SPB Git

spb/wp41_uqo

Public
152.3 KB · 614 lines
1{"event": "train", "step": 20, "loss": -5.979733943939209, "grad_norm": 1.1914286613464355, "lr": 1.5749999999999997e-05, "ex_per_s": 57.089669165147605, "data_frac": 0.9830834538698828, "mem_mb": 14.961152, "nll": -5.979733943939209, "elapsed_s": 22.4}2{"event": "train", "step": 40, "loss": -5.768097877502441, "grad_norm": 1.0859007835388184, "lr": 3.0749999999999995e-05, "ex_per_s": 90.71866160064457, "data_frac": 0.9802300838789231, "mem_mb": 14.961152, "nll": -5.768097877502441, "elapsed_s": 28.2}3{"event": "train", "step": 60, "loss": -5.878754615783691, "grad_norm": 1.0284702777862549, "lr": 4.5749999999999994e-05, "ex_per_s": 121.58492739291056, "data_frac": 0.9794813431533057, "mem_mb": 14.961152, "nll": -5.878754615783691, "elapsed_s": 31.6}4{"event": "train", "step": 80, "loss": -6.038109302520752, "grad_norm": 0.4551159143447876, "lr": 6.075e-05, "ex_per_s": 151.48512014349907, "data_frac": 0.9779891255640201, "mem_mb": 14.961152, "nll": -6.038109302520752, "elapsed_s": 33.8}5{"event": "train", "step": 100, "loss": -5.915530681610107, "grad_norm": 0.24097995460033417, "lr": 7.575e-05, "ex_per_s": 182.57612112573156, "data_frac": 0.9764409018896146, "mem_mb": 14.961152, "nll": -5.915530681610107, "elapsed_s": 35.1}6{"event": "train", "step": 120, "loss": -5.886272430419922, "grad_norm": 0.21974466741085052, "lr": 9.074999999999998e-05, "ex_per_s": 216.1250192165707, "data_frac": 0.9744213647593928, "mem_mb": 14.961152, "nll": -5.886272430419922, "elapsed_s": 35.5}7{"event": "train", "step": 140, "loss": -5.810540676116943, "grad_norm": 0.17165157198905945, "lr": 0.00010574999999999998, "ex_per_s": 249.33765389860838, "data_frac": 0.9722797712487208, "mem_mb": 14.961152, "nll": -5.810540676116943, "elapsed_s": 35.9}8{"event": "train", "step": 160, "loss": -5.7302021980285645, "grad_norm": 0.11076026409864426, "lr": 0.00012075, "ex_per_s": 282.403669221555, "data_frac": 0.9708794822743204, "mem_mb": 14.961152, "nll": -5.7302021980285645, "elapsed_s": 36.3}9{"event": "train", "step": 180, "loss": -5.975557804107666, "grad_norm": 0.5121926665306091, "lr": 0.00013575, "ex_per_s": 317.1078251929208, "data_frac": 0.9692229528386728, "mem_mb": 14.961152, "nll": -5.975557804107666, "elapsed_s": 36.3}10{"event": "train", "step": 200, "loss": -5.95612096786499, "grad_norm": 0.2518540024757385, "lr": 0.00015074999999999998, "ex_per_s": 350.50353255736957, "data_frac": 0.9675979695599295, "mem_mb": 14.961152, "nll": -5.95612096786499, "elapsed_s": 36.5}11{"event": "train", "step": 220, "loss": -6.024130821228027, "grad_norm": 0.5167198777198792, "lr": 0.00016575, "ex_per_s": 383.88843047455464, "data_frac": 0.9662761363097936, "mem_mb": 14.961152, "nll": -6.024130821228027, "elapsed_s": 36.7}12{"event": "train", "step": 240, "loss": -5.9017014503479, "grad_norm": 0.43794146180152893, "lr": 0.00018075, "ex_per_s": 417.3691534177702, "data_frac": 0.9648465149449192, "mem_mb": 14.961152, "nll": -5.9017014503479, "elapsed_s": 36.8}13{"event": "train", "step": 260, "loss": -5.880438804626465, "grad_norm": 0.10062180459499359, "lr": 0.00019574999999999996, "ex_per_s": 451.3326912534316, "data_frac": 0.963272751552294, "mem_mb": 14.961152, "nll": -5.880438804626465, "elapsed_s": 36.9}14{"event": "train", "step": 280, "loss": -5.947292804718018, "grad_norm": 0.18727588653564453, "lr": 0.00021074999999999997, "ex_per_s": 484.69946772644687, "data_frac": 0.9618459278705209, "mem_mb": 14.961152, "nll": -5.947292804718018, "elapsed_s": 37.0}15{"event": "train", "step": 300, "loss": -5.964990139007568, "grad_norm": 0.09265275299549103, "lr": 0.00022574999999999996, "ex_per_s": 518.1941048007341, "data_frac": 0.9603603627023756, "mem_mb": 14.961152, "nll": -5.964990139007568, "elapsed_s": 37.1}16{"event": "train", "step": 320, "loss": -5.968240261077881, "grad_norm": 0.5036424398422241, "lr": 0.00024074999999999997, "ex_per_s": 551.7961600900168, "data_frac": 0.9589355502509205, "mem_mb": 14.961152, "nll": -5.968240261077881, "elapsed_s": 37.1}17{"event": "train", "step": 340, "loss": -5.835287570953369, "grad_norm": 0.20869700610637665, "lr": 0.00025575, "ex_per_s": 585.3034128899063, "data_frac": 0.9575369838258757, "mem_mb": 14.961152, "nll": -5.835287570953369, "elapsed_s": 37.2}18{"event": "train", "step": 360, "loss": -5.98869514465332, "grad_norm": 0.5500252842903137, "lr": 0.00027074999999999994, "ex_per_s": 618.6923536392586, "data_frac": 0.9561956396081132, "mem_mb": 14.961152, "nll": -5.98869514465332, "elapsed_s": 37.2}19{"event": "train", "step": 380, "loss": -6.00840950012207, "grad_norm": 0.24983909726142883, "lr": 0.00028575, "ex_per_s": 651.9150673102939, "data_frac": 0.9547902689257806, "mem_mb": 14.961152, "nll": -6.00840950012207, "elapsed_s": 37.3}20{"event": "train", "step": 400, "loss": -6.001003265380859, "grad_norm": 0.44369855523109436, "lr": 0.0003, "ex_per_s": 684.9116554512526, "data_frac": 0.953218484176108, "mem_mb": 14.961152, "nll": -6.001003265380859, "elapsed_s": 37.4}21{"event": "train", "step": 420, "loss": -5.9664835929870605, "grad_norm": 0.2937568426132202, "lr": 0.0002999980196294046, "ex_per_s": 717.8680499910641, "data_frac": 0.9517671963064962, "mem_mb": 14.961152, "nll": -5.9664835929870605, "elapsed_s": 37.4}22{"event": "train", "step": 440, "loss": -5.943264961242676, "grad_norm": 0.20375287532806396, "lr": 0.00029999207857572027, "ex_per_s": 750.6329969039843, "data_frac": 0.9502788102081043, "mem_mb": 14.961152, "nll": -5.943264961242676, "elapsed_s": 37.5}23{"event": "train", "step": 460, "loss": -6.102592468261719, "grad_norm": 0.18124203383922577, "lr": 0.00029998217701325043, "ex_per_s": 783.2449989792722, "data_frac": 0.9487688901245765, "mem_mb": 14.961152, "nll": -6.102592468261719, "elapsed_s": 37.6}24{"event": "train", "step": 480, "loss": -5.877071857452393, "grad_norm": 0.14028720557689667, "lr": 0.00029996831523249534, "ex_per_s": 815.8012172657269, "data_frac": 0.9473352902370265, "mem_mb": 14.961152, "nll": -5.877071857452393, "elapsed_s": 37.7}25{"event": "train", "step": 500, "loss": -5.926998138427734, "grad_norm": 0.1615256667137146, "lr": 0.0002999504936401433, "ex_per_s": 848.1162395145574, "data_frac": 0.9457904015401071, "mem_mb": 14.961152, "nll": -5.926998138427734, "elapsed_s": 37.7}26{"event": "train", "step": 520, "loss": -6.038427829742432, "grad_norm": 0.12013476341962814, "lr": 0.0002999287127590589, "ex_per_s": 880.3890198994252, "data_frac": 0.9443437207193722, "mem_mb": 14.961152, "nll": -6.038427829742432, "elapsed_s": 37.8}27{"event": "train", "step": 540, "loss": -5.876936435699463, "grad_norm": 0.5297210216522217, "lr": 0.00029990297322826756, "ex_per_s": 912.5775984652167, "data_frac": 0.9429180596673864, "mem_mb": 14.961152, "nll": -5.876936435699463, "elapsed_s": 37.9}28{"event": "train", "step": 560, "loss": -5.974250316619873, "grad_norm": 0.14046959578990936, "lr": 0.00029987327580293684, "ex_per_s": 944.6431229874887, "data_frac": 0.9415095021264577, "mem_mb": 14.961152, "nll": -5.974250316619873, "elapsed_s": 37.9}29{"event": "train", "step": 580, "loss": -5.9239935874938965, "grad_norm": 0.15552447736263275, "lr": 0.00029983962135435424, "ex_per_s": 976.5628799415595, "data_frac": 0.940078351054604, "mem_mb": 14.961152, "nll": -5.9239935874938965, "elapsed_s": 38.0}30{"event": "train", "step": 600, "loss": -5.970829486846924, "grad_norm": 0.2475355714559555, "lr": 0.0002998020108699017, "ex_per_s": 1008.3372958121383, "data_frac": 0.9386251302600149, "mem_mb": 14.961152, "nll": -5.970829486846924, "elapsed_s": 38.1}31{"event": "train", "step": 620, "loss": -5.946911334991455, "grad_norm": 0.30326902866363525, "lr": 0.0002997604454530266, "ex_per_s": 1040.0675242178338, "data_frac": 0.9372339150412915, "mem_mb": 14.961152, "nll": -5.946911334991455, "elapsed_s": 38.2}32{"event": "train", "step": 640, "loss": -6.064590930938721, "grad_norm": 0.32644447684288025, "lr": 0.00029971492632320935, "ex_per_s": 1071.6530048477696, "data_frac": 0.9358332772187511, "mem_mb": 14.961152, "nll": -6.064590930938721, "elapsed_s": 38.2}33{"event": "train", "step": 660, "loss": -6.008476257324219, "grad_norm": 0.5084580183029175, "lr": 0.0002996654548159277, "ex_per_s": 1103.1341927420824, "data_frac": 0.9344415545769187, "mem_mb": 14.961152, "nll": -6.008476257324219, "elapsed_s": 38.3}34{"event": "train", "step": 680, "loss": -6.004378795623779, "grad_norm": 0.16566862165927887, "lr": 0.00029961203238261747, "ex_per_s": 1134.4927636539308, "data_frac": 0.9330433323070846, "mem_mb": 14.961152, "nll": -6.004378795623779, "elapsed_s": 38.4}35{"event": "train", "step": 700, "loss": -5.985960006713867, "grad_norm": 0.5905082821846008, "lr": 0.0002995546605906301, "ex_per_s": 1165.7146950909464, "data_frac": 0.9316344167053165, "mem_mb": 14.961152, "nll": -5.985960006713867, "elapsed_s": 38.4}36{"event": "train", "step": 720, "loss": -6.058958053588867, "grad_norm": 0.1745181530714035, "lr": 0.00029949334112318624, "ex_per_s": 1196.7854060210625, "data_frac": 0.9302096683683782, "mem_mb": 14.961152, "nll": -6.058958053588867, "elapsed_s": 38.5}37{"event": "train", "step": 740, "loss": -6.033651828765869, "grad_norm": 0.18515659868717194, "lr": 0.0002994280757793271, "ex_per_s": 1227.8187630610732, "data_frac": 0.9288386660794417, "mem_mb": 14.961152, "nll": -6.033651828765869, "elapsed_s": 38.6}38{"event": "train", "step": 760, "loss": -5.958449840545654, "grad_norm": 0.25198009610176086, "lr": 0.00029935886647386114, "ex_per_s": 1258.7060010765183, "data_frac": 0.9274454156108423, "mem_mb": 14.961152, "nll": -5.958449840545654, "elapsed_s": 38.6}39{"event": "train", "step": 780, "loss": -5.896336555480957, "grad_norm": 0.0923929512500763, "lr": 0.00029928571523730785, "ex_per_s": 1289.4579091067521, "data_frac": 0.9260497690679195, "mem_mb": 14.961152, "nll": -5.896336555480957, "elapsed_s": 38.7}40{"event": "train", "step": 800, "loss": -5.816554546356201, "grad_norm": 0.15681909024715424, "lr": 0.00029920862421583855, "ex_per_s": 1320.099871986528, "data_frac": 0.9246612129211782, "mem_mb": 14.961152, "nll": -5.816554546356201, "elapsed_s": 38.8}41{"event": "train", "step": 820, "loss": -5.936845302581787, "grad_norm": 0.23472976684570312, "lr": 0.00029912759567121304, "ex_per_s": 1350.70620727768, "data_frac": 0.9233216069506208, "mem_mb": 14.961152, "nll": -5.936845302581787, "elapsed_s": 38.9}42{"event": "train", "step": 840, "loss": -5.8066558837890625, "grad_norm": 0.45447686314582825, "lr": 0.0002990426319807135, "ex_per_s": 1381.1799435911373, "data_frac": 0.9219811563815645, "mem_mb": 14.961152, "nll": -5.8066558837890625, "elapsed_s": 38.9}43{"event": "train", "step": 860, "loss": -5.925714492797852, "grad_norm": 0.17719893157482147, "lr": 0.0002989537356370747, "ex_per_s": 1411.5146548261123, "data_frac": 0.920635202863451, "mem_mb": 14.961152, "nll": -5.925714492797852, "elapsed_s": 39.0}44{"event": "train", "step": 880, "loss": -5.872158527374268, "grad_norm": 0.1082216426730156, "lr": 0.00029886090924841085, "ex_per_s": 1441.790592001375, "data_frac": 0.9193033747037357, "mem_mb": 14.961152, "nll": -5.872158527374268, "elapsed_s": 39.1}45{"event": "train", "step": 900, "loss": -5.987218379974365, "grad_norm": 0.11622272431850433, "lr": 0.0002987641555381389, "ex_per_s": 1471.8337029560898, "data_frac": 0.917908474014774, "mem_mb": 14.961152, "nll": -5.987218379974365, "elapsed_s": 39.1}46{"event": "train", "step": 920, "loss": -6.028843879699707, "grad_norm": 0.09618596732616425, "lr": 0.0002986634773448989, "ex_per_s": 1501.8882842030619, "data_frac": 0.9165783730758131, "mem_mb": 14.961152, "nll": -6.028843879699707, "elapsed_s": 39.2}47{"event": "train", "step": 940, "loss": -6.0775299072265625, "grad_norm": 0.2290959358215332, "lr": 0.00029855887762247066, "ex_per_s": 1531.6389578968613, "data_frac": 0.9151633831782086, "mem_mb": 14.961152, "nll": -6.0775299072265625, "elapsed_s": 39.3}48{"event": "train", "step": 960, "loss": -5.9657416343688965, "grad_norm": 0.2642521858215332, "lr": 0.00029845035943968696, "ex_per_s": 1561.4716897300982, "data_frac": 0.9138353243905183, "mem_mb": 14.961152, "nll": -5.9657416343688965, "elapsed_s": 39.3}49{"event": "train", "step": 980, "loss": -5.928456783294678, "grad_norm": 0.15383940935134888, "lr": 0.0002983379259803436, "ex_per_s": 1591.266546001437, "data_frac": 0.9125565687825924, "mem_mb": 14.961152, "nll": -5.928456783294678, "elapsed_s": 39.4}50{"event": "train", "step": 1000, "loss": -6.068014144897461, "grad_norm": 0.09661108255386353, "lr": 0.00029822158054310604, "ex_per_s": 1620.7630237583705, "data_frac": 0.9111905425632895, "mem_mb": 14.961152, "nll": -6.068014144897461, "elapsed_s": 39.5}51{"event": "val", "step": 1000, "val_nll": -6.446475124806243, "best": -6.446475124806243, "bad": 0, "mem_mb": 14.961152, "elapsed_s": 104.2}52{"event": "train", "step": 1020, "loss": -6.0276360511779785, "grad_norm": 0.09735836833715439, "lr": 0.0002981013265414125, "ex_per_s": 623.9855748033261, "data_frac": 0.3440634726338288, "mem_mb": 14.961152, "nll": -6.0276360511779785, "elapsed_s": 104.6}53{"event": "train", "step": 1040, "loss": -5.977542877197266, "grad_norm": 0.14038924872875214, "lr": 0.000297977167503374, "ex_per_s": 635.7752423361729, "data_frac": 0.3439503889351497, "mem_mb": 14.961152, "nll": -5.977542877197266, "elapsed_s": 104.7}54{"event": "train", "step": 1060, "loss": -6.186011791229248, "grad_norm": 0.127073273062706, "lr": 0.00029784910707167064, "ex_per_s": 647.5761076233367, "data_frac": 0.3438446226176898, "mem_mb": 14.961152, "nll": -6.186011791229248, "elapsed_s": 104.8}55{"event": "train", "step": 1080, "loss": -5.814650058746338, "grad_norm": 0.09049536287784576, "lr": 0.0002977171490034448, "ex_per_s": 659.3568860806574, "data_frac": 0.34373431552779055, "mem_mb": 14.961152, "nll": -5.814650058746338, "elapsed_s": 104.8}56{"event": "train", "step": 1100, "loss": -6.050665855407715, "grad_norm": 0.11502651125192642, "lr": 0.000297581297170191, "ex_per_s": 671.1112137408259, "data_frac": 0.3436209286130533, "mem_mb": 14.961152, "nll": -6.050665855407715, "elapsed_s": 104.9}57{"event": "train", "step": 1120, "loss": -6.056844711303711, "grad_norm": 0.16898213326931, "lr": 0.0002974415555576423, "ex_per_s": 682.8566888531935, "data_frac": 0.3435083320631739, "mem_mb": 14.961152, "nll": -6.056844711303711, "elapsed_s": 105.0}58{"event": "train", "step": 1140, "loss": -6.021590232849121, "grad_norm": 0.10919110476970673, "lr": 0.000297297928265653, "ex_per_s": 694.5832047884285, "data_frac": 0.3433972267702308, "mem_mb": 14.961152, "nll": -6.021590232849121, "elapsed_s": 105.0}59{"event": "train", "step": 1160, "loss": -5.962672710418701, "grad_norm": 0.4101565480232239, "lr": 0.00029715041950807914, "ex_per_s": 706.297114277619, "data_frac": 0.3432935032825229, "mem_mb": 14.961152, "nll": -5.962672710418701, "elapsed_s": 105.1}60{"event": "train", "step": 1180, "loss": -5.95628023147583, "grad_norm": 0.17823253571987152, "lr": 0.000296999033612654, "ex_per_s": 717.9904243927753, "data_frac": 0.34318185899297715, "mem_mb": 14.961152, "nll": -5.95628023147583, "elapsed_s": 105.2}61{"event": "train", "step": 1200, "loss": -5.944998264312744, "grad_norm": 0.07371806353330612, "lr": 0.00029684377502086165, "ex_per_s": 729.6878493368245, "data_frac": 0.3430674414755308, "mem_mb": 14.961152, "nll": -5.944998264312744, "elapsed_s": 105.3}62{"event": "train", "step": 1220, "loss": -6.0360870361328125, "grad_norm": 0.09681524336338043, "lr": 0.00029668464828780665, "ex_per_s": 741.3648537758797, "data_frac": 0.3429586475836477, "mem_mb": 14.961152, "nll": -6.0360870361328125, "elapsed_s": 105.3}63{"event": "train", "step": 1240, "loss": -6.058070659637451, "grad_norm": 0.1108165830373764, "lr": 0.00029652165808208004, "ex_per_s": 753.0208215006918, "data_frac": 0.342845495178137, "mem_mb": 14.961152, "nll": -6.058070659637451, "elapsed_s": 105.4}64{"event": "train", "step": 1260, "loss": -6.047757148742676, "grad_norm": 0.29222574830055237, "lr": 0.0002963548091856228, "ex_per_s": 764.6580422179707, "data_frac": 0.34273451785326065, "mem_mb": 14.961152, "nll": -6.047757148742676, "elapsed_s": 105.5}65{"event": "train", "step": 1280, "loss": -6.048752784729004, "grad_norm": 0.2722478210926056, "lr": 0.0002961841064935852, "ex_per_s": 776.2783956146353, "data_frac": 0.34262221518736535, "mem_mb": 14.961152, "nll": -6.048752784729004, "elapsed_s": 105.5}66{"event": "train", "step": 1300, "loss": -5.868002891540527, "grad_norm": 0.2321803718805313, "lr": 0.0002960095550141835, "ex_per_s": 787.892396910609, "data_frac": 0.3425157417457001, "mem_mb": 14.961152, "nll": -5.868002891540527, "elapsed_s": 105.6}67{"event": "train", "step": 1320, "loss": -5.827094554901123, "grad_norm": 0.14087173342704773, "lr": 0.00029583115986855273, "ex_per_s": 799.4773856492078, "data_frac": 0.34240416841178395, "mem_mb": 14.961152, "nll": -5.827094554901123, "elapsed_s": 105.7}68{"event": "train", "step": 1340, "loss": -5.954585075378418, "grad_norm": 0.20417742431163788, "lr": 0.0002956489262905966, "ex_per_s": 811.0648692550262, "data_frac": 0.3422951819523634, "mem_mb": 14.961152, "nll": -5.954585075378418, "elapsed_s": 105.7}69{"event": "train", "step": 1360, "loss": -6.037257194519043, "grad_norm": 0.1883859485387802, "lr": 0.0002954628596268339, "ex_per_s": 822.6344618856947, "data_frac": 0.3421855551958938, "mem_mb": 14.961152, "nll": -6.037257194519043, "elapsed_s": 105.8}70{"event": "train", "step": 1380, "loss": -6.062933921813965, "grad_norm": 0.11544277518987656, "lr": 0.0002952729653362417, "ex_per_s": 834.1806179313213, "data_frac": 0.34207611681780764, "mem_mb": 14.961152, "nll": -6.062933921813965, "elapsed_s": 105.9}71{"event": "train", "step": 1400, "loss": -5.796350955963135, "grad_norm": 0.17549729347229004, "lr": 0.00029507924899009505, "ex_per_s": 845.720280510436, "data_frac": 0.3419668350003681, "mem_mb": 14.961152, "nll": -5.796350955963135, "elapsed_s": 105.9}72{"event": "train", "step": 1420, "loss": -5.988401889801025, "grad_norm": 0.1354461908340454, "lr": 0.00029488171627180373, "ex_per_s": 857.2367063899108, "data_frac": 0.3418547719579678, "mem_mb": 14.961152, "nll": -5.988401889801025, "elapsed_s": 106.0}73{"event": "train", "step": 1440, "loss": -5.745041847229004, "grad_norm": 0.1457028090953827, "lr": 0.00029468037297674526, "ex_per_s": 868.7321395357827, "data_frac": 0.3417423319280989, "mem_mb": 14.961152, "nll": -5.745041847229004, "elapsed_s": 106.1}74{"event": "train", "step": 1460, "loss": -5.80828332901001, "grad_norm": 0.1405736804008484, "lr": 0.00029447522501209515, "ex_per_s": 880.2174138940617, "data_frac": 0.341631433539495, "mem_mb": 14.961152, "nll": -5.80828332901001, "elapsed_s": 106.2}75{"event": "train", "step": 1480, "loss": -6.042961120605469, "grad_norm": 0.0747639611363411, "lr": 0.00029426627839665323, "ex_per_s": 891.6852869935371, "data_frac": 0.3415240187903769, "mem_mb": 14.961152, "nll": -6.042961120605469, "elapsed_s": 106.2}76{"event": "train", "step": 1500, "loss": -5.957233428955078, "grad_norm": 0.10305960476398468, "lr": 0.00029405353926066754, "ex_per_s": 903.1434864400464, "data_frac": 0.3414200600522305, "mem_mb": 14.961152, "nll": -5.957233428955078, "elapsed_s": 106.3}77{"event": "train", "step": 1520, "loss": -5.949239730834961, "grad_norm": 0.10277055203914642, "lr": 0.000293837013845654, "ex_per_s": 914.5794186008268, "data_frac": 0.3413119149848287, "mem_mb": 14.961152, "nll": -5.949239730834961, "elapsed_s": 106.4}78{"event": "train", "step": 1540, "loss": -5.879891395568848, "grad_norm": 0.10611680150032043, "lr": 0.00029361670850421354, "ex_per_s": 