{"event": "train", "step": 20, "loss": -6.110328674316406, "grad_norm": 1.1605756282806396, "lr": 3.149999999999999e-05, "ex_per_s": 91.2678943172381, "data_frac": 0.8718999725841483, "mem_mb": 90.587136, "nll": -6.110328674316406, "elapsed_s": 14.1} {"event": "train", "step": 40, "loss": -5.896738052368164, "grad_norm": 0.3131726384162903, "lr": 6.149999999999999e-05, "ex_per_s": 162.3737430002774, "data_frac": 0.7817950360029373, "mem_mb": 90.587136, "nll": -5.896738052368164, "elapsed_s": 15.8} {"event": "train", "step": 60, "loss": -5.976347923278809, "grad_norm": 0.5750947594642639, "lr": 9.149999999999999e-05, "ex_per_s": 220.11086415230054, "data_frac": 0.7072527746513882, "mem_mb": 90.587136, "nll": -5.976347923278809, "elapsed_s": 17.5} {"event": "train", "step": 80, "loss": -6.067798137664795, "grad_norm": 1.1805647611618042, "lr": 0.0001215, "ex_per_s": 267.6576562338672, "data_frac": 0.6456293583660562, "mem_mb": 90.587136, "nll": -6.067798137664795, "elapsed_s": 19.2} {"event": "train", "step": 100, "loss": -5.93828821182251, "grad_norm": 0.1374814212322235, "lr": 0.0001515, "ex_per_s": 307.6837735314187, "data_frac": 0.5943194589451327, "mem_mb": 90.587136, "nll": -5.93828821182251, "elapsed_s": 20.8} {"event": "train", "step": 120, "loss": -5.890838623046875, "grad_norm": 1.1565868854522705, "lr": 0.00018149999999999997, "ex_per_s": 341.9826927889891, "data_frac": 0.5509881100484214, "mem_mb": 90.587136, "nll": -5.890838623046875, "elapsed_s": 22.5} {"event": "train", "step": 140, "loss": -5.8151350021362305, "grad_norm": 0.5689684748649597, "lr": 0.00021149999999999996, "ex_per_s": 371.5878284238493, "data_frac": 0.5136239113858334, "mem_mb": 90.587136, "nll": -5.8151350021362305, "elapsed_s": 24.2} {"event": "train", "step": 160, "loss": -5.726049423217773, "grad_norm": 0.45477351546287537, "lr": 0.0002415, "ex_per_s": 397.40319103590735, "data_frac": 0.48106156740176453, "mem_mb": 90.587136, "nll": -5.726049423217773, "elapsed_s": 25.8} {"event": "train", "step": 180, "loss": -5.968119144439697, "grad_norm": 1.0592626333236694, "lr": 0.0002715, "ex_per_s": 420.16158693110407, "data_frac": 0.4524910855499809, "mem_mb": 90.587136, "nll": -5.968119144439697, "elapsed_s": 27.5} {"event": "train", "step": 200, "loss": -5.9680047035217285, "grad_norm": 0.6767856478691101, "lr": 0.0003, "ex_per_s": 440.29457721401565, "data_frac": 0.42711506331736415, "mem_mb": 90.587136, "nll": -5.9680047035217285, "elapsed_s": 29.1} {"event": "train", "step": 220, "loss": -6.034630298614502, "grad_norm": 0.7692214250564575, "lr": 0.00029998154617398326, "ex_per_s": 458.2977074133762, "data_frac": 0.40449758916788414, "mem_mb": 90.587136, "nll": -6.034630298614502, "elapsed_s": 30.8} {"event": "train", "step": 240, "loss": -5.9145989418029785, "grad_norm": 0.36909911036491394, "lr": 0.000299926189741025, "ex_per_s": 474.4460487401382, "data_frac": 0.38418912060598764, "mem_mb": 90.587136, "nll": -5.9145989418029785, "elapsed_s": 32.4} {"event": "train", "step": 260, "loss": -5.884219646453857, "grad_norm": 0.1772347241640091, "lr": 0.0002998339458350211, "ex_per_s": 488.99099053753935, "data_frac": 0.3658553373101739, "mem_mb": 90.587136, "nll": -5.884219646453857, "elapsed_s": 34.1} {"event": "train", "step": 280, "loss": -5.943273544311523, "grad_norm": 0.7572216391563416, "lr": 0.0002997048396745345, "ex_per_s": 502.2430811542127, "data_frac": 0.34923440966468383, "mem_mb": 90.587136, "nll": -5.943273544311523, "elapsed_s": 35.7} {"event": "train", "step": 300, "loss": -5.967167377471924, "grad_norm": 0.7032046318054199, "lr": 0.0002995389065559004, "ex_per_s": 514.3423830158322, "data_frac": 0.33408160946595056, "mem_mb": 90.587136, "nll": -5.967167377471924, "elapsed_s": 37.4} {"event": "train", "step": 320, "loss": -5.965827941894531, "grad_norm": 0.43770831823349, "lr": 0.0002993361918435765, "ex_per_s": 525.3637161985124, "data_frac": 0.3201804559209544, "mem_mb": 90.587136, "nll": -5.965827941894531, "elapsed_s": 39.0} {"event": "train", "step": 340, "loss": -5.835608005523682, "grad_norm": 0.211174875497818, "lr": 0.00029909675095774114, "ex_per_s": 535.5549432431397, "data_frac": 0.3074680523988922, "mem_mb": 90.587136, "nll": -5.835608005523682, "elapsed_s": 40.7} {"event": "train", "step": 360, "loss": -5.996823310852051, "grad_norm": 0.10928906500339508, "lr": 0.00029882064935914164, "ex_per_s": 544.9306820056707, "data_frac": 0.2957234326443547, "mem_mb": 90.587136, "nll": -5.996823310852051, "elapsed_s": 42.3} {"event": "train", "step": 380, "loss": -6.005190372467041, "grad_norm": 1.2386358976364136, "lr": 0.00029850796253119824, "ex_per_s": 553.6022161656806, "data_frac": 0.28485865643913594, "mem_mb": 90.587136, "nll": -6.005190372467041, "elapsed_s": 44.0} {"event": "train", "step": 400, "loss": -5.997462272644043, "grad_norm": 1.1981183290481567, "lr": 0.0002981587759593675, "ex_per_s": 561.6479612627916, "data_frac": 0.2747842264178428, "mem_mb": 90.587136, "nll": -5.997462272644043, "elapsed_s": 45.6} {"event": "train", "step": 420, "loss": -5.960166931152344, "grad_norm": 0.9327693581581116, "lr": 0.0002977731851077716, "ex_per_s": 569.1293707020199, "data_frac": 0.26541442368441454, "mem_mb": 90.587136, "nll": -5.960166931152344, "elapsed_s": 47.3} {"event": "train", "step": 440, "loss": -5.940631866455078, "grad_norm": 0.4679005742073059, "lr": 0.0002973512953930992, "ex_per_s": 576.1030027578039, "data_frac": 0.256673236815073, "mem_mb": 90.587136, "nll": -5.940631866455078, "elapsed_s": 48.9} {"event": "train", "step": 460, "loss": -6.0988874435424805, "grad_norm": 1.1395189762115479, "lr": 0.0002968932221557859, "ex_per_s": 582.6041112657258, "data_frac": 0.24850046934003306, "mem_mb": 90.587136, "nll": -6.0988874435424805, "elapsed_s": 50.6} {"event": "train", "step": 480, "loss": -5.875259876251221, "grad_norm": 0.339683473110199, "lr": 0.0002963990906284808, "ex_per_s": 588.7055690106757, "data_frac": 0.24085053885862368, "mem_mb": 90.587136, "nll": -5.875259876251221, "elapsed_s": 52.2} {"event": "train", "step": 500, "loss": -5.924335479736328, "grad_norm": 0.6591258645057678, "lr": 0.00029586903590180956, "ex_per_s": 594.3954609525592, "data_frac": 0.23366337956102112, "mem_mb": 90.587136, "nll": -5.924335479736328, "elapsed_s": 53.9} {"event": "val", "step": 500, "val_nll": -6.4400967215224885, "best": -6.4400967215224885, "bad": 0, "mem_mb": 90.587136, "elapsed_s": 118.0} {"event": "train", "step": 520, "loss": -6.039022922515869, "grad_norm": 0.3300909101963043, "lr": 0.00029530320288744205, "ex_per_s": 278.07737618112986, "data_frac": 