Permalink
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
3237 lines (3236 sloc) 210 KB
Namespace(batch_size=50, data_name='SST-2', dropout=0.5, epochs=40, gpu=0, log_interval=30, lr=0.0001, model_mode='non-static', save_prefix='sa-model')
Use gpu0
1614
53
Done! Tokenizing Time=4.31s, #Sentences=118038
Done! Tokenizing Time=0.74s, #Sentences=1745
SentimentNet(
(embedding): Embedding(17814 -> 300, float32)
(encoder): ConvolutionalEncoder(
(_convs): HybridConcurrent(
(0): HybridSequential(
(0): Conv1D(300 -> 100, kernel_size=(3,), stride=(1,))
(1): Activation(relu)
(2): HybridLambda(<lambda>)
)
(1): HybridSequential(
(0): Conv1D(300 -> 100, kernel_size=(4,), stride=(1,))
(1): Activation(relu)
(2): HybridLambda(<lambda>)
)
(2): HybridSequential(
(0): Conv1D(300 -> 100, kernel_size=(5,), stride=(1,))
(1): Activation(relu)
(2): HybridLambda(<lambda>)
)
)
)
(output): HybridSequential(
(0): Dropout(p = 0.5, axes=())
(1): Dense(None -> 2, linear)
)
)
[Epoch 0 Batch 30/2125] avg loss 0.0146453, throughput 3.78731K wps
[Epoch 0 Batch 60/2125] avg loss 0.014539, throughput 6.06875K wps
[Epoch 0 Batch 90/2125] avg loss 0.0141482, throughput 6.06143K wps
[Epoch 0 Batch 120/2125] avg loss 0.0134685, throughput 6.06698K wps
[Epoch 0 Batch 150/2125] avg loss 0.0136063, throughput 6.05797K wps
[Epoch 0 Batch 180/2125] avg loss 0.0133574, throughput 6.06772K wps
[Epoch 0 Batch 210/2125] avg loss 0.0135234, throughput 6.06618K wps
[Epoch 0 Batch 240/2125] avg loss 0.0132481, throughput 6.06549K wps
[Epoch 0 Batch 270/2125] avg loss 0.0129035, throughput 6.06342K wps
[Epoch 0 Batch 300/2125] avg loss 0.0129844, throughput 6.05879K wps
[Epoch 0 Batch 330/2125] avg loss 0.0127407, throughput 6.06172K wps
[Epoch 0 Batch 360/2125] avg loss 0.012595, throughput 6.06912K wps
[Epoch 0 Batch 390/2125] avg loss 0.0123208, throughput 6.05861K wps
[Epoch 0 Batch 420/2125] avg loss 0.0120249, throughput 6.05417K wps
[Epoch 0 Batch 450/2125] avg loss 0.0124691, throughput 6.04935K wps
[Epoch 0 Batch 480/2125] avg loss 0.0116782, throughput 6.05774K wps
[Epoch 0 Batch 510/2125] avg loss 0.0116368, throughput 6.06589K wps
[Epoch 0 Batch 540/2125] avg loss 0.0111518, throughput 6.06781K wps
[Epoch 0 Batch 570/2125] avg loss 0.0112152, throughput 6.05775K wps
[Epoch 0 Batch 600/2125] avg loss 0.0112537, throughput 6.06604K wps
[Epoch 0 Batch 630/2125] avg loss 0.0108123, throughput 6.05512K wps
[Epoch 0 Batch 660/2125] avg loss 0.010679, throughput 6.04928K wps
[Epoch 0 Batch 690/2125] avg loss 0.0102292, throughput 6.05303K wps
[Epoch 0 Batch 720/2125] avg loss 0.0103963, throughput 6.05583K wps
[Epoch 0 Batch 750/2125] avg loss 0.00990726, throughput 6.05314K wps
[Epoch 0 Batch 780/2125] avg loss 0.00980827, throughput 6.06013K wps
[Epoch 0 Batch 810/2125] avg loss 0.00958513, throughput 6.05506K wps
[Epoch 0 Batch 840/2125] avg loss 0.00927252, throughput 6.06102K wps
[Epoch 0 Batch 870/2125] avg loss 0.00887413, throughput 6.05244K wps
[Epoch 0 Batch 900/2125] avg loss 0.00894022, throughput 6.05989K wps
[Epoch 0 Batch 930/2125] avg loss 0.0086854, throughput 6.05053K wps
[Epoch 0 Batch 960/2125] avg loss 0.0087187, throughput 6.05007K wps
[Epoch 0 Batch 990/2125] avg loss 0.0083461, throughput 6.05783K wps
[Epoch 0 Batch 1020/2125] avg loss 0.0079939, throughput 6.05415K wps
[Epoch 0 Batch 1050/2125] avg loss 0.00812602, throughput 6.05618K wps
[Epoch 0 Batch 1080/2125] avg loss 0.00776078, throughput 6.044K wps
[Epoch 0 Batch 1110/2125] avg loss 0.00784576, throughput 6.04358K wps
[Epoch 0 Batch 1140/2125] avg loss 0.00760725, throughput 6.04737K wps
[Epoch 0 Batch 1170/2125] avg loss 0.00753368, throughput 6.04838K wps
[Epoch 0 Batch 1200/2125] avg loss 0.00737966, throughput 6.05171K wps
[Epoch 0 Batch 1230/2125] avg loss 0.00701846, throughput 6.04629K wps
[Epoch 0 Batch 1260/2125] avg loss 0.0071524, throughput 6.03264K wps
[Epoch 0 Batch 1290/2125] avg loss 0.00713113, throughput 6.05256K wps
[Epoch 0 Batch 1320/2125] avg loss 0.00696586, throughput 6.04338K wps
[Epoch 0 Batch 1350/2125] avg loss 0.00669066, throughput 6.04477K wps
[Epoch 0 Batch 1380/2125] avg loss 0.00680478, throughput 6.05472K wps
[Epoch 0 Batch 1410/2125] avg loss 0.00671461, throughput 6.05227K wps
[Epoch 0 Batch 1440/2125] avg loss 0.00665872, throughput 6.05157K wps
[Epoch 0 Batch 1470/2125] avg loss 0.00692924, throughput 6.04074K wps
[Epoch 0 Batch 1500/2125] avg loss 0.00661677, throughput 6.05028K wps
[Epoch 0 Batch 1530/2125] avg loss 0.00667788, throughput 6.04644K wps
[Epoch 0 Batch 1560/2125] avg loss 0.00681418, throughput 6.04478K wps
[Epoch 0 Batch 1590/2125] avg loss 0.00619376, throughput 6.03796K wps
[Epoch 0 Batch 1620/2125] avg loss 0.00632291, throughput 6.04602K wps
[Epoch 0 Batch 1650/2125] avg loss 0.00640708, throughput 6.04553K wps
[Epoch 0 Batch 1680/2125] avg loss 0.00634838, throughput 6.03631K wps
[Epoch 0 Batch 1710/2125] avg loss 0.00634249, throughput 6.03543K wps
[Epoch 0 Batch 1740/2125] avg loss 0.00624135, throughput 6.03105K wps
[Epoch 0 Batch 1770/2125] avg loss 0.00589104, throughput 6.04064K wps
[Epoch 0 Batch 1800/2125] avg loss 0.00650942, throughput 6.03518K wps
[Epoch 0 Batch 1830/2125] avg loss 0.00615124, throughput 6.04134K wps
[Epoch 0 Batch 1860/2125] avg loss 0.00593867, throughput 6.04025K wps
[Epoch 0 Batch 1890/2125] avg loss 0.00599553, throughput 6.04631K wps
[Epoch 0 Batch 1920/2125] avg loss 0.00539364, throughput 6.04263K wps
[Epoch 0 Batch 1950/2125] avg loss 0.00604897, throughput 6.03636K wps
[Epoch 0 Batch 1980/2125] avg loss 0.00627076, throughput 6.03568K wps
[Epoch 0 Batch 2010/2125] avg loss 0.00580447, throughput 6.0317K wps
[Epoch 0 Batch 2040/2125] avg loss 0.00590063, throughput 6.04459K wps
[Epoch 0 Batch 2070/2125] avg loss 0.00524726, throughput 6.04066K wps
[Epoch 0 Batch 2100/2125] avg loss 0.00588659, throughput 6.03332K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 0] train avg loss 0.00889854, test acc 0.8930, test avg loss 0.279222, throughput 5.96662K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 1 Batch 30/2125] avg loss 0.00561269, throughput 6.18097K wps
[Epoch 1 Batch 60/2125] avg loss 0.00532812, throughput 6.03743K wps
[Epoch 1 Batch 90/2125] avg loss 0.00541954, throughput 6.04267K wps
[Epoch 1 Batch 120/2125] avg loss 0.00544336, throughput 6.03444K wps
[Epoch 1 Batch 150/2125] avg loss 0.00520972, throughput 6.02785K wps
[Epoch 1 Batch 180/2125] avg loss 0.00485277, throughput 6.03183K wps
[Epoch 1 Batch 210/2125] avg loss 0.0052734, throughput 6.02935K wps
[Epoch 1 Batch 240/2125] avg loss 0.00510196, throughput 6.03061K wps
[Epoch 1 Batch 270/2125] avg loss 0.00541268, throughput 6.02883K wps
[Epoch 1 Batch 300/2125] avg loss 0.00487362, throughput 6.03165K wps
[Epoch 1 Batch 330/2125] avg loss 0.00518811, throughput 6.03327K wps
[Epoch 1 Batch 360/2125] avg loss 0.00480206, throughput 6.02454K wps
[Epoch 1 Batch 390/2125] avg loss 0.00541896, throughput 6.02393K wps
[Epoch 1 Batch 420/2125] avg loss 0.00449905, throughput 6.02714K wps
[Epoch 1 Batch 450/2125] avg loss 0.00527598, throughput 6.03365K wps
[Epoch 1 Batch 480/2125] avg loss 0.00492536, throughput 6.02611K wps
[Epoch 1 Batch 510/2125] avg loss 0.00469485, throughput 6.03017K wps
[Epoch 1 Batch 540/2125] avg loss 0.00481728, throughput 6.03267K wps
[Epoch 1 Batch 570/2125] avg loss 0.00513394, throughput 6.01846K wps
[Epoch 1 Batch 600/2125] avg loss 0.00492796, throughput 6.04245K wps
[Epoch 1 Batch 630/2125] avg loss 0.0048846, throughput 6.03608K wps
[Epoch 1 Batch 660/2125] avg loss 0.00514504, throughput 6.0344K wps
[Epoch 1 Batch 690/2125] avg loss 0.00525688, throughput 6.03488K wps
[Epoch 1 Batch 720/2125] avg loss 0.00463184, throughput 6.0419K wps
[Epoch 1 Batch 750/2125] avg loss 0.00481528, throughput 6.0314K wps
[Epoch 1 Batch 780/2125] avg loss 0.00503578, throughput 6.03532K wps
[Epoch 1 Batch 810/2125] avg loss 0.00512864, throughput 6.03828K wps
[Epoch 1 Batch 840/2125] avg loss 0.00476532, throughput 6.02524K wps
[Epoch 1 Batch 870/2125] avg loss 0.00437941, throughput 6.02447K wps
[Epoch 1 Batch 900/2125] avg loss 0.00449429, throughput 6.03371K wps
[Epoch 1 Batch 930/2125] avg loss 0.00448071, throughput 6.03466K wps
[Epoch 1 Batch 960/2125] avg loss 0.00436185, throughput 6.03271K wps
[Epoch 1 Batch 990/2125] avg loss 0.00476991, throughput 6.03082K wps
[Epoch 1 Batch 1020/2125] avg loss 0.00485248, throughput 6.0283K wps
[Epoch 1 Batch 1050/2125] avg loss 0.00486911, throughput 6.03283K wps
[Epoch 1 Batch 1080/2125] avg loss 0.00488294, throughput 6.02241K wps
[Epoch 1 Batch 1110/2125] avg loss 0.00481628, throughput 6.02468K wps
[Epoch 1 Batch 1140/2125] avg loss 0.00483072, throughput 6.03128K wps
[Epoch 1 Batch 1170/2125] avg loss 0.0049272, throughput 6.03886K wps
[Epoch 1 Batch 1200/2125] avg loss 0.0047501, throughput 6.02111K wps
[Epoch 1 Batch 1230/2125] avg loss 0.00417017, throughput 6.02575K wps
[Epoch 1 Batch 1260/2125] avg loss 0.00433226, throughput 6.04036K wps
[Epoch 1 Batch 1290/2125] avg loss 0.00452263, throughput 6.01695K wps
[Epoch 1 Batch 1320/2125] avg loss 0.00474539, throughput 6.01992K wps
[Epoch 1 Batch 1350/2125] avg loss 0.00485473, throughput 6.02199K wps
[Epoch 1 Batch 1380/2125] avg loss 0.00453053, throughput 6.02148K wps
[Epoch 1 Batch 1410/2125] avg loss 0.00464846, throughput 6.0261K wps
[Epoch 1 Batch 1440/2125] avg loss 0.00484329, throughput 6.02989K wps
[Epoch 1 Batch 1470/2125] avg loss 0.00461822, throughput 6.03254K wps
[Epoch 1 Batch 1500/2125] avg loss 0.004756, throughput 6.02205K wps
[Epoch 1 Batch 1530/2125] avg loss 0.0043247, throughput 6.01813K wps
[Epoch 1 Batch 1560/2125] avg loss 0.00459485, throughput 6.03643K wps
[Epoch 1 Batch 1590/2125] avg loss 0.00445727, throughput 6.03002K wps
[Epoch 1 Batch 1620/2125] avg loss 0.00475048, throughput 6.0224K wps
[Epoch 1 Batch 1650/2125] avg loss 0.00435204, throughput 6.04173K wps
[Epoch 1 Batch 1680/2125] avg loss 0.00463807, throughput 6.0345K wps
[Epoch 1 Batch 1710/2125] avg loss 0.0046142, throughput 6.02205K wps
[Epoch 1 Batch 1740/2125] avg loss 0.0047496, throughput 6.03595K wps
[Epoch 1 Batch 1770/2125] avg loss 0.00454747, throughput 6.02907K wps
[Epoch 1 Batch 1800/2125] avg loss 0.00466231, throughput 6.02229K wps
[Epoch 1 Batch 1830/2125] avg loss 0.004634, throughput 6.02346K wps
[Epoch 1 Batch 1860/2125] avg loss 0.00412036, throughput 5.99939K wps
[Epoch 1 Batch 1890/2125] avg loss 0.00490227, throughput 6.0146K wps
[Epoch 1 Batch 1920/2125] avg loss 0.00446759, throughput 6.03262K wps
[Epoch 1 Batch 1950/2125] avg loss 0.00500863, throughput 6.02351K wps
[Epoch 1 Batch 1980/2125] avg loss 0.00497393, throughput 6.02363K wps
[Epoch 1 Batch 2010/2125] avg loss 0.00456391, throughput 6.02503K wps
[Epoch 1 Batch 2040/2125] avg loss 0.00464973, throughput 6.02447K wps
[Epoch 1 Batch 2070/2125] avg loss 0.0043194, throughput 6.03011K wps
[Epoch 1 Batch 2100/2125] avg loss 0.00502493, throughput 6.03908K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 1] train avg loss 0.00480397, test acc 0.9084, test avg loss 0.241493, throughput 6.03121K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 2 Batch 30/2125] avg loss 0.00378288, throughput 6.16067K wps
[Epoch 2 Batch 60/2125] avg loss 0.00379975, throughput 6.02011K wps
[Epoch 2 Batch 90/2125] avg loss 0.0042167, throughput 6.01632K wps
[Epoch 2 Batch 120/2125] avg loss 0.00378315, throughput 6.02108K wps
[Epoch 2 Batch 150/2125] avg loss 0.00402631, throughput 6.02775K wps
[Epoch 2 Batch 180/2125] avg loss 0.00410664, throughput 6.0209K wps
[Epoch 2 Batch 210/2125] avg loss 0.00417997, throughput 6.02154K wps
[Epoch 2 Batch 240/2125] avg loss 0.00404273, throughput 6.02348K wps
[Epoch 2 Batch 270/2125] avg loss 0.00345837, throughput 6.02958K wps
[Epoch 2 Batch 300/2125] avg loss 0.00429986, throughput 6.03088K wps
[Epoch 2 Batch 330/2125] avg loss 0.0035167, throughput 6.02168K wps
[Epoch 2 Batch 360/2125] avg loss 0.00363477, throughput 6.01928K wps
[Epoch 2 Batch 390/2125] avg loss 0.00441074, throughput 6.02817K wps
[Epoch 2 Batch 420/2125] avg loss 0.00407198, throughput 6.01771K wps
[Epoch 2 Batch 450/2125] avg loss 0.00394415, throughput 6.02922K wps
[Epoch 2 Batch 480/2125] avg loss 0.00385617, throughput 6.02964K wps
[Epoch 2 Batch 510/2125] avg loss 0.00409874, throughput 6.03216K wps
[Epoch 2 Batch 540/2125] avg loss 0.00417068, throughput 6.03985K wps
[Epoch 2 Batch 570/2125] avg loss 0.00397655, throughput 6.01746K wps
[Epoch 2 Batch 600/2125] avg loss 0.00383564, throughput 6.02583K wps
[Epoch 2 Batch 630/2125] avg loss 0.00358196, throughput 6.0307K wps
[Epoch 2 Batch 660/2125] avg loss 0.00377499, throughput 6.02126K wps
[Epoch 2 Batch 690/2125] avg loss 0.00419209, throughput 6.02534K wps
[Epoch 2 Batch 720/2125] avg loss 0.00386005, throughput 6.02979K wps
[Epoch 2 Batch 750/2125] avg loss 0.00390685, throughput 6.02071K wps
[Epoch 2 Batch 780/2125] avg loss 0.00352134, throughput 6.02616K wps
[Epoch 2 Batch 810/2125] avg loss 0.00381646, throughput 6.01342K wps
[Epoch 2 Batch 840/2125] avg loss 0.00353567, throughput 6.01796K wps
[Epoch 2 Batch 870/2125] avg loss 0.00445868, throughput 6.01742K wps
[Epoch 2 Batch 900/2125] avg loss 0.00360173, throughput 6.01045K wps
[Epoch 2 Batch 930/2125] avg loss 0.00398804, throughput 6.02049K wps
[Epoch 2 Batch 960/2125] avg loss 0.00337316, throughput 6.03263K wps
[Epoch 2 Batch 990/2125] avg loss 0.00412185, throughput 6.02943K wps
[Epoch 2 Batch 1020/2125] avg loss 0.00388467, throughput 6.02132K wps
[Epoch 2 Batch 1050/2125] avg loss 0.00356656, throughput 6.01249K wps
[Epoch 2 Batch 1080/2125] avg loss 0.00390035, throughput 6.02519K wps
[Epoch 2 Batch 1110/2125] avg loss 0.0041124, throughput 6.02628K wps
[Epoch 2 Batch 1140/2125] avg loss 0.00413335, throughput 6.02239K wps
[Epoch 2 Batch 1170/2125] avg loss 0.00352991, throughput 6.01488K wps
[Epoch 2 Batch 1200/2125] avg loss 0.00438222, throughput 6.01225K wps
[Epoch 2 Batch 1230/2125] avg loss 0.0035305, throughput 6.01519K wps
[Epoch 2 Batch 1260/2125] avg loss 0.00394693, throughput 6.0281K wps
[Epoch 2 Batch 1290/2125] avg loss 0.0036885, throughput 6.02721K wps
[Epoch 2 Batch 1320/2125] avg loss 0.00382872, throughput 6.01345K wps
[Epoch 2 Batch 1350/2125] avg loss 0.0042578, throughput 6.01743K wps
[Epoch 2 Batch 1380/2125] avg loss 0.00352395, throughput 6.02091K wps
[Epoch 2 Batch 1410/2125] avg loss 0.00449212, throughput 6.02061K wps
[Epoch 2 Batch 1440/2125] avg loss 0.00367191, throughput 6.01238K wps
[Epoch 2 Batch 1470/2125] avg loss 0.00405565, throughput 6.01558K wps
[Epoch 2 Batch 1500/2125] avg loss 0.00362755, throughput 6.01427K wps
[Epoch 2 Batch 1530/2125] avg loss 0.00363849, throughput 6.01912K wps
[Epoch 2 Batch 1560/2125] avg loss 0.00419401, throughput 6.02055K wps
[Epoch 2 Batch 1590/2125] avg loss 0.00386007, throughput 6.01984K wps
[Epoch 2 Batch 1620/2125] avg loss 0.00354532, throughput 6.02261K wps
[Epoch 2 Batch 1650/2125] avg loss 0.00415341, throughput 6.0177K wps
[Epoch 2 Batch 1680/2125] avg loss 0.00411061, throughput 6.02489K wps
[Epoch 2 Batch 1710/2125] avg loss 0.00398758, throughput 6.0268K wps
[Epoch 2 Batch 1740/2125] avg loss 0.00401141, throughput 6.0102K wps
[Epoch 2 Batch 1770/2125] avg loss 0.00351123, throughput 6.02222K wps
[Epoch 2 Batch 1800/2125] avg loss 0.00408918, throughput 6.01202K wps
[Epoch 2 Batch 1830/2125] avg loss 0.00352344, throughput 6.02178K wps
[Epoch 2 Batch 1860/2125] avg loss 0.00370249, throughput 6.01689K wps
[Epoch 2 Batch 1890/2125] avg loss 0.00369327, throughput 6.01948K wps
[Epoch 2 Batch 1920/2125] avg loss 0.00402363, throughput 6.0237K wps
[Epoch 2 Batch 1950/2125] avg loss 0.00358791, throughput 6.03361K wps
[Epoch 2 Batch 1980/2125] avg loss 0.00360738, throughput 6.02192K wps
[Epoch 2 Batch 2010/2125] avg loss 0.00387046, throughput 6.02559K wps
[Epoch 2 Batch 2040/2125] avg loss 0.00372476, throughput 6.02428K wps
[Epoch 2 Batch 2070/2125] avg loss 0.00383992, throughput 6.03499K wps
[Epoch 2 Batch 2100/2125] avg loss 0.00408014, throughput 6.02132K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 2] train avg loss 0.00388178, test acc 0.9152, test avg loss 0.229892, throughput 6.02415K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 3 Batch 30/2125] avg loss 0.00329633, throughput 6.15708K wps
[Epoch 3 Batch 60/2125] avg loss 0.00341968, throughput 6.02623K wps
[Epoch 3 Batch 90/2125] avg loss 0.00343311, throughput 6.03134K wps
[Epoch 3 Batch 120/2125] avg loss 0.00349467, throughput 6.01985K wps
[Epoch 3 Batch 150/2125] avg loss 0.00319217, throughput 6.02449K wps
[Epoch 3 Batch 180/2125] avg loss 0.00372562, throughput 6.02767K wps
[Epoch 3 Batch 210/2125] avg loss 0.00319632, throughput 6.01599K wps
[Epoch 3 Batch 240/2125] avg loss 0.00300933, throughput 6.02374K wps
[Epoch 3 Batch 270/2125] avg loss 0.00288703, throughput 6.02617K wps
[Epoch 3 Batch 300/2125] avg loss 0.00316389, throughput 6.02573K wps
[Epoch 3 Batch 330/2125] avg loss 0.00312001, throughput 6.02461K wps
[Epoch 3 Batch 360/2125] avg loss 0.00350093, throughput 6.01783K wps
[Epoch 3 Batch 390/2125] avg loss 0.00296996, throughput 6.03313K wps
[Epoch 3 Batch 420/2125] avg loss 0.00321765, throughput 6.0324K wps
[Epoch 3 Batch 450/2125] avg loss 0.00364995, throughput 6.02378K wps
[Epoch 3 Batch 480/2125] avg loss 0.00361579, throughput 6.02895K wps
[Epoch 3 Batch 510/2125] avg loss 0.00366579, throughput 6.01317K wps
[Epoch 3 Batch 540/2125] avg loss 0.0036611, throughput 6.0246K wps
[Epoch 3 Batch 570/2125] avg loss 0.00355673, throughput 6.023K wps
[Epoch 3 Batch 600/2125] avg loss 0.00349155, throughput 6.02176K wps
[Epoch 3 Batch 630/2125] avg loss 0.0031149, throughput 6.01662K wps
[Epoch 3 Batch 660/2125] avg loss 0.00339862, throughput 6.02754K wps
[Epoch 3 Batch 690/2125] avg loss 0.00297542, throughput 6.02698K wps
[Epoch 3 Batch 720/2125] avg loss 0.00356097, throughput 6.02024K wps
[Epoch 3 Batch 750/2125] avg loss 0.00329657, throughput 6.01868K wps
[Epoch 3 Batch 780/2125] avg loss 0.00310237, throughput 6.01624K wps
[Epoch 3 Batch 810/2125] avg loss 0.00361155, throughput 6.02137K wps
[Epoch 3 Batch 840/2125] avg loss 0.00350197, throughput 6.02153K wps
[Epoch 3 Batch 870/2125] avg loss 0.00324632, throughput 6.01784K wps
[Epoch 3 Batch 900/2125] avg loss 0.00328363, throughput 6.02841K wps
[Epoch 3 Batch 930/2125] avg loss 0.0033988, throughput 6.02285K wps
[Epoch 3 Batch 960/2125] avg loss 0.00312493, throughput 6.00955K wps
[Epoch 3 Batch 990/2125] avg loss 0.00311125, throughput 6.02171K wps
[Epoch 3 Batch 1020/2125] avg loss 0.00317315, throughput 6.02494K wps
[Epoch 3 Batch 1050/2125] avg loss 0.00357406, throughput 6.02674K wps
[Epoch 3 Batch 1080/2125] avg loss 0.00341836, throughput 6.03112K wps
[Epoch 3 Batch 1110/2125] avg loss 0.00320742, throughput 6.01889K wps
[Epoch 3 Batch 1140/2125] avg loss 0.00361829, throughput 6.02816K wps
[Epoch 3 Batch 1170/2125] avg loss 0.0034175, throughput 6.0173K wps
[Epoch 3 Batch 1200/2125] avg loss 0.00366877, throughput 6.00778K wps
[Epoch 3 Batch 1230/2125] avg loss 0.00333758, throughput 6.00775K wps
[Epoch 3 Batch 1260/2125] avg loss 0.0033798, throughput 6.01779K wps
[Epoch 3 Batch 1290/2125] avg loss 0.00345485, throughput 6.02706K wps
[Epoch 3 Batch 1320/2125] avg loss 0.0033599, throughput 6.01528K wps
[Epoch 3 Batch 1350/2125] avg loss 0.0033127, throughput 6.02193K wps
[Epoch 3 Batch 1380/2125] avg loss 0.00291138, throughput 6.03433K wps
[Epoch 3 Batch 1410/2125] avg loss 0.00304158, throughput 6.02003K wps
[Epoch 3 Batch 1440/2125] avg loss 0.00325555, throughput 6.01303K wps
[Epoch 3 Batch 1470/2125] avg loss 0.00308925, throughput 6.01836K wps
[Epoch 3 Batch 1500/2125] avg loss 0.00342321, throughput 6.02697K wps
[Epoch 3 Batch 1530/2125] avg loss 0.00302186, throughput 6.0288K wps
[Epoch 3 Batch 1560/2125] avg loss 0.00342853, throughput 6.0139K wps
[Epoch 3 Batch 1590/2125] avg loss 0.00326577, throughput 6.01957K wps
[Epoch 3 Batch 1620/2125] avg loss 0.00320141, throughput 6.01603K wps
[Epoch 3 Batch 1650/2125] avg loss 0.00370802, throughput 6.02756K wps
[Epoch 3 Batch 1680/2125] avg loss 0.00363428, throughput 6.02024K wps
[Epoch 3 Batch 1710/2125] avg loss 0.00356219, throughput 6.01882K wps
[Epoch 3 Batch 1740/2125] avg loss 0.00374879, throughput 6.02289K wps
[Epoch 3 Batch 1770/2125] avg loss 0.00363636, throughput 6.02664K wps
[Epoch 3 Batch 1800/2125] avg loss 0.00329894, throughput 6.02246K wps
[Epoch 3 Batch 1830/2125] avg loss 0.00354369, throughput 6.02168K wps
[Epoch 3 Batch 1860/2125] avg loss 0.00309343, throughput 6.00595K wps
[Epoch 3 Batch 1890/2125] avg loss 0.00345957, throughput 6.01696K wps
[Epoch 3 Batch 1920/2125] avg loss 0.00397834, throughput 6.00512K wps
[Epoch 3 Batch 1950/2125] avg loss 0.00379751, throughput 6.01407K wps
[Epoch 3 Batch 1980/2125] avg loss 0.00315485, throughput 6.01925K wps
[Epoch 3 Batch 2010/2125] avg loss 0.00358171, throughput 6.0033K wps
[Epoch 3 Batch 2040/2125] avg loss 0.00307086, throughput 6.01343K wps
[Epoch 3 Batch 2070/2125] avg loss 0.00356996, throughput 6.01167K wps
[Epoch 3 Batch 2100/2125] avg loss 0.00297814, throughput 6.01457K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 3] train avg loss 0.00336021, test acc 0.9179, test avg loss 0.234045, throughput 6.02268K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 4 Batch 30/2125] avg loss 0.00266025, throughput 6.15885K wps
[Epoch 4 Batch 60/2125] avg loss 0.00259681, throughput 6.01388K wps
[Epoch 4 Batch 90/2125] avg loss 0.00287396, throughput 6.02441K wps
[Epoch 4 Batch 120/2125] avg loss 0.00308117, throughput 6.01646K wps
[Epoch 4 Batch 150/2125] avg loss 0.00333367, throughput 6.01788K wps
[Epoch 4 Batch 180/2125] avg loss 0.00245765, throughput 6.0311K wps
[Epoch 4 Batch 210/2125] avg loss 0.00289827, throughput 6.02066K wps
[Epoch 4 Batch 240/2125] avg loss 0.0029431, throughput 6.02755K wps
[Epoch 4 Batch 270/2125] avg loss 0.00279749, throughput 6.00569K wps
[Epoch 4 Batch 300/2125] avg loss 0.00270429, throughput 6.02326K wps
[Epoch 4 Batch 330/2125] avg loss 0.00286141, throughput 6.01547K wps
[Epoch 4 Batch 360/2125] avg loss 0.00300203, throughput 6.00617K wps
[Epoch 4 Batch 390/2125] avg loss 0.00296628, throughput 6.02361K wps
[Epoch 4 Batch 420/2125] avg loss 0.00292454, throughput 6.00537K wps
[Epoch 4 Batch 450/2125] avg loss 0.0030003, throughput 6.0249K wps
[Epoch 4 Batch 480/2125] avg loss 0.00294214, throughput 6.02609K wps
[Epoch 4 Batch 510/2125] avg loss 0.00266646, throughput 6.00804K wps
[Epoch 4 Batch 540/2125] avg loss 0.00266632, throughput 6.02019K wps
[Epoch 4 Batch 570/2125] avg loss 0.00316975, throughput 6.02607K wps
[Epoch 4 Batch 600/2125] avg loss 0.0032111, throughput 6.02006K wps
[Epoch 4 Batch 630/2125] avg loss 0.00284225, throughput 6.01752K wps
[Epoch 4 Batch 660/2125] avg loss 0.00283244, throughput 6.01101K wps
[Epoch 4 Batch 690/2125] avg loss 0.00292054, throughput 5.99128K wps
[Epoch 4 Batch 720/2125] avg loss 0.00276408, throughput 6.02985K wps
[Epoch 4 Batch 750/2125] avg loss 0.00270589, throughput 6.02231K wps
[Epoch 4 Batch 780/2125] avg loss 0.00256596, throughput 6.02188K wps
[Epoch 4 Batch 810/2125] avg loss 0.00332758, throughput 6.03238K wps
[Epoch 4 Batch 840/2125] avg loss 0.00345426, throughput 6.01958K wps
[Epoch 4 Batch 870/2125] avg loss 0.00294129, throughput 6.01155K wps
[Epoch 4 Batch 900/2125] avg loss 0.00283353, throughput 6.0167K wps
[Epoch 4 Batch 930/2125] avg loss 0.00300098, throughput 6.02054K wps
[Epoch 4 Batch 960/2125] avg loss 0.00315959, throughput 6.02545K wps
[Epoch 4 Batch 990/2125] avg loss 0.0030926, throughput 6.01709K wps
[Epoch 4 Batch 1020/2125] avg loss 0.00265336, throughput 6.01078K wps
[Epoch 4 Batch 1050/2125] avg loss 0.00330748, throughput 6.01059K wps
[Epoch 4 Batch 1080/2125] avg loss 0.00309418, throughput 6.01582K wps
[Epoch 4 Batch 1110/2125] avg loss 0.00301947, throughput 6.03444K wps
[Epoch 4 Batch 1140/2125] avg loss 0.00294563, throughput 6.01975K wps
[Epoch 4 Batch 1170/2125] avg loss 0.00291969, throughput 6.01953K wps
[Epoch 4 Batch 1200/2125] avg loss 0.00304187, throughput 6.02007K wps
[Epoch 4 Batch 1230/2125] avg loss 0.00327234, throughput 6.0126K wps
[Epoch 4 Batch 1260/2125] avg loss 0.00281882, throughput 6.02171K wps
[Epoch 4 Batch 1290/2125] avg loss 0.00297727, throughput 6.01382K wps
[Epoch 4 Batch 1320/2125] avg loss 0.00330625, throughput 6.02473K wps
[Epoch 4 Batch 1350/2125] avg loss 0.00269154, throughput 6.02435K wps
[Epoch 4 Batch 1380/2125] avg loss 0.00283565, throughput 6.01511K wps
[Epoch 4 Batch 1410/2125] avg loss 0.00313254, throughput 6.0156K wps
[Epoch 4 Batch 1440/2125] avg loss 0.00305659, throughput 6.01877K wps
[Epoch 4 Batch 1470/2125] avg loss 0.00307607, throughput 6.01865K wps
[Epoch 4 Batch 1500/2125] avg loss 0.00299291, throughput 6.02178K wps
[Epoch 4 Batch 1530/2125] avg loss 0.00275254, throughput 6.0312K wps
[Epoch 4 Batch 1560/2125] avg loss 0.00316133, throughput 6.01926K wps
[Epoch 4 Batch 1590/2125] avg loss 0.00303249, throughput 6.01204K wps
[Epoch 4 Batch 1620/2125] avg loss 0.00288925, throughput 6.02083K wps
[Epoch 4 Batch 1650/2125] avg loss 0.00307524, throughput 6.01999K wps
[Epoch 4 Batch 1680/2125] avg loss 0.00264711, throughput 6.01685K wps
[Epoch 4 Batch 1710/2125] avg loss 0.00303186, throughput 6.02136K wps
[Epoch 4 Batch 1740/2125] avg loss 0.00302763, throughput 6.01554K wps
[Epoch 4 Batch 1770/2125] avg loss 0.00265943, throughput 6.00421K wps
[Epoch 4 Batch 1800/2125] avg loss 0.00295478, throughput 6.01557K wps
[Epoch 4 Batch 1830/2125] avg loss 0.00312533, throughput 6.01973K wps
[Epoch 4 Batch 1860/2125] avg loss 0.00293816, throughput 6.01278K wps
[Epoch 4 Batch 1890/2125] avg loss 0.00351002, throughput 6.02173K wps
[Epoch 4 Batch 1920/2125] avg loss 0.00309331, throughput 6.01845K wps
[Epoch 4 Batch 1950/2125] avg loss 0.00332076, throughput 6.0262K wps
[Epoch 4 Batch 1980/2125] avg loss 0.0034739, throughput 6.01439K wps
