{"event": "train", "step": 20, "loss": -6.118469715118408, "grad_norm": 1.0747575759887695, "lr": 3.149999999999999e-05, "ex_per_s": 89.88170665360897, "data_frac": 0.8442207367879958, "mem_mb": 90.356224, "nll": -6.118469715118408, "elapsed_s": 14.3} {"event": "train", "step": 40, "loss": -5.898562908172607, "grad_norm": 0.35184842348098755, "lr": 6.149999999999999e-05, "ex_per_s": 160.4285649975061, "data_frac": 0.7542611757481081, "mem_mb": 90.356224, "nll": -5.898562908172607, "elapsed_s": 16.0} {"event": "train", "step": 60, "loss": -5.976221561431885, "grad_norm": 0.7525873184204102, "lr": 9.149999999999999e-05, "ex_per_s": 217.59658220700365, "data_frac": 0.6831959591826302, "mem_mb": 90.356224, "nll": -5.976221561431885, "elapsed_s": 17.7} {"event": "train", "step": 80, "loss": -6.065645217895508, "grad_norm": 1.3514533042907715, "lr": 0.0001215, "ex_per_s": 264.81302892489344, "data_frac": 0.6242422605722089, "mem_mb": 90.356224, "nll": -6.065645217895508, "elapsed_s": 19.4} {"event": "train", "step": 100, "loss": -5.943587779998779, "grad_norm": 0.1768481731414795, "lr": 0.0001515, "ex_per_s": 304.66805232974616, "data_frac": 0.5750874922510868, "mem_mb": 90.356224, "nll": -5.943587779998779, "elapsed_s": 21.0} {"event": "train", "step": 120, "loss": -5.901042938232422, "grad_norm": 1.011506199836731, "lr": 0.00018149999999999997, "ex_per_s": 338.5651499165287, "data_frac": 0.533069524653012, "mem_mb": 90.356224, "nll": -5.901042938232422, "elapsed_s": 22.7} {"event": "train", "step": 140, "loss": -5.831459999084473, "grad_norm": 0.27906763553619385, "lr": 0.00021149999999999996, "ex_per_s": 367.89064870918185, "data_frac": 0.49693491221006264, "mem_mb": 90.356224, "nll": -5.831459999084473, "elapsed_s": 24.4} {"event": "train", "step": 160, "loss": -5.732961654663086, "grad_norm": 1.1412919759750366, "lr": 0.0002415, "ex_per_s": 393.39309772586375, "data_frac": 0.46537488334935695, "mem_mb": 90.356224, "nll": -5.732961654663086, "elapsed_s": 26.1} {"event": "train", "step": 180, "loss": -5.975906848907471, "grad_norm": 1.1306798458099365, "lr": 0.0002715, "ex_per_s": 415.9111991593079, "data_frac": 0.43773700716875186, "mem_mb": 90.356224, "nll": -5.975906848907471, "elapsed_s": 27.7} {"event": "train", "step": 200, "loss": -5.9706597328186035, "grad_norm": 0.7245384454727173, "lr": 0.0003, "ex_per_s": 435.8221582675084, "data_frac": 0.41321747720923985, "mem_mb": 90.356224, "nll": -5.9706597328186035, "elapsed_s": 29.4} {"event": "train", "step": 220, "loss": -6.027897834777832, "grad_norm": 1.4985913038253784, "lr": 0.00029998154617398326, "ex_per_s": 453.6467408713751, "data_frac": 0.39138792966045843, "mem_mb": 90.356224, "nll": -6.027897834777832, "elapsed_s": 31.1} {"event": "train", "step": 240, "loss": -5.914258003234863, "grad_norm": 0.4416957199573517, "lr": 0.000299926189741025, "ex_per_s": 469.6321643761394, "data_frac": 0.3717629231254004, "mem_mb": 90.356224, "nll": -5.914258003234863, "elapsed_s": 32.8} {"event": "train", "step": 260, "loss": -5.886660099029541, "grad_norm": 0.12627601623535156, "lr": 0.0002998339458350211, "ex_per_s": 484.13549935791707, "data_frac": 0.3540939707855909, "mem_mb": 90.356224, "nll": -5.886660099029541, "elapsed_s": 34.4} {"event": "train", "step": 280, "loss": -5.945527076721191, "grad_norm": 0.2231270670890808, "lr": 0.0002997048396745345, "ex_per_s": 497.28659695812775, "data_frac": 0.3380471223865625, "mem_mb": 90.356224, "nll": -5.945527076721191, "elapsed_s": 36.1} {"event": "train", "step": 300, "loss": -5.967426300048828, "grad_norm": 0.09033843129873276, "lr": 0.0002995389065559004, "ex_per_s": 509.2333675266745, "data_frac": 0.3233914962443593, "mem_mb": 90.356224, "nll": -5.967426300048828, "elapsed_s": 37.7} {"event": "train", "step": 320, "loss": -5.965365409851074, "grad_norm": 0.8686078190803528, "lr": 0.0002993361918435765, "ex_per_s": 520.15685167721, "data_frac": 0.3099851855899566, "mem_mb": 90.356224, "nll": -5.965365409851074, "elapsed_s": 39.4} {"event": "train", "step": 340, "loss": -5.8346943855285645, "grad_norm": 0.23951901495456696, "lr": 0.00029909675095774114, "ex_per_s": 530.2628912592854, "data_frac": 0.29770323995904147, "mem_mb": 90.356224, "nll": -5.8346943855285645, "elapsed_s": 41.1} {"event": "train", "step": 360, "loss": -5.996823787689209, "grad_norm": 0.3420511484146118, "lr": 0.00029882064935914164, "ex_per_s": 539.4608214456969, "data_frac": 0.28632690542529865, "mem_mb": 90.356224, "nll": -5.996823787689209, "elapsed_s": 42.8} {"event": "train", "step": 380, "loss": -6.00636100769043, "grad_norm": 1.2859015464782715, "lr": 0.00029850796253119824, "ex_per_s": 548.0445186416313, "data_frac": 0.2758365351295363, "mem_mb": 90.356224, "nll": -6.00636100769043, "elapsed_s": 44.4} {"event": "train", "step": 400, "loss": -5.996966361999512, "grad_norm": 1.416736125946045, "lr": 0.0002981587759593675, "ex_per_s": 555.9986499965187, "data_frac": 0.2661017839617254, "mem_mb": 90.356224, "nll": -5.996966361999512, "elapsed_s": 46.1} {"event": "train", "step": 420, "loss": -5.962892055511475, "grad_norm": 0.7961797118186951, "lr": 0.0002977731851077716, "ex_per_s": 563.4200987693979, "data_frac": 0.25705500161440653, "mem_mb": 90.356224, "nll": -5.962892055511475, "elapsed_s": 47.8} {"event": "train", "step": 440, "loss": -5.938500881195068, "grad_norm": 0.9401934146881104, "lr": 0.0002973512953930992, "ex_per_s": 570.3370950283648, "data_frac": 0.2486114911241231, "mem_mb": 90.356224, "nll": -5.938500881195068, "elapsed_s": 49.4} {"event": "train", "step": 460, "loss": -6.1029372215271, "grad_norm": 0.5940672755241394, "lr": 0.0002968932221557859, "ex_per_s": 576.8039093417503, "data_frac": 0.24072012750417426, "mem_mb": 90.356224, "nll": -6.1029372215271, "elapsed_s": 51.1} {"event": "train", "step": 480, "loss": -5.878077507019043, "grad_norm": 0.3949568569660187, "lr": 0.0002963990906284808, "ex_per_s": 582.8396526878881, "data_frac": 0.23331682296078401, "mem_mb": 90.356224, "nll": -5.878077507019043, "elapsed_s": 52.8} {"event": "train", "step": 500, "loss": -5.924585819244385, "grad_norm": 0.6613417863845825, "lr": 0.00029586903590180956, "ex_per_s": 588.521072679939, "data_frac": 0.2263771208564529, "mem_mb": 90.356224, "nll": -5.924585819244385, "elapsed_s": 54.4} {"event": "val", "step": 500, "val_nll": -6.439468258477516, "best": -6.439468258477516, "bad": 0, "mem_mb": 90.356224, "elapsed_s": 118.2} {"event": "train", "step": 520, "loss": -6.0408735275268555, "grad_norm": 0.2678622007369995, "lr": 0.00029530320288744205, "ex_per_s": 277.70287904304473, "data_frac": 