mlstm_word2vec_embedding.log

(root)[junfeng@heartmaster reasoning_attention]$ python snli_match_lstm.py 
Loading data ...
Loading train ...
550152
550149
549364
Loading dev ...
10000
10000
9842
Loading test ...
10000
10000
9824
num_epochs: 20
k: 300
batch_size: 30
display_frequency: 100
save_frequency: 1000
load previous: False
Building network ...
unchanged_W.shape: (34283, 300)
oov_in_train_W.shape: (9166, 300)
apply dropout mask id 140256215453992 to embedding matrix ...
dropout rate is 0.3
input var is hypo_var
apply dropout mask id 140256215453992 to embedding matrix ...
dropout rate is 0.3
input var is premise_var
Computing updates ...
Compiling functions ...
Training ...
train_df.shape: (549364, 4)
dev_df.shape: (9842, 4)
test_df.shape: (9824, 4)
Starting training...
Seen 3000 samples, time used: 113.717s
  current training loss:		1.055556
  current training accuracy:		0.457000
Seen 6000 samples, time used: 116.094s
  current training loss:		1.039166
  current training accuracy:		0.480833
Seen 9000 samples, time used: 113.380s
  current training loss:		1.020142
  current training accuracy:		0.503778
Seen 12000 samples, time used: 113.598s
  current training loss:		1.008102
  current training accuracy:		0.515500
Seen 15000 samples, time used: 112.010s
  current training loss:		1.000636
  current training accuracy:		0.522933
Seen 18000 samples, time used: 112.207s
  current training loss:		0.990290
  current training accuracy:		0.533111
Seen 21000 samples, time used: 114.853s
  current training loss:		0.985887
  current training accuracy:		0.538429
Seen 24000 samples, time used: 114.861s
  current training loss:		0.981200
  current training accuracy:		0.543292
Seen 27000 samples, time used: 117.284s
  current training loss:		0.975245
  current training accuracy:		0.548519
Seen 30000 samples, time used: 116.270s
  current training loss:		0.968526
  current training accuracy:		0.554067
saving to ..., time used: 1144.274s
Seen 33000 samples, time used: 115.634s
  current training loss:		0.965249
  current training accuracy:		0.557879
Seen 36000 samples, time used: 119.970s
  current training loss:		0.961199
  current training accuracy:		0.560056
Seen 39000 samples, time used: 121.854s
  current training loss:		0.956373
  current training accuracy:		0.563282
Seen 42000 samples, time used: 121.315s
  current training loss:		0.952917
  current training accuracy:		0.565857
Seen 45000 samples, time used: 118.552s
  current training loss:		0.948475
  current training accuracy:		0.568644
Seen 48000 samples, time used: 118.751s
  current training loss:		0.944854
  current training accuracy:		0.571667
Seen 51000 samples, time used: 118.196s
  current training loss:		0.941263
  current training accuracy:		0.574314
Seen 54000 samples, time used: 121.252s
  current training loss:		0.939563
  current training accuracy:		0.575037
Seen 57000 samples, time used: 118.232s
  current training loss:		0.937118
  current training accuracy:		0.576614
Seen 60000 samples, time used: 118.307s
  current training loss:		0.934130
  current training accuracy:		0.578683
saving to ..., time used: 1191.687s
Seen 63000 samples, time used: 120.935s
  current training loss:		0.931266
  current training accuracy:		0.580302
Seen 66000 samples, time used: 120.415s
  current training loss:		0.928540
  current training accuracy:		0.581909
Seen 69000 samples, time used: 119.993s
  current training loss:		0.926223
  current training accuracy:		0.583493
Seen 72000 samples, time used: 119.140s
  current training loss:		0.923555
  current training accuracy:		0.585194
Seen 75000 samples, time used: 119.911s
  current training loss:		0.921569
  current training accuracy:		0.586320
Seen 78000 samples, time used: 121.211s
  current training loss:		0.919275
  current training accuracy:		0.588218
Seen 81000 samples, time used: 123.604s
  current training loss:		0.916753
  current training accuracy:		0.589827
Seen 84000 samples, time used: 118.953s
  current training loss:		0.915034
  current training accuracy:		0.591060
Seen 87000 samples, time used: 121.948s
  current training loss:		0.913255
  current training accuracy:		0.591874
Seen 90000 samples, time used: 120.588s
  current training loss:		0.911505
  current training accuracy:		0.592922
saving to ..., time used: 1206.219s
Seen 93000 samples, time used: 120.409s
  current training loss:		0.909816
  current training accuracy:		0.594108
Seen 96000 samples, time used: 118.475s
  current training loss:		0.907884
  current training accuracy:		0.595469
Seen 99000 samples, time used: 122.721s
  current training loss:		0.906557
  current training accuracy:		0.596333
Seen 102000 samples, time used: 121.617s
  current training loss:		0.905518
  current training accuracy:		0.597529
Seen 105000 samples, time used: 116.831s
  current training loss:		0.903430
  current training accuracy:		0.598905
Seen 108000 samples, time used: 119.690s
  current training loss:		0.902281
  current training accuracy:		0.599750
Seen 111000 samples, time used: 119.180s
  current training loss:		0.900586
  current training accuracy:		0.600811
Seen 114000 samples, time used: 122.361s
  current training loss:		0.899388
  current training accuracy:		0.601693
Seen 117000 samples, time used: 119.394s
  current training loss:		0.897500
  current training accuracy:		0.602991
Seen 120000 samples, time used: 122.555s
  current training loss:		0.895813
  current training accuracy:		0.604008
saving to ..., time used: 1202.652s
Seen 123000 samples, time used: 118.910s
  current training loss:		0.894277
  current training accuracy:		0.604789
Seen 126000 samples, time used: 123.527s
  current training loss:		0.893011
  current training accuracy:		0.605310
Seen 129000 samples, time used: 122.258s
  current training loss:		0.891226
  current training accuracy:		0.606519
Seen 132000 samples, time used: 123.195s
  current training loss:		0.889625
  current training accuracy:		0.607720
Seen 135000 samples, time used: 125.319s
  current training loss:		0.888271
  current training accuracy:		0.608607
Seen 138000 samples, time used: 121.250s
  current training loss:		0.886478
  current training accuracy:		0.609891
Seen 141000 samples, time used: 122.738s
  current training loss:		0.884963
  current training accuracy:		0.610929
Seen 144000 samples, time used: 120.876s
  current training loss:		0.883712
  current training accuracy:		0.611743
Seen 147000 samples, time used: 119.668s
  current training loss:		0.882778
  current training accuracy:		0.612279
Seen 150000 samples, time used: 120.729s
  current training loss:		0.881116
  current training accuracy:		0.613420
saving to ..., time used: 1218.269s
Seen 153000 samples, time used: 121.819s
  current training loss:		0.879520
  current training accuracy:		0.614477
Seen 156000 samples, time used: 125.764s
  current training loss:		0.877817
  current training accuracy:		0.615628
Seen 159000 samples, time used: 124.872s
  current training loss:		0.876970
  current training accuracy:		0.616170
Seen 162000 samples, time used: 124.079s
  current training loss:		0.875892
  current training accuracy:		0.616840
Seen 165000 samples, time used: 122.320s
  current training loss:		0.874768
  current training accuracy:		0.617679
Seen 168000 samples, time used: 120.716s
  current training loss:		0.873780
  current training accuracy:		0.618452
Seen 171000 samples, time used: 121.592s
  current training loss:		0.872167
  current training accuracy:		0.619322
Seen 174000 samples, time used: 122.147s
  current training loss:		0.870908
  current training accuracy:		0.620149
Seen 177000 samples, time used: 125.237s
  current training loss:		0.869541
  current training accuracy:		0.621028
Seen 180000 samples, time used: 121.913s
  current training loss:		0.868122
  current training accuracy:		0.621939
saving to ..., time used: 1230.259s
Seen 183000 samples, time used: 123.007s
  current training loss:		0.866726
  current training accuracy:		0.622689
Seen 186000 samples, time used: 121.683s
  current training loss:		0.865618
  current training accuracy:		0.623344
Seen 189000 samples, time used: 121.089s
  current training loss:		0.864227
  current training accuracy:		0.624243
Seen 192000 samples, time used: 123.213s
  current training loss:		0.863083
  current training accuracy:		0.624938
Seen 195000 samples, time used: 122.517s
  current training loss:		0.862133
  current training accuracy:		0.625569
Seen 198000 samples, time used: 123.873s
  current training loss:		0.860529
  current training accuracy:		0.626424
Seen 201000 samples, time used: 121.525s
  current training loss:		0.859022
  current training accuracy:		0.627363
Seen 204000 samples, time used: 120.627s
  current training loss:		0.857605
  current training accuracy:		0.628299
Seen 207000 samples, time used: 121.792s
  current training loss:		0.856297
  current training accuracy:		0.629063
Seen 210000 samples, time used: 122.873s
  current training loss:		0.854969
  current training accuracy:		0.629895
saving to ..., time used: 1221.999s
Seen 213000 samples, time used: 122.960s
  current training loss:		0.853590
  current training accuracy:		0.630493
Seen 216000 samples, time used: 122.127s
  current training loss:		0.852267
  current training accuracy:		0.631301
Seen 219000 samples, time used: 123.525s
  current training loss:		0.850964
  current training accuracy:		0.632000
Seen 222000 samples, time used: 123.463s
  current training loss:		0.849976
  current training accuracy:		0.632707
Seen 225000 samples, time used: 121.915s
  current training loss:		0.848996
  current training accuracy:		0.633178
Seen 228000 samples, time used: 122.220s
  current training loss:		0.847807
  current training accuracy:		0.633908
Seen 231000 samples, time used: 124.871s
  current training loss:		0.846406
  current training accuracy:		0.634810
Seen 234000 samples, time used: 121.232s
  current training loss:		0.845046
  current training accuracy:		0.635727
Seen 237000 samples, time used: 120.771s
  current training loss:		0.843380
  current training accuracy:		0.636751
Seen 240000 samples, time used: 121.574s
  current training loss:		0.841685
  current training accuracy:		0.637658
saving to ..., time used: 1224.459s
Seen 243000 samples, time used: 123.243s
  current training loss:		0.840422
  current training accuracy:		0.638494
Seen 246000 samples, time used: 120.193s
  current training loss:		0.839191
  current training accuracy:		0.639305
Seen 249000 samples, time used: 121.114s
  current training loss:		0.838233
  current training accuracy:		0.639924
Seen 252000 samples, time used: 121.881s
  current training loss:		0.836944
  current training accuracy:		0.640790
Seen 255000 samples, time used: 123.720s
  current training loss:		0.835541
  current training accuracy:		0.641604
Seen 258000 samples, time used: 124.214s
  current training loss:		0.834274
  current training accuracy:		0.642384
Seen 261000 samples, time used: 121.748s
  current training loss:		0.833093
  current training accuracy:		0.643188
Seen 264000 samples, time used: 124.034s
  current training loss:		0.831732
  current training accuracy:		0.643924
Seen 267000 samples, time used: 124.763s
  current training loss:		0.830350
  current training accuracy:		0.644708
Seen 270000 samples, time used: 121.751s
  current training loss:		0.829407
  current training accuracy:		0.645374
saving to ..., time used: 1226.460s
Seen 273000 samples, time used: 122.118s
  current training loss:		0.828195
  current training accuracy:		0.646114
Seen 276000 samples, time used: 119.052s
  current training loss:		0.826938
  current training accuracy:		0.646855
Seen 279000 samples, time used: 122.783s
  current training loss:		0.825529
  current training accuracy:		0.647581
Seen 282000 samples, time used: 123.553s
  current training loss:		0.824360
  current training accuracy:		0.648316
Seen 285000 samples, time used: 121.666s
  current training loss:		0.823088
  current training accuracy:		0.649151
Seen 288000 samples, time used: 125.444s
  current training loss:		0.821680
  current training accuracy:		0.650042
Seen 291000 samples, time used: 123.149s
  current training loss:		0.820333
  current training accuracy:		0.650828
Seen 294000 samples, time used: 123.951s
  current training loss:		0.819042
  current training accuracy:		0.651650
Seen 297000 samples, time used: 123.375s
  current training loss:		0.817833
  current training accuracy:		0.652508
Seen 300000 samples, time used: 123.242s
  current training loss:		0.816582
  current training accuracy:		0.653283
saving to ..., time used: 1228.130s
Seen 303000 samples, time used: 125.682s
  current training loss:		0.815524
  current training accuracy:		0.653947
Seen 306000 samples, time used: 124.738s
  current training loss:		0.814327
  current training accuracy:		0.654680
Seen 309000 samples, time used: 125.369s
  current training loss:		0.813227
  current training accuracy:		0.655359
Seen 312000 samples, time used: 122.588s
  current training loss:		0.811997
  current training accuracy:		0.656109
Seen 315000 samples, time used: 121.939s
  current training loss:		0.810775
  current training accuracy:		0.656860
Seen 318000 samples, time used: 125.391s
  current training loss:		0.809612
  current training accuracy:		0.657538
Seen 321000 samples, time used: 122.555s
  current training loss:		0.808341
  current training accuracy:		0.658287
Seen 324000 samples, time used: 118.184s
  current training loss:		0.807036
  current training accuracy:		0.659056
Seen 327000 samples, time used: 124.918s
  current training loss:		0.805978
  current training accuracy:		0.659691
Seen 330000 samples, time used: 124.013s
  current training loss:		0.805030
  current training accuracy:		0.660261
saving to ..., time used: 1235.156s
Seen 333000 samples, time used: 123.029s
  current training loss:		0.803763
  current training accuracy:		0.660988
Seen 336000 samples, time used: 124.069s
  current training loss:		0.802506
  current training accuracy:		0.661655
Seen 339000 samples, time used: 122.483s
  current training loss:		0.801164
  current training accuracy:		0.662378
Seen 342000 samples, time used: 123.777s
  current training loss:		0.799946
  current training accuracy:		0.663073
Seen 345000 samples, time used: 122.113s
  current training loss:		0.798706
  current training accuracy:		0.663806
Seen 348000 samples, time used: 126.011s
  current training loss:		0.797484
  current training accuracy:		0.664468
Seen 351000 samples, time used: 122.982s
  current training loss:		0.796238
  current training accuracy:		0.665217
Seen 354000 samples, time used: 122.017s
  current training loss:		0.794959
  current training accuracy:		0.665949
Seen 357000 samples, time used: 122.439s
  current training loss:		0.793822
  current training accuracy:		0.666625
Seen 360000 samples, time used: 123.041s
  current training loss:		0.792745
  current training accuracy:		0.667250
saving to ..., time used: 1231.761s
Seen 363000 samples, time used: 123.784s
  current training loss:		0.791565
  current training accuracy:		0.667871
Seen 366000 samples, time used: 126.724s
  current training loss:		0.790328
  current training accuracy:		0.668557
Seen 369000 samples, time used: 123.305s
  current training loss:		0.789312
  current training accuracy:		0.669184
Seen 372000 samples, time used: 124.038s
  current training loss:		0.788230
  current training accuracy:		0.669844
Seen 375000 samples, time used: 125.794s
  current training loss:		0.787027
  current training accuracy:		0.670541
Seen 378000 samples, time used: 123.254s
  current training loss:		0.785968
  current training accuracy:		0.671093
Seen 381000 samples, time used: 126.057s
  current training loss:		0.784709
  current training accuracy:		0.671864
Seen 384000 samples, time used: 121.305s
  current training loss:		0.783628
  current training accuracy:		0.672563
Seen 387000 samples, time used: 125.177s
  current training loss:		0.782466
  current training accuracy:		0.673194
Seen 390000 samples, time used: 123.002s
  current training loss:		0.781434
  current training accuracy:		0.673728
saving to ..., time used: 1242.239s
Seen 393000 samples, time used: 123.856s
  current training loss:		0.780396
  current training accuracy:		0.674349
Seen 396000 samples, time used: 124.928s
  current training loss:		0.779411
  current training accuracy:		0.674917
Seen 399000 samples, time used: 125.600s
  current training loss:		0.778450
  current training accuracy:		0.675559
Seen 402000 samples, time used: 124.035s
  current training loss:		0.777556
  current training accuracy:		0.676112
Seen 405000 samples, time used: 124.812s
  current training loss:		0.776470
  current training accuracy:		0.676714
Seen 408000 samples, time used: 123.139s
  current training loss:		0.775512
  current training accuracy:		0.677257
Seen 411000 samples, time used: 123.532s
  current training loss:		0.774459
  current training accuracy:		0.677888
Seen 414000 samples, time used: 126.011s
  current training loss:		0.773531
  current training accuracy:		0.678367
Seen 417000 samples, time used: 121.756s
  current training loss:		0.772426
  current training accuracy:		0.678921
Seen 420000 samples, time used: 125.095s
  current training loss:		0.771377
  current training accuracy:		0.679455
saving to ..., time used: 1242.562s
Seen 423000 samples, time used: 124.254s
  current training loss:		0.770408
  current training accuracy:		0.680000
Seen 426000 samples, time used: 123.470s
  current training loss:		0.769403
  current training accuracy:		0.680542
Seen 429000 samples, time used: 122.641s
  current training loss:		0.768523
  current training accuracy:		0.681096
Seen 432000 samples, time used: 122.907s
  current training loss:		0.767416
  current training accuracy:		0.681660
Seen 435000 samples, time used: 121.848s
  current training loss:		0.766504
  current training accuracy:		0.682126
Seen 438000 samples, time used: 124.139s
  current training loss:		0.765672
  current training accuracy:		0.682543
Seen 441000 samples, time used: 125.023s
  current training loss:		0.764798
  current training accuracy:		0.683091
Seen 444000 samples, time used: 125.742s
  current training loss:		0.763911
  current training accuracy:		0.683653
Seen 447000 samples, time used: 124.462s
  current training loss:		0.763041
  current training accuracy:		0.684105
Seen 450000 samples, time used: 123.657s
  current training loss:		0.761980
  current training accuracy:		0.684716
saving to ..., time used: 1237.939s
Seen 453000 samples, time used: 122.279s
  current training loss:		0.760903
  current training accuracy:		0.685283
Seen 456000 samples, time used: 122.928s
  current training loss:		0.759881
  current training accuracy:		0.685849
Seen 459000 samples, time used: 120.162s
  current training loss:		0.759038
  current training accuracy:		0.686296
Seen 462000 samples, time used: 125.776s
  current training loss:		0.758042
  current training accuracy:		0.686866
Seen 465000 samples, time used: 123.477s
  current training loss:		0.757185
  current training accuracy:		0.687357
Seen 468000 samples, time used: 124.706s
  current training loss:		0.756206
  current training accuracy:		0.687897
Seen 471000 samples, time used: 122.313s
  current training loss:		0.755157
  current training accuracy:		0.688427
Seen 474000 samples, time used: 122.192s
  current training loss:		0.754370
  current training accuracy:		0.688867
Seen 477000 samples, time used: 125.028s
  current training loss:		0.753586
  current training accuracy:		0.689289
Seen 480000 samples, time used: 124.759s
  current training loss:		0.752657
  current training accuracy:		0.689821
saving to ..., time used: 1233.419s
Seen 483000 samples, time used: 121.838s
  current training loss:		0.751635
  current training accuracy:		0.690321
Seen 486000 samples, time used: 123.389s
  current training loss:		0.750671
  current training accuracy:		0.690879
Seen 489000 samples, time used: 125.393s
  current training loss:		0.749880
  current training accuracy:		0.691317
Seen 492000 samples, time used: 125.058s
  current training loss:		0.749010
  current training accuracy:		0.691785
Seen 495000 samples, time used: 121.820s
  current training loss:		0.748090
  current training accuracy:		0.692313
Seen 498000 samples, time used: 123.121s
  current training loss:		0.747257
  current training accuracy:		0.692759
Seen 501000 samples, time used: 124.257s
  current training loss:		0.746500
  current training accuracy:		0.693140
Seen 504000 samples, time used: 124.249s
  current training loss:		0.745535
  current training accuracy:		0.693692
Seen 507000 samples, time used: 123.955s
  current training loss:		0.744505
  current training accuracy:		0.694233
Seen 510000 samples, time used: 127.118s
  current training loss:		0.743662
  current training accuracy:		0.694622
saving to ..., time used: 1239.996s
Seen 513000 samples, time used: 124.177s
  current training loss:		0.742706
  current training accuracy:		0.695170
Seen 516000 samples, time used: 125.315s
  current training loss:		0.741841
  current training accuracy:		0.695636
Seen 519000 samples, time used: 123.407s
  current training loss:		0.740943
  current training accuracy:		0.696135
Seen 522000 samples, time used: 126.236s
  current training loss:		0.740193
  current training accuracy:		0.696544

Seen 525000 samples, time used: 125.235s
  current training loss:		0.739390
  current training accuracy:		0.696977
Seen 528000 samples, time used: 124.031s
  current training loss:		0.738405
  current training accuracy:		0.697528
Seen 531000 samples, time used: 123.827s
  current training loss:		0.737563
  current training accuracy:		0.697994
Seen 534000 samples, time used: 124.134s
  current training loss:		0.736699
  current training accuracy:		0.698451
Seen 537000 samples, time used: 126.495s
  current training loss:		0.735948
  current training accuracy:		0.698888
Seen 540000 samples, time used: 126.930s
  current training loss:		0.735264
  current training accuracy:		0.699187
saving to ..., time used: 1249.585s
Seen 543000 samples, time used: 119.587s
  current training loss:		0.734478
  current training accuracy:		0.699621
Seen 546000 samples, time used: 128.906s
  current training loss:		0.733757
  current training accuracy:		0.699978
Seen 549000 samples, time used: 123.525s
  current training loss:		0.732967
  current training accuracy:		0.700421
Epoch 1 of 20 took 22428.237s
  training loss:		0.732875
  training accuracy:		70.05 %
  validation loss:		0.533690
  validation accuracy:		79.24 %
Seen 3000 samples, time used: 125.579s
  current training loss:		0.557400
  current training accuracy:		0.792000
Seen 6000 samples, time used: 122.986s
  current training loss:		0.567410
  current training accuracy:		0.790000
Seen 9000 samples, time used: 124.620s
  current training loss:		0.571989
  current training accuracy:		0.787556
Seen 12000 samples, time used: 124.700s
  current training loss:		0.570846
  current training accuracy:		0.789833
Seen 15000 samples, time used: 127.579s
  current training loss:		0.567116
  current training accuracy:		0.791333
Seen 18000 samples, time used: 123.973s
  current training loss:		0.566237
  current training accuracy:		0.792167
Seen 21000 samples, time used: 123.974s
  current training loss:		0.567234
  current training accuracy:		0.792095
Seen 24000 samples, time used: 123.111s
  current training loss:		0.563961
  current training accuracy:		0.793083
Seen 27000 samples, time used: 125.848s
  current training loss:		0.564304
  current training accuracy:		0.792815
Seen 30000 samples, time used: 123.011s
  current training loss:		0.566729
  current training accuracy:		0.791367
saving to ..., time used: 1245.384s
Seen 33000 samples, time used: 123.351s
  current training loss:		0.566130
  current training accuracy:		0.791545
Seen 36000 samples, time used: 123.709s
  current training loss:		0.564811
  current training accuracy:		0.791889
Seen 39000 samples, time used: 124.315s
  current training loss:		0.562805
  current training accuracy:		0.792641
Seen 42000 samples, time used: 126.277s
  current training loss:		0.562145
  current training accuracy:		0.792762
Seen 45000 samples, time used: 123.056s
  current training loss:		0.563214
  current training accuracy:		0.792311
Seen 48000 samples, time used: 125.306s
  current training loss:		0.562709
  current training accuracy:		0.792833
Seen 51000 samples, time used: 123.609s
  current training loss:		0.562533
  current training accuracy:		0.792843
Seen 54000 samples, time used: 124.674s
  current training loss:		0.562634
  current training accuracy:		0.792741
Seen 57000 samples, time used: 129.124s
  current training loss:		0.563577
  current training accuracy:		0.792123
Seen 60000 samples, time used: 122.974s
  current training loss:		0.562541
  current training accuracy:		0.793033
saving to ..., time used: 1246.181s
Seen 63000 samples, time used: 123.382s
  current training loss:		0.562816
  current training accuracy:		0.792746
Seen 66000 samples, time used: 127.295s
  current training loss:		0.563294
  current training accuracy:		0.792348
Seen 69000 samples, time used: 127.964s
  current training loss:		0.564582
  current training accuracy:		0.791870
Seen 72000 samples, time used: 125.874s
  current training loss:		0.564426
  current training accuracy:		0.792181
Seen 75000 samples, time used: 128.213s
  current training loss:		0.563493
  current training accuracy:		0.792507
Seen 78000 samples, time used: 123.910s
  current training loss:		0.563100
  current training accuracy:		0.792538
Seen 81000 samples, time used: 123.107s
  current training loss:		0.563725
  current training accuracy:		0.792667
Seen 84000 samples, time used: 122.250s
  current training loss:		0.563254
  current training accuracy:		0.792893

Seen 87000 samples, time used: 124.313s
  current training loss:		0.563425
  current training accuracy:		0.792966
Seen 90000 samples, time used: 121.288s
  current training loss:		0.563915
  current training accuracy:		0.792611
saving to ..., time used: 1247.393s
Seen 93000 samples, time used: 124.726s
  current training loss:		0.563595
  current training accuracy:		0.792989
Seen 96000 samples, time used: 126.783s
  current training loss:		0.562982
  current training accuracy:		0.793250
Seen 99000 samples, time used: 127.218s
  current training loss:		0.563152
  current training accuracy:		0.793263
Seen 102000 samples, time used: 125.325s
  current training loss:		0.562477
  current training accuracy:		0.793510
Seen 105000 samples, time used: 122.345s
  current training loss:		0.562052
  current training accuracy:		0.793667
Seen 108000 samples, time used: 125.245s
  current training loss:		0.561585
  current training accuracy:		0.793889
Seen 111000 samples, time used: 125.343s
  current training loss:		0.561511
  current training accuracy:		0.794207
Seen 114000 samples, time used: 123.046s
  current training loss:		0.560873
  current training accuracy:		0.794526
Seen 117000 samples, time used: 123.274s
  current training loss:		0.560337
  current training accuracy:		0.794786
Seen 120000 samples, time used: 124.271s
  current training loss:		0.560475
  current training accuracy:		0.794792
saving to ..., time used: 1247.120s
Seen 123000 samples, time used: 123.645s
  current training loss:		0.560191
  current training accuracy:		0.794756
Seen 126000 samples, time used: 121.857s
  current training loss:		0.559793
  current training accuracy:		0.795048
Seen 129000 samples, time used: 124.105s
  current training loss:		0.559643
  current training accuracy:		0.795109
Seen 132000 samples, time used: 126.508s
  current training loss:		0.559057
  current training accuracy:		0.795326
Seen 135000 samples, time used: 127.752s
  current training loss:		0.558942
  current training accuracy:		0.795163
Seen 138000 samples, time used: 127.490s
  current training loss:		0.558671
  current training accuracy:		0.795406
Seen 141000 samples, time used: 124.277s
  current training loss:		0.558010
  current training accuracy:		0.795667
Seen 144000 samples, time used: 125.189s
  current training loss:		0.557373
  current training accuracy:		0.795896
Seen 147000 samples, time used: 124.958s
  current training loss:		0.557018
  current training accuracy:		0.795939
Seen 150000 samples, time used: 126.339s
  current training loss:		0.556682
  current training accuracy:		0.796233
saving to ..., time used: 1251.914s
Seen 153000 samples, time used: 122.864s
  current training loss:		0.556046
  current training accuracy:		0.796477
Seen 156000 samples, time used: 124.007s
  current training loss:		0.555981
  current training accuracy:		0.796494
Seen 159000 samples, time used: 125.496s
  current training loss:		0.555637
  current training accuracy:		0.796553
Seen 162000 samples, time used: 123.663s
  current training loss:		0.554748
  current training accuracy:		0.797025
Seen 165000 samples, time used: 122.688s
  current training loss:		0.554503
  current training accuracy:		0.797079
Seen 168000 samples, time used: 124.208s
  current training loss:		0.553895
  current training accuracy:		0.797238
Seen 171000 samples, time used: 122.991s
  current training loss:		0.553506
  current training accuracy:		0.797333
Seen 174000 samples, time used: 126.740s
  current training loss:		0.553609
  current training accuracy:		0.797253
Seen 177000 samples, time used: 127.478s
  current training loss:		0.553652
  current training accuracy:		0.797322
Seen 180000 samples, time used: 123.268s
  current training loss:		0.553318
  current training accuracy:		0.797511
saving to ..., time used: 1243.202s
Seen 183000 samples, time used: 121.653s
  current training loss:		0.553227
  current training accuracy:		0.797443
Seen 186000 samples, time used: 123.950s
  current training loss:		0.552563
  current training accuracy:		0.797742
Seen 189000 samples, time used: 122.707s
  current training loss:		0.552437
  current training accuracy:		0.797942
Seen 192000 samples, time used: 125.109s
  current training loss:		0.551930
  current training accuracy:		0.798135
Seen 195000 samples, time used: 125.598s
  current training loss:		0.551489
  current training accuracy:		0.798349
Seen 198000 samples, time used: 127.895s
  current training loss:		0.551098
  current training accuracy:		0.798616
Seen 201000 samples, time used: 124.878s
  current training loss:		0.550763
  current training accuracy:		0.798736
Seen 204000 samples, time used: 124.177s
  current training loss:		0.550677
  current training accuracy:		0.798804
Seen 207000 samples, time used: 125.565s
  current training loss:		0.550874
