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='mobilenet0.25.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='mobilenet0.25', 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', resume_epoch=0, resume_params='', resume_states='', save_dir='params_mobilenet0.25', 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: 2983.071428 samples/sec	accuracy=0.000859	lr=0.003133
Epoch[0] Batch [99]	Speed: 5579.326897 samples/sec	accuracy=0.000996	lr=0.006331
Epoch[0] Batch [149]	Speed: 5762.866884 samples/sec	accuracy=0.001068	lr=0.009528
Epoch[0] Batch [199]	Speed: 5588.459939 samples/sec	accuracy=0.001074	lr=0.012726
Epoch[0] Batch [249]	Speed: 5583.985004 samples/sec	accuracy=0.001094	lr=0.015923
Epoch[0] Batch [299]	Speed: 5326.112999 samples/sec	accuracy=0.001104	lr=0.019121
Epoch[0] Batch [349]	Speed: 5571.488312 samples/sec	accuracy=0.001122	lr=0.022318
Epoch[0] Batch [399]	Speed: 4929.273054 samples/sec	accuracy=0.001165	lr=0.025516
Epoch[0] Batch [449]	Speed: 5225.309387 samples/sec	accuracy=0.001196	lr=0.028713
Epoch[0] Batch [499]	Speed: 5578.159817 samples/sec	accuracy=0.001242	lr=0.031910
Epoch[0] Batch [549]	Speed: 5558.446986 samples/sec	accuracy=0.001316	lr=0.035108
Epoch[0] Batch [599]	Speed: 5343.611912 samples/sec	accuracy=0.001390	lr=0.038305
Epoch[0] Batch [649]	Speed: 5707.251736 samples/sec	accuracy=0.001462	lr=0.041503
Epoch[0] Batch [699]	Speed: 5840.191424 samples/sec	accuracy=0.001542	lr=0.044700
Epoch[0] Batch [749]	Speed: 5071.370724 samples/sec	accuracy=0.001626	lr=0.047898
Epoch[0] Batch [799]	Speed: 5437.655327 samples/sec	accuracy=0.001743	lr=0.051095
Epoch[0] Batch [849]	Speed: 5319.632584 samples/sec	accuracy=0.001865	lr=0.054293
Epoch[0] Batch [899]	Speed: 5393.210340 samples/sec	accuracy=0.001967	lr=0.057490
Epoch[0] Batch [949]	Speed: 5496.266754 samples/sec	accuracy=0.002111	lr=0.060687
Epoch[0] Batch [999]	Speed: 5604.843561 samples/sec	accuracy=0.002258	lr=0.063885
Epoch[0] Batch [1049]	Speed: 5099.957186 samples/sec	accuracy=0.002453	lr=0.067082
Epoch[0] Batch [1099]	Speed: 5649.275590 samples/sec	accuracy=0.002684	lr=0.070280
Epoch[0] Batch [1149]	Speed: 5429.632447 samples/sec	accuracy=0.002914	lr=0.073477
Epoch[0] Batch [1199]	Speed: 5237.488941 samples/sec	accuracy=0.003162	lr=0.076675
Epoch[0] Batch [1249]	Speed: 5515.896828 samples/sec	accuracy=0.003490	lr=0.079872
[Epoch 0] training: accuracy=0.003506
[Epoch 0] speed: 5260 samples/sec	time cost: 253.525289
[Epoch 0] validation: err-top1=0.988859 err-top5=0.953364
Epoch[1] Batch [49]	Speed: 5237.931587 samples/sec	accuracy=0.013984	lr=0.083133
Epoch[1] Batch [99]	Speed: 5473.677979 samples/sec	accuracy=0.014277	lr=0.086331
Epoch[1] Batch [149]	Speed: 5555.237324 samples/sec	accuracy=0.014616	lr=0.089528
Epoch[1] Batch [199]	Speed: 5569.857275 samples/sec	accuracy=0.015405	lr=0.092726
Epoch[1] Batch [249]	Speed: 5393.260997 samples/sec	accuracy=0.016207	lr=0.095923
Epoch[1] Batch [299]	Speed: 5273.583172 samples/sec	accuracy=0.017142	lr=0.099121
Epoch[1] Batch [349]	Speed: 5341.360420 samples/sec	accuracy=0.018100	lr=0.102318
Epoch[1] Batch [399]	Speed: 5830.966259 samples/sec	accuracy=0.019143	lr=0.105516
Epoch[1] Batch [449]	Speed: 5818.434387 samples/sec	accuracy=0.020269	lr=0.108713
Epoch[1] Batch [499]	Speed: 5679.522704 samples/sec	accuracy=0.021475	lr=0.111910
Epoch[1] Batch [549]	Speed: 5380.262561 samples/sec	accuracy=0.022649	lr=0.115108
Epoch[1] Batch [599]	Speed: 5399.874496 samples/sec	accuracy=0.023908	lr=0.118305
Epoch[1] Batch [649]	Speed: 5687.393092 samples/sec	accuracy=0.025030	lr=0.121503
Epoch[1] Batch [699]	Speed: 5700.290153 samples/sec	accuracy=0.026239	lr=0.124700
Epoch[1] Batch [749]	Speed: 5742.921472 samples/sec	accuracy=0.027547	lr=0.127898
Epoch[1] Batch [799]	Speed: 5617.756924 samples/sec	accuracy=0.028907	lr=0.131095
Epoch[1] Batch [849]	Speed: 5681.373121 samples/sec	accuracy=0.030237	lr=0.134293
Epoch[1] Batch [899]	Speed: 5785.517937 samples/sec	accuracy=0.031638	lr=0.137490
Epoch[1] Batch [949]	Speed: 5662.329350 samples/sec	accuracy=0.033061	lr=0.140687
Epoch[1] Batch [999]	Speed: 5762.177386 samples/sec	accuracy=0.034548	lr=0.143885
Epoch[1] Batch [1049]	Speed: 5636.698629 samples/sec	accuracy=0.036001	lr=0.147082
Epoch[1] Batch [1099]	Speed: 5848.886468 samples/sec	accuracy=0.037466	lr=0.150280
Epoch[1] Batch [1149]	Speed: 5855.509381 samples/sec	accuracy=0.038957	lr=0.153477
Epoch[1] Batch [1199]	Speed: 5695.125675 samples/sec	accuracy=0.040420	lr=0.156675
Epoch[1] Batch [1249]	Speed: 5995.311834 samples/sec	accuracy=0.041871	lr=0.159872
[Epoch 1] training: accuracy=0.041909
[Epoch 1] speed: 5608 samples/sec	time cost: 237.778793
[Epoch 1] validation: err-top1=0.945173 err-top5=0.840322
Epoch[2] Batch [49]	Speed: 5546.845939 samples/sec	accuracy=0.081484	lr=0.163133
Epoch[2] Batch [99]	Speed: 5577.247276 samples/sec	accuracy=0.083535	lr=0.166331
Epoch[2] Batch [149]	Speed: 5705.112354 samples/sec	accuracy=0.084922	lr=0.169528
Epoch[2] Batch [199]	Speed: 5648.332652 samples/sec	accuracy=0.086870	lr=0.172726
Epoch[2] Batch [249]	Speed: 5666.211426 samples/sec	accuracy=0.088324	lr=0.175923
Epoch[2] Batch [299]	Speed: 5681.886914 samples/sec	accuracy=0.089899	lr=0.179121
Epoch[2] Batch [349]	Speed: 5551.927732 samples/sec	accuracy=0.091722	lr=0.182318
Epoch[2] Batch [399]	Speed: 5952.214202 samples/sec	accuracy=0.093052	lr=0.185516
Epoch[2] Batch [449]	Speed: 5543.974243 samples/sec	accuracy=0.094701	lr=0.188713
Epoch[2] Batch [499]	Speed: 6074.448235 samples/sec	accuracy=0.096211	lr=0.191910
Epoch[2] Batch [549]	Speed: 5893.112122 samples/sec	accuracy=0.097567	lr=0.195108
Epoch[2] Batch [599]	Speed: 5680.329888 samples/sec	accuracy=0.099189	lr=0.198305
Epoch[2] Batch [649]	Speed: 5381.822469 samples/sec	accuracy=0.100886	lr=0.201503
Epoch[2] Batch [699]	Speed: 5480.641075 samples/sec	accuracy=0.102291	lr=0.204700
Epoch[2] Batch [749]	Speed: 4331.797451 samples/sec	accuracy=0.103919	lr=0.207898
Epoch[2] Batch [799]	Speed: 5550.095825 samples/sec	accuracy=0.105162	lr=0.211095
Epoch[2] Batch [849]	Speed: 5299.752843 samples/sec	accuracy=0.106491	lr=0.214293
Epoch[2] Batch [899]	Speed: 4833.317578 samples/sec	accuracy=0.107923	lr=0.217490
Epoch[2] Batch [949]	Speed: 5701.547655 samples/sec	accuracy=0.109190	lr=0.220687
Epoch[2] Batch [999]	Speed: 5058.361923 samples/sec	accuracy=0.110583	lr=0.223885
Epoch[2] Batch [1049]	Speed: 5726.725379 samples/sec	accuracy=0.112026	lr=0.227082
Epoch[2] Batch [1099]	Speed: 5969.192915 samples/sec	accuracy=0.113181	lr=0.230280
Epoch[2] Batch [1149]	Speed: 5563.764695 samples/sec	accuracy=0.114647	lr=0.233477
Epoch[2] Batch [1199]	Speed: 5576.058479 samples/sec	accuracy=0.115937	lr=0.236675
Epoch[2] Batch [1249]	Speed: 5069.848521 samples/sec	accuracy=0.117194	lr=0.239872
[Epoch 2] training: accuracy=0.117220
[Epoch 2] speed: 5488 samples/sec	time cost: 244.979089
[Epoch 2] validation: err-top1=0.874123 err-top5=0.699577
Epoch[3] Batch [49]	Speed: 5576.565129 samples/sec	accuracy=0.154902	lr=0.243133
Epoch[3] Batch [99]	Speed: 4899.477914 samples/sec	accuracy=0.155020	lr=0.246331
Epoch[3] Batch [149]	Speed: 5596.513378 samples/sec	accuracy=0.156165	lr=0.249528
Epoch[3] Batch [199]	Speed: 5967.796519 samples/sec	accuracy=0.156689	lr=0.252726
Epoch[3] Batch [249]	Speed: 5213.425039 samples/sec	accuracy=0.157035	lr=0.255923
Epoch[3] Batch [299]	Speed: 4720.255210 samples/sec	accuracy=0.157780	lr=0.259121
Epoch[3] Batch [349]	Speed: 5727.287121 samples/sec	accuracy=0.158636	lr=0.262318
Epoch[3] Batch [399]	Speed: 5445.303025 samples/sec	accuracy=0.159585	lr=0.265516
Epoch[3] Batch [449]	Speed: 5573.760958 samples/sec	accuracy=0.160844	lr=0.268713
Epoch[3] Batch [499]	Speed: 5223.507495 samples/sec	accuracy=0.161879	lr=0.271910
Epoch[3] Batch [549]	Speed: 5138.378356 samples/sec	accuracy=0.162470	lr=0.275108
Epoch[3] Batch [599]	Speed: 4613.322039 samples/sec	accuracy=0.163237	lr=0.278305
Epoch[3] Batch [649]	Speed: 5444.289745 samples/sec	accuracy=0.164345	lr=0.281503
Epoch[3] Batch [699]	Speed: 6053.975224 samples/sec	accuracy=0.165229	lr=0.284700
Epoch[3] Batch [749]	Speed: 4921.435300 samples/sec	accuracy=0.166289	lr=0.287898
Epoch[3] Batch [799]	Speed: 5203.793615 samples/sec	accuracy=0.167195	lr=0.291095
Epoch[3] Batch [849]	Speed: 4666.864704 samples/sec	accuracy=0.167995	lr=0.294293
Epoch[3] Batch [899]	Speed: 5072.370459 samples/sec	accuracy=0.168803	lr=0.297490
Epoch[3] Batch [949]	Speed: 5466.388703 samples/sec	accuracy=0.169885	lr=0.300687
Epoch[3] Batch [999]	Speed: 5881.862348 samples/sec	accuracy=0.170848	lr=0.303885
Epoch[3] Batch [1049]	Speed: 5691.011142 samples/sec	accuracy=0.171713	lr=0.307082
Epoch[3] Batch [1099]	Speed: 5665.664590 samples/sec	accuracy=0.172438	lr=0.310280
Epoch[3] Batch [1149]	Speed: 4487.357170 samples/sec	accuracy=0.173128	lr=0.313477
Epoch[3] Batch [1199]	Speed: 4147.269829 samples/sec	accuracy=0.173793	lr=0.316675
Epoch[3] Batch [1249]	Speed: 5719.009002 samples/sec	accuracy=0.174686	lr=0.319872
[Epoch 3] training: accuracy=0.174703
[Epoch 3] speed: 5221 samples/sec	time cost: 255.609138
[Epoch 3] validation: err-top1=0.843870 err-top5=0.641303
Epoch[4] Batch [49]	Speed: 3655.209252 samples/sec	accuracy=0.193008	lr=0.323133
Epoch[4] Batch [99]	Speed: 5103.110707 samples/sec	accuracy=0.194775	lr=0.326331
Epoch[4] Batch [149]	Speed: 5783.040710 samples/sec	accuracy=0.197174	lr=0.329528
Epoch[4] Batch [199]	Speed: 4729.446507 samples/sec	accuracy=0.198667	lr=0.332726
Epoch[4] Batch [249]	Speed: 4505.218358 samples/sec	accuracy=0.199766	lr=0.335923
Epoch[4] Batch [299]	Speed: 4453.726524 samples/sec	accuracy=0.199427	lr=0.339121
Epoch[4] Batch [349]	Speed: 4913.909589 samples/sec	accuracy=0.200028	lr=0.342318
Epoch[4] Batch [399]	Speed: 5536.068198 samples/sec	accuracy=0.200671	lr=0.345516
Epoch[4] Batch [449]	Speed: 4799.783466 samples/sec	accuracy=0.201313	lr=0.348713
Epoch[4] Batch [499]	Speed: 5539.855047 samples/sec	accuracy=0.201764	lr=0.351910
Epoch[4] Batch [549]	Speed: 5589.943138 samples/sec	accuracy=0.202243	lr=0.355108
Epoch[4] Batch [599]	Speed: 5838.091692 samples/sec	accuracy=0.202723	lr=0.358305
Epoch[4] Batch [649]	Speed: 4911.289763 samples/sec	accuracy=0.203310	lr=0.361503
Epoch[4] Batch [699]	Speed: 4588.791762 samples/sec	accuracy=0.203945	lr=0.364700
Epoch[4] Batch [749]	Speed: 4866.386805 samples/sec	accuracy=0.204578	lr=0.367898
Epoch[4] Batch [799]	Speed: 5173.087771 samples/sec	accuracy=0.205099	lr=0.371095
Epoch[4] Batch [849]	Speed: 4687.268925 samples/sec	accuracy=0.205596	lr=0.374293
Epoch[4] Batch [899]	Speed: 4205.715700 samples/sec	accuracy=0.205634	lr=0.377490
Epoch[4] Batch [949]	Speed: 5619.889079 samples/sec	accuracy=0.206225	lr=0.380687
Epoch[4] Batch [999]	Speed: 4781.938489 samples/sec	accuracy=0.206621	lr=0.383885
Epoch[4] Batch [1049]	Speed: 5310.647020 samples/sec	accuracy=0.207115	lr=0.387082
Epoch[4] Batch [1099]	Speed: 4588.496540 samples/sec	accuracy=0.207679	lr=0.390280
Epoch[4] Batch [1149]	Speed: 4840.661015 samples/sec	accuracy=0.208187	lr=0.393477
Epoch[4] Batch [1199]	Speed: 4739.913917 samples/sec	accuracy=0.208771	lr=0.396675
Epoch[4] Batch [1249]	Speed: 5404.542725 samples/sec	accuracy=0.209216	lr=0.399872
[Epoch 4] training: accuracy=0.209220
[Epoch 4] speed: 4906 samples/sec	time cost: 270.078547
[Epoch 4] validation: err-top1=0.818599 err-top5=0.607601
Epoch[5] Batch [49]	Speed: 5355.896744 samples/sec	accuracy=0.226641	lr=0.400000
Epoch[5] Batch [99]	Speed: 4771.690901 samples/sec	accuracy=0.226855	lr=0.400000
Epoch[5] Batch [149]	Speed: 4868.469608 samples/sec	accuracy=0.226582	lr=0.399999
Epoch[5] Batch [199]	Speed: 4685.682255 samples/sec	accuracy=0.226621	lr=0.399998
Epoch[5] Batch [249]	Speed: 5300.041386 samples/sec	accuracy=0.226648	lr=0.399997
Epoch[5] Batch [299]	Speed: 4292.980848 samples/sec	accuracy=0.226989	lr=0.399996
Epoch[5] Batch [349]	Speed: 4700.419874 samples/sec	accuracy=0.226998	lr=0.399994
Epoch[5] Batch [399]	Speed: 5080.314164 samples/sec	accuracy=0.227932	lr=0.399992
Epoch[5] Batch [449]	Speed: 5427.116700 samples/sec	accuracy=0.228980	lr=0.399990
Epoch[5] Batch [499]	Speed: 5199.661999 samples/sec	accuracy=0.229367	lr=0.399988
Epoch[5] Batch [549]	Speed: 5861.003845 samples/sec	accuracy=0.229743	lr=0.399986
Epoch[5] Batch [599]	Speed: 5262.865268 samples/sec	accuracy=0.230298	lr=0.399983
Epoch[5] Batch [649]	Speed: 3922.578126 samples/sec	accuracy=0.230841	lr=0.399980
Epoch[5] Batch [699]	Speed: 5591.990303 samples/sec	accuracy=0.231233	lr=0.399977
Epoch[5] Batch [749]	Speed: 5973.615329 samples/sec	accuracy=0.231789	lr=0.399973
Epoch[5] Batch [799]	Speed: 5507.082272 samples/sec	accuracy=0.232225	lr=0.399970
Epoch[5] Batch [849]	Speed: 5616.844751 samples/sec	accuracy=0.232630	lr=0.399966
Epoch[5] Batch [899]	Speed: 6197.705357 samples/sec	accuracy=0.233281	lr=0.399961
Epoch[5] Batch [949]	Speed: 4931.855699 samples/sec	accuracy=0.233563	lr=0.399957
Epoch[5] Batch [999]	Speed: 5597.108800 samples/sec	accuracy=0.234133	lr=0.399952
Epoch[5] Batch [1049]	Speed: 5365.976899 samples/sec	accuracy=0.234562	lr=0.399948
Epoch[5] Batch [1099]	Speed: 5381.751256 samples/sec	accuracy=0.234934	lr=0.399942
Epoch[5] Batch [1149]	Speed: 5627.909233 samples/sec	accuracy=0.235239	lr=0.399937
Epoch[5] Batch [1199]	Speed: 5301.043421 samples/sec	accuracy=0.235680	lr=0.399931
Epoch[5] Batch [1249]	Speed: 5526.627670 samples/sec	accuracy=0.235984	lr=0.399926
[Epoch 5] training: accuracy=0.236002
[Epoch 5] speed: 5193 samples/sec	time cost: 255.640058
[Epoch 5] validation: err-top1=0.829264 err-top5=0.632772
Epoch[6] Batch [49]	Speed: 4841.422530 samples/sec	accuracy=0.246719	lr=0.399919
Epoch[6] Batch [99]	Speed: 5688.597442 samples/sec	accuracy=0.249805	lr=0.399913
Epoch[6] Batch [149]	Speed: 5604.061343 samples/sec	accuracy=0.247546	lr=0.399907
Epoch[6] Batch [199]	Speed: 5346.079300 samples/sec	accuracy=0.248037	lr=0.399900
Epoch[6] Batch [249]	Speed: 5285.416636 samples/sec	accuracy=0.248184	lr=0.399893
Epoch[6] Batch [299]	Speed: 5792.131138 samples/sec	accuracy=0.249144	lr=0.399885
Epoch[6] Batch [349]	Speed: 5954.332454 samples/sec	accuracy=0.249833	lr=0.399878
Epoch[6] Batch [399]	Speed: 5750.387342 samples/sec	accuracy=0.250254	lr=0.399870
Epoch[6] Batch [449]	Speed: 5192.442445 samples/sec	accuracy=0.250506	lr=0.399862
Epoch[6] Batch [499]	Speed: 5814.957224 samples/sec	accuracy=0.250773	lr=0.399854
Epoch[6] Batch [549]	Speed: 5714.045913 samples/sec	accuracy=0.250875	lr=0.399846
Epoch[6] Batch [599]	Speed: 5646.354931 samples/sec	accuracy=0.251128	lr=0.399837
Epoch[6] Batch [649]	Speed: 5941.335479 samples/sec	accuracy=0.251746	lr=0.399828
Epoch[6] Batch [699]	Speed: 5696.758073 samples/sec	accuracy=0.252063	lr=0.399819
Epoch[6] Batch [749]	Speed: 5630.799615 samples/sec	accuracy=0.252405	lr=0.399809
Epoch[6] Batch [799]	Speed: 5639.224851 samples/sec	accuracy=0.252716	lr=0.399800
Epoch[6] Batch [849]	Speed: 5747.649650 samples/sec	accuracy=0.252974	lr=0.399790
Epoch[6] Batch [899]	Speed: 5627.971770 samples/sec	accuracy=0.253226	lr=0.399780
Epoch[6] Batch [949]	Speed: 5699.856081 samples/sec	accuracy=0.253646	lr=0.399769
Epoch[6] Batch [999]	Speed: 5806.702286 samples/sec	accuracy=0.254023	lr=0.399759
Epoch[6] Batch [1049]	Speed: 4925.941485 samples/sec	accuracy=0.254355	lr=0.399748
Epoch[6] Batch [1099]	Speed: 6047.219721 samples/sec	accuracy=0.254684	lr=0.399737
Epoch[6] Batch [1149]	Speed: 4843.681530 samples/sec	accuracy=0.254943	lr=0.399725
Epoch[6] Batch [1199]	Speed: 5650.384755 samples/sec	accuracy=0.255259	lr=0.399714
Epoch[6] Batch [1249]	Speed: 5744.806830 samples/sec	accuracy=0.255427	lr=0.399702
[Epoch 6] training: accuracy=0.255430
[Epoch 6] speed: 5559 samples/sec	time cost: 239.831273
[Epoch 6] validation: err-top1=0.795998 err-top5=0.574976
Epoch[7] Batch [49]	Speed: 5649.314824 samples/sec	accuracy=0.266562	lr=0.399690
Epoch[7] Batch [99]	Speed: 5796.972022 samples/sec	accuracy=0.267148	lr=0.399677
Epoch[7] Batch [149]	Speed: 5910.209780 samples/sec	accuracy=0.267493	lr=0.399665
Epoch[7] Batch [199]	Speed: 5565.597272 samples/sec	accuracy=0.266338	lr=0.399652
Epoch[7] Batch [249]	Speed: 5664.080884 samples/sec	accuracy=0.265637	lr=0.399639
Epoch[7] Batch [299]	Speed: 6018.587805 samples/sec	accuracy=0.266527	lr=0.399626
Epoch[7] Batch [349]	Speed: 5479.865591 samples/sec	accuracy=0.266565	lr=0.399613
Epoch[7] Batch [399]	Speed: 5878.240117 samples/sec	accuracy=0.266682	lr=0.399599
Epoch[7] Batch [449]	Speed: 5560.730169 samples/sec	accuracy=0.267003	lr=0.399585
Epoch[7] Batch [499]	Speed: 6067.979943 samples/sec	accuracy=0.266945	lr=0.399571
Epoch[7] Batch [549]	Speed: 5306.230642 samples/sec	accuracy=0.267069	lr=0.399556
Epoch[7] Batch [599]	Speed: 4125.757917 samples/sec	accuracy=0.267070	lr=0.399542
Epoch[7] Batch [649]	Speed: 4966.372713 samples/sec	accuracy=0.267281	lr=0.399527
Epoch[7] Batch [699]	Speed: 5052.203001 samples/sec	accuracy=0.267626	lr=0.399512
Epoch[7] Batch [749]	Speed: 5549.677729 samples/sec	accuracy=0.267590	lr=0.399496
Epoch[7] Batch [799]	Speed: 6019.762373 samples/sec	accuracy=0.267626	lr=0.399481
Epoch[7] Batch [849]	Speed: 4998.168173 samples/sec	accuracy=0.267759	lr=0.399465
Epoch[7] Batch [899]	Speed: 5076.474864 samples/sec	accuracy=0.268118	lr=0.399449
Epoch[7] Batch [949]	Speed: 4806.802231 samples/sec	accuracy=0.268363	lr=0.399432
Epoch[7] Batch [999]	Speed: 4822.761542 samples/sec	accuracy=0.268432	lr=0.399416
Epoch[7] Batch [1049]	Speed: 5117.852464 samples/sec	accuracy=0.268597	lr=0.399399
Epoch[7] Batch [1099]	Speed: 5171.063697 samples/sec	accuracy=0.268740	lr=0.399382
Epoch[7] Batch [1149]	Speed: 5580.473873 samples/sec	accuracy=0.268931	lr=0.399365
Epoch[7] Batch [1199]	Speed: 4963.286190 samples/sec	accuracy=0.268947	lr=0.399347
Epoch[7] Batch [1249]	Speed: 5003.531269 samples/sec	accuracy=0.269177	lr=0.399329
[Epoch 7] training: accuracy=0.269197
[Epoch 7] speed: 5313 samples/sec	time cost: 251.152923
[Epoch 7] validation: err-top1=0.777722 err-top5=0.548230
Epoch[8] Batch [49]	Speed: 4930.583170 samples/sec	accuracy=0.278359	lr=0.399311
Epoch[8] Batch [99]	Speed: 4999.680689 samples/sec	accuracy=0.276484	lr=0.399293
Epoch[8] Batch [149]	Speed: 5511.400378 samples/sec	accuracy=0.276439	lr=0.399274
Epoch[8] Batch [199]	Speed: 5902.009862 samples/sec	accuracy=0.276724	lr=0.399256
Epoch[8] Batch [249]	Speed: 4411.688395 samples/sec	accuracy=0.276824	lr=0.399237
Epoch[8] Batch [299]	Speed: 5103.800562 samples/sec	accuracy=0.276426	lr=0.399218
Epoch[8] Batch [349]	Speed: 6013.964512 samples/sec	accuracy=0.276680	lr=0.399198
Epoch[8] Batch [399]	Speed: 4647.416387 samples/sec	accuracy=0.276689	lr=0.399179
Epoch[8] Batch [449]	Speed: 4322.791984 samples/sec	accuracy=0.276721	lr=0.399159
Epoch[8] Batch [499]	Speed: 5949.651540 samples/sec	accuracy=0.276885	lr=0.399138
Epoch[8] Batch [549]	Speed: 5025.810743 samples/sec	accuracy=0.277214	lr=0.399118
Epoch[8] Batch [599]	Speed: 5268.571618 samples/sec	accuracy=0.277165	lr=0.399098
Epoch[8] Batch [649]	Speed: 5371.507944 samples/sec	accuracy=0.277136	lr=0.399077
Epoch[8] Batch [699]	Speed: 5478.760852 samples/sec	accuracy=0.277275	lr=0.399056
Epoch[8] Batch [749]	Speed: 4820.074286 samples/sec	accuracy=0.277517	lr=0.399034
Epoch[8] Batch [799]	Speed: 5693.182069 samples/sec	accuracy=0.277716	lr=0.399013
Epoch[8] Batch [849]	Speed: 5029.558449 samples/sec	accuracy=0.277721	lr=0.398991
Epoch[8] Batch [899]	Speed: 5120.054810 samples/sec	accuracy=0.278063	lr=0.398969
Epoch[8] Batch [949]	Speed: 4991.076417 samples/sec	accuracy=0.278213	lr=0.398947
Epoch[8] Batch [999]	Speed: 5373.406686 samples/sec	accuracy=0.278343	lr=0.398924
Epoch[8] Batch [1049]	Speed: 5036.988898 samples/sec	accuracy=0.278497	lr=0.398901
Epoch[8] Batch [1099]	Speed: 5618.001475 samples/sec	accuracy=0.278610	lr=0.398878
Epoch[8] Batch [1149]	Speed: 5699.589528 samples/sec	accuracy=0.278684	lr=0.398855
Epoch[8] Batch [1199]	Speed: 5261.984627 samples/sec	accuracy=0.278845	lr=0.398832
Epoch[8] Batch [1249]	Speed: 5142.364107 samples/sec	accuracy=0.279089	lr=0.398808
[Epoch 8] training: accuracy=0.279094
[Epoch 8] speed: 5185 samples/sec	time cost: 256.543309
[Epoch 8] validation: err-top1=0.789222 err-top5=0.566825
Epoch[9] Batch [49]	Speed: 4703.122722 samples/sec	accuracy=0.288809	lr=0.398784
Epoch[9] Batch [99]	Speed: 5421.218637 samples/sec	accuracy=0.287666	lr=0.398760
Epoch[9] Batch [149]	Speed: 5643.995865 samples/sec	accuracy=0.286673	lr=0.398735
Epoch[9] Batch [199]	Speed: 5129.563607 samples/sec	accuracy=0.286533	lr=0.398710
Epoch[9] Batch [249]	Speed: 5457.971238 samples/sec	accuracy=0.286125	lr=0.398686
Epoch[9] Batch [299]	Speed: 5239.545930 samples/sec	accuracy=0.285885	lr=0.398660
Epoch[9] Batch [349]	Speed: 5516.622598 samples/sec	accuracy=0.285940	lr=0.398635
Epoch[9] Batch [399]	Speed: 5895.838015 samples/sec	accuracy=0.286182	lr=0.398610
Epoch[9] Batch [449]	Speed: 5774.008734 samples/sec	accuracy=0.285833	lr=0.398584
Epoch[9] Batch [499]	Speed: 4945.084956 samples/sec	accuracy=0.285480	lr=0.398558
Epoch[9] Batch [549]	Speed: 1993.293820 samples/sec	accuracy=0.285444	lr=0.398531
Epoch[9] Batch [599]	Speed: 1315.042125 samples/sec	accuracy=0.285501	lr=0.398505
Epoch[9] Batch [649]	Speed: 4746.964104 samples/sec	accuracy=0.285553	lr=0.398478
Epoch[9] Batch [699]	Speed: 5005.026619 samples/sec	accuracy=0.285307	lr=0.398451
Epoch[9] Batch [749]	Speed: 4940.338086 samples/sec	accuracy=0.285281	lr=0.398424
Epoch[9] Batch [799]	Speed: 5633.735828 samples/sec	accuracy=0.285457	lr=0.398396
Epoch[9] Batch [849]	Speed: 4333.416400 samples/sec	accuracy=0.285434	lr=0.398369
Epoch[9] Batch [899]	Speed: 4807.933404 samples/sec	accuracy=0.285553	lr=0.398341
Epoch[9] Batch [949]	Speed: 5082.825234 samples/sec	accuracy=0.285624	lr=0.398312
Epoch[9] Batch [999]	Speed: 4686.133887 samples/sec	accuracy=0.285730	lr=0.398284
Epoch[9] Batch [1049]	Speed: 5636.562664 samples/sec	accuracy=0.285816	lr=0.398255
Epoch[9] Batch [1099]	Speed: 4277.659202 samples/sec	accuracy=0.285906	lr=0.398226
Epoch[9] Batch [1149]	Speed: 5530.346777 samples/sec	accuracy=0.285926	lr=0.398197
Epoch[9] Batch [1199]	Speed: 4610.737529 samples/sec	accuracy=0.285981	lr=0.398168
Epoch[9] Batch [1249]	Speed: 4981.792247 samples/sec	accuracy=0.286169	lr=0.398138
[Epoch 9] training: accuracy=0.286172
[Epoch 9] speed: 4318 samples/sec	time cost: 305.930237
[Epoch 9] validation: err-top1=0.790717 err-top5=0.566028
Epoch[10] Batch [49]	Speed: 5494.793756 samples/sec	accuracy=0.294023	lr=0.398108
Epoch[10] Batch [99]	Speed: 5774.261333 samples/sec	accuracy=0.293320	lr=0.398078
Epoch[10] Batch [149]	Speed: 5676.355527 samples/sec	accuracy=0.292995	lr=0.398048
Epoch[10] Batch [199]	Speed: 5904.067853 samples/sec	accuracy=0.292451	lr=0.398017
Epoch[10] Batch [249]	Speed: 6078.895200 samples/sec	accuracy=0.292379	lr=0.397986
Epoch[10] Batch [299]	Speed: 5725.682675 samples/sec	accuracy=0.291852	lr=0.397955
Epoch[10] Batch [349]	Speed: 5119.859135 samples/sec	accuracy=0.292148	lr=0.397924
Epoch[10] Batch [399]	Speed: 4746.350654 samples/sec	accuracy=0.292083	lr=0.397892
Epoch[10] Batch [449]	Speed: 4556.779718 samples/sec	accuracy=0.291845	lr=0.397861
Epoch[10] Batch [499]	Speed: 5210.521972 samples/sec	accuracy=0.291732	lr=0.397829
Epoch[10] Batch [549]	Speed: 5414.551693 samples/sec	accuracy=0.291690	lr=0.397796
Epoch[10] Batch [599]	Speed: 5125.425298 samples/sec	accuracy=0.291654	lr=0.397764
Epoch[10] Batch [649]	Speed: 4942.110696 samples/sec	accuracy=0.291581	lr=0.397731
Epoch[10] Batch [699]	Speed: 5411.158168 samples/sec	accuracy=0.291624	lr=0.397698
Epoch[10] Batch [749]	Speed: 5697.492768 samples/sec	accuracy=0.291934	lr=0.397665
Epoch[10] Batch [799]	Speed: 5580.754926 samples/sec	accuracy=0.292257	lr=0.397632
Epoch[10] Batch [849]	Speed: 5592.094127 samples/sec	accuracy=0.292443	lr=0.397598
Epoch[10] Batch [899]	Speed: 5481.565789 samples/sec	accuracy=0.292278	lr=0.397564
Epoch[10] Batch [949]	Speed: 5354.000343 samples/sec	accuracy=0.292318	lr=0.397530
Epoch[10] Batch [999]	Speed: 5597.022440 samples/sec	accuracy=0.292366	lr=0.397496
Epoch[10] Batch [1049]	Speed: 5877.493138 samples/sec	accuracy=0.292426	lr=0.397461
Epoch[10] Batch [1099]	Speed: 5619.133679 samples/sec	accuracy=0.292437	lr=0.397427
Epoch[10] Batch [1149]	Speed: 5099.811972 samples/sec	accuracy=0.292415	lr=0.397392
Epoch[10] Batch [1199]	Speed: 4756.117184 samples/sec	accuracy=0.292397	lr=0.397356
Epoch[10] Batch [1249]	Speed: 5145.702849 samples/sec	accuracy=0.292502	lr=0.397321
[Epoch 10] training: accuracy=0.292501
[Epoch 10] speed: 5361 samples/sec	time cost: 249.414196
[Epoch 10] validation: err-top1=0.919842 err-top5=0.803412
Epoch[11] Batch [49]	Speed: 4700.838439 samples/sec	accuracy=0.299687	lr=0.397284
Epoch[11] Batch [99]	Speed: 4520.309136 samples/sec	accuracy=0.298242	lr=0.397248
Epoch[11] Batch [149]	Speed: 6009.254317 samples/sec	accuracy=0.297819	lr=0.397212
Epoch[11] Batch [199]	Speed: 4637.546772 samples/sec	accuracy=0.297744	lr=0.397176
Epoch[11] Batch [249]	Speed: 5686.277633 samples/sec	accuracy=0.296844	lr=0.397139
Epoch[11] Batch [299]	Speed: 5584.007510 samples/sec	accuracy=0.297041	lr=0.397102
Epoch[11] Batch [349]	Speed: 5918.724172 samples/sec	accuracy=0.297001	lr=0.397065
Epoch[11] Batch [399]	Speed: 4565.662795 samples/sec	accuracy=0.296626	lr=0.397028
Epoch[11] Batch [449]	Speed: 5665.432762 samples/sec	accuracy=0.296335	lr=0.396990
Epoch[11] Batch [499]	Speed: 5729.194477 samples/sec	accuracy=0.296357	lr=0.396952
Epoch[11] Batch [549]	Speed: 5391.005375 samples/sec	accuracy=0.296294	lr=0.396914
Epoch[11] Batch [599]	Speed: 5046.719579 samples/sec	accuracy=0.296613	lr=0.396876
Epoch[11] Batch [649]	Speed: 5617.106559 samples/sec	accuracy=0.296453	lr=0.396837
Epoch[11] Batch [699]	Speed: 5427.364137 samples/sec	accuracy=0.296631	lr=0.396798
Epoch[11] Batch [749]	Speed: 4636.887784 samples/sec	accuracy=0.296717	lr=0.396759
Epoch[11] Batch [799]	Speed: 5047.014557 samples/sec	accuracy=0.296565	lr=0.396720
Epoch[11] Batch [849]	Speed: 5375.877719 samples/sec	accuracy=0.296705	lr=0.396680
Epoch[11] Batch [899]	Speed: 4880.432245 samples/sec	accuracy=0.296709	lr=0.396641
Epoch[11] Batch [949]	Speed: 4362.768204 samples/sec	accuracy=0.296555	lr=0.396601
Epoch[11] Batch [999]	Speed: 4573.273780 samples/sec	accuracy=0.296571	lr=0.396561
Epoch[11] Batch [1049]	Speed: 5855.084873 samples/sec	accuracy=0.296527	lr=0.396520
Epoch[11] Batch [1099]	Speed: 5264.385323 samples/sec	accuracy=0.296605	lr=0.396479
Epoch[11] Batch [1149]	Speed: 5494.237193 samples/sec	accuracy=0.296701	lr=0.396439
Epoch[11] Batch [1199]	Speed: 5599.052166 samples/sec	accuracy=0.296723	lr=0.396397
Epoch[11] Batch [1249]	Speed: 4500.026483 samples/sec	accuracy=0.296768	lr=0.396356
[Epoch 11] training: accuracy=0.296774
[Epoch 11] speed: 5144 samples/sec	time cost: 258.434926
[Epoch 11] validation: err-top1=0.780823 err-top5=0.551860
Epoch[12] Batch [49]	Speed: 5046.400800 samples/sec	accuracy=0.302988	lr=0.396314
Epoch[12] Batch [99]	Speed: 5831.694806 samples/sec	accuracy=0.302285	lr=0.396272
Epoch[12] Batch [149]	Speed: 5667.699990 samples/sec	accuracy=0.301836	lr=0.396230
Epoch[12] Batch [199]	Speed: 5862.624538 samples/sec	accuracy=0.301592	lr=0.396187
Epoch[12] Batch [249]	Speed: 5591.952298 samples/sec	accuracy=0.301070	lr=0.396145
Epoch[12] Batch [299]	Speed: 5368.804282 samples/sec	accuracy=0.300892	lr=0.396102
Epoch[12] Batch [349]	Speed: 4973.359607 samples/sec	accuracy=0.301465	lr=0.396059
Epoch[12] Batch [399]	Speed: 4420.221158 samples/sec	accuracy=0.301428	lr=0.396016
Epoch[12] Batch [449]	Speed: 5734.097113 samples/sec	accuracy=0.301239	lr=0.395972
Epoch[12] Batch [499]	Speed: 5586.250275 samples/sec	accuracy=0.301105	lr=0.395928
Epoch[12] Batch [549]	Speed: 5664.845280 samples/sec	accuracy=0.301099	lr=0.395885
Epoch[12] Batch [599]	Speed: 5768.800221 samples/sec	accuracy=0.301216	lr=0.395840
Epoch[12] Batch [649]	Speed: 4882.795748 samples/sec	accuracy=0.301702	lr=0.395796
Epoch[12] Batch [699]	Speed: 5113.069331 samples/sec	accuracy=0.301883	lr=0.395751
Epoch[12] Batch [749]	Speed: 5278.155631 samples/sec	accuracy=0.301701	lr=0.395706
Epoch[12] Batch [799]	Speed: 5174.917765 samples/sec	accuracy=0.301584	lr=0.395661
Epoch[12] Batch [849]	Speed: 5553.950733 samples/sec	accuracy=0.301502	lr=0.395616
Epoch[12] Batch [899]	Speed: 5638.911226 samples/sec	accuracy=0.301478	lr=0.395570
Epoch[12] Batch [949]	Speed: 6109.361702 samples/sec	accuracy=0.301485	lr=0.395524
Epoch[12] Batch [999]	Speed: 4892.878073 samples/sec	accuracy=0.301401	lr=0.395478
Epoch[12] Batch [1049]	Speed: 5382.822747 samples/sec	accuracy=0.301502	lr=0.395432
Epoch[12] Batch [1099]	Speed: 5905.267643 samples/sec	accuracy=0.301382	lr=0.395386
Epoch[12] Batch [1149]	Speed: 5956.878165 samples/sec	accuracy=0.301413	lr=0.395339
Epoch[12] Batch [1199]	Speed: 5278.339592 samples/sec	accuracy=0.301593	lr=0.395292
Epoch[12] Batch [1249]	Speed: 4981.820677 samples/sec	accuracy=0.301684	lr=0.395245
[Epoch 12] training: accuracy=0.301674
[Epoch 12] speed: 5388 samples/sec	time cost: 247.899057
[Epoch 12] validation: err-top1=0.820073 err-top5=0.617746
Epoch[13] Batch [49]	Speed: 4561.110270 samples/sec	accuracy=0.303574	lr=0.395196
Epoch[13] Batch [99]	Speed: 4885.645335 samples/sec	accuracy=0.304727	lr=0.395149
Epoch[13] Batch [149]	Speed: 5377.299490 samples/sec	accuracy=0.304961	lr=0.395101
Epoch[13] Batch [199]	Speed: 5540.900070 samples/sec	accuracy=0.304556	lr=0.395052
Epoch[13] Batch [249]	Speed: 4716.975054 samples/sec	accuracy=0.305141	lr=0.395004
Epoch[13] Batch [299]	Speed: 5608.754854 samples/sec	accuracy=0.305166	lr=0.394955
Epoch[13] Batch [349]	Speed: 5648.699627 samples/sec	accuracy=0.305142	lr=0.394907
Epoch[13] Batch [399]	Speed: 5672.274803 samples/sec	accuracy=0.304866	lr=0.394858
Epoch[13] Batch [449]	Speed: 5241.011866 samples/sec	accuracy=0.305087	lr=0.394808
Epoch[13] Batch [499]	Speed: 4666.205470 samples/sec	accuracy=0.305254	lr=0.394759
Epoch[13] Batch [549]	Speed: 5976.327075 samples/sec	accuracy=0.305359	lr=0.394709
Epoch[13] Batch [599]	Speed: 5493.975469 samples/sec	accuracy=0.305578	lr=0.394659
Epoch[13] Batch [649]	Speed: 5338.115446 samples/sec	accuracy=0.305102	lr=0.394609
Epoch[13] Batch [699]	Speed: 5420.148360 samples/sec	accuracy=0.304980	lr=0.394558
Epoch[13] Batch [749]	Speed: 5206.544270 samples/sec	accuracy=0.304853	lr=0.394507
Epoch[13] Batch [799]	Speed: 5546.317171 samples/sec	accuracy=0.304554	lr=0.394457
Epoch[13] Batch [849]	Speed: 4762.064698 samples/sec	accuracy=0.304415	lr=0.394405
Epoch[13] Batch [899]	Speed: 5248.266007 samples/sec	accuracy=0.304376	lr=0.394354
Epoch[13] Batch [949]	Speed: 4212.964718 samples/sec	accuracy=0.304376	lr=0.394302
Epoch[13] Batch [999]	Speed: 5323.479903 samples/sec	accuracy=0.304284	lr=0.394250
Epoch[13] Batch [1049]	Speed: 5960.496741 samples/sec	accuracy=0.304231	lr=0.394198
Epoch[13] Batch [1099]	Speed: 5024.559574 samples/sec	accuracy=0.304279	lr=0.394146
Epoch[13] Batch [1149]	Speed: 5236.430472 samples/sec	accuracy=0.304293	lr=0.394093
Epoch[13] Batch [1199]	Speed: 4922.800497 samples/sec	accuracy=0.304297	lr=0.394041
Epoch[13] Batch [1249]	Speed: 4720.092531 samples/sec	accuracy=0.304304	lr=0.393988
[Epoch 13] training: accuracy=0.304304
[Epoch 13] speed: 5170 samples/sec	time cost: 257.268484
[Epoch 13] validation: err-top1=0.797592 err-top5=0.584124
Epoch[14] Batch [49]	Speed: 5256.017863 samples/sec	accuracy=0.305586	lr=0.393933
Epoch[14] Batch [99]	Speed: 5634.770739 samples/sec	accuracy=0.309326	lr=0.393880
Epoch[14] Batch [149]	Speed: 4587.279675 samples/sec	accuracy=0.310000	lr=0.393826
Epoch[14] Batch [199]	Speed: 5172.584003 samples/sec	accuracy=0.309868	lr=0.393772
Epoch[14] Batch [249]	Speed: 5905.821432 samples/sec	accuracy=0.309465	lr=0.393718
Epoch[14] Batch [299]	Speed: 5801.286234 samples/sec	accuracy=0.308971	lr=0.393664
Epoch[14] Batch [349]	Speed: 5605.958761 samples/sec	accuracy=0.308636	lr=0.393609
Epoch[14] Batch [399]	Speed: 5676.292061 samples/sec	accuracy=0.308762	lr=0.393554
Epoch[14] Batch [449]	Speed: 4405.951631 samples/sec	accuracy=0.308856	lr=0.393499
Epoch[14] Batch [499]	Speed: 5088.053127 samples/sec	accuracy=0.308598	lr=0.393444
Epoch[14] Batch [549]	Speed: 5871.740064 samples/sec	accuracy=0.308558	lr=0.393388
Epoch[14] Batch [599]	Speed: 5251.263689 samples/sec	accuracy=0.308498	lr=0.393332
Epoch[14] Batch [649]	Speed: 5956.544239 samples/sec	accuracy=0.308430	lr=0.393276
Epoch[14] Batch [699]	Speed: 5348.205171 samples/sec	accuracy=0.308273	lr=0.393220
Epoch[14] Batch [749]	Speed: 5535.696448 samples/sec	accuracy=0.308122	lr=0.393163
Epoch[14] Batch [799]	Speed: 5154.427953 samples/sec	accuracy=0.307899	lr=0.393107
Epoch[14] Batch [849]	Speed: 5184.549640 samples/sec	accuracy=0.307956	lr=0.393050
Epoch[14] Batch [899]	Speed: 4205.434850 samples/sec	accuracy=0.307961	lr=0.392993
Epoch[14] Batch [949]	Speed: 5019.574967 samples/sec	accuracy=0.307995	lr=0.392935
Epoch[14] Batch [999]	Speed: 5969.831281 samples/sec	accuracy=0.308005	lr=0.392877
Epoch[14] Batch [1049]	Speed: 5227.500988 samples/sec	accuracy=0.308113	lr=0.392820
Epoch[14] Batch [1099]	Speed: 5539.397197 samples/sec	accuracy=0.308066	lr=0.392762
Epoch[14] Batch [1149]	Speed: 5379.183601 samples/sec	accuracy=0.308077	lr=0.392703
Epoch[14] Batch [1199]	Speed: 5415.091545 samples/sec	accuracy=0.308033	lr=0.392645
Epoch[14] Batch [1249]	Speed: 5935.790013 samples/sec	accuracy=0.308020	lr=0.392586
[Epoch 14] training: accuracy=0.307995
[Epoch 14] speed: 5316 samples/sec	time cost: 250.945180
[Epoch 14] validation: err-top1=0.754923 err-top5=0.521225
