Namespace(batch_norm=False, batch_size=128, data_dir='~/.mxnet/datasets/imagenet', dtype='float16', input_size=224, label_smoothing=True, last_gamma=True, log_interval=50, logging_file='resnet50_v1b_best.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='resnet50_v1b', momentum=0.9, no_wd=True, num_epochs=120, num_gpus=8, num_workers=60, rec_train='/media/ramdisk/rec/train.rec', rec_train_idx='/media/ramdisk/rec/train.idx', rec_val='/media/ramdisk/rec/val.rec', rec_val_idx='/media/ramdisk/rec/val.idx', save_dir='params_resnet50_v1b_best', 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: 1934.811369 samples/sec	accuracy=0.868184	lr=0.003133
Epoch[0] Batch [99]	Speed: 5577.737626 samples/sec	accuracy=0.911337	lr=0.006331
Epoch[0] Batch [149]	Speed: 5190.855613 samples/sec	accuracy=0.935687	lr=0.009528
Epoch[0] Batch [199]	Speed: 5013.677458 samples/sec	accuracy=0.946683	lr=0.012726
Epoch[0] Batch [249]	Speed: 4915.822729 samples/sec	accuracy=0.950088	lr=0.015923
Epoch[0] Batch [299]	Speed: 4848.192835 samples/sec	accuracy=0.948317	lr=0.019121
Epoch[0] Batch [349]	Speed: 5026.933327 samples/sec	accuracy=0.942825	lr=0.022318
Epoch[0] Batch [399]	Speed: 4709.530087 samples/sec	accuracy=0.933959	lr=0.025516
Epoch[0] Batch [449]	Speed: 4757.985413 samples/sec	accuracy=0.923108	lr=0.028713
Epoch[0] Batch [499]	Speed: 4705.314993 samples/sec	accuracy=0.911047	lr=0.031910
Epoch[0] Batch [549]	Speed: 4982.186715 samples/sec	accuracy=0.898393	lr=0.035108
Epoch[0] Batch [599]	Speed: 4770.467350 samples/sec	accuracy=0.885758	lr=0.038305
Epoch[0] Batch [649]	Speed: 5068.222088 samples/sec	accuracy=0.873119	lr=0.041503
Epoch[0] Batch [699]	Speed: 5093.326214 samples/sec	accuracy=0.860779	lr=0.044700
Epoch[0] Batch [749]	Speed: 4902.432961 samples/sec	accuracy=0.848818	lr=0.047898
Epoch[0] Batch [799]	Speed: 5356.861751 samples/sec	accuracy=0.836935	lr=0.051095
Epoch[0] Batch [849]	Speed: 4907.948128 samples/sec	accuracy=0.825381	lr=0.054293
Epoch[0] Batch [899]	Speed: 4886.975843 samples/sec	accuracy=0.814066	lr=0.057490
Epoch[0] Batch [949]	Speed: 5267.342924 samples/sec	accuracy=0.803378	lr=0.060687
Epoch[0] Batch [999]	Speed: 5003.332741 samples/sec	accuracy=0.792740	lr=0.063885
Epoch[0] Batch [1049]	Speed: 4749.703217 samples/sec	accuracy=0.782670	lr=0.067082
Epoch[0] Batch [1099]	Speed: 5073.729585 samples/sec	accuracy=0.772834	lr=0.070280
Epoch[0] Batch [1149]	Speed: 4963.930037 samples/sec	accuracy=0.763329	lr=0.073477
Epoch[0] Batch [1199]	Speed: 4855.603644 samples/sec	accuracy=0.754164	lr=0.076675
Epoch[0] Batch [1249]	Speed: 5161.518788 samples/sec	accuracy=0.745333	lr=0.079872
[Epoch 0] training: accuracy=0.744984
[Epoch 0] speed: 4681 samples/sec	time cost: 283.533315
[Epoch 0] validation: err-top1=0.948103 err-top5=0.855469
Epoch[1] Batch [49]	Speed: 4586.254049 samples/sec	accuracy=0.519356	lr=0.083133
Epoch[1] Batch [99]	Speed: 4972.369848 samples/sec	accuracy=0.515757	lr=0.086331
Epoch[1] Batch [149]	Speed: 4908.614498 samples/sec	accuracy=0.512736	lr=0.089528
Epoch[1] Batch [199]	Speed: 5012.364466 samples/sec	accuracy=0.509551	lr=0.092726
Epoch[1] Batch [249]	Speed: 4888.337787 samples/sec	accuracy=0.506632	lr=0.095923
Epoch[1] Batch [299]	Speed: 4652.211652 samples/sec	accuracy=0.503131	lr=0.099121
Epoch[1] Batch [349]	Speed: 4768.792729 samples/sec	accuracy=0.500608	lr=0.102318
Epoch[1] Batch [399]	Speed: 5083.166920 samples/sec	accuracy=0.498323	lr=0.105516
Epoch[1] Batch [449]	Speed: 5285.064778 samples/sec	accuracy=0.495668	lr=0.108713
Epoch[1] Batch [499]	Speed: 4902.457359 samples/sec	accuracy=0.493413	lr=0.111910
Epoch[1] Batch [549]	Speed: 4867.538365 samples/sec	accuracy=0.491202	lr=0.115108
Epoch[1] Batch [599]	Speed: 4822.006751 samples/sec	accuracy=0.489164	lr=0.118305
Epoch[1] Batch [649]	Speed: 4944.071701 samples/sec	accuracy=0.487216	lr=0.121503
Epoch[1] Batch [699]	Speed: 4952.937391 samples/sec	accuracy=0.485487	lr=0.124700
Epoch[1] Batch [749]	Speed: 4982.094709 samples/sec	accuracy=0.483773	lr=0.127898
Epoch[1] Batch [799]	Speed: 4950.081016 samples/sec	accuracy=0.482097	lr=0.131095
Epoch[1] Batch [849]	Speed: 5045.661287 samples/sec	accuracy=0.480550	lr=0.134293
Epoch[1] Batch [899]	Speed: 5144.606092 samples/sec	accuracy=0.479161	lr=0.137490
Epoch[1] Batch [949]	Speed: 5022.719466 samples/sec	accuracy=0.477794	lr=0.140687
Epoch[1] Batch [999]	Speed: 5007.934980 samples/sec	accuracy=0.476567	lr=0.143885
Epoch[1] Batch [1049]	Speed: 4796.425527 samples/sec	accuracy=0.475301	lr=0.147082
Epoch[1] Batch [1099]	Speed: 5046.400800 samples/sec	accuracy=0.474160	lr=0.150280
Epoch[1] Batch [1149]	Speed: 5254.257854 samples/sec	accuracy=0.473054	lr=0.153477
Epoch[1] Batch [1199]	Speed: 5105.499315 samples/sec	accuracy=0.471952	lr=0.156675
Epoch[1] Batch [1249]	Speed: 5156.040248 samples/sec	accuracy=0.471048	lr=0.159872
[Epoch 1] training: accuracy=0.471029
[Epoch 1] speed: 4952 samples/sec	time cost: 268.207259
[Epoch 1] validation: err-top1=0.867825 err-top5=0.690171
Epoch[2] Batch [49]	Speed: 4688.139831 samples/sec	accuracy=0.446234	lr=0.163133
Epoch[2] Batch [99]	Speed: 4990.884212 samples/sec	accuracy=0.447271	lr=0.166331
Epoch[2] Batch [149]	Speed: 5077.617076 samples/sec	accuracy=0.447359	lr=0.169528
Epoch[2] Batch [199]	Speed: 4831.422126 samples/sec	accuracy=0.447068	lr=0.172726
Epoch[2] Batch [249]	Speed: 5011.399470 samples/sec	accuracy=0.447057	lr=0.175923
Epoch[2] Batch [299]	Speed: 4967.866503 samples/sec	accuracy=0.447364	lr=0.179121
Epoch[2] Batch [349]	Speed: 5027.073009 samples/sec	accuracy=0.447735	lr=0.182318
Epoch[2] Batch [399]	Speed: 5113.665683 samples/sec	accuracy=0.448008	lr=0.185516
Epoch[2] Batch [449]	Speed: 4925.003828 samples/sec	accuracy=0.448505	lr=0.188713
Epoch[2] Batch [499]	Speed: 5186.698282 samples/sec	accuracy=0.448863	lr=0.191910
Epoch[2] Batch [549]	Speed: 5206.510819 samples/sec	accuracy=0.449400	lr=0.195108
Epoch[2] Batch [599]	Speed: 5090.356828 samples/sec	accuracy=0.449776	lr=0.198305
Epoch[2] Batch [649]	Speed: 5053.385805 samples/sec	accuracy=0.450102	lr=0.201503
Epoch[2] Batch [699]	Speed: 5165.865099 samples/sec	accuracy=0.450543	lr=0.204700
Epoch[2] Batch [749]	Speed: 4867.734427 samples/sec	accuracy=0.451040	lr=0.207898
Epoch[2] Batch [799]	Speed: 5330.216826 samples/sec	accuracy=0.451634	lr=0.211095
Epoch[2] Batch [849]	Speed: 5021.527371 samples/sec	accuracy=0.452189	lr=0.214293
Epoch[2] Batch [899]	Speed: 4879.895036 samples/sec	accuracy=0.452657	lr=0.217490
Epoch[2] Batch [949]	Speed: 5232.384827 samples/sec	accuracy=0.453111	lr=0.220687
Epoch[2] Batch [999]	Speed: 5151.673000 samples/sec	accuracy=0.453676	lr=0.223885
Epoch[2] Batch [1049]	Speed: 5012.004275 samples/sec	accuracy=0.454244	lr=0.227082
Epoch[2] Batch [1099]	Speed: 5036.724269 samples/sec	accuracy=0.454862	lr=0.230280
Epoch[2] Batch [1149]	Speed: 5264.987291 samples/sec	accuracy=0.455425	lr=0.233477
Epoch[2] Batch [1199]	Speed: 5141.418573 samples/sec	accuracy=0.456006	lr=0.236675
Epoch[2] Batch [1249]	Speed: 5081.571132 samples/sec	accuracy=0.456646	lr=0.239872
[Epoch 2] training: accuracy=0.456661
[Epoch 2] speed: 5046 samples/sec	time cost: 263.426191
[Epoch 2] validation: err-top1=0.794284 err-top5=0.574717
Epoch[3] Batch [49]	Speed: 5080.530867 samples/sec	accuracy=0.472910	lr=0.243133
Epoch[3] Batch [99]	Speed: 5055.196452 samples/sec	accuracy=0.474220	lr=0.246331
Epoch[3] Batch [149]	Speed: 5240.323040 samples/sec	accuracy=0.473915	lr=0.249528
Epoch[3] Batch [199]	Speed: 4895.707523 samples/sec	accuracy=0.474701	lr=0.252726
Epoch[3] Batch [249]	Speed: 5174.045489 samples/sec	accuracy=0.475113	lr=0.255923
Epoch[3] Batch [299]	Speed: 5091.603078 samples/sec	accuracy=0.476051	lr=0.259121
Epoch[3] Batch [349]	Speed: 5064.901871 samples/sec	accuracy=0.476868	lr=0.262318
Epoch[3] Batch [399]	Speed: 5002.021316 samples/sec	accuracy=0.477324	lr=0.265516
Epoch[3] Batch [449]	Speed: 5060.472096 samples/sec	accuracy=0.478200	lr=0.268713
Epoch[3] Batch [499]	Speed: 4847.028857 samples/sec	accuracy=0.478942	lr=0.271910
Epoch[3] Batch [549]	Speed: 5062.915600 samples/sec	accuracy=0.479853	lr=0.275108
Epoch[3] Batch [599]	Speed: 4994.310159 samples/sec	accuracy=0.480666	lr=0.278305
Epoch[3] Batch [649]	Speed: 5143.268599 samples/sec	accuracy=0.481404	lr=0.281503
Epoch[3] Batch [699]	Speed: 5245.655997 samples/sec	accuracy=0.482215	lr=0.284700
Epoch[3] Batch [749]	Speed: 4982.819983 samples/sec	accuracy=0.482936	lr=0.287898
Epoch[3] Batch [799]	Speed: 5016.385539 samples/sec	accuracy=0.483767	lr=0.291095
Epoch[3] Batch [849]	Speed: 5078.387724 samples/sec	accuracy=0.484536	lr=0.294293
Epoch[3] Batch [899]	Speed: 5269.108868 samples/sec	accuracy=0.485264	lr=0.297490
Epoch[3] Batch [949]	Speed: 5150.424597 samples/sec	accuracy=0.486071	lr=0.300687
Epoch[3] Batch [999]	Speed: 5049.453034 samples/sec	accuracy=0.486853	lr=0.303885
Epoch[3] Batch [1049]	Speed: 5235.833866 samples/sec	accuracy=0.487765	lr=0.307082
Epoch[3] Batch [1099]	Speed: 5274.004092 samples/sec	accuracy=0.488586	lr=0.310280
Epoch[3] Batch [1149]	Speed: 5408.481879 samples/sec	accuracy=0.489365	lr=0.313477
Epoch[3] Batch [1199]	Speed: 4957.662848 samples/sec	accuracy=0.490183	lr=0.316675
Epoch[3] Batch [1249]	Speed: 5444.093896 samples/sec	accuracy=0.491000	lr=0.319872
[Epoch 3] training: accuracy=0.491015
[Epoch 3] speed: 5104 samples/sec	time cost: 260.152247
[Epoch 3] validation: err-top1=0.810945 err-top5=0.617546
Epoch[4] Batch [49]	Speed: 4901.111026 samples/sec	accuracy=0.509816	lr=0.323133
Epoch[4] Batch [99]	Speed: 5103.935815 samples/sec	accuracy=0.511046	lr=0.326331
Epoch[4] Batch [149]	Speed: 5307.518343 samples/sec	accuracy=0.512235	lr=0.329528
Epoch[4] Batch [199]	Speed: 5073.565243 samples/sec	accuracy=0.513098	lr=0.332726
Epoch[4] Batch [249]	Speed: 5052.364535 samples/sec	accuracy=0.514353	lr=0.335923
Epoch[4] Batch [299]	Speed: 4906.778379 samples/sec	accuracy=0.515116	lr=0.339121
Epoch[4] Batch [349]	Speed: 5197.759614 samples/sec	accuracy=0.515941	lr=0.342318
Epoch[4] Batch [399]	Speed: 5148.736065 samples/sec	accuracy=0.516779	lr=0.345516
Epoch[4] Batch [449]	Speed: 4982.217346 samples/sec	accuracy=0.517553	lr=0.348713
Epoch[4] Batch [499]	Speed: 5080.462837 samples/sec	accuracy=0.518247	lr=0.351910
Epoch[4] Batch [549]	Speed: 5264.087616 samples/sec	accuracy=0.519014	lr=0.355108
Epoch[4] Batch [599]	Speed: 5063.772416 samples/sec	accuracy=0.519854	lr=0.358305
Epoch[4] Batch [649]	Speed: 4959.947451 samples/sec	accuracy=0.520663	lr=0.361503
Epoch[4] Batch [699]	Speed: 5176.030979 samples/sec	accuracy=0.521379	lr=0.364700
Epoch[4] Batch [749]	Speed: 5091.678529 samples/sec	accuracy=0.522187	lr=0.367898
Epoch[4] Batch [799]	Speed: 4935.556308 samples/sec	accuracy=0.522936	lr=0.371095
Epoch[4] Batch [849]	Speed: 4922.089205 samples/sec	accuracy=0.523675	lr=0.374293
Epoch[4] Batch [899]	Speed: 5086.995024 samples/sec	accuracy=0.524551	lr=0.377490
Epoch[4] Batch [949]	Speed: 5047.124634 samples/sec	accuracy=0.525290	lr=0.380687
Epoch[4] Batch [999]	Speed: 5059.412558 samples/sec	accuracy=0.526034	lr=0.383885
Epoch[4] Batch [1049]	Speed: 4913.510231 samples/sec	accuracy=0.526847	lr=0.387082
Epoch[4] Batch [1099]	Speed: 4981.095232 samples/sec	accuracy=0.527557	lr=0.390280
Epoch[4] Batch [1149]	Speed: 5183.313778 samples/sec	accuracy=0.528390	lr=0.393477
Epoch[4] Batch [1199]	Speed: 5168.360345 samples/sec	accuracy=0.529148	lr=0.396675
Epoch[4] Batch [1249]	Speed: 5225.580725 samples/sec	accuracy=0.529923	lr=0.399872
[Epoch 4] training: accuracy=0.529940
[Epoch 4] speed: 5066 samples/sec	time cost: 262.136434
[Epoch 4] validation: err-top1=0.600765 err-top5=0.335918
Epoch[5] Batch [49]	Speed: 4965.324655 samples/sec	accuracy=0.548464	lr=0.400000
Epoch[5] Batch [99]	Speed: 5046.255062 samples/sec	accuracy=0.549148	lr=0.400000
Epoch[5] Batch [149]	Speed: 4882.178768 samples/sec	accuracy=0.549704	lr=0.399999
Epoch[5] Batch [199]	Speed: 4989.323111 samples/sec	accuracy=0.550055	lr=0.399998
Epoch[5] Batch [249]	Speed: 4951.513301 samples/sec	accuracy=0.551015	lr=0.399997
Epoch[5] Batch [299]	Speed: 5118.367220 samples/sec	accuracy=0.551770	lr=0.399996
Epoch[5] Batch [349]	Speed: 5206.416148 samples/sec	accuracy=0.552837	lr=0.399994
Epoch[5] Batch [399]	Speed: 4975.512641 samples/sec	accuracy=0.553720	lr=0.399992
Epoch[5] Batch [449]	Speed: 5212.514811 samples/sec	accuracy=0.554372	lr=0.399990
Epoch[5] Batch [499]	Speed: 5487.638315 samples/sec	accuracy=0.555097	lr=0.399988
Epoch[5] Batch [549]	Speed: 5224.925442 samples/sec	accuracy=0.555861	lr=0.399986
Epoch[5] Batch [599]	Speed: 5325.625079 samples/sec	accuracy=0.556567	lr=0.399983
Epoch[5] Batch [649]	Speed: 5126.710811 samples/sec	accuracy=0.557218	lr=0.399980
Epoch[5] Batch [699]	Speed: 5059.369051 samples/sec	accuracy=0.557961	lr=0.399977
Epoch[5] Batch [749]	Speed: 5174.833591 samples/sec	accuracy=0.558574	lr=0.399973
Epoch[5] Batch [799]	Speed: 4957.151643 samples/sec	accuracy=0.559288	lr=0.399970
Epoch[5] Batch [849]	Speed: 4896.508228 samples/sec	accuracy=0.560082	lr=0.399966
Epoch[5] Batch [899]	Speed: 5183.454278 samples/sec	accuracy=0.560776	lr=0.399961
Epoch[5] Batch [949]	Speed: 5020.555322 samples/sec	accuracy=0.561501	lr=0.399957
Epoch[5] Batch [999]	Speed: 5083.169928 samples/sec	accuracy=0.562183	lr=0.399952
Epoch[5] Batch [1049]	Speed: 5123.408151 samples/sec	accuracy=0.562816	lr=0.399948
Epoch[5] Batch [1099]	Speed: 4959.177284 samples/sec	accuracy=0.563530	lr=0.399942
Epoch[5] Batch [1149]	Speed: 5039.488058 samples/sec	accuracy=0.564233	lr=0.399937
Epoch[5] Batch [1199]	Speed: 4926.595232 samples/sec	accuracy=0.564883	lr=0.399931
Epoch[5] Batch [1249]	Speed: 5166.223759 samples/sec	accuracy=0.565576	lr=0.399926
[Epoch 5] training: accuracy=0.565587
[Epoch 5] speed: 5076 samples/sec	time cost: 261.809621
[Epoch 5] validation: err-top1=0.553121 err-top5=0.286947
Epoch[6] Batch [49]	Speed: 4875.106423 samples/sec	accuracy=0.580198	lr=0.399919
Epoch[6] Batch [99]	Speed: 5045.949620 samples/sec	accuracy=0.581574	lr=0.399913
Epoch[6] Batch [149]	Speed: 4816.033623 samples/sec	accuracy=0.582141	lr=0.399907
Epoch[6] Batch [199]	Speed: 5025.971182 samples/sec	accuracy=0.582037	lr=0.399900
Epoch[6] Batch [249]	Speed: 5088.552260 samples/sec	accuracy=0.582583	lr=0.399893
Epoch[6] Batch [299]	Speed: 5006.924642 samples/sec	accuracy=0.583276	lr=0.399885
Epoch[6] Batch [349]	Speed: 5178.466502 samples/sec	accuracy=0.584067	lr=0.399878
Epoch[6] Batch [399]	Speed: 5101.302661 samples/sec	accuracy=0.584676	lr=0.399870
Epoch[6] Batch [449]	Speed: 4997.971001 samples/sec	accuracy=0.585321	lr=0.399862
Epoch[6] Batch [499]	Speed: 5006.420852 samples/sec	accuracy=0.585907	lr=0.399854
Epoch[6] Batch [549]	Speed: 5337.496508 samples/sec	accuracy=0.586446	lr=0.399846
Epoch[6] Batch [599]	Speed: 5101.648896 samples/sec	accuracy=0.586988	lr=0.399837
Epoch[6] Batch [649]	Speed: 5479.885447 samples/sec	accuracy=0.587612	lr=0.399828
Epoch[6] Batch [699]	Speed: 4834.777668 samples/sec	accuracy=0.588210	lr=0.399819
Epoch[6] Batch [749]	Speed: 5332.730520 samples/sec	accuracy=0.588793	lr=0.399809
Epoch[6] Batch [799]	Speed: 5241.077228 samples/sec	accuracy=0.589338	lr=0.399800
Epoch[6] Batch [849]	Speed: 5174.033522 samples/sec	accuracy=0.589917	lr=0.399790
Epoch[6] Batch [899]	Speed: 5534.972926 samples/sec	accuracy=0.590482	lr=0.399780
Epoch[6] Batch [949]	Speed: 5315.267428 samples/sec	accuracy=0.591059	lr=0.399769
Epoch[6] Batch [999]	Speed: 5091.976251 samples/sec	accuracy=0.591609	lr=0.399759
Epoch[6] Batch [1049]	Speed: 4960.120897 samples/sec	accuracy=0.592114	lr=0.399748
Epoch[6] Batch [1099]	Speed: 5325.746324 samples/sec	accuracy=0.592608	lr=0.399737
Epoch[6] Batch [1149]	Speed: 5194.504779 samples/sec	accuracy=0.593189	lr=0.399725
Epoch[6] Batch [1199]	Speed: 4997.560771 samples/sec	accuracy=0.593632	lr=0.399714
Epoch[6] Batch [1249]	Speed: 4911.688423 samples/sec	accuracy=0.594111	lr=0.399702
[Epoch 6] training: accuracy=0.594123
[Epoch 6] speed: 5108 samples/sec	time cost: 260.369071
[Epoch 6] validation: err-top1=0.554110 err-top5=0.291235
Epoch[7] Batch [49]	Speed: 4871.556434 samples/sec	accuracy=0.603492	lr=0.399690
Epoch[7] Batch [99]	Speed: 5139.395585 samples/sec	accuracy=0.604684	lr=0.399677
Epoch[7] Batch [149]	Speed: 5104.621768 samples/sec	accuracy=0.605421	lr=0.399665
Epoch[7] Batch [199]	Speed: 5063.320155 samples/sec	accuracy=0.606061	lr=0.399652
Epoch[7] Batch [249]	Speed: 5144.632959 samples/sec	accuracy=0.606350	lr=0.399639
Epoch[7] Batch [299]	Speed: 5385.751253 samples/sec	accuracy=0.607079	lr=0.399626
Epoch[7] Batch [349]	Speed: 5215.677998 samples/sec	accuracy=0.607628	lr=0.399613
Epoch[7] Batch [399]	Speed: 5110.767411 samples/sec	accuracy=0.607976	lr=0.399599
Epoch[7] Batch [449]	Speed: 5007.899244 samples/sec	accuracy=0.608536	lr=0.399585
Epoch[7] Batch [499]	Speed: 5285.801847 samples/sec	accuracy=0.608792	lr=0.399571
Epoch[7] Batch [549]	Speed: 4976.064307 samples/sec	accuracy=0.609252	lr=0.399556
Epoch[7] Batch [599]	Speed: 4848.811874 samples/sec	accuracy=0.609731	lr=0.399542
Epoch[7] Batch [649]	Speed: 5058.693775 samples/sec	accuracy=0.610189	lr=0.399527
Epoch[7] Batch [699]	Speed: 5231.079930 samples/sec	accuracy=0.610633	lr=0.399512
Epoch[7] Batch [749]	Speed: 5112.749662 samples/sec	accuracy=0.611097	lr=0.399496
Epoch[7] Batch [799]	Speed: 5101.041409 samples/sec	accuracy=0.611555	lr=0.399481
Epoch[7] Batch [849]	Speed: 5150.226717 samples/sec	accuracy=0.611984	lr=0.399465
Epoch[7] Batch [899]	Speed: 5065.512371 samples/sec	accuracy=0.612358	lr=0.399449
Epoch[7] Batch [949]	Speed: 4987.643899 samples/sec	accuracy=0.612708	lr=0.399432
Epoch[7] Batch [999]	Speed: 5076.671078 samples/sec	accuracy=0.613134	lr=0.399416
Epoch[7] Batch [1049]	Speed: 5082.529062 samples/sec	accuracy=0.613473	lr=0.399399
Epoch[7] Batch [1099]	Speed: 5204.804996 samples/sec	accuracy=0.613825	lr=0.399382
Epoch[7] Batch [1149]	Speed: 5136.877719 samples/sec	accuracy=0.614229	lr=0.399365
Epoch[7] Batch [1199]	Speed: 5174.342947 samples/sec	accuracy=0.614623	lr=0.399347
Epoch[7] Batch [1249]	Speed: 5084.400502 samples/sec	accuracy=0.614952	lr=0.399329
[Epoch 7] training: accuracy=0.614958
[Epoch 7] speed: 5098 samples/sec	time cost: 260.839332
[Epoch 7] validation: err-top1=0.542809 err-top5=0.278101
Epoch[8] Batch [49]	Speed: 4919.722908 samples/sec	accuracy=0.620051	lr=0.399311
Epoch[8] Batch [99]	Speed: 5001.564525 samples/sec	accuracy=0.620783	lr=0.399293
Epoch[8] Batch [149]	Speed: 5004.740377 samples/sec	accuracy=0.620711	lr=0.399274
Epoch[8] Batch [199]	Speed: 5277.870633 samples/sec	accuracy=0.621998	lr=0.399256
Epoch[8] Batch [249]	Speed: 5119.044856 samples/sec	accuracy=0.622361	lr=0.399237
Epoch[8] Batch [299]	Speed: 5195.488293 samples/sec	accuracy=0.622727	lr=0.399218
Epoch[8] Batch [349]	Speed: 5423.988513 samples/sec	accuracy=0.623288	lr=0.399198
Epoch[8] Batch [399]	Speed: 5237.166936 samples/sec	accuracy=0.623696	lr=0.399179
Epoch[8] Batch [449]	Speed: 5224.965359 samples/sec	accuracy=0.624369	lr=0.399159
Epoch[8] Batch [499]	Speed: 4986.147802 samples/sec	accuracy=0.624611	lr=0.399138
Epoch[8] Batch [549]	Speed: 5342.224774 samples/sec	accuracy=0.624931	lr=0.399118
Epoch[8] Batch [599]	Speed: 5183.799618 samples/sec	accuracy=0.625271	lr=0.399098
Epoch[8] Batch [649]	Speed: 5174.676101 samples/sec	accuracy=0.625542	lr=0.399077
Epoch[8] Batch [699]	Speed: 5070.097489 samples/sec	accuracy=0.625989	lr=0.399056
Epoch[8] Batch [749]	Speed: 5185.176054 samples/sec	accuracy=0.626485	lr=0.399034
Epoch[8] Batch [799]	Speed: 5166.576129 samples/sec	accuracy=0.626900	lr=0.399013
Epoch[8] Batch [849]	Speed: 5166.148692 samples/sec	accuracy=0.627199	lr=0.398991
Epoch[8] Batch [899]	Speed: 4944.843672 samples/sec	accuracy=0.627504	lr=0.398969
Epoch[8] Batch [949]	Speed: 5149.823099 samples/sec	accuracy=0.627776	lr=0.398947
Epoch[8] Batch [999]	Speed: 4946.764799 samples/sec	accuracy=0.628080	lr=0.398924
Epoch[8] Batch [1049]	Speed: 4973.841213 samples/sec	accuracy=0.628376	lr=0.398901
Epoch[8] Batch [1099]	Speed: 5043.750134 samples/sec	accuracy=0.628683	lr=0.398878
Epoch[8] Batch [1149]	Speed: 5093.096701 samples/sec	accuracy=0.629043	lr=0.398855
Epoch[8] Batch [1199]	Speed: 5010.919333 samples/sec	accuracy=0.629372	lr=0.398832
Epoch[8] Batch [1249]	Speed: 5253.496655 samples/sec	accuracy=0.629684	lr=0.398808
[Epoch 8] training: accuracy=0.629685
[Epoch 8] speed: 5115 samples/sec	time cost: 260.260645
[Epoch 8] validation: err-top1=0.515725 err-top5=0.261300
Epoch[9] Batch [49]	Speed: 4949.836050 samples/sec	accuracy=0.635008	lr=0.398784
Epoch[9] Batch [99]	Speed: 4996.385818 samples/sec	accuracy=0.634179	lr=0.398760
Epoch[9] Batch [149]	Speed: 4969.699409 samples/sec	accuracy=0.634194	lr=0.398735
Epoch[9] Batch [199]	Speed: 5161.772127 samples/sec	accuracy=0.634735	lr=0.398710
Epoch[9] Batch [249]	Speed: 5091.104432 samples/sec	accuracy=0.635323	lr=0.398686
Epoch[9] Batch [299]	Speed: 5091.436490 samples/sec	accuracy=0.635670	lr=0.398660
Epoch[9] Batch [349]	Speed: 5217.240498 samples/sec	accuracy=0.635879	lr=0.398635
Epoch[9] Batch [399]	Speed: 5226.936313 samples/sec	accuracy=0.636332	lr=0.398610
Epoch[9] Batch [449]	Speed: 5201.841711 samples/sec	accuracy=0.636870	lr=0.398584
Epoch[9] Batch [499]	Speed: 5339.944308 samples/sec	accuracy=0.637159	lr=0.398558
Epoch[9] Batch [549]	Speed: 5220.612736 samples/sec	accuracy=0.637659	lr=0.398531
Epoch[9] Batch [599]	Speed: 5183.682623 samples/sec	accuracy=0.637887	lr=0.398505
Epoch[9] Batch [649]	Speed: 5081.795639 samples/sec	accuracy=0.638196	lr=0.398478
Epoch[9] Batch [699]	Speed: 5099.105880 samples/sec	accuracy=0.638471	lr=0.398451
Epoch[9] Batch [749]	Speed: 5262.529303 samples/sec	accuracy=0.638813	lr=0.398424
Epoch[9] Batch [799]	Speed: 5100.778730 samples/sec	accuracy=0.639070	lr=0.398396
Epoch[9] Batch [849]	Speed: 5151.024136 samples/sec	accuracy=0.639257	lr=0.398369
Epoch[9] Batch [899]	Speed: 5220.627077 samples/sec	accuracy=0.639491	lr=0.398341
Epoch[9] Batch [949]	Speed: 4873.273054 samples/sec	accuracy=0.639811	lr=0.398312
Epoch[9] Batch [999]	Speed: 5157.640422 samples/sec	accuracy=0.640211	lr=0.398284
Epoch[9] Batch [1049]	Speed: 4892.524482 samples/sec	accuracy=0.640545	lr=0.398255
Epoch[9] Batch [1099]	Speed: 5095.616436 samples/sec	accuracy=0.640857	lr=0.398226
Epoch[9] Batch [1149]	Speed: 4981.804728 samples/sec	accuracy=0.641080	lr=0.398197
Epoch[9] Batch [1199]	Speed: 4983.020702 samples/sec	accuracy=0.641333	lr=0.398168
Epoch[9] Batch [1249]	Speed: 5340.893480 samples/sec	accuracy=0.641566	lr=0.398138
[Epoch 9] training: accuracy=0.641577
[Epoch 9] speed: 5107 samples/sec	time cost: 260.329318
[Epoch 9] validation: err-top1=0.512197 err-top5=0.256099
Epoch[10] Batch [49]	Speed: 5031.414418 samples/sec	accuracy=0.644933	lr=0.398108
Epoch[10] Batch [99]	Speed: 4836.773035 samples/sec	accuracy=0.644438	lr=0.398078
Epoch[10] Batch [149]	Speed: 5130.400969 samples/sec	accuracy=0.644577	lr=0.398048
Epoch[10] Batch [199]	Speed: 5026.945212 samples/sec	accuracy=0.644502	lr=0.398017
Epoch[10] Batch [249]	Speed: 5135.879051 samples/sec	accuracy=0.645047	lr=0.397986
Epoch[10] Batch [299]	Speed: 4948.864413 samples/sec	accuracy=0.645482	lr=0.397955
Epoch[10] Batch [349]	Speed: 5204.342580 samples/sec	accuracy=0.645888	lr=0.397924
Epoch[10] Batch [399]	Speed: 5154.460119 samples/sec	accuracy=0.646351	lr=0.397892
Epoch[10] Batch [449]	Speed: 5238.835634 samples/sec	accuracy=0.646590	lr=0.397861
Epoch[10] Batch [499]	Speed: 5074.381304 samples/sec	accuracy=0.646866	lr=0.397829
Epoch[10] Batch [549]	Speed: 5170.785415 samples/sec	accuracy=0.647025	lr=0.397796
Epoch[10] Batch [599]	Speed: 5106.949123 samples/sec	accuracy=0.647347	lr=0.397764
Epoch[10] Batch [649]	Speed: 5178.838403 samples/sec	accuracy=0.647681	lr=0.397731
Epoch[10] Batch [699]	Speed: 5176.990411 samples/sec	accuracy=0.647934	lr=0.397698
Epoch[10] Batch [749]	Speed: 5140.303708 samples/sec	accuracy=0.648214	lr=0.397665
Epoch[10] Batch [799]	Speed: 5049.224490 samples/sec	accuracy=0.648416	lr=0.397632
Epoch[10] Batch [849]	Speed: 5063.732416 samples/sec	accuracy=0.648648	lr=0.397598
Epoch[10] Batch [899]	Speed: 5246.578864 samples/sec	accuracy=0.648823	lr=0.397564
Epoch[10] Batch [949]	Speed: 5157.642033 samples/sec	accuracy=0.648976	lr=0.397530
Epoch[10] Batch [999]	Speed: 5106.673692 samples/sec	accuracy=0.649150	lr=0.397496
Epoch[10] Batch [1049]	Speed: 4972.905962 samples/sec	accuracy=0.649484	lr=0.397461
Epoch[10] Batch [1099]	Speed: 4877.092026 samples/sec	accuracy=0.649680	lr=0.397427
Epoch[10] Batch [1149]	Speed: 4946.143510 samples/sec	accuracy=0.649901	lr=0.397392
Epoch[10] Batch [1199]	Speed: 5153.793978 samples/sec	accuracy=0.650117	lr=0.397356
Epoch[10] Batch [1249]	Speed: 5252.802873 samples/sec	accuracy=0.650284	lr=0.397321
[Epoch 10] training: accuracy=0.650292
[Epoch 10] speed: 5088 samples/sec	time cost: 261.836470
[Epoch 10] validation: err-top1=0.484534 err-top5=0.228137
Epoch[11] Batch [49]	Speed: 4862.654601 samples/sec	accuracy=0.653582	lr=0.397284
Epoch[11] Batch [99]	Speed: 4808.511841 samples/sec	accuracy=0.653229	lr=0.397248
Epoch[11] Batch [149]	Speed: 4973.999619 samples/sec	accuracy=0.652621	lr=0.397212
Epoch[11] Batch [199]	Speed: 5014.170886 samples/sec	accuracy=0.652992	lr=0.397176
Epoch[11] Batch [249]	Speed: 5114.303221 samples/sec	accuracy=0.652987	lr=0.397139
Epoch[11] Batch [299]	Speed: 4905.990899 samples/sec	accuracy=0.653180	lr=0.397102
Epoch[11] Batch [349]	Speed: 4948.452512 samples/sec	accuracy=0.653398	lr=0.397065
Epoch[11] Batch [399]	Speed: 5073.273146 samples/sec	accuracy=0.653645	lr=0.397028
Epoch[11] Batch [449]	Speed: 5022.463735 samples/sec	accuracy=0.654072	lr=0.396990
Epoch[11] Batch [499]	Speed: 5335.166793 samples/sec	accuracy=0.654077	lr=0.396952
Epoch[11] Batch [549]	Speed: 5195.547371 samples/sec	accuracy=0.654236	lr=0.396914
Epoch[11] Batch [599]	Speed: 4967.843403 samples/sec	accuracy=0.654453	lr=0.396876
Epoch[11] Batch [649]	Speed: 5533.826180 samples/sec	accuracy=0.654713	lr=0.396837
Epoch[11] Batch [699]	Speed: 5253.437408 samples/sec	accuracy=0.654961	lr=0.396798
Epoch[11] Batch [749]	Speed: 5064.003354 samples/sec	accuracy=0.655116	lr=0.396759
Epoch[11] Batch [799]	Speed: 5290.442186 samples/sec	accuracy=0.655367	lr=0.396720
Epoch[11] Batch [849]	Speed: 5195.722979 samples/sec	accuracy=0.655667	lr=0.396680
Epoch[11] Batch [899]	Speed: 5162.267091 samples/sec	accuracy=0.655995	lr=0.396641
Epoch[11] Batch [949]	Speed: 5211.252178 samples/sec	accuracy=0.656254	lr=0.396601
Epoch[11] Batch [999]	Speed: 5113.283237 samples/sec	accuracy=0.656514	lr=0.396561
Epoch[11] Batch [1049]	Speed: 5192.946950 samples/sec	accuracy=0.656728	lr=0.396520
Epoch[11] Batch [1099]	Speed: 4998.790849 samples/sec	accuracy=0.656961	lr=0.396479
Epoch[11] Batch [1149]	Speed: 5200.374430 samples/sec	accuracy=0.657293	lr=0.396439
Epoch[11] Batch [1199]	Speed: 4992.629563 samples/sec	accuracy=0.657509	lr=0.396397
Epoch[11] Batch [1249]	Speed: 5018.245986 samples/sec	accuracy=0.657747	lr=0.396356
[Epoch 11] training: accuracy=0.657752
[Epoch 11] speed: 5088 samples/sec	time cost: 261.109077
[Epoch 11] validation: err-top1=0.465800 err-top5=0.218323
Epoch[12] Batch [49]	Speed: 4876.986139 samples/sec	accuracy=0.660050	lr=0.396314
Epoch[12] Batch [99]	Speed: 5268.193696 samples/sec	accuracy=0.659126	lr=0.396272
Epoch[12] Batch [149]	Speed: 5074.047391 samples/sec	accuracy=0.659977	lr=0.396230
Epoch[12] Batch [199]	Speed: 5139.138165 samples/sec	accuracy=0.659664	lr=0.396187
Epoch[12] Batch [249]	Speed: 4962.644574 samples/sec	accuracy=0.659512	lr=0.396145
Epoch[12] Batch [299]	Speed: 5061.873532 samples/sec	accuracy=0.659888	lr=0.396102
Epoch[12] Batch [349]	Speed: 5036.771876 samples/sec	accuracy=0.660063	lr=0.396059
Epoch[12] Batch [399]	Speed: 5041.324499 samples/sec	accuracy=0.660296	lr=0.396016
Epoch[12] Batch [449]	Speed: 4979.131883 samples/sec	accuracy=0.660411	lr=0.395972
Epoch[12] Batch [499]	Speed: 4963.444497 samples/sec	accuracy=0.660540	lr=0.395928
Epoch[12] Batch [549]	Speed: 5368.453582 samples/sec	accuracy=0.660728	lr=0.395885
Epoch[12] Batch [599]	Speed: 5143.666632 samples/sec	accuracy=0.660924	lr=0.395840
Epoch[12] Batch [649]	Speed: 5025.207071 samples/sec	accuracy=0.661177	lr=0.395796
Epoch[12] Batch [699]	Speed: 5213.776536 samples/sec	accuracy=0.661536	lr=0.395751
Epoch[12] Batch [749]	Speed: 4994.769693 samples/sec	accuracy=0.661776	lr=0.395706
Epoch[12] Batch [799]	Speed: 5261.540743 samples/sec	accuracy=0.661936	lr=0.395661
Epoch[12] Batch [849]	Speed: 5162.136919 samples/sec	accuracy=0.662201	lr=0.395616
Epoch[12] Batch [899]	Speed: 4969.807404 samples/sec	accuracy=0.662469	lr=0.395570
Epoch[12] Batch [949]	Speed: 5021.448466 samples/sec	accuracy=0.662715	lr=0.395524
Epoch[12] Batch [999]	Speed: 5152.559630 samples/sec	accuracy=0.662913	lr=0.395478
Epoch[12] Batch [1049]	Speed: 4919.015434 samples/sec	accuracy=0.663112	lr=0.395432
Epoch[12] Batch [1099]	Speed: 5159.401365 samples/sec	accuracy=0.663290	lr=0.395386
Epoch[12] Batch [1149]	Speed: 5262.267009 samples/sec	accuracy=0.663426	lr=0.395339
Epoch[12] Batch [1199]	Speed: 5203.128404 samples/sec	accuracy=0.663609	lr=0.395292
Epoch[12] Batch [1249]	Speed: 5313.636861 samples/sec	accuracy=0.663805	lr=0.395245
[Epoch 12] training: accuracy=0.663807
[Epoch 12] speed: 5095 samples/sec	time cost: 261.288179
[Epoch 12] validation: err-top1=0.476064 err-top5=0.222836
Epoch[13] Batch [49]	Speed: 4829.269242 samples/sec	accuracy=0.663925	lr=0.395196
Epoch[13] Batch [99]	Speed: 4871.494769 samples/sec	accuracy=0.663746	lr=0.395149
Epoch[13] Batch [149]	Speed: 5028.086080 samples/sec	accuracy=0.663950	lr=0.395101
Epoch[13] Batch [199]	Speed: 5080.395290 samples/sec	accuracy=0.664979	lr=0.395052
Epoch[13] Batch [249]	Speed: 4867.119924 samples/sec	accuracy=0.665427	lr=0.395004
Epoch[13] Batch [299]	Speed: 5102.239072 samples/sec	accuracy=0.665644	lr=0.394955
Epoch[13] Batch [349]	Speed: 5130.080109 samples/sec	accuracy=0.665685	lr=0.394907
Epoch[13] Batch [399]	Speed: 5162.662112 samples/sec	accuracy=0.665825	lr=0.394858
Epoch[13] Batch [449]	Speed: 5166.008010 samples/sec	accuracy=0.666049	lr=0.394808
Epoch[13] Batch [499]	Speed: 5142.402896 samples/sec	accuracy=0.666396	lr=0.394759
Epoch[13] Batch [549]	Speed: 5078.621438 samples/sec	accuracy=0.666729	lr=0.394709
Epoch[13] Batch [599]	Speed: 5290.691003 samples/sec	accuracy=0.667040	lr=0.394659
Epoch[13] Batch [649]	Speed: 5192.535478 samples/sec	accuracy=0.667182	lr=0.394609
Epoch[13] Batch [699]	Speed: 4994.834402 samples/sec	accuracy=0.667320	lr=0.394558
Epoch[13] Batch [749]	Speed: 5130.711940 samples/sec	accuracy=0.667586	lr=0.394507
Epoch[13] Batch [799]	Speed: 5174.390699 samples/sec	accuracy=0.667799	lr=0.394457
Epoch[13] Batch [849]	Speed: 5064.879891 samples/sec	accuracy=0.668031	lr=0.394405
Epoch[13] Batch [899]	Speed: 5034.175804 samples/sec	accuracy=0.668279	lr=0.394354
Epoch[13] Batch [949]	Speed: 5319.982865 samples/sec	accuracy=0.668532	lr=0.394302
Epoch[13] Batch [999]	Speed: 4977.213224 samples/sec	accuracy=0.668773	lr=0.394250
Epoch[13] Batch [1049]	Speed: 5171.569909 samples/sec	accuracy=0.668909	lr=0.394198
Epoch[13] Batch [1099]	Speed: 5193.808150 samples/sec	accuracy=0.669125	lr=0.394146
Epoch[13] Batch [1149]	Speed: 5251.113967 samples/sec	accuracy=0.669319	lr=0.394093
Epoch[13] Batch [1199]	Speed: 5475.697542 samples/sec	accuracy=0.669435	lr=0.394041
Epoch[13] Batch [1249]	Speed: 5509.767572 samples/sec	accuracy=0.669564	lr=0.393988
[Epoch 13] training: accuracy=0.669569
[Epoch 13] speed: 5119 samples/sec	time cost: 259.709157
[Epoch 13] validation: err-top1=0.483598 err-top5=0.232023
Epoch[14] Batch [49]	Speed: 4865.631086 samples/sec	accuracy=0.671532	lr=0.393933
Epoch[14] Batch [99]	Speed: 4889.533716 samples/sec	accuracy=0.671482	lr=0.393880
Epoch[14] Batch [149]	Speed: 4821.255660 samples/sec	accuracy=0.670935	lr=0.393826
Epoch[14] Batch [199]	Speed: 5341.574323 samples/sec	accuracy=0.670893	lr=0.393772
Epoch[14] Batch [249]	Speed: 5358.545299 samples/sec	accuracy=0.671089	lr=0.393718
Epoch[14] Batch [299]	Speed: 5107.686910 samples/sec	accuracy=0.671214	lr=0.393664
Epoch[14] Batch [349]	Speed: 5427.020694 samples/sec	accuracy=0.671477	lr=0.393609
Epoch[14] Batch [399]	Speed: 5149.422260 samples/sec	accuracy=0.671495	lr=0.393554
Epoch[14] Batch [449]	Speed: 5397.293456 samples/sec	accuracy=0.671650	lr=0.393499
Epoch[14] Batch [499]	Speed: 5174.168906 samples/sec	accuracy=0.671755	lr=0.393444
Epoch[14] Batch [549]	Speed: 5057.262061 samples/sec	accuracy=0.671848	lr=0.393388
Epoch[14] Batch [599]	Speed: 5138.240657 samples/sec	accuracy=0.671990	lr=0.393332
Epoch[14] Batch [649]	Speed: 5104.696635 samples/sec	accuracy=0.672030	lr=0.393276
Epoch[14] Batch [699]	Speed: 5037.239258 samples/sec	accuracy=0.672128	lr=0.393220
Epoch[14] Batch [749]	Speed: 4943.332173 samples/sec	accuracy=0.672314	lr=0.393163
Epoch[14] Batch [799]	Speed: 4924.454843 samples/sec	accuracy=0.672513	lr=0.393107
Epoch[14] Batch [849]	Speed: 5011.483790 samples/sec	accuracy=0.672677	lr=0.393050
Epoch[14] Batch [899]	Speed: 4975.771223 samples/sec	accuracy=0.672992	lr=0.392993
Epoch[14] Batch [949]	Speed: 5140.778934 samples/sec	accuracy=0.673127	lr=0.392935
Epoch[14] Batch [999]	Speed: 5118.802648 samples/sec	accuracy=0.673259	lr=0.392877
Epoch[14] Batch [1049]	Speed: 5263.122720 samples/sec	accuracy=0.673438	lr=0.392820
Epoch[14] Batch [1099]	Speed: 5137.609184 samples/sec	accuracy=0.673560	lr=0.392762
Epoch[14] Batch [1149]	Speed: 5163.768197 samples/sec	accuracy=0.673686	lr=0.392703
Epoch[14] Batch [1199]	Speed: 5250.664341 samples/sec	accuracy=0.673948	lr=0.392645
Epoch[14] Batch [1249]	Speed: 5254.277266 samples/sec	accuracy=0.674189	lr=0.392586