0.10281304639209761, "mem_mb": 90.356224, "nll": -6.0408735275268555, "elapsed_s": 119.9} {"event": "train", "step": 540, "loss": -5.879276275634766, "grad_norm": 0.5546368956565857, "lr": 0.0002947017462784745, "ex_per_s": 284.4186942376919, "data_frac": 0.10149430101484876, "mem_mb": 90.356224, "nll": -5.879276275634766, "elapsed_s": 121.6} {"event": "train", "step": 560, "loss": -5.97292423248291, "grad_norm": 0.10086830705404282, "lr": 0.00029406483050713824, "ex_per_s": 290.9488533027576, "data_frac": 0.10021393192245893, "mem_mb": 90.356224, "nll": -5.97292423248291, "elapsed_s": 123.2} {"event": "train", "step": 580, "loss": -5.92419958114624, "grad_norm": 0.10033168643712997, "lr": 0.0002933926296998457, "ex_per_s": 297.3088431774561, "data_frac": 0.09896647798995814, "mem_mb": 90.356224, "nll": -5.92419958114624, "elapsed_s": 124.9} {"event": "train", "step": 600, "loss": -5.971851348876953, "grad_norm": 0.8955174684524536, "lr": 0.0002926853276295856, "ex_per_s": 303.49784587954787, "data_frac": 0.09775139082167965, "mem_mb": 90.356224, "nll": -5.971851348876953, "elapsed_s": 126.6} {"event": "train", "step": 620, "loss": -5.953142166137695, "grad_norm": 0.3640809655189514, "lr": 0.00029194311766568174, "ex_per_s": 309.52660561201196, "data_frac": 0.09657004388069669, "mem_mb": 90.356224, "nll": -5.953142166137695, "elapsed_s": 128.2} {"event": "train", "step": 640, "loss": -6.066600322723389, "grad_norm": 0.3582726716995239, "lr": 0.0002911662027209272, "ex_per_s": 315.3962981173624, "data_frac": 0.09541494702930242, "mem_mb": 90.356224, "nll": -6.066600322723389, "elapsed_s": 129.9} {"event": "train", "step": 660, "loss": -6.0075883865356445, "grad_norm": 1.126720905303955, "lr": 0.00029035479519611045, "ex_per_s": 321.12173627742925, "data_frac": 0.09429244005140452, "mem_mb": 90.356224, "nll": -6.0075883865356445, "elapsed_s": 131.6} {"event": "train", "step": 680, "loss": -6.008425235748291, "grad_norm": 0.09793967008590698, "lr": 0.0002895091169219469, "ex_per_s": 326.7038747008564, "data_frac": 0.09319506984648676, "mem_mb": 90.356224, "nll": -6.008425235748291, "elapsed_s": 133.3} {"event": "train", "step": 700, "loss": -5.987247943878174, "grad_norm": 0.7133572697639465, "lr": 0.00028862939909843273, "ex_per_s": 332.14206721776736, "data_frac": 0.09212835977101946, "mem_mb": 90.356224, "nll": -5.987247943878174, "elapsed_s": 134.9} {"event": "train", "step": 720, "loss": -6.063692569732666, "grad_norm": 0.3128707706928253, "lr": 0.0002877158822316369, "ex_per_s": 337.4517260434094, "data_frac": 0.09108551155927144, "mem_mb": 90.356224, "nll": -6.063692569732666, "elapsed_s": 136.6} {"event": "train", "step": 740, "loss": -6.032978057861328, "grad_norm": 0.29146668314933777, "lr": 0.00028676881606794945, "ex_per_s": 342.6330771476756, "data_frac": 0.09006850557628449, "mem_mb": 90.356224, "nll": -6.032978057861328, "elapsed_s": 138.3} {"event": "train", "step": 760, "loss": -5.958232879638672, "grad_norm": 0.988333523273468, "lr": 0.000285788459525803, "ex_per_s": 347.68359493100496, "data_frac": 0.08918244614293878, "mem_mb": 90.356224, "nll": -5.958232879638672, "elapsed_s": 139.9} {"event": "train", "step": 780, "loss": -5.89648962020874, "grad_norm": 0.216472327709198, "lr": 0.00028477508062488754, "ex_per_s": 352.61942337919226, "data_frac": 0.08821320888091583, "mem_mb": 90.356224, "nll": -5.89648962020874, "elapsed_s": 141.6} {"event": "train", "step": 800, "loss": -5.815075874328613, "grad_norm": 0.2756504714488983, "lr": 0.000283728956412876, "ex_per_s": 357.4397248588973, "data_frac": 0.08726513671889792, "mem_mb": 90.356224, "nll": -5.815075874328613, "elapsed_s": 143.3} {"event": "train", "step": 820, "loss": -5.934436321258545, "grad_norm": 0.48752984404563904, "lr": 0.00028265037288968236, "ex_per_s": 362.1628130788346, "data_frac": 0.08634295394045503, "mem_mb": 90.356224, "nll": -5.934436321258545, "elapsed_s": 145.0} {"event": "train", "step": 840, "loss": -5.807565689086914, "grad_norm": 0.5798416137695312, "lr": 0.0002815396249292725, "ex_per_s": 366.78073161607534, "data_frac": 0.08544246256136319, "mem_mb": 90.356224, "nll": -5.807565689086914, "elapsed_s": 146.6} {"event": "train", "step": 860, "loss": -5.913548946380615, "grad_norm": 1.6726946830749512, "lr": 0.00028039701619904774, "ex_per_s": 371.29833624219856, "data_frac": 0.08456250514244623, "mem_mb": 90.356224, "nll": -5.913548946380615, "elapsed_s": 148.3} {"event": "train", "step": 880, "loss": -5.869779586791992, "grad_norm": 0.47643154859542847, "lr": 0.00027922285907682625, "ex_per_s": 375.71479065272507, "data_frac": 0.0837015028335473, "mem_mb": 90.356224, "nll": -5.869779586791992, "elapsed_s": 149.9} {"event": "train", "step": 900, "loss": -5.986288547515869, "grad_norm": 0.3822539448738098, "lr": 0.00027801747456544134, "ex_per_s": 380.03913275689615, "data_frac": 0.08286041941113607, "mem_mb": 90.356224, "nll": -5.986288547515869, "elapsed_s": 151.6} {"event": "train", "step": 920, "loss": -6.025387763977051, "grad_norm": 0.5080310702323914, "lr": 0.00027678119220498265, "ex_per_s": 384.2692314683508, "data_frac": 0.08203914962328655, "mem_mb": 90.356224, "nll": -6.025387763977051, "elapsed_s": 153.3} {"event": "train", "step": 940, "loss": -6.081766605377197, "grad_norm": 0.7302393317222595, "lr": 0.0002755143499827035, "ex_per_s": 388.4109456986307, "data_frac": 0.08123432970742062, "mem_mb": 90.356224, "nll": -6.081766605377197, "elapsed_s": 154.9} {"event": "train", "step": 960, "loss": -5.96905517578125, "grad_norm": 0.31796613335609436, "lr": 0.00027421729424061787, "ex_per_s": 392.4640363671855, "data_frac": 0.08044398129917318, "mem_mb": 90.356224, "nll": -5.96905517578125, "elapsed_s": 156.6} {"event": "train", "step": 980, "loss": -5.926123142242432, "grad_norm": 0.2227289378643036, "lr": 0.00027289037958081483, "ex_per_s": 396.4302715903581, "data_frac": 0.07967331306427966, "mem_mb": 90.356224, "nll": -5.926123142242432, "elapsed_s": 158.3} {"event": "train", "step": 1000, "loss": -6.072259426116943, "grad_norm": 0.20720186829566956, "lr": 0.0002715339687685131, "ex_per_s": 400.3048494417522, "data_frac": 0.07891770887531481, "mem_mb": 90.356224, "nll": -6.072259426116943, "elapsed_s": 159.9} {"event": "val", "step": 1000, "val_nll": -6.44669000423706, "best": -6.44669000423706, "bad": 0, "mem_mb": 90.356224, "elapsed_s": 223.2} {"event": "train", "step": 1020, "loss": -6.032186508178711, "grad_norm": 0.1984001100063324, "lr": 0.0002701484326328852, "ex_per_s": 290.2878156091686, "data_frac": 0.056160932149910886, "mem_mb": 