925.9942405208867, "data_frac": 0.3412038570181049, "mem_mb": 14.961152, "nll": -5.879891395568848, "elapsed_s": 106.4}79{"event": "train", "step": 1560, "loss": -5.979611396789551, "grad_norm": 0.16689617931842804, "lr": 0.0002933926296998457, "ex_per_s": 937.3957179909872, "data_frac": 0.3410962269536454, "mem_mb": 14.961152, "nll": -5.979611396789551, "elapsed_s": 106.5}80{"event": "train", "step": 1580, "loss": -5.977689743041992, "grad_norm": 0.12154685705900192, "lr": 0.0002931647840067591, "ex_per_s": 948.8058701937715, "data_frac": 0.3409938577703945, "mem_mb": 14.961152, "nll": -5.977689743041992, "elapsed_s": 106.6}81{"event": "train", "step": 1600, "loss": -5.95443868637085, "grad_norm": 0.11999987810850143, "lr": 0.0002929331781096783, "ex_per_s": 960.1855130989214, "data_frac": 0.34088320990686133, "mem_mb": 14.961152, "nll": -5.95443868637085, "elapsed_s": 106.6}82{"event": "train", "step": 1620, "loss": -5.880439281463623, "grad_norm": 0.09687664359807968, "lr": 0.0002926978188036479, "ex_per_s": 971.5557297658349, "data_frac": 0.3407729943074754, "mem_mb": 14.961152, "nll": -5.880439281463623, "elapsed_s": 106.7}83{"event": "train", "step": 1640, "loss": -6.077223300933838, "grad_norm": 0.09101568907499313, "lr": 0.00029245871299383297, "ex_per_s": 982.8969259483306, "data_frac": 0.34066760999739937, "mem_mb": 14.961152, "nll": -6.077223300933838, "elapsed_s": 106.8}84{"event": "train", "step": 1660, "loss": -6.041635990142822, "grad_norm": 0.2621415853500366, "lr": 0.0002922158676953168, "ex_per_s": 994.2375200847912, "data_frac": 0.3405646461351405, "mem_mb": 14.961152, "nll": -6.041635990142822, "elapsed_s": 106.9}85{"event": "train", "step": 1680, "loss": -5.864790439605713, "grad_norm": 0.1187446340918541, "lr": 0.0002919692900328947, "ex_per_s": 1005.5614870998005, "data_frac": 0.3404596238652981, "mem_mb": 14.961152, "nll": -5.864790439605713, "elapsed_s": 106.9}86{"event": "train", "step": 1700, "loss": -5.924686431884766, "grad_norm": 0.22046978771686554, "lr": 0.0002917189872408653, "ex_per_s": 1016.8630373129326, "data_frac": 0.3403553937068966, "mem_mb": 14.961152, "nll": -5.924686431884766, "elapsed_s": 107.0}87{"event": "train", "step": 1720, "loss": -5.889966011047363, "grad_norm": 0.1997055560350418, "lr": 0.00029146496666281795, "ex_per_s": 1028.1460203308502, "data_frac": 0.3402506328497516, "mem_mb": 14.961152, "nll": -5.889966011047363, "elapsed_s": 107.1}88{"event": "train", "step": 1740, "loss": -6.270978927612305, "grad_norm": 0.14978958666324615, "lr": 0.00029120723575141777, "ex_per_s": 1039.4182362446154, "data_frac": 0.34015012479756523, "mem_mb": 14.961152, "nll": -6.270978927612305, "elapsed_s": 107.1}89{"event": "train", "step": 1760, "loss": -5.929815769195557, "grad_norm": 0.256658673286438, "lr": 0.00029094580206818636, "ex_per_s": 1050.6800698570462, "data_frac": 0.34004461811550624, "mem_mb": 14.961152, "nll": -5.929815769195557, "elapsed_s": 107.2}90{"event": "train", "step": 1780, "loss": -5.865959167480469, "grad_norm": 0.26115861535072327, "lr": 0.00029068067328328045, "ex_per_s": 1061.891968688112, "data_frac": 0.33993912515670077, "mem_mb": 14.961152, "nll": -5.865959167480469, "elapsed_s": 107.3}91{"event": "train", "step": 1800, "loss": -6.04735803604126, "grad_norm": 0.36345911026000977, "lr": 0.0002904118571752668, "ex_per_s": 1073.0874528865909, "data_frac": 0.3398346369720752, "mem_mb": 14.961152, "nll": -6.04735803604126, "elapsed_s": 107.4}92{"event": "train", "step": 1820, "loss": -5.934328556060791, "grad_norm": 0.23729102313518524, "lr": 0.00029013936163089376, "ex_per_s": 1084.2932411952295, "data_frac": 0.339728910293843, "mem_mb": 14.961152, "nll": -5.934328556060791, "elapsed_s": 107.4}93{"event": "train", "step": 1840, "loss": -6.075065612792969, "grad_norm": 0.21399345993995667, "lr": 0.00028986319464486, "ex_per_s": 1095.4778210112745, "data_frac": 0.3396221332709613, "mem_mb": 14.961152, "nll": -6.075065612792969, "elapsed_s": 107.5}94{"event": "train", "step": 1860, "loss": -6.048755168914795, "grad_norm": 0.0715738832950592, "lr": 0.00028958336431958003, "ex_per_s": 1106.6535506630112, "data_frac": 0.3395172729677641, "mem_mb": 14.961152, "nll": -6.048755168914795, "elapsed_s": 107.6}95{"event": "train", "step": 1880, "loss": -5.9473090171813965, "grad_norm": 0.07784496992826462, "lr": 0.0002892998788649464, "ex_per_s": 1117.8164613581787, "data_frac": 0.3394140235735374, "mem_mb": 14.961152, "nll": -5.9473090171813965, "elapsed_s": 107.6}96{"event": "train", "step": 1900, "loss": -5.982857704162598, "grad_norm": 0.0903676226735115, "lr": 0.0002890127465980889, "ex_per_s": 1128.9675610835852, "data_frac": 0.33930842255011445, "mem_mb": 14.961152, "nll": -5.982857704162598, "elapsed_s": 107.7}97{"event": "train", "step": 1920, "loss": -5.963586807250977, "grad_norm": 0.13586680591106415, "lr": 0.0002887219759431305, "ex_per_s": 1140.0913699274981, "data_frac": 0.33920225752728606, "mem_mb": 14.961152, "nll": -5.963586807250977, "elapsed_s": 107.8}98{"event": "train", "step": 1940, "loss": -6.044559478759766, "grad_norm": 0.12602558732032776, "lr": 0.0002884275754309401, "ex_per_s": 1151.215256831883, "data_frac": 0.33909378240201754, "mem_mb": 14.961152, "nll": -6.044559478759766, "elapsed_s": 107.9}99{"event": "train", "step": 1960, "loss": -5.9376959800720215, "grad_norm": 0.07197482883930206, "lr": 0.0002881295536988824, "ex_per_s": 1162.3180372245272, "data_frac": 0.3389866502210202, "mem_mb": 14.961152, "nll": -5.9376959800720215, "elapsed_s": 107.9}100{"event": "train", "step": 1980, "loss": -5.960682392120361, "grad_norm": 0.3401096761226654, "lr": 0.00028782791949056447, "ex_per_s": 1173.4114416334355, "data_frac": 0.3388845686259044, "mem_mb": 14.961152, "nll": -5.960682392120361, "elapsed_s": 108.0}101{"event": "train", "step": 2000, "loss": -5.914988040924072, "grad_norm": 0.11103449761867523, "lr": 0.00028752268165557917, "ex_per_s": 1184.5140557488287, "data_frac": 0.33878257706801673, "mem_mb": 14.961152, "nll": -5.914988040924072, "elapsed_s": 108.1}102{"event": "val", "step": 2000, "val_nll": -6.446842042212479, "best": -6.446842042212479, "bad": 0, "mem_mb": 14.961152, "elapsed_s": 171.3}103{"event": "train", "step": 2020, "loss": -5.919270038604736, "grad_norm": 0.4042128026485443, "lr": 0.00028721384914924547, "ex_per_s": 754.2179433862424, "data_frac": 0.2136498337480582, "mem_mb": 14.961152, "nll": -5.919270038604736, "elapsed_s": 171.4}104{"event": "train", "step": 2040, "loss": -5.874962329864502, "grad_norm": 0.061263445764780045, "lr": 0.0002869014310323458, "ex_per_s": 761.4032020969715, "data_frac": 0.21363912342818184, "mem_mb": 14.961152, "nll": -5.874962329864502, "elapsed_s": 171.5}105{"event": "train", "step": 2060, "loss": -6.145552635192871, "grad_norm": 0.1675245314836502, "lr": 0.0002865854364708602, "ex_per_s": 768.5944784796056, "data_frac": 0.2136287461166802, "mem_mb": 14.961152, "nll": -6.145552635192871, "elapsed_s": 171.5}106{"event": "train", "step": 2080, "loss": -5.956052303314209, "grad_norm": 0.33220118284225464, "lr": 0.00028626587473569746, "ex_per_s": 775.7694195265669, "data_frac": 0.21361729803107235, "mem_mb": 14.961152, "nll": -5.956052303314209, "elapsed_s": 171.6}107{"event": "train", "step": 2100, "loss": -5.89821720123291, "grad_norm": 0.2407446801662445, "lr": 0.0002859427552024228, "ex_per_s": 782.9271449351531, "data_frac": 0.2136039942732369, "mem_mb": 14.961152, "nll": -5.89821720123291, "elapsed_s": 171.7}108{"event": "train", "step": 2120, "loss": -6.1474833488464355, "grad_norm": 0.13943256437778473, "lr": 0.0002856160873509834, "ex_per_s": 790.0688903113388, "data_frac": 0.21358641401275646, "mem_mb": 14.961152, "nll": -6.1474833488464355, "elapsed_s": 171.7}109{"event": "train", "step": 2140, "loss": -5.979060649871826, "grad_norm": 0.124241903424263, "lr": 0.00028528588076542964, "ex_per_s": 797.2017512496743, "data_frac": 0.2135684457938635, "mem_mb": 14.961152, "nll": -5.979060649871826, "elapsed_s": 171.8}110{"event": "train", "step": 2160, "loss": -5.92366361618042, "grad_norm": 0.31827759742736816, "lr": 0.00028495214513363427, "ex_per_s": 804.3340755638905, "data_frac": 0.21355044000228457, "mem_mb": 14.961152, "nll": -5.92366361618042, "elapsed_s": 171.9}111{"event": "train", "step": 2180, "loss": -6.082777976989746, "grad_norm": 0.1256556361913681, "lr": 0.0002846148902470082, "ex_per_s": 811.4571971765632, "data_frac": 0.21353271957611186, "mem_mb": 14.961152, "nll": -6.082777976989746, "elapsed_s": 171.9}112{"event": "train", "step": 2200, "loss": -5.96387243270874, "grad_norm": 0.09161019325256348, "lr": 0.0002842741260002131, "ex_per_s": 818.5769912097139, "data_frac": 0.213513812594345, "mem_mb": 14.961152, "nll": -5.96387243270874, "elapsed_s": 172.0}113{"event": "train", "step": 2220, "loss": -5.821811676025391, "grad_norm": 0.2958911955356598, "lr": 0.00028392986239087094, "ex_per_s": 825.678398062288, "data_frac": 0.21349797584155844, "mem_mb": 14.961152, "nll": -5.821811676025391, "elapsed_s": 172.1}114{"event": "train", "step": 2240, "loss": -5.962153911590576, "grad_norm": 0.16935771703720093, "lr": 0.0002835821095192712, "ex_per_s": 832.7766624314718, "data_frac": 0.21348243690006782, "mem_mb": 14.961152, "nll": -5.962153911590576, "elapsed_s": 172.1}115{"event": "train", "step": 2260, "loss": -5.737075328826904, "grad_norm": 0.13316823542118073, "lr": 0.0002832308775880741, "ex_per_s": 839.8755582747701, "data_frac": 0.21346562043355558, "mem_mb": 14.961152, "nll": -5.737075328826904, "elapsed_s": 172.2}116{"event": "train", "step": 2280, "loss": -6.0102009773254395, "grad_norm": 0.09943722188472748, "lr": 0.0002828761769020111, "ex_per_s": 846.967337287882, "data_frac": 0.21345023779996042, "mem_mb": 14.961152, "nll": -6.0102009773254395, "elapsed_s": 172.3}117{"event": "train", "step": 2300, "loss": -5.998115062713623, "grad_norm": 0.09276874363422394, "lr": 0.0002825180178675832, "ex_per_s": 854.0556415965349, "data_frac": 0.213433668656123, "mem_mb": 14.961152, "nll": -5.998115062713623, "elapsed_s": 172.4}118{"event": "train", "step": 2320, "loss": -6.06288480758667, "grad_norm": 0.27025172114372253, "lr": 0.00028215641099275517, "ex_per_s": 861.1277113772234, "data_frac": 0.21341670987121775, "mem_mb": 14.961152, "nll": -6.06288480758667, "elapsed_s": 172.4}119{"event": "train", "step": 2340, "loss": -5.848974704742432, "grad_norm": 0.1938096433877945, "lr": 0.0002817913668866473, "ex_per_s": 868.2014836260032, "data_frac": 0.2133984911873493, "mem_mb": 14.961152, "nll": -5.848974704742432, "elapsed_s": 172.5}120{"event": "train", "step": 2360, "loss": -5.925163269042969, "grad_norm": 0.10296104848384857, "lr": 0.00028142289625922426, "ex_per_s": 875.2736136892627, "data_frac": 0.21338313199900835, "mem_mb": 14.961152, "nll": -5.925163269042969, "elapsed_s": 172.6}121{"event": "train", "step": 2380, "loss": -5.928082466125488, "grad_norm": 0.16139636933803558, "lr": 0.00028105100992098075, "ex_per_s": 882.3349660848645, "data_frac": 0.2133653666048639, "mem_mb": 14.961152, "nll": -5.928082466125488, "elapsed_s": 172.6}122{"event": "train", "step": 2400, "loss": -5.955637454986572, "grad_norm": 0.21835094690322876, "lr": 0.00028067571878262454, "ex_per_s": 889.395040331587, "data_frac": 0.2133465661229336, "mem_mb": 14.961152, "nll": -5.955637454986572, "elapsed_s": 172.7}123{"event": "train", "step": 2420, "loss": -5.997009754180908, "grad_norm": 0.07024311274290085, "lr": 0.00028029703385475607, "ex_per_s": 896.4517794868658, "data_frac": 0.21333147019946833, "mem_mb": 14.961152, "nll": -5.997009754180908, "elapsed_s": 172.8}124{"event": "train", "step": 2440, "loss": -6.094793319702148, "grad_norm": 0.16711103916168213, "lr": 0.0002799149662475457, "ex_per_s": 903.4991495335457, "data_frac": 0.213316143363966, "mem_mb": 14.961152, "nll": -6.094793319702148, "elapsed_s": 172.8}125{"event": "train", "step": 2460, "loss": -6.134577751159668, "grad_norm": 0.09287960827350616, "lr": 0.0002795295271704076, "ex_per_s": 910.5378088684417, "data_frac": 0.21329691139197166, "mem_mb": 14.961152, "nll": -6.134577751159668, "elapsed_s": 172.9}126{"event": "train", "step": 2480, "loss": -5.939916133880615, "grad_norm": 0.09106054157018661, "lr": 0.00027914072793167077, "ex_per_s": 917.573788714433, "data_frac": 0.21327885738956348, "mem_mb": 14.961152, "nll": -5.939916133880615, "elapsed_s": 173.0}127{"event": "train", "step": 2500, "loss": -5.7605485916137695, "grad_norm": 0.1219828873872757, "lr": 0.0002787485799382476, "ex_per_s": 924.6093203227738, "data_frac": 0.2132640264251121, "mem_mb": 14.961152, "nll": -5.7605485916137695, "elapsed_s": 173.0}128{"event": "train", "step": 2520, "loss": -5.9580559730529785, "grad_norm": 0.18860730528831482, "lr": 0.0002783530946952989, "ex_per_s": 931.6385481733422, "data_frac": 0.21324784905354768, "mem_mb": 14.961152, "nll": -5.9580559730529785, "elapsed_s": 173.1}129{"event": "train", "step": 2540, "loss": -6.026549816131592, "grad_norm": 0.09104927629232407, "lr": 0.00027795428380589644, "ex_per_s": 938.6604221655953, "data_frac": 0.2132336906326943, "mem_mb": 14.961152, "nll": -6.026549816131592, "elapsed_s": 173.2}130{"event": "train", "step": 2560, "loss": -5.946122169494629, "grad_norm": 0.0849103182554245, "lr": 0.00027755215897068257, "ex_per_s": 945.6748518618032, "data_frac": 0.21321648246258712, "mem_mb": 14.961152, "nll": -5.946122169494629, "elapsed_s": 173.3}131{"event": "train", "step": 2580, "loss": -5.958806037902832, "grad_norm": 0.10969379544258118, "lr": 0.0002771467319875269, "ex_per_s": 952.6726387608376, "data_frac": 0.213199623074437, "mem_mb": 14.961152, "nll": -5.958806037902832, "elapsed_s": 173.3}132{"event": "train", "step": 2600, "loss": -5.850579261779785, "grad_norm": 0.13378335535526276, "lr": 0.0002767380147511802, "ex_per_s": 959.6738252782079, "data_frac": 0.2131865489040903, "mem_mb": 14.961152, "nll": -5.850579261779785, "elapsed_s": 173.4}133{"event": "train", "step": 2620, "loss": -5.890560626983643, "grad_norm": 0.27378734946250916, "lr": 0.0002763260192529254, "ex_per_s": 966.6578666780322, "data_frac": 0.21317039415976216, "mem_mb": 14.961152, "nll": -5.890560626983643, "elapsed_s": 173.5}134{"event": "train", "step": 2640, "loss": -5.987339973449707, "grad_norm": 0.060761790722608566, "lr": 0.0002759107575802257, "ex_per_s": 973.6515844241901, "data_frac": 0.21315662775247923, "mem_mb": 14.961152, "nll": -5.987339973449707, "elapsed_s": 173.5}135{"event": "train", "step": 2660, "loss": -5.816434860229492, "grad_norm": 0.0714699849486351, "lr": 0.00027549224191637014, "ex_per_s": 980.6378851367606, "data_frac": 0.2131390262516643, "mem_mb": 14.961152, "nll": -5.816434860229492, "elapsed_s": 173.6}136{"event": "train", "step": 2680, "loss": -5.95816707611084, "grad_norm": 0.2048516720533371, "lr": 0.000275070484540116, "ex_per_s": 987.6103433642874, "data_frac": 0.21311918258620816, "mem_mb": 14.961152, "nll": -5.95816707611084, "elapsed_s": 173.7}137{"event": "train", "step": 2700, "loss": -6.126726150512695, "grad_norm": 0.29610535502433777, "lr": 0.00027464549782532844, "ex_per_s": 994.5742977991092, "data_frac": 0.21310333597934908, "mem_mb": 14.961152, "nll": -6.126726150512695, "elapsed_s": 173.7}138{"event": "train", "step": 2720, "loss": -5.881694316864014, "grad_norm": 0.12349110841751099, "lr": 0.00027421729424061787, "ex_per_s": 1001.5368935833454, "data_frac": 0.2130871113918103, "mem_mb": 14.961152, "nll": -5.881694316864014, "elapsed_s": 173.8}139{"event": "train", "step": 2740, "loss": -6.1250715255737305, "grad_norm": 0.09439755231142044, "lr": 0.0002737858863489736, "ex_per_s": 1008.4978768503261, "data_frac": 0.21307145616795023, "mem_mb": 14.961152, "nll": -6.1250715255737305, "elapsed_s": 173.9}140{"event": "train", "step": 2760, "loss": -5.963832855224609, "grad_norm": 0.1874709278345108, "lr": 0.00027335128680739583, "ex_per_s": 1015.4523276488188, "data_frac": 0.2130544846577905, "mem_mb": 14.961152, "nll": -5.963832855224609, "elapsed_s": 174.0}141{"event": "train", "step": 2780, "loss": -6.011934757232666, "grad_norm": 0.36960676312446594, "lr": 0.00027291350836652374, "ex_per_s": 1022.4043478882875, "data_frac": 0.21303594460224115, "mem_mb": 14.961152, "nll": -6.011934757232666, "elapsed_s": 174.0}142{"event": "train", "step": 2800, "loss": -5.978393077850342, "grad_norm": 0.200324147939682, "lr": 0.00027247256387026185, "ex_per_s": 1029.352106970628, "data_frac": 0.21301726922445405, "mem_mb": 14.961152, "nll": -5.978393077850342, "elapsed_s": 174.1}143{"event": "train", "step": 2820, "loss": -6.006741046905518, "grad_norm": 0.06736978143453598, "lr": 0.000272028466255403, "ex_per_s": 1036.2920874071901, "data_frac": 0.2129993214643242, "mem_mb": 14.961152, "nll": -6.006741046905518, "elapsed_s": 174.2}144{"event": "train", "step": 2840, "loss": -5.876975059509277, "grad_norm": 0.1001080796122551, "lr": 0.0002715812285512487, "ex_per_s": 1043.232769642814, "data_frac": 0.21298438126968186, "mem_mb": 14.961152, "nll": -5.876975059509277, "elapsed_s": 174.2}145{"event": "train", "step": 2860, "loss": -6.050123691558838, "grad_norm": 0.10529230535030365, "lr": 0.0002711308638792272, "ex_per_s": 1050.1707905444218, "data_frac": 0.2129696476445865, "mem_mb": 14.961152, "nll": -6.050123691558838, "elapsed_s": 174.3}146{"event": "train", "step": 2880, "loss": -5.7756781578063965, "grad_norm": 0.08994685113430023, "lr": 0.000270677385452508, "ex_per_s": 1057.0854952236314, "data_frac": 0.21295264855642138, "mem_mb": 14.961152, "nll": -5.7756781578063965, "elapsed_s": 174.4}147{"event": "train", "step": 2900, "loss": -5.790362358093262, "grad_norm": 0.1851780116558075, "lr": 0.00027022080657561484, "ex_per_s": 1064.0010310464195, "data_frac": 0.2129374634574732, "mem_mb": 14.961152, "nll": -5.790362358093262, "elapsed_s": 174.4}148{"event": "train", "step": 2920, "loss": -5.967151165008545, "grad_norm": 0.26720479130744934, "lr": 0.0002697611406440348, "ex_per_s": 1070.8972173254083, "data_frac": 0.212922760606717, "mem_mb": 14.961152, "nll": -5.967151165008545, "elapsed_s": 174.5}149{"event": "train", "step": 2940, "loss": -5.897181034088135, "grad_norm": 0.1676744669675827, "lr": 0.00026929840114382563, "ex_per_s": 1077.8110858236355, "data_frac": 0.21290632849273436, "mem_mb": 14.961152, "nll": -5.897181034088135, "elapsed_s": 174.6}150{"event": "train", "step": 2960, "loss": -5.936959743499756, "grad_norm": 0.10568694770336151, "lr": 0.0002688326016512199, "ex_per_s": 1084.7063953701677, "data_frac": 0.2128904987262222, "mem_mb": 14.961152, "nll": -5.936959743499756, "elapsed_s": 174.6}151{"event": "train", "step": 2980, "loss": -5.887735366821289, "grad_norm": 0.2200692743062973, "lr": 0.0002683637558322267, "ex_per_s": 1091.6066481351088, "data_frac": 0.21287313733000127, "mem_mb": 14.961152, "nll": -5.887735366821289, "elapsed_s": 174.7}152{"event": "train", "step": 3000, "loss": -5.99399995803833, "grad_norm": 0.07083873450756073, "lr": 0.0002678918774422309, "ex_per_s": 1098.488197010212, "data_frac": 0.21285651743839565, "mem_mb": 14.961152, "nll": -5.99399995803833, "elapsed_s": 174.8}153{"event": "val", "step": 3000, "val_nll": -6.446611885596821, "best": -6.446842042212479, "bad": 1, "mem_mb": 14.961152, "elapsed_s": 239.0}154{"event": "train", "step": 3020, "loss": -5.9623870849609375, "grad_norm": 0.3455778658390045, "lr": 0.0002674169803255893, "ex_per_s": 808.0093927246874, "data_frac": 0.15559201611384718, "mem_mb": 14.961152, "nll": -5.9623870849609375, "elapsed_s": 239.2}155{"event": "train", "step": 3040, "loss": -5.857821941375732, "grad_norm": 0.20951561629772186, "lr": 