0.10520308979880166, "mem_mb": 90.587136, "nll": -6.039022922515869, "elapsed_s": 119.7} {"event": "train", "step": 540, "loss": -5.880239486694336, "grad_norm": 0.44749534130096436, "lr": 0.0002947017462784745, "ex_per_s": 284.8155976531226, "data_frac": 0.10384894759549192, "mem_mb": 90.587136, "nll": -5.880239486694336, "elapsed_s": 121.4} {"event": "train", "step": 560, "loss": -5.973054885864258, "grad_norm": 0.11330600082874298, "lr": 0.00029406483050713824, "ex_per_s": 291.359899508687, "data_frac": 0.10253468936154252, "mem_mb": 90.587136, "nll": -5.973054885864258, "elapsed_s": 123.1} {"event": "train", "step": 580, "loss": -5.9250593185424805, "grad_norm": 0.16394247114658356, "lr": 0.0002933926296998457, "ex_per_s": 297.7439312031418, "data_frac": 0.10125525236198518, "mem_mb": 90.587136, "nll": -5.9250593185424805, "elapsed_s": 124.7} {"event": "train", "step": 600, "loss": -5.974524021148682, "grad_norm": 0.9586124420166016, "lr": 0.0002926853276295856, "ex_per_s": 303.9485962032029, "data_frac": 0.10000587812989611, "mem_mb": 90.587136, "nll": -5.974524021148682, "elapsed_s": 126.4} {"event": "train", "step": 620, "loss": -5.954524517059326, "grad_norm": 0.32521307468414307, "lr": 0.00029194311766568174, "ex_per_s": 309.99021592697335, "data_frac": 0.09879042766281078, "mem_mb": 90.587136, "nll": -5.954524517059326, "elapsed_s": 128.0} {"event": "train", "step": 640, "loss": -6.066202640533447, "grad_norm": 0.5233139991760254, "lr": 0.0002911662027209272, "ex_per_s": 315.87212370914466, "data_frac": 0.09760550522750099, "mem_mb": 90.587136, "nll": -6.066202640533447, "elapsed_s": 129.7} {"event": "train", "step": 660, "loss": -6.010836124420166, "grad_norm": 1.0010420083999634, "lr": 0.00029035479519611045, "ex_per_s": 321.61610883902927, "data_frac": 0.0964515935594922, "mem_mb": 90.587136, "nll": -6.010836124420166, "elapsed_s": 131.4} {"event": "train", "step": 680, "loss": -6.007375717163086, "grad_norm": 0.23944666981697083, "lr": 0.0002895091169219469, "ex_per_s": 327.2120644737735, "data_frac": 0.09532799045987464, "mem_mb": 90.587136, "nll": -6.007375717163086, "elapsed_s": 133.0} {"event": "train", "step": 700, "loss": -5.9861602783203125, "grad_norm": 0.937004566192627, "lr": 0.00028862939909843273, "ex_per_s": 332.6603696364198, "data_frac": 0.09423455093076498, "mem_mb": 90.587136, "nll": -5.9861602783203125, "elapsed_s": 134.7} {"event": "train", "step": 720, "loss": -6.062954902648926, "grad_norm": 0.5395597219467163, "lr": 0.0002877158822316369, "ex_per_s": 337.98674601311103, "data_frac": 0.09316417873301217, "mem_mb": 90.587136, "nll": -6.062954902648926, "elapsed_s": 136.4} {"event": "train", "step": 740, "loss": -6.032710552215576, "grad_norm": 0.2784220576286316, "lr": 0.00028676881606794945, "ex_per_s": 343.1840485082632, "data_frac": 0.0921185431632897, "mem_mb": 90.587136, "nll": -6.032710552215576, "elapsed_s": 138.0} {"event": "train", "step": 760, "loss": -5.954955577850342, "grad_norm": 1.2226871252059937, "lr": 0.000285788459525803, "ex_per_s": 348.2569728225811, "data_frac": 0.09109545875987494, "mem_mb": 90.587136, "nll": -5.954955577850342, "elapsed_s": 139.7} {"event": "train", "step": 780, "loss": -5.896873950958252, "grad_norm": 0.3362393379211426, "lr": 0.00028477508062488754, "ex_per_s": 353.2230278208635, "data_frac": 0.09010055578397037, "mem_mb": 90.587136, "nll": -5.896873950958252, "elapsed_s": 141.4} {"event": "train", "step": 800, "loss": -5.814384460449219, "grad_norm": 0.28393641114234924, "lr": 0.000283728956412876, "ex_per_s": 358.10499452580444, "data_frac": 0.08913814457340827, "mem_mb": 90.587136, "nll": -5.814384460449219, "elapsed_s": 143.0} {"event": "train", "step": 820, "loss": -5.934173583984375, "grad_norm": 0.744498610496521, "lr": 0.00028265037288968236, "ex_per_s": 362.88099442979666, "data_frac": 0.08819820786467966, "mem_mb": 90.587136, "nll": -5.934173583984375, "elapsed_s": 144.7} {"event": "train", "step": 840, "loss": -5.808391094207764, "grad_norm": 0.6341730952262878, "lr": 0.0002815396249292725, "ex_per_s": 367.5488643462515, "data_frac": 0.08738755117615432, "mem_mb": 90.587136, "nll": -5.808391094207764, "elapsed_s": 146.3} {"event": "train", "step": 860, "loss": -5.916451454162598, "grad_norm": 1.2756142616271973, "lr": 0.00028039701619904774, "ex_per_s": 372.11235136512414, "data_frac": 0.08648969565754319, "mem_mb": 90.587136, "nll": -5.916451454162598, "elapsed_s": 148.0} {"event": "train", "step": 880, "loss": -5.8720574378967285, "grad_norm": 0.10916467010974884, "lr": 0.00027922285907682625, "ex_per_s": 376.57556871493955, "data_frac": 0.08561276923634509, "mem_mb": 90.587136, "nll": -5.8720574378967285, "elapsed_s": 149.6} {"event": "train", "step": 900, "loss": -5.986175060272217, "grad_norm": 0.2624615728855133, "lr": 0.00027801747456544134, "ex_per_s": 380.9398115599836, "data_frac": 0.0847528150005876, "mem_mb": 90.587136, "nll": -5.986175060272217, "elapsed_s": 151.2} {"event": "train", "step": 920, "loss": -6.026350021362305, "grad_norm": 0.3811030387878418, "lr": 0.00027678119220498265, "ex_per_s": 385.2101725434486, "data_frac": 0.08391220871928806, "mem_mb": 90.587136, "nll": -6.026350021362305, "elapsed_s": 152.9} {"event": "train", "step": 940, "loss": -6.08756685256958, "grad_norm": 0.46402379870414734, "lr": 0.0002755143499827035, "ex_per_s": 389.37393113037785, "data_frac": 0.08309070914602247, "mem_mb": 90.587136, "nll": -6.08756685256958, "elapsed_s": 154.5} {"event": "train", "step": 960, "loss": -5.9662394523620605, "grad_norm": 0.719545841217041, "lr": 0.00027421729424061787, "ex_per_s": 393.46621692041384, "data_frac": 0.08228521642097665, "mem_mb": 90.587136, "nll": -5.9662394523620605, "elapsed_s": 156.2} {"event": "train", "step": 980, "loss": -5.924607276916504, "grad_norm": 0.14955389499664307, "lr": 0.00027289037958081483, "ex_per_s": 397.47144069226766, "data_frac": 0.08149701420043395, "mem_mb": 90.587136, "nll": -5.924607276916504, "elapsed_s": 157.8} {"event": "train", "step": 1000, "loss": -6.070557594299316, "grad_norm": 0.26036572456359863, "lr": 0.0002715339687685131, "ex_per_s": 401.3968740008927, "data_frac": 0.08072381048537142, "mem_mb": 90.587136, "nll": -6.070557594299316, "elapsed_s": 159.5} {"event": "val", "step": 1000, "val_nll": -6.447098318779918, "best": -6.447098318779918, "bad": 0, "mem_mb": 90.587136, "elapsed_s": 223.4} {"event": "train", "step": 1020, "loss": -6.029048919677734, "grad_norm": 0.7155010104179382, "lr": 0.0002701484326328852, "ex_per_s": 290.0661282551269, "data_frac": 0.057240910352274986, "mem_mb": 