Namespace(batch_norm=False, batch_size=128, data_dir='~/.mxnet/datasets/imagenet', dtype='float16', input_size=224, label_smoothing=True, last_gamma=False, log_interval=50, logging_file='darknet53.log', lr=0.4, lr_decay=0.1, lr_decay_epoch='40,60', lr_decay_period=0, lr_mode='cosine', mixup=False, mixup_alpha=0.2, mixup_off_epoch=0, mode='hybrid', model='darknet53', momentum=0.9, no_wd=True, num_epochs=120, num_gpus=8, num_workers=60, rec_train='/media/ramdisk/rec/train.rec', rec_train_idx='/media/ramdisk/rec/train.idx', rec_val='/media/ramdisk/rec/val.rec', rec_val_idx='/media/ramdisk/rec/val.idx', save_dir='params_darknet53', save_frequency=10, use_pretrained=False, use_rec=True, use_se=False, warmup_epochs=5, warmup_lr=0.0, wd=0.0001)
Epoch[0] Batch [49]	Speed: 1733.851843 samples/sec	accuracy=0.663267	lr=0.003133
Epoch[0] Batch [99]	Speed: 4382.510306 samples/sec	accuracy=0.815434	lr=0.006331
Epoch[0] Batch [149]	Speed: 4402.197881 samples/sec	accuracy=0.870426	lr=0.009528
Epoch[0] Batch [199]	Speed: 4514.777738 samples/sec	accuracy=0.898172	lr=0.012726
Epoch[0] Batch [249]	Speed: 4499.884758 samples/sec	accuracy=0.912390	lr=0.015923
Epoch[0] Batch [299]	Speed: 4499.580688 samples/sec	accuracy=0.915812	lr=0.019121
Epoch[0] Batch [349]	Speed: 4495.959903 samples/sec	accuracy=0.912161	lr=0.022318
Epoch[0] Batch [399]	Speed: 4452.558387 samples/sec	accuracy=0.905209	lr=0.025516
Epoch[0] Batch [449]	Speed: 4568.128109 samples/sec	accuracy=0.894158	lr=0.028713
Epoch[0] Batch [499]	Speed: 4522.061041 samples/sec	accuracy=0.880224	lr=0.031910
Epoch[0] Batch [549]	Speed: 4441.500848 samples/sec	accuracy=0.866221	lr=0.035108
Epoch[0] Batch [599]	Speed: 4465.771257 samples/sec	accuracy=0.848952	lr=0.038305
Epoch[0] Batch [649]	Speed: 4441.194054 samples/sec	accuracy=0.830924	lr=0.041503
Epoch[0] Batch [699]	Speed: 4456.205835 samples/sec	accuracy=0.813574	lr=0.044700
Epoch[0] Batch [749]	Speed: 4477.684589 samples/sec	accuracy=0.797068	lr=0.047898
Epoch[0] Batch [799]	Speed: 4485.178139 samples/sec	accuracy=0.781629	lr=0.051095
Epoch[0] Batch [849]	Speed: 4493.583225 samples/sec	accuracy=0.766996	lr=0.054293
Epoch[0] Batch [899]	Speed: 4474.731676 samples/sec	accuracy=0.753074	lr=0.057490
Epoch[0] Batch [949]	Speed: 4484.561834 samples/sec	accuracy=0.740063	lr=0.060687
Epoch[0] Batch [999]	Speed: 4528.217691 samples/sec	accuracy=0.727716	lr=0.063885
Epoch[0] Batch [1049]	Speed: 4521.184871 samples/sec	accuracy=0.716372	lr=0.067082
Epoch[0] Batch [1099]	Speed: 4491.813569 samples/sec	accuracy=0.705742	lr=0.070280
Epoch[0] Batch [1149]	Speed: 4536.756070 samples/sec	accuracy=0.696531	lr=0.073477
Epoch[0] Batch [1199]	Speed: 4558.686973 samples/sec	accuracy=0.687365	lr=0.076675
Epoch[0] Batch [1249]	Speed: 4573.249335 samples/sec	accuracy=0.678482	lr=0.079872
[Epoch 0] training: accuracy=0.678137
[Epoch 0] speed: 4217 samples/sec	time cost: 314.145355
[Epoch 0] validation: err-top1=0.905453 err-top5=0.752611
Epoch[1] Batch [49]	Speed: 4323.368019 samples/sec	accuracy=0.461503	lr=0.083133
Epoch[1] Batch [99]	Speed: 4491.459298 samples/sec	accuracy=0.460135	lr=0.086331
Epoch[1] Batch [149]	Speed: 4414.998478 samples/sec	accuracy=0.468768	lr=0.089528
Epoch[1] Batch [199]	Speed: 4483.108568 samples/sec	accuracy=0.476879	lr=0.092726
Epoch[1] Batch [249]	Speed: 4485.784119 samples/sec	accuracy=0.475443	lr=0.095923
Epoch[1] Batch [299]	Speed: 4472.903699 samples/sec	accuracy=0.473956	lr=0.099121
Epoch[1] Batch [349]	Speed: 3984.876777 samples/sec	accuracy=0.474503	lr=0.102318
Epoch[1] Batch [399]	Speed: 4467.083115 samples/sec	accuracy=0.473230	lr=0.105516
Epoch[1] Batch [449]	Speed: 4513.422639 samples/sec	accuracy=0.472204	lr=0.108713
Epoch[1] Batch [499]	Speed: 4469.915963 samples/sec	accuracy=0.471297	lr=0.111910
Epoch[1] Batch [549]	Speed: 4496.206247 samples/sec	accuracy=0.470408	lr=0.115108
Epoch[1] Batch [599]	Speed: 4512.928094 samples/sec	accuracy=0.469775	lr=0.118305
Epoch[1] Batch [649]	Speed: 4453.251623 samples/sec	accuracy=0.469377	lr=0.121503
Epoch[1] Batch [699]	Speed: 4523.595608 samples/sec	accuracy=0.468886	lr=0.124700
Epoch[1] Batch [749]	Speed: 4525.928410 samples/sec	accuracy=0.468843	lr=0.127898
Epoch[1] Batch [799]	Speed: 4479.084172 samples/sec	accuracy=0.468765	lr=0.131095
Epoch[1] Batch [849]	Speed: 4241.666598 samples/sec	accuracy=0.468776	lr=0.134293
Epoch[1] Batch [899]	Speed: 4153.021024 samples/sec	accuracy=0.468794	lr=0.137490
Epoch[1] Batch [949]	Speed: 4200.417001 samples/sec	accuracy=0.468908	lr=0.140687
Epoch[1] Batch [999]	Speed: 4104.341641 samples/sec	accuracy=0.469059	lr=0.143885
Epoch[1] Batch [1049]	Speed: 4139.445992 samples/sec	accuracy=0.469259	lr=0.147082
Epoch[1] Batch [1099]	Speed: 4281.675690 samples/sec	accuracy=0.469637	lr=0.150280
Epoch[1] Batch [1149]	Speed: 4217.879475 samples/sec	accuracy=0.469980	lr=0.153477
Epoch[1] Batch [1199]	Speed: 4243.670570 samples/sec	accuracy=0.470365	lr=0.156675
Epoch[1] Batch [1249]	Speed: 4301.039848 samples/sec	accuracy=0.470718	lr=0.159872
[Epoch 1] training: accuracy=0.470736
[Epoch 1] speed: 4349 samples/sec	time cost: 304.989674
[Epoch 1] validation: err-top1=0.763213 err-top5=0.526367
Epoch[2] Batch [49]	Speed: 4148.234293 samples/sec	accuracy=0.478946	lr=0.163133
Epoch[2] Batch [99]	Speed: 4123.129523 samples/sec	accuracy=0.480536	lr=0.166331
Epoch[2] Batch [149]	Speed: 4101.742493 samples/sec	accuracy=0.481555	lr=0.169528
Epoch[2] Batch [199]	Speed: 4157.987633 samples/sec	accuracy=0.481908	lr=0.172726
Epoch[2] Batch [249]	Speed: 4098.793075 samples/sec	accuracy=0.482438	lr=0.175923
Epoch[2] Batch [299]	Speed: 4181.123335 samples/sec	accuracy=0.483439	lr=0.179121
Epoch[2] Batch [349]	Speed: 4080.490900 samples/sec	accuracy=0.484073	lr=0.182318
Epoch[2] Batch [399]	Speed: 4096.802970 samples/sec	accuracy=0.484975	lr=0.185516
Epoch[2] Batch [449]	Speed: 4205.382966 samples/sec	accuracy=0.486044	lr=0.188713
Epoch[2] Batch [499]	Speed: 4182.994555 samples/sec	accuracy=0.487077	lr=0.191910
Epoch[2] Batch [549]	Speed: 4297.915322 samples/sec	accuracy=0.487840	lr=0.195108
Epoch[2] Batch [599]	Speed: 4259.121930 samples/sec	accuracy=0.488353	lr=0.198305
Epoch[2] Batch [649]	Speed: 4233.522376 samples/sec	accuracy=0.489006	lr=0.201503
Epoch[2] Batch [699]	Speed: 4290.797413 samples/sec	accuracy=0.489719	lr=0.204700
Epoch[2] Batch [749]	Speed: 4235.342734 samples/sec	accuracy=0.490450	lr=0.207898
Epoch[2] Batch [799]	Speed: 4127.596320 samples/sec	accuracy=0.491295	lr=0.211095
Epoch[2] Batch [849]	Speed: 4238.010637 samples/sec	accuracy=0.491908	lr=0.214293
Epoch[2] Batch [899]	Speed: 4296.147877 samples/sec	accuracy=0.492649	lr=0.217490
Epoch[2] Batch [949]	Speed: 4136.495984 samples/sec	accuracy=0.493391	lr=0.220687
Epoch[2] Batch [999]	Speed: 4121.679682 samples/sec	accuracy=0.494165	lr=0.223885
Epoch[2] Batch [1049]	Speed: 4178.695023 samples/sec	accuracy=0.494898	lr=0.227082
Epoch[2] Batch [1099]	Speed: 4087.792339 samples/sec	accuracy=0.495581	lr=0.230280
Epoch[2] Batch [1149]	Speed: 4053.715830 samples/sec	accuracy=0.496241	lr=0.233477
Epoch[2] Batch [1199]	Speed: 4170.132301 samples/sec	accuracy=0.496959	lr=0.236675
Epoch[2] Batch [1249]	Speed: 4415.958373 samples/sec	accuracy=0.497651	lr=0.239872
[Epoch 2] training: accuracy=0.497670
[Epoch 2] speed: 4175 samples/sec	time cost: 317.169433
[Epoch 2] validation: err-top1=0.654456 err-top5=0.387277
Epoch[3] Batch [49]	Speed: 3959.429483 samples/sec	accuracy=0.511575	lr=0.243133
Epoch[3] Batch [99]	Speed: 3975.016030 samples/sec	accuracy=0.513557	lr=0.246331
Epoch[3] Batch [149]	Speed: 4033.366379 samples/sec	accuracy=0.514128	lr=0.249528
Epoch[3] Batch [199]	Speed: 4183.424889 samples/sec	accuracy=0.514778	lr=0.252726
Epoch[3] Batch [249]	Speed: 4178.112265 samples/sec	accuracy=0.515709	lr=0.255923
Epoch[3] Batch [299]	Speed: 4205.927145 samples/sec	accuracy=0.516382	lr=0.259121
Epoch[3] Batch [349]	Speed: 4347.908033 samples/sec	accuracy=0.516769	lr=0.262318
Epoch[3] Batch [399]	Speed: 4157.388662 samples/sec	accuracy=0.517233	lr=0.265516
Epoch[3] Batch [449]	Speed: 4176.556084 samples/sec	accuracy=0.517830	lr=0.268713
Epoch[3] Batch [499]	Speed: 4179.418495 samples/sec	accuracy=0.518449	lr=0.271910
Epoch[3] Batch [549]	Speed: 4358.598210 samples/sec	accuracy=0.519067	lr=0.275108
Epoch[3] Batch [599]	Speed: 4253.393943 samples/sec	accuracy=0.519598	lr=0.278305
Epoch[3] Batch [649]	Speed: 4135.554411 samples/sec	accuracy=0.520146	lr=0.281503
Epoch[3] Batch [699]	Speed: 4191.897507 samples/sec	accuracy=0.520683	lr=0.284700
Epoch[3] Batch [749]	Speed: 4264.948965 samples/sec	accuracy=0.521257	lr=0.287898
Epoch[3] Batch [799]	Speed: 4076.361022 samples/sec	accuracy=0.521799	lr=0.291095
Epoch[3] Batch [849]	Speed: 4333.326946 samples/sec	accuracy=0.522271	lr=0.294293
Epoch[3] Batch [899]	Speed: 4198.025484 samples/sec	accuracy=0.523149	lr=0.297490
Epoch[3] Batch [949]	Speed: 4204.049101 samples/sec	accuracy=0.523660	lr=0.300687
Epoch[3] Batch [999]	Speed: 4121.459379 samples/sec	accuracy=0.524132	lr=0.303885
Epoch[3] Batch [1049]	Speed: 4066.023654 samples/sec	accuracy=0.524591	lr=0.307082
Epoch[3] Batch [1099]	Speed: 4079.360760 samples/sec	accuracy=0.525073	lr=0.310280
Epoch[3] Batch [1149]	Speed: 4186.330349 samples/sec	accuracy=0.525523	lr=0.313477
Epoch[3] Batch [1199]	Speed: 4257.067391 samples/sec	accuracy=0.525948	lr=0.316675
Epoch[3] Batch [1249]	Speed: 4384.890168 samples/sec	accuracy=0.526345	lr=0.319872
[Epoch 3] training: accuracy=0.526362
[Epoch 3] speed: 4173 samples/sec	time cost: 317.027865
[Epoch 3] validation: err-top1=0.601562 err-top5=0.336934
Epoch[4] Batch [49]	Speed: 4096.255641 samples/sec	accuracy=0.534788	lr=0.323133
Epoch[4] Batch [99]	Speed: 4011.378557 samples/sec	accuracy=0.535591	lr=0.326331
Epoch[4] Batch [149]	Speed: 4119.182422 samples/sec	accuracy=0.536077	lr=0.329528
Epoch[4] Batch [199]	Speed: 4177.617275 samples/sec	accuracy=0.536465	lr=0.332726
Epoch[4] Batch [249]	Speed: 4341.542759 samples/sec	accuracy=0.537018	lr=0.335923
Epoch[4] Batch [299]	Speed: 4270.623036 samples/sec	accuracy=0.537584	lr=0.339121
Epoch[4] Batch [349]	Speed: 4235.688914 samples/sec	accuracy=0.538149	lr=0.342318
Epoch[4] Batch [399]	Speed: 4208.611306 samples/sec	accuracy=0.538621	lr=0.345516
Epoch[4] Batch [449]	Speed: 4182.498408 samples/sec	accuracy=0.538880	lr=0.348713
Epoch[4] Batch [499]	Speed: 4116.394954 samples/sec	accuracy=0.538983	lr=0.351910
Epoch[4] Batch [549]	Speed: 4041.168193 samples/sec	accuracy=0.539311	lr=0.355108
Epoch[4] Batch [599]	Speed: 4313.557175 samples/sec	accuracy=0.539611	lr=0.358305
Epoch[4] Batch [649]	Speed: 4210.284249 samples/sec	accuracy=0.539912	lr=0.361503
Epoch[4] Batch [699]	Speed: 4024.594627 samples/sec	accuracy=0.540124	lr=0.364700
Epoch[4] Batch [749]	Speed: 4259.016765 samples/sec	accuracy=0.540467	lr=0.367898
Epoch[4] Batch [799]	Speed: 4286.398357 samples/sec	accuracy=0.540722	lr=0.371095
Epoch[4] Batch [849]	Speed: 4099.999292 samples/sec	accuracy=0.540987	lr=0.374293
Epoch[4] Batch [899]	Speed: 4210.396844 samples/sec	accuracy=0.541322	lr=0.377490
Epoch[4] Batch [949]	Speed: 4254.611961 samples/sec	accuracy=0.541657	lr=0.380687
Epoch[4] Batch [999]	Speed: 4325.085200 samples/sec	accuracy=0.541981	lr=0.383885
Epoch[4] Batch [1049]	Speed: 4243.641722 samples/sec	accuracy=0.542260	lr=0.387082
Epoch[4] Batch [1099]	Speed: 4228.379120 samples/sec	accuracy=0.542511	lr=0.390280
Epoch[4] Batch [1149]	Speed: 4138.845332 samples/sec	accuracy=0.542808	lr=0.393477
Epoch[4] Batch [1199]	Speed: 4070.562864 samples/sec	accuracy=0.543097	lr=0.396675
Epoch[4] Batch [1249]	Speed: 4408.234139 samples/sec	accuracy=0.543383	lr=0.399872
[Epoch 4] training: accuracy=0.543394
[Epoch 4] speed: 4188 samples/sec	time cost: 315.856608
[Epoch 4] validation: err-top1=0.589385 err-top5=0.324179
Epoch[5] Batch [49]	Speed: 4161.619760 samples/sec	accuracy=0.550532	lr=0.400000
Epoch[5] Batch [99]	Speed: 4055.734676 samples/sec	accuracy=0.549722	lr=0.400000
Epoch[5] Batch [149]	Speed: 4208.549364 samples/sec	accuracy=0.550233	lr=0.399999
Epoch[5] Batch [199]	Speed: 4245.613399 samples/sec	accuracy=0.550097	lr=0.399998
Epoch[5] Batch [249]	Speed: 4336.106891 samples/sec	accuracy=0.550227	lr=0.399997
Epoch[5] Batch [299]	Speed: 4211.655445 samples/sec	accuracy=0.550618	lr=0.399996
Epoch[5] Batch [349]	Speed: 4170.101043 samples/sec	accuracy=0.550968	lr=0.399994
Epoch[5] Batch [399]	Speed: 4335.069025 samples/sec	accuracy=0.551339	lr=0.399992
Epoch[5] Batch [449]	Speed: 4160.417884 samples/sec	accuracy=0.551347	lr=0.399990
Epoch[5] Batch [499]	Speed: 4302.001843 samples/sec	accuracy=0.551568	lr=0.399988
Epoch[5] Batch [549]	Speed: 4154.079368 samples/sec	accuracy=0.551951	lr=0.399986
Epoch[5] Batch [599]	Speed: 4210.060315 samples/sec	accuracy=0.552312	lr=0.399983
Epoch[5] Batch [649]	Speed: 4217.544565 samples/sec	accuracy=0.552492	lr=0.399980
Epoch[5] Batch [699]	Speed: 4291.763066 samples/sec	accuracy=0.552765	lr=0.399977
Epoch[5] Batch [749]	Speed: 4256.158958 samples/sec	accuracy=0.553079	lr=0.399973
Epoch[5] Batch [799]	Speed: 4178.046097 samples/sec	accuracy=0.553361	lr=0.399970
Epoch[5] Batch [849]	Speed: 4258.442887 samples/sec	accuracy=0.553800	lr=0.399966
Epoch[5] Batch [899]	Speed: 4267.694424 samples/sec	accuracy=0.554094	lr=0.399961
Epoch[5] Batch [949]	Speed: 4205.631359 samples/sec	accuracy=0.554392	lr=0.399957
Epoch[5] Batch [999]	Speed: 4183.327178 samples/sec	accuracy=0.554729	lr=0.399952
Epoch[5] Batch [1049]	Speed: 4116.946021 samples/sec	accuracy=0.554970	lr=0.399948
Epoch[5] Batch [1099]	Speed: 4177.193579 samples/sec	accuracy=0.555306	lr=0.399942
Epoch[5] Batch [1149]	Speed: 4075.289083 samples/sec	accuracy=0.555549	lr=0.399937
Epoch[5] Batch [1199]	Speed: 4237.088247 samples/sec	accuracy=0.555806	lr=0.399931
Epoch[5] Batch [1249]	Speed: 4344.948455 samples/sec	accuracy=0.556138	lr=0.399926
[Epoch 5] training: accuracy=0.556145
[Epoch 5] speed: 4209 samples/sec	time cost: 314.836725
[Epoch 5] validation: err-top1=0.527608 err-top5=0.265137
Epoch[6] Batch [49]	Speed: 4148.105528 samples/sec	accuracy=0.561200	lr=0.399919
Epoch[6] Batch [99]	Speed: 4191.415771 samples/sec	accuracy=0.562204	lr=0.399913
Epoch[6] Batch [149]	Speed: 4229.823530 samples/sec	accuracy=0.562681	lr=0.399907
Epoch[6] Batch [199]	Speed: 4114.045875 samples/sec	accuracy=0.562530	lr=0.399900
Epoch[6] Batch [249]	Speed: 4099.453378 samples/sec	accuracy=0.562798	lr=0.399893
Epoch[6] Batch [299]	Speed: 4178.965157 samples/sec	accuracy=0.563046	lr=0.399885
Epoch[6] Batch [349]	Speed: 4221.039006 samples/sec	accuracy=0.563285	lr=0.399878
Epoch[6] Batch [399]	Speed: 4250.741645 samples/sec	accuracy=0.563478	lr=0.399870
Epoch[6] Batch [449]	Speed: 4126.034250 samples/sec	accuracy=0.563879	lr=0.399862
Epoch[6] Batch [499]	Speed: 4216.659212 samples/sec	accuracy=0.564075	lr=0.399854
Epoch[6] Batch [549]	Speed: 4233.524546 samples/sec	accuracy=0.564251	lr=0.399846
Epoch[6] Batch [599]	Speed: 4278.198895 samples/sec	accuracy=0.564452	lr=0.399837
Epoch[6] Batch [649]	Speed: 4302.493907 samples/sec	accuracy=0.564733	lr=0.399828
Epoch[6] Batch [699]	Speed: 4268.212094 samples/sec	accuracy=0.565094	lr=0.399819
Epoch[6] Batch [749]	Speed: 4234.810291 samples/sec	accuracy=0.565450	lr=0.399809
Epoch[6] Batch [799]	Speed: 3986.497020 samples/sec	accuracy=0.565749	lr=0.399800
Epoch[6] Batch [849]	Speed: 4162.941117 samples/sec	accuracy=0.566039	lr=0.399790
Epoch[6] Batch [899]	Speed: 4244.070535 samples/sec	accuracy=0.566190	lr=0.399780
Epoch[6] Batch [949]	Speed: 4303.784287 samples/sec	accuracy=0.566512	lr=0.399769
Epoch[6] Batch [999]	Speed: 4217.819331 samples/sec	accuracy=0.566727	lr=0.399759
Epoch[6] Batch [1049]	Speed: 4249.702861 samples/sec	accuracy=0.566937	lr=0.399748
Epoch[6] Batch [1099]	Speed: 4295.509217 samples/sec	accuracy=0.567193	lr=0.399737
Epoch[6] Batch [1149]	Speed: 4298.630933 samples/sec	accuracy=0.567522	lr=0.399725
Epoch[6] Batch [1199]	Speed: 4063.060250 samples/sec	accuracy=0.567767	lr=0.399714
Epoch[6] Batch [1249]	Speed: 4367.831460 samples/sec	accuracy=0.567963	lr=0.399702
[Epoch 6] training: accuracy=0.567972
[Epoch 6] speed: 4205 samples/sec	time cost: 315.119747
[Epoch 6] validation: err-top1=0.528021 err-top5=0.267937
Epoch[7] Batch [49]	Speed: 4140.790750 samples/sec	accuracy=0.570853	lr=0.399690
Epoch[7] Batch [99]	Speed: 4138.628215 samples/sec	accuracy=0.571561	lr=0.399677
Epoch[7] Batch [149]	Speed: 4144.250039 samples/sec	accuracy=0.572319	lr=0.399665
Epoch[7] Batch [199]	Speed: 4264.992333 samples/sec	accuracy=0.572565	lr=0.399652
Epoch[7] Batch [249]	Speed: 4245.064273 samples/sec	accuracy=0.572789	lr=0.399639
Epoch[7] Batch [299]	Speed: 4237.552193 samples/sec	accuracy=0.573036	lr=0.399626
Epoch[7] Batch [349]	Speed: 4191.998483 samples/sec	accuracy=0.573254	lr=0.399613
Epoch[7] Batch [399]	Speed: 4118.958752 samples/sec	accuracy=0.573432	lr=0.399599
Epoch[7] Batch [449]	Speed: 4123.791039 samples/sec	accuracy=0.573614	lr=0.399585
Epoch[7] Batch [499]	Speed: 4066.970259 samples/sec	accuracy=0.573859	lr=0.399571
Epoch[7] Batch [549]	Speed: 4186.258698 samples/sec	accuracy=0.574062	lr=0.399556
Epoch[7] Batch [599]	Speed: 4279.205442 samples/sec	accuracy=0.574281	lr=0.399542
Epoch[7] Batch [649]	Speed: 4170.200162 samples/sec	accuracy=0.574589	lr=0.399527
Epoch[7] Batch [699]	Speed: 4226.420587 samples/sec	accuracy=0.574828	lr=0.399512
Epoch[7] Batch [749]	Speed: 4115.378200 samples/sec	accuracy=0.575065	lr=0.399496
Epoch[7] Batch [799]	Speed: 4079.207720 samples/sec	accuracy=0.575319	lr=0.399481
Epoch[7] Batch [849]	Speed: 4254.730396 samples/sec	accuracy=0.575530	lr=0.399465
Epoch[7] Batch [899]	Speed: 4040.272630 samples/sec	accuracy=0.575760	lr=0.399449
Epoch[7] Batch [949]	Speed: 4106.968905 samples/sec	accuracy=0.576015	lr=0.399432
Epoch[7] Batch [999]	Speed: 4223.622291 samples/sec	accuracy=0.576286	lr=0.399416
Epoch[7] Batch [1049]	Speed: 4286.702277 samples/sec	accuracy=0.576498	lr=0.399399
Epoch[7] Batch [1099]	Speed: 4171.095033 samples/sec	accuracy=0.576666	lr=0.399382
Epoch[7] Batch [1149]	Speed: 4119.338871 samples/sec	accuracy=0.576888	lr=0.399365
Epoch[7] Batch [1199]	Speed: 4140.532953 samples/sec	accuracy=0.577154	lr=0.399347
Epoch[7] Batch [1249]	Speed: 4177.171803 samples/sec	accuracy=0.577375	lr=0.399329
[Epoch 7] training: accuracy=0.577381
[Epoch 7] speed: 4165 samples/sec	time cost: 318.140364
[Epoch 7] validation: err-top1=0.489039 err-top5=0.233598
Epoch[8] Batch [49]	Speed: 4124.302900 samples/sec	accuracy=0.579658	lr=0.399311
Epoch[8] Batch [99]	Speed: 4316.717943 samples/sec	accuracy=0.580418	lr=0.399293
Epoch[8] Batch [149]	Speed: 4267.931741 samples/sec	accuracy=0.579972	lr=0.399274
Epoch[8] Batch [199]	Speed: 4274.402972 samples/sec	accuracy=0.580710	lr=0.399256
Epoch[8] Batch [249]	Speed: 4241.163472 samples/sec	accuracy=0.580869	lr=0.399237
Epoch[8] Batch [299]	Speed: 4245.398280 samples/sec	accuracy=0.580814	lr=0.399218
Epoch[8] Batch [349]	Speed: 4137.409209 samples/sec	accuracy=0.581022	lr=0.399198
Epoch[8] Batch [399]	Speed: 4260.506276 samples/sec	accuracy=0.581271	lr=0.399179
Epoch[8] Batch [449]	Speed: 4121.163490 samples/sec	accuracy=0.581618	lr=0.399159
Epoch[8] Batch [499]	Speed: 4030.798743 samples/sec	accuracy=0.581776	lr=0.399138
Epoch[8] Batch [549]	Speed: 4115.607397 samples/sec	accuracy=0.582038	lr=0.399118
Epoch[8] Batch [599]	Speed: 4142.030923 samples/sec	accuracy=0.582218	lr=0.399098
Epoch[8] Batch [649]	Speed: 4149.734714 samples/sec	accuracy=0.582359	lr=0.399077
Epoch[8] Batch [699]	Speed: 4227.246469 samples/sec	accuracy=0.582538	lr=0.399056
Epoch[8] Batch [749]	Speed: 4228.401016 samples/sec	accuracy=0.582771	lr=0.399034
Epoch[8] Batch [799]	Speed: 4112.082194 samples/sec	accuracy=0.583076	lr=0.399013
Epoch[8] Batch [849]	Speed: 4124.550203 samples/sec	accuracy=0.583171	lr=0.398991
Epoch[8] Batch [899]	Speed: 4091.707940 samples/sec	accuracy=0.583403	lr=0.398969
Epoch[8] Batch [949]	Speed: 4111.631065 samples/sec	accuracy=0.583597	lr=0.398947
Epoch[8] Batch [999]	Speed: 4214.254379 samples/sec	accuracy=0.583707	lr=0.398924
Epoch[8] Batch [1049]	Speed: 4283.936673 samples/sec	accuracy=0.583899	lr=0.398901
Epoch[8] Batch [1099]	Speed: 4234.373579 samples/sec	accuracy=0.584061	lr=0.398878
Epoch[8] Batch [1149]	Speed: 4187.862952 samples/sec	accuracy=0.584196	lr=0.398855
Epoch[8] Batch [1199]	Speed: 4153.113630 samples/sec	accuracy=0.584399	lr=0.398832
Epoch[8] Batch [1249]	Speed: 4189.219172 samples/sec	accuracy=0.584564	lr=0.398808
[Epoch 8] training: accuracy=0.584565
[Epoch 8] speed: 4178 samples/sec	time cost: 316.883608
[Epoch 8] validation: err-top1=0.478795 err-top5=0.223394
Epoch[9] Batch [49]	Speed: 4060.060019 samples/sec	accuracy=0.588379	lr=0.398784
Epoch[9] Batch [99]	Speed: 4185.953105 samples/sec	accuracy=0.587660	lr=0.398760
Epoch[9] Batch [149]	Speed: 4207.024085 samples/sec	accuracy=0.587116	lr=0.398735
Epoch[9] Batch [199]	Speed: 4305.695626 samples/sec	accuracy=0.587523	lr=0.398710
Epoch[9] Batch [249]	Speed: 4292.262570 samples/sec	accuracy=0.587731	lr=0.398686
Epoch[9] Batch [299]	Speed: 4200.456274 samples/sec	accuracy=0.587757	lr=0.398660
Epoch[9] Batch [349]	Speed: 4289.817456 samples/sec	accuracy=0.587899	lr=0.398635
Epoch[9] Batch [399]	Speed: 4226.340320 samples/sec	accuracy=0.588269	lr=0.398610
Epoch[9] Batch [449]	Speed: 4252.285491 samples/sec	accuracy=0.588404	lr=0.398584
Epoch[9] Batch [499]	Speed: 4271.436805 samples/sec	accuracy=0.588621	lr=0.398558
Epoch[9] Batch [549]	Speed: 4206.847389 samples/sec	accuracy=0.588838	lr=0.398531
Epoch[9] Batch [599]	Speed: 4208.018771 samples/sec	accuracy=0.589029	lr=0.398505
Epoch[9] Batch [649]	Speed: 4122.971203 samples/sec	accuracy=0.589234	lr=0.398478
Epoch[9] Batch [699]	Speed: 4222.087812 samples/sec	accuracy=0.589316	lr=0.398451
Epoch[9] Batch [749]	Speed: 4044.717776 samples/sec	accuracy=0.589554	lr=0.398424
Epoch[9] Batch [799]	Speed: 4160.775947 samples/sec	accuracy=0.589716	lr=0.398396
Epoch[9] Batch [849]	Speed: 4186.511039 samples/sec	accuracy=0.589848	lr=0.398369
Epoch[9] Batch [899]	Speed: 4058.100897 samples/sec	accuracy=0.589904	lr=0.398341
Epoch[9] Batch [949]	Speed: 4196.059953 samples/sec	accuracy=0.590074	lr=0.398312
Epoch[9] Batch [999]	Speed: 4194.377975 samples/sec	accuracy=0.590239	lr=0.398284
Epoch[9] Batch [1049]	Speed: 4253.250564 samples/sec	accuracy=0.590386	lr=0.398255
Epoch[9] Batch [1099]	Speed: 4193.986256 samples/sec	accuracy=0.590482	lr=0.398226
Epoch[9] Batch [1149]	Speed: 4116.518365 samples/sec	accuracy=0.590625	lr=0.398197
Epoch[9] Batch [1199]	Speed: 4114.036969 samples/sec	accuracy=0.590778	lr=0.398168
Epoch[9] Batch [1249]	Speed: 4295.526831 samples/sec	accuracy=0.590901	lr=0.398138
[Epoch 9] training: accuracy=0.590908
[Epoch 9] speed: 4189 samples/sec	time cost: 316.213605
[Epoch 9] validation: err-top1=0.451192 err-top5=0.202069
Epoch[10] Batch [49]	Speed: 4152.053049 samples/sec	accuracy=0.591077	lr=0.398108
Epoch[10] Batch [99]	Speed: 4170.818463 samples/sec	accuracy=0.590140	lr=0.398078
Epoch[10] Batch [149]	Speed: 4135.902870 samples/sec	accuracy=0.590192	lr=0.398048
Epoch[10] Batch [199]	Speed: 4215.868579 samples/sec	accuracy=0.590564	lr=0.398017
Epoch[10] Batch [249]	Speed: 4150.581114 samples/sec	accuracy=0.591120	lr=0.397986
Epoch[10] Batch [299]	Speed: 4035.269156 samples/sec	accuracy=0.591548	lr=0.397955
Epoch[10] Batch [349]	Speed: 4169.389941 samples/sec	accuracy=0.591921	lr=0.397924
Epoch[10] Batch [399]	Speed: 4174.013647 samples/sec	accuracy=0.592266	lr=0.397892
Epoch[10] Batch [449]	Speed: 4058.872658 samples/sec	accuracy=0.592627	lr=0.397861
Epoch[10] Batch [499]	Speed: 4187.942744 samples/sec	accuracy=0.592874	lr=0.397829
Epoch[10] Batch [549]	Speed: 4204.317008 samples/sec	accuracy=0.593140	lr=0.397796
Epoch[10] Batch [599]	Speed: 4204.479250 samples/sec	accuracy=0.593369	lr=0.397764
Epoch[10] Batch [649]	Speed: 4071.748660 samples/sec	accuracy=0.593598	lr=0.397731
Epoch[10] Batch [699]	Speed: 4046.680388 samples/sec	accuracy=0.593934	lr=0.397698
Epoch[10] Batch [749]	Speed: 4143.749927 samples/sec	accuracy=0.594075	lr=0.397665
Epoch[10] Batch [799]	Speed: 4054.633057 samples/sec	accuracy=0.594138	lr=0.397632
Epoch[10] Batch [849]	Speed: 4190.578396 samples/sec	accuracy=0.594183	lr=0.397598
Epoch[10] Batch [899]	Speed: 4135.619558 samples/sec	accuracy=0.594306	lr=0.397564
Epoch[10] Batch [949]	Speed: 4281.870595 samples/sec	accuracy=0.594491	lr=0.397530
Epoch[10] Batch [999]	Speed: 4281.827908 samples/sec	accuracy=0.594636	lr=0.397496
Epoch[10] Batch [1049]	Speed: 4179.300149 samples/sec	accuracy=0.594829	lr=0.397461
Epoch[10] Batch [1099]	Speed: 4170.755604 samples/sec	accuracy=0.594954	lr=0.397427
Epoch[10] Batch [1149]	Speed: 4178.337284 samples/sec	accuracy=0.595073	lr=0.397392
Epoch[10] Batch [1199]	Speed: 4298.219931 samples/sec	accuracy=0.595201	lr=0.397356
Epoch[10] Batch [1249]	Speed: 4270.841738 samples/sec	accuracy=0.595318	lr=0.397321
[Epoch 10] training: accuracy=0.595320
[Epoch 10] speed: 4161 samples/sec	time cost: 318.266435
[Epoch 10] validation: err-top1=0.444196 err-top5=0.199418
Epoch[11] Batch [49]	Speed: 4022.801069 samples/sec	accuracy=0.596002	lr=0.397284
Epoch[11] Batch [99]	Speed: 4148.831831 samples/sec	accuracy=0.595748	lr=0.397248
Epoch[11] Batch [149]	Speed: 4062.707585 samples/sec	accuracy=0.595929	lr=0.397212
Epoch[11] Batch [199]	Speed: 4154.393183 samples/sec	accuracy=0.596272	lr=0.397176
Epoch[11] Batch [249]	Speed: 4292.902753 samples/sec	accuracy=0.596491	lr=0.397139
Epoch[11] Batch [299]	Speed: 4183.567430 samples/sec	accuracy=0.596729	lr=0.397102
Epoch[11] Batch [349]	Speed: 4273.880908 samples/sec	accuracy=0.596706	lr=0.397065
Epoch[11] Batch [399]	Speed: 4085.930363 samples/sec	accuracy=0.596817	lr=0.397028
Epoch[11] Batch [449]	Speed: 4162.209060 samples/sec	accuracy=0.597070	lr=0.396990
Epoch[11] Batch [499]	Speed: 4077.111878 samples/sec	accuracy=0.597237	lr=0.396952
Epoch[11] Batch [549]	Speed: 4123.888681 samples/sec	accuracy=0.597396	lr=0.396914
Epoch[11] Batch [599]	Speed: 4161.213816 samples/sec	accuracy=0.597408	lr=0.396876
Epoch[11] Batch [649]	Speed: 4191.226723 samples/sec	accuracy=0.597534	lr=0.396837
Epoch[11] Batch [699]	Speed: 4137.708392 samples/sec	accuracy=0.597579	lr=0.396798
Epoch[11] Batch [749]	Speed: 4270.009772 samples/sec	accuracy=0.597757	lr=0.396759
Epoch[11] Batch [799]	Speed: 4121.720502 samples/sec	accuracy=0.597881	lr=0.396720
Epoch[11] Batch [849]	Speed: 4207.378676 samples/sec	accuracy=0.597977	lr=0.396680
Epoch[11] Batch [899]	Speed: 4256.144449 samples/sec	accuracy=0.598151	lr=0.396641
Epoch[11] Batch [949]	Speed: 4208.962452 samples/sec	accuracy=0.598281	lr=0.396601
Epoch[11] Batch [999]	Speed: 4156.830739 samples/sec	accuracy=0.598465	lr=0.396561
Epoch[11] Batch [1049]	Speed: 4224.085452 samples/sec	accuracy=0.598569	lr=0.396520
Epoch[11] Batch [1099]	Speed: 4170.373792 samples/sec	accuracy=0.598770	lr=0.396479
Epoch[11] Batch [1149]	Speed: 4212.184807 samples/sec	accuracy=0.598958	lr=0.396439
Epoch[11] Batch [1199]	Speed: 4153.417499 samples/sec	accuracy=0.599136	lr=0.396397
Epoch[11] Batch [1249]	Speed: 4176.577528 samples/sec	accuracy=0.599267	lr=0.396356
[Epoch 11] training: accuracy=0.599269
[Epoch 11] speed: 4164 samples/sec	time cost: 318.264971
[Epoch 11] validation: err-top1=0.449910 err-top5=0.203593
Epoch[12] Batch [49]	Speed: 4105.494222 samples/sec	accuracy=0.601787	lr=0.396314
Epoch[12] Batch [99]	Speed: 4105.150239 samples/sec	accuracy=0.600564	lr=0.396272
Epoch[12] Batch [149]	Speed: 4257.249597 samples/sec	accuracy=0.600981	lr=0.396230
Epoch[12] Batch [199]	Speed: 4084.614393 samples/sec	accuracy=0.600868	lr=0.396187
Epoch[12] Batch [249]	Speed: 4084.861388 samples/sec	accuracy=0.600576	lr=0.396145
Epoch[12] Batch [299]	Speed: 4090.046245 samples/sec	accuracy=0.600803	lr=0.396102
Epoch[12] Batch [349]	Speed: 4047.543778 samples/sec	accuracy=0.600908	lr=0.396059
Epoch[12] Batch [399]	Speed: 4112.694565 samples/sec	accuracy=0.601101	lr=0.396016
Epoch[12] Batch [449]	Speed: 4270.984692 samples/sec	accuracy=0.601057	lr=0.395972
Epoch[12] Batch [499]	Speed: 4082.033877 samples/sec	accuracy=0.601170	lr=0.395928
Epoch[12] Batch [549]	Speed: 4199.255266 samples/sec	accuracy=0.601310	lr=0.395885
Epoch[12] Batch [599]	Speed: 4201.100429 samples/sec	accuracy=0.601451	lr=0.395840
Epoch[12] Batch [649]	Speed: 4336.377183 samples/sec	accuracy=0.601648	lr=0.395796
Epoch[12] Batch [699]	Speed: 4156.779485 samples/sec	accuracy=0.601738	lr=0.395751
Epoch[12] Batch [749]	Speed: 4077.297275 samples/sec	accuracy=0.601872	lr=0.395706
Epoch[12] Batch [799]	Speed: 4229.529288 samples/sec	accuracy=0.601981	lr=0.395661
Epoch[12] Batch [849]	Speed: 4132.660366 samples/sec	accuracy=0.602082	lr=0.395616
Epoch[12] Batch [899]	Speed: 4138.240859 samples/sec	accuracy=0.602167	lr=0.395570
Epoch[12] Batch [949]	Speed: 4173.969107 samples/sec	accuracy=0.602252	lr=0.395524
Epoch[12] Batch [999]	Speed: 4103.519639 samples/sec	accuracy=0.602380	lr=0.395478
Epoch[12] Batch [1049]	Speed: 4113.212655 samples/sec	accuracy=0.602529	lr=0.395432
Epoch[12] Batch [1099]	Speed: 4164.627047 samples/sec	accuracy=0.602606	lr=0.395386
Epoch[12] Batch [1149]	Speed: 4100.122113 samples/sec	accuracy=0.602743	lr=0.395339
Epoch[12] Batch [1199]	Speed: 4137.074204 samples/sec	accuracy=0.602867	lr=0.395292
Epoch[12] Batch [1249]	Speed: 4319.395725 samples/sec	accuracy=0.602996	lr=0.395245
[Epoch 12] training: accuracy=0.602996
[Epoch 12] speed: 4147 samples/sec	time cost: 319.168194
[Epoch 12] validation: err-top1=0.444655 err-top5=0.196429
Epoch[13] Batch [49]	Speed: 4091.226039 samples/sec	accuracy=0.602982	lr=0.395196
Epoch[13] Batch [99]	Speed: 4012.518495 samples/sec	accuracy=0.602895	lr=0.395149
Epoch[13] Batch [149]	Speed: 4182.082925 samples/sec	accuracy=0.602753	lr=0.395101
Epoch[13] Batch [199]	Speed: 4199.892975 samples/sec	accuracy=0.603285	lr=0.395052
Epoch[13] Batch [249]	Speed: 4351.114513 samples/sec	accuracy=0.603541	lr=0.395004
Epoch[13] Batch [299]	Speed: 4153.266964 samples/sec	accuracy=0.603690	lr=0.394955
Epoch[13] Batch [349]	Speed: 4055.885653 samples/sec	accuracy=0.603793	lr=0.394907
Epoch[13] Batch [399]	Speed: 4141.266668 samples/sec	accuracy=0.603705	lr=0.394858
Epoch[13] Batch [449]	Speed: 4168.652943 samples/sec	accuracy=0.603945	lr=0.394808
Epoch[13] Batch [499]	Speed: 4180.727659 samples/sec	accuracy=0.604129	lr=0.394759
Epoch[13] Batch [549]	Speed: 4097.040655 samples/sec	accuracy=0.604237	lr=0.394709
Epoch[13] Batch [599]	Speed: 4177.167009 samples/sec	accuracy=0.604393	lr=0.394659
Epoch[13] Batch [649]	Speed: 4090.708563 samples/sec	accuracy=0.604535	lr=0.394609
Epoch[13] Batch [699]	Speed: 4122.444320 samples/sec	accuracy=0.604637	lr=0.394558
Epoch[13] Batch [749]	Speed: 4214.380667 samples/sec	accuracy=0.604680	lr=0.394507
Epoch[13] Batch [799]	Speed: 4193.173972 samples/sec	accuracy=0.604836	lr=0.394457
Epoch[13] Batch [849]	Speed: 4123.839899 samples/sec	accuracy=0.604978	lr=0.394405
Epoch[13] Batch [899]	Speed: 4183.748615 samples/sec	accuracy=0.605117	lr=0.394354
Epoch[13] Batch [949]	Speed: 4184.047364 samples/sec	accuracy=0.605232	lr=0.394302
Epoch[13] Batch [999]	Speed: 4091.707628 samples/sec	accuracy=0.605324	lr=0.394250
Epoch[13] Batch [1049]	Speed: 4196.646335 samples/sec	accuracy=0.605411	lr=0.394198
Epoch[13] Batch [1099]	Speed: 4152.687502 samples/sec	accuracy=0.605561	lr=0.394146
Epoch[13] Batch [1149]	Speed: 4180.666210 samples/sec	accuracy=0.605633	lr=0.394093
Epoch[13] Batch [1199]	Speed: 4278.870698 samples/sec	accuracy=0.605731	lr=0.394041
Epoch[13] Batch [1249]	Speed: 4221.260375 samples/sec	accuracy=0.605814	lr=0.393988
[Epoch 13] training: accuracy=0.605814
[Epoch 13] speed: 4156 samples/sec	time cost: 318.838241
[Epoch 13] validation: err-top1=0.431063 err-top5=0.189493
Epoch[14] Batch [49]	Speed: 4078.765479 samples/sec	accuracy=0.607528	lr=0.393933
Epoch[14] Batch [99]	Speed: 4274.072212 samples/sec	accuracy=0.606397	lr=0.393880
Epoch[14] Batch [149]	Speed: 4089.271306 samples/sec	accuracy=0.606764	lr=0.393826
Epoch[14] Batch [199]	Speed: 4165.552009 samples/sec	accuracy=0.606623	lr=0.393772
Epoch[14] Batch [249]	Speed: 4295.296573 samples/sec	accuracy=0.606311	lr=0.393718
Epoch[14] Batch [299]	Speed: 4124.325553 samples/sec	accuracy=0.606515	lr=0.393664
Epoch[14] Batch [349]	Speed: 4252.900075 samples/sec	accuracy=0.606508	lr=0.393609
Epoch[14] Batch [399]	Speed: 4124.711259 samples/sec	accuracy=0.606581	lr=0.393554
Epoch[14] Batch [449]	Speed: 4235.164821 samples/sec	accuracy=0.606492	lr=0.393499
Epoch[14] Batch [499]	Speed: 4031.738629 samples/sec	accuracy=0.606419	lr=0.393444
Epoch[14] Batch [549]	Speed: 4115.220711 samples/sec	accuracy=0.606668	lr=0.393388
Epoch[14] Batch [599]	Speed: 4106.519604 samples/sec	accuracy=0.606712	lr=0.393332
Epoch[14] Batch [649]	Speed: 4231.947087 samples/sec	accuracy=0.606703	lr=0.393276
Epoch[14] Batch [699]	Speed: 4058.456061 samples/sec	accuracy=0.606869	lr=0.393220
Epoch[14] Batch [749]	Speed: 4122.495127 samples/sec	accuracy=0.607093	lr=0.393163
Epoch[14] Batch [799]	Speed: 3979.978209 samples/sec	accuracy=0.607208	lr=0.393107
Epoch[14] Batch [849]	Speed: 4147.565554 samples/sec	accuracy=0.607338	lr=0.393050
Epoch[14] Batch [899]	Speed: 4110.577399 samples/sec	accuracy=0.607520	lr=0.392993
Epoch[14] Batch [949]	Speed: 4118.372315 samples/sec	accuracy=0.607590	lr=0.392935
Epoch[14] Batch [999]	Speed: 4225.747690 samples/sec	accuracy=0.607717	lr=0.392877
Epoch[14] Batch [1049]	Speed: 4236.645047 samples/sec	accuracy=0.607846	lr=0.392820
Epoch[14] Batch [1099]	Speed: 4153.963739 samples/sec	accuracy=0.607930	lr=0.392762
Epoch[14] Batch [1149]	Speed: 4085.320340 samples/sec	accuracy=0.608014	lr=0.392703