0.0002669390784152248, "ex_per_s": 813.105348727921, "data_frac": 0.15560114044102177, "mem_mb": 14.961152, "nll": -5.857821941375732, "elapsed_s": 239.3}156{"event": "train", "step": 3060, "loss": -6.068150520324707, "grad_norm": 0.20282618701457977, "lr": 0.00026645818573221706, "ex_per_s": 818.2154445157332, "data_frac": 0.15560789810178358, "mem_mb": 14.961152, "nll": -6.068150520324707, "elapsed_s": 239.4}157{"event": "train", "step": 3080, "loss": -5.96551513671875, "grad_norm": 0.10644841194152832, "lr": 0.00026597431638539183, "ex_per_s": 823.3232952161666, "data_frac": 0.15561261918058608, "mem_mb": 14.961152, "nll": -5.96551513671875, "elapsed_s": 239.4}158{"event": "train", "step": 3100, "loss": -6.009099006652832, "grad_norm": 0.2414611130952835, "lr": 0.00026548748457090647, "ex_per_s": 828.431691804352, "data_frac": 0.15561932806092132, "mem_mb": 14.961152, "nll": -6.009099006652832, "elapsed_s": 239.5}159{"event": "train", "step": 3120, "loss": -6.087316513061523, "grad_norm": 0.26880860328674316, "lr": 0.00026499770457183376, "ex_per_s": 833.5364855141049, "data_frac": 0.15562437710083846, "mem_mb": 14.961152, "nll": -6.087316513061523, "elapsed_s": 239.6}160{"event": "train", "step": 3140, "loss": -6.008022308349609, "grad_norm": 0.1738300323486328, "lr": 0.0002645049907577427, "ex_per_s": 838.6407035548465, "data_frac": 0.1556303042060705, "mem_mb": 14.961152, "nll": -6.008022308349609, "elapsed_s": 239.6}161{"event": "train", "step": 3160, "loss": -6.020836353302002, "grad_norm": 0.08528805524110794, "lr": 0.00026400935758427697, "ex_per_s": 843.7335708028802, "data_frac": 0.15563681493301879, "mem_mb": 14.961152, "nll": -6.020836353302002, "elapsed_s": 239.7}162{"event": "train", "step": 3180, "loss": -6.088943004608154, "grad_norm": 0.11990772932767868, "lr": 0.00026351081959273087, "ex_per_s": 848.8254046618506, "data_frac": 0.1556422766920708, "mem_mb": 14.961152, "nll": -6.088943004608154, "elapsed_s": 239.8}163{"event": "train", "step": 3200, "loss": -6.001389503479004, "grad_norm": 0.12861335277557373, "lr": 0.0002630093914096226, "ex_per_s": 853.9139877085034, "data_frac": 0.15564856206579125, "mem_mb": 14.961152, "nll": -6.001389503479004, "elapsed_s": 239.8}164{"event": "train", "step": 3220, "loss": -5.91422176361084, "grad_norm": 0.10820760577917099, "lr": 0.00026250508774626523, "ex_per_s": 859.0019984268226, "data_frac": 0.15565454709310267, "mem_mb": 14.961152, "nll": -5.91422176361084, "elapsed_s": 239.9}165{"event": "train", "step": 3240, "loss": -5.799646377563477, "grad_norm": 0.13758240640163422, "lr": 0.00026199792339833493, "ex_per_s": 864.0862675962917, "data_frac": 0.15565965040445923, "mem_mb": 14.961152, "nll": -5.799646377563477, "elapsed_s": 240.0}166{"event": "train", "step": 3260, "loss": -5.958094120025635, "grad_norm": 0.11537378281354904, "lr": 0.00026148791324543704, "ex_per_s": 869.1661818992878, "data_frac": 0.155664780901268, "mem_mb": 14.961152, "nll": -5.958094120025635, "elapsed_s": 240.0}167{"event": "train", "step": 3280, "loss": -6.015596866607666, "grad_norm": 0.13736435770988464, "lr": 0.0002609750722506695, "ex_per_s": 874.2431719889399, "data_frac": 0.15566924773395507, "mem_mb": 14.961152, "nll": -6.015596866607666, "elapsed_s": 240.1}168{"event": "train", "step": 3300, "loss": -5.807986259460449, "grad_norm": 0.1002291887998581, "lr": 0.0002604594154601839, "ex_per_s": 879.3075631626033, "data_frac": 0.15567992037542525, "mem_mb": 14.961152, "nll": -5.807986259460449, "elapsed_s": 240.2}169{"event": "train", "step": 3320, "loss": -5.98721981048584, "grad_norm": 0.13690806925296783, "lr": 0.0002599409580027438, "ex_per_s": 884.3823749073464, "data_frac": 0.15568462817891796, "mem_mb": 14.961152, "nll": -5.98721981048584, "elapsed_s": 240.3}170{"event": "train", "step": 3340, "loss": -5.9249444007873535, "grad_norm": 0.11766224354505539, "lr": 0.0002594197150892811, "ex_per_s": 889.45235301794, "data_frac": 0.15569052656409804, "mem_mb": 14.961152, "nll": -5.9249444007873535, "elapsed_s": 240.3}171{"event": "train", "step": 3360, "loss": -5.907783508300781, "grad_norm": 0.20187868177890778, "lr": 0.00025889570201244967, "ex_per_s": 894.5197932748273, "data_frac": 0.1556959492060669, "mem_mb": 14.961152, "nll": -5.907783508300781, "elapsed_s": 240.4}172{"event": "train", "step": 3380, "loss": -6.118409633636475, "grad_norm": 0.22412946820259094, "lr": 0.00025836893414617684, "ex_per_s": 899.5858908902458, "data_frac": 0.15570002112487238, "mem_mb": 14.961152, "nll": -6.118409633636475, "elapsed_s": 240.5}173{"event": "train", "step": 3400, "loss": -5.881747722625732, "grad_norm": 0.18055357038974762, "lr": 0.00025783942694521214, "ex_per_s": 904.6425588206725, "data_frac": 0.15570424208037895, "mem_mb": 14.961152, "nll": -5.881747722625732, "elapsed_s": 240.5}174{"event": "train", "step": 3420, "loss": -5.833810806274414, "grad_norm": 0.1700575351715088, "lr": 0.00025730719594467387, "ex_per_s": 909.7033138257694, "data_frac": 0.15570982203709574, "mem_mb": 14.961152, "nll": -5.833810806274414, "elapsed_s": 240.6}175{"event": "train", "step": 3440, "loss": -5.91472053527832, "grad_norm": 0.12473468482494354, "lr": 0.00025677225675959357, "ex_per_s": 914.7540231848029, "data_frac": 0.15571542640895128, "mem_mb": 14.961152, "nll": -5.91472053527832, "elapsed_s": 240.7}176{"event": "train", "step": 3460, "loss": -6.056624412536621, "grad_norm": 0.20722834765911102, "lr": 0.00025623462508445754, "ex_per_s": 919.8060037535853, "data_frac": 0.1557214861983432, "mem_mb": 14.961152, "nll": -6.056624412536621, "elapsed_s": 240.7}177{"event": "train", "step": 3480, "loss": -5.912749767303467, "grad_norm": 0.1442977339029312, "lr": 0.00025569431669274684, "ex_per_s": 924.8580414752942, "data_frac": 0.15572759203094552, "mem_mb": 14.961152, "nll": -5.912749767303467, "elapsed_s": 240.8}178{"event": "train", "step": 3500, "loss": -6.03176736831665, "grad_norm": 0.10039067268371582, "lr": 0.00025515134743647386, "ex_per_s": 929.8928452323262, "data_frac": 0.15573412990480892, "mem_mb": 14.961152, "nll": -6.03176736831665, "elapsed_s": 240.9}179{"event": "train", "step": 3520, "loss": -5.929340839385986, "grad_norm": 0.05733327567577362, "lr": 0.0002546057332457181, "ex_per_s": 934.9335896050441, "data_frac": 0.15574007426910147, "mem_mb": 14.961152, "nll": -5.929340839385986, "elapsed_s": 241.0}180{"event": "train", "step": 3540, "loss": -6.131101131439209, "grad_norm": 0.234149768948555, "lr": 0.00025405749012815786, "ex_per_s": 939.9763447741799, "data_frac": 0.15574517980852007, "mem_mb": 14.961152, "nll": -6.131101131439209, "elapsed_s": 241.0}181{"event": "train", "step": 3560, "loss": -6.1400556564331055, "grad_norm": 0.10604342073202133, "lr": 0.0002535066341686014, "ex_per_s": 945.0140199659503, "data_frac": 0.15574966314593855, "mem_mb": 14.961152, "nll": -6.1400556564331055, "elapsed_s": 241.1}182{"event": "train", "step": 3580, "loss": -5.902013301849365, "grad_norm": 0.1836054027080536, "lr": 0.0002529531815285146, "ex_per_s": 950.0498041968905, "data_frac": 0.15575562550175118, "mem_mb": 14.961152, "nll": -5.902013301849365, "elapsed_s": 241.2}183{"event": "train", "step": 3600, "loss": -6.093515396118164, "grad_norm": 0.1362764537334442, "lr": 0.00025239714844554667, "ex_per_s": 955.0764291160675, "data_frac": 0.1557604729359235, "mem_mb": 14.961152, "nll": -6.093515396118164, "elapsed_s": 241.2}184{"event": "train", "step": 3620, "loss": -5.901439666748047, "grad_norm": 0.4231445789337158, "lr": 0.0002518385512330544, "ex_per_s": 960.1001623755466, "data_frac": 0.1557658307364445, "mem_mb": 14.961152, "nll": -5.901439666748047, "elapsed_s": 241.3}185{"event": "train", "step": 3640, "loss": -6.051137447357178, "grad_norm": 0.17815792560577393, "lr": 0.00025127740627962274, "ex_per_s": 965.1233840641642, "data_frac": 0.15577251070951256, "mem_mb": 14.961152, "nll": -6.051137447357178, "elapsed_s": 241.4}186{"event": "train", "step": 3660, "loss": -5.83503532409668, "grad_norm": 0.12876804172992706, "lr": 0.00025071373004858453, "ex_per_s": 970.1405334021865, "data_frac": 0.15577974101195205, "mem_mb": 14.961152, "nll": -5.83503532409668, "elapsed_s": 241.5}187{"event": "train", "step": 3680, "loss": -5.990623950958252, "grad_norm": 0.12687444686889648, "lr": 0.00025014753907753737, "ex_per_s": 975.1373636371658, "data_frac": 0.15578766950879872, "mem_mb": 14.961152, "nll": -5.990623950958252, "elapsed_s": 241.5}188{"event": "train", "step": 3700, "loss": -6.062057018280029, "grad_norm": 0.287285715341568, "lr": 0.0002495788499778582, "ex_per_s": 980.1304089083172, "data_frac": 0.15579632889382516, "mem_mb": 14.961152, "nll": -6.062057018280029, "elapsed_s": 241.6}189{"event": "train", "step": 3720, "loss": -5.942983627319336, "grad_norm": 0.12044880539178848, "lr": 0.0002490076794342163, "ex_per_s": 985.1233371280747, "data_frac": 0.15580350677643692, "mem_mb": 14.961152, "nll": -5.942983627319336, "elapsed_s": 241.7}190{"event": "train", "step": 3740, "loss": -6.1878437995910645, "grad_norm": 0.13900400698184967, "lr": 0.00024843404420408344, "ex_per_s": 990.1132861492783, "data_frac": 0.15581060996022844, "mem_mb": 14.961152, "nll": -6.1878437995910645, "elapsed_s": 241.8}191{"event": "train", "step": 3760, "loss": -6.146939754486084, "grad_norm": 0.133475199341774, "lr": 0.0002478579611172425, "ex_per_s": 995.1215719701196, "data_frac": 0.15581439939480926, "mem_mb": 14.961152, "nll": -6.146939754486084, "elapsed_s": 241.8}192{"event": "train", "step": 3780, "loss": -5.962177276611328, "grad_norm": 0.08959439396858215, "lr": 0.00024727944707529366, "ex_per_s": 1000.1167794897614, "data_frac": 0.15582098214233925, "mem_mb": 14.961152, "nll": -5.962177276611328, "elapsed_s": 241.9}193{"event": "train", "step": 3800, "loss": -6.095065593719482, "grad_norm": 0.0981341153383255, "lr": 0.0002466985190511583, "ex_per_s": 1005.1108650714219, "data_frac": 0.1558287971435257, "mem_mb": 14.961152, "nll": -6.095065593719482, "elapsed_s": 242.0}194{"event": "train", "step": 3820, "loss": -5.971567630767822, "grad_norm": 0.14333578944206238, "lr": 0.0002461151940885812, "ex_per_s": 1010.0974285848853, "data_frac": 0.15583517496642324, "mem_mb": 14.961152, "nll": -5.971567630767822, "elapsed_s": 242.0}195{"event": "train", "step": 3840, "loss": -5.8684797286987305, "grad_norm": 0.260509192943573, "lr": 0.0002455294893016306, "ex_per_s": 1015.0798538224249, "data_frac": 0.15584253540722096, "mem_mb": 14.961152, "nll": -5.8684797286987305, "elapsed_s": 242.1}196{"event": "train", "step": 3860, "loss": -6.010839462280273, "grad_norm": 0.15305466949939728, "lr": 0.000244941421874196, "ex_per_s": 1020.0565430568215, "data_frac": 0.1558506531627317, "mem_mb": 14.961152, "nll": -6.010839462280273, "elapsed_s": 242.2}197{"event": "train", "step": 3880, "loss": -5.887011528015137, "grad_norm": 0.09382446110248566, "lr": 0.00024435100905948387, "ex_per_s": 1025.0324550409123, "data_frac": 0.15585794249107424, "mem_mb": 14.961152, "nll": -5.887011528015137, "elapsed_s": 242.3}198{"event": "train", "step": 3900, "loss": -6.17846155166626, "grad_norm": 0.1673019528388977, "lr": 0.0002437582681795117, "ex_per_s": 1029.9807498258808, "data_frac": 0.15585959875142477, "mem_mb": 14.961152, "nll": -6.17846155166626, "elapsed_s": 242.3}199{"event": "train", "step": 3920, "loss": -6.036172389984131, "grad_norm": 0.08741491287946701, "lr": 0.00024316321662459957, "ex_per_s": 1034.9639764338026, "data_frac": 0.15586424771804602, "mem_mb": 14.961152, "nll": -6.036172389984131, "elapsed_s": 242.4}200{"event": "train", "step": 3940, "loss": -5.915091514587402, "grad_norm": 0.09522268176078796, "lr": 0.00024256587185286015, "ex_per_s": 1039.9457616580612, "data_frac": 0.1558672652367266, "mem_mb": 14.961152, "nll": -5.915091514587402, "elapsed_s": 242.5}201{"event": "train", "step": 3960, "loss": -6.068744659423828, "grad_norm": 0.12843506038188934, "lr": 0.0002419662513896863, "ex_per_s": 1044.9265731223506, "data_frac": 0.15587259811413634, "mem_mb": 14.961152, "nll": -6.068744659423828, "elapsed_s": 242.5}202{"event": "train", "step": 3980, "loss": -5.803878307342529, "grad_norm": 0.10904710739850998, "lr": 0.00024136437282723714, "ex_per_s": 1049.9102179663646, "data_frac": 0.15587750456849944, "mem_mb": 14.961152, "nll": -5.803878307342529, "elapsed_s": 242.6}203{"event": "train", "step": 4000, "loss": -6.0539045333862305, "grad_norm": 0.21831023693084717, "lr": 0.0002407602538239216, "ex_per_s": 1054.8859859541622, "data_frac": 0.15588377923475072, "mem_mb": 14.961152, "nll": -6.0539045333862305, "elapsed_s": 242.7}204{"event": "val", "step": 4000, "val_nll": -6.4468271468885865, "best": -6.446842042212479, "bad": 2, "mem_mb": 14.961152, "elapsed_s": 305.8}205{"event": "train", "step": 4020, "loss": -5.986087322235107, "grad_norm": 0.09145530313253403, "lr": 0.00024015391210388054, "ex_per_s": 840.9862685810034, "data_frac": 0.12369910353983307, "mem_mb": 14.961152, "nll": -5.986087322235107, "elapsed_s": 305.9}206{"event": "train", "step": 4040, "loss": -6.129614353179932, "grad_norm": 0.10866452008485794, "lr": 0.00023954536545646676, "ex_per_s": 844.9721960492259, "data_frac": 0.12371238245208625, "mem_mb": 14.961152, "nll": -6.129614353179932, "elapsed_s": 306.0}207{"event": "train", "step": 4060, "loss": -5.864843368530273, "grad_norm": 0.19749793410301208, "lr": 0.0002389346317357231, "ex_per_s": 848.9589887803758, "data_frac": 0.1237244752806448, "mem_mb": 14.961152, "nll": -5.864843368530273, "elapsed_s": 306.1}208{"event": "train", "step": 4080, "loss": -5.928064823150635, "grad_norm": 0.1485171765089035, "lr": 0.0002383217288598584, "ex_per_s": 852.947257272101, "data_frac": 0.12373609685286836, "mem_mb": 14.961152, "nll": -5.928064823150635, "elapsed_s": 306.1}209{"event": "train", "step": 4100, "loss": -5.941948890686035, "grad_norm": 0.0779096931219101, "lr": 0.00023770667481072205, "ex_per_s": 856.9343100930106, "data_frac": 0.12374776337249752, "mem_mb": 14.961152, "nll": -5.941948890686035, "elapsed_s": 306.2}210{"event": "train", "step": 4120, "loss": -6.104601860046387, "grad_norm": 0.11398839950561523, "lr": 0.0002370894876332764, "ex_per_s": 860.9173202093156, "data_frac": 0.12376099765373164, "mem_mb": 14.961152, "nll": -6.104601860046387, "elapsed_s": 306.3}211{"event": "train", "step": 4140, "loss": -5.907988548278809, "grad_norm": 0.2973906695842743, "lr": 0.00023647018543506706, "ex_per_s": 864.9016274538806, "data_frac": 0.1237724279460505, "mem_mb": 14.961152, "nll": -5.907988548278809, "elapsed_s": 306.3}212{"event": "train", "step": 4160, "loss": -6.134659290313721, "grad_norm": 0.15679149329662323, "lr": 0.00023584878638569206, "ex_per_s": 868.8834022136487, "data_frac": 0.1237848707897401, "mem_mb": 14.961152, "nll": -6.134659290313721, "elapsed_s": 306.4}213{"event": "train", "step": 4180, "loss": -5.969607830047607, "grad_norm": 0.13009615242481232, "lr": 0.00023522530871626857, "ex_per_s": 872.8621580457659, "data_frac": 0.12379663028018605, "mem_mb": 14.961152, "nll": -5.969607830047607, "elapsed_s": 306.5}214{"event": "train", "step": 4200, "loss": -6.026494979858398, "grad_norm": 0.065815269947052, "lr": 0.0002345997707188979, "ex_per_s": 876.8350652857504, "data_frac": 0.12380965887526582, "mem_mb": 14.961152, "nll": -6.026494979858398, "elapsed_s": 306.6}215{"event": "train", "step": 4220, "loss": -5.965391159057617, "grad_norm": 0.1104944571852684, "lr": 0.00023397219074612914, "ex_per_s": 880.809648458265, "data_frac": 0.12382067549739958, "mem_mb": 14.961152, "nll": -5.965391159057617, "elapsed_s": 306.6}216{"event": "train", "step": 4240, "loss": -5.995607376098633, "grad_norm": 0.0785970613360405, "lr": 0.00023334258721042025, "ex_per_s": 884.7804763300851, "data_frac": 0.1238327163811293, "mem_mb": 14.961152, "nll": -5.995607376098633, "elapsed_s": 306.7}217{"event": "train", "step": 4260, "loss": -5.886627674102783, "grad_norm": 0.07702260464429855, "lr": 0.00023271097858359853, "ex_per_s": 888.7516869870916, "data_frac": 0.12384383010926071, "mem_mb": 14.961152, "nll": -5.886627674102783, "elapsed_s": 306.8}218{"event": "train", "step": 4280, "loss": -5.843306064605713, "grad_norm": 0.08603594452142715, "lr": 0.00023207738339631794, "ex_per_s": 892.7169509952195, "data_frac": 0.12385568847220689, "mem_mb": 14.961152, "nll": -5.843306064605713, "elapsed_s": 306.8}219{"event": "train", "step": 4300, "loss": -6.254765510559082, "grad_norm": 0.2550627887248993, "lr": 0.00023144182023751594, "ex_per_s": 896.6835390209992, "data_frac": 0.1238667588447953, "mem_mb": 14.961152, "nll": -6.254765510559082, "elapsed_s": 306.9}220{"event": "train", "step": 4320, "loss": -6.135263442993164, "grad_norm": 0.13485750555992126, "lr": 0.00023080430775386808, "ex_per_s": 900.6396881761203, "data_frac": 0.12387950665330782, "mem_mb": 14.961152, "nll": -6.135263442993164, "elapsed_s": 307.0}221{"event": "train", "step": 4340, "loss": -6.079629421234131, "grad_norm": 0.11729303002357483, "lr": 0.00023016486464924057, "ex_per_s": 904.6048840581013, "data_frac": 0.12389170632398849, "mem_mb": 14.961152, "nll": -6.079629421234131, "elapsed_s": 307.1}222{"event": "train", "step": 4360, "loss": -5.849893093109131, "grad_norm": 0.08832501620054245, "lr": 0.0002295235096841419, "ex_per_s": 908.5635227042716, "data_frac": 0.12390531938764558, "mem_mb": 14.961152, "nll": -5.849893093109131, "elapsed_s": 307.1}223{"event": "train", "step": 4380, "loss": -6.004482746124268, "grad_norm": 0.1702640950679779, "lr": 0.00022888026167517227, "ex_per_s": 912.5204423132533, "data_frac": 0.1239187735215248, "mem_mb": 14.961152, "nll": -6.004482746124268, "elapsed_s": 307.2}224{"event": "train", "step": 4400, "loss": -6.043374061584473, "grad_norm": 0.12842905521392822, "lr": 0.00022823513949447166, "ex_per_s": 916.4784096853324, "data_frac": 0.12393125931811227, "mem_mb": 14.961152, "nll": -6.043374061584473, "elapsed_s": 307.3}225{"event": "train", "step": 4420, "loss": -5.89332389831543, "grad_norm": 0.12288923561573029, "lr": 0.00022758816206916584, "ex_per_s": 920.4314384045801, "data_frac": 0.12394402773226373, "mem_mb": 14.961152, "nll": -5.89332389831543, "elapsed_s": 307.3}226{"event": "train", "step": 4440, "loss": -6.086895942687988, "grad_norm": 0.10166218131780624, "lr": 0.00022693934838081143, "ex_per_s": 924.3767642983022, "data_frac": 0.12395744409781172, "mem_mb": 14.961152, "nll": -6.086895942687988, "elapsed_s": 307.4}227{"event": "train", "step": 4460, "loss": -5.9322733879089355, "grad_norm": 0.08201119303703308, "lr": 0.0002262887174648388, "ex_per_s": 928.3309627213649, "data_frac": 0.12396862917817936, "mem_mb": 14.961152, "nll": -5.9322733879089355, "elapsed_s": 307.5}228{"event": "train", "step": 4480, "loss": -6.058394908905029, "grad_norm": 0.1776367872953415, "lr": 0.00022563628840999367, "ex_per_s": 932.2819968893174, "data_frac": 0.12398106930011839, "mem_mb": 14.961152, "nll": -6.058394908905029, "elapsed_s": 307.5}229{"event": "train", "step": 4500, "loss": -6.1488800048828125, "grad_norm": 0.27514439821243286, "lr": 0.000224982080357777, "ex_per_s": 936.2228085615501, "data_frac": 0.12399185055239562, "mem_mb": 14.961152, "nll": -6.1488800048828125, "elapsed_s": 307.6}230{"event": "train", "step": 4520, "loss": -6.011183738708496, "grad_norm": 0.13414056599140167, "lr": 0.00022432611250188336, "ex_per_s": 940.167756998624, "data_frac": 0.124004540567429, "mem_mb": 14.961152, "nll": -6.011183738708496, "elapsed_s": 307.7}231{"event": "train", "step": 4540, "loss": -6.0076212882995605, "grad_norm": 0.11041274666786194, "lr": 0.00022366840408763816, "ex_per_s": 944.10900625555, "data_frac": 0.12401812961530388, "mem_mb": 