90.587136, "nll": -6.029048919677734, "elapsed_s": 225.1} {"event": "train", "step": 1040, "loss": -5.973357677459717, "grad_norm": 0.2389991730451584, "lr": 0.00026873414996567596, "ex_per_s": 293.58396145741307, "data_frac": 0.05686829362183427, "mem_mb": 90.587136, "nll": -5.973357677459717, "elapsed_s": 226.8} {"event": "train", "step": 1060, "loss": -6.184004306793213, "grad_norm": 0.4276517629623413, "lr": 0.0002672915074176452, "ex_per_s": 297.0473098596255, "data_frac": 0.05650039107479461, "mem_mb": 90.587136, "nll": -6.184004306793213, "elapsed_s": 228.4} {"event": "train", "step": 1080, "loss": -5.812333106994629, "grad_norm": 0.6120077967643738, "lr": 0.00026582089939286075, "ex_per_s": 300.4618691337291, "data_frac": 0.05613937861409942, "mem_mb": 90.587136, "nll": -5.812333106994629, "elapsed_s": 230.1} {"event": "train", "step": 1100, "loss": -6.050426483154297, "grad_norm": 0.6818838715553284, "lr": 0.0002643227279408728, "ex_per_s": 303.82937235423344, "data_frac": 0.05578277982662251, "mem_mb": 90.587136, "nll": -6.050426483154297, "elapsed_s": 231.8} {"event": "train", "step": 1120, "loss": -6.053744316101074, "grad_norm": 0.12272989004850388, "lr": 0.000262797402646797, "ex_per_s": 307.1476663876095, "data_frac": 0.05543224053536517, "mem_mb": 90.587136, "nll": -6.053744316101074, "elapsed_s": 233.4} {"event": "train", "step": 1140, "loss": -6.021508693695068, "grad_norm": 0.2045598179101944, "lr": 0.0002612453405193386, "ex_per_s": 310.4137046829628, "data_frac": 0.05508815953280335, "mem_mb": 90.587136, "nll": -6.021508693695068, "elapsed_s": 235.1} {"event": "train", "step": 1160, "loss": -5.969545364379883, "grad_norm": 0.4935953617095947, "lr": 0.0002596669658767859, "ex_per_s": 313.64164731472937, "data_frac": 0.05474781324850234, "mem_mb": 90.587136, "nll": -5.969545364379883, "elapsed_s": 236.7} {"event": "train", "step": 1180, "loss": -5.956007957458496, "grad_norm": 0.7713842988014221, "lr": 0.00025806271023100637, "ex_per_s": 316.825229893178, "data_frac": 0.054411401793223076, "mem_mb": 90.587136, "nll": -5.956007957458496, "elapsed_s": 238.4} {"event": "train", "step": 1200, "loss": -5.947319507598877, "grad_norm": 0.3419705629348755, "lr": 0.00025643301216947504, "ex_per_s": 319.96269343164755, "data_frac": 0.054079352524274815, "mem_mb": 90.587136, "nll": -5.947319507598877, "elapsed_s": 240.1} {"event": "train", "step": 1220, "loss": -6.032900810241699, "grad_norm": 0.7673874497413635, "lr": 0.0002547783172353695, "ex_per_s": 323.0601984970959, "data_frac": 0.053753179358521426, "mem_mb": 90.587136, "nll": -6.032900810241699, "elapsed_s": 241.7} {"event": "train", "step": 1240, "loss": -6.0585832595825195, "grad_norm": 0.12547777593135834, "lr": 0.00025309907780576277, "ex_per_s": 326.11250178446585, "data_frac": 0.053432437750056246, "mem_mb": 90.587136, "nll": -6.0585832595825195, "elapsed_s": 243.4} {"event": "train", "step": 1260, "loss": -6.054042339324951, "grad_norm": 0.1846412718296051, "lr": 0.0002513957529679477, "ex_per_s": 329.13069704012594, "data_frac": 0.05311466961340365, "mem_mb": 90.587136, "nll": -6.054042339324951, "elapsed_s": 245.1} {"event": "train", "step": 1280, "loss": -6.046598434448242, "grad_norm": 0.4068543314933777, "lr": 0.00024966880839392746, "ex_per_s": 332.10331196435226, "data_frac": 0.05280413856612441, "mem_mb": 90.587136, "nll": -6.046598434448242, "elapsed_s": 246.7} {"event": "train", "step": 1300, "loss": -5.859414577484131, "grad_norm": 1.5984148979187012, "lr": 0.0002479187162131052, "ex_per_s": 335.0419387876392, "data_frac": 0.052495539966477725, "mem_mb": 90.587136, "nll": -5.859414577484131, "elapsed_s": 248.4} {"event": "train", "step": 1320, "loss": -5.828230381011963, "grad_norm": 0.3810129165649414, "lr": 0.00024614595488320833, "ex_per_s": 337.93674929694913, "data_frac": 0.052191413345282524, "mem_mb": 90.587136, "nll": -5.828230381011963, "elapsed_s": 250.0} {"event": "train", "step": 1340, "loss": -5.952639102935791, "grad_norm": 0.10180147737264633, "lr": 0.00024435100905948387, "ex_per_s": 340.7953620161362, "data_frac": 0.0518897229602631, "mem_mb": 90.587136, "nll": -5.952639102935791, "elapsed_s": 251.7} {"event": "train", "step": 1360, "loss": -6.037583827972412, "grad_norm": 0.7584078311920166, "lr": 0.0002425343694621974, "ex_per_s": 343.61617503797703, "data_frac": 0.051592628507933606, "mem_mb": 90.587136, "nll": -6.037583827972412, "elapsed_s": 253.3} {"event": "train", "step": 1380, "loss": -6.068718433380127, "grad_norm": 0.11085939407348633, "lr": 0.00024069653274247583, "ex_per_s": 346.39710897496104, "data_frac": 0.05130001702824968, "mem_mb": 90.587136, "nll": -6.068718433380127, "elapsed_s": 255.0} {"event": "train", "step": 1400, "loss": -5.798119068145752, "grad_norm": 0.08268816769123077, "lr": 0.0002388380013465276, "ex_per_s": 349.13915367173735, "data_frac": 0.05101200892778503, "mem_mb": 90.587136, "nll": -5.798119068145752, "elapsed_s": 256.7} {"event": "train", "step": 1420, "loss": -5.991825580596924, "grad_norm": 0.5008009076118469, "lr": 0.00023695928337827914, "ex_per_s": 351.8503895898282, "data_frac": 0.050726417099556304, "mem_mb": 90.587136, "nll": -5.991825580596924, "elapsed_s": 258.3} {"event": "train", "step": 1440, "loss": -5.743188381195068, "grad_norm": 0.12715363502502441, "lr": 0.00023506089246046403, "ex_per_s": 354.5276686647502, "data_frac": 0.05044464899785326, "mem_mb": 90.587136, "nll": -5.743188381195068, "elapsed_s": 260.0} {"event": "train", "step": 1460, "loss": -5.804590702056885, "grad_norm": 0.09564962238073349, "lr": 0.00023314334759420388, "ex_per_s": 357.1637955953646, "data_frac": 0.050164744765307585, "mem_mb": 90.587136, "nll": -5.804590702056885, "elapsed_s": 261.7} {"event": "train", "step": 1480, "loss": -6.040069580078125, "grad_norm": 1.2892457246780396, "lr": 0.0002312071730171186, "ex_per_s": 359.7682787668027, "data_frac": 0.0498892473055289, "mem_mb": 90.587136, "nll": -6.040069580078125, "elapsed_s": 263.3} {"event": "train", "step": 1500, "loss": -5.957092761993408, "grad_norm": 0.49252837896347046, "lr": 0.0002292528980600049, "ex_per_s": 362.34097570654757, "data_frac": 0.049617659368012976, "mem_mb": 90.587136, "nll": -5.957092761993408, "elapsed_s": 265.0} {"event": "val", "step": 1500, "val_nll": -6.448654807994589, "best": -6.448654807994589, "bad": 0, "mem_mb": 90.587136, "elapsed_s": 328.8} {"event": "train", "step": 1520, "loss": -5.949714183807373, "grad_norm": 0.6961193680763245, "lr": 0.00022728105700212285, "ex_per_s": 294.39672994732365, "data_frac": 0.039815428926819336, "mem_mb": 