[Epoch 14] training: accuracy=0.674188
[Epoch 14] speed: 5112 samples/sec	time cost: 260.246734
[Epoch 14] validation: err-top1=0.512357 err-top5=0.256138
Epoch[15] Batch [49]	Speed: 5049.624841 samples/sec	accuracy=0.674483	lr=0.392526
Epoch[15] Batch [99]	Speed: 5225.535458 samples/sec	accuracy=0.675183	lr=0.392466
Epoch[15] Batch [149]	Speed: 5020.774470 samples/sec	accuracy=0.675707	lr=0.392407
Epoch[15] Batch [199]	Speed: 5118.568394 samples/sec	accuracy=0.675759	lr=0.392347
Epoch[15] Batch [249]	Speed: 5200.564470 samples/sec	accuracy=0.675762	lr=0.392287
Epoch[15] Batch [299]	Speed: 5023.222078 samples/sec	accuracy=0.675542	lr=0.392227
Epoch[15] Batch [349]	Speed: 5119.719010 samples/sec	accuracy=0.675529	lr=0.392167
Epoch[15] Batch [399]	Speed: 4813.766450 samples/sec	accuracy=0.675613	lr=0.392106
Epoch[15] Batch [449]	Speed: 5198.270438 samples/sec	accuracy=0.675791	lr=0.392045
Epoch[15] Batch [499]	Speed: 5007.559661 samples/sec	accuracy=0.675937	lr=0.391984
Epoch[15] Batch [549]	Speed: 4950.693137 samples/sec	accuracy=0.676205	lr=0.391923
Epoch[15] Batch [599]	Speed: 5036.926873 samples/sec	accuracy=0.676310	lr=0.391861
Epoch[15] Batch [649]	Speed: 4865.424942 samples/sec	accuracy=0.676477	lr=0.391799
Epoch[15] Batch [699]	Speed: 4886.692158 samples/sec	accuracy=0.676549	lr=0.391737
Epoch[15] Batch [749]	Speed: 4980.974037 samples/sec	accuracy=0.676848	lr=0.391675
Epoch[15] Batch [799]	Speed: 4991.788412 samples/sec	accuracy=0.676897	lr=0.391613
Epoch[15] Batch [849]	Speed: 5081.460629 samples/sec	accuracy=0.676946	lr=0.391550
Epoch[15] Batch [899]	Speed: 5005.446825 samples/sec	accuracy=0.677019	lr=0.391487
Epoch[15] Batch [949]	Speed: 5093.999773 samples/sec	accuracy=0.677189	lr=0.391424
Epoch[15] Batch [999]	Speed: 5216.441581 samples/sec	accuracy=0.677303	lr=0.391361
Epoch[15] Batch [1049]	Speed: 5051.522506 samples/sec	accuracy=0.677542	lr=0.391297
Epoch[15] Batch [1099]	Speed: 5104.441587 samples/sec	accuracy=0.677754	lr=0.391233
Epoch[15] Batch [1149]	Speed: 5159.404463 samples/sec	accuracy=0.677978	lr=0.391169
Epoch[15] Batch [1199]	Speed: 5092.720345 samples/sec	accuracy=0.678174	lr=0.391105
Epoch[15] Batch [1249]	Speed: 5127.798849 samples/sec	accuracy=0.678414	lr=0.391040
[Epoch 15] training: accuracy=0.678419
[Epoch 15] speed: 5051 samples/sec	time cost: 263.113871
[Epoch 15] validation: err-top1=0.450893 err-top5=0.204640
Epoch[16] Batch [49]	Speed: 4690.072307 samples/sec	accuracy=0.679011	lr=0.390974
Epoch[16] Batch [99]	Speed: 5159.474004 samples/sec	accuracy=0.678533	lr=0.390909
Epoch[16] Batch [149]	Speed: 4962.960544 samples/sec	accuracy=0.678885	lr=0.390844
Epoch[16] Batch [199]	Speed: 4947.594149 samples/sec	accuracy=0.678847	lr=0.390779
Epoch[16] Batch [249]	Speed: 5020.552388 samples/sec	accuracy=0.678454	lr=0.390713
Epoch[16] Batch [299]	Speed: 4996.708657 samples/sec	accuracy=0.678779	lr=0.390647
Epoch[16] Batch [349]	Speed: 5240.703240 samples/sec	accuracy=0.679011	lr=0.390581
Epoch[16] Batch [399]	Speed: 5049.731470 samples/sec	accuracy=0.679015	lr=0.390515
Epoch[16] Batch [449]	Speed: 5081.823298 samples/sec	accuracy=0.679131	lr=0.390448
Epoch[16] Batch [499]	Speed: 5000.564441 samples/sec	accuracy=0.679299	lr=0.390381
Epoch[16] Batch [549]	Speed: 5171.790731 samples/sec	accuracy=0.679534	lr=0.390314
Epoch[16] Batch [599]	Speed: 5372.321335 samples/sec	accuracy=0.679725	lr=0.390247
Epoch[16] Batch [649]	Speed: 4906.571985 samples/sec	accuracy=0.679901	lr=0.390180
Epoch[16] Batch [699]	Speed: 5259.784119 samples/sec	accuracy=0.679992	lr=0.390112
Epoch[16] Batch [749]	Speed: 5177.268362 samples/sec	accuracy=0.680130	lr=0.390044
Epoch[16] Batch [799]	Speed: 5263.534490 samples/sec	accuracy=0.680259	lr=0.389976
Epoch[16] Batch [849]	Speed: 5541.179009 samples/sec	accuracy=0.680432	lr=0.389907
Epoch[16] Batch [899]	Speed: 4813.522059 samples/sec	accuracy=0.680661	lr=0.389839
Epoch[16] Batch [949]	Speed: 5160.979191 samples/sec	accuracy=0.680866	lr=0.389770
Epoch[16] Batch [999]	Speed: 5457.246949 samples/sec	accuracy=0.680979	lr=0.389701
Epoch[16] Batch [1049]	Speed: 4972.712275 samples/sec	accuracy=0.681211	lr=0.389632
Epoch[16] Batch [1099]	Speed: 5376.864614 samples/sec	accuracy=0.681393	lr=0.389562
Epoch[16] Batch [1149]	Speed: 5432.565945 samples/sec	accuracy=0.681492	lr=0.389492
Epoch[16] Batch [1199]	Speed: 5273.807352 samples/sec	accuracy=0.681689	lr=0.389422
Epoch[16] Batch [1249]	Speed: 5278.246572 samples/sec	accuracy=0.681870	lr=0.389352
[Epoch 16] training: accuracy=0.681876
[Epoch 16] speed: 5129 samples/sec	time cost: 259.096139
[Epoch 16] validation: err-top1=0.460579 err-top5=0.206075
Epoch[17] Batch [49]	Speed: 4733.057384 samples/sec	accuracy=0.682237	lr=0.389280
Epoch[17] Batch [99]	Speed: 5174.138114 samples/sec	accuracy=0.682361	lr=0.389210
Epoch[17] Batch [149]	Speed: 4946.259712 samples/sec	accuracy=0.681790	lr=0.389139
Epoch[17] Batch [199]	Speed: 5157.195761 samples/sec	accuracy=0.681790	lr=0.389068
Epoch[17] Batch [249]	Speed: 4946.099537 samples/sec	accuracy=0.681998	lr=0.388996
Epoch[17] Batch [299]	Speed: 5022.660847 samples/sec	accuracy=0.682363	lr=0.388925
Epoch[17] Batch [349]	Speed: 5190.592636 samples/sec	accuracy=0.682540	lr=0.388853
Epoch[17] Batch [399]	Speed: 5033.580975 samples/sec	accuracy=0.682701	lr=0.388781
Epoch[17] Batch [449]	Speed: 5152.664839 samples/sec	accuracy=0.682936	lr=0.388709
Epoch[17] Batch [499]	Speed: 5197.188139 samples/sec	accuracy=0.683178	lr=0.388636
Epoch[17] Batch [549]	Speed: 5078.579641 samples/sec	accuracy=0.683336	lr=0.388564
Epoch[17] Batch [599]	Speed: 5147.435907 samples/sec	accuracy=0.683585	lr=0.388491
Epoch[17] Batch [649]	Speed: 5025.596095 samples/sec	accuracy=0.683662	lr=0.388418
Epoch[17] Batch [699]	Speed: 5165.896911 samples/sec	accuracy=0.683741	lr=0.388344
Epoch[17] Batch [749]	Speed: 5047.873475 samples/sec	accuracy=0.683845	lr=0.388271
Epoch[17] Batch [799]	Speed: 5073.751762 samples/sec	accuracy=0.683984	lr=0.388197
Epoch[17] Batch [849]	Speed: 4923.511543 samples/sec	accuracy=0.684122	lr=0.388123
Epoch[17] Batch [899]	Speed: 5220.841573 samples/sec	accuracy=0.684262	lr=0.388049
Epoch[17] Batch [949]	Speed: 5242.711429 samples/sec	accuracy=0.684402	lr=0.387974
Epoch[17] Batch [999]	Speed: 5229.868531 samples/sec	accuracy=0.684564	lr=0.387900
Epoch[17] Batch [1049]	Speed: 5243.518414 samples/sec	accuracy=0.684741	lr=0.387825
Epoch[17] Batch [1099]	Speed: 5195.591241 samples/sec	accuracy=0.684917	lr=0.387750
Epoch[17] Batch [1149]	Speed: 5094.568842 samples/sec	accuracy=0.685031	lr=0.387674
Epoch[17] Batch [1199]	Speed: 5298.150863 samples/sec	accuracy=0.685173	lr=0.387599
Epoch[17] Batch [1249]	Speed: 4976.781597 samples/sec	accuracy=0.685321	lr=0.387523
[Epoch 17] training: accuracy=0.685328
[Epoch 17] speed: 5093 samples/sec	time cost: 260.739214
[Epoch 17] validation: err-top1=0.455648 err-top5=0.204081
Epoch[18] Batch [49]	Speed: 4946.413062 samples/sec	accuracy=0.686645	lr=0.387445
Epoch[18] Batch [99]	Speed: 4920.732971 samples/sec	accuracy=0.686846	lr=0.387369
Epoch[18] Batch [149]	Speed: 5157.193656 samples/sec	accuracy=0.686774	lr=0.387293
Epoch[18] Batch [199]	Speed: 5181.612993 samples/sec	accuracy=0.686584	lr=0.387216
Epoch[18] Batch [249]	Speed: 5210.721225 samples/sec	accuracy=0.686247	lr=0.387139
Epoch[18] Batch [299]	Speed: 5245.622169 samples/sec	accuracy=0.686276	lr=0.387062
Epoch[18] Batch [349]	Speed: 4922.748362 samples/sec	accuracy=0.686448	lr=0.386984
Epoch[18] Batch [399]	Speed: 5061.569776 samples/sec	accuracy=0.686544	lr=0.386907
Epoch[18] Batch [449]	Speed: 5180.716962 samples/sec	accuracy=0.686901	lr=0.386829
Epoch[18] Batch [499]	Speed: 5088.603746 samples/sec	accuracy=0.687044	lr=0.386751
Epoch[18] Batch [549]	Speed: 5130.965451 samples/sec	accuracy=0.687162	lr=0.386673
Epoch[18] Batch [599]	Speed: 5291.858371 samples/sec	accuracy=0.687174	lr=0.386594
Epoch[18] Batch [649]	Speed: 4946.017403 samples/sec	accuracy=0.687266	lr=0.386515
Epoch[18] Batch [699]	Speed: 4786.822819 samples/sec	accuracy=0.687177	lr=0.386436
Epoch[18] Batch [749]	Speed: 5121.353019 samples/sec	accuracy=0.687268	lr=0.386357
Epoch[18] Batch [799]	Speed: 5319.223453 samples/sec	accuracy=0.687337	lr=0.386278
Epoch[18] Batch [849]	Speed: 5157.499212 samples/sec	accuracy=0.687555	lr=0.386198
Epoch[18] Batch [899]	Speed: 5095.311760 samples/sec	accuracy=0.687618	lr=0.386119
Epoch[18] Batch [949]	Speed: 5236.979959 samples/sec	accuracy=0.687656	lr=0.386038
Epoch[18] Batch [999]	Speed: 5049.392364 samples/sec	accuracy=0.687814	lr=0.385958
Epoch[18] Batch [1049]	Speed: 5061.782974 samples/sec	accuracy=0.688005	lr=0.385878
Epoch[18] Batch [1099]	Speed: 5413.408991 samples/sec	accuracy=0.688105	lr=0.385797
Epoch[18] Batch [1149]	Speed: 5244.754846 samples/sec	accuracy=0.688204	lr=0.385716
Epoch[18] Batch [1199]	Speed: 5076.061605 samples/sec	accuracy=0.688424	lr=0.385635
Epoch[18] Batch [1249]	Speed: 5005.951003 samples/sec	accuracy=0.688510	lr=0.385554
[Epoch 18] training: accuracy=0.688510
[Epoch 18] speed: 5105 samples/sec	time cost: 260.641527
[Epoch 18] validation: err-top1=0.425801 err-top5=0.187281
Epoch[19] Batch [49]	Speed: 4830.992700 samples/sec	accuracy=0.687108	lr=0.385470
Epoch[19] Batch [99]	Speed: 5122.211770 samples/sec	accuracy=0.687456	lr=0.385388
Epoch[19] Batch [149]	Speed: 5118.322205 samples/sec	accuracy=0.688037	lr=0.385306
Epoch[19] Batch [199]	Speed: 5264.013291 samples/sec	accuracy=0.688638	lr=0.385224
Epoch[19] Batch [249]	Speed: 4965.810104 samples/sec	accuracy=0.688636	lr=0.385142
Epoch[19] Batch [299]	Speed: 5053.940959 samples/sec	accuracy=0.688948	lr=0.385059
Epoch[19] Batch [349]	Speed: 5001.802871 samples/sec	accuracy=0.689079	lr=0.384976
Epoch[19] Batch [399]	Speed: 5213.926288 samples/sec	accuracy=0.689142	lr=0.384893
Epoch[19] Batch [449]	Speed: 5029.058456 samples/sec	accuracy=0.689414	lr=0.384810
Epoch[19] Batch [499]	Speed: 5174.302304 samples/sec	accuracy=0.689502	lr=0.384726
Epoch[19] Batch [549]	Speed: 4878.526155 samples/sec	accuracy=0.689804	lr=0.384642
Epoch[19] Batch [599]	Speed: 4951.519580 samples/sec	accuracy=0.689905	lr=0.384558
Epoch[19] Batch [649]	Speed: 5114.975029 samples/sec	accuracy=0.690056	lr=0.384474
Epoch[19] Batch [699]	Speed: 4969.146164 samples/sec	accuracy=0.690161	lr=0.384389
Epoch[19] Batch [749]	Speed: 5117.377808 samples/sec	accuracy=0.690227	lr=0.384305
Epoch[19] Batch [799]	Speed: 5409.928311 samples/sec	accuracy=0.690436	lr=0.384220
Epoch[19] Batch [849]	Speed: 5090.184289 samples/sec	accuracy=0.690529	lr=0.384135
Epoch[19] Batch [899]	Speed: 5366.944871 samples/sec	accuracy=0.690640	lr=0.384049
Epoch[19] Batch [949]	Speed: 5221.634348 samples/sec	accuracy=0.690788	lr=0.383964
Epoch[19] Batch [999]	Speed: 5183.607548 samples/sec	accuracy=0.690821	lr=0.383878
Epoch[19] Batch [1049]	Speed: 5253.129502 samples/sec	accuracy=0.690957	lr=0.383792
Epoch[19] Batch [1099]	Speed: 5062.648836 samples/sec	accuracy=0.691149	lr=0.383706
Epoch[19] Batch [1149]	Speed: 5038.242364 samples/sec	accuracy=0.691293	lr=0.383619
Epoch[19] Batch [1199]	Speed: 5066.832676 samples/sec	accuracy=0.691397	lr=0.383533
Epoch[19] Batch [1249]	Speed: 5208.325886 samples/sec	accuracy=0.691528	lr=0.383446
[Epoch 19] training: accuracy=0.691519
[Epoch 19] speed: 5099 samples/sec	time cost: 260.941746
[Epoch 19] validation: err-top1=0.482302 err-top5=0.232940
Epoch[20] Batch [49]	Speed: 4914.932227 samples/sec	accuracy=0.688948	lr=0.383357
Epoch[20] Batch [99]	Speed: 4926.855423 samples/sec	accuracy=0.689216	lr=0.383270
Epoch[20] Batch [149]	Speed: 5302.779399 samples/sec	accuracy=0.689706	lr=0.383182
Epoch[20] Batch [199]	Speed: 5173.255633 samples/sec	accuracy=0.690070	lr=0.383094
Epoch[20] Batch [249]	Speed: 5052.926955 samples/sec	accuracy=0.690159	lr=0.383006
Epoch[20] Batch [299]	Speed: 5107.960020 samples/sec	accuracy=0.690144	lr=0.382918
Epoch[20] Batch [349]	Speed: 5043.246841 samples/sec	accuracy=0.690375	lr=0.382830
Epoch[20] Batch [399]	Speed: 5220.847920 samples/sec	accuracy=0.690554	lr=0.382741
Epoch[20] Batch [449]	Speed: 5079.786966 samples/sec	accuracy=0.690854	lr=0.382652
Epoch[20] Batch [499]	Speed: 5512.423372 samples/sec	accuracy=0.691042	lr=0.382563
Epoch[20] Batch [549]	Speed: 5290.915598 samples/sec	accuracy=0.691215	lr=0.382474
Epoch[20] Batch [599]	Speed: 5169.927974 samples/sec	accuracy=0.691385	lr=0.382384
Epoch[20] Batch [649]	Speed: 5152.876384 samples/sec	accuracy=0.691494	lr=0.382295
Epoch[20] Batch [699]	Speed: 5307.757618 samples/sec	accuracy=0.691687	lr=0.382205
Epoch[20] Batch [749]	Speed: 5187.341004 samples/sec	accuracy=0.691827	lr=0.382115
Epoch[20] Batch [799]	Speed: 5036.934906 samples/sec	accuracy=0.692070	lr=0.382024
Epoch[20] Batch [849]	Speed: 5094.325078 samples/sec	accuracy=0.692298	lr=0.381934
Epoch[20] Batch [899]	Speed: 5190.453630 samples/sec	accuracy=0.692433	lr=0.381843
Epoch[20] Batch [949]	Speed: 5290.841817 samples/sec	accuracy=0.692676	lr=0.381752
Epoch[20] Batch [999]	Speed: 5096.548705 samples/sec	accuracy=0.692886	lr=0.381661
Epoch[20] Batch [1049]	Speed: 5010.465240 samples/sec	accuracy=0.693068	lr=0.381569
Epoch[20] Batch [1099]	Speed: 5214.227337 samples/sec	accuracy=0.693264	lr=0.381477
Epoch[20] Batch [1149]	Speed: 5062.172671 samples/sec	accuracy=0.693392	lr=0.381385
Epoch[20] Batch [1199]	Speed: 5355.588063 samples/sec	accuracy=0.693587	lr=0.381293
Epoch[20] Batch [1249]	Speed: 5026.830248 samples/sec	accuracy=0.693764	lr=0.381201
[Epoch 20] training: accuracy=0.693763
[Epoch 20] speed: 5145 samples/sec	time cost: 258.209666
[Epoch 20] validation: err-top1=0.450713 err-top5=0.210698
Epoch[21] Batch [49]	Speed: 4935.460118 samples/sec	accuracy=0.691342	lr=0.381107
Epoch[21] Batch [99]	Speed: 4981.256527 samples/sec	accuracy=0.692143	lr=0.381014
Epoch[21] Batch [149]	Speed: 4874.605130 samples/sec	accuracy=0.693074	lr=0.380921
Epoch[21] Batch [199]	Speed: 5137.676171 samples/sec	accuracy=0.693472	lr=0.380828
Epoch[21] Batch [249]	Speed: 5230.987931 samples/sec	accuracy=0.693707	lr=0.380734
Epoch[21] Batch [299]	Speed: 5037.917682 samples/sec	accuracy=0.693856	lr=0.380641
Epoch[21] Batch [349]	Speed: 5164.261805 samples/sec	accuracy=0.694013	lr=0.380547
Epoch[21] Batch [399]	Speed: 5377.534461 samples/sec	accuracy=0.694115	lr=0.380453
Epoch[21] Batch [449]	Speed: 4910.487585 samples/sec	accuracy=0.694138	lr=0.380359
Epoch[21] Batch [499]	Speed: 5272.672272 samples/sec	accuracy=0.694188	lr=0.380264
Epoch[21] Batch [549]	Speed: 5423.969059 samples/sec	accuracy=0.694302	lr=0.380169
Epoch[21] Batch [599]	Speed: 5099.107454 samples/sec	accuracy=0.694435	lr=0.380074
Epoch[21] Batch [649]	Speed: 5219.526442 samples/sec	accuracy=0.694745	lr=0.379979
Epoch[21] Batch [699]	Speed: 5070.768390 samples/sec	accuracy=0.694809	lr=0.379884
Epoch[21] Batch [749]	Speed: 5185.714336 samples/sec	accuracy=0.694856	lr=0.379788
Epoch[21] Batch [799]	Speed: 4989.325082 samples/sec	accuracy=0.694918	lr=0.379693
Epoch[21] Batch [849]	Speed: 5149.601432 samples/sec	accuracy=0.694912	lr=0.379597
Epoch[21] Batch [899]	Speed: 4942.862145 samples/sec	accuracy=0.695104	lr=0.379501
Epoch[21] Batch [949]	Speed: 5207.944432 samples/sec	accuracy=0.695234	lr=0.379404
Epoch[21] Batch [999]	Speed: 5054.665412 samples/sec	accuracy=0.695469	lr=0.379307
Epoch[21] Batch [1049]	Speed: 5119.673361 samples/sec	accuracy=0.695606	lr=0.379211
Epoch[21] Batch [1099]	Speed: 5092.949824 samples/sec	accuracy=0.695790	lr=0.379114
Epoch[21] Batch [1149]	Speed: 5039.348159 samples/sec	accuracy=0.695997	lr=0.379016
Epoch[21] Batch [1199]	Speed: 4948.914594 samples/sec	accuracy=0.696202	lr=0.378919
Epoch[21] Batch [1249]	Speed: 5324.484748 samples/sec	accuracy=0.696272	lr=0.378821
[Epoch 21] training: accuracy=0.696274
[Epoch 21] speed: 5103 samples/sec	time cost: 260.454248
[Epoch 21] validation: err-top1=0.453145 err-top5=0.207629
Epoch[22] Batch [49]	Speed: 4701.958272 samples/sec	accuracy=0.694423	lr=0.378721
Epoch[22] Batch [99]	Speed: 4985.445286 samples/sec	accuracy=0.695093	lr=0.378623
Epoch[22] Batch [149]	Speed: 4807.610603 samples/sec	accuracy=0.695703	lr=0.378525
Epoch[22] Batch [199]	Speed: 5416.776511 samples/sec	accuracy=0.696086	lr=0.378426
Epoch[22] Batch [249]	Speed: 5210.447129 samples/sec	accuracy=0.695954	lr=0.378328
Epoch[22] Batch [299]	Speed: 5115.369061 samples/sec	accuracy=0.696127	lr=0.378229
Epoch[22] Batch [349]	Speed: 5239.999663 samples/sec	accuracy=0.695865	lr=0.378129
Epoch[22] Batch [399]	Speed: 5105.286059 samples/sec	accuracy=0.695901	lr=0.378030
Epoch[22] Batch [449]	Speed: 5165.877650 samples/sec	accuracy=0.696133	lr=0.377930
Epoch[22] Batch [499]	Speed: 4913.877656 samples/sec	accuracy=0.696353	lr=0.377831
Epoch[22] Batch [549]	Speed: 5016.169352 samples/sec	accuracy=0.696469	lr=0.377731
Epoch[22] Batch [599]	Speed: 5273.314207 samples/sec	accuracy=0.696576	lr=0.377630
Epoch[22] Batch [649]	Speed: 5144.065466 samples/sec	accuracy=0.696687	lr=0.377530
Epoch[22] Batch [699]	Speed: 5208.890211 samples/sec	accuracy=0.696854	lr=0.377429
Epoch[22] Batch [749]	Speed: 5262.787367 samples/sec	accuracy=0.696992	lr=0.377328
Epoch[22] Batch [799]	Speed: 5182.123274 samples/sec	accuracy=0.697082	lr=0.377227
Epoch[22] Batch [849]	Speed: 5097.077452 samples/sec	accuracy=0.697216	lr=0.377126
Epoch[22] Batch [899]	Speed: 5017.386921 samples/sec	accuracy=0.697415	lr=0.377024
Epoch[22] Batch [949]	Speed: 5308.960224 samples/sec	accuracy=0.697548	lr=0.376923
Epoch[22] Batch [999]	Speed: 4991.869636 samples/sec	accuracy=0.697715	lr=0.376821
Epoch[22] Batch [1049]	Speed: 5534.202386 samples/sec	accuracy=0.697826	lr=0.376719
Epoch[22] Batch [1099]	Speed: 5001.139379 samples/sec	accuracy=0.698029	lr=0.376616
Epoch[22] Batch [1149]	Speed: 5073.654186 samples/sec	accuracy=0.698107	lr=0.376514
Epoch[22] Batch [1199]	Speed: 5009.411813 samples/sec	accuracy=0.698213	lr=0.376411
Epoch[22] Batch [1249]	Speed: 5240.966203 samples/sec	accuracy=0.698316	lr=0.376308
[Epoch 22] training: accuracy=0.698317
[Epoch 22] speed: 5110 samples/sec	time cost: 259.823787
[Epoch 22] validation: err-top1=0.464047 err-top5=0.212492
Epoch[23] Batch [49]	Speed: 4894.715404 samples/sec	accuracy=0.698379	lr=0.376203
Epoch[23] Batch [99]	Speed: 5012.952887 samples/sec	accuracy=0.698722	lr=0.376099
Epoch[23] Batch [149]	Speed: 5178.335512 samples/sec	accuracy=0.698878	lr=0.375996
Epoch[23] Batch [199]	Speed: 5021.372616 samples/sec	accuracy=0.698620	lr=0.375892
Epoch[23] Batch [249]	Speed: 5220.207020 samples/sec	accuracy=0.699011	lr=0.375788
Epoch[23] Batch [299]	Speed: 4841.200970 samples/sec	accuracy=0.698747	lr=0.375683
Epoch[23] Batch [349]	Speed: 4931.056753 samples/sec	accuracy=0.698782	lr=0.375579
Epoch[23] Batch [399]	Speed: 5258.836512 samples/sec	accuracy=0.698637	lr=0.375474
Epoch[23] Batch [449]	Speed: 5169.103416 samples/sec	accuracy=0.698559	lr=0.375369
Epoch[23] Batch [499]	Speed: 5168.070290 samples/sec	accuracy=0.698756	lr=0.375264
Epoch[23] Batch [549]	Speed: 5129.877049 samples/sec	accuracy=0.698849	lr=0.375159
Epoch[23] Batch [599]	Speed: 5393.626732 samples/sec	accuracy=0.698835	lr=0.375053
Epoch[23] Batch [649]	Speed: 5020.752754 samples/sec	accuracy=0.698903	lr=0.374948
Epoch[23] Batch [699]	Speed: 5354.544075 samples/sec	accuracy=0.699033	lr=0.374842
Epoch[23] Batch [749]	Speed: 5358.470155 samples/sec	accuracy=0.699119	lr=0.374736
Epoch[23] Batch [799]	Speed: 5246.413900 samples/sec	accuracy=0.699082	lr=0.374629
Epoch[23] Batch [849]	Speed: 5230.281992 samples/sec	accuracy=0.699204	lr=0.374523
Epoch[23] Batch [899]	Speed: 5001.128664 samples/sec	accuracy=0.699384	lr=0.374416
Epoch[23] Batch [949]	Speed: 5049.583046 samples/sec	accuracy=0.699468	lr=0.374309
Epoch[23] Batch [999]	Speed: 5134.313463 samples/sec	accuracy=0.699590	lr=0.374202
Epoch[23] Batch [1049]	Speed: 5204.989178 samples/sec	accuracy=0.699657	lr=0.374095
Epoch[23] Batch [1099]	Speed: 5331.326392 samples/sec	accuracy=0.699713	lr=0.373987
Epoch[23] Batch [1149]	Speed: 4983.987985 samples/sec	accuracy=0.699895	lr=0.373879
Epoch[23] Batch [1199]	Speed: 4910.613459 samples/sec	accuracy=0.699964	lr=0.373771
Epoch[23] Batch [1249]	Speed: 5008.780530 samples/sec	accuracy=0.700123	lr=0.373663
[Epoch 23] training: accuracy=0.700133
[Epoch 23] speed: 5113 samples/sec	time cost: 259.882211
[Epoch 23] validation: err-top1=0.454325 err-top5=0.209595
Epoch[24] Batch [49]	Speed: 5112.471414 samples/sec	accuracy=0.699936	lr=0.373552
Epoch[24] Batch [99]	Speed: 5029.769902 samples/sec	accuracy=0.700068	lr=0.373444
Epoch[24] Batch [149]	Speed: 4750.462022 samples/sec	accuracy=0.699908	lr=0.373335
Epoch[24] Batch [199]	Speed: 5085.056049 samples/sec	accuracy=0.699363	lr=0.373226
Epoch[24] Batch [249]	Speed: 5222.719487 samples/sec	accuracy=0.699870	lr=0.373117
Epoch[24] Batch [299]	Speed: 5210.071811 samples/sec	accuracy=0.699833	lr=0.373007
Epoch[24] Batch [349]	Speed: 5400.222931 samples/sec	accuracy=0.699737	lr=0.372897
Epoch[24] Batch [399]	Speed: 4940.085560 samples/sec	accuracy=0.699704	lr=0.372788
Epoch[24] Batch [449]	Speed: 4852.053488 samples/sec	accuracy=0.699739	lr=0.372678
Epoch[24] Batch [499]	Speed: 5276.875134 samples/sec	accuracy=0.700213	lr=0.372567
Epoch[24] Batch [549]	Speed: 5267.624719 samples/sec	accuracy=0.700394	lr=0.372457
Epoch[24] Batch [599]	Speed: 5121.299280 samples/sec	accuracy=0.700561	lr=0.372346
Epoch[24] Batch [649]	Speed: 5164.496162 samples/sec	accuracy=0.700698	lr=0.372235
Epoch[24] Batch [699]	Speed: 5003.964051 samples/sec	accuracy=0.700908	lr=0.372124
Epoch[24] Batch [749]	Speed: 5341.070150 samples/sec	accuracy=0.701040	lr=0.372013
Epoch[24] Batch [799]	Speed: 5048.428487 samples/sec	accuracy=0.701255	lr=0.371901
Epoch[24] Batch [849]	Speed: 5265.174595 samples/sec	accuracy=0.701368	lr=0.371790
Epoch[24] Batch [899]	Speed: 5200.608424 samples/sec	accuracy=0.701344	lr=0.371678
Epoch[24] Batch [949]	Speed: 5059.288595 samples/sec	accuracy=0.701446	lr=0.371565
Epoch[24] Batch [999]	Speed: 5207.914247 samples/sec	accuracy=0.701637	lr=0.371453
Epoch[24] Batch [1049]	Speed: 5226.998780 samples/sec	accuracy=0.701784	lr=0.371341
Epoch[24] Batch [1099]	Speed: 5099.655866 samples/sec	accuracy=0.701944	lr=0.371228
Epoch[24] Batch [1149]	Speed: 5155.210584 samples/sec	accuracy=0.702076	lr=0.371115
Epoch[24] Batch [1199]	Speed: 5094.725603 samples/sec	accuracy=0.702240	lr=0.371002
Epoch[24] Batch [1249]	Speed: 5304.574297 samples/sec	accuracy=0.702279	lr=0.370888
[Epoch 24] training: accuracy=0.702282
[Epoch 24] speed: 5128 samples/sec	time cost: 259.553943
[Epoch 24] validation: err-top1=0.442861 err-top5=0.193180
Epoch[25] Batch [49]	Speed: 4659.075844 samples/sec	accuracy=0.701726	lr=0.370773
Epoch[25] Batch [99]	Speed: 5126.259719 samples/sec	accuracy=0.701846	lr=0.370659
Epoch[25] Batch [149]	Speed: 5101.556788 samples/sec	accuracy=0.701356	lr=0.370545
Epoch[25] Batch [199]	Speed: 4990.631828 samples/sec	accuracy=0.701859	lr=0.370431
Epoch[25] Batch [249]	Speed: 5100.338732 samples/sec	accuracy=0.701828	lr=0.370316
Epoch[25] Batch [299]	Speed: 4984.334906 samples/sec	accuracy=0.701843	lr=0.370202
Epoch[25] Batch [349]	Speed: 4936.905972 samples/sec	accuracy=0.702020	lr=0.370087
Epoch[25] Batch [399]	Speed: 5160.118428 samples/sec	accuracy=0.702147	lr=0.369972
Epoch[25] Batch [449]	Speed: 4843.160573 samples/sec	accuracy=0.701943	lr=0.369857
Epoch[25] Batch [499]	Speed: 4990.918894 samples/sec	accuracy=0.702231	lr=0.369741
Epoch[25] Batch [549]	Speed: 5443.704450 samples/sec	accuracy=0.702450	lr=0.369626
Epoch[25] Batch [599]	Speed: 5080.191217 samples/sec	accuracy=0.702643	lr=0.369510
Epoch[25] Batch [649]	Speed: 5253.588162 samples/sec	accuracy=0.702841	lr=0.369394
Epoch[25] Batch [699]	Speed: 5205.228382 samples/sec	accuracy=0.702998	lr=0.369278
Epoch[25] Batch [749]	Speed: 5275.538105 samples/sec	accuracy=0.703195	lr=0.369161
Epoch[25] Batch [799]	Speed: 5074.616687 samples/sec	accuracy=0.703265	lr=0.369045
Epoch[25] Batch [849]	Speed: 5158.245854 samples/sec	accuracy=0.703361	lr=0.368928
Epoch[25] Batch [899]	Speed: 5200.482105 samples/sec	accuracy=0.703336	lr=0.368811
Epoch[25] Batch [949]	Speed: 5245.619479 samples/sec	accuracy=0.703488	lr=0.368694
Epoch[25] Batch [999]	Speed: 5412.306739 samples/sec	accuracy=0.703591	lr=0.368576
Epoch[25] Batch [1049]	Speed: 5365.892027 samples/sec	accuracy=0.703657	lr=0.368459
Epoch[25] Batch [1099]	Speed: 5289.357778 samples/sec	accuracy=0.703802	lr=0.368341
Epoch[25] Batch [1149]	Speed: 4823.600319 samples/sec	accuracy=0.703937	lr=0.368223
Epoch[25] Batch [1199]	Speed: 5092.387033 samples/sec	accuracy=0.704058	lr=0.368105
Epoch[25] Batch [1249]	Speed: 5231.168619 samples/sec	accuracy=0.704155	lr=0.367986
[Epoch 25] training: accuracy=0.704154
[Epoch 25] speed: 5108 samples/sec	time cost: 259.855590
[Epoch 25] validation: err-top1=0.488221 err-top5=0.237085
Epoch[26] Batch [49]	Speed: 4771.039457 samples/sec	accuracy=0.703770	lr=0.367865
Epoch[26] Batch [99]	Speed: 5091.298278 samples/sec	accuracy=0.702530	lr=0.367747
Epoch[26] Batch [149]	Speed: 5141.327239 samples/sec	accuracy=0.702777	lr=0.367628
Epoch[26] Batch [199]	Speed: 5011.831625 samples/sec	accuracy=0.703379	lr=0.367508
Epoch[26] Batch [249]	Speed: 5091.649676 samples/sec	accuracy=0.703557	lr=0.367389
Epoch[26] Batch [299]	Speed: 5049.021845 samples/sec	accuracy=0.703820	lr=0.367269
Epoch[26] Batch [349]	Speed: 5161.795080 samples/sec	accuracy=0.703983	lr=0.367150
Epoch[26] Batch [399]	Speed: 5124.256221 samples/sec	accuracy=0.704171	lr=0.367030
Epoch[26] Batch [449]	Speed: 5129.689567 samples/sec	accuracy=0.704453	lr=0.366909
Epoch[26] Batch [499]	Speed: 5171.612129 samples/sec	accuracy=0.704590	lr=0.366789
Epoch[26] Batch [549]	Speed: 4824.575846 samples/sec	accuracy=0.704691	lr=0.366668
Epoch[26] Batch [599]	Speed: 5139.060809 samples/sec	accuracy=0.704840	lr=0.366548
Epoch[26] Batch [649]	Speed: 5037.909290 samples/sec	accuracy=0.705058	lr=0.366426
Epoch[26] Batch [699]	Speed: 5177.125451 samples/sec	accuracy=0.705300	lr=0.366305
Epoch[26] Batch [749]	Speed: 4934.674405 samples/sec	accuracy=0.705448	lr=0.366184
Epoch[26] Batch [799]	Speed: 5437.230320 samples/sec	accuracy=0.705564	lr=0.366062
Epoch[26] Batch [849]	Speed: 4928.590770 samples/sec	accuracy=0.705656	lr=0.365940
Epoch[26] Batch [899]	Speed: 5194.725051 samples/sec	accuracy=0.705799	lr=0.365819
Epoch[26] Batch [949]	Speed: 5180.142480 samples/sec	accuracy=0.706004	lr=0.365696
Epoch[26] Batch [999]	Speed: 4920.542425 samples/sec	accuracy=0.706175	lr=0.365574
Epoch[26] Batch [1049]	Speed: 5019.489671 samples/sec	accuracy=0.706328	lr=0.365451
Epoch[26] Batch [1099]	Speed: 5088.912927 samples/sec	accuracy=0.706377	lr=0.365329
Epoch[26] Batch [1149]	Speed: 5268.472350 samples/sec	accuracy=0.706482	lr=0.365206
Epoch[26] Batch [1199]	Speed: 5109.452559 samples/sec	accuracy=0.706655	lr=0.365082
Epoch[26] Batch [1249]	Speed: 5469.809663 samples/sec	accuracy=0.706714	lr=0.364959
[Epoch 26] training: accuracy=0.706713
[Epoch 26] speed: 5090 samples/sec	time cost: 260.834884
[Epoch 26] validation: err-top1=0.421516 err-top5=0.182318
Epoch[27] Batch [49]	Speed: 4860.600972 samples/sec	accuracy=0.705162	lr=0.364833
Epoch[27] Batch [99]	Speed: 5119.863285 samples/sec	accuracy=0.705356	lr=0.364709
Epoch[27] Batch [149]	Speed: 4942.459888 samples/sec	accuracy=0.704593	lr=0.364585
Epoch[27] Batch [199]	Speed: 4987.082714 samples/sec	accuracy=0.704975	lr=0.364461
Epoch[27] Batch [249]	Speed: 5109.453532 samples/sec	accuracy=0.705150	lr=0.364337
Epoch[27] Batch [299]	Speed: 5107.626776 samples/sec	accuracy=0.705527	lr=0.364212
Epoch[27] Batch [349]	Speed: 5255.426047 samples/sec	accuracy=0.705977	lr=0.364087
Epoch[27] Batch [399]	Speed: 5025.280919 samples/sec	accuracy=0.705879	lr=0.363962
Epoch[27] Batch [449]	Speed: 5005.420925 samples/sec	accuracy=0.705855	lr=0.363837
Epoch[27] Batch [499]	Speed: 5126.757564 samples/sec	accuracy=0.706059	lr=0.363712
Epoch[27] Batch [549]	Speed: 5143.832344 samples/sec	accuracy=0.706199	lr=0.363586
Epoch[27] Batch [599]	Speed: 5153.700967 samples/sec	accuracy=0.706388	lr=0.363461
Epoch[27] Batch [649]	Speed: 4968.024183 samples/sec	accuracy=0.706543	lr=0.363335
Epoch[27] Batch [699]	Speed: 4838.089806 samples/sec	accuracy=0.706615	lr=0.363209
Epoch[27] Batch [749]	Speed: 5044.260400 samples/sec	accuracy=0.706713	lr=0.363082
Epoch[27] Batch [799]	Speed: 5384.613383 samples/sec	accuracy=0.706833	lr=0.362956
Epoch[27] Batch [849]	Speed: 4791.609365 samples/sec	accuracy=0.707015	lr=0.362829
Epoch[27] Batch [899]	Speed: 5306.790549 samples/sec	accuracy=0.707136	lr=0.362702
Epoch[27] Batch [949]	Speed: 4915.508457 samples/sec	accuracy=0.707297	lr=0.362575
Epoch[27] Batch [999]	Speed: 5145.737866 samples/sec	accuracy=0.707424	lr=0.362448
Epoch[27] Batch [1049]	Speed: 5163.614980 samples/sec	accuracy=0.707535	lr=0.362320
Epoch[27] Batch [1099]	Speed: 5106.033442 samples/sec	accuracy=0.707722	lr=0.362193
Epoch[27] Batch [1149]	Speed: 5139.908902 samples/sec	accuracy=0.707867	lr=0.362065
Epoch[27] Batch [1199]	Speed: 5273.139401 samples/sec	accuracy=0.707993	lr=0.361937
Epoch[27] Batch [1249]	Speed: 5376.042714 samples/sec	accuracy=0.708146	lr=0.361809
[Epoch 27] training: accuracy=0.708149
[Epoch 27] speed: 5082 samples/sec	time cost: 261.053287
[Epoch 27] validation: err-top1=0.456633 err-top5=0.207589
Epoch[28] Batch [49]	Speed: 4933.501744 samples/sec	accuracy=0.706571	lr=0.361678
Epoch[28] Batch [99]	Speed: 5103.620440 samples/sec	accuracy=0.707464	lr=0.361549
Epoch[28] Batch [149]	Speed: 5020.704510 samples/sec	accuracy=0.707202	lr=0.361420
Epoch[28] Batch [199]	Speed: 4997.735347 samples/sec	accuracy=0.707195	lr=0.361291
Epoch[28] Batch [249]	Speed: 5169.213035 samples/sec	accuracy=0.707062	lr=0.361162
Epoch[28] Batch [299]	Speed: 5181.984346 samples/sec	accuracy=0.707155	lr=0.361032
Epoch[28] Batch [349]	Speed: 4785.951557 samples/sec	accuracy=0.707382	lr=0.360903
Epoch[28] Batch [399]	Speed: 5014.058027 samples/sec	accuracy=0.707262	lr=0.360773
Epoch[28] Batch [449]	Speed: 4979.109141 samples/sec	accuracy=0.707208	lr=0.360643
Epoch[28] Batch [499]	Speed: 5291.254545 samples/sec	accuracy=0.707248	lr=0.360513
Epoch[28] Batch [549]	Speed: 5121.653732 samples/sec	accuracy=0.707388	lr=0.360382
Epoch[28] Batch [599]	Speed: 5097.287966 samples/sec	accuracy=0.707495	lr=0.360252
Epoch[28] Batch [649]	Speed: 5080.440602 samples/sec	accuracy=0.707622	lr=0.360121
Epoch[28] Batch [699]	Speed: 5235.996761 samples/sec	accuracy=0.707721	lr=0.359990
Epoch[28] Batch [749]	Speed: 5102.310353 samples/sec	accuracy=0.707915	lr=0.359859
Epoch[28] Batch [799]	Speed: 5105.743179 samples/sec	accuracy=0.708131	lr=0.359728
Epoch[28] Batch [849]	Speed: 5053.714982 samples/sec	accuracy=0.708196	lr=0.359596
Epoch[28] Batch [899]	Speed: 5126.242220 samples/sec	accuracy=0.708354	lr=0.359465
Epoch[28] Batch [949]	Speed: 5214.267598 samples/sec	accuracy=0.708460	lr=0.359333
Epoch[28] Batch [999]	Speed: 5111.228189 samples/sec	accuracy=0.708561	lr=0.359201
Epoch[28] Batch [1049]	Speed: 4756.793535 samples/sec	accuracy=0.708622	lr=0.359068
Epoch[28] Batch [1099]	Speed: 5108.003395 samples/sec	accuracy=0.708765	lr=0.358936
Epoch[28] Batch [1149]	Speed: 5004.658616 samples/sec	accuracy=0.708888	lr=0.358803
Epoch[28] Batch [1199]	Speed: 5074.598580 samples/sec	accuracy=0.709031	lr=0.358670
Epoch[28] Batch [1249]	Speed: 5273.861749 samples/sec	accuracy=0.709068	lr=0.358537
[Epoch 28] training: accuracy=0.709065
[Epoch 28] speed: 5070 samples/sec	time cost: 262.169452
[Epoch 28] validation: err-top1=0.435427 err-top5=0.189792
Epoch[29] Batch [49]	Speed: 4697.801933 samples/sec	accuracy=0.709310	lr=0.358401
Epoch[29] Batch [99]	Speed: 4882.490791 samples/sec	accuracy=0.708291	lr=0.358268
Epoch[29] Batch [149]	Speed: 4718.900583 samples/sec	accuracy=0.708036	lr=0.358134
Epoch[29] Batch [199]	Speed: 5101.467956 samples/sec	accuracy=0.707756	lr=0.358001
Epoch[29] Batch [249]	Speed: 5110.229253 samples/sec	accuracy=0.708131	lr=0.357867
Epoch[29] Batch [299]	Speed: 5183.358317 samples/sec	accuracy=0.708495	lr=0.357732
Epoch[29] Batch [349]	Speed: 5570.754249 samples/sec	accuracy=0.708740	lr=0.357598
Epoch[29] Batch [399]	Speed: 5348.479033 samples/sec	accuracy=0.708837	lr=0.357464
Epoch[29] Batch [449]	Speed: 5168.728681 samples/sec	accuracy=0.709218	lr=0.357329
Epoch[29] Batch [499]	Speed: 5069.033439 samples/sec	accuracy=0.709307	lr=0.357194
Epoch[29] Batch [549]	Speed: 5234.067063 samples/sec	accuracy=0.709378	lr=0.357059
Epoch[29] Batch [599]	Speed: 4814.600806 samples/sec	accuracy=0.709513	lr=0.356924
Epoch[29] Batch [649]	Speed: 5098.531559 samples/sec	accuracy=0.709753	lr=0.356788
Epoch[29] Batch [699]	Speed: 5139.706171 samples/sec	accuracy=0.709872	lr=0.356652
Epoch[29] Batch [749]	Speed: 5151.910418 samples/sec	accuracy=0.709964	lr=0.356517
Epoch[29] Batch [799]	Speed: 5104.241159 samples/sec	accuracy=0.710201	lr=0.356380
Epoch[29] Batch [849]	Speed: 5148.027468 samples/sec	accuracy=0.710358	lr=0.356244
Epoch[29] Batch [899]	Speed: 5146.145530 samples/sec	accuracy=0.710497	lr=0.356108
Epoch[29] Batch [949]	Speed: 5129.470121 samples/sec	accuracy=0.710619	lr=0.355971
Epoch[29] Batch [999]	Speed: 5185.794731 samples/sec	accuracy=0.710676	lr=0.355834
Epoch[29] Batch [1049]	Speed: 5008.591398 samples/sec	accuracy=0.710813	lr=0.355698
Epoch[29] Batch [1099]	Speed: 5247.796478 samples/sec	accuracy=0.710868	lr=0.355560
Epoch[29] Batch [1149]	Speed: 5133.675591 samples/sec	accuracy=0.710925	lr=0.355423
Epoch[29] Batch [1199]	Speed: 4973.706893 samples/sec	accuracy=0.711050	lr=0.355285
Epoch[29] Batch [1249]	Speed: 5309.426060 samples/sec	accuracy=0.711180	lr=0.355148
[Epoch 29] training: accuracy=0.711191
[Epoch 29] speed: 5096 samples/sec	time cost: 260.688846