14.961152, "nll": -6.0076212882995605, "elapsed_s": 307.8}232{"event": "train", "step": 4560, "loss": -6.043184280395508, "grad_norm": 0.09590806066989899, "lr": 0.00022300897441143247, "ex_per_s": 948.0418609494109, "data_frac": 0.1240319628515417, "mem_mb": 14.961152, "nll": -6.043184280395508, "elapsed_s": 307.8}233{"event": "train", "step": 4580, "loss": -6.020596027374268, "grad_norm": 0.23691992461681366, "lr": 0.0002223478428201573, "ex_per_s": 951.9640030464165, "data_frac": 0.12404596753357726, "mem_mb": 14.961152, "nll": -6.020596027374268, "elapsed_s": 307.9}234{"event": "train", "step": 4600, "loss": -5.85875940322876, "grad_norm": 0.08900710940361023, "lr": 0.0002216850287106357, "ex_per_s": 955.889696999944, "data_frac": 0.12405978279257095, "mem_mb": 14.961152, "nll": -5.85875940322876, "elapsed_s": 308.0}235{"event": "train", "step": 4620, "loss": -6.023427963256836, "grad_norm": 0.11247517168521881, "lr": 0.00022102055152905406, "ex_per_s": 959.8119197661093, "data_frac": 0.12407474643409151, "mem_mb": 14.961152, "nll": -6.023427963256836, "elapsed_s": 308.1}236{"event": "train", "step": 4640, "loss": -5.8334221839904785, "grad_norm": 0.11775538325309753, "lr": 0.00022035443077039114, "ex_per_s": 963.7337410020643, "data_frac": 0.12409014385050586, "mem_mb": 14.961152, "nll": -5.8334221839904785, "elapsed_s": 308.1}237{"event": "train", "step": 4660, "loss": -5.977541923522949, "grad_norm": 0.14201083779335022, "lr": 0.00021968668597784646, "ex_per_s": 967.6544645552091, "data_frac": 0.12410418514980846, "mem_mb": 14.961152, "nll": -5.977541923522949, "elapsed_s": 308.2}238{"event": "train", "step": 4680, "loss": -6.02756404876709, "grad_norm": 0.1792544573545456, "lr": 0.00021901733674226665, "ex_per_s": 971.5695843425094, "data_frac": 0.12411801665899035, "mem_mb": 14.961152, "nll": -6.02756404876709, "elapsed_s": 308.3}239{"event": "train", "step": 4700, "loss": -5.905233860015869, "grad_norm": 0.11577939242124557, "lr": 0.00021834640270157097, "ex_per_s": 975.4959845848817, "data_frac": 0.12413171609016646, "mem_mb": 14.961152, "nll": -5.905233860015869, "elapsed_s": 308.4}240{"event": "train", "step": 4720, "loss": -6.062660217285156, "grad_norm": 0.06514213234186172, "lr": 0.00021767390354017485, "ex_per_s": 979.4288932127278, "data_frac": 0.12414234018371383, "mem_mb": 14.961152, "nll": -6.062660217285156, "elapsed_s": 308.4}241{"event": "train", "step": 4740, "loss": -6.038116455078125, "grad_norm": 0.10954656451940536, "lr": 0.0002169998589884126, "ex_per_s": 983.3427294000373, "data_frac": 0.12415695147520527, "mem_mb": 14.961152, "nll": -6.038116455078125, "elapsed_s": 308.5}242{"event": "train", "step": 4760, "loss": -5.991015911102295, "grad_norm": 0.26302286982536316, "lr": 0.00021632428882195843, "ex_per_s": 987.2706802712155, "data_frac": 0.12416829173849861, "mem_mb": 14.961152, "nll": -5.991015911102295, "elapsed_s": 308.6}243{"event": "train", "step": 4780, "loss": -6.060482978820801, "grad_norm": 0.274307519197464, "lr": 0.00021564721286124628, "ex_per_s": 991.1943590192317, "data_frac": 0.12418128193147907, "mem_mb": 14.961152, "nll": -6.060482978820801, "elapsed_s": 308.6}244{"event": "train", "step": 4800, "loss": -6.112823963165283, "grad_norm": 0.18395401537418365, "lr": 0.00021496865097088845, "ex_per_s": 995.1086195162402, "data_frac": 0.12419441231342834, "mem_mb": 14.961152, "nll": -6.112823963165283, "elapsed_s": 308.7}245{"event": "train", "step": 4820, "loss": -6.030040740966797, "grad_norm": 0.13292360305786133, "lr": 0.00021428862305909247, "ex_per_s": 999.0281661362446, "data_frac": 0.12420651748150947, "mem_mb": 14.961152, "nll": -6.030040740966797, "elapsed_s": 308.8}246{"event": "train", "step": 4840, "loss": -5.93540620803833, "grad_norm": 0.07460353523492813, "lr": 0.00021360714907707733, "ex_per_s": 1002.9425688389246, "data_frac": 0.12422046908913738, "mem_mb": 14.961152, "nll": -5.93540620803833, "elapsed_s": 308.9}247{"event": "train", "step": 4860, "loss": -5.921117305755615, "grad_norm": 0.2044782042503357, "lr": 0.00021292424901848804, "ex_per_s": 1006.8511934423959, "data_frac": 0.12423374460205339, "mem_mb": 14.961152, "nll": -5.921117305755615, "elapsed_s": 308.9}248{"event": "train", "step": 4880, "loss": -6.025149345397949, "grad_norm": 0.17959444224834442, "lr": 0.00021223994291880895, "ex_per_s": 1010.7386840940816, "data_frac": 0.1242495566426847, "mem_mb": 14.961152, "nll": -6.025149345397949, "elapsed_s": 309.0}249{"event": "train", "step": 4900, "loss": -5.9550862312316895, "grad_norm": 0.08433032780885696, "lr": 0.00021155425085477597, "ex_per_s": 1014.6363392286471, "data_frac": 0.12426384430859688, "mem_mb": 14.961152, "nll": -5.9550862312316895, "elapsed_s": 309.1}250{"event": "train", "step": 4920, "loss": -5.9417619705200195, "grad_norm": 0.0830589085817337, "lr": 0.00021086719294378776, "ex_per_s": 1018.5353341230536, "data_frac": 0.12427980422909464, "mem_mb": 14.961152, "nll": -5.9417619705200195, "elapsed_s": 309.2}251{"event": "train", "step": 4940, "loss": -6.0432915687561035, "grad_norm": 0.1279895156621933, "lr": 0.00021017878934331512, "ex_per_s": 1022.4295614016387, "data_frac": 0.12429417652363225, "mem_mb": 14.961152, "nll": -6.0432915687561035, "elapsed_s": 309.2}252{"event": "train", "step": 4960, "loss": -6.064329147338867, "grad_norm": 0.0851093977689743, "lr": 0.00020948906025030994, "ex_per_s": 1026.3310559756787, "data_frac": 0.12430658579842498, "mem_mb": 14.961152, "nll": -6.064329147338867, "elapsed_s": 309.3}253{"event": "train", "step": 4980, "loss": -6.068655014038086, "grad_norm": 0.12210638076066971, "lr": 0.00020879802590061233, "ex_per_s": 1030.2318653731593, "data_frac": 0.1243203686497404, "mem_mb": 14.961152, "nll": -6.068655014038086, "elapsed_s": 309.4}254{"event": "train", "step": 5000, "loss": -5.883049011230469, "grad_norm": 0.08950211107730865, "lr": 0.00020810570656835728, "ex_per_s": 1034.1155884752552, "data_frac": 0.12433614175846522, "mem_mb": 14.961152, "nll": -5.883049011230469, "elapsed_s": 309.4}255{"event": "val", "step": 5000, "val_nll": -6.448719176094779, "best": -6.448719176094779, "bad": 0, "mem_mb": 14.961152, "elapsed_s": 372.7}256{"event": "train", "step": 5020, "loss": -5.963367938995361, "grad_norm": 0.09375683218240738, "lr": 0.0002074121225653795, "ex_per_s": 861.7881661209315, "data_frac": 0.10323545901612059, "mem_mb": 14.961152, "nll": -5.963367938995361, "elapsed_s": 372.8}257{"event": "train", "step": 5040, "loss": -6.092452049255371, "grad_norm": 0.24026983976364136, "lr": 0.00020671729424061788, "ex_per_s": 865.0630109648905, "data_frac": 0.10324631372051714, "mem_mb": 14.961152, "nll": -6.092452049255371, "elapsed_s": 372.9}258{"event": "train", "step": 5060, "loss": -6.008312702178955, "grad_norm": 0.0990581288933754, "lr": 0.00020602124197951796, "ex_per_s": 868.3399474959441, "data_frac": 0.10325695048532482, "mem_mb": 14.961152, "nll": -6.008312702178955, "elapsed_s": 372.9}259{"event": "train", "step": 5080, "loss": -6.1302642822265625, "grad_norm": 0.11586963385343552, "lr": 0.00020532398620343433, "ex_per_s": 871.61326128171, "data_frac": 0.10326897712305826, "mem_mb": 14.961152, "nll": -6.1302642822265625, "elapsed_s": 373.0}260{"event": "train", "step": 5100, "loss": -6.05790901184082, "grad_norm": 0.08324051648378372, "lr": 0.0002046255473690312, "ex_per_s": 874.885340942606, "data_frac": 0.10328116384493564, "mem_mb": 14.961152, "nll": -6.05790901184082, "elapsed_s": 373.1}261{"event": "train", "step": 5120, "loss": -6.004499912261963, "grad_norm": 0.10619709640741348, "lr": 0.00020392594596768237, "ex_per_s": 878.159061170276, "data_frac": 0.10329332346121005, "mem_mb": 14.961152, "nll": -6.004499912261963, "elapsed_s": 373.1}262{"event": "train", "step": 5140, "loss": -5.942753791809082, "grad_norm": 0.25211775302886963, "lr": 0.00020322520252486997, "ex_per_s": 881.4278832513553, "data_frac": 0.10330524448296854, "mem_mb": 14.961152, "nll": -5.942753791809082, "elapsed_s": 373.2}263{"event": "train", "step": 5160, "loss": -6.159869194030762, "grad_norm": 0.193714439868927, "lr": 0.0002025233375995822, "ex_per_s": 884.6966403975958, "data_frac": 0.10331649626984703, "mem_mb": 14.961152, "nll": -6.159869194030762, "elapsed_s": 373.3}264{"event": "train", "step": 5180, "loss": -5.934330463409424, "grad_norm": 0.07786711305379868, "lr": 0.0002018203717837105, "ex_per_s": 887.9663783627423, "data_frac": 0.10332790893490094, "mem_mb": 14.961152, "nll": -5.934330463409424, "elapsed_s": 373.3}265{"event": "train", "step": 5200, "loss": -5.822294235229492, "grad_norm": 0.19671353697776794, "lr": 0.00020111632570144484, "ex_per_s": 891.2333819998298, "data_frac": 0.10333982218534195, "mem_mb": 14.961152, "nll": -5.822294235229492, "elapsed_s": 373.4}266{"event": "train", "step": 5220, "loss": -5.922746658325195, "grad_norm": 0.06877219676971436, "lr": 0.00020041122000866903, "ex_per_s": 894.5028701571526, "data_frac": 0.10335163334136328, "mem_mb": 14.961152, "nll": -5.922746658325195, "elapsed_s": 373.5}267{"event": "train", "step": 5240, "loss": -5.773873329162598, "grad_norm": 0.08582697063684464, "lr": 0.00019970507539235469, "ex_per_s": 897.7686193212268, "data_frac": 0.10336214354225609, "mem_mb": 14.961152, "nll": -5.773873329162598, "elapsed_s": 373.5}268{"event": "train", "step": 5260, "loss": -5.946480751037598, "grad_norm": 0.09470056742429733, "lr": 0.0001989979125699542, "ex_per_s": 901.0322247864768, "data_frac": 0.10337322091490293, "mem_mb": 14.961152, "nll": -5.946480751037598, "elapsed_s": 373.6}269{"event": "train", "step": 5280, "loss": -6.001761436462402, "grad_norm": 0.2779499888420105, "lr": 0.00019828975228879288, "ex_per_s": 904.2964362116521, "data_frac": 0.10338568649931318, "mem_mb": 14.961152, "nll": -6.001761436462402, "elapsed_s": 373.7}270{"event": "train", "step": 5300, "loss": -5.893306732177734, "grad_norm": 0.14016731083393097, "lr": 0.00019758061532546025, "ex_per_s": 907.5565710362675, "data_frac": 0.10339668453548877, "mem_mb": 14.961152, "nll": -5.893306732177734, "elapsed_s": 373.8}271{"event": "train", "step": 5320, "loss": -6.073132514953613, "grad_norm": 0.17396126687526703, "lr": 0.0001968705224852007, "ex_per_s": 910.8147779661655, "data_frac": 0.103408832365912, "mem_mb": 14.961152, "nll": -6.073132514953613, "elapsed_s": 373.8}272{"event": "train", "step": 5340, "loss": -5.97581672668457, "grad_norm": 0.0861905962228775, "lr": 0.00019615949460130276, "ex_per_s": 914.0744325673112, "data_frac": 0.10341978724575898, "mem_mb": 14.961152, "nll": -5.97581672668457, "elapsed_s": 373.9}273{"event": "train", "step": 5360, "loss": -6.009943962097168, "grad_norm": 0.06937984377145767, "lr": 0.00019544755253448823, "ex_per_s": 917.3314548309482, "data_frac": 0.10343212732601567, "mem_mb": 14.961152, "nll": -6.009943962097168, "elapsed_s": 374.0}274{"event": "train", "step": 5380, "loss": -5.858150959014893, "grad_norm": 0.09860311448574066, "lr": 0.0001947347171722997, "ex_per_s": 920.5892694333552, "data_frac": 0.10344312300817723, "mem_mb": 14.961152, "nll": -5.858150959014893, "elapsed_s": 374.0}275{"event": "train", "step": 5400, "loss": -5.9593682289123535, "grad_norm": 0.07989117503166199, "lr": 0.00019402100942848827, "ex_per_s": 923.8440514701997, "data_frac": 0.10345453497972114, "mem_mb": 14.961152, "nll": -5.9593682289123535, "elapsed_s": 374.1}276{"event": "train", "step": 5420, "loss": -5.834959983825684, "grad_norm": 0.1810283213853836, "lr": 0.00019330645024239946, "ex_per_s": 927.0998467313453, "data_frac": 0.10346630861936998, "mem_mb": 14.961152, "nll": -5.834959983825684, "elapsed_s": 374.2}277{"event": "train", "step": 5440, "loss": -6.014142036437988, "grad_norm": 0.10488606244325638, "lr": 0.00019259106057835926, "ex_per_s": 930.3522648852626, "data_frac": 0.10347733819255249, "mem_mb": 14.961152, "nll": -6.014142036437988, "elapsed_s": 374.2}278{"event": "train", "step": 5460, "loss": -6.094760894775391, "grad_norm": 0.09184689074754715, "lr": 0.00019187486142505883, "ex_per_s": 933.6032726176769, "data_frac": 0.1034887884092551, "mem_mb": 14.961152, "nll": -6.094760894775391, "elapsed_s": 374.3}279{"event": "train", "step": 5480, "loss": -6.095734596252441, "grad_norm": 0.09218557924032211, "lr": 0.00019115787379493887, "ex_per_s": 936.8560311456225, "data_frac": 0.1035000353155292, "mem_mb": 14.961152, "nll": -6.095734596252441, "elapsed_s": 374.4}280{"event": "train", "step": 5500, "loss": -6.086924076080322, "grad_norm": 0.23188430070877075, "lr": 0.00019044011872357296, "ex_per_s": 940.1062048906089, "data_frac": 0.10351220440240273, "mem_mb": 14.961152, "nll": -6.086924076080322, "elapsed_s": 374.4}281{"event": "train", "step": 5520, "loss": -5.920999050140381, "grad_norm": 0.10800237953662872, "lr": 0.00018972161726905055, "ex_per_s": 943.3528093304504, "data_frac": 0.10352401139245399, "mem_mb": 14.961152, "nll": -5.920999050140381, "elapsed_s": 374.5}282{"event": "train", "step": 5540, "loss": -5.918209552764893, "grad_norm": 0.17272450029850006, "lr": 0.00018900239051135913, "ex_per_s": 946.5996603719542, "data_frac": 0.10353607358541206, "mem_mb": 14.961152, "nll": -5.918209552764893, "elapsed_s": 374.6}283{"event": "train", "step": 5560, "loss": -5.976110935211182, "grad_norm": 0.17715729773044586, "lr": 0.00018828245955176568, "ex_per_s": 949.8462261288948, "data_frac": 0.10354702667905048, "mem_mb": 14.961152, "nll": -5.976110935211182, "elapsed_s": 374.6}284{"event": "train", "step": 5580, "loss": -6.032980442047119, "grad_norm": 0.08950994908809662, "lr": 0.00018756184551219765, "ex_per_s": 953.0910476422907, "data_frac": 0.10355882301067947, "mem_mb": 14.961152, "nll": -6.032980442047119, "elapsed_s": 374.7}285{"event": "train", "step": 5600, "loss": -6.083935260772705, "grad_norm": 0.12446063756942749, "lr": 0.00018684056953462326, "ex_per_s": 956.3292996066788, "data_frac": 0.10356968296577435, "mem_mb": 14.961152, "nll": -6.083935260772705, "elapsed_s": 374.8}286{"event": "train", "step": 5620, "loss": -5.817946434020996, "grad_norm": 0.09588411450386047, "lr": 0.00018611865278043117, "ex_per_s": 959.559578953797, "data_frac": 0.10358438085057167, "mem_mb": 14.961152, "nll": -5.817946434020996, "elapsed_s": 374.8}287{"event": "train", "step": 5640, "loss": -6.117686748504639, "grad_norm": 0.26208221912384033, "lr": 0.00018539611642980975, "ex_per_s": 962.7903508283559, "data_frac": 0.10359938036195256, "mem_mb": 14.961152, "nll": -6.117686748504639, "elapsed_s": 374.9}288{"event": "train", "step": 5660, "loss": -6.006849765777588, "grad_norm": 0.08483908325433731, "lr": 0.00018467298168112556, "ex_per_s": 966.0213124458993, "data_frac": 0.10361310459907416, "mem_mb": 14.961152, "nll": -6.006849765777588, "elapsed_s": 375.0}289{"event": "train", "step": 5680, "loss": -5.965648174285889, "grad_norm": 0.06973131000995636, "lr": 0.00018394926975030145, "ex_per_s": 969.2588890584913, "data_frac": 0.10362421952213152, "mem_mb": 14.961152, "nll": -5.965648174285889, "elapsed_s": 375.1}290{"event": "train", "step": 5700, "loss": -6.031251430511475, "grad_norm": 0.16786521673202515, "lr": 0.00018322500187019415, "ex_per_s": 972.4949656970713, "data_frac": 0.10363515530080021, "mem_mb": 14.961152, "nll": -6.031251430511475, "elapsed_s": 375.1}291{"event": "train", "step": 5720, "loss": -5.8318400382995605, "grad_norm": 0.14666233956813812, "lr": 0.00018250019928997124, "ex_per_s": 975.7326817110347, "data_frac": 0.10364784586698314, "mem_mb": 14.961152, "nll": -5.8318400382995605, "elapsed_s": 375.2}292{"event": "train", "step": 5740, "loss": -5.998716831207275, "grad_norm": 0.1572016477584839, "lr": 0.00018177488327448776, "ex_per_s": 978.9672736415184, "data_frac": 0.10365834457645766, "mem_mb": 14.961152, "nll": -5.998716831207275, "elapsed_s": 375.3}293{"event": "train", "step": 5760, "loss": -5.995051860809326, "grad_norm": 0.1405194252729416, "lr": 0.00018104907510366235, "ex_per_s": 982.1974833252301, "data_frac": 0.10367021491269013, "mem_mb": 14.961152, "nll": -5.995051860809326, "elapsed_s": 375.3}294{"event": "train", "step": 5780, "loss": -6.005426406860352, "grad_norm": 0.0839378833770752, "lr": 0.00018032279607185304, "ex_per_s": 985.4302936881057, "data_frac": 0.1036820458177938, "mem_mb": 14.961152, "nll": -6.005426406860352, "elapsed_s": 375.4}295{"event": "train", "step": 5800, "loss": -5.9831037521362305, "grad_norm": 0.20081938803195953, "lr": 0.00017959606748723213, "ex_per_s": 988.6636154686325, "data_frac": 0.10369304685957903, "mem_mb": 14.961152, "nll": -5.9831037521362305, "elapsed_s": 375.5}296{"event": "train", "step": 5820, "loss": -5.907012939453125, "grad_norm": 0.1120038703083992, "lr": 0.00017886891067116147, "ex_per_s": 991.8924054211508, "data_frac": 0.10370584373151229, "mem_mb": 14.961152, "nll": -5.907012939453125, "elapsed_s": 375.5}297{"event": "train", "step": 5840, "loss": -5.909180641174316, "grad_norm": 0.29457563161849976, "lr": 0.00017814134695756664, "ex_per_s": 995.1209253013226, "data_frac": 0.10371803571615491, "mem_mb": 14.961152, "nll": -5.909180641174316, "elapsed_s": 375.6}298{"event": "train", "step": 5860, "loss": -6.002812385559082, "grad_norm": 0.08721558749675751, "lr": 0.00017741339769231104, "ex_per_s": 998.3454653625512, "data_frac": 0.10372992921643236, "mem_mb": 14.961152, "nll": -6.002812385559082, "elapsed_s": 375.7}299{"event": "train", "step": 5880, "loss": -6.010322093963623, "grad_norm": 0.11468011885881424, "lr": 0.00017668508423256986, "ex_per_s": 1001.5721691312425, "data_frac": 0.1037412076026584, "mem_mb": 14.961152, "nll": -6.010322093963623, "elapsed_s": 375.7}300{"event": "train", "step": 5900, "loss": -6.133791446685791, "grad_norm": 0.16224823892116547, "lr": 0.0001759564279462032, "ex_per_s": 1004.7968157997564, "data_frac": 0.10375213586365278, "mem_mb": 14.961152, "nll": -6.133791446685791, "elapsed_s": 375.8}301{"event": "train", "step": 5920, "loss": -6.001173973083496, "grad_norm": 0.13934887945652008, "lr": 0.00017522745021112944, "ex_per_s": 1008.0182066721426, "data_frac": 0.10376436382410135, "mem_mb": 14.961152, "nll": -6.001173973083496, "elapsed_s": 375.9}302{"event": "train", "step": 5940, "loss": -5.939364910125732, "grad_norm": 0.08651816099882126, "lr": 0.0001744981724146976, "ex_per_s": 1011.2399739682224, "data_frac": 0.10377624538487558, "mem_mb": 14.961152, "nll": -5.939364910125732, "elapsed_s": 375.9}303{"event": "train", "step": 5960, "loss": -5.902010917663574, "grad_norm": 0.24994558095932007, "lr": 0.00017376861595306043, "ex_per_s": 1014.4577459963386, "data_frac": 0.10378790080361568, "mem_mb": 14.961152, "nll": -5.902010917663574, "elapsed_s": 376.0}304{"event": "train", "step": 5980, "loss": -6.029720306396484, "grad_norm": 0.14572849869728088, "lr": 0.00017303880223054632, "ex_per_s": 1017.6799576267236, "data_frac": 0.10379959955683625, "mem_mb": 14.961152, "nll": -6.029720306396484, "elapsed_s": 376.1}305{"event": "train", "step": 6000, "loss": -5.987454891204834, "grad_norm": 0.08628512173891068, "lr": 0.00017230875265903138, "ex_per_s": 1020.8980895995616, "data_frac": 0.1038116575537506, "mem_mb": 14.961152, "nll": -5.987454891204834, "elapsed_s": 376.1}306{"event": "val", "step": 6000, "val_nll": -6.44887495486764, "best": -6.44887495486764, "bad": 0, "mem_mb": 14.961152, "elapsed_s": 438.3}307{"event": "train", "step": 6020, "loss": -5.844367504119873, "grad_norm": 0.10072235018014908, "lr": 0.0001715784886573111, "ex_per_s": 878.7459697295716, "data_frac": 0.08908587981119494, "mem_mb": 14.961152, "nll": -5.844367504119873, "elapsed_s": 438.4}308{"event": "train", "step": 6040, "loss": -5.860933303833008, "grad_norm": 