Namespace(batch_norm=False, batch_size=128, data_dir='~/.mxnet/datasets/imagenet', dtype='float16', input_size=224, label_smoothing=True, last_gamma=True, log_interval=50, logging_file='resnet50_v1d_best.log', lr=0.4, lr_decay=0.1, lr_decay_epoch='40,60', lr_decay_period=0, lr_mode='cosine', mixup=False, mixup_alpha=0.2, mixup_off_epoch=0, mode='hybrid', model='resnet50_v1d', momentum=0.9, no_wd=True, num_epochs=120, num_gpus=8, num_workers=60, rec_train='/media/ramdisk/rec/train.rec', rec_train_idx='/media/ramdisk/rec/train.idx', rec_val='/media/ramdisk/rec/val.rec', rec_val_idx='/media/ramdisk/rec/val.idx', save_dir='params_resnet50_v1d_best', save_frequency=10, use_pretrained=False, use_rec=True, use_se=False, warmup_epochs=5, warmup_lr=0.0, wd=0.0001)
Epoch[0] Batch [49]	Speed: 1293.070153 samples/sec	accuracy=0.817061	lr=0.003133
Epoch[0] Batch [99]	Speed: 5306.646431 samples/sec	accuracy=0.859134	lr=0.006331
Epoch[0] Batch [149]	Speed: 5416.094121 samples/sec	accuracy=0.885739	lr=0.009528
Epoch[0] Batch [199]	Speed: 5083.178471 samples/sec	accuracy=0.896455	lr=0.012726
Epoch[0] Batch [249]	Speed: 5113.567662 samples/sec	accuracy=0.896354	lr=0.015923
Epoch[0] Batch [299]	Speed: 4951.878325 samples/sec	accuracy=0.889211	lr=0.019121
Epoch[0] Batch [349]	Speed: 4911.784026 samples/sec	accuracy=0.878012	lr=0.022318
Epoch[0] Batch [399]	Speed: 4648.891597 samples/sec	accuracy=0.864383	lr=0.025516
Epoch[0] Batch [449]	Speed: 4937.645848 samples/sec	accuracy=0.849552	lr=0.028713
Epoch[0] Batch [499]	Speed: 4982.542056 samples/sec	accuracy=0.833978	lr=0.031910
Epoch[0] Batch [549]	Speed: 4949.509201 samples/sec	accuracy=0.818352	lr=0.035108
Epoch[0] Batch [599]	Speed: 4887.187154 samples/sec	accuracy=0.803036	lr=0.038305
Epoch[0] Batch [649]	Speed: 5005.541212 samples/sec	accuracy=0.788242	lr=0.041503
Epoch[0] Batch [699]	Speed: 5000.768571 samples/sec	accuracy=0.774153	lr=0.044700
Epoch[0] Batch [749]	Speed: 4979.172520 samples/sec	accuracy=0.760985	lr=0.047898
Epoch[0] Batch [799]	Speed: 5118.229616 samples/sec	accuracy=0.748510	lr=0.051095
Epoch[0] Batch [849]	Speed: 4906.105110 samples/sec	accuracy=0.736818	lr=0.054293
Epoch[0] Batch [899]	Speed: 4802.735122 samples/sec	accuracy=0.725796	lr=0.057490
Epoch[0] Batch [949]	Speed: 5101.057767 samples/sec	accuracy=0.715446	lr=0.060687
Epoch[0] Batch [999]	Speed: 4910.781001 samples/sec	accuracy=0.705690	lr=0.063885
Epoch[0] Batch [1049]	Speed: 4765.436226 samples/sec	accuracy=0.696623	lr=0.067082
Epoch[0] Batch [1099]	Speed: 5104.154174 samples/sec	accuracy=0.687853	lr=0.070280
Epoch[0] Batch [1149]	Speed: 4910.432117 samples/sec	accuracy=0.679514	lr=0.073477
Epoch[0] Batch [1199]	Speed: 4693.563891 samples/sec	accuracy=0.671726	lr=0.076675
Epoch[0] Batch [1249]	Speed: 4960.267889 samples/sec	accuracy=0.664324	lr=0.079872
[Epoch 0] training: accuracy=0.664033
[Epoch 0] speed: 4456 samples/sec	time cost: 297.575884
[Epoch 0] validation: err-top1=0.908084 err-top5=0.771923
Epoch[1] Batch [49]	Speed: 4685.482080 samples/sec	accuracy=0.476718	lr=0.083133
Epoch[1] Batch [99]	Speed: 4947.710305 samples/sec	accuracy=0.476047	lr=0.086331
Epoch[1] Batch [149]	Speed: 5048.477740 samples/sec	accuracy=0.474199	lr=0.089528
Epoch[1] Batch [199]	Speed: 4949.020319 samples/sec	accuracy=0.471999	lr=0.092726
Epoch[1] Batch [249]	Speed: 4833.492398 samples/sec	accuracy=0.469627	lr=0.095923
Epoch[1] Batch [299]	Speed: 4809.917658 samples/sec	accuracy=0.467486	lr=0.099121
Epoch[1] Batch [349]	Speed: 4876.800295 samples/sec	accuracy=0.465968	lr=0.102318
Epoch[1] Batch [399]	Speed: 5003.941080 samples/sec	accuracy=0.464372	lr=0.105516
Epoch[1] Batch [449]	Speed: 4952.761134 samples/sec	accuracy=0.462707	lr=0.108713
Epoch[1] Batch [499]	Speed: 4867.810010 samples/sec	accuracy=0.460998	lr=0.111910
Epoch[1] Batch [549]	Speed: 4858.068877 samples/sec	accuracy=0.459309	lr=0.115108
Epoch[1] Batch [599]	Speed: 4996.768649 samples/sec	accuracy=0.457688	lr=0.118305
Epoch[1] Batch [649]	Speed: 4988.225487 samples/sec	accuracy=0.456351	lr=0.121503
Epoch[1] Batch [699]	Speed: 4937.132633 samples/sec	accuracy=0.455023	lr=0.124700
Epoch[1] Batch [749]	Speed: 4887.992418 samples/sec	accuracy=0.453743	lr=0.127898
Epoch[1] Batch [799]	Speed: 4978.675799 samples/sec	accuracy=0.452490	lr=0.131095
Epoch[1] Batch [849]	Speed: 4969.869283 samples/sec	accuracy=0.451351	lr=0.134293
Epoch[1] Batch [899]	Speed: 4833.794856 samples/sec	accuracy=0.450374	lr=0.137490
Epoch[1] Batch [949]	Speed: 5065.311284 samples/sec	accuracy=0.449421	lr=0.140687
Epoch[1] Batch [999]	Speed: 4993.584324 samples/sec	accuracy=0.448633	lr=0.143885
Epoch[1] Batch [1049]	Speed: 4870.993335 samples/sec	accuracy=0.447863	lr=0.147082
Epoch[1] Batch [1099]	Speed: 4980.525355 samples/sec	accuracy=0.447213	lr=0.150280
Epoch[1] Batch [1149]	Speed: 4873.732262 samples/sec	accuracy=0.446659	lr=0.153477
Epoch[1] Batch [1199]	Speed: 5111.926204 samples/sec	accuracy=0.446159	lr=0.156675
Epoch[1] Batch [1249]	Speed: 5068.313355 samples/sec	accuracy=0.445700	lr=0.159872
[Epoch 1] training: accuracy=0.445690
[Epoch 1] speed: 4929 samples/sec	time cost: 269.316038
[Epoch 1] validation: err-top1=0.835021 err-top5=0.642199
Epoch[2] Batch [49]	Speed: 4700.759824 samples/sec	accuracy=0.436194	lr=0.163133
Epoch[2] Batch [99]	Speed: 4844.295156 samples/sec	accuracy=0.437728	lr=0.166331
Epoch[2] Batch [149]	Speed: 4932.903385 samples/sec	accuracy=0.438139	lr=0.169528
Epoch[2] Batch [199]	Speed: 4936.780675 samples/sec	accuracy=0.438350	lr=0.172726
Epoch[2] Batch [249]	Speed: 4895.483421 samples/sec	accuracy=0.438799	lr=0.175923
Epoch[2] Batch [299]	Speed: 4999.237358 samples/sec	accuracy=0.439521	lr=0.179121
Epoch[2] Batch [349]	Speed: 4861.484219 samples/sec	accuracy=0.440206	lr=0.182318
Epoch[2] Batch [399]	Speed: 4945.102150 samples/sec	accuracy=0.440949	lr=0.185516
Epoch[2] Batch [449]	Speed: 4811.418188 samples/sec	accuracy=0.441860	lr=0.188713
Epoch[2] Batch [499]	Speed: 5005.334826 samples/sec	accuracy=0.442738	lr=0.191910
Epoch[2] Batch [549]	Speed: 5110.326782 samples/sec	accuracy=0.443624	lr=0.195108
Epoch[2] Batch [599]	Speed: 4867.878422 samples/sec	accuracy=0.444376	lr=0.198305
Epoch[2] Batch [649]	Speed: 4821.950882 samples/sec	accuracy=0.445148	lr=0.201503
Epoch[2] Batch [699]	Speed: 4977.196267 samples/sec	accuracy=0.445939	lr=0.204700
Epoch[2] Batch [749]	Speed: 5172.385413 samples/sec	accuracy=0.446876	lr=0.207898
Epoch[2] Batch [799]	Speed: 5157.172353 samples/sec	accuracy=0.447742	lr=0.211095
Epoch[2] Batch [849]	Speed: 4846.384789 samples/sec	accuracy=0.448642	lr=0.214293
Epoch[2] Batch [899]	Speed: 5106.554445 samples/sec	accuracy=0.449534	lr=0.217490
Epoch[2] Batch [949]	Speed: 5275.680021 samples/sec	accuracy=0.450361	lr=0.220687
Epoch[2] Batch [999]	Speed: 5010.221041 samples/sec	accuracy=0.451262	lr=0.223885
Epoch[2] Batch [1049]	Speed: 5093.217253 samples/sec	accuracy=0.452110	lr=0.227082
Epoch[2] Batch [1099]	Speed: 5049.416703 samples/sec	accuracy=0.453041	lr=0.230280
Epoch[2] Batch [1149]	Speed: 5314.906718 samples/sec	accuracy=0.453974	lr=0.233477
Epoch[2] Batch [1199]	Speed: 5344.241712 samples/sec	accuracy=0.454866	lr=0.236675
Epoch[2] Batch [1249]	Speed: 4991.101705 samples/sec	accuracy=0.455843	lr=0.239872
[Epoch 2] training: accuracy=0.455860
[Epoch 2] speed: 4989 samples/sec	time cost: 266.252412
[Epoch 2] validation: err-top1=0.689513 err-top5=0.438437
Epoch[3] Batch [49]	Speed: 4844.413506 samples/sec	accuracy=0.478300	lr=0.243133
Epoch[3] Batch [99]	Speed: 4795.830180 samples/sec	accuracy=0.480380	lr=0.246331
Epoch[3] Batch [149]	Speed: 4850.877246 samples/sec	accuracy=0.480756	lr=0.249528
Epoch[3] Batch [199]	Speed: 4861.314851 samples/sec	accuracy=0.481786	lr=0.252726
Epoch[3] Batch [249]	Speed: 4941.353790 samples/sec	accuracy=0.482342	lr=0.255923
Epoch[3] Batch [299]	Speed: 4843.791984 samples/sec	accuracy=0.483496	lr=0.259121
Epoch[3] Batch [349]	Speed: 4919.213412 samples/sec	accuracy=0.484403	lr=0.262318
Epoch[3] Batch [399]	Speed: 4950.506540 samples/sec	accuracy=0.485189	lr=0.265516
Epoch[3] Batch [449]	Speed: 4941.928499 samples/sec	accuracy=0.486060	lr=0.268713
Epoch[3] Batch [499]	Speed: 5082.522927 samples/sec	accuracy=0.487088	lr=0.271910
Epoch[3] Batch [549]	Speed: 4761.848546 samples/sec	accuracy=0.488113	lr=0.275108
Epoch[3] Batch [599]	Speed: 4840.659160 samples/sec	accuracy=0.489028	lr=0.278305
Epoch[3] Batch [649]	Speed: 4916.020562 samples/sec	accuracy=0.489891	lr=0.281503
Epoch[3] Batch [699]	Speed: 4970.643119 samples/sec	accuracy=0.490733	lr=0.284700
Epoch[3] Batch [749]	Speed: 4977.042848 samples/sec	accuracy=0.491599	lr=0.287898
Epoch[3] Batch [799]	Speed: 5092.378943 samples/sec	accuracy=0.492502	lr=0.291095
Epoch[3] Batch [849]	Speed: 4830.568348 samples/sec	accuracy=0.493469	lr=0.294293
Epoch[3] Batch [899]	Speed: 5066.966095 samples/sec	accuracy=0.494283	lr=0.297490
Epoch[3] Batch [949]	Speed: 5064.111904 samples/sec	accuracy=0.495142	lr=0.300687
Epoch[3] Batch [999]	Speed: 4948.843315 samples/sec	accuracy=0.496005	lr=0.303885
Epoch[3] Batch [1049]	Speed: 5285.710775 samples/sec	accuracy=0.496957	lr=0.307082
Epoch[3] Batch [1099]	Speed: 5371.288681 samples/sec	accuracy=0.497837	lr=0.310280
Epoch[3] Batch [1149]	Speed: 5043.344199 samples/sec	accuracy=0.498711	lr=0.313477
Epoch[3] Batch [1199]	Speed: 4891.371432 samples/sec	accuracy=0.499488	lr=0.316675
Epoch[3] Batch [1249]	Speed: 4790.685489 samples/sec	accuracy=0.500322	lr=0.319872
[Epoch 3] training: accuracy=0.500336
[Epoch 3] speed: 4943 samples/sec	time cost: 268.376661
[Epoch 3] validation: err-top1=0.597138 err-top5=0.336037
Epoch[4] Batch [49]	Speed: 4779.049768 samples/sec	accuracy=0.519956	lr=0.323133
Epoch[4] Batch [99]	Speed: 4870.912682 samples/sec	accuracy=0.521302	lr=0.326331
Epoch[4] Batch [149]	Speed: 4908.624371 samples/sec	accuracy=0.522518	lr=0.329528
Epoch[4] Batch [199]	Speed: 5010.415673 samples/sec	accuracy=0.523165	lr=0.332726
Epoch[4] Batch [249]	Speed: 5011.603550 samples/sec	accuracy=0.524376	lr=0.335923
Epoch[4] Batch [299]	Speed: 4971.099228 samples/sec	accuracy=0.525157	lr=0.339121
Epoch[4] Batch [349]	Speed: 4866.463559 samples/sec	accuracy=0.526025	lr=0.342318
Epoch[4] Batch [399]	Speed: 4867.011712 samples/sec	accuracy=0.526783	lr=0.345516
Epoch[4] Batch [449]	Speed: 4956.458188 samples/sec	accuracy=0.527544	lr=0.348713
Epoch[4] Batch [499]	Speed: 4962.150572 samples/sec	accuracy=0.528219	lr=0.351910
Epoch[4] Batch [549]	Speed: 5085.270628 samples/sec	accuracy=0.529080	lr=0.355108
Epoch[4] Batch [599]	Speed: 4769.310732 samples/sec	accuracy=0.529776	lr=0.358305
Epoch[4] Batch [649]	Speed: 4900.850192 samples/sec	accuracy=0.530588	lr=0.361503
Epoch[4] Batch [699]	Speed: 4930.347261 samples/sec	accuracy=0.531287	lr=0.364700
Epoch[4] Batch [749]	Speed: 4888.735401 samples/sec	accuracy=0.532173	lr=0.367898
Epoch[4] Batch [799]	Speed: 4948.099509 samples/sec	accuracy=0.532857	lr=0.371095
Epoch[4] Batch [849]	Speed: 5113.337782 samples/sec	accuracy=0.533570	lr=0.374293
Epoch[4] Batch [899]	Speed: 4830.229247 samples/sec	accuracy=0.534363	lr=0.377490
Epoch[4] Batch [949]	Speed: 4959.886850 samples/sec	accuracy=0.535087	lr=0.380687
Epoch[4] Batch [999]	Speed: 4783.623315 samples/sec	accuracy=0.535796	lr=0.383885
Epoch[4] Batch [1049]	Speed: 4953.333244 samples/sec	accuracy=0.536604	lr=0.387082
Epoch[4] Batch [1099]	Speed: 4929.194306 samples/sec	accuracy=0.537333	lr=0.390280
Epoch[4] Batch [1149]	Speed: 4874.391586 samples/sec	accuracy=0.538145	lr=0.393477
Epoch[4] Batch [1199]	Speed: 4829.470705 samples/sec	accuracy=0.538935	lr=0.396675
Epoch[4] Batch [1249]	Speed: 5098.562548 samples/sec	accuracy=0.539657	lr=0.399872
[Epoch 4] training: accuracy=0.539676
[Epoch 4] speed: 4918 samples/sec	time cost: 269.898887
[Epoch 4] validation: err-top1=0.553990 err-top5=0.295400
Epoch[5] Batch [49]	Speed: 4783.828873 samples/sec	accuracy=0.557830	lr=0.400000
Epoch[5] Batch [99]	Speed: 4816.651174 samples/sec	accuracy=0.558994	lr=0.400000
Epoch[5] Batch [149]	Speed: 5007.649924 samples/sec	accuracy=0.559384	lr=0.399999
Epoch[5] Batch [199]	Speed: 4853.257173 samples/sec	accuracy=0.559917	lr=0.399998
Epoch[5] Batch [249]	Speed: 4699.189921 samples/sec	accuracy=0.560646	lr=0.399997
Epoch[5] Batch [299]	Speed: 4764.201295 samples/sec	accuracy=0.561455	lr=0.399996
Epoch[5] Batch [349]	Speed: 4907.266576 samples/sec	accuracy=0.562302	lr=0.399994
Epoch[5] Batch [399]	Speed: 4868.212678 samples/sec	accuracy=0.563265	lr=0.399992
Epoch[5] Batch [449]	Speed: 4946.493728 samples/sec	accuracy=0.563953	lr=0.399990
Epoch[5] Batch [499]	Speed: 4805.930566 samples/sec	accuracy=0.564859	lr=0.399988
Epoch[5] Batch [549]	Speed: 4800.223027 samples/sec	accuracy=0.565469	lr=0.399986
Epoch[5] Batch [599]	Speed: 4913.825260 samples/sec	accuracy=0.566277	lr=0.399983
Epoch[5] Batch [649]	Speed: 4913.797375 samples/sec	accuracy=0.567144	lr=0.399980
Epoch[5] Batch [699]	Speed: 5059.591004 samples/sec	accuracy=0.567966	lr=0.399977
Epoch[5] Batch [749]	Speed: 5079.932484 samples/sec	accuracy=0.568586	lr=0.399973
Epoch[5] Batch [799]	Speed: 4740.977295 samples/sec	accuracy=0.569314	lr=0.399970
Epoch[5] Batch [849]	Speed: 5031.436109 samples/sec	accuracy=0.570110	lr=0.399966
Epoch[5] Batch [899]	Speed: 5077.743020 samples/sec	accuracy=0.570884	lr=0.399961
Epoch[5] Batch [949]	Speed: 5179.427212 samples/sec	accuracy=0.571640	lr=0.399957
Epoch[5] Batch [999]	Speed: 4983.606763 samples/sec	accuracy=0.572367	lr=0.399952
Epoch[5] Batch [1049]	Speed: 5224.051096 samples/sec	accuracy=0.573098	lr=0.399948
Epoch[5] Batch [1099]	Speed: 5053.580357 samples/sec	accuracy=0.573819	lr=0.399942
Epoch[5] Batch [1149]	Speed: 5344.467018 samples/sec	accuracy=0.574547	lr=0.399937
Epoch[5] Batch [1199]	Speed: 5055.619651 samples/sec	accuracy=0.575195	lr=0.399931
Epoch[5] Batch [1249]	Speed: 4953.115717 samples/sec	accuracy=0.575910	lr=0.399926
[Epoch 5] training: accuracy=0.575918
[Epoch 5] speed: 4943 samples/sec	time cost: 268.378407
[Epoch 5] validation: err-top1=0.526367 err-top5=0.262838
Epoch[6] Batch [49]	Speed: 4546.972709 samples/sec	accuracy=0.589669	lr=0.399919
Epoch[6] Batch [99]	Speed: 4987.673091 samples/sec	accuracy=0.591648	lr=0.399913
Epoch[6] Batch [149]	Speed: 4830.270749 samples/sec	accuracy=0.592467	lr=0.399907
Epoch[6] Batch [199]	Speed: 4959.098151 samples/sec	accuracy=0.592926	lr=0.399900
Epoch[6] Batch [249]	Speed: 5031.469234 samples/sec	accuracy=0.593709	lr=0.399893
Epoch[6] Batch [299]	Speed: 5190.612459 samples/sec	accuracy=0.593967	lr=0.399885
Epoch[6] Batch [349]	Speed: 4828.508722 samples/sec	accuracy=0.594640	lr=0.399878
Epoch[6] Batch [399]	Speed: 5193.824982 samples/sec	accuracy=0.595106	lr=0.399870
Epoch[6] Batch [449]	Speed: 4723.076801 samples/sec	accuracy=0.595764	lr=0.399862
Epoch[6] Batch [499]	Speed: 5049.082150 samples/sec	accuracy=0.596242	lr=0.399854
Epoch[6] Batch [549]	Speed: 4771.447158 samples/sec	accuracy=0.596780	lr=0.399846
Epoch[6] Batch [599]	Speed: 4992.898749 samples/sec	accuracy=0.597357	lr=0.399837
Epoch[6] Batch [649]	Speed: 4889.575353 samples/sec	accuracy=0.597930	lr=0.399828
Epoch[6] Batch [699]	Speed: 5028.497922 samples/sec	accuracy=0.598657	lr=0.399819
Epoch[6] Batch [749]	Speed: 5155.184596 samples/sec	accuracy=0.599235	lr=0.399809
Epoch[6] Batch [799]	Speed: 4855.454447 samples/sec	accuracy=0.599744	lr=0.399800
Epoch[6] Batch [849]	Speed: 5237.511423 samples/sec	accuracy=0.600323	lr=0.399790
Epoch[6] Batch [899]	Speed: 4934.613513 samples/sec	accuracy=0.600874	lr=0.399780
Epoch[6] Batch [949]	Speed: 5174.308039 samples/sec	accuracy=0.601494	lr=0.399769
Epoch[6] Batch [999]	Speed: 5144.523395 samples/sec	accuracy=0.601991	lr=0.399759
Epoch[6] Batch [1049]	Speed: 5197.612802 samples/sec	accuracy=0.602469	lr=0.399748
Epoch[6] Batch [1099]	Speed: 5042.987476 samples/sec	accuracy=0.602988	lr=0.399737
Epoch[6] Batch [1149]	Speed: 4931.310734 samples/sec	accuracy=0.603588	lr=0.399725
Epoch[6] Batch [1199]	Speed: 4917.452119 samples/sec	accuracy=0.604083	lr=0.399714
Epoch[6] Batch [1249]	Speed: 4949.168479 samples/sec	accuracy=0.604587	lr=0.399702
[Epoch 6] training: accuracy=0.604596
[Epoch 6] speed: 4972 samples/sec	time cost: 267.498487
[Epoch 6] validation: err-top1=0.488481 err-top5=0.235192
Epoch[7] Batch [49]	Speed: 4813.005735 samples/sec	accuracy=0.614122	lr=0.399690
Epoch[7] Batch [99]	Speed: 4922.005836 samples/sec	accuracy=0.615481	lr=0.399677
Epoch[7] Batch [149]	Speed: 4980.917774 samples/sec	accuracy=0.616406	lr=0.399665
Epoch[7] Batch [199]	Speed: 5005.908761 samples/sec	accuracy=0.616730	lr=0.399652
Epoch[7] Batch [249]	Speed: 4744.558845 samples/sec	accuracy=0.616892	lr=0.399639
Epoch[7] Batch [299]	Speed: 5034.056143 samples/sec	accuracy=0.617334	lr=0.399626
Epoch[7] Batch [349]	Speed: 4965.976726 samples/sec	accuracy=0.617918	lr=0.399613
Epoch[7] Batch [399]	Speed: 4834.974039 samples/sec	accuracy=0.618295	lr=0.399599
Epoch[7] Batch [449]	Speed: 5064.993257 samples/sec	accuracy=0.618758	lr=0.399585
Epoch[7] Batch [499]	Speed: 5095.155567 samples/sec	accuracy=0.619095	lr=0.399571
Epoch[7] Batch [549]	Speed: 4991.004962 samples/sec	accuracy=0.619430	lr=0.399556
Epoch[7] Batch [599]	Speed: 5102.440434 samples/sec	accuracy=0.619901	lr=0.399542
Epoch[7] Batch [649]	Speed: 4994.395532 samples/sec	accuracy=0.620397	lr=0.399527
Epoch[7] Batch [699]	Speed: 4861.251685 samples/sec	accuracy=0.620828	lr=0.399512
Epoch[7] Batch [749]	Speed: 4920.094644 samples/sec	accuracy=0.621119	lr=0.399496
Epoch[7] Batch [799]	Speed: 4976.568349 samples/sec	accuracy=0.621625	lr=0.399481
Epoch[7] Batch [849]	Speed: 4927.158939 samples/sec	accuracy=0.622108	lr=0.399465
Epoch[7] Batch [899]	Speed: 4996.603326 samples/sec	accuracy=0.622541	lr=0.399449
Epoch[7] Batch [949]	Speed: 4887.520618 samples/sec	accuracy=0.622926	lr=0.399432
Epoch[7] Batch [999]	Speed: 4854.073456 samples/sec	accuracy=0.623288	lr=0.399416
Epoch[7] Batch [1049]	Speed: 4988.973177 samples/sec	accuracy=0.623758	lr=0.399399
Epoch[7] Batch [1099]	Speed: 4838.926182 samples/sec	accuracy=0.624100	lr=0.399382
Epoch[7] Batch [1149]	Speed: 5004.019320 samples/sec	accuracy=0.624524	lr=0.399365
Epoch[7] Batch [1199]	Speed: 5071.943852 samples/sec	accuracy=0.624883	lr=0.399347
Epoch[7] Batch [1249]	Speed: 5128.722354 samples/sec	accuracy=0.625217	lr=0.399329
[Epoch 7] training: accuracy=0.625230
[Epoch 7] speed: 4954 samples/sec	time cost: 268.905468
[Epoch 7] validation: err-top1=0.470045 err-top5=0.218371
Epoch[8] Batch [49]	Speed: 4893.050539 samples/sec	accuracy=0.630379	lr=0.399311
Epoch[8] Batch [99]	Speed: 4700.835866 samples/sec	accuracy=0.631426	lr=0.399293
Epoch[8] Batch [149]	Speed: 4851.104734 samples/sec	accuracy=0.630940	lr=0.399274
Epoch[8] Batch [199]	Speed: 4836.122541 samples/sec	accuracy=0.632249	lr=0.399256
Epoch[8] Batch [249]	Speed: 4906.568733 samples/sec	accuracy=0.632698	lr=0.399237
Epoch[8] Batch [299]	Speed: 5013.552800 samples/sec	accuracy=0.633156	lr=0.399218
Epoch[8] Batch [349]	Speed: 5045.200401 samples/sec	accuracy=0.633768	lr=0.399198
Epoch[8] Batch [399]	Speed: 4747.792045 samples/sec	accuracy=0.634267	lr=0.399179
Epoch[8] Batch [449]	Speed: 5107.805116 samples/sec	accuracy=0.634763	lr=0.399159
Epoch[8] Batch [499]	Speed: 5017.879086 samples/sec	accuracy=0.635173	lr=0.399138
Epoch[8] Batch [549]	Speed: 4997.073980 samples/sec	accuracy=0.635517	lr=0.399118
Epoch[8] Batch [599]	Speed: 5103.467376 samples/sec	accuracy=0.635824	lr=0.399098
Epoch[8] Batch [649]	Speed: 4902.989248 samples/sec	accuracy=0.636118	lr=0.399077
Epoch[8] Batch [699]	Speed: 5163.632735 samples/sec	accuracy=0.636526	lr=0.399056
Epoch[8] Batch [749]	Speed: 4963.358459 samples/sec	accuracy=0.636915	lr=0.399034
Epoch[8] Batch [799]	Speed: 5025.949892 samples/sec	accuracy=0.637423	lr=0.399013
Epoch[8] Batch [849]	Speed: 5124.507505 samples/sec	accuracy=0.637733	lr=0.398991
Epoch[8] Batch [899]	Speed: 5250.187836 samples/sec	accuracy=0.638098	lr=0.398969
Epoch[8] Batch [949]	Speed: 4960.172338 samples/sec	accuracy=0.638473	lr=0.398947
Epoch[8] Batch [999]	Speed: 4940.278191 samples/sec	accuracy=0.638774	lr=0.398924
Epoch[8] Batch [1049]	Speed: 5090.972756 samples/sec	accuracy=0.639076	lr=0.398901
Epoch[8] Batch [1099]	Speed: 4908.553687 samples/sec	accuracy=0.639380	lr=0.398878
Epoch[8] Batch [1149]	Speed: 4916.311826 samples/sec	accuracy=0.639722	lr=0.398855
Epoch[8] Batch [1199]	Speed: 4921.914121 samples/sec	accuracy=0.640140	lr=0.398832
Epoch[8] Batch [1249]	Speed: 4993.029812 samples/sec	accuracy=0.640421	lr=0.398808
[Epoch 8] training: accuracy=0.640428
[Epoch 8] speed: 4966 samples/sec	time cost: 267.258119
[Epoch 8] validation: err-top1=0.488361 err-top5=0.235132
Epoch[9] Batch [49]	Speed: 4749.305315 samples/sec	accuracy=0.647527	lr=0.398784
Epoch[9] Batch [99]	Speed: 5018.766938 samples/sec	accuracy=0.647013	lr=0.398760
Epoch[9] Batch [149]	Speed: 4857.695576 samples/sec	accuracy=0.646777	lr=0.398735
Epoch[9] Batch [199]	Speed: 5045.080097 samples/sec	accuracy=0.647149	lr=0.398710
Epoch[9] Batch [249]	Speed: 4797.009663 samples/sec	accuracy=0.647320	lr=0.398686
Epoch[9] Batch [299]	Speed: 5066.948521 samples/sec	accuracy=0.647559	lr=0.398660
Epoch[9] Batch [349]	Speed: 5112.269137 samples/sec	accuracy=0.647695	lr=0.398635
Epoch[9] Batch [399]	Speed: 4853.186867 samples/sec	accuracy=0.648002	lr=0.398610
Epoch[9] Batch [449]	Speed: 4840.819780 samples/sec	accuracy=0.648315	lr=0.398584
Epoch[9] Batch [499]	Speed: 5225.916313 samples/sec	accuracy=0.648361	lr=0.398558
Epoch[9] Batch [549]	Speed: 5026.363855 samples/sec	accuracy=0.648774	lr=0.398531
Epoch[9] Batch [599]	Speed: 4928.241160 samples/sec	accuracy=0.648983	lr=0.398505
Epoch[9] Batch [649]	Speed: 5111.446929 samples/sec	accuracy=0.649249	lr=0.398478
Epoch[9] Batch [699]	Speed: 5027.761174 samples/sec	accuracy=0.649362	lr=0.398451
Epoch[9] Batch [749]	Speed: 4852.761350 samples/sec	accuracy=0.649685	lr=0.398424
Epoch[9] Batch [799]	Speed: 4976.819889 samples/sec	accuracy=0.649888	lr=0.398396
Epoch[9] Batch [849]	Speed: 5045.133315 samples/sec	accuracy=0.650009	lr=0.398369
Epoch[9] Batch [899]	Speed: 4986.122680 samples/sec	accuracy=0.650180	lr=0.398341
Epoch[9] Batch [949]	Speed: 5032.551418 samples/sec	accuracy=0.650445	lr=0.398312
Epoch[9] Batch [999]	Speed: 4807.469829 samples/sec	accuracy=0.650743	lr=0.398284
Epoch[9] Batch [1049]	Speed: 5000.493645 samples/sec	accuracy=0.650981	lr=0.398255
Epoch[9] Batch [1099]	Speed: 5057.842130 samples/sec	accuracy=0.651255	lr=0.398226
Epoch[9] Batch [1149]	Speed: 4998.946542 samples/sec	accuracy=0.651536	lr=0.398197
Epoch[9] Batch [1199]	Speed: 4906.075968 samples/sec	accuracy=0.651762	lr=0.398168
Epoch[9] Batch [1249]	Speed: 5163.204668 samples/sec	accuracy=0.651964	lr=0.398138
[Epoch 9] training: accuracy=0.651972
[Epoch 9] speed: 4972 samples/sec	time cost: 267.078502
[Epoch 9] validation: err-top1=0.432139 err-top5=0.189473
Epoch[10] Batch [49]	Speed: 4658.812037 samples/sec	accuracy=0.654272	lr=0.398108
Epoch[10] Batch [99]	Speed: 4686.294234 samples/sec	accuracy=0.654212	lr=0.398078
Epoch[10] Batch [149]	Speed: 4874.275638 samples/sec	accuracy=0.654860	lr=0.398048
Epoch[10] Batch [199]	Speed: 5076.809097 samples/sec	accuracy=0.654748	lr=0.398017
Epoch[10] Batch [249]	Speed: 5091.153435 samples/sec	accuracy=0.655434	lr=0.397986
Epoch[10] Batch [299]	Speed: 4829.599520 samples/sec	accuracy=0.655680	lr=0.397955
Epoch[10] Batch [349]	Speed: 4808.649123 samples/sec	accuracy=0.656250	lr=0.397924
Epoch[10] Batch [399]	Speed: 4993.837239 samples/sec	accuracy=0.656794	lr=0.397892
Epoch[10] Batch [449]	Speed: 4959.655346 samples/sec	accuracy=0.657243	lr=0.397861
Epoch[10] Batch [499]	Speed: 4848.050112 samples/sec	accuracy=0.657526	lr=0.397829
Epoch[10] Batch [549]	Speed: 4820.430034 samples/sec	accuracy=0.657789	lr=0.397796
Epoch[10] Batch [599]	Speed: 4893.306865 samples/sec	accuracy=0.658163	lr=0.397764
Epoch[10] Batch [649]	Speed: 4898.782732 samples/sec	accuracy=0.658362	lr=0.397731
Epoch[10] Batch [699]	Speed: 4943.316015 samples/sec	accuracy=0.658697	lr=0.397698
Epoch[10] Batch [749]	Speed: 4900.314964 samples/sec	accuracy=0.659001	lr=0.397665
Epoch[10] Batch [799]	Speed: 4964.637635 samples/sec	accuracy=0.659204	lr=0.397632
Epoch[10] Batch [849]	Speed: 4998.441331 samples/sec	accuracy=0.659413	lr=0.397598
Epoch[10] Batch [899]	Speed: 4840.599148 samples/sec	accuracy=0.659637	lr=0.397564
Epoch[10] Batch [949]	Speed: 5011.464259 samples/sec	accuracy=0.659791	lr=0.397530
Epoch[10] Batch [999]	Speed: 4930.672037 samples/sec	accuracy=0.660082	lr=0.397496
Epoch[10] Batch [1049]	Speed: 4960.566369 samples/sec	accuracy=0.660352	lr=0.397461
Epoch[10] Batch [1099]	Speed: 4983.210335 samples/sec	accuracy=0.660535	lr=0.397427
Epoch[10] Batch [1149]	Speed: 4992.696189 samples/sec	accuracy=0.660801	lr=0.397392
Epoch[10] Batch [1199]	Speed: 5038.313287 samples/sec	accuracy=0.661031	lr=0.397356
Epoch[10] Batch [1249]	Speed: 5056.511504 samples/sec	accuracy=0.661188	lr=0.397321
[Epoch 10] training: accuracy=0.661189
[Epoch 10] speed: 4915 samples/sec	time cost: 270.323582
[Epoch 10] validation: err-top1=0.439812 err-top5=0.196628
Epoch[11] Batch [49]	Speed: 4775.077034 samples/sec	accuracy=0.663181	lr=0.397284
Epoch[11] Batch [99]	Speed: 4723.499619 samples/sec	accuracy=0.663619	lr=0.397248
Epoch[11] Batch [149]	Speed: 4901.612192 samples/sec	accuracy=0.663955	lr=0.397212
Epoch[11] Batch [199]	Speed: 5003.919626 samples/sec	accuracy=0.664064	lr=0.397176
Epoch[11] Batch [249]	Speed: 4911.704937 samples/sec	accuracy=0.663944	lr=0.397139
Epoch[11] Batch [299]	Speed: 5207.435494 samples/sec	accuracy=0.664215	lr=0.397102
Epoch[11] Batch [349]	Speed: 5063.733730 samples/sec	accuracy=0.664557	lr=0.397065
Epoch[11] Batch [399]	Speed: 4977.325930 samples/sec	accuracy=0.664665	lr=0.397028
Epoch[11] Batch [449]	Speed: 4867.792907 samples/sec	accuracy=0.665144	lr=0.396990
Epoch[11] Batch [499]	Speed: 4888.046935 samples/sec	accuracy=0.665389	lr=0.396952
Epoch[11] Batch [549]	Speed: 4916.971351 samples/sec	accuracy=0.665430	lr=0.396914
Epoch[11] Batch [599]	Speed: 4991.694543 samples/sec	accuracy=0.665667	lr=0.396876
Epoch[11] Batch [649]	Speed: 5026.274563 samples/sec	accuracy=0.665962	lr=0.396837
Epoch[11] Batch [699]	Speed: 4933.578816 samples/sec	accuracy=0.666195	lr=0.396798
Epoch[11] Batch [749]	Speed: 5026.163511 samples/sec	accuracy=0.666484	lr=0.396759
Epoch[11] Batch [799]	Speed: 4802.329251 samples/sec	accuracy=0.666676	lr=0.396720