Epoch[15] Batch [49]	Speed: 5379.323736 samples/sec	accuracy=0.311445	lr=0.392526
Epoch[15] Batch [99]	Speed: 4949.095368 samples/sec	accuracy=0.310479	lr=0.392466
Epoch[15] Batch [149]	Speed: 4662.209238 samples/sec	accuracy=0.311413	lr=0.392407
Epoch[15] Batch [199]	Speed: 4454.060549 samples/sec	accuracy=0.310405	lr=0.392347
Epoch[15] Batch [249]	Speed: 5892.210033 samples/sec	accuracy=0.310301	lr=0.392287
Epoch[15] Batch [299]	Speed: 5389.925348 samples/sec	accuracy=0.310003	lr=0.392227
Epoch[15] Batch [349]	Speed: 5723.729144 samples/sec	accuracy=0.309866	lr=0.392167
Epoch[15] Batch [399]	Speed: 5391.979556 samples/sec	accuracy=0.310063	lr=0.392106
Epoch[15] Batch [449]	Speed: 5800.845263 samples/sec	accuracy=0.310428	lr=0.392045
Epoch[15] Batch [499]	Speed: 5562.407732 samples/sec	accuracy=0.310137	lr=0.391984
Epoch[15] Batch [549]	Speed: 5842.595783 samples/sec	accuracy=0.310133	lr=0.391923
Epoch[15] Batch [599]	Speed: 5868.305873 samples/sec	accuracy=0.309870	lr=0.391861
Epoch[15] Batch [649]	Speed: 5809.840113 samples/sec	accuracy=0.309962	lr=0.391799
Epoch[15] Batch [699]	Speed: 5441.688144 samples/sec	accuracy=0.309812	lr=0.391737
Epoch[15] Batch [749]	Speed: 5889.891158 samples/sec	accuracy=0.309967	lr=0.391675
Epoch[15] Batch [799]	Speed: 5857.195410 samples/sec	accuracy=0.309976	lr=0.391613
Epoch[15] Batch [849]	Speed: 5519.874045 samples/sec	accuracy=0.310072	lr=0.391550
Epoch[15] Batch [899]	Speed: 5537.888998 samples/sec	accuracy=0.309920	lr=0.391487
Epoch[15] Batch [949]	Speed: 5453.673259 samples/sec	accuracy=0.309827	lr=0.391424
Epoch[15] Batch [999]	Speed: 5839.330867 samples/sec	accuracy=0.309839	lr=0.391361
Epoch[15] Batch [1049]	Speed: 5846.712509 samples/sec	accuracy=0.309916	lr=0.391297
Epoch[15] Batch [1099]	Speed: 5838.464532 samples/sec	accuracy=0.309882	lr=0.391233
Epoch[15] Batch [1149]	Speed: 5613.676783 samples/sec	accuracy=0.310012	lr=0.391169
Epoch[15] Batch [1199]	Speed: 5720.806450 samples/sec	accuracy=0.310103	lr=0.391105
Epoch[15] Batch [1249]	Speed: 5313.114155 samples/sec	accuracy=0.310020	lr=0.391040
[Epoch 15] training: accuracy=0.310030
[Epoch 15] speed: 5509 samples/sec	time cost: 241.661558
[Epoch 15] validation: err-top1=0.803731 err-top5=0.595026
Epoch[16] Batch [49]	Speed: 5515.697211 samples/sec	accuracy=0.317246	lr=0.390974
Epoch[16] Batch [99]	Speed: 5736.654864 samples/sec	accuracy=0.315322	lr=0.390909
Epoch[16] Batch [149]	Speed: 5649.484993 samples/sec	accuracy=0.316061	lr=0.390844
Epoch[16] Batch [199]	Speed: 5181.866808 samples/sec	accuracy=0.315952	lr=0.390779
Epoch[16] Batch [249]	Speed: 5783.428669 samples/sec	accuracy=0.315215	lr=0.390713
Epoch[16] Batch [299]	Speed: 5556.652902 samples/sec	accuracy=0.314655	lr=0.390647
Epoch[16] Batch [349]	Speed: 5573.123343 samples/sec	accuracy=0.314311	lr=0.390581
Epoch[16] Batch [399]	Speed: 4872.087718 samples/sec	accuracy=0.314175	lr=0.390515
Epoch[16] Batch [449]	Speed: 5601.588730 samples/sec	accuracy=0.313854	lr=0.390448
Epoch[16] Batch [499]	Speed: 5850.617619 samples/sec	accuracy=0.314041	lr=0.390381
Epoch[16] Batch [549]	Speed: 5502.110682 samples/sec	accuracy=0.314007	lr=0.390314
Epoch[16] Batch [599]	Speed: 5544.442584 samples/sec	accuracy=0.314023	lr=0.390247
Epoch[16] Batch [649]	Speed: 5346.870363 samples/sec	accuracy=0.313885	lr=0.390180
Epoch[16] Batch [699]	Speed: 5993.991025 samples/sec	accuracy=0.313683	lr=0.390112
Epoch[16] Batch [749]	Speed: 5938.997166 samples/sec	accuracy=0.313625	lr=0.390044
Epoch[16] Batch [799]	Speed: 5888.763975 samples/sec	accuracy=0.313730	lr=0.389976
Epoch[16] Batch [849]	Speed: 5571.895823 samples/sec	accuracy=0.313844	lr=0.389907
Epoch[16] Batch [899]	Speed: 5812.947502 samples/sec	accuracy=0.313717	lr=0.389839
Epoch[16] Batch [949]	Speed: 5825.927787 samples/sec	accuracy=0.313540	lr=0.389770
Epoch[16] Batch [999]	Speed: 5869.355296 samples/sec	accuracy=0.313553	lr=0.389701
Epoch[16] Batch [1049]	Speed: 5925.888080 samples/sec	accuracy=0.313501	lr=0.389632
Epoch[16] Batch [1099]	Speed: 5657.823028 samples/sec	accuracy=0.313489	lr=0.389562
Epoch[16] Batch [1149]	Speed: 5611.457694 samples/sec	accuracy=0.313483	lr=0.389492
Epoch[16] Batch [1199]	Speed: 5732.985146 samples/sec	accuracy=0.313459	lr=0.389422
Epoch[16] Batch [1249]	Speed: 5370.120925 samples/sec	accuracy=0.313475	lr=0.389352
[Epoch 16] training: accuracy=0.313470
[Epoch 16] speed: 5616 samples/sec	time cost: 237.650997
[Epoch 16] validation: err-top1=0.760025 err-top5=0.529636
Epoch[17] Batch [49]	Speed: 4999.914548 samples/sec	accuracy=0.321914	lr=0.389280
Epoch[17] Batch [99]	Speed: 5948.786604 samples/sec	accuracy=0.318887	lr=0.389210
Epoch[17] Batch [149]	Speed: 5571.297371 samples/sec	accuracy=0.318887	lr=0.389139
Epoch[17] Batch [199]	Speed: 5841.090841 samples/sec	accuracy=0.318872	lr=0.389068
Epoch[17] Batch [249]	Speed: 5910.029072 samples/sec	accuracy=0.317656	lr=0.388996
Epoch[17] Batch [299]	Speed: 5494.664692 samples/sec	accuracy=0.317484	lr=0.388925
Epoch[17] Batch [349]	Speed: 5593.021294 samples/sec	accuracy=0.317121	lr=0.388853
Epoch[17] Batch [399]	Speed: 5578.101570 samples/sec	accuracy=0.317061	lr=0.388781
Epoch[17] Batch [449]	Speed: 5524.861883 samples/sec	accuracy=0.316558	lr=0.388709
Epoch[17] Batch [499]	Speed: 5779.970338 samples/sec	accuracy=0.316424	lr=0.388636
Epoch[17] Batch [549]	Speed: 6106.957537 samples/sec	accuracy=0.316273	lr=0.388564
Epoch[17] Batch [599]	Speed: 5542.714608 samples/sec	accuracy=0.316234	lr=0.388491
Epoch[17] Batch [649]	Speed: 4982.332128 samples/sec	accuracy=0.315688	lr=0.388418
Epoch[17] Batch [699]	Speed: 5710.600487 samples/sec	accuracy=0.315472	lr=0.388344
Epoch[17] Batch [749]	Speed: 5761.915331 samples/sec	accuracy=0.315201	lr=0.388271
Epoch[17] Batch [799]	Speed: 5782.424070 samples/sec	accuracy=0.315303	lr=0.388197
Epoch[17] Batch [849]	Speed: 5855.856345 samples/sec	accuracy=0.315031	lr=0.388123
Epoch[17] Batch [899]	Speed: 5727.546647 samples/sec	accuracy=0.314697	lr=0.388049
Epoch[17] Batch [949]	Speed: 6096.988099 samples/sec	accuracy=0.314618	lr=0.387974
Epoch[17] Batch [999]	Speed: 5849.792549 samples/sec	accuracy=0.314610	lr=0.387900
Epoch[17] Batch [1049]	Speed: 6081.275767 samples/sec	accuracy=0.314370	lr=0.387825
Epoch[17] Batch [1099]	Speed: 5822.761986 samples/sec	accuracy=0.314276	lr=0.387750
Epoch[17] Batch [1149]	Speed: 5554.659254 samples/sec	accuracy=0.314338	lr=0.387674
Epoch[17] Batch [1199]	Speed: 6080.724743 samples/sec	accuracy=0.314458	lr=0.387599
Epoch[17] Batch [1249]	Speed: 6020.977239 samples/sec	accuracy=0.314428	lr=0.387523
[Epoch 17] training: accuracy=0.314424
[Epoch 17] speed: 5706 samples/sec	time cost: 233.954918
[Epoch 17] validation: err-top1=0.719727 err-top5=0.473043
Epoch[18] Batch [49]	Speed: 5558.322683 samples/sec	accuracy=0.318477	lr=0.387445
Epoch[18] Batch [99]	Speed: 5856.489865 samples/sec	accuracy=0.319961	lr=0.387369
Epoch[18] Batch [149]	Speed: 5533.119828 samples/sec	accuracy=0.319720	lr=0.387293
Epoch[18] Batch [199]	Speed: 5788.019595 samples/sec	accuracy=0.318428	lr=0.387216
Epoch[18] Batch [249]	Speed: 5849.221337 samples/sec	accuracy=0.317934	lr=0.387139
Epoch[18] Batch [299]	Speed: 5745.294810 samples/sec	accuracy=0.317370	lr=0.387062
Epoch[18] Batch [349]	Speed: 5544.548373 samples/sec	accuracy=0.317561	lr=0.386984
Epoch[18] Batch [399]	Speed: 5711.561293 samples/sec	accuracy=0.317971	lr=0.386907
Epoch[18] Batch [449]	Speed: 5837.866011 samples/sec	accuracy=0.317917	lr=0.386829
Epoch[18] Batch [499]	Speed: 5879.266698 samples/sec	accuracy=0.317721	lr=0.386751
Epoch[18] Batch [549]	Speed: 5271.342935 samples/sec	accuracy=0.317264	lr=0.386673
Epoch[18] Batch [599]	Speed: 5808.260879 samples/sec	accuracy=0.317508	lr=0.386594
Epoch[18] Batch [649]	Speed: 5601.233256 samples/sec	accuracy=0.317450	lr=0.386515
Epoch[18] Batch [699]	Speed: 5929.912807 samples/sec	accuracy=0.317239	lr=0.386436
Epoch[18] Batch [749]	Speed: 6011.427336 samples/sec	accuracy=0.316999	lr=0.386357
Epoch[18] Batch [799]	Speed: 5977.542440 samples/sec	accuracy=0.316818	lr=0.386278
Epoch[18] Batch [849]	Speed: 5562.024657 samples/sec	accuracy=0.316905	lr=0.386198
Epoch[18] Batch [899]	Speed: 5803.409442 samples/sec	accuracy=0.316679	lr=0.386119
Epoch[18] Batch [949]	Speed: 5884.690074 samples/sec	accuracy=0.316557	lr=0.386038
Epoch[18] Batch [999]	Speed: 5609.339110 samples/sec	accuracy=0.316736	lr=0.385958
Epoch[18] Batch [1049]	Speed: 5847.697374 samples/sec	accuracy=0.316706	lr=0.385878
Epoch[18] Batch [1099]	Speed: 5712.415902 samples/sec	accuracy=0.316723	lr=0.385797
Epoch[18] Batch [1149]	Speed: 6064.015676 samples/sec	accuracy=0.316715	lr=0.385716
Epoch[18] Batch [1199]	Speed: 5760.478855 samples/sec	accuracy=0.316610	lr=0.385635
Epoch[18] Batch [1249]	Speed: 6149.603446 samples/sec	accuracy=0.316693	lr=0.385554
[Epoch 18] training: accuracy=0.316698
[Epoch 18] speed: 5759 samples/sec	time cost: 232.405369
[Epoch 18] validation: err-top1=0.757533 err-top5=0.529297
Epoch[19] Batch [49]	Speed: 6028.520433 samples/sec	accuracy=0.324551	lr=0.385470
Epoch[19] Batch [99]	Speed: 5597.225507 samples/sec	accuracy=0.322354	lr=0.385388
Epoch[19] Batch [149]	Speed: 5996.342044 samples/sec	accuracy=0.320859	lr=0.385306
Epoch[19] Batch [199]	Speed: 5586.300845 samples/sec	accuracy=0.320439	lr=0.385224
Epoch[19] Batch [249]	Speed: 6030.970944 samples/sec	accuracy=0.319844	lr=0.385142
Epoch[19] Batch [299]	Speed: 5773.528128 samples/sec	accuracy=0.319795	lr=0.385059
Epoch[19] Batch [349]	Speed: 5893.766022 samples/sec	accuracy=0.319358	lr=0.384976
Epoch[19] Batch [399]	Speed: 5675.249342 samples/sec	accuracy=0.319253	lr=0.384893
Epoch[19] Batch [449]	Speed: 5562.176353 samples/sec	accuracy=0.319219	lr=0.384810
Epoch[19] Batch [499]	Speed: 5872.678936 samples/sec	accuracy=0.319033	lr=0.384726
Epoch[19] Batch [549]	Speed: 5340.817635 samples/sec	accuracy=0.318887	lr=0.384642
Epoch[19] Batch [599]	Speed: 5928.374826 samples/sec	accuracy=0.318807	lr=0.384558
Epoch[19] Batch [649]	Speed: 5569.929507 samples/sec	accuracy=0.318609	lr=0.384474
Epoch[19] Batch [699]	Speed: 5674.925399 samples/sec	accuracy=0.318807	lr=0.384389
Epoch[19] Batch [749]	Speed: 5649.533594 samples/sec	accuracy=0.318676	lr=0.384305
Epoch[19] Batch [799]	Speed: 6223.992420 samples/sec	accuracy=0.318503	lr=0.384220
Epoch[19] Batch [849]	Speed: 5981.260972 samples/sec	accuracy=0.318486	lr=0.384135
Epoch[19] Batch [899]	Speed: 5985.162448 samples/sec	accuracy=0.318278	lr=0.384049
Epoch[19] Batch [949]	Speed: 5650.019346 samples/sec	accuracy=0.318507	lr=0.383964
Epoch[19] Batch [999]	Speed: 6015.568292 samples/sec	accuracy=0.318483	lr=0.383878
Epoch[19] Batch [1049]	Speed: 6057.257065 samples/sec	accuracy=0.318578	lr=0.383792
Epoch[19] Batch [1099]	Speed: 5892.806813 samples/sec	accuracy=0.318543	lr=0.383706
Epoch[19] Batch [1149]	Speed: 5656.115435 samples/sec	accuracy=0.318535	lr=0.383619
Epoch[19] Batch [1199]	Speed: 5592.730556 samples/sec	accuracy=0.318398	lr=0.383533
Epoch[19] Batch [1249]	Speed: 5539.438492 samples/sec	accuracy=0.318453	lr=0.383446
[Epoch 19] training: accuracy=0.318475
[Epoch 19] speed: 5775 samples/sec	time cost: 231.390515
[Epoch 19] validation: err-top1=0.825833 err-top5=0.627352
Epoch[20] Batch [49]	Speed: 5080.103848 samples/sec	accuracy=0.323477	lr=0.383357
Epoch[20] Batch [99]	Speed: 5729.536110 samples/sec	accuracy=0.321406	lr=0.383270
Epoch[20] Batch [149]	Speed: 5212.417264 samples/sec	accuracy=0.321400	lr=0.383182
Epoch[20] Batch [199]	Speed: 5991.936419 samples/sec	accuracy=0.321880	lr=0.383094
Epoch[20] Batch [249]	Speed: 6041.770511 samples/sec	accuracy=0.321441	lr=0.383006
Epoch[20] Batch [299]	Speed: 5563.854789 samples/sec	accuracy=0.320960	lr=0.382918
Epoch[20] Batch [349]	Speed: 6006.519670 samples/sec	accuracy=0.320868	lr=0.382830
Epoch[20] Batch [399]	Speed: 5568.740795 samples/sec	accuracy=0.320591	lr=0.382741
Epoch[20] Batch [449]	Speed: 5745.334928 samples/sec	accuracy=0.320614	lr=0.382652
Epoch[20] Batch [499]	Speed: 5463.782904 samples/sec	accuracy=0.321342	lr=0.382563
Epoch[20] Batch [549]	Speed: 6182.828948 samples/sec	accuracy=0.321049	lr=0.382474
Epoch[20] Batch [599]	Speed: 5616.535961 samples/sec	accuracy=0.320841	lr=0.382384
Epoch[20] Batch [649]	Speed: 6077.020154 samples/sec	accuracy=0.320829	lr=0.382295
Epoch[20] Batch [699]	Speed: 5775.959614 samples/sec	accuracy=0.320532	lr=0.382205
Epoch[20] Batch [749]	Speed: 6116.159115 samples/sec	accuracy=0.320298	lr=0.382115
Epoch[20] Batch [799]	Speed: 5996.171434 samples/sec	accuracy=0.320233	lr=0.382024
Epoch[20] Batch [849]	Speed: 5942.669361 samples/sec	accuracy=0.320156	lr=0.381934
Epoch[20] Batch [899]	Speed: 6011.555902 samples/sec	accuracy=0.320162	lr=0.381843
Epoch[20] Batch [949]	Speed: 5593.686783 samples/sec	accuracy=0.320006	lr=0.381752
Epoch[20] Batch [999]	Speed: 5867.544423 samples/sec	accuracy=0.320118	lr=0.381661
Epoch[20] Batch [1049]	Speed: 5836.799104 samples/sec	accuracy=0.320088	lr=0.381569
Epoch[20] Batch [1099]	Speed: 5686.313016 samples/sec	accuracy=0.320130	lr=0.381477
Epoch[20] Batch [1149]	Speed: 5541.323423 samples/sec	accuracy=0.320089	lr=0.381385
Epoch[20] Batch [1199]	Speed: 5508.321240 samples/sec	accuracy=0.319902	lr=0.381293
Epoch[20] Batch [1249]	Speed: 5761.173047 samples/sec	accuracy=0.319811	lr=0.381201
[Epoch 20] training: accuracy=0.319809
[Epoch 20] speed: 5737 samples/sec	time cost: 233.112876
[Epoch 20] validation: err-top1=0.756577 err-top5=0.525949
Epoch[21] Batch [49]	Speed: 5669.289461 samples/sec	accuracy=0.325488	lr=0.381107
Epoch[21] Batch [99]	Speed: 6030.058161 samples/sec	accuracy=0.323525	lr=0.381014
Epoch[21] Batch [149]	Speed: 5757.791136 samples/sec	accuracy=0.323581	lr=0.380921
Epoch[21] Batch [199]	Speed: 5919.647942 samples/sec	accuracy=0.322798	lr=0.380828
Epoch[21] Batch [249]	Speed: 5640.397774 samples/sec	accuracy=0.323223	lr=0.380734
Epoch[21] Batch [299]	Speed: 5565.242169 samples/sec	accuracy=0.323320	lr=0.380641
Epoch[21] Batch [349]	Speed: 6082.725091 samples/sec	accuracy=0.323044	lr=0.380547
Epoch[21] Batch [399]	Speed: 5394.574622 samples/sec	accuracy=0.323291	lr=0.380453
Epoch[21] Batch [449]	Speed: 5662.054502 samples/sec	accuracy=0.323229	lr=0.380359
Epoch[21] Batch [499]	Speed: 5886.460881 samples/sec	accuracy=0.323111	lr=0.380264
Epoch[21] Batch [549]	Speed: 5631.331029 samples/sec	accuracy=0.322956	lr=0.380169
Epoch[21] Batch [599]	Speed: 5879.286335 samples/sec	accuracy=0.323019	lr=0.380074
Epoch[21] Batch [649]	Speed: 5788.027551 samples/sec	accuracy=0.322719	lr=0.379979
Epoch[21] Batch [699]	Speed: 5504.315078 samples/sec	accuracy=0.322567	lr=0.379884
Epoch[21] Batch [749]	Speed: 5955.477284 samples/sec	accuracy=0.322504	lr=0.379788
Epoch[21] Batch [799]	Speed: 5387.651023 samples/sec	accuracy=0.322582	lr=0.379693
Epoch[21] Batch [849]	Speed: 5257.431250 samples/sec	accuracy=0.322503	lr=0.379597
Epoch[21] Batch [899]	Speed: 5262.714369 samples/sec	accuracy=0.322283	lr=0.379501
Epoch[21] Batch [949]	Speed: 5347.037310 samples/sec	accuracy=0.322260	lr=0.379404
Epoch[21] Batch [999]	Speed: 4783.817258 samples/sec	accuracy=0.322340	lr=0.379307
Epoch[21] Batch [1049]	Speed: 4506.788709 samples/sec	accuracy=0.322166	lr=0.379211
Epoch[21] Batch [1099]	Speed: 5687.072129 samples/sec	accuracy=0.322142	lr=0.379114
Epoch[21] Batch [1149]	Speed: 5806.348092 samples/sec	accuracy=0.322247	lr=0.379016
Epoch[21] Batch [1199]	Speed: 5919.512507 samples/sec	accuracy=0.322262	lr=0.378919
Epoch[21] Batch [1249]	Speed: 5336.064147 samples/sec	accuracy=0.322337	lr=0.378821
[Epoch 21] training: accuracy=0.322310
[Epoch 21] speed: 5553 samples/sec	time cost: 240.885373
[Epoch 21] validation: err-top1=0.693618 err-top5=0.440948
Epoch[22] Batch [49]	Speed: 5051.681621 samples/sec	accuracy=0.325645	lr=0.378721
Epoch[22] Batch [99]	Speed: 5538.422874 samples/sec	accuracy=0.326572	lr=0.378623
Epoch[22] Batch [149]	Speed: 6014.392496 samples/sec	accuracy=0.325723	lr=0.378525
Epoch[22] Batch [199]	Speed: 5740.600889 samples/sec	accuracy=0.325264	lr=0.378426
Epoch[22] Batch [249]	Speed: 5130.797871 samples/sec	accuracy=0.325004	lr=0.378328
Epoch[22] Batch [299]	Speed: 5056.708678 samples/sec	accuracy=0.324704	lr=0.378229
Epoch[22] Batch [349]	Speed: 5540.376009 samples/sec	accuracy=0.324657	lr=0.378129
Epoch[22] Batch [399]	Speed: 4681.072246 samples/sec	accuracy=0.324780	lr=0.378030
Epoch[22] Batch [449]	Speed: 5328.443934 samples/sec	accuracy=0.324379	lr=0.377930
Epoch[22] Batch [499]	Speed: 4871.621415 samples/sec	accuracy=0.324203	lr=0.377831
Epoch[22] Batch [549]	Speed: 4895.453401 samples/sec	accuracy=0.324249	lr=0.377731
Epoch[22] Batch [599]	Speed: 5610.305275 samples/sec	accuracy=0.324064	lr=0.377630
Epoch[22] Batch [649]	Speed: 4911.146670 samples/sec	accuracy=0.323726	lr=0.377530
Epoch[22] Batch [699]	Speed: 5780.847410 samples/sec	accuracy=0.323626	lr=0.377429
Epoch[22] Batch [749]	Speed: 4951.210317 samples/sec	accuracy=0.323589	lr=0.377328
Epoch[22] Batch [799]	Speed: 5209.779078 samples/sec	accuracy=0.323397	lr=0.377227
Epoch[22] Batch [849]	Speed: 4392.128199 samples/sec	accuracy=0.323336	lr=0.377126
Epoch[22] Batch [899]	Speed: 5131.574324 samples/sec	accuracy=0.323069	lr=0.377024
Epoch[22] Batch [949]	Speed: 5876.131914 samples/sec	accuracy=0.322925	lr=0.376923
Epoch[22] Batch [999]	Speed: 5104.059913 samples/sec	accuracy=0.322774	lr=0.376821
Epoch[22] Batch [1049]	Speed: 4898.646848 samples/sec	accuracy=0.322826	lr=0.376719
Epoch[22] Batch [1099]	Speed: 4945.514291 samples/sec	accuracy=0.322854	lr=0.376616
Epoch[22] Batch [1149]	Speed: 5546.071946 samples/sec	accuracy=0.322855	lr=0.376514
Epoch[22] Batch [1199]	Speed: 5200.799110 samples/sec	accuracy=0.322900	lr=0.376411
Epoch[22] Batch [1249]	Speed: 5273.336479 samples/sec	accuracy=0.322910	lr=0.376308
[Epoch 22] training: accuracy=0.322899
[Epoch 22] speed: 5191 samples/sec	time cost: 257.048055
[Epoch 22] validation: err-top1=0.715402 err-top5=0.470185
Epoch[23] Batch [49]	Speed: 5775.692420 samples/sec	accuracy=0.330371	lr=0.376203
Epoch[23] Batch [99]	Speed: 5388.933653 samples/sec	accuracy=0.328428	lr=0.376099
Epoch[23] Batch [149]	Speed: 5672.298625 samples/sec	accuracy=0.327630	lr=0.375996
Epoch[23] Batch [199]	Speed: 5570.067044 samples/sec	accuracy=0.326982	lr=0.375892
Epoch[23] Batch [249]	Speed: 5213.757422 samples/sec	accuracy=0.326879	lr=0.375788
Epoch[23] Batch [299]	Speed: 5874.543278 samples/sec	accuracy=0.326201	lr=0.375683
Epoch[23] Batch [349]	Speed: 5314.791885 samples/sec	accuracy=0.325530	lr=0.375579
Epoch[23] Batch [399]	Speed: 5484.637351 samples/sec	accuracy=0.325229	lr=0.375474
Epoch[23] Batch [449]	Speed: 5420.575764 samples/sec	accuracy=0.325200	lr=0.375369
Epoch[23] Batch [499]	Speed: 5211.355624 samples/sec	accuracy=0.324996	lr=0.375264
Epoch[23] Batch [549]	Speed: 4678.984567 samples/sec	accuracy=0.324762	lr=0.375159
Epoch[23] Batch [599]	Speed: 5052.676104 samples/sec	accuracy=0.324526	lr=0.375053
Epoch[23] Batch [649]	Speed: 5130.395208 samples/sec	accuracy=0.324603	lr=0.374948
Epoch[23] Batch [699]	Speed: 5925.927326 samples/sec	accuracy=0.324360	lr=0.374842
Epoch[23] Batch [749]	Speed: 5168.780434 samples/sec	accuracy=0.324236	lr=0.374736
Epoch[23] Batch [799]	Speed: 5548.980658 samples/sec	accuracy=0.323998	lr=0.374629
Epoch[23] Batch [849]	Speed: 5353.383988 samples/sec	accuracy=0.324000	lr=0.374523
Epoch[23] Batch [899]	Speed: 5784.148033 samples/sec	accuracy=0.324074	lr=0.374416
Epoch[23] Batch [949]	Speed: 5606.089447 samples/sec	accuracy=0.323787	lr=0.374309
Epoch[23] Batch [999]	Speed: 5592.621902 samples/sec	accuracy=0.323894	lr=0.374202
Epoch[23] Batch [1049]	Speed: 5889.049162 samples/sec	accuracy=0.323928	lr=0.374095
Epoch[23] Batch [1099]	Speed: 5613.497759 samples/sec	accuracy=0.323805	lr=0.373987
Epoch[23] Batch [1149]	Speed: 5604.373590 samples/sec	accuracy=0.323977	lr=0.373879
Epoch[23] Batch [1199]	Speed: 5271.863408 samples/sec	accuracy=0.323997	lr=0.373771
Epoch[23] Batch [1249]	Speed: 5620.008209 samples/sec	accuracy=0.323984	lr=0.373663
[Epoch 23] training: accuracy=0.323984
[Epoch 23] speed: 5448 samples/sec	time cost: 245.096916
[Epoch 23] validation: err-top1=0.795614 err-top5=0.582662
Epoch[24] Batch [49]	Speed: 5508.707693 samples/sec	accuracy=0.331406	lr=0.373552
Epoch[24] Batch [99]	Speed: 5689.476696 samples/sec	accuracy=0.328809	lr=0.373444
Epoch[24] Batch [149]	Speed: 5786.438011 samples/sec	accuracy=0.327500	lr=0.373335
Epoch[24] Batch [199]	Speed: 5474.217407 samples/sec	accuracy=0.327974	lr=0.373226
Epoch[24] Batch [249]	Speed: 5420.999950 samples/sec	accuracy=0.326789	lr=0.373117
Epoch[24] Batch [299]	Speed: 5413.731197 samples/sec	accuracy=0.326234	lr=0.373007
Epoch[24] Batch [349]	Speed: 5682.641838 samples/sec	accuracy=0.326074	lr=0.372897
Epoch[24] Batch [399]	Speed: 5537.919131 samples/sec	accuracy=0.326155	lr=0.372788
Epoch[24] Batch [449]	Speed: 5681.197719 samples/sec	accuracy=0.325751	lr=0.372678
Epoch[24] Batch [499]	Speed: 5669.565761 samples/sec	accuracy=0.325639	lr=0.372567
Epoch[24] Batch [549]	Speed: 5637.715388 samples/sec	accuracy=0.325639	lr=0.372457
Epoch[24] Batch [599]	Speed: 5348.048273 samples/sec	accuracy=0.325723	lr=0.372346
Epoch[24] Batch [649]	Speed: 4861.609905 samples/sec	accuracy=0.325470	lr=0.372235
Epoch[24] Batch [699]	Speed: 5363.569881 samples/sec	accuracy=0.325485	lr=0.372124
Epoch[24] Batch [749]	Speed: 5171.106656 samples/sec	accuracy=0.325591	lr=0.372013
Epoch[24] Batch [799]	Speed: 5703.740870 samples/sec	accuracy=0.325695	lr=0.371901
Epoch[24] Batch [849]	Speed: 5987.236781 samples/sec	accuracy=0.325508	lr=0.371790
Epoch[24] Batch [899]	Speed: 5321.962597 samples/sec	accuracy=0.325505	lr=0.371678
Epoch[24] Batch [949]	Speed: 5908.312331 samples/sec	accuracy=0.325238	lr=0.371565
Epoch[24] Batch [999]	Speed: 5387.555192 samples/sec	accuracy=0.325246	lr=0.371453
Epoch[24] Batch [1049]	Speed: 5934.416087 samples/sec	accuracy=0.325233	lr=0.371341
Epoch[24] Batch [1099]	Speed: 5605.068848 samples/sec	accuracy=0.325289	lr=0.371228
Epoch[24] Batch [1149]	Speed: 5425.590879 samples/sec	accuracy=0.325457	lr=0.371115
Epoch[24] Batch [1199]	Speed: 5279.853930 samples/sec	accuracy=0.325427	lr=0.371002
Epoch[24] Batch [1249]	Speed: 5283.768837 samples/sec	accuracy=0.325466	lr=0.370888
[Epoch 24] training: accuracy=0.325463
[Epoch 24] speed: 5504 samples/sec	time cost: 243.111322
[Epoch 24] validation: err-top1=0.772003 err-top5=0.545639
Epoch[25] Batch [49]	Speed: 5906.292316 samples/sec	accuracy=0.328184	lr=0.370773
Epoch[25] Batch [99]	Speed: 5599.192897 samples/sec	accuracy=0.330771	lr=0.370659
Epoch[25] Batch [149]	Speed: 5575.714201 samples/sec	accuracy=0.330020	lr=0.370545
Epoch[25] Batch [199]	Speed: 5938.270463 samples/sec	accuracy=0.329629	lr=0.370431
Epoch[25] Batch [249]	Speed: 6173.277820 samples/sec	accuracy=0.329648	lr=0.370316
Epoch[25] Batch [299]	Speed: 5294.366160 samples/sec	accuracy=0.329098	lr=0.370202
Epoch[25] Batch [349]	Speed: 5938.464069 samples/sec	accuracy=0.328605	lr=0.370087
Epoch[25] Batch [399]	Speed: 5681.496676 samples/sec	accuracy=0.328542	lr=0.369972
Epoch[25] Batch [449]	Speed: 5039.631395 samples/sec	accuracy=0.327988	lr=0.369857
Epoch[25] Batch [499]	Speed: 5761.825201 samples/sec	accuracy=0.327840	lr=0.369741
Epoch[25] Batch [549]	Speed: 5686.016112 samples/sec	accuracy=0.328015	lr=0.369626
Epoch[25] Batch [599]	Speed: 5716.843600 samples/sec	accuracy=0.327677	lr=0.369510
Epoch[25] Batch [649]	Speed: 5437.568723 samples/sec	accuracy=0.327686	lr=0.369394
Epoch[25] Batch [699]	Speed: 5987.199056 samples/sec	accuracy=0.327589	lr=0.369278
Epoch[25] Batch [749]	Speed: 5450.233439 samples/sec	accuracy=0.327337	lr=0.369161
Epoch[25] Batch [799]	Speed: 5987.333433 samples/sec	accuracy=0.327119	lr=0.369045
Epoch[25] Batch [849]	Speed: 5730.804562 samples/sec	accuracy=0.326988	lr=0.368928
Epoch[25] Batch [899]	Speed: 5899.480677 samples/sec	accuracy=0.327141	lr=0.368811
Epoch[25] Batch [949]	Speed: 5801.447188 samples/sec	accuracy=0.327019	lr=0.368694
Epoch[25] Batch [999]	Speed: 5734.937191 samples/sec	accuracy=0.327196	lr=0.368576
Epoch[25] Batch [1049]	Speed: 5894.636551 samples/sec	accuracy=0.327000	lr=0.368459
Epoch[25] Batch [1099]	Speed: 5457.139334 samples/sec	accuracy=0.326979	lr=0.368341
Epoch[25] Batch [1149]	Speed: 5336.962996 samples/sec	accuracy=0.326855	lr=0.368223
Epoch[25] Batch [1199]	Speed: 5887.950713 samples/sec	accuracy=0.326600	lr=0.368105
Epoch[25] Batch [1249]	Speed: 5970.103960 samples/sec	accuracy=0.326517	lr=0.367986
[Epoch 25] training: accuracy=0.326537
[Epoch 25] speed: 5695 samples/sec	time cost: 234.754978
[Epoch 25] validation: err-top1=0.745097 err-top5=0.504464
Epoch[26] Batch [49]	Speed: 5346.786626 samples/sec	accuracy=0.332461	lr=0.367865
Epoch[26] Batch [99]	Speed: 5320.422956 samples/sec	accuracy=0.331543	lr=0.367747
Epoch[26] Batch [149]	Speed: 5185.701313 samples/sec	accuracy=0.331458	lr=0.367628
Epoch[26] Batch [199]	Speed: 5341.386061 samples/sec	accuracy=0.330200	lr=0.367508
Epoch[26] Batch [249]	Speed: 5648.787589 samples/sec	accuracy=0.330281	lr=0.367389
Epoch[26] Batch [299]	Speed: 5840.661589 samples/sec	accuracy=0.329671	lr=0.367269
Epoch[26] Batch [349]	Speed: 5840.960248 samples/sec	accuracy=0.328979	lr=0.367150
Epoch[26] Batch [399]	Speed: 5979.096219 samples/sec	accuracy=0.328713	lr=0.367030
Epoch[26] Batch [449]	Speed: 5483.702779 samples/sec	accuracy=0.328559	lr=0.366909
Epoch[26] Batch [499]	Speed: 5436.921966 samples/sec	accuracy=0.328488	lr=0.366789
Epoch[26] Batch [549]	Speed: 5828.944825 samples/sec	accuracy=0.328478	lr=0.366668
Epoch[26] Batch [599]	Speed: 5943.269992 samples/sec	accuracy=0.328737	lr=0.366548
Epoch[26] Batch [649]	Speed: 5695.613559 samples/sec	accuracy=0.328173	lr=0.366426
Epoch[26] Batch [699]	Speed: 5783.250958 samples/sec	accuracy=0.328168	lr=0.366305
Epoch[26] Batch [749]	Speed: 5795.812701 samples/sec	accuracy=0.328178	lr=0.366184
Epoch[26] Batch [799]	Speed: 5531.715642 samples/sec	accuracy=0.328302	lr=0.366062
Epoch[26] Batch [849]	Speed: 5413.744845 samples/sec	accuracy=0.328285	lr=0.365940
Epoch[26] Batch [899]	Speed: 6109.436091 samples/sec	accuracy=0.328172	lr=0.365819
Epoch[26] Batch [949]	Speed: 6092.287001 samples/sec	accuracy=0.328095	lr=0.365696
Epoch[26] Batch [999]	Speed: 5783.267935 samples/sec	accuracy=0.328171	lr=0.365574
Epoch[26] Batch [1049]	Speed: 6010.686334 samples/sec	accuracy=0.328249	lr=0.365451
Epoch[26] Batch [1099]	Speed: 5444.974012 samples/sec	accuracy=0.328167	lr=0.365329
Epoch[26] Batch [1149]	Speed: 5883.965960 samples/sec	accuracy=0.327869	lr=0.365206
Epoch[26] Batch [1199]	Speed: 5711.674619 samples/sec	accuracy=0.327742	lr=0.365082
Epoch[26] Batch [1249]	Speed: 5709.106153 samples/sec	accuracy=0.327640	lr=0.364959
[Epoch 26] training: accuracy=0.327654
[Epoch 26] speed: 5669 samples/sec	time cost: 235.882235
[Epoch 26] validation: err-top1=0.823860 err-top5=0.624382
Epoch[27] Batch [49]	Speed: 5653.763241 samples/sec	accuracy=0.336738	lr=0.364833
Epoch[27] Batch [99]	Speed: 5704.872588 samples/sec	accuracy=0.335059	lr=0.364709
Epoch[27] Batch [149]	Speed: 5670.309778 samples/sec	accuracy=0.332910	lr=0.364585
Epoch[27] Batch [199]	Speed: 5518.918490 samples/sec	accuracy=0.331636	lr=0.364461
Epoch[27] Batch [249]	Speed: 5764.897832 samples/sec	accuracy=0.331203	lr=0.364337
Epoch[27] Batch [299]	Speed: 5521.586529 samples/sec	accuracy=0.331042	lr=0.364212
Epoch[27] Batch [349]	Speed: 6008.262026 samples/sec	accuracy=0.331035	lr=0.364087
Epoch[27] Batch [399]	Speed: 5595.731150 samples/sec	accuracy=0.331052	lr=0.363962
Epoch[27] Batch [449]	Speed: 5784.318476 samples/sec	accuracy=0.330762	lr=0.363837
Epoch[27] Batch [499]	Speed: 5208.732789 samples/sec	accuracy=0.330395	lr=0.363712
Epoch[27] Batch [549]	Speed: 5158.873736 samples/sec	accuracy=0.330471	lr=0.363586
Epoch[27] Batch [599]	Speed: 5458.745671 samples/sec	accuracy=0.330350	lr=0.363461
Epoch[27] Batch [649]	Speed: 5306.230118 samples/sec	accuracy=0.330101	lr=0.363335
Epoch[27] Batch [699]	Speed: 5637.714796 samples/sec	accuracy=0.330361	lr=0.363209
Epoch[27] Batch [749]	Speed: 5607.410120 samples/sec	accuracy=0.330135	lr=0.363082
Epoch[27] Batch [799]	Speed: 5585.533527 samples/sec	accuracy=0.329950	lr=0.362956
Epoch[27] Batch [849]	Speed: 5186.149275 samples/sec	accuracy=0.329881	lr=0.362829
Epoch[27] Batch [899]	Speed: 5192.816859 samples/sec	accuracy=0.329630	lr=0.362702
Epoch[27] Batch [949]	Speed: 5354.093382 samples/sec	accuracy=0.329524	lr=0.362575
Epoch[27] Batch [999]	Speed: 5115.110996 samples/sec	accuracy=0.329544	lr=0.362448
Epoch[27] Batch [1049]	Speed: 5485.878432 samples/sec	accuracy=0.329535	lr=0.362320
Epoch[27] Batch [1099]	Speed: 4819.763483 samples/sec	accuracy=0.329558	lr=0.362193
Epoch[27] Batch [1149]	Speed: 5052.773350 samples/sec	accuracy=0.329510	lr=0.362065
Epoch[27] Batch [1199]	Speed: 5815.421445 samples/sec	accuracy=0.329322	lr=0.361937
Epoch[27] Batch [1249]	Speed: 4872.947391 samples/sec	accuracy=0.329398	lr=0.361809
[Epoch 27] training: accuracy=0.329398
[Epoch 27] speed: 5420 samples/sec	time cost: 246.408170
[Epoch 27] validation: err-top1=0.719248 err-top5=0.477240
Epoch[28] Batch [49]	Speed: 4865.584013 samples/sec	accuracy=0.335020	lr=0.361678
Epoch[28] Batch [99]	Speed: 5160.733743 samples/sec	accuracy=0.335049	lr=0.361549
Epoch[28] Batch [149]	Speed: 4479.625804 samples/sec	accuracy=0.333066	lr=0.361420
Epoch[28] Batch [199]	Speed: 5423.370868 samples/sec	accuracy=0.332046	lr=0.361291
Epoch[28] Batch [249]	Speed: 4655.275570 samples/sec	accuracy=0.331914	lr=0.361162
Epoch[28] Batch [299]	Speed: 4858.661860 samples/sec	accuracy=0.331162	lr=0.361032
Epoch[28] Batch [349]	Speed: 5165.827073 samples/sec	accuracy=0.331071	lr=0.360903
Epoch[28] Batch [399]	Speed: 5451.702292 samples/sec	accuracy=0.331072	lr=0.360773
Epoch[28] Batch [449]	Speed: 5016.577721 samples/sec	accuracy=0.331107	lr=0.360643
Epoch[28] Batch [499]	Speed: 4870.492887 samples/sec	accuracy=0.330639	lr=0.360513
Epoch[28] Batch [549]	Speed: 5032.549414 samples/sec	accuracy=0.330755	lr=0.360382
Epoch[28] Batch [599]	Speed: 5692.266967 samples/sec	accuracy=0.330783	lr=0.360252
Epoch[28] Batch [649]	Speed: 5288.767678 samples/sec	accuracy=0.330721	lr=0.360121
Epoch[28] Batch [699]	Speed: 4645.627129 samples/sec	accuracy=0.330841	lr=0.359990
Epoch[28] Batch [749]	Speed: 4553.858042 samples/sec	accuracy=0.330708	lr=0.359859
Epoch[28] Batch [799]	Speed: 4907.602449 samples/sec	accuracy=0.330521	lr=0.359728
Epoch[28] Batch [849]	Speed: 5457.771630 samples/sec	accuracy=0.330446	lr=0.359596
Epoch[28] Batch [899]	Speed: 5795.561966 samples/sec	accuracy=0.330349	lr=0.359465
Epoch[28] Batch [949]	Speed: 5657.674864 samples/sec	accuracy=0.330285	lr=0.359333
Epoch[28] Batch [999]	Speed: 5489.731065 samples/sec	accuracy=0.330233	lr=0.359201
Epoch[28] Batch [1049]	Speed: 5251.931632 samples/sec	accuracy=0.330203	lr=0.359068
Epoch[28] Batch [1099]	Speed: 5561.264569 samples/sec	accuracy=0.330102	lr=0.358936
Epoch[28] Batch [1149]	Speed: 5323.910542 samples/sec	accuracy=0.330087	lr=0.358803
Epoch[28] Batch [1199]	Speed: 6018.543780 samples/sec	accuracy=0.330094	lr=0.358670
Epoch[28] Batch [1249]	Speed: 5481.895600 samples/sec	accuracy=0.330111	lr=0.358537
[Epoch 28] training: accuracy=0.330116
[Epoch 28] speed: 5168 samples/sec	time cost: 257.774496
[Epoch 28] validation: err-top1=0.702109 err-top5=0.454540
Epoch[29] Batch [49]	Speed: 4620.982411 samples/sec	accuracy=0.337168	lr=0.358401
Epoch[29] Batch [99]	Speed: 5362.364370 samples/sec	accuracy=0.336045	lr=0.358268
Epoch[29] Batch [149]	Speed: 5078.377155 samples/sec	accuracy=0.335072	lr=0.358134
Epoch[29] Batch [199]	Speed: 5217.116539 samples/sec	accuracy=0.334546	lr=0.358001
Epoch[29] Batch [249]	Speed: 5717.688829 samples/sec	accuracy=0.334266	lr=0.357867
Epoch[29] Batch [299]	Speed: 5871.109822 samples/sec	accuracy=0.333923	lr=0.357732
Epoch[29] Batch [349]	Speed: 5592.000641 samples/sec	accuracy=0.333906	lr=0.357598
Epoch[29] Batch [399]	Speed: 5430.894217 samples/sec	accuracy=0.334197	lr=0.357464
Epoch[29] Batch [449]	Speed: 4031.825602 samples/sec	accuracy=0.334017	lr=0.357329
Epoch[29] Batch [499]	Speed: 5726.989894 samples/sec	accuracy=0.333648	lr=0.357194
Epoch[29] Batch [549]	Speed: 4895.225864 samples/sec	accuracy=0.333537	lr=0.357059
Epoch[29] Batch [599]	Speed: 4687.425666 samples/sec	accuracy=0.333291	lr=0.356924
Epoch[29] Batch [649]	Speed: 5168.449034 samples/sec	accuracy=0.333191	lr=0.356788
Epoch[29] Batch [699]	Speed: 5286.364478 samples/sec	accuracy=0.333023	lr=0.356652
Epoch[29] Batch [749]	Speed: 4796.595332 samples/sec	accuracy=0.332882	lr=0.356517
Epoch[29] Batch [799]	Speed: 4840.468218 samples/sec	accuracy=0.332889	lr=0.356380
Epoch[29] Batch [849]	Speed: 5037.536083 samples/sec	accuracy=0.332864	lr=0.356244
Epoch[29] Batch [899]	Speed: 4609.837152 samples/sec	accuracy=0.332693	lr=0.356108
Epoch[29] Batch [949]	Speed: 4884.646068 samples/sec	accuracy=0.332571	lr=0.355971
Epoch[29] Batch [999]	Speed: 5269.579115 samples/sec	accuracy=0.332373	lr=0.355834
Epoch[29] Batch [1049]	Speed: 5986.608205 samples/sec	accuracy=0.332214	lr=0.355698
Epoch[29] Batch [1099]	Speed: 5065.065293 samples/sec	accuracy=0.331963	lr=0.355560
Epoch[29] Batch [1149]	Speed: 5113.870019 samples/sec	accuracy=0.331658	lr=0.355423
Epoch[29] Batch [1199]	Speed: 5630.414443 samples/sec	accuracy=0.331588	lr=0.355285
Epoch[29] Batch [1249]	Speed: 5033.647283 samples/sec	accuracy=0.331426	lr=0.355148
[Epoch 29] training: accuracy=0.331433
[Epoch 29] speed: 5114 samples/sec	time cost: 260.330293
[Epoch 29] validation: err-top1=0.681966 err-top5=0.430298
Epoch[30] Batch [49]	Speed: 5799.358304 samples/sec	accuracy=0.333789	lr=0.355007