Epoch[14] Batch [1199]	Speed: 4248.765203 samples/sec	accuracy=0.608120	lr=0.392645
Epoch[14] Batch [1249]	Speed: 4258.572175 samples/sec	accuracy=0.608302	lr=0.392586
[Epoch 14] training: accuracy=0.608303
[Epoch 14] speed: 4149 samples/sec	time cost: 319.548006
[Epoch 14] validation: err-top1=0.425124 err-top5=0.182537
Epoch[15] Batch [49]	Speed: 4115.222051 samples/sec	accuracy=0.609079	lr=0.392526
Epoch[15] Batch [99]	Speed: 4103.694270 samples/sec	accuracy=0.609613	lr=0.392466
Epoch[15] Batch [149]	Speed: 4061.409140 samples/sec	accuracy=0.610345	lr=0.392407
Epoch[15] Batch [199]	Speed: 4187.558432 samples/sec	accuracy=0.610000	lr=0.392347
Epoch[15] Batch [249]	Speed: 4248.890711 samples/sec	accuracy=0.609900	lr=0.392287
Epoch[15] Batch [299]	Speed: 4169.501330 samples/sec	accuracy=0.609662	lr=0.392227
Epoch[15] Batch [349]	Speed: 4124.064734 samples/sec	accuracy=0.609492	lr=0.392167
Epoch[15] Batch [399]	Speed: 4140.566084 samples/sec	accuracy=0.609620	lr=0.392106
Epoch[15] Batch [449]	Speed: 4079.482658 samples/sec	accuracy=0.609802	lr=0.392045
Epoch[15] Batch [499]	Speed: 4224.681357 samples/sec	accuracy=0.609844	lr=0.391984
Epoch[15] Batch [549]	Speed: 4249.630958 samples/sec	accuracy=0.609805	lr=0.391923
Epoch[15] Batch [599]	Speed: 4148.239261 samples/sec	accuracy=0.609715	lr=0.391861
Epoch[15] Batch [649]	Speed: 4103.589976 samples/sec	accuracy=0.609801	lr=0.391799
Epoch[15] Batch [699]	Speed: 4220.205014 samples/sec	accuracy=0.609742	lr=0.391737
Epoch[15] Batch [749]	Speed: 4280.765770 samples/sec	accuracy=0.609918	lr=0.391675
Epoch[15] Batch [799]	Speed: 4299.843577 samples/sec	accuracy=0.610048	lr=0.391613
Epoch[15] Batch [849]	Speed: 4161.708717 samples/sec	accuracy=0.610154	lr=0.391550
Epoch[15] Batch [899]	Speed: 4251.991230 samples/sec	accuracy=0.610325	lr=0.391487
Epoch[15] Batch [949]	Speed: 4194.402306 samples/sec	accuracy=0.610374	lr=0.391424
Epoch[15] Batch [999]	Speed: 4271.503075 samples/sec	accuracy=0.610532	lr=0.391361
Epoch[15] Batch [1049]	Speed: 4147.696849 samples/sec	accuracy=0.610578	lr=0.391297
Epoch[15] Batch [1099]	Speed: 4335.445618 samples/sec	accuracy=0.610695	lr=0.391233
Epoch[15] Batch [1149]	Speed: 4130.702477 samples/sec	accuracy=0.610791	lr=0.391169
Epoch[15] Batch [1199]	Speed: 4235.964379 samples/sec	accuracy=0.610946	lr=0.391105
Epoch[15] Batch [1249]	Speed: 4353.112605 samples/sec	accuracy=0.611068	lr=0.391040
[Epoch 15] training: accuracy=0.611067
[Epoch 15] speed: 4188 samples/sec	time cost: 316.423547
[Epoch 15] validation: err-top1=0.426379 err-top5=0.185029
Epoch[16] Batch [49]	Speed: 4123.683741 samples/sec	accuracy=0.611350	lr=0.390974
Epoch[16] Batch [99]	Speed: 4084.812904 samples/sec	accuracy=0.610746	lr=0.390909
Epoch[16] Batch [149]	Speed: 4166.544236 samples/sec	accuracy=0.610881	lr=0.390844
Epoch[16] Batch [199]	Speed: 4220.429447 samples/sec	accuracy=0.610531	lr=0.390779
Epoch[16] Batch [249]	Speed: 4284.380079 samples/sec	accuracy=0.610432	lr=0.390713
Epoch[16] Batch [299]	Speed: 4106.473195 samples/sec	accuracy=0.610891	lr=0.390647
Epoch[16] Batch [349]	Speed: 4206.718420 samples/sec	accuracy=0.610782	lr=0.390581
Epoch[16] Batch [399]	Speed: 4176.450815 samples/sec	accuracy=0.610992	lr=0.390515
Epoch[16] Batch [449]	Speed: 4181.531056 samples/sec	accuracy=0.611041	lr=0.390448
Epoch[16] Batch [499]	Speed: 4002.399870 samples/sec	accuracy=0.611196	lr=0.390381
Epoch[16] Batch [549]	Speed: 4145.711724 samples/sec	accuracy=0.611319	lr=0.390314
Epoch[16] Batch [599]	Speed: 4224.789902 samples/sec	accuracy=0.611446	lr=0.390247
Epoch[16] Batch [649]	Speed: 4217.033564 samples/sec	accuracy=0.611450	lr=0.390180
Epoch[16] Batch [699]	Speed: 4107.685194 samples/sec	accuracy=0.611543	lr=0.390112
Epoch[16] Batch [749]	Speed: 4180.797981 samples/sec	accuracy=0.611712	lr=0.390044
Epoch[16] Batch [799]	Speed: 4152.682121 samples/sec	accuracy=0.611848	lr=0.389976
Epoch[16] Batch [849]	Speed: 4173.439979 samples/sec	accuracy=0.611942	lr=0.389907
Epoch[16] Batch [899]	Speed: 4093.261754 samples/sec	accuracy=0.612083	lr=0.389839
Epoch[16] Batch [949]	Speed: 4184.960177 samples/sec	accuracy=0.612169	lr=0.389770
Epoch[16] Batch [999]	Speed: 4257.904958 samples/sec	accuracy=0.612271	lr=0.389701
Epoch[16] Batch [1049]	Speed: 4127.939313 samples/sec	accuracy=0.612465	lr=0.389632
Epoch[16] Batch [1099]	Speed: 4167.006850 samples/sec	accuracy=0.612579	lr=0.389562
Epoch[16] Batch [1149]	Speed: 4113.399616 samples/sec	accuracy=0.612643	lr=0.389492
Epoch[16] Batch [1199]	Speed: 4172.023926 samples/sec	accuracy=0.612741	lr=0.389422
Epoch[16] Batch [1249]	Speed: 4423.595181 samples/sec	accuracy=0.612845	lr=0.389352
[Epoch 16] training: accuracy=0.612846
[Epoch 16] speed: 4166 samples/sec	time cost: 317.969617
[Epoch 16] validation: err-top1=0.454301 err-top5=0.203284
Epoch[17] Batch [49]	Speed: 4077.368651 samples/sec	accuracy=0.613011	lr=0.389280
Epoch[17] Batch [99]	Speed: 4097.621736 samples/sec	accuracy=0.612915	lr=0.389210
Epoch[17] Batch [149]	Speed: 4151.002397 samples/sec	accuracy=0.612573	lr=0.389139
Epoch[17] Batch [199]	Speed: 4286.247012 samples/sec	accuracy=0.612585	lr=0.389068
Epoch[17] Batch [249]	Speed: 4118.379976 samples/sec	accuracy=0.612467	lr=0.388996
Epoch[17] Batch [299]	Speed: 4068.877513 samples/sec	accuracy=0.612597	lr=0.388925
Epoch[17] Batch [349]	Speed: 4168.923804 samples/sec	accuracy=0.612660	lr=0.388853
Epoch[17] Batch [399]	Speed: 4113.335245 samples/sec	accuracy=0.612917	lr=0.388781
Epoch[17] Batch [449]	Speed: 4100.087278 samples/sec	accuracy=0.613107	lr=0.388709
Epoch[17] Batch [499]	Speed: 4070.715025 samples/sec	accuracy=0.613268	lr=0.388636
Epoch[17] Batch [549]	Speed: 4266.569945 samples/sec	accuracy=0.613240	lr=0.388564
Epoch[17] Batch [599]	Speed: 4273.038830 samples/sec	accuracy=0.613248	lr=0.388491
Epoch[17] Batch [649]	Speed: 4130.082031 samples/sec	accuracy=0.613223	lr=0.388418
Epoch[17] Batch [699]	Speed: 4114.883375 samples/sec	accuracy=0.613296	lr=0.388344
Epoch[17] Batch [749]	Speed: 4133.879519 samples/sec	accuracy=0.613461	lr=0.388271
Epoch[17] Batch [799]	Speed: 4153.353476 samples/sec	accuracy=0.613610	lr=0.388197
Epoch[17] Batch [849]	Speed: 4253.484002 samples/sec	accuracy=0.613764	lr=0.388123
Epoch[17] Batch [899]	Speed: 4054.237460 samples/sec	accuracy=0.613864	lr=0.388049
Epoch[17] Batch [949]	Speed: 4166.174995 samples/sec	accuracy=0.614040	lr=0.387974
Epoch[17] Batch [999]	Speed: 4017.245249 samples/sec	accuracy=0.614093	lr=0.387900
Epoch[17] Batch [1049]	Speed: 4068.321049 samples/sec	accuracy=0.614196	lr=0.387825
Epoch[17] Batch [1099]	Speed: 4041.272761 samples/sec	accuracy=0.614280	lr=0.387750
Epoch[17] Batch [1149]	Speed: 4167.016633 samples/sec	accuracy=0.614327	lr=0.387674
Epoch[17] Batch [1199]	Speed: 4119.206284 samples/sec	accuracy=0.614314	lr=0.387599
Epoch[17] Batch [1249]	Speed: 4289.754901 samples/sec	accuracy=0.614420	lr=0.387523
[Epoch 17] training: accuracy=0.614426
[Epoch 17] speed: 4135 samples/sec	time cost: 320.166342
[Epoch 17] validation: err-top1=0.421834 err-top5=0.182170
Epoch[18] Batch [49]	Speed: 4160.051743 samples/sec	accuracy=0.614719	lr=0.387445
Epoch[18] Batch [99]	Speed: 4063.329171 samples/sec	accuracy=0.614551	lr=0.387369
Epoch[18] Batch [149]	Speed: 4138.716511 samples/sec	accuracy=0.614046	lr=0.387293
Epoch[18] Batch [199]	Speed: 4140.406740 samples/sec	accuracy=0.614209	lr=0.387216
Epoch[18] Batch [249]	Speed: 4132.818159 samples/sec	accuracy=0.614048	lr=0.387139
Epoch[18] Batch [299]	Speed: 4041.648261 samples/sec	accuracy=0.614240	lr=0.387062
Epoch[18] Batch [349]	Speed: 4138.731187 samples/sec	accuracy=0.614176	lr=0.386984
Epoch[18] Batch [399]	Speed: 4024.489336 samples/sec	accuracy=0.614162	lr=0.386907
Epoch[18] Batch [449]	Speed: 4156.991349 samples/sec	accuracy=0.614338	lr=0.386829
Epoch[18] Batch [499]	Speed: 4214.379179 samples/sec	accuracy=0.614437	lr=0.386751
Epoch[18] Batch [549]	Speed: 4198.522203 samples/sec	accuracy=0.614559	lr=0.386673
Epoch[18] Batch [599]	Speed: 4175.141000 samples/sec	accuracy=0.614676	lr=0.386594
Epoch[18] Batch [649]	Speed: 4068.158124 samples/sec	accuracy=0.614812	lr=0.386515
Epoch[18] Batch [699]	Speed: 4295.156025 samples/sec	accuracy=0.614966	lr=0.386436
Epoch[18] Batch [749]	Speed: 4232.853890 samples/sec	accuracy=0.615055	lr=0.386357
Epoch[18] Batch [799]	Speed: 4113.810153 samples/sec	accuracy=0.615105	lr=0.386278
Epoch[18] Batch [849]	Speed: 4262.829399 samples/sec	accuracy=0.615181	lr=0.386198
Epoch[18] Batch [899]	Speed: 4160.945327 samples/sec	accuracy=0.615161	lr=0.386119
Epoch[18] Batch [949]	Speed: 4132.066683 samples/sec	accuracy=0.615202	lr=0.386038
Epoch[18] Batch [999]	Speed: 4129.528871 samples/sec	accuracy=0.615306	lr=0.385958
Epoch[18] Batch [1049]	Speed: 3996.578285 samples/sec	accuracy=0.615373	lr=0.385878
Epoch[18] Batch [1099]	Speed: 4204.398827 samples/sec	accuracy=0.615483	lr=0.385797
Epoch[18] Batch [1149]	Speed: 4179.372132 samples/sec	accuracy=0.615596	lr=0.385716
Epoch[18] Batch [1199]	Speed: 4073.978643 samples/sec	accuracy=0.615723	lr=0.385635
Epoch[18] Batch [1249]	Speed: 4241.852180 samples/sec	accuracy=0.615843	lr=0.385554
[Epoch 18] training: accuracy=0.615844
[Epoch 18] speed: 4142 samples/sec	time cost: 319.839187
[Epoch 18] validation: err-top1=0.417769 err-top5=0.177535
Epoch[19] Batch [49]	Speed: 4121.884818 samples/sec	accuracy=0.615461	lr=0.385470
Epoch[19] Batch [99]	Speed: 4186.994097 samples/sec	accuracy=0.615216	lr=0.385388
Epoch[19] Batch [149]	Speed: 4224.688338 samples/sec	accuracy=0.615384	lr=0.385306
Epoch[19] Batch [199]	Speed: 4174.845469 samples/sec	accuracy=0.616010	lr=0.385224
Epoch[19] Batch [249]	Speed: 4224.579631 samples/sec	accuracy=0.615981	lr=0.385142
Epoch[19] Batch [299]	Speed: 4237.562729 samples/sec	accuracy=0.616244	lr=0.385059
Epoch[19] Batch [349]	Speed: 4300.885228 samples/sec	accuracy=0.616649	lr=0.384976
Epoch[19] Batch [399]	Speed: 4184.683478 samples/sec	accuracy=0.616747	lr=0.384893
Epoch[19] Batch [449]	Speed: 4197.217377 samples/sec	accuracy=0.616897	lr=0.384810
Epoch[19] Batch [499]	Speed: 4192.052982 samples/sec	accuracy=0.617042	lr=0.384726
Epoch[19] Batch [549]	Speed: 4240.467284 samples/sec	accuracy=0.617086	lr=0.384642
Epoch[19] Batch [599]	Speed: 4136.015107 samples/sec	accuracy=0.617162	lr=0.384558
Epoch[19] Batch [649]	Speed: 4191.529322 samples/sec	accuracy=0.617222	lr=0.384474
Epoch[19] Batch [699]	Speed: 4203.810523 samples/sec	accuracy=0.617328	lr=0.384389
Epoch[19] Batch [749]	Speed: 4441.486885 samples/sec	accuracy=0.617319	lr=0.384305
Epoch[19] Batch [799]	Speed: 3980.330600 samples/sec	accuracy=0.617413	lr=0.384220
Epoch[19] Batch [849]	Speed: 4103.914953 samples/sec	accuracy=0.617493	lr=0.384135
Epoch[19] Batch [899]	Speed: 4108.411951 samples/sec	accuracy=0.617642	lr=0.384049
Epoch[19] Batch [949]	Speed: 4166.076068 samples/sec	accuracy=0.617760	lr=0.383964
Epoch[19] Batch [999]	Speed: 4202.463592 samples/sec	accuracy=0.617851	lr=0.383878
Epoch[19] Batch [1049]	Speed: 4239.852437 samples/sec	accuracy=0.617955	lr=0.383792
Epoch[19] Batch [1099]	Speed: 4197.910842 samples/sec	accuracy=0.618002	lr=0.383706
Epoch[19] Batch [1149]	Speed: 4129.313366 samples/sec	accuracy=0.618089	lr=0.383619
Epoch[19] Batch [1199]	Speed: 4051.595185 samples/sec	accuracy=0.618214	lr=0.383533
Epoch[19] Batch [1249]	Speed: 4229.138389 samples/sec	accuracy=0.618321	lr=0.383446
[Epoch 19] training: accuracy=0.618321
[Epoch 19] speed: 4181 samples/sec	time cost: 316.372024
[Epoch 19] validation: err-top1=0.405792 err-top5=0.169703
Epoch[20] Batch [49]	Speed: 4121.521789 samples/sec	accuracy=0.617381	lr=0.383357
Epoch[20] Batch [99]	Speed: 4236.960259 samples/sec	accuracy=0.617194	lr=0.383270
Epoch[20] Batch [149]	Speed: 4109.196835 samples/sec	accuracy=0.617558	lr=0.383182
Epoch[20] Batch [199]	Speed: 4275.560527 samples/sec	accuracy=0.617338	lr=0.383094
Epoch[20] Batch [249]	Speed: 4008.263877 samples/sec	accuracy=0.617211	lr=0.383006
Epoch[20] Batch [299]	Speed: 4154.092868 samples/sec	accuracy=0.616835	lr=0.382918
Epoch[20] Batch [349]	Speed: 4205.637948 samples/sec	accuracy=0.616901	lr=0.382830
Epoch[20] Batch [399]	Speed: 4167.220485 samples/sec	accuracy=0.616850	lr=0.382741
Epoch[20] Batch [449]	Speed: 4100.070291 samples/sec	accuracy=0.616964	lr=0.382652
Epoch[20] Batch [499]	Speed: 4212.724880 samples/sec	accuracy=0.617147	lr=0.382563
Epoch[20] Batch [549]	Speed: 4213.069109 samples/sec	accuracy=0.617294	lr=0.382474
Epoch[20] Batch [599]	Speed: 4160.004519 samples/sec	accuracy=0.617464	lr=0.382384
Epoch[20] Batch [649]	Speed: 4120.729660 samples/sec	accuracy=0.617628	lr=0.382295
Epoch[20] Batch [699]	Speed: 4240.381962 samples/sec	accuracy=0.617686	lr=0.382205
Epoch[20] Batch [749]	Speed: 4272.700714 samples/sec	accuracy=0.617899	lr=0.382115
Epoch[20] Batch [799]	Speed: 4210.505565 samples/sec	accuracy=0.618001	lr=0.382024
Epoch[20] Batch [849]	Speed: 4175.031338 samples/sec	accuracy=0.618073	lr=0.381934
Epoch[20] Batch [899]	Speed: 4034.996658 samples/sec	accuracy=0.618140	lr=0.381843
Epoch[20] Batch [949]	Speed: 4228.917192 samples/sec	accuracy=0.618266	lr=0.381752
Epoch[20] Batch [999]	Speed: 4092.197830 samples/sec	accuracy=0.618394	lr=0.381661
Epoch[20] Batch [1049]	Speed: 4207.746600 samples/sec	accuracy=0.618539	lr=0.381569
Epoch[20] Batch [1099]	Speed: 4307.824342 samples/sec	accuracy=0.618633	lr=0.381477
Epoch[20] Batch [1149]	Speed: 4161.784450 samples/sec	accuracy=0.618706	lr=0.381385
Epoch[20] Batch [1199]	Speed: 4142.103944 samples/sec	accuracy=0.618866	lr=0.381293
Epoch[20] Batch [1249]	Speed: 4331.073725 samples/sec	accuracy=0.618989	lr=0.381201
[Epoch 20] training: accuracy=0.618992
[Epoch 20] speed: 4174 samples/sec	time cost: 317.018601
[Epoch 20] validation: err-top1=0.425821 err-top5=0.182079
Epoch[21] Batch [49]	Speed: 4170.682946 samples/sec	accuracy=0.618832	lr=0.381107
Epoch[21] Batch [99]	Speed: 4218.619729 samples/sec	accuracy=0.618409	lr=0.381014
Epoch[21] Batch [149]	Speed: 4223.385058 samples/sec	accuracy=0.618854	lr=0.380921
Epoch[21] Batch [199]	Speed: 4142.061042 samples/sec	accuracy=0.619122	lr=0.380828
Epoch[21] Batch [249]	Speed: 4124.507664 samples/sec	accuracy=0.619144	lr=0.380734
Epoch[21] Batch [299]	Speed: 4207.440995 samples/sec	accuracy=0.618938	lr=0.380641
Epoch[21] Batch [349]	Speed: 4201.463474 samples/sec	accuracy=0.619019	lr=0.380547
Epoch[21] Batch [399]	Speed: 4182.276036 samples/sec	accuracy=0.619300	lr=0.380453
Epoch[21] Batch [449]	Speed: 4096.513189 samples/sec	accuracy=0.619291	lr=0.380359
Epoch[21] Batch [499]	Speed: 4124.703971 samples/sec	accuracy=0.619219	lr=0.380264
Epoch[21] Batch [549]	Speed: 4131.709171 samples/sec	accuracy=0.619236	lr=0.380169
Epoch[21] Batch [599]	Speed: 4189.348050 samples/sec	accuracy=0.619224	lr=0.380074
Epoch[21] Batch [649]	Speed: 4123.552932 samples/sec	accuracy=0.619359	lr=0.379979
Epoch[21] Batch [699]	Speed: 4271.116782 samples/sec	accuracy=0.619490	lr=0.379884
Epoch[21] Batch [749]	Speed: 4025.735449 samples/sec	accuracy=0.619483	lr=0.379788
Epoch[21] Batch [799]	Speed: 4161.299772 samples/sec	accuracy=0.619508	lr=0.379693
Epoch[21] Batch [849]	Speed: 4146.541031 samples/sec	accuracy=0.619529	lr=0.379597
Epoch[21] Batch [899]	Speed: 4224.380932 samples/sec	accuracy=0.619605	lr=0.379501
Epoch[21] Batch [949]	Speed: 4226.406612 samples/sec	accuracy=0.619693	lr=0.379404
Epoch[21] Batch [999]	Speed: 4068.009776 samples/sec	accuracy=0.619821	lr=0.379307
Epoch[21] Batch [1049]	Speed: 4269.513057 samples/sec	accuracy=0.619962	lr=0.379211
Epoch[21] Batch [1099]	Speed: 4214.571066 samples/sec	accuracy=0.619987	lr=0.379114
Epoch[21] Batch [1149]	Speed: 4271.654145 samples/sec	accuracy=0.620107	lr=0.379016
Epoch[21] Batch [1199]	Speed: 4020.553503 samples/sec	accuracy=0.620232	lr=0.378919
Epoch[21] Batch [1249]	Speed: 4369.704982 samples/sec	accuracy=0.620376	lr=0.378821
[Epoch 21] training: accuracy=0.620386
[Epoch 21] speed: 4170 samples/sec	time cost: 317.574197
[Epoch 21] validation: err-top1=0.402065 err-top5=0.168068
Epoch[22] Batch [49]	Speed: 3974.968867 samples/sec	accuracy=0.619884	lr=0.378721
Epoch[22] Batch [99]	Speed: 4093.723922 samples/sec	accuracy=0.619719	lr=0.378623
Epoch[22] Batch [149]	Speed: 4102.543719 samples/sec	accuracy=0.619684	lr=0.378525
Epoch[22] Batch [199]	Speed: 4104.267592 samples/sec	accuracy=0.619696	lr=0.378426
Epoch[22] Batch [249]	Speed: 4126.205491 samples/sec	accuracy=0.619722	lr=0.378328
Epoch[22] Batch [299]	Speed: 4210.397835 samples/sec	accuracy=0.619893	lr=0.378229
Epoch[22] Batch [349]	Speed: 4145.263989 samples/sec	accuracy=0.619870	lr=0.378129
Epoch[22] Batch [399]	Speed: 4101.303812 samples/sec	accuracy=0.619871	lr=0.378030
Epoch[22] Batch [449]	Speed: 4145.554226 samples/sec	accuracy=0.619933	lr=0.377930
Epoch[22] Batch [499]	Speed: 4330.115100 samples/sec	accuracy=0.620166	lr=0.377831
Epoch[22] Batch [549]	Speed: 4153.848839 samples/sec	accuracy=0.620176	lr=0.377731
Epoch[22] Batch [599]	Speed: 4259.074204 samples/sec	accuracy=0.620312	lr=0.377630
Epoch[22] Batch [649]	Speed: 4286.461585 samples/sec	accuracy=0.620530	lr=0.377530
Epoch[22] Batch [699]	Speed: 4192.606322 samples/sec	accuracy=0.620583	lr=0.377429
Epoch[22] Batch [749]	Speed: 4227.199455 samples/sec	accuracy=0.620763	lr=0.377328
Epoch[22] Batch [799]	Speed: 4207.051695 samples/sec	accuracy=0.620899	lr=0.377227
Epoch[22] Batch [849]	Speed: 4140.134864 samples/sec	accuracy=0.620899	lr=0.377126
Epoch[22] Batch [899]	Speed: 4212.950502 samples/sec	accuracy=0.621012	lr=0.377024
Epoch[22] Batch [949]	Speed: 4052.360267 samples/sec	accuracy=0.621081	lr=0.376923
Epoch[22] Batch [999]	Speed: 4252.298458 samples/sec	accuracy=0.621144	lr=0.376821
Epoch[22] Batch [1049]	Speed: 4135.996387 samples/sec	accuracy=0.621165	lr=0.376719
Epoch[22] Batch [1099]	Speed: 4071.635638 samples/sec	accuracy=0.621244	lr=0.376616
Epoch[22] Batch [1149]	Speed: 4208.737256 samples/sec	accuracy=0.621339	lr=0.376514
Epoch[22] Batch [1199]	Speed: 4109.856798 samples/sec	accuracy=0.621383	lr=0.376411
Epoch[22] Batch [1249]	Speed: 4242.322869 samples/sec	accuracy=0.621501	lr=0.376308
[Epoch 22] training: accuracy=0.621506
[Epoch 22] speed: 4157 samples/sec	time cost: 317.898938
[Epoch 22] validation: err-top1=0.400610 err-top5=0.162787
Epoch[23] Batch [49]	Speed: 4125.041888 samples/sec	accuracy=0.622651	lr=0.376203
Epoch[23] Batch [99]	Speed: 4179.635520 samples/sec	accuracy=0.621899	lr=0.376099
Epoch[23] Batch [149]	Speed: 4174.127231 samples/sec	accuracy=0.621921	lr=0.375996
Epoch[23] Batch [199]	Speed: 4264.955741 samples/sec	accuracy=0.621795	lr=0.375892
Epoch[23] Batch [249]	Speed: 4244.537437 samples/sec	accuracy=0.621757	lr=0.375788
Epoch[23] Batch [299]	Speed: 4145.743978 samples/sec	accuracy=0.621582	lr=0.375683
Epoch[23] Batch [349]	Speed: 4210.056849 samples/sec	accuracy=0.621485	lr=0.375579
Epoch[23] Batch [399]	Speed: 4182.505251 samples/sec	accuracy=0.621429	lr=0.375474
Epoch[23] Batch [449]	Speed: 4151.615019 samples/sec	accuracy=0.621390	lr=0.375369
Epoch[23] Batch [499]	Speed: 4005.434839 samples/sec	accuracy=0.621504	lr=0.375264
Epoch[23] Batch [549]	Speed: 4239.831342 samples/sec	accuracy=0.621569	lr=0.375159
Epoch[23] Batch [599]	Speed: 4180.856992 samples/sec	accuracy=0.621597	lr=0.375053
Epoch[23] Batch [649]	Speed: 4222.806127 samples/sec	accuracy=0.621550	lr=0.374948
Epoch[23] Batch [699]	Speed: 3989.543634 samples/sec	accuracy=0.621612	lr=0.374842
Epoch[23] Batch [749]	Speed: 4141.467370 samples/sec	accuracy=0.621771	lr=0.374736
Epoch[23] Batch [799]	Speed: 4242.753427 samples/sec	accuracy=0.621912	lr=0.374629
Epoch[23] Batch [849]	Speed: 4196.735730 samples/sec	accuracy=0.621956	lr=0.374523
Epoch[23] Batch [899]	Speed: 4064.515679 samples/sec	accuracy=0.622027	lr=0.374416
Epoch[23] Batch [949]	Speed: 4082.548151 samples/sec	accuracy=0.622088	lr=0.374309
Epoch[23] Batch [999]	Speed: 4171.693179 samples/sec	accuracy=0.622175	lr=0.374202
Epoch[23] Batch [1049]	Speed: 4094.876090 samples/sec	accuracy=0.622302	lr=0.374095
Epoch[23] Batch [1099]	Speed: 4118.139414 samples/sec	accuracy=0.622300	lr=0.373987
Epoch[23] Batch [1149]	Speed: 4202.729488 samples/sec	accuracy=0.622445	lr=0.373879
Epoch[23] Batch [1199]	Speed: 4221.896983 samples/sec	accuracy=0.622582	lr=0.373771
Epoch[23] Batch [1249]	Speed: 4235.810225 samples/sec	accuracy=0.622727	lr=0.373663
[Epoch 23] training: accuracy=0.622732
[Epoch 23] speed: 4158 samples/sec	time cost: 318.586480
[Epoch 23] validation: err-top1=0.410238 err-top5=0.172872
Epoch[24] Batch [49]	Speed: 4140.149550 samples/sec	accuracy=0.624372	lr=0.373552
Epoch[24] Batch [99]	Speed: 4073.775774 samples/sec	accuracy=0.623860	lr=0.373444
Epoch[24] Batch [149]	Speed: 4188.456277 samples/sec	accuracy=0.623631	lr=0.373335
Epoch[24] Batch [199]	Speed: 4093.819131 samples/sec	accuracy=0.623625	lr=0.373226
Epoch[24] Batch [249]	Speed: 4076.778130 samples/sec	accuracy=0.623559	lr=0.373117
Epoch[24] Batch [299]	Speed: 4109.049332 samples/sec	accuracy=0.623642	lr=0.373007
Epoch[24] Batch [349]	Speed: 4172.542804 samples/sec	accuracy=0.623732	lr=0.372897
Epoch[24] Batch [399]	Speed: 4087.875911 samples/sec	accuracy=0.623719	lr=0.372788
Epoch[24] Batch [449]	Speed: 4165.405281 samples/sec	accuracy=0.623574	lr=0.372678
Epoch[24] Batch [499]	Speed: 4119.777072 samples/sec	accuracy=0.623576	lr=0.372567
Epoch[24] Batch [549]	Speed: 4305.930367 samples/sec	accuracy=0.623428	lr=0.372457
Epoch[24] Batch [599]	Speed: 4270.321732 samples/sec	accuracy=0.623454	lr=0.372346
Epoch[24] Batch [649]	Speed: 4127.749997 samples/sec	accuracy=0.623487	lr=0.372235
Epoch[24] Batch [699]	Speed: 4019.910545 samples/sec	accuracy=0.623634	lr=0.372124
Epoch[24] Batch [749]	Speed: 4163.449909 samples/sec	accuracy=0.623662	lr=0.372013
Epoch[24] Batch [799]	Speed: 4162.827737 samples/sec	accuracy=0.623734	lr=0.371901
Epoch[24] Batch [849]	Speed: 4178.948405 samples/sec	accuracy=0.623797	lr=0.371790
Epoch[24] Batch [899]	Speed: 4215.889105 samples/sec	accuracy=0.623773	lr=0.371678
Epoch[24] Batch [949]	Speed: 4321.948007 samples/sec	accuracy=0.623891	lr=0.371565
Epoch[24] Batch [999]	Speed: 4158.681321 samples/sec	accuracy=0.623947	lr=0.371453
Epoch[24] Batch [1049]	Speed: 4240.668923 samples/sec	accuracy=0.624055	lr=0.371341
Epoch[24] Batch [1099]	Speed: 4083.405176 samples/sec	accuracy=0.624185	lr=0.371228
Epoch[24] Batch [1149]	Speed: 4173.093438 samples/sec	accuracy=0.624238	lr=0.371115
Epoch[24] Batch [1199]	Speed: 4157.103123 samples/sec	accuracy=0.624278	lr=0.371002
Epoch[24] Batch [1249]	Speed: 4319.748484 samples/sec	accuracy=0.624322	lr=0.370888
[Epoch 24] training: accuracy=0.624323
[Epoch 24] speed: 4160 samples/sec	time cost: 318.621828
[Epoch 24] validation: err-top1=0.395189 err-top5=0.162109
Epoch[25] Batch [49]	Speed: 4070.117327 samples/sec	accuracy=0.624649	lr=0.370773
Epoch[25] Batch [99]	Speed: 4156.903962 samples/sec	accuracy=0.624370	lr=0.370659
Epoch[25] Batch [149]	Speed: 4183.446078 samples/sec	accuracy=0.624158	lr=0.370545
Epoch[25] Batch [199]	Speed: 4180.464784 samples/sec	accuracy=0.624409	lr=0.370431
Epoch[25] Batch [249]	Speed: 4183.206492 samples/sec	accuracy=0.624272	lr=0.370316
Epoch[25] Batch [299]	Speed: 4209.767826 samples/sec	accuracy=0.624171	lr=0.370202
Epoch[25] Batch [349]	Speed: 4243.547132 samples/sec	accuracy=0.624393	lr=0.370087
Epoch[25] Batch [399]	Speed: 4356.107354 samples/sec	accuracy=0.624460	lr=0.369972
Epoch[25] Batch [449]	Speed: 4157.049127 samples/sec	accuracy=0.624356	lr=0.369857
Epoch[25] Batch [499]	Speed: 4136.838546 samples/sec	accuracy=0.624492	lr=0.369741
Epoch[25] Batch [549]	Speed: 4227.242725 samples/sec	accuracy=0.624564	lr=0.369626
Epoch[25] Batch [599]	Speed: 4223.658177 samples/sec	accuracy=0.624552	lr=0.369510
Epoch[25] Batch [649]	Speed: 4245.475412 samples/sec	accuracy=0.624672	lr=0.369394
Epoch[25] Batch [699]	Speed: 4137.674271 samples/sec	accuracy=0.624754	lr=0.369278
Epoch[25] Batch [749]	Speed: 4127.527379 samples/sec	accuracy=0.624918	lr=0.369161
Epoch[25] Batch [799]	Speed: 4214.742455 samples/sec	accuracy=0.625025	lr=0.369045
Epoch[25] Batch [849]	Speed: 4226.217888 samples/sec	accuracy=0.625096	lr=0.368928
Epoch[25] Batch [899]	Speed: 4122.596665 samples/sec	accuracy=0.625144	lr=0.368811
Epoch[25] Batch [949]	Speed: 4168.680457 samples/sec	accuracy=0.625274	lr=0.368694
Epoch[25] Batch [999]	Speed: 4236.483489 samples/sec	accuracy=0.625384	lr=0.368576
Epoch[25] Batch [1049]	Speed: 4165.142713 samples/sec	accuracy=0.625487	lr=0.368459
Epoch[25] Batch [1099]	Speed: 4302.773732 samples/sec	accuracy=0.625630	lr=0.368341
Epoch[25] Batch [1149]	Speed: 4153.708959 samples/sec	accuracy=0.625752	lr=0.368223
Epoch[25] Batch [1199]	Speed: 4137.405144 samples/sec	accuracy=0.625805	lr=0.368105
Epoch[25] Batch [1249]	Speed: 4262.933059 samples/sec	accuracy=0.625875	lr=0.367986
[Epoch 25] training: accuracy=0.625874
[Epoch 25] speed: 4188 samples/sec	time cost: 316.322091
[Epoch 25] validation: err-top1=0.398099 err-top5=0.162628
Epoch[26] Batch [49]	Speed: 4157.302304 samples/sec	accuracy=0.626022	lr=0.367865
Epoch[26] Batch [99]	Speed: 4214.257108 samples/sec	accuracy=0.625367	lr=0.367747
Epoch[26] Batch [149]	Speed: 4149.886436 samples/sec	accuracy=0.625090	lr=0.367628
Epoch[26] Batch [199]	Speed: 4125.813799 samples/sec	accuracy=0.625521	lr=0.367508
Epoch[26] Batch [249]	Speed: 4262.708398 samples/sec	accuracy=0.625527	lr=0.367389
Epoch[26] Batch [299]	Speed: 4287.993732 samples/sec	accuracy=0.625248	lr=0.367269
Epoch[26] Batch [349]	Speed: 4236.692105 samples/sec	accuracy=0.625161	lr=0.367150
Epoch[26] Batch [399]	Speed: 4269.739455 samples/sec	accuracy=0.625309	lr=0.367030
Epoch[26] Batch [449]	Speed: 4132.065649 samples/sec	accuracy=0.625407	lr=0.366909
Epoch[26] Batch [499]	Speed: 4165.723880 samples/sec	accuracy=0.625561	lr=0.366789
Epoch[26] Batch [549]	Speed: 4092.381013 samples/sec	accuracy=0.625705	lr=0.366668
Epoch[26] Batch [599]	Speed: 4201.691345 samples/sec	accuracy=0.625696	lr=0.366548
Epoch[26] Batch [649]	Speed: 4191.680270 samples/sec	accuracy=0.625756	lr=0.366426
Epoch[26] Batch [699]	Speed: 4251.231477 samples/sec	accuracy=0.625886	lr=0.366305
Epoch[26] Batch [749]	Speed: 4171.345469 samples/sec	accuracy=0.625984	lr=0.366184
Epoch[26] Batch [799]	Speed: 4243.715268 samples/sec	accuracy=0.626163	lr=0.366062
Epoch[26] Batch [849]	Speed: 4268.190377 samples/sec	accuracy=0.626321	lr=0.365940
Epoch[26] Batch [899]	Speed: 4267.546517 samples/sec	accuracy=0.626381	lr=0.365819
Epoch[26] Batch [949]	Speed: 4165.070250 samples/sec	accuracy=0.626484	lr=0.365696
Epoch[26] Batch [999]	Speed: 4185.582456 samples/sec	accuracy=0.626527	lr=0.365574
Epoch[26] Batch [1049]	Speed: 4199.569128 samples/sec	accuracy=0.626618	lr=0.365451
Epoch[26] Batch [1099]	Speed: 4063.561066 samples/sec	accuracy=0.626687	lr=0.365329
Epoch[26] Batch [1149]	Speed: 4119.871837 samples/sec	accuracy=0.626747	lr=0.365206
Epoch[26] Batch [1199]	Speed: 4195.099348 samples/sec	accuracy=0.626795	lr=0.365082
Epoch[26] Batch [1249]	Speed: 4336.273685 samples/sec	accuracy=0.626855	lr=0.364959
[Epoch 26] training: accuracy=0.626855
[Epoch 26] speed: 4193 samples/sec	time cost: 315.444551
[Epoch 26] validation: err-top1=0.394950 err-top5=0.157685
Epoch[27] Batch [49]	Speed: 4136.604587 samples/sec	accuracy=0.626692	lr=0.364833
Epoch[27] Batch [99]	Speed: 4152.108923 samples/sec	accuracy=0.625891	lr=0.364709
Epoch[27] Batch [149]	Speed: 4094.558398 samples/sec	accuracy=0.625404	lr=0.364585
Epoch[27] Batch [199]	Speed: 4221.828425 samples/sec	accuracy=0.625696	lr=0.364461
Epoch[27] Batch [249]	Speed: 4147.461340 samples/sec	accuracy=0.625760	lr=0.364337
Epoch[27] Batch [299]	Speed: 4086.351687 samples/sec	accuracy=0.625975	lr=0.364212
Epoch[27] Batch [349]	Speed: 4196.729579 samples/sec	accuracy=0.626234	lr=0.364087
Epoch[27] Batch [399]	Speed: 4211.428144 samples/sec	accuracy=0.626328	lr=0.363962
Epoch[27] Batch [449]	Speed: 4043.903184 samples/sec	accuracy=0.626429	lr=0.363837
Epoch[27] Batch [499]	Speed: 4135.925572 samples/sec	accuracy=0.626452	lr=0.363712
Epoch[27] Batch [549]	Speed: 4091.771401 samples/sec	accuracy=0.626604	lr=0.363586
Epoch[27] Batch [599]	Speed: 4042.116955 samples/sec	accuracy=0.626764	lr=0.363461
Epoch[27] Batch [649]	Speed: 4187.673571 samples/sec	accuracy=0.626882	lr=0.363335
Epoch[27] Batch [699]	Speed: 4205.748812 samples/sec	accuracy=0.626840	lr=0.363209
Epoch[27] Batch [749]	Speed: 4132.455508 samples/sec	accuracy=0.627028	lr=0.363082
Epoch[27] Batch [799]	Speed: 4194.043592 samples/sec	accuracy=0.627112	lr=0.362956
Epoch[27] Batch [849]	Speed: 4067.409714 samples/sec	accuracy=0.627219	lr=0.362829
Epoch[27] Batch [899]	Speed: 4155.790858 samples/sec	accuracy=0.627284	lr=0.362702
Epoch[27] Batch [949]	Speed: 4158.519855 samples/sec	accuracy=0.627409	lr=0.362575
Epoch[27] Batch [999]	Speed: 4143.257690 samples/sec	accuracy=0.627495	lr=0.362448
Epoch[27] Batch [1049]	Speed: 4188.645810 samples/sec	accuracy=0.627596	lr=0.362320
Epoch[27] Batch [1099]	Speed: 4249.483544 samples/sec	accuracy=0.627641	lr=0.362193
Epoch[27] Batch [1149]	Speed: 4253.212320 samples/sec	accuracy=0.627737	lr=0.362065
Epoch[27] Batch [1199]	Speed: 4264.583419 samples/sec	accuracy=0.627781	lr=0.361937
Epoch[27] Batch [1249]	Speed: 4338.909866 samples/sec	accuracy=0.627842	lr=0.361809
[Epoch 27] training: accuracy=0.627845
[Epoch 27] speed: 4158 samples/sec	time cost: 317.866781
[Epoch 27] validation: err-top1=0.381437 err-top5=0.149952
Epoch[28] Batch [49]	Speed: 4111.600363 samples/sec	accuracy=0.627744	lr=0.361678
Epoch[28] Batch [99]	Speed: 4159.434373 samples/sec	accuracy=0.627533	lr=0.361549
Epoch[28] Batch [149]	Speed: 4421.013849 samples/sec	accuracy=0.627643	lr=0.361420
Epoch[28] Batch [199]	Speed: 4132.851167 samples/sec	accuracy=0.627588	lr=0.361291
Epoch[28] Batch [249]	Speed: 4247.989040 samples/sec	accuracy=0.627384	lr=0.361162
Epoch[28] Batch [299]	Speed: 4098.820613 samples/sec	accuracy=0.627248	lr=0.361032
Epoch[28] Batch [349]	Speed: 4197.164466 samples/sec	accuracy=0.627143	lr=0.360903
Epoch[28] Batch [399]	Speed: 4159.028936 samples/sec	accuracy=0.627216	lr=0.360773
Epoch[28] Batch [449]	Speed: 4106.157392 samples/sec	accuracy=0.627208	lr=0.360643
Epoch[28] Batch [499]	Speed: 4225.338950 samples/sec	accuracy=0.627242	lr=0.360513
Epoch[28] Batch [549]	Speed: 4207.657725 samples/sec	accuracy=0.627379	lr=0.360382
Epoch[28] Batch [599]	Speed: 4292.134745 samples/sec	accuracy=0.627451	lr=0.360252
Epoch[28] Batch [649]	Speed: 3998.157069 samples/sec	accuracy=0.627560	lr=0.360121
Epoch[28] Batch [699]	Speed: 4164.739232 samples/sec	accuracy=0.627570	lr=0.359990
Epoch[28] Batch [749]	Speed: 4125.216771 samples/sec	accuracy=0.627649	lr=0.359859
Epoch[28] Batch [799]	Speed: 4122.954263 samples/sec	accuracy=0.627911	lr=0.359728
Epoch[28] Batch [849]	Speed: 4174.078632 samples/sec	accuracy=0.627945	lr=0.359596
Epoch[28] Batch [899]	Speed: 4225.763405 samples/sec	accuracy=0.627955	lr=0.359465
Epoch[28] Batch [949]	Speed: 4216.128642 samples/sec	accuracy=0.628016	lr=0.359333
Epoch[28] Batch [999]	Speed: 4148.898119 samples/sec	accuracy=0.628156	lr=0.359201
Epoch[28] Batch [1049]	Speed: 4180.033757 samples/sec	accuracy=0.628222	lr=0.359068
Epoch[28] Batch [1099]	Speed: 4219.994287 samples/sec	accuracy=0.628312	lr=0.358936
Epoch[28] Batch [1149]	Speed: 4152.487478 samples/sec	accuracy=0.628455	lr=0.358803
Epoch[28] Batch [1199]	Speed: 4340.014652 samples/sec	accuracy=0.628561	lr=0.358670
Epoch[28] Batch [1249]	Speed: 4295.865647 samples/sec	accuracy=0.628642	lr=0.358537
[Epoch 28] training: accuracy=0.628648
[Epoch 28] speed: 4183 samples/sec	time cost: 316.407421
[Epoch 28] validation: err-top1=0.387974 err-top5=0.154397
Epoch[29] Batch [49]	Speed: 4121.076258 samples/sec	accuracy=0.630460	lr=0.358401
Epoch[29] Batch [99]	Speed: 4211.458620 samples/sec	accuracy=0.629422	lr=0.358268
Epoch[29] Batch [149]	Speed: 4212.543904 samples/sec	accuracy=0.628945	lr=0.358134
Epoch[29] Batch [199]	Speed: 4163.927983 samples/sec	accuracy=0.628406	lr=0.358001
Epoch[29] Batch [249]	Speed: 4220.850761 samples/sec	accuracy=0.628358	lr=0.357867
Epoch[29] Batch [299]	Speed: 4259.880703 samples/sec	accuracy=0.628134	lr=0.357732
Epoch[29] Batch [349]	Speed: 4259.968249 samples/sec	accuracy=0.628205	lr=0.357598
Epoch[29] Batch [399]	Speed: 4091.940746 samples/sec	accuracy=0.628319	lr=0.357464
Epoch[29] Batch [449]	Speed: 4280.034256 samples/sec	accuracy=0.628337	lr=0.357329
Epoch[29] Batch [499]	Speed: 4119.911277 samples/sec	accuracy=0.628217	lr=0.357194
Epoch[29] Batch [549]	Speed: 4212.497794 samples/sec	accuracy=0.628236	lr=0.357059
Epoch[29] Batch [599]	Speed: 4250.202043 samples/sec	accuracy=0.628242	lr=0.356924
Epoch[29] Batch [649]	Speed: 4037.289703 samples/sec	accuracy=0.628323	lr=0.356788
Epoch[29] Batch [699]	Speed: 4138.968098 samples/sec	accuracy=0.628506	lr=0.356652
Epoch[29] Batch [749]	Speed: 4284.639344 samples/sec	accuracy=0.628573	lr=0.356517
Epoch[29] Batch [799]	Speed: 4061.838943 samples/sec	accuracy=0.628619	lr=0.356380
Epoch[29] Batch [849]	Speed: 4127.384586 samples/sec	accuracy=0.628715	lr=0.356244
Epoch[29] Batch [899]	Speed: 4059.768045 samples/sec	accuracy=0.628862	lr=0.356108
Epoch[29] Batch [949]	Speed: 4206.791351 samples/sec	accuracy=0.629016	lr=0.355971
Epoch[29] Batch [999]	Speed: 4117.728091 samples/sec	accuracy=0.629045	lr=0.355834
Epoch[29] Batch [1049]	Speed: 4234.936395 samples/sec	accuracy=0.629096	lr=0.355698
Epoch[29] Batch [1099]	Speed: 4244.266477 samples/sec	accuracy=0.629197	lr=0.355560
Epoch[29] Batch [1149]	Speed: 4118.558481 samples/sec	accuracy=0.629284	lr=0.355423
Epoch[29] Batch [1199]	Speed: 4210.820285 samples/sec	accuracy=0.629360	lr=0.355285
