423016065, "bad": 2, "mem_mb": 63.172096, "elapsed_s": 1321.7} {"event": "train", "step": 9020, "loss": -6.005491733551025, "grad_norm": 0.37884387373924255, "lr": 7.163106585382312e-05, "ex_per_s": 436.2304634314798, "data_frac": 0.011949233542397908, "mem_mb": 63.172096, "nll": -6.005491733551025, "elapsed_s": 1323.4} {"event": "train", "step": 9040, "loss": -6.104473114013672, "grad_norm": 0.15318268537521362, "lr": 7.11042979875503e-05, "ex_per_s": 436.64645257385223, "data_frac": 0.011942356612452883, "mem_mb": 63.172096, "nll": -6.104473114013672, "elapsed_s": 1325.1} {"event": "train", "step": 9060, "loss": -6.083658695220947, "grad_norm": 0.06706541031599045, "lr": 7.058028491071894e-05, "ex_per_s": 437.0618005438599, "data_frac": 0.011935279692377688, "mem_mb": 63.172096, "nll": -6.083658695220947, "elapsed_s": 1326.7} {"event": "train", "step": 9080, "loss": -6.092719078063965, "grad_norm": 0.14692825078964233, "lr": 7.005904199725619e-05, "ex_per_s": 437.47629804726546, "data_frac": 0.01192842309258931, "mem_mb": 63.172096, "nll": -6.092719078063965, "elapsed_s": 1328.4} {"event": "train", "step": 9100, "loss": -6.168652057647705, "grad_norm": 0.3589339852333069, "lr": 6.954058453981609e-05, "ex_per_s": 437.88934127517894, "data_frac": 0.011922129745858018, "mem_mb": 63.172096, "nll": -6.168652057647705, "elapsed_s": 1330.1} {"event": "train", "step": 9120, "loss": -5.973888397216797, "grad_norm": 0.1678711473941803, "lr": 6.902492774933048e-05, "ex_per_s": 438.3007753895432, "data_frac": 0.011924300118283387, "mem_mb": 63.172096, "nll": -5.973888397216797, "elapsed_s": 1331.7} {"event": "train", "step": 9140, "loss": -6.124704360961914, "grad_norm": 0.11140580475330353, "lr": 6.851208675456295e-05, "ex_per_s": 438.71038447706945, "data_frac": 0.011915087259289216, "mem_mb": 63.172096, "nll": -6.124704360961914, "elapsed_s": 1333.4} {"event": "train", "step": 9160, "loss": -5.980803966522217, "grad_norm": 0.24631336331367493, "lr": 6.800207660166508e-05, "ex_per_s": 439.1165614223526, "data_frac": 0.011933779041582484, "mem_mb": 63.172096, "nll": -5.980803966522217, "elapsed_s": 1335.1} {"event": "train", "step": 9180, "loss": -5.958152770996094, "grad_norm": 0.2053661048412323, "lr": 6.749491225373478e-05, "ex_per_s": 439.52755444184936, "data_frac": 0.011924502233716326, "mem_mb": 63.172096, "nll": -5.958152770996094, "elapsed_s": 1336.8} {"event": "train", "step": 9200, "loss": -6.145275592803955, "grad_norm": 0.08126682043075562, "lr": 6.699060859037737e-05, "ex_per_s": 439.93554876296486, "data_frac": 0.011918515124025954, "mem_mb": 63.172096, "nll": -6.145275592803955, "elapsed_s": 1338.4} {"event": "train", "step": 9220, "loss": -5.969278812408447, "grad_norm": 0.39371079206466675, "lr": 6.648918040726909e-05, "ex_per_s": 440.3435152921691, "data_frac": 0.011912367576438144, "mem_mb": 63.172096, "nll": -5.969278812408447, "elapsed_s": 1340.1} {"event": "train", "step": 9240, "loss": -6.117040157318115, "grad_norm": 0.4418714642524719, "lr": 6.599064241572299e-05, "ex_per_s": 440.7503029850945, "data_frac": 0.01190495248720557, "mem_mb": 63.172096, "nll": -6.117040157318115, "elapsed_s": 1341.8} {"event": "train", "step": 9260, "loss": -6.025075435638428, "grad_norm": 0.5254652500152588, "lr": 6.549500924225728e-05, "ex_per_s": 441.156979021344, "data_frac": 0.011897266779102222, "mem_mb": 63.172096, "nll": -6.025075435638428, "elapsed_s": 1343.4} {"event": "train", "step": 9280, "loss": -5.821038246154785, "grad_norm": 0.09556449949741364, "lr": 6.500229542816624e-05, "ex_per_s": 441.56267866847764, "data_frac": 0.011889596299414306, "mem_mb": 63.172096, "nll": -5.821038246154785, "elapsed_s": 1345.1} {"event": "train", "step": 9300, "loss": -6.005321979522705, "grad_norm": 0.07790200412273407, "lr": 6.45125154290935e-05, "ex_per_s": 441.96668843585417, "data_frac": 0.011881944192744232, "mem_mb": 63.172096, "nll": -6.005321979522705, "elapsed_s": 1346.8} {"event": "train", "step": 9320, "loss": -5.825730800628662, "grad_norm": 0.612341046333313, "lr": 6.402568361460816e-05, "ex_per_s": 442.37015799711526, "data_frac": 0.011874270618118186, "mem_mb": 63.172096, "nll": -5.825730800628662, "elapsed_s": 1348.4} {"event": "train", "step": 9340, "loss": -6.051526069641113, "grad_norm": 0.15797333419322968, "lr": 6.354181426778293e-05, "ex_per_s": 442.77286853229566, "data_frac": 0.011866681644800117, "mem_mb": 63.172096, "nll": -6.051526069641113, "elapsed_s": 1350.1} {"event": "train", "step": 9360, "loss": -6.064432144165039, "grad_norm": 0.2730765640735626, "lr": 6.306092158477523e-05, "ex_per_s": 443.1743647943026, "data_frac": 0.011859795598780374, "mem_mb": 63.172096, "nll": -6.064432144165039, "elapsed_s": 1351.7} {"event": "train", "step": 9380, "loss": -5.955790042877197, "grad_norm": 0.13029389083385468, "lr": 6.258301967441065e-05, "ex_per_s": 443.5776729450744, "data_frac": 0.011852740233802022, "mem_mb": 63.172096, "nll": -5.955790042877197, "elapsed_s": 