90.356224, "nll": -6.032186508178711, "elapsed_s": 224.9} {"event": "train", "step": 1040, "loss": -5.972660541534424, "grad_norm": 0.3822425305843353, "lr": 0.00026873414996567596, "ex_per_s": 293.7912743669121, "data_frac": 0.05579693430272476, "mem_mb": 90.356224, "nll": -5.972660541534424, "elapsed_s": 226.6} {"event": "train", "step": 1060, "loss": -6.181791305541992, "grad_norm": 0.23389719426631927, "lr": 0.0002672915074176452, "ex_per_s": 297.2433347102272, "data_frac": 0.05543904269224888, "mem_mb": 90.356224, "nll": -6.181791305541992, "elapsed_s": 228.3} {"event": "train", "step": 1080, "loss": -5.814391136169434, "grad_norm": 0.5557996034622192, "lr": 0.00026582089939286075, "ex_per_s": 300.6477245275251, "data_frac": 0.055087158974373225, "mem_mb": 90.356224, "nll": -5.814391136169434, "elapsed_s": 229.9} {"event": "train", "step": 1100, "loss": -6.052751064300537, "grad_norm": 0.09823969006538391, "lr": 0.0002643227279408728, "ex_per_s": 304.0010873738082, "data_frac": 0.0547393582300585, "mem_mb": 90.356224, "nll": -6.052751064300537, "elapsed_s": 231.6} {"event": "train", "step": 1120, "loss": -6.051823139190674, "grad_norm": 0.4194229245185852, "lr": 0.000262797402646797, "ex_per_s": 307.3023335336001, "data_frac": 0.05439727025546397, "mem_mb": 90.356224, "nll": -6.051823139190674, "elapsed_s": 233.3} {"event": "train", "step": 1140, "loss": -6.021182537078857, "grad_norm": 0.22201509773731232, "lr": 0.0002612453405193386, "ex_per_s": 310.5606122845243, "data_frac": 0.05406066475630313, "mem_mb": 90.356224, "nll": -6.021182537078857, "elapsed_s": 235.0} {"event": "train", "step": 1160, "loss": -5.968923091888428, "grad_norm": 0.5865791440010071, "lr": 0.0002596669658767859, "ex_per_s": 313.77141674946324, "data_frac": 0.05372826507324145, "mem_mb": 90.356224, "nll": -5.968923091888428, "elapsed_s": 236.6} {"event": "train", "step": 1180, "loss": -5.958609580993652, "grad_norm": 0.19506745040416718, "lr": 0.00025806271023100637, "ex_per_s": 316.9380760852568, "data_frac": 0.05340041298046802, "mem_mb": 90.356224, "nll": -5.958609580993652, "elapsed_s": 238.3} {"event": "train", "step": 1200, "loss": -5.945618152618408, "grad_norm": 0.21302315592765808, "lr": 0.00025643301216947504, "ex_per_s": 320.0617826831741, "data_frac": 0.053078742765209755, "mem_mb": 90.356224, "nll": -5.945618152618408, "elapsed_s": 240.0} {"event": "train", "step": 1220, "loss": -6.032254695892334, "grad_norm": 0.7467830181121826, "lr": 0.0002547783172353695, "ex_per_s": 323.1423726619093, "data_frac": 0.05276426207353116, "mem_mb": 90.356224, "nll": -6.032254695892334, "elapsed_s": 241.7} {"event": "train", "step": 1240, "loss": -6.059254169464111, "grad_norm": 0.0753287523984909, "lr": 0.00025309907780576277, "ex_per_s": 326.18585544956585, "data_frac": 0.05245156609674275, "mem_mb": 90.356224, "nll": -6.059254169464111, "elapsed_s": 243.3} {"event": "train", "step": 1260, "loss": -6.052968502044678, "grad_norm": 0.21189221739768982, "lr": 0.0002513957529679477, "ex_per_s": 329.1869236627102, "data_frac": 0.05214038813045193, "mem_mb": 90.356224, "nll": -6.052968502044678, "elapsed_s": 245.0} {"event": "train", "step": 1280, "loss": -6.043802261352539, "grad_norm": 0.6110865473747253, "lr": 0.00024966880839392746, "ex_per_s": 332.1467965806565, "data_frac": 0.05183575060789054, "mem_mb": 90.356224, "nll": -6.043802261352539, "elapsed_s": 246.7} {"event": "train", "step": 1300, "loss": -5.869082450866699, "grad_norm": 0.7102776765823364, "lr": 0.0002479187162131052, "ex_per_s": 335.067285028485, "data_frac": 0.05153340405757919, "mem_mb": 90.356224, "nll": -5.869082450866699, "elapsed_s": 248.4} {"event": "train", "step": 1320, "loss": -5.828468322753906, "grad_norm": 0.23735885322093964, "lr": 0.00024614595488320833, "ex_per_s": 337.95111986759287, "data_frac": 0.05123607654747314, "mem_mb": 90.356224, "nll": -5.828468322753906, "elapsed_s": 250.0} {"event": "train", "step": 1340, "loss": -5.95186710357666, "grad_norm": 0.3828621208667755, "lr": 0.00024435100905948387, "ex_per_s": 340.7975127409654, "data_frac": 0.050943078415694384, "mem_mb": 90.356224, "nll": -5.95186710357666, "elapsed_s": 251.7} {"event": "train", "step": 1360, "loss": -6.039181709289551, "grad_norm": 0.5301186442375183, "lr": 0.0002425343694621974, "ex_per_s": 343.6058126135529, "data_frac": 0.050652709846804186, "mem_mb": 90.356224, "nll": -6.039181709289551, "elapsed_s": 253.4} {"event": "train", "step": 1380, "loss": -6.067743301391602, "grad_norm": 0.28117650747299194, "lr": 0.00024069653274247583, "ex_per_s": 346.37911649377816, "data_frac": 0.05036557678837715, "mem_mb": 90.356224, "nll": -6.067743301391602, "elapsed_s": 255.0} {"event": "train", "step": 1400, "loss": -5.795322895050049, "grad_norm": 0.1761929839849472, "lr": 0.0002388380013465276, "ex_per_s": 349.1183571805526, "data_frac": 0.05008314582033718, "mem_mb": 90.356224, "nll": -5.795322895050049, "elapsed_s": 256.7} {"event": "train", "step": 1420, "loss": -5.9884257316589355, "grad_norm": 0.7732142210006714, "lr": 0.00023695928337827914, "ex_per_s": 351.8187898140773, "data_frac": 0.049805069366348105, "mem_mb": 90.356224, "nll": -5.9884257316589355, "elapsed_s": 258.4} {"event": "train", "step": 1440, "loss": -5.743048667907715, "grad_norm": 0.2655065953731537, "lr": 0.00023506089246046403, "ex_per_s": 354.4853309845155, "data_frac": 0.04953174446775782, "mem_mb": 90.356224, "nll": -5.743048667907715, "elapsed_s": 260.0} {"event": "train", "step": 1460, "loss": -5.8027215003967285, "grad_norm": 0.3948141932487488, "lr": 0.00023314334759420388, "ex_per_s": 357.11547060313956, "data_frac": 0.04926151240952445, "mem_mb": 90.356224, "nll": -5.8027215003967285, "elapsed_s": 261.7} {"event": "train", "step": 1480, "loss": -6.045463562011719, "grad_norm": 0.150774285197258, "lr": 0.0002312071730171186, "ex_per_s": 359.7168134844276, "data_frac": 0.0489957945695629, "mem_mb": 90.356224, "nll": -6.045463562011719, "elapsed_s": 263.4} {"event": "train", "step": 1500, "loss": -5.957555294036865, "grad_norm": 0.32303503155708313, "lr": 0.0002292528980600049, "ex_per_s": 362.2887154025482, "data_frac": 0.048731238466230904, "mem_mb": 90.356224, "nll": -5.957555294036865, "elapsed_s": 265.0} {"event": "val", "step": 1500, "val_nll": -6.446729716320671, "best": -6.446729716320671, "bad": 0, "mem_mb": 90.356224, "elapsed_s": 328.2} {"event": "train", "step": 1520, "loss": -5.950323104858398, "grad_norm": 0.6510404944419861, "lr": 0.00022728105700212285, "ex_per_s": 294.8800994221057, "data_frac": 0.03917900893068083, "mem_mb": 