Epoch[30] Batch [99]	Speed: 5238.105216 samples/sec	accuracy=0.333730	lr=0.354869
Epoch[30] Batch [149]	Speed: 4405.014604 samples/sec	accuracy=0.333301	lr=0.354731
Epoch[30] Batch [199]	Speed: 5311.018185 samples/sec	accuracy=0.333169	lr=0.354592
Epoch[30] Batch [249]	Speed: 5767.125042 samples/sec	accuracy=0.332984	lr=0.354454
Epoch[30] Batch [299]	Speed: 5715.588318 samples/sec	accuracy=0.333789	lr=0.354315
Epoch[30] Batch [349]	Speed: 2276.739162 samples/sec	accuracy=0.333725	lr=0.354176
Epoch[30] Batch [399]	Speed: 683.375053 samples/sec	accuracy=0.334023	lr=0.354037
Epoch[30] Batch [449]	Speed: 3060.110850 samples/sec	accuracy=0.333733	lr=0.353897
Epoch[30] Batch [499]	Speed: 5784.495006 samples/sec	accuracy=0.333646	lr=0.353758
Epoch[30] Batch [549]	Speed: 1725.431715 samples/sec	accuracy=0.333647	lr=0.353618
Epoch[30] Batch [599]	Speed: 5259.049909 samples/sec	accuracy=0.333206	lr=0.353478
Epoch[30] Batch [649]	Speed: 4693.257496 samples/sec	accuracy=0.333185	lr=0.353338
Epoch[30] Batch [699]	Speed: 5345.144383 samples/sec	accuracy=0.333136	lr=0.353198
Epoch[30] Batch [749]	Speed: 5496.247200 samples/sec	accuracy=0.332758	lr=0.353057
Epoch[30] Batch [799]	Speed: 1342.663971 samples/sec	accuracy=0.332672	lr=0.352917
Epoch[30] Batch [849]	Speed: 5206.059837 samples/sec	accuracy=0.332590	lr=0.352776
Epoch[30] Batch [899]	Speed: 3175.933766 samples/sec	accuracy=0.332667	lr=0.352635
Epoch[30] Batch [949]	Speed: 1798.074660 samples/sec	accuracy=0.332563	lr=0.352493
Epoch[30] Batch [999]	Speed: 1182.956490 samples/sec	accuracy=0.332544	lr=0.352352
Epoch[30] Batch [1049]	Speed: 4892.832143 samples/sec	accuracy=0.332351	lr=0.352211
Epoch[30] Batch [1099]	Speed: 1986.354761 samples/sec	accuracy=0.332642	lr=0.352069
Epoch[30] Batch [1149]	Speed: 3516.448003 samples/sec	accuracy=0.332583	lr=0.351927
Epoch[30] Batch [1199]	Speed: 2988.137675 samples/sec	accuracy=0.332549	lr=0.351785
Epoch[30] Batch [1249]	Speed: 3956.152578 samples/sec	accuracy=0.332402	lr=0.351642
[Epoch 30] training: accuracy=0.332403
[Epoch 30] speed: 2741 samples/sec	time cost: 494.965524
[Epoch 30] validation: err-top1=0.785216 err-top5=0.570492
Epoch[31] Batch [49]	Speed: 4979.532858 samples/sec	accuracy=0.336641	lr=0.351497
Epoch[31] Batch [99]	Speed: 5663.009494 samples/sec	accuracy=0.336660	lr=0.351354
Epoch[31] Batch [149]	Speed: 1295.453862 samples/sec	accuracy=0.334486	lr=0.351212
Epoch[31] Batch [199]	Speed: 5153.124300 samples/sec	accuracy=0.333506	lr=0.351069
Epoch[31] Batch [249]	Speed: 1693.110493 samples/sec	accuracy=0.333859	lr=0.350925
Epoch[31] Batch [299]	Speed: 1567.353295 samples/sec	accuracy=0.334049	lr=0.350782
Epoch[31] Batch [349]	Speed: 5484.723080 samples/sec	accuracy=0.333917	lr=0.350639
Epoch[31] Batch [399]	Speed: 1622.208387 samples/sec	accuracy=0.334363	lr=0.350495
Epoch[31] Batch [449]	Speed: 5747.544430 samples/sec	accuracy=0.334351	lr=0.350351
Epoch[31] Batch [499]	Speed: 4035.695416 samples/sec	accuracy=0.333807	lr=0.350207
Epoch[31] Batch [549]	Speed: 2930.379957 samples/sec	accuracy=0.333700	lr=0.350062
Epoch[31] Batch [599]	Speed: 4721.327327 samples/sec	accuracy=0.333457	lr=0.349918
Epoch[31] Batch [649]	Speed: 5884.532530 samples/sec	accuracy=0.333475	lr=0.349773
Epoch[31] Batch [699]	Speed: 1409.857655 samples/sec	accuracy=0.333546	lr=0.349629
Epoch[31] Batch [749]	Speed: 2517.240383 samples/sec	accuracy=0.333672	lr=0.349484
Epoch[31] Batch [799]	Speed: 3019.239256 samples/sec	accuracy=0.333448	lr=0.349338
Epoch[31] Batch [849]	Speed: 5766.626224 samples/sec	accuracy=0.333466	lr=0.349193
Epoch[31] Batch [899]	Speed: 2830.569466 samples/sec	accuracy=0.333269	lr=0.349048
Epoch[31] Batch [949]	Speed: 2596.940033 samples/sec	accuracy=0.333154	lr=0.348902
Epoch[31] Batch [999]	Speed: 4251.670326 samples/sec	accuracy=0.333134	lr=0.348756
Epoch[31] Batch [1049]	Speed: 1659.363312 samples/sec	accuracy=0.333132	lr=0.348610
Epoch[31] Batch [1099]	Speed: 2359.653160 samples/sec	accuracy=0.333072	lr=0.348464
Epoch[31] Batch [1149]	Speed: 2159.992597 samples/sec	accuracy=0.333145	lr=0.348317
Epoch[31] Batch [1199]	Speed: 975.002311 samples/sec	accuracy=0.333171	lr=0.348171
Epoch[31] Batch [1249]	Speed: 819.172254 samples/sec	accuracy=0.333062	lr=0.348024
[Epoch 31] training: accuracy=0.333073
[Epoch 31] speed: 2305 samples/sec	time cost: 565.460676
[Epoch 31] validation: err-top1=0.808853 err-top5=0.605887
Epoch[32] Batch [49]	Speed: 4990.081218 samples/sec	accuracy=0.337402	lr=0.347874
Epoch[32] Batch [99]	Speed: 1777.402542 samples/sec	accuracy=0.336895	lr=0.347727
Epoch[32] Batch [149]	Speed: 811.939570 samples/sec	accuracy=0.336947	lr=0.347580
Epoch[32] Batch [199]	Speed: 2818.482625 samples/sec	accuracy=0.337358	lr=0.347432
Epoch[32] Batch [249]	Speed: 5784.922430 samples/sec	accuracy=0.337047	lr=0.347285
Epoch[32] Batch [299]	Speed: 5127.930845 samples/sec	accuracy=0.336491	lr=0.347137
Epoch[32] Batch [349]	Speed: 4840.274237 samples/sec	accuracy=0.335837	lr=0.346989
Epoch[32] Batch [399]	Speed: 5031.311155 samples/sec	accuracy=0.335632	lr=0.346841
Epoch[32] Batch [449]	Speed: 5293.123841 samples/sec	accuracy=0.335627	lr=0.346692
Epoch[32] Batch [499]	Speed: 5373.872605 samples/sec	accuracy=0.335654	lr=0.346544
Epoch[32] Batch [549]	Speed: 3104.548529 samples/sec	accuracy=0.335625	lr=0.346395
Epoch[32] Batch [599]	Speed: 4986.823882 samples/sec	accuracy=0.335348	lr=0.346246
Epoch[32] Batch [649]	Speed: 4643.311998 samples/sec	accuracy=0.335075	lr=0.346097
Epoch[32] Batch [699]	Speed: 1372.218067 samples/sec	accuracy=0.334933	lr=0.345948
Epoch[32] Batch [749]	Speed: 5699.531289 samples/sec	accuracy=0.334814	lr=0.345799
Epoch[32] Batch [799]	Speed: 5562.120024 samples/sec	accuracy=0.334719	lr=0.345649
Epoch[32] Batch [849]	Speed: 3219.065913 samples/sec	accuracy=0.334662	lr=0.345499
Epoch[32] Batch [899]	Speed: 2381.712708 samples/sec	accuracy=0.334622	lr=0.345349
Epoch[32] Batch [949]	Speed: 3159.674879 samples/sec	accuracy=0.334486	lr=0.345199
Epoch[32] Batch [999]	Speed: 1824.092496 samples/sec	accuracy=0.334362	lr=0.345049
Epoch[32] Batch [1049]	Speed: 2123.426479 samples/sec	accuracy=0.334345	lr=0.344899
Epoch[32] Batch [1099]	Speed: 1910.808891 samples/sec	accuracy=0.334347	lr=0.344748
Epoch[32] Batch [1149]	Speed: 2464.794940 samples/sec	accuracy=0.334350	lr=0.344597
Epoch[32] Batch [1199]	Speed: 5621.918061 samples/sec	accuracy=0.334378	lr=0.344446
Epoch[32] Batch [1249]	Speed: 3518.633116 samples/sec	accuracy=0.334194	lr=0.344295
[Epoch 32] training: accuracy=0.334205
[Epoch 32] speed: 2846 samples/sec	time cost: 459.896046
[Epoch 32] validation: err-top1=0.734136 err-top5=0.496313
Epoch[33] Batch [49]	Speed: 5660.303773 samples/sec	accuracy=0.339531	lr=0.344141
Epoch[33] Batch [99]	Speed: 5321.051653 samples/sec	accuracy=0.339717	lr=0.343989
Epoch[33] Batch [149]	Speed: 5514.021507 samples/sec	accuracy=0.338672	lr=0.343838
Epoch[33] Batch [199]	Speed: 4980.887275 samples/sec	accuracy=0.338711	lr=0.343686
Epoch[33] Batch [249]	Speed: 983.197707 samples/sec	accuracy=0.338621	lr=0.343534
Epoch[33] Batch [299]	Speed: 3244.753063 samples/sec	accuracy=0.338340	lr=0.343382
Epoch[33] Batch [349]	Speed: 1406.723182 samples/sec	accuracy=0.338214	lr=0.343229
Epoch[33] Batch [399]	Speed: 5466.159261 samples/sec	accuracy=0.337866	lr=0.343077
Epoch[33] Batch [449]	Speed: 2815.628671 samples/sec	accuracy=0.337719	lr=0.342924
Epoch[33] Batch [499]	Speed: 1199.159192 samples/sec	accuracy=0.337545	lr=0.342771
Epoch[33] Batch [549]	Speed: 669.905257 samples/sec	accuracy=0.337024	lr=0.342618
Epoch[33] Batch [599]	Speed: 4914.918616 samples/sec	accuracy=0.336978	lr=0.342465
Epoch[33] Batch [649]	Speed: 2091.151070 samples/sec	accuracy=0.336863	lr=0.342312
Epoch[33] Batch [699]	Speed: 2456.911224 samples/sec	accuracy=0.336858	lr=0.342158
Epoch[33] Batch [749]	Speed: 2175.707050 samples/sec	accuracy=0.336616	lr=0.342005
Epoch[33] Batch [799]	Speed: 3202.055104 samples/sec	accuracy=0.336499	lr=0.341851
Epoch[33] Batch [849]	Speed: 758.072737 samples/sec	accuracy=0.336507	lr=0.341697
Epoch[33] Batch [899]	Speed: 4671.941184 samples/sec	accuracy=0.336271	lr=0.341543
Epoch[33] Batch [949]	Speed: 1384.259852 samples/sec	accuracy=0.336229	lr=0.341388
Epoch[33] Batch [999]	Speed: 4876.891886 samples/sec	accuracy=0.336267	lr=0.341234
Epoch[33] Batch [1049]	Speed: 4291.593331 samples/sec	accuracy=0.336125	lr=0.341079
Epoch[33] Batch [1099]	Speed: 3642.142221 samples/sec	accuracy=0.336049	lr=0.340924
Epoch[33] Batch [1149]	Speed: 2738.608343 samples/sec	accuracy=0.336174	lr=0.340769
Epoch[33] Batch [1199]	Speed: 5043.691496 samples/sec	accuracy=0.336081	lr=0.340614
Epoch[33] Batch [1249]	Speed: 6036.075934 samples/sec	accuracy=0.336064	lr=0.340458
[Epoch 33] training: accuracy=0.336074
[Epoch 33] speed: 2229 samples/sec	time cost: 654.011899
[Epoch 33] validation: err-top1=0.899932 err-top5=0.748904
Epoch[34] Batch [49]	Speed: 2957.608774 samples/sec	accuracy=0.344727	lr=0.340300
Epoch[34] Batch [99]	Speed: 2592.196373 samples/sec	accuracy=0.342012	lr=0.340144
Epoch[34] Batch [149]	Speed: 2367.414620 samples/sec	accuracy=0.340104	lr=0.339988
Epoch[34] Batch [199]	Speed: 2526.886302 samples/sec	accuracy=0.339868	lr=0.339832
Epoch[34] Batch [249]	Speed: 5004.745276 samples/sec	accuracy=0.338613	lr=0.339676
Epoch[34] Batch [299]	Speed: 1456.900085 samples/sec	accuracy=0.339535	lr=0.339520
Epoch[34] Batch [349]	Speed: 1811.429555 samples/sec	accuracy=0.339093	lr=0.339363
Epoch[34] Batch [399]	Speed: 4050.730677 samples/sec	accuracy=0.338479	lr=0.339206
Epoch[34] Batch [449]	Speed: 3159.486701 samples/sec	accuracy=0.338329	lr=0.339049
Epoch[34] Batch [499]	Speed: 5573.652316 samples/sec	accuracy=0.338180	lr=0.338892
Epoch[34] Batch [549]	Speed: 2416.873137 samples/sec	accuracy=0.338024	lr=0.338735
Epoch[34] Batch [599]	Speed: 1266.520797 samples/sec	accuracy=0.338055	lr=0.338578
Epoch[34] Batch [649]	Speed: 4274.077571 samples/sec	accuracy=0.338032	lr=0.338420
Epoch[34] Batch [699]	Speed: 5959.325836 samples/sec	accuracy=0.337640	lr=0.338263
Epoch[34] Batch [749]	Speed: 3345.938799 samples/sec	accuracy=0.337527	lr=0.338105
Epoch[34] Batch [799]	Speed: 4813.040254 samples/sec	accuracy=0.337516	lr=0.337947
Epoch[34] Batch [849]	Speed: 3416.309831 samples/sec	accuracy=0.337546	lr=0.337788
Epoch[34] Batch [899]	Speed: 4360.860509 samples/sec	accuracy=0.337420	lr=0.337630
Epoch[34] Batch [949]	Speed: 2448.134938 samples/sec	accuracy=0.337284	lr=0.337472
Epoch[34] Batch [999]	Speed: 1890.851131 samples/sec	accuracy=0.337343	lr=0.337313
Epoch[34] Batch [1049]	Speed: 5253.365569 samples/sec	accuracy=0.337277	lr=0.337154
Epoch[34] Batch [1099]	Speed: 5313.148990 samples/sec	accuracy=0.337360	lr=0.336995
Epoch[34] Batch [1149]	Speed: 3445.024605 samples/sec	accuracy=0.337283	lr=0.336836
Epoch[34] Batch [1199]	Speed: 2500.334620 samples/sec	accuracy=0.337156	lr=0.336677
Epoch[34] Batch [1249]	Speed: 3539.283483 samples/sec	accuracy=0.337019	lr=0.336517
[Epoch 34] training: accuracy=0.337042
[Epoch 34] speed: 2890 samples/sec	time cost: 453.413701
[Epoch 34] validation: err-top1=0.746297 err-top5=0.508097
Epoch[35] Batch [49]	Speed: 4926.512727 samples/sec	accuracy=0.344043	lr=0.336354
Epoch[35] Batch [99]	Speed: 4856.050305 samples/sec	accuracy=0.342451	lr=0.336194
Epoch[35] Batch [149]	Speed: 1603.441077 samples/sec	accuracy=0.341029	lr=0.336034
Epoch[35] Batch [199]	Speed: 3832.506549 samples/sec	accuracy=0.340420	lr=0.335874
Epoch[35] Batch [249]	Speed: 2610.748520 samples/sec	accuracy=0.340027	lr=0.335714
Epoch[35] Batch [299]	Speed: 4173.835981 samples/sec	accuracy=0.339440	lr=0.335553
Epoch[35] Batch [349]	Speed: 5054.382505 samples/sec	accuracy=0.339754	lr=0.335393
Epoch[35] Batch [399]	Speed: 2124.737357 samples/sec	accuracy=0.339497	lr=0.335232
Epoch[35] Batch [449]	Speed: 2773.568972 samples/sec	accuracy=0.339551	lr=0.335071
Epoch[35] Batch [499]	Speed: 5577.386332 samples/sec	accuracy=0.339654	lr=0.334910
Epoch[35] Batch [549]	Speed: 3888.919741 samples/sec	accuracy=0.339242	lr=0.334749
Epoch[35] Batch [599]	Speed: 5745.195517 samples/sec	accuracy=0.339357	lr=0.334587
Epoch[35] Batch [649]	Speed: 5735.406186 samples/sec	accuracy=0.339059	lr=0.334425
Epoch[35] Batch [699]	Speed: 3447.584363 samples/sec	accuracy=0.338896	lr=0.334264
Epoch[35] Batch [749]	Speed: 4817.621083 samples/sec	accuracy=0.338835	lr=0.334102
Epoch[35] Batch [799]	Speed: 1206.237638 samples/sec	accuracy=0.338591	lr=0.333940
Epoch[35] Batch [849]	Speed: 2580.628364 samples/sec	accuracy=0.338482	lr=0.333777
Epoch[35] Batch [899]	Speed: 5495.121083 samples/sec	accuracy=0.338322	lr=0.333615
Epoch[35] Batch [949]	Speed: 5649.907116 samples/sec	accuracy=0.338262	lr=0.333452
Epoch[35] Batch [999]	Speed: 3695.270622 samples/sec	accuracy=0.338228	lr=0.333290
Epoch[35] Batch [1049]	Speed: 2338.044517 samples/sec	accuracy=0.338140	lr=0.333127
Epoch[35] Batch [1099]	Speed: 5037.067701 samples/sec	accuracy=0.337963	lr=0.332964
Epoch[35] Batch [1149]	Speed: 5455.947123 samples/sec	accuracy=0.337992	lr=0.332801
Epoch[35] Batch [1199]	Speed: 4226.696261 samples/sec	accuracy=0.338041	lr=0.332637
Epoch[35] Batch [1249]	Speed: 1705.013200 samples/sec	accuracy=0.338102	lr=0.332474
[Epoch 35] training: accuracy=0.338123
[Epoch 35] speed: 3243 samples/sec	time cost: 415.167047
[Epoch 35] validation: err-top1=0.779576 err-top5=0.560108
Epoch[36] Batch [49]	Speed: 5207.967293 samples/sec	accuracy=0.340098	lr=0.332307
Epoch[36] Batch [99]	Speed: 5705.088255 samples/sec	accuracy=0.342080	lr=0.332143
Epoch[36] Batch [149]	Speed: 2687.683222 samples/sec	accuracy=0.342441	lr=0.331979
Epoch[36] Batch [199]	Speed: 4293.294714 samples/sec	accuracy=0.341724	lr=0.331815
Epoch[36] Batch [249]	Speed: 4518.488424 samples/sec	accuracy=0.341082	lr=0.331650
Epoch[36] Batch [299]	Speed: 5840.916241 samples/sec	accuracy=0.340218	lr=0.331486
Epoch[36] Batch [349]	Speed: 2744.859112 samples/sec	accuracy=0.340611	lr=0.331321
Epoch[36] Batch [399]	Speed: 2557.322110 samples/sec	accuracy=0.340449	lr=0.331157
Epoch[36] Batch [449]	Speed: 5918.934940 samples/sec	accuracy=0.339850	lr=0.330992
Epoch[36] Batch [499]	Speed: 3065.328446 samples/sec	accuracy=0.340096	lr=0.330827
Epoch[36] Batch [549]	Speed: 3060.384893 samples/sec	accuracy=0.339814	lr=0.330661
Epoch[36] Batch [599]	Speed: 5496.274491 samples/sec	accuracy=0.339666	lr=0.330496
Epoch[36] Batch [649]	Speed: 5495.892031 samples/sec	accuracy=0.339561	lr=0.330330
Epoch[36] Batch [699]	Speed: 2468.562486 samples/sec	accuracy=0.339662	lr=0.330165
Epoch[36] Batch [749]	Speed: 2078.522954 samples/sec	accuracy=0.339242	lr=0.329999
Epoch[36] Batch [799]	Speed: 5655.826740 samples/sec	accuracy=0.339142	lr=0.329833
Epoch[36] Batch [849]	Speed: 1765.032229 samples/sec	accuracy=0.338973	lr=0.329667
Epoch[36] Batch [899]	Speed: 1542.437404 samples/sec	accuracy=0.339045	lr=0.329500
Epoch[36] Batch [949]	Speed: 5543.807365 samples/sec	accuracy=0.339110	lr=0.329334
Epoch[36] Batch [999]	Speed: 2049.350128 samples/sec	accuracy=0.339103	lr=0.329167
Epoch[36] Batch [1049]	Speed: 4025.568672 samples/sec	accuracy=0.338981	lr=0.329000
Epoch[36] Batch [1099]	Speed: 2038.450750 samples/sec	accuracy=0.338758	lr=0.328833
Epoch[36] Batch [1149]	Speed: 5264.913070 samples/sec	accuracy=0.338578	lr=0.328666
Epoch[36] Batch [1199]	Speed: 5630.236269 samples/sec	accuracy=0.338441	lr=0.328499
Epoch[36] Batch [1249]	Speed: 5006.430656 samples/sec	accuracy=0.338434	lr=0.328332
[Epoch 36] training: accuracy=0.338423
[Epoch 36] speed: 3304 samples/sec	time cost: 397.428541
[Epoch 36] validation: err-top1=0.732123 err-top5=0.491829
Epoch[37] Batch [49]	Speed: 5224.055290 samples/sec	accuracy=0.343828	lr=0.328161
Epoch[37] Batch [99]	Speed: 5708.538260 samples/sec	accuracy=0.344375	lr=0.327993
Epoch[37] Batch [149]	Speed: 2950.114926 samples/sec	accuracy=0.344264	lr=0.327825
Epoch[37] Batch [199]	Speed: 3116.139238 samples/sec	accuracy=0.344497	lr=0.327657
Epoch[37] Batch [249]	Speed: 4139.149430 samples/sec	accuracy=0.342969	lr=0.327489
Epoch[37] Batch [299]	Speed: 3288.506020 samples/sec	accuracy=0.342559	lr=0.327321
Epoch[37] Batch [349]	Speed: 5417.213222 samples/sec	accuracy=0.342218	lr=0.327152
Epoch[37] Batch [399]	Speed: 4746.043622 samples/sec	accuracy=0.341599	lr=0.326983
Epoch[37] Batch [449]	Speed: 3940.150372 samples/sec	accuracy=0.341115	lr=0.326815
Epoch[37] Batch [499]	Speed: 3391.324740 samples/sec	accuracy=0.340678	lr=0.326646
Epoch[37] Batch [549]	Speed: 3028.319833 samples/sec	accuracy=0.340517	lr=0.326477
Epoch[37] Batch [599]	Speed: 4527.673028 samples/sec	accuracy=0.340052	lr=0.326307
Epoch[37] Batch [649]	Speed: 3271.704087 samples/sec	accuracy=0.339913	lr=0.326138
Epoch[37] Batch [699]	Speed: 2452.284627 samples/sec	accuracy=0.339722	lr=0.325968
Epoch[37] Batch [749]	Speed: 3368.304228 samples/sec	accuracy=0.339966	lr=0.325799
Epoch[37] Batch [799]	Speed: 5669.757810 samples/sec	accuracy=0.339960	lr=0.325629
Epoch[37] Batch [849]	Speed: 5763.162588 samples/sec	accuracy=0.339733	lr=0.325459
Epoch[37] Batch [899]	Speed: 5596.975468 samples/sec	accuracy=0.339786	lr=0.325289
Epoch[37] Batch [949]	Speed: 2470.214363 samples/sec	accuracy=0.339588	lr=0.325118
Epoch[37] Batch [999]	Speed: 3358.868826 samples/sec	accuracy=0.339502	lr=0.324948
Epoch[37] Batch [1049]	Speed: 5623.614932 samples/sec	accuracy=0.339402	lr=0.324777
Epoch[37] Batch [1099]	Speed: 5906.793819 samples/sec	accuracy=0.339484	lr=0.324607
Epoch[37] Batch [1149]	Speed: 3566.056562 samples/sec	accuracy=0.339561	lr=0.324436
Epoch[37] Batch [1199]	Speed: 3500.238153 samples/sec	accuracy=0.339386	lr=0.324265
Epoch[37] Batch [1249]	Speed: 3392.908705 samples/sec	accuracy=0.339487	lr=0.324094
[Epoch 37] training: accuracy=0.339503
[Epoch 37] speed: 3833 samples/sec	time cost: 377.062128
[Epoch 37] validation: err-top1=0.809212 err-top5=0.605648
Epoch[38] Batch [49]	Speed: 2722.818558 samples/sec	accuracy=0.342969	lr=0.323919
Epoch[38] Batch [99]	Speed: 3470.131535 samples/sec	accuracy=0.342793	lr=0.323747
Epoch[38] Batch [149]	Speed: 2063.365399 samples/sec	accuracy=0.342559	lr=0.323576
Epoch[38] Batch [199]	Speed: 4091.877669 samples/sec	accuracy=0.342539	lr=0.323404
Epoch[38] Batch [249]	Speed: 3603.514849 samples/sec	accuracy=0.343180	lr=0.323232
Epoch[38] Batch [299]	Speed: 3797.695592 samples/sec	accuracy=0.342829	lr=0.323060
Epoch[38] Batch [349]	Speed: 5444.624886 samples/sec	accuracy=0.342093	lr=0.322888
Epoch[38] Batch [399]	Speed: 2100.057996 samples/sec	accuracy=0.342114	lr=0.322715
Epoch[38] Batch [449]	Speed: 3680.527834 samples/sec	accuracy=0.342595	lr=0.322543
Epoch[38] Batch [499]	Speed: 3812.134399 samples/sec	accuracy=0.342617	lr=0.322370
Epoch[38] Batch [549]	Speed: 2986.420213 samples/sec	accuracy=0.342205	lr=0.322197
Epoch[38] Batch [599]	Speed: 1975.042254 samples/sec	accuracy=0.342122	lr=0.322025
Epoch[38] Batch [649]	Speed: 4957.725798 samples/sec	accuracy=0.342115	lr=0.321851
Epoch[38] Batch [699]	Speed: 4323.131112 samples/sec	accuracy=0.341974	lr=0.321678
Epoch[38] Batch [749]	Speed: 2047.536687 samples/sec	accuracy=0.342056	lr=0.321505
Epoch[38] Batch [799]	Speed: 4382.870050 samples/sec	accuracy=0.341826	lr=0.321331
Epoch[38] Batch [849]	Speed: 1384.267945 samples/sec	accuracy=0.341282	lr=0.321158
Epoch[38] Batch [899]	Speed: 4806.270244 samples/sec	accuracy=0.341068	lr=0.320984
Epoch[38] Batch [949]	Speed: 3938.693269 samples/sec	accuracy=0.340931	lr=0.320810
Epoch[38] Batch [999]	Speed: 5214.621361 samples/sec	accuracy=0.341077	lr=0.320636
Epoch[38] Batch [1049]	Speed: 5814.365401 samples/sec	accuracy=0.341088	lr=0.320462
Epoch[38] Batch [1099]	Speed: 5046.125695 samples/sec	accuracy=0.341037	lr=0.320287
Epoch[38] Batch [1149]	Speed: 4773.165126 samples/sec	accuracy=0.340859	lr=0.320113
Epoch[38] Batch [1199]	Speed: 4919.080111 samples/sec	accuracy=0.340852	lr=0.319938
Epoch[38] Batch [1249]	Speed: 5041.995737 samples/sec	accuracy=0.340829	lr=0.319763
[Epoch 38] training: accuracy=0.340830
[Epoch 38] speed: 3345 samples/sec	time cost: 393.097440
[Epoch 38] validation: err-top1=0.759626 err-top5=0.524294
Epoch[39] Batch [49]	Speed: 5406.335186 samples/sec	accuracy=0.346055	lr=0.319585
Epoch[39] Batch [99]	Speed: 5247.791605 samples/sec	accuracy=0.344238	lr=0.319410
Epoch[39] Batch [149]	Speed: 5614.397689 samples/sec	accuracy=0.344499	lr=0.319234
Epoch[39] Batch [199]	Speed: 5163.293180 samples/sec	accuracy=0.344546	lr=0.319059
Epoch[39] Batch [249]	Speed: 4883.874005 samples/sec	accuracy=0.344316	lr=0.318883
Epoch[39] Batch [299]	Speed: 5523.818780 samples/sec	accuracy=0.344027	lr=0.318708
Epoch[39] Batch [349]	Speed: 5362.736238 samples/sec	accuracy=0.344062	lr=0.318532
Epoch[39] Batch [399]	Speed: 4245.678535 samples/sec	accuracy=0.343862	lr=0.318356
Epoch[39] Batch [449]	Speed: 5079.318624 samples/sec	accuracy=0.343442	lr=0.318180
Epoch[39] Batch [499]	Speed: 5574.091684 samples/sec	accuracy=0.343418	lr=0.318004
Epoch[39] Batch [549]	Speed: 5010.410763 samples/sec	accuracy=0.343263	lr=0.317827
Epoch[39] Batch [599]	Speed: 5513.303640 samples/sec	accuracy=0.343351	lr=0.317651
Epoch[39] Batch [649]	Speed: 5041.235740 samples/sec	accuracy=0.343344	lr=0.317474
Epoch[39] Batch [699]	Speed: 5339.642509 samples/sec	accuracy=0.343408	lr=0.317297
Epoch[39] Batch [749]	Speed: 5341.049029 samples/sec	accuracy=0.343143	lr=0.317120
Epoch[39] Batch [799]	Speed: 4961.386484 samples/sec	accuracy=0.342825	lr=0.316943
Epoch[39] Batch [849]	Speed: 5288.379820 samples/sec	accuracy=0.342926	lr=0.316766
Epoch[39] Batch [899]	Speed: 5188.793040 samples/sec	accuracy=0.342963	lr=0.316589
Epoch[39] Batch [949]	Speed: 5823.473798 samples/sec	accuracy=0.342730	lr=0.316411
Epoch[39] Batch [999]	Speed: 4728.450658 samples/sec	accuracy=0.342524	lr=0.316233
Epoch[39] Batch [1049]	Speed: 5330.158615 samples/sec	accuracy=0.342406	lr=0.316056
Epoch[39] Batch [1099]	Speed: 5427.556314 samples/sec	accuracy=0.342373	lr=0.315878
Epoch[39] Batch [1149]	Speed: 5242.565395 samples/sec	accuracy=0.342374	lr=0.315700
Epoch[39] Batch [1199]	Speed: 5576.582217 samples/sec	accuracy=0.342342	lr=0.315522
Epoch[39] Batch [1249]	Speed: 4780.171426 samples/sec	accuracy=0.342334	lr=0.315343
[Epoch 39] training: accuracy=0.342331
[Epoch 39] speed: 5199 samples/sec	time cost: 257.293068
[Epoch 39] validation: err-top1=0.803791 err-top5=0.589246
Epoch[40] Batch [49]	Speed: 5263.013468 samples/sec	accuracy=0.346387	lr=0.315161
Epoch[40] Batch [99]	Speed: 4772.325555 samples/sec	accuracy=0.345693	lr=0.314983
Epoch[40] Batch [149]	Speed: 4646.312126 samples/sec	accuracy=0.345807	lr=0.314804
Epoch[40] Batch [199]	Speed: 5552.541268 samples/sec	accuracy=0.344517	lr=0.314625
Epoch[40] Batch [249]	Speed: 5385.389962 samples/sec	accuracy=0.344734	lr=0.314446
Epoch[40] Batch [299]	Speed: 5376.936370 samples/sec	accuracy=0.345361	lr=0.314267
Epoch[40] Batch [349]	Speed: 2850.628098 samples/sec	accuracy=0.345273	lr=0.314087
Epoch[40] Batch [399]	Speed: 5607.229446 samples/sec	accuracy=0.344524	lr=0.313908
Epoch[40] Batch [449]	Speed: 3765.219709 samples/sec	accuracy=0.344614	lr=0.313728
Epoch[40] Batch [499]	Speed: 4932.441455 samples/sec	accuracy=0.344602	lr=0.313549
Epoch[40] Batch [549]	Speed: 4500.162087 samples/sec	accuracy=0.344462	lr=0.313369
Epoch[40] Batch [599]	Speed: 4603.228672 samples/sec	accuracy=0.344152	lr=0.313189
Epoch[40] Batch [649]	Speed: 4938.018141 samples/sec	accuracy=0.344193	lr=0.313009
Epoch[40] Batch [699]	Speed: 5069.603526 samples/sec	accuracy=0.344515	lr=0.312829
Epoch[40] Batch [749]	Speed: 5137.190581 samples/sec	accuracy=0.344408	lr=0.312648
Epoch[40] Batch [799]	Speed: 5062.367542 samples/sec	accuracy=0.344531	lr=0.312468
Epoch[40] Batch [849]	Speed: 5466.946040 samples/sec	accuracy=0.344256	lr=0.312287
Epoch[40] Batch [899]	Speed: 4932.935679 samples/sec	accuracy=0.344251	lr=0.312106
Epoch[40] Batch [949]	Speed: 5231.865620 samples/sec	accuracy=0.344151	lr=0.311925
Epoch[40] Batch [999]	Speed: 1331.984320 samples/sec	accuracy=0.344159	lr=0.311744
Epoch[40] Batch [1049]	Speed: 811.085062 samples/sec	accuracy=0.344059	lr=0.311563
Epoch[40] Batch [1099]	Speed: 4328.285922 samples/sec	accuracy=0.343917	lr=0.311382
Epoch[40] Batch [1149]	Speed: 2200.568309 samples/sec	accuracy=0.343845	lr=0.311201
Epoch[40] Batch [1199]	Speed: 5573.788878 samples/sec	accuracy=0.343653	lr=0.311019
Epoch[40] Batch [1249]	Speed: 3007.679547 samples/sec	accuracy=0.343591	lr=0.310837
[Epoch 40] training: accuracy=0.343571
[Epoch 40] speed: 3405 samples/sec	time cost: 403.536129
[Epoch 40] validation: err-top1=0.740621 err-top5=0.501180
Epoch[41] Batch [49]	Speed: 430.166267 samples/sec	accuracy=0.347734	lr=0.310652
Epoch[41] Batch [99]	Speed: 4421.210906 samples/sec	accuracy=0.347734	lr=0.310470
Epoch[41] Batch [149]	Speed: 3822.196453 samples/sec	accuracy=0.348171	lr=0.310288
Epoch[41] Batch [199]	Speed: 947.281229 samples/sec	accuracy=0.347515	lr=0.310105
Epoch[41] Batch [249]	Speed: 3081.380190 samples/sec	accuracy=0.346355	lr=0.309923
Epoch[41] Batch [299]	Speed: 3690.695546 samples/sec	accuracy=0.345365	lr=0.309741
Epoch[41] Batch [349]	Speed: 4244.113060 samples/sec	accuracy=0.344983	lr=0.309558
Epoch[41] Batch [399]	Speed: 576.116521 samples/sec	accuracy=0.344692	lr=0.309375
Epoch[41] Batch [449]	Speed: 5369.897881 samples/sec	accuracy=0.344607	lr=0.309192
Epoch[41] Batch [499]	Speed: 1001.547688 samples/sec	accuracy=0.344609	lr=0.309009
Epoch[41] Batch [549]	Speed: 3930.129953 samples/sec	accuracy=0.344547	lr=0.308826
Epoch[41] Batch [599]	Speed: 5837.479284 samples/sec	accuracy=0.344380	lr=0.308643
Epoch[41] Batch [649]	Speed: 1274.383034 samples/sec	accuracy=0.344280	lr=0.308459
Epoch[41] Batch [699]	Speed: 5937.685946 samples/sec	accuracy=0.344166	lr=0.308276
Epoch[41] Batch [749]	Speed: 4362.763063 samples/sec	accuracy=0.344245	lr=0.308092
Epoch[41] Batch [799]	Speed: 1843.759124 samples/sec	accuracy=0.344257	lr=0.307908
Epoch[41] Batch [849]	Speed: 5166.106437 samples/sec	accuracy=0.344334	lr=0.307725
Epoch[41] Batch [899]	Speed: 5546.797226 samples/sec	accuracy=0.344247	lr=0.307540
Epoch[41] Batch [949]	Speed: 1819.044503 samples/sec	accuracy=0.344252	lr=0.307356
Epoch[41] Batch [999]	Speed: 5605.376818 samples/sec	accuracy=0.344273	lr=0.307172
Epoch[41] Batch [1049]	Speed: 5627.398962 samples/sec	accuracy=0.344218	lr=0.306988
Epoch[41] Batch [1099]	Speed: 5693.316401 samples/sec	accuracy=0.344127	lr=0.306803
Epoch[41] Batch [1149]	Speed: 1176.111994 samples/sec	accuracy=0.344102	lr=0.306618
Epoch[41] Batch [1199]	Speed: 1797.894663 samples/sec	accuracy=0.343933	lr=0.306433
Epoch[41] Batch [1249]	Speed: 5297.759407 samples/sec	accuracy=0.343863	lr=0.306249
[Epoch 41] training: accuracy=0.343876
[Epoch 41] speed: 1948 samples/sec	time cost: 670.283640
[Epoch 41] validation: err-top1=0.718989 err-top5=0.474031
Epoch[42] Batch [49]	Speed: 1328.407301 samples/sec	accuracy=0.351562	lr=0.306060
Epoch[42] Batch [99]	Speed: 5255.308497 samples/sec	accuracy=0.350479	lr=0.305875
Epoch[42] Batch [149]	Speed: 1411.037822 samples/sec	accuracy=0.349635	lr=0.305689
Epoch[42] Batch [199]	Speed: 1019.118625 samples/sec	accuracy=0.348257	lr=0.305504
Epoch[42] Batch [249]	Speed: 1411.672487 samples/sec	accuracy=0.348187	lr=0.305318
Epoch[42] Batch [299]	Speed: 5918.614716 samples/sec	accuracy=0.347751	lr=0.305133
Epoch[42] Batch [349]	Speed: 5654.284111 samples/sec	accuracy=0.347109	lr=0.304947
Epoch[42] Batch [399]	Speed: 1413.053816 samples/sec	accuracy=0.347043	lr=0.304761
Epoch[42] Batch [449]	Speed: 5325.777891 samples/sec	accuracy=0.346899	lr=0.304575
Epoch[42] Batch [499]	Speed: 370.071473 samples/sec	accuracy=0.346674	lr=0.304388
Epoch[42] Batch [549]	Speed: 528.630924 samples/sec	accuracy=0.346119	lr=0.304202
Epoch[42] Batch [599]	Speed: 324.439986 samples/sec	accuracy=0.346076	lr=0.304016
Epoch[42] Batch [649]	Speed: 347.069379 samples/sec	accuracy=0.345868	lr=0.303829
Epoch[42] Batch [699]	Speed: 5662.383994 samples/sec	accuracy=0.346053	lr=0.303642
Epoch[42] Batch [749]	Speed: 2594.281746 samples/sec	accuracy=0.346016	lr=0.303456
Epoch[42] Batch [799]	Speed: 5852.696065 samples/sec	accuracy=0.345930	lr=0.303269
Epoch[42] Batch [849]	Speed: 1379.272034 samples/sec	accuracy=0.345851	lr=0.303082
Epoch[42] Batch [899]	Speed: 805.695968 samples/sec	accuracy=0.345630	lr=0.302894
Epoch[42] Batch [949]	Speed: 2793.587498 samples/sec	accuracy=0.345612	lr=0.302707
Epoch[42] Batch [999]	Speed: 2720.146003 samples/sec	accuracy=0.345487	lr=0.302520
Epoch[42] Batch [1049]	Speed: 806.075184 samples/sec	accuracy=0.345418	lr=0.302332
Epoch[42] Batch [1099]	Speed: 5380.051209 samples/sec	accuracy=0.345413	lr=0.302144
Epoch[42] Batch [1149]	Speed: 2561.087559 samples/sec	accuracy=0.345370	lr=0.301957
Epoch[42] Batch [1199]	Speed: 5864.335486 samples/sec	accuracy=0.345321	lr=0.301769
Epoch[42] Batch [1249]	Speed: 2513.673638 samples/sec	accuracy=0.345395	lr=0.301581
[Epoch 42] training: accuracy=0.345384
[Epoch 42] speed: 1193 samples/sec	time cost: 1083.129276
[Epoch 42] validation: err-top1=0.739636 err-top5=0.503508
Epoch[43] Batch [49]	Speed: 3188.823580 samples/sec	accuracy=0.351445	lr=0.301389
Epoch[43] Batch [99]	Speed: 1403.876292 samples/sec	accuracy=0.350205	lr=0.301200
Epoch[43] Batch [149]	Speed: 1823.337739 samples/sec	accuracy=0.349102	lr=0.301012
Epoch[43] Batch [199]	Speed: 2660.250463 samples/sec	accuracy=0.348794	lr=0.300823
Epoch[43] Batch [249]	Speed: 583.930745 samples/sec	accuracy=0.347809	lr=0.300635
Epoch[43] Batch [299]	Speed: 3660.502376 samples/sec	accuracy=0.347695	lr=0.300446
Epoch[43] Batch [349]	Speed: 5431.591332 samples/sec	accuracy=0.347586	lr=0.300257
Epoch[43] Batch [399]	Speed: 5951.646070 samples/sec	accuracy=0.347810	lr=0.300068
Epoch[43] Batch [449]	Speed: 817.249480 samples/sec	accuracy=0.347615	lr=0.299879
Epoch[43] Batch [499]	Speed: 5760.732898 samples/sec	accuracy=0.347443	lr=0.299690
Epoch[43] Batch [549]	Speed: 2966.678521 samples/sec	accuracy=0.347113	lr=0.299500
Epoch[43] Batch [599]	Speed: 5677.722128 samples/sec	accuracy=0.347119	lr=0.299311
Epoch[43] Batch [649]	Speed: 410.335851 samples/sec	accuracy=0.347012	lr=0.299121
Epoch[43] Batch [699]	Speed: 5828.102131 samples/sec	accuracy=0.347044	lr=0.298932
Epoch[43] Batch [749]	Speed: 503.505025 samples/sec	accuracy=0.347346	lr=0.298742
Epoch[43] Batch [799]	Speed: 5976.662557 samples/sec	accuracy=0.347172	lr=0.298552
Epoch[43] Batch [849]	Speed: 5372.740690 samples/sec	accuracy=0.347239	lr=0.298362
Epoch[43] Batch [899]	Speed: 5343.309299 samples/sec	accuracy=0.347171	lr=0.298171
Epoch[43] Batch [949]	Speed: 6023.781673 samples/sec	accuracy=0.347054	lr=0.297981
Epoch[43] Batch [999]	Speed: 290.863329 samples/sec	accuracy=0.346813	lr=0.297791
Epoch[43] Batch [1049]	Speed: 5403.614579 samples/sec	accuracy=0.346681	lr=0.297600
Epoch[43] Batch [1099]	Speed: 5891.869578 samples/sec	accuracy=0.346649	lr=0.297409
Epoch[43] Batch [1149]	Speed: 4333.398823 samples/sec	accuracy=0.346598	lr=0.297219
Epoch[43] Batch [1199]	Speed: 5061.371149 samples/sec	accuracy=0.346454	lr=0.297028
Epoch[43] Batch [1249]	Speed: 5940.814453 samples/sec	accuracy=0.346483	lr=0.296837
[Epoch 43] training: accuracy=0.346506
[Epoch 43] speed: 1572 samples/sec	time cost: 837.368208
[Epoch 43] validation: err-top1=0.820193 err-top5=0.619479
Epoch[44] Batch [49]	Speed: 4014.888963 samples/sec	accuracy=0.350762	lr=0.296642
Epoch[44] Batch [99]	Speed: 2140.381690 samples/sec	accuracy=0.351104	lr=0.296451
Epoch[44] Batch [149]	Speed: 3533.919090 samples/sec	accuracy=0.351035	lr=0.296259
Epoch[44] Batch [199]	Speed: 4753.772201 samples/sec	accuracy=0.349644	lr=0.296068
Epoch[44] Batch [249]	Speed: 809.834876 samples/sec	accuracy=0.348910	lr=0.295876
Epoch[44] Batch [299]	Speed: 3185.258851 samples/sec	accuracy=0.348838	lr=0.295684
Epoch[44] Batch [349]	Speed: 1828.421765 samples/sec	accuracy=0.348477	lr=0.295493
Epoch[44] Batch [399]	Speed: 5291.129521 samples/sec	accuracy=0.348364	lr=0.295301
Epoch[44] Batch [449]	Speed: 682.964785 samples/sec	accuracy=0.348531	lr=0.295109
Epoch[44] Batch [499]	Speed: 5615.793064 samples/sec	accuracy=0.348283	lr=0.294917
Epoch[44] Batch [549]	Speed: 5428.049919 samples/sec	accuracy=0.348427	lr=0.294724
Epoch[44] Batch [599]	Speed: 2519.849818 samples/sec	accuracy=0.348376	lr=0.294532
Epoch[44] Batch [649]	Speed: 208.141192 samples/sec	accuracy=0.348119	lr=0.294339
Epoch[44] Batch [699]	Speed: 797.050874 samples/sec	accuracy=0.347988	lr=0.294147
Epoch[44] Batch [749]	Speed: 3180.044657 samples/sec	accuracy=0.347980	lr=0.293954
Epoch[44] Batch [799]	Speed: 1168.457313 samples/sec	accuracy=0.348052	lr=0.293761
Epoch[44] Batch [849]	Speed: 3077.086009 samples/sec	accuracy=0.348164	lr=0.293568
Epoch[44] Batch [899]	Speed: 3546.876786 samples/sec	accuracy=0.348273	lr=0.293375
Epoch[44] Batch [949]	Speed: 2895.338374 samples/sec	accuracy=0.348241	lr=0.293182
Epoch[44] Batch [999]	Speed: 1594.498874 samples/sec	accuracy=0.348019	lr=0.292989
Epoch[44] Batch [1049]	Speed: 5304.652785 samples/sec	accuracy=0.347979	lr=0.292795
Epoch[44] Batch [1099]	Speed: 4193.032823 samples/sec	accuracy=0.347940	lr=0.292602
Epoch[44] Batch [1149]	Speed: 4757.063602 samples/sec	accuracy=0.348087	lr=0.292408
Epoch[44] Batch [1199]	Speed: 2259.997321 samples/sec	accuracy=0.348168	lr=0.292215
Epoch[44] Batch [1249]	Speed: 4021.317373 samples/sec	accuracy=0.348027	lr=0.292021
[Epoch 44] training: accuracy=0.348035
[Epoch 44] speed: 1575 samples/sec	time cost: 855.978822
[Epoch 44] validation: err-top1=0.713648 err-top5=0.466857
Epoch[45] Batch [49]	Speed: 1753.953245 samples/sec	accuracy=0.354336	lr=0.291823
Epoch[45] Batch [99]	Speed: 2492.929832 samples/sec	accuracy=0.352168	lr=0.291629