  current training accuracy:		0.798710
Seen 210000 samples, time used: 124.309s
  current training loss:		0.550728
  current training accuracy:		0.798781
saving to ..., time used: 1245.641s
Seen 213000 samples, time used: 122.939s
  current training loss:		0.550445
  current training accuracy:		0.798808
Seen 216000 samples, time used: 125.089s
  current training loss:		0.549908
  current training accuracy:		0.799093
Seen 219000 samples, time used: 124.284s
  current training loss:		0.549557
  current training accuracy:		0.799224

Seen 222000 samples, time used: 123.999s
  current training loss:		0.549403
  current training accuracy:		0.799320
Seen 225000 samples, time used: 133.539s
  current training loss:		0.549564
  current training accuracy:		0.799284
Seen 228000 samples, time used: 126.162s
  current training loss:		0.549147
  current training accuracy:		0.799478
Seen 231000 samples, time used: 123.660s
  current training loss:		0.548734
  current training accuracy:		0.799727
Seen 234000 samples, time used: 128.673s
  current training loss:		0.548342
  current training accuracy:		0.799932
Seen 237000 samples, time used: 124.071s
  current training loss:		0.548011
  current training accuracy:		0.800051
Seen 240000 samples, time used: 126.850s
  current training loss:		0.547945
  current training accuracy:		0.800108
saving to ..., time used: 1259.066s
Seen 243000 samples, time used: 127.058s
  current training loss:		0.547607
  current training accuracy:		0.800160
Seen 246000 samples, time used: 126.253s
  current training loss:		0.547440
  current training accuracy:		0.800130
Seen 249000 samples, time used: 125.352s
  current training loss:		0.547322
  current training accuracy:		0.800157
Seen 252000 samples, time used: 123.823s
  current training loss:		0.546847
  current training accuracy:		0.800361
Seen 255000 samples, time used: 124.388s
  current training loss:		0.546619
  current training accuracy:		0.800463
Seen 258000 samples, time used: 125.377s
  current training loss:		0.546632
  current training accuracy:		0.800535
Seen 261000 samples, time used: 128.114s
  current training loss:		0.546443
  current training accuracy:		0.800540
Seen 264000 samples, time used: 124.668s
  current training loss:		0.546481
  current training accuracy:		0.800534
Seen 267000 samples, time used: 127.507s
  current training loss:		0.546292
  current training accuracy:		0.800670
Seen 270000 samples, time used: 123.009s
  current training loss:		0.546276
  current training accuracy:		0.800704
saving to ..., time used: 1255.303s
Seen 273000 samples, time used: 123.079s
  current training loss:		0.546053
  current training accuracy:		0.800912
Seen 276000 samples, time used: 124.250s
  current training loss:		0.545938
  current training accuracy:		0.800953
Seen 279000 samples, time used: 123.827s
  current training loss:		0.545683
  current training accuracy:		0.801097
Seen 282000 samples, time used: 123.692s
  current training loss:		0.545361
  current training accuracy:		0.801209
Seen 285000 samples, time used: 124.439s
  current training loss:		0.545292
  current training accuracy:		0.801221
Seen 288000 samples, time used: 124.823s
  current training loss:		0.544954
  current training accuracy:		0.801375
Seen 291000 samples, time used: 124.521s
  current training loss:		0.544681
  current training accuracy:		0.801450
Seen 294000 samples, time used: 123.502s
  current training loss:		0.544374
  current training accuracy:		0.801578
Seen 297000 samples, time used: 124.037s
  current training loss:		0.544411
  current training accuracy:		0.801515
Seen 300000 samples, time used: 125.618s
  current training loss:		0.544075
  current training accuracy:		0.801657
saving to ..., time used: 1241.590s
Seen 303000 samples, time used: 127.595s
  current training loss:		0.543950
  current training accuracy:		0.801680
Seen 306000 samples, time used: 127.936s
  current training loss:		0.543954
  current training accuracy:		0.801739
Seen 309000 samples, time used: 123.266s
  current training loss:		0.543756
  current training accuracy:		0.801851
Seen 312000 samples, time used: 124.229s
  current training loss:		0.543623
  current training accuracy:		0.801859
Seen 315000 samples, time used: 125.194s
  current training loss:		0.543538
  current training accuracy:		0.801838
Seen 318000 samples, time used: 126.357s
  current training loss:		0.543352
  current training accuracy:		0.801950
Seen 321000 samples, time used: 125.642s
  current training loss:		0.543053
  current training accuracy:		0.802115
Seen 324000 samples, time used: 124.395s
  current training loss:		0.542812
  current training accuracy:		0.802235
Seen 327000 samples, time used: 124.673s
  current training loss:		0.542491
  current training accuracy:		0.802376
Seen 330000 samples, time used: 124.575s
  current training loss:		0.542353
  current training accuracy:		0.802467
saving to ..., time used: 1253.663s
Seen 333000 samples, time used: 128.186s
  current training loss:		0.541986
  current training accuracy:		0.802613
Seen 336000 samples, time used: 125.880s
  current training loss:		0.541965
  current training accuracy:		0.802637
Seen 339000 samples, time used: 125.476s
  current training loss:		0.541835
  current training accuracy:		0.802673
Seen 342000 samples, time used: 125.173s
  current training loss:		0.541755
  current training accuracy:		0.802620
Seen 345000 samples, time used: 124.179s
  current training loss:		0.541626
  current training accuracy:		0.802696

Seen 348000 samples, time used: 122.348s
  current training loss:		0.541486
  current training accuracy:		0.802787
Seen 351000 samples, time used: 123.786s
  current training loss:		0.541309
  current training accuracy:		0.802980
Seen 354000 samples, time used: 124.615s
  current training loss:		0.540942
  current training accuracy:		0.803198
Seen 357000 samples, time used: 123.575s
  current training loss:		0.540763
  current training accuracy:		0.803305
Seen 360000 samples, time used: 121.136s
  current training loss:		0.540776
  current training accuracy:		0.803369
saving to ..., time used: 1244.153s
Seen 363000 samples, time used: 125.144s
  current training loss:		0.540984
  current training accuracy:		0.803350
Seen 366000 samples, time used: 125.137s
  current training loss:		0.540820
  current training accuracy:		0.803484
Seen 369000 samples, time used: 124.781s
  current training loss:		0.540531
  current training accuracy:		0.803610
Seen 372000 samples, time used: 123.645s
  current training loss:		0.540383
  current training accuracy:		0.803675
Seen 375000 samples, time used: 123.889s
  current training loss:		0.540344
  current training accuracy:		0.803725
Seen 378000 samples, time used: 123.939s
  current training loss:		0.540111
  current training accuracy:		0.803828
Seen 381000 samples, time used: 122.081s
  current training loss:		0.540083
  current training accuracy:		0.803861
Seen 384000 samples, time used: 122.250s
  current training loss:		0.539778
  current training accuracy:		0.804023
Seen 387000 samples, time used: 127.664s
  current training loss:		0.539880
  current training accuracy:		0.803987
Seen 390000 samples, time used: 125.779s
  current training loss:		0.539805
  current training accuracy:		0.804010
saving to ..., time used: 1244.108s
Seen 393000 samples, time used: 128.561s
  current training loss:		0.539661
  current training accuracy:		0.804076
Seen 396000 samples, time used: 122.442s
  current training loss:		0.539418
  current training accuracy:		0.804152
Seen 399000 samples, time used: 123.810s
  current training loss:		0.539119
  current training accuracy:		0.804346
Seen 402000 samples, time used: 123.599s
  current training loss:		0.538887
  current training accuracy:		0.804463
Seen 405000 samples, time used: 122.523s
  current training loss:		0.538695
  current training accuracy:		0.804526
Seen 408000 samples, time used: 124.341s
  current training loss:		0.538575
  current training accuracy:		0.804542
Seen 411000 samples, time used: 124.093s
  current training loss:		0.538439
  current training accuracy:		0.804647
Seen 414000 samples, time used: 123.486s
  current training loss:		0.538165
  current training accuracy:		0.804744
Seen 417000 samples, time used: 124.832s
  current training loss:		0.538023
  current training accuracy:		0.804844
Seen 420000 samples, time used: 124.139s
  current training loss:		0.537937
  current training accuracy:		0.804933
saving to ..., time used: 1241.627s
Seen 423000 samples, time used: 124.984s
  current training loss:		0.537808
  current training accuracy:		0.804950
Seen 426000 samples, time used: 123.727s
  current training loss:		0.537768
  current training accuracy:		0.804974
Seen 429000 samples, time used: 124.216s
  current training loss:		0.537615
  current training accuracy:		0.805037
Seen 432000 samples, time used: 127.038s
  current training loss:		0.537486
  current training accuracy:		0.805116
Seen 435000 samples, time used: 125.426s
  current training loss:		0.537315
  current training accuracy:		0.805193
Seen 438000 samples, time used: 122.931s
  current training loss:		0.537248
  current training accuracy:		0.805235
Seen 441000 samples, time used: 128.490s
  current training loss:		0.537153
  current training accuracy:		0.805277
Seen 444000 samples, time used: 120.585s
  current training loss:		0.536925
  current training accuracy:		0.805381
Seen 447000 samples, time used: 124.810s
  current training loss:		0.536682
  current training accuracy:		0.805461
Seen 450000 samples, time used: 123.913s
  current training loss:		0.536559
  current training accuracy:		0.805484
saving to ..., time used: 1245.922s
Seen 453000 samples, time used: 125.127s
  current training loss:		0.536404
  current training accuracy:		0.805581
Seen 456000 samples, time used: 125.487s
  current training loss:		0.536341
  current training accuracy:		0.805564
Seen 459000 samples, time used: 126.971s
  current training loss:		0.536365
  current training accuracy:		0.805599
Seen 462000 samples, time used: 129.326s
  current training loss:		0.536325
  current training accuracy:		0.805589
Seen 465000 samples, time used: 122.735s
  current training loss:		0.536238
  current training accuracy:		0.805669
Seen 468000 samples, time used: 126.608s
  current training loss:		0.536205
  current training accuracy:		0.805699
Seen 471000 samples, time used: 123.212s
  current training loss:		0.536021
  current training accuracy:		0.805786
Seen 474000 samples, time used: 125.910s
  current training loss:		0.535795
  current training accuracy:		0.805884
Seen 477000 samples, time used: 124.123s
  current training loss:		0.535690
  current training accuracy:		0.805901
Seen 480000 samples, time used: 121.898s
  current training loss:		0.535470
  current training accuracy:		0.805937
saving to ..., time used: 1251.197s
Seen 483000 samples, time used: 125.552s
  current training loss:		0.535360
  current training accuracy:		0.805963
Seen 486000 samples, time used: 126.585s
  current training loss:		0.535347
  current training accuracy:		0.805965
Seen 489000 samples, time used: 125.950s
  current training loss:		0.535201
  current training accuracy:		0.806002
Seen 492000 samples, time used: 128.612s
  current training loss:		0.535235
  current training accuracy:		0.805941
Seen 495000 samples, time used: 127.719s
  current training loss:		0.535029
  current training accuracy:		0.805980
Seen 498000 samples, time used: 123.595s
  current training loss:		0.534940
  current training accuracy:		0.806014
Seen 501000 samples, time used: 122.273s
  current training loss:		0.534581
  current training accuracy:		0.806196
Seen 504000 samples, time used: 123.955s
  current training loss:		0.534288
  current training accuracy:		0.806298
Seen 507000 samples, time used: 123.670s
  current training loss:		0.534125
  current training accuracy:		0.806393
Seen 510000 samples, time used: 125.678s
  current training loss:		0.534025
  current training accuracy:		0.806412
saving to ..., time used: 1253.390s
Seen 513000 samples, time used: 123.115s
  current training loss:		0.533924
  current training accuracy:		0.806419
Seen 516000 samples, time used: 123.911s
  current training loss:		0.533827
  current training accuracy:		0.806467
Seen 519000 samples, time used: 125.446s
  current training loss:		0.533507
  current training accuracy:		0.806605
Seen 522000 samples, time used: 124.163s
  current training loss:		0.533246
  current training accuracy:		0.806730
Seen 525000 samples, time used: 122.244s
  current training loss:		0.533083
  current training accuracy:		0.806787
Seen 528000 samples, time used: 123.141s
  current training loss:		0.532903
  current training accuracy:		0.806850
Seen 531000 samples, time used: 124.582s
  current training loss:		0.532782
  current training accuracy:		0.806895
Seen 534000 samples, time used: 128.856s
  current training loss:		0.532677
  current training accuracy:		0.806948
Seen 537000 samples, time used: 123.619s
  current training loss:		0.532645
  current training accuracy:		0.806983
Seen 540000 samples, time used: 123.966s
  current training loss:		0.532566
  current training accuracy:		0.807035
saving to ..., time used: 1242.844s
Seen 543000 samples, time used: 123.165s
  current training loss:		0.532347
  current training accuracy:		0.807145
Seen 546000 samples, time used: 123.668s
  current training loss:		0.532230
  current training accuracy:		0.807183
Seen 549000 samples, time used: 126.168s
  current training loss:		0.532057
  current training accuracy:		0.807211
Epoch 2 of 20 took 22880.774s
  training loss:		0.532071
  training accuracy:		80.72 %
  validation loss:		0.467812
  validation accuracy:		82.10 %
Seen 3000 samples, time used: 125.763s
  current training loss:		0.484524
  current training accuracy:		0.826333
Seen 6000 samples, time used: 121.773s
  current training loss:		0.487100
  current training accuracy:		0.826833
Seen 9000 samples, time used: 127.167s
  current training loss:		0.486222
  current training accuracy:		0.827556
Seen 12000 samples, time used: 125.709s
  current training loss:		0.487551
  current training accuracy:		0.825417
Seen 15000 samples, time used: 125.025s
  current training loss:		0.487155
  current training accuracy:		0.826533
Seen 18000 samples, time used: 123.668s
  current training loss:		0.485685
  current training accuracy:		0.827500
Seen 21000 samples, time used: 123.952s
  current training loss:		0.484233
  current training accuracy:		0.828238
Seen 24000 samples, time used: 125.294s
  current training loss:		0.483477
  current training accuracy:		0.828292
Seen 27000 samples, time used: 126.602s
  current training loss:		0.482771
  current training accuracy:		0.829296
Seen 30000 samples, time used: 127.287s
  current training loss:		0.481835
  current training accuracy:		0.828833
saving to ..., time used: 1252.241s
Seen 33000 samples, time used: 124.572s
  current training loss:		0.482610
  current training accuracy:		0.827939
Seen 36000 samples, time used: 126.096s
  current training loss:		0.484539
  current training accuracy:		0.827444
Seen 39000 samples, time used: 121.386s
  current training loss:		0.485178
  current training accuracy:		0.828077
Seen 42000 samples, time used: 122.079s
  current training loss:		0.483497
  current training accuracy:		0.828762
Seen 45000 samples, time used: 124.887s
  current training loss:		0.484847
  current training accuracy:		0.828644
Seen 48000 samples, time used: 124.113s
  current training loss:		0.484659
  current training accuracy:		0.828979
Seen 51000 samples, time used: 125.868s
  current training loss:		0.483387
  current training accuracy:		0.829314
Seen 54000 samples, time used: 127.174s
  current training loss:		0.483210
  current training accuracy:		0.829463
Seen 57000 samples, time used: 124.583s
  current training loss:		0.482571
  current training accuracy:		0.829632
Seen 60000 samples, time used: 123.063s
  current training loss:		0.482922
  current training accuracy:		0.829567
saving to ..., time used: 1243.620s
Seen 63000 samples, time used: 125.518s
  current training loss:		0.482690
  current training accuracy:		0.829381
Seen 66000 samples, time used: 122.981s
  current training loss:		0.481710
  current training accuracy:		0.829894
Seen 69000 samples, time used: 123.589s
  current training loss:		0.482822
  current training accuracy:		0.829159
Seen 72000 samples, time used: 126.385s
  current training loss:		0.482424
  current training accuracy:		0.829194
Seen 75000 samples, time used: 125.896s
  current training loss:		0.482720
  current training accuracy:		0.829107
Seen 78000 samples, time used: 127.693s
  current training loss:		0.483153
  current training accuracy:		0.829308
Seen 81000 samples, time used: 124.901s
  current training loss:		0.481955
  current training accuracy:		0.829790
Seen 84000 samples, time used: 124.113s
  current training loss:		0.482094
  current training accuracy:		0.829607
Seen 87000 samples, time used: 124.831s
  current training loss:		0.481789
  current training accuracy:		0.829667
Seen 90000 samples, time used: 125.520s
  current training loss:		0.481237
  current training accuracy:		0.829911
saving to ..., time used: 1251.230s
Seen 93000 samples, time used: 125.160s
  current training loss:		0.480168
  current training accuracy:		0.830290
Seen 96000 samples, time used: 124.630s
  current training loss:		0.480015
  current training accuracy:		0.830156
Seen 99000 samples, time used: 123.629s
  current training loss:		0.479586
  current training accuracy:		0.830313
Seen 102000 samples, time used: 125.914s
  current training loss:		0.479666
  current training accuracy:		0.830353
Seen 105000 samples, time used: 122.759s
  current training loss:		0.479451
  current training accuracy:		0.830648
Seen 108000 samples, time used: 124.479s
  current training loss:		0.478835
  current training accuracy:		0.830963
Seen 111000 samples, time used: 122.440s
  current training loss:		0.478724
  current training accuracy:		0.830856
Seen 114000 samples, time used: 124.752s
  current training loss:		0.478622
  current training accuracy:		0.830851
Seen 117000 samples, time used: 124.558s
  current training loss:		0.478338
  current training accuracy:		0.830983
Seen 120000 samples, time used: 128.201s
  current training loss:		0.477880
  current training accuracy:		0.831217
saving to ..., time used: 1246.323s
Seen 123000 samples, time used: 124.443s
  current training loss:		0.478454
  current training accuracy:		0.830976
Seen 126000 samples, time used: 124.889s
  current training loss:		0.478429
  current training accuracy:		0.830968
Seen 129000 samples, time used: 122.692s
  current training loss:		0.478567
  current training accuracy:		0.830752
Seen 132000 samples, time used: 123.302s
  current training loss:		0.478313
  current training accuracy:		0.830955
Seen 135000 samples, time used: 124.613s
  current training loss:		0.478409
  current training accuracy:		0.831022
Seen 138000 samples, time used: 123.133s
  current training loss:		0.477934
  current training accuracy:		0.831297
Seen 141000 samples, time used: 124.943s
  current training loss:		0.478049
  current training accuracy:		0.831106
Seen 144000 samples, time used: 127.947s
  current training loss:		0.477900
  current training accuracy:		0.831174
Seen 147000 samples, time used: 129.797s
  current training loss:		0.477789
  current training accuracy:		0.831082
Seen 150000 samples, time used: 125.483s
  current training loss:		0.477766
  current training accuracy:		0.831220
saving to ..., time used: 1251.045s
Seen 153000 samples, time used: 126.850s
  current training loss:		0.477788
  current training accuracy:		0.831222
Seen 156000 samples, time used: 125.083s
  current training loss:		0.478297
  current training accuracy:		0.830891
Seen 159000 samples, time used: 126.671s
  current training loss:		0.478216
  current training accuracy:		0.830862
Seen 162000 samples, time used: 123.743s
  current training loss:		0.478212
  current training accuracy:		0.830815
Seen 165000 samples, time used: 126.702s
  current training loss:		0.478253
  current training accuracy:		0.830770
Seen 168000 samples, time used: 124.558s
  current training loss:		0.478422
  current training accuracy:		0.830679
Seen 171000 samples, time used: 126.045s
  current training loss:		0.479057
  current training accuracy:		0.830456
Seen 174000 samples, time used: 125.292s
  current training loss:		0.478283
  current training accuracy:		0.830759
Seen 177000 samples, time used: 124.862s
  current training loss:		0.478703
  current training accuracy:		0.830588
Seen 180000 samples, time used: 127.993s
  current training loss:		0.478958
  current training accuracy:		0.830511
saving to ..., time used: 1257.599s
Seen 183000 samples, time used: 122.769s
  current training loss:		0.478730
  current training accuracy:		0.830552
Seen 186000 samples, time used: 126.162s
  current training loss:		0.478524
  current training accuracy:		0.830640
Seen 189000 samples, time used: 124.732s
  current training loss:		0.479065
  current training accuracy:		0.830365
Seen 192000 samples, time used: 121.270s
  current training loss:		0.479302
  current training accuracy:		0.830385
Seen 195000 samples, time used: 123.129s
  current training loss:		0.479169
  current training accuracy:		0.830456
Seen 198000 samples, time used: 125.483s
  current training loss:		0.479210
  current training accuracy:		0.830510
Seen 201000 samples, time used: 123.963s
  current training loss:		0.479615
  current training accuracy:		0.830259
Seen 204000 samples, time used: 127.206s
  current training loss:		0.479857
  current training accuracy:		0.830270
Seen 207000 samples, time used: 125.019s
  current training loss:		0.479620
  current training accuracy:		0.830357
Seen 210000 samples, time used: 125.511s
  current training loss:		0.479526
  current training accuracy:		0.830414
saving to ..., time used: 1245.045s
Seen 213000 samples, time used: 128.462s
  current training loss:		0.479571
  current training accuracy:		0.830418
Seen 216000 samples, time used: 125.116s
  current training loss:		0.479334
  current training accuracy:		0.830495
Seen 219000 samples, time used: 127.725s
  current training loss:		0.479545
  current training accuracy:		0.830461
Seen 222000 samples, time used: 126.736s
  current training loss:		0.479249
  current training accuracy:		0.830568
Seen 225000 samples, time used: 122.848s
  current training loss:		0.478695
  current training accuracy:		0.830707
Seen 228000 samples, time used: 124.677s
  current training loss:		0.478584
  current training accuracy:		0.830781
Seen 231000 samples, time used: 123.937s
  current training loss:		0.478545
  current training accuracy:		0.830892
Seen 234000 samples, time used: 123.978s
  current training loss:		0.478362
  current training accuracy:		0.830919
Seen 237000 samples, time used: 123.644s
  current training loss:		0.478095
  current training accuracy:		0.831021
Seen 240000 samples, time used: 126.025s
  current training loss:		0.478301
  current training accuracy:		0.830921
saving to ..., time used: 1252.949s
Seen 243000 samples, time used: 125.990s
  current training loss:		0.478265
  current training accuracy:		0.830914
Seen 246000 samples, time used: 121.752s
  current training loss:		0.478114
  current training accuracy:		0.830915
Seen 249000 samples, time used: 124.096s
  current training loss:		0.478151
  current training accuracy:		0.830944
Seen 252000 samples, time used: 126.029s
  current training loss:		0.478291
  current training accuracy:		0.830869
Seen 255000 samples, time used: 126.584s
  current training loss:		0.478329
  current training accuracy:		0.830812
Seen 258000 samples, time used: 126.686s
  current training loss:		0.478441
  current training accuracy:		0.830860
Seen 261000 samples, time used: 125.780s
  current training loss:		0.478531
  current training accuracy:		0.830728
Seen 264000 samples, time used: 125.708s
  current training loss:		0.478602
  current training accuracy:		0.830697
Seen 267000 samples, time used: 125.335s
  current training loss:		0.478479
  current training accuracy:		0.830757
Seen 270000 samples, time used: 126.228s
  current training loss:		0.478339
  current training accuracy:		0.830863
saving to ..., time used: 1253.987s
Seen 273000 samples, time used: 125.062s
  current training loss:		0.478208
  current training accuracy:		0.830879
Seen 276000 samples, time used: 124.049s
  current training loss:		0.478160
  current training accuracy:		0.830880
Seen 279000 samples, time used: 125.524s
  current training loss:		0.478282
  current training accuracy:		0.830846
Seen 282000 samples, time used: 126.700s
  current training loss:		0.478232
  current training accuracy:		0.830848
Seen 285000 samples, time used: 125.775s
  current training loss:		0.478000
  current training accuracy:		0.830888
Seen 288000 samples, time used: 123.153s
  current training loss:		0.477965
  current training accuracy:		0.831007
Seen 291000 samples, time used: 123.648s
  current training loss:		0.477870
  current training accuracy:		0.831027
Seen 294000 samples, time used: 126.277s
  current training loss:		0.477627
  current training accuracy:		0.831187
Seen 297000 samples, time used: 126.470s
  current training loss:		0.477628
  current training accuracy:		0.831098
Seen 300000 samples, time used: 124.410s
  current training loss:		0.477584
  current training accuracy:		0.831077
saving to ..., time used: 1250.867s
Seen 303000 samples, time used: 126.978s
  current training loss:		0.477210
  current training accuracy:		0.831221
Seen 306000 samples, time used: 121.783s
  current training loss:		0.476774
  current training accuracy:		0.831353
Seen 309000 samples, time used: 128.511s
  current training loss:		0.476508
  current training accuracy:		0.831508
Seen 312000 samples, time used: 122.288s
  current training loss:		0.476363
  current training accuracy:		0.831481
Seen 315000 samples, time used: 124.249s
  current training loss:		0.476271
  current training accuracy:		0.831483
Seen 318000 samples, time used: 125.253s
  current training loss:		0.476250
  current training accuracy:		0.831478
Seen 321000 samples, time used: 124.257s
  current training loss:		0.476074
  current training accuracy:		0.831595
Seen 324000 samples, time used: 122.226s
  current training loss:		0.475885
  current training accuracy:		0.831630
Seen 327000 samples, time used: 125.876s
  current training loss:		0.475737
  current training accuracy:		0.831691
Seen 330000 samples, time used: 122.987s
  current training loss:		0.475941
  current training accuracy:		0.831615
saving to ..., time used: 1244.209s
Seen 333000 samples, time used: 125.179s
  current training loss:		0.475964
  current training accuracy:		0.831628
Seen 336000 samples, time used: 122.277s
  current training loss:		0.475980
  current training accuracy:		0.831646
Seen 339000 samples, time used: 126.199s
  current training loss:		0.476018
  current training accuracy:		0.831637
Seen 342000 samples, time used: 125.964s
  current training loss:		0.476089
  current training accuracy:		0.831649
Seen 345000 samples, time used: 125.983s
  current training loss:		0.475860
  current training accuracy:		0.831733
Seen 348000 samples, time used: 125.083s
  current training loss:		0.475945
  current training accuracy:		0.831698
Seen 351000 samples, time used: 124.799s
  current training loss:		0.475852
  current training accuracy:		0.831684
Seen 354000 samples, time used: 126.292s
  current training loss:		0.475894
  current training accuracy:		0.831636
Seen 357000 samples, time used: 125.788s
  current training loss:		0.475735
  current training accuracy:		0.831756
Seen 360000 samples, time used: 126.527s
  current training loss:		0.475864
  current training accuracy:		0.831656
saving to ..., time used: 1253.892s
Seen 363000 samples, time used: 125.398s
  current training loss:		0.475732
  current training accuracy:		0.831713
Seen 366000 samples, time used: 125.216s
  current training loss:		0.475803
  current training accuracy:		0.831631
Seen 369000 samples, time used: 124.562s
  current training loss:		0.475643
  current training accuracy:		0.831743
Seen 372000 samples, time used: 124.646s
  current training loss:		0.475547
  current training accuracy:		0.831761
Seen 375000 samples, time used: 122.906s
  current training loss:		0.475516
  current training accuracy:		0.831792
Seen 378000 samples, time used: 125.124s
  current training loss:		0.475644
  current training accuracy:		0.831738
Seen 381000 samples, time used: 125.302s
  current training loss:		0.475752
  current training accuracy:		0.831659
Seen 384000 samples, time used: 128.702s
  current training loss:		0.475787
  current training accuracy:		0.831661
Seen 387000 samples, time used: 124.384s
  current training loss:		0.475776
  current training accuracy:		0.831638
Seen 390000 samples, time used: 123.430s
  current training loss:		0.475579
  current training accuracy:		0.831662
saving to ..., time used: 1249.472s
Seen 393000 samples, time used: 125.176s
  current training loss:		0.475515
  current training accuracy:		0.831718
Seen 396000 samples, time used: 125.492s
  current training loss:		0.475342
  current training accuracy:		0.831801
Seen 399000 samples, time used: 125.647s
  current training loss:		0.475242
  current training accuracy:		0.831805
Seen 402000 samples, time used: 122.364s
  current training loss:		0.475263
  current training accuracy:		0.831838
Seen 405000 samples, time used: 125.268s
  current training loss:		0.475274
  current training accuracy:		0.831872
Seen 408000 samples, time used: 122.791s
  current training loss:		0.475278
  current training accuracy:		0.831980
Seen 411000 samples, time used: 120.579s
  current training loss:		0.475096
  current training accuracy:		0.832039
Seen 414000 samples, time used: 125.626s
  current training loss:		0.475031
  current training accuracy:		0.832087
Seen 417000 samples, time used: 125.537s
  current training loss:		0.474941
  current training accuracy:		0.832120
Seen 420000 samples, time used: 124.494s
  current training loss:		0.474847
  current training accuracy:		0.832176
saving to ..., time used: 1242.777s
Seen 423000 samples, time used: 125.581s
  current training loss:		0.474800
  current training accuracy:		0.832210
Seen 426000 samples, time used: 123.964s
  current training loss:		0.474846
  current training accuracy:		0.832216
Seen 429000 samples, time used: 124.658s
  current training loss:		0.474754
  current training accuracy:		0.832231
Seen 432000 samples, time used: 126.737s