Epoch[11] Batch [849]	Speed: 5003.746484 samples/sec	accuracy=0.666909	lr=0.396680
Epoch[11] Batch [899]	Speed: 5149.385464 samples/sec	accuracy=0.667180	lr=0.396641
Epoch[11] Batch [949]	Speed: 4927.502177 samples/sec	accuracy=0.667350	lr=0.396601
Epoch[11] Batch [999]	Speed: 4986.776751 samples/sec	accuracy=0.667547	lr=0.396561
Epoch[11] Batch [1049]	Speed: 5028.139410 samples/sec	accuracy=0.667729	lr=0.396520
Epoch[11] Batch [1099]	Speed: 5088.908827 samples/sec	accuracy=0.667936	lr=0.396479
Epoch[11] Batch [1149]	Speed: 4687.181453 samples/sec	accuracy=0.668222	lr=0.396439
Epoch[11] Batch [1199]	Speed: 5061.996071 samples/sec	accuracy=0.668469	lr=0.396397
Epoch[11] Batch [1249]	Speed: 5101.381550 samples/sec	accuracy=0.668752	lr=0.396356
[Epoch 11] training: accuracy=0.668756
[Epoch 11] speed: 4954 samples/sec	time cost: 268.391008
[Epoch 11] validation: err-top1=0.452454 err-top5=0.206441
Epoch[12] Batch [49]	Speed: 4636.566920 samples/sec	accuracy=0.672821	lr=0.396314
Epoch[12] Batch [99]	Speed: 4577.400896 samples/sec	accuracy=0.671722	lr=0.396272
Epoch[12] Batch [149]	Speed: 4951.004961 samples/sec	accuracy=0.672443	lr=0.396230
Epoch[12] Batch [199]	Speed: 4814.970212 samples/sec	accuracy=0.671960	lr=0.396187
Epoch[12] Batch [249]	Speed: 4981.175647 samples/sec	accuracy=0.672036	lr=0.396145
Epoch[12] Batch [299]	Speed: 4901.014272 samples/sec	accuracy=0.672294	lr=0.396102
Epoch[12] Batch [349]	Speed: 4900.618014 samples/sec	accuracy=0.672378	lr=0.396059
Epoch[12] Batch [399]	Speed: 5100.564294 samples/sec	accuracy=0.672449	lr=0.396016
Epoch[12] Batch [449]	Speed: 4969.667552 samples/sec	accuracy=0.672726	lr=0.395972
Epoch[12] Batch [499]	Speed: 5026.160688 samples/sec	accuracy=0.672755	lr=0.395928
Epoch[12] Batch [549]	Speed: 5056.341132 samples/sec	accuracy=0.672885	lr=0.395885
Epoch[12] Batch [599]	Speed: 4829.925823 samples/sec	accuracy=0.672939	lr=0.395840
Epoch[12] Batch [649]	Speed: 4899.585338 samples/sec	accuracy=0.673226	lr=0.395796
Epoch[12] Batch [699]	Speed: 5221.421437 samples/sec	accuracy=0.673467	lr=0.395751
Epoch[12] Batch [749]	Speed: 5079.288469 samples/sec	accuracy=0.673598	lr=0.395706
Epoch[12] Batch [799]	Speed: 5097.274294 samples/sec	accuracy=0.673824	lr=0.395661
Epoch[12] Batch [849]	Speed: 4868.716300 samples/sec	accuracy=0.674022	lr=0.395616
Epoch[12] Batch [899]	Speed: 4898.008540 samples/sec	accuracy=0.674213	lr=0.395570
Epoch[12] Batch [949]	Speed: 4972.113922 samples/sec	accuracy=0.674388	lr=0.395524
Epoch[12] Batch [999]	Speed: 4989.233623 samples/sec	accuracy=0.674568	lr=0.395478
Epoch[12] Batch [1049]	Speed: 4933.041176 samples/sec	accuracy=0.674754	lr=0.395432
Epoch[12] Batch [1099]	Speed: 4945.907933 samples/sec	accuracy=0.674836	lr=0.395386
Epoch[12] Batch [1149]	Speed: 4949.990078 samples/sec	accuracy=0.674984	lr=0.395339
Epoch[12] Batch [1199]	Speed: 4824.574437 samples/sec	accuracy=0.675142	lr=0.395292
Epoch[12] Batch [1249]	Speed: 5221.509798 samples/sec	accuracy=0.675300	lr=0.395245
[Epoch 12] training: accuracy=0.675299
[Epoch 12] speed: 4937 samples/sec	time cost: 269.590493
[Epoch 12] validation: err-top1=0.426678 err-top5=0.183713
Epoch[13] Batch [49]	Speed: 4529.326039 samples/sec	accuracy=0.676229	lr=0.395196
Epoch[13] Batch [99]	Speed: 4616.228651 samples/sec	accuracy=0.676608	lr=0.395149
Epoch[13] Batch [149]	Speed: 4819.878691 samples/sec	accuracy=0.676726	lr=0.395101
Epoch[13] Batch [199]	Speed: 4973.517982 samples/sec	accuracy=0.677249	lr=0.395052
Epoch[13] Batch [249]	Speed: 5065.598521 samples/sec	accuracy=0.677402	lr=0.395004
Epoch[13] Batch [299]	Speed: 5171.200047 samples/sec	accuracy=0.677408	lr=0.394955
Epoch[13] Batch [349]	Speed: 4936.268208 samples/sec	accuracy=0.677321	lr=0.394907
Epoch[13] Batch [399]	Speed: 5148.131999 samples/sec	accuracy=0.677452	lr=0.394858
Epoch[13] Batch [449]	Speed: 4916.801809 samples/sec	accuracy=0.677654	lr=0.394808
Epoch[13] Batch [499]	Speed: 4999.700127 samples/sec	accuracy=0.677909	lr=0.394759
Epoch[13] Batch [549]	Speed: 4947.492816 samples/sec	accuracy=0.677862	lr=0.394709
Epoch[13] Batch [599]	Speed: 5042.661708 samples/sec	accuracy=0.678045	lr=0.394659
Epoch[13] Batch [649]	Speed: 4900.400396 samples/sec	accuracy=0.678177	lr=0.394609
Epoch[13] Batch [699]	Speed: 4851.842573 samples/sec	accuracy=0.678277	lr=0.394558
Epoch[13] Batch [749]	Speed: 4934.567364 samples/sec	accuracy=0.678396	lr=0.394507
Epoch[13] Batch [799]	Speed: 5060.967860 samples/sec	accuracy=0.678584	lr=0.394457
Epoch[13] Batch [849]	Speed: 4897.582610 samples/sec	accuracy=0.678772	lr=0.394405
Epoch[13] Batch [899]	Speed: 4939.448792 samples/sec	accuracy=0.678982	lr=0.394354
Epoch[13] Batch [949]	Speed: 4762.408343 samples/sec	accuracy=0.679166	lr=0.394302
Epoch[13] Batch [999]	Speed: 4908.896357 samples/sec	accuracy=0.679360	lr=0.394250
Epoch[13] Batch [1049]	Speed: 5096.395581 samples/sec	accuracy=0.679495	lr=0.394198
Epoch[13] Batch [1099]	Speed: 4809.354177 samples/sec	accuracy=0.679619	lr=0.394146
Epoch[13] Batch [1149]	Speed: 4960.043566 samples/sec	accuracy=0.679850	lr=0.394093
Epoch[13] Batch [1199]	Speed: 4841.603068 samples/sec	accuracy=0.679990	lr=0.394041
Epoch[13] Batch [1249]	Speed: 5091.018860 samples/sec	accuracy=0.680116	lr=0.393988
[Epoch 13] training: accuracy=0.680114
[Epoch 13] speed: 4920 samples/sec	time cost: 270.085217
[Epoch 13] validation: err-top1=0.449059 err-top5=0.200534
Epoch[14] Batch [49]	Speed: 4673.325424 samples/sec	accuracy=0.679799	lr=0.393933
Epoch[14] Batch [99]	Speed: 4742.252052 samples/sec	accuracy=0.681036	lr=0.393880
Epoch[14] Batch [149]	Speed: 5033.294525 samples/sec	accuracy=0.681212	lr=0.393826
Epoch[14] Batch [199]	Speed: 4897.640692 samples/sec	accuracy=0.680961	lr=0.393772
Epoch[14] Batch [249]	Speed: 5000.145984 samples/sec	accuracy=0.681072	lr=0.393718
Epoch[14] Batch [299]	Speed: 4975.217663 samples/sec	accuracy=0.680888	lr=0.393664
Epoch[14] Batch [349]	Speed: 4890.601920 samples/sec	accuracy=0.680992	lr=0.393609
Epoch[14] Batch [399]	Speed: 4890.954230 samples/sec	accuracy=0.681300	lr=0.393554
Epoch[14] Batch [449]	Speed: 4939.076739 samples/sec	accuracy=0.681397	lr=0.393499
Epoch[14] Batch [499]	Speed: 4987.491341 samples/sec	accuracy=0.681619	lr=0.393444
Epoch[14] Batch [549]	Speed: 4946.754772 samples/sec	accuracy=0.681868	lr=0.393388
Epoch[14] Batch [599]	Speed: 4985.201668 samples/sec	accuracy=0.682002	lr=0.393332
Epoch[14] Batch [649]	Speed: 4889.668761 samples/sec	accuracy=0.682257	lr=0.393276
Epoch[14] Batch [699]	Speed: 4859.769075 samples/sec	accuracy=0.682485	lr=0.393220
Epoch[14] Batch [749]	Speed: 4934.983080 samples/sec	accuracy=0.682786	lr=0.393163
Epoch[14] Batch [799]	Speed: 5074.925010 samples/sec	accuracy=0.682926	lr=0.393107
Epoch[14] Batch [849]	Speed: 4781.662502 samples/sec	accuracy=0.683063	lr=0.393050
Epoch[14] Batch [899]	Speed: 4927.037076 samples/sec	accuracy=0.683254	lr=0.392993
Epoch[14] Batch [949]	Speed: 4885.698465 samples/sec	accuracy=0.683414	lr=0.392935
Epoch[14] Batch [999]	Speed: 5036.646539 samples/sec	accuracy=0.683626	lr=0.392877
Epoch[14] Batch [1049]	Speed: 4963.447480 samples/sec	accuracy=0.683814	lr=0.392820
Epoch[14] Batch [1099]	Speed: 4785.078054 samples/sec	accuracy=0.683990	lr=0.392762
Epoch[14] Batch [1149]	Speed: 5044.181371 samples/sec	accuracy=0.684141	lr=0.392703
Epoch[14] Batch [1199]	Speed: 4854.380909 samples/sec	accuracy=0.684405	lr=0.392645
Epoch[14] Batch [1249]	Speed: 4943.478741 samples/sec	accuracy=0.684604	lr=0.392586
[Epoch 14] training: accuracy=0.684610
[Epoch 14] speed: 4912 samples/sec	time cost: 270.636028
[Epoch 14] validation: err-top1=0.424785 err-top5=0.179149
Epoch[15] Batch [49]	Speed: 4569.059608 samples/sec	accuracy=0.684885	lr=0.392526
Epoch[15] Batch [99]	Speed: 4807.010217 samples/sec	accuracy=0.686007	lr=0.392466
Epoch[15] Batch [149]	Speed: 5051.797724 samples/sec	accuracy=0.686165	lr=0.392407
Epoch[15] Batch [199]	Speed: 5016.151659 samples/sec	accuracy=0.685946	lr=0.392347
Epoch[15] Batch [249]	Speed: 4926.233475 samples/sec	accuracy=0.685987	lr=0.392287
Epoch[15] Batch [299]	Speed: 5019.121181 samples/sec	accuracy=0.685987	lr=0.392227
Epoch[15] Batch [349]	Speed: 5064.863406 samples/sec	accuracy=0.686136	lr=0.392167
Epoch[15] Batch [399]	Speed: 5015.692635 samples/sec	accuracy=0.686416	lr=0.392106
Epoch[15] Batch [449]	Speed: 4760.007353 samples/sec	accuracy=0.686535	lr=0.392045
Epoch[15] Batch [499]	Speed: 4909.940484 samples/sec	accuracy=0.686638	lr=0.391984
Epoch[15] Batch [549]	Speed: 4982.651304 samples/sec	accuracy=0.686835	lr=0.391923
Epoch[15] Batch [599]	Speed: 5058.755741 samples/sec	accuracy=0.686910	lr=0.391861
Epoch[15] Batch [649]	Speed: 4979.515885 samples/sec	accuracy=0.687097	lr=0.391799
Epoch[15] Batch [699]	Speed: 4849.834647 samples/sec	accuracy=0.687194	lr=0.391737
Epoch[15] Batch [749]	Speed: 5018.427052 samples/sec	accuracy=0.687505	lr=0.391675
Epoch[15] Batch [799]	Speed: 4866.830929 samples/sec	accuracy=0.687628	lr=0.391613
Epoch[15] Batch [849]	Speed: 5100.063892 samples/sec	accuracy=0.687753	lr=0.391550
Epoch[15] Batch [899]	Speed: 4934.390484 samples/sec	accuracy=0.687863	lr=0.391487
Epoch[15] Batch [949]	Speed: 5061.877947 samples/sec	accuracy=0.687927	lr=0.391424
Epoch[15] Batch [999]	Speed: 5044.722654 samples/sec	accuracy=0.688047	lr=0.391361
Epoch[15] Batch [1049]	Speed: 5146.678329 samples/sec	accuracy=0.688230	lr=0.391297
Epoch[15] Batch [1099]	Speed: 5118.977743 samples/sec	accuracy=0.688379	lr=0.391233
Epoch[15] Batch [1149]	Speed: 4942.507664 samples/sec	accuracy=0.688511	lr=0.391169
Epoch[15] Batch [1199]	Speed: 5106.684257 samples/sec	accuracy=0.688676	lr=0.391105
Epoch[15] Batch [1249]	Speed: 5170.758274 samples/sec	accuracy=0.688832	lr=0.391040
[Epoch 15] training: accuracy=0.688840
[Epoch 15] speed: 4972 samples/sec	time cost: 267.192504
[Epoch 15] validation: err-top1=0.409997 err-top5=0.171676
Epoch[16] Batch [49]	Speed: 4774.480924 samples/sec	accuracy=0.689831	lr=0.390974
Epoch[16] Batch [99]	Speed: 4928.840311 samples/sec	accuracy=0.689552	lr=0.390909
Epoch[16] Batch [149]	Speed: 5143.037766 samples/sec	accuracy=0.689708	lr=0.390844
Epoch[16] Batch [199]	Speed: 4781.795593 samples/sec	accuracy=0.689311	lr=0.390779
Epoch[16] Batch [249]	Speed: 4750.280862 samples/sec	accuracy=0.689397	lr=0.390713
Epoch[16] Batch [299]	Speed: 4974.293992 samples/sec	accuracy=0.689806	lr=0.390647
Epoch[16] Batch [349]	Speed: 4872.396573 samples/sec	accuracy=0.689877	lr=0.390581
Epoch[16] Batch [399]	Speed: 4851.232514 samples/sec	accuracy=0.689906	lr=0.390515
Epoch[16] Batch [449]	Speed: 4989.824046 samples/sec	accuracy=0.690097	lr=0.390448
Epoch[16] Batch [499]	Speed: 5003.701015 samples/sec	accuracy=0.690290	lr=0.390381
Epoch[16] Batch [549]	Speed: 5178.553915 samples/sec	accuracy=0.690491	lr=0.390314
Epoch[16] Batch [599]	Speed: 4671.562708 samples/sec	accuracy=0.690695	lr=0.390247
Epoch[16] Batch [649]	Speed: 5119.378982 samples/sec	accuracy=0.690863	lr=0.390180
Epoch[16] Batch [699]	Speed: 5127.820888 samples/sec	accuracy=0.690915	lr=0.390112
Epoch[16] Batch [749]	Speed: 5044.902080 samples/sec	accuracy=0.691143	lr=0.390044
Epoch[16] Batch [799]	Speed: 4877.732316 samples/sec	accuracy=0.691377	lr=0.389976
Epoch[16] Batch [849]	Speed: 5114.607614 samples/sec	accuracy=0.691591	lr=0.389907
Epoch[16] Batch [899]	Speed: 5195.676593 samples/sec	accuracy=0.691826	lr=0.389839
Epoch[16] Batch [949]	Speed: 5181.538478 samples/sec	accuracy=0.691955	lr=0.389770
Epoch[16] Batch [999]	Speed: 5186.846482 samples/sec	accuracy=0.692045	lr=0.389701
Epoch[16] Batch [1049]	Speed: 4850.436467 samples/sec	accuracy=0.692196	lr=0.389632
Epoch[16] Batch [1099]	Speed: 4900.707930 samples/sec	accuracy=0.692331	lr=0.389562
Epoch[16] Batch [1149]	Speed: 4833.447142 samples/sec	accuracy=0.692403	lr=0.389492
Epoch[16] Batch [1199]	Speed: 5042.169642 samples/sec	accuracy=0.692568	lr=0.389422
Epoch[16] Batch [1249]	Speed: 4597.387786 samples/sec	accuracy=0.692661	lr=0.389352
[Epoch 16] training: accuracy=0.692660
[Epoch 16] speed: 4945 samples/sec	time cost: 268.397549
[Epoch 16] validation: err-top1=0.406768 err-top5=0.169145
Epoch[17] Batch [49]	Speed: 4607.438129 samples/sec	accuracy=0.693688	lr=0.389280
Epoch[17] Batch [99]	Speed: 4947.544451 samples/sec	accuracy=0.693675	lr=0.389210
Epoch[17] Batch [149]	Speed: 5027.288371 samples/sec	accuracy=0.693804	lr=0.389139
Epoch[17] Batch [199]	Speed: 4909.346816 samples/sec	accuracy=0.694041	lr=0.389068
Epoch[17] Batch [249]	Speed: 5005.288277 samples/sec	accuracy=0.693837	lr=0.388996
Epoch[17] Batch [299]	Speed: 5186.092665 samples/sec	accuracy=0.693861	lr=0.388925
Epoch[17] Batch [349]	Speed: 4791.437775 samples/sec	accuracy=0.694035	lr=0.388853
Epoch[17] Batch [399]	Speed: 4963.759651 samples/sec	accuracy=0.694224	lr=0.388781
Epoch[17] Batch [449]	Speed: 5003.866691 samples/sec	accuracy=0.694377	lr=0.388709
Epoch[17] Batch [499]	Speed: 4842.943441 samples/sec	accuracy=0.694532	lr=0.388636
Epoch[17] Batch [549]	Speed: 5078.636090 samples/sec	accuracy=0.694579	lr=0.388564
Epoch[17] Batch [599]	Speed: 4812.177539 samples/sec	accuracy=0.694866	lr=0.388491
Epoch[17] Batch [649]	Speed: 4894.647016 samples/sec	accuracy=0.695045	lr=0.388418
Epoch[17] Batch [699]	Speed: 4890.905107 samples/sec	accuracy=0.695081	lr=0.388344
Epoch[17] Batch [749]	Speed: 5143.990549 samples/sec	accuracy=0.695064	lr=0.388271
Epoch[17] Batch [799]	Speed: 4948.453994 samples/sec	accuracy=0.695123	lr=0.388197
Epoch[17] Batch [849]	Speed: 5055.115057 samples/sec	accuracy=0.695243	lr=0.388123
Epoch[17] Batch [899]	Speed: 4981.540666 samples/sec	accuracy=0.695351	lr=0.388049
Epoch[17] Batch [949]	Speed: 4859.550450 samples/sec	accuracy=0.695535	lr=0.387974
Epoch[17] Batch [999]	Speed: 4954.468716 samples/sec	accuracy=0.695593	lr=0.387900
Epoch[17] Batch [1049]	Speed: 5170.159361 samples/sec	accuracy=0.695687	lr=0.387825
Epoch[17] Batch [1099]	Speed: 4862.766473 samples/sec	accuracy=0.695856	lr=0.387750
Epoch[17] Batch [1149]	Speed: 4903.325432 samples/sec	accuracy=0.695908	lr=0.387674
Epoch[17] Batch [1199]	Speed: 5054.986307 samples/sec	accuracy=0.696018	lr=0.387599
Epoch[17] Batch [1249]	Speed: 5339.585552 samples/sec	accuracy=0.696139	lr=0.387523
[Epoch 17] training: accuracy=0.696139
[Epoch 17] speed: 4960 samples/sec	time cost: 267.938040
[Epoch 17] validation: err-top1=0.424601 err-top5=0.178691
Epoch[18] Batch [49]	Speed: 4733.165458 samples/sec	accuracy=0.694996	lr=0.387445
Epoch[18] Batch [99]	Speed: 4886.684263 samples/sec	accuracy=0.695920	lr=0.387369
Epoch[18] Batch [149]	Speed: 4880.656412 samples/sec	accuracy=0.695635	lr=0.387293
Epoch[18] Batch [199]	Speed: 4648.491086 samples/sec	accuracy=0.695407	lr=0.387216
Epoch[18] Batch [249]	Speed: 4895.497817 samples/sec	accuracy=0.695181	lr=0.387139
Epoch[18] Batch [299]	Speed: 5090.160400 samples/sec	accuracy=0.695689	lr=0.387062
Epoch[18] Batch [349]	Speed: 5047.626091 samples/sec	accuracy=0.695724	lr=0.386984
Epoch[18] Batch [399]	Speed: 4981.087491 samples/sec	accuracy=0.695905	lr=0.386907
Epoch[18] Batch [449]	Speed: 4845.220258 samples/sec	accuracy=0.696363	lr=0.386829
Epoch[18] Batch [499]	Speed: 4963.474898 samples/sec	accuracy=0.696455	lr=0.386751
Epoch[18] Batch [549]	Speed: 5052.424801 samples/sec	accuracy=0.696671	lr=0.386673
Epoch[18] Batch [599]	Speed: 5056.718680 samples/sec	accuracy=0.696792	lr=0.386594
Epoch[18] Batch [649]	Speed: 5086.560412 samples/sec	accuracy=0.696998	lr=0.386515
Epoch[18] Batch [699]	Speed: 5018.225347 samples/sec	accuracy=0.697037	lr=0.386436
Epoch[18] Batch [749]	Speed: 4868.013598 samples/sec	accuracy=0.697148	lr=0.386357
Epoch[18] Batch [799]	Speed: 5054.264736 samples/sec	accuracy=0.697246	lr=0.386278
Epoch[18] Batch [849]	Speed: 5021.012185 samples/sec	accuracy=0.697463	lr=0.386198
Epoch[18] Batch [899]	Speed: 4842.532930 samples/sec	accuracy=0.697562	lr=0.386119
Epoch[18] Batch [949]	Speed: 4851.112515 samples/sec	accuracy=0.697656	lr=0.386038
Epoch[18] Batch [999]	Speed: 4957.306241 samples/sec	accuracy=0.697863	lr=0.385958
Epoch[18] Batch [1049]	Speed: 5079.919867 samples/sec	accuracy=0.698068	lr=0.385878
Epoch[18] Batch [1099]	Speed: 4896.672465 samples/sec	accuracy=0.698156	lr=0.385797
Epoch[18] Batch [1149]	Speed: 5073.852339 samples/sec	accuracy=0.698282	lr=0.385716
Epoch[18] Batch [1199]	Speed: 5013.088867 samples/sec	accuracy=0.698453	lr=0.385635
Epoch[18] Batch [1249]	Speed: 4982.074714 samples/sec	accuracy=0.698648	lr=0.385554
[Epoch 18] training: accuracy=0.698653
[Epoch 18] speed: 4945 samples/sec	time cost: 268.834403
[Epoch 18] validation: err-top1=0.410116 err-top5=0.173131
Epoch[19] Batch [49]	Speed: 4968.039124 samples/sec	accuracy=0.698830	lr=0.385470
Epoch[19] Batch [99]	Speed: 4737.940472 samples/sec	accuracy=0.698454	lr=0.385388
Epoch[19] Batch [149]	Speed: 4902.626360 samples/sec	accuracy=0.698291	lr=0.385306
Epoch[19] Batch [199]	Speed: 5110.060957 samples/sec	accuracy=0.698865	lr=0.385224
Epoch[19] Batch [249]	Speed: 5226.893058 samples/sec	accuracy=0.698956	lr=0.385142
Epoch[19] Batch [299]	Speed: 5075.381266 samples/sec	accuracy=0.699320	lr=0.385059
Epoch[19] Batch [349]	Speed: 5030.336258 samples/sec	accuracy=0.699477	lr=0.384976
Epoch[19] Batch [399]	Speed: 4952.967206 samples/sec	accuracy=0.699703	lr=0.384893
Epoch[19] Batch [449]	Speed: 4792.233715 samples/sec	accuracy=0.699708	lr=0.384810
Epoch[19] Batch [499]	Speed: 5024.738040 samples/sec	accuracy=0.699867	lr=0.384726
Epoch[19] Batch [549]	Speed: 4674.414684 samples/sec	accuracy=0.700085	lr=0.384642
Epoch[19] Batch [599]	Speed: 4708.988641 samples/sec	accuracy=0.700153	lr=0.384558
Epoch[19] Batch [649]	Speed: 4771.397013 samples/sec	accuracy=0.700411	lr=0.384474
Epoch[19] Batch [699]	Speed: 4876.598297 samples/sec	accuracy=0.700647	lr=0.384389
Epoch[19] Batch [749]	Speed: 4783.721563 samples/sec	accuracy=0.700681	lr=0.384305
Epoch[19] Batch [799]	Speed: 5020.120722 samples/sec	accuracy=0.700797	lr=0.384220
Epoch[19] Batch [849]	Speed: 4961.712955 samples/sec	accuracy=0.700973	lr=0.384135
Epoch[19] Batch [899]	Speed: 4967.694468 samples/sec	accuracy=0.701053	lr=0.384049
Epoch[19] Batch [949]	Speed: 4941.138110 samples/sec	accuracy=0.701155	lr=0.383964
Epoch[19] Batch [999]	Speed: 5056.509480 samples/sec	accuracy=0.701258	lr=0.383878
Epoch[19] Batch [1049]	Speed: 4990.525245 samples/sec	accuracy=0.701408	lr=0.383792
Epoch[19] Batch [1099]	Speed: 5107.635037 samples/sec	accuracy=0.701494	lr=0.383706
Epoch[19] Batch [1149]	Speed: 4913.089244 samples/sec	accuracy=0.701644	lr=0.383619
Epoch[19] Batch [1199]	Speed: 4945.308497 samples/sec	accuracy=0.701801	lr=0.383533
Epoch[19] Batch [1249]	Speed: 5121.316745 samples/sec	accuracy=0.701991	lr=0.383446
[Epoch 19] training: accuracy=0.701989
[Epoch 19] speed: 4938 samples/sec	time cost: 269.057917
[Epoch 19] validation: err-top1=0.422553 err-top5=0.177376
Epoch[20] Batch [49]	Speed: 4728.228802 samples/sec	accuracy=0.702059	lr=0.383357
Epoch[20] Batch [99]	Speed: 4899.797741 samples/sec	accuracy=0.700609	lr=0.383270
Epoch[20] Batch [149]	Speed: 4830.970312 samples/sec	accuracy=0.701216	lr=0.383182
Epoch[20] Batch [199]	Speed: 4824.547123 samples/sec	accuracy=0.701076	lr=0.383094
Epoch[20] Batch [249]	Speed: 4881.398720 samples/sec	accuracy=0.700994	lr=0.383006
Epoch[20] Batch [299]	Speed: 4926.421184 samples/sec	accuracy=0.700807	lr=0.382918
Epoch[20] Batch [349]	Speed: 5028.251491 samples/sec	accuracy=0.701097	lr=0.382830
Epoch[20] Batch [399]	Speed: 4784.957681 samples/sec	accuracy=0.701111	lr=0.382741
Epoch[20] Batch [449]	Speed: 5064.810249 samples/sec	accuracy=0.701341	lr=0.382652
Epoch[20] Batch [499]	Speed: 4913.715748 samples/sec	accuracy=0.701472	lr=0.382563
Epoch[20] Batch [549]	Speed: 4881.329927 samples/sec	accuracy=0.701736	lr=0.382474
Epoch[20] Batch [599]	Speed: 4888.870735 samples/sec	accuracy=0.701817	lr=0.382384
Epoch[20] Batch [649]	Speed: 5049.167268 samples/sec	accuracy=0.701952	lr=0.382295
Epoch[20] Batch [699]	Speed: 5043.838626 samples/sec	accuracy=0.702049	lr=0.382205
Epoch[20] Batch [749]	Speed: 5055.993638 samples/sec	accuracy=0.702177	lr=0.382115
Epoch[20] Batch [799]	Speed: 4906.723331 samples/sec	accuracy=0.702423	lr=0.382024
Epoch[20] Batch [849]	Speed: 5297.871283 samples/sec	accuracy=0.702625	lr=0.381934
Epoch[20] Batch [899]	Speed: 5197.591165 samples/sec	accuracy=0.702785	lr=0.381843
Epoch[20] Batch [949]	Speed: 4702.550413 samples/sec	accuracy=0.702952	lr=0.381752
Epoch[20] Batch [999]	Speed: 4891.835952 samples/sec	accuracy=0.703124	lr=0.381661
Epoch[20] Batch [1049]	Speed: 4977.331237 samples/sec	accuracy=0.703370	lr=0.381569
Epoch[20] Batch [1099]	Speed: 4910.312877 samples/sec	accuracy=0.703520	lr=0.381477
Epoch[20] Batch [1149]	Speed: 5021.264364 samples/sec	accuracy=0.703635	lr=0.381385
Epoch[20] Batch [1199]	Speed: 4863.562827 samples/sec	accuracy=0.703841	lr=0.381293
Epoch[20] Batch [1249]	Speed: 5021.457859 samples/sec	accuracy=0.703983	lr=0.381201
[Epoch 20] training: accuracy=0.703982
[Epoch 20] speed: 4935 samples/sec	time cost: 269.357440
[Epoch 20] validation: err-top1=0.397282 err-top5=0.164401
Epoch[21] Batch [49]	Speed: 4855.140930 samples/sec	accuracy=0.701882	lr=0.381107
Epoch[21] Batch [99]	Speed: 4940.706238 samples/sec	accuracy=0.703534	lr=0.381014
Epoch[21] Batch [149]	Speed: 5040.355300 samples/sec	accuracy=0.704287	lr=0.380921
Epoch[21] Batch [199]	Speed: 5044.845430 samples/sec	accuracy=0.704343	lr=0.380828
Epoch[21] Batch [249]	Speed: 4941.621228 samples/sec	accuracy=0.704232	lr=0.380734
Epoch[21] Batch [299]	Speed: 4987.299876 samples/sec	accuracy=0.704282	lr=0.380641
Epoch[21] Batch [349]	Speed: 4783.036575 samples/sec	accuracy=0.704526	lr=0.380547
Epoch[21] Batch [399]	Speed: 4846.688534 samples/sec	accuracy=0.704701	lr=0.380453
Epoch[21] Batch [449]	Speed: 4877.789042 samples/sec	accuracy=0.704766	lr=0.380359
Epoch[21] Batch [499]	Speed: 5011.803203 samples/sec	accuracy=0.704780	lr=0.380264
Epoch[21] Batch [549]	Speed: 4837.600673 samples/sec	accuracy=0.704863	lr=0.380169
Epoch[21] Batch [599]	Speed: 5166.491730 samples/sec	accuracy=0.705016	lr=0.380074
Epoch[21] Batch [649]	Speed: 4837.573756 samples/sec	accuracy=0.705254	lr=0.379979
Epoch[21] Batch [699]	Speed: 4976.133951 samples/sec	accuracy=0.705386	lr=0.379884
Epoch[21] Batch [749]	Speed: 4909.804878 samples/sec	accuracy=0.705470	lr=0.379788
Epoch[21] Batch [799]	Speed: 5041.776390 samples/sec	accuracy=0.705515	lr=0.379693
Epoch[21] Batch [849]	Speed: 4999.675800 samples/sec	accuracy=0.705578	lr=0.379597
Epoch[21] Batch [899]	Speed: 5023.654394 samples/sec	accuracy=0.705719	lr=0.379501
Epoch[21] Batch [949]	Speed: 4959.504954 samples/sec	accuracy=0.705839	lr=0.379404
Epoch[21] Batch [999]	Speed: 4888.994056 samples/sec	accuracy=0.706051	lr=0.379307
Epoch[21] Batch [1049]	Speed: 4924.366764 samples/sec	accuracy=0.706177	lr=0.379211
Epoch[21] Batch [1099]	Speed: 5103.852358 samples/sec	accuracy=0.706328	lr=0.379114
Epoch[21] Batch [1149]	Speed: 4841.365774 samples/sec	accuracy=0.706477	lr=0.379016
Epoch[21] Batch [1199]	Speed: 4907.085929 samples/sec	accuracy=0.706617	lr=0.378919
Epoch[21] Batch [1249]	Speed: 5201.595888 samples/sec	accuracy=0.706714	lr=0.378821
[Epoch 21] training: accuracy=0.706725
[Epoch 21] speed: 4950 samples/sec	time cost: 268.201026
[Epoch 21] validation: err-top1=0.392997 err-top5=0.158781
Epoch[22] Batch [49]	Speed: 4804.063192 samples/sec	accuracy=0.703806	lr=0.378721
Epoch[22] Batch [99]	Speed: 5016.411670 samples/sec	accuracy=0.704940	lr=0.378623
Epoch[22] Batch [149]	Speed: 4966.709030 samples/sec	accuracy=0.704877	lr=0.378525
Epoch[22] Batch [199]	Speed: 4949.963379 samples/sec	accuracy=0.705404	lr=0.378426
Epoch[22] Batch [249]	Speed: 4896.400492 samples/sec	accuracy=0.705679	lr=0.378328
Epoch[22] Batch [299]	Speed: 4970.290854 samples/sec	accuracy=0.705692	lr=0.378229
Epoch[22] Batch [349]	Speed: 4884.679178 samples/sec	accuracy=0.705801	lr=0.378129
Epoch[22] Batch [399]	Speed: 4846.355369 samples/sec	accuracy=0.705890	lr=0.378030
Epoch[22] Batch [449]	Speed: 5034.583333 samples/sec	accuracy=0.705985	lr=0.377930
Epoch[22] Batch [499]	Speed: 4993.274660 samples/sec	accuracy=0.706169	lr=0.377831
Epoch[22] Batch [549]	Speed: 5045.430478 samples/sec	accuracy=0.706399	lr=0.377731
Epoch[22] Batch [599]	Speed: 4693.147441 samples/sec	accuracy=0.706418	lr=0.377630
Epoch[22] Batch [649]	Speed: 4996.600885 samples/sec	accuracy=0.706672	lr=0.377530
Epoch[22] Batch [699]	Speed: 4952.170999 samples/sec	accuracy=0.706837	lr=0.377429
Epoch[22] Batch [749]	Speed: 4999.108529 samples/sec	accuracy=0.707075	lr=0.377328
Epoch[22] Batch [799]	Speed: 5031.479137 samples/sec	accuracy=0.707168	lr=0.377227
Epoch[22] Batch [849]	Speed: 4888.407223 samples/sec	accuracy=0.707162	lr=0.377126
Epoch[22] Batch [899]	Speed: 4979.072891 samples/sec	accuracy=0.707366	lr=0.377024
Epoch[22] Batch [949]	Speed: 4916.178232 samples/sec	accuracy=0.707613	lr=0.376923
Epoch[22] Batch [999]	Speed: 5045.898993 samples/sec	accuracy=0.707782	lr=0.376821
Epoch[22] Batch [1049]	Speed: 4837.680227 samples/sec	accuracy=0.707948	lr=0.376719
Epoch[22] Batch [1099]	Speed: 5085.593976 samples/sec	accuracy=0.708081	lr=0.376616
Epoch[22] Batch [1149]	Speed: 5050.132971 samples/sec	accuracy=0.708194	lr=0.376514
Epoch[22] Batch [1199]	Speed: 5049.956616 samples/sec	accuracy=0.708308	lr=0.376411
Epoch[22] Batch [1249]	Speed: 5004.430727 samples/sec	accuracy=0.708417	lr=0.376308
[Epoch 22] training: accuracy=0.708419
[Epoch 22] speed: 4951 samples/sec	time cost: 268.170660
[Epoch 22] validation: err-top1=0.404257 err-top5=0.166932
Epoch[23] Batch [49]	Speed: 4808.118988 samples/sec	accuracy=0.710411	lr=0.376203
Epoch[23] Batch [99]	Speed: 4756.568590 samples/sec	accuracy=0.709885	lr=0.376099
Epoch[23] Batch [149]	Speed: 4910.813119 samples/sec	accuracy=0.709664	lr=0.375996
Epoch[23] Batch [199]	Speed: 4915.848273 samples/sec	accuracy=0.709325	lr=0.375892
Epoch[23] Batch [249]	Speed: 4920.163182 samples/sec	accuracy=0.709747	lr=0.375788
Epoch[23] Batch [299]	Speed: 4915.911065 samples/sec	accuracy=0.709759	lr=0.375683
Epoch[23] Batch [349]	Speed: 5009.732247 samples/sec	accuracy=0.709827	lr=0.375579
Epoch[23] Batch [399]	Speed: 4884.488969 samples/sec	accuracy=0.709819	lr=0.375474
Epoch[23] Batch [449]	Speed: 4803.872441 samples/sec	accuracy=0.709872	lr=0.375369
Epoch[23] Batch [499]	Speed: 5049.742988 samples/sec	accuracy=0.709987	lr=0.375264
Epoch[23] Batch [549]	Speed: 4797.187225 samples/sec	accuracy=0.710184	lr=0.375159
Epoch[23] Batch [599]	Speed: 4922.164905 samples/sec	accuracy=0.710180	lr=0.375053