90.356224, "nll": -5.950323104858398, "elapsed_s": 329.9} {"event": "train", "step": 1540, "loss": -5.879813194274902, "grad_norm": 0.11242252588272095, "lr": 0.00022529218892512886, "ex_per_s": 297.2505254770294, "data_frac": 0.03901535850435621, "mem_mb": 90.356224, "nll": -5.879813194274902, "elapsed_s": 331.6} {"event": "train", "step": 1560, "loss": -5.982844352722168, "grad_norm": 0.4550023078918457, "lr": 0.00022328683756569652, "ex_per_s": 299.59634252961973, "data_frac": 0.038854496494336414, "mem_mb": 90.356224, "nll": -5.982844352722168, "elapsed_s": 333.3} {"event": "train", "step": 1580, "loss": -5.981070518493652, "grad_norm": 0.11043626070022583, "lr": 0.0002212655511668642, "ex_per_s": 301.91878249605327, "data_frac": 0.03869500094423648, "mem_mb": 90.356224, "nll": -5.981070518493652, "elapsed_s": 335.0} {"event": "train", "step": 1600, "loss": -5.957505702972412, "grad_norm": 0.20179788768291473, "lr": 0.00021922888232815089, "ex_per_s": 304.2173973255792, "data_frac": 0.0385368300210768, "mem_mb": 90.356224, "nll": -5.957505702972412, "elapsed_s": 336.6} {"event": "train", "step": 1620, "loss": -5.882014274597168, "grad_norm": 0.144520565867424, "lr": 0.00021717738785448102, "ex_per_s": 306.4951210838355, "data_frac": 0.03838035072921668, "mem_mb": 90.356224, "nll": -5.882014274597168, "elapsed_s": 338.3} {"event": "train", "step": 1640, "loss": -6.077513217926025, "grad_norm": 0.14106272161006927, "lr": 0.00021511162860395977, "ex_per_s": 308.7473396925985, "data_frac": 0.03822464065239006, "mem_mb": 90.356224, "nll": -6.077513217926025, "elapsed_s": 340.0} {"event": "train", "step": 1660, "loss": -6.04099178314209, "grad_norm": 0.7721971869468689, "lr": 0.0002130321693345397, "ex_per_s": 310.9769659372601, "data_frac": 0.03807361146349552, "mem_mb": 90.356224, "nll": -6.04099178314209, "elapsed_s": 341.7} {"event": "train", "step": 1680, "loss": -5.86591911315918, "grad_norm": 0.17078141868114471, "lr": 0.0002109395785496218, "ex_per_s": 313.19260567139077, "data_frac": 0.03792285437664541, "mem_mb": 90.356224, "nll": -5.86591911315918, "elapsed_s": 343.3} {"event": "train", "step": 1700, "loss": -5.928846836090088, "grad_norm": 0.5537129044532776, "lr": 0.00020883442834263225, "ex_per_s": 315.38573223397316, "data_frac": 0.03777339407954396, "mem_mb": 90.356224, "nll": -5.928846836090088, "elapsed_s": 345.0} {"event": "train", "step": 1720, "loss": -5.888427734375, "grad_norm": 0.8697070479393005, "lr": 0.00020671729424061788, "ex_per_s": 317.55888599863886, "data_frac": 0.03762577270917897, "mem_mb": 90.356224, "nll": -5.888427734375, "elapsed_s": 346.7} {"event": "train", "step": 1740, "loss": -6.27331018447876, "grad_norm": 0.3618434965610504, "lr": 0.00020458875504690275, "ex_per_s": 319.7103449944509, "data_frac": 0.037478247381288764, "mem_mb": 90.356224, "nll": -6.27331018447876, "elapsed_s": 348.4} {"event": "train", "step": 1760, "loss": -5.930471897125244, "grad_norm": 0.3719712793827057, "lr": 0.000202449392682849, "ex_per_s": 321.8423564378505, "data_frac": 0.03733313042452792, "mem_mb": 90.356224, "nll": -5.930471897125244, "elapsed_s": 350.0} {"event": "train", "step": 1780, "loss": -5.870894908905029, "grad_norm": 0.563571572303772, "lr": 0.00020029979202876563, "ex_per_s": 323.9571448301234, "data_frac": 0.03718893181219956, "mem_mb": 90.356224, "nll": -5.870894908905029, "elapsed_s": 351.7} {"event": "train", "step": 1800, "loss": -6.0485053062438965, "grad_norm": 1.0295485258102417, "lr": 0.00019814054076400787, "ex_per_s": 326.04955960947325, "data_frac": 0.03704697995218168, "mem_mb": 90.356224, "nll": -6.0485053062438965, "elapsed_s": 353.4} {"event": "train", "step": 1820, "loss": -5.933297157287598, "grad_norm": 0.9859023094177246, "lr": 0.00019597222920631147, "ex_per_s": 328.122256971659, "data_frac": 0.03690725220196951, "mem_mb": 90.356224, "nll": -5.933297157287598, "elapsed_s": 355.0} {"event": "train", "step": 1840, "loss": -6.072739601135254, "grad_norm": 0.415211021900177, "lr": 0.0001937954501504058, "ex_per_s": 330.17314404159805, "data_frac": 0.036768678147112106, "mem_mb": 90.356224, "nll": -6.072739601135254, "elapsed_s": 356.7} {"event": "train", "step": 1860, "loss": -6.051424503326416, "grad_norm": 0.1582544445991516, "lr": 0.0001916107987059498, "ex_per_s": 332.20567595210116, "data_frac": 0.03662985235804546, "mem_mb": 90.356224, "nll": -6.051424503326416, "elapsed_s": 358.4} {"event": "train", "step": 1880, "loss": -5.94803524017334, "grad_norm": 0.3805929720401764, "lr": 0.00018941887213483484, "ex_per_s": 334.2193071125119, "data_frac": 0.03649166981691023, "mem_mb": 90.356224, "nll": -5.94803524017334, "elapsed_s": 360.0} {"event": "train", "step": 1900, "loss": -5.980976581573486, "grad_norm": 0.2401910126209259, "lr": 0.00018722026968789906, "ex_per_s": 336.2109329372424, "data_frac": 0.03635636879617305, "mem_mb": 90.356224, "nll": -5.980976581573486, "elapsed_s": 361.7} {"event": "train", "step": 1920, "loss": -5.963915824890137, "grad_norm": 0.43008455634117126, "lr": 0.00018501559244109864, "ex_per_s": 338.18743202140536, "data_frac": 0.03622136171143382, "mem_mb": 90.356224, "nll": -5.963915824890137, "elapsed_s": 363.4} {"event": "train", "step": 1940, "loss": -6.048807144165039, "grad_norm": 0.2015209197998047, "lr": 0.00018280544313117932, "ex_per_s": 340.1457845585801, "data_frac": 0.036087597303018026, "mem_mb": 90.356224, "nll": -6.048807144165039, "elapsed_s": 365.1} {"event": "train", "step": 1960, "loss": -5.939038276672363, "grad_norm": 0.0839996337890625, "lr": 0.00018059042599089415, "ex_per_s": 342.0868356708322, "data_frac": 0.03595582693421947, "mem_mb": 90.356224, "nll": -5.939038276672363, "elapsed_s": 366.7} {"event": "train", "step": 1980, "loss": -5.965105056762695, "grad_norm": 0.6296360492706299, "lr": 0.00017837114658381247, "ex_per_s": 344.0107320754799, "data_frac": 0.035825252131863855, "mem_mb": 90.356224, "nll": -5.965105056762695, "elapsed_s": 368.4} {"event": "train", "step": 2000, "loss": -5.915505409240723, "grad_norm": 0.4489603638648987, "lr": 0.00017614821163876486, "ex_per_s": 345.91590178169474, "data_frac": 0.03569488767513072, "mem_mb": 90.356224, "nll": -5.915505409240723, "elapsed_s": 370.1} {"event": "val", "step": 2000, "val_nll": -6.4484527915110705, "best": -6.4484527915110705, "bad": 0, "mem_mb": 90.356224, "elapsed_s": 433.6} {"event": "train", "step": 2020, "loss": -5.927671432495117, "grad_norm": 0.42518937587738037, "lr": 0.00017392222888396941, "ex_per_s": 297.03772969097093, "data_frac": 0.03037498256215464, "mem_mb": 90.356224, "nll": -5.927671432495117, "elapsed_s": 