Epoch[45] Batch [149]	Speed: 2967.004132 samples/sec	accuracy=0.352461	lr=0.291435
Epoch[45] Batch [199]	Speed: 5733.079426 samples/sec	accuracy=0.351729	lr=0.291240
Epoch[45] Batch [249]	Speed: 5759.315702 samples/sec	accuracy=0.351195	lr=0.291046
Epoch[45] Batch [299]	Speed: 3051.988643 samples/sec	accuracy=0.350798	lr=0.290852
Epoch[45] Batch [349]	Speed: 1334.669497 samples/sec	accuracy=0.350170	lr=0.290657
Epoch[45] Batch [399]	Speed: 1662.765016 samples/sec	accuracy=0.349480	lr=0.290462
Epoch[45] Batch [449]	Speed: 5242.374833 samples/sec	accuracy=0.349284	lr=0.290267
Epoch[45] Batch [499]	Speed: 5463.001065 samples/sec	accuracy=0.349318	lr=0.290073
Epoch[45] Batch [549]	Speed: 1839.615764 samples/sec	accuracy=0.349261	lr=0.289878
Epoch[45] Batch [599]	Speed: 1367.220025 samples/sec	accuracy=0.349040	lr=0.289682
Epoch[45] Batch [649]	Speed: 5276.911052 samples/sec	accuracy=0.348879	lr=0.289487
Epoch[45] Batch [699]	Speed: 4441.650678 samples/sec	accuracy=0.349014	lr=0.289292
Epoch[45] Batch [749]	Speed: 1715.974808 samples/sec	accuracy=0.349017	lr=0.289096
Epoch[45] Batch [799]	Speed: 3280.889949 samples/sec	accuracy=0.349203	lr=0.288901
Epoch[45] Batch [849]	Speed: 5622.330480 samples/sec	accuracy=0.349274	lr=0.288705
Epoch[45] Batch [899]	Speed: 3265.844104 samples/sec	accuracy=0.349311	lr=0.288509
Epoch[45] Batch [949]	Speed: 3510.773650 samples/sec	accuracy=0.349278	lr=0.288314
Epoch[45] Batch [999]	Speed: 5511.855451 samples/sec	accuracy=0.349347	lr=0.288118
Epoch[45] Batch [1049]	Speed: 1938.090071 samples/sec	accuracy=0.349221	lr=0.287921
Epoch[45] Batch [1099]	Speed: 2725.881006 samples/sec	accuracy=0.349118	lr=0.287725
Epoch[45] Batch [1149]	Speed: 3065.330634 samples/sec	accuracy=0.349101	lr=0.287529
Epoch[45] Batch [1199]	Speed: 4742.182727 samples/sec	accuracy=0.348891	lr=0.287333
Epoch[45] Batch [1249]	Speed: 2347.543641 samples/sec	accuracy=0.348794	lr=0.287136
[Epoch 45] training: accuracy=0.348794
[Epoch 45] speed: 2751 samples/sec	time cost: 484.456722
[Epoch 45] validation: err-top1=0.771086 err-top5=0.558056
Epoch[46] Batch [49]	Speed: 4349.279797 samples/sec	accuracy=0.352109	lr=0.286936
Epoch[46] Batch [99]	Speed: 4428.755390 samples/sec	accuracy=0.352393	lr=0.286739
Epoch[46] Batch [149]	Speed: 2450.200533 samples/sec	accuracy=0.352487	lr=0.286542
Epoch[46] Batch [199]	Speed: 3098.943993 samples/sec	accuracy=0.351357	lr=0.286345
Epoch[46] Batch [249]	Speed: 1592.503928 samples/sec	accuracy=0.351680	lr=0.286148
Epoch[46] Batch [299]	Speed: 4055.300343 samples/sec	accuracy=0.351344	lr=0.285951
Epoch[46] Batch [349]	Speed: 2178.794317 samples/sec	accuracy=0.351225	lr=0.285754
Epoch[46] Batch [399]	Speed: 3566.605055 samples/sec	accuracy=0.351116	lr=0.285556
Epoch[46] Batch [449]	Speed: 2007.629319 samples/sec	accuracy=0.351365	lr=0.285359
Epoch[46] Batch [499]	Speed: 1836.000476 samples/sec	accuracy=0.351393	lr=0.285161
Epoch[46] Batch [549]	Speed: 500.656510 samples/sec	accuracy=0.351312	lr=0.284964
Epoch[46] Batch [599]	Speed: 1514.178856 samples/sec	accuracy=0.351055	lr=0.284766
Epoch[46] Batch [649]	Speed: 2213.700043 samples/sec	accuracy=0.351010	lr=0.284568
Epoch[46] Batch [699]	Speed: 3822.135432 samples/sec	accuracy=0.350965	lr=0.284370
Epoch[46] Batch [749]	Speed: 1523.404735 samples/sec	accuracy=0.350971	lr=0.284172
Epoch[46] Batch [799]	Speed: 696.223191 samples/sec	accuracy=0.350881	lr=0.283974
Epoch[46] Batch [849]	Speed: 792.101054 samples/sec	accuracy=0.350749	lr=0.283776
Epoch[46] Batch [899]	Speed: 4944.089344 samples/sec	accuracy=0.350654	lr=0.283577
Epoch[46] Batch [949]	Speed: 5821.139428 samples/sec	accuracy=0.350650	lr=0.283379
Epoch[46] Batch [999]	Speed: 5686.360897 samples/sec	accuracy=0.350542	lr=0.283181
Epoch[46] Batch [1049]	Speed: 1383.055731 samples/sec	accuracy=0.350405	lr=0.282982
Epoch[46] Batch [1099]	Speed: 372.297760 samples/sec	accuracy=0.350358	lr=0.282783
Epoch[46] Batch [1149]	Speed: 5476.308310 samples/sec	accuracy=0.350472	lr=0.282584
Epoch[46] Batch [1199]	Speed: 5566.178775 samples/sec	accuracy=0.350155	lr=0.282385
Epoch[46] Batch [1249]	Speed: 504.383019 samples/sec	accuracy=0.349972	lr=0.282186
[Epoch 46] training: accuracy=0.349986
[Epoch 46] speed: 1480 samples/sec	time cost: 900.158950
[Epoch 46] validation: err-top1=0.737569 err-top5=0.496114
Epoch[47] Batch [49]	Speed: 3726.352962 samples/sec	accuracy=0.353184	lr=0.281983
Epoch[47] Batch [99]	Speed: 2675.142806 samples/sec	accuracy=0.353271	lr=0.281784
Epoch[47] Batch [149]	Speed: 878.248934 samples/sec	accuracy=0.353971	lr=0.281585
Epoch[47] Batch [199]	Speed: 1902.715205 samples/sec	accuracy=0.352935	lr=0.281385
Epoch[47] Batch [249]	Speed: 1487.316143 samples/sec	accuracy=0.353023	lr=0.281186
Epoch[47] Batch [299]	Speed: 1192.309783 samples/sec	accuracy=0.353571	lr=0.280986
Epoch[47] Batch [349]	Speed: 5739.887866 samples/sec	accuracy=0.353493	lr=0.280786
Epoch[47] Batch [399]	Speed: 5666.472324 samples/sec	accuracy=0.353306	lr=0.280587
Epoch[47] Batch [449]	Speed: 444.037984 samples/sec	accuracy=0.353047	lr=0.280387
Epoch[47] Batch [499]	Speed: 1566.388483 samples/sec	accuracy=0.352656	lr=0.280187
Epoch[47] Batch [549]	Speed: 1905.523191 samples/sec	accuracy=0.352377	lr=0.279987
Epoch[47] Batch [599]	Speed: 1459.563366 samples/sec	accuracy=0.352590	lr=0.279786
Epoch[47] Batch [649]	Speed: 1576.368229 samples/sec	accuracy=0.352691	lr=0.279586
Epoch[47] Batch [699]	Speed: 2757.280670 samples/sec	accuracy=0.352771	lr=0.279386
Epoch[47] Batch [749]	Speed: 5820.726356 samples/sec	accuracy=0.352660	lr=0.279185
Epoch[47] Batch [799]	Speed: 4054.505215 samples/sec	accuracy=0.352729	lr=0.278985
Epoch[47] Batch [849]	Speed: 753.110294 samples/sec	accuracy=0.352624	lr=0.278784
Epoch[47] Batch [899]	Speed: 1535.808580 samples/sec	accuracy=0.352645	lr=0.278583
Epoch[47] Batch [949]	Speed: 2949.865947 samples/sec	accuracy=0.352720	lr=0.278382
Epoch[47] Batch [999]	Speed: 1879.652672 samples/sec	accuracy=0.352572	lr=0.278181
Epoch[47] Batch [1049]	Speed: 5575.555106 samples/sec	accuracy=0.352382	lr=0.277980
Epoch[47] Batch [1099]	Speed: 6049.538727 samples/sec	accuracy=0.352227	lr=0.277779
Epoch[47] Batch [1149]	Speed: 3072.343720 samples/sec	accuracy=0.352386	lr=0.277578
Epoch[47] Batch [1199]	Speed: 2254.711504 samples/sec	accuracy=0.352235	lr=0.277377
Epoch[47] Batch [1249]	Speed: 751.077082 samples/sec	accuracy=0.352109	lr=0.277175
[Epoch 47] training: accuracy=0.352096
[Epoch 47] speed: 1668 samples/sec	time cost: 778.527964
[Epoch 47] validation: err-top1=0.725706 err-top5=0.482382
Epoch[48] Batch [49]	Speed: 1083.490771 samples/sec	accuracy=0.362871	lr=0.276970
Epoch[48] Batch [99]	Speed: 5076.023810 samples/sec	accuracy=0.359951	lr=0.276768
Epoch[48] Batch [149]	Speed: 1106.053296 samples/sec	accuracy=0.357858	lr=0.276566
Epoch[48] Batch [199]	Speed: 898.785723 samples/sec	accuracy=0.355776	lr=0.276365
Epoch[48] Batch [249]	Speed: 621.850667 samples/sec	accuracy=0.354492	lr=0.276163
Epoch[48] Batch [299]	Speed: 1578.582477 samples/sec	accuracy=0.354199	lr=0.275961
Epoch[48] Batch [349]	Speed: 5861.591601 samples/sec	accuracy=0.354124	lr=0.275759
Epoch[48] Batch [399]	Speed: 5799.286732 samples/sec	accuracy=0.354138	lr=0.275557
Epoch[48] Batch [449]	Speed: 5587.270719 samples/sec	accuracy=0.354234	lr=0.275354
Epoch[48] Batch [499]	Speed: 2371.834108 samples/sec	accuracy=0.353986	lr=0.275152
Epoch[48] Batch [549]	Speed: 1644.150060 samples/sec	accuracy=0.353968	lr=0.274950
Epoch[48] Batch [599]	Speed: 5993.279572 samples/sec	accuracy=0.353851	lr=0.274747
Epoch[48] Batch [649]	Speed: 5771.997426 samples/sec	accuracy=0.353822	lr=0.274545
Epoch[48] Batch [699]	Speed: 5556.526236 samples/sec	accuracy=0.353704	lr=0.274342
Epoch[48] Batch [749]	Speed: 5609.949428 samples/sec	accuracy=0.353646	lr=0.274139
Epoch[48] Batch [799]	Speed: 5847.206652 samples/sec	accuracy=0.353748	lr=0.273936
Epoch[48] Batch [849]	Speed: 5403.622057 samples/sec	accuracy=0.353593	lr=0.273733
Epoch[48] Batch [899]	Speed: 5236.497380 samples/sec	accuracy=0.353550	lr=0.273530
Epoch[48] Batch [949]	Speed: 5675.644423 samples/sec	accuracy=0.353432	lr=0.273327
Epoch[48] Batch [999]	Speed: 5764.849703 samples/sec	accuracy=0.353412	lr=0.273124
Epoch[48] Batch [1049]	Speed: 5287.081863 samples/sec	accuracy=0.353051	lr=0.272921
Epoch[48] Batch [1099]	Speed: 5910.824529 samples/sec	accuracy=0.353082	lr=0.272717
Epoch[48] Batch [1149]	Speed: 5664.794025 samples/sec	accuracy=0.352917	lr=0.272514
Epoch[48] Batch [1199]	Speed: 5760.445788 samples/sec	accuracy=0.352970	lr=0.272310
Epoch[48] Batch [1249]	Speed: 5797.274054 samples/sec	accuracy=0.352980	lr=0.272107
[Epoch 48] training: accuracy=0.352987
[Epoch 48] speed: 2656 samples/sec	time cost: 491.709521
[Epoch 48] validation: err-top1=0.706334 err-top5=0.458267
Epoch[49] Batch [49]	Speed: 5526.396841 samples/sec	accuracy=0.356563	lr=0.271899
Epoch[49] Batch [99]	Speed: 5909.664763 samples/sec	accuracy=0.356338	lr=0.271695
Epoch[49] Batch [149]	Speed: 5542.266298 samples/sec	accuracy=0.356530	lr=0.271491
Epoch[49] Batch [199]	Speed: 6132.743474 samples/sec	accuracy=0.355474	lr=0.271287
Epoch[49] Batch [249]	Speed: 5706.522710 samples/sec	accuracy=0.355465	lr=0.271083
Epoch[49] Batch [299]	Speed: 6189.017070 samples/sec	accuracy=0.355257	lr=0.270879
Epoch[49] Batch [349]	Speed: 5758.965257 samples/sec	accuracy=0.354916	lr=0.270675
Epoch[49] Batch [399]	Speed: 5984.971957 samples/sec	accuracy=0.354475	lr=0.270470
Epoch[49] Batch [449]	Speed: 5893.795138 samples/sec	accuracy=0.354280	lr=0.270266
Epoch[49] Batch [499]	Speed: 5704.376298 samples/sec	accuracy=0.354102	lr=0.270061
Epoch[49] Batch [549]	Speed: 6010.266615 samples/sec	accuracy=0.353951	lr=0.269857
Epoch[49] Batch [599]	Speed: 5671.898318 samples/sec	accuracy=0.354110	lr=0.269652
Epoch[49] Batch [649]	Speed: 5255.382318 samples/sec	accuracy=0.354159	lr=0.269447
Epoch[49] Batch [699]	Speed: 5916.713334 samples/sec	accuracy=0.354064	lr=0.269243
Epoch[49] Batch [749]	Speed: 5953.870551 samples/sec	accuracy=0.354051	lr=0.269038
Epoch[49] Batch [799]	Speed: 5872.301071 samples/sec	accuracy=0.354120	lr=0.268833
Epoch[49] Batch [849]	Speed: 5769.112808 samples/sec	accuracy=0.353876	lr=0.268628
Epoch[49] Batch [899]	Speed: 5634.168608 samples/sec	accuracy=0.353761	lr=0.268422
Epoch[49] Batch [949]	Speed: 6004.505827 samples/sec	accuracy=0.353651	lr=0.268217
Epoch[49] Batch [999]	Speed: 6115.322586 samples/sec	accuracy=0.353564	lr=0.268012
Epoch[49] Batch [1049]	Speed: 5838.453897 samples/sec	accuracy=0.353595	lr=0.267807
Epoch[49] Batch [1099]	Speed: 5748.209658 samples/sec	accuracy=0.353480	lr=0.267601
Epoch[49] Batch [1149]	Speed: 5469.217755 samples/sec	accuracy=0.353355	lr=0.267396
Epoch[49] Batch [1199]	Speed: 5577.714592 samples/sec	accuracy=0.353363	lr=0.267190
Epoch[49] Batch [1249]	Speed: 5764.504464 samples/sec	accuracy=0.353185	lr=0.266984
[Epoch 49] training: accuracy=0.353178
[Epoch 49] speed: 5784 samples/sec	time cost: 231.267520
[Epoch 49] validation: err-top1=0.674745 err-top5=0.422054
Epoch[50] Batch [49]	Speed: 5831.718561 samples/sec	accuracy=0.353750	lr=0.266774
Epoch[50] Batch [99]	Speed: 5809.217116 samples/sec	accuracy=0.354326	lr=0.266568
Epoch[50] Batch [149]	Speed: 5788.242842 samples/sec	accuracy=0.354720	lr=0.266362
Epoch[50] Batch [199]	Speed: 5671.777877 samples/sec	accuracy=0.354346	lr=0.266156
Epoch[50] Batch [249]	Speed: 5803.579767 samples/sec	accuracy=0.354535	lr=0.265950
Epoch[50] Batch [299]	Speed: 5395.377934 samples/sec	accuracy=0.354476	lr=0.265744
Epoch[50] Batch [349]	Speed: 5917.630932 samples/sec	accuracy=0.355078	lr=0.265538
Epoch[50] Batch [399]	Speed: 5811.740574 samples/sec	accuracy=0.354841	lr=0.265331
Epoch[50] Batch [449]	Speed: 5627.600552 samples/sec	accuracy=0.355098	lr=0.265125
Epoch[50] Batch [499]	Speed: 5805.455261 samples/sec	accuracy=0.355184	lr=0.264918
Epoch[50] Batch [549]	Speed: 5926.785463 samples/sec	accuracy=0.354966	lr=0.264712
Epoch[50] Batch [599]	Speed: 5917.745244 samples/sec	accuracy=0.354992	lr=0.264505
Epoch[50] Batch [649]	Speed: 5528.663301 samples/sec	accuracy=0.354869	lr=0.264298
Epoch[50] Batch [699]	Speed: 5575.320895 samples/sec	accuracy=0.354960	lr=0.264092
Epoch[50] Batch [749]	Speed: 5654.071077 samples/sec	accuracy=0.354768	lr=0.263885
Epoch[50] Batch [799]	Speed: 5853.972247 samples/sec	accuracy=0.354845	lr=0.263678
Epoch[50] Batch [849]	Speed: 5948.416183 samples/sec	accuracy=0.354979	lr=0.263471
Epoch[50] Batch [899]	Speed: 6096.477665 samples/sec	accuracy=0.355008	lr=0.263264
Epoch[50] Batch [949]	Speed: 6039.044748 samples/sec	accuracy=0.354946	lr=0.263056
Epoch[50] Batch [999]	Speed: 5910.337469 samples/sec	accuracy=0.354995	lr=0.262849
Epoch[50] Batch [1049]	Speed: 5757.431152 samples/sec	accuracy=0.354753	lr=0.262642
Epoch[50] Batch [1099]	Speed: 5850.350008 samples/sec	accuracy=0.354847	lr=0.262434
Epoch[50] Batch [1149]	Speed: 5566.512209 samples/sec	accuracy=0.354930	lr=0.262227
Epoch[50] Batch [1199]	Speed: 5791.792933 samples/sec	accuracy=0.354833	lr=0.262019
Epoch[50] Batch [1249]	Speed: 5879.196520 samples/sec	accuracy=0.354782	lr=0.261812
[Epoch 50] training: accuracy=0.354778
[Epoch 50] speed: 5776 samples/sec	time cost: 231.640129
[Epoch 50] validation: err-top1=0.900530 err-top5=0.771484
Epoch[51] Batch [49]	Speed: 5882.982218 samples/sec	accuracy=0.357129	lr=0.261600
Epoch[51] Batch [99]	Speed: 5915.311237 samples/sec	accuracy=0.357646	lr=0.261392
Epoch[51] Batch [149]	Speed: 5486.603192 samples/sec	accuracy=0.357611	lr=0.261184
Epoch[51] Batch [199]	Speed: 5764.978308 samples/sec	accuracy=0.357236	lr=0.260976
Epoch[51] Batch [249]	Speed: 5624.258556 samples/sec	accuracy=0.357254	lr=0.260768
Epoch[51] Batch [299]	Speed: 5782.742807 samples/sec	accuracy=0.357158	lr=0.260560
Epoch[51] Batch [349]	Speed: 5839.669089 samples/sec	accuracy=0.357045	lr=0.260352
Epoch[51] Batch [399]	Speed: 5809.133358 samples/sec	accuracy=0.356587	lr=0.260144
Epoch[51] Batch [449]	Speed: 6028.311603 samples/sec	accuracy=0.356404	lr=0.259935
Epoch[51] Batch [499]	Speed: 5881.358949 samples/sec	accuracy=0.356451	lr=0.259727
Epoch[51] Batch [549]	Speed: 6074.010803 samples/sec	accuracy=0.356365	lr=0.259519
Epoch[51] Batch [599]	Speed: 6066.890521 samples/sec	accuracy=0.356527	lr=0.259310
Epoch[51] Batch [649]	Speed: 6254.793321 samples/sec	accuracy=0.356481	lr=0.259102
Epoch[51] Batch [699]	Speed: 5947.152022 samples/sec	accuracy=0.356268	lr=0.258893
Epoch[51] Batch [749]	Speed: 6126.580097 samples/sec	accuracy=0.356158	lr=0.258684
Epoch[51] Batch [799]	Speed: 5840.022119 samples/sec	accuracy=0.356133	lr=0.258475
Epoch[51] Batch [849]	Speed: 5163.675943 samples/sec	accuracy=0.356157	lr=0.258267
Epoch[51] Batch [899]	Speed: 5708.749044 samples/sec	accuracy=0.355896	lr=0.258058
Epoch[51] Batch [949]	Speed: 5833.405810 samples/sec	accuracy=0.355760	lr=0.257849
Epoch[51] Batch [999]	Speed: 6004.441022 samples/sec	accuracy=0.355796	lr=0.257640
Epoch[51] Batch [1049]	Speed: 5441.934429 samples/sec	accuracy=0.355730	lr=0.257430
Epoch[51] Batch [1099]	Speed: 5457.393539 samples/sec	accuracy=0.355716	lr=0.257221
Epoch[51] Batch [1149]	Speed: 5726.771499 samples/sec	accuracy=0.355656	lr=0.257012
Epoch[51] Batch [1199]	Speed: 5782.083262 samples/sec	accuracy=0.355706	lr=0.256803
Epoch[51] Batch [1249]	Speed: 6080.474749 samples/sec	accuracy=0.355670	lr=0.256593
[Epoch 51] training: accuracy=0.355673
[Epoch 51] speed: 5802 samples/sec	time cost: 230.538382
[Epoch 51] validation: err-top1=0.680544 err-top5=0.424625
Epoch[52] Batch [49]	Speed: 6187.477257 samples/sec	accuracy=0.362539	lr=0.256379
Epoch[52] Batch [99]	Speed: 5379.939879 samples/sec	accuracy=0.361514	lr=0.256170
Epoch[52] Batch [149]	Speed: 5700.959167 samples/sec	accuracy=0.360814	lr=0.255960
Epoch[52] Batch [199]	Speed: 5674.882659 samples/sec	accuracy=0.359995	lr=0.255751
Epoch[52] Batch [249]	Speed: 5769.067398 samples/sec	accuracy=0.359555	lr=0.255541
Epoch[52] Batch [299]	Speed: 6073.074298 samples/sec	accuracy=0.358945	lr=0.255331
Epoch[52] Batch [349]	Speed: 5853.632207 samples/sec	accuracy=0.358733	lr=0.255121
Epoch[52] Batch [399]	Speed: 5838.915210 samples/sec	accuracy=0.358979	lr=0.254911
Epoch[52] Batch [449]	Speed: 5663.388534 samples/sec	accuracy=0.358789	lr=0.254701
Epoch[52] Batch [499]	Speed: 5849.275346 samples/sec	accuracy=0.358084	lr=0.254491
Epoch[52] Batch [549]	Speed: 5804.908365 samples/sec	accuracy=0.357882	lr=0.254281
Epoch[52] Batch [599]	Speed: 5918.408701 samples/sec	accuracy=0.357666	lr=0.254071
Epoch[52] Batch [649]	Speed: 5919.870362 samples/sec	accuracy=0.357625	lr=0.253861
Epoch[52] Batch [699]	Speed: 5801.278398 samples/sec	accuracy=0.357766	lr=0.253650
Epoch[52] Batch [749]	Speed: 5829.465877 samples/sec	accuracy=0.357530	lr=0.253440
Epoch[52] Batch [799]	Speed: 5838.282789 samples/sec	accuracy=0.357494	lr=0.253229
Epoch[52] Batch [849]	Speed: 6053.730496 samples/sec	accuracy=0.357333	lr=0.253019
Epoch[52] Batch [899]	Speed: 5726.901007 samples/sec	accuracy=0.357300	lr=0.252808
Epoch[52] Batch [949]	Speed: 5755.622805 samples/sec	accuracy=0.357090	lr=0.252598
Epoch[52] Batch [999]	Speed: 5518.098816 samples/sec	accuracy=0.357038	lr=0.252387
Epoch[52] Batch [1049]	Speed: 5858.148333 samples/sec	accuracy=0.356767	lr=0.252176
Epoch[52] Batch [1099]	Speed: 5649.351087 samples/sec	accuracy=0.356808	lr=0.251965
Epoch[52] Batch [1149]	Speed: 6214.213628 samples/sec	accuracy=0.356755	lr=0.251754
Epoch[52] Batch [1199]	Speed: 5860.389979 samples/sec	accuracy=0.356720	lr=0.251543
Epoch[52] Batch [1249]	Speed: 6065.435027 samples/sec	accuracy=0.356654	lr=0.251332
[Epoch 52] training: accuracy=0.356655
[Epoch 52] speed: 5817 samples/sec	time cost: 230.010851
[Epoch 52] validation: err-top1=0.735433 err-top5=0.493510
Epoch[53] Batch [49]	Speed: 5412.866473 samples/sec	accuracy=0.361426	lr=0.251117
Epoch[53] Batch [99]	Speed: 6186.229386 samples/sec	accuracy=0.361914	lr=0.250906
Epoch[53] Batch [149]	Speed: 5714.216203 samples/sec	accuracy=0.361999	lr=0.250695
Epoch[53] Batch [199]	Speed: 5915.318732 samples/sec	accuracy=0.361738	lr=0.250483
Epoch[53] Batch [249]	Speed: 6024.600442 samples/sec	accuracy=0.360258	lr=0.250272
Epoch[53] Batch [299]	Speed: 5536.611715 samples/sec	accuracy=0.360260	lr=0.250061
Epoch[53] Batch [349]	Speed: 5731.660503 samples/sec	accuracy=0.360075	lr=0.249849
Epoch[53] Batch [399]	Speed: 5612.946232 samples/sec	accuracy=0.359666	lr=0.249638
Epoch[53] Batch [449]	Speed: 5709.471504 samples/sec	accuracy=0.359538	lr=0.249426
Epoch[53] Batch [499]	Speed: 6037.893201 samples/sec	accuracy=0.359061	lr=0.249215
Epoch[53] Batch [549]	Speed: 5979.231730 samples/sec	accuracy=0.359206	lr=0.249003
Epoch[53] Batch [599]	Speed: 5750.704558 samples/sec	accuracy=0.359019	lr=0.248791
Epoch[53] Batch [649]	Speed: 5332.786669 samples/sec	accuracy=0.358618	lr=0.248579
Epoch[53] Batch [699]	Speed: 5835.297780 samples/sec	accuracy=0.358463	lr=0.248367
Epoch[53] Batch [749]	Speed: 5876.787359 samples/sec	accuracy=0.358555	lr=0.248156
Epoch[53] Batch [799]	Speed: 5885.794082 samples/sec	accuracy=0.358433	lr=0.247944
Epoch[53] Batch [849]	Speed: 5679.461570 samples/sec	accuracy=0.358178	lr=0.247732
Epoch[53] Batch [899]	Speed: 5212.340217 samples/sec	accuracy=0.358124	lr=0.247519
Epoch[53] Batch [949]	Speed: 6004.357919 samples/sec	accuracy=0.357980	lr=0.247307
Epoch[53] Batch [999]	Speed: 6019.443295 samples/sec	accuracy=0.357809	lr=0.247095
Epoch[53] Batch [1049]	Speed: 5649.875157 samples/sec	accuracy=0.357581	lr=0.246883
Epoch[53] Batch [1099]	Speed: 5622.528763 samples/sec	accuracy=0.357633	lr=0.246670
Epoch[53] Batch [1149]	Speed: 6154.002273 samples/sec	accuracy=0.357677	lr=0.246458
Epoch[53] Batch [1199]	Speed: 5886.674682 samples/sec	accuracy=0.357729	lr=0.246246
Epoch[53] Batch [1249]	Speed: 5813.012173 samples/sec	accuracy=0.357760	lr=0.246033
[Epoch 53] training: accuracy=0.357750
[Epoch 53] speed: 5765 samples/sec	time cost: 232.289376
[Epoch 53] validation: err-top1=0.843670 err-top5=0.658004
Epoch[54] Batch [49]	Speed: 5620.566567 samples/sec	accuracy=0.363848	lr=0.245816
Epoch[54] Batch [99]	Speed: 5682.432376 samples/sec	accuracy=0.361064	lr=0.245604
Epoch[54] Batch [149]	Speed: 6109.439741 samples/sec	accuracy=0.361439	lr=0.245391
Epoch[54] Batch [199]	Speed: 5595.230925 samples/sec	accuracy=0.360220	lr=0.245178
Epoch[54] Batch [249]	Speed: 5733.752945 samples/sec	accuracy=0.360848	lr=0.244966
Epoch[54] Batch [299]	Speed: 6093.449882 samples/sec	accuracy=0.360781	lr=0.244753
Epoch[54] Batch [349]	Speed: 6067.722424 samples/sec	accuracy=0.360938	lr=0.244540
Epoch[54] Batch [399]	Speed: 5798.727689 samples/sec	accuracy=0.360527	lr=0.244327
Epoch[54] Batch [449]	Speed: 5476.384421 samples/sec	accuracy=0.360410	lr=0.244114
Epoch[54] Batch [499]	Speed: 5997.286017 samples/sec	accuracy=0.360432	lr=0.243901
Epoch[54] Batch [549]	Speed: 6013.393625 samples/sec	accuracy=0.360167	lr=0.243688
Epoch[54] Batch [599]	Speed: 5800.816275 samples/sec	accuracy=0.360041	lr=0.243475
Epoch[54] Batch [649]	Speed: 5920.149594 samples/sec	accuracy=0.359725	lr=0.243262
Epoch[54] Batch [699]	Speed: 5782.825806 samples/sec	accuracy=0.359975	lr=0.243049
Epoch[54] Batch [749]	Speed: 5902.782069 samples/sec	accuracy=0.359931	lr=0.242835
Epoch[54] Batch [799]	Speed: 5854.071346 samples/sec	accuracy=0.359814	lr=0.242622
Epoch[54] Batch [849]	Speed: 5625.087237 samples/sec	accuracy=0.359655	lr=0.242409
Epoch[54] Batch [899]	Speed: 5895.799976 samples/sec	accuracy=0.359729	lr=0.242195
Epoch[54] Batch [949]	Speed: 5713.099013 samples/sec	accuracy=0.359749	lr=0.241982
Epoch[54] Batch [999]	Speed: 5727.841335 samples/sec	accuracy=0.359769	lr=0.241768
Epoch[54] Batch [1049]	Speed: 5720.368182 samples/sec	accuracy=0.359820	lr=0.241555
Epoch[54] Batch [1099]	Speed: 6020.305272 samples/sec	accuracy=0.359796	lr=0.241341
Epoch[54] Batch [1149]	Speed: 5532.914115 samples/sec	accuracy=0.359642	lr=0.241127
Epoch[54] Batch [1199]	Speed: 5634.427747 samples/sec	accuracy=0.359494	lr=0.240914
Epoch[54] Batch [1249]	Speed: 6092.292013 samples/sec	accuracy=0.359428	lr=0.240700
[Epoch 54] training: accuracy=0.359422
[Epoch 54] speed: 5803 samples/sec	time cost: 230.212016
[Epoch 54] validation: err-top1=0.687919 err-top5=0.442901
Epoch[55] Batch [49]	Speed: 5549.470783 samples/sec	accuracy=0.366406	lr=0.240482
Epoch[55] Batch [99]	Speed: 6157.378482 samples/sec	accuracy=0.364180	lr=0.240268
Epoch[55] Batch [149]	Speed: 6287.764088 samples/sec	accuracy=0.362663	lr=0.240054
Epoch[55] Batch [199]	Speed: 5558.870722 samples/sec	accuracy=0.362178	lr=0.239840
Epoch[55] Batch [249]	Speed: 5586.076629 samples/sec	accuracy=0.362746	lr=0.239626
Epoch[55] Batch [299]	Speed: 6069.713538 samples/sec	accuracy=0.362061	lr=0.239412
Epoch[55] Batch [349]	Speed: 5790.563389 samples/sec	accuracy=0.362461	lr=0.239198
Epoch[55] Batch [399]	Speed: 5594.743616 samples/sec	accuracy=0.362283	lr=0.238984
Epoch[55] Batch [449]	Speed: 5736.049302 samples/sec	accuracy=0.362578	lr=0.238769
Epoch[55] Batch [499]	Speed: 6029.787597 samples/sec	accuracy=0.362658	lr=0.238555
Epoch[55] Batch [549]	Speed: 5978.978858 samples/sec	accuracy=0.362401	lr=0.238341
Epoch[55] Batch [599]	Speed: 5885.036473 samples/sec	accuracy=0.362249	lr=0.238126
Epoch[55] Batch [649]	Speed: 5440.992984 samples/sec	accuracy=0.362036	lr=0.237912
Epoch[55] Batch [699]	Speed: 5881.410332 samples/sec	accuracy=0.362023	lr=0.237698
Epoch[55] Batch [749]	Speed: 5754.808271 samples/sec	accuracy=0.361708	lr=0.237483
Epoch[55] Batch [799]	Speed: 5819.245903 samples/sec	accuracy=0.361470	lr=0.237269
Epoch[55] Batch [849]	Speed: 5828.950046 samples/sec	accuracy=0.361398	lr=0.237054
Epoch[55] Batch [899]	Speed: 5681.294813 samples/sec	accuracy=0.361239	lr=0.236839
Epoch[55] Batch [949]	Speed: 5755.182734 samples/sec	accuracy=0.361103	lr=0.236625
Epoch[55] Batch [999]	Speed: 5831.259018 samples/sec	accuracy=0.361230	lr=0.236410
Epoch[55] Batch [1049]	Speed: 6145.601316 samples/sec	accuracy=0.361098	lr=0.236195
Epoch[55] Batch [1099]	Speed: 5780.967392 samples/sec	accuracy=0.361294	lr=0.235981
Epoch[55] Batch [1149]	Speed: 5092.460455 samples/sec	accuracy=0.361416	lr=0.235766
Epoch[55] Batch [1199]	Speed: 5557.134893 samples/sec	accuracy=0.361266	lr=0.235551
Epoch[55] Batch [1249]	Speed: 5892.832686 samples/sec	accuracy=0.360988	lr=0.235336
[Epoch 55] training: accuracy=0.360983
[Epoch 55] speed: 5770 samples/sec	time cost: 231.739240
[Epoch 55] validation: err-top1=0.789461 err-top5=0.578723
Epoch[56] Batch [49]	Speed: 6076.381870 samples/sec	accuracy=0.367617	lr=0.235117
Epoch[56] Batch [99]	Speed: 5740.241823 samples/sec	accuracy=0.365146	lr=0.234902
Epoch[56] Batch [149]	Speed: 5946.126131 samples/sec	accuracy=0.365176	lr=0.234687
Epoch[56] Batch [199]	Speed: 5497.359565 samples/sec	accuracy=0.365190	lr=0.234472
Epoch[56] Batch [249]	Speed: 5358.418545 samples/sec	accuracy=0.365211	lr=0.234256
Epoch[56] Batch [299]	Speed: 5853.377881 samples/sec	accuracy=0.364993	lr=0.234041
Epoch[56] Batch [349]	Speed: 6223.580080 samples/sec	accuracy=0.364498	lr=0.233826
Epoch[56] Batch [399]	Speed: 5994.360786 samples/sec	accuracy=0.364207	lr=0.233611
Epoch[56] Batch [449]	Speed: 5995.002372 samples/sec	accuracy=0.363798	lr=0.233396
Epoch[56] Batch [499]	Speed: 5726.311703 samples/sec	accuracy=0.363395	lr=0.233180
Epoch[56] Batch [549]	Speed: 5937.357452 samples/sec	accuracy=0.363599	lr=0.232965
Epoch[56] Batch [599]	Speed: 5504.613909 samples/sec	accuracy=0.363553	lr=0.232749
Epoch[56] Batch [649]	Speed: 5676.177435 samples/sec	accuracy=0.363501	lr=0.232534
Epoch[56] Batch [699]	Speed: 5670.022178 samples/sec	accuracy=0.363238	lr=0.232319
Epoch[56] Batch [749]	Speed: 5980.987606 samples/sec	accuracy=0.363301	lr=0.232103
Epoch[56] Batch [799]	Speed: 5786.743312 samples/sec	accuracy=0.363185	lr=0.231887
Epoch[56] Batch [849]	Speed: 5897.820594 samples/sec	accuracy=0.362989	lr=0.231672
Epoch[56] Batch [899]	Speed: 5505.271509 samples/sec	accuracy=0.363052	lr=0.231456
Epoch[56] Batch [949]	Speed: 5821.657033 samples/sec	accuracy=0.362812	lr=0.231241
Epoch[56] Batch [999]	Speed: 5919.817652 samples/sec	accuracy=0.362826	lr=0.231025
Epoch[56] Batch [1049]	Speed: 5991.081041 samples/sec	accuracy=0.362673	lr=0.230809
Epoch[56] Batch [1099]	Speed: 5619.241455 samples/sec	accuracy=0.362612	lr=0.230593
Epoch[56] Batch [1149]	Speed: 5681.953212 samples/sec	accuracy=0.362514	lr=0.230377
Epoch[56] Batch [1199]	Speed: 5826.121249 samples/sec	accuracy=0.362384	lr=0.230162
Epoch[56] Batch [1249]	Speed: 6021.742056 samples/sec	accuracy=0.362147	lr=0.229946
[Epoch 56] training: accuracy=0.362152
[Epoch 56] speed: 5797 samples/sec	time cost: 230.472836
[Epoch 56] validation: err-top1=0.741968 err-top5=0.504564
Epoch[57] Batch [49]	Speed: 5537.519002 samples/sec	accuracy=0.365684	lr=0.229725
Epoch[57] Batch [99]	Speed: 5693.022086 samples/sec	accuracy=0.365195	lr=0.229510
Epoch[57] Batch [149]	Speed: 5165.989742 samples/sec	accuracy=0.365091	lr=0.229294
Epoch[57] Batch [199]	Speed: 5866.170497 samples/sec	accuracy=0.364380	lr=0.229077
Epoch[57] Batch [249]	Speed: 5873.825033 samples/sec	accuracy=0.364234	lr=0.228861
Epoch[57] Batch [299]	Speed: 5834.822929 samples/sec	accuracy=0.364199	lr=0.228645
Epoch[57] Batch [349]	Speed: 5556.701213 samples/sec	accuracy=0.364093	lr=0.228429
Epoch[57] Batch [399]	Speed: 5957.738513 samples/sec	accuracy=0.364082	lr=0.228213
Epoch[57] Batch [449]	Speed: 5603.729683 samples/sec	accuracy=0.364212	lr=0.227997
Epoch[57] Batch [499]	Speed: 5726.122675 samples/sec	accuracy=0.363416	lr=0.227781
Epoch[57] Batch [549]	Speed: 5811.274424 samples/sec	accuracy=0.363503	lr=0.227564
Epoch[57] Batch [599]	Speed: 5824.992427 samples/sec	accuracy=0.363641	lr=0.227348
Epoch[57] Batch [649]	Speed: 5999.143179 samples/sec	accuracy=0.363654	lr=0.227132
Epoch[57] Batch [699]	Speed: 6057.324211 samples/sec	accuracy=0.364011	lr=0.226915
Epoch[57] Batch [749]	Speed: 5898.093538 samples/sec	accuracy=0.363954	lr=0.226699
Epoch[57] Batch [799]	Speed: 5803.719987 samples/sec	accuracy=0.363875	lr=0.226482
Epoch[57] Batch [849]	Speed: 5537.304539 samples/sec	accuracy=0.363724	lr=0.226266
Epoch[57] Batch [899]	Speed: 5366.455611 samples/sec	accuracy=0.363607	lr=0.226049
Epoch[57] Batch [949]	Speed: 5713.114364 samples/sec	accuracy=0.363592	lr=0.225833
Epoch[57] Batch [999]	Speed: 5799.160664 samples/sec	accuracy=0.363631	lr=0.225616
Epoch[57] Batch [1049]	Speed: 5351.352943 samples/sec	accuracy=0.363803	lr=0.225400
Epoch[57] Batch [1099]	Speed: 6151.985091 samples/sec	accuracy=0.363599	lr=0.225183
Epoch[57] Batch [1149]	Speed: 5428.886289 samples/sec	accuracy=0.363454	lr=0.224967
Epoch[57] Batch [1199]	Speed: 5915.799931 samples/sec	accuracy=0.363534	lr=0.224750
Epoch[57] Batch [1249]	Speed: 5788.774899 samples/sec	accuracy=0.363505	lr=0.224533
[Epoch 57] training: accuracy=0.363516
[Epoch 57] speed: 5714 samples/sec	time cost: 234.098407
[Epoch 57] validation: err-top1=0.739597 err-top5=0.498087
Epoch[58] Batch [49]	Speed: 5625.534752 samples/sec	accuracy=0.365352	lr=0.224312
Epoch[58] Batch [99]	Speed: 5686.986886 samples/sec	accuracy=0.365371	lr=0.224095
Epoch[58] Batch [149]	Speed: 4975.722917 samples/sec	accuracy=0.365436	lr=0.223879
Epoch[58] Batch [199]	Speed: 5672.057265 samples/sec	accuracy=0.365259	lr=0.223662
Epoch[58] Batch [249]	Speed: 5887.027932 samples/sec	accuracy=0.365320	lr=0.223445
Epoch[58] Batch [299]	Speed: 5893.378646 samples/sec	accuracy=0.365114	lr=0.223228
Epoch[58] Batch [349]	Speed: 6140.081056 samples/sec	accuracy=0.365198	lr=0.223011
Epoch[58] Batch [399]	Speed: 5907.510725 samples/sec	accuracy=0.364758	lr=0.222794
Epoch[58] Batch [449]	Speed: 6015.061627 samples/sec	accuracy=0.364720	lr=0.222577
Epoch[58] Batch [499]	Speed: 5411.367062 samples/sec	accuracy=0.364918	lr=0.222360
Epoch[58] Batch [549]	Speed: 5520.878187 samples/sec	accuracy=0.364789	lr=0.222143
Epoch[58] Batch [599]	Speed: 5556.432210 samples/sec	accuracy=0.364591	lr=0.221926
Epoch[58] Batch [649]	Speed: 5682.266531 samples/sec	accuracy=0.364252	lr=0.221709
Epoch[58] Batch [699]	Speed: 5692.292919 samples/sec	accuracy=0.364314	lr=0.221492
Epoch[58] Batch [749]	Speed: 5939.826891 samples/sec	accuracy=0.364473	lr=0.221275
Epoch[58] Batch [799]	Speed: 6360.897307 samples/sec	accuracy=0.364497	lr=0.221058
Epoch[58] Batch [849]	Speed: 5982.692178 samples/sec	accuracy=0.364602	lr=0.220841
Epoch[58] Batch [899]	Speed: 6094.796903 samples/sec	accuracy=0.364606	lr=0.220623
Epoch[58] Batch [949]	Speed: 5931.640651 samples/sec	accuracy=0.364640	lr=0.220406
Epoch[58] Batch [999]	Speed: 6033.936262 samples/sec	accuracy=0.364427	lr=0.220189
Epoch[58] Batch [1049]	Speed: 5713.391760 samples/sec	accuracy=0.364400	lr=0.219972
Epoch[58] Batch [1099]	Speed: 5783.331324 samples/sec	accuracy=0.364274	lr=0.219754
Epoch[58] Batch [1149]	Speed: 5868.407383 samples/sec	accuracy=0.364175	lr=0.219537
Epoch[58] Batch [1199]	Speed: 6072.299306 samples/sec	accuracy=0.364135	lr=0.219320
Epoch[58] Batch [1249]	Speed: 5850.977235 samples/sec	accuracy=0.364202	lr=0.219102
[Epoch 58] training: accuracy=0.364198
[Epoch 58] speed: 5791 samples/sec	time cost: 230.792423
[Epoch 58] validation: err-top1=0.785868 err-top5=0.568095
Epoch[59] Batch [49]	Speed: 5243.527632 samples/sec	accuracy=0.374805	lr=0.218881
Epoch[59] Batch [99]	Speed: 5733.663695 samples/sec	accuracy=0.371816	lr=0.218663
Epoch[59] Batch [149]	Speed: 6028.242730 samples/sec	accuracy=0.370710	lr=0.218446
Epoch[59] Batch [199]	Speed: 5948.065248 samples/sec	accuracy=0.370068	lr=0.218228
Epoch[59] Batch [249]	Speed: 5493.371293 samples/sec	accuracy=0.370086	lr=0.218011
Epoch[59] Batch [299]	Speed: 5535.384673 samples/sec	accuracy=0.369463	lr=0.217793
Epoch[59] Batch [349]	Speed: 5905.884776 samples/sec	accuracy=0.369037	lr=0.217576
Epoch[59] Batch [399]	Speed: 5881.035369 samples/sec	accuracy=0.368123	lr=0.217358
Epoch[59] Batch [449]	Speed: 5933.319338 samples/sec	accuracy=0.367928	lr=0.217141
Epoch[59] Batch [499]	Speed: 5804.508890 samples/sec	accuracy=0.368051	lr=0.216923
Epoch[59] Batch [549]	Speed: 6038.737886 samples/sec	accuracy=0.367843	lr=0.216706
Epoch[59] Batch [599]	Speed: 5820.545873 samples/sec	accuracy=0.367477	lr=0.216488
Epoch[59] Batch [649]	Speed: 5924.479670 samples/sec	accuracy=0.367505	lr=0.216270
Epoch[59] Batch [699]	Speed: 6042.976422 samples/sec	accuracy=0.367380	lr=0.216053
Epoch[59] Batch [749]	Speed: 5535.692881 samples/sec	accuracy=0.367320	lr=0.215835
Epoch[59] Batch [799]	Speed: 5931.168665 samples/sec	accuracy=0.366997	lr=0.215617
Epoch[59] Batch [849]	Speed: 5387.626018 samples/sec	accuracy=0.366960	lr=0.215400
Epoch[59] Batch [899]	Speed: 5898.027122 samples/sec	accuracy=0.366714	lr=0.215182
Epoch[59] Batch [949]	Speed: 5718.246364 samples/sec	accuracy=0.366637	lr=0.214964
Epoch[59] Batch [999]	Speed: 5741.735769 samples/sec	accuracy=0.366647	lr=0.214746
Epoch[59] Batch [1049]	Speed: 6071.139334 samples/sec	accuracy=0.366466	lr=0.214529
Epoch[59] Batch [1099]	Speed: 5968.240349 samples/sec	accuracy=0.366466	lr=0.214311
Epoch[59] Batch [1149]	Speed: 6347.663569 samples/sec	accuracy=0.366544	lr=0.214093
Epoch[59] Batch [1199]	Speed: 5916.234757 samples/sec	accuracy=0.366458	lr=0.213875
Epoch[59] Batch [1249]	Speed: 5907.062395 samples/sec	accuracy=0.366329	lr=0.213657
[Epoch 59] training: accuracy=0.366330
[Epoch 59] speed: 5812 samples/sec	time cost: 230.269580
[Epoch 59] validation: err-top1=0.649195 err-top5=0.390585
Epoch[60] Batch [49]	Speed: 5536.097455 samples/sec	accuracy=0.373711	lr=0.213435
Epoch[60] Batch [99]	Speed: 6019.926734 samples/sec	accuracy=0.372705	lr=0.213217
Epoch[60] Batch [149]	Speed: 6134.218838 samples/sec	accuracy=0.370124	lr=0.212999
