Namespace(batch_norm=False, batch_size=128, data_dir='~/.mxnet/datasets/imagenet', dtype='float16', input_size=224, label_smoothing=True, last_gamma=False, log_interval=50, logging_file='mobilenet0.75.log', lr=0.4, lr_decay=0.1, lr_decay_epoch='40,60', lr_decay_period=0, lr_mode='cosine', mixup=False, mixup_alpha=0.2, mixup_off_epoch=0, mode='hybrid', model='mobilenet0.75', momentum=0.9, no_wd=True, num_epochs=120, num_gpus=8, num_workers=60, rec_train='/media/ramdisk/rec/train.rec', rec_train_idx='/media/ramdisk/rec/train.idx', rec_val='/media/ramdisk/rec/val.rec', rec_val_idx='/media/ramdisk/rec/val.idx', resume_epoch=0, resume_params='', resume_states='', save_dir='params_mobilenet0.75', 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: 2419.912456 samples/sec	accuracy=0.001055	lr=0.003133
Epoch[0] Batch [99]	Speed: 5859.733870 samples/sec	accuracy=0.001133	lr=0.006331
Epoch[0] Batch [149]	Speed: 5869.096875 samples/sec	accuracy=0.001204	lr=0.009528
Epoch[0] Batch [199]	Speed: 5786.341812 samples/sec	accuracy=0.001230	lr=0.012726
Epoch[0] Batch [249]	Speed: 5875.977401 samples/sec	accuracy=0.001383	lr=0.015923
Epoch[0] Batch [299]	Speed: 5524.141759 samples/sec	accuracy=0.001507	lr=0.019121
Epoch[0] Batch [349]	Speed: 5602.793262 samples/sec	accuracy=0.001710	lr=0.022318
Epoch[0] Batch [399]	Speed: 4721.090051 samples/sec	accuracy=0.001863	lr=0.025516
Epoch[0] Batch [449]	Speed: 5349.465355 samples/sec	accuracy=0.002025	lr=0.028713
Epoch[0] Batch [499]	Speed: 5500.783481 samples/sec	accuracy=0.002258	lr=0.031910
Epoch[0] Batch [549]	Speed: 5273.069352 samples/sec	accuracy=0.002525	lr=0.035108
Epoch[0] Batch [599]	Speed: 5295.388639 samples/sec	accuracy=0.002829	lr=0.038305
Epoch[0] Batch [649]	Speed: 5393.717089 samples/sec	accuracy=0.003139	lr=0.041503
Epoch[0] Batch [699]	Speed: 5632.461670 samples/sec	accuracy=0.003520	lr=0.044700
Epoch[0] Batch [749]	Speed: 5399.691062 samples/sec	accuracy=0.004000	lr=0.047898
Epoch[0] Batch [799]	Speed: 5459.206107 samples/sec	accuracy=0.004481	lr=0.051095
Epoch[0] Batch [849]	Speed: 5187.542874 samples/sec	accuracy=0.005121	lr=0.054293
Epoch[0] Batch [899]	Speed: 5172.728781 samples/sec	accuracy=0.005818	lr=0.057490
Epoch[0] Batch [949]	Speed: 5072.834285 samples/sec	accuracy=0.006552	lr=0.060687
Epoch[0] Batch [999]	Speed: 5619.554807 samples/sec	accuracy=0.007434	lr=0.063885
Epoch[0] Batch [1049]	Speed: 5438.363407 samples/sec	accuracy=0.008306	lr=0.067082
Epoch[0] Batch [1099]	Speed: 5676.278707 samples/sec	accuracy=0.009244	lr=0.070280
Epoch[0] Batch [1149]	Speed: 5493.525733 samples/sec	accuracy=0.010299	lr=0.073477
Epoch[0] Batch [1199]	Speed: 5007.402613 samples/sec	accuracy=0.011395	lr=0.076675
Epoch[0] Batch [1249]	Speed: 5669.606625 samples/sec	accuracy=0.012546	lr=0.079872
[Epoch 0] training: accuracy=0.012587
[Epoch 0] speed: 5173 samples/sec	time cost: 257.774637
[Epoch 0] validation: err-top1=0.954859 err-top5=0.859754
Epoch[1] Batch [49]	Speed: 4855.940718 samples/sec	accuracy=0.044375	lr=0.083133
Epoch[1] Batch [99]	Speed: 5678.365585 samples/sec	accuracy=0.046084	lr=0.086331
Epoch[1] Batch [149]	Speed: 5457.238212 samples/sec	accuracy=0.047819	lr=0.089528
Epoch[1] Batch [199]	Speed: 5321.613110 samples/sec	accuracy=0.049863	lr=0.092726
Epoch[1] Batch [249]	Speed: 5387.971928 samples/sec	accuracy=0.051898	lr=0.095923
Epoch[1] Batch [299]	Speed: 5217.547761 samples/sec	accuracy=0.054092	lr=0.099121
Epoch[1] Batch [349]	Speed: 5244.020600 samples/sec	accuracy=0.056345	lr=0.102318
Epoch[1] Batch [399]	Speed: 5627.478151 samples/sec	accuracy=0.058826	lr=0.105516
Epoch[1] Batch [449]	Speed: 5972.408370 samples/sec	accuracy=0.061404	lr=0.108713
Epoch[1] Batch [499]	Speed: 5604.885838 samples/sec	accuracy=0.063818	lr=0.111910
Epoch[1] Batch [549]	Speed: 5582.333873 samples/sec	accuracy=0.066467	lr=0.115108
Epoch[1] Batch [599]	Speed: 5249.941915 samples/sec	accuracy=0.069097	lr=0.118305
Epoch[1] Batch [649]	Speed: 5621.862428 samples/sec	accuracy=0.071731	lr=0.121503
Epoch[1] Batch [699]	Speed: 5279.626899 samples/sec	accuracy=0.074443	lr=0.124700
Epoch[1] Batch [749]	Speed: 5698.183957 samples/sec	accuracy=0.077150	lr=0.127898
Epoch[1] Batch [799]	Speed: 5497.014383 samples/sec	accuracy=0.079979	lr=0.131095
Epoch[1] Batch [849]	Speed: 5666.335518 samples/sec	accuracy=0.082796	lr=0.134293
Epoch[1] Batch [899]	Speed: 5873.887691 samples/sec	accuracy=0.085579	lr=0.137490
Epoch[1] Batch [949]	Speed: 5605.402569 samples/sec	accuracy=0.088538	lr=0.140687
Epoch[1] Batch [999]	Speed: 5466.096373 samples/sec	accuracy=0.091387	lr=0.143885
Epoch[1] Batch [1049]	Speed: 5321.616275 samples/sec	accuracy=0.094156	lr=0.147082
Epoch[1] Batch [1099]	Speed: 5854.315039 samples/sec	accuracy=0.096917	lr=0.150280
Epoch[1] Batch [1149]	Speed: 5654.003046 samples/sec	accuracy=0.099698	lr=0.153477
Epoch[1] Batch [1199]	Speed: 5735.838642 samples/sec	accuracy=0.102528	lr=0.156675
Epoch[1] Batch [1249]	Speed: 5994.876018 samples/sec	accuracy=0.105295	lr=0.159872
[Epoch 1] training: accuracy=0.105353
[Epoch 1] speed: 5516 samples/sec	time cost: 242.077666
[Epoch 1] validation: err-top1=0.818040 err-top5=0.608518
Epoch[2] Batch [49]	Speed: 5042.573849 samples/sec	accuracy=0.177734	lr=0.163133
Epoch[2] Batch [99]	Speed: 5246.205372 samples/sec	accuracy=0.180869	lr=0.166331
Epoch[2] Batch [149]	Speed: 5587.551875 samples/sec	accuracy=0.183164	lr=0.169528
Epoch[2] Batch [199]	Speed: 5464.439404 samples/sec	accuracy=0.185391	lr=0.172726
Epoch[2] Batch [249]	Speed: 5702.040425 samples/sec	accuracy=0.187289	lr=0.175923
Epoch[2] Batch [299]	Speed: 5600.413633 samples/sec	accuracy=0.189062	lr=0.179121
Epoch[2] Batch [349]	Speed: 5566.633848 samples/sec	accuracy=0.191682	lr=0.182318
Epoch[2] Batch [399]	Speed: 5835.599061 samples/sec	accuracy=0.193601	lr=0.185516
Epoch[2] Batch [449]	Speed: 5379.199905 samples/sec	accuracy=0.196068	lr=0.188713
Epoch[2] Batch [499]	Speed: 5820.639111 samples/sec	accuracy=0.198049	lr=0.191910
Epoch[2] Batch [549]	Speed: 5795.970848 samples/sec	accuracy=0.200066	lr=0.195108
Epoch[2] Batch [599]	Speed: 5370.551891 samples/sec	accuracy=0.201961	lr=0.198305
Epoch[2] Batch [649]	Speed: 5721.667030 samples/sec	accuracy=0.204124	lr=0.201503
Epoch[2] Batch [699]	Speed: 5534.990902 samples/sec	accuracy=0.205914	lr=0.204700
Epoch[2] Batch [749]	Speed: 5459.127558 samples/sec	accuracy=0.207742	lr=0.207898
Epoch[2] Batch [799]	Speed: 5973.809751 samples/sec	accuracy=0.209550	lr=0.211095
Epoch[2] Batch [849]	Speed: 5405.189147 samples/sec	accuracy=0.211521	lr=0.214293
Epoch[2] Batch [899]	Speed: 5554.570751 samples/sec	accuracy=0.213414	lr=0.217490
Epoch[2] Batch [949]	Speed: 5838.235489 samples/sec	accuracy=0.215113	lr=0.220687
Epoch[2] Batch [999]	Speed: 5648.398021 samples/sec	accuracy=0.216996	lr=0.223885
Epoch[2] Batch [1049]	Speed: 5805.775286 samples/sec	accuracy=0.218683	lr=0.227082
Epoch[2] Batch [1099]	Speed: 5556.839247 samples/sec	accuracy=0.220303	lr=0.230280
Epoch[2] Batch [1149]	Speed: 5507.095971 samples/sec	accuracy=0.221985	lr=0.233477
Epoch[2] Batch [1199]	Speed: 5661.172510 samples/sec	accuracy=0.223752	lr=0.236675
Epoch[2] Batch [1249]	Speed: 5778.529356 samples/sec	accuracy=0.225434	lr=0.239872
[Epoch 2] training: accuracy=0.225469
[Epoch 2] speed: 5577 samples/sec	time cost: 239.610919
[Epoch 2] validation: err-top1=0.729951 err-top5=0.478536
Epoch[3] Batch [49]	Speed: 5331.967598 samples/sec	accuracy=0.275137	lr=0.243133
Epoch[3] Batch [99]	Speed: 5302.806373 samples/sec	accuracy=0.274561	lr=0.246331
Epoch[3] Batch [149]	Speed: 5749.988254 samples/sec	accuracy=0.275293	lr=0.249528
Epoch[3] Batch [199]	Speed: 5758.377056 samples/sec	accuracy=0.276084	lr=0.252726
Epoch[3] Batch [249]	Speed: 5600.137899 samples/sec	accuracy=0.276949	lr=0.255923
Epoch[3] Batch [299]	Speed: 5545.661616 samples/sec	accuracy=0.278138	lr=0.259121
Epoch[3] Batch [349]	Speed: 5523.591169 samples/sec	accuracy=0.278968	lr=0.262318
Epoch[3] Batch [399]	Speed: 5822.373942 samples/sec	accuracy=0.279846	lr=0.265516
Epoch[3] Batch [449]	Speed: 5868.558130 samples/sec	accuracy=0.280588	lr=0.268713
Epoch[3] Batch [499]	Speed: 5238.241803 samples/sec	accuracy=0.281836	lr=0.271910
Epoch[3] Batch [549]	Speed: 5757.743280 samples/sec	accuracy=0.282441	lr=0.275108
Epoch[3] Batch [599]	Speed: 5714.119197 samples/sec	accuracy=0.283009	lr=0.278305
Epoch[3] Batch [649]	Speed: 5536.324386 samples/sec	accuracy=0.284088	lr=0.281503
Epoch[3] Batch [699]	Speed: 5775.031221 samples/sec	accuracy=0.285266	lr=0.284700
Epoch[3] Batch [749]	Speed: 5694.566449 samples/sec	accuracy=0.286159	lr=0.287898
Epoch[3] Batch [799]	Speed: 5962.362472 samples/sec	accuracy=0.287057	lr=0.291095
Epoch[3] Batch [849]	Speed: 5476.583158 samples/sec	accuracy=0.288101	lr=0.294293
Epoch[3] Batch [899]	Speed: 5191.011956 samples/sec	accuracy=0.289124	lr=0.297490
Epoch[3] Batch [949]	Speed: 5491.753224 samples/sec	accuracy=0.290138	lr=0.300687
Epoch[3] Batch [999]	Speed: 5864.134674 samples/sec	accuracy=0.291147	lr=0.303885
Epoch[3] Batch [1049]	Speed: 5680.866785 samples/sec	accuracy=0.292122	lr=0.307082
Epoch[3] Batch [1099]	Speed: 5618.912847 samples/sec	accuracy=0.292954	lr=0.310280
Epoch[3] Batch [1149]	Speed: 5302.498282 samples/sec	accuracy=0.293869	lr=0.313477
Epoch[3] Batch [1199]	Speed: 5186.705798 samples/sec	accuracy=0.294666	lr=0.316675
Epoch[3] Batch [1249]	Speed: 5568.700940 samples/sec	accuracy=0.295521	lr=0.319872
[Epoch 3] training: accuracy=0.295527
[Epoch 3] speed: 5565 samples/sec	time cost: 240.170652
[Epoch 3] validation: err-top1=0.664262 err-top5=0.406967
Epoch[4] Batch [49]	Speed: 5339.242110 samples/sec	accuracy=0.322324	lr=0.323133
Epoch[4] Batch [99]	Speed: 5401.132116 samples/sec	accuracy=0.323047	lr=0.326331
Epoch[4] Batch [149]	Speed: 5930.647783 samples/sec	accuracy=0.324818	lr=0.329528
Epoch[4] Batch [199]	Speed: 5703.473953 samples/sec	accuracy=0.324951	lr=0.332726
Epoch[4] Batch [249]	Speed: 5468.407624 samples/sec	accuracy=0.325773	lr=0.335923
Epoch[4] Batch [299]	Speed: 5571.470822 samples/sec	accuracy=0.325192	lr=0.339121
Epoch[4] Batch [349]	Speed: 5918.251304 samples/sec	accuracy=0.325977	lr=0.342318
Epoch[4] Batch [399]	Speed: 5475.856156 samples/sec	accuracy=0.326267	lr=0.345516
Epoch[4] Batch [449]	Speed: 5554.035625 samples/sec	accuracy=0.326905	lr=0.348713
Epoch[4] Batch [499]	Speed: 5656.788424 samples/sec	accuracy=0.327258	lr=0.351910
Epoch[4] Batch [549]	Speed: 5946.406693 samples/sec	accuracy=0.327752	lr=0.355108
Epoch[4] Batch [599]	Speed: 5682.633568 samples/sec	accuracy=0.328127	lr=0.358305
Epoch[4] Batch [649]	Speed: 5720.814832 samples/sec	accuracy=0.328801	lr=0.361503
Epoch[4] Batch [699]	Speed: 6059.999127 samples/sec	accuracy=0.329371	lr=0.364700
Epoch[4] Batch [749]	Speed: 5738.891894 samples/sec	accuracy=0.329858	lr=0.367898
Epoch[4] Batch [799]	Speed: 5521.287271 samples/sec	accuracy=0.329982	lr=0.371095
Epoch[4] Batch [849]	Speed: 5708.322029 samples/sec	accuracy=0.330326	lr=0.374293
Epoch[4] Batch [899]	Speed: 5731.521908 samples/sec	accuracy=0.330732	lr=0.377490
Epoch[4] Batch [949]	Speed: 5723.733721 samples/sec	accuracy=0.331349	lr=0.380687
Epoch[4] Batch [999]	Speed: 5447.375210 samples/sec	accuracy=0.331638	lr=0.383885
Epoch[4] Batch [1049]	Speed: 5488.767820 samples/sec	accuracy=0.332042	lr=0.387082
Epoch[4] Batch [1099]	Speed: 5736.694709 samples/sec	accuracy=0.332481	lr=0.390280
Epoch[4] Batch [1149]	Speed: 5489.526321 samples/sec	accuracy=0.332849	lr=0.393477
Epoch[4] Batch [1199]	Speed: 6033.276316 samples/sec	accuracy=0.333271	lr=0.396675
Epoch[4] Batch [1249]	Speed: 5551.929454 samples/sec	accuracy=0.333678	lr=0.399872
[Epoch 4] training: accuracy=0.333690
[Epoch 4] speed: 5650 samples/sec	time cost: 236.779573
[Epoch 4] validation: err-top1=0.656091 err-top5=0.395109
Epoch[5] Batch [49]	Speed: 5651.094451 samples/sec	accuracy=0.352207	lr=0.400000
Epoch[5] Batch [99]	Speed: 5661.850435 samples/sec	accuracy=0.352646	lr=0.400000
Epoch[5] Batch [149]	Speed: 5612.537828 samples/sec	accuracy=0.352917	lr=0.399999
Epoch[5] Batch [199]	Speed: 5876.349629 samples/sec	accuracy=0.353545	lr=0.399998
Epoch[5] Batch [249]	Speed: 5584.143419 samples/sec	accuracy=0.353602	lr=0.399997
Epoch[5] Batch [299]	Speed: 5676.261153 samples/sec	accuracy=0.354229	lr=0.399996
Epoch[5] Batch [349]	Speed: 5474.789881 samples/sec	accuracy=0.353792	lr=0.399994
Epoch[5] Batch [399]	Speed: 5400.852972 samples/sec	accuracy=0.354226	lr=0.399992
Epoch[5] Batch [449]	Speed: 5562.234412 samples/sec	accuracy=0.355167	lr=0.399990
Epoch[5] Batch [499]	Speed: 5831.834171 samples/sec	accuracy=0.355383	lr=0.399988
Epoch[5] Batch [549]	Speed: 5720.855066 samples/sec	accuracy=0.355762	lr=0.399986
Epoch[5] Batch [599]	Speed: 5674.493382 samples/sec	accuracy=0.355949	lr=0.399983
Epoch[5] Batch [649]	Speed: 5849.032072 samples/sec	accuracy=0.356472	lr=0.399980
Epoch[5] Batch [699]	Speed: 5978.116692 samples/sec	accuracy=0.357044	lr=0.399977
Epoch[5] Batch [749]	Speed: 5686.111563 samples/sec	accuracy=0.357471	lr=0.399973
Epoch[5] Batch [799]	Speed: 5534.806590 samples/sec	accuracy=0.357645	lr=0.399970
Epoch[5] Batch [849]	Speed: 5508.357410 samples/sec	accuracy=0.358196	lr=0.399966
Epoch[5] Batch [899]	Speed: 5844.846220 samples/sec	accuracy=0.358571	lr=0.399961
Epoch[5] Batch [949]	Speed: 5476.220471 samples/sec	accuracy=0.359011	lr=0.399957
Epoch[5] Batch [999]	Speed: 5866.606071 samples/sec	accuracy=0.359571	lr=0.399952
Epoch[5] Batch [1049]	Speed: 5782.606869 samples/sec	accuracy=0.360009	lr=0.399948
Epoch[5] Batch [1099]	Speed: 5472.380774 samples/sec	accuracy=0.360512	lr=0.399942
Epoch[5] Batch [1149]	Speed: 5553.065767 samples/sec	accuracy=0.360767	lr=0.399937
Epoch[5] Batch [1199]	Speed: 5605.808325 samples/sec	accuracy=0.361169	lr=0.399931
Epoch[5] Batch [1249]	Speed: 5380.848988 samples/sec	accuracy=0.361565	lr=0.399926
[Epoch 5] training: accuracy=0.361576
[Epoch 5] speed: 5638 samples/sec	time cost: 236.673982
[Epoch 5] validation: err-top1=0.635722 err-top5=0.375549
Epoch[6] Batch [49]	Speed: 5047.708194 samples/sec	accuracy=0.374648	lr=0.399919
Epoch[6] Batch [99]	Speed: 5425.324278 samples/sec	accuracy=0.377773	lr=0.399913
Epoch[6] Batch [149]	Speed: 5451.611642 samples/sec	accuracy=0.377344	lr=0.399907
Epoch[6] Batch [199]	Speed: 5558.418068 samples/sec	accuracy=0.377812	lr=0.399900
Epoch[6] Batch [249]	Speed: 5775.963343 samples/sec	accuracy=0.378078	lr=0.399893
Epoch[6] Batch [299]	Speed: 5808.774310 samples/sec	accuracy=0.378242	lr=0.399885
Epoch[6] Batch [349]	Speed: 5867.240636 samples/sec	accuracy=0.378078	lr=0.399878
Epoch[6] Batch [399]	Speed: 5986.415787 samples/sec	accuracy=0.378259	lr=0.399870
Epoch[6] Batch [449]	Speed: 5315.886090 samples/sec	accuracy=0.378112	lr=0.399862
Epoch[6] Batch [499]	Speed: 5755.613704 samples/sec	accuracy=0.378402	lr=0.399854
Epoch[6] Batch [549]	Speed: 5678.877633 samples/sec	accuracy=0.378388	lr=0.399846
Epoch[6] Batch [599]	Speed: 5868.973046 samples/sec	accuracy=0.378576	lr=0.399837
Epoch[6] Batch [649]	Speed: 5961.490527 samples/sec	accuracy=0.378935	lr=0.399828
Epoch[6] Batch [699]	Speed: 5610.352764 samples/sec	accuracy=0.379216	lr=0.399819
Epoch[6] Batch [749]	Speed: 5443.613514 samples/sec	accuracy=0.379551	lr=0.399809
Epoch[6] Batch [799]	Speed: 5764.942712 samples/sec	accuracy=0.379978	lr=0.399800
Epoch[6] Batch [849]	Speed: 5561.784235 samples/sec	accuracy=0.380306	lr=0.399790
Epoch[6] Batch [899]	Speed: 5615.637254 samples/sec	accuracy=0.380526	lr=0.399780
Epoch[6] Batch [949]	Speed: 5602.489815 samples/sec	accuracy=0.380992	lr=0.399769
Epoch[6] Batch [999]	Speed: 5693.081550 samples/sec	accuracy=0.381224	lr=0.399759
Epoch[6] Batch [1049]	Speed: 5231.110767 samples/sec	accuracy=0.381551	lr=0.399748
Epoch[6] Batch [1099]	Speed: 5979.799479 samples/sec	accuracy=0.381909	lr=0.399737
Epoch[6] Batch [1149]	Speed: 5731.763154 samples/sec	accuracy=0.382150	lr=0.399725
Epoch[6] Batch [1199]	Speed: 5705.847386 samples/sec	accuracy=0.382485	lr=0.399714
Epoch[6] Batch [1249]	Speed: 5687.555470 samples/sec	accuracy=0.382644	lr=0.399702
[Epoch 6] training: accuracy=0.382639
[Epoch 6] speed: 5630 samples/sec	time cost: 237.411625
[Epoch 6] validation: err-top1=0.628089 err-top5=0.370376
Epoch[7] Batch [49]	Speed: 5438.325808 samples/sec	accuracy=0.395957	lr=0.399690
Epoch[7] Batch [99]	Speed: 5768.660444 samples/sec	accuracy=0.396641	lr=0.399677
Epoch[7] Batch [149]	Speed: 5519.692016 samples/sec	accuracy=0.396113	lr=0.399665
Epoch[7] Batch [199]	Speed: 5779.247347 samples/sec	accuracy=0.395752	lr=0.399652
Epoch[7] Batch [249]	Speed: 5618.228996 samples/sec	accuracy=0.395734	lr=0.399639
Epoch[7] Batch [299]	Speed: 5565.811480 samples/sec	accuracy=0.396110	lr=0.399626
Epoch[7] Batch [349]	Speed: 5558.723954 samples/sec	accuracy=0.395904	lr=0.399613
Epoch[7] Batch [399]	Speed: 5869.085486 samples/sec	accuracy=0.395996	lr=0.399599
Epoch[7] Batch [449]	Speed: 5432.688810 samples/sec	accuracy=0.396380	lr=0.399585
Epoch[7] Batch [499]	Speed: 5837.143220 samples/sec	accuracy=0.396074	lr=0.399571
Epoch[7] Batch [549]	Speed: 5477.788319 samples/sec	accuracy=0.396225	lr=0.399556
Epoch[7] Batch [599]	Speed: 5368.818375 samples/sec	accuracy=0.396029	lr=0.399542
Epoch[7] Batch [649]	Speed: 5988.677529 samples/sec	accuracy=0.396164	lr=0.399527
Epoch[7] Batch [699]	Speed: 5891.830459 samples/sec	accuracy=0.396403	lr=0.399512
Epoch[7] Batch [749]	Speed: 5782.861778 samples/sec	accuracy=0.396521	lr=0.399496
Epoch[7] Batch [799]	Speed: 5801.325100 samples/sec	accuracy=0.396647	lr=0.399481
Epoch[7] Batch [849]	Speed: 5660.152345 samples/sec	accuracy=0.396708	lr=0.399465
Epoch[7] Batch [899]	Speed: 5762.163780 samples/sec	accuracy=0.396858	lr=0.399449
Epoch[7] Batch [949]	Speed: 5320.486492 samples/sec	accuracy=0.396947	lr=0.399432
Epoch[7] Batch [999]	Speed: 5652.632960 samples/sec	accuracy=0.397013	lr=0.399416
Epoch[7] Batch [1049]	Speed: 5431.407523 samples/sec	accuracy=0.397158	lr=0.399399
Epoch[7] Batch [1099]	Speed: 5687.266118 samples/sec	accuracy=0.397373	lr=0.399382
Epoch[7] Batch [1149]	Speed: 5639.731048 samples/sec	accuracy=0.397503	lr=0.399365
Epoch[7] Batch [1199]	Speed: 5660.416863 samples/sec	accuracy=0.397644	lr=0.399347
Epoch[7] Batch [1249]	Speed: 5115.417557 samples/sec	accuracy=0.397989	lr=0.399329
[Epoch 7] training: accuracy=0.398001
[Epoch 7] speed: 5612 samples/sec	time cost: 238.293579
[Epoch 7] validation: err-top1=0.586097 err-top5=0.323661
Epoch[8] Batch [49]	Speed: 5791.832297 samples/sec	accuracy=0.409434	lr=0.399311
Epoch[8] Batch [99]	Speed: 5752.712148 samples/sec	accuracy=0.409600	lr=0.399293
Epoch[8] Batch [149]	Speed: 5495.362386 samples/sec	accuracy=0.408073	lr=0.399274
Epoch[8] Batch [199]	Speed: 5957.870910 samples/sec	accuracy=0.408740	lr=0.399256
Epoch[8] Batch [249]	Speed: 5173.651341 samples/sec	accuracy=0.409316	lr=0.399237
Epoch[8] Batch [299]	Speed: 5662.393251 samples/sec	accuracy=0.409023	lr=0.399218
Epoch[8] Batch [349]	Speed: 6072.438903 samples/sec	accuracy=0.409132	lr=0.399198
Epoch[8] Batch [399]	Speed: 5511.533200 samples/sec	accuracy=0.408816	lr=0.399179
Epoch[8] Batch [449]	Speed: 5548.573481 samples/sec	accuracy=0.408492	lr=0.399159
Epoch[8] Batch [499]	Speed: 6046.362617 samples/sec	accuracy=0.408574	lr=0.399138
Epoch[8] Batch [549]	Speed: 5922.335563 samples/sec	accuracy=0.408839	lr=0.399118
Epoch[8] Batch [599]	Speed: 5276.691924 samples/sec	accuracy=0.408530	lr=0.399098
Epoch[8] Batch [649]	Speed: 5614.191466 samples/sec	accuracy=0.408744	lr=0.399077
Epoch[8] Batch [699]	Speed: 5806.362535 samples/sec	accuracy=0.408484	lr=0.399056
Epoch[8] Batch [749]	Speed: 5422.043046 samples/sec	accuracy=0.408577	lr=0.399034
Epoch[8] Batch [799]	Speed: 5647.918934 samples/sec	accuracy=0.408645	lr=0.399013
Epoch[8] Batch [849]	Speed: 5659.560439 samples/sec	accuracy=0.408479	lr=0.398991
Epoch[8] Batch [899]	Speed: 5657.782483 samples/sec	accuracy=0.408689	lr=0.398969
Epoch[8] Batch [949]	Speed: 5637.606014 samples/sec	accuracy=0.408614	lr=0.398947
Epoch[8] Batch [999]	Speed: 5510.029107 samples/sec	accuracy=0.408828	lr=0.398924
Epoch[8] Batch [1049]	Speed: 5437.110691 samples/sec	accuracy=0.408887	lr=0.398901
Epoch[8] Batch [1099]	Speed: 5481.173762 samples/sec	accuracy=0.408968	lr=0.398878
Epoch[8] Batch [1149]	Speed: 5709.288292 samples/sec	accuracy=0.409061	lr=0.398855
Epoch[8] Batch [1199]	Speed: 5860.674344 samples/sec	accuracy=0.409119	lr=0.398832
Epoch[8] Batch [1249]	Speed: 5709.032846 samples/sec	accuracy=0.409394	lr=0.398808
[Epoch 8] training: accuracy=0.409411
[Epoch 8] speed: 5639 samples/sec	time cost: 236.917414
[Epoch 8] validation: err-top1=0.653559 err-top5=0.406908
Epoch[9] Batch [49]	Speed: 5268.802612 samples/sec	accuracy=0.426055	lr=0.398784
Epoch[9] Batch [99]	Speed: 5454.646946 samples/sec	accuracy=0.424268	lr=0.398760
Epoch[9] Batch [149]	Speed: 5567.928776 samples/sec	accuracy=0.422318	lr=0.398735
Epoch[9] Batch [199]	Speed: 5252.308252 samples/sec	accuracy=0.421426	lr=0.398710
Epoch[9] Batch [249]	Speed: 5627.728563 samples/sec	accuracy=0.421102	lr=0.398686
Epoch[9] Batch [299]	Speed: 5753.807115 samples/sec	accuracy=0.421146	lr=0.398660
Epoch[9] Batch [349]	Speed: 5691.284887 samples/sec	accuracy=0.420734	lr=0.398635
Epoch[9] Batch [399]	Speed: 5313.223920 samples/sec	accuracy=0.420669	lr=0.398610
Epoch[9] Batch [449]	Speed: 5852.253305 samples/sec	accuracy=0.420256	lr=0.398584
Epoch[9] Batch [499]	Speed: 5558.522376 samples/sec	accuracy=0.419789	lr=0.398558
Epoch[9] Batch [549]	Speed: 5540.939672 samples/sec	accuracy=0.419606	lr=0.398531
Epoch[9] Batch [599]	Speed: 5459.690494 samples/sec	accuracy=0.419435	lr=0.398505
Epoch[9] Batch [649]	Speed: 5846.946835 samples/sec	accuracy=0.419178	lr=0.398478
Epoch[9] Batch [699]	Speed: 5556.106019 samples/sec	accuracy=0.418994	lr=0.398451
Epoch[9] Batch [749]	Speed: 5492.286949 samples/sec	accuracy=0.418802	lr=0.398424
Epoch[9] Batch [799]	Speed: 5573.821140 samples/sec	accuracy=0.418616	lr=0.398396
Epoch[9] Batch [849]	Speed: 5649.253744 samples/sec	accuracy=0.418442	lr=0.398369
Epoch[9] Batch [899]	Speed: 5632.844019 samples/sec	accuracy=0.418363	lr=0.398341
Epoch[9] Batch [949]	Speed: 5650.824261 samples/sec	accuracy=0.418408	lr=0.398312
Epoch[9] Batch [999]	Speed: 5609.093262 samples/sec	accuracy=0.418280	lr=0.398284
Epoch[9] Batch [1049]	Speed: 5789.109318 samples/sec	accuracy=0.418174	lr=0.398255
Epoch[9] Batch [1099]	Speed: 5334.444118 samples/sec	accuracy=0.418239	lr=0.398226
Epoch[9] Batch [1149]	Speed: 5598.630894 samples/sec	accuracy=0.418216	lr=0.398197
Epoch[9] Batch [1199]	Speed: 5651.580918 samples/sec	accuracy=0.418186	lr=0.398168
Epoch[9] Batch [1249]	Speed: 5527.789505 samples/sec	accuracy=0.418159	lr=0.398138
[Epoch 9] training: accuracy=0.418169
[Epoch 9] speed: 5553 samples/sec	time cost: 240.376797
[Epoch 9] validation: err-top1=0.581433 err-top5=0.317961
Epoch[10] Batch [49]	Speed: 5418.552488 samples/sec	accuracy=0.429668	lr=0.398108
Epoch[10] Batch [99]	Speed: 5450.512040 samples/sec	accuracy=0.427705	lr=0.398078
Epoch[10] Batch [149]	Speed: 5623.647183 samples/sec	accuracy=0.427155	lr=0.398048
Epoch[10] Batch [199]	Speed: 5845.385393 samples/sec	accuracy=0.426211	lr=0.398017
Epoch[10] Batch [249]	Speed: 5931.607719 samples/sec	accuracy=0.425457	lr=0.397986
Epoch[10] Batch [299]	Speed: 5586.245334 samples/sec	accuracy=0.425238	lr=0.397955
Epoch[10] Batch [349]	Speed: 5465.425842 samples/sec	accuracy=0.425312	lr=0.397924
Epoch[10] Batch [399]	Speed: 5602.793262 samples/sec	accuracy=0.425056	lr=0.397892
Epoch[10] Batch [449]	Speed: 5467.366936 samples/sec	accuracy=0.424835	lr=0.397861
Epoch[10] Batch [499]	Speed: 5504.444736 samples/sec	accuracy=0.424701	lr=0.397829
Epoch[10] Batch [549]	Speed: 5692.041104 samples/sec	accuracy=0.424561	lr=0.397796
Epoch[10] Batch [599]	Speed: 5531.766940 samples/sec	accuracy=0.424611	lr=0.397764
Epoch[10] Batch [649]	Speed: 5417.383498 samples/sec	accuracy=0.424617	lr=0.397731
Epoch[10] Batch [699]	Speed: 5448.542113 samples/sec	accuracy=0.424559	lr=0.397698
Epoch[10] Batch [749]	Speed: 5682.262772 samples/sec	accuracy=0.424703	lr=0.397665
Epoch[10] Batch [799]	Speed: 5774.078441 samples/sec	accuracy=0.424678	lr=0.397632
Epoch[10] Batch [849]	Speed: 5580.354819 samples/sec	accuracy=0.424691	lr=0.397598
Epoch[10] Batch [899]	Speed: 5902.766493 samples/sec	accuracy=0.424654	lr=0.397564
Epoch[10] Batch [949]	Speed: 5524.427398 samples/sec	accuracy=0.424703	lr=0.397530
Epoch[10] Batch [999]	Speed: 5608.961263 samples/sec	accuracy=0.424854	lr=0.397496
Epoch[10] Batch [1049]	Speed: 5732.162317 samples/sec	accuracy=0.424720	lr=0.397461
Epoch[10] Batch [1099]	Speed: 5612.973519 samples/sec	accuracy=0.424633	lr=0.397427
Epoch[10] Batch [1149]	Speed: 5731.458425 samples/sec	accuracy=0.424642	lr=0.397392
Epoch[10] Batch [1199]	Speed: 5807.865654 samples/sec	accuracy=0.424662	lr=0.397356
Epoch[10] Batch [1249]	Speed: 5526.620558 samples/sec	accuracy=0.424695	lr=0.397321
[Epoch 10] training: accuracy=0.424706
[Epoch 10] speed: 5605 samples/sec	time cost: 238.678051
[Epoch 10] validation: err-top1=0.545360 err-top5=0.286272
Epoch[11] Batch [49]	Speed: 5440.616662 samples/sec	accuracy=0.431816	lr=0.397284
Epoch[11] Batch [99]	Speed: 5605.831007 samples/sec	accuracy=0.434072	lr=0.397248
Epoch[11] Batch [149]	Speed: 5944.847474 samples/sec	accuracy=0.432467	lr=0.397212
Epoch[11] Batch [199]	Speed: 5639.924340 samples/sec	accuracy=0.432026	lr=0.397176
Epoch[11] Batch [249]	Speed: 5984.659391 samples/sec	accuracy=0.431469	lr=0.397139
Epoch[11] Batch [299]	Speed: 5933.624433 samples/sec	accuracy=0.431566	lr=0.397102
Epoch[11] Batch [349]	Speed: 5652.177404 samples/sec	accuracy=0.431152	lr=0.397065
Epoch[11] Batch [399]	Speed: 5604.537552 samples/sec	accuracy=0.430957	lr=0.397028
Epoch[11] Batch [449]	Speed: 5774.979040 samples/sec	accuracy=0.430799	lr=0.396990
Epoch[11] Batch [499]	Speed: 5108.490044 samples/sec	accuracy=0.430670	lr=0.396952
Epoch[11] Batch [549]	Speed: 5721.869180 samples/sec	accuracy=0.430726	lr=0.396914
Epoch[11] Batch [599]	Speed: 5752.616913 samples/sec	accuracy=0.430628	lr=0.396876
Epoch[11] Batch [649]	Speed: 5844.690167 samples/sec	accuracy=0.430422	lr=0.396837
Epoch[11] Batch [699]	Speed: 5911.760319 samples/sec	accuracy=0.430516	lr=0.396798
Epoch[11] Batch [749]	Speed: 5656.562239 samples/sec	accuracy=0.430668	lr=0.396759
Epoch[11] Batch [799]	Speed: 5556.550390 samples/sec	accuracy=0.430634	lr=0.396720
Epoch[11] Batch [849]	Speed: 5885.704068 samples/sec	accuracy=0.430723	lr=0.396680
Epoch[11] Batch [899]	Speed: 5744.197856 samples/sec	accuracy=0.430711	lr=0.396641
Epoch[11] Batch [949]	Speed: 5251.307862 samples/sec	accuracy=0.430297	lr=0.396601
Epoch[11] Batch [999]	Speed: 5760.215718 samples/sec	accuracy=0.430140	lr=0.396561
Epoch[11] Batch [1049]	Speed: 5846.377292 samples/sec	accuracy=0.430024	lr=0.396520
Epoch[11] Batch [1099]	Speed: 5190.616724 samples/sec	accuracy=0.430207	lr=0.396479
Epoch[11] Batch [1149]	Speed: 5439.832895 samples/sec	accuracy=0.430199	lr=0.396439
Epoch[11] Batch [1199]	Speed: 5361.944892 samples/sec	accuracy=0.430181	lr=0.396397
Epoch[11] Batch [1249]	Speed: 5533.578067 samples/sec	accuracy=0.430243	lr=0.396356
[Epoch 11] training: accuracy=0.430248
[Epoch 11] speed: 5628 samples/sec	time cost: 237.517826
[Epoch 11] validation: err-top1=0.573303 err-top5=0.313558
Epoch[12] Batch [49]	Speed: 5196.060903 samples/sec	accuracy=0.439023	lr=0.396314
Epoch[12] Batch [99]	Speed: 5957.645460 samples/sec	accuracy=0.439219	lr=0.396272
Epoch[12] Batch [149]	Speed: 5549.514666 samples/sec	accuracy=0.437832	lr=0.396230
Epoch[12] Batch [199]	Speed: 6016.220661 samples/sec	accuracy=0.437285	lr=0.396187
Epoch[12] Batch [249]	Speed: 5579.077730 samples/sec	accuracy=0.436246	lr=0.396145
Epoch[12] Batch [299]	Speed: 5664.778933 samples/sec	accuracy=0.435840	lr=0.396102
Epoch[12] Batch [349]	Speed: 5269.322065 samples/sec	accuracy=0.435904	lr=0.396059
Epoch[12] Batch [399]	Speed: 5679.723089 samples/sec	accuracy=0.435312	lr=0.396016
Epoch[12] Batch [449]	Speed: 5838.845358 samples/sec	accuracy=0.435406	lr=0.395972
Epoch[12] Batch [499]	Speed: 5434.262218 samples/sec	accuracy=0.435135	lr=0.395928
Epoch[12] Batch [549]	Speed: 5792.085052 samples/sec	accuracy=0.434950	lr=0.395885
Epoch[12] Batch [599]	Speed: 5649.315419 samples/sec	accuracy=0.435186	lr=0.395840
Epoch[12] Batch [649]	Speed: 5158.288848 samples/sec	accuracy=0.435168	lr=0.395796
Epoch[12] Batch [699]	Speed: 5351.967097 samples/sec	accuracy=0.435137	lr=0.395751
Epoch[12] Batch [749]	Speed: 5584.057458 samples/sec	accuracy=0.434964	lr=0.395706
Epoch[12] Batch [799]	Speed: 5642.256877 samples/sec	accuracy=0.434890	lr=0.395661
Epoch[12] Batch [849]	Speed: 5684.198778 samples/sec	accuracy=0.434574	lr=0.395616
Epoch[12] Batch [899]	Speed: 5726.776539 samples/sec	accuracy=0.434457	lr=0.395570
Epoch[12] Batch [949]	Speed: 5965.549193 samples/sec	accuracy=0.434338	lr=0.395524
Epoch[12] Batch [999]	Speed: 5754.263782 samples/sec	accuracy=0.434205	lr=0.395478
Epoch[12] Batch [1049]	Speed: 5578.372675 samples/sec	accuracy=0.434046	lr=0.395432
Epoch[12] Batch [1099]	Speed: 5639.137927 samples/sec	accuracy=0.433883	lr=0.395386
Epoch[12] Batch [1149]	Speed: 5667.701037 samples/sec	accuracy=0.433953	lr=0.395339
Epoch[12] Batch [1199]	Speed: 5730.385097 samples/sec	accuracy=0.434035	lr=0.395292
Epoch[12] Batch [1249]	Speed: 5840.460489 samples/sec	accuracy=0.434264	lr=0.395245
[Epoch 12] training: accuracy=0.434272
[Epoch 12] speed: 5623 samples/sec	time cost: 237.707804
[Epoch 12] validation: err-top1=0.606884 err-top5=0.346720
Epoch[13] Batch [49]	Speed: 5653.835135 samples/sec	accuracy=0.440352	lr=0.395196
Epoch[13] Batch [99]	Speed: 5833.311371 samples/sec	accuracy=0.441484	lr=0.395149
Epoch[13] Batch [149]	Speed: 5258.786803 samples/sec	accuracy=0.441868	lr=0.395101
Epoch[13] Batch [199]	Speed: 5854.792909 samples/sec	accuracy=0.442007	lr=0.395052
Epoch[13] Batch [249]	Speed: 5569.078869 samples/sec	accuracy=0.441570	lr=0.395004
Epoch[13] Batch [299]	Speed: 5517.442401 samples/sec	accuracy=0.441064	lr=0.394955
Epoch[13] Batch [349]	Speed: 5554.807532 samples/sec	accuracy=0.440499	lr=0.394907
Epoch[13] Batch [399]	Speed: 5650.395311 samples/sec	accuracy=0.440273	lr=0.394858
Epoch[13] Batch [449]	Speed: 5487.571146 samples/sec	accuracy=0.440538	lr=0.394808
Epoch[13] Batch [499]	Speed: 5498.652591 samples/sec	accuracy=0.440520	lr=0.394759
Epoch[13] Batch [549]	Speed: 5931.573149 samples/sec	accuracy=0.440376	lr=0.394709
Epoch[13] Batch [599]	Speed: 5690.410201 samples/sec	accuracy=0.440223	lr=0.394659
Epoch[13] Batch [649]	Speed: 5692.787712 samples/sec	accuracy=0.439775	lr=0.394609
Epoch[13] Batch [699]	Speed: 5925.137935 samples/sec	accuracy=0.439710	lr=0.394558
Epoch[13] Batch [749]	Speed: 5542.922767 samples/sec	accuracy=0.439730	lr=0.394507
Epoch[13] Batch [799]	Speed: 5672.709030 samples/sec	accuracy=0.439613	lr=0.394457
Epoch[13] Batch [849]	Speed: 5312.119641 samples/sec	accuracy=0.439431	lr=0.394405
Epoch[13] Batch [899]	Speed: 5506.501331 samples/sec	accuracy=0.438844	lr=0.394354
Epoch[13] Batch [949]	Speed: 5620.424614 samples/sec	accuracy=0.438659	lr=0.394302
Epoch[13] Batch [999]	Speed: 5706.800073 samples/sec	accuracy=0.438800	lr=0.394250
Epoch[13] Batch [1049]	Speed: 5976.488075 samples/sec	accuracy=0.438618	lr=0.394198
Epoch[13] Batch [1099]	Speed: 6053.693123 samples/sec	accuracy=0.438597	lr=0.394146
Epoch[13] Batch [1149]	Speed: 5817.078477 samples/sec	accuracy=0.438531	lr=0.394093
Epoch[13] Batch [1199]	Speed: 5542.865826 samples/sec	accuracy=0.438452	lr=0.394041
Epoch[13] Batch [1249]	Speed: 5754.407952 samples/sec	accuracy=0.438250	lr=0.393988
[Epoch 13] training: accuracy=0.438261
[Epoch 13] speed: 5653 samples/sec	time cost: 236.378475
[Epoch 13] validation: err-top1=0.550522 err-top5=0.289182
Epoch[14] Batch [49]	Speed: 5355.567361 samples/sec	accuracy=0.443281	lr=0.393933
Epoch[14] Batch [99]	Speed: 5621.924389 samples/sec	accuracy=0.446943	lr=0.393880
Epoch[14] Batch [149]	Speed: 5641.400307 samples/sec	accuracy=0.446042	lr=0.393826
Epoch[14] Batch [199]	Speed: 5939.151561 samples/sec	accuracy=0.444551	lr=0.393772
Epoch[14] Batch [249]	Speed: 5710.082247 samples/sec	accuracy=0.444363	lr=0.393718
Epoch[14] Batch [299]	Speed: 5801.803449 samples/sec	accuracy=0.443411	lr=0.393664
Epoch[14] Batch [349]	Speed: 5481.663735 samples/sec	accuracy=0.442743	lr=0.393609
Epoch[14] Batch [399]	Speed: 5933.751168 samples/sec	accuracy=0.443000	lr=0.393554
Epoch[14] Batch [449]	Speed: 5638.723630 samples/sec	accuracy=0.443153	lr=0.393499
Epoch[14] Batch [499]	Speed: 5999.983259 samples/sec	accuracy=0.443051	lr=0.393444
Epoch[14] Batch [549]	Speed: 5651.300122 samples/sec	accuracy=0.442740	lr=0.393388
Epoch[14] Batch [599]	Speed: 5789.666352 samples/sec	accuracy=0.442695	lr=0.393332
Epoch[14] Batch [649]	Speed: 5863.680575 samples/sec	accuracy=0.442529	lr=0.393276
Epoch[14] Batch [699]	Speed: 5040.151355 samples/sec	accuracy=0.442347	lr=0.393220
Epoch[14] Batch [749]	Speed: 5323.141168 samples/sec	accuracy=0.442259	lr=0.393163
Epoch[14] Batch [799]	Speed: 5168.651551 samples/sec	accuracy=0.442472	lr=0.393107
Epoch[14] Batch [849]	Speed: 5459.656210 samples/sec	accuracy=0.442498	lr=0.393050
Epoch[14] Batch [899]	Speed: 5730.133723 samples/sec	accuracy=0.442274	lr=0.392993
Epoch[14] Batch [949]	Speed: 5952.611331 samples/sec	accuracy=0.442231	lr=0.392935
Epoch[14] Batch [999]	Speed: 5767.772814 samples/sec	accuracy=0.442341	lr=0.392877
Epoch[14] Batch [1049]	Speed: 5534.987906 samples/sec	accuracy=0.442398	lr=0.392820
Epoch[14] Batch [1099]	Speed: 5321.014077 samples/sec	accuracy=0.442305	lr=0.392762
Epoch[14] Batch [1149]	Speed: 5709.483041 samples/sec	accuracy=0.442238	lr=0.392703