1353.4} {"event": "train", "step": 9400, "loss": -6.073474407196045, "grad_norm": 0.29360780119895935, "lr": 6.210812255776909e-05, "ex_per_s": 443.9773799755245, "data_frac": 0.011845832760523776, "mem_mb": 63.172096, "nll": -6.073474407196045, "elapsed_s": 1355.1} {"event": "train", "step": 9420, "loss": -5.940446376800537, "grad_norm": 0.31278249621391296, "lr": 6.163624416777327e-05, "ex_per_s": 444.37847041788814, "data_frac": 0.011838840033593725, "mem_mb": 63.172096, "nll": -5.940446376800537, "elapsed_s": 1356.7} {"event": "train", "step": 9440, "loss": -5.92750358581543, "grad_norm": 0.20375260710716248, "lr": 6.116739834878009e-05, "ex_per_s": 444.7780859409765, "data_frac": 0.011832272777284844, "mem_mb": 63.172096, "nll": -5.92750358581543, "elapsed_s": 1358.4} {"event": "train", "step": 9460, "loss": -6.003002166748047, "grad_norm": 0.22051163017749786, "lr": 6.0701598856174347e-05, "ex_per_s": 445.1763898138503, "data_frac": 0.011825629500864791, "mem_mb": 63.172096, "nll": -6.003002166748047, "elapsed_s": 1360.0} {"event": "train", "step": 9480, "loss": -5.92727518081665, "grad_norm": 0.09606792777776718, "lr": 6.023885935596515e-05, "ex_per_s": 445.57471753192215, "data_frac": 0.011818753739775121, "mem_mb": 63.172096, "nll": -5.92727518081665, "elapsed_s": 1361.7} {"event": "train", "step": 9500, "loss": -5.918548107147217, "grad_norm": 0.10032342374324799, "lr": 5.977919342438512e-05, "ex_per_s": 445.97256191387737, "data_frac": 0.011811878153815403, "mem_mb": 63.172096, "nll": -5.918548107147217, "elapsed_s": 1363.4} {"event": "train", "step": 9520, "loss": -5.987831115722656, "grad_norm": 0.17094384133815765, "lr": 5.9322614547491995e-05, "ex_per_s": 446.3657532383693, "data_frac": 0.011805274014065113, "mem_mb": 63.172096, "nll": -5.987831115722656, "elapsed_s": 1365.0} {"event": "train", "step": 9540, "loss": -6.0075907707214355, "grad_norm": 0.19550101459026337, "lr": 5.886913612077285e-05, "ex_per_s": 446.7553575853242, "data_frac": 0.01179982013949831, "mem_mb": 63.172096, "nll": -6.0075907707214355, "elapsed_s": 1366.7} {"event": "train", "step": 9560, "loss": -5.966063976287842, "grad_norm": 0.17421859502792358, "lr": 5.841877144875127e-05, "ex_per_s": 447.14498788694885, "data_frac": 0.011793608128383965, "mem_mb": 63.172096, "nll": -5.966063976287842, "elapsed_s": 1368.4} {"event": "train", "step": 9580, "loss": -5.910111904144287, "grad_norm": 0.2661150395870209, "lr": 5.7971533744597e-05, "ex_per_s": 447.53778698659573, "data_frac": 0.011786325068088412, "mem_mb": 63.172096, "nll": -5.910111904144287, "elapsed_s": 1370.0} {"event": "train", "step": 9600, "loss": -6.154843807220459, "grad_norm": 0.7703678011894226, "lr": 5.752743612973809e-05, "ex_per_s": 447.9303131292899, "data_frac": 0.011778937512508695, "mem_mb": 63.172096, "nll": -6.154843807220459, "elapsed_s": 1371.7} {"event": "train", "step": 9620, "loss": -6.026759624481201, "grad_norm": 0.11496666818857193, "lr": 5.7086491633476214e-05, "ex_per_s": 448.3181727024881, "data_frac": 0.011772768720051618, "mem_mb": 63.172096, "nll": -6.026759624481201, "elapsed_s": 1373.4} {"event": "train", "step": 9640, "loss": -6.000673770904541, "grad_norm": 0.3960995376110077, "lr": 5.6648713192604105e-05, "ex_per_s": 448.6998265620723, "data_frac": 0.011766892476880673, "mem_mb": 63.172096, "nll": -6.000673770904541, "elapsed_s": 1375.0} {"event": "train", "step": 9660, "loss": -6.050432205200195, "grad_norm": 0.29903939366340637, "lr": 5.621411365102632e-05, "ex_per_s": 449.08401107604567, "data_frac": 0.011760909373049544, "mem_mb": 63.172096, "nll": -6.050432205200195, "elapsed_s": 1376.7} {"event": "train", "step": 9680, "loss": -5.925705432891846, "grad_norm": 0.2912720739841461, "lr": 5.578270575938211e-05, "ex_per_s": 449.47142726854315, "data_frac": 0.011753647059532489, "mem_mb": 63.172096, "nll": -5.925705432891846, "elapsed_s": 1378.4} {"event": "train", "step": 9700, "loss": -6.194322109222412, "grad_norm": 0.678977370262146, "lr": 5.5354502174671515e-05, "ex_per_s": 449.85778732271626, "data_frac": 0.011746363590113766, "mem_mb": 63.172096, "nll": -6.194322109222412, "elapsed_s": 1380.0} {"event": "train", "step": 9720, "loss": -6.038232326507568, "grad_norm": 0.3793227970600128, "lr": 5.492951545988398e-05, "ex_per_s": 450.2430530872303, "data_frac": 0.011738959429404878, "mem_mb": 63.172096, "nll": -6.038232326507568, "elapsed_s": 1381.7} {"event": "train", "step": 9740, "loss": -5.921197891235352, "grad_norm": 0.19880180060863495, "lr": 5.4507758083629804e-05, "ex_per_s": 450.6275680317014, "data_frac": 0.011731707743129859, "mem_mb": 63.172096, "nll": -5.921197891235352, "elapsed_s": 1383.4} {"event": "train", "step": 9760, "loss": -6.0422492027282715, "grad_norm": 0.6721118688583374, "lr": 5.408924241977427e-05, "ex_per_s": 451.01067032062224, "data_frac": 0.011724875416565589, "mem_mb": 63.172096, "nll": -6.0422492027282715, "elapsed_s": 1385.0} {"event": "train", "step": 