Epoch[29] Batch [1249]	Speed: 4115.648412 samples/sec	accuracy=0.629454	lr=0.355148
[Epoch 29] training: accuracy=0.629459
[Epoch 29] speed: 4169 samples/sec	time cost: 317.190155
[Epoch 29] validation: err-top1=0.381104 err-top5=0.152730
Epoch[30] Batch [49]	Speed: 4191.477536 samples/sec	accuracy=0.629254	lr=0.355007
Epoch[30] Batch [99]	Speed: 4201.103223 samples/sec	accuracy=0.628896	lr=0.354869
Epoch[30] Batch [149]	Speed: 4175.605443 samples/sec	accuracy=0.629390	lr=0.354731
Epoch[30] Batch [199]	Speed: 4312.239887 samples/sec	accuracy=0.629366	lr=0.354592
Epoch[30] Batch [249]	Speed: 4152.700992 samples/sec	accuracy=0.629488	lr=0.354454
Epoch[30] Batch [299]	Speed: 4227.399751 samples/sec	accuracy=0.629502	lr=0.354315
Epoch[30] Batch [349]	Speed: 4276.087259 samples/sec	accuracy=0.629394	lr=0.354176
Epoch[30] Batch [399]	Speed: 4166.333742 samples/sec	accuracy=0.629388	lr=0.354037
Epoch[30] Batch [449]	Speed: 4140.410732 samples/sec	accuracy=0.629468	lr=0.353897
Epoch[30] Batch [499]	Speed: 4181.891379 samples/sec	accuracy=0.629544	lr=0.353758
Epoch[30] Batch [549]	Speed: 4164.957724 samples/sec	accuracy=0.629575	lr=0.353618
Epoch[30] Batch [599]	Speed: 4135.791596 samples/sec	accuracy=0.629661	lr=0.353478
Epoch[30] Batch [649]	Speed: 4114.595998 samples/sec	accuracy=0.629703	lr=0.353338
Epoch[30] Batch [699]	Speed: 4089.637788 samples/sec	accuracy=0.629828	lr=0.353198
Epoch[30] Batch [749]	Speed: 4157.699597 samples/sec	accuracy=0.629884	lr=0.353057
Epoch[30] Batch [799]	Speed: 4256.142256 samples/sec	accuracy=0.629983	lr=0.352917
Epoch[30] Batch [849]	Speed: 4201.903125 samples/sec	accuracy=0.629975	lr=0.352776
Epoch[30] Batch [899]	Speed: 4089.494100 samples/sec	accuracy=0.629990	lr=0.352635
Epoch[30] Batch [949]	Speed: 4215.251495 samples/sec	accuracy=0.630091	lr=0.352493
Epoch[30] Batch [999]	Speed: 4313.099393 samples/sec	accuracy=0.630199	lr=0.352352
Epoch[30] Batch [1049]	Speed: 4098.112964 samples/sec	accuracy=0.630295	lr=0.352211
Epoch[30] Batch [1099]	Speed: 4017.667409 samples/sec	accuracy=0.630329	lr=0.352069
Epoch[30] Batch [1149]	Speed: 4324.266189 samples/sec	accuracy=0.630414	lr=0.351927
Epoch[30] Batch [1199]	Speed: 4099.601758 samples/sec	accuracy=0.630508	lr=0.351785
Epoch[30] Batch [1249]	Speed: 4159.325937 samples/sec	accuracy=0.630560	lr=0.351642
[Epoch 30] training: accuracy=0.630559
[Epoch 30] speed: 4173 samples/sec	time cost: 317.729846
[Epoch 30] validation: err-top1=0.378109 err-top5=0.146624
Epoch[31] Batch [49]	Speed: 3916.391504 samples/sec	accuracy=0.631133	lr=0.351497
Epoch[31] Batch [99]	Speed: 4287.287050 samples/sec	accuracy=0.631034	lr=0.351354
Epoch[31] Batch [149]	Speed: 4186.847241 samples/sec	accuracy=0.630612	lr=0.351212
Epoch[31] Batch [199]	Speed: 4224.113287 samples/sec	accuracy=0.630065	lr=0.351069
Epoch[31] Batch [249]	Speed: 4300.355297 samples/sec	accuracy=0.630026	lr=0.350925
Epoch[31] Batch [299]	Speed: 4182.910471 samples/sec	accuracy=0.630014	lr=0.350782
Epoch[31] Batch [349]	Speed: 4256.330203 samples/sec	accuracy=0.630102	lr=0.350639
Epoch[31] Batch [399]	Speed: 4331.818684 samples/sec	accuracy=0.629812	lr=0.350495
Epoch[31] Batch [449]	Speed: 4293.281754 samples/sec	accuracy=0.629820	lr=0.350351
Epoch[31] Batch [499]	Speed: 4226.958576 samples/sec	accuracy=0.629843	lr=0.350207
Epoch[31] Batch [549]	Speed: 4115.336480 samples/sec	accuracy=0.630021	lr=0.350062
Epoch[31] Batch [599]	Speed: 4256.292326 samples/sec	accuracy=0.630262	lr=0.349918
Epoch[31] Batch [649]	Speed: 4130.826827 samples/sec	accuracy=0.630366	lr=0.349773
Epoch[31] Batch [699]	Speed: 4171.060035 samples/sec	accuracy=0.630595	lr=0.349629
Epoch[31] Batch [749]	Speed: 4204.766477 samples/sec	accuracy=0.630759	lr=0.349484
Epoch[31] Batch [799]	Speed: 4191.973034 samples/sec	accuracy=0.630834	lr=0.349338
Epoch[31] Batch [849]	Speed: 4272.365287 samples/sec	accuracy=0.630845	lr=0.349193
Epoch[31] Batch [899]	Speed: 4098.076912 samples/sec	accuracy=0.630918	lr=0.349048
Epoch[31] Batch [949]	Speed: 4249.652066 samples/sec	accuracy=0.631031	lr=0.348902
Epoch[31] Batch [999]	Speed: 4239.995667 samples/sec	accuracy=0.631102	lr=0.348756
Epoch[31] Batch [1049]	Speed: 4263.631650 samples/sec	accuracy=0.631142	lr=0.348610
Epoch[31] Batch [1099]	Speed: 4215.465968 samples/sec	accuracy=0.631231	lr=0.348464
Epoch[31] Batch [1149]	Speed: 4124.100928 samples/sec	accuracy=0.631365	lr=0.348317
Epoch[31] Batch [1199]	Speed: 4111.384521 samples/sec	accuracy=0.631440	lr=0.348171
Epoch[31] Batch [1249]	Speed: 4299.144173 samples/sec	accuracy=0.631534	lr=0.348024
[Epoch 31] training: accuracy=0.631537
[Epoch 31] speed: 4200 samples/sec	time cost: 315.484501
[Epoch 31] validation: err-top1=0.380939 err-top5=0.151885
Epoch[32] Batch [49]	Speed: 4143.988373 samples/sec	accuracy=0.631345	lr=0.347874
Epoch[32] Batch [99]	Speed: 4299.919427 samples/sec	accuracy=0.631422	lr=0.347727
Epoch[32] Batch [149]	Speed: 4213.980242 samples/sec	accuracy=0.630885	lr=0.347580
Epoch[32] Batch [199]	Speed: 4147.956981 samples/sec	accuracy=0.630350	lr=0.347432
Epoch[32] Batch [249]	Speed: 4243.792254 samples/sec	accuracy=0.630450	lr=0.347285
Epoch[32] Batch [299]	Speed: 4209.801744 samples/sec	accuracy=0.630787	lr=0.347137
Epoch[32] Batch [349]	Speed: 4193.905907 samples/sec	accuracy=0.631004	lr=0.346989
Epoch[32] Batch [399]	Speed: 4249.128043 samples/sec	accuracy=0.631221	lr=0.346841
Epoch[32] Batch [449]	Speed: 4332.292946 samples/sec	accuracy=0.631361	lr=0.346692
Epoch[32] Batch [499]	Speed: 4233.348788 samples/sec	accuracy=0.631519	lr=0.346544
Epoch[32] Batch [549]	Speed: 4175.283301 samples/sec	accuracy=0.631557	lr=0.346395
Epoch[32] Batch [599]	Speed: 4236.725706 samples/sec	accuracy=0.631753	lr=0.346246
Epoch[32] Batch [649]	Speed: 4219.265985 samples/sec	accuracy=0.631771	lr=0.346097
Epoch[32] Batch [699]	Speed: 4300.909432 samples/sec	accuracy=0.631806	lr=0.345948
Epoch[32] Batch [749]	Speed: 4142.233935 samples/sec	accuracy=0.631848	lr=0.345799
Epoch[32] Batch [799]	Speed: 4134.359581 samples/sec	accuracy=0.632021	lr=0.345649
Epoch[32] Batch [849]	Speed: 4146.710295 samples/sec	accuracy=0.632043	lr=0.345499
Epoch[32] Batch [899]	Speed: 4238.437308 samples/sec	accuracy=0.632147	lr=0.345349
Epoch[32] Batch [949]	Speed: 4269.360271 samples/sec	accuracy=0.632184	lr=0.345199
Epoch[32] Batch [999]	Speed: 4113.363294 samples/sec	accuracy=0.632165	lr=0.345049
Epoch[32] Batch [1049]	Speed: 4261.885179 samples/sec	accuracy=0.632271	lr=0.344899
Epoch[32] Batch [1099]	Speed: 4257.819946 samples/sec	accuracy=0.632321	lr=0.344748
Epoch[32] Batch [1149]	Speed: 4241.525349 samples/sec	accuracy=0.632337	lr=0.344597
Epoch[32] Batch [1199]	Speed: 4216.200491 samples/sec	accuracy=0.632486	lr=0.344446
Epoch[32] Batch [1249]	Speed: 4317.801645 samples/sec	accuracy=0.632584	lr=0.344295
[Epoch 32] training: accuracy=0.632589
[Epoch 32] speed: 4216 samples/sec	time cost: 314.662595
[Epoch 32] validation: err-top1=0.377073 err-top5=0.147501
Epoch[33] Batch [49]	Speed: 4053.119901 samples/sec	accuracy=0.633602	lr=0.344141
Epoch[33] Batch [99]	Speed: 4142.819835 samples/sec	accuracy=0.632718	lr=0.343989
Epoch[33] Batch [149]	Speed: 4219.985579 samples/sec	accuracy=0.632233	lr=0.343838
Epoch[33] Batch [199]	Speed: 4202.540158 samples/sec	accuracy=0.632032	lr=0.343686
Epoch[33] Batch [249]	Speed: 4045.306512 samples/sec	accuracy=0.632203	lr=0.343534
Epoch[33] Batch [299]	Speed: 4140.946210 samples/sec	accuracy=0.632151	lr=0.343382
Epoch[33] Batch [349]	Speed: 4205.150908 samples/sec	accuracy=0.632179	lr=0.343229
Epoch[33] Batch [399]	Speed: 4215.769678 samples/sec	accuracy=0.632181	lr=0.343077
Epoch[33] Batch [449]	Speed: 4125.352757 samples/sec	accuracy=0.632107	lr=0.342924
Epoch[33] Batch [499]	Speed: 4261.122986 samples/sec	accuracy=0.632106	lr=0.342771
Epoch[33] Batch [549]	Speed: 4128.797165 samples/sec	accuracy=0.632066	lr=0.342618
Epoch[33] Batch [599]	Speed: 4284.827509 samples/sec	accuracy=0.632234	lr=0.342465
Epoch[33] Batch [649]	Speed: 4279.897179 samples/sec	accuracy=0.632339	lr=0.342312
Epoch[33] Batch [699]	Speed: 4139.498894 samples/sec	accuracy=0.632423	lr=0.342158
Epoch[33] Batch [749]	Speed: 4132.610899 samples/sec	accuracy=0.632582	lr=0.342005
Epoch[33] Batch [799]	Speed: 4253.821864 samples/sec	accuracy=0.632705	lr=0.341851
Epoch[33] Batch [849]	Speed: 4184.041250 samples/sec	accuracy=0.632729	lr=0.341697
Epoch[33] Batch [899]	Speed: 4224.688338 samples/sec	accuracy=0.632861	lr=0.341543
Epoch[33] Batch [949]	Speed: 4331.551843 samples/sec	accuracy=0.633034	lr=0.341388
Epoch[33] Batch [999]	Speed: 4372.202160 samples/sec	accuracy=0.633098	lr=0.341234
Epoch[33] Batch [1049]	Speed: 4173.712192 samples/sec	accuracy=0.633194	lr=0.341079
Epoch[33] Batch [1099]	Speed: 4135.590170 samples/sec	accuracy=0.633298	lr=0.340924
Epoch[33] Batch [1149]	Speed: 4186.544583 samples/sec	accuracy=0.633405	lr=0.340769
Epoch[33] Batch [1199]	Speed: 4174.932314 samples/sec	accuracy=0.633503	lr=0.340614
Epoch[33] Batch [1249]	Speed: 4362.061649 samples/sec	accuracy=0.633594	lr=0.340458
[Epoch 33] training: accuracy=0.633597
[Epoch 33] speed: 4193 samples/sec	time cost: 316.034183
[Epoch 33] validation: err-top1=0.388034 err-top5=0.156250
Epoch[34] Batch [49]	Speed: 4089.910006 samples/sec	accuracy=0.633372	lr=0.340300
Epoch[34] Batch [99]	Speed: 4221.179226 samples/sec	accuracy=0.633093	lr=0.340144
Epoch[34] Batch [149]	Speed: 4156.095118 samples/sec	accuracy=0.632731	lr=0.339988
Epoch[34] Batch [199]	Speed: 4205.071529 samples/sec	accuracy=0.632653	lr=0.339832
Epoch[34] Batch [249]	Speed: 4096.202901 samples/sec	accuracy=0.632741	lr=0.339676
Epoch[34] Batch [299]	Speed: 4244.972053 samples/sec	accuracy=0.632673	lr=0.339520
Epoch[34] Batch [349]	Speed: 4175.161374 samples/sec	accuracy=0.632907	lr=0.339363
Epoch[34] Batch [399]	Speed: 4226.525894 samples/sec	accuracy=0.632938	lr=0.339206
Epoch[34] Batch [449]	Speed: 4257.014226 samples/sec	accuracy=0.632915	lr=0.339049
Epoch[34] Batch [499]	Speed: 4095.769622 samples/sec	accuracy=0.633014	lr=0.338892
Epoch[34] Batch [549]	Speed: 4053.234192 samples/sec	accuracy=0.632974	lr=0.338735
Epoch[34] Batch [599]	Speed: 4174.923386 samples/sec	accuracy=0.633037	lr=0.338578
Epoch[34] Batch [649]	Speed: 4163.276854 samples/sec	accuracy=0.633106	lr=0.338420
Epoch[34] Batch [699]	Speed: 4156.747703 samples/sec	accuracy=0.633101	lr=0.338263
Epoch[34] Batch [749]	Speed: 4251.687246 samples/sec	accuracy=0.633223	lr=0.338105
Epoch[34] Batch [799]	Speed: 4266.315743 samples/sec	accuracy=0.633326	lr=0.337947
Epoch[34] Batch [849]	Speed: 4193.320617 samples/sec	accuracy=0.633373	lr=0.337788
Epoch[34] Batch [899]	Speed: 4218.633320 samples/sec	accuracy=0.633586	lr=0.337630
Epoch[34] Batch [949]	Speed: 4294.373899 samples/sec	accuracy=0.633698	lr=0.337472
Epoch[34] Batch [999]	Speed: 4091.774052 samples/sec	accuracy=0.633853	lr=0.337313
Epoch[34] Batch [1049]	Speed: 4227.965293 samples/sec	accuracy=0.634001	lr=0.337154
Epoch[34] Batch [1099]	Speed: 4188.156571 samples/sec	accuracy=0.634090	lr=0.336995
Epoch[34] Batch [1149]	Speed: 4234.481789 samples/sec	accuracy=0.634212	lr=0.336836
Epoch[34] Batch [1199]	Speed: 4265.988691 samples/sec	accuracy=0.634333	lr=0.336677
Epoch[34] Batch [1249]	Speed: 4206.013146 samples/sec	accuracy=0.634389	lr=0.336517
[Epoch 34] training: accuracy=0.634389
[Epoch 34] speed: 4185 samples/sec	time cost: 316.002159
[Epoch 34] validation: err-top1=0.377136 err-top5=0.148641
Epoch[35] Batch [49]	Speed: 4112.878249 samples/sec	accuracy=0.635738	lr=0.336354
Epoch[35] Batch [99]	Speed: 4141.063032 samples/sec	accuracy=0.634958	lr=0.336194
Epoch[35] Batch [149]	Speed: 4217.809722 samples/sec	accuracy=0.634076	lr=0.336034
Epoch[35] Batch [199]	Speed: 4354.636882 samples/sec	accuracy=0.633858	lr=0.335874
Epoch[35] Batch [249]	Speed: 4216.974935 samples/sec	accuracy=0.633950	lr=0.335714
Epoch[35] Batch [299]	Speed: 4211.310704 samples/sec	accuracy=0.634047	lr=0.335553
Epoch[35] Batch [349]	Speed: 4235.500112 samples/sec	accuracy=0.634111	lr=0.335393
Epoch[35] Batch [399]	Speed: 4277.564453 samples/sec	accuracy=0.634095	lr=0.335232
Epoch[35] Batch [449]	Speed: 4226.124073 samples/sec	accuracy=0.633971	lr=0.335071
Epoch[35] Batch [499]	Speed: 4147.522137 samples/sec	accuracy=0.634013	lr=0.334910
Epoch[35] Batch [549]	Speed: 4283.328123 samples/sec	accuracy=0.634144	lr=0.334749
Epoch[35] Batch [599]	Speed: 4026.931060 samples/sec	accuracy=0.634199	lr=0.334587
Epoch[35] Batch [649]	Speed: 4222.305057 samples/sec	accuracy=0.634311	lr=0.334425
Epoch[35] Batch [699]	Speed: 4178.403136 samples/sec	accuracy=0.634406	lr=0.334264
Epoch[35] Batch [749]	Speed: 4137.544566 samples/sec	accuracy=0.634532	lr=0.334102
Epoch[35] Batch [799]	Speed: 4207.200384 samples/sec	accuracy=0.634610	lr=0.333940
Epoch[35] Batch [849]	Speed: 4295.535166 samples/sec	accuracy=0.634718	lr=0.333777
Epoch[35] Batch [899]	Speed: 4155.931200 samples/sec	accuracy=0.634922	lr=0.333615
Epoch[35] Batch [949]	Speed: 4049.697985 samples/sec	accuracy=0.635073	lr=0.333452
Epoch[35] Batch [999]	Speed: 4171.640909 samples/sec	accuracy=0.635188	lr=0.333290
Epoch[35] Batch [1049]	Speed: 4107.660444 samples/sec	accuracy=0.635242	lr=0.333127
Epoch[35] Batch [1099]	Speed: 4281.126158 samples/sec	accuracy=0.635344	lr=0.332964
Epoch[35] Batch [1149]	Speed: 4055.041825 samples/sec	accuracy=0.635368	lr=0.332801
Epoch[35] Batch [1199]	Speed: 4136.122410 samples/sec	accuracy=0.635428	lr=0.332637
Epoch[35] Batch [1249]	Speed: 4189.892174 samples/sec	accuracy=0.635551	lr=0.332474
[Epoch 35] training: accuracy=0.635556
[Epoch 35] speed: 4180 samples/sec	time cost: 317.921051
[Epoch 35] validation: err-top1=0.382593 err-top5=0.153799
Epoch[36] Batch [49]	Speed: 4080.963450 samples/sec	accuracy=0.636634	lr=0.332307
Epoch[36] Batch [99]	Speed: 4172.700252 samples/sec	accuracy=0.636007	lr=0.332143
Epoch[36] Batch [149]	Speed: 4239.485907 samples/sec	accuracy=0.635705	lr=0.331979
Epoch[36] Batch [199]	Speed: 4096.725910 samples/sec	accuracy=0.635473	lr=0.331815
Epoch[36] Batch [249]	Speed: 4131.435494 samples/sec	accuracy=0.635580	lr=0.331650
Epoch[36] Batch [299]	Speed: 4196.706122 samples/sec	accuracy=0.635418	lr=0.331486
Epoch[36] Batch [349]	Speed: 4328.401601 samples/sec	accuracy=0.635461	lr=0.331321
Epoch[36] Batch [399]	Speed: 4252.563371 samples/sec	accuracy=0.635356	lr=0.331157
Epoch[36] Batch [449]	Speed: 4248.680219 samples/sec	accuracy=0.635454	lr=0.330992
Epoch[36] Batch [499]	Speed: 4097.909953 samples/sec	accuracy=0.635473	lr=0.330827
Epoch[36] Batch [549]	Speed: 4228.377621 samples/sec	accuracy=0.635483	lr=0.330661
Epoch[36] Batch [599]	Speed: 4181.390690 samples/sec	accuracy=0.635602	lr=0.330496
Epoch[36] Batch [649]	Speed: 4287.448569 samples/sec	accuracy=0.635671	lr=0.330330
Epoch[36] Batch [699]	Speed: 4204.188935 samples/sec	accuracy=0.635667	lr=0.330165
Epoch[36] Batch [749]	Speed: 4184.107608 samples/sec	accuracy=0.635617	lr=0.329999
Epoch[36] Batch [799]	Speed: 4227.114916 samples/sec	accuracy=0.635711	lr=0.329833
Epoch[36] Batch [849]	Speed: 4148.544421 samples/sec	accuracy=0.635803	lr=0.329667
Epoch[36] Batch [899]	Speed: 4197.998649 samples/sec	accuracy=0.635932	lr=0.329500
Epoch[36] Batch [949]	Speed: 4298.685401 samples/sec	accuracy=0.636039	lr=0.329334
Epoch[36] Batch [999]	Speed: 4143.354177 samples/sec	accuracy=0.636082	lr=0.329167
Epoch[36] Batch [1049]	Speed: 4236.687507 samples/sec	accuracy=0.636128	lr=0.329000
Epoch[36] Batch [1099]	Speed: 4192.579310 samples/sec	accuracy=0.636154	lr=0.328833
Epoch[36] Batch [1149]	Speed: 4325.248186 samples/sec	accuracy=0.636230	lr=0.328666
Epoch[36] Batch [1199]	Speed: 4195.334069 samples/sec	accuracy=0.636315	lr=0.328499
Epoch[36] Batch [1249]	Speed: 4279.440800 samples/sec	accuracy=0.636355	lr=0.328332
[Epoch 36] training: accuracy=0.636357
[Epoch 36] speed: 4202 samples/sec	time cost: 315.452528
[Epoch 36] validation: err-top1=0.379863 err-top5=0.154137
Epoch[37] Batch [49]	Speed: 4223.872095 samples/sec	accuracy=0.637734	lr=0.328161
Epoch[37] Batch [99]	Speed: 4102.403354 samples/sec	accuracy=0.636301	lr=0.327993
Epoch[37] Batch [149]	Speed: 4271.630609 samples/sec	accuracy=0.636249	lr=0.327825
Epoch[37] Batch [199]	Speed: 4103.291786 samples/sec	accuracy=0.636603	lr=0.327657
Epoch[37] Batch [249]	Speed: 4240.562072 samples/sec	accuracy=0.636462	lr=0.327489
Epoch[37] Batch [299]	Speed: 4176.781261 samples/sec	accuracy=0.636321	lr=0.327321
Epoch[37] Batch [349]	Speed: 4108.333510 samples/sec	accuracy=0.636336	lr=0.327152
Epoch[37] Batch [399]	Speed: 4121.171794 samples/sec	accuracy=0.636322	lr=0.326983
Epoch[37] Batch [449]	Speed: 4148.802254 samples/sec	accuracy=0.636467	lr=0.326815
Epoch[37] Batch [499]	Speed: 4053.418341 samples/sec	accuracy=0.636556	lr=0.326646
Epoch[37] Batch [549]	Speed: 4227.796988 samples/sec	accuracy=0.636489	lr=0.326477
Epoch[37] Batch [599]	Speed: 4212.535558 samples/sec	accuracy=0.636585	lr=0.326307
Epoch[37] Batch [649]	Speed: 4222.005634 samples/sec	accuracy=0.636675	lr=0.326138
Epoch[37] Batch [699]	Speed: 4230.248138 samples/sec	accuracy=0.636783	lr=0.325968
Epoch[37] Batch [749]	Speed: 4297.364452 samples/sec	accuracy=0.636827	lr=0.325799
Epoch[37] Batch [799]	Speed: 4204.949174 samples/sec	accuracy=0.636826	lr=0.325629
Epoch[37] Batch [849]	Speed: 4200.619287 samples/sec	accuracy=0.636878	lr=0.325459
Epoch[37] Batch [899]	Speed: 4275.506303 samples/sec	accuracy=0.636969	lr=0.325289
Epoch[37] Batch [949]	Speed: 4297.136576 samples/sec	accuracy=0.637064	lr=0.325118
Epoch[37] Batch [999]	Speed: 4193.268787 samples/sec	accuracy=0.637181	lr=0.324948
Epoch[37] Batch [1049]	Speed: 4276.886331 samples/sec	accuracy=0.637263	lr=0.324777
Epoch[37] Batch [1099]	Speed: 4145.307598 samples/sec	accuracy=0.637369	lr=0.324607
Epoch[37] Batch [1149]	Speed: 4309.953349 samples/sec	accuracy=0.637412	lr=0.324436
Epoch[37] Batch [1199]	Speed: 4296.000484 samples/sec	accuracy=0.637486	lr=0.324265
Epoch[37] Batch [1249]	Speed: 4344.150376 samples/sec	accuracy=0.637587	lr=0.324094
[Epoch 37] training: accuracy=0.637591
[Epoch 37] speed: 4206 samples/sec	time cost: 316.105566
[Epoch 37] validation: err-top1=0.370436 err-top5=0.144471
Epoch[38] Batch [49]	Speed: 4115.510147 samples/sec	accuracy=0.637440	lr=0.323919
Epoch[38] Batch [99]	Speed: 4159.332946 samples/sec	accuracy=0.636812	lr=0.323747
Epoch[38] Batch [149]	Speed: 4175.824021 samples/sec	accuracy=0.636389	lr=0.323576
Epoch[38] Batch [199]	Speed: 4314.522264 samples/sec	accuracy=0.636383	lr=0.323404
Epoch[38] Batch [249]	Speed: 4163.942839 samples/sec	accuracy=0.636628	lr=0.323232
Epoch[38] Batch [299]	Speed: 4320.373599 samples/sec	accuracy=0.636737	lr=0.323060
Epoch[38] Batch [349]	Speed: 4141.192638 samples/sec	accuracy=0.636818	lr=0.322888
Epoch[38] Batch [399]	Speed: 4174.638112 samples/sec	accuracy=0.636901	lr=0.322715
Epoch[38] Batch [449]	Speed: 4108.107402 samples/sec	accuracy=0.637056	lr=0.322543
Epoch[38] Batch [499]	Speed: 4335.661905 samples/sec	accuracy=0.637269	lr=0.322370
Epoch[38] Batch [549]	Speed: 4117.753199 samples/sec	accuracy=0.637416	lr=0.322197
Epoch[38] Batch [599]	Speed: 4217.549038 samples/sec	accuracy=0.637520	lr=0.322025
Epoch[38] Batch [649]	Speed: 4208.944881 samples/sec	accuracy=0.637554	lr=0.321851
Epoch[38] Batch [699]	Speed: 4278.757736 samples/sec	accuracy=0.637562	lr=0.321678
Epoch[38] Batch [749]	Speed: 4261.405996 samples/sec	accuracy=0.637638	lr=0.321505
Epoch[38] Batch [799]	Speed: 4199.062144 samples/sec	accuracy=0.637727	lr=0.321331
Epoch[38] Batch [849]	Speed: 4233.800062 samples/sec	accuracy=0.637843	lr=0.321158
Epoch[38] Batch [899]	Speed: 4141.191041 samples/sec	accuracy=0.637962	lr=0.320984
Epoch[38] Batch [949]	Speed: 4209.341791 samples/sec	accuracy=0.638030	lr=0.320810
Epoch[38] Batch [999]	Speed: 4208.765301 samples/sec	accuracy=0.638067	lr=0.320636
Epoch[38] Batch [1049]	Speed: 4234.024776 samples/sec	accuracy=0.638208	lr=0.320462
Epoch[38] Batch [1099]	Speed: 4243.180383 samples/sec	accuracy=0.638306	lr=0.320287
Epoch[38] Batch [1149]	Speed: 4268.813895 samples/sec	accuracy=0.638440	lr=0.320113
Epoch[38] Batch [1199]	Speed: 4115.810194 samples/sec	accuracy=0.638520	lr=0.319938
Epoch[38] Batch [1249]	Speed: 4301.113932 samples/sec	accuracy=0.638624	lr=0.319763
[Epoch 38] training: accuracy=0.638623
[Epoch 38] speed: 4205 samples/sec	time cost: 315.471785
[Epoch 38] validation: err-top1=0.377970 err-top5=0.150450
Epoch[39] Batch [49]	Speed: 4070.237979 samples/sec	accuracy=0.639181	lr=0.319585
Epoch[39] Batch [99]	Speed: 4248.056266 samples/sec	accuracy=0.637908	lr=0.319410
Epoch[39] Batch [149]	Speed: 4229.717225 samples/sec	accuracy=0.637605	lr=0.319234
Epoch[39] Batch [199]	Speed: 4147.220092 samples/sec	accuracy=0.637746	lr=0.319059
Epoch[39] Batch [249]	Speed: 4245.735950 samples/sec	accuracy=0.637901	lr=0.318883
Epoch[39] Batch [299]	Speed: 4247.488613 samples/sec	accuracy=0.637691	lr=0.318708
Epoch[39] Batch [349]	Speed: 4198.901962 samples/sec	accuracy=0.637911	lr=0.318532
Epoch[39] Batch [399]	Speed: 4101.143169 samples/sec	accuracy=0.638122	lr=0.318356
Epoch[39] Batch [449]	Speed: 4288.721802 samples/sec	accuracy=0.638324	lr=0.318180
Epoch[39] Batch [499]	Speed: 4263.133286 samples/sec	accuracy=0.638304	lr=0.318004
Epoch[39] Batch [549]	Speed: 4289.751987 samples/sec	accuracy=0.638377	lr=0.317827
Epoch[39] Batch [599]	Speed: 4219.937566 samples/sec	accuracy=0.638488	lr=0.317651
Epoch[39] Batch [649]	Speed: 4131.866573 samples/sec	accuracy=0.638601	lr=0.317474
Epoch[39] Batch [699]	Speed: 4239.650624 samples/sec	accuracy=0.638683	lr=0.317297
Epoch[39] Batch [749]	Speed: 4239.367650 samples/sec	accuracy=0.638749	lr=0.317120
Epoch[39] Batch [799]	Speed: 4198.520807 samples/sec	accuracy=0.638861	lr=0.316943
Epoch[39] Batch [849]	Speed: 4221.175077 samples/sec	accuracy=0.638878	lr=0.316766
Epoch[39] Batch [899]	Speed: 4217.995210 samples/sec	accuracy=0.638968	lr=0.316589
Epoch[39] Batch [949]	Speed: 4177.135971 samples/sec	accuracy=0.639127	lr=0.316411
Epoch[39] Batch [999]	Speed: 4171.903972 samples/sec	accuracy=0.639213	lr=0.316233
Epoch[39] Batch [1049]	Speed: 4368.066540 samples/sec	accuracy=0.639347	lr=0.316056
Epoch[39] Batch [1099]	Speed: 4118.956065 samples/sec	accuracy=0.639466	lr=0.315878
Epoch[39] Batch [1149]	Speed: 4176.327195 samples/sec	accuracy=0.639525	lr=0.315700
Epoch[39] Batch [1199]	Speed: 4178.423542 samples/sec	accuracy=0.639611	lr=0.315522
Epoch[39] Batch [1249]	Speed: 4326.675423 samples/sec	accuracy=0.639687	lr=0.315343
[Epoch 39] training: accuracy=0.639690
[Epoch 39] speed: 4207 samples/sec	time cost: 314.576524
[Epoch 39] validation: err-top1=0.365852 err-top5=0.140226
Epoch[40] Batch [49]	Speed: 4023.533302 samples/sec	accuracy=0.639084	lr=0.315161
Epoch[40] Batch [99]	Speed: 4160.352678 samples/sec	accuracy=0.639322	lr=0.314983
Epoch[40] Batch [149]	Speed: 4086.481468 samples/sec	accuracy=0.639485	lr=0.314804
Epoch[40] Batch [199]	Speed: 4268.138800 samples/sec	accuracy=0.639671	lr=0.314625
Epoch[40] Batch [249]	Speed: 4221.850337 samples/sec	accuracy=0.639683	lr=0.314446
Epoch[40] Batch [299]	Speed: 4117.025817 samples/sec	accuracy=0.639538	lr=0.314267
Epoch[40] Batch [349]	Speed: 4229.459399 samples/sec	accuracy=0.639241	lr=0.314087
Epoch[40] Batch [399]	Speed: 4250.763858 samples/sec	accuracy=0.639218	lr=0.313908
Epoch[40] Batch [449]	Speed: 4316.313972 samples/sec	accuracy=0.639190	lr=0.313728
Epoch[40] Batch [499]	Speed: 4293.248365 samples/sec	accuracy=0.639178	lr=0.313549
Epoch[40] Batch [549]	Speed: 4258.393994 samples/sec	accuracy=0.639264	lr=0.313369
Epoch[40] Batch [599]	Speed: 4181.505083 samples/sec	accuracy=0.639316	lr=0.313189
Epoch[40] Batch [649]	Speed: 4201.274505 samples/sec	accuracy=0.639382	lr=0.313009
Epoch[40] Batch [699]	Speed: 4266.027843 samples/sec	accuracy=0.639407	lr=0.312829
Epoch[40] Batch [749]	Speed: 4135.653248 samples/sec	accuracy=0.639463	lr=0.312648
Epoch[40] Batch [799]	Speed: 4211.534523 samples/sec	accuracy=0.639591	lr=0.312468
Epoch[40] Batch [849]	Speed: 4124.628551 samples/sec	accuracy=0.639700	lr=0.312287
Epoch[40] Batch [899]	Speed: 4247.097496 samples/sec	accuracy=0.639816	lr=0.312106
Epoch[40] Batch [949]	Speed: 4240.265999 samples/sec	accuracy=0.639951	lr=0.311925
Epoch[40] Batch [999]	Speed: 4103.965617 samples/sec	accuracy=0.640019	lr=0.311744
Epoch[40] Batch [1049]	Speed: 4233.000904 samples/sec	accuracy=0.640191	lr=0.311563
Epoch[40] Batch [1099]	Speed: 4326.965465 samples/sec	accuracy=0.640337	lr=0.311382
Epoch[40] Batch [1149]	Speed: 4212.814217 samples/sec	accuracy=0.640456	lr=0.311201
Epoch[40] Batch [1199]	Speed: 4144.996114 samples/sec	accuracy=0.640626	lr=0.311019
Epoch[40] Batch [1249]	Speed: 4261.536818 samples/sec	accuracy=0.640703	lr=0.310837
[Epoch 40] training: accuracy=0.640700
[Epoch 40] speed: 4199 samples/sec	time cost: 315.087571
[Epoch 40] validation: err-top1=0.366699 err-top5=0.138102
Epoch[41] Batch [49]	Speed: 4201.132810 samples/sec	accuracy=0.640535	lr=0.310652
Epoch[41] Batch [99]	Speed: 4124.724252 samples/sec	accuracy=0.640450	lr=0.310470
Epoch[41] Batch [149]	Speed: 4251.846178 samples/sec	accuracy=0.640660	lr=0.310288
Epoch[41] Batch [199]	Speed: 4238.154496 samples/sec	accuracy=0.640544	lr=0.310105
Epoch[41] Batch [249]	Speed: 4226.282763 samples/sec	accuracy=0.640555	lr=0.309923
Epoch[41] Batch [299]	Speed: 4332.033474 samples/sec	accuracy=0.640129	lr=0.309741
Epoch[41] Batch [349]	Speed: 4171.026414 samples/sec	accuracy=0.640137	lr=0.309558
Epoch[41] Batch [399]	Speed: 4388.648798 samples/sec	accuracy=0.639933	lr=0.309375
Epoch[41] Batch [449]	Speed: 4115.358326 samples/sec	accuracy=0.640042	lr=0.309192
Epoch[41] Batch [499]	Speed: 4230.901463 samples/sec	accuracy=0.640195	lr=0.309009
Epoch[41] Batch [549]	Speed: 4233.024350 samples/sec	accuracy=0.640273	lr=0.308826
Epoch[41] Batch [599]	Speed: 4245.279274 samples/sec	accuracy=0.640291	lr=0.308643
Epoch[41] Batch [649]	Speed: 4357.441592 samples/sec	accuracy=0.640468	lr=0.308459
Epoch[41] Batch [699]	Speed: 4370.424511 samples/sec	accuracy=0.640448	lr=0.308276
Epoch[41] Batch [749]	Speed: 4199.760982 samples/sec	accuracy=0.640544	lr=0.308092
Epoch[41] Batch [799]	Speed: 4193.053045 samples/sec	accuracy=0.640692	lr=0.307908
Epoch[41] Batch [849]	Speed: 4105.124892 samples/sec	accuracy=0.640790	lr=0.307725
Epoch[41] Batch [899]	Speed: 4193.143760 samples/sec	accuracy=0.640874	lr=0.307540
Epoch[41] Batch [949]	Speed: 4236.894722 samples/sec	accuracy=0.641022	lr=0.307356
Epoch[41] Batch [999]	Speed: 4206.651755 samples/sec	accuracy=0.641174	lr=0.307172
Epoch[41] Batch [1049]	Speed: 4149.873044 samples/sec	accuracy=0.641299	lr=0.306988
Epoch[41] Batch [1099]	Speed: 4256.489735 samples/sec	accuracy=0.641400	lr=0.306803
Epoch[41] Batch [1149]	Speed: 4196.064380 samples/sec	accuracy=0.641517	lr=0.306618
Epoch[41] Batch [1199]	Speed: 4073.635130 samples/sec	accuracy=0.641656	lr=0.306433
Epoch[41] Batch [1249]	Speed: 4374.702817 samples/sec	accuracy=0.641752	lr=0.306249
[Epoch 41] training: accuracy=0.641753
[Epoch 41] speed: 4222 samples/sec	time cost: 313.366954
[Epoch 41] validation: err-top1=0.368682 err-top5=0.141183
Epoch[42] Batch [49]	Speed: 4154.514784 samples/sec	accuracy=0.643241	lr=0.306060
Epoch[42] Batch [99]	Speed: 4103.219029 samples/sec	accuracy=0.642209	lr=0.305875
Epoch[42] Batch [149]	Speed: 4268.558323 samples/sec	accuracy=0.641263	lr=0.305689
Epoch[42] Batch [199]	Speed: 4236.825593 samples/sec	accuracy=0.641159	lr=0.305504
Epoch[42] Batch [249]	Speed: 4154.520892 samples/sec	accuracy=0.641304	lr=0.305318
Epoch[42] Batch [299]	Speed: 4239.490427 samples/sec	accuracy=0.641441	lr=0.305133
Epoch[42] Batch [349]	Speed: 4194.498241 samples/sec	accuracy=0.641316	lr=0.304947
Epoch[42] Batch [399]	Speed: 4140.261698 samples/sec	accuracy=0.641391	lr=0.304761
Epoch[42] Batch [449]	Speed: 4273.670825 samples/sec	accuracy=0.641445	lr=0.304575
Epoch[42] Batch [499]	Speed: 4144.981073 samples/sec	accuracy=0.641401	lr=0.304388
Epoch[42] Batch [549]	Speed: 4268.288275 samples/sec	accuracy=0.641258	lr=0.304202
Epoch[42] Batch [599]	Speed: 4117.025186 samples/sec	accuracy=0.641369	lr=0.304016
Epoch[42] Batch [649]	Speed: 4308.685720 samples/sec	accuracy=0.641412	lr=0.303829
Epoch[42] Batch [699]	Speed: 4310.845606 samples/sec	accuracy=0.641486	lr=0.303642
Epoch[42] Batch [749]	Speed: 4113.649868 samples/sec	accuracy=0.641572	lr=0.303456
Epoch[42] Batch [799]	Speed: 4277.065638 samples/sec	accuracy=0.641692	lr=0.303269
Epoch[42] Batch [849]	Speed: 4109.983121 samples/sec	accuracy=0.641759	lr=0.303082
Epoch[42] Batch [899]	Speed: 4139.261284 samples/sec	accuracy=0.641895	lr=0.302894
Epoch[42] Batch [949]	Speed: 4182.579055 samples/sec	accuracy=0.641997	lr=0.302707
Epoch[42] Batch [999]	Speed: 4132.404773 samples/sec	accuracy=0.642060	lr=0.302520
Epoch[42] Batch [1049]	Speed: 4329.548700 samples/sec	accuracy=0.642123	lr=0.302332
Epoch[42] Batch [1099]	Speed: 4119.795645 samples/sec	accuracy=0.642157	lr=0.302144
Epoch[42] Batch [1149]	Speed: 4111.916060 samples/sec	accuracy=0.642259	lr=0.301957
Epoch[42] Batch [1199]	Speed: 4216.167132 samples/sec	accuracy=0.642415	lr=0.301769
Epoch[42] Batch [1249]	Speed: 4379.585888 samples/sec	accuracy=0.642498	lr=0.301581
[Epoch 42] training: accuracy=0.642502
[Epoch 42] speed: 4195 samples/sec	time cost: 316.841189
[Epoch 42] validation: err-top1=0.364078 err-top5=0.138971
Epoch[43] Batch [49]	Speed: 4141.835199 samples/sec	accuracy=0.642845	lr=0.301389
Epoch[43] Batch [99]	Speed: 4162.377266 samples/sec	accuracy=0.641634	lr=0.301200
Epoch[43] Batch [149]	Speed: 4108.042804 samples/sec	accuracy=0.641943	lr=0.301012
Epoch[43] Batch [199]	Speed: 4269.948812 samples/sec	accuracy=0.641974	lr=0.300823
Epoch[43] Batch [249]	Speed: 4189.246875 samples/sec	accuracy=0.641873	lr=0.300635
Epoch[43] Batch [299]	Speed: 4245.214906 samples/sec	accuracy=0.641845	lr=0.300446
Epoch[43] Batch [349]	Speed: 4235.598187 samples/sec	accuracy=0.641876	lr=0.300257
Epoch[43] Batch [399]	Speed: 4330.164955 samples/sec	accuracy=0.641973	lr=0.300068
Epoch[43] Batch [449]	Speed: 4346.890732 samples/sec	accuracy=0.642136	lr=0.299879
Epoch[43] Batch [499]	Speed: 4297.188513 samples/sec	accuracy=0.642165	lr=0.299690
Epoch[43] Batch [549]	Speed: 4112.328900 samples/sec	accuracy=0.642328	lr=0.299500
Epoch[43] Batch [599]	Speed: 4196.960053 samples/sec	accuracy=0.642469	lr=0.299311
Epoch[43] Batch [649]	Speed: 4293.500379 samples/sec	accuracy=0.642595	lr=0.299121
Epoch[43] Batch [699]	Speed: 4340.288414 samples/sec	accuracy=0.642687	lr=0.298932
Epoch[43] Batch [749]	Speed: 4256.602284 samples/sec	accuracy=0.642799	lr=0.298742
Epoch[43] Batch [799]	Speed: 4252.094027 samples/sec	accuracy=0.642830	lr=0.298552
Epoch[43] Batch [849]	Speed: 4191.734843 samples/sec	accuracy=0.642861	lr=0.298362
Epoch[43] Batch [899]	Speed: 4301.309405 samples/sec	accuracy=0.642941	lr=0.298171
Epoch[43] Batch [949]	Speed: 4182.202894 samples/sec	accuracy=0.643091	lr=0.297981
Epoch[43] Batch [999]	Speed: 4272.363162 samples/sec	accuracy=0.643161	lr=0.297791
Epoch[43] Batch [1049]	Speed: 4262.712967 samples/sec	accuracy=0.643284	lr=0.297600
Epoch[43] Batch [1099]	Speed: 4236.118962 samples/sec	accuracy=0.643361	lr=0.297409
Epoch[43] Batch [1149]	Speed: 4197.843389 samples/sec	accuracy=0.643497	lr=0.297219
Epoch[43] Batch [1199]	Speed: 4225.727816 samples/sec	accuracy=0.643564	lr=0.297028
Epoch[43] Batch [1249]	Speed: 4234.152252 samples/sec	accuracy=0.643680	lr=0.296837
[Epoch 43] training: accuracy=0.643683
[Epoch 43] speed: 4230 samples/sec	time cost: 313.679536
[Epoch 43] validation: err-top1=0.358438 err-top5=0.137376
Epoch[44] Batch [49]	Speed: 4135.295514 samples/sec	accuracy=0.644721	lr=0.296642
Epoch[44] Batch [99]	Speed: 4169.470082 samples/sec	accuracy=0.644036	lr=0.296451
Epoch[44] Batch [149]	Speed: 4177.115171 samples/sec	accuracy=0.644183	lr=0.296259
Epoch[44] Batch [199]	Speed: 4082.463167 samples/sec	accuracy=0.644190	lr=0.296068
Epoch[44] Batch [249]	Speed: 4196.048556 samples/sec	accuracy=0.643823	lr=0.295876
Epoch[44] Batch [299]	Speed: 4200.317016 samples/sec	accuracy=0.643764	lr=0.295684
Epoch[44] Batch [349]	Speed: 4236.127652 samples/sec	accuracy=0.643511	lr=0.295493
Epoch[44] Batch [399]	Speed: 4244.980863 samples/sec	accuracy=0.643555	lr=0.295301
Epoch[44] Batch [449]	Speed: 4288.715635 samples/sec	accuracy=0.643475	lr=0.295109
Epoch[44] Batch [499]	Speed: 4144.651240 samples/sec	accuracy=0.643535	lr=0.294917
Epoch[44] Batch [549]	Speed: 4095.441092 samples/sec	accuracy=0.643697	lr=0.294724
Epoch[44] Batch [599]	Speed: 4104.718126 samples/sec	accuracy=0.643770	lr=0.294532
Epoch[44] Batch [649]	Speed: 4178.938565 samples/sec	accuracy=0.643810	lr=0.294339
Epoch[44] Batch [699]	Speed: 4172.392826 samples/sec	accuracy=0.643861	lr=0.294147
Epoch[44] Batch [749]	Speed: 4129.569450 samples/sec	accuracy=0.643843	lr=0.293954
Epoch[44] Batch [799]	Speed: 4205.804164 samples/sec	accuracy=0.643878	lr=0.293761
Epoch[44] Batch [849]	Speed: 4298.442845 samples/sec	accuracy=0.643928	lr=0.293568
Epoch[44] Batch [899]	Speed: 4212.495398 samples/sec	accuracy=0.644031	lr=0.293375
Epoch[44] Batch [949]	Speed: 4317.547552 samples/sec	accuracy=0.644192	lr=0.293182
Epoch[44] Batch [999]	Speed: 4199.086530 samples/sec	accuracy=0.644333	lr=0.292989
Epoch[44] Batch [1049]	Speed: 4291.394024 samples/sec	accuracy=0.644359	lr=0.292795
Epoch[44] Batch [1099]	Speed: 4075.649195 samples/sec	accuracy=0.644442	lr=0.292602
Epoch[44] Batch [1149]	Speed: 4253.188312 samples/sec	accuracy=0.644563	lr=0.292408
Epoch[44] Batch [1199]	Speed: 4166.325254 samples/sec	accuracy=0.644701	lr=0.292215
Epoch[44] Batch [1249]	Speed: 4343.229787 samples/sec	accuracy=0.644797	lr=0.292021
[Epoch 44] training: accuracy=0.644801
[Epoch 44] speed: 4192 samples/sec	time cost: 316.907386