Epoch[14] Batch [1199]	Speed: 5581.895815 samples/sec	accuracy=0.442111	lr=0.392645
Epoch[14] Batch [1249]	Speed: 5589.219478 samples/sec	accuracy=0.441942	lr=0.392586
[Epoch 14] training: accuracy=0.441934
[Epoch 14] speed: 5607 samples/sec	time cost: 238.491864
[Epoch 14] validation: err-top1=0.545898 err-top5=0.286173
Epoch[15] Batch [49]	Speed: 5358.694390 samples/sec	accuracy=0.451309	lr=0.392526
Epoch[15] Batch [99]	Speed: 5557.632932 samples/sec	accuracy=0.448984	lr=0.392466
Epoch[15] Batch [149]	Speed: 5331.199202 samples/sec	accuracy=0.448737	lr=0.392407
Epoch[15] Batch [199]	Speed: 5626.790887 samples/sec	accuracy=0.448086	lr=0.392347
Epoch[15] Batch [249]	Speed: 5769.854039 samples/sec	accuracy=0.447551	lr=0.392287
Epoch[15] Batch [299]	Speed: 5352.042592 samples/sec	accuracy=0.446885	lr=0.392227
Epoch[15] Batch [349]	Speed: 5861.559282 samples/sec	accuracy=0.446850	lr=0.392167
Epoch[15] Batch [399]	Speed: 5098.012797 samples/sec	accuracy=0.446504	lr=0.392106
Epoch[15] Batch [449]	Speed: 5755.758095 samples/sec	accuracy=0.446287	lr=0.392045
Epoch[15] Batch [499]	Speed: 5532.635580 samples/sec	accuracy=0.446023	lr=0.391984
Epoch[15] Batch [549]	Speed: 5801.360519 samples/sec	accuracy=0.446009	lr=0.391923
Epoch[15] Batch [599]	Speed: 5947.893914 samples/sec	accuracy=0.445674	lr=0.391861
Epoch[15] Batch [649]	Speed: 5720.887071 samples/sec	accuracy=0.445616	lr=0.391799
Epoch[15] Batch [699]	Speed: 4967.641033 samples/sec	accuracy=0.445555	lr=0.391737
Epoch[15] Batch [749]	Speed: 5528.654334 samples/sec	accuracy=0.445384	lr=0.391675
Epoch[15] Batch [799]	Speed: 5642.726996 samples/sec	accuracy=0.445610	lr=0.391613
Epoch[15] Batch [849]	Speed: 5259.571692 samples/sec	accuracy=0.445728	lr=0.391550
Epoch[15] Batch [899]	Speed: 5736.845815 samples/sec	accuracy=0.445730	lr=0.391487
Epoch[15] Batch [949]	Speed: 5129.332409 samples/sec	accuracy=0.445843	lr=0.391424
Epoch[15] Batch [999]	Speed: 5723.783607 samples/sec	accuracy=0.445770	lr=0.391361
Epoch[15] Batch [1049]	Speed: 5745.711387 samples/sec	accuracy=0.445621	lr=0.391297
Epoch[15] Batch [1099]	Speed: 5795.969440 samples/sec	accuracy=0.445447	lr=0.391233
Epoch[15] Batch [1149]	Speed: 5639.511260 samples/sec	accuracy=0.445532	lr=0.391169
Epoch[15] Batch [1199]	Speed: 5675.131759 samples/sec	accuracy=0.445576	lr=0.391105
Epoch[15] Batch [1249]	Speed: 5477.861537 samples/sec	accuracy=0.445407	lr=0.391040
[Epoch 15] training: accuracy=0.445409
[Epoch 15] speed: 5543 samples/sec	time cost: 240.867455
[Epoch 15] validation: err-top1=0.553890 err-top5=0.293148
Epoch[16] Batch [49]	Speed: 5298.151386 samples/sec	accuracy=0.452305	lr=0.390974
Epoch[16] Batch [99]	Speed: 5542.819043 samples/sec	accuracy=0.450361	lr=0.390909
Epoch[16] Batch [149]	Speed: 5440.712047 samples/sec	accuracy=0.451582	lr=0.390844
Epoch[16] Batch [199]	Speed: 5552.811043 samples/sec	accuracy=0.451230	lr=0.390779
Epoch[16] Batch [249]	Speed: 5989.260270 samples/sec	accuracy=0.450102	lr=0.390713
Epoch[16] Batch [299]	Speed: 5395.554838 samples/sec	accuracy=0.449154	lr=0.390647
Epoch[16] Batch [349]	Speed: 5682.363511 samples/sec	accuracy=0.449300	lr=0.390581
Epoch[16] Batch [399]	Speed: 5392.494605 samples/sec	accuracy=0.449216	lr=0.390515
Epoch[16] Batch [449]	Speed: 5793.655501 samples/sec	accuracy=0.449178	lr=0.390448
Epoch[16] Batch [499]	Speed: 5710.658497 samples/sec	accuracy=0.449080	lr=0.390381
Epoch[16] Batch [549]	Speed: 5571.029410 samples/sec	accuracy=0.448851	lr=0.390314
Epoch[16] Batch [599]	Speed: 5404.278460 samples/sec	accuracy=0.448618	lr=0.390247
Epoch[16] Batch [649]	Speed: 5217.792304 samples/sec	accuracy=0.448313	lr=0.390180
Epoch[16] Batch [699]	Speed: 5846.285774 samples/sec	accuracy=0.448436	lr=0.390112
Epoch[16] Batch [749]	Speed: 5773.978772 samples/sec	accuracy=0.448029	lr=0.390044
Epoch[16] Batch [799]	Speed: 6094.318140 samples/sec	accuracy=0.447781	lr=0.389976
Epoch[16] Batch [849]	Speed: 5457.079566 samples/sec	accuracy=0.447668	lr=0.389907
Epoch[16] Batch [899]	Speed: 5797.646864 samples/sec	accuracy=0.447673	lr=0.389839
Epoch[16] Batch [949]	Speed: 5707.137827 samples/sec	accuracy=0.447643	lr=0.389770
Epoch[16] Batch [999]	Speed: 5852.133535 samples/sec	accuracy=0.447628	lr=0.389701
Epoch[16] Batch [1049]	Speed: 5810.616216 samples/sec	accuracy=0.447356	lr=0.389632
Epoch[16] Batch [1099]	Speed: 5651.977024 samples/sec	accuracy=0.447466	lr=0.389562
Epoch[16] Batch [1149]	Speed: 5616.688883 samples/sec	accuracy=0.447600	lr=0.389492
Epoch[16] Batch [1199]	Speed: 5827.513798 samples/sec	accuracy=0.447557	lr=0.389422
Epoch[16] Batch [1249]	Speed: 5363.622127 samples/sec	accuracy=0.447622	lr=0.389352
[Epoch 16] training: accuracy=0.447628
[Epoch 16] speed: 5613 samples/sec	time cost: 238.342772
[Epoch 16] validation: err-top1=0.527982 err-top5=0.272142
Epoch[17] Batch [49]	Speed: 5176.614531 samples/sec	accuracy=0.457383	lr=0.389280
Epoch[17] Batch [99]	Speed: 5826.002547 samples/sec	accuracy=0.454150	lr=0.389210
Epoch[17] Batch [149]	Speed: 5558.195940 samples/sec	accuracy=0.453503	lr=0.389139
Epoch[17] Batch [199]	Speed: 5630.882001 samples/sec	accuracy=0.453511	lr=0.389068
Epoch[17] Batch [249]	Speed: 5579.798186 samples/sec	accuracy=0.452344	lr=0.388996
Epoch[17] Batch [299]	Speed: 5799.335908 samples/sec	accuracy=0.451930	lr=0.388925
Epoch[17] Batch [349]	Speed: 5538.099366 samples/sec	accuracy=0.452003	lr=0.388853
Epoch[17] Batch [399]	Speed: 5543.701176 samples/sec	accuracy=0.452112	lr=0.388781
Epoch[17] Batch [449]	Speed: 5221.086933 samples/sec	accuracy=0.451966	lr=0.388709
Epoch[17] Batch [499]	Speed: 5558.882090 samples/sec	accuracy=0.451635	lr=0.388636
Epoch[17] Batch [549]	Speed: 5858.722250 samples/sec	accuracy=0.451486	lr=0.388564
Epoch[17] Batch [599]	Speed: 5492.248602 samples/sec	accuracy=0.451864	lr=0.388491
Epoch[17] Batch [649]	Speed: 5462.675886 samples/sec	accuracy=0.451310	lr=0.388418
Epoch[17] Batch [699]	Speed: 5528.945138 samples/sec	accuracy=0.450993	lr=0.388344
Epoch[17] Batch [749]	Speed: 5415.131690 samples/sec	accuracy=0.450969	lr=0.388271
Epoch[17] Batch [799]	Speed: 5541.674907 samples/sec	accuracy=0.450929	lr=0.388197
Epoch[17] Batch [849]	Speed: 5803.845626 samples/sec	accuracy=0.450700	lr=0.388123
Epoch[17] Batch [899]	Speed: 5626.731030 samples/sec	accuracy=0.450533	lr=0.388049
Epoch[17] Batch [949]	Speed: 5800.845263 samples/sec	accuracy=0.450663	lr=0.387974
Epoch[17] Batch [999]	Speed: 5818.607645 samples/sec	accuracy=0.450696	lr=0.387900
Epoch[17] Batch [1049]	Speed: 6019.207593 samples/sec	accuracy=0.450658	lr=0.387825
Epoch[17] Batch [1099]	Speed: 5452.984311 samples/sec	accuracy=0.450518	lr=0.387750
Epoch[17] Batch [1149]	Speed: 5511.903551 samples/sec	accuracy=0.450383	lr=0.387674
Epoch[17] Batch [1199]	Speed: 5815.683036 samples/sec	accuracy=0.450470	lr=0.387599
Epoch[17] Batch [1249]	Speed: 5831.952478 samples/sec	accuracy=0.450333	lr=0.387523
[Epoch 17] training: accuracy=0.450318
[Epoch 17] speed: 5602 samples/sec	time cost: 238.833775
[Epoch 17] validation: err-top1=0.533752 err-top5=0.278300
Epoch[18] Batch [49]	Speed: 5546.111622 samples/sec	accuracy=0.455254	lr=0.387445
Epoch[18] Batch [99]	Speed: 5726.861757 samples/sec	accuracy=0.455303	lr=0.387369
Epoch[18] Batch [149]	Speed: 5579.103385 samples/sec	accuracy=0.455645	lr=0.387293
Epoch[18] Batch [199]	Speed: 5642.517946 samples/sec	accuracy=0.456016	lr=0.387216
Epoch[18] Batch [249]	Speed: 5886.128350 samples/sec	accuracy=0.455914	lr=0.387139
Epoch[18] Batch [299]	Speed: 5717.449985 samples/sec	accuracy=0.455964	lr=0.387062
Epoch[18] Batch [349]	Speed: 5529.606001 samples/sec	accuracy=0.456099	lr=0.386984
Epoch[18] Batch [399]	Speed: 5810.839166 samples/sec	accuracy=0.456191	lr=0.386907
Epoch[18] Batch [449]	Speed: 5622.187849 samples/sec	accuracy=0.455816	lr=0.386829
Epoch[18] Batch [499]	Speed: 5884.058661 samples/sec	accuracy=0.455371	lr=0.386751
Epoch[18] Batch [549]	Speed: 5443.304436 samples/sec	accuracy=0.454913	lr=0.386673
Epoch[18] Batch [599]	Speed: 5845.355322 samples/sec	accuracy=0.454938	lr=0.386594
Epoch[18] Batch [649]	Speed: 5563.468488 samples/sec	accuracy=0.454443	lr=0.386515
Epoch[18] Batch [699]	Speed: 5962.952851 samples/sec	accuracy=0.454280	lr=0.386436
Epoch[18] Batch [749]	Speed: 5914.356730 samples/sec	accuracy=0.454194	lr=0.386357
Epoch[18] Batch [799]	Speed: 5802.589323 samples/sec	accuracy=0.454126	lr=0.386278
Epoch[18] Batch [849]	Speed: 5384.115092 samples/sec	accuracy=0.453892	lr=0.386198
Epoch[18] Batch [899]	Speed: 5504.113335 samples/sec	accuracy=0.453618	lr=0.386119
Epoch[18] Batch [949]	Speed: 5822.728200 samples/sec	accuracy=0.453583	lr=0.386038
Epoch[18] Batch [999]	Speed: 5496.765901 samples/sec	accuracy=0.453685	lr=0.385958
Epoch[18] Batch [1049]	Speed: 5928.877467 samples/sec	accuracy=0.453631	lr=0.385878
Epoch[18] Batch [1099]	Speed: 5584.182915 samples/sec	accuracy=0.453466	lr=0.385797
Epoch[18] Batch [1149]	Speed: 5951.309103 samples/sec	accuracy=0.453411	lr=0.385716
Epoch[18] Batch [1199]	Speed: 5494.775057 samples/sec	accuracy=0.453344	lr=0.385635
Epoch[18] Batch [1249]	Speed: 6185.393179 samples/sec	accuracy=0.453396	lr=0.385554
[Epoch 18] training: accuracy=0.453387
[Epoch 18] speed: 5701 samples/sec	time cost: 234.543132
[Epoch 18] validation: err-top1=0.530871 err-top5=0.271464
Epoch[19] Batch [49]	Speed: 5937.887559 samples/sec	accuracy=0.464668	lr=0.385470
Epoch[19] Batch [99]	Speed: 5536.815275 samples/sec	accuracy=0.460479	lr=0.385388
Epoch[19] Batch [149]	Speed: 5874.194256 samples/sec	accuracy=0.460046	lr=0.385306
Epoch[19] Batch [199]	Speed: 5592.808773 samples/sec	accuracy=0.458213	lr=0.385224
Epoch[19] Batch [249]	Speed: 5849.058836 samples/sec	accuracy=0.458059	lr=0.385142
Epoch[19] Batch [299]	Speed: 5769.200530 samples/sec	accuracy=0.458649	lr=0.385059
Epoch[19] Batch [349]	Speed: 5779.058851 samples/sec	accuracy=0.458172	lr=0.384976
Epoch[19] Batch [399]	Speed: 5750.853631 samples/sec	accuracy=0.457717	lr=0.384893
Epoch[19] Batch [449]	Speed: 5721.250121 samples/sec	accuracy=0.457504	lr=0.384810
Epoch[19] Batch [499]	Speed: 5683.257383 samples/sec	accuracy=0.457324	lr=0.384726
Epoch[19] Batch [549]	Speed: 5342.241519 samples/sec	accuracy=0.456731	lr=0.384642
Epoch[19] Batch [599]	Speed: 5905.807140 samples/sec	accuracy=0.456597	lr=0.384558
Epoch[19] Batch [649]	Speed: 5611.937506 samples/sec	accuracy=0.456525	lr=0.384474
Epoch[19] Batch [699]	Speed: 5424.099481 samples/sec	accuracy=0.456426	lr=0.384389
Epoch[19] Batch [749]	Speed: 5728.746975 samples/sec	accuracy=0.456324	lr=0.384305
Epoch[19] Batch [799]	Speed: 6004.408452 samples/sec	accuracy=0.456124	lr=0.384220
Epoch[19] Batch [849]	Speed: 5839.404066 samples/sec	accuracy=0.456129	lr=0.384135
Epoch[19] Batch [899]	Speed: 6007.562811 samples/sec	accuracy=0.456022	lr=0.384049
Epoch[19] Batch [949]	Speed: 5706.809627 samples/sec	accuracy=0.455949	lr=0.383964
Epoch[19] Batch [999]	Speed: 5980.418465 samples/sec	accuracy=0.455862	lr=0.383878
Epoch[19] Batch [1049]	Speed: 5956.458327 samples/sec	accuracy=0.455760	lr=0.383792
Epoch[19] Batch [1099]	Speed: 5927.385832 samples/sec	accuracy=0.455566	lr=0.383706
Epoch[19] Batch [1149]	Speed: 5871.774421 samples/sec	accuracy=0.455386	lr=0.383619
Epoch[19] Batch [1199]	Speed: 5899.630918 samples/sec	accuracy=0.455124	lr=0.383533
Epoch[19] Batch [1249]	Speed: 5138.282458 samples/sec	accuracy=0.455038	lr=0.383446
[Epoch 19] training: accuracy=0.455052
[Epoch 19] speed: 5738 samples/sec	time cost: 232.676737
[Epoch 19] validation: err-top1=0.580576 err-top5=0.319535
Epoch[20] Batch [49]	Speed: 5383.760094 samples/sec	accuracy=0.460957	lr=0.383357
Epoch[20] Batch [99]	Speed: 5726.239327 samples/sec	accuracy=0.461543	lr=0.383270
Epoch[20] Batch [149]	Speed: 5427.030431 samples/sec	accuracy=0.462643	lr=0.383182
Epoch[20] Batch [199]	Speed: 5841.398123 samples/sec	accuracy=0.461890	lr=0.383094
Epoch[20] Batch [249]	Speed: 5956.242731 samples/sec	accuracy=0.461832	lr=0.383006
Epoch[20] Batch [299]	Speed: 5605.381207 samples/sec	accuracy=0.461585	lr=0.382918
Epoch[20] Batch [349]	Speed: 6019.918128 samples/sec	accuracy=0.461035	lr=0.382830
Epoch[20] Batch [399]	Speed: 5764.968558 samples/sec	accuracy=0.460737	lr=0.382741
Epoch[20] Batch [449]	Speed: 5845.897453 samples/sec	accuracy=0.460399	lr=0.382652
Epoch[20] Batch [499]	Speed: 5554.965445 samples/sec	accuracy=0.460305	lr=0.382563
Epoch[20] Batch [549]	Speed: 6058.724876 samples/sec	accuracy=0.460030	lr=0.382474
Epoch[20] Batch [599]	Speed: 5667.002268 samples/sec	accuracy=0.459562	lr=0.382384
Epoch[20] Batch [649]	Speed: 5860.370468 samples/sec	accuracy=0.459434	lr=0.382295
Epoch[20] Batch [699]	Speed: 5726.483335 samples/sec	accuracy=0.458869	lr=0.382205
Epoch[20] Batch [749]	Speed: 6153.210016 samples/sec	accuracy=0.458964	lr=0.382115
Epoch[20] Batch [799]	Speed: 5762.032209 samples/sec	accuracy=0.458746	lr=0.382024
Epoch[20] Batch [849]	Speed: 5683.490220 samples/sec	accuracy=0.458440	lr=0.381934
Epoch[20] Batch [899]	Speed: 5842.528545 samples/sec	accuracy=0.458369	lr=0.381843
Epoch[20] Batch [949]	Speed: 5773.331003 samples/sec	accuracy=0.458134	lr=0.381752
Epoch[20] Batch [999]	Speed: 5837.832684 samples/sec	accuracy=0.457920	lr=0.381661
Epoch[20] Batch [1049]	Speed: 5733.974475 samples/sec	accuracy=0.457729	lr=0.381569
Epoch[20] Batch [1099]	Speed: 5657.322371 samples/sec	accuracy=0.457503	lr=0.381477
Epoch[20] Batch [1149]	Speed: 5235.271218 samples/sec	accuracy=0.457469	lr=0.381385
Epoch[20] Batch [1199]	Speed: 5531.633711 samples/sec	accuracy=0.457463	lr=0.381293
Epoch[20] Batch [1249]	Speed: 5564.279927 samples/sec	accuracy=0.457608	lr=0.381201
[Epoch 20] training: accuracy=0.457599
[Epoch 20] speed: 5716 samples/sec	time cost: 233.578967
[Epoch 20] validation: err-top1=0.511958 err-top5=0.255560
Epoch[21] Batch [49]	Speed: 5655.166489 samples/sec	accuracy=0.466230	lr=0.381107
Epoch[21] Batch [99]	Speed: 5714.450216 samples/sec	accuracy=0.465713	lr=0.381014
Epoch[21] Batch [149]	Speed: 5590.536434 samples/sec	accuracy=0.463997	lr=0.380921
Epoch[21] Batch [199]	Speed: 5926.554345 samples/sec	accuracy=0.462939	lr=0.380828
Epoch[21] Batch [249]	Speed: 5637.098274 samples/sec	accuracy=0.463008	lr=0.380734
Epoch[21] Batch [299]	Speed: 5720.533668 samples/sec	accuracy=0.462575	lr=0.380641
Epoch[21] Batch [349]	Speed: 5789.868029 samples/sec	accuracy=0.462492	lr=0.380547
Epoch[21] Batch [399]	Speed: 5603.609488 samples/sec	accuracy=0.462063	lr=0.380453
Epoch[21] Batch [449]	Speed: 5586.085056 samples/sec	accuracy=0.462044	lr=0.380359
Epoch[21] Batch [499]	Speed: 5806.122189 samples/sec	accuracy=0.462094	lr=0.380264
Epoch[21] Batch [549]	Speed: 5487.885832 samples/sec	accuracy=0.461676	lr=0.380169
Epoch[21] Batch [599]	Speed: 5405.283838 samples/sec	accuracy=0.461463	lr=0.380074
Epoch[21] Batch [649]	Speed: 5480.395190 samples/sec	accuracy=0.461248	lr=0.379979
Epoch[21] Batch [699]	Speed: 5401.354366 samples/sec	accuracy=0.461007	lr=0.379884
Epoch[21] Batch [749]	Speed: 5916.702249 samples/sec	accuracy=0.460837	lr=0.379788
Epoch[21] Batch [799]	Speed: 5526.691532 samples/sec	accuracy=0.460659	lr=0.379693
Epoch[21] Batch [849]	Speed: 5682.430421 samples/sec	accuracy=0.460679	lr=0.379597
Epoch[21] Batch [899]	Speed: 5973.149101 samples/sec	accuracy=0.460387	lr=0.379501
Epoch[21] Batch [949]	Speed: 5571.934568 samples/sec	accuracy=0.460199	lr=0.379404
Epoch[21] Batch [999]	Speed: 5719.997322 samples/sec	accuracy=0.460008	lr=0.379307
Epoch[21] Batch [1049]	Speed: 5722.545099 samples/sec	accuracy=0.459767	lr=0.379211
Epoch[21] Batch [1099]	Speed: 5545.733795 samples/sec	accuracy=0.459698	lr=0.379114
Epoch[21] Batch [1149]	Speed: 5678.712596 samples/sec	accuracy=0.459794	lr=0.379016
Epoch[21] Batch [1199]	Speed: 5942.018211 samples/sec	accuracy=0.459578	lr=0.378919
Epoch[21] Batch [1249]	Speed: 5627.614710 samples/sec	accuracy=0.459652	lr=0.378821
[Epoch 21] training: accuracy=0.459638
[Epoch 21] speed: 5658 samples/sec	time cost: 235.806049
[Epoch 21] validation: err-top1=0.523398 err-top5=0.265087
Epoch[22] Batch [49]	Speed: 5498.539817 samples/sec	accuracy=0.467344	lr=0.378721
Epoch[22] Batch [99]	Speed: 5553.703109 samples/sec	accuracy=0.467969	lr=0.378623
Epoch[22] Batch [149]	Speed: 5811.635196 samples/sec	accuracy=0.467064	lr=0.378525
Epoch[22] Batch [199]	Speed: 5424.518328 samples/sec	accuracy=0.466694	lr=0.378426
Epoch[22] Batch [249]	Speed: 5560.376263 samples/sec	accuracy=0.465953	lr=0.378328
Epoch[22] Batch [299]	Speed: 5518.190414 samples/sec	accuracy=0.465417	lr=0.378229
Epoch[22] Batch [349]	Speed: 5649.008249 samples/sec	accuracy=0.465086	lr=0.378129
Epoch[22] Batch [399]	Speed: 5658.095677 samples/sec	accuracy=0.464885	lr=0.378030
Epoch[22] Batch [449]	Speed: 5730.580065 samples/sec	accuracy=0.464058	lr=0.377930
Epoch[22] Batch [499]	Speed: 5655.308416 samples/sec	accuracy=0.464115	lr=0.377831
Epoch[22] Batch [549]	Speed: 5537.784321 samples/sec	accuracy=0.463981	lr=0.377731
Epoch[22] Batch [599]	Speed: 5869.789898 samples/sec	accuracy=0.463740	lr=0.377630
Epoch[22] Batch [649]	Speed: 5736.798766 samples/sec	accuracy=0.463199	lr=0.377530
Epoch[22] Batch [699]	Speed: 5856.781998 samples/sec	accuracy=0.462879	lr=0.377429
Epoch[22] Batch [749]	Speed: 5794.298146 samples/sec	accuracy=0.462785	lr=0.377328
Epoch[22] Batch [799]	Speed: 5948.509608 samples/sec	accuracy=0.462582	lr=0.377227
Epoch[22] Batch [849]	Speed: 5558.731292 samples/sec	accuracy=0.462407	lr=0.377126
Epoch[22] Batch [899]	Speed: 5600.848613 samples/sec	accuracy=0.462148	lr=0.377024
Epoch[22] Batch [949]	Speed: 5704.056597 samples/sec	accuracy=0.461987	lr=0.376923
Epoch[22] Batch [999]	Speed: 5342.806260 samples/sec	accuracy=0.461955	lr=0.376821
Epoch[22] Batch [1049]	Speed: 5680.211943 samples/sec	accuracy=0.461679	lr=0.376719
Epoch[22] Batch [1099]	Speed: 5489.264484 samples/sec	accuracy=0.461588	lr=0.376616
Epoch[22] Batch [1149]	Speed: 5762.931064 samples/sec	accuracy=0.461457	lr=0.376514
Epoch[22] Batch [1199]	Speed: 5447.477189 samples/sec	accuracy=0.461300	lr=0.376411
Epoch[22] Batch [1249]	Speed: 5799.224872 samples/sec	accuracy=0.461153	lr=0.376308
[Epoch 22] training: accuracy=0.461140
[Epoch 22] speed: 5637 samples/sec	time cost: 236.593913
[Epoch 22] validation: err-top1=0.571867 err-top5=0.311444
Epoch[23] Batch [49]	Speed: 5753.980554 samples/sec	accuracy=0.469609	lr=0.376203
Epoch[23] Batch [99]	Speed: 5583.850409 samples/sec	accuracy=0.469307	lr=0.376099
Epoch[23] Batch [149]	Speed: 5768.683998 samples/sec	accuracy=0.469401	lr=0.375996
Epoch[23] Batch [199]	Speed: 5653.292025 samples/sec	accuracy=0.467651	lr=0.375892
Epoch[23] Batch [249]	Speed: 5236.077446 samples/sec	accuracy=0.467234	lr=0.375788
Epoch[23] Batch [299]	Speed: 5650.598552 samples/sec	accuracy=0.466803	lr=0.375683
Epoch[23] Batch [349]	Speed: 5261.380638 samples/sec	accuracy=0.466543	lr=0.375579
Epoch[23] Batch [399]	Speed: 5678.471441 samples/sec	accuracy=0.466086	lr=0.375474
Epoch[23] Batch [449]	Speed: 5926.526213 samples/sec	accuracy=0.466157	lr=0.375369
Epoch[23] Batch [499]	Speed: 5770.474978 samples/sec	accuracy=0.465434	lr=0.375264
Epoch[23] Batch [549]	Speed: 5893.547662 samples/sec	accuracy=0.465156	lr=0.375159
Epoch[23] Batch [599]	Speed: 5293.556629 samples/sec	accuracy=0.464907	lr=0.375053
Epoch[23] Batch [649]	Speed: 5873.164787 samples/sec	accuracy=0.464803	lr=0.374948
Epoch[23] Batch [699]	Speed: 5910.632071 samples/sec	accuracy=0.464381	lr=0.374842
Epoch[23] Batch [749]	Speed: 5666.562635 samples/sec	accuracy=0.464340	lr=0.374736
Epoch[23] Batch [799]	Speed: 5633.766274 samples/sec	accuracy=0.464121	lr=0.374629
Epoch[23] Batch [849]	Speed: 5736.612263 samples/sec	accuracy=0.463943	lr=0.374523
Epoch[23] Batch [899]	Speed: 5728.336063 samples/sec	accuracy=0.464018	lr=0.374416
Epoch[23] Batch [949]	Speed: 5523.623988 samples/sec	accuracy=0.463451	lr=0.374309
Epoch[23] Batch [999]	Speed: 6055.993701 samples/sec	accuracy=0.463320	lr=0.374202
Epoch[23] Batch [1049]	Speed: 6024.825241 samples/sec	accuracy=0.463370	lr=0.374095
Epoch[23] Batch [1099]	Speed: 5537.334523 samples/sec	accuracy=0.463326	lr=0.373987
Epoch[23] Batch [1149]	Speed: 5461.000715 samples/sec	accuracy=0.463258	lr=0.373879
Epoch[23] Batch [1199]	Speed: 5271.800381 samples/sec	accuracy=0.463179	lr=0.373771
Epoch[23] Batch [1249]	Speed: 5431.031977 samples/sec	accuracy=0.463080	lr=0.373663
[Epoch 23] training: accuracy=0.463080
[Epoch 23] speed: 5637 samples/sec	time cost: 236.772501
[Epoch 23] validation: err-top1=0.543152 err-top5=0.282715
Epoch[24] Batch [49]	Speed: 5654.672555 samples/sec	accuracy=0.467832	lr=0.373552
Epoch[24] Batch [99]	Speed: 5553.882074 samples/sec	accuracy=0.468369	lr=0.373444
Epoch[24] Batch [149]	Speed: 5875.543811 samples/sec	accuracy=0.467969	lr=0.373335
Epoch[24] Batch [199]	Speed: 5478.157503 samples/sec	accuracy=0.468047	lr=0.373226
Epoch[24] Batch [249]	Speed: 5356.039542 samples/sec	accuracy=0.466832	lr=0.373117
Epoch[24] Batch [299]	Speed: 5659.205773 samples/sec	accuracy=0.466950	lr=0.373007
Epoch[24] Batch [349]	Speed: 5564.888986 samples/sec	accuracy=0.466660	lr=0.372897
Epoch[24] Batch [399]	Speed: 5688.311148 samples/sec	accuracy=0.466709	lr=0.372788
Epoch[24] Batch [449]	Speed: 5508.000109 samples/sec	accuracy=0.466500	lr=0.372678
Epoch[24] Batch [499]	Speed: 5811.352739 samples/sec	accuracy=0.466396	lr=0.372567
Epoch[24] Batch [549]	Speed: 5983.013206 samples/sec	accuracy=0.466415	lr=0.372457
Epoch[24] Batch [599]	Speed: 5512.906635 samples/sec	accuracy=0.466213	lr=0.372346
Epoch[24] Batch [649]	Speed: 5706.545608 samples/sec	accuracy=0.465944	lr=0.372235
Epoch[24] Batch [699]	Speed: 5798.848571 samples/sec	accuracy=0.466057	lr=0.372124
Epoch[24] Batch [749]	Speed: 5123.132286 samples/sec	accuracy=0.465948	lr=0.372013
Epoch[24] Batch [799]	Speed: 5676.989370 samples/sec	accuracy=0.465924	lr=0.371901
Epoch[24] Batch [849]	Speed: 6054.066874 samples/sec	accuracy=0.465588	lr=0.371790
Epoch[24] Batch [899]	Speed: 5663.274578 samples/sec	accuracy=0.465461	lr=0.371678
Epoch[24] Batch [949]	Speed: 6013.327955 samples/sec	accuracy=0.465330	lr=0.371565
Epoch[24] Batch [999]	Speed: 5678.597271 samples/sec	accuracy=0.465174	lr=0.371453
Epoch[24] Batch [1049]	Speed: 5810.418280 samples/sec	accuracy=0.465053	lr=0.371341
Epoch[24] Batch [1099]	Speed: 6029.610168 samples/sec	accuracy=0.465065	lr=0.371228
Epoch[24] Batch [1149]	Speed: 5586.467239 samples/sec	accuracy=0.464992	lr=0.371115
Epoch[24] Batch [1199]	Speed: 5356.842509 samples/sec	accuracy=0.465128	lr=0.371002
Epoch[24] Batch [1249]	Speed: 5727.512124 samples/sec	accuracy=0.464897	lr=0.370888
[Epoch 24] training: accuracy=0.464891
[Epoch 24] speed: 5659 samples/sec	time cost: 236.168701
[Epoch 24] validation: err-top1=0.540976 err-top5=0.277065
Epoch[25] Batch [49]	Speed: 5866.161043 samples/sec	accuracy=0.470742	lr=0.370773
Epoch[25] Batch [99]	Speed: 5680.904054 samples/sec	accuracy=0.473701	lr=0.370659
Epoch[25] Batch [149]	Speed: 5513.419426 samples/sec	accuracy=0.472891	lr=0.370545
Epoch[25] Batch [199]	Speed: 6044.014404 samples/sec	accuracy=0.472383	lr=0.370431
Epoch[25] Batch [249]	Speed: 5874.583775 samples/sec	accuracy=0.471496	lr=0.370316
Epoch[25] Batch [299]	Speed: 5418.806528 samples/sec	accuracy=0.470990	lr=0.370202
Epoch[25] Batch [349]	Speed: 5954.185522 samples/sec	accuracy=0.470798	lr=0.370087
Epoch[25] Batch [399]	Speed: 5608.139231 samples/sec	accuracy=0.470107	lr=0.369972
Epoch[25] Batch [449]	Speed: 5286.923454 samples/sec	accuracy=0.469911	lr=0.369857
Epoch[25] Batch [499]	Speed: 5741.127292 samples/sec	accuracy=0.469877	lr=0.369741
Epoch[25] Batch [549]	Speed: 5742.225223 samples/sec	accuracy=0.469574	lr=0.369626
Epoch[25] Batch [599]	Speed: 5768.550734 samples/sec	accuracy=0.469316	lr=0.369510
Epoch[25] Batch [649]	Speed: 5484.873531 samples/sec	accuracy=0.469037	lr=0.369394
Epoch[25] Batch [699]	Speed: 6005.779878 samples/sec	accuracy=0.468825	lr=0.369278
Epoch[25] Batch [749]	Speed: 5478.340436 samples/sec	accuracy=0.468667	lr=0.369161
Epoch[25] Batch [799]	Speed: 5718.940618 samples/sec	accuracy=0.468390	lr=0.369045
Epoch[25] Batch [849]	Speed: 5589.418924 samples/sec	accuracy=0.468273	lr=0.368928
Epoch[25] Batch [899]	Speed: 5740.916873 samples/sec	accuracy=0.468205	lr=0.368811
Epoch[25] Batch [949]	Speed: 5749.369714 samples/sec	accuracy=0.467986	lr=0.368694
Epoch[25] Batch [999]	Speed: 5849.128296 samples/sec	accuracy=0.467947	lr=0.368576
Epoch[25] Batch [1049]	Speed: 5684.428534 samples/sec	accuracy=0.467758	lr=0.368459
Epoch[25] Batch [1099]	Speed: 5595.619900 samples/sec	accuracy=0.467690	lr=0.368341
Epoch[25] Batch [1149]	Speed: 5490.808644 samples/sec	accuracy=0.467566	lr=0.368223
Epoch[25] Batch [1199]	Speed: 5750.442929 samples/sec	accuracy=0.467349	lr=0.368105
Epoch[25] Batch [1249]	Speed: 5868.999351 samples/sec	accuracy=0.467188	lr=0.367986
[Epoch 25] training: accuracy=0.467204
[Epoch 25] speed: 5686 samples/sec	time cost: 234.606881
[Epoch 25] validation: err-top1=0.524055 err-top5=0.262436
Epoch[26] Batch [49]	Speed: 5264.967800 samples/sec	accuracy=0.473594	lr=0.367865
Epoch[26] Batch [99]	Speed: 5560.852132 samples/sec	accuracy=0.473994	lr=0.367747
Epoch[26] Batch [149]	Speed: 5520.313773 samples/sec	accuracy=0.473281	lr=0.367628
Epoch[26] Batch [199]	Speed: 5316.850027 samples/sec	accuracy=0.472725	lr=0.367508
Epoch[26] Batch [249]	Speed: 5674.192764 samples/sec	accuracy=0.472902	lr=0.367389
Epoch[26] Batch [299]	Speed: 5832.668598 samples/sec	accuracy=0.472126	lr=0.367269
Epoch[26] Batch [349]	Speed: 5697.292337 samples/sec	accuracy=0.471992	lr=0.367150
Epoch[26] Batch [399]	Speed: 5908.578281 samples/sec	accuracy=0.471497	lr=0.367030
Epoch[26] Batch [449]	Speed: 5495.627758 samples/sec	accuracy=0.471198	lr=0.366909
Epoch[26] Batch [499]	Speed: 5598.992752 samples/sec	accuracy=0.471170	lr=0.366789
Epoch[26] Batch [549]	Speed: 5869.571225 samples/sec	accuracy=0.470900	lr=0.366668
Epoch[26] Batch [599]	Speed: 5836.040732 samples/sec	accuracy=0.470741	lr=0.366548
Epoch[26] Batch [649]	Speed: 5619.780543 samples/sec	accuracy=0.470209	lr=0.366426
Epoch[26] Batch [699]	Speed: 5837.899973 samples/sec	accuracy=0.470123	lr=0.366305
Epoch[26] Batch [749]	Speed: 5673.087421 samples/sec	accuracy=0.469931	lr=0.366184
Epoch[26] Batch [799]	Speed: 5575.474476 samples/sec	accuracy=0.469723	lr=0.366062
Epoch[26] Batch [849]	Speed: 5325.884350 samples/sec	accuracy=0.469660	lr=0.365940
Epoch[26] Batch [899]	Speed: 5890.934741 samples/sec	accuracy=0.469361	lr=0.365819
Epoch[26] Batch [949]	Speed: 5856.680730 samples/sec	accuracy=0.469221	lr=0.365696
Epoch[26] Batch [999]	Speed: 5823.307040 samples/sec	accuracy=0.468980	lr=0.365574
Epoch[26] Batch [1049]	Speed: 5888.821623 samples/sec	accuracy=0.468876	lr=0.365451
Epoch[26] Batch [1099]	Speed: 5498.119457 samples/sec	accuracy=0.468580	lr=0.365329
Epoch[26] Batch [1149]	Speed: 5723.008257 samples/sec	accuracy=0.468207	lr=0.365206
Epoch[26] Batch [1199]	Speed: 5712.900218 samples/sec	accuracy=0.468069	lr=0.365082
Epoch[26] Batch [1249]	Speed: 5637.162199 samples/sec	accuracy=0.467988	lr=0.364959
[Epoch 26] training: accuracy=0.467983
[Epoch 26] speed: 5654 samples/sec	time cost: 236.091767
[Epoch 26] validation: err-top1=0.523717 err-top5=0.266761
Epoch[27] Batch [49]	Speed: 5495.266762 samples/sec	accuracy=0.479941	lr=0.364833
Epoch[27] Batch [99]	Speed: 5605.754913 samples/sec	accuracy=0.479141	lr=0.364709
Epoch[27] Batch [149]	Speed: 5697.029198 samples/sec	accuracy=0.475482	lr=0.364585
Epoch[27] Batch [199]	Speed: 5576.323304 samples/sec	accuracy=0.475361	lr=0.364461
Epoch[27] Batch [249]	Speed: 5852.958467 samples/sec	accuracy=0.474445	lr=0.364337
Epoch[27] Batch [299]	Speed: 5521.240710 samples/sec	accuracy=0.473587	lr=0.364212
Epoch[27] Batch [349]	Speed: 5863.373825 samples/sec	accuracy=0.473343	lr=0.364087
Epoch[27] Batch [399]	Speed: 5616.355433 samples/sec	accuracy=0.472896	lr=0.363962
Epoch[27] Batch [449]	Speed: 5901.350726 samples/sec	accuracy=0.472858	lr=0.363837
Epoch[27] Batch [499]	Speed: 5313.869587 samples/sec	accuracy=0.472727	lr=0.363712
Epoch[27] Batch [549]	Speed: 5802.660349 samples/sec	accuracy=0.472818	lr=0.363586
Epoch[27] Batch [599]	Speed: 5883.587286 samples/sec	accuracy=0.472251	lr=0.363461
Epoch[27] Batch [649]	Speed: 5619.372909 samples/sec	accuracy=0.472021	lr=0.363335
Epoch[27] Batch [699]	Speed: 5785.691266 samples/sec	accuracy=0.471927	lr=0.363209
Epoch[27] Batch [749]	Speed: 5719.213553 samples/sec	accuracy=0.471730	lr=0.363082
Epoch[27] Batch [799]	Speed: 5524.284859 samples/sec	accuracy=0.471324	lr=0.362956
Epoch[27] Batch [849]	Speed: 5618.654987 samples/sec	accuracy=0.471019	lr=0.362829
Epoch[27] Batch [899]	Speed: 5895.511383 samples/sec	accuracy=0.470681	lr=0.362702
Epoch[27] Batch [949]	Speed: 5544.048239 samples/sec	accuracy=0.470547	lr=0.362575
Epoch[27] Batch [999]	Speed: 5735.931177 samples/sec	accuracy=0.470484	lr=0.362448
Epoch[27] Batch [1049]	Speed: 5721.582119 samples/sec	accuracy=0.470521	lr=0.362320
Epoch[27] Batch [1099]	Speed: 5198.336878 samples/sec	accuracy=0.470373	lr=0.362193
Epoch[27] Batch [1149]	Speed: 5597.442740 samples/sec	accuracy=0.470200	lr=0.362065
Epoch[27] Batch [1199]	Speed: 5792.931895 samples/sec	accuracy=0.470069	lr=0.361937
Epoch[27] Batch [1249]	Speed: 5624.624471 samples/sec	accuracy=0.470097	lr=0.361809
[Epoch 27] training: accuracy=0.470102
[Epoch 27] speed: 5649 samples/sec	time cost: 236.044398
[Epoch 27] validation: err-top1=0.532685 err-top5=0.269591
Epoch[28] Batch [49]	Speed: 5394.450223 samples/sec	accuracy=0.480684	lr=0.361678
Epoch[28] Batch [99]	Speed: 5837.769523 samples/sec	accuracy=0.477070	lr=0.361549
Epoch[28] Batch [149]	Speed: 5630.082609 samples/sec	accuracy=0.476061	lr=0.361420
Epoch[28] Batch [199]	Speed: 5691.784785 samples/sec	accuracy=0.475166	lr=0.361291
Epoch[28] Batch [249]	Speed: 5605.304540 samples/sec	accuracy=0.475191	lr=0.361162
Epoch[28] Batch [299]	Speed: 5779.697639 samples/sec	accuracy=0.474740	lr=0.361032
Epoch[28] Batch [349]	Speed: 5410.917387 samples/sec	accuracy=0.474336	lr=0.360903
Epoch[28] Batch [399]	Speed: 5671.240000 samples/sec	accuracy=0.474067	lr=0.360773
Epoch[28] Batch [449]	Speed: 5808.355137 samples/sec	accuracy=0.474234	lr=0.360643
Epoch[28] Batch [499]	Speed: 5358.020405 samples/sec	accuracy=0.474029	lr=0.360513
Epoch[28] Batch [549]	Speed: 5755.042536 samples/sec	accuracy=0.473681	lr=0.360382
Epoch[28] Batch [599]	Speed: 5915.591505 samples/sec	accuracy=0.473597	lr=0.360252
Epoch[28] Batch [649]	Speed: 5483.074400 samples/sec	accuracy=0.473436	lr=0.360121
Epoch[28] Batch [699]	Speed: 5303.945165 samples/sec	accuracy=0.473288	lr=0.359990
Epoch[28] Batch [749]	Speed: 5431.996496 samples/sec	accuracy=0.473173	lr=0.359859
Epoch[28] Batch [799]	Speed: 5895.119084 samples/sec	accuracy=0.472753	lr=0.359728
Epoch[28] Batch [849]	Speed: 5437.284561 samples/sec	accuracy=0.472815	lr=0.359596
Epoch[28] Batch [899]	Speed: 5586.408091 samples/sec	accuracy=0.472716	lr=0.359465
Epoch[28] Batch [949]	Speed: 5727.517776 samples/sec	accuracy=0.472595	lr=0.359333
Epoch[28] Batch [999]	Speed: 5709.849807 samples/sec	accuracy=0.472539	lr=0.359201
Epoch[28] Batch [1049]	Speed: 5877.401930 samples/sec	accuracy=0.472520	lr=0.359068
Epoch[28] Batch [1099]	Speed: 5705.020204 samples/sec	accuracy=0.472304	lr=0.358936
Epoch[28] Batch [1149]	Speed: 5629.788596 samples/sec	accuracy=0.472105	lr=0.358803
Epoch[28] Batch [1199]	Speed: 6052.076800 samples/sec	accuracy=0.472003	lr=0.358670
Epoch[28] Batch [1249]	Speed: 5590.087194 samples/sec	accuracy=0.471833	lr=0.358537
[Epoch 28] training: accuracy=0.471867
[Epoch 28] speed: 5638 samples/sec	time cost: 236.487271
[Epoch 28] validation: err-top1=0.509786 err-top5=0.252531
Epoch[29] Batch [49]	Speed: 5125.783993 samples/sec	accuracy=0.476738	lr=0.358401
Epoch[29] Batch [99]	Speed: 5493.251711 samples/sec	accuracy=0.478086	lr=0.358268
Epoch[29] Batch [149]	Speed: 5880.678007 samples/sec	accuracy=0.476673	lr=0.358134
Epoch[29] Batch [199]	Speed: 5849.064890 samples/sec	accuracy=0.475435	lr=0.358001
Epoch[29] Batch [249]	Speed: 5915.419919 samples/sec	accuracy=0.475145	lr=0.357867
Epoch[29] Batch [299]	Speed: 5887.332642 samples/sec	accuracy=0.475527	lr=0.357732
Epoch[29] Batch [349]	Speed: 5659.336419 samples/sec	accuracy=0.475402	lr=0.357598
Epoch[29] Batch [399]	Speed: 5363.325414 samples/sec	accuracy=0.475215	lr=0.357464
Epoch[29] Batch [449]	Speed: 5662.886892 samples/sec	accuracy=0.474809	lr=0.357329
Epoch[29] Batch [499]	Speed: 5977.688364 samples/sec	accuracy=0.474623	lr=0.357194
Epoch[29] Batch [549]	Speed: 5617.899919 samples/sec	accuracy=0.474600	lr=0.357059
Epoch[29] Batch [599]	Speed: 5452.608544 samples/sec	accuracy=0.474476	lr=0.356924
Epoch[29] Batch [649]	Speed: 5421.690010 samples/sec	accuracy=0.474292	lr=0.356788
Epoch[29] Batch [699]	Speed: 5859.022598 samples/sec	accuracy=0.474227	lr=0.356652
Epoch[29] Batch [749]	Speed: 5382.531326 samples/sec	accuracy=0.474187	lr=0.356517
Epoch[29] Batch [799]	Speed: 5377.994766 samples/sec	accuracy=0.474110	lr=0.356380
Epoch[29] Batch [849]	Speed: 5582.621789 samples/sec	accuracy=0.474080	lr=0.356244
Epoch[29] Batch [899]	Speed: 5525.239287 samples/sec	accuracy=0.473872	lr=0.356108
Epoch[29] Batch [949]	Speed: 5458.248133 samples/sec	accuracy=0.473664	lr=0.355971
Epoch[29] Batch [999]	Speed: 5458.555441 samples/sec	accuracy=0.473485	lr=0.355834
Epoch[29] Batch [1049]	Speed: 5950.166039 samples/sec	accuracy=0.473285	lr=0.355698
Epoch[29] Batch [1099]	Speed: 5729.179650 samples/sec	accuracy=0.473139	lr=0.355560
Epoch[29] Batch [1149]	Speed: 5767.222617 samples/sec	accuracy=0.472909	lr=0.355423
Epoch[29] Batch [1199]	Speed: 5614.809593 samples/sec	accuracy=0.472892	lr=0.355285
Epoch[29] Batch [1249]	Speed: 5815.180349 samples/sec	accuracy=0.472874	lr=0.355148
