SPB Git

spb/wp41_uqo

Public
152.8 KB · 614 lines
1{"event": "train", "step": 20, "loss": -6.074785232543945, "grad_norm": 2.0882441997528076, "lr": 1.5749999999999997e-05, "ex_per_s": 55.21651807349261, "data_frac": 0.9725663633807002, "mem_mb": 199.197696, "nll": -6.074785232543945, "elapsed_s": 23.2}2{"event": "train", "step": 40, "loss": -5.862731456756592, "grad_norm": 1.6481750011444092, "lr": 3.0749999999999995e-05, "ex_per_s": 87.97733394931183, "data_frac": 0.9671997681690494, "mem_mb": 199.197696, "nll": -5.862731456756592, "elapsed_s": 29.1}3{"event": "train", "step": 60, "loss": -5.957578182220459, "grad_norm": 0.9313188195228577, "lr": 4.5749999999999994e-05, "ex_per_s": 120.67998686152552, "data_frac": 0.9632453971113054, "mem_mb": 199.197696, "nll": -5.957578182220459, "elapsed_s": 31.8}4{"event": "train", "step": 80, "loss": -6.0685954093933105, "grad_norm": 0.18229295313358307, "lr": 6.075e-05, "ex_per_s": 148.45454585486831, "data_frac": 0.9594792391943503, "mem_mb": 199.197696, "nll": -6.0685954093933105, "elapsed_s": 34.5}5{"event": "train", "step": 100, "loss": -5.9327616691589355, "grad_norm": 0.43501946330070496, "lr": 7.575e-05, "ex_per_s": 178.94679217312785, "data_frac": 0.9546574324185774, "mem_mb": 199.197696, "nll": -5.9327616691589355, "elapsed_s": 35.8}6{"event": "train", "step": 120, "loss": -5.894950866699219, "grad_norm": 0.4139050245285034, "lr": 9.074999999999998e-05, "ex_per_s": 212.15446110996515, "data_frac": 0.9486251077682067, "mem_mb": 199.197696, "nll": -5.894950866699219, "elapsed_s": 36.2}7{"event": "train", "step": 140, "loss": -5.814927577972412, "grad_norm": 0.23503628373146057, "lr": 0.00010574999999999998, "ex_per_s": 245.64651153870707, "data_frac": 0.9429970253676333, "mem_mb": 199.197696, "nll": -5.814927577972412, "elapsed_s": 36.5}8{"event": "train", "step": 160, "loss": -5.730831146240234, "grad_norm": 0.15382201969623566, "lr": 0.00012075, "ex_per_s": 277.9979963279389, "data_frac": 0.9372827472806152, "mem_mb": 199.197696, "nll": -5.730831146240234, "elapsed_s": 36.8}9{"event": "train", "step": 180, "loss": -5.970207214355469, "grad_norm": 0.9913017749786377, "lr": 0.00013575, "ex_per_s": 310.90117452642187, "data_frac": 0.9319352126660335, "mem_mb": 199.197696, "nll": -5.970207214355469, "elapsed_s": 37.1}10{"event": "train", "step": 200, "loss": -5.962467193603516, "grad_norm": 0.5681918859481812, "lr": 0.00015074999999999998, "ex_per_s": 343.3002054451961, "data_frac": 0.9264606514253408, "mem_mb": 199.197696, "nll": -5.962467193603516, "elapsed_s": 37.3}11{"event": "train", "step": 220, "loss": -6.0237650871276855, "grad_norm": 1.2080389261245728, "lr": 0.00016575, "ex_per_s": 375.42024467252963, "data_frac": 0.9212242769180184, "mem_mb": 199.197696, "nll": -6.0237650871276855, "elapsed_s": 37.5}12{"event": "train", "step": 240, "loss": -5.905878067016602, "grad_norm": 0.7234233021736145, "lr": 0.00018075, "ex_per_s": 407.1976214345555, "data_frac": 0.9161844137582559, "mem_mb": 199.197696, "nll": -5.905878067016602, "elapsed_s": 37.7}13{"event": "train", "step": 260, "loss": -5.885006427764893, "grad_norm": 0.12898588180541992, "lr": 0.00019574999999999996, "ex_per_s": 438.5556181106107, "data_frac": 0.9111292221041267, "mem_mb": 199.197696, "nll": -5.885006427764893, "elapsed_s": 37.9}14{"event": "train", "step": 280, "loss": -5.946396350860596, "grad_norm": 0.2684224247932434, "lr": 0.00021074999999999997, "ex_per_s": 469.6323854335842, "data_frac": 0.9063214328169082, "mem_mb": 199.197696, "nll": -5.946396350860596, "elapsed_s": 38.2}15{"event": "train", "step": 300, "loss": -5.96865177154541, "grad_norm": 0.17942094802856445, "lr": 0.00022574999999999996, "ex_per_s": 500.3392426080634, "data_frac": 0.9015064733246363, "mem_mb": 199.197696, "nll": -5.96865177154541, "elapsed_s": 38.4}16{"event": "train", "step": 320, "loss": -5.965313911437988, "grad_norm": 1.2768847942352295, "lr": 0.00024074999999999997, "ex_per_s": 530.6956511650008, "data_frac": 0.8967660902987936, "mem_mb": 199.197696, "nll": -5.965313911437988, "elapsed_s": 38.6}17{"event": "train", "step": 340, "loss": -5.833019733428955, "grad_norm": 0.39559969305992126, "lr": 0.00025575, "ex_per_s": 560.7211808298791, "data_frac": 0.8920814689023316, "mem_mb": 199.197696, "nll": -5.833019733428955, "elapsed_s": 38.8}18{"event": "train", "step": 360, "loss": -5.993951320648193, "grad_norm": 0.9508464932441711, "lr": 0.00027074999999999994, "ex_per_s": 590.4195042327513, "data_frac": 0.8874554742711107, "mem_mb": 199.197696, "nll": -5.993951320648193, "elapsed_s": 39.0}19{"event": "train", "step": 380, "loss": -6.0088605880737305, "grad_norm": 0.24148868024349213, "lr": 0.00028575, "ex_per_s": 619.794760023424, "data_frac": 0.882888049477967, "mem_mb": 199.197696, "nll": -6.0088605880737305, "elapsed_s": 39.2}20{"event": "train", "step": 400, "loss": -5.996034145355225, "grad_norm": 1.207338809967041, "lr": 0.0003, "ex_per_s": 648.8715655329104, "data_frac": 0.8783934662526113, "mem_mb": 199.197696, "nll": -5.996034145355225, "elapsed_s": 39.5}21{"event": "train", "step": 420, "loss": -5.956755638122559, "grad_norm": 1.2939516305923462, "lr": 0.0002999980196294046, "ex_per_s": 677.6171097766618, "data_frac": 0.8739377569412622, "mem_mb": 199.197696, "nll": -5.956755638122559, "elapsed_s": 39.7}22{"event": "train", "step": 440, "loss": -5.943725109100342, "grad_norm": 0.19923467934131622, "lr": 0.00029999207857572027, "ex_per_s": 706.0608150375584, "data_frac": 0.8695363276995991, "mem_mb": 199.197696, "nll": -5.943725109100342, "elapsed_s": 39.9}23{"event": "train", "step": 460, "loss": -6.102391719818115, "grad_norm": 0.20417994260787964, "lr": 0.00029998217701325043, "ex_per_s": 734.1995079650785, "data_frac": 0.8651865567551117, "mem_mb": 199.197696, "nll": -6.102391719818115, "elapsed_s": 40.1}24{"event": "train", "step": 480, "loss": -5.877411842346191, "grad_norm": 0.21606390178203583, "lr": 0.00029996831523249534, "ex_per_s": 762.0100882742627, "data_frac": 0.8608576843982945, "mem_mb": 199.197696, "nll": -5.877411842346191, "elapsed_s": 40.3}25{"event": "train", "step": 500, "loss": -5.92513370513916, "grad_norm": 0.2452145367860794, "lr": 0.0002999504936401433, "ex_per_s": 789.5519809628041, "data_frac": 0.8565969970056814, "mem_mb": 199.197696, "nll": -5.92513370513916, "elapsed_s": 40.5}26{"event": "train", "step": 520, "loss": -6.040017127990723, "grad_norm": 0.21397247910499573, "lr": 0.0002999287127590589, "ex_per_s": 816.803110538664, "data_frac": 0.8523816419216035, "mem_mb": 199.197696, "nll": -6.040017127990723, "elapsed_s": 40.7}27{"event": "train", "step": 540, "loss": -5.880002021789551, "grad_norm": 0.46581125259399414, "lr": 0.00029990297322826756, "ex_per_s": 843.692575626322, "data_frac": 0.8481319682742111, "mem_mb": 199.197696, "nll": -5.880002021789551, "elapsed_s": 41.0}28{"event": "train", "step": 560, "loss": -5.974058151245117, "grad_norm": 0.2900279760360718, "lr": 0.00029987327580293684, "ex_per_s": 870.3832294839875, "data_frac": 0.844008037914987, "mem_mb": 199.197696, "nll": -5.974058151245117, "elapsed_s": 41.2}29{"event": "train", "step": 580, "loss": -5.926791667938232, "grad_norm": 0.361263632774353, "lr": 0.00029983962135435424, "ex_per_s": 896.5282555755247, "data_frac": 0.8399831506828419, "mem_mb": 199.197696, "nll": -5.926791667938232, "elapsed_s": 41.4}30{"event": "train", "step": 600, "loss": -5.972969055175781, "grad_norm": 0.6762101650238037, "lr": 0.0002998020108699017, "ex_per_s": 922.6679769907239, "data_frac": 0.8359521740503075, "mem_mb": 199.197696, "nll": -5.972969055175781, "elapsed_s": 41.6}31{"event": "train", "step": 620, "loss": -5.945714473724365, "grad_norm": 1.1603693962097168, "lr": 0.0002997604454530266, "ex_per_s": 948.4872206849315, "data_frac": 0.8319168826707908, "mem_mb": 199.197696, "nll": -5.945714473724365, "elapsed_s": 41.8}32{"event": "train", "step": 640, "loss": -6.066171646118164, "grad_norm": 0.7239569425582886, "lr": 0.00029971492632320935, "ex_per_s": 974.0383568113139, "data_frac": 0.827934252794746, "mem_mb": 199.197696, "nll": -6.066171646118164, "elapsed_s": 42.1}33{"event": "train", "step": 660, "loss": -6.011338710784912, "grad_norm": 1.5237735509872437, "lr": 0.0002996654548159277, "ex_per_s": 999.3958181477273, "data_frac": 0.8240348286513638, "mem_mb": 199.197696, "nll": -6.011338710784912, "elapsed_s": 42.3}34{"event": "train", "step": 680, "loss": -6.004473686218262, "grad_norm": 0.7184116840362549, "lr": 0.00029961203238261747, "ex_per_s": 1024.437255795463, "data_frac": 0.8201391123848287, "mem_mb": 199.197696, "nll": -6.004473686218262, "elapsed_s": 42.5}35{"event": "train", "step": 700, "loss": -5.980607986450195, "grad_norm": 1.7571313381195068, "lr": 0.0002995546605906301, "ex_per_s": 1049.2699234216773, "data_frac": 0.8163096064960104, "mem_mb": 199.197696, "nll": -5.980607986450195, "elapsed_s": 42.7}36{"event": "train", "step": 720, "loss": -6.0626983642578125, "grad_norm": 0.5219307541847229, "lr": 0.00029949334112318624, "ex_per_s": 1073.8280822235604, "data_frac": 0.8125022442775001, "mem_mb": 199.197696, "nll": -6.0626983642578125, "elapsed_s": 42.9}37{"event": "train", "step": 740, "loss": -6.036559581756592, "grad_norm": 0.30032598972320557, "lr": 0.0002994280757793271, "ex_per_s": 1098.1706903678098, "data_frac": 0.8087523284519448, "mem_mb": 199.197696, "nll": -6.036559581756592, "elapsed_s": 43.1}38{"event": "train", "step": 760, "loss": -5.958896160125732, "grad_norm": 0.6852020025253296, "lr": 0.00029935886647386114, "ex_per_s": 1122.2177749944603, "data_frac": 0.8049898639193082, "mem_mb": 199.197696, "nll": -5.958896160125732, "elapsed_s": 43.3}39{"event": "train", "step": 780, "loss": -5.895640850067139, "grad_norm": 0.14441102743148804, "lr": 0.00029928571523730785, "ex_per_s": 1146.0816515058284, "data_frac": 0.8013162652823546, "mem_mb": 199.197696, "nll": -5.895640850067139, "elapsed_s": 43.6}40{"event": "train", "step": 800, "loss": -5.81829309463501, "grad_norm": 0.25993070006370544, "lr": 0.00029920862421583855, "ex_per_s": 1169.7006242975342, "data_frac": 0.7976705628225035, "mem_mb": 199.197696, "nll": -5.81829309463501, "elapsed_s": 43.8}41{"event": "train", "step": 820, "loss": -5.938891887664795, "grad_norm": 0.4973072409629822, "lr": 0.00029912759567121304, "ex_per_s": 1193.0828132542574, "data_frac": 0.794058035805752, "mem_mb": 199.197696, "nll": -5.938891887664795, "elapsed_s": 44.0}42{"event": "train", "step": 840, "loss": -5.81005334854126, "grad_norm": 0.967096209526062, "lr": 0.0002990426319807135, "ex_per_s": 1216.2498607225195, "data_frac": 0.7904811989007666, "mem_mb": 199.197696, "nll": -5.81005334854126, "elapsed_s": 44.2}43{"event": "train", "step": 860, "loss": -5.924115180969238, "grad_norm": 0.821315348148346, "lr": 0.0002989537356370747, "ex_per_s": 1239.2072884304346, "data_frac": 0.7869523872543132, "mem_mb": 199.197696, "nll": -5.924115180969238, "elapsed_s": 44.4}44{"event": "train", "step": 880, "loss": -5.87129020690918, "grad_norm": 0.4687993824481964, "lr": 0.00029886090924841085, "ex_per_s": 1261.9173908744053, "data_frac": 0.7834459601511206, "mem_mb": 199.197696, "nll": -5.87129020690918, "elapsed_s": 44.6}45{"event": "train", "step": 900, "loss": -5.988379001617432, "grad_norm": 0.6411498188972473, "lr": 0.0002987641555381389, "ex_per_s": 1284.4434589153495, "data_frac": 0.7799878943530296, "mem_mb": 199.197696, "nll": -5.988379001617432, "elapsed_s": 44.8}46{"event": "train", "step": 920, "loss": -6.028652191162109, "grad_norm": 0.15523895621299744, "lr": 0.0002986634773448989, "ex_per_s": 1306.6979383066255, "data_frac": 0.7765327099520772, "mem_mb": 199.197696, "nll": -6.028652191162109, "elapsed_s": 45.1}47{"event": "train", "step": 940, "loss": -6.078057765960693, "grad_norm": 0.9625306725502014, "lr": 0.00029855887762247066, "ex_per_s": 1328.7452207330093, "data_frac": 0.7731085222226424, "mem_mb": 199.197696, "nll": -6.078057765960693, "elapsed_s": 45.3}48{"event": "train", "step": 960, "loss": -5.968510627746582, "grad_norm": 0.4143795371055603, "lr": 0.00029845035943968696, "ex_per_s": 1350.6213152213618, "data_frac": 0.769720797613945, "mem_mb": 199.197696, "nll": -5.968510627746582, "elapsed_s": 45.5}49{"event": "train", "step": 980, "loss": -5.927549362182617, "grad_norm": 0.6391553282737732, "lr": 0.0002983379259803436, "ex_per_s": 1372.3257360555176, "data_frac": 0.7663835715340221, "mem_mb": 199.197696, "nll": -5.927549362182617, "elapsed_s": 45.7}50{"event": "train", "step": 1000, "loss": -6.069548606872559, "grad_norm": 0.21454739570617676, "lr": 0.00029822158054310604, "ex_per_s": 1393.8646020186713, "data_frac": 0.7630962388047643, "mem_mb": 199.197696, "nll": -6.069548606872559, "elapsed_s": 45.9}51{"event": "val", "step": 1000, "val_nll": -6.447678925720307, "best": -6.447678925720307, "bad": 0, "mem_mb": 199.197696, "elapsed_s": 109.4}52{"event": "train", "step": 1020, "loss": -6.026237487792969, "grad_norm": 0.21872439980506897, "lr": 0.0002981013265414125, "ex_per_s": 595.128734831091, "data_frac": 0.3195535167840106, "mem_mb": 199.197696, "nll": -6.026237487792969, "elapsed_s": 109.7}53{"event": "train", "step": 1040, "loss": -5.974891185760498, "grad_norm": 0.389641135931015, "lr": 0.000297977167503374, "ex_per_s": 605.5975638000345, "data_frac": 0.3190479247888239, "mem_mb": 199.197696, "nll": -5.974891185760498, "elapsed_s": 109.9}54{"event": "train", "step": 1060, "loss": -6.1846723556518555, "grad_norm": 0.5431748032569885, "lr": 0.00029784910707167064, "ex_per_s": 616.028735504375, "data_frac": 0.31854454103287116, "mem_mb": 199.197696, "nll": -6.1846723556518555, "elapsed_s": 110.1}55{"event": "train", "step": 1080, "loss": -5.817030429840088, "grad_norm": 0.21255825459957123, "lr": 0.0002977171490034448, "ex_per_s": 626.4248795979148, "data_frac": 0.3180435831466964, "mem_mb": 199.197696, "nll": -5.817030429840088, "elapsed_s": 110.3}56{"event": "train", "step": 1100, "loss": -6.050743103027344, "grad_norm": 0.19914235174655914, "lr": 0.000297581297170191, "ex_per_s": 636.7689938721301, "data_frac": 0.31754707041948343, "mem_mb": 199.197696, "nll": -6.050743103027344, "elapsed_s": 110.6}57{"event": "train", "step": 1120, "loss": -6.053983211517334, "grad_norm": 0.38820651173591614, "lr": 0.0002974415555576423, "ex_per_s": 647.0767150465781, "data_frac": 0.3170531394507278, "mem_mb": 199.197696, "nll": -6.053983211517334, "elapsed_s": 110.8}58{"event": "train", "step": 1140, "loss": -6.018344879150391, "grad_norm": 0.33665165305137634, "lr": 0.000297297928265653, "ex_per_s": 657.3413334641358, "data_frac": 0.3165614459642165, "mem_mb": 199.197696, "nll": -6.018344879150391, "elapsed_s": 111.0}59{"event": "train", "step": 1160, "loss": -5.963735103607178, "grad_norm": 1.3502252101898193, "lr": 0.00029715041950807914, "ex_per_s": 667.5685252692007, "data_frac": 0.3160714721761666, "mem_mb": 199.197696, "nll": -5.963735103607178, "elapsed_s": 111.2}60{"event": "train", "step": 1180, "loss": -5.957385540008545, "grad_norm": 0.5510579943656921, "lr": 0.000296999033612654, "ex_per_s": 677.773950958951, "data_frac": 0.3155861959923869, "mem_mb": 199.197696, "nll": -5.957385540008545, "elapsed_s": 111.4}61{"event": "train", "step": 1200, "loss": -5.948732376098633, "grad_norm": 0.19615575671195984, "lr": 0.00029684377502086165, "ex_per_s": 687.9232997811214, "data_frac": 0.31509722427556713, "mem_mb": 199.197696, "nll": -5.948732376098633, "elapsed_s": 111.6}62{"event": "train", "step": 1220, "loss": -6.035130977630615, "grad_norm": 0.17106734216213226, "lr": 0.00029668464828780665, "ex_per_s": 698.0371806908852, "data_frac": 0.3146063650455935, "mem_mb": 199.197696, "nll": -6.035130977630615, "elapsed_s": 111.9}63{"event": "train", "step": 1240, "loss": -6.058520317077637, "grad_norm": 0.12410221248865128, "lr": 0.00029652165808208004, "ex_per_s": 708.1205207466006, "data_frac": 0.3141250207948996, "mem_mb": 199.197696, "nll": -6.058520317077637, "elapsed_s": 112.1}64{"event": "train", "step": 1260, "loss": -6.0528669357299805, "grad_norm": 0.5712079405784607, "lr": 0.0002963548091856228, "ex_per_s": 718.1613477168502, "data_frac": 0.3136420638603939, "mem_mb": 199.197696, "nll": -6.0528669357299805, "elapsed_s": 112.3}65{"event": "train", "step": 1280, "loss": -6.047211170196533, "grad_norm": 0.6370858550071716, "lr": 0.0002961841064935852, "ex_per_s": 728.1650141126172, "data_frac": 0.31315831050367104, "mem_mb": 199.197696, "nll": -6.047211170196533, "elapsed_s": 112.5}66{"event": "train", "step": 1300, "loss": -5.8716020584106445, "grad_norm": 0.3899727165699005, "lr": 0.0002960095550141835, "ex_per_s": 738.127923377669, "data_frac": 0.3126750644949059, "mem_mb": 199.197696, "nll": -5.8716020584106445, "elapsed_s": 112.7}67{"event": "train", "step": 1320, "loss": -5.826051235198975, "grad_norm": 0.26944446563720703, "lr": 0.00029583115986855273, "ex_per_s": 748.0546528885029, "data_frac": 0.3121981191074803, "mem_mb": 199.197696, "nll": -5.826051235198975, "elapsed_s": 112.9}68{"event": "train", "step": 1340, "loss": -5.95355749130249, "grad_norm": 0.7209857702255249, "lr": 0.0002956489262905966, "ex_per_s": 757.9446559608685, "data_frac": 0.3117246924249203, "mem_mb": 199.197696, "nll": -5.95355749130249, "elapsed_s": 113.2}69{"event": "train", "step": 1360, "loss": -6.0370893478393555, "grad_norm": 0.3401136100292206, "lr": 0.0002954628596268339, "ex_per_s": 767.7909246373299, "data_frac": 0.3112570083761575, "mem_mb": 199.197696, "nll": -6.0370893478393555, "elapsed_s": 113.4}70{"event": "train", "step": 1380, "loss": -6.065449237823486, "grad_norm": 0.23736390471458435, "lr": 0.0002952729653362417, "ex_per_s": 777.591360379637, "data_frac": 0.31078489244029756, "mem_mb": 199.197696, "nll": -6.065449237823486, "elapsed_s": 113.6}71{"event": "train", "step": 1400, "loss": -5.797794818878174, "grad_norm": 0.17134438455104828, "lr": 0.00029507924899009505, "ex_per_s": 787.3514607537393, "data_frac": 0.3103166046117428, "mem_mb": 199.197696, "nll": -5.797794818878174, "elapsed_s": 113.8}72{"event": "train", "step": 1420, "loss": -5.989433288574219, "grad_norm": 0.5646085143089294, "lr": 0.00029488171627180373, "ex_per_s": 797.0843197498435, "data_frac": 0.3098481618676621, "mem_mb": 199.197696, "nll": -5.989433288574219, "elapsed_s": 114.0}73{"event": "train", "step": 1440, "loss": -5.74404239654541, "grad_norm": 0.4992274045944214, "lr": 0.00029468037297674526, "ex_per_s": 806.7735158926577, "data_frac": 0.3093806077369424, "mem_mb": 199.197696, "nll": -5.74404239654541, "elapsed_s": 114.2}74{"event": "train", "step": 1460, "loss": -5.8055596351623535, "grad_norm": 0.489360511302948, "lr": 0.00029447522501209515, "ex_per_s": 816.4300535690935, "data_frac": 0.3089159572650912, "mem_mb": 199.197696, "nll": -5.8055596351623535, "elapsed_s": 114.5}75{"event": "train", "step": 1480, "loss": -6.045531749725342, "grad_norm": 0.4943753480911255, "lr": 0.00029426627839665323, "ex_per_s": 826.0471946377417, "data_frac": 0.30845119369742074, "mem_mb": 199.197696, "nll": -6.045531749725342, "elapsed_s": 114.7}76{"event": "train", "step": 1500, "loss": -5.9613566398620605, "grad_norm": 0.3051784336566925, "lr": 0.00029405353926066754, "ex_per_s": 835.6237485367387, "data_frac": 0.30798684938471194, "mem_mb": 199.197696, "nll": -5.9613566398620605, "elapsed_s": 114.9}77{"event": "train", "step": 1520, "loss": -5.9512152671813965, "grad_norm": 0.39722442626953125, "lr": 0.000293837013845654, "ex_per_s": 845.1830629704301, "data_frac": 0.3075247256793286, "mem_mb": 199.197696, "nll": -5.9512152671813965, "elapsed_s": 115.1}78{"event": "train", "step": 1540, "loss": -5.879307746887207, "grad_norm": 0.2741452157497406, "lr": 