[Epoch 4 Batch 2010/2125] avg loss 0.00320883, throughput 6.0242K wps
[Epoch 4 Batch 2040/2125] avg loss 0.00279056, throughput 6.01514K wps
[Epoch 4 Batch 2070/2125] avg loss 0.00349205, throughput 6.01928K wps
[Epoch 4 Batch 2100/2125] avg loss 0.00317856, throughput 6.02242K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 4] train avg loss 0.00298122, test acc 0.9177, test avg loss 0.233191, throughput 6.02063K wps
[Epoch 5 Batch 30/2125] avg loss 0.002266, throughput 6.15136K wps
[Epoch 5 Batch 60/2125] avg loss 0.00258292, throughput 6.00632K wps
[Epoch 5 Batch 90/2125] avg loss 0.00261513, throughput 6.01767K wps
[Epoch 5 Batch 120/2125] avg loss 0.00236579, throughput 6.00488K wps
[Epoch 5 Batch 150/2125] avg loss 0.00283505, throughput 6.01023K wps
[Epoch 5 Batch 180/2125] avg loss 0.00255751, throughput 6.01691K wps
[Epoch 5 Batch 210/2125] avg loss 0.00271899, throughput 6.0121K wps
[Epoch 5 Batch 240/2125] avg loss 0.00231081, throughput 6.00793K wps
[Epoch 5 Batch 270/2125] avg loss 0.002578, throughput 6.02047K wps
[Epoch 5 Batch 300/2125] avg loss 0.00286929, throughput 6.00367K wps
[Epoch 5 Batch 330/2125] avg loss 0.00285696, throughput 6.01059K wps
[Epoch 5 Batch 360/2125] avg loss 0.00217682, throughput 6.02371K wps
[Epoch 5 Batch 390/2125] avg loss 0.00242748, throughput 6.01082K wps
[Epoch 5 Batch 420/2125] avg loss 0.0029081, throughput 6.01749K wps
[Epoch 5 Batch 450/2125] avg loss 0.002882, throughput 6.0142K wps
[Epoch 5 Batch 480/2125] avg loss 0.00277555, throughput 6.0193K wps
[Epoch 5 Batch 510/2125] avg loss 0.00214411, throughput 6.01853K wps
[Epoch 5 Batch 540/2125] avg loss 0.00236509, throughput 6.02171K wps
[Epoch 5 Batch 570/2125] avg loss 0.00269093, throughput 6.01898K wps
[Epoch 5 Batch 600/2125] avg loss 0.00262393, throughput 6.01465K wps
[Epoch 5 Batch 630/2125] avg loss 0.00248691, throughput 6.01901K wps
[Epoch 5 Batch 660/2125] avg loss 0.00271584, throughput 6.01917K wps
[Epoch 5 Batch 690/2125] avg loss 0.00282762, throughput 6.01094K wps
[Epoch 5 Batch 720/2125] avg loss 0.00255462, throughput 6.00591K wps
[Epoch 5 Batch 750/2125] avg loss 0.00282782, throughput 6.01582K wps
[Epoch 5 Batch 780/2125] avg loss 0.00254651, throughput 6.01097K wps
[Epoch 5 Batch 810/2125] avg loss 0.00290707, throughput 6.02047K wps
[Epoch 5 Batch 840/2125] avg loss 0.00262968, throughput 6.01899K wps
[Epoch 5 Batch 870/2125] avg loss 0.00246314, throughput 6.02158K wps
[Epoch 5 Batch 900/2125] avg loss 0.00299079, throughput 6.02227K wps
[Epoch 5 Batch 930/2125] avg loss 0.00251947, throughput 6.02393K wps
[Epoch 5 Batch 960/2125] avg loss 0.00286057, throughput 6.0196K wps
[Epoch 5 Batch 990/2125] avg loss 0.00295726, throughput 6.0171K wps
[Epoch 5 Batch 1020/2125] avg loss 0.00330064, throughput 6.02635K wps
[Epoch 5 Batch 1050/2125] avg loss 0.00292193, throughput 6.01763K wps
[Epoch 5 Batch 1080/2125] avg loss 0.00297889, throughput 6.023K wps
[Epoch 5 Batch 1110/2125] avg loss 0.00231156, throughput 6.01802K wps
[Epoch 5 Batch 1140/2125] avg loss 0.00279298, throughput 6.01194K wps
[Epoch 5 Batch 1170/2125] avg loss 0.00335635, throughput 6.00787K wps
[Epoch 5 Batch 1200/2125] avg loss 0.00251146, throughput 6.0251K wps
[Epoch 5 Batch 1230/2125] avg loss 0.00279162, throughput 6.02171K wps
[Epoch 5 Batch 1260/2125] avg loss 0.00265361, throughput 6.02782K wps
[Epoch 5 Batch 1290/2125] avg loss 0.00251453, throughput 6.02191K wps
[Epoch 5 Batch 1320/2125] avg loss 0.00264515, throughput 6.02679K wps
[Epoch 5 Batch 1350/2125] avg loss 0.00248954, throughput 6.01827K wps
[Epoch 5 Batch 1380/2125] avg loss 0.00270039, throughput 6.01786K wps
[Epoch 5 Batch 1410/2125] avg loss 0.00286901, throughput 6.01535K wps
[Epoch 5 Batch 1440/2125] avg loss 0.00319263, throughput 6.00976K wps
[Epoch 5 Batch 1470/2125] avg loss 0.00258001, throughput 6.02117K wps
[Epoch 5 Batch 1500/2125] avg loss 0.00267841, throughput 6.02077K wps
[Epoch 5 Batch 1530/2125] avg loss 0.00258607, throughput 6.02191K wps
[Epoch 5 Batch 1560/2125] avg loss 0.00267857, throughput 6.02158K wps
[Epoch 5 Batch 1590/2125] avg loss 0.00293231, throughput 6.01361K wps
[Epoch 5 Batch 1620/2125] avg loss 0.00250702, throughput 6.01928K wps
[Epoch 5 Batch 1650/2125] avg loss 0.00250209, throughput 6.01549K wps
[Epoch 5 Batch 1680/2125] avg loss 0.00296369, throughput 6.02271K wps
[Epoch 5 Batch 1710/2125] avg loss 0.00238726, throughput 6.01431K wps
[Epoch 5 Batch 1740/2125] avg loss 0.00261131, throughput 6.01672K wps
[Epoch 5 Batch 1770/2125] avg loss 0.00254073, throughput 6.01258K wps
[Epoch 5 Batch 1800/2125] avg loss 0.00258452, throughput 6.01479K wps
[Epoch 5 Batch 1830/2125] avg loss 0.00227937, throughput 6.01754K wps
[Epoch 5 Batch 1860/2125] avg loss 0.00298834, throughput 6.01824K wps
[Epoch 5 Batch 1890/2125] avg loss 0.00305386, throughput 6.02103K wps
[Epoch 5 Batch 1920/2125] avg loss 0.00277241, throughput 6.01351K wps
[Epoch 5 Batch 1950/2125] avg loss 0.00300737, throughput 6.02004K wps
[Epoch 5 Batch 1980/2125] avg loss 0.00319979, throughput 6.02786K wps
[Epoch 5 Batch 2010/2125] avg loss 0.00300186, throughput 6.01881K wps
[Epoch 5 Batch 2040/2125] avg loss 0.00299978, throughput 6.00717K wps
[Epoch 5 Batch 2070/2125] avg loss 0.00280821, throughput 6.02016K wps
[Epoch 5 Batch 2100/2125] avg loss 0.00222502, throughput 6.01845K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 5] train avg loss 0.00269941, test acc 0.9201, test avg loss 0.240024, throughput 6.01902K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 6 Batch 30/2125] avg loss 0.00192233, throughput 6.16515K wps
[Epoch 6 Batch 60/2125] avg loss 0.00239781, throughput 6.02379K wps
[Epoch 6 Batch 90/2125] avg loss 0.00207097, throughput 6.01697K wps
[Epoch 6 Batch 120/2125] avg loss 0.00235802, throughput 6.01453K wps
[Epoch 6 Batch 150/2125] avg loss 0.00220505, throughput 6.01069K wps
[Epoch 6 Batch 180/2125] avg loss 0.0022343, throughput 6.01581K wps
[Epoch 6 Batch 210/2125] avg loss 0.0024081, throughput 6.0074K wps
[Epoch 6 Batch 240/2125] avg loss 0.00193505, throughput 6.02289K wps
[Epoch 6 Batch 270/2125] avg loss 0.00250878, throughput 6.00724K wps
[Epoch 6 Batch 300/2125] avg loss 0.00199056, throughput 6.0081K wps
[Epoch 6 Batch 330/2125] avg loss 0.00239016, throughput 5.99979K wps
[Epoch 6 Batch 360/2125] avg loss 0.00195498, throughput 6.02173K wps
[Epoch 6 Batch 390/2125] avg loss 0.00246945, throughput 6.02171K wps
[Epoch 6 Batch 420/2125] avg loss 0.00296545, throughput 6.00897K wps
[Epoch 6 Batch 450/2125] avg loss 0.00233941, throughput 6.01583K wps
[Epoch 6 Batch 480/2125] avg loss 0.00253441, throughput 6.01456K wps
[Epoch 6 Batch 510/2125] avg loss 0.00221608, throughput 6.00995K wps
[Epoch 6 Batch 540/2125] avg loss 0.00218651, throughput 6.00781K wps
[Epoch 6 Batch 570/2125] avg loss 0.00248294, throughput 6.01802K wps
[Epoch 6 Batch 600/2125] avg loss 0.00226155, throughput 6.02535K wps
[Epoch 6 Batch 630/2125] avg loss 0.00242754, throughput 6.02488K wps
[Epoch 6 Batch 660/2125] avg loss 0.00248287, throughput 6.01794K wps
[Epoch 6 Batch 690/2125] avg loss 0.00260461, throughput 6.01234K wps
[Epoch 6 Batch 720/2125] avg loss 0.00264918, throughput 6.01408K wps
[Epoch 6 Batch 750/2125] avg loss 0.00222153, throughput 6.02575K wps
[Epoch 6 Batch 780/2125] avg loss 0.00252996, throughput 6.02087K wps
[Epoch 6 Batch 810/2125] avg loss 0.00227176, throughput 6.03148K wps
[Epoch 6 Batch 840/2125] avg loss 0.00212412, throughput 6.02182K wps
[Epoch 6 Batch 870/2125] avg loss 0.00245276, throughput 6.02794K wps
[Epoch 6 Batch 900/2125] avg loss 0.00302922, throughput 6.02133K wps
[Epoch 6 Batch 930/2125] avg loss 0.00247494, throughput 6.01733K wps
[Epoch 6 Batch 960/2125] avg loss 0.00235907, throughput 6.01794K wps
[Epoch 6 Batch 990/2125] avg loss 0.0020372, throughput 6.02545K wps
[Epoch 6 Batch 1020/2125] avg loss 0.00237273, throughput 6.01851K wps
[Epoch 6 Batch 1050/2125] avg loss 0.00256401, throughput 6.02599K wps
[Epoch 6 Batch 1080/2125] avg loss 0.00228862, throughput 6.01342K wps
[Epoch 6 Batch 1110/2125] avg loss 0.00229383, throughput 6.03221K wps
[Epoch 6 Batch 1140/2125] avg loss 0.00200327, throughput 6.02054K wps
[Epoch 6 Batch 1170/2125] avg loss 0.00243308, throughput 6.0191K wps
[Epoch 6 Batch 1200/2125] avg loss 0.00233369, throughput 6.01964K wps
[Epoch 6 Batch 1230/2125] avg loss 0.00211744, throughput 6.02201K wps
[Epoch 6 Batch 1260/2125] avg loss 0.00237833, throughput 6.01458K wps
[Epoch 6 Batch 1290/2125] avg loss 0.0025499, throughput 6.02268K wps
[Epoch 6 Batch 1320/2125] avg loss 0.00278492, throughput 6.02132K wps
[Epoch 6 Batch 1350/2125] avg loss 0.00274881, throughput 6.01752K wps
[Epoch 6 Batch 1380/2125] avg loss 0.00211643, throughput 6.01807K wps
[Epoch 6 Batch 1410/2125] avg loss 0.00252623, throughput 6.02834K wps
[Epoch 6 Batch 1440/2125] avg loss 0.00271443, throughput 6.01857K wps
[Epoch 6 Batch 1470/2125] avg loss 0.00240694, throughput 6.02077K wps
[Epoch 6 Batch 1500/2125] avg loss 0.00257398, throughput 6.01377K wps
[Epoch 6 Batch 1530/2125] avg loss 0.00274423, throughput 6.02598K wps
[Epoch 6 Batch 1560/2125] avg loss 0.00227469, throughput 6.01697K wps
[Epoch 6 Batch 1590/2125] avg loss 0.0025055, throughput 6.02679K wps
[Epoch 6 Batch 1620/2125] avg loss 0.00270308, throughput 6.01831K wps
[Epoch 6 Batch 1650/2125] avg loss 0.00236972, throughput 6.01252K wps
[Epoch 6 Batch 1680/2125] avg loss 0.00246744, throughput 6.00294K wps
[Epoch 6 Batch 1710/2125] avg loss 0.00276843, throughput 5.93829K wps
[Epoch 6 Batch 1740/2125] avg loss 0.00239753, throughput 5.99672K wps
[Epoch 6 Batch 1770/2125] avg loss 0.00240108, throughput 6.0319K wps
[Epoch 6 Batch 1800/2125] avg loss 0.0029332, throughput 6.02184K wps
[Epoch 6 Batch 1830/2125] avg loss 0.00309994, throughput 6.01496K wps
[Epoch 6 Batch 1860/2125] avg loss 0.00240021, throughput 6.01516K wps
[Epoch 6 Batch 1890/2125] avg loss 0.00212872, throughput 6.02349K wps
[Epoch 6 Batch 1920/2125] avg loss 0.00302542, throughput 6.02583K wps
[Epoch 6 Batch 1950/2125] avg loss 0.00261946, throughput 6.02423K wps
[Epoch 6 Batch 1980/2125] avg loss 0.00284279, throughput 6.02674K wps
[Epoch 6 Batch 2010/2125] avg loss 0.00285437, throughput 6.01909K wps
[Epoch 6 Batch 2040/2125] avg loss 0.00229951, throughput 6.0278K wps
[Epoch 6 Batch 2070/2125] avg loss 0.00283838, throughput 6.02163K wps
[Epoch 6 Batch 2100/2125] avg loss 0.00301136, throughput 6.02464K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 6] train avg loss 0.00244919, test acc 0.9212, test avg loss 0.247174, throughput 6.01968K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 7 Batch 30/2125] avg loss 0.00202062, throughput 6.15686K wps
[Epoch 7 Batch 60/2125] avg loss 0.00217135, throughput 6.01831K wps
[Epoch 7 Batch 90/2125] avg loss 0.00201642, throughput 6.01379K wps
[Epoch 7 Batch 120/2125] avg loss 0.00209621, throughput 6.02374K wps
[Epoch 7 Batch 150/2125] avg loss 0.00202836, throughput 6.02347K wps
[Epoch 7 Batch 180/2125] avg loss 0.00199345, throughput 6.02604K wps
[Epoch 7 Batch 210/2125] avg loss 0.0020707, throughput 6.01924K wps
[Epoch 7 Batch 240/2125] avg loss 0.00162263, throughput 6.00325K wps
[Epoch 7 Batch 270/2125] avg loss 0.0020139, throughput 6.00906K wps
[Epoch 7 Batch 300/2125] avg loss 0.00244518, throughput 6.01157K wps
[Epoch 7 Batch 330/2125] avg loss 0.00224331, throughput 6.02949K wps
[Epoch 7 Batch 360/2125] avg loss 0.00213068, throughput 6.02496K wps
[Epoch 7 Batch 390/2125] avg loss 0.00184593, throughput 6.01844K wps
[Epoch 7 Batch 420/2125] avg loss 0.00222306, throughput 6.01983K wps
[Epoch 7 Batch 450/2125] avg loss 0.00179457, throughput 6.01813K wps
[Epoch 7 Batch 480/2125] avg loss 0.0018563, throughput 6.01723K wps
[Epoch 7 Batch 510/2125] avg loss 0.00229117, throughput 6.0144K wps
[Epoch 7 Batch 540/2125] avg loss 0.0028272, throughput 6.01897K wps
[Epoch 7 Batch 570/2125] avg loss 0.00232186, throughput 6.01392K wps
[Epoch 7 Batch 600/2125] avg loss 0.0020769, throughput 6.01538K wps
[Epoch 7 Batch 630/2125] avg loss 0.00214815, throughput 6.0288K wps
[Epoch 7 Batch 660/2125] avg loss 0.00189695, throughput 6.02869K wps
[Epoch 7 Batch 690/2125] avg loss 0.00217327, throughput 6.00771K wps
[Epoch 7 Batch 720/2125] avg loss 0.00217847, throughput 6.01944K wps
[Epoch 7 Batch 750/2125] avg loss 0.00243964, throughput 6.02905K wps
[Epoch 7 Batch 780/2125] avg loss 0.0019589, throughput 6.03117K wps
[Epoch 7 Batch 810/2125] avg loss 0.00237444, throughput 6.00244K wps
[Epoch 7 Batch 840/2125] avg loss 0.00210078, throughput 6.01605K wps
[Epoch 7 Batch 870/2125] avg loss 0.00212865, throughput 6.01132K wps
[Epoch 7 Batch 900/2125] avg loss 0.00221865, throughput 6.01313K wps
[Epoch 7 Batch 930/2125] avg loss 0.00179696, throughput 6.0099K wps
[Epoch 7 Batch 960/2125] avg loss 0.00263186, throughput 6.01365K wps
[Epoch 7 Batch 990/2125] avg loss 0.00193493, throughput 6.00968K wps
[Epoch 7 Batch 1020/2125] avg loss 0.00207612, throughput 6.01253K wps
[Epoch 7 Batch 1050/2125] avg loss 0.00215651, throughput 6.00794K wps
[Epoch 7 Batch 1080/2125] avg loss 0.00237432, throughput 6.00503K wps
[Epoch 7 Batch 1110/2125] avg loss 0.00276681, throughput 6.00578K wps
[Epoch 7 Batch 1140/2125] avg loss 0.00253792, throughput 6.01817K wps
[Epoch 7 Batch 1170/2125] avg loss 0.00240757, throughput 6.00991K wps
[Epoch 7 Batch 1200/2125] avg loss 0.00204832, throughput 5.99969K wps
[Epoch 7 Batch 1230/2125] avg loss 0.00228504, throughput 6.01623K wps
[Epoch 7 Batch 1260/2125] avg loss 0.00227115, throughput 6.00309K wps
[Epoch 7 Batch 1290/2125] avg loss 0.00189741, throughput 6.01114K wps
[Epoch 7 Batch 1320/2125] avg loss 0.00247494, throughput 6.01152K wps
[Epoch 7 Batch 1350/2125] avg loss 0.00213737, throughput 6.02014K wps
[Epoch 7 Batch 1380/2125] avg loss 0.00245954, throughput 6.00678K wps
[Epoch 7 Batch 1410/2125] avg loss 0.00222627, throughput 6.01284K wps
[Epoch 7 Batch 1440/2125] avg loss 0.00210314, throughput 6.01753K wps
[Epoch 7 Batch 1470/2125] avg loss 0.00203042, throughput 6.02036K wps
[Epoch 7 Batch 1500/2125] avg loss 0.00183952, throughput 6.01651K wps
[Epoch 7 Batch 1530/2125] avg loss 0.00272776, throughput 6.02245K wps
[Epoch 7 Batch 1560/2125] avg loss 0.00228069, throughput 6.01601K wps
[Epoch 7 Batch 1590/2125] avg loss 0.00248911, throughput 6.01353K wps
[Epoch 7 Batch 1620/2125] avg loss 0.00241448, throughput 6.01958K wps
[Epoch 7 Batch 1650/2125] avg loss 0.00278445, throughput 6.01073K wps
[Epoch 7 Batch 1680/2125] avg loss 0.00255007, throughput 6.01385K wps
[Epoch 7 Batch 1710/2125] avg loss 0.00226164, throughput 6.01572K wps
[Epoch 7 Batch 1740/2125] avg loss 0.00287883, throughput 6.00687K wps
[Epoch 7 Batch 1770/2125] avg loss 0.00197915, throughput 6.01578K wps
[Epoch 7 Batch 1800/2125] avg loss 0.00244848, throughput 6.01259K wps
[Epoch 7 Batch 1830/2125] avg loss 0.00274109, throughput 6.01434K wps
[Epoch 7 Batch 1860/2125] avg loss 0.0025687, throughput 6.01017K wps
[Epoch 7 Batch 1890/2125] avg loss 0.00208887, throughput 6.01575K wps
[Epoch 7 Batch 1920/2125] avg loss 0.00243895, throughput 6.0217K wps
[Epoch 7 Batch 1950/2125] avg loss 0.00220696, throughput 6.00715K wps
[Epoch 7 Batch 1980/2125] avg loss 0.00212949, throughput 6.00976K wps
[Epoch 7 Batch 2010/2125] avg loss 0.00243932, throughput 6.01639K wps
[Epoch 7 Batch 2040/2125] avg loss 0.00227523, throughput 6.0148K wps
[Epoch 7 Batch 2070/2125] avg loss 0.00239413, throughput 6.00751K wps
[Epoch 7 Batch 2100/2125] avg loss 0.00216554, throughput 6.01479K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 7] train avg loss 0.0022391, test acc 0.9229, test avg loss 0.257001, throughput 6.01708K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 8 Batch 30/2125] avg loss 0.00194165, throughput 6.15192K wps
[Epoch 8 Batch 60/2125] avg loss 0.00169439, throughput 6.0209K wps
[Epoch 8 Batch 90/2125] avg loss 0.00169657, throughput 6.01627K wps
[Epoch 8 Batch 120/2125] avg loss 0.00178722, throughput 6.01381K wps
[Epoch 8 Batch 150/2125] avg loss 0.00172245, throughput 6.02335K wps
[Epoch 8 Batch 180/2125] avg loss 0.00194368, throughput 6.01437K wps
[Epoch 8 Batch 210/2125] avg loss 0.00200053, throughput 6.0118K wps
[Epoch 8 Batch 240/2125] avg loss 0.0017583, throughput 6.01847K wps
[Epoch 8 Batch 270/2125] avg loss 0.00222153, throughput 6.01632K wps
[Epoch 8 Batch 300/2125] avg loss 0.0018808, throughput 6.02446K wps
[Epoch 8 Batch 330/2125] avg loss 0.00184873, throughput 6.00352K wps
[Epoch 8 Batch 360/2125] avg loss 0.00180156, throughput 6.01192K wps
[Epoch 8 Batch 390/2125] avg loss 0.00207927, throughput 6.01913K wps
[Epoch 8 Batch 420/2125] avg loss 0.00187912, throughput 6.01394K wps
[Epoch 8 Batch 450/2125] avg loss 0.00185151, throughput 6.01195K wps
[Epoch 8 Batch 480/2125] avg loss 0.00199105, throughput 6.01203K wps
[Epoch 8 Batch 510/2125] avg loss 0.0021623, throughput 6.01186K wps
[Epoch 8 Batch 540/2125] avg loss 0.00185061, throughput 6.0147K wps
[Epoch 8 Batch 570/2125] avg loss 0.00211765, throughput 6.0122K wps
[Epoch 8 Batch 600/2125] avg loss 0.00190702, throughput 6.01718K wps
[Epoch 8 Batch 630/2125] avg loss 0.00180521, throughput 6.01021K wps
[Epoch 8 Batch 660/2125] avg loss 0.00217044, throughput 6.00487K wps
[Epoch 8 Batch 690/2125] avg loss 0.00219558, throughput 6.02234K wps
[Epoch 8 Batch 720/2125] avg loss 0.00169814, throughput 6.02171K wps
[Epoch 8 Batch 750/2125] avg loss 0.00217971, throughput 6.01822K wps
[Epoch 8 Batch 780/2125] avg loss 0.00207785, throughput 6.01445K wps
[Epoch 8 Batch 810/2125] avg loss 0.00186725, throughput 6.01464K wps
[Epoch 8 Batch 840/2125] avg loss 0.00208753, throughput 6.01564K wps
[Epoch 8 Batch 870/2125] avg loss 0.00233762, throughput 6.01727K wps
[Epoch 8 Batch 900/2125] avg loss 0.00210107, throughput 6.01232K wps
[Epoch 8 Batch 930/2125] avg loss 0.00180357, throughput 6.01352K wps
[Epoch 8 Batch 960/2125] avg loss 0.00217358, throughput 6.02522K wps
[Epoch 8 Batch 990/2125] avg loss 0.00217225, throughput 6.01535K wps
[Epoch 8 Batch 1020/2125] avg loss 0.00222565, throughput 6.02205K wps
[Epoch 8 Batch 1050/2125] avg loss 0.00207359, throughput 6.01442K wps
[Epoch 8 Batch 1080/2125] avg loss 0.00199008, throughput 6.02177K wps
[Epoch 8 Batch 1110/2125] avg loss 0.00201866, throughput 6.01239K wps
[Epoch 8 Batch 1140/2125] avg loss 0.0019341, throughput 6.01308K wps
[Epoch 8 Batch 1170/2125] avg loss 0.00215001, throughput 6.01423K wps
[Epoch 8 Batch 1200/2125] avg loss 0.00204672, throughput 6.01649K wps
[Epoch 8 Batch 1230/2125] avg loss 0.00198817, throughput 6.01448K wps
[Epoch 8 Batch 1260/2125] avg loss 0.00200644, throughput 6.01718K wps
[Epoch 8 Batch 1290/2125] avg loss 0.00213092, throughput 6.00849K wps
[Epoch 8 Batch 1320/2125] avg loss 0.00226168, throughput 6.01938K wps
[Epoch 8 Batch 1350/2125] avg loss 0.00189901, throughput 6.02392K wps
[Epoch 8 Batch 1380/2125] avg loss 0.00223916, throughput 6.01614K wps
[Epoch 8 Batch 1410/2125] avg loss 0.0021712, throughput 6.01774K wps
[Epoch 8 Batch 1440/2125] avg loss 0.00191818, throughput 6.00829K wps
[Epoch 8 Batch 1470/2125] avg loss 0.00204638, throughput 6.0144K wps
[Epoch 8 Batch 1500/2125] avg loss 0.00211871, throughput 6.01829K wps
[Epoch 8 Batch 1530/2125] avg loss 0.00241585, throughput 6.01949K wps
[Epoch 8 Batch 1560/2125] avg loss 0.0018464, throughput 6.01623K wps
[Epoch 8 Batch 1590/2125] avg loss 0.00203842, throughput 6.01809K wps
[Epoch 8 Batch 1620/2125] avg loss 0.00185013, throughput 6.01573K wps
[Epoch 8 Batch 1650/2125] avg loss 0.00215092, throughput 6.02168K wps
[Epoch 8 Batch 1680/2125] avg loss 0.00171337, throughput 6.02643K wps
[Epoch 8 Batch 1710/2125] avg loss 0.0020024, throughput 6.02015K wps
[Epoch 8 Batch 1740/2125] avg loss 0.00236147, throughput 6.01545K wps
[Epoch 8 Batch 1770/2125] avg loss 0.00257322, throughput 6.0183K wps
[Epoch 8 Batch 1800/2125] avg loss 0.00239423, throughput 6.0105K wps
[Epoch 8 Batch 1830/2125] avg loss 0.00230529, throughput 6.01204K wps
[Epoch 8 Batch 1860/2125] avg loss 0.00228218, throughput 6.01464K wps
[Epoch 8 Batch 1890/2125] avg loss 0.00196181, throughput 6.02066K wps
[Epoch 8 Batch 1920/2125] avg loss 0.00194806, throughput 6.00895K wps
[Epoch 8 Batch 1950/2125] avg loss 0.00203694, throughput 6.01834K wps
[Epoch 8 Batch 1980/2125] avg loss 0.00246739, throughput 6.00691K wps
[Epoch 8 Batch 2010/2125] avg loss 0.00228341, throughput 6.01209K wps
[Epoch 8 Batch 2040/2125] avg loss 0.00246577, throughput 6.00999K wps
[Epoch 8 Batch 2070/2125] avg loss 0.00242343, throughput 6.01807K wps
[Epoch 8 Batch 2100/2125] avg loss 0.0019853, throughput 6.01106K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 8] train avg loss 0.00205443, test acc 0.9249, test avg loss 0.265395, throughput 6.01758K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 9 Batch 30/2125] avg loss 0.00161072, throughput 6.15676K wps
[Epoch 9 Batch 60/2125] avg loss 0.00187389, throughput 6.01498K wps
[Epoch 9 Batch 90/2125] avg loss 0.00159963, throughput 6.02048K wps
[Epoch 9 Batch 120/2125] avg loss 0.00175688, throughput 6.02217K wps
[Epoch 9 Batch 150/2125] avg loss 0.00177901, throughput 6.01476K wps
[Epoch 9 Batch 180/2125] avg loss 0.00134942, throughput 6.02033K wps
[Epoch 9 Batch 210/2125] avg loss 0.00139904, throughput 6.01668K wps
[Epoch 9 Batch 240/2125] avg loss 0.00180726, throughput 6.02901K wps
[Epoch 9 Batch 270/2125] avg loss 0.00190185, throughput 6.02032K wps
[Epoch 9 Batch 300/2125] avg loss 0.00177372, throughput 6.02627K wps
[Epoch 9 Batch 330/2125] avg loss 0.00149637, throughput 6.02372K wps
[Epoch 9 Batch 360/2125] avg loss 0.00204333, throughput 6.03007K wps
[Epoch 9 Batch 390/2125] avg loss 0.0017299, throughput 6.01475K wps
[Epoch 9 Batch 420/2125] avg loss 0.00207305, throughput 6.01912K wps
[Epoch 9 Batch 450/2125] avg loss 0.00188322, throughput 6.01139K wps
[Epoch 9 Batch 480/2125] avg loss 0.00193118, throughput 6.01905K wps
[Epoch 9 Batch 510/2125] avg loss 0.00192007, throughput 5.98578K wps
[Epoch 9 Batch 540/2125] avg loss 0.00205725, throughput 5.97893K wps
[Epoch 9 Batch 570/2125] avg loss 0.00175896, throughput 6.0144K wps
[Epoch 9 Batch 600/2125] avg loss 0.00186995, throughput 6.01726K wps
[Epoch 9 Batch 630/2125] avg loss 0.00171525, throughput 6.01671K wps
[Epoch 9 Batch 660/2125] avg loss 0.00149189, throughput 6.02242K wps
[Epoch 9 Batch 690/2125] avg loss 0.00196265, throughput 6.02281K wps
[Epoch 9 Batch 720/2125] avg loss 0.00184022, throughput 6.03107K wps
[Epoch 9 Batch 750/2125] avg loss 0.00179706, throughput 6.02065K wps
[Epoch 9 Batch 780/2125] avg loss 0.00183945, throughput 6.01745K wps
[Epoch 9 Batch 810/2125] avg loss 0.00175176, throughput 6.02051K wps
[Epoch 9 Batch 840/2125] avg loss 0.00169189, throughput 6.01104K wps
[Epoch 9 Batch 870/2125] avg loss 0.00173344, throughput 6.02022K wps
[Epoch 9 Batch 900/2125] avg loss 0.00169748, throughput 6.02006K wps
[Epoch 9 Batch 930/2125] avg loss 0.00161149, throughput 6.01691K wps
[Epoch 9 Batch 960/2125] avg loss 0.00184808, throughput 6.00892K wps
[Epoch 9 Batch 990/2125] avg loss 0.00162143, throughput 6.01801K wps
[Epoch 9 Batch 1020/2125] avg loss 0.00184479, throughput 6.01646K wps
[Epoch 9 Batch 1050/2125] avg loss 0.00173171, throughput 6.01623K wps
[Epoch 9 Batch 1080/2125] avg loss 0.00211823, throughput 6.00728K wps
[Epoch 9 Batch 1110/2125] avg loss 0.00169557, throughput 6.01603K wps
[Epoch 9 Batch 1140/2125] avg loss 0.00178106, throughput 6.01177K wps
[Epoch 9 Batch 1170/2125] avg loss 0.00181508, throughput 6.01133K wps
[Epoch 9 Batch 1200/2125] avg loss 0.00214408, throughput 6.01744K wps
[Epoch 9 Batch 1230/2125] avg loss 0.00174044, throughput 6.02186K wps
[Epoch 9 Batch 1260/2125] avg loss 0.00184899, throughput 6.01214K wps
[Epoch 9 Batch 1290/2125] avg loss 0.00188608, throughput 6.01121K wps
[Epoch 9 Batch 1320/2125] avg loss 0.00180929, throughput 6.00484K wps
[Epoch 9 Batch 1350/2125] avg loss 0.00185828, throughput 6.00721K wps
[Epoch 9 Batch 1380/2125] avg loss 0.00191972, throughput 6.00954K wps
[Epoch 9 Batch 1410/2125] avg loss 0.00186677, throughput 6.01917K wps
[Epoch 9 Batch 1440/2125] avg loss 0.00177678, throughput 6.02776K wps
[Epoch 9 Batch 1470/2125] avg loss 0.00218766, throughput 6.01177K wps
[Epoch 9 Batch 1500/2125] avg loss 0.00209133, throughput 6.01661K wps
[Epoch 9 Batch 1530/2125] avg loss 0.00206169, throughput 6.01529K wps
[Epoch 9 Batch 1560/2125] avg loss 0.00191061, throughput 6.00576K wps
[Epoch 9 Batch 1590/2125] avg loss 0.00192632, throughput 6.00981K wps
[Epoch 9 Batch 1620/2125] avg loss 0.00176926, throughput 6.0155K wps
[Epoch 9 Batch 1650/2125] avg loss 0.00182262, throughput 6.00243K wps
[Epoch 9 Batch 1680/2125] avg loss 0.00211868, throughput 6.0129K wps
[Epoch 9 Batch 1710/2125] avg loss 0.00199809, throughput 6.01661K wps
[Epoch 9 Batch 1740/2125] avg loss 0.00236236, throughput 6.00889K wps
[Epoch 9 Batch 1770/2125] avg loss 0.00220667, throughput 6.013K wps
[Epoch 9 Batch 1800/2125] avg loss 0.00214932, throughput 6.02276K wps
[Epoch 9 Batch 1830/2125] avg loss 0.00215349, throughput 6.01814K wps
[Epoch 9 Batch 1860/2125] avg loss 0.00258506, throughput 6.01505K wps
[Epoch 9 Batch 1890/2125] avg loss 0.00209567, throughput 6.00682K wps
[Epoch 9 Batch 1920/2125] avg loss 0.00215111, throughput 6.0179K wps
[Epoch 9 Batch 1950/2125] avg loss 0.00221761, throughput 6.01464K wps
[Epoch 9 Batch 1980/2125] avg loss 0.00201257, throughput 6.00873K wps
[Epoch 9 Batch 2010/2125] avg loss 0.00188507, throughput 6.01782K wps
[Epoch 9 Batch 2040/2125] avg loss 0.00211831, throughput 6.01089K wps
[Epoch 9 Batch 2070/2125] avg loss 0.00197462, throughput 6.01522K wps
[Epoch 9 Batch 2100/2125] avg loss 0.00182952, throughput 6.0116K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 9] train avg loss 0.0018832, test acc 0.9238, test avg loss 0.273964, throughput 6.01705K wps
[Epoch 10 Batch 30/2125] avg loss 0.00154547, throughput 6.15872K wps