Epoch[60] Batch [199]	Speed: 5591.973120 samples/sec	accuracy=0.369805	lr=0.212781
Epoch[60] Batch [249]	Speed: 6027.620403 samples/sec	accuracy=0.369055	lr=0.212563
Epoch[60] Batch [299]	Speed: 5262.208596 samples/sec	accuracy=0.368665	lr=0.212345
Epoch[60] Batch [349]	Speed: 5578.117363 samples/sec	accuracy=0.368471	lr=0.212127
Epoch[60] Batch [399]	Speed: 5936.067303 samples/sec	accuracy=0.368462	lr=0.211909
Epoch[60] Batch [449]	Speed: 6149.003701 samples/sec	accuracy=0.368357	lr=0.211691
Epoch[60] Batch [499]	Speed: 5950.019973 samples/sec	accuracy=0.367953	lr=0.211473
Epoch[60] Batch [549]	Speed: 5901.469599 samples/sec	accuracy=0.367901	lr=0.211255
Epoch[60] Batch [599]	Speed: 5729.785597 samples/sec	accuracy=0.367918	lr=0.211037
Epoch[60] Batch [649]	Speed: 5666.027092 samples/sec	accuracy=0.367967	lr=0.210819
Epoch[60] Batch [699]	Speed: 5968.412193 samples/sec	accuracy=0.367510	lr=0.210601
Epoch[60] Batch [749]	Speed: 5865.503160 samples/sec	accuracy=0.367358	lr=0.210383
Epoch[60] Batch [799]	Speed: 5993.599396 samples/sec	accuracy=0.367407	lr=0.210165
Epoch[60] Batch [849]	Speed: 5617.733999 samples/sec	accuracy=0.367487	lr=0.209947
Epoch[60] Batch [899]	Speed: 5604.735021 samples/sec	accuracy=0.367406	lr=0.209729
Epoch[60] Batch [949]	Speed: 6095.435601 samples/sec	accuracy=0.367467	lr=0.209511
Epoch[60] Batch [999]	Speed: 6358.848618 samples/sec	accuracy=0.367630	lr=0.209293
Epoch[60] Batch [1049]	Speed: 6213.874143 samples/sec	accuracy=0.367467	lr=0.209075
Epoch[60] Batch [1099]	Speed: 5776.520025 samples/sec	accuracy=0.367537	lr=0.208856
Epoch[60] Batch [1149]	Speed: 6080.331167 samples/sec	accuracy=0.367441	lr=0.208638
Epoch[60] Batch [1199]	Speed: 5665.433359 samples/sec	accuracy=0.367544	lr=0.208420
Epoch[60] Batch [1249]	Speed: 5695.304959 samples/sec	accuracy=0.367476	lr=0.208202
[Epoch 60] training: accuracy=0.367486
[Epoch 60] speed: 5838 samples/sec	time cost: 229.142670
[Epoch 60] validation: err-top1=0.681382 err-top5=0.428890
Epoch[61] Batch [49]	Speed: 5838.606284 samples/sec	accuracy=0.374316	lr=0.207979
Epoch[61] Batch [99]	Speed: 5988.031620 samples/sec	accuracy=0.372129	lr=0.207761
Epoch[61] Batch [149]	Speed: 5251.777121 samples/sec	accuracy=0.371693	lr=0.207543
Epoch[61] Batch [199]	Speed: 5233.097200 samples/sec	accuracy=0.371694	lr=0.207325
Epoch[61] Batch [249]	Speed: 6083.679914 samples/sec	accuracy=0.370637	lr=0.207106
Epoch[61] Batch [299]	Speed: 5813.735768 samples/sec	accuracy=0.370423	lr=0.206888
Epoch[61] Batch [349]	Speed: 5530.587480 samples/sec	accuracy=0.369827	lr=0.206670
Epoch[61] Batch [399]	Speed: 5996.617318 samples/sec	accuracy=0.370234	lr=0.206452
Epoch[61] Batch [449]	Speed: 5486.596744 samples/sec	accuracy=0.370111	lr=0.206233
Epoch[61] Batch [499]	Speed: 5434.471111 samples/sec	accuracy=0.370340	lr=0.206015
Epoch[61] Batch [549]	Speed: 5946.221624 samples/sec	accuracy=0.370046	lr=0.205797
Epoch[61] Batch [599]	Speed: 5626.031715 samples/sec	accuracy=0.369569	lr=0.205579
Epoch[61] Batch [649]	Speed: 5859.982192 samples/sec	accuracy=0.369593	lr=0.205360
Epoch[61] Batch [699]	Speed: 5903.329388 samples/sec	accuracy=0.369088	lr=0.205142
Epoch[61] Batch [749]	Speed: 5715.520016 samples/sec	accuracy=0.369234	lr=0.204924
Epoch[61] Batch [799]	Speed: 5782.930920 samples/sec	accuracy=0.369137	lr=0.204705
Epoch[61] Batch [849]	Speed: 5898.091594 samples/sec	accuracy=0.369053	lr=0.204487
Epoch[61] Batch [899]	Speed: 5724.531852 samples/sec	accuracy=0.369172	lr=0.204269
Epoch[61] Batch [949]	Speed: 6072.737178 samples/sec	accuracy=0.369190	lr=0.204051
Epoch[61] Batch [999]	Speed: 5376.867441 samples/sec	accuracy=0.369249	lr=0.203832
Epoch[61] Batch [1049]	Speed: 5796.152470 samples/sec	accuracy=0.369071	lr=0.203614
Epoch[61] Batch [1099]	Speed: 5739.970099 samples/sec	accuracy=0.368981	lr=0.203396
Epoch[61] Batch [1149]	Speed: 5934.367873 samples/sec	accuracy=0.368869	lr=0.203177
Epoch[61] Batch [1199]	Speed: 5507.756989 samples/sec	accuracy=0.368901	lr=0.202959
Epoch[61] Batch [1249]	Speed: 5692.985261 samples/sec	accuracy=0.368855	lr=0.202740
[Epoch 61] training: accuracy=0.368846
[Epoch 61] speed: 5711 samples/sec	time cost: 234.120560
[Epoch 61] validation: err-top1=0.758949 err-top5=0.528938
Epoch[62] Batch [49]	Speed: 5376.334510 samples/sec	accuracy=0.373027	lr=0.202518
Epoch[62] Batch [99]	Speed: 5649.750448 samples/sec	accuracy=0.370937	lr=0.202299
Epoch[62] Batch [149]	Speed: 5789.888322 samples/sec	accuracy=0.369928	lr=0.202081
Epoch[62] Batch [199]	Speed: 5518.679937 samples/sec	accuracy=0.370132	lr=0.201863
Epoch[62] Batch [249]	Speed: 5777.014651 samples/sec	accuracy=0.369977	lr=0.201644
Epoch[62] Batch [299]	Speed: 5280.579807 samples/sec	accuracy=0.369840	lr=0.201426
Epoch[62] Batch [349]	Speed: 5742.433282 samples/sec	accuracy=0.370081	lr=0.201208
Epoch[62] Batch [399]	Speed: 6047.030367 samples/sec	accuracy=0.369746	lr=0.200989
Epoch[62] Batch [449]	Speed: 5599.973172 samples/sec	accuracy=0.370022	lr=0.200771
Epoch[62] Batch [499]	Speed: 5504.196852 samples/sec	accuracy=0.370146	lr=0.200552
Epoch[62] Batch [549]	Speed: 5772.996232 samples/sec	accuracy=0.370073	lr=0.200334
Epoch[62] Batch [599]	Speed: 5781.306357 samples/sec	accuracy=0.370449	lr=0.200116
Epoch[62] Batch [649]	Speed: 5906.587327 samples/sec	accuracy=0.370428	lr=0.199897
Epoch[62] Batch [699]	Speed: 5795.137504 samples/sec	accuracy=0.370218	lr=0.199679
Epoch[62] Batch [749]	Speed: 5787.142373 samples/sec	accuracy=0.370100	lr=0.199461
Epoch[62] Batch [799]	Speed: 5695.519752 samples/sec	accuracy=0.370002	lr=0.199242
Epoch[62] Batch [849]	Speed: 5970.823866 samples/sec	accuracy=0.369867	lr=0.199024
Epoch[62] Batch [899]	Speed: 5789.307522 samples/sec	accuracy=0.369707	lr=0.198806
Epoch[62] Batch [949]	Speed: 5740.031929 samples/sec	accuracy=0.369632	lr=0.198587
Epoch[62] Batch [999]	Speed: 5865.867333 samples/sec	accuracy=0.369718	lr=0.198369
Epoch[62] Batch [1049]	Speed: 5940.744935 samples/sec	accuracy=0.369714	lr=0.198150
Epoch[62] Batch [1099]	Speed: 6191.748353 samples/sec	accuracy=0.369762	lr=0.197932
Epoch[62] Batch [1149]	Speed: 5933.494587 samples/sec	accuracy=0.369738	lr=0.197714
Epoch[62] Batch [1199]	Speed: 6274.912903 samples/sec	accuracy=0.369770	lr=0.197495
Epoch[62] Batch [1249]	Speed: 5867.325918 samples/sec	accuracy=0.369755	lr=0.197277
[Epoch 62] training: accuracy=0.369734
[Epoch 62] speed: 5769 samples/sec	time cost: 231.581158
[Epoch 62] validation: err-top1=0.756896 err-top5=0.533741
Epoch[63] Batch [49]	Speed: 5945.566239 samples/sec	accuracy=0.371719	lr=0.197054
Epoch[63] Batch [99]	Speed: 5762.183261 samples/sec	accuracy=0.373027	lr=0.196836
Epoch[63] Batch [149]	Speed: 5893.744347 samples/sec	accuracy=0.374225	lr=0.196618
Epoch[63] Batch [199]	Speed: 6043.432866 samples/sec	accuracy=0.371812	lr=0.196399
Epoch[63] Batch [249]	Speed: 5598.714239 samples/sec	accuracy=0.372074	lr=0.196181
Epoch[63] Batch [299]	Speed: 5604.014838 samples/sec	accuracy=0.371777	lr=0.195963
Epoch[63] Batch [349]	Speed: 5557.629480 samples/sec	accuracy=0.371663	lr=0.195744
Epoch[63] Batch [399]	Speed: 5810.010187 samples/sec	accuracy=0.371521	lr=0.195526
Epoch[63] Batch [449]	Speed: 5422.507717 samples/sec	accuracy=0.371695	lr=0.195308
Epoch[63] Batch [499]	Speed: 5632.142593 samples/sec	accuracy=0.371574	lr=0.195089
Epoch[63] Batch [549]	Speed: 5397.134207 samples/sec	accuracy=0.371550	lr=0.194871
Epoch[63] Batch [599]	Speed: 5688.838253 samples/sec	accuracy=0.371379	lr=0.194653
Epoch[63] Batch [649]	Speed: 5753.081714 samples/sec	accuracy=0.371124	lr=0.194434
Epoch[63] Batch [699]	Speed: 5744.854626 samples/sec	accuracy=0.371318	lr=0.194216
Epoch[63] Batch [749]	Speed: 6173.271254 samples/sec	accuracy=0.371176	lr=0.193998
Epoch[63] Batch [799]	Speed: 5577.680402 samples/sec	accuracy=0.371028	lr=0.193780
Epoch[63] Batch [849]	Speed: 6000.666961 samples/sec	accuracy=0.371060	lr=0.193561
Epoch[63] Batch [899]	Speed: 5658.511483 samples/sec	accuracy=0.371039	lr=0.193343
Epoch[63] Batch [949]	Speed: 5692.379075 samples/sec	accuracy=0.371200	lr=0.193125
Epoch[63] Batch [999]	Speed: 5901.058833 samples/sec	accuracy=0.371181	lr=0.192907
Epoch[63] Batch [1049]	Speed: 6118.098659 samples/sec	accuracy=0.371159	lr=0.192688
Epoch[63] Batch [1099]	Speed: 5412.208392 samples/sec	accuracy=0.371244	lr=0.192470
Epoch[63] Batch [1149]	Speed: 5766.466577 samples/sec	accuracy=0.371301	lr=0.192252
Epoch[63] Batch [1199]	Speed: 6039.908951 samples/sec	accuracy=0.371210	lr=0.192034
Epoch[63] Batch [1249]	Speed: 5760.924064 samples/sec	accuracy=0.371151	lr=0.191816
[Epoch 63] training: accuracy=0.371157
[Epoch 63] speed: 5742 samples/sec	time cost: 232.238257
[Epoch 63] validation: err-top1=0.728190 err-top5=0.485942
Epoch[64] Batch [49]	Speed: 5820.005121 samples/sec	accuracy=0.375078	lr=0.191593
Epoch[64] Batch [99]	Speed: 5827.062822 samples/sec	accuracy=0.374219	lr=0.191375
Epoch[64] Batch [149]	Speed: 6104.900437 samples/sec	accuracy=0.375098	lr=0.191157
Epoch[64] Batch [199]	Speed: 5964.522743 samples/sec	accuracy=0.375098	lr=0.190939
Epoch[64] Batch [249]	Speed: 6086.472004 samples/sec	accuracy=0.374789	lr=0.190720
Epoch[64] Batch [299]	Speed: 5759.150436 samples/sec	accuracy=0.374196	lr=0.190502
Epoch[64] Batch [349]	Speed: 6155.817138 samples/sec	accuracy=0.373278	lr=0.190284
Epoch[64] Batch [399]	Speed: 5640.923445 samples/sec	accuracy=0.373252	lr=0.190066
Epoch[64] Batch [449]	Speed: 5742.575016 samples/sec	accuracy=0.373216	lr=0.189848
Epoch[64] Batch [499]	Speed: 5629.815900 samples/sec	accuracy=0.373477	lr=0.189630
Epoch[64] Batch [549]	Speed: 5556.869443 samples/sec	accuracy=0.373672	lr=0.189412
Epoch[64] Batch [599]	Speed: 5891.705669 samples/sec	accuracy=0.373457	lr=0.189194
Epoch[64] Batch [649]	Speed: 5877.244134 samples/sec	accuracy=0.373460	lr=0.188976
Epoch[64] Batch [699]	Speed: 6007.241328 samples/sec	accuracy=0.373446	lr=0.188758
Epoch[64] Batch [749]	Speed: 5506.738268 samples/sec	accuracy=0.373395	lr=0.188540
Epoch[64] Batch [799]	Speed: 5735.755761 samples/sec	accuracy=0.373250	lr=0.188322
Epoch[64] Batch [849]	Speed: 6192.201302 samples/sec	accuracy=0.373295	lr=0.188104
Epoch[64] Batch [899]	Speed: 5595.016341 samples/sec	accuracy=0.373148	lr=0.187886
Epoch[64] Batch [949]	Speed: 6121.495604 samples/sec	accuracy=0.373156	lr=0.187668
Epoch[64] Batch [999]	Speed: 5921.184826 samples/sec	accuracy=0.373104	lr=0.187450
Epoch[64] Batch [1049]	Speed: 6084.073062 samples/sec	accuracy=0.373008	lr=0.187232
Epoch[64] Batch [1099]	Speed: 5785.418028 samples/sec	accuracy=0.373061	lr=0.187014
Epoch[64] Batch [1149]	Speed: 5998.107315 samples/sec	accuracy=0.372859	lr=0.186796
Epoch[64] Batch [1199]	Speed: 5376.677760 samples/sec	accuracy=0.372943	lr=0.186578
Epoch[64] Batch [1249]	Speed: 5743.130964 samples/sec	accuracy=0.373045	lr=0.186360
[Epoch 64] training: accuracy=0.373022
[Epoch 64] speed: 5830 samples/sec	time cost: 229.978566
[Epoch 64] validation: err-top1=0.709064 err-top5=0.473473
Epoch[65] Batch [49]	Speed: 5876.493870 samples/sec	accuracy=0.377383	lr=0.186138
Epoch[65] Batch [99]	Speed: 5775.325535 samples/sec	accuracy=0.377549	lr=0.185920
Epoch[65] Batch [149]	Speed: 5655.129556 samples/sec	accuracy=0.377272	lr=0.185702
Epoch[65] Batch [199]	Speed: 5842.826599 samples/sec	accuracy=0.376084	lr=0.185485
Epoch[65] Batch [249]	Speed: 5939.683467 samples/sec	accuracy=0.375508	lr=0.185267
Epoch[65] Batch [299]	Speed: 5802.825456 samples/sec	accuracy=0.375579	lr=0.185049
Epoch[65] Batch [349]	Speed: 6107.021621 samples/sec	accuracy=0.375592	lr=0.184831
Epoch[65] Batch [399]	Speed: 5856.929433 samples/sec	accuracy=0.375374	lr=0.184614
Epoch[65] Batch [449]	Speed: 5882.380658 samples/sec	accuracy=0.375171	lr=0.184396
Epoch[65] Batch [499]	Speed: 5773.050394 samples/sec	accuracy=0.375229	lr=0.184178
Epoch[65] Batch [549]	Speed: 5707.971390 samples/sec	accuracy=0.374915	lr=0.183960
Epoch[65] Batch [599]	Speed: 5610.411687 samples/sec	accuracy=0.374679	lr=0.183743
Epoch[65] Batch [649]	Speed: 5652.707802 samples/sec	accuracy=0.374823	lr=0.183525
Epoch[65] Batch [699]	Speed: 6171.119413 samples/sec	accuracy=0.374600	lr=0.183308
Epoch[65] Batch [749]	Speed: 5680.224564 samples/sec	accuracy=0.374568	lr=0.183090
Epoch[65] Batch [799]	Speed: 5604.782269 samples/sec	accuracy=0.374541	lr=0.182872
Epoch[65] Batch [849]	Speed: 5925.489766 samples/sec	accuracy=0.374355	lr=0.182655
Epoch[65] Batch [899]	Speed: 5664.117187 samples/sec	accuracy=0.374257	lr=0.182437
Epoch[65] Batch [949]	Speed: 5618.248104 samples/sec	accuracy=0.374222	lr=0.182220
Epoch[65] Batch [999]	Speed: 6168.995121 samples/sec	accuracy=0.373996	lr=0.182002
Epoch[65] Batch [1049]	Speed: 5693.775443 samples/sec	accuracy=0.374170	lr=0.181785
Epoch[65] Batch [1099]	Speed: 5706.810234 samples/sec	accuracy=0.374019	lr=0.181567
Epoch[65] Batch [1149]	Speed: 5980.439284 samples/sec	accuracy=0.374029	lr=0.181350
Epoch[65] Batch [1199]	Speed: 5694.938247 samples/sec	accuracy=0.374089	lr=0.181132
Epoch[65] Batch [1249]	Speed: 5676.421546 samples/sec	accuracy=0.374206	lr=0.180915
[Epoch 65] training: accuracy=0.374218
[Epoch 65] speed: 5792 samples/sec	time cost: 231.065297
[Epoch 65] validation: err-top1=0.681441 err-top5=0.436005
Epoch[66] Batch [49]	Speed: 5711.995783 samples/sec	accuracy=0.374844	lr=0.180693
Epoch[66] Batch [99]	Speed: 5503.448255 samples/sec	accuracy=0.376729	lr=0.180476
Epoch[66] Batch [149]	Speed: 5280.278059 samples/sec	accuracy=0.377728	lr=0.180259
Epoch[66] Batch [199]	Speed: 5942.138071 samples/sec	accuracy=0.377866	lr=0.180041
Epoch[66] Batch [249]	Speed: 5535.425480 samples/sec	accuracy=0.377801	lr=0.179824
Epoch[66] Batch [299]	Speed: 5877.858157 samples/sec	accuracy=0.377708	lr=0.179607
Epoch[66] Batch [349]	Speed: 5432.099137 samples/sec	accuracy=0.377380	lr=0.179390
Epoch[66] Batch [399]	Speed: 6006.084071 samples/sec	accuracy=0.377715	lr=0.179172
Epoch[66] Batch [449]	Speed: 5582.460558 samples/sec	accuracy=0.378242	lr=0.178955
Epoch[66] Batch [499]	Speed: 5622.404522 samples/sec	accuracy=0.377844	lr=0.178738
Epoch[66] Batch [549]	Speed: 5789.627174 samples/sec	accuracy=0.378031	lr=0.178521
Epoch[66] Batch [599]	Speed: 5875.856015 samples/sec	accuracy=0.377744	lr=0.178304
Epoch[66] Batch [649]	Speed: 6113.402902 samples/sec	accuracy=0.377459	lr=0.178087
Epoch[66] Batch [699]	Speed: 5683.661702 samples/sec	accuracy=0.377208	lr=0.177870
Epoch[66] Batch [749]	Speed: 5497.753912 samples/sec	accuracy=0.377004	lr=0.177653
Epoch[66] Batch [799]	Speed: 5965.203027 samples/sec	accuracy=0.376808	lr=0.177436
Epoch[66] Batch [849]	Speed: 5771.038358 samples/sec	accuracy=0.376581	lr=0.177219
Epoch[66] Batch [899]	Speed: 5828.185804 samples/sec	accuracy=0.376475	lr=0.177002
Epoch[66] Batch [949]	Speed: 6170.849341 samples/sec	accuracy=0.376293	lr=0.176785
Epoch[66] Batch [999]	Speed: 5986.691485 samples/sec	accuracy=0.376144	lr=0.176568
Epoch[66] Batch [1049]	Speed: 5597.986263 samples/sec	accuracy=0.376090	lr=0.176351
Epoch[66] Batch [1099]	Speed: 5918.494171 samples/sec	accuracy=0.376098	lr=0.176134
Epoch[66] Batch [1149]	Speed: 5725.267776 samples/sec	accuracy=0.376040	lr=0.175918
Epoch[66] Batch [1199]	Speed: 5749.109438 samples/sec	accuracy=0.375922	lr=0.175701
Epoch[66] Batch [1249]	Speed: 5150.277483 samples/sec	accuracy=0.375887	lr=0.175484
[Epoch 66] training: accuracy=0.375902
[Epoch 66] speed: 5712 samples/sec	time cost: 233.876581
[Epoch 66] validation: err-top1=0.827966 err-top5=0.625917
Epoch[67] Batch [49]	Speed: 5772.430142 samples/sec	accuracy=0.379336	lr=0.175263
Epoch[67] Batch [99]	Speed: 5880.268036 samples/sec	accuracy=0.377207	lr=0.175046
Epoch[67] Batch [149]	Speed: 6126.096852 samples/sec	accuracy=0.377988	lr=0.174830
Epoch[67] Batch [199]	Speed: 5729.508442 samples/sec	accuracy=0.378252	lr=0.174613
Epoch[67] Batch [249]	Speed: 5888.177538 samples/sec	accuracy=0.378105	lr=0.174397
Epoch[67] Batch [299]	Speed: 5940.249974 samples/sec	accuracy=0.377428	lr=0.174180
Epoch[67] Batch [349]	Speed: 5751.024890 samples/sec	accuracy=0.377946	lr=0.173964
Epoch[67] Batch [399]	Speed: 5935.877299 samples/sec	accuracy=0.377734	lr=0.173747
Epoch[67] Batch [449]	Speed: 5956.803479 samples/sec	accuracy=0.377374	lr=0.173531
Epoch[67] Batch [499]	Speed: 5783.859207 samples/sec	accuracy=0.377811	lr=0.173314
Epoch[67] Batch [549]	Speed: 5851.599016 samples/sec	accuracy=0.377807	lr=0.173098
Epoch[67] Batch [599]	Speed: 5761.121270 samples/sec	accuracy=0.377642	lr=0.172881
Epoch[67] Batch [649]	Speed: 5161.294500 samples/sec	accuracy=0.377931	lr=0.172665
Epoch[67] Batch [699]	Speed: 5742.193593 samples/sec	accuracy=0.377718	lr=0.172449
Epoch[67] Batch [749]	Speed: 5883.750904 samples/sec	accuracy=0.377577	lr=0.172232
Epoch[67] Batch [799]	Speed: 6134.455573 samples/sec	accuracy=0.377426	lr=0.172016
Epoch[67] Batch [849]	Speed: 5530.062377 samples/sec	accuracy=0.377207	lr=0.171800
Epoch[67] Batch [899]	Speed: 5557.353053 samples/sec	accuracy=0.377103	lr=0.171584
Epoch[67] Batch [949]	Speed: 6020.981459 samples/sec	accuracy=0.377067	lr=0.171368
Epoch[67] Batch [999]	Speed: 5804.388400 samples/sec	accuracy=0.376937	lr=0.171152
Epoch[67] Batch [1049]	Speed: 5663.336559 samples/sec	accuracy=0.377157	lr=0.170935
Epoch[67] Batch [1099]	Speed: 6012.303346 samples/sec	accuracy=0.377220	lr=0.170719
Epoch[67] Batch [1149]	Speed: 5859.628023 samples/sec	accuracy=0.377229	lr=0.170503
Epoch[67] Batch [1199]	Speed: 5765.583337 samples/sec	accuracy=0.377205	lr=0.170287
Epoch[67] Batch [1249]	Speed: 5727.045030 samples/sec	accuracy=0.377166	lr=0.170072
[Epoch 67] training: accuracy=0.377159
[Epoch 67] speed: 5795 samples/sec	time cost: 230.797858
[Epoch 67] validation: err-top1=0.643296 err-top5=0.386739
Epoch[68] Batch [49]	Speed: 5432.358984 samples/sec	accuracy=0.382285	lr=0.169851
Epoch[68] Batch [99]	Speed: 5393.010294 samples/sec	accuracy=0.384150	lr=0.169635
Epoch[68] Batch [149]	Speed: 5273.067151 samples/sec	accuracy=0.382324	lr=0.169420
Epoch[68] Batch [199]	Speed: 6060.027172 samples/sec	accuracy=0.381592	lr=0.169204
Epoch[68] Batch [249]	Speed: 5830.733213 samples/sec	accuracy=0.381059	lr=0.168988
Epoch[68] Batch [299]	Speed: 5829.585195 samples/sec	accuracy=0.380557	lr=0.168772
Epoch[68] Batch [349]	Speed: 5960.067957 samples/sec	accuracy=0.380020	lr=0.168557
Epoch[68] Batch [399]	Speed: 5548.325763 samples/sec	accuracy=0.379910	lr=0.168341
Epoch[68] Batch [449]	Speed: 5648.874068 samples/sec	accuracy=0.379857	lr=0.168126
Epoch[68] Batch [499]	Speed: 5609.062643 samples/sec	accuracy=0.379812	lr=0.167910
Epoch[68] Batch [549]	Speed: 5987.261319 samples/sec	accuracy=0.379609	lr=0.167694
Epoch[68] Batch [599]	Speed: 5477.957114 samples/sec	accuracy=0.379596	lr=0.167479
Epoch[68] Batch [649]	Speed: 5569.942221 samples/sec	accuracy=0.379495	lr=0.167263
Epoch[68] Batch [699]	Speed: 6204.169395 samples/sec	accuracy=0.379449	lr=0.167048
Epoch[68] Batch [749]	Speed: 5939.768896 samples/sec	accuracy=0.379658	lr=0.166833
Epoch[68] Batch [799]	Speed: 5973.059220 samples/sec	accuracy=0.379589	lr=0.166617
Epoch[68] Batch [849]	Speed: 6030.534670 samples/sec	accuracy=0.379455	lr=0.166402
Epoch[68] Batch [899]	Speed: 5949.358311 samples/sec	accuracy=0.379266	lr=0.166187
Epoch[68] Batch [949]	Speed: 5840.436504 samples/sec	accuracy=0.379124	lr=0.165972
Epoch[68] Batch [999]	Speed: 6041.559573 samples/sec	accuracy=0.379025	lr=0.165756
Epoch[68] Batch [1049]	Speed: 5889.045770 samples/sec	accuracy=0.378985	lr=0.165541
Epoch[68] Batch [1099]	Speed: 5869.947455 samples/sec	accuracy=0.378853	lr=0.165326
Epoch[68] Batch [1149]	Speed: 5907.347894 samples/sec	accuracy=0.378818	lr=0.165111
Epoch[68] Batch [1199]	Speed: 5849.993177 samples/sec	accuracy=0.378838	lr=0.164896
Epoch[68] Batch [1249]	Speed: 5662.237383 samples/sec	accuracy=0.378641	lr=0.164681
[Epoch 68] training: accuracy=0.378650
[Epoch 68] speed: 5772 samples/sec	time cost: 231.530442
[Epoch 68] validation: err-top1=0.671397 err-top5=0.415478
Epoch[69] Batch [49]	Speed: 5780.390559 samples/sec	accuracy=0.384238	lr=0.164462
Epoch[69] Batch [99]	Speed: 5865.696697 samples/sec	accuracy=0.382227	lr=0.164247
Epoch[69] Batch [149]	Speed: 5749.339391 samples/sec	accuracy=0.381471	lr=0.164032
Epoch[69] Batch [199]	Speed: 5982.245030 samples/sec	accuracy=0.381367	lr=0.163818
Epoch[69] Batch [249]	Speed: 5714.042720 samples/sec	accuracy=0.381566	lr=0.163603
Epoch[69] Batch [299]	Speed: 5937.517837 samples/sec	accuracy=0.381533	lr=0.163388
Epoch[69] Batch [349]	Speed: 5667.765807 samples/sec	accuracy=0.381164	lr=0.163173
Epoch[69] Batch [399]	Speed: 5503.052527 samples/sec	accuracy=0.381235	lr=0.162959
Epoch[69] Batch [449]	Speed: 5956.711941 samples/sec	accuracy=0.381042	lr=0.162744
Epoch[69] Batch [499]	Speed: 5919.912955 samples/sec	accuracy=0.380920	lr=0.162530
Epoch[69] Batch [549]	Speed: 6038.816849 samples/sec	accuracy=0.380531	lr=0.162315
Epoch[69] Batch [599]	Speed: 5955.475797 samples/sec	accuracy=0.380267	lr=0.162101
Epoch[69] Batch [649]	Speed: 5631.423324 samples/sec	accuracy=0.380135	lr=0.161886
Epoch[69] Batch [699]	Speed: 5406.374520 samples/sec	accuracy=0.380035	lr=0.161672
Epoch[69] Batch [749]	Speed: 5912.122771 samples/sec	accuracy=0.379801	lr=0.161458
Epoch[69] Batch [799]	Speed: 5808.264806 samples/sec	accuracy=0.379922	lr=0.161244
Epoch[69] Batch [849]	Speed: 5912.501709 samples/sec	accuracy=0.379871	lr=0.161029
Epoch[69] Batch [899]	Speed: 5217.511633 samples/sec	accuracy=0.379776	lr=0.160815
Epoch[69] Batch [949]	Speed: 5644.886312 samples/sec	accuracy=0.379683	lr=0.160601
Epoch[69] Batch [999]	Speed: 5524.426687 samples/sec	accuracy=0.379841	lr=0.160387
Epoch[69] Batch [1049]	Speed: 5765.563833 samples/sec	accuracy=0.379839	lr=0.160173
Epoch[69] Batch [1099]	Speed: 5396.029623 samples/sec	accuracy=0.379902	lr=0.159959
Epoch[69] Batch [1149]	Speed: 5366.333981 samples/sec	accuracy=0.379808	lr=0.159745
Epoch[69] Batch [1199]	Speed: 5602.873515 samples/sec	accuracy=0.379925	lr=0.159531
Epoch[69] Batch [1249]	Speed: 5319.702689 samples/sec	accuracy=0.379955	lr=0.159317
[Epoch 69] training: accuracy=0.379970
[Epoch 69] speed: 5687 samples/sec	time cost: 234.590888
[Epoch 69] validation: err-top1=0.677592 err-top5=0.421407
Epoch[70] Batch [49]	Speed: 5848.569159 samples/sec	accuracy=0.388730	lr=0.159099
Epoch[70] Batch [99]	Speed: 5818.061579 samples/sec	accuracy=0.386914	lr=0.158886
Epoch[70] Batch [149]	Speed: 6087.534474 samples/sec	accuracy=0.386348	lr=0.158672
Epoch[70] Batch [199]	Speed: 6069.049517 samples/sec	accuracy=0.385854	lr=0.158458
Epoch[70] Batch [249]	Speed: 5835.780004 samples/sec	accuracy=0.384480	lr=0.158245
Epoch[70] Batch [299]	Speed: 5963.572991 samples/sec	accuracy=0.383535	lr=0.158031
Epoch[70] Batch [349]	Speed: 5523.605944 samples/sec	accuracy=0.383412	lr=0.157818
Epoch[70] Batch [399]	Speed: 5972.439099 samples/sec	accuracy=0.383508	lr=0.157604
Epoch[70] Batch [449]	Speed: 5365.012090 samples/sec	accuracy=0.383173	lr=0.157391
Epoch[70] Batch [499]	Speed: 5486.224740 samples/sec	accuracy=0.383229	lr=0.157177
Epoch[70] Batch [549]	Speed: 5911.666092 samples/sec	accuracy=0.382814	lr=0.156964
Epoch[70] Batch [599]	Speed: 5646.841770 samples/sec	accuracy=0.382443	lr=0.156751
Epoch[70] Batch [649]	Speed: 5831.925237 samples/sec	accuracy=0.381872	lr=0.156538
Epoch[70] Batch [699]	Speed: 5564.793956 samples/sec	accuracy=0.381875	lr=0.156325
Epoch[70] Batch [749]	Speed: 5783.726020 samples/sec	accuracy=0.381839	lr=0.156112
Epoch[70] Batch [799]	Speed: 5881.761662 samples/sec	accuracy=0.381786	lr=0.155899
Epoch[70] Batch [849]	Speed: 5500.205559 samples/sec	accuracy=0.381660	lr=0.155686
Epoch[70] Batch [899]	Speed: 6318.869651 samples/sec	accuracy=0.381699	lr=0.155473
Epoch[70] Batch [949]	Speed: 5792.319862 samples/sec	accuracy=0.381657	lr=0.155260
Epoch[70] Batch [999]	Speed: 5766.995103 samples/sec	accuracy=0.381469	lr=0.155047
Epoch[70] Batch [1049]	Speed: 5519.468149 samples/sec	accuracy=0.381414	lr=0.154834
Epoch[70] Batch [1099]	Speed: 5867.134358 samples/sec	accuracy=0.381501	lr=0.154622
Epoch[70] Batch [1149]	Speed: 5954.236700 samples/sec	accuracy=0.381394	lr=0.154409
Epoch[70] Batch [1199]	Speed: 5582.179914 samples/sec	accuracy=0.381516	lr=0.154196
Epoch[70] Batch [1249]	Speed: 5574.673806 samples/sec	accuracy=0.381502	lr=0.153984
[Epoch 70] training: accuracy=0.381512
[Epoch 70] speed: 5764 samples/sec	time cost: 232.091502
[Epoch 70] validation: err-top1=0.676080 err-top5=0.423509
Epoch[71] Batch [49]	Speed: 5493.272226 samples/sec	accuracy=0.383848	lr=0.153767
Epoch[71] Batch [99]	Speed: 5442.967802 samples/sec	accuracy=0.384951	lr=0.153555
Epoch[71] Batch [149]	Speed: 5941.155000 samples/sec	accuracy=0.386419	lr=0.153342
Epoch[71] Batch [199]	Speed: 5259.286509 samples/sec	accuracy=0.385557	lr=0.153130
Epoch[71] Batch [249]	Speed: 5955.147479 samples/sec	accuracy=0.384809	lr=0.152918
Epoch[71] Batch [299]	Speed: 5610.833414 samples/sec	accuracy=0.384079	lr=0.152705
Epoch[71] Batch [349]	Speed: 5529.912427 samples/sec	accuracy=0.383987	lr=0.152493
Epoch[71] Batch [399]	Speed: 6139.016662 samples/sec	accuracy=0.384038	lr=0.152281
Epoch[71] Batch [449]	Speed: 5939.372657 samples/sec	accuracy=0.383811	lr=0.152069
Epoch[71] Batch [499]	Speed: 5997.307120 samples/sec	accuracy=0.383885	lr=0.151857
Epoch[71] Batch [549]	Speed: 5733.108354 samples/sec	accuracy=0.383944	lr=0.151645
Epoch[71] Batch [599]	Speed: 5550.403235 samples/sec	accuracy=0.383607	lr=0.151433
Epoch[71] Batch [649]	Speed: 5549.124331 samples/sec	accuracy=0.383413	lr=0.151222
Epoch[71] Batch [699]	Speed: 5692.607983 samples/sec	accuracy=0.383461	lr=0.151010
Epoch[71] Batch [749]	Speed: 5543.672983 samples/sec	accuracy=0.383421	lr=0.150798
Epoch[71] Batch [799]	Speed: 5511.793063 samples/sec	accuracy=0.383484	lr=0.150587
Epoch[71] Batch [849]	Speed: 5528.970191 samples/sec	accuracy=0.383285	lr=0.150375
Epoch[71] Batch [899]	Speed: 6096.341980 samples/sec	accuracy=0.383149	lr=0.150163
Epoch[71] Batch [949]	Speed: 5872.506618 samples/sec	accuracy=0.383187	lr=0.149952
Epoch[71] Batch [999]	Speed: 6078.673059 samples/sec	accuracy=0.383094	lr=0.149741
Epoch[71] Batch [1049]	Speed: 5818.569966 samples/sec	accuracy=0.383156	lr=0.149529
Epoch[71] Batch [1099]	Speed: 5978.169114 samples/sec	accuracy=0.383034	lr=0.149318
Epoch[71] Batch [1149]	Speed: 5623.589013 samples/sec	accuracy=0.382971	lr=0.149107
Epoch[71] Batch [1199]	Speed: 5760.201967 samples/sec	accuracy=0.382969	lr=0.148896
Epoch[71] Batch [1249]	Speed: 5915.271643 samples/sec	accuracy=0.382955	lr=0.148685
[Epoch 71] training: accuracy=0.382983
[Epoch 71] speed: 5723 samples/sec	time cost: 233.632131
[Epoch 71] validation: err-top1=0.731226 err-top5=0.499581
Epoch[72] Batch [49]	Speed: 5636.432476 samples/sec	accuracy=0.385762	lr=0.148469
Epoch[72] Batch [99]	Speed: 5691.409476 samples/sec	accuracy=0.388574	lr=0.148258
Epoch[72] Batch [149]	Speed: 5788.144555 samples/sec	accuracy=0.388555	lr=0.148047
Epoch[72] Batch [199]	Speed: 5940.915035 samples/sec	accuracy=0.388452	lr=0.147837
Epoch[72] Batch [249]	Speed: 5878.271815 samples/sec	accuracy=0.387602	lr=0.147626
Epoch[72] Batch [299]	Speed: 5862.321740 samples/sec	accuracy=0.387744	lr=0.147415
Epoch[72] Batch [349]	Speed: 5755.939211 samples/sec	accuracy=0.387157	lr=0.147204
Epoch[72] Batch [399]	Speed: 5725.160932 samples/sec	accuracy=0.387075	lr=0.146994
Epoch[72] Batch [449]	Speed: 5691.027430 samples/sec	accuracy=0.387057	lr=0.146783
Epoch[72] Batch [499]	Speed: 6105.218747 samples/sec	accuracy=0.387076	lr=0.146573
Epoch[72] Batch [549]	Speed: 5605.247920 samples/sec	accuracy=0.386467	lr=0.146362
Epoch[72] Batch [599]	Speed: 6048.046748 samples/sec	accuracy=0.386185	lr=0.146152
Epoch[72] Batch [649]	Speed: 5779.831885 samples/sec	accuracy=0.385817	lr=0.145942
Epoch[72] Batch [699]	Speed: 5885.159045 samples/sec	accuracy=0.385771	lr=0.145732
Epoch[72] Batch [749]	Speed: 5976.705804 samples/sec	accuracy=0.385701	lr=0.145521
Epoch[72] Batch [799]	Speed: 5854.648933 samples/sec	accuracy=0.385696	lr=0.145311
Epoch[72] Batch [849]	Speed: 5703.829191 samples/sec	accuracy=0.385759	lr=0.145101
Epoch[72] Batch [899]	Speed: 5516.401956 samples/sec	accuracy=0.385547	lr=0.144891
Epoch[72] Batch [949]	Speed: 6091.440402 samples/sec	accuracy=0.385508	lr=0.144682
Epoch[72] Batch [999]	Speed: 5431.253260 samples/sec	accuracy=0.385506	lr=0.144472
Epoch[72] Batch [1049]	Speed: 5824.945975 samples/sec	accuracy=0.385480	lr=0.144262
Epoch[72] Batch [1099]	Speed: 5965.090519 samples/sec	accuracy=0.385421	lr=0.144052
Epoch[72] Batch [1149]	Speed: 6026.835656 samples/sec	accuracy=0.385439	lr=0.143843
Epoch[72] Batch [1199]	Speed: 5462.121642 samples/sec	accuracy=0.385185	lr=0.143633
Epoch[72] Batch [1249]	Speed: 5872.777224 samples/sec	accuracy=0.385293	lr=0.143424
[Epoch 72] training: accuracy=0.385304
[Epoch 72] speed: 5790 samples/sec	time cost: 230.902176
[Epoch 72] validation: err-top1=0.634805 err-top5=0.381238
Epoch[73] Batch [49]	Speed: 5612.919384 samples/sec	accuracy=0.386348	lr=0.143210
Epoch[73] Batch [99]	Speed: 5853.241314 samples/sec	accuracy=0.387852	lr=0.143001
Epoch[73] Batch [149]	Speed: 5846.412785 samples/sec	accuracy=0.389271	lr=0.142791
Epoch[73] Batch [199]	Speed: 5871.701051 samples/sec	accuracy=0.388574	lr=0.142582
Epoch[73] Batch [249]	Speed: 5889.736082 samples/sec	accuracy=0.388371	lr=0.142373
Epoch[73] Batch [299]	Speed: 5721.688677 samples/sec	accuracy=0.387969	lr=0.142164
Epoch[73] Batch [349]	Speed: 5576.039223 samples/sec	accuracy=0.387966	lr=0.141955
Epoch[73] Batch [399]	Speed: 5658.758402 samples/sec	accuracy=0.387825	lr=0.141746
Epoch[73] Batch [449]	Speed: 5588.599555 samples/sec	accuracy=0.387650	lr=0.141537
Epoch[73] Batch [499]	Speed: 5874.361370 samples/sec	accuracy=0.387547	lr=0.141328
Epoch[73] Batch [549]	Speed: 5845.953470 samples/sec	accuracy=0.387037	lr=0.141120
Epoch[73] Batch [599]	Speed: 5532.199872 samples/sec	accuracy=0.387121	lr=0.140911
Epoch[73] Batch [649]	Speed: 6007.106560 samples/sec	accuracy=0.386789	lr=0.140702
Epoch[73] Batch [699]	Speed: 5538.666708 samples/sec	accuracy=0.386775	lr=0.140494
Epoch[73] Batch [749]	Speed: 5856.910584 samples/sec	accuracy=0.386612	lr=0.140285
Epoch[73] Batch [799]	Speed: 5731.203133 samples/sec	accuracy=0.386388	lr=0.140077
Epoch[73] Batch [849]	Speed: 5243.260060 samples/sec	accuracy=0.386258	lr=0.139869
Epoch[73] Batch [899]	Speed: 5473.791410 samples/sec	accuracy=0.386258	lr=0.139661
Epoch[73] Batch [949]	Speed: 5939.101957 samples/sec	accuracy=0.386217	lr=0.139452
Epoch[73] Batch [999]	Speed: 5656.304785 samples/sec	accuracy=0.386025	lr=0.139244
Epoch[73] Batch [1049]	Speed: 5892.029616 samples/sec	accuracy=0.386015	lr=0.139036
Epoch[73] Batch [1099]	Speed: 5982.208534 samples/sec	accuracy=0.385851	lr=0.138828
Epoch[73] Batch [1149]	Speed: 5896.381294 samples/sec	accuracy=0.385864	lr=0.138620
Epoch[73] Batch [1199]	Speed: 5818.413893 samples/sec	accuracy=0.385890	lr=0.138413
Epoch[73] Batch [1249]	Speed: 5887.755705 samples/sec	accuracy=0.385875	lr=0.138205
[Epoch 73] training: accuracy=0.385873
[Epoch 73] speed: 5738 samples/sec	time cost: 232.703368
[Epoch 73] validation: err-top1=0.625120 err-top5=0.368025
Epoch[74] Batch [49]	Speed: 5563.189894 samples/sec	accuracy=0.392793	lr=0.137993
Epoch[74] Batch [99]	Speed: 5654.561332 samples/sec	accuracy=0.392559	lr=0.137786
Epoch[74] Batch [149]	Speed: 5882.748863 samples/sec	accuracy=0.391263	lr=0.137578
Epoch[74] Batch [199]	Speed: 6079.664820 samples/sec	accuracy=0.391084	lr=0.137371
Epoch[74] Batch [249]	Speed: 5738.112292 samples/sec	accuracy=0.390652	lr=0.137163
Epoch[74] Batch [299]	Speed: 5734.334748 samples/sec	accuracy=0.390133	lr=0.136956
Epoch[74] Batch [349]	Speed: 5625.869146 samples/sec	accuracy=0.389612	lr=0.136749
Epoch[74] Batch [399]	Speed: 5375.768848 samples/sec	accuracy=0.389595	lr=0.136542
Epoch[74] Batch [449]	Speed: 5799.997515 samples/sec	accuracy=0.389219	lr=0.136335
Epoch[74] Batch [499]	Speed: 5473.637380 samples/sec	accuracy=0.389426	lr=0.136128
Epoch[74] Batch [549]	Speed: 5733.709314 samples/sec	accuracy=0.389460	lr=0.135921
Epoch[74] Batch [599]	Speed: 5827.893828 samples/sec	accuracy=0.389355	lr=0.135714
Epoch[74] Batch [649]	Speed: 5753.715389 samples/sec	accuracy=0.389145	lr=0.135507
Epoch[74] Batch [699]	Speed: 5786.121362 samples/sec	accuracy=0.389251	lr=0.135301
Epoch[74] Batch [749]	Speed: 6092.784978 samples/sec	accuracy=0.388971	lr=0.135094
Epoch[74] Batch [799]	Speed: 5910.989829 samples/sec	accuracy=0.389031	lr=0.134887
Epoch[74] Batch [849]	Speed: 5876.076604 samples/sec	accuracy=0.388770	lr=0.134681
Epoch[74] Batch [899]	Speed: 5827.442794 samples/sec	accuracy=0.388544	lr=0.134475
Epoch[74] Batch [949]	Speed: 5765.423593 samples/sec	accuracy=0.388431	lr=0.134268
Epoch[74] Batch [999]	Speed: 5480.990639 samples/sec	accuracy=0.388404	lr=0.134062
Epoch[74] Batch [1049]	Speed: 5532.737782 samples/sec	accuracy=0.388464	lr=0.133856
Epoch[74] Batch [1099]	Speed: 6010.315229 samples/sec	accuracy=0.388329	lr=0.133650
Epoch[74] Batch [1149]	Speed: 5802.215093 samples/sec	accuracy=0.388245	lr=0.133444
Epoch[74] Batch [1199]	Speed: 5653.992477 samples/sec	accuracy=0.388162	lr=0.133238
Epoch[74] Batch [1249]	Speed: 5399.345411 samples/sec	accuracy=0.388062	lr=0.133032
[Epoch 74] training: accuracy=0.388071
[Epoch 74] speed: 5724 samples/sec	time cost: 233.536099
[Epoch 74] validation: err-top1=0.640246 err-top5=0.380600
Epoch[75] Batch [49]	Speed: 5947.772175 samples/sec	accuracy=0.392695	lr=0.132822
Epoch[75] Batch [99]	Speed: 5939.602804 samples/sec	accuracy=0.391631	lr=0.132617
Epoch[75] Batch [149]	Speed: 5651.484837 samples/sec	accuracy=0.390768	lr=0.132411
Epoch[75] Batch [199]	Speed: 5926.163301 samples/sec	accuracy=0.391172	lr=0.132206