435.3} {"event": "train", "step": 2040, "loss": -5.872918128967285, "grad_norm": 0.3643059730529785, "lr": 0.000171693806880885, "ex_per_s": 298.8278093998596, "data_frac": 0.030284642555440073, "mem_mb": 90.356224, "nll": -5.872918128967285, "elapsed_s": 437.0} {"event": "train", "step": 2060, "loss": -6.143669605255127, "grad_norm": 0.35460859537124634, "lr": 0.0001694635548578366, "ex_per_s": 300.6048427416768, "data_frac": 0.030195234257890176, "mem_mb": 90.356224, "nll": -6.143669605255127, "elapsed_s": 438.6} {"event": "train", "step": 2080, "loss": -5.960995674133301, "grad_norm": 0.5695834755897522, "lr": 0.0001672320825434584, "ex_per_s": 302.36824600169876, "data_frac": 0.030106501811813964, "mem_mb": 90.356224, "nll": -5.960995674133301, "elapsed_s": 440.3} {"event": "train", "step": 2100, "loss": -5.899832725524902, "grad_norm": 0.9322943091392517, "lr": 0.000165, "ex_per_s": 304.11665472823165, "data_frac": 0.030020826110253498, "mem_mb": 90.356224, "nll": -5.899832725524902, "elapsed_s": 442.0} {"event": "train", "step": 2120, "loss": -6.150994300842285, "grad_norm": 0.5278491973876953, "lr": 0.0001627679174565416, "ex_per_s": 305.8535000510983, "data_frac": 0.029934010754721574, "mem_mb": 90.356224, "nll": -6.150994300842285, "elapsed_s": 443.7} {"event": "train", "step": 2140, "loss": -5.978010654449463, "grad_norm": 0.11705812066793442, "lr": 0.00016053644514216336, "ex_per_s": 307.5788802617711, "data_frac": 0.029848314424623115, "mem_mb": 90.356224, "nll": -5.978010654449463, "elapsed_s": 445.3} {"event": "train", "step": 2160, "loss": -5.927591800689697, "grad_norm": 0.466929167509079, "lr": 0.00015830619311911498, "ex_per_s": 309.290875013888, "data_frac": 0.02976303215593198, "mem_mb": 90.356224, "nll": -5.927591800689697, "elapsed_s": 447.0} {"event": "train", "step": 2180, "loss": -6.083564281463623, "grad_norm": 0.3217177391052246, "lr": 0.00015607777111603058, "ex_per_s": 310.9900934286229, "data_frac": 0.029677600030407862, "mem_mb": 90.356224, "nll": -6.083564281463623, "elapsed_s": 448.7} {"event": "train", "step": 2200, "loss": -5.964006423950195, "grad_norm": 0.5351234078407288, "lr": 0.00015385178836123514, "ex_per_s": 312.6778569866338, "data_frac": 0.029593223057275426, "mem_mb": 90.356224, "nll": -5.964006423950195, "elapsed_s": 450.3} {"event": "train", "step": 2220, "loss": -5.82370662689209, "grad_norm": 0.836722731590271, "lr": 0.00015162885341618752, "ex_per_s": 314.352010524483, "data_frac": 0.02950960027727738, "mem_mb": 90.356224, "nll": -5.82370662689209, "elapsed_s": 452.0} {"event": "train", "step": 2240, "loss": -5.961542129516602, "grad_norm": 0.713954746723175, "lr": 0.00014940957400910585, "ex_per_s": 316.01504667936473, "data_frac": 0.029426685409593655, "mem_mb": 90.356224, "nll": -5.961542129516602, "elapsed_s": 453.7} {"event": "train", "step": 2260, "loss": -5.738166809082031, "grad_norm": 0.8245008587837219, "lr": 0.00014719455686882068, "ex_per_s": 317.66604694177744, "data_frac": 0.029343596762674614, "mem_mb": 90.356224, "nll": -5.738166809082031, "elapsed_s": 455.4} {"event": "train", "step": 2280, "loss": -6.011317253112793, "grad_norm": 0.4356158673763275, "lr": 0.0001449844075589013, "ex_per_s": 319.30535604084884, "data_frac": 0.02926202795073939, "mem_mb": 90.356224, "nll": -6.011317253112793, "elapsed_s": 457.0} {"event": "train", "step": 2300, "loss": -6.00209379196167, "grad_norm": 0.1455961912870407, "lr": 0.0001427797303121009, "ex_per_s": 320.9321207728311, "data_frac": 0.029180956418527446, "mem_mb": 90.356224, "nll": -6.00209379196167, "elapsed_s": 458.7} {"event": "train", "step": 2320, "loss": -6.069661617279053, "grad_norm": 0.3406854569911957, "lr": 0.00014058112786516518, "ex_per_s": 322.54779382209324, "data_frac": 0.02909976285084284, "mem_mb": 90.356224, "nll": -6.069661617279053, "elapsed_s": 460.4} {"event": "train", "step": 2340, "loss": -5.851537227630615, "grad_norm": 0.3108377754688263, "lr": 0.00013838920129405016, "ex_per_s": 324.15001712232623, "data_frac": 0.029021117634787864, "mem_mb": 90.356224, "nll": -5.851537227630615, "elapsed_s": 462.1} {"event": "train", "step": 2360, "loss": -5.930434703826904, "grad_norm": 0.11273012310266495, "lr": 0.0001362045498495942, "ex_per_s": 325.7464124349718, "data_frac": 0.028942460750130836, "mem_mb": 90.356224, "nll": -5.930434703826904, "elapsed_s": 463.7} {"event": "train", "step": 2380, "loss": -5.931563377380371, "grad_norm": 0.09806700050830841, "lr": 0.00013402777079368853, "ex_per_s": 327.3312025957175, "data_frac": 0.028863914989941588, "mem_mb": 90.356224, "nll": -5.931563377380371, "elapsed_s": 465.4} {"event": "train", "step": 2400, "loss": -5.957581043243408, "grad_norm": 0.8913285732269287, "lr": 0.00013185945923599213, "ex_per_s": 328.90431472733655, "data_frac": 0.028786449634937067, "mem_mb": 90.356224, "nll": -5.957581043243408, "elapsed_s": 467.0} {"event": "train", "step": 2420, "loss": -5.996944904327393, "grad_norm": 0.47923529148101807, "lr": 0.00012970020797123432, "ex_per_s": 330.4665515645331, "data_frac": 0.028709847772016874, "mem_mb": 90.356224, "nll": -5.996944904327393, "elapsed_s": 468.7} {"event": "train", "step": 2440, "loss": -6.099518775939941, "grad_norm": 0.37945640087127686, "lr": 0.000127550607317151, "ex_per_s": 332.01645096526306, "data_frac": 0.028633477447899243, "mem_mb": 90.356224, "nll": -6.099518775939941, "elapsed_s": 470.4} {"event": "train", "step": 2460, "loss": -6.133598804473877, "grad_norm": 0.18308228254318237, "lr": 0.00012541124495309727, "ex_per_s": 333.5573397742982, "data_frac": 0.02855758950869887, "mem_mb": 90.356224, "nll": -6.133598804473877, "elapsed_s": 472.0} {"event": "train", "step": 2480, "loss": -5.9455389976501465, "grad_norm": 0.16401207447052002, "lr": 0.00012328270575938212, "ex_per_s": 335.083294703183, "data_frac": 0.0284819309983917, "mem_mb": 90.356224, "nll": -5.9455389976501465, "elapsed_s": 473.7} {"event": "train", "step": 2500, "loss": -5.7559943199157715, "grad_norm": 0.7244518995285034, "lr": 0.00012116557165736775, "ex_per_s": 336.6012107757025, "data_frac": 0.028407911216349103, "mem_mb": 90.356224, "nll": -5.7559943199157715, "elapsed_s": 475.4} {"event": "val", "step": 2500, "val_nll": -6.450148327536085, "best": -6.450148327536085, "bad": 0, "mem_mb": 90.356224, "elapsed_s": 538.7} {"event": "train", "step": 2520, "loss": -5.959945201873779, "grad_norm": 0.8075306415557861, "lr": 0.00011906042145037817, "ex_per_s": 298.4787825435957, "data_frac": 0.025013270821702304, "mem_mb": 90.356224, "nll": -5.959945201873779, "elapsed_s": 