[Epoch 29] validation: err-top1=0.433370 err-top5=0.192424
Epoch[30] Batch [49]	Speed: 4904.150583 samples/sec	accuracy=0.709635	lr=0.355007
Epoch[30] Batch [99]	Speed: 5021.313206 samples/sec	accuracy=0.710088	lr=0.354869
Epoch[30] Batch [149]	Speed: 5271.299459 samples/sec	accuracy=0.710055	lr=0.354731
Epoch[30] Batch [199]	Speed: 5196.723302 samples/sec	accuracy=0.710056	lr=0.354592
Epoch[30] Batch [249]	Speed: 5009.956295 samples/sec	accuracy=0.710488	lr=0.354454
Epoch[30] Batch [299]	Speed: 5154.953063 samples/sec	accuracy=0.710440	lr=0.354315
Epoch[30] Batch [349]	Speed: 5055.495992 samples/sec	accuracy=0.710584	lr=0.354176
Epoch[30] Batch [399]	Speed: 5105.683698 samples/sec	accuracy=0.710420	lr=0.354037
Epoch[30] Batch [449]	Speed: 4890.885502 samples/sec	accuracy=0.710459	lr=0.353897
Epoch[30] Batch [499]	Speed: 4962.897461 samples/sec	accuracy=0.710504	lr=0.353758
Epoch[30] Batch [549]	Speed: 5328.611981 samples/sec	accuracy=0.710694	lr=0.353618
Epoch[30] Batch [599]	Speed: 5068.845233 samples/sec	accuracy=0.710805	lr=0.353478
Epoch[30] Batch [649]	Speed: 5141.799700 samples/sec	accuracy=0.710835	lr=0.353338
Epoch[30] Batch [699]	Speed: 5073.951242 samples/sec	accuracy=0.711084	lr=0.353198
Epoch[30] Batch [749]	Speed: 5236.109108 samples/sec	accuracy=0.711267	lr=0.353057
Epoch[30] Batch [799]	Speed: 5247.412940 samples/sec	accuracy=0.711411	lr=0.352917
Epoch[30] Batch [849]	Speed: 5089.921038 samples/sec	accuracy=0.711536	lr=0.352776
Epoch[30] Batch [899]	Speed: 5133.374445 samples/sec	accuracy=0.711604	lr=0.352635
Epoch[30] Batch [949]	Speed: 5103.015150 samples/sec	accuracy=0.711639	lr=0.352493
Epoch[30] Batch [999]	Speed: 5119.836187 samples/sec	accuracy=0.711816	lr=0.352352
Epoch[30] Batch [1049]	Speed: 5113.501301 samples/sec	accuracy=0.711931	lr=0.352211
Epoch[30] Batch [1099]	Speed: 5082.807549 samples/sec	accuracy=0.712117	lr=0.352069
Epoch[30] Batch [1149]	Speed: 5162.455597 samples/sec	accuracy=0.712321	lr=0.351927
Epoch[30] Batch [1199]	Speed: 5127.869254 samples/sec	accuracy=0.712423	lr=0.351785
Epoch[30] Batch [1249]	Speed: 5467.763395 samples/sec	accuracy=0.712518	lr=0.351642
[Epoch 30] training: accuracy=0.712522
[Epoch 30] speed: 5114 samples/sec	time cost: 259.924873
[Epoch 30] validation: err-top1=0.473553 err-top5=0.225227
Epoch[31] Batch [49]	Speed: 4928.676285 samples/sec	accuracy=0.711131	lr=0.351497
Epoch[31] Batch [99]	Speed: 4902.871824 samples/sec	accuracy=0.711343	lr=0.351354
Epoch[31] Batch [149]	Speed: 4866.760009 samples/sec	accuracy=0.711393	lr=0.351212
Epoch[31] Batch [199]	Speed: 5343.783709 samples/sec	accuracy=0.711513	lr=0.351069
Epoch[31] Batch [249]	Speed: 5146.022953 samples/sec	accuracy=0.712053	lr=0.350925
Epoch[31] Batch [299]	Speed: 5084.818130 samples/sec	accuracy=0.712153	lr=0.350782
Epoch[31] Batch [349]	Speed: 5214.254937 samples/sec	accuracy=0.712110	lr=0.350639
Epoch[31] Batch [399]	Speed: 5332.319638 samples/sec	accuracy=0.711989	lr=0.350495
Epoch[31] Batch [449]	Speed: 5271.602253 samples/sec	accuracy=0.711935	lr=0.350351
Epoch[31] Batch [499]	Speed: 5132.220260 samples/sec	accuracy=0.711975	lr=0.350207
Epoch[31] Batch [549]	Speed: 5052.161044 samples/sec	accuracy=0.712121	lr=0.350062
Epoch[31] Batch [599]	Speed: 5077.388977 samples/sec	accuracy=0.712242	lr=0.349918
Epoch[31] Batch [649]	Speed: 5258.206079 samples/sec	accuracy=0.712306	lr=0.349773
Epoch[31] Batch [699]	Speed: 5515.038959 samples/sec	accuracy=0.712462	lr=0.349629
Epoch[31] Batch [749]	Speed: 5247.469486 samples/sec	accuracy=0.712657	lr=0.349484
Epoch[31] Batch [799]	Speed: 5294.870431 samples/sec	accuracy=0.712771	lr=0.349338
Epoch[31] Batch [849]	Speed: 5174.111436 samples/sec	accuracy=0.713015	lr=0.349193
Epoch[31] Batch [899]	Speed: 5273.469988 samples/sec	accuracy=0.713137	lr=0.349048
Epoch[31] Batch [949]	Speed: 5109.800267 samples/sec	accuracy=0.713396	lr=0.348902
Epoch[31] Batch [999]	Speed: 5229.242605 samples/sec	accuracy=0.713444	lr=0.348756
Epoch[31] Batch [1049]	Speed: 5167.865207 samples/sec	accuracy=0.713536	lr=0.348610
Epoch[31] Batch [1099]	Speed: 4781.892489 samples/sec	accuracy=0.713630	lr=0.348464
Epoch[31] Batch [1149]	Speed: 5117.907838 samples/sec	accuracy=0.713790	lr=0.348317
Epoch[31] Batch [1199]	Speed: 5182.678808 samples/sec	accuracy=0.713909	lr=0.348171
Epoch[31] Batch [1249]	Speed: 5121.636876 samples/sec	accuracy=0.714018	lr=0.348024
[Epoch 31] training: accuracy=0.714024
[Epoch 31] speed: 5143 samples/sec	time cost: 258.349213
[Epoch 31] validation: err-top1=0.421038 err-top5=0.180126
Epoch[32] Batch [49]	Speed: 4850.208603 samples/sec	accuracy=0.712519	lr=0.347874
Epoch[32] Batch [99]	Speed: 5150.170765 samples/sec	accuracy=0.713021	lr=0.347727
Epoch[32] Batch [149]	Speed: 5344.772290 samples/sec	accuracy=0.712837	lr=0.347580
Epoch[32] Batch [199]	Speed: 4964.739672 samples/sec	accuracy=0.712875	lr=0.347432
Epoch[32] Batch [249]	Speed: 4973.817943 samples/sec	accuracy=0.712980	lr=0.347285
Epoch[32] Batch [299]	Speed: 5130.543273 samples/sec	accuracy=0.713486	lr=0.347137
Epoch[32] Batch [349]	Speed: 5152.561608 samples/sec	accuracy=0.713731	lr=0.346989
Epoch[32] Batch [399]	Speed: 5202.044585 samples/sec	accuracy=0.713882	lr=0.346841
Epoch[32] Batch [449]	Speed: 5250.037021 samples/sec	accuracy=0.713868	lr=0.346692
Epoch[32] Batch [499]	Speed: 4927.147069 samples/sec	accuracy=0.714088	lr=0.346544
Epoch[32] Batch [549]	Speed: 5180.795077 samples/sec	accuracy=0.714242	lr=0.346395
Epoch[32] Batch [599]	Speed: 5299.427583 samples/sec	accuracy=0.714578	lr=0.346246
Epoch[32] Batch [649]	Speed: 5270.773148 samples/sec	accuracy=0.714669	lr=0.346097
Epoch[32] Batch [699]	Speed: 5107.997684 samples/sec	accuracy=0.714952	lr=0.345948
Epoch[32] Batch [749]	Speed: 5147.648996 samples/sec	accuracy=0.715087	lr=0.345799
Epoch[32] Batch [799]	Speed: 5272.557703 samples/sec	accuracy=0.715170	lr=0.345649
Epoch[32] Batch [849]	Speed: 5194.388305 samples/sec	accuracy=0.715142	lr=0.345499
Epoch[32] Batch [899]	Speed: 5357.349931 samples/sec	accuracy=0.715240	lr=0.345349
Epoch[32] Batch [949]	Speed: 5395.754395 samples/sec	accuracy=0.715342	lr=0.345199
Epoch[32] Batch [999]	Speed: 5228.690284 samples/sec	accuracy=0.715342	lr=0.345049
Epoch[32] Batch [1049]	Speed: 5114.497253 samples/sec	accuracy=0.715394	lr=0.344899
Epoch[32] Batch [1099]	Speed: 5117.372199 samples/sec	accuracy=0.715515	lr=0.344748
Epoch[32] Batch [1149]	Speed: 4931.449003 samples/sec	accuracy=0.715669	lr=0.344597
Epoch[32] Batch [1199]	Speed: 5120.772700 samples/sec	accuracy=0.715815	lr=0.344446
Epoch[32] Batch [1249]	Speed: 5081.920587 samples/sec	accuracy=0.715934	lr=0.344295
[Epoch 32] training: accuracy=0.715939
[Epoch 32] speed: 5141 samples/sec	time cost: 258.785280
[Epoch 32] validation: err-top1=0.432996 err-top5=0.187101
Epoch[33] Batch [49]	Speed: 5074.795248 samples/sec	accuracy=0.714619	lr=0.344141
Epoch[33] Batch [99]	Speed: 5036.954754 samples/sec	accuracy=0.713648	lr=0.343989
Epoch[33] Batch [149]	Speed: 4899.801989 samples/sec	accuracy=0.713275	lr=0.343838
Epoch[33] Batch [199]	Speed: 5214.660615 samples/sec	accuracy=0.713591	lr=0.343686
Epoch[33] Batch [249]	Speed: 5281.240034 samples/sec	accuracy=0.713881	lr=0.343534
Epoch[33] Batch [299]	Speed: 5085.675753 samples/sec	accuracy=0.714368	lr=0.343382
Epoch[33] Batch [349]	Speed: 5023.820455 samples/sec	accuracy=0.714571	lr=0.343229
Epoch[33] Batch [399]	Speed: 5284.732475 samples/sec	accuracy=0.714944	lr=0.343077
Epoch[33] Batch [449]	Speed: 4833.669299 samples/sec	accuracy=0.715172	lr=0.342924
Epoch[33] Batch [499]	Speed: 5242.249292 samples/sec	accuracy=0.715186	lr=0.342771
Epoch[33] Batch [549]	Speed: 5181.825920 samples/sec	accuracy=0.715072	lr=0.342618
Epoch[33] Batch [599]	Speed: 5360.540862 samples/sec	accuracy=0.715376	lr=0.342465
Epoch[33] Batch [649]	Speed: 5021.856286 samples/sec	accuracy=0.715389	lr=0.342312
Epoch[33] Batch [699]	Speed: 5242.023948 samples/sec	accuracy=0.715579	lr=0.342158
Epoch[33] Batch [749]	Speed: 5229.572169 samples/sec	accuracy=0.715653	lr=0.342005
Epoch[33] Batch [799]	Speed: 5214.501451 samples/sec	accuracy=0.715733	lr=0.341851
Epoch[33] Batch [849]	Speed: 5231.486191 samples/sec	accuracy=0.715821	lr=0.341697
Epoch[33] Batch [899]	Speed: 5256.692936 samples/sec	accuracy=0.715959	lr=0.341543
Epoch[33] Batch [949]	Speed: 5155.019019 samples/sec	accuracy=0.716174	lr=0.341388
Epoch[33] Batch [999]	Speed: 5155.851715 samples/sec	accuracy=0.716354	lr=0.341234
Epoch[33] Batch [1049]	Speed: 5071.558040 samples/sec	accuracy=0.716532	lr=0.341079
Epoch[33] Batch [1099]	Speed: 4942.750995 samples/sec	accuracy=0.716636	lr=0.340924
Epoch[33] Batch [1149]	Speed: 4981.370456 samples/sec	accuracy=0.716730	lr=0.340769
Epoch[33] Batch [1199]	Speed: 4953.918856 samples/sec	accuracy=0.716839	lr=0.340614
Epoch[33] Batch [1249]	Speed: 5324.424946 samples/sec	accuracy=0.716981	lr=0.340458
[Epoch 33] training: accuracy=0.716978
[Epoch 33] speed: 5123 samples/sec	time cost: 259.201198
[Epoch 33] validation: err-top1=0.439055 err-top5=0.196030
Epoch[34] Batch [49]	Speed: 4927.051658 samples/sec	accuracy=0.715414	lr=0.340300
Epoch[34] Batch [99]	Speed: 5157.224866 samples/sec	accuracy=0.715621	lr=0.340144
Epoch[34] Batch [149]	Speed: 5383.835409 samples/sec	accuracy=0.716278	lr=0.339988
Epoch[34] Batch [199]	Speed: 5004.146536 samples/sec	accuracy=0.716119	lr=0.339832
Epoch[34] Batch [249]	Speed: 4960.625038 samples/sec	accuracy=0.716426	lr=0.339676
Epoch[34] Batch [299]	Speed: 4812.494591 samples/sec	accuracy=0.716438	lr=0.339520
Epoch[34] Batch [349]	Speed: 5074.522076 samples/sec	accuracy=0.716496	lr=0.339363
Epoch[34] Batch [399]	Speed: 5000.937198 samples/sec	accuracy=0.716753	lr=0.339206
Epoch[34] Batch [449]	Speed: 5123.482347 samples/sec	accuracy=0.716717	lr=0.339049
Epoch[34] Batch [499]	Speed: 5046.832015 samples/sec	accuracy=0.716889	lr=0.338892
Epoch[34] Batch [549]	Speed: 5201.037929 samples/sec	accuracy=0.716906	lr=0.338735
Epoch[34] Batch [599]	Speed: 5002.981189 samples/sec	accuracy=0.717031	lr=0.338578
Epoch[34] Batch [649]	Speed: 5005.010871 samples/sec	accuracy=0.717258	lr=0.338420
Epoch[34] Batch [699]	Speed: 5171.940323 samples/sec	accuracy=0.717383	lr=0.338263
Epoch[34] Batch [749]	Speed: 5165.911451 samples/sec	accuracy=0.717596	lr=0.338105
Epoch[34] Batch [799]	Speed: 5069.677848 samples/sec	accuracy=0.717809	lr=0.337947
Epoch[34] Batch [849]	Speed: 5151.095551 samples/sec	accuracy=0.717968	lr=0.337788
Epoch[34] Batch [899]	Speed: 5311.650704 samples/sec	accuracy=0.718073	lr=0.337630
Epoch[34] Batch [949]	Speed: 5375.456797 samples/sec	accuracy=0.718279	lr=0.337472
Epoch[34] Batch [999]	Speed: 5168.309720 samples/sec	accuracy=0.718407	lr=0.337313
Epoch[34] Batch [1049]	Speed: 5178.428165 samples/sec	accuracy=0.718532	lr=0.337154
Epoch[34] Batch [1099]	Speed: 5098.746914 samples/sec	accuracy=0.718546	lr=0.336995
Epoch[34] Batch [1149]	Speed: 5019.173384 samples/sec	accuracy=0.718580	lr=0.336836
Epoch[34] Batch [1199]	Speed: 5014.352126 samples/sec	accuracy=0.718700	lr=0.336677
Epoch[34] Batch [1249]	Speed: 5254.996642 samples/sec	accuracy=0.718765	lr=0.336517
[Epoch 34] training: accuracy=0.718769
[Epoch 34] speed: 5099 samples/sec	time cost: 260.203193
[Epoch 34] validation: err-top1=0.416056 err-top5=0.174845
Epoch[35] Batch [49]	Speed: 4942.434635 samples/sec	accuracy=0.715967	lr=0.336354
Epoch[35] Batch [99]	Speed: 5140.175134 samples/sec	accuracy=0.716289	lr=0.336194
Epoch[35] Batch [149]	Speed: 5003.578251 samples/sec	accuracy=0.716706	lr=0.336034
Epoch[35] Batch [199]	Speed: 5142.802643 samples/sec	accuracy=0.716661	lr=0.335874
Epoch[35] Batch [249]	Speed: 5130.345814 samples/sec	accuracy=0.716552	lr=0.335714
Epoch[35] Batch [299]	Speed: 5244.707581 samples/sec	accuracy=0.716913	lr=0.335553
Epoch[35] Batch [349]	Speed: 5302.281344 samples/sec	accuracy=0.717258	lr=0.335393
Epoch[35] Batch [399]	Speed: 5359.343533 samples/sec	accuracy=0.717360	lr=0.335232
Epoch[35] Batch [449]	Speed: 5108.147011 samples/sec	accuracy=0.717483	lr=0.335071
Epoch[35] Batch [499]	Speed: 5135.088893 samples/sec	accuracy=0.717745	lr=0.334910
Epoch[35] Batch [549]	Speed: 5082.272737 samples/sec	accuracy=0.717751	lr=0.334749
Epoch[35] Batch [599]	Speed: 5060.059053 samples/sec	accuracy=0.717983	lr=0.334587
Epoch[35] Batch [649]	Speed: 5051.409505 samples/sec	accuracy=0.718110	lr=0.334425
Epoch[35] Batch [699]	Speed: 4989.256922 samples/sec	accuracy=0.718196	lr=0.334264
Epoch[35] Batch [749]	Speed: 5288.171849 samples/sec	accuracy=0.718242	lr=0.334102
Epoch[35] Batch [799]	Speed: 5266.672088 samples/sec	accuracy=0.718421	lr=0.333940
Epoch[35] Batch [849]	Speed: 5083.030601 samples/sec	accuracy=0.718586	lr=0.333777
Epoch[35] Batch [899]	Speed: 5031.877740 samples/sec	accuracy=0.718723	lr=0.333615
Epoch[35] Batch [949]	Speed: 5067.588932 samples/sec	accuracy=0.718855	lr=0.333452
Epoch[35] Batch [999]	Speed: 5179.938062 samples/sec	accuracy=0.718940	lr=0.333290
Epoch[35] Batch [1049]	Speed: 5184.909396 samples/sec	accuracy=0.719081	lr=0.333127
Epoch[35] Batch [1099]	Speed: 5248.724460 samples/sec	accuracy=0.719226	lr=0.332964
Epoch[35] Batch [1149]	Speed: 5270.762798 samples/sec	accuracy=0.719363	lr=0.332801
Epoch[35] Batch [1199]	Speed: 5219.966057 samples/sec	accuracy=0.719478	lr=0.332637
Epoch[35] Batch [1249]	Speed: 5276.825084 samples/sec	accuracy=0.719661	lr=0.332474
[Epoch 35] training: accuracy=0.719673
[Epoch 35] speed: 5145 samples/sec	time cost: 258.637944
[Epoch 35] validation: err-top1=0.414640 err-top5=0.174167
Epoch[36] Batch [49]	Speed: 4740.164389 samples/sec	accuracy=0.720107	lr=0.332307
Epoch[36] Batch [99]	Speed: 4888.628118 samples/sec	accuracy=0.720028	lr=0.332143
Epoch[36] Batch [149]	Speed: 5277.434310 samples/sec	accuracy=0.719192	lr=0.331979
Epoch[36] Batch [199]	Speed: 5422.261133 samples/sec	accuracy=0.719195	lr=0.331815
Epoch[36] Batch [249]	Speed: 5018.282456 samples/sec	accuracy=0.719790	lr=0.331650
Epoch[36] Batch [299]	Speed: 5116.771934 samples/sec	accuracy=0.719588	lr=0.331486
Epoch[36] Batch [349]	Speed: 5305.349323 samples/sec	accuracy=0.719588	lr=0.331321
Epoch[36] Batch [399]	Speed: 5018.045820 samples/sec	accuracy=0.719691	lr=0.331157
Epoch[36] Batch [449]	Speed: 5093.783369 samples/sec	accuracy=0.719845	lr=0.330992
Epoch[36] Batch [499]	Speed: 4925.235950 samples/sec	accuracy=0.719901	lr=0.330827
Epoch[36] Batch [549]	Speed: 5132.221363 samples/sec	accuracy=0.719861	lr=0.330661
Epoch[36] Batch [599]	Speed: 5222.510551 samples/sec	accuracy=0.720099	lr=0.330496
Epoch[36] Batch [649]	Speed: 5140.627325 samples/sec	accuracy=0.720131	lr=0.330330
Epoch[36] Batch [699]	Speed: 5088.211899 samples/sec	accuracy=0.720308	lr=0.330165
Epoch[36] Batch [749]	Speed: 5173.117056 samples/sec	accuracy=0.720390	lr=0.329999
Epoch[36] Batch [799]	Speed: 5096.117538 samples/sec	accuracy=0.720391	lr=0.329833
Epoch[36] Batch [849]	Speed: 5222.650136 samples/sec	accuracy=0.720521	lr=0.329667
Epoch[36] Batch [899]	Speed: 5012.640231 samples/sec	accuracy=0.720653	lr=0.329500
Epoch[36] Batch [949]	Speed: 5162.568905 samples/sec	accuracy=0.720810	lr=0.329334
Epoch[36] Batch [999]	Speed: 5397.954156 samples/sec	accuracy=0.720845	lr=0.329167
Epoch[36] Batch [1049]	Speed: 5001.613334 samples/sec	accuracy=0.720881	lr=0.329000
Epoch[36] Batch [1099]	Speed: 5126.806277 samples/sec	accuracy=0.720944	lr=0.328833
Epoch[36] Batch [1149]	Speed: 5293.228345 samples/sec	accuracy=0.721050	lr=0.328666
Epoch[36] Batch [1199]	Speed: 4992.734262 samples/sec	accuracy=0.721130	lr=0.328499
Epoch[36] Batch [1249]	Speed: 5072.751363 samples/sec	accuracy=0.721232	lr=0.328332
[Epoch 36] training: accuracy=0.721236
[Epoch 36] speed: 5108 samples/sec	time cost: 260.145853
[Epoch 36] validation: err-top1=0.463628 err-top5=0.218331
Epoch[37] Batch [49]	Speed: 4864.245734 samples/sec	accuracy=0.721759	lr=0.328161
Epoch[37] Batch [99]	Speed: 5065.198500 samples/sec	accuracy=0.720466	lr=0.327993
Epoch[37] Batch [149]	Speed: 5242.919168 samples/sec	accuracy=0.720995	lr=0.327825
Epoch[37] Batch [199]	Speed: 5100.590219 samples/sec	accuracy=0.721332	lr=0.327657
Epoch[37] Batch [249]	Speed: 5072.502133 samples/sec	accuracy=0.721132	lr=0.327489
Epoch[37] Batch [299]	Speed: 5052.320674 samples/sec	accuracy=0.721014	lr=0.327321
Epoch[37] Batch [349]	Speed: 5006.842810 samples/sec	accuracy=0.721067	lr=0.327152
Epoch[37] Batch [399]	Speed: 5084.382446 samples/sec	accuracy=0.721226	lr=0.326983
Epoch[37] Batch [449]	Speed: 5084.008820 samples/sec	accuracy=0.721205	lr=0.326815
Epoch[37] Batch [499]	Speed: 5240.941005 samples/sec	accuracy=0.720978	lr=0.326646
Epoch[37] Batch [549]	Speed: 4873.171425 samples/sec	accuracy=0.720895	lr=0.326477
Epoch[37] Batch [599]	Speed: 5151.245307 samples/sec	accuracy=0.721037	lr=0.326307
Epoch[37] Batch [649]	Speed: 4951.091485 samples/sec	accuracy=0.721187	lr=0.326138
Epoch[37] Batch [699]	Speed: 5154.654737 samples/sec	accuracy=0.721338	lr=0.325968
Epoch[37] Batch [749]	Speed: 5204.270185 samples/sec	accuracy=0.721492	lr=0.325799
Epoch[37] Batch [799]	Speed: 5174.218026 samples/sec	accuracy=0.721590	lr=0.325629
Epoch[37] Batch [849]	Speed: 5246.904847 samples/sec	accuracy=0.721695	lr=0.325459
Epoch[37] Batch [899]	Speed: 5210.209594 samples/sec	accuracy=0.721761	lr=0.325289
Epoch[37] Batch [949]	Speed: 4970.103582 samples/sec	accuracy=0.721928	lr=0.325118
Epoch[37] Batch [999]	Speed: 5065.031246 samples/sec	accuracy=0.722027	lr=0.324948
Epoch[37] Batch [1049]	Speed: 5233.561934 samples/sec	accuracy=0.722153	lr=0.324777
Epoch[37] Batch [1099]	Speed: 5149.347187 samples/sec	accuracy=0.722294	lr=0.324607
Epoch[37] Batch [1149]	Speed: 5082.108322 samples/sec	accuracy=0.722403	lr=0.324436
Epoch[37] Batch [1199]	Speed: 5050.640490 samples/sec	accuracy=0.722539	lr=0.324265
Epoch[37] Batch [1249]	Speed: 5383.717579 samples/sec	accuracy=0.722649	lr=0.324094
[Epoch 37] training: accuracy=0.722653
[Epoch 37] speed: 5101 samples/sec	time cost: 260.722758
[Epoch 37] validation: err-top1=0.433056 err-top5=0.190031
Epoch[38] Batch [49]	Speed: 4708.074981 samples/sec	accuracy=0.721148	lr=0.323919
Epoch[38] Batch [99]	Speed: 5403.548499 samples/sec	accuracy=0.720930	lr=0.323747
Epoch[38] Batch [149]	Speed: 4877.615212 samples/sec	accuracy=0.721438	lr=0.323576
Epoch[38] Batch [199]	Speed: 4908.107187 samples/sec	accuracy=0.721369	lr=0.323404
Epoch[38] Batch [249]	Speed: 5234.918987 samples/sec	accuracy=0.721579	lr=0.323232
Epoch[38] Batch [299]	Speed: 5133.274685 samples/sec	accuracy=0.721562	lr=0.323060
Epoch[38] Batch [349]	Speed: 5135.441327 samples/sec	accuracy=0.721673	lr=0.322888
Epoch[38] Batch [399]	Speed: 5006.402178 samples/sec	accuracy=0.721734	lr=0.322715
Epoch[38] Batch [449]	Speed: 5109.005472 samples/sec	accuracy=0.721884	lr=0.322543
Epoch[38] Batch [499]	Speed: 4906.717726 samples/sec	accuracy=0.722013	lr=0.322370
Epoch[38] Batch [549]	Speed: 5137.028247 samples/sec	accuracy=0.722309	lr=0.322197
Epoch[38] Batch [599]	Speed: 5228.687738 samples/sec	accuracy=0.722345	lr=0.322025
Epoch[38] Batch [649]	Speed: 5218.416760 samples/sec	accuracy=0.722409	lr=0.321851
Epoch[38] Batch [699]	Speed: 4834.330777 samples/sec	accuracy=0.722495	lr=0.321678
Epoch[38] Batch [749]	Speed: 5299.411498 samples/sec	accuracy=0.722627	lr=0.321505
Epoch[38] Batch [799]	Speed: 5034.248619 samples/sec	accuracy=0.722863	lr=0.321331
Epoch[38] Batch [849]	Speed: 5272.329617 samples/sec	accuracy=0.723062	lr=0.321158
Epoch[38] Batch [899]	Speed: 5038.751634 samples/sec	accuracy=0.723190	lr=0.320984
Epoch[38] Batch [949]	Speed: 5057.463939 samples/sec	accuracy=0.723376	lr=0.320810
Epoch[38] Batch [999]	Speed: 5129.651460 samples/sec	accuracy=0.723556	lr=0.320636
Epoch[38] Batch [1049]	Speed: 5169.330871 samples/sec	accuracy=0.723666	lr=0.320462
Epoch[38] Batch [1099]	Speed: 5270.468639 samples/sec	accuracy=0.723703	lr=0.320287
Epoch[38] Batch [1149]	Speed: 5243.094793 samples/sec	accuracy=0.723815	lr=0.320113
Epoch[38] Batch [1199]	Speed: 5050.033807 samples/sec	accuracy=0.723920	lr=0.319938
Epoch[38] Batch [1249]	Speed: 5170.693533 samples/sec	accuracy=0.723995	lr=0.319763
[Epoch 38] training: accuracy=0.723998
[Epoch 38] speed: 5093 samples/sec	time cost: 260.802869
[Epoch 38] validation: err-top1=0.425542 err-top5=0.183853
Epoch[39] Batch [49]	Speed: 4961.882627 samples/sec	accuracy=0.722511	lr=0.319585
Epoch[39] Batch [99]	Speed: 5303.299549 samples/sec	accuracy=0.722666	lr=0.319410
Epoch[39] Batch [149]	Speed: 5125.972290 samples/sec	accuracy=0.722131	lr=0.319234
Epoch[39] Batch [199]	Speed: 4896.295552 samples/sec	accuracy=0.722532	lr=0.319059
Epoch[39] Batch [249]	Speed: 5276.297668 samples/sec	accuracy=0.722562	lr=0.318883
Epoch[39] Batch [299]	Speed: 5020.623283 samples/sec	accuracy=0.722522	lr=0.318708
Epoch[39] Batch [349]	Speed: 5003.388579 samples/sec	accuracy=0.722572	lr=0.318532
Epoch[39] Batch [399]	Speed: 4918.434214 samples/sec	accuracy=0.722972	lr=0.318356
Epoch[39] Batch [449]	Speed: 5003.041448 samples/sec	accuracy=0.723199	lr=0.318180
Epoch[39] Batch [499]	Speed: 5070.433516 samples/sec	accuracy=0.723322	lr=0.318004
Epoch[39] Batch [549]	Speed: 5190.885726 samples/sec	accuracy=0.723423	lr=0.317827
Epoch[39] Batch [599]	Speed: 5270.880523 samples/sec	accuracy=0.723625	lr=0.317651
Epoch[39] Batch [649]	Speed: 4946.700760 samples/sec	accuracy=0.723782	lr=0.317474
Epoch[39] Batch [699]	Speed: 5075.644934 samples/sec	accuracy=0.723919	lr=0.317297
Epoch[39] Batch [749]	Speed: 5055.607273 samples/sec	accuracy=0.724161	lr=0.317120
Epoch[39] Batch [799]	Speed: 5108.697005 samples/sec	accuracy=0.724303	lr=0.316943
Epoch[39] Batch [849]	Speed: 5210.952484 samples/sec	accuracy=0.724403	lr=0.316766
Epoch[39] Batch [899]	Speed: 5067.270500 samples/sec	accuracy=0.724554	lr=0.316589
Epoch[39] Batch [949]	Speed: 5200.432989 samples/sec	accuracy=0.724609	lr=0.316411
Epoch[39] Batch [999]	Speed: 5174.311155 samples/sec	accuracy=0.724716	lr=0.316233
Epoch[39] Batch [1049]	Speed: 5241.013401 samples/sec	accuracy=0.724866	lr=0.316056
Epoch[39] Batch [1099]	Speed: 5196.108930 samples/sec	accuracy=0.724902	lr=0.315878
Epoch[39] Batch [1149]	Speed: 4976.434919 samples/sec	accuracy=0.724954	lr=0.315700
Epoch[39] Batch [1199]	Speed: 5080.294454 samples/sec	accuracy=0.725091	lr=0.315522
Epoch[39] Batch [1249]	Speed: 5201.651830 samples/sec	accuracy=0.725263	lr=0.315343
[Epoch 39] training: accuracy=0.725264
[Epoch 39] speed: 5096 samples/sec	time cost: 260.501352
[Epoch 39] validation: err-top1=0.409060 err-top5=0.173988
Epoch[40] Batch [49]	Speed: 4935.520577 samples/sec	accuracy=0.724093	lr=0.315161
Epoch[40] Batch [99]	Speed: 5160.606750 samples/sec	accuracy=0.724079	lr=0.314983
Epoch[40] Batch [149]	Speed: 4907.490972 samples/sec	accuracy=0.724364	lr=0.314804
Epoch[40] Batch [199]	Speed: 4914.494014 samples/sec	accuracy=0.724380	lr=0.314625
Epoch[40] Batch [249]	Speed: 5060.168031 samples/sec	accuracy=0.724615	lr=0.314446
Epoch[40] Batch [299]	Speed: 5104.587187 samples/sec	accuracy=0.724862	lr=0.314267
Epoch[40] Batch [349]	Speed: 4992.751674 samples/sec	accuracy=0.724879	lr=0.314087
Epoch[40] Batch [399]	Speed: 5201.148907 samples/sec	accuracy=0.724991	lr=0.313908
Epoch[40] Batch [449]	Speed: 5072.464871 samples/sec	accuracy=0.724793	lr=0.313728
Epoch[40] Batch [499]	Speed: 5134.797895 samples/sec	accuracy=0.724926	lr=0.313549
Epoch[40] Batch [549]	Speed: 5367.881259 samples/sec	accuracy=0.725084	lr=0.313369
Epoch[40] Batch [599]	Speed: 5122.603863 samples/sec	accuracy=0.725031	lr=0.313189
Epoch[40] Batch [649]	Speed: 5089.025683 samples/sec	accuracy=0.725113	lr=0.313009
Epoch[40] Batch [699]	Speed: 5106.949123 samples/sec	accuracy=0.725101	lr=0.312829
Epoch[40] Batch [749]	Speed: 5024.255696 samples/sec	accuracy=0.725250	lr=0.312648
Epoch[40] Batch [799]	Speed: 4961.943161 samples/sec	accuracy=0.725419	lr=0.312468
Epoch[40] Batch [849]	Speed: 5173.410918 samples/sec	accuracy=0.725444	lr=0.312287
Epoch[40] Batch [899]	Speed: 4853.616080 samples/sec	accuracy=0.725507	lr=0.312106
Epoch[40] Batch [949]	Speed: 5206.761272 samples/sec	accuracy=0.725589	lr=0.311925
Epoch[40] Batch [999]	Speed: 5131.702345 samples/sec	accuracy=0.725700	lr=0.311744
Epoch[40] Batch [1049]	Speed: 5260.136355 samples/sec	accuracy=0.725872	lr=0.311563
Epoch[40] Batch [1099]	Speed: 5241.280104 samples/sec	accuracy=0.726011	lr=0.311382
Epoch[40] Batch [1149]	Speed: 5197.609657 samples/sec	accuracy=0.726107	lr=0.311201
Epoch[40] Batch [1199]	Speed: 4972.898707 samples/sec	accuracy=0.726237	lr=0.311019
Epoch[40] Batch [1249]	Speed: 5531.707948 samples/sec	accuracy=0.726281	lr=0.310837
[Epoch 40] training: accuracy=0.726283
[Epoch 40] speed: 5099 samples/sec	time cost: 260.216769
[Epoch 40] validation: err-top1=0.427470 err-top5=0.187866
Epoch[41] Batch [49]	Speed: 4709.844189 samples/sec	accuracy=0.726593	lr=0.310652
Epoch[41] Batch [99]	Speed: 5139.041255 samples/sec	accuracy=0.725525	lr=0.310470
Epoch[41] Batch [149]	Speed: 5176.271271 samples/sec	accuracy=0.726056	lr=0.310288
Epoch[41] Batch [199]	Speed: 5362.005541 samples/sec	accuracy=0.725977	lr=0.310105
Epoch[41] Batch [249]	Speed: 5371.351422 samples/sec	accuracy=0.725912	lr=0.309923
Epoch[41] Batch [299]	Speed: 5237.893004 samples/sec	accuracy=0.725908	lr=0.309741
Epoch[41] Batch [349]	Speed: 5316.010314 samples/sec	accuracy=0.725951	lr=0.309558
Epoch[41] Batch [399]	Speed: 5212.250520 samples/sec	accuracy=0.726024	lr=0.309375
Epoch[41] Batch [449]	Speed: 5048.689837 samples/sec	accuracy=0.726141	lr=0.309192
Epoch[41] Batch [499]	Speed: 5123.841870 samples/sec	accuracy=0.726283	lr=0.309009
Epoch[41] Batch [549]	Speed: 5049.211074 samples/sec	accuracy=0.726268	lr=0.308826
Epoch[41] Batch [599]	Speed: 5096.374778 samples/sec	accuracy=0.726429	lr=0.308643
Epoch[41] Batch [649]	Speed: 5000.280223 samples/sec	accuracy=0.726566	lr=0.308459
Epoch[41] Batch [699]	Speed: 5386.825013 samples/sec	accuracy=0.726636	lr=0.308276
Epoch[41] Batch [749]	Speed: 5218.067807 samples/sec	accuracy=0.726813	lr=0.308092
Epoch[41] Batch [799]	Speed: 5072.309237 samples/sec	accuracy=0.726934	lr=0.307908
Epoch[41] Batch [849]	Speed: 5129.547678 samples/sec	accuracy=0.727018	lr=0.307725
Epoch[41] Batch [899]	Speed: 5075.498582 samples/sec	accuracy=0.727068	lr=0.307540
Epoch[41] Batch [949]	Speed: 4958.945731 samples/sec	accuracy=0.727160	lr=0.307356
Epoch[41] Batch [999]	Speed: 5213.307716 samples/sec	accuracy=0.727238	lr=0.307172
Epoch[41] Batch [1049]	Speed: 5285.246489 samples/sec	accuracy=0.727451	lr=0.306988
Epoch[41] Batch [1099]	Speed: 5209.073040 samples/sec	accuracy=0.727545	lr=0.306803
Epoch[41] Batch [1149]	Speed: 4983.682401 samples/sec	accuracy=0.727680	lr=0.306618
Epoch[41] Batch [1199]	Speed: 5047.742127 samples/sec	accuracy=0.727774	lr=0.306433
Epoch[41] Batch [1249]	Speed: 5041.804562 samples/sec	accuracy=0.727852	lr=0.306249
[Epoch 41] training: accuracy=0.727857
[Epoch 41] speed: 5129 samples/sec	time cost: 259.121663
[Epoch 41] validation: err-top1=0.403819 err-top5=0.167849
Epoch[42] Batch [49]	Speed: 4888.458633 samples/sec	accuracy=0.726547	lr=0.306060
Epoch[42] Batch [99]	Speed: 5251.585375 samples/sec	accuracy=0.727012	lr=0.305875
Epoch[42] Batch [149]	Speed: 4996.447779 samples/sec	accuracy=0.726535	lr=0.305689
Epoch[42] Batch [199]	Speed: 5307.668019 samples/sec	accuracy=0.726590	lr=0.305504
Epoch[42] Batch [249]	Speed: 5263.559905 samples/sec	accuracy=0.726373	lr=0.305318
Epoch[42] Batch [299]	Speed: 5062.399524 samples/sec	accuracy=0.726767	lr=0.305133
Epoch[42] Batch [349]	Speed: 4980.250687 samples/sec	accuracy=0.727131	lr=0.304947
Epoch[42] Batch [399]	Speed: 5097.417428 samples/sec	accuracy=0.727290	lr=0.304761
Epoch[42] Batch [449]	Speed: 5218.404460 samples/sec	accuracy=0.727289	lr=0.304575
Epoch[42] Batch [499]	Speed: 4912.932671 samples/sec	accuracy=0.727313	lr=0.304388
Epoch[42] Batch [549]	Speed: 5094.793049 samples/sec	accuracy=0.727402	lr=0.304202
Epoch[42] Batch [599]	Speed: 5252.355655 samples/sec	accuracy=0.727600	lr=0.304016
Epoch[42] Batch [649]	Speed: 5222.912179 samples/sec	accuracy=0.727655	lr=0.303829
Epoch[42] Batch [699]	Speed: 5126.670789 samples/sec	accuracy=0.727717	lr=0.303642
Epoch[42] Batch [749]	Speed: 5326.287107 samples/sec	accuracy=0.727816	lr=0.303456
Epoch[42] Batch [799]	Speed: 4949.185360 samples/sec	accuracy=0.727922	lr=0.303269
Epoch[42] Batch [849]	Speed: 5183.244970 samples/sec	accuracy=0.727981	lr=0.303082
Epoch[42] Batch [899]	Speed: 5091.696758 samples/sec	accuracy=0.728055	lr=0.302894
Epoch[42] Batch [949]	Speed: 4996.043262 samples/sec	accuracy=0.728191	lr=0.302707
Epoch[42] Batch [999]	Speed: 5155.086090 samples/sec	accuracy=0.728306	lr=0.302520
Epoch[42] Batch [1049]	Speed: 5164.557146 samples/sec	accuracy=0.728411	lr=0.302332
Epoch[42] Batch [1099]	Speed: 4863.145620 samples/sec	accuracy=0.728402	lr=0.302144
Epoch[42] Batch [1149]	Speed: 5032.561679 samples/sec	accuracy=0.728599	lr=0.301957
Epoch[42] Batch [1199]	Speed: 5193.205518 samples/sec	accuracy=0.728730	lr=0.301769
Epoch[42] Batch [1249]	Speed: 5111.815472 samples/sec	accuracy=0.728852	lr=0.301581
[Epoch 42] training: accuracy=0.728855
[Epoch 42] speed: 5102 samples/sec	time cost: 260.627993
[Epoch 42] validation: err-top1=0.410654 err-top5=0.174984
Epoch[43] Batch [49]	Speed: 5012.014685 samples/sec	accuracy=0.728656	lr=0.301389
Epoch[43] Batch [99]	Speed: 5095.260017 samples/sec	accuracy=0.729084	lr=0.301200
Epoch[43] Batch [149]	Speed: 4919.403855 samples/sec	accuracy=0.728511	lr=0.301012
Epoch[43] Batch [199]	Speed: 5283.663666 samples/sec	accuracy=0.728139	lr=0.300823
Epoch[43] Batch [249]	Speed: 5070.289858 samples/sec	accuracy=0.728210	lr=0.300635
Epoch[43] Batch [299]	Speed: 5015.157680 samples/sec	accuracy=0.727952	lr=0.300446
Epoch[43] Batch [349]	Speed: 5329.326331 samples/sec	accuracy=0.728015	lr=0.300257
Epoch[43] Batch [399]	Speed: 5044.189428 samples/sec	accuracy=0.728117	lr=0.300068
Epoch[43] Batch [449]	Speed: 5481.916031 samples/sec	accuracy=0.728404	lr=0.299879
Epoch[43] Batch [499]	Speed: 5255.218344 samples/sec	accuracy=0.728621	lr=0.299690
Epoch[43] Batch [549]	Speed: 5044.211229 samples/sec	accuracy=0.728865	lr=0.299500
Epoch[43] Batch [599]	Speed: 5161.587393 samples/sec	accuracy=0.728999	lr=0.299311
Epoch[43] Batch [649]	Speed: 5164.873513 samples/sec	accuracy=0.729174	lr=0.299121
Epoch[43] Batch [699]	Speed: 5268.950629 samples/sec	accuracy=0.729383	lr=0.298932
Epoch[43] Batch [749]	Speed: 5200.631094 samples/sec	accuracy=0.729446	lr=0.298742
Epoch[43] Batch [799]	Speed: 5207.981565 samples/sec	accuracy=0.729630	lr=0.298552
Epoch[43] Batch [849]	Speed: 5223.288460 samples/sec	accuracy=0.729687	lr=0.298362
Epoch[43] Batch [899]	Speed: 5271.851501 samples/sec	accuracy=0.729685	lr=0.298171
Epoch[43] Batch [949]	Speed: 5121.642495 samples/sec	accuracy=0.729786	lr=0.297981
Epoch[43] Batch [999]	Speed: 5278.427814 samples/sec	accuracy=0.730001	lr=0.297791
Epoch[43] Batch [1049]	Speed: 5162.394042 samples/sec	accuracy=0.730168	lr=0.297600
Epoch[43] Batch [1099]	Speed: 5256.021079 samples/sec	accuracy=0.730275	lr=0.297409
Epoch[43] Batch [1149]	Speed: 5036.277416 samples/sec	accuracy=0.730356	lr=0.297219
Epoch[43] Batch [1199]	Speed: 5433.197782 samples/sec	accuracy=0.730395	lr=0.297028
Epoch[43] Batch [1249]	Speed: 5181.086562 samples/sec	accuracy=0.730417	lr=0.296837
[Epoch 43] training: accuracy=0.730418
[Epoch 43] speed: 5172 samples/sec	time cost: 256.895748
[Epoch 43] validation: err-top1=0.440330 err-top5=0.199239
Epoch[44] Batch [49]	Speed: 4916.583764 samples/sec	accuracy=0.729742	lr=0.296642
Epoch[44] Batch [99]	Speed: 5030.013065 samples/sec	accuracy=0.728998	lr=0.296451
Epoch[44] Batch [149]	Speed: 5069.760669 samples/sec	accuracy=0.728686	lr=0.296259
Epoch[44] Batch [199]	Speed: 5323.714549 samples/sec	accuracy=0.728821	lr=0.296068
Epoch[44] Batch [249]	Speed: 5324.011910 samples/sec	accuracy=0.728900	lr=0.295876
Epoch[44] Batch [299]	Speed: 5044.053177 samples/sec	accuracy=0.728843	lr=0.295684
Epoch[44] Batch [349]	Speed: 5034.783286 samples/sec	accuracy=0.728880	lr=0.295493
Epoch[44] Batch [399]	Speed: 4899.833740 samples/sec	accuracy=0.729069	lr=0.295301
Epoch[44] Batch [449]	Speed: 4962.275554 samples/sec	accuracy=0.729098	lr=0.295109
Epoch[44] Batch [499]	Speed: 5305.197550 samples/sec	accuracy=0.729294	lr=0.294917
Epoch[44] Batch [549]	Speed: 4900.066402 samples/sec	accuracy=0.729462	lr=0.294724
Epoch[44] Batch [599]	Speed: 5263.388325 samples/sec	accuracy=0.729691	lr=0.294532
Epoch[44] Batch [649]	Speed: 4983.180733 samples/sec	accuracy=0.729807	lr=0.294339
Epoch[44] Batch [699]	Speed: 5067.428455 samples/sec	accuracy=0.730105	lr=0.294147
Epoch[44] Batch [749]	Speed: 5184.233110 samples/sec	accuracy=0.730252	lr=0.293954
Epoch[44] Batch [799]	Speed: 4999.159851 samples/sec	accuracy=0.730347	lr=0.293761
Epoch[44] Batch [849]	Speed: 5353.490351 samples/sec	accuracy=0.730532	lr=0.293568
Epoch[44] Batch [899]	Speed: 5023.830797 samples/sec	accuracy=0.730660	lr=0.293375
Epoch[44] Batch [949]	Speed: 5086.498365 samples/sec	accuracy=0.730789	lr=0.293182
Epoch[44] Batch [999]	Speed: 5187.024133 samples/sec	accuracy=0.730822	lr=0.292989
Epoch[44] Batch [1049]	Speed: 5179.540392 samples/sec	accuracy=0.730865	lr=0.292795
Epoch[44] Batch [1099]	Speed: 5182.313107 samples/sec	accuracy=0.730958	lr=0.292602
Epoch[44] Batch [1149]	Speed: 5299.862449 samples/sec	accuracy=0.731104	lr=0.292408
Epoch[44] Batch [1199]	Speed: 5238.699017 samples/sec	accuracy=0.731189	lr=0.292215
Epoch[44] Batch [1249]	Speed: 5266.674284 samples/sec	accuracy=0.731317	lr=0.292021
[Epoch 44] training: accuracy=0.731316
[Epoch 44] speed: 5116 samples/sec	time cost: 259.521734
[Epoch 44] validation: err-top1=0.393395 err-top5=0.159837