Epoch[23] Batch [649]	Speed: 4879.928969 samples/sec	accuracy=0.710440	lr=0.374948
Epoch[23] Batch [699]	Speed: 4896.751629 samples/sec	accuracy=0.710478	lr=0.374842
Epoch[23] Batch [749]	Speed: 4931.398609 samples/sec	accuracy=0.710587	lr=0.374736
Epoch[23] Batch [799]	Speed: 4974.958101 samples/sec	accuracy=0.710649	lr=0.374629
Epoch[23] Batch [849]	Speed: 5012.663866 samples/sec	accuracy=0.710746	lr=0.374523
Epoch[23] Batch [899]	Speed: 4978.121016 samples/sec	accuracy=0.710818	lr=0.374416
Epoch[23] Batch [949]	Speed: 4765.170600 samples/sec	accuracy=0.710894	lr=0.374309
Epoch[23] Batch [999]	Speed: 4845.682614 samples/sec	accuracy=0.711000	lr=0.374202
Epoch[23] Batch [1049]	Speed: 5120.404451 samples/sec	accuracy=0.711077	lr=0.374095
Epoch[23] Batch [1099]	Speed: 5082.610861 samples/sec	accuracy=0.711115	lr=0.373987
Epoch[23] Batch [1149]	Speed: 5232.142101 samples/sec	accuracy=0.711120	lr=0.373879
Epoch[23] Batch [1199]	Speed: 5160.435120 samples/sec	accuracy=0.711235	lr=0.373771
Epoch[23] Batch [1249]	Speed: 5062.749928 samples/sec	accuracy=0.711304	lr=0.373663
[Epoch 23] training: accuracy=0.711310
[Epoch 23] speed: 4938 samples/sec	time cost: 268.879605
[Epoch 23] validation: err-top1=0.393290 err-top5=0.159037
Epoch[24] Batch [49]	Speed: 4666.755884 samples/sec	accuracy=0.710357	lr=0.373552
Epoch[24] Batch [99]	Speed: 4997.788152 samples/sec	accuracy=0.711124	lr=0.373444
Epoch[24] Batch [149]	Speed: 4965.643493 samples/sec	accuracy=0.711056	lr=0.373335
Epoch[24] Batch [199]	Speed: 4832.509564 samples/sec	accuracy=0.711022	lr=0.373226
Epoch[24] Batch [249]	Speed: 4986.052293 samples/sec	accuracy=0.711133	lr=0.373117
Epoch[24] Batch [299]	Speed: 4958.362364 samples/sec	accuracy=0.711214	lr=0.373007
Epoch[24] Batch [349]	Speed: 5055.406495 samples/sec	accuracy=0.710905	lr=0.372897
Epoch[24] Batch [399]	Speed: 4960.629850 samples/sec	accuracy=0.710942	lr=0.372788
Epoch[24] Batch [449]	Speed: 4776.377846 samples/sec	accuracy=0.711013	lr=0.372678
Epoch[24] Batch [499]	Speed: 4786.220039 samples/sec	accuracy=0.711250	lr=0.372567
Epoch[24] Batch [549]	Speed: 5096.345751 samples/sec	accuracy=0.711398	lr=0.372457
Epoch[24] Batch [599]	Speed: 4944.283198 samples/sec	accuracy=0.711575	lr=0.372346
Epoch[24] Batch [649]	Speed: 5272.466570 samples/sec	accuracy=0.711661	lr=0.372235
Epoch[24] Batch [699]	Speed: 5260.392509 samples/sec	accuracy=0.711788	lr=0.372124
Epoch[24] Batch [749]	Speed: 5076.880989 samples/sec	accuracy=0.712042	lr=0.372013
Epoch[24] Batch [799]	Speed: 4995.473794 samples/sec	accuracy=0.712214	lr=0.371901
Epoch[24] Batch [849]	Speed: 5072.580973 samples/sec	accuracy=0.712349	lr=0.371790
Epoch[24] Batch [899]	Speed: 5070.517919 samples/sec	accuracy=0.712361	lr=0.371678
Epoch[24] Batch [949]	Speed: 4930.220487 samples/sec	accuracy=0.712420	lr=0.371565
Epoch[24] Batch [999]	Speed: 4991.404487 samples/sec	accuracy=0.712590	lr=0.371453
Epoch[24] Batch [1049]	Speed: 4910.125045 samples/sec	accuracy=0.712629	lr=0.371341
Epoch[24] Batch [1099]	Speed: 5042.575270 samples/sec	accuracy=0.712743	lr=0.371228
Epoch[24] Batch [1149]	Speed: 4851.167199 samples/sec	accuracy=0.712780	lr=0.371115
Epoch[24] Batch [1199]	Speed: 5250.900186 samples/sec	accuracy=0.712898	lr=0.371002
Epoch[24] Batch [1249]	Speed: 5010.605060 samples/sec	accuracy=0.712991	lr=0.370888
[Epoch 24] training: accuracy=0.712998
[Epoch 24] speed: 4981 samples/sec	time cost: 266.800822
[Epoch 24] validation: err-top1=0.397461 err-top5=0.161452
Epoch[25] Batch [49]	Speed: 4903.482289 samples/sec	accuracy=0.713181	lr=0.370773
Epoch[25] Batch [99]	Speed: 4771.716666 samples/sec	accuracy=0.713180	lr=0.370659
Epoch[25] Batch [149]	Speed: 4813.379535 samples/sec	accuracy=0.713209	lr=0.370545
Epoch[25] Batch [199]	Speed: 5140.918861 samples/sec	accuracy=0.713453	lr=0.370431
Epoch[25] Batch [249]	Speed: 5178.712141 samples/sec	accuracy=0.713341	lr=0.370316
Epoch[25] Batch [299]	Speed: 4862.017052 samples/sec	accuracy=0.713499	lr=0.370202
Epoch[25] Batch [349]	Speed: 5148.044499 samples/sec	accuracy=0.713730	lr=0.370087
Epoch[25] Batch [399]	Speed: 5094.378252 samples/sec	accuracy=0.713848	lr=0.369972
Epoch[25] Batch [449]	Speed: 4828.198241 samples/sec	accuracy=0.713826	lr=0.369857
Epoch[25] Batch [499]	Speed: 5146.607406 samples/sec	accuracy=0.713874	lr=0.369741
Epoch[25] Batch [549]	Speed: 4900.029394 samples/sec	accuracy=0.714086	lr=0.369626
Epoch[25] Batch [599]	Speed: 5220.490266 samples/sec	accuracy=0.714186	lr=0.369510
Epoch[25] Batch [649]	Speed: 4935.422119 samples/sec	accuracy=0.714265	lr=0.369394
Epoch[25] Batch [699]	Speed: 4953.330730 samples/sec	accuracy=0.714230	lr=0.369278
Epoch[25] Batch [749]	Speed: 4979.044723 samples/sec	accuracy=0.714312	lr=0.369161
Epoch[25] Batch [799]	Speed: 4961.159080 samples/sec	accuracy=0.714334	lr=0.369045
Epoch[25] Batch [849]	Speed: 4915.064516 samples/sec	accuracy=0.714398	lr=0.368928
Epoch[25] Batch [899]	Speed: 5120.917767 samples/sec	accuracy=0.714438	lr=0.368811
Epoch[25] Batch [949]	Speed: 4728.276899 samples/sec	accuracy=0.714605	lr=0.368694
Epoch[25] Batch [999]	Speed: 5038.092724 samples/sec	accuracy=0.714729	lr=0.368576
Epoch[25] Batch [1049]	Speed: 4875.495799 samples/sec	accuracy=0.714872	lr=0.368459
Epoch[25] Batch [1099]	Speed: 5125.408295 samples/sec	accuracy=0.715019	lr=0.368341
Epoch[25] Batch [1149]	Speed: 4870.804301 samples/sec	accuracy=0.715225	lr=0.368223
Epoch[25] Batch [1199]	Speed: 4920.225070 samples/sec	accuracy=0.715314	lr=0.368105
Epoch[25] Batch [1249]	Speed: 5021.557313 samples/sec	accuracy=0.715416	lr=0.367986
[Epoch 25] training: accuracy=0.715418
[Epoch 25] speed: 4969 samples/sec	time cost: 267.063727
[Epoch 25] validation: err-top1=0.402065 err-top5=0.162508
Epoch[26] Batch [49]	Speed: 4734.696975 samples/sec	accuracy=0.715914	lr=0.367865
Epoch[26] Batch [99]	Speed: 4745.780048 samples/sec	accuracy=0.715208	lr=0.367747
Epoch[26] Batch [149]	Speed: 4943.085031 samples/sec	accuracy=0.714847	lr=0.367628
Epoch[26] Batch [199]	Speed: 5120.267226 samples/sec	accuracy=0.714740	lr=0.367508
Epoch[26] Batch [249]	Speed: 4852.759047 samples/sec	accuracy=0.714864	lr=0.367389
Epoch[26] Batch [299]	Speed: 4841.403866 samples/sec	accuracy=0.715106	lr=0.367269
Epoch[26] Batch [349]	Speed: 4921.831773 samples/sec	accuracy=0.714977	lr=0.367150
Epoch[26] Batch [399]	Speed: 4861.379450 samples/sec	accuracy=0.715282	lr=0.367030
Epoch[26] Batch [449]	Speed: 4900.478226 samples/sec	accuracy=0.715721	lr=0.366909
Epoch[26] Batch [499]	Speed: 5105.200010 samples/sec	accuracy=0.715986	lr=0.366789
Epoch[26] Batch [549]	Speed: 4901.728884 samples/sec	accuracy=0.715921	lr=0.366668
Epoch[26] Batch [599]	Speed: 5138.242870 samples/sec	accuracy=0.715955	lr=0.366548
Epoch[26] Batch [649]	Speed: 5046.108858 samples/sec	accuracy=0.716135	lr=0.366426
Epoch[26] Batch [699]	Speed: 5071.941217 samples/sec	accuracy=0.716316	lr=0.366305
Epoch[26] Batch [749]	Speed: 4910.650627 samples/sec	accuracy=0.716482	lr=0.366184
Epoch[26] Batch [799]	Speed: 4987.447557 samples/sec	accuracy=0.716585	lr=0.366062
Epoch[26] Batch [849]	Speed: 4971.981077 samples/sec	accuracy=0.716707	lr=0.365940
Epoch[26] Batch [899]	Speed: 4974.367850 samples/sec	accuracy=0.716782	lr=0.365819
Epoch[26] Batch [949]	Speed: 5095.585000 samples/sec	accuracy=0.716901	lr=0.365696
Epoch[26] Batch [999]	Speed: 5004.384662 samples/sec	accuracy=0.717070	lr=0.365574
Epoch[26] Batch [1049]	Speed: 4977.217031 samples/sec	accuracy=0.717181	lr=0.365451
Epoch[26] Batch [1099]	Speed: 4874.222755 samples/sec	accuracy=0.717258	lr=0.365329
Epoch[26] Batch [1149]	Speed: 4954.829831 samples/sec	accuracy=0.717410	lr=0.365206
Epoch[26] Batch [1199]	Speed: 4843.919379 samples/sec	accuracy=0.717489	lr=0.365082
Epoch[26] Batch [1249]	Speed: 5034.508975 samples/sec	accuracy=0.717590	lr=0.364959
[Epoch 26] training: accuracy=0.717590
[Epoch 26] speed: 4945 samples/sec	time cost: 268.649455
[Epoch 26] validation: err-top1=0.399952 err-top5=0.167391
Epoch[27] Batch [49]	Speed: 4936.233827 samples/sec	accuracy=0.716733	lr=0.364833
Epoch[27] Batch [99]	Speed: 4765.065923 samples/sec	accuracy=0.717052	lr=0.364709
Epoch[27] Batch [149]	Speed: 4747.788162 samples/sec	accuracy=0.716909	lr=0.364585
Epoch[27] Batch [199]	Speed: 4873.155169 samples/sec	accuracy=0.717185	lr=0.364461
Epoch[27] Batch [249]	Speed: 4922.193787 samples/sec	accuracy=0.717213	lr=0.364337
Epoch[27] Batch [299]	Speed: 4937.528688 samples/sec	accuracy=0.717432	lr=0.364212
Epoch[27] Batch [349]	Speed: 4878.033909 samples/sec	accuracy=0.717448	lr=0.364087
Epoch[27] Batch [399]	Speed: 4886.502349 samples/sec	accuracy=0.717622	lr=0.363962
Epoch[27] Batch [449]	Speed: 5063.135954 samples/sec	accuracy=0.717717	lr=0.363837
Epoch[27] Batch [499]	Speed: 5109.265838 samples/sec	accuracy=0.717828	lr=0.363712
Epoch[27] Batch [549]	Speed: 4914.044971 samples/sec	accuracy=0.718055	lr=0.363586
Epoch[27] Batch [599]	Speed: 5086.699932 samples/sec	accuracy=0.718399	lr=0.363461
Epoch[27] Batch [649]	Speed: 5088.594824 samples/sec	accuracy=0.718480	lr=0.363335
Epoch[27] Batch [699]	Speed: 5107.631878 samples/sec	accuracy=0.718582	lr=0.363209
Epoch[27] Batch [749]	Speed: 4983.111354 samples/sec	accuracy=0.718664	lr=0.363082
Epoch[27] Batch [799]	Speed: 4976.709628 samples/sec	accuracy=0.718725	lr=0.362956
Epoch[27] Batch [849]	Speed: 5055.767001 samples/sec	accuracy=0.718842	lr=0.362829
Epoch[27] Batch [899]	Speed: 4929.602102 samples/sec	accuracy=0.718890	lr=0.362702
Epoch[27] Batch [949]	Speed: 4996.966656 samples/sec	accuracy=0.719112	lr=0.362575
Epoch[27] Batch [999]	Speed: 4952.126347 samples/sec	accuracy=0.719192	lr=0.362448
Epoch[27] Batch [1049]	Speed: 4901.709752 samples/sec	accuracy=0.719216	lr=0.362320
Epoch[27] Batch [1099]	Speed: 4980.307859 samples/sec	accuracy=0.719417	lr=0.362193
Epoch[27] Batch [1149]	Speed: 4903.840713 samples/sec	accuracy=0.719576	lr=0.362065
Epoch[27] Batch [1199]	Speed: 4913.101384 samples/sec	accuracy=0.719654	lr=0.361937
Epoch[27] Batch [1249]	Speed: 4985.737080 samples/sec	accuracy=0.719790	lr=0.361809
[Epoch 27] training: accuracy=0.719795
[Epoch 27] speed: 4949 samples/sec	time cost: 267.921654
[Epoch 27] validation: err-top1=0.406449 err-top5=0.174825
Epoch[28] Batch [49]	Speed: 4700.590254 samples/sec	accuracy=0.720335	lr=0.361678
Epoch[28] Batch [99]	Speed: 4727.859887 samples/sec	accuracy=0.720058	lr=0.361549
Epoch[28] Batch [149]	Speed: 4800.155537 samples/sec	accuracy=0.719517	lr=0.361420
Epoch[28] Batch [199]	Speed: 4950.777481 samples/sec	accuracy=0.719451	lr=0.361291
Epoch[28] Batch [249]	Speed: 5098.603221 samples/sec	accuracy=0.719465	lr=0.361162
Epoch[28] Batch [299]	Speed: 4898.715236 samples/sec	accuracy=0.719423	lr=0.361032
Epoch[28] Batch [349]	Speed: 4908.873803 samples/sec	accuracy=0.719479	lr=0.360903
Epoch[28] Batch [399]	Speed: 4985.016165 samples/sec	accuracy=0.719302	lr=0.360773
Epoch[28] Batch [449]	Speed: 4978.675799 samples/sec	accuracy=0.719371	lr=0.360643
Epoch[28] Batch [499]	Speed: 4819.559260 samples/sec	accuracy=0.719372	lr=0.360513
Epoch[28] Batch [549]	Speed: 4921.948190 samples/sec	accuracy=0.719565	lr=0.360382
Epoch[28] Batch [599]	Speed: 4934.081316 samples/sec	accuracy=0.719628	lr=0.360252
Epoch[28] Batch [649]	Speed: 4980.863592 samples/sec	accuracy=0.719803	lr=0.360121
Epoch[28] Batch [699]	Speed: 5115.129394 samples/sec	accuracy=0.719857	lr=0.359990
Epoch[28] Batch [749]	Speed: 5102.091667 samples/sec	accuracy=0.719951	lr=0.359859
Epoch[28] Batch [799]	Speed: 4748.391379 samples/sec	accuracy=0.720152	lr=0.359728
Epoch[28] Batch [849]	Speed: 4909.910399 samples/sec	accuracy=0.720125	lr=0.359596
Epoch[28] Batch [899]	Speed: 5052.000473 samples/sec	accuracy=0.720182	lr=0.359465
Epoch[28] Batch [949]	Speed: 5044.990494 samples/sec	accuracy=0.720251	lr=0.359333
Epoch[28] Batch [999]	Speed: 4965.547160 samples/sec	accuracy=0.720313	lr=0.359201
Epoch[28] Batch [1049]	Speed: 4901.622485 samples/sec	accuracy=0.720475	lr=0.359068
Epoch[28] Batch [1099]	Speed: 4911.269995 samples/sec	accuracy=0.720612	lr=0.358936
Epoch[28] Batch [1149]	Speed: 4896.869765 samples/sec	accuracy=0.720699	lr=0.358803
Epoch[28] Batch [1199]	Speed: 5016.321677 samples/sec	accuracy=0.720858	lr=0.358670
Epoch[28] Batch [1249]	Speed: 5197.339078 samples/sec	accuracy=0.720970	lr=0.358537
[Epoch 28] training: accuracy=0.720971
[Epoch 28] speed: 4934 samples/sec	time cost: 268.858816
[Epoch 28] validation: err-top1=0.378069 err-top5=0.147700
Epoch[29] Batch [49]	Speed: 4681.054287 samples/sec	accuracy=0.720695	lr=0.358401
Epoch[29] Batch [99]	Speed: 4762.401372 samples/sec	accuracy=0.720871	lr=0.358268
Epoch[29] Batch [149]	Speed: 4775.204026 samples/sec	accuracy=0.720371	lr=0.358134
Epoch[29] Batch [199]	Speed: 4959.340713 samples/sec	accuracy=0.719755	lr=0.358001
Epoch[29] Batch [249]	Speed: 4653.528458 samples/sec	accuracy=0.720083	lr=0.357867
Epoch[29] Batch [299]	Speed: 4928.088482 samples/sec	accuracy=0.720028	lr=0.357732
Epoch[29] Batch [349]	Speed: 5000.200703 samples/sec	accuracy=0.720089	lr=0.357598
Epoch[29] Batch [399]	Speed: 5006.802187 samples/sec	accuracy=0.720403	lr=0.357464
Epoch[29] Batch [449]	Speed: 4929.479552 samples/sec	accuracy=0.720677	lr=0.357329
Epoch[29] Batch [499]	Speed: 4990.665347 samples/sec	accuracy=0.720816	lr=0.357194
Epoch[29] Batch [549]	Speed: 5157.124053 samples/sec	accuracy=0.720920	lr=0.357059
Epoch[29] Batch [599]	Speed: 5188.049060 samples/sec	accuracy=0.721072	lr=0.356924
Epoch[29] Batch [649]	Speed: 5127.056832 samples/sec	accuracy=0.721121	lr=0.356788
Epoch[29] Batch [699]	Speed: 5270.160543 samples/sec	accuracy=0.721212	lr=0.356652
Epoch[29] Batch [749]	Speed: 4736.463477 samples/sec	accuracy=0.721361	lr=0.356517
Epoch[29] Batch [799]	Speed: 4833.667993 samples/sec	accuracy=0.721457	lr=0.356380
Epoch[29] Batch [849]	Speed: 4783.301107 samples/sec	accuracy=0.721622	lr=0.356244
Epoch[29] Batch [899]	Speed: 5026.191156 samples/sec	accuracy=0.721737	lr=0.356108
Epoch[29] Batch [949]	Speed: 4921.191244 samples/sec	accuracy=0.721911	lr=0.355971
Epoch[29] Batch [999]	Speed: 4779.962990 samples/sec	accuracy=0.721996	lr=0.355834
Epoch[29] Batch [1049]	Speed: 5014.911033 samples/sec	accuracy=0.722161	lr=0.355698
Epoch[29] Batch [1099]	Speed: 5033.840790 samples/sec	accuracy=0.722244	lr=0.355560
Epoch[29] Batch [1149]	Speed: 5137.763319 samples/sec	accuracy=0.722333	lr=0.355423
Epoch[29] Batch [1199]	Speed: 4905.143505 samples/sec	accuracy=0.722472	lr=0.355285
Epoch[29] Batch [1249]	Speed: 5001.327016 samples/sec	accuracy=0.722571	lr=0.355148
[Epoch 29] training: accuracy=0.722572
[Epoch 29] speed: 4934 samples/sec	time cost: 268.845618
[Epoch 29] validation: err-top1=0.390238 err-top5=0.158162
Epoch[30] Batch [49]	Speed: 4790.961984 samples/sec	accuracy=0.722865	lr=0.355007
Epoch[30] Batch [99]	Speed: 4963.291008 samples/sec	accuracy=0.722691	lr=0.354869
Epoch[30] Batch [149]	Speed: 5031.734254 samples/sec	accuracy=0.722460	lr=0.354731
Epoch[30] Batch [199]	Speed: 4834.214768 samples/sec	accuracy=0.722367	lr=0.354592
Epoch[30] Batch [249]	Speed: 4727.373639 samples/sec	accuracy=0.722419	lr=0.354454
Epoch[30] Batch [299]	Speed: 4926.250991 samples/sec	accuracy=0.722615	lr=0.354315
Epoch[30] Batch [349]	Speed: 5001.222073 samples/sec	accuracy=0.722658	lr=0.354176
Epoch[30] Batch [399]	Speed: 5099.838858 samples/sec	accuracy=0.722695	lr=0.354037
Epoch[30] Batch [449]	Speed: 4998.435979 samples/sec	accuracy=0.722695	lr=0.353897
Epoch[30] Batch [499]	Speed: 5126.899789 samples/sec	accuracy=0.722771	lr=0.353758
Epoch[30] Batch [549]	Speed: 5070.042307 samples/sec	accuracy=0.722988	lr=0.353618
Epoch[30] Batch [599]	Speed: 5012.583836 samples/sec	accuracy=0.723109	lr=0.353478
Epoch[30] Batch [649]	Speed: 4903.236316 samples/sec	accuracy=0.723157	lr=0.353338
Epoch[30] Batch [699]	Speed: 5134.914904 samples/sec	accuracy=0.723159	lr=0.353198
Epoch[30] Batch [749]	Speed: 4841.707860 samples/sec	accuracy=0.723233	lr=0.353057
Epoch[30] Batch [799]	Speed: 5177.451099 samples/sec	accuracy=0.723455	lr=0.352917
Epoch[30] Batch [849]	Speed: 5126.768947 samples/sec	accuracy=0.723596	lr=0.352776
Epoch[30] Batch [899]	Speed: 5150.059359 samples/sec	accuracy=0.723666	lr=0.352635
Epoch[30] Batch [949]	Speed: 4990.741315 samples/sec	accuracy=0.723753	lr=0.352493
Epoch[30] Batch [999]	Speed: 4965.215821 samples/sec	accuracy=0.723915	lr=0.352352
Epoch[30] Batch [1049]	Speed: 5025.752051 samples/sec	accuracy=0.724133	lr=0.352211
Epoch[30] Batch [1099]	Speed: 4996.844339 samples/sec	accuracy=0.724232	lr=0.352069
Epoch[30] Batch [1149]	Speed: 4901.470334 samples/sec	accuracy=0.724407	lr=0.351927
Epoch[30] Batch [1199]	Speed: 4925.338294 samples/sec	accuracy=0.724514	lr=0.351785
Epoch[30] Batch [1249]	Speed: 4990.059303 samples/sec	accuracy=0.724570	lr=0.351642
[Epoch 30] training: accuracy=0.724576
[Epoch 30] speed: 4981 samples/sec	time cost: 266.733903
[Epoch 30] validation: err-top1=0.398756 err-top5=0.163764
Epoch[31] Batch [49]	Speed: 4627.790890 samples/sec	accuracy=0.721440	lr=0.351497
Epoch[31] Batch [99]	Speed: 4910.574831 samples/sec	accuracy=0.723042	lr=0.351354
Epoch[31] Batch [149]	Speed: 4914.042723 samples/sec	accuracy=0.723617	lr=0.351212
Epoch[31] Batch [199]	Speed: 4848.549133 samples/sec	accuracy=0.724125	lr=0.351069
Epoch[31] Batch [249]	Speed: 4934.250237 samples/sec	accuracy=0.724204	lr=0.350925
Epoch[31] Batch [299]	Speed: 4929.352370 samples/sec	accuracy=0.724208	lr=0.350782
Epoch[31] Batch [349]	Speed: 5035.017490 samples/sec	accuracy=0.724206	lr=0.350639
Epoch[31] Batch [399]	Speed: 4991.224438 samples/sec	accuracy=0.724010	lr=0.350495
Epoch[31] Batch [449]	Speed: 5095.689467 samples/sec	accuracy=0.724038	lr=0.350351
Epoch[31] Batch [499]	Speed: 4877.033987 samples/sec	accuracy=0.724243	lr=0.350207
Epoch[31] Batch [549]	Speed: 4849.786346 samples/sec	accuracy=0.724307	lr=0.350062
Epoch[31] Batch [599]	Speed: 5012.292401 samples/sec	accuracy=0.724577	lr=0.349918
Epoch[31] Batch [649]	Speed: 4816.816580 samples/sec	accuracy=0.724667	lr=0.349773
Epoch[31] Batch [699]	Speed: 5047.545533 samples/sec	accuracy=0.724880	lr=0.349629
Epoch[31] Batch [749]	Speed: 4862.265952 samples/sec	accuracy=0.725056	lr=0.349484
Epoch[31] Batch [799]	Speed: 4776.382308 samples/sec	accuracy=0.725141	lr=0.349338
Epoch[31] Batch [849]	Speed: 5050.701784 samples/sec	accuracy=0.725239	lr=0.349193
Epoch[31] Batch [899]	Speed: 4967.408689 samples/sec	accuracy=0.725350	lr=0.349048
Epoch[31] Batch [949]	Speed: 5024.174824 samples/sec	accuracy=0.725512	lr=0.348902
Epoch[31] Batch [999]	Speed: 5140.606160 samples/sec	accuracy=0.725615	lr=0.348756
Epoch[31] Batch [1049]	Speed: 5106.674664 samples/sec	accuracy=0.725705	lr=0.348610
Epoch[31] Batch [1099]	Speed: 4904.624702 samples/sec	accuracy=0.725852	lr=0.348464
Epoch[31] Batch [1149]	Speed: 4779.714466 samples/sec	accuracy=0.725967	lr=0.348317
Epoch[31] Batch [1199]	Speed: 5134.935531 samples/sec	accuracy=0.726055	lr=0.348171
Epoch[31] Batch [1249]	Speed: 5177.796638 samples/sec	accuracy=0.726131	lr=0.348024
[Epoch 31] training: accuracy=0.726138
[Epoch 31] speed: 4943 samples/sec	time cost: 268.470582
[Epoch 31] validation: err-top1=0.388692 err-top5=0.157366
Epoch[32] Batch [49]	Speed: 4656.139169 samples/sec	accuracy=0.724298	lr=0.347874
Epoch[32] Batch [99]	Speed: 5129.005068 samples/sec	accuracy=0.725254	lr=0.347727
Epoch[32] Batch [149]	Speed: 4880.233606 samples/sec	accuracy=0.725456	lr=0.347580
Epoch[32] Batch [199]	Speed: 4643.768554 samples/sec	accuracy=0.725266	lr=0.347432
Epoch[32] Batch [249]	Speed: 4837.087887 samples/sec	accuracy=0.725339	lr=0.347285
Epoch[32] Batch [299]	Speed: 5041.903403 samples/sec	accuracy=0.725732	lr=0.347137
Epoch[32] Batch [349]	Speed: 4987.200732 samples/sec	accuracy=0.725842	lr=0.346989
Epoch[32] Batch [399]	Speed: 4840.562596 samples/sec	accuracy=0.726030	lr=0.346841
Epoch[32] Batch [449]	Speed: 4892.287521 samples/sec	accuracy=0.726207	lr=0.346692
Epoch[32] Batch [499]	Speed: 4910.653884 samples/sec	accuracy=0.726388	lr=0.346544
Epoch[32] Batch [549]	Speed: 4880.590524 samples/sec	accuracy=0.726415	lr=0.346395
Epoch[32] Batch [599]	Speed: 5013.904786 samples/sec	accuracy=0.726496	lr=0.346246
Epoch[32] Batch [649]	Speed: 4943.601874 samples/sec	accuracy=0.726412	lr=0.346097
Epoch[32] Batch [699]	Speed: 4917.829481 samples/sec	accuracy=0.726477	lr=0.345948
Epoch[32] Batch [749]	Speed: 4932.205822 samples/sec	accuracy=0.726359	lr=0.345799
Epoch[32] Batch [799]	Speed: 5087.600134 samples/sec	accuracy=0.726460	lr=0.345649
Epoch[32] Batch [849]	Speed: 5064.398409 samples/sec	accuracy=0.726552	lr=0.345499
Epoch[32] Batch [899]	Speed: 5107.604423 samples/sec	accuracy=0.726591	lr=0.345349
Epoch[32] Batch [949]	Speed: 4951.565591 samples/sec	accuracy=0.726722	lr=0.345199
Epoch[32] Batch [999]	Speed: 4848.460026 samples/sec	accuracy=0.726811	lr=0.345049
Epoch[32] Batch [1049]	Speed: 5025.401811 samples/sec	accuracy=0.726987	lr=0.344899
Epoch[32] Batch [1099]	Speed: 5000.794540 samples/sec	accuracy=0.727177	lr=0.344748
Epoch[32] Batch [1149]	Speed: 4978.606199 samples/sec	accuracy=0.727323	lr=0.344597
Epoch[32] Batch [1199]	Speed: 4975.427106 samples/sec	accuracy=0.727458	lr=0.344446
Epoch[32] Batch [1249]	Speed: 5199.374085 samples/sec	accuracy=0.727608	lr=0.344295
[Epoch 32] training: accuracy=0.727611
[Epoch 32] speed: 4942 samples/sec	time cost: 268.394011
[Epoch 32] validation: err-top1=0.379305 err-top5=0.148178
Epoch[33] Batch [49]	Speed: 4842.397419 samples/sec	accuracy=0.727816	lr=0.344141
Epoch[33] Batch [99]	Speed: 5004.679144 samples/sec	accuracy=0.728127	lr=0.343989
Epoch[33] Batch [149]	Speed: 4927.847498 samples/sec	accuracy=0.727333	lr=0.343838
Epoch[33] Batch [199]	Speed: 4868.250973 samples/sec	accuracy=0.727164	lr=0.343686
Epoch[33] Batch [249]	Speed: 4901.843456 samples/sec	accuracy=0.727068	lr=0.343534
Epoch[33] Batch [299]	Speed: 4956.324348 samples/sec	accuracy=0.727233	lr=0.343382
Epoch[33] Batch [349]	Speed: 5024.423442 samples/sec	accuracy=0.727289	lr=0.343229
Epoch[33] Batch [399]	Speed: 4918.692643 samples/sec	accuracy=0.727500	lr=0.343077
Epoch[33] Batch [449]	Speed: 5006.603867 samples/sec	accuracy=0.727505	lr=0.342924
Epoch[33] Batch [499]	Speed: 5030.551311 samples/sec	accuracy=0.727635	lr=0.342771
Epoch[33] Batch [549]	Speed: 5049.360901 samples/sec	accuracy=0.727785	lr=0.342618
Epoch[33] Batch [599]	Speed: 4995.143446 samples/sec	accuracy=0.727943	lr=0.342465
Epoch[33] Batch [649]	Speed: 5019.988702 samples/sec	accuracy=0.728080	lr=0.342312
Epoch[33] Batch [699]	Speed: 5010.689821 samples/sec	accuracy=0.728287	lr=0.342158
Epoch[33] Batch [749]	Speed: 4980.461017 samples/sec	accuracy=0.728353	lr=0.342005
Epoch[33] Batch [799]	Speed: 5008.456714 samples/sec	accuracy=0.728422	lr=0.341851
Epoch[33] Batch [849]	Speed: 5025.730880 samples/sec	accuracy=0.728506	lr=0.341697
Epoch[33] Batch [899]	Speed: 5028.578461 samples/sec	accuracy=0.728618	lr=0.341543
Epoch[33] Batch [949]	Speed: 4989.740337 samples/sec	accuracy=0.728805	lr=0.341388
Epoch[33] Batch [999]	Speed: 5090.389769 samples/sec	accuracy=0.728946	lr=0.341234
Epoch[33] Batch [1049]	Speed: 4966.558670 samples/sec	accuracy=0.729080	lr=0.341079
Epoch[33] Batch [1099]	Speed: 5153.156451 samples/sec	accuracy=0.729195	lr=0.340924
Epoch[33] Batch [1149]	Speed: 5035.641470 samples/sec	accuracy=0.729190	lr=0.340769
Epoch[33] Batch [1199]	Speed: 5049.084168 samples/sec	accuracy=0.729333	lr=0.340614
Epoch[33] Batch [1249]	Speed: 5155.948765 samples/sec	accuracy=0.729441	lr=0.340458
[Epoch 33] training: accuracy=0.729438
[Epoch 33] speed: 4996 samples/sec	time cost: 265.700689
[Epoch 33] validation: err-top1=0.398936 err-top5=0.168288
Epoch[34] Batch [49]	Speed: 4770.626526 samples/sec	accuracy=0.728110	lr=0.340300
Epoch[34] Batch [99]	Speed: 5118.813995 samples/sec	accuracy=0.727796	lr=0.340144
Epoch[34] Batch [149]	Speed: 4918.847667 samples/sec	accuracy=0.727962	lr=0.339988
Epoch[34] Batch [199]	Speed: 4883.657427 samples/sec	accuracy=0.728012	lr=0.339832
Epoch[34] Batch [249]	Speed: 4697.982606 samples/sec	accuracy=0.728348	lr=0.339676
Epoch[34] Batch [299]	Speed: 4790.544208 samples/sec	accuracy=0.728432	lr=0.339520
Epoch[34] Batch [349]	Speed: 5096.597934 samples/sec	accuracy=0.728614	lr=0.339363
Epoch[34] Batch [399]	Speed: 5025.248463 samples/sec	accuracy=0.728804	lr=0.339206
Epoch[34] Batch [449]	Speed: 4985.735691 samples/sec	accuracy=0.728935	lr=0.339049
Epoch[34] Batch [499]	Speed: 4957.752580 samples/sec	accuracy=0.729124	lr=0.338892
Epoch[34] Batch [549]	Speed: 5061.926628 samples/sec	accuracy=0.729228	lr=0.338735
Epoch[34] Batch [599]	Speed: 4880.302479 samples/sec	accuracy=0.729292	lr=0.338578
Epoch[34] Batch [649]	Speed: 4995.079543 samples/sec	accuracy=0.729502	lr=0.338420
Epoch[34] Batch [699]	Speed: 4991.122238 samples/sec	accuracy=0.729631	lr=0.338263
Epoch[34] Batch [749]	Speed: 5002.690520 samples/sec	accuracy=0.729794	lr=0.338105
Epoch[34] Batch [799]	Speed: 4772.130316 samples/sec	accuracy=0.729925	lr=0.337947
Epoch[34] Batch [849]	Speed: 4921.768830 samples/sec	accuracy=0.730116	lr=0.337788
Epoch[34] Batch [899]	Speed: 4949.658987 samples/sec	accuracy=0.730208	lr=0.337630
Epoch[34] Batch [949]	Speed: 4904.214533 samples/sec	accuracy=0.730390	lr=0.337472
Epoch[34] Batch [999]	Speed: 5026.070227 samples/sec	accuracy=0.730446	lr=0.337313
Epoch[34] Batch [1049]	Speed: 4843.937188 samples/sec	accuracy=0.730538	lr=0.337154
Epoch[34] Batch [1099]	Speed: 4927.372722 samples/sec	accuracy=0.730567	lr=0.336995
Epoch[34] Batch [1149]	Speed: 4891.322411 samples/sec	accuracy=0.730661	lr=0.336836
Epoch[34] Batch [1199]	Speed: 4938.781294 samples/sec	accuracy=0.730724	lr=0.336677
Epoch[34] Batch [1249]	Speed: 5157.752652 samples/sec	accuracy=0.730814	lr=0.336517
[Epoch 34] training: accuracy=0.730817
[Epoch 34] speed: 4934 samples/sec	time cost: 268.387625
[Epoch 34] validation: err-top1=0.385335 err-top5=0.151937
Epoch[35] Batch [49]	Speed: 4647.266634 samples/sec	accuracy=0.730524	lr=0.336354
Epoch[35] Batch [99]	Speed: 5079.763775 samples/sec	accuracy=0.729407	lr=0.336194
Epoch[35] Batch [149]	Speed: 4675.114712 samples/sec	accuracy=0.728964	lr=0.336034
Epoch[35] Batch [199]	Speed: 4693.317808 samples/sec	accuracy=0.729200	lr=0.335874
Epoch[35] Batch [249]	Speed: 4881.565273 samples/sec	accuracy=0.729498	lr=0.335714
Epoch[35] Batch [299]	Speed: 5052.199910 samples/sec	accuracy=0.729681	lr=0.335553
Epoch[35] Batch [349]	Speed: 4804.384119 samples/sec	accuracy=0.729721	lr=0.335393