90.587136, "nll": -5.949714183807373, "elapsed_s": 330.5} {"event": "train", "step": 1540, "loss": -5.8810625076293945, "grad_norm": 0.13812057673931122, "lr": 0.00022529218892512886, "ex_per_s": 296.7704179437142, "data_frac": 0.03964768182905842, "mem_mb": 90.587136, "nll": -5.8810625076293945, "elapsed_s": 332.2} {"event": "train", "step": 1560, "loss": -5.984130382537842, "grad_norm": 0.22336342930793762, "lr": 0.00022328683756569652, "ex_per_s": 299.11964631810855, "data_frac": 0.03948138891014645, "mem_mb": 90.587136, "nll": -5.984130382537842, "elapsed_s": 333.8} {"event": "train", "step": 1580, "loss": -5.983509063720703, "grad_norm": 0.1876135617494583, "lr": 0.0002212655511668642, "ex_per_s": 301.44447909038234, "data_frac": 0.0393176751154522, "mem_mb": 90.587136, "nll": -5.983509063720703, "elapsed_s": 335.5} {"event": "train", "step": 1600, "loss": -5.9582977294921875, "grad_norm": 0.28975942730903625, "lr": 0.00021922888232815089, "ex_per_s": 303.74897248083084, "data_frac": 0.039157875240479964, "mem_mb": 90.587136, "nll": -5.9582977294921875, "elapsed_s": 337.2} {"event": "train", "step": 1620, "loss": -5.881650924682617, "grad_norm": 0.24355770647525787, "lr": 0.00021717738785448102, "ex_per_s": 306.0314440230334, "data_frac": 0.03899728933184949, "mem_mb": 90.587136, "nll": -5.881650924682617, "elapsed_s": 338.8} {"event": "train", "step": 1640, "loss": -6.078200817108154, "grad_norm": 0.28681421279907227, "lr": 0.00021511162860395977, "ex_per_s": 308.2885277380397, "data_frac": 0.0388382861023846, "mem_mb": 90.587136, "nll": -6.078200817108154, "elapsed_s": 340.5} {"event": "train", "step": 1660, "loss": -6.0416789054870605, "grad_norm": 0.9503816962242126, "lr": 0.0002130321693345397, "ex_per_s": 310.52834932339897, "data_frac": 0.038684131901132804, "mem_mb": 90.587136, "nll": -6.0416789054870605, "elapsed_s": 342.2} {"event": "train", "step": 1680, "loss": -5.868173599243164, "grad_norm": 0.16268715262413025, "lr": 0.0002109395785496218, "ex_per_s": 312.7470635349712, "data_frac": 0.038528460417247834, "mem_mb": 90.587136, "nll": -5.868173599243164, "elapsed_s": 343.8} {"event": "train", "step": 1700, "loss": -5.928957462310791, "grad_norm": 0.43875235319137573, "lr": 0.00020883442834263225, "ex_per_s": 314.94403318708913, "data_frac": 0.03837477200129992, "mem_mb": 90.587136, "nll": -5.928957462310791, "elapsed_s": 345.5} {"event": "train", "step": 1720, "loss": -5.889745235443115, "grad_norm": 0.8626999855041504, "lr": 0.00020671729424061788, "ex_per_s": 317.1212407962519, "data_frac": 0.038221689130062464, "mem_mb": 90.587136, "nll": -5.889745235443115, "elapsed_s": 347.2} {"event": "train", "step": 1740, "loss": -6.275541305541992, "grad_norm": 0.2991836667060852, "lr": 0.00020458875504690275, "ex_per_s": 319.27865231407935, "data_frac": 0.03807043982040267, "mem_mb": 90.587136, "nll": -6.275541305541992, "elapsed_s": 348.8} {"event": "train", "step": 1760, "loss": -5.932249069213867, "grad_norm": 0.19309498369693756, "lr": 0.000202449392682849, "ex_per_s": 321.41498855407986, "data_frac": 0.03792079575960729, "mem_mb": 90.587136, "nll": -5.932249069213867, "elapsed_s": 350.5} {"event": "train", "step": 1780, "loss": -5.870885848999023, "grad_norm": 0.43099600076675415, "lr": 0.00020029979202876563, "ex_per_s": 323.5318951760344, "data_frac": 0.037773422457859265, "mem_mb": 90.587136, "nll": -5.870885848999023, "elapsed_s": 352.2} {"event": "train", "step": 1800, "loss": -6.054300785064697, "grad_norm": 0.766976535320282, "lr": 0.00019814054076400787, "ex_per_s": 325.62868810509076, "data_frac": 0.03762656440659765, "mem_mb": 90.587136, "nll": -6.054300785064697, "elapsed_s": 353.8} {"event": "train", "step": 1820, "loss": -5.939777374267578, "grad_norm": 0.1445099264383316, "lr": 0.00019597222920631147, "ex_per_s": 327.7041721126535, "data_frac": 0.03748187595866012, "mem_mb": 90.587136, "nll": -5.939777374267578, "elapsed_s": 355.5} {"event": "train", "step": 1840, "loss": -6.0736589431762695, "grad_norm": 0.5474252104759216, "lr": 0.0001937954501504058, "ex_per_s": 329.76368800288634, "data_frac": 0.03734070408162993, "mem_mb": 90.587136, "nll": -6.0736589431762695, "elapsed_s": 357.1} {"event": "train", "step": 1860, "loss": -6.05275821685791, "grad_norm": 0.15224459767341614, "lr": 0.0001916107987059498, "ex_per_s": 331.8055486797205, "data_frac": 0.037199754715734765, "mem_mb": 90.587136, "nll": -6.05275821685791, "elapsed_s": 358.8} {"event": "train", "step": 1880, "loss": -5.950157642364502, "grad_norm": 0.2617349326610565, "lr": 0.00018941887213483484, "ex_per_s": 333.8299034585442, "data_frac": 0.03705846053012929, "mem_mb": 90.587136, "nll": -5.950157642364502, "elapsed_s": 360.5} {"event": "train", "step": 1900, "loss": -5.983439922332764, "grad_norm": 0.22470097243785858, "lr": 0.00018722026968789906, "ex_per_s": 335.8371262268652, "data_frac": 0.036919460012517044, "mem_mb": 90.587136, "nll": -5.983439922332764, "elapsed_s": 362.1} {"event": "train", "step": 1920, "loss": -5.966001987457275, "grad_norm": 0.47159910202026367, "lr": 0.00018501559244109864, "ex_per_s": 337.82314716742655, "data_frac": 0.036781378238809455, "mem_mb": 90.587136, "nll": -5.966001987457275, "elapsed_s": 363.8} {"event": "train", "step": 1940, "loss": -6.050561428070068, "grad_norm": 0.36435723304748535, "lr": 0.00018280544313117932, "ex_per_s": 339.7941390907985, "data_frac": 0.036644514820966624, "mem_mb": 90.587136, "nll": -6.050561428070068, "elapsed_s": 365.4} {"event": "train", "step": 1960, "loss": -5.942800998687744, "grad_norm": 0.28399553894996643, "lr": 0.00018059042599089415, "ex_per_s": 341.74644764825103, "data_frac": 0.03650917238146328, "mem_mb": 90.587136, "nll": -5.942800998687744, "elapsed_s": 367.1} {"event": "train", "step": 1980, "loss": -5.964433670043945, "grad_norm": 0.7620005011558533, "lr": 0.00017837114658381247, "ex_per_s": 343.68029619736274, "data_frac": 0.03637445125848536, "mem_mb": 90.587136, "nll": -5.964433670043945, "elapsed_s": 368.8} {"event": "train", "step": 2000, "loss": -5.9165873527526855, "grad_norm": 0.09619174152612686, "lr": 0.00017614821163876486, "ex_per_s": 345.5949599279343, "data_frac": 0.036242826336247605, "mem_mb": 90.587136, "nll": -5.9165873527526855, "elapsed_s": 370.4} {"event": "val", "step": 2000, "val_nll": -6.4493568058476765, "best": -6.4493568058476765, "bad": 0, "mem_mb": 90.587136, "elapsed_s": 433.8} {"event": "train", "step": 2020, "loss": -5.92819356918335, "grad_norm": 0.456305593252182, "lr": 0.00017392222888396941, "ex_per_s": 296.84318998557063, "data_frac": 0.03085039112259908, "mem_mb": 