Epoch[45] Batch [49]	Speed: 5097.572671 samples/sec	accuracy=0.730222	lr=0.291823
Epoch[45] Batch [99]	Speed: 4946.393466 samples/sec	accuracy=0.730566	lr=0.291629
Epoch[45] Batch [149]	Speed: 5033.077704 samples/sec	accuracy=0.730574	lr=0.291435
Epoch[45] Batch [199]	Speed: 4982.155276 samples/sec	accuracy=0.730031	lr=0.291240
Epoch[45] Batch [249]	Speed: 5165.873549 samples/sec	accuracy=0.730200	lr=0.291046
Epoch[45] Batch [299]	Speed: 5149.159638 samples/sec	accuracy=0.730358	lr=0.290852
Epoch[45] Batch [349]	Speed: 5233.647008 samples/sec	accuracy=0.730122	lr=0.290657
Epoch[45] Batch [399]	Speed: 5263.440830 samples/sec	accuracy=0.730465	lr=0.290462
Epoch[45] Batch [449]	Speed: 5031.548808 samples/sec	accuracy=0.730468	lr=0.290267
Epoch[45] Batch [499]	Speed: 4979.277695 samples/sec	accuracy=0.730685	lr=0.290073
Epoch[45] Batch [549]	Speed: 5096.364498 samples/sec	accuracy=0.730874	lr=0.289878
Epoch[45] Batch [599]	Speed: 5086.954656 samples/sec	accuracy=0.731185	lr=0.289682
Epoch[45] Batch [649]	Speed: 5095.610391 samples/sec	accuracy=0.731337	lr=0.289487
Epoch[45] Batch [699]	Speed: 5224.120993 samples/sec	accuracy=0.731516	lr=0.289292
Epoch[45] Batch [749]	Speed: 5189.682962 samples/sec	accuracy=0.731563	lr=0.289096
Epoch[45] Batch [799]	Speed: 5097.263163 samples/sec	accuracy=0.731685	lr=0.288901
Epoch[45] Batch [849]	Speed: 5098.089891 samples/sec	accuracy=0.731808	lr=0.288705
Epoch[45] Batch [899]	Speed: 5133.617543 samples/sec	accuracy=0.731905	lr=0.288509
Epoch[45] Batch [949]	Speed: 5178.795441 samples/sec	accuracy=0.731948	lr=0.288314
Epoch[45] Batch [999]	Speed: 5278.037062 samples/sec	accuracy=0.732110	lr=0.288118
Epoch[45] Batch [1049]	Speed: 5070.746240 samples/sec	accuracy=0.732167	lr=0.287921
Epoch[45] Batch [1099]	Speed: 5035.156086 samples/sec	accuracy=0.732271	lr=0.287725
Epoch[45] Batch [1149]	Speed: 5495.793013 samples/sec	accuracy=0.732415	lr=0.287529
Epoch[45] Batch [1199]	Speed: 5137.823550 samples/sec	accuracy=0.732493	lr=0.287333
Epoch[45] Batch [1249]	Speed: 5169.626294 samples/sec	accuracy=0.732629	lr=0.287136
[Epoch 45] training: accuracy=0.732633
[Epoch 45] speed: 5124 samples/sec	time cost: 259.203217
[Epoch 45] validation: err-top1=0.396385 err-top5=0.165936
Epoch[46] Batch [49]	Speed: 5078.371151 samples/sec	accuracy=0.732326	lr=0.286936
Epoch[46] Batch [99]	Speed: 4859.006285 samples/sec	accuracy=0.732253	lr=0.286739
Epoch[46] Batch [149]	Speed: 5144.545085 samples/sec	accuracy=0.732281	lr=0.286542
Epoch[46] Batch [199]	Speed: 5065.141991 samples/sec	accuracy=0.732490	lr=0.286345
Epoch[46] Batch [249]	Speed: 5324.097179 samples/sec	accuracy=0.732343	lr=0.286148
Epoch[46] Batch [299]	Speed: 4942.529619 samples/sec	accuracy=0.732241	lr=0.285951
Epoch[46] Batch [349]	Speed: 5122.949330 samples/sec	accuracy=0.732191	lr=0.285754
Epoch[46] Batch [399]	Speed: 5314.210957 samples/sec	accuracy=0.732225	lr=0.285556
Epoch[46] Batch [449]	Speed: 5185.619919 samples/sec	accuracy=0.732290	lr=0.285359
Epoch[46] Batch [499]	Speed: 5255.691904 samples/sec	accuracy=0.732137	lr=0.285161
Epoch[46] Batch [549]	Speed: 5208.469514 samples/sec	accuracy=0.732294	lr=0.284964
Epoch[46] Batch [599]	Speed: 4991.624346 samples/sec	accuracy=0.732323	lr=0.284766
Epoch[46] Batch [649]	Speed: 5337.176680 samples/sec	accuracy=0.732461	lr=0.284568
Epoch[46] Batch [699]	Speed: 5350.921720 samples/sec	accuracy=0.732662	lr=0.284370
Epoch[46] Batch [749]	Speed: 5245.897928 samples/sec	accuracy=0.732871	lr=0.284172
Epoch[46] Batch [799]	Speed: 4990.500543 samples/sec	accuracy=0.732991	lr=0.283974
Epoch[46] Batch [849]	Speed: 5400.584721 samples/sec	accuracy=0.733105	lr=0.283776
Epoch[46] Batch [899]	Speed: 4865.563178 samples/sec	accuracy=0.733208	lr=0.283577
Epoch[46] Batch [949]	Speed: 5046.688269 samples/sec	accuracy=0.733333	lr=0.283379
Epoch[46] Batch [999]	Speed: 5024.201037 samples/sec	accuracy=0.733450	lr=0.283181
Epoch[46] Batch [1049]	Speed: 5177.933843 samples/sec	accuracy=0.733461	lr=0.282982
Epoch[46] Batch [1099]	Speed: 5206.251428 samples/sec	accuracy=0.733556	lr=0.282783
Epoch[46] Batch [1149]	Speed: 5004.655234 samples/sec	accuracy=0.733683	lr=0.282584
Epoch[46] Batch [1199]	Speed: 5062.365036 samples/sec	accuracy=0.733766	lr=0.282385
Epoch[46] Batch [1249]	Speed: 5101.554365 samples/sec	accuracy=0.733781	lr=0.282186
[Epoch 46] training: accuracy=0.733780
[Epoch 46] speed: 5123 samples/sec	time cost: 259.342217
[Epoch 46] validation: err-top1=0.396423 err-top5=0.163045
Epoch[47] Batch [49]	Speed: 5017.971128 samples/sec	accuracy=0.732832	lr=0.281983
Epoch[47] Batch [99]	Speed: 5104.613760 samples/sec	accuracy=0.732445	lr=0.281784
Epoch[47] Batch [149]	Speed: 5119.484915 samples/sec	accuracy=0.732507	lr=0.281585
Epoch[47] Batch [199]	Speed: 5145.585964 samples/sec	accuracy=0.732686	lr=0.281385
Epoch[47] Batch [249]	Speed: 5143.463974 samples/sec	accuracy=0.732753	lr=0.281186
Epoch[47] Batch [299]	Speed: 5238.802917 samples/sec	accuracy=0.732697	lr=0.280986
Epoch[47] Batch [349]	Speed: 5088.325709 samples/sec	accuracy=0.732964	lr=0.280786
Epoch[47] Batch [399]	Speed: 5429.275951 samples/sec	accuracy=0.732891	lr=0.280587
Epoch[47] Batch [449]	Speed: 4792.225694 samples/sec	accuracy=0.732962	lr=0.280387
Epoch[47] Batch [499]	Speed: 4925.296158 samples/sec	accuracy=0.732988	lr=0.280187
Epoch[47] Batch [549]	Speed: 4969.358317 samples/sec	accuracy=0.733043	lr=0.279987
Epoch[47] Batch [599]	Speed: 5026.331855 samples/sec	accuracy=0.733255	lr=0.279786
Epoch[47] Batch [649]	Speed: 5161.290283 samples/sec	accuracy=0.733318	lr=0.279586
Epoch[47] Batch [699]	Speed: 5074.154574 samples/sec	accuracy=0.733488	lr=0.279386
Epoch[47] Batch [749]	Speed: 5217.695701 samples/sec	accuracy=0.733696	lr=0.279185
Epoch[47] Batch [799]	Speed: 5048.204426 samples/sec	accuracy=0.733912	lr=0.278985
Epoch[47] Batch [849]	Speed: 5183.580772 samples/sec	accuracy=0.733986	lr=0.278784
Epoch[47] Batch [899]	Speed: 5080.157928 samples/sec	accuracy=0.734171	lr=0.278583
Epoch[47] Batch [949]	Speed: 5249.449373 samples/sec	accuracy=0.734296	lr=0.278382
Epoch[47] Batch [999]	Speed: 4892.882198 samples/sec	accuracy=0.734361	lr=0.278181
Epoch[47] Batch [1049]	Speed: 5038.411991 samples/sec	accuracy=0.734463	lr=0.277980
Epoch[47] Batch [1099]	Speed: 5242.808065 samples/sec	accuracy=0.734574	lr=0.277779
Epoch[47] Batch [1149]	Speed: 5028.986498 samples/sec	accuracy=0.734694	lr=0.277578
Epoch[47] Batch [1199]	Speed: 5270.585316 samples/sec	accuracy=0.734922	lr=0.277377
Epoch[47] Batch [1249]	Speed: 5296.306622 samples/sec	accuracy=0.735086	lr=0.277175
[Epoch 47] training: accuracy=0.735089
[Epoch 47] speed: 5103 samples/sec	time cost: 260.543786
[Epoch 47] validation: err-top1=0.435507 err-top5=0.195412
Epoch[48] Batch [49]	Speed: 5075.619382 samples/sec	accuracy=0.734469	lr=0.276970
Epoch[48] Batch [99]	Speed: 5010.040683 samples/sec	accuracy=0.734945	lr=0.276768
Epoch[48] Batch [149]	Speed: 5187.463929 samples/sec	accuracy=0.734564	lr=0.276566
Epoch[48] Batch [199]	Speed: 4964.110073 samples/sec	accuracy=0.734390	lr=0.276365
Epoch[48] Batch [249]	Speed: 5259.341508 samples/sec	accuracy=0.734021	lr=0.276163
Epoch[48] Batch [299]	Speed: 4864.346551 samples/sec	accuracy=0.734222	lr=0.275961
Epoch[48] Batch [349]	Speed: 5167.433950 samples/sec	accuracy=0.734453	lr=0.275759
Epoch[48] Batch [399]	Speed: 5163.046892 samples/sec	accuracy=0.734836	lr=0.275557
Epoch[48] Batch [449]	Speed: 5101.321080 samples/sec	accuracy=0.735051	lr=0.275354
Epoch[48] Batch [499]	Speed: 5184.500950 samples/sec	accuracy=0.735056	lr=0.275152
Epoch[48] Batch [549]	Speed: 5293.545277 samples/sec	accuracy=0.735328	lr=0.274950
Epoch[48] Batch [599]	Speed: 5045.836749 samples/sec	accuracy=0.735564	lr=0.274747
Epoch[48] Batch [649]	Speed: 5089.837074 samples/sec	accuracy=0.735633	lr=0.274545
Epoch[48] Batch [699]	Speed: 5178.533060 samples/sec	accuracy=0.735626	lr=0.274342
Epoch[48] Batch [749]	Speed: 5138.858636 samples/sec	accuracy=0.735647	lr=0.274139
Epoch[48] Batch [799]	Speed: 5014.915483 samples/sec	accuracy=0.735761	lr=0.273936
Epoch[48] Batch [849]	Speed: 5191.400973 samples/sec	accuracy=0.735784	lr=0.273733
Epoch[48] Batch [899]	Speed: 4985.645753 samples/sec	accuracy=0.735888	lr=0.273530
Epoch[48] Batch [949]	Speed: 5157.382782 samples/sec	accuracy=0.736061	lr=0.273327
Epoch[48] Batch [999]	Speed: 5392.714384 samples/sec	accuracy=0.736224	lr=0.273124
Epoch[48] Batch [1049]	Speed: 5035.652334 samples/sec	accuracy=0.736267	lr=0.272921
Epoch[48] Batch [1099]	Speed: 5141.805978 samples/sec	accuracy=0.736301	lr=0.272717
Epoch[48] Batch [1149]	Speed: 5281.082495 samples/sec	accuracy=0.736399	lr=0.272514
Epoch[48] Batch [1199]	Speed: 4972.089747 samples/sec	accuracy=0.736511	lr=0.272310
Epoch[48] Batch [1249]	Speed: 5432.976478 samples/sec	accuracy=0.736611	lr=0.272107
[Epoch 48] training: accuracy=0.736615
[Epoch 48] speed: 5125 samples/sec	time cost: 259.205418
[Epoch 48] validation: err-top1=0.404118 err-top5=0.169802
Epoch[49] Batch [49]	Speed: 4715.088882 samples/sec	accuracy=0.735590	lr=0.271899
Epoch[49] Batch [99]	Speed: 5116.092828 samples/sec	accuracy=0.736289	lr=0.271695
Epoch[49] Batch [149]	Speed: 5324.104307 samples/sec	accuracy=0.735872	lr=0.271491
Epoch[49] Batch [199]	Speed: 5100.562235 samples/sec	accuracy=0.735830	lr=0.271287
Epoch[49] Batch [249]	Speed: 5158.912155 samples/sec	accuracy=0.735466	lr=0.271083
Epoch[49] Batch [299]	Speed: 5399.537374 samples/sec	accuracy=0.735680	lr=0.270879
Epoch[49] Batch [349]	Speed: 5103.465314 samples/sec	accuracy=0.735738	lr=0.270675
Epoch[49] Batch [399]	Speed: 5156.480624 samples/sec	accuracy=0.735745	lr=0.270470
Epoch[49] Batch [449]	Speed: 5191.487317 samples/sec	accuracy=0.735925	lr=0.270266
Epoch[49] Batch [499]	Speed: 5116.051875 samples/sec	accuracy=0.736013	lr=0.270061
Epoch[49] Batch [549]	Speed: 5108.617038 samples/sec	accuracy=0.736130	lr=0.269857
Epoch[49] Batch [599]	Speed: 5107.081141 samples/sec	accuracy=0.736339	lr=0.269652
Epoch[49] Batch [649]	Speed: 5372.002829 samples/sec	accuracy=0.736502	lr=0.269447
Epoch[49] Batch [699]	Speed: 5206.228204 samples/sec	accuracy=0.736644	lr=0.269243
Epoch[49] Batch [749]	Speed: 4995.449391 samples/sec	accuracy=0.736638	lr=0.269038
Epoch[49] Batch [799]	Speed: 4965.448879 samples/sec	accuracy=0.736741	lr=0.268833
Epoch[49] Batch [849]	Speed: 5188.402784 samples/sec	accuracy=0.736881	lr=0.268628
Epoch[49] Batch [899]	Speed: 5125.880893 samples/sec	accuracy=0.736918	lr=0.268422
Epoch[49] Batch [949]	Speed: 5191.090884 samples/sec	accuracy=0.736991	lr=0.268217
Epoch[49] Batch [999]	Speed: 5156.234366 samples/sec	accuracy=0.737025	lr=0.268012
Epoch[49] Batch [1049]	Speed: 5115.917077 samples/sec	accuracy=0.737075	lr=0.267807
Epoch[49] Batch [1099]	Speed: 5104.099704 samples/sec	accuracy=0.737112	lr=0.267601
Epoch[49] Batch [1149]	Speed: 5388.799778 samples/sec	accuracy=0.737185	lr=0.267396
Epoch[49] Batch [1199]	Speed: 5056.036492 samples/sec	accuracy=0.737338	lr=0.267190
Epoch[49] Batch [1249]	Speed: 5352.954171 samples/sec	accuracy=0.737462	lr=0.266984
[Epoch 49] training: accuracy=0.737463
[Epoch 49] speed: 5143 samples/sec	time cost: 258.195323
[Epoch 49] validation: err-top1=0.425602 err-top5=0.186922
Epoch[50] Batch [49]	Speed: 4840.183253 samples/sec	accuracy=0.738261	lr=0.266774
Epoch[50] Batch [99]	Speed: 5160.617663 samples/sec	accuracy=0.737219	lr=0.266568
Epoch[50] Batch [149]	Speed: 5397.755522 samples/sec	accuracy=0.736954	lr=0.266362
Epoch[50] Batch [199]	Speed: 4873.168660 samples/sec	accuracy=0.736986	lr=0.266156
Epoch[50] Batch [249]	Speed: 5075.838205 samples/sec	accuracy=0.736936	lr=0.265950
Epoch[50] Batch [299]	Speed: 5239.137905 samples/sec	accuracy=0.737234	lr=0.265744
Epoch[50] Batch [349]	Speed: 5258.035491 samples/sec	accuracy=0.737237	lr=0.265538
Epoch[50] Batch [399]	Speed: 5143.366901 samples/sec	accuracy=0.737386	lr=0.265331
Epoch[50] Batch [449]	Speed: 5073.206749 samples/sec	accuracy=0.737433	lr=0.265125
Epoch[50] Batch [499]	Speed: 5161.622875 samples/sec	accuracy=0.737671	lr=0.264918
Epoch[50] Batch [549]	Speed: 5277.772701 samples/sec	accuracy=0.737693	lr=0.264712
Epoch[50] Batch [599]	Speed: 5133.829490 samples/sec	accuracy=0.737776	lr=0.264505
Epoch[50] Batch [649]	Speed: 5008.222522 samples/sec	accuracy=0.737736	lr=0.264298
Epoch[50] Batch [699]	Speed: 5034.169903 samples/sec	accuracy=0.737784	lr=0.264092
Epoch[50] Batch [749]	Speed: 5104.449838 samples/sec	accuracy=0.737922	lr=0.263885
Epoch[50] Batch [799]	Speed: 5374.656581 samples/sec	accuracy=0.737961	lr=0.263678
Epoch[50] Batch [849]	Speed: 5191.452553 samples/sec	accuracy=0.738096	lr=0.263471
Epoch[50] Batch [899]	Speed: 5235.328524 samples/sec	accuracy=0.738166	lr=0.263264
Epoch[50] Batch [949]	Speed: 5280.904316 samples/sec	accuracy=0.738342	lr=0.263056
Epoch[50] Batch [999]	Speed: 5216.999049 samples/sec	accuracy=0.738524	lr=0.262849
Epoch[50] Batch [1049]	Speed: 4880.179707 samples/sec	accuracy=0.738612	lr=0.262642
Epoch[50] Batch [1099]	Speed: 4977.332275 samples/sec	accuracy=0.738657	lr=0.262434
Epoch[50] Batch [1149]	Speed: 5187.420447 samples/sec	accuracy=0.738727	lr=0.262227
Epoch[50] Batch [1199]	Speed: 5170.087291 samples/sec	accuracy=0.738826	lr=0.262019
Epoch[50] Batch [1249]	Speed: 5274.734451 samples/sec	accuracy=0.738941	lr=0.261812
[Epoch 50] training: accuracy=0.738951
[Epoch 50] speed: 5134 samples/sec	time cost: 258.744357
[Epoch 50] validation: err-top1=0.405513 err-top5=0.174047
Epoch[51] Batch [49]	Speed: 5008.128500 samples/sec	accuracy=0.737609	lr=0.261600
Epoch[51] Batch [99]	Speed: 5311.503037 samples/sec	accuracy=0.737674	lr=0.261392
Epoch[51] Batch [149]	Speed: 5023.019047 samples/sec	accuracy=0.738381	lr=0.261184
Epoch[51] Batch [199]	Speed: 4771.239483 samples/sec	accuracy=0.738447	lr=0.260976
Epoch[51] Batch [249]	Speed: 5318.772626 samples/sec	accuracy=0.738451	lr=0.260768
Epoch[51] Batch [299]	Speed: 4930.845820 samples/sec	accuracy=0.738133	lr=0.260560
Epoch[51] Batch [349]	Speed: 5137.309174 samples/sec	accuracy=0.738147	lr=0.260352
Epoch[51] Batch [399]	Speed: 5223.721212 samples/sec	accuracy=0.738318	lr=0.260144
Epoch[51] Batch [449]	Speed: 5274.641947 samples/sec	accuracy=0.738453	lr=0.259935
Epoch[51] Batch [499]	Speed: 4988.917661 samples/sec	accuracy=0.738299	lr=0.259727
Epoch[51] Batch [549]	Speed: 5160.688848 samples/sec	accuracy=0.738440	lr=0.259519
Epoch[51] Batch [599]	Speed: 5201.666697 samples/sec	accuracy=0.738712	lr=0.259310
Epoch[51] Batch [649]	Speed: 5164.982829 samples/sec	accuracy=0.738754	lr=0.259102
Epoch[51] Batch [699]	Speed: 5245.193340 samples/sec	accuracy=0.738773	lr=0.258893
Epoch[51] Batch [749]	Speed: 5182.684561 samples/sec	accuracy=0.738828	lr=0.258684
Epoch[51] Batch [799]	Speed: 5281.876782 samples/sec	accuracy=0.738812	lr=0.258475
Epoch[51] Batch [849]	Speed: 5159.618173 samples/sec	accuracy=0.738856	lr=0.258267
Epoch[51] Batch [899]	Speed: 5030.486970 samples/sec	accuracy=0.738931	lr=0.258058
Epoch[51] Batch [949]	Speed: 5305.521814 samples/sec	accuracy=0.739030	lr=0.257849
Epoch[51] Batch [999]	Speed: 5142.931348 samples/sec	accuracy=0.739180	lr=0.257640
Epoch[51] Batch [1049]	Speed: 5115.629223 samples/sec	accuracy=0.739334	lr=0.257430
Epoch[51] Batch [1099]	Speed: 5255.066726 samples/sec	accuracy=0.739476	lr=0.257221
Epoch[51] Batch [1149]	Speed: 5229.287937 samples/sec	accuracy=0.739486	lr=0.257012
Epoch[51] Batch [1199]	Speed: 5239.181363 samples/sec	accuracy=0.739618	lr=0.256803
Epoch[51] Batch [1249]	Speed: 5300.285482 samples/sec	accuracy=0.739740	lr=0.256593
[Epoch 51] training: accuracy=0.739744
[Epoch 51] speed: 5152 samples/sec	time cost: 257.776656
[Epoch 51] validation: err-top1=0.406788 err-top5=0.173609
Epoch[52] Batch [49]	Speed: 5022.128751 samples/sec	accuracy=0.739217	lr=0.256379
Epoch[52] Batch [99]	Speed: 5059.949961 samples/sec	accuracy=0.740139	lr=0.256170
Epoch[52] Batch [149]	Speed: 5142.060464 samples/sec	accuracy=0.739650	lr=0.255960
Epoch[52] Batch [199]	Speed: 5103.339668 samples/sec	accuracy=0.739468	lr=0.255751
Epoch[52] Batch [249]	Speed: 5041.930275 samples/sec	accuracy=0.739415	lr=0.255541
Epoch[52] Batch [299]	Speed: 5025.674190 samples/sec	accuracy=0.739486	lr=0.255331
Epoch[52] Batch [349]	Speed: 5011.762967 samples/sec	accuracy=0.739489	lr=0.255121
Epoch[52] Batch [399]	Speed: 5126.026617 samples/sec	accuracy=0.739498	lr=0.254911
Epoch[52] Batch [449]	Speed: 5348.929448 samples/sec	accuracy=0.739649	lr=0.254701
Epoch[52] Batch [499]	Speed: 5185.970682 samples/sec	accuracy=0.739757	lr=0.254491
Epoch[52] Batch [549]	Speed: 5274.851187 samples/sec	accuracy=0.739906	lr=0.254281
Epoch[52] Batch [599]	Speed: 5205.259546 samples/sec	accuracy=0.740046	lr=0.254071
Epoch[52] Batch [649]	Speed: 5219.901601 samples/sec	accuracy=0.740194	lr=0.253861
Epoch[52] Batch [699]	Speed: 5020.275282 samples/sec	accuracy=0.740278	lr=0.253650
Epoch[52] Batch [749]	Speed: 5084.163247 samples/sec	accuracy=0.740281	lr=0.253440
Epoch[52] Batch [799]	Speed: 5144.562216 samples/sec	accuracy=0.740268	lr=0.253229
Epoch[52] Batch [849]	Speed: 5053.860081 samples/sec	accuracy=0.740365	lr=0.253019
Epoch[52] Batch [899]	Speed: 5132.468277 samples/sec	accuracy=0.740497	lr=0.252808
Epoch[52] Batch [949]	Speed: 5136.909176 samples/sec	accuracy=0.740619	lr=0.252598
Epoch[52] Batch [999]	Speed: 5191.858456 samples/sec	accuracy=0.740727	lr=0.252387
Epoch[52] Batch [1049]	Speed: 5035.089266 samples/sec	accuracy=0.740878	lr=0.252176
Epoch[52] Batch [1099]	Speed: 5090.634122 samples/sec	accuracy=0.740970	lr=0.251965
Epoch[52] Batch [1149]	Speed: 5119.033020 samples/sec	accuracy=0.741080	lr=0.251754
Epoch[52] Batch [1199]	Speed: 5126.588668 samples/sec	accuracy=0.741151	lr=0.251543
Epoch[52] Batch [1249]	Speed: 5467.289823 samples/sec	accuracy=0.741310	lr=0.251332
[Epoch 52] training: accuracy=0.741316
[Epoch 52] speed: 5128 samples/sec	time cost: 259.101807
[Epoch 52] validation: err-top1=0.416117 err-top5=0.177979
Epoch[53] Batch [49]	Speed: 4788.602384 samples/sec	accuracy=0.741199	lr=0.251117
Epoch[53] Batch [99]	Speed: 4976.634547 samples/sec	accuracy=0.740814	lr=0.250906
Epoch[53] Batch [149]	Speed: 5105.152678 samples/sec	accuracy=0.740286	lr=0.250695
Epoch[53] Batch [199]	Speed: 5263.655117 samples/sec	accuracy=0.739926	lr=0.250483
Epoch[53] Batch [249]	Speed: 4847.182680 samples/sec	accuracy=0.739944	lr=0.250272
Epoch[53] Batch [299]	Speed: 5010.534798 samples/sec	accuracy=0.740143	lr=0.250061
Epoch[53] Batch [349]	Speed: 5198.408604 samples/sec	accuracy=0.740323	lr=0.249849
Epoch[53] Batch [399]	Speed: 5137.701984 samples/sec	accuracy=0.740185	lr=0.249638
Epoch[53] Batch [449]	Speed: 5086.861150 samples/sec	accuracy=0.740351	lr=0.249426
Epoch[53] Batch [499]	Speed: 5360.201943 samples/sec	accuracy=0.740570	lr=0.249215
Epoch[53] Batch [549]	Speed: 5301.631683 samples/sec	accuracy=0.740592	lr=0.249003
Epoch[53] Batch [599]	Speed: 5072.022316 samples/sec	accuracy=0.740789	lr=0.248791
Epoch[53] Batch [649]	Speed: 5419.362957 samples/sec	accuracy=0.740835	lr=0.248579
Epoch[53] Batch [699]	Speed: 4997.825721 samples/sec	accuracy=0.741027	lr=0.248367
Epoch[53] Batch [749]	Speed: 5129.333879 samples/sec	accuracy=0.741224	lr=0.248156
Epoch[53] Batch [799]	Speed: 5006.564765 samples/sec	accuracy=0.741347	lr=0.247944
Epoch[53] Batch [849]	Speed: 5223.009230 samples/sec	accuracy=0.741502	lr=0.247732
Epoch[53] Batch [899]	Speed: 5134.594093 samples/sec	accuracy=0.741614	lr=0.247519
Epoch[53] Batch [949]	Speed: 4934.955976 samples/sec	accuracy=0.741749	lr=0.247307
Epoch[53] Batch [999]	Speed: 5068.588253 samples/sec	accuracy=0.741865	lr=0.247095
Epoch[53] Batch [1049]	Speed: 5011.328835 samples/sec	accuracy=0.741971	lr=0.246883
Epoch[53] Batch [1099]	Speed: 5039.331958 samples/sec	accuracy=0.742149	lr=0.246670
Epoch[53] Batch [1149]	Speed: 5227.601263 samples/sec	accuracy=0.742310	lr=0.246458
Epoch[53] Batch [1199]	Speed: 5163.990464 samples/sec	accuracy=0.742415	lr=0.246246
Epoch[53] Batch [1249]	Speed: 5186.948962 samples/sec	accuracy=0.742531	lr=0.246033
[Epoch 53] training: accuracy=0.742525
[Epoch 53] speed: 5099 samples/sec	time cost: 260.669849
[Epoch 53] validation: err-top1=0.381776 err-top5=0.156509
Epoch[54] Batch [49]	Speed: 4777.746226 samples/sec	accuracy=0.741884	lr=0.245816
Epoch[54] Batch [99]	Speed: 5104.309342 samples/sec	accuracy=0.741029	lr=0.245604
Epoch[54] Batch [149]	Speed: 5187.669944 samples/sec	accuracy=0.741299	lr=0.245391
Epoch[54] Batch [199]	Speed: 4983.362169 samples/sec	accuracy=0.741787	lr=0.245178
Epoch[54] Batch [249]	Speed: 5154.009821 samples/sec	accuracy=0.741733	lr=0.244966
Epoch[54] Batch [299]	Speed: 5105.984516 samples/sec	accuracy=0.741712	lr=0.244753
Epoch[54] Batch [349]	Speed: 5189.754450 samples/sec	accuracy=0.742110	lr=0.244540
Epoch[54] Batch [399]	Speed: 5233.462451 samples/sec	accuracy=0.742256	lr=0.244327
Epoch[54] Batch [449]	Speed: 5027.609566 samples/sec	accuracy=0.742360	lr=0.244114
Epoch[54] Batch [499]	Speed: 5296.503869 samples/sec	accuracy=0.742392	lr=0.243901
Epoch[54] Batch [549]	Speed: 5270.958147 samples/sec	accuracy=0.742527	lr=0.243688
Epoch[54] Batch [599]	Speed: 5340.836629 samples/sec	accuracy=0.742609	lr=0.243475
Epoch[54] Batch [649]	Speed: 5102.694797 samples/sec	accuracy=0.742780	lr=0.243262
Epoch[54] Batch [699]	Speed: 5224.548417 samples/sec	accuracy=0.742830	lr=0.243049
Epoch[54] Batch [749]	Speed: 5170.079450 samples/sec	accuracy=0.742981	lr=0.242835
Epoch[54] Batch [799]	Speed: 5258.101923 samples/sec	accuracy=0.743118	lr=0.242622
Epoch[54] Batch [849]	Speed: 5163.311430 samples/sec	accuracy=0.743221	lr=0.242409
Epoch[54] Batch [899]	Speed: 5012.718275 samples/sec	accuracy=0.743295	lr=0.242195
Epoch[54] Batch [949]	Speed: 5266.614094 samples/sec	accuracy=0.743316	lr=0.241982
Epoch[54] Batch [999]	Speed: 5251.996753 samples/sec	accuracy=0.743386	lr=0.241768
Epoch[54] Batch [1049]	Speed: 5301.718068 samples/sec	accuracy=0.743483	lr=0.241555
Epoch[54] Batch [1099]	Speed: 4982.192610 samples/sec	accuracy=0.743638	lr=0.241341
Epoch[54] Batch [1149]	Speed: 4996.922705 samples/sec	accuracy=0.743727	lr=0.241127
Epoch[54] Batch [1199]	Speed: 5111.883858 samples/sec	accuracy=0.743778	lr=0.240914
Epoch[54] Batch [1249]	Speed: 5266.671572 samples/sec	accuracy=0.743901	lr=0.240700
[Epoch 54] training: accuracy=0.743910
[Epoch 54] speed: 5143 samples/sec	time cost: 258.319596
[Epoch 54] validation: err-top1=0.412986 err-top5=0.178272
Epoch[55] Batch [49]	Speed: 5083.407572 samples/sec	accuracy=0.741472	lr=0.240482
Epoch[55] Batch [99]	Speed: 5205.432656 samples/sec	accuracy=0.741554	lr=0.240268
Epoch[55] Batch [149]	Speed: 4992.066791 samples/sec	accuracy=0.742068	lr=0.240054
Epoch[55] Batch [199]	Speed: 5244.220887 samples/sec	accuracy=0.741893	lr=0.239840
Epoch[55] Batch [249]	Speed: 5290.415989 samples/sec	accuracy=0.742060	lr=0.239626
Epoch[55] Batch [299]	Speed: 5398.947953 samples/sec	accuracy=0.742225	lr=0.239412
Epoch[55] Batch [349]	Speed: 4954.889736 samples/sec	accuracy=0.742366	lr=0.239198
Epoch[55] Batch [399]	Speed: 4900.108666 samples/sec	accuracy=0.742504	lr=0.238984
Epoch[55] Batch [449]	Speed: 4831.890332 samples/sec	accuracy=0.742767	lr=0.238769
Epoch[55] Batch [499]	Speed: 4866.060298 samples/sec	accuracy=0.742955	lr=0.238555
Epoch[55] Batch [549]	Speed: 5134.232078 samples/sec	accuracy=0.743067	lr=0.238341
Epoch[55] Batch [599]	Speed: 4986.283837 samples/sec	accuracy=0.743218	lr=0.238126
Epoch[55] Batch [649]	Speed: 5157.231554 samples/sec	accuracy=0.743425	lr=0.237912
Epoch[55] Batch [699]	Speed: 5181.900443 samples/sec	accuracy=0.743494	lr=0.237698
Epoch[55] Batch [749]	Speed: 5316.330374 samples/sec	accuracy=0.743547	lr=0.237483
Epoch[55] Batch [799]	Speed: 5312.983494 samples/sec	accuracy=0.743801	lr=0.237269
Epoch[55] Batch [849]	Speed: 5251.332132 samples/sec	accuracy=0.743914	lr=0.237054
Epoch[55] Batch [899]	Speed: 5198.353740 samples/sec	accuracy=0.743923	lr=0.236839
Epoch[55] Batch [949]	Speed: 5130.133052 samples/sec	accuracy=0.744072	lr=0.236625
Epoch[55] Batch [999]	Speed: 4845.951715 samples/sec	accuracy=0.744207	lr=0.236410
Epoch[55] Batch [1049]	Speed: 5008.275899 samples/sec	accuracy=0.744303	lr=0.236195
Epoch[55] Batch [1099]	Speed: 5312.733759 samples/sec	accuracy=0.744435	lr=0.235981
Epoch[55] Batch [1149]	Speed: 5314.715201 samples/sec	accuracy=0.744568	lr=0.235766
Epoch[55] Batch [1199]	Speed: 5399.397269 samples/sec	accuracy=0.744660	lr=0.235551
Epoch[55] Batch [1249]	Speed: 5262.669487 samples/sec	accuracy=0.744707	lr=0.235336
[Epoch 55] training: accuracy=0.744701
[Epoch 55] speed: 5132 samples/sec	time cost: 258.894350
[Epoch 55] validation: err-top1=0.412010 err-top5=0.178532
Epoch[56] Batch [49]	Speed: 5061.724871 samples/sec	accuracy=0.744632	lr=0.235117
Epoch[56] Batch [99]	Speed: 5119.085124 samples/sec	accuracy=0.743576	lr=0.234902
Epoch[56] Batch [149]	Speed: 5172.597583 samples/sec	accuracy=0.744106	lr=0.234687
Epoch[56] Batch [199]	Speed: 5155.912495 samples/sec	accuracy=0.743897	lr=0.234472
Epoch[56] Batch [249]	Speed: 5123.149641 samples/sec	accuracy=0.743883	lr=0.234256
Epoch[56] Batch [299]	Speed: 5289.898884 samples/sec	accuracy=0.743914	lr=0.234041
Epoch[56] Batch [349]	Speed: 5037.190105 samples/sec	accuracy=0.744192	lr=0.233826
Epoch[56] Batch [399]	Speed: 5353.611801 samples/sec	accuracy=0.744059	lr=0.233611
Epoch[56] Batch [449]	Speed: 4916.593220 samples/sec	accuracy=0.743958	lr=0.233396
Epoch[56] Batch [499]	Speed: 5192.540249 samples/sec	accuracy=0.744054	lr=0.233180
Epoch[56] Batch [549]	Speed: 5045.896978 samples/sec	accuracy=0.744196	lr=0.232965
Epoch[56] Batch [599]	Speed: 5186.165181 samples/sec	accuracy=0.744314	lr=0.232749
Epoch[56] Batch [649]	Speed: 5259.693813 samples/sec	accuracy=0.744454	lr=0.232534
Epoch[56] Batch [699]	Speed: 5345.571349 samples/sec	accuracy=0.744632	lr=0.232319
Epoch[56] Batch [749]	Speed: 5006.726312 samples/sec	accuracy=0.744919	lr=0.232103
Epoch[56] Batch [799]	Speed: 4904.567351 samples/sec	accuracy=0.744992	lr=0.231887
Epoch[56] Batch [849]	Speed: 5113.833242 samples/sec	accuracy=0.745094	lr=0.231672
Epoch[56] Batch [899]	Speed: 5076.547228 samples/sec	accuracy=0.745223	lr=0.231456
Epoch[56] Batch [949]	Speed: 5185.923468 samples/sec	accuracy=0.745368	lr=0.231241
Epoch[56] Batch [999]	Speed: 5089.137962 samples/sec	accuracy=0.745475	lr=0.231025
Epoch[56] Batch [1049]	Speed: 5049.314124 samples/sec	accuracy=0.745582	lr=0.230809
Epoch[56] Batch [1099]	Speed: 4935.107718 samples/sec	accuracy=0.745629	lr=0.230593
Epoch[56] Batch [1149]	Speed: 5207.826345 samples/sec	accuracy=0.745715	lr=0.230377
Epoch[56] Batch [1199]	Speed: 5242.919296 samples/sec	accuracy=0.745904	lr=0.230162
Epoch[56] Batch [1249]	Speed: 5179.701801 samples/sec	accuracy=0.745971	lr=0.229946
[Epoch 56] training: accuracy=0.745974
[Epoch 56] speed: 5122 samples/sec	time cost: 258.992669
[Epoch 56] validation: err-top1=0.373565 err-top5=0.147082
Epoch[57] Batch [49]	Speed: 4960.895253 samples/sec	accuracy=0.747078	lr=0.229725
Epoch[57] Batch [99]	Speed: 4950.187248 samples/sec	accuracy=0.745191	lr=0.229510
Epoch[57] Batch [149]	Speed: 4901.309355 samples/sec	accuracy=0.744606	lr=0.229294
Epoch[57] Batch [199]	Speed: 5033.334046 samples/sec	accuracy=0.744224	lr=0.229077
Epoch[57] Batch [249]	Speed: 5011.005624 samples/sec	accuracy=0.744246	lr=0.228861
Epoch[57] Batch [299]	Speed: 5019.316036 samples/sec	accuracy=0.744554	lr=0.228645
Epoch[57] Batch [349]	Speed: 5300.748096 samples/sec	accuracy=0.744377	lr=0.228429
Epoch[57] Batch [399]	Speed: 5384.699389 samples/sec	accuracy=0.744509	lr=0.228213
Epoch[57] Batch [449]	Speed: 5352.393419 samples/sec	accuracy=0.744747	lr=0.227997
Epoch[57] Batch [499]	Speed: 5025.094067 samples/sec	accuracy=0.744821	lr=0.227781
Epoch[57] Batch [549]	Speed: 5285.727688 samples/sec	accuracy=0.745086	lr=0.227564
Epoch[57] Batch [599]	Speed: 5144.497144 samples/sec	accuracy=0.745303	lr=0.227348
Epoch[57] Batch [649]	Speed: 5159.761607 samples/sec	accuracy=0.745535	lr=0.227132
Epoch[57] Batch [699]	Speed: 5145.826891 samples/sec	accuracy=0.745641	lr=0.226915
Epoch[57] Batch [749]	Speed: 5218.890686 samples/sec	accuracy=0.745807	lr=0.226699
Epoch[57] Batch [799]	Speed: 5314.326817 samples/sec	accuracy=0.745895	lr=0.226482
Epoch[57] Batch [849]	Speed: 5163.933095 samples/sec	accuracy=0.745968	lr=0.226266
Epoch[57] Batch [899]	Speed: 5070.546892 samples/sec	accuracy=0.746128	lr=0.226049
Epoch[57] Batch [949]	Speed: 4894.380734 samples/sec	accuracy=0.746332	lr=0.225833
Epoch[57] Batch [999]	Speed: 5211.922757 samples/sec	accuracy=0.746408	lr=0.225616
Epoch[57] Batch [1049]	Speed: 5222.984205 samples/sec	accuracy=0.746489	lr=0.225400
Epoch[57] Batch [1099]	Speed: 4951.722695 samples/sec	accuracy=0.746615	lr=0.225183
Epoch[57] Batch [1149]	Speed: 5298.039106 samples/sec	accuracy=0.746719	lr=0.224967
Epoch[57] Batch [1199]	Speed: 5153.574938 samples/sec	accuracy=0.746836	lr=0.224750
Epoch[57] Batch [1249]	Speed: 5191.237328 samples/sec	accuracy=0.746930	lr=0.224533
[Epoch 57] training: accuracy=0.746942
[Epoch 57] speed: 5126 samples/sec	time cost: 259.133757
[Epoch 57] validation: err-top1=0.379903 err-top5=0.149833
Epoch[58] Batch [49]	Speed: 4853.257282 samples/sec	accuracy=0.747609	lr=0.224312
Epoch[58] Batch [99]	Speed: 4888.974467 samples/sec	accuracy=0.746984	lr=0.224095
Epoch[58] Batch [149]	Speed: 4936.395747 samples/sec	accuracy=0.746314	lr=0.223879
Epoch[58] Batch [199]	Speed: 5166.510001 samples/sec	accuracy=0.746197	lr=0.223662
Epoch[58] Batch [249]	Speed: 4958.396939 samples/sec	accuracy=0.746387	lr=0.223445
Epoch[58] Batch [299]	Speed: 4975.253049 samples/sec	accuracy=0.746273	lr=0.223228
Epoch[58] Batch [349]	Speed: 4876.647466 samples/sec	accuracy=0.746368	lr=0.223011
Epoch[58] Batch [399]	Speed: 5078.334282 samples/sec	accuracy=0.746232	lr=0.222794
Epoch[58] Batch [449]	Speed: 5009.161875 samples/sec	accuracy=0.746252	lr=0.222577
Epoch[58] Batch [499]	Speed: 5025.832974 samples/sec	accuracy=0.746407	lr=0.222360
Epoch[58] Batch [549]	Speed: 5030.085759 samples/sec	accuracy=0.746394	lr=0.222143
Epoch[58] Batch [599]	Speed: 5255.218987 samples/sec	accuracy=0.746560	lr=0.221926
Epoch[58] Batch [649]	Speed: 5135.825498 samples/sec	accuracy=0.746811	lr=0.221709
Epoch[58] Batch [699]	Speed: 5090.113467 samples/sec	accuracy=0.746973	lr=0.221492
Epoch[58] Batch [749]	Speed: 5198.908983 samples/sec	accuracy=0.747166	lr=0.221275
Epoch[58] Batch [799]	Speed: 4898.381583 samples/sec	accuracy=0.747354	lr=0.221058
Epoch[58] Batch [849]	Speed: 4845.683379 samples/sec	accuracy=0.747473	lr=0.220841
Epoch[58] Batch [899]	Speed: 5008.844668 samples/sec	accuracy=0.747612	lr=0.220623
Epoch[58] Batch [949]	Speed: 5121.646525 samples/sec	accuracy=0.747768	lr=0.220406
Epoch[58] Batch [999]	Speed: 5024.008270 samples/sec	accuracy=0.747901	lr=0.220189
Epoch[58] Batch [1049]	Speed: 5206.456793 samples/sec	accuracy=0.747971	lr=0.219972
Epoch[58] Batch [1099]	Speed: 5103.432204 samples/sec	accuracy=0.748059	lr=0.219754
Epoch[58] Batch [1149]	Speed: 5095.542924 samples/sec	accuracy=0.748149	lr=0.219537
Epoch[58] Batch [1199]	Speed: 5142.658549 samples/sec	accuracy=0.748262	lr=0.219320
Epoch[58] Batch [1249]	Speed: 5415.465027 samples/sec	accuracy=0.748339	lr=0.219102
[Epoch 58] training: accuracy=0.748342
[Epoch 58] speed: 5046 samples/sec	time cost: 263.106447
[Epoch 58] validation: err-top1=0.366821 err-top5=0.139587
Epoch[59] Batch [49]	Speed: 4954.543929 samples/sec	accuracy=0.747504	lr=0.218881
Epoch[59] Batch [99]	Speed: 4820.582930 samples/sec	accuracy=0.747961	lr=0.218663
Epoch[59] Batch [149]	Speed: 5037.408699 samples/sec	accuracy=0.746952	lr=0.218446
Epoch[59] Batch [199]	Speed: 4957.189519 samples/sec	accuracy=0.747057	lr=0.218228
Epoch[59] Batch [249]	Speed: 4954.766040 samples/sec	accuracy=0.747242	lr=0.218011
Epoch[59] Batch [299]	Speed: 5226.383080 samples/sec	accuracy=0.747151	lr=0.217793
Epoch[59] Batch [349]	Speed: 5418.442976 samples/sec	accuracy=0.747435	lr=0.217576
Epoch[59] Batch [399]	Speed: 5217.624329 samples/sec	accuracy=0.747578	lr=0.217358
Epoch[59] Batch [449]	Speed: 5389.192362 samples/sec	accuracy=0.747697	lr=0.217141
Epoch[59] Batch [499]	Speed: 5476.791687 samples/sec	accuracy=0.747888	lr=0.216923
Epoch[59] Batch [549]	Speed: 5259.400888 samples/sec	accuracy=0.747955	lr=0.216706
Epoch[59] Batch [599]	Speed: 5200.684495 samples/sec	accuracy=0.748104	lr=0.216488
Epoch[59] Batch [649]	Speed: 5133.561338 samples/sec	accuracy=0.748258	lr=0.216270
Epoch[59] Batch [699]	Speed: 5326.008380 samples/sec	accuracy=0.748513	lr=0.216053
Epoch[59] Batch [749]	Speed: 5222.051968 samples/sec	accuracy=0.748505	lr=0.215835
Epoch[59] Batch [799]	Speed: 5124.687026 samples/sec	accuracy=0.748594	lr=0.215617
Epoch[59] Batch [849]	Speed: 5209.247920 samples/sec	accuracy=0.748730	lr=0.215400
Epoch[59] Batch [899]	Speed: 5193.681659 samples/sec	accuracy=0.748908	lr=0.215182
Epoch[59] Batch [949]	Speed: 5111.883371 samples/sec	accuracy=0.748959	lr=0.214964
Epoch[59] Batch [999]	Speed: 5113.804138 samples/sec	accuracy=0.749108	lr=0.214746
Epoch[59] Batch [1049]	Speed: 5214.675050 samples/sec	accuracy=0.749271	lr=0.214529
Epoch[59] Batch [1099]	Speed: 5276.981721 samples/sec	accuracy=0.749344	lr=0.214311
Epoch[59] Batch [1149]	Speed: 5118.886107 samples/sec	accuracy=0.749554	lr=0.214093
Epoch[59] Batch [1199]	Speed: 4915.791671 samples/sec	accuracy=0.749704	lr=0.213875
Epoch[59] Batch [1249]	Speed: 5139.372830 samples/sec	accuracy=0.749806	lr=0.213657
[Epoch 59] training: accuracy=0.749807
[Epoch 59] speed: 5151 samples/sec	time cost: 258.220740
[Epoch 59] validation: err-top1=0.384068 err-top5=0.153101
Epoch[60] Batch [49]	Speed: 4842.804630 samples/sec	accuracy=0.749773	lr=0.213435
Epoch[60] Batch [99]	Speed: 5133.357021 samples/sec	accuracy=0.750104	lr=0.213217