9780, "loss": -6.0604400634765625, "grad_norm": 0.1795816421508789, "lr": 5.367398074707458e-05, "ex_per_s": 451.3903985032235, "data_frac": 0.011719196734297435, "mem_mb": 63.172096, "nll": -6.0604400634765625, "elapsed_s": 1386.7} {"event": "train", "step": 9800, "loss": -5.857109069824219, "grad_norm": 0.770349383354187, "lr": 5.326198524881979e-05, "ex_per_s": 451.76974651755535, "data_frac": 0.011713336558375292, "mem_mb": 63.172096, "nll": -5.857109069824219, "elapsed_s": 1388.4} {"event": "train", "step": 9820, "loss": -6.096351146697998, "grad_norm": 0.2864855229854584, "lr": 5.285326801247307e-05, "ex_per_s": 452.1538508267686, "data_frac": 0.011706339245442758, "mem_mb": 63.172096, "nll": -6.096351146697998, "elapsed_s": 1390.0} {"event": "train", "step": 9840, "loss": -5.954110622406006, "grad_norm": 0.4064166247844696, "lr": 5.244784102931745e-05, "ex_per_s": 452.537600825074, "data_frac": 0.01169940841450569, "mem_mb": 63.172096, "nll": -5.954110622406006, "elapsed_s": 1391.7} {"event": "train", "step": 9860, "loss": -5.883657932281494, "grad_norm": 0.49218472838401794, "lr": 5.2045716194103555e-05, "ex_per_s": 452.91686752625543, "data_frac": 0.011694492297042803, "mem_mb": 63.172096, "nll": -5.883657932281494, "elapsed_s": 1393.3} {"event": "train", "step": 9880, "loss": -5.891746520996094, "grad_norm": 0.08602099865674973, "lr": 5.164690530470111e-05, "ex_per_s": 453.2921101935983, "data_frac": 0.011689863761796608, "mem_mb": 63.172096, "nll": -5.891746520996094, "elapsed_s": 1395.0} {"event": "train", "step": 9900, "loss": -5.937680721282959, "grad_norm": 0.35325005650520325, "lr": 5.125142006175238e-05, "ex_per_s": 453.6730647366101, "data_frac": 0.01168340587720512, "mem_mb": 63.172096, "nll": -5.937680721282959, "elapsed_s": 1396.6} {"event": "train", "step": 9920, "loss": -5.8828840255737305, "grad_norm": 0.10392209142446518, "lr": 5.085927206832921e-05, "ex_per_s": 454.05362641964143, "data_frac": 0.011676382530265471, "mem_mb": 63.172096, "nll": -5.8828840255737305, "elapsed_s": 1398.3} {"event": "train", "step": 9940, "loss": -5.958414554595947, "grad_norm": 0.12999464571475983, "lr": 5.047047282959241e-05, "ex_per_s": 454.4329321590637, "data_frac": 0.011669267538065504, "mem_mb": 63.172096, "nll": -5.958414554595947, "elapsed_s": 1399.9} {"event": "train", "step": 9960, "loss": -6.137436866760254, "grad_norm": 0.5038516521453857, "lr": 5.0085033752454274e-05, "ex_per_s": 454.81168211467497, "data_frac": 0.011662222847408116, "mem_mb": 63.172096, "nll": -6.137436866760254, "elapsed_s": 1401.6} {"event": "train", "step": 9980, "loss": -5.8065972328186035, "grad_norm": 0.9859757423400879, "lr": 4.970296614524392e-05, "ex_per_s": 455.18976918376603, "data_frac": 0.011655188800336365, "mem_mb": 63.172096, "nll": -5.8065972328186035, "elapsed_s": 1403.2} {"event": "train", "step": 10000, "loss": -5.944288730621338, "grad_norm": 0.34993618726730347, "lr": 4.9324281217375474e-05, "ex_per_s": 455.5662636898868, "data_frac": 0.011648116556750389, "mem_mb": 63.172096, "nll": -5.944288730621338, "elapsed_s": 1404.9} {"event": "val", "step": 10000, "val_nll": -6.4525959497856835, "best": -6.452620423016065, "bad": 3, "mem_mb": 63.172096, "elapsed_s": 1467.7} {"event": "train", "step": 10020, "loss": -6.075506210327148, "grad_norm": 0.4470657706260681, "lr": 4.8948990079019224e-05, "ex_per_s": 436.4346662241242, "data_frac": 0.011143361517321001, "mem_mb": 63.172096, "nll": -6.075506210327148, "elapsed_s": 1469.4} {"event": "train", "step": 10040, "loss": -6.08160924911499, "grad_norm": 0.1299087554216385, "lr": 4.857710374077574e-05, "ex_per_s": 436.8089448899537, "data_frac": 0.011137299163807428, "mem_mb": 63.172096, "nll": -6.08160924911499, "elapsed_s": 1471.1} {"event": "train", "step": 10060, "loss": -5.758434772491455, "grad_norm": 0.14472365379333496, "lr": 4.820863311335266e-05, "ex_per_s": 437.1827589231529, "data_frac": 0.011131260855516828, "mem_mb": 63.172096, "nll": -5.758434772491455, "elapsed_s": 1472.7} {"event": "train", "step": 10080, "loss": -6.1788411140441895, "grad_norm": 0.13944683969020844, "lr": 4.78435890072448e-05, "ex_per_s": 437.55617556032524, "data_frac": 0.011125107937692207, "mem_mb": 63.172096, "nll": -6.1788411140441895, "elapsed_s": 1474.4} {"event": "train", "step": 10100, "loss": -5.980792999267578, "grad_norm": 0.5682160258293152, "lr": 4.748198213241678e-05, "ex_per_s": 437.9274958684039, "data_frac": 0.011119434314780068, "mem_mb": 63.172096, "nll": -5.980792999267578, "elapsed_s": 1476.1} {"event": "train", "step": 10120, "loss": -6.067307949066162, "grad_norm": 0.4765165448188782, "lr": 4.71238230979889e-05, "ex_per_s": 438.29943299220565, "data_frac": 0.011113277253977778, "mem_mb": 63.172096, "nll": -6.067307949066162, "elapsed_s": 1477.8} {"event": "train", "step": 10140, "loss": -5.972808361053467, "grad_norm": 0.20539209246635437, "lr": 4.676912241192593e-05, "ex_per_s": 438.67053311416475, "data_frac": 