0.00029361670850421354, "ex_per_s": 854.6940877563852, "data_frac": 0.307067572736875, "mem_mb": 199.197696, "nll": -5.879307746887207, "elapsed_s": 115.3}79{"event": "train", "step": 1560, "loss": -5.982054233551025, "grad_norm": 0.13679097592830658, "lr": 0.0002933926296998457, "ex_per_s": 864.1697259538007, "data_frac": 0.30661228758218273, "mem_mb": 199.197696, "nll": -5.982054233551025, "elapsed_s": 115.5}80{"event": "train", "step": 1580, "loss": -5.979308605194092, "grad_norm": 0.4099767804145813, "lr": 0.0002931647840067591, "ex_per_s": 873.6007070536481, "data_frac": 0.3061598585404953, "mem_mb": 199.197696, "nll": -5.979308605194092, "elapsed_s": 115.8}81{"event": "train", "step": 1600, "loss": -5.9534196853637695, "grad_norm": 0.1871674358844757, "lr": 0.0002929331781096783, "ex_per_s": 883.0117569171792, "data_frac": 0.30570886599855535, "mem_mb": 199.197696, "nll": -5.9534196853637695, "elapsed_s": 116.0}82{"event": "train", "step": 1620, "loss": -5.883004188537598, "grad_norm": 0.11423912644386292, "lr": 0.0002926978188036479, "ex_per_s": 892.3935013264671, "data_frac": 0.3052581810212004, "mem_mb": 199.197696, "nll": -5.883004188537598, "elapsed_s": 116.2}83{"event": "train", "step": 1640, "loss": -6.079347133636475, "grad_norm": 0.17040717601776123, "lr": 0.00029245871299383297, "ex_per_s": 901.731902209685, "data_frac": 0.30481000972750283, "mem_mb": 199.197696, "nll": -6.079347133636475, "elapsed_s": 116.4}84{"event": "train", "step": 1660, "loss": -6.044588088989258, "grad_norm": 0.7359273433685303, "lr": 0.0002922158676953168, "ex_per_s": 911.0376367148267, "data_frac": 0.304362454081677, "mem_mb": 199.197696, "nll": -6.044588088989258, "elapsed_s": 116.6}85{"event": "train", "step": 1680, "loss": -5.865784168243408, "grad_norm": 0.2968960404396057, "lr": 0.0002919692900328947, "ex_per_s": 920.2577334683946, "data_frac": 0.3039025142781582, "mem_mb": 199.197696, "nll": -5.865784168243408, "elapsed_s": 116.8}86{"event": "train", "step": 1700, "loss": -5.930179119110107, "grad_norm": 0.6575830578804016, "lr": 0.0002917189872408653, "ex_per_s": 929.3923615264616, "data_frac": 0.30344726824012924, "mem_mb": 199.197696, "nll": -5.930179119110107, "elapsed_s": 117.1}87{"event": "train", "step": 1720, "loss": -5.884696960449219, "grad_norm": 0.8367131948471069, "lr": 0.00029146496666281795, "ex_per_s": 938.5718455611524, "data_frac": 0.30300803150202954, "mem_mb": 199.197696, "nll": -5.884696960449219, "elapsed_s": 117.3}88{"event": "train", "step": 1740, "loss": -6.272566795349121, "grad_norm": 0.44146084785461426, "lr": 0.00029120723575141777, "ex_per_s": 947.7203797230039, "data_frac": 0.3025660382392047, "mem_mb": 199.197696, "nll": -6.272566795349121, "elapsed_s": 117.5}89{"event": "train", "step": 1760, "loss": -5.933366775512695, "grad_norm": 0.18536236882209778, "lr": 0.00029094580206818636, "ex_per_s": 956.8372455795554, "data_frac": 0.30212376104430383, "mem_mb": 199.197696, "nll": -5.933366775512695, "elapsed_s": 117.7}90{"event": "train", "step": 1780, "loss": -5.868925094604492, "grad_norm": 0.6969295740127563, "lr": 0.00029068067328328045, "ex_per_s": 965.9233654736425, "data_frac": 0.30168838145199217, "mem_mb": 199.197696, "nll": -5.868925094604492, "elapsed_s": 117.9}91{"event": "train", "step": 1800, "loss": -6.051147937774658, "grad_norm": 0.9744719862937927, "lr": 0.0002904118571752668, "ex_per_s": 974.9669279014449, "data_frac": 0.30125146603050634, "mem_mb": 199.197696, "nll": -6.051147937774658, "elapsed_s": 118.2}92{"event": "train", "step": 1820, "loss": -5.937679290771484, "grad_norm": 0.6725231409072876, "lr": 0.00029013936163089376, "ex_per_s": 983.79508481928, "data_frac": 0.30075424966516656, "mem_mb": 199.197696, "nll": -5.937679290771484, "elapsed_s": 118.4}93{"event": "train", "step": 1840, "loss": -6.073995590209961, "grad_norm": 0.4657009243965149, "lr": 0.00028986319464486, "ex_per_s": 992.7315878333098, "data_frac": 0.3003104929485855, "mem_mb": 199.197696, "nll": -6.073995590209961, "elapsed_s": 118.6}94{"event": "train", "step": 1860, "loss": -6.050817966461182, "grad_norm": 0.16184860467910767, "lr": 0.00028958336431958003, "ex_per_s": 1001.5967759208479, "data_frac": 0.29985622193038985, "mem_mb": 199.197696, "nll": -6.050817966461182, "elapsed_s": 118.9}95{"event": "train", "step": 1880, "loss": -5.945365905761719, "grad_norm": 0.193020761013031, "lr": 0.0002892998788649464, "ex_per_s": 1010.5106467930108, "data_frac": 0.2994293020513785, "mem_mb": 199.197696, "nll": -5.945365905761719, "elapsed_s": 119.1}96{"event": "train", "step": 1900, "loss": -5.986899375915527, "grad_norm": 0.22295698523521423, "lr": 0.0002890127465980889, "ex_per_s": 1019.4054979041168, "data_frac": 0.29900641199235595, "mem_mb": 199.197696, "nll": -5.986899375915527, "elapsed_s": 119.3}97{"event": "train", "step": 1920, "loss": -5.9652299880981445, "grad_norm": 0.5106670260429382, "lr": 0.0002887219759431305, "ex_per_s": 1028.2497910925263, "data_frac": 0.2985784452112738, "mem_mb": 199.197696, "nll": -5.9652299880981445, "elapsed_s": 119.5}98{"event": "train", "step": 1940, "loss": -6.048130989074707, "grad_norm": 0.32099467515945435, "lr": 0.0002884275754309401, "ex_per_s": 1037.0663158644247, "data_frac": 0.2981525379777291, "mem_mb": 199.197696, "nll": -6.048130989074707, "elapsed_s": 119.7}99{"event": "train", "step": 1960, "loss": -5.9413228034973145, "grad_norm": 0.1326863169670105, "lr": 0.0002881295536988824, "ex_per_s": 1045.8349436496433, "data_frac": 0.29772282934565414, "mem_mb": 199.197696, "nll": -5.9413228034973145, "elapsed_s": 119.9}100{"event": "train", "step": 1980, "loss": -5.960785865783691, "grad_norm": 1.2589647769927979, "lr": 0.00028782791949056447, "ex_per_s": 1054.5642723168824, "data_frac": 0.29729612468224215, "mem_mb": 199.197696, "nll": -5.960785865783691, "elapsed_s": 120.2}101{"event": "train", "step": 2000, "loss": -5.917927265167236, "grad_norm": 0.23653535544872284, "lr": 0.00028752268165557917, "ex_per_s": 1063.2823642934343, "data_frac": 0.29687554888192375, "mem_mb": 199.197696, "nll": -5.917927265167236, "elapsed_s": 120.4}102{"event": "val", "step": 2000, "val_nll": -6.449277231022185, "best": -6.449277231022185, "bad": 0, "mem_mb": 199.197696, "elapsed_s": 184.8}103{"event": "train", "step": 2020, "loss": -5.929020881652832, "grad_norm": 0.5315039157867432, "lr": 0.00028721384914924547, "ex_per_s": 698.1242156557885, "data_frac": 0.19306878145922513, "mem_mb": 199.197696, "nll": -5.929020881652832, "elapsed_s": 185.2}104{"event": "train", "step": 2040, "loss": -5.874599456787109, "grad_norm": 0.48756521940231323, "lr": 0.0002869014310323458, "ex_per_s": 704.2106756418641, "data_frac": 0.19291595541259426, "mem_mb": 199.197696, "nll": -5.874599456787109, "elapsed_s": 185.4}105{"event": "train", "step": 2060, "loss": -6.1442646980285645, "grad_norm": 0.24986663460731506, "lr": 0.0002865854364708602, "ex_per_s": 710.2790658335769, "data_frac": 0.19276152241638053, "mem_mb": 199.197696, "nll": -6.1442646980285645, "elapsed_s": 185.6}106{"event": "train", "step": 2080, "loss": -5.960216999053955, "grad_norm": 0.7432281374931335, "lr": 0.00028626587473569746, "ex_per_s": 716.3362839766221, "data_frac": 0.19260890692530802, "mem_mb": 199.197696, "nll": -5.960216999053955, "elapsed_s": 185.8}107{"event": "train", "step": 2100, "loss": -5.900568962097168, "grad_norm": 0.772300124168396, "lr": 0.0002859427552024228, "ex_per_s": 722.3737047860585, "data_frac": 0.1924584343317627, "mem_mb": 199.197696, "nll": -5.900568962097168, "elapsed_s": 186.1}108{"event": "train", "step": 2120, "loss": -6.15016508102417, "grad_norm": 0.8180380463600159, "lr": 0.0002856160873509834, "ex_per_s": 728.397714673719, "data_frac": 0.19230805669946918, "mem_mb": 199.197696, "nll": -6.15016508102417, "elapsed_s": 186.3}109{"event": "train", "step": 2140, "loss": -5.979053020477295, "grad_norm": 0.5553125739097595, "lr": 0.00028528588076542964, "ex_per_s": 734.4150424319322, "data_frac": 0.19215821243525158, "mem_mb": 199.197696, "nll": -5.979053020477295, "elapsed_s": 186.5}110{"event": "train", "step": 2160, "loss": -5.926481246948242, "grad_norm": 0.9076982140541077, "lr": 0.00028495214513363427, "ex_per_s": 740.4151350824769, "data_frac": 0.1920076834535949, "mem_mb": 199.197696, "nll": -5.926481246948242, "elapsed_s": 186.7}111{"event": "train", "step": 2180, "loss": -6.08289909362793, "grad_norm": 0.298287957906723, "lr": 0.0002846148902470082, "ex_per_s": 746.4045824039064, "data_frac": 0.19185601989127563, "mem_mb": 199.197696, "nll": -6.08289909362793, "elapsed_s": 186.9}112{"event": "train", "step": 2200, "loss": -5.9618611335754395, "grad_norm": 0.5838539004325867, "lr": 0.0002842741260002131, "ex_per_s": 752.3769782817592, "data_frac": 0.19170694818628525, "mem_mb": 199.197696, "nll": -5.9618611335754395, "elapsed_s": 187.1}113{"event": "train", "step": 2220, "loss": -5.825697422027588, "grad_norm": 1.0048506259918213, "lr": 0.00028392986239087094, "ex_per_s": 758.3311715840435, "data_frac": 0.19155825047736233, "mem_mb": 199.197696, "nll": -5.825697422027588, "elapsed_s": 187.4}114{"event": "train", "step": 2240, "loss": -5.9631853103637695, "grad_norm": 0.8368867039680481, "lr": 0.0002835821095192712, "ex_per_s": 764.2738645514717, "data_frac": 0.19141070231861623, "mem_mb": 199.197696, "nll": -5.9631853103637695, "elapsed_s": 187.6}115{"event": "train", "step": 2260, "loss": -5.733707427978516, "grad_norm": 0.8787540793418884, "lr": 0.0002832308775880741, "ex_per_s": 770.20090486399, "data_frac": 0.19126121855153247, "mem_mb": 199.197696, "nll": -5.733707427978516, "elapsed_s": 187.8}116{"event": "train", "step": 2280, "loss": -6.012693881988525, "grad_norm": 0.26425591111183167, "lr": 0.0002828761769020111, "ex_per_s": 776.0792391702036, "data_frac": 0.19110267442147175, "mem_mb": 199.197696, "nll": -6.012693881988525, "elapsed_s": 188.0}117{"event": "train", "step": 2300, "loss": -6.0013508796691895, "grad_norm": 0.19240310788154602, "lr": 0.0002825180178675832, "ex_per_s": 781.9773682410481, "data_frac": 0.1909524021841827, "mem_mb": 199.197696, "nll": -6.0013508796691895, "elapsed_s": 188.2}118{"event": "train", "step": 2320, "loss": -6.067874908447266, "grad_norm": 0.5061754584312439, "lr": 0.00028215641099275517, "ex_per_s": 787.8749359292858, "data_frac": 0.19080522282998463, "mem_mb": 199.197696, "nll": -6.067874908447266, "elapsed_s": 188.5}119{"event": "train", "step": 2340, "loss": -5.848104000091553, "grad_norm": 0.41527754068374634, "lr": 0.0002817913668866473, "ex_per_s": 793.7511102980585, "data_frac": 0.1906562089994414, "mem_mb": 199.197696, "nll": -5.848104000091553, "elapsed_s": 188.7}120{"event": "train", "step": 2360, "loss": -5.92477560043335, "grad_norm": 0.12284264713525772, "lr": 0.00028142289625922426, "ex_per_s": 799.6102999817499, "data_frac": 0.19050807430638622, "mem_mb": 199.197696, "nll": -5.92477560043335, "elapsed_s": 188.9}121{"event": "train", "step": 2380, "loss": -5.928670883178711, "grad_norm": 0.23551250994205475, "lr": 0.00028105100992098075, "ex_per_s": 805.4654047751569, "data_frac": 0.19036026968881228, "mem_mb": 199.197696, "nll": -5.928670883178711, "elapsed_s": 189.1}122{"event": "train", "step": 2400, "loss": -5.9567952156066895, "grad_norm": 0.9595536589622498, "lr": 0.00028067571878262454, "ex_per_s": 811.3010011225834, "data_frac": 0.19021374797787274, "mem_mb": 199.197696, "nll": -5.9567952156066895, "elapsed_s": 189.3}123{"event": "train", "step": 2420, "loss": -5.998198986053467, "grad_norm": 0.1389225274324417, "lr": 0.00028029703385475607, "ex_per_s": 817.1299476193726, "data_frac": 0.19006651929147278, "mem_mb": 199.197696, "nll": -5.998198986053467, "elapsed_s": 189.5}124{"event": "train", "step": 2440, "loss": -6.098771095275879, "grad_norm": 0.39147552847862244, "lr": 0.0002799149662475457, "ex_per_s": 822.9411199710494, "data_frac": 0.18992119715031225, "mem_mb": 199.197696, "nll": -6.098771095275879, "elapsed_s": 189.8}125{"event": "train", "step": 2460, "loss": -6.134539604187012, "grad_norm": 0.1990252137184143, "lr": 0.0002795295271704076, "ex_per_s": 828.7351274526067, "data_frac": 0.1897757160505173, "mem_mb": 199.197696, "nll": -6.134539604187012, "elapsed_s": 190.0}126{"event": "train", "step": 2480, "loss": -5.942187786102295, "grad_norm": 0.4262833595275879, "lr": 0.00027914072793167077, "ex_per_s": 834.4553259130495, "data_frac": 0.18962015407116395, "mem_mb": 199.197696, "nll": -5.942187786102295, "elapsed_s": 190.2}127{"event": "train", "step": 2500, "loss": -5.761035919189453, "grad_norm": 0.49565863609313965, "lr": 0.0002787485799382476, "ex_per_s": 840.2231714591306, "data_frac": 0.18947179715077214, "mem_mb": 199.197696, "nll": -5.761035919189453, "elapsed_s": 190.4}128{"event": "train", "step": 2520, "loss": -5.9579548835754395, "grad_norm": 0.7768881916999817, "lr": 0.0002783530946952989, "ex_per_s": 845.9827451859721, "data_frac": 0.18932674792018067, "mem_mb": 199.197696, "nll": -5.9579548835754395, "elapsed_s": 190.6}129{"event": "train", "step": 2540, "loss": -6.027801513671875, "grad_norm": 0.3396403193473816, "lr": 0.00027795428380589644, "ex_per_s": 851.7338677430553, "data_frac": 0.18917932291012068, "mem_mb": 199.197696, "nll": -6.027801513671875, "elapsed_s": 190.9}130{"event": "train", "step": 2560, "loss": -5.948874473571777, "grad_norm": 0.24640551209449768, "lr": 0.00027755215897068257, "ex_per_s": 857.4654685969423, "data_frac": 0.18903770888832663, "mem_mb": 199.197696, "nll": -5.948874473571777, "elapsed_s": 191.1}131{"event": "train", "step": 2580, "loss": -5.962703704833984, "grad_norm": 0.2659573554992676, "lr": 0.0002771467319875269, "ex_per_s": 863.1729119273937, "data_frac": 0.18889567050040593, "mem_mb": 199.197696, "nll": -5.962703704833984, "elapsed_s": 191.3}132{"event": "train", "step": 2600, "loss": -5.851249694824219, "grad_norm": 0.31454166769981384, "lr": 0.0002767380147511802, "ex_per_s": 868.872966161812, "data_frac": 0.1887541013559444, "mem_mb": 199.197696, "nll": -5.851249694824219, "elapsed_s": 191.5}133{"event": "train", "step": 2620, "loss": -5.887082576751709, "grad_norm": 1.025826096534729, "lr": 0.0002763260192529254, "ex_per_s": 874.5549823928978, "data_frac": 0.18860927818142376, "mem_mb": 199.197696, "nll": -5.887082576751709, "elapsed_s": 191.7}134{"event": "train", "step": 2640, "loss": -5.9860663414001465, "grad_norm": 0.537056028842926, "lr": 0.0002759107575802257, "ex_per_s": 880.2368575903967, "data_frac": 0.1884659896761587, "mem_mb": 199.197696, "nll": -5.9860663414001465, "elapsed_s": 192.0}135{"event": "train", "step": 2660, "loss": -5.820494174957275, "grad_norm": 0.17502231895923615, "lr": 0.00027549224191637014, "ex_per_s": 885.9042307100167, "data_frac": 0.18832323777833299, "mem_mb": 199.197696, "nll": -5.820494174957275, "elapsed_s": 192.2}136{"event": "train", "step": 2680, "loss": -5.960418701171875, "grad_norm": 0.4191732108592987, "lr": 0.000275070484540116, "ex_per_s": 891.5172529873904, "data_frac": 0.1881718115328964, "mem_mb": 199.197696, "nll": -5.960418701171875, "elapsed_s": 192.4}137{"event": "train", "step": 2700, "loss": -6.126676559448242, "grad_norm": 0.9769043922424316, "lr": 0.00027464549782532844, "ex_per_s": 897.1597237304155, "data_frac": 0.18802962434449166, "mem_mb": 199.197696, "nll": -6.126676559448242, "elapsed_s": 192.6}138{"event": "train", "step": 2720, "loss": -5.884420394897461, "grad_norm": 0.3094206154346466, "lr": 0.00027421729424061787, "ex_per_s": 902.7960363417176, "data_frac": 0.18788697334879687, "mem_mb": 199.197696, "nll": -5.884420394897461, "elapsed_s": 192.8}139{"event": "train", "step": 2740, "loss": -6.128833770751953, "grad_norm": 0.2123863697052002, "lr": 0.0002737858863489736, "ex_per_s": 908.4209289242101, "data_frac": 0.1877419681163985, "mem_mb": 199.197696, "nll": -6.128833770751953, "elapsed_s": 193.0}140{"event": "train", "step": 2760, "loss": -5.963802337646484, "grad_norm": 0.16129069030284882, "lr": 0.00027335128680739583, "ex_per_s": 914.0354714109216, "data_frac": 0.18760183081309525, "mem_mb": 199.197696, "nll": -5.963802337646484, "elapsed_s": 193.3}141{"event": "train", "step": 2780, "loss": -6.009857177734375, "grad_norm": 1.070116639137268, "lr": 0.00027291350836652374, "ex_per_s": 919.6394911434058, "data_frac": 0.1874592809514291, "mem_mb": 199.197696, "nll": -6.009857177734375, "elapsed_s": 193.5}142{"event": "train", "step": 2800, "loss": -5.980702877044678, "grad_norm": 0.6836847066879272, "lr": 0.00027247256387026185, "ex_per_s": 925.228720127647, "data_frac": 0.18731841559526713, "mem_mb": 199.197696, "nll": -5.980702877044678, "elapsed_s": 193.7}143{"event": "train", "step": 2820, "loss": -6.010733604431152, "grad_norm": 0.19501887261867523, "lr": 0.000272028466255403, "ex_per_s": 930.804552478699, "data_frac": 0.18717722725559732, "mem_mb": 199.197696, "nll": -6.010733604431152, "elapsed_s": 193.9}144{"event": "train", "step": 2840, "loss": -5.879223346710205, "grad_norm": 0.317897230386734, "lr": 0.0002715812285512487, "ex_per_s": 936.3673479606324, "data_frac": 0.18703621397748238, "mem_mb": 199.197696, "nll": -5.879223346710205, "elapsed_s": 194.1}145{"event": "train", "step": 2860, "loss": -6.052853584289551, "grad_norm": 0.22201384603977203, "lr": 0.0002711308638792272, "ex_per_s": 941.9183960806675, "data_frac": 0.18689525467706175, "mem_mb": 199.197696, "nll": -6.052853584289551, "elapsed_s": 194.3}146{"event": "train", "step": 2880, "loss": -5.775929927825928, "grad_norm": 0.20098692178726196, "lr": 0.000270677385452508, "ex_per_s": 947.4561352717751, "data_frac": 0.1867549898431218, "mem_mb": 199.197696, "nll": -5.775929927825928, "elapsed_s": 194.5}147{"event": "train", "step": 2900, "loss": -5.793023109436035, "grad_norm": 0.7258414626121521, "lr": 0.00027022080657561484, "ex_per_s": 952.9827426598462, "data_frac": 0.18661526230985648, "mem_mb": 199.197696, "nll": -5.793023109436035, "elapsed_s": 194.8}148{"event": "train", "step": 2920, "loss": -5.971313953399658, "grad_norm": 0.7536779642105103, "lr": 0.0002697611406440348, "ex_per_s": 958.4936843902967, "data_frac": 0.18647837700349634, "mem_mb": 199.197696, "nll": -5.971313953399658, "elapsed_s": 195.0}149{"event": "train", "step": 2940, "loss": -5.897683620452881, "grad_norm": 0.7197476625442505, "lr": 0.00026929840114382563, "ex_per_s": 963.9990318433152, "data_frac": 0.18634035140008864, "mem_mb": 199.197696, "nll": -5.897683620452881, "elapsed_s": 195.2}150{"event": "train", "step": 2960, "loss": -5.938149929046631, "grad_norm": 0.40631604194641113, "lr": 0.0002688326016512199, "ex_per_s": 969.4938200150506, "data_frac": 0.1862010190776313, "mem_mb": 199.197696, "nll": -5.938149929046631, "elapsed_s": 195.4}151{"event": "train", "step": 2980, "loss": -5.890077590942383, "grad_norm": 0.5396569967269897, "lr": 0.0002683637558322267, "ex_per_s": 974.9716817577373, "data_frac": 0.18606334001521715, "mem_mb": 199.197696, "nll": -5.890077590942383, "elapsed_s": 195.6}152{"event": "train", "step": 3000, "loss": -5.996424198150635, "grad_norm": 0.264538437128067, "lr": 0.0002678918774422309, "ex_per_s": 980.4361771806434, "data_frac": 0.18592480006848025, "mem_mb": 199.197696, "nll": -5.996424198150635, "elapsed_s": 195.8}153{"event": "val", "step": 3000, "val_nll": -6.448561137257443, "best": -6.449277231022185, "bad": 1, "mem_mb": 199.197696, "elapsed_s": 259.3}154{"event": "train", "step": 3020, "loss": -5.961047649383545, "grad_norm": 1.020127773284912, "lr": 0.0002674169803255893, "ex_per_s": 744.4672512090024, "data_frac": 0.1403030263672905, "mem_mb": 199.197696, "nll": -5.961047649383545, "elapsed_s": 259.6}155{"event": "train", "step": 3040, "loss": -5.859514236450195, "grad_norm": 0.6286898851394653, "lr": 