  current training loss:		0.474707
  current training accuracy:		0.832264
Seen 435000 samples, time used: 126.266s
  current training loss:		0.474622
  current training accuracy:		0.832343
Seen 438000 samples, time used: 125.433s
  current training loss:		0.474732
  current training accuracy:		0.832256
Seen 441000 samples, time used: 125.353s
  current training loss:		0.474763
  current training accuracy:		0.832249
Seen 444000 samples, time used: 127.999s
  current training loss:		0.474650
  current training accuracy:		0.832313
Seen 447000 samples, time used: 122.079s
  current training loss:		0.474373
  current training accuracy:		0.832407
Seen 450000 samples, time used: 125.642s
  current training loss:		0.474302
  current training accuracy:		0.832420
saving to ..., time used: 1253.513s
Seen 453000 samples, time used: 124.783s
  current training loss:		0.474289
  current training accuracy:		0.832428
Seen 456000 samples, time used: 125.196s
  current training loss:		0.474384
  current training accuracy:		0.832388
Seen 459000 samples, time used: 122.878s
  current training loss:		0.474343
  current training accuracy:		0.832420
Seen 462000 samples, time used: 125.535s
  current training loss:		0.474247
  current training accuracy:		0.832457
Seen 465000 samples, time used: 122.334s
  current training loss:		0.474171
  current training accuracy:		0.832514
Seen 468000 samples, time used: 125.217s
  current training loss:		0.474189
  current training accuracy:		0.832509
Seen 471000 samples, time used: 124.270s
  current training loss:		0.474116
  current training accuracy:		0.832533
Seen 474000 samples, time used: 125.283s
  current training loss:		0.473983
  current training accuracy:		0.832622
Seen 477000 samples, time used: 124.673s
  current training loss:		0.474144
  current training accuracy:		0.832551
Seen 480000 samples, time used: 123.930s
  current training loss:		0.474155
  current training accuracy:		0.832583
saving to ..., time used: 1243.902s
Seen 483000 samples, time used: 125.179s
  current training loss:		0.474137
  current training accuracy:		0.832578
Seen 486000 samples, time used: 125.613s
  current training loss:		0.474135
  current training accuracy:		0.832623
Seen 489000 samples, time used: 125.019s
  current training loss:		0.474153
  current training accuracy:		0.832630
Seen 492000 samples, time used: 120.338s
  current training loss:		0.474140
  current training accuracy:		0.832624
Seen 495000 samples, time used: 126.706s
  current training loss:		0.474017
  current training accuracy:		0.832628
Seen 498000 samples, time used: 122.798s
  current training loss:		0.474028
  current training accuracy:		0.832618
Seen 501000 samples, time used: 126.389s
  current training loss:		0.474103
  current training accuracy:		0.832563
Seen 504000 samples, time used: 125.049s
  current training loss:		0.474009
  current training accuracy:		0.832605
Seen 507000 samples, time used: 124.364s
  current training loss:		0.473981
  current training accuracy:		0.832574
Seen 510000 samples, time used: 123.809s
  current training loss:		0.473791
  current training accuracy:		0.832635
saving to ..., time used: 1245.067s
Seen 513000 samples, time used: 128.428s
  current training loss:		0.473863
  current training accuracy:		0.832647
Seen 516000 samples, time used: 125.543s
  current training loss:		0.473787
  current training accuracy:		0.832678
Seen 519000 samples, time used: 124.899s
  current training loss:		0.473827
  current training accuracy:		0.832655
Seen 522000 samples, time used: 123.218s
  current training loss:		0.473823
  current training accuracy:		0.832670
Seen 525000 samples, time used: 124.172s
  current training loss:		0.473855
  current training accuracy:		0.832680
Seen 528000 samples, time used: 123.692s
  current training loss:		0.473860
  current training accuracy:		0.832652
Seen 531000 samples, time used: 126.802s
  current training loss:		0.473790
  current training accuracy:		0.832699
Seen 534000 samples, time used: 125.490s
  current training loss:		0.473700
  current training accuracy:		0.832725
Seen 537000 samples, time used: 126.066s
  current training loss:		0.473609
  current training accuracy:		0.832747
Seen 540000 samples, time used: 125.196s
  current training loss:		0.473514
  current training accuracy:		0.832772
saving to ..., time used: 1253.306s
Seen 543000 samples, time used: 124.037s
  current training loss:		0.473401
  current training accuracy:		0.832834
Seen 546000 samples, time used: 124.296s
  current training loss:		0.473486
  current training accuracy:		0.832839
Seen 549000 samples, time used: 125.614s
  current training loss:		0.473597
  current training accuracy:		0.832803
Epoch 3 of 20 took 22913.076s
  training loss:		0.473640
  training accuracy:		83.28 %
  validation loss:		0.432471
  validation accuracy:		83.19 %
Seen 3000 samples, time used: 126.233s
  current training loss:		0.440715
  current training accuracy:		0.844333
Seen 6000 samples, time used: 122.224s
  current training loss:		0.440946
  current training accuracy:		0.849333
Seen 9000 samples, time used: 128.777s
  current training loss:		0.441248
  current training accuracy:		0.850111
Seen 12000 samples, time used: 124.099s
  current training loss:		0.442640
  current training accuracy:		0.848000
Seen 15000 samples, time used: 126.164s
  current training loss:		0.441153
  current training accuracy:		0.848667
Seen 18000 samples, time used: 122.313s
  current training loss:		0.439964
  current training accuracy:		0.848444
Seen 21000 samples, time used: 122.809s
  current training loss:		0.440067
  current training accuracy:		0.848810
Seen 24000 samples, time used: 125.795s
  current training loss:		0.440989
  current training accuracy:		0.848250
Seen 27000 samples, time used: 124.825s
  current training loss:		0.439604
  current training accuracy:		0.848926
Seen 30000 samples, time used: 124.331s
  current training loss:		0.438707
  current training accuracy:		0.850000
saving to ..., time used: 1247.571s
Seen 33000 samples, time used: 124.080s
  current training loss:		0.439665
  current training accuracy:		0.849667
Seen 36000 samples, time used: 125.454s
  current training loss:		0.439019
  current training accuracy:		0.849917
Seen 39000 samples, time used: 124.486s
  current training loss:		0.440147
  current training accuracy:		0.849667
Seen 42000 samples, time used: 124.426s
  current training loss:		0.439775
  current training accuracy:		0.849881
Seen 45000 samples, time used: 124.491s
  current training loss:		0.440411
  current training accuracy:		0.849378
Seen 48000 samples, time used: 124.201s
  current training loss:		0.440363
  current training accuracy:		0.849479
Seen 51000 samples, time used: 127.657s
  current training loss:		0.440361
  current training accuracy:		0.849549
Seen 54000 samples, time used: 126.035s
  current training loss:		0.440498
  current training accuracy:		0.849593
Seen 57000 samples, time used: 122.532s
  current training loss:		0.439060
  current training accuracy:		0.850351
Seen 60000 samples, time used: 125.296s
  current training loss:		0.439316
  current training accuracy:		0.850383
saving to ..., time used: 1248.458s
Seen 63000 samples, time used: 123.739s
  current training loss:		0.437911
  current training accuracy:		0.850889
Seen 66000 samples, time used: 124.551s
  current training loss:		0.437821
  current training accuracy:		0.850879
Seen 69000 samples, time used: 122.801s
  current training loss:		0.437386
  current training accuracy:		0.850971
Seen 72000 samples, time used: 128.990s
  current training loss:		0.436855
  current training accuracy:		0.850958
Seen 75000 samples, time used: 128.512s
  current training loss:		0.435912
  current training accuracy:		0.851267
Seen 78000 samples, time used: 128.804s
  current training loss:		0.435785
  current training accuracy:		0.851231
Seen 81000 samples, time used: 123.226s
  current training loss:		0.436330
  current training accuracy:		0.850901
Seen 84000 samples, time used: 125.532s
  current training loss:		0.436175
  current training accuracy:		0.851131
Seen 87000 samples, time used: 123.136s
  current training loss:		0.436427
  current training accuracy:		0.850816
Seen 90000 samples, time used: 125.268s
  current training loss:		0.437160
  current training accuracy:		0.850600
saving to ..., time used: 1254.360s
Seen 93000 samples, time used: 125.355s
  current training loss:		0.437652
  current training accuracy:		0.850376
Seen 96000 samples, time used: 128.391s
  current training loss:		0.437776
  current training accuracy:		0.850198
Seen 99000 samples, time used: 129.545s
  current training loss:		0.437670
  current training accuracy:		0.850222
Seen 102000 samples, time used: 123.669s
  current training loss:		0.437352
  current training accuracy:		0.850196
Seen 105000 samples, time used: 122.100s
  current training loss:		0.437336
  current training accuracy:		0.850295
Seen 108000 samples, time used: 123.468s
  current training loss:		0.437294
  current training accuracy:		0.850380
Seen 111000 samples, time used: 122.307s
  current training loss:		0.437203
  current training accuracy:		0.850351
Seen 114000 samples, time used: 126.248s
  current training loss:		0.436738
  current training accuracy:		0.850544
Seen 117000 samples, time used: 124.735s
  current training loss:		0.437500
  current training accuracy:		0.850239
Seen 120000 samples, time used: 123.360s
  current training loss:		0.437302
  current training accuracy:		0.850375
saving to ..., time used: 1248.976s
Seen 123000 samples, time used: 127.353s
  current training loss:		0.437615
  current training accuracy:		0.850293
Seen 126000 samples, time used: 124.715s
  current training loss:		0.438351
  current training accuracy:		0.850079
Seen 129000 samples, time used: 125.284s
  current training loss:		0.438554
  current training accuracy:		0.850039
Seen 132000 samples, time used: 126.592s
  current training loss:		0.438191
  current training accuracy:		0.850000
Seen 135000 samples, time used: 125.429s
  current training loss:		0.438866
  current training accuracy:		0.849881
Seen 138000 samples, time used: 125.796s
  current training loss:		0.439357
  current training accuracy:		0.849551
Seen 141000 samples, time used: 126.516s
  current training loss:		0.439715
  current training accuracy:		0.849511
Seen 144000 samples, time used: 125.574s
  current training loss:		0.439690
  current training accuracy:		0.849507
Seen 147000 samples, time used: 123.174s
  current training loss:		0.439592
  current training accuracy:		0.849388
Seen 150000 samples, time used: 126.291s
  current training loss:		0.439195
  current training accuracy:		0.849420
saving to ..., time used: 1256.525s
Seen 153000 samples, time used: 122.268s
  current training loss:		0.439371
  current training accuracy:		0.849275
Seen 156000 samples, time used: 126.185s
  current training loss:		0.439411
  current training accuracy:		0.849250
Seen 159000 samples, time used: 123.096s
  current training loss:		0.439688
  current training accuracy:		0.849044
Seen 162000 samples, time used: 125.495s
  current training loss:		0.439407
  current training accuracy:		0.849216
Seen 165000 samples, time used: 127.109s
  current training loss:		0.439645
  current training accuracy:		0.849079
Seen 168000 samples, time used: 127.215s
  current training loss:		0.439506
  current training accuracy:		0.849167
Seen 171000 samples, time used: 126.645s
  current training loss:		0.439687
  current training accuracy:		0.849058
Seen 174000 samples, time used: 123.882s
  current training loss:		0.439577
  current training accuracy:		0.849138
Seen 177000 samples, time used: 126.080s
  current training loss:		0.439591
  current training accuracy:		0.849090
Seen 180000 samples, time used: 125.144s
  current training loss:		0.439470
  current training accuracy:		0.849133
saving to ..., time used: 1252.918s
Seen 183000 samples, time used: 127.339s
  current training loss:		0.439690
  current training accuracy:		0.849109
Seen 186000 samples, time used: 126.733s
  current training loss:		0.439853
  current training accuracy:		0.848909
Seen 189000 samples, time used: 126.185s
  current training loss:		0.440133
  current training accuracy:		0.848762
Seen 192000 samples, time used: 123.564s
  current training loss:		0.440311
  current training accuracy:		0.848630
Seen 195000 samples, time used: 128.529s
  current training loss:		0.440289
  current training accuracy:		0.848703
Seen 198000 samples, time used: 125.774s
  current training loss:		0.440110
  current training accuracy:		0.848758
Seen 201000 samples, time used: 126.318s
  current training loss:		0.440159
  current training accuracy:		0.848766
Seen 204000 samples, time used: 124.700s
  current training loss:		0.440181
  current training accuracy:		0.848725
Seen 207000 samples, time used: 124.251s
  current training loss:		0.440241
  current training accuracy:		0.848778
Seen 210000 samples, time used: 123.635s
  current training loss:		0.439953
  current training accuracy:		0.848924
saving to ..., time used: 1256.828s
Seen 213000 samples, time used: 125.030s
  current training loss:		0.439833
  current training accuracy:		0.848953
Seen 216000 samples, time used: 121.249s
  current training loss:		0.439936
  current training accuracy:		0.848894
Seen 219000 samples, time used: 127.404s
  current training loss:		0.439911
  current training accuracy:		0.848909
Seen 222000 samples, time used: 124.687s
  current training loss:		0.439706
  current training accuracy:		0.848982
Seen 225000 samples, time used: 125.241s
  current training loss:		0.439820
  current training accuracy:		0.848938
Seen 228000 samples, time used: 125.296s
  current training loss:		0.439857
  current training accuracy:		0.848982
Seen 231000 samples, time used: 123.078s
  current training loss:		0.439878
  current training accuracy:		0.848978
Seen 234000 samples, time used: 125.157s
  current training loss:		0.439977
  current training accuracy:		0.848927
Seen 237000 samples, time used: 127.188s
  current training loss:		0.439653
  current training accuracy:		0.849046
Seen 240000 samples, time used: 127.622s
  current training loss:		0.439137
  current training accuracy:		0.849246
saving to ..., time used: 1251.752s
Seen 243000 samples, time used: 124.456s
  current training loss:		0.439362
  current training accuracy:		0.849144
Seen 246000 samples, time used: 123.204s
  current training loss:		0.439341
  current training accuracy:		0.849224
Seen 249000 samples, time used: 127.601s
  current training loss:		0.439368
  current training accuracy:		0.849141
Seen 252000 samples, time used: 123.470s
  current training loss:		0.439003
  current training accuracy:		0.849198
Seen 255000 samples, time used: 124.468s
  current training loss:		0.438863
  current training accuracy:		0.849275
Seen 258000 samples, time used: 125.163s
  current training loss:		0.438843
  current training accuracy:		0.849291
Seen 261000 samples, time used: 124.793s
  current training loss:		0.438772
  current training accuracy:		0.849310
Seen 264000 samples, time used: 122.778s
  current training loss:		0.438672
  current training accuracy:		0.849318
Seen 267000 samples, time used: 123.462s
  current training loss:		0.438737
  current training accuracy:		0.849416
Seen 270000 samples, time used: 126.117s
  current training loss:		0.438991
  current training accuracy:		0.849337
saving to ..., time used: 1245.312s
Seen 273000 samples, time used: 129.766s
  current training loss:		0.438799
  current training accuracy:		0.849462
Seen 276000 samples, time used: 123.233s
  current training loss:		0.438737
  current training accuracy:		0.849554
Seen 279000 samples, time used: 126.489s
  current training loss:		0.438721
  current training accuracy:		0.849552
Seen 282000 samples, time used: 127.731s
  current training loss:		0.438677
  current training accuracy:		0.849564
Seen 285000 samples, time used: 126.260s
  current training loss:		0.438630
  current training accuracy:		0.849614
Seen 288000 samples, time used: 126.010s
  current training loss:		0.438561
  current training accuracy:		0.849646
Seen 291000 samples, time used: 127.604s
  current training loss:		0.438672
  current training accuracy:		0.849540
Seen 294000 samples, time used: 124.753s
  current training loss:		0.438533
  current training accuracy:		0.849514
Seen 297000 samples, time used: 124.345s
  current training loss:		0.438266
  current training accuracy:		0.849606
Seen 300000 samples, time used: 126.971s
  current training loss:		0.438156
  current training accuracy:		0.849613
saving to ..., time used: 1262.964s
Seen 303000 samples, time used: 126.099s
  current training loss:		0.437926
  current training accuracy:		0.849673
Seen 306000 samples, time used: 124.590s
  current training loss:		0.438159
  current training accuracy:		0.849601
Seen 309000 samples, time used: 121.588s
  current training loss:		0.438056
  current training accuracy:		0.849663
Seen 312000 samples, time used: 123.299s
  current training loss:		0.438306
  current training accuracy:		0.849554
Seen 315000 samples, time used: 127.387s
  current training loss:		0.438249
  current training accuracy:		0.849552
Seen 318000 samples, time used: 121.951s
  current training loss:		0.438061
  current training accuracy:		0.849572
Seen 321000 samples, time used: 122.452s
  current training loss:		0.437976
  current training accuracy:		0.849611
Seen 324000 samples, time used: 127.730s
  current training loss:		0.437991
  current training accuracy:		0.849633
Seen 327000 samples, time used: 124.893s
  current training loss:		0.438008
  current training accuracy:		0.849679
Seen 330000 samples, time used: 128.629s
  current training loss:		0.438097
  current training accuracy:		0.849627
saving to ..., time used: 1248.417s
Seen 333000 samples, time used: 125.166s
  current training loss:		0.438258
  current training accuracy:		0.849511
Seen 336000 samples, time used: 124.912s
  current training loss:		0.438221
  current training accuracy:		0.849551
Seen 339000 samples, time used: 127.145s
  current training loss:		0.438152
  current training accuracy:		0.849584
Seen 342000 samples, time used: 126.092s
  current training loss:		0.438484
  current training accuracy:		0.849482
Seen 345000 samples, time used: 123.814s
  current training loss:		0.438481
  current training accuracy:		0.849443
Seen 348000 samples, time used: 124.803s
  current training loss:		0.438374
  current training accuracy:		0.849443
Seen 351000 samples, time used: 127.892s
  current training loss:		0.438482
  current training accuracy:		0.849436
Seen 354000 samples, time used: 123.721s
  current training loss:		0.438437
  current training accuracy:		0.849494
Seen 357000 samples, time used: 127.040s
  current training loss:		0.438241
  current training accuracy:		0.849507
Seen 360000 samples, time used: 123.460s
  current training loss:		0.438191
  current training accuracy:		0.849572
saving to ..., time used: 1253.846s
Seen 363000 samples, time used: 124.185s
  current training loss:		0.438213
  current training accuracy:		0.849592
Seen 366000 samples, time used: 125.259s
  current training loss:		0.438238
  current training accuracy:		0.849615
Seen 369000 samples, time used: 121.839s
  current training loss:		0.438295
  current training accuracy:		0.849591
Seen 372000 samples, time used: 123.094s
  current training loss:		0.438450
  current training accuracy:		0.849546
Seen 375000 samples, time used: 125.929s
  current training loss:		0.438464
  current training accuracy:		0.849557
Seen 378000 samples, time used: 122.916s
  current training loss:		0.438698
  current training accuracy:		0.849421
Seen 381000 samples, time used: 122.483s
  current training loss:		0.438900
  current training accuracy:		0.849409

Seen 384000 samples, time used: 127.422s
  current training loss:		0.439102
  current training accuracy:		0.849284
Seen 387000 samples, time used: 122.458s
  current training loss:		0.439154
  current training accuracy:		0.849256
Seen 390000 samples, time used: 123.510s
  current training loss:		0.439089
  current training accuracy:		0.849300
saving to ..., time used: 1238.896s
Seen 393000 samples, time used: 123.617s
  current training loss:		0.438908
  current training accuracy:		0.849387
Seen 396000 samples, time used: 125.812s
  current training loss:		0.438985
  current training accuracy:		0.849361
Seen 399000 samples, time used: 124.960s
  current training loss:		0.438933
  current training accuracy:		0.849381
Seen 402000 samples, time used: 124.455s
  current training loss:		0.438746
  current training accuracy:		0.849438
Seen 405000 samples, time used: 122.975s
  current training loss:		0.438758
  current training accuracy:		0.849435
Seen 408000 samples, time used: 125.015s
  current training loss:		0.438734
  current training accuracy:		0.849456
Seen 411000 samples, time used: 125.677s
  current training loss:		0.438787
  current training accuracy:		0.849387
Seen 414000 samples, time used: 124.476s
  current training loss:		0.438709
  current training accuracy:		0.849406
Seen 417000 samples, time used: 122.971s
  current training loss:		0.438749
  current training accuracy:		0.849369
Seen 420000 samples, time used: 124.739s
  current training loss:		0.438854
  current training accuracy:		0.849319
saving to ..., time used: 1244.498s
Seen 423000 samples, time used: 124.162s
  current training loss:		0.439004
  current training accuracy:		0.849348
Seen 426000 samples, time used: 125.359s
  current training loss:		0.438947
  current training accuracy:		0.849387
Seen 429000 samples, time used: 125.864s
  current training loss:		0.439103
  current training accuracy:		0.849331
Seen 432000 samples, time used: 124.696s
  current training loss:		0.439122
  current training accuracy:		0.849352
Seen 435000 samples, time used: 124.423s
  current training loss:		0.439050
  current training accuracy:		0.849361
Seen 438000 samples, time used: 125.356s
  current training loss:		0.439045
  current training accuracy:		0.849358
Seen 441000 samples, time used: 123.933s
  current training loss:		0.438968
  current training accuracy:		0.849349
Seen 444000 samples, time used: 125.860s
  current training loss:		0.439141
  current training accuracy:		0.849284
Seen 447000 samples, time used: 126.780s
  current training loss:		0.439059
  current training accuracy:		0.849295
Seen 450000 samples, time used: 124.521s
  current training loss:		0.439175
  current training accuracy:		0.849231
saving to ..., time used: 1250.753s
Seen 453000 samples, time used: 125.468s
  current training loss:		0.439218
  current training accuracy:		0.849183
Seen 456000 samples, time used: 125.893s
  current training loss:		0.439264
  current training accuracy:		0.849208
Seen 459000 samples, time used: 123.091s
  current training loss:		0.439276
  current training accuracy:		0.849181
Seen 462000 samples, time used: 123.076s
  current training loss:		0.439190
  current training accuracy:		0.849201
Seen 465000 samples, time used: 126.026s
  current training loss:		0.439315
  current training accuracy:		0.849168
Seen 468000 samples, time used: 126.172s
  current training loss:		0.439298
  current training accuracy:		0.849109
Seen 471000 samples, time used: 129.003s
  current training loss:		0.439268
  current training accuracy:		0.849113
Seen 474000 samples, time used: 124.202s
  current training loss:		0.439283
  current training accuracy:		0.849055
Seen 477000 samples, time used: 122.698s
  current training loss:		0.439217
  current training accuracy:		0.849065
Seen 480000 samples, time used: 126.540s
  current training loss:		0.439341
  current training accuracy:		0.849042
saving to ..., time used: 1251.970s
Seen 483000 samples, time used: 124.469s
  current training loss:		0.439351
  current training accuracy:		0.848975
Seen 486000 samples, time used: 128.059s
  current training loss:		0.439412
  current training accuracy:		0.848973
Seen 489000 samples, time used: 124.760s
  current training loss:		0.439481
  current training accuracy:		0.848939
Seen 492000 samples, time used: 123.756s
  current training loss:		0.439435
  current training accuracy:		0.848972
Seen 495000 samples, time used: 126.607s
  current training loss:		0.439440
  current training accuracy:		0.848941
Seen 498000 samples, time used: 124.396s
  current training loss:		0.439400
  current training accuracy:		0.848988
Seen 501000 samples, time used: 125.397s
  current training loss:		0.439361
  current training accuracy:		0.848970
Seen 504000 samples, time used: 124.332s
  current training loss:		0.439485
  current training accuracy:		0.848937
Seen 507000 samples, time used: 124.212s
  current training loss:		0.439547
  current training accuracy:		0.848913
Seen 510000 samples, time used: 121.654s
  current training loss:		0.439572
  current training accuracy:		0.848937
saving to ..., time used: 1247.443s
Seen 513000 samples, time used: 122.348s
  current training loss:		0.439536
  current training accuracy:		0.848994
Seen 516000 samples, time used: 123.777s
  current training loss:		0.439490
  current training accuracy:		0.849031
Seen 519000 samples, time used: 126.524s
  current training loss:		0.439531
  current training accuracy:		0.849012
Seen 522000 samples, time used: 126.148s
  current training loss:		0.439509
  current training accuracy:		0.849036
Seen 525000 samples, time used: 125.144s
  current training loss:		0.439506
  current training accuracy:		0.849067
Seen 528000 samples, time used: 125.590s
  current training loss:		0.439426
  current training accuracy:		0.849062
Seen 531000 samples, time used: 124.489s
  current training loss:		0.439415
  current training accuracy:		0.849083
Seen 534000 samples, time used: 123.516s
  current training loss:		0.439375
  current training accuracy:		0.849114
Seen 537000 samples, time used: 125.378s
  current training loss:		0.439595
  current training accuracy:		0.849022
Seen 540000 samples, time used: 125.192s
  current training loss:		0.439577
  current training accuracy:		0.849002
saving to ..., time used: 1247.905s
Seen 543000 samples, time used: 125.711s
  current training loss:		0.439629
  current training accuracy:		0.848996
Seen 546000 samples, time used: 127.092s
  current training loss:		0.439681
  current training accuracy:		0.848971
Seen 549000 samples, time used: 124.908s
  current training loss:		0.439673
  current training accuracy:		0.848967
Epoch 4 of 20 took 22934.916s
  training loss:		0.439629
  training accuracy:		84.90 %
  validation loss:		0.422672
  validation accuracy:		84.03 %
Seen 3000 samples, time used: 127.102s
  current training loss:		0.411354
  current training accuracy:		0.856667
Seen 6000 samples, time used: 128.612s
  current training loss:		0.402033
  current training accuracy:		0.863833
Seen 9000 samples, time used: 125.352s
  current training loss:		0.397874
  current training accuracy:		0.865444
Seen 12000 samples, time used: 121.394s
  current training loss:		0.401446
  current training accuracy:		0.862167
Seen 15000 samples, time used: 122.685s
  current training loss:		0.402179
  current training accuracy:		0.863200
Seen 18000 samples, time used: 123.148s
  current training loss:		0.401028
  current training accuracy:		0.865444
Seen 21000 samples, time used: 121.874s
  current training loss:		0.401879
  current training accuracy:		0.865048
Seen 24000 samples, time used: 121.676s
  current training loss:		0.401750
  current training accuracy:		0.865417
Seen 27000 samples, time used: 125.923s
  current training loss:		0.399601
  current training accuracy:		0.866111
Seen 30000 samples, time used: 122.261s
  current training loss:		0.399822
  current training accuracy:		0.866067
saving to ..., time used: 1240.028s
Seen 33000 samples, time used: 123.971s
  current training loss:		0.400222
  current training accuracy:		0.865697
Seen 36000 samples, time used: 123.795s
  current training loss:		0.401131
  current training accuracy:		0.865389
Seen 39000 samples, time used: 124.846s
  current training loss:		0.402538
  current training accuracy:		0.864872
Seen 42000 samples, time used: 119.879s
  current training loss:		0.402730
  current training accuracy:		0.865048
Seen 45000 samples, time used: 123.253s
  current training loss:		0.404300
  current training accuracy:		0.864444
Seen 48000 samples, time used: 127.853s
  current training loss:		0.404461
  current training accuracy:		0.864271
Seen 51000 samples, time used: 125.392s
  current training loss:		0.402950
  current training accuracy:		0.864745
Seen 54000 samples, time used: 122.091s
  current training loss:		0.402865
  current training accuracy:		0.865167
Seen 57000 samples, time used: 123.154s
  current training loss:		0.402416
  current training accuracy:		0.865526
Seen 60000 samples, time used: 125.252s
  current training loss:		0.403408
  current training accuracy:		0.864967
saving to ..., time used: 1239.288s
Seen 63000 samples, time used: 126.529s
  current training loss:		0.403894
  current training accuracy:		0.864444
Seen 66000 samples, time used: 124.030s
  current training loss:		0.404330
  current training accuracy:		0.864273
Seen 69000 samples, time used: 122.323s
  current training loss:		0.404611
  current training accuracy:		0.863913
Seen 72000 samples, time used: 124.569s
  current training loss:		0.404845
  current training accuracy:		0.864125
Seen 75000 samples, time used: 125.931s
  current training loss:		0.405189
  current training accuracy:		0.863880
Seen 78000 samples, time used: 124.721s
  current training loss:		0.405048
  current training accuracy:		0.863808
Seen 81000 samples, time used: 127.170s
  current training loss:		0.404427
  current training accuracy:		0.864111
Seen 84000 samples, time used: 125.878s
  current training loss:		0.405140
  current training accuracy:		0.864036
Seen 87000 samples, time used: 124.949s
  current training loss:		0.404739
  current training accuracy:		0.864195
Seen 90000 samples, time used: 124.357s
  current training loss:		0.405576
  current training accuracy:		0.863856
saving to ..., time used: 1250.257s
Seen 93000 samples, time used: 124.541s
  current training loss:		0.404772
  current training accuracy:		0.864161
Seen 96000 samples, time used: 125.703s
  current training loss:		0.403922
  current training accuracy:		0.864479
Seen 99000 samples, time used: 126.999s
  current training loss:		0.404964
  current training accuracy:		0.864242
Seen 102000 samples, time used: 122.169s
  current training loss:		0.405560
  current training accuracy:		0.863922