0.20887157320976257, "lr": 0.0001708480316504724, "ex_per_s": 881.5261066803747, "data_frac": 0.08909832390762715, "mem_mb": 14.961152, "nll": -5.860933303833008, "elapsed_s": 438.5}309{"event": "train", "step": 6060, "loss": -6.02325963973999, "grad_norm": 0.13839182257652283, "lr": 0.00017011740306926445, "ex_per_s": 884.3078088673188, "data_frac": 0.08911069881195931, "mem_mb": 14.961152, "nll": -6.02325963973999, "elapsed_s": 438.6}310{"event": "train", "step": 6080, "loss": -6.005561828613281, "grad_norm": 0.12107596546411514, "lr": 0.00016938662434947037, "ex_per_s": 887.0895368974114, "data_frac": 0.08912310810946628, "mem_mb": 14.961152, "nll": -6.005561828613281, "elapsed_s": 438.6}311{"event": "train", "step": 6100, "loss": -6.136677265167236, "grad_norm": 0.1918133944272995, "lr": 0.00016865571693127818, "ex_per_s": 889.8683878698441, "data_frac": 0.08913538031228549, "mem_mb": 14.961152, "nll": -6.136677265167236, "elapsed_s": 438.7}312{"event": "train", "step": 6120, "loss": -5.9946980476379395, "grad_norm": 0.17034538090229034, "lr": 0.00016792470225865166, "ex_per_s": 892.6517792538637, "data_frac": 0.08914713642784705, "mem_mb": 14.961152, "nll": -5.9946980476379395, "elapsed_s": 438.8}313{"event": "train", "step": 6140, "loss": -5.996318340301514, "grad_norm": 0.11628734320402145, "lr": 0.00016719360177870138, "ex_per_s": 895.4300670864019, "data_frac": 0.08915945564321812, "mem_mb": 14.961152, "nll": -5.996318340301514, "elapsed_s": 438.9}314{"event": "train", "step": 6160, "loss": -6.063632011413574, "grad_norm": 0.05932343751192093, "lr": 0.0001664624369410554, "ex_per_s": 898.2104691259871, "data_frac": 0.0891717193580036, "mem_mb": 14.961152, "nll": -6.063632011413574, "elapsed_s": 438.9}315{"event": "train", "step": 6180, "loss": -6.040646553039551, "grad_norm": 0.10294584929943085, "lr": 0.00016573122919722992, "ex_per_s": 900.9900603255, "data_frac": 0.08918382617997178, "mem_mb": 14.961152, "nll": -6.040646553039551, "elapsed_s": 439.0}316{"event": "train", "step": 6200, "loss": -5.988370895385742, "grad_norm": 0.0687442496418953, "lr": 0.000165, "ex_per_s": 903.7663217211689, "data_frac": 0.08919568444699812, "mem_mb": 14.961152, "nll": -5.988370895385742, "elapsed_s": 439.1}317{"event": "train", "step": 6220, "loss": -5.930605888366699, "grad_norm": 0.0766947939991951, "lr": 0.00016426877080277007, "ex_per_s": 906.5406012484394, "data_frac": 0.08920722966054724, "mem_mb": 14.961152, "nll": -5.930605888366699, "elapsed_s": 439.1}318{"event": "train", "step": 6240, "loss": -5.899495601654053, "grad_norm": 0.12078344076871872, "lr": 0.0001635375630589446, "ex_per_s": 909.3160043292924, "data_frac": 0.08921898251244571, "mem_mb": 14.961152, "nll": -5.899495601654053, "elapsed_s": 439.2}319{"event": "train", "step": 6260, "loss": -5.906186580657959, "grad_norm": 0.07096575200557709, "lr": 0.00016280639822129864, "ex_per_s": 912.0907683810129, "data_frac": 0.08923117427659734, "mem_mb": 14.961152, "nll": -5.906186580657959, "elapsed_s": 439.3}320{"event": "train", "step": 6280, "loss": -5.901670932769775, "grad_norm": 0.1033553034067154, "lr": 0.00016207529774134834, "ex_per_s": 914.8642700526107, "data_frac": 0.08924300512502424, "mem_mb": 14.961152, "nll": -5.901670932769775, "elapsed_s": 439.3}321{"event": "train", "step": 6300, "loss": -5.981564044952393, "grad_norm": 0.06331702321767807, "lr": 0.00016134428306872187, "ex_per_s": 917.6378953770877, "data_frac": 0.08925492667359163, "mem_mb": 14.961152, "nll": -5.981564044952393, "elapsed_s": 439.4}322{"event": "train", "step": 6320, "loss": -5.821831226348877, "grad_norm": 0.10978025943040848, "lr": 0.00016061337565052966, "ex_per_s": 920.4110369408568, "data_frac": 0.08926631329583443, "mem_mb": 14.961152, "nll": -5.821831226348877, "elapsed_s": 439.5}323{"event": "train", "step": 6340, "loss": -5.880457878112793, "grad_norm": 0.09370540827512741, "lr": 0.00015988259693073558, "ex_per_s": 923.1820917364493, "data_frac": 0.08927884499352869, "mem_mb": 14.961152, "nll": -5.880457878112793, "elapsed_s": 439.5}324{"event": "train", "step": 6360, "loss": -6.1601881980896, "grad_norm": 0.11077423393726349, "lr": 0.0001591519683495276, "ex_per_s": 925.9510163834135, "data_frac": 0.08929048738126277, "mem_mb": 14.961152, "nll": -6.1601881980896, "elapsed_s": 439.6}325{"event": "train", "step": 6380, "loss": -5.89462423324585, "grad_norm": 0.1705595701932907, "lr": 0.0001584215113426889, "ex_per_s": 928.7187013737472, "data_frac": 0.08930281222099691, "mem_mb": 14.961152, "nll": -5.89462423324585, "elapsed_s": 439.7}326{"event": "train", "step": 6400, "loss": -5.887009620666504, "grad_norm": 0.20853863656520844, "lr": 0.00015769124734096862, "ex_per_s": 931.4873812276229, "data_frac": 0.08931537927869787, "mem_mb": 14.961152, "nll": -5.887009620666504, "elapsed_s": 439.7}327{"event": "train", "step": 6420, "loss": -5.982292175292969, "grad_norm": 0.13201476633548737, "lr": 0.00015696119776945368, "ex_per_s": 934.2473751466907, "data_frac": 0.08932834286887596, "mem_mb": 14.961152, "nll": -5.982292175292969, "elapsed_s": 439.8}328{"event": "train", "step": 6440, "loss": -5.96655797958374, "grad_norm": 0.2233273833990097, "lr": 0.00015623138404693952, "ex_per_s": 937.0121690545708, "data_frac": 0.089340019926863, "mem_mb": 14.961152, "nll": -5.96655797958374, "elapsed_s": 439.9}329{"event": "train", "step": 6460, "loss": -5.926481246948242, "grad_norm": 0.13844650983810425, "lr": 0.0001555018275853024, "ex_per_s": 939.7790007251062, "data_frac": 0.08935218881774848, "mem_mb": 14.961152, "nll": -5.926481246948242, "elapsed_s": 439.9}330{"event": "train", "step": 6480, "loss": -5.930569171905518, "grad_norm": 0.20076122879981995, "lr": 0.0001547725497888706, "ex_per_s": 942.5434662852969, "data_frac": 0.08936510741799356, "mem_mb": 14.961152, "nll": -5.930569171905518, "elapsed_s": 440.0}331{"event": "train", "step": 6500, "loss": -5.956127643585205, "grad_norm": 0.10795086622238159, "lr": 0.0001540435720537968, "ex_per_s": 945.3052707214139, "data_frac": 0.08937709130003663, "mem_mb": 14.961152, "nll": -5.956127643585205, "elapsed_s": 440.1}332{"event": "train", "step": 6520, "loss": -6.037714958190918, "grad_norm": 0.08652206510305405, "lr": 0.0001533149157674301, "ex_per_s": 948.0653941713067, "data_frac": 0.08938987919346031, "mem_mb": 14.961152, "nll": -6.037714958190918, "elapsed_s": 440.1}333{"event": "train", "step": 6540, "loss": -6.152820110321045, "grad_norm": 0.11197800934314728, "lr": 0.00015258660230768893, "ex_per_s": 950.8247180683944, "data_frac": 0.08940264047782462, "mem_mb": 14.961152, "nll": -6.152820110321045, "elapsed_s": 440.2}334{"event": "train", "step": 6560, "loss": -6.046351909637451, "grad_norm": 0.14575771987438202, "lr": 0.0001518586530424334, "ex_per_s": 953.5861862393112, "data_frac": 0.08941423666025332, "mem_mb": 14.961152, "nll": -6.046351909637451, "elapsed_s": 440.3}335{"event": "train", "step": 6580, "loss": -6.04160213470459, "grad_norm": 0.1783641129732132, "lr": 0.00015113108932883853, "ex_per_s": 956.3485474336874, "data_frac": 0.08942602770871666, "mem_mb": 14.961152, "nll": -6.04160213470459, "elapsed_s": 440.3}336{"event": "train", "step": 6600, "loss": -6.010291576385498, "grad_norm": 0.117018923163414, "lr": 0.00015040393251276787, "ex_per_s": 959.1067030299811, "data_frac": 0.08943743392260403, "mem_mb": 14.961152, "nll": -6.010291576385498, "elapsed_s": 440.4}337{"event": "train", "step": 6620, "loss": -6.201038837432861, "grad_norm": 0.07971938699483871, "lr": 0.00014967720392814694, "ex_per_s": 961.8657827604338, "data_frac": 0.0894498864802115, "mem_mb": 14.961152, "nll": -6.201038837432861, "elapsed_s": 440.5}338{"event": "train", "step": 6640, "loss": -6.096444129943848, "grad_norm": 0.10934649407863617, "lr": 0.00014895092489633762, "ex_per_s": 964.6264337068933, "data_frac": 0.08946166849377345, "mem_mb": 14.961152, "nll": -6.096444129943848, "elapsed_s": 440.5}339{"event": "train", "step": 6660, "loss": -5.963422775268555, "grad_norm": 0.09031274169683456, "lr": 0.00014822511672551224, "ex_per_s": 967.381056399058, "data_frac": 0.08947320546718075, "mem_mb": 14.961152, "nll": -5.963422775268555, "elapsed_s": 440.6}340{"event": "train", "step": 6680, "loss": -5.786128520965576, "grad_norm": 0.1460196077823639, "lr": 0.0001474998007100288, "ex_per_s": 970.1381414627149, "data_frac": 0.08948525133387683, "mem_mb": 14.961152, "nll": -5.786128520965576, "elapsed_s": 440.7}341{"event": "train", "step": 6700, "loss": -5.858173370361328, "grad_norm": 0.16738221049308777, "lr": 0.00014677499812980582, "ex_per_s": 972.8892035250321, "data_frac": 0.0894978397248024, "mem_mb": 14.961152, "nll": -5.858173370361328, "elapsed_s": 440.8}342{"event": "train", "step": 6720, "loss": -6.211953163146973, "grad_norm": 0.1454598754644394, "lr": 0.0001460507302496985, "ex_per_s": 975.6452843243957, "data_frac": 0.08950967050105557, "mem_mb": 14.961152, "nll": -6.211953163146973, "elapsed_s": 440.8}343{"event": "train", "step": 6740, "loss": -5.967777729034424, "grad_norm": 0.09378531575202942, "lr": 0.0001453270183188744, "ex_per_s": 978.3946236471755, "data_frac": 0.08952236824975025, "mem_mb": 14.961152, "nll": -5.967777729034424, "elapsed_s": 440.9}344{"event": "train", "step": 6760, "loss": -5.901087760925293, "grad_norm": 0.11874868720769882, "lr": 0.00014460388357019025, "ex_per_s": 981.1474324700313, "data_frac": 0.08953448559072277, "mem_mb": 14.961152, "nll": -5.901087760925293, "elapsed_s": 441.0}345{"event": "train", "step": 6780, "loss": -5.995107650756836, "grad_norm": 0.13502541184425354, "lr": 0.0001438813472195688, "ex_per_s": 983.902216997264, "data_frac": 0.08954637253276358, "mem_mb": 14.961152, "nll": -5.995107650756836, "elapsed_s": 441.0}346{"event": "train", "step": 6800, "loss": -6.085047245025635, "grad_norm": 0.1488775610923767, "lr": 0.00014315943046537674, "ex_per_s": 986.6542407202967, "data_frac": 0.0895580486242501, "mem_mb": 14.961152, "nll": -6.085047245025635, "elapsed_s": 441.1}347{"event": "train", "step": 6820, "loss": -6.1025261878967285, "grad_norm": 0.20075729489326477, "lr": 0.00014243815448780235, "ex_per_s": 989.4028417301129, "data_frac": 0.089569253195136, "mem_mb": 14.961152, "nll": -6.1025261878967285, "elapsed_s": 441.2}348{"event": "train", "step": 6840, "loss": -5.82643985748291, "grad_norm": 0.15100093185901642, "lr": 0.00014171754044823431, "ex_per_s": 992.1521388985282, "data_frac": 0.08958151535329548, "mem_mb": 14.961152, "nll": -5.82643985748291, "elapsed_s": 441.2}349{"event": "train", "step": 6860, "loss": -5.950536727905273, "grad_norm": 0.1239013522863388, "lr": 0.0001409976094886409, "ex_per_s": 994.897988125192, "data_frac": 0.0895940799941263, "mem_mb": 14.961152, "nll": -5.950536727905273, "elapsed_s": 441.3}350{"event": "train", "step": 6880, "loss": -6.027340888977051, "grad_norm": 0.08280550688505173, "lr": 0.00014027838273094942, "ex_per_s": 997.6464651560569, "data_frac": 0.08960574848481043, "mem_mb": 14.961152, "nll": -6.027340888977051, "elapsed_s": 441.4}351{"event": "train", "step": 6900, "loss": -6.004117488861084, "grad_norm": 0.13005660474300385, "lr": 0.00013955988127642706, "ex_per_s": 1000.3934786888125, "data_frac": 0.08961791638113098, "mem_mb": 14.961152, "nll": -6.004117488861084, "elapsed_s": 441.4}352{"event": "train", "step": 6920, "loss": -6.06626558303833, "grad_norm": 0.15813501179218292, "lr": 0.0001388421262050611, "ex_per_s": 1003.1393741941032, "data_frac": 0.08963008277290847, "mem_mb": 14.961152, "nll": -6.06626558303833, "elapsed_s": 441.5}353{"event": "train", "step": 6940, "loss": -6.06243371963501, "grad_norm": 0.14344558119773865, "lr": 0.00013812513857494116, "ex_per_s": 1005.8803643696856, "data_frac": 0.08964373745477533, "mem_mb": 14.961152, "nll": -6.06243371963501, "elapsed_s": 441.6}354{"event": "train", "step": 6960, "loss": -5.740185260772705, "grad_norm": 0.12310043722391129, "lr": 0.00013740893942164073, "ex_per_s": 1008.6187994478312, "data_frac": 0.08965779179142966, "mem_mb": 14.961152, "nll": -5.740185260772705, "elapsed_s": 441.6}355{"event": "train", "step": 6980, "loss": -6.196485996246338, "grad_norm": 0.07646717131137848, "lr": 0.0001366935497576006, "ex_per_s": 1011.3584892540267, "data_frac": 0.08967087551985303, "mem_mb": 14.961152, "nll": -6.196485996246338, "elapsed_s": 441.7}356{"event": "train", "step": 7000, "loss": -6.122445583343506, "grad_norm": 0.10091147571802139, "lr": 0.00013597899057151173, "ex_per_s": 1014.0981117792866, "data_frac": 0.08968342936704088, "mem_mb": 14.961152, "nll": -6.122445583343506, "elapsed_s": 441.8}357{"event": "val", "step": 7000, "val_nll": -6.449065738354402, "best": -6.449065738354402, "bad": 0, "mem_mb": 14.961152, "elapsed_s": 504.3}358{"event": "train", "step": 7020, "loss": -6.1408491134643555, "grad_norm": 0.10167623311281204, "lr": 0.0001352652828277003, "ex_per_s": 890.6727736185084, "data_frac": 0.07856695771887783, "mem_mb": 14.961152, "nll": -6.1408491134643555, "elapsed_s": 504.4}359{"event": "train", "step": 7040, "loss": -6.057835102081299, "grad_norm": 0.14093327522277832, "lr": 0.00013455244746551177, "ex_per_s": 893.089311857437, "data_frac": 0.07857921739288241, "mem_mb": 14.961152, "nll": -6.057835102081299, "elapsed_s": 504.5}360{"event": "train", "step": 7060, "loss": -5.985753059387207, "grad_norm": 0.13353019952774048, "lr": 0.00013384050539869724, "ex_per_s": 895.5066460655229, "data_frac": 0.07859146638012234, "mem_mb": 14.961152, "nll": -5.985753059387207, "elapsed_s": 504.6}361{"event": "train", "step": 7080, "loss": -6.067689418792725, "grad_norm": 0.21548551321029663, "lr": 0.00013312947751479927, "ex_per_s": 897.9218528185088, "data_frac": 0.07860323805727487, "mem_mb": 14.961152, "nll": -6.067689418792725, "elapsed_s": 504.6}362{"event": "train", "step": 7100, "loss": -5.960017681121826, "grad_norm": 0.10629422217607498, "lr": 0.00013241938467453975, "ex_per_s": 900.3367052321285, "data_frac": 0.07861484282442123, "mem_mb": 14.961152, "nll": -5.960017681121826, "elapsed_s": 504.7}363{"event": "train", "step": 7120, "loss": -6.044051170349121, "grad_norm": 0.08966662734746933, "lr": 0.00013171024771120715, "ex_per_s": 902.7521980941759, "data_frac": 0.07862711157798373, "mem_mb": 14.961152, "nll": -6.044051170349121, "elapsed_s": 504.8}364{"event": "train", "step": 7140, "loss": -5.94336462020874, "grad_norm": 0.08161349594593048, "lr": 0.0001310020874300458, "ex_per_s": 905.1641782695345, "data_frac": 0.07863941085085073, "mem_mb": 14.961152, "nll": -5.94336462020874, "elapsed_s": 504.8}365{"event": "train", "step": 7160, "loss": -6.029740333557129, "grad_norm": 0.08646877110004425, "lr": 0.00013029492460764526, "ex_per_s": 907.5790672735723, "data_frac": 0.07865061545918428, "mem_mb": 14.961152, "nll": -6.029740333557129, "elapsed_s": 504.9}366{"event": "train", "step": 7180, "loss": -5.938041687011719, "grad_norm": 0.15620258450508118, "lr": 0.00012958877999133094, "ex_per_s": 909.9944077668936, "data_frac": 0.07866226757083146, "mem_mb": 14.961152, "nll": -5.938041687011719, "elapsed_s": 505.0}367{"event": "train", "step": 7200, "loss": -5.957479476928711, "grad_norm": 0.09727372229099274, "lr": 0.0001288836742985552, "ex_per_s": 912.4090840624461, "data_frac": 0.07867323247804939, "mem_mb": 14.961152, "nll": -5.957479476928711, "elapsed_s": 505.0}368{"event": "train", "step": 7220, "loss": -6.058040618896484, "grad_norm": 0.15365929901599884, "lr": 0.00012817962821628953, "ex_per_s": 914.8225689516968, "data_frac": 0.07868475182778598, "mem_mb": 14.961152, "nll": -6.058040618896484, "elapsed_s": 505.1}369{"event": "train", "step": 7240, "loss": -5.954905986785889, "grad_norm": 0.13783538341522217, "lr": 0.00012747666240041779, "ex_per_s": 917.2335468390569, "data_frac": 0.07869635887360278, "mem_mb": 14.961152, "nll": -5.954905986785889, "elapsed_s": 505.2}370{"event": "train", "step": 7260, "loss": -6.069413661956787, "grad_norm": 0.20502126216888428, "lr": 0.00012677479747513008, "ex_per_s": 919.6447237062232, "data_frac": 0.07870836413916071, "mem_mb": 14.961152, "nll": -6.069413661956787, "elapsed_s": 505.2}371{"event": "train", "step": 7280, "loss": -5.862281322479248, "grad_norm": 0.15028740465641022, "lr": 0.00012607405403231763, "ex_per_s": 922.0536089242146, "data_frac": 0.07872018438470325, "mem_mb": 14.961152, "nll": -5.862281322479248, "elapsed_s": 505.3}372{"event": "train", "step": 7300, "loss": -5.9214558601379395, "grad_norm": 0.12626081705093384, "lr": 0.0001253744526309688, "ex_per_s": 924.4618115899941, "data_frac": 0.07873207892593354, "mem_mb": 14.961152, "nll": -5.9214558601379395, "elapsed_s": 505.4}373{"event": "train", "step": 7320, "loss": -6.070229530334473, "grad_norm": 0.16319192945957184, "lr": 0.00012467601379656567, "ex_per_s": 926.8692648388829, "data_frac": 0.07874387034907704, "mem_mb": 14.961152, "nll": -6.070229530334473, "elapsed_s": 505.4}374{"event": "train", "step": 7340, "loss": -6.015079498291016, "grad_norm": 0.14070017635822296, "lr": 0.000123978758020482, "ex_per_s": 929.2773391525546, "data_frac": 0.07875553951883453, "mem_mb": 14.961152, "nll": -6.015079498291016, "elapsed_s": 505.5}375{"event": "train", "step": 7360, "loss": -5.989681720733643, "grad_norm": 0.10702266544103622, "lr": 0.00012328270575938212, "ex_per_s": 931.6833491357203, "data_frac": 0.07876703599948834, "mem_mb": 14.961152, "nll": -5.989681720733643, "elapsed_s": 505.6}376{"event": "train", "step": 7380, "loss": -6.076192378997803, "grad_norm": 0.1157410517334938, "lr": 0.00012258787743462041, "ex_per_s": 934.088330653238, "data_frac": 0.07877926091232747, "mem_mb": 14.961152, "nll": -6.076192378997803, "elapsed_s": 505.6}377{"event": "train", "step": 7400, "loss": -5.991371154785156, "grad_norm": 0.07339850813150406, "lr": 0.00012189429343164274, "ex_per_s": 936.4958578640465, "data_frac": 0.07879006003724627, "mem_mb": 14.961152, "nll": -5.991371154785156, "elapsed_s": 505.7}378{"event": "train", "step": 7420, "loss": -6.051268100738525, "grad_norm": 0.11950414627790451, "lr": 0.00012120197409938763, "ex_per_s": 938.9026478522304, "data_frac": 0.07880206263589867, "mem_mb": 14.961152, "nll": -6.051268100738525, "elapsed_s": 505.8}379{"event": "train", "step": 7440, "loss": -5.782158374786377, "grad_norm": 0.15085375308990479, "lr": 0.00012051093974969008, "ex_per_s": 941.3057529066326, "data_frac": 0.07881375359984763, "mem_mb": 14.961152, "nll": -5.782158374786377, "elapsed_s": 505.9}380{"event": "train", "step": 7460, "loss": -6.21289587020874, "grad_norm": 0.08874277025461197, "lr": 0.00011982121065668482, "ex_per_s": 943.7097937181214, "data_frac": 0.07882476906217252, "mem_mb": 14.961152, "nll": -6.21289587020874, "elapsed_s": 505.9}381{"event": "train", "step": 7480, "loss": -5.961880683898926, "grad_norm": 0.17047332227230072, "lr": 0.00011913280705621224, "ex_per_s": 946.1115664498259, "data_frac": 0.07883697403992206, "mem_mb": 14.961152, "nll": -5.961880683898926, "elapsed_s": 506.0}382{"event": "train", "step": 7500, "loss": -5.769440174102783, "grad_norm": 0.1634596437215805, "lr": 0.00011844574914522403, "ex_per_s": 948.5164324328041, "data_frac": 0.07884871007722725, "mem_mb": 14.961152, "nll": -5.769440174102783, "elapsed_s": 506.1}383{"event": "train", "step": 7520, "loss": -6.060462951660156, "grad_norm": 0.09834663569927216, "lr": 0.00011776005708119105, "ex_per_s": 950.9181901509346, "data_frac": 0.07886050813591312, "mem_mb": 14.961152, "nll": -6.060462951660156, "elapsed_s": 506.1}384{"event": "train", "step": 7540, "loss": -5.840465068817139, "grad_norm": 0.10704607516527176, "lr": 0.00011707575098151199, "ex_per_s": 953.31919516512, "data_frac": 