Epoch[75] Batch [249]	Speed: 5549.175522 samples/sec	accuracy=0.391082	lr=0.132000
Epoch[75] Batch [299]	Speed: 5977.489364 samples/sec	accuracy=0.390492	lr=0.131795
Epoch[75] Batch [349]	Speed: 6132.929300 samples/sec	accuracy=0.390455	lr=0.131590
Epoch[75] Batch [399]	Speed: 6101.289415 samples/sec	accuracy=0.390527	lr=0.131385
Epoch[75] Batch [449]	Speed: 5873.101984 samples/sec	accuracy=0.390747	lr=0.131180
Epoch[75] Batch [499]	Speed: 5649.878279 samples/sec	accuracy=0.390334	lr=0.130975
Epoch[75] Batch [549]	Speed: 5806.830724 samples/sec	accuracy=0.390270	lr=0.130770
Epoch[75] Batch [599]	Speed: 5910.439137 samples/sec	accuracy=0.390099	lr=0.130565
Epoch[75] Batch [649]	Speed: 5663.858448 samples/sec	accuracy=0.389818	lr=0.130360
Epoch[75] Batch [699]	Speed: 5431.265211 samples/sec	accuracy=0.390045	lr=0.130155
Epoch[75] Batch [749]	Speed: 5624.181520 samples/sec	accuracy=0.389934	lr=0.129951
Epoch[75] Batch [799]	Speed: 6109.990071 samples/sec	accuracy=0.389844	lr=0.129746
Epoch[75] Batch [849]	Speed: 5952.782277 samples/sec	accuracy=0.389615	lr=0.129542
Epoch[75] Batch [899]	Speed: 5770.762469 samples/sec	accuracy=0.389692	lr=0.129338
Epoch[75] Batch [949]	Speed: 5682.995087 samples/sec	accuracy=0.389672	lr=0.129133
Epoch[75] Batch [999]	Speed: 5568.115733 samples/sec	accuracy=0.389703	lr=0.128929
Epoch[75] Batch [1049]	Speed: 5421.404083 samples/sec	accuracy=0.389801	lr=0.128725
Epoch[75] Batch [1099]	Speed: 5792.393137 samples/sec	accuracy=0.389858	lr=0.128521
Epoch[75] Batch [1149]	Speed: 5894.276240 samples/sec	accuracy=0.389683	lr=0.128317
Epoch[75] Batch [1199]	Speed: 5795.533031 samples/sec	accuracy=0.389604	lr=0.128113
Epoch[75] Batch [1249]	Speed: 5728.446388 samples/sec	accuracy=0.389422	lr=0.127910
[Epoch 75] training: accuracy=0.389437
[Epoch 75] speed: 5783 samples/sec	time cost: 231.324395
[Epoch 75] validation: err-top1=0.599548 err-top5=0.336914
Epoch[76] Batch [49]	Speed: 5361.734977 samples/sec	accuracy=0.391445	lr=0.127702
Epoch[76] Batch [99]	Speed: 5743.733260 samples/sec	accuracy=0.392949	lr=0.127498
Epoch[76] Batch [149]	Speed: 5733.733043 samples/sec	accuracy=0.392129	lr=0.127295
Epoch[76] Batch [199]	Speed: 5708.952104 samples/sec	accuracy=0.392544	lr=0.127092
Epoch[76] Batch [249]	Speed: 5892.710602 samples/sec	accuracy=0.391359	lr=0.126888
Epoch[76] Batch [299]	Speed: 5741.331587 samples/sec	accuracy=0.391432	lr=0.126685
Epoch[76] Batch [349]	Speed: 6158.007939 samples/sec	accuracy=0.391203	lr=0.126482
Epoch[76] Batch [399]	Speed: 5353.148855 samples/sec	accuracy=0.391426	lr=0.126279
Epoch[76] Batch [449]	Speed: 5598.374892 samples/sec	accuracy=0.391638	lr=0.126076
Epoch[76] Batch [499]	Speed: 6266.425668 samples/sec	accuracy=0.391482	lr=0.125873
Epoch[76] Batch [549]	Speed: 6114.629571 samples/sec	accuracy=0.391790	lr=0.125670
Epoch[76] Batch [599]	Speed: 5636.357325 samples/sec	accuracy=0.391538	lr=0.125468
Epoch[76] Batch [649]	Speed: 5794.401645 samples/sec	accuracy=0.391451	lr=0.125265
Epoch[76] Batch [699]	Speed: 5201.279415 samples/sec	accuracy=0.391463	lr=0.125062
Epoch[76] Batch [749]	Speed: 5334.681454 samples/sec	accuracy=0.391310	lr=0.124860
Epoch[76] Batch [799]	Speed: 5602.314573 samples/sec	accuracy=0.391226	lr=0.124658
Epoch[76] Batch [849]	Speed: 5645.591065 samples/sec	accuracy=0.391209	lr=0.124455
Epoch[76] Batch [899]	Speed: 5647.902001 samples/sec	accuracy=0.391379	lr=0.124253
Epoch[76] Batch [949]	Speed: 5633.437148 samples/sec	accuracy=0.391314	lr=0.124051
Epoch[76] Batch [999]	Speed: 5943.771707 samples/sec	accuracy=0.391355	lr=0.123849
Epoch[76] Batch [1049]	Speed: 5980.239434 samples/sec	accuracy=0.391483	lr=0.123647
Epoch[76] Batch [1099]	Speed: 5758.686043 samples/sec	accuracy=0.391485	lr=0.123446
Epoch[76] Batch [1149]	Speed: 5987.327256 samples/sec	accuracy=0.391467	lr=0.123244
Epoch[76] Batch [1199]	Speed: 5883.868425 samples/sec	accuracy=0.391391	lr=0.123042
Epoch[76] Batch [1249]	Speed: 5642.988702 samples/sec	accuracy=0.391294	lr=0.122841
[Epoch 76] training: accuracy=0.391328
[Epoch 76] speed: 5717 samples/sec	time cost: 233.695861
[Epoch 76] validation: err-top1=0.681382 err-top5=0.425562
Epoch[77] Batch [49]	Speed: 5425.991173 samples/sec	accuracy=0.393711	lr=0.122635
Epoch[77] Batch [99]	Speed: 5894.582833 samples/sec	accuracy=0.397031	lr=0.122434
Epoch[77] Batch [149]	Speed: 5556.382035 samples/sec	accuracy=0.396979	lr=0.122233
Epoch[77] Batch [199]	Speed: 5939.701374 samples/sec	accuracy=0.396504	lr=0.122032
Epoch[77] Batch [249]	Speed: 6061.019359 samples/sec	accuracy=0.395336	lr=0.121831
Epoch[77] Batch [299]	Speed: 5505.438615 samples/sec	accuracy=0.394847	lr=0.121630
Epoch[77] Batch [349]	Speed: 5944.987032 samples/sec	accuracy=0.394330	lr=0.121429
Epoch[77] Batch [399]	Speed: 5883.784596 samples/sec	accuracy=0.393955	lr=0.121228
Epoch[77] Batch [449]	Speed: 5974.190987 samples/sec	accuracy=0.393856	lr=0.121027
Epoch[77] Batch [499]	Speed: 5683.935042 samples/sec	accuracy=0.393543	lr=0.120827
Epoch[77] Batch [549]	Speed: 6029.461529 samples/sec	accuracy=0.393702	lr=0.120626
Epoch[77] Batch [599]	Speed: 5503.994977 samples/sec	accuracy=0.393525	lr=0.120426
Epoch[77] Batch [649]	Speed: 5984.718099 samples/sec	accuracy=0.393559	lr=0.120226
Epoch[77] Batch [699]	Speed: 5677.355575 samples/sec	accuracy=0.393384	lr=0.120025
Epoch[77] Batch [749]	Speed: 5664.216088 samples/sec	accuracy=0.393117	lr=0.119825
Epoch[77] Batch [799]	Speed: 5831.290844 samples/sec	accuracy=0.393157	lr=0.119625
Epoch[77] Batch [849]	Speed: 5718.709435 samples/sec	accuracy=0.393255	lr=0.119425
Epoch[77] Batch [899]	Speed: 5710.967396 samples/sec	accuracy=0.393086	lr=0.119226
Epoch[77] Batch [949]	Speed: 5965.455232 samples/sec	accuracy=0.393127	lr=0.119026
Epoch[77] Batch [999]	Speed: 5851.576375 samples/sec	accuracy=0.393188	lr=0.118826
Epoch[77] Batch [1049]	Speed: 5958.637799 samples/sec	accuracy=0.393012	lr=0.118627
Epoch[77] Batch [1099]	Speed: 5744.418965 samples/sec	accuracy=0.393029	lr=0.118427
Epoch[77] Batch [1149]	Speed: 5889.797627 samples/sec	accuracy=0.393121	lr=0.118228
Epoch[77] Batch [1199]	Speed: 5524.014296 samples/sec	accuracy=0.393198	lr=0.118029
Epoch[77] Batch [1249]	Speed: 5496.541217 samples/sec	accuracy=0.393158	lr=0.117830
[Epoch 77] training: accuracy=0.393141
[Epoch 77] speed: 5763 samples/sec	time cost: 231.836494
[Epoch 77] validation: err-top1=0.643335 err-top5=0.386539
Epoch[78] Batch [49]	Speed: 5706.114373 samples/sec	accuracy=0.397852	lr=0.117627
Epoch[78] Batch [99]	Speed: 5843.867722 samples/sec	accuracy=0.396152	lr=0.117428
Epoch[78] Batch [149]	Speed: 5892.602106 samples/sec	accuracy=0.395957	lr=0.117229
Epoch[78] Batch [199]	Speed: 6002.626887 samples/sec	accuracy=0.395444	lr=0.117030
Epoch[78] Batch [249]	Speed: 5877.016702 samples/sec	accuracy=0.395031	lr=0.116831
Epoch[78] Batch [299]	Speed: 5893.516931 samples/sec	accuracy=0.395039	lr=0.116633
Epoch[78] Batch [349]	Speed: 5747.091596 samples/sec	accuracy=0.394682	lr=0.116434
Epoch[78] Batch [399]	Speed: 5815.696265 samples/sec	accuracy=0.394778	lr=0.116236
Epoch[78] Batch [449]	Speed: 6036.925033 samples/sec	accuracy=0.394813	lr=0.116038
Epoch[78] Batch [499]	Speed: 5961.895350 samples/sec	accuracy=0.394588	lr=0.115840
Epoch[78] Batch [549]	Speed: 5367.814842 samples/sec	accuracy=0.394400	lr=0.115642
Epoch[78] Batch [599]	Speed: 5415.556254 samples/sec	accuracy=0.394533	lr=0.115444
Epoch[78] Batch [649]	Speed: 5468.269214 samples/sec	accuracy=0.394428	lr=0.115246
Epoch[78] Batch [699]	Speed: 5822.743356 samples/sec	accuracy=0.394176	lr=0.115048
Epoch[78] Batch [749]	Speed: 5893.868414 samples/sec	accuracy=0.394296	lr=0.114850
Epoch[78] Batch [799]	Speed: 5939.053010 samples/sec	accuracy=0.394347	lr=0.114653
Epoch[78] Batch [849]	Speed: 5928.065198 samples/sec	accuracy=0.394283	lr=0.114455
Epoch[78] Batch [899]	Speed: 5822.589743 samples/sec	accuracy=0.394185	lr=0.114258
Epoch[78] Batch [949]	Speed: 5768.677334 samples/sec	accuracy=0.394141	lr=0.114061
Epoch[78] Batch [999]	Speed: 5319.337424 samples/sec	accuracy=0.394337	lr=0.113864
Epoch[78] Batch [1049]	Speed: 5764.372631 samples/sec	accuracy=0.394481	lr=0.113667
Epoch[78] Batch [1099]	Speed: 5949.117848 samples/sec	accuracy=0.394420	lr=0.113470
Epoch[78] Batch [1149]	Speed: 5945.051219 samples/sec	accuracy=0.394385	lr=0.113273
Epoch[78] Batch [1199]	Speed: 5601.983997 samples/sec	accuracy=0.394582	lr=0.113076
Epoch[78] Batch [1249]	Speed: 5595.932374 samples/sec	accuracy=0.394623	lr=0.112880
[Epoch 78] training: accuracy=0.394651
[Epoch 78] speed: 5762 samples/sec	time cost: 232.060669
[Epoch 78] validation: err-top1=0.639449 err-top5=0.382015
Epoch[79] Batch [49]	Speed: 5446.838020 samples/sec	accuracy=0.398027	lr=0.112679
Epoch[79] Batch [99]	Speed: 5655.985533 samples/sec	accuracy=0.398682	lr=0.112483
Epoch[79] Batch [149]	Speed: 5723.326120 samples/sec	accuracy=0.399160	lr=0.112287
Epoch[79] Batch [199]	Speed: 5520.692402 samples/sec	accuracy=0.397573	lr=0.112090
Epoch[79] Batch [249]	Speed: 5592.307322 samples/sec	accuracy=0.397590	lr=0.111894
Epoch[79] Batch [299]	Speed: 5831.346107 samples/sec	accuracy=0.397386	lr=0.111698
Epoch[79] Batch [349]	Speed: 5964.938251 samples/sec	accuracy=0.397854	lr=0.111502
Epoch[79] Batch [399]	Speed: 5976.506870 samples/sec	accuracy=0.398096	lr=0.111307
Epoch[79] Batch [449]	Speed: 5604.437506 samples/sec	accuracy=0.398173	lr=0.111111
Epoch[79] Batch [499]	Speed: 5759.968826 samples/sec	accuracy=0.397633	lr=0.110915
Epoch[79] Batch [549]	Speed: 5414.580225 samples/sec	accuracy=0.397585	lr=0.110720
Epoch[79] Batch [599]	Speed: 5893.107270 samples/sec	accuracy=0.397700	lr=0.110525
Epoch[79] Batch [649]	Speed: 5568.869175 samples/sec	accuracy=0.397512	lr=0.110329
Epoch[79] Batch [699]	Speed: 5639.444765 samples/sec	accuracy=0.397217	lr=0.110134
Epoch[79] Batch [749]	Speed: 6017.410155 samples/sec	accuracy=0.396870	lr=0.109939
Epoch[79] Batch [799]	Speed: 5448.367660 samples/sec	accuracy=0.396548	lr=0.109744
Epoch[79] Batch [849]	Speed: 5731.710528 samples/sec	accuracy=0.396536	lr=0.109549
Epoch[79] Batch [899]	Speed: 6188.205608 samples/sec	accuracy=0.396482	lr=0.109355
Epoch[79] Batch [949]	Speed: 5836.279596 samples/sec	accuracy=0.396613	lr=0.109160
Epoch[79] Batch [999]	Speed: 5661.805952 samples/sec	accuracy=0.396493	lr=0.108966
Epoch[79] Batch [1049]	Speed: 5564.962676 samples/sec	accuracy=0.396479	lr=0.108771
Epoch[79] Batch [1099]	Speed: 5717.771189 samples/sec	accuracy=0.396562	lr=0.108577
Epoch[79] Batch [1149]	Speed: 5956.238766 samples/sec	accuracy=0.396668	lr=0.108383
Epoch[79] Batch [1199]	Speed: 5887.247907 samples/sec	accuracy=0.396794	lr=0.108189
Epoch[79] Batch [1249]	Speed: 6019.998962 samples/sec	accuracy=0.396944	lr=0.107995
[Epoch 79] training: accuracy=0.396932
[Epoch 79] speed: 5727 samples/sec	time cost: 233.282512
[Epoch 79] validation: err-top1=0.642339 err-top5=0.390745
Epoch[80] Batch [49]	Speed: 5639.695798 samples/sec	accuracy=0.404531	lr=0.107797
Epoch[80] Batch [99]	Speed: 5601.381693 samples/sec	accuracy=0.402158	lr=0.107603
Epoch[80] Batch [149]	Speed: 5942.949925 samples/sec	accuracy=0.400736	lr=0.107410
Epoch[80] Batch [199]	Speed: 5871.242408 samples/sec	accuracy=0.401523	lr=0.107216
Epoch[80] Batch [249]	Speed: 5779.883532 samples/sec	accuracy=0.401723	lr=0.107023
Epoch[80] Batch [299]	Speed: 6050.869300 samples/sec	accuracy=0.401361	lr=0.106830
Epoch[80] Batch [349]	Speed: 6083.674916 samples/sec	accuracy=0.400773	lr=0.106636
Epoch[80] Batch [399]	Speed: 5900.129345 samples/sec	accuracy=0.400823	lr=0.106443
Epoch[80] Batch [449]	Speed: 5754.321912 samples/sec	accuracy=0.400710	lr=0.106250
Epoch[80] Batch [499]	Speed: 5556.831338 samples/sec	accuracy=0.400496	lr=0.106057
Epoch[80] Batch [549]	Speed: 5933.044928 samples/sec	accuracy=0.400192	lr=0.105865
Epoch[80] Batch [599]	Speed: 5908.133040 samples/sec	accuracy=0.400374	lr=0.105672
Epoch[80] Batch [649]	Speed: 5997.505097 samples/sec	accuracy=0.399889	lr=0.105480
Epoch[80] Batch [699]	Speed: 5748.871347 samples/sec	accuracy=0.399503	lr=0.105287
Epoch[80] Batch [749]	Speed: 5736.304567 samples/sec	accuracy=0.399199	lr=0.105095
Epoch[80] Batch [799]	Speed: 5889.539018 samples/sec	accuracy=0.399175	lr=0.104903
Epoch[80] Batch [849]	Speed: 5817.545243 samples/sec	accuracy=0.398984	lr=0.104711
Epoch[80] Batch [899]	Speed: 5406.555005 samples/sec	accuracy=0.398853	lr=0.104519
Epoch[80] Batch [949]	Speed: 6021.564932 samples/sec	accuracy=0.398772	lr=0.104327
Epoch[80] Batch [999]	Speed: 5765.568322 samples/sec	accuracy=0.398418	lr=0.104135
Epoch[80] Batch [1049]	Speed: 5862.077380 samples/sec	accuracy=0.398477	lr=0.103944
Epoch[80] Batch [1099]	Speed: 5680.478039 samples/sec	accuracy=0.398444	lr=0.103752
Epoch[80] Batch [1149]	Speed: 5915.799931 samples/sec	accuracy=0.398289	lr=0.103561
Epoch[80] Batch [1199]	Speed: 5262.314075 samples/sec	accuracy=0.398382	lr=0.103370
Epoch[80] Batch [1249]	Speed: 5617.132565 samples/sec	accuracy=0.398313	lr=0.103178
[Epoch 80] training: accuracy=0.398324
[Epoch 80] speed: 5774 samples/sec	time cost: 231.336023
[Epoch 80] validation: err-top1=0.730170 err-top5=0.496353
Epoch[81] Batch [49]	Speed: 5482.239446 samples/sec	accuracy=0.400508	lr=0.102984
Epoch[81] Batch [99]	Speed: 5856.835029 samples/sec	accuracy=0.403496	lr=0.102793
Epoch[81] Batch [149]	Speed: 5741.270957 samples/sec	accuracy=0.401536	lr=0.102602
Epoch[81] Batch [199]	Speed: 5863.970223 samples/sec	accuracy=0.400850	lr=0.102411
Epoch[81] Batch [249]	Speed: 6160.708028 samples/sec	accuracy=0.401039	lr=0.102221
Epoch[81] Batch [299]	Speed: 6179.435092 samples/sec	accuracy=0.400540	lr=0.102030
Epoch[81] Batch [349]	Speed: 5655.326436 samples/sec	accuracy=0.401253	lr=0.101840
Epoch[81] Batch [399]	Speed: 5850.802044 samples/sec	accuracy=0.401250	lr=0.101650
Epoch[81] Batch [449]	Speed: 5821.139270 samples/sec	accuracy=0.401233	lr=0.101460
Epoch[81] Batch [499]	Speed: 6194.254595 samples/sec	accuracy=0.401061	lr=0.101270
Epoch[81] Batch [549]	Speed: 5943.586803 samples/sec	accuracy=0.400996	lr=0.101080
Epoch[81] Batch [599]	Speed: 6040.528891 samples/sec	accuracy=0.401211	lr=0.100890
Epoch[81] Batch [649]	Speed: 6013.880978 samples/sec	accuracy=0.401212	lr=0.100701
Epoch[81] Batch [699]	Speed: 5463.898843 samples/sec	accuracy=0.400989	lr=0.100511
Epoch[81] Batch [749]	Speed: 5949.763631 samples/sec	accuracy=0.400730	lr=0.100322
Epoch[81] Batch [799]	Speed: 5937.770497 samples/sec	accuracy=0.400529	lr=0.100132
Epoch[81] Batch [849]	Speed: 5968.586038 samples/sec	accuracy=0.400271	lr=0.099943
Epoch[81] Batch [899]	Speed: 5965.236332 samples/sec	accuracy=0.400033	lr=0.099754
Epoch[81] Batch [949]	Speed: 6041.404906 samples/sec	accuracy=0.399947	lr=0.099565
Epoch[81] Batch [999]	Speed: 5964.580394 samples/sec	accuracy=0.399826	lr=0.099377
Epoch[81] Batch [1049]	Speed: 5882.830728 samples/sec	accuracy=0.399794	lr=0.099188
Epoch[81] Batch [1099]	Speed: 5867.328643 samples/sec	accuracy=0.399666	lr=0.098999
Epoch[81] Batch [1149]	Speed: 5671.331062 samples/sec	accuracy=0.399708	lr=0.098811
Epoch[81] Batch [1199]	Speed: 5382.428931 samples/sec	accuracy=0.399733	lr=0.098623
Epoch[81] Batch [1249]	Speed: 5771.861216 samples/sec	accuracy=0.399592	lr=0.098434
[Epoch 81] training: accuracy=0.399612
[Epoch 81] speed: 5853 samples/sec	time cost: 228.479511
[Epoch 81] validation: err-top1=0.677979 err-top5=0.427755
Epoch[82] Batch [49]	Speed: 5835.436683 samples/sec	accuracy=0.400547	lr=0.098243
Epoch[82] Batch [99]	Speed: 5830.399825 samples/sec	accuracy=0.400791	lr=0.098055
Epoch[82] Batch [149]	Speed: 5745.445140 samples/sec	accuracy=0.401999	lr=0.097867
Epoch[82] Batch [199]	Speed: 5590.089522 samples/sec	accuracy=0.402373	lr=0.097679
Epoch[82] Batch [249]	Speed: 5801.756269 samples/sec	accuracy=0.402230	lr=0.097492
Epoch[82] Batch [299]	Speed: 6003.187173 samples/sec	accuracy=0.402624	lr=0.097304
Epoch[82] Batch [349]	Speed: 5408.286555 samples/sec	accuracy=0.402441	lr=0.097117
Epoch[82] Batch [399]	Speed: 5846.508764 samples/sec	accuracy=0.402346	lr=0.096930
Epoch[82] Batch [449]	Speed: 5795.929394 samples/sec	accuracy=0.402465	lr=0.096743
Epoch[82] Batch [499]	Speed: 5708.213843 samples/sec	accuracy=0.402453	lr=0.096556
Epoch[82] Batch [549]	Speed: 5931.132790 samples/sec	accuracy=0.401813	lr=0.096369
Epoch[82] Batch [599]	Speed: 5627.178807 samples/sec	accuracy=0.401908	lr=0.096182
Epoch[82] Batch [649]	Speed: 5801.309272 samples/sec	accuracy=0.401964	lr=0.095996
Epoch[82] Batch [699]	Speed: 5945.208069 samples/sec	accuracy=0.401772	lr=0.095809
Epoch[82] Batch [749]	Speed: 5676.148479 samples/sec	accuracy=0.401995	lr=0.095623
Epoch[82] Batch [799]	Speed: 5766.166354 samples/sec	accuracy=0.401871	lr=0.095437
Epoch[82] Batch [849]	Speed: 5361.476087 samples/sec	accuracy=0.401893	lr=0.095250
Epoch[82] Batch [899]	Speed: 6051.998002 samples/sec	accuracy=0.401772	lr=0.095064
Epoch[82] Batch [949]	Speed: 5590.388135 samples/sec	accuracy=0.401759	lr=0.094879
Epoch[82] Batch [999]	Speed: 5416.814359 samples/sec	accuracy=0.401872	lr=0.094693
Epoch[82] Batch [1049]	Speed: 6045.847178 samples/sec	accuracy=0.401813	lr=0.094507
Epoch[82] Batch [1099]	Speed: 5996.032978 samples/sec	accuracy=0.401700	lr=0.094322
Epoch[82] Batch [1149]	Speed: 5838.587077 samples/sec	accuracy=0.401601	lr=0.094137
Epoch[82] Batch [1199]	Speed: 6148.234557 samples/sec	accuracy=0.401465	lr=0.093951
Epoch[82] Batch [1249]	Speed: 5796.761401 samples/sec	accuracy=0.401526	lr=0.093766
[Epoch 82] training: accuracy=0.401512
[Epoch 82] speed: 5768 samples/sec	time cost: 231.816039
[Epoch 82] validation: err-top1=0.603695 err-top5=0.340780
Epoch[83] Batch [49]	Speed: 5218.285390 samples/sec	accuracy=0.404531	lr=0.093578
Epoch[83] Batch [99]	Speed: 5717.478755 samples/sec	accuracy=0.406191	lr=0.093393
Epoch[83] Batch [149]	Speed: 5771.645281 samples/sec	accuracy=0.405378	lr=0.093208
Epoch[83] Batch [199]	Speed: 6159.732055 samples/sec	accuracy=0.404175	lr=0.093024
Epoch[83] Batch [249]	Speed: 5879.741245 samples/sec	accuracy=0.405004	lr=0.092839
Epoch[83] Batch [299]	Speed: 6136.443563 samples/sec	accuracy=0.404655	lr=0.092655
Epoch[83] Batch [349]	Speed: 5751.601115 samples/sec	accuracy=0.405059	lr=0.092471
Epoch[83] Batch [399]	Speed: 5657.159181 samples/sec	accuracy=0.404885	lr=0.092287
Epoch[83] Batch [449]	Speed: 5880.857729 samples/sec	accuracy=0.404631	lr=0.092103
Epoch[83] Batch [499]	Speed: 6063.471371 samples/sec	accuracy=0.404508	lr=0.091919
Epoch[83] Batch [549]	Speed: 6213.196362 samples/sec	accuracy=0.404583	lr=0.091735
Epoch[83] Batch [599]	Speed: 5711.854338 samples/sec	accuracy=0.404339	lr=0.091552
Epoch[83] Batch [649]	Speed: 5372.785721 samples/sec	accuracy=0.404244	lr=0.091368
Epoch[83] Batch [699]	Speed: 5686.656481 samples/sec	accuracy=0.403993	lr=0.091185
Epoch[83] Batch [749]	Speed: 5572.567139 samples/sec	accuracy=0.403918	lr=0.091002
Epoch[83] Batch [799]	Speed: 5832.965805 samples/sec	accuracy=0.403856	lr=0.090819
Epoch[83] Batch [849]	Speed: 5860.592454 samples/sec	accuracy=0.404090	lr=0.090636
Epoch[83] Batch [899]	Speed: 5717.553497 samples/sec	accuracy=0.403919	lr=0.090453
Epoch[83] Batch [949]	Speed: 5959.647007 samples/sec	accuracy=0.403659	lr=0.090270
Epoch[83] Batch [999]	Speed: 5800.218241 samples/sec	accuracy=0.403660	lr=0.090088
Epoch[83] Batch [1049]	Speed: 5744.297115 samples/sec	accuracy=0.403661	lr=0.089905
Epoch[83] Batch [1099]	Speed: 5970.700522 samples/sec	accuracy=0.403687	lr=0.089723
Epoch[83] Batch [1149]	Speed: 5839.837102 samples/sec	accuracy=0.403522	lr=0.089541
Epoch[83] Batch [1199]	Speed: 5664.168579 samples/sec	accuracy=0.403599	lr=0.089359
Epoch[83] Batch [1249]	Speed: 5604.078746 samples/sec	accuracy=0.403532	lr=0.089177
[Epoch 83] training: accuracy=0.403523
[Epoch 83] speed: 5776 samples/sec	time cost: 231.562799
[Epoch 83] validation: err-top1=0.641562 err-top5=0.391263
Epoch[84] Batch [49]	Speed: 5777.682211 samples/sec	accuracy=0.408984	lr=0.088992
Epoch[84] Batch [99]	Speed: 6161.967901 samples/sec	accuracy=0.409111	lr=0.088810
Epoch[84] Batch [149]	Speed: 5674.372681 samples/sec	accuracy=0.408750	lr=0.088629
Epoch[84] Batch [199]	Speed: 5781.632753 samples/sec	accuracy=0.407100	lr=0.088448
Epoch[84] Batch [249]	Speed: 6169.570413 samples/sec	accuracy=0.406941	lr=0.088266
Epoch[84] Batch [299]	Speed: 5868.424863 samples/sec	accuracy=0.406950	lr=0.088085
Epoch[84] Batch [349]	Speed: 5870.720122 samples/sec	accuracy=0.406370	lr=0.087904
Epoch[84] Batch [399]	Speed: 5817.852575 samples/sec	accuracy=0.406453	lr=0.087724
Epoch[84] Batch [449]	Speed: 6072.729279 samples/sec	accuracy=0.406517	lr=0.087543
Epoch[84] Batch [499]	Speed: 5604.351066 samples/sec	accuracy=0.406432	lr=0.087363
Epoch[84] Batch [549]	Speed: 5056.091965 samples/sec	accuracy=0.405975	lr=0.087182
Epoch[84] Batch [599]	Speed: 5181.829547 samples/sec	accuracy=0.405843	lr=0.087002
Epoch[84] Batch [649]	Speed: 6163.835767 samples/sec	accuracy=0.405518	lr=0.086822
Epoch[84] Batch [699]	Speed: 5972.120035 samples/sec	accuracy=0.405791	lr=0.086642
Epoch[84] Batch [749]	Speed: 5635.978936 samples/sec	accuracy=0.405948	lr=0.086462
Epoch[84] Batch [799]	Speed: 5973.203430 samples/sec	accuracy=0.405889	lr=0.086282
Epoch[84] Batch [849]	Speed: 5261.541774 samples/sec	accuracy=0.405866	lr=0.086103
Epoch[84] Batch [899]	Speed: 5696.636423 samples/sec	accuracy=0.405839	lr=0.085923
Epoch[84] Batch [949]	Speed: 5655.127769 samples/sec	accuracy=0.405837	lr=0.085744
Epoch[84] Batch [999]	Speed: 5772.518586 samples/sec	accuracy=0.406017	lr=0.085565
Epoch[84] Batch [1049]	Speed: 5903.994322 samples/sec	accuracy=0.406113	lr=0.085386
Epoch[84] Batch [1099]	Speed: 5806.400213 samples/sec	accuracy=0.406009	lr=0.085207
Epoch[84] Batch [1149]	Speed: 5763.976859 samples/sec	accuracy=0.406050	lr=0.085028
Epoch[84] Batch [1199]	Speed: 5921.284908 samples/sec	accuracy=0.406016	lr=0.084850
Epoch[84] Batch [1249]	Speed: 5459.031387 samples/sec	accuracy=0.406045	lr=0.084671
[Epoch 84] training: accuracy=0.406034
[Epoch 84] speed: 5742 samples/sec	time cost: 232.807700
[Epoch 84] validation: err-top1=0.731645 err-top5=0.507533
Epoch[85] Batch [49]	Speed: 5861.062071 samples/sec	accuracy=0.413945	lr=0.084489
Epoch[85] Batch [99]	Speed: 5953.539604 samples/sec	accuracy=0.410908	lr=0.084311
Epoch[85] Batch [149]	Speed: 5819.946288 samples/sec	accuracy=0.411237	lr=0.084133
Epoch[85] Batch [199]	Speed: 5661.661460 samples/sec	accuracy=0.409907	lr=0.083955
Epoch[85] Batch [249]	Speed: 5782.264170 samples/sec	accuracy=0.408781	lr=0.083777
Epoch[85] Batch [299]	Speed: 6179.717830 samples/sec	accuracy=0.408600	lr=0.083600
Epoch[85] Batch [349]	Speed: 5984.270481 samples/sec	accuracy=0.408650	lr=0.083422
Epoch[85] Batch [399]	Speed: 5563.971555 samples/sec	accuracy=0.408555	lr=0.083245
Epoch[85] Batch [449]	Speed: 5695.296198 samples/sec	accuracy=0.408314	lr=0.083067
Epoch[85] Batch [499]	Speed: 5854.090017 samples/sec	accuracy=0.407834	lr=0.082890
Epoch[85] Batch [549]	Speed: 5579.293558 samples/sec	accuracy=0.407379	lr=0.082713
Epoch[85] Batch [599]	Speed: 5308.928856 samples/sec	accuracy=0.407503	lr=0.082537
Epoch[85] Batch [649]	Speed: 5872.732094 samples/sec	accuracy=0.407640	lr=0.082360
Epoch[85] Batch [699]	Speed: 5933.075908 samples/sec	accuracy=0.407402	lr=0.082183
Epoch[85] Batch [749]	Speed: 5644.287655 samples/sec	accuracy=0.407552	lr=0.082007
Epoch[85] Batch [799]	Speed: 5592.526650 samples/sec	accuracy=0.407623	lr=0.081831
Epoch[85] Batch [849]	Speed: 5724.692390 samples/sec	accuracy=0.407505	lr=0.081655
Epoch[85] Batch [899]	Speed: 6006.870648 samples/sec	accuracy=0.407497	lr=0.081479
Epoch[85] Batch [949]	Speed: 5781.983316 samples/sec	accuracy=0.407650	lr=0.081303
Epoch[85] Batch [999]	Speed: 5683.763393 samples/sec	accuracy=0.407645	lr=0.081127
Epoch[85] Batch [1049]	Speed: 5475.950965 samples/sec	accuracy=0.407593	lr=0.080952
Epoch[85] Batch [1099]	Speed: 5668.855160 samples/sec	accuracy=0.407559	lr=0.080776
Epoch[85] Batch [1149]	Speed: 5582.602342 samples/sec	accuracy=0.407683	lr=0.080601
Epoch[85] Batch [1199]	Speed: 5860.531518 samples/sec	accuracy=0.407692	lr=0.080426
Epoch[85] Batch [1249]	Speed: 6120.086350 samples/sec	accuracy=0.407512	lr=0.080251
[Epoch 85] training: accuracy=0.407521
[Epoch 85] speed: 5755 samples/sec	time cost: 232.086043
[Epoch 85] validation: err-top1=0.593830 err-top5=0.334443
Epoch[86] Batch [49]	Speed: 5597.206541 samples/sec	accuracy=0.409238	lr=0.080073
Epoch[86] Batch [99]	Speed: 6012.501641 samples/sec	accuracy=0.409922	lr=0.079898
Epoch[86] Batch [149]	Speed: 5520.286386 samples/sec	accuracy=0.410553	lr=0.079723
Epoch[86] Batch [199]	Speed: 5611.289661 samples/sec	accuracy=0.411621	lr=0.079549
Epoch[86] Batch [249]	Speed: 5883.909696 samples/sec	accuracy=0.411027	lr=0.079375
Epoch[86] Batch [299]	Speed: 6172.032121 samples/sec	accuracy=0.410436	lr=0.079201
Epoch[86] Batch [349]	Speed: 6065.401450 samples/sec	accuracy=0.410098	lr=0.079027
Epoch[86] Batch [399]	Speed: 5803.770650 samples/sec	accuracy=0.410154	lr=0.078853
Epoch[86] Batch [449]	Speed: 5768.283450 samples/sec	accuracy=0.410564	lr=0.078679
Epoch[86] Batch [499]	Speed: 5613.971757 samples/sec	accuracy=0.410166	lr=0.078506
Epoch[86] Batch [549]	Speed: 5698.858073 samples/sec	accuracy=0.410059	lr=0.078332
Epoch[86] Batch [599]	Speed: 5824.726681 samples/sec	accuracy=0.410151	lr=0.078159
Epoch[86] Batch [649]	Speed: 5686.587664 samples/sec	accuracy=0.410293	lr=0.077986
Epoch[86] Batch [699]	Speed: 5595.830156 samples/sec	accuracy=0.410198	lr=0.077813
Epoch[86] Batch [749]	Speed: 5986.548793 samples/sec	accuracy=0.410092	lr=0.077640
Epoch[86] Batch [799]	Speed: 5752.802455 samples/sec	accuracy=0.410115	lr=0.077467
Epoch[86] Batch [849]	Speed: 5500.638074 samples/sec	accuracy=0.410226	lr=0.077295
Epoch[86] Batch [899]	Speed: 5849.997958 samples/sec	accuracy=0.410135	lr=0.077123
Epoch[86] Batch [949]	Speed: 5669.370582 samples/sec	accuracy=0.410119	lr=0.076950
Epoch[86] Batch [999]	Speed: 5833.366513 samples/sec	accuracy=0.410073	lr=0.076778
Epoch[86] Batch [1049]	Speed: 5685.003529 samples/sec	accuracy=0.410059	lr=0.076606
Epoch[86] Batch [1099]	Speed: 5482.149737 samples/sec	accuracy=0.409854	lr=0.076435
Epoch[86] Batch [1149]	Speed: 5777.452161 samples/sec	accuracy=0.409696	lr=0.076263
Epoch[86] Batch [1199]	Speed: 5844.376653 samples/sec	accuracy=0.409469	lr=0.076091
Epoch[86] Batch [1249]	Speed: 5941.179655 samples/sec	accuracy=0.409529	lr=0.075920
[Epoch 86] training: accuracy=0.409540
[Epoch 86] speed: 5755 samples/sec	time cost: 232.176323
[Epoch 86] validation: err-top1=0.647321 err-top5=0.390964
Epoch[87] Batch [49]	Speed: 5675.734426 samples/sec	accuracy=0.412793	lr=0.075745
Epoch[87] Batch [99]	Speed: 5293.714261 samples/sec	accuracy=0.413066	lr=0.075574
Epoch[87] Batch [149]	Speed: 5978.066600 samples/sec	accuracy=0.413151	lr=0.075404
Epoch[87] Batch [199]	Speed: 5593.463722 samples/sec	accuracy=0.413052	lr=0.075233
Epoch[87] Batch [249]	Speed: 6135.961917 samples/sec	accuracy=0.412934	lr=0.075062
Epoch[87] Batch [299]	Speed: 5634.871722 samples/sec	accuracy=0.413421	lr=0.074892
Epoch[87] Batch [349]	Speed: 5343.432946 samples/sec	accuracy=0.412804	lr=0.074721
Epoch[87] Batch [399]	Speed: 5655.277736 samples/sec	accuracy=0.412805	lr=0.074551
Epoch[87] Batch [449]	Speed: 5709.404108 samples/sec	accuracy=0.412665	lr=0.074381
Epoch[87] Batch [499]	Speed: 5751.527482 samples/sec	accuracy=0.412574	lr=0.074211
Epoch[87] Batch [549]	Speed: 5953.864939 samples/sec	accuracy=0.412544	lr=0.074042
Epoch[87] Batch [599]	Speed: 5958.422871 samples/sec	accuracy=0.412397	lr=0.073872
Epoch[87] Batch [649]	Speed: 5903.520559 samples/sec	accuracy=0.412320	lr=0.073703
Epoch[87] Batch [699]	Speed: 6110.776974 samples/sec	accuracy=0.411899	lr=0.073534
Epoch[87] Batch [749]	Speed: 5743.884583 samples/sec	accuracy=0.411602	lr=0.073364
Epoch[87] Batch [799]	Speed: 5869.888250 samples/sec	accuracy=0.411606	lr=0.073196
Epoch[87] Batch [849]	Speed: 6105.304491 samples/sec	accuracy=0.411512	lr=0.073027
Epoch[87] Batch [899]	Speed: 5760.860083 samples/sec	accuracy=0.411725	lr=0.072858
Epoch[87] Batch [949]	Speed: 5610.314509 samples/sec	accuracy=0.411674	lr=0.072690
Epoch[87] Batch [999]	Speed: 5835.106879 samples/sec	accuracy=0.411500	lr=0.072521
Epoch[87] Batch [1049]	Speed: 5876.275341 samples/sec	accuracy=0.411535	lr=0.072353
Epoch[87] Batch [1099]	Speed: 5902.000292 samples/sec	accuracy=0.411588	lr=0.072185
Epoch[87] Batch [1149]	Speed: 5744.153298 samples/sec	accuracy=0.411588	lr=0.072017
Epoch[87] Batch [1199]	Speed: 5700.445703 samples/sec	accuracy=0.411708	lr=0.071849
Epoch[87] Batch [1249]	Speed: 5689.747429 samples/sec	accuracy=0.411681	lr=0.071682
[Epoch 87] training: accuracy=0.411689
[Epoch 87] speed: 5768 samples/sec	time cost: 231.675370
[Epoch 87] validation: err-top1=0.662659 err-top5=0.412699
Epoch[88] Batch [49]	Speed: 5951.918740 samples/sec	accuracy=0.412891	lr=0.071511
Epoch[88] Batch [99]	Speed: 5674.862564 samples/sec	accuracy=0.415410	lr=0.071344
Epoch[88] Batch [149]	Speed: 5807.914976 samples/sec	accuracy=0.415540	lr=0.071177
Epoch[88] Batch [199]	Speed: 5379.086993 samples/sec	accuracy=0.415337	lr=0.071010
Epoch[88] Batch [249]	Speed: 5607.028141 samples/sec	accuracy=0.415391	lr=0.070843
Epoch[88] Batch [299]	Speed: 5654.842898 samples/sec	accuracy=0.414860	lr=0.070676
Epoch[88] Batch [349]	Speed: 6058.728465 samples/sec	accuracy=0.414891	lr=0.070510
Epoch[88] Batch [399]	Speed: 5576.819575 samples/sec	accuracy=0.415112	lr=0.070343
Epoch[88] Batch [449]	Speed: 5534.805449 samples/sec	accuracy=0.415167	lr=0.070177
Epoch[88] Batch [499]	Speed: 5982.741180 samples/sec	accuracy=0.415102	lr=0.070011
Epoch[88] Batch [549]	Speed: 5982.219699 samples/sec	accuracy=0.415012	lr=0.069845
Epoch[88] Batch [599]	Speed: 5984.285156 samples/sec	accuracy=0.414746	lr=0.069680
Epoch[88] Batch [649]	Speed: 5836.559088 samples/sec	accuracy=0.414449	lr=0.069514
Epoch[88] Batch [699]	Speed: 6056.399164 samples/sec	accuracy=0.414495	lr=0.069349
Epoch[88] Batch [749]	Speed: 5892.390461 samples/sec	accuracy=0.414279	lr=0.069183
Epoch[88] Batch [799]	Speed: 5636.106145 samples/sec	accuracy=0.414537	lr=0.069018
Epoch[88] Batch [849]	Speed: 5793.414175 samples/sec	accuracy=0.414420	lr=0.068853
Epoch[88] Batch [899]	Speed: 5901.443813 samples/sec	accuracy=0.414225	lr=0.068689
Epoch[88] Batch [949]	Speed: 5681.414606 samples/sec	accuracy=0.414005	lr=0.068524
Epoch[88] Batch [999]	Speed: 5951.236040 samples/sec	accuracy=0.413879	lr=0.068359
Epoch[88] Batch [1049]	Speed: 5640.636299 samples/sec	accuracy=0.413774	lr=0.068195
Epoch[88] Batch [1099]	Speed: 5669.953015 samples/sec	accuracy=0.413508	lr=0.068031
Epoch[88] Batch [1149]	Speed: 6103.356917 samples/sec	accuracy=0.413499	lr=0.067867
Epoch[88] Batch [1199]	Speed: 5574.241725 samples/sec	accuracy=0.413417	lr=0.067703
Epoch[88] Batch [1249]	Speed: 5834.782978 samples/sec	accuracy=0.413379	lr=0.067539
[Epoch 88] training: accuracy=0.413371
[Epoch 88] speed: 5779 samples/sec	time cost: 231.438672
[Epoch 88] validation: err-top1=0.646883 err-top5=0.397939
Epoch[89] Batch [49]	Speed: 5552.472788 samples/sec	accuracy=0.419648	lr=0.067373
Epoch[89] Batch [99]	Speed: 5496.950642 samples/sec	accuracy=0.418574	lr=0.067209
Epoch[89] Batch [149]	Speed: 5720.654969 samples/sec	accuracy=0.418783	lr=0.067046
Epoch[89] Batch [199]	Speed: 5916.030863 samples/sec	accuracy=0.417109	lr=0.066883
Epoch[89] Batch [249]	Speed: 5981.272134 samples/sec	accuracy=0.416375	lr=0.066720
Epoch[89] Batch [299]	Speed: 5852.754446 samples/sec	accuracy=0.416126	lr=0.066557
Epoch[89] Batch [349]	Speed: 6158.093230 samples/sec	accuracy=0.416108	lr=0.066395
Epoch[89] Batch [399]	Speed: 5905.571484 samples/sec	accuracy=0.415747	lr=0.066232
Epoch[89] Batch [449]	Speed: 5816.780365 samples/sec	accuracy=0.415794	lr=0.066070
Epoch[89] Batch [499]	Speed: 5747.200645 samples/sec	accuracy=0.415625	lr=0.065908
Epoch[89] Batch [549]	Speed: 6185.672721 samples/sec	accuracy=0.415339	lr=0.065746
Epoch[89] Batch [599]	Speed: 6005.796674 samples/sec	accuracy=0.415241	lr=0.065584
Epoch[89] Batch [649]	Speed: 5856.206545 samples/sec	accuracy=0.415380	lr=0.065423
Epoch[89] Batch [699]	Speed: 5731.935113 samples/sec	accuracy=0.415205	lr=0.065261
Epoch[89] Batch [749]	Speed: 5476.454110 samples/sec	accuracy=0.415108	lr=0.065100
Epoch[89] Batch [799]	Speed: 5722.046187 samples/sec	accuracy=0.415088	lr=0.064939
Epoch[89] Batch [849]	Speed: 6229.042515 samples/sec	accuracy=0.414998	lr=0.064778
Epoch[89] Batch [899]	Speed: 5936.803806 samples/sec	accuracy=0.414969	lr=0.064617
Epoch[89] Batch [949]	Speed: 5886.878332 samples/sec	accuracy=0.415056	lr=0.064456
Epoch[89] Batch [999]	Speed: 6065.999218 samples/sec	accuracy=0.414941	lr=0.064296
Epoch[89] Batch [1049]	Speed: 5385.312847 samples/sec	accuracy=0.414947	lr=0.064135
Epoch[89] Batch [1099]	Speed: 5730.255585 samples/sec	accuracy=0.414882	lr=0.063975
Epoch[89] Batch [1149]	Speed: 6259.434287 samples/sec	accuracy=0.414974	lr=0.063815
Epoch[89] Batch [1199]	Speed: 6095.460169 samples/sec	accuracy=0.415113	lr=0.063655
Epoch[89] Batch [1249]	Speed: 6034.521401 samples/sec	accuracy=0.415111	lr=0.063496
[Epoch 89] training: accuracy=0.415076
[Epoch 89] speed: 5851 samples/sec	time cost: 228.488751
[Epoch 89] validation: err-top1=0.561563 err-top5=0.308375
Epoch[90] Batch [49]	Speed: 5048.805803 samples/sec	accuracy=0.421211	lr=0.063333
Epoch[90] Batch [99]	Speed: 5143.105388 samples/sec	accuracy=0.420088	lr=0.063174
Epoch[90] Batch [149]	Speed: 5949.581650 samples/sec	accuracy=0.420872	lr=0.063015
Epoch[90] Batch [199]	Speed: 5684.552221 samples/sec	accuracy=0.418838	lr=0.062855
Epoch[90] Batch [249]	Speed: 5783.570409 samples/sec	accuracy=0.417980	lr=0.062697