Seen 105000 samples, time used: 124.358s
  current training loss:		0.405711
  current training accuracy:		0.863933
Seen 108000 samples, time used: 128.124s
  current training loss:		0.405689
  current training accuracy:		0.863898
Seen 111000 samples, time used: 124.435s
  current training loss:		0.405679
  current training accuracy:		0.863883
Seen 114000 samples, time used: 124.136s
  current training loss:		0.405930
  current training accuracy:		0.863719
Seen 117000 samples, time used: 124.867s
  current training loss:		0.406168
  current training accuracy:		0.863615
Seen 120000 samples, time used: 123.510s
  current training loss:		0.405952
  current training accuracy:		0.863808
saving to ..., time used: 1248.645s
Seen 123000 samples, time used: 125.832s
  current training loss:		0.406397
  current training accuracy:		0.863659
Seen 126000 samples, time used: 123.549s
  current training loss:		0.406439
  current training accuracy:		0.863667
Seen 129000 samples, time used: 125.562s
  current training loss:		0.406138
  current training accuracy:		0.863853
Seen 132000 samples, time used: 124.917s
  current training loss:		0.406000
  current training accuracy:		0.863894
Seen 135000 samples, time used: 125.379s
  current training loss:		0.406448
  current training accuracy:		0.863726
Seen 138000 samples, time used: 125.592s
  current training loss:		0.406976
  current training accuracy:		0.863667
Seen 141000 samples, time used: 127.104s
  current training loss:		0.406996
  current training accuracy:		0.863574
Seen 144000 samples, time used: 123.998s
  current training loss:		0.407150
  current training accuracy:		0.863486
Seen 147000 samples, time used: 125.583s
  current training loss:		0.407445
  current training accuracy:		0.863449
Seen 150000 samples, time used: 126.788s
  current training loss:		0.407420
  current training accuracy:		0.863487
saving to ..., time used: 1254.104s
Seen 153000 samples, time used: 126.832s
  current training loss:		0.407829
  current training accuracy:		0.863392
Seen 156000 samples, time used: 127.130s
  current training loss:		0.407977
  current training accuracy:		0.863333
Seen 159000 samples, time used: 127.053s
  current training loss:		0.408128
  current training accuracy:		0.863283
Seen 162000 samples, time used: 124.982s
  current training loss:		0.408097
  current training accuracy:		0.863352
Seen 165000 samples, time used: 123.790s
  current training loss:		0.407890
  current training accuracy:		0.863485
Seen 168000 samples, time used: 122.087s
  current training loss:		0.408008
  current training accuracy:		0.863393
Seen 171000 samples, time used: 123.093s
  current training loss:		0.408423
  current training accuracy:		0.863193
Seen 174000 samples, time used: 126.791s
  current training loss:		0.408005
  current training accuracy:		0.863356
Seen 177000 samples, time used: 127.127s
  current training loss:		0.408584
  current training accuracy:		0.863282
Seen 180000 samples, time used: 128.050s
  current training loss:		0.408761
  current training accuracy:		0.863117
saving to ..., time used: 1256.733s
Seen 183000 samples, time used: 123.472s
  current training loss:		0.409171
  current training accuracy:		0.863066
Seen 186000 samples, time used: 124.893s
  current training loss:		0.409354
  current training accuracy:		0.862978
Seen 189000 samples, time used: 124.112s
  current training loss:		0.409459
  current training accuracy:		0.863032
Seen 192000 samples, time used: 127.553s
  current training loss:		0.409642
  current training accuracy:		0.862953
Seen 195000 samples, time used: 121.908s
  current training loss:		0.409860
  current training accuracy:		0.862903
Seen 198000 samples, time used: 126.614s
  current training loss:		0.409895
  current training accuracy:		0.862955
Seen 201000 samples, time used: 122.072s
  current training loss:		0.410212
  current training accuracy:		0.862786
Seen 204000 samples, time used: 124.585s
  current training loss:		0.410077
  current training accuracy:		0.862833
Seen 207000 samples, time used: 125.844s
  current training loss:		0.410544
  current training accuracy:		0.862686
Seen 210000 samples, time used: 125.393s
  current training loss:		0.410825
  current training accuracy:		0.862529
saving to ..., time used: 1246.247s
Seen 213000 samples, time used: 125.727s
  current training loss:		0.410852
  current training accuracy:		0.862493
Seen 216000 samples, time used: 124.037s
  current training loss:		0.410978
  current training accuracy:		0.862449
Seen 219000 samples, time used: 124.321s
  current training loss:		0.410648
  current training accuracy:		0.862616
Seen 222000 samples, time used: 129.541s
  current training loss:		0.410733
  current training accuracy:		0.862532
Seen 225000 samples, time used: 123.644s
  current training loss:		0.410714
  current training accuracy:		0.862542
Seen 228000 samples, time used: 124.389s
  current training loss:		0.410719
  current training accuracy:		0.862623
Seen 231000 samples, time used: 127.493s
  current training loss:		0.410799
  current training accuracy:		0.862541
Seen 234000 samples, time used: 124.785s
  current training loss:		0.410545
  current training accuracy:		0.862479
Seen 237000 samples, time used: 123.944s
  current training loss:		0.410288
  current training accuracy:		0.862494
Seen 240000 samples, time used: 125.100s
  current training loss:		0.410388
  current training accuracy:		0.862450
saving to ..., time used: 1252.783s
Seen 243000 samples, time used: 127.605s
  current training loss:		0.410520
  current training accuracy:		0.862465
Seen 246000 samples, time used: 127.366s
  current training loss:		0.410624
  current training accuracy:		0.862467
Seen 249000 samples, time used: 127.858s
  current training loss:		0.410643
  current training accuracy:		0.862430
Seen 252000 samples, time used: 126.143s
  current training loss:		0.410472
  current training accuracy:		0.862448
Seen 255000 samples, time used: 123.695s
  current training loss:		0.410224
  current training accuracy:		0.862580
Seen 258000 samples, time used: 127.928s
  current training loss:		0.410326
  current training accuracy:		0.862500
Seen 261000 samples, time used: 125.289s
  current training loss:		0.410585
  current training accuracy:		0.862421
Seen 264000 samples, time used: 124.947s
  current training loss:		0.410541
  current training accuracy:		0.862413
Seen 267000 samples, time used: 124.080s
  current training loss:		0.410533
  current training accuracy:		0.862352
Seen 270000 samples, time used: 126.395s
  current training loss:		0.410639
  current training accuracy:		0.862256
saving to ..., time used: 1261.108s
Seen 273000 samples, time used: 123.155s
  current training loss:		0.410828
  current training accuracy:		0.862183
Seen 276000 samples, time used: 126.125s
  current training loss:		0.411074
  current training accuracy:		0.862123
Seen 279000 samples, time used: 124.634s
  current training loss:		0.410846
  current training accuracy:		0.862254
Seen 282000 samples, time used: 123.043s
  current training loss:		0.410758
  current training accuracy:		0.862266
Seen 285000 samples, time used: 125.686s
  current training loss:		0.410702
  current training accuracy:		0.862253
Seen 288000 samples, time used: 128.628s
  current training loss:		0.410763
  current training accuracy:		0.862194
Seen 291000 samples, time used: 126.182s
  current training loss:		0.410874
  current training accuracy:		0.862127
Seen 294000 samples, time used: 121.868s
  current training loss:		0.410959
  current training accuracy:		0.862065
Seen 297000 samples, time used: 125.991s
  current training loss:		0.411178
  current training accuracy:		0.861973
Seen 300000 samples, time used: 125.394s
  current training loss:		0.411251
  current training accuracy:		0.862000
saving to ..., time used: 1250.506s
Seen 303000 samples, time used: 126.436s
  current training loss:		0.411267
  current training accuracy:		0.861904
Seen 306000 samples, time used: 126.220s
  current training loss:		0.411361
  current training accuracy:		0.861863
Seen 309000 samples, time used: 126.448s
  current training loss:		0.411483
  current training accuracy:		0.861803
Seen 312000 samples, time used: 125.651s
  current training loss:		0.411578
  current training accuracy:		0.861776
Seen 315000 samples, time used: 124.806s
  current training loss:		0.411745
  current training accuracy:		0.861714
Seen 318000 samples, time used: 125.672s
  current training loss:		0.411826
  current training accuracy:		0.861604
Seen 321000 samples, time used: 126.230s
  current training loss:		0.411847
  current training accuracy:		0.861626
Seen 324000 samples, time used: 124.807s
  current training loss:		0.412079
  current training accuracy:		0.861466
Seen 327000 samples, time used: 125.690s
  current training loss:		0.412272
  current training accuracy:		0.861373
Seen 330000 samples, time used: 123.594s
  current training loss:		0.412294
  current training accuracy:		0.861391
saving to ..., time used: 1255.355s
Seen 333000 samples, time used: 123.514s
  current training loss:		0.412317
  current training accuracy:		0.861453
Seen 336000 samples, time used: 127.918s
  current training loss:		0.412537
  current training accuracy:		0.861345
Seen 339000 samples, time used: 125.750s
  current training loss:		0.412665
  current training accuracy:		0.861265
Seen 342000 samples, time used: 123.829s
  current training loss:		0.412624
  current training accuracy:		0.861263
Seen 345000 samples, time used: 125.205s
  current training loss:		0.412725
  current training accuracy:		0.861226
Seen 348000 samples, time used: 123.605s
  current training loss:		0.412593
  current training accuracy:		0.861264
Seen 351000 samples, time used: 127.096s
  current training loss:		0.412631
  current training accuracy:		0.861288
Seen 354000 samples, time used: 127.454s
  current training loss:		0.412457
  current training accuracy:		0.861384
Seen 357000 samples, time used: 128.652s
  current training loss:		0.412599
  current training accuracy:		0.861319
Seen 360000 samples, time used: 128.422s
  current training loss:		0.412616
  current training accuracy:		0.861308
saving to ..., time used: 1261.246s
Seen 363000 samples, time used: 127.466s
  current training loss:		0.412701
  current training accuracy:		0.861309

Seen 366000 samples, time used: 124.684s
  current training loss:		0.413025
  current training accuracy:		0.861205
Seen 369000 samples, time used: 121.479s
  current training loss:		0.413027
  current training accuracy:		0.861225
Seen 372000 samples, time used: 123.478s
  current training loss:		0.412993
  current training accuracy:		0.861199
Seen 375000 samples, time used: 125.914s
  current training loss:		0.413083
  current training accuracy:		0.861099
Seen 378000 samples, time used: 125.760s
  current training loss:		0.413112
  current training accuracy:		0.861116
Seen 381000 samples, time used: 128.118s
  current training loss:		0.413068
  current training accuracy:		0.861186
Seen 384000 samples, time used: 126.114s
  current training loss:		0.412967
  current training accuracy:		0.861201
Seen 387000 samples, time used: 126.087s
  current training loss:		0.413061
  current training accuracy:		0.861181
Seen 390000 samples, time used: 128.907s
  current training loss:		0.413139
  current training accuracy:		0.861185
saving to ..., time used: 1257.810s
Seen 393000 samples, time used: 122.057s
  current training loss:		0.413271
  current training accuracy:		0.861168
Seen 396000 samples, time used: 122.357s
  current training loss:		0.413450
  current training accuracy:		0.861028
Seen 399000 samples, time used: 125.993s
  current training loss:		0.413429
  current training accuracy:		0.861073
Seen 402000 samples, time used: 124.666s
  current training loss:		0.413468
  current training accuracy:		0.861085
Seen 405000 samples, time used: 129.685s
  current training loss:		0.413425
  current training accuracy:		0.861035
Seen 408000 samples, time used: 127.765s
  current training loss:		0.413399
  current training accuracy:		0.861051
Seen 411000 samples, time used: 122.967s
  current training loss:		0.413443
  current training accuracy:		0.861036
Seen 414000 samples, time used: 123.723s
  current training loss:		0.413484
  current training accuracy:		0.861036
Seen 417000 samples, time used: 124.278s
  current training loss:		0.413366
  current training accuracy:		0.861137
Seen 420000 samples, time used: 127.106s
  current training loss:		0.413270
  current training accuracy:		0.861145
saving to ..., time used: 1250.399s
Seen 423000 samples, time used: 125.124s
  current training loss:		0.413129
  current training accuracy:		0.861139
Seen 426000 samples, time used: 124.235s
  current training loss:		0.413123
  current training accuracy:		0.861153
Seen 429000 samples, time used: 124.347s
  current training loss:		0.413284
  current training accuracy:		0.861112
Seen 432000 samples, time used: 125.361s
  current training loss:		0.413250
  current training accuracy:		0.861120
Seen 435000 samples, time used: 128.908s
  current training loss:		0.413288
  current training accuracy:		0.861124
Seen 438000 samples, time used: 124.665s
  current training loss:		0.413445
  current training accuracy:		0.861110
Seen 441000 samples, time used: 121.635s
  current training loss:		0.413397
  current training accuracy:		0.861109
Seen 444000 samples, time used: 124.162s
  current training loss:		0.413417
  current training accuracy:		0.861119
Seen 447000 samples, time used: 122.629s
  current training loss:		0.413430
  current training accuracy:		0.861103
Seen 450000 samples, time used: 124.463s
  current training loss:		0.413587
  current training accuracy:		0.861056
saving to ..., time used: 1245.328s
Seen 453000 samples, time used: 126.696s
  current training loss:		0.413672
  current training accuracy:		0.860982
Seen 456000 samples, time used: 124.495s
  current training loss:		0.413766
  current training accuracy:		0.860914
Seen 459000 samples, time used: 123.872s
  current training loss:		0.413859
  current training accuracy:		0.860869
Seen 462000 samples, time used: 124.232s
  current training loss:		0.413842
  current training accuracy:		0.860855
Seen 465000 samples, time used: 125.370s
  current training loss:		0.413933
  current training accuracy:		0.860733
Seen 468000 samples, time used: 124.908s
  current training loss:		0.414090
  current training accuracy:		0.860677
Seen 471000 samples, time used: 127.063s
  current training loss:		0.413992
  current training accuracy:		0.860677
Seen 474000 samples, time used: 121.898s
  current training loss:		0.414014
  current training accuracy:		0.860637
Seen 477000 samples, time used: 122.187s
  current training loss:		0.413990
  current training accuracy:		0.860644
Seen 480000 samples, time used: 126.655s
  current training loss:		0.413948
  current training accuracy:		0.860635
saving to ..., time used: 1247.174s
Seen 483000 samples, time used: 127.721s
  current training loss:		0.414005
  current training accuracy:		0.860598
Seen 486000 samples, time used: 126.696s
  current training loss:		0.413985
  current training accuracy:		0.860568
Seen 489000 samples, time used: 125.332s
  current training loss:		0.413913
  current training accuracy:		0.860575
Seen 492000 samples, time used: 125.237s
  current training loss:		0.413973
  current training accuracy:		0.860518
Seen 495000 samples, time used: 123.053s
  current training loss:		0.414028
  current training accuracy:		0.860515
Seen 498000 samples, time used: 125.695s
  current training loss:		0.414036
  current training accuracy:		0.860506
Seen 501000 samples, time used: 124.234s
  current training loss:		0.414096
  current training accuracy:		0.860495
Seen 504000 samples, time used: 128.417s
  current training loss:		0.414125
  current training accuracy:		0.860492
Seen 507000 samples, time used: 122.379s
  current training loss:		0.414165
  current training accuracy:		0.860471
Seen 510000 samples, time used: 126.387s
  current training loss:		0.414206
  current training accuracy:		0.860445
saving to ..., time used: 1254.950s
Seen 513000 samples, time used: 125.892s
  current training loss:		0.414432
  current training accuracy:		0.860343
Seen 516000 samples, time used: 123.350s
  current training loss:		0.414489
  current training accuracy:		0.860304
Seen 519000 samples, time used: 126.759s
  current training loss:		0.414518
  current training accuracy:		0.860256
Seen 522000 samples, time used: 121.556s
  current training loss:		0.414606
  current training accuracy:		0.860261
Seen 525000 samples, time used: 124.787s
  current training loss:		0.414554
  current training accuracy:		0.860255
Seen 528000 samples, time used: 125.953s
  current training loss:		0.414424
  current training accuracy:		0.860278
Seen 531000 samples, time used: 126.366s
  current training loss:		0.414428
  current training accuracy:		0.860307
Seen 534000 samples, time used: 126.473s
  current training loss:		0.414418
  current training accuracy:		0.860283
Seen 537000 samples, time used: 127.100s
  current training loss:		0.414510
  current training accuracy:		0.860257
Seen 540000 samples, time used: 123.471s
  current training loss:		0.414469
  current training accuracy:		0.860274
saving to ..., time used: 1251.507s
Seen 543000 samples, time used: 126.691s
  current training loss:		0.414546
  current training accuracy:		0.860232
Seen 546000 samples, time used: 127.836s
  current training loss:		0.414659
  current training accuracy:		0.860212
Seen 549000 samples, time used: 124.271s
  current training loss:		0.414561
  current training accuracy:		0.860250
Epoch 5 of 20 took 22949.636s
  training loss:		0.414590
  training accuracy:		86.02 %
  validation loss:		0.416372
  validation accuracy:		84.44 %
Seen 3000 samples, time used: 125.705s
  current training loss:		0.391363
  current training accuracy:		0.876333
Seen 6000 samples, time used: 123.252s
  current training loss:		0.394175
  current training accuracy:		0.871333
Seen 9000 samples, time used: 123.580s
  current training loss:		0.389003
  current training accuracy:		0.874667
Seen 12000 samples, time used: 124.377s
  current training loss:		0.384139
  current training accuracy:		0.877250
Seen 15000 samples, time used: 126.787s
  current training loss:		0.386894
  current training accuracy:		0.875267
Seen 18000 samples, time used: 127.659s
  current training loss:		0.385073
  current training accuracy:		0.873944
Seen 21000 samples, time used: 130.131s
  current training loss:		0.386411
  current training accuracy:		0.873190
Seen 24000 samples, time used: 124.701s
  current training loss:		0.384739
  current training accuracy:		0.873625
Seen 27000 samples, time used: 127.092s
  current training loss:		0.384398
  current training accuracy:		0.874037
Seen 30000 samples, time used: 127.768s
  current training loss:		0.383779
  current training accuracy:		0.874900
saving to ..., time used: 1261.052s
Seen 33000 samples, time used: 124.491s
  current training loss:		0.382277
  current training accuracy:		0.875636
Seen 36000 samples, time used: 125.454s
  current training loss:		0.382805
  current training accuracy:		0.875250
Seen 39000 samples, time used: 122.685s
  current training loss:		0.383078
  current training accuracy:		0.875077
Seen 42000 samples, time used: 126.030s
  current training loss:		0.383021
  current training accuracy:		0.875429
Seen 45000 samples, time used: 125.108s
  current training loss:		0.384735
  current training accuracy:		0.875156
Seen 48000 samples, time used: 123.669s
  current training loss:		0.385689
  current training accuracy:		0.874604
Seen 51000 samples, time used: 127.372s
  current training loss:		0.385644
  current training accuracy:		0.874706
Seen 54000 samples, time used: 123.302s
  current training loss:		0.385383
  current training accuracy:		0.874630
Seen 57000 samples, time used: 123.239s
  current training loss:		0.383336
  current training accuracy:		0.875526
Seen 60000 samples, time used: 124.298s
  current training loss:		0.383964
  current training accuracy:		0.875567
saving to ..., time used: 1245.449s
Seen 63000 samples, time used: 124.069s
  current training loss:		0.383189
  current training accuracy:		0.875905
Seen 66000 samples, time used: 125.683s
  current training loss:		0.383399
  current training accuracy:		0.875621
Seen 69000 samples, time used: 124.483s
  current training loss:		0.384039
  current training accuracy:		0.875449
Seen 72000 samples, time used: 124.173s
  current training loss:		0.384453
  current training accuracy:		0.875083
Seen 75000 samples, time used: 122.883s
  current training loss:		0.384053
  current training accuracy:		0.875333
Seen 78000 samples, time used: 124.521s
  current training loss:		0.384240
  current training accuracy:		0.875474
Seen 81000 samples, time used: 120.576s
  current training loss:		0.384452
  current training accuracy:		0.875383
Seen 84000 samples, time used: 123.972s
  current training loss:		0.384629
  current training accuracy:		0.875262
Seen 87000 samples, time used: 128.473s
  current training loss:		0.384467
  current training accuracy:		0.875299
Seen 90000 samples, time used: 122.271s
  current training loss:		0.384430
  current training accuracy:		0.875111
saving to ..., time used: 1240.906s
Seen 93000 samples, time used: 120.521s
  current training loss:		0.384780
  current training accuracy:		0.874946
Seen 96000 samples, time used: 123.999s
  current training loss:		0.384680
  current training accuracy:		0.875062
Seen 99000 samples, time used: 125.345s
  current training loss:		0.384698
  current training accuracy:		0.875010
Seen 102000 samples, time used: 125.237s
  current training loss:		0.385126
  current training accuracy:		0.874853
Seen 105000 samples, time used: 125.289s
  current training loss:		0.385333
  current training accuracy:		0.875076
Seen 108000 samples, time used: 124.686s
  current training loss:		0.385699
  current training accuracy:		0.874972
Seen 111000 samples, time used: 126.247s
  current training loss:		0.386001
  current training accuracy:		0.874991
Seen 114000 samples, time used: 124.753s
  current training loss:		0.385462
  current training accuracy:		0.875149
Seen 117000 samples, time used: 124.833s
  current training loss:		0.386030
  current training accuracy:		0.875017
Seen 120000 samples, time used: 124.735s
  current training loss:		0.386492
  current training accuracy:		0.874725
saving to ..., time used: 1245.443s
Seen 123000 samples, time used: 127.187s
  current training loss:		0.386643
  current training accuracy:		0.874707
Seen 126000 samples, time used: 122.295s
  current training loss:		0.386940
  current training accuracy:		0.874508
Seen 129000 samples, time used: 122.819s
  current training loss:		0.386782
  current training accuracy:		0.874411
Seen 132000 samples, time used: 124.709s
  current training loss:		0.386519
  current training accuracy:		0.874477
Seen 135000 samples, time used: 124.901s
  current training loss:		0.386759
  current training accuracy:		0.874341
Seen 138000 samples, time used: 127.026s
  current training loss:		0.386926
  current training accuracy:		0.874348
Seen 141000 samples, time used: 124.400s
  current training loss:		0.387464
  current training accuracy:		0.874177
Seen 144000 samples, time used: 123.692s
  current training loss:		0.387421
  current training accuracy:		0.874146
Seen 147000 samples, time used: 124.318s
  current training loss:		0.387403
  current training accuracy:		0.874034
Seen 150000 samples, time used: 125.711s
  current training loss:		0.387317
  current training accuracy:		0.873993
saving to ..., time used: 1246.860s
Seen 153000 samples, time used: 121.283s
  current training loss:		0.387495
  current training accuracy:		0.873915
Seen 156000 samples, time used: 124.054s
  current training loss:		0.387346
  current training accuracy:		0.874038
Seen 159000 samples, time used: 128.853s
  current training loss:		0.387261
  current training accuracy:		0.874013
Seen 162000 samples, time used: 123.193s
  current training loss:		0.387269
  current training accuracy:		0.873975
Seen 165000 samples, time used: 128.040s
  current training loss:		0.387186
  current training accuracy:		0.874006
Seen 168000 samples, time used: 123.124s
  current training loss:		0.387262
  current training accuracy:		0.873929
Seen 171000 samples, time used: 124.356s
  current training loss:		0.387258
  current training accuracy:		0.873906
Seen 174000 samples, time used: 123.708s
  current training loss:		0.387171
  current training accuracy:		0.873925
Seen 177000 samples, time used: 126.843s
  current training loss:		0.387059
  current training accuracy:		0.873983
Seen 180000 samples, time used: 124.202s
  current training loss:		0.387319
  current training accuracy:		0.873889
saving to ..., time used: 1247.456s
Seen 183000 samples, time used: 125.417s
  current training loss:		0.387621
  current training accuracy:		0.873781
Seen 186000 samples, time used: 126.781s
  current training loss:		0.387882
  current training accuracy:		0.873602
Seen 189000 samples, time used: 126.523s
  current training loss:		0.388072
  current training accuracy:		0.873556
Seen 192000 samples, time used: 129.069s
  current training loss:		0.388266
  current training accuracy:		0.873526
Seen 195000 samples, time used: 142.821s
  current training loss:		0.388113
  current training accuracy:		0.873564
Seen 198000 samples, time used: 135.930s
  current training loss:		0.388199
  current training accuracy:		0.873611
Seen 201000 samples, time used: 132.369s
  current training loss:		0.388419
  current training accuracy:		0.873537
Seen 204000 samples, time used: 129.452s
  current training loss:		0.388793
  current training accuracy:		0.873324
Seen 207000 samples, time used: 128.153s
  current training loss:		0.388801
  current training accuracy:		0.873401
Seen 210000 samples, time used: 130.479s
  current training loss:		0.388848
  current training accuracy:		0.873433
saving to ..., time used: 1306.795s
Seen 213000 samples, time used: 124.067s
  current training loss:		0.388724
  current training accuracy:		0.873460
Seen 216000 samples, time used: 125.774s
  current training loss:		0.388961
  current training accuracy:		0.873407
Seen 219000 samples, time used: 126.828s
  current training loss:		0.389028
  current training accuracy:		0.873361
Seen 222000 samples, time used: 125.149s
  current training loss:		0.389088
  current training accuracy:		0.873387
Seen 225000 samples, time used: 127.242s
  current training loss:		0.389245
  current training accuracy:		0.873356
Seen 228000 samples, time used: 126.448s
  current training loss:		0.389376
  current training accuracy:		0.873272
Seen 231000 samples, time used: 129.199s
  current training loss:		0.389388
  current training accuracy:		0.873182
Seen 234000 samples, time used: 123.887s
  current training loss:		0.389577
  current training accuracy:		0.873124
Seen 237000 samples, time used: 127.509s
  current training loss:		0.389461
  current training accuracy:		0.873350
Seen 240000 samples, time used: 125.632s
  current training loss:		0.389408
  current training accuracy:		0.873450
saving to ..., time used: 1261.330s
Seen 243000 samples, time used: 122.293s
  current training loss:		0.389430
  current training accuracy:		0.873403
Seen 246000 samples, time used: 123.607s
  current training loss:		0.389339
  current training accuracy:		0.873415
Seen 249000 samples, time used: 122.231s
  current training loss:		0.389340
  current training accuracy:		0.873378
Seen 252000 samples, time used: 126.535s
  current training loss:		0.389458
  current training accuracy:		0.873214
Seen 255000 samples, time used: 127.858s
  current training loss:		0.389462
  current training accuracy:		0.873196
Seen 258000 samples, time used: 124.898s
  current training loss:		0.389587
  current training accuracy:		0.873109
Seen 261000 samples, time used: 126.784s
  current training loss:		0.389530
  current training accuracy:		0.873149
Seen 264000 samples, time used: 124.567s
  current training loss:		0.389403
  current training accuracy:		0.873239
Seen 267000 samples, time used: 124.720s
  current training loss:		0.389563
  current training accuracy:		0.873112
Seen 270000 samples, time used: 127.728s
  current training loss:		0.389528
  current training accuracy:		0.873111
saving to ..., time used: 1251.015s
Seen 273000 samples, time used: 122.801s
  current training loss:		0.389755
  current training accuracy:		0.873095
Seen 276000 samples, time used: 122.156s
  current training loss:		0.389787
  current training accuracy:		0.873080
Seen 279000 samples, time used: 125.592s
  current training loss:		0.389925
  current training accuracy:		0.873018
Seen 282000 samples, time used: 123.292s
  current training loss:		0.390239
  current training accuracy:		0.872904
Seen 285000 samples, time used: 126.353s
  current training loss:		0.390288
  current training accuracy:		0.872818
Seen 288000 samples, time used: 127.121s
  current training loss:		0.390562
  current training accuracy:		0.872708
Seen 291000 samples, time used: 126.098s
  current training loss:		0.390381
  current training accuracy:		0.872777
Seen 294000 samples, time used: 124.527s
  current training loss:		0.390314
  current training accuracy:		0.872793
Seen 297000 samples, time used: 123.369s
  current training loss:		0.390423
  current training accuracy:		0.872771
Seen 300000 samples, time used: 127.978s
  current training loss:		0.390678
  current training accuracy:		0.872733
saving to ..., time used: 1248.866s
Seen 303000 samples, time used: 126.116s
  current training loss:		0.390776
  current training accuracy:		0.872677
Seen 306000 samples, time used: 128.700s
  current training loss:		0.390957
  current training accuracy:		0.872601
Seen 309000 samples, time used: 124.039s
  current training loss:		0.390892
  current training accuracy:		0.872657
Seen 312000 samples, time used: 125.422s
  current training loss:		0.390819
  current training accuracy:		0.872737
Seen 315000 samples, time used: 123.061s
  current training loss:		0.390910
  current training accuracy:		0.872702
Seen 318000 samples, time used: 125.684s
  current training loss:		0.390848
  current training accuracy:		0.872695
Seen 321000 samples, time used: 126.759s
  current training loss:		0.391051
  current training accuracy:		0.872651
Seen 324000 samples, time used: 125.660s
  current training loss:		0.391209
  current training accuracy:		0.872534
Seen 327000 samples, time used: 123.115s
  current training loss:		0.391229
  current training accuracy:		0.872419