0.0002669390784152248, "ex_per_s": 748.7669712886405, "data_frac": 0.14023896571754066, "mem_mb": 199.197696, "nll": -5.859514236450195, "elapsed_s": 259.8}156{"event": "train", "step": 3060, "loss": -6.067919731140137, "grad_norm": 0.4888870418071747, "lr": 0.00026645818573221706, "ex_per_s": 753.0636500259988, "data_frac": 0.14017406233979676, "mem_mb": 199.197696, "nll": -6.067919731140137, "elapsed_s": 260.1}157{"event": "train", "step": 3080, "loss": -5.965974807739258, "grad_norm": 0.32920658588409424, "lr": 0.00026597431638539183, "ex_per_s": 757.3534899321118, "data_frac": 0.14010836404749566, "mem_mb": 199.197696, "nll": -5.965974807739258, "elapsed_s": 260.3}158{"event": "train", "step": 3100, "loss": -6.014957904815674, "grad_norm": 0.45732447504997253, "lr": 0.00026548748457090647, "ex_per_s": 761.6408023055027, "data_frac": 0.14004229151144415, "mem_mb": 199.197696, "nll": -6.014957904815674, "elapsed_s": 260.5}159{"event": "train", "step": 3120, "loss": -6.086752891540527, "grad_norm": 1.1954331398010254, "lr": 0.00026499770457183376, "ex_per_s": 765.9122474590381, "data_frac": 0.13997661205666417, "mem_mb": 199.197696, "nll": -6.086752891540527, "elapsed_s": 260.7}160{"event": "train", "step": 3140, "loss": -6.009483814239502, "grad_norm": 0.26316094398498535, "lr": 0.0002645049907577427, "ex_per_s": 770.1790914973619, "data_frac": 0.13991111597658895, "mem_mb": 199.197696, "nll": -6.009483814239502, "elapsed_s": 260.9}161{"event": "train", "step": 3160, "loss": -6.022395610809326, "grad_norm": 0.15147849917411804, "lr": 0.00026400935758427697, "ex_per_s": 774.4470192858074, "data_frac": 0.13984404624299093, "mem_mb": 199.197696, "nll": -6.022395610809326, "elapsed_s": 261.1}162{"event": "train", "step": 3180, "loss": -6.09152889251709, "grad_norm": 0.3600750267505646, "lr": 0.00026351081959273087, "ex_per_s": 778.7091940389912, "data_frac": 0.13977807533612752, "mem_mb": 199.197696, "nll": -6.09152889251709, "elapsed_s": 261.4}163{"event": "train", "step": 3200, "loss": -6.006505966186523, "grad_norm": 0.3882271945476532, "lr": 0.0002630093914096226, "ex_per_s": 782.9587259900483, "data_frac": 0.13971207943828834, "mem_mb": 199.197696, "nll": -6.006505966186523, "elapsed_s": 261.6}164{"event": "train", "step": 3220, "loss": -5.914494514465332, "grad_norm": 0.3230733275413513, "lr": 0.00026250508774626523, "ex_per_s": 787.2036894299948, "data_frac": 0.13964568442073955, "mem_mb": 199.197696, "nll": -5.914494514465332, "elapsed_s": 261.8}165{"event": "train", "step": 3240, "loss": -5.801637172698975, "grad_norm": 0.1621481329202652, "lr": 0.00026199792339833493, "ex_per_s": 791.4391601599315, "data_frac": 0.13957918255998572, "mem_mb": 199.197696, "nll": -5.801637172698975, "elapsed_s": 262.0}166{"event": "train", "step": 3260, "loss": -5.959612846374512, "grad_norm": 0.19655068218708038, "lr": 0.00026148791324543704, "ex_per_s": 795.6714739979045, "data_frac": 0.13951353223191698, "mem_mb": 199.197696, "nll": -5.959612846374512, "elapsed_s": 262.2}167{"event": "train", "step": 3280, "loss": -6.017915725708008, "grad_norm": 0.34692245721817017, "lr": 0.0002609750722506695, "ex_per_s": 799.8966876741606, "data_frac": 0.13944701526031986, "mem_mb": 199.197696, "nll": -6.017915725708008, "elapsed_s": 262.4}168{"event": "train", "step": 3300, "loss": -5.811666488647461, "grad_norm": 0.4381527304649353, "lr": 0.0002604594154601839, "ex_per_s": 804.1137058205145, "data_frac": 0.13938391474185133, "mem_mb": 199.197696, "nll": -5.811666488647461, "elapsed_s": 262.7}169{"event": "train", "step": 3320, "loss": -5.9898762702941895, "grad_norm": 0.509434700012207, "lr": 0.0002599409580027438, "ex_per_s": 808.3219266480653, "data_frac": 0.13931891900503962, "mem_mb": 199.197696, "nll": -5.9898762702941895, "elapsed_s": 262.9}170{"event": "train", "step": 3340, "loss": -5.929775714874268, "grad_norm": 0.1633261889219284, "lr": 0.0002594197150892811, "ex_per_s": 812.5273562705607, "data_frac": 0.13925265873835146, "mem_mb": 199.197696, "nll": -5.929775714874268, "elapsed_s": 263.1}171{"event": "train", "step": 3360, "loss": -5.91081428527832, "grad_norm": 0.44515788555145264, "lr": 0.00025889570201244967, "ex_per_s": 816.7257375686983, "data_frac": 0.1391885240140244, "mem_mb": 199.197696, "nll": -5.91081428527832, "elapsed_s": 263.3}172{"event": "train", "step": 3380, "loss": -6.123254299163818, "grad_norm": 0.29455095529556274, "lr": 0.00025836893414617684, "ex_per_s": 820.9183446165366, "data_frac": 0.1391237002944648, "mem_mb": 199.197696, "nll": -6.123254299163818, "elapsed_s": 263.5}173{"event": "train", "step": 3400, "loss": -5.884012699127197, "grad_norm": 0.7265268564224243, "lr": 0.00025783942694521214, "ex_per_s": 825.0984796633096, "data_frac": 0.13905886585650068, "mem_mb": 199.197696, "nll": -5.884012699127197, "elapsed_s": 263.7}174{"event": "train", "step": 3420, "loss": -5.838011741638184, "grad_norm": 0.5353617072105408, "lr": 0.00025730719594467387, "ex_per_s": 829.2728030151536, "data_frac": 0.13899408626528922, "mem_mb": 199.197696, "nll": -5.838011741638184, "elapsed_s": 263.9}175{"event": "train", "step": 3440, "loss": -5.916312217712402, "grad_norm": 0.4432368278503418, "lr": 0.00025677225675959357, "ex_per_s": 833.4439338183878, "data_frac": 0.13892924940057602, "mem_mb": 199.197696, "nll": -5.916312217712402, "elapsed_s": 264.2}176{"event": "train", "step": 3460, "loss": -6.058072090148926, "grad_norm": 0.21169666945934296, "lr": 0.00025623462508445754, "ex_per_s": 837.6062954454264, "data_frac": 0.13886585734136223, "mem_mb": 199.197696, "nll": -6.058072090148926, "elapsed_s": 264.4}177{"event": "train", "step": 3480, "loss": -5.915340900421143, "grad_norm": 0.2067326009273529, "lr": 0.00025569431669274684, "ex_per_s": 841.7599722201257, "data_frac": 0.13880165241751521, "mem_mb": 199.197696, "nll": -5.915340900421143, "elapsed_s": 264.6}178{"event": "train", "step": 3500, "loss": -6.039490699768066, "grad_norm": 0.45337262749671936, "lr": 0.00025515134743647386, "ex_per_s": 845.9079183869452, "data_frac": 0.13873835620506741, "mem_mb": 199.197696, "nll": -6.039490699768066, "elapsed_s": 264.8}179{"event": "train", "step": 3520, "loss": -5.934995651245117, "grad_norm": 0.22336681187152863, "lr": 0.0002546057332457181, "ex_per_s": 850.0501885813684, "data_frac": 0.13867397949418095, "mem_mb": 199.197696, "nll": -5.934995651245117, "elapsed_s": 265.0}180{"event": "train", "step": 3540, "loss": -6.136158466339111, "grad_norm": 0.16638237237930298, "lr": 0.00025405749012815786, "ex_per_s": 854.1821102336661, "data_frac": 0.1386106974389412, "mem_mb": 199.197696, "nll": -6.136158466339111, "elapsed_s": 265.2}181{"event": "train", "step": 3560, "loss": -6.141542911529541, "grad_norm": 0.4907991290092468, "lr": 0.0002535066341686014, "ex_per_s": 858.3044108759296, "data_frac": 0.13854727108035464, "mem_mb": 199.197696, "nll": -6.141542911529541, "elapsed_s": 265.5}182{"event": "train", "step": 3580, "loss": -5.907487392425537, "grad_norm": 0.23731844127178192, "lr": 0.0002529531815285146, "ex_per_s": 862.4189734868986, "data_frac": 0.1384860177539167, "mem_mb": 199.197696, "nll": -5.907487392425537, "elapsed_s": 265.7}183{"event": "train", "step": 3600, "loss": -6.093670845031738, "grad_norm": 0.3130287230014801, "lr": 0.00025239714844554667, "ex_per_s": 866.5341689405566, "data_frac": 0.13842140910630685, "mem_mb": 199.197696, "nll": -6.093670845031738, "elapsed_s": 265.9}184{"event": "train", "step": 3620, "loss": -5.909344673156738, "grad_norm": 1.521618366241455, "lr": 0.0002518385512330544, "ex_per_s": 870.6431821683235, "data_frac": 0.13835754705250988, "mem_mb": 199.197696, "nll": -5.909344673156738, "elapsed_s": 266.1}185{"event": "train", "step": 3640, "loss": -6.057831764221191, "grad_norm": 0.364615261554718, "lr": 0.00025127740627962274, "ex_per_s": 874.7493475528089, "data_frac": 0.13829337685336368, "mem_mb": 199.197696, "nll": -6.057831764221191, "elapsed_s": 266.3}186{"event": "train", "step": 3660, "loss": -5.8379645347595215, "grad_norm": 0.24465550482273102, "lr": 0.00025071373004858453, "ex_per_s": 878.8441295851221, "data_frac": 0.13822791140622853, "mem_mb": 199.197696, "nll": -5.8379645347595215, "elapsed_s": 266.5}187{"event": "train", "step": 3680, "loss": -5.996062278747559, "grad_norm": 0.23887404799461365, "lr": 0.00025014753907753737, "ex_per_s": 882.9437704063648, "data_frac": 0.1381608710486588, "mem_mb": 199.197696, "nll": -5.996062278747559, "elapsed_s": 266.7}188{"event": "train", "step": 3700, "loss": -6.0686116218566895, "grad_norm": 0.8122957944869995, "lr": 0.0002495788499778582, "ex_per_s": 887.0289883131086, "data_frac": 0.13809603751466623, "mem_mb": 199.197696, "nll": -6.0686116218566895, "elapsed_s": 267.0}189{"event": "train", "step": 3720, "loss": -5.948728561401367, "grad_norm": 0.12405149638652802, "lr": 0.0002490076794342163, "ex_per_s": 891.1109442423343, "data_frac": 0.13803056167112013, "mem_mb": 199.197696, "nll": -5.948728561401367, "elapsed_s": 267.2}190{"event": "train", "step": 3740, "loss": -6.191690444946289, "grad_norm": 0.4052141010761261, "lr": 0.00024843404420408344, "ex_per_s": 895.1858982114288, "data_frac": 0.13796591236147576, "mem_mb": 199.197696, "nll": -6.191690444946289, "elapsed_s": 267.4}191{"event": "train", "step": 3760, "loss": -6.147166728973389, "grad_norm": 0.196609228849411, "lr": 0.0002478579611172425, "ex_per_s": 899.2535030835852, "data_frac": 0.13790089516669227, "mem_mb": 199.197696, "nll": -6.147166728973389, "elapsed_s": 267.6}192{"event": "train", "step": 3780, "loss": -5.962629795074463, "grad_norm": 0.1726135015487671, "lr": 0.00024727944707529366, "ex_per_s": 903.3148796389833, "data_frac": 0.13783525034578553, "mem_mb": 199.197696, "nll": -5.962629795074463, "elapsed_s": 267.8}193{"event": "train", "step": 3800, "loss": -6.096767902374268, "grad_norm": 0.15420402586460114, "lr": 0.0002466985190511583, "ex_per_s": 907.3664651901121, "data_frac": 0.13777060158995486, "mem_mb": 199.197696, "nll": -6.096767902374268, "elapsed_s": 268.0}194{"event": "train", "step": 3820, "loss": -5.973278045654297, "grad_norm": 0.6350702047348022, "lr": 0.0002461151940885812, "ex_per_s": 911.4144489973881, "data_frac": 0.13770564123893447, "mem_mb": 199.197696, "nll": -5.973278045654297, "elapsed_s": 268.2}195{"event": "train", "step": 3840, "loss": -5.87188196182251, "grad_norm": 1.055260419845581, "lr": 0.0002455294893016306, "ex_per_s": 915.4603781654941, "data_frac": 0.13764008340074513, "mem_mb": 199.197696, "nll": -5.87188196182251, "elapsed_s": 268.5}196{"event": "train", "step": 3860, "loss": -6.012255668640137, "grad_norm": 0.19942662119865417, "lr": 0.000244941421874196, "ex_per_s": 919.4950692361881, "data_frac": 0.13757603614962188, "mem_mb": 199.197696, "nll": -6.012255668640137, "elapsed_s": 268.7}197{"event": "train", "step": 3880, "loss": -5.8875226974487305, "grad_norm": 0.4666946828365326, "lr": 0.00024435100905948387, "ex_per_s": 923.5216684889025, "data_frac": 0.13751131961357577, "mem_mb": 199.197696, "nll": -5.8875226974487305, "elapsed_s": 268.9}198{"event": "train", "step": 3900, "loss": -6.183870792388916, "grad_norm": 0.6346753835678101, "lr": 0.0002437582681795117, "ex_per_s": 927.5455003767919, "data_frac": 0.1374466294101738, "mem_mb": 199.197696, "nll": -6.183870792388916, "elapsed_s": 269.1}199{"event": "train", "step": 3920, "loss": -6.037327766418457, "grad_norm": 0.4144543409347534, "lr": 0.00024316321662459957, "ex_per_s": 931.563888395931, "data_frac": 0.13738281559226384, "mem_mb": 199.197696, "nll": -6.037327766418457, "elapsed_s": 269.3}200{"event": "train", "step": 3940, "loss": -5.918807029724121, "grad_norm": 0.28282430768013, "lr": 0.00024256587185286015, "ex_per_s": 935.574345809799, "data_frac": 0.13731884289278576, "mem_mb": 199.197696, "nll": -5.918807029724121, "elapsed_s": 269.5}201{"event": "train", "step": 3960, "loss": -6.067635536193848, "grad_norm": 0.33338430523872375, "lr": 0.0002419662513896863, "ex_per_s": 939.5780605235459, "data_frac": 0.13725572808517858, "mem_mb": 199.197696, "nll": -6.067635536193848, "elapsed_s": 269.7}202{"event": "train", "step": 3980, "loss": -5.807090759277344, "grad_norm": 0.22171254456043243, "lr": 0.00024136437282723714, "ex_per_s": 943.5744452569107, "data_frac": 0.13719311372144702, "mem_mb": 199.197696, "nll": -5.807090759277344, "elapsed_s": 270.0}203{"event": "train", "step": 4000, "loss": -6.0557332038879395, "grad_norm": 0.8144536018371582, "lr": 0.0002407602538239216, "ex_per_s": 947.5585737781413, "data_frac": 0.13713173774497303, "mem_mb": 199.197696, "nll": -6.0557332038879395, "elapsed_s": 270.2}204{"event": "val", "step": 4000, "val_nll": -6.449701234277575, "best": -6.449701234277575, "bad": 0, "mem_mb": 199.197696, "elapsed_s": 333.7}205{"event": "train", "step": 4020, "loss": -5.989889621734619, "grad_norm": 0.282842755317688, "lr": 0.00024015391210388054, "ex_per_s": 770.3188508342646, "data_frac": 0.11096768324812313, "mem_mb": 199.197696, "nll": -5.989889621734619, "elapsed_s": 334.0}206{"event": "train", "step": 4040, "loss": -6.1330952644348145, "grad_norm": 0.45741981267929077, "lr": 0.00023954536545646676, "ex_per_s": 773.6462454557222, "data_frac": 0.11093561659769408, "mem_mb": 199.197696, "nll": -6.1330952644348145, "elapsed_s": 334.2}207{"event": "train", "step": 4060, "loss": -5.868185043334961, "grad_norm": 0.2092069834470749, "lr": 0.0002389346317357231, "ex_per_s": 776.9691884865912, "data_frac": 0.11090501286683999, "mem_mb": 199.197696, "nll": -5.868185043334961, "elapsed_s": 334.4}208{"event": "train", "step": 4080, "loss": -5.934335231781006, "grad_norm": 0.20174744725227356, "lr": 0.0002383217288598584, "ex_per_s": 780.2918248884297, "data_frac": 0.11087524171410496, "mem_mb": 199.197696, "nll": -5.934335231781006, "elapsed_s": 334.6}209{"event": "train", "step": 4100, "loss": -5.94481086730957, "grad_norm": 0.18288759887218475, "lr": 0.00023770667481072205, "ex_per_s": 783.6065965201993, "data_frac": 0.11084415018151776, "mem_mb": 199.197696, "nll": -5.94481086730957, "elapsed_s": 334.9}210{"event": "train", "step": 4120, "loss": -6.106508255004883, "grad_norm": 0.32424646615982056, "lr": 0.0002370894876332764, "ex_per_s": 786.9165456610689, "data_frac": 0.11081257021596216, "mem_mb": 199.197696, "nll": -6.106508255004883, "elapsed_s": 335.1}211{"event": "train", "step": 4140, "loss": -5.91709566116333, "grad_norm": 0.7348207235336304, "lr": 0.00023647018543506706, "ex_per_s": 790.2213263293271, "data_frac": 0.11078217753319414, "mem_mb": 199.197696, "nll": -5.91709566116333, "elapsed_s": 335.3}212{"event": "train", "step": 4160, "loss": -6.136982440948486, "grad_norm": 0.3663749098777771, "lr": 0.00023584878638569206, "ex_per_s": 793.5158764883837, "data_frac": 0.11075221588739659, "mem_mb": 199.197696, "nll": -6.136982440948486, "elapsed_s": 335.5}213{"event": "train", "step": 4180, "loss": -5.973779678344727, "grad_norm": 0.14537198841571808, "lr": 0.00023522530871626857, "ex_per_s": 796.8079453777098, "data_frac": 0.11072130806222524, "mem_mb": 199.197696, "nll": -5.973779678344727, "elapsed_s": 335.7}214{"event": "train", "step": 4200, "loss": -6.026877403259277, "grad_norm": 0.40490972995758057, "lr": 0.0002345997707188979, "ex_per_s": 800.1006173463164, "data_frac": 0.110691305695079, "mem_mb": 199.197696, "nll": -6.026877403259277, "elapsed_s": 336.0}215{"event": "train", "step": 4220, "loss": -5.9667510986328125, "grad_norm": 0.6357144713401794, "lr": 0.00023397219074612914, "ex_per_s": 803.388196628343, "data_frac": 0.11066139598933573, "mem_mb": 199.197696, "nll": -5.9667510986328125, "elapsed_s": 336.2}216{"event": "train", "step": 4240, "loss": -5.995819091796875, "grad_norm": 0.20145680010318756, "lr": 0.00023334258721042025, "ex_per_s": 806.6705950413966, "data_frac": 0.11063116563077292, "mem_mb": 199.197696, "nll": -5.995819091796875, "elapsed_s": 336.4}217{"event": "train", "step": 4260, "loss": -5.888527870178223, "grad_norm": 0.37102970480918884, "lr": 0.00023271097858359853, "ex_per_s": 809.9505732521322, "data_frac": 0.11060161150364002, "mem_mb": 199.197696, "nll": -5.888527870178223, "elapsed_s": 336.6}218{"event": "train", "step": 4280, "loss": -5.84810209274292, "grad_norm": 0.19046826660633087, "lr": 0.00023207738339631794, "ex_per_s": 813.2266634949916, "data_frac": 0.11057135242345535, "mem_mb": 199.197696, "nll": -5.84810209274292, "elapsed_s": 336.8}219{"event": "train", "step": 4300, "loss": -6.262007713317871, "grad_norm": 0.4452481269836426, "lr": 0.00023144182023751594, "ex_per_s": 816.4974718525782, "data_frac": 0.11054045546383726, "mem_mb": 199.197696, "nll": -6.262007713317871, "elapsed_s": 337.1}220{"event": "train", "step": 4320, "loss": -6.135000228881836, "grad_norm": 0.19161632657051086, "lr": 0.00023080430775386808, "ex_per_s": 819.7606073392692, "data_frac": 0.1105103729001945, "mem_mb": 199.197696, "nll": -6.135000228881836, "elapsed_s": 337.3}221{"event": "train", "step": 4340, "loss": -6.083557605743408, "grad_norm": 0.17831040918827057, "lr": 0.00023016486464924057, "ex_per_s": 823.0214313570928, "data_frac": 0.1104804759926762, "mem_mb": 199.197696, "nll": -6.083557605743408, "elapsed_s": 337.5}222{"event": "train", "step": 4360, "loss": -5.844722270965576, "grad_norm": 0.5800654888153076, "lr": 0.0002295235096841419, "ex_per_s": 826.2786454793405, "data_frac": 0.11045037654052832, "mem_mb": 199.197696, "nll": -5.844722270965576, "elapsed_s": 337.7}223{"event": "train", "step": 4380, "loss": -6.009580135345459, "grad_norm": 0.47670063376426697, "lr": 0.00022888026167517227, "ex_per_s": 829.5311432494325, "data_frac": 0.11042031372527217, "mem_mb": 199.197696, "nll": -6.009580135345459, "elapsed_s": 337.9}224{"event": "train", "step": 4400, "loss": -6.047191143035889, "grad_norm": 0.4869578182697296, "lr": 0.00022823513949447166, "ex_per_s": 832.7796601595019, "data_frac": 0.11039115594486786, "mem_mb": 199.197696, "nll": -6.047191143035889, "elapsed_s": 338.2}225{"event": "train", "step": 4420, "loss": -5.89776086807251, "grad_norm": 0.23263615369796753, "lr": 0.00022758816206916584, "ex_per_s": 836.024990700626, "data_frac": 0.11036167670811263, "mem_mb": 199.197696, "nll": -5.89776086807251, "elapsed_s": 338.4}226{"event": "train", "step": 4440, "loss": -6.0906524658203125, "grad_norm": 0.2324923872947693, "lr": 0.00022693934838081143, "ex_per_s": 839.2665063201508, "data_frac": 0.11033271258773765, "mem_mb": 199.197696, "nll": -6.0906524658203125, "elapsed_s": 338.6}227{"event": "train", "step": 4460, "loss": -5.9326605796813965, "grad_norm": 0.19416509568691254, "lr": 0.0002262887174648388, "ex_per_s": 842.5023404517361, "data_frac": 0.11030353248549937, "mem_mb": 199.197696, "nll": -5.9326605796813965, "elapsed_s": 338.8}228{"event": "train", "step": 4480, "loss": -6.057425022125244, "grad_norm": 0.3755660653114319, "lr": 0.00022563628840999367, "ex_per_s": 845.7347564166934, "data_frac": 0.11027370152130817, "mem_mb": 199.197696, "nll": -6.057425022125244, "elapsed_s": 339.0}229{"event": "train", "step": 4500, "loss": -6.153731822967529, "grad_norm": 0.6001710295677185, "lr": 0.000224982080357777, "ex_per_s": 848.9611257459785, "data_frac": 0.11024456921465728, "mem_mb": 199.197696, "nll": -6.153731822967529, "elapsed_s": 339.2}230{"event": "train", "step": 4520, "loss": -6.019184589385986, "grad_norm": 0.1605006754398346, "lr": 0.00022432611250188336, "ex_per_s": 852.1857661183843, "data_frac": 0.11021532730382975, "mem_mb": 199.197696, "nll": -6.019184589385986, "elapsed_s": 339.5}231{"event": "train", "step": 4540, "loss": -6.0103678703308105, "grad_norm": 0.17642882466316223, "lr": 0.00022366840408763816, "ex_per_s": 855.4071186555157, "data_frac": 