Epoch[90] Batch [299]	Speed: 5631.290863 samples/sec	accuracy=0.417448	lr=0.062538
Epoch[90] Batch [349]	Speed: 6166.004661 samples/sec	accuracy=0.417274	lr=0.062379
Epoch[90] Batch [399]	Speed: 6010.416665 samples/sec	accuracy=0.417576	lr=0.062221
Epoch[90] Batch [449]	Speed: 6197.051128 samples/sec	accuracy=0.417812	lr=0.062063
Epoch[90] Batch [499]	Speed: 5415.420370 samples/sec	accuracy=0.417822	lr=0.061905
Epoch[90] Batch [549]	Speed: 5984.501286 samples/sec	accuracy=0.417955	lr=0.061747
Epoch[90] Batch [599]	Speed: 5781.240366 samples/sec	accuracy=0.417736	lr=0.061589
Epoch[90] Batch [649]	Speed: 5937.373211 samples/sec	accuracy=0.418059	lr=0.061432
Epoch[90] Batch [699]	Speed: 6022.203741 samples/sec	accuracy=0.417958	lr=0.061274
Epoch[90] Batch [749]	Speed: 5836.101319 samples/sec	accuracy=0.417743	lr=0.061117
Epoch[90] Batch [799]	Speed: 5962.451534 samples/sec	accuracy=0.417633	lr=0.060960
Epoch[90] Batch [849]	Speed: 5918.456003 samples/sec	accuracy=0.417553	lr=0.060803
Epoch[90] Batch [899]	Speed: 6027.726992 samples/sec	accuracy=0.417636	lr=0.060646
Epoch[90] Batch [949]	Speed: 6012.255374 samples/sec	accuracy=0.417596	lr=0.060490
Epoch[90] Batch [999]	Speed: 5662.500304 samples/sec	accuracy=0.417464	lr=0.060333
Epoch[90] Batch [1049]	Speed: 5936.530714 samples/sec	accuracy=0.417404	lr=0.060177
Epoch[90] Batch [1099]	Speed: 6108.484113 samples/sec	accuracy=0.417327	lr=0.060021
Epoch[90] Batch [1149]	Speed: 6016.265326 samples/sec	accuracy=0.417379	lr=0.059865
Epoch[90] Batch [1199]	Speed: 5671.006218 samples/sec	accuracy=0.417336	lr=0.059710
Epoch[90] Batch [1249]	Speed: 5374.933696 samples/sec	accuracy=0.417312	lr=0.059554
[Epoch 90] training: accuracy=0.417303
[Epoch 90] speed: 5790 samples/sec	time cost: 230.792800
[Epoch 90] validation: err-top1=0.586994 err-top5=0.331673
Epoch[91] Batch [49]	Speed: 5836.976948 samples/sec	accuracy=0.420742	lr=0.059396
Epoch[91] Batch [99]	Speed: 5707.882182 samples/sec	accuracy=0.421924	lr=0.059240
Epoch[91] Batch [149]	Speed: 5589.260937 samples/sec	accuracy=0.421797	lr=0.059085
Epoch[91] Batch [199]	Speed: 5638.611849 samples/sec	accuracy=0.422554	lr=0.058930
Epoch[91] Batch [249]	Speed: 5657.436088 samples/sec	accuracy=0.422199	lr=0.058776
Epoch[91] Batch [299]	Speed: 6087.217316 samples/sec	accuracy=0.421452	lr=0.058621
Epoch[91] Batch [349]	Speed: 5496.628444 samples/sec	accuracy=0.421222	lr=0.058467
Epoch[91] Batch [399]	Speed: 5561.086424 samples/sec	accuracy=0.421191	lr=0.058313
Epoch[91] Batch [449]	Speed: 5474.850736 samples/sec	accuracy=0.420536	lr=0.058158
Epoch[91] Batch [499]	Speed: 5917.310197 samples/sec	accuracy=0.420551	lr=0.058005
Epoch[91] Batch [549]	Speed: 5894.314906 samples/sec	accuracy=0.420368	lr=0.057851
Epoch[91] Batch [599]	Speed: 5638.220130 samples/sec	accuracy=0.419998	lr=0.057697
Epoch[91] Batch [649]	Speed: 5663.907296 samples/sec	accuracy=0.420053	lr=0.057544
Epoch[91] Batch [699]	Speed: 5381.762585 samples/sec	accuracy=0.419944	lr=0.057391
Epoch[91] Batch [749]	Speed: 6111.895262 samples/sec	accuracy=0.419790	lr=0.057238
Epoch[91] Batch [799]	Speed: 6327.881113 samples/sec	accuracy=0.419525	lr=0.057085
Epoch[91] Batch [849]	Speed: 5803.580551 samples/sec	accuracy=0.419650	lr=0.056932
Epoch[91] Batch [899]	Speed: 5871.846509 samples/sec	accuracy=0.419331	lr=0.056780
Epoch[91] Batch [949]	Speed: 5787.691853 samples/sec	accuracy=0.419268	lr=0.056627
Epoch[91] Batch [999]	Speed: 5728.199003 samples/sec	accuracy=0.419045	lr=0.056475
Epoch[91] Batch [1049]	Speed: 5696.586858 samples/sec	accuracy=0.419011	lr=0.056323
Epoch[91] Batch [1099]	Speed: 5594.621474 samples/sec	accuracy=0.418967	lr=0.056171
Epoch[91] Batch [1149]	Speed: 5708.006892 samples/sec	accuracy=0.419059	lr=0.056020
Epoch[91] Batch [1199]	Speed: 5848.748199 samples/sec	accuracy=0.419199	lr=0.055868
Epoch[91] Batch [1249]	Speed: 5686.659794 samples/sec	accuracy=0.419148	lr=0.055717
[Epoch 91] training: accuracy=0.419161
[Epoch 91] speed: 5733 samples/sec	time cost: 232.818881
[Epoch 91] validation: err-top1=0.653121 err-top5=0.409638
Epoch[92] Batch [49]	Speed: 5619.754660 samples/sec	accuracy=0.422656	lr=0.055563
Epoch[92] Batch [99]	Speed: 5918.710469 samples/sec	accuracy=0.421279	lr=0.055412
Epoch[92] Batch [149]	Speed: 4979.070929 samples/sec	accuracy=0.421302	lr=0.055261
Epoch[92] Batch [199]	Speed: 5204.140031 samples/sec	accuracy=0.421812	lr=0.055111
Epoch[92] Batch [249]	Speed: 5950.231656 samples/sec	accuracy=0.421453	lr=0.054960
Epoch[92] Batch [299]	Speed: 5743.155385 samples/sec	accuracy=0.421680	lr=0.054810
Epoch[92] Batch [349]	Speed: 6010.392946 samples/sec	accuracy=0.421730	lr=0.054660
Epoch[92] Batch [399]	Speed: 5738.481519 samples/sec	accuracy=0.421746	lr=0.054510
Epoch[92] Batch [449]	Speed: 6078.298500 samples/sec	accuracy=0.421745	lr=0.054360
Epoch[92] Batch [499]	Speed: 5873.087849 samples/sec	accuracy=0.422111	lr=0.054210
Epoch[92] Batch [549]	Speed: 5896.798211 samples/sec	accuracy=0.421921	lr=0.054061
Epoch[92] Batch [599]	Speed: 6079.661033 samples/sec	accuracy=0.421950	lr=0.053912
Epoch[92] Batch [649]	Speed: 5727.918946 samples/sec	accuracy=0.421680	lr=0.053763
Epoch[92] Batch [699]	Speed: 6265.761605 samples/sec	accuracy=0.421786	lr=0.053614
Epoch[92] Batch [749]	Speed: 5884.589451 samples/sec	accuracy=0.421771	lr=0.053465
Epoch[92] Batch [799]	Speed: 5845.798471 samples/sec	accuracy=0.421926	lr=0.053317
Epoch[92] Batch [849]	Speed: 5905.941298 samples/sec	accuracy=0.422058	lr=0.053168
Epoch[92] Batch [899]	Speed: 5687.987068 samples/sec	accuracy=0.421877	lr=0.053020
Epoch[92] Batch [949]	Speed: 5604.644183 samples/sec	accuracy=0.421722	lr=0.052872
Epoch[92] Batch [999]	Speed: 5312.141717 samples/sec	accuracy=0.421758	lr=0.052724
Epoch[92] Batch [1049]	Speed: 6114.612683 samples/sec	accuracy=0.421760	lr=0.052577
Epoch[92] Batch [1099]	Speed: 5673.766405 samples/sec	accuracy=0.421703	lr=0.052429
Epoch[92] Batch [1149]	Speed: 5684.918951 samples/sec	accuracy=0.421580	lr=0.052282
Epoch[92] Batch [1199]	Speed: 5601.080152 samples/sec	accuracy=0.421648	lr=0.052135
Epoch[92] Batch [1249]	Speed: 5726.344837 samples/sec	accuracy=0.421695	lr=0.051988
[Epoch 92] training: accuracy=0.421703
[Epoch 92] speed: 5745 samples/sec	time cost: 232.336805
[Epoch 92] validation: err-top1=0.601762 err-top5=0.346381
Epoch[93] Batch [49]	Speed: 5419.329040 samples/sec	accuracy=0.427988	lr=0.051838
Epoch[93] Batch [99]	Speed: 5730.193354 samples/sec	accuracy=0.425928	lr=0.051691
Epoch[93] Batch [149]	Speed: 6325.866303 samples/sec	accuracy=0.424427	lr=0.051545
Epoch[93] Batch [199]	Speed: 5861.780879 samples/sec	accuracy=0.424653	lr=0.051399
Epoch[93] Batch [249]	Speed: 5689.630299 samples/sec	accuracy=0.424664	lr=0.051253
Epoch[93] Batch [299]	Speed: 5930.658757 samples/sec	accuracy=0.424176	lr=0.051107
Epoch[93] Batch [349]	Speed: 6147.440265 samples/sec	accuracy=0.424291	lr=0.050961
Epoch[93] Batch [399]	Speed: 5834.448017 samples/sec	accuracy=0.424075	lr=0.050816
Epoch[93] Batch [449]	Speed: 5779.451097 samples/sec	accuracy=0.424167	lr=0.050670
Epoch[93] Batch [499]	Speed: 5782.467823 samples/sec	accuracy=0.423852	lr=0.050525
Epoch[93] Batch [549]	Speed: 5851.833893 samples/sec	accuracy=0.423498	lr=0.050380
Epoch[93] Batch [599]	Speed: 5454.076877 samples/sec	accuracy=0.423561	lr=0.050235
Epoch[93] Batch [649]	Speed: 5898.280482 samples/sec	accuracy=0.423532	lr=0.050091
Epoch[93] Batch [699]	Speed: 5573.552936 samples/sec	accuracy=0.423775	lr=0.049946
Epoch[93] Batch [749]	Speed: 5873.192897 samples/sec	accuracy=0.423737	lr=0.049802
Epoch[93] Batch [799]	Speed: 6021.783763 samples/sec	accuracy=0.423705	lr=0.049658
Epoch[93] Batch [849]	Speed: 5266.678030 samples/sec	accuracy=0.423660	lr=0.049514
Epoch[93] Batch [899]	Speed: 5368.506593 samples/sec	accuracy=0.423727	lr=0.049370
Epoch[93] Batch [949]	Speed: 5825.769897 samples/sec	accuracy=0.423717	lr=0.049227
Epoch[93] Batch [999]	Speed: 5589.647919 samples/sec	accuracy=0.423652	lr=0.049083
Epoch[93] Batch [1049]	Speed: 5518.777512 samples/sec	accuracy=0.423462	lr=0.048940
Epoch[93] Batch [1099]	Speed: 5937.164739 samples/sec	accuracy=0.423548	lr=0.048797
Epoch[93] Batch [1149]	Speed: 5927.948687 samples/sec	accuracy=0.423520	lr=0.048654
Epoch[93] Batch [1199]	Speed: 6206.705976 samples/sec	accuracy=0.423613	lr=0.048511
Epoch[93] Batch [1249]	Speed: 5762.595642 samples/sec	accuracy=0.423532	lr=0.048369
[Epoch 93] training: accuracy=0.423549
[Epoch 93] speed: 5762 samples/sec	time cost: 231.906552
[Epoch 93] validation: err-top1=0.610819 err-top5=0.361267
Epoch[94] Batch [49]	Speed: 5759.504147 samples/sec	accuracy=0.426855	lr=0.048224
Epoch[94] Batch [99]	Speed: 5811.692446 samples/sec	accuracy=0.427070	lr=0.048082
Epoch[94] Batch [149]	Speed: 5413.296140 samples/sec	accuracy=0.427786	lr=0.047940
Epoch[94] Batch [199]	Speed: 5770.967483 samples/sec	accuracy=0.427285	lr=0.047798
Epoch[94] Batch [249]	Speed: 5488.905586 samples/sec	accuracy=0.427414	lr=0.047656
Epoch[94] Batch [299]	Speed: 6257.228168 samples/sec	accuracy=0.427191	lr=0.047515
Epoch[94] Batch [349]	Speed: 5762.245571 samples/sec	accuracy=0.427118	lr=0.047374
Epoch[94] Batch [399]	Speed: 5977.288380 samples/sec	accuracy=0.427202	lr=0.047233
Epoch[94] Batch [449]	Speed: 5907.307595 samples/sec	accuracy=0.427203	lr=0.047092
Epoch[94] Batch [499]	Speed: 5525.155841 samples/sec	accuracy=0.426891	lr=0.046951
Epoch[94] Batch [549]	Speed: 5810.407118 samples/sec	accuracy=0.426909	lr=0.046811
Epoch[94] Batch [599]	Speed: 5667.086315 samples/sec	accuracy=0.426663	lr=0.046670
Epoch[94] Batch [649]	Speed: 5629.049714 samples/sec	accuracy=0.426587	lr=0.046530
Epoch[94] Batch [699]	Speed: 6056.485934 samples/sec	accuracy=0.426539	lr=0.046390
Epoch[94] Batch [749]	Speed: 5670.297950 samples/sec	accuracy=0.426478	lr=0.046251
Epoch[94] Batch [799]	Speed: 6014.801673 samples/sec	accuracy=0.426516	lr=0.046111
Epoch[94] Batch [849]	Speed: 6244.070699 samples/sec	accuracy=0.426364	lr=0.045972
Epoch[94] Batch [899]	Speed: 5227.443344 samples/sec	accuracy=0.426233	lr=0.045833
Epoch[94] Batch [949]	Speed: 5995.680419 samples/sec	accuracy=0.426293	lr=0.045693
Epoch[94] Batch [999]	Speed: 5769.630348 samples/sec	accuracy=0.426050	lr=0.045555
Epoch[94] Batch [1049]	Speed: 5806.317007 samples/sec	accuracy=0.426057	lr=0.045416
Epoch[94] Batch [1099]	Speed: 5754.613656 samples/sec	accuracy=0.426127	lr=0.045278
Epoch[94] Batch [1149]	Speed: 5798.887404 samples/sec	accuracy=0.426098	lr=0.045139
Epoch[94] Batch [1199]	Speed: 5905.903941 samples/sec	accuracy=0.426271	lr=0.045001
Epoch[94] Batch [1249]	Speed: 5886.247579 samples/sec	accuracy=0.426079	lr=0.044863
[Epoch 94] training: accuracy=0.426076
[Epoch 94] speed: 5781 samples/sec	time cost: 231.647696
[Epoch 94] validation: err-top1=0.690390 err-top5=0.461595
Epoch[95] Batch [49]	Speed: 6053.806950 samples/sec	accuracy=0.431523	lr=0.044723
Epoch[95] Batch [99]	Speed: 5825.745401 samples/sec	accuracy=0.430459	lr=0.044585
Epoch[95] Batch [149]	Speed: 5749.327693 samples/sec	accuracy=0.429492	lr=0.044448
Epoch[95] Batch [199]	Speed: 5853.612103 samples/sec	accuracy=0.429731	lr=0.044311
Epoch[95] Batch [249]	Speed: 5742.084274 samples/sec	accuracy=0.430262	lr=0.044174
Epoch[95] Batch [299]	Speed: 5331.823697 samples/sec	accuracy=0.430120	lr=0.044037
Epoch[95] Batch [349]	Speed: 5576.519514 samples/sec	accuracy=0.429799	lr=0.043900
Epoch[95] Batch [399]	Speed: 5759.856666 samples/sec	accuracy=0.430225	lr=0.043764
Epoch[95] Batch [449]	Speed: 6056.404288 samples/sec	accuracy=0.430410	lr=0.043628
Epoch[95] Batch [499]	Speed: 5604.665100 samples/sec	accuracy=0.430465	lr=0.043492
Epoch[95] Batch [549]	Speed: 5785.965934 samples/sec	accuracy=0.430501	lr=0.043356
Epoch[95] Batch [599]	Speed: 5779.509114 samples/sec	accuracy=0.430083	lr=0.043220
Epoch[95] Batch [649]	Speed: 5602.508816 samples/sec	accuracy=0.429965	lr=0.043085
Epoch[95] Batch [699]	Speed: 5498.838022 samples/sec	accuracy=0.429806	lr=0.042949
Epoch[95] Batch [749]	Speed: 5826.729222 samples/sec	accuracy=0.429736	lr=0.042814
Epoch[95] Batch [799]	Speed: 6031.493844 samples/sec	accuracy=0.429652	lr=0.042679
Epoch[95] Batch [849]	Speed: 6065.383462 samples/sec	accuracy=0.429611	lr=0.042545
Epoch[95] Batch [899]	Speed: 5703.495615 samples/sec	accuracy=0.429744	lr=0.042410
Epoch[95] Batch [949]	Speed: 5789.092932 samples/sec	accuracy=0.429557	lr=0.042276
Epoch[95] Batch [999]	Speed: 5735.723590 samples/sec	accuracy=0.429454	lr=0.042141
Epoch[95] Batch [1049]	Speed: 6041.813347 samples/sec	accuracy=0.429153	lr=0.042007
Epoch[95] Batch [1099]	Speed: 5983.299760 samples/sec	accuracy=0.429167	lr=0.041874
Epoch[95] Batch [1149]	Speed: 6138.319671 samples/sec	accuracy=0.429174	lr=0.041740
Epoch[95] Batch [1199]	Speed: 5982.892858 samples/sec	accuracy=0.429113	lr=0.041607
Epoch[95] Batch [1249]	Speed: 5686.224634 samples/sec	accuracy=0.428966	lr=0.041473
[Epoch 95] training: accuracy=0.428957
[Epoch 95] speed: 5794 samples/sec	time cost: 230.573809
[Epoch 95] validation: err-top1=0.553472 err-top5=0.300203
Epoch[96] Batch [49]	Speed: 5312.663180 samples/sec	accuracy=0.430410	lr=0.041338
Epoch[96] Batch [99]	Speed: 5729.482302 samples/sec	accuracy=0.431572	lr=0.041205
Epoch[96] Batch [149]	Speed: 5860.270515 samples/sec	accuracy=0.431328	lr=0.041072
Epoch[96] Batch [199]	Speed: 6127.012373 samples/sec	accuracy=0.430352	lr=0.040940
Epoch[96] Batch [249]	Speed: 5722.823259 samples/sec	accuracy=0.430465	lr=0.040807
Epoch[96] Batch [299]	Speed: 5914.392891 samples/sec	accuracy=0.430537	lr=0.040675
Epoch[96] Batch [349]	Speed: 5736.896543 samples/sec	accuracy=0.431007	lr=0.040543
Epoch[96] Batch [399]	Speed: 5690.334358 samples/sec	accuracy=0.431282	lr=0.040412
Epoch[96] Batch [449]	Speed: 5865.100589 samples/sec	accuracy=0.430642	lr=0.040280
Epoch[96] Batch [499]	Speed: 6145.598502 samples/sec	accuracy=0.430311	lr=0.040149
Epoch[96] Batch [549]	Speed: 5521.300331 samples/sec	accuracy=0.430231	lr=0.040018
Epoch[96] Batch [599]	Speed: 5554.603221 samples/sec	accuracy=0.430483	lr=0.039887
Epoch[96] Batch [649]	Speed: 5946.970201 samples/sec	accuracy=0.430261	lr=0.039756
Epoch[96] Batch [699]	Speed: 5864.716971 samples/sec	accuracy=0.430297	lr=0.039625
Epoch[96] Batch [749]	Speed: 6150.590480 samples/sec	accuracy=0.430049	lr=0.039495
Epoch[96] Batch [799]	Speed: 5477.573148 samples/sec	accuracy=0.430104	lr=0.039365
Epoch[96] Batch [849]	Speed: 5837.352184 samples/sec	accuracy=0.430074	lr=0.039235
Epoch[96] Batch [899]	Speed: 6089.182227 samples/sec	accuracy=0.429940	lr=0.039105
Epoch[96] Batch [949]	Speed: 5539.409628 samples/sec	accuracy=0.429750	lr=0.038975
Epoch[96] Batch [999]	Speed: 5554.558682 samples/sec	accuracy=0.429784	lr=0.038846
Epoch[96] Batch [1049]	Speed: 5891.892209 samples/sec	accuracy=0.429971	lr=0.038717
Epoch[96] Batch [1099]	Speed: 5936.240909 samples/sec	accuracy=0.430051	lr=0.038588
Epoch[96] Batch [1149]	Speed: 5852.295250 samples/sec	accuracy=0.430037	lr=0.038459
Epoch[96] Batch [1199]	Speed: 5913.670081 samples/sec	accuracy=0.430116	lr=0.038330
Epoch[96] Batch [1249]	Speed: 5475.022979 samples/sec	accuracy=0.430080	lr=0.038202
[Epoch 96] training: accuracy=0.430079
[Epoch 96] speed: 5771 samples/sec	time cost: 231.604942
[Epoch 96] validation: err-top1=0.657246 err-top5=0.414760
Epoch[97] Batch [49]	Speed: 5457.309772 samples/sec	accuracy=0.440156	lr=0.038071
Epoch[97] Batch [99]	Speed: 5292.434945 samples/sec	accuracy=0.439033	lr=0.037943
Epoch[97] Batch [149]	Speed: 5795.818488 samples/sec	accuracy=0.436361	lr=0.037815
Epoch[97] Batch [199]	Speed: 6193.688982 samples/sec	accuracy=0.435449	lr=0.037687
Epoch[97] Batch [249]	Speed: 5804.061469 samples/sec	accuracy=0.435383	lr=0.037560
Epoch[97] Batch [299]	Speed: 5400.955525 samples/sec	accuracy=0.434626	lr=0.037432
Epoch[97] Batch [349]	Speed: 6027.682157 samples/sec	accuracy=0.433998	lr=0.037305
Epoch[97] Batch [399]	Speed: 5686.359692 samples/sec	accuracy=0.433955	lr=0.037178
Epoch[97] Batch [449]	Speed: 5843.584827 samples/sec	accuracy=0.433813	lr=0.037052
Epoch[97] Batch [499]	Speed: 5923.204913 samples/sec	accuracy=0.433166	lr=0.036925
Epoch[97] Batch [549]	Speed: 5832.374430 samples/sec	accuracy=0.433514	lr=0.036799
Epoch[97] Batch [599]	Speed: 5637.779919 samples/sec	accuracy=0.433187	lr=0.036673
Epoch[97] Batch [649]	Speed: 5744.712164 samples/sec	accuracy=0.433221	lr=0.036547
Epoch[97] Batch [699]	Speed: 5375.562290 samples/sec	accuracy=0.433326	lr=0.036421
Epoch[97] Batch [749]	Speed: 6144.515668 samples/sec	accuracy=0.433414	lr=0.036296
Epoch[97] Batch [799]	Speed: 5751.120534 samples/sec	accuracy=0.433273	lr=0.036170
Epoch[97] Batch [849]	Speed: 5970.038568 samples/sec	accuracy=0.433171	lr=0.036045
Epoch[97] Batch [899]	Speed: 5853.730018 samples/sec	accuracy=0.433030	lr=0.035920
Epoch[97] Batch [949]	Speed: 6201.838167 samples/sec	accuracy=0.432981	lr=0.035795
Epoch[97] Batch [999]	Speed: 5680.433863 samples/sec	accuracy=0.433098	lr=0.035671
Epoch[97] Batch [1049]	Speed: 5713.026971 samples/sec	accuracy=0.432911	lr=0.035546
Epoch[97] Batch [1099]	Speed: 6089.319839 samples/sec	accuracy=0.432782	lr=0.035422
Epoch[97] Batch [1149]	Speed: 5915.311888 samples/sec	accuracy=0.432835	lr=0.035298
Epoch[97] Batch [1199]	Speed: 5342.069289 samples/sec	accuracy=0.432970	lr=0.035174
Epoch[97] Batch [1249]	Speed: 5733.269526 samples/sec	accuracy=0.432803	lr=0.035051
[Epoch 97] training: accuracy=0.432827
[Epoch 97] speed: 5758 samples/sec	time cost: 232.010940
[Epoch 97] validation: err-top1=0.529775 err-top5=0.279735
Epoch[98] Batch [49]	Speed: 5409.920406 samples/sec	accuracy=0.441211	lr=0.034925
Epoch[98] Batch [99]	Speed: 5836.476443 samples/sec	accuracy=0.439482	lr=0.034802
Epoch[98] Batch [149]	Speed: 5774.391600 samples/sec	accuracy=0.438366	lr=0.034679
Epoch[98] Batch [199]	Speed: 5950.121691 samples/sec	accuracy=0.436270	lr=0.034556
Epoch[98] Batch [249]	Speed: 5658.918105 samples/sec	accuracy=0.435914	lr=0.034433
Epoch[98] Batch [299]	Speed: 5627.784311 samples/sec	accuracy=0.435794	lr=0.034311
Epoch[98] Batch [349]	Speed: 5906.434619 samples/sec	accuracy=0.435734	lr=0.034189
Epoch[98] Batch [399]	Speed: 5988.535911 samples/sec	accuracy=0.435857	lr=0.034067
Epoch[98] Batch [449]	Speed: 5461.614181 samples/sec	accuracy=0.435790	lr=0.033945
Epoch[98] Batch [499]	Speed: 5951.912471 samples/sec	accuracy=0.435422	lr=0.033823
Epoch[98] Batch [549]	Speed: 5925.188124 samples/sec	accuracy=0.435430	lr=0.033702
Epoch[98] Batch [599]	Speed: 5561.114362 samples/sec	accuracy=0.435363	lr=0.033581
Epoch[98] Batch [649]	Speed: 5731.210628 samples/sec	accuracy=0.435554	lr=0.033460
Epoch[98] Batch [699]	Speed: 5838.718834 samples/sec	accuracy=0.435417	lr=0.033339
Epoch[98] Batch [749]	Speed: 5771.932888 samples/sec	accuracy=0.435096	lr=0.033218
Epoch[98] Batch [799]	Speed: 5547.772062 samples/sec	accuracy=0.435037	lr=0.033098
Epoch[98] Batch [849]	Speed: 5551.509933 samples/sec	accuracy=0.434984	lr=0.032978
Epoch[98] Batch [899]	Speed: 5561.612539 samples/sec	accuracy=0.434906	lr=0.032858
Epoch[98] Batch [949]	Speed: 6060.373657 samples/sec	accuracy=0.435029	lr=0.032738
Epoch[98] Batch [999]	Speed: 5398.520425 samples/sec	accuracy=0.435167	lr=0.032618
Epoch[98] Batch [1049]	Speed: 5598.233473 samples/sec	accuracy=0.435120	lr=0.032499
Epoch[98] Batch [1099]	Speed: 5893.031911 samples/sec	accuracy=0.435193	lr=0.032380
Epoch[98] Batch [1149]	Speed: 5490.214990 samples/sec	accuracy=0.435053	lr=0.032261
Epoch[98] Batch [1199]	Speed: 5399.364688 samples/sec	accuracy=0.435082	lr=0.032142
Epoch[98] Batch [1249]	Speed: 5693.875382 samples/sec	accuracy=0.434995	lr=0.032023
[Epoch 98] training: accuracy=0.434997
[Epoch 98] speed: 5689 samples/sec	time cost: 234.365985
[Epoch 98] validation: err-top1=0.731934 err-top5=0.515503
Epoch[99] Batch [49]	Speed: 5259.129117 samples/sec	accuracy=0.439629	lr=0.031902
Epoch[99] Batch [99]	Speed: 6030.561089 samples/sec	accuracy=0.439385	lr=0.031784
Epoch[99] Batch [149]	Speed: 5951.492673 samples/sec	accuracy=0.440827	lr=0.031666
Epoch[99] Batch [199]	Speed: 6018.367688 samples/sec	accuracy=0.440884	lr=0.031548
Epoch[99] Batch [249]	Speed: 6153.796651 samples/sec	accuracy=0.440797	lr=0.031431
Epoch[99] Batch [299]	Speed: 5397.835164 samples/sec	accuracy=0.440114	lr=0.031313
Epoch[99] Batch [349]	Speed: 5857.470678 samples/sec	accuracy=0.440257	lr=0.031196
Epoch[99] Batch [399]	Speed: 5692.318720 samples/sec	accuracy=0.439631	lr=0.031079
Epoch[99] Batch [449]	Speed: 5655.059713 samples/sec	accuracy=0.439286	lr=0.030962
Epoch[99] Batch [499]	Speed: 5660.789290 samples/sec	accuracy=0.439201	lr=0.030846
Epoch[99] Batch [549]	Speed: 5678.220847 samples/sec	accuracy=0.438945	lr=0.030729
Epoch[99] Batch [599]	Speed: 5884.590096 samples/sec	accuracy=0.438896	lr=0.030613
Epoch[99] Batch [649]	Speed: 5768.567624 samples/sec	accuracy=0.438567	lr=0.030497
Epoch[99] Batch [699]	Speed: 5932.131228 samples/sec	accuracy=0.438302	lr=0.030381
Epoch[99] Batch [749]	Speed: 5963.248747 samples/sec	accuracy=0.438294	lr=0.030266
Epoch[99] Batch [799]	Speed: 5840.647134 samples/sec	accuracy=0.438112	lr=0.030150
Epoch[99] Batch [849]	Speed: 6007.063712 samples/sec	accuracy=0.438139	lr=0.030035
Epoch[99] Batch [899]	Speed: 5664.448115 samples/sec	accuracy=0.437817	lr=0.029920
Epoch[99] Batch [949]	Speed: 6168.933983 samples/sec	accuracy=0.437728	lr=0.029805
Epoch[99] Batch [999]	Speed: 6114.446941 samples/sec	accuracy=0.437673	lr=0.029691
Epoch[99] Batch [1049]	Speed: 5386.529780 samples/sec	accuracy=0.437688	lr=0.029576
Epoch[99] Batch [1099]	Speed: 6138.902416 samples/sec	accuracy=0.437670	lr=0.029462
Epoch[99] Batch [1149]	Speed: 5943.055514 samples/sec	accuracy=0.437592	lr=0.029348
Epoch[99] Batch [1199]	Speed: 5406.884972 samples/sec	accuracy=0.437559	lr=0.029234
Epoch[99] Batch [1249]	Speed: 5541.121961 samples/sec	accuracy=0.437242	lr=0.029121
[Epoch 99] training: accuracy=0.437242
[Epoch 99] speed: 5787 samples/sec	time cost: 231.186254
[Epoch 99] validation: err-top1=0.652762 err-top5=0.416434
Epoch[100] Batch [49]	Speed: 5348.346094 samples/sec	accuracy=0.440078	lr=0.029005
Epoch[100] Batch [99]	Speed: 5718.513904 samples/sec	accuracy=0.441025	lr=0.028892
Epoch[100] Batch [149]	Speed: 5716.513065 samples/sec	accuracy=0.441921	lr=0.028779
Epoch[100] Batch [199]	Speed: 5485.608115 samples/sec	accuracy=0.441499	lr=0.028666
Epoch[100] Batch [249]	Speed: 6079.959330 samples/sec	accuracy=0.440445	lr=0.028554
Epoch[100] Batch [299]	Speed: 5797.249796 samples/sec	accuracy=0.439987	lr=0.028441
Epoch[100] Batch [349]	Speed: 5801.334190 samples/sec	accuracy=0.440148	lr=0.028329
Epoch[100] Batch [399]	Speed: 5771.273170 samples/sec	accuracy=0.440261	lr=0.028217
Epoch[100] Batch [449]	Speed: 6002.475549 samples/sec	accuracy=0.440265	lr=0.028105
Epoch[100] Batch [499]	Speed: 5969.722582 samples/sec	accuracy=0.440180	lr=0.027994
Epoch[100] Batch [549]	Speed: 6202.916216 samples/sec	accuracy=0.440204	lr=0.027883
Epoch[100] Batch [599]	Speed: 6048.284543 samples/sec	accuracy=0.439977	lr=0.027772
Epoch[100] Batch [649]	Speed: 5611.256525 samples/sec	accuracy=0.439910	lr=0.027661
Epoch[100] Batch [699]	Speed: 5688.198748 samples/sec	accuracy=0.439771	lr=0.027550
Epoch[100] Batch [749]	Speed: 6172.233109 samples/sec	accuracy=0.439693	lr=0.027439
Epoch[100] Batch [799]	Speed: 5764.337043 samples/sec	accuracy=0.439413	lr=0.027329
Epoch[100] Batch [849]	Speed: 5643.190966 samples/sec	accuracy=0.439558	lr=0.027219
Epoch[100] Batch [899]	Speed: 5875.025420 samples/sec	accuracy=0.439684	lr=0.027109
Epoch[100] Batch [949]	Speed: 6020.958838 samples/sec	accuracy=0.439657	lr=0.026999
Epoch[100] Batch [999]	Speed: 5618.988710 samples/sec	accuracy=0.439744	lr=0.026890
Epoch[100] Batch [1049]	Speed: 5642.899882 samples/sec	accuracy=0.439819	lr=0.026781
Epoch[100] Batch [1099]	Speed: 6041.213197 samples/sec	accuracy=0.439728	lr=0.026672
Epoch[100] Batch [1149]	Speed: 5875.936242 samples/sec	accuracy=0.439571	lr=0.026563
Epoch[100] Batch [1199]	Speed: 5823.299618 samples/sec	accuracy=0.439318	lr=0.026454
Epoch[100] Batch [1249]	Speed: 5723.133629 samples/sec	accuracy=0.439276	lr=0.026346
[Epoch 100] training: accuracy=0.439279
[Epoch 100] speed: 5803 samples/sec	time cost: 230.659706
[Epoch 100] validation: err-top1=0.570213 err-top5=0.320372
Epoch[101] Batch [49]	Speed: 5635.364864 samples/sec	accuracy=0.449766	lr=0.026235
Epoch[101] Batch [99]	Speed: 5981.021255 samples/sec	accuracy=0.446504	lr=0.026127
Epoch[101] Batch [149]	Speed: 6085.380416 samples/sec	accuracy=0.446491	lr=0.026020
Epoch[101] Batch [199]	Speed: 5507.167009 samples/sec	accuracy=0.445542	lr=0.025912
Epoch[101] Batch [249]	Speed: 5918.249347 samples/sec	accuracy=0.444551	lr=0.025805
Epoch[101] Batch [299]	Speed: 5573.740415 samples/sec	accuracy=0.443919	lr=0.025697
Epoch[101] Batch [349]	Speed: 5988.608055 samples/sec	accuracy=0.443055	lr=0.025590
Epoch[101] Batch [399]	Speed: 5978.732500 samples/sec	accuracy=0.442957	lr=0.025484
Epoch[101] Batch [449]	Speed: 5715.113585 samples/sec	accuracy=0.442539	lr=0.025377
Epoch[101] Batch [499]	Speed: 5972.987450 samples/sec	accuracy=0.442244	lr=0.025271
Epoch[101] Batch [549]	Speed: 5614.406350 samples/sec	accuracy=0.441886	lr=0.025165
Epoch[101] Batch [599]	Speed: 5773.376480 samples/sec	accuracy=0.441888	lr=0.025059
Epoch[101] Batch [649]	Speed: 5708.193967 samples/sec	accuracy=0.441696	lr=0.024953
Epoch[101] Batch [699]	Speed: 5097.545808 samples/sec	accuracy=0.441797	lr=0.024847
Epoch[101] Batch [749]	Speed: 5804.188377 samples/sec	accuracy=0.441780	lr=0.024742
Epoch[101] Batch [799]	Speed: 5779.265699 samples/sec	accuracy=0.441755	lr=0.024637
Epoch[101] Batch [849]	Speed: 5774.059501 samples/sec	accuracy=0.441762	lr=0.024532
Epoch[101] Batch [899]	Speed: 5876.622841 samples/sec	accuracy=0.441710	lr=0.024427
Epoch[101] Batch [949]	Speed: 5549.079881 samples/sec	accuracy=0.441474	lr=0.024323
Epoch[101] Batch [999]	Speed: 5776.393546 samples/sec	accuracy=0.441617	lr=0.024219
Epoch[101] Batch [1049]	Speed: 5702.521771 samples/sec	accuracy=0.441575	lr=0.024115
Epoch[101] Batch [1099]	Speed: 5983.118056 samples/sec	accuracy=0.441689	lr=0.024011
Epoch[101] Batch [1149]	Speed: 5264.882220 samples/sec	accuracy=0.441632	lr=0.023907
Epoch[101] Batch [1199]	Speed: 5643.344601 samples/sec	accuracy=0.441776	lr=0.023804
Epoch[101] Batch [1249]	Speed: 5971.406790 samples/sec	accuracy=0.441691	lr=0.023700
[Epoch 101] training: accuracy=0.441683
[Epoch 101] speed: 5728 samples/sec	time cost: 233.399696
[Epoch 101] validation: err-top1=0.518156 err-top5=0.269412
Epoch[102] Batch [49]	Speed: 5863.208617 samples/sec	accuracy=0.448691	lr=0.023595
Epoch[102] Batch [99]	Speed: 5649.443973 samples/sec	accuracy=0.448115	lr=0.023493
Epoch[102] Batch [149]	Speed: 5408.216820 samples/sec	accuracy=0.449499	lr=0.023390
Epoch[102] Batch [199]	Speed: 5716.973116 samples/sec	accuracy=0.447808	lr=0.023288
Epoch[102] Batch [249]	Speed: 5797.751265 samples/sec	accuracy=0.447180	lr=0.023185
Epoch[102] Batch [299]	Speed: 5570.265848 samples/sec	accuracy=0.446582	lr=0.023083
Epoch[102] Batch [349]	Speed: 5542.007844 samples/sec	accuracy=0.445770	lr=0.022982
Epoch[102] Batch [399]	Speed: 5737.674434 samples/sec	accuracy=0.445483	lr=0.022880
Epoch[102] Batch [449]	Speed: 6205.462530 samples/sec	accuracy=0.445456	lr=0.022779
Epoch[102] Batch [499]	Speed: 5865.170430 samples/sec	accuracy=0.445574	lr=0.022678
Epoch[102] Batch [549]	Speed: 5753.481232 samples/sec	accuracy=0.445352	lr=0.022577
Epoch[102] Batch [599]	Speed: 5664.674184 samples/sec	accuracy=0.445186	lr=0.022476
Epoch[102] Batch [649]	Speed: 5705.743085 samples/sec	accuracy=0.445074	lr=0.022376
Epoch[102] Batch [699]	Speed: 5613.828228 samples/sec	accuracy=0.444923	lr=0.022276
Epoch[102] Batch [749]	Speed: 5274.967929 samples/sec	accuracy=0.444810	lr=0.022175
Epoch[102] Batch [799]	Speed: 6140.441320 samples/sec	accuracy=0.444603	lr=0.022076
Epoch[102] Batch [849]	Speed: 5700.449334 samples/sec	accuracy=0.444530	lr=0.021976
Epoch[102] Batch [899]	Speed: 5979.349101 samples/sec	accuracy=0.444625	lr=0.021877
Epoch[102] Batch [949]	Speed: 6071.364701 samples/sec	accuracy=0.444505	lr=0.021777
Epoch[102] Batch [999]	Speed: 5702.312052 samples/sec	accuracy=0.444366	lr=0.021678
Epoch[102] Batch [1049]	Speed: 5892.266456 samples/sec	accuracy=0.444454	lr=0.021580
Epoch[102] Batch [1099]	Speed: 5199.891648 samples/sec	accuracy=0.444493	lr=0.021481
Epoch[102] Batch [1149]	Speed: 5575.765304 samples/sec	accuracy=0.444479	lr=0.021383
Epoch[102] Batch [1199]	Speed: 5728.640612 samples/sec	accuracy=0.444502	lr=0.021285
Epoch[102] Batch [1249]	Speed: 5881.683853 samples/sec	accuracy=0.444529	lr=0.021187
[Epoch 102] training: accuracy=0.444524
[Epoch 102] speed: 5714 samples/sec	time cost: 233.876402
[Epoch 102] validation: err-top1=0.579261 err-top5=0.327427
Epoch[103] Batch [49]	Speed: 5719.583398 samples/sec	accuracy=0.447207	lr=0.021087
Epoch[103] Batch [99]	Speed: 5760.579295 samples/sec	accuracy=0.447383	lr=0.020990
Epoch[103] Batch [149]	Speed: 5852.185685 samples/sec	accuracy=0.448874	lr=0.020892
Epoch[103] Batch [199]	Speed: 6220.394673 samples/sec	accuracy=0.448437	lr=0.020795
Epoch[103] Batch [249]	Speed: 5973.995211 samples/sec	accuracy=0.447918	lr=0.020698
Epoch[103] Batch [299]	Speed: 6157.412556 samples/sec	accuracy=0.447402	lr=0.020602
Epoch[103] Batch [349]	Speed: 5923.402276 samples/sec	accuracy=0.447330	lr=0.020505
Epoch[103] Batch [399]	Speed: 5770.169376 samples/sec	accuracy=0.447100	lr=0.020409
Epoch[103] Batch [449]	Speed: 5866.890879 samples/sec	accuracy=0.446816	lr=0.020313
Epoch[103] Batch [499]	Speed: 5951.697369 samples/sec	accuracy=0.446813	lr=0.020217
Epoch[103] Batch [549]	Speed: 6024.107802 samples/sec	accuracy=0.446175	lr=0.020122
Epoch[103] Batch [599]	Speed: 5583.731065 samples/sec	accuracy=0.446154	lr=0.020026
Epoch[103] Batch [649]	Speed: 5342.245904 samples/sec	accuracy=0.445916	lr=0.019931
Epoch[103] Batch [699]	Speed: 5877.490403 samples/sec	accuracy=0.446009	lr=0.019836
Epoch[103] Batch [749]	Speed: 5763.738154 samples/sec	accuracy=0.445991	lr=0.019742
Epoch[103] Batch [799]	Speed: 5927.807800 samples/sec	accuracy=0.446075	lr=0.019647
Epoch[103] Batch [849]	Speed: 5453.421063 samples/sec	accuracy=0.446021	lr=0.019553
Epoch[103] Batch [899]	Speed: 5481.647783 samples/sec	accuracy=0.446055	lr=0.019459
Epoch[103] Batch [949]	Speed: 5491.835804 samples/sec	accuracy=0.445980	lr=0.019365
Epoch[103] Batch [999]	Speed: 5951.741576 samples/sec	accuracy=0.445821	lr=0.019271
Epoch[103] Batch [1049]	Speed: 5785.117854 samples/sec	accuracy=0.445878	lr=0.019178
Epoch[103] Batch [1099]	Speed: 5734.350060 samples/sec	accuracy=0.446156	lr=0.019085
Epoch[103] Batch [1149]	Speed: 5943.097619 samples/sec	accuracy=0.446155	lr=0.018992
Epoch[103] Batch [1199]	Speed: 5936.875529 samples/sec	accuracy=0.446276	lr=0.018899
Epoch[103] Batch [1249]	Speed: 5724.145193 samples/sec	accuracy=0.446127	lr=0.018806
[Epoch 103] training: accuracy=0.446126
[Epoch 103] speed: 5793 samples/sec	time cost: 230.831392
[Epoch 103] validation: err-top1=0.558275 err-top5=0.308494
Epoch[104] Batch [49]	Speed: 5512.984192 samples/sec	accuracy=0.447188	lr=0.018712
Epoch[104] Batch [99]	Speed: 5941.244581 samples/sec	accuracy=0.448076	lr=0.018620
Epoch[104] Batch [149]	Speed: 5563.343384 samples/sec	accuracy=0.448542	lr=0.018528
Epoch[104] Batch [199]	Speed: 5779.537735 samples/sec	accuracy=0.448882	lr=0.018436
Epoch[104] Batch [249]	Speed: 6117.901181 samples/sec	accuracy=0.449199	lr=0.018345
Epoch[104] Batch [299]	Speed: 5619.603482 samples/sec	accuracy=0.449167	lr=0.018254
Epoch[104] Batch [349]	Speed: 6219.586852 samples/sec	accuracy=0.449344	lr=0.018163
Epoch[104] Batch [399]	Speed: 5980.085892 samples/sec	accuracy=0.449553	lr=0.018072
Epoch[104] Batch [449]	Speed: 5974.652391 samples/sec	accuracy=0.449742	lr=0.017981
Epoch[104] Batch [499]	Speed: 5922.797159 samples/sec	accuracy=0.449637	lr=0.017891
Epoch[104] Batch [549]	Speed: 6137.265184 samples/sec	accuracy=0.449396	lr=0.017801
Epoch[104] Batch [599]	Speed: 5871.774421 samples/sec	accuracy=0.449707	lr=0.017711
Epoch[104] Batch [649]	Speed: 5911.242189 samples/sec	accuracy=0.449211	lr=0.017621
Epoch[104] Batch [699]	Speed: 6235.184676 samples/sec	accuracy=0.449208	lr=0.017532
Epoch[104] Batch [749]	Speed: 6082.720094 samples/sec	accuracy=0.449091	lr=0.017442
Epoch[104] Batch [799]	Speed: 6084.034796 samples/sec	accuracy=0.449065	lr=0.017353
Epoch[104] Batch [849]	Speed: 6179.772958 samples/sec	accuracy=0.448925	lr=0.017264
Epoch[104] Batch [899]	Speed: 5653.634787 samples/sec	accuracy=0.448886	lr=0.017176
Epoch[104] Batch [949]	Speed: 5191.737331 samples/sec	accuracy=0.448745	lr=0.017087
Epoch[104] Batch [999]	Speed: 5724.858278 samples/sec	accuracy=0.448609	lr=0.016999
Epoch[104] Batch [1049]	Speed: 6146.715321 samples/sec	accuracy=0.448487	lr=0.016911
Epoch[104] Batch [1099]	Speed: 5840.715759 samples/sec	accuracy=0.448489	lr=0.016823
Epoch[104] Batch [1149]	Speed: 5731.901149 samples/sec	accuracy=0.448553	lr=0.016736
Epoch[104] Batch [1199]	Speed: 5933.526229 samples/sec	accuracy=0.448629	lr=0.016648
Epoch[104] Batch [1249]	Speed: 5978.411431 samples/sec	accuracy=0.448594	lr=0.016561
[Epoch 104] training: accuracy=0.448599
[Epoch 104] speed: 5875 samples/sec	time cost: 227.731379
[Epoch 104] validation: err-top1=0.542847 err-top5=0.292216
Epoch[105] Batch [49]	Speed: 5612.278793 samples/sec	accuracy=0.453184	lr=0.016473
Epoch[105] Batch [99]	Speed: 6092.024476 samples/sec	accuracy=0.451895	lr=0.016386
Epoch[105] Batch [149]	Speed: 5733.254679 samples/sec	accuracy=0.450983	lr=0.016299
Epoch[105] Batch [199]	Speed: 6239.985685 samples/sec	accuracy=0.450898	lr=0.016213