Seen 330000 samples, time used: 124.597s
  current training loss:		0.391274
  current training accuracy:		0.872397
saving to ..., time used: 1252.949s
Seen 333000 samples, time used: 128.535s
  current training loss:		0.391544
  current training accuracy:		0.872300
Seen 336000 samples, time used: 124.786s
  current training loss:		0.391677
  current training accuracy:		0.872220
Seen 339000 samples, time used: 124.545s
  current training loss:		0.391544
  current training accuracy:		0.872265
Seen 342000 samples, time used: 125.129s
  current training loss:		0.391700
  current training accuracy:		0.872155
Seen 345000 samples, time used: 123.695s
  current training loss:		0.391964
  current training accuracy:		0.872041
Seen 348000 samples, time used: 125.325s
  current training loss:		0.391928
  current training accuracy:		0.872020
Seen 351000 samples, time used: 124.908s
  current training loss:		0.392109
  current training accuracy:		0.871954
Seen 354000 samples, time used: 121.589s
  current training loss:		0.392210
  current training accuracy:		0.871915
Seen 357000 samples, time used: 125.242s
  current training loss:		0.392209
  current training accuracy:		0.871930
Seen 360000 samples, time used: 126.443s
  current training loss:		0.392326
  current training accuracy:		0.871886
saving to ..., time used: 1249.780s
Seen 363000 samples, time used: 125.866s
  current training loss:		0.392489
  current training accuracy:		0.871810
Seen 366000 samples, time used: 126.223s
  current training loss:		0.392434
  current training accuracy:		0.871760
Seen 369000 samples, time used: 124.354s
  current training loss:		0.392497
  current training accuracy:		0.871740
Seen 372000 samples, time used: 123.505s
  current training loss:		0.392640
  current training accuracy:		0.871745
Seen 375000 samples, time used: 128.394s
  current training loss:		0.392776
  current training accuracy:		0.871691
Seen 378000 samples, time used: 121.723s
  current training loss:		0.392657
  current training accuracy:		0.871772
Seen 381000 samples, time used: 127.390s
  current training loss:		0.392737
  current training accuracy:		0.871735
Seen 384000 samples, time used: 127.262s
  current training loss:		0.392902
  current training accuracy:		0.871695
Seen 387000 samples, time used: 123.459s
  current training loss:		0.392931
  current training accuracy:		0.871687
Seen 390000 samples, time used: 127.094s
  current training loss:		0.392973
  current training accuracy:		0.871692
saving to ..., time used: 1255.067s
Seen 393000 samples, time used: 124.973s
  current training loss:		0.393018
  current training accuracy:		0.871662
Seen 396000 samples, time used: 125.355s
  current training loss:		0.393065
  current training accuracy:		0.871601
Seen 399000 samples, time used: 123.267s
  current training loss:		0.393079
  current training accuracy:		0.871599
Seen 402000 samples, time used: 126.241s
  current training loss:		0.393051
  current training accuracy:		0.871619
Seen 405000 samples, time used: 124.447s
  current training loss:		0.393114
  current training accuracy:		0.871595
Seen 408000 samples, time used: 124.409s
  current training loss:		0.393059
  current training accuracy:		0.871544
Seen 411000 samples, time used: 126.400s
  current training loss:		0.392997
  current training accuracy:		0.871545
Seen 414000 samples, time used: 124.137s
  current training loss:		0.393005
  current training accuracy:		0.871539
Seen 417000 samples, time used: 123.464s
  current training loss:		0.393039
  current training accuracy:		0.871568
Seen 420000 samples, time used: 127.412s
  current training loss:		0.392982
  current training accuracy:		0.871545
saving to ..., time used: 1249.903s
Seen 423000 samples, time used: 127.929s
  current training loss:		0.392942
  current training accuracy:		0.871527
Seen 426000 samples, time used: 124.890s
  current training loss:		0.393172
  current training accuracy:		0.871455
Seen 429000 samples, time used: 127.666s
  current training loss:		0.393342
  current training accuracy:		0.871354
Seen 432000 samples, time used: 125.035s
  current training loss:		0.393351
  current training accuracy:		0.871326
Seen 435000 samples, time used: 128.599s
  current training loss:		0.393370
  current training accuracy:		0.871361
Seen 438000 samples, time used: 126.225s
  current training loss:		0.393338
  current training accuracy:		0.871374
Seen 441000 samples, time used: 123.827s
  current training loss:		0.393324
  current training accuracy:		0.871426
Seen 444000 samples, time used: 126.966s
  current training loss:		0.393196
  current training accuracy:		0.871471
Seen 447000 samples, time used: 128.426s
  current training loss:		0.393238
  current training accuracy:		0.871416
Seen 450000 samples, time used: 125.123s
  current training loss:		0.393259
  current training accuracy:		0.871427
saving to ..., time used: 1264.482s
Seen 453000 samples, time used: 126.940s
  current training loss:		0.393186
  current training accuracy:		0.871448
Seen 456000 samples, time used: 125.566s
  current training loss:		0.393229
  current training accuracy:		0.871467
Seen 459000 samples, time used: 124.634s
  current training loss:		0.393256
  current training accuracy:		0.871451
Seen 462000 samples, time used: 124.207s
  current training loss:		0.393353
  current training accuracy:		0.871416
Seen 465000 samples, time used: 125.580s
  current training loss:		0.393452
  current training accuracy:		0.871361
Seen 468000 samples, time used: 124.419s
  current training loss:		0.393412
  current training accuracy:		0.871410
Seen 471000 samples, time used: 123.091s
  current training loss:		0.393435
  current training accuracy:		0.871439
Seen 474000 samples, time used: 122.770s
  current training loss:		0.393393
  current training accuracy:		0.871439
Seen 477000 samples, time used: 128.270s
  current training loss:		0.393452
  current training accuracy:		0.871405
Seen 480000 samples, time used: 126.167s
  current training loss:		0.393335
  current training accuracy:		0.871427
saving to ..., time used: 1251.418s
Seen 483000 samples, time used: 125.313s
  current training loss:		0.393376
  current training accuracy:		0.871462
Seen 486000 samples, time used: 126.486s
  current training loss:		0.393470
  current training accuracy:		0.871449
Seen 489000 samples, time used: 121.766s
  current training loss:		0.393361
  current training accuracy:		0.871440
Seen 492000 samples, time used: 121.928s
  current training loss:		0.393379
  current training accuracy:		0.871457
Seen 495000 samples, time used: 126.073s
  current training loss:		0.393360
  current training accuracy:		0.871477
Seen 498000 samples, time used: 125.930s
  current training loss:		0.393455
  current training accuracy:		0.871420
Seen 501000 samples, time used: 126.555s
  current training loss:		0.393461
  current training accuracy:		0.871401
Seen 504000 samples, time used: 127.198s
  current training loss:		0.393334
  current training accuracy:		0.871407
Seen 507000 samples, time used: 126.194s
  current training loss:		0.393379
  current training accuracy:		0.871389
Seen 510000 samples, time used: 124.123s
  current training loss:		0.393416
  current training accuracy:		0.871375
saving to ..., time used: 1251.361s
Seen 513000 samples, time used: 124.664s
  current training loss:		0.393367
  current training accuracy:		0.871396
Seen 516000 samples, time used: 124.549s
  current training loss:		0.393447
  current training accuracy:		0.871378
Seen 519000 samples, time used: 125.481s
  current training loss:		0.393568
  current training accuracy:		0.871299
Seen 522000 samples, time used: 128.190s
  current training loss:		0.393728
  current training accuracy:		0.871203
Seen 525000 samples, time used: 125.710s
  current training loss:		0.393776
  current training accuracy:		0.871175
Seen 528000 samples, time used: 122.664s
  current training loss:		0.393779
  current training accuracy:		0.871231
Seen 531000 samples, time used: 123.694s
  current training loss:		0.393813
  current training accuracy:		0.871224
Seen 534000 samples, time used: 124.910s
  current training loss:		0.393711
  current training accuracy:		0.871255
Seen 537000 samples, time used: 125.674s
  current training loss:		0.393866
  current training accuracy:		0.871212
Seen 540000 samples, time used: 124.653s
  current training loss:		0.393905
  current training accuracy:		0.871193
saving to ..., time used: 1249.983s
Seen 543000 samples, time used: 123.758s
  current training loss:		0.394053
  current training accuracy:		0.871123
Seen 546000 samples, time used: 123.459s
  current training loss:		0.394054
  current training accuracy:		0.871143
Seen 549000 samples, time used: 124.432s
  current training loss:		0.394139
  current training accuracy:		0.871124
Epoch 6 of 20 took 23001.248s
  training loss:		0.394135
  training accuracy:		87.11 %
  validation loss:		0.412982
  validation accuracy:		84.51 %
Seen 3000 samples, time used: 125.262s
  current training loss:		0.367110
  current training accuracy:		0.886000
Seen 6000 samples, time used: 124.281s
  current training loss:		0.365065
  current training accuracy:		0.882333
Seen 9000 samples, time used: 126.028s
  current training loss:		0.357451
  current training accuracy:		0.884333
Seen 12000 samples, time used: 124.021s
  current training loss:		0.356992
  current training accuracy:		0.886667
Seen 15000 samples, time used: 128.340s
  current training loss:		0.361040
  current training accuracy:		0.887133
Seen 18000 samples, time used: 123.867s
  current training loss:		0.363064
  current training accuracy:		0.886556
Seen 21000 samples, time used: 126.667s
  current training loss:		0.361499
  current training accuracy:		0.886143
Seen 24000 samples, time used: 123.577s
  current training loss:		0.361403
  current training accuracy:		0.886083
Seen 27000 samples, time used: 123.469s
  current training loss:		0.359988
  current training accuracy:		0.886667
Seen 30000 samples, time used: 128.549s
  current training loss:		0.359968
  current training accuracy:		0.886267
saving to ..., time used: 1254.063s
Seen 33000 samples, time used: 125.971s
  current training loss:		0.359122
  current training accuracy:		0.886939
Seen 36000 samples, time used: 124.668s
  current training loss:		0.358880
  current training accuracy:		0.887444
Seen 39000 samples, time used: 124.647s
  current training loss:		0.358561
  current training accuracy:		0.887513
Seen 42000 samples, time used: 124.841s
  current training loss:		0.360216
  current training accuracy:		0.886976
Seen 45000 samples, time used: 123.831s
  current training loss:		0.361183
  current training accuracy:		0.886111
Seen 48000 samples, time used: 127.476s
  current training loss:		0.362429
  current training accuracy:		0.885833
Seen 51000 samples, time used: 129.029s
  current training loss:		0.362531
  current training accuracy:		0.885373
Seen 54000 samples, time used: 125.750s
  current training loss:		0.363288
  current training accuracy:		0.885204
Seen 57000 samples, time used: 126.765s
  current training loss:		0.363735
  current training accuracy:		0.885368
Seen 60000 samples, time used: 124.762s
  current training loss:		0.363727
  current training accuracy:		0.885333
saving to ..., time used: 1257.537s
Seen 63000 samples, time used: 127.139s
  current training loss:		0.364439
  current training accuracy:		0.884889
Seen 66000 samples, time used: 127.570s
  current training loss:		0.364905
  current training accuracy:		0.884803
Seen 69000 samples, time used: 122.797s
  current training loss:		0.365677
  current training accuracy:		0.884710
Seen 72000 samples, time used: 128.053s
  current training loss:		0.366256
  current training accuracy:		0.884361
Seen 75000 samples, time used: 125.280s
  current training loss:		0.367278
  current training accuracy:		0.884173
Seen 78000 samples, time used: 124.865s
  current training loss:		0.366834
  current training accuracy:		0.884385
Seen 81000 samples, time used: 125.387s
  current training loss:		0.367178
  current training accuracy:		0.884284
Seen 84000 samples, time used: 126.685s
  current training loss:		0.367504
  current training accuracy:		0.884190
Seen 87000 samples, time used: 125.476s
  current training loss:		0.367144
  current training accuracy:		0.884310
Seen 90000 samples, time used: 125.149s
  current training loss:		0.367982
  current training accuracy:		0.883967
saving to ..., time used: 1258.197s
Seen 93000 samples, time used: 124.470s
  current training loss:		0.367873
  current training accuracy:		0.883817
Seen 96000 samples, time used: 126.962s
  current training loss:		0.368298
  current training accuracy:		0.883667
Seen 99000 samples, time used: 121.915s
  current training loss:		0.368510
  current training accuracy:		0.883707
Seen 102000 samples, time used: 126.366s
  current training loss:		0.368480
  current training accuracy:		0.883706
Seen 105000 samples, time used: 124.770s
  current training loss:		0.368432
  current training accuracy:		0.883724
Seen 108000 samples, time used: 124.763s
  current training loss:		0.367964
  current training accuracy:		0.883935
Seen 111000 samples, time used: 124.253s
  current training loss:		0.368343
  current training accuracy:		0.883820
Seen 114000 samples, time used: 125.116s
  current training loss:		0.368471
  current training accuracy:		0.883807
Seen 117000 samples, time used: 125.400s
  current training loss:		0.368881
  current training accuracy:		0.883598
Seen 120000 samples, time used: 121.578s
  current training loss:		0.368702
  current training accuracy:		0.883650
saving to ..., time used: 1245.390s
Seen 123000 samples, time used: 124.173s
  current training loss:		0.368531
  current training accuracy:		0.883650
Seen 126000 samples, time used: 124.581s
  current training loss:		0.368856
  current training accuracy:		0.883437
Seen 129000 samples, time used: 125.827s
  current training loss:		0.368629
  current training accuracy:		0.883535
Seen 132000 samples, time used: 126.595s
  current training loss:		0.369237
  current training accuracy:		0.883424
Seen 135000 samples, time used: 126.099s
  current training loss:		0.369047
  current training accuracy:		0.883526
Seen 138000 samples, time used: 123.373s
  current training loss:		0.369022
  current training accuracy:		0.883522
Seen 141000 samples, time used: 121.675s
  current training loss:		0.369295
  current training accuracy:		0.883397
Seen 144000 samples, time used: 123.303s
  current training loss:		0.368974
  current training accuracy:		0.883632
Seen 147000 samples, time used: 123.371s
  current training loss:		0.369173
  current training accuracy:		0.883551
Seen 150000 samples, time used: 125.255s
  current training loss:		0.369189
  current training accuracy:		0.883527
saving to ..., time used: 1244.050s
Seen 153000 samples, time used: 124.623s
  current training loss:		0.369409
  current training accuracy:		0.883438
Seen 156000 samples, time used: 124.387s
  current training loss:		0.369503
  current training accuracy:		0.883487
Seen 159000 samples, time used: 124.938s
  current training loss:		0.369502
  current training accuracy:		0.883384
Seen 162000 samples, time used: 124.494s
  current training loss:		0.369243
  current training accuracy:		0.883451
Seen 165000 samples, time used: 124.218s
  current training loss:		0.369132
  current training accuracy:		0.883388
Seen 168000 samples, time used: 123.301s
  current training loss:		0.369157
  current training accuracy:		0.883321
Seen 171000 samples, time used: 124.383s
  current training loss:		0.369509
  current training accuracy:		0.883322
Seen 174000 samples, time used: 125.836s
  current training loss:		0.369614
  current training accuracy:		0.883230
Seen 177000 samples, time used: 124.769s
  current training loss:		0.369548
  current training accuracy:		0.883254
Seen 180000 samples, time used: 128.856s
  current training loss:		0.369673
  current training accuracy:		0.883200
saving to ..., time used: 1249.602s
Seen 183000 samples, time used: 127.298s
  current training loss:		0.369866
  current training accuracy:		0.883175
Seen 186000 samples, time used: 125.297s
  current training loss:		0.369877
  current training accuracy:		0.883065
Seen 189000 samples, time used: 124.856s
  current training loss:		0.369669
  current training accuracy:		0.883037
Seen 192000 samples, time used: 128.124s
  current training loss:		0.369769
  current training accuracy:		0.882880
Seen 195000 samples, time used: 125.404s
  current training loss:		0.369872
  current training accuracy:		0.882805
Seen 198000 samples, time used: 123.993s
  current training loss:		0.369784
  current training accuracy:		0.882808
Seen 201000 samples, time used: 130.456s
  current training loss:		0.369693
  current training accuracy:		0.882801
Seen 204000 samples, time used: 124.360s
  current training loss:		0.369818
  current training accuracy:		0.882696
Seen 207000 samples, time used: 127.756s
  current training loss:		0.370225
  current training accuracy:		0.882623
Seen 210000 samples, time used: 126.868s
  current training loss:		0.370622
  current training accuracy:		0.882495
saving to ..., time used: 1264.206s
Seen 213000 samples, time used: 123.959s
  current training loss:		0.370976
  current training accuracy:		0.882324
Seen 216000 samples, time used: 126.438s
  current training loss:		0.371009
  current training accuracy:		0.882236
Seen 219000 samples, time used: 124.479s
  current training loss:		0.371115
  current training accuracy:		0.882205
Seen 222000 samples, time used: 126.923s
  current training loss:		0.371499
  current training accuracy:		0.882185
Seen 225000 samples, time used: 124.297s
  current training loss:		0.371501
  current training accuracy:		0.882182
Seen 228000 samples, time used: 125.738s
  current training loss:		0.371529
  current training accuracy:		0.882215
Seen 231000 samples, time used: 124.528s
  current training loss:		0.371880
  current training accuracy:		0.882164
Seen 234000 samples, time used: 122.220s
  current training loss:		0.371921
  current training accuracy:		0.882205
Seen 237000 samples, time used: 125.669s
  current training loss:		0.372039
  current training accuracy:		0.882181
Seen 240000 samples, time used: 124.513s
  current training loss:		0.372235
  current training accuracy:		0.882062
saving to ..., time used: 1248.560s
Seen 243000 samples, time used: 127.238s
  current training loss:		0.372156
  current training accuracy:		0.882115
Seen 246000 samples, time used: 124.938s
  current training loss:		0.372230
  current training accuracy:		0.882114
Seen 249000 samples, time used: 124.182s
  current training loss:		0.372016
  current training accuracy:		0.882193
Seen 252000 samples, time used: 125.158s
  current training loss:		0.372186
  current training accuracy:		0.882087
Seen 255000 samples, time used: 128.843s
  current training loss:		0.372256
  current training accuracy:		0.882118
Seen 258000 samples, time used: 126.366s
  current training loss:		0.372090
  current training accuracy:		0.882209
Seen 261000 samples, time used: 126.108s
  current training loss:		0.372011
  current training accuracy:		0.882241
Seen 264000 samples, time used: 123.811s
  current training loss:		0.372221
  current training accuracy:		0.882170
Seen 267000 samples, time used: 123.438s
  current training loss:		0.372501
  current training accuracy:		0.882064
Seen 270000 samples, time used: 123.905s
  current training loss:		0.372373
  current training accuracy:		0.882067
saving to ..., time used: 1253.785s
Seen 273000 samples, time used: 123.331s
  current training loss:		0.372380
  current training accuracy:		0.882099
Seen 276000 samples, time used: 126.371s
  current training loss:		0.372229
  current training accuracy:		0.882159
Seen 279000 samples, time used: 128.764s
  current training loss:		0.372010
  current training accuracy:		0.882287
Seen 282000 samples, time used: 124.334s
  current training loss:		0.372057
  current training accuracy:		0.882177
Seen 285000 samples, time used: 125.619s
  current training loss:		0.372082
  current training accuracy:		0.882102
Seen 288000 samples, time used: 123.104s
  current training loss:		0.372176
  current training accuracy:		0.882073
Seen 291000 samples, time used: 124.219s
  current training loss:		0.372277
  current training accuracy:		0.881990
Seen 294000 samples, time used: 124.587s
  current training loss:		0.372357
  current training accuracy:		0.881901
Seen 297000 samples, time used: 125.360s
  current training loss:		0.372476
  current training accuracy:		0.881828
Seen 300000 samples, time used: 122.339s
  current training loss:		0.372420
  current training accuracy:		0.881847
saving to ..., time used: 1247.821s
Seen 303000 samples, time used: 126.879s
  current training loss:		0.372497
  current training accuracy:		0.881832
Seen 306000 samples, time used: 125.752s
  current training loss:		0.372570
  current training accuracy:		0.881817
Seen 309000 samples, time used: 124.626s
  current training loss:		0.372806
  current training accuracy:		0.881673
Seen 312000 samples, time used: 122.432s
  current training loss:		0.372711
  current training accuracy:		0.881747
Seen 315000 samples, time used: 124.870s
  current training loss:		0.372898
  current training accuracy:		0.881670
Seen 318000 samples, time used: 124.676s
  current training loss:		0.373236
  current training accuracy:		0.881528
Seen 321000 samples, time used: 124.492s
  current training loss:		0.373317
  current training accuracy:		0.881467
Seen 324000 samples, time used: 124.615s
  current training loss:		0.373358
  current training accuracy:		0.881497
Seen 327000 samples, time used: 123.086s
  current training loss:		0.373461
  current training accuracy:		0.881541
Seen 330000 samples, time used: 124.999s
  current training loss:		0.373470
  current training accuracy:		0.881536
saving to ..., time used: 1246.222s
Seen 333000 samples, time used: 125.194s
  current training loss:		0.373539
  current training accuracy:		0.881453
Seen 336000 samples, time used: 127.354s
  current training loss:		0.373492
  current training accuracy:		0.881455
Seen 339000 samples, time used: 122.933s
  current training loss:		0.373328
  current training accuracy:		0.881519
Seen 342000 samples, time used: 128.715s
  current training loss:		0.373336
  current training accuracy:		0.881477
Seen 345000 samples, time used: 122.279s
  current training loss:		0.373265
  current training accuracy:		0.881417
Seen 348000 samples, time used: 126.497s
  current training loss:		0.373377
  current training accuracy:		0.881388
Seen 351000 samples, time used: 124.721s
  current training loss:		0.373279
  current training accuracy:		0.881447
Seen 354000 samples, time used: 124.807s
  current training loss:		0.373439
  current training accuracy:		0.881362
Seen 357000 samples, time used: 126.797s
  current training loss:		0.373488
  current training accuracy:		0.881370
Seen 360000 samples, time used: 126.762s
  current training loss:		0.373661
  current training accuracy:		0.881314
saving to ..., time used: 1255.857s
Seen 363000 samples, time used: 126.199s
  current training loss:		0.373568
  current training accuracy:		0.881298
Seen 366000 samples, time used: 124.576s
  current training loss:		0.373735
  current training accuracy:		0.881227
Seen 369000 samples, time used: 122.959s
  current training loss:		0.373637
  current training accuracy:		0.881279
Seen 372000 samples, time used: 126.373s
  current training loss:		0.373760
  current training accuracy:		0.881194
Seen 375000 samples, time used: 128.691s
  current training loss:		0.373842
  current training accuracy:		0.881155
Seen 378000 samples, time used: 123.402s
  current training loss:		0.373847
  current training accuracy:		0.881175
Seen 381000 samples, time used: 125.014s
  current training loss:		0.373877
  current training accuracy:		0.881142
Seen 384000 samples, time used: 124.912s
  current training loss:		0.373773
  current training accuracy:		0.881182
Seen 387000 samples, time used: 124.588s
  current training loss:		0.373707
  current training accuracy:		0.881220
Seen 390000 samples, time used: 128.253s
  current training loss:		0.373684
  current training accuracy:		0.881226
saving to ..., time used: 1254.765s
Seen 393000 samples, time used: 127.135s
  current training loss:		0.374000
  current training accuracy:		0.881069
Seen 396000 samples, time used: 123.380s
  current training loss:		0.374124
  current training accuracy:		0.881018
Seen 399000 samples, time used: 125.840s
  current training loss:		0.374282
  current training accuracy:		0.880940
Seen 402000 samples, time used: 124.767s
  current training loss:		0.374350
  current training accuracy:		0.880863
Seen 405000 samples, time used: 125.896s
  current training loss:		0.374464
  current training accuracy:		0.880798
Seen 408000 samples, time used: 125.562s
  current training loss:		0.374338
  current training accuracy:		0.880816
Seen 411000 samples, time used: 125.687s
  current training loss:		0.374367
  current training accuracy:		0.880798
Seen 414000 samples, time used: 128.637s
  current training loss:		0.374326
  current training accuracy:		0.880778
Seen 417000 samples, time used: 126.913s
  current training loss:		0.374351
  current training accuracy:		0.880763
Seen 420000 samples, time used: 123.559s
  current training loss:		0.374340
  current training accuracy:		0.880767
saving to ..., time used: 1257.173s
Seen 423000 samples, time used: 128.808s
  current training loss:		0.374394
  current training accuracy:		0.880740
Seen 426000 samples, time used: 126.257s
  current training loss:		0.374526
  current training accuracy:		0.880688
Seen 429000 samples, time used: 122.761s
  current training loss:		0.374779
  current training accuracy:		0.880611
Seen 432000 samples, time used: 123.349s
  current training loss:		0.374763
  current training accuracy:		0.880562
Seen 435000 samples, time used: 125.249s
  current training loss:		0.374877
  current training accuracy:		0.880552
Seen 438000 samples, time used: 127.269s
  current training loss:		0.374915
  current training accuracy:		0.880511
Seen 441000 samples, time used: 123.552s
  current training loss:		0.375076
  current training accuracy:		0.880478
Seen 444000 samples, time used: 128.979s
  current training loss:		0.375143
  current training accuracy:		0.880489
Seen 447000 samples, time used: 123.739s
  current training loss:		0.374964
  current training accuracy:		0.880526
Seen 450000 samples, time used: 124.794s
  current training loss:		0.375013
  current training accuracy:		0.880507
saving to ..., time used: 1254.556s
Seen 453000 samples, time used: 123.719s
  current training loss:		0.375006
  current training accuracy:		0.880435
Seen 456000 samples, time used: 122.804s
  current training loss:		0.374987
  current training accuracy:		0.880439
Seen 459000 samples, time used: 125.018s
  current training loss:		0.375084
  current training accuracy:		0.880390
Seen 462000 samples, time used: 123.461s
  current training loss:		0.375140
  current training accuracy:		0.880377
Seen 465000 samples, time used: 125.714s
  current training loss:		0.375106
  current training accuracy:		0.880366
Seen 468000 samples, time used: 123.479s
  current training loss:		0.375215
  current training accuracy:		0.880314
Seen 471000 samples, time used: 125.613s
  current training loss:		0.375370
  current training accuracy:		0.880251
Seen 474000 samples, time used: 124.133s
  current training loss:		0.375333
  current training accuracy:		0.880230
Seen 477000 samples, time used: 125.035s
  current training loss:		0.375486
  current training accuracy:		0.880203
Seen 480000 samples, time used: 122.064s
  current training loss:		0.375638
  current training accuracy:		0.880112
saving to ..., time used: 1240.836s
Seen 483000 samples, time used: 125.175s
  current training loss:		0.375656
  current training accuracy:		0.880110
Seen 486000 samples, time used: 125.799s
  current training loss:		0.375721
  current training accuracy:		0.880091
Seen 489000 samples, time used: 128.439s
  current training loss:		0.375649
  current training accuracy:		0.880121
Seen 492000 samples, time used: 122.835s
  current training loss:		0.375595
  current training accuracy:		0.880163
Seen 495000 samples, time used: 122.074s
  current training loss:		0.375677
  current training accuracy:		0.880158
Seen 498000 samples, time used: 129.015s
  current training loss:		0.375737
  current training accuracy:		0.880106
Seen 501000 samples, time used: 125.831s
  current training loss:		0.375797
  current training accuracy:		0.880104
Seen 504000 samples, time used: 123.696s
  current training loss:		0.375868
  current training accuracy:		0.880089
Seen 507000 samples, time used: 125.480s
  current training loss:		0.376003
  current training accuracy:		0.880081
Seen 510000 samples, time used: 124.265s
  current training loss:		0.376083
  current training accuracy:		0.880075
saving to ..., time used: 1252.405s
Seen 513000 samples, time used: 122.992s
  current training loss:		0.376181
  current training accuracy:		0.880019
Seen 516000 samples, time used: 124.471s
  current training loss:		0.376093
  current training accuracy:		0.880017
Seen 519000 samples, time used: 123.716s
  current training loss:		0.375998
  current training accuracy:		0.880046
Seen 522000 samples, time used: 126.215s
  current training loss:		0.376319
  current training accuracy:		0.879914
Seen 525000 samples, time used: 132.982s
  current training loss:		0.376392
  current training accuracy:		0.879926
Seen 528000 samples, time used: 124.114s
  current training loss:		0.376458
  current training accuracy:		0.879930
Seen 531000 samples, time used: 124.451s
  current training loss:		0.376408
  current training accuracy:		0.879962
Seen 534000 samples, time used: 126.297s
  current training loss:		0.376438
  current training accuracy:		0.879993
Seen 537000 samples, time used: 122.296s
  current training loss:		0.376582
  current training accuracy:		0.879955
Seen 540000 samples, time used: 124.091s
  current training loss:		0.376676
  current training accuracy:		0.879939
saving to ..., time used: 1251.421s
Seen 543000 samples, time used: 124.098s
  current training loss:		0.376862
  current training accuracy:		0.879866
Seen 546000 samples, time used: 124.502s
  current training loss:		0.376998
  current training accuracy:		0.879804
Seen 549000 samples, time used: 122.707s
  current training loss:		0.377031
  current training accuracy:		0.879763
Epoch 7 of 20 took 22956.110s
  training loss:		0.377049
  training accuracy:		87.98 %