Epoch[35] Batch [399]	Speed: 4855.899323 samples/sec	accuracy=0.729932	lr=0.335232
Epoch[35] Batch [449]	Speed: 5082.143682 samples/sec	accuracy=0.730006	lr=0.335071
Epoch[35] Batch [499]	Speed: 5004.520761 samples/sec	accuracy=0.730122	lr=0.334910
Epoch[35] Batch [549]	Speed: 4895.274516 samples/sec	accuracy=0.730080	lr=0.334749
Epoch[35] Batch [599]	Speed: 4953.263665 samples/sec	accuracy=0.730219	lr=0.334587
Epoch[35] Batch [649]	Speed: 4956.032670 samples/sec	accuracy=0.730306	lr=0.334425
Epoch[35] Batch [699]	Speed: 4944.840826 samples/sec	accuracy=0.730446	lr=0.334264
Epoch[35] Batch [749]	Speed: 5097.336725 samples/sec	accuracy=0.730537	lr=0.334102
Epoch[35] Batch [799]	Speed: 4941.602238 samples/sec	accuracy=0.730715	lr=0.333940
Epoch[35] Batch [849]	Speed: 4910.320960 samples/sec	accuracy=0.730809	lr=0.333777
Epoch[35] Batch [899]	Speed: 4969.704239 samples/sec	accuracy=0.731005	lr=0.333615
Epoch[35] Batch [949]	Speed: 5082.394341 samples/sec	accuracy=0.731160	lr=0.333452
Epoch[35] Batch [999]	Speed: 5170.038251 samples/sec	accuracy=0.731269	lr=0.333290
Epoch[35] Batch [1049]	Speed: 4915.503731 samples/sec	accuracy=0.731333	lr=0.333127
Epoch[35] Batch [1099]	Speed: 5062.127207 samples/sec	accuracy=0.731536	lr=0.332964
Epoch[35] Batch [1149]	Speed: 4960.243485 samples/sec	accuracy=0.731591	lr=0.332801
Epoch[35] Batch [1199]	Speed: 5032.260015 samples/sec	accuracy=0.731645	lr=0.332637
Epoch[35] Batch [1249]	Speed: 5173.465631 samples/sec	accuracy=0.731839	lr=0.332474
[Epoch 35] training: accuracy=0.731847
[Epoch 35] speed: 4945 samples/sec	time cost: 269.109158
[Epoch 35] validation: err-top1=0.393614 err-top5=0.160096
Epoch[36] Batch [49]	Speed: 4474.433979 samples/sec	accuracy=0.730937	lr=0.332307
Epoch[36] Batch [99]	Speed: 5265.765769 samples/sec	accuracy=0.730298	lr=0.332143
Epoch[36] Batch [149]	Speed: 5086.832834 samples/sec	accuracy=0.730580	lr=0.331979
Epoch[36] Batch [199]	Speed: 5002.597056 samples/sec	accuracy=0.730408	lr=0.331815
Epoch[36] Batch [249]	Speed: 5087.654735 samples/sec	accuracy=0.730679	lr=0.331650
Epoch[36] Batch [299]	Speed: 5089.335156 samples/sec	accuracy=0.730821	lr=0.331486
Epoch[36] Batch [349]	Speed: 5125.248294 samples/sec	accuracy=0.731091	lr=0.331321
Epoch[36] Batch [399]	Speed: 4856.975289 samples/sec	accuracy=0.731305	lr=0.331157
Epoch[36] Batch [449]	Speed: 5012.945047 samples/sec	accuracy=0.731667	lr=0.330992
Epoch[36] Batch [499]	Speed: 5164.459771 samples/sec	accuracy=0.731859	lr=0.330827
Epoch[36] Batch [549]	Speed: 5071.502706 samples/sec	accuracy=0.732056	lr=0.330661
Epoch[36] Batch [599]	Speed: 5056.998513 samples/sec	accuracy=0.732140	lr=0.330496
Epoch[36] Batch [649]	Speed: 4969.872503 samples/sec	accuracy=0.732129	lr=0.330330
Epoch[36] Batch [699]	Speed: 4958.482003 samples/sec	accuracy=0.732300	lr=0.330165
Epoch[36] Batch [749]	Speed: 4986.311508 samples/sec	accuracy=0.732405	lr=0.329999
Epoch[36] Batch [799]	Speed: 4931.725902 samples/sec	accuracy=0.732485	lr=0.329833
Epoch[36] Batch [849]	Speed: 5086.831388 samples/sec	accuracy=0.732670	lr=0.329667
Epoch[36] Batch [899]	Speed: 4804.797218 samples/sec	accuracy=0.732778	lr=0.329500
Epoch[36] Batch [949]	Speed: 4960.252078 samples/sec	accuracy=0.732909	lr=0.329334
Epoch[36] Batch [999]	Speed: 4928.387286 samples/sec	accuracy=0.733030	lr=0.329167
Epoch[36] Batch [1049]	Speed: 5073.917075 samples/sec	accuracy=0.733189	lr=0.329000
Epoch[36] Batch [1099]	Speed: 4959.548707 samples/sec	accuracy=0.733272	lr=0.328833
Epoch[36] Batch [1149]	Speed: 5062.436042 samples/sec	accuracy=0.733356	lr=0.328666
Epoch[36] Batch [1199]	Speed: 5130.458944 samples/sec	accuracy=0.733416	lr=0.328499
Epoch[36] Batch [1249]	Speed: 5233.736422 samples/sec	accuracy=0.733527	lr=0.328332
[Epoch 36] training: accuracy=0.733525
[Epoch 36] speed: 5006 samples/sec	time cost: 265.243405
[Epoch 36] validation: err-top1=0.410236 err-top5=0.174087
Epoch[37] Batch [49]	Speed: 4852.454102 samples/sec	accuracy=0.733117	lr=0.328161
Epoch[37] Batch [99]	Speed: 4822.933325 samples/sec	accuracy=0.732815	lr=0.327993
Epoch[37] Batch [149]	Speed: 4863.495417 samples/sec	accuracy=0.733039	lr=0.327825
Epoch[37] Batch [199]	Speed: 4814.193791 samples/sec	accuracy=0.733419	lr=0.327657
Epoch[37] Batch [249]	Speed: 4903.056414 samples/sec	accuracy=0.733593	lr=0.327489
Epoch[37] Batch [299]	Speed: 4832.430289 samples/sec	accuracy=0.733730	lr=0.327321
Epoch[37] Batch [349]	Speed: 4797.968997 samples/sec	accuracy=0.733682	lr=0.327152
Epoch[37] Batch [399]	Speed: 4963.919481 samples/sec	accuracy=0.733444	lr=0.326983
Epoch[37] Batch [449]	Speed: 4991.746988 samples/sec	accuracy=0.733239	lr=0.326815
Epoch[37] Batch [499]	Speed: 4905.436732 samples/sec	accuracy=0.733120	lr=0.326646
Epoch[37] Batch [549]	Speed: 4932.585905 samples/sec	accuracy=0.733336	lr=0.326477
Epoch[37] Batch [599]	Speed: 4812.595862 samples/sec	accuracy=0.733392	lr=0.326307
Epoch[37] Batch [649]	Speed: 4902.064000 samples/sec	accuracy=0.733478	lr=0.326138
Epoch[37] Batch [699]	Speed: 5054.134483 samples/sec	accuracy=0.733561	lr=0.325968
Epoch[37] Batch [749]	Speed: 5004.206240 samples/sec	accuracy=0.733708	lr=0.325799
Epoch[37] Batch [799]	Speed: 5017.213784 samples/sec	accuracy=0.733798	lr=0.325629
Epoch[37] Batch [849]	Speed: 5000.341814 samples/sec	accuracy=0.733904	lr=0.325459
Epoch[37] Batch [899]	Speed: 5142.248975 samples/sec	accuracy=0.734050	lr=0.325289
Epoch[37] Batch [949]	Speed: 5258.687388 samples/sec	accuracy=0.734126	lr=0.325118
Epoch[37] Batch [999]	Speed: 5022.138616 samples/sec	accuracy=0.734187	lr=0.324948
Epoch[37] Batch [1049]	Speed: 4892.011243 samples/sec	accuracy=0.734303	lr=0.324777
Epoch[37] Batch [1099]	Speed: 5050.341168 samples/sec	accuracy=0.734446	lr=0.324607
Epoch[37] Batch [1149]	Speed: 5065.095877 samples/sec	accuracy=0.734540	lr=0.324436
Epoch[37] Batch [1199]	Speed: 5042.320591 samples/sec	accuracy=0.734623	lr=0.324265
Epoch[37] Batch [1249]	Speed: 5217.800925 samples/sec	accuracy=0.734773	lr=0.324094
[Epoch 37] training: accuracy=0.734775
[Epoch 37] speed: 4959 samples/sec	time cost: 267.595611
[Epoch 37] validation: err-top1=0.384088 err-top5=0.153659
Epoch[38] Batch [49]	Speed: 4946.131207 samples/sec	accuracy=0.734253	lr=0.323919
Epoch[38] Batch [99]	Speed: 5000.837045 samples/sec	accuracy=0.733857	lr=0.323747
Epoch[38] Batch [149]	Speed: 5025.293855 samples/sec	accuracy=0.734061	lr=0.323576
Epoch[38] Batch [199]	Speed: 5045.485244 samples/sec	accuracy=0.733855	lr=0.323404
Epoch[38] Batch [249]	Speed: 5100.009993 samples/sec	accuracy=0.733849	lr=0.323232
Epoch[38] Batch [299]	Speed: 4924.407416 samples/sec	accuracy=0.733968	lr=0.323060
Epoch[38] Batch [349]	Speed: 5147.702796 samples/sec	accuracy=0.734183	lr=0.322888
Epoch[38] Batch [399]	Speed: 4989.368899 samples/sec	accuracy=0.734027	lr=0.322715
Epoch[38] Batch [449]	Speed: 4989.233160 samples/sec	accuracy=0.734270	lr=0.322543
Epoch[38] Batch [499]	Speed: 5223.368373 samples/sec	accuracy=0.734430	lr=0.322370
Epoch[38] Batch [549]	Speed: 4967.396624 samples/sec	accuracy=0.734643	lr=0.322197
Epoch[38] Batch [599]	Speed: 5101.338894 samples/sec	accuracy=0.734722	lr=0.322025
Epoch[38] Batch [649]	Speed: 5128.806626 samples/sec	accuracy=0.734798	lr=0.321851
Epoch[38] Batch [699]	Speed: 5097.985446 samples/sec	accuracy=0.734962	lr=0.321678
Epoch[38] Batch [749]	Speed: 5085.952537 samples/sec	accuracy=0.735099	lr=0.321505
Epoch[38] Batch [799]	Speed: 5008.563480 samples/sec	accuracy=0.735257	lr=0.321331
Epoch[38] Batch [849]	Speed: 4866.176737 samples/sec	accuracy=0.735415	lr=0.321158
Epoch[38] Batch [899]	Speed: 5052.165085 samples/sec	accuracy=0.735493	lr=0.320984
Epoch[38] Batch [949]	Speed: 5106.069135 samples/sec	accuracy=0.735654	lr=0.320810
Epoch[38] Batch [999]	Speed: 5052.694055 samples/sec	accuracy=0.735825	lr=0.320636
Epoch[38] Batch [1049]	Speed: 5054.720379 samples/sec	accuracy=0.736018	lr=0.320462
Epoch[38] Batch [1099]	Speed: 5012.717807 samples/sec	accuracy=0.736167	lr=0.320287
Epoch[38] Batch [1149]	Speed: 5127.715711 samples/sec	accuracy=0.736253	lr=0.320113
Epoch[38] Batch [1199]	Speed: 5124.285444 samples/sec	accuracy=0.736323	lr=0.319938
Epoch[38] Batch [1249]	Speed: 5220.468437 samples/sec	accuracy=0.736375	lr=0.319763
[Epoch 38] training: accuracy=0.736378
[Epoch 38] speed: 5050 samples/sec	time cost: 262.778884
[Epoch 38] validation: err-top1=0.381577 err-top5=0.150530
Epoch[39] Batch [49]	Speed: 4991.084769 samples/sec	accuracy=0.735599	lr=0.319585
Epoch[39] Batch [99]	Speed: 5016.925795 samples/sec	accuracy=0.736176	lr=0.319410
Epoch[39] Batch [149]	Speed: 4801.009438 samples/sec	accuracy=0.735929	lr=0.319234
Epoch[39] Batch [199]	Speed: 4966.186311 samples/sec	accuracy=0.735882	lr=0.319059
Epoch[39] Batch [249]	Speed: 5009.881376 samples/sec	accuracy=0.735870	lr=0.318883
Epoch[39] Batch [299]	Speed: 5032.973429 samples/sec	accuracy=0.735749	lr=0.318708
Epoch[39] Batch [349]	Speed: 5006.741020 samples/sec	accuracy=0.735938	lr=0.318532
Epoch[39] Batch [399]	Speed: 5013.475667 samples/sec	accuracy=0.736146	lr=0.318356
Epoch[39] Batch [449]	Speed: 4914.265265 samples/sec	accuracy=0.736157	lr=0.318180
Epoch[39] Batch [499]	Speed: 5027.901020 samples/sec	accuracy=0.736185	lr=0.318004
Epoch[39] Batch [549]	Speed: 5060.759739 samples/sec	accuracy=0.736270	lr=0.317827
Epoch[39] Batch [599]	Speed: 5001.366619 samples/sec	accuracy=0.736374	lr=0.317651
Epoch[39] Batch [649]	Speed: 5133.199591 samples/sec	accuracy=0.736467	lr=0.317474
Epoch[39] Batch [699]	Speed: 4974.899554 samples/sec	accuracy=0.736575	lr=0.317297
Epoch[39] Batch [749]	Speed: 4995.786753 samples/sec	accuracy=0.736688	lr=0.317120
Epoch[39] Batch [799]	Speed: 5076.295705 samples/sec	accuracy=0.736755	lr=0.316943
Epoch[39] Batch [849]	Speed: 5168.048898 samples/sec	accuracy=0.736843	lr=0.316766
Epoch[39] Batch [899]	Speed: 5086.062753 samples/sec	accuracy=0.737009	lr=0.316589
Epoch[39] Batch [949]	Speed: 4916.675280 samples/sec	accuracy=0.737151	lr=0.316411
Epoch[39] Batch [999]	Speed: 5033.778370 samples/sec	accuracy=0.737271	lr=0.316233
Epoch[39] Batch [1049]	Speed: 4884.478193 samples/sec	accuracy=0.737293	lr=0.316056
Epoch[39] Batch [1099]	Speed: 4992.182144 samples/sec	accuracy=0.737384	lr=0.315878
Epoch[39] Batch [1149]	Speed: 4995.652756 samples/sec	accuracy=0.737444	lr=0.315700
Epoch[39] Batch [1199]	Speed: 4823.289602 samples/sec	accuracy=0.737574	lr=0.315522
Epoch[39] Batch [1249]	Speed: 5151.763219 samples/sec	accuracy=0.737683	lr=0.315343
[Epoch 39] training: accuracy=0.737683
[Epoch 39] speed: 4996 samples/sec	time cost: 265.451338
[Epoch 39] validation: err-top1=0.371273 err-top5=0.142160
Epoch[40] Batch [49]	Speed: 4610.965128 samples/sec	accuracy=0.736719	lr=0.315161
Epoch[40] Batch [99]	Speed: 4932.216244 samples/sec	accuracy=0.736836	lr=0.314983
Epoch[40] Batch [149]	Speed: 5068.652017 samples/sec	accuracy=0.737001	lr=0.314804
Epoch[40] Batch [199]	Speed: 4793.314495 samples/sec	accuracy=0.736778	lr=0.314625
Epoch[40] Batch [249]	Speed: 4904.736497 samples/sec	accuracy=0.736734	lr=0.314446
Epoch[40] Batch [299]	Speed: 5054.392022 samples/sec	accuracy=0.736953	lr=0.314267
Epoch[40] Batch [349]	Speed: 4954.416022 samples/sec	accuracy=0.737050	lr=0.314087
Epoch[40] Batch [399]	Speed: 4978.487549 samples/sec	accuracy=0.737010	lr=0.313908
Epoch[40] Batch [449]	Speed: 4908.148581 samples/sec	accuracy=0.736997	lr=0.313728
Epoch[40] Batch [499]	Speed: 5241.146557 samples/sec	accuracy=0.737072	lr=0.313549
Epoch[40] Batch [549]	Speed: 5181.688509 samples/sec	accuracy=0.737219	lr=0.313369
Epoch[40] Batch [599]	Speed: 4971.208320 samples/sec	accuracy=0.737242	lr=0.313189
Epoch[40] Batch [649]	Speed: 4996.616580 samples/sec	accuracy=0.737346	lr=0.313009
Epoch[40] Batch [699]	Speed: 5130.508830 samples/sec	accuracy=0.737552	lr=0.312829
Epoch[40] Batch [749]	Speed: 4906.674003 samples/sec	accuracy=0.737675	lr=0.312648
Epoch[40] Batch [799]	Speed: 5049.600975 samples/sec	accuracy=0.737727	lr=0.312468
Epoch[40] Batch [849]	Speed: 5004.098144 samples/sec	accuracy=0.737805	lr=0.312287
Epoch[40] Batch [899]	Speed: 4919.453102 samples/sec	accuracy=0.737892	lr=0.312106
Epoch[40] Batch [949]	Speed: 5012.988227 samples/sec	accuracy=0.738017	lr=0.311925
Epoch[40] Batch [999]	Speed: 4937.943996 samples/sec	accuracy=0.738213	lr=0.311744
Epoch[40] Batch [1049]	Speed: 5005.854034 samples/sec	accuracy=0.738423	lr=0.311563
Epoch[40] Batch [1099]	Speed: 4947.130946 samples/sec	accuracy=0.738586	lr=0.311382
Epoch[40] Batch [1149]	Speed: 5037.201448 samples/sec	accuracy=0.738659	lr=0.311201
Epoch[40] Batch [1199]	Speed: 4906.828719 samples/sec	accuracy=0.738740	lr=0.311019
Epoch[40] Batch [1249]	Speed: 5130.724198 samples/sec	accuracy=0.738798	lr=0.310837
[Epoch 40] training: accuracy=0.738793
[Epoch 40] speed: 4975 samples/sec	time cost: 266.423123
[Epoch 40] validation: err-top1=0.377502 err-top5=0.148132
Epoch[41] Batch [49]	Speed: 4598.538628 samples/sec	accuracy=0.737993	lr=0.310652
Epoch[41] Batch [99]	Speed: 4827.952708 samples/sec	accuracy=0.737369	lr=0.310470
Epoch[41] Batch [149]	Speed: 5055.574305 samples/sec	accuracy=0.737396	lr=0.310288
Epoch[41] Batch [199]	Speed: 5015.442069 samples/sec	accuracy=0.737676	lr=0.310105
Epoch[41] Batch [249]	Speed: 5032.109197 samples/sec	accuracy=0.737572	lr=0.309923
Epoch[41] Batch [299]	Speed: 5030.651007 samples/sec	accuracy=0.737578	lr=0.309741
Epoch[41] Batch [349]	Speed: 5001.471569 samples/sec	accuracy=0.737703	lr=0.309558
Epoch[41] Batch [399]	Speed: 5036.395293 samples/sec	accuracy=0.737836	lr=0.309375
Epoch[41] Batch [449]	Speed: 5144.647010 samples/sec	accuracy=0.738116	lr=0.309192
Epoch[41] Batch [499]	Speed: 5129.419642 samples/sec	accuracy=0.738181	lr=0.309009
Epoch[41] Batch [549]	Speed: 5095.469413 samples/sec	accuracy=0.738415	lr=0.308826
Epoch[41] Batch [599]	Speed: 5040.357311 samples/sec	accuracy=0.738716	lr=0.308643
Epoch[41] Batch [649]	Speed: 4983.134249 samples/sec	accuracy=0.738723	lr=0.308459
Epoch[41] Batch [699]	Speed: 5140.842559 samples/sec	accuracy=0.738772	lr=0.308276
Epoch[41] Batch [749]	Speed: 5035.040746 samples/sec	accuracy=0.738919	lr=0.308092
Epoch[41] Batch [799]	Speed: 4996.184138 samples/sec	accuracy=0.739045	lr=0.307908
Epoch[41] Batch [849]	Speed: 5099.863202 samples/sec	accuracy=0.739155	lr=0.307725
Epoch[41] Batch [899]	Speed: 5126.409503 samples/sec	accuracy=0.739179	lr=0.307540
Epoch[41] Batch [949]	Speed: 5075.802453 samples/sec	accuracy=0.739235	lr=0.307356
Epoch[41] Batch [999]	Speed: 4955.903084 samples/sec	accuracy=0.739271	lr=0.307172
Epoch[41] Batch [1049]	Speed: 5026.664930 samples/sec	accuracy=0.739373	lr=0.306988
Epoch[41] Batch [1099]	Speed: 5048.718561 samples/sec	accuracy=0.739500	lr=0.306803
Epoch[41] Batch [1149]	Speed: 5121.716273 samples/sec	accuracy=0.739589	lr=0.306618
Epoch[41] Batch [1199]	Speed: 5028.293523 samples/sec	accuracy=0.739733	lr=0.306433
Epoch[41] Batch [1249]	Speed: 5299.432422 samples/sec	accuracy=0.739829	lr=0.306249
[Epoch 41] training: accuracy=0.739838
[Epoch 41] speed: 5030 samples/sec	time cost: 263.984056
[Epoch 41] validation: err-top1=0.377611 err-top5=0.150291
Epoch[42] Batch [49]	Speed: 4807.254808 samples/sec	accuracy=0.738935	lr=0.306060
Epoch[42] Batch [99]	Speed: 4902.394014 samples/sec	accuracy=0.739544	lr=0.305875
Epoch[42] Batch [149]	Speed: 4993.450445 samples/sec	accuracy=0.739642	lr=0.305689
Epoch[42] Batch [199]	Speed: 4928.345438 samples/sec	accuracy=0.738793	lr=0.305504
Epoch[42] Batch [249]	Speed: 5024.408042 samples/sec	accuracy=0.738808	lr=0.305318
Epoch[42] Batch [299]	Speed: 5039.235228 samples/sec	accuracy=0.739291	lr=0.305133
Epoch[42] Batch [349]	Speed: 5006.315578 samples/sec	accuracy=0.739610	lr=0.304947
Epoch[42] Batch [399]	Speed: 5106.890100 samples/sec	accuracy=0.739548	lr=0.304761
Epoch[42] Batch [449]	Speed: 5049.806754 samples/sec	accuracy=0.739633	lr=0.304575
Epoch[42] Batch [499]	Speed: 5062.454063 samples/sec	accuracy=0.739740	lr=0.304388
Epoch[42] Batch [549]	Speed: 4965.763828 samples/sec	accuracy=0.739769	lr=0.304202
Epoch[42] Batch [599]	Speed: 5019.644310 samples/sec	accuracy=0.739860	lr=0.304016
Epoch[42] Batch [649]	Speed: 4889.185395 samples/sec	accuracy=0.740072	lr=0.303829
Epoch[42] Batch [699]	Speed: 5160.470462 samples/sec	accuracy=0.740184	lr=0.303642
Epoch[42] Batch [749]	Speed: 5026.726584 samples/sec	accuracy=0.740354	lr=0.303456
Epoch[42] Batch [799]	Speed: 5045.559335 samples/sec	accuracy=0.740597	lr=0.303269
Epoch[42] Batch [849]	Speed: 5142.639461 samples/sec	accuracy=0.740656	lr=0.303082
Epoch[42] Batch [899]	Speed: 5139.821681 samples/sec	accuracy=0.740762	lr=0.302894
Epoch[42] Batch [949]	Speed: 5021.331757 samples/sec	accuracy=0.740927	lr=0.302707
Epoch[42] Batch [999]	Speed: 5051.617213 samples/sec	accuracy=0.741013	lr=0.302520
Epoch[42] Batch [1049]	Speed: 4986.411543 samples/sec	accuracy=0.741168	lr=0.302332
Epoch[42] Batch [1099]	Speed: 5112.827688 samples/sec	accuracy=0.741271	lr=0.302144
Epoch[42] Batch [1149]	Speed: 5136.069811 samples/sec	accuracy=0.741378	lr=0.301957
Epoch[42] Batch [1199]	Speed: 5178.311412 samples/sec	accuracy=0.741534	lr=0.301769
Epoch[42] Batch [1249]	Speed: 5155.199694 samples/sec	accuracy=0.741651	lr=0.301581
[Epoch 42] training: accuracy=0.741652
[Epoch 42] speed: 5032 samples/sec	time cost: 264.014470
[Epoch 42] validation: err-top1=0.362883 err-top5=0.140864
Epoch[43] Batch [49]	Speed: 4895.825943 samples/sec	accuracy=0.741350	lr=0.301389
Epoch[43] Batch [99]	Speed: 5021.898798 samples/sec	accuracy=0.741304	lr=0.301200
Epoch[43] Batch [149]	Speed: 5035.691891 samples/sec	accuracy=0.741028	lr=0.301012
Epoch[43] Batch [199]	Speed: 5026.679166 samples/sec	accuracy=0.740772	lr=0.300823
Epoch[43] Batch [249]	Speed: 4968.005794 samples/sec	accuracy=0.740761	lr=0.300635
Epoch[43] Batch [299]	Speed: 5012.487662 samples/sec	accuracy=0.740859	lr=0.300446
Epoch[43] Batch [349]	Speed: 5002.235820 samples/sec	accuracy=0.741011	lr=0.300257
Epoch[43] Batch [399]	Speed: 4980.198829 samples/sec	accuracy=0.740980	lr=0.300068
Epoch[43] Batch [449]	Speed: 5094.665774 samples/sec	accuracy=0.741171	lr=0.299879
Epoch[43] Batch [499]	Speed: 5129.837469 samples/sec	accuracy=0.741374	lr=0.299690
Epoch[43] Batch [549]	Speed: 5150.651029 samples/sec	accuracy=0.741664	lr=0.299500
Epoch[43] Batch [599]	Speed: 5074.917934 samples/sec	accuracy=0.741784	lr=0.299311
Epoch[43] Batch [649]	Speed: 5052.251258 samples/sec	accuracy=0.741781	lr=0.299121
Epoch[43] Batch [699]	Speed: 5158.594039 samples/sec	accuracy=0.741838	lr=0.298932
Epoch[43] Batch [749]	Speed: 5063.291026 samples/sec	accuracy=0.741940	lr=0.298742
Epoch[43] Batch [799]	Speed: 5010.937456 samples/sec	accuracy=0.742166	lr=0.298552
Epoch[43] Batch [849]	Speed: 5041.002969 samples/sec	accuracy=0.742302	lr=0.298362
Epoch[43] Batch [899]	Speed: 5149.936101 samples/sec	accuracy=0.742391	lr=0.298171
Epoch[43] Batch [949]	Speed: 4980.500174 samples/sec	accuracy=0.742429	lr=0.297981
Epoch[43] Batch [999]	Speed: 5082.234970 samples/sec	accuracy=0.742581	lr=0.297791
Epoch[43] Batch [1049]	Speed: 5042.038473 samples/sec	accuracy=0.742664	lr=0.297600
Epoch[43] Batch [1099]	Speed: 5052.994250 samples/sec	accuracy=0.742732	lr=0.297409
Epoch[43] Batch [1149]	Speed: 5032.266501 samples/sec	accuracy=0.742796	lr=0.297219
Epoch[43] Batch [1199]	Speed: 5034.141699 samples/sec	accuracy=0.742859	lr=0.297028
Epoch[43] Batch [1249]	Speed: 5095.877253 samples/sec	accuracy=0.742922	lr=0.296837
[Epoch 43] training: accuracy=0.742921
[Epoch 43] speed: 5042 samples/sec	time cost: 263.361898
[Epoch 43] validation: err-top1=0.372130 err-top5=0.145089
Epoch[44] Batch [49]	Speed: 4812.372941 samples/sec	accuracy=0.741299	lr=0.296642
Epoch[44] Batch [99]	Speed: 5078.934331 samples/sec	accuracy=0.740670	lr=0.296451
Epoch[44] Batch [149]	Speed: 5090.906015 samples/sec	accuracy=0.741122	lr=0.296259
Epoch[44] Batch [199]	Speed: 4979.768647 samples/sec	accuracy=0.741731	lr=0.296068
Epoch[44] Batch [249]	Speed: 5122.625858 samples/sec	accuracy=0.741691	lr=0.295876
Epoch[44] Batch [299]	Speed: 5060.426543 samples/sec	accuracy=0.741653	lr=0.295684
Epoch[44] Batch [349]	Speed: 4819.192286 samples/sec	accuracy=0.741624	lr=0.295493
Epoch[44] Batch [399]	Speed: 5096.893207 samples/sec	accuracy=0.741706	lr=0.295301
Epoch[44] Batch [449]	Speed: 5005.661622 samples/sec	accuracy=0.741896	lr=0.295109
Epoch[44] Batch [499]	Speed: 5046.336645 samples/sec	accuracy=0.742068	lr=0.294917
Epoch[44] Batch [549]	Speed: 5008.565465 samples/sec	accuracy=0.742226	lr=0.294724
Epoch[44] Batch [599]	Speed: 5046.090005 samples/sec	accuracy=0.742362	lr=0.294532
Epoch[44] Batch [649]	Speed: 5005.195416 samples/sec	accuracy=0.742500	lr=0.294339
Epoch[44] Batch [699]	Speed: 5230.551427 samples/sec	accuracy=0.742710	lr=0.294147
Epoch[44] Batch [749]	Speed: 5206.110194 samples/sec	accuracy=0.742814	lr=0.293954
Epoch[44] Batch [799]	Speed: 5067.672761 samples/sec	accuracy=0.742907	lr=0.293761
Epoch[44] Batch [849]	Speed: 5195.694569 samples/sec	accuracy=0.742926	lr=0.293568
Epoch[44] Batch [899]	Speed: 4911.095792 samples/sec	accuracy=0.743049	lr=0.293375
Epoch[44] Batch [949]	Speed: 5107.974843 samples/sec	accuracy=0.743092	lr=0.293182
Epoch[44] Batch [999]	Speed: 4957.970973 samples/sec	accuracy=0.743161	lr=0.292989
Epoch[44] Batch [1049]	Speed: 4958.610579 samples/sec	accuracy=0.743176	lr=0.292795
Epoch[44] Batch [1099]	Speed: 5143.137906 samples/sec	accuracy=0.743214	lr=0.292602
Epoch[44] Batch [1149]	Speed: 5044.863207 samples/sec	accuracy=0.743357	lr=0.292408
Epoch[44] Batch [1199]	Speed: 5043.429716 samples/sec	accuracy=0.743514	lr=0.292215
Epoch[44] Batch [1249]	Speed: 5189.310504 samples/sec	accuracy=0.743655	lr=0.292021
[Epoch 44] training: accuracy=0.743653
[Epoch 44] speed: 5042 samples/sec	time cost: 263.439366
[Epoch 44] validation: err-top1=0.355748 err-top5=0.134367
Epoch[45] Batch [49]	Speed: 4920.989837 samples/sec	accuracy=0.744030	lr=0.291823
Epoch[45] Batch [99]	Speed: 4993.046297 samples/sec	accuracy=0.743354	lr=0.291629
Epoch[45] Batch [149]	Speed: 4894.056483 samples/sec	accuracy=0.743279	lr=0.291435
Epoch[45] Batch [199]	Speed: 4994.645857 samples/sec	accuracy=0.743306	lr=0.291240
Epoch[45] Batch [249]	Speed: 5072.598108 samples/sec	accuracy=0.743235	lr=0.291046
Epoch[45] Batch [299]	Speed: 5025.114292 samples/sec	accuracy=0.743525	lr=0.290852
Epoch[45] Batch [349]	Speed: 5062.459553 samples/sec	accuracy=0.743646	lr=0.290657
Epoch[45] Batch [399]	Speed: 5096.029862 samples/sec	accuracy=0.743774	lr=0.290462
Epoch[45] Batch [449]	Speed: 5078.957875 samples/sec	accuracy=0.744031	lr=0.290267
Epoch[45] Batch [499]	Speed: 4892.710078 samples/sec	accuracy=0.744169	lr=0.290073
Epoch[45] Batch [549]	Speed: 5052.693104 samples/sec	accuracy=0.744220	lr=0.289878
Epoch[45] Batch [599]	Speed: 5178.120748 samples/sec	accuracy=0.744330	lr=0.289682
Epoch[45] Batch [649]	Speed: 4955.381264 samples/sec	accuracy=0.744594	lr=0.289487
Epoch[45] Batch [699]	Speed: 5112.654109 samples/sec	accuracy=0.744658	lr=0.289292
Epoch[45] Batch [749]	Speed: 5054.552745 samples/sec	accuracy=0.744669	lr=0.289096
Epoch[45] Batch [799]	Speed: 5088.788358 samples/sec	accuracy=0.744802	lr=0.288901
Epoch[45] Batch [849]	Speed: 5033.861085 samples/sec	accuracy=0.744850	lr=0.288705
Epoch[45] Batch [899]	Speed: 5017.823510 samples/sec	accuracy=0.744959	lr=0.288509
Epoch[45] Batch [949]	Speed: 5048.852571 samples/sec	accuracy=0.745086	lr=0.288314
Epoch[45] Batch [999]	Speed: 5059.192527 samples/sec	accuracy=0.745154	lr=0.288118
Epoch[45] Batch [1049]	Speed: 4984.072889 samples/sec	accuracy=0.745181	lr=0.287921
Epoch[45] Batch [1099]	Speed: 5021.982884 samples/sec	accuracy=0.745217	lr=0.287725
Epoch[45] Batch [1149]	Speed: 5080.711888 samples/sec	accuracy=0.745316	lr=0.287529
Epoch[45] Batch [1199]	Speed: 5240.098373 samples/sec	accuracy=0.745346	lr=0.287333
Epoch[45] Batch [1249]	Speed: 5008.697703 samples/sec	accuracy=0.745473	lr=0.287136
[Epoch 45] training: accuracy=0.745468
[Epoch 45] speed: 5033 samples/sec	time cost: 263.717357
[Epoch 45] validation: err-top1=0.357402 err-top5=0.137357
Epoch[46] Batch [49]	Speed: 4832.358303 samples/sec	accuracy=0.744578	lr=0.286936
Epoch[46] Batch [99]	Speed: 5030.436064 samples/sec	accuracy=0.744114	lr=0.286739
Epoch[46] Batch [149]	Speed: 4984.524178 samples/sec	accuracy=0.744450	lr=0.286542
Epoch[46] Batch [199]	Speed: 5029.519930 samples/sec	accuracy=0.744612	lr=0.286345
Epoch[46] Batch [249]	Speed: 5098.655516 samples/sec	accuracy=0.744965	lr=0.286148
Epoch[46] Batch [299]	Speed: 5109.608657 samples/sec	accuracy=0.744794	lr=0.285951
Epoch[46] Batch [349]	Speed: 4997.277942 samples/sec	accuracy=0.745028	lr=0.285754
Epoch[46] Batch [399]	Speed: 5212.610336 samples/sec	accuracy=0.744965	lr=0.285556
Epoch[46] Batch [449]	Speed: 5171.773543 samples/sec	accuracy=0.745111	lr=0.285359
Epoch[46] Batch [499]	Speed: 5043.450563 samples/sec	accuracy=0.745147	lr=0.285161
Epoch[46] Batch [549]	Speed: 5001.641175 samples/sec	accuracy=0.745254	lr=0.284964
Epoch[46] Batch [599]	Speed: 4986.399849 samples/sec	accuracy=0.745471	lr=0.284766
Epoch[46] Batch [649]	Speed: 5042.911092 samples/sec	accuracy=0.745589	lr=0.284568
Epoch[46] Batch [699]	Speed: 5035.175448 samples/sec	accuracy=0.745659	lr=0.284370
Epoch[46] Batch [749]	Speed: 4993.653995 samples/sec	accuracy=0.745716	lr=0.284172
Epoch[46] Batch [799]	Speed: 4979.157974 samples/sec	accuracy=0.745786	lr=0.283974
Epoch[46] Batch [849]	Speed: 4984.439721 samples/sec	accuracy=0.745889	lr=0.283776
Epoch[46] Batch [899]	Speed: 5065.582510 samples/sec	accuracy=0.746007	lr=0.283577
Epoch[46] Batch [949]	Speed: 5031.161808 samples/sec	accuracy=0.746109	lr=0.283379
Epoch[46] Batch [999]	Speed: 5086.942968 samples/sec	accuracy=0.746164	lr=0.283181
Epoch[46] Batch [1049]	Speed: 5045.416609 samples/sec	accuracy=0.746185	lr=0.282982
Epoch[46] Batch [1099]	Speed: 5016.689287 samples/sec	accuracy=0.746219	lr=0.282783
Epoch[46] Batch [1149]	Speed: 4969.277363 samples/sec	accuracy=0.746279	lr=0.282584
Epoch[46] Batch [1199]	Speed: 5077.893584 samples/sec	accuracy=0.746375	lr=0.282385
Epoch[46] Batch [1249]	Speed: 5192.037455 samples/sec	accuracy=0.746487	lr=0.282186
[Epoch 46] training: accuracy=0.746494
[Epoch 46] speed: 5035 samples/sec	time cost: 263.583653
[Epoch 46] validation: err-top1=0.369303 err-top5=0.141744
Epoch[47] Batch [49]	Speed: 4865.233034 samples/sec	accuracy=0.745362	lr=0.281983
Epoch[47] Batch [99]	Speed: 4920.790138 samples/sec	accuracy=0.744869	lr=0.281784