[Epoch 44] validation: err-top1=0.360332 err-top5=0.138632
Epoch[45] Batch [49]	Speed: 4252.492466 samples/sec	accuracy=0.644643	lr=0.291823
Epoch[45] Batch [99]	Speed: 4287.386511 samples/sec	accuracy=0.644961	lr=0.291629
Epoch[45] Batch [149]	Speed: 4102.459232 samples/sec	accuracy=0.644646	lr=0.291435
Epoch[45] Batch [199]	Speed: 4185.750517 samples/sec	accuracy=0.644381	lr=0.291240
Epoch[45] Batch [249]	Speed: 4333.109143 samples/sec	accuracy=0.644452	lr=0.291046
Epoch[45] Batch [299]	Speed: 4209.866858 samples/sec	accuracy=0.644365	lr=0.290852
Epoch[45] Batch [349]	Speed: 4215.458687 samples/sec	accuracy=0.644546	lr=0.290657
Epoch[45] Batch [399]	Speed: 4262.120413 samples/sec	accuracy=0.644619	lr=0.290462
Epoch[45] Batch [449]	Speed: 4147.355210 samples/sec	accuracy=0.644648	lr=0.290267
Epoch[45] Batch [499]	Speed: 4226.667893 samples/sec	accuracy=0.644850	lr=0.290073
Epoch[45] Batch [549]	Speed: 4185.074766 samples/sec	accuracy=0.644972	lr=0.289878
Epoch[45] Batch [599]	Speed: 4058.008570 samples/sec	accuracy=0.645126	lr=0.289682
Epoch[45] Batch [649]	Speed: 4133.958301 samples/sec	accuracy=0.645303	lr=0.289487
Epoch[45] Batch [699]	Speed: 4241.063380 samples/sec	accuracy=0.645393	lr=0.289292
Epoch[45] Batch [749]	Speed: 4298.020782 samples/sec	accuracy=0.645477	lr=0.289096
Epoch[45] Batch [799]	Speed: 4203.681823 samples/sec	accuracy=0.645488	lr=0.288901
Epoch[45] Batch [849]	Speed: 4205.900786 samples/sec	accuracy=0.645521	lr=0.288705
Epoch[45] Batch [899]	Speed: 4162.206397 samples/sec	accuracy=0.645637	lr=0.288509
Epoch[45] Batch [949]	Speed: 4165.595158 samples/sec	accuracy=0.645719	lr=0.288314
Epoch[45] Batch [999]	Speed: 4161.804775 samples/sec	accuracy=0.645845	lr=0.288118
Epoch[45] Batch [1049]	Speed: 4299.281797 samples/sec	accuracy=0.645857	lr=0.287921
Epoch[45] Batch [1099]	Speed: 4246.144700 samples/sec	accuracy=0.645926	lr=0.287725
Epoch[45] Batch [1149]	Speed: 4263.123638 samples/sec	accuracy=0.645976	lr=0.287529
Epoch[45] Batch [1199]	Speed: 4191.752025 samples/sec	accuracy=0.646059	lr=0.287333
Epoch[45] Batch [1249]	Speed: 4297.995492 samples/sec	accuracy=0.646114	lr=0.287136
[Epoch 45] training: accuracy=0.646113
[Epoch 45] speed: 4208 samples/sec	time cost: 314.933140
[Epoch 45] validation: err-top1=0.365772 err-top5=0.141821
Epoch[46] Batch [49]	Speed: 4046.544125 samples/sec	accuracy=0.644869	lr=0.286936
Epoch[46] Batch [99]	Speed: 4281.272277 samples/sec	accuracy=0.645380	lr=0.286739
Epoch[46] Batch [149]	Speed: 4222.920389 samples/sec	accuracy=0.645115	lr=0.286542
Epoch[46] Batch [199]	Speed: 4177.173184 samples/sec	accuracy=0.645219	lr=0.286345
Epoch[46] Batch [249]	Speed: 4223.906988 samples/sec	accuracy=0.645186	lr=0.286148
Epoch[46] Batch [299]	Speed: 4270.897712 samples/sec	accuracy=0.645229	lr=0.285951
Epoch[46] Batch [349]	Speed: 4167.407050 samples/sec	accuracy=0.645268	lr=0.285754
Epoch[46] Batch [399]	Speed: 4198.459491 samples/sec	accuracy=0.645404	lr=0.285556
Epoch[46] Batch [449]	Speed: 4273.011027 samples/sec	accuracy=0.645440	lr=0.285359
Epoch[46] Batch [499]	Speed: 4309.573561 samples/sec	accuracy=0.645470	lr=0.285161
Epoch[46] Batch [549]	Speed: 4123.706309 samples/sec	accuracy=0.645563	lr=0.284964
Epoch[46] Batch [599]	Speed: 4257.369276 samples/sec	accuracy=0.645739	lr=0.284766
Epoch[46] Batch [649]	Speed: 4077.171946 samples/sec	accuracy=0.645795	lr=0.284568
Epoch[46] Batch [699]	Speed: 4270.532165 samples/sec	accuracy=0.645878	lr=0.284370
Epoch[46] Batch [749]	Speed: 4271.903630 samples/sec	accuracy=0.645938	lr=0.284172
Epoch[46] Batch [799]	Speed: 4096.928882 samples/sec	accuracy=0.645988	lr=0.283974
Epoch[46] Batch [849]	Speed: 4157.815112 samples/sec	accuracy=0.646109	lr=0.283776
Epoch[46] Batch [899]	Speed: 4052.863267 samples/sec	accuracy=0.646164	lr=0.283577
Epoch[46] Batch [949]	Speed: 4046.350307 samples/sec	accuracy=0.646277	lr=0.283379
Epoch[46] Batch [999]	Speed: 4217.514332 samples/sec	accuracy=0.646377	lr=0.283181
Epoch[46] Batch [1049]	Speed: 4221.728082 samples/sec	accuracy=0.646452	lr=0.282982
Epoch[46] Batch [1099]	Speed: 4188.316834 samples/sec	accuracy=0.646514	lr=0.282783
Epoch[46] Batch [1149]	Speed: 4253.198842 samples/sec	accuracy=0.646599	lr=0.282584
Epoch[46] Batch [1199]	Speed: 4230.733341 samples/sec	accuracy=0.646724	lr=0.282385
Epoch[46] Batch [1249]	Speed: 4307.047444 samples/sec	accuracy=0.646753	lr=0.282186
[Epoch 46] training: accuracy=0.646756
[Epoch 46] speed: 4192 samples/sec	time cost: 316.177685
[Epoch 46] validation: err-top1=0.359151 err-top5=0.136841
Epoch[47] Batch [49]	Speed: 4172.224215 samples/sec	accuracy=0.646965	lr=0.281983
Epoch[47] Batch [99]	Speed: 4275.257078 samples/sec	accuracy=0.645995	lr=0.281784
Epoch[47] Batch [149]	Speed: 4241.208787 samples/sec	accuracy=0.646186	lr=0.281585
Epoch[47] Batch [199]	Speed: 4135.512281 samples/sec	accuracy=0.646383	lr=0.281385
Epoch[47] Batch [249]	Speed: 4066.328155 samples/sec	accuracy=0.646334	lr=0.281186
Epoch[47] Batch [299]	Speed: 4191.507642 samples/sec	accuracy=0.646316	lr=0.280986
Epoch[47] Batch [349]	Speed: 4167.220162 samples/sec	accuracy=0.646502	lr=0.280786
Epoch[47] Batch [399]	Speed: 4147.814774 samples/sec	accuracy=0.646515	lr=0.280587
Epoch[47] Batch [449]	Speed: 4228.130447 samples/sec	accuracy=0.646647	lr=0.280387
Epoch[47] Batch [499]	Speed: 4268.440984 samples/sec	accuracy=0.646752	lr=0.280187
Epoch[47] Batch [549]	Speed: 4141.673762 samples/sec	accuracy=0.646853	lr=0.279987
Epoch[47] Batch [599]	Speed: 4225.555116 samples/sec	accuracy=0.646937	lr=0.279786
Epoch[47] Batch [649]	Speed: 4242.469787 samples/sec	accuracy=0.646888	lr=0.279586
Epoch[47] Batch [699]	Speed: 4287.736200 samples/sec	accuracy=0.646983	lr=0.279386
Epoch[47] Batch [749]	Speed: 4252.717399 samples/sec	accuracy=0.647157	lr=0.279185
Epoch[47] Batch [799]	Speed: 4322.350075 samples/sec	accuracy=0.647246	lr=0.278985
Epoch[47] Batch [849]	Speed: 4238.647789 samples/sec	accuracy=0.647277	lr=0.278784
Epoch[47] Batch [899]	Speed: 4105.486294 samples/sec	accuracy=0.647407	lr=0.278583
Epoch[47] Batch [949]	Speed: 4224.364562 samples/sec	accuracy=0.647520	lr=0.278382
Epoch[47] Batch [999]	Speed: 4248.043661 samples/sec	accuracy=0.647534	lr=0.278181
Epoch[47] Batch [1049]	Speed: 4196.990320 samples/sec	accuracy=0.647627	lr=0.277980
Epoch[47] Batch [1099]	Speed: 4333.059569 samples/sec	accuracy=0.647686	lr=0.277779
Epoch[47] Batch [1149]	Speed: 4104.744096 samples/sec	accuracy=0.647763	lr=0.277578
Epoch[47] Batch [1199]	Speed: 4297.571968 samples/sec	accuracy=0.647845	lr=0.277377
Epoch[47] Batch [1249]	Speed: 4343.247179 samples/sec	accuracy=0.647854	lr=0.277175
[Epoch 47] training: accuracy=0.647858
[Epoch 47] speed: 4213 samples/sec	time cost: 315.632043
[Epoch 47] validation: err-top1=0.356565 err-top5=0.134945
Epoch[48] Batch [49]	Speed: 4058.318697 samples/sec	accuracy=0.647700	lr=0.276970
Epoch[48] Batch [99]	Speed: 4084.593183 samples/sec	accuracy=0.647188	lr=0.276768
Epoch[48] Batch [149]	Speed: 4143.369286 samples/sec	accuracy=0.647501	lr=0.276566
Epoch[48] Batch [199]	Speed: 4235.449990 samples/sec	accuracy=0.647550	lr=0.276365
Epoch[48] Batch [249]	Speed: 4083.731156 samples/sec	accuracy=0.647326	lr=0.276163
Epoch[48] Batch [299]	Speed: 4174.022490 samples/sec	accuracy=0.647452	lr=0.275961
Epoch[48] Batch [349]	Speed: 4203.825501 samples/sec	accuracy=0.647288	lr=0.275759
Epoch[48] Batch [399]	Speed: 4167.983672 samples/sec	accuracy=0.647300	lr=0.275557
Epoch[48] Batch [449]	Speed: 4193.102496 samples/sec	accuracy=0.647435	lr=0.275354
Epoch[48] Batch [499]	Speed: 3955.568156 samples/sec	accuracy=0.647421	lr=0.275152
Epoch[48] Batch [549]	Speed: 4233.777609 samples/sec	accuracy=0.647511	lr=0.274950
Epoch[48] Batch [599]	Speed: 4189.757704 samples/sec	accuracy=0.647651	lr=0.274747
Epoch[48] Batch [649]	Speed: 4206.005156 samples/sec	accuracy=0.647672	lr=0.274545
Epoch[48] Batch [699]	Speed: 4085.671423 samples/sec	accuracy=0.647652	lr=0.274342
Epoch[48] Batch [749]	Speed: 4251.191670 samples/sec	accuracy=0.647750	lr=0.274139
Epoch[48] Batch [799]	Speed: 4163.987972 samples/sec	accuracy=0.647760	lr=0.273936
Epoch[48] Batch [849]	Speed: 4288.484052 samples/sec	accuracy=0.647883	lr=0.273733
Epoch[48] Batch [899]	Speed: 4201.626154 samples/sec	accuracy=0.647939	lr=0.273530
Epoch[48] Batch [949]	Speed: 4307.325703 samples/sec	accuracy=0.648067	lr=0.273327
Epoch[48] Batch [999]	Speed: 4222.237814 samples/sec	accuracy=0.648215	lr=0.273124
Epoch[48] Batch [1049]	Speed: 4261.378260 samples/sec	accuracy=0.648313	lr=0.272921
Epoch[48] Batch [1099]	Speed: 4096.891131 samples/sec	accuracy=0.648409	lr=0.272717
Epoch[48] Batch [1149]	Speed: 4231.792474 samples/sec	accuracy=0.648572	lr=0.272514
Epoch[48] Batch [1199]	Speed: 4255.052183 samples/sec	accuracy=0.648648	lr=0.272310
Epoch[48] Batch [1249]	Speed: 4348.145903 samples/sec	accuracy=0.648771	lr=0.272107
[Epoch 48] training: accuracy=0.648774
[Epoch 48] speed: 4180 samples/sec	time cost: 316.270963
[Epoch 48] validation: err-top1=0.355927 err-top5=0.136420
Epoch[49] Batch [49]	Speed: 4158.671173 samples/sec	accuracy=0.651496	lr=0.271899
Epoch[49] Batch [99]	Speed: 4200.277747 samples/sec	accuracy=0.650270	lr=0.271695
Epoch[49] Batch [149]	Speed: 4142.849006 samples/sec	accuracy=0.649748	lr=0.271491
Epoch[49] Batch [199]	Speed: 4152.237937 samples/sec	accuracy=0.649724	lr=0.271287
Epoch[49] Batch [249]	Speed: 4155.564883 samples/sec	accuracy=0.649473	lr=0.271083
Epoch[49] Batch [299]	Speed: 4223.081662 samples/sec	accuracy=0.649368	lr=0.270879
Epoch[49] Batch [349]	Speed: 4243.630150 samples/sec	accuracy=0.649313	lr=0.270675
Epoch[49] Batch [399]	Speed: 4174.188244 samples/sec	accuracy=0.649414	lr=0.270470
Epoch[49] Batch [449]	Speed: 4121.415242 samples/sec	accuracy=0.649432	lr=0.270266
Epoch[49] Batch [499]	Speed: 4235.861692 samples/sec	accuracy=0.649302	lr=0.270061
Epoch[49] Batch [549]	Speed: 4220.224504 samples/sec	accuracy=0.649294	lr=0.269857
Epoch[49] Batch [599]	Speed: 4286.824045 samples/sec	accuracy=0.649323	lr=0.269652
Epoch[49] Batch [649]	Speed: 4255.127136 samples/sec	accuracy=0.649427	lr=0.269447
Epoch[49] Batch [699]	Speed: 4180.782923 samples/sec	accuracy=0.649558	lr=0.269243
Epoch[49] Batch [749]	Speed: 4139.947380 samples/sec	accuracy=0.649754	lr=0.269038
Epoch[49] Batch [799]	Speed: 4209.815526 samples/sec	accuracy=0.649753	lr=0.268833
Epoch[49] Batch [849]	Speed: 4259.124464 samples/sec	accuracy=0.649762	lr=0.268628
Epoch[49] Batch [899]	Speed: 4167.922030 samples/sec	accuracy=0.649838	lr=0.268422
Epoch[49] Batch [949]	Speed: 4241.305786 samples/sec	accuracy=0.649951	lr=0.268217
Epoch[49] Batch [999]	Speed: 4345.676122 samples/sec	accuracy=0.649992	lr=0.268012
Epoch[49] Batch [1049]	Speed: 4253.080577 samples/sec	accuracy=0.650092	lr=0.267807
Epoch[49] Batch [1099]	Speed: 4166.575440 samples/sec	accuracy=0.650093	lr=0.267601
Epoch[49] Batch [1149]	Speed: 4080.886054 samples/sec	accuracy=0.650156	lr=0.267396
Epoch[49] Batch [1199]	Speed: 4133.830103 samples/sec	accuracy=0.650268	lr=0.267190
Epoch[49] Batch [1249]	Speed: 4300.887468 samples/sec	accuracy=0.650341	lr=0.266984
[Epoch 49] training: accuracy=0.650338
[Epoch 49] speed: 4196 samples/sec	time cost: 316.603394
[Epoch 49] validation: err-top1=0.357601 err-top5=0.136500
Epoch[50] Batch [49]	Speed: 4005.448361 samples/sec	accuracy=0.649134	lr=0.266774
Epoch[50] Batch [99]	Speed: 4112.983961 samples/sec	accuracy=0.649106	lr=0.266568
Epoch[50] Batch [149]	Speed: 4155.893399 samples/sec	accuracy=0.648831	lr=0.266362
Epoch[50] Batch [199]	Speed: 4048.972613 samples/sec	accuracy=0.649154	lr=0.266156
Epoch[50] Batch [249]	Speed: 4046.665976 samples/sec	accuracy=0.649320	lr=0.265950
Epoch[50] Batch [299]	Speed: 4019.663968 samples/sec	accuracy=0.649656	lr=0.265744
Epoch[50] Batch [349]	Speed: 4227.571437 samples/sec	accuracy=0.649558	lr=0.265538
Epoch[50] Batch [399]	Speed: 4105.137997 samples/sec	accuracy=0.649560	lr=0.265331
Epoch[50] Batch [449]	Speed: 4133.212854 samples/sec	accuracy=0.649707	lr=0.265125
Epoch[50] Batch [499]	Speed: 4118.727127 samples/sec	accuracy=0.649805	lr=0.264918
Epoch[50] Batch [549]	Speed: 4199.419993 samples/sec	accuracy=0.649908	lr=0.264712
Epoch[50] Batch [599]	Speed: 4120.271335 samples/sec	accuracy=0.649967	lr=0.264505
Epoch[50] Batch [649]	Speed: 4172.413984 samples/sec	accuracy=0.650005	lr=0.264298
Epoch[50] Batch [699]	Speed: 4186.262860 samples/sec	accuracy=0.650150	lr=0.264092
Epoch[50] Batch [749]	Speed: 4170.592876 samples/sec	accuracy=0.650285	lr=0.263885
Epoch[50] Batch [799]	Speed: 4154.078645 samples/sec	accuracy=0.650330	lr=0.263678
Epoch[50] Batch [849]	Speed: 4243.652624 samples/sec	accuracy=0.650437	lr=0.263471
Epoch[50] Batch [899]	Speed: 4238.847833 samples/sec	accuracy=0.650496	lr=0.263264
Epoch[50] Batch [949]	Speed: 4224.138380 samples/sec	accuracy=0.650582	lr=0.263056
Epoch[50] Batch [999]	Speed: 4205.132298 samples/sec	accuracy=0.650667	lr=0.262849
Epoch[50] Batch [1049]	Speed: 4108.529382 samples/sec	accuracy=0.650795	lr=0.262642
Epoch[50] Batch [1099]	Speed: 4069.674125 samples/sec	accuracy=0.650847	lr=0.262434
Epoch[50] Batch [1149]	Speed: 4079.914357 samples/sec	accuracy=0.650971	lr=0.262227
Epoch[50] Batch [1199]	Speed: 4291.968841 samples/sec	accuracy=0.651070	lr=0.262019
Epoch[50] Batch [1249]	Speed: 4194.639817 samples/sec	accuracy=0.651099	lr=0.261812
[Epoch 50] training: accuracy=0.651108
[Epoch 50] speed: 4141 samples/sec	time cost: 319.498264
[Epoch 50] validation: err-top1=0.349310 err-top5=0.130122
Epoch[51] Batch [49]	Speed: 4047.637538 samples/sec	accuracy=0.650863	lr=0.261600
Epoch[51] Batch [99]	Speed: 4141.758433 samples/sec	accuracy=0.651904	lr=0.261392
Epoch[51] Batch [149]	Speed: 4104.748333 samples/sec	accuracy=0.651950	lr=0.261184
Epoch[51] Batch [199]	Speed: 4213.880438 samples/sec	accuracy=0.651932	lr=0.260976
Epoch[51] Batch [249]	Speed: 4189.884245 samples/sec	accuracy=0.651630	lr=0.260768
Epoch[51] Batch [299]	Speed: 4065.468355 samples/sec	accuracy=0.651386	lr=0.260560
Epoch[51] Batch [349]	Speed: 4226.508093 samples/sec	accuracy=0.651118	lr=0.260352
Epoch[51] Batch [399]	Speed: 4133.647566 samples/sec	accuracy=0.651102	lr=0.260144
Epoch[51] Batch [449]	Speed: 4168.891188 samples/sec	accuracy=0.651149	lr=0.259935
Epoch[51] Batch [499]	Speed: 4190.345842 samples/sec	accuracy=0.651028	lr=0.259727
Epoch[51] Batch [549]	Speed: 4172.809386 samples/sec	accuracy=0.651167	lr=0.259519
Epoch[51] Batch [599]	Speed: 4216.626839 samples/sec	accuracy=0.651241	lr=0.259310
Epoch[51] Batch [649]	Speed: 4199.410302 samples/sec	accuracy=0.651260	lr=0.259102
Epoch[51] Batch [699]	Speed: 4281.812113 samples/sec	accuracy=0.651308	lr=0.258893
Epoch[51] Batch [749]	Speed: 4207.709582 samples/sec	accuracy=0.651373	lr=0.258684
Epoch[51] Batch [799]	Speed: 4171.300581 samples/sec	accuracy=0.651466	lr=0.258475
Epoch[51] Batch [849]	Speed: 4142.661597 samples/sec	accuracy=0.651476	lr=0.258267
Epoch[51] Batch [899]	Speed: 4201.461337 samples/sec	accuracy=0.651504	lr=0.258058
Epoch[51] Batch [949]	Speed: 4256.595197 samples/sec	accuracy=0.651612	lr=0.257849
Epoch[51] Batch [999]	Speed: 4272.072236 samples/sec	accuracy=0.651675	lr=0.257640
Epoch[51] Batch [1049]	Speed: 4122.266032 samples/sec	accuracy=0.651780	lr=0.257430
Epoch[51] Batch [1099]	Speed: 4157.711430 samples/sec	accuracy=0.651883	lr=0.257221
Epoch[51] Batch [1149]	Speed: 4149.852434 samples/sec	accuracy=0.651958	lr=0.257012
Epoch[51] Batch [1199]	Speed: 4021.228970 samples/sec	accuracy=0.652074	lr=0.256803
Epoch[51] Batch [1249]	Speed: 4168.035041 samples/sec	accuracy=0.652217	lr=0.256593
[Epoch 51] training: accuracy=0.652222
[Epoch 51] speed: 4164 samples/sec	time cost: 317.694889
[Epoch 51] validation: err-top1=0.351981 err-top5=0.131876
Epoch[52] Batch [49]	Speed: 4129.170767 samples/sec	accuracy=0.651905	lr=0.256379
Epoch[52] Batch [99]	Speed: 4210.565748 samples/sec	accuracy=0.652185	lr=0.256170
Epoch[52] Batch [149]	Speed: 4207.415770 samples/sec	accuracy=0.652033	lr=0.255960
Epoch[52] Batch [199]	Speed: 4136.951789 samples/sec	accuracy=0.651751	lr=0.255751
Epoch[52] Batch [249]	Speed: 4173.795907 samples/sec	accuracy=0.651671	lr=0.255541
Epoch[52] Batch [299]	Speed: 4094.371429 samples/sec	accuracy=0.651801	lr=0.255331
Epoch[52] Batch [349]	Speed: 4205.388155 samples/sec	accuracy=0.651934	lr=0.255121
Epoch[52] Batch [399]	Speed: 4141.501234 samples/sec	accuracy=0.651914	lr=0.254911
Epoch[52] Batch [449]	Speed: 4162.301107 samples/sec	accuracy=0.652145	lr=0.254701
Epoch[52] Batch [499]	Speed: 4189.264446 samples/sec	accuracy=0.652282	lr=0.254491
Epoch[52] Batch [549]	Speed: 4017.898105 samples/sec	accuracy=0.652340	lr=0.254281
Epoch[52] Batch [599]	Speed: 4110.811728 samples/sec	accuracy=0.652405	lr=0.254071
Epoch[52] Batch [649]	Speed: 4150.527286 samples/sec	accuracy=0.652511	lr=0.253861
Epoch[52] Batch [699]	Speed: 4242.764743 samples/sec	accuracy=0.652602	lr=0.253650
Epoch[52] Batch [749]	Speed: 4182.305032 samples/sec	accuracy=0.652672	lr=0.253440
Epoch[52] Batch [799]	Speed: 4228.265228 samples/sec	accuracy=0.652751	lr=0.253229
Epoch[52] Batch [849]	Speed: 4276.321508 samples/sec	accuracy=0.652811	lr=0.253019
Epoch[52] Batch [899]	Speed: 4151.889770 samples/sec	accuracy=0.652929	lr=0.252808
Epoch[52] Batch [949]	Speed: 4163.079521 samples/sec	accuracy=0.653024	lr=0.252598
Epoch[52] Batch [999]	Speed: 4061.133638 samples/sec	accuracy=0.653153	lr=0.252387
Epoch[52] Batch [1049]	Speed: 4181.560694 samples/sec	accuracy=0.653218	lr=0.252176
Epoch[52] Batch [1099]	Speed: 4157.164204 samples/sec	accuracy=0.653343	lr=0.251965
Epoch[52] Batch [1149]	Speed: 4320.785980 samples/sec	accuracy=0.653403	lr=0.251754
Epoch[52] Batch [1199]	Speed: 4368.485855 samples/sec	accuracy=0.653525	lr=0.251543
Epoch[52] Batch [1249]	Speed: 4281.944703 samples/sec	accuracy=0.653616	lr=0.251332
[Epoch 52] training: accuracy=0.653619
[Epoch 52] speed: 4177 samples/sec	time cost: 317.155063
[Epoch 52] validation: err-top1=0.354106 err-top5=0.133260
Epoch[53] Batch [49]	Speed: 4122.030792 samples/sec	accuracy=0.654581	lr=0.251117
Epoch[53] Batch [99]	Speed: 4230.090483 samples/sec	accuracy=0.654519	lr=0.250906
Epoch[53] Batch [149]	Speed: 4166.716835 samples/sec	accuracy=0.653781	lr=0.250695
Epoch[53] Batch [199]	Speed: 4151.727788 samples/sec	accuracy=0.653449	lr=0.250483
Epoch[53] Batch [249]	Speed: 4162.312563 samples/sec	accuracy=0.653117	lr=0.250272
Epoch[53] Batch [299]	Speed: 4055.270477 samples/sec	accuracy=0.653247	lr=0.250061
Epoch[53] Batch [349]	Speed: 4140.997393 samples/sec	accuracy=0.653165	lr=0.249849
Epoch[53] Batch [399]	Speed: 4203.037534 samples/sec	accuracy=0.653024	lr=0.249638
Epoch[53] Batch [449]	Speed: 4099.447509 samples/sec	accuracy=0.653094	lr=0.249426
Epoch[53] Batch [499]	Speed: 4211.480176 samples/sec	accuracy=0.653074	lr=0.249215
Epoch[53] Batch [549]	Speed: 4153.055640 samples/sec	accuracy=0.653107	lr=0.249003
Epoch[53] Batch [599]	Speed: 4180.212276 samples/sec	accuracy=0.653131	lr=0.248791
Epoch[53] Batch [649]	Speed: 4133.069985 samples/sec	accuracy=0.653274	lr=0.248579
Epoch[53] Batch [699]	Speed: 4216.277559 samples/sec	accuracy=0.653562	lr=0.248367
Epoch[53] Batch [749]	Speed: 4192.520541 samples/sec	accuracy=0.653680	lr=0.248156
Epoch[53] Batch [799]	Speed: 4134.870406 samples/sec	accuracy=0.653801	lr=0.247944
Epoch[53] Batch [849]	Speed: 4267.326287 samples/sec	accuracy=0.653937	lr=0.247732
Epoch[53] Batch [899]	Speed: 4083.201289 samples/sec	accuracy=0.654079	lr=0.247519
Epoch[53] Batch [949]	Speed: 4261.485740 samples/sec	accuracy=0.654076	lr=0.247307
Epoch[53] Batch [999]	Speed: 4161.275098 samples/sec	accuracy=0.654150	lr=0.247095
Epoch[53] Batch [1049]	Speed: 4143.124837 samples/sec	accuracy=0.654218	lr=0.246883
Epoch[53] Batch [1099]	Speed: 4241.378832 samples/sec	accuracy=0.654262	lr=0.246670
Epoch[53] Batch [1149]	Speed: 4141.421126 samples/sec	accuracy=0.654381	lr=0.246458
Epoch[53] Batch [1199]	Speed: 4245.698932 samples/sec	accuracy=0.654491	lr=0.246246
Epoch[53] Batch [1249]	Speed: 4251.304275 samples/sec	accuracy=0.654545	lr=0.246033
[Epoch 53] training: accuracy=0.654549
[Epoch 53] speed: 4168 samples/sec	time cost: 317.858228
[Epoch 53] validation: err-top1=0.351483 err-top5=0.131597
Epoch[54] Batch [49]	Speed: 4147.580613 samples/sec	accuracy=0.655025	lr=0.245816
Epoch[54] Batch [99]	Speed: 4350.355941 samples/sec	accuracy=0.654368	lr=0.245604
Epoch[54] Batch [149]	Speed: 4121.307514 samples/sec	accuracy=0.654143	lr=0.245391
Epoch[54] Batch [199]	Speed: 4308.474622 samples/sec	accuracy=0.654036	lr=0.245178
Epoch[54] Batch [249]	Speed: 4245.130063 samples/sec	accuracy=0.654155	lr=0.244966
Epoch[54] Batch [299]	Speed: 4138.046211 samples/sec	accuracy=0.654203	lr=0.244753
Epoch[54] Batch [349]	Speed: 4250.322083 samples/sec	accuracy=0.654256	lr=0.244540
Epoch[54] Batch [399]	Speed: 4139.439290 samples/sec	accuracy=0.654145	lr=0.244327
Epoch[54] Batch [449]	Speed: 3984.122031 samples/sec	accuracy=0.654164	lr=0.244114
Epoch[54] Batch [499]	Speed: 4103.881072 samples/sec	accuracy=0.654143	lr=0.243901
Epoch[54] Batch [549]	Speed: 4051.016233 samples/sec	accuracy=0.654198	lr=0.243688
Epoch[54] Batch [599]	Speed: 4122.862602 samples/sec	accuracy=0.654343	lr=0.243475
Epoch[54] Batch [649]	Speed: 4119.732971 samples/sec	accuracy=0.654429	lr=0.243262
Epoch[54] Batch [699]	Speed: 4124.526834 samples/sec	accuracy=0.654617	lr=0.243049
Epoch[54] Batch [749]	Speed: 4071.742175 samples/sec	accuracy=0.654778	lr=0.242835
Epoch[54] Batch [799]	Speed: 4194.187595 samples/sec	accuracy=0.654831	lr=0.242622
Epoch[54] Batch [849]	Speed: 4092.661005 samples/sec	accuracy=0.654915	lr=0.242409
Epoch[54] Batch [899]	Speed: 4044.184884 samples/sec	accuracy=0.654936	lr=0.242195
Epoch[54] Batch [949]	Speed: 4086.378047 samples/sec	accuracy=0.655050	lr=0.241982
Epoch[54] Batch [999]	Speed: 4151.048213 samples/sec	accuracy=0.655130	lr=0.241768
Epoch[54] Batch [1049]	Speed: 4123.348343 samples/sec	accuracy=0.655217	lr=0.241555
Epoch[54] Batch [1099]	Speed: 4054.673785 samples/sec	accuracy=0.655346	lr=0.241341
Epoch[54] Batch [1149]	Speed: 4189.848195 samples/sec	accuracy=0.655449	lr=0.241127
Epoch[54] Batch [1199]	Speed: 4099.573819 samples/sec	accuracy=0.655550	lr=0.240914
Epoch[54] Batch [1249]	Speed: 4231.971856 samples/sec	accuracy=0.655673	lr=0.240700
[Epoch 54] training: accuracy=0.655676
[Epoch 54] speed: 4136 samples/sec	time cost: 319.650552
[Epoch 54] validation: err-top1=0.346640 err-top5=0.130859
Epoch[55] Batch [49]	Speed: 4155.180058 samples/sec	accuracy=0.655586	lr=0.240482
Epoch[55] Batch [99]	Speed: 4029.392230 samples/sec	accuracy=0.655287	lr=0.240268
Epoch[55] Batch [149]	Speed: 4087.023621 samples/sec	accuracy=0.654737	lr=0.240054
Epoch[55] Batch [199]	Speed: 4069.131090 samples/sec	accuracy=0.654969	lr=0.239840
Epoch[55] Batch [249]	Speed: 4210.416243 samples/sec	accuracy=0.654928	lr=0.239626
Epoch[55] Batch [299]	Speed: 4094.607193 samples/sec	accuracy=0.654957	lr=0.239412
Epoch[55] Batch [349]	Speed: 4080.767329 samples/sec	accuracy=0.655002	lr=0.239198
Epoch[55] Batch [399]	Speed: 4128.842096 samples/sec	accuracy=0.654897	lr=0.238984
Epoch[55] Batch [449]	Speed: 4119.027011 samples/sec	accuracy=0.655071	lr=0.238769
Epoch[55] Batch [499]	Speed: 4116.755661 samples/sec	accuracy=0.655317	lr=0.238555
Epoch[55] Batch [549]	Speed: 4231.780216 samples/sec	accuracy=0.655324	lr=0.238341
Epoch[55] Batch [599]	Speed: 4075.300451 samples/sec	accuracy=0.655382	lr=0.238126
Epoch[55] Batch [649]	Speed: 4149.249472 samples/sec	accuracy=0.655446	lr=0.237912
Epoch[55] Batch [699]	Speed: 4175.523685 samples/sec	accuracy=0.655541	lr=0.237698
Epoch[55] Batch [749]	Speed: 4044.257086 samples/sec	accuracy=0.655612	lr=0.237483
Epoch[55] Batch [799]	Speed: 4164.677929 samples/sec	accuracy=0.655737	lr=0.237269
Epoch[55] Batch [849]	Speed: 4138.027473 samples/sec	accuracy=0.655880	lr=0.237054
Epoch[55] Batch [899]	Speed: 4113.125759 samples/sec	accuracy=0.656006	lr=0.236839
Epoch[55] Batch [949]	Speed: 4137.296499 samples/sec	accuracy=0.656047	lr=0.236625
Epoch[55] Batch [999]	Speed: 4152.014997 samples/sec	accuracy=0.656138	lr=0.236410
Epoch[55] Batch [1049]	Speed: 4114.214152 samples/sec	accuracy=0.656215	lr=0.236195
Epoch[55] Batch [1099]	Speed: 4192.005275 samples/sec	accuracy=0.656300	lr=0.235981
Epoch[55] Batch [1149]	Speed: 4154.522259 samples/sec	accuracy=0.656399	lr=0.235766
Epoch[55] Batch [1199]	Speed: 4180.065407 samples/sec	accuracy=0.656509	lr=0.235551
Epoch[55] Batch [1249]	Speed: 4270.272566 samples/sec	accuracy=0.656629	lr=0.235336
[Epoch 55] training: accuracy=0.656627
[Epoch 55] speed: 4131 samples/sec	time cost: 320.144204
[Epoch 55] validation: err-top1=0.350865 err-top5=0.130102
Epoch[56] Batch [49]	Speed: 3865.996133 samples/sec	accuracy=0.657096	lr=0.235117
Epoch[56] Batch [99]	Speed: 4057.567080 samples/sec	accuracy=0.656915	lr=0.234902
Epoch[56] Batch [149]	Speed: 4245.055798 samples/sec	accuracy=0.656555	lr=0.234687
Epoch[56] Batch [199]	Speed: 4147.952975 samples/sec	accuracy=0.656382	lr=0.234472
Epoch[56] Batch [249]	Speed: 4153.959319 samples/sec	accuracy=0.656665	lr=0.234256
Epoch[56] Batch [299]	Speed: 4061.145926 samples/sec	accuracy=0.656893	lr=0.234041
Epoch[56] Batch [349]	Speed: 4009.171276 samples/sec	accuracy=0.656877	lr=0.233826
Epoch[56] Batch [399]	Speed: 4171.430871 samples/sec	accuracy=0.656898	lr=0.233611
Epoch[56] Batch [449]	Speed: 4186.646525 samples/sec	accuracy=0.656739	lr=0.233396
Epoch[56] Batch [499]	Speed: 4066.536520 samples/sec	accuracy=0.656656	lr=0.233180
Epoch[56] Batch [549]	Speed: 4183.979051 samples/sec	accuracy=0.656704	lr=0.232965
Epoch[56] Batch [599]	Speed: 4114.025699 samples/sec	accuracy=0.656778	lr=0.232749
Epoch[56] Batch [649]	Speed: 4146.637751 samples/sec	accuracy=0.656859	lr=0.232534
Epoch[56] Batch [699]	Speed: 4127.265519 samples/sec	accuracy=0.656918	lr=0.232319
Epoch[56] Batch [749]	Speed: 4099.685422 samples/sec	accuracy=0.657017	lr=0.232103
Epoch[56] Batch [799]	Speed: 4115.486643 samples/sec	accuracy=0.657069	lr=0.231887
Epoch[56] Batch [849]	Speed: 4048.068933 samples/sec	accuracy=0.657143	lr=0.231672
Epoch[56] Batch [899]	Speed: 4125.289201 samples/sec	accuracy=0.657255	lr=0.231456
Epoch[56] Batch [949]	Speed: 4163.992332 samples/sec	accuracy=0.657350	lr=0.231241
Epoch[56] Batch [999]	Speed: 4101.803524 samples/sec	accuracy=0.657470	lr=0.231025
Epoch[56] Batch [1049]	Speed: 4051.640362 samples/sec	accuracy=0.657529	lr=0.230809
Epoch[56] Batch [1099]	Speed: 4068.950599 samples/sec	accuracy=0.657684	lr=0.230593
Epoch[56] Batch [1149]	Speed: 4158.429666 samples/sec	accuracy=0.657802	lr=0.230377
Epoch[56] Batch [1199]	Speed: 4101.145127 samples/sec	accuracy=0.657892	lr=0.230162
Epoch[56] Batch [1249]	Speed: 4286.395106 samples/sec	accuracy=0.657964	lr=0.229946
[Epoch 56] training: accuracy=0.657969
[Epoch 56] speed: 4108 samples/sec	time cost: 321.809825
[Epoch 56] validation: err-top1=0.344966 err-top5=0.126594
Epoch[57] Batch [49]	Speed: 3970.970346 samples/sec	accuracy=0.659607	lr=0.229725
Epoch[57] Batch [99]	Speed: 4110.682364 samples/sec	accuracy=0.658891	lr=0.229510
Epoch[57] Batch [149]	Speed: 4037.188529 samples/sec	accuracy=0.658376	lr=0.229294
Epoch[57] Batch [199]	Speed: 4123.602024 samples/sec	accuracy=0.658320	lr=0.229077
Epoch[57] Batch [249]	Speed: 4051.705491 samples/sec	accuracy=0.658179	lr=0.228861
Epoch[57] Batch [299]	Speed: 4189.477346 samples/sec	accuracy=0.658062	lr=0.228645
Epoch[57] Batch [349]	Speed: 4131.989642 samples/sec	accuracy=0.657939	lr=0.228429
Epoch[57] Batch [399]	Speed: 4090.223626 samples/sec	accuracy=0.658066	lr=0.228213
Epoch[57] Batch [449]	Speed: 4206.653568 samples/sec	accuracy=0.658043	lr=0.227997
Epoch[57] Batch [499]	Speed: 4112.165738 samples/sec	accuracy=0.658161	lr=0.227781
Epoch[57] Batch [549]	Speed: 4146.680028 samples/sec	accuracy=0.658202	lr=0.227564
Epoch[57] Batch [599]	Speed: 4295.535853 samples/sec	accuracy=0.658236	lr=0.227348
Epoch[57] Batch [649]	Speed: 4249.012358 samples/sec	accuracy=0.658311	lr=0.227132
Epoch[57] Batch [699]	Speed: 4277.112405 samples/sec	accuracy=0.658313	lr=0.226915
Epoch[57] Batch [749]	Speed: 4231.842760 samples/sec	accuracy=0.658365	lr=0.226699
Epoch[57] Batch [799]	Speed: 4159.023217 samples/sec	accuracy=0.658425	lr=0.226482
Epoch[57] Batch [849]	Speed: 4052.988712 samples/sec	accuracy=0.658450	lr=0.226266
Epoch[57] Batch [899]	Speed: 4075.323111 samples/sec	accuracy=0.658559	lr=0.226049
Epoch[57] Batch [949]	Speed: 4138.741875 samples/sec	accuracy=0.658722	lr=0.225833
Epoch[57] Batch [999]	Speed: 4161.169872 samples/sec	accuracy=0.658801	lr=0.225616
Epoch[57] Batch [1049]	Speed: 4133.537447 samples/sec	accuracy=0.658973	lr=0.225400
Epoch[57] Batch [1099]	Speed: 4134.480092 samples/sec	accuracy=0.659079	lr=0.225183
Epoch[57] Batch [1149]	Speed: 4159.703312 samples/sec	accuracy=0.659188	lr=0.224967
Epoch[57] Batch [1199]	Speed: 4184.358222 samples/sec	accuracy=0.659278	lr=0.224750
Epoch[57] Batch [1249]	Speed: 4415.296216 samples/sec	accuracy=0.659391	lr=0.224533
[Epoch 57] training: accuracy=0.659396
[Epoch 57] speed: 4148 samples/sec	time cost: 319.148759
[Epoch 57] validation: err-top1=0.373565 err-top5=0.151566
Epoch[58] Batch [49]	Speed: 4081.378243 samples/sec	accuracy=0.660472	lr=0.224312
Epoch[58] Batch [99]	Speed: 4081.282293 samples/sec	accuracy=0.659952	lr=0.224095
Epoch[58] Batch [149]	Speed: 4164.415292 samples/sec	accuracy=0.659995	lr=0.223879
Epoch[58] Batch [199]	Speed: 4119.601701 samples/sec	accuracy=0.659724	lr=0.223662
Epoch[58] Batch [249]	Speed: 4236.438943 samples/sec	accuracy=0.659823	lr=0.223445
Epoch[58] Batch [299]	Speed: 4211.600682 samples/sec	accuracy=0.659943	lr=0.223228
Epoch[58] Batch [349]	Speed: 4077.299210 samples/sec	accuracy=0.659943	lr=0.223011
Epoch[58] Batch [399]	Speed: 4200.349796 samples/sec	accuracy=0.659764	lr=0.222794
Epoch[58] Batch [449]	Speed: 4178.025369 samples/sec	accuracy=0.659862	lr=0.222577
Epoch[58] Batch [499]	Speed: 4122.384256 samples/sec	accuracy=0.659680	lr=0.222360
Epoch[58] Batch [549]	Speed: 4285.304107 samples/sec	accuracy=0.659781	lr=0.222143
Epoch[58] Batch [599]	Speed: 4106.200103 samples/sec	accuracy=0.659866	lr=0.221926
Epoch[58] Batch [649]	Speed: 4145.720768 samples/sec	accuracy=0.659888	lr=0.221709
Epoch[58] Batch [699]	Speed: 4203.066573 samples/sec	accuracy=0.660050	lr=0.221492
Epoch[58] Batch [749]	Speed: 4143.837162 samples/sec	accuracy=0.660247	lr=0.221275
Epoch[58] Batch [799]	Speed: 4163.188879 samples/sec	accuracy=0.660344	lr=0.221058
Epoch[58] Batch [849]	Speed: 4121.409310 samples/sec	accuracy=0.660387	lr=0.220841
Epoch[58] Batch [899]	Speed: 4231.156714 samples/sec	accuracy=0.660455	lr=0.220623
Epoch[58] Batch [949]	Speed: 4219.613439 samples/sec	accuracy=0.660474	lr=0.220406
Epoch[58] Batch [999]	Speed: 4161.226314 samples/sec	accuracy=0.660546	lr=0.220189
Epoch[58] Batch [1049]	Speed: 4166.671400 samples/sec	accuracy=0.660558	lr=0.219972
Epoch[58] Batch [1099]	Speed: 4143.179512 samples/sec	accuracy=0.660602	lr=0.219754
Epoch[58] Batch [1149]	Speed: 4208.538312 samples/sec	accuracy=0.660727	lr=0.219537
Epoch[58] Batch [1199]	Speed: 4140.387662 samples/sec	accuracy=0.660826	lr=0.219320
Epoch[58] Batch [1249]	Speed: 4315.430637 samples/sec	accuracy=0.660881	lr=0.219102
[Epoch 58] training: accuracy=0.660879
[Epoch 58] speed: 4165 samples/sec	time cost: 318.232634
[Epoch 58] validation: err-top1=0.350220 err-top5=0.128296
Epoch[59] Batch [49]	Speed: 4121.171162 samples/sec	accuracy=0.661310	lr=0.218881
Epoch[59] Batch [99]	Speed: 4164.883168 samples/sec	accuracy=0.661551	lr=0.218663
Epoch[59] Batch [149]	Speed: 3994.888379 samples/sec	accuracy=0.660849	lr=0.218446
Epoch[59] Batch [199]	Speed: 4196.851702 samples/sec	accuracy=0.661139	lr=0.218228
Epoch[59] Batch [249]	Speed: 4218.030007 samples/sec	accuracy=0.660826	lr=0.218011
Epoch[59] Batch [299]	Speed: 4135.591843 samples/sec	accuracy=0.660897	lr=0.217793
Epoch[59] Batch [349]	Speed: 4251.417897 samples/sec	accuracy=0.660921	lr=0.217576
Epoch[59] Batch [399]	Speed: 4156.207166 samples/sec	accuracy=0.660904	lr=0.217358
Epoch[59] Batch [449]	Speed: 4162.417847 samples/sec	accuracy=0.660909	lr=0.217141
Epoch[59] Batch [499]	Speed: 4270.428984 samples/sec	accuracy=0.660946	lr=0.216923
Epoch[59] Batch [549]	Speed: 4082.792646 samples/sec	accuracy=0.661049	lr=0.216706
Epoch[59] Batch [599]	Speed: 4121.214898 samples/sec	accuracy=0.661073	lr=0.216488
Epoch[59] Batch [649]	Speed: 4208.813721 samples/sec	accuracy=0.661025	lr=0.216270
Epoch[59] Batch [699]	Speed: 4147.664645 samples/sec	accuracy=0.661094	lr=0.216053
Epoch[59] Batch [749]	Speed: 4139.026652 samples/sec	accuracy=0.661104	lr=0.215835
Epoch[59] Batch [799]	Speed: 4117.237674 samples/sec	accuracy=0.661135	lr=0.215617
Epoch[59] Batch [849]	Speed: 4235.095831 samples/sec	accuracy=0.661269	lr=0.215400
Epoch[59] Batch [899]	Speed: 4203.303914 samples/sec	accuracy=0.661370	lr=0.215182
Epoch[59] Batch [949]	Speed: 4141.709227 samples/sec	accuracy=0.661464	lr=0.214964
Epoch[59] Batch [999]	Speed: 4121.736166 samples/sec	accuracy=0.661578	lr=0.214746
Epoch[59] Batch [1049]	Speed: 4254.655794 samples/sec	accuracy=0.661673	lr=0.214529
Epoch[59] Batch [1099]	Speed: 4249.908575 samples/sec	accuracy=0.661735	lr=0.214311
Epoch[59] Batch [1149]	Speed: 4318.720865 samples/sec	accuracy=0.661864	lr=0.214093
Epoch[59] Batch [1199]	Speed: 4188.187038 samples/sec	accuracy=0.661957	lr=0.213875
Epoch[59] Batch [1249]	Speed: 4318.926367 samples/sec	accuracy=0.662026	lr=0.213657
[Epoch 59] training: accuracy=0.662027
[Epoch 59] speed: 4176 samples/sec	time cost: 317.193004
[Epoch 59] validation: err-top1=0.338429 err-top5=0.122030
Epoch[60] Batch [49]	Speed: 4239.505910 samples/sec	accuracy=0.662704	lr=0.213435