  validation loss:		0.414955
  validation accuracy:		84.57 %
Seen 3000 samples, time used: 124.448s
  current training loss:		0.362737
  current training accuracy:		0.893667
Seen 6000 samples, time used: 128.076s
  current training loss:		0.362740
  current training accuracy:		0.890000
Seen 9000 samples, time used: 123.832s
  current training loss:		0.355843
  current training accuracy:		0.890556
Seen 12000 samples, time used: 123.591s
  current training loss:		0.352137
  current training accuracy:		0.894083
Seen 15000 samples, time used: 124.433s
  current training loss:		0.349932
  current training accuracy:		0.894400
Seen 18000 samples, time used: 124.569s
  current training loss:		0.347650
  current training accuracy:		0.894944
Seen 21000 samples, time used: 125.487s
  current training loss:		0.347031
  current training accuracy:		0.894095
Seen 24000 samples, time used: 124.088s
  current training loss:		0.348646
  current training accuracy:		0.893500
Seen 27000 samples, time used: 125.752s
  current training loss:		0.347165
  current training accuracy:		0.894074
Seen 30000 samples, time used: 125.328s
  current training loss:		0.347296
  current training accuracy:		0.894000
saving to ..., time used: 1249.607s
Seen 33000 samples, time used: 122.133s
  current training loss:		0.346001
  current training accuracy:		0.894606
Seen 36000 samples, time used: 125.653s
  current training loss:		0.346405
  current training accuracy:		0.894556
Seen 39000 samples, time used: 127.057s
  current training loss:		0.347999
  current training accuracy:		0.893821
Seen 42000 samples, time used: 125.484s
  current training loss:		0.350147
  current training accuracy:		0.892976
Seen 45000 samples, time used: 126.641s
  current training loss:		0.350164
  current training accuracy:		0.893067
Seen 48000 samples, time used: 125.719s
  current training loss:		0.351610
  current training accuracy:		0.892500
Seen 51000 samples, time used: 125.961s
  current training loss:		0.350375
  current training accuracy:		0.892804
Seen 54000 samples, time used: 126.251s
  current training loss:		0.350550
  current training accuracy:		0.892907
Seen 57000 samples, time used: 124.327s
  current training loss:		0.350046
  current training accuracy:		0.893088
Seen 60000 samples, time used: 126.309s
  current training loss:		0.350135
  current training accuracy:		0.892683
saving to ..., time used: 1255.330s
Seen 63000 samples, time used: 126.867s
  current training loss:		0.350444
  current training accuracy:		0.892794
Seen 66000 samples, time used: 124.982s
  current training loss:		0.350668
  current training accuracy:		0.892894
Seen 69000 samples, time used: 125.648s
  current training loss:		0.350437
  current training accuracy:		0.892623
Seen 72000 samples, time used: 124.019s
  current training loss:		0.350617
  current training accuracy:		0.892486
Seen 75000 samples, time used: 124.215s
  current training loss:		0.350836
  current training accuracy:		0.892427
Seen 78000 samples, time used: 126.658s
  current training loss:		0.349941
  current training accuracy:		0.892487
Seen 81000 samples, time used: 124.347s
  current training loss:		0.350103
  current training accuracy:		0.892407
Seen 84000 samples, time used: 124.847s
  current training loss:		0.349788
  current training accuracy:		0.892571
Seen 87000 samples, time used: 123.070s
  current training loss:		0.350762
  current training accuracy:		0.892287
Seen 90000 samples, time used: 126.882s
  current training loss:		0.351383
  current training accuracy:		0.892044
saving to ..., time used: 1251.331s
Seen 93000 samples, time used: 128.978s
  current training loss:		0.350754
  current training accuracy:		0.892290
Seen 96000 samples, time used: 126.417s
  current training loss:		0.350819
  current training accuracy:		0.892187
Seen 99000 samples, time used: 126.510s
  current training loss:		0.351362
  current training accuracy:		0.892162
Seen 102000 samples, time used: 126.795s
  current training loss:		0.351640
  current training accuracy:		0.892333
Seen 105000 samples, time used: 123.514s
  current training loss:		0.351828
  current training accuracy:		0.892295
Seen 108000 samples, time used: 125.540s
  current training loss:		0.351946
  current training accuracy:		0.892324
Seen 111000 samples, time used: 126.763s
  current training loss:		0.351368
  current training accuracy:		0.892631
Seen 114000 samples, time used: 122.902s
  current training loss:		0.352019
  current training accuracy:		0.892412
Seen 117000 samples, time used: 123.571s
  current training loss:		0.351999
  current training accuracy:		0.892513
Seen 120000 samples, time used: 125.664s
  current training loss:		0.352124
  current training accuracy:		0.892200
saving to ..., time used: 1256.448s
Seen 123000 samples, time used: 128.466s
  current training loss:		0.352108
  current training accuracy:		0.892293
Seen 126000 samples, time used: 121.505s
  current training loss:		0.351919
  current training accuracy:		0.892389
Seen 129000 samples, time used: 124.639s
  current training loss:		0.352263
  current training accuracy:		0.892256
Seen 132000 samples, time used: 126.414s
  current training loss:		0.352826
  current training accuracy:		0.892136
Seen 135000 samples, time used: 125.144s
  current training loss:		0.352853
  current training accuracy:		0.892178
Seen 138000 samples, time used: 130.101s
  current training loss:		0.352989
  current training accuracy:		0.892188
Seen 141000 samples, time used: 126.561s
  current training loss:		0.353214
  current training accuracy:		0.891950
Seen 144000 samples, time used: 126.484s
  current training loss:		0.352891
  current training accuracy:		0.891986
Seen 147000 samples, time used: 127.245s
  current training loss:		0.353462
  current training accuracy:		0.891687
Seen 150000 samples, time used: 124.518s
  current training loss:		0.353690
  current training accuracy:		0.891633
saving to ..., time used: 1260.874s
Seen 153000 samples, time used: 126.990s
  current training loss:		0.353511
  current training accuracy:		0.891569
Seen 156000 samples, time used: 129.241s
  current training loss:		0.354067
  current training accuracy:		0.891327
Seen 159000 samples, time used: 121.965s
  current training loss:		0.354544
  current training accuracy:		0.891157
Seen 162000 samples, time used: 128.648s
  current training loss:		0.354188
  current training accuracy:		0.891272
Seen 165000 samples, time used: 127.262s
  current training loss:		0.353901
  current training accuracy:		0.891352
Seen 168000 samples, time used: 124.745s
  current training loss:		0.354124
  current training accuracy:		0.891214
Seen 171000 samples, time used: 129.013s
  current training loss:		0.354297
  current training accuracy:		0.891117
Seen 174000 samples, time used: 123.357s
  current training loss:		0.354589
  current training accuracy:		0.891006
Seen 177000 samples, time used: 126.623s
  current training loss:		0.354643
  current training accuracy:		0.890966
Seen 180000 samples, time used: 127.493s
  current training loss:		0.354676
  current training accuracy:		0.890933
saving to ..., time used: 1265.132s
Seen 183000 samples, time used: 129.575s
  current training loss:		0.354701
  current training accuracy:		0.890995
Seen 186000 samples, time used: 122.652s
  current training loss:		0.354553
  current training accuracy:		0.891005
Seen 189000 samples, time used: 128.035s
  current training loss:		0.354801
  current training accuracy:		0.891000
Seen 192000 samples, time used: 126.754s
  current training loss:		0.354840
  current training accuracy:		0.890875
Seen 195000 samples, time used: 125.313s
  current training loss:		0.355190
  current training accuracy:		0.890856
Seen 198000 samples, time used: 124.822s
  current training loss:		0.355390
  current training accuracy:		0.890747
Seen 201000 samples, time used: 127.219s
  current training loss:		0.355205
  current training accuracy:		0.890841
Seen 204000 samples, time used: 125.827s
  current training loss:		0.355391
  current training accuracy:		0.890755
Seen 207000 samples, time used: 123.715s
  current training loss:		0.355323
  current training accuracy:		0.890807
Seen 210000 samples, time used: 127.669s
  current training loss:		0.355112
  current training accuracy:		0.890876
saving to ..., time used: 1261.377s
Seen 213000 samples, time used: 126.621s
  current training loss:		0.355141
  current training accuracy:		0.890840
Seen 216000 samples, time used: 123.098s
  current training loss:		0.355357
  current training accuracy:		0.890824
Seen 219000 samples, time used: 124.193s
  current training loss:		0.355339
  current training accuracy:		0.890790
Seen 222000 samples, time used: 123.441s
  current training loss:		0.355231
  current training accuracy:		0.890887
Seen 225000 samples, time used: 123.631s
  current training loss:		0.355178
  current training accuracy:		0.890889
Seen 228000 samples, time used: 124.519s
  current training loss:		0.354801
  current training accuracy:		0.890965
Seen 231000 samples, time used: 126.613s
  current training loss:		0.354950
  current training accuracy:		0.891043
Seen 234000 samples, time used: 123.121s
  current training loss:		0.355214
  current training accuracy:		0.890902
Seen 237000 samples, time used: 123.885s
  current training loss:		0.355317
  current training accuracy:		0.890886
Seen 240000 samples, time used: 124.100s
  current training loss:		0.355492
  current training accuracy:		0.890829
saving to ..., time used: 1243.019s
Seen 243000 samples, time used: 125.896s
  current training loss:		0.355767
  current training accuracy:		0.890778
Seen 246000 samples, time used: 123.370s
  current training loss:		0.355856
  current training accuracy:		0.890785
Seen 249000 samples, time used: 124.799s
  current training loss:		0.355960
  current training accuracy:		0.890779
Seen 252000 samples, time used: 131.153s
  current training loss:		0.356066
  current training accuracy:		0.890802
Seen 255000 samples, time used: 126.804s
  current training loss:		0.356143
  current training accuracy:		0.890788
Seen 258000 samples, time used: 123.619s
  current training loss:		0.356222
  current training accuracy:		0.890787
Seen 261000 samples, time used: 125.261s
  current training loss:		0.356407
  current training accuracy:		0.890766
Seen 264000 samples, time used: 122.880s
  current training loss:		0.356648
  current training accuracy:		0.890674
Seen 267000 samples, time used: 127.940s
  current training loss:		0.356573
  current training accuracy:		0.890610
Seen 270000 samples, time used: 124.049s
  current training loss:		0.356477
  current training accuracy:		0.890633
saving to ..., time used: 1255.567s
Seen 273000 samples, time used: 124.633s
  current training loss:		0.356620
  current training accuracy:		0.890557
Seen 276000 samples, time used: 126.209s
  current training loss:		0.356971
  current training accuracy:		0.890504
Seen 279000 samples, time used: 123.632s
  current training loss:		0.357238
  current training accuracy:		0.890366
Seen 282000 samples, time used: 123.966s
  current training loss:		0.357237
  current training accuracy:		0.890404
Seen 285000 samples, time used: 123.043s
  current training loss:		0.357330
  current training accuracy:		0.890337
Seen 288000 samples, time used: 124.127s
  current training loss:		0.357483
  current training accuracy:		0.890233
Seen 291000 samples, time used: 130.303s
  current training loss:		0.357672
  current training accuracy:		0.890186
Seen 294000 samples, time used: 123.757s
  current training loss:		0.357763
  current training accuracy:		0.890116
Seen 297000 samples, time used: 123.237s
  current training loss:		0.357499
  current training accuracy:		0.890232
Seen 300000 samples, time used: 126.449s
  current training loss:		0.357452
  current training accuracy:		0.890217
saving to ..., time used: 1249.151s
Seen 303000 samples, time used: 124.246s
  current training loss:		0.357780
  current training accuracy:		0.890096
Seen 306000 samples, time used: 123.621s
  current training loss:		0.357964
  current training accuracy:		0.889938
Seen 309000 samples, time used: 128.182s
  current training loss:		0.358230
  current training accuracy:		0.889838
Seen 312000 samples, time used: 127.446s
  current training loss:		0.358334
  current training accuracy:		0.889721
Seen 315000 samples, time used: 123.392s
  current training loss:		0.358332
  current training accuracy:		0.889721
Seen 318000 samples, time used: 127.489s
  current training loss:		0.358459
  current training accuracy:		0.889635
Seen 321000 samples, time used: 121.870s
  current training loss:		0.358650
  current training accuracy:		0.889570
Seen 324000 samples, time used: 129.141s
  current training loss:		0.358742
  current training accuracy:		0.889540
Seen 327000 samples, time used: 121.028s
  current training loss:		0.358936
  current training accuracy:		0.889453
Seen 330000 samples, time used: 128.160s
  current training loss:		0.359139
  current training accuracy:		0.889303
saving to ..., time used: 1254.372s
Seen 333000 samples, time used: 127.679s
  current training loss:		0.359099
  current training accuracy:		0.889312
Seen 336000 samples, time used: 122.606s
  current training loss:		0.358986
  current training accuracy:		0.889298
Seen 339000 samples, time used: 127.573s
  current training loss:		0.359075
  current training accuracy:		0.889251
Seen 342000 samples, time used: 128.807s
  current training loss:		0.359304
  current training accuracy:		0.889117
Seen 345000 samples, time used: 128.378s
  current training loss:		0.359420
  current training accuracy:		0.889032
Seen 348000 samples, time used: 126.558s
  current training loss:		0.359394
  current training accuracy:		0.889043
Seen 351000 samples, time used: 127.848s
  current training loss:		0.359322
  current training accuracy:		0.889071
Seen 354000 samples, time used: 131.038s
  current training loss:		0.359298
  current training accuracy:		0.889062
Seen 357000 samples, time used: 126.147s
  current training loss:		0.359405
  current training accuracy:		0.889003
Seen 360000 samples, time used: 125.935s
  current training loss:		0.359423
  current training accuracy:		0.888989
saving to ..., time used: 1272.365s
Seen 363000 samples, time used: 127.273s
  current training loss:		0.359506
  current training accuracy:		0.888959
Seen 366000 samples, time used: 131.612s
  current training loss:		0.359490
  current training accuracy:		0.888934
Seen 369000 samples, time used: 124.908s
  current training loss:		0.359535
  current training accuracy:		0.888908
Seen 372000 samples, time used: 130.511s
  current training loss:		0.359595
  current training accuracy:		0.888860
Seen 375000 samples, time used: 129.585s
  current training loss:		0.359585
  current training accuracy:		0.888827
Seen 378000 samples, time used: 128.755s
  current training loss:		0.359701
  current training accuracy:		0.888722
Seen 381000 samples, time used: 126.022s
  current training loss:		0.359881
  current training accuracy:		0.888598
Seen 384000 samples, time used: 126.606s
  current training loss:		0.360015
  current training accuracy:		0.888542
Seen 387000 samples, time used: 128.982s
  current training loss:		0.360162
  current training accuracy:		0.888514
Seen 390000 samples, time used: 124.956s
  current training loss:		0.360124
  current training accuracy:		0.888490
saving to ..., time used: 1278.476s
Seen 393000 samples, time used: 126.206s
  current training loss:		0.360252
  current training accuracy:		0.888420
Seen 396000 samples, time used: 124.305s
  current training loss:		0.360420
  current training accuracy:		0.888364
Seen 399000 samples, time used: 125.731s
  current training loss:		0.360568
  current training accuracy:		0.888273
Seen 402000 samples, time used: 123.588s
  current training loss:		0.360723
  current training accuracy:		0.888244
Seen 405000 samples, time used: 127.715s
  current training loss:		0.360891
  current training accuracy:		0.888185
Seen 408000 samples, time used: 128.215s
  current training loss:		0.360889
  current training accuracy:		0.888184
Seen 411000 samples, time used: 128.148s
  current training loss:		0.360932
  current training accuracy:		0.888136
Seen 414000 samples, time used: 128.086s
  current training loss:		0.360966
  current training accuracy:		0.888128
Seen 417000 samples, time used: 122.807s
  current training loss:		0.360995
  current training accuracy:		0.888113
Seen 420000 samples, time used: 127.110s
  current training loss:		0.361007
  current training accuracy:		0.888117
saving to ..., time used: 1261.518s
Seen 423000 samples, time used: 125.515s
  current training loss:		0.360993
  current training accuracy:		0.888047
Seen 426000 samples, time used: 127.217s
  current training loss:		0.361158
  current training accuracy:		0.888002
Seen 429000 samples, time used: 125.984s
  current training loss:		0.361231
  current training accuracy:		0.887949
Seen 432000 samples, time used: 126.368s
  current training loss:		0.361191
  current training accuracy:		0.887956
Seen 435000 samples, time used: 129.457s
  current training loss:		0.361288
  current training accuracy:		0.887952
Seen 438000 samples, time used: 124.408s
  current training loss:		0.361300
  current training accuracy:		0.887934
Seen 441000 samples, time used: 123.046s
  current training loss:		0.361414
  current training accuracy:		0.887893
Seen 444000 samples, time used: 122.052s
  current training loss:		0.361510
  current training accuracy:		0.887845
Seen 447000 samples, time used: 125.821s
  current training loss:		0.361485
  current training accuracy:		0.887843
Seen 450000 samples, time used: 126.822s
  current training loss:		0.361464
  current training accuracy:		0.887858
saving to ..., time used: 1256.390s
Seen 453000 samples, time used: 126.127s
  current training loss:		0.361513
  current training accuracy:		0.887819
Seen 456000 samples, time used: 127.613s
  current training loss:		0.361621
  current training accuracy:		0.887752
Seen 459000 samples, time used: 125.854s
  current training loss:		0.361697
  current training accuracy:		0.887695
Seen 462000 samples, time used: 125.025s
  current training loss:		0.361546
  current training accuracy:		0.887747
Seen 465000 samples, time used: 125.587s
  current training loss:		0.361506
  current training accuracy:		0.887757
Seen 468000 samples, time used: 126.393s
  current training loss:		0.361661
  current training accuracy:		0.887699
Seen 471000 samples, time used: 124.542s
  current training loss:		0.361547
  current training accuracy:		0.887732
Seen 474000 samples, time used: 125.183s
  current training loss:		0.361740
  current training accuracy:		0.887646
Seen 477000 samples, time used: 124.920s
  current training loss:		0.361777
  current training accuracy:		0.887606
Seen 480000 samples, time used: 124.285s
  current training loss:		0.361854
  current training accuracy:		0.887546
saving to ..., time used: 1255.324s
Seen 483000 samples, time used: 124.033s
  current training loss:		0.361806
  current training accuracy:		0.887551
Seen 486000 samples, time used: 125.652s
  current training loss:		0.361912
  current training accuracy:		0.887494
Seen 489000 samples, time used: 121.825s
  current training loss:		0.361873
  current training accuracy:		0.887481
Seen 492000 samples, time used: 125.554s
  current training loss:		0.361897
  current training accuracy:		0.887443
Seen 495000 samples, time used: 126.382s
  current training loss:		0.361930
  current training accuracy:		0.887402
Seen 498000 samples, time used: 129.208s
  current training loss:		0.362068
  current training accuracy:		0.887333
Seen 501000 samples, time used: 129.793s
  current training loss:		0.362100
  current training accuracy:		0.887285
Seen 504000 samples, time used: 126.809s
  current training loss:		0.362167
  current training accuracy:		0.887240
Seen 507000 samples, time used: 126.648s
  current training loss:		0.362206
  current training accuracy:		0.887266
Seen 510000 samples, time used: 126.932s
  current training loss:		0.362114
  current training accuracy:		0.887245
saving to ..., time used: 1262.631s
Seen 513000 samples, time used: 125.795s
  current training loss:		0.362453
  current training accuracy:		0.887097
Seen 516000 samples, time used: 126.995s
  current training loss:		0.362466
  current training accuracy:		0.887101
Seen 519000 samples, time used: 126.481s
  current training loss:		0.362532
  current training accuracy:		0.887064
Seen 522000 samples, time used: 124.754s
  current training loss:		0.362493
  current training accuracy:		0.887048
Seen 525000 samples, time used: 126.561s
  current training loss:		0.362627
  current training accuracy:		0.886960
Seen 528000 samples, time used: 124.525s
  current training loss:		0.362718
  current training accuracy:		0.886919
Seen 531000 samples, time used: 122.111s
  current training loss:		0.362713
  current training accuracy:		0.886911
Seen 534000 samples, time used: 128.773s
  current training loss:		0.362748
  current training accuracy:		0.886934
Seen 537000 samples, time used: 124.108s
  current training loss:		0.362764
  current training accuracy:		0.886907
Seen 540000 samples, time used: 126.538s
  current training loss:		0.362801
  current training accuracy:		0.886891
saving to ..., time used: 1256.161s
Seen 543000 samples, time used: 124.988s
  current training loss:		0.362875
  current training accuracy:		0.886858
Seen 546000 samples, time used: 128.818s
  current training loss:		0.362924
  current training accuracy:		0.886866
Seen 549000 samples, time used: 122.055s
  current training loss:		0.362943
  current training accuracy:		0.886851
Epoch 8 of 20 took 23070.145s
  training loss:		0.362947
  training accuracy:		88.68 %
  validation loss:		0.415264
  validation accuracy:		84.91 %
Seen 3000 samples, time used: 125.282s
  current training loss:		0.327025
  current training accuracy:		0.898333
Seen 6000 samples, time used: 123.731s
  current training loss:		0.338462
  current training accuracy:		0.894333
Seen 9000 samples, time used: 123.632s
  current training loss:		0.339741
  current training accuracy:		0.897000
Seen 12000 samples, time used: 128.339s
  current training loss:		0.342002
  current training accuracy:		0.896333
Seen 15000 samples, time used: 122.464s
  current training loss:		0.336319
  current training accuracy:		0.898267
Seen 18000 samples, time used: 119.743s
  current training loss:		0.335286
  current training accuracy:		0.899500
Seen 21000 samples, time used: 125.292s
  current training loss:		0.334326
  current training accuracy:		0.900571
Seen 24000 samples, time used: 127.237s
  current training loss:		0.336381
  current training accuracy:		0.900167
Seen 27000 samples, time used: 123.155s
  current training loss:		0.336920
  current training accuracy:		0.899556
Seen 30000 samples, time used: 129.279s
  current training loss:		0.338014
  current training accuracy:		0.898933
saving to ..., time used: 1248.155s
Seen 33000 samples, time used: 126.766s
  current training loss:		0.338139
  current training accuracy:		0.898636
Seen 36000 samples, time used: 121.580s
  current training loss:		0.337693
  current training accuracy:		0.898611
Seen 39000 samples, time used: 129.324s
  current training loss:		0.336128
  current training accuracy:		0.899205
Seen 42000 samples, time used: 125.147s
  current training loss:		0.335193
  current training accuracy:		0.899214
Seen 45000 samples, time used: 125.833s
  current training loss:		0.335802
  current training accuracy:		0.899067
Seen 48000 samples, time used: 130.277s
  current training loss:		0.337302
  current training accuracy:		0.898500
Seen 51000 samples, time used: 122.979s
  current training loss:		0.336986
  current training accuracy:		0.898745
Seen 54000 samples, time used: 123.877s
  current training loss:		0.337554
  current training accuracy:		0.899000
Seen 57000 samples, time used: 127.172s
  current training loss:		0.338186
  current training accuracy:		0.898789
Seen 60000 samples, time used: 127.550s
  current training loss:		0.337194
  current training accuracy:		0.899067
saving to ..., time used: 1260.303s
Seen 63000 samples, time used: 122.322s
  current training loss:		0.337756
  current training accuracy:		0.898873
Seen 66000 samples, time used: 124.948s
  current training loss:		0.337845
  current training accuracy:		0.898773
Seen 69000 samples, time used: 123.932s
  current training loss:		0.338108
  current training accuracy:		0.898855
Seen 72000 samples, time used: 122.697s
  current training loss:		0.338135
  current training accuracy:		0.898861
Seen 75000 samples, time used: 127.390s
  current training loss:		0.338085
  current training accuracy:		0.898813
Seen 78000 samples, time used: 125.692s
  current training loss:		0.338344
  current training accuracy:		0.898641
Seen 81000 samples, time used: 123.854s
  current training loss:		0.338761
  current training accuracy:		0.898556
Seen 84000 samples, time used: 121.063s
  current training loss:		0.339173
  current training accuracy:		0.898286
Seen 87000 samples, time used: 127.432s
  current training loss:		0.340397
  current training accuracy:		0.897885
Seen 90000 samples, time used: 125.680s
  current training loss:		0.339770
  current training accuracy:		0.898444
saving to ..., time used: 1244.809s
Seen 93000 samples, time used: 124.223s
  current training loss:		0.339696
  current training accuracy:		0.898473
Seen 96000 samples, time used: 123.431s
  current training loss:		0.339767
  current training accuracy:		0.898552
Seen 99000 samples, time used: 125.588s
  current training loss:		0.339998
  current training accuracy:		0.898667
Seen 102000 samples, time used: 124.422s
  current training loss:		0.340226
  current training accuracy:		0.898490
Seen 105000 samples, time used: 125.106s
  current training loss:		0.340283
  current training accuracy:		0.898400
Seen 108000 samples, time used: 126.286s
  current training loss:		0.341078
  current training accuracy:		0.898185
Seen 111000 samples, time used: 123.589s
  current training loss:		0.341017
  current training accuracy:		0.898126
Seen 114000 samples, time used: 125.171s
  current training loss:		0.341091
  current training accuracy:		0.897912
Seen 117000 samples, time used: 125.337s
  current training loss:		0.340779
  current training accuracy:		0.898188

Seen 120000 samples, time used: 125.553s
  current training loss:		0.340893
  current training accuracy:		0.898058
saving to ..., time used: 1248.505s
Seen 123000 samples, time used: 124.305s
  current training loss:		0.341681
  current training accuracy:		0.897699
Seen 126000 samples, time used: 124.862s
  current training loss:		0.342094
  current training accuracy:		0.897421
Seen 129000 samples, time used: 122.937s
  current training loss:		0.342065
  current training accuracy:		0.897388
Seen 132000 samples, time used: 124.890s
  current training loss:		0.341825
  current training accuracy:		0.897424
Seen 135000 samples, time used: 125.879s
  current training loss:		0.342333
  current training accuracy:		0.897341
Seen 138000 samples, time used: 125.024s
  current training loss:		0.342575
  current training accuracy:		0.897232
Seen 141000 samples, time used: 123.646s
  current training loss:		0.342543
  current training accuracy:		0.897312
Seen 144000 samples, time used: 124.597s
  current training loss:		0.342607
  current training accuracy:		0.897264
Seen 147000 samples, time used: 126.397s
  current training loss:		0.342749
  current training accuracy:		0.897156
Seen 150000 samples, time used: 128.476s
  current training loss:		0.342766
  current training accuracy:		0.897200
saving to ..., time used: 1250.812s
Seen 153000 samples, time used: 125.602s
  current training loss:		0.343065
  current training accuracy:		0.897137
Seen 156000 samples, time used: 123.634s
  current training loss:		0.343445
  current training accuracy:		0.896942
Seen 159000 samples, time used: 126.201s
  current training loss:		0.343073
  current training accuracy:		0.897157
Seen 162000 samples, time used: 128.901s
  current training loss:		0.343291
  current training accuracy:		0.897123
Seen 165000 samples, time used: 121.065s
  current training loss:		0.342943
  current training accuracy:		0.897297
Seen 168000 samples, time used: 125.553s
  current training loss:		0.343038
  current training accuracy:		0.897232
Seen 171000 samples, time used: 127.298s
  current training loss:		0.342842
  current training accuracy:		0.897275
Seen 174000 samples, time used: 123.573s
  current training loss:		0.342801
  current training accuracy:		0.897287
Seen 177000 samples, time used: 124.713s
  current training loss:		0.342705
  current training accuracy:		0.897249
Seen 180000 samples, time used: 126.159s
  current training loss:		0.342432
  current training accuracy:		0.897383
saving to ..., time used: 1252.499s
Seen 183000 samples, time used: 125.389s
  current training loss:		0.342217
  current training accuracy:		0.897355
Seen 186000 samples, time used: 126.224s
  current training loss:		0.342431
  current training accuracy:		0.897290
Seen 189000 samples, time used: 123.326s
  current training loss:		0.342463
  current training accuracy:		0.897212
Seen 192000 samples, time used: 121.854s
  current training loss:		0.342528
  current training accuracy:		0.897323
Seen 195000 samples, time used: 129.004s
  current training loss:		0.342942
  current training accuracy:		0.896985
Seen 198000 samples, time used: 123.436s
  current training loss:		0.342723
  current training accuracy:		0.897096
Seen 201000 samples, time used: 122.159s
  current training loss:		0.342951
  current training accuracy:		0.897005
Seen 204000 samples, time used: 123.548s
  current training loss:		0.342736
  current training accuracy:		0.897069
Seen 207000 samples, time used: 123.389s
  current training loss:		0.343019
  current training accuracy:		0.896976
Seen 210000 samples, time used: 125.447s
  current training loss:		0.343381
  current training accuracy:		0.896824
saving to ..., time used: 1243.574s
Seen 213000 samples, time used: 124.524s
  current training loss:		0.343353
  current training accuracy:		0.896798
Seen 216000 samples, time used: 122.482s
  current training loss:		0.343448
  current training accuracy:		0.896630
Seen 219000 samples, time used: 127.393s
  current training loss:		0.343421
  current training accuracy:		0.896516
Seen 222000 samples, time used: 122.862s
  current training loss:		0.343459
  current training accuracy:		0.896536
Seen 225000 samples, time used: 122.488s
  current training loss:		0.343724