0.11018533717709106, "mem_mb": 199.197696, "nll": -6.0103678703308105, "elapsed_s": 339.7}232{"event": "train", "step": 4560, "loss": -6.047405242919922, "grad_norm": 0.39461854100227356, "lr": 0.00022300897441143247, "ex_per_s": 858.6224457297609, "data_frac": 0.1101559437249959, "mem_mb": 199.197696, "nll": -6.047405242919922, "elapsed_s": 339.9}233{"event": "train", "step": 4580, "loss": -6.021829605102539, "grad_norm": 0.6774302124977112, "lr": 0.0002223478428201573, "ex_per_s": 861.8334649633683, "data_frac": 0.1101264462898116, "mem_mb": 199.197696, "nll": -6.021829605102539, "elapsed_s": 340.1}234{"event": "train", "step": 4600, "loss": -5.860202312469482, "grad_norm": 0.2193460911512375, "lr": 0.0002216850287106357, "ex_per_s": 865.0419619783416, "data_frac": 0.11009704281614559, "mem_mb": 199.197696, "nll": -5.860202312469482, "elapsed_s": 340.3}235{"event": "train", "step": 4620, "loss": -6.031363487243652, "grad_norm": 0.17533130943775177, "lr": 0.00022102055152905406, "ex_per_s": 868.2468835781235, "data_frac": 0.11006711139504265, "mem_mb": 199.197696, "nll": -6.031363487243652, "elapsed_s": 340.6}236{"event": "train", "step": 4640, "loss": -5.835496425628662, "grad_norm": 0.23643207550048828, "lr": 0.00022035443077039114, "ex_per_s": 871.4449574371687, "data_frac": 0.11003831544323195, "mem_mb": 199.197696, "nll": -5.835496425628662, "elapsed_s": 340.8}237{"event": "train", "step": 4660, "loss": -5.978397846221924, "grad_norm": 0.24769409000873566, "lr": 0.00021968668597784646, "ex_per_s": 874.6392346622903, "data_frac": 0.1100090972647236, "mem_mb": 199.197696, "nll": -5.978397846221924, "elapsed_s": 341.0}238{"event": "train", "step": 4680, "loss": -6.02909517288208, "grad_norm": 0.4418308138847351, "lr": 0.00021901733674226665, "ex_per_s": 877.8315741237905, "data_frac": 0.1099794439231093, "mem_mb": 199.197696, "nll": -6.02909517288208, "elapsed_s": 341.2}239{"event": "train", "step": 4700, "loss": -5.911900043487549, "grad_norm": 0.26847365498542786, "lr": 0.00021834640270157097, "ex_per_s": 881.0186716582359, "data_frac": 0.10994979640373513, "mem_mb": 199.197696, "nll": -5.911900043487549, "elapsed_s": 341.4}240{"event": "train", "step": 4720, "loss": -6.0656538009643555, "grad_norm": 0.18242809176445007, "lr": 0.00021767390354017485, "ex_per_s": 884.1983891064997, "data_frac": 0.10992041899882768, "mem_mb": 199.197696, "nll": -6.0656538009643555, "elapsed_s": 341.6}241{"event": "train", "step": 4740, "loss": -6.041344165802002, "grad_norm": 0.27594324946403503, "lr": 0.0002169998589884126, "ex_per_s": 887.3792627697986, "data_frac": 0.10989181266242298, "mem_mb": 199.197696, "nll": -6.041344165802002, "elapsed_s": 341.9}242{"event": "train", "step": 4760, "loss": -5.997693061828613, "grad_norm": 0.4585180878639221, "lr": 0.00021632428882195843, "ex_per_s": 890.5533864231962, "data_frac": 0.10986330910900022, "mem_mb": 199.197696, "nll": -5.997693061828613, "elapsed_s": 342.1}243{"event": "train", "step": 4780, "loss": -6.061470031738281, "grad_norm": 1.0809181928634644, "lr": 0.00021564721286124628, "ex_per_s": 893.7226415373145, "data_frac": 0.10983459480997809, "mem_mb": 199.197696, "nll": -6.061470031738281, "elapsed_s": 342.3}244{"event": "train", "step": 4800, "loss": -6.114729881286621, "grad_norm": 0.4510861039161682, "lr": 0.00021496865097088845, "ex_per_s": 896.8881500939917, "data_frac": 0.10980573304166721, "mem_mb": 199.197696, "nll": -6.114729881286621, "elapsed_s": 342.5}245{"event": "train", "step": 4820, "loss": -6.036983489990234, "grad_norm": 0.173771470785141, "lr": 0.00021428862305909247, "ex_per_s": 900.047761867558, "data_frac": 0.10977668660460424, "mem_mb": 199.197696, "nll": -6.036983489990234, "elapsed_s": 342.7}246{"event": "train", "step": 4840, "loss": -5.9347310066223145, "grad_norm": 0.19823427498340607, "lr": 0.00021360714907707733, "ex_per_s": 903.2072930653912, "data_frac": 0.10974794893313022, "mem_mb": 199.197696, "nll": -5.9347310066223145, "elapsed_s": 343.0}247{"event": "train", "step": 4860, "loss": -5.926506519317627, "grad_norm": 0.8342258930206299, "lr": 0.00021292424901848804, "ex_per_s": 906.3607618158095, "data_frac": 0.1097192338136935, "mem_mb": 199.197696, "nll": -5.926506519317627, "elapsed_s": 343.2}248{"event": "train", "step": 4880, "loss": -6.026139259338379, "grad_norm": 0.5201253890991211, "lr": 0.00021223994291880895, "ex_per_s": 909.5119270326395, "data_frac": 0.10968865263992097, "mem_mb": 199.197696, "nll": -6.026139259338379, "elapsed_s": 343.4}249{"event": "train", "step": 4900, "loss": -5.958730220794678, "grad_norm": 0.14442573487758636, "lr": 0.00021155425085477597, "ex_per_s": 912.6582449952409, "data_frac": 0.10966063817524134, "mem_mb": 199.197696, "nll": -5.958730220794678, "elapsed_s": 343.6}250{"event": "train", "step": 4920, "loss": -5.945367813110352, "grad_norm": 0.3662509024143219, "lr": 0.00021086719294378776, "ex_per_s": 915.7963690024665, "data_frac": 0.10963154108305252, "mem_mb": 199.197696, "nll": -5.945367813110352, "elapsed_s": 343.8}251{"event": "train", "step": 4940, "loss": -6.047619819641113, "grad_norm": 0.20515312254428864, "lr": 0.00021017878934331512, "ex_per_s": 918.9291551093272, "data_frac": 0.10960399385817937, "mem_mb": 199.197696, "nll": -6.047619819641113, "elapsed_s": 344.1}252{"event": "train", "step": 4960, "loss": -6.069591999053955, "grad_norm": 0.16597281396389008, "lr": 0.00020948906025030994, "ex_per_s": 922.0590763873911, "data_frac": 0.10957566947564863, "mem_mb": 199.197696, "nll": -6.069591999053955, "elapsed_s": 344.3}253{"event": "train", "step": 4980, "loss": -6.071577548980713, "grad_norm": 0.17931872606277466, "lr": 0.00020879802590061233, "ex_per_s": 925.1901795396162, "data_frac": 0.10954454452927564, "mem_mb": 199.197696, "nll": -6.071577548980713, "elapsed_s": 344.5}254{"event": "train", "step": 5000, "loss": -5.889564037322998, "grad_norm": 0.20179662108421326, "lr": 0.00020810570656835728, "ex_per_s": 928.3195883488758, "data_frac": 0.10951625583761457, "mem_mb": 199.197696, "nll": -5.889564037322998, "elapsed_s": 344.7}255{"event": "val", "step": 5000, "val_nll": -6.45052985700391, "best": -6.45052985700391, "bad": 0, "mem_mb": 199.197696, "elapsed_s": 408.8}256{"event": "train", "step": 5020, "loss": -5.967226028442383, "grad_norm": 0.19862742722034454, "lr": 0.0002074121225653795, "ex_per_s": 785.3885078829056, "data_frac": 0.09231825477813842, "mem_mb": 199.197696, "nll": -5.967226028442383, "elapsed_s": 409.1}257{"event": "train", "step": 5040, "loss": -6.091768741607666, "grad_norm": 0.6416983604431152, "lr": 0.00020671729424061788, "ex_per_s": 788.1003183823706, "data_frac": 0.09230199096612442, "mem_mb": 199.197696, "nll": -6.091768741607666, "elapsed_s": 409.3}258{"event": "train", "step": 5060, "loss": -6.009261608123779, "grad_norm": 0.19142867624759674, "lr": 0.00020602124197951796, "ex_per_s": 790.8094401685858, "data_frac": 0.09228553999477708, "mem_mb": 199.197696, "nll": -6.009261608123779, "elapsed_s": 409.5}259{"event": "train", "step": 5080, "loss": -6.132133483886719, "grad_norm": 0.507061243057251, "lr": 0.00020532398620343433, "ex_per_s": 793.5152855366069, "data_frac": 0.09226918049901689, "mem_mb": 199.197696, "nll": -6.132133483886719, "elapsed_s": 409.7}260{"event": "train", "step": 5100, "loss": -6.061741352081299, "grad_norm": 0.2086382806301117, "lr": 0.0002046255473690312, "ex_per_s": 796.2161022439683, "data_frac": 0.0922537334151952, "mem_mb": 199.197696, "nll": -6.061741352081299, "elapsed_s": 409.9}261{"event": "train", "step": 5120, "loss": -6.005159378051758, "grad_norm": 0.27760589122772217, "lr": 0.00020392594596768237, "ex_per_s": 798.9149129433333, "data_frac": 0.09223714439826784, "mem_mb": 199.197696, "nll": -6.005159378051758, "elapsed_s": 410.2}262{"event": "train", "step": 5140, "loss": -5.947878360748291, "grad_norm": 0.7969412803649902, "lr": 0.00020322520252486997, "ex_per_s": 801.612387154856, "data_frac": 0.09222195363631097, "mem_mb": 199.197696, "nll": -5.947878360748291, "elapsed_s": 410.4}263{"event": "train", "step": 5160, "loss": -6.164603233337402, "grad_norm": 0.395868718624115, "lr": 0.0002025233375995822, "ex_per_s": 804.3079870423772, "data_frac": 0.09220644024507853, "mem_mb": 199.197696, "nll": -6.164603233337402, "elapsed_s": 410.6}264{"event": "train", "step": 5180, "loss": -5.940816879272461, "grad_norm": 0.2955150902271271, "lr": 0.0002018203717837105, "ex_per_s": 806.9922786262002, "data_frac": 0.09219210637584517, "mem_mb": 199.197696, "nll": -5.940816879272461, "elapsed_s": 410.8}265{"event": "train", "step": 5200, "loss": -5.826324939727783, "grad_norm": 0.5863794684410095, "lr": 0.00020111632570144484, "ex_per_s": 809.6790888112857, "data_frac": 0.09217704384520754, "mem_mb": 199.197696, "nll": -5.826324939727783, "elapsed_s": 411.0}266{"event": "train", "step": 5220, "loss": -5.9274725914001465, "grad_norm": 0.14115136861801147, "lr": 0.00020041122000866903, "ex_per_s": 812.3619721687544, "data_frac": 0.09216170281612875, "mem_mb": 199.197696, "nll": -5.9274725914001465, "elapsed_s": 411.3}267{"event": "train", "step": 5240, "loss": -5.7769880294799805, "grad_norm": 0.16782984137535095, "lr": 0.00019970507539235469, "ex_per_s": 815.0422503522229, "data_frac": 0.09214772288352363, "mem_mb": 199.197696, "nll": -5.7769880294799805, "elapsed_s": 411.5}268{"event": "train", "step": 5260, "loss": -5.951096534729004, "grad_norm": 0.18954667448997498, "lr": 0.0001989979125699542, "ex_per_s": 817.7157335961182, "data_frac": 0.09213215096310351, "mem_mb": 199.197696, "nll": -5.951096534729004, "elapsed_s": 411.7}269{"event": "train", "step": 5280, "loss": -6.004115581512451, "grad_norm": 1.07816743850708, "lr": 0.00019828975228879288, "ex_per_s": 820.3924368001315, "data_frac": 0.09211694692331172, "mem_mb": 199.197696, "nll": -6.004115581512451, "elapsed_s": 411.9}270{"event": "train", "step": 5300, "loss": -5.892044544219971, "grad_norm": 0.4070219397544861, "lr": 0.00019758061532546025, "ex_per_s": 823.0651728257102, "data_frac": 0.09210288177111374, "mem_mb": 199.197696, "nll": -5.892044544219971, "elapsed_s": 412.1}271{"event": "train", "step": 5320, "loss": -6.076975345611572, "grad_norm": 0.4351007342338562, "lr": 0.0001968705224852007, "ex_per_s": 825.7392527486525, "data_frac": 0.09208717214723297, "mem_mb": 199.197696, "nll": -6.076975345611572, "elapsed_s": 412.3}272{"event": "train", "step": 5340, "loss": -5.972102642059326, "grad_norm": 0.5540052056312561, "lr": 0.00019615949460130276, "ex_per_s": 828.4069007182268, "data_frac": 0.09207038052581759, "mem_mb": 199.197696, "nll": -5.972102642059326, "elapsed_s": 412.6}273{"event": "train", "step": 5360, "loss": -6.012178421020508, "grad_norm": 0.13035671412944794, "lr": 0.00019544755253448823, "ex_per_s": 831.0675315724848, "data_frac": 0.09205686951622405, "mem_mb": 199.197696, "nll": -6.012178421020508, "elapsed_s": 412.8}274{"event": "train", "step": 5380, "loss": -5.863175392150879, "grad_norm": 0.36855894327163696, "lr": 0.0001947347171722997, "ex_per_s": 833.7280844801713, "data_frac": 0.09204131695143421, "mem_mb": 199.197696, "nll": -5.863175392150879, "elapsed_s": 413.0}275{"event": "train", "step": 5400, "loss": -5.959009170532227, "grad_norm": 0.30694296956062317, "lr": 0.00019402100942848827, "ex_per_s": 836.38587997097, "data_frac": 0.09202645626973864, "mem_mb": 199.197696, "nll": -5.959009170532227, "elapsed_s": 413.2}276{"event": "train", "step": 5420, "loss": -5.8375983238220215, "grad_norm": 0.6624240279197693, "lr": 0.00019330645024239946, "ex_per_s": 839.0434128837536, "data_frac": 0.09201014135849878, "mem_mb": 199.197696, "nll": -5.8375983238220215, "elapsed_s": 413.4}277{"event": "train", "step": 5440, "loss": -6.02055025100708, "grad_norm": 0.5293865203857422, "lr": 0.00019259106057835926, "ex_per_s": 841.6983145202693, "data_frac": 0.0919943538011387, "mem_mb": 199.197696, "nll": -6.02055025100708, "elapsed_s": 413.6}278{"event": "train", "step": 5460, "loss": -6.102335453033447, "grad_norm": 0.20912352204322815, "lr": 0.00019187486142505883, "ex_per_s": 844.3527725672517, "data_frac": 0.09197832173029319, "mem_mb": 199.197696, "nll": -6.102335453033447, "elapsed_s": 413.9}279{"event": "train", "step": 5480, "loss": -6.100974082946777, "grad_norm": 0.1842484325170517, "lr": 0.00019115787379493887, "ex_per_s": 847.0068536381234, "data_frac": 0.09196213940833607, "mem_mb": 199.197696, "nll": -6.100974082946777, "elapsed_s": 414.1}280{"event": "train", "step": 5500, "loss": -6.082758903503418, "grad_norm": 0.5675257444381714, "lr": 0.00019044011872357296, "ex_per_s": 849.6575667005371, "data_frac": 0.0919472014971666, "mem_mb": 199.197696, "nll": -6.082758903503418, "elapsed_s": 414.3}281{"event": "train", "step": 5520, "loss": -5.926143646240234, "grad_norm": 0.1556575745344162, "lr": 0.00018972161726905055, "ex_per_s": 852.3042444496497, "data_frac": 0.09193087915966565, "mem_mb": 199.197696, "nll": -5.926143646240234, "elapsed_s": 414.5}282{"event": "train", "step": 5540, "loss": -5.919712066650391, "grad_norm": 0.6695707440376282, "lr": 0.00018900239051135913, "ex_per_s": 854.9491873171429, "data_frac": 0.09191453843144946, "mem_mb": 199.197696, "nll": -5.919712066650391, "elapsed_s": 414.7}283{"event": "train", "step": 5560, "loss": -5.977831840515137, "grad_norm": 0.6847689151763916, "lr": 0.00018828245955176568, "ex_per_s": 857.5894532332995, "data_frac": 0.0918992584201236, "mem_mb": 199.197696, "nll": -5.977831840515137, "elapsed_s": 414.9}284{"event": "train", "step": 5580, "loss": -6.031119346618652, "grad_norm": 0.21894977986812592, "lr": 0.00018756184551219765, "ex_per_s": 860.2255656425174, "data_frac": 0.09188430093756149, "mem_mb": 199.197696, "nll": -6.031119346618652, "elapsed_s": 415.2}285{"event": "train", "step": 5600, "loss": -6.08457612991333, "grad_norm": 0.3729437291622162, "lr": 0.00018684056953462326, "ex_per_s": 862.8562214522749, "data_frac": 0.09187005155830331, "mem_mb": 199.197696, "nll": -6.08457612991333, "elapsed_s": 415.4}286{"event": "train", "step": 5620, "loss": -5.821881294250488, "grad_norm": 0.19570551812648773, "lr": 0.00018611865278043117, "ex_per_s": 865.486039043637, "data_frac": 0.09185509822756277, "mem_mb": 199.197696, "nll": -5.821881294250488, "elapsed_s": 415.6}287{"event": "train", "step": 5640, "loss": -6.1180572509765625, "grad_norm": 0.8967223167419434, "lr": 0.00018539611642980975, "ex_per_s": 868.1158167288311, "data_frac": 0.09184012591214022, "mem_mb": 199.197696, "nll": -6.1180572509765625, "elapsed_s": 415.8}288{"event": "train", "step": 5660, "loss": -6.009692668914795, "grad_norm": 0.24035488069057465, "lr": 0.00018467298168112556, "ex_per_s": 870.7424302562645, "data_frac": 0.09182421901600978, "mem_mb": 199.197696, "nll": -6.009692668914795, "elapsed_s": 416.0}289{"event": "train", "step": 5680, "loss": -5.9685821533203125, "grad_norm": 0.13926005363464355, "lr": 0.00018394926975030145, "ex_per_s": 873.3653828061981, "data_frac": 0.09180919060320945, "mem_mb": 199.197696, "nll": -5.9685821533203125, "elapsed_s": 416.2}290{"event": "train", "step": 5700, "loss": -6.038414478302002, "grad_norm": 0.27770712971687317, "lr": 0.00018322500187019415, "ex_per_s": 875.9862223004868, "data_frac": 0.09179382381954516, "mem_mb": 199.197696, "nll": -6.038414478302002, "elapsed_s": 416.5}291{"event": "train", "step": 5720, "loss": -5.836523532867432, "grad_norm": 0.41704556345939636, "lr": 0.00018250019928997124, "ex_per_s": 878.6015895952917, "data_frac": 0.09177957038767545, "mem_mb": 199.197696, "nll": -5.836523532867432, "elapsed_s": 416.7}292{"event": "train", "step": 5740, "loss": -6.003908157348633, "grad_norm": 0.29748138785362244, "lr": 0.00018177488327448776, "ex_per_s": 881.2164061104003, "data_frac": 0.09176481821166471, "mem_mb": 199.197696, "nll": -6.003908157348633, "elapsed_s": 416.9}293{"event": "train", "step": 5760, "loss": -5.997725009918213, "grad_norm": 0.43733716011047363, "lr": 0.00018104907510366235, "ex_per_s": 883.829760436325, "data_frac": 0.09174964673477906, "mem_mb": 199.197696, "nll": -5.997725009918213, "elapsed_s": 417.1}294{"event": "train", "step": 5780, "loss": -6.009850978851318, "grad_norm": 0.24805018305778503, "lr": 0.00018032279607185304, "ex_per_s": 886.4385244379488, "data_frac": 0.09173389884272387, "mem_mb": 199.197696, "nll": -6.009850978851318, "elapsed_s": 417.3}295{"event": "train", "step": 5800, "loss": -5.985660076141357, "grad_norm": 0.6498529314994812, "lr": 0.00017959606748723213, "ex_per_s": 889.0477063678189, "data_frac": 0.09171911932708032, "mem_mb": 199.197696, "nll": -5.985660076141357, "elapsed_s": 417.5}296{"event": "train", "step": 5820, "loss": -5.910436153411865, "grad_norm": 0.3800836503505707, "lr": 0.00017886891067116147, "ex_per_s": 891.6486843797111, "data_frac": 0.09170418018256546, "mem_mb": 199.197696, "nll": -5.910436153411865, "elapsed_s": 417.7}297{"event": "train", "step": 5840, "loss": -5.919267654418945, "grad_norm": 0.5849001407623291, "lr": 0.00017814134695756664, "ex_per_s": 894.2489251545773, "data_frac": 0.09168954539729174, "mem_mb": 199.197696, "nll": -5.919267654418945, "elapsed_s": 418.0}298{"event": "train", "step": 5860, "loss": -6.008349895477295, "grad_norm": 0.23204055428504944, "lr": 0.00017741339769231104, "ex_per_s": 896.8507071045218, "data_frac": 0.09167430172142557, "mem_mb": 199.197696, "nll": -6.008349895477295, "elapsed_s": 418.2}299{"event": "train", "step": 5880, "loss": -6.012391090393066, "grad_norm": 0.34141355752944946, "lr": 0.00017668508423256986, "ex_per_s": 899.4462754669354, "data_frac": 0.09166048369721413, "mem_mb": 199.197696, "nll": -6.012391090393066, "elapsed_s": 418.4}300{"event": "train", "step": 5900, "loss": -6.143407344818115, "grad_norm": 0.4452528655529022, "lr": 0.0001759564279462032, "ex_per_s": 902.038359352768, "data_frac": 0.09164585298532153, "mem_mb": 199.197696, "nll": -6.143407344818115, "elapsed_s": 418.6}301{"event": "train", "step": 5920, "loss": -6.002052307128906, "grad_norm": 0.4386497735977173, "lr": 0.00017522745021112944, "ex_per_s": 904.6271756107683, "data_frac": 0.09163177701743241, "mem_mb": 199.197696, "nll": -6.002052307128906, "elapsed_s": 418.8}302{"event": "train", "step": 5940, "loss": -5.939102649688721, "grad_norm": 0.2270265519618988, "lr": 0.0001744981724146976, "ex_per_s": 907.2127454604542, "data_frac": 0.09161709812183264, "mem_mb": 199.197696, "nll": -5.939102649688721, "elapsed_s": 419.0}303{"event": "train", "step": 5960, "loss": -5.905972480773926, "grad_norm": 0.6783722043037415, "lr": 0.00017376861595306043, "ex_per_s": 909.7983259727775, "data_frac": 0.0916027041772964, "mem_mb": 199.197696, "nll": -5.905972480773926, "elapsed_s": 419.3}304{"event": "train", "step": 5980, "loss": -6.030241012573242, "grad_norm": 0.8015355467796326, "lr": 0.00017303880223054632, "ex_per_s": 912.3795430539685, "data_frac": 0.09158837318885828, "mem_mb": 199.197696, "nll": -6.030241012573242, "elapsed_s": 419.5}305{"event": "train", "step": 6000, "loss": -5.991155624389648, "grad_norm": 0.19305042922496796, "lr": 0.00017230875265903138, "ex_per_s": 914.9568043366769, "data_frac": 0.09157425804781824, "mem_mb": 199.197696, "nll": -5.991155624389648, "elapsed_s": 419.7}306{"event": "val", "step": 6000, "val_nll": -6.451241575266094, "best": -6.451241575266094, "bad": 0, "mem_mb": 199.197696, "elapsed_s": 483.0}307{"event": "train", "step": 6020, "loss": -5.8484368324279785, "grad_norm": 0.2519277036190033, "lr": 0.0001715784886573111, "ex_per_s": 797.244510335618, "data_frac": 0.0795554719787698, "mem_mb": 199.197696, "nll": -5.8484368324279785, "elapsed_s": 483.3}308{"event": "train", "step": 6040, "loss": -5.863344192504883, "grad_norm": 