0.07887253381811958, "mem_mb": 14.961152, "nll": -5.840465068817139, "elapsed_s": 506.2}385{"event": "train", "step": 7560, "loss": -5.894818305969238, "grad_norm": 0.09507128596305847, "lr": 0.00011639285092292267, "ex_per_s": 955.7206149454863, "data_frac": 0.07888374939995905, "mem_mb": 14.961152, "nll": -5.894818305969238, "elapsed_s": 506.3}386{"event": "train", "step": 7580, "loss": -5.867343425750732, "grad_norm": 0.11407788842916489, "lr": 0.00011571137694090757, "ex_per_s": 958.1188629141441, "data_frac": 0.07889574844158381, "mem_mb": 14.961152, "nll": -5.867343425750732, "elapsed_s": 506.3}387{"event": "train", "step": 7600, "loss": -6.054860591888428, "grad_norm": 0.07850214093923569, "lr": 0.00011503134902911152, "ex_per_s": 960.5174821765322, "data_frac": 0.07890658338452657, "mem_mb": 14.961152, "nll": -6.054860591888428, "elapsed_s": 506.4}388{"event": "train", "step": 7620, "loss": -5.946222305297852, "grad_norm": 0.1655166745185852, "lr": 0.00011435278713875375, "ex_per_s": 962.9159915883112, "data_frac": 0.0789185976400942, "mem_mb": 14.961152, "nll": -5.946222305297852, "elapsed_s": 506.5}389{"event": "train", "step": 7640, "loss": -5.909646034240723, "grad_norm": 0.26417461037635803, "lr": 0.00011367571117804159, "ex_per_s": 965.3158096388927, "data_frac": 0.07893037572150718, "mem_mb": 14.961152, "nll": -5.909646034240723, "elapsed_s": 506.5}390{"event": "train", "step": 7660, "loss": -6.043819427490234, "grad_norm": 0.24757875502109528, "lr": 0.00011300014101158744, "ex_per_s": 967.7138757222772, "data_frac": 0.07894152814299617, "mem_mb": 14.961152, "nll": -6.043819427490234, "elapsed_s": 506.6}391{"event": "train", "step": 7680, "loss": -5.956801891326904, "grad_norm": 0.06779872626066208, "lr": 0.00011232609645982515, "ex_per_s": 970.1097613345094, "data_frac": 0.07895285140922831, "mem_mb": 14.961152, "nll": -5.956801891326904, "elapsed_s": 506.7}392{"event": "train", "step": 7700, "loss": -5.903902530670166, "grad_norm": 0.14860878884792328, "lr": 0.00011165359729842902, "ex_per_s": 972.5036892944921, "data_frac": 0.07896449937890583, "mem_mb": 14.961152, "nll": -5.903902530670166, "elapsed_s": 506.7}393{"event": "train", "step": 7720, "loss": -6.037815093994141, "grad_norm": 0.0786558985710144, "lr": 0.00011098266325773331, "ex_per_s": 974.9010047496455, "data_frac": 0.07897636083522086, "mem_mb": 14.961152, "nll": -6.037815093994141, "elapsed_s": 506.8}394{"event": "train", "step": 7740, "loss": -6.035426616668701, "grad_norm": 0.17839154601097107, "lr": 0.00011031331402215353, "ex_per_s": 977.2947896433408, "data_frac": 0.0789884703693483, "mem_mb": 14.961152, "nll": -6.035426616668701, "elapsed_s": 506.9}395{"event": "train", "step": 7760, "loss": -5.967434406280518, "grad_norm": 0.07800859212875366, "lr": 0.00010964556922960883, "ex_per_s": 979.6919700641291, "data_frac": 0.07899926959335167, "mem_mb": 14.961152, "nll": -5.967434406280518, "elapsed_s": 506.9}396{"event": "train", "step": 7780, "loss": -6.043909549713135, "grad_norm": 0.08227361738681793, "lr": 0.00010897944847094595, "ex_per_s": 982.0842618725533, "data_frac": 0.07901194741630763, "mem_mb": 14.961152, "nll": -6.043909549713135, "elapsed_s": 507.0}397{"event": "train", "step": 7800, "loss": -6.056555271148682, "grad_norm": 0.1141757220029831, "lr": 0.00010831497128936428, "ex_per_s": 984.4761263657766, "data_frac": 0.07902375073577933, "mem_mb": 14.961152, "nll": -6.056555271148682, "elapsed_s": 507.1}398{"event": "train", "step": 7820, "loss": -5.894943714141846, "grad_norm": 0.07790659368038177, "lr": 0.00010765215717984271, "ex_per_s": 986.8697185086243, "data_frac": 0.07903494073860468, "mem_mb": 14.961152, "nll": -5.894943714141846, "elapsed_s": 507.1}399{"event": "train", "step": 7840, "loss": -5.953675270080566, "grad_norm": 0.07442066818475723, "lr": 0.00010699102558856754, "ex_per_s": 989.2594328458381, "data_frac": 0.0790471513709059, "mem_mb": 14.961152, "nll": -5.953675270080566, "elapsed_s": 507.2}400{"event": "train", "step": 7860, "loss": -6.114635467529297, "grad_norm": 0.12587891519069672, "lr": 0.0001063315959123618, "ex_per_s": 991.6471779119976, "data_frac": 0.07905962985255548, "mem_mb": 14.961152, "nll": -6.114635467529297, "elapsed_s": 507.3}401{"event": "train", "step": 7880, "loss": -6.06193733215332, "grad_norm": 0.05905122682452202, "lr": 0.0001056738874981166, "ex_per_s": 994.0376755994556, "data_frac": 0.07907200701554878, "mem_mb": 14.961152, "nll": -6.06193733215332, "elapsed_s": 507.3}402{"event": "train", "step": 7900, "loss": -6.049515724182129, "grad_norm": 0.20055194199085236, "lr": 0.00010501791964222304, "ex_per_s": 996.4255595405413, "data_frac": 0.0790839388862102, "mem_mb": 14.961152, "nll": -6.049515724182129, "elapsed_s": 507.4}403{"event": "train", "step": 7920, "loss": -6.077322006225586, "grad_norm": 0.07299190014600754, "lr": 0.00010436371159000636, "ex_per_s": 998.8161346020618, "data_frac": 0.0790954046157394, "mem_mb": 14.961152, "nll": -6.077322006225586, "elapsed_s": 507.5}404{"event": "train", "step": 7940, "loss": -6.048182964324951, "grad_norm": 0.09752815961837769, "lr": 0.00010371128253516118, "ex_per_s": 1001.2077559468415, "data_frac": 0.07910719390805, "mem_mb": 14.961152, "nll": -6.048182964324951, "elapsed_s": 507.5}405{"event": "train", "step": 7960, "loss": -5.991184234619141, "grad_norm": 0.11893235892057419, "lr": 0.00010306065161918857, "ex_per_s": 1003.5967014230325, "data_frac": 0.07911907293189152, "mem_mb": 14.961152, "nll": -5.991184234619141, "elapsed_s": 507.6}406{"event": "train", "step": 7980, "loss": -5.835948467254639, "grad_norm": 0.11893469840288162, "lr": 0.00010241183793083414, "ex_per_s": 1005.98292106602, "data_frac": 0.0791310358995771, "mem_mb": 14.961152, "nll": -5.835948467254639, "elapsed_s": 507.7}407{"event": "train", "step": 8000, "loss": -6.047072410583496, "grad_norm": 0.07877729088068008, "lr": 0.00010176486050552833, "ex_per_s": 1008.3700104794037, "data_frac": 0.07914351531079009, "mem_mb": 14.961152, "nll": -6.047072410583496, "elapsed_s": 507.8}408{"event": "val", "step": 8000, "val_nll": -6.448821791776209, "best": -6.449065738354402, "bad": 1, "mem_mb": 14.961152, "elapsed_s": 571.0}409{"event": "train", "step": 8020, "loss": -6.037694931030273, "grad_norm": 0.11559657007455826, "lr": 0.00010111973832482765, "ex_per_s": 898.3033878551048, "data_frac": 0.07035057793424411, "mem_mb": 14.961152, "nll": -6.037694931030273, "elapsed_s": 571.4}410{"event": "train", "step": 8040, "loss": -6.0320634841918945, "grad_norm": 0.07275978475809097, "lr": 0.00010047649031585807, "ex_per_s": 900.4336070934351, "data_frac": 0.0703625692660511, "mem_mb": 14.961152, "nll": -6.0320634841918945, "elapsed_s": 571.5}411{"event": "train", "step": 8060, "loss": -5.8281121253967285, "grad_norm": 0.21502023935317993, "lr": 9.983513535075943e-05, "ex_per_s": 902.5657916503876, "data_frac": 0.07037418286873641, "mem_mb": 14.961152, "nll": -5.8281121253967285, "elapsed_s": 571.5}412{"event": "train", "step": 8080, "loss": -6.001591205596924, "grad_norm": 0.14749744534492493, "lr": 9.919569224613192e-05, "ex_per_s": 904.695928983559, "data_frac": 0.07038653867576736, "mem_mb": 14.961152, "nll": -6.001591205596924, "elapsed_s": 571.6}413{"event": "train", "step": 8100, "loss": -5.942444801330566, "grad_norm": 0.11810959130525589, "lr": 9.855817976248399e-05, "ex_per_s": 906.8275043046773, "data_frac": 0.07039792708506479, "mem_mb": 14.961152, "nll": -5.942444801330566, "elapsed_s": 571.7}414{"event": "train", "step": 8120, "loss": -5.894966125488281, "grad_norm": 0.21302050352096558, "lr": 9.792261660368207e-05, "ex_per_s": 908.9586932538725, "data_frac": 0.07041008814620665, "mem_mb": 14.961152, "nll": -5.894966125488281, "elapsed_s": 571.7}415{"event": "train", "step": 8140, "loss": -5.889225006103516, "grad_norm": 0.1008606031537056, "lr": 9.72890214164015e-05, "ex_per_s": 911.0861993810066, "data_frac": 0.07042248192838073, "mem_mb": 14.961152, "nll": -5.889225006103516, "elapsed_s": 571.8}416{"event": "train", "step": 8160, "loss": -6.0356879234313965, "grad_norm": 0.0913577675819397, "lr": 9.66574127895797e-05, "ex_per_s": 913.2158733593906, "data_frac": 0.0704342307232264, "mem_mb": 14.961152, "nll": -6.0356879234313965, "elapsed_s": 571.9}417{"event": "train", "step": 8180, "loss": -5.939759254455566, "grad_norm": 0.12792326509952545, "lr": 9.602780925387089e-05, "ex_per_s": 915.3437144922821, "data_frac": 0.07044613735112087, "mem_mb": 14.961152, "nll": -5.939759254455566, "elapsed_s": 571.9}418{"event": "train", "step": 8200, "loss": -6.1641950607299805, "grad_norm": 0.15360376238822937, "lr": 9.540022928110205e-05, "ex_per_s": 917.4738140429092, "data_frac": 0.070458117666653, "mem_mb": 14.961152, "nll": -6.1641950607299805, "elapsed_s": 572.0}419{"event": "train", "step": 8220, "loss": -5.899892807006836, "grad_norm": 0.0906054750084877, "lr": 9.477469128373143e-05, "ex_per_s": 919.6033306499442, "data_frac": 0.07046962627319883, "mem_mb": 14.961152, "nll": -5.899892807006836, "elapsed_s": 572.1}420{"event": "train", "step": 8240, "loss": -5.838497638702393, "grad_norm": 0.10989989340305328, "lr": 9.415121361430792e-05, "ex_per_s": 921.7275115245413, "data_frac": 0.07048232432226577, "mem_mb": 14.961152, "nll": -5.838497638702393, "elapsed_s": 572.1}421{"event": "train", "step": 8260, "loss": -6.185311794281006, "grad_norm": 0.13307593762874603, "lr": 9.352981456493298e-05, "ex_per_s": 923.8541897448164, "data_frac": 0.07049421841136978, "mem_mb": 14.961152, "nll": -6.185311794281006, "elapsed_s": 572.2}422{"event": "train", "step": 8280, "loss": -6.148599147796631, "grad_norm": 0.08755917847156525, "lr": 9.291051236672361e-05, "ex_per_s": 925.9812469078861, "data_frac": 0.07050550108258324, "mem_mb": 14.961152, "nll": -6.148599147796631, "elapsed_s": 572.3}423{"event": "train", "step": 8300, "loss": -6.040492057800293, "grad_norm": 0.15179698169231415, "lr": 9.229332518927795e-05, "ex_per_s": 928.1081759563033, "data_frac": 0.07051768238709502, "mem_mb": 14.961152, "nll": -6.040492057800293, "elapsed_s": 572.3}424{"event": "train", "step": 8320, "loss": -5.875080585479736, "grad_norm": 0.08485366404056549, "lr": 9.167827114014155e-05, "ex_per_s": 930.2356893435926, "data_frac": 0.07052853917330267, "mem_mb": 14.961152, "nll": -5.875080585479736, "elapsed_s": 572.4}425{"event": "train", "step": 8340, "loss": -6.05177116394043, "grad_norm": 0.1370873898267746, "lr": 9.106536826427686e-05, "ex_per_s": 932.3607652223625, "data_frac": 0.0705411191041518, "mem_mb": 14.961152, "nll": -6.05177116394043, "elapsed_s": 572.5}426{"event": "train", "step": 8360, "loss": -6.009487152099609, "grad_norm": 0.1573295146226883, "lr": 9.045463454353313e-05, "ex_per_s": 934.4824199304865, "data_frac": 0.07055413305730504, "mem_mb": 14.961152, "nll": -6.009487152099609, "elapsed_s": 572.6}427{"event": "train", "step": 8380, "loss": -6.0481486320495605, "grad_norm": 0.13094478845596313, "lr": 8.984608789611947e-05, "ex_per_s": 936.6060647663182, "data_frac": 0.07056663876310999, "mem_mb": 14.961152, "nll": -6.0481486320495605, "elapsed_s": 572.6}428{"event": "train", "step": 8400, "loss": -6.033065319061279, "grad_norm": 0.17649827897548676, "lr": 8.92397461760784e-05, "ex_per_s": 938.7305678852175, "data_frac": 0.0705778882367749, "mem_mb": 14.961152, "nll": -6.033065319061279, "elapsed_s": 572.7}429{"event": "train", "step": 8420, "loss": -5.997753620147705, "grad_norm": 0.10751121491193771, "lr": 8.863562717276283e-05, "ex_per_s": 940.8522415186167, "data_frac": 0.07058978501774969, "mem_mb": 14.961152, "nll": -5.997753620147705, "elapsed_s": 572.8}430{"event": "train", "step": 8440, "loss": -6.024221420288086, "grad_norm": 0.08606202900409698, "lr": 8.803374861031363e-05, "ex_per_s": 942.9745200445768, "data_frac": 0.07060169549330267, "mem_mb": 14.961152, "nll": -6.024221420288086, "elapsed_s": 572.8}431{"event": "train", "step": 8460, "loss": -5.899958610534668, "grad_norm": 0.16255751252174377, "lr": 8.743412814713985e-05, "ex_per_s": 945.0972389026625, "data_frac": 0.07061373295691814, "mem_mb": 14.961152, "nll": -5.899958610534668, "elapsed_s": 572.9}432{"event": "train", "step": 8480, "loss": -5.879104137420654, "grad_norm": 0.11957480013370514, "lr": 8.683678337540044e-05, "ex_per_s": 947.2173070599193, "data_frac": 0.07062576882904269, "mem_mb": 14.961152, "nll": -5.879104137420654, "elapsed_s": 573.0}433{"event": "train", "step": 8500, "loss": -5.965920925140381, "grad_norm": 0.22760112583637238, "lr": 8.62417318204883e-05, "ex_per_s": 949.3386277313793, "data_frac": 0.07063665528687148, "mem_mb": 14.961152, "nll": -5.965920925140381, "elapsed_s": 573.0}434{"event": "train", "step": 8520, "loss": -5.934305667877197, "grad_norm": 0.11351485550403595, "lr": 8.564899094051614e-05, "ex_per_s": 951.4585056518912, "data_frac": 0.07064921793387378, "mem_mb": 14.961152, "nll": -5.934305667877197, "elapsed_s": 573.1}435{"event": "train", "step": 8540, "loss": -5.906402587890625, "grad_norm": 0.11162666231393814, "lr": 8.505857812580402e-05, "ex_per_s": 953.5797509290759, "data_frac": 0.07066123138790008, "mem_mb": 14.961152, "nll": -5.906402587890625, "elapsed_s": 573.2}436{"event": "train", "step": 8560, "loss": -5.965624809265137, "grad_norm": 0.13989055156707764, "lr": 8.447051069836941e-05, "ex_per_s": 955.6987901974627, "data_frac": 0.07067319981135754, "mem_mb": 14.961152, "nll": -5.965624809265137, "elapsed_s": 573.2}437{"event": "train", "step": 8580, "loss": -5.9539031982421875, "grad_norm": 0.08601970970630646, "lr": 8.38848059114188e-05, "ex_per_s": 957.8139001044754, "data_frac": 0.07068561306948036, "mem_mb": 14.961152, "nll": -5.9539031982421875, "elapsed_s": 573.3}438{"event": "train", "step": 8600, "loss": -6.014431953430176, "grad_norm": 0.1593427211046219, "lr": 8.330148094884171e-05, "ex_per_s": 959.9280410859702, "data_frac": 0.07069858526312826, "mem_mb": 14.961152, "nll": -6.014431953430176, "elapsed_s": 573.4}439{"event": "train", "step": 8620, "loss": -5.9904656410217285, "grad_norm": 0.1338035613298416, "lr": 8.27205529247063e-05, "ex_per_s": 962.0387038575586, "data_frac": 0.07071240796891541, "mem_mb": 14.961152, "nll": -5.9904656410217285, "elapsed_s": 573.5}440{"event": "train", "step": 8640, "loss": -5.958902359008789, "grad_norm": 0.21416765451431274, "lr": 8.214203888275743e-05, "ex_per_s": 964.1526186655814, "data_frac": 0.07072429747747017, "mem_mb": 14.961152, "nll": -5.958902359008789, "elapsed_s": 573.5}441{"event": "train", "step": 8660, "loss": -5.945700168609619, "grad_norm": 0.14628197252750397, "lr": 8.156595579591652e-05, "ex_per_s": 966.2660349997125, "data_frac": 0.07073764531273424, "mem_mb": 14.961152, "nll": -5.945700168609619, "elapsed_s": 573.6}442{"event": "train", "step": 8680, "loss": -5.896286964416504, "grad_norm": 0.07468996942043304, "lr": 8.09923205657837e-05, "ex_per_s": 968.3787370471626, "data_frac": 0.07075034795703758, "mem_mb": 14.961152, "nll": -5.896286964416504, "elapsed_s": 573.7}443{"event": "train", "step": 8700, "loss": -5.931527614593506, "grad_norm": 0.08553469181060791, "lr": 8.042115002214181e-05, "ex_per_s": 970.4918364798137, "data_frac": 0.07076307021612759, "mem_mb": 14.961152, "nll": -5.931527614593506, "elapsed_s": 573.7}444{"event": "train", "step": 8720, "loss": -6.162047863006592, "grad_norm": 0.06742456555366516, "lr": 7.985246092246267e-05, "ex_per_s": 972.6072389350726, "data_frac": 0.07077531422072098, "mem_mb": 14.961152, "nll": -6.162047863006592, "elapsed_s": 573.8}445{"event": "train", "step": 8740, "loss": -5.896264553070068, "grad_norm": 0.08874150365591049, "lr": 7.928626995141546e-05, "ex_per_s": 974.7159828636559, "data_frac": 0.07078887811504221, "mem_mb": 14.961152, "nll": -5.896264553070068, "elapsed_s": 573.9}446{"event": "train", "step": 8760, "loss": -6.071666240692139, "grad_norm": 0.12609773874282837, "lr": 7.872259372037726e-05, "ex_per_s": 976.8254025552328, "data_frac": 0.07080270530537403, "mem_mb": 14.961152, "nll": -6.071666240692139, "elapsed_s": 573.9}447{"event": "train", "step": 8780, "loss": -5.978584289550781, "grad_norm": 0.09736111760139465, "lr": 7.816144876694563e-05, "ex_per_s": 978.9366513717005, "data_frac": 0.0708152849141711, "mem_mb": 14.961152, "nll": -5.978584289550781, "elapsed_s": 574.0}448{"event": "train", "step": 8800, "loss": -5.951244354248047, "grad_norm": 0.1480400562286377, "lr": 7.760285155445326e-05, "ex_per_s": 981.04964607631, "data_frac": 0.07082704691404165, "mem_mb": 14.961152, "nll": -5.951244354248047, "elapsed_s": 574.1}449{"event": "train", "step": 8820, "loss": -5.9820146560668945, "grad_norm": 0.08177927136421204, "lr": 7.70468184714854e-05, "ex_per_s": 983.162957826519, "data_frac": 0.07083880461006332, "mem_mb": 14.961152, "nll": -5.9820146560668945, "elapsed_s": 574.1}450{"event": "train", "step": 8840, "loss": -5.955790996551514, "grad_norm": 0.08740390837192535, "lr": 7.649336583139853e-05, "ex_per_s": 985.2774770833142, "data_frac": 0.07085042446912938, "mem_mb": 14.961152, "nll": -5.955790996551514, "elapsed_s": 574.2}451{"event": "train", "step": 8860, "loss": -5.992856025695801, "grad_norm": 0.11355536431074142, "lr": 7.594250987184208e-05, "ex_per_s": 987.3909745802439, "data_frac": 0.07086109073170012, "mem_mb": 14.961152, "nll": -5.992856025695801, "elapsed_s": 574.3}452{"event": "train", "step": 8880, "loss": -5.968749046325684, "grad_norm": 0.11426465958356857, "lr": 7.53942667542819e-05, "ex_per_s": 989.5030114711384, "data_frac": 0.07087334743550747, "mem_mb": 14.961152, "nll": -5.968749046325684, "elapsed_s": 574.4}453{"event": "train", "step": 8900, "loss": -5.943839073181152, "grad_norm": 0.1636037528514862, "lr": 7.484865256352612e-05, "ex_per_s": 991.6171665600816, "data_frac": 0.07088407036434824, "mem_mb": 14.961152, "nll": -5.943839073181152, "elapsed_s": 574.4}454{"event": "train", "step": 8920, "loss": -6.093650817871094, "grad_norm": 0.11364617198705673, "lr": 7.430568330725319e-05, "ex_per_s": 993.7301437507873, "data_frac": 0.07089439505727234, "mem_mb": 14.961152, "nll": -6.093650817871094, "elapsed_s": 574.5}455{"event": "train", "step": 8940, "loss": -6.107052326202393, "grad_norm": 0.1051553338766098, "lr": 7.376537491554246e-05, "ex_per_s": 995.8420353378706, "data_frac": 0.07090552566539968, "mem_mb": 14.961152, "nll": -6.107052326202393, "elapsed_s": 574.6}456{"event": "train", "step": 8960, "loss": -6.052927017211914, "grad_norm": 0.11280475556850433, "lr": 7.322774324040643e-05, "ex_per_s": 997.9540863337799, "data_frac": 0.07091612411358539, "mem_mb": 14.961152, "nll": -6.052927017211914, "elapsed_s": 574.6}457{"event": "train", "step": 8980, "loss": -5.9131388664245605, "grad_norm": 0.2870902121067047, "lr": 7.26928040553261e-05, "ex_per_s": 1000.0629297881026, "data_frac": 0.07092746621612964, "mem_mb": 14.961152, "nll": -5.9131388664245605, "elapsed_s": 574.7}458{"event": "train", "step": 9000, "loss": -6.092285633087158, "grad_norm": 0.08908605575561523, "lr": 7.216057305478784e-05, "ex_per_s": 1002.1721643371561, "data_frac": 0.07093918152119484, "mem_mb": 14.961152, "nll": -6.092285633087158, "elapsed_s": 574.8}459{"event": "val", "step": 9000, "val_nll": -6.449216434135477, "best": -6.449216434135477, "bad": 0, "mem_mb": 14.961152, "elapsed_s": 637.5}460{"event": "train", "step": 9020, "loss": -5.998085021972656, "grad_norm": 0.20215682685375214, "lr": 7.163106585382312e-05, "ex_per_s": 905.4001450627127, "data_frac": 0.06396776058658937, "mem_mb": 14.961152, "nll": -5.998085021972656, "elapsed_s": 637.6}461{"event": "train", "step": 9040, "loss": -6.101938247680664, "grad_norm": 0.10357975959777832, "lr": 7.11042979875503e-05, "ex_per_s": 