90.587136, "nll": -5.92819356918335, "elapsed_s": 435.6} {"event": "train", "step": 2040, "loss": -5.875527381896973, "grad_norm": 0.26204895973205566, "lr": 0.000171693806880885, "ex_per_s": 298.6308654089833, "data_frac": 0.030760951066246593, "mem_mb": 90.587136, "nll": -5.875527381896973, "elapsed_s": 437.2} {"event": "train", "step": 2060, "loss": -6.144644260406494, "grad_norm": 0.1853942722082138, "lr": 0.0001694635548578366, "ex_per_s": 300.40767805544107, "data_frac": 0.030671965789852446, "mem_mb": 90.587136, "nll": -6.144644260406494, "elapsed_s": 438.9} {"event": "train", "step": 2080, "loss": -5.963052749633789, "grad_norm": 0.41003677248954773, "lr": 0.0001672320825434584, "ex_per_s": 302.17217919626296, "data_frac": 0.030581531591184284, "mem_mb": 90.587136, "nll": -5.963052749633789, "elapsed_s": 440.6} {"event": "train", "step": 2100, "loss": -5.898340225219727, "grad_norm": 1.1166911125183105, "lr": 0.000165, "ex_per_s": 303.9269567986206, "data_frac": 0.03049074033578876, "mem_mb": 90.587136, "nll": -5.898340225219727, "elapsed_s": 442.3} {"event": "train", "step": 2120, "loss": -6.14968204498291, "grad_norm": 0.602655827999115, "lr": 0.0001627679174565416, "ex_per_s": 305.6688536198582, "data_frac": 0.030400482033253986, "mem_mb": 90.587136, "nll": -6.14968204498291, "elapsed_s": 443.9} {"event": "train", "step": 2140, "loss": -5.9812912940979, "grad_norm": 0.1489732563495636, "lr": 0.00016053644514216336, "ex_per_s": 307.4000890230319, "data_frac": 0.030311998507583194, "mem_mb": 90.587136, "nll": -5.9812912940979, "elapsed_s": 445.6} {"event": "train", "step": 2160, "loss": -5.927874565124512, "grad_norm": 0.3962521255016327, "lr": 0.00015830619311911498, "ex_per_s": 309.1175710232609, "data_frac": 0.030226001880680663, "mem_mb": 90.587136, "nll": -5.927874565124512, "elapsed_s": 447.3} {"event": "train", "step": 2180, "loss": -6.083937644958496, "grad_norm": 0.40149423480033875, "lr": 0.00015607777111603058, "ex_per_s": 310.81990208976964, "data_frac": 0.03014430057761571, "mem_mb": 90.587136, "nll": -6.083937644958496, "elapsed_s": 448.9} {"event": "train", "step": 2200, "loss": -5.96399450302124, "grad_norm": 0.6236323118209839, "lr": 0.00015385178836123514, "ex_per_s": 312.51482600277376, "data_frac": 0.030057988317958813, "mem_mb": 90.587136, "nll": -5.96399450302124, "elapsed_s": 450.6} {"event": "train", "step": 2220, "loss": -5.82511043548584, "grad_norm": 0.6660694479942322, "lr": 0.00015162885341618752, "ex_per_s": 314.19820703952234, "data_frac": 0.029971479753003658, "mem_mb": 90.587136, "nll": -5.82511043548584, "elapsed_s": 452.2} {"event": "train", "step": 2240, "loss": -5.962779521942139, "grad_norm": 0.545706033706665, "lr": 0.00014940957400910585, "ex_per_s": 315.872004900631, "data_frac": 0.029886557910790326, "mem_mb": 90.587136, "nll": -5.962779521942139, "elapsed_s": 453.9} {"event": "train", "step": 2260, "loss": -5.738633632659912, "grad_norm": 0.6488598585128784, "lr": 0.00014719455686882068, "ex_per_s": 317.5361157356732, "data_frac": 0.02980163246275197, "mem_mb": 90.587136, "nll": -5.738633632659912, "elapsed_s": 455.6} {"event": "train", "step": 2280, "loss": -6.0126166343688965, "grad_norm": 0.3685223460197449, "lr": 0.0001449844075589013, "ex_per_s": 319.1873099374479, "data_frac": 0.029720305040435928, "mem_mb": 90.587136, "nll": -6.0126166343688965, "elapsed_s": 457.2} {"event": "train", "step": 2300, "loss": -6.001496315002441, "grad_norm": 0.2411249577999115, "lr": 0.0001427797303121009, "ex_per_s": 320.82849361724874, "data_frac": 0.029637404022537873, "mem_mb": 90.587136, "nll": -6.001496315002441, "elapsed_s": 458.9} {"event": "train", "step": 2320, "loss": -6.071399211883545, "grad_norm": 0.42445871233940125, "lr": 0.00014058112786516518, "ex_per_s": 322.4590869633632, "data_frac": 0.029555344372622257, "mem_mb": 90.587136, "nll": -6.071399211883545, "elapsed_s": 460.5} {"event": "train", "step": 2340, "loss": -5.853727340698242, "grad_norm": 0.39282578229904175, "lr": 0.00013838920129405016, "ex_per_s": 324.07480440107827, "data_frac": 0.029474804129493486, "mem_mb": 90.587136, "nll": -5.853727340698242, "elapsed_s": 462.2} {"event": "train", "step": 2360, "loss": -5.93007755279541, "grad_norm": 0.1663031280040741, "lr": 0.0001362045498495942, "ex_per_s": 325.6815933893477, "data_frac": 0.029395250585653364, "mem_mb": 90.587136, "nll": -5.93007755279541, "elapsed_s": 463.8} {"event": "train", "step": 2380, "loss": -5.931504726409912, "grad_norm": 0.05828210711479187, "lr": 0.00013402777079368853, "ex_per_s": 327.2699809179027, "data_frac": 0.02931895304343313, "mem_mb": 90.587136, "nll": -5.931504726409912, "elapsed_s": 465.5} {"event": "train", "step": 2400, "loss": -5.958134174346924, "grad_norm": 0.9089797735214233, "lr": 0.00013185945923599213, "ex_per_s": 328.8420636728866, "data_frac": 0.029253132245272404, "mem_mb": 90.587136, "nll": -5.958134174346924, "elapsed_s": 467.1} {"event": "train", "step": 2420, "loss": -5.996464252471924, "grad_norm": 0.43496984243392944, "lr": 0.00012970020797123432, "ex_per_s": 330.40520327089524, "data_frac": 0.029188168890700972, "mem_mb": 90.587136, "nll": -5.996464252471924, "elapsed_s": 468.8} {"event": "train", "step": 2440, "loss": -6.099253177642822, "grad_norm": 0.24610909819602966, "lr": 0.000127550607317151, "ex_per_s": 331.95769650289776, "data_frac": 0.02912401264877737, "mem_mb": 90.587136, "nll": -6.099253177642822, "elapsed_s": 470.5} {"event": "train", "step": 2460, "loss": -6.134093761444092, "grad_norm": 0.1689298450946808, "lr": 0.00012541124495309727, "ex_per_s": 333.50001927892697, "data_frac": 0.029060186427281982, "mem_mb": 90.587136, "nll": -6.134093761444092, "elapsed_s": 472.1} {"event": "train", "step": 2480, "loss": -5.9468512535095215, "grad_norm": 0.179790660738945, "lr": 0.00012328270575938212, "ex_per_s": 335.0286157204167, "data_frac": 0.028995955691066947, "mem_mb": 90.587136, "nll": -5.9468512535095215, "elapsed_s": 473.8} {"event": "train", "step": 2500, "loss": -5.7570319175720215, "grad_norm": 0.6756011843681335, "lr": 0.00012116557165736775, "ex_per_s": 336.5466852626393, "data_frac": 0.02893280650797846, "mem_mb": 90.587136, "nll": -5.7570319175720215, "elapsed_s": 475.5} {"event": "val", "step": 2500, "val_nll": -6.450321103123497, "best": -6.450321103123497, "bad": 0, "mem_mb": 90.587136, "elapsed_s": 539.7} {"event": "train", "step": 2520, "loss": -5.959056854248047, "grad_norm": 0.8921460509300232, "lr": 0.00011906042145037817, "ex_per_s": 297.8909410993082, "data_frac": 0.02542758614867064, "mem_mb": 90.587136, "nll": -5.959056854248047, "elapsed_s": 