Epoch[60] Batch [99]	Speed: 4121.797872 samples/sec	accuracy=0.661678	lr=0.213217
Epoch[60] Batch [149]	Speed: 4070.175878 samples/sec	accuracy=0.661529	lr=0.212999
Epoch[60] Batch [199]	Speed: 4141.407389 samples/sec	accuracy=0.661351	lr=0.212781
Epoch[60] Batch [249]	Speed: 4110.681026 samples/sec	accuracy=0.661041	lr=0.212563
Epoch[60] Batch [299]	Speed: 4001.059619 samples/sec	accuracy=0.661154	lr=0.212345
Epoch[60] Batch [349]	Speed: 4147.496024 samples/sec	accuracy=0.661281	lr=0.212127
Epoch[60] Batch [399]	Speed: 4165.929302 samples/sec	accuracy=0.661233	lr=0.211909
Epoch[60] Batch [449]	Speed: 4365.398105 samples/sec	accuracy=0.661385	lr=0.211691
Epoch[60] Batch [499]	Speed: 4016.420577 samples/sec	accuracy=0.661472	lr=0.211473
Epoch[60] Batch [549]	Speed: 4167.420960 samples/sec	accuracy=0.661454	lr=0.211255
Epoch[60] Batch [599]	Speed: 4019.115015 samples/sec	accuracy=0.661607	lr=0.211037
Epoch[60] Batch [649]	Speed: 4174.648662 samples/sec	accuracy=0.661748	lr=0.210819
Epoch[60] Batch [699]	Speed: 4084.504851 samples/sec	accuracy=0.661800	lr=0.210601
Epoch[60] Batch [749]	Speed: 4156.089407 samples/sec	accuracy=0.661874	lr=0.210383
Epoch[60] Batch [799]	Speed: 4117.444263 samples/sec	accuracy=0.661921	lr=0.210165
Epoch[60] Batch [849]	Speed: 4293.862571 samples/sec	accuracy=0.662078	lr=0.209947
Epoch[60] Batch [899]	Speed: 4259.689992 samples/sec	accuracy=0.662241	lr=0.209729
Epoch[60] Batch [949]	Speed: 4215.246696 samples/sec	accuracy=0.662300	lr=0.209511
Epoch[60] Batch [999]	Speed: 4268.694251 samples/sec	accuracy=0.662392	lr=0.209293
Epoch[60] Batch [1049]	Speed: 4135.513317 samples/sec	accuracy=0.662522	lr=0.209075
Epoch[60] Batch [1099]	Speed: 4192.810639 samples/sec	accuracy=0.662642	lr=0.208856
Epoch[60] Batch [1149]	Speed: 4199.252557 samples/sec	accuracy=0.662748	lr=0.208638
Epoch[60] Batch [1199]	Speed: 4136.429853 samples/sec	accuracy=0.662937	lr=0.208420
Epoch[60] Batch [1249]	Speed: 4389.111096 samples/sec	accuracy=0.663043	lr=0.208202
[Epoch 60] training: accuracy=0.663047
[Epoch 60] speed: 4161 samples/sec	time cost: 319.380538
[Epoch 60] validation: err-top1=0.326331 err-top5=0.117307
Epoch[61] Batch [49]	Speed: 4344.239223 samples/sec	accuracy=0.664291	lr=0.207979
Epoch[61] Batch [99]	Speed: 4188.383736 samples/sec	accuracy=0.662855	lr=0.207761
Epoch[61] Batch [149]	Speed: 4293.463983 samples/sec	accuracy=0.662436	lr=0.207543
Epoch[61] Batch [199]	Speed: 4213.155732 samples/sec	accuracy=0.662368	lr=0.207325
Epoch[61] Batch [249]	Speed: 4204.294208 samples/sec	accuracy=0.662359	lr=0.207106
Epoch[61] Batch [299]	Speed: 4157.271158 samples/sec	accuracy=0.662475	lr=0.206888
Epoch[61] Batch [349]	Speed: 4233.059061 samples/sec	accuracy=0.662500	lr=0.206670
Epoch[61] Batch [399]	Speed: 4277.321804 samples/sec	accuracy=0.662569	lr=0.206452
Epoch[61] Batch [449]	Speed: 4137.211292 samples/sec	accuracy=0.662583	lr=0.206233
Epoch[61] Batch [499]	Speed: 4192.614998 samples/sec	accuracy=0.662695	lr=0.206015
Epoch[61] Batch [549]	Speed: 4110.940470 samples/sec	accuracy=0.662876	lr=0.205797
Epoch[61] Batch [599]	Speed: 4222.785202 samples/sec	accuracy=0.663006	lr=0.205579
Epoch[61] Batch [649]	Speed: 4237.257627 samples/sec	accuracy=0.663174	lr=0.205360
Epoch[61] Batch [699]	Speed: 4261.565909 samples/sec	accuracy=0.663329	lr=0.205142
Epoch[61] Batch [749]	Speed: 4235.497940 samples/sec	accuracy=0.663397	lr=0.204924
Epoch[61] Batch [799]	Speed: 4175.396954 samples/sec	accuracy=0.663460	lr=0.204705
Epoch[61] Batch [849]	Speed: 4127.015352 samples/sec	accuracy=0.663610	lr=0.204487
Epoch[61] Batch [899]	Speed: 4199.198855 samples/sec	accuracy=0.663697	lr=0.204269
Epoch[61] Batch [949]	Speed: 4223.048526 samples/sec	accuracy=0.663823	lr=0.204051
Epoch[61] Batch [999]	Speed: 4311.039108 samples/sec	accuracy=0.663935	lr=0.203832
Epoch[61] Batch [1049]	Speed: 4219.654481 samples/sec	accuracy=0.664055	lr=0.203614
Epoch[61] Batch [1099]	Speed: 4223.926180 samples/sec	accuracy=0.664100	lr=0.203396
Epoch[61] Batch [1149]	Speed: 4243.439801 samples/sec	accuracy=0.664190	lr=0.203177
Epoch[61] Batch [1199]	Speed: 4243.985738 samples/sec	accuracy=0.664304	lr=0.202959
Epoch[61] Batch [1249]	Speed: 4357.789892 samples/sec	accuracy=0.664407	lr=0.202740
[Epoch 61] training: accuracy=0.664414
[Epoch 61] speed: 4219 samples/sec	time cost: 314.644660
[Epoch 61] validation: err-top1=0.330038 err-top5=0.118742
Epoch[62] Batch [49]	Speed: 4046.635246 samples/sec	accuracy=0.665397	lr=0.202518
Epoch[62] Batch [99]	Speed: 4193.436319 samples/sec	accuracy=0.664216	lr=0.202299
Epoch[62] Batch [149]	Speed: 4038.412522 samples/sec	accuracy=0.663728	lr=0.202081
Epoch[62] Batch [199]	Speed: 4246.650100 samples/sec	accuracy=0.663374	lr=0.201863
Epoch[62] Batch [249]	Speed: 4177.933925 samples/sec	accuracy=0.663250	lr=0.201644
Epoch[62] Batch [299]	Speed: 4117.648267 samples/sec	accuracy=0.663342	lr=0.201426
Epoch[62] Batch [349]	Speed: 4181.833234 samples/sec	accuracy=0.663403	lr=0.201208
Epoch[62] Batch [399]	Speed: 4146.452080 samples/sec	accuracy=0.663467	lr=0.200989
Epoch[62] Batch [449]	Speed: 4227.050098 samples/sec	accuracy=0.663725	lr=0.200771
Epoch[62] Batch [499]	Speed: 4242.861729 samples/sec	accuracy=0.663824	lr=0.200552
Epoch[62] Batch [549]	Speed: 4270.821863 samples/sec	accuracy=0.663915	lr=0.200334
Epoch[62] Batch [599]	Speed: 4190.834856 samples/sec	accuracy=0.664155	lr=0.200116
Epoch[62] Batch [649]	Speed: 4247.351764 samples/sec	accuracy=0.664224	lr=0.199897
Epoch[62] Batch [699]	Speed: 4055.007752 samples/sec	accuracy=0.664333	lr=0.199679
Epoch[62] Batch [749]	Speed: 4201.007314 samples/sec	accuracy=0.664484	lr=0.199461
Epoch[62] Batch [799]	Speed: 4160.828751 samples/sec	accuracy=0.664603	lr=0.199242
Epoch[62] Batch [849]	Speed: 4068.703211 samples/sec	accuracy=0.664668	lr=0.199024
Epoch[62] Batch [899]	Speed: 4175.435760 samples/sec	accuracy=0.664751	lr=0.198806
Epoch[62] Batch [949]	Speed: 4271.059103 samples/sec	accuracy=0.664879	lr=0.198587
Epoch[62] Batch [999]	Speed: 4266.200562 samples/sec	accuracy=0.665000	lr=0.198369
Epoch[62] Batch [1049]	Speed: 4109.561550 samples/sec	accuracy=0.665099	lr=0.198150
Epoch[62] Batch [1099]	Speed: 4246.151416 samples/sec	accuracy=0.665229	lr=0.197932
Epoch[62] Batch [1149]	Speed: 4163.183795 samples/sec	accuracy=0.665365	lr=0.197714
Epoch[62] Batch [1199]	Speed: 4258.083099 samples/sec	accuracy=0.665493	lr=0.197495
Epoch[62] Batch [1249]	Speed: 4270.074894 samples/sec	accuracy=0.665687	lr=0.197277
[Epoch 62] training: accuracy=0.665688
[Epoch 62] speed: 4178 samples/sec	time cost: 316.741171
[Epoch 62] validation: err-top1=0.333486 err-top5=0.118862
Epoch[63] Batch [49]	Speed: 4091.720336 samples/sec	accuracy=0.666831	lr=0.197054
Epoch[63] Batch [99]	Speed: 4226.702417 samples/sec	accuracy=0.667032	lr=0.196836
Epoch[63] Batch [149]	Speed: 4089.179268 samples/sec	accuracy=0.666860	lr=0.196618
Epoch[63] Batch [199]	Speed: 4173.843526 samples/sec	accuracy=0.666641	lr=0.196399
Epoch[63] Batch [249]	Speed: 4245.689782 samples/sec	accuracy=0.666667	lr=0.196181
Epoch[63] Batch [299]	Speed: 4218.539427 samples/sec	accuracy=0.666613	lr=0.195963
Epoch[63] Batch [349]	Speed: 4173.801342 samples/sec	accuracy=0.666642	lr=0.195744
Epoch[63] Batch [399]	Speed: 4095.551455 samples/sec	accuracy=0.666574	lr=0.195526
Epoch[63] Batch [449]	Speed: 4261.087813 samples/sec	accuracy=0.666504	lr=0.195308
Epoch[63] Batch [499]	Speed: 4239.591532 samples/sec	accuracy=0.666512	lr=0.195089
Epoch[63] Batch [549]	Speed: 4179.939540 samples/sec	accuracy=0.666572	lr=0.194871
Epoch[63] Batch [599]	Speed: 4155.360481 samples/sec	accuracy=0.666524	lr=0.194653
Epoch[63] Batch [649]	Speed: 4097.259606 samples/sec	accuracy=0.666506	lr=0.194434
Epoch[63] Batch [699]	Speed: 4155.241403 samples/sec	accuracy=0.666577	lr=0.194216
Epoch[63] Batch [749]	Speed: 4161.911405 samples/sec	accuracy=0.666648	lr=0.193998
Epoch[63] Batch [799]	Speed: 4326.971655 samples/sec	accuracy=0.666748	lr=0.193780
Epoch[63] Batch [849]	Speed: 4181.993502 samples/sec	accuracy=0.666802	lr=0.193561
Epoch[63] Batch [899]	Speed: 4201.333767 samples/sec	accuracy=0.666808	lr=0.193343
Epoch[63] Batch [949]	Speed: 4180.959065 samples/sec	accuracy=0.666952	lr=0.193125
Epoch[63] Batch [999]	Speed: 4181.816948 samples/sec	accuracy=0.667119	lr=0.192907
Epoch[63] Batch [1049]	Speed: 4305.141724 samples/sec	accuracy=0.667188	lr=0.192688
Epoch[63] Batch [1099]	Speed: 4179.412720 samples/sec	accuracy=0.667279	lr=0.192470
Epoch[63] Batch [1149]	Speed: 4179.103653 samples/sec	accuracy=0.667360	lr=0.192252
Epoch[63] Batch [1199]	Speed: 4111.468982 samples/sec	accuracy=0.667436	lr=0.192034
Epoch[63] Batch [1249]	Speed: 4248.237113 samples/sec	accuracy=0.667538	lr=0.191816
[Epoch 63] training: accuracy=0.667539
[Epoch 63] speed: 4181 samples/sec	time cost: 316.361362
[Epoch 63] validation: err-top1=0.326803 err-top5=0.115845
Epoch[64] Batch [49]	Speed: 4159.519933 samples/sec	accuracy=0.668484	lr=0.191593
Epoch[64] Batch [99]	Speed: 4194.480381 samples/sec	accuracy=0.667175	lr=0.191375
Epoch[64] Batch [149]	Speed: 4085.011667 samples/sec	accuracy=0.667179	lr=0.191157
Epoch[64] Batch [199]	Speed: 4206.799344 samples/sec	accuracy=0.667478	lr=0.190939
Epoch[64] Batch [249]	Speed: 4094.265344 samples/sec	accuracy=0.667389	lr=0.190720
Epoch[64] Batch [299]	Speed: 4111.108493 samples/sec	accuracy=0.667680	lr=0.190502
Epoch[64] Batch [349]	Speed: 4287.608388 samples/sec	accuracy=0.667557	lr=0.190284
Epoch[64] Batch [399]	Speed: 4063.157958 samples/sec	accuracy=0.667513	lr=0.190066
Epoch[64] Batch [449]	Speed: 4215.285170 samples/sec	accuracy=0.667507	lr=0.189848
Epoch[64] Batch [499]	Speed: 4218.634314 samples/sec	accuracy=0.667571	lr=0.189630
Epoch[64] Batch [549]	Speed: 4181.742112 samples/sec	accuracy=0.667556	lr=0.189412
Epoch[64] Batch [599]	Speed: 4161.734122 samples/sec	accuracy=0.667582	lr=0.189194
Epoch[64] Batch [649]	Speed: 4290.250509 samples/sec	accuracy=0.667752	lr=0.188976
Epoch[64] Batch [699]	Speed: 4190.169563 samples/sec	accuracy=0.667955	lr=0.188758
Epoch[64] Batch [749]	Speed: 4181.152723 samples/sec	accuracy=0.668096	lr=0.188540
Epoch[64] Batch [799]	Speed: 4056.782251 samples/sec	accuracy=0.668172	lr=0.188322
Epoch[64] Batch [849]	Speed: 4206.988811 samples/sec	accuracy=0.668271	lr=0.188104
Epoch[64] Batch [899]	Speed: 4049.199359 samples/sec	accuracy=0.668335	lr=0.187886
Epoch[64] Batch [949]	Speed: 4175.501927 samples/sec	accuracy=0.668464	lr=0.187668
Epoch[64] Batch [999]	Speed: 4152.840403 samples/sec	accuracy=0.668504	lr=0.187450
Epoch[64] Batch [1049]	Speed: 4157.268825 samples/sec	accuracy=0.668603	lr=0.187232
Epoch[64] Batch [1099]	Speed: 4137.226277 samples/sec	accuracy=0.668702	lr=0.187014
Epoch[64] Batch [1149]	Speed: 4100.249795 samples/sec	accuracy=0.668807	lr=0.186796
Epoch[64] Batch [1199]	Speed: 4205.395814 samples/sec	accuracy=0.668904	lr=0.186578
Epoch[64] Batch [1249]	Speed: 4108.380983 samples/sec	accuracy=0.668994	lr=0.186360
[Epoch 64] training: accuracy=0.668999
[Epoch 64] speed: 4155 samples/sec	time cost: 319.107941
[Epoch 64] validation: err-top1=0.325056 err-top5=0.114457
Epoch[65] Batch [49]	Speed: 4230.326803 samples/sec	accuracy=0.669083	lr=0.186138
Epoch[65] Batch [99]	Speed: 4124.603359 samples/sec	accuracy=0.669107	lr=0.185920
Epoch[65] Batch [149]	Speed: 4268.640541 samples/sec	accuracy=0.668763	lr=0.185702
Epoch[65] Batch [199]	Speed: 4200.151316 samples/sec	accuracy=0.668436	lr=0.185485
Epoch[65] Batch [249]	Speed: 3989.442022 samples/sec	accuracy=0.668391	lr=0.185267
Epoch[65] Batch [299]	Speed: 4210.044056 samples/sec	accuracy=0.668421	lr=0.185049
Epoch[65] Batch [349]	Speed: 4074.865397 samples/sec	accuracy=0.668358	lr=0.184831
Epoch[65] Batch [399]	Speed: 4095.117456 samples/sec	accuracy=0.668379	lr=0.184614
Epoch[65] Batch [449]	Speed: 4056.909854 samples/sec	accuracy=0.668331	lr=0.184396
Epoch[65] Batch [499]	Speed: 4153.640910 samples/sec	accuracy=0.668291	lr=0.184178
Epoch[65] Batch [549]	Speed: 4146.777396 samples/sec	accuracy=0.668288	lr=0.183960
Epoch[65] Batch [599]	Speed: 4112.323939 samples/sec	accuracy=0.668349	lr=0.183743
Epoch[65] Batch [649]	Speed: 4174.644685 samples/sec	accuracy=0.668512	lr=0.183525
Epoch[65] Batch [699]	Speed: 4262.501442 samples/sec	accuracy=0.668608	lr=0.183308
Epoch[65] Batch [749]	Speed: 4140.963218 samples/sec	accuracy=0.668756	lr=0.183090
Epoch[65] Batch [799]	Speed: 4136.765311 samples/sec	accuracy=0.668831	lr=0.182872
Epoch[65] Batch [849]	Speed: 4127.690096 samples/sec	accuracy=0.668913	lr=0.182655
Epoch[65] Batch [899]	Speed: 4048.333965 samples/sec	accuracy=0.669000	lr=0.182437
Epoch[65] Batch [949]	Speed: 4175.679978 samples/sec	accuracy=0.669082	lr=0.182220
Epoch[65] Batch [999]	Speed: 4282.291713 samples/sec	accuracy=0.669181	lr=0.182002
Epoch[65] Batch [1049]	Speed: 4115.657641 samples/sec	accuracy=0.669315	lr=0.181785
Epoch[65] Batch [1099]	Speed: 4250.498832 samples/sec	accuracy=0.669445	lr=0.181567
Epoch[65] Batch [1149]	Speed: 4183.737285 samples/sec	accuracy=0.669552	lr=0.181350
Epoch[65] Batch [1199]	Speed: 4061.359521 samples/sec	accuracy=0.669622	lr=0.181132
Epoch[65] Batch [1249]	Speed: 4383.661033 samples/sec	accuracy=0.669702	lr=0.180915
[Epoch 65] training: accuracy=0.669702
[Epoch 65] speed: 4154 samples/sec	time cost: 318.626540
[Epoch 65] validation: err-top1=0.326371 err-top5=0.115354
Epoch[66] Batch [49]	Speed: 3923.779698 samples/sec	accuracy=0.671171	lr=0.180693
Epoch[66] Batch [99]	Speed: 4080.988810 samples/sec	accuracy=0.670517	lr=0.180476
Epoch[66] Batch [149]	Speed: 4332.243129 samples/sec	accuracy=0.670394	lr=0.180259
Epoch[66] Batch [199]	Speed: 4137.593035 samples/sec	accuracy=0.670392	lr=0.180041
Epoch[66] Batch [249]	Speed: 4211.753492 samples/sec	accuracy=0.670290	lr=0.179824
Epoch[66] Batch [299]	Speed: 4215.200445 samples/sec	accuracy=0.670215	lr=0.179607
Epoch[66] Batch [349]	Speed: 4189.668933 samples/sec	accuracy=0.670311	lr=0.179390
Epoch[66] Batch [399]	Speed: 4225.297465 samples/sec	accuracy=0.670389	lr=0.179172
Epoch[66] Batch [449]	Speed: 4071.526251 samples/sec	accuracy=0.670365	lr=0.178955
Epoch[66] Batch [499]	Speed: 4170.421333 samples/sec	accuracy=0.670398	lr=0.178738
Epoch[66] Batch [549]	Speed: 4120.731004 samples/sec	accuracy=0.670484	lr=0.178521
Epoch[66] Batch [599]	Speed: 4140.384868 samples/sec	accuracy=0.670553	lr=0.178304
Epoch[66] Batch [649]	Speed: 4108.642495 samples/sec	accuracy=0.670541	lr=0.178087
Epoch[66] Batch [699]	Speed: 4068.490539 samples/sec	accuracy=0.670587	lr=0.177870
Epoch[66] Batch [749]	Speed: 4228.637064 samples/sec	accuracy=0.670744	lr=0.177653
Epoch[66] Batch [799]	Speed: 4190.749557 samples/sec	accuracy=0.670749	lr=0.177436
Epoch[66] Batch [849]	Speed: 4209.839789 samples/sec	accuracy=0.670776	lr=0.177219
Epoch[66] Batch [899]	Speed: 4124.141162 samples/sec	accuracy=0.670957	lr=0.177002
Epoch[66] Batch [949]	Speed: 4220.233875 samples/sec	accuracy=0.671052	lr=0.176785
Epoch[66] Batch [999]	Speed: 4123.242413 samples/sec	accuracy=0.671180	lr=0.176568
Epoch[66] Batch [1049]	Speed: 4197.572860 samples/sec	accuracy=0.671272	lr=0.176351
Epoch[66] Batch [1099]	Speed: 4171.723082 samples/sec	accuracy=0.671355	lr=0.176134
Epoch[66] Batch [1149]	Speed: 4106.683575 samples/sec	accuracy=0.671458	lr=0.175918
Epoch[66] Batch [1199]	Speed: 4042.071001 samples/sec	accuracy=0.671538	lr=0.175701
Epoch[66] Batch [1249]	Speed: 4172.403689 samples/sec	accuracy=0.671614	lr=0.175484
[Epoch 66] training: accuracy=0.671615
[Epoch 66] speed: 4146 samples/sec	time cost: 319.949813
[Epoch 66] validation: err-top1=0.335459 err-top5=0.120297
Epoch[67] Batch [49]	Speed: 4011.564617 samples/sec	accuracy=0.671900	lr=0.175263
Epoch[67] Batch [99]	Speed: 4138.485051 samples/sec	accuracy=0.671598	lr=0.175046
Epoch[67] Batch [149]	Speed: 4160.935492 samples/sec	accuracy=0.670846	lr=0.174830
Epoch[67] Batch [199]	Speed: 4237.157385 samples/sec	accuracy=0.670768	lr=0.174613
Epoch[67] Batch [249]	Speed: 4154.194281 samples/sec	accuracy=0.670762	lr=0.174397
Epoch[67] Batch [299]	Speed: 4215.937689 samples/sec	accuracy=0.671027	lr=0.174180
Epoch[67] Batch [349]	Speed: 4186.555846 samples/sec	accuracy=0.671025	lr=0.173964
Epoch[67] Batch [399]	Speed: 4121.348010 samples/sec	accuracy=0.671205	lr=0.173747
Epoch[67] Batch [449]	Speed: 4049.800857 samples/sec	accuracy=0.671139	lr=0.173531
Epoch[67] Batch [499]	Speed: 4101.297233 samples/sec	accuracy=0.671212	lr=0.173314
Epoch[67] Batch [549]	Speed: 4158.103406 samples/sec	accuracy=0.671175	lr=0.173098
Epoch[67] Batch [599]	Speed: 4087.870386 samples/sec	accuracy=0.671285	lr=0.172881
Epoch[67] Batch [649]	Speed: 4075.561559 samples/sec	accuracy=0.671409	lr=0.172665
Epoch[67] Batch [699]	Speed: 4188.022620 samples/sec	accuracy=0.671402	lr=0.172449
Epoch[67] Batch [749]	Speed: 4056.523544 samples/sec	accuracy=0.671529	lr=0.172232
Epoch[67] Batch [799]	Speed: 4083.869935 samples/sec	accuracy=0.671683	lr=0.172016
Epoch[67] Batch [849]	Speed: 4021.253292 samples/sec	accuracy=0.671885	lr=0.171800
Epoch[67] Batch [899]	Speed: 4075.001641 samples/sec	accuracy=0.672055	lr=0.171584
Epoch[67] Batch [949]	Speed: 4193.866757 samples/sec	accuracy=0.672180	lr=0.171368
Epoch[67] Batch [999]	Speed: 4190.534402 samples/sec	accuracy=0.672238	lr=0.171152
Epoch[67] Batch [1049]	Speed: 4142.395816 samples/sec	accuracy=0.672349	lr=0.170935
Epoch[67] Batch [1099]	Speed: 4125.053219 samples/sec	accuracy=0.672505	lr=0.170719
Epoch[67] Batch [1149]	Speed: 4111.299591 samples/sec	accuracy=0.672642	lr=0.170503
Epoch[67] Batch [1199]	Speed: 4231.401157 samples/sec	accuracy=0.672802	lr=0.170287
Epoch[67] Batch [1249]	Speed: 4338.365002 samples/sec	accuracy=0.672903	lr=0.170072
[Epoch 67] training: accuracy=0.672904
[Epoch 67] speed: 4133 samples/sec	time cost: 321.113222
[Epoch 67] validation: err-top1=0.327786 err-top5=0.115693
Epoch[68] Batch [49]	Speed: 4045.499849 samples/sec	accuracy=0.673721	lr=0.169851
Epoch[68] Batch [99]	Speed: 4008.024786 samples/sec	accuracy=0.673196	lr=0.169635
Epoch[68] Batch [149]	Speed: 4146.010909 samples/sec	accuracy=0.673187	lr=0.169420
Epoch[68] Batch [199]	Speed: 4139.523949 samples/sec	accuracy=0.673010	lr=0.169204
Epoch[68] Batch [249]	Speed: 4091.333292 samples/sec	accuracy=0.673153	lr=0.168988
Epoch[68] Batch [299]	Speed: 4198.962224 samples/sec	accuracy=0.673140	lr=0.168772
Epoch[68] Batch [349]	Speed: 4090.460549 samples/sec	accuracy=0.673183	lr=0.168557
Epoch[68] Batch [399]	Speed: 3982.067354 samples/sec	accuracy=0.673328	lr=0.168341
Epoch[68] Batch [449]	Speed: 4060.667051 samples/sec	accuracy=0.673617	lr=0.168126
Epoch[68] Batch [499]	Speed: 4241.776689 samples/sec	accuracy=0.673809	lr=0.167910
Epoch[68] Batch [549]	Speed: 4234.190738 samples/sec	accuracy=0.673885	lr=0.167694
Epoch[68] Batch [599]	Speed: 4167.080431 samples/sec	accuracy=0.673935	lr=0.167479
Epoch[68] Batch [649]	Speed: 4106.904029 samples/sec	accuracy=0.674002	lr=0.167263
Epoch[68] Batch [699]	Speed: 4085.366272 samples/sec	accuracy=0.674084	lr=0.167048
Epoch[68] Batch [749]	Speed: 4190.099906 samples/sec	accuracy=0.674157	lr=0.166833
Epoch[68] Batch [799]	Speed: 4076.533427 samples/sec	accuracy=0.674272	lr=0.166617
Epoch[68] Batch [849]	Speed: 4058.564440 samples/sec	accuracy=0.674462	lr=0.166402
Epoch[68] Batch [899]	Speed: 4027.667819 samples/sec	accuracy=0.674608	lr=0.166187
Epoch[68] Batch [949]	Speed: 4116.982722 samples/sec	accuracy=0.674707	lr=0.165972
Epoch[68] Batch [999]	Speed: 4050.240047 samples/sec	accuracy=0.674764	lr=0.165756
Epoch[68] Batch [1049]	Speed: 4166.325254 samples/sec	accuracy=0.674823	lr=0.165541
Epoch[68] Batch [1099]	Speed: 4073.218974 samples/sec	accuracy=0.674911	lr=0.165326
Epoch[68] Batch [1149]	Speed: 4098.169039 samples/sec	accuracy=0.674935	lr=0.165111
Epoch[68] Batch [1199]	Speed: 4078.751767 samples/sec	accuracy=0.675043	lr=0.164896
Epoch[68] Batch [1249]	Speed: 4282.692500 samples/sec	accuracy=0.675083	lr=0.164681
[Epoch 68] training: accuracy=0.675085
[Epoch 68] speed: 4107 samples/sec	time cost: 322.473331
[Epoch 68] validation: err-top1=0.319396 err-top5=0.111567
Epoch[69] Batch [49]	Speed: 4132.494235 samples/sec	accuracy=0.675571	lr=0.164462
Epoch[69] Batch [99]	Speed: 4017.149436 samples/sec	accuracy=0.674846	lr=0.164247
Epoch[69] Batch [149]	Speed: 4096.190712 samples/sec	accuracy=0.674687	lr=0.164032
Epoch[69] Batch [199]	Speed: 4116.251984 samples/sec	accuracy=0.674184	lr=0.163818
Epoch[69] Batch [249]	Speed: 4189.706697 samples/sec	accuracy=0.674299	lr=0.163603
Epoch[69] Batch [299]	Speed: 4064.360135 samples/sec	accuracy=0.674114	lr=0.163388
Epoch[69] Batch [349]	Speed: 4060.141463 samples/sec	accuracy=0.674160	lr=0.163173
Epoch[69] Batch [399]	Speed: 4155.476831 samples/sec	accuracy=0.674164	lr=0.162959
Epoch[69] Batch [449]	Speed: 4062.893211 samples/sec	accuracy=0.674255	lr=0.162744
Epoch[69] Batch [499]	Speed: 4126.479506 samples/sec	accuracy=0.674251	lr=0.162530
Epoch[69] Batch [549]	Speed: 4113.560669 samples/sec	accuracy=0.674339	lr=0.162315
Epoch[69] Batch [599]	Speed: 4219.537659 samples/sec	accuracy=0.674421	lr=0.162101
Epoch[69] Batch [649]	Speed: 4139.274927 samples/sec	accuracy=0.674510	lr=0.161886
Epoch[69] Batch [699]	Speed: 4150.815292 samples/sec	accuracy=0.674666	lr=0.161672
Epoch[69] Batch [749]	Speed: 4171.709873 samples/sec	accuracy=0.674729	lr=0.161458
Epoch[69] Batch [799]	Speed: 4153.252426 samples/sec	accuracy=0.674825	lr=0.161244
Epoch[69] Batch [849]	Speed: 3990.406837 samples/sec	accuracy=0.674963	lr=0.161029
Epoch[69] Batch [899]	Speed: 4019.983689 samples/sec	accuracy=0.675100	lr=0.160815
Epoch[69] Batch [949]	Speed: 4094.935355 samples/sec	accuracy=0.675236	lr=0.160601
Epoch[69] Batch [999]	Speed: 4135.760214 samples/sec	accuracy=0.675415	lr=0.160387
Epoch[69] Batch [1049]	Speed: 4170.398251 samples/sec	accuracy=0.675553	lr=0.160173
Epoch[69] Batch [1099]	Speed: 4058.532609 samples/sec	accuracy=0.675600	lr=0.159959
Epoch[69] Batch [1149]	Speed: 4047.447811 samples/sec	accuracy=0.675678	lr=0.159745
Epoch[69] Batch [1199]	Speed: 4153.405047 samples/sec	accuracy=0.675773	lr=0.159531
Epoch[69] Batch [1249]	Speed: 4228.924187 samples/sec	accuracy=0.675880	lr=0.159317
[Epoch 69] training: accuracy=0.675882
[Epoch 69] speed: 4110 samples/sec	time cost: 321.364510
[Epoch 69] validation: err-top1=0.326355 err-top5=0.116252
Epoch[70] Batch [49]	Speed: 3944.097394 samples/sec	accuracy=0.676533	lr=0.159099
Epoch[70] Batch [99]	Speed: 4041.033214 samples/sec	accuracy=0.675535	lr=0.158886
Epoch[70] Batch [149]	Speed: 4132.296470 samples/sec	accuracy=0.675791	lr=0.158672
Epoch[70] Batch [199]	Speed: 4056.943960 samples/sec	accuracy=0.675716	lr=0.158458
Epoch[70] Batch [249]	Speed: 4092.898600 samples/sec	accuracy=0.675980	lr=0.158245
Epoch[70] Batch [299]	Speed: 4078.282905 samples/sec	accuracy=0.676086	lr=0.158031
Epoch[70] Batch [349]	Speed: 4110.776474 samples/sec	accuracy=0.676163	lr=0.157818
Epoch[70] Batch [399]	Speed: 4156.003023 samples/sec	accuracy=0.676242	lr=0.157604
Epoch[70] Batch [449]	Speed: 4005.202435 samples/sec	accuracy=0.676402	lr=0.157391
Epoch[70] Batch [499]	Speed: 4067.626511 samples/sec	accuracy=0.676497	lr=0.157177
Epoch[70] Batch [549]	Speed: 4213.589484 samples/sec	accuracy=0.676566	lr=0.156964
Epoch[70] Batch [599]	Speed: 4156.269507 samples/sec	accuracy=0.676590	lr=0.156751
Epoch[70] Batch [649]	Speed: 4166.297772 samples/sec	accuracy=0.676556	lr=0.156538
Epoch[70] Batch [699]	Speed: 4114.831810 samples/sec	accuracy=0.676590	lr=0.156325
Epoch[70] Batch [749]	Speed: 4053.724248 samples/sec	accuracy=0.676663	lr=0.156112
Epoch[70] Batch [799]	Speed: 4082.850631 samples/sec	accuracy=0.676800	lr=0.155899
Epoch[70] Batch [849]	Speed: 4165.453758 samples/sec	accuracy=0.676862	lr=0.155686
Epoch[70] Batch [899]	Speed: 4104.315990 samples/sec	accuracy=0.676998	lr=0.155473
Epoch[70] Batch [949]	Speed: 4111.846697 samples/sec	accuracy=0.677145	lr=0.155260
Epoch[70] Batch [999]	Speed: 4204.066055 samples/sec	accuracy=0.677276	lr=0.155047
Epoch[70] Batch [1049]	Speed: 4113.908742 samples/sec	accuracy=0.677409	lr=0.154834
Epoch[70] Batch [1099]	Speed: 4138.831931 samples/sec	accuracy=0.677514	lr=0.154622
Epoch[70] Batch [1149]	Speed: 4103.162816 samples/sec	accuracy=0.677632	lr=0.154409
Epoch[70] Batch [1199]	Speed: 4176.475020 samples/sec	accuracy=0.677751	lr=0.154196
Epoch[70] Batch [1249]	Speed: 4261.766769 samples/sec	accuracy=0.677870	lr=0.153984
[Epoch 70] training: accuracy=0.677872
[Epoch 70] speed: 4109 samples/sec	time cost: 322.500081
[Epoch 70] validation: err-top1=0.317124 err-top5=0.110930
Epoch[71] Batch [49]	Speed: 4227.207111 samples/sec	accuracy=0.678628	lr=0.153767
Epoch[71] Batch [99]	Speed: 4089.591137 samples/sec	accuracy=0.678533	lr=0.153555
Epoch[71] Batch [149]	Speed: 4015.009816 samples/sec	accuracy=0.678440	lr=0.153342
Epoch[71] Batch [199]	Speed: 4133.472445 samples/sec	accuracy=0.678366	lr=0.153130
Epoch[71] Batch [249]	Speed: 4058.530921 samples/sec	accuracy=0.678232	lr=0.152918
Epoch[71] Batch [299]	Speed: 4151.955353 samples/sec	accuracy=0.678075	lr=0.152705
Epoch[71] Batch [349]	Speed: 4050.371058 samples/sec	accuracy=0.678031	lr=0.152493
Epoch[71] Batch [399]	Speed: 4095.623550 samples/sec	accuracy=0.678113	lr=0.152281
Epoch[71] Batch [449]	Speed: 4086.868995 samples/sec	accuracy=0.678072	lr=0.152069
Epoch[71] Batch [499]	Speed: 4213.906815 samples/sec	accuracy=0.678017	lr=0.151857
Epoch[71] Batch [549]	Speed: 4116.327097 samples/sec	accuracy=0.678197	lr=0.151645
Epoch[71] Batch [599]	Speed: 3887.817055 samples/sec	accuracy=0.678219	lr=0.151433
Epoch[71] Batch [649]	Speed: 4158.120313 samples/sec	accuracy=0.678281	lr=0.151222
Epoch[71] Batch [699]	Speed: 4096.042891 samples/sec	accuracy=0.678324	lr=0.151010
Epoch[71] Batch [749]	Speed: 4109.039426 samples/sec	accuracy=0.678440	lr=0.150798
Epoch[71] Batch [799]	Speed: 4186.827732 samples/sec	accuracy=0.678539	lr=0.150587
Epoch[71] Batch [849]	Speed: 4125.302593 samples/sec	accuracy=0.678572	lr=0.150375
Epoch[71] Batch [899]	Speed: 4057.588930 samples/sec	accuracy=0.678691	lr=0.150163
Epoch[71] Batch [949]	Speed: 4051.195060 samples/sec	accuracy=0.678733	lr=0.149952
Epoch[71] Batch [999]	Speed: 4072.773165 samples/sec	accuracy=0.678798	lr=0.149741
Epoch[71] Batch [1049]	Speed: 4185.213993 samples/sec	accuracy=0.678918	lr=0.149529
Epoch[71] Batch [1099]	Speed: 4144.415997 samples/sec	accuracy=0.678973	lr=0.149318
Epoch[71] Batch [1149]	Speed: 4077.033932 samples/sec	accuracy=0.679012	lr=0.149107
Epoch[71] Batch [1199]	Speed: 4052.079645 samples/sec	accuracy=0.679113	lr=0.148896
Epoch[71] Batch [1249]	Speed: 4215.655554 samples/sec	accuracy=0.679246	lr=0.148685
[Epoch 71] training: accuracy=0.679252
[Epoch 71] speed: 4102 samples/sec	time cost: 323.026689
[Epoch 71] validation: err-top1=0.311982 err-top5=0.107063
Epoch[72] Batch [49]	Speed: 3999.791559 samples/sec	accuracy=0.679094	lr=0.148469
Epoch[72] Batch [99]	Speed: 4064.786332 samples/sec	accuracy=0.679353	lr=0.148258
Epoch[72] Batch [149]	Speed: 4125.831079 samples/sec	accuracy=0.679088	lr=0.148047
Epoch[72] Batch [199]	Speed: 4118.007614 samples/sec	accuracy=0.679270	lr=0.147837
Epoch[72] Batch [249]	Speed: 4120.405335 samples/sec	accuracy=0.679273	lr=0.147626
Epoch[72] Batch [299]	Speed: 4032.810043 samples/sec	accuracy=0.679468	lr=0.147415
Epoch[72] Batch [349]	Speed: 4053.998669 samples/sec	accuracy=0.679430	lr=0.147204
Epoch[72] Batch [399]	Speed: 4043.593124 samples/sec	accuracy=0.679587	lr=0.146994
Epoch[72] Batch [449]	Speed: 4157.460616 samples/sec	accuracy=0.679617	lr=0.146783
Epoch[72] Batch [499]	Speed: 4195.029855 samples/sec	accuracy=0.679690	lr=0.146573
Epoch[72] Batch [549]	Speed: 4058.512666 samples/sec	accuracy=0.679819	lr=0.146362
Epoch[72] Batch [599]	Speed: 4159.433729 samples/sec	accuracy=0.679851	lr=0.146152
Epoch[72] Batch [649]	Speed: 4076.171224 samples/sec	accuracy=0.679913	lr=0.145942
Epoch[72] Batch [699]	Speed: 4074.783748 samples/sec	accuracy=0.680058	lr=0.145732
Epoch[72] Batch [749]	Speed: 4007.771586 samples/sec	accuracy=0.680191	lr=0.145521
Epoch[72] Batch [799]	Speed: 3995.332984 samples/sec	accuracy=0.680341	lr=0.145311
Epoch[72] Batch [849]	Speed: 4066.635628 samples/sec	accuracy=0.680467	lr=0.145101
Epoch[72] Batch [899]	Speed: 4087.714684 samples/sec	accuracy=0.680553	lr=0.144891
Epoch[72] Batch [949]	Speed: 4086.162123 samples/sec	accuracy=0.680661	lr=0.144682
Epoch[72] Batch [999]	Speed: 3973.070403 samples/sec	accuracy=0.680778	lr=0.144472
Epoch[72] Batch [1049]	Speed: 4050.645102 samples/sec	accuracy=0.680885	lr=0.144262
Epoch[72] Batch [1099]	Speed: 4147.374673 samples/sec	accuracy=0.681022	lr=0.144052
Epoch[72] Batch [1149]	Speed: 4160.103883 samples/sec	accuracy=0.681092	lr=0.143843
Epoch[72] Batch [1199]	Speed: 4111.638701 samples/sec	accuracy=0.681189	lr=0.143633
Epoch[72] Batch [1249]	Speed: 4295.110323 samples/sec	accuracy=0.681277	lr=0.143424
[Epoch 72] training: accuracy=0.681277
[Epoch 72] speed: 4086 samples/sec	time cost: 323.585524
[Epoch 72] validation: err-top1=0.309212 err-top5=0.106306
Epoch[73] Batch [49]	Speed: 3770.606341 samples/sec	accuracy=0.680851	lr=0.143210
Epoch[73] Batch [99]	Speed: 3982.514280 samples/sec	accuracy=0.680323	lr=0.143001
Epoch[73] Batch [149]	Speed: 4066.199959 samples/sec	accuracy=0.680250	lr=0.142791
Epoch[73] Batch [199]	Speed: 4163.835540 samples/sec	accuracy=0.680000	lr=0.142582
Epoch[73] Batch [249]	Speed: 4113.092357 samples/sec	accuracy=0.680176	lr=0.142373
Epoch[73] Batch [299]	Speed: 4113.349112 samples/sec	accuracy=0.680104	lr=0.142164
Epoch[73] Batch [349]	Speed: 4011.769206 samples/sec	accuracy=0.680143	lr=0.141955
Epoch[73] Batch [399]	Speed: 4150.488300 samples/sec	accuracy=0.680295	lr=0.141746
Epoch[73] Batch [449]	Speed: 4072.262585 samples/sec	accuracy=0.680657	lr=0.141537
Epoch[73] Batch [499]	Speed: 3962.017857 samples/sec	accuracy=0.680848	lr=0.141328
Epoch[73] Batch [549]	Speed: 4011.401785 samples/sec	accuracy=0.680858	lr=0.141120
Epoch[73] Batch [599]	Speed: 4203.284251 samples/sec	accuracy=0.680935	lr=0.140911
Epoch[73] Batch [649]	Speed: 4085.792066 samples/sec	accuracy=0.680981	lr=0.140702
Epoch[73] Batch [699]	Speed: 4254.070028 samples/sec	accuracy=0.681084	lr=0.140494
Epoch[73] Batch [749]	Speed: 4114.711733 samples/sec	accuracy=0.681172	lr=0.140285
Epoch[73] Batch [799]	Speed: 4101.241386 samples/sec	accuracy=0.681293	lr=0.140077
Epoch[73] Batch [849]	Speed: 4053.848368 samples/sec	accuracy=0.681384	lr=0.139869
Epoch[73] Batch [899]	Speed: 4131.896545 samples/sec	accuracy=0.681528	lr=0.139661
Epoch[73] Batch [949]	Speed: 4113.849478 samples/sec	accuracy=0.681649	lr=0.139452
Epoch[73] Batch [999]	Speed: 4083.027078 samples/sec	accuracy=0.681732	lr=0.139244
Epoch[73] Batch [1049]	Speed: 4090.903693 samples/sec	accuracy=0.681830	lr=0.139036
Epoch[73] Batch [1099]	Speed: 4058.189625 samples/sec	accuracy=0.681910	lr=0.138828
Epoch[73] Batch [1149]	Speed: 4037.767939 samples/sec	accuracy=0.682034	lr=0.138620
Epoch[73] Batch [1199]	Speed: 4141.303485 samples/sec	accuracy=0.682094	lr=0.138413
Epoch[73] Batch [1249]	Speed: 4235.457007 samples/sec	accuracy=0.682239	lr=0.138205
[Epoch 73] training: accuracy=0.682241
[Epoch 73] speed: 4079 samples/sec	time cost: 323.817028
[Epoch 73] validation: err-top1=0.306860 err-top5=0.105329
Epoch[74] Batch [49]	Speed: 4020.204039 samples/sec	accuracy=0.683418	lr=0.137993
Epoch[74] Batch [99]	Speed: 4010.994501 samples/sec	accuracy=0.682805	lr=0.137786
Epoch[74] Batch [149]	Speed: 4118.542920 samples/sec	accuracy=0.682557	lr=0.137578
Epoch[74] Batch [199]	Speed: 4043.641701 samples/sec	accuracy=0.682302	lr=0.137371
Epoch[74] Batch [249]	Speed: 4033.082019 samples/sec	accuracy=0.682598	lr=0.137163
Epoch[74] Batch [299]	Speed: 4025.999150 samples/sec	accuracy=0.682329	lr=0.136956
Epoch[74] Batch [349]	Speed: 4090.144866 samples/sec	accuracy=0.682278	lr=0.136749
Epoch[74] Batch [399]	Speed: 4057.272552 samples/sec	accuracy=0.682377	lr=0.136542
Epoch[74] Batch [449]	Speed: 4083.861160 samples/sec	accuracy=0.682497	lr=0.136335
Epoch[74] Batch [499]	Speed: 4086.796508 samples/sec	accuracy=0.682480	lr=0.136128
Epoch[74] Batch [549]	Speed: 4061.510994 samples/sec	accuracy=0.682553	lr=0.135921
Epoch[74] Batch [599]	Speed: 4112.854539 samples/sec	accuracy=0.682702	lr=0.135714
Epoch[74] Batch [649]	Speed: 4150.616812 samples/sec	accuracy=0.682818	lr=0.135507
Epoch[74] Batch [699]	Speed: 4108.796101 samples/sec	accuracy=0.682885	lr=0.135301
Epoch[74] Batch [749]	Speed: 4176.046114 samples/sec	accuracy=0.682988	lr=0.135094
Epoch[74] Batch [799]	Speed: 4160.650191 samples/sec	accuracy=0.683141	lr=0.134887
Epoch[74] Batch [849]	Speed: 4162.196959 samples/sec	accuracy=0.683215	lr=0.134681
Epoch[74] Batch [899]	Speed: 4062.127988 samples/sec	accuracy=0.683326	lr=0.134475
Epoch[74] Batch [949]	Speed: 4122.868222 samples/sec	accuracy=0.683346	lr=0.134268
Epoch[74] Batch [999]	Speed: 4032.717272 samples/sec	accuracy=0.683426	lr=0.134062
Epoch[74] Batch [1049]	Speed: 4046.362125 samples/sec	accuracy=0.683504	lr=0.133856
Epoch[74] Batch [1099]	Speed: 4096.784291 samples/sec	accuracy=0.683580	lr=0.133650
Epoch[74] Batch [1149]	Speed: 4003.959220 samples/sec	accuracy=0.683659	lr=0.133444
Epoch[74] Batch [1199]	Speed: 4138.108088 samples/sec	accuracy=0.683749	lr=0.133238
Epoch[74] Batch [1249]	Speed: 4262.565744 samples/sec	accuracy=0.683841	lr=0.133032
[Epoch 74] training: accuracy=0.683843
[Epoch 74] speed: 4086 samples/sec	time cost: 323.526540
[Epoch 74] validation: err-top1=0.304090 err-top5=0.103217
Epoch[75] Batch [49]	Speed: 4047.234380 samples/sec	accuracy=0.684633	lr=0.132822
Epoch[75] Batch [99]	Speed: 4052.919410 samples/sec	accuracy=0.684273	lr=0.132617