0.011107285608598468, "mem_mb": 63.172096, "nll": -5.972808361053467, "elapsed_s": 1479.4} {"event": "train", "step": 10160, "loss": -5.776061534881592, "grad_norm": 0.1621466428041458, "lr": 4.641789048072876e-05, "ex_per_s": 439.0410763308718, "data_frac": 0.011101269596128472, "mem_mb": 63.172096, "nll": -5.776061534881592, "elapsed_s": 1481.1} {"event": "train", "step": 10180, "loss": -6.011735916137695, "grad_norm": 0.15991269052028656, "lr": 4.6070137609129025e-05, "ex_per_s": 439.4092176981422, "data_frac": 0.011095960257475763, "mem_mb": 63.172096, "nll": -6.011735916137695, "elapsed_s": 1482.8} {"event": "train", "step": 10200, "loss": -5.957472324371338, "grad_norm": 0.29377037286758423, "lr": 4.57258739997869e-05, "ex_per_s": 439.77903143034194, "data_frac": 0.011089978456413904, "mem_mb": 63.172096, "nll": -5.957472324371338, "elapsed_s": 1484.4} {"event": "train", "step": 10220, "loss": -5.78178071975708, "grad_norm": 0.09641289710998535, "lr": 4.538510975299176e-05, "ex_per_s": 440.14788766214406, "data_frac": 0.01108421568031607, "mem_mb": 63.172096, "nll": -5.78178071975708, "elapsed_s": 1486.1} {"event": "train", "step": 10240, "loss": -6.024240970611572, "grad_norm": 0.38451626896858215, "lr": 4.504785486636568e-05, "ex_per_s": 440.51500996838183, "data_frac": 0.0110787810888656, "mem_mb": 63.172096, "nll": -6.024240970611572, "elapsed_s": 1487.8} {"event": "train", "step": 10260, "loss": -6.0319695472717285, "grad_norm": 0.15895234048366547, "lr": 4.471411923457034e-05, "ex_per_s": 440.88274749386676, "data_frac": 0.01107272034516525, "mem_mb": 63.172096, "nll": -6.0319695472717285, "elapsed_s": 1489.4} {"event": "train", "step": 10280, "loss": -6.065873622894287, "grad_norm": 0.1867392659187317, "lr": 4.438391264901654e-05, "ex_per_s": 441.25018764154055, "data_frac": 0.011066710208272523, "mem_mb": 63.172096, "nll": -6.065873622894287, "elapsed_s": 1491.1} {"event": "train", "step": 10300, "loss": -5.9920654296875, "grad_norm": 0.11666863411664963, "lr": 4.405724479757712e-05, "ex_per_s": 441.61638730669546, "data_frac": 0.011060736073377209, "mem_mb": 63.172096, "nll": -5.9920654296875, "elapsed_s": 1492.7} {"event": "train", "step": 10320, "loss": -5.9075422286987305, "grad_norm": 0.38074374198913574, "lr": 4.373412526430252e-05, "ex_per_s": 441.9818263372846, "data_frac": 0.011054803220805647, "mem_mb": 63.172096, "nll": -5.9075422286987305, "elapsed_s": 1494.4} {"event": "train", "step": 10340, "loss": -5.9490156173706055, "grad_norm": 0.07999832183122635, "lr": 4.341456352913975e-05, "ex_per_s": 442.3442776229755, "data_frac": 0.011049189053118818, "mem_mb": 63.172096, "nll": -5.9490156173706055, "elapsed_s": 1496.1} {"event": "train", "step": 10360, "loss": -5.7982306480407715, "grad_norm": 0.23543377220630646, "lr": 4.3098568967654146e-05, "ex_per_s": 442.7065322859891, "data_frac": 0.011043170288236922, "mem_mb": 63.172096, "nll": -5.7982306480407715, "elapsed_s": 1497.7} {"event": "train", "step": 10380, "loss": -5.9836344718933105, "grad_norm": 0.5839988589286804, "lr": 4.27861508507545e-05, "ex_per_s": 443.0684769368809, "data_frac": 0.011037209532694525, "mem_mb": 63.172096, "nll": -5.9836344718933105, "elapsed_s": 1499.4} {"event": "train", "step": 10400, "loss": -5.974215030670166, "grad_norm": 0.7241143584251404, "lr": 4.247731834442082e-05, "ex_per_s": 443.42845322443674, "data_frac": 0.011031351034122727, "mem_mb": 63.172096, "nll": -5.974215030670166, "elapsed_s": 1501.1} {"event": "train", "step": 10420, "loss": -6.03476095199585, "grad_norm": 0.384708434343338, "lr": 4.2172080509435485e-05, "ex_per_s": 443.7876492609329, "data_frac": 0.011025457012801946, "mem_mb": 63.172096, "nll": -6.03476095199585, "elapsed_s": 1502.7} {"event": "train", "step": 10440, "loss": -6.128844738006592, "grad_norm": 0.5997136235237122, "lr": 4.1870446301117574e-05, "ex_per_s": 444.1466807190784, "data_frac": 0.011019787653920681, "mem_mb": 63.172096, "nll": -6.128844738006592, "elapsed_s": 1504.4} {"event": "train", "step": 10460, "loss": -5.974477767944336, "grad_norm": 0.15828408300876617, "lr": 4.157242456905989e-05, "ex_per_s": 444.5048069318496, "data_frac": 0.011014369059064418, "mem_mb": 63.172096, "nll": -5.974477767944336, "elapsed_s": 1506.1} {"event": "train", "step": 10480, "loss": -5.97821044921875, "grad_norm": 0.3214838206768036, "lr": 4.1278024056869486e-05, "ex_per_s": 444.86305912140085, "data_frac": 0.011008400551370279, "mem_mb": 63.172096, "nll": -5.97821044921875, "elapsed_s": 1507.7} {"event": "train", "step": 10500, "loss": -5.996211051940918, "grad_norm": 0.4190135896205902, "lr": 4.0987253401911067e-05, "ex_per_s": 445.21984780025815, "data_frac": 0.011002825676375186, "mem_mb": 63.172096, "nll": -5.996211051940918, "elapsed_s": 1509.4} {"event": "train", "step": 10520, "loss": -6.006505966186523, "grad_norm": 0.5200763940811157, "lr": 4.070012113505359e-05, "ex_per_s": 445.57749367083943, "data_frac": 0.01099692973115075, "mem_mb": 