  current training accuracy:		0.896556
Seen 228000 samples, time used: 124.607s
  current training loss:		0.343920
  current training accuracy:		0.896447
Seen 231000 samples, time used: 124.071s
  current training loss:		0.344071
  current training accuracy:		0.896385
Seen 234000 samples, time used: 124.015s
  current training loss:		0.344295
  current training accuracy:		0.896303
Seen 237000 samples, time used: 125.638s
  current training loss:		0.344550
  current training accuracy:		0.896198
Seen 240000 samples, time used: 123.821s
  current training loss:		0.344770
  current training accuracy:		0.896092
saving to ..., time used: 1241.696s
Seen 243000 samples, time used: 125.534s
  current training loss:		0.344745
  current training accuracy:		0.896078
Seen 246000 samples, time used: 124.691s
  current training loss:		0.344933
  current training accuracy:		0.896045
Seen 249000 samples, time used: 122.767s
  current training loss:		0.345033
  current training accuracy:		0.895936
Seen 252000 samples, time used: 124.190s
  current training loss:		0.345065
  current training accuracy:		0.895960
Seen 255000 samples, time used: 124.005s
  current training loss:		0.345341
  current training accuracy:		0.895867
Seen 258000 samples, time used: 124.014s
  current training loss:		0.345579
  current training accuracy:		0.895705
Seen 261000 samples, time used: 124.193s
  current training loss:		0.345931
  current training accuracy:		0.895579
Seen 264000 samples, time used: 124.204s
  current training loss:		0.346035
  current training accuracy:		0.895462
Seen 267000 samples, time used: 123.122s
  current training loss:		0.345975
  current training accuracy:		0.895468
Seen 270000 samples, time used: 124.789s
  current training loss:		0.346141
  current training accuracy:		0.895448
saving to ..., time used: 1241.284s
Seen 273000 samples, time used: 126.235s
  current training loss:		0.346394
  current training accuracy:		0.895443
Seen 276000 samples, time used: 127.701s
  current training loss:		0.346547
  current training accuracy:		0.895409
Seen 279000 samples, time used: 124.433s
  current training loss:		0.346680
  current training accuracy:		0.895301
Seen 282000 samples, time used: 125.602s
  current training loss:		0.346889
  current training accuracy:		0.895213
Seen 285000 samples, time used: 124.970s
  current training loss:		0.346946
  current training accuracy:		0.895246
Seen 288000 samples, time used: 127.557s
  current training loss:		0.347357
  current training accuracy:		0.895184
Seen 291000 samples, time used: 125.853s
  current training loss:		0.347478
  current training accuracy:		0.895055
Seen 294000 samples, time used: 124.669s
  current training loss:		0.347554
  current training accuracy:		0.894980
Seen 297000 samples, time used: 127.113s
  current training loss:		0.347645
  current training accuracy:		0.894949
Seen 300000 samples, time used: 123.251s
  current training loss:		0.347411
  current training accuracy:		0.895017
saving to ..., time used: 1257.184s
Seen 303000 samples, time used: 124.003s
  current training loss:		0.347369
  current training accuracy:		0.895063
Seen 306000 samples, time used: 124.058s
  current training loss:		0.347430
  current training accuracy:		0.894987
Seen 309000 samples, time used: 123.191s
  current training loss:		0.347508
  current training accuracy:		0.894929
Seen 312000 samples, time used: 125.868s
  current training loss:		0.347433
  current training accuracy:		0.894923
Seen 315000 samples, time used: 127.045s
  current training loss:		0.347374
  current training accuracy:		0.894968
Seen 318000 samples, time used: 127.980s
  current training loss:		0.347289
  current training accuracy:		0.894991
Seen 321000 samples, time used: 125.539s
  current training loss:		0.347309
  current training accuracy:		0.894988
Seen 324000 samples, time used: 127.960s
  current training loss:		0.347437
  current training accuracy:		0.894985
Seen 327000 samples, time used: 124.172s
  current training loss:		0.347489
  current training accuracy:		0.894963
Seen 330000 samples, time used: 124.363s
  current training loss:		0.347554
  current training accuracy:		0.894888
saving to ..., time used: 1253.977s
Seen 333000 samples, time used: 124.116s
  current training loss:		0.347505
  current training accuracy:		0.894943
Seen 336000 samples, time used: 123.269s
  current training loss:		0.347581
  current training accuracy:		0.894881
Seen 339000 samples, time used: 121.290s
  current training loss:		0.347447
  current training accuracy:		0.894947
Seen 342000 samples, time used: 127.906s
  current training loss:		0.347658
  current training accuracy:		0.894915
Seen 345000 samples, time used: 123.372s
  current training loss:		0.347555
  current training accuracy:		0.894916
Seen 348000 samples, time used: 125.155s
  current training loss:		0.347694
  current training accuracy:		0.894856
Seen 351000 samples, time used: 125.233s
  current training loss:		0.347878
  current training accuracy:		0.894880
Seen 354000 samples, time used: 127.955s
  current training loss:		0.348021
  current training accuracy:		0.894850
Seen 357000 samples, time used: 124.595s
  current training loss:		0.348175
  current training accuracy:		0.894796
Seen 360000 samples, time used: 125.939s
  current training loss:		0.348172
  current training accuracy:		0.894831
saving to ..., time used: 1248.625s
Seen 363000 samples, time used: 127.001s
  current training loss:		0.348274
  current training accuracy:		0.894741
Seen 366000 samples, time used: 123.119s
  current training loss:		0.348342
  current training accuracy:		0.894691
Seen 369000 samples, time used: 124.653s
  current training loss:		0.348460
  current training accuracy:		0.894667
Seen 372000 samples, time used: 122.700s
  current training loss:		0.348464
  current training accuracy:		0.894618
Seen 375000 samples, time used: 124.248s
  current training loss:		0.348621
  current training accuracy:		0.894568
Seen 378000 samples, time used: 124.406s
  current training loss:		0.348676
  current training accuracy:		0.894587
Seen 381000 samples, time used: 123.375s
  current training loss:		0.348977
  current training accuracy:		0.894486
Seen 384000 samples, time used: 123.534s
  current training loss:		0.348970
  current training accuracy:		0.894487
Seen 387000 samples, time used: 124.861s
  current training loss:		0.349004
  current training accuracy:		0.894463
Seen 390000 samples, time used: 125.131s
  current training loss:		0.349089
  current training accuracy:		0.894485
saving to ..., time used: 1242.827s
Seen 393000 samples, time used: 125.181s
  current training loss:		0.349253
  current training accuracy:		0.894448
Seen 396000 samples, time used: 123.196s
  current training loss:		0.349226
  current training accuracy:		0.894419
Seen 399000 samples, time used: 129.467s
  current training loss:		0.349249
  current training accuracy:		0.894398
Seen 402000 samples, time used: 123.729s
  current training loss:		0.349224
  current training accuracy:		0.894373
Seen 405000 samples, time used: 120.830s
  current training loss:		0.349250
  current training accuracy:		0.894341
Seen 408000 samples, time used: 123.533s
  current training loss:		0.349173
  current training accuracy:		0.894355
Seen 411000 samples, time used: 128.366s
  current training loss:		0.348988
  current training accuracy:		0.894380
Seen 414000 samples, time used: 125.000s
  current training loss:		0.349008
  current training accuracy:		0.894367
Seen 417000 samples, time used: 126.402s
  current training loss:		0.348985
  current training accuracy:		0.894362
Seen 420000 samples, time used: 124.686s
  current training loss:		0.349005
  current training accuracy:		0.894348
saving to ..., time used: 1250.191s
Seen 423000 samples, time used: 125.258s
  current training loss:		0.348963
  current training accuracy:		0.894385
Seen 426000 samples, time used: 124.570s
  current training loss:		0.348976
  current training accuracy:		0.894380
Seen 429000 samples, time used: 124.636s
  current training loss:		0.349210
  current training accuracy:		0.894298
Seen 432000 samples, time used: 127.326s
  current training loss:		0.349270
  current training accuracy:		0.894326
Seen 435000 samples, time used: 126.010s
  current training loss:		0.349278
  current training accuracy:		0.894294
Seen 438000 samples, time used: 123.789s
  current training loss:		0.349355
  current training accuracy:		0.894235
Seen 441000 samples, time used: 124.437s
  current training loss:		0.349473
  current training accuracy:		0.894163
Seen 444000 samples, time used: 126.319s
  current training loss:		0.349545
  current training accuracy:		0.894113
Seen 447000 samples, time used: 125.435s
  current training loss:		0.349608
  current training accuracy:		0.894058
Seen 450000 samples, time used: 124.395s
  current training loss:		0.349598
  current training accuracy:		0.894042
saving to ..., time used: 1251.976s
Seen 453000 samples, time used: 122.914s
  current training loss:		0.349585
  current training accuracy:		0.894068
Seen 456000 samples, time used: 128.976s
  current training loss:		0.349556
  current training accuracy:		0.894011
Seen 459000 samples, time used: 128.722s
  current training loss:		0.349597
  current training accuracy:		0.893996
Seen 462000 samples, time used: 125.920s
  current training loss:		0.349693
  current training accuracy:		0.893931
Seen 465000 samples, time used: 123.282s
  current training loss:		0.349704
  current training accuracy:		0.893899
Seen 468000 samples, time used: 123.559s
  current training loss:		0.349671
  current training accuracy:		0.893878
Seen 471000 samples, time used: 126.483s
  current training loss:		0.349823
  current training accuracy:		0.893830
Seen 474000 samples, time used: 130.259s
  current training loss:		0.349781
  current training accuracy:		0.893810
Seen 477000 samples, time used: 126.086s
  current training loss:		0.349811
  current training accuracy:		0.893786
Seen 480000 samples, time used: 124.034s
  current training loss:		0.349806
  current training accuracy:		0.893750
saving to ..., time used: 1260.033s
Seen 483000 samples, time used: 126.179s
  current training loss:		0.349833
  current training accuracy:		0.893710
Seen 486000 samples, time used: 125.727s
  current training loss:		0.349993
  current training accuracy:		0.893636
Seen 489000 samples, time used: 125.681s
  current training loss:		0.350137
  current training accuracy:		0.893583
Seen 492000 samples, time used: 125.984s
  current training loss:		0.350317
  current training accuracy:		0.893500
Seen 495000 samples, time used: 123.898s
  current training loss:		0.350344
  current training accuracy:		0.893473
Seen 498000 samples, time used: 123.154s
  current training loss:		0.350505
  current training accuracy:		0.893404
Seen 501000 samples, time used: 124.318s
  current training loss:		0.350551
  current training accuracy:		0.893321
Seen 504000 samples, time used: 126.404s
  current training loss:		0.350546
  current training accuracy:		0.893327
Seen 507000 samples, time used: 126.031s
  current training loss:		0.350611
  current training accuracy:		0.893329
Seen 510000 samples, time used: 125.097s
  current training loss:		0.350639
  current training accuracy:		0.893290
saving to ..., time used: 1252.272s
Seen 513000 samples, time used: 120.707s
  current training loss:		0.350807
  current training accuracy:		0.893226
Seen 516000 samples, time used: 126.624s
  current training loss:		0.350851
  current training accuracy:		0.893178
Seen 519000 samples, time used: 125.103s
  current training loss:		0.350918
  current training accuracy:		0.893148
Seen 522000 samples, time used: 124.473s
  current training loss:		0.350799
  current training accuracy:		0.893165
Seen 525000 samples, time used: 125.784s
  current training loss:		0.350763
  current training accuracy:		0.893185
Seen 528000 samples, time used: 124.730s
  current training loss:		0.350855
  current training accuracy:		0.893148
Seen 531000 samples, time used: 130.830s
  current training loss:		0.350870
  current training accuracy:		0.893156
Seen 534000 samples, time used: 122.849s
  current training loss:		0.350993
  current training accuracy:		0.893155
Seen 537000 samples, time used: 130.295s
  current training loss:		0.351132
  current training accuracy:		0.893084
Seen 540000 samples, time used: 124.438s
  current training loss:		0.351239
  current training accuracy:		0.893046
saving to ..., time used: 1255.630s
Seen 543000 samples, time used: 125.559s
  current training loss:		0.351335
  current training accuracy:		0.892965
Seen 546000 samples, time used: 130.216s
  current training loss:		0.351424
  current training accuracy:		0.892949
Seen 549000 samples, time used: 124.115s
  current training loss:		0.351388
  current training accuracy:		0.892949
Epoch 9 of 20 took 22931.299s
  training loss:		0.351361
  training accuracy:		89.30 %
  validation loss:		0.416265
  validation accuracy:		84.85 %
Seen 3000 samples, time used: 128.045s
  current training loss:		0.316356
  current training accuracy:		0.908667
Seen 6000 samples, time used: 126.573s
  current training loss:		0.310911
  current training accuracy:		0.910500
Seen 9000 samples, time used: 128.721s
  current training loss:		0.311586
  current training accuracy:		0.910778
Seen 12000 samples, time used: 121.946s
  current training loss:		0.317131
  current training accuracy:		0.909500
Seen 15000 samples, time used: 127.102s
  current training loss:		0.319721
  current training accuracy:		0.908400
Seen 18000 samples, time used: 124.376s
  current training loss:		0.320206
  current training accuracy:		0.907667
Seen 21000 samples, time used: 123.142s
  current training loss:		0.321245
  current training accuracy:		0.907429
Seen 24000 samples, time used: 123.137s
  current training loss:		0.323211
  current training accuracy:		0.905417
Seen 27000 samples, time used: 123.436s
  current training loss:		0.321346
  current training accuracy:		0.905963
Seen 30000 samples, time used: 127.168s
  current training loss:		0.322705
  current training accuracy:		0.904933
saving to ..., time used: 1253.647s
Seen 33000 samples, time used: 126.213s
  current training loss:		0.322580
  current training accuracy:		0.905000
Seen 36000 samples, time used: 124.166s
  current training loss:		0.323161
  current training accuracy:		0.904556
Seen 39000 samples, time used: 124.819s
  current training loss:		0.323444
  current training accuracy:		0.904795
Seen 42000 samples, time used: 124.093s
  current training loss:		0.324694
  current training accuracy:		0.904452
Seen 45000 samples, time used: 127.278s
  current training loss:		0.325053
  current training accuracy:		0.904556
Seen 48000 samples, time used: 125.172s
  current training loss:		0.323565
  current training accuracy:		0.904937
Seen 51000 samples, time used: 122.790s
  current training loss:		0.325090
  current training accuracy:		0.904569
Seen 54000 samples, time used: 128.814s
  current training loss:		0.324584
  current training accuracy:		0.904759
Seen 57000 samples, time used: 126.153s
  current training loss:		0.324361
  current training accuracy:		0.905351
Seen 60000 samples, time used: 127.079s
  current training loss:		0.325354
  current training accuracy:		0.905100
saving to ..., time used: 1256.372s
Seen 63000 samples, time used: 123.940s
  current training loss:		0.325747
  current training accuracy:		0.904952
Seen 66000 samples, time used: 125.552s
  current training loss:		0.325918
  current training accuracy:		0.904955
Seen 69000 samples, time used: 124.030s
  current training loss:		0.326308
  current training accuracy:		0.904870
Seen 72000 samples, time used: 123.811s
  current training loss:		0.326914
  current training accuracy:		0.904736
Seen 75000 samples, time used: 126.042s
  current training loss:		0.327419
  current training accuracy:		0.905000
Seen 78000 samples, time used: 125.113s
  current training loss:		0.327532
  current training accuracy:		0.905000
Seen 81000 samples, time used: 126.145s
  current training loss:		0.326326
  current training accuracy:		0.905395
Seen 84000 samples, time used: 129.517s
  current training loss:		0.326479
  current training accuracy:		0.905274
Seen 87000 samples, time used: 123.752s
  current training loss:		0.326818
  current training accuracy:		0.905207
Seen 90000 samples, time used: 122.575s
  current training loss:		0.326597
  current training accuracy:		0.905289
saving to ..., time used: 1250.247s
Seen 93000 samples, time used: 129.647s
  current training loss:		0.326090
  current training accuracy:		0.905355
Seen 96000 samples, time used: 127.584s
  current training loss:		0.326554
  current training accuracy:		0.904948
Seen 99000 samples, time used: 128.700s
  current training loss:		0.326889
  current training accuracy:		0.904808
Seen 102000 samples, time used: 124.650s
  current training loss:		0.327890
  current training accuracy:		0.904529
Seen 105000 samples, time used: 124.541s
  current training loss:		0.327825
  current training accuracy:		0.904743
Seen 108000 samples, time used: 124.555s
  current training loss:		0.327772
  current training accuracy:		0.904722
Seen 111000 samples, time used: 122.711s
  current training loss:		0.327599
  current training accuracy:		0.904883
Seen 114000 samples, time used: 126.053s
  current training loss:		0.327570
  current training accuracy:		0.904842
Seen 117000 samples, time used: 124.616s
  current training loss:		0.327716
  current training accuracy:		0.904641
Seen 120000 samples, time used: 124.379s
  current training loss:		0.327888
  current training accuracy:		0.904642
saving to ..., time used: 1257.234s
Seen 123000 samples, time used: 124.144s
  current training loss:		0.327698
  current training accuracy:		0.904569
Seen 126000 samples, time used: 124.666s
  current training loss:		0.327407
  current training accuracy:		0.904762
Seen 129000 samples, time used: 126.565s
  current training loss:		0.326798
  current training accuracy:		0.904876
Seen 132000 samples, time used: 126.026s
  current training loss:		0.327043
  current training accuracy:		0.904697
Seen 135000 samples, time used: 125.304s
  current training loss:		0.326815
  current training accuracy:		0.904741
Seen 138000 samples, time used: 122.272s
  current training loss:		0.327186
  current training accuracy:		0.904703
Seen 141000 samples, time used: 125.060s
  current training loss:		0.326783
  current training accuracy:		0.904794
Seen 144000 samples, time used: 121.394s
  current training loss:		0.326783
  current training accuracy:		0.904910
Seen 147000 samples, time used: 126.634s
  current training loss:		0.326840
  current training accuracy:		0.904782
Seen 150000 samples, time used: 128.165s
  current training loss:		0.327098
  current training accuracy:		0.904647
saving to ..., time used: 1250.026s
Seen 153000 samples, time used: 128.690s
  current training loss:		0.327019
  current training accuracy:		0.904654
Seen 156000 samples, time used: 122.882s
  current training loss:		0.327108
  current training accuracy:		0.904590
Seen 159000 samples, time used: 126.272s
  current training loss:		0.327420
  current training accuracy:		0.904497
Seen 162000 samples, time used: 123.943s
  current training loss:		0.327412
  current training accuracy:		0.904494
Seen 165000 samples, time used: 123.603s
  current training loss:		0.327512
  current training accuracy:		0.904370
Seen 168000 samples, time used: 126.185s
  current training loss:		0.327684
  current training accuracy:		0.904321
Seen 171000 samples, time used: 124.027s
  current training loss:		0.327928
  current training accuracy:		0.904216
Seen 174000 samples, time used: 125.235s
  current training loss:		0.328207
  current training accuracy:		0.904098
Seen 177000 samples, time used: 123.765s
  current training loss:		0.328620
  current training accuracy:		0.904045
Seen 180000 samples, time used: 124.233s
  current training loss:		0.328870
  current training accuracy:		0.903917
saving to ..., time used: 1248.634s
Seen 183000 samples, time used: 125.904s
  current training loss:		0.328664
  current training accuracy:		0.904033
Seen 186000 samples, time used: 128.437s
  current training loss:		0.328582
  current training accuracy:		0.903925
Seen 189000 samples, time used: 125.745s
  current training loss:		0.328510
  current training accuracy:		0.903952
Seen 192000 samples, time used: 126.009s
  current training loss:		0.328645
  current training accuracy:		0.903922
Seen 195000 samples, time used: 127.146s
  current training loss:		0.328444
  current training accuracy:		0.903836
Seen 198000 samples, time used: 127.646s
  current training loss:		0.328835
  current training accuracy:		0.903747
Seen 201000 samples, time used: 124.244s
  current training loss:		0.328918
  current training accuracy:		0.903786
Seen 204000 samples, time used: 122.775s
  current training loss:		0.329286
  current training accuracy:		0.903706
Seen 207000 samples, time used: 125.598s
  current training loss:		0.329403
  current training accuracy:		0.903725
Seen 210000 samples, time used: 125.242s
  current training loss:		0.329589
  current training accuracy:		0.903657
saving to ..., time used: 1258.542s
Seen 213000 samples, time used: 129.490s
  current training loss:		0.329405
  current training accuracy:		0.903671
Seen 216000 samples, time used: 128.999s
  current training loss:		0.329721
  current training accuracy:		0.903519
Seen 219000 samples, time used: 125.346s
  current training loss:		0.329851
  current training accuracy:		0.903502
Seen 222000 samples, time used: 123.069s
  current training loss:		0.329881
  current training accuracy:		0.903473
Seen 225000 samples, time used: 124.916s
  current training loss:		0.329745
  current training accuracy:		0.903507
Seen 228000 samples, time used: 127.834s
  current training loss:		0.330130
  current training accuracy:		0.903395
Seen 231000 samples, time used: 127.252s
  current training loss:		0.330346
  current training accuracy:		0.903351
Seen 234000 samples, time used: 125.162s
  current training loss:		0.330388
  current training accuracy:		0.903342
Seen 237000 samples, time used: 123.635s
  current training loss:		0.330287
  current training accuracy:		0.903342
Seen 240000 samples, time used: 127.503s
  current training loss:		0.330450
  current training accuracy:		0.903283
saving to ..., time used: 1263.004s
Seen 243000 samples, time used: 125.209s
  current training loss:		0.330598
  current training accuracy:		0.903119
Seen 246000 samples, time used: 126.368s
  current training loss:		0.330631
  current training accuracy:		0.903122
Seen 249000 samples, time used: 123.828s
  current training loss:		0.330724
  current training accuracy:		0.903020
Seen 252000 samples, time used: 126.911s
  current training loss:		0.330883
  current training accuracy:		0.902964
Seen 255000 samples, time used: 125.940s
  current training loss:		0.330932
  current training accuracy:		0.902922
Seen 258000 samples, time used: 124.185s
  current training loss:		0.331346
  current training accuracy:		0.902849
Seen 261000 samples, time used: 127.010s
  current training loss:		0.331291
  current training accuracy:		0.902828
Seen 264000 samples, time used: 125.147s
  current training loss:		0.331406
  current training accuracy:		0.902723
Seen 267000 samples, time used: 123.553s
  current training loss:		0.331571
  current training accuracy:		0.902689
Seen 270000 samples, time used: 126.897s
  current training loss:		0.331581
  current training accuracy:		0.902678
saving to ..., time used: 1254.848s
Seen 273000 samples, time used: 126.582s
  current training loss:		0.331740
  current training accuracy:		0.902575
Seen 276000 samples, time used: 126.916s
  current training loss:		0.331827
  current training accuracy:		0.902576
Seen 279000 samples, time used: 127.681s
  current training loss:		0.331905
  current training accuracy:		0.902609
Seen 282000 samples, time used: 123.793s
  current training loss:		0.331943
  current training accuracy:		0.902514
Seen 285000 samples, time used: 126.000s
  current training loss:		0.332272
  current training accuracy:		0.902439
Seen 288000 samples, time used: 123.701s
  current training loss:		0.332347
  current training accuracy:		0.902469
Seen 291000 samples, time used: 121.119s
  current training loss:		0.332517
  current training accuracy:		0.902440
Seen 294000 samples, time used: 128.516s
  current training loss:		0.332565
  current training accuracy:		0.902388
Seen 297000 samples, time used: 123.085s
  current training loss:		0.332588
  current training accuracy:		0.902354
Seen 300000 samples, time used: 122.572s
  current training loss:		0.332626
  current training accuracy:		0.902317
saving to ..., time used: 1249.765s
Seen 303000 samples, time used: 128.793s
  current training loss:		0.332657
  current training accuracy:		0.902224
Seen 306000 samples, time used: 120.720s
  current training loss:		0.333024
  current training accuracy:		0.902052
Seen 309000 samples, time used: 124.598s
  current training loss:		0.333142
  current training accuracy:		0.901971
Seen 312000 samples, time used: 127.053s
  current training loss:		0.333130
  current training accuracy:		0.901952
Seen 315000 samples, time used: 125.771s
  current training loss:		0.333332
  current training accuracy:		0.901902
Seen 318000 samples, time used: 126.225s
  current training loss:		0.333576
  current training accuracy:		0.901777
Seen 321000 samples, time used: 123.489s
  current training loss:		0.333630
  current training accuracy:		0.901701
Seen 324000 samples, time used: 125.072s
  current training loss:		0.333921
  current training accuracy:		0.901577
Seen 327000 samples, time used: 124.863s
  current training loss:		0.334195
  current training accuracy:		0.901489
Seen 330000 samples, time used: 123.889s
  current training loss:		0.334336
  current training accuracy:		0.901409
saving to ..., time used: 1250.275s
Seen 333000 samples, time used: 127.476s
  current training loss:		0.334556
  current training accuracy:		0.901285
Seen 336000 samples, time used: 122.828s
  current training loss:		0.334468
  current training accuracy:		0.901286
Seen 339000 samples, time used: 127.515s
  current training loss:		0.334441
  current training accuracy:		0.901280
Seen 342000 samples, time used: 124.929s
  current training loss:		0.334443
  current training accuracy:		0.901216
Seen 345000 samples, time used: 122.478s
  current training loss:		0.334374
  current training accuracy:		0.901232
Seen 348000 samples, time used: 127.290s
  current training loss:		0.334543
  current training accuracy:		0.901109
Seen 351000 samples, time used: 125.866s
  current training loss:		0.334448
  current training accuracy:		0.901111
Seen 354000 samples, time used: 124.386s
  current training loss:		0.334533
  current training accuracy:		0.901090
Seen 357000 samples, time used: 124.959s
  current training loss:		0.334664
  current training accuracy:		0.901022
Seen 360000 samples, time used: 124.298s
  current training loss:		0.334858
  current training accuracy:		0.900950
saving to ..., time used: 1251.824s
Seen 363000 samples, time used: 126.338s
  current training loss:		0.335022
  current training accuracy:		0.900832
Seen 366000 samples, time used: 124.807s
  current training loss:		0.335149
  current training accuracy:		0.900798
Seen 369000 samples, time used: 124.349s
  current training loss:		0.335309
  current training accuracy:		0.900729
Seen 372000 samples, time used: 123.014s
  current training loss:		0.335533
  current training accuracy:		0.900613
Seen 375000 samples, time used: 126.772s
  current training loss:		0.335375
  current training accuracy:		0.900645
Seen 378000 samples, time used: 123.184s
  current training loss:		0.335594
  current training accuracy:		0.900579
Seen 381000 samples, time used: 124.381s
  current training loss:		0.335560
  current training accuracy:		0.900601
Seen 384000 samples, time used: 125.186s
  current training loss:		0.335792
  current training accuracy:		0.900531
Seen 387000 samples, time used: 124.514s
  current training loss:		0.335869
  current training accuracy:		0.900460
Seen 390000 samples, time used: 123.543s
  current training loss:		0.335905
  current training accuracy:		0.900410
saving to ..., time used: 1245.887s
Seen 393000 samples, time used: 124.367s
  current training loss:		0.336131
  current training accuracy:		0.900331
Seen 396000 samples, time used: 122.485s
  current training loss:		0.336136
  current training accuracy:		0.900348
Seen 399000 samples, time used: 122.990s
  current training loss:		0.336092
  current training accuracy:		0.900346
Seen 402000 samples, time used: 125.961s
  current training loss:		0.336059
  current training accuracy:		0.900381
Seen 405000 samples, time used: 126.943s
  current training loss:		0.336249
  current training accuracy:		0.900294
Seen 408000 samples, time used: 124.827s
  current training loss:		0.336232
  current training accuracy:		0.900287
Seen 411000 samples, time used: 124.545s
  current training loss:		0.336422
  current training accuracy:		0.900204
Seen 414000 samples, time used: 123.950s
  current training loss:		0.336498
  current training accuracy:		0.900171
Seen 417000 samples, time used: 121.473s
  current training loss:		0.336556
  current training accuracy:		0.900139
Seen 420000 samples, time used: 124.630s
  current training loss:		0.336794
  current training accuracy:		0.900048
saving to ..., time used: 1241.969s
Seen 423000 samples, time used: 123.627s
  current training loss:		0.336850
  current training accuracy:		0.900050
Seen 426000 samples, time used: 128.293s
  current training loss:		0.336805
  current training accuracy:		0.900047
Seen 429000 samples, time used: 122.755s
  current training loss:		0.336723
  current training accuracy:		0.900030
Seen 432000 samples, time used: 125.355s
  current training loss:		0.336773
  current training accuracy:		0.900019
Seen 435000 samples, time used: 127.258s
  current training loss:		0.336797
  current training accuracy:		0.899968
Seen 438000 samples, time used: 124.831s
  current training loss:		0.336895
  current training accuracy:		0.899966
Seen 441000 samples, time used: 123.875s
  current training loss:		0.336827
  current training accuracy:		0.900005
Seen 444000 samples, time used: 127.942s
  current training loss:		0.336984
  current training accuracy:		0.899926
Seen 447000 samples, time used: 124.457s
  current training loss:		0.337064
  current training accuracy:		0.899904
Seen 450000 samples, time used: 123.170s