[Epoch 29] training: accuracy=0.472870
[Epoch 29] speed: 5617 samples/sec	time cost: 237.541883
[Epoch 29] validation: err-top1=0.497498 err-top5=0.242615
Epoch[30] Batch [49]	Speed: 5517.321626 samples/sec	accuracy=0.478398	lr=0.355007
Epoch[30] Batch [99]	Speed: 5435.394205 samples/sec	accuracy=0.478945	lr=0.354869
Epoch[30] Batch [149]	Speed: 5638.110885 samples/sec	accuracy=0.478229	lr=0.354731
Epoch[30] Batch [199]	Speed: 5376.061825 samples/sec	accuracy=0.477690	lr=0.354592
Epoch[30] Batch [249]	Speed: 5820.101181 samples/sec	accuracy=0.476789	lr=0.354454
Epoch[30] Batch [299]	Speed: 5926.405346 samples/sec	accuracy=0.477907	lr=0.354315
Epoch[30] Batch [349]	Speed: 5581.924978 samples/sec	accuracy=0.477891	lr=0.354176
Epoch[30] Batch [399]	Speed: 5536.855390 samples/sec	accuracy=0.477568	lr=0.354037
Epoch[30] Batch [449]	Speed: 5946.559333 samples/sec	accuracy=0.477142	lr=0.353897
Epoch[30] Batch [499]	Speed: 5919.258462 samples/sec	accuracy=0.477193	lr=0.353758
Epoch[30] Batch [549]	Speed: 5606.713110 samples/sec	accuracy=0.477111	lr=0.353618
Epoch[30] Batch [599]	Speed: 5578.251682 samples/sec	accuracy=0.476880	lr=0.353478
Epoch[30] Batch [649]	Speed: 5627.781952 samples/sec	accuracy=0.476573	lr=0.353338
Epoch[30] Batch [699]	Speed: 5683.314237 samples/sec	accuracy=0.476341	lr=0.353198
Epoch[30] Batch [749]	Speed: 5702.441062 samples/sec	accuracy=0.475707	lr=0.353057
Epoch[30] Batch [799]	Speed: 5929.409499 samples/sec	accuracy=0.475546	lr=0.352917
Epoch[30] Batch [849]	Speed: 5968.492147 samples/sec	accuracy=0.475335	lr=0.352776
Epoch[30] Batch [899]	Speed: 5777.333413 samples/sec	accuracy=0.475378	lr=0.352635
Epoch[30] Batch [949]	Speed: 5906.517958 samples/sec	accuracy=0.475247	lr=0.352493
Epoch[30] Batch [999]	Speed: 5414.524798 samples/sec	accuracy=0.475095	lr=0.352352
Epoch[30] Batch [1049]	Speed: 5426.100990 samples/sec	accuracy=0.474930	lr=0.352211
Epoch[30] Batch [1099]	Speed: 5701.187706 samples/sec	accuracy=0.474907	lr=0.352069
Epoch[30] Batch [1149]	Speed: 5738.723351 samples/sec	accuracy=0.474676	lr=0.351927
Epoch[30] Batch [1199]	Speed: 5823.599030 samples/sec	accuracy=0.474665	lr=0.351785
Epoch[30] Batch [1249]	Speed: 5626.484393 samples/sec	accuracy=0.474507	lr=0.351642
[Epoch 30] training: accuracy=0.474517
[Epoch 30] speed: 5674 samples/sec	time cost: 235.341228
[Epoch 30] validation: err-top1=0.500817 err-top5=0.247290
Epoch[31] Batch [49]	Speed: 5669.810651 samples/sec	accuracy=0.483594	lr=0.351497
Epoch[31] Batch [99]	Speed: 5560.246547 samples/sec	accuracy=0.481738	lr=0.351354
Epoch[31] Batch [149]	Speed: 5178.887112 samples/sec	accuracy=0.480286	lr=0.351212
Epoch[31] Batch [199]	Speed: 5926.004020 samples/sec	accuracy=0.479648	lr=0.351069
Epoch[31] Batch [249]	Speed: 5834.920112 samples/sec	accuracy=0.479723	lr=0.350925
Epoch[31] Batch [299]	Speed: 5475.969257 samples/sec	accuracy=0.478708	lr=0.350782
Epoch[31] Batch [349]	Speed: 6135.944384 samples/sec	accuracy=0.478571	lr=0.350639
Epoch[31] Batch [399]	Speed: 6151.123934 samples/sec	accuracy=0.478533	lr=0.350495
Epoch[31] Batch [449]	Speed: 6039.046106 samples/sec	accuracy=0.478234	lr=0.350351
Epoch[31] Batch [499]	Speed: 5852.550279 samples/sec	accuracy=0.477457	lr=0.350207
Epoch[31] Batch [549]	Speed: 5793.051910 samples/sec	accuracy=0.477388	lr=0.350062
Epoch[31] Batch [599]	Speed: 5863.726206 samples/sec	accuracy=0.477184	lr=0.349918
Epoch[31] Batch [649]	Speed: 5965.367902 samples/sec	accuracy=0.477258	lr=0.349773
Epoch[31] Batch [699]	Speed: 6008.418027 samples/sec	accuracy=0.477306	lr=0.349629
Epoch[31] Batch [749]	Speed: 5820.134148 samples/sec	accuracy=0.477057	lr=0.349484
Epoch[31] Batch [799]	Speed: 5958.475444 samples/sec	accuracy=0.476877	lr=0.349338
Epoch[31] Batch [849]	Speed: 5975.896675 samples/sec	accuracy=0.476943	lr=0.349193
Epoch[31] Batch [899]	Speed: 5730.358950 samples/sec	accuracy=0.476849	lr=0.349048
Epoch[31] Batch [949]	Speed: 5577.192524 samples/sec	accuracy=0.476475	lr=0.348902
Epoch[31] Batch [999]	Speed: 5492.334147 samples/sec	accuracy=0.476451	lr=0.348756
Epoch[31] Batch [1049]	Speed: 5832.135253 samples/sec	accuracy=0.476202	lr=0.348610
Epoch[31] Batch [1099]	Speed: 5387.378271 samples/sec	accuracy=0.476098	lr=0.348464
Epoch[31] Batch [1149]	Speed: 5635.282939 samples/sec	accuracy=0.476111	lr=0.348317
Epoch[31] Batch [1199]	Speed: 5665.171959 samples/sec	accuracy=0.476013	lr=0.348171
Epoch[31] Batch [1249]	Speed: 5664.643553 samples/sec	accuracy=0.475781	lr=0.348024
[Epoch 31] training: accuracy=0.475792
[Epoch 31] speed: 5752 samples/sec	time cost: 232.130013
[Epoch 31] validation: err-top1=0.486169 err-top5=0.235092
Epoch[32] Batch [49]	Speed: 5131.104721 samples/sec	accuracy=0.480820	lr=0.347874
Epoch[32] Batch [99]	Speed: 5924.050659 samples/sec	accuracy=0.481260	lr=0.347727
Epoch[32] Batch [149]	Speed: 5786.360833 samples/sec	accuracy=0.480241	lr=0.347580
Epoch[32] Batch [199]	Speed: 5585.476869 samples/sec	accuracy=0.480225	lr=0.347432
Epoch[32] Batch [249]	Speed: 5957.156272 samples/sec	accuracy=0.480258	lr=0.347285
Epoch[32] Batch [299]	Speed: 6142.751562 samples/sec	accuracy=0.480254	lr=0.347137
Epoch[32] Batch [349]	Speed: 5951.660750 samples/sec	accuracy=0.479997	lr=0.346989
Epoch[32] Batch [399]	Speed: 5779.557178 samples/sec	accuracy=0.479939	lr=0.346841
Epoch[32] Batch [449]	Speed: 5750.103109 samples/sec	accuracy=0.479447	lr=0.346692
Epoch[32] Batch [499]	Speed: 5580.249979 samples/sec	accuracy=0.479418	lr=0.346544
Epoch[32] Batch [549]	Speed: 5430.362606 samples/sec	accuracy=0.479384	lr=0.346395
Epoch[32] Batch [599]	Speed: 6226.598497 samples/sec	accuracy=0.479230	lr=0.346246
Epoch[32] Batch [649]	Speed: 5681.002490 samples/sec	accuracy=0.478924	lr=0.346097
Epoch[32] Batch [699]	Speed: 5595.128733 samples/sec	accuracy=0.478617	lr=0.345948
Epoch[32] Batch [749]	Speed: 5592.299895 samples/sec	accuracy=0.478341	lr=0.345799
Epoch[32] Batch [799]	Speed: 5574.409137 samples/sec	accuracy=0.478125	lr=0.345649
Epoch[32] Batch [849]	Speed: 5831.788401 samples/sec	accuracy=0.477930	lr=0.345499
Epoch[32] Batch [899]	Speed: 5843.343457 samples/sec	accuracy=0.477747	lr=0.345349
Epoch[32] Batch [949]	Speed: 5787.177151 samples/sec	accuracy=0.477746	lr=0.345199
Epoch[32] Batch [999]	Speed: 5494.362723 samples/sec	accuracy=0.477821	lr=0.345049
Epoch[32] Batch [1049]	Speed: 5812.167472 samples/sec	accuracy=0.477617	lr=0.344899
Epoch[32] Batch [1099]	Speed: 5630.080837 samples/sec	accuracy=0.477673	lr=0.344748
Epoch[32] Batch [1149]	Speed: 5573.893186 samples/sec	accuracy=0.477717	lr=0.344597
Epoch[32] Batch [1199]	Speed: 5352.620081 samples/sec	accuracy=0.477700	lr=0.344446
Epoch[32] Batch [1249]	Speed: 5457.768162 samples/sec	accuracy=0.477495	lr=0.344295
[Epoch 32] training: accuracy=0.477505
[Epoch 32] speed: 5681 samples/sec	time cost: 234.853031
[Epoch 32] validation: err-top1=0.511400 err-top5=0.258191
Epoch[33] Batch [49]	Speed: 5918.234178 samples/sec	accuracy=0.487148	lr=0.344141
Epoch[33] Batch [99]	Speed: 5621.315731 samples/sec	accuracy=0.485352	lr=0.343989
Epoch[33] Batch [149]	Speed: 5341.617770 samples/sec	accuracy=0.483971	lr=0.343838
Epoch[33] Batch [199]	Speed: 5408.433251 samples/sec	accuracy=0.483130	lr=0.343686
Epoch[33] Batch [249]	Speed: 5802.820909 samples/sec	accuracy=0.482687	lr=0.343534
Epoch[33] Batch [299]	Speed: 5618.447863 samples/sec	accuracy=0.482194	lr=0.343382
Epoch[33] Batch [349]	Speed: 5655.245717 samples/sec	accuracy=0.481560	lr=0.343229
Epoch[33] Batch [399]	Speed: 5868.682583 samples/sec	accuracy=0.481365	lr=0.343077
Epoch[33] Batch [449]	Speed: 5841.486627 samples/sec	accuracy=0.481239	lr=0.342924
Epoch[33] Batch [499]	Speed: 5731.021429 samples/sec	accuracy=0.481029	lr=0.342771
Epoch[33] Batch [549]	Speed: 5571.068142 samples/sec	accuracy=0.480904	lr=0.342618
Epoch[33] Batch [599]	Speed: 5917.436074 samples/sec	accuracy=0.480532	lr=0.342465
Epoch[33] Batch [649]	Speed: 5385.339992 samples/sec	accuracy=0.480128	lr=0.342312
Epoch[33] Batch [699]	Speed: 5331.300186 samples/sec	accuracy=0.480059	lr=0.342158
Epoch[33] Batch [749]	Speed: 5763.574490 samples/sec	accuracy=0.479591	lr=0.342005
Epoch[33] Batch [799]	Speed: 5749.617699 samples/sec	accuracy=0.479664	lr=0.341851
Epoch[33] Batch [849]	Speed: 5695.826411 samples/sec	accuracy=0.479555	lr=0.341697
Epoch[33] Batch [899]	Speed: 5619.151470 samples/sec	accuracy=0.479487	lr=0.341543
Epoch[33] Batch [949]	Speed: 5513.734678 samples/sec	accuracy=0.479106	lr=0.341388
Epoch[33] Batch [999]	Speed: 5742.258082 samples/sec	accuracy=0.479193	lr=0.341234
Epoch[33] Batch [1049]	Speed: 5386.209048 samples/sec	accuracy=0.479080	lr=0.341079
Epoch[33] Batch [1099]	Speed: 5882.928873 samples/sec	accuracy=0.478985	lr=0.340924
Epoch[33] Batch [1149]	Speed: 5700.354157 samples/sec	accuracy=0.478975	lr=0.340769
Epoch[33] Batch [1199]	Speed: 5609.528419 samples/sec	accuracy=0.478876	lr=0.340614
Epoch[33] Batch [1249]	Speed: 6015.034839 samples/sec	accuracy=0.478830	lr=0.340458
[Epoch 33] training: accuracy=0.478829
[Epoch 33] speed: 5653 samples/sec	time cost: 235.716558
[Epoch 33] validation: err-top1=0.494579 err-top5=0.238341
Epoch[34] Batch [49]	Speed: 5593.769834 samples/sec	accuracy=0.489902	lr=0.340300
Epoch[34] Batch [99]	Speed: 5619.615982 samples/sec	accuracy=0.488389	lr=0.340144
Epoch[34] Batch [149]	Speed: 5768.658584 samples/sec	accuracy=0.487324	lr=0.339988
Epoch[34] Batch [199]	Speed: 5207.168562 samples/sec	accuracy=0.486709	lr=0.339832
Epoch[34] Batch [249]	Speed: 5926.381304 samples/sec	accuracy=0.485582	lr=0.339676
Epoch[34] Batch [299]	Speed: 5727.416652 samples/sec	accuracy=0.485202	lr=0.339520
Epoch[34] Batch [349]	Speed: 5651.149325 samples/sec	accuracy=0.484464	lr=0.339363
Epoch[34] Batch [399]	Speed: 5807.324115 samples/sec	accuracy=0.483999	lr=0.339206
Epoch[34] Batch [449]	Speed: 5838.025192 samples/sec	accuracy=0.483533	lr=0.339049
Epoch[34] Batch [499]	Speed: 5799.637560 samples/sec	accuracy=0.483266	lr=0.338892
Epoch[34] Batch [549]	Speed: 5585.004766 samples/sec	accuracy=0.483113	lr=0.338735
Epoch[34] Batch [599]	Speed: 5602.501362 samples/sec	accuracy=0.482798	lr=0.338578
Epoch[34] Batch [649]	Speed: 5863.564499 samples/sec	accuracy=0.482700	lr=0.338420
Epoch[34] Batch [699]	Speed: 6145.688198 samples/sec	accuracy=0.482287	lr=0.338263
Epoch[34] Batch [749]	Speed: 5921.545006 samples/sec	accuracy=0.482043	lr=0.338105
Epoch[34] Batch [799]	Speed: 5670.095535 samples/sec	accuracy=0.481847	lr=0.337947
Epoch[34] Batch [849]	Speed: 5798.025044 samples/sec	accuracy=0.481418	lr=0.337788
Epoch[34] Batch [899]	Speed: 5838.391992 samples/sec	accuracy=0.481200	lr=0.337630
Epoch[34] Batch [949]	Speed: 5679.311068 samples/sec	accuracy=0.481015	lr=0.337472
Epoch[34] Batch [999]	Speed: 5460.704795 samples/sec	accuracy=0.480957	lr=0.337313
Epoch[34] Batch [1049]	Speed: 5542.851376 samples/sec	accuracy=0.480951	lr=0.337154
Epoch[34] Batch [1099]	Speed: 5589.518434 samples/sec	accuracy=0.480768	lr=0.336995
Epoch[34] Batch [1149]	Speed: 5669.744038 samples/sec	accuracy=0.480655	lr=0.336836
Epoch[34] Batch [1199]	Speed: 5781.439432 samples/sec	accuracy=0.480428	lr=0.336677
Epoch[34] Batch [1249]	Speed: 5567.619278 samples/sec	accuracy=0.480344	lr=0.336517
[Epoch 34] training: accuracy=0.480334
[Epoch 34] speed: 5696 samples/sec	time cost: 233.650012
[Epoch 34] validation: err-top1=0.497274 err-top5=0.244893
Epoch[35] Batch [49]	Speed: 5546.598375 samples/sec	accuracy=0.488203	lr=0.336354
Epoch[35] Batch [99]	Speed: 5772.083219 samples/sec	accuracy=0.486855	lr=0.336194
Epoch[35] Batch [149]	Speed: 5946.645619 samples/sec	accuracy=0.485898	lr=0.336034
Epoch[35] Batch [199]	Speed: 5827.577528 samples/sec	accuracy=0.486094	lr=0.335874
Epoch[35] Batch [249]	Speed: 5566.920147 samples/sec	accuracy=0.485645	lr=0.335714
Epoch[35] Batch [299]	Speed: 5809.733860 samples/sec	accuracy=0.484993	lr=0.335553
Epoch[35] Batch [349]	Speed: 5787.813524 samples/sec	accuracy=0.485226	lr=0.335393
Epoch[35] Batch [399]	Speed: 5577.405743 samples/sec	accuracy=0.484775	lr=0.335232
Epoch[35] Batch [449]	Speed: 5227.723940 samples/sec	accuracy=0.484388	lr=0.335071
Epoch[35] Batch [499]	Speed: 5895.445672 samples/sec	accuracy=0.484236	lr=0.334910
Epoch[35] Batch [549]	Speed: 5538.950137 samples/sec	accuracy=0.484217	lr=0.334749
Epoch[35] Batch [599]	Speed: 5889.156235 samples/sec	accuracy=0.484093	lr=0.334587
Epoch[35] Batch [649]	Speed: 5897.038673 samples/sec	accuracy=0.483516	lr=0.334425
Epoch[35] Batch [699]	Speed: 5401.134969 samples/sec	accuracy=0.482985	lr=0.334264
Epoch[35] Batch [749]	Speed: 5888.856665 samples/sec	accuracy=0.482986	lr=0.334102
Epoch[35] Batch [799]	Speed: 5357.362761 samples/sec	accuracy=0.482711	lr=0.333940
Epoch[35] Batch [849]	Speed: 5640.475996 samples/sec	accuracy=0.482738	lr=0.333777
Epoch[35] Batch [899]	Speed: 5648.291500 samples/sec	accuracy=0.482589	lr=0.333615
Epoch[35] Batch [949]	Speed: 5867.682300 samples/sec	accuracy=0.482741	lr=0.333452
Epoch[35] Batch [999]	Speed: 5649.105434 samples/sec	accuracy=0.482552	lr=0.333290
Epoch[35] Batch [1049]	Speed: 5834.214059 samples/sec	accuracy=0.482537	lr=0.333127
Epoch[35] Batch [1099]	Speed: 5893.627726 samples/sec	accuracy=0.482244	lr=0.332964
Epoch[35] Batch [1149]	Speed: 5525.186831 samples/sec	accuracy=0.482227	lr=0.332801
Epoch[35] Batch [1199]	Speed: 5687.900592 samples/sec	accuracy=0.482057	lr=0.332637
Epoch[35] Batch [1249]	Speed: 5334.456442 samples/sec	accuracy=0.482179	lr=0.332474
[Epoch 35] training: accuracy=0.482179
[Epoch 35] speed: 5666 samples/sec	time cost: 235.702032
[Epoch 35] validation: err-top1=0.524713 err-top5=0.264210
Epoch[36] Batch [49]	Speed: 5101.622960 samples/sec	accuracy=0.488867	lr=0.332307
Epoch[36] Batch [99]	Speed: 5820.220116 samples/sec	accuracy=0.489082	lr=0.332143
Epoch[36] Batch [149]	Speed: 5916.767945 samples/sec	accuracy=0.488509	lr=0.331979
Epoch[36] Batch [199]	Speed: 5951.688297 samples/sec	accuracy=0.487686	lr=0.331815
Epoch[36] Batch [249]	Speed: 5630.363809 samples/sec	accuracy=0.486609	lr=0.331650
Epoch[36] Batch [299]	Speed: 5986.888928 samples/sec	accuracy=0.486156	lr=0.331486
Epoch[36] Batch [349]	Speed: 6010.122629 samples/sec	accuracy=0.486127	lr=0.331321
Epoch[36] Batch [399]	Speed: 5981.328109 samples/sec	accuracy=0.486377	lr=0.331157
Epoch[36] Batch [449]	Speed: 6096.042938 samples/sec	accuracy=0.485829	lr=0.330992
Epoch[36] Batch [499]	Speed: 5551.779034 samples/sec	accuracy=0.485994	lr=0.330827
Epoch[36] Batch [549]	Speed: 5589.859036 samples/sec	accuracy=0.485872	lr=0.330661
Epoch[36] Batch [599]	Speed: 5598.806636 samples/sec	accuracy=0.485425	lr=0.330496
Epoch[36] Batch [649]	Speed: 5509.408250 samples/sec	accuracy=0.484962	lr=0.330330
Epoch[36] Batch [699]	Speed: 5340.814314 samples/sec	accuracy=0.484675	lr=0.330165
Epoch[36] Batch [749]	Speed: 5997.014199 samples/sec	accuracy=0.484154	lr=0.329999
Epoch[36] Batch [799]	Speed: 5733.998819 samples/sec	accuracy=0.483956	lr=0.329833
Epoch[36] Batch [849]	Speed: 5792.046466 samples/sec	accuracy=0.483850	lr=0.329667
Epoch[36] Batch [899]	Speed: 5388.181604 samples/sec	accuracy=0.483719	lr=0.329500
Epoch[36] Batch [949]	Speed: 5487.387175 samples/sec	accuracy=0.483726	lr=0.329334
Epoch[36] Batch [999]	Speed: 5756.499601 samples/sec	accuracy=0.483604	lr=0.329167
Epoch[36] Batch [1049]	Speed: 5770.475753 samples/sec	accuracy=0.483450	lr=0.329000
Epoch[36] Batch [1099]	Speed: 5559.078801 samples/sec	accuracy=0.483382	lr=0.328833
Epoch[36] Batch [1149]	Speed: 5466.395660 samples/sec	accuracy=0.483401	lr=0.328666
Epoch[36] Batch [1199]	Speed: 5485.173757 samples/sec	accuracy=0.483290	lr=0.328499
Epoch[36] Batch [1249]	Speed: 5661.051629 samples/sec	accuracy=0.483261	lr=0.328332
[Epoch 36] training: accuracy=0.483263
[Epoch 36] speed: 5670 samples/sec	time cost: 235.353225
[Epoch 36] validation: err-top1=0.519471 err-top5=0.262197
Epoch[37] Batch [49]	Speed: 5860.292585 samples/sec	accuracy=0.496250	lr=0.328161
Epoch[37] Batch [99]	Speed: 5896.248864 samples/sec	accuracy=0.494395	lr=0.327993
Epoch[37] Batch [149]	Speed: 5963.911349 samples/sec	accuracy=0.492467	lr=0.327825
Epoch[37] Batch [199]	Speed: 5591.259999 samples/sec	accuracy=0.491396	lr=0.327657
Epoch[37] Batch [249]	Speed: 5504.397472 samples/sec	accuracy=0.490871	lr=0.327489
Epoch[37] Batch [299]	Speed: 5454.890388 samples/sec	accuracy=0.489701	lr=0.327321
Epoch[37] Batch [349]	Speed: 5736.109822 samples/sec	accuracy=0.488965	lr=0.327152
Epoch[37] Batch [399]	Speed: 5480.762347 samples/sec	accuracy=0.488584	lr=0.326983
Epoch[37] Batch [449]	Speed: 5452.779391 samples/sec	accuracy=0.488073	lr=0.326815
Epoch[37] Batch [499]	Speed: 5473.062916 samples/sec	accuracy=0.487496	lr=0.326646
Epoch[37] Batch [549]	Speed: 5542.365423 samples/sec	accuracy=0.487044	lr=0.326477
Epoch[37] Batch [599]	Speed: 5520.973284 samples/sec	accuracy=0.486867	lr=0.326307
Epoch[37] Batch [649]	Speed: 5452.785068 samples/sec	accuracy=0.486717	lr=0.326138
Epoch[37] Batch [699]	Speed: 5712.009760 samples/sec	accuracy=0.486323	lr=0.325968
Epoch[37] Batch [749]	Speed: 5648.628011 samples/sec	accuracy=0.486375	lr=0.325799
Epoch[37] Batch [799]	Speed: 5742.449405 samples/sec	accuracy=0.486007	lr=0.325629
Epoch[37] Batch [849]	Speed: 5699.470934 samples/sec	accuracy=0.485815	lr=0.325459
Epoch[37] Batch [899]	Speed: 5693.816656 samples/sec	accuracy=0.485752	lr=0.325289
Epoch[37] Batch [949]	Speed: 5380.300035 samples/sec	accuracy=0.485620	lr=0.325118
Epoch[37] Batch [999]	Speed: 5448.265510 samples/sec	accuracy=0.485417	lr=0.324948
Epoch[37] Batch [1049]	Speed: 5794.115859 samples/sec	accuracy=0.485311	lr=0.324777
Epoch[37] Batch [1099]	Speed: 5593.383739 samples/sec	accuracy=0.485357	lr=0.324607
Epoch[37] Batch [1149]	Speed: 6126.926717 samples/sec	accuracy=0.485322	lr=0.324436
Epoch[37] Batch [1199]	Speed: 5742.468599 samples/sec	accuracy=0.485202	lr=0.324265
Epoch[37] Batch [1249]	Speed: 5699.679082 samples/sec	accuracy=0.485279	lr=0.324094
[Epoch 37] training: accuracy=0.485277
[Epoch 37] speed: 5635 samples/sec	time cost: 236.748006
[Epoch 37] validation: err-top1=0.481764 err-top5=0.227479
Epoch[38] Batch [49]	Speed: 5506.560069 samples/sec	accuracy=0.495000	lr=0.323919
Epoch[38] Batch [99]	Speed: 5657.352178 samples/sec	accuracy=0.491924	lr=0.323747
Epoch[38] Batch [149]	Speed: 5129.855972 samples/sec	accuracy=0.490898	lr=0.323576
Epoch[38] Batch [199]	Speed: 5525.054914 samples/sec	accuracy=0.490981	lr=0.323404
Epoch[38] Batch [249]	Speed: 5562.600514 samples/sec	accuracy=0.490863	lr=0.323232
Epoch[38] Batch [299]	Speed: 5678.365585 samples/sec	accuracy=0.489867	lr=0.323060
Epoch[38] Batch [349]	Speed: 5671.880191 samples/sec	accuracy=0.488532	lr=0.322888
Epoch[38] Batch [399]	Speed: 5519.033803 samples/sec	accuracy=0.488584	lr=0.322715
Epoch[38] Batch [449]	Speed: 5650.788723 samples/sec	accuracy=0.488553	lr=0.322543
Epoch[38] Batch [499]	Speed: 5951.800300 samples/sec	accuracy=0.488342	lr=0.322370
Epoch[38] Batch [549]	Speed: 5714.989324 samples/sec	accuracy=0.487951	lr=0.322197
Epoch[38] Batch [599]	Speed: 5817.205483 samples/sec	accuracy=0.487687	lr=0.322025
Epoch[38] Batch [649]	Speed: 5552.535238 samples/sec	accuracy=0.487471	lr=0.321851
Epoch[38] Batch [699]	Speed: 5551.955434 samples/sec	accuracy=0.487358	lr=0.321678
Epoch[38] Batch [749]	Speed: 5314.594458 samples/sec	accuracy=0.487495	lr=0.321505
Epoch[38] Batch [799]	Speed: 5371.468443 samples/sec	accuracy=0.487363	lr=0.321331
Epoch[38] Batch [849]	Speed: 5765.173005 samples/sec	accuracy=0.487300	lr=0.321158
Epoch[38] Batch [899]	Speed: 5840.367886 samples/sec	accuracy=0.487118	lr=0.320984
Epoch[38] Batch [949]	Speed: 5815.438768 samples/sec	accuracy=0.486786	lr=0.320810
Epoch[38] Batch [999]	Speed: 5589.455004 samples/sec	accuracy=0.486667	lr=0.320636
Epoch[38] Batch [1049]	Speed: 5865.479610 samples/sec	accuracy=0.486428	lr=0.320462
Epoch[38] Batch [1099]	Speed: 5930.281746 samples/sec	accuracy=0.486306	lr=0.320287
Epoch[38] Batch [1149]	Speed: 5876.153299 samples/sec	accuracy=0.486290	lr=0.320113
Epoch[38] Batch [1199]	Speed: 5371.167638 samples/sec	accuracy=0.486206	lr=0.319938
Epoch[38] Batch [1249]	Speed: 5618.264273 samples/sec	accuracy=0.486180	lr=0.319763
[Epoch 38] training: accuracy=0.486179
[Epoch 38] speed: 5619 samples/sec	time cost: 237.520744
[Epoch 38] validation: err-top1=0.505620 err-top5=0.251694
Epoch[39] Batch [49]	Speed: 5697.852250 samples/sec	accuracy=0.494141	lr=0.319585
Epoch[39] Batch [99]	Speed: 5679.597659 samples/sec	accuracy=0.494707	lr=0.319410
Epoch[39] Batch [149]	Speed: 5708.562388 samples/sec	accuracy=0.494629	lr=0.319234
Epoch[39] Batch [199]	Speed: 5536.108873 samples/sec	accuracy=0.493525	lr=0.319059
Epoch[39] Batch [249]	Speed: 5691.167995 samples/sec	accuracy=0.492844	lr=0.318883
Epoch[39] Batch [299]	Speed: 5759.243262 samples/sec	accuracy=0.492347	lr=0.318708
Epoch[39] Batch [349]	Speed: 5676.119523 samples/sec	accuracy=0.491942	lr=0.318532
Epoch[39] Batch [399]	Speed: 5233.218605 samples/sec	accuracy=0.491157	lr=0.318356
Epoch[39] Batch [449]	Speed: 5654.877891 samples/sec	accuracy=0.490256	lr=0.318180
Epoch[39] Batch [499]	Speed: 5879.361343 samples/sec	accuracy=0.489768	lr=0.318004
Epoch[39] Batch [549]	Speed: 5708.337202 samples/sec	accuracy=0.489519	lr=0.317827
Epoch[39] Batch [599]	Speed: 5663.287123 samples/sec	accuracy=0.489403	lr=0.317651
Epoch[39] Batch [649]	Speed: 5317.732538 samples/sec	accuracy=0.489220	lr=0.317474
Epoch[39] Batch [699]	Speed: 5787.858605 samples/sec	accuracy=0.488902	lr=0.317297
Epoch[39] Batch [749]	Speed: 5552.329228 samples/sec	accuracy=0.488693	lr=0.317120
Epoch[39] Batch [799]	Speed: 5758.430327 samples/sec	accuracy=0.488589	lr=0.316943
Epoch[39] Batch [849]	Speed: 5628.942004 samples/sec	accuracy=0.488338	lr=0.316766
Epoch[39] Batch [899]	Speed: 5763.363659 samples/sec	accuracy=0.488166	lr=0.316589
Epoch[39] Batch [949]	Speed: 5831.741841 samples/sec	accuracy=0.488220	lr=0.316411
Epoch[39] Batch [999]	Speed: 5152.492625 samples/sec	accuracy=0.488252	lr=0.316233
Epoch[39] Batch [1049]	Speed: 5625.667532 samples/sec	accuracy=0.488053	lr=0.316056
Epoch[39] Batch [1099]	Speed: 5529.141586 samples/sec	accuracy=0.487936	lr=0.315878
Epoch[39] Batch [1149]	Speed: 5674.432206 samples/sec	accuracy=0.487756	lr=0.315700
Epoch[39] Batch [1199]	Speed: 5586.667796 samples/sec	accuracy=0.487568	lr=0.315522
Epoch[39] Batch [1249]	Speed: 6094.988394 samples/sec	accuracy=0.487510	lr=0.315343
[Epoch 39] training: accuracy=0.487500
[Epoch 39] speed: 5635 samples/sec	time cost: 236.587533
[Epoch 39] validation: err-top1=0.504006 err-top5=0.249980
Epoch[40] Batch [49]	Speed: 5334.441468 samples/sec	accuracy=0.493887	lr=0.315161
Epoch[40] Batch [99]	Speed: 6044.529531 samples/sec	accuracy=0.495889	lr=0.314983
Epoch[40] Batch [149]	Speed: 5439.325848 samples/sec	accuracy=0.495319	lr=0.314804
Epoch[40] Batch [199]	Speed: 5760.445788 samples/sec	accuracy=0.493994	lr=0.314625
Epoch[40] Batch [249]	Speed: 5561.272490 samples/sec	accuracy=0.493277	lr=0.314446
Epoch[40] Batch [299]	Speed: 5694.952746 samples/sec	accuracy=0.493034	lr=0.314267
Epoch[40] Batch [349]	Speed: 5596.421786 samples/sec	accuracy=0.493016	lr=0.314087
Epoch[40] Batch [399]	Speed: 5515.685877 samples/sec	accuracy=0.492261	lr=0.313908
Epoch[40] Batch [449]	Speed: 5514.368263 samples/sec	accuracy=0.491465	lr=0.313728
Epoch[40] Batch [499]	Speed: 5466.774443 samples/sec	accuracy=0.491760	lr=0.313549
Epoch[40] Batch [549]	Speed: 5660.886284 samples/sec	accuracy=0.491523	lr=0.313369
Epoch[40] Batch [599]	Speed: 5488.178394 samples/sec	accuracy=0.491006	lr=0.313189
Epoch[40] Batch [649]	Speed: 5117.679153 samples/sec	accuracy=0.490910	lr=0.313009
Epoch[40] Batch [699]	Speed: 5140.540572 samples/sec	accuracy=0.490880	lr=0.312829
Epoch[40] Batch [749]	Speed: 5523.793205 samples/sec	accuracy=0.490482	lr=0.312648
Epoch[40] Batch [799]	Speed: 5587.689120 samples/sec	accuracy=0.490527	lr=0.312468
Epoch[40] Batch [849]	Speed: 5488.563709 samples/sec	accuracy=0.490579	lr=0.312287
Epoch[40] Batch [899]	Speed: 5400.863159 samples/sec	accuracy=0.490221	lr=0.312106
Epoch[40] Batch [949]	Speed: 5833.618310 samples/sec	accuracy=0.489971	lr=0.311925
Epoch[40] Batch [999]	Speed: 5512.665488 samples/sec	accuracy=0.489970	lr=0.311744
Epoch[40] Batch [1049]	Speed: 5298.469953 samples/sec	accuracy=0.489804	lr=0.311563
Epoch[40] Batch [1099]	Speed: 5421.409284 samples/sec	accuracy=0.489798	lr=0.311382
Epoch[40] Batch [1149]	Speed: 4476.303515 samples/sec	accuracy=0.489611	lr=0.311201
Epoch[40] Batch [1199]	Speed: 4980.173189 samples/sec	accuracy=0.489303	lr=0.311019
Epoch[40] Batch [1249]	Speed: 5697.796314 samples/sec	accuracy=0.489190	lr=0.310837
[Epoch 40] training: accuracy=0.489185
[Epoch 40] speed: 5435 samples/sec	time cost: 244.945909
[Epoch 40] validation: err-top1=0.494019 err-top5=0.240763
Epoch[41] Batch [49]	Speed: 5719.528406 samples/sec	accuracy=0.494453	lr=0.310652
Epoch[41] Batch [99]	Speed: 5693.717776 samples/sec	accuracy=0.496230	lr=0.310470
Epoch[41] Batch [149]	Speed: 5787.803072 samples/sec	accuracy=0.496816	lr=0.310288
Epoch[41] Batch [199]	Speed: 6023.620818 samples/sec	accuracy=0.496499	lr=0.310105
Epoch[41] Batch [249]	Speed: 5510.792930 samples/sec	accuracy=0.495730	lr=0.309923
Epoch[41] Batch [299]	Speed: 5723.278072 samples/sec	accuracy=0.495615	lr=0.309741
Epoch[41] Batch [349]	Speed: 5621.232301 samples/sec	accuracy=0.495354	lr=0.309558
Epoch[41] Batch [399]	Speed: 5561.037606 samples/sec	accuracy=0.495393	lr=0.309375
Epoch[41] Batch [449]	Speed: 5686.362252 samples/sec	accuracy=0.495152	lr=0.309192
Epoch[41] Batch [499]	Speed: 4591.140581 samples/sec	accuracy=0.494770	lr=0.309009
Epoch[41] Batch [549]	Speed: 5236.565438 samples/sec	accuracy=0.494377	lr=0.308826
Epoch[41] Batch [599]	Speed: 5541.819490 samples/sec	accuracy=0.494237	lr=0.308643
Epoch[41] Batch [649]	Speed: 4791.728684 samples/sec	accuracy=0.493920	lr=0.308459
Epoch[41] Batch [699]	Speed: 5542.131418 samples/sec	accuracy=0.493722	lr=0.308276
Epoch[41] Batch [749]	Speed: 4298.928329 samples/sec	accuracy=0.493714	lr=0.308092
Epoch[41] Batch [799]	Speed: 4742.548644 samples/sec	accuracy=0.493351	lr=0.307908
Epoch[41] Batch [849]	Speed: 5972.896743 samples/sec	accuracy=0.493166	lr=0.307725
Epoch[41] Batch [899]	Speed: 5743.757072 samples/sec	accuracy=0.492836	lr=0.307540
Epoch[41] Batch [949]	Speed: 5498.402271 samples/sec	accuracy=0.492609	lr=0.307356
Epoch[41] Batch [999]	Speed: 5324.431547 samples/sec	accuracy=0.492743	lr=0.307172
Epoch[41] Batch [1049]	Speed: 5695.775047 samples/sec	accuracy=0.492795	lr=0.306988
Epoch[41] Batch [1099]	Speed: 5509.511434 samples/sec	accuracy=0.492604	lr=0.306803
Epoch[41] Batch [1149]	Speed: 4889.423058 samples/sec	accuracy=0.492224	lr=0.306618
Epoch[41] Batch [1199]	Speed: 5645.765908 samples/sec	accuracy=0.491816	lr=0.306433
Epoch[41] Batch [1249]	Speed: 5643.673700 samples/sec	accuracy=0.491641	lr=0.306249
[Epoch 41] training: accuracy=0.491636
[Epoch 41] speed: 5395 samples/sec	time cost: 247.058935
[Epoch 41] validation: err-top1=0.478974 err-top5=0.227997
Epoch[42] Batch [49]	Speed: 5304.154379 samples/sec	accuracy=0.502422	lr=0.306060
Epoch[42] Batch [99]	Speed: 5501.821566 samples/sec	accuracy=0.498477	lr=0.305875
Epoch[42] Batch [149]	Speed: 5914.095635 samples/sec	accuracy=0.498073	lr=0.305689
Epoch[42] Batch [199]	Speed: 5066.487325 samples/sec	accuracy=0.497944	lr=0.305504
Epoch[42] Batch [249]	Speed: 5039.235583 samples/sec	accuracy=0.497109	lr=0.305318
Epoch[42] Batch [299]	Speed: 5399.855622 samples/sec	accuracy=0.496816	lr=0.305133
Epoch[42] Batch [349]	Speed: 5992.916132 samples/sec	accuracy=0.496105	lr=0.304947
Epoch[42] Batch [399]	Speed: 5788.505114 samples/sec	accuracy=0.495413	lr=0.304761
Epoch[42] Batch [449]	Speed: 5376.035850 samples/sec	accuracy=0.494924	lr=0.304575
Epoch[42] Batch [499]	Speed: 4292.085504 samples/sec	accuracy=0.494498	lr=0.304388
Epoch[42] Batch [549]	Speed: 5073.324923 samples/sec	accuracy=0.494377	lr=0.304202
Epoch[42] Batch [599]	Speed: 5503.557139 samples/sec	accuracy=0.494180	lr=0.304016
Epoch[42] Batch [649]	Speed: 5466.501275 samples/sec	accuracy=0.493933	lr=0.303829
Epoch[42] Batch [699]	Speed: 5749.021094 samples/sec	accuracy=0.494169	lr=0.303642
Epoch[42] Batch [749]	Speed: 5747.036228 samples/sec	accuracy=0.494251	lr=0.303456
Epoch[42] Batch [799]	Speed: 5910.961682 samples/sec	accuracy=0.493878	lr=0.303269
Epoch[42] Batch [849]	Speed: 5345.201858 samples/sec	accuracy=0.493709	lr=0.303082
Epoch[42] Batch [899]	Speed: 4914.899381 samples/sec	accuracy=0.493675	lr=0.302894
Epoch[42] Batch [949]	Speed: 5115.842368 samples/sec	accuracy=0.493590	lr=0.302707
Epoch[42] Batch [999]	Speed: 4782.031876 samples/sec	accuracy=0.493357	lr=0.302520
Epoch[42] Batch [1049]	Speed: 4622.775221 samples/sec	accuracy=0.493230	lr=0.302332
Epoch[42] Batch [1099]	Speed: 5038.378301 samples/sec	accuracy=0.493121	lr=0.302144
Epoch[42] Batch [1149]	Speed: 5848.565177 samples/sec	accuracy=0.492988	lr=0.301957
Epoch[42] Batch [1199]	Speed: 5869.638446 samples/sec	accuracy=0.492900	lr=0.301769
Epoch[42] Batch [1249]	Speed: 5581.555458 samples/sec	accuracy=0.492806	lr=0.301581
[Epoch 42] training: accuracy=0.492804
[Epoch 42] speed: 5327 samples/sec	time cost: 250.776593
[Epoch 42] validation: err-top1=0.468909 err-top5=0.219448
Epoch[43] Batch [49]	Speed: 5581.494094 samples/sec	accuracy=0.501582	lr=0.301389
Epoch[43] Batch [99]	Speed: 5654.225901 samples/sec	accuracy=0.501045	lr=0.301200
Epoch[43] Batch [149]	Speed: 5765.416473 samples/sec	accuracy=0.498307	lr=0.301012
Epoch[43] Batch [199]	Speed: 5553.635031 samples/sec	accuracy=0.497363	lr=0.300823
Epoch[43] Batch [249]	Speed: 5243.425977 samples/sec	accuracy=0.497094	lr=0.300635
Epoch[43] Batch [299]	Speed: 5497.664820 samples/sec	accuracy=0.496872	lr=0.300446
Epoch[43] Batch [349]	Speed: 5314.520015 samples/sec	accuracy=0.497134	lr=0.300257
Epoch[43] Batch [399]	Speed: 5739.266282 samples/sec	accuracy=0.497065	lr=0.300068
Epoch[43] Batch [449]	Speed: 5086.156230 samples/sec	accuracy=0.496678	lr=0.299879
Epoch[43] Batch [499]	Speed: 5613.404730 samples/sec	accuracy=0.496330	lr=0.299690
Epoch[43] Batch [549]	Speed: 5547.342278 samples/sec	accuracy=0.496097	lr=0.299500
Epoch[43] Batch [599]	Speed: 5366.373406 samples/sec	accuracy=0.496061	lr=0.299311
Epoch[43] Batch [649]	Speed: 5877.329706 samples/sec	accuracy=0.495987	lr=0.299121
Epoch[43] Batch [699]	Speed: 5272.497508 samples/sec	accuracy=0.495926	lr=0.298932
Epoch[43] Batch [749]	Speed: 5064.244703 samples/sec	accuracy=0.495814	lr=0.298742
Epoch[43] Batch [799]	Speed: 5376.584607 samples/sec	accuracy=0.495627	lr=0.298552
Epoch[43] Batch [849]	Speed: 5231.105543 samples/sec	accuracy=0.495418	lr=0.298362
Epoch[43] Batch [899]	Speed: 4093.456190 samples/sec	accuracy=0.495217	lr=0.298171
Epoch[43] Batch [949]	Speed: 5896.192527 samples/sec	accuracy=0.494884	lr=0.297981
Epoch[43] Batch [999]	Speed: 4647.744589 samples/sec	accuracy=0.494696	lr=0.297791
Epoch[43] Batch [1049]	Speed: 5315.932542 samples/sec	accuracy=0.494643	lr=0.297600
Epoch[43] Batch [1099]	Speed: 4667.319107 samples/sec	accuracy=0.494693	lr=0.297409
Epoch[43] Batch [1149]	Speed: 5658.278749 samples/sec	accuracy=0.494609	lr=0.297219
Epoch[43] Batch [1199]	Speed: 4891.185493 samples/sec	accuracy=0.494346	lr=0.297028
Epoch[43] Batch [1249]	Speed: 5791.553168 samples/sec	accuracy=0.494264	lr=0.296837
[Epoch 43] training: accuracy=0.494278
[Epoch 43] speed: 5306 samples/sec	time cost: 250.761987
[Epoch 43] validation: err-top1=0.492267 err-top5=0.242227
Epoch[44] Batch [49]	Speed: 5518.299883 samples/sec	accuracy=0.500059	lr=0.296642
Epoch[44] Batch [99]	Speed: 5173.445815 samples/sec	accuracy=0.500635	lr=0.296451
Epoch[44] Batch [149]	Speed: 5795.919539 samples/sec	accuracy=0.499264	lr=0.296259
Epoch[44] Batch [199]	Speed: 5569.153826 samples/sec	accuracy=0.497671	lr=0.296068
Epoch[44] Batch [249]	Speed: 5678.947014 samples/sec	accuracy=0.497707	lr=0.295876
Epoch[44] Batch [299]	Speed: 5707.091416 samples/sec	accuracy=0.497692	lr=0.295684
Epoch[44] Batch [349]	Speed: 5883.800233 samples/sec	accuracy=0.497048	lr=0.295493
Epoch[44] Batch [399]	Speed: 5765.115121 samples/sec	accuracy=0.497012	lr=0.295301
Epoch[44] Batch [449]	Speed: 5284.910912 samples/sec	accuracy=0.497112	lr=0.295109
Epoch[44] Batch [499]	Speed: 5374.147220 samples/sec	accuracy=0.496576	lr=0.294917
Epoch[44] Batch [549]	Speed: 5581.340761 samples/sec	accuracy=0.496477	lr=0.294724
Epoch[44] Batch [599]	Speed: 5715.347823 samples/sec	accuracy=0.496462	lr=0.294532
Epoch[44] Batch [649]	Speed: 5618.652635 samples/sec	accuracy=0.496173	lr=0.294339
Epoch[44] Batch [699]	Speed: 5450.526981 samples/sec	accuracy=0.496250	lr=0.294147
Epoch[44] Batch [749]	Speed: 5540.929521 samples/sec	accuracy=0.496353	lr=0.293954
Epoch[44] Batch [799]	Speed: 5224.836964 samples/sec	accuracy=0.496307	lr=0.293761
Epoch[44] Batch [849]	Speed: 5631.320545 samples/sec	accuracy=0.496247	lr=0.293568
Epoch[44] Batch [899]	Speed: 5620.221625 samples/sec	accuracy=0.496149	lr=0.293375
Epoch[44] Batch [949]	Speed: 5456.272333 samples/sec	accuracy=0.496089	lr=0.293182
Epoch[44] Batch [999]	Speed: 5591.329439 samples/sec	accuracy=0.495803	lr=0.292989
Epoch[44] Batch [1049]	Speed: 5652.344472 samples/sec	accuracy=0.495508	lr=0.292795
Epoch[44] Batch [1099]	Speed: 5460.680356 samples/sec	accuracy=0.495456	lr=0.292602
Epoch[44] Batch [1149]	Speed: 5613.205476 samples/sec	accuracy=0.495372	lr=0.292408
Epoch[44] Batch [1199]	Speed: 5475.942308 samples/sec	accuracy=0.495267	lr=0.292215
Epoch[44] Batch [1249]	Speed: 4509.650848 samples/sec	accuracy=0.495026	lr=0.292021
[Epoch 44] training: accuracy=0.495034
[Epoch 44] speed: 5494 samples/sec	time cost: 242.655834