63.172096, "nll": -6.006505966186523, "elapsed_s": 1511.1} {"event": "train", "step": 10540, "loss": -5.7781243324279785, "grad_norm": 0.37430495023727417, "lr": 4.041663568041998e-05, "ex_per_s": 445.933344328393, "data_frac": 0.01099162802476442, "mem_mb": 63.172096, "nll": -5.7781243324279785, "elapsed_s": 1512.7} {"event": "train", "step": 10560, "loss": -5.870539665222168, "grad_norm": 0.15975074470043182, "lr": 4.013680535514001e-05, "ex_per_s": 446.28761013209504, "data_frac": 0.010986447002162017, "mem_mb": 63.172096, "nll": -5.870539665222168, "elapsed_s": 1514.4} {"event": "train", "step": 10580, "loss": -5.871490478515625, "grad_norm": 0.2041986882686615, "lr": 3.986063836910623e-05, "ex_per_s": 446.6418825549872, "data_frac": 0.010981826053336132, "mem_mb": 63.172096, "nll": -5.871490478515625, "elapsed_s": 1516.1} {"event": "train", "step": 10600, "loss": -5.896048069000244, "grad_norm": 0.25817468762397766, "lr": 3.958814282473316e-05, "ex_per_s": 446.996935445086, "data_frac": 0.010976230027592055, "mem_mb": 63.172096, "nll": -5.896048069000244, "elapsed_s": 1517.7} {"event": "train", "step": 10620, "loss": -6.044766902923584, "grad_norm": 0.19144213199615479, "lr": 3.9319326716719496e-05, "ex_per_s": 447.3514180492293, "data_frac": 0.010970439727989029, "mem_mb": 63.172096, "nll": -6.044766902923584, "elapsed_s": 1519.4} {"event": "train", "step": 10640, "loss": -5.815213680267334, "grad_norm": 0.40217503905296326, "lr": 3.905419793181364e-05, "ex_per_s": 447.703670552977, "data_frac": 0.010965162305091028, "mem_mb": 63.172096, "nll": -5.815213680267334, "elapsed_s": 1521.0} {"event": "train", "step": 10660, "loss": -5.851538181304932, "grad_norm": 0.2379125952720642, "lr": 3.87927642485822e-05, "ex_per_s": 448.0568970765587, "data_frac": 0.010960077797537954, "mem_mb": 63.172096, "nll": -5.851538181304932, "elapsed_s": 1522.7} {"event": "train", "step": 10680, "loss": -5.950348377227783, "grad_norm": 0.13598987460136414, "lr": 3.853503333718199e-05, "ex_per_s": 448.4084886726861, "data_frac": 0.01095501242399772, "mem_mb": 63.172096, "nll": -5.950348377227783, "elapsed_s": 1524.4} {"event": "train", "step": 10700, "loss": -5.80232572555542, "grad_norm": 0.7373836040496826, "lr": 3.8281012759134686e-05, "ex_per_s": 448.7612080462344, "data_frac": 0.01095000551546777, "mem_mb": 63.172096, "nll": -5.80232572555542, "elapsed_s": 1526.0} {"event": "train", "step": 10720, "loss": -5.880722999572754, "grad_norm": 0.11987880617380142, "lr": 3.803070996710526e-05, "ex_per_s": 449.1128554735322, "data_frac": 0.010945104013023404, "mem_mb": 63.172096, "nll": -5.880722999572754, "elapsed_s": 1527.7} {"event": "train", "step": 10740, "loss": -6.101659297943115, "grad_norm": 0.5411739945411682, "lr": 3.778413230468319e-05, "ex_per_s": 449.46455626884716, "data_frac": 0.010940147358482604, "mem_mb": 63.172096, "nll": -6.101659297943115, "elapsed_s": 1529.3} {"event": "train", "step": 10760, "loss": -5.962273120880127, "grad_norm": 0.28983187675476074, "lr": 3.754128700616701e-05, "ex_per_s": 449.8151406991539, "data_frac": 0.010935109872277082, "mem_mb": 63.172096, "nll": -5.962273120880127, "elapsed_s": 1531.0} {"event": "train", "step": 10780, "loss": -6.020055294036865, "grad_norm": 0.4682409167289734, "lr": 3.7302181196352115e-05, "ex_per_s": 450.16615837178875, "data_frac": 0.010929710008363848, "mem_mb": 63.172096, "nll": -6.020055294036865, "elapsed_s": 1532.6} {"event": "train", "step": 10800, "loss": -5.934305667877197, "grad_norm": 0.46633264422416687, "lr": 3.7066821890321684e-05, "ex_per_s": 450.5151793763366, "data_frac": 0.010924496303485036, "mem_mb": 63.172096, "nll": -5.934305667877197, "elapsed_s": 1534.3} {"event": "train", "step": 10820, "loss": -6.01970911026001, "grad_norm": 0.24249720573425293, "lr": 3.683521599324089e-05, "ex_per_s": 450.8642110310604, "data_frac": 0.010919722875196545, "mem_mb": 63.172096, "nll": -6.01970911026001, "elapsed_s": 1535.9} {"event": "train", "step": 10840, "loss": -5.963700771331787, "grad_norm": 0.1928597390651703, "lr": 3.660737030015427e-05, "ex_per_s": 451.21261752986226, "data_frac": 0.010914339974018295, "mem_mb": 63.172096, "nll": -5.963700771331787, "elapsed_s": 1537.6} {"event": "train", "step": 10860, "loss": -5.843427658081055, "grad_norm": 0.10267985612154007, "lr": 3.638329149578648e-05, "ex_per_s": 451.5604546810307, "data_frac": 0.01090925901694235, "mem_mb": 63.172096, "nll": -5.843427658081055, "elapsed_s": 1539.2} {"event": "train", "step": 10880, "loss": -5.999927997589111, "grad_norm": 0.19769184291362762, "lr": 3.6162986154346006e-05, "ex_per_s": 451.9073350346584, "data_frac": 0.010903807216224568, "mem_mb": 63.172096, "nll": -5.999927997589111, "elapsed_s": 1540.9} {"event": "train", "step": 10900, "loss": -6.124413013458252, "grad_norm": 0.18128782510757446, "lr": 3.594646073933244e-05, "ex_per_s": 452.2537070888645, "data_frac": 0.010898239494238623, "mem_mb": 63.172096, "nll": -6.124413013458252, "elapsed_s": 