Epoch[60] Batch [149]	Speed: 4967.177860 samples/sec	accuracy=0.750064	lr=0.212999
Epoch[60] Batch [199]	Speed: 5044.014910 samples/sec	accuracy=0.749618	lr=0.212781
Epoch[60] Batch [249]	Speed: 5192.716282 samples/sec	accuracy=0.749440	lr=0.212563
Epoch[60] Batch [299]	Speed: 5261.199017 samples/sec	accuracy=0.749394	lr=0.212345
Epoch[60] Batch [349]	Speed: 5243.896389 samples/sec	accuracy=0.749257	lr=0.212127
Epoch[60] Batch [399]	Speed: 4899.386031 samples/sec	accuracy=0.749397	lr=0.211909
Epoch[60] Batch [449]	Speed: 4879.602527 samples/sec	accuracy=0.749718	lr=0.211691
Epoch[60] Batch [499]	Speed: 5242.354997 samples/sec	accuracy=0.749789	lr=0.211473
Epoch[60] Batch [549]	Speed: 5180.429143 samples/sec	accuracy=0.749833	lr=0.211255
Epoch[60] Batch [599]	Speed: 5164.087324 samples/sec	accuracy=0.749910	lr=0.211037
Epoch[60] Batch [649]	Speed: 5047.649583 samples/sec	accuracy=0.750032	lr=0.210819
Epoch[60] Batch [699]	Speed: 4822.335061 samples/sec	accuracy=0.750052	lr=0.210601
Epoch[60] Batch [749]	Speed: 5176.945856 samples/sec	accuracy=0.750139	lr=0.210383
Epoch[60] Batch [799]	Speed: 5083.135517 samples/sec	accuracy=0.750213	lr=0.210165
Epoch[60] Batch [849]	Speed: 5036.803536 samples/sec	accuracy=0.750340	lr=0.209947
Epoch[60] Batch [899]	Speed: 5170.662159 samples/sec	accuracy=0.750515	lr=0.209729
Epoch[60] Batch [949]	Speed: 5015.750155 samples/sec	accuracy=0.750640	lr=0.209511
Epoch[60] Batch [999]	Speed: 5165.711882 samples/sec	accuracy=0.750781	lr=0.209293
Epoch[60] Batch [1049]	Speed: 5114.700072 samples/sec	accuracy=0.750920	lr=0.209075
Epoch[60] Batch [1099]	Speed: 5178.748732 samples/sec	accuracy=0.751085	lr=0.208856
Epoch[60] Batch [1149]	Speed: 5192.829416 samples/sec	accuracy=0.751098	lr=0.208638
Epoch[60] Batch [1199]	Speed: 5052.155101 samples/sec	accuracy=0.751197	lr=0.208420
Epoch[60] Batch [1249]	Speed: 5149.395342 samples/sec	accuracy=0.751299	lr=0.208202
[Epoch 60] training: accuracy=0.751305
[Epoch 60] speed: 5083 samples/sec	time cost: 261.138904
[Epoch 60] validation: err-top1=0.384686 err-top5=0.156310
Epoch[61] Batch [49]	Speed: 4650.674004 samples/sec	accuracy=0.751847	lr=0.207979
Epoch[61] Batch [99]	Speed: 5123.451177 samples/sec	accuracy=0.750496	lr=0.207761
Epoch[61] Batch [149]	Speed: 5025.894137 samples/sec	accuracy=0.750500	lr=0.207543
Epoch[61] Batch [199]	Speed: 5107.758222 samples/sec	accuracy=0.750728	lr=0.207325
Epoch[61] Batch [249]	Speed: 5454.330242 samples/sec	accuracy=0.750600	lr=0.207106
Epoch[61] Batch [299]	Speed: 5350.124393 samples/sec	accuracy=0.750462	lr=0.206888
Epoch[61] Batch [349]	Speed: 5052.182438 samples/sec	accuracy=0.750616	lr=0.206670
Epoch[61] Batch [399]	Speed: 4937.156356 samples/sec	accuracy=0.750722	lr=0.206452
Epoch[61] Batch [449]	Speed: 5238.429892 samples/sec	accuracy=0.750826	lr=0.206233
Epoch[61] Batch [499]	Speed: 5124.031125 samples/sec	accuracy=0.751125	lr=0.206015
Epoch[61] Batch [549]	Speed: 5104.997459 samples/sec	accuracy=0.751195	lr=0.205797
Epoch[61] Batch [599]	Speed: 5124.116711 samples/sec	accuracy=0.751432	lr=0.205579
Epoch[61] Batch [649]	Speed: 5235.922206 samples/sec	accuracy=0.751572	lr=0.205360
Epoch[61] Batch [699]	Speed: 5236.709479 samples/sec	accuracy=0.751648	lr=0.205142
Epoch[61] Batch [749]	Speed: 5245.913947 samples/sec	accuracy=0.751684	lr=0.204924
Epoch[61] Batch [799]	Speed: 5188.859237 samples/sec	accuracy=0.751759	lr=0.204705
Epoch[61] Batch [849]	Speed: 5044.633656 samples/sec	accuracy=0.751848	lr=0.204487
Epoch[61] Batch [899]	Speed: 5462.280720 samples/sec	accuracy=0.752040	lr=0.204269
Epoch[61] Batch [949]	Speed: 5309.873071 samples/sec	accuracy=0.752142	lr=0.204051
Epoch[61] Batch [999]	Speed: 5058.462725 samples/sec	accuracy=0.752180	lr=0.203832
Epoch[61] Batch [1049]	Speed: 4936.214879 samples/sec	accuracy=0.752200	lr=0.203614
Epoch[61] Batch [1099]	Speed: 5211.515228 samples/sec	accuracy=0.752262	lr=0.203396
Epoch[61] Batch [1149]	Speed: 5130.681540 samples/sec	accuracy=0.752358	lr=0.203177
Epoch[61] Batch [1199]	Speed: 5294.273749 samples/sec	accuracy=0.752459	lr=0.202959
Epoch[61] Batch [1249]	Speed: 5239.414133 samples/sec	accuracy=0.752541	lr=0.202740
[Epoch 61] training: accuracy=0.752541
[Epoch 61] speed: 5145 samples/sec	time cost: 258.395181
[Epoch 61] validation: err-top1=0.381796 err-top5=0.155513
Epoch[62] Batch [49]	Speed: 4914.512909 samples/sec	accuracy=0.752460	lr=0.202518
Epoch[62] Batch [99]	Speed: 5259.795327 samples/sec	accuracy=0.752162	lr=0.202299
Epoch[62] Batch [149]	Speed: 5063.382712 samples/sec	accuracy=0.751866	lr=0.202081
Epoch[62] Batch [199]	Speed: 5027.940338 samples/sec	accuracy=0.750925	lr=0.201863
Epoch[62] Batch [249]	Speed: 5209.458450 samples/sec	accuracy=0.750823	lr=0.201644
Epoch[62] Batch [299]	Speed: 5076.919757 samples/sec	accuracy=0.750943	lr=0.201426
Epoch[62] Batch [349]	Speed: 5132.135998 samples/sec	accuracy=0.751119	lr=0.201208
Epoch[62] Batch [399]	Speed: 5062.445470 samples/sec	accuracy=0.751334	lr=0.200989
Epoch[62] Batch [449]	Speed: 5042.413887 samples/sec	accuracy=0.751468	lr=0.200771
Epoch[62] Batch [499]	Speed: 5176.732331 samples/sec	accuracy=0.751650	lr=0.200552
Epoch[62] Batch [549]	Speed: 5253.094936 samples/sec	accuracy=0.751696	lr=0.200334
Epoch[62] Batch [599]	Speed: 5285.660557 samples/sec	accuracy=0.751863	lr=0.200116
Epoch[62] Batch [649]	Speed: 5310.314775 samples/sec	accuracy=0.752044	lr=0.199897
Epoch[62] Batch [699]	Speed: 5147.469960 samples/sec	accuracy=0.752154	lr=0.199679
Epoch[62] Batch [749]	Speed: 5277.895538 samples/sec	accuracy=0.752304	lr=0.199461
Epoch[62] Batch [799]	Speed: 5076.026330 samples/sec	accuracy=0.752347	lr=0.199242
Epoch[62] Batch [849]	Speed: 4953.992225 samples/sec	accuracy=0.752381	lr=0.199024
Epoch[62] Batch [899]	Speed: 5458.671159 samples/sec	accuracy=0.752495	lr=0.198806
Epoch[62] Batch [949]	Speed: 5186.169315 samples/sec	accuracy=0.752644	lr=0.198587
Epoch[62] Batch [999]	Speed: 4912.784650 samples/sec	accuracy=0.752718	lr=0.198369
Epoch[62] Batch [1049]	Speed: 5387.063113 samples/sec	accuracy=0.752855	lr=0.198150
Epoch[62] Batch [1099]	Speed: 5234.975264 samples/sec	accuracy=0.752905	lr=0.197932
Epoch[62] Batch [1149]	Speed: 5125.523164 samples/sec	accuracy=0.753051	lr=0.197714
Epoch[62] Batch [1199]	Speed: 5244.863470 samples/sec	accuracy=0.753170	lr=0.197495
Epoch[62] Batch [1249]	Speed: 5313.501704 samples/sec	accuracy=0.753339	lr=0.197277
[Epoch 62] training: accuracy=0.753342
[Epoch 62] speed: 5156 samples/sec	time cost: 257.737614
[Epoch 62] validation: err-top1=0.372030 err-top5=0.147381
Epoch[63] Batch [49]	Speed: 4978.628822 samples/sec	accuracy=0.752845	lr=0.197054
Epoch[63] Batch [99]	Speed: 5168.570568 samples/sec	accuracy=0.753249	lr=0.196836
Epoch[63] Batch [149]	Speed: 5134.855601 samples/sec	accuracy=0.753057	lr=0.196618
Epoch[63] Batch [199]	Speed: 5071.127498 samples/sec	accuracy=0.752503	lr=0.196399
Epoch[63] Batch [249]	Speed: 5320.439038 samples/sec	accuracy=0.752586	lr=0.196181
Epoch[63] Batch [299]	Speed: 5131.319158 samples/sec	accuracy=0.752715	lr=0.195963
Epoch[63] Batch [349]	Speed: 5043.547455 samples/sec	accuracy=0.752826	lr=0.195744
Epoch[63] Batch [399]	Speed: 5172.831700 samples/sec	accuracy=0.752885	lr=0.195526
Epoch[63] Batch [449]	Speed: 5027.231410 samples/sec	accuracy=0.752766	lr=0.195308
Epoch[63] Batch [499]	Speed: 5206.679847 samples/sec	accuracy=0.752796	lr=0.195089
Epoch[63] Batch [549]	Speed: 4963.400675 samples/sec	accuracy=0.753068	lr=0.194871
Epoch[63] Batch [599]	Speed: 5356.408262 samples/sec	accuracy=0.753159	lr=0.194653
Epoch[63] Batch [649]	Speed: 5081.071925 samples/sec	accuracy=0.753103	lr=0.194434
Epoch[63] Batch [699]	Speed: 5138.478437 samples/sec	accuracy=0.753382	lr=0.194216
Epoch[63] Batch [749]	Speed: 5197.762130 samples/sec	accuracy=0.753525	lr=0.193998
Epoch[63] Batch [799]	Speed: 5364.331959 samples/sec	accuracy=0.753685	lr=0.193780
Epoch[63] Batch [849]	Speed: 5069.531839 samples/sec	accuracy=0.753742	lr=0.193561
Epoch[63] Batch [899]	Speed: 5072.847946 samples/sec	accuracy=0.753831	lr=0.193343
Epoch[63] Batch [949]	Speed: 5284.281884 samples/sec	accuracy=0.753926	lr=0.193125
Epoch[63] Batch [999]	Speed: 5128.026847 samples/sec	accuracy=0.754054	lr=0.192907
Epoch[63] Batch [1049]	Speed: 5194.938053 samples/sec	accuracy=0.754100	lr=0.192688
Epoch[63] Batch [1099]	Speed: 5047.709262 samples/sec	accuracy=0.754156	lr=0.192470
Epoch[63] Batch [1149]	Speed: 5068.187400 samples/sec	accuracy=0.754201	lr=0.192252
Epoch[63] Batch [1199]	Speed: 5170.797368 samples/sec	accuracy=0.754300	lr=0.192034
Epoch[63] Batch [1249]	Speed: 5421.545880 samples/sec	accuracy=0.754337	lr=0.191816
[Epoch 63] training: accuracy=0.754337
[Epoch 63] speed: 5145 samples/sec	time cost: 257.543022
[Epoch 63] validation: err-top1=0.376668 err-top5=0.150879
Epoch[64] Batch [49]	Speed: 4923.125183 samples/sec	accuracy=0.753861	lr=0.191593
Epoch[64] Batch [99]	Speed: 5253.269700 samples/sec	accuracy=0.753704	lr=0.191375
Epoch[64] Batch [149]	Speed: 4794.044383 samples/sec	accuracy=0.753691	lr=0.191157
Epoch[64] Batch [199]	Speed: 4959.509306 samples/sec	accuracy=0.753772	lr=0.190939
Epoch[64] Batch [249]	Speed: 5158.671489 samples/sec	accuracy=0.753888	lr=0.190720
Epoch[64] Batch [299]	Speed: 5161.733169 samples/sec	accuracy=0.754051	lr=0.190502
Epoch[64] Batch [349]	Speed: 5262.111759 samples/sec	accuracy=0.753982	lr=0.190284
Epoch[64] Batch [399]	Speed: 5311.288777 samples/sec	accuracy=0.754000	lr=0.190066
Epoch[64] Batch [449]	Speed: 5155.299813 samples/sec	accuracy=0.754248	lr=0.189848
Epoch[64] Batch [499]	Speed: 4931.162622 samples/sec	accuracy=0.754540	lr=0.189630
Epoch[64] Batch [549]	Speed: 5339.164055 samples/sec	accuracy=0.754641	lr=0.189412
Epoch[64] Batch [599]	Speed: 4922.332899 samples/sec	accuracy=0.754766	lr=0.189194
Epoch[64] Batch [649]	Speed: 5414.093025 samples/sec	accuracy=0.755035	lr=0.188976
Epoch[64] Batch [699]	Speed: 5324.147602 samples/sec	accuracy=0.755064	lr=0.188758
Epoch[64] Batch [749]	Speed: 5141.338194 samples/sec	accuracy=0.755164	lr=0.188540
Epoch[64] Batch [799]	Speed: 5140.823484 samples/sec	accuracy=0.755215	lr=0.188322
Epoch[64] Batch [849]	Speed: 5238.143675 samples/sec	accuracy=0.755305	lr=0.188104
Epoch[64] Batch [899]	Speed: 4964.234695 samples/sec	accuracy=0.755337	lr=0.187886
Epoch[64] Batch [949]	Speed: 5279.303456 samples/sec	accuracy=0.755387	lr=0.187668
Epoch[64] Batch [999]	Speed: 5118.383079 samples/sec	accuracy=0.755555	lr=0.187450
Epoch[64] Batch [1049]	Speed: 5188.865005 samples/sec	accuracy=0.755597	lr=0.187232
Epoch[64] Batch [1099]	Speed: 5219.220976 samples/sec	accuracy=0.755663	lr=0.187014
Epoch[64] Batch [1149]	Speed: 5142.515942 samples/sec	accuracy=0.755753	lr=0.186796
Epoch[64] Batch [1199]	Speed: 4992.071201 samples/sec	accuracy=0.755866	lr=0.186578
Epoch[64] Batch [1249]	Speed: 5183.326289 samples/sec	accuracy=0.755963	lr=0.186360
[Epoch 64] training: accuracy=0.755966
[Epoch 64] speed: 5131 samples/sec	time cost: 259.328935
[Epoch 64] validation: err-top1=0.354811 err-top5=0.135324
Epoch[65] Batch [49]	Speed: 4767.866409 samples/sec	accuracy=0.755583	lr=0.186138
Epoch[65] Batch [99]	Speed: 4878.107817 samples/sec	accuracy=0.755175	lr=0.185920
Epoch[65] Batch [149]	Speed: 5169.318428 samples/sec	accuracy=0.754313	lr=0.185702
Epoch[65] Batch [199]	Speed: 5240.832672 samples/sec	accuracy=0.754689	lr=0.185485
Epoch[65] Batch [249]	Speed: 5041.695901 samples/sec	accuracy=0.754958	lr=0.185267
Epoch[65] Batch [299]	Speed: 5027.041471 samples/sec	accuracy=0.754999	lr=0.185049
Epoch[65] Batch [349]	Speed: 5146.416479 samples/sec	accuracy=0.755116	lr=0.184831
Epoch[65] Batch [399]	Speed: 5061.897992 samples/sec	accuracy=0.755393	lr=0.184614
Epoch[65] Batch [449]	Speed: 5123.651650 samples/sec	accuracy=0.755394	lr=0.184396
Epoch[65] Batch [499]	Speed: 5108.931937 samples/sec	accuracy=0.755462	lr=0.184178
Epoch[65] Batch [549]	Speed: 5279.473999 samples/sec	accuracy=0.755448	lr=0.183960
Epoch[65] Batch [599]	Speed: 5285.871714 samples/sec	accuracy=0.755597	lr=0.183743
Epoch[65] Batch [649]	Speed: 5064.532775 samples/sec	accuracy=0.755681	lr=0.183525
Epoch[65] Batch [699]	Speed: 5116.790222 samples/sec	accuracy=0.755665	lr=0.183308
Epoch[65] Batch [749]	Speed: 5067.192661 samples/sec	accuracy=0.755776	lr=0.183090
Epoch[65] Batch [799]	Speed: 5132.661238 samples/sec	accuracy=0.755936	lr=0.182872
Epoch[65] Batch [849]	Speed: 5402.983350 samples/sec	accuracy=0.756025	lr=0.182655
Epoch[65] Batch [899]	Speed: 4933.371520 samples/sec	accuracy=0.756128	lr=0.182437
Epoch[65] Batch [949]	Speed: 5111.443401 samples/sec	accuracy=0.756238	lr=0.182220
Epoch[65] Batch [999]	Speed: 4953.586669 samples/sec	accuracy=0.756396	lr=0.182002
Epoch[65] Batch [1049]	Speed: 5080.989094 samples/sec	accuracy=0.756531	lr=0.181785
Epoch[65] Batch [1099]	Speed: 5099.534040 samples/sec	accuracy=0.756649	lr=0.181567
Epoch[65] Batch [1149]	Speed: 5103.910947 samples/sec	accuracy=0.756712	lr=0.181350
Epoch[65] Batch [1199]	Speed: 5339.130736 samples/sec	accuracy=0.756878	lr=0.181132
Epoch[65] Batch [1249]	Speed: 5285.669274 samples/sec	accuracy=0.756950	lr=0.180915
[Epoch 65] training: accuracy=0.756954
[Epoch 65] speed: 5104 samples/sec	time cost: 260.510730
[Epoch 65] validation: err-top1=0.370217 err-top5=0.144511
Epoch[66] Batch [49]	Speed: 4988.034081 samples/sec	accuracy=0.757179	lr=0.180693
Epoch[66] Batch [99]	Speed: 5099.684446 samples/sec	accuracy=0.757183	lr=0.180476
Epoch[66] Batch [149]	Speed: 5038.537416 samples/sec	accuracy=0.756732	lr=0.180259
Epoch[66] Batch [199]	Speed: 5129.860629 samples/sec	accuracy=0.756648	lr=0.180041
Epoch[66] Batch [249]	Speed: 5109.564161 samples/sec	accuracy=0.756432	lr=0.179824
Epoch[66] Batch [299]	Speed: 5149.241990 samples/sec	accuracy=0.756297	lr=0.179607
Epoch[66] Batch [349]	Speed: 4791.985303 samples/sec	accuracy=0.756351	lr=0.179390
Epoch[66] Batch [399]	Speed: 5303.135845 samples/sec	accuracy=0.756390	lr=0.179172
Epoch[66] Batch [449]	Speed: 5105.669981 samples/sec	accuracy=0.756558	lr=0.178955
Epoch[66] Batch [499]	Speed: 5224.542570 samples/sec	accuracy=0.756644	lr=0.178738
Epoch[66] Batch [549]	Speed: 5132.683442 samples/sec	accuracy=0.756831	lr=0.178521
Epoch[66] Batch [599]	Speed: 5190.838298 samples/sec	accuracy=0.756920	lr=0.178304
Epoch[66] Batch [649]	Speed: 5120.121829 samples/sec	accuracy=0.757111	lr=0.178087
Epoch[66] Batch [699]	Speed: 5233.123470 samples/sec	accuracy=0.757180	lr=0.177870
Epoch[66] Batch [749]	Speed: 5385.636310 samples/sec	accuracy=0.757283	lr=0.177653
Epoch[66] Batch [799]	Speed: 5142.832078 samples/sec	accuracy=0.757325	lr=0.177436
Epoch[66] Batch [849]	Speed: 5069.747145 samples/sec	accuracy=0.757393	lr=0.177219
Epoch[66] Batch [899]	Speed: 5020.357437 samples/sec	accuracy=0.757553	lr=0.177002
Epoch[66] Batch [949]	Speed: 5158.873612 samples/sec	accuracy=0.757671	lr=0.176785
Epoch[66] Batch [999]	Speed: 5201.784001 samples/sec	accuracy=0.757772	lr=0.176568
Epoch[66] Batch [1049]	Speed: 5146.186350 samples/sec	accuracy=0.757911	lr=0.176351
Epoch[66] Batch [1099]	Speed: 5135.186882 samples/sec	accuracy=0.757986	lr=0.176134
Epoch[66] Batch [1149]	Speed: 5230.142636 samples/sec	accuracy=0.758114	lr=0.175918
Epoch[66] Batch [1199]	Speed: 5171.004676 samples/sec	accuracy=0.758297	lr=0.175701
Epoch[66] Batch [1249]	Speed: 5285.972679 samples/sec	accuracy=0.758436	lr=0.175484
[Epoch 66] training: accuracy=0.758438
[Epoch 66] speed: 5135 samples/sec	time cost: 258.562207
[Epoch 66] validation: err-top1=0.358598 err-top5=0.136878
Epoch[67] Batch [49]	Speed: 4753.646452 samples/sec	accuracy=0.758025	lr=0.175263
Epoch[67] Batch [99]	Speed: 4928.706035 samples/sec	accuracy=0.758073	lr=0.175046
Epoch[67] Batch [149]	Speed: 5189.382609 samples/sec	accuracy=0.757174	lr=0.174830
Epoch[67] Batch [199]	Speed: 5217.980703 samples/sec	accuracy=0.757265	lr=0.174613
Epoch[67] Batch [249]	Speed: 5227.148656 samples/sec	accuracy=0.757458	lr=0.174397
Epoch[67] Batch [299]	Speed: 5061.726422 samples/sec	accuracy=0.757584	lr=0.174180
Epoch[67] Batch [349]	Speed: 5084.609007 samples/sec	accuracy=0.757591	lr=0.173964
Epoch[67] Batch [399]	Speed: 5091.304796 samples/sec	accuracy=0.757675	lr=0.173747
Epoch[67] Batch [449]	Speed: 5303.509367 samples/sec	accuracy=0.757556	lr=0.173531
Epoch[67] Batch [499]	Speed: 5073.604919 samples/sec	accuracy=0.757646	lr=0.173314
Epoch[67] Batch [549]	Speed: 5181.472967 samples/sec	accuracy=0.757636	lr=0.173098
Epoch[67] Batch [599]	Speed: 5167.655041 samples/sec	accuracy=0.757670	lr=0.172881
Epoch[67] Batch [649]	Speed: 5081.788544 samples/sec	accuracy=0.757771	lr=0.172665
Epoch[67] Batch [699]	Speed: 5098.880446 samples/sec	accuracy=0.757875	lr=0.172449
Epoch[67] Batch [749]	Speed: 4946.599350 samples/sec	accuracy=0.758022	lr=0.172232
Epoch[67] Batch [799]	Speed: 4902.479518 samples/sec	accuracy=0.758116	lr=0.172016
Epoch[67] Batch [849]	Speed: 5182.875062 samples/sec	accuracy=0.758138	lr=0.171800
Epoch[67] Batch [899]	Speed: 5195.992385 samples/sec	accuracy=0.758327	lr=0.171584
Epoch[67] Batch [949]	Speed: 5171.848773 samples/sec	accuracy=0.758534	lr=0.171368
Epoch[67] Batch [999]	Speed: 5249.117684 samples/sec	accuracy=0.758672	lr=0.171152
Epoch[67] Batch [1049]	Speed: 5288.823946 samples/sec	accuracy=0.758808	lr=0.170935
Epoch[67] Batch [1099]	Speed: 5153.812655 samples/sec	accuracy=0.758979	lr=0.170719
Epoch[67] Batch [1149]	Speed: 5134.109577 samples/sec	accuracy=0.759117	lr=0.170503
Epoch[67] Batch [1199]	Speed: 4977.420298 samples/sec	accuracy=0.759297	lr=0.170287
Epoch[67] Batch [1249]	Speed: 5134.478204 samples/sec	accuracy=0.759415	lr=0.170072
[Epoch 67] training: accuracy=0.759416
[Epoch 67] speed: 5103 samples/sec	time cost: 260.252169
[Epoch 67] validation: err-top1=0.387636 err-top5=0.159359
Epoch[68] Batch [49]	Speed: 4841.090306 samples/sec	accuracy=0.758788	lr=0.169851
Epoch[68] Batch [99]	Speed: 5351.945756 samples/sec	accuracy=0.758051	lr=0.169635
Epoch[68] Batch [149]	Speed: 5080.726553 samples/sec	accuracy=0.757853	lr=0.169420
Epoch[68] Batch [199]	Speed: 5077.758148 samples/sec	accuracy=0.758455	lr=0.169204
Epoch[68] Batch [249]	Speed: 5032.823394 samples/sec	accuracy=0.758696	lr=0.168988
Epoch[68] Batch [299]	Speed: 4969.007269 samples/sec	accuracy=0.758542	lr=0.168772
Epoch[68] Batch [349]	Speed: 5340.898394 samples/sec	accuracy=0.758626	lr=0.168557
Epoch[68] Batch [399]	Speed: 5115.914396 samples/sec	accuracy=0.758863	lr=0.168341
Epoch[68] Batch [449]	Speed: 5102.483716 samples/sec	accuracy=0.759067	lr=0.168126
Epoch[68] Batch [499]	Speed: 5008.981477 samples/sec	accuracy=0.759309	lr=0.167910
Epoch[68] Batch [549]	Speed: 5194.202234 samples/sec	accuracy=0.759361	lr=0.167694
Epoch[68] Batch [599]	Speed: 5423.751246 samples/sec	accuracy=0.759448	lr=0.167479
Epoch[68] Batch [649]	Speed: 5358.911422 samples/sec	accuracy=0.759614	lr=0.167263
Epoch[68] Batch [699]	Speed: 5403.323894 samples/sec	accuracy=0.759740	lr=0.167048
Epoch[68] Batch [749]	Speed: 5161.411604 samples/sec	accuracy=0.759968	lr=0.166833
Epoch[68] Batch [799]	Speed: 5078.642096 samples/sec	accuracy=0.760109	lr=0.166617
Epoch[68] Batch [849]	Speed: 5220.994397 samples/sec	accuracy=0.760208	lr=0.166402
Epoch[68] Batch [899]	Speed: 5141.597311 samples/sec	accuracy=0.760385	lr=0.166187
Epoch[68] Batch [949]	Speed: 5132.850164 samples/sec	accuracy=0.760406	lr=0.165972
Epoch[68] Batch [999]	Speed: 4936.082129 samples/sec	accuracy=0.760454	lr=0.165756
Epoch[68] Batch [1049]	Speed: 5045.533966 samples/sec	accuracy=0.760485	lr=0.165541
Epoch[68] Batch [1099]	Speed: 5286.161480 samples/sec	accuracy=0.760519	lr=0.165326
Epoch[68] Batch [1149]	Speed: 5217.304255 samples/sec	accuracy=0.760611	lr=0.165111
Epoch[68] Batch [1199]	Speed: 5192.329202 samples/sec	accuracy=0.760717	lr=0.164896
Epoch[68] Batch [1249]	Speed: 5133.566983 samples/sec	accuracy=0.760817	lr=0.164681
[Epoch 68] training: accuracy=0.760815
[Epoch 68] speed: 5145 samples/sec	time cost: 258.220982
[Epoch 68] validation: err-top1=0.348214 err-top5=0.131597
Epoch[69] Batch [49]	Speed: 5111.349966 samples/sec	accuracy=0.759563	lr=0.164462
Epoch[69] Batch [99]	Speed: 5121.762814 samples/sec	accuracy=0.759001	lr=0.164247
Epoch[69] Batch [149]	Speed: 5192.450856 samples/sec	accuracy=0.759548	lr=0.164032
Epoch[69] Batch [199]	Speed: 5144.950342 samples/sec	accuracy=0.759825	lr=0.163818
Epoch[69] Batch [249]	Speed: 5223.523504 samples/sec	accuracy=0.759970	lr=0.163603
Epoch[69] Batch [299]	Speed: 5012.964589 samples/sec	accuracy=0.759856	lr=0.163388
Epoch[69] Batch [349]	Speed: 4930.401482 samples/sec	accuracy=0.759873	lr=0.163173
Epoch[69] Batch [399]	Speed: 5203.815304 samples/sec	accuracy=0.759850	lr=0.162959
Epoch[69] Batch [449]	Speed: 5237.908974 samples/sec	accuracy=0.759891	lr=0.162744
Epoch[69] Batch [499]	Speed: 4966.111088 samples/sec	accuracy=0.760076	lr=0.162530
Epoch[69] Batch [549]	Speed: 5071.294796 samples/sec	accuracy=0.760249	lr=0.162315
Epoch[69] Batch [599]	Speed: 5176.701882 samples/sec	accuracy=0.760356	lr=0.162101
Epoch[69] Batch [649]	Speed: 4975.059064 samples/sec	accuracy=0.760412	lr=0.161886
Epoch[69] Batch [699]	Speed: 5163.107841 samples/sec	accuracy=0.760520	lr=0.161672
Epoch[69] Batch [749]	Speed: 5252.537822 samples/sec	accuracy=0.760628	lr=0.161458
Epoch[69] Batch [799]	Speed: 4896.413554 samples/sec	accuracy=0.760713	lr=0.161244
Epoch[69] Batch [849]	Speed: 5108.448241 samples/sec	accuracy=0.760937	lr=0.161029
Epoch[69] Batch [899]	Speed: 5308.482528 samples/sec	accuracy=0.761114	lr=0.160815
Epoch[69] Batch [949]	Speed: 5315.052206 samples/sec	accuracy=0.761226	lr=0.160601
Epoch[69] Batch [999]	Speed: 5072.921763 samples/sec	accuracy=0.761367	lr=0.160387
Epoch[69] Batch [1049]	Speed: 5193.200244 samples/sec	accuracy=0.761514	lr=0.160173
Epoch[69] Batch [1099]	Speed: 5089.119148 samples/sec	accuracy=0.761597	lr=0.159959
Epoch[69] Batch [1149]	Speed: 5138.679596 samples/sec	accuracy=0.761642	lr=0.159745
Epoch[69] Batch [1199]	Speed: 5169.598169 samples/sec	accuracy=0.761745	lr=0.159531
Epoch[69] Batch [1249]	Speed: 5387.265556 samples/sec	accuracy=0.761886	lr=0.159317
[Epoch 69] training: accuracy=0.761891
[Epoch 69] speed: 5131 samples/sec	time cost: 258.560138
[Epoch 69] validation: err-top1=0.343608 err-top5=0.126750
Epoch[70] Batch [49]	Speed: 4952.434355 samples/sec	accuracy=0.762818	lr=0.159099
Epoch[70] Batch [99]	Speed: 5150.614710 samples/sec	accuracy=0.762340	lr=0.158886
Epoch[70] Batch [149]	Speed: 5119.489065 samples/sec	accuracy=0.761705	lr=0.158672
Epoch[70] Batch [199]	Speed: 4990.866234 samples/sec	accuracy=0.761408	lr=0.158458
Epoch[70] Batch [249]	Speed: 5145.799024 samples/sec	accuracy=0.761524	lr=0.158245
Epoch[70] Batch [299]	Speed: 5126.375605 samples/sec	accuracy=0.761323	lr=0.158031
Epoch[70] Batch [349]	Speed: 5206.461084 samples/sec	accuracy=0.761381	lr=0.157818
Epoch[70] Batch [399]	Speed: 4947.147699 samples/sec	accuracy=0.761364	lr=0.157604
Epoch[70] Batch [449]	Speed: 5100.350966 samples/sec	accuracy=0.761473	lr=0.157391
Epoch[70] Batch [499]	Speed: 5191.934774 samples/sec	accuracy=0.761530	lr=0.157177
Epoch[70] Batch [549]	Speed: 5154.379208 samples/sec	accuracy=0.761458	lr=0.156964
Epoch[70] Batch [599]	Speed: 4975.394599 samples/sec	accuracy=0.761617	lr=0.156751
Epoch[70] Batch [649]	Speed: 5100.632863 samples/sec	accuracy=0.761601	lr=0.156538
Epoch[70] Batch [699]	Speed: 5157.075753 samples/sec	accuracy=0.761735	lr=0.156325
Epoch[70] Batch [749]	Speed: 5294.268267 samples/sec	accuracy=0.761822	lr=0.156112
Epoch[70] Batch [799]	Speed: 5331.194173 samples/sec	accuracy=0.762009	lr=0.155899
Epoch[70] Batch [849]	Speed: 5126.555869 samples/sec	accuracy=0.762053	lr=0.155686
Epoch[70] Batch [899]	Speed: 5166.771787 samples/sec	accuracy=0.762106	lr=0.155473
Epoch[70] Batch [949]	Speed: 5170.763503 samples/sec	accuracy=0.762196	lr=0.155260
Epoch[70] Batch [999]	Speed: 5060.388146 samples/sec	accuracy=0.762392	lr=0.155047
Epoch[70] Batch [1049]	Speed: 5269.395764 samples/sec	accuracy=0.762455	lr=0.154834
Epoch[70] Batch [1099]	Speed: 5231.629825 samples/sec	accuracy=0.762524	lr=0.154622
Epoch[70] Batch [1149]	Speed: 5133.964988 samples/sec	accuracy=0.762631	lr=0.154409
Epoch[70] Batch [1199]	Speed: 4983.600287 samples/sec	accuracy=0.762787	lr=0.154196
Epoch[70] Batch [1249]	Speed: 5171.943935 samples/sec	accuracy=0.762927	lr=0.153984
[Epoch 70] training: accuracy=0.762931
[Epoch 70] speed: 5124 samples/sec	time cost: 259.224770
[Epoch 70] validation: err-top1=0.383092 err-top5=0.159220
Epoch[71] Batch [49]	Speed: 4765.444263 samples/sec	accuracy=0.763706	lr=0.153767
Epoch[71] Batch [99]	Speed: 5258.119174 samples/sec	accuracy=0.762620	lr=0.153555
Epoch[71] Batch [149]	Speed: 5070.381799 samples/sec	accuracy=0.762349	lr=0.153342
Epoch[71] Batch [199]	Speed: 5160.224442 samples/sec	accuracy=0.761939	lr=0.153130
Epoch[71] Batch [249]	Speed: 5142.437992 samples/sec	accuracy=0.762115	lr=0.152918
Epoch[71] Batch [299]	Speed: 5166.811815 samples/sec	accuracy=0.762178	lr=0.152705
Epoch[71] Batch [349]	Speed: 4881.225520 samples/sec	accuracy=0.762383	lr=0.152493
Epoch[71] Batch [399]	Speed: 4941.277044 samples/sec	accuracy=0.762449	lr=0.152281
Epoch[71] Batch [449]	Speed: 5222.462670 samples/sec	accuracy=0.762484	lr=0.152069
Epoch[71] Batch [499]	Speed: 4998.395609 samples/sec	accuracy=0.762633	lr=0.151857
Epoch[71] Batch [549]	Speed: 5096.562978 samples/sec	accuracy=0.762889	lr=0.151645
Epoch[71] Batch [599]	Speed: 5234.760498 samples/sec	accuracy=0.763019	lr=0.151433
Epoch[71] Batch [649]	Speed: 5057.340071 samples/sec	accuracy=0.762989	lr=0.151222
Epoch[71] Batch [699]	Speed: 5178.490103 samples/sec	accuracy=0.763001	lr=0.151010
Epoch[71] Batch [749]	Speed: 5054.350505 samples/sec	accuracy=0.763158	lr=0.150798
Epoch[71] Batch [799]	Speed: 5175.901360 samples/sec	accuracy=0.763272	lr=0.150587
Epoch[71] Batch [849]	Speed: 5310.021829 samples/sec	accuracy=0.763410	lr=0.150375
Epoch[71] Batch [899]	Speed: 4965.129263 samples/sec	accuracy=0.763645	lr=0.150163
Epoch[71] Batch [949]	Speed: 5250.720829 samples/sec	accuracy=0.763758	lr=0.149952
Epoch[71] Batch [999]	Speed: 5183.750317 samples/sec	accuracy=0.763786	lr=0.149741
Epoch[71] Batch [1049]	Speed: 5339.318441 samples/sec	accuracy=0.763895	lr=0.149529
Epoch[71] Batch [1099]	Speed: 5223.764796 samples/sec	accuracy=0.763983	lr=0.149318
Epoch[71] Batch [1149]	Speed: 5324.041213 samples/sec	accuracy=0.764155	lr=0.149107
Epoch[71] Batch [1199]	Speed: 5265.816901 samples/sec	accuracy=0.764318	lr=0.148896
Epoch[71] Batch [1249]	Speed: 5325.180826 samples/sec	accuracy=0.764442	lr=0.148685
[Epoch 71] training: accuracy=0.764450
[Epoch 71] speed: 5135 samples/sec	time cost: 258.904552
[Epoch 71] validation: err-top1=0.336137 err-top5=0.126156
Epoch[72] Batch [49]	Speed: 5003.481839 samples/sec	accuracy=0.765126	lr=0.148469
Epoch[72] Batch [99]	Speed: 5106.782137 samples/sec	accuracy=0.764691	lr=0.148258
Epoch[72] Batch [149]	Speed: 5246.669233 samples/sec	accuracy=0.763852	lr=0.148047
Epoch[72] Batch [199]	Speed: 5171.366665 samples/sec	accuracy=0.763530	lr=0.147837
Epoch[72] Batch [249]	Speed: 4950.799052 samples/sec	accuracy=0.763293	lr=0.147626
Epoch[72] Batch [299]	Speed: 5043.252290 samples/sec	accuracy=0.763587	lr=0.147415
Epoch[72] Batch [349]	Speed: 5205.248443 samples/sec	accuracy=0.763681	lr=0.147204
Epoch[72] Batch [399]	Speed: 5068.164913 samples/sec	accuracy=0.763710	lr=0.146994
Epoch[72] Batch [449]	Speed: 5189.402548 samples/sec	accuracy=0.763850	lr=0.146783
Epoch[72] Batch [499]	Speed: 5031.349583 samples/sec	accuracy=0.764017	lr=0.146573
Epoch[72] Batch [549]	Speed: 5321.234133 samples/sec	accuracy=0.764134	lr=0.146362
Epoch[72] Batch [599]	Speed: 4948.123907 samples/sec	accuracy=0.764150	lr=0.146152
Epoch[72] Batch [649]	Speed: 5339.646890 samples/sec	accuracy=0.764289	lr=0.145942
Epoch[72] Batch [699]	Speed: 5072.444263 samples/sec	accuracy=0.764382	lr=0.145732
Epoch[72] Batch [749]	Speed: 4926.096968 samples/sec	accuracy=0.764392	lr=0.145521
Epoch[72] Batch [799]	Speed: 5134.527063 samples/sec	accuracy=0.764525	lr=0.145311
Epoch[72] Batch [849]	Speed: 4819.746176 samples/sec	accuracy=0.764683	lr=0.145101
Epoch[72] Batch [899]	Speed: 5111.900650 samples/sec	accuracy=0.764801	lr=0.144891
Epoch[72] Batch [949]	Speed: 5029.436887 samples/sec	accuracy=0.764884	lr=0.144682
Epoch[72] Batch [999]	Speed: 5340.715758 samples/sec	accuracy=0.764936	lr=0.144472
Epoch[72] Batch [1049]	Speed: 4850.062475 samples/sec	accuracy=0.765019	lr=0.144262
Epoch[72] Batch [1099]	Speed: 5287.296388 samples/sec	accuracy=0.765101	lr=0.144052
Epoch[72] Batch [1149]	Speed: 5233.374322 samples/sec	accuracy=0.765178	lr=0.143843
Epoch[72] Batch [1199]	Speed: 5266.333054 samples/sec	accuracy=0.765237	lr=0.143633
Epoch[72] Batch [1249]	Speed: 5251.447449 samples/sec	accuracy=0.765351	lr=0.143424
[Epoch 72] training: accuracy=0.765354
[Epoch 72] speed: 5108 samples/sec	time cost: 259.930174
[Epoch 72] validation: err-top1=0.346321 err-top5=0.132135
Epoch[73] Batch [49]	Speed: 4968.486019 samples/sec	accuracy=0.765935	lr=0.143210
Epoch[73] Batch [99]	Speed: 5007.071970 samples/sec	accuracy=0.765204	lr=0.143001
Epoch[73] Batch [149]	Speed: 5032.775389 samples/sec	accuracy=0.765608	lr=0.142791
Epoch[73] Batch [199]	Speed: 5209.092119 samples/sec	accuracy=0.765097	lr=0.142582
Epoch[73] Batch [249]	Speed: 5111.558253 samples/sec	accuracy=0.764962	lr=0.142373
Epoch[73] Batch [299]	Speed: 5189.710805 samples/sec	accuracy=0.764890	lr=0.142164
Epoch[73] Batch [349]	Speed: 5235.836164 samples/sec	accuracy=0.764982	lr=0.141955
Epoch[73] Batch [399]	Speed: 5071.208451 samples/sec	accuracy=0.764951	lr=0.141746
Epoch[73] Batch [449]	Speed: 5095.206704 samples/sec	accuracy=0.765050	lr=0.141537
Epoch[73] Batch [499]	Speed: 5340.080149 samples/sec	accuracy=0.765232	lr=0.141328
Epoch[73] Batch [549]	Speed: 5354.127422 samples/sec	accuracy=0.765285	lr=0.141120
Epoch[73] Batch [599]	Speed: 4941.175856 samples/sec	accuracy=0.765270	lr=0.140911
Epoch[73] Batch [649]	Speed: 5076.209550 samples/sec	accuracy=0.765278	lr=0.140702
Epoch[73] Batch [699]	Speed: 5483.491343 samples/sec	accuracy=0.765359	lr=0.140494
Epoch[73] Batch [749]	Speed: 5251.229789 samples/sec	accuracy=0.765402	lr=0.140285
Epoch[73] Batch [799]	Speed: 5090.723784 samples/sec	accuracy=0.765497	lr=0.140077
Epoch[73] Batch [849]	Speed: 5185.148886 samples/sec	accuracy=0.765653	lr=0.139869
Epoch[73] Batch [899]	Speed: 5223.051405 samples/sec	accuracy=0.765781	lr=0.139661
Epoch[73] Batch [949]	Speed: 5441.428369 samples/sec	accuracy=0.765890	lr=0.139452
Epoch[73] Batch [999]	Speed: 5379.036062 samples/sec	accuracy=0.766009	lr=0.139244
Epoch[73] Batch [1049]	Speed: 5133.202659 samples/sec	accuracy=0.766157	lr=0.139036
Epoch[73] Batch [1099]	Speed: 5107.487198 samples/sec	accuracy=0.766248	lr=0.138828
Epoch[73] Batch [1149]	Speed: 5223.753614 samples/sec	accuracy=0.766350	lr=0.138620
Epoch[73] Batch [1199]	Speed: 5152.478161 samples/sec	accuracy=0.766417	lr=0.138413
Epoch[73] Batch [1249]	Speed: 5278.859370 samples/sec	accuracy=0.766501	lr=0.138205
[Epoch 73] training: accuracy=0.766502
[Epoch 73] speed: 5175 samples/sec	time cost: 256.620528
[Epoch 73] validation: err-top1=0.353436 err-top5=0.135623
Epoch[74] Batch [49]	Speed: 4943.498314 samples/sec	accuracy=0.764439	lr=0.137993
Epoch[74] Batch [99]	Speed: 5122.732414 samples/sec	accuracy=0.764635	lr=0.137786
Epoch[74] Batch [149]	Speed: 4949.846090 samples/sec	accuracy=0.765105	lr=0.137578
Epoch[74] Batch [199]	Speed: 5063.443361 samples/sec	accuracy=0.765404	lr=0.137371
Epoch[74] Batch [249]	Speed: 4911.392988 samples/sec	accuracy=0.765398	lr=0.137163
Epoch[74] Batch [299]	Speed: 5156.103013 samples/sec	accuracy=0.765626	lr=0.136956
Epoch[74] Batch [349]	Speed: 5211.013432 samples/sec	accuracy=0.765769	lr=0.136749
Epoch[74] Batch [399]	Speed: 5246.074649 samples/sec	accuracy=0.765940	lr=0.136542
Epoch[74] Batch [449]	Speed: 5029.218043 samples/sec	accuracy=0.765939	lr=0.136335
Epoch[74] Batch [499]	Speed: 5156.026136 samples/sec	accuracy=0.765958	lr=0.136128
Epoch[74] Batch [549]	Speed: 5200.959202 samples/sec	accuracy=0.766103	lr=0.135921
Epoch[74] Batch [599]	Speed: 5045.452408 samples/sec	accuracy=0.766248	lr=0.135714
Epoch[74] Batch [649]	Speed: 4920.236455 samples/sec	accuracy=0.766437	lr=0.135507
Epoch[74] Batch [699]	Speed: 4990.369612 samples/sec	accuracy=0.766544	lr=0.135301
Epoch[74] Batch [749]	Speed: 4876.760758 samples/sec	accuracy=0.766679	lr=0.135094
Epoch[74] Batch [799]	Speed: 4996.262600 samples/sec	accuracy=0.766816	lr=0.134887
Epoch[74] Batch [849]	Speed: 4981.953124 samples/sec	accuracy=0.766944	lr=0.134681
Epoch[74] Batch [899]	Speed: 5114.321978 samples/sec	accuracy=0.767081	lr=0.134475
Epoch[74] Batch [949]	Speed: 5335.327443 samples/sec	accuracy=0.767081	lr=0.134268
Epoch[74] Batch [999]	Speed: 5104.533678 samples/sec	accuracy=0.767193	lr=0.134062
Epoch[74] Batch [1049]	Speed: 4952.739545 samples/sec	accuracy=0.767272	lr=0.133856
Epoch[74] Batch [1099]	Speed: 5129.538857 samples/sec	accuracy=0.767294	lr=0.133650
Epoch[74] Batch [1149]	Speed: 4918.658732 samples/sec	accuracy=0.767458	lr=0.133444
Epoch[74] Batch [1199]	Speed: 5054.407012 samples/sec	accuracy=0.767607	lr=0.133238
Epoch[74] Batch [1249]	Speed: 5202.397070 samples/sec	accuracy=0.767727	lr=0.133032
[Epoch 74] training: accuracy=0.767728
[Epoch 74] speed: 5056 samples/sec	time cost: 262.560329
[Epoch 74] validation: err-top1=0.336655 err-top5=0.123545
Epoch[75] Batch [49]	Speed: 4577.232695 samples/sec	accuracy=0.767920	lr=0.132822
Epoch[75] Batch [99]	Speed: 5304.690261 samples/sec	accuracy=0.768523	lr=0.132617
Epoch[75] Batch [149]	Speed: 5031.247384 samples/sec	accuracy=0.767978	lr=0.132411