540.4} {"event": "train", "step": 2540, "loss": -6.031347751617432, "grad_norm": 0.5700039863586426, "lr": 0.00011696783066546026, "ex_per_s": 299.9166920367298, "data_frac": 0.024957679503404762, "mem_mb": 90.356224, "nll": -6.031347751617432, "elapsed_s": 542.1} {"event": "train", "step": 2560, "loss": -5.948004245758057, "grad_norm": 0.30018162727355957, "lr": 0.00011488837139604017, "ex_per_s": 301.3469863641897, "data_frac": 0.024903708489937938, "mem_mb": 90.356224, "nll": -5.948004245758057, "elapsed_s": 543.7} {"event": "train", "step": 2580, "loss": -5.961906433105469, "grad_norm": 0.23195162415504456, "lr": 0.00011282261214551898, "ex_per_s": 302.7701565744424, "data_frac": 0.02484879113013905, "mem_mb": 90.356224, "nll": -5.961906433105469, "elapsed_s": 545.4} {"event": "train", "step": 2600, "loss": -5.8500542640686035, "grad_norm": 0.32378673553466797, "lr": 0.00011077111767184915, "ex_per_s": 304.18328123660166, "data_frac": 0.024794297217787198, "mem_mb": 90.356224, "nll": -5.8500542640686035, "elapsed_s": 547.1} {"event": "train", "step": 2620, "loss": -5.890227794647217, "grad_norm": 0.9162752032279968, "lr": 0.00010873444883313583, "ex_per_s": 305.5888085058241, "data_frac": 0.024740083470679533, "mem_mb": 90.356224, "nll": -5.890227794647217, "elapsed_s": 548.8} {"event": "train", "step": 2640, "loss": -5.988448143005371, "grad_norm": 0.31018921732902527, "lr": 0.00010671316243430345, "ex_per_s": 306.98536694930925, "data_frac": 0.02468601377291158, "mem_mb": 90.356224, "nll": -5.988448143005371, "elapsed_s": 550.4} {"event": "train", "step": 2660, "loss": -5.821882724761963, "grad_norm": 0.09596081078052521, "lr": 0.00010470781107487112, "ex_per_s": 308.3748455482297, "data_frac": 0.02463287925471993, "mem_mb": 90.356224, "nll": -5.821882724761963, "elapsed_s": 552.1} {"event": "train", "step": 2680, "loss": -5.960226535797119, "grad_norm": 0.3232385814189911, "lr": 0.00010271894299787715, "ex_per_s": 309.7541476264325, "data_frac": 0.024579174088291474, "mem_mb": 90.356224, "nll": -5.960226535797119, "elapsed_s": 553.8} {"event": "train", "step": 2700, "loss": -6.134474277496338, "grad_norm": 0.6933901309967041, "lr": 0.00010074710193999506, "ex_per_s": 311.12642704552695, "data_frac": 0.024526246536447114, "mem_mb": 90.356224, "nll": -6.134474277496338, "elapsed_s": 555.4} {"event": "train", "step": 2720, "loss": -5.8814849853515625, "grad_norm": 0.5849605202674866, "lr": 9.879282698288144e-05, "ex_per_s": 312.4907006966191, "data_frac": 0.024473862282938155, "mem_mb": 90.356224, "nll": -5.8814849853515625, "elapsed_s": 557.1} {"event": "train", "step": 2740, "loss": -6.12577486038208, "grad_norm": 0.08987097442150116, "lr": 9.685665240579613e-05, "ex_per_s": 313.84631785537175, "data_frac": 0.024421476420402023, "mem_mb": 90.356224, "nll": -6.12577486038208, "elapsed_s": 558.8} {"event": "train", "step": 2760, "loss": -5.959655284881592, "grad_norm": 0.24881775677204132, "lr": 9.4939107539536e-05, "ex_per_s": 315.19326889700847, "data_frac": 0.02436919528256802, "mem_mb": 90.356224, "nll": -5.959655284881592, "elapsed_s": 560.5} {"event": "train", "step": 2780, "loss": -6.014822006225586, "grad_norm": 0.729365885257721, "lr": 9.304071662172087e-05, "ex_per_s": 316.53267484870867, "data_frac": 0.024317911659508664, "mem_mb": 90.356224, "nll": -6.014822006225586, "elapsed_s": 562.1} {"event": "train", "step": 2800, "loss": -5.980257511138916, "grad_norm": 0.8955816030502319, "lr": 9.116199865347241e-05, "ex_per_s": 317.86417009141337, "data_frac": 0.02426765314818355, "mem_mb": 90.356224, "nll": -5.980257511138916, "elapsed_s": 563.8} {"event": "train", "step": 2820, "loss": -6.008753299713135, "grad_norm": 0.23394238948822021, "lr": 8.930346725752419e-05, "ex_per_s": 319.1874493658209, "data_frac": 0.02421683422033321, "mem_mb": 90.356224, "nll": -6.008753299713135, "elapsed_s": 565.5} {"event": "train", "step": 2840, "loss": -5.88114070892334, "grad_norm": 0.11364301294088364, "lr": 8.746563053780262e-05, "ex_per_s": 320.505795888718, "data_frac": 0.024166946809857485, "mem_mb": 90.356224, "nll": -5.88114070892334, "elapsed_s": 567.1} {"event": "train", "step": 2860, "loss": -6.047072410583496, "grad_norm": 0.24148449301719666, "lr": 8.564899094051614e-05, "ex_per_s": 321.81648815237725, "data_frac": 0.024117526302667384, "mem_mb": 90.356224, "nll": -6.047072410583496, "elapsed_s": 568.8} {"event": "train", "step": 2880, "loss": -5.772184371948242, "grad_norm": 0.4068450629711151, "lr": 8.385404511679161e-05, "ex_per_s": 323.1189589619526, "data_frac": 0.02406752860334145, "mem_mb": 90.356224, "nll": -5.772184371948242, "elapsed_s": 570.5} {"event": "train", "step": 2900, "loss": -5.795387268066406, "grad_norm": 0.6149458289146423, "lr": 8.208128378689484e-05, "ex_per_s": 324.41440295376646, "data_frac": 0.024017737334638165, "mem_mb": 90.356224, "nll": -5.795387268066406, "elapsed_s": 572.2} {"event": "train", "step": 2920, "loss": -5.9704203605651855, "grad_norm": 0.7079156041145325, "lr": 8.03311916060725e-05, "ex_per_s": 325.7018906469775, "data_frac": 0.02396811275894684, "mem_mb": 90.356224, "nll": -5.9704203605651855, "elapsed_s": 573.8} {"event": "train", "step": 2940, "loss": -5.905200004577637, "grad_norm": 0.4983508586883545, "lr": 7.860424703205225e-05, "ex_per_s": 326.9829943273125, "data_frac": 0.023918793355624055, "mem_mb": 90.356224, "nll": -5.905200004577637, "elapsed_s": 575.5} {"event": "train", "step": 2960, "loss": -5.938331127166748, "grad_norm": 0.36914879083633423, "lr": 7.690092219423722e-05, "ex_per_s": 328.2574144728704, "data_frac": 0.023870452823048872, "mem_mb": 90.356224, "nll": -5.938331127166748, "elapsed_s": 577.2} {"event": "train", "step": 2980, "loss": -5.895800590515137, "grad_norm": 0.4409834146499634, "lr": 7.522168276463048e-05, "ex_per_s": 329.5250983082414, "data_frac": 0.023822824998791937, "mem_mb": 90.356224, "nll": -5.895800590515137, "elapsed_s": 578.8} {"event": "train", "step": 3000, "loss": -5.995471954345703, "grad_norm": 0.11588819324970245, "lr": 7.356698783052496e-05, "ex_per_s": 330.78461428863557, "data_frac": 0.023774044886151112, "mem_mb": 90.356224, "nll": -5.995471954345703, "elapsed_s": 580.5} {"event": "val", "step": 3000, "val_nll": -6.448145964421379, "best": -6.450148327536085, "bad": 1, "mem_mb": 90.356224, "elapsed_s": 643.6} {"event": "train", "step": 3020, "loss": -5.962202072143555, "grad_norm": 0.8950678110122681, "lr": 7.193728976899362e-05, "ex_per_s": 299.516131839418, "data_frac": 0.02140308582502994, "mem_mb": 90.356224, "nll": -5.962202072143555, "elapsed_s": 