1542.5} {"event": "train", "step": 10920, "loss": -5.900936603546143, "grad_norm": 0.31502705812454224, "lr": 3.5733721603346695e-05, "ex_per_s": 452.5998036917151, "data_frac": 0.010892671082043285, "mem_mb": 63.172096, "nll": -5.900936603546143, "elapsed_s": 1544.2} {"event": "train", "step": 10940, "loss": -5.8987603187561035, "grad_norm": 0.2269534468650818, "lr": 3.552477498790482e-05, "ex_per_s": 452.9445078285411, "data_frac": 0.010887423644106331, "mem_mb": 63.172096, "nll": -5.8987603187561035, "elapsed_s": 1545.8} {"event": "train", "step": 10960, "loss": -6.126065254211426, "grad_norm": 0.1712750494480133, "lr": 3.531962702325468e-05, "ex_per_s": 453.28908047607973, "data_frac": 0.010882055841072318, "mem_mb": 63.172096, "nll": -6.126065254211426, "elapsed_s": 1547.5} {"event": "train", "step": 10980, "loss": -5.895640850067139, "grad_norm": 0.10163795202970505, "lr": 3.511828372819624e-05, "ex_per_s": 453.63261719285003, "data_frac": 0.010876545383939656, "mem_mb": 63.172096, "nll": -5.895640850067139, "elapsed_s": 1549.1} {"event": "train", "step": 11000, "loss": -5.921398639678955, "grad_norm": 0.06576220691204071, "lr": 3.4920751009904904e-05, "ex_per_s": 453.9751693164411, "data_frac": 0.010871094234666865, "mem_mb": 63.172096, "nll": -5.921398639678955, "elapsed_s": 1550.8} {"event": "val", "step": 11000, "val_nll": -6.453004997058924, "best": -6.453004997058924, "bad": 0, "mem_mb": 63.172096, "elapsed_s": 1613.7} {"event": "train", "step": 11020, "loss": -5.806365013122559, "grad_norm": 0.35421398282051086, "lr": 3.472703466375828e-05, "ex_per_s": 436.6149221054281, "data_frac": 0.010442642449210981, "mem_mb": 63.172096, "nll": -5.806365013122559, "elapsed_s": 1615.4} {"event": "train", "step": 11040, "loss": -5.984075546264648, "grad_norm": 0.24564407765865326, "lr": 3.4537140373166085e-05, "ex_per_s": 436.95656988606055, "data_frac": 0.010438115508608873, "mem_mb": 63.172096, "nll": -5.984075546264648, "elapsed_s": 1617.0} {"event": "train", "step": 11060, "loss": -5.982636451721191, "grad_norm": 0.47516685724258423, "lr": 3.4351073709403385e-05, "ex_per_s": 437.29757717156565, "data_frac": 0.010433242204562368, "mem_mb": 63.172096, "nll": -5.982636451721191, "elapsed_s": 1618.7} {"event": "train", "step": 11080, "loss": -6.0313215255737305, "grad_norm": 0.37395018339157104, "lr": 3.4168840131447255e-05, "ex_per_s": 437.637864303503, "data_frac": 0.010428425263940649, "mem_mb": 63.172096, "nll": -6.0313215255737305, "elapsed_s": 1620.4} {"event": "train", "step": 11100, "loss": -6.140146255493164, "grad_norm": 0.48653942346572876, "lr": 3.399044498581645e-05, "ex_per_s": 437.9774992520247, "data_frac": 0.010423528937521829, "mem_mb": 63.172096, "nll": -6.140146255493164, "elapsed_s": 1622.0} {"event": "train", "step": 11120, "loss": -5.907780647277832, "grad_norm": 0.480852872133255, "lr": 3.3815893506414745e-05, "ex_per_s": 438.31683297723316, "data_frac": 0.01041869288777952, "mem_mb": 63.172096, "nll": -5.907780647277832, "elapsed_s": 1623.7} {"event": "train", "step": 11140, "loss": -5.883107662200928, "grad_norm": 0.07971934229135513, "lr": 3.364519081437718e-05, "ex_per_s": 438.65593836103716, "data_frac": 0.01041382442827674, "mem_mb": 63.172096, "nll": -5.883107662200928, "elapsed_s": 1625.4} {"event": "train", "step": 11160, "loss": -6.079113960266113, "grad_norm": 0.25860127806663513, "lr": 3.347834191791993e-05, "ex_per_s": 438.9941098757026, "data_frac": 0.010408985856762427, "mem_mb": 63.172096, "nll": -6.079113960266113, "elapsed_s": 1627.0} {"event": "train", "step": 11180, "loss": -6.0708231925964355, "grad_norm": 0.37125658988952637, "lr": 3.331535171219332e-05, "ex_per_s": 439.3309001472849, "data_frac": 0.010404470491429264, "mem_mb": 63.172096, "nll": -6.0708231925964355, "elapsed_s": 1628.7} {"event": "train", "step": 11200, "loss": -6.179490566253662, "grad_norm": 0.10487707704305649, "lr": 3.3156224979138305e-05, "ex_per_s": 439.6679132678437, "data_frac": 0.010399703814157066, "mem_mb": 63.172096, "nll": -6.179490566253662, "elapsed_s": 1630.4} {"event": "train", "step": 11220, "loss": -5.980915546417236, "grad_norm": 0.16054755449295044, "lr": 3.300096638734598e-05, "ex_per_s": 440.00419434286874, "data_frac": 0.010394942421595834, "mem_mb": 63.172096, "nll": -5.980915546417236, "elapsed_s": 1632.0} {"event": "train", "step": 11240, "loss": -5.974283218383789, "grad_norm": 0.18192674219608307, "lr": 3.2849580491920845e-05, "ex_per_s": 440.3405365989807, "data_frac": 0.01039019869007774, "mem_mb": 63.172096, "nll": -5.974283218383789, "elapsed_s": 1633.7} {"event": "train", "step": 11260, "loss": -6.033377647399902, "grad_norm": 0.3171460032463074, "lr": 3.2702071734346924e-05, "ex_per_s": 440.6757072202051, "data_frac": 0.010385453074430409, "mem_mb": 63.172096, "nll": -6.033377647399902, "elapsed_s": 1635.4} {"event": "train", "step": 11280, "loss": -6.107831001281738, "grad_norm": 0.33806470036506653, "lr": 