[Epoch 10 Batch 60/2125] avg loss 0.00167571, throughput 6.02243K wps
[Epoch 10 Batch 90/2125] avg loss 0.0019125, throughput 6.01319K wps
[Epoch 10 Batch 120/2125] avg loss 0.00189433, throughput 6.01156K wps
[Epoch 10 Batch 150/2125] avg loss 0.00147028, throughput 6.0123K wps
[Epoch 10 Batch 180/2125] avg loss 0.00158474, throughput 6.01473K wps
[Epoch 10 Batch 210/2125] avg loss 0.00162858, throughput 6.01975K wps
[Epoch 10 Batch 240/2125] avg loss 0.0015955, throughput 6.01969K wps
[Epoch 10 Batch 270/2125] avg loss 0.00188014, throughput 6.01387K wps
[Epoch 10 Batch 300/2125] avg loss 0.00194126, throughput 6.0222K wps
[Epoch 10 Batch 330/2125] avg loss 0.00155786, throughput 6.02043K wps
[Epoch 10 Batch 360/2125] avg loss 0.00149925, throughput 6.02116K wps
[Epoch 10 Batch 390/2125] avg loss 0.00184137, throughput 6.01091K wps
[Epoch 10 Batch 420/2125] avg loss 0.00164687, throughput 6.03128K wps
[Epoch 10 Batch 450/2125] avg loss 0.00210684, throughput 6.01889K wps
[Epoch 10 Batch 480/2125] avg loss 0.00184784, throughput 6.00839K wps
[Epoch 10 Batch 510/2125] avg loss 0.00151014, throughput 6.0251K wps
[Epoch 10 Batch 540/2125] avg loss 0.00171735, throughput 6.00679K wps
[Epoch 10 Batch 570/2125] avg loss 0.00133102, throughput 6.0087K wps
[Epoch 10 Batch 600/2125] avg loss 0.0014377, throughput 6.01966K wps
[Epoch 10 Batch 630/2125] avg loss 0.00179768, throughput 6.01358K wps
[Epoch 10 Batch 660/2125] avg loss 0.00180245, throughput 6.01282K wps
[Epoch 10 Batch 690/2125] avg loss 0.00161994, throughput 6.01009K wps
[Epoch 10 Batch 720/2125] avg loss 0.00195253, throughput 6.00801K wps
[Epoch 10 Batch 750/2125] avg loss 0.00176513, throughput 6.01395K wps
[Epoch 10 Batch 780/2125] avg loss 0.00153525, throughput 6.01043K wps
[Epoch 10 Batch 810/2125] avg loss 0.00167136, throughput 6.02303K wps
[Epoch 10 Batch 840/2125] avg loss 0.0018898, throughput 6.01672K wps
[Epoch 10 Batch 870/2125] avg loss 0.0016315, throughput 6.01429K wps
[Epoch 10 Batch 900/2125] avg loss 0.00171821, throughput 6.00551K wps
[Epoch 10 Batch 930/2125] avg loss 0.00168842, throughput 6.00767K wps
[Epoch 10 Batch 960/2125] avg loss 0.00203871, throughput 6.01479K wps
[Epoch 10 Batch 990/2125] avg loss 0.00186281, throughput 6.01315K wps
[Epoch 10 Batch 1020/2125] avg loss 0.00186208, throughput 6.0111K wps
[Epoch 10 Batch 1050/2125] avg loss 0.00181951, throughput 6.01348K wps
[Epoch 10 Batch 1080/2125] avg loss 0.00197001, throughput 6.01845K wps
[Epoch 10 Batch 1110/2125] avg loss 0.00157876, throughput 6.00847K wps
[Epoch 10 Batch 1140/2125] avg loss 0.00167403, throughput 6.01918K wps
[Epoch 10 Batch 1170/2125] avg loss 0.00181912, throughput 6.02596K wps
[Epoch 10 Batch 1200/2125] avg loss 0.00155661, throughput 6.02112K wps
[Epoch 10 Batch 1230/2125] avg loss 0.00162699, throughput 6.01918K wps
[Epoch 10 Batch 1260/2125] avg loss 0.00181716, throughput 6.02262K wps
[Epoch 10 Batch 1290/2125] avg loss 0.00150949, throughput 6.02562K wps
[Epoch 10 Batch 1320/2125] avg loss 0.00198742, throughput 6.01591K wps
[Epoch 10 Batch 1350/2125] avg loss 0.00189704, throughput 6.01407K wps
[Epoch 10 Batch 1380/2125] avg loss 0.00177395, throughput 6.01233K wps
[Epoch 10 Batch 1410/2125] avg loss 0.00161222, throughput 6.00983K wps
[Epoch 10 Batch 1440/2125] avg loss 0.00146055, throughput 6.01504K wps
[Epoch 10 Batch 1470/2125] avg loss 0.00173187, throughput 6.02377K wps
[Epoch 10 Batch 1500/2125] avg loss 0.00187839, throughput 6.01569K wps
[Epoch 10 Batch 1530/2125] avg loss 0.00180196, throughput 6.01806K wps
[Epoch 10 Batch 1560/2125] avg loss 0.00195319, throughput 6.02626K wps
[Epoch 10 Batch 1590/2125] avg loss 0.0017562, throughput 6.01581K wps
[Epoch 10 Batch 1620/2125] avg loss 0.0022276, throughput 6.01434K wps
[Epoch 10 Batch 1650/2125] avg loss 0.00176203, throughput 6.01642K wps
[Epoch 10 Batch 1680/2125] avg loss 0.00222834, throughput 6.01839K wps
[Epoch 10 Batch 1710/2125] avg loss 0.00177621, throughput 6.01352K wps
[Epoch 10 Batch 1740/2125] avg loss 0.00171831, throughput 6.01335K wps
[Epoch 10 Batch 1770/2125] avg loss 0.00190618, throughput 6.02386K wps
[Epoch 10 Batch 1800/2125] avg loss 0.0018506, throughput 6.02006K wps
[Epoch 10 Batch 1830/2125] avg loss 0.00187726, throughput 6.0193K wps
[Epoch 10 Batch 1860/2125] avg loss 0.00178552, throughput 6.02088K wps
[Epoch 10 Batch 1890/2125] avg loss 0.00229936, throughput 6.00844K wps
[Epoch 10 Batch 1920/2125] avg loss 0.00166587, throughput 6.02135K wps
[Epoch 10 Batch 1950/2125] avg loss 0.0018016, throughput 6.01547K wps
[Epoch 10 Batch 1980/2125] avg loss 0.00176483, throughput 6.01671K wps
[Epoch 10 Batch 2010/2125] avg loss 0.0022677, throughput 6.01963K wps
[Epoch 10 Batch 2040/2125] avg loss 0.00219904, throughput 6.00966K wps
[Epoch 10 Batch 2070/2125] avg loss 0.00174707, throughput 6.03142K wps
[Epoch 10 Batch 2100/2125] avg loss 0.00165896, throughput 6.02159K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 10] train avg loss 0.00177478, test acc 0.9258, test avg loss 0.285094, throughput 6.0188K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 11 Batch 30/2125] avg loss 0.00132532, throughput 6.15719K wps
[Epoch 11 Batch 60/2125] avg loss 0.00165941, throughput 6.0128K wps
[Epoch 11 Batch 90/2125] avg loss 0.00139682, throughput 6.0263K wps
[Epoch 11 Batch 120/2125] avg loss 0.00133825, throughput 6.01315K wps
[Epoch 11 Batch 150/2125] avg loss 0.00137215, throughput 6.00446K wps
[Epoch 11 Batch 180/2125] avg loss 0.00161289, throughput 6.01728K wps
[Epoch 11 Batch 210/2125] avg loss 0.00139325, throughput 6.01773K wps
[Epoch 11 Batch 240/2125] avg loss 0.00137649, throughput 6.01657K wps
[Epoch 11 Batch 270/2125] avg loss 0.00132649, throughput 6.02108K wps
[Epoch 11 Batch 300/2125] avg loss 0.00160659, throughput 6.02857K wps
[Epoch 11 Batch 330/2125] avg loss 0.00158187, throughput 6.01518K wps
[Epoch 11 Batch 360/2125] avg loss 0.00141766, throughput 6.02101K wps
[Epoch 11 Batch 390/2125] avg loss 0.00126441, throughput 6.01843K wps
[Epoch 11 Batch 420/2125] avg loss 0.00112723, throughput 6.0247K wps
[Epoch 11 Batch 450/2125] avg loss 0.00161123, throughput 6.01845K wps
[Epoch 11 Batch 480/2125] avg loss 0.00163232, throughput 6.01057K wps
[Epoch 11 Batch 510/2125] avg loss 0.00149205, throughput 6.02054K wps
[Epoch 11 Batch 540/2125] avg loss 0.00162793, throughput 6.01641K wps
[Epoch 11 Batch 570/2125] avg loss 0.00195476, throughput 6.02482K wps
[Epoch 11 Batch 600/2125] avg loss 0.00142364, throughput 6.01781K wps
[Epoch 11 Batch 630/2125] avg loss 0.00172855, throughput 6.00782K wps
[Epoch 11 Batch 660/2125] avg loss 0.00162091, throughput 6.01927K wps
[Epoch 11 Batch 690/2125] avg loss 0.00150445, throughput 6.02924K wps
[Epoch 11 Batch 720/2125] avg loss 0.00148111, throughput 6.01856K wps
[Epoch 11 Batch 750/2125] avg loss 0.0018917, throughput 6.01606K wps
[Epoch 11 Batch 780/2125] avg loss 0.00154718, throughput 6.02375K wps
[Epoch 11 Batch 810/2125] avg loss 0.00146236, throughput 6.01783K wps
[Epoch 11 Batch 840/2125] avg loss 0.00187273, throughput 6.01286K wps
[Epoch 11 Batch 870/2125] avg loss 0.0016205, throughput 6.01173K wps
[Epoch 11 Batch 900/2125] avg loss 0.00167477, throughput 6.02126K wps
[Epoch 11 Batch 930/2125] avg loss 0.00161951, throughput 6.02244K wps
[Epoch 11 Batch 960/2125] avg loss 0.00184429, throughput 6.01957K wps
[Epoch 11 Batch 990/2125] avg loss 0.00158349, throughput 6.01205K wps
[Epoch 11 Batch 1020/2125] avg loss 0.00182522, throughput 6.02397K wps
[Epoch 11 Batch 1050/2125] avg loss 0.00147489, throughput 6.0172K wps
[Epoch 11 Batch 1080/2125] avg loss 0.00216514, throughput 6.02499K wps
[Epoch 11 Batch 1110/2125] avg loss 0.00178658, throughput 6.01444K wps
[Epoch 11 Batch 1140/2125] avg loss 0.00175479, throughput 6.00862K wps
[Epoch 11 Batch 1170/2125] avg loss 0.00185798, throughput 6.00503K wps
[Epoch 11 Batch 1200/2125] avg loss 0.00176919, throughput 6.00138K wps
[Epoch 11 Batch 1230/2125] avg loss 0.00158506, throughput 6.01287K wps
[Epoch 11 Batch 1260/2125] avg loss 0.00170506, throughput 5.99999K wps
[Epoch 11 Batch 1290/2125] avg loss 0.00136324, throughput 6.01975K wps
[Epoch 11 Batch 1320/2125] avg loss 0.00197814, throughput 6.01208K wps
[Epoch 11 Batch 1350/2125] avg loss 0.00186294, throughput 6.01347K wps
[Epoch 11 Batch 1380/2125] avg loss 0.00191181, throughput 6.01992K wps
[Epoch 11 Batch 1410/2125] avg loss 0.00165113, throughput 6.01371K wps
[Epoch 11 Batch 1440/2125] avg loss 0.00153531, throughput 6.01549K wps
[Epoch 11 Batch 1470/2125] avg loss 0.00186237, throughput 6.01142K wps
[Epoch 11 Batch 1500/2125] avg loss 0.00150011, throughput 6.01936K wps
[Epoch 11 Batch 1530/2125] avg loss 0.00175135, throughput 6.00509K wps
[Epoch 11 Batch 1560/2125] avg loss 0.00169371, throughput 5.98422K wps
[Epoch 11 Batch 1590/2125] avg loss 0.00186827, throughput 6.01304K wps
[Epoch 11 Batch 1620/2125] avg loss 0.00183257, throughput 6.01847K wps
[Epoch 11 Batch 1650/2125] avg loss 0.00192661, throughput 6.00869K wps
[Epoch 11 Batch 1680/2125] avg loss 0.00173807, throughput 6.01085K wps
[Epoch 11 Batch 1710/2125] avg loss 0.00168934, throughput 6.02028K wps
[Epoch 11 Batch 1740/2125] avg loss 0.00162175, throughput 6.02028K wps
[Epoch 11 Batch 1770/2125] avg loss 0.00188944, throughput 6.02316K wps
[Epoch 11 Batch 1800/2125] avg loss 0.00178978, throughput 6.0203K wps
[Epoch 11 Batch 1830/2125] avg loss 0.00177428, throughput 6.02234K wps
[Epoch 11 Batch 1860/2125] avg loss 0.00144609, throughput 6.00569K wps
[Epoch 11 Batch 1890/2125] avg loss 0.00160884, throughput 6.01805K wps
[Epoch 11 Batch 1920/2125] avg loss 0.00163485, throughput 6.02521K wps
[Epoch 11 Batch 1950/2125] avg loss 0.00180822, throughput 6.01919K wps
[Epoch 11 Batch 1980/2125] avg loss 0.00183987, throughput 6.02414K wps
[Epoch 11 Batch 2010/2125] avg loss 0.00162786, throughput 6.01507K wps
[Epoch 11 Batch 2040/2125] avg loss 0.00148403, throughput 6.02096K wps
[Epoch 11 Batch 2070/2125] avg loss 0.00180243, throughput 6.0122K wps
[Epoch 11 Batch 2100/2125] avg loss 0.0019833, throughput 6.01985K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 11] train avg loss 0.001649, test acc 0.9262, test avg loss 0.297033, throughput 6.01817K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 12 Batch 30/2125] avg loss 0.00135836, throughput 6.15529K wps
[Epoch 12 Batch 60/2125] avg loss 0.00120175, throughput 6.01546K wps
[Epoch 12 Batch 90/2125] avg loss 0.00138461, throughput 6.01946K wps
[Epoch 12 Batch 120/2125] avg loss 0.0014593, throughput 6.0239K wps
[Epoch 12 Batch 150/2125] avg loss 0.00155391, throughput 6.01293K wps
[Epoch 12 Batch 180/2125] avg loss 0.00115072, throughput 6.01377K wps
[Epoch 12 Batch 210/2125] avg loss 0.00152083, throughput 6.02769K wps
[Epoch 12 Batch 240/2125] avg loss 0.0014112, throughput 6.0114K wps
[Epoch 12 Batch 270/2125] avg loss 0.00125972, throughput 6.01969K wps
[Epoch 12 Batch 300/2125] avg loss 0.00152881, throughput 6.01443K wps
[Epoch 12 Batch 330/2125] avg loss 0.00157118, throughput 6.01523K wps
[Epoch 12 Batch 360/2125] avg loss 0.00148575, throughput 6.00734K wps
[Epoch 12 Batch 390/2125] avg loss 0.00180032, throughput 6.00543K wps
[Epoch 12 Batch 420/2125] avg loss 0.00131008, throughput 6.01787K wps
[Epoch 12 Batch 450/2125] avg loss 0.00163455, throughput 6.01237K wps
[Epoch 12 Batch 480/2125] avg loss 0.00142842, throughput 6.00081K wps
[Epoch 12 Batch 510/2125] avg loss 0.00150354, throughput 6.01129K wps
[Epoch 12 Batch 540/2125] avg loss 0.00131576, throughput 6.01423K wps
[Epoch 12 Batch 570/2125] avg loss 0.00131994, throughput 6.00817K wps
[Epoch 12 Batch 600/2125] avg loss 0.00152034, throughput 6.01881K wps
[Epoch 12 Batch 630/2125] avg loss 0.00155987, throughput 6.02034K wps
[Epoch 12 Batch 660/2125] avg loss 0.00135031, throughput 6.02051K wps
[Epoch 12 Batch 690/2125] avg loss 0.00143739, throughput 6.01154K wps
[Epoch 12 Batch 720/2125] avg loss 0.00126977, throughput 6.0115K wps
[Epoch 12 Batch 750/2125] avg loss 0.00129467, throughput 6.02132K wps
[Epoch 12 Batch 780/2125] avg loss 0.00120783, throughput 6.01028K wps
[Epoch 12 Batch 810/2125] avg loss 0.0015139, throughput 6.01049K wps
[Epoch 12 Batch 840/2125] avg loss 0.00134214, throughput 6.01455K wps
[Epoch 12 Batch 870/2125] avg loss 0.0016193, throughput 6.01592K wps
[Epoch 12 Batch 900/2125] avg loss 0.00156318, throughput 6.01263K wps
[Epoch 12 Batch 930/2125] avg loss 0.00130542, throughput 6.00736K wps
[Epoch 12 Batch 960/2125] avg loss 0.00164973, throughput 6.01624K wps
[Epoch 12 Batch 990/2125] avg loss 0.00155723, throughput 6.01077K wps
[Epoch 12 Batch 1020/2125] avg loss 0.00141351, throughput 6.00774K wps
[Epoch 12 Batch 1050/2125] avg loss 0.00137504, throughput 6.00357K wps
[Epoch 12 Batch 1080/2125] avg loss 0.00207983, throughput 6.01169K wps
[Epoch 12 Batch 1110/2125] avg loss 0.00173751, throughput 6.01937K wps
[Epoch 12 Batch 1140/2125] avg loss 0.00185003, throughput 6.01516K wps
[Epoch 12 Batch 1170/2125] avg loss 0.00167818, throughput 6.01609K wps
[Epoch 12 Batch 1200/2125] avg loss 0.00172262, throughput 6.02098K wps
[Epoch 12 Batch 1230/2125] avg loss 0.00135962, throughput 6.01654K wps
[Epoch 12 Batch 1260/2125] avg loss 0.00155797, throughput 6.02526K wps
[Epoch 12 Batch 1290/2125] avg loss 0.00175506, throughput 6.007K wps
[Epoch 12 Batch 1320/2125] avg loss 0.00135091, throughput 6.01461K wps
[Epoch 12 Batch 1350/2125] avg loss 0.00166648, throughput 6.01346K wps
[Epoch 12 Batch 1380/2125] avg loss 0.00184981, throughput 6.02169K wps
[Epoch 12 Batch 1410/2125] avg loss 0.00140199, throughput 6.02614K wps
[Epoch 12 Batch 1440/2125] avg loss 0.00188289, throughput 6.01365K wps
[Epoch 12 Batch 1470/2125] avg loss 0.00165532, throughput 6.01657K wps
[Epoch 12 Batch 1500/2125] avg loss 0.00154695, throughput 6.02241K wps
[Epoch 12 Batch 1530/2125] avg loss 0.00162295, throughput 6.0213K wps
[Epoch 12 Batch 1560/2125] avg loss 0.00159939, throughput 6.01468K wps
[Epoch 12 Batch 1590/2125] avg loss 0.00151947, throughput 6.0144K wps
[Epoch 12 Batch 1620/2125] avg loss 0.00149057, throughput 6.01455K wps
[Epoch 12 Batch 1650/2125] avg loss 0.00172545, throughput 6.0296K wps
[Epoch 12 Batch 1680/2125] avg loss 0.00148571, throughput 6.02376K wps
[Epoch 12 Batch 1710/2125] avg loss 0.00167975, throughput 6.01376K wps
[Epoch 12 Batch 1740/2125] avg loss 0.00154928, throughput 6.01785K wps
[Epoch 12 Batch 1770/2125] avg loss 0.00180392, throughput 6.01532K wps
[Epoch 12 Batch 1800/2125] avg loss 0.00171294, throughput 6.01584K wps
[Epoch 12 Batch 1830/2125] avg loss 0.00138392, throughput 6.01737K wps
[Epoch 12 Batch 1860/2125] avg loss 0.00156324, throughput 6.01657K wps
[Epoch 12 Batch 1890/2125] avg loss 0.00143407, throughput 6.01247K wps
[Epoch 12 Batch 1920/2125] avg loss 0.00169738, throughput 6.02549K wps
[Epoch 12 Batch 1950/2125] avg loss 0.00158776, throughput 6.02169K wps
[Epoch 12 Batch 1980/2125] avg loss 0.00190532, throughput 6.02543K wps
[Epoch 12 Batch 2010/2125] avg loss 0.00160872, throughput 6.02056K wps
[Epoch 12 Batch 2040/2125] avg loss 0.00178711, throughput 6.02088K wps
[Epoch 12 Batch 2070/2125] avg loss 0.00176616, throughput 6.0179K wps
[Epoch 12 Batch 2100/2125] avg loss 0.00190201, throughput 6.02677K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 12] train avg loss 0.00154826, test acc 0.9266, test avg loss 0.307455, throughput 6.01817K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 13 Batch 30/2125] avg loss 0.00128801, throughput 6.15553K wps
[Epoch 13 Batch 60/2125] avg loss 0.00168189, throughput 6.01696K wps
[Epoch 13 Batch 90/2125] avg loss 0.00129239, throughput 6.01446K wps
[Epoch 13 Batch 120/2125] avg loss 0.00140122, throughput 6.02104K wps
[Epoch 13 Batch 150/2125] avg loss 0.00153953, throughput 6.02559K wps
[Epoch 13 Batch 180/2125] avg loss 0.00134643, throughput 6.01906K wps
[Epoch 13 Batch 210/2125] avg loss 0.00138116, throughput 5.9984K wps
[Epoch 13 Batch 240/2125] avg loss 0.0013128, throughput 6.01788K wps
[Epoch 13 Batch 270/2125] avg loss 0.00124161, throughput 6.02478K wps
[Epoch 13 Batch 300/2125] avg loss 0.00134867, throughput 6.02352K wps
[Epoch 13 Batch 330/2125] avg loss 0.00137784, throughput 6.02256K wps
[Epoch 13 Batch 360/2125] avg loss 0.00141771, throughput 6.02454K wps
[Epoch 13 Batch 390/2125] avg loss 0.00177789, throughput 6.01811K wps
[Epoch 13 Batch 420/2125] avg loss 0.00129432, throughput 6.01832K wps
[Epoch 13 Batch 450/2125] avg loss 0.00134653, throughput 6.02199K wps
[Epoch 13 Batch 480/2125] avg loss 0.00138692, throughput 6.01751K wps
[Epoch 13 Batch 510/2125] avg loss 0.0015443, throughput 6.02087K wps
[Epoch 13 Batch 540/2125] avg loss 0.00164271, throughput 6.02158K wps
[Epoch 13 Batch 570/2125] avg loss 0.00124664, throughput 6.01577K wps
[Epoch 13 Batch 600/2125] avg loss 0.00114388, throughput 6.01666K wps
[Epoch 13 Batch 630/2125] avg loss 0.00117034, throughput 6.00838K wps
[Epoch 13 Batch 660/2125] avg loss 0.00121591, throughput 6.01689K wps
[Epoch 13 Batch 690/2125] avg loss 0.00130056, throughput 6.02361K wps
[Epoch 13 Batch 720/2125] avg loss 0.00112288, throughput 6.02012K wps
[Epoch 13 Batch 750/2125] avg loss 0.00141603, throughput 6.01446K wps
[Epoch 13 Batch 780/2125] avg loss 0.00131729, throughput 6.0149K wps
[Epoch 13 Batch 810/2125] avg loss 0.00183771, throughput 6.01317K wps
[Epoch 13 Batch 840/2125] avg loss 0.0014071, throughput 6.00573K wps
[Epoch 13 Batch 870/2125] avg loss 0.0015478, throughput 6.01607K wps
[Epoch 13 Batch 900/2125] avg loss 0.00133206, throughput 6.00512K wps
[Epoch 13 Batch 930/2125] avg loss 0.00154598, throughput 6.01841K wps
[Epoch 13 Batch 960/2125] avg loss 0.00146045, throughput 6.0232K wps
[Epoch 13 Batch 990/2125] avg loss 0.00131551, throughput 6.02223K wps
[Epoch 13 Batch 1020/2125] avg loss 0.00122389, throughput 6.01659K wps
[Epoch 13 Batch 1050/2125] avg loss 0.00150081, throughput 6.02026K wps
[Epoch 13 Batch 1080/2125] avg loss 0.00171368, throughput 6.01272K wps
[Epoch 13 Batch 1110/2125] avg loss 0.00138504, throughput 6.00567K wps
[Epoch 13 Batch 1140/2125] avg loss 0.001318, throughput 6.01827K wps
[Epoch 13 Batch 1170/2125] avg loss 0.0011701, throughput 6.02135K wps
[Epoch 13 Batch 1200/2125] avg loss 0.00117667, throughput 6.0263K wps
[Epoch 13 Batch 1230/2125] avg loss 0.00134757, throughput 6.02087K wps
[Epoch 13 Batch 1260/2125] avg loss 0.00133882, throughput 6.02306K wps
[Epoch 13 Batch 1290/2125] avg loss 0.00128412, throughput 6.02357K wps
[Epoch 13 Batch 1320/2125] avg loss 0.00142275, throughput 6.01635K wps
[Epoch 13 Batch 1350/2125] avg loss 0.00148268, throughput 6.01987K wps
[Epoch 13 Batch 1380/2125] avg loss 0.00168731, throughput 6.01954K wps
[Epoch 13 Batch 1410/2125] avg loss 0.00117432, throughput 6.02055K wps
[Epoch 13 Batch 1440/2125] avg loss 0.00161463, throughput 6.01786K wps
[Epoch 13 Batch 1470/2125] avg loss 0.00176441, throughput 6.01642K wps
[Epoch 13 Batch 1500/2125] avg loss 0.00175982, throughput 6.01139K wps
[Epoch 13 Batch 1530/2125] avg loss 0.0014997, throughput 6.0231K wps
[Epoch 13 Batch 1560/2125] avg loss 0.00155769, throughput 6.01527K wps
[Epoch 13 Batch 1590/2125] avg loss 0.00124253, throughput 6.01743K wps
[Epoch 13 Batch 1620/2125] avg loss 0.00156624, throughput 6.0245K wps
[Epoch 13 Batch 1650/2125] avg loss 0.00161178, throughput 6.01618K wps
[Epoch 13 Batch 1680/2125] avg loss 0.0017939, throughput 6.02426K wps
[Epoch 13 Batch 1710/2125] avg loss 0.00146152, throughput 6.02021K wps
[Epoch 13 Batch 1740/2125] avg loss 0.00162905, throughput 6.01632K wps
[Epoch 13 Batch 1770/2125] avg loss 0.00163544, throughput 6.01961K wps
[Epoch 13 Batch 1800/2125] avg loss 0.00149404, throughput 6.02254K wps
[Epoch 13 Batch 1830/2125] avg loss 0.00142675, throughput 6.01966K wps
[Epoch 13 Batch 1860/2125] avg loss 0.00153581, throughput 6.02157K wps
[Epoch 13 Batch 1890/2125] avg loss 0.00145309, throughput 6.02095K wps
[Epoch 13 Batch 1920/2125] avg loss 0.00132817, throughput 6.01794K wps
[Epoch 13 Batch 1950/2125] avg loss 0.00153138, throughput 6.01673K wps
[Epoch 13 Batch 1980/2125] avg loss 0.00134782, throughput 6.02358K wps
[Epoch 13 Batch 2010/2125] avg loss 0.00149999, throughput 6.02612K wps
[Epoch 13 Batch 2040/2125] avg loss 0.00123677, throughput 6.01013K wps
[Epoch 13 Batch 2070/2125] avg loss 0.00130083, throughput 6.01502K wps
[Epoch 13 Batch 2100/2125] avg loss 0.00173719, throughput 6.01564K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 13] train avg loss 0.00143298, test acc 0.9271, test avg loss 0.322869, throughput 6.01993K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 14 Batch 30/2125] avg loss 0.00105788, throughput 6.14324K wps
[Epoch 14 Batch 60/2125] avg loss 0.00110231, throughput 6.01168K wps
[Epoch 14 Batch 90/2125] avg loss 0.00116197, throughput 6.01725K wps
[Epoch 14 Batch 120/2125] avg loss 0.00106596, throughput 6.01501K wps
[Epoch 14 Batch 150/2125] avg loss 0.00136135, throughput 6.03013K wps
[Epoch 14 Batch 180/2125] avg loss 0.00106975, throughput 6.02169K wps
[Epoch 14 Batch 210/2125] avg loss 0.00137637, throughput 6.01729K wps
[Epoch 14 Batch 240/2125] avg loss 0.000968779, throughput 6.01322K wps
[Epoch 14 Batch 270/2125] avg loss 0.00133395, throughput 6.01528K wps
[Epoch 14 Batch 300/2125] avg loss 0.00158395, throughput 6.01671K wps
[Epoch 14 Batch 330/2125] avg loss 0.00130716, throughput 6.02582K wps
[Epoch 14 Batch 360/2125] avg loss 0.00104071, throughput 5.98804K wps
[Epoch 14 Batch 390/2125] avg loss 0.00115179, throughput 5.99609K wps
[Epoch 14 Batch 420/2125] avg loss 0.00105869, throughput 6.03073K wps
[Epoch 14 Batch 450/2125] avg loss 0.00124814, throughput 6.02067K wps
[Epoch 14 Batch 480/2125] avg loss 0.00122411, throughput 6.02176K wps
[Epoch 14 Batch 510/2125] avg loss 0.00140525, throughput 6.01582K wps
[Epoch 14 Batch 540/2125] avg loss 0.0012576, throughput 6.01677K wps
[Epoch 14 Batch 570/2125] avg loss 0.00121425, throughput 6.02661K wps
[Epoch 14 Batch 600/2125] avg loss 0.00139294, throughput 6.0298K wps
[Epoch 14 Batch 630/2125] avg loss 0.00143343, throughput 6.01555K wps
[Epoch 14 Batch 660/2125] avg loss 0.00138246, throughput 6.01019K wps
[Epoch 14 Batch 690/2125] avg loss 0.00139832, throughput 6.01243K wps
[Epoch 14 Batch 720/2125] avg loss 0.00124792, throughput 6.02473K wps
[Epoch 14 Batch 750/2125] avg loss 0.00111424, throughput 6.01307K wps
[Epoch 14 Batch 780/2125] avg loss 0.00163054, throughput 6.01329K wps
[Epoch 14 Batch 810/2125] avg loss 0.00140883, throughput 6.01644K wps
[Epoch 14 Batch 840/2125] avg loss 0.00118788, throughput 6.00899K wps
[Epoch 14 Batch 870/2125] avg loss 0.00127881, throughput 6.01501K wps
[Epoch 14 Batch 900/2125] avg loss 0.00128197, throughput 6.02039K wps
[Epoch 14 Batch 930/2125] avg loss 0.00117978, throughput 6.01572K wps
[Epoch 14 Batch 960/2125] avg loss 0.00118249, throughput 6.00749K wps
[Epoch 14 Batch 990/2125] avg loss 0.00137611, throughput 6.00237K wps
[Epoch 14 Batch 1020/2125] avg loss 0.00147356, throughput 6.01673K wps
[Epoch 14 Batch 1050/2125] avg loss 0.00115542, throughput 6.00998K wps
[Epoch 14 Batch 1080/2125] avg loss 0.0010051, throughput 6.01631K wps
[Epoch 14 Batch 1110/2125] avg loss 0.00140924, throughput 6.01081K wps
[Epoch 14 Batch 1140/2125] avg loss 0.00131716, throughput 6.00809K wps
[Epoch 14 Batch 1170/2125] avg loss 0.00153626, throughput 6.01221K wps
[Epoch 14 Batch 1200/2125] avg loss 0.00108101, throughput 6.01124K wps
[Epoch 14 Batch 1230/2125] avg loss 0.00133403, throughput 6.01229K wps
[Epoch 14 Batch 1260/2125] avg loss 0.00122994, throughput 6.0102K wps
[Epoch 14 Batch 1290/2125] avg loss 0.00127956, throughput 6.02097K wps
[Epoch 14 Batch 1320/2125] avg loss 0.00167607, throughput 6.01859K wps
[Epoch 14 Batch 1350/2125] avg loss 0.00133129, throughput 6.01306K wps
[Epoch 14 Batch 1380/2125] avg loss 0.00104568, throughput 6.01711K wps
[Epoch 14 Batch 1410/2125] avg loss 0.00132797, throughput 6.02124K wps
[Epoch 14 Batch 1440/2125] avg loss 0.0012349, throughput 6.02136K wps
[Epoch 14 Batch 1470/2125] avg loss 0.00155973, throughput 6.01572K wps
[Epoch 14 Batch 1500/2125] avg loss 0.00133455, throughput 6.01844K wps
[Epoch 14 Batch 1530/2125] avg loss 0.00176581, throughput 6.01022K wps
[Epoch 14 Batch 1560/2125] avg loss 0.00143533, throughput 6.02154K wps
[Epoch 14 Batch 1590/2125] avg loss 0.00131201, throughput 6.02399K wps
[Epoch 14 Batch 1620/2125] avg loss 0.00156912, throughput 6.02675K wps
[Epoch 14 Batch 1650/2125] avg loss 0.00147671, throughput 6.02218K wps
[Epoch 14 Batch 1680/2125] avg loss 0.00137821, throughput 6.01311K wps
[Epoch 14 Batch 1710/2125] avg loss 0.00133733, throughput 6.02557K wps
[Epoch 14 Batch 1740/2125] avg loss 0.00163633, throughput 6.01314K wps
[Epoch 14 Batch 1770/2125] avg loss 0.00144688, throughput 6.01863K wps
[Epoch 14 Batch 1800/2125] avg loss 0.001308, throughput 6.01543K wps
[Epoch 14 Batch 1830/2125] avg loss 0.00152428, throughput 6.01206K wps
[Epoch 14 Batch 1860/2125] avg loss 0.00155492, throughput 6.01688K wps
[Epoch 14 Batch 1890/2125] avg loss 0.00141973, throughput 6.01026K wps
[Epoch 14 Batch 1920/2125] avg loss 0.00155178, throughput 6.02049K wps
[Epoch 14 Batch 1950/2125] avg loss 0.00138436, throughput 6.01114K wps
[Epoch 14 Batch 1980/2125] avg loss 0.00158883, throughput 6.02309K wps
[Epoch 14 Batch 2010/2125] avg loss 0.00121132, throughput 6.00983K wps
[Epoch 14 Batch 2040/2125] avg loss 0.00166829, throughput 6.01494K wps
[Epoch 14 Batch 2070/2125] avg loss 0.00148368, throughput 6.00556K wps
[Epoch 14 Batch 2100/2125] avg loss 0.00196371, throughput 6.01749K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 14] train avg loss 0.00134307, test acc 0.9277, test avg loss 0.326684, throughput 6.01755K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 15 Batch 30/2125] avg loss 0.00139543, throughput 6.15557K wps
[Epoch 15 Batch 60/2125] avg loss 0.00108212, throughput 6.02454K wps