0.5574017763137817, "lr": 0.0001708480316504724, "ex_per_s": 799.5353083772995, "data_frac": 0.07954688336526096, "mem_mb": 199.197696, "nll": -5.863344192504883, "elapsed_s": 483.5}309{"event": "train", "step": 6060, "loss": -6.026839256286621, "grad_norm": 0.4049450755119324, "lr": 0.00017011740306926445, "ex_per_s": 801.8252642562778, "data_frac": 0.0795386907621807, "mem_mb": 199.197696, "nll": -6.026839256286621, "elapsed_s": 483.7}310{"event": "train", "step": 6080, "loss": -6.008134841918945, "grad_norm": 0.34801432490348816, "lr": 0.00016938662434947037, "ex_per_s": 804.112815412759, "data_frac": 0.07953135575801, "mem_mb": 199.197696, "nll": -6.008134841918945, "elapsed_s": 483.9}311{"event": "train", "step": 6100, "loss": -6.141883850097656, "grad_norm": 0.4427778720855713, "lr": 0.00016865571693127818, "ex_per_s": 806.3990758950739, "data_frac": 0.07952310681727726, "mem_mb": 199.197696, "nll": -6.141883850097656, "elapsed_s": 484.1}312{"event": "train", "step": 6120, "loss": -5.9950385093688965, "grad_norm": 0.3609658479690552, "lr": 0.00016792470225865166, "ex_per_s": 808.6832763464618, "data_frac": 0.07951498167485511, "mem_mb": 199.197696, "nll": -5.9950385093688965, "elapsed_s": 484.3}313{"event": "train", "step": 6140, "loss": -6.00015926361084, "grad_norm": 0.2431037575006485, "lr": 0.00016719360177870138, "ex_per_s": 810.9646898734712, "data_frac": 0.07950755026535886, "mem_mb": 199.197696, "nll": -6.00015926361084, "elapsed_s": 484.6}314{"event": "train", "step": 6160, "loss": -6.0664496421813965, "grad_norm": 0.2501370906829834, "lr": 0.0001664624369410554, "ex_per_s": 813.2445925014736, "data_frac": 0.07949937271699573, "mem_mb": 199.197696, "nll": -6.0664496421813965, "elapsed_s": 484.8}315{"event": "train", "step": 6180, "loss": -6.044407844543457, "grad_norm": 0.1963062286376953, "lr": 0.00016573122919722992, "ex_per_s": 815.5222369137572, "data_frac": 0.07949073196690801, "mem_mb": 199.197696, "nll": -6.044407844543457, "elapsed_s": 485.0}316{"event": "train", "step": 6200, "loss": -5.987377643585205, "grad_norm": 0.28203123807907104, "lr": 0.000165, "ex_per_s": 817.79661112351, "data_frac": 0.07948307752381933, "mem_mb": 199.197696, "nll": -5.987377643585205, "elapsed_s": 485.2}317{"event": "train", "step": 6220, "loss": -5.930910110473633, "grad_norm": 0.21562498807907104, "lr": 0.00016426877080277007, "ex_per_s": 820.0709043904842, "data_frac": 0.07947513526133909, "mem_mb": 199.197696, "nll": -5.930910110473633, "elapsed_s": 485.4}318{"event": "train", "step": 6240, "loss": -5.902710914611816, "grad_norm": 0.35176149010658264, "lr": 0.0001635375630589446, "ex_per_s": 822.34144057461, "data_frac": 0.0794667784484161, "mem_mb": 199.197696, "nll": -5.902710914611816, "elapsed_s": 485.6}319{"event": "train", "step": 6260, "loss": -5.907347679138184, "grad_norm": 0.14377400279045105, "lr": 0.00016280639822129864, "ex_per_s": 824.6119581748169, "data_frac": 0.07945850230989361, "mem_mb": 199.197696, "nll": -5.907347679138184, "elapsed_s": 485.9}320{"event": "train", "step": 6280, "loss": -5.905317783355713, "grad_norm": 0.5052511096000671, "lr": 0.00016207529774134834, "ex_per_s": 826.8803512999182, "data_frac": 0.07945015375697949, "mem_mb": 199.197696, "nll": -5.905317783355713, "elapsed_s": 486.1}321{"event": "train", "step": 6300, "loss": -5.982355117797852, "grad_norm": 0.3888099789619446, "lr": 0.00016134428306872187, "ex_per_s": 829.142936325353, "data_frac": 0.0794408684088302, "mem_mb": 199.197696, "nll": -5.982355117797852, "elapsed_s": 486.3}322{"event": "train", "step": 6320, "loss": -5.82638692855835, "grad_norm": 0.30647698044776917, "lr": 0.00016061337565052966, "ex_per_s": 831.4057212996075, "data_frac": 0.079431696791444, "mem_mb": 199.197696, "nll": -5.82638692855835, "elapsed_s": 486.5}323{"event": "train", "step": 6340, "loss": -5.885511875152588, "grad_norm": 0.2619370222091675, "lr": 0.00015988259693073558, "ex_per_s": 833.669015259799, "data_frac": 0.07942216970982674, "mem_mb": 199.197696, "nll": -5.885511875152588, "elapsed_s": 486.7}324{"event": "train", "step": 6360, "loss": -6.167451858520508, "grad_norm": 0.15096226334571838, "lr": 0.0001591519683495276, "ex_per_s": 835.9310129506559, "data_frac": 0.07941407809827651, "mem_mb": 199.197696, "nll": -6.167451858520508, "elapsed_s": 486.9}325{"event": "train", "step": 6380, "loss": -5.899844646453857, "grad_norm": 0.28817102313041687, "lr": 0.0001584215113426889, "ex_per_s": 838.1901899597011, "data_frac": 0.07940561138658078, "mem_mb": 199.197696, "nll": -5.899844646453857, "elapsed_s": 487.2}326{"event": "train", "step": 6400, "loss": -5.888009071350098, "grad_norm": 0.673487663269043, "lr": 0.00015769124734096862, "ex_per_s": 840.4471390504627, "data_frac": 0.07939622171968992, "mem_mb": 199.197696, "nll": -5.888009071350098, "elapsed_s": 487.4}327{"event": "train", "step": 6420, "loss": -5.983592987060547, "grad_norm": 0.519534707069397, "lr": 0.00015696119776945368, "ex_per_s": 842.702281155426, "data_frac": 0.07938721734186176, "mem_mb": 199.197696, "nll": -5.983592987060547, "elapsed_s": 487.6}328{"event": "train", "step": 6440, "loss": -5.974699020385742, "grad_norm": 0.4431053102016449, "lr": 0.00015623138404693952, "ex_per_s": 844.9554429452933, "data_frac": 0.0793782125585093, "mem_mb": 199.197696, "nll": -5.974699020385742, "elapsed_s": 487.8}329{"event": "train", "step": 6460, "loss": -5.930586338043213, "grad_norm": 0.5685964226722717, "lr": 0.0001555018275853024, "ex_per_s": 847.206989659133, "data_frac": 0.07936994828332902, "mem_mb": 199.197696, "nll": -5.930586338043213, "elapsed_s": 488.0}330{"event": "train", "step": 6480, "loss": -5.9396138191223145, "grad_norm": 0.4134977161884308, "lr": 0.0001547725497888706, "ex_per_s": 849.4555740183633, "data_frac": 0.07936059872433668, "mem_mb": 199.197696, "nll": -5.9396138191223145, "elapsed_s": 488.2}331{"event": "train", "step": 6500, "loss": -5.9584832191467285, "grad_norm": 0.29049164056777954, "lr": 0.0001540435720537968, "ex_per_s": 851.7035327452149, "data_frac": 0.07935131964804117, "mem_mb": 199.197696, "nll": -5.9584832191467285, "elapsed_s": 488.4}332{"event": "train", "step": 6520, "loss": -6.040899276733398, "grad_norm": 0.3915294408798218, "lr": 0.0001533149157674301, "ex_per_s": 853.9504128111296, "data_frac": 0.07934212093611548, "mem_mb": 199.197696, "nll": -6.040899276733398, "elapsed_s": 488.7}333{"event": "train", "step": 6540, "loss": -6.160860061645508, "grad_norm": 0.24131719768047333, "lr": 0.00015258660230768893, "ex_per_s": 856.1931635713551, "data_frac": 0.07933313745321106, "mem_mb": 199.197696, "nll": -6.160860061645508, "elapsed_s": 488.9}334{"event": "train", "step": 6560, "loss": -6.050800323486328, "grad_norm": 0.6001623868942261, "lr": 0.0001518586530424334, "ex_per_s": 858.4344810033499, "data_frac": 0.07932449340433824, "mem_mb": 199.197696, "nll": -6.050800323486328, "elapsed_s": 489.1}335{"event": "train", "step": 6580, "loss": -6.043028354644775, "grad_norm": 0.7197692394256592, "lr": 0.00015113108932883853, "ex_per_s": 860.6733667869078, "data_frac": 0.0793160309555219, "mem_mb": 199.197696, "nll": -6.043028354644775, "elapsed_s": 489.3}336{"event": "train", "step": 6600, "loss": -6.012094497680664, "grad_norm": 0.21036382019519806, "lr": 0.00015040393251276787, "ex_per_s": 862.9092253734208, "data_frac": 0.079306752726067, "mem_mb": 199.197696, "nll": -6.012094497680664, "elapsed_s": 489.5}337{"event": "train", "step": 6620, "loss": -6.207972049713135, "grad_norm": 0.2587510645389557, "lr": 0.00014967720392814694, "ex_per_s": 865.1461186955028, "data_frac": 0.07929804078067466, "mem_mb": 199.197696, "nll": -6.207972049713135, "elapsed_s": 489.7}338{"event": "train", "step": 6640, "loss": -6.1017746925354, "grad_norm": 0.24645493924617767, "lr": 0.00014895092489633762, "ex_per_s": 867.3787966904436, "data_frac": 0.07928992089087952, "mem_mb": 199.197696, "nll": -6.1017746925354, "elapsed_s": 489.9}339{"event": "train", "step": 6660, "loss": -5.968257427215576, "grad_norm": 0.1872396320104599, "lr": 0.00014822511672551224, "ex_per_s": 869.6055724006391, "data_frac": 0.07928145746855601, "mem_mb": 199.197696, "nll": -5.968257427215576, "elapsed_s": 490.2}340{"event": "train", "step": 6680, "loss": -5.7969651222229, "grad_norm": 0.3267176151275635, "lr": 0.0001474998007100288, "ex_per_s": 871.8356594398392, "data_frac": 0.07927322202091437, "mem_mb": 199.197696, "nll": -5.7969651222229, "elapsed_s": 490.4}341{"event": "train", "step": 6700, "loss": -5.867527008056641, "grad_norm": 0.6363363862037659, "lr": 0.00014677499812980582, "ex_per_s": 874.0605124347728, "data_frac": 0.07926544090290785, "mem_mb": 199.197696, "nll": -5.867527008056641, "elapsed_s": 490.6}342{"event": "train", "step": 6720, "loss": -6.21061897277832, "grad_norm": 0.3791656494140625, "lr": 0.0001460507302496985, "ex_per_s": 876.2839328168167, "data_frac": 0.07925758905299633, "mem_mb": 199.197696, "nll": -6.21061897277832, "elapsed_s": 490.8}343{"event": "train", "step": 6740, "loss": -5.973897457122803, "grad_norm": 0.33077433705329895, "lr": 0.0001453270183188744, "ex_per_s": 878.503009620919, "data_frac": 0.07925104849998892, "mem_mb": 199.197696, "nll": -5.973897457122803, "elapsed_s": 491.0}344{"event": "train", "step": 6760, "loss": -5.906261920928955, "grad_norm": 0.21028174459934235, "lr": 0.00014460388357019025, "ex_per_s": 880.7229629888252, "data_frac": 0.0792434937369873, "mem_mb": 199.197696, "nll": -5.906261920928955, "elapsed_s": 491.2}345{"event": "train", "step": 6780, "loss": -6.001175403594971, "grad_norm": 0.14503253996372223, "lr": 0.0001438813472195688, "ex_per_s": 882.9385587180411, "data_frac": 0.07923652455212117, "mem_mb": 199.197696, "nll": -6.001175403594971, "elapsed_s": 491.5}346{"event": "train", "step": 6800, "loss": -6.0929365158081055, "grad_norm": 0.45992153882980347, "lr": 0.00014315943046537674, "ex_per_s": 885.1520486185559, "data_frac": 0.07922986621820025, "mem_mb": 199.197696, "nll": -6.0929365158081055, "elapsed_s": 491.7}347{"event": "train", "step": 6820, "loss": -6.112729072570801, "grad_norm": 0.6652604341506958, "lr": 0.00014243815448780235, "ex_per_s": 887.3662657739412, "data_frac": 0.07922235480328918, "mem_mb": 199.197696, "nll": -6.112729072570801, "elapsed_s": 491.9}348{"event": "train", "step": 6840, "loss": -5.833058834075928, "grad_norm": 0.38248327374458313, "lr": 0.00014171754044823431, "ex_per_s": 889.5777084826605, "data_frac": 0.07921380236029377, "mem_mb": 199.197696, "nll": -5.833058834075928, "elapsed_s": 492.1}349{"event": "train", "step": 6860, "loss": -5.953714370727539, "grad_norm": 0.15213283896446228, "lr": 0.0001409976094886409, "ex_per_s": 891.7876520999481, "data_frac": 0.0792057146378447, "mem_mb": 199.197696, "nll": -5.953714370727539, "elapsed_s": 492.3}350{"event": "train", "step": 6880, "loss": -6.027513027191162, "grad_norm": 0.41715532541275024, "lr": 0.00014027838273094942, "ex_per_s": 893.9925874866847, "data_frac": 0.07919694175637243, "mem_mb": 199.197696, "nll": -6.027513027191162, "elapsed_s": 492.5}351{"event": "train", "step": 6900, "loss": -6.005640029907227, "grad_norm": 0.4163166582584381, "lr": 0.00013955988127642706, "ex_per_s": 896.1971205869373, "data_frac": 0.07918923097248975, "mem_mb": 199.197696, "nll": -6.005640029907227, "elapsed_s": 492.8}352{"event": "train", "step": 6920, "loss": -6.0707106590271, "grad_norm": 0.4486375153064728, "lr": 0.0001388421262050611, "ex_per_s": 898.400427772419, "data_frac": 0.07918164631982179, "mem_mb": 199.197696, "nll": -6.0707106590271, "elapsed_s": 493.0}353{"event": "train", "step": 6940, "loss": -6.065920352935791, "grad_norm": 0.4079975485801697, "lr": 0.00013812513857494116, "ex_per_s": 900.5995836288907, "data_frac": 0.07917493633024604, "mem_mb": 199.197696, "nll": -6.065920352935791, "elapsed_s": 493.2}354{"event": "train", "step": 6960, "loss": -5.746312141418457, "grad_norm": 0.42632800340652466, "lr": 0.00013740893942164073, "ex_per_s": 902.7934632366405, "data_frac": 0.07916819329193427, "mem_mb": 199.197696, "nll": -5.746312141418457, "elapsed_s": 493.4}355{"event": "train", "step": 6980, "loss": -6.199116230010986, "grad_norm": 0.3047567307949066, "lr": 0.0001366935497576006, "ex_per_s": 904.9874003673044, "data_frac": 0.07916128212909139, "mem_mb": 199.197696, "nll": -6.199116230010986, "elapsed_s": 493.6}356{"event": "train", "step": 7000, "loss": -6.12650728225708, "grad_norm": 0.3954755663871765, "lr": 0.00013597899057151173, "ex_per_s": 907.1815715874079, "data_frac": 0.07915336310643949, "mem_mb": 199.197696, "nll": -6.12650728225708, "elapsed_s": 493.8}357{"event": "val", "step": 7000, "val_nll": -6.451356390340603, "best": -6.451356390340603, "bad": 0, "mem_mb": 199.197696, "elapsed_s": 557.1}358{"event": "train", "step": 7020, "loss": -6.146768569946289, "grad_norm": 0.18179325759410858, "lr": 0.0001352652828277003, "ex_per_s": 806.0448428615043, "data_frac": 0.07015176245915831, "mem_mb": 199.197696, "nll": -6.146768569946289, "elapsed_s": 557.4}359{"event": "train", "step": 7040, "loss": -6.059642314910889, "grad_norm": 0.2840631306171417, "lr": 0.00013455244746551177, "ex_per_s": 808.0275936315443, "data_frac": 0.07014781207223514, "mem_mb": 199.197696, "nll": -6.059642314910889, "elapsed_s": 557.6}360{"event": "train", "step": 7060, "loss": -5.989737033843994, "grad_norm": 0.5075242519378662, "lr": 0.00013384050539869724, "ex_per_s": 810.0114969268901, "data_frac": 0.07014374685188167, "mem_mb": 199.197696, "nll": -5.989737033843994, "elapsed_s": 557.8}361{"event": "train", "step": 7080, "loss": -6.067172050476074, "grad_norm": 0.8520718812942505, "lr": 0.00013312947751479927, "ex_per_s": 811.9918873777643, "data_frac": 0.07014009712252078, "mem_mb": 199.197696, "nll": -6.067172050476074, "elapsed_s": 558.0}362{"event": "train", "step": 7100, "loss": -5.960032939910889, "grad_norm": 0.16685454547405243, "lr": 0.00013241938467453975, "ex_per_s": 813.9737614687431, "data_frac": 0.07013577558033417, "mem_mb": 199.197696, "nll": -5.960032939910889, "elapsed_s": 558.3}363{"event": "train", "step": 7120, "loss": -6.043019771575928, "grad_norm": 0.26027485728263855, "lr": 0.00013171024771120715, "ex_per_s": 815.9533258037352, "data_frac": 0.07013179910498318, "mem_mb": 199.197696, "nll": -6.043019771575928, "elapsed_s": 558.5}364{"event": "train", "step": 7140, "loss": -5.944083213806152, "grad_norm": 0.2716039717197418, "lr": 0.0001310020874300458, "ex_per_s": 817.927582390955, "data_frac": 0.07012969222810914, "mem_mb": 199.197696, "nll": -5.944083213806152, "elapsed_s": 558.7}365{"event": "train", "step": 7160, "loss": -6.030919551849365, "grad_norm": 0.36017143726348877, "lr": 0.00013029492460764526, "ex_per_s": 819.9011642175625, "data_frac": 0.07012606493084458, "mem_mb": 199.197696, "nll": -6.030919551849365, "elapsed_s": 558.9}366{"event": "train", "step": 7180, "loss": -5.9498395919799805, "grad_norm": 0.46407878398895264, "lr": 0.00012958877999133094, "ex_per_s": 821.8751251042368, "data_frac": 0.0701222803043381, "mem_mb": 199.197696, "nll": -5.9498395919799805, "elapsed_s": 559.1}367{"event": "train", "step": 7200, "loss": -5.95897912979126, "grad_norm": 0.46309542655944824, "lr": 0.0001288836742985552, "ex_per_s": 823.8452554901762, "data_frac": 0.07011918370102553, "mem_mb": 199.197696, "nll": -5.95897912979126, "elapsed_s": 559.3}368{"event": "train", "step": 7220, "loss": -6.064138412475586, "grad_norm": 0.5892592072486877, "lr": 0.00012817962821628953, "ex_per_s": 825.8126618439392, "data_frac": 0.07011594086771518, "mem_mb": 199.197696, "nll": -6.064138412475586, "elapsed_s": 559.6}369{"event": "train", "step": 7240, "loss": -5.956246852874756, "grad_norm": 0.2932833731174469, "lr": 0.00012747666240041779, "ex_per_s": 827.7799978892186, "data_frac": 0.07011181759864414, "mem_mb": 199.197696, "nll": -5.956246852874756, "elapsed_s": 559.8}370{"event": "train", "step": 7260, "loss": -6.072705268859863, "grad_norm": 0.4188159704208374, "lr": 0.00012677479747513008, "ex_per_s": 829.7474457933399, "data_frac": 0.07010900079850084, "mem_mb": 199.197696, "nll": -6.072705268859863, "elapsed_s": 560.0}371{"event": "train", "step": 7280, "loss": -5.871745586395264, "grad_norm": 0.20709998905658722, "lr": 0.00012607405403231763, "ex_per_s": 831.7129522820239, "data_frac": 0.07010542693589554, "mem_mb": 199.197696, "nll": -5.871745586395264, "elapsed_s": 560.2}372{"event": "train", "step": 7300, "loss": -5.928674221038818, "grad_norm": 0.4534956216812134, "lr": 0.0001253744526309688, "ex_per_s": 833.6763193585174, "data_frac": 0.07010176015592404, "mem_mb": 199.197696, "nll": -5.928674221038818, "elapsed_s": 560.4}373{"event": "train", "step": 7320, "loss": -6.072600841522217, "grad_norm": 0.6130055785179138, "lr": 0.00012467601379656567, "ex_per_s": 835.6383652729818, "data_frac": 0.0700982692556467, "mem_mb": 199.197696, "nll": -6.072600841522217, "elapsed_s": 560.6}374{"event": "train", "step": 7340, "loss": -6.020956516265869, "grad_norm": 0.4568396210670471, "lr": 0.000123978758020482, "ex_per_s": 837.5999250436932, "data_frac": 0.07009478007979719, "mem_mb": 199.197696, "nll": -6.020956516265869, "elapsed_s": 560.8}375{"event": "train", "step": 7360, "loss": -5.992308139801025, "grad_norm": 0.36458539962768555, "lr": 0.00012328270575938212, "ex_per_s": 839.5591712126271, "data_frac": 0.07009075142393668, "mem_mb": 199.197696, "nll": -5.992308139801025, "elapsed_s": 561.1}376{"event": "train", "step": 7380, "loss": -6.079230308532715, "grad_norm": 0.17244796454906464, "lr": 0.00012258787743462041, "ex_per_s": 841.5161165190135, "data_frac": 0.07008761929623228, "mem_mb": 199.197696, "nll": -6.079230308532715, "elapsed_s": 561.3}377{"event": "train", "step": 7400, "loss": -6.006875991821289, "grad_norm": 0.21788007020950317, "lr": 0.00012189429343164274, "ex_per_s": 843.4716663598203, "data_frac": 0.07008387855875897, "mem_mb": 199.197696, "nll": -6.006875991821289, "elapsed_s": 561.5}378{"event": "train", "step": 7420, "loss": -6.051244258880615, "grad_norm": 0.2510816752910614, "lr": 0.00012120197409938763, "ex_per_s": 845.4249428821382, "data_frac": 0.07007978674829035, "mem_mb": 199.197696, "nll": -6.051244258880615, "elapsed_s": 561.7}379{"event": "train", "step": 7440, "loss": -5.788137912750244, "grad_norm": 0.570601224899292, "lr": 0.00012051093974969008, "ex_per_s": 847.3787150520026, "data_frac": 0.07007583909605243, "mem_mb": 199.197696, "nll": -5.788137912750244, "elapsed_s": 561.9}380{"event": "train", "step": 7460, "loss": -6.218047142028809, "grad_norm": 0.1796950101852417, "lr": 0.00011982121065668482, "ex_per_s": 849.331441405801, "data_frac": 0.07007210223802779, "mem_mb": 199.197696, "nll": -6.218047142028809, "elapsed_s": 562.1}381{"event": "train", "step": 7480, "loss": -5.959985733032227, "grad_norm": 0.6467498540878296, "lr": 0.00011913280705621224, "ex_per_s": 851.2802670885097, "data_frac": 0.07006926398794334, "mem_mb": 199.197696, "nll": -5.959985733032227, "elapsed_s": 562.4}382{"event": "train", "step": 7500, "loss": -5.776193141937256, "grad_norm": 0.3709909915924072, "lr": 0.00011844574914522403, "ex_per_s": 853.2293149813667, "data_frac": 0.07006616432168197, "mem_mb": 199.197696, "nll": -5.776193141937256, "elapsed_s": 562.6}383{"event": "train", "step": 7520, "loss": -6.064426898956299, "grad_norm": 0.4137631058692932, "lr": 0.00011776005708119105, "ex_per_s": 855.1754313243672, "data_frac": 0.07006235525045293, "mem_mb": 199.197696, "nll": -6.064426898956299, "elapsed_s": 562.8}384{"event": "train", "step": 7540, "loss": -5.844050407409668, "grad_norm": 0.4170382618904114, "lr": 0.00011707575098151199, "ex_per_s": 857.1196270904283, "data_frac": 0.07005978280117993, "mem_mb": 