3.255844444235771e-05, "ex_per_s": 441.0103556652723, "data_frac": 0.01038129292479136, "mem_mb": 63.172096, "nll": -6.107831001281738, "elapsed_s": 1637.0} {"event": "train", "step": 11300, "loss": -5.936168670654297, "grad_norm": 0.08373088389635086, "lr": 3.2418702829808946e-05, "ex_per_s": 441.34507723116843, "data_frac": 0.010376601190034862, "mem_mb": 63.172096, "nll": -5.936168670654297, "elapsed_s": 1638.7} {"event": "train", "step": 11320, "loss": -5.8606977462768555, "grad_norm": 0.34471845626831055, "lr": 3.228285099655516e-05, "ex_per_s": 441.6787545457743, "data_frac": 0.010371827904973083, "mem_mb": 63.172096, "nll": -5.8606977462768555, "elapsed_s": 1640.3} {"event": "train", "step": 11340, "loss": -5.930577754974365, "grad_norm": 0.6696545481681824, "lr": 3.215089292832933e-05, "ex_per_s": 442.01128218315137, "data_frac": 0.01036742964798837, "mem_mb": 63.172096, "nll": -5.930577754974365, "elapsed_s": 1642.0} {"event": "train", "step": 11360, "loss": -5.961962699890137, "grad_norm": 0.10562596470117569, "lr": 3.202283249662597e-05, "ex_per_s": 442.3429065159472, "data_frac": 0.01036304125905114, "mem_mb": 63.172096, "nll": -5.961962699890137, "elapsed_s": 1643.7} {"event": "train", "step": 11380, "loss": -5.910901069641113, "grad_norm": 0.09511138498783112, "lr": 3.189867345858747e-05, "ex_per_s": 442.67510466674725, "data_frac": 0.010358398565488019, "mem_mb": 63.172096, "nll": -5.910901069641113, "elapsed_s": 1645.3} {"event": "train", "step": 11400, "loss": -6.077014923095703, "grad_norm": 0.2515811324119568, "lr": 3.177841945689397e-05, "ex_per_s": 443.00677789494864, "data_frac": 0.010353880778616287, "mem_mb": 63.172096, "nll": -6.077014923095703, "elapsed_s": 1647.0} {"event": "train", "step": 11420, "loss": -5.8339715003967285, "grad_norm": 0.24958164989948273, "lr": 3.1662074019656414e-05, "ex_per_s": 443.3381724516754, "data_frac": 0.01034922813243361, "mem_mb": 63.172096, "nll": -5.8339715003967285, "elapsed_s": 1648.6} {"event": "train", "step": 11440, "loss": -5.88530969619751, "grad_norm": 0.7630122900009155, "lr": 3.154964056031306e-05, "ex_per_s": 443.6692562013778, "data_frac": 0.010344566647478395, "mem_mb": 63.172096, "nll": -5.88530969619751, "elapsed_s": 1650.3} {"event": "train", "step": 11460, "loss": -6.141327857971191, "grad_norm": 0.6373913288116455, "lr": 3.1441122377529333e-05, "ex_per_s": 443.99977799999016, "data_frac": 0.010339966945481694, "mem_mb": 63.172096, "nll": -6.141327857971191, "elapsed_s": 1651.9} {"event": "train", "step": 11480, "loss": -5.96607780456543, "grad_norm": 0.20248475670814514, "lr": 3.133652265510106e-05, "ex_per_s": 444.330260363021, "data_frac": 0.010335172184097897, "mem_mb": 63.172096, "nll": -5.96607780456543, "elapsed_s": 1653.6} {"event": "train", "step": 11500, "loss": -5.9416608810424805, "grad_norm": 0.31231915950775146, "lr": 3.123584446186108e-05, "ex_per_s": 444.65937736068463, "data_frac": 0.01033062195451142, "mem_mb": 63.172096, "nll": -5.9416608810424805, "elapsed_s": 1655.2} {"event": "train", "step": 11520, "loss": -5.912590980529785, "grad_norm": 0.4162837564945221, "lr": 3.113909075158913e-05, "ex_per_s": 444.98727303109143, "data_frac": 0.01032682031471068, "mem_mb": 63.172096, "nll": -5.912590980529785, "elapsed_s": 1656.9} {"event": "train", "step": 11540, "loss": -6.05053186416626, "grad_norm": 0.3417062759399414, "lr": 3.104626436292525e-05, "ex_per_s": 445.31623721202544, "data_frac": 0.010322797836278635, "mem_mb": 63.172096, "nll": -6.05053186416626, "elapsed_s": 1658.5} {"event": "train", "step": 11560, "loss": -6.117702007293701, "grad_norm": 0.08748973906040192, "lr": 3.095736801928647e-05, "ex_per_s": 445.6439606415809, "data_frac": 0.010318677428559502, "mem_mb": 63.172096, "nll": -6.117702007293701, "elapsed_s": 1660.2} {"event": "train", "step": 11580, "loss": -5.960811614990234, "grad_norm": 0.0980251282453537, "lr": 3.087240432878696e-05, "ex_per_s": 445.9714685608287, "data_frac": 0.010314544030569167, "mem_mb": 63.172096, "nll": -5.960811614990234, "elapsed_s": 1661.9} {"event": "train", "step": 11600, "loss": -6.074941158294678, "grad_norm": 0.1840716302394867, "lr": 3.0791375784161455e-05, "ex_per_s": 446.2986528689606, "data_frac": 0.010310153628685925, "mem_mb": 63.172096, "nll": -6.074941158294678, "elapsed_s": 1663.5} {"event": "train", "step": 11620, "loss": -6.028655052185059, "grad_norm": 0.13898836076259613, "lr": 3.0714284762692116e-05, "ex_per_s": 446.62548052030286, "data_frac": 0.010305511616966698, "mem_mb": 63.172096, "nll": -6.028655052185059, "elapsed_s": 1665.2} {"event": "train", "step": 11640, "loss": -5.963109016418457, "grad_norm": 0.0872587040066719, "lr": 3.0641133526138836e-05, "ex_per_s": 446.95137058721735, "data_frac": 0.010301074916595467, "mem_mb": 63.172096, "nll": -5.963109016418457, "elapsed_s": 1666.8} {"event": "train", "step": 11660, "loss": -5.92079496383667, "grad_norm": 0.267625629901886, "lr": 3.0571924220672835e-05, "ex_per_s": 