[Epoch 15 Batch 90/2125] avg loss 0.00105397, throughput 6.0202K wps
[Epoch 15 Batch 120/2125] avg loss 0.00121987, throughput 6.01633K wps
[Epoch 15 Batch 150/2125] avg loss 0.00107644, throughput 6.02491K wps
[Epoch 15 Batch 180/2125] avg loss 0.00101311, throughput 6.02074K wps
[Epoch 15 Batch 210/2125] avg loss 0.00120645, throughput 6.01981K wps
[Epoch 15 Batch 240/2125] avg loss 0.00104356, throughput 6.01162K wps
[Epoch 15 Batch 270/2125] avg loss 0.00121972, throughput 6.02672K wps
[Epoch 15 Batch 300/2125] avg loss 0.0012096, throughput 6.02178K wps
[Epoch 15 Batch 330/2125] avg loss 0.00109145, throughput 6.01523K wps
[Epoch 15 Batch 360/2125] avg loss 0.00105086, throughput 6.0257K wps
[Epoch 15 Batch 390/2125] avg loss 0.00116527, throughput 6.01177K wps
[Epoch 15 Batch 420/2125] avg loss 0.00125013, throughput 6.02269K wps
[Epoch 15 Batch 450/2125] avg loss 0.0011548, throughput 6.01184K wps
[Epoch 15 Batch 480/2125] avg loss 0.00138427, throughput 6.01061K wps
[Epoch 15 Batch 510/2125] avg loss 0.0013103, throughput 6.0189K wps
[Epoch 15 Batch 540/2125] avg loss 0.00111735, throughput 6.01372K wps
[Epoch 15 Batch 570/2125] avg loss 0.00121926, throughput 6.01614K wps
[Epoch 15 Batch 600/2125] avg loss 0.00145133, throughput 6.01052K wps
[Epoch 15 Batch 630/2125] avg loss 0.00136436, throughput 6.01953K wps
[Epoch 15 Batch 660/2125] avg loss 0.00122863, throughput 6.01999K wps
[Epoch 15 Batch 690/2125] avg loss 0.00122896, throughput 6.0231K wps
[Epoch 15 Batch 720/2125] avg loss 0.00109679, throughput 6.02065K wps
[Epoch 15 Batch 750/2125] avg loss 0.00105243, throughput 6.01739K wps
[Epoch 15 Batch 780/2125] avg loss 0.00119097, throughput 6.02536K wps
[Epoch 15 Batch 810/2125] avg loss 0.00142336, throughput 6.02183K wps
[Epoch 15 Batch 840/2125] avg loss 0.00142099, throughput 6.01971K wps
[Epoch 15 Batch 870/2125] avg loss 0.00144078, throughput 6.01482K wps
[Epoch 15 Batch 900/2125] avg loss 0.00141252, throughput 6.01523K wps
[Epoch 15 Batch 930/2125] avg loss 0.00115066, throughput 6.01047K wps
[Epoch 15 Batch 960/2125] avg loss 0.00113436, throughput 6.01694K wps
[Epoch 15 Batch 990/2125] avg loss 0.00121192, throughput 6.01546K wps
[Epoch 15 Batch 1020/2125] avg loss 0.00112482, throughput 6.02401K wps
[Epoch 15 Batch 1050/2125] avg loss 0.00139836, throughput 6.01055K wps
[Epoch 15 Batch 1080/2125] avg loss 0.00116062, throughput 6.00418K wps
[Epoch 15 Batch 1110/2125] avg loss 0.00124224, throughput 6.01116K wps
[Epoch 15 Batch 1140/2125] avg loss 0.00130143, throughput 6.01312K wps
[Epoch 15 Batch 1170/2125] avg loss 0.00109039, throughput 6.0072K wps
[Epoch 15 Batch 1200/2125] avg loss 0.00137149, throughput 6.02695K wps
[Epoch 15 Batch 1230/2125] avg loss 0.0013831, throughput 6.01274K wps
[Epoch 15 Batch 1260/2125] avg loss 0.00144567, throughput 6.02628K wps
[Epoch 15 Batch 1290/2125] avg loss 0.00128547, throughput 6.00156K wps
[Epoch 15 Batch 1320/2125] avg loss 0.00136907, throughput 6.01867K wps
[Epoch 15 Batch 1350/2125] avg loss 0.00108266, throughput 6.02277K wps
[Epoch 15 Batch 1380/2125] avg loss 0.00144198, throughput 6.02021K wps
[Epoch 15 Batch 1410/2125] avg loss 0.00108195, throughput 6.02061K wps
[Epoch 15 Batch 1440/2125] avg loss 0.00108859, throughput 6.02156K wps
[Epoch 15 Batch 1470/2125] avg loss 0.00125979, throughput 6.01924K wps
[Epoch 15 Batch 1500/2125] avg loss 0.00145254, throughput 6.02009K wps
[Epoch 15 Batch 1530/2125] avg loss 0.00124668, throughput 6.01554K wps
[Epoch 15 Batch 1560/2125] avg loss 0.00152923, throughput 6.02254K wps
[Epoch 15 Batch 1590/2125] avg loss 0.00144515, throughput 6.01806K wps
[Epoch 15 Batch 1620/2125] avg loss 0.00146725, throughput 6.02363K wps
[Epoch 15 Batch 1650/2125] avg loss 0.00135033, throughput 6.02609K wps
[Epoch 15 Batch 1680/2125] avg loss 0.00136663, throughput 6.02002K wps
[Epoch 15 Batch 1710/2125] avg loss 0.00149468, throughput 6.02363K wps
[Epoch 15 Batch 1740/2125] avg loss 0.00100883, throughput 6.02839K wps
[Epoch 15 Batch 1770/2125] avg loss 0.00119081, throughput 6.01717K wps
[Epoch 15 Batch 1800/2125] avg loss 0.00106411, throughput 6.0176K wps
[Epoch 15 Batch 1830/2125] avg loss 0.00129282, throughput 6.01457K wps
[Epoch 15 Batch 1860/2125] avg loss 0.00153609, throughput 6.02193K wps
[Epoch 15 Batch 1890/2125] avg loss 0.00122608, throughput 6.01631K wps
[Epoch 15 Batch 1920/2125] avg loss 0.00165437, throughput 6.01889K wps
[Epoch 15 Batch 1950/2125] avg loss 0.00145851, throughput 6.03063K wps
[Epoch 15 Batch 1980/2125] avg loss 0.00134381, throughput 6.01668K wps
[Epoch 15 Batch 2010/2125] avg loss 0.00150068, throughput 6.02066K wps
[Epoch 15 Batch 2040/2125] avg loss 0.00101282, throughput 6.02369K wps
[Epoch 15 Batch 2070/2125] avg loss 0.00146956, throughput 6.0222K wps
[Epoch 15 Batch 2100/2125] avg loss 0.00149269, throughput 6.02032K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 15] train avg loss 0.00127184, test acc 0.9272, test avg loss 0.34703, throughput 6.0205K wps
[Epoch 16 Batch 30/2125] avg loss 0.00109304, throughput 6.14094K wps
[Epoch 16 Batch 60/2125] avg loss 0.00101887, throughput 6.01325K wps
[Epoch 16 Batch 90/2125] avg loss 0.0010207, throughput 6.01692K wps
[Epoch 16 Batch 120/2125] avg loss 0.00105435, throughput 6.00852K wps
[Epoch 16 Batch 150/2125] avg loss 0.001306, throughput 6.00227K wps
[Epoch 16 Batch 180/2125] avg loss 0.000984744, throughput 6.00999K wps
[Epoch 16 Batch 210/2125] avg loss 0.00101001, throughput 6.00006K wps
[Epoch 16 Batch 240/2125] avg loss 0.00111195, throughput 6.01363K wps
[Epoch 16 Batch 270/2125] avg loss 0.00113135, throughput 6.02236K wps
[Epoch 16 Batch 300/2125] avg loss 0.00111818, throughput 6.01929K wps
[Epoch 16 Batch 330/2125] avg loss 0.00128709, throughput 6.0164K wps
[Epoch 16 Batch 360/2125] avg loss 0.00122084, throughput 6.0202K wps
[Epoch 16 Batch 390/2125] avg loss 0.00122089, throughput 6.01859K wps
[Epoch 16 Batch 420/2125] avg loss 0.000810615, throughput 6.0259K wps
[Epoch 16 Batch 450/2125] avg loss 0.00118425, throughput 6.02011K wps
[Epoch 16 Batch 480/2125] avg loss 0.00106158, throughput 6.02579K wps
[Epoch 16 Batch 510/2125] avg loss 0.00095936, throughput 6.02046K wps
[Epoch 16 Batch 540/2125] avg loss 0.000939449, throughput 6.02328K wps
[Epoch 16 Batch 570/2125] avg loss 0.00101176, throughput 6.02239K wps
[Epoch 16 Batch 600/2125] avg loss 0.0013104, throughput 6.01222K wps
[Epoch 16 Batch 630/2125] avg loss 0.000971847, throughput 6.01555K wps
[Epoch 16 Batch 660/2125] avg loss 0.0007907, throughput 6.01345K wps
[Epoch 16 Batch 690/2125] avg loss 0.00112267, throughput 6.01573K wps
[Epoch 16 Batch 720/2125] avg loss 0.00117494, throughput 6.02197K wps
[Epoch 16 Batch 750/2125] avg loss 0.000936253, throughput 6.0281K wps
[Epoch 16 Batch 780/2125] avg loss 0.000987352, throughput 6.02388K wps
[Epoch 16 Batch 810/2125] avg loss 0.0014418, throughput 6.02555K wps
[Epoch 16 Batch 840/2125] avg loss 0.0012049, throughput 6.01543K wps
[Epoch 16 Batch 870/2125] avg loss 0.00100001, throughput 6.02254K wps
[Epoch 16 Batch 900/2125] avg loss 0.00110334, throughput 6.01799K wps
[Epoch 16 Batch 930/2125] avg loss 0.00125007, throughput 6.02418K wps
[Epoch 16 Batch 960/2125] avg loss 0.00098449, throughput 6.01928K wps
[Epoch 16 Batch 990/2125] avg loss 0.00159226, throughput 6.01804K wps
[Epoch 16 Batch 1020/2125] avg loss 0.0010858, throughput 6.02859K wps
[Epoch 16 Batch 1050/2125] avg loss 0.00163871, throughput 6.02847K wps
[Epoch 16 Batch 1080/2125] avg loss 0.00133311, throughput 6.02073K wps
[Epoch 16 Batch 1110/2125] avg loss 0.00124191, throughput 6.01872K wps
[Epoch 16 Batch 1140/2125] avg loss 0.00128125, throughput 6.02267K wps
[Epoch 16 Batch 1170/2125] avg loss 0.00123693, throughput 6.01338K wps
[Epoch 16 Batch 1200/2125] avg loss 0.00101443, throughput 6.02471K wps
[Epoch 16 Batch 1230/2125] avg loss 0.00123647, throughput 6.02045K wps
[Epoch 16 Batch 1260/2125] avg loss 0.00094031, throughput 6.02021K wps
[Epoch 16 Batch 1290/2125] avg loss 0.00108316, throughput 6.01567K wps
[Epoch 16 Batch 1320/2125] avg loss 0.000910961, throughput 6.0037K wps
[Epoch 16 Batch 1350/2125] avg loss 0.00141035, throughput 6.0198K wps
[Epoch 16 Batch 1380/2125] avg loss 0.00132467, throughput 6.01615K wps
[Epoch 16 Batch 1410/2125] avg loss 0.00136275, throughput 5.99416K wps
[Epoch 16 Batch 1440/2125] avg loss 0.0011498, throughput 6.00495K wps
[Epoch 16 Batch 1470/2125] avg loss 0.00114509, throughput 6.02677K wps
[Epoch 16 Batch 1500/2125] avg loss 0.00117565, throughput 6.01055K wps
[Epoch 16 Batch 1530/2125] avg loss 0.00107754, throughput 6.02191K wps
[Epoch 16 Batch 1560/2125] avg loss 0.00116241, throughput 6.01117K wps
[Epoch 16 Batch 1590/2125] avg loss 0.00130518, throughput 6.01412K wps
[Epoch 16 Batch 1620/2125] avg loss 0.00165557, throughput 6.03367K wps
[Epoch 16 Batch 1650/2125] avg loss 0.00145014, throughput 6.01976K wps
[Epoch 16 Batch 1680/2125] avg loss 0.00152163, throughput 6.02165K wps
[Epoch 16 Batch 1710/2125] avg loss 0.0011961, throughput 6.03305K wps
[Epoch 16 Batch 1740/2125] avg loss 0.00136172, throughput 6.02285K wps
[Epoch 16 Batch 1770/2125] avg loss 0.00137843, throughput 6.02293K wps
[Epoch 16 Batch 1800/2125] avg loss 0.00129435, throughput 6.02573K wps
[Epoch 16 Batch 1830/2125] avg loss 0.00113423, throughput 6.02066K wps
[Epoch 16 Batch 1860/2125] avg loss 0.00137744, throughput 6.01463K wps
[Epoch 16 Batch 1890/2125] avg loss 0.00125024, throughput 6.01393K wps
[Epoch 16 Batch 1920/2125] avg loss 0.00133241, throughput 6.01036K wps
[Epoch 16 Batch 1950/2125] avg loss 0.00153178, throughput 6.01638K wps
[Epoch 16 Batch 1980/2125] avg loss 0.00141413, throughput 6.02418K wps
[Epoch 16 Batch 2010/2125] avg loss 0.00123913, throughput 6.01853K wps
[Epoch 16 Batch 2040/2125] avg loss 0.00104419, throughput 6.0157K wps
[Epoch 16 Batch 2070/2125] avg loss 0.00123038, throughput 6.01485K wps
[Epoch 16 Batch 2100/2125] avg loss 0.00123554, throughput 6.01079K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 16] train avg loss 0.00118697, test acc 0.9285, test avg loss 0.355477, throughput 6.01975K wps
Observed Improvement.
Begin Testing...
[Batch 30/35] elapsed 0.27 s
[Epoch 17 Batch 30/2125] avg loss 0.00117438, throughput 6.1511K wps
[Epoch 17 Batch 60/2125] avg loss 0.00104685, throughput 6.01309K wps
[Epoch 17 Batch 90/2125] avg loss 0.00103128, throughput 6.0094K wps
[Epoch 17 Batch 120/2125] avg loss 0.000883869, throughput 6.00953K wps
[Epoch 17 Batch 150/2125] avg loss 0.000804002, throughput 6.01728K wps
[Epoch 17 Batch 180/2125] avg loss 0.00110635, throughput 6.02031K wps
[Epoch 17 Batch 210/2125] avg loss 0.000838765, throughput 6.0155K wps
[Epoch 17 Batch 240/2125] avg loss 0.0010489, throughput 6.0155K wps
[Epoch 17 Batch 270/2125] avg loss 0.00116714, throughput 6.01461K wps
[Epoch 17 Batch 300/2125] avg loss 0.00104019, throughput 6.01314K wps
[Epoch 17 Batch 330/2125] avg loss 0.00106901, throughput 6.01193K wps
[Epoch 17 Batch 360/2125] avg loss 0.000920718, throughput 6.01811K wps
[Epoch 17 Batch 390/2125] avg loss 0.000840775, throughput 6.02748K wps
[Epoch 17 Batch 420/2125] avg loss 0.000988947, throughput 6.02998K wps
[Epoch 17 Batch 450/2125] avg loss 0.00084346, throughput 6.03629K wps
[Epoch 17 Batch 480/2125] avg loss 0.00113987, throughput 6.01756K wps
[Epoch 17 Batch 510/2125] avg loss 0.00121444, throughput 6.01327K wps
[Epoch 17 Batch 540/2125] avg loss 0.000926791, throughput 6.01717K wps
[Epoch 17 Batch 570/2125] avg loss 0.00105735, throughput 6.01083K wps
[Epoch 17 Batch 600/2125] avg loss 0.00098716, throughput 6.01631K wps
[Epoch 17 Batch 630/2125] avg loss 0.00128537, throughput 6.01761K wps
[Epoch 17 Batch 660/2125] avg loss 0.00120699, throughput 6.01375K wps
[Epoch 17 Batch 690/2125] avg loss 0.000986114, throughput 6.0222K wps
[Epoch 17 Batch 720/2125] avg loss 0.00110155, throughput 6.01695K wps
[Epoch 17 Batch 750/2125] avg loss 0.00113507, throughput 6.00476K wps
[Epoch 17 Batch 780/2125] avg loss 0.000894686, throughput 6.016K wps
[Epoch 17 Batch 810/2125] avg loss 0.000881319, throughput 6.00985K wps
[Epoch 17 Batch 840/2125] avg loss 0.00108711, throughput 6.01607K wps
[Epoch 17 Batch 870/2125] avg loss 0.00108492, throughput 6.01325K wps
[Epoch 17 Batch 900/2125] avg loss 0.00126494, throughput 6.01945K wps
[Epoch 17 Batch 930/2125] avg loss 0.0011265, throughput 6.01019K wps
[Epoch 17 Batch 960/2125] avg loss 0.00121778, throughput 6.0129K wps
[Epoch 17 Batch 990/2125] avg loss 0.0013408, throughput 6.01432K wps
[Epoch 17 Batch 1020/2125] avg loss 0.00122753, throughput 6.0239K wps
[Epoch 17 Batch 1050/2125] avg loss 0.00126528, throughput 6.01135K wps
[Epoch 17 Batch 1080/2125] avg loss 0.0013045, throughput 6.02739K wps
[Epoch 17 Batch 1110/2125] avg loss 0.00128109, throughput 6.02776K wps
[Epoch 17 Batch 1140/2125] avg loss 0.00102235, throughput 6.00619K wps
[Epoch 17 Batch 1170/2125] avg loss 0.00110282, throughput 6.01571K wps
[Epoch 17 Batch 1200/2125] avg loss 0.00118616, throughput 6.02568K wps
[Epoch 17 Batch 1230/2125] avg loss 0.00142121, throughput 6.01922K wps
[Epoch 17 Batch 1260/2125] avg loss 0.00110921, throughput 6.02399K wps
[Epoch 17 Batch 1290/2125] avg loss 0.000899998, throughput 6.02917K wps
[Epoch 17 Batch 1320/2125] avg loss 0.00100246, throughput 6.02727K wps
[Epoch 17 Batch 1350/2125] avg loss 0.000991138, throughput 6.0193K wps
[Epoch 17 Batch 1380/2125] avg loss 0.0010534, throughput 6.02476K wps
[Epoch 17 Batch 1410/2125] avg loss 0.00121981, throughput 6.01382K wps
[Epoch 17 Batch 1440/2125] avg loss 0.000957027, throughput 6.02428K wps
[Epoch 17 Batch 1470/2125] avg loss 0.00100467, throughput 6.02741K wps
[Epoch 17 Batch 1500/2125] avg loss 0.00159456, throughput 6.02795K wps
[Epoch 17 Batch 1530/2125] avg loss 0.00141238, throughput 6.01866K wps
[Epoch 17 Batch 1560/2125] avg loss 0.00109313, throughput 6.02479K wps
[Epoch 17 Batch 1590/2125] avg loss 0.001352, throughput 6.01428K wps
[Epoch 17 Batch 1620/2125] avg loss 0.0013158, throughput 6.01387K wps
[Epoch 17 Batch 1650/2125] avg loss 0.00120518, throughput 6.01128K wps
[Epoch 17 Batch 1680/2125] avg loss 0.001375, throughput 6.0112K wps
[Epoch 17 Batch 1710/2125] avg loss 0.00114793, throughput 6.00557K wps
[Epoch 17 Batch 1740/2125] avg loss 0.00129936, throughput 6.01116K wps
[Epoch 17 Batch 1770/2125] avg loss 0.00107526, throughput 6.01132K wps
[Epoch 17 Batch 1800/2125] avg loss 0.00149049, throughput 6.01398K wps
[Epoch 17 Batch 1830/2125] avg loss 0.00114994, throughput 6.01399K wps
[Epoch 17 Batch 1860/2125] avg loss 0.00133185, throughput 6.01437K wps
[Epoch 17 Batch 1890/2125] avg loss 0.00152777, throughput 6.0113K wps
[Epoch 17 Batch 1920/2125] avg loss 0.00134692, throughput 6.01186K wps
[Epoch 17 Batch 1950/2125] avg loss 0.00146533, throughput 6.0159K wps
[Epoch 17 Batch 1980/2125] avg loss 0.00118309, throughput 6.02111K wps
[Epoch 17 Batch 2010/2125] avg loss 0.00108206, throughput 6.00513K wps
[Epoch 17 Batch 2040/2125] avg loss 0.000997473, throughput 6.0152K wps
[Epoch 17 Batch 2070/2125] avg loss 0.00129077, throughput 6.02237K wps
[Epoch 17 Batch 2100/2125] avg loss 0.00136845, throughput 6.01773K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 17] train avg loss 0.00114284, test acc 0.9273, test avg loss 0.361693, throughput 6.0191K wps
[Epoch 18 Batch 30/2125] avg loss 0.00121641, throughput 6.15438K wps
[Epoch 18 Batch 60/2125] avg loss 0.0011193, throughput 6.01678K wps
[Epoch 18 Batch 90/2125] avg loss 0.000771981, throughput 6.01278K wps
[Epoch 18 Batch 120/2125] avg loss 0.00108096, throughput 6.01364K wps
[Epoch 18 Batch 150/2125] avg loss 0.000832984, throughput 6.01563K wps
[Epoch 18 Batch 180/2125] avg loss 0.000632041, throughput 6.02188K wps
[Epoch 18 Batch 210/2125] avg loss 0.000619666, throughput 6.02411K wps
[Epoch 18 Batch 240/2125] avg loss 0.000751692, throughput 6.01004K wps
[Epoch 18 Batch 270/2125] avg loss 0.00121601, throughput 6.02271K wps
[Epoch 18 Batch 300/2125] avg loss 0.000846375, throughput 6.02179K wps
[Epoch 18 Batch 330/2125] avg loss 0.000947245, throughput 6.02063K wps
[Epoch 18 Batch 360/2125] avg loss 0.00111381, throughput 6.01382K wps
[Epoch 18 Batch 390/2125] avg loss 0.00105886, throughput 6.02173K wps
[Epoch 18 Batch 420/2125] avg loss 0.00107209, throughput 6.01092K wps
[Epoch 18 Batch 450/2125] avg loss 0.00110599, throughput 6.02052K wps
[Epoch 18 Batch 480/2125] avg loss 0.000933346, throughput 6.02053K wps
[Epoch 18 Batch 510/2125] avg loss 0.00103028, throughput 6.01692K wps
[Epoch 18 Batch 540/2125] avg loss 0.00101708, throughput 6.02444K wps
[Epoch 18 Batch 570/2125] avg loss 0.000778076, throughput 6.01595K wps
[Epoch 18 Batch 600/2125] avg loss 0.00103866, throughput 6.01874K wps
[Epoch 18 Batch 630/2125] avg loss 0.000885341, throughput 6.02788K wps
[Epoch 18 Batch 660/2125] avg loss 0.000993672, throughput 6.01442K wps
[Epoch 18 Batch 690/2125] avg loss 0.00101195, throughput 6.01453K wps
[Epoch 18 Batch 720/2125] avg loss 0.00118302, throughput 6.01501K wps
[Epoch 18 Batch 750/2125] avg loss 0.00105879, throughput 6.02406K wps
[Epoch 18 Batch 780/2125] avg loss 0.00125803, throughput 6.02645K wps
[Epoch 18 Batch 810/2125] avg loss 0.00122268, throughput 6.01883K wps
[Epoch 18 Batch 840/2125] avg loss 0.00119952, throughput 6.01161K wps
[Epoch 18 Batch 870/2125] avg loss 0.00112216, throughput 6.01757K wps
[Epoch 18 Batch 900/2125] avg loss 0.00137437, throughput 6.02784K wps
[Epoch 18 Batch 930/2125] avg loss 0.00109045, throughput 6.02384K wps
[Epoch 18 Batch 960/2125] avg loss 0.00106619, throughput 6.01371K wps
[Epoch 18 Batch 990/2125] avg loss 0.00104749, throughput 6.01823K wps
[Epoch 18 Batch 1020/2125] avg loss 0.00122606, throughput 6.02848K wps
[Epoch 18 Batch 1050/2125] avg loss 0.000983595, throughput 6.03297K wps
[Epoch 18 Batch 1080/2125] avg loss 0.000923525, throughput 6.02235K wps
[Epoch 18 Batch 1110/2125] avg loss 0.000977297, throughput 6.01609K wps
[Epoch 18 Batch 1140/2125] avg loss 0.00122825, throughput 6.02583K wps
[Epoch 18 Batch 1170/2125] avg loss 0.00107073, throughput 6.02134K wps
[Epoch 18 Batch 1200/2125] avg loss 0.00118714, throughput 6.00729K wps
[Epoch 18 Batch 1230/2125] avg loss 0.00113762, throughput 6.00586K wps
[Epoch 18 Batch 1260/2125] avg loss 0.00124474, throughput 6.01508K wps
[Epoch 18 Batch 1290/2125] avg loss 0.00111304, throughput 6.0253K wps
[Epoch 18 Batch 1320/2125] avg loss 0.000980166, throughput 6.01463K wps
[Epoch 18 Batch 1350/2125] avg loss 0.00120758, throughput 6.01239K wps
[Epoch 18 Batch 1380/2125] avg loss 0.00129161, throughput 6.01418K wps
[Epoch 18 Batch 1410/2125] avg loss 0.00113604, throughput 6.01661K wps
[Epoch 18 Batch 1440/2125] avg loss 0.00107991, throughput 6.02216K wps
[Epoch 18 Batch 1470/2125] avg loss 0.00119865, throughput 6.02518K wps
[Epoch 18 Batch 1500/2125] avg loss 0.00158996, throughput 6.0144K wps
[Epoch 18 Batch 1530/2125] avg loss 0.00110104, throughput 6.01582K wps
[Epoch 18 Batch 1560/2125] avg loss 0.00105103, throughput 6.02287K wps
[Epoch 18 Batch 1590/2125] avg loss 0.00126462, throughput 6.02873K wps
[Epoch 18 Batch 1620/2125] avg loss 0.000908892, throughput 6.01512K wps
[Epoch 18 Batch 1650/2125] avg loss 0.00127227, throughput 6.02295K wps
[Epoch 18 Batch 1680/2125] avg loss 0.000966295, throughput 6.02446K wps
[Epoch 18 Batch 1710/2125] avg loss 0.00111314, throughput 6.01527K wps
[Epoch 18 Batch 1740/2125] avg loss 0.00124689, throughput 6.01781K wps
[Epoch 18 Batch 1770/2125] avg loss 0.00115228, throughput 6.0213K wps
[Epoch 18 Batch 1800/2125] avg loss 0.00101937, throughput 6.01047K wps
[Epoch 18 Batch 1830/2125] avg loss 0.00108011, throughput 6.02389K wps
[Epoch 18 Batch 1860/2125] avg loss 0.00106205, throughput 6.02521K wps
[Epoch 18 Batch 1890/2125] avg loss 0.00123929, throughput 6.01923K wps
[Epoch 18 Batch 1920/2125] avg loss 0.0014739, throughput 6.01937K wps
[Epoch 18 Batch 1950/2125] avg loss 0.00124162, throughput 6.034K wps
[Epoch 18 Batch 1980/2125] avg loss 0.00105689, throughput 6.02171K wps
[Epoch 18 Batch 2010/2125] avg loss 0.00130178, throughput 6.02201K wps
[Epoch 18 Batch 2040/2125] avg loss 0.00111735, throughput 6.01733K wps
[Epoch 18 Batch 2070/2125] avg loss 0.0013211, throughput 6.01883K wps
[Epoch 18 Batch 2100/2125] avg loss 0.00139475, throughput 6.01069K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 18] train avg loss 0.00109693, test acc 0.9277, test avg loss 0.371421, throughput 6.02114K wps
[Epoch 19 Batch 30/2125] avg loss 0.000709907, throughput 6.14712K wps
[Epoch 19 Batch 60/2125] avg loss 0.00099514, throughput 6.00149K wps
[Epoch 19 Batch 90/2125] avg loss 0.000763144, throughput 6.00421K wps
[Epoch 19 Batch 120/2125] avg loss 0.000810787, throughput 6.01303K wps
[Epoch 19 Batch 150/2125] avg loss 0.00077597, throughput 6.01854K wps
[Epoch 19 Batch 180/2125] avg loss 0.000854127, throughput 6.01275K wps
[Epoch 19 Batch 210/2125] avg loss 0.0011791, throughput 5.9983K wps
[Epoch 19 Batch 240/2125] avg loss 0.00124338, throughput 5.96746K wps
[Epoch 19 Batch 270/2125] avg loss 0.00102799, throughput 6.0083K wps
[Epoch 19 Batch 300/2125] avg loss 0.00094047, throughput 6.02111K wps
[Epoch 19 Batch 330/2125] avg loss 0.000891375, throughput 6.00818K wps
[Epoch 19 Batch 360/2125] avg loss 0.000778277, throughput 6.01588K wps
[Epoch 19 Batch 390/2125] avg loss 0.000949477, throughput 6.0094K wps
[Epoch 19 Batch 420/2125] avg loss 0.000969605, throughput 6.01407K wps
[Epoch 19 Batch 450/2125] avg loss 0.000922991, throughput 6.01444K wps
[Epoch 19 Batch 480/2125] avg loss 0.00128131, throughput 6.01599K wps
[Epoch 19 Batch 510/2125] avg loss 0.00108032, throughput 6.0058K wps
[Epoch 19 Batch 540/2125] avg loss 0.00098382, throughput 6.00903K wps
[Epoch 19 Batch 570/2125] avg loss 0.00114365, throughput 6.01013K wps
[Epoch 19 Batch 600/2125] avg loss 0.00105989, throughput 6.00664K wps
[Epoch 19 Batch 630/2125] avg loss 0.00097341, throughput 6.01309K wps
[Epoch 19 Batch 660/2125] avg loss 0.00103617, throughput 6.00882K wps
[Epoch 19 Batch 690/2125] avg loss 0.00133535, throughput 6.00608K wps
[Epoch 19 Batch 720/2125] avg loss 0.00115783, throughput 6.00994K wps
[Epoch 19 Batch 750/2125] avg loss 0.000940525, throughput 6.00862K wps
[Epoch 19 Batch 780/2125] avg loss 0.00080173, throughput 6.01375K wps
[Epoch 19 Batch 810/2125] avg loss 0.00102424, throughput 6.01293K wps
[Epoch 19 Batch 840/2125] avg loss 0.00096557, throughput 6.02269K wps
[Epoch 19 Batch 870/2125] avg loss 0.000973778, throughput 6.01317K wps
[Epoch 19 Batch 900/2125] avg loss 0.00110342, throughput 6.02627K wps
[Epoch 19 Batch 930/2125] avg loss 0.00107776, throughput 6.01586K wps
[Epoch 19 Batch 960/2125] avg loss 0.0010197, throughput 6.02829K wps
[Epoch 19 Batch 990/2125] avg loss 0.000991303, throughput 6.01571K wps
[Epoch 19 Batch 1020/2125] avg loss 0.00110074, throughput 6.01467K wps
[Epoch 19 Batch 1050/2125] avg loss 0.000898195, throughput 6.02339K wps
[Epoch 19 Batch 1080/2125] avg loss 0.00100657, throughput 6.01127K wps
[Epoch 19 Batch 1110/2125] avg loss 0.000962801, throughput 6.02768K wps
[Epoch 19 Batch 1140/2125] avg loss 0.00101767, throughput 6.01927K wps
[Epoch 19 Batch 1170/2125] avg loss 0.00091216, throughput 6.01428K wps
[Epoch 19 Batch 1200/2125] avg loss 0.00107986, throughput 6.01908K wps
[Epoch 19 Batch 1230/2125] avg loss 0.00109747, throughput 6.00948K wps
[Epoch 19 Batch 1260/2125] avg loss 0.000847505, throughput 6.01085K wps
[Epoch 19 Batch 1290/2125] avg loss 0.00123716, throughput 6.01412K wps
[Epoch 19 Batch 1320/2125] avg loss 0.0010004, throughput 6.00896K wps
[Epoch 19 Batch 1350/2125] avg loss 0.000947529, throughput 6.01144K wps
[Epoch 19 Batch 1380/2125] avg loss 0.000926755, throughput 6.01804K wps
[Epoch 19 Batch 1410/2125] avg loss 0.000951126, throughput 6.02096K wps
[Epoch 19 Batch 1440/2125] avg loss 0.00111626, throughput 6.01389K wps
[Epoch 19 Batch 1470/2125] avg loss 0.00113345, throughput 6.01372K wps
[Epoch 19 Batch 1500/2125] avg loss 0.00112871, throughput 6.0217K wps
[Epoch 19 Batch 1530/2125] avg loss 0.00129302, throughput 6.0169K wps
[Epoch 19 Batch 1560/2125] avg loss 0.00118624, throughput 6.01198K wps
[Epoch 19 Batch 1590/2125] avg loss 0.00129351, throughput 6.02756K wps
[Epoch 19 Batch 1620/2125] avg loss 0.00104542, throughput 6.02231K wps
[Epoch 19 Batch 1650/2125] avg loss 0.00104802, throughput 6.00834K wps
[Epoch 19 Batch 1680/2125] avg loss 0.00107737, throughput 6.02346K wps
[Epoch 19 Batch 1710/2125] avg loss 0.00101448, throughput 6.0323K wps
[Epoch 19 Batch 1740/2125] avg loss 0.00108237, throughput 6.02163K wps
[Epoch 19 Batch 1770/2125] avg loss 0.00114523, throughput 6.02063K wps
[Epoch 19 Batch 1800/2125] avg loss 0.00106179, throughput 6.02483K wps
[Epoch 19 Batch 1830/2125] avg loss 0.00106158, throughput 6.02814K wps
[Epoch 19 Batch 1860/2125] avg loss 0.00100421, throughput 6.01527K wps
[Epoch 19 Batch 1890/2125] avg loss 0.00131481, throughput 6.01623K wps
[Epoch 19 Batch 1920/2125] avg loss 0.000995391, throughput 6.02521K wps
[Epoch 19 Batch 1950/2125] avg loss 0.00107049, throughput 6.01499K wps
[Epoch 19 Batch 1980/2125] avg loss 0.00110572, throughput 6.01994K wps
[Epoch 19 Batch 2010/2125] avg loss 0.00084731, throughput 6.0073K wps
[Epoch 19 Batch 2040/2125] avg loss 0.00121508, throughput 6.01814K wps
[Epoch 19 Batch 2070/2125] avg loss 0.00132456, throughput 6.01303K wps
[Epoch 19 Batch 2100/2125] avg loss 0.00136898, throughput 6.01105K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 19] train avg loss 0.00103943, test acc 0.9277, test avg loss 0.382672, throughput 6.01635K wps
[Epoch 20 Batch 30/2125] avg loss 0.000766517, throughput 6.16949K wps
[Epoch 20 Batch 60/2125] avg loss 0.00127279, throughput 6.02369K wps
[Epoch 20 Batch 90/2125] avg loss 0.000826111, throughput 6.02059K wps
[Epoch 20 Batch 120/2125] avg loss 0.00091852, throughput 6.02042K wps
[Epoch 20 Batch 150/2125] avg loss 0.00081213, throughput 6.02415K wps