Epoch[47] Batch [149]	Speed: 4983.632323 samples/sec	accuracy=0.744929	lr=0.281585
Epoch[47] Batch [199]	Speed: 4994.214221 samples/sec	accuracy=0.744965	lr=0.281385
Epoch[47] Batch [249]	Speed: 5079.376171 samples/sec	accuracy=0.745217	lr=0.281186
Epoch[47] Batch [299]	Speed: 5118.832908 samples/sec	accuracy=0.745347	lr=0.280986
Epoch[47] Batch [349]	Speed: 5130.724198 samples/sec	accuracy=0.745607	lr=0.280786
Epoch[47] Batch [399]	Speed: 5006.506055 samples/sec	accuracy=0.745617	lr=0.280587
Epoch[47] Batch [449]	Speed: 5083.468581 samples/sec	accuracy=0.745709	lr=0.280387
Epoch[47] Batch [499]	Speed: 5030.066436 samples/sec	accuracy=0.745795	lr=0.280187
Epoch[47] Batch [549]	Speed: 4806.989557 samples/sec	accuracy=0.745893	lr=0.279987
Epoch[47] Batch [599]	Speed: 5049.187568 samples/sec	accuracy=0.746101	lr=0.279786
Epoch[47] Batch [649]	Speed: 5032.281831 samples/sec	accuracy=0.746124	lr=0.279586
Epoch[47] Batch [699]	Speed: 5098.570174 samples/sec	accuracy=0.746163	lr=0.279386
Epoch[47] Batch [749]	Speed: 5101.363130 samples/sec	accuracy=0.746373	lr=0.279185
Epoch[47] Batch [799]	Speed: 5300.833668 samples/sec	accuracy=0.746622	lr=0.278985
Epoch[47] Batch [849]	Speed: 4905.074377 samples/sec	accuracy=0.746623	lr=0.278784
Epoch[47] Batch [899]	Speed: 4988.682511 samples/sec	accuracy=0.746755	lr=0.278583
Epoch[47] Batch [949]	Speed: 5014.256353 samples/sec	accuracy=0.746898	lr=0.278382
Epoch[47] Batch [999]	Speed: 4919.929622 samples/sec	accuracy=0.747026	lr=0.278181
Epoch[47] Batch [1049]	Speed: 4984.817252 samples/sec	accuracy=0.747099	lr=0.277980
Epoch[47] Batch [1099]	Speed: 4974.120014 samples/sec	accuracy=0.747240	lr=0.277779
Epoch[47] Batch [1149]	Speed: 5032.054249 samples/sec	accuracy=0.747294	lr=0.277578
Epoch[47] Batch [1199]	Speed: 5035.405674 samples/sec	accuracy=0.747407	lr=0.277377
Epoch[47] Batch [1249]	Speed: 5219.613217 samples/sec	accuracy=0.747492	lr=0.277175
[Epoch 47] training: accuracy=0.747493
[Epoch 47] speed: 5020 samples/sec	time cost: 264.638995
[Epoch 47] validation: err-top1=0.358339 err-top5=0.135862
Epoch[48] Batch [49]	Speed: 4959.389617 samples/sec	accuracy=0.747632	lr=0.276970
Epoch[48] Batch [99]	Speed: 4905.819761 samples/sec	accuracy=0.747114	lr=0.276768
Epoch[48] Batch [149]	Speed: 4821.939080 samples/sec	accuracy=0.746856	lr=0.276566
Epoch[48] Batch [199]	Speed: 4948.799750 samples/sec	accuracy=0.746967	lr=0.276365
Epoch[48] Batch [249]	Speed: 5132.304769 samples/sec	accuracy=0.746747	lr=0.276163
Epoch[48] Batch [299]	Speed: 5115.331287 samples/sec	accuracy=0.746802	lr=0.275961
Epoch[48] Batch [349]	Speed: 4951.587512 samples/sec	accuracy=0.746772	lr=0.275759
Epoch[48] Batch [399]	Speed: 5153.289756 samples/sec	accuracy=0.747015	lr=0.275557
Epoch[48] Batch [449]	Speed: 4964.846074 samples/sec	accuracy=0.747188	lr=0.275354
Epoch[48] Batch [499]	Speed: 4880.708547 samples/sec	accuracy=0.747484	lr=0.275152
Epoch[48] Batch [549]	Speed: 4958.159735 samples/sec	accuracy=0.747661	lr=0.274950
Epoch[48] Batch [599]	Speed: 4938.059245 samples/sec	accuracy=0.747913	lr=0.274747
Epoch[48] Batch [649]	Speed: 4969.375796 samples/sec	accuracy=0.748015	lr=0.274545
Epoch[48] Batch [699]	Speed: 5016.119673 samples/sec	accuracy=0.748122	lr=0.274342
Epoch[48] Batch [749]	Speed: 5085.639381 samples/sec	accuracy=0.748224	lr=0.274139
Epoch[48] Batch [799]	Speed: 4969.309675 samples/sec	accuracy=0.748387	lr=0.273936
Epoch[48] Batch [849]	Speed: 5007.231098 samples/sec	accuracy=0.748596	lr=0.273733
Epoch[48] Batch [899]	Speed: 5055.773548 samples/sec	accuracy=0.748659	lr=0.273530
Epoch[48] Batch [949]	Speed: 4925.794600 samples/sec	accuracy=0.748827	lr=0.273327
Epoch[48] Batch [999]	Speed: 4965.753723 samples/sec	accuracy=0.748987	lr=0.273124
Epoch[48] Batch [1049]	Speed: 5104.793831 samples/sec	accuracy=0.749059	lr=0.272921
Epoch[48] Batch [1099]	Speed: 4875.634165 samples/sec	accuracy=0.749118	lr=0.272717
Epoch[48] Batch [1149]	Speed: 5014.480805 samples/sec	accuracy=0.749222	lr=0.272514
Epoch[48] Batch [1199]	Speed: 4960.618735 samples/sec	accuracy=0.749346	lr=0.272310
Epoch[48] Batch [1249]	Speed: 5067.210835 samples/sec	accuracy=0.749391	lr=0.272107
[Epoch 48] training: accuracy=0.749398
[Epoch 48] speed: 4984 samples/sec	time cost: 266.118299
[Epoch 48] validation: err-top1=0.350287 err-top5=0.130700
Epoch[49] Batch [49]	Speed: 4891.729870 samples/sec	accuracy=0.749305	lr=0.271899
Epoch[49] Batch [99]	Speed: 4934.820800 samples/sec	accuracy=0.749140	lr=0.271695
Epoch[49] Batch [149]	Speed: 5084.524857 samples/sec	accuracy=0.749275	lr=0.271491
Epoch[49] Batch [199]	Speed: 5105.269796 samples/sec	accuracy=0.749102	lr=0.271287
Epoch[49] Batch [249]	Speed: 4888.582379 samples/sec	accuracy=0.748717	lr=0.271083
Epoch[49] Batch [299]	Speed: 5026.684579 samples/sec	accuracy=0.748747	lr=0.270879
Epoch[49] Batch [349]	Speed: 5097.273205 samples/sec	accuracy=0.748834	lr=0.270675
Epoch[49] Batch [399]	Speed: 5061.781900 samples/sec	accuracy=0.748873	lr=0.270470
Epoch[49] Batch [449]	Speed: 5102.118092 samples/sec	accuracy=0.749012	lr=0.270266
Epoch[49] Batch [499]	Speed: 5023.847134 samples/sec	accuracy=0.748950	lr=0.270061
Epoch[49] Batch [549]	Speed: 5027.010875 samples/sec	accuracy=0.748916	lr=0.269857
Epoch[49] Batch [599]	Speed: 5060.909894 samples/sec	accuracy=0.749115	lr=0.269652
Epoch[49] Batch [649]	Speed: 4997.715923 samples/sec	accuracy=0.749192	lr=0.269447
Epoch[49] Batch [699]	Speed: 4994.544098 samples/sec	accuracy=0.749318	lr=0.269243
Epoch[49] Batch [749]	Speed: 4974.016093 samples/sec	accuracy=0.749478	lr=0.269038
Epoch[49] Batch [799]	Speed: 4813.755228 samples/sec	accuracy=0.749555	lr=0.268833
Epoch[49] Batch [849]	Speed: 5053.416128 samples/sec	accuracy=0.749626	lr=0.268628
Epoch[49] Batch [899]	Speed: 5260.396246 samples/sec	accuracy=0.749672	lr=0.268422
Epoch[49] Batch [949]	Speed: 5049.307238 samples/sec	accuracy=0.749797	lr=0.268217
Epoch[49] Batch [999]	Speed: 5000.798965 samples/sec	accuracy=0.749833	lr=0.268012
Epoch[49] Batch [1049]	Speed: 5092.030462 samples/sec	accuracy=0.749984	lr=0.267807
Epoch[49] Batch [1099]	Speed: 4999.991730 samples/sec	accuracy=0.750038	lr=0.267601
Epoch[49] Batch [1149]	Speed: 5105.183261 samples/sec	accuracy=0.750032	lr=0.267396
Epoch[49] Batch [1199]	Speed: 4927.914329 samples/sec	accuracy=0.750106	lr=0.267190
Epoch[49] Batch [1249]	Speed: 5247.734667 samples/sec	accuracy=0.750219	lr=0.266984
[Epoch 49] training: accuracy=0.750222
[Epoch 49] speed: 5026 samples/sec	time cost: 264.046391
[Epoch 49] validation: err-top1=0.361169 err-top5=0.136021
Epoch[50] Batch [49]	Speed: 4860.162164 samples/sec	accuracy=0.751131	lr=0.266774
Epoch[50] Batch [99]	Speed: 4929.216595 samples/sec	accuracy=0.748923	lr=0.266568
Epoch[50] Batch [149]	Speed: 5050.643578 samples/sec	accuracy=0.748623	lr=0.266362
Epoch[50] Batch [199]	Speed: 4951.559768 samples/sec	accuracy=0.748869	lr=0.266156
Epoch[50] Batch [249]	Speed: 5064.310388 samples/sec	accuracy=0.748652	lr=0.265950
Epoch[50] Batch [299]	Speed: 5053.145371 samples/sec	accuracy=0.748673	lr=0.265744
Epoch[50] Batch [349]	Speed: 4995.272536 samples/sec	accuracy=0.748755	lr=0.265538
Epoch[50] Batch [399]	Speed: 5065.891169 samples/sec	accuracy=0.749047	lr=0.265331
Epoch[50] Batch [449]	Speed: 5037.281203 samples/sec	accuracy=0.749473	lr=0.265125
Epoch[50] Batch [499]	Speed: 5015.395918 samples/sec	accuracy=0.749652	lr=0.264918
Epoch[50] Batch [549]	Speed: 4988.841400 samples/sec	accuracy=0.749814	lr=0.264712
Epoch[50] Batch [599]	Speed: 5033.366135 samples/sec	accuracy=0.749792	lr=0.264505
Epoch[50] Batch [649]	Speed: 4984.490626 samples/sec	accuracy=0.749944	lr=0.264298
Epoch[50] Batch [699]	Speed: 5085.331200 samples/sec	accuracy=0.750059	lr=0.264092
Epoch[50] Batch [749]	Speed: 5130.018467 samples/sec	accuracy=0.750146	lr=0.263885
Epoch[50] Batch [799]	Speed: 5045.267016 samples/sec	accuracy=0.750215	lr=0.263678
Epoch[50] Batch [849]	Speed: 4912.285017 samples/sec	accuracy=0.750302	lr=0.263471
Epoch[50] Batch [899]	Speed: 4943.183339 samples/sec	accuracy=0.750364	lr=0.263264
Epoch[50] Batch [949]	Speed: 5000.645136 samples/sec	accuracy=0.750477	lr=0.263056
Epoch[50] Batch [999]	Speed: 5016.831916 samples/sec	accuracy=0.750614	lr=0.262849
Epoch[50] Batch [1049]	Speed: 5094.980527 samples/sec	accuracy=0.750717	lr=0.262642
Epoch[50] Batch [1099]	Speed: 5115.876614 samples/sec	accuracy=0.750848	lr=0.262434
Epoch[50] Batch [1149]	Speed: 4934.217812 samples/sec	accuracy=0.750983	lr=0.262227
Epoch[50] Batch [1199]	Speed: 5009.609655 samples/sec	accuracy=0.751110	lr=0.262019
Epoch[50] Batch [1249]	Speed: 5093.016618 samples/sec	accuracy=0.751173	lr=0.261812
[Epoch 50] training: accuracy=0.751179
[Epoch 50] speed: 5011 samples/sec	time cost: 264.972748
[Epoch 50] validation: err-top1=0.345285 err-top5=0.129464
Epoch[51] Batch [49]	Speed: 4880.910997 samples/sec	accuracy=0.750015	lr=0.261600
Epoch[51] Batch [99]	Speed: 4899.201829 samples/sec	accuracy=0.749888	lr=0.261392
Epoch[51] Batch [149]	Speed: 4989.260516 samples/sec	accuracy=0.750388	lr=0.261184
Epoch[51] Batch [199]	Speed: 5063.901615 samples/sec	accuracy=0.750237	lr=0.260976
Epoch[51] Batch [249]	Speed: 4826.993716 samples/sec	accuracy=0.750316	lr=0.260768
Epoch[51] Batch [299]	Speed: 4883.020244 samples/sec	accuracy=0.750656	lr=0.260560
Epoch[51] Batch [349]	Speed: 4928.568034 samples/sec	accuracy=0.750616	lr=0.260352
Epoch[51] Batch [399]	Speed: 4929.949981 samples/sec	accuracy=0.750813	lr=0.260144
Epoch[51] Batch [449]	Speed: 5061.116476 samples/sec	accuracy=0.750747	lr=0.259935
Epoch[51] Batch [499]	Speed: 5026.526212 samples/sec	accuracy=0.750802	lr=0.259727
Epoch[51] Batch [549]	Speed: 5022.664253 samples/sec	accuracy=0.750935	lr=0.259519
Epoch[51] Batch [599]	Speed: 5009.475499 samples/sec	accuracy=0.751158	lr=0.259310
Epoch[51] Batch [649]	Speed: 5066.205484 samples/sec	accuracy=0.751171	lr=0.259102
Epoch[51] Batch [699]	Speed: 5077.981718 samples/sec	accuracy=0.751261	lr=0.258893
Epoch[51] Batch [749]	Speed: 5002.230460 samples/sec	accuracy=0.751343	lr=0.258684
Epoch[51] Batch [799]	Speed: 5112.621975 samples/sec	accuracy=0.751375	lr=0.258475
Epoch[51] Batch [849]	Speed: 5053.652069 samples/sec	accuracy=0.751467	lr=0.258267
Epoch[51] Batch [899]	Speed: 4738.737767 samples/sec	accuracy=0.751550	lr=0.258058
Epoch[51] Batch [949]	Speed: 5001.160926 samples/sec	accuracy=0.751721	lr=0.257849
Epoch[51] Batch [999]	Speed: 5085.161048 samples/sec	accuracy=0.751868	lr=0.257640
Epoch[51] Batch [1049]	Speed: 5040.704670 samples/sec	accuracy=0.751912	lr=0.257430
Epoch[51] Batch [1099]	Speed: 4944.464544 samples/sec	accuracy=0.752016	lr=0.257221
Epoch[51] Batch [1149]	Speed: 4989.644111 samples/sec	accuracy=0.752116	lr=0.257012
Epoch[51] Batch [1199]	Speed: 4949.596242 samples/sec	accuracy=0.752244	lr=0.256803
Epoch[51] Batch [1249]	Speed: 5221.869751 samples/sec	accuracy=0.752360	lr=0.256593
[Epoch 51] training: accuracy=0.752362
[Epoch 51] speed: 4986 samples/sec	time cost: 266.207991
[Epoch 51] validation: err-top1=0.349510 err-top5=0.129803
Epoch[52] Batch [49]	Speed: 4794.754153 samples/sec	accuracy=0.752594	lr=0.256379
Epoch[52] Batch [99]	Speed: 5049.502901 samples/sec	accuracy=0.752481	lr=0.256170
Epoch[52] Batch [149]	Speed: 5160.038827 samples/sec	accuracy=0.752233	lr=0.255960
Epoch[52] Batch [199]	Speed: 5091.405105 samples/sec	accuracy=0.752010	lr=0.255751
Epoch[52] Batch [249]	Speed: 4897.653761 samples/sec	accuracy=0.751808	lr=0.255541
Epoch[52] Batch [299]	Speed: 5109.312882 samples/sec	accuracy=0.751967	lr=0.255331
Epoch[52] Batch [349]	Speed: 5011.324859 samples/sec	accuracy=0.752002	lr=0.255121
Epoch[52] Batch [399]	Speed: 4993.542058 samples/sec	accuracy=0.752242	lr=0.254911
Epoch[52] Batch [449]	Speed: 5076.706842 samples/sec	accuracy=0.752206	lr=0.254701
Epoch[52] Batch [499]	Speed: 5074.789732 samples/sec	accuracy=0.752410	lr=0.254491
Epoch[52] Batch [549]	Speed: 5044.520370 samples/sec	accuracy=0.752551	lr=0.254281
Epoch[52] Batch [599]	Speed: 5065.491222 samples/sec	accuracy=0.752663	lr=0.254071
Epoch[52] Batch [649]	Speed: 5058.541249 samples/sec	accuracy=0.752860	lr=0.253861
Epoch[52] Batch [699]	Speed: 5139.897584 samples/sec	accuracy=0.752980	lr=0.253650
Epoch[52] Batch [749]	Speed: 5096.762320 samples/sec	accuracy=0.753173	lr=0.253440
Epoch[52] Batch [799]	Speed: 5045.649787 samples/sec	accuracy=0.753292	lr=0.253229
Epoch[52] Batch [849]	Speed: 4984.906466 samples/sec	accuracy=0.753398	lr=0.253019
Epoch[52] Batch [899]	Speed: 4964.405113 samples/sec	accuracy=0.753533	lr=0.252808
Epoch[52] Batch [949]	Speed: 4915.058329 samples/sec	accuracy=0.753584	lr=0.252598
Epoch[52] Batch [999]	Speed: 4978.084319 samples/sec	accuracy=0.753724	lr=0.252387
Epoch[52] Batch [1049]	Speed: 4925.795165 samples/sec	accuracy=0.753858	lr=0.252176
Epoch[52] Batch [1099]	Speed: 5012.809075 samples/sec	accuracy=0.753957	lr=0.251965
Epoch[52] Batch [1149]	Speed: 4943.292574 samples/sec	accuracy=0.754083	lr=0.251754
Epoch[52] Batch [1199]	Speed: 4905.486932 samples/sec	accuracy=0.754165	lr=0.251543
Epoch[52] Batch [1249]	Speed: 5121.045135 samples/sec	accuracy=0.754235	lr=0.251332
[Epoch 52] training: accuracy=0.754234
[Epoch 52] speed: 5012 samples/sec	time cost: 264.849426
[Epoch 52] validation: err-top1=0.348165 err-top5=0.127686
Epoch[53] Batch [49]	Speed: 4893.446690 samples/sec	accuracy=0.752756	lr=0.251117
Epoch[53] Batch [99]	Speed: 5143.249013 samples/sec	accuracy=0.752867	lr=0.250906
Epoch[53] Batch [149]	Speed: 4964.692039 samples/sec	accuracy=0.753166	lr=0.250695
Epoch[53] Batch [199]	Speed: 4904.150583 samples/sec	accuracy=0.752892	lr=0.250483
Epoch[53] Batch [249]	Speed: 5075.412574 samples/sec	accuracy=0.752982	lr=0.250272
Epoch[53] Batch [299]	Speed: 4872.121542 samples/sec	accuracy=0.753102	lr=0.250061
Epoch[53] Batch [349]	Speed: 5008.720834 samples/sec	accuracy=0.752948	lr=0.249849
Epoch[53] Batch [399]	Speed: 4950.883742 samples/sec	accuracy=0.752972	lr=0.249638
Epoch[53] Batch [449]	Speed: 5066.019520 samples/sec	accuracy=0.752906	lr=0.249426
Epoch[53] Batch [499]	Speed: 5007.710529 samples/sec	accuracy=0.752922	lr=0.249215
Epoch[53] Batch [549]	Speed: 5091.015360 samples/sec	accuracy=0.752982	lr=0.249003
Epoch[53] Batch [599]	Speed: 5055.521937 samples/sec	accuracy=0.753099	lr=0.248791
Epoch[53] Batch [649]	Speed: 5096.535521 samples/sec	accuracy=0.753128	lr=0.248579
Epoch[53] Batch [699]	Speed: 5064.270260 samples/sec	accuracy=0.753315	lr=0.248367
Epoch[53] Batch [749]	Speed: 5116.434248 samples/sec	accuracy=0.753550	lr=0.248156
Epoch[53] Batch [799]	Speed: 5058.740726 samples/sec	accuracy=0.753631	lr=0.247944
Epoch[53] Batch [849]	Speed: 4939.561726 samples/sec	accuracy=0.753786	lr=0.247732
Epoch[53] Batch [899]	Speed: 5030.756011 samples/sec	accuracy=0.753822	lr=0.247519
Epoch[53] Batch [949]	Speed: 4928.972219 samples/sec	accuracy=0.753876	lr=0.247307
Epoch[53] Batch [999]	Speed: 5022.120882 samples/sec	accuracy=0.753955	lr=0.247095
Epoch[53] Batch [1049]	Speed: 4929.101305 samples/sec	accuracy=0.753983	lr=0.246883
Epoch[53] Batch [1099]	Speed: 4961.672602 samples/sec	accuracy=0.754095	lr=0.246670
Epoch[53] Batch [1149]	Speed: 4985.054583 samples/sec	accuracy=0.754211	lr=0.246458
Epoch[53] Batch [1199]	Speed: 5022.594828 samples/sec	accuracy=0.754298	lr=0.246246
Epoch[53] Batch [1249]	Speed: 5172.739745 samples/sec	accuracy=0.754345	lr=0.246033
[Epoch 53] training: accuracy=0.754346
[Epoch 53] speed: 5009 samples/sec	time cost: 265.226427
[Epoch 53] validation: err-top1=0.337950 err-top5=0.123326
Epoch[54] Batch [49]	Speed: 4812.263484 samples/sec	accuracy=0.754361	lr=0.245816
Epoch[54] Batch [99]	Speed: 5115.857115 samples/sec	accuracy=0.754155	lr=0.245604
Epoch[54] Batch [149]	Speed: 5002.499867 samples/sec	accuracy=0.753831	lr=0.245391
Epoch[54] Batch [199]	Speed: 5036.830353 samples/sec	accuracy=0.753845	lr=0.245178
Epoch[54] Batch [249]	Speed: 4827.133467 samples/sec	accuracy=0.754082	lr=0.244966
Epoch[54] Batch [299]	Speed: 4982.590610 samples/sec	accuracy=0.754062	lr=0.244753
Epoch[54] Batch [349]	Speed: 5017.060936 samples/sec	accuracy=0.754201	lr=0.244540
Epoch[54] Batch [399]	Speed: 5117.796847 samples/sec	accuracy=0.754301	lr=0.244327
Epoch[54] Batch [449]	Speed: 5058.974065 samples/sec	accuracy=0.754498	lr=0.244114
Epoch[54] Batch [499]	Speed: 5043.895135 samples/sec	accuracy=0.754582	lr=0.243901
Epoch[54] Batch [549]	Speed: 5023.600571 samples/sec	accuracy=0.754579	lr=0.243688
Epoch[54] Batch [599]	Speed: 5026.767531 samples/sec	accuracy=0.754718	lr=0.243475
Epoch[54] Batch [649]	Speed: 4955.963129 samples/sec	accuracy=0.754894	lr=0.243262
Epoch[54] Batch [699]	Speed: 5035.561649 samples/sec	accuracy=0.755006	lr=0.243049
Epoch[54] Batch [749]	Speed: 5001.561496 samples/sec	accuracy=0.755154	lr=0.242835
Epoch[54] Batch [799]	Speed: 5034.651084 samples/sec	accuracy=0.755372	lr=0.242622
Epoch[54] Batch [849]	Speed: 5001.256083 samples/sec	accuracy=0.755454	lr=0.242409
Epoch[54] Batch [899]	Speed: 4998.266241 samples/sec	accuracy=0.755541	lr=0.242195
Epoch[54] Batch [949]	Speed: 5066.196401 samples/sec	accuracy=0.755642	lr=0.241982
Epoch[54] Batch [999]	Speed: 5159.703960 samples/sec	accuracy=0.755699	lr=0.241768
Epoch[54] Batch [1049]	Speed: 5069.608672 samples/sec	accuracy=0.755866	lr=0.241555
Epoch[54] Batch [1099]	Speed: 4663.151557 samples/sec	accuracy=0.755951	lr=0.241341
Epoch[54] Batch [1149]	Speed: 5142.199352 samples/sec	accuracy=0.756105	lr=0.241127
Epoch[54] Batch [1199]	Speed: 5021.168327 samples/sec	accuracy=0.756185	lr=0.240914
Epoch[54] Batch [1249]	Speed: 5041.115151 samples/sec	accuracy=0.756287	lr=0.240700
[Epoch 54] training: accuracy=0.756285
[Epoch 54] speed: 5004 samples/sec	time cost: 265.174517
[Epoch 54] validation: err-top1=0.353715 err-top5=0.133610
Epoch[55] Batch [49]	Speed: 4932.109763 samples/sec	accuracy=0.756048	lr=0.240482
Epoch[55] Batch [99]	Speed: 5058.224906 samples/sec	accuracy=0.755961	lr=0.240268
Epoch[55] Batch [149]	Speed: 4798.347970 samples/sec	accuracy=0.755900	lr=0.240054
Epoch[55] Batch [199]	Speed: 4855.816863 samples/sec	accuracy=0.755643	lr=0.239840
Epoch[55] Batch [249]	Speed: 4913.589378 samples/sec	accuracy=0.755564	lr=0.239626
Epoch[55] Batch [299]	Speed: 5047.968520 samples/sec	accuracy=0.755821	lr=0.239412
Epoch[55] Batch [349]	Speed: 5038.081850 samples/sec	accuracy=0.755998	lr=0.239198
Epoch[55] Batch [399]	Speed: 4894.271307 samples/sec	accuracy=0.755908	lr=0.238984
Epoch[55] Batch [449]	Speed: 5057.024831 samples/sec	accuracy=0.756127	lr=0.238769
Epoch[55] Batch [499]	Speed: 5135.350082 samples/sec	accuracy=0.756211	lr=0.238555
Epoch[55] Batch [549]	Speed: 4971.557723 samples/sec	accuracy=0.756309	lr=0.238341
Epoch[55] Batch [599]	Speed: 4921.305713 samples/sec	accuracy=0.756350	lr=0.238126
Epoch[55] Batch [649]	Speed: 4978.043585 samples/sec	accuracy=0.756497	lr=0.237912
Epoch[55] Batch [699]	Speed: 4876.490439 samples/sec	accuracy=0.756562	lr=0.237698
Epoch[55] Batch [749]	Speed: 4880.939397 samples/sec	accuracy=0.756604	lr=0.237483
Epoch[55] Batch [799]	Speed: 5037.747143 samples/sec	accuracy=0.756789	lr=0.237269
Epoch[55] Batch [849]	Speed: 4943.367563 samples/sec	accuracy=0.756841	lr=0.237054
Epoch[55] Batch [899]	Speed: 4999.517151 samples/sec	accuracy=0.756959	lr=0.236839
Epoch[55] Batch [949]	Speed: 5122.244147 samples/sec	accuracy=0.757116	lr=0.236625
Epoch[55] Batch [999]	Speed: 4968.371759 samples/sec	accuracy=0.757238	lr=0.236410
Epoch[55] Batch [1049]	Speed: 5170.787407 samples/sec	accuracy=0.757386	lr=0.236195
Epoch[55] Batch [1099]	Speed: 4957.048431 samples/sec	accuracy=0.757396	lr=0.235981
Epoch[55] Batch [1149]	Speed: 5038.841370 samples/sec	accuracy=0.757487	lr=0.235766
Epoch[55] Batch [1199]	Speed: 4898.591982 samples/sec	accuracy=0.757618	lr=0.235551
Epoch[55] Batch [1249]	Speed: 5195.643407 samples/sec	accuracy=0.757718	lr=0.235336
[Epoch 55] training: accuracy=0.757712
[Epoch 55] speed: 4981 samples/sec	time cost: 266.496074
[Epoch 55] validation: err-top1=0.344268 err-top5=0.127890
Epoch[56] Batch [49]	Speed: 4978.408606 samples/sec	accuracy=0.756507	lr=0.235117
Epoch[56] Batch [99]	Speed: 5089.186566 samples/sec	accuracy=0.756271	lr=0.234902
Epoch[56] Batch [149]	Speed: 4926.183301 samples/sec	accuracy=0.756843	lr=0.234687
Epoch[56] Batch [199]	Speed: 4929.977935 samples/sec	accuracy=0.756603	lr=0.234472
Epoch[56] Batch [249]	Speed: 5060.755326 samples/sec	accuracy=0.756585	lr=0.234256
Epoch[56] Batch [299]	Speed: 4727.654322 samples/sec	accuracy=0.756573	lr=0.234041
Epoch[56] Batch [349]	Speed: 5096.678613 samples/sec	accuracy=0.756637	lr=0.233826
Epoch[56] Batch [399]	Speed: 5090.650292 samples/sec	accuracy=0.756714	lr=0.233611
Epoch[56] Batch [449]	Speed: 5012.651113 samples/sec	accuracy=0.756733	lr=0.233396
Epoch[56] Batch [499]	Speed: 5004.464431 samples/sec	accuracy=0.756782	lr=0.233180
Epoch[56] Batch [549]	Speed: 5021.265303 samples/sec	accuracy=0.756976	lr=0.232965
Epoch[56] Batch [599]	Speed: 5067.972706 samples/sec	accuracy=0.757228	lr=0.232749
Epoch[56] Batch [649]	Speed: 5101.200992 samples/sec	accuracy=0.757357	lr=0.232534
Epoch[56] Batch [699]	Speed: 4995.436376 samples/sec	accuracy=0.757466	lr=0.232319
Epoch[56] Batch [749]	Speed: 4966.497448 samples/sec	accuracy=0.757685	lr=0.232103
Epoch[56] Batch [799]	Speed: 4964.977965 samples/sec	accuracy=0.757777	lr=0.231887
Epoch[56] Batch [849]	Speed: 4947.260985 samples/sec	accuracy=0.757872	lr=0.231672
Epoch[56] Batch [899]	Speed: 5046.821104 samples/sec	accuracy=0.757978	lr=0.231456
Epoch[56] Batch [949]	Speed: 5066.435687 samples/sec	accuracy=0.758042	lr=0.231241
Epoch[56] Batch [999]	Speed: 5043.542480 samples/sec	accuracy=0.758126	lr=0.231025
Epoch[56] Batch [1049]	Speed: 5020.708970 samples/sec	accuracy=0.758147	lr=0.230809
Epoch[56] Batch [1099]	Speed: 5073.224367 samples/sec	accuracy=0.758225	lr=0.230593
Epoch[56] Batch [1149]	Speed: 5017.371213 samples/sec	accuracy=0.758364	lr=0.230377
Epoch[56] Batch [1199]	Speed: 4996.666803 samples/sec	accuracy=0.758518	lr=0.230162
Epoch[56] Batch [1249]	Speed: 5238.911806 samples/sec	accuracy=0.758590	lr=0.229946
[Epoch 56] training: accuracy=0.758595
[Epoch 56] speed: 5013 samples/sec	time cost: 264.579969
[Epoch 56] validation: err-top1=0.345663 err-top5=0.128069
Epoch[57] Batch [49]	Speed: 4873.284997 samples/sec	accuracy=0.759304	lr=0.229725
Epoch[57] Batch [99]	Speed: 4960.749943 samples/sec	accuracy=0.757114	lr=0.229510
Epoch[57] Batch [149]	Speed: 4989.876568 samples/sec	accuracy=0.757311	lr=0.229294
Epoch[57] Batch [199]	Speed: 5017.192216 samples/sec	accuracy=0.757146	lr=0.229077
Epoch[57] Batch [249]	Speed: 5043.690904 samples/sec	accuracy=0.757375	lr=0.228861
Epoch[57] Batch [299]	Speed: 4997.991125 samples/sec	accuracy=0.757668	lr=0.228645
Epoch[57] Batch [349]	Speed: 5119.132715 samples/sec	accuracy=0.757535	lr=0.228429
Epoch[57] Batch [399]	Speed: 4898.212763 samples/sec	accuracy=0.757792	lr=0.228213
Epoch[57] Batch [449]	Speed: 4925.934480 samples/sec	accuracy=0.757812	lr=0.227997
Epoch[57] Batch [499]	Speed: 4978.558185 samples/sec	accuracy=0.757957	lr=0.227781
Epoch[57] Batch [549]	Speed: 5065.591471 samples/sec	accuracy=0.758067	lr=0.227564
Epoch[57] Batch [599]	Speed: 5039.585271 samples/sec	accuracy=0.758205	lr=0.227348
Epoch[57] Batch [649]	Speed: 5148.968027 samples/sec	accuracy=0.758362	lr=0.227132
Epoch[57] Batch [699]	Speed: 5072.025311 samples/sec	accuracy=0.758530	lr=0.226915
Epoch[57] Batch [749]	Speed: 5024.992944 samples/sec	accuracy=0.758670	lr=0.226699
Epoch[57] Batch [799]	Speed: 5045.295345 samples/sec	accuracy=0.758727	lr=0.226482
Epoch[57] Batch [849]	Speed: 4978.844325 samples/sec	accuracy=0.758751	lr=0.226266
Epoch[57] Batch [899]	Speed: 5079.513736 samples/sec	accuracy=0.758853	lr=0.226049
Epoch[57] Batch [949]	Speed: 5036.709502 samples/sec	accuracy=0.759046	lr=0.225833
Epoch[57] Batch [999]	Speed: 5050.096036 samples/sec	accuracy=0.759166	lr=0.225616
Epoch[57] Batch [1049]	Speed: 5018.528849 samples/sec	accuracy=0.759257	lr=0.225400
Epoch[57] Batch [1099]	Speed: 5108.907628 samples/sec	accuracy=0.759417	lr=0.225183
Epoch[57] Batch [1149]	Speed: 5106.486688 samples/sec	accuracy=0.759471	lr=0.224967
Epoch[57] Batch [1199]	Speed: 5121.378667 samples/sec	accuracy=0.759576	lr=0.224750
Epoch[57] Batch [1249]	Speed: 5084.098731 samples/sec	accuracy=0.759664	lr=0.224533
[Epoch 57] training: accuracy=0.759673
[Epoch 57] speed: 5026 samples/sec	time cost: 264.015529
[Epoch 57] validation: err-top1=0.339166 err-top5=0.123964
Epoch[58] Batch [49]	Speed: 4948.554797 samples/sec	accuracy=0.760409	lr=0.224312
Epoch[58] Batch [99]	Speed: 4788.429514 samples/sec	accuracy=0.759516	lr=0.224095
Epoch[58] Batch [149]	Speed: 4855.539639 samples/sec	accuracy=0.758781	lr=0.223879
Epoch[58] Batch [199]	Speed: 4900.529332 samples/sec	accuracy=0.758578	lr=0.223662
Epoch[58] Batch [249]	Speed: 4895.438893 samples/sec	accuracy=0.758434	lr=0.223445
Epoch[58] Batch [299]	Speed: 4971.176443 samples/sec	accuracy=0.758584	lr=0.223228
Epoch[58] Batch [349]	Speed: 5070.072472 samples/sec	accuracy=0.758758	lr=0.223011
Epoch[58] Batch [399]	Speed: 5020.139382 samples/sec	accuracy=0.758809	lr=0.222794
Epoch[58] Batch [449]	Speed: 5048.317878 samples/sec	accuracy=0.758896	lr=0.222577
Epoch[58] Batch [499]	Speed: 5078.494490 samples/sec	accuracy=0.758996	lr=0.222360
Epoch[58] Batch [549]	Speed: 5007.197824 samples/sec	accuracy=0.759105	lr=0.222143
Epoch[58] Batch [599]	Speed: 5056.390064 samples/sec	accuracy=0.759295	lr=0.221926
Epoch[58] Batch [649]	Speed: 4985.505008 samples/sec	accuracy=0.759580	lr=0.221709
Epoch[58] Batch [699]	Speed: 5015.772765 samples/sec	accuracy=0.759616	lr=0.221492
Epoch[58] Batch [749]	Speed: 5143.859696 samples/sec	accuracy=0.759723	lr=0.221275
Epoch[58] Batch [799]	Speed: 5133.648346 samples/sec	accuracy=0.759753	lr=0.221058
Epoch[58] Batch [849]	Speed: 5020.169307 samples/sec	accuracy=0.759919	lr=0.220841
Epoch[58] Batch [899]	Speed: 5094.097046 samples/sec	accuracy=0.760062	lr=0.220623
Epoch[58] Batch [949]	Speed: 5188.266653 samples/sec	accuracy=0.760109	lr=0.220406
Epoch[58] Batch [999]	Speed: 5030.884473 samples/sec	accuracy=0.760199	lr=0.220189
Epoch[58] Batch [1049]	Speed: 4985.816719 samples/sec	accuracy=0.760320	lr=0.219972
Epoch[58] Batch [1099]	Speed: 4943.504460 samples/sec	accuracy=0.760479	lr=0.219754
Epoch[58] Batch [1149]	Speed: 5056.603302 samples/sec	accuracy=0.760617	lr=0.219537
Epoch[58] Batch [1199]	Speed: 5065.060276 samples/sec	accuracy=0.760697	lr=0.219320