Epoch[105] Batch [249]	Speed: 5633.685725 samples/sec	accuracy=0.450609	lr=0.016127
Epoch[105] Batch [299]	Speed: 5493.989665 samples/sec	accuracy=0.450833	lr=0.016041
Epoch[105] Batch [349]	Speed: 6200.954761 samples/sec	accuracy=0.450818	lr=0.015956
Epoch[105] Batch [399]	Speed: 6070.685903 samples/sec	accuracy=0.450837	lr=0.015870
Epoch[105] Batch [449]	Speed: 6003.807318 samples/sec	accuracy=0.450879	lr=0.015785
Epoch[105] Batch [499]	Speed: 5728.979582 samples/sec	accuracy=0.450805	lr=0.015700
Epoch[105] Batch [549]	Speed: 5283.991934 samples/sec	accuracy=0.451007	lr=0.015616
Epoch[105] Batch [599]	Speed: 5427.582240 samples/sec	accuracy=0.450877	lr=0.015531
Epoch[105] Batch [649]	Speed: 6138.987179 samples/sec	accuracy=0.450784	lr=0.015447
Epoch[105] Batch [699]	Speed: 5880.251935 samples/sec	accuracy=0.451138	lr=0.015363
Epoch[105] Batch [749]	Speed: 6119.324075 samples/sec	accuracy=0.451219	lr=0.015279
Epoch[105] Batch [799]	Speed: 6008.936183 samples/sec	accuracy=0.451200	lr=0.015195
Epoch[105] Batch [849]	Speed: 5792.006005 samples/sec	accuracy=0.451017	lr=0.015112
Epoch[105] Batch [899]	Speed: 6013.631734 samples/sec	accuracy=0.451234	lr=0.015029
Epoch[105] Batch [949]	Speed: 5534.653530 samples/sec	accuracy=0.451091	lr=0.014946
Epoch[105] Batch [999]	Speed: 5922.499220 samples/sec	accuracy=0.451204	lr=0.014863
Epoch[105] Batch [1049]	Speed: 5722.284197 samples/sec	accuracy=0.451302	lr=0.014781
Epoch[105] Batch [1099]	Speed: 5776.980150 samples/sec	accuracy=0.451297	lr=0.014699
Epoch[105] Batch [1149]	Speed: 6025.184616 samples/sec	accuracy=0.451067	lr=0.014616
Epoch[105] Batch [1199]	Speed: 5811.724846 samples/sec	accuracy=0.451117	lr=0.014535
Epoch[105] Batch [1249]	Speed: 5957.378710 samples/sec	accuracy=0.451145	lr=0.014453
[Epoch 105] training: accuracy=0.451154
[Epoch 105] speed: 5831 samples/sec	time cost: 229.458771
[Epoch 105] validation: err-top1=0.555265 err-top5=0.304329
Epoch[106] Batch [49]	Speed: 5437.385060 samples/sec	accuracy=0.450937	lr=0.014370
Epoch[106] Batch [99]	Speed: 5322.737831 samples/sec	accuracy=0.452139	lr=0.014289
Epoch[106] Batch [149]	Speed: 5278.034986 samples/sec	accuracy=0.452949	lr=0.014208
Epoch[106] Batch [199]	Speed: 5266.098532 samples/sec	accuracy=0.452573	lr=0.014127
Epoch[106] Batch [249]	Speed: 5641.216844 samples/sec	accuracy=0.453117	lr=0.014047
Epoch[106] Batch [299]	Speed: 5760.618082 samples/sec	accuracy=0.452855	lr=0.013966
Epoch[106] Batch [349]	Speed: 5786.812703 samples/sec	accuracy=0.452344	lr=0.013886
Epoch[106] Batch [399]	Speed: 5950.282766 samples/sec	accuracy=0.452336	lr=0.013806
Epoch[106] Batch [449]	Speed: 5735.679317 samples/sec	accuracy=0.452342	lr=0.013727
Epoch[106] Batch [499]	Speed: 5997.057072 samples/sec	accuracy=0.452375	lr=0.013647
Epoch[106] Batch [549]	Speed: 5728.452806 samples/sec	accuracy=0.452679	lr=0.013568
Epoch[106] Batch [599]	Speed: 5691.603762 samples/sec	accuracy=0.452682	lr=0.013489
Epoch[106] Batch [649]	Speed: 6125.394754 samples/sec	accuracy=0.453042	lr=0.013411
Epoch[106] Batch [699]	Speed: 6116.046415 samples/sec	accuracy=0.452937	lr=0.013332
Epoch[106] Batch [749]	Speed: 6181.282428 samples/sec	accuracy=0.452914	lr=0.013254
Epoch[106] Batch [799]	Speed: 6112.959145 samples/sec	accuracy=0.452797	lr=0.013176
Epoch[106] Batch [849]	Speed: 5931.404894 samples/sec	accuracy=0.452798	lr=0.013098
Epoch[106] Batch [899]	Speed: 5789.498091 samples/sec	accuracy=0.452713	lr=0.013020
Epoch[106] Batch [949]	Speed: 5985.531120 samples/sec	accuracy=0.452796	lr=0.012943
Epoch[106] Batch [999]	Speed: 5913.567650 samples/sec	accuracy=0.452742	lr=0.012866
Epoch[106] Batch [1049]	Speed: 5916.493759 samples/sec	accuracy=0.452920	lr=0.012789
Epoch[106] Batch [1099]	Speed: 5305.524043 samples/sec	accuracy=0.452923	lr=0.012712
Epoch[106] Batch [1149]	Speed: 5187.471823 samples/sec	accuracy=0.453049	lr=0.012636
Epoch[106] Batch [1199]	Speed: 6051.808520 samples/sec	accuracy=0.453155	lr=0.012559
Epoch[106] Batch [1249]	Speed: 6114.848603 samples/sec	accuracy=0.453277	lr=0.012483
[Epoch 106] training: accuracy=0.453277
[Epoch 106] speed: 5751 samples/sec	time cost: 232.467489
[Epoch 106] validation: err-top1=0.538684 err-top5=0.289919
Epoch[107] Batch [49]	Speed: 5739.098331 samples/sec	accuracy=0.456895	lr=0.012406
Epoch[107] Batch [99]	Speed: 5825.335626 samples/sec	accuracy=0.456689	lr=0.012330
Epoch[107] Batch [149]	Speed: 5714.314428 samples/sec	accuracy=0.456178	lr=0.012255
Epoch[107] Batch [199]	Speed: 5797.625264 samples/sec	accuracy=0.456499	lr=0.012180
Epoch[107] Batch [249]	Speed: 5663.642899 samples/sec	accuracy=0.456637	lr=0.012105
Epoch[107] Batch [299]	Speed: 5857.562546 samples/sec	accuracy=0.456491	lr=0.012030
Epoch[107] Batch [349]	Speed: 5918.709164 samples/sec	accuracy=0.456599	lr=0.011956
Epoch[107] Batch [399]	Speed: 5724.806389 samples/sec	accuracy=0.456521	lr=0.011881
Epoch[107] Batch [449]	Speed: 5937.135029 samples/sec	accuracy=0.455757	lr=0.011807
Epoch[107] Batch [499]	Speed: 6226.730294 samples/sec	accuracy=0.455707	lr=0.011734
Epoch[107] Batch [549]	Speed: 5680.755881 samples/sec	accuracy=0.455440	lr=0.011660
Epoch[107] Batch [599]	Speed: 5904.494136 samples/sec	accuracy=0.455549	lr=0.011587
Epoch[107] Batch [649]	Speed: 5891.880570 samples/sec	accuracy=0.455572	lr=0.011513
Epoch[107] Batch [699]	Speed: 5860.590535 samples/sec	accuracy=0.455618	lr=0.011441
Epoch[107] Batch [749]	Speed: 5788.486234 samples/sec	accuracy=0.455452	lr=0.011368
Epoch[107] Batch [799]	Speed: 5873.625498 samples/sec	accuracy=0.455430	lr=0.011295
Epoch[107] Batch [849]	Speed: 5842.070475 samples/sec	accuracy=0.455348	lr=0.011223
Epoch[107] Batch [899]	Speed: 5820.570957 samples/sec	accuracy=0.455356	lr=0.011151
Epoch[107] Batch [949]	Speed: 5283.851131 samples/sec	accuracy=0.455393	lr=0.011079
Epoch[107] Batch [999]	Speed: 5539.630257 samples/sec	accuracy=0.455214	lr=0.011008
Epoch[107] Batch [1049]	Speed: 5931.224854 samples/sec	accuracy=0.455251	lr=0.010936
Epoch[107] Batch [1099]	Speed: 5589.619839 samples/sec	accuracy=0.455305	lr=0.010865
Epoch[107] Batch [1149]	Speed: 5813.041913 samples/sec	accuracy=0.455215	lr=0.010794
Epoch[107] Batch [1199]	Speed: 6050.630620 samples/sec	accuracy=0.455295	lr=0.010724
Epoch[107] Batch [1249]	Speed: 6022.513316 samples/sec	accuracy=0.455209	lr=0.010653
[Epoch 107] training: accuracy=0.455197
[Epoch 107] speed: 5795 samples/sec	time cost: 230.906704
[Epoch 107] validation: err-top1=0.497250 err-top5=0.249880
Epoch[108] Batch [49]	Speed: 5361.997775 samples/sec	accuracy=0.457734	lr=0.010582
Epoch[108] Batch [99]	Speed: 5785.240818 samples/sec	accuracy=0.457598	lr=0.010512
Epoch[108] Batch [149]	Speed: 5664.003799 samples/sec	accuracy=0.458262	lr=0.010442
Epoch[108] Batch [199]	Speed: 5850.133418 samples/sec	accuracy=0.458428	lr=0.010373
Epoch[108] Batch [249]	Speed: 5960.752188 samples/sec	accuracy=0.458707	lr=0.010303
Epoch[108] Batch [299]	Speed: 5677.203234 samples/sec	accuracy=0.458118	lr=0.010234
Epoch[108] Batch [349]	Speed: 6430.893150 samples/sec	accuracy=0.457972	lr=0.010165
Epoch[108] Batch [399]	Speed: 5898.587654 samples/sec	accuracy=0.458042	lr=0.010097
Epoch[108] Batch [449]	Speed: 5799.365508 samples/sec	accuracy=0.458377	lr=0.010028
Epoch[108] Batch [499]	Speed: 5959.264814 samples/sec	accuracy=0.457955	lr=0.009960
Epoch[108] Batch [549]	Speed: 5622.687900 samples/sec	accuracy=0.457488	lr=0.009892
Epoch[108] Batch [599]	Speed: 6241.281098 samples/sec	accuracy=0.457762	lr=0.009824
Epoch[108] Batch [649]	Speed: 5914.629252 samples/sec	accuracy=0.457712	lr=0.009757
Epoch[108] Batch [699]	Speed: 5941.413559 samples/sec	accuracy=0.457610	lr=0.009690
Epoch[108] Batch [749]	Speed: 5492.245231 samples/sec	accuracy=0.457488	lr=0.009623
Epoch[108] Batch [799]	Speed: 5295.128413 samples/sec	accuracy=0.457450	lr=0.009556
Epoch[108] Batch [849]	Speed: 5998.140989 samples/sec	accuracy=0.457298	lr=0.009489
Epoch[108] Batch [899]	Speed: 6055.516746 samples/sec	accuracy=0.457355	lr=0.009423
Epoch[108] Batch [949]	Speed: 5370.280867 samples/sec	accuracy=0.457418	lr=0.009357
Epoch[108] Batch [999]	Speed: 5144.161580 samples/sec	accuracy=0.457454	lr=0.009291
Epoch[108] Batch [1049]	Speed: 6114.688593 samples/sec	accuracy=0.457519	lr=0.009225
Epoch[108] Batch [1099]	Speed: 5868.270755 samples/sec	accuracy=0.457536	lr=0.009160
Epoch[108] Batch [1149]	Speed: 5613.954439 samples/sec	accuracy=0.457521	lr=0.009095
Epoch[108] Batch [1199]	Speed: 5551.902470 samples/sec	accuracy=0.457701	lr=0.009030
Epoch[108] Batch [1249]	Speed: 6210.329567 samples/sec	accuracy=0.457732	lr=0.008965
[Epoch 108] training: accuracy=0.457729
[Epoch 108] speed: 5769 samples/sec	time cost: 231.671106
[Epoch 108] validation: err-top1=0.499621 err-top5=0.251774
Epoch[109] Batch [49]	Speed: 5664.347862 samples/sec	accuracy=0.462441	lr=0.008899
Epoch[109] Batch [99]	Speed: 5913.176364 samples/sec	accuracy=0.461064	lr=0.008835
Epoch[109] Batch [149]	Speed: 5744.501482 samples/sec	accuracy=0.461035	lr=0.008771
Epoch[109] Batch [199]	Speed: 5847.775560 samples/sec	accuracy=0.460342	lr=0.008707
Epoch[109] Batch [249]	Speed: 5773.819804 samples/sec	accuracy=0.461125	lr=0.008643
Epoch[109] Batch [299]	Speed: 5782.110662 samples/sec	accuracy=0.460915	lr=0.008580
Epoch[109] Batch [349]	Speed: 5281.956939 samples/sec	accuracy=0.460446	lr=0.008517
Epoch[109] Batch [399]	Speed: 5667.891164 samples/sec	accuracy=0.460537	lr=0.008454
Epoch[109] Batch [449]	Speed: 5425.328115 samples/sec	accuracy=0.460970	lr=0.008391
Epoch[109] Batch [499]	Speed: 5741.306414 samples/sec	accuracy=0.460828	lr=0.008329
Epoch[109] Batch [549]	Speed: 5997.976643 samples/sec	accuracy=0.460526	lr=0.008266
Epoch[109] Batch [599]	Speed: 6004.586751 samples/sec	accuracy=0.460405	lr=0.008204
Epoch[109] Batch [649]	Speed: 5845.173944 samples/sec	accuracy=0.460771	lr=0.008143
Epoch[109] Batch [699]	Speed: 5845.121919 samples/sec	accuracy=0.460794	lr=0.008081
Epoch[109] Batch [749]	Speed: 5568.373740 samples/sec	accuracy=0.460717	lr=0.008020
Epoch[109] Batch [799]	Speed: 5551.382927 samples/sec	accuracy=0.460820	lr=0.007959
Epoch[109] Batch [849]	Speed: 6015.892689 samples/sec	accuracy=0.460607	lr=0.007898
Epoch[109] Batch [899]	Speed: 5869.353852 samples/sec	accuracy=0.460591	lr=0.007837
Epoch[109] Batch [949]	Speed: 5942.646502 samples/sec	accuracy=0.460377	lr=0.007777
Epoch[109] Batch [999]	Speed: 5882.220499 samples/sec	accuracy=0.460400	lr=0.007716
Epoch[109] Batch [1049]	Speed: 5858.810641 samples/sec	accuracy=0.460461	lr=0.007656
Epoch[109] Batch [1099]	Speed: 5742.330863 samples/sec	accuracy=0.460504	lr=0.007597
Epoch[109] Batch [1149]	Speed: 5832.429555 samples/sec	accuracy=0.460415	lr=0.007537
Epoch[109] Batch [1199]	Speed: 5544.210837 samples/sec	accuracy=0.460291	lr=0.007478
Epoch[109] Batch [1249]	Speed: 5786.706668 samples/sec	accuracy=0.460322	lr=0.007419
[Epoch 109] training: accuracy=0.460320
[Epoch 109] speed: 5753 samples/sec	time cost: 232.180787
[Epoch 109] validation: err-top1=0.485352 err-top5=0.240593
Epoch[110] Batch [49]	Speed: 5889.473926 samples/sec	accuracy=0.462285	lr=0.007359
Epoch[110] Batch [99]	Speed: 5685.302736 samples/sec	accuracy=0.463213	lr=0.007300
Epoch[110] Batch [149]	Speed: 5880.413274 samples/sec	accuracy=0.462904	lr=0.007242
Epoch[110] Batch [199]	Speed: 5834.622230 samples/sec	accuracy=0.462729	lr=0.007184
Epoch[110] Batch [249]	Speed: 5638.035097 samples/sec	accuracy=0.462770	lr=0.007126
Epoch[110] Batch [299]	Speed: 5613.540900 samples/sec	accuracy=0.462731	lr=0.007068
Epoch[110] Batch [349]	Speed: 5870.258262 samples/sec	accuracy=0.462539	lr=0.007011
Epoch[110] Batch [399]	Speed: 5992.140395 samples/sec	accuracy=0.462561	lr=0.006954
Epoch[110] Batch [449]	Speed: 5683.108485 samples/sec	accuracy=0.462791	lr=0.006897
Epoch[110] Batch [499]	Speed: 5987.166173 samples/sec	accuracy=0.462813	lr=0.006840
Epoch[110] Batch [549]	Speed: 6033.996619 samples/sec	accuracy=0.462564	lr=0.006784
Epoch[110] Batch [599]	Speed: 5658.249379 samples/sec	accuracy=0.462686	lr=0.006727
Epoch[110] Batch [649]	Speed: 5828.191024 samples/sec	accuracy=0.462758	lr=0.006671
Epoch[110] Batch [699]	Speed: 5842.951234 samples/sec	accuracy=0.462648	lr=0.006615
Epoch[110] Batch [749]	Speed: 5767.319885 samples/sec	accuracy=0.462456	lr=0.006560
Epoch[110] Batch [799]	Speed: 5818.972798 samples/sec	accuracy=0.462428	lr=0.006504
Epoch[110] Batch [849]	Speed: 5546.234806 samples/sec	accuracy=0.462236	lr=0.006449
Epoch[110] Batch [899]	Speed: 5671.898467 samples/sec	accuracy=0.462201	lr=0.006394
Epoch[110] Batch [949]	Speed: 5807.406407 samples/sec	accuracy=0.462134	lr=0.006340
Epoch[110] Batch [999]	Speed: 5879.322712 samples/sec	accuracy=0.462327	lr=0.006285
Epoch[110] Batch [1049]	Speed: 6139.900062 samples/sec	accuracy=0.462184	lr=0.006231
Epoch[110] Batch [1099]	Speed: 5961.817559 samples/sec	accuracy=0.462132	lr=0.006177
Epoch[110] Batch [1149]	Speed: 5693.235348 samples/sec	accuracy=0.461855	lr=0.006123
Epoch[110] Batch [1199]	Speed: 5643.115634 samples/sec	accuracy=0.461807	lr=0.006070
Epoch[110] Batch [1249]	Speed: 5753.258809 samples/sec	accuracy=0.461883	lr=0.006017
[Epoch 110] training: accuracy=0.461873
[Epoch 110] speed: 5795 samples/sec	time cost: 230.782098
[Epoch 110] validation: err-top1=0.491414 err-top5=0.249471
Epoch[111] Batch [49]	Speed: 5610.271565 samples/sec	accuracy=0.467813	lr=0.005963
Epoch[111] Batch [99]	Speed: 5516.716415 samples/sec	accuracy=0.466562	lr=0.005910
Epoch[111] Batch [149]	Speed: 5958.496276 samples/sec	accuracy=0.466927	lr=0.005857
Epoch[111] Batch [199]	Speed: 5653.741509 samples/sec	accuracy=0.464648	lr=0.005805
Epoch[111] Batch [249]	Speed: 5977.439116 samples/sec	accuracy=0.464418	lr=0.005753
Epoch[111] Batch [299]	Speed: 5933.121806 samples/sec	accuracy=0.464561	lr=0.005701
Epoch[111] Batch [349]	Speed: 6091.502778 samples/sec	accuracy=0.464004	lr=0.005649
Epoch[111] Batch [399]	Speed: 5942.281613 samples/sec	accuracy=0.464014	lr=0.005598
Epoch[111] Batch [449]	Speed: 6129.564446 samples/sec	accuracy=0.464273	lr=0.005547
Epoch[111] Batch [499]	Speed: 5547.360907 samples/sec	accuracy=0.464137	lr=0.005496
Epoch[111] Batch [549]	Speed: 6098.385348 samples/sec	accuracy=0.464047	lr=0.005445
Epoch[111] Batch [599]	Speed: 5937.365003 samples/sec	accuracy=0.464023	lr=0.005394
Epoch[111] Batch [649]	Speed: 6022.897923 samples/sec	accuracy=0.463914	lr=0.005344
Epoch[111] Batch [699]	Speed: 5539.997677 samples/sec	accuracy=0.463590	lr=0.005294
Epoch[111] Batch [749]	Speed: 5958.822318 samples/sec	accuracy=0.463928	lr=0.005244
Epoch[111] Batch [799]	Speed: 5831.739940 samples/sec	accuracy=0.463890	lr=0.005195
Epoch[111] Batch [849]	Speed: 5946.516521 samples/sec	accuracy=0.463850	lr=0.005145
Epoch[111] Batch [899]	Speed: 5608.609102 samples/sec	accuracy=0.463662	lr=0.005096
Epoch[111] Batch [949]	Speed: 5607.940204 samples/sec	accuracy=0.463418	lr=0.005047
Epoch[111] Batch [999]	Speed: 6038.873058 samples/sec	accuracy=0.463319	lr=0.004999
Epoch[111] Batch [1049]	Speed: 5935.218943 samples/sec	accuracy=0.463371	lr=0.004950
Epoch[111] Batch [1099]	Speed: 5577.469046 samples/sec	accuracy=0.463383	lr=0.004902
Epoch[111] Batch [1149]	Speed: 5784.329694 samples/sec	accuracy=0.463542	lr=0.004854
Epoch[111] Batch [1199]	Speed: 5690.966802 samples/sec	accuracy=0.463475	lr=0.004807
Epoch[111] Batch [1249]	Speed: 5670.453065 samples/sec	accuracy=0.463495	lr=0.004759
[Epoch 111] training: accuracy=0.463483
[Epoch 111] speed: 5812 samples/sec	time cost: 230.447047
[Epoch 111] validation: err-top1=0.482820 err-top5=0.240872
Epoch[112] Batch [49]	Speed: 5265.740332 samples/sec	accuracy=0.466914	lr=0.004711
Epoch[112] Batch [99]	Speed: 5678.885892 samples/sec	accuracy=0.464629	lr=0.004664
Epoch[112] Batch [149]	Speed: 5883.643382 samples/sec	accuracy=0.464876	lr=0.004617
Epoch[112] Batch [199]	Speed: 5965.831424 samples/sec	accuracy=0.464590	lr=0.004571
Epoch[112] Batch [249]	Speed: 5786.859797 samples/sec	accuracy=0.464723	lr=0.004524
Epoch[112] Batch [299]	Speed: 5829.485815 samples/sec	accuracy=0.464326	lr=0.004478
Epoch[112] Batch [349]	Speed: 6055.945712 samples/sec	accuracy=0.464383	lr=0.004432
Epoch[112] Batch [399]	Speed: 5711.461796 samples/sec	accuracy=0.464507	lr=0.004387
Epoch[112] Batch [449]	Speed: 5762.640486 samples/sec	accuracy=0.464418	lr=0.004341
Epoch[112] Batch [499]	Speed: 6036.311432 samples/sec	accuracy=0.464699	lr=0.004296
Epoch[112] Batch [549]	Speed: 6132.210226 samples/sec	accuracy=0.464867	lr=0.004251
Epoch[112] Batch [599]	Speed: 5493.217425 samples/sec	accuracy=0.464910	lr=0.004207
Epoch[112] Batch [649]	Speed: 6251.193370 samples/sec	accuracy=0.464961	lr=0.004162
Epoch[112] Batch [699]	Speed: 5837.171779 samples/sec	accuracy=0.464703	lr=0.004118
Epoch[112] Batch [749]	Speed: 5719.850453 samples/sec	accuracy=0.464548	lr=0.004074
Epoch[112] Batch [799]	Speed: 6151.091516 samples/sec	accuracy=0.464615	lr=0.004030
Epoch[112] Batch [849]	Speed: 5791.423531 samples/sec	accuracy=0.464729	lr=0.003987
Epoch[112] Batch [899]	Speed: 5580.753041 samples/sec	accuracy=0.464608	lr=0.003944
Epoch[112] Batch [949]	Speed: 6194.995980 samples/sec	accuracy=0.464645	lr=0.003901
Epoch[112] Batch [999]	Speed: 5836.696147 samples/sec	accuracy=0.464633	lr=0.003858
Epoch[112] Batch [1049]	Speed: 5749.041102 samples/sec	accuracy=0.464595	lr=0.003815
Epoch[112] Batch [1099]	Speed: 5938.180972 samples/sec	accuracy=0.464640	lr=0.003773
Epoch[112] Batch [1149]	Speed: 5901.373105 samples/sec	accuracy=0.464426	lr=0.003731
Epoch[112] Batch [1199]	Speed: 5935.929803 samples/sec	accuracy=0.464635	lr=0.003689
Epoch[112] Batch [1249]	Speed: 5599.410283 samples/sec	accuracy=0.464615	lr=0.003647
[Epoch 112] training: accuracy=0.464610
[Epoch 112] speed: 5828 samples/sec	time cost: 229.314740
[Epoch 112] validation: err-top1=0.481565 err-top5=0.238740
Epoch[113] Batch [49]	Speed: 5836.077370 samples/sec	accuracy=0.468672	lr=0.003605
Epoch[113] Batch [99]	Speed: 5697.733728 samples/sec	accuracy=0.467998	lr=0.003564
Epoch[113] Batch [149]	Speed: 5781.377485 samples/sec	accuracy=0.469264	lr=0.003523
Epoch[113] Batch [199]	Speed: 5828.296529 samples/sec	accuracy=0.469053	lr=0.003482
Epoch[113] Batch [249]	Speed: 5872.149489 samples/sec	accuracy=0.468918	lr=0.003442
Epoch[113] Batch [299]	Speed: 5721.397213 samples/sec	accuracy=0.469089	lr=0.003402
Epoch[113] Batch [349]	Speed: 5657.246512 samples/sec	accuracy=0.468873	lr=0.003362
Epoch[113] Batch [399]	Speed: 5450.629077 samples/sec	accuracy=0.468811	lr=0.003322
Epoch[113] Batch [449]	Speed: 6045.795094 samples/sec	accuracy=0.469164	lr=0.003282
Epoch[113] Batch [499]	Speed: 5714.925447 samples/sec	accuracy=0.468408	lr=0.003243
Epoch[113] Batch [549]	Speed: 5881.631982 samples/sec	accuracy=0.468237	lr=0.003204
Epoch[113] Batch [599]	Speed: 5757.774309 samples/sec	accuracy=0.468395	lr=0.003165
Epoch[113] Batch [649]	Speed: 5721.716880 samples/sec	accuracy=0.468265	lr=0.003127
Epoch[113] Batch [699]	Speed: 5531.438937 samples/sec	accuracy=0.468165	lr=0.003088
Epoch[113] Batch [749]	Speed: 5381.746670 samples/sec	accuracy=0.467742	lr=0.003050
Epoch[113] Batch [799]	Speed: 5830.140233 samples/sec	accuracy=0.467524	lr=0.003012
Epoch[113] Batch [849]	Speed: 5553.964522 samples/sec	accuracy=0.467390	lr=0.002975
Epoch[113] Batch [899]	Speed: 6001.886544 samples/sec	accuracy=0.467263	lr=0.002937
Epoch[113] Batch [949]	Speed: 5995.304302 samples/sec	accuracy=0.467516	lr=0.002900
Epoch[113] Batch [999]	Speed: 6013.654300 samples/sec	accuracy=0.467368	lr=0.002863
Epoch[113] Batch [1049]	Speed: 5963.561896 samples/sec	accuracy=0.467336	lr=0.002826
Epoch[113] Batch [1099]	Speed: 5817.962276 samples/sec	accuracy=0.467273	lr=0.002790
Epoch[113] Batch [1149]	Speed: 5912.671010 samples/sec	accuracy=0.467271	lr=0.002754
Epoch[113] Batch [1199]	Speed: 5804.878709 samples/sec	accuracy=0.467274	lr=0.002718
Epoch[113] Batch [1249]	Speed: 5887.228378 samples/sec	accuracy=0.467195	lr=0.002682
[Epoch 113] training: accuracy=0.467201
[Epoch 113] speed: 5777 samples/sec	time cost: 231.360530
[Epoch 113] validation: err-top1=0.476303 err-top5=0.234275
Epoch[114] Batch [49]	Speed: 5072.066879 samples/sec	accuracy=0.471758	lr=0.002646
Epoch[114] Batch [99]	Speed: 5555.005248 samples/sec	accuracy=0.469814	lr=0.002611
Epoch[114] Batch [149]	Speed: 5619.305711 samples/sec	accuracy=0.469160	lr=0.002575
Epoch[114] Batch [199]	Speed: 5795.390704 samples/sec	accuracy=0.469287	lr=0.002541
Epoch[114] Batch [249]	Speed: 5546.829749 samples/sec	accuracy=0.468426	lr=0.002506
Epoch[114] Batch [299]	Speed: 6030.249162 samples/sec	accuracy=0.468096	lr=0.002472
Epoch[114] Batch [349]	Speed: 5718.220175 samples/sec	accuracy=0.468178	lr=0.002438
Epoch[114] Batch [399]	Speed: 6037.120033 samples/sec	accuracy=0.468025	lr=0.002404
Epoch[114] Batch [449]	Speed: 5962.106390 samples/sec	accuracy=0.468184	lr=0.002370
Epoch[114] Batch [499]	Speed: 6050.295988 samples/sec	accuracy=0.468150	lr=0.002337
Epoch[114] Batch [549]	Speed: 5948.607155 samples/sec	accuracy=0.467979	lr=0.002304
Epoch[114] Batch [599]	Speed: 5782.675538 samples/sec	accuracy=0.467648	lr=0.002271
Epoch[114] Batch [649]	Speed: 5891.877822 samples/sec	accuracy=0.467776	lr=0.002238
Epoch[114] Batch [699]	Speed: 5422.888384 samples/sec	accuracy=0.467345	lr=0.002205
Epoch[114] Batch [749]	Speed: 6072.593961 samples/sec	accuracy=0.467438	lr=0.002173
Epoch[114] Batch [799]	Speed: 5514.246065 samples/sec	accuracy=0.467522	lr=0.002141
Epoch[114] Batch [849]	Speed: 5703.402760 samples/sec	accuracy=0.467526	lr=0.002110
Epoch[114] Batch [899]	Speed: 5702.252395 samples/sec	accuracy=0.467575	lr=0.002078
Epoch[114] Batch [949]	Speed: 6019.851640 samples/sec	accuracy=0.467657	lr=0.002047
Epoch[114] Batch [999]	Speed: 5902.211168 samples/sec	accuracy=0.467759	lr=0.002016
Epoch[114] Batch [1049]	Speed: 6094.351519 samples/sec	accuracy=0.467754	lr=0.001985
Epoch[114] Batch [1099]	Speed: 5814.405703 samples/sec	accuracy=0.467749	lr=0.001954
Epoch[114] Batch [1149]	Speed: 5722.433172 samples/sec	accuracy=0.467825	lr=0.001924
Epoch[114] Batch [1199]	Speed: 5994.886562 samples/sec	accuracy=0.467792	lr=0.001894
Epoch[114] Batch [1249]	Speed: 5451.014534 samples/sec	accuracy=0.467785	lr=0.001864
[Epoch 114] training: accuracy=0.467779
[Epoch 114] speed: 5757 samples/sec	time cost: 232.175549
[Epoch 114] validation: err-top1=0.474530 err-top5=0.232641
Epoch[115] Batch [49]	Speed: 5039.101255 samples/sec	accuracy=0.471719	lr=0.001834
Epoch[115] Batch [99]	Speed: 5734.838255 samples/sec	accuracy=0.468672	lr=0.001804
Epoch[115] Batch [149]	Speed: 6094.352211 samples/sec	accuracy=0.468079	lr=0.001775
Epoch[115] Batch [199]	Speed: 5673.088171 samples/sec	accuracy=0.468037	lr=0.001746
Epoch[115] Batch [249]	Speed: 5770.097902 samples/sec	accuracy=0.468793	lr=0.001718
Epoch[115] Batch [299]	Speed: 5739.568006 samples/sec	accuracy=0.468174	lr=0.001689
Epoch[115] Batch [349]	Speed: 5441.603618 samples/sec	accuracy=0.468267	lr=0.001661
Epoch[115] Batch [399]	Speed: 6214.898645 samples/sec	accuracy=0.468215	lr=0.001633
Epoch[115] Batch [449]	Speed: 6191.931703 samples/sec	accuracy=0.468420	lr=0.001605
Epoch[115] Batch [499]	Speed: 5931.417345 samples/sec	accuracy=0.468943	lr=0.001578
Epoch[115] Batch [549]	Speed: 5858.057246 samples/sec	accuracy=0.468947	lr=0.001551
Epoch[115] Batch [599]	Speed: 5952.101359 samples/sec	accuracy=0.468703	lr=0.001524
Epoch[115] Batch [649]	Speed: 5954.970783 samples/sec	accuracy=0.468636	lr=0.001497
Epoch[115] Batch [699]	Speed: 5994.339034 samples/sec	accuracy=0.468648	lr=0.001470
Epoch[115] Batch [749]	Speed: 6110.211030 samples/sec	accuracy=0.468874	lr=0.001444
Epoch[115] Batch [799]	Speed: 5440.306133 samples/sec	accuracy=0.468939	lr=0.001418
Epoch[115] Batch [849]	Speed: 5568.740940 samples/sec	accuracy=0.468682	lr=0.001392
Epoch[115] Batch [899]	Speed: 5994.160173 samples/sec	accuracy=0.468559	lr=0.001366
Epoch[115] Batch [949]	Speed: 6106.629843 samples/sec	accuracy=0.468553	lr=0.001341
Epoch[115] Batch [999]	Speed: 5501.320232 samples/sec	accuracy=0.468645	lr=0.001316
Epoch[115] Batch [1049]	Speed: 5856.259566 samples/sec	accuracy=0.468652	lr=0.001291
Epoch[115] Batch [1099]	Speed: 5930.258164 samples/sec	accuracy=0.468627	lr=0.001266
Epoch[115] Batch [1149]	Speed: 5796.796921 samples/sec	accuracy=0.468509	lr=0.001242
Epoch[115] Batch [1199]	Speed: 6050.658408 samples/sec	accuracy=0.468621	lr=0.001218
Epoch[115] Batch [1249]	Speed: 6094.947050 samples/sec	accuracy=0.468717	lr=0.001194
[Epoch 115] training: accuracy=0.468716
[Epoch 115] speed: 5821 samples/sec	time cost: 229.739551
[Epoch 115] validation: err-top1=0.472856 err-top5=0.231665
Epoch[116] Batch [49]	Speed: 5245.571557 samples/sec	accuracy=0.470469	lr=0.001170
Epoch[116] Batch [99]	Speed: 5882.755793 samples/sec	accuracy=0.468613	lr=0.001146
Epoch[116] Batch [149]	Speed: 5478.501159 samples/sec	accuracy=0.468066	lr=0.001123
Epoch[116] Batch [199]	Speed: 5751.882877 samples/sec	accuracy=0.469189	lr=0.001100
Epoch[116] Batch [249]	Speed: 6053.870607 samples/sec	accuracy=0.468609	lr=0.001077
Epoch[116] Batch [299]	Speed: 6022.316556 samples/sec	accuracy=0.469053	lr=0.001055
Epoch[116] Batch [349]	Speed: 6075.176856 samples/sec	accuracy=0.468831	lr=0.001032
Epoch[116] Batch [399]	Speed: 5980.847351 samples/sec	accuracy=0.468704	lr=0.001010
Epoch[116] Batch [449]	Speed: 6134.552655 samples/sec	accuracy=0.468906	lr=0.000989
Epoch[116] Batch [499]	Speed: 5381.687193 samples/sec	accuracy=0.469211	lr=0.000967
Epoch[116] Batch [549]	Speed: 5913.047901 samples/sec	accuracy=0.469203	lr=0.000946
Epoch[116] Batch [599]	Speed: 5876.664010 samples/sec	accuracy=0.468822	lr=0.000925
Epoch[116] Batch [649]	Speed: 5660.211722 samples/sec	accuracy=0.469133	lr=0.000904
Epoch[116] Batch [699]	Speed: 5718.422538 samples/sec	accuracy=0.469676	lr=0.000883
Epoch[116] Batch [749]	Speed: 5827.468096 samples/sec	accuracy=0.470061	lr=0.000863
Epoch[116] Batch [799]	Speed: 6103.702649 samples/sec	accuracy=0.469952	lr=0.000843
Epoch[116] Batch [849]	Speed: 5625.450607 samples/sec	accuracy=0.469963	lr=0.000823
Epoch[116] Batch [899]	Speed: 5494.307478 samples/sec	accuracy=0.469966	lr=0.000803
Epoch[116] Batch [949]	Speed: 5907.797893 samples/sec	accuracy=0.469837	lr=0.000784
Epoch[116] Batch [999]	Speed: 5728.816205 samples/sec	accuracy=0.470021	lr=0.000764
Epoch[116] Batch [1049]	Speed: 5557.993250 samples/sec	accuracy=0.470058	lr=0.000745
Epoch[116] Batch [1099]	Speed: 5997.652667 samples/sec	accuracy=0.470075	lr=0.000727
Epoch[116] Batch [1149]	Speed: 5673.107504 samples/sec	accuracy=0.470042	lr=0.000708
Epoch[116] Batch [1199]	Speed: 5758.748277 samples/sec	accuracy=0.469964	lr=0.000690
Epoch[116] Batch [1249]	Speed: 6100.767862 samples/sec	accuracy=0.470026	lr=0.000672
[Epoch 116] training: accuracy=0.470022
[Epoch 116] speed: 5781 samples/sec	time cost: 231.393518
[Epoch 116] validation: err-top1=0.471456 err-top5=0.231506
Epoch[117] Batch [49]	Speed: 5399.675177 samples/sec	accuracy=0.470098	lr=0.000654
Epoch[117] Batch [99]	Speed: 5919.754663 samples/sec	accuracy=0.471064	lr=0.000636
Epoch[117] Batch [149]	Speed: 5740.383910 samples/sec	accuracy=0.471387	lr=0.000619
Epoch[117] Batch [199]	Speed: 5964.515123 samples/sec	accuracy=0.471318	lr=0.000602
Epoch[117] Batch [249]	Speed: 5403.721316 samples/sec	accuracy=0.470984	lr=0.000585
Epoch[117] Batch [299]	Speed: 5645.027723 samples/sec	accuracy=0.470853	lr=0.000569
Epoch[117] Batch [349]	Speed: 5601.175695 samples/sec	accuracy=0.471136	lr=0.000552
Epoch[117] Batch [399]	Speed: 5875.502819 samples/sec	accuracy=0.471550	lr=0.000536
Epoch[117] Batch [449]	Speed: 5856.792061 samples/sec	accuracy=0.471280	lr=0.000520
Epoch[117] Batch [499]	Speed: 6197.886913 samples/sec	accuracy=0.471270	lr=0.000505
Epoch[117] Batch [549]	Speed: 5770.402877 samples/sec	accuracy=0.470952	lr=0.000489
Epoch[117] Batch [599]	Speed: 5606.050080 samples/sec	accuracy=0.470913	lr=0.000474
Epoch[117] Batch [649]	Speed: 5805.424343 samples/sec	accuracy=0.470769	lr=0.000459
Epoch[117] Batch [699]	Speed: 5519.095646 samples/sec	accuracy=0.470639	lr=0.000445
Epoch[117] Batch [749]	Speed: 5844.319076 samples/sec	accuracy=0.470708	lr=0.000430
Epoch[117] Batch [799]	Speed: 5703.586352 samples/sec	accuracy=0.470802	lr=0.000416
Epoch[117] Batch [849]	Speed: 5718.177390 samples/sec	accuracy=0.470809	lr=0.000402
Epoch[117] Batch [899]	Speed: 5819.192289 samples/sec	accuracy=0.470894	lr=0.000388
Epoch[117] Batch [949]	Speed: 5967.087788 samples/sec	accuracy=0.470663	lr=0.000375
Epoch[117] Batch [999]	Speed: 6257.501842 samples/sec	accuracy=0.470654	lr=0.000362
Epoch[117] Batch [1049]	Speed: 5842.215263 samples/sec	accuracy=0.470525	lr=0.000349
Epoch[117] Batch [1099]	Speed: 5964.742915 samples/sec	accuracy=0.470574	lr=0.000336
Epoch[117] Batch [1149]	Speed: 5399.391025 samples/sec	accuracy=0.470810	lr=0.000323
Epoch[117] Batch [1199]	Speed: 5745.786869 samples/sec	accuracy=0.470774	lr=0.000311
Epoch[117] Batch [1249]	Speed: 5760.853438 samples/sec	accuracy=0.470880	lr=0.000299
[Epoch 117] training: accuracy=0.470864
[Epoch 117] speed: 5756 samples/sec	time cost: 232.542830
[Epoch 117] validation: err-top1=0.469906 err-top5=0.230429
Epoch[118] Batch [49]	Speed: 4908.189974 samples/sec	accuracy=0.472930	lr=0.000287
Epoch[118] Batch [99]	Speed: 5891.263613 samples/sec	accuracy=0.471680	lr=0.000275
Epoch[118] Batch [149]	Speed: 5598.426412 samples/sec	accuracy=0.470742	lr=0.000264
Epoch[118] Batch [199]	Speed: 6217.493331 samples/sec	accuracy=0.470972	lr=0.000253
Epoch[118] Batch [249]	Speed: 6216.367022 samples/sec	accuracy=0.470699	lr=0.000242
Epoch[118] Batch [299]	Speed: 5730.849525 samples/sec	accuracy=0.471455	lr=0.000231
Epoch[118] Batch [349]	Speed: 5353.422556 samples/sec	accuracy=0.471370	lr=0.000221
Epoch[118] Batch [399]	Speed: 6007.375933 samples/sec	accuracy=0.471716	lr=0.000211
Epoch[118] Batch [449]	Speed: 5405.870017 samples/sec	accuracy=0.471476	lr=0.000201
Epoch[118] Batch [499]	Speed: 5961.332485 samples/sec	accuracy=0.471373	lr=0.000191
Epoch[118] Batch [549]	Speed: 6126.657004 samples/sec	accuracy=0.471552	lr=0.000182
Epoch[118] Batch [599]	Speed: 5799.470598 samples/sec	accuracy=0.471361	lr=0.000173
Epoch[118] Batch [649]	Speed: 5938.575739 samples/sec	accuracy=0.471737	lr=0.000164
Epoch[118] Batch [699]	Speed: 5153.361234 samples/sec	accuracy=0.471627	lr=0.000155
Epoch[118] Batch [749]	Speed: 5468.940164 samples/sec	accuracy=0.471693	lr=0.000147
Epoch[118] Batch [799]	Speed: 5713.541641 samples/sec	accuracy=0.471620	lr=0.000138
Epoch[118] Batch [849]	Speed: 5703.201912 samples/sec	accuracy=0.471491	lr=0.000130
Epoch[118] Batch [899]	Speed: 5785.764529 samples/sec	accuracy=0.471327	lr=0.000123
Epoch[118] Batch [949]	Speed: 5528.787562 samples/sec	accuracy=0.471307	lr=0.000115
Epoch[118] Batch [999]	Speed: 5615.018358 samples/sec	accuracy=0.471153	lr=0.000108
Epoch[118] Batch [1049]	Speed: 5551.824102 samples/sec	accuracy=0.471232	lr=0.000101
Epoch[118] Batch [1099]	Speed: 5573.396857 samples/sec	accuracy=0.471095	lr=0.000094
Epoch[118] Batch [1149]	Speed: 5694.938247 samples/sec	accuracy=0.471123	lr=0.000087
Epoch[118] Batch [1199]	Speed: 6032.469422 samples/sec	accuracy=0.471071	lr=0.000081
Epoch[118] Batch [1249]	Speed: 5989.487283 samples/sec	accuracy=0.470971	lr=0.000075
[Epoch 118] training: accuracy=0.470940
[Epoch 118] speed: 5691 samples/sec	time cost: 234.600160
[Epoch 118] validation: err-top1=0.470623 err-top5=0.230230
Epoch[119] Batch [49]	Speed: 5425.128147 samples/sec	accuracy=0.474043	lr=0.000069
Epoch[119] Batch [99]	Speed: 5635.267412 samples/sec	accuracy=0.472295	lr=0.000063
Epoch[119] Batch [149]	Speed: 5849.706502 samples/sec	accuracy=0.470579	lr=0.000058
Epoch[119] Batch [199]	Speed: 5829.676191 samples/sec	accuracy=0.470317	lr=0.000053
Epoch[119] Batch [249]	Speed: 5968.608433 samples/sec	accuracy=0.471258	lr=0.000048
Epoch[119] Batch [299]	Speed: 5799.598090 samples/sec	accuracy=0.471335	lr=0.000043
Epoch[119] Batch [349]	Speed: 5860.510727 samples/sec	accuracy=0.471150	lr=0.000039
Epoch[119] Batch [399]	Speed: 5758.500122 samples/sec	accuracy=0.471387	lr=0.000035
Epoch[119] Batch [449]	Speed: 5506.853637 samples/sec	accuracy=0.471528	lr=0.000031
Epoch[119] Batch [499]	Speed: 6120.846548 samples/sec	accuracy=0.471719	lr=0.000027
Epoch[119] Batch [549]	Speed: 5914.509685 samples/sec	accuracy=0.471681	lr=0.000023
Epoch[119] Batch [599]	Speed: 5967.180972 samples/sec	accuracy=0.471755	lr=0.000020
Epoch[119] Batch [649]	Speed: 5866.987691 samples/sec	accuracy=0.471956	lr=0.000017
Epoch[119] Batch [699]	Speed: 5218.294393 samples/sec	accuracy=0.472052	lr=0.000015
Epoch[119] Batch [749]	Speed: 5825.789653 samples/sec	accuracy=0.471944	lr=0.000012
Epoch[119] Batch [799]	Speed: 5233.797904 samples/sec	accuracy=0.471794	lr=0.000010
Epoch[119] Batch [849]	Speed: 5314.475693 samples/sec	accuracy=0.471602	lr=0.000008
Epoch[119] Batch [899]	Speed: 5917.600276 samples/sec	accuracy=0.471670	lr=0.000006
Epoch[119] Batch [949]	Speed: 5828.161762 samples/sec	accuracy=0.471766	lr=0.000004
Epoch[119] Batch [999]	Speed: 5945.816951 samples/sec	accuracy=0.471746	lr=0.000003
Epoch[119] Batch [1049]	Speed: 5908.810925 samples/sec	accuracy=0.471796	lr=0.000002
Epoch[119] Batch [1099]	Speed: 5612.189617 samples/sec	accuracy=0.471709	lr=0.000001
Epoch[119] Batch [1149]	Speed: 5778.963053 samples/sec	accuracy=0.471773	lr=0.000000
Epoch[119] Batch [1199]	Speed: 5982.925695 samples/sec	accuracy=0.471901	lr=0.000000
Epoch[119] Batch [1249]	Speed: 5661.253846 samples/sec	accuracy=0.471822	lr=0.000000
[Epoch 119] training: accuracy=0.471829
[Epoch 119] speed: 5731 samples/sec	time cost: 233.121466
[Epoch 119] validation: err-top1=0.471122 err-top5=0.230170