541.5} {"event": "train", "step": 2540, "loss": -6.031774044036865, "grad_norm": 0.21482045948505402, "lr": 0.00011696783066546026, "ex_per_s": 299.33164933565047, "data_frac": 0.025369565258257846, "mem_mb": 90.587136, "nll": -6.031774044036865, "elapsed_s": 543.1} {"event": "train", "step": 2560, "loss": -5.94840145111084, "grad_norm": 0.17889054119586945, "lr": 0.00011488837139604017, "ex_per_s": 300.76334838136506, "data_frac": 0.025312180103744388, "mem_mb": 90.587136, "nll": -5.94840145111084, "elapsed_s": 544.8} {"event": "train", "step": 2580, "loss": -5.963006019592285, "grad_norm": 0.1597909778356552, "lr": 0.00011282261214551898, "ex_per_s": 302.18483534327856, "data_frac": 0.025254318966074524, "mem_mb": 90.587136, "nll": -5.963006019592285, "elapsed_s": 546.5} {"event": "train", "step": 2600, "loss": -5.850461006164551, "grad_norm": 0.15184129774570465, "lr": 0.00011077111767184915, "ex_per_s": 303.59889913947546, "data_frac": 0.025197162905642723, "mem_mb": 90.587136, "nll": -5.850461006164551, "elapsed_s": 548.1} {"event": "train", "step": 2620, "loss": -5.8891496658325195, "grad_norm": 1.0519516468048096, "lr": 0.00010873444883313583, "ex_per_s": 305.003685307614, "data_frac": 0.025140802508413034, "mem_mb": 90.587136, "nll": -5.8891496658325195, "elapsed_s": 549.8} {"event": "train", "step": 2640, "loss": -5.988500595092773, "grad_norm": 0.42157837748527527, "lr": 0.00010671316243430345, "ex_per_s": 306.39957898969953, "data_frac": 0.025085045743545316, "mem_mb": 90.587136, "nll": -5.988500595092773, "elapsed_s": 551.5} {"event": "train", "step": 2660, "loss": -5.822550296783447, "grad_norm": 0.06934574991464615, "lr": 0.00010470781107487112, "ex_per_s": 307.7893291586707, "data_frac": 0.025028854681979, "mem_mb": 90.587136, "nll": -5.822550296783447, "elapsed_s": 553.1} {"event": "train", "step": 2680, "loss": -5.9583516120910645, "grad_norm": 0.43225008249282837, "lr": 0.00010271894299787715, "ex_per_s": 309.16984050827523, "data_frac": 0.024973599897334612, "mem_mb": 90.587136, "nll": -5.9583516120910645, "elapsed_s": 554.8} {"event": "train", "step": 2700, "loss": -6.13491678237915, "grad_norm": 0.6783389449119568, "lr": 0.00010074710193999506, "ex_per_s": 310.54075940177466, "data_frac": 0.024918097884287845, "mem_mb": 90.587136, "nll": -6.13491678237915, "elapsed_s": 556.5} {"event": "train", "step": 2720, "loss": -5.883577346801758, "grad_norm": 0.6422591209411621, "lr": 9.879282698288144e-05, "ex_per_s": 311.905236822733, "data_frac": 0.024865893802593412, "mem_mb": 90.587136, "nll": -5.883577346801758, "elapsed_s": 558.2} {"event": "train", "step": 2740, "loss": -6.126973628997803, "grad_norm": 0.05770125240087509, "lr": 9.685665240579613e-05, "ex_per_s": 313.26122570454055, "data_frac": 0.02481120459330484, "mem_mb": 90.587136, "nll": -6.126973628997803, "elapsed_s": 559.8} {"event": "train", "step": 2760, "loss": -5.96104097366333, "grad_norm": 0.2643730044364929, "lr": 9.4939107539536e-05, "ex_per_s": 314.6106159235282, "data_frac": 0.02475745017868692, "mem_mb": 90.587136, "nll": -5.96104097366333, "elapsed_s": 561.5} {"event": "train", "step": 2780, "loss": -6.015292167663574, "grad_norm": 0.8982701301574707, "lr": 9.304071662172087e-05, "ex_per_s": 315.95202671228844, "data_frac": 0.024703183132343605, "mem_mb": 90.587136, "nll": -6.015292167663574, "elapsed_s": 563.2} {"event": "train", "step": 2800, "loss": -5.979710578918457, "grad_norm": 0.8294841647148132, "lr": 9.116199865347241e-05, "ex_per_s": 317.2867619385624, "data_frac": 0.024649745705973745, "mem_mb": 90.587136, "nll": -5.979710578918457, "elapsed_s": 564.8} {"event": "train", "step": 2820, "loss": -6.008241176605225, "grad_norm": 0.2969299256801605, "lr": 8.930346725752419e-05, "ex_per_s": 318.61460997045396, "data_frac": 0.024596968544993984, "mem_mb": 90.587136, "nll": -6.008241176605225, "elapsed_s": 566.5} {"event": "train", "step": 2840, "loss": -5.8794379234313965, "grad_norm": 0.16628247499465942, "lr": 8.746563053780262e-05, "ex_per_s": 319.9348515209012, "data_frac": 0.024543936066496352, "mem_mb": 90.587136, "nll": -5.8794379234313965, "elapsed_s": 568.2} {"event": "train", "step": 2860, "loss": -6.049481391906738, "grad_norm": 0.24883082509040833, "lr": 8.564899094051614e-05, "ex_per_s": 321.2474983087843, "data_frac": 0.02449177461136243, "mem_mb": 90.587136, "nll": -6.049481391906738, "elapsed_s": 569.8} {"event": "train", "step": 2880, "loss": -5.772535800933838, "grad_norm": 0.2698267996311188, "lr": 8.385404511679161e-05, "ex_per_s": 322.5518138030373, "data_frac": 0.024440194871326947, "mem_mb": 90.587136, "nll": -5.772535800933838, "elapsed_s": 571.5} {"event": "train", "step": 2900, "loss": -5.796339511871338, "grad_norm": 0.6047093868255615, "lr": 8.208128378689484e-05, "ex_per_s": 323.8506038824426, "data_frac": 0.024389498011539085, "mem_mb": 90.587136, "nll": -5.796339511871338, "elapsed_s": 573.1} {"event": "train", "step": 2920, "loss": -5.9708571434021, "grad_norm": 0.6782792210578918, "lr": 8.03311916060725e-05, "ex_per_s": 325.14249667598625, "data_frac": 0.024339667681373398, "mem_mb": 90.587136, "nll": -5.9708571434021, "elapsed_s": 574.8} {"event": "train", "step": 2940, "loss": -5.900778293609619, "grad_norm": 0.7649500966072083, "lr": 7.860424703205225e-05, "ex_per_s": 326.4271336845284, "data_frac": 0.024288484704286212, "mem_mb": 90.587136, "nll": -5.900778293609619, "elapsed_s": 576.5} {"event": "train", "step": 2960, "loss": -5.940110206604004, "grad_norm": 0.3150622844696045, "lr": 7.690092219423722e-05, "ex_per_s": 327.70210149060955, "data_frac": 0.02423961913539305, "mem_mb": 90.587136, "nll": -5.940110206604004, "elapsed_s": 578.1} {"event": "train", "step": 2980, "loss": -5.8953657150268555, "grad_norm": 0.479099303483963, "lr": 7.522168276463048e-05, "ex_per_s": 328.9713267585475, "data_frac": 0.024189447927462422, "mem_mb": 90.587136, "nll": -5.8953657150268555, "elapsed_s": 579.8} {"event": "train", "step": 3000, "loss": -5.996098041534424, "grad_norm": 0.22216247022151947, "lr": 7.356698783052496e-05, "ex_per_s": 330.23244999724716, "data_frac": 0.024139858206426102, "mem_mb": 90.587136, "nll": -5.996098041534424, "elapsed_s": 581.5} {"event": "val", "step": 3000, "val_nll": -6.449514239172613, "best": -6.450321103123497, "bad": 1, "mem_mb": 90.587136, "elapsed_s": 644.7} {"event": "train", "step": 3020, "loss": -5.962279796600342, "grad_norm": 0.8931116461753845, "lr": 7.193728976899362e-05, "ex_per_s": 299.00996336843826, "data_frac": 0.021730314420394466, "mem_mb": 90.587136, "nll": -5.962279796600342, "elapsed_s": 