[Epoch 44] validation: err-top1=0.480608 err-top5=0.231605
Epoch[45] Batch [49]	Speed: 5622.688489 samples/sec	accuracy=0.505410	lr=0.291823
Epoch[45] Batch [99]	Speed: 5893.864370 samples/sec	accuracy=0.503643	lr=0.291629
Epoch[45] Batch [149]	Speed: 5532.555474 samples/sec	accuracy=0.503268	lr=0.291435
Epoch[45] Batch [199]	Speed: 5814.241353 samples/sec	accuracy=0.502520	lr=0.291240
Epoch[45] Batch [249]	Speed: 5531.633141 samples/sec	accuracy=0.502594	lr=0.291046
Epoch[45] Batch [299]	Speed: 5304.001626 samples/sec	accuracy=0.502454	lr=0.290852
Epoch[45] Batch [349]	Speed: 5467.189606 samples/sec	accuracy=0.501688	lr=0.290657
Epoch[45] Batch [399]	Speed: 5037.629793 samples/sec	accuracy=0.500674	lr=0.290462
Epoch[45] Batch [449]	Speed: 5111.475642 samples/sec	accuracy=0.500143	lr=0.290267
Epoch[45] Batch [499]	Speed: 5192.903627 samples/sec	accuracy=0.500199	lr=0.290073
Epoch[45] Batch [549]	Speed: 5539.731575 samples/sec	accuracy=0.500073	lr=0.289878
Epoch[45] Batch [599]	Speed: 5310.364280 samples/sec	accuracy=0.499546	lr=0.289682
Epoch[45] Batch [649]	Speed: 5716.002424 samples/sec	accuracy=0.499096	lr=0.289487
Epoch[45] Batch [699]	Speed: 5437.350505 samples/sec	accuracy=0.498963	lr=0.289292
Epoch[45] Batch [749]	Speed: 5136.826234 samples/sec	accuracy=0.498717	lr=0.289096
Epoch[45] Batch [799]	Speed: 4992.638384 samples/sec	accuracy=0.498596	lr=0.288901
Epoch[45] Batch [849]	Speed: 5738.787608 samples/sec	accuracy=0.498636	lr=0.288705
Epoch[45] Batch [899]	Speed: 4391.044401 samples/sec	accuracy=0.498541	lr=0.288509
Epoch[45] Batch [949]	Speed: 4643.556280 samples/sec	accuracy=0.498265	lr=0.288314
Epoch[45] Batch [999]	Speed: 4962.042794 samples/sec	accuracy=0.498122	lr=0.288118
Epoch[45] Batch [1049]	Speed: 5222.209688 samples/sec	accuracy=0.497925	lr=0.287921
Epoch[45] Batch [1099]	Speed: 5588.172002 samples/sec	accuracy=0.497985	lr=0.287725
Epoch[45] Batch [1149]	Speed: 5504.994620 samples/sec	accuracy=0.497761	lr=0.287529
Epoch[45] Batch [1199]	Speed: 5457.047394 samples/sec	accuracy=0.497491	lr=0.287333
Epoch[45] Batch [1249]	Speed: 5294.367335 samples/sec	accuracy=0.497365	lr=0.287136
[Epoch 45] training: accuracy=0.497357
[Epoch 45] speed: 5307 samples/sec	time cost: 251.794223
[Epoch 45] validation: err-top1=0.477878 err-top5=0.226961
Epoch[46] Batch [49]	Speed: 5089.358555 samples/sec	accuracy=0.505957	lr=0.286936
Epoch[46] Batch [99]	Speed: 5304.693013 samples/sec	accuracy=0.504424	lr=0.286739
Epoch[46] Batch [149]	Speed: 5260.388644 samples/sec	accuracy=0.502493	lr=0.286542
Epoch[46] Batch [199]	Speed: 5472.292503 samples/sec	accuracy=0.501460	lr=0.286345
Epoch[46] Batch [249]	Speed: 4328.618932 samples/sec	accuracy=0.502012	lr=0.286148
Epoch[46] Batch [299]	Speed: 5577.857873 samples/sec	accuracy=0.501621	lr=0.285951
Epoch[46] Batch [349]	Speed: 5530.304336 samples/sec	accuracy=0.501278	lr=0.285754
Epoch[46] Batch [399]	Speed: 4951.249472 samples/sec	accuracy=0.501096	lr=0.285556
Epoch[46] Batch [449]	Speed: 5154.151830 samples/sec	accuracy=0.501217	lr=0.285359
Epoch[46] Batch [499]	Speed: 5527.813836 samples/sec	accuracy=0.501271	lr=0.285161
Epoch[46] Batch [549]	Speed: 5537.069819 samples/sec	accuracy=0.500804	lr=0.284964
Epoch[46] Batch [599]	Speed: 5487.250606 samples/sec	accuracy=0.500757	lr=0.284766
Epoch[46] Batch [649]	Speed: 5625.480964 samples/sec	accuracy=0.500711	lr=0.284568
Epoch[46] Batch [699]	Speed: 5774.512712 samples/sec	accuracy=0.500525	lr=0.284370
Epoch[46] Batch [749]	Speed: 5194.346718 samples/sec	accuracy=0.500349	lr=0.284172
Epoch[46] Batch [799]	Speed: 5113.568149 samples/sec	accuracy=0.500005	lr=0.283974
Epoch[46] Batch [849]	Speed: 5315.648845 samples/sec	accuracy=0.499745	lr=0.283776
Epoch[46] Batch [899]	Speed: 5188.624294 samples/sec	accuracy=0.499637	lr=0.283577
Epoch[46] Batch [949]	Speed: 5119.430605 samples/sec	accuracy=0.499669	lr=0.283379
Epoch[46] Batch [999]	Speed: 5608.434063 samples/sec	accuracy=0.499675	lr=0.283181
Epoch[46] Batch [1049]	Speed: 5066.064575 samples/sec	accuracy=0.499621	lr=0.282982
Epoch[46] Batch [1099]	Speed: 5089.720301 samples/sec	accuracy=0.499616	lr=0.282783
Epoch[46] Batch [1149]	Speed: 5545.259222 samples/sec	accuracy=0.499368	lr=0.282584
Epoch[46] Batch [1199]	Speed: 5547.275502 samples/sec	accuracy=0.499037	lr=0.282385
Epoch[46] Batch [1249]	Speed: 5100.069948 samples/sec	accuracy=0.499013	lr=0.282186
[Epoch 46] training: accuracy=0.499031
[Epoch 46] speed: 5276 samples/sec	time cost: 253.809062
[Epoch 46] validation: err-top1=0.482585 err-top5=0.230062
Epoch[47] Batch [49]	Speed: 5033.833828 samples/sec	accuracy=0.504180	lr=0.281983
Epoch[47] Batch [99]	Speed: 5579.792242 samples/sec	accuracy=0.505713	lr=0.281784
Epoch[47] Batch [149]	Speed: 4972.410030 samples/sec	accuracy=0.503926	lr=0.281585
Epoch[47] Batch [199]	Speed: 5372.053492 samples/sec	accuracy=0.503965	lr=0.281385
Epoch[47] Batch [249]	Speed: 4600.435180 samples/sec	accuracy=0.503352	lr=0.281186
Epoch[47] Batch [299]	Speed: 5139.780471 samples/sec	accuracy=0.503945	lr=0.280986
Epoch[47] Batch [349]	Speed: 5843.836394 samples/sec	accuracy=0.503747	lr=0.280786
Epoch[47] Batch [399]	Speed: 5797.152768 samples/sec	accuracy=0.503328	lr=0.280587
Epoch[47] Batch [449]	Speed: 4453.539027 samples/sec	accuracy=0.502745	lr=0.280387
Epoch[47] Batch [499]	Speed: 4949.402086 samples/sec	accuracy=0.502799	lr=0.280187
Epoch[47] Batch [549]	Speed: 5376.481227 samples/sec	accuracy=0.502152	lr=0.279987
Epoch[47] Batch [599]	Speed: 5635.240055 samples/sec	accuracy=0.501965	lr=0.279786
Epoch[47] Batch [649]	Speed: 5572.876465 samples/sec	accuracy=0.501549	lr=0.279586
Epoch[47] Batch [699]	Speed: 4992.080485 samples/sec	accuracy=0.501448	lr=0.279386
Epoch[47] Batch [749]	Speed: 4950.103608 samples/sec	accuracy=0.501318	lr=0.279185
Epoch[47] Batch [799]	Speed: 4927.550568 samples/sec	accuracy=0.501348	lr=0.278985
Epoch[47] Batch [849]	Speed: 5005.091593 samples/sec	accuracy=0.501142	lr=0.278784
Epoch[47] Batch [899]	Speed: 4811.842416 samples/sec	accuracy=0.501068	lr=0.278583
Epoch[47] Batch [949]	Speed: 5310.129760 samples/sec	accuracy=0.501028	lr=0.278382
Epoch[47] Batch [999]	Speed: 5761.559468 samples/sec	accuracy=0.500900	lr=0.278181
Epoch[47] Batch [1049]	Speed: 5845.211810 samples/sec	accuracy=0.500812	lr=0.277980
Epoch[47] Batch [1099]	Speed: 6057.731390 samples/sec	accuracy=0.500649	lr=0.277779
Epoch[47] Batch [1149]	Speed: 5520.034093 samples/sec	accuracy=0.500543	lr=0.277578
Epoch[47] Batch [1199]	Speed: 5475.521766 samples/sec	accuracy=0.500305	lr=0.277377
Epoch[47] Batch [1249]	Speed: 5651.000171 samples/sec	accuracy=0.500125	lr=0.277175
[Epoch 47] training: accuracy=0.500098
[Epoch 47] speed: 5266 samples/sec	time cost: 252.639417
[Epoch 47] validation: err-top1=0.476981 err-top5=0.227380
Epoch[48] Batch [49]	Speed: 5287.822880 samples/sec	accuracy=0.512383	lr=0.276970
Epoch[48] Batch [99]	Speed: 5340.858413 samples/sec	accuracy=0.509268	lr=0.276768
Epoch[48] Batch [149]	Speed: 5525.690677 samples/sec	accuracy=0.507812	lr=0.276566
Epoch[48] Batch [199]	Speed: 5815.482864 samples/sec	accuracy=0.507163	lr=0.276365
Epoch[48] Batch [249]	Speed: 5618.303077 samples/sec	accuracy=0.506477	lr=0.276163
Epoch[48] Batch [299]	Speed: 5004.607765 samples/sec	accuracy=0.505785	lr=0.275961
Epoch[48] Batch [349]	Speed: 5397.756200 samples/sec	accuracy=0.505868	lr=0.275759
Epoch[48] Batch [399]	Speed: 5732.185421 samples/sec	accuracy=0.505471	lr=0.275557
Epoch[48] Batch [449]	Speed: 5754.952467 samples/sec	accuracy=0.505106	lr=0.275354
Epoch[48] Batch [499]	Speed: 5633.160664 samples/sec	accuracy=0.504932	lr=0.275152
Epoch[48] Batch [549]	Speed: 4636.631590 samples/sec	accuracy=0.504959	lr=0.274950
Epoch[48] Batch [599]	Speed: 5798.535415 samples/sec	accuracy=0.504840	lr=0.274747
Epoch[48] Batch [649]	Speed: 5634.288196 samples/sec	accuracy=0.504624	lr=0.274545
Epoch[48] Batch [699]	Speed: 5182.539725 samples/sec	accuracy=0.504438	lr=0.274342
Epoch[48] Batch [749]	Speed: 5259.213479 samples/sec	accuracy=0.504306	lr=0.274139
Epoch[48] Batch [799]	Speed: 5191.626755 samples/sec	accuracy=0.504227	lr=0.273936
Epoch[48] Batch [849]	Speed: 5408.002040 samples/sec	accuracy=0.503845	lr=0.273733
Epoch[48] Batch [899]	Speed: 4776.162199 samples/sec	accuracy=0.503622	lr=0.273530
Epoch[48] Batch [949]	Speed: 5737.736215 samples/sec	accuracy=0.503594	lr=0.273327
Epoch[48] Batch [999]	Speed: 5757.841618 samples/sec	accuracy=0.503406	lr=0.273124
Epoch[48] Batch [1049]	Speed: 5315.350838 samples/sec	accuracy=0.503092	lr=0.272921
Epoch[48] Batch [1099]	Speed: 5750.123894 samples/sec	accuracy=0.502864	lr=0.272717
Epoch[48] Batch [1149]	Speed: 5690.968612 samples/sec	accuracy=0.502589	lr=0.272514
Epoch[48] Batch [1199]	Speed: 5562.051740 samples/sec	accuracy=0.502322	lr=0.272310
Epoch[48] Batch [1249]	Speed: 5723.250617 samples/sec	accuracy=0.502289	lr=0.272107
[Epoch 48] training: accuracy=0.502295
[Epoch 48] speed: 5437 samples/sec	time cost: 244.910623
[Epoch 48] validation: err-top1=0.480509 err-top5=0.228496
Epoch[49] Batch [49]	Speed: 5381.259158 samples/sec	accuracy=0.511816	lr=0.271899
Epoch[49] Batch [99]	Speed: 5937.449052 samples/sec	accuracy=0.508877	lr=0.271695
Epoch[49] Batch [149]	Speed: 5394.992444 samples/sec	accuracy=0.509414	lr=0.271491
Epoch[49] Batch [199]	Speed: 4965.207097 samples/sec	accuracy=0.508481	lr=0.271287
Epoch[49] Batch [249]	Speed: 5685.530623 samples/sec	accuracy=0.508297	lr=0.271083
Epoch[49] Batch [299]	Speed: 5138.635207 samples/sec	accuracy=0.507982	lr=0.270879
Epoch[49] Batch [349]	Speed: 5325.873783 samples/sec	accuracy=0.507921	lr=0.270675
Epoch[49] Batch [399]	Speed: 5679.629203 samples/sec	accuracy=0.507441	lr=0.270470
Epoch[49] Batch [449]	Speed: 4717.383101 samples/sec	accuracy=0.507012	lr=0.270266
Epoch[49] Batch [499]	Speed: 5328.310271 samples/sec	accuracy=0.506447	lr=0.270061
Epoch[49] Batch [549]	Speed: 5283.797178 samples/sec	accuracy=0.506096	lr=0.269857
Epoch[49] Batch [599]	Speed: 5633.116630 samples/sec	accuracy=0.506006	lr=0.269652
Epoch[49] Batch [649]	Speed: 5537.574691 samples/sec	accuracy=0.505783	lr=0.269447
Epoch[49] Batch [699]	Speed: 5804.904442 samples/sec	accuracy=0.505484	lr=0.269243
Epoch[49] Batch [749]	Speed: 5703.883730 samples/sec	accuracy=0.505542	lr=0.269038
Epoch[49] Batch [799]	Speed: 5682.368022 samples/sec	accuracy=0.505624	lr=0.268833
Epoch[49] Batch [849]	Speed: 5400.368239 samples/sec	accuracy=0.505184	lr=0.268628
Epoch[49] Batch [899]	Speed: 5811.010399 samples/sec	accuracy=0.504814	lr=0.268422
Epoch[49] Batch [949]	Speed: 5788.472348 samples/sec	accuracy=0.504668	lr=0.268217
Epoch[49] Batch [999]	Speed: 5930.187583 samples/sec	accuracy=0.504449	lr=0.268012
Epoch[49] Batch [1049]	Speed: 5731.670753 samples/sec	accuracy=0.504422	lr=0.267807
Epoch[49] Batch [1099]	Speed: 5281.907312 samples/sec	accuracy=0.504270	lr=0.267601
Epoch[49] Batch [1149]	Speed: 5059.762549 samples/sec	accuracy=0.504148	lr=0.267396
Epoch[49] Batch [1199]	Speed: 5285.441092 samples/sec	accuracy=0.503991	lr=0.267190
Epoch[49] Batch [1249]	Speed: 4997.799783 samples/sec	accuracy=0.503720	lr=0.266984
[Epoch 49] training: accuracy=0.503706
[Epoch 49] speed: 5436 samples/sec	time cost: 245.486442
[Epoch 49] validation: err-top1=0.491231 err-top5=0.235730
Epoch[50] Batch [49]	Speed: 5155.842431 samples/sec	accuracy=0.511445	lr=0.266774
Epoch[50] Batch [99]	Speed: 5491.021767 samples/sec	accuracy=0.509922	lr=0.266568
Epoch[50] Batch [149]	Speed: 5717.135970 samples/sec	accuracy=0.509421	lr=0.266362
Epoch[50] Batch [199]	Speed: 5671.804841 samples/sec	accuracy=0.508667	lr=0.266156
Epoch[50] Batch [249]	Speed: 5468.130115 samples/sec	accuracy=0.508461	lr=0.265950
Epoch[50] Batch [299]	Speed: 5372.437995 samples/sec	accuracy=0.508613	lr=0.265744
Epoch[50] Batch [349]	Speed: 5314.372715 samples/sec	accuracy=0.509062	lr=0.265538
Epoch[50] Batch [399]	Speed: 5557.918449 samples/sec	accuracy=0.508521	lr=0.265331
Epoch[50] Batch [449]	Speed: 5313.689847 samples/sec	accuracy=0.508184	lr=0.265125
Epoch[50] Batch [499]	Speed: 5213.183057 samples/sec	accuracy=0.507887	lr=0.264918
Epoch[50] Batch [549]	Speed: 4809.803787 samples/sec	accuracy=0.507365	lr=0.264712
Epoch[50] Batch [599]	Speed: 5184.423223 samples/sec	accuracy=0.507235	lr=0.264505
Epoch[50] Batch [649]	Speed: 4764.807421 samples/sec	accuracy=0.506798	lr=0.264298
Epoch[50] Batch [699]	Speed: 4715.069316 samples/sec	accuracy=0.506543	lr=0.264092
Epoch[50] Batch [749]	Speed: 4450.714901 samples/sec	accuracy=0.506557	lr=0.263885
Epoch[50] Batch [799]	Speed: 5419.410004 samples/sec	accuracy=0.506373	lr=0.263678
Epoch[50] Batch [849]	Speed: 5652.858831 samples/sec	accuracy=0.506371	lr=0.263471
Epoch[50] Batch [899]	Speed: 5950.537998 samples/sec	accuracy=0.506280	lr=0.263264
Epoch[50] Batch [949]	Speed: 5975.467003 samples/sec	accuracy=0.506170	lr=0.263056
Epoch[50] Batch [999]	Speed: 5832.254681 samples/sec	accuracy=0.506077	lr=0.262849
Epoch[50] Batch [1049]	Speed: 5812.028259 samples/sec	accuracy=0.505955	lr=0.262642
Epoch[50] Batch [1099]	Speed: 5732.943976 samples/sec	accuracy=0.505850	lr=0.262434
Epoch[50] Batch [1149]	Speed: 5720.516296 samples/sec	accuracy=0.505741	lr=0.262227
Epoch[50] Batch [1199]	Speed: 5797.188762 samples/sec	accuracy=0.505579	lr=0.262019
Epoch[50] Batch [1249]	Speed: 5764.317548 samples/sec	accuracy=0.505397	lr=0.261812
[Epoch 50] training: accuracy=0.505390
[Epoch 50] speed: 5382 samples/sec	time cost: 249.793490
[Epoch 50] validation: err-top1=0.513493 err-top5=0.256676
Epoch[51] Batch [49]	Speed: 5761.704003 samples/sec	accuracy=0.513047	lr=0.261600
Epoch[51] Batch [99]	Speed: 5618.234729 samples/sec	accuracy=0.509609	lr=0.261392
Epoch[51] Batch [149]	Speed: 3915.067828 samples/sec	accuracy=0.510384	lr=0.261184
Epoch[51] Batch [199]	Speed: 5274.489335 samples/sec	accuracy=0.510234	lr=0.260976
Epoch[51] Batch [249]	Speed: 5896.055897 samples/sec	accuracy=0.510758	lr=0.260768
Epoch[51] Batch [299]	Speed: 5525.313921 samples/sec	accuracy=0.510081	lr=0.260560
Epoch[51] Batch [349]	Speed: 5764.440713 samples/sec	accuracy=0.509933	lr=0.260352
Epoch[51] Batch [399]	Speed: 5666.621697 samples/sec	accuracy=0.509409	lr=0.260144
Epoch[51] Batch [449]	Speed: 5818.550732 samples/sec	accuracy=0.508956	lr=0.259935
Epoch[51] Batch [499]	Speed: 5694.423148 samples/sec	accuracy=0.508908	lr=0.259727
Epoch[51] Batch [549]	Speed: 6043.282524 samples/sec	accuracy=0.508970	lr=0.259519
Epoch[51] Batch [599]	Speed: 5836.787364 samples/sec	accuracy=0.509023	lr=0.259310
Epoch[51] Batch [649]	Speed: 6030.908446 samples/sec	accuracy=0.508926	lr=0.259102
Epoch[51] Batch [699]	Speed: 5652.946923 samples/sec	accuracy=0.508770	lr=0.258893
Epoch[51] Batch [749]	Speed: 5971.534979 samples/sec	accuracy=0.508680	lr=0.258684
Epoch[51] Batch [799]	Speed: 5844.416735 samples/sec	accuracy=0.508414	lr=0.258475
Epoch[51] Batch [849]	Speed: 5347.546073 samples/sec	accuracy=0.508372	lr=0.258267
Epoch[51] Batch [899]	Speed: 5774.213668 samples/sec	accuracy=0.508153	lr=0.258058
Epoch[51] Batch [949]	Speed: 5834.303138 samples/sec	accuracy=0.508021	lr=0.257849
Epoch[51] Batch [999]	Speed: 5763.897031 samples/sec	accuracy=0.507795	lr=0.257640
Epoch[51] Batch [1049]	Speed: 5452.382057 samples/sec	accuracy=0.507731	lr=0.257430
Epoch[51] Batch [1099]	Speed: 5532.731225 samples/sec	accuracy=0.507455	lr=0.257221
Epoch[51] Batch [1149]	Speed: 5618.921522 samples/sec	accuracy=0.507317	lr=0.257012
Epoch[51] Batch [1199]	Speed: 5363.020430 samples/sec	accuracy=0.507319	lr=0.256803
Epoch[51] Batch [1249]	Speed: 5815.958510 samples/sec	accuracy=0.507173	lr=0.256593
[Epoch 51] training: accuracy=0.507194
[Epoch 51] speed: 5585 samples/sec	time cost: 240.463351
[Epoch 51] validation: err-top1=0.525590 err-top5=0.273019
Epoch[52] Batch [49]	Speed: 5890.684112 samples/sec	accuracy=0.520059	lr=0.256379
Epoch[52] Batch [99]	Speed: 4520.081645 samples/sec	accuracy=0.516729	lr=0.256170
Epoch[52] Batch [149]	Speed: 5571.182176 samples/sec	accuracy=0.514954	lr=0.255960
Epoch[52] Batch [199]	Speed: 5662.727711 samples/sec	accuracy=0.513867	lr=0.255751
Epoch[52] Batch [249]	Speed: 4682.564717 samples/sec	accuracy=0.513500	lr=0.255541
Epoch[52] Batch [299]	Speed: 5764.356385 samples/sec	accuracy=0.512995	lr=0.255331
Epoch[52] Batch [349]	Speed: 5702.470286 samples/sec	accuracy=0.512190	lr=0.255121
Epoch[52] Batch [399]	Speed: 5685.067793 samples/sec	accuracy=0.512117	lr=0.254911
Epoch[52] Batch [449]	Speed: 5585.002877 samples/sec	accuracy=0.512216	lr=0.254701
Epoch[52] Batch [499]	Speed: 5928.950800 samples/sec	accuracy=0.511732	lr=0.254491
Epoch[52] Batch [549]	Speed: 5612.590049 samples/sec	accuracy=0.511246	lr=0.254281
Epoch[52] Batch [599]	Speed: 5866.454782 samples/sec	accuracy=0.511074	lr=0.254071
Epoch[52] Batch [649]	Speed: 5649.610880 samples/sec	accuracy=0.511122	lr=0.253861
Epoch[52] Batch [699]	Speed: 5875.674669 samples/sec	accuracy=0.510921	lr=0.253650
Epoch[52] Batch [749]	Speed: 5713.600167 samples/sec	accuracy=0.510539	lr=0.253440
Epoch[52] Batch [799]	Speed: 5577.452097 samples/sec	accuracy=0.510579	lr=0.253229
Epoch[52] Batch [849]	Speed: 5791.063858 samples/sec	accuracy=0.510229	lr=0.253019
Epoch[52] Batch [899]	Speed: 5524.622815 samples/sec	accuracy=0.510138	lr=0.252808
Epoch[52] Batch [949]	Speed: 5309.444175 samples/sec	accuracy=0.509779	lr=0.252598
Epoch[52] Batch [999]	Speed: 5420.117854 samples/sec	accuracy=0.509682	lr=0.252387
Epoch[52] Batch [1049]	Speed: 5586.994533 samples/sec	accuracy=0.509364	lr=0.252176
Epoch[52] Batch [1099]	Speed: 5555.697221 samples/sec	accuracy=0.509202	lr=0.251965
Epoch[52] Batch [1149]	Speed: 5941.259703 samples/sec	accuracy=0.509050	lr=0.251754
Epoch[52] Batch [1199]	Speed: 5404.781714 samples/sec	accuracy=0.508857	lr=0.251543
Epoch[52] Batch [1249]	Speed: 5943.607366 samples/sec	accuracy=0.508796	lr=0.251332
[Epoch 52] training: accuracy=0.508791
[Epoch 52] speed: 5559 samples/sec	time cost: 241.599074
[Epoch 52] validation: err-top1=0.468160 err-top5=0.219747
Epoch[53] Batch [49]	Speed: 4395.784076 samples/sec	accuracy=0.522813	lr=0.251117
Epoch[53] Batch [99]	Speed: 5980.859178 samples/sec	accuracy=0.519268	lr=0.250906
Epoch[53] Batch [149]	Speed: 5702.537065 samples/sec	accuracy=0.517617	lr=0.250695
Epoch[53] Batch [199]	Speed: 5836.306243 samples/sec	accuracy=0.516069	lr=0.250483
Epoch[53] Batch [249]	Speed: 5750.546254 samples/sec	accuracy=0.514594	lr=0.250272
Epoch[53] Batch [299]	Speed: 4826.596862 samples/sec	accuracy=0.514486	lr=0.250061
Epoch[53] Batch [349]	Speed: 5538.239476 samples/sec	accuracy=0.514358	lr=0.249849
Epoch[53] Batch [399]	Speed: 5085.327708 samples/sec	accuracy=0.513665	lr=0.249638
Epoch[53] Batch [449]	Speed: 5552.926628 samples/sec	accuracy=0.513700	lr=0.249426
Epoch[53] Batch [499]	Speed: 4748.492175 samples/sec	accuracy=0.513008	lr=0.249215
Epoch[53] Batch [549]	Speed: 5442.318519 samples/sec	accuracy=0.512617	lr=0.249003
Epoch[53] Batch [599]	Speed: 5683.259338 samples/sec	accuracy=0.512546	lr=0.248791
Epoch[53] Batch [649]	Speed: 5131.245103 samples/sec	accuracy=0.512489	lr=0.248579
Epoch[53] Batch [699]	Speed: 5727.310339 samples/sec	accuracy=0.512347	lr=0.248367
Epoch[53] Batch [749]	Speed: 5345.911747 samples/sec	accuracy=0.512072	lr=0.248156
Epoch[53] Batch [799]	Speed: 5746.198290 samples/sec	accuracy=0.511947	lr=0.247944
Epoch[53] Batch [849]	Speed: 5659.169384 samples/sec	accuracy=0.511628	lr=0.247732
Epoch[53] Batch [899]	Speed: 5216.526607 samples/sec	accuracy=0.511356	lr=0.247519
Epoch[53] Batch [949]	Speed: 5790.954388 samples/sec	accuracy=0.511170	lr=0.247307
Epoch[53] Batch [999]	Speed: 5511.659521 samples/sec	accuracy=0.510914	lr=0.247095
Epoch[53] Batch [1049]	Speed: 4671.997087 samples/sec	accuracy=0.510679	lr=0.246883
Epoch[53] Batch [1099]	Speed: 5026.098929 samples/sec	accuracy=0.510653	lr=0.246670
Epoch[53] Batch [1149]	Speed: 5600.492065 samples/sec	accuracy=0.510503	lr=0.246458
Epoch[53] Batch [1199]	Speed: 5857.657772 samples/sec	accuracy=0.510479	lr=0.246246
Epoch[53] Batch [1249]	Speed: 5790.060508 samples/sec	accuracy=0.510188	lr=0.246033
[Epoch 53] training: accuracy=0.510198
[Epoch 53] speed: 5384 samples/sec	time cost: 247.700384
[Epoch 53] validation: err-top1=0.461436 err-top5=0.218072
Epoch[54] Batch [49]	Speed: 5543.496536 samples/sec	accuracy=0.521680	lr=0.245816
Epoch[54] Batch [99]	Speed: 5612.056006 samples/sec	accuracy=0.519473	lr=0.245604
Epoch[54] Batch [149]	Speed: 5774.676842 samples/sec	accuracy=0.519941	lr=0.245391
Epoch[54] Batch [199]	Speed: 5824.131762 samples/sec	accuracy=0.518433	lr=0.245178
Epoch[54] Batch [249]	Speed: 5325.541875 samples/sec	accuracy=0.518414	lr=0.244966
Epoch[54] Batch [299]	Speed: 5755.278671 samples/sec	accuracy=0.517669	lr=0.244753
Epoch[54] Batch [349]	Speed: 5919.365168 samples/sec	accuracy=0.517642	lr=0.244540
Epoch[54] Batch [399]	Speed: 5545.548195 samples/sec	accuracy=0.517090	lr=0.244327
Epoch[54] Batch [449]	Speed: 4733.654776 samples/sec	accuracy=0.516664	lr=0.244114
Epoch[54] Batch [499]	Speed: 5719.571821 samples/sec	accuracy=0.516365	lr=0.243901
Epoch[54] Batch [549]	Speed: 5340.779381 samples/sec	accuracy=0.515714	lr=0.243688
Epoch[54] Batch [599]	Speed: 5528.001382 samples/sec	accuracy=0.515125	lr=0.243475
Epoch[54] Batch [649]	Speed: 6043.990760 samples/sec	accuracy=0.514722	lr=0.243262
Epoch[54] Batch [699]	Speed: 5166.988096 samples/sec	accuracy=0.514548	lr=0.243049
Epoch[54] Batch [749]	Speed: 5023.330532 samples/sec	accuracy=0.514449	lr=0.242835
Epoch[54] Batch [799]	Speed: 5818.782805 samples/sec	accuracy=0.514272	lr=0.242622
Epoch[54] Batch [849]	Speed: 5350.861855 samples/sec	accuracy=0.514211	lr=0.242409
Epoch[54] Batch [899]	Speed: 5843.326763 samples/sec	accuracy=0.513925	lr=0.242195
Epoch[54] Batch [949]	Speed: 5880.229554 samples/sec	accuracy=0.513651	lr=0.241982
Epoch[54] Batch [999]	Speed: 5713.295391 samples/sec	accuracy=0.513599	lr=0.241768
Epoch[54] Batch [1049]	Speed: 5603.218085 samples/sec	accuracy=0.513498	lr=0.241555
Epoch[54] Batch [1099]	Speed: 5777.637909 samples/sec	accuracy=0.513293	lr=0.241341
Epoch[54] Batch [1149]	Speed: 5035.226450 samples/sec	accuracy=0.513038	lr=0.241127
Epoch[54] Batch [1199]	Speed: 4817.790987 samples/sec	accuracy=0.512837	lr=0.240914
Epoch[54] Batch [1249]	Speed: 5698.167325 samples/sec	accuracy=0.512681	lr=0.240700
[Epoch 54] training: accuracy=0.512674
[Epoch 54] speed: 5506 samples/sec	time cost: 244.664762
[Epoch 54] validation: err-top1=0.469248 err-top5=0.222975
Epoch[55] Batch [49]	Speed: 5382.387111 samples/sec	accuracy=0.521465	lr=0.240482
Epoch[55] Batch [99]	Speed: 4657.576688 samples/sec	accuracy=0.521641	lr=0.240268
Epoch[55] Batch [149]	Speed: 6081.284549 samples/sec	accuracy=0.519010	lr=0.240054
Epoch[55] Batch [199]	Speed: 5389.116897 samples/sec	accuracy=0.518940	lr=0.239840
Epoch[55] Batch [249]	Speed: 4882.021052 samples/sec	accuracy=0.518590	lr=0.239626
Epoch[55] Batch [299]	Speed: 5866.364077 samples/sec	accuracy=0.517754	lr=0.239412
Epoch[55] Batch [349]	Speed: 5437.115922 samples/sec	accuracy=0.516992	lr=0.239198
Epoch[55] Batch [399]	Speed: 5814.370754 samples/sec	accuracy=0.516558	lr=0.238984
Epoch[55] Batch [449]	Speed: 5640.710972 samples/sec	accuracy=0.516354	lr=0.238769
Epoch[55] Batch [499]	Speed: 5808.732830 samples/sec	accuracy=0.516387	lr=0.238555
Epoch[55] Batch [549]	Speed: 5515.451853 samples/sec	accuracy=0.516559	lr=0.238341
Epoch[55] Batch [599]	Speed: 5703.139207 samples/sec	accuracy=0.516146	lr=0.238126
Epoch[55] Batch [649]	Speed: 5260.506034 samples/sec	accuracy=0.515936	lr=0.237912
Epoch[55] Batch [699]	Speed: 5350.507896 samples/sec	accuracy=0.515688	lr=0.237698
Epoch[55] Batch [749]	Speed: 5407.903169 samples/sec	accuracy=0.515272	lr=0.237483
Epoch[55] Batch [799]	Speed: 4601.078818 samples/sec	accuracy=0.515203	lr=0.237269
Epoch[55] Batch [849]	Speed: 5789.656675 samples/sec	accuracy=0.514926	lr=0.237054
Epoch[55] Batch [899]	Speed: 5486.665011 samples/sec	accuracy=0.514793	lr=0.236839
Epoch[55] Batch [949]	Speed: 5257.070239 samples/sec	accuracy=0.514463	lr=0.236625
Epoch[55] Batch [999]	Speed: 5678.249073 samples/sec	accuracy=0.514206	lr=0.236410
Epoch[55] Batch [1049]	Speed: 6052.725000 samples/sec	accuracy=0.514299	lr=0.236195
Epoch[55] Batch [1099]	Speed: 5410.242742 samples/sec	accuracy=0.514433	lr=0.235981
Epoch[55] Batch [1149]	Speed: 4411.674347 samples/sec	accuracy=0.514567	lr=0.235766
Epoch[55] Batch [1199]	Speed: 5478.135283 samples/sec	accuracy=0.514534	lr=0.235551
Epoch[55] Batch [1249]	Speed: 5665.283003 samples/sec	accuracy=0.514291	lr=0.235336
[Epoch 55] training: accuracy=0.514306
[Epoch 55] speed: 5399 samples/sec	time cost: 247.975979
[Epoch 55] validation: err-top1=0.461894 err-top5=0.213768
Epoch[56] Batch [49]	Speed: 4982.979539 samples/sec	accuracy=0.522676	lr=0.235117
Epoch[56] Batch [99]	Speed: 4443.470111 samples/sec	accuracy=0.522061	lr=0.234902
Epoch[56] Batch [149]	Speed: 5489.039010 samples/sec	accuracy=0.520775	lr=0.234687
Epoch[56] Batch [199]	Speed: 5639.327327 samples/sec	accuracy=0.520835	lr=0.234472
Epoch[56] Batch [249]	Speed: 4808.837885 samples/sec	accuracy=0.519586	lr=0.234256
Epoch[56] Batch [299]	Speed: 5443.398260 samples/sec	accuracy=0.519574	lr=0.234041
Epoch[56] Batch [349]	Speed: 5987.648615 samples/sec	accuracy=0.519124	lr=0.233826
Epoch[56] Batch [399]	Speed: 5859.654884 samples/sec	accuracy=0.518875	lr=0.233611
Epoch[56] Batch [449]	Speed: 5892.551174 samples/sec	accuracy=0.518388	lr=0.233396
Epoch[56] Batch [499]	Speed: 5370.533759 samples/sec	accuracy=0.518215	lr=0.233180
Epoch[56] Batch [549]	Speed: 5240.560642 samples/sec	accuracy=0.518189	lr=0.232965
Epoch[56] Batch [599]	Speed: 4777.035531 samples/sec	accuracy=0.517754	lr=0.232749
Epoch[56] Batch [649]	Speed: 4842.746968 samples/sec	accuracy=0.517483	lr=0.232534
Epoch[56] Batch [699]	Speed: 4498.061898 samples/sec	accuracy=0.517319	lr=0.232319
Epoch[56] Batch [749]	Speed: 5863.074952 samples/sec	accuracy=0.517194	lr=0.232103
Epoch[56] Batch [799]	Speed: 5903.383103 samples/sec	accuracy=0.517041	lr=0.231887
Epoch[56] Batch [849]	Speed: 5896.883058 samples/sec	accuracy=0.516733	lr=0.231672
Epoch[56] Batch [899]	Speed: 4797.644317 samples/sec	accuracy=0.516465	lr=0.231456
Epoch[56] Batch [949]	Speed: 5894.629108 samples/sec	accuracy=0.516428	lr=0.231241
Epoch[56] Batch [999]	Speed: 5728.081049 samples/sec	accuracy=0.516159	lr=0.231025
Epoch[56] Batch [1049]	Speed: 5363.722735 samples/sec	accuracy=0.516011	lr=0.230809
Epoch[56] Batch [1099]	Speed: 4815.320671 samples/sec	accuracy=0.515896	lr=0.230593
Epoch[56] Batch [1149]	Speed: 5760.075738 samples/sec	accuracy=0.515667	lr=0.230377
Epoch[56] Batch [1199]	Speed: 5760.639561 samples/sec	accuracy=0.515719	lr=0.230162
Epoch[56] Batch [1249]	Speed: 4518.302564 samples/sec	accuracy=0.515543	lr=0.229946
[Epoch 56] training: accuracy=0.515547
[Epoch 56] speed: 5287 samples/sec	time cost: 253.048476
[Epoch 56] validation: err-top1=0.450295 err-top5=0.204759
Epoch[57] Batch [49]	Speed: 5287.618598 samples/sec	accuracy=0.526563	lr=0.229725
Epoch[57] Batch [99]	Speed: 5642.543594 samples/sec	accuracy=0.523701	lr=0.229510
Epoch[57] Batch [149]	Speed: 5222.019460 samples/sec	accuracy=0.523327	lr=0.229294
Epoch[57] Batch [199]	Speed: 5755.585937 samples/sec	accuracy=0.522451	lr=0.229077
Epoch[57] Batch [249]	Speed: 5789.885044 samples/sec	accuracy=0.521898	lr=0.228861
Epoch[57] Batch [299]	Speed: 5342.154073 samples/sec	accuracy=0.521556	lr=0.228645
Epoch[57] Batch [349]	Speed: 5216.335272 samples/sec	accuracy=0.521789	lr=0.228429
Epoch[57] Batch [399]	Speed: 5940.519958 samples/sec	accuracy=0.521304	lr=0.228213
Epoch[57] Batch [449]	Speed: 5235.392851 samples/sec	accuracy=0.521161	lr=0.227997
Epoch[57] Batch [499]	Speed: 5042.577756 samples/sec	accuracy=0.520379	lr=0.227781
Epoch[57] Batch [549]	Speed: 5291.689505 samples/sec	accuracy=0.520020	lr=0.227564
Epoch[57] Batch [599]	Speed: 5001.489624 samples/sec	accuracy=0.519513	lr=0.227348
Epoch[57] Batch [649]	Speed: 5438.464497 samples/sec	accuracy=0.519225	lr=0.227132
Epoch[57] Batch [699]	Speed: 4826.242590 samples/sec	accuracy=0.519247	lr=0.226915
Epoch[57] Batch [749]	Speed: 5545.232875 samples/sec	accuracy=0.519139	lr=0.226699
Epoch[57] Batch [799]	Speed: 5832.128442 samples/sec	accuracy=0.518738	lr=0.226482
Epoch[57] Batch [849]	Speed: 5832.405636 samples/sec	accuracy=0.518526	lr=0.226266
Epoch[57] Batch [899]	Speed: 5656.108582 samples/sec	accuracy=0.518406	lr=0.226049
Epoch[57] Batch [949]	Speed: 5781.319275 samples/sec	accuracy=0.518363	lr=0.225833
Epoch[57] Batch [999]	Speed: 5789.558964 samples/sec	accuracy=0.518160	lr=0.225616
Epoch[57] Batch [1049]	Speed: 5647.601817 samples/sec	accuracy=0.518050	lr=0.225400
Epoch[57] Batch [1099]	Speed: 5212.169935 samples/sec	accuracy=0.517867	lr=0.225183
Epoch[57] Batch [1149]	Speed: 4731.195959 samples/sec	accuracy=0.517744	lr=0.224967
Epoch[57] Batch [1199]	Speed: 5799.017376 samples/sec	accuracy=0.517725	lr=0.224750
Epoch[57] Batch [1249]	Speed: 5932.101077 samples/sec	accuracy=0.517547	lr=0.224533
[Epoch 57] training: accuracy=0.517569
[Epoch 57] speed: 5445 samples/sec	time cost: 245.944062
[Epoch 57] validation: err-top1=0.459941 err-top5=0.213110
Epoch[58] Batch [49]	Speed: 5082.033154 samples/sec	accuracy=0.527266	lr=0.224312
Epoch[58] Batch [99]	Speed: 5890.989847 samples/sec	accuracy=0.526396	lr=0.224095
Epoch[58] Batch [149]	Speed: 5321.683268 samples/sec	accuracy=0.526074	lr=0.223879
Epoch[58] Batch [199]	Speed: 5904.408907 samples/sec	accuracy=0.525215	lr=0.223662
Epoch[58] Batch [249]	Speed: 4897.719440 samples/sec	accuracy=0.525125	lr=0.223445
Epoch[58] Batch [299]	Speed: 5168.295167 samples/sec	accuracy=0.524115	lr=0.223228
Epoch[58] Batch [349]	Speed: 6144.795044 samples/sec	accuracy=0.523797	lr=0.223011
Epoch[58] Batch [399]	Speed: 5274.752201 samples/sec	accuracy=0.523398	lr=0.222794
Epoch[58] Batch [449]	Speed: 4796.156328 samples/sec	accuracy=0.522576	lr=0.222577
Epoch[58] Batch [499]	Speed: 4792.194789 samples/sec	accuracy=0.522391	lr=0.222360
Epoch[58] Batch [549]	Speed: 5701.478931 samples/sec	accuracy=0.521982	lr=0.222143
Epoch[58] Batch [599]	Speed: 5317.645103 samples/sec	accuracy=0.521553	lr=0.221926
Epoch[58] Batch [649]	Speed: 4411.163519 samples/sec	accuracy=0.521372	lr=0.221709
Epoch[58] Batch [699]	Speed: 5679.442043 samples/sec	accuracy=0.521250	lr=0.221492
Epoch[58] Batch [749]	Speed: 5292.283388 samples/sec	accuracy=0.521335	lr=0.221275
Epoch[58] Batch [799]	Speed: 6383.628299 samples/sec	accuracy=0.521140	lr=0.221058
Epoch[58] Batch [849]	Speed: 5416.427440 samples/sec	accuracy=0.520832	lr=0.220841
Epoch[58] Batch [899]	Speed: 5672.906386 samples/sec	accuracy=0.520634	lr=0.220623
Epoch[58] Batch [949]	Speed: 4944.704766 samples/sec	accuracy=0.520659	lr=0.220406
Epoch[58] Batch [999]	Speed: 5877.678618 samples/sec	accuracy=0.520531	lr=0.220189
Epoch[58] Batch [1049]	Speed: 4957.965707 samples/sec	accuracy=0.520441	lr=0.219972
Epoch[58] Batch [1099]	Speed: 5402.193400 samples/sec	accuracy=0.520218	lr=0.219754
Epoch[58] Batch [1149]	Speed: 5733.446627 samples/sec	accuracy=0.519963	lr=0.219537
Epoch[58] Batch [1199]	Speed: 5862.924646 samples/sec	accuracy=0.519749	lr=0.219320
Epoch[58] Batch [1249]	Speed: 5107.903403 samples/sec	accuracy=0.519567	lr=0.219102
[Epoch 58] training: accuracy=0.519587
[Epoch 58] speed: 5354 samples/sec	time cost: 250.518653
[Epoch 58] validation: err-top1=0.527181 err-top5=0.271505
Epoch[59] Batch [49]	Speed: 5032.908554 samples/sec	accuracy=0.531133	lr=0.218881
Epoch[59] Batch [99]	Speed: 5777.944925 samples/sec	accuracy=0.530469	lr=0.218663
Epoch[59] Batch [149]	Speed: 5557.064286 samples/sec	accuracy=0.529388	lr=0.218446
Epoch[59] Batch [199]	Speed: 5854.682453 samples/sec	accuracy=0.527021	lr=0.218228
Epoch[59] Batch [249]	Speed: 5442.365965 samples/sec	accuracy=0.526922	lr=0.218011
Epoch[59] Batch [299]	Speed: 4745.521432 samples/sec	accuracy=0.526439	lr=0.217793
Epoch[59] Batch [349]	Speed: 4712.490671 samples/sec	accuracy=0.525388	lr=0.217576
Epoch[59] Batch [399]	Speed: 5278.472446 samples/sec	accuracy=0.524448	lr=0.217358
Epoch[59] Batch [449]	Speed: 5036.886351 samples/sec	accuracy=0.524034	lr=0.217141
Epoch[59] Batch [499]	Speed: 5230.294858 samples/sec	accuracy=0.523648	lr=0.216923
Epoch[59] Batch [549]	Speed: 5205.646915 samples/sec	accuracy=0.523301	lr=0.216706
Epoch[59] Batch [599]	Speed: 5873.287187 samples/sec	accuracy=0.523234	lr=0.216488
Epoch[59] Batch [649]	Speed: 4709.800805 samples/sec	accuracy=0.523000	lr=0.216270
Epoch[59] Batch [699]	Speed: 5776.390283 samples/sec	accuracy=0.522520	lr=0.216053
Epoch[59] Batch [749]	Speed: 5100.855423 samples/sec	accuracy=0.522583	lr=0.215835
Epoch[59] Batch [799]	Speed: 5297.196438 samples/sec	accuracy=0.522279	lr=0.215617
Epoch[59] Batch [849]	Speed: 5168.739629 samples/sec	accuracy=0.522158	lr=0.215400
Epoch[59] Batch [899]	Speed: 5667.847932 samples/sec	accuracy=0.522012	lr=0.215182
Epoch[59] Batch [949]	Speed: 4783.175496 samples/sec	accuracy=0.521884	lr=0.214964
Epoch[59] Batch [999]	Speed: 5609.577653 samples/sec	accuracy=0.521807	lr=0.214746
Epoch[59] Batch [1049]	Speed: 5914.315683 samples/sec	accuracy=0.521644	lr=0.214529
Epoch[59] Batch [1099]	Speed: 5123.588573 samples/sec	accuracy=0.521567	lr=0.214311
Epoch[59] Batch [1149]	Speed: 5626.863424 samples/sec	accuracy=0.521699	lr=0.214093
Epoch[59] Batch [1199]	Speed: 5143.241253 samples/sec	accuracy=0.521647	lr=0.213875
Epoch[59] Batch [1249]	Speed: 5807.361021 samples/sec	accuracy=0.521483	lr=0.213657
[Epoch 59] training: accuracy=0.521462
[Epoch 59] speed: 5304 samples/sec	time cost: 251.655791
[Epoch 59] validation: err-top1=0.456114 err-top5=0.209662
Epoch[60] Batch [49]	Speed: 5265.459385 samples/sec	accuracy=0.532832	lr=0.213435