907.3032186094454, "data_frac": 0.06398194993062076, "mem_mb": 14.961152, "nll": -6.101938247680664, "elapsed_s": 637.7}462{"event": "train", "step": 9060, "loss": -6.077178478240967, "grad_norm": 0.13733920454978943, "lr": 7.058028491071894e-05, "ex_per_s": 909.208328871112, "data_frac": 0.0639954735323599, "mem_mb": 14.961152, "nll": -6.077178478240967, "elapsed_s": 637.7}463{"event": "train", "step": 9080, "loss": -6.092523574829102, "grad_norm": 0.1296919733285904, "lr": 7.005904199725619e-05, "ex_per_s": 911.1153864984815, "data_frac": 0.06400730727570791, "mem_mb": 14.961152, "nll": -6.092523574829102, "elapsed_s": 637.8}464{"event": "train", "step": 9100, "loss": -6.16492223739624, "grad_norm": 0.07300448417663574, "lr": 6.954058453981609e-05, "ex_per_s": 913.0191702073913, "data_frac": 0.06401964313560378, "mem_mb": 14.961152, "nll": -6.16492223739624, "elapsed_s": 637.9}465{"event": "train", "step": 9120, "loss": -5.969149112701416, "grad_norm": 0.08538708835840225, "lr": 6.902492774933048e-05, "ex_per_s": 914.8992268302153, "data_frac": 0.06405316577422569, "mem_mb": 14.961152, "nll": -5.969149112701416, "elapsed_s": 638.0}466{"event": "train", "step": 9140, "loss": -6.118272304534912, "grad_norm": 0.10226363688707352, "lr": 6.851208675456295e-05, "ex_per_s": 916.8151164457273, "data_frac": 0.06406103321204426, "mem_mb": 14.961152, "nll": -6.118272304534912, "elapsed_s": 638.0}467{"event": "train", "step": 9160, "loss": -5.982449054718018, "grad_norm": 0.06375555694103241, "lr": 6.800207660166508e-05, "ex_per_s": 918.6825598114953, "data_frac": 0.064118769539684, "mem_mb": 14.961152, "nll": -5.982449054718018, "elapsed_s": 638.1}468{"event": "train", "step": 9180, "loss": -5.954491138458252, "grad_norm": 0.07034506648778915, "lr": 6.749491225373478e-05, "ex_per_s": 920.585356019623, "data_frac": 0.06412569285934056, "mem_mb": 14.961152, "nll": -5.954491138458252, "elapsed_s": 638.2}469{"event": "train", "step": 9200, "loss": -6.140054225921631, "grad_norm": 0.0890924483537674, "lr": 6.699060859037737e-05, "ex_per_s": 922.4784167841177, "data_frac": 0.06414520093301182, "mem_mb": 14.961152, "nll": -6.140054225921631, "elapsed_s": 638.3}470{"event": "train", "step": 9220, "loss": -5.96195650100708, "grad_norm": 0.23874375224113464, "lr": 6.648918040726909e-05, "ex_per_s": 924.3808816790327, "data_frac": 0.06415826325771092, "mem_mb": 14.961152, "nll": -5.96195650100708, "elapsed_s": 638.4}471{"event": "train", "step": 9240, "loss": -6.109521865844727, "grad_norm": 0.14820478856563568, "lr": 6.599064241572299e-05, "ex_per_s": 926.2877918647741, "data_frac": 0.06416569521407844, "mem_mb": 14.961152, "nll": -6.109521865844727, "elapsed_s": 638.4}472{"event": "train", "step": 9260, "loss": -6.025632858276367, "grad_norm": 0.07190240174531937, "lr": 6.549500924225728e-05, "ex_per_s": 928.1899668612283, "data_frac": 0.06417597365112926, "mem_mb": 14.961152, "nll": -6.025632858276367, "elapsed_s": 638.5}473{"event": "train", "step": 9280, "loss": -5.814825057983398, "grad_norm": 0.12854187190532684, "lr": 6.500229542816624e-05, "ex_per_s": 930.0942755530283, "data_frac": 0.06418722774928708, "mem_mb": 14.961152, "nll": -5.814825057983398, "elapsed_s": 638.6}474{"event": "train", "step": 9300, "loss": -6.005171775817871, "grad_norm": 0.08099321275949478, "lr": 6.45125154290935e-05, "ex_per_s": 931.9971223366474, "data_frac": 0.06419987459333332, "mem_mb": 14.961152, "nll": -6.005171775817871, "elapsed_s": 638.6}475{"event": "train", "step": 9320, "loss": -5.821127891540527, "grad_norm": 0.08015812188386917, "lr": 6.402568361460816e-05, "ex_per_s": 933.9017461559786, "data_frac": 0.06421182502105142, "mem_mb": 14.961152, "nll": -5.821127891540527, "elapsed_s": 638.7}476{"event": "train", "step": 9340, "loss": -6.047384262084961, "grad_norm": 0.20655816793441772, "lr": 6.354181426778293e-05, "ex_per_s": 935.8038363535221, "data_frac": 0.06422406659922482, "mem_mb": 14.961152, "nll": -6.047384262084961, "elapsed_s": 638.8}477{"event": "train", "step": 9360, "loss": -6.057175636291504, "grad_norm": 0.19730782508850098, "lr": 6.306092158477523e-05, "ex_per_s": 937.7021167435233, "data_frac": 0.0642374937166958, "mem_mb": 14.961152, "nll": -6.057175636291504, "elapsed_s": 638.8}478{"event": "train", "step": 9380, "loss": -5.953861236572266, "grad_norm": 0.09222159534692764, "lr": 6.258301967441065e-05, "ex_per_s": 939.6026271461417, "data_frac": 0.06424991286062742, "mem_mb": 14.961152, "nll": -5.953861236572266, "elapsed_s": 638.9}479{"event": "train", "step": 9400, "loss": -6.068983554840088, "grad_norm": 0.16349394619464874, "lr": 6.210812255776909e-05, "ex_per_s": 941.5028869723892, "data_frac": 0.06426224475149102, "mem_mb": 14.961152, "nll": -6.068983554840088, "elapsed_s": 639.0}480{"event": "train", "step": 9420, "loss": -5.935633182525635, "grad_norm": 0.07973375171422958, "lr": 6.163624416777327e-05, "ex_per_s": 943.3999404357866, "data_frac": 0.06427649177033024, "mem_mb": 14.961152, "nll": -5.935633182525635, "elapsed_s": 639.1}481{"event": "train", "step": 9440, "loss": -5.9220404624938965, "grad_norm": 0.08637393265962601, "lr": 6.116739834878009e-05, "ex_per_s": 945.2993342390059, "data_frac": 0.06428861166660939, "mem_mb": 14.961152, "nll": -5.9220404624938965, "elapsed_s": 639.1}482{"event": "train", "step": 9460, "loss": -6.001434803009033, "grad_norm": 0.07421818375587463, "lr": 6.0701598856174347e-05, "ex_per_s": 947.1921919679282, "data_frac": 0.06430271668468686, "mem_mb": 14.961152, "nll": -6.001434803009033, "elapsed_s": 639.2}483{"event": "train", "step": 9480, "loss": -5.924988746643066, "grad_norm": 0.09442289918661118, "lr": 6.023885935596515e-05, "ex_per_s": 949.0869817973033, "data_frac": 0.06431659982539907, "mem_mb": 14.961152, "nll": -5.924988746643066, "elapsed_s": 639.3}484{"event": "train", "step": 9500, "loss": -5.9141387939453125, "grad_norm": 0.08575707674026489, "lr": 5.977919342438512e-05, "ex_per_s": 950.9848368565891, "data_frac": 0.064329720202147, "mem_mb": 14.961152, "nll": -5.9141387939453125, "elapsed_s": 639.3}485{"event": "train", "step": 9520, "loss": -5.988900661468506, "grad_norm": 0.08716656267642975, "lr": 5.9322614547491995e-05, "ex_per_s": 952.8828788226153, "data_frac": 0.06434138099578005, "mem_mb": 14.961152, "nll": -5.988900661468506, "elapsed_s": 639.4}486{"event": "train", "step": 9540, "loss": -6.0007476806640625, "grad_norm": 0.1315421462059021, "lr": 5.886913612077285e-05, "ex_per_s": 954.7777540535608, "data_frac": 0.06435444721768495, "mem_mb": 14.961152, "nll": -6.0007476806640625, "elapsed_s": 639.5}487{"event": "train", "step": 9560, "loss": -5.961147308349609, "grad_norm": 0.07103128731250763, "lr": 5.841877144875127e-05, "ex_per_s": 956.6765735959275, "data_frac": 0.06436663053583686, "mem_mb": 14.961152, "nll": -5.961147308349609, "elapsed_s": 639.5}488{"event": "train", "step": 9580, "loss": -5.906157970428467, "grad_norm": 0.09659575670957565, "lr": 5.7971533744597e-05, "ex_per_s": 958.5735233474287, "data_frac": 0.06437885657239358, "mem_mb": 14.961152, "nll": -5.906157970428467, "elapsed_s": 639.6}489{"event": "train", "step": 9600, "loss": -6.155018329620361, "grad_norm": 0.1009039431810379, "lr": 5.752743612973809e-05, "ex_per_s": 960.4696262171714, "data_frac": 0.06439139713037488, "mem_mb": 14.961152, "nll": -6.155018329620361, "elapsed_s": 639.7}490{"event": "train", "step": 9620, "loss": -6.022785663604736, "grad_norm": 0.09432908147573471, "lr": 5.7086491633476214e-05, "ex_per_s": 962.3661527481395, "data_frac": 0.06440411925877682, "mem_mb": 14.961152, "nll": -6.022785663604736, "elapsed_s": 639.8}491{"event": "train", "step": 9640, "loss": -5.998536586761475, "grad_norm": 0.09081383794546127, "lr": 5.6648713192604105e-05, "ex_per_s": 964.2634516228975, "data_frac": 0.06441649180129264, "mem_mb": 14.961152, "nll": -5.998536586761475, "elapsed_s": 639.8}492{"event": "train", "step": 9660, "loss": -6.050878524780273, "grad_norm": 0.09928731620311737, "lr": 5.621411365102632e-05, "ex_per_s": 966.159624552358, "data_frac": 0.06442902836233065, "mem_mb": 14.961152, "nll": -6.050878524780273, "elapsed_s": 639.9}493{"event": "train", "step": 9680, "loss": -5.92808198928833, "grad_norm": 0.07900100201368332, "lr": 5.578270575938211e-05, "ex_per_s": 968.0549370739419, "data_frac": 0.06444044212387198, "mem_mb": 14.961152, "nll": -5.92808198928833, "elapsed_s": 640.0}494{"event": "train", "step": 9700, "loss": -6.192510604858398, "grad_norm": 0.17853248119354248, "lr": 5.5354502174671515e-05, "ex_per_s": 969.9493048014255, "data_frac": 0.06445187866740784, "mem_mb": 14.961152, "nll": -6.192510604858398, "elapsed_s": 640.0}495{"event": "train", "step": 9720, "loss": -6.0374579429626465, "grad_norm": 0.18541286885738373, "lr": 5.492951545988398e-05, "ex_per_s": 971.844011037633, "data_frac": 0.06446355736524241, "mem_mb": 14.961152, "nll": -6.0374579429626465, "elapsed_s": 640.1}496{"event": "train", "step": 9740, "loss": -5.917126178741455, "grad_norm": 0.1312480866909027, "lr": 5.4507758083629804e-05, "ex_per_s": 973.7369218942864, "data_frac": 0.0644761510890841, "mem_mb": 14.961152, "nll": -5.917126178741455, "elapsed_s": 640.2}497{"event": "train", "step": 9760, "loss": -6.031574249267578, "grad_norm": 0.2296123206615448, "lr": 5.408924241977427e-05, "ex_per_s": 975.6259891687281, "data_frac": 0.06448909709220614, "mem_mb": 14.961152, "nll": -6.031574249267578, "elapsed_s": 640.2}498{"event": "train", "step": 9780, "loss": -6.054408550262451, "grad_norm": 0.07987184822559357, "lr": 5.367398074707458e-05, "ex_per_s": 977.5193257575863, "data_frac": 0.06450106735085123, "mem_mb": 14.961152, "nll": -6.054408550262451, "elapsed_s": 640.3}499{"event": "train", "step": 9800, "loss": -5.846637725830078, "grad_norm": 0.24454675614833832, "lr": 5.326198524881979e-05, "ex_per_s": 979.4120604150289, "data_frac": 0.06451268321241958, "mem_mb": 14.961152, "nll": -5.846637725830078, "elapsed_s": 640.4}500{"event": "train", "step": 9820, "loss": -6.093161582946777, "grad_norm": 0.11450838297605515, "lr": 5.285326801247307e-05, "ex_per_s": 981.3028380403333, "data_frac": 0.0645255238608557, "mem_mb": 14.961152, "nll": -6.093161582946777, "elapsed_s": 640.5}501{"event": "train", "step": 9840, "loss": -5.952985763549805, "grad_norm": 0.08263888210058212, "lr": 5.244784102931745e-05, "ex_per_s": 983.193958864598, "data_frac": 0.06453771584362115, "mem_mb": 14.961152, "nll": -5.952985763549805, "elapsed_s": 640.5}502{"event": "train", "step": 9860, "loss": -5.876458168029785, "grad_norm": 0.07881748676300049, "lr": 5.2045716194103555e-05, "ex_per_s": 985.0871166023388, "data_frac": 0.06454968720603968, "mem_mb": 14.961152, "nll": -5.876458168029785, "elapsed_s": 640.6}503{"event": "train", "step": 9880, "loss": -5.886932373046875, "grad_norm": 0.09037894755601883, "lr": 5.164690530470111e-05, "ex_per_s": 986.9768332418281, "data_frac": 0.06456166799658815, "mem_mb": 14.961152, "nll": -5.886932373046875, "elapsed_s": 640.7}504{"event": "train", "step": 9900, "loss": -5.93922758102417, "grad_norm": 0.07355788350105286, "lr": 5.125142006175238e-05, "ex_per_s": 988.8667106116592, "data_frac": 0.06457440117286793, "mem_mb": 14.961152, "nll": -5.93922758102417, "elapsed_s": 640.7}505{"event": "train", "step": 9920, "loss": -5.877810478210449, "grad_norm": 0.09550517052412033, "lr": 5.085927206832921e-05, "ex_per_s": 990.7557577849877, "data_frac": 0.0645864528845146, "mem_mb": 14.961152, "nll": -5.877810478210449, "elapsed_s": 640.8}506{"event": "train", "step": 9940, "loss": -5.954232215881348, "grad_norm": 0.096140056848526, "lr": 5.047047282959241e-05, "ex_per_s": 992.6441920032039, "data_frac": 0.06459939452281042, "mem_mb": 14.961152, "nll": -5.954232215881348, "elapsed_s": 640.9}507{"event": "train", "step": 9960, "loss": -6.130692005157471, "grad_norm": 0.2109522968530655, "lr": 5.0085033752454274e-05, "ex_per_s": 994.5337851188348, "data_frac": 0.0646111507398814, "mem_mb": 14.961152, "nll": -6.130692005157471, "elapsed_s": 640.9}508{"event": "train", "step": 9980, "loss": -5.805058479309082, "grad_norm": 0.30132344365119934, "lr": 4.970296614524392e-05, "ex_per_s": 996.4218417536891, "data_frac": 0.06462338399752554, "mem_mb": 14.961152, "nll": -5.805058479309082, "elapsed_s": 641.0}509{"event": "train", "step": 10000, "loss": -5.933992862701416, "grad_norm": 0.17210976779460907, "lr": 4.9324281217375474e-05, "ex_per_s": 998.3054682763664, "data_frac": 0.06463669741275056, "mem_mb": 14.961152, "nll": -5.933992862701416, "elapsed_s": 641.1}510{"event": "val", "step": 10000, "val_nll": -6.449462538869622, "best": -6.449462538869622, "bad": 0, "mem_mb": 14.961152, "elapsed_s": 704.6}511{"event": "train", "step": 10020, "loss": -6.071625232696533, "grad_norm": 0.1969016045331955, "lr": 4.8948990079019224e-05, "ex_per_s": 909.8478010720024, "data_frac": 0.05880821609074556, "mem_mb": 14.961152, "nll": -6.071625232696533, "elapsed_s": 704.8}512{"event": "train", "step": 10040, "loss": -6.07761287689209, "grad_norm": 0.1282041370868683, "lr": 4.857710374077574e-05, "ex_per_s": 911.5752628152769, "data_frac": 0.0588188827948049, "mem_mb": 14.961152, "nll": -6.07761287689209, "elapsed_s": 704.9}513{"event": "train", "step": 10060, "loss": -5.756709098815918, "grad_norm": 0.07184922695159912, "lr": 4.820863311335266e-05, "ex_per_s": 913.3022075049116, "data_frac": 0.05882932572294305, "mem_mb": 14.961152, "nll": -5.756709098815918, "elapsed_s": 705.0}514{"event": "train", "step": 10080, "loss": -6.177216053009033, "grad_norm": 0.07178782671689987, "lr": 4.78435890072448e-05, "ex_per_s": 915.0294103263012, "data_frac": 0.05883982580773889, "mem_mb": 14.961152, "nll": -6.177216053009033, "elapsed_s": 705.0}515{"event": "train", "step": 10100, "loss": -5.973994731903076, "grad_norm": 0.25569355487823486, "lr": 4.748198213241678e-05, "ex_per_s": 916.758467299568, "data_frac": 0.058849905052487994, "mem_mb": 14.961152, "nll": -5.973994731903076, "elapsed_s": 705.1}516{"event": "train", "step": 10120, "loss": -6.06099271774292, "grad_norm": 0.18646569550037384, "lr": 4.71238230979889e-05, "ex_per_s": 918.486482944748, "data_frac": 0.05886055800241384, "mem_mb": 14.961152, "nll": -6.06099271774292, "elapsed_s": 705.2}517{"event": "train", "step": 10140, "loss": -5.968737602233887, "grad_norm": 0.139174684882164, "lr": 4.676912241192593e-05, "ex_per_s": 920.2139808366322, "data_frac": 0.05887092636693221, "mem_mb": 14.961152, "nll": -5.968737602233887, "elapsed_s": 705.2}518{"event": "train", "step": 10160, "loss": -5.770279407501221, "grad_norm": 0.11582019925117493, "lr": 4.641789048072876e-05, "ex_per_s": 921.9406253418397, "data_frac": 0.058881122370080134, "mem_mb": 14.961152, "nll": -5.770279407501221, "elapsed_s": 705.3}519{"event": "train", "step": 10180, "loss": -6.009319305419922, "grad_norm": 0.13556210696697235, "lr": 4.6070137609129025e-05, "ex_per_s": 923.6672308072443, "data_frac": 0.05889116230836963, "mem_mb": 14.961152, "nll": -6.009319305419922, "elapsed_s": 705.4}520{"event": "train", "step": 10200, "loss": -5.959250450134277, "grad_norm": 0.07025151699781418, "lr": 4.57258739997869e-05, "ex_per_s": 925.3907663565014, "data_frac": 0.058902710729627966, "mem_mb": 14.961152, "nll": -5.959250450134277, "elapsed_s": 705.4}521{"event": "train", "step": 10220, "loss": -5.778154373168945, "grad_norm": 0.1350811868906021, "lr": 4.538510975299176e-05, "ex_per_s": 927.115457618153, "data_frac": 0.05891414033339047, "mem_mb": 14.961152, "nll": -5.778154373168945, "elapsed_s": 705.5}522{"event": "train", "step": 10240, "loss": -6.020912170410156, "grad_norm": 0.14443591237068176, "lr": 4.504785486636568e-05, "ex_per_s": 928.841255162342, "data_frac": 0.05892449330646207, "mem_mb": 14.961152, "nll": -6.020912170410156, "elapsed_s": 705.6}523{"event": "train", "step": 10260, "loss": -6.027403354644775, "grad_norm": 0.0799226313829422, "lr": 4.471411923457034e-05, "ex_per_s": 930.5656227412272, "data_frac": 0.05893482390225023, "mem_mb": 14.961152, "nll": -6.027403354644775, "elapsed_s": 705.6}524{"event": "train", "step": 10280, "loss": -6.060848712921143, "grad_norm": 0.12741905450820923, "lr": 4.438391264901654e-05, "ex_per_s": 932.2908992827189, "data_frac": 0.058944815795176096, "mem_mb": 14.961152, "nll": -6.060848712921143, "elapsed_s": 705.7}525{"event": "train", "step": 10300, "loss": -5.9885663986206055, "grad_norm": 0.09316706657409668, "lr": 4.405724479757712e-05, "ex_per_s": 934.0179824608327, "data_frac": 0.058955039517137854, "mem_mb": 14.961152, "nll": -5.9885663986206055, "elapsed_s": 705.8}526{"event": "train", "step": 10320, "loss": -5.900144100189209, "grad_norm": 0.13372132182121277, "lr": 4.373412526430252e-05, "ex_per_s": 935.7428150481338, "data_frac": 0.05896495124282667, "mem_mb": 14.961152, "nll": -5.900144100189209, "elapsed_s": 705.8}527{"event": "train", "step": 10340, "loss": -5.947429656982422, "grad_norm": 0.12956006824970245, "lr": 4.341456352913975e-05, "ex_per_s": 937.466444554363, "data_frac": 0.058975384220443704, "mem_mb": 14.961152, "nll": -5.947429656982422, "elapsed_s": 705.9}528{"event": "train", "step": 10360, "loss": -5.798385143280029, "grad_norm": 0.09540167450904846, "lr": 4.3098568967654146e-05, "ex_per_s": 939.18979361402, "data_frac": 0.05898616887754722, "mem_mb": 14.961152, "nll": -5.798385143280029, "elapsed_s": 706.0}529{"event": "train", "step": 10380, "loss": -5.9796552658081055, "grad_norm": 0.1424694061279297, "lr": 4.27861508507545e-05, "ex_per_s": 940.9136413446431, "data_frac": 0.058996324341617626, "mem_mb": 14.961152, "nll": -5.9796552658081055, "elapsed_s": 706.0}530{"event": "train", "step": 10400, "loss": -5.97393274307251, "grad_norm": 0.11834895610809326, "lr": 4.247731834442082e-05, "ex_per_s": 942.6371168744832, "data_frac": 0.059006728627895924, "mem_mb": 14.961152, "nll": -5.97393274307251, "elapsed_s": 706.1}531{"event": "train", "step": 10420, "loss": -6.036853313446045, "grad_norm": 0.15191765129566193, "lr": 4.2172080509435485e-05, "ex_per_s": 944.360567274476, "data_frac": 0.05901706945519345, "mem_mb": 14.961152, "nll": -6.036853313446045, "elapsed_s": 706.2}532{"event": "train", "step": 10440, "loss": -6.124643325805664, "grad_norm": 0.13624662160873413, "lr": 4.1870446301117574e-05, "ex_per_s": 946.0838099176809, "data_frac": 0.059027323875484344, "mem_mb": 14.961152, "nll": -6.124643325805664, "elapsed_s": 706.2}533{"event": "train", "step": 10460, "loss": -5.967655658721924, "grad_norm": 0.06849592179059982, "lr": 4.157242456905989e-05, "ex_per_s": 947.8035370238499, "data_frac": 0.059037675039346656, "mem_mb": 14.961152, "nll": -5.967655658721924, "elapsed_s": 706.3}534{"event": "train", "step": 10480, "loss": -5.964616775512695, "grad_norm": 0.2083885222673416, "lr": 4.1278024056869486e-05, "ex_per_s": 949.52450461777, "data_frac": 0.059047695267029875, "mem_mb": 14.961152, "nll": -5.964616775512695, "elapsed_s": 706.4}535{"event": "train", "step": 10500, "loss": -5.9944610595703125, "grad_norm": 0.10666468739509583, "lr": 4.0987253401911067e-05, "ex_per_s": 951.2449102606755, "data_frac": 0.059058752192230855, "mem_mb": 14.961152, "nll": -5.9944610595703125, "elapsed_s": 706.4}536{"event": "train", "step": 10520, "loss": -6.004821300506592, "grad_norm": 0.13664610683918, "lr": 4.070012113505359e-05, "ex_per_s": 952.9603358883357, "data_frac": 0.05907114910500573, "mem_mb": 14.961152, "nll": -6.004821300506592, "elapsed_s": 706.5}537{"event": "train", "step": 10540, "loss": -5.7752556800842285, "grad_norm": 0.15749676525592804, "lr": 4.041663568041998e-05, "ex_per_s": 954.6737350166537, "data_frac": 0.05908339132982528, "mem_mb": 