646.4} {"event": "train", "step": 3040, "loss": -5.861953258514404, "grad_norm": 0.7823965549468994, "lr": 7.033303412321404e-05, "ex_per_s": 300.2138117452867, "data_frac": 0.021691040363839675, "mem_mb": 90.587136, "nll": -5.861953258514404, "elapsed_s": 648.1} {"event": "train", "step": 3060, "loss": -6.072576999664307, "grad_norm": 0.27439793944358826, "lr": 6.875465948066139e-05, "ex_per_s": 301.4115451785778, "data_frac": 0.02165199609997594, "mem_mb": 90.587136, "nll": -6.072576999664307, "elapsed_s": 649.8} {"event": "train", "step": 3080, "loss": -5.968700885772705, "grad_norm": 0.3267911374568939, "lr": 6.720259735320299e-05, "ex_per_s": 302.60389879553946, "data_frac": 0.021613286368208907, "mem_mb": 90.587136, "nll": -5.968700885772705, "elapsed_s": 651.5} {"event": "train", "step": 3100, "loss": -6.01653528213501, "grad_norm": 0.2151486724615097, "lr": 6.567727205912722e-05, "ex_per_s": 303.78957231318583, "data_frac": 0.02157471491944035, "mem_mb": 90.587136, "nll": -6.01653528213501, "elapsed_s": 653.1} {"event": "train", "step": 3120, "loss": -6.09162712097168, "grad_norm": 0.6038528084754944, "lr": 6.417910060713926e-05, "ex_per_s": 304.96956487258586, "data_frac": 0.02153626349948577, "mem_mb": 90.587136, "nll": -6.09162712097168, "elapsed_s": 654.8} {"event": "train", "step": 3140, "loss": -6.00908899307251, "grad_norm": 0.29610979557037354, "lr": 6.270849258235485e-05, "ex_per_s": 306.1439576031732, "data_frac": 0.021498373611296844, "mem_mb": 90.587136, "nll": -6.00908899307251, "elapsed_s": 656.5} {"event": "train", "step": 3160, "loss": -6.021483898162842, "grad_norm": 0.17532937228679657, "lr": 6.126585003432406e-05, "ex_per_s": 307.31049324833094, "data_frac": 0.021462034861410573, "mem_mb": 90.587136, "nll": -6.021483898162842, "elapsed_s": 658.1} {"event": "train", "step": 3180, "loss": -6.097679615020752, "grad_norm": 0.11180613189935684, "lr": 5.985156736711483e-05, "ex_per_s": 308.4730527633611, "data_frac": 0.021424384799557102, "mem_mb": 90.587136, "nll": -6.097679615020752, "elapsed_s": 659.8} {"event": "train", "step": 3200, "loss": -6.002908706665039, "grad_norm": 0.36486878991127014, "lr": 5.846603123148688e-05, "ex_per_s": 309.62901724437296, "data_frac": 0.021387200039879078, "mem_mb": 90.587136, "nll": -6.002908706665039, "elapsed_s": 661.5} {"event": "train", "step": 3220, "loss": -5.915755271911621, "grad_norm": 0.15891994535923004, "lr": 5.710962041918516e-05, "ex_per_s": 310.77922489986815, "data_frac": 0.02135012000322648, "mem_mb": 90.587136, "nll": -5.915755271911621, "elapsed_s": 663.2} {"event": "train", "step": 3240, "loss": -5.80257511138916, "grad_norm": 0.5694242715835571, "lr": 5.578270575938211e-05, "ex_per_s": 311.92327397999105, "data_frac": 0.0213125033548404, "mem_mb": 90.587136, "nll": -5.80257511138916, "elapsed_s": 664.8} {"event": "train", "step": 3260, "loss": -5.962348937988281, "grad_norm": 0.5170437097549438, "lr": 5.448565001729653e-05, "ex_per_s": 313.0631843120033, "data_frac": 0.021275701776034293, "mem_mb": 90.587136, "nll": -5.962348937988281, "elapsed_s": 666.5} {"event": "train", "step": 3280, "loss": -6.017902374267578, "grad_norm": 0.3385675847530365, "lr": 5.3218807795017296e-05, "ex_per_s": 314.19900130080214, "data_frac": 0.021239097459650105, "mem_mb": 90.587136, "nll": -6.017902374267578, "elapsed_s": 668.2} {"event": "train", "step": 3300, "loss": -5.804913520812988, "grad_norm": 0.13826704025268555, "lr": 5.198252543455864e-05, "ex_per_s": 315.3286874308234, "data_frac": 0.021202514699966236, "mem_mb": 90.587136, "nll": -5.804913520812988, "elapsed_s": 669.8} {"event": "train", "step": 3320, "loss": -5.990378379821777, "grad_norm": 0.2195957750082016, "lr": 5.07771409231737e-05, "ex_per_s": 316.4537813265703, "data_frac": 0.021166356599401755, "mem_mb": 90.587136, "nll": -5.990378379821777, "elapsed_s": 671.5} {"event": "train", "step": 3340, "loss": -5.932602405548096, "grad_norm": 0.24434122443199158, "lr": 4.960298380095221e-05, "ex_per_s": 317.5721528805523, "data_frac": 0.02113161513793191, "mem_mb": 90.587136, "nll": -5.932602405548096, "elapsed_s": 673.2} {"event": "train", "step": 3360, "loss": -5.906638145446777, "grad_norm": 0.4261169731616974, "lr": 4.846037507072751e-05, "ex_per_s": 318.6874113814924, "data_frac": 0.021095449372453032, "mem_mb": 90.587136, "nll": -5.906638145446777, "elapsed_s": 674.8} {"event": "train", "step": 3380, "loss": -6.122256278991699, "grad_norm": 0.4827311933040619, "lr": 4.7349627110317556e-05, "ex_per_s": 319.79633841601736, "data_frac": 0.021060816204417777, "mem_mb": 90.587136, "nll": -6.122256278991699, "elapsed_s": 676.5} {"event": "train", "step": 3400, "loss": -5.8819684982299805, "grad_norm": 0.45949870347976685, "lr": 4.6271043587123986e-05, "ex_per_s": 320.89905055757697, "data_frac": 0.021027255185156885, "mem_mb": 90.587136, "nll": -5.8819684982299805, "elapsed_s": 678.1} {"event": "train", "step": 3420, "loss": -5.83981990814209, "grad_norm": 0.2743491232395172, "lr": 4.52249193751124e-05, "ex_per_s": 321.9981553582218, "data_frac": 0.020992277442967743, "mem_mb": 90.587136, "nll": -5.83981990814209, "elapsed_s": 679.8} {"event": "train", "step": 3440, "loss": -5.916119575500488, "grad_norm": 0.17352905869483948, "lr": 4.4211540474196926e-05, "ex_per_s": 323.0890349502117, "data_frac": 0.02095719875504174, "mem_mb": 90.587136, "nll": -5.916119575500488, "elapsed_s": 681.5} {"event": "train", "step": 3460, "loss": -6.058788299560547, "grad_norm": 0.15001080930233002, "lr": 4.3231183932050546e-05, "ex_per_s": 324.1743756311533, "data_frac": 0.020921938056793946, "mem_mb": 90.587136, "nll": -6.058788299560547, "elapsed_s": 683.1} {"event": "train", "step": 3480, "loss": -5.917154788970947, "grad_norm": 0.1408236026763916, "lr": 4.228411776836306e-05, "ex_per_s": 325.2551308452064, "data_frac": 0.020887073260212774, "mem_mb": 90.587136, "nll": -5.917154788970947, "elapsed_s": 684.8} {"event": "train", "step": 3500, "loss": -6.040772438049316, "grad_norm": 0.07941709458827972, "lr": 4.137060090156723e-05, "ex_per_s": 326.33028068198695, "data_frac": 0.020852211529145827, "mem_mb": 90.587136, "nll": -6.040772438049316, "elapsed_s": 686.5} {"event": "val", "step": 3500, "val_nll": -6.450862349156063, "best": -6.450862349156063, "bad": 0, "mem_mb": 90.587136, "elapsed_s": 749.9} {"event": "train", "step": 3520, "loss": -5.939716339111328, "grad_norm": 0.4028024971485138, "lr": 4.049088307805306e-05, "ex_per_s": 299.75665152250207, "data_frac": 0.019060524113183014, "mem_mb": 90.587136, "nll": -5.939716339111328, "elapsed_s": 