Epoch[75] Batch [149]	Speed: 4128.009934 samples/sec	accuracy=0.684000	lr=0.132411
Epoch[75] Batch [199]	Speed: 4028.020244 samples/sec	accuracy=0.684168	lr=0.132206
Epoch[75] Batch [249]	Speed: 4104.833306 samples/sec	accuracy=0.684263	lr=0.132000
Epoch[75] Batch [299]	Speed: 4187.031731 samples/sec	accuracy=0.684159	lr=0.131795
Epoch[75] Batch [349]	Speed: 4070.050679 samples/sec	accuracy=0.684348	lr=0.131590
Epoch[75] Batch [399]	Speed: 4135.102895 samples/sec	accuracy=0.684357	lr=0.131385
Epoch[75] Batch [449]	Speed: 4137.383781 samples/sec	accuracy=0.684443	lr=0.131180
Epoch[75] Batch [499]	Speed: 4062.140897 samples/sec	accuracy=0.684285	lr=0.130975
Epoch[75] Batch [549]	Speed: 4099.198668 samples/sec	accuracy=0.684284	lr=0.130770
Epoch[75] Batch [599]	Speed: 4056.702245 samples/sec	accuracy=0.684406	lr=0.130565
Epoch[75] Batch [649]	Speed: 4096.831184 samples/sec	accuracy=0.684444	lr=0.130360
Epoch[75] Batch [699]	Speed: 4072.946116 samples/sec	accuracy=0.684576	lr=0.130155
Epoch[75] Batch [749]	Speed: 4048.631319 samples/sec	accuracy=0.684726	lr=0.129951
Epoch[75] Batch [799]	Speed: 3945.565966 samples/sec	accuracy=0.684831	lr=0.129746
Epoch[75] Batch [849]	Speed: 4043.672005 samples/sec	accuracy=0.684892	lr=0.129542
Epoch[75] Batch [899]	Speed: 4080.901952 samples/sec	accuracy=0.685025	lr=0.129338
Epoch[75] Batch [949]	Speed: 4047.408067 samples/sec	accuracy=0.685161	lr=0.129133
Epoch[75] Batch [999]	Speed: 4118.696635 samples/sec	accuracy=0.685309	lr=0.128929
Epoch[75] Batch [1049]	Speed: 4106.811430 samples/sec	accuracy=0.685451	lr=0.128725
Epoch[75] Batch [1099]	Speed: 4155.271072 samples/sec	accuracy=0.685550	lr=0.128521
Epoch[75] Batch [1149]	Speed: 4136.408022 samples/sec	accuracy=0.685625	lr=0.128317
Epoch[75] Batch [1199]	Speed: 4024.970202 samples/sec	accuracy=0.685732	lr=0.128113
Epoch[75] Batch [1249]	Speed: 4331.123952 samples/sec	accuracy=0.685876	lr=0.127910
[Epoch 75] training: accuracy=0.685881
[Epoch 75] speed: 4087 samples/sec	time cost: 323.285048
[Epoch 75] validation: err-top1=0.301046 err-top5=0.100932
Epoch[76] Batch [49]	Speed: 3955.107080 samples/sec	accuracy=0.687775	lr=0.127702
Epoch[76] Batch [99]	Speed: 4221.410319 samples/sec	accuracy=0.686847	lr=0.127498
Epoch[76] Batch [149]	Speed: 4121.496239 samples/sec	accuracy=0.686030	lr=0.127295
Epoch[76] Batch [199]	Speed: 4184.888328 samples/sec	accuracy=0.685924	lr=0.127092
Epoch[76] Batch [249]	Speed: 4145.836579 samples/sec	accuracy=0.686134	lr=0.126888
Epoch[76] Batch [299]	Speed: 4237.470500 samples/sec	accuracy=0.686090	lr=0.126685
Epoch[76] Batch [349]	Speed: 4082.841160 samples/sec	accuracy=0.686310	lr=0.126482
Epoch[76] Batch [399]	Speed: 4074.972490 samples/sec	accuracy=0.686433	lr=0.126279
Epoch[76] Batch [449]	Speed: 4221.990610 samples/sec	accuracy=0.686334	lr=0.126076
Epoch[76] Batch [499]	Speed: 4091.319573 samples/sec	accuracy=0.686417	lr=0.125873
Epoch[76] Batch [549]	Speed: 4103.684076 samples/sec	accuracy=0.686433	lr=0.125670
Epoch[76] Batch [599]	Speed: 3986.546085 samples/sec	accuracy=0.686471	lr=0.125468
Epoch[76] Batch [649]	Speed: 4165.222288 samples/sec	accuracy=0.686480	lr=0.125265
Epoch[76] Batch [699]	Speed: 4143.518544 samples/sec	accuracy=0.686550	lr=0.125062
Epoch[76] Batch [749]	Speed: 4085.839330 samples/sec	accuracy=0.686641	lr=0.124860
Epoch[76] Batch [799]	Speed: 3998.118958 samples/sec	accuracy=0.686637	lr=0.124658
Epoch[76] Batch [849]	Speed: 4084.932797 samples/sec	accuracy=0.686744	lr=0.124455
Epoch[76] Batch [899]	Speed: 4157.134508 samples/sec	accuracy=0.686852	lr=0.124253
Epoch[76] Batch [949]	Speed: 4244.908365 samples/sec	accuracy=0.686877	lr=0.124051
Epoch[76] Batch [999]	Speed: 4129.991482 samples/sec	accuracy=0.686987	lr=0.123849
Epoch[76] Batch [1049]	Speed: 4198.391364 samples/sec	accuracy=0.687050	lr=0.123647
Epoch[76] Batch [1099]	Speed: 4098.215338 samples/sec	accuracy=0.687165	lr=0.123446
Epoch[76] Batch [1149]	Speed: 4198.400639 samples/sec	accuracy=0.687228	lr=0.123244
Epoch[76] Batch [1199]	Speed: 4299.460060 samples/sec	accuracy=0.687379	lr=0.123042
Epoch[76] Batch [1249]	Speed: 4258.208059 samples/sec	accuracy=0.687429	lr=0.122841
[Epoch 76] training: accuracy=0.687431
[Epoch 76] speed: 4134 samples/sec	time cost: 320.214766
[Epoch 76] validation: err-top1=0.295520 err-top5=0.098872
Epoch[77] Batch [49]	Speed: 4103.319698 samples/sec	accuracy=0.689843	lr=0.122635
Epoch[77] Batch [99]	Speed: 4108.157148 samples/sec	accuracy=0.689253	lr=0.122434
Epoch[77] Batch [149]	Speed: 4041.031693 samples/sec	accuracy=0.688588	lr=0.122233
Epoch[77] Batch [199]	Speed: 4144.415757 samples/sec	accuracy=0.688293	lr=0.122032
Epoch[77] Batch [249]	Speed: 4136.507696 samples/sec	accuracy=0.688110	lr=0.121831
Epoch[77] Batch [299]	Speed: 4089.426815 samples/sec	accuracy=0.687525	lr=0.121630
Epoch[77] Batch [349]	Speed: 4115.160857 samples/sec	accuracy=0.687459	lr=0.121429
Epoch[77] Batch [399]	Speed: 4079.027961 samples/sec	accuracy=0.687597	lr=0.121228
Epoch[77] Batch [449]	Speed: 4213.382145 samples/sec	accuracy=0.687584	lr=0.121027
Epoch[77] Batch [499]	Speed: 4196.915678 samples/sec	accuracy=0.687653	lr=0.120827
Epoch[77] Batch [549]	Speed: 4099.066435 samples/sec	accuracy=0.687678	lr=0.120626
Epoch[77] Batch [599]	Speed: 4144.795870 samples/sec	accuracy=0.687670	lr=0.120426
Epoch[77] Batch [649]	Speed: 4111.202230 samples/sec	accuracy=0.687639	lr=0.120226
Epoch[77] Batch [699]	Speed: 4214.566185 samples/sec	accuracy=0.687760	lr=0.120025
Epoch[77] Batch [749]	Speed: 4230.807356 samples/sec	accuracy=0.687862	lr=0.119825
Epoch[77] Batch [799]	Speed: 4172.407418 samples/sec	accuracy=0.687947	lr=0.119625
Epoch[77] Batch [849]	Speed: 4177.700415 samples/sec	accuracy=0.688043	lr=0.119425
Epoch[77] Batch [899]	Speed: 4122.086256 samples/sec	accuracy=0.688165	lr=0.119226
Epoch[77] Batch [949]	Speed: 4055.724642 samples/sec	accuracy=0.688283	lr=0.119026
Epoch[77] Batch [999]	Speed: 4092.757881 samples/sec	accuracy=0.688445	lr=0.118826
Epoch[77] Batch [1049]	Speed: 4114.050525 samples/sec	accuracy=0.688595	lr=0.118627
Epoch[77] Batch [1099]	Speed: 4193.055092 samples/sec	accuracy=0.688698	lr=0.118427
Epoch[77] Batch [1149]	Speed: 4051.401877 samples/sec	accuracy=0.688827	lr=0.118228
Epoch[77] Batch [1199]	Speed: 4078.734879 samples/sec	accuracy=0.688898	lr=0.118029
Epoch[77] Batch [1249]	Speed: 4280.128007 samples/sec	accuracy=0.689025	lr=0.117830
[Epoch 77] training: accuracy=0.689028
[Epoch 77] speed: 4130 samples/sec	time cost: 320.274194
[Epoch 77] validation: err-top1=0.295619 err-top5=0.097856
Epoch[78] Batch [49]	Speed: 4151.044843 samples/sec	accuracy=0.690712	lr=0.117627
Epoch[78] Batch [99]	Speed: 4055.359770 samples/sec	accuracy=0.690633	lr=0.117428
Epoch[78] Batch [149]	Speed: 4278.070969 samples/sec	accuracy=0.689893	lr=0.117229
Epoch[78] Batch [199]	Speed: 4160.209700 samples/sec	accuracy=0.689841	lr=0.117030
Epoch[78] Batch [249]	Speed: 4129.944541 samples/sec	accuracy=0.689719	lr=0.116831
Epoch[78] Batch [299]	Speed: 4097.551447 samples/sec	accuracy=0.689455	lr=0.116633
Epoch[78] Batch [349]	Speed: 4202.847519 samples/sec	accuracy=0.689104	lr=0.116434
Epoch[78] Batch [399]	Speed: 4199.584978 samples/sec	accuracy=0.688967	lr=0.116236
Epoch[78] Batch [449]	Speed: 4097.305885 samples/sec	accuracy=0.689125	lr=0.116038
Epoch[78] Batch [499]	Speed: 4077.428804 samples/sec	accuracy=0.689162	lr=0.115840
Epoch[78] Batch [549]	Speed: 4120.973055 samples/sec	accuracy=0.689186	lr=0.115642
Epoch[78] Batch [599]	Speed: 4205.185081 samples/sec	accuracy=0.689345	lr=0.115444
Epoch[78] Batch [649]	Speed: 4060.763262 samples/sec	accuracy=0.689503	lr=0.115246
Epoch[78] Batch [699]	Speed: 4099.940193 samples/sec	accuracy=0.689581	lr=0.115048
Epoch[78] Batch [749]	Speed: 4173.946960 samples/sec	accuracy=0.689656	lr=0.114850
Epoch[78] Batch [799]	Speed: 4025.640815 samples/sec	accuracy=0.689689	lr=0.114653
Epoch[78] Batch [849]	Speed: 4277.057460 samples/sec	accuracy=0.689764	lr=0.114455
Epoch[78] Batch [899]	Speed: 4272.904410 samples/sec	accuracy=0.689892	lr=0.114258
Epoch[78] Batch [949]	Speed: 4139.324155 samples/sec	accuracy=0.689958	lr=0.114061
Epoch[78] Batch [999]	Speed: 4292.368953 samples/sec	accuracy=0.690043	lr=0.113864
Epoch[78] Batch [1049]	Speed: 4186.822263 samples/sec	accuracy=0.690147	lr=0.113667
Epoch[78] Batch [1099]	Speed: 4201.589244 samples/sec	accuracy=0.690226	lr=0.113470
Epoch[78] Batch [1149]	Speed: 4195.277763 samples/sec	accuracy=0.690350	lr=0.113273
Epoch[78] Batch [1199]	Speed: 4110.882079 samples/sec	accuracy=0.690448	lr=0.113076
Epoch[78] Batch [1249]	Speed: 4204.963006 samples/sec	accuracy=0.690556	lr=0.112880
[Epoch 78] training: accuracy=0.690568
[Epoch 78] speed: 4155 samples/sec	time cost: 318.412352
[Epoch 78] validation: err-top1=0.292730 err-top5=0.096261
Epoch[79] Batch [49]	Speed: 4088.554652 samples/sec	accuracy=0.691955	lr=0.112679
Epoch[79] Batch [99]	Speed: 4089.810617 samples/sec	accuracy=0.691777	lr=0.112483
Epoch[79] Batch [149]	Speed: 4183.061939 samples/sec	accuracy=0.691195	lr=0.112287
Epoch[79] Batch [199]	Speed: 4171.710602 samples/sec	accuracy=0.691209	lr=0.112090
Epoch[79] Batch [249]	Speed: 4056.122521 samples/sec	accuracy=0.691069	lr=0.111894
Epoch[79] Batch [299]	Speed: 4097.073094 samples/sec	accuracy=0.691045	lr=0.111698
Epoch[79] Batch [349]	Speed: 4182.398948 samples/sec	accuracy=0.690959	lr=0.111502
Epoch[79] Batch [399]	Speed: 4199.586374 samples/sec	accuracy=0.690967	lr=0.111307
Epoch[79] Batch [449]	Speed: 4177.509921 samples/sec	accuracy=0.691164	lr=0.111111
Epoch[79] Batch [499]	Speed: 4256.030828 samples/sec	accuracy=0.691153	lr=0.110915
Epoch[79] Batch [549]	Speed: 4258.519057 samples/sec	accuracy=0.691205	lr=0.110720
Epoch[79] Batch [599]	Speed: 4122.003891 samples/sec	accuracy=0.691375	lr=0.110525
Epoch[79] Batch [649]	Speed: 4032.401959 samples/sec	accuracy=0.691399	lr=0.110329
Epoch[79] Batch [699]	Speed: 4229.142470 samples/sec	accuracy=0.691403	lr=0.110134
Epoch[79] Batch [749]	Speed: 4034.520367 samples/sec	accuracy=0.691410	lr=0.109939
Epoch[79] Batch [799]	Speed: 4249.608084 samples/sec	accuracy=0.691476	lr=0.109744
Epoch[79] Batch [849]	Speed: 4142.635465 samples/sec	accuracy=0.691611	lr=0.109549
Epoch[79] Batch [899]	Speed: 4009.989006 samples/sec	accuracy=0.691719	lr=0.109355
Epoch[79] Batch [949]	Speed: 4049.779320 samples/sec	accuracy=0.691854	lr=0.109160
Epoch[79] Batch [999]	Speed: 4057.978050 samples/sec	accuracy=0.691912	lr=0.108966
Epoch[79] Batch [1049]	Speed: 4283.884373 samples/sec	accuracy=0.691981	lr=0.108771
Epoch[79] Batch [1099]	Speed: 4166.862686 samples/sec	accuracy=0.692054	lr=0.108577
Epoch[79] Batch [1149]	Speed: 4077.897702 samples/sec	accuracy=0.692179	lr=0.108383
Epoch[79] Batch [1199]	Speed: 4226.997348 samples/sec	accuracy=0.692283	lr=0.108189
Epoch[79] Batch [1249]	Speed: 4234.418749 samples/sec	accuracy=0.692397	lr=0.107995
[Epoch 79] training: accuracy=0.692399
[Epoch 79] speed: 4142 samples/sec	time cost: 319.535478
[Epoch 79] validation: err-top1=0.292391 err-top5=0.096341
Epoch[80] Batch [49]	Speed: 4051.931550 samples/sec	accuracy=0.693628	lr=0.107797
Epoch[80] Batch [99]	Speed: 4194.472762 samples/sec	accuracy=0.693481	lr=0.107603
Epoch[80] Batch [149]	Speed: 4056.833061 samples/sec	accuracy=0.693007	lr=0.107410
Epoch[80] Batch [199]	Speed: 4137.742196 samples/sec	accuracy=0.692867	lr=0.107216
Epoch[80] Batch [249]	Speed: 4107.542906 samples/sec	accuracy=0.692950	lr=0.107023
Epoch[80] Batch [299]	Speed: 4041.612815 samples/sec	accuracy=0.692802	lr=0.106830
Epoch[80] Batch [349]	Speed: 4059.598436 samples/sec	accuracy=0.692708	lr=0.106636
Epoch[80] Batch [399]	Speed: 4177.123702 samples/sec	accuracy=0.692778	lr=0.106443
Epoch[80] Batch [449]	Speed: 4195.488406 samples/sec	accuracy=0.692919	lr=0.106250
Epoch[80] Batch [499]	Speed: 4101.406032 samples/sec	accuracy=0.693050	lr=0.106057
Epoch[80] Batch [549]	Speed: 4043.626930 samples/sec	accuracy=0.693078	lr=0.105865
Epoch[80] Batch [599]	Speed: 4256.860983 samples/sec	accuracy=0.693136	lr=0.105672
Epoch[80] Batch [649]	Speed: 4116.290014 samples/sec	accuracy=0.693130	lr=0.105480
Epoch[80] Batch [699]	Speed: 4095.559969 samples/sec	accuracy=0.693289	lr=0.105287
Epoch[80] Batch [749]	Speed: 4129.947242 samples/sec	accuracy=0.693445	lr=0.105095
Epoch[80] Batch [799]	Speed: 4107.631138 samples/sec	accuracy=0.693583	lr=0.104903
Epoch[80] Batch [849]	Speed: 4116.964569 samples/sec	accuracy=0.693702	lr=0.104711
Epoch[80] Batch [899]	Speed: 4137.733824 samples/sec	accuracy=0.693731	lr=0.104519
Epoch[80] Batch [949]	Speed: 4178.931084 samples/sec	accuracy=0.693801	lr=0.104327
Epoch[80] Batch [999]	Speed: 4109.274601 samples/sec	accuracy=0.693923	lr=0.104135
Epoch[80] Batch [1049]	Speed: 4087.090127 samples/sec	accuracy=0.693999	lr=0.103944
Epoch[80] Batch [1099]	Speed: 4182.821338 samples/sec	accuracy=0.694111	lr=0.103752
Epoch[80] Batch [1149]	Speed: 4174.344437 samples/sec	accuracy=0.694163	lr=0.103561
Epoch[80] Batch [1199]	Speed: 4208.220140 samples/sec	accuracy=0.694219	lr=0.103370
Epoch[80] Batch [1249]	Speed: 4261.795779 samples/sec	accuracy=0.694300	lr=0.103178
[Epoch 80] training: accuracy=0.694303
[Epoch 80] speed: 4128 samples/sec	time cost: 320.634314
[Epoch 80] validation: err-top1=0.290398 err-top5=0.096620
Epoch[81] Batch [49]	Speed: 4023.950300 samples/sec	accuracy=0.694203	lr=0.102984
Epoch[81] Batch [99]	Speed: 4050.764985 samples/sec	accuracy=0.694659	lr=0.102793
Epoch[81] Batch [149]	Speed: 4136.292179 samples/sec	accuracy=0.694525	lr=0.102602
Epoch[81] Batch [199]	Speed: 4217.995542 samples/sec	accuracy=0.694185	lr=0.102411
Epoch[81] Batch [249]	Speed: 4123.716682 samples/sec	accuracy=0.694339	lr=0.102221
Epoch[81] Batch [299]	Speed: 4018.663222 samples/sec	accuracy=0.694374	lr=0.102030
Epoch[81] Batch [349]	Speed: 4088.335774 samples/sec	accuracy=0.694388	lr=0.101840
Epoch[81] Batch [399]	Speed: 4073.203754 samples/sec	accuracy=0.694298	lr=0.101650
Epoch[81] Batch [449]	Speed: 4089.901827 samples/sec	accuracy=0.694354	lr=0.101460
Epoch[81] Batch [499]	Speed: 4109.401596 samples/sec	accuracy=0.694443	lr=0.101270
Epoch[81] Batch [549]	Speed: 4151.232289 samples/sec	accuracy=0.694537	lr=0.101080
Epoch[81] Batch [599]	Speed: 4022.318537 samples/sec	accuracy=0.694569	lr=0.100890
Epoch[81] Batch [649]	Speed: 4249.693358 samples/sec	accuracy=0.694777	lr=0.100701
Epoch[81] Batch [699]	Speed: 4063.645265 samples/sec	accuracy=0.694818	lr=0.100511
Epoch[81] Batch [749]	Speed: 4057.228476 samples/sec	accuracy=0.694854	lr=0.100322
Epoch[81] Batch [799]	Speed: 4118.573726 samples/sec	accuracy=0.694851	lr=0.100132
Epoch[81] Batch [849]	Speed: 4135.303875 samples/sec	accuracy=0.694955	lr=0.099943
Epoch[81] Batch [899]	Speed: 4111.382160 samples/sec	accuracy=0.695074	lr=0.099754
Epoch[81] Batch [949]	Speed: 4125.929848 samples/sec	accuracy=0.695146	lr=0.099565
Epoch[81] Batch [999]	Speed: 4171.086527 samples/sec	accuracy=0.695232	lr=0.099377
Epoch[81] Batch [1049]	Speed: 4119.635130 samples/sec	accuracy=0.695336	lr=0.099188
Epoch[81] Batch [1099]	Speed: 4093.196295 samples/sec	accuracy=0.695448	lr=0.098999
Epoch[81] Batch [1149]	Speed: 4187.417416 samples/sec	accuracy=0.695560	lr=0.098811
Epoch[81] Batch [1199]	Speed: 4167.830785 samples/sec	accuracy=0.695742	lr=0.098623
Epoch[81] Batch [1249]	Speed: 4248.105930 samples/sec	accuracy=0.695873	lr=0.098434
[Epoch 81] training: accuracy=0.695877
[Epoch 81] speed: 4113 samples/sec	time cost: 322.024631
[Epoch 81] validation: err-top1=0.289754 err-top5=0.094808
Epoch[82] Batch [49]	Speed: 4059.054481 samples/sec	accuracy=0.697253	lr=0.098243
Epoch[82] Batch [99]	Speed: 4103.155839 samples/sec	accuracy=0.696225	lr=0.098055
Epoch[82] Batch [149]	Speed: 4006.178924 samples/sec	accuracy=0.696225	lr=0.097867
Epoch[82] Batch [199]	Speed: 4016.477292 samples/sec	accuracy=0.696206	lr=0.097679
Epoch[82] Batch [249]	Speed: 4136.444832 samples/sec	accuracy=0.696160	lr=0.097492
Epoch[82] Batch [299]	Speed: 4078.267647 samples/sec	accuracy=0.696252	lr=0.097304
Epoch[82] Batch [349]	Speed: 4131.801862 samples/sec	accuracy=0.696392	lr=0.097117
Epoch[82] Batch [399]	Speed: 3998.321284 samples/sec	accuracy=0.696442	lr=0.096930
Epoch[82] Batch [449]	Speed: 4211.513379 samples/sec	accuracy=0.696505	lr=0.096743
Epoch[82] Batch [499]	Speed: 4151.375694 samples/sec	accuracy=0.696414	lr=0.096556
Epoch[82] Batch [549]	Speed: 4107.911964 samples/sec	accuracy=0.696426	lr=0.096369
Epoch[82] Batch [599]	Speed: 4187.150597 samples/sec	accuracy=0.696517	lr=0.096182
Epoch[82] Batch [649]	Speed: 4146.464810 samples/sec	accuracy=0.696419	lr=0.095996
Epoch[82] Batch [699]	Speed: 4106.958302 samples/sec	accuracy=0.696523	lr=0.095809
Epoch[82] Batch [749]	Speed: 4202.945815 samples/sec	accuracy=0.696596	lr=0.095623
Epoch[82] Batch [799]	Speed: 4062.826184 samples/sec	accuracy=0.696734	lr=0.095437
Epoch[82] Batch [849]	Speed: 4194.643176 samples/sec	accuracy=0.696821	lr=0.095250
Epoch[82] Batch [899]	Speed: 4255.726603 samples/sec	accuracy=0.696863	lr=0.095064
Epoch[82] Batch [949]	Speed: 4089.333446 samples/sec	accuracy=0.696938	lr=0.094879
Epoch[82] Batch [999]	Speed: 4094.082070 samples/sec	accuracy=0.697004	lr=0.094693
Epoch[82] Batch [1049]	Speed: 4121.444825 samples/sec	accuracy=0.697138	lr=0.094507
Epoch[82] Batch [1099]	Speed: 4144.449910 samples/sec	accuracy=0.697246	lr=0.094322
Epoch[82] Batch [1149]	Speed: 4161.892046 samples/sec	accuracy=0.697333	lr=0.094137
Epoch[82] Batch [1199]	Speed: 4116.404896 samples/sec	accuracy=0.697468	lr=0.093951
Epoch[82] Batch [1249]	Speed: 4307.612724 samples/sec	accuracy=0.697577	lr=0.093766
[Epoch 82] training: accuracy=0.697579
[Epoch 82] speed: 4123 samples/sec	time cost: 321.342017
[Epoch 82] validation: err-top1=0.290119 err-top5=0.095743
Epoch[83] Batch [49]	Speed: 3942.343586 samples/sec	accuracy=0.698793	lr=0.093578
Epoch[83] Batch [99]	Speed: 4059.451326 samples/sec	accuracy=0.698695	lr=0.093393
Epoch[83] Batch [149]	Speed: 4080.599451 samples/sec	accuracy=0.698444	lr=0.093208
Epoch[83] Batch [199]	Speed: 4095.479988 samples/sec	accuracy=0.698035	lr=0.093024
Epoch[83] Batch [249]	Speed: 4067.632983 samples/sec	accuracy=0.697967	lr=0.092839
Epoch[83] Batch [299]	Speed: 4211.362733 samples/sec	accuracy=0.697973	lr=0.092655
Epoch[83] Batch [349]	Speed: 4173.327406 samples/sec	accuracy=0.698055	lr=0.092471
Epoch[83] Batch [399]	Speed: 4069.418552 samples/sec	accuracy=0.698247	lr=0.092287
Epoch[83] Batch [449]	Speed: 4125.090857 samples/sec	accuracy=0.698413	lr=0.092103
Epoch[83] Batch [499]	Speed: 4281.105419 samples/sec	accuracy=0.698450	lr=0.091919
Epoch[83] Batch [549]	Speed: 4041.042111 samples/sec	accuracy=0.698465	lr=0.091735
Epoch[83] Batch [599]	Speed: 4159.830784 samples/sec	accuracy=0.698574	lr=0.091552
Epoch[83] Batch [649]	Speed: 4126.593928 samples/sec	accuracy=0.698700	lr=0.091368
Epoch[83] Batch [699]	Speed: 4189.042334 samples/sec	accuracy=0.698846	lr=0.091185
Epoch[83] Batch [749]	Speed: 4074.509366 samples/sec	accuracy=0.699008	lr=0.091002
Epoch[83] Batch [799]	Speed: 4117.966157 samples/sec	accuracy=0.699038	lr=0.090819
Epoch[83] Batch [849]	Speed: 4089.744023 samples/sec	accuracy=0.699034	lr=0.090636
Epoch[83] Batch [899]	Speed: 4162.941763 samples/sec	accuracy=0.699167	lr=0.090453
Epoch[83] Batch [949]	Speed: 4173.707487 samples/sec	accuracy=0.699260	lr=0.090270
Epoch[83] Batch [999]	Speed: 3982.157735 samples/sec	accuracy=0.699290	lr=0.090088
Epoch[83] Batch [1049]	Speed: 4192.271322 samples/sec	accuracy=0.699432	lr=0.089905
Epoch[83] Batch [1099]	Speed: 4041.485488 samples/sec	accuracy=0.699585	lr=0.089723
Epoch[83] Batch [1149]	Speed: 4125.290548 samples/sec	accuracy=0.699720	lr=0.089541
Epoch[83] Batch [1199]	Speed: 4262.363117 samples/sec	accuracy=0.699840	lr=0.089359
Epoch[83] Batch [1249]	Speed: 4289.774267 samples/sec	accuracy=0.699976	lr=0.089177
[Epoch 83] training: accuracy=0.699980
[Epoch 83] speed: 4120 samples/sec	time cost: 321.574406
[Epoch 83] validation: err-top1=0.283044 err-top5=0.092335
Epoch[84] Batch [49]	Speed: 4097.240688 samples/sec	accuracy=0.701775	lr=0.088992
Epoch[84] Batch [99]	Speed: 4141.736706 samples/sec	accuracy=0.700799	lr=0.088810
Epoch[84] Batch [149]	Speed: 4199.276288 samples/sec	accuracy=0.700458	lr=0.088629
Epoch[84] Batch [199]	Speed: 3960.981240 samples/sec	accuracy=0.700412	lr=0.088448
Epoch[84] Batch [249]	Speed: 4142.962979 samples/sec	accuracy=0.700146	lr=0.088266
Epoch[84] Batch [299]	Speed: 4185.380964 samples/sec	accuracy=0.700034	lr=0.088085
Epoch[84] Batch [349]	Speed: 4240.377021 samples/sec	accuracy=0.700066	lr=0.087904
Epoch[84] Batch [399]	Speed: 4110.386997 samples/sec	accuracy=0.699903	lr=0.087724
Epoch[84] Batch [449]	Speed: 4087.478080 samples/sec	accuracy=0.699919	lr=0.087543
Epoch[84] Batch [499]	Speed: 4142.205172 samples/sec	accuracy=0.699984	lr=0.087363
Epoch[84] Batch [549]	Speed: 4141.229853 samples/sec	accuracy=0.699925	lr=0.087182
Epoch[84] Batch [599]	Speed: 4109.106728 samples/sec	accuracy=0.700002	lr=0.087002
Epoch[84] Batch [649]	Speed: 4221.347668 samples/sec	accuracy=0.699923	lr=0.086822
Epoch[84] Batch [699]	Speed: 4218.033735 samples/sec	accuracy=0.700023	lr=0.086642
Epoch[84] Batch [749]	Speed: 4142.331494 samples/sec	accuracy=0.700087	lr=0.086462
Epoch[84] Batch [799]	Speed: 4061.448161 samples/sec	accuracy=0.700142	lr=0.086282
Epoch[84] Batch [849]	Speed: 4257.516562 samples/sec	accuracy=0.700184	lr=0.086103
Epoch[84] Batch [899]	Speed: 4093.311219 samples/sec	accuracy=0.700263	lr=0.085923
Epoch[84] Batch [949]	Speed: 4250.197837 samples/sec	accuracy=0.700353	lr=0.085744
Epoch[84] Batch [999]	Speed: 4089.896842 samples/sec	accuracy=0.700484	lr=0.085565
Epoch[84] Batch [1049]	Speed: 4236.776861 samples/sec	accuracy=0.700632	lr=0.085386
Epoch[84] Batch [1099]	Speed: 4155.705853 samples/sec	accuracy=0.700768	lr=0.085207
Epoch[84] Batch [1149]	Speed: 4047.225150 samples/sec	accuracy=0.700870	lr=0.085028
Epoch[84] Batch [1199]	Speed: 4243.207379 samples/sec	accuracy=0.701012	lr=0.084850
Epoch[84] Batch [1249]	Speed: 4384.917655 samples/sec	accuracy=0.701145	lr=0.084671
[Epoch 84] training: accuracy=0.701150
[Epoch 84] speed: 4152 samples/sec	time cost: 318.605153
[Epoch 84] validation: err-top1=0.282526 err-top5=0.090242
Epoch[85] Batch [49]	Speed: 4087.654305 samples/sec	accuracy=0.700967	lr=0.084489
Epoch[85] Batch [99]	Speed: 4192.460791 samples/sec	accuracy=0.700938	lr=0.084311
Epoch[85] Batch [149]	Speed: 4132.779346 samples/sec	accuracy=0.701137	lr=0.084133
Epoch[85] Batch [199]	Speed: 4089.035534 samples/sec	accuracy=0.701040	lr=0.083955
Epoch[85] Batch [249]	Speed: 4172.462220 samples/sec	accuracy=0.700990	lr=0.083777
Epoch[85] Batch [299]	Speed: 4090.119081 samples/sec	accuracy=0.701153	lr=0.083600
Epoch[85] Batch [349]	Speed: 4127.585609 samples/sec	accuracy=0.701360	lr=0.083422
Epoch[85] Batch [399]	Speed: 4147.155299 samples/sec	accuracy=0.701376	lr=0.083245
Epoch[85] Batch [449]	Speed: 4077.979631 samples/sec	accuracy=0.701492	lr=0.083067
Epoch[85] Batch [499]	Speed: 4175.496081 samples/sec	accuracy=0.701489	lr=0.082890
Epoch[85] Batch [549]	Speed: 4214.438893 samples/sec	accuracy=0.701520	lr=0.082713
Epoch[85] Batch [599]	Speed: 4203.260886 samples/sec	accuracy=0.701564	lr=0.082537
Epoch[85] Batch [649]	Speed: 4176.942116 samples/sec	accuracy=0.701604	lr=0.082360
Epoch[85] Batch [699]	Speed: 4159.486740 samples/sec	accuracy=0.701773	lr=0.082183
Epoch[85] Batch [749]	Speed: 4235.291614 samples/sec	accuracy=0.701852	lr=0.082007
Epoch[85] Batch [799]	Speed: 4148.288943 samples/sec	accuracy=0.701886	lr=0.081831
Epoch[85] Batch [849]	Speed: 4180.571639 samples/sec	accuracy=0.701920	lr=0.081655
Epoch[85] Batch [899]	Speed: 4198.554298 samples/sec	accuracy=0.701953	lr=0.081479
Epoch[85] Batch [949]	Speed: 4047.131029 samples/sec	accuracy=0.702044	lr=0.081303
Epoch[85] Batch [999]	Speed: 4225.570581 samples/sec	accuracy=0.702156	lr=0.081127
Epoch[85] Batch [1049]	Speed: 4059.044201 samples/sec	accuracy=0.702291	lr=0.080952
Epoch[85] Batch [1099]	Speed: 4314.930234 samples/sec	accuracy=0.702388	lr=0.080776
Epoch[85] Batch [1149]	Speed: 4133.823498 samples/sec	accuracy=0.702522	lr=0.080601
Epoch[85] Batch [1199]	Speed: 4275.611602 samples/sec	accuracy=0.702631	lr=0.080426
Epoch[85] Batch [1249]	Speed: 4257.713748 samples/sec	accuracy=0.702771	lr=0.080251
[Epoch 85] training: accuracy=0.702778
[Epoch 85] speed: 4160 samples/sec	time cost: 318.138851
[Epoch 85] validation: err-top1=0.277982 err-top5=0.090462
Epoch[86] Batch [49]	Speed: 4141.446923 samples/sec	accuracy=0.704443	lr=0.080073
Epoch[86] Batch [99]	Speed: 4170.033509 samples/sec	accuracy=0.703436	lr=0.079898
Epoch[86] Batch [149]	Speed: 4202.198963 samples/sec	accuracy=0.703196	lr=0.079723
Epoch[86] Batch [199]	Speed: 4098.878428 samples/sec	accuracy=0.703362	lr=0.079549
Epoch[86] Batch [249]	Speed: 4127.927649 samples/sec	accuracy=0.703390	lr=0.079375
Epoch[86] Batch [299]	Speed: 4177.488467 samples/sec	accuracy=0.703308	lr=0.079201
Epoch[86] Batch [349]	Speed: 4158.409776 samples/sec	accuracy=0.703372	lr=0.079027
Epoch[86] Batch [399]	Speed: 4235.229636 samples/sec	accuracy=0.703416	lr=0.078853
Epoch[86] Batch [449]	Speed: 4087.653605 samples/sec	accuracy=0.703484	lr=0.078679
Epoch[86] Batch [499]	Speed: 4205.821050 samples/sec	accuracy=0.703571	lr=0.078506
Epoch[86] Batch [549]	Speed: 4059.525686 samples/sec	accuracy=0.703724	lr=0.078332
Epoch[86] Batch [599]	Speed: 4082.093081 samples/sec	accuracy=0.703772	lr=0.078159
Epoch[86] Batch [649]	Speed: 4122.285657 samples/sec	accuracy=0.703840	lr=0.077986
Epoch[86] Batch [699]	Speed: 4124.525883 samples/sec	accuracy=0.703910	lr=0.077813
Epoch[86] Batch [749]	Speed: 4163.048450 samples/sec	accuracy=0.703996	lr=0.077640
Epoch[86] Batch [799]	Speed: 4272.260148 samples/sec	accuracy=0.704130	lr=0.077467
Epoch[86] Batch [849]	Speed: 4243.318137 samples/sec	accuracy=0.704239	lr=0.077295
Epoch[86] Batch [899]	Speed: 4180.399599 samples/sec	accuracy=0.704318	lr=0.077123
Epoch[86] Batch [949]	Speed: 4126.452944 samples/sec	accuracy=0.704354	lr=0.076950
Epoch[86] Batch [999]	Speed: 4122.329812 samples/sec	accuracy=0.704422	lr=0.076778
Epoch[86] Batch [1049]	Speed: 4265.068314 samples/sec	accuracy=0.704534	lr=0.076606
Epoch[86] Batch [1099]	Speed: 4113.483293 samples/sec	accuracy=0.704585	lr=0.076435
Epoch[86] Batch [1149]	Speed: 4154.983492 samples/sec	accuracy=0.704633	lr=0.076263
Epoch[86] Batch [1199]	Speed: 4182.861993 samples/sec	accuracy=0.704669	lr=0.076091
Epoch[86] Batch [1249]	Speed: 4219.824874 samples/sec	accuracy=0.704766	lr=0.075920
[Epoch 86] training: accuracy=0.704770
[Epoch 86] speed: 4157 samples/sec	time cost: 318.377766
[Epoch 86] validation: err-top1=0.280353 err-top5=0.089405
Epoch[87] Batch [49]	Speed: 4124.194941 samples/sec	accuracy=0.705469	lr=0.075745
Epoch[87] Batch [99]	Speed: 4202.182764 samples/sec	accuracy=0.705735	lr=0.075574
Epoch[87] Batch [149]	Speed: 3987.556293 samples/sec	accuracy=0.705442	lr=0.075404
Epoch[87] Batch [199]	Speed: 4057.530434 samples/sec	accuracy=0.705075	lr=0.075233
Epoch[87] Batch [249]	Speed: 4078.347655 samples/sec	accuracy=0.704825	lr=0.075062
Epoch[87] Batch [299]	Speed: 4341.719803 samples/sec	accuracy=0.704858	lr=0.074892
Epoch[87] Batch [349]	Speed: 4216.232775 samples/sec	accuracy=0.704949	lr=0.074721
Epoch[87] Batch [399]	Speed: 4109.832730 samples/sec	accuracy=0.705136	lr=0.074551
Epoch[87] Batch [449]	Speed: 4181.351447 samples/sec	accuracy=0.705160	lr=0.074381
Epoch[87] Batch [499]	Speed: 4232.385383 samples/sec	accuracy=0.705264	lr=0.074211
Epoch[87] Batch [549]	Speed: 4104.844133 samples/sec	accuracy=0.705347	lr=0.074042
Epoch[87] Batch [599]	Speed: 4059.953786 samples/sec	accuracy=0.705380	lr=0.073872
Epoch[87] Batch [649]	Speed: 4125.915183 samples/sec	accuracy=0.705499	lr=0.073703
Epoch[87] Batch [699]	Speed: 4137.508135 samples/sec	accuracy=0.705544	lr=0.073534
Epoch[87] Batch [749]	Speed: 4155.662748 samples/sec	accuracy=0.705597	lr=0.073364
Epoch[87] Batch [799]	Speed: 4269.856611 samples/sec	accuracy=0.705670	lr=0.073196
Epoch[87] Batch [849]	Speed: 4204.392242 samples/sec	accuracy=0.705748	lr=0.073027
Epoch[87] Batch [899]	Speed: 4224.738289 samples/sec	accuracy=0.705837	lr=0.072858
Epoch[87] Batch [949]	Speed: 4235.684737 samples/sec	accuracy=0.705858	lr=0.072690
Epoch[87] Batch [999]	Speed: 4244.228143 samples/sec	accuracy=0.705954	lr=0.072521
Epoch[87] Batch [1049]	Speed: 4330.023949 samples/sec	accuracy=0.706103	lr=0.072353
Epoch[87] Batch [1099]	Speed: 4157.826221 samples/sec	accuracy=0.706252	lr=0.072185
Epoch[87] Batch [1149]	Speed: 4220.541756 samples/sec	accuracy=0.706398	lr=0.072017
Epoch[87] Batch [1199]	Speed: 4154.248525 samples/sec	accuracy=0.706497	lr=0.071849
Epoch[87] Batch [1249]	Speed: 4289.656701 samples/sec	accuracy=0.706613	lr=0.071682
[Epoch 87] training: accuracy=0.706613
[Epoch 87] speed: 4172 samples/sec	time cost: 316.887340
[Epoch 87] validation: err-top1=0.282410 err-top5=0.090902
Epoch[88] Batch [49]	Speed: 4106.294951 samples/sec	accuracy=0.707287	lr=0.071511
Epoch[88] Batch [99]	Speed: 4203.919316 samples/sec	accuracy=0.707221	lr=0.071344
Epoch[88] Batch [149]	Speed: 4235.425097 samples/sec	accuracy=0.707104	lr=0.071177
Epoch[88] Batch [199]	Speed: 4276.534406 samples/sec	accuracy=0.707074	lr=0.071010
Epoch[88] Batch [249]	Speed: 4180.551862 samples/sec	accuracy=0.706953	lr=0.070843
Epoch[88] Batch [299]	Speed: 4206.397393 samples/sec	accuracy=0.707131	lr=0.070676
Epoch[88] Batch [349]	Speed: 4172.699198 samples/sec	accuracy=0.707013	lr=0.070510
Epoch[88] Batch [399]	Speed: 4212.034774 samples/sec	accuracy=0.707064	lr=0.070343
Epoch[88] Batch [449]	Speed: 4054.061502 samples/sec	accuracy=0.707127	lr=0.070177
Epoch[88] Batch [499]	Speed: 4260.959471 samples/sec	accuracy=0.707221	lr=0.070011
Epoch[88] Batch [549]	Speed: 4165.797820 samples/sec	accuracy=0.707255	lr=0.069845
Epoch[88] Batch [599]	Speed: 4283.681509 samples/sec	accuracy=0.707338	lr=0.069680
Epoch[88] Batch [649]	Speed: 4132.274365 samples/sec	accuracy=0.707455	lr=0.069514
Epoch[88] Batch [699]	Speed: 4190.168827 samples/sec	accuracy=0.707493	lr=0.069349
Epoch[88] Batch [749]	Speed: 4219.680847 samples/sec	accuracy=0.707558	lr=0.069183
Epoch[88] Batch [799]	Speed: 4204.103996 samples/sec	accuracy=0.707687	lr=0.069018
Epoch[88] Batch [849]	Speed: 4103.063802 samples/sec	accuracy=0.707744	lr=0.068853
Epoch[88] Batch [899]	Speed: 4207.957590 samples/sec	accuracy=0.707824	lr=0.068689
Epoch[88] Batch [949]	Speed: 4099.645429 samples/sec	accuracy=0.707909	lr=0.068524
Epoch[88] Batch [999]	Speed: 4140.777017 samples/sec	accuracy=0.708038	lr=0.068359
Epoch[88] Batch [1049]	Speed: 4141.777205 samples/sec	accuracy=0.708125	lr=0.068195
Epoch[88] Batch [1099]	Speed: 3998.112407 samples/sec	accuracy=0.708280	lr=0.068031
Epoch[88] Batch [1149]	Speed: 4157.308984 samples/sec	accuracy=0.708441	lr=0.067867
Epoch[88] Batch [1199]	Speed: 4236.330634 samples/sec	accuracy=0.708580	lr=0.067703
Epoch[88] Batch [1249]	Speed: 4252.883399 samples/sec	accuracy=0.708697	lr=0.067539
[Epoch 88] training: accuracy=0.708699
[Epoch 88] speed: 4172 samples/sec	time cost: 318.391883
[Epoch 88] validation: err-top1=0.277503 err-top5=0.088130
Epoch[89] Batch [49]	Speed: 4124.281989 samples/sec	accuracy=0.709317	lr=0.067373
Epoch[89] Batch [99]	Speed: 4077.463178 samples/sec	accuracy=0.709463	lr=0.067209
Epoch[89] Batch [149]	Speed: 4151.151563 samples/sec	accuracy=0.708945	lr=0.067046
Epoch[89] Batch [199]	Speed: 4184.508735 samples/sec	accuracy=0.708847	lr=0.066883
Epoch[89] Batch [249]	Speed: 4046.272313 samples/sec	accuracy=0.708787	lr=0.066720
Epoch[89] Batch [299]	Speed: 4100.789030 samples/sec	accuracy=0.708864	lr=0.066557
Epoch[89] Batch [349]	Speed: 4185.047770 samples/sec	accuracy=0.709085	lr=0.066395
Epoch[89] Batch [399]	Speed: 4103.587388 samples/sec	accuracy=0.709165	lr=0.066232
Epoch[89] Batch [449]	Speed: 4164.212522 samples/sec	accuracy=0.709160	lr=0.066070
Epoch[89] Batch [499]	Speed: 4141.974600 samples/sec	accuracy=0.709174	lr=0.065908
Epoch[89] Batch [549]	Speed: 4204.827731 samples/sec	accuracy=0.709258	lr=0.065746
Epoch[89] Batch [599]	Speed: 4122.328467 samples/sec	accuracy=0.709302	lr=0.065584
Epoch[89] Batch [649]	Speed: 4158.114597 samples/sec	accuracy=0.709410	lr=0.065423
Epoch[89] Batch [699]	Speed: 4129.013967 samples/sec	accuracy=0.709529	lr=0.065261
Epoch[89] Batch [749]	Speed: 4206.167199 samples/sec	accuracy=0.709586	lr=0.065100
Epoch[89] Batch [799]	Speed: 4183.690500 samples/sec	accuracy=0.709651	lr=0.064939
Epoch[89] Batch [849]	Speed: 4108.401812 samples/sec	accuracy=0.709758	lr=0.064778
Epoch[89] Batch [899]	Speed: 4006.771146 samples/sec	accuracy=0.709853	lr=0.064617
Epoch[89] Batch [949]	Speed: 4094.520066 samples/sec	accuracy=0.709940	lr=0.064456
Epoch[89] Batch [999]	Speed: 4120.026362 samples/sec	accuracy=0.710082	lr=0.064296
Epoch[89] Batch [1049]	Speed: 4135.104248 samples/sec	accuracy=0.710222	lr=0.064135
Epoch[89] Batch [1099]	Speed: 4120.562826 samples/sec	accuracy=0.710279	lr=0.063975
Epoch[89] Batch [1149]	Speed: 4101.243657 samples/sec	accuracy=0.710344	lr=0.063815
Epoch[89] Batch [1199]	Speed: 4129.087165 samples/sec	accuracy=0.710410	lr=0.063655
Epoch[89] Batch [1249]	Speed: 4368.965247 samples/sec	accuracy=0.710528	lr=0.063496
[Epoch 89] training: accuracy=0.710529
[Epoch 89] speed: 4134 samples/sec	time cost: 319.778873
[Epoch 89] validation: err-top1=0.276925 err-top5=0.089485
Epoch[90] Batch [49]	Speed: 4007.963372 samples/sec	accuracy=0.711137	lr=0.063333
Epoch[90] Batch [99]	Speed: 4136.202234 samples/sec	accuracy=0.711535	lr=0.063174
Epoch[90] Batch [149]	Speed: 4111.804812 samples/sec	accuracy=0.711382	lr=0.063015
Epoch[90] Batch [199]	Speed: 4122.084516 samples/sec	accuracy=0.711387	lr=0.062855