645.4} {"event": "train", "step": 3040, "loss": -5.860806941986084, "grad_norm": 0.9212675094604492, "lr": 7.033303412321404e-05, "ex_per_s": 300.71914563506084, "data_frac": 0.021366594253451093, "mem_mb": 90.356224, "nll": -5.860806941986084, "elapsed_s": 647.0} {"event": "train", "step": 3060, "loss": -6.0727081298828125, "grad_norm": 0.25212937593460083, "lr": 6.875465948066139e-05, "ex_per_s": 301.9153788364488, "data_frac": 0.021329936390286363, "mem_mb": 90.356224, "nll": -6.0727081298828125, "elapsed_s": 648.7} {"event": "train", "step": 3080, "loss": -5.968886375427246, "grad_norm": 0.30598780512809753, "lr": 6.720259735320299e-05, "ex_per_s": 303.1052992881994, "data_frac": 0.02129314704963422, "mem_mb": 90.356224, "nll": -5.968886375427246, "elapsed_s": 650.4} {"event": "train", "step": 3100, "loss": -6.014895439147949, "grad_norm": 0.32084921002388, "lr": 6.567727205912722e-05, "ex_per_s": 304.28908058767763, "data_frac": 0.021256752830490264, "mem_mb": 90.356224, "nll": -6.014895439147949, "elapsed_s": 652.1} {"event": "train", "step": 3120, "loss": -6.091645240783691, "grad_norm": 0.6220258474349976, "lr": 6.417910060713926e-05, "ex_per_s": 305.466870788112, "data_frac": 0.021219690889201184, "mem_mb": 90.356224, "nll": -6.091645240783691, "elapsed_s": 653.7} {"event": "train", "step": 3140, "loss": -6.007205009460449, "grad_norm": 0.5265828371047974, "lr": 6.270849258235485e-05, "ex_per_s": 306.63764443441283, "data_frac": 0.02118397205586792, "mem_mb": 90.356224, "nll": -6.007205009460449, "elapsed_s": 655.4} {"event": "train", "step": 3160, "loss": -6.022627830505371, "grad_norm": 0.10873537510633469, "lr": 6.126585003432406e-05, "ex_per_s": 307.8042473070918, "data_frac": 0.021147995875762347, "mem_mb": 90.356224, "nll": -6.022627830505371, "elapsed_s": 657.1} {"event": "train", "step": 3180, "loss": -6.0957489013671875, "grad_norm": 0.13887763023376465, "lr": 5.985156736711483e-05, "ex_per_s": 308.9663346469637, "data_frac": 0.021112970056940534, "mem_mb": 90.356224, "nll": -6.0957489013671875, "elapsed_s": 658.8} {"event": "train", "step": 3200, "loss": -6.000300407409668, "grad_norm": 0.37632736563682556, "lr": 5.846603123148688e-05, "ex_per_s": 310.1222210149547, "data_frac": 0.02107699266391383, "mem_mb": 90.356224, "nll": -6.000300407409668, "elapsed_s": 660.4} {"event": "train", "step": 3220, "loss": -5.915248394012451, "grad_norm": 0.19907976686954498, "lr": 5.710962041918516e-05, "ex_per_s": 311.2722748148563, "data_frac": 0.021041458189997336, "mem_mb": 90.356224, "nll": -5.915248394012451, "elapsed_s": 662.1} {"event": "train", "step": 3240, "loss": -5.803150653839111, "grad_norm": 0.43263423442840576, "lr": 5.578270575938211e-05, "ex_per_s": 312.41556723133925, "data_frac": 0.021006776331107417, "mem_mb": 90.356224, "nll": -5.803150653839111, "elapsed_s": 663.8} {"event": "train", "step": 3260, "loss": -5.9615559577941895, "grad_norm": 0.3909398913383484, "lr": 5.448565001729653e-05, "ex_per_s": 313.55462397178286, "data_frac": 0.02097197602988103, "mem_mb": 90.356224, "nll": -5.9615559577941895, "elapsed_s": 665.4} {"event": "train", "step": 3280, "loss": -6.015501976013184, "grad_norm": 0.31555891036987305, "lr": 5.3218807795017296e-05, "ex_per_s": 314.6884805468031, "data_frac": 0.02093699028682622, "mem_mb": 90.356224, "nll": -6.015501976013184, "elapsed_s": 667.1} {"event": "train", "step": 3300, "loss": -5.801801681518555, "grad_norm": 0.247976154088974, "lr": 5.198252543455864e-05, "ex_per_s": 315.8158679667448, "data_frac": 0.02090229241529089, "mem_mb": 90.356224, "nll": -5.801801681518555, "elapsed_s": 668.8} {"event": "train", "step": 3320, "loss": -5.98851203918457, "grad_norm": 0.2969433665275574, "lr": 5.07771409231737e-05, "ex_per_s": 316.9385553515685, "data_frac": 0.02086806196019839, "mem_mb": 90.356224, "nll": -5.98851203918457, "elapsed_s": 670.5} {"event": "train", "step": 3340, "loss": -5.932383060455322, "grad_norm": 0.3498874604701996, "lr": 4.960298380095221e-05, "ex_per_s": 318.056727643641, "data_frac": 0.020833231210656042, "mem_mb": 90.356224, "nll": -5.932383060455322, "elapsed_s": 672.1} {"event": "train", "step": 3360, "loss": -5.90700626373291, "grad_norm": 0.35921230912208557, "lr": 4.846037507072751e-05, "ex_per_s": 319.16796048866536, "data_frac": 0.02079891235308556, "mem_mb": 90.356224, "nll": -5.90700626373291, "elapsed_s": 673.8} {"event": "train", "step": 3380, "loss": -6.123401165008545, "grad_norm": 0.4600759744644165, "lr": 4.7349627110317556e-05, "ex_per_s": 320.27374166962824, "data_frac": 0.02076520670324948, "mem_mb": 90.356224, "nll": -6.123401165008545, "elapsed_s": 675.5} {"event": "train", "step": 3400, "loss": -5.882874488830566, "grad_norm": 0.33039361238479614, "lr": 4.6271043587123986e-05, "ex_per_s": 321.3733259733388, "data_frac": 0.020731761107739256, "mem_mb": 90.356224, "nll": -5.882874488830566, "elapsed_s": 677.1} {"event": "train", "step": 3420, "loss": -5.840181827545166, "grad_norm": 0.20531807839870453, "lr": 4.52249193751124e-05, "ex_per_s": 322.46923840409227, "data_frac": 0.020698018344055286, "mem_mb": 90.356224, "nll": -5.840181827545166, "elapsed_s": 678.8} {"event": "train", "step": 3440, "loss": -5.915998458862305, "grad_norm": 0.12337134033441544, "lr": 4.4211540474196926e-05, "ex_per_s": 323.55986721239503, "data_frac": 0.020664270205040764, "mem_mb": 90.356224, "nll": -5.915998458862305, "elapsed_s": 680.5} {"event": "train", "step": 3460, "loss": -6.056445121765137, "grad_norm": 0.2048177719116211, "lr": 4.3231183932050546e-05, "ex_per_s": 324.64530563241453, "data_frac": 0.0206311136194352, "mem_mb": 90.356224, "nll": -6.056445121765137, "elapsed_s": 682.1} {"event": "train", "step": 3480, "loss": -5.916863441467285, "grad_norm": 0.08080132305622101, "lr": 4.228411776836306e-05, "ex_per_s": 325.7249052717397, "data_frac": 0.020598650652253483, "mem_mb": 90.356224, "nll": -5.916863441467285, "elapsed_s": 683.8} {"event": "train", "step": 3500, "loss": -6.037835121154785, "grad_norm": 0.14311856031417847, "lr": 4.137060090156723e-05, "ex_per_s": 326.8006124007797, "data_frac": 0.02056703932576697, "mem_mb": 90.356224, "nll": -6.037835121154785, "elapsed_s": 685.5} {"event": "val", "step": 3500, "val_nll": -6.450088758579578, "best": -6.450148327536085, "bad": 2, "mem_mb": 90.356224, "elapsed_s": 748.4} {"event": "train", "step": 3520, "loss": -5.937657833099365, "grad_norm": 0.4189371168613434, "lr": 4.049088307805306e-05, "ex_per_s": 300.3265071794884, "data_frac": 0.018809195384566928, "mem_mb": 90.356224, "nll": -5.937657833099365, "elapsed_s": 