Epoch[60] Batch [99]	Speed: 5552.940699 samples/sec	accuracy=0.531104	lr=0.213217
Epoch[60] Batch [149]	Speed: 5056.522934 samples/sec	accuracy=0.529414	lr=0.212999
Epoch[60] Batch [199]	Speed: 4908.861908 samples/sec	accuracy=0.528643	lr=0.212781
Epoch[60] Batch [249]	Speed: 5890.496356 samples/sec	accuracy=0.527937	lr=0.212563
Epoch[60] Batch [299]	Speed: 4241.841288 samples/sec	accuracy=0.527783	lr=0.212345
Epoch[60] Batch [349]	Speed: 4760.379509 samples/sec	accuracy=0.526903	lr=0.212127
Epoch[60] Batch [399]	Speed: 5730.273475 samples/sec	accuracy=0.526438	lr=0.211909
Epoch[60] Batch [449]	Speed: 5942.856839 samples/sec	accuracy=0.526152	lr=0.211691
Epoch[60] Batch [499]	Speed: 5693.864060 samples/sec	accuracy=0.525715	lr=0.211473
Epoch[60] Batch [549]	Speed: 5754.755837 samples/sec	accuracy=0.525703	lr=0.211255
Epoch[60] Batch [599]	Speed: 5564.661006 samples/sec	accuracy=0.525457	lr=0.211037
Epoch[60] Batch [649]	Speed: 5544.495979 samples/sec	accuracy=0.524953	lr=0.210819
Epoch[60] Batch [699]	Speed: 5986.225383 samples/sec	accuracy=0.524537	lr=0.210601
Epoch[60] Batch [749]	Speed: 5656.173684 samples/sec	accuracy=0.524098	lr=0.210383
Epoch[60] Batch [799]	Speed: 5518.274359 samples/sec	accuracy=0.523790	lr=0.210165
Epoch[60] Batch [849]	Speed: 5233.812828 samples/sec	accuracy=0.523534	lr=0.209947
Epoch[60] Batch [899]	Speed: 5016.405225 samples/sec	accuracy=0.523372	lr=0.209729
Epoch[60] Batch [949]	Speed: 5843.804271 samples/sec	accuracy=0.523433	lr=0.209511
Epoch[60] Batch [999]	Speed: 6117.108259 samples/sec	accuracy=0.523339	lr=0.209293
Epoch[60] Batch [1049]	Speed: 5856.063937 samples/sec	accuracy=0.523057	lr=0.209075
Epoch[60] Batch [1099]	Speed: 5404.711525 samples/sec	accuracy=0.522973	lr=0.208856
Epoch[60] Batch [1149]	Speed: 5805.454005 samples/sec	accuracy=0.522835	lr=0.208638
Epoch[60] Batch [1199]	Speed: 5495.937040 samples/sec	accuracy=0.522866	lr=0.208420
Epoch[60] Batch [1249]	Speed: 4958.241815 samples/sec	accuracy=0.522753	lr=0.208202
[Epoch 60] training: accuracy=0.522752
[Epoch 60] speed: 5427 samples/sec	time cost: 246.266900
[Epoch 60] validation: err-top1=0.461874 err-top5=0.218690
Epoch[61] Batch [49]	Speed: 5638.434932 samples/sec	accuracy=0.534922	lr=0.207979
Epoch[61] Batch [99]	Speed: 5902.057389 samples/sec	accuracy=0.531992	lr=0.207761
Epoch[61] Batch [149]	Speed: 4940.560857 samples/sec	accuracy=0.531699	lr=0.207543
Epoch[61] Batch [199]	Speed: 4666.952738 samples/sec	accuracy=0.530713	lr=0.207325
Epoch[61] Batch [249]	Speed: 5946.211910 samples/sec	accuracy=0.529937	lr=0.207106
Epoch[61] Batch [299]	Speed: 5493.003569 samples/sec	accuracy=0.529137	lr=0.206888
Epoch[61] Batch [349]	Speed: 5287.021336 samples/sec	accuracy=0.528744	lr=0.206670
Epoch[61] Batch [399]	Speed: 5968.735672 samples/sec	accuracy=0.528359	lr=0.206452
Epoch[61] Batch [449]	Speed: 5642.568947 samples/sec	accuracy=0.527873	lr=0.206233
Epoch[61] Batch [499]	Speed: 5418.763457 samples/sec	accuracy=0.527738	lr=0.206015
Epoch[61] Batch [549]	Speed: 5872.040464 samples/sec	accuracy=0.527198	lr=0.205797
Epoch[61] Batch [599]	Speed: 5301.529202 samples/sec	accuracy=0.526820	lr=0.205579
Epoch[61] Batch [649]	Speed: 5447.250713 samples/sec	accuracy=0.526934	lr=0.205360
Epoch[61] Batch [699]	Speed: 5788.174197 samples/sec	accuracy=0.526196	lr=0.205142
Epoch[61] Batch [749]	Speed: 5712.472581 samples/sec	accuracy=0.526141	lr=0.204924
Epoch[61] Batch [799]	Speed: 5687.347001 samples/sec	accuracy=0.525868	lr=0.204705
Epoch[61] Batch [849]	Speed: 5894.449352 samples/sec	accuracy=0.525865	lr=0.204487
Epoch[61] Batch [899]	Speed: 5749.742854 samples/sec	accuracy=0.525999	lr=0.204269
Epoch[61] Batch [949]	Speed: 5134.719687 samples/sec	accuracy=0.525828	lr=0.204051
Epoch[61] Batch [999]	Speed: 4344.509036 samples/sec	accuracy=0.525783	lr=0.203832
Epoch[61] Batch [1049]	Speed: 5659.029351 samples/sec	accuracy=0.525607	lr=0.203614
Epoch[61] Batch [1099]	Speed: 5432.029885 samples/sec	accuracy=0.525531	lr=0.203396
Epoch[61] Batch [1149]	Speed: 5355.574039 samples/sec	accuracy=0.525178	lr=0.203177
Epoch[61] Batch [1199]	Speed: 4570.229670 samples/sec	accuracy=0.525177	lr=0.202959
Epoch[61] Batch [1249]	Speed: 4794.392124 samples/sec	accuracy=0.525177	lr=0.202740
[Epoch 61] training: accuracy=0.525170
[Epoch 61] speed: 5378 samples/sec	time cost: 248.927169
[Epoch 61] validation: err-top1=0.459403 err-top5=0.215202
Epoch[62] Batch [49]	Speed: 5318.415128 samples/sec	accuracy=0.532305	lr=0.202518
Epoch[62] Batch [99]	Speed: 5520.262546 samples/sec	accuracy=0.532490	lr=0.202299
Epoch[62] Batch [149]	Speed: 5673.737473 samples/sec	accuracy=0.532741	lr=0.202081
Epoch[62] Batch [199]	Speed: 5312.063532 samples/sec	accuracy=0.531313	lr=0.201863
Epoch[62] Batch [249]	Speed: 5352.786320 samples/sec	accuracy=0.530727	lr=0.201644
Epoch[62] Batch [299]	Speed: 4919.610880 samples/sec	accuracy=0.530215	lr=0.201426
Epoch[62] Batch [349]	Speed: 5526.182100 samples/sec	accuracy=0.530173	lr=0.201208
Epoch[62] Batch [399]	Speed: 5945.346164 samples/sec	accuracy=0.529700	lr=0.200989
Epoch[62] Batch [449]	Speed: 4741.796239 samples/sec	accuracy=0.529648	lr=0.200771
Epoch[62] Batch [499]	Speed: 5373.724148 samples/sec	accuracy=0.529742	lr=0.200552
Epoch[62] Batch [549]	Speed: 5025.306085 samples/sec	accuracy=0.529368	lr=0.200334
Epoch[62] Batch [599]	Speed: 4832.510978 samples/sec	accuracy=0.529621	lr=0.200116
Epoch[62] Batch [649]	Speed: 5093.443274 samples/sec	accuracy=0.529407	lr=0.199897
Epoch[62] Batch [699]	Speed: 4492.807255 samples/sec	accuracy=0.529058	lr=0.199679
Epoch[62] Batch [749]	Speed: 5710.685376 samples/sec	accuracy=0.528690	lr=0.199461
Epoch[62] Batch [799]	Speed: 5674.247487 samples/sec	accuracy=0.528491	lr=0.199242
Epoch[62] Batch [849]	Speed: 5921.108094 samples/sec	accuracy=0.528398	lr=0.199024
Epoch[62] Batch [899]	Speed: 5754.076913 samples/sec	accuracy=0.528269	lr=0.198806
Epoch[62] Batch [949]	Speed: 5617.383967 samples/sec	accuracy=0.528215	lr=0.198587
Epoch[62] Batch [999]	Speed: 5819.702924 samples/sec	accuracy=0.528155	lr=0.198369
Epoch[62] Batch [1049]	Speed: 5687.630035 samples/sec	accuracy=0.528047	lr=0.198150
Epoch[62] Batch [1099]	Speed: 6012.753148 samples/sec	accuracy=0.528026	lr=0.197932
Epoch[62] Batch [1149]	Speed: 5234.220024 samples/sec	accuracy=0.528055	lr=0.197714
Epoch[62] Batch [1199]	Speed: 6049.681882 samples/sec	accuracy=0.527964	lr=0.197495
Epoch[62] Batch [1249]	Speed: 5361.321621 samples/sec	accuracy=0.527761	lr=0.197277
[Epoch 62] training: accuracy=0.527749
[Epoch 62] speed: 5402 samples/sec	time cost: 247.433162
[Epoch 62] validation: err-top1=0.456334 err-top5=0.210320
Epoch[63] Batch [49]	Speed: 5728.255996 samples/sec	accuracy=0.536797	lr=0.197054
Epoch[63] Batch [99]	Speed: 5605.730330 samples/sec	accuracy=0.535889	lr=0.196836
Epoch[63] Batch [149]	Speed: 4734.676410 samples/sec	accuracy=0.535944	lr=0.196618
Epoch[63] Batch [199]	Speed: 5995.834930 samples/sec	accuracy=0.534116	lr=0.196399
Epoch[63] Batch [249]	Speed: 5885.029700 samples/sec	accuracy=0.534348	lr=0.196181
Epoch[63] Batch [299]	Speed: 5746.798000 samples/sec	accuracy=0.533724	lr=0.195963
Epoch[63] Batch [349]	Speed: 5714.519100 samples/sec	accuracy=0.533075	lr=0.195744
Epoch[63] Batch [399]	Speed: 5064.697010 samples/sec	accuracy=0.533137	lr=0.195526
Epoch[63] Batch [449]	Speed: 4656.744667 samples/sec	accuracy=0.532802	lr=0.195308
Epoch[63] Batch [499]	Speed: 5538.560144 samples/sec	accuracy=0.532279	lr=0.195089
Epoch[63] Batch [549]	Speed: 5148.619166 samples/sec	accuracy=0.532042	lr=0.194871
Epoch[63] Batch [599]	Speed: 5709.847226 samples/sec	accuracy=0.531951	lr=0.194653
Epoch[63] Batch [649]	Speed: 5620.296347 samples/sec	accuracy=0.531581	lr=0.194434
Epoch[63] Batch [699]	Speed: 5628.404255 samples/sec	accuracy=0.531476	lr=0.194216
Epoch[63] Batch [749]	Speed: 5734.443925 samples/sec	accuracy=0.531305	lr=0.193998
Epoch[63] Batch [799]	Speed: 5560.292904 samples/sec	accuracy=0.531241	lr=0.193780
Epoch[63] Batch [849]	Speed: 5916.629871 samples/sec	accuracy=0.531089	lr=0.193561
Epoch[63] Batch [899]	Speed: 5681.792957 samples/sec	accuracy=0.530890	lr=0.193343
Epoch[63] Batch [949]	Speed: 5556.453919 samples/sec	accuracy=0.530903	lr=0.193125
Epoch[63] Batch [999]	Speed: 5456.779356 samples/sec	accuracy=0.530445	lr=0.192907
Epoch[63] Batch [1049]	Speed: 5527.406914 samples/sec	accuracy=0.530276	lr=0.192688
Epoch[63] Batch [1099]	Speed: 5056.513885 samples/sec	accuracy=0.530100	lr=0.192470
Epoch[63] Batch [1149]	Speed: 4473.604031 samples/sec	accuracy=0.530053	lr=0.192252
Epoch[63] Batch [1199]	Speed: 5159.046750 samples/sec	accuracy=0.530089	lr=0.192034
Epoch[63] Batch [1249]	Speed: 5165.414794 samples/sec	accuracy=0.529930	lr=0.191816
[Epoch 63] training: accuracy=0.529928
[Epoch 63] speed: 5406 samples/sec	time cost: 246.347148
[Epoch 63] validation: err-top1=0.464457 err-top5=0.217590
Epoch[64] Batch [49]	Speed: 5070.771863 samples/sec	accuracy=0.541543	lr=0.191593
Epoch[64] Batch [99]	Speed: 5723.187318 samples/sec	accuracy=0.537373	lr=0.191375
Epoch[64] Batch [149]	Speed: 5681.595583 samples/sec	accuracy=0.536934	lr=0.191157
Epoch[64] Batch [199]	Speed: 5899.671761 samples/sec	accuracy=0.536367	lr=0.190939
Epoch[64] Batch [249]	Speed: 5291.636305 samples/sec	accuracy=0.535426	lr=0.190720
Epoch[64] Batch [299]	Speed: 4809.326173 samples/sec	accuracy=0.534701	lr=0.190502
Epoch[64] Batch [349]	Speed: 6091.877410 samples/sec	accuracy=0.534316	lr=0.190284
Epoch[64] Batch [399]	Speed: 5551.295676 samples/sec	accuracy=0.533950	lr=0.190066
Epoch[64] Batch [449]	Speed: 5754.936428 samples/sec	accuracy=0.533416	lr=0.189848
Epoch[64] Batch [499]	Speed: 5551.842904 samples/sec	accuracy=0.533469	lr=0.189630
Epoch[64] Batch [549]	Speed: 5236.393060 samples/sec	accuracy=0.533525	lr=0.189412
Epoch[64] Batch [599]	Speed: 5326.234398 samples/sec	accuracy=0.533250	lr=0.189194
Epoch[64] Batch [649]	Speed: 5762.690435 samples/sec	accuracy=0.533017	lr=0.188976
Epoch[64] Batch [699]	Speed: 5534.584063 samples/sec	accuracy=0.532833	lr=0.188758
Epoch[64] Batch [749]	Speed: 4861.493024 samples/sec	accuracy=0.532720	lr=0.188540
Epoch[64] Batch [799]	Speed: 4267.424485 samples/sec	accuracy=0.532251	lr=0.188322
Epoch[64] Batch [849]	Speed: 5963.379235 samples/sec	accuracy=0.532144	lr=0.188104
Epoch[64] Batch [899]	Speed: 4759.904168 samples/sec	accuracy=0.531962	lr=0.187886
Epoch[64] Batch [949]	Speed: 5491.405235 samples/sec	accuracy=0.531993	lr=0.187668
Epoch[64] Batch [999]	Speed: 5308.545647 samples/sec	accuracy=0.531951	lr=0.187450
Epoch[64] Batch [1049]	Speed: 5874.976238 samples/sec	accuracy=0.531899	lr=0.187232
Epoch[64] Batch [1099]	Speed: 5769.739633 samples/sec	accuracy=0.531804	lr=0.187014
Epoch[64] Batch [1149]	Speed: 5261.036742 samples/sec	accuracy=0.531523	lr=0.186796
Epoch[64] Batch [1199]	Speed: 4833.604129 samples/sec	accuracy=0.531352	lr=0.186578
Epoch[64] Batch [1249]	Speed: 4496.140258 samples/sec	accuracy=0.531316	lr=0.186360
[Epoch 64] training: accuracy=0.531340
[Epoch 64] speed: 5318 samples/sec	time cost: 252.086881
[Epoch 64] validation: err-top1=0.431820 err-top5=0.188576
Epoch[65] Batch [49]	Speed: 5135.476327 samples/sec	accuracy=0.541660	lr=0.186138
Epoch[65] Batch [99]	Speed: 5618.508426 samples/sec	accuracy=0.540762	lr=0.185920
Epoch[65] Batch [149]	Speed: 5552.321476 samples/sec	accuracy=0.538372	lr=0.185702
Epoch[65] Batch [199]	Speed: 5747.472901 samples/sec	accuracy=0.537847	lr=0.185485
Epoch[65] Batch [249]	Speed: 5761.135489 samples/sec	accuracy=0.537758	lr=0.185267
Epoch[65] Batch [299]	Speed: 4562.092211 samples/sec	accuracy=0.537064	lr=0.185049
Epoch[65] Batch [349]	Speed: 5402.871748 samples/sec	accuracy=0.536931	lr=0.184831
Epoch[65] Batch [399]	Speed: 5457.327246 samples/sec	accuracy=0.536597	lr=0.184614
Epoch[65] Batch [449]	Speed: 5794.265315 samples/sec	accuracy=0.536530	lr=0.184396
Epoch[65] Batch [499]	Speed: 5035.717633 samples/sec	accuracy=0.536154	lr=0.184178
Epoch[65] Batch [549]	Speed: 4747.481258 samples/sec	accuracy=0.535929	lr=0.183960
Epoch[65] Batch [599]	Speed: 4387.126522 samples/sec	accuracy=0.535469	lr=0.183743
Epoch[65] Batch [649]	Speed: 5561.568464 samples/sec	accuracy=0.535179	lr=0.183525
Epoch[65] Batch [699]	Speed: 5113.904726 samples/sec	accuracy=0.535091	lr=0.183308
Epoch[65] Batch [749]	Speed: 5671.780873 samples/sec	accuracy=0.534823	lr=0.183090
Epoch[65] Batch [799]	Speed: 4885.360137 samples/sec	accuracy=0.534878	lr=0.182872
Epoch[65] Batch [849]	Speed: 5066.248631 samples/sec	accuracy=0.534712	lr=0.182655
Epoch[65] Batch [899]	Speed: 5173.578551 samples/sec	accuracy=0.534436	lr=0.182437
Epoch[65] Batch [949]	Speed: 5107.779604 samples/sec	accuracy=0.534277	lr=0.182220
Epoch[65] Batch [999]	Speed: 5230.851086 samples/sec	accuracy=0.534160	lr=0.182002
Epoch[65] Batch [1049]	Speed: 5374.357032 samples/sec	accuracy=0.534041	lr=0.181785
Epoch[65] Batch [1099]	Speed: 5963.214636 samples/sec	accuracy=0.533687	lr=0.181567
Epoch[65] Batch [1149]	Speed: 5896.382589 samples/sec	accuracy=0.533603	lr=0.181350
Epoch[65] Batch [1199]	Speed: 4394.431923 samples/sec	accuracy=0.533612	lr=0.181132
Epoch[65] Batch [1249]	Speed: 5440.136756 samples/sec	accuracy=0.533684	lr=0.180915
[Epoch 65] training: accuracy=0.533712
[Epoch 65] speed: 5239 samples/sec	time cost: 255.064313
[Epoch 65] validation: err-top1=0.446608 err-top5=0.203683
Epoch[66] Batch [49]	Speed: 4727.994788 samples/sec	accuracy=0.540371	lr=0.180693
Epoch[66] Batch [99]	Speed: 5553.410988 samples/sec	accuracy=0.541143	lr=0.180476
Epoch[66] Batch [149]	Speed: 4316.880993 samples/sec	accuracy=0.541673	lr=0.180259
Epoch[66] Batch [199]	Speed: 5823.333569 samples/sec	accuracy=0.539746	lr=0.180041
Epoch[66] Batch [249]	Speed: 5631.308731 samples/sec	accuracy=0.539730	lr=0.179824
Epoch[66] Batch [299]	Speed: 5880.555138 samples/sec	accuracy=0.539714	lr=0.179607
Epoch[66] Batch [349]	Speed: 5374.077286 samples/sec	accuracy=0.539129	lr=0.179390
Epoch[66] Batch [399]	Speed: 5940.435986 samples/sec	accuracy=0.539026	lr=0.179172
Epoch[66] Batch [449]	Speed: 5597.533344 samples/sec	accuracy=0.539191	lr=0.178955
Epoch[66] Batch [499]	Speed: 5318.899751 samples/sec	accuracy=0.538781	lr=0.178738
Epoch[66] Batch [549]	Speed: 4940.333426 samples/sec	accuracy=0.538768	lr=0.178521
Epoch[66] Batch [599]	Speed: 5047.232106 samples/sec	accuracy=0.538421	lr=0.178304
Epoch[66] Batch [649]	Speed: 5897.921021 samples/sec	accuracy=0.538000	lr=0.178087
Epoch[66] Batch [699]	Speed: 5510.413396 samples/sec	accuracy=0.538069	lr=0.177870
Epoch[66] Batch [749]	Speed: 5127.029658 samples/sec	accuracy=0.537699	lr=0.177653
Epoch[66] Batch [799]	Speed: 4917.357647 samples/sec	accuracy=0.537179	lr=0.177436
Epoch[66] Batch [849]	Speed: 5437.809541 samples/sec	accuracy=0.536801	lr=0.177219
Epoch[66] Batch [899]	Speed: 4893.409670 samples/sec	accuracy=0.536670	lr=0.177002
Epoch[66] Batch [949]	Speed: 5686.761442 samples/sec	accuracy=0.536182	lr=0.176785
Epoch[66] Batch [999]	Speed: 5337.019765 samples/sec	accuracy=0.536052	lr=0.176568
Epoch[66] Batch [1049]	Speed: 4914.067124 samples/sec	accuracy=0.536043	lr=0.176351
Epoch[66] Batch [1099]	Speed: 5923.830049 samples/sec	accuracy=0.536066	lr=0.176134
Epoch[66] Batch [1149]	Speed: 5403.455500 samples/sec	accuracy=0.536008	lr=0.175918
Epoch[66] Batch [1199]	Speed: 4614.144464 samples/sec	accuracy=0.535929	lr=0.175701
Epoch[66] Batch [1249]	Speed: 5337.566156 samples/sec	accuracy=0.535787	lr=0.175484
[Epoch 66] training: accuracy=0.535797
[Epoch 66] speed: 5282 samples/sec	time cost: 252.179198
[Epoch 66] validation: err-top1=0.455716 err-top5=0.208685
Epoch[67] Batch [49]	Speed: 5142.508800 samples/sec	accuracy=0.540078	lr=0.175263
Epoch[67] Batch [99]	Speed: 5494.174079 samples/sec	accuracy=0.541924	lr=0.175046
Epoch[67] Batch [149]	Speed: 5955.009754 samples/sec	accuracy=0.542363	lr=0.174830
Epoch[67] Batch [199]	Speed: 5031.118668 samples/sec	accuracy=0.541929	lr=0.174613
Epoch[67] Batch [249]	Speed: 5591.244713 samples/sec	accuracy=0.542191	lr=0.174397
Epoch[67] Batch [299]	Speed: 5126.321271 samples/sec	accuracy=0.541940	lr=0.174180
Epoch[67] Batch [349]	Speed: 5016.034142 samples/sec	accuracy=0.541713	lr=0.173964
Epoch[67] Batch [399]	Speed: 5729.377746 samples/sec	accuracy=0.541543	lr=0.173747
Epoch[67] Batch [449]	Speed: 5453.325232 samples/sec	accuracy=0.540940	lr=0.173531
Epoch[67] Batch [499]	Speed: 4945.732176 samples/sec	accuracy=0.540637	lr=0.173314
Epoch[67] Batch [549]	Speed: 4895.312903 samples/sec	accuracy=0.540517	lr=0.173098
Epoch[67] Batch [599]	Speed: 5682.040258 samples/sec	accuracy=0.540101	lr=0.172881
Epoch[67] Batch [649]	Speed: 5353.812805 samples/sec	accuracy=0.540009	lr=0.172665
Epoch[67] Batch [699]	Speed: 5334.118429 samples/sec	accuracy=0.539406	lr=0.172449
Epoch[67] Batch [749]	Speed: 5819.452800 samples/sec	accuracy=0.539220	lr=0.172232
Epoch[67] Batch [799]	Speed: 6035.666742 samples/sec	accuracy=0.538994	lr=0.172016
Epoch[67] Batch [849]	Speed: 4868.317632 samples/sec	accuracy=0.538719	lr=0.171800
Epoch[67] Batch [899]	Speed: 4214.167958 samples/sec	accuracy=0.538580	lr=0.171584
Epoch[67] Batch [949]	Speed: 5069.582941 samples/sec	accuracy=0.538440	lr=0.171368
Epoch[67] Batch [999]	Speed: 4721.622553 samples/sec	accuracy=0.538283	lr=0.171152
Epoch[67] Batch [1049]	Speed: 5196.699157 samples/sec	accuracy=0.538465	lr=0.170935
Epoch[67] Batch [1099]	Speed: 5382.490044 samples/sec	accuracy=0.538370	lr=0.170719
Epoch[67] Batch [1149]	Speed: 4613.485470 samples/sec	accuracy=0.538192	lr=0.170503
Epoch[67] Batch [1199]	Speed: 4890.621968 samples/sec	accuracy=0.538164	lr=0.170287
Epoch[67] Batch [1249]	Speed: 4597.591726 samples/sec	accuracy=0.537851	lr=0.170072
[Epoch 67] training: accuracy=0.537878
[Epoch 67] speed: 5161 samples/sec	time cost: 257.773913
[Epoch 67] validation: err-top1=0.433893 err-top5=0.193658
Epoch[68] Batch [49]	Speed: 5423.930975 samples/sec	accuracy=0.546406	lr=0.169851
Epoch[68] Batch [99]	Speed: 4556.948837 samples/sec	accuracy=0.547842	lr=0.169635
Epoch[68] Batch [149]	Speed: 5011.678637 samples/sec	accuracy=0.545586	lr=0.169420
Epoch[68] Batch [199]	Speed: 5767.875987 samples/sec	accuracy=0.545552	lr=0.169204
Epoch[68] Batch [249]	Speed: 5455.655354 samples/sec	accuracy=0.544605	lr=0.168988
Epoch[68] Batch [299]	Speed: 5480.755913 samples/sec	accuracy=0.544294	lr=0.168772
Epoch[68] Batch [349]	Speed: 5686.097110 samples/sec	accuracy=0.544088	lr=0.168557
Epoch[68] Batch [399]	Speed: 4468.637589 samples/sec	accuracy=0.543330	lr=0.168341
Epoch[68] Batch [449]	Speed: 5109.589205 samples/sec	accuracy=0.542804	lr=0.168126
Epoch[68] Batch [499]	Speed: 5578.299789 samples/sec	accuracy=0.542508	lr=0.167910
Epoch[68] Batch [549]	Speed: 4971.368054 samples/sec	accuracy=0.542100	lr=0.167694
Epoch[68] Batch [599]	Speed: 4326.780120 samples/sec	accuracy=0.542028	lr=0.167479
Epoch[68] Batch [649]	Speed: 5566.474117 samples/sec	accuracy=0.541686	lr=0.167263
Epoch[68] Batch [699]	Speed: 6081.565438 samples/sec	accuracy=0.541932	lr=0.167048
Epoch[68] Batch [749]	Speed: 5335.540333 samples/sec	accuracy=0.541690	lr=0.166833
Epoch[68] Batch [799]	Speed: 5419.066336 samples/sec	accuracy=0.541902	lr=0.166617
Epoch[68] Batch [849]	Speed: 5221.136567 samples/sec	accuracy=0.541429	lr=0.166402
Epoch[68] Batch [899]	Speed: 5160.451241 samples/sec	accuracy=0.541238	lr=0.166187
Epoch[68] Batch [949]	Speed: 5235.756253 samples/sec	accuracy=0.541036	lr=0.165972
Epoch[68] Batch [999]	Speed: 5225.861884 samples/sec	accuracy=0.540942	lr=0.165756
Epoch[68] Batch [1049]	Speed: 5754.547656 samples/sec	accuracy=0.540925	lr=0.165541
Epoch[68] Batch [1099]	Speed: 5553.760417 samples/sec	accuracy=0.540619	lr=0.165326
Epoch[68] Batch [1149]	Speed: 5420.217173 samples/sec	accuracy=0.540614	lr=0.165111
Epoch[68] Batch [1199]	Speed: 5350.202635 samples/sec	accuracy=0.540612	lr=0.164896
Epoch[68] Batch [1249]	Speed: 4903.831754 samples/sec	accuracy=0.540366	lr=0.164681
[Epoch 68] training: accuracy=0.540376
[Epoch 68] speed: 5243 samples/sec	time cost: 255.668327
[Epoch 68] validation: err-top1=0.452129 err-top5=0.204022
Epoch[69] Batch [49]	Speed: 5799.841184 samples/sec	accuracy=0.549883	lr=0.164462
Epoch[69] Batch [99]	Speed: 5876.079337 samples/sec	accuracy=0.548477	lr=0.164247
Epoch[69] Batch [149]	Speed: 5698.282842 samples/sec	accuracy=0.546875	lr=0.164032
Epoch[69] Batch [199]	Speed: 5864.328760 samples/sec	accuracy=0.547041	lr=0.163818
Epoch[69] Batch [249]	Speed: 5251.267027 samples/sec	accuracy=0.546230	lr=0.163603
Epoch[69] Batch [299]	Speed: 5833.965221 samples/sec	accuracy=0.545944	lr=0.163388
Epoch[69] Batch [349]	Speed: 5606.390211 samples/sec	accuracy=0.545360	lr=0.163173
Epoch[69] Batch [399]	Speed: 5594.719129 samples/sec	accuracy=0.545134	lr=0.162959
Epoch[69] Batch [449]	Speed: 5904.711360 samples/sec	accuracy=0.544633	lr=0.162744
Epoch[69] Batch [499]	Speed: 5788.017567 samples/sec	accuracy=0.544594	lr=0.162530
Epoch[69] Batch [549]	Speed: 5589.689821 samples/sec	accuracy=0.544203	lr=0.162315
Epoch[69] Batch [599]	Speed: 5071.052655 samples/sec	accuracy=0.544238	lr=0.162101
Epoch[69] Batch [649]	Speed: 5726.199628 samples/sec	accuracy=0.543857	lr=0.161886
Epoch[69] Batch [699]	Speed: 5635.648811 samples/sec	accuracy=0.543581	lr=0.161672
Epoch[69] Batch [749]	Speed: 5956.246200 samples/sec	accuracy=0.543289	lr=0.161458
Epoch[69] Batch [799]	Speed: 5854.269395 samples/sec	accuracy=0.543336	lr=0.161244
Epoch[69] Batch [849]	Speed: 5916.389927 samples/sec	accuracy=0.543201	lr=0.161029
Epoch[69] Batch [899]	Speed: 5534.968219 samples/sec	accuracy=0.543190	lr=0.160815
Epoch[69] Batch [949]	Speed: 5685.350600 samples/sec	accuracy=0.543051	lr=0.160601
Epoch[69] Batch [999]	Speed: 5546.305138 samples/sec	accuracy=0.542830	lr=0.160387
Epoch[69] Batch [1049]	Speed: 5711.536988 samples/sec	accuracy=0.542687	lr=0.160173
Epoch[69] Batch [1099]	Speed: 5467.318636 samples/sec	accuracy=0.542276	lr=0.159959
Epoch[69] Batch [1149]	Speed: 5158.283768 samples/sec	accuracy=0.541843	lr=0.159745
Epoch[69] Batch [1199]	Speed: 5684.096621 samples/sec	accuracy=0.541796	lr=0.159531
Epoch[69] Batch [1249]	Speed: 5711.449188 samples/sec	accuracy=0.541632	lr=0.159317
[Epoch 69] training: accuracy=0.541635
[Epoch 69] speed: 5644 samples/sec	time cost: 237.343948
[Epoch 69] validation: err-top1=0.447713 err-top5=0.206238
Epoch[70] Batch [49]	Speed: 5243.700480 samples/sec	accuracy=0.552266	lr=0.159099
Epoch[70] Batch [99]	Speed: 4604.428742 samples/sec	accuracy=0.551279	lr=0.158886
Epoch[70] Batch [149]	Speed: 5130.147513 samples/sec	accuracy=0.551380	lr=0.158672
Epoch[70] Batch [199]	Speed: 5568.725633 samples/sec	accuracy=0.550698	lr=0.158458
Epoch[70] Batch [249]	Speed: 5398.493689 samples/sec	accuracy=0.549723	lr=0.158245
Epoch[70] Batch [299]	Speed: 5659.698560 samples/sec	accuracy=0.549473	lr=0.158031
Epoch[70] Batch [349]	Speed: 5530.092994 samples/sec	accuracy=0.548976	lr=0.157818
Epoch[70] Batch [399]	Speed: 5641.331248 samples/sec	accuracy=0.548538	lr=0.157604
Epoch[70] Batch [449]	Speed: 4743.590886 samples/sec	accuracy=0.547878	lr=0.157391
Epoch[70] Batch [499]	Speed: 5661.168033 samples/sec	accuracy=0.547549	lr=0.157177
Epoch[70] Batch [549]	Speed: 5914.035536 samples/sec	accuracy=0.546911	lr=0.156964
Epoch[70] Batch [599]	Speed: 4328.309389 samples/sec	accuracy=0.546455	lr=0.156751
Epoch[70] Batch [649]	Speed: 4201.449336 samples/sec	accuracy=0.546098	lr=0.156538
Epoch[70] Batch [699]	Speed: 5364.280102 samples/sec	accuracy=0.545984	lr=0.156325
Epoch[70] Batch [749]	Speed: 5312.135672 samples/sec	accuracy=0.545953	lr=0.156112
Epoch[70] Batch [799]	Speed: 5884.585420 samples/sec	accuracy=0.545822	lr=0.155899
Epoch[70] Batch [849]	Speed: 5039.733699 samples/sec	accuracy=0.545708	lr=0.155686
Epoch[70] Batch [899]	Speed: 5813.674229 samples/sec	accuracy=0.545844	lr=0.155473
Epoch[70] Batch [949]	Speed: 5759.567171 samples/sec	accuracy=0.545560	lr=0.155260
Epoch[70] Batch [999]	Speed: 4970.565460 samples/sec	accuracy=0.545321	lr=0.155047
Epoch[70] Batch [1049]	Speed: 5054.899922 samples/sec	accuracy=0.545135	lr=0.154834
Epoch[70] Batch [1099]	Speed: 5659.683047 samples/sec	accuracy=0.545120	lr=0.154622
Epoch[70] Batch [1149]	Speed: 5930.959810 samples/sec	accuracy=0.544972	lr=0.154409
Epoch[70] Batch [1199]	Speed: 5604.781684 samples/sec	accuracy=0.544846	lr=0.154196
Epoch[70] Batch [1249]	Speed: 5803.892997 samples/sec	accuracy=0.544664	lr=0.153984
[Epoch 70] training: accuracy=0.544667
[Epoch 70] speed: 5300 samples/sec	time cost: 252.573308
[Epoch 70] validation: err-top1=0.441307 err-top5=0.200953
Epoch[71] Batch [49]	Speed: 5319.810486 samples/sec	accuracy=0.550195	lr=0.153767
Epoch[71] Batch [99]	Speed: 5093.605523 samples/sec	accuracy=0.549092	lr=0.153555
Epoch[71] Batch [149]	Speed: 5345.929047 samples/sec	accuracy=0.549499	lr=0.153342
Epoch[71] Batch [199]	Speed: 5402.240013 samples/sec	accuracy=0.549536	lr=0.153130
Epoch[71] Batch [249]	Speed: 5772.344183 samples/sec	accuracy=0.549871	lr=0.152918
Epoch[71] Batch [299]	Speed: 5126.169779 samples/sec	accuracy=0.549811	lr=0.152705
Epoch[71] Batch [349]	Speed: 5410.198717 samples/sec	accuracy=0.549545	lr=0.152493
Epoch[71] Batch [399]	Speed: 5977.650593 samples/sec	accuracy=0.549656	lr=0.152281
Epoch[71] Batch [449]	Speed: 5980.624989 samples/sec	accuracy=0.549395	lr=0.152069
Epoch[71] Batch [499]	Speed: 5972.876476 samples/sec	accuracy=0.548840	lr=0.151857
Epoch[71] Batch [549]	Speed: 5682.233604 samples/sec	accuracy=0.548968	lr=0.151645
Epoch[71] Batch [599]	Speed: 5268.469636 samples/sec	accuracy=0.548468	lr=0.151433
Epoch[71] Batch [649]	Speed: 5524.652662 samples/sec	accuracy=0.548164	lr=0.151222
Epoch[71] Batch [699]	Speed: 4933.091603 samples/sec	accuracy=0.548164	lr=0.151010
Epoch[71] Batch [749]	Speed: 5348.639820 samples/sec	accuracy=0.548094	lr=0.150798
Epoch[71] Batch [799]	Speed: 5163.909750 samples/sec	accuracy=0.548008	lr=0.150587
Epoch[71] Batch [849]	Speed: 5712.549016 samples/sec	accuracy=0.547876	lr=0.150375
Epoch[71] Batch [899]	Speed: 5568.996983 samples/sec	accuracy=0.547618	lr=0.150163
Epoch[71] Batch [949]	Speed: 5366.869491 samples/sec	accuracy=0.547295	lr=0.149952
Epoch[71] Batch [999]	Speed: 5300.495061 samples/sec	accuracy=0.547121	lr=0.149741
Epoch[71] Batch [1049]	Speed: 5673.892777 samples/sec	accuracy=0.547010	lr=0.149529
Epoch[71] Batch [1099]	Speed: 5746.426012 samples/sec	accuracy=0.546843	lr=0.149318
Epoch[71] Batch [1149]	Speed: 5597.409184 samples/sec	accuracy=0.546727	lr=0.149107
Epoch[71] Batch [1199]	Speed: 5028.210049 samples/sec	accuracy=0.546555	lr=0.148896
Epoch[71] Batch [1249]	Speed: 4971.248368 samples/sec	accuracy=0.546610	lr=0.148685
[Epoch 71] training: accuracy=0.546634
[Epoch 71] speed: 5416 samples/sec	time cost: 247.427164
[Epoch 71] validation: err-top1=0.462651 err-top5=0.217973
Epoch[72] Batch [49]	Speed: 5507.663618 samples/sec	accuracy=0.553809	lr=0.148469
Epoch[72] Batch [99]	Speed: 5062.820469 samples/sec	accuracy=0.553750	lr=0.148258
Epoch[72] Batch [149]	Speed: 4525.846475 samples/sec	accuracy=0.553672	lr=0.148047
Epoch[72] Batch [199]	Speed: 5338.432069 samples/sec	accuracy=0.553716	lr=0.147837
Epoch[72] Batch [249]	Speed: 5263.475533 samples/sec	accuracy=0.553605	lr=0.147626
Epoch[72] Batch [299]	Speed: 5718.267377 samples/sec	accuracy=0.552972	lr=0.147415
Epoch[72] Batch [349]	Speed: 3784.723446 samples/sec	accuracy=0.552662	lr=0.147204
Epoch[72] Batch [399]	Speed: 5084.675823 samples/sec	accuracy=0.551841	lr=0.146994
Epoch[72] Batch [449]	Speed: 5582.610904 samples/sec	accuracy=0.551949	lr=0.146783
Epoch[72] Batch [499]	Speed: 5794.201685 samples/sec	accuracy=0.551697	lr=0.146573
Epoch[72] Batch [549]	Speed: 4928.024813 samples/sec	accuracy=0.551582	lr=0.146362
Epoch[72] Batch [599]	Speed: 5888.107148 samples/sec	accuracy=0.551058	lr=0.146152
Epoch[72] Batch [649]	Speed: 5731.498809 samples/sec	accuracy=0.550753	lr=0.145942
Epoch[72] Batch [699]	Speed: 5790.672844 samples/sec	accuracy=0.550491	lr=0.145732
Epoch[72] Batch [749]	Speed: 5864.888992 samples/sec	accuracy=0.550276	lr=0.145521
Epoch[72] Batch [799]	Speed: 5860.522242 samples/sec	accuracy=0.550135	lr=0.145311
Epoch[72] Batch [849]	Speed: 5729.910655 samples/sec	accuracy=0.549992	lr=0.145101
Epoch[72] Batch [899]	Speed: 5327.771456 samples/sec	accuracy=0.549843	lr=0.144891
Epoch[72] Batch [949]	Speed: 5886.986618 samples/sec	accuracy=0.549738	lr=0.144682
Epoch[72] Batch [999]	Speed: 4095.720175 samples/sec	accuracy=0.549669	lr=0.144472
Epoch[72] Batch [1049]	Speed: 5647.498891 samples/sec	accuracy=0.549424	lr=0.144262
Epoch[72] Batch [1099]	Speed: 5972.023043 samples/sec	accuracy=0.549123	lr=0.144052
Epoch[72] Batch [1149]	Speed: 6027.108155 samples/sec	accuracy=0.549063	lr=0.143843
Epoch[72] Batch [1199]	Speed: 5313.054607 samples/sec	accuracy=0.548916	lr=0.143633
Epoch[72] Batch [1249]	Speed: 5338.772753 samples/sec	accuracy=0.548987	lr=0.143424
[Epoch 72] training: accuracy=0.548999
[Epoch 72] speed: 5325 samples/sec	time cost: 252.378268
[Epoch 72] validation: err-top1=0.422592 err-top5=0.185866
Epoch[73] Batch [49]	Speed: 5139.983208 samples/sec	accuracy=0.557344	lr=0.143210
Epoch[73] Batch [99]	Speed: 5843.215307 samples/sec	accuracy=0.556777	lr=0.143001
Epoch[73] Batch [149]	Speed: 5627.585068 samples/sec	accuracy=0.557233	lr=0.142791
Epoch[73] Batch [199]	Speed: 5804.139747 samples/sec	accuracy=0.555947	lr=0.142582
Epoch[73] Batch [249]	Speed: 5883.558109 samples/sec	accuracy=0.555613	lr=0.142373
Epoch[73] Batch [299]	Speed: 5281.720504 samples/sec	accuracy=0.555195	lr=0.142164
Epoch[73] Batch [349]	Speed: 5734.461994 samples/sec	accuracy=0.554453	lr=0.141955
Epoch[73] Batch [399]	Speed: 5650.292432 samples/sec	accuracy=0.554031	lr=0.141746
Epoch[73] Batch [449]	Speed: 5731.569329 samples/sec	accuracy=0.554173	lr=0.141537
Epoch[73] Batch [499]	Speed: 5466.834007 samples/sec	accuracy=0.553918	lr=0.141328
Epoch[73] Batch [549]	Speed: 5670.962938 samples/sec	accuracy=0.553505	lr=0.141120
Epoch[73] Batch [599]	Speed: 5480.240928 samples/sec	accuracy=0.553441	lr=0.140911
Epoch[73] Batch [649]	Speed: 5806.598975 samples/sec	accuracy=0.553000	lr=0.140702
Epoch[73] Batch [699]	Speed: 5562.314083 samples/sec	accuracy=0.552893	lr=0.140494
Epoch[73] Batch [749]	Speed: 5540.981561 samples/sec	accuracy=0.552753	lr=0.140285
Epoch[73] Batch [799]	Speed: 5705.577089 samples/sec	accuracy=0.552452	lr=0.140077
Epoch[73] Batch [849]	Speed: 5031.708081 samples/sec	accuracy=0.552224	lr=0.139869
Epoch[73] Batch [899]	Speed: 5093.452334 samples/sec	accuracy=0.552089	lr=0.139661
Epoch[73] Batch [949]	Speed: 5227.914075 samples/sec	accuracy=0.552044	lr=0.139452
Epoch[73] Batch [999]	Speed: 5261.557501 samples/sec	accuracy=0.551808	lr=0.139244
Epoch[73] Batch [1049]	Speed: 5548.580649 samples/sec	accuracy=0.551729	lr=0.139036
Epoch[73] Batch [1099]	Speed: 5957.406475 samples/sec	accuracy=0.551710	lr=0.138828
Epoch[73] Batch [1149]	Speed: 5211.950332 samples/sec	accuracy=0.551508	lr=0.138620
Epoch[73] Batch [1199]	Speed: 5555.612853 samples/sec	accuracy=0.551420	lr=0.138413
Epoch[73] Batch [1249]	Speed: 5350.239558 samples/sec	accuracy=0.551196	lr=0.138205
[Epoch 73] training: accuracy=0.551225
[Epoch 73] speed: 5507 samples/sec	time cost: 242.311755
[Epoch 73] validation: err-top1=0.425084 err-top5=0.186623
Epoch[74] Batch [49]	Speed: 5113.758107 samples/sec	accuracy=0.561836	lr=0.137993
Epoch[74] Batch [99]	Speed: 4833.616858 samples/sec	accuracy=0.559092	lr=0.137786
Epoch[74] Batch [149]	Speed: 5706.030378 samples/sec	accuracy=0.558854	lr=0.137578
Epoch[74] Batch [199]	Speed: 5623.364739 samples/sec	accuracy=0.558750	lr=0.137371
Epoch[74] Batch [249]	Speed: 5392.255617 samples/sec	accuracy=0.558094	lr=0.137163
Epoch[74] Batch [299]	Speed: 5611.568841 samples/sec	accuracy=0.557832	lr=0.136956
Epoch[74] Batch [349]	Speed: 5020.119197 samples/sec	accuracy=0.557112	lr=0.136749
Epoch[74] Batch [399]	Speed: 5639.295192 samples/sec	accuracy=0.556860	lr=0.136542
Epoch[74] Batch [449]	Speed: 5474.767130 samples/sec	accuracy=0.555751	lr=0.136335
Epoch[74] Batch [499]	Speed: 4591.519784 samples/sec	accuracy=0.556027	lr=0.136128
Epoch[74] Batch [549]	Speed: 5720.357820 samples/sec	accuracy=0.556035	lr=0.135921
Epoch[74] Batch [599]	Speed: 5056.335179 samples/sec	accuracy=0.556076	lr=0.135714
Epoch[74] Batch [649]	Speed: 5392.218654 samples/sec	accuracy=0.556037	lr=0.135507
Epoch[74] Batch [699]	Speed: 4755.912104 samples/sec	accuracy=0.556097	lr=0.135301
Epoch[74] Batch [749]	Speed: 5330.311687 samples/sec	accuracy=0.555650	lr=0.135094
Epoch[74] Batch [799]	Speed: 5150.916893 samples/sec	accuracy=0.555514	lr=0.134887
Epoch[74] Batch [849]	Speed: 5399.251471 samples/sec	accuracy=0.555315	lr=0.134681
Epoch[74] Batch [899]	Speed: 5432.602226 samples/sec	accuracy=0.554982	lr=0.134475
Epoch[74] Batch [949]	Speed: 4484.337927 samples/sec	accuracy=0.554895	lr=0.134268
Epoch[74] Batch [999]	Speed: 5506.741092 samples/sec	accuracy=0.554642	lr=0.134062
Epoch[74] Batch [1049]	Speed: 5196.932318 samples/sec	accuracy=0.554541	lr=0.133856
Epoch[74] Batch [1099]	Speed: 5584.270622 samples/sec	accuracy=0.554455	lr=0.133650
Epoch[74] Batch [1149]	Speed: 5775.285773 samples/sec	accuracy=0.554276	lr=0.133444
Epoch[74] Batch [1199]	Speed: 4243.746213 samples/sec	accuracy=0.554091	lr=0.133238
Epoch[74] Batch [1249]	Speed: 5089.791594 samples/sec	accuracy=0.553910	lr=0.133032
[Epoch 74] training: accuracy=0.553904
[Epoch 74] speed: 5207 samples/sec	time cost: 256.893441
[Epoch 74] validation: err-top1=0.412966 err-top5=0.181103
Epoch[75] Batch [49]	Speed: 5206.326655 samples/sec	accuracy=0.561758	lr=0.132822
Epoch[75] Batch [99]	Speed: 4635.027290 samples/sec	accuracy=0.561943	lr=0.132617