Epoch[90] Batch [249]	Speed: 4220.643370 samples/sec	accuracy=0.711373	lr=0.062697
Epoch[90] Batch [299]	Speed: 4237.856334 samples/sec	accuracy=0.711379	lr=0.062538
Epoch[90] Batch [349]	Speed: 4089.923715 samples/sec	accuracy=0.711428	lr=0.062379
Epoch[90] Batch [399]	Speed: 4163.200663 samples/sec	accuracy=0.711403	lr=0.062221
Epoch[90] Batch [449]	Speed: 4192.987467 samples/sec	accuracy=0.711317	lr=0.062063
Epoch[90] Batch [499]	Speed: 4265.296783 samples/sec	accuracy=0.711412	lr=0.061905
Epoch[90] Batch [549]	Speed: 4115.769885 samples/sec	accuracy=0.711410	lr=0.061747
Epoch[90] Batch [599]	Speed: 4096.511236 samples/sec	accuracy=0.711491	lr=0.061589
Epoch[90] Batch [649]	Speed: 4185.845322 samples/sec	accuracy=0.711499	lr=0.061432
Epoch[90] Batch [699]	Speed: 4222.918645 samples/sec	accuracy=0.711634	lr=0.061274
Epoch[90] Batch [749]	Speed: 4203.224934 samples/sec	accuracy=0.711713	lr=0.061117
Epoch[90] Batch [799]	Speed: 4210.343848 samples/sec	accuracy=0.711753	lr=0.060960
Epoch[90] Batch [849]	Speed: 4308.169423 samples/sec	accuracy=0.711851	lr=0.060803
Epoch[90] Batch [899]	Speed: 4203.381992 samples/sec	accuracy=0.711895	lr=0.060646
Epoch[90] Batch [949]	Speed: 4162.985744 samples/sec	accuracy=0.711949	lr=0.060490
Epoch[90] Batch [999]	Speed: 4102.808408 samples/sec	accuracy=0.711995	lr=0.060333
Epoch[90] Batch [1049]	Speed: 4163.512144 samples/sec	accuracy=0.712056	lr=0.060177
Epoch[90] Batch [1099]	Speed: 4253.286871 samples/sec	accuracy=0.712140	lr=0.060021
Epoch[90] Batch [1149]	Speed: 4183.995762 samples/sec	accuracy=0.712252	lr=0.059865
Epoch[90] Batch [1199]	Speed: 4212.110368 samples/sec	accuracy=0.712381	lr=0.059710
Epoch[90] Batch [1249]	Speed: 4263.109843 samples/sec	accuracy=0.712468	lr=0.059554
[Epoch 90] training: accuracy=0.712471
[Epoch 90] speed: 4172 samples/sec	time cost: 318.143867
[Epoch 90] validation: err-top1=0.267738 err-top5=0.083845
Epoch[91] Batch [49]	Speed: 3877.272246 samples/sec	accuracy=0.713827	lr=0.059396
Epoch[91] Batch [99]	Speed: 4126.715334 samples/sec	accuracy=0.713330	lr=0.059240
Epoch[91] Batch [149]	Speed: 4258.776891 samples/sec	accuracy=0.713342	lr=0.059085
Epoch[91] Batch [199]	Speed: 4163.396715 samples/sec	accuracy=0.713376	lr=0.058930
Epoch[91] Batch [249]	Speed: 4037.415172 samples/sec	accuracy=0.713131	lr=0.058776
Epoch[91] Batch [299]	Speed: 4173.862184 samples/sec	accuracy=0.713055	lr=0.058621
Epoch[91] Batch [349]	Speed: 4111.436630 samples/sec	accuracy=0.713112	lr=0.058467
Epoch[91] Batch [399]	Speed: 4196.098734 samples/sec	accuracy=0.713139	lr=0.058313
Epoch[91] Batch [449]	Speed: 3999.049547 samples/sec	accuracy=0.713305	lr=0.058158
Epoch[91] Batch [499]	Speed: 4162.241570 samples/sec	accuracy=0.713498	lr=0.058005
Epoch[91] Batch [549]	Speed: 4163.871790 samples/sec	accuracy=0.713509	lr=0.057851
Epoch[91] Batch [599]	Speed: 4106.634571 samples/sec	accuracy=0.713579	lr=0.057697
Epoch[91] Batch [649]	Speed: 4155.040088 samples/sec	accuracy=0.713640	lr=0.057544
Epoch[91] Batch [699]	Speed: 4315.620214 samples/sec	accuracy=0.713718	lr=0.057391
Epoch[91] Batch [749]	Speed: 4217.208301 samples/sec	accuracy=0.713756	lr=0.057238
Epoch[91] Batch [799]	Speed: 4125.999291 samples/sec	accuracy=0.713886	lr=0.057085
Epoch[91] Batch [849]	Speed: 4118.229523 samples/sec	accuracy=0.713948	lr=0.056932
Epoch[91] Batch [899]	Speed: 4173.525062 samples/sec	accuracy=0.714022	lr=0.056780
Epoch[91] Batch [949]	Speed: 4112.207394 samples/sec	accuracy=0.714114	lr=0.056627
Epoch[91] Batch [999]	Speed: 4127.315096 samples/sec	accuracy=0.714138	lr=0.056475
Epoch[91] Batch [1049]	Speed: 4041.680437 samples/sec	accuracy=0.714204	lr=0.056323
Epoch[91] Batch [1099]	Speed: 4174.822663 samples/sec	accuracy=0.714268	lr=0.056171
Epoch[91] Batch [1149]	Speed: 4208.963936 samples/sec	accuracy=0.714379	lr=0.056020
Epoch[91] Batch [1199]	Speed: 4171.577944 samples/sec	accuracy=0.714477	lr=0.055868
Epoch[91] Batch [1249]	Speed: 4343.895721 samples/sec	accuracy=0.714575	lr=0.055717
[Epoch 91] training: accuracy=0.714577
[Epoch 91] speed: 4141 samples/sec	time cost: 319.888031
[Epoch 91] validation: err-top1=0.266303 err-top5=0.082609
Epoch[92] Batch [49]	Speed: 3985.663470 samples/sec	accuracy=0.716204	lr=0.055563
Epoch[92] Batch [99]	Speed: 4113.967456 samples/sec	accuracy=0.715676	lr=0.055412
Epoch[92] Batch [149]	Speed: 4068.256463 samples/sec	accuracy=0.715238	lr=0.055261
Epoch[92] Batch [199]	Speed: 4195.722269 samples/sec	accuracy=0.715401	lr=0.055111
Epoch[92] Batch [249]	Speed: 4036.252473 samples/sec	accuracy=0.715415	lr=0.054960
Epoch[92] Batch [299]	Speed: 4159.301609 samples/sec	accuracy=0.715310	lr=0.054810
Epoch[92] Batch [349]	Speed: 4148.617752 samples/sec	accuracy=0.715261	lr=0.054660
Epoch[92] Batch [399]	Speed: 4267.740308 samples/sec	accuracy=0.715208	lr=0.054510
Epoch[92] Batch [449]	Speed: 4194.663905 samples/sec	accuracy=0.715247	lr=0.054360
Epoch[92] Batch [499]	Speed: 4159.994365 samples/sec	accuracy=0.715303	lr=0.054210
Epoch[92] Batch [549]	Speed: 4079.565581 samples/sec	accuracy=0.715320	lr=0.054061
Epoch[92] Batch [599]	Speed: 4031.056752 samples/sec	accuracy=0.715432	lr=0.053912
Epoch[92] Batch [649]	Speed: 4170.913566 samples/sec	accuracy=0.715530	lr=0.053763
Epoch[92] Batch [699]	Speed: 4255.774001 samples/sec	accuracy=0.715575	lr=0.053614
Epoch[92] Batch [749]	Speed: 4234.346611 samples/sec	accuracy=0.715714	lr=0.053465
Epoch[92] Batch [799]	Speed: 4143.251295 samples/sec	accuracy=0.715850	lr=0.053317
Epoch[92] Batch [849]	Speed: 4097.095058 samples/sec	accuracy=0.715950	lr=0.053168
Epoch[92] Batch [899]	Speed: 4026.883186 samples/sec	accuracy=0.716047	lr=0.053020
Epoch[92] Batch [949]	Speed: 4075.974248 samples/sec	accuracy=0.716190	lr=0.052872
Epoch[92] Batch [999]	Speed: 4081.380880 samples/sec	accuracy=0.716347	lr=0.052724
Epoch[92] Batch [1049]	Speed: 4124.007315 samples/sec	accuracy=0.716473	lr=0.052577
Epoch[92] Batch [1099]	Speed: 4109.612040 samples/sec	accuracy=0.716584	lr=0.052429
Epoch[92] Batch [1149]	Speed: 4153.430352 samples/sec	accuracy=0.716672	lr=0.052282
Epoch[92] Batch [1199]	Speed: 4209.738034 samples/sec	accuracy=0.716758	lr=0.052135
Epoch[92] Batch [1249]	Speed: 4319.683228 samples/sec	accuracy=0.716848	lr=0.051988
[Epoch 92] training: accuracy=0.716848
[Epoch 92] speed: 4132 samples/sec	time cost: 319.873937
[Epoch 92] validation: err-top1=0.261141 err-top5=0.080198
Epoch[93] Batch [49]	Speed: 4057.047506 samples/sec	accuracy=0.717976	lr=0.051838
Epoch[93] Batch [99]	Speed: 4215.859972 samples/sec	accuracy=0.717242	lr=0.051691
Epoch[93] Batch [149]	Speed: 4073.253045 samples/sec	accuracy=0.717196	lr=0.051545
Epoch[93] Batch [199]	Speed: 4266.440933 samples/sec	accuracy=0.717028	lr=0.051399
Epoch[93] Batch [249]	Speed: 4204.096342 samples/sec	accuracy=0.716941	lr=0.051253
Epoch[93] Batch [299]	Speed: 4085.537418 samples/sec	accuracy=0.716821	lr=0.051107
Epoch[93] Batch [349]	Speed: 4276.457931 samples/sec	accuracy=0.716913	lr=0.050961
Epoch[93] Batch [399]	Speed: 4171.489213 samples/sec	accuracy=0.716964	lr=0.050816
Epoch[93] Batch [449]	Speed: 4066.278107 samples/sec	accuracy=0.716887	lr=0.050670
Epoch[93] Batch [499]	Speed: 4223.756784 samples/sec	accuracy=0.716859	lr=0.050525
Epoch[93] Batch [549]	Speed: 4094.170427 samples/sec	accuracy=0.716850	lr=0.050380
Epoch[93] Batch [599]	Speed: 4050.601782 samples/sec	accuracy=0.716984	lr=0.050235
Epoch[93] Batch [649]	Speed: 4105.155183 samples/sec	accuracy=0.717100	lr=0.050091
Epoch[93] Batch [699]	Speed: 4181.490753 samples/sec	accuracy=0.717200	lr=0.049946
Epoch[93] Batch [749]	Speed: 4257.626718 samples/sec	accuracy=0.717241	lr=0.049802
Epoch[93] Batch [799]	Speed: 4252.677059 samples/sec	accuracy=0.717303	lr=0.049658
Epoch[93] Batch [849]	Speed: 4240.194750 samples/sec	accuracy=0.717427	lr=0.049514
Epoch[93] Batch [899]	Speed: 4278.950841 samples/sec	accuracy=0.717521	lr=0.049370
Epoch[93] Batch [949]	Speed: 4184.987906 samples/sec	accuracy=0.717643	lr=0.049227
Epoch[93] Batch [999]	Speed: 4239.399201 samples/sec	accuracy=0.717760	lr=0.049083
Epoch[93] Batch [1049]	Speed: 4302.411069 samples/sec	accuracy=0.717867	lr=0.048940
Epoch[93] Batch [1099]	Speed: 4268.755515 samples/sec	accuracy=0.717882	lr=0.048797
Epoch[93] Batch [1149]	Speed: 4197.288256 samples/sec	accuracy=0.717939	lr=0.048654
Epoch[93] Batch [1199]	Speed: 4172.269771 samples/sec	accuracy=0.718020	lr=0.048511
Epoch[93] Batch [1249]	Speed: 4226.304055 samples/sec	accuracy=0.718087	lr=0.048369
[Epoch 93] training: accuracy=0.718090
[Epoch 93] speed: 4182 samples/sec	time cost: 316.873313
[Epoch 93] validation: err-top1=0.262309 err-top5=0.079915
Epoch[94] Batch [49]	Speed: 4116.576996 samples/sec	accuracy=0.719342	lr=0.048224
Epoch[94] Batch [99]	Speed: 4138.937386 samples/sec	accuracy=0.718349	lr=0.048082
Epoch[94] Batch [149]	Speed: 4266.792894 samples/sec	accuracy=0.718039	lr=0.047940
Epoch[94] Batch [199]	Speed: 4206.256752 samples/sec	accuracy=0.717972	lr=0.047798
Epoch[94] Batch [249]	Speed: 4151.612370 samples/sec	accuracy=0.718087	lr=0.047656
Epoch[94] Batch [299]	Speed: 4176.780286 samples/sec	accuracy=0.718031	lr=0.047515
Epoch[94] Batch [349]	Speed: 4246.951937 samples/sec	accuracy=0.718142	lr=0.047374
Epoch[94] Batch [399]	Speed: 4149.285468 samples/sec	accuracy=0.718230	lr=0.047233
Epoch[94] Batch [449]	Speed: 4214.461721 samples/sec	accuracy=0.718402	lr=0.047092
Epoch[94] Batch [499]	Speed: 4225.968805 samples/sec	accuracy=0.718541	lr=0.046951
Epoch[94] Batch [549]	Speed: 4179.507564 samples/sec	accuracy=0.718682	lr=0.046811
Epoch[94] Batch [599]	Speed: 4269.111423 samples/sec	accuracy=0.718744	lr=0.046670
Epoch[94] Batch [649]	Speed: 4151.206450 samples/sec	accuracy=0.718748	lr=0.046530
Epoch[94] Batch [699]	Speed: 4210.248259 samples/sec	accuracy=0.718832	lr=0.046390
Epoch[94] Batch [749]	Speed: 4092.507668 samples/sec	accuracy=0.718914	lr=0.046251
Epoch[94] Batch [799]	Speed: 4110.745392 samples/sec	accuracy=0.719011	lr=0.046111
Epoch[94] Batch [849]	Speed: 4198.907544 samples/sec	accuracy=0.719080	lr=0.045972
Epoch[94] Batch [899]	Speed: 4152.080584 samples/sec	accuracy=0.719170	lr=0.045833
Epoch[94] Batch [949]	Speed: 4263.286896 samples/sec	accuracy=0.719230	lr=0.045693
Epoch[94] Batch [999]	Speed: 4037.964428 samples/sec	accuracy=0.719298	lr=0.045555
Epoch[94] Batch [1049]	Speed: 4202.994183 samples/sec	accuracy=0.719412	lr=0.045416
Epoch[94] Batch [1099]	Speed: 4299.251844 samples/sec	accuracy=0.719511	lr=0.045278
Epoch[94] Batch [1149]	Speed: 4192.355701 samples/sec	accuracy=0.719663	lr=0.045139
Epoch[94] Batch [1199]	Speed: 4280.841973 samples/sec	accuracy=0.719739	lr=0.045001
Epoch[94] Batch [1249]	Speed: 4214.968294 samples/sec	accuracy=0.719892	lr=0.044863
[Epoch 94] training: accuracy=0.719893
[Epoch 94] speed: 4185 samples/sec	time cost: 316.693516
[Epoch 94] validation: err-top1=0.258131 err-top5=0.078185
Epoch[95] Batch [49]	Speed: 4050.265332 samples/sec	accuracy=0.721694	lr=0.044723
Epoch[95] Batch [99]	Speed: 4256.259004 samples/sec	accuracy=0.720676	lr=0.044585
Epoch[95] Batch [149]	Speed: 4146.367777 samples/sec	accuracy=0.720619	lr=0.044448
Epoch[95] Batch [199]	Speed: 4114.665769 samples/sec	accuracy=0.720536	lr=0.044311
Epoch[95] Batch [249]	Speed: 4057.755150 samples/sec	accuracy=0.720592	lr=0.044174
Epoch[95] Batch [299]	Speed: 4142.966336 samples/sec	accuracy=0.720709	lr=0.044037
Epoch[95] Batch [349]	Speed: 4173.116874 samples/sec	accuracy=0.720749	lr=0.043900
Epoch[95] Batch [399]	Speed: 4200.967046 samples/sec	accuracy=0.720835	lr=0.043764
Epoch[95] Batch [449]	Speed: 4169.238166 samples/sec	accuracy=0.720895	lr=0.043628
Epoch[95] Batch [499]	Speed: 4121.213316 samples/sec	accuracy=0.721005	lr=0.043492
Epoch[95] Batch [549]	Speed: 4194.532897 samples/sec	accuracy=0.721100	lr=0.043356
Epoch[95] Batch [599]	Speed: 4154.464632 samples/sec	accuracy=0.721150	lr=0.043220
Epoch[95] Batch [649]	Speed: 4212.799341 samples/sec	accuracy=0.721187	lr=0.043085
Epoch[95] Batch [699]	Speed: 4125.601693 samples/sec	accuracy=0.721288	lr=0.042949
Epoch[95] Batch [749]	Speed: 4129.338933 samples/sec	accuracy=0.721397	lr=0.042814
Epoch[95] Batch [799]	Speed: 4164.870971 samples/sec	accuracy=0.721464	lr=0.042679
Epoch[95] Batch [849]	Speed: 4339.613412 samples/sec	accuracy=0.721559	lr=0.042545
Epoch[95] Batch [899]	Speed: 4120.948857 samples/sec	accuracy=0.721617	lr=0.042410
Epoch[95] Batch [949]	Speed: 4049.451406 samples/sec	accuracy=0.721679	lr=0.042276
Epoch[95] Batch [999]	Speed: 4102.421379 samples/sec	accuracy=0.721741	lr=0.042141
Epoch[95] Batch [1049]	Speed: 4195.570620 samples/sec	accuracy=0.721791	lr=0.042007
Epoch[95] Batch [1099]	Speed: 4263.919650 samples/sec	accuracy=0.721954	lr=0.041874
Epoch[95] Batch [1149]	Speed: 4263.304670 samples/sec	accuracy=0.722036	lr=0.041740
Epoch[95] Batch [1199]	Speed: 4172.053834 samples/sec	accuracy=0.722148	lr=0.041607
Epoch[95] Batch [1249]	Speed: 4313.374795 samples/sec	accuracy=0.722216	lr=0.041473
[Epoch 95] training: accuracy=0.722216
[Epoch 95] speed: 4164 samples/sec	time cost: 317.639601
[Epoch 95] validation: err-top1=0.255142 err-top5=0.076610
Epoch[96] Batch [49]	Speed: 4064.094000 samples/sec	accuracy=0.722589	lr=0.041338
Epoch[96] Batch [99]	Speed: 4228.641977 samples/sec	accuracy=0.722334	lr=0.041205
Epoch[96] Batch [149]	Speed: 4176.159565 samples/sec	accuracy=0.722522	lr=0.041072
Epoch[96] Batch [199]	Speed: 4155.437672 samples/sec	accuracy=0.722650	lr=0.040940
Epoch[96] Batch [249]	Speed: 4147.421371 samples/sec	accuracy=0.722449	lr=0.040807
Epoch[96] Batch [299]	Speed: 4174.222241 samples/sec	accuracy=0.722287	lr=0.040675
Epoch[96] Batch [349]	Speed: 4212.179354 samples/sec	accuracy=0.722315	lr=0.040543
Epoch[96] Batch [399]	Speed: 4129.250720 samples/sec	accuracy=0.722307	lr=0.040412
Epoch[96] Batch [449]	Speed: 4218.699702 samples/sec	accuracy=0.722246	lr=0.040280
Epoch[96] Batch [499]	Speed: 4179.122765 samples/sec	accuracy=0.722282	lr=0.040149
Epoch[96] Batch [549]	Speed: 4156.643430 samples/sec	accuracy=0.722328	lr=0.040018
Epoch[96] Batch [599]	Speed: 4112.300235 samples/sec	accuracy=0.722444	lr=0.039887
Epoch[96] Batch [649]	Speed: 4147.254051 samples/sec	accuracy=0.722508	lr=0.039756
Epoch[96] Batch [699]	Speed: 4187.628985 samples/sec	accuracy=0.722630	lr=0.039625
Epoch[96] Batch [749]	Speed: 4075.980437 samples/sec	accuracy=0.722620	lr=0.039495
Epoch[96] Batch [799]	Speed: 4200.945679 samples/sec	accuracy=0.722750	lr=0.039365
Epoch[96] Batch [849]	Speed: 4120.881482 samples/sec	accuracy=0.722868	lr=0.039235
Epoch[96] Batch [899]	Speed: 4085.390754 samples/sec	accuracy=0.722893	lr=0.039105
Epoch[96] Batch [949]	Speed: 4171.301310 samples/sec	accuracy=0.722963	lr=0.038975
Epoch[96] Batch [999]	Speed: 4150.325706 samples/sec	accuracy=0.723134	lr=0.038846
Epoch[96] Batch [1049]	Speed: 4140.138216 samples/sec	accuracy=0.723258	lr=0.038717
Epoch[96] Batch [1099]	Speed: 4149.517496 samples/sec	accuracy=0.723304	lr=0.038588
Epoch[96] Batch [1149]	Speed: 4141.676078 samples/sec	accuracy=0.723441	lr=0.038459
Epoch[96] Batch [1199]	Speed: 4096.011094 samples/sec	accuracy=0.723537	lr=0.038330
Epoch[96] Batch [1249]	Speed: 4393.335663 samples/sec	accuracy=0.723685	lr=0.038202
[Epoch 96] training: accuracy=0.723686
[Epoch 96] speed: 4156 samples/sec	time cost: 319.061035
[Epoch 96] validation: err-top1=0.257713 err-top5=0.077467
Epoch[97] Batch [49]	Speed: 3899.234671 samples/sec	accuracy=0.725953	lr=0.038071
Epoch[97] Batch [99]	Speed: 4219.620735 samples/sec	accuracy=0.725233	lr=0.037943
Epoch[97] Batch [149]	Speed: 4144.149751 samples/sec	accuracy=0.725005	lr=0.037815
Epoch[97] Batch [199]	Speed: 4121.630952 samples/sec	accuracy=0.724949	lr=0.037687
Epoch[97] Batch [249]	Speed: 4345.260119 samples/sec	accuracy=0.724890	lr=0.037560
Epoch[97] Batch [299]	Speed: 4123.914577 samples/sec	accuracy=0.724851	lr=0.037432
Epoch[97] Batch [349]	Speed: 4205.490111 samples/sec	accuracy=0.724764	lr=0.037305
Epoch[97] Batch [399]	Speed: 4120.810235 samples/sec	accuracy=0.724796	lr=0.037178
Epoch[97] Batch [449]	Speed: 4219.646107 samples/sec	accuracy=0.724798	lr=0.037052
Epoch[97] Batch [499]	Speed: 3961.526776 samples/sec	accuracy=0.724772	lr=0.036925
Epoch[97] Batch [549]	Speed: 4136.437501 samples/sec	accuracy=0.724775	lr=0.036799
Epoch[97] Batch [599]	Speed: 4124.310979 samples/sec	accuracy=0.724851	lr=0.036673
Epoch[97] Batch [649]	Speed: 4207.252559 samples/sec	accuracy=0.724981	lr=0.036547
Epoch[97] Batch [699]	Speed: 4280.671992 samples/sec	accuracy=0.725111	lr=0.036421
Epoch[97] Batch [749]	Speed: 4192.176634 samples/sec	accuracy=0.725159	lr=0.036296
Epoch[97] Batch [799]	Speed: 4302.275141 samples/sec	accuracy=0.725271	lr=0.036170
Epoch[97] Batch [849]	Speed: 4253.392174 samples/sec	accuracy=0.725391	lr=0.036045
Epoch[97] Batch [899]	Speed: 4161.287274 samples/sec	accuracy=0.725464	lr=0.035920
Epoch[97] Batch [949]	Speed: 4173.345410 samples/sec	accuracy=0.725554	lr=0.035795
Epoch[97] Batch [999]	Speed: 4039.333016 samples/sec	accuracy=0.725671	lr=0.035671
Epoch[97] Batch [1049]	Speed: 4195.220639 samples/sec	accuracy=0.725800	lr=0.035546
Epoch[97] Batch [1099]	Speed: 4069.962357 samples/sec	accuracy=0.725889	lr=0.035422
Epoch[97] Batch [1149]	Speed: 4131.816569 samples/sec	accuracy=0.725989	lr=0.035298
Epoch[97] Batch [1199]	Speed: 4153.466661 samples/sec	accuracy=0.726053	lr=0.035174
Epoch[97] Batch [1249]	Speed: 4197.186779 samples/sec	accuracy=0.726085	lr=0.035051
[Epoch 97] training: accuracy=0.726087
[Epoch 97] speed: 4153 samples/sec	time cost: 318.441711
[Epoch 97] validation: err-top1=0.250698 err-top5=0.073980
Epoch[98] Batch [49]	Speed: 4138.104340 samples/sec	accuracy=0.726665	lr=0.034925
Epoch[98] Batch [99]	Speed: 4087.058157 samples/sec	accuracy=0.726675	lr=0.034802
Epoch[98] Batch [149]	Speed: 4026.092518 samples/sec	accuracy=0.726405	lr=0.034679
Epoch[98] Batch [199]	Speed: 4063.877195 samples/sec	accuracy=0.726430	lr=0.034556
Epoch[98] Batch [249]	Speed: 4072.559989 samples/sec	accuracy=0.726391	lr=0.034433
Epoch[98] Batch [299]	Speed: 4145.228782 samples/sec	accuracy=0.726403	lr=0.034311
Epoch[98] Batch [349]	Speed: 4163.238919 samples/sec	accuracy=0.726363	lr=0.034189
Epoch[98] Batch [399]	Speed: 4131.517204 samples/sec	accuracy=0.726373	lr=0.034067
Epoch[98] Batch [449]	Speed: 3999.622828 samples/sec	accuracy=0.726424	lr=0.033945
Epoch[98] Batch [499]	Speed: 4091.983318 samples/sec	accuracy=0.726503	lr=0.033823
Epoch[98] Batch [549]	Speed: 4201.779968 samples/sec	accuracy=0.726573	lr=0.033702
Epoch[98] Batch [599]	Speed: 4097.200664 samples/sec	accuracy=0.726631	lr=0.033581
Epoch[98] Batch [649]	Speed: 4145.009635 samples/sec	accuracy=0.726711	lr=0.033460
Epoch[98] Batch [699]	Speed: 4098.377082 samples/sec	accuracy=0.726780	lr=0.033339
Epoch[98] Batch [749]	Speed: 4151.460201 samples/sec	accuracy=0.726852	lr=0.033218
Epoch[98] Batch [799]	Speed: 4146.535986 samples/sec	accuracy=0.726880	lr=0.033098
Epoch[98] Batch [849]	Speed: 4148.554919 samples/sec	accuracy=0.726920	lr=0.032978
Epoch[98] Batch [899]	Speed: 4102.493089 samples/sec	accuracy=0.726966	lr=0.032858
Epoch[98] Batch [949]	Speed: 4065.757609 samples/sec	accuracy=0.727037	lr=0.032738
Epoch[98] Batch [999]	Speed: 4065.267564 samples/sec	accuracy=0.727130	lr=0.032618
Epoch[98] Batch [1049]	Speed: 4066.479691 samples/sec	accuracy=0.727215	lr=0.032499
Epoch[98] Batch [1099]	Speed: 4171.129060 samples/sec	accuracy=0.727231	lr=0.032380
Epoch[98] Batch [1149]	Speed: 4138.105058 samples/sec	accuracy=0.727314	lr=0.032261
Epoch[98] Batch [1199]	Speed: 4146.246893 samples/sec	accuracy=0.727424	lr=0.032142
Epoch[98] Batch [1249]	Speed: 4234.241164 samples/sec	accuracy=0.727505	lr=0.032023
[Epoch 98] training: accuracy=0.727505
[Epoch 98] speed: 4111 samples/sec	time cost: 321.564528
[Epoch 98] validation: err-top1=0.254252 err-top5=0.076416
Epoch[99] Batch [49]	Speed: 3752.688759 samples/sec	accuracy=0.729215	lr=0.031902
Epoch[99] Batch [99]	Speed: 4149.397390 samples/sec	accuracy=0.728689	lr=0.031784
Epoch[99] Batch [149]	Speed: 4114.929344 samples/sec	accuracy=0.728377	lr=0.031666
Epoch[99] Batch [199]	Speed: 4063.735542 samples/sec	accuracy=0.728099	lr=0.031548
Epoch[99] Batch [249]	Speed: 4157.690822 samples/sec	accuracy=0.728030	lr=0.031431
Epoch[99] Batch [299]	Speed: 4128.363076 samples/sec	accuracy=0.728206	lr=0.031313
Epoch[99] Batch [349]	Speed: 4153.389222 samples/sec	accuracy=0.728123	lr=0.031196
Epoch[99] Batch [399]	Speed: 4141.748767 samples/sec	accuracy=0.728094	lr=0.031079
Epoch[99] Batch [449]	Speed: 4116.950994 samples/sec	accuracy=0.728136	lr=0.030962
Epoch[99] Batch [499]	Speed: 4059.142330 samples/sec	accuracy=0.728107	lr=0.030846
Epoch[99] Batch [549]	Speed: 4076.999565 samples/sec	accuracy=0.728141	lr=0.030729
Epoch[99] Batch [599]	Speed: 4145.084122 samples/sec	accuracy=0.728206	lr=0.030613
Epoch[99] Batch [649]	Speed: 4070.725365 samples/sec	accuracy=0.728285	lr=0.030497
Epoch[99] Batch [699]	Speed: 4100.455388 samples/sec	accuracy=0.728390	lr=0.030381
Epoch[99] Batch [749]	Speed: 4150.779349 samples/sec	accuracy=0.728479	lr=0.030266
Epoch[99] Batch [799]	Speed: 4076.554863 samples/sec	accuracy=0.728535	lr=0.030150
Epoch[99] Batch [849]	Speed: 4127.240533 samples/sec	accuracy=0.728619	lr=0.030035
Epoch[99] Batch [899]	Speed: 4145.483483 samples/sec	accuracy=0.728675	lr=0.029920
Epoch[99] Batch [949]	Speed: 4070.316283 samples/sec	accuracy=0.728783	lr=0.029805
Epoch[99] Batch [999]	Speed: 4145.787917 samples/sec	accuracy=0.728849	lr=0.029691
Epoch[99] Batch [1049]	Speed: 4131.271846 samples/sec	accuracy=0.728927	lr=0.029576
Epoch[99] Batch [1099]	Speed: 4133.304260 samples/sec	accuracy=0.729000	lr=0.029462
Epoch[99] Batch [1149]	Speed: 4096.269393 samples/sec	accuracy=0.729089	lr=0.029348
Epoch[99] Batch [1199]	Speed: 4086.724490 samples/sec	accuracy=0.729170	lr=0.029234
Epoch[99] Batch [1249]	Speed: 4194.192756 samples/sec	accuracy=0.729272	lr=0.029121
[Epoch 99] training: accuracy=0.729274
[Epoch 99] speed: 4098 samples/sec	time cost: 323.616536
[Epoch 99] validation: err-top1=0.245596 err-top5=0.071867
Epoch[100] Batch [49]	Speed: 3941.818731 samples/sec	accuracy=0.729767	lr=0.029005
Epoch[100] Batch [99]	Speed: 4011.914755 samples/sec	accuracy=0.730081	lr=0.028892
Epoch[100] Batch [149]	Speed: 4125.131982 samples/sec	accuracy=0.729817	lr=0.028779
Epoch[100] Batch [199]	Speed: 4302.524250 samples/sec	accuracy=0.729691	lr=0.028666
Epoch[100] Batch [249]	Speed: 4183.085080 samples/sec	accuracy=0.729371	lr=0.028554
Epoch[100] Batch [299]	Speed: 4133.569512 samples/sec	accuracy=0.729296	lr=0.028441
Epoch[100] Batch [349]	Speed: 4169.475506 samples/sec	accuracy=0.729364	lr=0.028329
Epoch[100] Batch [399]	Speed: 4093.340634 samples/sec	accuracy=0.729441	lr=0.028217
Epoch[100] Batch [449]	Speed: 4106.295658 samples/sec	accuracy=0.729601	lr=0.028105
Epoch[100] Batch [499]	Speed: 3986.681372 samples/sec	accuracy=0.729706	lr=0.027994
Epoch[100] Batch [549]	Speed: 4067.701479 samples/sec	accuracy=0.729681	lr=0.027883
Epoch[100] Batch [599]	Speed: 4194.784106 samples/sec	accuracy=0.729691	lr=0.027772
Epoch[100] Batch [649]	Speed: 4287.233213 samples/sec	accuracy=0.729713	lr=0.027661
Epoch[100] Batch [699]	Speed: 4167.081078 samples/sec	accuracy=0.729783	lr=0.027550
Epoch[100] Batch [749]	Speed: 4217.463973 samples/sec	accuracy=0.729822	lr=0.027439
Epoch[100] Batch [799]	Speed: 4235.206833 samples/sec	accuracy=0.729883	lr=0.027329
Epoch[100] Batch [849]	Speed: 3963.834361 samples/sec	accuracy=0.730009	lr=0.027219
Epoch[100] Batch [899]	Speed: 4095.983984 samples/sec	accuracy=0.730130	lr=0.027109
Epoch[100] Batch [949]	Speed: 4139.780025 samples/sec	accuracy=0.730255	lr=0.026999
Epoch[100] Batch [999]	Speed: 4141.898548 samples/sec	accuracy=0.730300	lr=0.026890
Epoch[100] Batch [1049]	Speed: 4221.325430 samples/sec	accuracy=0.730388	lr=0.026781
Epoch[100] Batch [1099]	Speed: 4137.632177 samples/sec	accuracy=0.730427	lr=0.026672
Epoch[100] Batch [1149]	Speed: 4130.249001 samples/sec	accuracy=0.730468	lr=0.026563
Epoch[100] Batch [1199]	Speed: 4122.650720 samples/sec	accuracy=0.730499	lr=0.026454
Epoch[100] Batch [1249]	Speed: 4223.764178 samples/sec	accuracy=0.730546	lr=0.026346
[Epoch 100] training: accuracy=0.730551
[Epoch 100] speed: 4130 samples/sec	time cost: 321.519529
[Epoch 100] validation: err-top1=0.247230 err-top5=0.072305
Epoch[101] Batch [49]	Speed: 3909.369810 samples/sec	accuracy=0.731746	lr=0.026235
Epoch[101] Batch [99]	Speed: 4136.908037 samples/sec	accuracy=0.731193	lr=0.026127
Epoch[101] Batch [149]	Speed: 4012.869399 samples/sec	accuracy=0.731225	lr=0.026020
Epoch[101] Batch [199]	Speed: 4106.376376 samples/sec	accuracy=0.731099	lr=0.025912
Epoch[101] Batch [249]	Speed: 4172.306816 samples/sec	accuracy=0.731064	lr=0.025805
Epoch[101] Batch [299]	Speed: 4090.569008 samples/sec	accuracy=0.731208	lr=0.025697
Epoch[101] Batch [349]	Speed: 4092.415717 samples/sec	accuracy=0.731228	lr=0.025590
Epoch[101] Batch [399]	Speed: 4201.404867 samples/sec	accuracy=0.731205	lr=0.025484
Epoch[101] Batch [449]	Speed: 4106.591379 samples/sec	accuracy=0.731297	lr=0.025377
Epoch[101] Batch [499]	Speed: 4242.553014 samples/sec	accuracy=0.731382	lr=0.025271
Epoch[101] Batch [549]	Speed: 4093.572689 samples/sec	accuracy=0.731502	lr=0.025165
Epoch[101] Batch [599]	Speed: 4131.096767 samples/sec	accuracy=0.731604	lr=0.025059
Epoch[101] Batch [649]	Speed: 4236.585788 samples/sec	accuracy=0.731639	lr=0.024953
Epoch[101] Batch [699]	Speed: 4081.226913 samples/sec	accuracy=0.731680	lr=0.024847
Epoch[101] Batch [749]	Speed: 4050.368537 samples/sec	accuracy=0.731700	lr=0.024742
Epoch[101] Batch [799]	Speed: 4179.547829 samples/sec	accuracy=0.731770	lr=0.024637
Epoch[101] Batch [849]	Speed: 4115.499894 samples/sec	accuracy=0.731796	lr=0.024532
Epoch[101] Batch [899]	Speed: 4067.075319 samples/sec	accuracy=0.731836	lr=0.024427
Epoch[101] Batch [949]	Speed: 4242.226829 samples/sec	accuracy=0.731914	lr=0.024323
Epoch[101] Batch [999]	Speed: 4152.655863 samples/sec	accuracy=0.732002	lr=0.024219
Epoch[101] Batch [1049]	Speed: 4146.314379 samples/sec	accuracy=0.732067	lr=0.024115
Epoch[101] Batch [1099]	Speed: 4137.333244 samples/sec	accuracy=0.732154	lr=0.024011
Epoch[101] Batch [1149]	Speed: 4106.923821 samples/sec	accuracy=0.732243	lr=0.023907
Epoch[101] Batch [1199]	Speed: 4210.586552 samples/sec	accuracy=0.732313	lr=0.023804
Epoch[101] Batch [1249]	Speed: 4185.802488 samples/sec	accuracy=0.732383	lr=0.023700
[Epoch 101] training: accuracy=0.732383
[Epoch 101] speed: 4123 samples/sec	time cost: 322.786745
[Epoch 101] validation: err-top1=0.245018 err-top5=0.072186
Epoch[102] Batch [49]	Speed: 4120.723018 samples/sec	accuracy=0.734403	lr=0.023595
Epoch[102] Batch [99]	Speed: 4118.521752 samples/sec	accuracy=0.733798	lr=0.023493
Epoch[102] Batch [149]	Speed: 4140.683284 samples/sec	accuracy=0.733519	lr=0.023390
Epoch[102] Batch [199]	Speed: 4258.068239 samples/sec	accuracy=0.733313	lr=0.023288
Epoch[102] Batch [249]	Speed: 4085.918857 samples/sec	accuracy=0.733118	lr=0.023185
Epoch[102] Batch [299]	Speed: 4230.817525 samples/sec	accuracy=0.733031	lr=0.023083
Epoch[102] Batch [349]	Speed: 4140.633704 samples/sec	accuracy=0.732933	lr=0.022982
Epoch[102] Batch [399]	Speed: 4238.762911 samples/sec	accuracy=0.733008	lr=0.022880
Epoch[102] Batch [449]	Speed: 4095.943985 samples/sec	accuracy=0.733044	lr=0.022779
Epoch[102] Batch [499]	Speed: 4227.431208 samples/sec	accuracy=0.733105	lr=0.022678
Epoch[102] Batch [549]	Speed: 4247.417038 samples/sec	accuracy=0.733160	lr=0.022577
Epoch[102] Batch [599]	Speed: 4180.186644 samples/sec	accuracy=0.733218	lr=0.022476
Epoch[102] Batch [649]	Speed: 4228.840078 samples/sec	accuracy=0.733225	lr=0.022376
Epoch[102] Batch [699]	Speed: 4105.137370 samples/sec	accuracy=0.733235	lr=0.022276
Epoch[102] Batch [749]	Speed: 4054.709461 samples/sec	accuracy=0.733303	lr=0.022175
Epoch[102] Batch [799]	Speed: 4160.678969 samples/sec	accuracy=0.733345	lr=0.022076
Epoch[102] Batch [849]	Speed: 4254.186241 samples/sec	accuracy=0.733397	lr=0.021976
Epoch[102] Batch [899]	Speed: 4171.298312 samples/sec	accuracy=0.733491	lr=0.021877
Epoch[102] Batch [949]	Speed: 4092.146129 samples/sec	accuracy=0.733533	lr=0.021777
Epoch[102] Batch [999]	Speed: 4186.873608 samples/sec	accuracy=0.733600	lr=0.021678
Epoch[102] Batch [1049]	Speed: 4283.820880 samples/sec	accuracy=0.733701	lr=0.021580
Epoch[102] Batch [1099]	Speed: 4099.599097 samples/sec	accuracy=0.733780	lr=0.021481
Epoch[102] Batch [1149]	Speed: 4232.560644 samples/sec	accuracy=0.733828	lr=0.021383
Epoch[102] Batch [1199]	Speed: 4120.339322 samples/sec	accuracy=0.733875	lr=0.021285
Epoch[102] Batch [1249]	Speed: 4363.159021 samples/sec	accuracy=0.733946	lr=0.021187
[Epoch 102] training: accuracy=0.733945
[Epoch 102] speed: 4173 samples/sec	time cost: 317.183032
[Epoch 102] validation: err-top1=0.239138 err-top5=0.070552
Epoch[103] Batch [49]	Speed: 3995.976878 samples/sec	accuracy=0.735753	lr=0.021087
Epoch[103] Batch [99]	Speed: 4052.911379 samples/sec	accuracy=0.735433	lr=0.020990
Epoch[103] Batch [149]	Speed: 4157.663132 samples/sec	accuracy=0.735235	lr=0.020892
Epoch[103] Batch [199]	Speed: 4038.466746 samples/sec	accuracy=0.735253	lr=0.020795
Epoch[103] Batch [249]	Speed: 4071.269134 samples/sec	accuracy=0.735182	lr=0.020698
Epoch[103] Batch [299]	Speed: 4251.066279 samples/sec	accuracy=0.735227	lr=0.020602
Epoch[103] Batch [349]	Speed: 4242.232109 samples/sec	accuracy=0.735269	lr=0.020505
Epoch[103] Batch [399]	Speed: 4279.367802 samples/sec	accuracy=0.735192	lr=0.020409
Epoch[103] Batch [449]	Speed: 4293.288706 samples/sec	accuracy=0.735220	lr=0.020313
Epoch[103] Batch [499]	Speed: 4204.795540 samples/sec	accuracy=0.735150	lr=0.020217
Epoch[103] Batch [549]	Speed: 4140.764322 samples/sec	accuracy=0.735115	lr=0.020122
Epoch[103] Batch [599]	Speed: 4188.081509 samples/sec	accuracy=0.735144	lr=0.020026
Epoch[103] Batch [649]	Speed: 4151.213111 samples/sec	accuracy=0.735165	lr=0.019931
Epoch[103] Batch [699]	Speed: 4208.325945 samples/sec	accuracy=0.735180	lr=0.019836
Epoch[103] Batch [749]	Speed: 4091.563716 samples/sec	accuracy=0.735166	lr=0.019742
Epoch[103] Batch [799]	Speed: 4158.856974 samples/sec	accuracy=0.735221	lr=0.019647
Epoch[103] Batch [849]	Speed: 4257.127225 samples/sec	accuracy=0.735265	lr=0.019553
Epoch[103] Batch [899]	Speed: 4267.703414 samples/sec	accuracy=0.735347	lr=0.019459
Epoch[103] Batch [949]	Speed: 4205.461698 samples/sec	accuracy=0.735444	lr=0.019365
Epoch[103] Batch [999]	Speed: 4064.631767 samples/sec	accuracy=0.735508	lr=0.019271
Epoch[103] Batch [1049]	Speed: 4311.297975 samples/sec	accuracy=0.735580	lr=0.019178
Epoch[103] Batch [1099]	Speed: 4211.701453 samples/sec	accuracy=0.735652	lr=0.019085
Epoch[103] Batch [1149]	Speed: 4064.353674 samples/sec	accuracy=0.735732	lr=0.018992
Epoch[103] Batch [1199]	Speed: 4198.631378 samples/sec	accuracy=0.735816	lr=0.018899
Epoch[103] Batch [1249]	Speed: 4253.222260 samples/sec	accuracy=0.735862	lr=0.018806
[Epoch 103] training: accuracy=0.735861
[Epoch 103] speed: 4169 samples/sec	time cost: 318.045033
[Epoch 103] validation: err-top1=0.236288 err-top5=0.068300
Epoch[104] Batch [49]	Speed: 4115.204229 samples/sec	accuracy=0.736618	lr=0.018712
Epoch[104] Batch [99]	Speed: 4094.063962 samples/sec	accuracy=0.736340	lr=0.018620
Epoch[104] Batch [149]	Speed: 4166.931169 samples/sec	accuracy=0.736317	lr=0.018528
Epoch[104] Batch [199]	Speed: 4169.708421 samples/sec	accuracy=0.736317	lr=0.018436
Epoch[104] Batch [249]	Speed: 4173.725090 samples/sec	accuracy=0.736135	lr=0.018345
Epoch[104] Batch [299]	Speed: 4129.631391 samples/sec	accuracy=0.736114	lr=0.018254
Epoch[104] Batch [349]	Speed: 4177.322369 samples/sec	accuracy=0.736115	lr=0.018163
Epoch[104] Batch [399]	Speed: 4077.053825 samples/sec	accuracy=0.736195	lr=0.018072
Epoch[104] Batch [449]	Speed: 4093.003990 samples/sec	accuracy=0.736253	lr=0.017981
Epoch[104] Batch [499]	Speed: 4130.900487 samples/sec	accuracy=0.736321	lr=0.017891
Epoch[104] Batch [549]	Speed: 4229.962585 samples/sec	accuracy=0.736459	lr=0.017801
Epoch[104] Batch [599]	Speed: 4185.792289 samples/sec	accuracy=0.736569	lr=0.017711
Epoch[104] Batch [649]	Speed: 4116.087879 samples/sec	accuracy=0.736608	lr=0.017621
Epoch[104] Batch [699]	Speed: 4219.649174 samples/sec	accuracy=0.736595	lr=0.017532
Epoch[104] Batch [749]	Speed: 3961.155640 samples/sec	accuracy=0.736597	lr=0.017442
Epoch[104] Batch [799]	Speed: 4325.482363 samples/sec	accuracy=0.736657	lr=0.017353
Epoch[104] Batch [849]	Speed: 4212.256605 samples/sec	accuracy=0.736696	lr=0.017264
Epoch[104] Batch [899]	Speed: 4062.853317 samples/sec	accuracy=0.736737	lr=0.017176
Epoch[104] Batch [949]	Speed: 4199.224310 samples/sec	accuracy=0.736794	lr=0.017087
Epoch[104] Batch [999]	Speed: 4131.939634 samples/sec	accuracy=0.736830	lr=0.016999
Epoch[104] Batch [1049]	Speed: 4239.787145 samples/sec	accuracy=0.736899	lr=0.016911
Epoch[104] Batch [1099]	Speed: 4154.548541 samples/sec	accuracy=0.736981	lr=0.016823
Epoch[104] Batch [1149]	Speed: 4078.535022 samples/sec	accuracy=0.737069	lr=0.016736
Epoch[104] Batch [1199]	Speed: 4157.236070 samples/sec	accuracy=0.737127	lr=0.016648
Epoch[104] Batch [1249]	Speed: 4336.467463 samples/sec	accuracy=0.737215	lr=0.016561
[Epoch 104] training: accuracy=0.737217
[Epoch 104] speed: 4152 samples/sec	time cost: 318.635365
[Epoch 104] validation: err-top1=0.235372 err-top5=0.067708
Epoch[105] Batch [49]	Speed: 3961.369223 samples/sec	accuracy=0.738569	lr=0.016473
Epoch[105] Batch [99]	Speed: 4162.324826 samples/sec	accuracy=0.738168	lr=0.016386
Epoch[105] Batch [149]	Speed: 4270.163624 samples/sec	accuracy=0.737981	lr=0.016299