447.2763190790451, "data_frac": 0.010296648010873885, "mem_mb": 63.172096, "nll": -5.92079496383667, "elapsed_s": 1668.5} {"event": "train", "step": 11680, "loss": -5.964593410491943, "grad_norm": 0.44957485795021057, "lr": 3.0506658876813734e-05, "ex_per_s": 447.6008224689444, "data_frac": 0.010292210838944642, "mem_mb": 63.172096, "nll": -5.964593410491943, "elapsed_s": 1670.1} {"event": "train", "step": 11700, "loss": -6.048835754394531, "grad_norm": 0.07592904567718506, "lr": 3.0445339409369934e-05, "ex_per_s": 447.92392215411377, "data_frac": 0.010287831075996547, "mem_mb": 63.172096, "nll": -6.048835754394531, "elapsed_s": 1671.8} {"event": "train", "step": 11720, "loss": -5.7662434577941895, "grad_norm": 0.27608850598335266, "lr": 3.0387967617382475e-05, "ex_per_s": 448.2473181256006, "data_frac": 0.010283914115251211, "mem_mb": 63.172096, "nll": -5.7662434577941895, "elapsed_s": 1673.4} {"event": "train", "step": 11740, "loss": -6.043995380401611, "grad_norm": 0.46590015292167664, "lr": 3.0334545184072252e-05, "ex_per_s": 448.5683843475477, "data_frac": 0.010279753421388309, "mem_mb": 63.172096, "nll": -6.043995380401611, "elapsed_s": 1675.1} {"event": "train", "step": 11760, "loss": -5.900750160217285, "grad_norm": 0.273719847202301, "lr": 3.0285073676790606e-05, "ex_per_s": 448.88810905884696, "data_frac": 0.01027592232877482, "mem_mb": 63.172096, "nll": -5.900750160217285, "elapsed_s": 1676.7} {"event": "train", "step": 11780, "loss": -6.084139347076416, "grad_norm": 0.24382658302783966, "lr": 3.0239554546973377e-05, "ex_per_s": 449.20824128576777, "data_frac": 0.010271368394911039, "mem_mb": 63.172096, "nll": -6.084139347076416, "elapsed_s": 1678.4} {"event": "train", "step": 11800, "loss": -5.968868255615234, "grad_norm": 0.3446606397628784, "lr": 3.0197989130098285e-05, "ex_per_s": 449.52768909467926, "data_frac": 0.01026687173038303, "mem_mb": 63.172096, "nll": -5.968868255615234, "elapsed_s": 1680.0} {"event": "train", "step": 11820, "loss": -5.9716291427612305, "grad_norm": 0.677733838558197, "lr": 3.0160378645645744e-05, "ex_per_s": 449.84633200129275, "data_frac": 0.010262415197901742, "mem_mb": 63.172096, "nll": -5.9716291427612305, "elapsed_s": 1681.7} {"event": "train", "step": 11840, "loss": -6.032918930053711, "grad_norm": 0.1543138176202774, "lr": 3.0126724197063143e-05, "ex_per_s": 450.1648939013703, "data_frac": 0.01025779262086478, "mem_mb": 63.172096, "nll": -6.032918930053711, "elapsed_s": 1683.3} {"event": "train", "step": 11860, "loss": -5.989689350128174, "grad_norm": 0.35392823815345764, "lr": 3.0097026771732407e-05, "ex_per_s": 450.48152223126425, "data_frac": 0.01025403257122679, "mem_mb": 63.172096, "nll": -5.989689350128174, "elapsed_s": 1685.0} {"event": "train", "step": 11880, "loss": -6.0297627449035645, "grad_norm": 0.7967751622200012, "lr": 3.0071287240941065e-05, "ex_per_s": 450.7987096552931, "data_frac": 0.010249560970171991, "mem_mb": 63.172096, "nll": -6.0297627449035645, "elapsed_s": 1686.7} {"event": "train", "step": 11900, "loss": -5.890211582183838, "grad_norm": 0.19056308269500732, "lr": 3.0049506359856666e-05, "ex_per_s": 451.1153567329298, "data_frac": 0.010245012696982156, "mem_mb": 63.172096, "nll": -5.890211582183838, "elapsed_s": 1688.3} {"event": "train", "step": 11920, "loss": -6.011999130249023, "grad_norm": 0.09476378560066223, "lr": 3.0031684767504643e-05, "ex_per_s": 451.43103671870585, "data_frac": 0.010240811238337613, "mem_mb": 63.172096, "nll": -6.011999130249023, "elapsed_s": 1690.0} {"event": "train", "step": 11940, "loss": -6.0682172775268555, "grad_norm": 0.7864689826965332, "lr": 3.0017822986749544e-05, "ex_per_s": 451.74643599963935, "data_frac": 0.01023640998923699, "mem_mb": 63.172096, "nll": -6.0682172775268555, "elapsed_s": 1691.6} {"event": "train", "step": 11960, "loss": -6.165287017822266, "grad_norm": 0.08346591889858246, "lr": 3.000792142427973e-05, "ex_per_s": 452.06131377344525, "data_frac": 0.01023190809723447, "mem_mb": 63.172096, "nll": -6.165287017822266, "elapsed_s": 1693.3} {"event": "train", "step": 11980, "loss": -6.093001842498779, "grad_norm": 0.09915713220834732, "lr": 3.000198037059536e-05, "ex_per_s": 452.3755396555692, "data_frac": 0.0102275536824706, "mem_mb": 63.172096, "nll": -6.093001842498779, "elapsed_s": 1694.9} {"event": "train", "step": 12000, "loss": -5.93685245513916, "grad_norm": 0.15613403916358948, "lr": 2.9999999999999997e-05, "ex_per_s": 452.68829811299906, "data_frac": 0.010223515748459169, "mem_mb": 63.172096, "nll": -5.93685245513916, "elapsed_s": 1696.6} {"event": "val", "step": 12000, "val_nll": -6.453461130264837, "best": -6.453461130264837, "bad": 0, "mem_mb": 63.172096, "elapsed_s": 1759.4} {"event": "done", "steps": 12000, "best_val": -6.453461130264837, "wall_s": 1759.3876690864563, "ex_per_s": 436.5155067835482, "peak_mem_mb": 63.172096, "n_params": 3455619, "device": "mps", "data_time_frac": 0.009858269313898334, "elapsed_s": 1759.4}