14.961152, "nll": -5.7752556800842285, "elapsed_s": 706.6}538{"event": "train", "step": 10560, "loss": -5.866751194000244, "grad_norm": 0.19282737374305725, "lr": 4.013680535514001e-05, "ex_per_s": 956.387551547432, "data_frac": 0.05909544785253254, "mem_mb": 14.961152, "nll": -5.866751194000244, "elapsed_s": 706.7}539{"event": "train", "step": 10580, "loss": -5.865234375, "grad_norm": 0.11533369868993759, "lr": 3.986063836910623e-05, "ex_per_s": 958.1050905466566, "data_frac": 0.05910608088870761, "mem_mb": 14.961152, "nll": -5.865234375, "elapsed_s": 706.7}540{"event": "train", "step": 10600, "loss": -5.893553256988525, "grad_norm": 0.12264243513345718, "lr": 3.958814282473316e-05, "ex_per_s": 959.824825516226, "data_frac": 0.05911591452986879, "mem_mb": 14.961152, "nll": -5.893553256988525, "elapsed_s": 706.8}541{"event": "train", "step": 10620, "loss": -6.041429042816162, "grad_norm": 0.06406545639038086, "lr": 3.9319326716719496e-05, "ex_per_s": 961.5446859983924, "data_frac": 0.059126021054990825, "mem_mb": 14.961152, "nll": -6.041429042816162, "elapsed_s": 706.9}542{"event": "train", "step": 10640, "loss": -5.816895961761475, "grad_norm": 0.17188169062137604, "lr": 3.905419793181364e-05, "ex_per_s": 963.2619233961508, "data_frac": 0.059136547524054714, "mem_mb": 14.961152, "nll": -5.816895961761475, "elapsed_s": 706.9}543{"event": "train", "step": 10660, "loss": -5.848128795623779, "grad_norm": 0.10115062445402145, "lr": 3.87927642485822e-05, "ex_per_s": 964.979247635597, "data_frac": 0.059147463608423115, "mem_mb": 14.961152, "nll": -5.848128795623779, "elapsed_s": 707.0}544{"event": "train", "step": 10680, "loss": -5.950589179992676, "grad_norm": 0.12784157693386078, "lr": 3.853503333718199e-05, "ex_per_s": 966.6980606263885, "data_frac": 0.05915744505824469, "mem_mb": 14.961152, "nll": -5.950589179992676, "elapsed_s": 707.1}545{"event": "train", "step": 10700, "loss": -5.799734592437744, "grad_norm": 0.21147777140140533, "lr": 3.8281012759134686e-05, "ex_per_s": 968.4153866709911, "data_frac": 0.05916808396901803, "mem_mb": 14.961152, "nll": -5.799734592437744, "elapsed_s": 707.1}546{"event": "train", "step": 10720, "loss": -5.879059791564941, "grad_norm": 0.12701421976089478, "lr": 3.803070996710526e-05, "ex_per_s": 970.1316337512222, "data_frac": 0.059178298894183044, "mem_mb": 14.961152, "nll": -5.879059791564941, "elapsed_s": 707.2}547{"event": "train", "step": 10740, "loss": -6.095066070556641, "grad_norm": 0.16681446135044098, "lr": 3.778413230468319e-05, "ex_per_s": 971.8480037944252, "data_frac": 0.0591884158420873, "mem_mb": 14.961152, "nll": -6.095066070556641, "elapsed_s": 707.3}548{"event": "train", "step": 10760, "loss": -5.964556694030762, "grad_norm": 0.07930019497871399, "lr": 3.754128700616701e-05, "ex_per_s": 973.5647810599203, "data_frac": 0.059199111293137194, "mem_mb": 14.961152, "nll": -5.964556694030762, "elapsed_s": 707.3}549{"event": "train", "step": 10780, "loss": -6.016637802124023, "grad_norm": 0.13120895624160767, "lr": 3.7302181196352115e-05, "ex_per_s": 975.2786261211807, "data_frac": 0.05920947740383693, "mem_mb": 14.961152, "nll": -6.016637802124023, "elapsed_s": 707.4}550{"event": "train", "step": 10800, "loss": -5.930477142333984, "grad_norm": 0.08734068274497986, "lr": 3.7066821890321684e-05, "ex_per_s": 976.9955578775374, "data_frac": 0.059219804311313365, "mem_mb": 14.961152, "nll": -5.930477142333984, "elapsed_s": 707.5}551{"event": "train", "step": 10820, "loss": -6.01542329788208, "grad_norm": 0.0942450761795044, "lr": 3.683521599324089e-05, "ex_per_s": 978.7117846946322, "data_frac": 0.05922980974040558, "mem_mb": 14.961152, "nll": -6.01542329788208, "elapsed_s": 707.5}552{"event": "train", "step": 10840, "loss": -5.958430290222168, "grad_norm": 0.16879256069660187, "lr": 3.660737030015427e-05, "ex_per_s": 980.4288141023972, "data_frac": 0.05923958732642035, "mem_mb": 14.961152, "nll": -5.958430290222168, "elapsed_s": 707.6}553{"event": "train", "step": 10860, "loss": -5.840099811553955, "grad_norm": 0.0701942965388298, "lr": 3.638329149578648e-05, "ex_per_s": 982.1454341329345, "data_frac": 0.05924916646552519, "mem_mb": 14.961152, "nll": -5.840099811553955, "elapsed_s": 707.7}554{"event": "train", "step": 10880, "loss": -5.998306751251221, "grad_norm": 0.08829799294471741, "lr": 3.6162986154346006e-05, "ex_per_s": 983.8616229020421, "data_frac": 0.05925962525225163, "mem_mb": 14.961152, "nll": -5.998306751251221, "elapsed_s": 707.7}555{"event": "train", "step": 10900, "loss": -6.1187262535095215, "grad_norm": 0.09664370864629745, "lr": 3.594646073933244e-05, "ex_per_s": 985.5767651315222, "data_frac": 0.059269790971293146, "mem_mb": 14.961152, "nll": -6.1187262535095215, "elapsed_s": 707.8}556{"event": "train", "step": 10920, "loss": -5.8950910568237305, "grad_norm": 0.13074186444282532, "lr": 3.5733721603346695e-05, "ex_per_s": 987.2904795897256, "data_frac": 0.059280406567003945, "mem_mb": 14.961152, "nll": -5.8950910568237305, "elapsed_s": 707.9}557{"event": "train", "step": 10940, "loss": -5.8964033126831055, "grad_norm": 0.07985105365514755, "lr": 3.552477498790482e-05, "ex_per_s": 989.0037151974984, "data_frac": 0.059290548054961686, "mem_mb": 14.961152, "nll": -5.8964033126831055, "elapsed_s": 707.9}558{"event": "train", "step": 10960, "loss": -6.126322269439697, "grad_norm": 0.1398288756608963, "lr": 3.531962702325468e-05, "ex_per_s": 990.7182841909835, "data_frac": 0.05930056586072222, "mem_mb": 14.961152, "nll": -6.126322269439697, "elapsed_s": 708.0}559{"event": "train", "step": 10980, "loss": -5.895016670227051, "grad_norm": 0.08947473019361496, "lr": 3.511828372819624e-05, "ex_per_s": 992.4313520879056, "data_frac": 0.059311007376434965, "mem_mb": 14.961152, "nll": -5.895016670227051, "elapsed_s": 708.1}560{"event": "train", "step": 11000, "loss": -5.918903827667236, "grad_norm": 0.1050722673535347, "lr": 3.4920751009904904e-05, "ex_per_s": 994.1453619449978, "data_frac": 0.05932117921898912, "mem_mb": 14.961152, "nll": -5.918903827667236, "elapsed_s": 708.1}561{"event": "val", "step": 11000, "val_nll": -6.449690639457265, "best": -6.449690639457265, "bad": 0, "mem_mb": 14.961152, "elapsed_s": 771.6}562{"event": "train", "step": 11020, "loss": -5.795467853546143, "grad_norm": 0.09662161767482758, "lr": 3.472703466375828e-05, "ex_per_s": 913.5818984419958, "data_frac": 0.054432275654962406, "mem_mb": 14.961152, "nll": -5.795467853546143, "elapsed_s": 772.0}563{"event": "train", "step": 11040, "loss": -5.98204231262207, "grad_norm": 0.13620322942733765, "lr": 3.4537140373166085e-05, "ex_per_s": 915.1555372387522, "data_frac": 0.054443639553188546, "mem_mb": 14.961152, "nll": -5.98204231262207, "elapsed_s": 772.1}564{"event": "train", "step": 11060, "loss": -5.982024192810059, "grad_norm": 0.10576959699392319, "lr": 3.4351073709403385e-05, "ex_per_s": 916.7326763611696, "data_frac": 0.05445338002117385, "mem_mb": 14.961152, "nll": -5.982024192810059, "elapsed_s": 772.1}565{"event": "train", "step": 11080, "loss": -6.0222296714782715, "grad_norm": 0.11565357446670532, "lr": 3.4168840131447255e-05, "ex_per_s": 918.3089564491668, "data_frac": 0.05446380523646369, "mem_mb": 14.961152, "nll": -6.0222296714782715, "elapsed_s": 772.2}566{"event": "train", "step": 11100, "loss": -6.142801761627197, "grad_norm": 0.1591458022594452, "lr": 3.399044498581645e-05, "ex_per_s": 919.8845516645525, "data_frac": 0.05447412618792466, "mem_mb": 14.961152, "nll": -6.142801761627197, "elapsed_s": 772.3}567{"event": "train", "step": 11120, "loss": -5.906047344207764, "grad_norm": 0.19413793087005615, "lr": 3.3815893506414745e-05, "ex_per_s": 921.4614305060679, "data_frac": 0.054484588431465183, "mem_mb": 14.961152, "nll": -5.906047344207764, "elapsed_s": 772.3}568{"event": "train", "step": 11140, "loss": -5.880709648132324, "grad_norm": 0.11368779093027115, "lr": 3.364519081437718e-05, "ex_per_s": 923.0343696550314, "data_frac": 0.05449524737957965, "mem_mb": 14.961152, "nll": -5.880709648132324, "elapsed_s": 772.4}569{"event": "train", "step": 11160, "loss": -6.0770263671875, "grad_norm": 0.06606385111808777, "lr": 3.347834191791993e-05, "ex_per_s": 924.6100048890687, "data_frac": 0.05450515371967773, "mem_mb": 14.961152, "nll": -6.0770263671875, "elapsed_s": 772.5}570{"event": "train", "step": 11180, "loss": -6.061487674713135, "grad_norm": 0.12614141404628754, "lr": 3.331535171219332e-05, "ex_per_s": 926.1850067634663, "data_frac": 0.054515658933315496, "mem_mb": 14.961152, "nll": -6.061487674713135, "elapsed_s": 772.5}571{"event": "train", "step": 11200, "loss": -6.179257869720459, "grad_norm": 0.09712888300418854, "lr": 3.3156224979138305e-05, "ex_per_s": 927.7613854274518, "data_frac": 0.054526059010400425, "mem_mb": 14.961152, "nll": -6.179257869720459, "elapsed_s": 772.6}572{"event": "train", "step": 11220, "loss": -5.979396820068359, "grad_norm": 0.09579291194677353, "lr": 3.300096638734598e-05, "ex_per_s": 929.3362326405406, "data_frac": 0.05453628040074758, "mem_mb": 14.961152, "nll": -5.979396820068359, "elapsed_s": 772.7}573{"event": "train", "step": 11240, "loss": -5.964810848236084, "grad_norm": 0.12609966099262238, "lr": 3.2849580491920845e-05, "ex_per_s": 930.9117289938478, "data_frac": 0.05454628740075869, "mem_mb": 14.961152, "nll": -5.964810848236084, "elapsed_s": 772.7}574{"event": "train", "step": 11260, "loss": -6.072934627532959, "grad_norm": 4.501899719238281, "lr": 3.2702071734346924e-05, "ex_per_s": 932.4865732487699, "data_frac": 0.05455689405059546, "mem_mb": 14.961152, "nll": -6.072934627532959, "elapsed_s": 772.8}575{"event": "train", "step": 11280, "loss": -6.10233211517334, "grad_norm": 0.11552902311086655, "lr": 3.255844444235771e-05, "ex_per_s": 934.0610065636506, "data_frac": 0.05456716641293783, "mem_mb": 14.961152, "nll": -6.10233211517334, "elapsed_s": 772.9}576{"event": "train", "step": 11300, "loss": -5.935162544250488, "grad_norm": 0.0904155895113945, "lr": 3.2418702829808946e-05, "ex_per_s": 935.6361746778568, "data_frac": 0.054577365145995704, "mem_mb": 14.961152, "nll": -5.935162544250488, "elapsed_s": 773.0}577{"event": "train", "step": 11320, "loss": -5.855699062347412, "grad_norm": 0.13063247501850128, "lr": 3.228285099655516e-05, "ex_per_s": 937.2103923682872, "data_frac": 0.0545873872343148, "mem_mb": 14.961152, "nll": -5.855699062347412, "elapsed_s": 773.0}578{"event": "train", "step": 11340, "loss": -5.923004150390625, "grad_norm": 0.1168212965130806, "lr": 3.215089292832933e-05, "ex_per_s": 938.7839737819736, "data_frac": 0.05459773565110226, "mem_mb": 14.961152, "nll": -5.923004150390625, "elapsed_s": 773.1}579{"event": "train", "step": 11360, "loss": -5.951563835144043, "grad_norm": 0.10074255615472794, "lr": 3.202283249662597e-05, "ex_per_s": 940.3579153782982, "data_frac": 0.054607942794961256, "mem_mb": 14.961152, "nll": -5.951563835144043, "elapsed_s": 773.2}580{"event": "train", "step": 11380, "loss": -5.907245635986328, "grad_norm": 0.10432118922472, "lr": 3.189867345858747e-05, "ex_per_s": 941.9301802066277, "data_frac": 0.05461812847317051, "mem_mb": 14.961152, "nll": -5.907245635986328, "elapsed_s": 773.2}581{"event": "train", "step": 11400, "loss": -6.068364143371582, "grad_norm": 0.12911184132099152, "lr": 3.177841945689397e-05, "ex_per_s": 943.503103997503, "data_frac": 0.054627864043160224, "mem_mb": 14.961152, "nll": -6.068364143371582, "elapsed_s": 773.3}582{"event": "train", "step": 11420, "loss": -5.828389644622803, "grad_norm": 0.10224208235740662, "lr": 3.1662074019656414e-05, "ex_per_s": 945.075198926529, "data_frac": 0.054637723734987556, "mem_mb": 14.961152, "nll": -5.828389644622803, "elapsed_s": 773.4}583{"event": "train", "step": 11440, "loss": -5.8841552734375, "grad_norm": 0.1761786937713623, "lr": 3.154964056031306e-05, "ex_per_s": 946.6464948403453, "data_frac": 0.05464827287389562, "mem_mb": 14.961152, "nll": -5.8841552734375, "elapsed_s": 773.4}584{"event": "train", "step": 11460, "loss": -6.136636734008789, "grad_norm": 0.18700498342514038, "lr": 3.1441122377529333e-05, "ex_per_s": 948.2167730672236, "data_frac": 0.05465881080936713, "mem_mb": 14.961152, "nll": -6.136636734008789, "elapsed_s": 773.5}585{"event": "train", "step": 11480, "loss": -5.959365367889404, "grad_norm": 0.11776512116193771, "lr": 3.133652265510106e-05, "ex_per_s": 949.7881156318456, "data_frac": 0.05466909695183839, "mem_mb": 14.961152, "nll": -5.959365367889404, "elapsed_s": 773.6}586{"event": "train", "step": 11500, "loss": -5.936694622039795, "grad_norm": 0.18018895387649536, "lr": 3.123584446186108e-05, "ex_per_s": 951.3583686918072, "data_frac": 0.05467933947209315, "mem_mb": 14.961152, "nll": -5.936694622039795, "elapsed_s": 773.6}587{"event": "train", "step": 11520, "loss": -5.906948089599609, "grad_norm": 0.20399193465709686, "lr": 3.113909075158913e-05, "ex_per_s": 952.9281968541246, "data_frac": 0.05469022321955842, "mem_mb": 14.961152, "nll": -5.906948089599609, "elapsed_s": 773.7}588{"event": "train", "step": 11540, "loss": -6.04364013671875, "grad_norm": 0.06573447585105896, "lr": 3.104626436292525e-05, "ex_per_s": 954.498214779735, "data_frac": 0.05470073996181909, "mem_mb": 14.961152, "nll": -6.04364013671875, "elapsed_s": 773.8}589{"event": "train", "step": 11560, "loss": -6.117554664611816, "grad_norm": 0.07983458787202835, "lr": 3.095736801928647e-05, "ex_per_s": 956.0669345325776, "data_frac": 0.054711188426855295, "mem_mb": 14.961152, "nll": -6.117554664611816, "elapsed_s": 773.8}590{"event": "train", "step": 11580, "loss": -5.952788352966309, "grad_norm": 0.10128949582576752, "lr": 3.087240432878696e-05, "ex_per_s": 957.636915460242, "data_frac": 0.05472102164365519, "mem_mb": 14.961152, "nll": -5.952788352966309, "elapsed_s": 773.9}591{"event": "train", "step": 11600, "loss": -6.067084312438965, "grad_norm": 0.10234636068344116, "lr": 3.0791375784161455e-05, "ex_per_s": 959.2053314648575, "data_frac": 0.05473095700888367, "mem_mb": 14.961152, "nll": -6.067084312438965, "elapsed_s": 774.0}592{"event": "train", "step": 11620, "loss": -6.018622875213623, "grad_norm": 0.10488464683294296, "lr": 3.0714284762692116e-05, "ex_per_s": 960.7750623025768, "data_frac": 0.054740587803823244, "mem_mb": 14.961152, "nll": -6.018622875213623, "elapsed_s": 774.0}593{"event": "train", "step": 11640, "loss": -5.955671310424805, "grad_norm": 0.11232548952102661, "lr": 3.0641133526138836e-05, "ex_per_s": 962.3449989272568, "data_frac": 0.05475118331218456, "mem_mb": 14.961152, "nll": -5.955671310424805, "elapsed_s": 774.1}594{"event": "train", "step": 11660, "loss": -5.917551040649414, "grad_norm": 0.08978218585252762, "lr": 3.0571924220672835e-05, "ex_per_s": 963.9138084004014, "data_frac": 0.054761446362812556, "mem_mb": 14.961152, "nll": -5.917551040649414, "elapsed_s": 774.2}595{"event": "train", "step": 11680, "loss": -5.962039470672607, "grad_norm": 0.11944154649972916, "lr": 3.0506658876813734e-05, "ex_per_s": 965.4803359753778, "data_frac": 0.05477195254261217, "mem_mb": 14.961152, "nll": -5.962039470672607, "elapsed_s": 774.2}596{"event": "train", "step": 11700, "loss": -6.046535015106201, "grad_norm": 0.05619003623723984, "lr": 3.0445339409369934e-05, "ex_per_s": 967.0467393120031, "data_frac": 0.05478283401454873, "mem_mb": 14.961152, "nll": -6.046535015106201, "elapsed_s": 774.3}597{"event": "train", "step": 11720, "loss": -5.7561750411987305, "grad_norm": 0.07605654001235962, "lr": 3.0387967617382475e-05, "ex_per_s": 968.6129363046317, "data_frac": 0.05479390772995423, "mem_mb": 14.961152, "nll": -5.7561750411987305, "elapsed_s": 774.4}598{"event": "train", "step": 11740, "loss": -6.041703224182129, "grad_norm": 0.18512314558029175, "lr": 3.0334545184072252e-05, "ex_per_s": 970.1754887638767, "data_frac": 0.05480592542348735, "mem_mb": 14.961152, "nll": -6.041703224182129, "elapsed_s": 774.5}599{"event": "train", "step": 11760, "loss": -5.895691871643066, "grad_norm": 0.09545912593603134, "lr": 3.0285073676790606e-05, "ex_per_s": 971.7383938925196, "data_frac": 0.0548175308475147, "mem_mb": 14.961152, "nll": -5.895691871643066, "elapsed_s": 774.5}600{"event": "train", "step": 11780, "loss": -6.078613758087158, "grad_norm": 0.07746712118387222, "lr": 3.0239554546973377e-05, "ex_per_s": 973.3042844002327, "data_frac": 0.054828799329959356, "mem_mb": 14.961152, "nll": -6.078613758087158, "elapsed_s": 774.6}601{"event": "train", "step": 11800, "loss": -5.962878227233887, "grad_norm": 0.17389607429504395, "lr": 3.0197989130098285e-05, "ex_per_s": 974.8662693963379, "data_frac": 0.05484085804448142, "mem_mb": 14.961152, "nll": -5.962878227233887, "elapsed_s": 774.7}602{"event": "train", "step": 11820, "loss": -5.9688215255737305, "grad_norm": 0.2230576127767563, "lr": 3.0160378645645744e-05, "ex_per_s": 976.4289837358155, "data_frac": 0.05485232614170246, "mem_mb": 14.961152, "nll": -5.9688215255737305, "elapsed_s": 774.7}603{"event": "train", "step": 11840, "loss": -6.027766227722168, "grad_norm": 0.07063502073287964, "lr": 3.0126724197063143e-05, "ex_per_s": 977.991005057481, "data_frac": 0.05486483279345893, "mem_mb": 14.961152, "nll": -6.027766227722168, "elapsed_s": 774.8}604{"event": "train", "step": 11860, "loss": -5.983198642730713, "grad_norm": 0.18655037879943848, "lr": 3.0097026771732407e-05, "ex_per_s": 979.549383405606, "data_frac": 0.054878041456160744, "mem_mb": 14.961152, "nll": -5.983198642730713, "elapsed_s": 774.9}605{"event": "train", "step": 11880, "loss": -6.024130821228027, "grad_norm": 0.23580080270767212, "lr": 3.0071287240941065e-05, "ex_per_s": 981.1115813022985, "data_frac": 0.054889651109030366, "mem_mb": 14.961152, "nll": -6.024130821228027, "elapsed_s": 775.0}606{"event": "train", "step": 11900, "loss": -5.882590293884277, "grad_norm": 0.06675788760185242, "lr": 3.0049506359856666e-05, "ex_per_s": 982.6694777583633, "data_frac": 0.054902269074521615, "mem_mb": 14.961152, "nll": -5.882590293884277, "elapsed_s": 775.0}607{"event": "train", "step": 11920, "loss": -6.001864910125732, "grad_norm": 0.12144723534584045, "lr": 3.0031684767504643e-05, "ex_per_s": 984.2294515585753, "data_frac": 0.05491423183723371, "mem_mb": 14.961152, "nll": -6.001864910125732, "elapsed_s": 775.1}608{"event": "train", "step": 11940, "loss": -6.06478214263916, "grad_norm": 0.3318183124065399, "lr": 3.0017822986749544e-05, "ex_per_s": 985.7898970693742, "data_frac": 0.05492647811345724, "mem_mb": 14.961152, "nll": -6.06478214263916, "elapsed_s": 775.2}609{"event": "train", "step": 11960, "loss": -6.154135227203369, "grad_norm": 0.12007695436477661, "lr": 3.000792142427973e-05, "ex_per_s": 987.3518234648474, "data_frac": 0.05493762975580761, "mem_mb": 14.961152, "nll": -6.154135227203369, "elapsed_s": 775.2}610{"event": "train", "step": 11980, "loss": -6.091974258422852, "grad_norm": 0.06267210841178894, "lr": 3.000198037059536e-05, "ex_per_s": 988.9092941568381, "data_frac": 0.05494985318813909, "mem_mb": 14.961152, "nll": -6.091974258422852, "elapsed_s": 775.3}611{"event": "train", "step": 12000, "loss": -5.936828136444092, "grad_norm": 0.10667138546705246, "lr": 2.9999999999999997e-05, "ex_per_s": 990.4679991627519, "data_frac": 0.05496178526281797, "mem_mb": 14.961152, "nll": -5.936828136444092, "elapsed_s": 775.4}612{"event": "val", "step": 12000, "val_nll": -6.449677905491562, "best": -6.449690639457265, "bad": 1, "mem_mb": 14.961152, "elapsed_s": 838.6}613{"event": "done", "steps": 12000, "best_val": -6.449690639457265, "wall_s": 838.5925340652466, "ex_per_s": 915.8202211471702, "peak_mem_mb": 14.961152, "n_params": 301827, "device": "mps", "data_time_frac": 0.05081952609936494, "elapsed_s": 838.6}614
Files · a19bfab00d86abc6a1086d679b6b076e598d824b
Menu

Theme