[Epoch 20 Batch 180/2125] avg loss 0.000824622, throughput 6.01718K wps
[Epoch 20 Batch 210/2125] avg loss 0.0008642, throughput 6.01877K wps
[Epoch 20 Batch 240/2125] avg loss 0.000822092, throughput 6.01369K wps
[Epoch 20 Batch 270/2125] avg loss 0.000720896, throughput 6.02486K wps
[Epoch 20 Batch 300/2125] avg loss 0.000789478, throughput 6.00871K wps
[Epoch 20 Batch 330/2125] avg loss 0.000783648, throughput 6.0176K wps
[Epoch 20 Batch 360/2125] avg loss 0.000830137, throughput 6.0117K wps
[Epoch 20 Batch 390/2125] avg loss 0.000765871, throughput 6.02238K wps
[Epoch 20 Batch 420/2125] avg loss 0.000927282, throughput 6.02578K wps
[Epoch 20 Batch 450/2125] avg loss 0.00099298, throughput 6.01011K wps
[Epoch 20 Batch 480/2125] avg loss 0.000951018, throughput 6.01599K wps
[Epoch 20 Batch 510/2125] avg loss 0.000920504, throughput 6.02587K wps
[Epoch 20 Batch 540/2125] avg loss 0.00101553, throughput 6.02409K wps
[Epoch 20 Batch 570/2125] avg loss 0.000782682, throughput 6.02642K wps
[Epoch 20 Batch 600/2125] avg loss 0.000699617, throughput 6.02135K wps
[Epoch 20 Batch 630/2125] avg loss 0.00113477, throughput 6.027K wps
[Epoch 20 Batch 660/2125] avg loss 0.000901179, throughput 6.02885K wps
[Epoch 20 Batch 690/2125] avg loss 0.00112075, throughput 6.00809K wps
[Epoch 20 Batch 720/2125] avg loss 0.00104315, throughput 6.02296K wps
[Epoch 20 Batch 750/2125] avg loss 0.00121782, throughput 6.02199K wps
[Epoch 20 Batch 780/2125] avg loss 0.000719691, throughput 6.01393K wps
[Epoch 20 Batch 810/2125] avg loss 0.000881129, throughput 6.01608K wps
[Epoch 20 Batch 840/2125] avg loss 0.000823492, throughput 6.02107K wps
[Epoch 20 Batch 870/2125] avg loss 0.00111129, throughput 6.01566K wps
[Epoch 20 Batch 900/2125] avg loss 0.000979171, throughput 6.00926K wps
[Epoch 20 Batch 930/2125] avg loss 0.00110865, throughput 6.01256K wps
[Epoch 20 Batch 960/2125] avg loss 0.00104865, throughput 6.02927K wps
[Epoch 20 Batch 990/2125] avg loss 0.000853041, throughput 6.01533K wps
[Epoch 20 Batch 1020/2125] avg loss 0.000994174, throughput 6.01927K wps
[Epoch 20 Batch 1050/2125] avg loss 0.000913339, throughput 6.01496K wps
[Epoch 20 Batch 1080/2125] avg loss 0.000812404, throughput 6.01511K wps
[Epoch 20 Batch 1110/2125] avg loss 0.000933515, throughput 6.02816K wps
[Epoch 20 Batch 1140/2125] avg loss 0.00124534, throughput 6.01138K wps
[Epoch 20 Batch 1170/2125] avg loss 0.00111946, throughput 6.02339K wps
[Epoch 20 Batch 1200/2125] avg loss 0.00104956, throughput 6.01825K wps
[Epoch 20 Batch 1230/2125] avg loss 0.0010214, throughput 6.00831K wps
[Epoch 20 Batch 1260/2125] avg loss 0.00120769, throughput 6.01305K wps
[Epoch 20 Batch 1290/2125] avg loss 0.000966026, throughput 6.01622K wps
[Epoch 20 Batch 1320/2125] avg loss 0.0011422, throughput 6.02183K wps
[Epoch 20 Batch 1350/2125] avg loss 0.00109952, throughput 6.00976K wps
[Epoch 20 Batch 1380/2125] avg loss 0.000930208, throughput 6.00392K wps
[Epoch 20 Batch 1410/2125] avg loss 0.00121201, throughput 6.01356K wps
[Epoch 20 Batch 1440/2125] avg loss 0.000865655, throughput 6.0157K wps
[Epoch 20 Batch 1470/2125] avg loss 0.000748427, throughput 6.01392K wps
[Epoch 20 Batch 1500/2125] avg loss 0.00116758, throughput 6.01345K wps
[Epoch 20 Batch 1530/2125] avg loss 0.000964328, throughput 6.02158K wps
[Epoch 20 Batch 1560/2125] avg loss 0.00126029, throughput 5.99913K wps
[Epoch 20 Batch 1590/2125] avg loss 0.000978653, throughput 6.02113K wps
[Epoch 20 Batch 1620/2125] avg loss 0.00116591, throughput 6.01807K wps
[Epoch 20 Batch 1650/2125] avg loss 0.00104375, throughput 6.02263K wps
[Epoch 20 Batch 1680/2125] avg loss 0.00128772, throughput 6.02327K wps
[Epoch 20 Batch 1710/2125] avg loss 0.00119439, throughput 6.02088K wps
[Epoch 20 Batch 1740/2125] avg loss 0.000967488, throughput 6.02804K wps
[Epoch 20 Batch 1770/2125] avg loss 0.000948422, throughput 6.00198K wps
[Epoch 20 Batch 1800/2125] avg loss 0.00088143, throughput 6.00651K wps
[Epoch 20 Batch 1830/2125] avg loss 0.00105678, throughput 6.01923K wps
[Epoch 20 Batch 1860/2125] avg loss 0.00105807, throughput 6.02098K wps
[Epoch 20 Batch 1890/2125] avg loss 0.000910123, throughput 6.0173K wps
[Epoch 20 Batch 1920/2125] avg loss 0.00129143, throughput 6.02084K wps
[Epoch 20 Batch 1950/2125] avg loss 0.000987933, throughput 6.01039K wps
[Epoch 20 Batch 1980/2125] avg loss 0.00101744, throughput 6.02203K wps
[Epoch 20 Batch 2010/2125] avg loss 0.00111551, throughput 6.01858K wps
[Epoch 20 Batch 2040/2125] avg loss 0.00113568, throughput 6.02188K wps
[Epoch 20 Batch 2070/2125] avg loss 0.00114575, throughput 6.02232K wps
[Epoch 20 Batch 2100/2125] avg loss 0.000970812, throughput 6.02219K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 20] train avg loss 0.000987143, test acc 0.9282, test avg loss 0.394505, throughput 6.01994K wps
[Epoch 21 Batch 30/2125] avg loss 0.000871003, throughput 6.14624K wps
[Epoch 21 Batch 60/2125] avg loss 0.00066755, throughput 6.01257K wps
[Epoch 21 Batch 90/2125] avg loss 0.000867658, throughput 6.01802K wps
[Epoch 21 Batch 120/2125] avg loss 0.000845353, throughput 6.02016K wps
[Epoch 21 Batch 150/2125] avg loss 0.000965835, throughput 6.02014K wps
[Epoch 21 Batch 180/2125] avg loss 0.000909754, throughput 6.00574K wps
[Epoch 21 Batch 210/2125] avg loss 0.000875807, throughput 6.01536K wps
[Epoch 21 Batch 240/2125] avg loss 0.000904203, throughput 6.01866K wps
[Epoch 21 Batch 270/2125] avg loss 0.000860118, throughput 6.01386K wps
[Epoch 21 Batch 300/2125] avg loss 0.000738827, throughput 6.02323K wps
[Epoch 21 Batch 330/2125] avg loss 0.00080597, throughput 6.01993K wps
[Epoch 21 Batch 360/2125] avg loss 0.000932411, throughput 6.02217K wps
[Epoch 21 Batch 390/2125] avg loss 0.000874892, throughput 6.02249K wps
[Epoch 21 Batch 420/2125] avg loss 0.000724263, throughput 6.0199K wps
[Epoch 21 Batch 450/2125] avg loss 0.000794124, throughput 6.02632K wps
[Epoch 21 Batch 480/2125] avg loss 0.000741047, throughput 6.01436K wps
[Epoch 21 Batch 510/2125] avg loss 0.000848311, throughput 6.01767K wps
[Epoch 21 Batch 540/2125] avg loss 0.000960267, throughput 6.0159K wps
[Epoch 21 Batch 570/2125] avg loss 0.000809379, throughput 6.00954K wps
[Epoch 21 Batch 600/2125] avg loss 0.000761192, throughput 6.01942K wps
[Epoch 21 Batch 630/2125] avg loss 0.000805568, throughput 6.02755K wps
[Epoch 21 Batch 660/2125] avg loss 0.00117109, throughput 6.00088K wps
[Epoch 21 Batch 690/2125] avg loss 0.00086334, throughput 6.00851K wps
[Epoch 21 Batch 720/2125] avg loss 0.00101747, throughput 6.01819K wps
[Epoch 21 Batch 750/2125] avg loss 0.000798602, throughput 6.01516K wps
[Epoch 21 Batch 780/2125] avg loss 0.000982956, throughput 6.01714K wps
[Epoch 21 Batch 810/2125] avg loss 0.000695188, throughput 6.01337K wps
[Epoch 21 Batch 840/2125] avg loss 0.000840892, throughput 6.01422K wps
[Epoch 21 Batch 870/2125] avg loss 0.000996045, throughput 6.01933K wps
[Epoch 21 Batch 900/2125] avg loss 0.000802486, throughput 6.01906K wps
[Epoch 21 Batch 930/2125] avg loss 0.000870642, throughput 6.02746K wps
[Epoch 21 Batch 960/2125] avg loss 0.000930234, throughput 6.0189K wps
[Epoch 21 Batch 990/2125] avg loss 0.000715667, throughput 6.02627K wps
[Epoch 21 Batch 1020/2125] avg loss 0.00101723, throughput 6.0161K wps
[Epoch 21 Batch 1050/2125] avg loss 0.000825354, throughput 6.01634K wps
[Epoch 21 Batch 1080/2125] avg loss 0.00122644, throughput 6.01258K wps
[Epoch 21 Batch 1110/2125] avg loss 0.000789749, throughput 6.02209K wps
[Epoch 21 Batch 1140/2125] avg loss 0.000971948, throughput 6.01857K wps
[Epoch 21 Batch 1170/2125] avg loss 0.000979618, throughput 6.01939K wps
[Epoch 21 Batch 1200/2125] avg loss 0.000782288, throughput 6.01194K wps
[Epoch 21 Batch 1230/2125] avg loss 0.00099474, throughput 6.0153K wps
[Epoch 21 Batch 1260/2125] avg loss 0.000705412, throughput 5.99093K wps
[Epoch 21 Batch 1290/2125] avg loss 0.00107997, throughput 5.98749K wps
[Epoch 21 Batch 1320/2125] avg loss 0.00116493, throughput 6.00725K wps
[Epoch 21 Batch 1350/2125] avg loss 0.00137123, throughput 6.01727K wps
[Epoch 21 Batch 1380/2125] avg loss 0.00102958, throughput 6.02055K wps
[Epoch 21 Batch 1410/2125] avg loss 0.00108592, throughput 6.01284K wps
[Epoch 21 Batch 1440/2125] avg loss 0.00090357, throughput 6.01151K wps
[Epoch 21 Batch 1470/2125] avg loss 0.000873058, throughput 6.01399K wps
[Epoch 21 Batch 1500/2125] avg loss 0.000760589, throughput 6.01607K wps
[Epoch 21 Batch 1530/2125] avg loss 0.000935511, throughput 6.01442K wps
[Epoch 21 Batch 1560/2125] avg loss 0.00104259, throughput 6.0091K wps
[Epoch 21 Batch 1590/2125] avg loss 0.000970954, throughput 6.01353K wps
[Epoch 21 Batch 1620/2125] avg loss 0.00106496, throughput 5.99084K wps
[Epoch 21 Batch 1650/2125] avg loss 0.00116455, throughput 6.00424K wps
[Epoch 21 Batch 1680/2125] avg loss 0.0012147, throughput 6.01877K wps
[Epoch 21 Batch 1710/2125] avg loss 0.00103372, throughput 6.02721K wps
[Epoch 21 Batch 1740/2125] avg loss 0.00111338, throughput 6.02785K wps
[Epoch 21 Batch 1770/2125] avg loss 0.000871248, throughput 6.01448K wps
[Epoch 21 Batch 1800/2125] avg loss 0.000909905, throughput 6.02656K wps
[Epoch 21 Batch 1830/2125] avg loss 0.00122069, throughput 6.02234K wps
[Epoch 21 Batch 1860/2125] avg loss 0.000865826, throughput 6.02326K wps
[Epoch 21 Batch 1890/2125] avg loss 0.000880791, throughput 6.03123K wps
[Epoch 21 Batch 1920/2125] avg loss 0.00069663, throughput 6.01651K wps
[Epoch 21 Batch 1950/2125] avg loss 0.00101694, throughput 6.02253K wps
[Epoch 21 Batch 1980/2125] avg loss 0.00123288, throughput 6.01416K wps
[Epoch 21 Batch 2010/2125] avg loss 0.00114745, throughput 6.02764K wps
[Epoch 21 Batch 2040/2125] avg loss 0.00116874, throughput 6.02004K wps
[Epoch 21 Batch 2070/2125] avg loss 0.00120104, throughput 6.01431K wps
[Epoch 21 Batch 2100/2125] avg loss 0.00101232, throughput 6.01175K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 21] train avg loss 0.000940903, test acc 0.9283, test avg loss 0.404603, throughput 6.01807K wps
[Epoch 22 Batch 30/2125] avg loss 0.000700278, throughput 6.14802K wps
[Epoch 22 Batch 60/2125] avg loss 0.000514831, throughput 6.01044K wps
[Epoch 22 Batch 90/2125] avg loss 0.000751283, throughput 6.01959K wps
[Epoch 22 Batch 120/2125] avg loss 0.000653659, throughput 6.01159K wps
[Epoch 22 Batch 150/2125] avg loss 0.00060791, throughput 6.0152K wps
[Epoch 22 Batch 180/2125] avg loss 0.000742712, throughput 6.01145K wps
[Epoch 22 Batch 210/2125] avg loss 0.000862335, throughput 6.01238K wps
[Epoch 22 Batch 240/2125] avg loss 0.000983504, throughput 6.01465K wps
[Epoch 22 Batch 270/2125] avg loss 0.000672759, throughput 6.01619K wps
[Epoch 22 Batch 300/2125] avg loss 0.000703842, throughput 6.01318K wps
[Epoch 22 Batch 330/2125] avg loss 0.000682151, throughput 6.00242K wps
[Epoch 22 Batch 360/2125] avg loss 0.00104764, throughput 5.99926K wps
[Epoch 22 Batch 390/2125] avg loss 0.000838399, throughput 6.01048K wps
[Epoch 22 Batch 420/2125] avg loss 0.000793955, throughput 6.00912K wps
[Epoch 22 Batch 450/2125] avg loss 0.000833868, throughput 6.01313K wps
[Epoch 22 Batch 480/2125] avg loss 0.000914517, throughput 6.01393K wps
[Epoch 22 Batch 510/2125] avg loss 0.00110132, throughput 6.02079K wps
[Epoch 22 Batch 540/2125] avg loss 0.000808055, throughput 6.01904K wps
[Epoch 22 Batch 570/2125] avg loss 0.00105182, throughput 6.01484K wps
[Epoch 22 Batch 600/2125] avg loss 0.000905248, throughput 6.01382K wps
[Epoch 22 Batch 630/2125] avg loss 0.00143976, throughput 6.02292K wps
[Epoch 22 Batch 660/2125] avg loss 0.000969646, throughput 6.01394K wps
[Epoch 22 Batch 690/2125] avg loss 0.000930404, throughput 6.02229K wps
[Epoch 22 Batch 720/2125] avg loss 0.000775092, throughput 6.0183K wps
[Epoch 22 Batch 750/2125] avg loss 0.000895447, throughput 6.01683K wps
[Epoch 22 Batch 780/2125] avg loss 0.000904602, throughput 6.00239K wps
[Epoch 22 Batch 810/2125] avg loss 0.000721593, throughput 6.01142K wps
[Epoch 22 Batch 840/2125] avg loss 0.000675888, throughput 6.00909K wps
[Epoch 22 Batch 870/2125] avg loss 0.000933884, throughput 6.01608K wps
[Epoch 22 Batch 900/2125] avg loss 0.000830793, throughput 6.01744K wps
[Epoch 22 Batch 930/2125] avg loss 0.00106188, throughput 6.01415K wps
[Epoch 22 Batch 960/2125] avg loss 0.000891207, throughput 6.01632K wps
[Epoch 22 Batch 990/2125] avg loss 0.000891825, throughput 6.01191K wps
[Epoch 22 Batch 1020/2125] avg loss 0.00080095, throughput 6.00767K wps
[Epoch 22 Batch 1050/2125] avg loss 0.000957851, throughput 6.01445K wps
[Epoch 22 Batch 1080/2125] avg loss 0.000771127, throughput 6.01705K wps
[Epoch 22 Batch 1110/2125] avg loss 0.00105982, throughput 6.01487K wps
[Epoch 22 Batch 1140/2125] avg loss 0.000818904, throughput 6.01716K wps
[Epoch 22 Batch 1170/2125] avg loss 0.000861342, throughput 6.01069K wps
[Epoch 22 Batch 1200/2125] avg loss 0.000984939, throughput 6.01166K wps
[Epoch 22 Batch 1230/2125] avg loss 0.000791102, throughput 6.01613K wps
[Epoch 22 Batch 1260/2125] avg loss 0.000922868, throughput 6.01862K wps
[Epoch 22 Batch 1290/2125] avg loss 0.000912361, throughput 6.00912K wps
[Epoch 22 Batch 1320/2125] avg loss 0.000855276, throughput 6.01668K wps
[Epoch 22 Batch 1350/2125] avg loss 0.000832863, throughput 6.01024K wps
[Epoch 22 Batch 1380/2125] avg loss 0.00110448, throughput 6.00844K wps
[Epoch 22 Batch 1410/2125] avg loss 0.000869952, throughput 6.00462K wps
[Epoch 22 Batch 1440/2125] avg loss 0.00103367, throughput 6.01306K wps
[Epoch 22 Batch 1470/2125] avg loss 0.00090721, throughput 6.00839K wps
[Epoch 22 Batch 1500/2125] avg loss 0.000873132, throughput 6.00521K wps
[Epoch 22 Batch 1530/2125] avg loss 0.00105319, throughput 6.00707K wps
[Epoch 22 Batch 1560/2125] avg loss 0.000985223, throughput 6.00424K wps
[Epoch 22 Batch 1590/2125] avg loss 0.000954859, throughput 6.01859K wps
[Epoch 22 Batch 1620/2125] avg loss 0.00109711, throughput 6.00592K wps
[Epoch 22 Batch 1650/2125] avg loss 0.0008739, throughput 5.99851K wps
[Epoch 22 Batch 1680/2125] avg loss 0.000928549, throughput 6.01816K wps
[Epoch 22 Batch 1710/2125] avg loss 0.00125052, throughput 6.02361K wps
[Epoch 22 Batch 1740/2125] avg loss 0.000888419, throughput 6.02366K wps
[Epoch 22 Batch 1770/2125] avg loss 0.00113149, throughput 6.01843K wps
[Epoch 22 Batch 1800/2125] avg loss 0.000990359, throughput 6.00571K wps
[Epoch 22 Batch 1830/2125] avg loss 0.00115914, throughput 6.01004K wps
[Epoch 22 Batch 1860/2125] avg loss 0.000980899, throughput 6.01824K wps
[Epoch 22 Batch 1890/2125] avg loss 0.000954596, throughput 6.01308K wps
[Epoch 22 Batch 1920/2125] avg loss 0.001036, throughput 4.91555K wps
[Epoch 22 Batch 1950/2125] avg loss 0.000959444, throughput 6.01256K wps
[Epoch 22 Batch 1980/2125] avg loss 0.00105384, throughput 6.00573K wps
[Epoch 22 Batch 2010/2125] avg loss 0.00117421, throughput 6.01378K wps
[Epoch 22 Batch 2040/2125] avg loss 0.000643583, throughput 6.0093K wps
[Epoch 22 Batch 2070/2125] avg loss 0.00102081, throughput 6.01898K wps
[Epoch 22 Batch 2100/2125] avg loss 0.0010996, throughput 6.01761K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 22] train avg loss 0.000911913, test acc 0.9283, test avg loss 0.411143, throughput 5.99602K wps
[Epoch 23 Batch 30/2125] avg loss 0.000714227, throughput 6.15604K wps
[Epoch 23 Batch 60/2125] avg loss 0.00064315, throughput 6.01829K wps
[Epoch 23 Batch 90/2125] avg loss 0.000644621, throughput 6.02084K wps
[Epoch 23 Batch 120/2125] avg loss 0.000721235, throughput 6.0113K wps
[Epoch 23 Batch 150/2125] avg loss 0.00071406, throughput 6.02163K wps
[Epoch 23 Batch 180/2125] avg loss 0.000682732, throughput 6.01475K wps
[Epoch 23 Batch 210/2125] avg loss 0.000686617, throughput 6.01561K wps
[Epoch 23 Batch 240/2125] avg loss 0.00109406, throughput 6.01868K wps
[Epoch 23 Batch 270/2125] avg loss 0.000567516, throughput 6.01328K wps
[Epoch 23 Batch 300/2125] avg loss 0.000530952, throughput 6.02828K wps
[Epoch 23 Batch 330/2125] avg loss 0.000904731, throughput 6.02314K wps
[Epoch 23 Batch 360/2125] avg loss 0.000614482, throughput 6.01246K wps
[Epoch 23 Batch 390/2125] avg loss 0.000797762, throughput 6.02228K wps
[Epoch 23 Batch 420/2125] avg loss 0.000654905, throughput 6.02829K wps
[Epoch 23 Batch 450/2125] avg loss 0.000838655, throughput 6.01146K wps
[Epoch 23 Batch 480/2125] avg loss 0.000765333, throughput 6.00749K wps
[Epoch 23 Batch 510/2125] avg loss 0.000789932, throughput 6.01269K wps
[Epoch 23 Batch 540/2125] avg loss 0.000740501, throughput 6.01156K wps
[Epoch 23 Batch 570/2125] avg loss 0.000790974, throughput 6.00404K wps
[Epoch 23 Batch 600/2125] avg loss 0.000896102, throughput 6.00525K wps
[Epoch 23 Batch 630/2125] avg loss 0.000684406, throughput 6.013K wps
[Epoch 23 Batch 660/2125] avg loss 0.000724695, throughput 6.01806K wps
[Epoch 23 Batch 690/2125] avg loss 0.000847814, throughput 6.02499K wps
[Epoch 23 Batch 720/2125] avg loss 0.000985351, throughput 6.019K wps
[Epoch 23 Batch 750/2125] avg loss 0.000972872, throughput 6.01151K wps
[Epoch 23 Batch 780/2125] avg loss 0.000923572, throughput 6.01484K wps
[Epoch 23 Batch 810/2125] avg loss 0.00094659, throughput 6.02824K wps
[Epoch 23 Batch 840/2125] avg loss 0.000769842, throughput 6.01472K wps
[Epoch 23 Batch 870/2125] avg loss 0.00100737, throughput 6.02318K wps
[Epoch 23 Batch 900/2125] avg loss 0.000877051, throughput 6.00427K wps
[Epoch 23 Batch 930/2125] avg loss 0.000706285, throughput 6.0095K wps
[Epoch 23 Batch 960/2125] avg loss 0.000800328, throughput 6.01108K wps
[Epoch 23 Batch 990/2125] avg loss 0.000862063, throughput 6.02517K wps
[Epoch 23 Batch 1020/2125] avg loss 0.000752986, throughput 6.01098K wps
[Epoch 23 Batch 1050/2125] avg loss 0.000936651, throughput 6.02459K wps
[Epoch 23 Batch 1080/2125] avg loss 0.000984744, throughput 6.01355K wps
[Epoch 23 Batch 1110/2125] avg loss 0.000773674, throughput 6.02383K wps
[Epoch 23 Batch 1140/2125] avg loss 0.000838275, throughput 6.02091K wps
[Epoch 23 Batch 1170/2125] avg loss 0.000834419, throughput 6.01778K wps
[Epoch 23 Batch 1200/2125] avg loss 0.000845889, throughput 6.01455K wps
[Epoch 23 Batch 1230/2125] avg loss 0.000956615, throughput 6.01433K wps
[Epoch 23 Batch 1260/2125] avg loss 0.000886108, throughput 6.02082K wps
[Epoch 23 Batch 1290/2125] avg loss 0.000852967, throughput 6.0123K wps
[Epoch 23 Batch 1320/2125] avg loss 0.00101928, throughput 6.0178K wps
[Epoch 23 Batch 1350/2125] avg loss 0.000780119, throughput 6.01428K wps
[Epoch 23 Batch 1380/2125] avg loss 0.000822504, throughput 6.01817K wps
[Epoch 23 Batch 1410/2125] avg loss 0.000797503, throughput 6.01307K wps
[Epoch 23 Batch 1440/2125] avg loss 0.000819985, throughput 6.02912K wps
[Epoch 23 Batch 1470/2125] avg loss 0.000998892, throughput 6.01867K wps
[Epoch 23 Batch 1500/2125] avg loss 0.000648695, throughput 6.01201K wps
[Epoch 23 Batch 1530/2125] avg loss 0.000932707, throughput 6.01904K wps
[Epoch 23 Batch 1560/2125] avg loss 0.00103111, throughput 6.02257K wps
[Epoch 23 Batch 1590/2125] avg loss 0.000770372, throughput 6.02552K wps
[Epoch 23 Batch 1620/2125] avg loss 0.000912959, throughput 6.01847K wps
[Epoch 23 Batch 1650/2125] avg loss 0.000997372, throughput 6.01856K wps
[Epoch 23 Batch 1680/2125] avg loss 0.00076577, throughput 6.01776K wps
[Epoch 23 Batch 1710/2125] avg loss 0.000898312, throughput 6.0127K wps
[Epoch 23 Batch 1740/2125] avg loss 0.000782327, throughput 6.02457K wps
[Epoch 23 Batch 1770/2125] avg loss 0.00114321, throughput 6.0126K wps
[Epoch 23 Batch 1800/2125] avg loss 0.00077433, throughput 6.009K wps
[Epoch 23 Batch 1830/2125] avg loss 0.000865845, throughput 6.02545K wps
[Epoch 23 Batch 1860/2125] avg loss 0.000886153, throughput 6.01354K wps
[Epoch 23 Batch 1890/2125] avg loss 0.00111454, throughput 6.01037K wps
[Epoch 23 Batch 1920/2125] avg loss 0.000939486, throughput 6.02496K wps
[Epoch 23 Batch 1950/2125] avg loss 0.00097274, throughput 6.01557K wps
[Epoch 23 Batch 1980/2125] avg loss 0.000774415, throughput 6.0094K wps
[Epoch 23 Batch 2010/2125] avg loss 0.00107038, throughput 6.01602K wps
[Epoch 23 Batch 2040/2125] avg loss 0.000960621, throughput 6.01842K wps
[Epoch 23 Batch 2070/2125] avg loss 0.00104719, throughput 6.01609K wps
[Epoch 23 Batch 2100/2125] avg loss 0.00143183, throughput 6.02093K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 23] train avg loss 0.000848727, test acc 0.9280, test avg loss 0.421381, throughput 6.01899K wps
[Epoch 24 Batch 30/2125] avg loss 0.000766409, throughput 6.16036K wps
[Epoch 24 Batch 60/2125] avg loss 0.00071408, throughput 6.01743K wps
[Epoch 24 Batch 90/2125] avg loss 0.000850352, throughput 6.00932K wps
[Epoch 24 Batch 120/2125] avg loss 0.000660364, throughput 5.99789K wps
[Epoch 24 Batch 150/2125] avg loss 0.000835349, throughput 5.99678K wps
[Epoch 24 Batch 180/2125] avg loss 0.000661297, throughput 5.99909K wps
[Epoch 24 Batch 210/2125] avg loss 0.00070094, throughput 6.02001K wps
[Epoch 24 Batch 240/2125] avg loss 0.000885145, throughput 6.0113K wps
[Epoch 24 Batch 270/2125] avg loss 0.000955705, throughput 6.0163K wps
[Epoch 24 Batch 300/2125] avg loss 0.000928279, throughput 6.01888K wps
[Epoch 24 Batch 330/2125] avg loss 0.000824076, throughput 6.01463K wps
[Epoch 24 Batch 360/2125] avg loss 0.00070531, throughput 6.0191K wps
[Epoch 24 Batch 390/2125] avg loss 0.00071363, throughput 6.02617K wps
[Epoch 24 Batch 420/2125] avg loss 0.000986132, throughput 6.02117K wps
[Epoch 24 Batch 450/2125] avg loss 0.000584413, throughput 6.01741K wps
[Epoch 24 Batch 480/2125] avg loss 0.000638687, throughput 6.02708K wps
[Epoch 24 Batch 510/2125] avg loss 0.000781355, throughput 6.02168K wps
[Epoch 24 Batch 540/2125] avg loss 0.000988329, throughput 6.02237K wps
[Epoch 24 Batch 570/2125] avg loss 0.000795261, throughput 6.02779K wps
[Epoch 24 Batch 600/2125] avg loss 0.000791865, throughput 6.02039K wps
[Epoch 24 Batch 630/2125] avg loss 0.000691959, throughput 6.02001K wps
[Epoch 24 Batch 660/2125] avg loss 0.000839818, throughput 6.01511K wps
[Epoch 24 Batch 690/2125] avg loss 0.00107749, throughput 6.01085K wps
[Epoch 24 Batch 720/2125] avg loss 0.00073378, throughput 6.0111K wps
[Epoch 24 Batch 750/2125] avg loss 0.000688474, throughput 6.00998K wps
[Epoch 24 Batch 780/2125] avg loss 0.000745237, throughput 6.01223K wps
[Epoch 24 Batch 810/2125] avg loss 0.000975536, throughput 6.02199K wps
[Epoch 24 Batch 840/2125] avg loss 0.000789578, throughput 6.01179K wps
[Epoch 24 Batch 870/2125] avg loss 0.00082909, throughput 6.02427K wps
[Epoch 24 Batch 900/2125] avg loss 0.00091849, throughput 6.01394K wps
[Epoch 24 Batch 930/2125] avg loss 0.000893966, throughput 6.01759K wps
[Epoch 24 Batch 960/2125] avg loss 0.00073347, throughput 6.02445K wps
[Epoch 24 Batch 990/2125] avg loss 0.000758338, throughput 6.01124K wps
[Epoch 24 Batch 1020/2125] avg loss 0.000556278, throughput 6.01329K wps
[Epoch 24 Batch 1050/2125] avg loss 0.000756888, throughput 6.01071K wps
[Epoch 24 Batch 1080/2125] avg loss 0.00102254, throughput 6.00998K wps
[Epoch 24 Batch 1110/2125] avg loss 0.0011134, throughput 6.01608K wps
[Epoch 24 Batch 1140/2125] avg loss 0.000849966, throughput 6.01623K wps
[Epoch 24 Batch 1170/2125] avg loss 0.000691135, throughput 6.01049K wps
[Epoch 24 Batch 1200/2125] avg loss 0.00078924, throughput 6.0158K wps
[Epoch 24 Batch 1230/2125] avg loss 0.00085702, throughput 6.02156K wps
[Epoch 24 Batch 1260/2125] avg loss 0.000706021, throughput 6.01731K wps
[Epoch 24 Batch 1290/2125] avg loss 0.00079267, throughput 6.02633K wps
[Epoch 24 Batch 1320/2125] avg loss 0.000942899, throughput 6.01458K wps
[Epoch 24 Batch 1350/2125] avg loss 0.000957885, throughput 5.99809K wps
[Epoch 24 Batch 1380/2125] avg loss 0.000918017, throughput 6.0082K wps
[Epoch 24 Batch 1410/2125] avg loss 0.00075605, throughput 6.0176K wps
[Epoch 24 Batch 1440/2125] avg loss 0.000867871, throughput 6.0176K wps
[Epoch 24 Batch 1470/2125] avg loss 0.0010563, throughput 6.00604K wps
[Epoch 24 Batch 1500/2125] avg loss 0.000935312, throughput 6.01992K wps
[Epoch 24 Batch 1530/2125] avg loss 0.000668538, throughput 6.01401K wps
[Epoch 24 Batch 1560/2125] avg loss 0.000824972, throughput 6.01853K wps
[Epoch 24 Batch 1590/2125] avg loss 0.000928604, throughput 6.02958K wps
[Epoch 24 Batch 1620/2125] avg loss 0.000790669, throughput 6.02036K wps
[Epoch 24 Batch 1650/2125] avg loss 0.00107377, throughput 6.01604K wps
[Epoch 24 Batch 1680/2125] avg loss 0.000866434, throughput 6.01237K wps
[Epoch 24 Batch 1710/2125] avg loss 0.000683923, throughput 6.00983K wps
[Epoch 24 Batch 1740/2125] avg loss 0.000906712, throughput 6.02793K wps
[Epoch 24 Batch 1770/2125] avg loss 0.00102727, throughput 6.02533K wps
[Epoch 24 Batch 1800/2125] avg loss 0.00125812, throughput 6.02012K wps
[Epoch 24 Batch 1830/2125] avg loss 0.00076795, throughput 6.01261K wps
[Epoch 24 Batch 1860/2125] avg loss 0.000898253, throughput 6.01303K wps
[Epoch 24 Batch 1890/2125] avg loss 0.000862278, throughput 6.01489K wps
[Epoch 24 Batch 1920/2125] avg loss 0.000818708, throughput 6.02553K wps
[Epoch 24 Batch 1950/2125] avg loss 0.000905088, throughput 6.01958K wps
[Epoch 24 Batch 1980/2125] avg loss 0.000856147, throughput 6.01056K wps
[Epoch 24 Batch 2010/2125] avg loss 0.00101175, throughput 6.01382K wps
[Epoch 24 Batch 2040/2125] avg loss 0.000916202, throughput 6.02248K wps
[Epoch 24 Batch 2070/2125] avg loss 0.000749344, throughput 6.02604K wps
[Epoch 24 Batch 2100/2125] avg loss 0.00115099, throughput 6.01694K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 24] train avg loss 0.000845206, test acc 0.9273, test avg loss 0.43326, throughput 6.01814K wps
[Epoch 25 Batch 30/2125] avg loss 0.000551763, throughput 6.14868K wps
[Epoch 25 Batch 60/2125] avg loss 0.000671989, throughput 6.01464K wps
[Epoch 25 Batch 90/2125] avg loss 0.000717775, throughput 6.01172K wps
[Epoch 25 Batch 120/2125] avg loss 0.00067645, throughput 5.99665K wps
[Epoch 25 Batch 150/2125] avg loss 0.000779451, throughput 6.01462K wps
[Epoch 25 Batch 180/2125] avg loss 0.000827227, throughput 6.01097K wps