Epoch[75] Batch [149]	Speed: 5508.303438 samples/sec	accuracy=0.561432	lr=0.132411
Epoch[75] Batch [199]	Speed: 5875.887688 samples/sec	accuracy=0.561250	lr=0.132206
Epoch[75] Batch [249]	Speed: 5605.046757 samples/sec	accuracy=0.560855	lr=0.132000
Epoch[75] Batch [299]	Speed: 5967.524548 samples/sec	accuracy=0.560303	lr=0.131795
Epoch[75] Batch [349]	Speed: 6068.210049 samples/sec	accuracy=0.559350	lr=0.131590
Epoch[75] Batch [399]	Speed: 6182.896059 samples/sec	accuracy=0.558691	lr=0.131385
Epoch[75] Batch [449]	Speed: 5862.225882 samples/sec	accuracy=0.558401	lr=0.131180
Epoch[75] Batch [499]	Speed: 5844.811541 samples/sec	accuracy=0.558064	lr=0.130975
Epoch[75] Batch [549]	Speed: 5885.003412 samples/sec	accuracy=0.557747	lr=0.130770
Epoch[75] Batch [599]	Speed: 5938.826518 samples/sec	accuracy=0.557650	lr=0.130565
Epoch[75] Batch [649]	Speed: 5656.628394 samples/sec	accuracy=0.557471	lr=0.130360
Epoch[75] Batch [699]	Speed: 5548.658783 samples/sec	accuracy=0.557342	lr=0.130155
Epoch[75] Batch [749]	Speed: 5529.486687 samples/sec	accuracy=0.557310	lr=0.129951
Epoch[75] Batch [799]	Speed: 5898.716138 samples/sec	accuracy=0.557173	lr=0.129746
Epoch[75] Batch [849]	Speed: 6079.734701 samples/sec	accuracy=0.556830	lr=0.129542
Epoch[75] Batch [899]	Speed: 5609.160216 samples/sec	accuracy=0.556882	lr=0.129338
Epoch[75] Batch [949]	Speed: 5107.531537 samples/sec	accuracy=0.557001	lr=0.129133
Epoch[75] Batch [999]	Speed: 4836.969132 samples/sec	accuracy=0.557018	lr=0.128929
Epoch[75] Batch [1049]	Speed: 5158.540507 samples/sec	accuracy=0.556938	lr=0.128725
Epoch[75] Batch [1099]	Speed: 5081.259958 samples/sec	accuracy=0.556955	lr=0.128521
Epoch[75] Batch [1149]	Speed: 5209.052318 samples/sec	accuracy=0.556654	lr=0.128317
Epoch[75] Batch [1199]	Speed: 5230.680612 samples/sec	accuracy=0.556436	lr=0.128113
Epoch[75] Batch [1249]	Speed: 4716.603956 samples/sec	accuracy=0.556086	lr=0.127910
[Epoch 75] training: accuracy=0.556103
[Epoch 75] speed: 5477 samples/sec	time cost: 244.262325
[Epoch 75] validation: err-top1=0.434224 err-top5=0.193197
Epoch[76] Batch [49]	Speed: 5229.581083 samples/sec	accuracy=0.565059	lr=0.127702
Epoch[76] Batch [99]	Speed: 4642.284349 samples/sec	accuracy=0.564707	lr=0.127498
Epoch[76] Batch [149]	Speed: 4879.422470 samples/sec	accuracy=0.562025	lr=0.127295
Epoch[76] Batch [199]	Speed: 5137.654047 samples/sec	accuracy=0.562930	lr=0.127092
Epoch[76] Batch [249]	Speed: 6074.849642 samples/sec	accuracy=0.562070	lr=0.126888
Epoch[76] Batch [299]	Speed: 5607.943865 samples/sec	accuracy=0.561523	lr=0.126685
Epoch[76] Batch [349]	Speed: 5560.080992 samples/sec	accuracy=0.561116	lr=0.126482
Epoch[76] Batch [399]	Speed: 5212.615144 samples/sec	accuracy=0.560652	lr=0.126279
Epoch[76] Batch [449]	Speed: 5436.902282 samples/sec	accuracy=0.560703	lr=0.126076
Epoch[76] Batch [499]	Speed: 6087.641639 samples/sec	accuracy=0.560244	lr=0.125873
Epoch[76] Batch [549]	Speed: 5961.893861 samples/sec	accuracy=0.560565	lr=0.125670
Epoch[76] Batch [599]	Speed: 5750.361012 samples/sec	accuracy=0.560446	lr=0.125468
Epoch[76] Batch [649]	Speed: 5717.195330 samples/sec	accuracy=0.560338	lr=0.125265
Epoch[76] Batch [699]	Speed: 5440.718939 samples/sec	accuracy=0.560494	lr=0.125062
Epoch[76] Batch [749]	Speed: 4478.989818 samples/sec	accuracy=0.560413	lr=0.124860
Epoch[76] Batch [799]	Speed: 5465.115395 samples/sec	accuracy=0.560515	lr=0.124658
Epoch[76] Batch [849]	Speed: 5608.159442 samples/sec	accuracy=0.560393	lr=0.124455
Epoch[76] Batch [899]	Speed: 5903.765628 samples/sec	accuracy=0.560291	lr=0.124253
Epoch[76] Batch [949]	Speed: 5605.267671 samples/sec	accuracy=0.559904	lr=0.124051
Epoch[76] Batch [999]	Speed: 5732.454267 samples/sec	accuracy=0.559807	lr=0.123849
Epoch[76] Batch [1049]	Speed: 5718.756645 samples/sec	accuracy=0.559498	lr=0.123647
Epoch[76] Batch [1099]	Speed: 5691.366488 samples/sec	accuracy=0.559347	lr=0.123446
Epoch[76] Batch [1149]	Speed: 5525.076521 samples/sec	accuracy=0.559316	lr=0.123244
Epoch[76] Batch [1199]	Speed: 5171.494189 samples/sec	accuracy=0.559264	lr=0.123042
Epoch[76] Batch [1249]	Speed: 4265.666010 samples/sec	accuracy=0.559155	lr=0.122841
[Epoch 76] training: accuracy=0.559172
[Epoch 76] speed: 5385 samples/sec	time cost: 247.711103
[Epoch 76] validation: err-top1=0.423948 err-top5=0.185766
Epoch[77] Batch [49]	Speed: 5218.758785 samples/sec	accuracy=0.567617	lr=0.122635
Epoch[77] Batch [99]	Speed: 5972.868336 samples/sec	accuracy=0.568154	lr=0.122434
Epoch[77] Batch [149]	Speed: 5343.100574 samples/sec	accuracy=0.567214	lr=0.122233
Epoch[77] Batch [199]	Speed: 5777.245443 samples/sec	accuracy=0.565801	lr=0.122032
Epoch[77] Batch [249]	Speed: 5953.157533 samples/sec	accuracy=0.564449	lr=0.121831
Epoch[77] Batch [299]	Speed: 5344.447599 samples/sec	accuracy=0.564251	lr=0.121630
Epoch[77] Batch [349]	Speed: 5818.302914 samples/sec	accuracy=0.563504	lr=0.121429
Epoch[77] Batch [399]	Speed: 5788.218036 samples/sec	accuracy=0.563528	lr=0.121228
Epoch[77] Batch [449]	Speed: 6026.509908 samples/sec	accuracy=0.563444	lr=0.121027
Epoch[77] Batch [499]	Speed: 5802.380959 samples/sec	accuracy=0.563586	lr=0.120827
Epoch[77] Batch [549]	Speed: 5949.486379 samples/sec	accuracy=0.563313	lr=0.120626
Epoch[77] Batch [599]	Speed: 5449.625984 samples/sec	accuracy=0.563031	lr=0.120426
Epoch[77] Batch [649]	Speed: 5907.285170 samples/sec	accuracy=0.562883	lr=0.120226
Epoch[77] Batch [699]	Speed: 5714.237186 samples/sec	accuracy=0.562839	lr=0.120025
Epoch[77] Batch [749]	Speed: 5397.139361 samples/sec	accuracy=0.562357	lr=0.119825
Epoch[77] Batch [799]	Speed: 5841.933958 samples/sec	accuracy=0.562214	lr=0.119625
Epoch[77] Batch [849]	Speed: 5869.649355 samples/sec	accuracy=0.562255	lr=0.119425
Epoch[77] Batch [899]	Speed: 5100.909824 samples/sec	accuracy=0.562026	lr=0.119226
Epoch[77] Batch [949]	Speed: 4125.673423 samples/sec	accuracy=0.561758	lr=0.119026
Epoch[77] Batch [999]	Speed: 5823.603768 samples/sec	accuracy=0.561685	lr=0.118826
Epoch[77] Batch [1049]	Speed: 5878.419529 samples/sec	accuracy=0.561523	lr=0.118627
Epoch[77] Batch [1099]	Speed: 4760.796472 samples/sec	accuracy=0.561341	lr=0.118427
Epoch[77] Batch [1149]	Speed: 5716.320879 samples/sec	accuracy=0.561299	lr=0.118228
Epoch[77] Batch [1199]	Speed: 5829.824322 samples/sec	accuracy=0.561400	lr=0.118029
Epoch[77] Batch [1249]	Speed: 5837.542439 samples/sec	accuracy=0.561369	lr=0.117830
[Epoch 77] training: accuracy=0.561377
[Epoch 77] speed: 5564 samples/sec	time cost: 239.810893
[Epoch 77] validation: err-top1=0.401965 err-top5=0.171058
Epoch[78] Batch [49]	Speed: 5791.610179 samples/sec	accuracy=0.570586	lr=0.117627
Epoch[78] Batch [99]	Speed: 5730.739719 samples/sec	accuracy=0.570000	lr=0.117428
Epoch[78] Batch [149]	Speed: 5591.443139 samples/sec	accuracy=0.569792	lr=0.117229
Epoch[78] Batch [199]	Speed: 4773.633674 samples/sec	accuracy=0.569980	lr=0.117030
Epoch[78] Batch [249]	Speed: 5248.549612 samples/sec	accuracy=0.569504	lr=0.116831
Epoch[78] Batch [299]	Speed: 5378.189120 samples/sec	accuracy=0.567913	lr=0.116633
Epoch[78] Batch [349]	Speed: 5175.097717 samples/sec	accuracy=0.567260	lr=0.116434
Epoch[78] Batch [399]	Speed: 5714.098823 samples/sec	accuracy=0.567002	lr=0.116236
Epoch[78] Batch [449]	Speed: 5581.778876 samples/sec	accuracy=0.566970	lr=0.116038
Epoch[78] Batch [499]	Speed: 5380.404640 samples/sec	accuracy=0.566117	lr=0.115840
Epoch[78] Batch [549]	Speed: 5948.570410 samples/sec	accuracy=0.565726	lr=0.115642
Epoch[78] Batch [599]	Speed: 4623.836064 samples/sec	accuracy=0.565757	lr=0.115444
Epoch[78] Batch [649]	Speed: 4787.579228 samples/sec	accuracy=0.565337	lr=0.115246
Epoch[78] Batch [699]	Speed: 5756.803912 samples/sec	accuracy=0.565074	lr=0.115048
Epoch[78] Batch [749]	Speed: 5828.983272 samples/sec	accuracy=0.564928	lr=0.114850
Epoch[78] Batch [799]	Speed: 5680.158607 samples/sec	accuracy=0.564728	lr=0.114653
Epoch[78] Batch [849]	Speed: 5907.728171 samples/sec	accuracy=0.564528	lr=0.114455
Epoch[78] Batch [899]	Speed: 5643.918732 samples/sec	accuracy=0.564291	lr=0.114258
Epoch[78] Batch [949]	Speed: 5535.884386 samples/sec	accuracy=0.564076	lr=0.114061
Epoch[78] Batch [999]	Speed: 4693.835444 samples/sec	accuracy=0.563891	lr=0.113864
Epoch[78] Batch [1049]	Speed: 5234.715071 samples/sec	accuracy=0.563850	lr=0.113667
Epoch[78] Batch [1099]	Speed: 5733.091670 samples/sec	accuracy=0.563816	lr=0.113470
Epoch[78] Batch [1149]	Speed: 5935.670245 samples/sec	accuracy=0.563757	lr=0.113273
Epoch[78] Batch [1199]	Speed: 5646.959670 samples/sec	accuracy=0.563696	lr=0.113076
Epoch[78] Batch [1249]	Speed: 5693.441079 samples/sec	accuracy=0.563711	lr=0.112880
[Epoch 78] training: accuracy=0.563709
[Epoch 78] speed: 5445 samples/sec	time cost: 245.747210
[Epoch 78] validation: err-top1=0.412189 err-top5=0.177396
Epoch[79] Batch [49]	Speed: 4777.801500 samples/sec	accuracy=0.569570	lr=0.112679
Epoch[79] Batch [99]	Speed: 5607.065034 samples/sec	accuracy=0.572227	lr=0.112483
Epoch[79] Batch [149]	Speed: 5730.063850 samples/sec	accuracy=0.571439	lr=0.112287
Epoch[79] Batch [199]	Speed: 5525.005020 samples/sec	accuracy=0.570537	lr=0.112090
Epoch[79] Batch [249]	Speed: 5528.474714 samples/sec	accuracy=0.570562	lr=0.111894
Epoch[79] Batch [299]	Speed: 5668.129776 samples/sec	accuracy=0.569919	lr=0.111698
Epoch[79] Batch [349]	Speed: 5803.697715 samples/sec	accuracy=0.570000	lr=0.111502
Epoch[79] Batch [399]	Speed: 6062.719538 samples/sec	accuracy=0.569548	lr=0.111307
Epoch[79] Batch [449]	Speed: 5700.031729 samples/sec	accuracy=0.569230	lr=0.111111
Epoch[79] Batch [499]	Speed: 5638.944689 samples/sec	accuracy=0.568201	lr=0.110915
Epoch[79] Batch [549]	Speed: 5078.234128 samples/sec	accuracy=0.568345	lr=0.110720
Epoch[79] Batch [599]	Speed: 5762.003607 samples/sec	accuracy=0.568319	lr=0.110525
Epoch[79] Batch [649]	Speed: 5232.943667 samples/sec	accuracy=0.568187	lr=0.110329
Epoch[79] Batch [699]	Speed: 4897.040726 samples/sec	accuracy=0.567875	lr=0.110134
Epoch[79] Batch [749]	Speed: 4858.617999 samples/sec	accuracy=0.567522	lr=0.109939
Epoch[79] Batch [799]	Speed: 5488.138842 samples/sec	accuracy=0.567150	lr=0.109744
Epoch[79] Batch [849]	Speed: 4751.584072 samples/sec	accuracy=0.566938	lr=0.109549
Epoch[79] Batch [899]	Speed: 5814.298182 samples/sec	accuracy=0.566800	lr=0.109355
Epoch[79] Batch [949]	Speed: 5613.505829 samples/sec	accuracy=0.566741	lr=0.109160
Epoch[79] Batch [999]	Speed: 5009.595982 samples/sec	accuracy=0.566594	lr=0.108966
Epoch[79] Batch [1049]	Speed: 4868.427005 samples/sec	accuracy=0.566472	lr=0.108771
Epoch[79] Batch [1099]	Speed: 5793.942180 samples/sec	accuracy=0.566325	lr=0.108577
Epoch[79] Batch [1149]	Speed: 5739.763600 samples/sec	accuracy=0.566233	lr=0.108383
Epoch[79] Batch [1199]	Speed: 5868.122909 samples/sec	accuracy=0.566118	lr=0.108189
Epoch[79] Batch [1249]	Speed: 5498.320760 samples/sec	accuracy=0.566205	lr=0.107995
[Epoch 79] training: accuracy=0.566207
[Epoch 79] speed: 5413 samples/sec	time cost: 247.577888
[Epoch 79] validation: err-top1=0.392478 err-top5=0.165956
Epoch[80] Batch [49]	Speed: 5647.495921 samples/sec	accuracy=0.577148	lr=0.107797
Epoch[80] Batch [99]	Speed: 5517.780088 samples/sec	accuracy=0.576602	lr=0.107603
Epoch[80] Batch [149]	Speed: 5999.839932 samples/sec	accuracy=0.574154	lr=0.107410
Epoch[80] Batch [199]	Speed: 5145.365155 samples/sec	accuracy=0.573838	lr=0.107216
Epoch[80] Batch [249]	Speed: 5416.809713 samples/sec	accuracy=0.573438	lr=0.107023
Epoch[80] Batch [299]	Speed: 6049.125833 samples/sec	accuracy=0.572952	lr=0.106830
Epoch[80] Batch [349]	Speed: 5927.005639 samples/sec	accuracy=0.572302	lr=0.106636
Epoch[80] Batch [399]	Speed: 4954.349842 samples/sec	accuracy=0.572051	lr=0.106443
Epoch[80] Batch [449]	Speed: 5756.746195 samples/sec	accuracy=0.571901	lr=0.106250
Epoch[80] Batch [499]	Speed: 5739.658054 samples/sec	accuracy=0.571561	lr=0.106057
Epoch[80] Batch [549]	Speed: 5189.206928 samples/sec	accuracy=0.571456	lr=0.105865
Epoch[80] Batch [599]	Speed: 5252.108375 samples/sec	accuracy=0.571450	lr=0.105672
Epoch[80] Batch [649]	Speed: 5696.616778 samples/sec	accuracy=0.570918	lr=0.105480
Epoch[80] Batch [699]	Speed: 5615.757378 samples/sec	accuracy=0.570758	lr=0.105287
Epoch[80] Batch [749]	Speed: 5401.788593 samples/sec	accuracy=0.570491	lr=0.105095
Epoch[80] Batch [799]	Speed: 5481.214753 samples/sec	accuracy=0.570496	lr=0.104903
Epoch[80] Batch [849]	Speed: 5641.007460 samples/sec	accuracy=0.570249	lr=0.104711
Epoch[80] Batch [899]	Speed: 4710.367234 samples/sec	accuracy=0.570242	lr=0.104519
Epoch[80] Batch [949]	Speed: 4870.271861 samples/sec	accuracy=0.570012	lr=0.104327
Epoch[80] Batch [999]	Speed: 5535.115875 samples/sec	accuracy=0.569634	lr=0.104135
Epoch[80] Batch [1049]	Speed: 5799.493621 samples/sec	accuracy=0.569603	lr=0.103944
Epoch[80] Batch [1099]	Speed: 5572.160253 samples/sec	accuracy=0.569317	lr=0.103752
Epoch[80] Batch [1149]	Speed: 5845.935964 samples/sec	accuracy=0.568975	lr=0.103561
Epoch[80] Batch [1199]	Speed: 5698.705786 samples/sec	accuracy=0.568968	lr=0.103370
Epoch[80] Batch [1249]	Speed: 5727.125063 samples/sec	accuracy=0.568868	lr=0.103178
[Epoch 80] training: accuracy=0.568868
[Epoch 80] speed: 5498 samples/sec	time cost: 242.908655
[Epoch 80] validation: err-top1=0.407246 err-top5=0.173848
Epoch[81] Batch [49]	Speed: 5242.183261 samples/sec	accuracy=0.578047	lr=0.102984
Epoch[81] Batch [99]	Speed: 5531.290764 samples/sec	accuracy=0.581221	lr=0.102793
Epoch[81] Batch [149]	Speed: 5155.640917 samples/sec	accuracy=0.578073	lr=0.102602
Epoch[81] Batch [199]	Speed: 5637.817217 samples/sec	accuracy=0.577812	lr=0.102411
Epoch[81] Batch [249]	Speed: 5592.123397 samples/sec	accuracy=0.576594	lr=0.102221
Epoch[81] Batch [299]	Speed: 5057.867623 samples/sec	accuracy=0.575999	lr=0.102030
Epoch[81] Batch [349]	Speed: 5070.458059 samples/sec	accuracy=0.575622	lr=0.101840
Epoch[81] Batch [399]	Speed: 5674.753094 samples/sec	accuracy=0.575291	lr=0.101650
Epoch[81] Batch [449]	Speed: 5818.300076 samples/sec	accuracy=0.574794	lr=0.101460
Epoch[81] Batch [499]	Speed: 6131.868085 samples/sec	accuracy=0.574697	lr=0.101270
Epoch[81] Batch [549]	Speed: 5856.426299 samples/sec	accuracy=0.574556	lr=0.101080
Epoch[81] Batch [599]	Speed: 5964.228211 samples/sec	accuracy=0.574551	lr=0.100890
Epoch[81] Batch [649]	Speed: 6060.128411 samples/sec	accuracy=0.574483	lr=0.100701
Epoch[81] Batch [699]	Speed: 4821.326018 samples/sec	accuracy=0.574003	lr=0.100511
Epoch[81] Batch [749]	Speed: 5122.583334 samples/sec	accuracy=0.573493	lr=0.100322
Epoch[81] Batch [799]	Speed: 5621.382977 samples/sec	accuracy=0.573120	lr=0.100132
Epoch[81] Batch [849]	Speed: 5115.977284 samples/sec	accuracy=0.572814	lr=0.099943
Epoch[81] Batch [899]	Speed: 5548.968471 samples/sec	accuracy=0.572524	lr=0.099754
Epoch[81] Batch [949]	Speed: 5607.164441 samples/sec	accuracy=0.572481	lr=0.099565
Epoch[81] Batch [999]	Speed: 5254.078267 samples/sec	accuracy=0.572336	lr=0.099377
Epoch[81] Batch [1049]	Speed: 4744.086135 samples/sec	accuracy=0.572375	lr=0.099188
Epoch[81] Batch [1099]	Speed: 5035.929250 samples/sec	accuracy=0.572218	lr=0.098999
Epoch[81] Batch [1149]	Speed: 5841.822079 samples/sec	accuracy=0.572206	lr=0.098811
Epoch[81] Batch [1199]	Speed: 4264.003044 samples/sec	accuracy=0.571997	lr=0.098623
Epoch[81] Batch [1249]	Speed: 5033.952653 samples/sec	accuracy=0.571799	lr=0.098434
[Epoch 81] training: accuracy=0.571818
[Epoch 81] speed: 5346 samples/sec	time cost: 249.556794
[Epoch 81] validation: err-top1=0.386027 err-top5=0.157633
Epoch[82] Batch [49]	Speed: 5142.998967 samples/sec	accuracy=0.582187	lr=0.098243
Epoch[82] Batch [99]	Speed: 5022.989440 samples/sec	accuracy=0.579619	lr=0.098055
Epoch[82] Batch [149]	Speed: 5481.894341 samples/sec	accuracy=0.578783	lr=0.097867
Epoch[82] Batch [199]	Speed: 5303.329148 samples/sec	accuracy=0.579541	lr=0.097679
Epoch[82] Batch [249]	Speed: 4999.873921 samples/sec	accuracy=0.578812	lr=0.097492
Epoch[82] Batch [299]	Speed: 5568.159335 samples/sec	accuracy=0.578682	lr=0.097304
Epoch[82] Batch [349]	Speed: 4808.260338 samples/sec	accuracy=0.578493	lr=0.097117
Epoch[82] Batch [399]	Speed: 5506.654109 samples/sec	accuracy=0.578147	lr=0.096930
Epoch[82] Batch [449]	Speed: 5465.550754 samples/sec	accuracy=0.577871	lr=0.096743
Epoch[82] Batch [499]	Speed: 4996.246442 samples/sec	accuracy=0.577982	lr=0.096556
Epoch[82] Batch [549]	Speed: 5374.838855 samples/sec	accuracy=0.577417	lr=0.096369
Epoch[82] Batch [599]	Speed: 5426.454601 samples/sec	accuracy=0.577067	lr=0.096182
Epoch[82] Batch [649]	Speed: 5120.131107 samples/sec	accuracy=0.576928	lr=0.095996
Epoch[82] Batch [699]	Speed: 5234.536818 samples/sec	accuracy=0.576812	lr=0.095809
Epoch[82] Batch [749]	Speed: 5602.723390 samples/sec	accuracy=0.576710	lr=0.095623
Epoch[82] Batch [799]	Speed: 5695.035660 samples/sec	accuracy=0.576599	lr=0.095437
Epoch[82] Batch [849]	Speed: 5520.582270 samples/sec	accuracy=0.576374	lr=0.095250
Epoch[82] Batch [899]	Speed: 6092.634246 samples/sec	accuracy=0.575949	lr=0.095064
Epoch[82] Batch [949]	Speed: 5622.148990 samples/sec	accuracy=0.575776	lr=0.094879
Epoch[82] Batch [999]	Speed: 5327.737619 samples/sec	accuracy=0.575637	lr=0.094693
Epoch[82] Batch [1049]	Speed: 6065.675048 samples/sec	accuracy=0.575208	lr=0.094507
Epoch[82] Batch [1099]	Speed: 5564.772182 samples/sec	accuracy=0.575021	lr=0.094322
Epoch[82] Batch [1149]	Speed: 4999.535658 samples/sec	accuracy=0.574788	lr=0.094137
Epoch[82] Batch [1199]	Speed: 5627.161408 samples/sec	accuracy=0.574559	lr=0.093951
Epoch[82] Batch [1249]	Speed: 5021.329879 samples/sec	accuracy=0.574501	lr=0.093766
[Epoch 82] training: accuracy=0.574497
[Epoch 82] speed: 5359 samples/sec	time cost: 249.706462
[Epoch 82] validation: err-top1=0.393714 err-top5=0.164322
Epoch[83] Batch [49]	Speed: 5426.033125 samples/sec	accuracy=0.584688	lr=0.093578
Epoch[83] Batch [99]	Speed: 5726.568392 samples/sec	accuracy=0.582920	lr=0.093393
Epoch[83] Batch [149]	Speed: 5825.017549 samples/sec	accuracy=0.582650	lr=0.093208
Epoch[83] Batch [199]	Speed: 5966.384696 samples/sec	accuracy=0.581216	lr=0.093024
Epoch[83] Batch [249]	Speed: 5897.708508 samples/sec	accuracy=0.581441	lr=0.092839
Epoch[83] Batch [299]	Speed: 5676.089367 samples/sec	accuracy=0.580436	lr=0.092655
Epoch[83] Batch [349]	Speed: 5170.957735 samples/sec	accuracy=0.579855	lr=0.092471
Epoch[83] Batch [399]	Speed: 5668.040014 samples/sec	accuracy=0.579651	lr=0.092287
Epoch[83] Batch [449]	Speed: 4966.255219 samples/sec	accuracy=0.579698	lr=0.092103
Epoch[83] Batch [499]	Speed: 4912.390532 samples/sec	accuracy=0.579592	lr=0.091919
Epoch[83] Batch [549]	Speed: 6053.919416 samples/sec	accuracy=0.579451	lr=0.091735
Epoch[83] Batch [599]	Speed: 5751.447382 samples/sec	accuracy=0.579090	lr=0.091552
Epoch[83] Batch [649]	Speed: 4845.423491 samples/sec	accuracy=0.578935	lr=0.091368
Epoch[83] Batch [699]	Speed: 5501.341372 samples/sec	accuracy=0.578750	lr=0.091185
Epoch[83] Batch [749]	Speed: 4528.600896 samples/sec	accuracy=0.578478	lr=0.091002
Epoch[83] Batch [799]	Speed: 5749.914662 samples/sec	accuracy=0.578145	lr=0.090819
Epoch[83] Batch [849]	Speed: 4378.891466 samples/sec	accuracy=0.578163	lr=0.090636
Epoch[83] Batch [899]	Speed: 5058.104456 samples/sec	accuracy=0.578166	lr=0.090453
Epoch[83] Batch [949]	Speed: 5774.947048 samples/sec	accuracy=0.578029	lr=0.090270
Epoch[83] Batch [999]	Speed: 5708.501386 samples/sec	accuracy=0.577824	lr=0.090088
Epoch[83] Batch [1049]	Speed: 5635.804699 samples/sec	accuracy=0.577785	lr=0.089905
Epoch[83] Batch [1099]	Speed: 5791.136476 samples/sec	accuracy=0.577801	lr=0.089723
Epoch[83] Batch [1149]	Speed: 5656.660429 samples/sec	accuracy=0.577623	lr=0.089541
Epoch[83] Batch [1199]	Speed: 5262.291251 samples/sec	accuracy=0.577527	lr=0.089359
Epoch[83] Batch [1249]	Speed: 4965.187466 samples/sec	accuracy=0.577346	lr=0.089177
[Epoch 83] training: accuracy=0.577346
[Epoch 83] speed: 5389 samples/sec	time cost: 246.960867
[Epoch 83] validation: err-top1=0.390605 err-top5=0.163644
Epoch[84] Batch [49]	Speed: 5745.590251 samples/sec	accuracy=0.583867	lr=0.088992
Epoch[84] Batch [99]	Speed: 6053.905080 samples/sec	accuracy=0.583184	lr=0.088810
Epoch[84] Batch [149]	Speed: 5814.852831 samples/sec	accuracy=0.584199	lr=0.088629
Epoch[84] Batch [199]	Speed: 5438.351838 samples/sec	accuracy=0.583228	lr=0.088448
Epoch[84] Batch [249]	Speed: 5851.543529 samples/sec	accuracy=0.583266	lr=0.088266
Epoch[84] Batch [299]	Speed: 5622.584850 samples/sec	accuracy=0.582832	lr=0.088085
Epoch[84] Batch [349]	Speed: 5609.688140 samples/sec	accuracy=0.582531	lr=0.087904
Epoch[84] Batch [399]	Speed: 4890.356235 samples/sec	accuracy=0.582410	lr=0.087724
Epoch[84] Batch [449]	Speed: 5960.737628 samples/sec	accuracy=0.582001	lr=0.087543
Epoch[84] Batch [499]	Speed: 4890.339419 samples/sec	accuracy=0.581895	lr=0.087363
Epoch[84] Batch [549]	Speed: 4558.930174 samples/sec	accuracy=0.581438	lr=0.087182
Epoch[84] Batch [599]	Speed: 4661.754211 samples/sec	accuracy=0.581546	lr=0.087002
Epoch[84] Batch [649]	Speed: 6144.555402 samples/sec	accuracy=0.581064	lr=0.086822
Epoch[84] Batch [699]	Speed: 5813.260957 samples/sec	accuracy=0.580977	lr=0.086642
Epoch[84] Batch [749]	Speed: 5693.261309 samples/sec	accuracy=0.580944	lr=0.086462
Epoch[84] Batch [799]	Speed: 5860.006338 samples/sec	accuracy=0.580935	lr=0.086282
Epoch[84] Batch [849]	Speed: 5798.065745 samples/sec	accuracy=0.580725	lr=0.086103
Epoch[84] Batch [899]	Speed: 5468.812451 samples/sec	accuracy=0.580436	lr=0.085923
Epoch[84] Batch [949]	Speed: 3761.058490 samples/sec	accuracy=0.580409	lr=0.085744
Epoch[84] Batch [999]	Speed: 5352.893728 samples/sec	accuracy=0.580543	lr=0.085565
Epoch[84] Batch [1049]	Speed: 4927.839356 samples/sec	accuracy=0.580459	lr=0.085386
Epoch[84] Batch [1099]	Speed: 5685.233952 samples/sec	accuracy=0.580384	lr=0.085207
Epoch[84] Batch [1149]	Speed: 5662.042111 samples/sec	accuracy=0.580228	lr=0.085028
Epoch[84] Batch [1199]	Speed: 6031.241445 samples/sec	accuracy=0.580123	lr=0.084850
Epoch[84] Batch [1249]	Speed: 5479.756383 samples/sec	accuracy=0.580134	lr=0.084671
[Epoch 84] training: accuracy=0.580140
[Epoch 84] speed: 5401 samples/sec	time cost: 247.829845
[Epoch 84] validation: err-top1=0.394491 err-top5=0.162986
Epoch[85] Batch [49]	Speed: 5195.162131 samples/sec	accuracy=0.591387	lr=0.084489
Epoch[85] Batch [99]	Speed: 5346.162482 samples/sec	accuracy=0.590566	lr=0.084311
Epoch[85] Batch [149]	Speed: 5619.179112 samples/sec	accuracy=0.589167	lr=0.084133
Epoch[85] Batch [199]	Speed: 4906.638240 samples/sec	accuracy=0.588555	lr=0.083955
Epoch[85] Batch [249]	Speed: 5968.070670 samples/sec	accuracy=0.587879	lr=0.083777
Epoch[85] Batch [299]	Speed: 5844.281858 samples/sec	accuracy=0.588369	lr=0.083600
Epoch[85] Batch [349]	Speed: 5853.568544 samples/sec	accuracy=0.587670	lr=0.083422
Epoch[85] Batch [399]	Speed: 5614.531852 samples/sec	accuracy=0.587087	lr=0.083245
Epoch[85] Batch [449]	Speed: 5656.437234 samples/sec	accuracy=0.586424	lr=0.083067
Epoch[85] Batch [499]	Speed: 5731.494832 samples/sec	accuracy=0.585930	lr=0.082890
Epoch[85] Batch [549]	Speed: 5206.202835 samples/sec	accuracy=0.585321	lr=0.082713
Epoch[85] Batch [599]	Speed: 5691.324707 samples/sec	accuracy=0.585116	lr=0.082537
Epoch[85] Batch [649]	Speed: 5989.101921 samples/sec	accuracy=0.585033	lr=0.082360
Epoch[85] Batch [699]	Speed: 5874.551474 samples/sec	accuracy=0.584759	lr=0.082183
Epoch[85] Batch [749]	Speed: 5678.327148 samples/sec	accuracy=0.584625	lr=0.082007
Epoch[85] Batch [799]	Speed: 4433.036249 samples/sec	accuracy=0.584517	lr=0.081831
Epoch[85] Batch [849]	Speed: 5762.079518 samples/sec	accuracy=0.584125	lr=0.081655
Epoch[85] Batch [899]	Speed: 6129.557798 samples/sec	accuracy=0.583993	lr=0.081479
Epoch[85] Batch [949]	Speed: 4955.339985 samples/sec	accuracy=0.583958	lr=0.081303
Epoch[85] Batch [999]	Speed: 5579.187164 samples/sec	accuracy=0.583818	lr=0.081127
Epoch[85] Batch [1049]	Speed: 5882.605121 samples/sec	accuracy=0.583761	lr=0.080952
Epoch[85] Batch [1099]	Speed: 5295.050076 samples/sec	accuracy=0.583576	lr=0.080776
Epoch[85] Batch [1149]	Speed: 4509.748487 samples/sec	accuracy=0.583545	lr=0.080601
Epoch[85] Batch [1199]	Speed: 4782.658525 samples/sec	accuracy=0.583384	lr=0.080426
Epoch[85] Batch [1249]	Speed: 5586.943950 samples/sec	accuracy=0.583072	lr=0.080251
[Epoch 85] training: accuracy=0.583094
[Epoch 85] speed: 5437 samples/sec	time cost: 245.403610
[Epoch 85] validation: err-top1=0.397401 err-top5=0.168288
Epoch[86] Batch [49]	Speed: 5656.036629 samples/sec	accuracy=0.588320	lr=0.080073
Epoch[86] Batch [99]	Speed: 5542.792004 samples/sec	accuracy=0.588906	lr=0.079898
Epoch[86] Batch [149]	Speed: 5657.494811 samples/sec	accuracy=0.589102	lr=0.079723
Epoch[86] Batch [199]	Speed: 5627.062470 samples/sec	accuracy=0.588862	lr=0.079549
Epoch[86] Batch [249]	Speed: 5199.191055 samples/sec	accuracy=0.587918	lr=0.079375
Epoch[86] Batch [299]	Speed: 6102.009752 samples/sec	accuracy=0.587822	lr=0.079201
Epoch[86] Batch [349]	Speed: 5722.815634 samples/sec	accuracy=0.587812	lr=0.079027
Epoch[86] Batch [399]	Speed: 5373.305580 samples/sec	accuracy=0.587307	lr=0.078853
Epoch[86] Batch [449]	Speed: 5489.107338 samples/sec	accuracy=0.587431	lr=0.078679
Epoch[86] Batch [499]	Speed: 5774.485384 samples/sec	accuracy=0.587064	lr=0.078506
Epoch[86] Batch [549]	Speed: 5715.404560 samples/sec	accuracy=0.587024	lr=0.078332
Epoch[86] Batch [599]	Speed: 5784.541438 samples/sec	accuracy=0.587005	lr=0.078159
Epoch[86] Batch [649]	Speed: 5534.864650 samples/sec	accuracy=0.586926	lr=0.077986
Epoch[86] Batch [699]	Speed: 4966.000841 samples/sec	accuracy=0.586733	lr=0.077813
Epoch[86] Batch [749]	Speed: 5625.458123 samples/sec	accuracy=0.586531	lr=0.077640
Epoch[86] Batch [799]	Speed: 5660.906280 samples/sec	accuracy=0.586467	lr=0.077467
Epoch[86] Batch [849]	Speed: 4978.333589 samples/sec	accuracy=0.586426	lr=0.077295
Epoch[86] Batch [899]	Speed: 5730.453755 samples/sec	accuracy=0.586217	lr=0.077123
Epoch[86] Batch [949]	Speed: 5724.760911 samples/sec	accuracy=0.586103	lr=0.076950
Epoch[86] Batch [999]	Speed: 5223.101964 samples/sec	accuracy=0.586070	lr=0.076778
Epoch[86] Batch [1049]	Speed: 5247.901252 samples/sec	accuracy=0.586045	lr=0.076606
Epoch[86] Batch [1099]	Speed: 4097.076689 samples/sec	accuracy=0.585888	lr=0.076435
Epoch[86] Batch [1149]	Speed: 5165.796753 samples/sec	accuracy=0.585817	lr=0.076263
Epoch[86] Batch [1199]	Speed: 5643.051574 samples/sec	accuracy=0.585767	lr=0.076091
Epoch[86] Batch [1249]	Speed: 5920.806733 samples/sec	accuracy=0.585805	lr=0.075920
[Epoch 86] training: accuracy=0.585828
[Epoch 86] speed: 5447 samples/sec	time cost: 244.504673
[Epoch 86] validation: err-top1=0.385065 err-top5=0.157087
Epoch[87] Batch [49]	Speed: 5332.013139 samples/sec	accuracy=0.591191	lr=0.075745
Epoch[87] Batch [99]	Speed: 5387.962330 samples/sec	accuracy=0.592275	lr=0.075574
Epoch[87] Batch [149]	Speed: 5849.658221 samples/sec	accuracy=0.592650	lr=0.075404
Epoch[87] Batch [199]	Speed: 5566.169685 samples/sec	accuracy=0.592661	lr=0.075233
Epoch[87] Batch [249]	Speed: 5981.150523 samples/sec	accuracy=0.592680	lr=0.075062
Epoch[87] Batch [299]	Speed: 4733.997776 samples/sec	accuracy=0.591595	lr=0.074892
Epoch[87] Batch [349]	Speed: 5708.514132 samples/sec	accuracy=0.591122	lr=0.074721
Epoch[87] Batch [399]	Speed: 4847.423608 samples/sec	accuracy=0.590630	lr=0.074551
Epoch[87] Batch [449]	Speed: 4685.891956 samples/sec	accuracy=0.590106	lr=0.074381
Epoch[87] Batch [499]	Speed: 4902.550811 samples/sec	accuracy=0.590127	lr=0.074211
Epoch[87] Batch [549]	Speed: 6133.641186 samples/sec	accuracy=0.590352	lr=0.074042
Epoch[87] Batch [599]	Speed: 5610.284463 samples/sec	accuracy=0.590015	lr=0.073872
Epoch[87] Batch [649]	Speed: 5884.834241 samples/sec	accuracy=0.589620	lr=0.073703
Epoch[87] Batch [699]	Speed: 6112.392796 samples/sec	accuracy=0.589372	lr=0.073534
Epoch[87] Batch [749]	Speed: 4995.869037 samples/sec	accuracy=0.589102	lr=0.073364
Epoch[87] Batch [799]	Speed: 5188.908135 samples/sec	accuracy=0.588961	lr=0.073196
Epoch[87] Batch [849]	Speed: 5450.714437 samples/sec	accuracy=0.588703	lr=0.073027
Epoch[87] Batch [899]	Speed: 5050.976912 samples/sec	accuracy=0.588792	lr=0.072858
Epoch[87] Batch [949]	Speed: 5582.847034 samples/sec	accuracy=0.588553	lr=0.072690
Epoch[87] Batch [999]	Speed: 5819.427725 samples/sec	accuracy=0.588442	lr=0.072521
Epoch[87] Batch [1049]	Speed: 5855.404964 samples/sec	accuracy=0.588378	lr=0.072353
Epoch[87] Batch [1099]	Speed: 5864.310664 samples/sec	accuracy=0.588331	lr=0.072185
Epoch[87] Batch [1149]	Speed: 5425.092513 samples/sec	accuracy=0.588299	lr=0.072017
Epoch[87] Batch [1199]	Speed: 5687.508774 samples/sec	accuracy=0.588286	lr=0.071849
Epoch[87] Batch [1249]	Speed: 5687.372908 samples/sec	accuracy=0.588157	lr=0.071682
[Epoch 87] training: accuracy=0.588175
[Epoch 87] speed: 5454 samples/sec	time cost: 244.486481
[Epoch 87] validation: err-top1=0.375610 err-top5=0.153158
Epoch[88] Batch [49]	Speed: 5598.980490 samples/sec	accuracy=0.595293	lr=0.071511
Epoch[88] Batch [99]	Speed: 5549.168782 samples/sec	accuracy=0.597412	lr=0.071344
Epoch[88] Batch [149]	Speed: 4969.182844 samples/sec	accuracy=0.597376	lr=0.071177
Epoch[88] Batch [199]	Speed: 3812.472855 samples/sec	accuracy=0.596851	lr=0.071010
Epoch[88] Batch [249]	Speed: 4849.905403 samples/sec	accuracy=0.596281	lr=0.070843
Epoch[88] Batch [299]	Speed: 5664.283916 samples/sec	accuracy=0.595234	lr=0.070676
Epoch[88] Batch [349]	Speed: 5982.091051 samples/sec	accuracy=0.594939	lr=0.070510
Epoch[88] Batch [399]	Speed: 5791.113675 samples/sec	accuracy=0.594941	lr=0.070343
Epoch[88] Batch [449]	Speed: 5097.713280 samples/sec	accuracy=0.594744	lr=0.070177
Epoch[88] Batch [499]	Speed: 5592.500289 samples/sec	accuracy=0.594437	lr=0.070011
Epoch[88] Batch [549]	Speed: 5599.022824 samples/sec	accuracy=0.594176	lr=0.069845
Epoch[88] Batch [599]	Speed: 5737.623692 samples/sec	accuracy=0.594141	lr=0.069680
Epoch[88] Batch [649]	Speed: 5835.708799 samples/sec	accuracy=0.593498	lr=0.069514
Epoch[88] Batch [699]	Speed: 6074.380537 samples/sec	accuracy=0.593403	lr=0.069349
Epoch[88] Batch [749]	Speed: 5876.041392 samples/sec	accuracy=0.593263	lr=0.069183
Epoch[88] Batch [799]	Speed: 5701.969115 samples/sec	accuracy=0.593173	lr=0.069018
Epoch[88] Batch [849]	Speed: 5845.344661 samples/sec	accuracy=0.593147	lr=0.068853
Epoch[88] Batch [899]	Speed: 6183.621907 samples/sec	accuracy=0.593188	lr=0.068689
Epoch[88] Batch [949]	Speed: 5515.865942 samples/sec	accuracy=0.592963	lr=0.068524
Epoch[88] Batch [999]	Speed: 4840.079835 samples/sec	accuracy=0.592695	lr=0.068359
Epoch[88] Batch [1049]	Speed: 4741.517015 samples/sec	accuracy=0.592526	lr=0.068195
Epoch[88] Batch [1099]	Speed: 5665.629015 samples/sec	accuracy=0.592388	lr=0.068031
Epoch[88] Batch [1149]	Speed: 5949.039071 samples/sec	accuracy=0.592311	lr=0.067867
Epoch[88] Batch [1199]	Speed: 4481.672238 samples/sec	accuracy=0.592173	lr=0.067703
Epoch[88] Batch [1249]	Speed: 5804.756947 samples/sec	accuracy=0.591966	lr=0.067539
[Epoch 88] training: accuracy=0.591968
[Epoch 88] speed: 5398 samples/sec	time cost: 247.547893
[Epoch 88] validation: err-top1=0.372967 err-top5=0.150929
Epoch[89] Batch [49]	Speed: 5340.789343 samples/sec	accuracy=0.602969	lr=0.067373
Epoch[89] Batch [99]	Speed: 5453.634203 samples/sec	accuracy=0.600986	lr=0.067209
Epoch[89] Batch [149]	Speed: 5634.259075 samples/sec	accuracy=0.601042	lr=0.067046
Epoch[89] Batch [199]	Speed: 5997.553505 samples/sec	accuracy=0.599917	lr=0.066883
Epoch[89] Batch [249]	Speed: 5891.078568 samples/sec	accuracy=0.598637	lr=0.066720
Epoch[89] Batch [299]	Speed: 5897.630277 samples/sec	accuracy=0.597982	lr=0.066557
Epoch[89] Batch [349]	Speed: 6141.285263 samples/sec	accuracy=0.597377	lr=0.066395
Epoch[89] Batch [399]	Speed: 5518.439703 samples/sec	accuracy=0.596799	lr=0.066232
Epoch[89] Batch [449]	Speed: 4520.740966 samples/sec	accuracy=0.596220	lr=0.066070
Epoch[89] Batch [499]	Speed: 5069.837509 samples/sec	accuracy=0.596158	lr=0.065908
Epoch[89] Batch [549]	Speed: 5900.960244 samples/sec	accuracy=0.595740	lr=0.065746
Epoch[89] Batch [599]	Speed: 5326.732603 samples/sec	accuracy=0.595267	lr=0.065584
Epoch[89] Batch [649]	Speed: 5031.429153 samples/sec	accuracy=0.595508	lr=0.065423
Epoch[89] Batch [699]	Speed: 5652.902430 samples/sec	accuracy=0.595328	lr=0.065261
Epoch[89] Batch [749]	Speed: 5065.019300 samples/sec	accuracy=0.595402	lr=0.065100
Epoch[89] Batch [799]	Speed: 5592.861210 samples/sec	accuracy=0.595208	lr=0.064939
Epoch[89] Batch [849]	Speed: 6205.766485 samples/sec	accuracy=0.595139	lr=0.064778
Epoch[89] Batch [899]	Speed: 5920.076316 samples/sec	accuracy=0.595183	lr=0.064617
Epoch[89] Batch [949]	Speed: 5868.135095 samples/sec	accuracy=0.595193	lr=0.064456
Epoch[89] Batch [999]	Speed: 6041.702860 samples/sec	accuracy=0.595121	lr=0.064296
Epoch[89] Batch [1049]	Speed: 5077.796329 samples/sec	accuracy=0.595057	lr=0.064135
Epoch[89] Batch [1099]	Speed: 5748.870731 samples/sec	accuracy=0.594880	lr=0.063975
Epoch[89] Batch [1149]	Speed: 6147.207631 samples/sec	accuracy=0.594909	lr=0.063815
Epoch[89] Batch [1199]	Speed: 5839.746424 samples/sec	accuracy=0.594896	lr=0.063655
Epoch[89] Batch [1249]	Speed: 5769.043221 samples/sec	accuracy=0.594892	lr=0.063496
[Epoch 89] training: accuracy=0.594899
[Epoch 89] speed: 5585 samples/sec	time cost: 240.487508
[Epoch 89] validation: err-top1=0.377033 err-top5=0.154337
Epoch[90] Batch [49]	Speed: 5103.131079 samples/sec	accuracy=0.603340	lr=0.063333
Epoch[90] Batch [99]	Speed: 5523.467852 samples/sec	accuracy=0.602822	lr=0.063174
Epoch[90] Batch [149]	Speed: 6024.734305 samples/sec	accuracy=0.603542	lr=0.063015
Epoch[90] Batch [199]	Speed: 5662.398328 samples/sec	accuracy=0.602344	lr=0.062855