  current training loss:		0.337209
  current training accuracy:		0.899860
saving to ..., time used: 1251.362s
Seen 453000 samples, time used: 124.404s
  current training loss:		0.337395
  current training accuracy:		0.899766
Seen 456000 samples, time used: 121.595s
  current training loss:		0.337574
  current training accuracy:		0.899717
Seen 459000 samples, time used: 123.432s
  current training loss:		0.337489
  current training accuracy:		0.899736
Seen 462000 samples, time used: 125.124s
  current training loss:		0.337687
  current training accuracy:		0.899632
Seen 465000 samples, time used: 125.854s
  current training loss:		0.337740
  current training accuracy:		0.899587
Seen 468000 samples, time used: 125.697s
  current training loss:		0.337649
  current training accuracy:		0.899620
Seen 471000 samples, time used: 124.768s
  current training loss:		0.337707
  current training accuracy:		0.899582
Seen 474000 samples, time used: 126.535s
  current training loss:		0.337814
  current training accuracy:		0.899549
Seen 477000 samples, time used: 127.079s
  current training loss:		0.337901
  current training accuracy:		0.899488
Seen 480000 samples, time used: 124.583s
  current training loss:		0.337925
  current training accuracy:		0.899506
saving to ..., time used: 1248.868s
Seen 483000 samples, time used: 123.437s
  current training loss:		0.337953
  current training accuracy:		0.899491
Seen 486000 samples, time used: 123.870s
  current training loss:		0.337936
  current training accuracy:		0.899486
Seen 489000 samples, time used: 123.131s
  current training loss:		0.338101
  current training accuracy:		0.899448
Seen 492000 samples, time used: 126.014s
  current training loss:		0.338418
  current training accuracy:		0.899270
Seen 495000 samples, time used: 126.439s
  current training loss:		0.338550
  current training accuracy:		0.899271
Seen 498000 samples, time used: 125.897s
  current training loss:		0.338660
  current training accuracy:		0.899193
Seen 501000 samples, time used: 121.975s
  current training loss:		0.338601
  current training accuracy:		0.899212
Seen 504000 samples, time used: 126.143s
  current training loss:		0.338672
  current training accuracy:		0.899165
Seen 507000 samples, time used: 128.668s
  current training loss:		0.338716
  current training accuracy:		0.899128
Seen 510000 samples, time used: 124.580s
  current training loss:		0.338816
  current training accuracy:		0.899080
saving to ..., time used: 1249.950s
Seen 513000 samples, time used: 124.645s
  current training loss:		0.338780
  current training accuracy:		0.899041
Seen 516000 samples, time used: 126.011s
  current training loss:		0.338795
  current training accuracy:		0.898973
Seen 519000 samples, time used: 123.000s
  current training loss:		0.338885
  current training accuracy:		0.898925
Seen 522000 samples, time used: 125.938s
  current training loss:		0.338995
  current training accuracy:		0.898835
Seen 525000 samples, time used: 124.524s
  current training loss:		0.338981
  current training accuracy:		0.898846
Seen 528000 samples, time used: 126.389s
  current training loss:		0.339120
  current training accuracy:		0.898780
Seen 531000 samples, time used: 125.464s
  current training loss:		0.339109
  current training accuracy:		0.898761
Seen 534000 samples, time used: 125.677s
  current training loss:		0.339243
  current training accuracy:		0.898700
Seen 537000 samples, time used: 123.871s
  current training loss:		0.339303
  current training accuracy:		0.898693
Seen 540000 samples, time used: 125.442s
  current training loss:		0.339218
  current training accuracy:		0.898709
saving to ..., time used: 1250.761s
Seen 543000 samples, time used: 124.504s
  current training loss:		0.339261
  current training accuracy:		0.898670
Seen 546000 samples, time used: 126.888s
  current training loss:		0.339287
  current training accuracy:		0.898692
Seen 549000 samples, time used: 124.106s
  current training loss:		0.339404
  current training accuracy:		0.898668
Epoch 10 of 20 took 22956.534s
  training loss:		0.339479
  training accuracy:		89.87 %
  validation loss:		0.418939
  validation accuracy:		84.77 %
Seen 3000 samples, time used: 123.374s
  current training loss:		0.312605
  current training accuracy:		0.911000
Seen 6000 samples, time used: 125.099s
  current training loss:		0.308263
  current training accuracy:		0.913500
Seen 9000 samples, time used: 124.449s
  current training loss:		0.314952
  current training accuracy:		0.911111
Seen 12000 samples, time used: 122.540s
  current training loss:		0.317160
  current training accuracy:		0.911667
Seen 15000 samples, time used: 127.195s
  current training loss:		0.317937
  current training accuracy:		0.909933
Seen 18000 samples, time used: 123.102s
  current training loss:		0.320150
  current training accuracy:		0.909278
Seen 21000 samples, time used: 124.039s
  current training loss:		0.318596
  current training accuracy:		0.909333
Seen 24000 samples, time used: 128.796s
  current training loss:		0.319100
  current training accuracy:		0.909458
Seen 27000 samples, time used: 123.645s
  current training loss:		0.318745
  current training accuracy:		0.909259
Seen 30000 samples, time used: 124.506s
  current training loss:		0.320939
  current training accuracy:		0.908400
saving to ..., time used: 1246.745s
Seen 33000 samples, time used: 122.622s
  current training loss:		0.318992
  current training accuracy:		0.908909
Seen 36000 samples, time used: 124.774s
  current training loss:		0.317164
  current training accuracy:		0.909639
Seen 39000 samples, time used: 126.591s
  current training loss:		0.316021
  current training accuracy:		0.910077
Seen 42000 samples, time used: 122.871s
  current training loss:		0.314773
  current training accuracy:		0.910214
Seen 45000 samples, time used: 127.295s
  current training loss:		0.317436
  current training accuracy:		0.909600
Seen 48000 samples, time used: 128.698s
  current training loss:		0.317928
  current training accuracy:		0.909458
Seen 51000 samples, time used: 124.340s
  current training loss:		0.317139
  current training accuracy:		0.910235
Seen 54000 samples, time used: 124.270s
  current training loss:		0.316798
  current training accuracy:		0.910722
Seen 57000 samples, time used: 124.132s
  current training loss:		0.316229
  current training accuracy:		0.910509
Seen 60000 samples, time used: 126.617s
  current training loss:		0.316866
  current training accuracy:		0.910617
saving to ..., time used: 1252.010s
Seen 63000 samples, time used: 127.400s
  current training loss:		0.316822
  current training accuracy:		0.910397
Seen 66000 samples, time used: 123.974s
  current training loss:		0.316634
  current training accuracy:		0.910288
Seen 69000 samples, time used: 127.274s
  current training loss:		0.316573
  current training accuracy:		0.910522
Seen 72000 samples, time used: 126.209s
  current training loss:		0.316132
  current training accuracy:		0.910986
Seen 75000 samples, time used: 127.964s
  current training loss:		0.316062
  current training accuracy:		0.911040
Seen 78000 samples, time used: 124.166s
  current training loss:		0.315977
  current training accuracy:		0.910949
Seen 81000 samples, time used: 127.153s
  current training loss:		0.316340
  current training accuracy:		0.910839
Seen 84000 samples, time used: 123.281s
  current training loss:		0.316971
  current training accuracy:		0.910667
Seen 87000 samples, time used: 123.079s
  current training loss:		0.317381
  current training accuracy:		0.910586
Seen 90000 samples, time used: 127.485s
  current training loss:		0.317369
  current training accuracy:		0.910644
saving to ..., time used: 1257.785s
Seen 93000 samples, time used: 121.863s
  current training loss:		0.317307
  current training accuracy:		0.910817
Seen 96000 samples, time used: 121.394s
  current training loss:		0.316774
  current training accuracy:		0.910865
Seen 99000 samples, time used: 124.797s
  current training loss:		0.316568
  current training accuracy:		0.910879
Seen 102000 samples, time used: 125.616s
  current training loss:		0.316897
  current training accuracy:		0.910873
Seen 105000 samples, time used: 122.797s
  current training loss:		0.316886
  current training accuracy:		0.910810
Seen 108000 samples, time used: 124.441s
  current training loss:		0.317007
  current training accuracy:		0.910759
Seen 111000 samples, time used: 125.972s
  current training loss:		0.317340
  current training accuracy:		0.910667
Seen 114000 samples, time used: 127.380s
  current training loss:		0.317607
  current training accuracy:		0.910430
Seen 117000 samples, time used: 125.972s
  current training loss:		0.317512
  current training accuracy:		0.910299
Seen 120000 samples, time used: 125.989s
  current training loss:		0.317245
  current training accuracy:		0.910467
saving to ..., time used: 1246.022s
Seen 123000 samples, time used: 125.654s
  current training loss:		0.317109
  current training accuracy:		0.910398
Seen 126000 samples, time used: 127.575s
  current training loss:		0.316973
  current training accuracy:		0.910516
Seen 129000 samples, time used: 128.761s
  current training loss:		0.316982
  current training accuracy:		0.910496
Seen 132000 samples, time used: 123.256s
  current training loss:		0.317605
  current training accuracy:		0.910235
Seen 135000 samples, time used: 123.234s
  current training loss:		0.317424
  current training accuracy:		0.910163
Seen 138000 samples, time used: 128.985s
  current training loss:		0.318279
  current training accuracy:		0.909841
Seen 141000 samples, time used: 125.492s
  current training loss:		0.318226
  current training accuracy:		0.909837
Seen 144000 samples, time used: 128.532s
  current training loss:		0.318164
  current training accuracy:		0.909812
Seen 147000 samples, time used: 125.199s
  current training loss:		0.317971
  current training accuracy:		0.909755
Seen 150000 samples, time used: 123.846s
  current training loss:		0.318198
  current training accuracy:		0.909733
saving to ..., time used: 1260.334s
Seen 153000 samples, time used: 126.699s
  current training loss:		0.318602
  current training accuracy:		0.909490
Seen 156000 samples, time used: 124.730s
  current training loss:		0.318510
  current training accuracy:		0.909365
Seen 159000 samples, time used: 122.759s
  current training loss:		0.318486
  current training accuracy:		0.909371
Seen 162000 samples, time used: 127.456s
  current training loss:		0.318478
  current training accuracy:		0.909377
Seen 165000 samples, time used: 124.221s
  current training loss:		0.318784
  current training accuracy:		0.909164
Seen 168000 samples, time used: 120.683s
  current training loss:		0.318585
  current training accuracy:		0.909190
Seen 171000 samples, time used: 122.749s
  current training loss:		0.318472
  current training accuracy:		0.909175
Seen 174000 samples, time used: 127.943s
  current training loss:		0.318970
  current training accuracy:		0.908971
Seen 177000 samples, time used: 124.117s
  current training loss:		0.318868
  current training accuracy:		0.909023
Seen 180000 samples, time used: 124.688s
  current training loss:		0.319219
  current training accuracy:		0.908883
saving to ..., time used: 1245.846s
Seen 183000 samples, time used: 125.901s
  current training loss:		0.319254
  current training accuracy:		0.908858
Seen 186000 samples, time used: 123.398s
  current training loss:		0.319679
  current training accuracy:		0.908855
Seen 189000 samples, time used: 124.747s
  current training loss:		0.319870
  current training accuracy:		0.908704
Seen 192000 samples, time used: 126.409s
  current training loss:		0.319664
  current training accuracy:		0.908823
Seen 195000 samples, time used: 125.819s
  current training loss:		0.319785
  current training accuracy:		0.908810
Seen 198000 samples, time used: 126.019s
  current training loss:		0.319877
  current training accuracy:		0.908672
Seen 201000 samples, time used: 124.924s
  current training loss:		0.320077
  current training accuracy:		0.908532
Seen 204000 samples, time used: 124.753s
  current training loss:		0.320198
  current training accuracy:		0.908490
Seen 207000 samples, time used: 124.966s
  current training loss:		0.320279
  current training accuracy:		0.908440
Seen 210000 samples, time used: 126.271s
  current training loss:		0.320597
  current training accuracy:		0.908205
saving to ..., time used: 1253.005s
Seen 213000 samples, time used: 125.257s
  current training loss:		0.320780
  current training accuracy:		0.908202
Seen 216000 samples, time used: 122.895s
  current training loss:		0.320829
  current training accuracy:		0.908167
Seen 219000 samples, time used: 126.189s
  current training loss:		0.321077
  current training accuracy:		0.908160
Seen 222000 samples, time used: 124.977s
  current training loss:		0.321092
  current training accuracy:		0.908135
Seen 225000 samples, time used: 124.747s
  current training loss:		0.321011
  current training accuracy:		0.908111
Seen 228000 samples, time used: 122.627s
  current training loss:		0.321025
  current training accuracy:		0.908136
Seen 231000 samples, time used: 125.158s
  current training loss:		0.320857
  current training accuracy:		0.908134
Seen 234000 samples, time used: 125.111s
  current training loss:		0.320943
  current training accuracy:		0.908073
Seen 237000 samples, time used: 125.235s
  current training loss:		0.320835
  current training accuracy:		0.908131
Seen 240000 samples, time used: 125.467s
  current training loss:		0.320987
  current training accuracy:		0.908079
saving to ..., time used: 1247.463s
Seen 243000 samples, time used: 126.611s
  current training loss:		0.320648
  current training accuracy:		0.908058
Seen 246000 samples, time used: 125.873s
  current training loss:		0.320836
  current training accuracy:		0.907927
Seen 249000 samples, time used: 125.003s
  current training loss:		0.321001
  current training accuracy:		0.907827
Seen 252000 samples, time used: 124.836s
  current training loss:		0.321337
  current training accuracy:		0.907714
Seen 255000 samples, time used: 122.721s
  current training loss:		0.321505
  current training accuracy:		0.907690
Seen 258000 samples, time used: 124.117s
  current training loss:		0.321821
  current training accuracy:		0.907601
Seen 261000 samples, time used: 124.153s
  current training loss:		0.322082
  current training accuracy:		0.907498
Seen 264000 samples, time used: 125.158s
  current training loss:		0.322434
  current training accuracy:		0.907352
Seen 267000 samples, time used: 129.831s
  current training loss:		0.322577
  current training accuracy:		0.907303
Seen 270000 samples, time used: 124.482s
  current training loss:		0.322717
  current training accuracy:		0.907270
saving to ..., time used: 1252.584s
Seen 273000 samples, time used: 128.680s
  current training loss:		0.323047
  current training accuracy:		0.907147
Seen 276000 samples, time used: 126.350s
  current training loss:		0.323182
  current training accuracy:		0.907083
Seen 279000 samples, time used: 122.581s
  current training loss:		0.323121
  current training accuracy:		0.907104
Seen 282000 samples, time used: 124.046s
  current training loss:		0.323423
  current training accuracy:		0.907014
Seen 285000 samples, time used: 126.447s
  current training loss:		0.323405
  current training accuracy:		0.906965
Seen 288000 samples, time used: 125.562s
  current training loss:		0.323483
  current training accuracy:		0.907035
Seen 291000 samples, time used: 123.453s
  current training loss:		0.323657
  current training accuracy:		0.906931
Seen 294000 samples, time used: 128.367s
  current training loss:		0.323708
  current training accuracy:		0.906901
Seen 297000 samples, time used: 124.631s
  current training loss:		0.323714
  current training accuracy:		0.906879
Seen 300000 samples, time used: 121.571s
  current training loss:		0.324050
  current training accuracy:		0.906737
saving to ..., time used: 1251.488s
Seen 303000 samples, time used: 123.324s
  current training loss:		0.324097
  current training accuracy:		0.906686
Seen 306000 samples, time used: 125.494s
  current training loss:		0.324178
  current training accuracy:		0.906654
Seen 309000 samples, time used: 124.726s
  current training loss:		0.324291
  current training accuracy:		0.906605
Seen 312000 samples, time used: 125.687s
  current training loss:		0.324316
  current training accuracy:		0.906567
Seen 315000 samples, time used: 124.214s
  current training loss:		0.324469
  current training accuracy:		0.906517
Seen 318000 samples, time used: 123.344s
  current training loss:		0.324401
  current training accuracy:		0.906531
Seen 321000 samples, time used: 125.412s
  current training loss:		0.324436
  current training accuracy:		0.906523
Seen 324000 samples, time used: 126.546s
  current training loss:		0.324686
  current training accuracy:		0.906444
Seen 327000 samples, time used: 123.601s
  current training loss:		0.324867
  current training accuracy:		0.906370
Seen 330000 samples, time used: 124.035s
  current training loss:		0.324928
  current training accuracy:		0.906364
saving to ..., time used: 1246.181s
Seen 333000 samples, time used: 126.986s
  current training loss:		0.324950
  current training accuracy:		0.906336
Seen 336000 samples, time used: 123.158s
  current training loss:		0.325217
  current training accuracy:		0.906244
Seen 339000 samples, time used: 124.927s
  current training loss:		0.325394
  current training accuracy:		0.906189
Seen 342000 samples, time used: 126.110s
  current training loss:		0.325399
  current training accuracy:		0.906231
Seen 345000 samples, time used: 123.839s
  current training loss:		0.325339
  current training accuracy:		0.906241
Seen 348000 samples, time used: 123.670s
  current training loss:		0.325477
  current training accuracy:		0.906224
Seen 351000 samples, time used: 127.274s
  current training loss:		0.325615
  current training accuracy:		0.906179
Seen 354000 samples, time used: 124.870s
  current training loss:		0.325781
  current training accuracy:		0.906105
Seen 357000 samples, time used: 125.362s
  current training loss:		0.325797
  current training accuracy:		0.906137
Seen 360000 samples, time used: 121.872s
  current training loss:		0.325925
  current training accuracy:		0.906111
saving to ..., time used: 1247.867s
Seen 363000 samples, time used: 128.125s
  current training loss:		0.325977
  current training accuracy:		0.906099
Seen 366000 samples, time used: 127.922s
  current training loss:		0.326116
  current training accuracy:		0.906025
Seen 369000 samples, time used: 121.574s
  current training loss:		0.326189
  current training accuracy:		0.905900
Seen 372000 samples, time used: 126.020s
  current training loss:		0.326314
  current training accuracy:		0.905806
Seen 375000 samples, time used: 127.901s
  current training loss:		0.326194
  current training accuracy:		0.905795
Seen 378000 samples, time used: 126.827s
  current training loss:		0.326306
  current training accuracy:		0.905759
Seen 381000 samples, time used: 123.008s
  current training loss:		0.326553
  current training accuracy:		0.905661
Seen 384000 samples, time used: 126.705s
  current training loss:		0.326766
  current training accuracy:		0.905536
Seen 387000 samples, time used: 124.266s
  current training loss:		0.326848
  current training accuracy:		0.905470
Seen 390000 samples, time used: 125.439s
  current training loss:		0.326758
  current training accuracy:		0.905533
saving to ..., time used: 1257.586s
Seen 393000 samples, time used: 125.992s
  current training loss:		0.326974
  current training accuracy:		0.905501
Seen 396000 samples, time used: 127.565s
  current training loss:		0.327130
  current training accuracy:		0.905449
Seen 399000 samples, time used: 122.824s
  current training loss:		0.327279
  current training accuracy:		0.905363
Seen 402000 samples, time used: 125.023s
  current training loss:		0.327387
  current training accuracy:		0.905306
Seen 405000 samples, time used: 127.258s
  current training loss:		0.327448
  current training accuracy:		0.905267
Seen 408000 samples, time used: 124.032s
  current training loss:		0.327461
  current training accuracy:		0.905260
Seen 411000 samples, time used: 128.008s
  current training loss:		0.327602
  current training accuracy:		0.905168
Seen 414000 samples, time used: 124.177s
  current training loss:		0.327672
  current training accuracy:		0.905109
Seen 417000 samples, time used: 123.564s
  current training loss:		0.327574
  current training accuracy:		0.905115
Seen 420000 samples, time used: 123.096s
  current training loss:		0.327542
  current training accuracy:		0.905093
saving to ..., time used: 1251.338s
Seen 423000 samples, time used: 127.923s
  current training loss:		0.327560
  current training accuracy:		0.905130
Seen 426000 samples, time used: 126.659s
  current training loss:		0.327631
  current training accuracy:		0.905138
Seen 429000 samples, time used: 131.777s
  current training loss:		0.327796
  current training accuracy:		0.905093
Seen 432000 samples, time used: 125.309s
  current training loss:		0.327756
  current training accuracy:		0.905072
Seen 435000 samples, time used: 125.121s
  current training loss:		0.327844
  current training accuracy:		0.905039
Seen 438000 samples, time used: 128.950s
  current training loss:		0.327904
  current training accuracy:		0.904986
Seen 441000 samples, time used: 125.423s
  current training loss:		0.327937
  current training accuracy:		0.904986
Seen 444000 samples, time used: 124.763s
  current training loss:		0.328004
  current training accuracy:		0.904984
Seen 447000 samples, time used: 125.924s
  current training loss:		0.328086
  current training accuracy:		0.904940
Seen 450000 samples, time used: 126.448s
  current training loss:		0.328280
  current training accuracy:		0.904860
saving to ..., time used: 1268.096s
Seen 453000 samples, time used: 124.095s
  current training loss:		0.328405
  current training accuracy:		0.904815
Seen 456000 samples, time used: 126.366s
  current training loss:		0.328468
  current training accuracy:		0.904778
Seen 459000 samples, time used: 123.422s
  current training loss:		0.328493
  current training accuracy:		0.904810
Seen 462000 samples, time used: 125.807s
  current training loss:		0.328574
  current training accuracy:		0.904758
Seen 465000 samples, time used: 125.904s
  current training loss:		0.328817
  current training accuracy:		0.904649
Seen 468000 samples, time used: 126.948s
  current training loss:		0.328853
  current training accuracy:		0.904656
Seen 471000 samples, time used: 125.278s
  current training loss:		0.328880
  current training accuracy:		0.904639
Seen 474000 samples, time used: 128.635s
  current training loss:		0.328693
  current training accuracy:		0.904690
Seen 477000 samples, time used: 123.863s
  current training loss:		0.328740
  current training accuracy:		0.904677
Seen 480000 samples, time used: 124.073s
  current training loss:		0.328892
  current training accuracy:		0.904623
saving to ..., time used: 1254.190s
Seen 483000 samples, time used: 123.408s
  current training loss:		0.328923
  current training accuracy:		0.904607
Seen 486000 samples, time used: 122.856s
  current training loss:		0.328984
  current training accuracy:		0.904564
Seen 489000 samples, time used: 126.946s
  current training loss:		0.328990
  current training accuracy:		0.904587
Seen 492000 samples, time used: 124.301s
  current training loss:		0.329025
  current training accuracy:		0.904585
Seen 495000 samples, time used: 126.164s
  current training loss:		0.329086
  current training accuracy:		0.904549
Seen 498000 samples, time used: 122.807s
  current training loss:		0.329102
  current training accuracy:		0.904546
Seen 501000 samples, time used: 123.325s
  current training loss:		0.329241
  current training accuracy:		0.904499
Seen 504000 samples, time used: 127.142s
  current training loss:		0.329388
  current training accuracy:		0.904413
Seen 507000 samples, time used: 126.330s
  current training loss:		0.329430
  current training accuracy:		0.904371
Seen 510000 samples, time used: 125.843s
  current training loss:		0.329504
  current training accuracy:		0.904337
saving to ..., time used: 1248.921s
Seen 513000 samples, time used: 127.396s
  current training loss:		0.329500
  current training accuracy:		0.904347
Seen 516000 samples, time used: 126.499s
  current training loss:		0.329574
  current training accuracy:		0.904331
Seen 519000 samples, time used: 125.760s
  current training loss:		0.329510
  current training accuracy:		0.904347
Seen 522000 samples, time used: 125.722s
  current training loss:		0.329514
  current training accuracy:		0.904391
Seen 525000 samples, time used: 127.501s
  current training loss:		0.329596
  current training accuracy:		0.904354
Seen 528000 samples, time used: 124.397s
  current training loss:		0.329743
  current training accuracy:		0.904263
Seen 531000 samples, time used: 129.022s
  current training loss:		0.329817
  current training accuracy:		0.904213
Seen 534000 samples, time used: 123.701s
  current training loss:		0.329935
  current training accuracy:		0.904210
Seen 537000 samples, time used: 129.907s
  current training loss:		0.330013
  current training accuracy:		0.904147
Seen 540000 samples, time used: 122.883s
  current training loss:		0.330176
  current training accuracy:		0.904091
saving to ..., time used: 1262.587s
Seen 543000 samples, time used: 125.956s
  current training loss:		0.330208
  current training accuracy:		0.904039
Seen 546000 samples, time used: 125.860s
  current training loss:		0.330356
  current training accuracy:		0.904020
Seen 549000 samples, time used: 128.146s
  current training loss:		0.330597
  current training accuracy:		0.903958
Epoch 11 of 20 took 22978.833s
  training loss:		0.330614
  training accuracy:		90.39 %
  validation loss:		0.414340
  validation accuracy:		84.80 %
Seen 3000 samples, time used: 128.271s
  current training loss:		0.295965
  current training accuracy:		0.914667
Seen 6000 samples, time used: 124.421s
  current training loss:		0.300610
  current training accuracy:		0.916500
Seen 9000 samples, time used: 125.312s
  current training loss:		0.298869
  current training accuracy:		0.918222
Seen 12000 samples, time used: 126.694s
  current training loss:		0.301446
  current training accuracy:		0.918167
Seen 15000 samples, time used: 124.369s
  current training loss:		0.301406
  current training accuracy:		0.917467
Seen 18000 samples, time used: 126.664s
  current training loss:		0.300756
  current training accuracy:		0.918111
Seen 21000 samples, time used: 122.355s
  current training loss:		0.300683
  current training accuracy:		0.917476
Seen 24000 samples, time used: 125.493s
  current training loss:		0.303203
  current training accuracy:		0.915708
Seen 27000 samples, time used: 124.977s
  current training loss:		0.303155
  current training accuracy:		0.915444
Seen 30000 samples, time used: 128.788s
  current training loss:		0.304299
  current training accuracy:		0.915133
saving to ..., time used: 1257.345s
Seen 33000 samples, time used: 123.288s
  current training loss:		0.305005
  current training accuracy:		0.914848
Seen 36000 samples, time used: 126.624s
  current training loss:		0.305055
  current training accuracy:		0.915222
Seen 39000 samples, time used: 125.196s
  current training loss:		0.304843
  current training accuracy:		0.915359
Seen 42000 samples, time used: 122.655s
  current training loss:		0.305130
  current training accuracy:		0.915667
Seen 45000 samples, time used: 129.237s
  current training loss:		0.305354
  current training accuracy:		0.915600
Seen 48000 samples, time used: 125.099s
  current training loss:		0.305531
  current training accuracy:		0.915771
Seen 51000 samples, time used: 124.524s
  current training loss:		0.306001
  current training accuracy:		0.915608
Seen 54000 samples, time used: 121.661s
  current training loss:		0.306176
  current training accuracy:		0.916000
Seen 57000 samples, time used: 127.604s
  current training loss:		0.306196
  current training accuracy:		0.916088
Seen 60000 samples, time used: 124.836s
  current training loss:		0.306653
  current training accuracy:		0.916100
saving to ..., time used: 1250.522s
Seen 63000 samples, time used: 123.543s
  current training loss:		0.306697
  current training accuracy:		0.915873
Seen 66000 samples, time used: 125.057s
  current training loss:		0.307423
  current training accuracy:		0.915939
Seen 69000 samples, time used: 124.338s
  current training loss:		0.307893
  current training accuracy:		0.915594
Seen 72000 samples, time used: 125.040s
  current training loss:		0.308331
  current training accuracy:		0.915333
Seen 75000 samples, time used: 126.868s
  current training loss:		0.309121
  current training accuracy:		0.915267
Seen 78000 samples, time used: 125.966s
  current training loss:		0.308997
  current training accuracy:		0.915423
Seen 81000 samples, time used: 128.320s
  current training loss:		0.309149
  current training accuracy:		0.915407
Seen 84000 samples, time used: 127.507s
  current training loss:		0.309155
  current training accuracy:		0.915429
Seen 87000 samples, time used: 123.545s
  current training loss:		0.309873
  current training accuracy:		0.915310
Seen 90000 samples, time used: 128.866s
  current training loss:		0.310217
  current training accuracy:		0.915178
saving to ..., time used: 1258.848s
Seen 93000 samples, time used: 125.698s
  current training loss:		0.310859
  current training accuracy:		0.914946
Seen 96000 samples, time used: 124.381s
  current training loss:		0.310570
  current training accuracy:		0.914937
Seen 99000 samples, time used: 124.443s
  current training loss:		0.311448
  current training accuracy:		0.914717
Seen 102000 samples, time used: 124.566s
  current training loss:		0.311055
  current training accuracy:		0.914578
Seen 105000 samples, time used: 126.504s
  current training loss:		0.311165
  current training accuracy:		0.914448
Seen 108000 samples, time used: 126.206s
  current training loss:		0.311116
  current training accuracy:		0.914667
Seen 111000 samples, time used: 124.095s
  current training loss:		0.311280
  current training accuracy:		0.914523
Seen 114000 samples, time used: 129.749s
  current training loss:		0.311596
  current training accuracy:		0.914377
Seen 117000 samples, time used: 126.405s
  current training loss:		0.311680
  current training accuracy:		0.914453
Seen 120000 samples, time used: 123.497s
  current training loss:		0.311436
  current training accuracy:		0.914500
saving to ..., time used: 1255.343s
^Cexit ...