[Epoch 25 Batch 210/2125] avg loss 0.000729926, throughput 6.00893K wps
[Epoch 25 Batch 240/2125] avg loss 0.00069148, throughput 6.01466K wps
[Epoch 25 Batch 270/2125] avg loss 0.00100808, throughput 6.02253K wps
[Epoch 25 Batch 300/2125] avg loss 0.000581355, throughput 6.01301K wps
[Epoch 25 Batch 330/2125] avg loss 0.000697924, throughput 6.01273K wps
[Epoch 25 Batch 360/2125] avg loss 0.000819709, throughput 6.01704K wps
[Epoch 25 Batch 390/2125] avg loss 0.000792394, throughput 6.00901K wps
[Epoch 25 Batch 420/2125] avg loss 0.000631541, throughput 6.01309K wps
[Epoch 25 Batch 450/2125] avg loss 0.000552123, throughput 6.0082K wps
[Epoch 25 Batch 480/2125] avg loss 0.00094136, throughput 6.02164K wps
[Epoch 25 Batch 510/2125] avg loss 0.000814547, throughput 6.0059K wps
[Epoch 25 Batch 540/2125] avg loss 0.000641334, throughput 6.01982K wps
[Epoch 25 Batch 570/2125] avg loss 0.000825223, throughput 6.02176K wps
[Epoch 25 Batch 600/2125] avg loss 0.000644232, throughput 6.00944K wps
[Epoch 25 Batch 630/2125] avg loss 0.000812206, throughput 6.00673K wps
[Epoch 25 Batch 660/2125] avg loss 0.000782062, throughput 6.01006K wps
[Epoch 25 Batch 690/2125] avg loss 0.000806199, throughput 6.01476K wps
[Epoch 25 Batch 720/2125] avg loss 0.00068811, throughput 6.02148K wps
[Epoch 25 Batch 750/2125] avg loss 0.000823386, throughput 6.01548K wps
[Epoch 25 Batch 780/2125] avg loss 0.00100212, throughput 6.01569K wps
[Epoch 25 Batch 810/2125] avg loss 0.000928582, throughput 6.02592K wps
[Epoch 25 Batch 840/2125] avg loss 0.000744177, throughput 6.01227K wps
[Epoch 25 Batch 870/2125] avg loss 0.000622366, throughput 6.02368K wps
[Epoch 25 Batch 900/2125] avg loss 0.000823298, throughput 6.01335K wps
[Epoch 25 Batch 930/2125] avg loss 0.000863438, throughput 6.01081K wps
[Epoch 25 Batch 960/2125] avg loss 0.00070977, throughput 5.99231K wps
[Epoch 25 Batch 990/2125] avg loss 0.000760581, throughput 6.00134K wps
[Epoch 25 Batch 1020/2125] avg loss 0.000648461, throughput 6.02266K wps
[Epoch 25 Batch 1050/2125] avg loss 0.000729106, throughput 6.00546K wps
[Epoch 25 Batch 1080/2125] avg loss 0.000882836, throughput 6.02338K wps
[Epoch 25 Batch 1110/2125] avg loss 0.000706135, throughput 6.0342K wps
[Epoch 25 Batch 1140/2125] avg loss 0.000817255, throughput 6.01608K wps
[Epoch 25 Batch 1170/2125] avg loss 0.000786423, throughput 6.0173K wps
[Epoch 25 Batch 1200/2125] avg loss 0.00114047, throughput 6.01812K wps
[Epoch 25 Batch 1230/2125] avg loss 0.000794105, throughput 6.02064K wps
[Epoch 25 Batch 1260/2125] avg loss 0.000835851, throughput 6.01609K wps
[Epoch 25 Batch 1290/2125] avg loss 0.000723683, throughput 6.02407K wps
[Epoch 25 Batch 1320/2125] avg loss 0.000652058, throughput 6.02361K wps
[Epoch 25 Batch 1350/2125] avg loss 0.000912328, throughput 6.01734K wps
[Epoch 25 Batch 1380/2125] avg loss 0.000882918, throughput 6.02589K wps
[Epoch 25 Batch 1410/2125] avg loss 0.00086279, throughput 6.00904K wps
[Epoch 25 Batch 1440/2125] avg loss 0.000776909, throughput 6.00841K wps
[Epoch 25 Batch 1470/2125] avg loss 0.000881755, throughput 6.02266K wps
[Epoch 25 Batch 1500/2125] avg loss 0.000799945, throughput 6.02362K wps
[Epoch 25 Batch 1530/2125] avg loss 0.000882916, throughput 6.02019K wps
[Epoch 25 Batch 1560/2125] avg loss 0.000754581, throughput 6.01726K wps
[Epoch 25 Batch 1590/2125] avg loss 0.000999865, throughput 6.02047K wps
[Epoch 25 Batch 1620/2125] avg loss 0.000724849, throughput 6.01976K wps
[Epoch 25 Batch 1650/2125] avg loss 0.000866219, throughput 6.01033K wps
[Epoch 25 Batch 1680/2125] avg loss 0.000757675, throughput 6.01406K wps
[Epoch 25 Batch 1710/2125] avg loss 0.000895048, throughput 6.0232K wps
[Epoch 25 Batch 1740/2125] avg loss 0.00088353, throughput 6.01698K wps
[Epoch 25 Batch 1770/2125] avg loss 0.000632618, throughput 6.01281K wps
[Epoch 25 Batch 1800/2125] avg loss 0.00116591, throughput 6.02357K wps
[Epoch 25 Batch 1830/2125] avg loss 0.000851285, throughput 6.02025K wps
[Epoch 25 Batch 1860/2125] avg loss 0.000814403, throughput 6.01886K wps
[Epoch 25 Batch 1890/2125] avg loss 0.00103039, throughput 6.0212K wps
[Epoch 25 Batch 1920/2125] avg loss 0.00088961, throughput 6.01452K wps
[Epoch 25 Batch 1950/2125] avg loss 0.000925942, throughput 6.01559K wps
[Epoch 25 Batch 1980/2125] avg loss 0.000700979, throughput 6.00868K wps
[Epoch 25 Batch 2010/2125] avg loss 0.00120384, throughput 6.01607K wps
[Epoch 25 Batch 2040/2125] avg loss 0.000872584, throughput 6.01019K wps
[Epoch 25 Batch 2070/2125] avg loss 0.000950239, throughput 6.01553K wps
[Epoch 25 Batch 2100/2125] avg loss 0.000921461, throughput 6.02417K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 25] train avg loss 0.000808695, test acc 0.9265, test avg loss 0.44394, throughput 6.01757K wps
[Epoch 26 Batch 30/2125] avg loss 0.000629606, throughput 6.14547K wps
[Epoch 26 Batch 60/2125] avg loss 0.000435417, throughput 6.02043K wps
[Epoch 26 Batch 90/2125] avg loss 0.000850694, throughput 6.00297K wps
[Epoch 26 Batch 120/2125] avg loss 0.000656353, throughput 6.00761K wps
[Epoch 26 Batch 150/2125] avg loss 0.00069811, throughput 6.00983K wps
[Epoch 26 Batch 180/2125] avg loss 0.000608497, throughput 6.01432K wps
[Epoch 26 Batch 210/2125] avg loss 0.000859798, throughput 6.01647K wps
[Epoch 26 Batch 240/2125] avg loss 0.000657705, throughput 6.01529K wps
[Epoch 26 Batch 270/2125] avg loss 0.000869454, throughput 6.01619K wps
[Epoch 26 Batch 300/2125] avg loss 0.00054994, throughput 6.02457K wps
[Epoch 26 Batch 330/2125] avg loss 0.000699989, throughput 6.01294K wps
[Epoch 26 Batch 360/2125] avg loss 0.000523244, throughput 6.01271K wps
[Epoch 26 Batch 390/2125] avg loss 0.000536608, throughput 6.0172K wps
[Epoch 26 Batch 420/2125] avg loss 0.000647217, throughput 6.0037K wps
[Epoch 26 Batch 450/2125] avg loss 0.000647666, throughput 6.02015K wps
[Epoch 26 Batch 480/2125] avg loss 0.000902901, throughput 6.01528K wps
[Epoch 26 Batch 510/2125] avg loss 0.000772693, throughput 6.01448K wps
[Epoch 26 Batch 540/2125] avg loss 0.000553242, throughput 6.01394K wps
[Epoch 26 Batch 570/2125] avg loss 0.000745872, throughput 6.02448K wps
[Epoch 26 Batch 600/2125] avg loss 0.000857267, throughput 6.04143K wps
[Epoch 26 Batch 630/2125] avg loss 0.000625189, throughput 6.02701K wps
[Epoch 26 Batch 660/2125] avg loss 0.00062701, throughput 6.01866K wps
[Epoch 26 Batch 690/2125] avg loss 0.000753243, throughput 6.01503K wps
[Epoch 26 Batch 720/2125] avg loss 0.000819675, throughput 6.0182K wps
[Epoch 26 Batch 750/2125] avg loss 0.000761072, throughput 6.01732K wps
[Epoch 26 Batch 780/2125] avg loss 0.000536526, throughput 6.0005K wps
[Epoch 26 Batch 810/2125] avg loss 0.000828371, throughput 6.0118K wps
[Epoch 26 Batch 840/2125] avg loss 0.000697047, throughput 6.01482K wps
[Epoch 26 Batch 870/2125] avg loss 0.000776477, throughput 6.01268K wps
[Epoch 26 Batch 900/2125] avg loss 0.000620965, throughput 6.01901K wps
[Epoch 26 Batch 930/2125] avg loss 0.000588181, throughput 6.00836K wps
[Epoch 26 Batch 960/2125] avg loss 0.000590234, throughput 6.00822K wps
[Epoch 26 Batch 990/2125] avg loss 0.000630713, throughput 6.02145K wps
[Epoch 26 Batch 1020/2125] avg loss 0.000592393, throughput 5.99278K wps
[Epoch 26 Batch 1050/2125] avg loss 0.000698983, throughput 6.00831K wps
[Epoch 26 Batch 1080/2125] avg loss 0.000932948, throughput 6.01678K wps
[Epoch 26 Batch 1110/2125] avg loss 0.000900698, throughput 6.01411K wps
[Epoch 26 Batch 1140/2125] avg loss 0.000965135, throughput 5.99854K wps
[Epoch 26 Batch 1170/2125] avg loss 0.000621407, throughput 5.97477K wps
[Epoch 26 Batch 1200/2125] avg loss 0.000771057, throughput 6.00809K wps
[Epoch 26 Batch 1230/2125] avg loss 0.00082819, throughput 6.00885K wps
[Epoch 26 Batch 1260/2125] avg loss 0.000827468, throughput 6.01734K wps
[Epoch 26 Batch 1290/2125] avg loss 0.000642073, throughput 6.01397K wps
[Epoch 26 Batch 1320/2125] avg loss 0.000780492, throughput 6.01051K wps
[Epoch 26 Batch 1350/2125] avg loss 0.000655371, throughput 6.00712K wps
[Epoch 26 Batch 1380/2125] avg loss 0.00100117, throughput 6.01059K wps
[Epoch 26 Batch 1410/2125] avg loss 0.000692377, throughput 6.01088K wps
[Epoch 26 Batch 1440/2125] avg loss 0.000876032, throughput 6.02502K wps
[Epoch 26 Batch 1470/2125] avg loss 0.000899107, throughput 6.02395K wps
[Epoch 26 Batch 1500/2125] avg loss 0.000874944, throughput 6.02588K wps
[Epoch 26 Batch 1530/2125] avg loss 0.000741515, throughput 6.02699K wps
[Epoch 26 Batch 1560/2125] avg loss 0.00085555, throughput 6.0137K wps
[Epoch 26 Batch 1590/2125] avg loss 0.00098268, throughput 6.00615K wps
[Epoch 26 Batch 1620/2125] avg loss 0.00110888, throughput 6.01427K wps
[Epoch 26 Batch 1650/2125] avg loss 0.000640516, throughput 6.01605K wps
[Epoch 26 Batch 1680/2125] avg loss 0.000787004, throughput 6.01313K wps
[Epoch 26 Batch 1710/2125] avg loss 0.00103007, throughput 6.01318K wps
[Epoch 26 Batch 1740/2125] avg loss 0.000754781, throughput 6.00807K wps
[Epoch 26 Batch 1770/2125] avg loss 0.00120696, throughput 6.01504K wps
[Epoch 26 Batch 1800/2125] avg loss 0.00089088, throughput 6.00939K wps
[Epoch 26 Batch 1830/2125] avg loss 0.000926681, throughput 6.00669K wps
[Epoch 26 Batch 1860/2125] avg loss 0.000950341, throughput 6.01195K wps
[Epoch 26 Batch 1890/2125] avg loss 0.000901411, throughput 6.01847K wps
[Epoch 26 Batch 1920/2125] avg loss 0.00115569, throughput 6.01258K wps
[Epoch 26 Batch 1950/2125] avg loss 0.000791565, throughput 6.01356K wps
[Epoch 26 Batch 1980/2125] avg loss 0.00103533, throughput 6.00927K wps
[Epoch 26 Batch 2010/2125] avg loss 0.00089401, throughput 6.0132K wps
[Epoch 26 Batch 2040/2125] avg loss 0.000924346, throughput 6.01573K wps
[Epoch 26 Batch 2070/2125] avg loss 0.00102417, throughput 6.01027K wps
[Epoch 26 Batch 2100/2125] avg loss 0.000885243, throughput 6.01292K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 26] train avg loss 0.000781961, test acc 0.9265, test avg loss 0.447689, throughput 6.01529K wps
[Epoch 27 Batch 30/2125] avg loss 0.000688855, throughput 6.14417K wps
[Epoch 27 Batch 60/2125] avg loss 0.000537112, throughput 6.01447K wps
[Epoch 27 Batch 90/2125] avg loss 0.000564553, throughput 6.0036K wps
[Epoch 27 Batch 120/2125] avg loss 0.000534479, throughput 6.0064K wps
[Epoch 27 Batch 150/2125] avg loss 0.000530759, throughput 6.01454K wps
[Epoch 27 Batch 180/2125] avg loss 0.000596648, throughput 6.01297K wps
[Epoch 27 Batch 210/2125] avg loss 0.000648678, throughput 6.012K wps
[Epoch 27 Batch 240/2125] avg loss 0.000533177, throughput 6.01098K wps
[Epoch 27 Batch 270/2125] avg loss 0.000731812, throughput 6.00823K wps
[Epoch 27 Batch 300/2125] avg loss 0.000710919, throughput 6.01849K wps
[Epoch 27 Batch 330/2125] avg loss 0.000725028, throughput 6.00918K wps
[Epoch 27 Batch 360/2125] avg loss 0.00059991, throughput 6.00961K wps
[Epoch 27 Batch 390/2125] avg loss 0.000681704, throughput 6.0143K wps
[Epoch 27 Batch 420/2125] avg loss 0.00080549, throughput 6.01527K wps
[Epoch 27 Batch 450/2125] avg loss 0.000614105, throughput 6.02K wps
[Epoch 27 Batch 480/2125] avg loss 0.000737609, throughput 6.01382K wps
[Epoch 27 Batch 510/2125] avg loss 0.000703727, throughput 6.00506K wps
[Epoch 27 Batch 540/2125] avg loss 0.000613277, throughput 6.01122K wps
[Epoch 27 Batch 570/2125] avg loss 0.000918097, throughput 6.01257K wps
[Epoch 27 Batch 600/2125] avg loss 0.000701596, throughput 6.00913K wps
[Epoch 27 Batch 630/2125] avg loss 0.000739438, throughput 6.01162K wps
[Epoch 27 Batch 660/2125] avg loss 0.000595729, throughput 6.00597K wps
[Epoch 27 Batch 690/2125] avg loss 0.0009774, throughput 6.01255K wps
[Epoch 27 Batch 720/2125] avg loss 0.000485377, throughput 6.01423K wps
[Epoch 27 Batch 750/2125] avg loss 0.00066499, throughput 6.01115K wps
[Epoch 27 Batch 780/2125] avg loss 0.000960176, throughput 6.00673K wps
[Epoch 27 Batch 810/2125] avg loss 0.000674006, throughput 6.01124K wps
[Epoch 27 Batch 840/2125] avg loss 0.000686894, throughput 6.01247K wps
[Epoch 27 Batch 870/2125] avg loss 0.00087837, throughput 6.01798K wps
[Epoch 27 Batch 900/2125] avg loss 0.00071839, throughput 6.01697K wps
[Epoch 27 Batch 930/2125] avg loss 0.000683446, throughput 6.01584K wps
[Epoch 27 Batch 960/2125] avg loss 0.000762941, throughput 6.0147K wps
[Epoch 27 Batch 990/2125] avg loss 0.000571686, throughput 6.00109K wps
[Epoch 27 Batch 1020/2125] avg loss 0.000517739, throughput 6.01332K wps
[Epoch 27 Batch 1050/2125] avg loss 0.000891731, throughput 6.01727K wps
[Epoch 27 Batch 1080/2125] avg loss 0.000678282, throughput 6.00834K wps
[Epoch 27 Batch 1110/2125] avg loss 0.000655698, throughput 6.01548K wps
[Epoch 27 Batch 1140/2125] avg loss 0.000697578, throughput 6.00418K wps
[Epoch 27 Batch 1170/2125] avg loss 0.000946663, throughput 6.01692K wps
[Epoch 27 Batch 1200/2125] avg loss 0.000942827, throughput 6.0157K wps
[Epoch 27 Batch 1230/2125] avg loss 0.000730621, throughput 6.01412K wps
[Epoch 27 Batch 1260/2125] avg loss 0.00102974, throughput 6.0075K wps
[Epoch 27 Batch 1290/2125] avg loss 0.000674686, throughput 6.01354K wps
[Epoch 27 Batch 1320/2125] avg loss 0.000634021, throughput 6.01575K wps
[Epoch 27 Batch 1350/2125] avg loss 0.000849801, throughput 6.01144K wps
[Epoch 27 Batch 1380/2125] avg loss 0.000665925, throughput 6.01716K wps
[Epoch 27 Batch 1410/2125] avg loss 0.000850234, throughput 6.00761K wps
[Epoch 27 Batch 1440/2125] avg loss 0.0010447, throughput 6.00114K wps
[Epoch 27 Batch 1470/2125] avg loss 0.000973234, throughput 6.00415K wps
[Epoch 27 Batch 1500/2125] avg loss 0.000681446, throughput 6.00891K wps
[Epoch 27 Batch 1530/2125] avg loss 0.000676349, throughput 6.01738K wps
[Epoch 27 Batch 1560/2125] avg loss 0.000921224, throughput 6.02099K wps
[Epoch 27 Batch 1590/2125] avg loss 0.000807595, throughput 6.01754K wps
[Epoch 27 Batch 1620/2125] avg loss 0.00106917, throughput 6.0173K wps
[Epoch 27 Batch 1650/2125] avg loss 0.000728101, throughput 6.00053K wps
[Epoch 27 Batch 1680/2125] avg loss 0.0010972, throughput 6.01118K wps
[Epoch 27 Batch 1710/2125] avg loss 0.00082429, throughput 6.02185K wps
[Epoch 27 Batch 1740/2125] avg loss 0.00100423, throughput 6.02222K wps
[Epoch 27 Batch 1770/2125] avg loss 0.000832505, throughput 6.01849K wps
[Epoch 27 Batch 1800/2125] avg loss 0.00061203, throughput 6.0203K wps
[Epoch 27 Batch 1830/2125] avg loss 0.00108565, throughput 6.01364K wps
[Epoch 27 Batch 1860/2125] avg loss 0.00063851, throughput 6.03338K wps
[Epoch 27 Batch 1890/2125] avg loss 0.000617071, throughput 6.02571K wps
[Epoch 27 Batch 1920/2125] avg loss 0.000572293, throughput 6.00484K wps
[Epoch 27 Batch 1950/2125] avg loss 0.000838784, throughput 6.01509K wps
[Epoch 27 Batch 1980/2125] avg loss 0.000991486, throughput 6.01345K wps
[Epoch 27 Batch 2010/2125] avg loss 0.000853908, throughput 6.01494K wps
[Epoch 27 Batch 2040/2125] avg loss 0.000874445, throughput 6.02899K wps
[Epoch 27 Batch 2070/2125] avg loss 0.000960624, throughput 6.02167K wps
[Epoch 27 Batch 2100/2125] avg loss 0.000749336, throughput 6.02317K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 27] train avg loss 0.000755504, test acc 0.9259, test avg loss 0.454235, throughput 6.01554K wps
[Epoch 28 Batch 30/2125] avg loss 0.000652985, throughput 6.15064K wps
[Epoch 28 Batch 60/2125] avg loss 0.000585453, throughput 6.014K wps
[Epoch 28 Batch 90/2125] avg loss 0.000529279, throughput 6.01432K wps
[Epoch 28 Batch 120/2125] avg loss 0.000757707, throughput 6.02115K wps
[Epoch 28 Batch 150/2125] avg loss 0.000551183, throughput 6.01314K wps
[Epoch 28 Batch 180/2125] avg loss 0.000649029, throughput 6.00614K wps
[Epoch 28 Batch 210/2125] avg loss 0.000441257, throughput 5.99929K wps
[Epoch 28 Batch 240/2125] avg loss 0.000413088, throughput 6.01168K wps
[Epoch 28 Batch 270/2125] avg loss 0.000634083, throughput 6.00878K wps
[Epoch 28 Batch 300/2125] avg loss 0.000681087, throughput 6.01312K wps
[Epoch 28 Batch 330/2125] avg loss 0.000541111, throughput 6.01047K wps
[Epoch 28 Batch 360/2125] avg loss 0.000802887, throughput 6.00908K wps
[Epoch 28 Batch 390/2125] avg loss 0.000662646, throughput 6.00867K wps
[Epoch 28 Batch 420/2125] avg loss 0.00071305, throughput 6.01639K wps
[Epoch 28 Batch 450/2125] avg loss 0.000815833, throughput 6.00739K wps
[Epoch 28 Batch 480/2125] avg loss 0.000438958, throughput 6.01129K wps
[Epoch 28 Batch 510/2125] avg loss 0.000582378, throughput 6.01144K wps
[Epoch 28 Batch 540/2125] avg loss 0.000553454, throughput 6.00554K wps
[Epoch 28 Batch 570/2125] avg loss 0.000780062, throughput 6.01904K wps
[Epoch 28 Batch 600/2125] avg loss 0.000757877, throughput 6.01369K wps
[Epoch 28 Batch 630/2125] avg loss 0.000810768, throughput 6.01125K wps
[Epoch 28 Batch 660/2125] avg loss 0.000581669, throughput 6.01646K wps
[Epoch 28 Batch 690/2125] avg loss 0.000656247, throughput 6.00681K wps
[Epoch 28 Batch 720/2125] avg loss 0.000438711, throughput 6.00039K wps
[Epoch 28 Batch 750/2125] avg loss 0.000507503, throughput 6.00526K wps
[Epoch 28 Batch 780/2125] avg loss 0.000998637, throughput 6.00403K wps
[Epoch 28 Batch 810/2125] avg loss 0.000649149, throughput 6.01177K wps
[Epoch 28 Batch 840/2125] avg loss 0.000755024, throughput 6.01242K wps
[Epoch 28 Batch 870/2125] avg loss 0.000906334, throughput 6.01818K wps
[Epoch 28 Batch 900/2125] avg loss 0.000739633, throughput 6.01087K wps
[Epoch 28 Batch 930/2125] avg loss 0.000676357, throughput 6.00644K wps
[Epoch 28 Batch 960/2125] avg loss 0.000600721, throughput 6.00257K wps
[Epoch 28 Batch 990/2125] avg loss 0.000689061, throughput 6.00787K wps
[Epoch 28 Batch 1020/2125] avg loss 0.0006502, throughput 6.00163K wps
[Epoch 28 Batch 1050/2125] avg loss 0.000735226, throughput 6.00965K wps
[Epoch 28 Batch 1080/2125] avg loss 0.000726931, throughput 6.00763K wps
[Epoch 28 Batch 1110/2125] avg loss 0.000642691, throughput 6.01196K wps
[Epoch 28 Batch 1140/2125] avg loss 0.000650554, throughput 6.01293K wps
[Epoch 28 Batch 1170/2125] avg loss 0.00065096, throughput 6.00049K wps
[Epoch 28 Batch 1200/2125] avg loss 0.000921847, throughput 6.00926K wps
[Epoch 28 Batch 1230/2125] avg loss 0.000841361, throughput 6.00839K wps
[Epoch 28 Batch 1260/2125] avg loss 0.000762647, throughput 6.01883K wps
[Epoch 28 Batch 1290/2125] avg loss 0.00076074, throughput 6.02017K wps
[Epoch 28 Batch 1320/2125] avg loss 0.000539957, throughput 6.01987K wps
[Epoch 28 Batch 1350/2125] avg loss 0.000636336, throughput 6.01911K wps
[Epoch 28 Batch 1380/2125] avg loss 0.000858473, throughput 6.00234K wps
[Epoch 28 Batch 1410/2125] avg loss 0.000738765, throughput 6.01211K wps
[Epoch 28 Batch 1440/2125] avg loss 0.000507318, throughput 6.01435K wps
[Epoch 28 Batch 1470/2125] avg loss 0.00084414, throughput 6.01096K wps
[Epoch 28 Batch 1500/2125] avg loss 0.00100492, throughput 6.01174K wps
[Epoch 28 Batch 1530/2125] avg loss 0.000744925, throughput 6.01496K wps
[Epoch 28 Batch 1560/2125] avg loss 0.000621551, throughput 6.01056K wps
[Epoch 28 Batch 1590/2125] avg loss 0.000628667, throughput 6.01404K wps
[Epoch 28 Batch 1620/2125] avg loss 0.000498443, throughput 6.01617K wps
[Epoch 28 Batch 1650/2125] avg loss 0.000849961, throughput 6.01381K wps
[Epoch 28 Batch 1680/2125] avg loss 0.000847643, throughput 6.01586K wps
[Epoch 28 Batch 1710/2125] avg loss 0.000817519, throughput 6.01745K wps
[Epoch 28 Batch 1740/2125] avg loss 0.000739191, throughput 6.02343K wps
[Epoch 28 Batch 1770/2125] avg loss 0.000818391, throughput 6.02011K wps
[Epoch 28 Batch 1800/2125] avg loss 0.00081638, throughput 6.01545K wps
[Epoch 28 Batch 1830/2125] avg loss 0.000709638, throughput 6.02051K wps
[Epoch 28 Batch 1860/2125] avg loss 0.000738696, throughput 6.0178K wps
[Epoch 28 Batch 1890/2125] avg loss 0.000942373, throughput 6.02372K wps
[Epoch 28 Batch 1920/2125] avg loss 0.000716514, throughput 6.01556K wps
[Epoch 28 Batch 1950/2125] avg loss 0.000908092, throughput 6.01108K wps
[Epoch 28 Batch 1980/2125] avg loss 0.00124693, throughput 6.01636K wps
[Epoch 28 Batch 2010/2125] avg loss 0.000779676, throughput 6.02409K wps
[Epoch 28 Batch 2040/2125] avg loss 0.000901963, throughput 6.02043K wps
[Epoch 28 Batch 2070/2125] avg loss 0.000735962, throughput 6.02338K wps
[Epoch 28 Batch 2100/2125] avg loss 0.000843925, throughput 6.0162K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.29 s
[Batch 120/237] elapsed 0.28 s
[Batch 150/237] elapsed 0.28 s
[Batch 180/237] elapsed 0.29 s
[Batch 210/237] elapsed 0.30 s
[Epoch 28] train avg loss 0.000716701, test acc 0.9254, test avg loss 0.469635, throughput 6.01457K wps
[Epoch 29 Batch 30/2125] avg loss 0.000789425, throughput 6.13696K wps
[Epoch 29 Batch 60/2125] avg loss 0.000645757, throughput 6.00907K wps
[Epoch 29 Batch 90/2125] avg loss 0.000581918, throughput 6.00664K wps
[Epoch 29 Batch 120/2125] avg loss 0.000446242, throughput 6.0182K wps
[Epoch 29 Batch 150/2125] avg loss 0.000458789, throughput 6.01743K wps
[Epoch 29 Batch 180/2125] avg loss 0.000615038, throughput 6.02644K wps
[Epoch 29 Batch 210/2125] avg loss 0.000718819, throughput 6.01327K wps
[Epoch 29 Batch 240/2125] avg loss 0.000558854, throughput 6.00143K wps
[Epoch 29 Batch 270/2125] avg loss 0.000587444, throughput 6.02352K wps
[Epoch 29 Batch 300/2125] avg loss 0.000555708, throughput 6.02179K wps
[Epoch 29 Batch 330/2125] avg loss 0.000543366, throughput 6.02209K wps
[Epoch 29 Batch 360/2125] avg loss 0.000598584, throughput 6.0273K wps
[Epoch 29 Batch 390/2125] avg loss 0.000846851, throughput 6.01397K wps
[Epoch 29 Batch 420/2125] avg loss 0.000499227, throughput 6.02105K wps
[Epoch 29 Batch 450/2125] avg loss 0.000642242, throughput 6.01782K wps
[Epoch 29 Batch 480/2125] avg loss 0.000736861, throughput 6.02049K wps
[Epoch 29 Batch 510/2125] avg loss 0.000679866, throughput 6.02139K wps
[Epoch 29 Batch 540/2125] avg loss 0.000429393, throughput 6.01674K wps
[Epoch 29 Batch 570/2125] avg loss 0.000665637, throughput 6.0189K wps
[Epoch 29 Batch 600/2125] avg loss 0.000566837, throughput 6.01665K wps
[Epoch 29 Batch 630/2125] avg loss 0.000594157, throughput 6.0076K wps
[Epoch 29 Batch 660/2125] avg loss 0.000808534, throughput 6.01572K wps
[Epoch 29 Batch 690/2125] avg loss 0.000591572, throughput 6.015K wps
[Epoch 29 Batch 720/2125] avg loss 0.000785182, throughput 6.0128K wps
[Epoch 29 Batch 750/2125] avg loss 0.000673816, throughput 6.01523K wps
[Epoch 29 Batch 780/2125] avg loss 0.000454437, throughput 6.01319K wps
[Epoch 29 Batch 810/2125] avg loss 0.000816988, throughput 6.00862K wps
[Epoch 29 Batch 840/2125] avg loss 0.00074785, throughput 6.01999K wps
[Epoch 29 Batch 870/2125] avg loss 0.000545004, throughput 6.01595K wps
[Epoch 29 Batch 900/2125] avg loss 0.000640378, throughput 6.0129K wps
[Epoch 29 Batch 930/2125] avg loss 0.000711453, throughput 6.01982K wps
[Epoch 29 Batch 960/2125] avg loss 0.000522378, throughput 6.0054K wps
[Epoch 29 Batch 990/2125] avg loss 0.000724593, throughput 6.01878K wps
[Epoch 29 Batch 1020/2125] avg loss 0.000653246, throughput 6.02234K wps
[Epoch 29 Batch 1050/2125] avg loss 0.000706048, throughput 6.01755K wps
[Epoch 29 Batch 1080/2125] avg loss 0.000866256, throughput 6.02079K wps
[Epoch 29 Batch 1110/2125] avg loss 0.000648801, throughput 6.02531K wps
[Epoch 29 Batch 1140/2125] avg loss 0.000794723, throughput 6.01821K wps
[Epoch 29 Batch 1170/2125] avg loss 0.000658505, throughput 6.018K wps
[Epoch 29 Batch 1200/2125] avg loss 0.000583408, throughput 6.01358K wps
[Epoch 29 Batch 1230/2125] avg loss 0.000803476, throughput 6.0143K wps
[Epoch 29 Batch 1260/2125] avg loss 0.000876664, throughput 6.0268K wps
[Epoch 29 Batch 1290/2125] avg loss 0.000773859, throughput 6.01419K wps
[Epoch 29 Batch 1320/2125] avg loss 0.000730729, throughput 6.01604K wps
[Epoch 29 Batch 1350/2125] avg loss 0.000633321, throughput 6.01681K wps
[Epoch 29 Batch 1380/2125] avg loss 0.000803545, throughput 6.02355K wps
[Epoch 29 Batch 1410/2125] avg loss 0.000724541, throughput 6.01652K wps
[Epoch 29 Batch 1440/2125] avg loss 0.000660004, throughput 6.01012K wps
[Epoch 29 Batch 1470/2125] avg loss 0.000851641, throughput 6.00933K wps
[Epoch 29 Batch 1500/2125] avg loss 0.000726226, throughput 6.01759K wps
[Epoch 29 Batch 1530/2125] avg loss 0.000517603, throughput 6.01618K wps
[Epoch 29 Batch 1560/2125] avg loss 0.000987538, throughput 6.01225K wps
[Epoch 29 Batch 1590/2125] avg loss 0.000760709, throughput 6.01146K wps
[Epoch 29 Batch 1620/2125] avg loss 0.000835697, throughput 6.01341K wps
[Epoch 29 Batch 1650/2125] avg loss 0.00084539, throughput 6.01865K wps
[Epoch 29 Batch 1680/2125] avg loss 0.000777263, throughput 6.02016K wps
[Epoch 29 Batch 1710/2125] avg loss 0.000772582, throughput 6.02335K wps
[Epoch 29 Batch 1740/2125] avg loss 0.000991925, throughput 6.01962K wps
[Epoch 29 Batch 1770/2125] avg loss 0.00110307, throughput 6.01695K wps
[Epoch 29 Batch 1800/2125] avg loss 0.000735052, throughput 6.02128K wps
[Epoch 29 Batch 1830/2125] avg loss 0.00105085, throughput 6.01549K wps
[Epoch 29 Batch 1860/2125] avg loss 0.000758522, throughput 6.01212K wps
[Epoch 29 Batch 1890/2125] avg loss 0.000567539, throughput 6.02064K wps
[Epoch 29 Batch 1920/2125] avg loss 0.000978181, throughput 6.02807K wps
[Epoch 29 Batch 1950/2125] avg loss 0.000652792, throughput 6.01523K wps
[Epoch 29 Batch 1980/2125] avg loss 0.000842137, throughput 6.02319K wps
[Epoch 29 Batch 2010/2125] avg loss 0.000635, throughput 6.01711K wps
[Epoch 29 Batch 2040/2125] avg loss 0.000868212, throughput 6.01226K wps
[Epoch 29 Batch 2070/2125] avg loss 0.000666248, throughput 6.01367K wps
[Epoch 29 Batch 2100/2125] avg loss 0.0008764, throughput 6.02041K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 29] train avg loss 0.000708242, test acc 0.9254, test avg loss 0.477164, throughput 6.01859K wps
[Epoch 30 Batch 30/2125] avg loss 0.000408318, throughput 6.15579K wps
[Epoch 30 Batch 60/2125] avg loss 0.000543177, throughput 6.00818K wps
[Epoch 30 Batch 90/2125] avg loss 0.000676124, throughput 6.01811K wps
[Epoch 30 Batch 120/2125] avg loss 0.000553361, throughput 6.00957K wps