Epoch[75] Batch [199]	Speed: 5084.076945 samples/sec	accuracy=0.767844	lr=0.132206
Epoch[75] Batch [249]	Speed: 4949.053737 samples/sec	accuracy=0.767524	lr=0.132000
Epoch[75] Batch [299]	Speed: 5224.641206 samples/sec	accuracy=0.767513	lr=0.131795
Epoch[75] Batch [349]	Speed: 4970.817774 samples/sec	accuracy=0.767521	lr=0.131590
Epoch[75] Batch [399]	Speed: 5051.344154 samples/sec	accuracy=0.767481	lr=0.131385
Epoch[75] Batch [449]	Speed: 5012.180913 samples/sec	accuracy=0.767650	lr=0.131180
Epoch[75] Batch [499]	Speed: 5415.636968 samples/sec	accuracy=0.767616	lr=0.130975
Epoch[75] Batch [549]	Speed: 5073.827164 samples/sec	accuracy=0.767702	lr=0.130770
Epoch[75] Batch [599]	Speed: 5128.560064 samples/sec	accuracy=0.767890	lr=0.130565
Epoch[75] Batch [649]	Speed: 4870.901634 samples/sec	accuracy=0.767984	lr=0.130360
Epoch[75] Batch [699]	Speed: 5031.069281 samples/sec	accuracy=0.768118	lr=0.130155
Epoch[75] Batch [749]	Speed: 5111.095956 samples/sec	accuracy=0.768293	lr=0.129951
Epoch[75] Batch [799]	Speed: 5128.027949 samples/sec	accuracy=0.768442	lr=0.129746
Epoch[75] Batch [849]	Speed: 4878.795258 samples/sec	accuracy=0.768481	lr=0.129542
Epoch[75] Batch [899]	Speed: 5030.880466 samples/sec	accuracy=0.768491	lr=0.129338
Epoch[75] Batch [949]	Speed: 5398.430042 samples/sec	accuracy=0.768614	lr=0.129133
Epoch[75] Batch [999]	Speed: 5085.917245 samples/sec	accuracy=0.768731	lr=0.128929
Epoch[75] Batch [1049]	Speed: 5010.266044 samples/sec	accuracy=0.768804	lr=0.128725
Epoch[75] Batch [1099]	Speed: 4999.613410 samples/sec	accuracy=0.768905	lr=0.128521
Epoch[75] Batch [1149]	Speed: 5098.365002 samples/sec	accuracy=0.768984	lr=0.128317
Epoch[75] Batch [1199]	Speed: 4964.722340 samples/sec	accuracy=0.769081	lr=0.128113
Epoch[75] Batch [1249]	Speed: 5319.233994 samples/sec	accuracy=0.769175	lr=0.127910
[Epoch 75] training: accuracy=0.769179
[Epoch 75] speed: 5059 samples/sec	time cost: 262.467053
[Epoch 75] validation: err-top1=0.350606 err-top5=0.135864
Epoch[76] Batch [49]	Speed: 4957.092141 samples/sec	accuracy=0.770286	lr=0.127702
Epoch[76] Batch [99]	Speed: 5260.589796 samples/sec	accuracy=0.769050	lr=0.127498
Epoch[76] Batch [149]	Speed: 4921.817560 samples/sec	accuracy=0.767921	lr=0.127295
Epoch[76] Batch [199]	Speed: 4867.913622 samples/sec	accuracy=0.767962	lr=0.127092
Epoch[76] Batch [249]	Speed: 5112.020147 samples/sec	accuracy=0.768222	lr=0.126888
Epoch[76] Batch [299]	Speed: 5170.340601 samples/sec	accuracy=0.768259	lr=0.126685
Epoch[76] Batch [349]	Speed: 5510.906206 samples/sec	accuracy=0.768283	lr=0.126482
Epoch[76] Batch [399]	Speed: 5487.787664 samples/sec	accuracy=0.768340	lr=0.126279
Epoch[76] Batch [449]	Speed: 5304.252245 samples/sec	accuracy=0.768535	lr=0.126076
Epoch[76] Batch [499]	Speed: 5012.584304 samples/sec	accuracy=0.768658	lr=0.125873
Epoch[76] Batch [549]	Speed: 5232.715297 samples/sec	accuracy=0.768723	lr=0.125670
Epoch[76] Batch [599]	Speed: 5226.545768 samples/sec	accuracy=0.768803	lr=0.125468
Epoch[76] Batch [649]	Speed: 5194.161655 samples/sec	accuracy=0.768907	lr=0.125265
Epoch[76] Batch [699]	Speed: 5151.692774 samples/sec	accuracy=0.768919	lr=0.125062
Epoch[76] Batch [749]	Speed: 5378.467677 samples/sec	accuracy=0.768981	lr=0.124860
Epoch[76] Batch [799]	Speed: 5198.290949 samples/sec	accuracy=0.769040	lr=0.124658
Epoch[76] Batch [849]	Speed: 5005.725797 samples/sec	accuracy=0.769029	lr=0.124455
Epoch[76] Batch [899]	Speed: 5164.954133 samples/sec	accuracy=0.769158	lr=0.124253
Epoch[76] Batch [949]	Speed: 5129.449414 samples/sec	accuracy=0.769284	lr=0.124051
Epoch[76] Batch [999]	Speed: 5281.234449 samples/sec	accuracy=0.769393	lr=0.123849
Epoch[76] Batch [1049]	Speed: 5343.747540 samples/sec	accuracy=0.769520	lr=0.123647
Epoch[76] Batch [1099]	Speed: 4772.185884 samples/sec	accuracy=0.769621	lr=0.123446
Epoch[76] Batch [1149]	Speed: 5365.613027 samples/sec	accuracy=0.769740	lr=0.123244
Epoch[76] Batch [1199]	Speed: 5064.370342 samples/sec	accuracy=0.769895	lr=0.123042
Epoch[76] Batch [1249]	Speed: 5256.775032 samples/sec	accuracy=0.769919	lr=0.122841
[Epoch 76] training: accuracy=0.769922
[Epoch 76] speed: 5164 samples/sec	time cost: 257.483253
[Epoch 76] validation: err-top1=0.327368 err-top5=0.117327
Epoch[77] Batch [49]	Speed: 4633.459390 samples/sec	accuracy=0.770220	lr=0.122635
Epoch[77] Batch [99]	Speed: 4829.551295 samples/sec	accuracy=0.768601	lr=0.122434
Epoch[77] Batch [149]	Speed: 4947.139950 samples/sec	accuracy=0.768816	lr=0.122233
Epoch[77] Batch [199]	Speed: 5197.832331 samples/sec	accuracy=0.769060	lr=0.122032
Epoch[77] Batch [249]	Speed: 5125.879424 samples/sec	accuracy=0.769061	lr=0.121831
Epoch[77] Batch [299]	Speed: 5233.173715 samples/sec	accuracy=0.768952	lr=0.121630
Epoch[77] Batch [349]	Speed: 5540.923373 samples/sec	accuracy=0.768958	lr=0.121429
Epoch[77] Batch [399]	Speed: 5560.784598 samples/sec	accuracy=0.769100	lr=0.121228
Epoch[77] Batch [449]	Speed: 5244.695797 samples/sec	accuracy=0.769367	lr=0.121027
Epoch[77] Batch [499]	Speed: 5115.777899 samples/sec	accuracy=0.769611	lr=0.120827
Epoch[77] Batch [549]	Speed: 5222.886139 samples/sec	accuracy=0.769725	lr=0.120626
Epoch[77] Batch [599]	Speed: 5197.158581 samples/sec	accuracy=0.769842	lr=0.120426
Epoch[77] Batch [649]	Speed: 5096.601079 samples/sec	accuracy=0.770080	lr=0.120226
Epoch[77] Batch [699]	Speed: 5128.774288 samples/sec	accuracy=0.770160	lr=0.120025
Epoch[77] Batch [749]	Speed: 5172.448701 samples/sec	accuracy=0.770276	lr=0.119825
Epoch[77] Batch [799]	Speed: 5130.117733 samples/sec	accuracy=0.770390	lr=0.119625
Epoch[77] Batch [849]	Speed: 5242.917504 samples/sec	accuracy=0.770558	lr=0.119425
Epoch[77] Batch [899]	Speed: 5252.836279 samples/sec	accuracy=0.770658	lr=0.119226
Epoch[77] Batch [949]	Speed: 4804.367030 samples/sec	accuracy=0.770813	lr=0.119026
Epoch[77] Batch [999]	Speed: 4982.144526 samples/sec	accuracy=0.770908	lr=0.118826
Epoch[77] Batch [1049]	Speed: 5153.092150 samples/sec	accuracy=0.771016	lr=0.118627
Epoch[77] Batch [1099]	Speed: 5259.466838 samples/sec	accuracy=0.771157	lr=0.118427
Epoch[77] Batch [1149]	Speed: 5054.673859 samples/sec	accuracy=0.771265	lr=0.118228
Epoch[77] Batch [1199]	Speed: 5216.224152 samples/sec	accuracy=0.771318	lr=0.118029
Epoch[77] Batch [1249]	Speed: 5417.260231 samples/sec	accuracy=0.771385	lr=0.117830
[Epoch 77] training: accuracy=0.771383
[Epoch 77] speed: 5137 samples/sec	time cost: 258.697880
[Epoch 77] validation: err-top1=0.318778 err-top5=0.112285
Epoch[78] Batch [49]	Speed: 5039.603247 samples/sec	accuracy=0.770837	lr=0.117627
Epoch[78] Batch [99]	Speed: 4999.425901 samples/sec	accuracy=0.770945	lr=0.117428
Epoch[78] Batch [149]	Speed: 5124.687026 samples/sec	accuracy=0.770709	lr=0.117229
Epoch[78] Batch [199]	Speed: 5312.753079 samples/sec	accuracy=0.770734	lr=0.117030
Epoch[78] Batch [249]	Speed: 5420.823152 samples/sec	accuracy=0.770674	lr=0.116831
Epoch[78] Batch [299]	Speed: 5205.974521 samples/sec	accuracy=0.770558	lr=0.116633
Epoch[78] Batch [349]	Speed: 5112.883319 samples/sec	accuracy=0.770491	lr=0.116434
Epoch[78] Batch [399]	Speed: 4983.654644 samples/sec	accuracy=0.770591	lr=0.116236
Epoch[78] Batch [449]	Speed: 5410.900890 samples/sec	accuracy=0.770719	lr=0.116038
Epoch[78] Batch [499]	Speed: 5225.134952 samples/sec	accuracy=0.770821	lr=0.115840
Epoch[78] Batch [549]	Speed: 5224.228001 samples/sec	accuracy=0.770934	lr=0.115642
Epoch[78] Batch [599]	Speed: 5184.134492 samples/sec	accuracy=0.770992	lr=0.115444
Epoch[78] Batch [649]	Speed: 4982.523213 samples/sec	accuracy=0.771200	lr=0.115246
Epoch[78] Batch [699]	Speed: 5214.741403 samples/sec	accuracy=0.771209	lr=0.115048
Epoch[78] Batch [749]	Speed: 5073.931701 samples/sec	accuracy=0.771286	lr=0.114850
Epoch[78] Batch [799]	Speed: 4984.332014 samples/sec	accuracy=0.771458	lr=0.114653
Epoch[78] Batch [849]	Speed: 4904.217892 samples/sec	accuracy=0.771507	lr=0.114455
Epoch[78] Batch [899]	Speed: 5150.726140 samples/sec	accuracy=0.771603	lr=0.114258
Epoch[78] Batch [949]	Speed: 5176.660702 samples/sec	accuracy=0.771641	lr=0.114061
Epoch[78] Batch [999]	Speed: 5343.025863 samples/sec	accuracy=0.771754	lr=0.113864
Epoch[78] Batch [1049]	Speed: 5110.836741 samples/sec	accuracy=0.771856	lr=0.113667
Epoch[78] Batch [1099]	Speed: 4953.528737 samples/sec	accuracy=0.771984	lr=0.113470
Epoch[78] Batch [1149]	Speed: 5042.495347 samples/sec	accuracy=0.772119	lr=0.113273
Epoch[78] Batch [1199]	Speed: 5216.697173 samples/sec	accuracy=0.772253	lr=0.113076
Epoch[78] Batch [1249]	Speed: 5318.680415 samples/sec	accuracy=0.772359	lr=0.112880
[Epoch 78] training: accuracy=0.772364
[Epoch 78] speed: 5140 samples/sec	time cost: 258.536714
[Epoch 78] validation: err-top1=0.335399 err-top5=0.124043
Epoch[79] Batch [49]	Speed: 4793.818470 samples/sec	accuracy=0.772140	lr=0.112679
Epoch[79] Batch [99]	Speed: 4915.543899 samples/sec	accuracy=0.772575	lr=0.112483
Epoch[79] Batch [149]	Speed: 5075.286626 samples/sec	accuracy=0.772369	lr=0.112287
Epoch[79] Batch [199]	Speed: 5130.766489 samples/sec	accuracy=0.771971	lr=0.112090
Epoch[79] Batch [249]	Speed: 5113.499231 samples/sec	accuracy=0.771784	lr=0.111894
Epoch[79] Batch [299]	Speed: 5120.767083 samples/sec	accuracy=0.771844	lr=0.111698
Epoch[79] Batch [349]	Speed: 5371.810132 samples/sec	accuracy=0.771890	lr=0.111502
Epoch[79] Batch [399]	Speed: 5227.362289 samples/sec	accuracy=0.771954	lr=0.111307
Epoch[79] Batch [449]	Speed: 5057.093902 samples/sec	accuracy=0.772034	lr=0.111111
Epoch[79] Batch [499]	Speed: 5293.991706 samples/sec	accuracy=0.771978	lr=0.110915
Epoch[79] Batch [549]	Speed: 5334.286834 samples/sec	accuracy=0.772037	lr=0.110720
Epoch[79] Batch [599]	Speed: 4977.853421 samples/sec	accuracy=0.772170	lr=0.110525
Epoch[79] Batch [649]	Speed: 5076.911715 samples/sec	accuracy=0.772242	lr=0.110329
Epoch[79] Batch [699]	Speed: 5246.877285 samples/sec	accuracy=0.772264	lr=0.110134
Epoch[79] Batch [749]	Speed: 5366.402104 samples/sec	accuracy=0.772403	lr=0.109939
Epoch[79] Batch [799]	Speed: 5263.038878 samples/sec	accuracy=0.772529	lr=0.109744
Epoch[79] Batch [849]	Speed: 5154.482884 samples/sec	accuracy=0.772700	lr=0.109549
Epoch[79] Batch [899]	Speed: 5230.483524 samples/sec	accuracy=0.772818	lr=0.109355
Epoch[79] Batch [949]	Speed: 5297.959507 samples/sec	accuracy=0.772978	lr=0.109160
Epoch[79] Batch [999]	Speed: 5417.795567 samples/sec	accuracy=0.773143	lr=0.108966
Epoch[79] Batch [1049]	Speed: 5136.669575 samples/sec	accuracy=0.773242	lr=0.108771
Epoch[79] Batch [1099]	Speed: 5278.338035 samples/sec	accuracy=0.773418	lr=0.108577
Epoch[79] Batch [1149]	Speed: 5433.089464 samples/sec	accuracy=0.773529	lr=0.108383
Epoch[79] Batch [1199]	Speed: 5241.148732 samples/sec	accuracy=0.773630	lr=0.108189
Epoch[79] Batch [1249]	Speed: 5176.724968 samples/sec	accuracy=0.773714	lr=0.107995
[Epoch 79] training: accuracy=0.773716
[Epoch 79] speed: 5180 samples/sec	time cost: 256.301605
[Epoch 79] validation: err-top1=0.313995 err-top5=0.112743
Epoch[80] Batch [49]	Speed: 4968.902527 samples/sec	accuracy=0.773964	lr=0.107797
Epoch[80] Batch [99]	Speed: 5278.083373 samples/sec	accuracy=0.773089	lr=0.107603
Epoch[80] Batch [149]	Speed: 5000.480488 samples/sec	accuracy=0.773287	lr=0.107410
Epoch[80] Batch [199]	Speed: 5296.113831 samples/sec	accuracy=0.773437	lr=0.107216
Epoch[80] Batch [249]	Speed: 5078.653146 samples/sec	accuracy=0.773738	lr=0.107023
Epoch[80] Batch [299]	Speed: 5155.469617 samples/sec	accuracy=0.773655	lr=0.106830
Epoch[80] Batch [349]	Speed: 4951.440577 samples/sec	accuracy=0.773553	lr=0.106636
Epoch[80] Batch [399]	Speed: 5047.814741 samples/sec	accuracy=0.773646	lr=0.106443
Epoch[80] Batch [449]	Speed: 5068.278666 samples/sec	accuracy=0.773730	lr=0.106250
Epoch[80] Batch [499]	Speed: 5317.565045 samples/sec	accuracy=0.773752	lr=0.106057
Epoch[80] Batch [549]	Speed: 5093.454992 samples/sec	accuracy=0.773789	lr=0.105865
Epoch[80] Batch [599]	Speed: 5166.237306 samples/sec	accuracy=0.773913	lr=0.105672
Epoch[80] Batch [649]	Speed: 5201.287856 samples/sec	accuracy=0.773903	lr=0.105480
Epoch[80] Batch [699]	Speed: 5306.532610 samples/sec	accuracy=0.774041	lr=0.105287
Epoch[80] Batch [749]	Speed: 5159.014526 samples/sec	accuracy=0.774194	lr=0.105095
Epoch[80] Batch [799]	Speed: 5167.649818 samples/sec	accuracy=0.774320	lr=0.104903
Epoch[80] Batch [849]	Speed: 5128.404398 samples/sec	accuracy=0.774349	lr=0.104711
Epoch[80] Batch [899]	Speed: 5168.358230 samples/sec	accuracy=0.774340	lr=0.104519
Epoch[80] Batch [949]	Speed: 4966.644015 samples/sec	accuracy=0.774417	lr=0.104327
Epoch[80] Batch [999]	Speed: 5244.195146 samples/sec	accuracy=0.774481	lr=0.104135
Epoch[80] Batch [1049]	Speed: 5063.761789 samples/sec	accuracy=0.774614	lr=0.103944
Epoch[80] Batch [1099]	Speed: 5136.127545 samples/sec	accuracy=0.774689	lr=0.103752
Epoch[80] Batch [1149]	Speed: 5234.802863 samples/sec	accuracy=0.774738	lr=0.103561
Epoch[80] Batch [1199]	Speed: 5382.837453 samples/sec	accuracy=0.774837	lr=0.103370
Epoch[80] Batch [1249]	Speed: 5352.580857 samples/sec	accuracy=0.774935	lr=0.103178
[Epoch 80] training: accuracy=0.774939
[Epoch 80] speed: 5150 samples/sec	time cost: 257.882873
[Epoch 80] validation: err-top1=0.324597 err-top5=0.117168
Epoch[81] Batch [49]	Speed: 4762.469494 samples/sec	accuracy=0.774297	lr=0.102984
Epoch[81] Batch [99]	Speed: 5183.739806 samples/sec	accuracy=0.774375	lr=0.102793
Epoch[81] Batch [149]	Speed: 4893.792718 samples/sec	accuracy=0.774369	lr=0.102602
Epoch[81] Batch [199]	Speed: 5011.737001 samples/sec	accuracy=0.773875	lr=0.102411
Epoch[81] Batch [249]	Speed: 5017.983909 samples/sec	accuracy=0.773994	lr=0.102221
Epoch[81] Batch [299]	Speed: 5100.881351 samples/sec	accuracy=0.774064	lr=0.102030
Epoch[81] Batch [349]	Speed: 5085.533759 samples/sec	accuracy=0.774253	lr=0.101840
Epoch[81] Batch [399]	Speed: 4999.714328 samples/sec	accuracy=0.774409	lr=0.101650
Epoch[81] Batch [449]	Speed: 5335.329697 samples/sec	accuracy=0.774286	lr=0.101460
Epoch[81] Batch [499]	Speed: 5257.738631 samples/sec	accuracy=0.774259	lr=0.101270
Epoch[81] Batch [549]	Speed: 5140.765151 samples/sec	accuracy=0.774425	lr=0.101080
Epoch[81] Batch [599]	Speed: 4976.928656 samples/sec	accuracy=0.774462	lr=0.100890
Epoch[81] Batch [649]	Speed: 5086.465956 samples/sec	accuracy=0.774465	lr=0.100701
Epoch[81] Batch [699]	Speed: 5080.005547 samples/sec	accuracy=0.774537	lr=0.100511
Epoch[81] Batch [749]	Speed: 4987.312153 samples/sec	accuracy=0.774525	lr=0.100322
Epoch[81] Batch [799]	Speed: 5038.349931 samples/sec	accuracy=0.774545	lr=0.100132
Epoch[81] Batch [849]	Speed: 5174.480094 samples/sec	accuracy=0.774611	lr=0.099943
Epoch[81] Batch [899]	Speed: 5087.511666 samples/sec	accuracy=0.774719	lr=0.099754
Epoch[81] Batch [949]	Speed: 5119.505907 samples/sec	accuracy=0.774872	lr=0.099565
Epoch[81] Batch [999]	Speed: 5319.382750 samples/sec	accuracy=0.775001	lr=0.099377
Epoch[81] Batch [1049]	Speed: 5156.575098 samples/sec	accuracy=0.775147	lr=0.099188
Epoch[81] Batch [1099]	Speed: 5045.533492 samples/sec	accuracy=0.775238	lr=0.098999
Epoch[81] Batch [1149]	Speed: 5097.198677 samples/sec	accuracy=0.775359	lr=0.098811
Epoch[81] Batch [1199]	Speed: 5228.733569 samples/sec	accuracy=0.775542	lr=0.098623
Epoch[81] Batch [1249]	Speed: 5269.376887 samples/sec	accuracy=0.775649	lr=0.098434
[Epoch 81] training: accuracy=0.775653
[Epoch 81] speed: 5090 samples/sec	time cost: 260.925176
[Epoch 81] validation: err-top1=0.324402 err-top5=0.119324
Epoch[82] Batch [49]	Speed: 4880.375901 samples/sec	accuracy=0.774593	lr=0.098243
Epoch[82] Batch [99]	Speed: 5097.878706 samples/sec	accuracy=0.774510	lr=0.098055
Epoch[82] Batch [149]	Speed: 4957.371356 samples/sec	accuracy=0.774241	lr=0.097867
Epoch[82] Batch [199]	Speed: 4827.985162 samples/sec	accuracy=0.774380	lr=0.097679
Epoch[82] Batch [249]	Speed: 4900.264758 samples/sec	accuracy=0.774478	lr=0.097492
Epoch[82] Batch [299]	Speed: 5197.988466 samples/sec	accuracy=0.774510	lr=0.097304
Epoch[82] Batch [349]	Speed: 4994.523653 samples/sec	accuracy=0.774644	lr=0.097117
Epoch[82] Batch [399]	Speed: 4983.342856 samples/sec	accuracy=0.774780	lr=0.096930
Epoch[82] Batch [449]	Speed: 5020.071551 samples/sec	accuracy=0.774908	lr=0.096743
Epoch[82] Batch [499]	Speed: 5043.030702 samples/sec	accuracy=0.774982	lr=0.096556
Epoch[82] Batch [549]	Speed: 5138.610984 samples/sec	accuracy=0.774993	lr=0.096369
Epoch[82] Batch [599]	Speed: 5134.246440 samples/sec	accuracy=0.775106	lr=0.096182
Epoch[82] Batch [649]	Speed: 5330.388292 samples/sec	accuracy=0.775165	lr=0.095996
Epoch[82] Batch [699]	Speed: 5306.346679 samples/sec	accuracy=0.775281	lr=0.095809
Epoch[82] Batch [749]	Speed: 5217.178391 samples/sec	accuracy=0.775460	lr=0.095623
Epoch[82] Batch [799]	Speed: 5350.412315 samples/sec	accuracy=0.775650	lr=0.095437
Epoch[82] Batch [849]	Speed: 5308.082327 samples/sec	accuracy=0.775892	lr=0.095250
Epoch[82] Batch [899]	Speed: 5261.070253 samples/sec	accuracy=0.775931	lr=0.095064
Epoch[82] Batch [949]	Speed: 5070.593585 samples/sec	accuracy=0.776014	lr=0.094879
Epoch[82] Batch [999]	Speed: 5122.764798 samples/sec	accuracy=0.776099	lr=0.094693
Epoch[82] Batch [1049]	Speed: 5242.398508 samples/sec	accuracy=0.776251	lr=0.094507
Epoch[82] Batch [1099]	Speed: 5242.311486 samples/sec	accuracy=0.776335	lr=0.094322
Epoch[82] Batch [1149]	Speed: 4975.929982 samples/sec	accuracy=0.776451	lr=0.094137
Epoch[82] Batch [1199]	Speed: 5041.889435 samples/sec	accuracy=0.776634	lr=0.093951
Epoch[82] Batch [1249]	Speed: 5252.691993 samples/sec	accuracy=0.776760	lr=0.093766
[Epoch 82] training: accuracy=0.776762
[Epoch 82] speed: 5107 samples/sec	time cost: 260.229125
[Epoch 82] validation: err-top1=0.318937 err-top5=0.114856
Epoch[83] Batch [49]	Speed: 5074.174117 samples/sec	accuracy=0.777069	lr=0.093578
Epoch[83] Batch [99]	Speed: 5086.919229 samples/sec	accuracy=0.777895	lr=0.093393
Epoch[83] Batch [149]	Speed: 5193.878244 samples/sec	accuracy=0.777525	lr=0.093208
Epoch[83] Batch [199]	Speed: 5065.901208 samples/sec	accuracy=0.776793	lr=0.093024
Epoch[83] Batch [249]	Speed: 5051.311716 samples/sec	accuracy=0.776906	lr=0.092839
Epoch[83] Batch [299]	Speed: 5050.939015 samples/sec	accuracy=0.776935	lr=0.092655
Epoch[83] Batch [349]	Speed: 4938.842743 samples/sec	accuracy=0.777050	lr=0.092471
Epoch[83] Batch [399]	Speed: 5096.259519 samples/sec	accuracy=0.777051	lr=0.092287
Epoch[83] Batch [449]	Speed: 4933.562154 samples/sec	accuracy=0.777032	lr=0.092103
Epoch[83] Batch [499]	Speed: 5244.424519 samples/sec	accuracy=0.777074	lr=0.091919
Epoch[83] Batch [549]	Speed: 4905.830183 samples/sec	accuracy=0.777199	lr=0.091735
Epoch[83] Batch [599]	Speed: 5267.592158 samples/sec	accuracy=0.777303	lr=0.091552
Epoch[83] Batch [649]	Speed: 5257.858350 samples/sec	accuracy=0.777291	lr=0.091368
Epoch[83] Batch [699]	Speed: 4874.405526 samples/sec	accuracy=0.777324	lr=0.091185
Epoch[83] Batch [749]	Speed: 5160.686368 samples/sec	accuracy=0.777327	lr=0.091002
Epoch[83] Batch [799]	Speed: 5146.979927 samples/sec	accuracy=0.777373	lr=0.090819
Epoch[83] Batch [849]	Speed: 5141.704166 samples/sec	accuracy=0.777370	lr=0.090636
Epoch[83] Batch [899]	Speed: 5027.613098 samples/sec	accuracy=0.777446	lr=0.090453
Epoch[83] Batch [949]	Speed: 5165.535191 samples/sec	accuracy=0.777526	lr=0.090270
Epoch[83] Batch [999]	Speed: 5166.282670 samples/sec	accuracy=0.777581	lr=0.090088
Epoch[83] Batch [1049]	Speed: 5347.798559 samples/sec	accuracy=0.777686	lr=0.089905
Epoch[83] Batch [1099]	Speed: 5276.535303 samples/sec	accuracy=0.777826	lr=0.089723
Epoch[83] Batch [1149]	Speed: 5097.602679 samples/sec	accuracy=0.777868	lr=0.089541
Epoch[83] Batch [1199]	Speed: 5060.876261 samples/sec	accuracy=0.777932	lr=0.089359
Epoch[83] Batch [1249]	Speed: 5351.865862 samples/sec	accuracy=0.778002	lr=0.089177
[Epoch 83] training: accuracy=0.778005
[Epoch 83] speed: 5112 samples/sec	time cost: 259.727258
[Epoch 83] validation: err-top1=0.316346 err-top5=0.111567
Epoch[84] Batch [49]	Speed: 5075.271632 samples/sec	accuracy=0.778047	lr=0.088992
Epoch[84] Batch [99]	Speed: 5287.002202 samples/sec	accuracy=0.777614	lr=0.088810
Epoch[84] Batch [149]	Speed: 5198.533187 samples/sec	accuracy=0.777046	lr=0.088629
Epoch[84] Batch [199]	Speed: 4964.612615 samples/sec	accuracy=0.777321	lr=0.088448
Epoch[84] Batch [249]	Speed: 5031.721521 samples/sec	accuracy=0.777361	lr=0.088266
Epoch[84] Batch [299]	Speed: 5169.151320 samples/sec	accuracy=0.777561	lr=0.088085
Epoch[84] Batch [349]	Speed: 5256.795492 samples/sec	accuracy=0.777586	lr=0.087904
Epoch[84] Batch [399]	Speed: 4985.405588 samples/sec	accuracy=0.777602	lr=0.087724
Epoch[84] Batch [449]	Speed: 4975.858844 samples/sec	accuracy=0.777809	lr=0.087543
Epoch[84] Batch [499]	Speed: 5314.541980 samples/sec	accuracy=0.778015	lr=0.087363
Epoch[84] Batch [549]	Speed: 5381.397109 samples/sec	accuracy=0.777979	lr=0.087182
Epoch[84] Batch [599]	Speed: 5145.260613 samples/sec	accuracy=0.778125	lr=0.087002
Epoch[84] Batch [649]	Speed: 5419.872855 samples/sec	accuracy=0.778129	lr=0.086822
Epoch[84] Batch [699]	Speed: 5377.478578 samples/sec	accuracy=0.778293	lr=0.086642
Epoch[84] Batch [749]	Speed: 5174.536950 samples/sec	accuracy=0.778408	lr=0.086462
Epoch[84] Batch [799]	Speed: 5193.595241 samples/sec	accuracy=0.778534	lr=0.086282
Epoch[84] Batch [849]	Speed: 5106.327260 samples/sec	accuracy=0.778594	lr=0.086103
Epoch[84] Batch [899]	Speed: 5215.857503 samples/sec	accuracy=0.778678	lr=0.085923
Epoch[84] Batch [949]	Speed: 5214.564127 samples/sec	accuracy=0.778690	lr=0.085744
Epoch[84] Batch [999]	Speed: 5145.809258 samples/sec	accuracy=0.778792	lr=0.085565
Epoch[84] Batch [1049]	Speed: 5284.829625 samples/sec	accuracy=0.778906	lr=0.085386
Epoch[84] Batch [1099]	Speed: 4863.071283 samples/sec	accuracy=0.778996	lr=0.085207
Epoch[84] Batch [1149]	Speed: 5161.670763 samples/sec	accuracy=0.779050	lr=0.085028
Epoch[84] Batch [1199]	Speed: 5075.527252 samples/sec	accuracy=0.779142	lr=0.084850
Epoch[84] Batch [1249]	Speed: 5121.463675 samples/sec	accuracy=0.779236	lr=0.084671
[Epoch 84] training: accuracy=0.779239
[Epoch 84] speed: 5157 samples/sec	time cost: 257.582502
[Epoch 84] validation: err-top1=0.309889 err-top5=0.105748
Epoch[85] Batch [49]	Speed: 4941.389606 samples/sec	accuracy=0.779189	lr=0.084489
Epoch[85] Batch [99]	Speed: 5111.025037 samples/sec	accuracy=0.778700	lr=0.084311
Epoch[85] Batch [149]	Speed: 5080.768865 samples/sec	accuracy=0.778990	lr=0.084133
Epoch[85] Batch [199]	Speed: 5154.476698 samples/sec	accuracy=0.778836	lr=0.083955
Epoch[85] Batch [249]	Speed: 5310.995068 samples/sec	accuracy=0.778663	lr=0.083777
Epoch[85] Batch [299]	Speed: 5282.251733 samples/sec	accuracy=0.778723	lr=0.083600
Epoch[85] Batch [349]	Speed: 5162.157766 samples/sec	accuracy=0.778815	lr=0.083422
Epoch[85] Batch [399]	Speed: 5167.346290 samples/sec	accuracy=0.779007	lr=0.083245
Epoch[85] Batch [449]	Speed: 5051.914309 samples/sec	accuracy=0.779119	lr=0.083067
Epoch[85] Batch [499]	Speed: 4946.707939 samples/sec	accuracy=0.779176	lr=0.082890
Epoch[85] Batch [549]	Speed: 5066.058719 samples/sec	accuracy=0.779242	lr=0.082713
Epoch[85] Batch [599]	Speed: 5251.914421 samples/sec	accuracy=0.779326	lr=0.082537
Epoch[85] Batch [649]	Speed: 5187.567561 samples/sec	accuracy=0.779308	lr=0.082360
Epoch[85] Batch [699]	Speed: 5099.007447 samples/sec	accuracy=0.779364	lr=0.082183
Epoch[85] Batch [749]	Speed: 5198.012246 samples/sec	accuracy=0.779359	lr=0.082007
Epoch[85] Batch [799]	Speed: 5354.204580 samples/sec	accuracy=0.779375	lr=0.081831
Epoch[85] Batch [849]	Speed: 5129.732577 samples/sec	accuracy=0.779439	lr=0.081655
Epoch[85] Batch [899]	Speed: 5257.184136 samples/sec	accuracy=0.779542	lr=0.081479
Epoch[85] Batch [949]	Speed: 5123.562413 samples/sec	accuracy=0.779674	lr=0.081303
Epoch[85] Batch [999]	Speed: 5175.334680 samples/sec	accuracy=0.779806	lr=0.081127
Epoch[85] Batch [1049]	Speed: 5131.832458 samples/sec	accuracy=0.779933	lr=0.080952
Epoch[85] Batch [1099]	Speed: 5188.287210 samples/sec	accuracy=0.779958	lr=0.080776
Epoch[85] Batch [1149]	Speed: 5095.593705 samples/sec	accuracy=0.780067	lr=0.080601
Epoch[85] Batch [1199]	Speed: 5080.591927 samples/sec	accuracy=0.780210	lr=0.080426
Epoch[85] Batch [1249]	Speed: 5191.472006 samples/sec	accuracy=0.780328	lr=0.080251
[Epoch 85] training: accuracy=0.780331
[Epoch 85] speed: 5144 samples/sec	time cost: 258.136882
[Epoch 85] validation: err-top1=0.313058 err-top5=0.110750
Epoch[86] Batch [49]	Speed: 4766.113221 samples/sec	accuracy=0.780413	lr=0.080073
Epoch[86] Batch [99]	Speed: 4954.488376 samples/sec	accuracy=0.779871	lr=0.079898
Epoch[86] Batch [149]	Speed: 5146.923675 samples/sec	accuracy=0.779380	lr=0.079723
Epoch[86] Batch [199]	Speed: 5116.352576 samples/sec	accuracy=0.779504	lr=0.079549
Epoch[86] Batch [249]	Speed: 5232.648996 samples/sec	accuracy=0.779547	lr=0.079375
Epoch[86] Batch [299]	Speed: 5253.830311 samples/sec	accuracy=0.779314	lr=0.079201
Epoch[86] Batch [349]	Speed: 4917.032258 samples/sec	accuracy=0.779194	lr=0.079027
Epoch[86] Batch [399]	Speed: 5271.011321 samples/sec	accuracy=0.779259	lr=0.078853
Epoch[86] Batch [449]	Speed: 5225.831363 samples/sec	accuracy=0.779278	lr=0.078679
Epoch[86] Batch [499]	Speed: 4995.968526 samples/sec	accuracy=0.779338	lr=0.078506
Epoch[86] Batch [549]	Speed: 5073.594850 samples/sec	accuracy=0.779475	lr=0.078332
Epoch[86] Batch [599]	Speed: 5010.667724 samples/sec	accuracy=0.779595	lr=0.078159
Epoch[86] Batch [649]	Speed: 5338.211119 samples/sec	accuracy=0.779747	lr=0.077986
Epoch[86] Batch [699]	Speed: 4805.853881 samples/sec	accuracy=0.779837	lr=0.077813
Epoch[86] Batch [749]	Speed: 5255.945053 samples/sec	accuracy=0.780001	lr=0.077640
Epoch[86] Batch [799]	Speed: 5516.450136 samples/sec	accuracy=0.780008	lr=0.077467
Epoch[86] Batch [849]	Speed: 5309.549850 samples/sec	accuracy=0.780152	lr=0.077295
Epoch[86] Batch [899]	Speed: 5182.862928 samples/sec	accuracy=0.780272	lr=0.077123
Epoch[86] Batch [949]	Speed: 5299.520174 samples/sec	accuracy=0.780424	lr=0.076950
Epoch[86] Batch [999]	Speed: 5177.403417 samples/sec	accuracy=0.780570	lr=0.076778
Epoch[86] Batch [1049]	Speed: 5058.055015 samples/sec	accuracy=0.780708	lr=0.076606
Epoch[86] Batch [1099]	Speed: 4969.033713 samples/sec	accuracy=0.780833	lr=0.076435
Epoch[86] Batch [1149]	Speed: 5107.357224 samples/sec	accuracy=0.780856	lr=0.076263
Epoch[86] Batch [1199]	Speed: 5161.254434 samples/sec	accuracy=0.780896	lr=0.076091
Epoch[86] Batch [1249]	Speed: 5037.094875 samples/sec	accuracy=0.780975	lr=0.075920
[Epoch 86] training: accuracy=0.780978
[Epoch 86] speed: 5117 samples/sec	time cost: 259.362245
[Epoch 86] validation: err-top1=0.300761 err-top5=0.100845
Epoch[87] Batch [49]	Speed: 4713.452392 samples/sec	accuracy=0.780515	lr=0.075745
Epoch[87] Batch [99]	Speed: 5152.817778 samples/sec	accuracy=0.780799	lr=0.075574
Epoch[87] Batch [149]	Speed: 5030.230093 samples/sec	accuracy=0.780438	lr=0.075404
Epoch[87] Batch [199]	Speed: 5124.757224 samples/sec	accuracy=0.780456	lr=0.075233
Epoch[87] Batch [249]	Speed: 5106.098759 samples/sec	accuracy=0.780508	lr=0.075062
Epoch[87] Batch [299]	Speed: 5316.389468 samples/sec	accuracy=0.780508	lr=0.074892
Epoch[87] Batch [349]	Speed: 5055.739506 samples/sec	accuracy=0.780618	lr=0.074721
Epoch[87] Batch [399]	Speed: 5126.524539 samples/sec	accuracy=0.780714	lr=0.074551
Epoch[87] Batch [449]	Speed: 5304.344087 samples/sec	accuracy=0.780846	lr=0.074381
Epoch[87] Batch [499]	Speed: 4979.814260 samples/sec	accuracy=0.780868	lr=0.074211
Epoch[87] Batch [549]	Speed: 5405.959560 samples/sec	accuracy=0.781052	lr=0.074042
Epoch[87] Batch [599]	Speed: 5139.153169 samples/sec	accuracy=0.781190	lr=0.073872
Epoch[87] Batch [649]	Speed: 5185.480178 samples/sec	accuracy=0.781244	lr=0.073703
Epoch[87] Batch [699]	Speed: 5048.432878 samples/sec	accuracy=0.781228	lr=0.073534
Epoch[87] Batch [749]	Speed: 5219.690988 samples/sec	accuracy=0.781286	lr=0.073364
Epoch[87] Batch [799]	Speed: 5225.555675 samples/sec	accuracy=0.781311	lr=0.073196
Epoch[87] Batch [849]	Speed: 5080.672101 samples/sec	accuracy=0.781375	lr=0.073027
Epoch[87] Batch [899]	Speed: 4890.312803 samples/sec	accuracy=0.781502	lr=0.072858
Epoch[87] Batch [949]	Speed: 5071.530852 samples/sec	accuracy=0.781597	lr=0.072690
Epoch[87] Batch [999]	Speed: 5069.223095 samples/sec	accuracy=0.781645	lr=0.072521
Epoch[87] Batch [1049]	Speed: 5132.884025 samples/sec	accuracy=0.781731	lr=0.072353
Epoch[87] Batch [1099]	Speed: 5274.036085 samples/sec	accuracy=0.781872	lr=0.072185
Epoch[87] Batch [1149]	Speed: 5261.534684 samples/sec	accuracy=0.781953	lr=0.072017
Epoch[87] Batch [1199]	Speed: 5338.502803 samples/sec	accuracy=0.782025	lr=0.071849
Epoch[87] Batch [1249]	Speed: 4972.439965 samples/sec	accuracy=0.782135	lr=0.071682
[Epoch 87] training: accuracy=0.782136
[Epoch 87] speed: 5120 samples/sec	time cost: 259.452494
[Epoch 87] validation: err-top1=0.308594 err-top5=0.108297
Epoch[88] Batch [49]	Speed: 4954.474888 samples/sec	accuracy=0.782815	lr=0.071511
Epoch[88] Batch [99]	Speed: 5158.492057 samples/sec	accuracy=0.782426	lr=0.071344
Epoch[88] Batch [149]	Speed: 5089.159671 samples/sec	accuracy=0.782174	lr=0.071177
Epoch[88] Batch [199]	Speed: 5238.929954 samples/sec	accuracy=0.782211	lr=0.071010
Epoch[88] Batch [249]	Speed: 5173.569826 samples/sec	accuracy=0.781699	lr=0.070843
Epoch[88] Batch [299]	Speed: 5198.384318 samples/sec	accuracy=0.781775	lr=0.070676
Epoch[88] Batch [349]	Speed: 5007.191987 samples/sec	accuracy=0.781837	lr=0.070510
Epoch[88] Batch [399]	Speed: 5003.154278 samples/sec	accuracy=0.781956	lr=0.070343
Epoch[88] Batch [449]	Speed: 5137.268987 samples/sec	accuracy=0.782028	lr=0.070177
Epoch[88] Batch [499]	Speed: 5079.523828 samples/sec	accuracy=0.782015	lr=0.070011
Epoch[88] Batch [549]	Speed: 5042.556443 samples/sec	accuracy=0.782102	lr=0.069845
Epoch[88] Batch [599]	Speed: 5257.260713 samples/sec	accuracy=0.782180	lr=0.069680
Epoch[88] Batch [649]	Speed: 4947.857589 samples/sec	accuracy=0.782400	lr=0.069514
Epoch[88] Batch [699]	Speed: 5004.061413 samples/sec	accuracy=0.782462	lr=0.069349
Epoch[88] Batch [749]	Speed: 5089.218768 samples/sec	accuracy=0.782453	lr=0.069183
Epoch[88] Batch [799]	Speed: 5193.955249 samples/sec	accuracy=0.782572	lr=0.069018
Epoch[88] Batch [849]	Speed: 5141.620455 samples/sec	accuracy=0.782560	lr=0.068853
Epoch[88] Batch [899]	Speed: 5097.613328 samples/sec	accuracy=0.782563	lr=0.068689
Epoch[88] Batch [949]	Speed: 5013.688812 samples/sec	accuracy=0.782534	lr=0.068524
Epoch[88] Batch [999]	Speed: 5266.846078 samples/sec	accuracy=0.782631	lr=0.068359
Epoch[88] Batch [1049]	Speed: 5020.846310 samples/sec	accuracy=0.782654	lr=0.068195
Epoch[88] Batch [1099]	Speed: 5073.436031 samples/sec	accuracy=0.782794	lr=0.068031
Epoch[88] Batch [1149]	Speed: 4990.915994 samples/sec	accuracy=0.782875	lr=0.067867
Epoch[88] Batch [1199]	Speed: 5190.713707 samples/sec	accuracy=0.782891	lr=0.067703
Epoch[88] Batch [1249]	Speed: 5110.722043 samples/sec	accuracy=0.782990	lr=0.067539
[Epoch 88] training: accuracy=0.782993
[Epoch 88] speed: 5093 samples/sec	time cost: 261.022563
[Epoch 88] validation: err-top1=0.330417 err-top5=0.122748
Epoch[89] Batch [49]	Speed: 4972.285342 samples/sec	accuracy=0.782845	lr=0.067373
Epoch[89] Batch [99]	Speed: 4939.671370 samples/sec	accuracy=0.782699	lr=0.067209
Epoch[89] Batch [149]	Speed: 5169.910799 samples/sec	accuracy=0.782748	lr=0.067046
Epoch[89] Batch [199]	Speed: 5027.454672 samples/sec	accuracy=0.782581	lr=0.066883
Epoch[89] Batch [249]	Speed: 5150.869698 samples/sec	accuracy=0.782652	lr=0.066720
Epoch[89] Batch [299]	Speed: 5176.064039 samples/sec	accuracy=0.782707	lr=0.066557
Epoch[89] Batch [349]	Speed: 5103.751194 samples/sec	accuracy=0.782817	lr=0.066395
Epoch[89] Batch [399]	Speed: 5091.411744 samples/sec	accuracy=0.782763	lr=0.066232
Epoch[89] Batch [449]	Speed: 5187.655908 samples/sec	accuracy=0.782805	lr=0.066070
Epoch[89] Batch [499]	Speed: 5320.826602 samples/sec	accuracy=0.782924	lr=0.065908
Epoch[89] Batch [549]	Speed: 5189.980591 samples/sec	accuracy=0.782987	lr=0.065746
Epoch[89] Batch [599]	Speed: 5167.601943 samples/sec	accuracy=0.783100	lr=0.065584
Epoch[89] Batch [649]	Speed: 5042.757148 samples/sec	accuracy=0.783193	lr=0.065423
Epoch[89] Batch [699]	Speed: 5091.626860 samples/sec	accuracy=0.783274	lr=0.065261
Epoch[89] Batch [749]	Speed: 4949.935083 samples/sec	accuracy=0.783297	lr=0.065100
Epoch[89] Batch [799]	Speed: 5050.877715 samples/sec	accuracy=0.783380	lr=0.064939
Epoch[89] Batch [849]	Speed: 5155.346100 samples/sec	accuracy=0.783534	lr=0.064778
Epoch[89] Batch [899]	Speed: 5125.496006 samples/sec	accuracy=0.783595	lr=0.064617
Epoch[89] Batch [949]	Speed: 5155.074581 samples/sec	accuracy=0.783711	lr=0.064456
Epoch[89] Batch [999]	Speed: 5202.519323 samples/sec	accuracy=0.783787	lr=0.064296
Epoch[89] Batch [1049]	Speed: 5191.965779 samples/sec	accuracy=0.783867	lr=0.064135
Epoch[89] Batch [1099]	Speed: 5290.827478 samples/sec	accuracy=0.783931	lr=0.063975
Epoch[89] Batch [1149]	Speed: 4864.934454 samples/sec	accuracy=0.784017	lr=0.063815
Epoch[89] Batch [1199]	Speed: 5108.399391 samples/sec	accuracy=0.784041	lr=0.063655
Epoch[89] Batch [1249]	Speed: 5355.340984 samples/sec	accuracy=0.784148	lr=0.063496
[Epoch 89] training: accuracy=0.784151
[Epoch 89] speed: 5116 samples/sec	time cost: 259.432557
[Epoch 89] validation: err-top1=0.291235 err-top5=0.097258
Epoch[90] Batch [49]	Speed: 4973.585596 samples/sec	accuracy=0.784362	lr=0.063333
Epoch[90] Batch [99]	Speed: 5038.360334 samples/sec	accuracy=0.783628	lr=0.063174
Epoch[90] Batch [149]	Speed: 4840.604276 samples/sec	accuracy=0.783432	lr=0.063015
Epoch[90] Batch [199]	Speed: 5217.029466 samples/sec	accuracy=0.783395	lr=0.062855