750.2} {"event": "train", "step": 3540, "loss": -6.135462760925293, "grad_norm": 0.6327059864997864, "lr": 3.964520480388956e-05, "ex_per_s": 301.3592463384943, "data_frac": 0.018783048493103388, "mem_mb": 90.356224, "nll": -6.135462760925293, "elapsed_s": 751.8} {"event": "train", "step": 3560, "loss": -6.139243125915527, "grad_norm": 0.08099903911352158, "lr": 3.88337972790728e-05, "ex_per_s": 302.3880753888549, "data_frac": 0.018756604275705253, "mem_mb": 90.356224, "nll": -6.139243125915527, "elapsed_s": 753.5} {"event": "train", "step": 3580, "loss": -5.902944087982178, "grad_norm": 0.2149488776922226, "lr": 3.805688233431824e-05, "ex_per_s": 303.41182377243166, "data_frac": 0.018730430841097625, "mem_mb": 90.356224, "nll": -5.902944087982178, "elapsed_s": 755.2} {"event": "train", "step": 3600, "loss": -6.100104808807373, "grad_norm": 0.36847686767578125, "lr": 3.7314672370414325e-05, "ex_per_s": 304.4311222155283, "data_frac": 0.018704628417073675, "mem_mb": 90.356224, "nll": -6.100104808807373, "elapsed_s": 756.9} {"event": "train", "step": 3620, "loss": -5.910183906555176, "grad_norm": 0.9214469790458679, "lr": 3.660737030015427e-05, "ex_per_s": 305.44546408708277, "data_frac": 0.018678688931788142, "mem_mb": 90.356224, "nll": -5.910183906555176, "elapsed_s": 758.5} {"event": "train", "step": 3640, "loss": -6.053597927093506, "grad_norm": 0.5015806555747986, "lr": 3.5935169492861716e-05, "ex_per_s": 306.4552745679467, "data_frac": 0.018653223748066456, "mem_mb": 90.356224, "nll": -6.053597927093506, "elapsed_s": 760.2} {"event": "train", "step": 3660, "loss": -5.83387565612793, "grad_norm": 0.4858696758747101, "lr": 3.52982537215255e-05, "ex_per_s": 307.46145151472746, "data_frac": 0.01862735729595289, "mem_mb": 90.356224, "nll": -5.83387565612793, "elapsed_s": 761.9} {"event": "train", "step": 3680, "loss": -5.99337911605835, "grad_norm": 0.14528261125087738, "lr": 3.469679711255794e-05, "ex_per_s": 308.46244455520593, "data_frac": 0.018602255625592673, "mem_mb": 90.356224, "nll": -5.99337911605835, "elapsed_s": 763.6} {"event": "train", "step": 3700, "loss": -6.066461086273193, "grad_norm": 0.8579258322715759, "lr": 3.41309640981904e-05, "ex_per_s": 309.46065658117453, "data_frac": 0.018577475461760015, "mem_mb": 90.356224, "nll": -6.066461086273193, "elapsed_s": 765.2} {"event": "train", "step": 3720, "loss": -5.949582099914551, "grad_norm": 0.5303481817245483, "lr": 3.3600909371519225e-05, "ex_per_s": 310.4547664714114, "data_frac": 0.018552277978904987, "mem_mb": 90.356224, "nll": -5.949582099914551, "elapsed_s": 766.9} {"event": "train", "step": 3740, "loss": -6.190715312957764, "grad_norm": 0.13367168605327606, "lr": 3.31067778442141e-05, "ex_per_s": 311.4447938814697, "data_frac": 0.018526859778999843, "mem_mb": 90.356224, "nll": -6.190715312957764, "elapsed_s": 768.6} {"event": "train", "step": 3760, "loss": -6.148360252380371, "grad_norm": 0.2636023163795471, "lr": 3.2648704606900735e-05, "ex_per_s": 312.43126453960605, "data_frac": 0.018502297619045423, "mem_mb": 90.356224, "nll": -6.148360252380371, "elapsed_s": 770.3} {"event": "train", "step": 3780, "loss": -5.964633464813232, "grad_norm": 0.27789703011512756, "lr": 3.222681489222838e-05, "ex_per_s": 313.4130947894993, "data_frac": 0.018477286137936446, "mem_mb": 90.356224, "nll": -5.964633464813232, "elapsed_s": 771.9} {"event": "train", "step": 3800, "loss": -6.098912715911865, "grad_norm": 0.13780957460403442, "lr": 3.1841224040632484e-05, "ex_per_s": 314.3910464897908, "data_frac": 0.018453020804733822, "mem_mb": 90.356224, "nll": -6.098912715911865, "elapsed_s": 773.6} {"event": "train", "step": 3820, "loss": -5.972224235534668, "grad_norm": 0.1681126356124878, "lr": 3.149203746880175e-05, "ex_per_s": 315.3634550604956, "data_frac": 0.01842841762902553, "mem_mb": 90.356224, "nll": -5.972224235534668, "elapsed_s": 775.3} {"event": "train", "step": 3840, "loss": -5.8747663497924805, "grad_norm": 0.6072670221328735, "lr": 3.117935064085835e-05, "ex_per_s": 316.33199260556, "data_frac": 0.018404205964032423, "mem_mb": 90.356224, "nll": -5.8747663497924805, "elapsed_s": 776.9} {"event": "train", "step": 3860, "loss": -6.014779090881348, "grad_norm": 0.32728108763694763, "lr": 3.090324904225885e-05, "ex_per_s": 317.29710538171696, "data_frac": 0.018379918410260165, "mem_mb": 90.356224, "nll": -6.014779090881348, "elapsed_s": 778.6} {"event": "train", "step": 3880, "loss": -5.891053676605225, "grad_norm": 0.14199498295783997, "lr": 3.0663808156423456e-05, "ex_per_s": 318.25802533741273, "data_frac": 0.01835545952534968, "mem_mb": 90.356224, "nll": -5.891053676605225, "elapsed_s": 780.3} {"event": "train", "step": 3900, "loss": -6.1804351806640625, "grad_norm": 0.7613977193832397, "lr": 3.046109344409957e-05, "ex_per_s": 319.2131984288633, "data_frac": 0.01833187242058145, "mem_mb": 90.356224, "nll": -6.1804351806640625, "elapsed_s": 782.0} {"event": "train", "step": 3920, "loss": -6.038137912750244, "grad_norm": 0.21197770535945892, "lr": 3.0295160325465435e-05, "ex_per_s": 320.16438555443193, "data_frac": 0.01830780027701006, "mem_mb": 90.356224, "nll": -6.038137912750244, "elapsed_s": 783.6} {"event": "train", "step": 3940, "loss": -5.920352935791016, "grad_norm": 0.2985157072544098, "lr": 3.016605416497886e-05, "ex_per_s": 321.1119404735222, "data_frac": 0.018284787739357256, "mem_mb": 90.356224, "nll": -5.920352935791016, "elapsed_s": 785.3} {"event": "train", "step": 3960, "loss": -6.06809139251709, "grad_norm": 0.13579335808753967, "lr": 3.0073810258974985e-05, "ex_per_s": 322.05632597825235, "data_frac": 0.01826144508166231, "mem_mb": 90.356224, "nll": -6.06809139251709, "elapsed_s": 787.0} {"event": "train", "step": 3980, "loss": -5.805756092071533, "grad_norm": 0.17946605384349823, "lr": 3.0018453826016683e-05, "ex_per_s": 322.9953803756461, "data_frac": 0.01823834714884298, "mem_mb": 90.356224, "nll": -5.805756092071533, "elapsed_s": 788.7} {"event": "train", "step": 4000, "loss": -6.056830883026123, "grad_norm": 0.7695972323417664, "lr": 2.9999999999999997e-05, "ex_per_s": 323.93188067258313, "data_frac": 0.01821533593098843, "mem_mb": 90.356224, "nll": -6.056830883026123, "elapsed_s": 790.3} {"event": "val", "step": 4000, "val_nll": -6.449693275106555, "best": -6.450148327536085, "bad": 3, "mem_mb": 90.356224, "elapsed_s": 853.3} {"event": "done", "steps": 4000, "best_val": -6.450148327536085, "wall_s": 853.3193740844727, "ex_per_s": 300.004907628709, "peak_mem_mb": 90.356224, "n_params": 3421643, "device": "mps", "data_time_frac": 0.01686987449971178, "elapsed_s": 853.3}