Epoch[90] Batch [249]	Speed: 5633.911267 samples/sec	accuracy=0.601629	lr=0.062697
Epoch[90] Batch [299]	Speed: 5483.281044 samples/sec	accuracy=0.601494	lr=0.062538
Epoch[90] Batch [349]	Speed: 6041.676514 samples/sec	accuracy=0.601306	lr=0.062379
Epoch[90] Batch [399]	Speed: 5948.851038 samples/sec	accuracy=0.601201	lr=0.062221
Epoch[90] Batch [449]	Speed: 5868.288394 samples/sec	accuracy=0.600647	lr=0.062063
Epoch[90] Batch [499]	Speed: 4843.249922 samples/sec	accuracy=0.600482	lr=0.061905
Epoch[90] Batch [549]	Speed: 6005.238084 samples/sec	accuracy=0.600451	lr=0.061747
Epoch[90] Batch [599]	Speed: 5641.529539 samples/sec	accuracy=0.600277	lr=0.061589
Epoch[90] Batch [649]	Speed: 5934.829543 samples/sec	accuracy=0.600427	lr=0.061432
Epoch[90] Batch [699]	Speed: 5785.081230 samples/sec	accuracy=0.600389	lr=0.061274
Epoch[90] Batch [749]	Speed: 5680.509593 samples/sec	accuracy=0.600336	lr=0.061117
Epoch[90] Batch [799]	Speed: 5812.664288 samples/sec	accuracy=0.600098	lr=0.060960
Epoch[90] Batch [849]	Speed: 5927.766566 samples/sec	accuracy=0.599784	lr=0.060803
Epoch[90] Batch [899]	Speed: 5735.393932 samples/sec	accuracy=0.599707	lr=0.060646
Epoch[90] Batch [949]	Speed: 5960.965960 samples/sec	accuracy=0.599693	lr=0.060490
Epoch[90] Batch [999]	Speed: 5980.379994 samples/sec	accuracy=0.599493	lr=0.060333
Epoch[90] Batch [1049]	Speed: 5974.586234 samples/sec	accuracy=0.599288	lr=0.060177
Epoch[90] Batch [1099]	Speed: 5955.608754 samples/sec	accuracy=0.599107	lr=0.060021
Epoch[90] Batch [1149]	Speed: 5932.541088 samples/sec	accuracy=0.598995	lr=0.059865
Epoch[90] Batch [1199]	Speed: 5470.967521 samples/sec	accuracy=0.598831	lr=0.059710
Epoch[90] Batch [1249]	Speed: 3819.199121 samples/sec	accuracy=0.599016	lr=0.059554
[Epoch 90] training: accuracy=0.598999
[Epoch 90] speed: 5612 samples/sec	time cost: 238.359574
[Epoch 90] validation: err-top1=0.359335 err-top5=0.143575
Epoch[91] Batch [49]	Speed: 5330.770559 samples/sec	accuracy=0.608105	lr=0.059396
Epoch[91] Batch [99]	Speed: 5628.040355 samples/sec	accuracy=0.605195	lr=0.059240
Epoch[91] Batch [149]	Speed: 5465.659257 samples/sec	accuracy=0.604811	lr=0.059085
Epoch[91] Batch [199]	Speed: 5542.745080 samples/sec	accuracy=0.605000	lr=0.058930
Epoch[91] Batch [249]	Speed: 5679.993345 samples/sec	accuracy=0.604590	lr=0.058776
Epoch[91] Batch [299]	Speed: 6060.959657 samples/sec	accuracy=0.604046	lr=0.058621
Epoch[91] Batch [349]	Speed: 5501.071924 samples/sec	accuracy=0.603931	lr=0.058467
Epoch[91] Batch [399]	Speed: 5630.059140 samples/sec	accuracy=0.603630	lr=0.058313
Epoch[91] Batch [449]	Speed: 5976.783153 samples/sec	accuracy=0.603411	lr=0.058158
Epoch[91] Batch [499]	Speed: 5512.965227 samples/sec	accuracy=0.603129	lr=0.058005
Epoch[91] Batch [549]	Speed: 5325.143720 samples/sec	accuracy=0.603116	lr=0.057851
Epoch[91] Batch [599]	Speed: 5453.085392 samples/sec	accuracy=0.603112	lr=0.057697
Epoch[91] Batch [649]	Speed: 5188.948757 samples/sec	accuracy=0.603105	lr=0.057544
Epoch[91] Batch [699]	Speed: 5240.224961 samples/sec	accuracy=0.602879	lr=0.057391
Epoch[91] Batch [749]	Speed: 5893.945737 samples/sec	accuracy=0.602552	lr=0.057238
Epoch[91] Batch [799]	Speed: 6291.431178 samples/sec	accuracy=0.602159	lr=0.057085
Epoch[91] Batch [849]	Speed: 5741.535590 samples/sec	accuracy=0.602148	lr=0.056932
Epoch[91] Batch [899]	Speed: 6012.655001 samples/sec	accuracy=0.601883	lr=0.056780
Epoch[91] Batch [949]	Speed: 5702.682440 samples/sec	accuracy=0.601714	lr=0.056627
Epoch[91] Batch [999]	Speed: 4820.514001 samples/sec	accuracy=0.601625	lr=0.056475
Epoch[91] Batch [1049]	Speed: 4763.844499 samples/sec	accuracy=0.601671	lr=0.056323
Epoch[91] Batch [1099]	Speed: 5712.597492 samples/sec	accuracy=0.601466	lr=0.056171
Epoch[91] Batch [1149]	Speed: 4216.034529 samples/sec	accuracy=0.601137	lr=0.056020
Epoch[91] Batch [1199]	Speed: 5699.307269 samples/sec	accuracy=0.600972	lr=0.055868
Epoch[91] Batch [1249]	Speed: 5445.022609 samples/sec	accuracy=0.600744	lr=0.055717
[Epoch 91] training: accuracy=0.600744
[Epoch 91] speed: 5470 samples/sec	time cost: 245.828912
[Epoch 91] validation: err-top1=0.370177 err-top5=0.148876
Epoch[92] Batch [49]	Speed: 5130.104742 samples/sec	accuracy=0.612871	lr=0.055563
Epoch[92] Batch [99]	Speed: 5057.974600 samples/sec	accuracy=0.611406	lr=0.055412
Epoch[92] Batch [149]	Speed: 4523.994900 samples/sec	accuracy=0.610065	lr=0.055261
Epoch[92] Batch [199]	Speed: 5565.028004 samples/sec	accuracy=0.609448	lr=0.055111
Epoch[92] Batch [249]	Speed: 5892.387066 samples/sec	accuracy=0.608430	lr=0.054960
Epoch[92] Batch [299]	Speed: 5789.536644 samples/sec	accuracy=0.608125	lr=0.054810
Epoch[92] Batch [349]	Speed: 5912.763152 samples/sec	accuracy=0.608253	lr=0.054660
Epoch[92] Batch [399]	Speed: 5679.990190 samples/sec	accuracy=0.607288	lr=0.054510
Epoch[92] Batch [449]	Speed: 5867.673643 samples/sec	accuracy=0.606977	lr=0.054360
Epoch[92] Batch [499]	Speed: 5779.515647 samples/sec	accuracy=0.606666	lr=0.054210
Epoch[92] Batch [549]	Speed: 5198.072891 samples/sec	accuracy=0.606577	lr=0.054061
Epoch[92] Batch [599]	Speed: 5951.496632 samples/sec	accuracy=0.606284	lr=0.053912
Epoch[92] Batch [649]	Speed: 5702.002877 samples/sec	accuracy=0.606020	lr=0.053763
Epoch[92] Batch [699]	Speed: 6073.265973 samples/sec	accuracy=0.606095	lr=0.053614
Epoch[92] Batch [749]	Speed: 5835.832655 samples/sec	accuracy=0.605755	lr=0.053465
Epoch[92] Batch [799]	Speed: 5429.133887 samples/sec	accuracy=0.605627	lr=0.053317
Epoch[92] Batch [849]	Speed: 5903.433411 samples/sec	accuracy=0.605256	lr=0.053168
Epoch[92] Batch [899]	Speed: 4794.172706 samples/sec	accuracy=0.605086	lr=0.053020
Epoch[92] Batch [949]	Speed: 5418.046256 samples/sec	accuracy=0.604866	lr=0.052872
Epoch[92] Batch [999]	Speed: 5348.547103 samples/sec	accuracy=0.604835	lr=0.052724
Epoch[92] Batch [1049]	Speed: 5927.317937 samples/sec	accuracy=0.604667	lr=0.052577
Epoch[92] Batch [1099]	Speed: 5220.006406 samples/sec	accuracy=0.604671	lr=0.052429
Epoch[92] Batch [1149]	Speed: 5765.355488 samples/sec	accuracy=0.604428	lr=0.052282
Epoch[92] Batch [1199]	Speed: 5482.958205 samples/sec	accuracy=0.604271	lr=0.052135
Epoch[92] Batch [1249]	Speed: 5749.622163 samples/sec	accuracy=0.604352	lr=0.051988
[Epoch 92] training: accuracy=0.604355
[Epoch 92] speed: 5526 samples/sec	time cost: 241.293229
[Epoch 92] validation: err-top1=0.368264 err-top5=0.149773
Epoch[93] Batch [49]	Speed: 4907.239327 samples/sec	accuracy=0.615723	lr=0.051838
Epoch[93] Batch [99]	Speed: 5356.351081 samples/sec	accuracy=0.615791	lr=0.051691
Epoch[93] Batch [149]	Speed: 5687.144120 samples/sec	accuracy=0.614447	lr=0.051545
Epoch[93] Batch [199]	Speed: 5225.387200 samples/sec	accuracy=0.613257	lr=0.051399
Epoch[93] Batch [249]	Speed: 5696.036107 samples/sec	accuracy=0.612289	lr=0.051253
Epoch[93] Batch [299]	Speed: 4862.456195 samples/sec	accuracy=0.611374	lr=0.051107
Epoch[93] Batch [349]	Speed: 5402.307285 samples/sec	accuracy=0.610868	lr=0.050961
Epoch[93] Batch [399]	Speed: 4763.351244 samples/sec	accuracy=0.610754	lr=0.050816
Epoch[93] Batch [449]	Speed: 5858.840851 samples/sec	accuracy=0.610694	lr=0.050670
Epoch[93] Batch [499]	Speed: 5215.645063 samples/sec	accuracy=0.610230	lr=0.050525
Epoch[93] Batch [549]	Speed: 4960.918747 samples/sec	accuracy=0.609893	lr=0.050380
Epoch[93] Batch [599]	Speed: 4793.809053 samples/sec	accuracy=0.609451	lr=0.050235
Epoch[93] Batch [649]	Speed: 5482.448546 samples/sec	accuracy=0.609237	lr=0.050091
Epoch[93] Batch [699]	Speed: 4889.628569 samples/sec	accuracy=0.609229	lr=0.049946
Epoch[93] Batch [749]	Speed: 4984.902184 samples/sec	accuracy=0.609116	lr=0.049802
Epoch[93] Batch [799]	Speed: 5551.320932 samples/sec	accuracy=0.608805	lr=0.049658
Epoch[93] Batch [849]	Speed: 5189.604579 samples/sec	accuracy=0.608532	lr=0.049514
Epoch[93] Batch [899]	Speed: 4727.043252 samples/sec	accuracy=0.608270	lr=0.049370
Epoch[93] Batch [949]	Speed: 5729.731020 samples/sec	accuracy=0.608079	lr=0.049227
Epoch[93] Batch [999]	Speed: 4331.572375 samples/sec	accuracy=0.607887	lr=0.049083
Epoch[93] Batch [1049]	Speed: 5041.130061 samples/sec	accuracy=0.607769	lr=0.048940
Epoch[93] Batch [1099]	Speed: 5853.025946 samples/sec	accuracy=0.607592	lr=0.048797
Epoch[93] Batch [1149]	Speed: 5508.043903 samples/sec	accuracy=0.607565	lr=0.048654
Epoch[93] Batch [1199]	Speed: 6014.857436 samples/sec	accuracy=0.607442	lr=0.048511
Epoch[93] Batch [1249]	Speed: 4749.177071 samples/sec	accuracy=0.607363	lr=0.048369
[Epoch 93] training: accuracy=0.607376
[Epoch 93] speed: 5190 samples/sec	time cost: 258.028211
[Epoch 93] validation: err-top1=0.354472 err-top5=0.139425
Epoch[94] Batch [49]	Speed: 5585.013190 samples/sec	accuracy=0.615273	lr=0.048224
Epoch[94] Batch [99]	Speed: 5648.718942 samples/sec	accuracy=0.614404	lr=0.048082
Epoch[94] Batch [149]	Speed: 5063.672716 samples/sec	accuracy=0.616328	lr=0.047940
Epoch[94] Batch [199]	Speed: 5480.246802 samples/sec	accuracy=0.615762	lr=0.047798
Epoch[94] Batch [249]	Speed: 4977.131438 samples/sec	accuracy=0.616059	lr=0.047656
Epoch[94] Batch [299]	Speed: 6027.378986 samples/sec	accuracy=0.615296	lr=0.047515
Epoch[94] Batch [349]	Speed: 5500.337420 samples/sec	accuracy=0.614311	lr=0.047374
Epoch[94] Batch [399]	Speed: 5955.855358 samples/sec	accuracy=0.614265	lr=0.047233
Epoch[94] Batch [449]	Speed: 5830.182019 samples/sec	accuracy=0.614041	lr=0.047092
Epoch[94] Batch [499]	Speed: 4565.063962 samples/sec	accuracy=0.613594	lr=0.046951
Epoch[94] Batch [549]	Speed: 5071.175758 samples/sec	accuracy=0.613244	lr=0.046811
Epoch[94] Batch [599]	Speed: 5724.692390 samples/sec	accuracy=0.613032	lr=0.046670
Epoch[94] Batch [649]	Speed: 5228.805628 samples/sec	accuracy=0.613102	lr=0.046530
Epoch[94] Batch [699]	Speed: 5537.753047 samples/sec	accuracy=0.612800	lr=0.046390
Epoch[94] Batch [749]	Speed: 4724.700752 samples/sec	accuracy=0.612637	lr=0.046251
Epoch[94] Batch [799]	Speed: 5969.398830 samples/sec	accuracy=0.612549	lr=0.046111
Epoch[94] Batch [849]	Speed: 5936.384003 samples/sec	accuracy=0.612344	lr=0.045972
Epoch[94] Batch [899]	Speed: 4383.558219 samples/sec	accuracy=0.612042	lr=0.045833
Epoch[94] Batch [949]	Speed: 5626.465081 samples/sec	accuracy=0.611685	lr=0.045693
Epoch[94] Batch [999]	Speed: 4793.460327 samples/sec	accuracy=0.611508	lr=0.045555
Epoch[94] Batch [1049]	Speed: 5626.480707 samples/sec	accuracy=0.611337	lr=0.045416
Epoch[94] Batch [1099]	Speed: 5162.209264 samples/sec	accuracy=0.611301	lr=0.045278
Epoch[94] Batch [1149]	Speed: 5558.272618 samples/sec	accuracy=0.611268	lr=0.045139
Epoch[94] Batch [1199]	Speed: 5009.068169 samples/sec	accuracy=0.611266	lr=0.045001
Epoch[94] Batch [1249]	Speed: 5382.439589 samples/sec	accuracy=0.611019	lr=0.044863
[Epoch 94] training: accuracy=0.611021
[Epoch 94] speed: 5329 samples/sec	time cost: 251.040878
[Epoch 94] validation: err-top1=0.358299 err-top5=0.142478
Epoch[95] Batch [49]	Speed: 5855.599112 samples/sec	accuracy=0.618398	lr=0.044723
Epoch[95] Batch [99]	Speed: 4858.660431 samples/sec	accuracy=0.616641	lr=0.044585
Epoch[95] Batch [149]	Speed: 5533.544986 samples/sec	accuracy=0.617116	lr=0.044448
Epoch[95] Batch [199]	Speed: 5461.689884 samples/sec	accuracy=0.618027	lr=0.044311
Epoch[95] Batch [249]	Speed: 5205.672910 samples/sec	accuracy=0.618055	lr=0.044174
Epoch[95] Batch [299]	Speed: 4796.047607 samples/sec	accuracy=0.617653	lr=0.044037
Epoch[95] Batch [349]	Speed: 5695.150445 samples/sec	accuracy=0.617132	lr=0.043900
Epoch[95] Batch [399]	Speed: 5762.874616 samples/sec	accuracy=0.616841	lr=0.043764
Epoch[95] Batch [449]	Speed: 4866.245765 samples/sec	accuracy=0.617029	lr=0.043628
Epoch[95] Batch [499]	Speed: 5019.538009 samples/sec	accuracy=0.617049	lr=0.043492
Epoch[95] Batch [549]	Speed: 5419.610235 samples/sec	accuracy=0.616655	lr=0.043356
Epoch[95] Batch [599]	Speed: 4938.775615 samples/sec	accuracy=0.616374	lr=0.043220
Epoch[95] Batch [649]	Speed: 5328.938585 samples/sec	accuracy=0.615804	lr=0.043085
Epoch[95] Batch [699]	Speed: 5206.689314 samples/sec	accuracy=0.615771	lr=0.042949
Epoch[95] Batch [749]	Speed: 5886.231283 samples/sec	accuracy=0.615668	lr=0.042814
Epoch[95] Batch [799]	Speed: 5981.578847 samples/sec	accuracy=0.615558	lr=0.042679
Epoch[95] Batch [849]	Speed: 6111.864473 samples/sec	accuracy=0.615527	lr=0.042545
Epoch[95] Batch [899]	Speed: 4985.781761 samples/sec	accuracy=0.615570	lr=0.042410
Epoch[95] Batch [949]	Speed: 5402.673296 samples/sec	accuracy=0.615392	lr=0.042276
Epoch[95] Batch [999]	Speed: 5362.288048 samples/sec	accuracy=0.615123	lr=0.042141
Epoch[95] Batch [1049]	Speed: 5637.002686 samples/sec	accuracy=0.614867	lr=0.042007
Epoch[95] Batch [1099]	Speed: 5315.428593 samples/sec	accuracy=0.614868	lr=0.041874
Epoch[95] Batch [1149]	Speed: 5528.752831 samples/sec	accuracy=0.614702	lr=0.041740
Epoch[95] Batch [1199]	Speed: 5349.320641 samples/sec	accuracy=0.614591	lr=0.041607
Epoch[95] Batch [1249]	Speed: 5292.188180 samples/sec	accuracy=0.614457	lr=0.041473
[Epoch 95] training: accuracy=0.614443
[Epoch 95] speed: 5361 samples/sec	time cost: 250.235752
[Epoch 95] validation: err-top1=0.354014 err-top5=0.137675
Epoch[96] Batch [49]	Speed: 5051.745435 samples/sec	accuracy=0.621758	lr=0.041338
Epoch[96] Batch [99]	Speed: 5778.532621 samples/sec	accuracy=0.622207	lr=0.041205
Epoch[96] Batch [149]	Speed: 5860.593734 samples/sec	accuracy=0.621191	lr=0.041072
Epoch[96] Batch [199]	Speed: 6042.201952 samples/sec	accuracy=0.620562	lr=0.040940
Epoch[96] Batch [249]	Speed: 5212.828983 samples/sec	accuracy=0.620621	lr=0.040807
Epoch[96] Batch [299]	Speed: 5673.130134 samples/sec	accuracy=0.620462	lr=0.040675
Epoch[96] Batch [349]	Speed: 5177.862681 samples/sec	accuracy=0.620352	lr=0.040543
Epoch[96] Batch [399]	Speed: 5855.256967 samples/sec	accuracy=0.620413	lr=0.040412
Epoch[96] Batch [449]	Speed: 5814.796622 samples/sec	accuracy=0.619865	lr=0.040280
Epoch[96] Batch [499]	Speed: 5174.249443 samples/sec	accuracy=0.619613	lr=0.040149
Epoch[96] Batch [549]	Speed: 4399.592239 samples/sec	accuracy=0.619036	lr=0.040018
Epoch[96] Batch [599]	Speed: 5439.683113 samples/sec	accuracy=0.619246	lr=0.039887
Epoch[96] Batch [649]	Speed: 5056.717132 samples/sec	accuracy=0.619172	lr=0.039756
Epoch[96] Batch [699]	Speed: 5028.039107 samples/sec	accuracy=0.619187	lr=0.039625
Epoch[96] Batch [749]	Speed: 4809.709313 samples/sec	accuracy=0.619064	lr=0.039495
Epoch[96] Batch [799]	Speed: 4798.965498 samples/sec	accuracy=0.618928	lr=0.039365
Epoch[96] Batch [849]	Speed: 5819.179674 samples/sec	accuracy=0.618990	lr=0.039235
Epoch[96] Batch [899]	Speed: 5868.512264 samples/sec	accuracy=0.618677	lr=0.039105
Epoch[96] Batch [949]	Speed: 5431.064255 samples/sec	accuracy=0.618530	lr=0.038975
Epoch[96] Batch [999]	Speed: 5646.237948 samples/sec	accuracy=0.618445	lr=0.038846
Epoch[96] Batch [1049]	Speed: 5799.641475 samples/sec	accuracy=0.618466	lr=0.038717
Epoch[96] Batch [1099]	Speed: 5439.938037 samples/sec	accuracy=0.618366	lr=0.038588
Epoch[96] Batch [1149]	Speed: 5402.034813 samples/sec	accuracy=0.618160	lr=0.038459
Epoch[96] Batch [1199]	Speed: 5237.848928 samples/sec	accuracy=0.618187	lr=0.038330
Epoch[96] Batch [1249]	Speed: 5452.116692 samples/sec	accuracy=0.618038	lr=0.038202
[Epoch 96] training: accuracy=0.618040
[Epoch 96] speed: 5373 samples/sec	time cost: 249.812883
[Epoch 96] validation: err-top1=0.343272 err-top5=0.129564
Epoch[97] Batch [49]	Speed: 4320.418449 samples/sec	accuracy=0.630469	lr=0.038071
Epoch[97] Batch [99]	Speed: 4560.804359 samples/sec	accuracy=0.629062	lr=0.037943
Epoch[97] Batch [149]	Speed: 5346.015019 samples/sec	accuracy=0.627747	lr=0.037815
Epoch[97] Batch [199]	Speed: 5926.174258 samples/sec	accuracy=0.626841	lr=0.037687
Epoch[97] Batch [249]	Speed: 5866.846000 samples/sec	accuracy=0.626602	lr=0.037560
Epoch[97] Batch [299]	Speed: 3726.725313 samples/sec	accuracy=0.625527	lr=0.037432
Epoch[97] Batch [349]	Speed: 5832.640083 samples/sec	accuracy=0.624919	lr=0.037305
Epoch[97] Batch [399]	Speed: 5600.416700 samples/sec	accuracy=0.624441	lr=0.037178
Epoch[97] Batch [449]	Speed: 5440.804541 samples/sec	accuracy=0.624212	lr=0.037052
Epoch[97] Batch [499]	Speed: 5136.673629 samples/sec	accuracy=0.623947	lr=0.036925
Epoch[97] Batch [549]	Speed: 5787.124126 samples/sec	accuracy=0.623722	lr=0.036799
Epoch[97] Batch [599]	Speed: 5582.271911 samples/sec	accuracy=0.623340	lr=0.036673
Epoch[97] Batch [649]	Speed: 4634.039003 samples/sec	accuracy=0.623235	lr=0.036547
Epoch[97] Batch [699]	Speed: 5523.638195 samples/sec	accuracy=0.623292	lr=0.036421
Epoch[97] Batch [749]	Speed: 5569.223007 samples/sec	accuracy=0.623177	lr=0.036296
Epoch[97] Batch [799]	Speed: 5426.519734 samples/sec	accuracy=0.622961	lr=0.036170
Epoch[97] Batch [849]	Speed: 5162.730872 samples/sec	accuracy=0.622768	lr=0.036045
Epoch[97] Batch [899]	Speed: 4734.090135 samples/sec	accuracy=0.622439	lr=0.035920
Epoch[97] Batch [949]	Speed: 6221.384918 samples/sec	accuracy=0.622593	lr=0.035795
Epoch[97] Batch [999]	Speed: 5718.133387 samples/sec	accuracy=0.622679	lr=0.035671
Epoch[97] Batch [1049]	Speed: 5482.577736 samples/sec	accuracy=0.622539	lr=0.035546
Epoch[97] Batch [1099]	Speed: 5813.746156 samples/sec	accuracy=0.622396	lr=0.035422
Epoch[97] Batch [1149]	Speed: 5466.053660 samples/sec	accuracy=0.622308	lr=0.035298
Epoch[97] Batch [1199]	Speed: 5231.306501 samples/sec	accuracy=0.622257	lr=0.035174
Epoch[97] Batch [1249]	Speed: 5807.031556 samples/sec	accuracy=0.622175	lr=0.035051
[Epoch 97] training: accuracy=0.622180
[Epoch 97] speed: 5282 samples/sec	time cost: 253.968030
[Epoch 97] validation: err-top1=0.344049 err-top5=0.132075
Epoch[98] Batch [49]	Speed: 5307.535396 samples/sec	accuracy=0.629453	lr=0.034925
Epoch[98] Batch [99]	Speed: 5743.572113 samples/sec	accuracy=0.631436	lr=0.034802
Epoch[98] Batch [149]	Speed: 5919.711582 samples/sec	accuracy=0.630221	lr=0.034679
Epoch[98] Batch [199]	Speed: 6042.107771 samples/sec	accuracy=0.629268	lr=0.034556
Epoch[98] Batch [249]	Speed: 5950.392737 samples/sec	accuracy=0.628473	lr=0.034433
Epoch[98] Batch [299]	Speed: 5534.264712 samples/sec	accuracy=0.628372	lr=0.034311
Epoch[98] Batch [349]	Speed: 5625.311649 samples/sec	accuracy=0.627720	lr=0.034189
Epoch[98] Batch [399]	Speed: 5957.307978 samples/sec	accuracy=0.627090	lr=0.034067
Epoch[98] Batch [449]	Speed: 4543.032498 samples/sec	accuracy=0.626860	lr=0.033945
Epoch[98] Batch [499]	Speed: 5741.826499 samples/sec	accuracy=0.626391	lr=0.033823
Epoch[98] Batch [549]	Speed: 5319.246247 samples/sec	accuracy=0.626007	lr=0.033702
Epoch[98] Batch [599]	Speed: 4316.277622 samples/sec	accuracy=0.625895	lr=0.033581
Epoch[98] Batch [649]	Speed: 5273.849833 samples/sec	accuracy=0.625786	lr=0.033460
Epoch[98] Batch [699]	Speed: 5769.607716 samples/sec	accuracy=0.625451	lr=0.033339
Epoch[98] Batch [749]	Speed: 5727.434524 samples/sec	accuracy=0.625215	lr=0.033218
Epoch[98] Batch [799]	Speed: 5333.381203 samples/sec	accuracy=0.625067	lr=0.033098
Epoch[98] Batch [849]	Speed: 4998.237972 samples/sec	accuracy=0.624992	lr=0.032978
Epoch[98] Batch [899]	Speed: 5594.653102 samples/sec	accuracy=0.624757	lr=0.032858
Epoch[98] Batch [949]	Speed: 5495.997100 samples/sec	accuracy=0.624828	lr=0.032738
Epoch[98] Batch [999]	Speed: 4808.283808 samples/sec	accuracy=0.624797	lr=0.032618
Epoch[98] Batch [1049]	Speed: 5034.754602 samples/sec	accuracy=0.624783	lr=0.032499
Epoch[98] Batch [1099]	Speed: 5499.160480 samples/sec	accuracy=0.624637	lr=0.032380
Epoch[98] Batch [1149]	Speed: 5090.170052 samples/sec	accuracy=0.624537	lr=0.032261
Epoch[98] Batch [1199]	Speed: 5006.116946 samples/sec	accuracy=0.624549	lr=0.032142
Epoch[98] Batch [1249]	Speed: 5270.818167 samples/sec	accuracy=0.624364	lr=0.032023
[Epoch 98] training: accuracy=0.624355
[Epoch 98] speed: 5350 samples/sec	time cost: 250.224008
[Epoch 98] validation: err-top1=0.342916 err-top5=0.130554
Epoch[99] Batch [49]	Speed: 4741.461216 samples/sec	accuracy=0.633574	lr=0.031902
Epoch[99] Batch [99]	Speed: 5514.099237 samples/sec	accuracy=0.632715	lr=0.031784
Epoch[99] Batch [149]	Speed: 5730.227451 samples/sec	accuracy=0.633099	lr=0.031666
Epoch[99] Batch [199]	Speed: 5849.420492 samples/sec	accuracy=0.632451	lr=0.031548
Epoch[99] Batch [249]	Speed: 5939.396804 samples/sec	accuracy=0.632668	lr=0.031431
Epoch[99] Batch [299]	Speed: 4452.064722 samples/sec	accuracy=0.632373	lr=0.031313
Epoch[99] Batch [349]	Speed: 5308.253028 samples/sec	accuracy=0.632040	lr=0.031196
Epoch[99] Batch [399]	Speed: 5512.397195 samples/sec	accuracy=0.631318	lr=0.031079
Epoch[99] Batch [449]	Speed: 5655.509479 samples/sec	accuracy=0.630818	lr=0.030962
Epoch[99] Batch [499]	Speed: 5717.904705 samples/sec	accuracy=0.630115	lr=0.030846
Epoch[99] Batch [549]	Speed: 5352.891460 samples/sec	accuracy=0.629938	lr=0.030729
Epoch[99] Batch [599]	Speed: 5721.116601 samples/sec	accuracy=0.630143	lr=0.030613
Epoch[99] Batch [649]	Speed: 5768.856165 samples/sec	accuracy=0.629833	lr=0.030497
Epoch[99] Batch [699]	Speed: 5960.892991 samples/sec	accuracy=0.629623	lr=0.030381
Epoch[99] Batch [749]	Speed: 6015.590366 samples/sec	accuracy=0.629461	lr=0.030266
Epoch[99] Batch [799]	Speed: 5677.271224 samples/sec	accuracy=0.629235	lr=0.030150
Epoch[99] Batch [849]	Speed: 5465.272979 samples/sec	accuracy=0.629106	lr=0.030035
Epoch[99] Batch [899]	Speed: 5076.085242 samples/sec	accuracy=0.628889	lr=0.029920
Epoch[99] Batch [949]	Speed: 5615.510967 samples/sec	accuracy=0.628713	lr=0.029805
Epoch[99] Batch [999]	Speed: 4627.044446 samples/sec	accuracy=0.628709	lr=0.029691
Epoch[99] Batch [1049]	Speed: 5515.283714 samples/sec	accuracy=0.628815	lr=0.029576
Epoch[99] Batch [1099]	Speed: 5492.864473 samples/sec	accuracy=0.628752	lr=0.029462
Epoch[99] Batch [1149]	Speed: 5866.927263 samples/sec	accuracy=0.628691	lr=0.029348
Epoch[99] Batch [1199]	Speed: 5381.083190 samples/sec	accuracy=0.628443	lr=0.029234
Epoch[99] Batch [1249]	Speed: 5599.535262 samples/sec	accuracy=0.628120	lr=0.029121
[Epoch 99] training: accuracy=0.628102
[Epoch 99] speed: 5464 samples/sec	time cost: 244.138199
[Epoch 99] validation: err-top1=0.339126 err-top5=0.130441
Epoch[100] Batch [49]	Speed: 5377.694306 samples/sec	accuracy=0.639961	lr=0.029005
Epoch[100] Batch [99]	Speed: 5903.141474 samples/sec	accuracy=0.638467	lr=0.028892
Epoch[100] Batch [149]	Speed: 5766.681971 samples/sec	accuracy=0.637240	lr=0.028779
Epoch[100] Batch [199]	Speed: 4786.875636 samples/sec	accuracy=0.636274	lr=0.028666
Epoch[100] Batch [249]	Speed: 5996.990920 samples/sec	accuracy=0.635527	lr=0.028554
Epoch[100] Batch [299]	Speed: 4789.117864 samples/sec	accuracy=0.634854	lr=0.028441
Epoch[100] Batch [349]	Speed: 5048.692329 samples/sec	accuracy=0.634492	lr=0.028329
Epoch[100] Batch [399]	Speed: 5474.379005 samples/sec	accuracy=0.634180	lr=0.028217
Epoch[100] Batch [449]	Speed: 5823.297092 samples/sec	accuracy=0.633698	lr=0.028105
Epoch[100] Batch [499]	Speed: 5951.008948 samples/sec	accuracy=0.633836	lr=0.027994
Epoch[100] Batch [549]	Speed: 5957.694879 samples/sec	accuracy=0.633656	lr=0.027883
Epoch[100] Batch [599]	Speed: 5917.901147 samples/sec	accuracy=0.633413	lr=0.027772
Epoch[100] Batch [649]	Speed: 5659.403384 samples/sec	accuracy=0.633209	lr=0.027661
Epoch[100] Batch [699]	Speed: 5044.202342 samples/sec	accuracy=0.632962	lr=0.027550
Epoch[100] Batch [749]	Speed: 5488.204202 samples/sec	accuracy=0.632664	lr=0.027439
Epoch[100] Batch [799]	Speed: 5428.972067 samples/sec	accuracy=0.632354	lr=0.027329
Epoch[100] Batch [849]	Speed: 4926.016852 samples/sec	accuracy=0.632173	lr=0.027219
Epoch[100] Batch [899]	Speed: 5239.460280 samples/sec	accuracy=0.632174	lr=0.027109
Epoch[100] Batch [949]	Speed: 5043.128999 samples/sec	accuracy=0.632075	lr=0.026999
Epoch[100] Batch [999]	Speed: 5225.748196 samples/sec	accuracy=0.632036	lr=0.026890
Epoch[100] Batch [1049]	Speed: 4722.605038 samples/sec	accuracy=0.632259	lr=0.026781
Epoch[100] Batch [1099]	Speed: 5310.529088 samples/sec	accuracy=0.632146	lr=0.026672
Epoch[100] Batch [1149]	Speed: 5312.206237 samples/sec	accuracy=0.632110	lr=0.026563
Epoch[100] Batch [1199]	Speed: 5329.645880 samples/sec	accuracy=0.632044	lr=0.026454
Epoch[100] Batch [1249]	Speed: 5222.886647 samples/sec	accuracy=0.631995	lr=0.026346
[Epoch 100] training: accuracy=0.631979
[Epoch 100] speed: 5353 samples/sec	time cost: 251.050269
[Epoch 100] validation: err-top1=0.349450 err-top5=0.135144
Epoch[101] Batch [49]	Speed: 5496.679092 samples/sec	accuracy=0.642207	lr=0.026235
Epoch[101] Batch [99]	Speed: 5945.396531 samples/sec	accuracy=0.639922	lr=0.026127
Epoch[101] Batch [149]	Speed: 6096.168401 samples/sec	accuracy=0.640352	lr=0.026020
Epoch[101] Batch [199]	Speed: 4841.616276 samples/sec	accuracy=0.639678	lr=0.025912
Epoch[101] Batch [249]	Speed: 5667.481008 samples/sec	accuracy=0.639094	lr=0.025805
Epoch[101] Batch [299]	Speed: 5040.736971 samples/sec	accuracy=0.638805	lr=0.025697
Epoch[101] Batch [349]	Speed: 5615.910111 samples/sec	accuracy=0.638577	lr=0.025590
Epoch[101] Batch [399]	Speed: 5607.726401 samples/sec	accuracy=0.638044	lr=0.025484
Epoch[101] Batch [449]	Speed: 5031.146956 samples/sec	accuracy=0.637441	lr=0.025377
Epoch[101] Batch [499]	Speed: 5873.461157 samples/sec	accuracy=0.637051	lr=0.025271
Epoch[101] Batch [549]	Speed: 5663.914765 samples/sec	accuracy=0.636596	lr=0.025165
Epoch[101] Batch [599]	Speed: 5919.508428 samples/sec	accuracy=0.636457	lr=0.025059
Epoch[101] Batch [649]	Speed: 5260.959024 samples/sec	accuracy=0.636420	lr=0.024953
Epoch[101] Batch [699]	Speed: 4710.549599 samples/sec	accuracy=0.636346	lr=0.024847
Epoch[101] Batch [749]	Speed: 5050.108625 samples/sec	accuracy=0.636260	lr=0.024742
Epoch[101] Batch [799]	Speed: 5822.163365 samples/sec	accuracy=0.636241	lr=0.024637
Epoch[101] Batch [849]	Speed: 5399.992356 samples/sec	accuracy=0.636124	lr=0.024532
Epoch[101] Batch [899]	Speed: 4514.454949 samples/sec	accuracy=0.636015	lr=0.024427
Epoch[101] Batch [949]	Speed: 4573.677897 samples/sec	accuracy=0.635916	lr=0.024323
Epoch[101] Batch [999]	Speed: 5641.824780 samples/sec	accuracy=0.635922	lr=0.024219
Epoch[101] Batch [1049]	Speed: 5580.680672 samples/sec	accuracy=0.635901	lr=0.024115
Epoch[101] Batch [1099]	Speed: 4520.521520 samples/sec	accuracy=0.635837	lr=0.024011
Epoch[101] Batch [1149]	Speed: 4558.780843 samples/sec	accuracy=0.635681	lr=0.023907
Epoch[101] Batch [1199]	Speed: 5615.841821 samples/sec	accuracy=0.635733	lr=0.023804
Epoch[101] Batch [1249]	Speed: 5763.638222 samples/sec	accuracy=0.635670	lr=0.023700
[Epoch 101] training: accuracy=0.635672
[Epoch 101] speed: 5298 samples/sec	time cost: 251.417529
[Epoch 101] validation: err-top1=0.327627 err-top5=0.122987
Epoch[102] Batch [49]	Speed: 5724.349960 samples/sec	accuracy=0.642988	lr=0.023595
Epoch[102] Batch [99]	Speed: 5613.419696 samples/sec	accuracy=0.641338	lr=0.023493
Epoch[102] Batch [149]	Speed: 5604.178633 samples/sec	accuracy=0.641471	lr=0.023390
Epoch[102] Batch [199]	Speed: 5732.052614 samples/sec	accuracy=0.641509	lr=0.023288
Epoch[102] Batch [249]	Speed: 5697.212531 samples/sec	accuracy=0.641094	lr=0.023185
Epoch[102] Batch [299]	Speed: 5922.515554 samples/sec	accuracy=0.640788	lr=0.023083
Epoch[102] Batch [349]	Speed: 4843.953687 samples/sec	accuracy=0.640293	lr=0.022982
Epoch[102] Batch [399]	Speed: 5896.164035 samples/sec	accuracy=0.640398	lr=0.022880
Epoch[102] Batch [449]	Speed: 5057.939099 samples/sec	accuracy=0.640267	lr=0.022779
Epoch[102] Batch [499]	Speed: 5885.253719 samples/sec	accuracy=0.639996	lr=0.022678
Epoch[102] Batch [549]	Speed: 5339.000386 samples/sec	accuracy=0.639751	lr=0.022577
Epoch[102] Batch [599]	Speed: 5237.155186 samples/sec	accuracy=0.639736	lr=0.022476
Epoch[102] Batch [649]	Speed: 5553.873600 samples/sec	accuracy=0.639672	lr=0.022376
Epoch[102] Batch [699]	Speed: 5006.885886 samples/sec	accuracy=0.639348	lr=0.022276
Epoch[102] Batch [749]	Speed: 4126.882587 samples/sec	accuracy=0.639171	lr=0.022175
Epoch[102] Batch [799]	Speed: 5817.110937 samples/sec	accuracy=0.639176	lr=0.022076
Epoch[102] Batch [849]	Speed: 4685.959645 samples/sec	accuracy=0.638987	lr=0.021976
Epoch[102] Batch [899]	Speed: 5885.688744 samples/sec	accuracy=0.639090	lr=0.021877
Epoch[102] Batch [949]	Speed: 6116.235064 samples/sec	accuracy=0.639045	lr=0.021777
Epoch[102] Batch [999]	Speed: 5023.466958 samples/sec	accuracy=0.638981	lr=0.021678
Epoch[102] Batch [1049]	Speed: 5281.136392 samples/sec	accuracy=0.638947	lr=0.021580
Epoch[102] Batch [1099]	Speed: 4892.890559 samples/sec	accuracy=0.639025	lr=0.021481
Epoch[102] Batch [1149]	Speed: 4456.393927 samples/sec	accuracy=0.638817	lr=0.021383
Epoch[102] Batch [1199]	Speed: 5725.699468 samples/sec	accuracy=0.638689	lr=0.021285
Epoch[102] Batch [1249]	Speed: 5718.795631 samples/sec	accuracy=0.638725	lr=0.021187
[Epoch 102] training: accuracy=0.638741
[Epoch 102] speed: 5336 samples/sec	time cost: 249.650750
[Epoch 102] validation: err-top1=0.325415 err-top5=0.119639
Epoch[103] Batch [49]	Speed: 5231.516140 samples/sec	accuracy=0.649219	lr=0.021087
Epoch[103] Batch [99]	Speed: 4579.811708 samples/sec	accuracy=0.646504	lr=0.020990
Epoch[103] Batch [149]	Speed: 5729.967682 samples/sec	accuracy=0.646126	lr=0.020892
Epoch[103] Batch [199]	Speed: 6071.266004 samples/sec	accuracy=0.646064	lr=0.020795
Epoch[103] Batch [249]	Speed: 4453.213299 samples/sec	accuracy=0.645473	lr=0.020698
Epoch[103] Batch [299]	Speed: 5907.741822 samples/sec	accuracy=0.644756	lr=0.020602
Epoch[103] Batch [349]	Speed: 5767.745549 samples/sec	accuracy=0.644919	lr=0.020505
Epoch[103] Batch [399]	Speed: 5766.935323 samples/sec	accuracy=0.644253	lr=0.020409
Epoch[103] Batch [449]	Speed: 5960.969269 samples/sec	accuracy=0.643431	lr=0.020313
Epoch[103] Batch [499]	Speed: 5433.697088 samples/sec	accuracy=0.643291	lr=0.020217
Epoch[103] Batch [549]	Speed: 5935.765238 samples/sec	accuracy=0.643505	lr=0.020122
Epoch[103] Batch [599]	Speed: 4624.997199 samples/sec	accuracy=0.643625	lr=0.020026
Epoch[103] Batch [649]	Speed: 4821.888193 samples/sec	accuracy=0.643382	lr=0.019931
Epoch[103] Batch [699]	Speed: 5783.446269 samples/sec	accuracy=0.643255	lr=0.019836
Epoch[103] Batch [749]	Speed: 5853.287580 samples/sec	accuracy=0.643193	lr=0.019742
Epoch[103] Batch [799]	Speed: 4761.126317 samples/sec	accuracy=0.643097	lr=0.019647
Epoch[103] Batch [849]	Speed: 5125.049775 samples/sec	accuracy=0.643073	lr=0.019553
Epoch[103] Batch [899]	Speed: 5042.225640 samples/sec	accuracy=0.642841	lr=0.019459
Epoch[103] Batch [949]	Speed: 5239.087034 samples/sec	accuracy=0.642742	lr=0.019365
Epoch[103] Batch [999]	Speed: 6100.267192 samples/sec	accuracy=0.642718	lr=0.019271
Epoch[103] Batch [1049]	Speed: 5874.273795 samples/sec	accuracy=0.642600	lr=0.019178
Epoch[103] Batch [1099]	Speed: 5659.488399 samples/sec	accuracy=0.642509	lr=0.019085
Epoch[103] Batch [1149]	Speed: 5971.645737 samples/sec	accuracy=0.642608	lr=0.018992
Epoch[103] Batch [1199]	Speed: 5999.170664 samples/sec	accuracy=0.642547	lr=0.018899
Epoch[103] Batch [1249]	Speed: 5641.610313 samples/sec	accuracy=0.642513	lr=0.018806
[Epoch 103] training: accuracy=0.642517
[Epoch 103] speed: 5435 samples/sec	time cost: 245.146840
[Epoch 103] validation: err-top1=0.335918 err-top5=0.129185
Epoch[104] Batch [49]	Speed: 5116.422058 samples/sec	accuracy=0.643730	lr=0.018712
Epoch[104] Batch [99]	Speed: 5945.186508 samples/sec	accuracy=0.645693	lr=0.018620
Epoch[104] Batch [149]	Speed: 4672.824197 samples/sec	accuracy=0.646120	lr=0.018528
Epoch[104] Batch [199]	Speed: 5029.335707 samples/sec	accuracy=0.646650	lr=0.018436
Epoch[104] Batch [249]	Speed: 6023.579085 samples/sec	accuracy=0.646477	lr=0.018345
Epoch[104] Batch [299]	Speed: 5666.612875 samples/sec	accuracy=0.646624	lr=0.018254
Epoch[104] Batch [349]	Speed: 6050.846454 samples/sec	accuracy=0.646719	lr=0.018163
Epoch[104] Batch [399]	Speed: 5447.843542 samples/sec	accuracy=0.646750	lr=0.018072
Epoch[104] Batch [449]	Speed: 4801.980144 samples/sec	accuracy=0.647088	lr=0.017981
Epoch[104] Batch [499]	Speed: 6017.656845 samples/sec	accuracy=0.646959	lr=0.017891
Epoch[104] Batch [549]	Speed: 5242.886400 samples/sec	accuracy=0.646818	lr=0.017801
Epoch[104] Batch [599]	Speed: 5126.708853 samples/sec	accuracy=0.646784	lr=0.017711
Epoch[104] Batch [649]	Speed: 5040.627291 samples/sec	accuracy=0.646668	lr=0.017621
Epoch[104] Batch [699]	Speed: 6081.034681 samples/sec	accuracy=0.646494	lr=0.017532
Epoch[104] Batch [749]	Speed: 6012.125599 samples/sec	accuracy=0.646197	lr=0.017442
Epoch[104] Batch [799]	Speed: 6006.299091 samples/sec	accuracy=0.646261	lr=0.017353
Epoch[104] Batch [849]	Speed: 5521.061572 samples/sec	accuracy=0.646344	lr=0.017264
Epoch[104] Batch [899]	Speed: 5048.269696 samples/sec	accuracy=0.646235	lr=0.017176
Epoch[104] Batch [949]	Speed: 5569.022545 samples/sec	accuracy=0.646070	lr=0.017087
Epoch[104] Batch [999]	Speed: 5673.430190 samples/sec	accuracy=0.646072	lr=0.016999
Epoch[104] Batch [1049]	Speed: 6114.598058 samples/sec	accuracy=0.646004	lr=0.016911
Epoch[104] Batch [1099]	Speed: 5725.296014 samples/sec	accuracy=0.645938	lr=0.016823
Epoch[104] Batch [1149]	Speed: 3982.935598 samples/sec	accuracy=0.645951	lr=0.016736
Epoch[104] Batch [1199]	Speed: 5259.100525 samples/sec	accuracy=0.645961	lr=0.016648
Epoch[104] Batch [1249]	Speed: 5457.914780 samples/sec	accuracy=0.645948	lr=0.016561
[Epoch 104] training: accuracy=0.645936
[Epoch 104] speed: 5401 samples/sec	time cost: 248.413119
[Epoch 104] validation: err-top1=0.320272 err-top5=0.118917
Epoch[105] Batch [49]	Speed: 5329.926708 samples/sec	accuracy=0.656680	lr=0.016473
Epoch[105] Batch [99]	Speed: 5520.680480 samples/sec	accuracy=0.656074	lr=0.016386
Epoch[105] Batch [149]	Speed: 4964.884068 samples/sec	accuracy=0.654987	lr=0.016299
Epoch[105] Batch [199]	Speed: 5135.207880 samples/sec	accuracy=0.654458	lr=0.016213