Epoch[90] Batch [249]	Speed: 5123.273209 samples/sec	accuracy=0.783427	lr=0.062697
Epoch[90] Batch [299]	Speed: 5144.071627 samples/sec	accuracy=0.783526	lr=0.062538
Epoch[90] Batch [349]	Speed: 5048.288091 samples/sec	accuracy=0.783471	lr=0.062379
Epoch[90] Batch [399]	Speed: 5139.450811 samples/sec	accuracy=0.783582	lr=0.062221
Epoch[90] Batch [449]	Speed: 5243.432506 samples/sec	accuracy=0.783667	lr=0.062063
Epoch[90] Batch [499]	Speed: 5162.648088 samples/sec	accuracy=0.783791	lr=0.061905
Epoch[90] Batch [549]	Speed: 5016.166306 samples/sec	accuracy=0.783951	lr=0.061747
Epoch[90] Batch [599]	Speed: 5178.349122 samples/sec	accuracy=0.784013	lr=0.061589
Epoch[90] Batch [649]	Speed: 5379.606993 samples/sec	accuracy=0.784122	lr=0.061432
Epoch[90] Batch [699]	Speed: 5150.146310 samples/sec	accuracy=0.784308	lr=0.061274
Epoch[90] Batch [749]	Speed: 5057.256702 samples/sec	accuracy=0.784482	lr=0.061117
Epoch[90] Batch [799]	Speed: 5136.720442 samples/sec	accuracy=0.784547	lr=0.060960
Epoch[90] Batch [849]	Speed: 5418.730915 samples/sec	accuracy=0.784701	lr=0.060803
Epoch[90] Batch [899]	Speed: 5032.895697 samples/sec	accuracy=0.784853	lr=0.060646
Epoch[90] Batch [949]	Speed: 5077.757548 samples/sec	accuracy=0.784902	lr=0.060490
Epoch[90] Batch [999]	Speed: 5326.791928 samples/sec	accuracy=0.784907	lr=0.060333
Epoch[90] Batch [1049]	Speed: 5095.417184 samples/sec	accuracy=0.784948	lr=0.060177
Epoch[90] Batch [1099]	Speed: 5145.959817 samples/sec	accuracy=0.784998	lr=0.060021
Epoch[90] Batch [1149]	Speed: 5351.131456 samples/sec	accuracy=0.785079	lr=0.059865
Epoch[90] Batch [1199]	Speed: 5469.721336 samples/sec	accuracy=0.785214	lr=0.059710
Epoch[90] Batch [1249]	Speed: 5263.693048 samples/sec	accuracy=0.785280	lr=0.059554
[Epoch 90] training: accuracy=0.785283
[Epoch 90] speed: 5152 samples/sec	time cost: 257.908262
[Epoch 90] validation: err-top1=0.296795 err-top5=0.101244
Epoch[91] Batch [49]	Speed: 4790.192431 samples/sec	accuracy=0.785792	lr=0.059396
Epoch[91] Batch [99]	Speed: 5153.636653 samples/sec	accuracy=0.785379	lr=0.059240
Epoch[91] Batch [149]	Speed: 5150.768144 samples/sec	accuracy=0.785310	lr=0.059085
Epoch[91] Batch [199]	Speed: 5141.826538 samples/sec	accuracy=0.785567	lr=0.058930
Epoch[91] Batch [249]	Speed: 4965.433953 samples/sec	accuracy=0.785357	lr=0.058776
Epoch[91] Batch [299]	Speed: 5092.783631 samples/sec	accuracy=0.785398	lr=0.058621
Epoch[91] Batch [349]	Speed: 5043.002398 samples/sec	accuracy=0.785451	lr=0.058467
Epoch[91] Batch [399]	Speed: 5247.888941 samples/sec	accuracy=0.785318	lr=0.058313
Epoch[91] Batch [449]	Speed: 5123.363047 samples/sec	accuracy=0.785355	lr=0.058158
Epoch[91] Batch [499]	Speed: 5047.791248 samples/sec	accuracy=0.785489	lr=0.058005
Epoch[91] Batch [549]	Speed: 5153.948592 samples/sec	accuracy=0.785545	lr=0.057851
Epoch[91] Batch [599]	Speed: 5074.612730 samples/sec	accuracy=0.785635	lr=0.057697
Epoch[91] Batch [649]	Speed: 5169.405533 samples/sec	accuracy=0.785652	lr=0.057544
Epoch[91] Batch [699]	Speed: 4895.311564 samples/sec	accuracy=0.785664	lr=0.057391
Epoch[91] Batch [749]	Speed: 5157.522994 samples/sec	accuracy=0.785655	lr=0.057238
Epoch[91] Batch [799]	Speed: 5195.935434 samples/sec	accuracy=0.785754	lr=0.057085
Epoch[91] Batch [849]	Speed: 5002.615119 samples/sec	accuracy=0.785830	lr=0.056932
Epoch[91] Batch [899]	Speed: 5067.006744 samples/sec	accuracy=0.785866	lr=0.056780
Epoch[91] Batch [949]	Speed: 5114.778280 samples/sec	accuracy=0.785905	lr=0.056627
Epoch[91] Batch [999]	Speed: 5338.489400 samples/sec	accuracy=0.785956	lr=0.056475
Epoch[91] Batch [1049]	Speed: 5100.430674 samples/sec	accuracy=0.785973	lr=0.056323
Epoch[91] Batch [1099]	Speed: 5117.912838 samples/sec	accuracy=0.786103	lr=0.056171
Epoch[91] Batch [1149]	Speed: 5175.302751 samples/sec	accuracy=0.786202	lr=0.056020
Epoch[91] Batch [1199]	Speed: 5184.859448 samples/sec	accuracy=0.786293	lr=0.055868
Epoch[91] Batch [1249]	Speed: 5213.836537 samples/sec	accuracy=0.786337	lr=0.055717
[Epoch 91] training: accuracy=0.786339
[Epoch 91] speed: 5102 samples/sec	time cost: 260.115816
[Epoch 91] validation: err-top1=0.282605 err-top5=0.091438
Epoch[92] Batch [49]	Speed: 4763.902094 samples/sec	accuracy=0.786328	lr=0.055563
Epoch[92] Batch [99]	Speed: 5147.794480 samples/sec	accuracy=0.785838	lr=0.055412
Epoch[92] Batch [149]	Speed: 5266.298830 samples/sec	accuracy=0.785970	lr=0.055261
Epoch[92] Batch [199]	Speed: 4909.869425 samples/sec	accuracy=0.786112	lr=0.055111
Epoch[92] Batch [249]	Speed: 5146.628128 samples/sec	accuracy=0.786248	lr=0.054960
Epoch[92] Batch [299]	Speed: 4941.210191 samples/sec	accuracy=0.786217	lr=0.054810
Epoch[92] Batch [349]	Speed: 5070.143216 samples/sec	accuracy=0.786245	lr=0.054660
Epoch[92] Batch [399]	Speed: 4993.627288 samples/sec	accuracy=0.786198	lr=0.054510
Epoch[92] Batch [449]	Speed: 5357.284978 samples/sec	accuracy=0.786398	lr=0.054360
Epoch[92] Batch [499]	Speed: 5103.834163 samples/sec	accuracy=0.786430	lr=0.054210
Epoch[92] Batch [549]	Speed: 4957.505825 samples/sec	accuracy=0.786501	lr=0.054061
Epoch[92] Batch [599]	Speed: 5172.878676 samples/sec	accuracy=0.786528	lr=0.053912
Epoch[92] Batch [649]	Speed: 5197.765905 samples/sec	accuracy=0.786604	lr=0.053763
Epoch[92] Batch [699]	Speed: 5183.925503 samples/sec	accuracy=0.786609	lr=0.053614
Epoch[92] Batch [749]	Speed: 5225.307098 samples/sec	accuracy=0.786686	lr=0.053465
Epoch[92] Batch [799]	Speed: 5346.621957 samples/sec	accuracy=0.786811	lr=0.053317
Epoch[92] Batch [849]	Speed: 5109.344974 samples/sec	accuracy=0.786857	lr=0.053168
Epoch[92] Batch [899]	Speed: 4970.287403 samples/sec	accuracy=0.786871	lr=0.053020
Epoch[92] Batch [949]	Speed: 5000.940109 samples/sec	accuracy=0.786934	lr=0.052872
Epoch[92] Batch [999]	Speed: 5074.411040 samples/sec	accuracy=0.787007	lr=0.052724
Epoch[92] Batch [1049]	Speed: 5141.379430 samples/sec	accuracy=0.787110	lr=0.052577
Epoch[92] Batch [1099]	Speed: 5243.424569 samples/sec	accuracy=0.787156	lr=0.052429
Epoch[92] Batch [1149]	Speed: 5203.476749 samples/sec	accuracy=0.787213	lr=0.052282
Epoch[92] Batch [1199]	Speed: 5336.635275 samples/sec	accuracy=0.787215	lr=0.052135
Epoch[92] Batch [1249]	Speed: 5371.188461 samples/sec	accuracy=0.787277	lr=0.051988
[Epoch 92] training: accuracy=0.787280
[Epoch 92] speed: 5120 samples/sec	time cost: 259.477851
[Epoch 92] validation: err-top1=0.291434 err-top5=0.099908
Epoch[93] Batch [49]	Speed: 4811.674118 samples/sec	accuracy=0.788032	lr=0.051838
Epoch[93] Batch [99]	Speed: 5120.581731 samples/sec	accuracy=0.787719	lr=0.051691
Epoch[93] Batch [149]	Speed: 5225.493879 samples/sec	accuracy=0.787389	lr=0.051545
Epoch[93] Batch [199]	Speed: 5389.303535 samples/sec	accuracy=0.787096	lr=0.051399
Epoch[93] Batch [249]	Speed: 5283.885974 samples/sec	accuracy=0.787061	lr=0.051253
Epoch[93] Batch [299]	Speed: 5280.514495 samples/sec	accuracy=0.786940	lr=0.051107
Epoch[93] Batch [349]	Speed: 5328.834385 samples/sec	accuracy=0.787030	lr=0.050961
Epoch[93] Batch [399]	Speed: 5139.782439 samples/sec	accuracy=0.787060	lr=0.050816
Epoch[93] Batch [449]	Speed: 5234.633790 samples/sec	accuracy=0.787086	lr=0.050670
Epoch[93] Batch [499]	Speed: 5023.892852 samples/sec	accuracy=0.787153	lr=0.050525
Epoch[93] Batch [549]	Speed: 5204.859493 samples/sec	accuracy=0.787216	lr=0.050380
Epoch[93] Batch [599]	Speed: 4998.822499 samples/sec	accuracy=0.787260	lr=0.050235
Epoch[93] Batch [649]	Speed: 5228.606008 samples/sec	accuracy=0.787260	lr=0.050091
Epoch[93] Batch [699]	Speed: 5415.093320 samples/sec	accuracy=0.787375	lr=0.049946
Epoch[93] Batch [749]	Speed: 5113.125941 samples/sec	accuracy=0.787525	lr=0.049802
Epoch[93] Batch [799]	Speed: 5102.019907 samples/sec	accuracy=0.787593	lr=0.049658
Epoch[93] Batch [849]	Speed: 5116.912507 samples/sec	accuracy=0.787677	lr=0.049514
Epoch[93] Batch [899]	Speed: 5082.355970 samples/sec	accuracy=0.787764	lr=0.049370
Epoch[93] Batch [949]	Speed: 5241.375024 samples/sec	accuracy=0.787785	lr=0.049227
Epoch[93] Batch [999]	Speed: 5291.678422 samples/sec	accuracy=0.787822	lr=0.049083
Epoch[93] Batch [1049]	Speed: 5134.992135 samples/sec	accuracy=0.787855	lr=0.048940
Epoch[93] Batch [1099]	Speed: 5021.409249 samples/sec	accuracy=0.787921	lr=0.048797
Epoch[93] Batch [1149]	Speed: 5364.579869 samples/sec	accuracy=0.787992	lr=0.048654
Epoch[93] Batch [1199]	Speed: 5160.373241 samples/sec	accuracy=0.788058	lr=0.048511
Epoch[93] Batch [1249]	Speed: 5023.603626 samples/sec	accuracy=0.788138	lr=0.048369
[Epoch 93] training: accuracy=0.788141
[Epoch 93] speed: 5166 samples/sec	time cost: 257.390293
[Epoch 93] validation: err-top1=0.301168 err-top5=0.102946
Epoch[94] Batch [49]	Speed: 4849.874515 samples/sec	accuracy=0.788156	lr=0.048224
Epoch[94] Batch [99]	Speed: 5040.270006 samples/sec	accuracy=0.787767	lr=0.048082
Epoch[94] Batch [149]	Speed: 4995.783382 samples/sec	accuracy=0.787985	lr=0.047940
Epoch[94] Batch [199]	Speed: 5026.930856 samples/sec	accuracy=0.787991	lr=0.047798
Epoch[94] Batch [249]	Speed: 5161.557742 samples/sec	accuracy=0.787937	lr=0.047656
Epoch[94] Batch [299]	Speed: 5195.259661 samples/sec	accuracy=0.787981	lr=0.047515
Epoch[94] Batch [349]	Speed: 5211.015581 samples/sec	accuracy=0.787856	lr=0.047374
Epoch[94] Batch [399]	Speed: 5300.424938 samples/sec	accuracy=0.788011	lr=0.047233
Epoch[94] Batch [449]	Speed: 5155.828939 samples/sec	accuracy=0.787949	lr=0.047092
Epoch[94] Batch [499]	Speed: 5063.175945 samples/sec	accuracy=0.787901	lr=0.046951
Epoch[94] Batch [549]	Speed: 5212.611981 samples/sec	accuracy=0.788029	lr=0.046811
Epoch[94] Batch [599]	Speed: 5249.115632 samples/sec	accuracy=0.788057	lr=0.046670
Epoch[94] Batch [649]	Speed: 5164.755632 samples/sec	accuracy=0.788107	lr=0.046530
Epoch[94] Batch [699]	Speed: 5278.327656 samples/sec	accuracy=0.788229	lr=0.046390
Epoch[94] Batch [749]	Speed: 5255.239435 samples/sec	accuracy=0.788321	lr=0.046251
Epoch[94] Batch [799]	Speed: 5191.999545 samples/sec	accuracy=0.788347	lr=0.046111
Epoch[94] Batch [849]	Speed: 5288.047231 samples/sec	accuracy=0.788330	lr=0.045972
Epoch[94] Batch [899]	Speed: 4942.250936 samples/sec	accuracy=0.788392	lr=0.045833
Epoch[94] Batch [949]	Speed: 5324.738362 samples/sec	accuracy=0.788464	lr=0.045693
Epoch[94] Batch [999]	Speed: 5210.404652 samples/sec	accuracy=0.788538	lr=0.045555
Epoch[94] Batch [1049]	Speed: 5065.205669 samples/sec	accuracy=0.788608	lr=0.045416
Epoch[94] Batch [1099]	Speed: 5195.333943 samples/sec	accuracy=0.788752	lr=0.045278
Epoch[94] Batch [1149]	Speed: 5144.796391 samples/sec	accuracy=0.788802	lr=0.045139
Epoch[94] Batch [1199]	Speed: 5003.223750 samples/sec	accuracy=0.788884	lr=0.045001
Epoch[94] Batch [1249]	Speed: 5281.889903 samples/sec	accuracy=0.788960	lr=0.044863
[Epoch 94] training: accuracy=0.788962
[Epoch 94] speed: 5145 samples/sec	time cost: 258.489049
[Epoch 94] validation: err-top1=0.295340 err-top5=0.099709
Epoch[95] Batch [49]	Speed: 5169.089356 samples/sec	accuracy=0.788865	lr=0.044723
Epoch[95] Batch [99]	Speed: 5188.121505 samples/sec	accuracy=0.788194	lr=0.044585
Epoch[95] Batch [149]	Speed: 5011.416076 samples/sec	accuracy=0.787838	lr=0.044448
Epoch[95] Batch [199]	Speed: 5336.403467 samples/sec	accuracy=0.787956	lr=0.044311
Epoch[95] Batch [249]	Speed: 5216.572859 samples/sec	accuracy=0.788126	lr=0.044174
Epoch[95] Batch [299]	Speed: 4972.002488 samples/sec	accuracy=0.788104	lr=0.044037
Epoch[95] Batch [349]	Speed: 5190.301084 samples/sec	accuracy=0.788275	lr=0.043900
Epoch[95] Batch [399]	Speed: 4929.503315 samples/sec	accuracy=0.788379	lr=0.043764
Epoch[95] Batch [449]	Speed: 5161.124313 samples/sec	accuracy=0.788548	lr=0.043628
Epoch[95] Batch [499]	Speed: 5161.477229 samples/sec	accuracy=0.788541	lr=0.043492
Epoch[95] Batch [549]	Speed: 5076.308185 samples/sec	accuracy=0.788609	lr=0.043356
Epoch[95] Batch [599]	Speed: 5087.877730 samples/sec	accuracy=0.788750	lr=0.043220
Epoch[95] Batch [649]	Speed: 5324.123710 samples/sec	accuracy=0.788830	lr=0.043085
Epoch[95] Batch [699]	Speed: 4992.928002 samples/sec	accuracy=0.789016	lr=0.042949
Epoch[95] Batch [749]	Speed: 4810.181077 samples/sec	accuracy=0.789167	lr=0.042814
Epoch[95] Batch [799]	Speed: 4987.520416 samples/sec	accuracy=0.789220	lr=0.042679
Epoch[95] Batch [849]	Speed: 5155.884024 samples/sec	accuracy=0.789351	lr=0.042545
Epoch[95] Batch [899]	Speed: 5223.247425 samples/sec	accuracy=0.789495	lr=0.042410
Epoch[95] Batch [949]	Speed: 5239.489682 samples/sec	accuracy=0.789532	lr=0.042276
Epoch[95] Batch [999]	Speed: 5150.246356 samples/sec	accuracy=0.789545	lr=0.042141
Epoch[95] Batch [1049]	Speed: 5203.505874 samples/sec	accuracy=0.789612	lr=0.042007
Epoch[95] Batch [1099]	Speed: 5126.616939 samples/sec	accuracy=0.789753	lr=0.041874
Epoch[95] Batch [1149]	Speed: 5017.810848 samples/sec	accuracy=0.789887	lr=0.041740
Epoch[95] Batch [1199]	Speed: 5159.145036 samples/sec	accuracy=0.790018	lr=0.041607
Epoch[95] Batch [1249]	Speed: 5319.493696 samples/sec	accuracy=0.790059	lr=0.041473
[Epoch 95] training: accuracy=0.790059
[Epoch 95] speed: 5121 samples/sec	time cost: 259.704626
[Epoch 95] validation: err-top1=0.277244 err-top5=0.088947
Epoch[96] Batch [49]	Speed: 4784.430198 samples/sec	accuracy=0.790727	lr=0.041338
Epoch[96] Batch [99]	Speed: 5130.563252 samples/sec	accuracy=0.790165	lr=0.041205
Epoch[96] Batch [149]	Speed: 4747.552627 samples/sec	accuracy=0.790053	lr=0.041072
Epoch[96] Batch [199]	Speed: 5255.959074 samples/sec	accuracy=0.790197	lr=0.040940
Epoch[96] Batch [249]	Speed: 5146.447683 samples/sec	accuracy=0.790037	lr=0.040807
Epoch[96] Batch [299]	Speed: 5390.875321 samples/sec	accuracy=0.789893	lr=0.040675
Epoch[96] Batch [349]	Speed: 5007.719404 samples/sec	accuracy=0.790070	lr=0.040543
Epoch[96] Batch [399]	Speed: 5063.509980 samples/sec	accuracy=0.790155	lr=0.040412
Epoch[96] Batch [449]	Speed: 5385.025073 samples/sec	accuracy=0.790227	lr=0.040280
Epoch[96] Batch [499]	Speed: 5466.884804 samples/sec	accuracy=0.790268	lr=0.040149
Epoch[96] Batch [549]	Speed: 5153.781981 samples/sec	accuracy=0.790284	lr=0.040018
Epoch[96] Batch [599]	Speed: 4968.602699 samples/sec	accuracy=0.790428	lr=0.039887
Epoch[96] Batch [649]	Speed: 5206.726429 samples/sec	accuracy=0.790440	lr=0.039756
Epoch[96] Batch [699]	Speed: 5075.279189 samples/sec	accuracy=0.790577	lr=0.039625
Epoch[96] Batch [749]	Speed: 5220.306635 samples/sec	accuracy=0.790581	lr=0.039495
Epoch[96] Batch [799]	Speed: 5438.291241 samples/sec	accuracy=0.790616	lr=0.039365
Epoch[96] Batch [849]	Speed: 5003.100077 samples/sec	accuracy=0.790652	lr=0.039235
Epoch[96] Batch [899]	Speed: 5044.590877 samples/sec	accuracy=0.790662	lr=0.039105
Epoch[96] Batch [949]	Speed: 5427.030431 samples/sec	accuracy=0.790664	lr=0.038975
Epoch[96] Batch [999]	Speed: 5274.752849 samples/sec	accuracy=0.790729	lr=0.038846
Epoch[96] Batch [1049]	Speed: 5181.457840 samples/sec	accuracy=0.790812	lr=0.038717
Epoch[96] Batch [1099]	Speed: 5073.820691 samples/sec	accuracy=0.790864	lr=0.038588
Epoch[96] Batch [1149]	Speed: 5183.645336 samples/sec	accuracy=0.790934	lr=0.038459
Epoch[96] Batch [1199]	Speed: 5161.881435 samples/sec	accuracy=0.791034	lr=0.038330
Epoch[96] Batch [1249]	Speed: 5276.029074 samples/sec	accuracy=0.791152	lr=0.038202
[Epoch 96] training: accuracy=0.791153
[Epoch 96] speed: 5151 samples/sec	time cost: 258.008859
[Epoch 96] validation: err-top1=0.266263 err-top5=0.082689
Epoch[97] Batch [49]	Speed: 4915.662831 samples/sec	accuracy=0.791282	lr=0.038071
Epoch[97] Batch [99]	Speed: 5092.800419 samples/sec	accuracy=0.790913	lr=0.037943
Epoch[97] Batch [149]	Speed: 5168.882575 samples/sec	accuracy=0.790870	lr=0.037815
Epoch[97] Batch [199]	Speed: 5144.163675 samples/sec	accuracy=0.790694	lr=0.037687
Epoch[97] Batch [249]	Speed: 5045.707048 samples/sec	accuracy=0.790726	lr=0.037560
Epoch[97] Batch [299]	Speed: 4966.190675 samples/sec	accuracy=0.790702	lr=0.037432
Epoch[97] Batch [349]	Speed: 5029.454673 samples/sec	accuracy=0.790697	lr=0.037305
Epoch[97] Batch [399]	Speed: 5095.701196 samples/sec	accuracy=0.790794	lr=0.037178
Epoch[97] Batch [449]	Speed: 5186.500736 samples/sec	accuracy=0.790862	lr=0.037052
Epoch[97] Batch [499]	Speed: 5272.490518 samples/sec	accuracy=0.790921	lr=0.036925
Epoch[97] Batch [549]	Speed: 5168.698949 samples/sec	accuracy=0.790924	lr=0.036799
Epoch[97] Batch [599]	Speed: 5153.107112 samples/sec	accuracy=0.790997	lr=0.036673
Epoch[97] Batch [649]	Speed: 5136.031608 samples/sec	accuracy=0.791087	lr=0.036547
Epoch[97] Batch [699]	Speed: 5111.975244 samples/sec	accuracy=0.791153	lr=0.036421
Epoch[97] Batch [749]	Speed: 5387.885142 samples/sec	accuracy=0.791243	lr=0.036296
Epoch[97] Batch [799]	Speed: 5062.886594 samples/sec	accuracy=0.791309	lr=0.036170
Epoch[97] Batch [849]	Speed: 5173.270214 samples/sec	accuracy=0.791344	lr=0.036045
Epoch[97] Batch [899]	Speed: 5140.471910 samples/sec	accuracy=0.791352	lr=0.035920
Epoch[97] Batch [949]	Speed: 5142.252546 samples/sec	accuracy=0.791433	lr=0.035795
Epoch[97] Batch [999]	Speed: 5239.266620 samples/sec	accuracy=0.791492	lr=0.035671
Epoch[97] Batch [1049]	Speed: 5189.721716 samples/sec	accuracy=0.791585	lr=0.035546
Epoch[97] Batch [1099]	Speed: 5232.273023 samples/sec	accuracy=0.791673	lr=0.035422
Epoch[97] Batch [1149]	Speed: 5224.625190 samples/sec	accuracy=0.791746	lr=0.035298
Epoch[97] Batch [1199]	Speed: 5066.181461 samples/sec	accuracy=0.791854	lr=0.035174
Epoch[97] Batch [1249]	Speed: 5207.529192 samples/sec	accuracy=0.791911	lr=0.035051
[Epoch 97] training: accuracy=0.791909
[Epoch 97] speed: 5135 samples/sec	time cost: 258.448457
[Epoch 97] validation: err-top1=0.265924 err-top5=0.082848
Epoch[98] Batch [49]	Speed: 5021.487918 samples/sec	accuracy=0.792753	lr=0.034925
Epoch[98] Batch [99]	Speed: 5177.173129 samples/sec	accuracy=0.792426	lr=0.034802
Epoch[98] Batch [149]	Speed: 4920.241641 samples/sec	accuracy=0.792457	lr=0.034679
Epoch[98] Batch [199]	Speed: 5063.954633 samples/sec	accuracy=0.792290	lr=0.034556
Epoch[98] Batch [249]	Speed: 5123.996648 samples/sec	accuracy=0.792223	lr=0.034433
Epoch[98] Batch [299]	Speed: 4994.762723 samples/sec	accuracy=0.792089	lr=0.034311
Epoch[98] Batch [349]	Speed: 5234.336250 samples/sec	accuracy=0.792116	lr=0.034189
Epoch[98] Batch [399]	Speed: 5118.386251 samples/sec	accuracy=0.792057	lr=0.034067
Epoch[98] Batch [449]	Speed: 5175.308489 samples/sec	accuracy=0.792125	lr=0.033945
Epoch[98] Batch [499]	Speed: 5096.954056 samples/sec	accuracy=0.792155	lr=0.033823
Epoch[98] Batch [549]	Speed: 5229.153472 samples/sec	accuracy=0.792159	lr=0.033702
Epoch[98] Batch [599]	Speed: 5079.178066 samples/sec	accuracy=0.792223	lr=0.033581
Epoch[98] Batch [649]	Speed: 5037.052814 samples/sec	accuracy=0.792253	lr=0.033460
Epoch[98] Batch [699]	Speed: 5050.593213 samples/sec	accuracy=0.792253	lr=0.033339
Epoch[98] Batch [749]	Speed: 5318.372058 samples/sec	accuracy=0.792280	lr=0.033218
Epoch[98] Batch [799]	Speed: 5055.108156 samples/sec	accuracy=0.792283	lr=0.033098
Epoch[98] Batch [849]	Speed: 5360.201810 samples/sec	accuracy=0.792279	lr=0.032978
Epoch[98] Batch [899]	Speed: 5032.528067 samples/sec	accuracy=0.792348	lr=0.032858
Epoch[98] Batch [949]	Speed: 5151.626409 samples/sec	accuracy=0.792409	lr=0.032738
Epoch[98] Batch [999]	Speed: 5097.347856 samples/sec	accuracy=0.792498	lr=0.032618
Epoch[98] Batch [1049]	Speed: 5133.855755 samples/sec	accuracy=0.792549	lr=0.032499
Epoch[98] Batch [1099]	Speed: 5275.711645 samples/sec	accuracy=0.792602	lr=0.032380
Epoch[98] Batch [1149]	Speed: 5176.038838 samples/sec	accuracy=0.792623	lr=0.032261
Epoch[98] Batch [1199]	Speed: 5311.175283 samples/sec	accuracy=0.792676	lr=0.032142
Epoch[98] Batch [1249]	Speed: 5176.182812 samples/sec	accuracy=0.792725	lr=0.032023
[Epoch 98] training: accuracy=0.792726
[Epoch 98] speed: 5130 samples/sec	time cost: 258.324038
[Epoch 98] validation: err-top1=0.264750 err-top5=0.079895
Epoch[99] Batch [49]	Speed: 4776.459330 samples/sec	accuracy=0.794183	lr=0.031902
Epoch[99] Batch [99]	Speed: 4996.977470 samples/sec	accuracy=0.793752	lr=0.031784
Epoch[99] Batch [149]	Speed: 5181.292571 samples/sec	accuracy=0.793746	lr=0.031666
Epoch[99] Batch [199]	Speed: 4940.285238 samples/sec	accuracy=0.793496	lr=0.031548
Epoch[99] Batch [249]	Speed: 5185.558437 samples/sec	accuracy=0.793344	lr=0.031431
Epoch[99] Batch [299]	Speed: 5180.523996 samples/sec	accuracy=0.793357	lr=0.031313
Epoch[99] Batch [349]	Speed: 5173.907992 samples/sec	accuracy=0.793370	lr=0.031196
Epoch[99] Batch [399]	Speed: 5213.426179 samples/sec	accuracy=0.793186	lr=0.031079
Epoch[99] Batch [449]	Speed: 5155.863723 samples/sec	accuracy=0.793161	lr=0.030962
Epoch[99] Batch [499]	Speed: 5216.932765 samples/sec	accuracy=0.793092	lr=0.030846
Epoch[99] Batch [549]	Speed: 5178.517825 samples/sec	accuracy=0.793074	lr=0.030729
Epoch[99] Batch [599]	Speed: 5165.004816 samples/sec	accuracy=0.793043	lr=0.030613
Epoch[99] Batch [649]	Speed: 5233.770097 samples/sec	accuracy=0.793087	lr=0.030497
Epoch[99] Batch [699]	Speed: 5162.866783 samples/sec	accuracy=0.793109	lr=0.030381
Epoch[99] Batch [749]	Speed: 5132.012983 samples/sec	accuracy=0.793178	lr=0.030266
Epoch[99] Batch [799]	Speed: 5169.231326 samples/sec	accuracy=0.793209	lr=0.030150
Epoch[99] Batch [849]	Speed: 4989.810365 samples/sec	accuracy=0.793316	lr=0.030035
Epoch[99] Batch [899]	Speed: 4930.839253 samples/sec	accuracy=0.793401	lr=0.029920
Epoch[99] Batch [949]	Speed: 5095.094157 samples/sec	accuracy=0.793485	lr=0.029805
Epoch[99] Batch [999]	Speed: 5310.219442 samples/sec	accuracy=0.793522	lr=0.029691
Epoch[99] Batch [1049]	Speed: 5037.116260 samples/sec	accuracy=0.793532	lr=0.029576
Epoch[99] Batch [1099]	Speed: 5245.866148 samples/sec	accuracy=0.793606	lr=0.029462
Epoch[99] Batch [1149]	Speed: 5020.413655 samples/sec	accuracy=0.793677	lr=0.029348
Epoch[99] Batch [1199]	Speed: 5254.093822 samples/sec	accuracy=0.793788	lr=0.029234
Epoch[99] Batch [1249]	Speed: 5061.719860 samples/sec	accuracy=0.793875	lr=0.029121
[Epoch 99] training: accuracy=0.793876
[Epoch 99] speed: 5113 samples/sec	time cost: 259.925838
[Epoch 99] validation: err-top1=0.268455 err-top5=0.085818
Epoch[100] Batch [49]	Speed: 4660.676405 samples/sec	accuracy=0.793851	lr=0.029005
Epoch[100] Batch [99]	Speed: 5000.499583 samples/sec	accuracy=0.793962	lr=0.028892
Epoch[100] Batch [149]	Speed: 5022.186888 samples/sec	accuracy=0.794264	lr=0.028779
Epoch[100] Batch [199]	Speed: 5168.468937 samples/sec	accuracy=0.794312	lr=0.028666
Epoch[100] Batch [249]	Speed: 5014.642864 samples/sec	accuracy=0.794426	lr=0.028554
Epoch[100] Batch [299]	Speed: 5190.113801 samples/sec	accuracy=0.794405	lr=0.028441
Epoch[100] Batch [349]	Speed: 5126.995752 samples/sec	accuracy=0.794394	lr=0.028329
Epoch[100] Batch [399]	Speed: 5089.374235 samples/sec	accuracy=0.794270	lr=0.028217
Epoch[100] Batch [449]	Speed: 5264.443139 samples/sec	accuracy=0.794264	lr=0.028105
Epoch[100] Batch [499]	Speed: 5127.481987 samples/sec	accuracy=0.794264	lr=0.027994
Epoch[100] Batch [549]	Speed: 5090.583560 samples/sec	accuracy=0.794299	lr=0.027883
Epoch[100] Batch [599]	Speed: 5156.760959 samples/sec	accuracy=0.794323	lr=0.027772
Epoch[100] Batch [649]	Speed: 5016.749877 samples/sec	accuracy=0.794280	lr=0.027661
Epoch[100] Batch [699]	Speed: 5073.463239 samples/sec	accuracy=0.794273	lr=0.027550
Epoch[100] Batch [749]	Speed: 5111.650114 samples/sec	accuracy=0.794231	lr=0.027439
Epoch[100] Batch [799]	Speed: 5180.730085 samples/sec	accuracy=0.794188	lr=0.027329
Epoch[100] Batch [849]	Speed: 5237.937336 samples/sec	accuracy=0.794246	lr=0.027219
Epoch[100] Batch [899]	Speed: 4981.029262 samples/sec	accuracy=0.794290	lr=0.027109
Epoch[100] Batch [949]	Speed: 5118.443467 samples/sec	accuracy=0.794397	lr=0.026999
Epoch[100] Batch [999]	Speed: 5244.450262 samples/sec	accuracy=0.794429	lr=0.026890
Epoch[100] Batch [1049]	Speed: 4979.935745 samples/sec	accuracy=0.794494	lr=0.026781
Epoch[100] Batch [1099]	Speed: 5187.261188 samples/sec	accuracy=0.794520	lr=0.026672
Epoch[100] Batch [1149]	Speed: 4995.719579 samples/sec	accuracy=0.794567	lr=0.026563
Epoch[100] Batch [1199]	Speed: 5146.334217 samples/sec	accuracy=0.794613	lr=0.026454
Epoch[100] Batch [1249]	Speed: 5316.009261 samples/sec	accuracy=0.794672	lr=0.026346
[Epoch 100] training: accuracy=0.794676
[Epoch 100] speed: 5093 samples/sec	time cost: 261.189029
[Epoch 100] validation: err-top1=0.264429 err-top5=0.083048
Epoch[101] Batch [49]	Speed: 4973.919435 samples/sec	accuracy=0.795768	lr=0.026235
Epoch[101] Batch [99]	Speed: 5240.835869 samples/sec	accuracy=0.795580	lr=0.026127
Epoch[101] Batch [149]	Speed: 4945.414979 samples/sec	accuracy=0.795162	lr=0.026020
Epoch[101] Batch [199]	Speed: 5043.198638 samples/sec	accuracy=0.795072	lr=0.025912
Epoch[101] Batch [249]	Speed: 5178.845148 samples/sec	accuracy=0.795015	lr=0.025805
Epoch[101] Batch [299]	Speed: 5054.318148 samples/sec	accuracy=0.794911	lr=0.025697
Epoch[101] Batch [349]	Speed: 5204.202963 samples/sec	accuracy=0.794789	lr=0.025590
Epoch[101] Batch [399]	Speed: 4910.834793 samples/sec	accuracy=0.794705	lr=0.025484
Epoch[101] Batch [449]	Speed: 5241.380909 samples/sec	accuracy=0.794738	lr=0.025377
Epoch[101] Batch [499]	Speed: 5192.156460 samples/sec	accuracy=0.794758	lr=0.025271
Epoch[101] Batch [549]	Speed: 5031.539966 samples/sec	accuracy=0.794724	lr=0.025165
Epoch[101] Batch [599]	Speed: 5193.464113 samples/sec	accuracy=0.794793	lr=0.025059
Epoch[101] Batch [649]	Speed: 5300.486296 samples/sec	accuracy=0.794823	lr=0.024953
Epoch[101] Batch [699]	Speed: 5130.129375 samples/sec	accuracy=0.794871	lr=0.024847
Epoch[101] Batch [749]	Speed: 5101.248372 samples/sec	accuracy=0.794931	lr=0.024742
Epoch[101] Batch [799]	Speed: 5342.554777 samples/sec	accuracy=0.795035	lr=0.024637
Epoch[101] Batch [849]	Speed: 5219.247614 samples/sec	accuracy=0.795162	lr=0.024532
Epoch[101] Batch [899]	Speed: 5256.141363 samples/sec	accuracy=0.795302	lr=0.024427
Epoch[101] Batch [949]	Speed: 5324.325806 samples/sec	accuracy=0.795313	lr=0.024323
Epoch[101] Batch [999]	Speed: 5277.177531 samples/sec	accuracy=0.795340	lr=0.024219
Epoch[101] Batch [1049]	Speed: 5309.351631 samples/sec	accuracy=0.795411	lr=0.024115
Epoch[101] Batch [1099]	Speed: 5183.113738 samples/sec	accuracy=0.795485	lr=0.024011
Epoch[101] Batch [1149]	Speed: 5224.111207 samples/sec	accuracy=0.795559	lr=0.023907
Epoch[101] Batch [1199]	Speed: 5179.461815 samples/sec	accuracy=0.795658	lr=0.023804
Epoch[101] Batch [1249]	Speed: 5309.862699 samples/sec	accuracy=0.795715	lr=0.023700
[Epoch 101] training: accuracy=0.795716
[Epoch 101] speed: 5165 samples/sec	time cost: 257.220854
[Epoch 101] validation: err-top1=0.256956 err-top5=0.079082
Epoch[102] Batch [49]	Speed: 4967.798584 samples/sec	accuracy=0.796628	lr=0.023595
Epoch[102] Batch [99]	Speed: 5195.273738 samples/sec	accuracy=0.795786	lr=0.023493
Epoch[102] Batch [149]	Speed: 5309.553788 samples/sec	accuracy=0.796036	lr=0.023390
Epoch[102] Batch [199]	Speed: 5014.496496 samples/sec	accuracy=0.796037	lr=0.023288
Epoch[102] Batch [249]	Speed: 5282.914977 samples/sec	accuracy=0.795979	lr=0.023185
Epoch[102] Batch [299]	Speed: 5203.487340 samples/sec	accuracy=0.795889	lr=0.023083
Epoch[102] Batch [349]	Speed: 5085.984939 samples/sec	accuracy=0.795768	lr=0.022982
Epoch[102] Batch [399]	Speed: 4917.777113 samples/sec	accuracy=0.795864	lr=0.022880
Epoch[102] Batch [449]	Speed: 4963.355018 samples/sec	accuracy=0.795847	lr=0.022779
Epoch[102] Batch [499]	Speed: 5229.047663 samples/sec	accuracy=0.795861	lr=0.022678
Epoch[102] Batch [549]	Speed: 5244.707197 samples/sec	accuracy=0.795843	lr=0.022577
Epoch[102] Batch [599]	Speed: 5111.920606 samples/sec	accuracy=0.795908	lr=0.022476
Epoch[102] Batch [649]	Speed: 5264.462110 samples/sec	accuracy=0.795981	lr=0.022376
Epoch[102] Batch [699]	Speed: 5321.209213 samples/sec	accuracy=0.795993	lr=0.022276
Epoch[102] Batch [749]	Speed: 5243.798689 samples/sec	accuracy=0.795957	lr=0.022175
Epoch[102] Batch [799]	Speed: 5209.215066 samples/sec	accuracy=0.795997	lr=0.022076
Epoch[102] Batch [849]	Speed: 5048.176539 samples/sec	accuracy=0.796034	lr=0.021976
Epoch[102] Batch [899]	Speed: 5178.075176 samples/sec	accuracy=0.796116	lr=0.021877
Epoch[102] Batch [949]	Speed: 5060.486167 samples/sec	accuracy=0.796173	lr=0.021777
Epoch[102] Batch [999]	Speed: 5300.586774 samples/sec	accuracy=0.796233	lr=0.021678
Epoch[102] Batch [1049]	Speed: 5285.752798 samples/sec	accuracy=0.796269	lr=0.021580
Epoch[102] Batch [1099]	Speed: 5158.069178 samples/sec	accuracy=0.796312	lr=0.021481
Epoch[102] Batch [1149]	Speed: 5156.067731 samples/sec	accuracy=0.796345	lr=0.021383
Epoch[102] Batch [1199]	Speed: 5230.744316 samples/sec	accuracy=0.796406	lr=0.021285
Epoch[102] Batch [1249]	Speed: 5261.562271 samples/sec	accuracy=0.796496	lr=0.021187
[Epoch 102] training: accuracy=0.796496
[Epoch 102] speed: 5163 samples/sec	time cost: 257.313206
[Epoch 102] validation: err-top1=0.248485 err-top5=0.075415
Epoch[103] Batch [49]	Speed: 4836.003724 samples/sec	accuracy=0.797972	lr=0.021087
Epoch[103] Batch [99]	Speed: 5167.651435 samples/sec	accuracy=0.797583	lr=0.020990
Epoch[103] Batch [149]	Speed: 5110.012927 samples/sec	accuracy=0.797091	lr=0.020892
Epoch[103] Batch [199]	Speed: 5167.365066 samples/sec	accuracy=0.797149	lr=0.020795
Epoch[103] Batch [249]	Speed: 5303.964291 samples/sec	accuracy=0.797047	lr=0.020698
Epoch[103] Batch [299]	Speed: 5031.276499 samples/sec	accuracy=0.797068	lr=0.020602
Epoch[103] Batch [349]	Speed: 5140.737216 samples/sec	accuracy=0.796964	lr=0.020505
Epoch[103] Batch [399]	Speed: 5173.261241 samples/sec	accuracy=0.796868	lr=0.020409
Epoch[103] Batch [449]	Speed: 5281.421872 samples/sec	accuracy=0.796912	lr=0.020313
Epoch[103] Batch [499]	Speed: 5122.701742 samples/sec	accuracy=0.796991	lr=0.020217
Epoch[103] Batch [549]	Speed: 5462.873490 samples/sec	accuracy=0.797053	lr=0.020122
Epoch[103] Batch [599]	Speed: 5314.976830 samples/sec	accuracy=0.797077	lr=0.020026
Epoch[103] Batch [649]	Speed: 5078.039114 samples/sec	accuracy=0.797181	lr=0.019931
Epoch[103] Batch [699]	Speed: 5020.956658 samples/sec	accuracy=0.797185	lr=0.019836
Epoch[103] Batch [749]	Speed: 5235.888504 samples/sec	accuracy=0.797185	lr=0.019742
Epoch[103] Batch [799]	Speed: 5255.833911 samples/sec	accuracy=0.797190	lr=0.019647
Epoch[103] Batch [849]	Speed: 5218.856695 samples/sec	accuracy=0.797213	lr=0.019553
Epoch[103] Batch [899]	Speed: 5142.501041 samples/sec	accuracy=0.797242	lr=0.019459
Epoch[103] Batch [949]	Speed: 5067.978327 samples/sec	accuracy=0.797261	lr=0.019365
Epoch[103] Batch [999]	Speed: 5142.388366 samples/sec	accuracy=0.797302	lr=0.019271
Epoch[103] Batch [1049]	Speed: 5214.407882 samples/sec	accuracy=0.797343	lr=0.019178
Epoch[103] Batch [1099]	Speed: 5163.562337 samples/sec	accuracy=0.797377	lr=0.019085
Epoch[103] Batch [1149]	Speed: 5088.301355 samples/sec	accuracy=0.797396	lr=0.018992
Epoch[103] Batch [1199]	Speed: 5050.654981 samples/sec	accuracy=0.797477	lr=0.018899
Epoch[103] Batch [1249]	Speed: 5239.625446 samples/sec	accuracy=0.797509	lr=0.018806
[Epoch 103] training: accuracy=0.797511
[Epoch 103] speed: 5153 samples/sec	time cost: 257.656253
[Epoch 103] validation: err-top1=0.247887 err-top5=0.073641
Epoch[104] Batch [49]	Speed: 4682.067938 samples/sec	accuracy=0.798036	lr=0.018712
Epoch[104] Batch [99]	Speed: 5008.336052 samples/sec	accuracy=0.797805	lr=0.018620
Epoch[104] Batch [149]	Speed: 5028.856955 samples/sec	accuracy=0.797609	lr=0.018528
Epoch[104] Batch [199]	Speed: 5076.612152 samples/sec	accuracy=0.797441	lr=0.018436
Epoch[104] Batch [249]	Speed: 5068.489918 samples/sec	accuracy=0.797520	lr=0.018345
Epoch[104] Batch [299]	Speed: 5000.306186 samples/sec	accuracy=0.797579	lr=0.018254
Epoch[104] Batch [349]	Speed: 4881.278111 samples/sec	accuracy=0.797538	lr=0.018163
Epoch[104] Batch [399]	Speed: 5210.887239 samples/sec	accuracy=0.797553	lr=0.018072
Epoch[104] Batch [449]	Speed: 5400.390375 samples/sec	accuracy=0.797628	lr=0.017981
Epoch[104] Batch [499]	Speed: 4886.419848 samples/sec	accuracy=0.797702	lr=0.017891
Epoch[104] Batch [549]	Speed: 5112.859947 samples/sec	accuracy=0.797753	lr=0.017801
Epoch[104] Batch [599]	Speed: 5203.138363 samples/sec	accuracy=0.797811	lr=0.017711
Epoch[104] Batch [649]	Speed: 5074.456725 samples/sec	accuracy=0.797874	lr=0.017621
Epoch[104] Batch [699]	Speed: 5263.956646 samples/sec	accuracy=0.797879	lr=0.017532
Epoch[104] Batch [749]	Speed: 5281.223020 samples/sec	accuracy=0.798007	lr=0.017442
Epoch[104] Batch [799]	Speed: 5401.612502 samples/sec	accuracy=0.798018	lr=0.017353
Epoch[104] Batch [849]	Speed: 5210.048933 samples/sec	accuracy=0.798043	lr=0.017264
Epoch[104] Batch [899]	Speed: 5214.104154 samples/sec	accuracy=0.798110	lr=0.017176
Epoch[104] Batch [949]	Speed: 5339.744344 samples/sec	accuracy=0.798169	lr=0.017087
Epoch[104] Batch [999]	Speed: 5258.584758 samples/sec	accuracy=0.798242	lr=0.016999
Epoch[104] Batch [1049]	Speed: 5239.297170 samples/sec	accuracy=0.798315	lr=0.016911
Epoch[104] Batch [1099]	Speed: 5266.244589 samples/sec	accuracy=0.798345	lr=0.016823
Epoch[104] Batch [1149]	Speed: 5235.507853 samples/sec	accuracy=0.798388	lr=0.016736
Epoch[104] Batch [1199]	Speed: 5268.773268 samples/sec	accuracy=0.798418	lr=0.016648
Epoch[104] Batch [1249]	Speed: 5329.800643 samples/sec	accuracy=0.798462	lr=0.016561
[Epoch 104] training: accuracy=0.798465
[Epoch 104] speed: 5147 samples/sec	time cost: 258.007037
[Epoch 104] validation: err-top1=0.243876 err-top5=0.072001
Epoch[105] Batch [49]	Speed: 5024.641751 samples/sec	accuracy=0.798475	lr=0.016473
Epoch[105] Batch [99]	Speed: 5318.862733 samples/sec	accuracy=0.798118	lr=0.016386
Epoch[105] Batch [149]	Speed: 5022.809689 samples/sec	accuracy=0.798133	lr=0.016299
Epoch[105] Batch [199]	Speed: 5044.400098 samples/sec	accuracy=0.798219	lr=0.016213