Epoch[58] Batch [1249]	Speed: 5065.643689 samples/sec	accuracy=0.760771	lr=0.219102
[Epoch 58] training: accuracy=0.760773
[Epoch 58] speed: 5009 samples/sec	time cost: 264.989862
[Epoch 58] validation: err-top1=0.343709 err-top5=0.124390
Epoch[59] Batch [49]	Speed: 4777.126920 samples/sec	accuracy=0.761405	lr=0.218881
Epoch[59] Batch [99]	Speed: 4910.056787 samples/sec	accuracy=0.760894	lr=0.218663
Epoch[59] Batch [149]	Speed: 5079.943660 samples/sec	accuracy=0.760145	lr=0.218446
Epoch[59] Batch [199]	Speed: 5094.737328 samples/sec	accuracy=0.760055	lr=0.218228
Epoch[59] Batch [249]	Speed: 5024.610714 samples/sec	accuracy=0.760025	lr=0.218011
Epoch[59] Batch [299]	Speed: 5095.124016 samples/sec	accuracy=0.760249	lr=0.217793
Epoch[59] Batch [349]	Speed: 4919.056561 samples/sec	accuracy=0.760349	lr=0.217576
Epoch[59] Batch [399]	Speed: 4874.608338 samples/sec	accuracy=0.760493	lr=0.217358
Epoch[59] Batch [449]	Speed: 5081.627046 samples/sec	accuracy=0.760460	lr=0.217141
Epoch[59] Batch [499]	Speed: 5050.268481 samples/sec	accuracy=0.760468	lr=0.216923
Epoch[59] Batch [549]	Speed: 5042.129864 samples/sec	accuracy=0.760621	lr=0.216706
Epoch[59] Batch [599]	Speed: 5065.603062 samples/sec	accuracy=0.760693	lr=0.216488
Epoch[59] Batch [649]	Speed: 5122.760154 samples/sec	accuracy=0.760743	lr=0.216270
Epoch[59] Batch [699]	Speed: 5047.175404 samples/sec	accuracy=0.760800	lr=0.216053
Epoch[59] Batch [749]	Speed: 4991.271189 samples/sec	accuracy=0.760811	lr=0.215835
Epoch[59] Batch [799]	Speed: 5061.523727 samples/sec	accuracy=0.760917	lr=0.215617
Epoch[59] Batch [849]	Speed: 4998.381299 samples/sec	accuracy=0.761097	lr=0.215400
Epoch[59] Batch [899]	Speed: 5091.645692 samples/sec	accuracy=0.761184	lr=0.215182
Epoch[59] Batch [949]	Speed: 5137.774136 samples/sec	accuracy=0.761261	lr=0.214964
Epoch[59] Batch [999]	Speed: 5140.440902 samples/sec	accuracy=0.761362	lr=0.214746
Epoch[59] Batch [1049]	Speed: 4991.003106 samples/sec	accuracy=0.761515	lr=0.214529
Epoch[59] Batch [1099]	Speed: 4935.447868 samples/sec	accuracy=0.761577	lr=0.214311
Epoch[59] Batch [1149]	Speed: 5044.324975 samples/sec	accuracy=0.761670	lr=0.214093
Epoch[59] Batch [1199]	Speed: 4973.847088 samples/sec	accuracy=0.761749	lr=0.213875
Epoch[59] Batch [1249]	Speed: 5112.699146 samples/sec	accuracy=0.761840	lr=0.213657
[Epoch 59] training: accuracy=0.761841
[Epoch 59] speed: 5020 samples/sec	time cost: 264.620780
[Epoch 59] validation: err-top1=0.329102 err-top5=0.117626
Epoch[60] Batch [49]	Speed: 4812.823980 samples/sec	accuracy=0.761173	lr=0.213435
Epoch[60] Batch [99]	Speed: 4922.401498 samples/sec	accuracy=0.762144	lr=0.213217
Epoch[60] Batch [149]	Speed: 4977.720730 samples/sec	accuracy=0.761836	lr=0.212999
Epoch[60] Batch [199]	Speed: 4938.789472 samples/sec	accuracy=0.761172	lr=0.212781
Epoch[60] Batch [249]	Speed: 4934.516453 samples/sec	accuracy=0.760958	lr=0.212563
Epoch[60] Batch [299]	Speed: 5135.269647 samples/sec	accuracy=0.761069	lr=0.212345
Epoch[60] Batch [349]	Speed: 4983.217620 samples/sec	accuracy=0.760989	lr=0.212127
Epoch[60] Batch [399]	Speed: 4946.856530 samples/sec	accuracy=0.760966	lr=0.211909
Epoch[60] Batch [449]	Speed: 4932.990410 samples/sec	accuracy=0.761191	lr=0.211691
Epoch[60] Batch [499]	Speed: 5046.929512 samples/sec	accuracy=0.761431	lr=0.211473
Epoch[60] Batch [549]	Speed: 4977.964194 samples/sec	accuracy=0.761443	lr=0.211255
Epoch[60] Batch [599]	Speed: 5156.290450 samples/sec	accuracy=0.761597	lr=0.211037
Epoch[60] Batch [649]	Speed: 4966.068826 samples/sec	accuracy=0.761759	lr=0.210819
Epoch[60] Batch [699]	Speed: 4952.479469 samples/sec	accuracy=0.761851	lr=0.210601
Epoch[60] Batch [749]	Speed: 5008.289098 samples/sec	accuracy=0.761948	lr=0.210383
Epoch[60] Batch [799]	Speed: 5058.491799 samples/sec	accuracy=0.762109	lr=0.210165
Epoch[60] Batch [849]	Speed: 5060.317555 samples/sec	accuracy=0.762269	lr=0.209947
Epoch[60] Batch [899]	Speed: 4952.082382 samples/sec	accuracy=0.762402	lr=0.209729
Epoch[60] Batch [949]	Speed: 4995.745844 samples/sec	accuracy=0.762494	lr=0.209511
Epoch[60] Batch [999]	Speed: 5049.740494 samples/sec	accuracy=0.762771	lr=0.209293
Epoch[60] Batch [1049]	Speed: 5098.115791 samples/sec	accuracy=0.762932	lr=0.209075
Epoch[60] Batch [1099]	Speed: 5019.212683 samples/sec	accuracy=0.763086	lr=0.208856
Epoch[60] Batch [1149]	Speed: 5036.190252 samples/sec	accuracy=0.763172	lr=0.208638
Epoch[60] Batch [1199]	Speed: 4939.708979 samples/sec	accuracy=0.763302	lr=0.208420
Epoch[60] Batch [1249]	Speed: 5043.938021 samples/sec	accuracy=0.763332	lr=0.208202
[Epoch 60] training: accuracy=0.763339
[Epoch 60] speed: 4992 samples/sec	time cost: 265.623530
[Epoch 60] validation: err-top1=0.330576 err-top5=0.117207
Epoch[61] Batch [49]	Speed: 4736.132653 samples/sec	accuracy=0.762986	lr=0.207979
Epoch[61] Batch [99]	Speed: 4893.006613 samples/sec	accuracy=0.762653	lr=0.207761
Epoch[61] Batch [149]	Speed: 4952.865996 samples/sec	accuracy=0.762294	lr=0.207543
Epoch[61] Batch [199]	Speed: 5116.957985 samples/sec	accuracy=0.762653	lr=0.207325
Epoch[61] Batch [249]	Speed: 4670.908039 samples/sec	accuracy=0.762695	lr=0.207106
Epoch[61] Batch [299]	Speed: 5010.716711 samples/sec	accuracy=0.762872	lr=0.206888
Epoch[61] Batch [349]	Speed: 4886.934250 samples/sec	accuracy=0.762933	lr=0.206670
Epoch[61] Batch [399]	Speed: 5032.279826 samples/sec	accuracy=0.763091	lr=0.206452
Epoch[61] Batch [449]	Speed: 5025.376174 samples/sec	accuracy=0.763208	lr=0.206233
Epoch[61] Batch [499]	Speed: 4846.648500 samples/sec	accuracy=0.763316	lr=0.206015
Epoch[61] Batch [549]	Speed: 5036.374033 samples/sec	accuracy=0.763514	lr=0.205797
Epoch[61] Batch [599]	Speed: 5042.189413 samples/sec	accuracy=0.763676	lr=0.205579
Epoch[61] Batch [649]	Speed: 4951.956314 samples/sec	accuracy=0.763790	lr=0.205360
Epoch[61] Batch [699]	Speed: 5018.841301 samples/sec	accuracy=0.763922	lr=0.205142
Epoch[61] Batch [749]	Speed: 5049.037277 samples/sec	accuracy=0.763978	lr=0.204924
Epoch[61] Batch [799]	Speed: 5024.025548 samples/sec	accuracy=0.763965	lr=0.204705
Epoch[61] Batch [849]	Speed: 5051.872000 samples/sec	accuracy=0.764096	lr=0.204487
Epoch[61] Batch [899]	Speed: 4956.942820 samples/sec	accuracy=0.764246	lr=0.204269
Epoch[61] Batch [949]	Speed: 4899.414199 samples/sec	accuracy=0.764342	lr=0.204051
Epoch[61] Batch [999]	Speed: 4861.286680 samples/sec	accuracy=0.764384	lr=0.203832
Epoch[61] Batch [1049]	Speed: 4914.062513 samples/sec	accuracy=0.764424	lr=0.203614
Epoch[61] Batch [1099]	Speed: 4937.666284 samples/sec	accuracy=0.764450	lr=0.203396
Epoch[61] Batch [1149]	Speed: 4978.007351 samples/sec	accuracy=0.764555	lr=0.203177
Epoch[61] Batch [1199]	Speed: 4974.089597 samples/sec	accuracy=0.764639	lr=0.202959
Epoch[61] Batch [1249]	Speed: 5063.713312 samples/sec	accuracy=0.764697	lr=0.202740
[Epoch 61] training: accuracy=0.764698
[Epoch 61] speed: 4950 samples/sec	time cost: 268.156788
[Epoch 61] validation: err-top1=0.334642 err-top5=0.120536
Epoch[62] Batch [49]	Speed: 4769.067656 samples/sec	accuracy=0.764613	lr=0.202518
Epoch[62] Batch [99]	Speed: 4837.700279 samples/sec	accuracy=0.763942	lr=0.202299
Epoch[62] Batch [149]	Speed: 4961.970677 samples/sec	accuracy=0.763521	lr=0.202081
Epoch[62] Batch [199]	Speed: 4708.424401 samples/sec	accuracy=0.763276	lr=0.201863
Epoch[62] Batch [249]	Speed: 5038.390240 samples/sec	accuracy=0.762725	lr=0.201644
Epoch[62] Batch [299]	Speed: 4759.670173 samples/sec	accuracy=0.763210	lr=0.201426
Epoch[62] Batch [349]	Speed: 4961.002408 samples/sec	accuracy=0.763317	lr=0.201208
Epoch[62] Batch [399]	Speed: 5020.128116 samples/sec	accuracy=0.763428	lr=0.200989
Epoch[62] Batch [449]	Speed: 4957.110335 samples/sec	accuracy=0.763494	lr=0.200771
Epoch[62] Batch [499]	Speed: 5200.354910 samples/sec	accuracy=0.763603	lr=0.200552
Epoch[62] Batch [549]	Speed: 4954.399219 samples/sec	accuracy=0.763669	lr=0.200334
Epoch[62] Batch [599]	Speed: 5282.634923 samples/sec	accuracy=0.763845	lr=0.200116
Epoch[62] Batch [649]	Speed: 4981.129547 samples/sec	accuracy=0.764012	lr=0.199897
Epoch[62] Batch [699]	Speed: 5062.688342 samples/sec	accuracy=0.764170	lr=0.199679
Epoch[62] Batch [749]	Speed: 5177.058429 samples/sec	accuracy=0.764307	lr=0.199461
Epoch[62] Batch [799]	Speed: 5058.469755 samples/sec	accuracy=0.764401	lr=0.199242
Epoch[62] Batch [849]	Speed: 4949.091946 samples/sec	accuracy=0.764455	lr=0.199024
Epoch[62] Batch [899]	Speed: 4977.898883 samples/sec	accuracy=0.764568	lr=0.198806
Epoch[62] Batch [949]	Speed: 5003.510285 samples/sec	accuracy=0.764683	lr=0.198587
Epoch[62] Batch [999]	Speed: 5043.451037 samples/sec	accuracy=0.764705	lr=0.198369
Epoch[62] Batch [1049]	Speed: 4923.826840 samples/sec	accuracy=0.764816	lr=0.198150
Epoch[62] Batch [1099]	Speed: 5076.917236 samples/sec	accuracy=0.765000	lr=0.197932
Epoch[62] Batch [1149]	Speed: 5055.583231 samples/sec	accuracy=0.765182	lr=0.197714
Epoch[62] Batch [1199]	Speed: 5000.318877 samples/sec	accuracy=0.765283	lr=0.197495
Epoch[62] Batch [1249]	Speed: 5182.813645 samples/sec	accuracy=0.765465	lr=0.197277
[Epoch 62] training: accuracy=0.765465
[Epoch 62] speed: 4989 samples/sec	time cost: 266.021947
[Epoch 62] validation: err-top1=0.361009 err-top5=0.135643
Epoch[63] Batch [49]	Speed: 4960.409734 samples/sec	accuracy=0.766609	lr=0.197054
Epoch[63] Batch [99]	Speed: 5024.433669 samples/sec	accuracy=0.766170	lr=0.196836
Epoch[63] Batch [149]	Speed: 5018.261231 samples/sec	accuracy=0.765881	lr=0.196618
Epoch[63] Batch [199]	Speed: 5122.971328 samples/sec	accuracy=0.765555	lr=0.196399
Epoch[63] Batch [249]	Speed: 5003.773883 samples/sec	accuracy=0.765709	lr=0.196181
Epoch[63] Batch [299]	Speed: 5016.475535 samples/sec	accuracy=0.765577	lr=0.195963
Epoch[63] Batch [349]	Speed: 4998.257400 samples/sec	accuracy=0.765549	lr=0.195744
Epoch[63] Batch [399]	Speed: 5000.628484 samples/sec	accuracy=0.765650	lr=0.195526
Epoch[63] Batch [449]	Speed: 5023.644523 samples/sec	accuracy=0.765692	lr=0.195308
Epoch[63] Batch [499]	Speed: 5027.858406 samples/sec	accuracy=0.765789	lr=0.195089
Epoch[63] Batch [549]	Speed: 4995.216066 samples/sec	accuracy=0.765893	lr=0.194871
Epoch[63] Batch [599]	Speed: 5023.118563 samples/sec	accuracy=0.765923	lr=0.194653
Epoch[63] Batch [649]	Speed: 4965.956055 samples/sec	accuracy=0.765873	lr=0.194434
Epoch[63] Batch [699]	Speed: 4985.378274 samples/sec	accuracy=0.765952	lr=0.194216
Epoch[63] Batch [749]	Speed: 4946.394377 samples/sec	accuracy=0.766162	lr=0.193998
Epoch[63] Batch [799]	Speed: 4937.158740 samples/sec	accuracy=0.766223	lr=0.193780
Epoch[63] Batch [849]	Speed: 5053.493544 samples/sec	accuracy=0.766255	lr=0.193561
Epoch[63] Batch [899]	Speed: 5049.493521 samples/sec	accuracy=0.766313	lr=0.193343
Epoch[63] Batch [949]	Speed: 4992.452791 samples/sec	accuracy=0.766312	lr=0.193125
Epoch[63] Batch [999]	Speed: 5034.183829 samples/sec	accuracy=0.766491	lr=0.192907
Epoch[63] Batch [1049]	Speed: 5082.157273 samples/sec	accuracy=0.766553	lr=0.192688
Epoch[63] Batch [1099]	Speed: 5023.422657 samples/sec	accuracy=0.766589	lr=0.192470
Epoch[63] Batch [1149]	Speed: 5030.621428 samples/sec	accuracy=0.766725	lr=0.192252
Epoch[63] Batch [1199]	Speed: 5086.455957 samples/sec	accuracy=0.766848	lr=0.192034
Epoch[63] Batch [1249]	Speed: 5030.542355 samples/sec	accuracy=0.766903	lr=0.191816
[Epoch 63] training: accuracy=0.766899
[Epoch 63] speed: 5013 samples/sec	time cost: 264.260561
[Epoch 63] validation: err-top1=0.331665 err-top5=0.118429
Epoch[64] Batch [49]	Speed: 4925.933915 samples/sec	accuracy=0.765343	lr=0.191593
Epoch[64] Batch [99]	Speed: 5087.270987 samples/sec	accuracy=0.765333	lr=0.191375
Epoch[64] Batch [149]	Speed: 5151.001278 samples/sec	accuracy=0.765687	lr=0.191157
Epoch[64] Batch [199]	Speed: 4909.395301 samples/sec	accuracy=0.765545	lr=0.190939
Epoch[64] Batch [249]	Speed: 4836.090958 samples/sec	accuracy=0.765697	lr=0.190720
Epoch[64] Batch [299]	Speed: 4906.619854 samples/sec	accuracy=0.766039	lr=0.190502
Epoch[64] Batch [349]	Speed: 4948.719921 samples/sec	accuracy=0.765966	lr=0.190284
Epoch[64] Batch [399]	Speed: 4966.252348 samples/sec	accuracy=0.766204	lr=0.190066
Epoch[64] Batch [449]	Speed: 5101.113141 samples/sec	accuracy=0.766243	lr=0.189848
Epoch[64] Batch [499]	Speed: 5033.352332 samples/sec	accuracy=0.766400	lr=0.189630
Epoch[64] Batch [549]	Speed: 4965.291017 samples/sec	accuracy=0.766460	lr=0.189412
Epoch[64] Batch [599]	Speed: 5056.277201 samples/sec	accuracy=0.766613	lr=0.189194
Epoch[64] Batch [649]	Speed: 4890.632326 samples/sec	accuracy=0.766871	lr=0.188976
Epoch[64] Batch [699]	Speed: 5058.183921 samples/sec	accuracy=0.766957	lr=0.188758
Epoch[64] Batch [749]	Speed: 5024.213379 samples/sec	accuracy=0.767116	lr=0.188540
Epoch[64] Batch [799]	Speed: 4995.157040 samples/sec	accuracy=0.767306	lr=0.188322
Epoch[64] Batch [849]	Speed: 5019.501051 samples/sec	accuracy=0.767329	lr=0.188104
Epoch[64] Batch [899]	Speed: 5154.972367 samples/sec	accuracy=0.767400	lr=0.187886
Epoch[64] Batch [949]	Speed: 4948.937062 samples/sec	accuracy=0.767507	lr=0.187668
Epoch[64] Batch [999]	Speed: 5059.163565 samples/sec	accuracy=0.767597	lr=0.187450
Epoch[64] Batch [1049]	Speed: 5008.524230 samples/sec	accuracy=0.767669	lr=0.187232
Epoch[64] Batch [1099]	Speed: 5041.784795 samples/sec	accuracy=0.767810	lr=0.187014
Epoch[64] Batch [1149]	Speed: 5028.438225 samples/sec	accuracy=0.767873	lr=0.186796
Epoch[64] Batch [1199]	Speed: 5003.572422 samples/sec	accuracy=0.767937	lr=0.186578
Epoch[64] Batch [1249]	Speed: 5239.820538 samples/sec	accuracy=0.768000	lr=0.186360
[Epoch 64] training: accuracy=0.768007
[Epoch 64] speed: 5009 samples/sec	time cost: 265.678206
[Epoch 64] validation: err-top1=0.331932 err-top5=0.119380
Epoch[65] Batch [49]	Speed: 4718.198268 samples/sec	accuracy=0.768575	lr=0.186138
Epoch[65] Batch [99]	Speed: 4950.445258 samples/sec	accuracy=0.767727	lr=0.185920
Epoch[65] Batch [149]	Speed: 5015.995830 samples/sec	accuracy=0.767220	lr=0.185702
Epoch[65] Batch [199]	Speed: 4863.921828 samples/sec	accuracy=0.767011	lr=0.185485
Epoch[65] Batch [249]	Speed: 4952.428530 samples/sec	accuracy=0.766952	lr=0.185267
Epoch[65] Batch [299]	Speed: 4798.860725 samples/sec	accuracy=0.767130	lr=0.185049
Epoch[65] Batch [349]	Speed: 4944.705335 samples/sec	accuracy=0.767389	lr=0.184831
Epoch[65] Batch [399]	Speed: 4904.278932 samples/sec	accuracy=0.767554	lr=0.184614
Epoch[65] Batch [449]	Speed: 5049.231969 samples/sec	accuracy=0.767629	lr=0.184396
Epoch[65] Batch [499]	Speed: 4814.900903 samples/sec	accuracy=0.767773	lr=0.184178
Epoch[65] Batch [549]	Speed: 4956.410714 samples/sec	accuracy=0.767719	lr=0.183960
Epoch[65] Batch [599]	Speed: 4933.184754 samples/sec	accuracy=0.767870	lr=0.183743
Epoch[65] Batch [649]	Speed: 4816.068077 samples/sec	accuracy=0.767933	lr=0.183525
Epoch[65] Batch [699]	Speed: 5054.269257 samples/sec	accuracy=0.768140	lr=0.183308
Epoch[65] Batch [749]	Speed: 4953.702306 samples/sec	accuracy=0.768135	lr=0.183090
Epoch[65] Batch [799]	Speed: 4887.882720 samples/sec	accuracy=0.768237	lr=0.182872
Epoch[65] Batch [849]	Speed: 5089.773379 samples/sec	accuracy=0.768395	lr=0.182655
Epoch[65] Batch [899]	Speed: 5100.750259 samples/sec	accuracy=0.768537	lr=0.182437
Epoch[65] Batch [949]	Speed: 4952.997250 samples/sec	accuracy=0.768567	lr=0.182220
Epoch[65] Batch [999]	Speed: 5212.445351 samples/sec	accuracy=0.768628	lr=0.182002
Epoch[65] Batch [1049]	Speed: 5070.390298 samples/sec	accuracy=0.768812	lr=0.181785
Epoch[65] Batch [1099]	Speed: 4967.297351 samples/sec	accuracy=0.768942	lr=0.181567
Epoch[65] Batch [1149]	Speed: 5160.118800 samples/sec	accuracy=0.769060	lr=0.181350
Epoch[65] Batch [1199]	Speed: 5259.221078 samples/sec	accuracy=0.769170	lr=0.181132
Epoch[65] Batch [1249]	Speed: 5320.813814 samples/sec	accuracy=0.769294	lr=0.180915
[Epoch 65] training: accuracy=0.769299
[Epoch 65] speed: 4980 samples/sec	time cost: 266.986721
[Epoch 65] validation: err-top1=0.322246 err-top5=0.111089
Epoch[66] Batch [49]	Speed: 4770.894032 samples/sec	accuracy=0.769000	lr=0.180693
Epoch[66] Batch [99]	Speed: 4959.311393 samples/sec	accuracy=0.768098	lr=0.180476
Epoch[66] Batch [149]	Speed: 5235.880461 samples/sec	accuracy=0.767976	lr=0.180259
Epoch[66] Batch [199]	Speed: 5232.304639 samples/sec	accuracy=0.767958	lr=0.180041
Epoch[66] Batch [249]	Speed: 5151.969993 samples/sec	accuracy=0.767788	lr=0.179824
Epoch[66] Batch [299]	Speed: 5031.667054 samples/sec	accuracy=0.767730	lr=0.179607
Epoch[66] Batch [349]	Speed: 4920.433629 samples/sec	accuracy=0.767880	lr=0.179390
Epoch[66] Batch [399]	Speed: 5093.836532 samples/sec	accuracy=0.767996	lr=0.179172
Epoch[66] Batch [449]	Speed: 4961.963340 samples/sec	accuracy=0.768197	lr=0.178955
Epoch[66] Batch [499]	Speed: 5074.667532 samples/sec	accuracy=0.768523	lr=0.178738
Epoch[66] Batch [549]	Speed: 4994.217705 samples/sec	accuracy=0.768651	lr=0.178521
Epoch[66] Batch [599]	Speed: 4887.140553 samples/sec	accuracy=0.768783	lr=0.178304
Epoch[66] Batch [649]	Speed: 5073.491886 samples/sec	accuracy=0.768914	lr=0.178087
Epoch[66] Batch [699]	Speed: 4984.654223 samples/sec	accuracy=0.768991	lr=0.177870
Epoch[66] Batch [749]	Speed: 5183.853175 samples/sec	accuracy=0.768928	lr=0.177653
Epoch[66] Batch [799]	Speed: 5010.529420 samples/sec	accuracy=0.768854	lr=0.177436
Epoch[66] Batch [849]	Speed: 4969.902983 samples/sec	accuracy=0.768983	lr=0.177219
Epoch[66] Batch [899]	Speed: 5078.805205 samples/sec	accuracy=0.769213	lr=0.177002
Epoch[66] Batch [949]	Speed: 5047.516111 samples/sec	accuracy=0.769352	lr=0.176785
Epoch[66] Batch [999]	Speed: 5014.394862 samples/sec	accuracy=0.769597	lr=0.176568
Epoch[66] Batch [1049]	Speed: 5044.979116 samples/sec	accuracy=0.769674	lr=0.176351
Epoch[66] Batch [1099]	Speed: 5043.693510 samples/sec	accuracy=0.769836	lr=0.176134
Epoch[66] Batch [1149]	Speed: 5102.468440 samples/sec	accuracy=0.769969	lr=0.175918
Epoch[66] Batch [1199]	Speed: 4983.254046 samples/sec	accuracy=0.770112	lr=0.175701
Epoch[66] Batch [1249]	Speed: 5243.704706 samples/sec	accuracy=0.770190	lr=0.175484
[Epoch 66] training: accuracy=0.770191
[Epoch 66] speed: 5037 samples/sec	time cost: 263.613118
[Epoch 66] validation: err-top1=0.319655 err-top5=0.111288
Epoch[67] Batch [49]	Speed: 4720.625638 samples/sec	accuracy=0.770368	lr=0.175263
Epoch[67] Batch [99]	Speed: 4891.302692 samples/sec	accuracy=0.769344	lr=0.175046
Epoch[67] Batch [149]	Speed: 5009.338196 samples/sec	accuracy=0.768770	lr=0.174830
Epoch[67] Batch [199]	Speed: 5082.209111 samples/sec	accuracy=0.769098	lr=0.174613
Epoch[67] Batch [249]	Speed: 5023.236882 samples/sec	accuracy=0.769591	lr=0.174397
Epoch[67] Batch [299]	Speed: 4963.007111 samples/sec	accuracy=0.769822	lr=0.174180
Epoch[67] Batch [349]	Speed: 5142.481461 samples/sec	accuracy=0.769888	lr=0.173964
Epoch[67] Batch [399]	Speed: 5147.044692 samples/sec	accuracy=0.769958	lr=0.173747
Epoch[67] Batch [449]	Speed: 5052.742440 samples/sec	accuracy=0.770065	lr=0.173531
Epoch[67] Batch [499]	Speed: 5182.599635 samples/sec	accuracy=0.770082	lr=0.173314
Epoch[67] Batch [549]	Speed: 5054.312081 samples/sec	accuracy=0.770211	lr=0.173098
Epoch[67] Batch [599]	Speed: 5040.523531 samples/sec	accuracy=0.770177	lr=0.172881
Epoch[67] Batch [649]	Speed: 5127.844030 samples/sec	accuracy=0.770197	lr=0.172665
Epoch[67] Batch [699]	Speed: 5203.837498 samples/sec	accuracy=0.770249	lr=0.172449
Epoch[67] Batch [749]	Speed: 5247.048432 samples/sec	accuracy=0.770334	lr=0.172232
Epoch[67] Batch [799]	Speed: 4850.148902 samples/sec	accuracy=0.770396	lr=0.172016
Epoch[67] Batch [849]	Speed: 5063.832000 samples/sec	accuracy=0.770508	lr=0.171800
Epoch[67] Batch [899]	Speed: 4791.802780 samples/sec	accuracy=0.770710	lr=0.171584
Epoch[67] Batch [949]	Speed: 4874.075177 samples/sec	accuracy=0.770833	lr=0.171368
Epoch[67] Batch [999]	Speed: 4972.229276 samples/sec	accuracy=0.770959	lr=0.171152
Epoch[67] Batch [1049]	Speed: 5004.695822 samples/sec	accuracy=0.771129	lr=0.170935
Epoch[67] Batch [1099]	Speed: 5067.744396 samples/sec	accuracy=0.771326	lr=0.170719
Epoch[67] Batch [1149]	Speed: 5190.230083 samples/sec	accuracy=0.771457	lr=0.170503
Epoch[67] Batch [1199]	Speed: 5059.394559 samples/sec	accuracy=0.771603	lr=0.170287
Epoch[67] Batch [1249]	Speed: 5014.866883 samples/sec	accuracy=0.771650	lr=0.170072
[Epoch 67] training: accuracy=0.771651
[Epoch 67] speed: 5023 samples/sec	time cost: 264.232921
[Epoch 67] validation: err-top1=0.323461 err-top5=0.115214
Epoch[68] Batch [49]	Speed: 4625.083660 samples/sec	accuracy=0.771003	lr=0.169851
Epoch[68] Batch [99]	Speed: 4966.017722 samples/sec	accuracy=0.770975	lr=0.169635
Epoch[68] Batch [149]	Speed: 4805.180926 samples/sec	accuracy=0.771188	lr=0.169420
Epoch[68] Batch [199]	Speed: 4912.371653 samples/sec	accuracy=0.770957	lr=0.169204
Epoch[68] Batch [249]	Speed: 5051.624818 samples/sec	accuracy=0.771217	lr=0.168988
Epoch[68] Batch [299]	Speed: 5050.014924 samples/sec	accuracy=0.771120	lr=0.168772
Epoch[68] Batch [349]	Speed: 5075.225933 samples/sec	accuracy=0.771172	lr=0.168557
Epoch[68] Batch [399]	Speed: 5072.096110 samples/sec	accuracy=0.771305	lr=0.168341
Epoch[68] Batch [449]	Speed: 4964.277614 samples/sec	accuracy=0.771292	lr=0.168126
Epoch[68] Batch [499]	Speed: 4958.758512 samples/sec	accuracy=0.771347	lr=0.167910
Epoch[68] Batch [549]	Speed: 5095.097663 samples/sec	accuracy=0.771481	lr=0.167694
Epoch[68] Batch [599]	Speed: 4990.356972 samples/sec	accuracy=0.771504	lr=0.167479
Epoch[68] Batch [649]	Speed: 4787.200140 samples/sec	accuracy=0.771552	lr=0.167263
Epoch[68] Batch [699]	Speed: 4895.796923 samples/sec	accuracy=0.771667	lr=0.167048
Epoch[68] Batch [749]	Speed: 4948.455819 samples/sec	accuracy=0.771772	lr=0.166833
Epoch[68] Batch [799]	Speed: 4893.040282 samples/sec	accuracy=0.771878	lr=0.166617
Epoch[68] Batch [849]	Speed: 4815.325638 samples/sec	accuracy=0.771913	lr=0.166402
Epoch[68] Batch [899]	Speed: 5039.701529 samples/sec	accuracy=0.771976	lr=0.166187
Epoch[68] Batch [949]	Speed: 4995.503543 samples/sec	accuracy=0.772003	lr=0.165972
Epoch[68] Batch [999]	Speed: 4826.216992 samples/sec	accuracy=0.772022	lr=0.165756
Epoch[68] Batch [1049]	Speed: 5143.269215 samples/sec	accuracy=0.772068	lr=0.165541
Epoch[68] Batch [1099]	Speed: 5130.581394 samples/sec	accuracy=0.772161	lr=0.165326
Epoch[68] Batch [1149]	Speed: 5056.077442 samples/sec	accuracy=0.772246	lr=0.165111
Epoch[68] Batch [1199]	Speed: 4887.410163 samples/sec	accuracy=0.772310	lr=0.164896
Epoch[68] Batch [1249]	Speed: 5039.738075 samples/sec	accuracy=0.772391	lr=0.164681
[Epoch 68] training: accuracy=0.772393
[Epoch 68] speed: 4954 samples/sec	time cost: 267.714568
[Epoch 68] validation: err-top1=0.317921 err-top5=0.109375
Epoch[69] Batch [49]	Speed: 4811.235152 samples/sec	accuracy=0.772707	lr=0.164462
Epoch[69] Batch [99]	Speed: 4700.531093 samples/sec	accuracy=0.771743	lr=0.164247
Epoch[69] Batch [149]	Speed: 4938.122038 samples/sec	accuracy=0.772070	lr=0.164032
Epoch[69] Batch [199]	Speed: 4861.539247 samples/sec	accuracy=0.772101	lr=0.163818
Epoch[69] Batch [249]	Speed: 4995.321803 samples/sec	accuracy=0.772160	lr=0.163603
Epoch[69] Batch [299]	Speed: 4829.510348 samples/sec	accuracy=0.772057	lr=0.163388
Epoch[69] Batch [349]	Speed: 4984.533896 samples/sec	accuracy=0.771978	lr=0.163173
Epoch[69] Batch [399]	Speed: 5026.229154 samples/sec	accuracy=0.771856	lr=0.162959
Epoch[69] Batch [449]	Speed: 5030.537405 samples/sec	accuracy=0.772040	lr=0.162744
Epoch[69] Batch [499]	Speed: 5018.935725 samples/sec	accuracy=0.772002	lr=0.162530
Epoch[69] Batch [549]	Speed: 5091.576278 samples/sec	accuracy=0.772127	lr=0.162315
Epoch[69] Batch [599]	Speed: 5116.072351 samples/sec	accuracy=0.772308	lr=0.162101
Epoch[69] Batch [649]	Speed: 4891.503236 samples/sec	accuracy=0.772390	lr=0.161886
Epoch[69] Batch [699]	Speed: 4977.602237 samples/sec	accuracy=0.772468	lr=0.161672
Epoch[69] Batch [749]	Speed: 4948.919840 samples/sec	accuracy=0.772483	lr=0.161458
Epoch[69] Batch [799]	Speed: 5050.457804 samples/sec	accuracy=0.772572	lr=0.161244
Epoch[69] Batch [849]	Speed: 4926.322073 samples/sec	accuracy=0.772687	lr=0.161029
Epoch[69] Batch [899]	Speed: 5035.307677 samples/sec	accuracy=0.772840	lr=0.160815
Epoch[69] Batch [949]	Speed: 5094.649578 samples/sec	accuracy=0.772873	lr=0.160601
Epoch[69] Batch [999]	Speed: 5057.448931 samples/sec	accuracy=0.773054	lr=0.160387
Epoch[69] Batch [1049]	Speed: 4974.949918 samples/sec	accuracy=0.773267	lr=0.160173
Epoch[69] Batch [1099]	Speed: 5197.399707 samples/sec	accuracy=0.773357	lr=0.159959
Epoch[69] Batch [1149]	Speed: 4932.866332 samples/sec	accuracy=0.773419	lr=0.159745
Epoch[69] Batch [1199]	Speed: 5217.565382 samples/sec	accuracy=0.773569	lr=0.159531
Epoch[69] Batch [1249]	Speed: 4987.062447 samples/sec	accuracy=0.773636	lr=0.159317
[Epoch 69] training: accuracy=0.773640
[Epoch 69] speed: 4980 samples/sec	time cost: 266.141284
[Epoch 69] validation: err-top1=0.316386 err-top5=0.111410
Epoch[70] Batch [49]	Speed: 4923.138501 samples/sec	accuracy=0.774263	lr=0.159099
Epoch[70] Batch [99]	Speed: 5170.131728 samples/sec	accuracy=0.773846	lr=0.158886
Epoch[70] Batch [149]	Speed: 4891.653209 samples/sec	accuracy=0.773938	lr=0.158672
Epoch[70] Batch [199]	Speed: 4834.866599 samples/sec	accuracy=0.773769	lr=0.158458
Epoch[70] Batch [249]	Speed: 4914.684542 samples/sec	accuracy=0.773776	lr=0.158245
Epoch[70] Batch [299]	Speed: 5015.148779 samples/sec	accuracy=0.773523	lr=0.158031
Epoch[70] Batch [349]	Speed: 5108.034499 samples/sec	accuracy=0.773542	lr=0.157818
Epoch[70] Batch [399]	Speed: 5015.348246 samples/sec	accuracy=0.773772	lr=0.157604
Epoch[70] Batch [449]	Speed: 5220.194330 samples/sec	accuracy=0.773810	lr=0.157391
Epoch[70] Batch [499]	Speed: 4844.439078 samples/sec	accuracy=0.773803	lr=0.157177
Epoch[70] Batch [549]	Speed: 5114.761956 samples/sec	accuracy=0.773784	lr=0.156964
Epoch[70] Batch [599]	Speed: 5105.335457 samples/sec	accuracy=0.773912	lr=0.156751
Epoch[70] Batch [649]	Speed: 5114.110542 samples/sec	accuracy=0.773937	lr=0.156538
Epoch[70] Batch [699]	Speed: 5012.451393 samples/sec	accuracy=0.774027	lr=0.156325
Epoch[70] Batch [749]	Speed: 5028.650996 samples/sec	accuracy=0.774019	lr=0.156112
Epoch[70] Batch [799]	Speed: 5165.129294 samples/sec	accuracy=0.774166	lr=0.155899
Epoch[70] Batch [849]	Speed: 5093.676443 samples/sec	accuracy=0.774230	lr=0.155686
Epoch[70] Batch [899]	Speed: 5081.869717 samples/sec	accuracy=0.774265	lr=0.155473
Epoch[70] Batch [949]	Speed: 5036.583105 samples/sec	accuracy=0.774359	lr=0.155260
Epoch[70] Batch [999]	Speed: 5122.543866 samples/sec	accuracy=0.774455	lr=0.155047