751.6} {"event": "train", "step": 3540, "loss": -6.135863780975342, "grad_norm": 0.6118470430374146, "lr": 3.964520480388956e-05, "ex_per_s": 300.79114365320623, "data_frac": 0.019032464238339965, "mem_mb": 90.587136, "nll": -6.135863780975342, "elapsed_s": 753.3} {"event": "train", "step": 3560, "loss": -6.139781475067139, "grad_norm": 0.08687904477119446, "lr": 3.88337972790728e-05, "ex_per_s": 301.8205556785201, "data_frac": 0.019005117633631587, "mem_mb": 90.587136, "nll": -6.139781475067139, "elapsed_s": 754.9} {"event": "train", "step": 3580, "loss": -5.9040937423706055, "grad_norm": 0.22211135923862457, "lr": 3.805688233431824e-05, "ex_per_s": 302.84603984540126, "data_frac": 0.01897763654286137, "mem_mb": 90.587136, "nll": -5.9040937423706055, "elapsed_s": 756.6} {"event": "train", "step": 3600, "loss": -6.100482940673828, "grad_norm": 0.40555813908576965, "lr": 3.7314672370414325e-05, "ex_per_s": 303.86604345918124, "data_frac": 0.018950714921456727, "mem_mb": 90.587136, "nll": -6.100482940673828, "elapsed_s": 758.3} {"event": "train", "step": 3620, "loss": -5.912384986877441, "grad_norm": 0.8846305012702942, "lr": 3.660737030015427e-05, "ex_per_s": 304.88279065686, "data_frac": 0.01892425764189401, "mem_mb": 90.587136, "nll": -5.912384986877441, "elapsed_s": 759.9} {"event": "train", "step": 3640, "loss": -6.053845405578613, "grad_norm": 0.4929547905921936, "lr": 3.5935169492861716e-05, "ex_per_s": 305.89483456291214, "data_frac": 0.018897425707592605, "mem_mb": 90.587136, "nll": -6.053845405578613, "elapsed_s": 761.6} {"event": "train", "step": 3660, "loss": -5.83349084854126, "grad_norm": 0.527417778968811, "lr": 3.52982537215255e-05, "ex_per_s": 306.90188774290544, "data_frac": 0.018871146846930364, "mem_mb": 90.587136, "nll": -5.83349084854126, "elapsed_s": 763.3} {"event": "train", "step": 3680, "loss": -5.993903636932373, "grad_norm": 0.08807270973920822, "lr": 3.469679711255794e-05, "ex_per_s": 307.9052608422689, "data_frac": 0.018844647386007232, "mem_mb": 90.587136, "nll": -5.993903636932373, "elapsed_s": 765.0} {"event": "train", "step": 3700, "loss": -6.0687103271484375, "grad_norm": 0.7822890877723694, "lr": 3.41309640981904e-05, "ex_per_s": 308.90430196036476, "data_frac": 0.01881819581253249, "mem_mb": 90.587136, "nll": -6.0687103271484375, "elapsed_s": 766.6} {"event": "train", "step": 3720, "loss": -5.951270580291748, "grad_norm": 0.5874511003494263, "lr": 3.3600909371519225e-05, "ex_per_s": 309.89880919913156, "data_frac": 0.018791265144296423, "mem_mb": 90.587136, "nll": -5.951270580291748, "elapsed_s": 768.3} {"event": "train", "step": 3740, "loss": -6.190686225891113, "grad_norm": 0.18538519740104675, "lr": 3.31067778442141e-05, "ex_per_s": 310.8888911012126, "data_frac": 0.018764642240295754, "mem_mb": 90.587136, "nll": -6.190686225891113, "elapsed_s": 770.0} {"event": "train", "step": 3760, "loss": -6.1487345695495605, "grad_norm": 0.2070234715938568, "lr": 3.2648704606900735e-05, "ex_per_s": 311.87398925171635, "data_frac": 0.01873861250543476, "mem_mb": 90.587136, "nll": -6.1487345695495605, "elapsed_s": 771.6} {"event": "train", "step": 3780, "loss": -5.965912342071533, "grad_norm": 0.3024000823497772, "lr": 3.222681489222838e-05, "ex_per_s": 312.85654933768734, "data_frac": 0.018712415459963307, "mem_mb": 90.587136, "nll": -5.965912342071533, "elapsed_s": 773.3} {"event": "train", "step": 3800, "loss": -6.100894927978516, "grad_norm": 0.093129001557827, "lr": 3.1841224040632484e-05, "ex_per_s": 313.83438778588356, "data_frac": 0.018686329301847174, "mem_mb": 90.587136, "nll": -6.100894927978516, "elapsed_s": 775.0} {"event": "train", "step": 3820, "loss": -5.9731669425964355, "grad_norm": 0.1358669549226761, "lr": 3.149203746880175e-05, "ex_per_s": 314.80746473761224, "data_frac": 0.01866117814231169, "mem_mb": 90.587136, "nll": -5.9731669425964355, "elapsed_s": 776.6} {"event": "train", "step": 3840, "loss": -5.874337673187256, "grad_norm": 0.6975057125091553, "lr": 3.117935064085835e-05, "ex_per_s": 315.77658161288747, "data_frac": 0.01863668145253469, "mem_mb": 90.587136, "nll": -5.874337673187256, "elapsed_s": 778.3} {"event": "train", "step": 3860, "loss": -6.014374732971191, "grad_norm": 0.3294009566307068, "lr": 3.090324904225885e-05, "ex_per_s": 316.74213610578647, "data_frac": 0.018611547472678364, "mem_mb": 90.587136, "nll": -6.014374732971191, "elapsed_s": 780.0} {"event": "train", "step": 3880, "loss": -5.890832901000977, "grad_norm": 0.10115442425012589, "lr": 3.0663808156423456e-05, "ex_per_s": 317.70467749109594, "data_frac": 0.018585940269370434, "mem_mb": 90.587136, "nll": -5.890832901000977, "elapsed_s": 781.6} {"event": "train", "step": 3900, "loss": -6.184309482574463, "grad_norm": 0.7105429172515869, "lr": 3.046109344409957e-05, "ex_per_s": 318.66385802838636, "data_frac": 0.018560440280566026, "mem_mb": 90.587136, "nll": -6.184309482574463, "elapsed_s": 783.3} {"event": "train", "step": 3920, "loss": -6.037186145782471, "grad_norm": 0.12338800728321075, "lr": 3.0295160325465435e-05, "ex_per_s": 319.61996539994243, "data_frac": 0.018534999744818446, "mem_mb": 90.587136, "nll": -6.037186145782471, "elapsed_s": 785.0} {"event": "train", "step": 3940, "loss": -5.922738075256348, "grad_norm": 0.16757924854755402, "lr": 3.016605416497886e-05, "ex_per_s": 320.5731889884874, "data_frac": 0.018510352065532986, "mem_mb": 90.587136, "nll": -5.922738075256348, "elapsed_s": 786.6} {"event": "train", "step": 3960, "loss": -6.069593906402588, "grad_norm": 0.1193820908665657, "lr": 3.0073810258974985e-05, "ex_per_s": 321.52231200523, "data_frac": 0.018485418907739224, "mem_mb": 90.587136, "nll": -6.069593906402588, "elapsed_s": 788.3} {"event": "train", "step": 3980, "loss": -5.803100109100342, "grad_norm": 0.23878537118434906, "lr": 3.0018453826016683e-05, "ex_per_s": 322.46777500984064, "data_frac": 0.018460857306795093, "mem_mb": 90.587136, "nll": -5.803100109100342, "elapsed_s": 790.0} {"event": "train", "step": 4000, "loss": -6.056766033172607, "grad_norm": 0.8348932266235352, "lr": 2.9999999999999997e-05, "ex_per_s": 323.40870429046475, "data_frac": 0.01843627321673212, "mem_mb": 90.587136, "nll": -6.056766033172607, "elapsed_s": 791.6} {"event": "val", "step": 4000, "val_nll": -6.450718539128351, "best": -6.450862349156063, "bad": 1, "mem_mb": 90.587136, "elapsed_s": 855.0} {"event": "done", "steps": 4000, "best_val": -6.450862349156063, "wall_s": 854.9879620075226, "ex_per_s": 299.4194203610876, "peak_mem_mb": 90.587136, "n_params": 3416643, "device": "mps", "data_time_frac": 0.01706873738071907, "elapsed_s": 855.0}