Epoch[105] Batch [249]	Speed: 5136.134179 samples/sec	accuracy=0.798200	lr=0.016127
Epoch[105] Batch [299]	Speed: 5323.990264 samples/sec	accuracy=0.798204	lr=0.016041
Epoch[105] Batch [349]	Speed: 5218.273725 samples/sec	accuracy=0.798315	lr=0.015956
Epoch[105] Batch [399]	Speed: 5058.531835 samples/sec	accuracy=0.798515	lr=0.015870
Epoch[105] Batch [449]	Speed: 5131.064141 samples/sec	accuracy=0.798560	lr=0.015785
Epoch[105] Batch [499]	Speed: 5204.287086 samples/sec	accuracy=0.798582	lr=0.015700
Epoch[105] Batch [549]	Speed: 5290.118459 samples/sec	accuracy=0.798656	lr=0.015616
Epoch[105] Batch [599]	Speed: 5303.017723 samples/sec	accuracy=0.798780	lr=0.015531
Epoch[105] Batch [649]	Speed: 5260.226676 samples/sec	accuracy=0.798834	lr=0.015447
Epoch[105] Batch [699]	Speed: 4974.985186 samples/sec	accuracy=0.798913	lr=0.015363
Epoch[105] Batch [749]	Speed: 5100.670056 samples/sec	accuracy=0.798942	lr=0.015279
Epoch[105] Batch [799]	Speed: 5140.199987 samples/sec	accuracy=0.798963	lr=0.015195
Epoch[105] Batch [849]	Speed: 5265.010397 samples/sec	accuracy=0.799019	lr=0.015112
Epoch[105] Batch [899]	Speed: 5002.921164 samples/sec	accuracy=0.799076	lr=0.015029
Epoch[105] Batch [949]	Speed: 5048.183422 samples/sec	accuracy=0.799114	lr=0.014946
Epoch[105] Batch [999]	Speed: 4941.257829 samples/sec	accuracy=0.799151	lr=0.014863
Epoch[105] Batch [1049]	Speed: 5236.669382 samples/sec	accuracy=0.799201	lr=0.014781
Epoch[105] Batch [1099]	Speed: 5164.564474 samples/sec	accuracy=0.799248	lr=0.014699
Epoch[105] Batch [1149]	Speed: 5298.992527 samples/sec	accuracy=0.799241	lr=0.014616
Epoch[105] Batch [1199]	Speed: 5268.212695 samples/sec	accuracy=0.799288	lr=0.014535
Epoch[105] Batch [1249]	Speed: 5500.917624 samples/sec	accuracy=0.799330	lr=0.014453
[Epoch 105] training: accuracy=0.799330
[Epoch 105] speed: 5163 samples/sec	time cost: 257.355808
[Epoch 105] validation: err-top1=0.239995 err-top5=0.069774
Epoch[106] Batch [49]	Speed: 4846.419242 samples/sec	accuracy=0.799501	lr=0.014370
Epoch[106] Batch [99]	Speed: 4919.553290 samples/sec	accuracy=0.799530	lr=0.014289
Epoch[106] Batch [149]	Speed: 5132.654000 samples/sec	accuracy=0.799751	lr=0.014208
Epoch[106] Batch [199]	Speed: 4973.192374 samples/sec	accuracy=0.799984	lr=0.014127
Epoch[106] Batch [249]	Speed: 5152.152061 samples/sec	accuracy=0.799898	lr=0.014047
Epoch[106] Batch [299]	Speed: 5065.283804 samples/sec	accuracy=0.799728	lr=0.013966
Epoch[106] Batch [349]	Speed: 5006.968770 samples/sec	accuracy=0.799611	lr=0.013886
Epoch[106] Batch [399]	Speed: 4960.498649 samples/sec	accuracy=0.799610	lr=0.013806
Epoch[106] Batch [449]	Speed: 5364.489279 samples/sec	accuracy=0.799725	lr=0.013727
Epoch[106] Batch [499]	Speed: 5070.244129 samples/sec	accuracy=0.799672	lr=0.013647
Epoch[106] Batch [549]	Speed: 5238.760615 samples/sec	accuracy=0.799707	lr=0.013568
Epoch[106] Batch [599]	Speed: 5265.470746 samples/sec	accuracy=0.799737	lr=0.013489
Epoch[106] Batch [649]	Speed: 5163.430735 samples/sec	accuracy=0.799701	lr=0.013411
Epoch[106] Batch [699]	Speed: 5271.725580 samples/sec	accuracy=0.799735	lr=0.013332
Epoch[106] Batch [749]	Speed: 5095.883783 samples/sec	accuracy=0.799740	lr=0.013254
Epoch[106] Batch [799]	Speed: 5033.690349 samples/sec	accuracy=0.799781	lr=0.013176
Epoch[106] Batch [849]	Speed: 4988.698040 samples/sec	accuracy=0.799805	lr=0.013098
Epoch[106] Batch [899]	Speed: 5292.001688 samples/sec	accuracy=0.799833	lr=0.013020
Epoch[106] Batch [949]	Speed: 5327.697702 samples/sec	accuracy=0.799823	lr=0.012943
Epoch[106] Batch [999]	Speed: 5189.750688 samples/sec	accuracy=0.799881	lr=0.012866
Epoch[106] Batch [1049]	Speed: 5318.255495 samples/sec	accuracy=0.799939	lr=0.012789
Epoch[106] Batch [1099]	Speed: 5261.142819 samples/sec	accuracy=0.799984	lr=0.012712
Epoch[106] Batch [1149]	Speed: 5194.377751 samples/sec	accuracy=0.800048	lr=0.012636
Epoch[106] Batch [1199]	Speed: 5168.941671 samples/sec	accuracy=0.800127	lr=0.012559
Epoch[106] Batch [1249]	Speed: 4992.298546 samples/sec	accuracy=0.800230	lr=0.012483
[Epoch 106] training: accuracy=0.800233
[Epoch 106] speed: 5122 samples/sec	time cost: 259.185056
[Epoch 106] validation: err-top1=0.239935 err-top5=0.069575
Epoch[107] Batch [49]	Speed: 4976.834306 samples/sec	accuracy=0.800662	lr=0.012406
Epoch[107] Batch [99]	Speed: 4830.022832 samples/sec	accuracy=0.800874	lr=0.012330
Epoch[107] Batch [149]	Speed: 5137.278205 samples/sec	accuracy=0.800552	lr=0.012255
Epoch[107] Batch [199]	Speed: 5127.579074 samples/sec	accuracy=0.800469	lr=0.012180
Epoch[107] Batch [249]	Speed: 5166.268626 samples/sec	accuracy=0.800455	lr=0.012105
Epoch[107] Batch [299]	Speed: 5215.825706 samples/sec	accuracy=0.800571	lr=0.012030
Epoch[107] Batch [349]	Speed: 5201.764345 samples/sec	accuracy=0.800618	lr=0.011956
Epoch[107] Batch [399]	Speed: 5494.503160 samples/sec	accuracy=0.800687	lr=0.011881
Epoch[107] Batch [449]	Speed: 4951.959169 samples/sec	accuracy=0.800703	lr=0.011807
Epoch[107] Batch [499]	Speed: 5116.833746 samples/sec	accuracy=0.800763	lr=0.011734
Epoch[107] Batch [549]	Speed: 5244.596914 samples/sec	accuracy=0.800802	lr=0.011660
Epoch[107] Batch [599]	Speed: 5239.009835 samples/sec	accuracy=0.800880	lr=0.011587
Epoch[107] Batch [649]	Speed: 5225.258149 samples/sec	accuracy=0.800964	lr=0.011513
Epoch[107] Batch [699]	Speed: 5249.573591 samples/sec	accuracy=0.800961	lr=0.011441
Epoch[107] Batch [749]	Speed: 5174.382220 samples/sec	accuracy=0.800945	lr=0.011368
Epoch[107] Batch [799]	Speed: 5257.572837 samples/sec	accuracy=0.800991	lr=0.011295
Epoch[107] Batch [849]	Speed: 5271.289238 samples/sec	accuracy=0.800998	lr=0.011223
Epoch[107] Batch [899]	Speed: 5114.176553 samples/sec	accuracy=0.801045	lr=0.011151
Epoch[107] Batch [949]	Speed: 5043.622436 samples/sec	accuracy=0.801096	lr=0.011079
Epoch[107] Batch [999]	Speed: 5126.751934 samples/sec	accuracy=0.801106	lr=0.011008
Epoch[107] Batch [1049]	Speed: 5137.013747 samples/sec	accuracy=0.801142	lr=0.010936
Epoch[107] Batch [1099]	Speed: 5337.769274 samples/sec	accuracy=0.801172	lr=0.010865
Epoch[107] Batch [1149]	Speed: 5194.962810 samples/sec	accuracy=0.801204	lr=0.010794
Epoch[107] Batch [1199]	Speed: 5023.770271 samples/sec	accuracy=0.801235	lr=0.010724
Epoch[107] Batch [1249]	Speed: 5213.201660 samples/sec	accuracy=0.801284	lr=0.010653
[Epoch 107] training: accuracy=0.801286
[Epoch 107] speed: 5155 samples/sec	time cost: 258.149542
[Epoch 107] validation: err-top1=0.237604 err-top5=0.067981
Epoch[108] Batch [49]	Speed: 4940.906420 samples/sec	accuracy=0.801998	lr=0.010582
Epoch[108] Batch [99]	Speed: 5170.641742 samples/sec	accuracy=0.801708	lr=0.010512
Epoch[108] Batch [149]	Speed: 5140.790379 samples/sec	accuracy=0.801329	lr=0.010442
Epoch[108] Batch [199]	Speed: 5131.622638 samples/sec	accuracy=0.801450	lr=0.010373
Epoch[108] Batch [249]	Speed: 5088.312929 samples/sec	accuracy=0.801565	lr=0.010303
Epoch[108] Batch [299]	Speed: 5049.038345 samples/sec	accuracy=0.801550	lr=0.010234
Epoch[108] Batch [349]	Speed: 4923.477905 samples/sec	accuracy=0.801605	lr=0.010165
Epoch[108] Batch [399]	Speed: 5030.835562 samples/sec	accuracy=0.801613	lr=0.010097
Epoch[108] Batch [449]	Speed: 5126.196088 samples/sec	accuracy=0.801645	lr=0.010028
Epoch[108] Batch [499]	Speed: 5238.328690 samples/sec	accuracy=0.801611	lr=0.009960
Epoch[108] Batch [549]	Speed: 5297.078711 samples/sec	accuracy=0.801611	lr=0.009892
Epoch[108] Batch [599]	Speed: 5090.147249 samples/sec	accuracy=0.801631	lr=0.009824
Epoch[108] Batch [649]	Speed: 4997.858405 samples/sec	accuracy=0.801626	lr=0.009757
Epoch[108] Batch [699]	Speed: 5169.383756 samples/sec	accuracy=0.801615	lr=0.009690
Epoch[108] Batch [749]	Speed: 5177.818611 samples/sec	accuracy=0.801659	lr=0.009623
Epoch[108] Batch [799]	Speed: 5195.683884 samples/sec	accuracy=0.801654	lr=0.009556
Epoch[108] Batch [849]	Speed: 5052.124793 samples/sec	accuracy=0.801697	lr=0.009489
Epoch[108] Batch [899]	Speed: 5166.327414 samples/sec	accuracy=0.801715	lr=0.009423
Epoch[108] Batch [949]	Speed: 5374.952261 samples/sec	accuracy=0.801719	lr=0.009357
Epoch[108] Batch [999]	Speed: 5192.502834 samples/sec	accuracy=0.801746	lr=0.009291
Epoch[108] Batch [1049]	Speed: 5266.621198 samples/sec	accuracy=0.801815	lr=0.009225
Epoch[108] Batch [1099]	Speed: 4942.480932 samples/sec	accuracy=0.801865	lr=0.009160
Epoch[108] Batch [1149]	Speed: 5236.321175 samples/sec	accuracy=0.801903	lr=0.009095
Epoch[108] Batch [1199]	Speed: 5199.259155 samples/sec	accuracy=0.801954	lr=0.009030
Epoch[108] Batch [1249]	Speed: 5229.854776 samples/sec	accuracy=0.801974	lr=0.008965
[Epoch 108] training: accuracy=0.801975
[Epoch 108] speed: 5130 samples/sec	time cost: 258.983150
[Epoch 108] validation: err-top1=0.234734 err-top5=0.067303
Epoch[109] Batch [49]	Speed: 5027.623456 samples/sec	accuracy=0.802352	lr=0.008899
Epoch[109] Batch [99]	Speed: 5003.104973 samples/sec	accuracy=0.802190	lr=0.008835
Epoch[109] Batch [149]	Speed: 5160.692569 samples/sec	accuracy=0.802072	lr=0.008771
Epoch[109] Batch [199]	Speed: 5183.485682 samples/sec	accuracy=0.801960	lr=0.008707
Epoch[109] Batch [249]	Speed: 5023.467075 samples/sec	accuracy=0.801936	lr=0.008643
Epoch[109] Batch [299]	Speed: 4956.607938 samples/sec	accuracy=0.802052	lr=0.008580
Epoch[109] Batch [349]	Speed: 5439.514326 samples/sec	accuracy=0.802171	lr=0.008517
Epoch[109] Batch [399]	Speed: 5153.000895 samples/sec	accuracy=0.802331	lr=0.008454
Epoch[109] Batch [449]	Speed: 5200.678072 samples/sec	accuracy=0.802420	lr=0.008391
Epoch[109] Batch [499]	Speed: 5323.375256 samples/sec	accuracy=0.802430	lr=0.008329
Epoch[109] Batch [549]	Speed: 5365.103100 samples/sec	accuracy=0.802469	lr=0.008266
Epoch[109] Batch [599]	Speed: 5318.362180 samples/sec	accuracy=0.802479	lr=0.008204
Epoch[109] Batch [649]	Speed: 5141.743930 samples/sec	accuracy=0.802528	lr=0.008143
Epoch[109] Batch [699]	Speed: 5141.204399 samples/sec	accuracy=0.802559	lr=0.008081
Epoch[109] Batch [749]	Speed: 5208.526234 samples/sec	accuracy=0.802604	lr=0.008020
Epoch[109] Batch [799]	Speed: 5093.567950 samples/sec	accuracy=0.802622	lr=0.007959
Epoch[109] Batch [849]	Speed: 5019.104993 samples/sec	accuracy=0.802650	lr=0.007898
Epoch[109] Batch [899]	Speed: 5228.411622 samples/sec	accuracy=0.802697	lr=0.007837
Epoch[109] Batch [949]	Speed: 5303.112666 samples/sec	accuracy=0.802708	lr=0.007777
Epoch[109] Batch [999]	Speed: 5291.090020 samples/sec	accuracy=0.802787	lr=0.007716
Epoch[109] Batch [1049]	Speed: 5160.961455 samples/sec	accuracy=0.802847	lr=0.007656
Epoch[109] Batch [1099]	Speed: 5013.410357 samples/sec	accuracy=0.802891	lr=0.007597
Epoch[109] Batch [1149]	Speed: 5260.083916 samples/sec	accuracy=0.802921	lr=0.007537
Epoch[109] Batch [1199]	Speed: 5188.717316 samples/sec	accuracy=0.802971	lr=0.007478
Epoch[109] Batch [1249]	Speed: 5340.323033 samples/sec	accuracy=0.803003	lr=0.007419
[Epoch 109] training: accuracy=0.803004
[Epoch 109] speed: 5174 samples/sec	time cost: 256.618408
[Epoch 109] validation: err-top1=0.233020 err-top5=0.067184
Epoch[110] Batch [49]	Speed: 4841.022972 samples/sec	accuracy=0.803527	lr=0.007359
Epoch[110] Batch [99]	Speed: 5241.354044 samples/sec	accuracy=0.803681	lr=0.007300
Epoch[110] Batch [149]	Speed: 5001.863917 samples/sec	accuracy=0.803596	lr=0.007242
Epoch[110] Batch [199]	Speed: 5061.040378 samples/sec	accuracy=0.803588	lr=0.007184
Epoch[110] Batch [249]	Speed: 4992.951800 samples/sec	accuracy=0.803566	lr=0.007126
Epoch[110] Batch [299]	Speed: 5280.112009 samples/sec	accuracy=0.803444	lr=0.007068
Epoch[110] Batch [349]	Speed: 4936.097333 samples/sec	accuracy=0.803467	lr=0.007011
Epoch[110] Batch [399]	Speed: 5078.728694 samples/sec	accuracy=0.803479	lr=0.006954
Epoch[110] Batch [449]	Speed: 5214.793955 samples/sec	accuracy=0.803438	lr=0.006897
Epoch[110] Batch [499]	Speed: 4949.821675 samples/sec	accuracy=0.803453	lr=0.006840
Epoch[110] Batch [549]	Speed: 5078.073457 samples/sec	accuracy=0.803411	lr=0.006784
Epoch[110] Batch [599]	Speed: 5392.286759 samples/sec	accuracy=0.803380	lr=0.006727
Epoch[110] Batch [649]	Speed: 5111.459339 samples/sec	accuracy=0.803377	lr=0.006671
Epoch[110] Batch [699]	Speed: 5153.925709 samples/sec	accuracy=0.803402	lr=0.006615
Epoch[110] Batch [749]	Speed: 5079.304087 samples/sec	accuracy=0.803404	lr=0.006560
Epoch[110] Batch [799]	Speed: 5073.702494 samples/sec	accuracy=0.803408	lr=0.006504
Epoch[110] Batch [849]	Speed: 4929.679844 samples/sec	accuracy=0.803433	lr=0.006449
Epoch[110] Batch [899]	Speed: 5190.786227 samples/sec	accuracy=0.803445	lr=0.006394
Epoch[110] Batch [949]	Speed: 5180.874820 samples/sec	accuracy=0.803448	lr=0.006340
Epoch[110] Batch [999]	Speed: 5379.200039 samples/sec	accuracy=0.803465	lr=0.006285
Epoch[110] Batch [1049]	Speed: 5374.881769 samples/sec	accuracy=0.803473	lr=0.006231
Epoch[110] Batch [1099]	Speed: 4971.564974 samples/sec	accuracy=0.803490	lr=0.006177
Epoch[110] Batch [1149]	Speed: 5355.997597 samples/sec	accuracy=0.803532	lr=0.006123
Epoch[110] Batch [1199]	Speed: 5158.993704 samples/sec	accuracy=0.803561	lr=0.006070
Epoch[110] Batch [1249]	Speed: 5333.729985 samples/sec	accuracy=0.803601	lr=0.006017
[Epoch 110] training: accuracy=0.803601
[Epoch 110] speed: 5125 samples/sec	time cost: 259.092238
[Epoch 110] validation: err-top1=0.232076 err-top5=0.065877
Epoch[111] Batch [49]	Speed: 5082.674858 samples/sec	accuracy=0.804118	lr=0.005963
Epoch[111] Batch [99]	Speed: 5245.454576 samples/sec	accuracy=0.804346	lr=0.005910
Epoch[111] Batch [149]	Speed: 5146.265524 samples/sec	accuracy=0.804171	lr=0.005857
Epoch[111] Batch [199]	Speed: 5124.750130 samples/sec	accuracy=0.804058	lr=0.005805
Epoch[111] Batch [249]	Speed: 5241.145918 samples/sec	accuracy=0.803989	lr=0.005753
Epoch[111] Batch [299]	Speed: 5076.601111 samples/sec	accuracy=0.803969	lr=0.005701
Epoch[111] Batch [349]	Speed: 5127.079967 samples/sec	accuracy=0.803982	lr=0.005649
Epoch[111] Batch [399]	Speed: 5108.464646 samples/sec	accuracy=0.804051	lr=0.005598
Epoch[111] Batch [449]	Speed: 5028.360163 samples/sec	accuracy=0.804086	lr=0.005547
Epoch[111] Batch [499]	Speed: 4996.794460 samples/sec	accuracy=0.804171	lr=0.005496
Epoch[111] Batch [549]	Speed: 5314.292097 samples/sec	accuracy=0.804116	lr=0.005445
Epoch[111] Batch [599]	Speed: 5151.239252 samples/sec	accuracy=0.804129	lr=0.005394
Epoch[111] Batch [649]	Speed: 5024.154607 samples/sec	accuracy=0.804186	lr=0.005344
Epoch[111] Batch [699]	Speed: 5085.676717 samples/sec	accuracy=0.804244	lr=0.005294
Epoch[111] Batch [749]	Speed: 4969.693199 samples/sec	accuracy=0.804272	lr=0.005244
Epoch[111] Batch [799]	Speed: 5033.193072 samples/sec	accuracy=0.804286	lr=0.005195
Epoch[111] Batch [849]	Speed: 5094.775281 samples/sec	accuracy=0.804309	lr=0.005145
Epoch[111] Batch [899]	Speed: 5294.886358 samples/sec	accuracy=0.804335	lr=0.005096
Epoch[111] Batch [949]	Speed: 5123.198408 samples/sec	accuracy=0.804313	lr=0.005047
Epoch[111] Batch [999]	Speed: 5342.109953 samples/sec	accuracy=0.804301	lr=0.004999
Epoch[111] Batch [1049]	Speed: 5184.216465 samples/sec	accuracy=0.804317	lr=0.004950
Epoch[111] Batch [1099]	Speed: 5219.006105 samples/sec	accuracy=0.804337	lr=0.004902
Epoch[111] Batch [1149]	Speed: 5167.249930 samples/sec	accuracy=0.804359	lr=0.004854
Epoch[111] Batch [1199]	Speed: 5170.620951 samples/sec	accuracy=0.804400	lr=0.004807
Epoch[111] Batch [1249]	Speed: 5263.164901 samples/sec	accuracy=0.804433	lr=0.004759
[Epoch 111] training: accuracy=0.804433
[Epoch 111] speed: 5137 samples/sec	time cost: 258.951449
[Epoch 111] validation: err-top1=0.230489 err-top5=0.066008
Epoch[112] Batch [49]	Speed: 4908.868978 samples/sec	accuracy=0.804724	lr=0.004711
Epoch[112] Batch [99]	Speed: 5169.185412 samples/sec	accuracy=0.804705	lr=0.004664
Epoch[112] Batch [149]	Speed: 5123.572193 samples/sec	accuracy=0.804615	lr=0.004617
Epoch[112] Batch [199]	Speed: 5234.377970 samples/sec	accuracy=0.804712	lr=0.004571
Epoch[112] Batch [249]	Speed: 5156.375630 samples/sec	accuracy=0.804701	lr=0.004524
Epoch[112] Batch [299]	Speed: 4890.281065 samples/sec	accuracy=0.804682	lr=0.004478
Epoch[112] Batch [349]	Speed: 4959.289748 samples/sec	accuracy=0.804735	lr=0.004432
Epoch[112] Batch [399]	Speed: 5304.292335 samples/sec	accuracy=0.804736	lr=0.004387
Epoch[112] Batch [449]	Speed: 5293.762936 samples/sec	accuracy=0.804688	lr=0.004341
Epoch[112] Batch [499]	Speed: 5227.726740 samples/sec	accuracy=0.804685	lr=0.004296
Epoch[112] Batch [549]	Speed: 5175.125528 samples/sec	accuracy=0.804673	lr=0.004251
Epoch[112] Batch [599]	Speed: 5161.840863 samples/sec	accuracy=0.804708	lr=0.004207
Epoch[112] Batch [649]	Speed: 5250.021490 samples/sec	accuracy=0.804679	lr=0.004162
Epoch[112] Batch [699]	Speed: 5042.379670 samples/sec	accuracy=0.804659	lr=0.004118
Epoch[112] Batch [749]	Speed: 5114.652563 samples/sec	accuracy=0.804693	lr=0.004074
Epoch[112] Batch [799]	Speed: 5022.307983 samples/sec	accuracy=0.804697	lr=0.004030
Epoch[112] Batch [849]	Speed: 5217.101709 samples/sec	accuracy=0.804691	lr=0.003987
Epoch[112] Batch [899]	Speed: 5007.593524 samples/sec	accuracy=0.804673	lr=0.003944
Epoch[112] Batch [949]	Speed: 5062.816052 samples/sec	accuracy=0.804661	lr=0.003901
Epoch[112] Batch [999]	Speed: 5228.471451 samples/sec	accuracy=0.804638	lr=0.003858
Epoch[112] Batch [1049]	Speed: 5392.568946 samples/sec	accuracy=0.804622	lr=0.003815
Epoch[112] Batch [1099]	Speed: 5088.283150 samples/sec	accuracy=0.804625	lr=0.003773
Epoch[112] Batch [1149]	Speed: 4367.929452 samples/sec	accuracy=0.804664	lr=0.003731
Epoch[112] Batch [1199]	Speed: 5051.197655 samples/sec	accuracy=0.804702	lr=0.003689
Epoch[112] Batch [1249]	Speed: 5177.130194 samples/sec	accuracy=0.804724	lr=0.003647
[Epoch 112] training: accuracy=0.804725
[Epoch 112] speed: 5092 samples/sec	time cost: 260.791107
[Epoch 112] validation: err-top1=0.227061 err-top5=0.065111
Epoch[113] Batch [49]	Speed: 4970.919605 samples/sec	accuracy=0.805042	lr=0.003605
Epoch[113] Batch [99]	Speed: 5070.563055 samples/sec	accuracy=0.805014	lr=0.003564
Epoch[113] Batch [149]	Speed: 4992.904089 samples/sec	accuracy=0.804932	lr=0.003523
Epoch[113] Batch [199]	Speed: 4884.117262 samples/sec	accuracy=0.804924	lr=0.003482
Epoch[113] Batch [249]	Speed: 5209.690102 samples/sec	accuracy=0.804906	lr=0.003442
Epoch[113] Batch [299]	Speed: 5219.192308 samples/sec	accuracy=0.804918	lr=0.003402
Epoch[113] Batch [349]	Speed: 5006.447346 samples/sec	accuracy=0.804915	lr=0.003362
Epoch[113] Batch [399]	Speed: 5215.947957 samples/sec	accuracy=0.804910	lr=0.003322
Epoch[113] Batch [449]	Speed: 5299.660113 samples/sec	accuracy=0.804946	lr=0.003282
Epoch[113] Batch [499]	Speed: 4958.890652 samples/sec	accuracy=0.804919	lr=0.003243
Epoch[113] Batch [549]	Speed: 4957.417589 samples/sec	accuracy=0.804870	lr=0.003204
Epoch[113] Batch [599]	Speed: 5295.026575 samples/sec	accuracy=0.804897	lr=0.003165
Epoch[113] Batch [649]	Speed: 4999.157872 samples/sec	accuracy=0.804938	lr=0.003127
Epoch[113] Batch [699]	Speed: 5154.730460 samples/sec	accuracy=0.804944	lr=0.003088
Epoch[113] Batch [749]	Speed: 5200.655905 samples/sec	accuracy=0.804938	lr=0.003050
Epoch[113] Batch [799]	Speed: 5352.869177 samples/sec	accuracy=0.804972	lr=0.003012
Epoch[113] Batch [849]	Speed: 5350.184241 samples/sec	accuracy=0.804976	lr=0.002975
Epoch[113] Batch [899]	Speed: 4989.038199 samples/sec	accuracy=0.804995	lr=0.002937
Epoch[113] Batch [949]	Speed: 5154.089113 samples/sec	accuracy=0.805009	lr=0.002900
Epoch[113] Batch [999]	Speed: 5208.879093 samples/sec	accuracy=0.804999	lr=0.002863
Epoch[113] Batch [1049]	Speed: 5246.749350 samples/sec	accuracy=0.805019	lr=0.002826
Epoch[113] Batch [1099]	Speed: 5153.941294 samples/sec	accuracy=0.805033	lr=0.002790
Epoch[113] Batch [1149]	Speed: 5261.821788 samples/sec	accuracy=0.805037	lr=0.002754
Epoch[113] Batch [1199]	Speed: 5147.183849 samples/sec	accuracy=0.805028	lr=0.002718
Epoch[113] Batch [1249]	Speed: 5264.212011 samples/sec	accuracy=0.805041	lr=0.002682
[Epoch 113] training: accuracy=0.805041
[Epoch 113] speed: 5133 samples/sec	time cost: 258.842062
[Epoch 113] validation: err-top1=0.225088 err-top5=0.063556
Epoch[114] Batch [49]	Speed: 5062.663874 samples/sec	accuracy=0.805521	lr=0.002646
Epoch[114] Batch [99]	Speed: 5073.097688 samples/sec	accuracy=0.805465	lr=0.002611
Epoch[114] Batch [149]	Speed: 5147.842606 samples/sec	accuracy=0.805495	lr=0.002575
Epoch[114] Batch [199]	Speed: 4932.927973 samples/sec	accuracy=0.805450	lr=0.002541
Epoch[114] Batch [249]	Speed: 5005.632219 samples/sec	accuracy=0.805321	lr=0.002506
Epoch[114] Batch [299]	Speed: 5227.283654 samples/sec	accuracy=0.805212	lr=0.002472
Epoch[114] Batch [349]	Speed: 5240.701194 samples/sec	accuracy=0.805288	lr=0.002438
Epoch[114] Batch [399]	Speed: 5176.317311 samples/sec	accuracy=0.805326	lr=0.002404
Epoch[114] Batch [449]	Speed: 5126.373035 samples/sec	accuracy=0.805333	lr=0.002370
Epoch[114] Batch [499]	Speed: 5235.382002 samples/sec	accuracy=0.805327	lr=0.002337
Epoch[114] Batch [549]	Speed: 5169.100803 samples/sec	accuracy=0.805299	lr=0.002304
Epoch[114] Batch [599]	Speed: 5291.943265 samples/sec	accuracy=0.805304	lr=0.002271
Epoch[114] Batch [649]	Speed: 5206.253574 samples/sec	accuracy=0.805325	lr=0.002238
Epoch[114] Batch [699]	Speed: 5047.172794 samples/sec	accuracy=0.805352	lr=0.002205
Epoch[114] Batch [749]	Speed: 5170.041362 samples/sec	accuracy=0.805339	lr=0.002173
Epoch[114] Batch [799]	Speed: 5086.262359 samples/sec	accuracy=0.805332	lr=0.002141
Epoch[114] Batch [849]	Speed: 5028.219468 samples/sec	accuracy=0.805332	lr=0.002110
Epoch[114] Batch [899]	Speed: 5257.797975 samples/sec	accuracy=0.805349	lr=0.002078
Epoch[114] Batch [949]	Speed: 5316.045582 samples/sec	accuracy=0.805368	lr=0.002047
Epoch[114] Batch [999]	Speed: 5171.579873 samples/sec	accuracy=0.805383	lr=0.002016
Epoch[114] Batch [1049]	Speed: 5118.391253 samples/sec	accuracy=0.805416	lr=0.001985
Epoch[114] Batch [1099]	Speed: 5107.550851 samples/sec	accuracy=0.805437	lr=0.001954
Epoch[114] Batch [1149]	Speed: 5218.311385 samples/sec	accuracy=0.805452	lr=0.001924
Epoch[114] Batch [1199]	Speed: 5258.971477 samples/sec	accuracy=0.805462	lr=0.001894
Epoch[114] Batch [1249]	Speed: 5261.309870 samples/sec	accuracy=0.805488	lr=0.001864
[Epoch 114] training: accuracy=0.805487
[Epoch 114] speed: 5152 samples/sec	time cost: 257.953064
[Epoch 114] validation: err-top1=0.225347 err-top5=0.062520
Epoch[115] Batch [49]	Speed: 5000.836579 samples/sec	accuracy=0.805976	lr=0.001834
Epoch[115] Batch [99]	Speed: 5069.741639 samples/sec	accuracy=0.805773	lr=0.001804
Epoch[115] Batch [149]	Speed: 5018.763067 samples/sec	accuracy=0.805818	lr=0.001775
Epoch[115] Batch [199]	Speed: 5144.287150 samples/sec	accuracy=0.805786	lr=0.001746
Epoch[115] Batch [249]	Speed: 4987.470955 samples/sec	accuracy=0.805727	lr=0.001718
Epoch[115] Batch [299]	Speed: 5225.026762 samples/sec	accuracy=0.805677	lr=0.001689
Epoch[115] Batch [349]	Speed: 5036.808616 samples/sec	accuracy=0.805656	lr=0.001661
Epoch[115] Batch [399]	Speed: 5048.008746 samples/sec	accuracy=0.805656	lr=0.001633
Epoch[115] Batch [449]	Speed: 4846.760511 samples/sec	accuracy=0.805658	lr=0.001605
Epoch[115] Batch [499]	Speed: 5194.631311 samples/sec	accuracy=0.805707	lr=0.001578
Epoch[115] Batch [549]	Speed: 5185.434977 samples/sec	accuracy=0.805717	lr=0.001551
Epoch[115] Batch [599]	Speed: 5172.679690 samples/sec	accuracy=0.805704	lr=0.001524
Epoch[115] Batch [649]	Speed: 5260.637348 samples/sec	accuracy=0.805686	lr=0.001497
Epoch[115] Batch [699]	Speed: 5340.161020 samples/sec	accuracy=0.805695	lr=0.001470
Epoch[115] Batch [749]	Speed: 5244.290171 samples/sec	accuracy=0.805681	lr=0.001444
Epoch[115] Batch [799]	Speed: 5084.878210 samples/sec	accuracy=0.805692	lr=0.001418
Epoch[115] Batch [849]	Speed: 5277.933156 samples/sec	accuracy=0.805701	lr=0.001392
Epoch[115] Batch [899]	Speed: 5195.715059 samples/sec	accuracy=0.805703	lr=0.001366
Epoch[115] Batch [949]	Speed: 5242.981506 samples/sec	accuracy=0.805709	lr=0.001341
Epoch[115] Batch [999]	Speed: 5080.735088 samples/sec	accuracy=0.805720	lr=0.001316
Epoch[115] Batch [1049]	Speed: 5306.970872 samples/sec	accuracy=0.805742	lr=0.001291
Epoch[115] Batch [1099]	Speed: 5239.499909 samples/sec	accuracy=0.805760	lr=0.001266
Epoch[115] Batch [1149]	Speed: 5108.098652 samples/sec	accuracy=0.805765	lr=0.001242
Epoch[115] Batch [1199]	Speed: 4961.190713 samples/sec	accuracy=0.805774	lr=0.001218
Epoch[115] Batch [1249]	Speed: 5230.685963 samples/sec	accuracy=0.805784	lr=0.001194
[Epoch 115] training: accuracy=0.805785
[Epoch 115] speed: 5132 samples/sec	time cost: 258.914405
[Epoch 115] validation: err-top1=0.224011 err-top5=0.062241
Epoch[116] Batch [49]	Speed: 4957.592690 samples/sec	accuracy=0.805961	lr=0.001170
Epoch[116] Batch [99]	Speed: 4876.681353 samples/sec	accuracy=0.805973	lr=0.001146
Epoch[116] Batch [149]	Speed: 4894.786918 samples/sec	accuracy=0.805937	lr=0.001123
Epoch[116] Batch [199]	Speed: 5062.422080 samples/sec	accuracy=0.805888	lr=0.001100
Epoch[116] Batch [249]	Speed: 5141.484675 samples/sec	accuracy=0.805833	lr=0.001077
Epoch[116] Batch [299]	Speed: 5218.130697 samples/sec	accuracy=0.805734	lr=0.001055
Epoch[116] Batch [349]	Speed: 5013.281850 samples/sec	accuracy=0.805675	lr=0.001032
Epoch[116] Batch [399]	Speed: 5254.465481 samples/sec	accuracy=0.805617	lr=0.001010
Epoch[116] Batch [449]	Speed: 5178.383711 samples/sec	accuracy=0.805615	lr=0.000989
Epoch[116] Batch [499]	Speed: 5249.585525 samples/sec	accuracy=0.805646	lr=0.000967
Epoch[116] Batch [549]	Speed: 5293.684770 samples/sec	accuracy=0.805690	lr=0.000946
Epoch[116] Batch [599]	Speed: 5191.142584 samples/sec	accuracy=0.805710	lr=0.000925
Epoch[116] Batch [649]	Speed: 5351.113988 samples/sec	accuracy=0.805704	lr=0.000904
Epoch[116] Batch [699]	Speed: 5159.926001 samples/sec	accuracy=0.805708	lr=0.000883
Epoch[116] Batch [749]	Speed: 5035.001671 samples/sec	accuracy=0.805727	lr=0.000863
Epoch[116] Batch [799]	Speed: 5009.829367 samples/sec	accuracy=0.805727	lr=0.000843
Epoch[116] Batch [849]	Speed: 5184.946076 samples/sec	accuracy=0.805728	lr=0.000823
Epoch[116] Batch [899]	Speed: 5096.213924 samples/sec	accuracy=0.805740	lr=0.000803
Epoch[116] Batch [949]	Speed: 5231.995635 samples/sec	accuracy=0.805761	lr=0.000784
Epoch[116] Batch [999]	Speed: 5134.557018 samples/sec	accuracy=0.805768	lr=0.000764
Epoch[116] Batch [1049]	Speed: 5234.317495 samples/sec	accuracy=0.805787	lr=0.000745
Epoch[116] Batch [1099]	Speed: 5285.809393 samples/sec	accuracy=0.805801	lr=0.000727
Epoch[116] Batch [1149]	Speed: 5393.932361 samples/sec	accuracy=0.805827	lr=0.000708
Epoch[116] Batch [1199]	Speed: 5169.712538 samples/sec	accuracy=0.805837	lr=0.000690
Epoch[116] Batch [1249]	Speed: 5170.647967 samples/sec	accuracy=0.805857	lr=0.000672
[Epoch 116] training: accuracy=0.805857
[Epoch 116] speed: 5143 samples/sec	time cost: 258.361198
[Epoch 116] validation: err-top1=0.223755 err-top5=0.062642
Epoch[117] Batch [49]	Speed: 4871.987576 samples/sec	accuracy=0.806278	lr=0.000654
Epoch[117] Batch [99]	Speed: 4950.325207 samples/sec	accuracy=0.806501	lr=0.000636
Epoch[117] Batch [149]	Speed: 5247.411273 samples/sec	accuracy=0.806447	lr=0.000619
Epoch[117] Batch [199]	Speed: 4920.680880 samples/sec	accuracy=0.806367	lr=0.000602
Epoch[117] Batch [249]	Speed: 5106.379228 samples/sec	accuracy=0.806402	lr=0.000585
Epoch[117] Batch [299]	Speed: 5128.769144 samples/sec	accuracy=0.806327	lr=0.000569
Epoch[117] Batch [349]	Speed: 5103.284244 samples/sec	accuracy=0.806335	lr=0.000552
Epoch[117] Batch [399]	Speed: 5081.089958 samples/sec	accuracy=0.806268	lr=0.000536
Epoch[117] Batch [449]	Speed: 4938.303501 samples/sec	accuracy=0.806200	lr=0.000520
Epoch[117] Batch [499]	Speed: 5086.995144 samples/sec	accuracy=0.806181	lr=0.000505
Epoch[117] Batch [549]	Speed: 5103.924776 samples/sec	accuracy=0.806191	lr=0.000489
Epoch[117] Batch [599]	Speed: 5243.459392 samples/sec	accuracy=0.806217	lr=0.000474
Epoch[117] Batch [649]	Speed: 5241.691150 samples/sec	accuracy=0.806198	lr=0.000459
Epoch[117] Batch [699]	Speed: 5000.526015 samples/sec	accuracy=0.806199	lr=0.000445
Epoch[117] Batch [749]	Speed: 5121.273144 samples/sec	accuracy=0.806194	lr=0.000430
Epoch[117] Batch [799]	Speed: 5222.605936 samples/sec	accuracy=0.806209	lr=0.000416
Epoch[117] Batch [849]	Speed: 5433.583526 samples/sec	accuracy=0.806224	lr=0.000402
Epoch[117] Batch [899]	Speed: 4818.634847 samples/sec	accuracy=0.806217	lr=0.000388
Epoch[117] Batch [949]	Speed: 5034.729814 samples/sec	accuracy=0.806221	lr=0.000375
Epoch[117] Batch [999]	Speed: 5183.828524 samples/sec	accuracy=0.806222	lr=0.000362
Epoch[117] Batch [1049]	Speed: 5181.513349 samples/sec	accuracy=0.806215	lr=0.000349
Epoch[117] Batch [1099]	Speed: 5414.252185 samples/sec	accuracy=0.806224	lr=0.000336
Epoch[117] Batch [1149]	Speed: 5225.813305 samples/sec	accuracy=0.806232	lr=0.000323
Epoch[117] Batch [1199]	Speed: 4972.982312 samples/sec	accuracy=0.806221	lr=0.000311
Epoch[117] Batch [1249]	Speed: 5274.081678 samples/sec	accuracy=0.806236	lr=0.000299
[Epoch 117] training: accuracy=0.806237
[Epoch 117] speed: 5107 samples/sec	time cost: 260.423261
[Epoch 117] validation: err-top1=0.223633 err-top5=0.062101
Epoch[118] Batch [49]	Speed: 4891.944379 samples/sec	accuracy=0.806484	lr=0.000287
Epoch[118] Batch [99]	Speed: 4914.188346 samples/sec	accuracy=0.806249	lr=0.000275
Epoch[118] Batch [149]	Speed: 5043.638309 samples/sec	accuracy=0.806235	lr=0.000264
Epoch[118] Batch [199]	Speed: 5205.580794 samples/sec	accuracy=0.806288	lr=0.000253
Epoch[118] Batch [249]	Speed: 5134.929392 samples/sec	accuracy=0.806278	lr=0.000242
Epoch[118] Batch [299]	Speed: 5430.306718 samples/sec	accuracy=0.806326	lr=0.000231
Epoch[118] Batch [349]	Speed: 5036.652918 samples/sec	accuracy=0.806324	lr=0.000221
Epoch[118] Batch [399]	Speed: 5289.797768 samples/sec	accuracy=0.806322	lr=0.000211
Epoch[118] Batch [449]	Speed: 5421.490721 samples/sec	accuracy=0.806315	lr=0.000201
Epoch[118] Batch [499]	Speed: 5019.126577 samples/sec	accuracy=0.806353	lr=0.000191
Epoch[118] Batch [549]	Speed: 4970.377707 samples/sec	accuracy=0.806359	lr=0.000182
Epoch[118] Batch [599]	Speed: 5018.590421 samples/sec	accuracy=0.806365	lr=0.000173
Epoch[118] Batch [649]	Speed: 4891.095703 samples/sec	accuracy=0.806369	lr=0.000164
Epoch[118] Batch [699]	Speed: 4964.745526 samples/sec	accuracy=0.806380	lr=0.000155
Epoch[118] Batch [749]	Speed: 5178.363233 samples/sec	accuracy=0.806369	lr=0.000147
Epoch[118] Batch [799]	Speed: 5008.927967 samples/sec	accuracy=0.806373	lr=0.000138
Epoch[118] Batch [849]	Speed: 5001.936266 samples/sec	accuracy=0.806361	lr=0.000130
Epoch[118] Batch [899]	Speed: 5216.077436 samples/sec	accuracy=0.806346	lr=0.000123
Epoch[118] Batch [949]	Speed: 5314.884093 samples/sec	accuracy=0.806316	lr=0.000115
Epoch[118] Batch [999]	Speed: 5458.749279 samples/sec	accuracy=0.806319	lr=0.000108
Epoch[118] Batch [1049]	Speed: 5246.846262 samples/sec	accuracy=0.806310	lr=0.000101
Epoch[118] Batch [1099]	Speed: 5084.431320 samples/sec	accuracy=0.806310	lr=0.000094
Epoch[118] Batch [1149]	Speed: 5126.733453 samples/sec	accuracy=0.806319	lr=0.000087
Epoch[118] Batch [1199]	Speed: 5226.430016 samples/sec	accuracy=0.806317	lr=0.000081
Epoch[118] Batch [1249]	Speed: 5266.252854 samples/sec	accuracy=0.806311	lr=0.000075
[Epoch 118] training: accuracy=0.806310
[Epoch 118] speed: 5124 samples/sec	time cost: 258.986983
[Epoch 118] validation: err-top1=0.222935 err-top5=0.061802
Epoch[119] Batch [49]	Speed: 4930.779815 samples/sec	accuracy=0.806156	lr=0.000069
Epoch[119] Batch [99]	Speed: 5121.308562 samples/sec	accuracy=0.806226	lr=0.000063
Epoch[119] Batch [149]	Speed: 5196.187511 samples/sec	accuracy=0.806283	lr=0.000058
Epoch[119] Batch [199]	Speed: 5185.595126 samples/sec	accuracy=0.806316	lr=0.000053
Epoch[119] Batch [249]	Speed: 5176.532923 samples/sec	accuracy=0.806309	lr=0.000048
Epoch[119] Batch [299]	Speed: 4959.471280 samples/sec	accuracy=0.806320	lr=0.000043
Epoch[119] Batch [349]	Speed: 5051.911457 samples/sec	accuracy=0.806313	lr=0.000039
Epoch[119] Batch [399]	Speed: 5317.481698 samples/sec	accuracy=0.806292	lr=0.000035
Epoch[119] Batch [449]	Speed: 5250.994042 samples/sec	accuracy=0.806268	lr=0.000031
Epoch[119] Batch [499]	Speed: 5221.746714 samples/sec	accuracy=0.806254	lr=0.000027
Epoch[119] Batch [549]	Speed: 4972.964808 samples/sec	accuracy=0.806283	lr=0.000023
Epoch[119] Batch [599]	Speed: 5126.214565 samples/sec	accuracy=0.806302	lr=0.000020
Epoch[119] Batch [649]	Speed: 5101.303751 samples/sec	accuracy=0.806306	lr=0.000017
Epoch[119] Batch [699]	Speed: 5410.169549 samples/sec	accuracy=0.806319	lr=0.000015
Epoch[119] Batch [749]	Speed: 5227.760846 samples/sec	accuracy=0.806321	lr=0.000012
Epoch[119] Batch [799]	Speed: 5297.306723 samples/sec	accuracy=0.806317	lr=0.000010
Epoch[119] Batch [849]	Speed: 5078.825505 samples/sec	accuracy=0.806304	lr=0.000008
Epoch[119] Batch [899]	Speed: 5227.926420 samples/sec	accuracy=0.806304	lr=0.000006
Epoch[119] Batch [949]	Speed: 5249.263699 samples/sec	accuracy=0.806303	lr=0.000004
Epoch[119] Batch [999]	Speed: 5175.028129 samples/sec	accuracy=0.806308	lr=0.000003
Epoch[119] Batch [1049]	Speed: 5179.348013 samples/sec	accuracy=0.806306	lr=0.000002
Epoch[119] Batch [1099]	Speed: 5067.690819 samples/sec	accuracy=0.806297	lr=0.000001
Epoch[119] Batch [1149]	Speed: 5190.806679 samples/sec	accuracy=0.806282	lr=0.000000
Epoch[119] Batch [1199]	Speed: 5008.581586 samples/sec	accuracy=0.806286	lr=0.000000
Epoch[119] Batch [1249]	Speed: 5326.299790 samples/sec	accuracy=0.806293	lr=0.000000
[Epoch 119] training: accuracy=0.806293
[Epoch 119] speed: 5154 samples/sec	time cost: 257.641256
[Epoch 119] validation: err-top1=0.223374 err-top5=0.062161