[Epoch 30 Batch 150/2125] avg loss 0.000636714, throughput 6.0186K wps
[Epoch 30 Batch 180/2125] avg loss 0.000596402, throughput 6.02326K wps
[Epoch 30 Batch 210/2125] avg loss 0.000619897, throughput 6.01557K wps
[Epoch 30 Batch 240/2125] avg loss 0.000735503, throughput 6.0171K wps
[Epoch 30 Batch 270/2125] avg loss 0.000606803, throughput 6.0297K wps
[Epoch 30 Batch 300/2125] avg loss 0.000717699, throughput 6.01564K wps
[Epoch 30 Batch 330/2125] avg loss 0.000749229, throughput 6.02226K wps
[Epoch 30 Batch 360/2125] avg loss 0.000511468, throughput 6.02174K wps
[Epoch 30 Batch 390/2125] avg loss 0.000558803, throughput 6.01443K wps
[Epoch 30 Batch 420/2125] avg loss 0.000638625, throughput 6.01841K wps
[Epoch 30 Batch 450/2125] avg loss 0.00047411, throughput 6.00932K wps
[Epoch 30 Batch 480/2125] avg loss 0.0006531, throughput 6.02596K wps
[Epoch 30 Batch 510/2125] avg loss 0.000645494, throughput 6.01428K wps
[Epoch 30 Batch 540/2125] avg loss 0.000524001, throughput 6.0199K wps
[Epoch 30 Batch 570/2125] avg loss 0.000508976, throughput 6.02805K wps
[Epoch 30 Batch 600/2125] avg loss 0.000567181, throughput 6.0202K wps
[Epoch 30 Batch 630/2125] avg loss 0.000559612, throughput 6.01726K wps
[Epoch 30 Batch 660/2125] avg loss 0.000633111, throughput 6.01472K wps
[Epoch 30 Batch 690/2125] avg loss 0.000942358, throughput 6.01022K wps
[Epoch 30 Batch 720/2125] avg loss 0.000562323, throughput 6.02148K wps
[Epoch 30 Batch 750/2125] avg loss 0.000692842, throughput 6.02622K wps
[Epoch 30 Batch 780/2125] avg loss 0.000628181, throughput 6.01795K wps
[Epoch 30 Batch 810/2125] avg loss 0.000538982, throughput 6.00342K wps
[Epoch 30 Batch 840/2125] avg loss 0.000461209, throughput 6.00552K wps
[Epoch 30 Batch 870/2125] avg loss 0.000658543, throughput 5.98827K wps
[Epoch 30 Batch 900/2125] avg loss 0.000666378, throughput 6.0054K wps
[Epoch 30 Batch 930/2125] avg loss 0.000677583, throughput 6.00853K wps
[Epoch 30 Batch 960/2125] avg loss 0.000670944, throughput 6.01185K wps
[Epoch 30 Batch 990/2125] avg loss 0.000560314, throughput 6.00291K wps
[Epoch 30 Batch 1020/2125] avg loss 0.000811236, throughput 6.01527K wps
[Epoch 30 Batch 1050/2125] avg loss 0.00069372, throughput 6.00895K wps
[Epoch 30 Batch 1080/2125] avg loss 0.000599903, throughput 6.00733K wps
[Epoch 30 Batch 1110/2125] avg loss 0.00080983, throughput 6.0068K wps
[Epoch 30 Batch 1140/2125] avg loss 0.000683367, throughput 6.00484K wps
[Epoch 30 Batch 1170/2125] avg loss 0.00046915, throughput 6.00654K wps
[Epoch 30 Batch 1200/2125] avg loss 0.000881885, throughput 6.00594K wps
[Epoch 30 Batch 1230/2125] avg loss 0.00081683, throughput 6.00748K wps
[Epoch 30 Batch 1260/2125] avg loss 0.000553104, throughput 6.01147K wps
[Epoch 30 Batch 1290/2125] avg loss 0.000708445, throughput 6.01036K wps
[Epoch 30 Batch 1320/2125] avg loss 0.000840633, throughput 6.00634K wps
[Epoch 30 Batch 1350/2125] avg loss 0.000913818, throughput 6.01459K wps
[Epoch 30 Batch 1380/2125] avg loss 0.000797624, throughput 6.01207K wps
[Epoch 30 Batch 1410/2125] avg loss 0.00064481, throughput 6.0021K wps
[Epoch 30 Batch 1440/2125] avg loss 0.000915911, throughput 6.00909K wps
[Epoch 30 Batch 1470/2125] avg loss 0.000751459, throughput 6.01188K wps
[Epoch 30 Batch 1500/2125] avg loss 0.000498891, throughput 6.00521K wps
[Epoch 30 Batch 1530/2125] avg loss 0.00053709, throughput 6.00268K wps
[Epoch 30 Batch 1560/2125] avg loss 0.000813883, throughput 6.0068K wps
[Epoch 30 Batch 1590/2125] avg loss 0.000863406, throughput 6.01456K wps
[Epoch 30 Batch 1620/2125] avg loss 0.000607705, throughput 6.00619K wps
[Epoch 30 Batch 1650/2125] avg loss 0.000846425, throughput 6.0192K wps
[Epoch 30 Batch 1680/2125] avg loss 0.000781885, throughput 6.01993K wps
[Epoch 30 Batch 1710/2125] avg loss 0.000654774, throughput 6.01641K wps
[Epoch 30 Batch 1740/2125] avg loss 0.000864521, throughput 6.02482K wps
[Epoch 30 Batch 1770/2125] avg loss 0.000597843, throughput 6.01712K wps
[Epoch 30 Batch 1800/2125] avg loss 0.000983231, throughput 6.00744K wps
[Epoch 30 Batch 1830/2125] avg loss 0.000813255, throughput 6.01766K wps
[Epoch 30 Batch 1860/2125] avg loss 0.000839037, throughput 6.02222K wps
[Epoch 30 Batch 1890/2125] avg loss 0.000674818, throughput 6.01643K wps
[Epoch 30 Batch 1920/2125] avg loss 0.000622968, throughput 6.01222K wps
[Epoch 30 Batch 1950/2125] avg loss 0.000800402, throughput 6.00764K wps
[Epoch 30 Batch 1980/2125] avg loss 0.000650672, throughput 6.01293K wps
[Epoch 30 Batch 2010/2125] avg loss 0.000793957, throughput 6.01726K wps
[Epoch 30 Batch 2040/2125] avg loss 0.000646945, throughput 6.01553K wps
[Epoch 30 Batch 2070/2125] avg loss 0.000908168, throughput 6.01352K wps
[Epoch 30 Batch 2100/2125] avg loss 0.000834982, throughput 6.03285K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 30] train avg loss 0.000682547, test acc 0.9255, test avg loss 0.478963, throughput 6.01565K wps
[Epoch 31 Batch 30/2125] avg loss 0.00052264, throughput 6.15303K wps
[Epoch 31 Batch 60/2125] avg loss 0.000494237, throughput 6.01795K wps
[Epoch 31 Batch 90/2125] avg loss 0.000625686, throughput 6.01979K wps
[Epoch 31 Batch 120/2125] avg loss 0.000495298, throughput 6.01403K wps
[Epoch 31 Batch 150/2125] avg loss 0.000556815, throughput 6.01568K wps
[Epoch 31 Batch 180/2125] avg loss 0.000551182, throughput 6.01511K wps
[Epoch 31 Batch 210/2125] avg loss 0.000649478, throughput 6.00925K wps
[Epoch 31 Batch 240/2125] avg loss 0.000581145, throughput 6.00792K wps
[Epoch 31 Batch 270/2125] avg loss 0.000577683, throughput 6.01997K wps
[Epoch 31 Batch 300/2125] avg loss 0.000580381, throughput 6.00357K wps
[Epoch 31 Batch 330/2125] avg loss 0.00052845, throughput 6.01536K wps
[Epoch 31 Batch 360/2125] avg loss 0.000467264, throughput 6.01532K wps
[Epoch 31 Batch 390/2125] avg loss 0.000621706, throughput 6.01698K wps
[Epoch 31 Batch 420/2125] avg loss 0.000620819, throughput 6.00141K wps
[Epoch 31 Batch 450/2125] avg loss 0.000586773, throughput 6.01155K wps
[Epoch 31 Batch 480/2125] avg loss 0.00050997, throughput 6.01038K wps
[Epoch 31 Batch 510/2125] avg loss 0.000510021, throughput 6.0094K wps
[Epoch 31 Batch 540/2125] avg loss 0.000834522, throughput 6.01555K wps
[Epoch 31 Batch 570/2125] avg loss 0.000549193, throughput 6.01416K wps
[Epoch 31 Batch 600/2125] avg loss 0.000669813, throughput 6.02729K wps
[Epoch 31 Batch 630/2125] avg loss 0.000774181, throughput 6.02218K wps
[Epoch 31 Batch 660/2125] avg loss 0.000615948, throughput 6.00407K wps
[Epoch 31 Batch 690/2125] avg loss 0.000545713, throughput 6.02069K wps
[Epoch 31 Batch 720/2125] avg loss 0.000501098, throughput 6.018K wps
[Epoch 31 Batch 750/2125] avg loss 0.000616352, throughput 6.01592K wps
[Epoch 31 Batch 780/2125] avg loss 0.000756492, throughput 6.02552K wps
[Epoch 31 Batch 810/2125] avg loss 0.000660723, throughput 6.02447K wps
[Epoch 31 Batch 840/2125] avg loss 0.000618211, throughput 6.01702K wps
[Epoch 31 Batch 870/2125] avg loss 0.000696506, throughput 6.01571K wps
[Epoch 31 Batch 900/2125] avg loss 0.000511351, throughput 6.01617K wps
[Epoch 31 Batch 930/2125] avg loss 0.000594551, throughput 6.00967K wps
[Epoch 31 Batch 960/2125] avg loss 0.00062294, throughput 6.0075K wps
[Epoch 31 Batch 990/2125] avg loss 0.000712211, throughput 6.00023K wps
[Epoch 31 Batch 1020/2125] avg loss 0.000551513, throughput 6.00154K wps
[Epoch 31 Batch 1050/2125] avg loss 0.000896425, throughput 5.999K wps
[Epoch 31 Batch 1080/2125] avg loss 0.000872409, throughput 6.01929K wps
[Epoch 31 Batch 1110/2125] avg loss 0.000645092, throughput 6.01206K wps
[Epoch 31 Batch 1140/2125] avg loss 0.000478724, throughput 6.02177K wps
[Epoch 31 Batch 1170/2125] avg loss 0.000623312, throughput 6.02237K wps
[Epoch 31 Batch 1200/2125] avg loss 0.000605625, throughput 6.00683K wps
[Epoch 31 Batch 1230/2125] avg loss 0.000680051, throughput 6.01138K wps
[Epoch 31 Batch 1260/2125] avg loss 0.000859788, throughput 6.0264K wps
[Epoch 31 Batch 1290/2125] avg loss 0.000637198, throughput 6.01835K wps
[Epoch 31 Batch 1320/2125] avg loss 0.000721791, throughput 6.02843K wps
[Epoch 31 Batch 1350/2125] avg loss 0.00063052, throughput 6.02226K wps
[Epoch 31 Batch 1380/2125] avg loss 0.000691143, throughput 6.0357K wps
[Epoch 31 Batch 1410/2125] avg loss 0.000752613, throughput 6.02127K wps
[Epoch 31 Batch 1440/2125] avg loss 0.00087562, throughput 6.02762K wps
[Epoch 31 Batch 1470/2125] avg loss 0.000934925, throughput 6.01116K wps
[Epoch 31 Batch 1500/2125] avg loss 0.000846471, throughput 6.01485K wps
[Epoch 31 Batch 1530/2125] avg loss 0.000812978, throughput 6.00892K wps
[Epoch 31 Batch 1560/2125] avg loss 0.00070314, throughput 6.01436K wps
[Epoch 31 Batch 1590/2125] avg loss 0.000768871, throughput 6.01459K wps
[Epoch 31 Batch 1620/2125] avg loss 0.000550724, throughput 6.01928K wps
[Epoch 31 Batch 1650/2125] avg loss 0.000939346, throughput 6.01153K wps
[Epoch 31 Batch 1680/2125] avg loss 0.00062201, throughput 6.01231K wps
[Epoch 31 Batch 1710/2125] avg loss 0.000826567, throughput 6.0118K wps
[Epoch 31 Batch 1740/2125] avg loss 0.000888527, throughput 6.01208K wps
[Epoch 31 Batch 1770/2125] avg loss 0.000755578, throughput 6.01521K wps
[Epoch 31 Batch 1800/2125] avg loss 0.000623762, throughput 6.01476K wps
[Epoch 31 Batch 1830/2125] avg loss 0.000739893, throughput 6.00535K wps
[Epoch 31 Batch 1860/2125] avg loss 0.000678825, throughput 6.01701K wps
[Epoch 31 Batch 1890/2125] avg loss 0.000804798, throughput 6.00715K wps
[Epoch 31 Batch 1920/2125] avg loss 0.000726188, throughput 6.0251K wps
[Epoch 31 Batch 1950/2125] avg loss 0.000575069, throughput 6.02473K wps
[Epoch 31 Batch 1980/2125] avg loss 0.000881537, throughput 6.0122K wps
[Epoch 31 Batch 2010/2125] avg loss 0.000732232, throughput 6.00843K wps
[Epoch 31 Batch 2040/2125] avg loss 0.000806634, throughput 6.01779K wps
[Epoch 31 Batch 2070/2125] avg loss 0.000852313, throughput 6.01068K wps
[Epoch 31 Batch 2100/2125] avg loss 0.000488458, throughput 6.02657K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 31] train avg loss 0.000670571, test acc 0.9257, test avg loss 0.48574, throughput 6.01712K wps
[Epoch 32 Batch 30/2125] avg loss 0.000496184, throughput 6.15626K wps
[Epoch 32 Batch 60/2125] avg loss 0.000558525, throughput 6.01356K wps
[Epoch 32 Batch 90/2125] avg loss 0.000622697, throughput 6.01224K wps
[Epoch 32 Batch 120/2125] avg loss 0.000540502, throughput 6.01998K wps
[Epoch 32 Batch 150/2125] avg loss 0.000537749, throughput 6.0162K wps
[Epoch 32 Batch 180/2125] avg loss 0.000534343, throughput 6.00576K wps
[Epoch 32 Batch 210/2125] avg loss 0.000537163, throughput 5.99891K wps
[Epoch 32 Batch 240/2125] avg loss 0.000641625, throughput 6.00285K wps
[Epoch 32 Batch 270/2125] avg loss 0.000458067, throughput 6.01908K wps
[Epoch 32 Batch 300/2125] avg loss 0.00049918, throughput 6.0281K wps
[Epoch 32 Batch 330/2125] avg loss 0.000477373, throughput 6.01781K wps
[Epoch 32 Batch 360/2125] avg loss 0.000715142, throughput 6.01363K wps
[Epoch 32 Batch 390/2125] avg loss 0.000394898, throughput 6.01633K wps
[Epoch 32 Batch 420/2125] avg loss 0.000723061, throughput 6.01717K wps
[Epoch 32 Batch 450/2125] avg loss 0.000692485, throughput 6.02562K wps
[Epoch 32 Batch 480/2125] avg loss 0.000529428, throughput 6.01007K wps
[Epoch 32 Batch 510/2125] avg loss 0.000484905, throughput 6.01052K wps
[Epoch 32 Batch 540/2125] avg loss 0.00049419, throughput 6.01842K wps
[Epoch 32 Batch 570/2125] avg loss 0.000448736, throughput 6.00531K wps
[Epoch 32 Batch 600/2125] avg loss 0.000421037, throughput 6.02198K wps
[Epoch 32 Batch 630/2125] avg loss 0.000614088, throughput 6.02264K wps
[Epoch 32 Batch 660/2125] avg loss 0.000591461, throughput 6.02572K wps
[Epoch 32 Batch 690/2125] avg loss 0.000435738, throughput 6.01254K wps
[Epoch 32 Batch 720/2125] avg loss 0.000614504, throughput 6.01491K wps
[Epoch 32 Batch 750/2125] avg loss 0.000417642, throughput 6.01754K wps
[Epoch 32 Batch 780/2125] avg loss 0.000500724, throughput 6.01403K wps
[Epoch 32 Batch 810/2125] avg loss 0.000527586, throughput 6.00405K wps
[Epoch 32 Batch 840/2125] avg loss 0.0007102, throughput 6.01674K wps
[Epoch 32 Batch 870/2125] avg loss 0.000451559, throughput 6.02393K wps
[Epoch 32 Batch 900/2125] avg loss 0.000584525, throughput 6.01444K wps
[Epoch 32 Batch 930/2125] avg loss 0.000615148, throughput 6.02139K wps
[Epoch 32 Batch 960/2125] avg loss 0.000547572, throughput 6.02295K wps
[Epoch 32 Batch 990/2125] avg loss 0.00047706, throughput 6.01819K wps
[Epoch 32 Batch 1020/2125] avg loss 0.000698309, throughput 6.01161K wps
[Epoch 32 Batch 1050/2125] avg loss 0.000675559, throughput 6.01101K wps
[Epoch 32 Batch 1080/2125] avg loss 0.000759682, throughput 6.01722K wps
[Epoch 32 Batch 1110/2125] avg loss 0.000474633, throughput 6.01889K wps
[Epoch 32 Batch 1140/2125] avg loss 0.000626398, throughput 6.02444K wps
[Epoch 32 Batch 1170/2125] avg loss 0.00079957, throughput 6.01694K wps
[Epoch 32 Batch 1200/2125] avg loss 0.000773996, throughput 6.01K wps
[Epoch 32 Batch 1230/2125] avg loss 0.000898933, throughput 6.02093K wps
[Epoch 32 Batch 1260/2125] avg loss 0.000702933, throughput 5.99958K wps
[Epoch 32 Batch 1290/2125] avg loss 0.000781662, throughput 6.02262K wps
[Epoch 32 Batch 1320/2125] avg loss 0.000527773, throughput 6.01599K wps
[Epoch 32 Batch 1350/2125] avg loss 0.00058435, throughput 6.01739K wps
[Epoch 32 Batch 1380/2125] avg loss 0.000675593, throughput 6.02105K wps
[Epoch 32 Batch 1410/2125] avg loss 0.000767952, throughput 6.00868K wps
[Epoch 32 Batch 1440/2125] avg loss 0.000645237, throughput 6.01087K wps
[Epoch 32 Batch 1470/2125] avg loss 0.000560413, throughput 6.0244K wps
[Epoch 32 Batch 1500/2125] avg loss 0.000797828, throughput 6.01697K wps
[Epoch 32 Batch 1530/2125] avg loss 0.000929923, throughput 6.01098K wps
[Epoch 32 Batch 1560/2125] avg loss 0.000726379, throughput 6.00762K wps
[Epoch 32 Batch 1590/2125] avg loss 0.000867345, throughput 6.0181K wps
[Epoch 32 Batch 1620/2125] avg loss 0.000635559, throughput 6.02126K wps
[Epoch 32 Batch 1650/2125] avg loss 0.000578466, throughput 6.01059K wps
[Epoch 32 Batch 1680/2125] avg loss 0.000693019, throughput 6.02462K wps
[Epoch 32 Batch 1710/2125] avg loss 0.00085532, throughput 6.0081K wps
[Epoch 32 Batch 1740/2125] avg loss 0.000736983, throughput 6.01585K wps
[Epoch 32 Batch 1770/2125] avg loss 0.000578091, throughput 6.01913K wps
[Epoch 32 Batch 1800/2125] avg loss 0.000757684, throughput 6.0163K wps
[Epoch 32 Batch 1830/2125] avg loss 0.000697875, throughput 6.02065K wps
[Epoch 32 Batch 1860/2125] avg loss 0.000659345, throughput 6.02257K wps
[Epoch 32 Batch 1890/2125] avg loss 0.000877323, throughput 6.00653K wps
[Epoch 32 Batch 1920/2125] avg loss 0.000648036, throughput 6.01932K wps
[Epoch 32 Batch 1950/2125] avg loss 0.000495325, throughput 6.00836K wps
[Epoch 32 Batch 1980/2125] avg loss 0.000746669, throughput 6.02004K wps
[Epoch 32 Batch 2010/2125] avg loss 0.000643451, throughput 6.01874K wps
[Epoch 32 Batch 2040/2125] avg loss 0.000883268, throughput 6.0082K wps
[Epoch 32 Batch 2070/2125] avg loss 0.000685272, throughput 6.01762K wps
[Epoch 32 Batch 2100/2125] avg loss 0.000805963, throughput 6.01983K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 32] train avg loss 0.000634932, test acc 0.9245, test avg loss 0.495249, throughput 6.01772K wps
[Epoch 33 Batch 30/2125] avg loss 0.00059043, throughput 6.14085K wps
[Epoch 33 Batch 60/2125] avg loss 0.000627103, throughput 6.01351K wps
[Epoch 33 Batch 90/2125] avg loss 0.000413329, throughput 6.00927K wps
[Epoch 33 Batch 120/2125] avg loss 0.00056645, throughput 6.0166K wps
[Epoch 33 Batch 150/2125] avg loss 0.000478819, throughput 6.01338K wps
[Epoch 33 Batch 180/2125] avg loss 0.000482331, throughput 6.0057K wps
[Epoch 33 Batch 210/2125] avg loss 0.000479504, throughput 6.01285K wps
[Epoch 33 Batch 240/2125] avg loss 0.000524419, throughput 6.02085K wps
[Epoch 33 Batch 270/2125] avg loss 0.000495416, throughput 6.02041K wps
[Epoch 33 Batch 300/2125] avg loss 0.000513743, throughput 6.0024K wps
[Epoch 33 Batch 330/2125] avg loss 0.000495744, throughput 6.01786K wps
[Epoch 33 Batch 360/2125] avg loss 0.000563229, throughput 6.01266K wps
[Epoch 33 Batch 390/2125] avg loss 0.000338148, throughput 6.01558K wps
[Epoch 33 Batch 420/2125] avg loss 0.000559212, throughput 6.0083K wps
[Epoch 33 Batch 450/2125] avg loss 0.000523012, throughput 6.00951K wps
[Epoch 33 Batch 480/2125] avg loss 0.000531719, throughput 6.02122K wps
[Epoch 33 Batch 510/2125] avg loss 0.000507449, throughput 6.01828K wps
[Epoch 33 Batch 540/2125] avg loss 0.000631633, throughput 6.01657K wps
[Epoch 33 Batch 570/2125] avg loss 0.000538151, throughput 6.01899K wps
[Epoch 33 Batch 600/2125] avg loss 0.00045673, throughput 6.02353K wps
[Epoch 33 Batch 630/2125] avg loss 0.00053725, throughput 6.01983K wps
[Epoch 33 Batch 660/2125] avg loss 0.000446499, throughput 6.01992K wps
[Epoch 33 Batch 690/2125] avg loss 0.000473488, throughput 6.01744K wps
[Epoch 33 Batch 720/2125] avg loss 0.000497968, throughput 6.02275K wps
[Epoch 33 Batch 750/2125] avg loss 0.000456823, throughput 6.01745K wps
[Epoch 33 Batch 780/2125] avg loss 0.00060339, throughput 6.02487K wps
[Epoch 33 Batch 810/2125] avg loss 0.00057335, throughput 6.02708K wps
[Epoch 33 Batch 840/2125] avg loss 0.000831872, throughput 6.01324K wps
[Epoch 33 Batch 870/2125] avg loss 0.000439588, throughput 6.01003K wps
[Epoch 33 Batch 900/2125] avg loss 0.000643511, throughput 6.0123K wps
[Epoch 33 Batch 930/2125] avg loss 0.000384268, throughput 6.02138K wps
[Epoch 33 Batch 960/2125] avg loss 0.000613229, throughput 6.0149K wps
[Epoch 33 Batch 990/2125] avg loss 0.000735673, throughput 6.02688K wps
[Epoch 33 Batch 1020/2125] avg loss 0.000481716, throughput 6.01152K wps
[Epoch 33 Batch 1050/2125] avg loss 0.000572609, throughput 6.02132K wps
[Epoch 33 Batch 1080/2125] avg loss 0.000541297, throughput 6.02317K wps
[Epoch 33 Batch 1110/2125] avg loss 0.00057833, throughput 6.01418K wps
[Epoch 33 Batch 1140/2125] avg loss 0.000605059, throughput 6.01305K wps
[Epoch 33 Batch 1170/2125] avg loss 0.000760499, throughput 6.0202K wps
[Epoch 33 Batch 1200/2125] avg loss 0.000678698, throughput 6.01548K wps
[Epoch 33 Batch 1230/2125] avg loss 0.000687359, throughput 6.02228K wps
[Epoch 33 Batch 1260/2125] avg loss 0.000590909, throughput 6.02208K wps
[Epoch 33 Batch 1290/2125] avg loss 0.000625023, throughput 6.01575K wps
[Epoch 33 Batch 1320/2125] avg loss 0.000609067, throughput 6.02388K wps
[Epoch 33 Batch 1350/2125] avg loss 0.000620318, throughput 6.0194K wps
[Epoch 33 Batch 1380/2125] avg loss 0.000716272, throughput 6.01312K wps
[Epoch 33 Batch 1410/2125] avg loss 0.000785574, throughput 6.01365K wps
[Epoch 33 Batch 1440/2125] avg loss 0.000789248, throughput 6.0243K wps
[Epoch 33 Batch 1470/2125] avg loss 0.000613298, throughput 6.01963K wps
[Epoch 33 Batch 1500/2125] avg loss 0.000691636, throughput 6.0191K wps
[Epoch 33 Batch 1530/2125] avg loss 0.000756811, throughput 6.0223K wps
[Epoch 33 Batch 1560/2125] avg loss 0.000688654, throughput 6.01354K wps
[Epoch 33 Batch 1590/2125] avg loss 0.000701075, throughput 6.01918K wps
[Epoch 33 Batch 1620/2125] avg loss 0.000892637, throughput 6.01751K wps
[Epoch 33 Batch 1650/2125] avg loss 0.000479701, throughput 6.01872K wps
[Epoch 33 Batch 1680/2125] avg loss 0.000859681, throughput 6.0253K wps
[Epoch 33 Batch 1710/2125] avg loss 0.000712761, throughput 6.01574K wps
[Epoch 33 Batch 1740/2125] avg loss 0.000629237, throughput 6.01973K wps
[Epoch 33 Batch 1770/2125] avg loss 0.000704793, throughput 6.0274K wps
[Epoch 33 Batch 1800/2125] avg loss 0.000618274, throughput 6.02013K wps
[Epoch 33 Batch 1830/2125] avg loss 0.000828679, throughput 6.02493K wps
[Epoch 33 Batch 1860/2125] avg loss 0.000467051, throughput 6.00942K wps
[Epoch 33 Batch 1890/2125] avg loss 0.000539805, throughput 6.00347K wps
[Epoch 33 Batch 1920/2125] avg loss 0.000501023, throughput 6.016K wps
[Epoch 33 Batch 1950/2125] avg loss 0.000745058, throughput 6.01681K wps
[Epoch 33 Batch 1980/2125] avg loss 0.000513568, throughput 6.01812K wps
[Epoch 33 Batch 2010/2125] avg loss 0.000977539, throughput 6.01034K wps
[Epoch 33 Batch 2040/2125] avg loss 0.000520984, throughput 6.0093K wps
[Epoch 33 Batch 2070/2125] avg loss 0.000505899, throughput 5.99134K wps
[Epoch 33 Batch 2100/2125] avg loss 0.000889527, throughput 5.99984K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 33] train avg loss 0.000599575, test acc 0.9250, test avg loss 0.505249, throughput 6.01831K wps
[Epoch 34 Batch 30/2125] avg loss 0.00041241, throughput 6.16221K wps
[Epoch 34 Batch 60/2125] avg loss 0.000632089, throughput 6.01591K wps
[Epoch 34 Batch 90/2125] avg loss 0.000381954, throughput 6.01417K wps
[Epoch 34 Batch 120/2125] avg loss 0.000434291, throughput 6.01475K wps
[Epoch 34 Batch 150/2125] avg loss 0.00052834, throughput 6.01264K wps
[Epoch 34 Batch 180/2125] avg loss 0.000483241, throughput 6.01354K wps
[Epoch 34 Batch 210/2125] avg loss 0.000397727, throughput 6.0125K wps
[Epoch 34 Batch 240/2125] avg loss 0.00061295, throughput 6.00861K wps
[Epoch 34 Batch 270/2125] avg loss 0.000578967, throughput 6.00567K wps
[Epoch 34 Batch 300/2125] avg loss 0.000510717, throughput 6.00162K wps
[Epoch 34 Batch 330/2125] avg loss 0.000685159, throughput 6.00927K wps
[Epoch 34 Batch 360/2125] avg loss 0.000471567, throughput 6.00577K wps
[Epoch 34 Batch 390/2125] avg loss 0.000576568, throughput 6.01557K wps
[Epoch 34 Batch 420/2125] avg loss 0.000728058, throughput 6.00726K wps
[Epoch 34 Batch 450/2125] avg loss 0.000469102, throughput 6.00664K wps
[Epoch 34 Batch 480/2125] avg loss 0.000429479, throughput 6.01535K wps
[Epoch 34 Batch 510/2125] avg loss 0.000655993, throughput 6.00807K wps
[Epoch 34 Batch 540/2125] avg loss 0.000504561, throughput 6.01319K wps
[Epoch 34 Batch 570/2125] avg loss 0.000529864, throughput 6.01826K wps
[Epoch 34 Batch 600/2125] avg loss 0.000547423, throughput 6.01395K wps
[Epoch 34 Batch 630/2125] avg loss 0.000528733, throughput 6.02289K wps
[Epoch 34 Batch 660/2125] avg loss 0.000444657, throughput 6.02076K wps
[Epoch 34 Batch 690/2125] avg loss 0.000471187, throughput 6.01589K wps
[Epoch 34 Batch 720/2125] avg loss 0.000447345, throughput 6.01542K wps
[Epoch 34 Batch 750/2125] avg loss 0.000639345, throughput 6.01032K wps
[Epoch 34 Batch 780/2125] avg loss 0.000585334, throughput 6.01457K wps
[Epoch 34 Batch 810/2125] avg loss 0.000635712, throughput 6.02254K wps
[Epoch 34 Batch 840/2125] avg loss 0.000548026, throughput 6.00567K wps
[Epoch 34 Batch 870/2125] avg loss 0.000595277, throughput 6.00599K wps
[Epoch 34 Batch 900/2125] avg loss 0.000353535, throughput 6.00941K wps
[Epoch 34 Batch 930/2125] avg loss 0.000744956, throughput 6.00728K wps
[Epoch 34 Batch 960/2125] avg loss 0.000700419, throughput 6.02233K wps
[Epoch 34 Batch 990/2125] avg loss 0.000581348, throughput 6.03309K wps
[Epoch 34 Batch 1020/2125] avg loss 0.00052928, throughput 6.02464K wps
[Epoch 34 Batch 1050/2125] avg loss 0.000882126, throughput 6.02561K wps
[Epoch 34 Batch 1080/2125] avg loss 0.000573498, throughput 6.02285K wps
[Epoch 34 Batch 1110/2125] avg loss 0.000481131, throughput 6.00621K wps
[Epoch 34 Batch 1140/2125] avg loss 0.000896409, throughput 5.9995K wps
[Epoch 34 Batch 1170/2125] avg loss 0.000649732, throughput 6.00342K wps
[Epoch 34 Batch 1200/2125] avg loss 0.000857829, throughput 6.01233K wps
[Epoch 34 Batch 1230/2125] avg loss 0.00064753, throughput 6.015K wps
[Epoch 34 Batch 1260/2125] avg loss 0.000812424, throughput 6.0113K wps
[Epoch 34 Batch 1290/2125] avg loss 0.000716579, throughput 6.01363K wps
[Epoch 34 Batch 1320/2125] avg loss 0.000757782, throughput 6.01554K wps
[Epoch 34 Batch 1350/2125] avg loss 0.000488576, throughput 6.01853K wps
[Epoch 34 Batch 1380/2125] avg loss 0.000452974, throughput 6.01364K wps
[Epoch 34 Batch 1410/2125] avg loss 0.00068463, throughput 6.0176K wps
[Epoch 34 Batch 1440/2125] avg loss 0.000765328, throughput 6.01981K wps
[Epoch 34 Batch 1470/2125] avg loss 0.000897466, throughput 6.02448K wps
[Epoch 34 Batch 1500/2125] avg loss 0.000682385, throughput 6.02012K wps
[Epoch 34 Batch 1530/2125] avg loss 0.000652992, throughput 6.02177K wps
[Epoch 34 Batch 1560/2125] avg loss 0.000834237, throughput 6.02538K wps
[Epoch 34 Batch 1590/2125] avg loss 0.00071918, throughput 6.02014K wps
[Epoch 34 Batch 1620/2125] avg loss 0.000718685, throughput 6.00774K wps
[Epoch 34 Batch 1650/2125] avg loss 0.0005035, throughput 6.00852K wps
[Epoch 34 Batch 1680/2125] avg loss 0.000592173, throughput 6.01541K wps
[Epoch 34 Batch 1710/2125] avg loss 0.00054061, throughput 6.00607K wps
[Epoch 34 Batch 1740/2125] avg loss 0.000734323, throughput 6.01535K wps
[Epoch 34 Batch 1770/2125] avg loss 0.000868618, throughput 6.01012K wps
[Epoch 34 Batch 1800/2125] avg loss 0.000386935, throughput 6.01135K wps
[Epoch 34 Batch 1830/2125] avg loss 0.000726003, throughput 6.01251K wps
[Epoch 34 Batch 1860/2125] avg loss 0.000672533, throughput 6.01405K wps
[Epoch 34 Batch 1890/2125] avg loss 0.000532447, throughput 6.01748K wps
[Epoch 34 Batch 1920/2125] avg loss 0.000605161, throughput 6.01602K wps
[Epoch 34 Batch 1950/2125] avg loss 0.000514869, throughput 6.01626K wps
[Epoch 34 Batch 1980/2125] avg loss 0.000549538, throughput 6.01507K wps
[Epoch 34 Batch 2010/2125] avg loss 0.00058906, throughput 6.00964K wps
[Epoch 34 Batch 2040/2125] avg loss 0.000574093, throughput 6.00208K wps
[Epoch 34 Batch 2070/2125] avg loss 0.000584121, throughput 6.01092K wps
[Epoch 34 Batch 2100/2125] avg loss 0.000744063, throughput 6.00884K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 34] train avg loss 0.00060661, test acc 0.9245, test avg loss 0.511161, throughput 6.01596K wps
[Epoch 35 Batch 30/2125] avg loss 0.000393299, throughput 6.14751K wps
[Epoch 35 Batch 60/2125] avg loss 0.000420977, throughput 6.0202K wps
[Epoch 35 Batch 90/2125] avg loss 0.000536559, throughput 6.0193K wps
[Epoch 35 Batch 120/2125] avg loss 0.000365617, throughput 6.01969K wps