Epoch[70] Batch [1049]	Speed: 5039.246107 samples/sec	accuracy=0.774562	lr=0.154834
Epoch[70] Batch [1099]	Speed: 5047.297707 samples/sec	accuracy=0.774656	lr=0.154622
Epoch[70] Batch [1149]	Speed: 5059.767079 samples/sec	accuracy=0.774762	lr=0.154409
Epoch[70] Batch [1199]	Speed: 4945.986304 samples/sec	accuracy=0.774865	lr=0.154196
Epoch[70] Batch [1249]	Speed: 5205.133505 samples/sec	accuracy=0.774984	lr=0.153984
[Epoch 70] training: accuracy=0.774988
[Epoch 70] speed: 5038 samples/sec	time cost: 263.656223
[Epoch 70] validation: err-top1=0.313257 err-top5=0.108518
Epoch[71] Batch [49]	Speed: 4803.069085 samples/sec	accuracy=0.774796	lr=0.153767
Epoch[71] Batch [99]	Speed: 4965.014124 samples/sec	accuracy=0.773814	lr=0.153555
Epoch[71] Batch [149]	Speed: 5153.310036 samples/sec	accuracy=0.773597	lr=0.153342
Epoch[71] Batch [199]	Speed: 5065.153460 samples/sec	accuracy=0.773862	lr=0.153130
Epoch[71] Batch [249]	Speed: 5076.961046 samples/sec	accuracy=0.773797	lr=0.152918
Epoch[71] Batch [299]	Speed: 5074.768146 samples/sec	accuracy=0.773883	lr=0.152705
Epoch[71] Batch [349]	Speed: 4997.027701 samples/sec	accuracy=0.774083	lr=0.152493
Epoch[71] Batch [399]	Speed: 5024.405691 samples/sec	accuracy=0.774314	lr=0.152281
Epoch[71] Batch [449]	Speed: 4923.226987 samples/sec	accuracy=0.774290	lr=0.152069
Epoch[71] Batch [499]	Speed: 4999.681620 samples/sec	accuracy=0.774279	lr=0.151857
Epoch[71] Batch [549]	Speed: 5072.246339 samples/sec	accuracy=0.774378	lr=0.151645
Epoch[71] Batch [599]	Speed: 5009.423498 samples/sec	accuracy=0.774537	lr=0.151433
Epoch[71] Batch [649]	Speed: 5008.381723 samples/sec	accuracy=0.774702	lr=0.151222
Epoch[71] Batch [699]	Speed: 5053.559783 samples/sec	accuracy=0.774807	lr=0.151010
Epoch[71] Batch [749]	Speed: 5146.292409 samples/sec	accuracy=0.774910	lr=0.150798
Epoch[71] Batch [799]	Speed: 4977.624043 samples/sec	accuracy=0.774993	lr=0.150587
Epoch[71] Batch [849]	Speed: 4971.358847 samples/sec	accuracy=0.775187	lr=0.150375
Epoch[71] Batch [899]	Speed: 5092.947167 samples/sec	accuracy=0.775293	lr=0.150163
Epoch[71] Batch [949]	Speed: 5051.668549 samples/sec	accuracy=0.775344	lr=0.149952
Epoch[71] Batch [999]	Speed: 4970.056881 samples/sec	accuracy=0.775398	lr=0.149741
Epoch[71] Batch [1049]	Speed: 5046.882305 samples/sec	accuracy=0.775556	lr=0.149529
Epoch[71] Batch [1099]	Speed: 5094.955263 samples/sec	accuracy=0.775619	lr=0.149318
Epoch[71] Batch [1149]	Speed: 5115.521377 samples/sec	accuracy=0.775768	lr=0.149107
Epoch[71] Batch [1199]	Speed: 5001.432547 samples/sec	accuracy=0.775858	lr=0.148896
Epoch[71] Batch [1249]	Speed: 5149.663670 samples/sec	accuracy=0.775937	lr=0.148685
[Epoch 71] training: accuracy=0.775943
[Epoch 71] speed: 5028 samples/sec	time cost: 264.143998
[Epoch 71] validation: err-top1=0.326830 err-top5=0.117606
Epoch[72] Batch [49]	Speed: 4836.317170 samples/sec	accuracy=0.775955	lr=0.148469
Epoch[72] Batch [99]	Speed: 4985.052616 samples/sec	accuracy=0.776166	lr=0.148258
Epoch[72] Batch [149]	Speed: 5043.423083 samples/sec	accuracy=0.775969	lr=0.148047
Epoch[72] Batch [199]	Speed: 4840.936979 samples/sec	accuracy=0.775266	lr=0.147837
Epoch[72] Batch [249]	Speed: 5086.869223 samples/sec	accuracy=0.775257	lr=0.147626
Epoch[72] Batch [299]	Speed: 4814.473545 samples/sec	accuracy=0.775317	lr=0.147415
Epoch[72] Batch [349]	Speed: 5096.343695 samples/sec	accuracy=0.775579	lr=0.147204
Epoch[72] Batch [399]	Speed: 5105.462658 samples/sec	accuracy=0.775687	lr=0.146994
Epoch[72] Batch [449]	Speed: 5039.923653 samples/sec	accuracy=0.775724	lr=0.146783
Epoch[72] Batch [499]	Speed: 5098.159483 samples/sec	accuracy=0.775807	lr=0.146573
Epoch[72] Batch [549]	Speed: 4968.781695 samples/sec	accuracy=0.775958	lr=0.146362
Epoch[72] Batch [599]	Speed: 5023.685420 samples/sec	accuracy=0.776028	lr=0.146152
Epoch[72] Batch [649]	Speed: 4988.656668 samples/sec	accuracy=0.776180	lr=0.145942
Epoch[72] Batch [699]	Speed: 5181.515974 samples/sec	accuracy=0.776333	lr=0.145732
Epoch[72] Batch [749]	Speed: 4992.099865 samples/sec	accuracy=0.776422	lr=0.145521
Epoch[72] Batch [799]	Speed: 4829.511651 samples/sec	accuracy=0.776536	lr=0.145311
Epoch[72] Batch [849]	Speed: 4927.653913 samples/sec	accuracy=0.776688	lr=0.145101
Epoch[72] Batch [899]	Speed: 4920.338141 samples/sec	accuracy=0.776757	lr=0.144891
Epoch[72] Batch [949]	Speed: 4831.422996 samples/sec	accuracy=0.776818	lr=0.144682
Epoch[72] Batch [999]	Speed: 4778.902898 samples/sec	accuracy=0.776939	lr=0.144472
Epoch[72] Batch [1049]	Speed: 4975.821605 samples/sec	accuracy=0.776971	lr=0.144262
Epoch[72] Batch [1099]	Speed: 4887.652660 samples/sec	accuracy=0.777018	lr=0.144052
Epoch[72] Batch [1149]	Speed: 5066.850250 samples/sec	accuracy=0.777089	lr=0.143843
Epoch[72] Batch [1199]	Speed: 5050.835305 samples/sec	accuracy=0.777142	lr=0.143633
Epoch[72] Batch [1249]	Speed: 5053.771355 samples/sec	accuracy=0.777216	lr=0.143424
[Epoch 72] training: accuracy=0.777217
[Epoch 72] speed: 4969 samples/sec	time cost: 267.018403
[Epoch 72] validation: err-top1=0.308076 err-top5=0.105847
Epoch[73] Batch [49]	Speed: 4776.863177 samples/sec	accuracy=0.776701	lr=0.143210
Epoch[73] Batch [99]	Speed: 4992.747379 samples/sec	accuracy=0.775766	lr=0.143001
Epoch[73] Batch [149]	Speed: 4861.147696 samples/sec	accuracy=0.776061	lr=0.142791
Epoch[73] Batch [199]	Speed: 5034.508502 samples/sec	accuracy=0.776266	lr=0.142582
Epoch[73] Batch [249]	Speed: 4858.551166 samples/sec	accuracy=0.776162	lr=0.142373
Epoch[73] Batch [299]	Speed: 4994.206904 samples/sec	accuracy=0.776253	lr=0.142164
Epoch[73] Batch [349]	Speed: 5044.969161 samples/sec	accuracy=0.776482	lr=0.141955
Epoch[73] Batch [399]	Speed: 4964.655884 samples/sec	accuracy=0.776512	lr=0.141746
Epoch[73] Batch [449]	Speed: 4925.969508 samples/sec	accuracy=0.776735	lr=0.141537
Epoch[73] Batch [499]	Speed: 5018.578576 samples/sec	accuracy=0.776892	lr=0.141328
Epoch[73] Batch [549]	Speed: 5041.806575 samples/sec	accuracy=0.776977	lr=0.141120
Epoch[73] Batch [599]	Speed: 4931.612874 samples/sec	accuracy=0.776946	lr=0.140911
Epoch[73] Batch [649]	Speed: 4994.683379 samples/sec	accuracy=0.777075	lr=0.140702
Epoch[73] Batch [699]	Speed: 4886.874642 samples/sec	accuracy=0.777183	lr=0.140494
Epoch[73] Batch [749]	Speed: 4997.449938 samples/sec	accuracy=0.777242	lr=0.140285
Epoch[73] Batch [799]	Speed: 4964.315599 samples/sec	accuracy=0.777260	lr=0.140077
Epoch[73] Batch [849]	Speed: 5183.548241 samples/sec	accuracy=0.777332	lr=0.139869
Epoch[73] Batch [899]	Speed: 5242.454179 samples/sec	accuracy=0.777522	lr=0.139661
Epoch[73] Batch [949]	Speed: 5255.002043 samples/sec	accuracy=0.777641	lr=0.139452
Epoch[73] Batch [999]	Speed: 5037.372541 samples/sec	accuracy=0.777724	lr=0.139244
Epoch[73] Batch [1049]	Speed: 5099.255655 samples/sec	accuracy=0.777893	lr=0.139036
Epoch[73] Batch [1099]	Speed: 4897.544746 samples/sec	accuracy=0.777936	lr=0.138828
Epoch[73] Batch [1149]	Speed: 4877.517168 samples/sec	accuracy=0.778052	lr=0.138620
Epoch[73] Batch [1199]	Speed: 5028.233242 samples/sec	accuracy=0.778159	lr=0.138413
Epoch[73] Batch [1249]	Speed: 5086.492220 samples/sec	accuracy=0.778262	lr=0.138205
[Epoch 73] training: accuracy=0.778262
[Epoch 73] speed: 4993 samples/sec	time cost: 265.739505
[Epoch 73] validation: err-top1=0.304787 err-top5=0.103396
Epoch[74] Batch [49]	Speed: 4806.379859 samples/sec	accuracy=0.777220	lr=0.137993
Epoch[74] Batch [99]	Speed: 4964.509550 samples/sec	accuracy=0.777259	lr=0.137786
Epoch[74] Batch [149]	Speed: 4798.374131 samples/sec	accuracy=0.777898	lr=0.137578
Epoch[74] Batch [199]	Speed: 4871.949001 samples/sec	accuracy=0.777863	lr=0.137371
Epoch[74] Batch [249]	Speed: 5026.839661 samples/sec	accuracy=0.777694	lr=0.137163
Epoch[74] Batch [299]	Speed: 4882.985269 samples/sec	accuracy=0.777605	lr=0.136956
Epoch[74] Batch [349]	Speed: 5068.577845 samples/sec	accuracy=0.777590	lr=0.136749
Epoch[74] Batch [399]	Speed: 5070.717624 samples/sec	accuracy=0.777853	lr=0.136542
Epoch[74] Batch [449]	Speed: 4950.980420 samples/sec	accuracy=0.777967	lr=0.136335
Epoch[74] Batch [499]	Speed: 4774.751093 samples/sec	accuracy=0.777970	lr=0.136128
Epoch[74] Batch [549]	Speed: 5125.496006 samples/sec	accuracy=0.778178	lr=0.135921
Epoch[74] Batch [599]	Speed: 5108.542543 samples/sec	accuracy=0.778311	lr=0.135714
Epoch[74] Batch [649]	Speed: 4947.938416 samples/sec	accuracy=0.778463	lr=0.135507
Epoch[74] Batch [699]	Speed: 4882.303639 samples/sec	accuracy=0.778636	lr=0.135301
Epoch[74] Batch [749]	Speed: 5000.916119 samples/sec	accuracy=0.778767	lr=0.135094
Epoch[74] Batch [799]	Speed: 5188.146824 samples/sec	accuracy=0.778841	lr=0.134887
Epoch[74] Batch [849]	Speed: 5027.743400 samples/sec	accuracy=0.778886	lr=0.134681
Epoch[74] Batch [899]	Speed: 5111.282203 samples/sec	accuracy=0.779045	lr=0.134475
Epoch[74] Batch [949]	Speed: 5027.107136 samples/sec	accuracy=0.779069	lr=0.134268
Epoch[74] Batch [999]	Speed: 4876.155156 samples/sec	accuracy=0.779180	lr=0.134062
Epoch[74] Batch [1049]	Speed: 5068.631440 samples/sec	accuracy=0.779334	lr=0.133856
Epoch[74] Batch [1099]	Speed: 4796.752613 samples/sec	accuracy=0.779406	lr=0.133650
Epoch[74] Batch [1149]	Speed: 4992.954702 samples/sec	accuracy=0.779497	lr=0.133444
Epoch[74] Batch [1199]	Speed: 4959.927289 samples/sec	accuracy=0.779580	lr=0.133238
Epoch[74] Batch [1249]	Speed: 5178.127490 samples/sec	accuracy=0.779668	lr=0.133032
[Epoch 74] training: accuracy=0.779671
[Epoch 74] speed: 4973 samples/sec	time cost: 266.910525
[Epoch 74] validation: err-top1=0.310587 err-top5=0.105907
Epoch[75] Batch [49]	Speed: 4767.686989 samples/sec	accuracy=0.780365	lr=0.132822
Epoch[75] Batch [99]	Speed: 4858.059756 samples/sec	accuracy=0.780265	lr=0.132617
Epoch[75] Batch [149]	Speed: 4952.105221 samples/sec	accuracy=0.780416	lr=0.132411
Epoch[75] Batch [199]	Speed: 5029.698748 samples/sec	accuracy=0.779913	lr=0.132206
Epoch[75] Batch [249]	Speed: 5038.000533 samples/sec	accuracy=0.779610	lr=0.132000
Epoch[75] Batch [299]	Speed: 5176.084999 samples/sec	accuracy=0.779208	lr=0.131795
Epoch[75] Batch [349]	Speed: 5101.732160 samples/sec	accuracy=0.779168	lr=0.131590
Epoch[75] Batch [399]	Speed: 5017.871230 samples/sec	accuracy=0.779121	lr=0.131385
Epoch[75] Batch [449]	Speed: 5166.895852 samples/sec	accuracy=0.779147	lr=0.131180
Epoch[75] Batch [499]	Speed: 5116.931893 samples/sec	accuracy=0.779312	lr=0.130975
Epoch[75] Batch [549]	Speed: 5084.386900 samples/sec	accuracy=0.779299	lr=0.130770
Epoch[75] Batch [599]	Speed: 5134.241898 samples/sec	accuracy=0.779330	lr=0.130565
Epoch[75] Batch [649]	Speed: 5174.450794 samples/sec	accuracy=0.779447	lr=0.130360
Epoch[75] Batch [699]	Speed: 5051.515496 samples/sec	accuracy=0.779518	lr=0.130155
Epoch[75] Batch [749]	Speed: 4877.730986 samples/sec	accuracy=0.779692	lr=0.129951
Epoch[75] Batch [799]	Speed: 4878.352827 samples/sec	accuracy=0.779862	lr=0.129746
Epoch[75] Batch [849]	Speed: 5088.743982 samples/sec	accuracy=0.779932	lr=0.129542
Epoch[75] Batch [899]	Speed: 5089.760712 samples/sec	accuracy=0.779974	lr=0.129338
Epoch[75] Batch [949]	Speed: 4974.582408 samples/sec	accuracy=0.780078	lr=0.129133
Epoch[75] Batch [999]	Speed: 4839.497378 samples/sec	accuracy=0.780200	lr=0.128929
Epoch[75] Batch [1049]	Speed: 4913.895984 samples/sec	accuracy=0.780348	lr=0.128725
Epoch[75] Batch [1099]	Speed: 4741.531881 samples/sec	accuracy=0.780532	lr=0.128521
Epoch[75] Batch [1149]	Speed: 5125.660303 samples/sec	accuracy=0.780637	lr=0.128317
Epoch[75] Batch [1199]	Speed: 4982.940690 samples/sec	accuracy=0.780780	lr=0.128113
Epoch[75] Batch [1249]	Speed: 5114.197867 samples/sec	accuracy=0.780910	lr=0.127910
[Epoch 75] training: accuracy=0.780914
[Epoch 75] speed: 5004 samples/sec	time cost: 265.259938
[Epoch 75] validation: err-top1=0.310974 err-top5=0.108988
Epoch[76] Batch [49]	Speed: 4691.569601 samples/sec	accuracy=0.781520	lr=0.127702
Epoch[76] Batch [99]	Speed: 4948.640892 samples/sec	accuracy=0.781064	lr=0.127498
Epoch[76] Batch [149]	Speed: 5094.187314 samples/sec	accuracy=0.780758	lr=0.127295
Epoch[76] Batch [199]	Speed: 5066.024539 samples/sec	accuracy=0.780502	lr=0.127092
Epoch[76] Batch [249]	Speed: 4807.582620 samples/sec	accuracy=0.780759	lr=0.126888
Epoch[76] Batch [299]	Speed: 5008.482645 samples/sec	accuracy=0.780482	lr=0.126685
Epoch[76] Batch [349]	Speed: 4955.925158 samples/sec	accuracy=0.780530	lr=0.126482
Epoch[76] Batch [399]	Speed: 4875.071451 samples/sec	accuracy=0.780567	lr=0.126279
Epoch[76] Batch [449]	Speed: 4941.673764 samples/sec	accuracy=0.780595	lr=0.126076
Epoch[76] Batch [499]	Speed: 5018.643903 samples/sec	accuracy=0.780787	lr=0.125873
Epoch[76] Batch [549]	Speed: 5128.644331 samples/sec	accuracy=0.780830	lr=0.125670
Epoch[76] Batch [599]	Speed: 5012.775258 samples/sec	accuracy=0.780833	lr=0.125468
Epoch[76] Batch [649]	Speed: 4868.844126 samples/sec	accuracy=0.780891	lr=0.125265
Epoch[76] Batch [699]	Speed: 4913.887888 samples/sec	accuracy=0.780943	lr=0.125062
Epoch[76] Batch [749]	Speed: 4935.252778 samples/sec	accuracy=0.780927	lr=0.124860
Epoch[76] Batch [799]	Speed: 5001.691034 samples/sec	accuracy=0.781007	lr=0.124658
Epoch[76] Batch [849]	Speed: 5045.624536 samples/sec	accuracy=0.781134	lr=0.124455
Epoch[76] Batch [899]	Speed: 4920.561705 samples/sec	accuracy=0.781338	lr=0.124253
Epoch[76] Batch [949]	Speed: 5054.712764 samples/sec	accuracy=0.781439	lr=0.124051
Epoch[76] Batch [999]	Speed: 4942.827332 samples/sec	accuracy=0.781540	lr=0.123849
Epoch[76] Batch [1049]	Speed: 4920.573543 samples/sec	accuracy=0.781585	lr=0.123647
Epoch[76] Batch [1099]	Speed: 4889.178939 samples/sec	accuracy=0.781675	lr=0.123446
Epoch[76] Batch [1149]	Speed: 4847.278523 samples/sec	accuracy=0.781755	lr=0.123244
Epoch[76] Batch [1199]	Speed: 4859.805917 samples/sec	accuracy=0.781883	lr=0.123042
Epoch[76] Batch [1249]	Speed: 5104.260934 samples/sec	accuracy=0.781968	lr=0.122841
[Epoch 76] training: accuracy=0.781975
[Epoch 76] speed: 4947 samples/sec	time cost: 268.429030
[Epoch 76] validation: err-top1=0.303272 err-top5=0.099868
Epoch[77] Batch [49]	Speed: 4876.418019 samples/sec	accuracy=0.782357	lr=0.122635
Epoch[77] Batch [99]	Speed: 4913.789392 samples/sec	accuracy=0.781737	lr=0.122434
Epoch[77] Batch [149]	Speed: 4783.748417 samples/sec	accuracy=0.781228	lr=0.122233
Epoch[77] Batch [199]	Speed: 5015.702007 samples/sec	accuracy=0.781262	lr=0.122032
Epoch[77] Batch [249]	Speed: 4982.455355 samples/sec	accuracy=0.781433	lr=0.121831
Epoch[77] Batch [299]	Speed: 4832.268377 samples/sec	accuracy=0.781480	lr=0.121630
Epoch[77] Batch [349]	Speed: 5037.340755 samples/sec	accuracy=0.781462	lr=0.121429
Epoch[77] Batch [399]	Speed: 4969.970499 samples/sec	accuracy=0.781534	lr=0.121228
Epoch[77] Batch [449]	Speed: 4864.816973 samples/sec	accuracy=0.781814	lr=0.121027
Epoch[77] Batch [499]	Speed: 4854.483951 samples/sec	accuracy=0.781904	lr=0.120827
Epoch[77] Batch [549]	Speed: 4955.774534 samples/sec	accuracy=0.781801	lr=0.120626
Epoch[77] Batch [599]	Speed: 5082.544219 samples/sec	accuracy=0.781790	lr=0.120426
Epoch[77] Batch [649]	Speed: 4893.937813 samples/sec	accuracy=0.781972	lr=0.120226
Epoch[77] Batch [699]	Speed: 4987.985421 samples/sec	accuracy=0.782114	lr=0.120025
Epoch[77] Batch [749]	Speed: 4967.463613 samples/sec	accuracy=0.782222	lr=0.119825
Epoch[77] Batch [799]	Speed: 5006.772304 samples/sec	accuracy=0.782269	lr=0.119625
Epoch[77] Batch [849]	Speed: 5042.945317 samples/sec	accuracy=0.782296	lr=0.119425
Epoch[77] Batch [899]	Speed: 4914.011013 samples/sec	accuracy=0.782329	lr=0.119226
Epoch[77] Batch [949]	Speed: 5007.701304 samples/sec	accuracy=0.782454	lr=0.119026
Epoch[77] Batch [999]	Speed: 4919.222427 samples/sec	accuracy=0.782511	lr=0.118826
Epoch[77] Batch [1049]	Speed: 5015.778622 samples/sec	accuracy=0.782624	lr=0.118627
Epoch[77] Batch [1099]	Speed: 5060.470188 samples/sec	accuracy=0.782744	lr=0.118427
Epoch[77] Batch [1149]	Speed: 4947.627548 samples/sec	accuracy=0.782883	lr=0.118228
Epoch[77] Batch [1199]	Speed: 4963.743818 samples/sec	accuracy=0.782926	lr=0.118029
Epoch[77] Batch [1249]	Speed: 5115.062993 samples/sec	accuracy=0.783013	lr=0.117830
[Epoch 77] training: accuracy=0.783013
[Epoch 77] speed: 4954 samples/sec	time cost: 267.806011
[Epoch 77] validation: err-top1=0.301997 err-top5=0.100845
Epoch[78] Batch [49]	Speed: 4867.747888 samples/sec	accuracy=0.783136	lr=0.117627
Epoch[78] Batch [99]	Speed: 5023.076735 samples/sec	accuracy=0.782810	lr=0.117428
Epoch[78] Batch [149]	Speed: 4626.418540 samples/sec	accuracy=0.783074	lr=0.117229
Epoch[78] Batch [199]	Speed: 5040.610964 samples/sec	accuracy=0.783068	lr=0.117030
Epoch[78] Batch [249]	Speed: 4951.723608 samples/sec	accuracy=0.783037	lr=0.116831
Epoch[78] Batch [299]	Speed: 4920.416042 samples/sec	accuracy=0.782938	lr=0.116633
Epoch[78] Batch [349]	Speed: 4926.570593 samples/sec	accuracy=0.782933	lr=0.116434
Epoch[78] Batch [399]	Speed: 5026.267622 samples/sec	accuracy=0.783052	lr=0.116236
Epoch[78] Batch [449]	Speed: 4805.528564 samples/sec	accuracy=0.783124	lr=0.116038
Epoch[78] Batch [499]	Speed: 5002.556036 samples/sec	accuracy=0.783013	lr=0.115840
Epoch[78] Batch [549]	Speed: 4861.472334 samples/sec	accuracy=0.783134	lr=0.115642
Epoch[78] Batch [599]	Speed: 4974.650627 samples/sec	accuracy=0.783225	lr=0.115444
Epoch[78] Batch [649]	Speed: 4941.858331 samples/sec	accuracy=0.783298	lr=0.115246
Epoch[78] Batch [699]	Speed: 4984.544424 samples/sec	accuracy=0.783409	lr=0.115048
Epoch[78] Batch [749]	Speed: 4884.120483 samples/sec	accuracy=0.783426	lr=0.114850
Epoch[78] Batch [799]	Speed: 4883.493396 samples/sec	accuracy=0.783466	lr=0.114653
Epoch[78] Batch [849]	Speed: 5020.532200 samples/sec	accuracy=0.783590	lr=0.114455
Epoch[78] Batch [899]	Speed: 5092.010782 samples/sec	accuracy=0.783709	lr=0.114258
Epoch[78] Batch [949]	Speed: 4883.807364 samples/sec	accuracy=0.783790	lr=0.114061
Epoch[78] Batch [999]	Speed: 4933.708597 samples/sec	accuracy=0.783846	lr=0.113864
Epoch[78] Batch [1049]	Speed: 4821.793676 samples/sec	accuracy=0.783876	lr=0.113667
Epoch[78] Batch [1099]	Speed: 4926.690512 samples/sec	accuracy=0.783999	lr=0.113470
Epoch[78] Batch [1149]	Speed: 4956.226432 samples/sec	accuracy=0.784114	lr=0.113273
Epoch[78] Batch [1199]	Speed: 4873.159371 samples/sec	accuracy=0.784223	lr=0.113076
Epoch[78] Batch [1249]	Speed: 5169.402920 samples/sec	accuracy=0.784308	lr=0.112880
[Epoch 78] training: accuracy=0.784316
[Epoch 78] speed: 4929 samples/sec	time cost: 269.655592
[Epoch 78] validation: err-top1=0.293886 err-top5=0.098892
Epoch[79] Batch [49]	Speed: 4657.670231 samples/sec	accuracy=0.784224	lr=0.112679
Epoch[79] Batch [99]	Speed: 5135.222370 samples/sec	accuracy=0.783961	lr=0.112483
Epoch[79] Batch [149]	Speed: 4698.119200 samples/sec	accuracy=0.783763	lr=0.112287
Epoch[79] Batch [199]	Speed: 5017.294549 samples/sec	accuracy=0.783544	lr=0.112090
Epoch[79] Batch [249]	Speed: 4859.888622 samples/sec	accuracy=0.783336	lr=0.111894
Epoch[79] Batch [299]	Speed: 4997.509599 samples/sec	accuracy=0.783222	lr=0.111698
Epoch[79] Batch [349]	Speed: 4971.149285 samples/sec	accuracy=0.783230	lr=0.111502
Epoch[79] Batch [399]	Speed: 5156.784363 samples/sec	accuracy=0.783229	lr=0.111307
Epoch[79] Batch [449]	Speed: 4892.379137 samples/sec	accuracy=0.783362	lr=0.111111
Epoch[79] Batch [499]	Speed: 5175.363492 samples/sec	accuracy=0.783438	lr=0.110915
Epoch[79] Batch [549]	Speed: 5075.151329 samples/sec	accuracy=0.783607	lr=0.110720
Epoch[79] Batch [599]	Speed: 5029.522404 samples/sec	accuracy=0.783787	lr=0.110525
Epoch[79] Batch [649]	Speed: 4920.538141 samples/sec	accuracy=0.783975	lr=0.110329
Epoch[79] Batch [699]	Speed: 5135.384099 samples/sec	accuracy=0.784075	lr=0.110134
Epoch[79] Batch [749]	Speed: 5130.240534 samples/sec	accuracy=0.784189	lr=0.109939
Epoch[79] Batch [799]	Speed: 4958.937143 samples/sec	accuracy=0.784341	lr=0.109744
Epoch[79] Batch [849]	Speed: 5333.673419 samples/sec	accuracy=0.784472	lr=0.109549
Epoch[79] Batch [899]	Speed: 5048.890912 samples/sec	accuracy=0.784677	lr=0.109355
Epoch[79] Batch [949]	Speed: 4939.313029 samples/sec	accuracy=0.784804	lr=0.109160
Epoch[79] Batch [999]	Speed: 4899.214906 samples/sec	accuracy=0.784939	lr=0.108966
Epoch[79] Batch [1049]	Speed: 5214.046805 samples/sec	accuracy=0.785076	lr=0.108771
Epoch[79] Batch [1099]	Speed: 5327.243587 samples/sec	accuracy=0.785173	lr=0.108577
Epoch[79] Batch [1149]	Speed: 5199.365273 samples/sec	accuracy=0.785274	lr=0.108383
Epoch[79] Batch [1199]	Speed: 5247.755057 samples/sec	accuracy=0.785360	lr=0.108189
Epoch[79] Batch [1249]	Speed: 4991.552411 samples/sec	accuracy=0.785489	lr=0.107995
[Epoch 79] training: accuracy=0.785491
[Epoch 79] speed: 5029 samples/sec	time cost: 264.080080
[Epoch 79] validation: err-top1=0.293447 err-top5=0.098772
Epoch[80] Batch [49]	Speed: 4773.773535 samples/sec	accuracy=0.785415	lr=0.107797
Epoch[80] Batch [99]	Speed: 5157.097054 samples/sec	accuracy=0.785803	lr=0.107603
Epoch[80] Batch [149]	Speed: 5255.967050 samples/sec	accuracy=0.785345	lr=0.107410
Epoch[80] Batch [199]	Speed: 4942.020051 samples/sec	accuracy=0.785275	lr=0.107216
Epoch[80] Batch [249]	Speed: 4879.174692 samples/sec	accuracy=0.785512	lr=0.107023
Epoch[80] Batch [299]	Speed: 5173.357577 samples/sec	accuracy=0.785506	lr=0.106830
Epoch[80] Batch [349]	Speed: 5108.279577 samples/sec	accuracy=0.785455	lr=0.106636
Epoch[80] Batch [399]	Speed: 4996.278176 samples/sec	accuracy=0.785495	lr=0.106443
Epoch[80] Batch [449]	Speed: 5069.458479 samples/sec	accuracy=0.785498	lr=0.106250
Epoch[80] Batch [499]	Speed: 5036.692137 samples/sec	accuracy=0.785510	lr=0.106057
Epoch[80] Batch [549]	Speed: 4925.100966 samples/sec	accuracy=0.785600	lr=0.105865
Epoch[80] Batch [599]	Speed: 5108.121130 samples/sec	accuracy=0.785739	lr=0.105672
Epoch[80] Batch [649]	Speed: 4975.024833 samples/sec	accuracy=0.785778	lr=0.105480
Epoch[80] Batch [699]	Speed: 5087.819026 samples/sec	accuracy=0.785858	lr=0.105287
Epoch[80] Batch [749]	Speed: 4985.826905 samples/sec	accuracy=0.786039	lr=0.105095
Epoch[80] Batch [799]	Speed: 5004.911138 samples/sec	accuracy=0.786148	lr=0.104903
Epoch[80] Batch [849]	Speed: 5008.636490 samples/sec	accuracy=0.786200	lr=0.104711
Epoch[80] Batch [899]	Speed: 4966.116830 samples/sec	accuracy=0.786234	lr=0.104519
Epoch[80] Batch [949]	Speed: 4996.515554 samples/sec	accuracy=0.786289	lr=0.104327
Epoch[80] Batch [999]	Speed: 5069.916386 samples/sec	accuracy=0.786305	lr=0.104135
Epoch[80] Batch [1049]	Speed: 5117.901739 samples/sec	accuracy=0.786354	lr=0.103944
Epoch[80] Batch [1099]	Speed: 5048.687700 samples/sec	accuracy=0.786459	lr=0.103752
Epoch[80] Batch [1149]	Speed: 5212.559979 samples/sec	accuracy=0.786512	lr=0.103561
Epoch[80] Batch [1199]	Speed: 5090.882602 samples/sec	accuracy=0.786547	lr=0.103370
Epoch[80] Batch [1249]	Speed: 5096.803448 samples/sec	accuracy=0.786632	lr=0.103178
[Epoch 80] training: accuracy=0.786635
[Epoch 80] speed: 5036 samples/sec	time cost: 263.475036
[Epoch 80] validation: err-top1=0.290398 err-top5=0.095683
Epoch[81] Batch [49]	Speed: 4833.396012 samples/sec	accuracy=0.785348	lr=0.102984
Epoch[81] Batch [99]	Speed: 4995.358406 samples/sec	accuracy=0.785783	lr=0.102793
Epoch[81] Batch [149]	Speed: 5105.630287 samples/sec	accuracy=0.785469	lr=0.102602
Epoch[81] Batch [199]	Speed: 5013.691622 samples/sec	accuracy=0.785305	lr=0.102411
Epoch[81] Batch [249]	Speed: 4987.213356 samples/sec	accuracy=0.785635	lr=0.102221
Epoch[81] Batch [299]	Speed: 5170.485752 samples/sec	accuracy=0.785640	lr=0.102030
Epoch[81] Batch [349]	Speed: 5120.401887 samples/sec	accuracy=0.785726	lr=0.101840
Epoch[81] Batch [399]	Speed: 4940.173635 samples/sec	accuracy=0.785828	lr=0.101650
Epoch[81] Batch [449]	Speed: 4978.978345 samples/sec	accuracy=0.785812	lr=0.101460
Epoch[81] Batch [499]	Speed: 4997.733951 samples/sec	accuracy=0.785848	lr=0.101270
Epoch[81] Batch [549]	Speed: 4948.380676 samples/sec	accuracy=0.785997	lr=0.101080
Epoch[81] Batch [599]	Speed: 5161.216972 samples/sec	accuracy=0.786070	lr=0.100890
Epoch[81] Batch [649]	Speed: 4985.419129 samples/sec	accuracy=0.786166	lr=0.100701
Epoch[81] Batch [699]	Speed: 5044.360285 samples/sec	accuracy=0.786236	lr=0.100511
Epoch[81] Batch [749]	Speed: 5046.831541 samples/sec	accuracy=0.786291	lr=0.100322
Epoch[81] Batch [799]	Speed: 5226.329658 samples/sec	accuracy=0.786338	lr=0.100132
Epoch[81] Batch [849]	Speed: 5088.744344 samples/sec	accuracy=0.786413	lr=0.099943
Epoch[81] Batch [899]	Speed: 5120.145390 samples/sec	accuracy=0.786500	lr=0.099754
Epoch[81] Batch [949]	Speed: 5060.766776 samples/sec	accuracy=0.786633	lr=0.099565
Epoch[81] Batch [999]	Speed: 5084.263636 samples/sec	accuracy=0.786775	lr=0.099377
Epoch[81] Batch [1049]	Speed: 5049.332526 samples/sec	accuracy=0.786940	lr=0.099188
Epoch[81] Batch [1099]	Speed: 5079.025023 samples/sec	accuracy=0.787031	lr=0.098999
Epoch[81] Batch [1149]	Speed: 4948.188895 samples/sec	accuracy=0.787147	lr=0.098811
Epoch[81] Batch [1199]	Speed: 5027.562250 samples/sec	accuracy=0.787294	lr=0.098623
Epoch[81] Batch [1249]	Speed: 5141.770768 samples/sec	accuracy=0.787397	lr=0.098434
[Epoch 81] training: accuracy=0.787401
[Epoch 81] speed: 5040 samples/sec	time cost: 263.594021
[Epoch 81] validation: err-top1=0.283040 err-top5=0.092265
Epoch[82] Batch [49]	Speed: 4931.691246 samples/sec	accuracy=0.786964	lr=0.098243
Epoch[82] Batch [99]	Speed: 5056.527458 samples/sec	accuracy=0.787551	lr=0.098055
Epoch[82] Batch [149]	Speed: 5062.197253 samples/sec	accuracy=0.787462	lr=0.097867
Epoch[82] Batch [199]	Speed: 5017.227616 samples/sec	accuracy=0.787274	lr=0.097679
Epoch[82] Batch [249]	Speed: 5062.995813 samples/sec	accuracy=0.787077	lr=0.097492
Epoch[82] Batch [299]	Speed: 5138.315530 samples/sec	accuracy=0.787086	lr=0.097304
Epoch[82] Batch [349]	Speed: 5086.486196 samples/sec	accuracy=0.787081	lr=0.097117
Epoch[82] Batch [399]	Speed: 5113.683583 samples/sec	accuracy=0.787202	lr=0.096930
Epoch[82] Batch [449]	Speed: 5074.189703 samples/sec	accuracy=0.787317	lr=0.096743
Epoch[82] Batch [499]	Speed: 5098.277854 samples/sec	accuracy=0.787479	lr=0.096556
Epoch[82] Batch [549]	Speed: 5057.261704 samples/sec	accuracy=0.787511	lr=0.096369
Epoch[82] Batch [599]	Speed: 5068.156301 samples/sec	accuracy=0.787529	lr=0.096182
Epoch[82] Batch [649]	Speed: 5102.187673 samples/sec	accuracy=0.787551	lr=0.095996
Epoch[82] Batch [699]	Speed: 5099.469497 samples/sec	accuracy=0.787627	lr=0.095809
Epoch[82] Batch [749]	Speed: 4976.910893 samples/sec	accuracy=0.787751	lr=0.095623