Epoch[105] Batch [249]	Speed: 4988.029678 samples/sec	accuracy=0.653133	lr=0.016127
Epoch[105] Batch [299]	Speed: 4075.027932 samples/sec	accuracy=0.652360	lr=0.016041
Epoch[105] Batch [349]	Speed: 5884.653630 samples/sec	accuracy=0.652257	lr=0.015956
Epoch[105] Batch [399]	Speed: 5617.144760 samples/sec	accuracy=0.651897	lr=0.015870
Epoch[105] Batch [449]	Speed: 5622.641675 samples/sec	accuracy=0.651793	lr=0.015785
Epoch[105] Batch [499]	Speed: 5214.863097 samples/sec	accuracy=0.651516	lr=0.015700
Epoch[105] Batch [549]	Speed: 5581.387036 samples/sec	accuracy=0.651555	lr=0.015616
Epoch[105] Batch [599]	Speed: 5311.195643 samples/sec	accuracy=0.651356	lr=0.015531
Epoch[105] Batch [649]	Speed: 5968.760888 samples/sec	accuracy=0.651100	lr=0.015447
Epoch[105] Batch [699]	Speed: 5815.290108 samples/sec	accuracy=0.651069	lr=0.015363
Epoch[105] Batch [749]	Speed: 5540.175187 samples/sec	accuracy=0.650897	lr=0.015279
Epoch[105] Batch [799]	Speed: 5207.974113 samples/sec	accuracy=0.650850	lr=0.015195
Epoch[105] Batch [849]	Speed: 5049.677680 samples/sec	accuracy=0.650563	lr=0.015112
Epoch[105] Batch [899]	Speed: 5464.821532 samples/sec	accuracy=0.650632	lr=0.015029
Epoch[105] Batch [949]	Speed: 4982.204285 samples/sec	accuracy=0.650311	lr=0.014946
Epoch[105] Batch [999]	Speed: 5398.961662 samples/sec	accuracy=0.650320	lr=0.014863
Epoch[105] Batch [1049]	Speed: 5807.343274 samples/sec	accuracy=0.650304	lr=0.014781
Epoch[105] Batch [1099]	Speed: 5281.447591 samples/sec	accuracy=0.650186	lr=0.014699
Epoch[105] Batch [1149]	Speed: 5532.009761 samples/sec	accuracy=0.649989	lr=0.014616
Epoch[105] Batch [1199]	Speed: 5717.743177 samples/sec	accuracy=0.649941	lr=0.014535
Epoch[105] Batch [1249]	Speed: 5717.931500 samples/sec	accuracy=0.649827	lr=0.014453
[Epoch 105] training: accuracy=0.649842
[Epoch 105] speed: 5352 samples/sec	time cost: 251.340167
[Epoch 105] validation: err-top1=0.315669 err-top5=0.116908
Epoch[106] Batch [49]	Speed: 4518.346104 samples/sec	accuracy=0.654355	lr=0.014370
Epoch[106] Batch [99]	Speed: 4915.072391 samples/sec	accuracy=0.654932	lr=0.014289
Epoch[106] Batch [149]	Speed: 5037.657446 samples/sec	accuracy=0.655111	lr=0.014208
Epoch[106] Batch [199]	Speed: 5413.524441 samples/sec	accuracy=0.654844	lr=0.014127
Epoch[106] Batch [249]	Speed: 5146.045150 samples/sec	accuracy=0.654754	lr=0.014047
Epoch[106] Batch [299]	Speed: 5811.203658 samples/sec	accuracy=0.654329	lr=0.013966
Epoch[106] Batch [349]	Speed: 5395.845366 samples/sec	accuracy=0.654434	lr=0.013886
Epoch[106] Batch [399]	Speed: 5226.456600 samples/sec	accuracy=0.654502	lr=0.013806
Epoch[106] Batch [449]	Speed: 5247.456535 samples/sec	accuracy=0.654638	lr=0.013727
Epoch[106] Batch [499]	Speed: 5835.738137 samples/sec	accuracy=0.654459	lr=0.013647
Epoch[106] Batch [549]	Speed: 5466.102077 samples/sec	accuracy=0.654522	lr=0.013568
Epoch[106] Batch [599]	Speed: 5064.144626 samples/sec	accuracy=0.654461	lr=0.013489
Epoch[106] Batch [649]	Speed: 5335.132331 samples/sec	accuracy=0.654225	lr=0.013411
Epoch[106] Batch [699]	Speed: 6045.688547 samples/sec	accuracy=0.653951	lr=0.013332
Epoch[106] Batch [749]	Speed: 6443.940525 samples/sec	accuracy=0.654117	lr=0.013254
Epoch[106] Batch [799]	Speed: 5640.992494 samples/sec	accuracy=0.654045	lr=0.013176
Epoch[106] Batch [849]	Speed: 4848.208487 samples/sec	accuracy=0.653851	lr=0.013098
Epoch[106] Batch [899]	Speed: 5518.600235 samples/sec	accuracy=0.653637	lr=0.013020
Epoch[106] Batch [949]	Speed: 6113.608096 samples/sec	accuracy=0.653679	lr=0.012943
Epoch[106] Batch [999]	Speed: 5892.211488 samples/sec	accuracy=0.653544	lr=0.012866
Epoch[106] Batch [1049]	Speed: 5406.788591 samples/sec	accuracy=0.653507	lr=0.012789
Epoch[106] Batch [1099]	Speed: 5274.601267 samples/sec	accuracy=0.653485	lr=0.012712
Epoch[106] Batch [1149]	Speed: 5130.931983 samples/sec	accuracy=0.653492	lr=0.012636
Epoch[106] Batch [1199]	Speed: 5570.497900 samples/sec	accuracy=0.653570	lr=0.012559
Epoch[106] Batch [1249]	Speed: 6023.914148 samples/sec	accuracy=0.653477	lr=0.012483
[Epoch 106] training: accuracy=0.653480
[Epoch 106] speed: 5412 samples/sec	time cost: 246.267396
[Epoch 106] validation: err-top1=0.314712 err-top5=0.116629
Epoch[107] Batch [49]	Speed: 5333.173916 samples/sec	accuracy=0.658555	lr=0.012406
Epoch[107] Batch [99]	Speed: 5537.830304 samples/sec	accuracy=0.659102	lr=0.012330
Epoch[107] Batch [149]	Speed: 4488.158743 samples/sec	accuracy=0.658626	lr=0.012255
Epoch[107] Batch [199]	Speed: 5389.767616 samples/sec	accuracy=0.658657	lr=0.012180
Epoch[107] Batch [249]	Speed: 5174.474857 samples/sec	accuracy=0.658984	lr=0.012105
Epoch[107] Batch [299]	Speed: 5649.096666 samples/sec	accuracy=0.658457	lr=0.012030
Epoch[107] Batch [349]	Speed: 5963.752185 samples/sec	accuracy=0.658605	lr=0.011956
Epoch[107] Batch [399]	Speed: 4516.783175 samples/sec	accuracy=0.658188	lr=0.011881
Epoch[107] Batch [449]	Speed: 5094.155291 samples/sec	accuracy=0.657793	lr=0.011807
Epoch[107] Batch [499]	Speed: 6086.527033 samples/sec	accuracy=0.657867	lr=0.011734
Epoch[107] Batch [549]	Speed: 4943.630553 samples/sec	accuracy=0.657804	lr=0.011660
Epoch[107] Batch [599]	Speed: 5292.347427 samples/sec	accuracy=0.657759	lr=0.011587
Epoch[107] Batch [649]	Speed: 5679.409449 samples/sec	accuracy=0.657611	lr=0.011513
Epoch[107] Batch [699]	Speed: 5896.273310 samples/sec	accuracy=0.657430	lr=0.011441
Epoch[107] Batch [749]	Speed: 5252.433248 samples/sec	accuracy=0.657487	lr=0.011368
Epoch[107] Batch [799]	Speed: 5157.898708 samples/sec	accuracy=0.657212	lr=0.011295
Epoch[107] Batch [849]	Speed: 4202.995499 samples/sec	accuracy=0.657028	lr=0.011223
Epoch[107] Batch [899]	Speed: 5833.017455 samples/sec	accuracy=0.656985	lr=0.011151
Epoch[107] Batch [949]	Speed: 5144.044272 samples/sec	accuracy=0.656779	lr=0.011079
Epoch[107] Batch [999]	Speed: 5645.859864 samples/sec	accuracy=0.656804	lr=0.011008
Epoch[107] Batch [1049]	Speed: 6221.864206 samples/sec	accuracy=0.656744	lr=0.010936
Epoch[107] Batch [1099]	Speed: 5586.109468 samples/sec	accuracy=0.656660	lr=0.010865
Epoch[107] Batch [1149]	Speed: 5885.687292 samples/sec	accuracy=0.656606	lr=0.010794
Epoch[107] Batch [1199]	Speed: 5685.327872 samples/sec	accuracy=0.656610	lr=0.010724
Epoch[107] Batch [1249]	Speed: 5896.156426 samples/sec	accuracy=0.656521	lr=0.010653
[Epoch 107] training: accuracy=0.656514
[Epoch 107] speed: 5363 samples/sec	time cost: 250.939607
[Epoch 107] validation: err-top1=0.312859 err-top5=0.113042
Epoch[108] Batch [49]	Speed: 5153.524107 samples/sec	accuracy=0.661074	lr=0.010582
Epoch[108] Batch [99]	Speed: 5844.655967 samples/sec	accuracy=0.660986	lr=0.010512
Epoch[108] Batch [149]	Speed: 5474.329185 samples/sec	accuracy=0.661888	lr=0.010442
Epoch[108] Batch [199]	Speed: 5428.061992 samples/sec	accuracy=0.662271	lr=0.010373
Epoch[108] Batch [249]	Speed: 5335.623718 samples/sec	accuracy=0.662438	lr=0.010303
Epoch[108] Batch [299]	Speed: 5182.565490 samples/sec	accuracy=0.662194	lr=0.010234
Epoch[108] Batch [349]	Speed: 6319.141120 samples/sec	accuracy=0.661872	lr=0.010165
Epoch[108] Batch [399]	Speed: 5751.031359 samples/sec	accuracy=0.661460	lr=0.010097
Epoch[108] Batch [449]	Speed: 5514.321960 samples/sec	accuracy=0.661788	lr=0.010028
Epoch[108] Batch [499]	Speed: 5296.091888 samples/sec	accuracy=0.661471	lr=0.009960
Epoch[108] Batch [549]	Speed: 5828.655621 samples/sec	accuracy=0.661158	lr=0.009892
Epoch[108] Batch [599]	Speed: 5765.533339 samples/sec	accuracy=0.661287	lr=0.009824
Epoch[108] Batch [649]	Speed: 4909.924431 samples/sec	accuracy=0.661232	lr=0.009757
Epoch[108] Batch [699]	Speed: 5439.348442 samples/sec	accuracy=0.660939	lr=0.009690
Epoch[108] Batch [749]	Speed: 4078.168669 samples/sec	accuracy=0.660926	lr=0.009623
Epoch[108] Batch [799]	Speed: 5041.269941 samples/sec	accuracy=0.660779	lr=0.009556
Epoch[108] Batch [849]	Speed: 5629.181480 samples/sec	accuracy=0.660604	lr=0.009489
Epoch[108] Batch [899]	Speed: 4758.422728 samples/sec	accuracy=0.660736	lr=0.009423
Epoch[108] Batch [949]	Speed: 5041.109115 samples/sec	accuracy=0.660747	lr=0.009357
Epoch[108] Batch [999]	Speed: 4683.783325 samples/sec	accuracy=0.660610	lr=0.009291
Epoch[108] Batch [1049]	Speed: 6066.744666 samples/sec	accuracy=0.660701	lr=0.009225
Epoch[108] Batch [1099]	Speed: 4399.219111 samples/sec	accuracy=0.660670	lr=0.009160
Epoch[108] Batch [1149]	Speed: 5128.335570 samples/sec	accuracy=0.660459	lr=0.009095
Epoch[108] Batch [1199]	Speed: 5599.305603 samples/sec	accuracy=0.660487	lr=0.009030
Epoch[108] Batch [1249]	Speed: 6209.031890 samples/sec	accuracy=0.660505	lr=0.008965
[Epoch 108] training: accuracy=0.660504
[Epoch 108] speed: 5293 samples/sec	time cost: 251.602953
[Epoch 108] validation: err-top1=0.307677 err-top5=0.111388
Epoch[109] Batch [49]	Speed: 5413.706631 samples/sec	accuracy=0.665039	lr=0.008899
Epoch[109] Batch [99]	Speed: 6087.723439 samples/sec	accuracy=0.663662	lr=0.008835
Epoch[109] Batch [149]	Speed: 5793.592041 samples/sec	accuracy=0.664010	lr=0.008771
Epoch[109] Batch [199]	Speed: 5763.586865 samples/sec	accuracy=0.663291	lr=0.008707
Epoch[109] Batch [249]	Speed: 5855.332641 samples/sec	accuracy=0.664242	lr=0.008643
Epoch[109] Batch [299]	Speed: 5878.599114 samples/sec	accuracy=0.664134	lr=0.008580
Epoch[109] Batch [349]	Speed: 5376.408540 samples/sec	accuracy=0.664060	lr=0.008517
Epoch[109] Batch [399]	Speed: 5509.592570 samples/sec	accuracy=0.663936	lr=0.008454
Epoch[109] Batch [449]	Speed: 5446.805692 samples/sec	accuracy=0.664097	lr=0.008391
Epoch[109] Batch [499]	Speed: 4961.281719 samples/sec	accuracy=0.664064	lr=0.008329
Epoch[109] Batch [549]	Speed: 5972.329640 samples/sec	accuracy=0.664062	lr=0.008266
Epoch[109] Batch [599]	Speed: 6068.641501 samples/sec	accuracy=0.663905	lr=0.008204
Epoch[109] Batch [649]	Speed: 5769.513936 samples/sec	accuracy=0.664004	lr=0.008143
Epoch[109] Batch [699]	Speed: 4294.734178 samples/sec	accuracy=0.663725	lr=0.008081
Epoch[109] Batch [749]	Speed: 5591.223168 samples/sec	accuracy=0.663824	lr=0.008020
Epoch[109] Batch [799]	Speed: 5337.412667 samples/sec	accuracy=0.663925	lr=0.007959
Epoch[109] Batch [849]	Speed: 4791.142733 samples/sec	accuracy=0.663736	lr=0.007898
Epoch[109] Batch [899]	Speed: 5676.042109 samples/sec	accuracy=0.663529	lr=0.007837
Epoch[109] Batch [949]	Speed: 5594.170994 samples/sec	accuracy=0.663378	lr=0.007777
Epoch[109] Batch [999]	Speed: 5577.293482 samples/sec	accuracy=0.663374	lr=0.007716
Epoch[109] Batch [1049]	Speed: 5774.861170 samples/sec	accuracy=0.663557	lr=0.007656
Epoch[109] Batch [1099]	Speed: 5599.502119 samples/sec	accuracy=0.663584	lr=0.007597
Epoch[109] Batch [1149]	Speed: 5527.824650 samples/sec	accuracy=0.663400	lr=0.007537
Epoch[109] Batch [1199]	Speed: 4533.091595 samples/sec	accuracy=0.663331	lr=0.007478
Epoch[109] Batch [1249]	Speed: 5332.293025 samples/sec	accuracy=0.663219	lr=0.007419
[Epoch 109] training: accuracy=0.663217
[Epoch 109] speed: 5455 samples/sec	time cost: 244.874120
[Epoch 109] validation: err-top1=0.307458 err-top5=0.112962
Epoch[110] Batch [49]	Speed: 5430.975526 samples/sec	accuracy=0.668477	lr=0.007359
Epoch[110] Batch [99]	Speed: 5082.315676 samples/sec	accuracy=0.667705	lr=0.007300
Epoch[110] Batch [149]	Speed: 5204.910836 samples/sec	accuracy=0.667383	lr=0.007242
Epoch[110] Batch [199]	Speed: 5135.660179 samples/sec	accuracy=0.667017	lr=0.007184
Epoch[110] Batch [249]	Speed: 5594.956284 samples/sec	accuracy=0.666430	lr=0.007126
Epoch[110] Batch [299]	Speed: 5361.250950 samples/sec	accuracy=0.666130	lr=0.007068
Epoch[110] Batch [349]	Speed: 5041.517176 samples/sec	accuracy=0.666164	lr=0.007011
Epoch[110] Batch [399]	Speed: 5446.628865 samples/sec	accuracy=0.666431	lr=0.006954
Epoch[110] Batch [449]	Speed: 4884.703843 samples/sec	accuracy=0.666526	lr=0.006897
Epoch[110] Batch [499]	Speed: 5956.381338 samples/sec	accuracy=0.666553	lr=0.006840
Epoch[110] Batch [549]	Speed: 5050.382026 samples/sec	accuracy=0.666562	lr=0.006784
Epoch[110] Batch [599]	Speed: 5474.322905 samples/sec	accuracy=0.666493	lr=0.006727
Epoch[110] Batch [649]	Speed: 4920.423595 samples/sec	accuracy=0.666406	lr=0.006671
Epoch[110] Batch [699]	Speed: 5154.606978 samples/sec	accuracy=0.666408	lr=0.006615
Epoch[110] Batch [749]	Speed: 5678.171001 samples/sec	accuracy=0.666260	lr=0.006560
Epoch[110] Batch [799]	Speed: 5739.210297 samples/sec	accuracy=0.666200	lr=0.006504
Epoch[110] Batch [849]	Speed: 5871.841853 samples/sec	accuracy=0.666313	lr=0.006449
Epoch[110] Batch [899]	Speed: 5672.961685 samples/sec	accuracy=0.666399	lr=0.006394
Epoch[110] Batch [949]	Speed: 5637.721604 samples/sec	accuracy=0.666387	lr=0.006340
Epoch[110] Batch [999]	Speed: 5450.865934 samples/sec	accuracy=0.666198	lr=0.006285
Epoch[110] Batch [1049]	Speed: 5256.704774 samples/sec	accuracy=0.666083	lr=0.006231
Epoch[110] Batch [1099]	Speed: 5620.605992 samples/sec	accuracy=0.666207	lr=0.006177
Epoch[110] Batch [1149]	Speed: 5402.950318 samples/sec	accuracy=0.666127	lr=0.006123
Epoch[110] Batch [1199]	Speed: 5423.354706 samples/sec	accuracy=0.666091	lr=0.006070
Epoch[110] Batch [1249]	Speed: 5436.184262 samples/sec	accuracy=0.666190	lr=0.006017
[Epoch 110] training: accuracy=0.666183
[Epoch 110] speed: 5376 samples/sec	time cost: 248.268957
[Epoch 110] validation: err-top1=0.306600 err-top5=0.110799
Epoch[111] Batch [49]	Speed: 5603.402741 samples/sec	accuracy=0.674102	lr=0.005963
Epoch[111] Batch [99]	Speed: 5571.599327 samples/sec	accuracy=0.671777	lr=0.005910
Epoch[111] Batch [149]	Speed: 5782.761026 samples/sec	accuracy=0.672285	lr=0.005857
Epoch[111] Batch [199]	Speed: 5781.878236 samples/sec	accuracy=0.670869	lr=0.005805
Epoch[111] Batch [249]	Speed: 5764.647909 samples/sec	accuracy=0.670414	lr=0.005753
Epoch[111] Batch [299]	Speed: 4986.310466 samples/sec	accuracy=0.670179	lr=0.005701
Epoch[111] Batch [349]	Speed: 5505.121771 samples/sec	accuracy=0.669944	lr=0.005649
Epoch[111] Batch [399]	Speed: 5167.686378 samples/sec	accuracy=0.669971	lr=0.005598
Epoch[111] Batch [449]	Speed: 5887.587183 samples/sec	accuracy=0.669850	lr=0.005547
Epoch[111] Batch [499]	Speed: 5659.074388 samples/sec	accuracy=0.669791	lr=0.005496
Epoch[111] Batch [549]	Speed: 5911.784894 samples/sec	accuracy=0.669554	lr=0.005445
Epoch[111] Batch [599]	Speed: 5755.252604 samples/sec	accuracy=0.669422	lr=0.005394
Epoch[111] Batch [649]	Speed: 4996.735514 samples/sec	accuracy=0.669426	lr=0.005344
Epoch[111] Batch [699]	Speed: 5434.977117 samples/sec	accuracy=0.669248	lr=0.005294
Epoch[111] Batch [749]	Speed: 5959.118300 samples/sec	accuracy=0.669240	lr=0.005244
Epoch[111] Batch [799]	Speed: 5892.530802 samples/sec	accuracy=0.669344	lr=0.005195
Epoch[111] Batch [849]	Speed: 5172.267313 samples/sec	accuracy=0.669264	lr=0.005145
Epoch[111] Batch [899]	Speed: 5579.259639 samples/sec	accuracy=0.669238	lr=0.005096
Epoch[111] Batch [949]	Speed: 5058.441278 samples/sec	accuracy=0.669385	lr=0.005047
Epoch[111] Batch [999]	Speed: 5846.610953 samples/sec	accuracy=0.669369	lr=0.004999
Epoch[111] Batch [1049]	Speed: 5870.552412 samples/sec	accuracy=0.669408	lr=0.004950
Epoch[111] Batch [1099]	Speed: 5099.541669 samples/sec	accuracy=0.669482	lr=0.004902
Epoch[111] Batch [1149]	Speed: 5336.667634 samples/sec	accuracy=0.669552	lr=0.004854
Epoch[111] Batch [1199]	Speed: 5662.549128 samples/sec	accuracy=0.669548	lr=0.004807
Epoch[111] Batch [1249]	Speed: 5021.664756 samples/sec	accuracy=0.669620	lr=0.004759
[Epoch 111] training: accuracy=0.669627
[Epoch 111] speed: 5507 samples/sec	time cost: 242.799735
[Epoch 111] validation: err-top1=0.304349 err-top5=0.109714
Epoch[112] Batch [49]	Speed: 5150.538614 samples/sec	accuracy=0.670977	lr=0.004711
Epoch[112] Batch [99]	Speed: 5506.966187 samples/sec	accuracy=0.672109	lr=0.004664
Epoch[112] Batch [149]	Speed: 5647.475277 samples/sec	accuracy=0.671947	lr=0.004617
Epoch[112] Batch [199]	Speed: 5916.582109 samples/sec	accuracy=0.672886	lr=0.004571
Epoch[112] Batch [249]	Speed: 5643.692981 samples/sec	accuracy=0.673160	lr=0.004524
Epoch[112] Batch [299]	Speed: 5828.689950 samples/sec	accuracy=0.673154	lr=0.004478
Epoch[112] Batch [349]	Speed: 6166.065564 samples/sec	accuracy=0.673072	lr=0.004432
Epoch[112] Batch [399]	Speed: 5617.160922 samples/sec	accuracy=0.672900	lr=0.004387
Epoch[112] Batch [449]	Speed: 5356.925090 samples/sec	accuracy=0.672795	lr=0.004341
Epoch[112] Batch [499]	Speed: 5829.034377 samples/sec	accuracy=0.672967	lr=0.004296
Epoch[112] Batch [549]	Speed: 6129.424134 samples/sec	accuracy=0.672923	lr=0.004251
Epoch[112] Batch [599]	Speed: 5166.008010 samples/sec	accuracy=0.673091	lr=0.004207
Epoch[112] Batch [649]	Speed: 5737.090574 samples/sec	accuracy=0.672897	lr=0.004162
Epoch[112] Batch [699]	Speed: 5746.832910 samples/sec	accuracy=0.672892	lr=0.004118
Epoch[112] Batch [749]	Speed: 5077.566292 samples/sec	accuracy=0.672835	lr=0.004074
Epoch[112] Batch [799]	Speed: 5831.763062 samples/sec	accuracy=0.672738	lr=0.004030
Epoch[112] Batch [849]	Speed: 5177.171506 samples/sec	accuracy=0.672679	lr=0.003987
Epoch[112] Batch [899]	Speed: 4333.346708 samples/sec	accuracy=0.672604	lr=0.003944
Epoch[112] Batch [949]	Speed: 5476.813337 samples/sec	accuracy=0.672628	lr=0.003901
Epoch[112] Batch [999]	Speed: 4646.776109 samples/sec	accuracy=0.672586	lr=0.003858
Epoch[112] Batch [1049]	Speed: 5323.990264 samples/sec	accuracy=0.672467	lr=0.003815
Epoch[112] Batch [1099]	Speed: 5169.193127 samples/sec	accuracy=0.672508	lr=0.003773
Epoch[112] Batch [1149]	Speed: 5713.758876 samples/sec	accuracy=0.672385	lr=0.003731
Epoch[112] Batch [1199]	Speed: 5853.222648 samples/sec	accuracy=0.672354	lr=0.003689
Epoch[112] Batch [1249]	Speed: 5322.307909 samples/sec	accuracy=0.672253	lr=0.003647
[Epoch 112] training: accuracy=0.672241
[Epoch 112] speed: 5453 samples/sec	time cost: 245.184811
[Epoch 112] validation: err-top1=0.304030 err-top5=0.108658
Epoch[113] Batch [49]	Speed: 5046.629563 samples/sec	accuracy=0.678730	lr=0.003605
Epoch[113] Batch [99]	Speed: 5522.060041 samples/sec	accuracy=0.678066	lr=0.003564
Epoch[113] Batch [149]	Speed: 5728.756603 samples/sec	accuracy=0.676634	lr=0.003523
Epoch[113] Batch [199]	Speed: 4364.686973 samples/sec	accuracy=0.675923	lr=0.003482
Epoch[113] Batch [249]	Speed: 5754.006917 samples/sec	accuracy=0.675742	lr=0.003442
Epoch[113] Batch [299]	Speed: 5765.710582 samples/sec	accuracy=0.675192	lr=0.003402
Epoch[113] Batch [349]	Speed: 5943.253544 samples/sec	accuracy=0.675497	lr=0.003362
Epoch[113] Batch [399]	Speed: 5309.694126 samples/sec	accuracy=0.675593	lr=0.003322
Epoch[113] Batch [449]	Speed: 5957.422340 samples/sec	accuracy=0.676061	lr=0.003282
Epoch[113] Batch [499]	Speed: 5666.269136 samples/sec	accuracy=0.675203	lr=0.003243
Epoch[113] Batch [549]	Speed: 5693.737401 samples/sec	accuracy=0.675059	lr=0.003204
Epoch[113] Batch [599]	Speed: 5640.070243 samples/sec	accuracy=0.675145	lr=0.003165
Epoch[113] Batch [649]	Speed: 5232.367107 samples/sec	accuracy=0.675146	lr=0.003127
Epoch[113] Batch [699]	Speed: 4110.750350 samples/sec	accuracy=0.674973	lr=0.003088
Epoch[113] Batch [749]	Speed: 5438.896034 samples/sec	accuracy=0.674565	lr=0.003050
Epoch[113] Batch [799]	Speed: 5803.514051 samples/sec	accuracy=0.674371	lr=0.003012
Epoch[113] Batch [849]	Speed: 4768.770491 samples/sec	accuracy=0.674382	lr=0.002975
Epoch[113] Batch [899]	Speed: 4843.352492 samples/sec	accuracy=0.674378	lr=0.002937
Epoch[113] Batch [949]	Speed: 5665.143115 samples/sec	accuracy=0.674543	lr=0.002900
Epoch[113] Batch [999]	Speed: 5592.413635 samples/sec	accuracy=0.674445	lr=0.002863
Epoch[113] Batch [1049]	Speed: 5773.687079 samples/sec	accuracy=0.674385	lr=0.002826
Epoch[113] Batch [1099]	Speed: 5599.119465 samples/sec	accuracy=0.674308	lr=0.002790
Epoch[113] Batch [1149]	Speed: 5659.799991 samples/sec	accuracy=0.674346	lr=0.002754
Epoch[113] Batch [1199]	Speed: 5789.701161 samples/sec	accuracy=0.674325	lr=0.002718
Epoch[113] Batch [1249]	Speed: 5769.269037 samples/sec	accuracy=0.674203	lr=0.002682
[Epoch 113] training: accuracy=0.674215
[Epoch 113] speed: 5406 samples/sec	time cost: 248.502598
[Epoch 113] validation: err-top1=0.302615 err-top5=0.107143
Epoch[114] Batch [49]	Speed: 4633.807920 samples/sec	accuracy=0.680586	lr=0.002646
Epoch[114] Batch [99]	Speed: 5444.590790 samples/sec	accuracy=0.680293	lr=0.002611
Epoch[114] Batch [149]	Speed: 5710.298764 samples/sec	accuracy=0.679004	lr=0.002575
Epoch[114] Batch [199]	Speed: 5725.387599 samples/sec	accuracy=0.678364	lr=0.002541
Epoch[114] Batch [249]	Speed: 5640.149789 samples/sec	accuracy=0.677949	lr=0.002506
Epoch[114] Batch [299]	Speed: 6078.981928 samples/sec	accuracy=0.676986	lr=0.002472
Epoch[114] Batch [349]	Speed: 5668.085792 samples/sec	accuracy=0.677031	lr=0.002438
Epoch[114] Batch [399]	Speed: 6030.855772 samples/sec	accuracy=0.676931	lr=0.002404
Epoch[114] Batch [449]	Speed: 5784.822541 samples/sec	accuracy=0.676584	lr=0.002370
Epoch[114] Batch [499]	Speed: 6020.606887 samples/sec	accuracy=0.676438	lr=0.002337
Epoch[114] Batch [549]	Speed: 5964.196739 samples/sec	accuracy=0.676234	lr=0.002304
Epoch[114] Batch [599]	Speed: 5789.055009 samples/sec	accuracy=0.676121	lr=0.002271
Epoch[114] Batch [649]	Speed: 5485.135649 samples/sec	accuracy=0.676298	lr=0.002238
Epoch[114] Batch [699]	Speed: 5163.831647 samples/sec	accuracy=0.676203	lr=0.002205
Epoch[114] Batch [749]	Speed: 6201.274033 samples/sec	accuracy=0.676349	lr=0.002173
Epoch[114] Batch [799]	Speed: 5697.864345 samples/sec	accuracy=0.676389	lr=0.002141
Epoch[114] Batch [849]	Speed: 5691.249442 samples/sec	accuracy=0.676340	lr=0.002110
Epoch[114] Batch [899]	Speed: 5979.136172 samples/sec	accuracy=0.676376	lr=0.002078
Epoch[114] Batch [949]	Speed: 6051.105271 samples/sec	accuracy=0.676546	lr=0.002047
Epoch[114] Batch [999]	Speed: 5874.500210 samples/sec	accuracy=0.676517	lr=0.002016
Epoch[114] Batch [1049]	Speed: 6011.122433 samples/sec	accuracy=0.676395	lr=0.001985
Epoch[114] Batch [1099]	Speed: 5909.911643 samples/sec	accuracy=0.676410	lr=0.001954
Epoch[114] Batch [1149]	Speed: 4712.056587 samples/sec	accuracy=0.676324	lr=0.001924
Epoch[114] Batch [1199]	Speed: 5054.357048 samples/sec	accuracy=0.676187	lr=0.001894
Epoch[114] Batch [1249]	Speed: 4707.662454 samples/sec	accuracy=0.676238	lr=0.001864
[Epoch 114] training: accuracy=0.676231
[Epoch 114] speed: 5594 samples/sec	time cost: 238.876381
[Epoch 114] validation: err-top1=0.300084 err-top5=0.105429
Epoch[115] Batch [49]	Speed: 5246.096180 samples/sec	accuracy=0.679297	lr=0.001834
Epoch[115] Batch [99]	Speed: 5904.779551 samples/sec	accuracy=0.677061	lr=0.001804
Epoch[115] Batch [149]	Speed: 5735.589853 samples/sec	accuracy=0.677663	lr=0.001775
Epoch[115] Batch [199]	Speed: 5201.275636 samples/sec	accuracy=0.678047	lr=0.001746
Epoch[115] Batch [249]	Speed: 5761.938830 samples/sec	accuracy=0.677895	lr=0.001718
Epoch[115] Batch [299]	Speed: 5528.567796 samples/sec	accuracy=0.677594	lr=0.001689
Epoch[115] Batch [349]	Speed: 4996.252835 samples/sec	accuracy=0.678133	lr=0.001661
Epoch[115] Batch [399]	Speed: 4651.286644 samples/sec	accuracy=0.677966	lr=0.001633
Epoch[115] Batch [449]	Speed: 5860.958096 samples/sec	accuracy=0.678811	lr=0.001605
Epoch[115] Batch [499]	Speed: 5788.609186 samples/sec	accuracy=0.679193	lr=0.001578
Epoch[115] Batch [549]	Speed: 5050.914542 samples/sec	accuracy=0.679306	lr=0.001551
Epoch[115] Batch [599]	Speed: 5163.256186 samples/sec	accuracy=0.679118	lr=0.001524
Epoch[115] Batch [649]	Speed: 4347.155330 samples/sec	accuracy=0.679099	lr=0.001497
Epoch[115] Batch [699]	Speed: 5803.156168 samples/sec	accuracy=0.678817	lr=0.001470
Epoch[115] Batch [749]	Speed: 6095.001541 samples/sec	accuracy=0.678609	lr=0.001444
Epoch[115] Batch [799]	Speed: 5497.435137 samples/sec	accuracy=0.678417	lr=0.001418
Epoch[115] Batch [849]	Speed: 5369.760519 samples/sec	accuracy=0.678182	lr=0.001392
Epoch[115] Batch [899]	Speed: 5680.179941 samples/sec	accuracy=0.678110	lr=0.001366
Epoch[115] Batch [949]	Speed: 5996.039339 samples/sec	accuracy=0.677905	lr=0.001341
Epoch[115] Batch [999]	Speed: 4588.452912 samples/sec	accuracy=0.677956	lr=0.001316
Epoch[115] Batch [1049]	Speed: 4803.129781 samples/sec	accuracy=0.677914	lr=0.001291
Epoch[115] Batch [1099]	Speed: 4622.555608 samples/sec	accuracy=0.677830	lr=0.001266
Epoch[115] Batch [1149]	Speed: 5417.490644 samples/sec	accuracy=0.677750	lr=0.001242
Epoch[115] Batch [1199]	Speed: 5890.068698 samples/sec	accuracy=0.677914	lr=0.001218
Epoch[115] Batch [1249]	Speed: 5820.210967 samples/sec	accuracy=0.677972	lr=0.001194
[Epoch 115] training: accuracy=0.677977
[Epoch 115] speed: 5339 samples/sec	time cost: 249.534047
[Epoch 115] validation: err-top1=0.298111 err-top5=0.106405
Epoch[116] Batch [49]	Speed: 3853.579046 samples/sec	accuracy=0.680605	lr=0.001170
Epoch[116] Batch [99]	Speed: 5917.484665 samples/sec	accuracy=0.678711	lr=0.001146
Epoch[116] Batch [149]	Speed: 5458.979765 samples/sec	accuracy=0.679076	lr=0.001123
Epoch[116] Batch [199]	Speed: 5705.174041 samples/sec	accuracy=0.679263	lr=0.001100
Epoch[116] Batch [249]	Speed: 5928.501828 samples/sec	accuracy=0.678934	lr=0.001077
Epoch[116] Batch [299]	Speed: 5299.389528 samples/sec	accuracy=0.679333	lr=0.001055
Epoch[116] Batch [349]	Speed: 5805.166343 samples/sec	accuracy=0.679213	lr=0.001032
Epoch[116] Batch [399]	Speed: 5580.790313 samples/sec	accuracy=0.679414	lr=0.001010
Epoch[116] Batch [449]	Speed: 5150.685249 samples/sec	accuracy=0.679464	lr=0.000989
Epoch[116] Batch [499]	Speed: 5289.900578 samples/sec	accuracy=0.679324	lr=0.000967
Epoch[116] Batch [549]	Speed: 5360.191240 samples/sec	accuracy=0.679300	lr=0.000946
Epoch[116] Batch [599]	Speed: 5387.996261 samples/sec	accuracy=0.679095	lr=0.000925
Epoch[116] Batch [649]	Speed: 4915.674983 samples/sec	accuracy=0.679157	lr=0.000904
Epoch[116] Batch [699]	Speed: 5709.560003 samples/sec	accuracy=0.679568	lr=0.000883
Epoch[116] Batch [749]	Speed: 5826.788350 samples/sec	accuracy=0.679714	lr=0.000863
Epoch[116] Batch [799]	Speed: 5135.305382 samples/sec	accuracy=0.679789	lr=0.000843
Epoch[116] Batch [849]	Speed: 5216.594148 samples/sec	accuracy=0.679797	lr=0.000823
Epoch[116] Batch [899]	Speed: 5680.476686 samples/sec	accuracy=0.679678	lr=0.000803
Epoch[116] Batch [949]	Speed: 5242.959361 samples/sec	accuracy=0.679520	lr=0.000784
Epoch[116] Batch [999]	Speed: 5621.534103 samples/sec	accuracy=0.679596	lr=0.000764
Epoch[116] Batch [1049]	Speed: 4612.797930 samples/sec	accuracy=0.679604	lr=0.000745
Epoch[116] Batch [1099]	Speed: 5012.915207 samples/sec	accuracy=0.679656	lr=0.000727
Epoch[116] Batch [1149]	Speed: 4789.309689 samples/sec	accuracy=0.679804	lr=0.000708
Epoch[116] Batch [1199]	Speed: 5271.004852 samples/sec	accuracy=0.679875	lr=0.000690
Epoch[116] Batch [1249]	Speed: 4933.275529 samples/sec	accuracy=0.679948	lr=0.000672
[Epoch 116] training: accuracy=0.679957
[Epoch 116] speed: 5257 samples/sec	time cost: 255.375383
[Epoch 116] validation: err-top1=0.298645 err-top5=0.105652
Epoch[117] Batch [49]	Speed: 5305.158494 samples/sec	accuracy=0.678926	lr=0.000654
Epoch[117] Batch [99]	Speed: 5480.648068 samples/sec	accuracy=0.680537	lr=0.000636
Epoch[117] Batch [149]	Speed: 5129.043043 samples/sec	accuracy=0.679850	lr=0.000619
Epoch[117] Batch [199]	Speed: 5473.870939 samples/sec	accuracy=0.679556	lr=0.000602
Epoch[117] Batch [249]	Speed: 5686.156430 samples/sec	accuracy=0.680359	lr=0.000585
Epoch[117] Batch [299]	Speed: 5676.385086 samples/sec	accuracy=0.679899	lr=0.000569
Epoch[117] Batch [349]	Speed: 4872.359429 samples/sec	accuracy=0.679886	lr=0.000552
Epoch[117] Batch [399]	Speed: 5475.277038 samples/sec	accuracy=0.680408	lr=0.000536
Epoch[117] Batch [449]	Speed: 5333.675539 samples/sec	accuracy=0.680332	lr=0.000520
Epoch[117] Batch [499]	Speed: 5746.692965 samples/sec	accuracy=0.680289	lr=0.000505
Epoch[117] Batch [549]	Speed: 5489.981999 samples/sec	accuracy=0.680249	lr=0.000489
Epoch[117] Batch [599]	Speed: 4225.149240 samples/sec	accuracy=0.680218	lr=0.000474
Epoch[117] Batch [649]	Speed: 5367.380425 samples/sec	accuracy=0.680141	lr=0.000459
Epoch[117] Batch [699]	Speed: 5126.901869 samples/sec	accuracy=0.679907	lr=0.000445
Epoch[117] Batch [749]	Speed: 5492.961699 samples/sec	accuracy=0.679927	lr=0.000430
Epoch[117] Batch [799]	Speed: 4601.192977 samples/sec	accuracy=0.679911	lr=0.000416
Epoch[117] Batch [849]	Speed: 5641.631507 samples/sec	accuracy=0.679971	lr=0.000402
Epoch[117] Batch [899]	Speed: 5196.820011 samples/sec	accuracy=0.679807	lr=0.000388
Epoch[117] Batch [949]	Speed: 5017.582111 samples/sec	accuracy=0.679847	lr=0.000375
Epoch[117] Batch [999]	Speed: 6072.462771 samples/sec	accuracy=0.679925	lr=0.000362
Epoch[117] Batch [1049]	Speed: 5930.177266 samples/sec	accuracy=0.679887	lr=0.000349
Epoch[117] Batch [1099]	Speed: 4987.776540 samples/sec	accuracy=0.679986	lr=0.000336
Epoch[117] Batch [1149]	Speed: 5149.862001 samples/sec	accuracy=0.679990	lr=0.000323
Epoch[117] Batch [1199]	Speed: 5845.626454 samples/sec	accuracy=0.679967	lr=0.000311
Epoch[117] Batch [1249]	Speed: 5762.203979 samples/sec	accuracy=0.680125	lr=0.000299
[Epoch 117] training: accuracy=0.680123
[Epoch 117] speed: 5322 samples/sec	time cost: 250.743669
[Epoch 117] validation: err-top1=0.298350 err-top5=0.106067
Epoch[118] Batch [49]	Speed: 4924.011995 samples/sec	accuracy=0.683613	lr=0.000287
Epoch[118] Batch [99]	Speed: 5903.441038 samples/sec	accuracy=0.681455	lr=0.000275
Epoch[118] Batch [149]	Speed: 4773.341776 samples/sec	accuracy=0.680951	lr=0.000264
Epoch[118] Batch [199]	Speed: 6105.237492 samples/sec	accuracy=0.680732	lr=0.000253
Epoch[118] Batch [249]	Speed: 5544.625104 samples/sec	accuracy=0.680918	lr=0.000242
Epoch[118] Batch [299]	Speed: 5778.692937 samples/sec	accuracy=0.680788	lr=0.000231
Epoch[118] Batch [349]	Speed: 5243.216150 samples/sec	accuracy=0.680949	lr=0.000221
Epoch[118] Batch [399]	Speed: 5690.745415 samples/sec	accuracy=0.680740	lr=0.000211
Epoch[118] Batch [449]	Speed: 5343.137399 samples/sec	accuracy=0.680666	lr=0.000201
Epoch[118] Batch [499]	Speed: 5066.711935 samples/sec	accuracy=0.680648	lr=0.000191
Epoch[118] Batch [549]	Speed: 6039.775602 samples/sec	accuracy=0.680737	lr=0.000182
Epoch[118] Batch [599]	Speed: 5594.412329 samples/sec	accuracy=0.680724	lr=0.000173
Epoch[118] Batch [649]	Speed: 5891.556478 samples/sec	accuracy=0.681082	lr=0.000164
Epoch[118] Batch [699]	Speed: 5453.221926 samples/sec	accuracy=0.680960	lr=0.000155
Epoch[118] Batch [749]	Speed: 4266.814936 samples/sec	accuracy=0.680837	lr=0.000147
Epoch[118] Batch [799]	Speed: 5354.549149 samples/sec	accuracy=0.680924	lr=0.000138
Epoch[118] Batch [849]	Speed: 5596.139153 samples/sec	accuracy=0.680932	lr=0.000130
Epoch[118] Batch [899]	Speed: 5289.108565 samples/sec	accuracy=0.681120	lr=0.000123
Epoch[118] Batch [949]	Speed: 5416.919842 samples/sec	accuracy=0.681118	lr=0.000115
Epoch[118] Batch [999]	Speed: 5140.108082 samples/sec	accuracy=0.680998	lr=0.000108
Epoch[118] Batch [1049]	Speed: 4701.859338 samples/sec	accuracy=0.680949	lr=0.000101
Epoch[118] Batch [1099]	Speed: 5592.513106 samples/sec	accuracy=0.680931	lr=0.000094
Epoch[118] Batch [1149]	Speed: 5447.851696 samples/sec	accuracy=0.681090	lr=0.000087
Epoch[118] Batch [1199]	Speed: 5928.234737 samples/sec	accuracy=0.680956	lr=0.000081
Epoch[118] Batch [1249]	Speed: 5353.603393 samples/sec	accuracy=0.680830	lr=0.000075
[Epoch 118] training: accuracy=0.680816
[Epoch 118] speed: 5374 samples/sec	time cost: 248.715787
[Epoch 118] validation: err-top1=0.297573 err-top5=0.105150
Epoch[119] Batch [49]	Speed: 4634.449229 samples/sec	accuracy=0.679180	lr=0.000069
Epoch[119] Batch [99]	Speed: 5620.831223 samples/sec	accuracy=0.679180	lr=0.000063
Epoch[119] Batch [149]	Speed: 6010.147860 samples/sec	accuracy=0.679193	lr=0.000058
Epoch[119] Batch [199]	Speed: 5613.290428 samples/sec	accuracy=0.679619	lr=0.000053
Epoch[119] Batch [249]	Speed: 5769.617327 samples/sec	accuracy=0.680934	lr=0.000048
Epoch[119] Batch [299]	Speed: 5221.920415 samples/sec	accuracy=0.680775	lr=0.000043
Epoch[119] Batch [349]	Speed: 5777.390766 samples/sec	accuracy=0.681311	lr=0.000039
Epoch[119] Batch [399]	Speed: 5240.096199 samples/sec	accuracy=0.681802	lr=0.000035
Epoch[119] Batch [449]	Speed: 4814.800399 samples/sec	accuracy=0.681819	lr=0.000031
Epoch[119] Batch [499]	Speed: 5341.233813 samples/sec	accuracy=0.681863	lr=0.000027
Epoch[119] Batch [549]	Speed: 4812.988152 samples/sec	accuracy=0.681863	lr=0.000023
Epoch[119] Batch [599]	Speed: 6158.650241 samples/sec	accuracy=0.682002	lr=0.000020
Epoch[119] Batch [649]	Speed: 5134.313954 samples/sec	accuracy=0.681953	lr=0.000017
Epoch[119] Batch [699]	Speed: 4877.119384 samples/sec	accuracy=0.681948	lr=0.000015
Epoch[119] Batch [749]	Speed: 4725.769690 samples/sec	accuracy=0.681871	lr=0.000012
Epoch[119] Batch [799]	Speed: 5002.414334 samples/sec	accuracy=0.681719	lr=0.000010
Epoch[119] Batch [849]	Speed: 4996.424413 samples/sec	accuracy=0.681704	lr=0.000008
Epoch[119] Batch [899]	Speed: 5992.705749 samples/sec	accuracy=0.681687	lr=0.000006
Epoch[119] Batch [949]	Speed: 5751.431208 samples/sec	accuracy=0.681831	lr=0.000004
Epoch[119] Batch [999]	Speed: 5268.099483 samples/sec	accuracy=0.681657	lr=0.000003
Epoch[119] Batch [1049]	Speed: 5805.725216 samples/sec	accuracy=0.681799	lr=0.000002
Epoch[119] Batch [1099]	Speed: 4793.183864 samples/sec	accuracy=0.681829	lr=0.000001
Epoch[119] Batch [1149]	Speed: 5797.106602 samples/sec	accuracy=0.681848	lr=0.000000
Epoch[119] Batch [1199]	Speed: 5219.948927 samples/sec	accuracy=0.681839	lr=0.000000
Epoch[119] Batch [1249]	Speed: 5532.297212 samples/sec	accuracy=0.681825	lr=0.000000
[Epoch 119] training: accuracy=0.681838
[Epoch 119] speed: 5313 samples/sec	time cost: 253.002997
[Epoch 119] validation: err-top1=0.297333 err-top5=0.105389