199.197696, "nll": -5.844050407409668, "elapsed_s": 563.0}385{"event": "train", "step": 7560, "loss": -5.9034271240234375, "grad_norm": 0.22872020304203033, "lr": 0.00011639285092292267, "ex_per_s": 859.0624300044333, "data_frac": 0.07005706799764212, "mem_mb": 199.197696, "nll": -5.9034271240234375, "elapsed_s": 563.2}386{"event": "train", "step": 7580, "loss": -5.870643615722656, "grad_norm": 0.3666591942310333, "lr": 0.00011571137694090757, "ex_per_s": 861.0043520843801, "data_frac": 0.07005356950737883, "mem_mb": 199.197696, "nll": -5.870643615722656, "elapsed_s": 563.4}387{"event": "train", "step": 7600, "loss": -6.055402755737305, "grad_norm": 0.3926175534725189, "lr": 0.00011503134902911152, "ex_per_s": 862.9429890258838, "data_frac": 0.0700499521501369, "mem_mb": 199.197696, "nll": -6.055402755737305, "elapsed_s": 563.7}388{"event": "train", "step": 7620, "loss": -5.945544719696045, "grad_norm": 0.5763276219367981, "lr": 0.00011435278713875375, "ex_per_s": 864.8814668972399, "data_frac": 0.07004753799508587, "mem_mb": 199.197696, "nll": -5.945544719696045, "elapsed_s": 563.9}389{"event": "train", "step": 7640, "loss": -5.9180755615234375, "grad_norm": 0.5324267745018005, "lr": 0.00011367571117804159, "ex_per_s": 866.8176193434936, "data_frac": 0.0700452339329681, "mem_mb": 199.197696, "nll": -5.9180755615234375, "elapsed_s": 564.1}390{"event": "train", "step": 7660, "loss": -6.045597553253174, "grad_norm": 0.9865278601646423, "lr": 0.00011300014101158744, "ex_per_s": 868.7507078841809, "data_frac": 0.07004216933897285, "mem_mb": 199.197696, "nll": -6.045597553253174, "elapsed_s": 564.3}391{"event": "train", "step": 7680, "loss": -5.96113395690918, "grad_norm": 0.20937462151050568, "lr": 0.00011232609645982515, "ex_per_s": 870.6794153894776, "data_frac": 0.07004034379285137, "mem_mb": 199.197696, "nll": -5.96113395690918, "elapsed_s": 564.5}392{"event": "train", "step": 7700, "loss": -5.908987522125244, "grad_norm": 0.20445457100868225, "lr": 0.00011165359729842902, "ex_per_s": 872.6116378692166, "data_frac": 0.0700372713314972, "mem_mb": 199.197696, "nll": -5.908987522125244, "elapsed_s": 564.7}393{"event": "train", "step": 7720, "loss": -6.040693759918213, "grad_norm": 0.2716276943683624, "lr": 0.00011098266325773331, "ex_per_s": 874.5420630720992, "data_frac": 0.07003414273829438, "mem_mb": 199.197696, "nll": -6.040693759918213, "elapsed_s": 565.0}394{"event": "train", "step": 7740, "loss": -6.047130107879639, "grad_norm": 0.5702710747718811, "lr": 0.00011031331402215353, "ex_per_s": 876.4696313221527, "data_frac": 0.07003186517743454, "mem_mb": 199.197696, "nll": -6.047130107879639, "elapsed_s": 565.2}395{"event": "train", "step": 7760, "loss": -5.970617294311523, "grad_norm": 0.4582119286060333, "lr": 0.00010964556922960883, "ex_per_s": 878.3962963086673, "data_frac": 0.07002857359739748, "mem_mb": 199.197696, "nll": -5.970617294311523, "elapsed_s": 565.4}396{"event": "train", "step": 7780, "loss": -6.051660537719727, "grad_norm": 0.186898872256279, "lr": 0.00010897944847094595, "ex_per_s": 880.3198468418337, "data_frac": 0.07002571785742467, "mem_mb": 199.197696, "nll": -6.051660537719727, "elapsed_s": 565.6}397{"event": "train", "step": 7800, "loss": -6.060385227203369, "grad_norm": 0.27900657057762146, "lr": 0.00010831497128936428, "ex_per_s": 882.2437646286011, "data_frac": 0.07002268572346051, "mem_mb": 199.197696, "nll": -6.060385227203369, "elapsed_s": 565.8}398{"event": "train", "step": 7820, "loss": -5.89963960647583, "grad_norm": 0.5097573399543762, "lr": 0.00010765215717984271, "ex_per_s": 884.1698660765985, "data_frac": 0.07001945603083663, "mem_mb": 199.197696, "nll": -5.89963960647583, "elapsed_s": 566.1}399{"event": "train", "step": 7840, "loss": -5.955586910247803, "grad_norm": 0.16589206457138062, "lr": 0.00010699102558856754, "ex_per_s": 886.0907162049026, "data_frac": 0.07001646459652353, "mem_mb": 199.197696, "nll": -5.955586910247803, "elapsed_s": 566.3}400{"event": "train", "step": 7860, "loss": -6.11777925491333, "grad_norm": 0.2647417485713959, "lr": 0.0001063315959123618, "ex_per_s": 888.0112858931719, "data_frac": 0.0700143836138445, "mem_mb": 199.197696, "nll": -6.11777925491333, "elapsed_s": 566.5}401{"event": "train", "step": 7880, "loss": -6.057584762573242, "grad_norm": 0.4305945932865143, "lr": 0.0001056738874981166, "ex_per_s": 889.9307681211571, "data_frac": 0.07001141076391294, "mem_mb": 199.197696, "nll": -6.057584762573242, "elapsed_s": 566.7}402{"event": "train", "step": 7900, "loss": -6.060110569000244, "grad_norm": 0.43378332257270813, "lr": 0.00010501791964222304, "ex_per_s": 891.8472237282672, "data_frac": 0.07000825711436767, "mem_mb": 199.197696, "nll": -6.060110569000244, "elapsed_s": 566.9}403{"event": "train", "step": 7920, "loss": -6.086141586303711, "grad_norm": 0.19338393211364746, "lr": 0.00010436371159000636, "ex_per_s": 893.7637429274032, "data_frac": 0.07000606971948552, "mem_mb": 199.197696, "nll": -6.086141586303711, "elapsed_s": 567.1}404{"event": "train", "step": 7940, "loss": -6.05014181137085, "grad_norm": 0.17552322149276733, "lr": 0.00010371128253516118, "ex_per_s": 895.6779103893838, "data_frac": 0.070003090444962, "mem_mb": 199.197696, "nll": -6.05014181137085, "elapsed_s": 567.4}405{"event": "train", "step": 7960, "loss": -5.9949517250061035, "grad_norm": 0.2957487106323242, "lr": 0.00010306065161918857, "ex_per_s": 897.5916438285112, "data_frac": 0.07000111962756779, "mem_mb": 199.197696, "nll": -5.9949517250061035, "elapsed_s": 567.6}406{"event": "train", "step": 7980, "loss": -5.839671611785889, "grad_norm": 0.15575507283210754, "lr": 0.00010241183793083414, "ex_per_s": 899.5036834727593, "data_frac": 0.0699983820636234, "mem_mb": 199.197696, "nll": -5.839671611785889, "elapsed_s": 567.8}407{"event": "train", "step": 8000, "loss": -6.053898811340332, "grad_norm": 0.16893762350082397, "lr": 0.00010176486050552833, "ex_per_s": 901.4133021535001, "data_frac": 0.06999587781292675, "mem_mb": 199.197696, "nll": -6.053898811340332, "elapsed_s": 568.0}408{"event": "val", "step": 8000, "val_nll": -6.451161267979316, "best": -6.451356390340603, "bad": 1, "mem_mb": 199.197696, "elapsed_s": 631.1}409{"event": "train", "step": 8020, "loss": -6.045139312744141, "grad_norm": 0.12427534908056259, "lr": 0.00010111973832482765, "ex_per_s": 812.9870318629493, "data_frac": 0.06299405436613344, "mem_mb": 199.197696, "nll": -6.045139312744141, "elapsed_s": 631.4}410{"event": "train", "step": 8040, "loss": -6.034376621246338, "grad_norm": 0.14963825047016144, "lr": 0.00010047649031585807, "ex_per_s": 814.73440337864, "data_frac": 0.06299431848459684, "mem_mb": 199.197696, "nll": -6.034376621246338, "elapsed_s": 631.6}411{"event": "train", "step": 8060, "loss": -5.831313610076904, "grad_norm": 0.7285075783729553, "lr": 9.983513535075943e-05, "ex_per_s": 816.4816714627449, "data_frac": 0.06299454629214812, "mem_mb": 199.197696, "nll": -5.831313610076904, "elapsed_s": 631.8}412{"event": "train", "step": 8080, "loss": -6.004344463348389, "grad_norm": 0.23246079683303833, "lr": 9.919569224613192e-05, "ex_per_s": 818.2276486991851, "data_frac": 0.06299470697227132, "mem_mb": 199.197696, "nll": -6.004344463348389, "elapsed_s": 632.0}413{"event": "train", "step": 8100, "loss": -5.941213607788086, "grad_norm": 0.5103816986083984, "lr": 9.855817976248399e-05, "ex_per_s": 819.9713085584829, "data_frac": 0.06299425179017913, "mem_mb": 199.197696, "nll": -5.941213607788086, "elapsed_s": 632.2}414{"event": "train", "step": 8120, "loss": -5.904944896697998, "grad_norm": 0.6870570778846741, "lr": 9.792261660368207e-05, "ex_per_s": 821.7147573754255, "data_frac": 0.06299395167680463, "mem_mb": 199.197696, "nll": -5.904944896697998, "elapsed_s": 632.4}415{"event": "train", "step": 8140, "loss": -5.895094871520996, "grad_norm": 0.15653373301029205, "lr": 9.72890214164015e-05, "ex_per_s": 823.4572416520236, "data_frac": 0.06299351204411079, "mem_mb": 199.197696, "nll": -5.895094871520996, "elapsed_s": 632.7}416{"event": "train", "step": 8160, "loss": -6.044752597808838, "grad_norm": 0.21605417132377625, "lr": 9.66574127895797e-05, "ex_per_s": 825.1981616558409, "data_frac": 0.06299326454904962, "mem_mb": 199.197696, "nll": -6.044752597808838, "elapsed_s": 632.9}417{"event": "train", "step": 8180, "loss": -5.946181774139404, "grad_norm": 0.5066234469413757, "lr": 9.602780925387089e-05, "ex_per_s": 826.937901545476, "data_frac": 0.06299340929294119, "mem_mb": 199.197696, "nll": -5.946181774139404, "elapsed_s": 633.1}418{"event": "train", "step": 8200, "loss": -6.171606540679932, "grad_norm": 0.5172151923179626, "lr": 9.540022928110205e-05, "ex_per_s": 828.6759144604323, "data_frac": 0.0629931674529186, "mem_mb": 199.197696, "nll": -6.171606540679932, "elapsed_s": 633.3}419{"event": "train", "step": 8220, "loss": -5.907508850097656, "grad_norm": 0.15454810857772827, "lr": 9.477469128373143e-05, "ex_per_s": 830.4102277415805, "data_frac": 0.06299272108827406, "mem_mb": 199.197696, "nll": -5.907508850097656, "elapsed_s": 633.5}420{"event": "train", "step": 8240, "loss": -5.842297077178955, "grad_norm": 0.2265644520521164, "lr": 9.415121361430792e-05, "ex_per_s": 832.145918309904, "data_frac": 0.06299286645495339, "mem_mb": 199.197696, "nll": -5.842297077178955, "elapsed_s": 633.7}421{"event": "train", "step": 8260, "loss": -6.187610149383545, "grad_norm": 0.1561296284198761, "lr": 9.352981456493298e-05, "ex_per_s": 833.8802003268685, "data_frac": 0.06299342711604258, "mem_mb": 199.197696, "nll": -6.187610149383545, "elapsed_s": 634.0}422{"event": "train", "step": 8280, "loss": -6.15413761138916, "grad_norm": 0.4148308336734772, "lr": 9.291051236672361e-05, "ex_per_s": 835.6092167131548, "data_frac": 0.062993886670576, "mem_mb": 199.197696, "nll": -6.15413761138916, "elapsed_s": 634.2}423{"event": "train", "step": 8300, "loss": -6.048221111297607, "grad_norm": 0.45408329367637634, "lr": 9.229332518927795e-05, "ex_per_s": 837.3378007639327, "data_frac": 0.0629938744671571, "mem_mb": 199.197696, "nll": -6.048221111297607, "elapsed_s": 634.4}424{"event": "train", "step": 8320, "loss": -5.896383285522461, "grad_norm": 0.5587940216064453, "lr": 9.167827114014155e-05, "ex_per_s": 839.0679969122709, "data_frac": 0.06299361096550526, "mem_mb": 199.197696, "nll": -5.896383285522461, "elapsed_s": 634.6}425{"event": "train", "step": 8340, "loss": -6.060928821563721, "grad_norm": 0.4353545010089874, "lr": 9.106536826427686e-05, "ex_per_s": 840.7934877073944, "data_frac": 0.06299509652615491, "mem_mb": 199.197696, "nll": -6.060928821563721, "elapsed_s": 634.8}426{"event": "train", "step": 8360, "loss": -6.007114887237549, "grad_norm": 0.38545963168144226, "lr": 9.045463454353313e-05, "ex_per_s": 842.5203930833188, "data_frac": 0.06299459194791884, "mem_mb": 199.197696, "nll": -6.007114887237549, "elapsed_s": 635.1}427{"event": "train", "step": 8380, "loss": -6.053290367126465, "grad_norm": 0.3625524342060089, "lr": 8.984608789611947e-05, "ex_per_s": 844.2463930666012, "data_frac": 0.06299514933871794, "mem_mb": 199.197696, "nll": -6.053290367126465, "elapsed_s": 635.3}428{"event": "train", "step": 8400, "loss": -6.03920841217041, "grad_norm": 0.2159765213727951, "lr": 8.92397461760784e-05, "ex_per_s": 845.9692501778519, "data_frac": 0.06299458534475438, "mem_mb": 199.197696, "nll": -6.03920841217041, "elapsed_s": 635.5}429{"event": "train", "step": 8420, "loss": -6.001802444458008, "grad_norm": 0.17769992351531982, "lr": 8.863562717276283e-05, "ex_per_s": 847.6929176688427, "data_frac": 0.0629938863875974, "mem_mb": 199.197696, "nll": -6.001802444458008, "elapsed_s": 635.7}430{"event": "train", "step": 8440, "loss": -6.026605129241943, "grad_norm": 0.36799079179763794, "lr": 8.803374861031363e-05, "ex_per_s": 849.4141872413187, "data_frac": 0.062993316596377, "mem_mb": 199.197696, "nll": -6.026605129241943, "elapsed_s": 635.9}431{"event": "train", "step": 8460, "loss": -5.907409191131592, "grad_norm": 0.5527023077011108, "lr": 8.743412814713985e-05, "ex_per_s": 851.1315266236963, "data_frac": 0.06299416938216601, "mem_mb": 199.197696, "nll": -5.907409191131592, "elapsed_s": 636.1}432{"event": "train", "step": 8480, "loss": -5.878860950469971, "grad_norm": 0.30565744638442993, "lr": 8.683678337540044e-05, "ex_per_s": 852.8526411560389, "data_frac": 0.0629942614623673, "mem_mb": 199.197696, "nll": -5.878860950469971, "elapsed_s": 636.4}433{"event": "train", "step": 8500, "loss": -5.966801166534424, "grad_norm": 0.7053433060646057, "lr": 8.62417318204883e-05, "ex_per_s": 854.5738327248713, "data_frac": 0.06299443407263525, "mem_mb": 199.197696, "nll": -5.966801166534424, "elapsed_s": 636.6}434{"event": "train", "step": 8520, "loss": -5.931637287139893, "grad_norm": 0.3116026222705841, "lr": 8.564899094051614e-05, "ex_per_s": 856.2921374695908, "data_frac": 0.06299426336631168, "mem_mb": 199.197696, "nll": -5.931637287139893, "elapsed_s": 636.8}435{"event": "train", "step": 8540, "loss": -5.908463478088379, "grad_norm": 0.3809453845024109, "lr": 8.505857812580402e-05, "ex_per_s": 858.0103996720943, "data_frac": 0.06299399199800645, "mem_mb": 199.197696, "nll": -5.908463478088379, "elapsed_s": 637.0}436{"event": "train", "step": 8560, "loss": -5.970465183258057, "grad_norm": 0.2280689775943756, "lr": 8.447051069836941e-05, "ex_per_s": 859.7258652890146, "data_frac": 0.06299426560080912, "mem_mb": 199.197696, "nll": -5.970465183258057, "elapsed_s": 637.2}437{"event": "train", "step": 8580, "loss": -5.957803249359131, "grad_norm": 0.17284850776195526, "lr": 8.38848059114188e-05, "ex_per_s": 861.4400221153297, "data_frac": 0.06299485966263649, "mem_mb": 199.197696, "nll": -5.957803249359131, "elapsed_s": 637.4}438{"event": "train", "step": 8600, "loss": -6.0261454582214355, "grad_norm": 0.2886722683906555, "lr": 8.330148094884171e-05, "ex_per_s": 863.1534473933416, "data_frac": 0.06299476230919898, "mem_mb": 199.197696, "nll": -6.0261454582214355, "elapsed_s": 637.7}439{"event": "train", "step": 8620, "loss": -5.993616580963135, "grad_norm": 0.5515847206115723, "lr": 8.27205529247063e-05, "ex_per_s": 864.8654016416741, "data_frac": 0.06299530121159776, "mem_mb": 199.197696, "nll": -5.993616580963135, "elapsed_s": 637.9}440{"event": "train", "step": 8640, "loss": -5.9649128913879395, "grad_norm": 0.751103937625885, "lr": 8.214203888275743e-05, "ex_per_s": 866.5778823434548, "data_frac": 0.06299589122753851, "mem_mb": 199.197696, "nll": -5.9649128913879395, "elapsed_s": 638.1}441{"event": "train", "step": 8660, "loss": -5.949605941772461, "grad_norm": 0.414027601480484, "lr": 8.156595579591652e-05, "ex_per_s": 868.2864205069806, "data_frac": 0.06299555709731922, "mem_mb": 199.197696, "nll": -5.949605941772461, "elapsed_s": 638.3}442{"event": "train", "step": 8680, "loss": -5.902663707733154, "grad_norm": 0.14198079705238342, "lr": 8.09923205657837e-05, "ex_per_s": 869.9934693168492, "data_frac": 0.06299572241579139, "mem_mb": 199.197696, "nll": -5.902663707733154, "elapsed_s": 638.5}443{"event": "train", "step": 8700, "loss": -5.933640956878662, "grad_norm": 0.12423156946897507, "lr": 8.042115002214181e-05, "ex_per_s": 871.701444282128, "data_frac": 0.06299574755852405, "mem_mb": 199.197696, "nll": -5.933640956878662, "elapsed_s": 638.8}444{"event": "train", "step": 8720, "loss": -6.165403842926025, "grad_norm": 0.17173735797405243, "lr": 7.985246092246267e-05, "ex_per_s": 873.4090639412176, "data_frac": 0.06299577412128701, "mem_mb": 199.197696, "nll": -6.165403842926025, "elapsed_s": 639.0}445{"event": "train", "step": 8740, "loss": -5.898627281188965, "grad_norm": 0.17379140853881836, "lr": 7.928626995141546e-05, "ex_per_s": 875.1137491025357, "data_frac": 0.06299587792028509, "mem_mb": 199.197696, "nll": -5.898627281188965, "elapsed_s": 639.2}446{"event": "train", "step": 8760, "loss": -6.0755157470703125, "grad_norm": 0.31601905822753906, "lr": 7.872259372037726e-05, "ex_per_s": 876.8187033099865, "data_frac": 0.06299509002322609, "mem_mb": 199.197696, "nll": -6.0755157470703125, "elapsed_s": 639.4}447{"event": "train", "step": 8780, "loss": -5.980639934539795, "grad_norm": 0.21533653140068054, "lr": 7.816144876694563e-05, "ex_per_s": 878.5231866937318, "data_frac": 0.06299514578990513, "mem_mb": 199.197696, "nll": -5.980639934539795, "elapsed_s": 639.6}448{"event": "train", "step": 8800, "loss": -5.952752590179443, "grad_norm": 0.3599402606487274, "lr": 7.760285155445326e-05, "ex_per_s": 880.2255070074087, "data_frac": 0.06299482145594741, "mem_mb": 199.197696, "nll": -5.952752590179443, "elapsed_s": 639.8}449{"event": "train", "step": 8820, "loss": -5.985782623291016, "grad_norm": 0.26832664012908936, "lr": 7.70468184714854e-05, "ex_per_s": 881.9259983093623, "data_frac": 0.0629943858901163, "mem_mb": 199.197696, "nll": -5.985782623291016, "elapsed_s": 640.1}450{"event": "train", "step": 8840, "loss": -5.955906867980957, "grad_norm": 0.18901723623275757, "lr": 7.649336583139853e-05, "ex_per_s": 883.624619742532, "data_frac": 0.06299348715507011, "mem_mb": 199.197696, "nll": -5.955906867980957, "elapsed_s": 640.3}451{"event": "train", "step": 8860, "loss": -5.99208402633667, "grad_norm": 0.45043671131134033, "lr": 7.594250987184208e-05, "ex_per_s": 885.3234677165293, "data_frac": 0.0629934442664277, "mem_mb": 199.197696, "nll": -5.99208402633667, "elapsed_s": 640.5}452{"event": "train", "step": 8880, "loss": -5.97582483291626, "grad_norm": 0.33611416816711426, "lr": 7.53942667542819e-05, "ex_per_s": 887.0209022185444, "data_frac": 0.06299316029141733, "mem_mb": 199.197696, "nll": -5.97582483291626, "elapsed_s": 640.7}453{"event": "train", "step": 8900, "loss": -5.946056365966797, "grad_norm": 0.2975800037384033, "lr": 7.484865256352612e-05, "ex_per_s": 888.7174332937823, "data_frac": 0.06299345393190714, "mem_mb": 199.197696, "nll": -5.946056365966797, "elapsed_s": 640.9}454{"event": "train", "step": 8920, "loss": -6.097597599029541, "grad_norm": 0.2645953595638275, "lr": 7.430568330725319e-05, "ex_per_s": 890.4115057195671, "data_frac": 0.0629939210059597, "mem_mb": 199.197696, "nll": -6.097597599029541, "elapsed_s": 641.1}455{"event": "train", "step": 8940, "loss": -6.111641883850098, "grad_norm": 0.3833582401275635, "lr": 7.376537491554246e-05, "ex_per_s": 892.1037930482031, "data_frac": 0.06299460924681725, "mem_mb": 199.197696, "nll": -6.111641883850098, "elapsed_s": 641.4}456{"event": "train", "step": 8960, "loss": -6.058460712432861, "grad_norm": 0.18517132103443146, "lr": 7.322774324040643e-05, "ex_per_s": 893.7997779107662, "data_frac": 0.06299437763885808, "mem_mb": 199.197696, "nll": -6.058460712432861, "elapsed_s": 641.6}457{"event": "train", "step": 8980, "loss": -5.912591934204102, "grad_norm": 1.2155793905258179, "lr": 7.26928040553261e-05, "ex_per_s": 895.4933364817683, "data_frac": 0.0629937538321916, "mem_mb": 199.197696, "nll": -5.912591934204102, "elapsed_s": 641.8}458{"event": "train", "step": 9000, "loss": -6.092541694641113, "grad_norm": 0.17395059764385223, "lr": 7.216057305478784e-05, "ex_per_s": 897.1872772921906, "data_frac": 0.06299279612516409, "mem_mb": 199.197696, "nll": -6.092541694641113, "elapsed_s": 642.0}459{"event": "val", "step": 9000, "val_nll": -6.451411052838157, "best": -6.451411052838157, "bad": 0, "mem_mb": 199.197696, "elapsed_s": 705.9}460{"event": "train", "step": 9020, "loss": -6.007840633392334, "grad_norm": 0.35179656744003296, "lr": 7.163106585382312e-05, "ex_per_s": 817.4312744378805, "data_frac": 0.05728562154310674, "mem_mb": 199.197696, "nll": -6.007840633392334, "elapsed_s": 706.2}461{"event": "train", "step": 9040, "loss": -6.105551719665527, "grad_norm": 0.3281143009662628, "lr": 7.11042979875503e-05, "ex_per_s": 818.9928355881227, "data_frac": 