Epoch[105] Batch [199]	Speed: 4116.495482 samples/sec	accuracy=0.737825	lr=0.016213
Epoch[105] Batch [249]	Speed: 4167.728301 samples/sec	accuracy=0.737815	lr=0.016127
Epoch[105] Batch [299]	Speed: 4104.597932 samples/sec	accuracy=0.737690	lr=0.016041
Epoch[105] Batch [349]	Speed: 4117.813207 samples/sec	accuracy=0.737698	lr=0.015956
Epoch[105] Batch [399]	Speed: 4162.847024 samples/sec	accuracy=0.737718	lr=0.015870
Epoch[105] Batch [449]	Speed: 4064.073926 samples/sec	accuracy=0.737706	lr=0.015785
Epoch[105] Batch [499]	Speed: 4168.235190 samples/sec	accuracy=0.737654	lr=0.015700
Epoch[105] Batch [549]	Speed: 4078.221565 samples/sec	accuracy=0.737650	lr=0.015616
Epoch[105] Batch [599]	Speed: 4039.296699 samples/sec	accuracy=0.737692	lr=0.015531
Epoch[105] Batch [649]	Speed: 4104.436089 samples/sec	accuracy=0.737817	lr=0.015447
Epoch[105] Batch [699]	Speed: 4255.655845 samples/sec	accuracy=0.737933	lr=0.015363
Epoch[105] Batch [749]	Speed: 4078.677709 samples/sec	accuracy=0.738013	lr=0.015279
Epoch[105] Batch [799]	Speed: 4192.148073 samples/sec	accuracy=0.738086	lr=0.015195
Epoch[105] Batch [849]	Speed: 4188.634781 samples/sec	accuracy=0.738141	lr=0.015112
Epoch[105] Batch [899]	Speed: 4137.858917 samples/sec	accuracy=0.738208	lr=0.015029
Epoch[105] Batch [949]	Speed: 4156.867189 samples/sec	accuracy=0.738289	lr=0.014946
Epoch[105] Batch [999]	Speed: 4288.794349 samples/sec	accuracy=0.738390	lr=0.014863
Epoch[105] Batch [1049]	Speed: 4094.520378 samples/sec	accuracy=0.738459	lr=0.014781
Epoch[105] Batch [1099]	Speed: 4185.028930 samples/sec	accuracy=0.738542	lr=0.014699
Epoch[105] Batch [1149]	Speed: 4091.205774 samples/sec	accuracy=0.738607	lr=0.014616
Epoch[105] Batch [1199]	Speed: 4233.287201 samples/sec	accuracy=0.738661	lr=0.014535
Epoch[105] Batch [1249]	Speed: 4309.464680 samples/sec	accuracy=0.738734	lr=0.014453
[Epoch 105] training: accuracy=0.738736
[Epoch 105] speed: 4143 samples/sec	time cost: 319.557989
[Epoch 105] validation: err-top1=0.233837 err-top5=0.066785
Epoch[106] Batch [49]	Speed: 4115.658666 samples/sec	accuracy=0.740383	lr=0.014370
Epoch[106] Batch [99]	Speed: 4140.764322 samples/sec	accuracy=0.739843	lr=0.014289
Epoch[106] Batch [149]	Speed: 4220.218947 samples/sec	accuracy=0.739680	lr=0.014208
Epoch[106] Batch [199]	Speed: 4230.921385 samples/sec	accuracy=0.739604	lr=0.014127
Epoch[106] Batch [249]	Speed: 4090.860208 samples/sec	accuracy=0.739625	lr=0.014047
Epoch[106] Batch [299]	Speed: 4042.062099 samples/sec	accuracy=0.739636	lr=0.013966
Epoch[106] Batch [349]	Speed: 4246.381389 samples/sec	accuracy=0.739699	lr=0.013886
Epoch[106] Batch [399]	Speed: 4230.215556 samples/sec	accuracy=0.739720	lr=0.013806
Epoch[106] Batch [449]	Speed: 4345.401327 samples/sec	accuracy=0.739754	lr=0.013727
Epoch[106] Batch [499]	Speed: 4111.802451 samples/sec	accuracy=0.739747	lr=0.013647
Epoch[106] Batch [549]	Speed: 4147.921969 samples/sec	accuracy=0.739822	lr=0.013568
Epoch[106] Batch [599]	Speed: 4167.603742 samples/sec	accuracy=0.739844	lr=0.013489
Epoch[106] Batch [649]	Speed: 4271.672754 samples/sec	accuracy=0.739857	lr=0.013411
Epoch[106] Batch [699]	Speed: 4138.426513 samples/sec	accuracy=0.739901	lr=0.013332
Epoch[106] Batch [749]	Speed: 4177.806154 samples/sec	accuracy=0.739896	lr=0.013254
Epoch[106] Batch [799]	Speed: 4234.441460 samples/sec	accuracy=0.739917	lr=0.013176
Epoch[106] Batch [849]	Speed: 4254.226610 samples/sec	accuracy=0.739945	lr=0.013098
Epoch[106] Batch [899]	Speed: 4293.580298 samples/sec	accuracy=0.739994	lr=0.013020
Epoch[106] Batch [949]	Speed: 4218.846729 samples/sec	accuracy=0.740027	lr=0.012943
Epoch[106] Batch [999]	Speed: 4313.093156 samples/sec	accuracy=0.740030	lr=0.012866
Epoch[106] Batch [1049]	Speed: 4188.112710 samples/sec	accuracy=0.740063	lr=0.012789
Epoch[106] Batch [1099]	Speed: 4142.509843 samples/sec	accuracy=0.740129	lr=0.012712
Epoch[106] Batch [1149]	Speed: 4203.020260 samples/sec	accuracy=0.740173	lr=0.012636
Epoch[106] Batch [1199]	Speed: 4190.224424 samples/sec	accuracy=0.740230	lr=0.012559
Epoch[106] Batch [1249]	Speed: 4145.791999 samples/sec	accuracy=0.740274	lr=0.012483
[Epoch 106] training: accuracy=0.740274
[Epoch 106] speed: 4189 samples/sec	time cost: 316.269121
[Epoch 106] validation: err-top1=0.231864 err-top5=0.065529
Epoch[107] Batch [49]	Speed: 4075.549570 samples/sec	accuracy=0.741342	lr=0.012406
Epoch[107] Batch [99]	Speed: 3959.706253 samples/sec	accuracy=0.741702	lr=0.012330
Epoch[107] Batch [149]	Speed: 4147.735542 samples/sec	accuracy=0.741356	lr=0.012255
Epoch[107] Batch [199]	Speed: 4152.403973 samples/sec	accuracy=0.741302	lr=0.012180
Epoch[107] Batch [249]	Speed: 4067.221981 samples/sec	accuracy=0.741331	lr=0.012105
Epoch[107] Batch [299]	Speed: 4096.430280 samples/sec	accuracy=0.741301	lr=0.012030
Epoch[107] Batch [349]	Speed: 4098.003088 samples/sec	accuracy=0.741386	lr=0.011956
Epoch[107] Batch [399]	Speed: 4309.217015 samples/sec	accuracy=0.741277	lr=0.011881
Epoch[107] Batch [449]	Speed: 4094.578384 samples/sec	accuracy=0.741165	lr=0.011807
Epoch[107] Batch [499]	Speed: 4124.202228 samples/sec	accuracy=0.741172	lr=0.011734
Epoch[107] Batch [549]	Speed: 4207.795079 samples/sec	accuracy=0.741201	lr=0.011660
Epoch[107] Batch [599]	Speed: 4182.749563 samples/sec	accuracy=0.741216	lr=0.011587
Epoch[107] Batch [649]	Speed: 4232.156674 samples/sec	accuracy=0.741253	lr=0.011513
Epoch[107] Batch [699]	Speed: 4102.537762 samples/sec	accuracy=0.741290	lr=0.011441
Epoch[107] Batch [749]	Speed: 4101.035401 samples/sec	accuracy=0.741296	lr=0.011368
Epoch[107] Batch [799]	Speed: 4132.376544 samples/sec	accuracy=0.741293	lr=0.011295
Epoch[107] Batch [849]	Speed: 4221.258633 samples/sec	accuracy=0.741295	lr=0.011223
Epoch[107] Batch [899]	Speed: 4146.351686 samples/sec	accuracy=0.741288	lr=0.011151
Epoch[107] Batch [949]	Speed: 4149.436756 samples/sec	accuracy=0.741309	lr=0.011079
Epoch[107] Batch [999]	Speed: 4250.107496 samples/sec	accuracy=0.741364	lr=0.011008
Epoch[107] Batch [1049]	Speed: 4247.741585 samples/sec	accuracy=0.741452	lr=0.010936
Epoch[107] Batch [1099]	Speed: 4222.321080 samples/sec	accuracy=0.741522	lr=0.010865
Epoch[107] Batch [1149]	Speed: 4192.753664 samples/sec	accuracy=0.741592	lr=0.010794
Epoch[107] Batch [1199]	Speed: 4278.266825 samples/sec	accuracy=0.741615	lr=0.010724
Epoch[107] Batch [1249]	Speed: 4233.087682 samples/sec	accuracy=0.741668	lr=0.010653
[Epoch 107] training: accuracy=0.741674
[Epoch 107] speed: 4156 samples/sec	time cost: 318.370951
[Epoch 107] validation: err-top1=0.229771 err-top5=0.065788
Epoch[108] Batch [49]	Speed: 3949.705453 samples/sec	accuracy=0.742944	lr=0.010582
Epoch[108] Batch [99]	Speed: 4199.818559 samples/sec	accuracy=0.742753	lr=0.010512
Epoch[108] Batch [149]	Speed: 4228.044206 samples/sec	accuracy=0.742453	lr=0.010442
Epoch[108] Batch [199]	Speed: 4156.141368 samples/sec	accuracy=0.742514	lr=0.010373
Epoch[108] Batch [249]	Speed: 4139.351282 samples/sec	accuracy=0.742466	lr=0.010303
Epoch[108] Batch [299]	Speed: 4216.325820 samples/sec	accuracy=0.742425	lr=0.010234
Epoch[108] Batch [349]	Speed: 4008.615384 samples/sec	accuracy=0.742431	lr=0.010165
Epoch[108] Batch [399]	Speed: 4119.418760 samples/sec	accuracy=0.742372	lr=0.010097
Epoch[108] Batch [449]	Speed: 4060.735926 samples/sec	accuracy=0.742352	lr=0.010028
Epoch[108] Batch [499]	Speed: 4227.955137 samples/sec	accuracy=0.742329	lr=0.009960
Epoch[108] Batch [549]	Speed: 4240.689942 samples/sec	accuracy=0.742389	lr=0.009892
Epoch[108] Batch [599]	Speed: 4307.813109 samples/sec	accuracy=0.742452	lr=0.009824
Epoch[108] Batch [649]	Speed: 4145.615047 samples/sec	accuracy=0.742468	lr=0.009757
Epoch[108] Batch [699]	Speed: 4194.733632 samples/sec	accuracy=0.742490	lr=0.009690
Epoch[108] Batch [749]	Speed: 3772.310970 samples/sec	accuracy=0.742506	lr=0.009623
Epoch[108] Batch [799]	Speed: 4107.025929 samples/sec	accuracy=0.742517	lr=0.009556
Epoch[108] Batch [849]	Speed: 3716.574048 samples/sec	accuracy=0.742504	lr=0.009489
Epoch[108] Batch [899]	Speed: 3629.166629 samples/sec	accuracy=0.742549	lr=0.009423
Epoch[108] Batch [949]	Speed: 4222.927032 samples/sec	accuracy=0.742581	lr=0.009357
Epoch[108] Batch [999]	Speed: 3388.133554 samples/sec	accuracy=0.742597	lr=0.009291
Epoch[108] Batch [1049]	Speed: 4047.058419 samples/sec	accuracy=0.742654	lr=0.009225
Epoch[108] Batch [1099]	Speed: 3586.536525 samples/sec	accuracy=0.742718	lr=0.009160
Epoch[108] Batch [1149]	Speed: 3647.393947 samples/sec	accuracy=0.742788	lr=0.009095
Epoch[108] Batch [1199]	Speed: 3377.818981 samples/sec	accuracy=0.742837	lr=0.009030
Epoch[108] Batch [1249]	Speed: 3133.430814 samples/sec	accuracy=0.742873	lr=0.008965
[Epoch 108] training: accuracy=0.742875
[Epoch 108] speed: 3921 samples/sec	time cost: 336.513051
[Epoch 108] validation: err-top1=0.228276 err-top5=0.064852
Epoch[109] Batch [49]	Speed: 4138.110082 samples/sec	accuracy=0.743364	lr=0.008899
Epoch[109] Batch [99]	Speed: 2983.736695 samples/sec	accuracy=0.743421	lr=0.008835
Epoch[109] Batch [149]	Speed: 3793.574165 samples/sec	accuracy=0.743478	lr=0.008771
Epoch[109] Batch [199]	Speed: 4199.146797 samples/sec	accuracy=0.743458	lr=0.008707
Epoch[109] Batch [249]	Speed: 4096.433874 samples/sec	accuracy=0.743482	lr=0.008643
Epoch[109] Batch [299]	Speed: 4273.642334 samples/sec	accuracy=0.743591	lr=0.008580
Epoch[109] Batch [349]	Speed: 3796.245424 samples/sec	accuracy=0.743599	lr=0.008517
Epoch[109] Batch [399]	Speed: 4253.401020 samples/sec	accuracy=0.743641	lr=0.008454
Epoch[109] Batch [449]	Speed: 4186.133844 samples/sec	accuracy=0.743647	lr=0.008391
Epoch[109] Batch [499]	Speed: 4142.534935 samples/sec	accuracy=0.743654	lr=0.008329
Epoch[109] Batch [549]	Speed: 4205.289415 samples/sec	accuracy=0.743639	lr=0.008266
Epoch[109] Batch [599]	Speed: 3651.374742 samples/sec	accuracy=0.743677	lr=0.008204
Epoch[109] Batch [649]	Speed: 4168.799254 samples/sec	accuracy=0.743674	lr=0.008143
Epoch[109] Batch [699]	Speed: 4204.073627 samples/sec	accuracy=0.743702	lr=0.008081
Epoch[109] Batch [749]	Speed: 1375.708119 samples/sec	accuracy=0.743740	lr=0.008020
Epoch[109] Batch [799]	Speed: 3895.236421 samples/sec	accuracy=0.743782	lr=0.007959
Epoch[109] Batch [849]	Speed: 294.756204 samples/sec	accuracy=0.743803	lr=0.007898
Epoch[109] Batch [899]	Speed: 4231.571750 samples/sec	accuracy=0.743813	lr=0.007837
Epoch[109] Batch [949]	Speed: 4058.599878 samples/sec	accuracy=0.743856	lr=0.007777
Epoch[109] Batch [999]	Speed: 2385.940852 samples/sec	accuracy=0.743878	lr=0.007716
Epoch[109] Batch [1049]	Speed: 4153.410108 samples/sec	accuracy=0.743906	lr=0.007656
Epoch[109] Batch [1099]	Speed: 4097.783354 samples/sec	accuracy=0.743940	lr=0.007597
Epoch[109] Batch [1149]	Speed: 4189.433783 samples/sec	accuracy=0.743942	lr=0.007537
Epoch[109] Batch [1199]	Speed: 4152.602302 samples/sec	accuracy=0.743954	lr=0.007478
Epoch[109] Batch [1249]	Speed: 4138.473646 samples/sec	accuracy=0.743984	lr=0.007419
[Epoch 109] training: accuracy=0.743984
[Epoch 109] speed: 2496 samples/sec	time cost: 535.594312
[Epoch 109] validation: err-top1=0.227220 err-top5=0.063636
Epoch[110] Batch [49]	Speed: 1575.442737 samples/sec	accuracy=0.744139	lr=0.007359
Epoch[110] Batch [99]	Speed: 4111.942121 samples/sec	accuracy=0.744108	lr=0.007300
Epoch[110] Batch [149]	Speed: 4155.220901 samples/sec	accuracy=0.744192	lr=0.007242
Epoch[110] Batch [199]	Speed: 4159.892669 samples/sec	accuracy=0.744115	lr=0.007184
Epoch[110] Batch [249]	Speed: 3631.899595 samples/sec	accuracy=0.744066	lr=0.007126
Epoch[110] Batch [299]	Speed: 757.308731 samples/sec	accuracy=0.743995	lr=0.007068
Epoch[110] Batch [349]	Speed: 1137.155855 samples/sec	accuracy=0.743944	lr=0.007011
Epoch[110] Batch [399]	Speed: 4237.166080 samples/sec	accuracy=0.743987	lr=0.006954
Epoch[110] Batch [449]	Speed: 1209.358500 samples/sec	accuracy=0.744042	lr=0.006897
Epoch[110] Batch [499]	Speed: 4288.397043 samples/sec	accuracy=0.744054	lr=0.006840
Epoch[110] Batch [549]	Speed: 4317.139433 samples/sec	accuracy=0.744139	lr=0.006784
Epoch[110] Batch [599]	Speed: 4259.027746 samples/sec	accuracy=0.744186	lr=0.006727
Epoch[110] Batch [649]	Speed: 4192.747115 samples/sec	accuracy=0.744223	lr=0.006671
Epoch[110] Batch [699]	Speed: 4211.149998 samples/sec	accuracy=0.744239	lr=0.006615
Epoch[110] Batch [749]	Speed: 4277.119560 samples/sec	accuracy=0.744291	lr=0.006560
Epoch[110] Batch [799]	Speed: 1103.615719 samples/sec	accuracy=0.744285	lr=0.006504
Epoch[110] Batch [849]	Speed: 1700.122211 samples/sec	accuracy=0.744307	lr=0.006449
Epoch[110] Batch [899]	Speed: 4092.241031 samples/sec	accuracy=0.744320	lr=0.006394
Epoch[110] Batch [949]	Speed: 1102.413476 samples/sec	accuracy=0.744335	lr=0.006340
Epoch[110] Batch [999]	Speed: 1777.469126 samples/sec	accuracy=0.744367	lr=0.006285
Epoch[110] Batch [1049]	Speed: 4136.369301 samples/sec	accuracy=0.744398	lr=0.006231
Epoch[110] Batch [1099]	Speed: 4128.241890 samples/sec	accuracy=0.744426	lr=0.006177
Epoch[110] Batch [1149]	Speed: 4255.244841 samples/sec	accuracy=0.744450	lr=0.006123
Epoch[110] Batch [1199]	Speed: 1504.741754 samples/sec	accuracy=0.744487	lr=0.006070
Epoch[110] Batch [1249]	Speed: 4175.905791 samples/sec	accuracy=0.744521	lr=0.006017
[Epoch 110] training: accuracy=0.744522
[Epoch 110] speed: 2243 samples/sec	time cost: 583.285193
[Epoch 110] validation: err-top1=0.225952 err-top5=0.063558
Epoch[111] Batch [49]	Speed: 4179.848744 samples/sec	accuracy=0.745068	lr=0.005963
Epoch[111] Batch [99]	Speed: 4114.386069 samples/sec	accuracy=0.745385	lr=0.005910
Epoch[111] Batch [149]	Speed: 1517.365737 samples/sec	accuracy=0.745306	lr=0.005857
Epoch[111] Batch [199]	Speed: 2744.479204 samples/sec	accuracy=0.745262	lr=0.005805
Epoch[111] Batch [249]	Speed: 4350.205862 samples/sec	accuracy=0.745308	lr=0.005753
Epoch[111] Batch [299]	Speed: 3481.252342 samples/sec	accuracy=0.745205	lr=0.005701
Epoch[111] Batch [349]	Speed: 4217.699546 samples/sec	accuracy=0.745174	lr=0.005649
Epoch[111] Batch [399]	Speed: 3929.262864 samples/sec	accuracy=0.745176	lr=0.005598
Epoch[111] Batch [449]	Speed: 4221.106128 samples/sec	accuracy=0.745125	lr=0.005547
Epoch[111] Batch [499]	Speed: 2079.913277 samples/sec	accuracy=0.745184	lr=0.005496
Epoch[111] Batch [549]	Speed: 4129.079543 samples/sec	accuracy=0.745248	lr=0.005445
Epoch[111] Batch [599]	Speed: 4222.869568 samples/sec	accuracy=0.745272	lr=0.005394
Epoch[111] Batch [649]	Speed: 4216.069624 samples/sec	accuracy=0.745266	lr=0.005344
Epoch[111] Batch [699]	Speed: 1989.719615 samples/sec	accuracy=0.745277	lr=0.005294
Epoch[111] Batch [749]	Speed: 1477.798232 samples/sec	accuracy=0.745300	lr=0.005244
Epoch[111] Batch [799]	Speed: 4212.926286 samples/sec	accuracy=0.745309	lr=0.005195
Epoch[111] Batch [849]	Speed: 4095.696976 samples/sec	accuracy=0.745350	lr=0.005145
Epoch[111] Batch [899]	Speed: 1475.162274 samples/sec	accuracy=0.745395	lr=0.005096
Epoch[111] Batch [949]	Speed: 4158.798260 samples/sec	accuracy=0.745398	lr=0.005047
Epoch[111] Batch [999]	Speed: 978.429905 samples/sec	accuracy=0.745390	lr=0.004999
Epoch[111] Batch [1049]	Speed: 3859.648988 samples/sec	accuracy=0.745423	lr=0.004950
Epoch[111] Batch [1099]	Speed: 4304.366051 samples/sec	accuracy=0.745439	lr=0.004902
Epoch[111] Batch [1149]	Speed: 4146.304292 samples/sec	accuracy=0.745484	lr=0.004854
Epoch[111] Batch [1199]	Speed: 4170.652004 samples/sec	accuracy=0.745513	lr=0.004807
Epoch[111] Batch [1249]	Speed: 4334.552248 samples/sec	accuracy=0.745540	lr=0.004759
[Epoch 111] training: accuracy=0.745542
[Epoch 111] speed: 2854 samples/sec	time cost: 469.977688
[Epoch 111] validation: err-top1=0.224550 err-top5=0.062839
Epoch[112] Batch [49]	Speed: 2830.602634 samples/sec	accuracy=0.745916	lr=0.004711
Epoch[112] Batch [99]	Speed: 2725.475893 samples/sec	accuracy=0.746086	lr=0.004664
Epoch[112] Batch [149]	Speed: 4226.190026 samples/sec	accuracy=0.746126	lr=0.004617
Epoch[112] Batch [199]	Speed: 4020.112825 samples/sec	accuracy=0.746200	lr=0.004571
Epoch[112] Batch [249]	Speed: 4236.820744 samples/sec	accuracy=0.746239	lr=0.004524
Epoch[112] Batch [299]	Speed: 3657.025026 samples/sec	accuracy=0.746265	lr=0.004478
Epoch[112] Batch [349]	Speed: 4024.798058 samples/sec	accuracy=0.746282	lr=0.004432
Epoch[112] Batch [399]	Speed: 4274.763056 samples/sec	accuracy=0.746281	lr=0.004387
Epoch[112] Batch [449]	Speed: 1369.856637 samples/sec	accuracy=0.746255	lr=0.004341
Epoch[112] Batch [499]	Speed: 4200.516580 samples/sec	accuracy=0.746237	lr=0.004296
Epoch[112] Batch [549]	Speed: 4080.012568 samples/sec	accuracy=0.746234	lr=0.004251
Epoch[112] Batch [599]	Speed: 4103.696858 samples/sec	accuracy=0.746214	lr=0.004207
Epoch[112] Batch [649]	Speed: 4318.036926 samples/sec	accuracy=0.746242	lr=0.004162
Epoch[112] Batch [699]	Speed: 2109.695002 samples/sec	accuracy=0.746261	lr=0.004118
Epoch[112] Batch [749]	Speed: 2805.875961 samples/sec	accuracy=0.746320	lr=0.004074
Epoch[112] Batch [799]	Speed: 2255.554291 samples/sec	accuracy=0.746366	lr=0.004030
Epoch[112] Batch [849]	Speed: 3238.298999 samples/sec	accuracy=0.746407	lr=0.003987
Epoch[112] Batch [899]	Speed: 2505.042725 samples/sec	accuracy=0.746422	lr=0.003944
Epoch[112] Batch [949]	Speed: 2896.046861 samples/sec	accuracy=0.746414	lr=0.003901
Epoch[112] Batch [999]	Speed: 4068.535168 samples/sec	accuracy=0.746415	lr=0.003858
Epoch[112] Batch [1049]	Speed: 4103.013238 samples/sec	accuracy=0.746420	lr=0.003815
Epoch[112] Batch [1099]	Speed: 2892.092634 samples/sec	accuracy=0.746432	lr=0.003773
Epoch[112] Batch [1149]	Speed: 4229.352779 samples/sec	accuracy=0.746471	lr=0.003731
Epoch[112] Batch [1199]	Speed: 4321.450093 samples/sec	accuracy=0.746515	lr=0.003689
Epoch[112] Batch [1249]	Speed: 4238.253447 samples/sec	accuracy=0.746551	lr=0.003647
[Epoch 112] training: accuracy=0.746551
[Epoch 112] speed: 3222 samples/sec	time cost: 407.191327
[Epoch 112] validation: err-top1=0.222078 err-top5=0.062121
Epoch[113] Batch [49]	Speed: 2164.471066 samples/sec	accuracy=0.747085	lr=0.003605
Epoch[113] Batch [99]	Speed: 2266.194460 samples/sec	accuracy=0.747139	lr=0.003564
Epoch[113] Batch [149]	Speed: 4245.290855 samples/sec	accuracy=0.746927	lr=0.003523
Epoch[113] Batch [199]	Speed: 3832.768800 samples/sec	accuracy=0.746915	lr=0.003482
Epoch[113] Batch [249]	Speed: 3171.579689 samples/sec	accuracy=0.746864	lr=0.003442
Epoch[113] Batch [299]	Speed: 4289.489618 samples/sec	accuracy=0.746838	lr=0.003402
Epoch[113] Batch [349]	Speed: 2930.104073 samples/sec	accuracy=0.746794	lr=0.003362
Epoch[113] Batch [399]	Speed: 3049.319015 samples/sec	accuracy=0.746723	lr=0.003322
Epoch[113] Batch [449]	Speed: 2872.492246 samples/sec	accuracy=0.746729	lr=0.003282
Epoch[113] Batch [499]	Speed: 4206.334528 samples/sec	accuracy=0.746656	lr=0.003243
Epoch[113] Batch [549]	Speed: 1754.733365 samples/sec	accuracy=0.746668	lr=0.003204
Epoch[113] Batch [599]	Speed: 4275.814044 samples/sec	accuracy=0.746662	lr=0.003165
Epoch[113] Batch [649]	Speed: 2000.240504 samples/sec	accuracy=0.746671	lr=0.003127
Epoch[113] Batch [699]	Speed: 2142.772420 samples/sec	accuracy=0.746669	lr=0.003088
Epoch[113] Batch [749]	Speed: 3132.945018 samples/sec	accuracy=0.746634	lr=0.003050
Epoch[113] Batch [799]	Speed: 4193.076788 samples/sec	accuracy=0.746660	lr=0.003012
Epoch[113] Batch [849]	Speed: 2800.828410 samples/sec	accuracy=0.746689	lr=0.002975
Epoch[113] Batch [899]	Speed: 3238.865548 samples/sec	accuracy=0.746726	lr=0.002937
Epoch[113] Batch [949]	Speed: 2910.171145 samples/sec	accuracy=0.746745	lr=0.002900
Epoch[113] Batch [999]	Speed: 3729.682775 samples/sec	accuracy=0.746754	lr=0.002863
Epoch[113] Batch [1049]	Speed: 3346.652433 samples/sec	accuracy=0.746770	lr=0.002826
Epoch[113] Batch [1099]	Speed: 4152.128511 samples/sec	accuracy=0.746799	lr=0.002790
Epoch[113] Batch [1149]	Speed: 4087.240180 samples/sec	accuracy=0.746840	lr=0.002754
Epoch[113] Batch [1199]	Speed: 4158.556335 samples/sec	accuracy=0.746866	lr=0.002718
Epoch[113] Batch [1249]	Speed: 4247.354957 samples/sec	accuracy=0.746895	lr=0.002682
[Epoch 113] training: accuracy=0.746895
[Epoch 113] speed: 3096 samples/sec	time cost: 426.775419
[Epoch 113] validation: err-top1=0.221700 err-top5=0.061922
Epoch[114] Batch [49]	Speed: 4208.503425 samples/sec	accuracy=0.747882	lr=0.002646
Epoch[114] Batch [99]	Speed: 2609.886949 samples/sec	accuracy=0.747658	lr=0.002611
Epoch[114] Batch [149]	Speed: 4313.427818 samples/sec	accuracy=0.747495	lr=0.002575
Epoch[114] Batch [199]	Speed: 4183.430349 samples/sec	accuracy=0.747499	lr=0.002541
Epoch[114] Batch [249]	Speed: 2938.953010 samples/sec	accuracy=0.747391	lr=0.002506
Epoch[114] Batch [299]	Speed: 3275.175056 samples/sec	accuracy=0.747331	lr=0.002472
Epoch[114] Batch [349]	Speed: 4242.869106 samples/sec	accuracy=0.747320	lr=0.002438
Epoch[114] Batch [399]	Speed: 4226.705578 samples/sec	accuracy=0.747343	lr=0.002404
Epoch[114] Batch [449]	Speed: 4297.618067 samples/sec	accuracy=0.747331	lr=0.002370
Epoch[114] Batch [499]	Speed: 3971.625214 samples/sec	accuracy=0.747307	lr=0.002337
Epoch[114] Batch [549]	Speed: 4227.226915 samples/sec	accuracy=0.747326	lr=0.002304
Epoch[114] Batch [599]	Speed: 3416.713141 samples/sec	accuracy=0.747296	lr=0.002271
Epoch[114] Batch [649]	Speed: 3963.198589 samples/sec	accuracy=0.747321	lr=0.002238
Epoch[114] Batch [699]	Speed: 3434.151374 samples/sec	accuracy=0.747325	lr=0.002205
Epoch[114] Batch [749]	Speed: 4085.966513 samples/sec	accuracy=0.747322	lr=0.002173
Epoch[114] Batch [799]	Speed: 4181.824765 samples/sec	accuracy=0.747302	lr=0.002141
Epoch[114] Batch [849]	Speed: 4226.956829 samples/sec	accuracy=0.747316	lr=0.002110
Epoch[114] Batch [899]	Speed: 4234.881443 samples/sec	accuracy=0.747340	lr=0.002078
Epoch[114] Batch [949]	Speed: 3833.701746 samples/sec	accuracy=0.747368	lr=0.002047
Epoch[114] Batch [999]	Speed: 3441.700055 samples/sec	accuracy=0.747367	lr=0.002016
Epoch[114] Batch [1049]	Speed: 3427.909529 samples/sec	accuracy=0.747393	lr=0.001985
Epoch[114] Batch [1099]	Speed: 3411.938775 samples/sec	accuracy=0.747388	lr=0.001954
Epoch[114] Batch [1149]	Speed: 2660.727565 samples/sec	accuracy=0.747405	lr=0.001924
Epoch[114] Batch [1199]	Speed: 3404.958047 samples/sec	accuracy=0.747406	lr=0.001894
Epoch[114] Batch [1249]	Speed: 3393.540357 samples/sec	accuracy=0.747417	lr=0.001864
[Epoch 114] training: accuracy=0.747416
[Epoch 114] speed: 3597 samples/sec	time cost: 366.065181
[Epoch 114] validation: err-top1=0.220065 err-top5=0.062500
Epoch[115] Batch [49]	Speed: 4177.117609 samples/sec	accuracy=0.747695	lr=0.001834
Epoch[115] Batch [99]	Speed: 4241.712678 samples/sec	accuracy=0.747803	lr=0.001804
Epoch[115] Batch [149]	Speed: 3412.352658 samples/sec	accuracy=0.747807	lr=0.001775
Epoch[115] Batch [199]	Speed: 3137.277785 samples/sec	accuracy=0.747730	lr=0.001746
Epoch[115] Batch [249]	Speed: 4018.798215 samples/sec	accuracy=0.747657	lr=0.001718
Epoch[115] Batch [299]	Speed: 4252.872871 samples/sec	accuracy=0.747623	lr=0.001689
Epoch[115] Batch [349]	Speed: 3442.278770 samples/sec	accuracy=0.747589	lr=0.001661
Epoch[115] Batch [399]	Speed: 4167.176738 samples/sec	accuracy=0.747576	lr=0.001633
Epoch[115] Batch [449]	Speed: 4260.316692 samples/sec	accuracy=0.747602	lr=0.001605
Epoch[115] Batch [499]	Speed: 4104.719774 samples/sec	accuracy=0.747654	lr=0.001578
Epoch[115] Batch [549]	Speed: 3471.772028 samples/sec	accuracy=0.747666	lr=0.001551
Epoch[115] Batch [599]	Speed: 2471.909311 samples/sec	accuracy=0.747655	lr=0.001524
Epoch[115] Batch [649]	Speed: 3551.079493 samples/sec	accuracy=0.747640	lr=0.001497
Epoch[115] Batch [699]	Speed: 4249.631631 samples/sec	accuracy=0.747649	lr=0.001470
Epoch[115] Batch [749]	Speed: 4087.126997 samples/sec	accuracy=0.747647	lr=0.001444
Epoch[115] Batch [799]	Speed: 4150.021327 samples/sec	accuracy=0.747650	lr=0.001418
Epoch[115] Batch [849]	Speed: 4185.718699 samples/sec	accuracy=0.747666	lr=0.001392
Epoch[115] Batch [899]	Speed: 4248.678118 samples/sec	accuracy=0.747689	lr=0.001366
Epoch[115] Batch [949]	Speed: 4152.712717 samples/sec	accuracy=0.747695	lr=0.001341
Epoch[115] Batch [999]	Speed: 3350.110923 samples/sec	accuracy=0.747705	lr=0.001316
Epoch[115] Batch [1049]	Speed: 4128.066830 samples/sec	accuracy=0.747728	lr=0.001291
Epoch[115] Batch [1099]	Speed: 3438.359563 samples/sec	accuracy=0.747740	lr=0.001266
Epoch[115] Batch [1149]	Speed: 2933.218657 samples/sec	accuracy=0.747751	lr=0.001242
Epoch[115] Batch [1199]	Speed: 4289.903665 samples/sec	accuracy=0.747758	lr=0.001218
Epoch[115] Batch [1249]	Speed: 4081.949613 samples/sec	accuracy=0.747774	lr=0.001194
[Epoch 115] training: accuracy=0.747775
[Epoch 115] speed: 3761 samples/sec	time cost: 351.365132
[Epoch 115] validation: err-top1=0.220026 err-top5=0.061663
Epoch[116] Batch [49]	Speed: 2782.189560 samples/sec	accuracy=0.747944	lr=0.001170
Epoch[116] Batch [99]	Speed: 3336.035812 samples/sec	accuracy=0.747914	lr=0.001146
Epoch[116] Batch [149]	Speed: 3183.247562 samples/sec	accuracy=0.748000	lr=0.001123
Epoch[116] Batch [199]	Speed: 3391.385849 samples/sec	accuracy=0.747994	lr=0.001100
Epoch[116] Batch [249]	Speed: 4161.106336 samples/sec	accuracy=0.748027	lr=0.001077
Epoch[116] Batch [299]	Speed: 4140.025756 samples/sec	accuracy=0.748014	lr=0.001055
Epoch[116] Batch [349]	Speed: 4067.607404 samples/sec	accuracy=0.747997	lr=0.001032
Epoch[116] Batch [399]	Speed: 4120.049206 samples/sec	accuracy=0.748006	lr=0.001010
Epoch[116] Batch [449]	Speed: 4345.919904 samples/sec	accuracy=0.748038	lr=0.000989
Epoch[116] Batch [499]	Speed: 3589.317341 samples/sec	accuracy=0.748068	lr=0.000967
Epoch[116] Batch [549]	Speed: 3438.304622 samples/sec	accuracy=0.748119	lr=0.000946
Epoch[116] Batch [599]	Speed: 4085.212159 samples/sec	accuracy=0.748116	lr=0.000925
Epoch[116] Batch [649]	Speed: 2062.974635 samples/sec	accuracy=0.748098	lr=0.000904
Epoch[116] Batch [699]	Speed: 4212.407397 samples/sec	accuracy=0.748070	lr=0.000883
Epoch[116] Batch [749]	Speed: 4279.329258 samples/sec	accuracy=0.748069	lr=0.000863
Epoch[116] Batch [799]	Speed: 2883.107098 samples/sec	accuracy=0.748051	lr=0.000843
Epoch[116] Batch [849]	Speed: 4216.567890 samples/sec	accuracy=0.748028	lr=0.000823
Epoch[116] Batch [899]	Speed: 4036.563685 samples/sec	accuracy=0.748013	lr=0.000803
Epoch[116] Batch [949]	Speed: 3484.991497 samples/sec	accuracy=0.748001	lr=0.000784
Epoch[116] Batch [999]	Speed: 4120.100981 samples/sec	accuracy=0.748009	lr=0.000764
Epoch[116] Batch [1049]	Speed: 2805.180816 samples/sec	accuracy=0.748012	lr=0.000745
Epoch[116] Batch [1099]	Speed: 4261.570899 samples/sec	accuracy=0.748014	lr=0.000727
Epoch[116] Batch [1149]	Speed: 3494.778660 samples/sec	accuracy=0.748015	lr=0.000708
Epoch[116] Batch [1199]	Speed: 3354.162748 samples/sec	accuracy=0.748012	lr=0.000690
Epoch[116] Batch [1249]	Speed: 4310.511431 samples/sec	accuracy=0.748020	lr=0.000672
[Epoch 116] training: accuracy=0.748021
[Epoch 116] speed: 3564 samples/sec	time cost: 382.070313
[Epoch 116] validation: err-top1=0.219523 err-top5=0.061584
Epoch[117] Batch [49]	Speed: 4214.266867 samples/sec	accuracy=0.748430	lr=0.000654
Epoch[117] Batch [99]	Speed: 4135.456932 samples/sec	accuracy=0.748432	lr=0.000636
Epoch[117] Batch [149]	Speed: 4183.613886 samples/sec	accuracy=0.748450	lr=0.000619
Epoch[117] Batch [199]	Speed: 4081.652775 samples/sec	accuracy=0.748354	lr=0.000602
Epoch[117] Batch [249]	Speed: 4160.981930 samples/sec	accuracy=0.748329	lr=0.000585
Epoch[117] Batch [299]	Speed: 4291.450452 samples/sec	accuracy=0.748266	lr=0.000569
Epoch[117] Batch [349]	Speed: 3361.923728 samples/sec	accuracy=0.748291	lr=0.000552
Epoch[117] Batch [399]	Speed: 3349.550347 samples/sec	accuracy=0.748275	lr=0.000536
Epoch[117] Batch [449]	Speed: 4323.603821 samples/sec	accuracy=0.748272	lr=0.000520
Epoch[117] Batch [499]	Speed: 4356.925478 samples/sec	accuracy=0.748286	lr=0.000505
Epoch[117] Batch [549]	Speed: 4327.233137 samples/sec	accuracy=0.748267	lr=0.000489
Epoch[117] Batch [599]	Speed: 4176.261734 samples/sec	accuracy=0.748261	lr=0.000474
Epoch[117] Batch [649]	Speed: 4245.104637 samples/sec	accuracy=0.748251	lr=0.000459
Epoch[117] Batch [699]	Speed: 4163.328026 samples/sec	accuracy=0.748255	lr=0.000445
Epoch[117] Batch [749]	Speed: 2491.934077 samples/sec	accuracy=0.748253	lr=0.000430
Epoch[117] Batch [799]	Speed: 3478.927796 samples/sec	accuracy=0.748266	lr=0.000416
Epoch[117] Batch [849]	Speed: 4272.239750 samples/sec	accuracy=0.748273	lr=0.000402
Epoch[117] Batch [899]	Speed: 4104.348230 samples/sec	accuracy=0.748272	lr=0.000388
Epoch[117] Batch [949]	Speed: 4166.427669 samples/sec	accuracy=0.748298	lr=0.000375
Epoch[117] Batch [999]	Speed: 3785.717766 samples/sec	accuracy=0.748309	lr=0.000362
Epoch[117] Batch [1049]	Speed: 3434.438890 samples/sec	accuracy=0.748317	lr=0.000349
Epoch[117] Batch [1099]	Speed: 4256.014295 samples/sec	accuracy=0.748334	lr=0.000336
Epoch[117] Batch [1149]	Speed: 2689.540498 samples/sec	accuracy=0.748333	lr=0.000323
Epoch[117] Batch [1199]	Speed: 4347.698708 samples/sec	accuracy=0.748328	lr=0.000311
Epoch[117] Batch [1249]	Speed: 4293.653352 samples/sec	accuracy=0.748323	lr=0.000299
[Epoch 117] training: accuracy=0.748322
[Epoch 117] speed: 3859 samples/sec	time cost: 349.783054
[Epoch 117] validation: err-top1=0.219547 err-top5=0.061224
Epoch[118] Batch [49]	Speed: 3627.279702 samples/sec	accuracy=0.748277	lr=0.000287
Epoch[118] Batch [99]	Speed: 3411.003382 samples/sec	accuracy=0.748170	lr=0.000275
Epoch[118] Batch [149]	Speed: 4154.994265 samples/sec	accuracy=0.748139	lr=0.000264
Epoch[118] Batch [199]	Speed: 4411.455756 samples/sec	accuracy=0.748230	lr=0.000253
Epoch[118] Batch [249]	Speed: 4179.834751 samples/sec	accuracy=0.748282	lr=0.000242
Epoch[118] Batch [299]	Speed: 3375.303380 samples/sec	accuracy=0.748311	lr=0.000231
Epoch[118] Batch [349]	Speed: 4281.720509 samples/sec	accuracy=0.748315	lr=0.000221
Epoch[118] Batch [399]	Speed: 4183.117592 samples/sec	accuracy=0.748314	lr=0.000211
Epoch[118] Batch [449]	Speed: 4237.437890 samples/sec	accuracy=0.748345	lr=0.000201
Epoch[118] Batch [499]	Speed: 4143.275116 samples/sec	accuracy=0.748401	lr=0.000191
Epoch[118] Batch [549]	Speed: 4237.095938 samples/sec	accuracy=0.748397	lr=0.000182
Epoch[118] Batch [599]	Speed: 4215.200527 samples/sec	accuracy=0.748396	lr=0.000173
Epoch[118] Batch [649]	Speed: 2853.209040 samples/sec	accuracy=0.748390	lr=0.000164
Epoch[118] Batch [699]	Speed: 4226.842431 samples/sec	accuracy=0.748403	lr=0.000155
Epoch[118] Batch [749]	Speed: 4111.658854 samples/sec	accuracy=0.748390	lr=0.000147
Epoch[118] Batch [799]	Speed: 2424.822695 samples/sec	accuracy=0.748407	lr=0.000138
Epoch[118] Batch [849]	Speed: 4039.891078 samples/sec	accuracy=0.748401	lr=0.000130
Epoch[118] Batch [899]	Speed: 4189.105418 samples/sec	accuracy=0.748408	lr=0.000123
Epoch[118] Batch [949]	Speed: 3990.396233 samples/sec	accuracy=0.748397	lr=0.000115
Epoch[118] Batch [999]	Speed: 3388.149911 samples/sec	accuracy=0.748384	lr=0.000108
Epoch[118] Batch [1049]	Speed: 3732.077135 samples/sec	accuracy=0.748382	lr=0.000101
Epoch[118] Batch [1099]	Speed: 4221.307672 samples/sec	accuracy=0.748379	lr=0.000094
Epoch[118] Batch [1149]	Speed: 4233.566192 samples/sec	accuracy=0.748391	lr=0.000087
Epoch[118] Batch [1199]	Speed: 4224.600242 samples/sec	accuracy=0.748397	lr=0.000081
Epoch[118] Batch [1249]	Speed: 4291.005925 samples/sec	accuracy=0.748394	lr=0.000075
[Epoch 118] training: accuracy=0.748395
[Epoch 118] speed: 3853 samples/sec	time cost: 342.430120
[Epoch 118] validation: err-top1=0.219228 err-top5=0.061484
Epoch[119] Batch [49]	Speed: 4131.378506 samples/sec	accuracy=0.747843	lr=0.000069
Epoch[119] Batch [99]	Speed: 4015.943929 samples/sec	accuracy=0.748156	lr=0.000063
Epoch[119] Batch [149]	Speed: 4246.393312 samples/sec	accuracy=0.748322	lr=0.000058
Epoch[119] Batch [199]	Speed: 4068.075741 samples/sec	accuracy=0.748395	lr=0.000053
Epoch[119] Batch [249]	Speed: 4152.323442 samples/sec	accuracy=0.748379	lr=0.000048
Epoch[119] Batch [299]	Speed: 4157.171688 samples/sec	accuracy=0.748411	lr=0.000043
Epoch[119] Batch [349]	Speed: 4168.513116 samples/sec	accuracy=0.748407	lr=0.000039
Epoch[119] Batch [399]	Speed: 4192.435418 samples/sec	accuracy=0.748381	lr=0.000035
Epoch[119] Batch [449]	Speed: 4185.841895 samples/sec	accuracy=0.748363	lr=0.000031
Epoch[119] Batch [499]	Speed: 4059.122074 samples/sec	accuracy=0.748369	lr=0.000027
Epoch[119] Batch [549]	Speed: 4266.942953 samples/sec	accuracy=0.748392	lr=0.000023
Epoch[119] Batch [599]	Speed: 4235.826350 samples/sec	accuracy=0.748400	lr=0.000020
Epoch[119] Batch [649]	Speed: 4175.080040 samples/sec	accuracy=0.748409	lr=0.000017
Epoch[119] Batch [699]	Speed: 4145.407622 samples/sec	accuracy=0.748428	lr=0.000015
Epoch[119] Batch [749]	Speed: 4115.794023 samples/sec	accuracy=0.748430	lr=0.000012
Epoch[119] Batch [799]	Speed: 4188.097273 samples/sec	accuracy=0.748438	lr=0.000010
Epoch[119] Batch [849]	Speed: 4101.897855 samples/sec	accuracy=0.748419	lr=0.000008
Epoch[119] Batch [899]	Speed: 4089.932517 samples/sec	accuracy=0.748448	lr=0.000006
Epoch[119] Batch [949]	Speed: 4348.383711 samples/sec	accuracy=0.748446	lr=0.000004
Epoch[119] Batch [999]	Speed: 4334.561434 samples/sec	accuracy=0.748427	lr=0.000003
Epoch[119] Batch [1049]	Speed: 4236.168013 samples/sec	accuracy=0.748414	lr=0.000002
Epoch[119] Batch [1099]	Speed: 4170.091164 samples/sec	accuracy=0.748415	lr=0.000001
Epoch[119] Batch [1149]	Speed: 4309.611615 samples/sec	accuracy=0.748411	lr=0.000000
Epoch[119] Batch [1199]	Speed: 4223.343279 samples/sec	accuracy=0.748422	lr=0.000000
Epoch[119] Batch [1249]	Speed: 4317.400856 samples/sec	accuracy=0.748423	lr=0.000000
[Epoch 119] training: accuracy=0.748425
[Epoch 119] speed: 4179 samples/sec	time cost: 316.454825
[Epoch 119] validation: err-top1=0.218909 err-top5=0.061304