[Epoch 35 Batch 150/2125] avg loss 0.000607257, throughput 6.01686K wps
[Epoch 35 Batch 180/2125] avg loss 0.00060767, throughput 6.0226K wps
[Epoch 35 Batch 210/2125] avg loss 0.000537716, throughput 6.007K wps
[Epoch 35 Batch 240/2125] avg loss 0.00036277, throughput 6.00298K wps
[Epoch 35 Batch 270/2125] avg loss 0.000684197, throughput 6.00834K wps
[Epoch 35 Batch 300/2125] avg loss 0.000494366, throughput 6.01592K wps
[Epoch 35 Batch 330/2125] avg loss 0.000337785, throughput 6.01686K wps
[Epoch 35 Batch 360/2125] avg loss 0.000368297, throughput 6.02694K wps
[Epoch 35 Batch 390/2125] avg loss 0.000594161, throughput 6.00798K wps
[Epoch 35 Batch 420/2125] avg loss 0.000473225, throughput 6.00431K wps
[Epoch 35 Batch 450/2125] avg loss 0.000633843, throughput 6.01643K wps
[Epoch 35 Batch 480/2125] avg loss 0.000508844, throughput 6.01645K wps
[Epoch 35 Batch 510/2125] avg loss 0.000404699, throughput 5.99407K wps
[Epoch 35 Batch 540/2125] avg loss 0.000440117, throughput 6.01841K wps
[Epoch 35 Batch 570/2125] avg loss 0.00037304, throughput 5.99771K wps
[Epoch 35 Batch 600/2125] avg loss 0.00058828, throughput 6.02012K wps
[Epoch 35 Batch 630/2125] avg loss 0.000583833, throughput 6.01324K wps
[Epoch 35 Batch 660/2125] avg loss 0.000517223, throughput 6.01459K wps
[Epoch 35 Batch 690/2125] avg loss 0.000862242, throughput 6.01042K wps
[Epoch 35 Batch 720/2125] avg loss 0.000540344, throughput 6.01373K wps
[Epoch 35 Batch 750/2125] avg loss 0.000671603, throughput 6.0214K wps
[Epoch 35 Batch 780/2125] avg loss 0.000416303, throughput 6.01241K wps
[Epoch 35 Batch 810/2125] avg loss 0.00060358, throughput 6.02964K wps
[Epoch 35 Batch 840/2125] avg loss 0.000422919, throughput 6.0225K wps
[Epoch 35 Batch 870/2125] avg loss 0.000406335, throughput 6.02016K wps
[Epoch 35 Batch 900/2125] avg loss 0.000451264, throughput 6.01316K wps
[Epoch 35 Batch 930/2125] avg loss 0.000431306, throughput 6.01732K wps
[Epoch 35 Batch 960/2125] avg loss 0.000666011, throughput 6.01404K wps
[Epoch 35 Batch 990/2125] avg loss 0.000606255, throughput 6.02584K wps
[Epoch 35 Batch 1020/2125] avg loss 0.000588525, throughput 6.01351K wps
[Epoch 35 Batch 1050/2125] avg loss 0.000543886, throughput 5.99995K wps
[Epoch 35 Batch 1080/2125] avg loss 0.000570307, throughput 6.00016K wps
[Epoch 35 Batch 1110/2125] avg loss 0.000511496, throughput 6.01136K wps
[Epoch 35 Batch 1140/2125] avg loss 0.000576439, throughput 6.01486K wps
[Epoch 35 Batch 1170/2125] avg loss 0.000550234, throughput 6.00995K wps
[Epoch 35 Batch 1200/2125] avg loss 0.000465925, throughput 6.00676K wps
[Epoch 35 Batch 1230/2125] avg loss 0.000669386, throughput 6.00859K wps
[Epoch 35 Batch 1260/2125] avg loss 0.000655409, throughput 6.0194K wps
[Epoch 35 Batch 1290/2125] avg loss 0.000657788, throughput 6.00436K wps
[Epoch 35 Batch 1320/2125] avg loss 0.000524513, throughput 6.01935K wps
[Epoch 35 Batch 1350/2125] avg loss 0.000630156, throughput 6.02381K wps
[Epoch 35 Batch 1380/2125] avg loss 0.00051516, throughput 6.02068K wps
[Epoch 35 Batch 1410/2125] avg loss 0.000585186, throughput 6.01426K wps
[Epoch 35 Batch 1440/2125] avg loss 0.000508147, throughput 6.00775K wps
[Epoch 35 Batch 1470/2125] avg loss 0.000537277, throughput 6.00087K wps
[Epoch 35 Batch 1500/2125] avg loss 0.00068911, throughput 6.01798K wps
[Epoch 35 Batch 1530/2125] avg loss 0.000510652, throughput 6.02366K wps
[Epoch 35 Batch 1560/2125] avg loss 0.000662983, throughput 6.00754K wps
[Epoch 35 Batch 1590/2125] avg loss 0.000561588, throughput 6.01436K wps
[Epoch 35 Batch 1620/2125] avg loss 0.000804644, throughput 6.01711K wps
[Epoch 35 Batch 1650/2125] avg loss 0.000761352, throughput 6.01842K wps
[Epoch 35 Batch 1680/2125] avg loss 0.000866941, throughput 6.02222K wps
[Epoch 35 Batch 1710/2125] avg loss 0.000556907, throughput 6.01335K wps
[Epoch 35 Batch 1740/2125] avg loss 0.00082518, throughput 6.01018K wps
[Epoch 35 Batch 1770/2125] avg loss 0.000600096, throughput 6.01311K wps
[Epoch 35 Batch 1800/2125] avg loss 0.000559009, throughput 6.00133K wps
[Epoch 35 Batch 1830/2125] avg loss 0.00058899, throughput 6.0184K wps
[Epoch 35 Batch 1860/2125] avg loss 0.000423576, throughput 6.0105K wps
[Epoch 35 Batch 1890/2125] avg loss 0.000705978, throughput 6.00526K wps
[Epoch 35 Batch 1920/2125] avg loss 0.000579624, throughput 6.00324K wps
[Epoch 35 Batch 1950/2125] avg loss 0.000492475, throughput 6.00305K wps
[Epoch 35 Batch 1980/2125] avg loss 0.00103379, throughput 6.00215K wps
[Epoch 35 Batch 2010/2125] avg loss 0.000714982, throughput 6.00469K wps
[Epoch 35 Batch 2040/2125] avg loss 0.000609381, throughput 6.00117K wps
[Epoch 35 Batch 2070/2125] avg loss 0.000571714, throughput 6.01119K wps
[Epoch 35 Batch 2100/2125] avg loss 0.000720676, throughput 6.0099K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 35] train avg loss 0.000571206, test acc 0.9244, test avg loss 0.513567, throughput 6.01475K wps
[Epoch 36 Batch 30/2125] avg loss 0.000532397, throughput 6.15475K wps
[Epoch 36 Batch 60/2125] avg loss 0.000479445, throughput 6.02002K wps
[Epoch 36 Batch 90/2125] avg loss 0.000370319, throughput 6.01108K wps
[Epoch 36 Batch 120/2125] avg loss 0.000467572, throughput 6.01079K wps
[Epoch 36 Batch 150/2125] avg loss 0.000638487, throughput 6.00423K wps
[Epoch 36 Batch 180/2125] avg loss 0.00061505, throughput 6.0169K wps
[Epoch 36 Batch 210/2125] avg loss 0.000374733, throughput 6.01686K wps
[Epoch 36 Batch 240/2125] avg loss 0.000390769, throughput 6.0101K wps
[Epoch 36 Batch 270/2125] avg loss 0.000477661, throughput 6.01794K wps
[Epoch 36 Batch 300/2125] avg loss 0.000412183, throughput 6.01041K wps
[Epoch 36 Batch 330/2125] avg loss 0.000427784, throughput 6.02039K wps
[Epoch 36 Batch 360/2125] avg loss 0.00039547, throughput 6.01091K wps
[Epoch 36 Batch 390/2125] avg loss 0.000574342, throughput 6.01082K wps
[Epoch 36 Batch 420/2125] avg loss 0.000382665, throughput 6.011K wps
[Epoch 36 Batch 450/2125] avg loss 0.000635657, throughput 6.00973K wps
[Epoch 36 Batch 480/2125] avg loss 0.000377353, throughput 6.0088K wps
[Epoch 36 Batch 510/2125] avg loss 0.000480935, throughput 6.01689K wps
[Epoch 36 Batch 540/2125] avg loss 0.000679141, throughput 6.0131K wps
[Epoch 36 Batch 570/2125] avg loss 0.000715302, throughput 6.01315K wps
[Epoch 36 Batch 600/2125] avg loss 0.000406305, throughput 6.01525K wps
[Epoch 36 Batch 630/2125] avg loss 0.000571318, throughput 6.01361K wps
[Epoch 36 Batch 660/2125] avg loss 0.000572003, throughput 6.02573K wps
[Epoch 36 Batch 690/2125] avg loss 0.00038122, throughput 6.02617K wps
[Epoch 36 Batch 720/2125] avg loss 0.0005546, throughput 6.02491K wps
[Epoch 36 Batch 750/2125] avg loss 0.000493875, throughput 6.02637K wps
[Epoch 36 Batch 780/2125] avg loss 0.000601267, throughput 6.01867K wps
[Epoch 36 Batch 810/2125] avg loss 0.000600034, throughput 6.02155K wps
[Epoch 36 Batch 840/2125] avg loss 0.000561251, throughput 6.02015K wps
[Epoch 36 Batch 870/2125] avg loss 0.000609266, throughput 6.02208K wps
[Epoch 36 Batch 900/2125] avg loss 0.000571439, throughput 6.00442K wps
[Epoch 36 Batch 930/2125] avg loss 0.000690201, throughput 5.97917K wps
[Epoch 36 Batch 960/2125] avg loss 0.00050258, throughput 6.00953K wps
[Epoch 36 Batch 990/2125] avg loss 0.000678321, throughput 6.01774K wps
[Epoch 36 Batch 1020/2125] avg loss 0.000882472, throughput 6.00689K wps
[Epoch 36 Batch 1050/2125] avg loss 0.000759386, throughput 6.01399K wps
[Epoch 36 Batch 1080/2125] avg loss 0.000578425, throughput 6.01735K wps
[Epoch 36 Batch 1110/2125] avg loss 0.000515704, throughput 6.01228K wps
[Epoch 36 Batch 1140/2125] avg loss 0.000715035, throughput 6.01897K wps
[Epoch 36 Batch 1170/2125] avg loss 0.000509565, throughput 6.01993K wps
[Epoch 36 Batch 1200/2125] avg loss 0.000575523, throughput 6.01383K wps
[Epoch 36 Batch 1230/2125] avg loss 0.00050788, throughput 6.01332K wps
[Epoch 36 Batch 1260/2125] avg loss 0.000505249, throughput 6.0222K wps
[Epoch 36 Batch 1290/2125] avg loss 0.000584597, throughput 6.02445K wps
[Epoch 36 Batch 1320/2125] avg loss 0.000413439, throughput 6.00326K wps
[Epoch 36 Batch 1350/2125] avg loss 0.000487794, throughput 6.01213K wps
[Epoch 36 Batch 1380/2125] avg loss 0.000695774, throughput 6.00717K wps
[Epoch 36 Batch 1410/2125] avg loss 0.000442477, throughput 6.01037K wps
[Epoch 36 Batch 1440/2125] avg loss 0.000567078, throughput 6.03363K wps
[Epoch 36 Batch 1470/2125] avg loss 0.000751545, throughput 6.01555K wps
[Epoch 36 Batch 1500/2125] avg loss 0.000551378, throughput 6.02188K wps
[Epoch 36 Batch 1530/2125] avg loss 0.000675961, throughput 6.02238K wps
[Epoch 36 Batch 1560/2125] avg loss 0.000632717, throughput 6.01679K wps
[Epoch 36 Batch 1590/2125] avg loss 0.000737359, throughput 6.01166K wps
[Epoch 36 Batch 1620/2125] avg loss 0.000757794, throughput 6.02114K wps
[Epoch 36 Batch 1650/2125] avg loss 0.000693377, throughput 6.01815K wps
[Epoch 36 Batch 1680/2125] avg loss 0.00065942, throughput 6.01144K wps
[Epoch 36 Batch 1710/2125] avg loss 0.000702785, throughput 6.02919K wps
[Epoch 36 Batch 1740/2125] avg loss 0.000580382, throughput 6.02234K wps
[Epoch 36 Batch 1770/2125] avg loss 0.000582596, throughput 6.0213K wps
[Epoch 36 Batch 1800/2125] avg loss 0.000741857, throughput 6.02092K wps
[Epoch 36 Batch 1830/2125] avg loss 0.000417293, throughput 6.01529K wps
[Epoch 36 Batch 1860/2125] avg loss 0.000652528, throughput 6.01917K wps
[Epoch 36 Batch 1890/2125] avg loss 0.000618775, throughput 6.01101K wps
[Epoch 36 Batch 1920/2125] avg loss 0.000536887, throughput 6.02207K wps
[Epoch 36 Batch 1950/2125] avg loss 0.000513555, throughput 6.02524K wps
[Epoch 36 Batch 1980/2125] avg loss 0.000623775, throughput 6.01368K wps
[Epoch 36 Batch 2010/2125] avg loss 0.000646737, throughput 6.02035K wps
[Epoch 36 Batch 2040/2125] avg loss 0.000601659, throughput 6.00808K wps
[Epoch 36 Batch 2070/2125] avg loss 0.000623801, throughput 6.00121K wps
[Epoch 36 Batch 2100/2125] avg loss 0.000624114, throughput 6.00511K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 36] train avg loss 0.000570415, test acc 0.9236, test avg loss 0.521287, throughput 6.01736K wps
[Epoch 37 Batch 30/2125] avg loss 0.000603371, throughput 6.15037K wps
[Epoch 37 Batch 60/2125] avg loss 0.000588885, throughput 6.0036K wps
[Epoch 37 Batch 90/2125] avg loss 0.000454255, throughput 6.01523K wps
[Epoch 37 Batch 120/2125] avg loss 0.000308072, throughput 6.01504K wps
[Epoch 37 Batch 150/2125] avg loss 0.000333251, throughput 6.01548K wps
[Epoch 37 Batch 180/2125] avg loss 0.000432285, throughput 6.01188K wps
[Epoch 37 Batch 210/2125] avg loss 0.000396201, throughput 6.00677K wps
[Epoch 37 Batch 240/2125] avg loss 0.000351201, throughput 6.01189K wps
[Epoch 37 Batch 270/2125] avg loss 0.000436187, throughput 6.00175K wps
[Epoch 37 Batch 300/2125] avg loss 0.000508804, throughput 6.00173K wps
[Epoch 37 Batch 330/2125] avg loss 0.000476252, throughput 6.00674K wps
[Epoch 37 Batch 360/2125] avg loss 0.000393043, throughput 6.01109K wps
[Epoch 37 Batch 390/2125] avg loss 0.000542721, throughput 6.01777K wps
[Epoch 37 Batch 420/2125] avg loss 0.000541639, throughput 6.01108K wps
[Epoch 37 Batch 450/2125] avg loss 0.000399138, throughput 6.00826K wps
[Epoch 37 Batch 480/2125] avg loss 0.000349381, throughput 6.01323K wps
[Epoch 37 Batch 510/2125] avg loss 0.000550487, throughput 6.01074K wps
[Epoch 37 Batch 540/2125] avg loss 0.000599141, throughput 6.014K wps
[Epoch 37 Batch 570/2125] avg loss 0.000450884, throughput 6.00864K wps
[Epoch 37 Batch 600/2125] avg loss 0.000729657, throughput 6.01066K wps
[Epoch 37 Batch 630/2125] avg loss 0.000347742, throughput 6.01638K wps
[Epoch 37 Batch 660/2125] avg loss 0.000556575, throughput 6.01046K wps
[Epoch 37 Batch 690/2125] avg loss 0.00055366, throughput 6.01475K wps
[Epoch 37 Batch 720/2125] avg loss 0.000668882, throughput 6.00868K wps
[Epoch 37 Batch 750/2125] avg loss 0.000624635, throughput 6.01549K wps
[Epoch 37 Batch 780/2125] avg loss 0.00041244, throughput 6.01517K wps
[Epoch 37 Batch 810/2125] avg loss 0.000543704, throughput 6.00653K wps
[Epoch 37 Batch 840/2125] avg loss 0.000607868, throughput 6.01431K wps
[Epoch 37 Batch 870/2125] avg loss 0.000516883, throughput 5.99916K wps
[Epoch 37 Batch 900/2125] avg loss 0.000518334, throughput 6.00608K wps
[Epoch 37 Batch 930/2125] avg loss 0.000534979, throughput 6.00692K wps
[Epoch 37 Batch 960/2125] avg loss 0.000783952, throughput 6.00717K wps
[Epoch 37 Batch 990/2125] avg loss 0.000592942, throughput 6.00779K wps
[Epoch 37 Batch 1020/2125] avg loss 0.000431216, throughput 6.00234K wps
[Epoch 37 Batch 1050/2125] avg loss 0.000568429, throughput 6.00148K wps
[Epoch 37 Batch 1080/2125] avg loss 0.000420278, throughput 6.00432K wps
[Epoch 37 Batch 1110/2125] avg loss 0.000574999, throughput 6.00381K wps
[Epoch 37 Batch 1140/2125] avg loss 0.000617172, throughput 6.00342K wps
[Epoch 37 Batch 1170/2125] avg loss 0.000374898, throughput 6.00404K wps
[Epoch 37 Batch 1200/2125] avg loss 0.000473523, throughput 6.01031K wps
[Epoch 37 Batch 1230/2125] avg loss 0.000466122, throughput 6.00701K wps
[Epoch 37 Batch 1260/2125] avg loss 0.0003866, throughput 6.00782K wps
[Epoch 37 Batch 1290/2125] avg loss 0.000489353, throughput 6.01197K wps
[Epoch 37 Batch 1320/2125] avg loss 0.0005771, throughput 6.01445K wps
[Epoch 37 Batch 1350/2125] avg loss 0.000590424, throughput 6.01309K wps
[Epoch 37 Batch 1380/2125] avg loss 0.000740914, throughput 6.0019K wps
[Epoch 37 Batch 1410/2125] avg loss 0.000461295, throughput 6.01288K wps
[Epoch 37 Batch 1440/2125] avg loss 0.000555085, throughput 6.01371K wps
[Epoch 37 Batch 1470/2125] avg loss 0.000435373, throughput 6.01099K wps
[Epoch 37 Batch 1500/2125] avg loss 0.00058061, throughput 6.00458K wps
[Epoch 37 Batch 1530/2125] avg loss 0.000530163, throughput 6.01134K wps
[Epoch 37 Batch 1560/2125] avg loss 0.000695706, throughput 6.02065K wps
[Epoch 37 Batch 1590/2125] avg loss 0.000759215, throughput 6.02097K wps
[Epoch 37 Batch 1620/2125] avg loss 0.000523717, throughput 6.01701K wps
[Epoch 37 Batch 1650/2125] avg loss 0.000434628, throughput 6.01322K wps
[Epoch 37 Batch 1680/2125] avg loss 0.000664142, throughput 6.01605K wps
[Epoch 37 Batch 1710/2125] avg loss 0.000453585, throughput 6.00582K wps
[Epoch 37 Batch 1740/2125] avg loss 0.00072307, throughput 6.01107K wps
[Epoch 37 Batch 1770/2125] avg loss 0.000549535, throughput 6.01274K wps
[Epoch 37 Batch 1800/2125] avg loss 0.000473674, throughput 6.00656K wps
[Epoch 37 Batch 1830/2125] avg loss 0.000718493, throughput 6.01911K wps
[Epoch 37 Batch 1860/2125] avg loss 0.000766276, throughput 6.01162K wps
[Epoch 37 Batch 1890/2125] avg loss 0.000522833, throughput 6.00582K wps
[Epoch 37 Batch 1920/2125] avg loss 0.000363313, throughput 6.01265K wps
[Epoch 37 Batch 1950/2125] avg loss 0.000482573, throughput 6.01531K wps
[Epoch 37 Batch 1980/2125] avg loss 0.000677745, throughput 6.01906K wps
[Epoch 37 Batch 2010/2125] avg loss 0.000609724, throughput 6.0132K wps
[Epoch 37 Batch 2040/2125] avg loss 0.000618674, throughput 6.01116K wps
[Epoch 37 Batch 2070/2125] avg loss 0.000895776, throughput 6.0185K wps
[Epoch 37 Batch 2100/2125] avg loss 0.000552121, throughput 6.00842K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 37] train avg loss 0.000536597, test acc 0.9239, test avg loss 0.529902, throughput 6.01247K wps
[Epoch 38 Batch 30/2125] avg loss 0.000471407, throughput 6.14571K wps
[Epoch 38 Batch 60/2125] avg loss 0.00052364, throughput 6.02067K wps
[Epoch 38 Batch 90/2125] avg loss 0.000430367, throughput 6.0105K wps
[Epoch 38 Batch 120/2125] avg loss 0.000395363, throughput 6.01389K wps
[Epoch 38 Batch 150/2125] avg loss 0.00064386, throughput 6.00818K wps
[Epoch 38 Batch 180/2125] avg loss 0.000536463, throughput 6.01086K wps
[Epoch 38 Batch 210/2125] avg loss 0.000374822, throughput 6.01454K wps
[Epoch 38 Batch 240/2125] avg loss 0.000455382, throughput 6.01054K wps
[Epoch 38 Batch 270/2125] avg loss 0.000334256, throughput 6.01834K wps
[Epoch 38 Batch 300/2125] avg loss 0.000301762, throughput 6.00959K wps
[Epoch 38 Batch 330/2125] avg loss 0.000592086, throughput 6.01151K wps
[Epoch 38 Batch 360/2125] avg loss 0.000547125, throughput 6.0213K wps
[Epoch 38 Batch 390/2125] avg loss 0.000451575, throughput 6.01578K wps
[Epoch 38 Batch 420/2125] avg loss 0.000325717, throughput 6.0155K wps
[Epoch 38 Batch 450/2125] avg loss 0.00057174, throughput 6.0145K wps
[Epoch 38 Batch 480/2125] avg loss 0.00038295, throughput 6.00384K wps
[Epoch 38 Batch 510/2125] avg loss 0.000458335, throughput 6.01148K wps
[Epoch 38 Batch 540/2125] avg loss 0.000455242, throughput 6.01643K wps
[Epoch 38 Batch 570/2125] avg loss 0.000602857, throughput 6.01471K wps
[Epoch 38 Batch 600/2125] avg loss 0.000438918, throughput 6.00926K wps
[Epoch 38 Batch 630/2125] avg loss 0.00052738, throughput 6.01124K wps
[Epoch 38 Batch 660/2125] avg loss 0.000416921, throughput 6.0173K wps
[Epoch 38 Batch 690/2125] avg loss 0.000387253, throughput 6.01796K wps
[Epoch 38 Batch 720/2125] avg loss 0.000364714, throughput 6.01565K wps
[Epoch 38 Batch 750/2125] avg loss 0.000461608, throughput 6.00592K wps
[Epoch 38 Batch 780/2125] avg loss 0.000417736, throughput 6.01926K wps
[Epoch 38 Batch 810/2125] avg loss 0.000406864, throughput 6.01906K wps
[Epoch 38 Batch 840/2125] avg loss 0.000635632, throughput 6.01876K wps
[Epoch 38 Batch 870/2125] avg loss 0.000593846, throughput 6.01297K wps
[Epoch 38 Batch 900/2125] avg loss 0.000421801, throughput 6.01743K wps
[Epoch 38 Batch 930/2125] avg loss 0.000462469, throughput 6.01055K wps
[Epoch 38 Batch 960/2125] avg loss 0.000603024, throughput 6.00963K wps
[Epoch 38 Batch 990/2125] avg loss 0.000512374, throughput 6.01099K wps
[Epoch 38 Batch 1020/2125] avg loss 0.000611201, throughput 6.02366K wps
[Epoch 38 Batch 1050/2125] avg loss 0.000704818, throughput 6.00168K wps
[Epoch 38 Batch 1080/2125] avg loss 0.000351437, throughput 6.02079K wps
[Epoch 38 Batch 1110/2125] avg loss 0.000715524, throughput 6.01184K wps
[Epoch 38 Batch 1140/2125] avg loss 0.000543164, throughput 6.0122K wps
[Epoch 38 Batch 1170/2125] avg loss 0.000717271, throughput 6.00793K wps
[Epoch 38 Batch 1200/2125] avg loss 0.000716009, throughput 6.01766K wps
[Epoch 38 Batch 1230/2125] avg loss 0.000703358, throughput 6.01785K wps
[Epoch 38 Batch 1260/2125] avg loss 0.000450422, throughput 6.02K wps
[Epoch 38 Batch 1290/2125] avg loss 0.000392084, throughput 6.01902K wps
[Epoch 38 Batch 1320/2125] avg loss 0.000323456, throughput 6.00622K wps
[Epoch 38 Batch 1350/2125] avg loss 0.000469721, throughput 6.02544K wps
[Epoch 38 Batch 1380/2125] avg loss 0.000702203, throughput 6.02241K wps
[Epoch 38 Batch 1410/2125] avg loss 0.000398366, throughput 6.02116K wps
[Epoch 38 Batch 1440/2125] avg loss 0.000697384, throughput 6.02329K wps
[Epoch 38 Batch 1470/2125] avg loss 0.00073853, throughput 6.02253K wps
[Epoch 38 Batch 1500/2125] avg loss 0.000561293, throughput 6.01761K wps
[Epoch 38 Batch 1530/2125] avg loss 0.000540367, throughput 6.01328K wps
[Epoch 38 Batch 1560/2125] avg loss 0.00067038, throughput 6.02638K wps
[Epoch 38 Batch 1590/2125] avg loss 0.000617359, throughput 6.02155K wps
[Epoch 38 Batch 1620/2125] avg loss 0.000702422, throughput 6.01177K wps
[Epoch 38 Batch 1650/2125] avg loss 0.000654858, throughput 6.00956K wps
[Epoch 38 Batch 1680/2125] avg loss 0.000752322, throughput 6.02579K wps
[Epoch 38 Batch 1710/2125] avg loss 0.000529251, throughput 6.02365K wps
[Epoch 38 Batch 1740/2125] avg loss 0.000479282, throughput 6.02546K wps
[Epoch 38 Batch 1770/2125] avg loss 0.00031302, throughput 6.01463K wps
[Epoch 38 Batch 1800/2125] avg loss 0.000494277, throughput 6.0187K wps
[Epoch 38 Batch 1830/2125] avg loss 0.000510413, throughput 6.00982K wps
[Epoch 38 Batch 1860/2125] avg loss 0.000575342, throughput 6.00467K wps
[Epoch 38 Batch 1890/2125] avg loss 0.000518029, throughput 6.01009K wps
[Epoch 38 Batch 1920/2125] avg loss 0.000383929, throughput 6.01008K wps
[Epoch 38 Batch 1950/2125] avg loss 0.000401971, throughput 5.99346K wps
[Epoch 38 Batch 1980/2125] avg loss 0.000531556, throughput 6.0072K wps
[Epoch 38 Batch 2010/2125] avg loss 0.000614242, throughput 6.01583K wps
[Epoch 38 Batch 2040/2125] avg loss 0.000454553, throughput 6.01221K wps
[Epoch 38 Batch 2070/2125] avg loss 0.000635757, throughput 6.01191K wps
[Epoch 38 Batch 2100/2125] avg loss 0.000569957, throughput 6.0109K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 38] train avg loss 0.000517904, test acc 0.9233, test avg loss 0.534516, throughput 6.01637K wps
[Epoch 39 Batch 30/2125] avg loss 0.000489643, throughput 6.15234K wps
[Epoch 39 Batch 60/2125] avg loss 0.000359365, throughput 6.01426K wps
[Epoch 39 Batch 90/2125] avg loss 0.00043376, throughput 6.01115K wps
[Epoch 39 Batch 120/2125] avg loss 0.000463177, throughput 6.00492K wps
[Epoch 39 Batch 150/2125] avg loss 0.000367602, throughput 6.00972K wps
[Epoch 39 Batch 180/2125] avg loss 0.000448156, throughput 6.01312K wps
[Epoch 39 Batch 210/2125] avg loss 0.000419215, throughput 6.01236K wps
[Epoch 39 Batch 240/2125] avg loss 0.000383456, throughput 6.01869K wps
[Epoch 39 Batch 270/2125] avg loss 0.000621901, throughput 6.01565K wps
[Epoch 39 Batch 300/2125] avg loss 0.000324581, throughput 6.0098K wps
[Epoch 39 Batch 330/2125] avg loss 0.000449386, throughput 6.02159K wps
[Epoch 39 Batch 360/2125] avg loss 0.000452623, throughput 6.00514K wps
[Epoch 39 Batch 390/2125] avg loss 0.000435343, throughput 6.00827K wps
[Epoch 39 Batch 420/2125] avg loss 0.000581014, throughput 6.01359K wps
[Epoch 39 Batch 450/2125] avg loss 0.000349295, throughput 6.01453K wps
[Epoch 39 Batch 480/2125] avg loss 0.000643916, throughput 6.01889K wps
[Epoch 39 Batch 510/2125] avg loss 0.00050218, throughput 6.02007K wps
[Epoch 39 Batch 540/2125] avg loss 0.000488158, throughput 6.01961K wps
[Epoch 39 Batch 570/2125] avg loss 0.00069974, throughput 6.00991K wps
[Epoch 39 Batch 600/2125] avg loss 0.000418279, throughput 6.01349K wps
[Epoch 39 Batch 630/2125] avg loss 0.000466817, throughput 6.01772K wps
[Epoch 39 Batch 660/2125] avg loss 0.000571344, throughput 6.01109K wps
[Epoch 39 Batch 690/2125] avg loss 0.000394456, throughput 6.01301K wps
[Epoch 39 Batch 720/2125] avg loss 0.000517208, throughput 6.00625K wps
[Epoch 39 Batch 750/2125] avg loss 0.000399686, throughput 6.0218K wps
[Epoch 39 Batch 780/2125] avg loss 0.000512348, throughput 6.01758K wps
[Epoch 39 Batch 810/2125] avg loss 0.000607629, throughput 6.01742K wps
[Epoch 39 Batch 840/2125] avg loss 0.000479988, throughput 6.02508K wps
[Epoch 39 Batch 870/2125] avg loss 0.000616144, throughput 6.01983K wps
[Epoch 39 Batch 900/2125] avg loss 0.000451256, throughput 6.01449K wps
[Epoch 39 Batch 930/2125] avg loss 0.00040134, throughput 6.01273K wps
[Epoch 39 Batch 960/2125] avg loss 0.000513545, throughput 6.0142K wps
[Epoch 39 Batch 990/2125] avg loss 0.000520326, throughput 6.0134K wps
[Epoch 39 Batch 1020/2125] avg loss 0.000400551, throughput 6.00115K wps
[Epoch 39 Batch 1050/2125] avg loss 0.000476942, throughput 6.00611K wps
[Epoch 39 Batch 1080/2125] avg loss 0.000685628, throughput 6.01598K wps
[Epoch 39 Batch 1110/2125] avg loss 0.000482413, throughput 6.01614K wps
[Epoch 39 Batch 1140/2125] avg loss 0.00043441, throughput 6.01396K wps
[Epoch 39 Batch 1170/2125] avg loss 0.000508974, throughput 6.0151K wps
[Epoch 39 Batch 1200/2125] avg loss 0.000345961, throughput 6.01K wps
[Epoch 39 Batch 1230/2125] avg loss 0.000634677, throughput 6.02297K wps
[Epoch 39 Batch 1260/2125] avg loss 0.000566852, throughput 6.0217K wps
[Epoch 39 Batch 1290/2125] avg loss 0.000887121, throughput 6.0167K wps
[Epoch 39 Batch 1320/2125] avg loss 0.000707437, throughput 6.01884K wps
[Epoch 39 Batch 1350/2125] avg loss 0.000650818, throughput 6.01066K wps
[Epoch 39 Batch 1380/2125] avg loss 0.000546391, throughput 6.01671K wps
[Epoch 39 Batch 1410/2125] avg loss 0.000595805, throughput 6.0178K wps
[Epoch 39 Batch 1440/2125] avg loss 0.000653003, throughput 6.01495K wps
[Epoch 39 Batch 1470/2125] avg loss 0.000545232, throughput 6.01307K wps
[Epoch 39 Batch 1500/2125] avg loss 0.000532379, throughput 6.01562K wps
[Epoch 39 Batch 1530/2125] avg loss 0.000366373, throughput 6.01992K wps
[Epoch 39 Batch 1560/2125] avg loss 0.000429599, throughput 6.02839K wps
[Epoch 39 Batch 1590/2125] avg loss 0.000785293, throughput 6.02423K wps
[Epoch 39 Batch 1620/2125] avg loss 0.000680564, throughput 6.01903K wps
[Epoch 39 Batch 1650/2125] avg loss 0.000515132, throughput 6.01183K wps
[Epoch 39 Batch 1680/2125] avg loss 0.000497922, throughput 6.00931K wps
[Epoch 39 Batch 1710/2125] avg loss 0.000630309, throughput 6.00447K wps
[Epoch 39 Batch 1740/2125] avg loss 0.000607757, throughput 6.00989K wps
[Epoch 39 Batch 1770/2125] avg loss 0.000390059, throughput 6.00983K wps
[Epoch 39 Batch 1800/2125] avg loss 0.000522211, throughput 6.02094K wps
[Epoch 39 Batch 1830/2125] avg loss 0.00062392, throughput 6.01781K wps
[Epoch 39 Batch 1860/2125] avg loss 0.000514757, throughput 6.02271K wps
[Epoch 39 Batch 1890/2125] avg loss 0.000784329, throughput 6.01516K wps
[Epoch 39 Batch 1920/2125] avg loss 0.000649471, throughput 6.01517K wps
[Epoch 39 Batch 1950/2125] avg loss 0.000706892, throughput 6.01799K wps
[Epoch 39 Batch 1980/2125] avg loss 0.000604257, throughput 6.0121K wps
[Epoch 39 Batch 2010/2125] avg loss 0.000517775, throughput 6.01325K wps
[Epoch 39 Batch 2040/2125] avg loss 0.000629394, throughput 6.01748K wps
[Epoch 39 Batch 2070/2125] avg loss 0.000721338, throughput 6.01626K wps
[Epoch 39 Batch 2100/2125] avg loss 0.000549781, throughput 6.01542K wps
Begin Testing...
[Batch 30/237] elapsed 0.29 s
[Batch 60/237] elapsed 0.27 s
[Batch 90/237] elapsed 0.27 s
[Batch 120/237] elapsed 0.27 s
[Batch 150/237] elapsed 0.27 s
[Batch 180/237] elapsed 0.27 s
[Batch 210/237] elapsed 0.27 s
[Epoch 39] train avg loss 0.000531707, test acc 0.9247, test avg loss 0.540785, throughput 6.01683K wps
Test loss 0.286665, test acc 0.9387
Total time cost 2918.61s