0.057286460837318774, "mem_mb": 199.197696, "nll": -6.105551719665527, "elapsed_s": 706.4}462{"event": "train", "step": 9060, "loss": -6.081694602966309, "grad_norm": 0.3289381265640259, "lr": 7.058028491071894e-05, "ex_per_s": 820.5521425802792, "data_frac": 0.05728828968075526, "mem_mb": 199.197696, "nll": -6.081694602966309, "elapsed_s": 706.7}463{"event": "train", "step": 9080, "loss": -6.0953288078308105, "grad_norm": 0.16721753776073456, "lr": 7.005904199725619e-05, "ex_per_s": 822.1124540847965, "data_frac": 0.057289224515762883, "mem_mb": 199.197696, "nll": -6.0953288078308105, "elapsed_s": 706.9}464{"event": "train", "step": 9100, "loss": -6.169569492340088, "grad_norm": 0.22093935310840607, "lr": 6.954058453981609e-05, "ex_per_s": 823.67114689512, "data_frac": 0.05729059900009737, "mem_mb": 199.197696, "nll": -6.169569492340088, "elapsed_s": 707.1}465{"event": "train", "step": 9120, "loss": -5.974195957183838, "grad_norm": 0.16760164499282837, "lr": 6.902492774933048e-05, "ex_per_s": 825.2144583320286, "data_frac": 0.05731015889444627, "mem_mb": 199.197696, "nll": -5.974195957183838, "elapsed_s": 707.3}466{"event": "train", "step": 9140, "loss": -6.122684955596924, "grad_norm": 0.18388749659061432, "lr": 6.851208675456295e-05, "ex_per_s": 826.7737856162983, "data_frac": 0.05730728201141104, "mem_mb": 199.197696, "nll": -6.122684955596924, "elapsed_s": 707.5}467{"event": "train", "step": 9160, "loss": -5.980203628540039, "grad_norm": 0.165159210562706, "lr": 6.800207660166508e-05, "ex_per_s": 828.2942980734437, "data_frac": 0.057348119317746304, "mem_mb": 199.197696, "nll": -5.980203628540039, "elapsed_s": 707.8}468{"event": "train", "step": 9180, "loss": -5.958505153656006, "grad_norm": 0.15378999710083008, "lr": 6.749491225373478e-05, "ex_per_s": 829.8447707506105, "data_frac": 0.057347408741779904, "mem_mb": 199.197696, "nll": -5.958505153656006, "elapsed_s": 708.0}469{"event": "train", "step": 9200, "loss": -6.1455535888671875, "grad_norm": 0.2485022395849228, "lr": 6.699060859037737e-05, "ex_per_s": 831.3926662667913, "data_frac": 0.057349638367720286, "mem_mb": 199.197696, "nll": -6.1455535888671875, "elapsed_s": 708.2}470{"event": "train", "step": 9220, "loss": -5.971134185791016, "grad_norm": 0.5134848952293396, "lr": 6.648918040726909e-05, "ex_per_s": 832.938130803757, "data_frac": 0.0573485262730572, "mem_mb": 199.197696, "nll": -5.971134185791016, "elapsed_s": 708.4}471{"event": "train", "step": 9240, "loss": -6.123229026794434, "grad_norm": 0.3187539279460907, "lr": 6.599064241572299e-05, "ex_per_s": 834.4888402080255, "data_frac": 0.057347623043063664, "mem_mb": 199.197696, "nll": -6.123229026794434, "elapsed_s": 708.7}472{"event": "train", "step": 9260, "loss": -6.0256147384643555, "grad_norm": 0.2474401742219925, "lr": 6.549500924225728e-05, "ex_per_s": 836.0381420228522, "data_frac": 0.057346656101867906, "mem_mb": 199.197696, "nll": -6.0256147384643555, "elapsed_s": 708.9}473{"event": "train", "step": 9280, "loss": -5.820409297943115, "grad_norm": 0.3502484858036041, "lr": 6.500229542816624e-05, "ex_per_s": 837.5864681481534, "data_frac": 0.0573456753388425, "mem_mb": 199.197696, "nll": -5.820409297943115, "elapsed_s": 709.1}474{"event": "train", "step": 9300, "loss": -6.006431579589844, "grad_norm": 0.19321858882904053, "lr": 6.45125154290935e-05, "ex_per_s": 839.1301526019777, "data_frac": 0.05734577941218472, "mem_mb": 199.197696, "nll": -6.006431579589844, "elapsed_s": 709.3}475{"event": "train", "step": 9320, "loss": -5.823249340057373, "grad_norm": 0.45986005663871765, "lr": 6.402568361460816e-05, "ex_per_s": 840.6742739952135, "data_frac": 0.05734484169597841, "mem_mb": 199.197696, "nll": -5.823249340057373, "elapsed_s": 709.5}476{"event": "train", "step": 9340, "loss": -6.050802707672119, "grad_norm": 0.6952289342880249, "lr": 6.354181426778293e-05, "ex_per_s": 842.2191533827817, "data_frac": 0.05734459301456309, "mem_mb": 199.197696, "nll": -6.050802707672119, "elapsed_s": 709.7}477{"event": "train", "step": 9360, "loss": -6.064287185668945, "grad_norm": 0.5112816095352173, "lr": 6.306092158477523e-05, "ex_per_s": 843.7627072258427, "data_frac": 0.05734371449690238, "mem_mb": 199.197696, "nll": -6.064287185668945, "elapsed_s": 710.0}478{"event": "train", "step": 9380, "loss": -5.954927921295166, "grad_norm": 0.38810333609580994, "lr": 6.258301967441065e-05, "ex_per_s": 845.3048396219914, "data_frac": 0.05734272231358864, "mem_mb": 199.197696, "nll": -5.954927921295166, "elapsed_s": 710.2}479{"event": "train", "step": 9400, "loss": -6.069336891174316, "grad_norm": 0.24465437233448029, "lr": 6.210812255776909e-05, "ex_per_s": 846.8483800582893, "data_frac": 0.05734199115590519, "mem_mb": 199.197696, "nll": -6.069336891174316, "elapsed_s": 710.4}480{"event": "train", "step": 9420, "loss": -5.9392499923706055, "grad_norm": 0.2173706293106079, "lr": 6.163624416777327e-05, "ex_per_s": 848.3907336523922, "data_frac": 0.057341229530506604, "mem_mb": 199.197696, "nll": -5.9392499923706055, "elapsed_s": 710.6}481{"event": "train", "step": 9440, "loss": -5.922790050506592, "grad_norm": 0.21937479078769684, "lr": 6.116739834878009e-05, "ex_per_s": 849.9310583421623, "data_frac": 0.05733994665643073, "mem_mb": 199.197696, "nll": -5.922790050506592, "elapsed_s": 710.8}482{"event": "train", "step": 9460, "loss": -6.001981735229492, "grad_norm": 0.22447530925273895, "lr": 6.0701598856174347e-05, "ex_per_s": 851.4710982155112, "data_frac": 0.05733928877987553, "mem_mb": 199.197696, "nll": -6.001981735229492, "elapsed_s": 711.1}483{"event": "train", "step": 9480, "loss": -5.926907062530518, "grad_norm": 0.23037730157375336, "lr": 6.023885935596515e-05, "ex_per_s": 853.0085974110008, "data_frac": 0.057338942867422235, "mem_mb": 199.197696, "nll": -5.926907062530518, "elapsed_s": 711.3}484{"event": "train", "step": 9500, "loss": -5.917639255523682, "grad_norm": 0.15797008574008942, "lr": 5.977919342438512e-05, "ex_per_s": 854.5473412786497, "data_frac": 0.05733817417206472, "mem_mb": 199.197696, "nll": -5.917639255523682, "elapsed_s": 711.5}485{"event": "train", "step": 9520, "loss": -5.990250110626221, "grad_norm": 0.17903968691825867, "lr": 5.9322614547491995e-05, "ex_per_s": 856.0842363407862, "data_frac": 0.05733727197021396, "mem_mb": 199.197696, "nll": -5.990250110626221, "elapsed_s": 711.7}486{"event": "train", "step": 9540, "loss": -6.004460334777832, "grad_norm": 0.24510756134986877, "lr": 5.886913612077285e-05, "ex_per_s": 857.6220656367663, "data_frac": 0.05733672172237408, "mem_mb": 199.197696, "nll": -6.004460334777832, "elapsed_s": 711.9}487{"event": "train", "step": 9560, "loss": -5.963859558105469, "grad_norm": 0.26264047622680664, "lr": 5.841877144875127e-05, "ex_per_s": 859.1585157013088, "data_frac": 0.05733605138432387, "mem_mb": 199.197696, "nll": -5.963859558105469, "elapsed_s": 712.1}488{"event": "train", "step": 9580, "loss": -5.911973476409912, "grad_norm": 0.21278227865695953, "lr": 5.7971533744597e-05, "ex_per_s": 860.6941575055785, "data_frac": 0.057335411298518706, "mem_mb": 199.197696, "nll": -5.911973476409912, "elapsed_s": 712.4}489{"event": "train", "step": 9600, "loss": -6.158555507659912, "grad_norm": 0.3894377648830414, "lr": 5.752743612973809e-05, "ex_per_s": 862.2240549468839, "data_frac": 0.05733514763920542, "mem_mb": 199.197696, "nll": -6.158555507659912, "elapsed_s": 712.6}490{"event": "train", "step": 9620, "loss": -6.025304794311523, "grad_norm": 0.2554040253162384, "lr": 5.7086491633476214e-05, "ex_per_s": 863.754355405988, "data_frac": 0.0573349275820145, "mem_mb": 199.197696, "nll": -6.025304794311523, "elapsed_s": 712.8}491{"event": "train", "step": 9640, "loss": -6.002772808074951, "grad_norm": 0.19570909440517426, "lr": 5.6648713192604105e-05, "ex_per_s": 865.2883583316883, "data_frac": 0.05733380293297793, "mem_mb": 199.197696, "nll": -6.002772808074951, "elapsed_s": 713.0}492{"event": "train", "step": 9660, "loss": -6.051549911499023, "grad_norm": 0.2571553885936737, "lr": 5.621411365102632e-05, "ex_per_s": 866.8223951996877, "data_frac": 0.05733344722101531, "mem_mb": 199.197696, "nll": -6.051549911499023, "elapsed_s": 713.2}493{"event": "train", "step": 9680, "loss": -5.9368157386779785, "grad_norm": 0.19179736077785492, "lr": 5.578270575938211e-05, "ex_per_s": 868.3535898058033, "data_frac": 0.057333045596618976, "mem_mb": 199.197696, "nll": -5.9368157386779785, "elapsed_s": 713.4}494{"event": "train", "step": 9700, "loss": -6.192706108093262, "grad_norm": 0.6158312559127808, "lr": 5.5354502174671515e-05, "ex_per_s": 869.8842583057328, "data_frac": 0.057332488670271714, "mem_mb": 199.197696, "nll": -6.192706108093262, "elapsed_s": 713.7}495{"event": "train", "step": 9720, "loss": -6.034080505371094, "grad_norm": 0.24731945991516113, "lr": 5.492951545988398e-05, "ex_per_s": 871.4140721050137, "data_frac": 0.05733158012155398, "mem_mb": 199.197696, "nll": -6.034080505371094, "elapsed_s": 713.9}496{"event": "train", "step": 9740, "loss": -5.920910358428955, "grad_norm": 0.2964785695075989, "lr": 5.4507758083629804e-05, "ex_per_s": 872.9419359017471, "data_frac": 0.05733105156795467, "mem_mb": 199.197696, "nll": -5.920910358428955, "elapsed_s": 714.1}497{"event": "train", "step": 9760, "loss": -6.048220157623291, "grad_norm": 0.755578875541687, "lr": 5.408924241977427e-05, "ex_per_s": 874.4704207377736, "data_frac": 0.05733071765838761, "mem_mb": 199.197696, "nll": -6.048220157623291, "elapsed_s": 714.3}498{"event": "train", "step": 9780, "loss": -6.060220241546631, "grad_norm": 0.17987856268882751, "lr": 5.367398074707458e-05, "ex_per_s": 875.9963405097945, "data_frac": 0.057329905728507984, "mem_mb": 199.197696, "nll": -6.060220241546631, "elapsed_s": 714.5}499{"event": "train", "step": 9800, "loss": -5.854714393615723, "grad_norm": 0.5165004730224609, "lr": 5.326198524881979e-05, "ex_per_s": 877.5193877658955, "data_frac": 0.05732982141997776, "mem_mb": 199.197696, "nll": -5.854714393615723, "elapsed_s": 714.7}500{"event": "train", "step": 9820, "loss": -6.09135627746582, "grad_norm": 0.287311315536499, "lr": 5.285326801247307e-05, "ex_per_s": 879.0403573160642, "data_frac": 0.05733002852486827, "mem_mb": 199.197696, "nll": -6.09135627746582, "elapsed_s": 715.0}501{"event": "train", "step": 9840, "loss": -5.95468282699585, "grad_norm": 0.16592954099178314, "lr": 5.244784102931745e-05, "ex_per_s": 880.561997456232, "data_frac": 0.05733008890453378, "mem_mb": 199.197696, "nll": -5.95468282699585, "elapsed_s": 715.2}502{"event": "train", "step": 9860, "loss": -5.878232955932617, "grad_norm": 0.24001412093639374, "lr": 5.2045716194103555e-05, "ex_per_s": 882.0807059454694, "data_frac": 0.05733028863195039, "mem_mb": 199.197696, "nll": -5.878232955932617, "elapsed_s": 715.4}503{"event": "train", "step": 9880, "loss": -5.8918352127075195, "grad_norm": 0.2937087416648865, "lr": 5.164690530470111e-05, "ex_per_s": 883.6037396606889, "data_frac": 0.05732929982673044, "mem_mb": 199.197696, "nll": -5.8918352127075195, "elapsed_s": 715.6}504{"event": "train", "step": 9900, "loss": -5.9424262046813965, "grad_norm": 0.15483412146568298, "lr": 5.125142006175238e-05, "ex_per_s": 885.1253973202778, "data_frac": 0.057329017962284665, "mem_mb": 199.197696, "nll": -5.9424262046813965, "elapsed_s": 715.8}505{"event": "train", "step": 9920, "loss": -5.882998943328857, "grad_norm": 0.33925095200538635, "lr": 5.085927206832921e-05, "ex_per_s": 886.6457028960839, "data_frac": 0.05732862859094856, "mem_mb": 199.197696, "nll": -5.882998943328857, "elapsed_s": 716.1}506{"event": "train", "step": 9940, "loss": -5.962901592254639, "grad_norm": 0.33542951941490173, "lr": 5.047047282959241e-05, "ex_per_s": 888.1642275894587, "data_frac": 0.05732831090694346, "mem_mb": 199.197696, "nll": -5.962901592254639, "elapsed_s": 716.3}507{"event": "train", "step": 9960, "loss": -6.136005401611328, "grad_norm": 0.6240231990814209, "lr": 5.0085033752454274e-05, "ex_per_s": 889.677662582034, "data_frac": 0.057328699694874495, "mem_mb": 199.197696, "nll": -6.136005401611328, "elapsed_s": 716.5}508{"event": "train", "step": 9980, "loss": -5.807998180389404, "grad_norm": 1.100656270980835, "lr": 4.970296614524392e-05, "ex_per_s": 891.1954499523504, "data_frac": 0.05732820421152139, "mem_mb": 199.197696, "nll": -5.807998180389404, "elapsed_s": 716.7}509{"event": "train", "step": 10000, "loss": -5.943007469177246, "grad_norm": 0.5241106748580933, "lr": 4.9324281217375474e-05, "ex_per_s": 892.7109043845641, "data_frac": 0.05732773568004234, "mem_mb": 199.197696, "nll": -5.943007469177246, "elapsed_s": 716.9}510{"event": "val", "step": 10000, "val_nll": -6.451684055723319, "best": -6.451684055723319, "bad": 0, "mem_mb": 199.197696, "elapsed_s": 779.9}511{"event": "train", "step": 10020, "loss": -6.075315475463867, "grad_norm": 0.5745468735694885, "lr": 4.8948990079019224e-05, "ex_per_s": 822.0036081761924, "data_frac": 0.05269823636939964, "mem_mb": 199.197696, "nll": -6.075315475463867, "elapsed_s": 780.1}512{"event": "train", "step": 10040, "loss": -6.077851295471191, "grad_norm": 0.27413713932037354, "lr": 4.857710374077574e-05, "ex_per_s": 823.4150410827294, "data_frac": 0.052699090083525586, "mem_mb": 199.197696, "nll": -6.077851295471191, "elapsed_s": 780.4}513{"event": "train", "step": 10060, "loss": -5.756053447723389, "grad_norm": 0.21443793177604675, "lr": 4.820863311335266e-05, "ex_per_s": 824.8246188977218, "data_frac": 0.052700028017108554, "mem_mb": 199.197696, "nll": -5.756053447723389, "elapsed_s": 780.6}514{"event": "train", "step": 10080, "loss": -6.182244300842285, "grad_norm": 0.15446701645851135, "lr": 4.78435890072448e-05, "ex_per_s": 826.233083976691, "data_frac": 0.052701123652266965, "mem_mb": 199.197696, "nll": -6.182244300842285, "elapsed_s": 780.8}515{"event": "train", "step": 10100, "loss": -5.976241588592529, "grad_norm": 0.7938621044158936, "lr": 4.748198213241678e-05, "ex_per_s": 827.6435508003837, "data_frac": 0.05270170215693444, "mem_mb": 199.197696, "nll": -5.976241588592529, "elapsed_s": 781.0}516{"event": "train", "step": 10120, "loss": -6.0634026527404785, "grad_norm": 0.4212263524532318, "lr": 4.71238230979889e-05, "ex_per_s": 829.0531263479841, "data_frac": 0.05270215636327982, "mem_mb": 199.197696, "nll": -6.0634026527404785, "elapsed_s": 781.2}517{"event": "train", "step": 10140, "loss": -5.970485687255859, "grad_norm": 0.3382403552532196, "lr": 4.676912241192593e-05, "ex_per_s": 830.4633404094351, "data_frac": 0.05270225119219254, "mem_mb": 199.197696, "nll": -5.970485687255859, "elapsed_s": 781.4}518{"event": "train", "step": 10160, "loss": -5.775796413421631, "grad_norm": 0.22028888761997223, "lr": 4.641789048072876e-05, "ex_per_s": 831.8717432708734, "data_frac": 0.05270283238154081, "mem_mb": 199.197696, "nll": -5.775796413421631, "elapsed_s": 781.7}519{"event": "train", "step": 10180, "loss": -6.011634349822998, "grad_norm": 0.3414095342159271, "lr": 4.6070137609129025e-05, "ex_per_s": 833.2790535547252, "data_frac": 0.05270321764938261, "mem_mb": 199.197696, "nll": -6.011634349822998, "elapsed_s": 781.9}520{"event": "train", "step": 10200, "loss": -5.959774017333984, "grad_norm": 0.21353963017463684, "lr": 4.57258739997869e-05, "ex_per_s": 834.6851168787765, "data_frac": 0.05270420956712697, "mem_mb": 199.197696, "nll": -5.959774017333984, "elapsed_s": 782.1}521{"event": "train", "step": 10220, "loss": -5.783790588378906, "grad_norm": 0.1573418825864792, "lr": 4.538510975299176e-05, "ex_per_s": 836.0906578569688, "data_frac": 0.0527047583781867, "mem_mb": 199.197696, "nll": -5.783790588378906, "elapsed_s": 782.3}522{"event": "train", "step": 10240, "loss": -6.02581787109375, "grad_norm": 0.21529781818389893, "lr": 4.504785486636568e-05, "ex_per_s": 837.4968672928813, "data_frac": 0.05270543347130386, "mem_mb": 199.197696, "nll": -6.02581787109375, "elapsed_s": 782.5}523{"event": "train", "step": 10260, "loss": -6.033553123474121, "grad_norm": 0.2232770323753357, "lr": 4.471411923457034e-05, "ex_per_s": 838.9014731281358, "data_frac": 0.052705590211154056, "mem_mb": 199.197696, "nll": -6.033553123474121, "elapsed_s": 782.7}524{"event": "train", "step": 10280, "loss": -6.064607620239258, "grad_norm": 0.2886476218700409, "lr": 4.438391264901654e-05, "ex_per_s": 840.305873213182, "data_frac": 0.05270586804646606, "mem_mb": 199.197696, "nll": -6.064607620239258, "elapsed_s": 783.0}525{"event": "train", "step": 10300, "loss": -5.994340419769287, "grad_norm": 0.2422480434179306, "lr": 4.405724479757712e-05, "ex_per_s": 841.7079927137562, "data_frac": 0.0527064521543016, "mem_mb": 199.197696, "nll": -5.994340419769287, "elapsed_s": 783.2}526{"event": "train", "step": 10320, "loss": -5.90533447265625, "grad_norm": 0.2637682557106018, "lr": 4.373412526430252e-05, "ex_per_s": 843.1099713689144, "data_frac": 0.05270673927675428, "mem_mb": 199.197696, "nll": -5.90533447265625, "elapsed_s": 783.4}527{"event": "train", "step": 10340, "loss": -5.950149059295654, "grad_norm": 0.2507370710372925, "lr": 4.341456352913975e-05, "ex_per_s": 844.511612805989, "data_frac": 0.05270761628854195, "mem_mb": 199.197696, "nll": -5.950149059295654, "elapsed_s": 783.6}528{"event": "train", "step": 10360, "loss": -5.7980194091796875, "grad_norm": 0.12763892114162445, "lr": 4.3098568967654146e-05, "ex_per_s": 845.9136548539698, "data_frac": 0.05270833310361631, "mem_mb": 199.197696, "nll": -5.7980194091796875, "elapsed_s": 783.8}529{"event": "train", "step": 10380, "loss": -5.984067440032959, "grad_norm": 0.392332524061203, "lr": 4.27861508507545e-05, "ex_per_s": 847.3128832497987, "data_frac": 0.05270939810230124, "mem_mb": 199.197696, "nll": -5.984067440032959, "elapsed_s": 784.0}530{"event": "train", "step": 10400, "loss": -5.974179744720459, "grad_norm": 0.4604610800743103, "lr": 4.247731834442082e-05, "ex_per_s": 848.7111437665063, "data_frac": 0.05271030805866854, "mem_mb": 199.197696, "nll": -5.974179744720459, "elapsed_s": 784.3}531{"event": "train", "step": 10420, "loss": -6.037562370300293, "grad_norm": 0.5857800245285034, "lr": 4.2172080509435485e-05, "ex_per_s": 850.1100476893328, "data_frac": 0.05271099889881943, "mem_mb": 199.197696, "nll": -6.037562370300293, "elapsed_s": 784.5}532{"event": "train", "step": 10440, "loss": -6.13021993637085, "grad_norm": 0.687355101108551, "lr": 4.1870446301117574e-05, "ex_per_s": 851.507807217509, "data_frac": 0.052711481576431414, "mem_mb": 199.197696, "nll": -6.13021993637085, "elapsed_s": 784.7}533{"event": "train", "step": 10460, "loss": -5.972776889801025, "grad_norm": 0.14646418392658234, "lr": 4.157242456905989e-05, "ex_per_s": 852.9037548315002, "data_frac": 0.05271262603615956, "mem_mb": 199.197696, "nll": -5.972776889801025, "elapsed_s": 784.9}534{"event": "train", "step": 10480, "loss": -5.96760892868042, "grad_norm": 0.7113385796546936, "lr": 4.1278024056869486e-05, "ex_per_s": 854.3002599017562, "data_frac": 0.05271332154406751, "mem_mb": 199.197696, "nll": -5.96760892868042, "elapsed_s": 785.1}535{"event": "train", "step": 10500, "loss": -5.9984331130981445, "grad_norm": 0.20053569972515106, "lr": 4.0987253401911067e-05, "ex_per_s": 855.6931316988605, "data_frac": 0.0527138876718542, "mem_mb": 199.197696, "nll": -5.9984331130981445, "elapsed_s": 785.3}536{"event": "train", "step": 10520, "loss": -6.00521993637085, "grad_norm": 0.35068824887275696, "lr": 4.070012113505359e-05, "ex_per_s": 857.0898245055043, "data_frac": 0.05271424979640747, "mem_mb": 199.197696, "nll": -6.00521993637085, "elapsed_s": 785.5}537{"event": "train", "step": 10540, "loss": -5.779522895812988, "grad_norm": 0.5470522046089172, "lr": 4.041663568041998e-05, "ex_per_s": 858.48619244771, "data_frac": 0.05271383555438532, "mem_mb": 199.197696, "nll": -5.779522895812988, "elapsed_s": 785.8}538{"event": "train", "step": 10560, "loss": -5.868712425231934, "grad_norm": 0.37769240140914917, "lr": 4.013680535514001e-05, "ex_per_s": 859.881269673883, "data_frac": 0.05271362638700649, "mem_mb": 199.197696, "nll": -5.868712425231934, "elapsed_s": 786.0}539{"event": "train", "step": 10580, "loss": -5.864528179168701, "grad_norm": 0.6197869181632996, "lr": 3.986063836910623e-05, "ex_per_s": 861.2762989595226, "data_frac": 0.05271353664642747, "mem_mb": 199.197696, "nll": -5.864528179168701, "elapsed_s": 786.2}540{"event": "train", "step": 10600, "loss": -5.8906755447387695, "grad_norm": 0.5011329650878906, "lr": 3.958814282473316e-05, "ex_per_s": 862.6701876661366, "data_frac": 0.05271315545133231, "mem_mb": 199.197696, "nll": -5.8906755447387695, "elapsed_s": 786.4}541{"event": "train", "step": 10620, "loss": -6.043774127960205, "grad_norm": 0.18423713743686676, "lr": 3.9319326716719496e-05, "ex_per_s": 864.0622209353616, "data_frac": 0.05271294232306817, "mem_mb": 199.197696, "nll": -6.043774127960205, "elapsed_s": 786.6}542{"event": "train", "step": 10640, "loss": -5.8184309005737305, "grad_norm": 0.3067433536052704, "lr": 3.905419793181364e-05, "ex_per_s": 865.453506626687, "data_frac": 0.05271242365098396, "mem_mb": 199.197696, "nll": -5.8184309005737305, "elapsed_s": 786.8}543{"event": "train", "step": 10660, "loss": -5.851380825042725, "grad_norm": 0.3107651174068451, "lr": 3.87927642485822e-05, "ex_per_s": 866.8447717412995, "data_frac": 0.052712054627245826, "mem_mb": 199.197696, "nll": -5.851380825042725, "elapsed_s": 787.0}544{"event": "train", "step": 10680, "loss": -5.950866222381592, "grad_norm": 0.2854960262775421, "lr": 3.853503333718199e-05, "ex_per_s": 868.2354095523426, "data_frac": 0.052711921639623134, "mem_mb": 199.197696, "nll": -5.950866222381592, "elapsed_s": 787.3}545{"event": "train", "step": 10700, "loss": -5.8038763999938965, "grad_norm": 0.6423882246017456, "lr": 3.8281012759134686e-05, "ex_per_s": 869.6243369439229, "data_frac": 0.052712202199317985, "mem_mb": 199.197696, "nll": -5.8038763999938965, "elapsed_s": 787.5}546{"event": "train", "step": 10720, "loss": -5.8781585693359375, "grad_norm": 0.33448633551597595, "lr": 3.803070996710526e-05, "ex_per_s": 871.0137004704593, "data_frac": 0.05271241857969714, "mem_mb": 199.197696, "nll": -5.8781585693359375, "elapsed_s": 787.7}547{"event": "train", "step": 10740, "loss": -6.1008405685424805, "grad_norm": 0.4104260504245758, "lr": 3.778413230468319e-05, "ex_per_s": 872.4014460873441, "data_frac": 0.05271242854695518, "mem_mb": 199.197696, "nll": -6.1008405685424805, "elapsed_s": 787.9}548{"event": "train", "step": 10760, "loss": -5.963177680969238, "grad_norm": 0.33241793513298035, "lr": 3.754128700616701e-05, "ex_per_s": 873.7898228878454, "data_frac": 0.05271237749603393, "mem_mb": 199.197696, "nll": -5.963177680969238, "elapsed_s": 788.1}549{"event": "train", "step": 10780, "loss": -6.021379470825195, "grad_norm": 0.48837655782699585, "lr": 3.7302181196352115e-05, "ex_per_s": 875.1747517813864, "data_frac": 0.05271263276843511, "mem_mb": 199.197696, "nll": -6.021379470825195, "elapsed_s": 788.3}550{"event": "train", "step": 10800, "loss": -5.931765556335449, "grad_norm": 0.3982764780521393, "lr": 3.7066821890321684e-05, "ex_per_s": 876.5597580469654, "data_frac": 0.05271234302269708, "mem_mb": 199.197696, "nll": -5.931765556335449, "elapsed_s": 788.5}551{"event": "train", "step": 10820, "loss": -6.020315647125244, "grad_norm": 0.22086375951766968, "lr": 3.683521599324089e-05, "ex_per_s": 877.9465148139503, "data_frac": 0.052711543330116066, "mem_mb": 199.197696, "nll": -6.020315647125244, "elapsed_s": 788.8}552{"event": "train", "step": 10840, "loss": -5.962944030761719, "grad_norm": 0.6199668645858765, "lr": 3.660737030015427e-05, "ex_per_s": 879.3327276020185, "data_frac": 0.052710797320043266, "mem_mb": 199.197696, "nll": -5.962944030761719, "elapsed_s": 789.0}553{"event": "train", "step": 10860, "loss": -5.848615646362305, "grad_norm": 0.22301025688648224, "lr": 3.638329149578648e-05, "ex_per_s": 880.7179055697779, "data_frac": 0.052710610830064, "mem_mb": 199.197696, "nll": -5.848615646362305, "elapsed_s": 789.2}554{"event": "train", "step": 10880, "loss": -5.998461723327637, "grad_norm": 0.16047967970371246, "lr": 3.6162986154346006e-05, "ex_per_s": 882.1025525243698, "data_frac": 0.05271000959741892, "mem_mb": 199.197696, "nll": -5.998461723327637, "elapsed_s": 789.4}555{"event": "train", "step": 10900, "loss": -6.120952129364014, "grad_norm": 0.34282130002975464, "lr": 3.594646073933244e-05, "ex_per_s": 883.4856873998583, "data_frac": 0.05270961442233977, "mem_mb": 199.197696, "nll": -6.120952129364014, "elapsed_s": 789.6}556{"event": "train", "step": 10920, "loss": -5.895169734954834, "grad_norm": 0.2747628390789032, "lr": 3.5733721603346695e-05, "ex_per_s": 884.8677971657729, "data_frac": 0.052709072124726114, "mem_mb": 199.197696, "nll": -5.895169734954834, "elapsed_s": 789.8}557{"event": "train", "step": 10940, "loss": -5.895615577697754, "grad_norm": 0.3635508120059967, "lr": 3.552477498790482e-05, "ex_per_s": 886.2457582585273, "data_frac": 0.0527088957143321, "mem_mb": 199.197696, "nll": -5.895615577697754, "elapsed_s": 790.0}558{"event": "train", "step": 10960, "loss": -6.125649452209473, "grad_norm": 0.41625797748565674, "lr": 3.531962702325468e-05, "ex_per_s": 887.6260204379786, "data_frac": 0.05270829350263232, "mem_mb": 199.197696, "nll": -6.125649452209473, "elapsed_s": 790.2}559{"event": "train", "step": 10980, "loss": -5.906558036804199, "grad_norm": 0.2395995408296585, "lr": 3.511828372819624e-05, "ex_per_s": 889.0059733699532, "data_frac": 0.05270823538221671, "mem_mb": 199.197696, "nll": -5.906558036804199, "elapsed_s": 790.5}560{"event": "train", "step": 11000, "loss": -5.91797399520874, "grad_norm": 0.14804349839687347, "lr": 3.4920751009904904e-05, "ex_per_s": 890.3838548419686, "data_frac": 0.05270811318901544, "mem_mb": 199.197696, "nll": -5.91797399520874, "elapsed_s": 790.7}561{"event": "val", "step": 11000, "val_nll": -6.451640294365718, "best": -6.451684055723319, "bad": 1, "mem_mb": 199.197696, "elapsed_s": 853.6}562{"event": "train", "step": 11020, "loss": -5.807902812957764, "grad_norm": 0.35360926389694214, "lr": 3.472703466375828e-05, "ex_per_s": 825.9846558066432, "data_frac": 0.04882279109212525, "mem_mb": 199.197696, "nll": -5.807902812957764, "elapsed_s": 853.9}563{"event": "train", "step": 11040, "loss": -5.983648777008057, "grad_norm": 0.5074911713600159, "lr": 3.4537140373166085e-05, "ex_per_s": 827.2703293780842, "data_frac": 0.04882512770978099, "mem_mb": 199.197696, "nll": -5.983648777008057, "elapsed_s": 854.1}564{"event": "train", "step": 11060, "loss": -5.984896183013916, "grad_norm": 0.2853129506111145, "lr": 3.4351073709403385e-05, "ex_per_s": 828.5557546218839, "data_frac": 0.048827864311580496, "mem_mb": 199.197696, "nll": -5.984896183013916, "elapsed_s": 854.3}565{"event": "train", "step": 11080, "loss": -6.027288436889648, "grad_norm": 0.3855433464050293, "lr": 3.4168840131447255e-05, "ex_per_s": 829.8408137255313, "data_frac": 0.04883010573523416, "mem_mb": 199.197696, "nll": -6.027288436889648, "elapsed_s": 854.5}566{"event": "train", "step": 11100, "loss": -6.142112731933594, "grad_norm": 0.6927661299705505, "lr": 3.399044498581645e-05, "ex_per_s": 831.1245736031129, "data_frac": 0.048832657960081925, "mem_mb": 199.197696, "nll": -6.142112731933594, "elapsed_s": 854.8}567{"event": "train", "step": 11120, "loss": -5.910508155822754, "grad_norm": 0.6371989250183105, "lr": 3.3815893506414745e-05, "ex_per_s": 832.4073656328181, "data_frac": 0.04883544994449904, "mem_mb": 199.197696, "nll": -5.910508155822754, "elapsed_s": 855.0}568{"event": "train", "step": 11140, "loss": -5.884253025054932, "grad_norm": 0.3529949188232422, "lr": 3.364519081437718e-05, "ex_per_s": 833.6896249435076, "data_frac": 0.048837330808681946, "mem_mb": 199.197696, "nll": -5.884253025054932, "elapsed_s": 855.2}569{"event": "train", "step": 11160, "loss": -6.078900337219238, "grad_norm": 0.16069845855236053, "lr": 3.347834191791993e-05, "ex_per_s": 834.9710625682599, "data_frac": 0.048839890836041296, "mem_mb": 199.197696, "nll": -6.078900337219238, "elapsed_s": 855.4}570{"event": "train", "step": 11180, "loss": -6.065895080566406, "grad_norm": 0.25485071539878845, "lr": 3.331535171219332e-05, "ex_per_s": 836.2537066080414, "data_frac": 0.04884215733249092, "mem_mb": 199.197696, "nll": -6.065895080566406, "elapsed_s": 855.6}571{"event": "train", "step": 11200, "loss": -6.1770548820495605, "grad_norm": 0.17675036191940308, "lr": 3.3156224979138305e-05, "ex_per_s": 837.5328769014152, "data_frac": 0.048844640784866965, "mem_mb": 199.197696, "nll": -6.1770548820495605, "elapsed_s": 855.9}572{"event": "train", "step": 11220, "loss": -5.9883036613464355, "grad_norm": 0.2942616939544678, "lr": 3.300096638734598e-05, "ex_per_s": 838.8127159208055, "data_frac": 0.0488471881871424, "mem_mb": 199.197696, "nll": -5.9883036613464355, "elapsed_s": 856.1}573{"event": "train", "step": 11240, "loss": -5.966488838195801, "grad_norm": 0.3189751207828522, "lr": 3.2849580491920845e-05, "ex_per_s": 840.0899559160518, "data_frac": 0.04884976397301564, "mem_mb": 199.197696, "nll": -5.966488838195801, "elapsed_s": 856.3}574{"event": "train", "step": 11260, "loss": -6.058408260345459, "grad_norm": 3.1072635650634766, "lr": 3.2702071734346924e-05, "ex_per_s": 841.3685879018996, "data_frac": 0.04885233077713954, "mem_mb": 199.197696, "nll": -6.058408260345459, "elapsed_s": 856.5}575{"event": "train", "step": 11280, "loss": -6.106337547302246, "grad_norm": 0.4310399889945984, "lr": 3.255844444235771e-05, "ex_per_s": 842.646411648586, "data_frac": 0.04885516409947056, "mem_mb": 199.197696, "nll": -6.106337547302246, "elapsed_s": 856.7}576{"event": "train", "step": 11300, "loss": -5.938758373260498, "grad_norm": 0.2991997301578522, "lr": 3.2418702829808946e-05, "ex_per_s": 843.9257510916424, "data_frac": 0.048857376382893676, "mem_mb": 199.197696, "nll": -5.938758373260498, "elapsed_s": 857.0}577{"event": "train", "step": 11320, "loss": -5.861535549163818, "grad_norm": 0.31312936544418335, "lr": 3.228285099655516e-05, "ex_per_s": 845.2037870288901, "data_frac": 0.04885922409057216, "mem_mb": 199.197696, "nll": -5.861535549163818, "elapsed_s": 857.2}578{"event": "train", "step": 11340, "loss": -5.928648471832275, "grad_norm": 0.4681607186794281, "lr": 3.215089292832933e-05, "ex_per_s": 846.4813253429627, "data_frac": 0.048861563052613896, "mem_mb": 199.197696, "nll": -5.928648471832275, "elapsed_s": 857.4}579{"event": "train", "step": 11360, "loss": -5.9611616134643555, "grad_norm": 0.23459981381893158, "lr": 3.202283249662597e-05, "ex_per_s": 847.7560397611494, "data_frac": 0.04886386398026978, "mem_mb": 199.197696, "nll": -5.9611616134643555, "elapsed_s": 857.6}580{"event": "train", "step": 11380, "loss": -5.909996032714844, "grad_norm": 0.1668248325586319, "lr": 3.189867345858747e-05, "ex_per_s": 849.0303324235676, "data_frac": 0.04886596374398161, "mem_mb": 199.197696, "nll": -5.909996032714844, "elapsed_s": 857.8}581{"event": "train", "step": 11400, "loss": -6.081756591796875, "grad_norm": 0.41980525851249695, "lr": 3.177841945689397e-05, "ex_per_s": 850.3043920691533, "data_frac": 0.04886790522247877, "mem_mb": 199.197696, "nll": -6.081756591796875, "elapsed_s": 858.1}582{"event": "train", "step": 11420, "loss": -5.832974433898926, "grad_norm": 0.17915330827236176, "lr": 3.1662074019656414e-05, "ex_per_s": 851.5797404428038, "data_frac": 0.04887016632520255, "mem_mb": 199.197696, "nll": -5.832974433898926, "elapsed_s": 858.3}583{"event": "train", "step": 11440, "loss": -5.884792327880859, "grad_norm": 0.6790859699249268, "lr": 3.154964056031306e-05, "ex_per_s": 852.8537414974714, "data_frac": 0.048872004054468274, "mem_mb": 199.197696, "nll": -5.884792327880859, "elapsed_s": 858.5}584{"event": "train", "step": 11460, "loss": -6.137067794799805, "grad_norm": 0.5836895704269409, "lr": 3.1441122377529333e-05, "ex_per_s": 854.1265790534246, "data_frac": 0.04887319475369488, "mem_mb": 199.197696, "nll": -6.137067794799805, "elapsed_s": 858.7}585{"event": "train", "step": 11480, "loss": -5.973583221435547, "grad_norm": 0.2574062645435333, "lr": 3.133652265510106e-05, "ex_per_s": 855.4001709392994, "data_frac": 0.04887501662812914, "mem_mb": 199.197696, "nll": -5.973583221435547, "elapsed_s": 858.9}586{"event": "train", "step": 11500, "loss": -5.937276840209961, "grad_norm": 0.4883059859275818, "lr": 3.123584446186108e-05, "ex_per_s": 856.669180255945, "data_frac": 0.048877982478776526, "mem_mb": 199.197696, "nll": -5.937276840209961, "elapsed_s": 859.1}587{"event": "train", "step": 11520, "loss": -5.906184196472168, "grad_norm": 0.687404990196228, "lr": 3.113909075158913e-05, "ex_per_s": 857.9392039642203, "data_frac": 0.04888042739520704, "mem_mb": 199.197696, "nll": -5.906184196472168, "elapsed_s": 859.4}588{"event": "train", "step": 11540, "loss": -6.0453362464904785, "grad_norm": 0.14672644436359406, "lr": 3.104626436292525e-05, "ex_per_s": 859.2080171117681, "data_frac": 0.04888287053933303, "mem_mb": 199.197696, "nll": -6.0453362464904785, "elapsed_s": 859.6}589{"event": "train", "step": 11560, "loss": -6.114807605743408, "grad_norm": 0.22907289862632751, "lr": 3.095736801928647e-05, "ex_per_s": 860.4783404027132, "data_frac": 0.04888520978896826, "mem_mb": 199.197696, "nll": -6.114807605743408, "elapsed_s": 859.8}590{"event": "train", "step": 11580, "loss": -5.959295272827148, "grad_norm": 0.2049648016691208, "lr": 3.087240432878696e-05, "ex_per_s": 861.7438622299454, "data_frac": 0.0488885019282573, "mem_mb": 199.197696, "nll": -5.959295272827148, "elapsed_s": 860.0}591{"event": "train", "step": 11600, "loss": -6.073004245758057, "grad_norm": 0.21902291476726532, "lr": 3.0791375784161455e-05, "ex_per_s": 863.0096726683256, "data_frac": 0.0488912327732441, "mem_mb": 199.197696, "nll": -6.073004245758057, "elapsed_s": 860.3}592{"event": "train", "step": 11620, "loss": -6.02445125579834, "grad_norm": 0.3206517994403839, "lr": 3.0714284762692116e-05, "ex_per_s": 864.2782285875535, "data_frac": 0.048893766523654465, "mem_mb": 199.197696, "nll": -6.02445125579834, "elapsed_s": 860.5}593{"event": "train", "step": 11640, "loss": -5.965457916259766, "grad_norm": 0.1713353544473648, "lr": 3.0641133526138836e-05, "ex_per_s": 865.5446072625226, "data_frac": 0.04889590431989016, "mem_mb": 199.197696, "nll": -5.965457916259766, "elapsed_s": 860.7}594{"event": "train", "step": 11660, "loss": -5.919635772705078, "grad_norm": 0.3123888075351715, "lr": 3.0571924220672835e-05, "ex_per_s": 866.8104124742641, "data_frac": 0.048898175883527774, "mem_mb": 199.197696, "nll": -5.919635772705078, "elapsed_s": 860.9}595{"event": "train", "step": 11680, "loss": -5.9642863273620605, "grad_norm": 0.29894450306892395, "lr": 3.0506658876813734e-05, "ex_per_s": 868.0771420860525, "data_frac": 0.04890033846448027, "mem_mb": 199.197696, "nll": -5.9642863273620605, "elapsed_s": 861.1}596{"event": "train", "step": 11700, "loss": -6.048731803894043, "grad_norm": 0.18587692081928253, "lr": 3.0445339409369934e-05, "ex_per_s": 869.3444670651745, "data_frac": 0.048902157397392794, "mem_mb": 199.197696, "nll": -6.048731803894043, "elapsed_s": 861.3}597{"event": "train", "step": 11720, "loss": -5.767541408538818, "grad_norm": 0.29071491956710815, "lr": 3.0387967617382475e-05, "ex_per_s": 870.6090929138187, "data_frac": 0.04890443609247213, "mem_mb": 199.197696, "nll": -5.767541408538818, "elapsed_s": 861.6}598{"event": "train", "step": 11740, "loss": -6.04293155670166, "grad_norm": 0.5333802103996277, "lr": 3.0334545184072252e-05, "ex_per_s": 871.874517759372, "data_frac": 0.04890647884565498, "mem_mb": 199.197696, "nll": -6.04293155670166, "elapsed_s": 861.8}599{"event": "train", "step": 11760, "loss": -5.9006428718566895, "grad_norm": 0.2929174304008484, "lr": 3.0285073676790606e-05, "ex_per_s": 873.1383170424714, "data_frac": 0.0489085325183384, "mem_mb": 199.197696, "nll": -5.9006428718566895, "elapsed_s": 862.0}600{"event": "train", "step": 11780, "loss": -6.0862202644348145, "grad_norm": 0.14244844019412994, "lr": 3.0239554546973377e-05, "ex_per_s": 874.4006265035777, "data_frac": 0.048910495633361716, "mem_mb": 199.197696, "nll": -6.0862202644348145, "elapsed_s": 862.2}601{"event": "train", "step": 11800, "loss": -5.965174674987793, "grad_norm": 0.33696284890174866, "lr": 3.0197989130098285e-05, "ex_per_s": 875.6636581152953, "data_frac": 0.04891240127610053, "mem_mb": 199.197696, "nll": -5.965174674987793, "elapsed_s": 862.4}602{"event": "train", "step": 11820, "loss": -5.970404148101807, "grad_norm": 0.7969751358032227, "lr": 3.0160378645645744e-05, "ex_per_s": 876.9264291119985, "data_frac": 0.048914007549988855, "mem_mb": 199.197696, "nll": -5.970404148101807, "elapsed_s": 862.7}603{"event": "train", "step": 11840, "loss": -6.0319743156433105, "grad_norm": 0.17215025424957275, "lr": 3.0126724197063143e-05, "ex_per_s": 878.1864398288543, "data_frac": 0.04891601450057858, "mem_mb": 199.197696, "nll": -6.0319743156433105, "elapsed_s": 862.9}604{"event": "train", "step": 11860, "loss": -5.986623764038086, "grad_norm": 0.5427295565605164, "lr": 3.0097026771732407e-05, "ex_per_s": 879.4473178167699, "data_frac": 0.04891829746320929, "mem_mb": 199.197696, "nll": -5.986623764038086, "elapsed_s": 863.1}605{"event": "train", "step": 11880, "loss": -6.028807163238525, "grad_norm": 0.6392213702201843, "lr": 3.0071287240941065e-05, "ex_per_s": 880.7082842631629, "data_frac": 0.04892032328308076, "mem_mb": 199.197696, "nll": -6.028807163238525, "elapsed_s": 863.3}606{"event": "train", "step": 11900, "loss": -5.885828018188477, "grad_norm": 0.14996542036533356, "lr": 3.0049506359856666e-05, "ex_per_s": 881.9659583919228, "data_frac": 0.04892239094816451, "mem_mb": 199.197696, "nll": -5.885828018188477, "elapsed_s": 863.5}607{"event": "train", "step": 11920, "loss": -6.011484622955322, "grad_norm": 0.4300556778907776, "lr": 3.0031684767504643e-05, "ex_per_s": 883.2258084561219, "data_frac": 0.04892403920039532, "mem_mb": 199.197696, "nll": -6.011484622955322, "elapsed_s": 863.7}608{"event": "train", "step": 11940, "loss": -6.069141387939453, "grad_norm": 0.7685412764549255, "lr": 3.0017822986749544e-05, "ex_per_s": 884.4826192993814, "data_frac": 0.048925991815950566, "mem_mb": 199.197696, "nll": -6.069141387939453, "elapsed_s": 864.0}609{"event": "train", "step": 11960, "loss": -6.159242630004883, "grad_norm": 0.4395144283771515, "lr": 3.000792142427973e-05, "ex_per_s": 885.7366324879056, "data_frac": 0.048928103141199726, "mem_mb": 199.197696, "nll": -6.159242630004883, "elapsed_s": 864.2}610{"event": "train", "step": 11980, "loss": -6.0917277336120605, "grad_norm": 0.14595115184783936, "lr": 3.000198037059536e-05, "ex_per_s": 886.9937448558156, "data_frac": 0.04892985416476889, "mem_mb": 199.197696, "nll": -6.0917277336120605, "elapsed_s": 864.4}611{"event": "train", "step": 12000, "loss": -5.936335563659668, "grad_norm": 0.24455416202545166, "lr": 2.9999999999999997e-05, "ex_per_s": 888.2509930208824, "data_frac": 0.04893154086494521, "mem_mb": 199.197696, "nll": -5.936335563659668, "elapsed_s": 864.6}612{"event": "val", "step": 12000, "val_nll": -6.451788332870508, "best": -6.451788332870508, "bad": 0, "mem_mb": 199.197696, "elapsed_s": 927.4}613{"event": "done", "steps": 12000, "best_val": -6.451788332870508, "wall_s": 927.3784019947052, "ex_per_s": 828.1409167477947, "peak_mem_mb": 199.197696, "n_params": 231459, "device": "mps", "data_time_frac": 0.045620226071422174, "elapsed_s": 927.4}614
Files · e976cfd0ab0ee7e59d46d9e75fd8b8fb0a861c50
Menu

Theme