Epoch[82] Batch [799]	Speed: 4996.686335 samples/sec	accuracy=0.787781	lr=0.095437
Epoch[82] Batch [849]	Speed: 5136.830904 samples/sec	accuracy=0.787863	lr=0.095250
Epoch[82] Batch [899]	Speed: 5129.241872 samples/sec	accuracy=0.787901	lr=0.095064
Epoch[82] Batch [949]	Speed: 5063.967172 samples/sec	accuracy=0.787981	lr=0.094879
Epoch[82] Batch [999]	Speed: 5019.773730 samples/sec	accuracy=0.788020	lr=0.094693
Epoch[82] Batch [1049]	Speed: 4821.306534 samples/sec	accuracy=0.788172	lr=0.094507
Epoch[82] Batch [1099]	Speed: 5083.023502 samples/sec	accuracy=0.788301	lr=0.094322
Epoch[82] Batch [1149]	Speed: 4990.428525 samples/sec	accuracy=0.788377	lr=0.094137
Epoch[82] Batch [1199]	Speed: 4985.894392 samples/sec	accuracy=0.788501	lr=0.093951
Epoch[82] Batch [1249]	Speed: 5134.710847 samples/sec	accuracy=0.788638	lr=0.093766
[Epoch 82] training: accuracy=0.788642
[Epoch 82] speed: 5047 samples/sec	time cost: 263.295603
[Epoch 82] validation: err-top1=0.284479 err-top5=0.091996
Epoch[83] Batch [49]	Speed: 4995.104407 samples/sec	accuracy=0.788996	lr=0.093578
Epoch[83] Batch [99]	Speed: 5083.007982 samples/sec	accuracy=0.789541	lr=0.093393
Epoch[83] Batch [149]	Speed: 4944.965507 samples/sec	accuracy=0.788888	lr=0.093208
Epoch[83] Batch [199]	Speed: 5043.051663 samples/sec	accuracy=0.788750	lr=0.093024
Epoch[83] Batch [249]	Speed: 5021.299352 samples/sec	accuracy=0.788873	lr=0.092839
Epoch[83] Batch [299]	Speed: 4944.002496 samples/sec	accuracy=0.788896	lr=0.092655
Epoch[83] Batch [349]	Speed: 5142.953025 samples/sec	accuracy=0.788906	lr=0.092471
Epoch[83] Batch [399]	Speed: 5077.489338 samples/sec	accuracy=0.788967	lr=0.092287
Epoch[83] Batch [449]	Speed: 5080.756725 samples/sec	accuracy=0.788955	lr=0.092103
Epoch[83] Batch [499]	Speed: 4878.000889 samples/sec	accuracy=0.789058	lr=0.091919
Epoch[83] Batch [549]	Speed: 4927.849307 samples/sec	accuracy=0.788996	lr=0.091735
Epoch[83] Batch [599]	Speed: 5059.900961 samples/sec	accuracy=0.788973	lr=0.091552
Epoch[83] Batch [649]	Speed: 5034.176866 samples/sec	accuracy=0.789044	lr=0.091368
Epoch[83] Batch [699]	Speed: 4929.909690 samples/sec	accuracy=0.789000	lr=0.091185
Epoch[83] Batch [749]	Speed: 4971.986948 samples/sec	accuracy=0.789075	lr=0.091002
Epoch[83] Batch [799]	Speed: 5066.328352 samples/sec	accuracy=0.789099	lr=0.090819
Epoch[83] Batch [849]	Speed: 5065.625048 samples/sec	accuracy=0.789124	lr=0.090636
Epoch[83] Batch [899]	Speed: 5020.730099 samples/sec	accuracy=0.789180	lr=0.090453
Epoch[83] Batch [949]	Speed: 5005.464909 samples/sec	accuracy=0.789253	lr=0.090270
Epoch[83] Batch [999]	Speed: 5053.897546 samples/sec	accuracy=0.789366	lr=0.090088
Epoch[83] Batch [1049]	Speed: 5018.423651 samples/sec	accuracy=0.789502	lr=0.089905
Epoch[83] Batch [1099]	Speed: 5059.281562 samples/sec	accuracy=0.789584	lr=0.089723
Epoch[83] Batch [1149]	Speed: 4864.423681 samples/sec	accuracy=0.789679	lr=0.089541
Epoch[83] Batch [1199]	Speed: 5036.160962 samples/sec	accuracy=0.789724	lr=0.089359
Epoch[83] Batch [1249]	Speed: 5147.556454 samples/sec	accuracy=0.789809	lr=0.089177
[Epoch 83] training: accuracy=0.789809
[Epoch 83] speed: 5013 samples/sec	time cost: 264.703319
[Epoch 83] validation: err-top1=0.276088 err-top5=0.088010
Epoch[84] Batch [49]	Speed: 4737.022232 samples/sec	accuracy=0.790785	lr=0.088992
Epoch[84] Batch [99]	Speed: 4900.120741 samples/sec	accuracy=0.790107	lr=0.088810
Epoch[84] Batch [149]	Speed: 4902.086492 samples/sec	accuracy=0.789739	lr=0.088629
Epoch[84] Batch [199]	Speed: 4807.015704 samples/sec	accuracy=0.789213	lr=0.088448
Epoch[84] Batch [249]	Speed: 4999.688022 samples/sec	accuracy=0.789158	lr=0.088266
Epoch[84] Batch [299]	Speed: 5126.081067 samples/sec	accuracy=0.789069	lr=0.088085
Epoch[84] Batch [349]	Speed: 4819.388258 samples/sec	accuracy=0.789269	lr=0.087904
Epoch[84] Batch [399]	Speed: 5136.896273 samples/sec	accuracy=0.789215	lr=0.087724
Epoch[84] Batch [449]	Speed: 5000.059601 samples/sec	accuracy=0.789347	lr=0.087543
Epoch[84] Batch [499]	Speed: 5153.719273 samples/sec	accuracy=0.789415	lr=0.087363
Epoch[84] Batch [549]	Speed: 5036.705486 samples/sec	accuracy=0.789366	lr=0.087182
Epoch[84] Batch [599]	Speed: 5062.179234 samples/sec	accuracy=0.789534	lr=0.087002
Epoch[84] Batch [649]	Speed: 4904.742210 samples/sec	accuracy=0.789546	lr=0.086822
Epoch[84] Batch [699]	Speed: 4955.086267 samples/sec	accuracy=0.789738	lr=0.086642
Epoch[84] Batch [749]	Speed: 5194.984425 samples/sec	accuracy=0.789785	lr=0.086462
Epoch[84] Batch [799]	Speed: 5062.784900 samples/sec	accuracy=0.789938	lr=0.086282
Epoch[84] Batch [849]	Speed: 4943.476351 samples/sec	accuracy=0.790036	lr=0.086103
Epoch[84] Batch [899]	Speed: 4959.627397 samples/sec	accuracy=0.790132	lr=0.085923
Epoch[84] Batch [949]	Speed: 5051.872476 samples/sec	accuracy=0.790244	lr=0.085744
Epoch[84] Batch [999]	Speed: 4819.860085 samples/sec	accuracy=0.790337	lr=0.085565
Epoch[84] Batch [1049]	Speed: 4999.323248 samples/sec	accuracy=0.790480	lr=0.085386
Epoch[84] Batch [1099]	Speed: 5035.064475 samples/sec	accuracy=0.790572	lr=0.085207
Epoch[84] Batch [1149]	Speed: 5012.562658 samples/sec	accuracy=0.790682	lr=0.085028
Epoch[84] Batch [1199]	Speed: 5071.712789 samples/sec	accuracy=0.790737	lr=0.084850
Epoch[84] Batch [1249]	Speed: 5179.644207 samples/sec	accuracy=0.790861	lr=0.084671
[Epoch 84] training: accuracy=0.790864
[Epoch 84] speed: 4987 samples/sec	time cost: 266.245262
[Epoch 84] validation: err-top1=0.284239 err-top5=0.090840
Epoch[85] Batch [49]	Speed: 4797.146289 samples/sec	accuracy=0.790661	lr=0.084489
Epoch[85] Batch [99]	Speed: 5063.983292 samples/sec	accuracy=0.790013	lr=0.084311
Epoch[85] Batch [149]	Speed: 5003.131315 samples/sec	accuracy=0.790279	lr=0.084133
Epoch[85] Batch [199]	Speed: 5100.928119 samples/sec	accuracy=0.790097	lr=0.083955
Epoch[85] Batch [249]	Speed: 5088.140649 samples/sec	accuracy=0.789870	lr=0.083777
Epoch[85] Batch [299]	Speed: 4971.652908 samples/sec	accuracy=0.790129	lr=0.083600
Epoch[85] Batch [349]	Speed: 5131.915238 samples/sec	accuracy=0.790183	lr=0.083422
Epoch[85] Batch [399]	Speed: 5062.137231 samples/sec	accuracy=0.790244	lr=0.083245
Epoch[85] Batch [449]	Speed: 5037.875016 samples/sec	accuracy=0.790306	lr=0.083067
Epoch[85] Batch [499]	Speed: 5013.156626 samples/sec	accuracy=0.790408	lr=0.082890
Epoch[85] Batch [549]	Speed: 5050.421340 samples/sec	accuracy=0.790417	lr=0.082713
Epoch[85] Batch [599]	Speed: 5002.146917 samples/sec	accuracy=0.790539	lr=0.082537
Epoch[85] Batch [649]	Speed: 5151.269279 samples/sec	accuracy=0.790585	lr=0.082360
Epoch[85] Batch [699]	Speed: 5077.446000 samples/sec	accuracy=0.790736	lr=0.082183
Epoch[85] Batch [749]	Speed: 5025.493188 samples/sec	accuracy=0.790826	lr=0.082007
Epoch[85] Batch [799]	Speed: 5039.308544 samples/sec	accuracy=0.790981	lr=0.081831
Epoch[85] Batch [849]	Speed: 5076.149675 samples/sec	accuracy=0.791118	lr=0.081655
Epoch[85] Batch [899]	Speed: 5188.422841 samples/sec	accuracy=0.791193	lr=0.081479
Epoch[85] Batch [949]	Speed: 5102.325022 samples/sec	accuracy=0.791251	lr=0.081303
Epoch[85] Batch [999]	Speed: 5028.282573 samples/sec	accuracy=0.791365	lr=0.081127
Epoch[85] Batch [1049]	Speed: 5004.878828 samples/sec	accuracy=0.791546	lr=0.080952
Epoch[85] Batch [1099]	Speed: 4982.753042 samples/sec	accuracy=0.791630	lr=0.080776
Epoch[85] Batch [1149]	Speed: 4990.052578 samples/sec	accuracy=0.791730	lr=0.080601
Epoch[85] Batch [1199]	Speed: 5090.542049 samples/sec	accuracy=0.791826	lr=0.080426
Epoch[85] Batch [1249]	Speed: 5112.816489 samples/sec	accuracy=0.791969	lr=0.080251
[Epoch 85] training: accuracy=0.791980
[Epoch 85] speed: 5043 samples/sec	time cost: 263.165939
[Epoch 85] validation: err-top1=0.284379 err-top5=0.094547
Epoch[86] Batch [49]	Speed: 4917.867209 samples/sec	accuracy=0.794064	lr=0.080073
Epoch[86] Batch [99]	Speed: 4938.963259 samples/sec	accuracy=0.792533	lr=0.079898
Epoch[86] Batch [149]	Speed: 4947.007182 samples/sec	accuracy=0.791832	lr=0.079723
Epoch[86] Batch [199]	Speed: 5074.730850 samples/sec	accuracy=0.791861	lr=0.079549
Epoch[86] Batch [249]	Speed: 4874.301084 samples/sec	accuracy=0.791824	lr=0.079375
Epoch[86] Batch [299]	Speed: 4946.570295 samples/sec	accuracy=0.791979	lr=0.079201
Epoch[86] Batch [349]	Speed: 4906.670751 samples/sec	accuracy=0.791955	lr=0.079027
Epoch[86] Batch [399]	Speed: 5012.967632 samples/sec	accuracy=0.791939	lr=0.078853
Epoch[86] Batch [449]	Speed: 4799.255606 samples/sec	accuracy=0.791925	lr=0.078679
Epoch[86] Batch [499]	Speed: 5056.518052 samples/sec	accuracy=0.791917	lr=0.078506
Epoch[86] Batch [549]	Speed: 5038.591560 samples/sec	accuracy=0.792102	lr=0.078332
Epoch[86] Batch [599]	Speed: 4990.360915 samples/sec	accuracy=0.792205	lr=0.078159
Epoch[86] Batch [649]	Speed: 5053.583330 samples/sec	accuracy=0.792383	lr=0.077986
Epoch[86] Batch [699]	Speed: 4817.685930 samples/sec	accuracy=0.792472	lr=0.077813
Epoch[86] Batch [749]	Speed: 4760.848717 samples/sec	accuracy=0.792602	lr=0.077640
Epoch[86] Batch [799]	Speed: 5023.987584 samples/sec	accuracy=0.792670	lr=0.077467
Epoch[86] Batch [849]	Speed: 5002.603932 samples/sec	accuracy=0.792688	lr=0.077295
Epoch[86] Batch [899]	Speed: 5165.859507 samples/sec	accuracy=0.792789	lr=0.077123
Epoch[86] Batch [949]	Speed: 5107.911057 samples/sec	accuracy=0.792850	lr=0.076950
Epoch[86] Batch [999]	Speed: 5204.838930 samples/sec	accuracy=0.792953	lr=0.076778
Epoch[86] Batch [1049]	Speed: 4942.046323 samples/sec	accuracy=0.793058	lr=0.076606
Epoch[86] Batch [1099]	Speed: 4910.023333 samples/sec	accuracy=0.793096	lr=0.076435
Epoch[86] Batch [1149]	Speed: 5061.173969 samples/sec	accuracy=0.793138	lr=0.076263
Epoch[86] Batch [1199]	Speed: 4963.657655 samples/sec	accuracy=0.793194	lr=0.076091
Epoch[86] Batch [1249]	Speed: 5048.986470 samples/sec	accuracy=0.793262	lr=0.075920
[Epoch 86] training: accuracy=0.793264
[Epoch 86] speed: 4976 samples/sec	time cost: 266.755492
[Epoch 86] validation: err-top1=0.275391 err-top5=0.086157
Epoch[87] Batch [49]	Speed: 4879.687127 samples/sec	accuracy=0.793063	lr=0.075745
Epoch[87] Batch [99]	Speed: 4939.344498 samples/sec	accuracy=0.793541	lr=0.075574
Epoch[87] Batch [149]	Speed: 4970.806729 samples/sec	accuracy=0.793502	lr=0.075404
Epoch[87] Batch [199]	Speed: 4947.046840 samples/sec	accuracy=0.793480	lr=0.075233
Epoch[87] Batch [249]	Speed: 5095.823443 samples/sec	accuracy=0.793240	lr=0.075062
Epoch[87] Batch [299]	Speed: 4849.006868 samples/sec	accuracy=0.793107	lr=0.074892
Epoch[87] Batch [349]	Speed: 4999.475250 samples/sec	accuracy=0.793227	lr=0.074721
Epoch[87] Batch [399]	Speed: 4916.056912 samples/sec	accuracy=0.793413	lr=0.074551
Epoch[87] Batch [449]	Speed: 5023.750997 samples/sec	accuracy=0.793461	lr=0.074381
Epoch[87] Batch [499]	Speed: 4990.849299 samples/sec	accuracy=0.793398	lr=0.074211
Epoch[87] Batch [549]	Speed: 5220.569585 samples/sec	accuracy=0.793360	lr=0.074042
Epoch[87] Batch [599]	Speed: 5061.993088 samples/sec	accuracy=0.793360	lr=0.073872
Epoch[87] Batch [649]	Speed: 5020.836450 samples/sec	accuracy=0.793456	lr=0.073703
Epoch[87] Batch [699]	Speed: 4879.693558 samples/sec	accuracy=0.793415	lr=0.073534
Epoch[87] Batch [749]	Speed: 5200.183145 samples/sec	accuracy=0.793457	lr=0.073364
Epoch[87] Batch [799]	Speed: 5144.219497 samples/sec	accuracy=0.793520	lr=0.073196
Epoch[87] Batch [849]	Speed: 5063.223576 samples/sec	accuracy=0.793564	lr=0.073027
Epoch[87] Batch [899]	Speed: 5353.374379 samples/sec	accuracy=0.793662	lr=0.072858
Epoch[87] Batch [949]	Speed: 5328.666324 samples/sec	accuracy=0.793729	lr=0.072690
Epoch[87] Batch [999]	Speed: 5141.567275 samples/sec	accuracy=0.793848	lr=0.072521
Epoch[87] Batch [1049]	Speed: 5065.492775 samples/sec	accuracy=0.793889	lr=0.072353
Epoch[87] Batch [1099]	Speed: 5287.798271 samples/sec	accuracy=0.794016	lr=0.072185
Epoch[87] Batch [1149]	Speed: 4996.208897 samples/sec	accuracy=0.794096	lr=0.072017
Epoch[87] Batch [1199]	Speed: 5053.316479 samples/sec	accuracy=0.794161	lr=0.071849
Epoch[87] Batch [1249]	Speed: 5027.143147 samples/sec	accuracy=0.794249	lr=0.071682
[Epoch 87] training: accuracy=0.794250
[Epoch 87] speed: 5048 samples/sec	time cost: 262.999702
[Epoch 87] validation: err-top1=0.271220 err-top5=0.084635
Epoch[88] Batch [49]	Speed: 5136.040944 samples/sec	accuracy=0.793947	lr=0.071511
Epoch[88] Batch [99]	Speed: 5161.450061 samples/sec	accuracy=0.793728	lr=0.071344
Epoch[88] Batch [149]	Speed: 5061.878066 samples/sec	accuracy=0.793449	lr=0.071177
Epoch[88] Batch [199]	Speed: 4901.774422 samples/sec	accuracy=0.793414	lr=0.071010
Epoch[88] Batch [249]	Speed: 4866.665710 samples/sec	accuracy=0.793416	lr=0.070843
Epoch[88] Batch [299]	Speed: 4848.413722 samples/sec	accuracy=0.793608	lr=0.070676
Epoch[88] Batch [349]	Speed: 5051.877467 samples/sec	accuracy=0.793598	lr=0.070510
Epoch[88] Batch [399]	Speed: 5111.839444 samples/sec	accuracy=0.793693	lr=0.070343
Epoch[88] Batch [449]	Speed: 5193.989419 samples/sec	accuracy=0.793720	lr=0.070177
Epoch[88] Batch [499]	Speed: 5123.048200 samples/sec	accuracy=0.793717	lr=0.070011
Epoch[88] Batch [549]	Speed: 5038.252412 samples/sec	accuracy=0.793700	lr=0.069845
Epoch[88] Batch [599]	Speed: 5034.381744 samples/sec	accuracy=0.793914	lr=0.069680
Epoch[88] Batch [649]	Speed: 5126.508140 samples/sec	accuracy=0.794037	lr=0.069514
Epoch[88] Batch [699]	Speed: 5082.274180 samples/sec	accuracy=0.794173	lr=0.069349
Epoch[88] Batch [749]	Speed: 4933.880213 samples/sec	accuracy=0.794216	lr=0.069183
Epoch[88] Batch [799]	Speed: 5057.714194 samples/sec	accuracy=0.794348	lr=0.069018
Epoch[88] Batch [849]	Speed: 5160.191707 samples/sec	accuracy=0.794384	lr=0.068853
Epoch[88] Batch [899]	Speed: 5120.202156 samples/sec	accuracy=0.794448	lr=0.068689
Epoch[88] Batch [949]	Speed: 5022.967235 samples/sec	accuracy=0.794570	lr=0.068524
Epoch[88] Batch [999]	Speed: 5058.036549 samples/sec	accuracy=0.794730	lr=0.068359
Epoch[88] Batch [1049]	Speed: 5117.412075 samples/sec	accuracy=0.794795	lr=0.068195
Epoch[88] Batch [1099]	Speed: 4960.713502 samples/sec	accuracy=0.794877	lr=0.068031
Epoch[88] Batch [1149]	Speed: 5171.873311 samples/sec	accuracy=0.794995	lr=0.067867
Epoch[88] Batch [1199]	Speed: 5227.568686 samples/sec	accuracy=0.795103	lr=0.067703
Epoch[88] Batch [1249]	Speed: 5201.506562 samples/sec	accuracy=0.795186	lr=0.067539
[Epoch 88] training: accuracy=0.795189
[Epoch 88] speed: 5064 samples/sec	time cost: 262.717983
[Epoch 88] validation: err-top1=0.268156 err-top5=0.084263
Epoch[89] Batch [49]	Speed: 4780.339231 samples/sec	accuracy=0.796292	lr=0.067373
Epoch[89] Batch [99]	Speed: 4965.103663 samples/sec	accuracy=0.795628	lr=0.067209
Epoch[89] Batch [149]	Speed: 5024.270977 samples/sec	accuracy=0.795726	lr=0.067046
Epoch[89] Batch [199]	Speed: 4930.940924 samples/sec	accuracy=0.795580	lr=0.066883
Epoch[89] Batch [249]	Speed: 4893.849261 samples/sec	accuracy=0.795596	lr=0.066720
Epoch[89] Batch [299]	Speed: 5056.100536 samples/sec	accuracy=0.795523	lr=0.066557
Epoch[89] Batch [349]	Speed: 5109.928420 samples/sec	accuracy=0.795662	lr=0.066395
Epoch[89] Batch [399]	Speed: 5145.447386 samples/sec	accuracy=0.795608	lr=0.066232
Epoch[89] Batch [449]	Speed: 4963.348708 samples/sec	accuracy=0.795709	lr=0.066070
Epoch[89] Batch [499]	Speed: 5030.612118 samples/sec	accuracy=0.795739	lr=0.065908
Epoch[89] Batch [549]	Speed: 5165.223587 samples/sec	accuracy=0.795753	lr=0.065746
Epoch[89] Batch [599]	Speed: 5130.795174 samples/sec	accuracy=0.795818	lr=0.065584
Epoch[89] Batch [649]	Speed: 5016.234733 samples/sec	accuracy=0.795903	lr=0.065423
Epoch[89] Batch [699]	Speed: 5184.622113 samples/sec	accuracy=0.795989	lr=0.065261
Epoch[89] Batch [749]	Speed: 5184.428480 samples/sec	accuracy=0.796072	lr=0.065100
Epoch[89] Batch [799]	Speed: 5046.080045 samples/sec	accuracy=0.796196	lr=0.064939
Epoch[89] Batch [849]	Speed: 4931.704950 samples/sec	accuracy=0.796217	lr=0.064778
Epoch[89] Batch [899]	Speed: 5039.432594 samples/sec	accuracy=0.796270	lr=0.064617
Epoch[89] Batch [949]	Speed: 4980.441150 samples/sec	accuracy=0.796276	lr=0.064456
Epoch[89] Batch [999]	Speed: 5094.991769 samples/sec	accuracy=0.796418	lr=0.064296
Epoch[89] Batch [1049]	Speed: 5113.504223 samples/sec	accuracy=0.796511	lr=0.064135
Epoch[89] Batch [1099]	Speed: 4936.378159 samples/sec	accuracy=0.796562	lr=0.063975
Epoch[89] Batch [1149]	Speed: 5082.115418 samples/sec	accuracy=0.796642	lr=0.063815
Epoch[89] Batch [1199]	Speed: 5187.280108 samples/sec	accuracy=0.796707	lr=0.063655
Epoch[89] Batch [1249]	Speed: 5180.189588 samples/sec	accuracy=0.796813	lr=0.063496
[Epoch 89] training: accuracy=0.796816
[Epoch 89] speed: 5040 samples/sec	time cost: 263.520363
[Epoch 89] validation: err-top1=0.267817 err-top5=0.082948
Epoch[90] Batch [49]	Speed: 4669.544015 samples/sec	accuracy=0.796270	lr=0.063333
Epoch[90] Batch [99]	Speed: 4803.894148 samples/sec	accuracy=0.796093	lr=0.063174
Epoch[90] Batch [149]	Speed: 5005.177218 samples/sec	accuracy=0.795897	lr=0.063015
Epoch[90] Batch [199]	Speed: 5007.028425 samples/sec	accuracy=0.796224	lr=0.062855
Epoch[90] Batch [249]	Speed: 5031.911461 samples/sec	accuracy=0.796376	lr=0.062697
Epoch[90] Batch [299]	Speed: 4943.860564 samples/sec	accuracy=0.796458	lr=0.062538
Epoch[90] Batch [349]	Speed: 5182.229319 samples/sec	accuracy=0.796575	lr=0.062379
Epoch[90] Batch [399]	Speed: 5151.054160 samples/sec	accuracy=0.796761	lr=0.062221
Epoch[90] Batch [449]	Speed: 5020.908879 samples/sec	accuracy=0.796873	lr=0.062063
Epoch[90] Batch [499]	Speed: 4963.264050 samples/sec	accuracy=0.796929	lr=0.061905
Epoch[90] Batch [549]	Speed: 4929.311071 samples/sec	accuracy=0.796897	lr=0.061747
Epoch[90] Batch [599]	Speed: 4996.102192 samples/sec	accuracy=0.796839	lr=0.061589
Epoch[90] Batch [649]	Speed: 4971.321215 samples/sec	accuracy=0.796879	lr=0.061432
Epoch[90] Batch [699]	Speed: 4993.040957 samples/sec	accuracy=0.796925	lr=0.061274
Epoch[90] Batch [749]	Speed: 4975.020914 samples/sec	accuracy=0.797025	lr=0.061117
Epoch[90] Batch [799]	Speed: 5017.165256 samples/sec	accuracy=0.797061	lr=0.060960
Epoch[90] Batch [849]	Speed: 5053.495565 samples/sec	accuracy=0.797131	lr=0.060803
Epoch[90] Batch [899]	Speed: 5027.063241 samples/sec	accuracy=0.797155	lr=0.060646
Epoch[90] Batch [949]	Speed: 4976.854722 samples/sec	accuracy=0.797239	lr=0.060490
Epoch[90] Batch [999]	Speed: 5027.321795 samples/sec	accuracy=0.797320	lr=0.060333
Epoch[90] Batch [1049]	Speed: 4828.441303 samples/sec	accuracy=0.797354	lr=0.060177
Epoch[90] Batch [1099]	Speed: 4904.084395 samples/sec	accuracy=0.797439	lr=0.060021
Epoch[90] Batch [1149]	Speed: 4992.555045 samples/sec	accuracy=0.797514	lr=0.059865
Epoch[90] Batch [1199]	Speed: 5144.931853 samples/sec	accuracy=0.797620	lr=0.059710
Epoch[90] Batch [1249]	Speed: 5097.448887 samples/sec	accuracy=0.797707	lr=0.059554
[Epoch 90] training: accuracy=0.797709
[Epoch 90] speed: 4982 samples/sec	time cost: 266.421733
[Epoch 90] validation: err-top1=0.270029 err-top5=0.085539
Epoch[91] Batch [49]	Speed: 4790.585031 samples/sec	accuracy=0.798813	lr=0.059396
Epoch[91] Batch [99]	Speed: 5071.107500 samples/sec	accuracy=0.798632	lr=0.059240
Epoch[91] Batch [149]	Speed: 5054.008280 samples/sec	accuracy=0.798428	lr=0.059085
Epoch[91] Batch [199]	Speed: 4910.729794 samples/sec	accuracy=0.798480	lr=0.058930
Epoch[91] Batch [249]	Speed: 4828.810230 samples/sec	accuracy=0.798172	lr=0.058776
Epoch[91] Batch [299]	Speed: 4958.913668 samples/sec	accuracy=0.798020	lr=0.058621
Epoch[91] Batch [349]	Speed: 5014.519563 samples/sec	accuracy=0.797952	lr=0.058467
Epoch[91] Batch [399]	Speed: 5194.217059 samples/sec	accuracy=0.797854	lr=0.058313
Epoch[91] Batch [449]	Speed: 4834.480965 samples/sec	accuracy=0.797931	lr=0.058158
Epoch[91] Batch [499]	Speed: 5253.796250 samples/sec	accuracy=0.798008	lr=0.058005
Epoch[91] Batch [549]	Speed: 5128.575496 samples/sec	accuracy=0.798175	lr=0.057851
Epoch[91] Batch [599]	Speed: 5193.477301 samples/sec	accuracy=0.798212	lr=0.057697
Epoch[91] Batch [649]	Speed: 4883.462634 samples/sec	accuracy=0.798263	lr=0.057544
Epoch[91] Batch [699]	Speed: 5012.812000 samples/sec	accuracy=0.798298	lr=0.057391
Epoch[91] Batch [749]	Speed: 4816.914035 samples/sec	accuracy=0.798304	lr=0.057238
Epoch[91] Batch [799]	Speed: 5001.130644 samples/sec	accuracy=0.798334	lr=0.057085
Epoch[91] Batch [849]	Speed: 4941.697189 samples/sec	accuracy=0.798398	lr=0.056932
Epoch[91] Batch [899]	Speed: 5057.670240 samples/sec	accuracy=0.798396	lr=0.056780
Epoch[91] Batch [949]	Speed: 5057.442023 samples/sec	accuracy=0.798457	lr=0.056627
Epoch[91] Batch [999]	Speed: 5045.919268 samples/sec	accuracy=0.798511	lr=0.056475
Epoch[91] Batch [1049]	Speed: 5148.509308 samples/sec	accuracy=0.798544	lr=0.056323
Epoch[91] Batch [1099]	Speed: 5000.052732 samples/sec	accuracy=0.798619	lr=0.056171
Epoch[91] Batch [1149]	Speed: 4915.643928 samples/sec	accuracy=0.798692	lr=0.056020
Epoch[91] Batch [1199]	Speed: 5015.478850 samples/sec	accuracy=0.798748	lr=0.055868
Epoch[91] Batch [1249]	Speed: 5176.505222 samples/sec	accuracy=0.798785	lr=0.055717
[Epoch 91] training: accuracy=0.798786
[Epoch 91] speed: 5004 samples/sec	time cost: 264.938834
[Epoch 91] validation: err-top1=0.265904 err-top5=0.081234
Epoch[92] Batch [49]	Speed: 4917.963615 samples/sec	accuracy=0.798903	lr=0.055563
Epoch[92] Batch [99]	Speed: 5176.990910 samples/sec	accuracy=0.798634	lr=0.055412
Epoch[92] Batch [149]	Speed: 4830.127558 samples/sec	accuracy=0.798587	lr=0.055261
Epoch[92] Batch [199]	Speed: 4883.203675 samples/sec	accuracy=0.798435	lr=0.055111
Epoch[92] Batch [249]	Speed: 4974.824643 samples/sec	accuracy=0.798339	lr=0.054960
Epoch[92] Batch [299]	Speed: 5043.401289 samples/sec	accuracy=0.798396	lr=0.054810
Epoch[92] Batch [349]	Speed: 5025.378643 samples/sec	accuracy=0.798582	lr=0.054660
Epoch[92] Batch [399]	Speed: 4920.977884 samples/sec	accuracy=0.798780	lr=0.054510
Epoch[92] Batch [449]	Speed: 5018.537176 samples/sec	accuracy=0.798953	lr=0.054360
Epoch[92] Batch [499]	Speed: 5151.238635 samples/sec	accuracy=0.799062	lr=0.054210
Epoch[92] Batch [549]	Speed: 5063.363730 samples/sec	accuracy=0.799115	lr=0.054061
Epoch[92] Batch [599]	Speed: 5066.938000 samples/sec	accuracy=0.799140	lr=0.053912
Epoch[92] Batch [649]	Speed: 5059.932436 samples/sec	accuracy=0.799172	lr=0.053763
Epoch[92] Batch [699]	Speed: 5147.099466 samples/sec	accuracy=0.799189	lr=0.053614
Epoch[92] Batch [749]	Speed: 5135.920813 samples/sec	accuracy=0.799265	lr=0.053465
Epoch[92] Batch [799]	Speed: 5048.361789 samples/sec	accuracy=0.799309	lr=0.053317
Epoch[92] Batch [849]	Speed: 5072.494585 samples/sec	accuracy=0.799327	lr=0.053168
Epoch[92] Batch [899]	Speed: 5147.089720 samples/sec	accuracy=0.799393	lr=0.053020
Epoch[92] Batch [949]	Speed: 5018.807755 samples/sec	accuracy=0.799484	lr=0.052872
Epoch[92] Batch [999]	Speed: 5164.664585 samples/sec	accuracy=0.799553	lr=0.052724
Epoch[92] Batch [1049]	Speed: 5077.519471 samples/sec	accuracy=0.799637	lr=0.052577
Epoch[92] Batch [1099]	Speed: 5001.460386 samples/sec	accuracy=0.799738	lr=0.052429
Epoch[92] Batch [1149]	Speed: 4989.521920 samples/sec	accuracy=0.799782	lr=0.052282
Epoch[92] Batch [1199]	Speed: 5018.041481 samples/sec	accuracy=0.799820	lr=0.052135
Epoch[92] Batch [1249]	Speed: 5115.975700 samples/sec	accuracy=0.799873	lr=0.051988
[Epoch 92] training: accuracy=0.799873
[Epoch 92] speed: 5037 samples/sec	time cost: 263.476577
[Epoch 92] validation: err-top1=0.258331 err-top5=0.078663
Epoch[93] Batch [49]	Speed: 4815.109698 samples/sec	accuracy=0.801042	lr=0.051838
Epoch[93] Batch [99]	Speed: 5012.519017 samples/sec	accuracy=0.800834	lr=0.051691
Epoch[93] Batch [149]	Speed: 4904.227300 samples/sec	accuracy=0.800440	lr=0.051545
Epoch[93] Batch [199]	Speed: 5070.547969 samples/sec	accuracy=0.800211	lr=0.051399
Epoch[93] Batch [249]	Speed: 5037.034502 samples/sec	accuracy=0.800031	lr=0.051253
Epoch[93] Batch [299]	Speed: 5017.276731 samples/sec	accuracy=0.799784	lr=0.051107
Epoch[93] Batch [349]	Speed: 5122.405915 samples/sec	accuracy=0.799857	lr=0.050961
Epoch[93] Batch [399]	Speed: 5128.456326 samples/sec	accuracy=0.799902	lr=0.050816
Epoch[93] Batch [449]	Speed: 5139.946055 samples/sec	accuracy=0.799931	lr=0.050670
Epoch[93] Batch [499]	Speed: 4885.394146 samples/sec	accuracy=0.799956	lr=0.050525
Epoch[93] Batch [549]	Speed: 5082.096295 samples/sec	accuracy=0.799928	lr=0.050380
Epoch[93] Batch [599]	Speed: 5093.380455 samples/sec	accuracy=0.799987	lr=0.050235
Epoch[93] Batch [649]	Speed: 5144.911638 samples/sec	accuracy=0.800022	lr=0.050091
Epoch[93] Batch [699]	Speed: 4905.447041 samples/sec	accuracy=0.800075	lr=0.049946
Epoch[93] Batch [749]	Speed: 5038.177472 samples/sec	accuracy=0.800126	lr=0.049802
Epoch[93] Batch [799]	Speed: 4982.602749 samples/sec	accuracy=0.800087	lr=0.049658
Epoch[93] Batch [849]	Speed: 4931.226713 samples/sec	accuracy=0.800130	lr=0.049514
Epoch[93] Batch [899]	Speed: 5253.423786 samples/sec	accuracy=0.800149	lr=0.049370
Epoch[93] Batch [949]	Speed: 4839.997039 samples/sec	accuracy=0.800156	lr=0.049227
Epoch[93] Batch [999]	Speed: 5070.250234 samples/sec	accuracy=0.800213	lr=0.049083
Epoch[93] Batch [1049]	Speed: 4940.501753 samples/sec	accuracy=0.800279	lr=0.048940
Epoch[93] Batch [1099]	Speed: 5037.933637 samples/sec	accuracy=0.800295	lr=0.048797
Epoch[93] Batch [1149]	Speed: 4752.080991 samples/sec	accuracy=0.800360	lr=0.048654
Epoch[93] Batch [1199]	Speed: 5205.201382 samples/sec	accuracy=0.800409	lr=0.048511
Epoch[93] Batch [1249]	Speed: 5072.953880 samples/sec	accuracy=0.800466	lr=0.048369
[Epoch 93] training: accuracy=0.800465
[Epoch 93] speed: 5013 samples/sec	time cost: 265.103249
[Epoch 93] validation: err-top1=0.274984 err-top5=0.085836
Epoch[94] Batch [49]	Speed: 5023.194465 samples/sec	accuracy=0.801883	lr=0.048224
Epoch[94] Batch [99]	Speed: 4995.732130 samples/sec	accuracy=0.801478	lr=0.048082
Epoch[94] Batch [149]	Speed: 4760.718055 samples/sec	accuracy=0.801255	lr=0.047940
Epoch[94] Batch [199]	Speed: 4767.550660 samples/sec	accuracy=0.801038	lr=0.047798
Epoch[94] Batch [249]	Speed: 5118.598040 samples/sec	accuracy=0.801043	lr=0.047656
Epoch[94] Batch [299]	Speed: 4893.416249 samples/sec	accuracy=0.800898	lr=0.047515
Epoch[94] Batch [349]	Speed: 4990.817286 samples/sec	accuracy=0.801009	lr=0.047374
Epoch[94] Batch [399]	Speed: 5043.302508 samples/sec	accuracy=0.801160	lr=0.047233
Epoch[94] Batch [449]	Speed: 5022.715002 samples/sec	accuracy=0.801303	lr=0.047092
Epoch[94] Batch [499]	Speed: 4927.371252 samples/sec	accuracy=0.801379	lr=0.046951
Epoch[94] Batch [549]	Speed: 4962.683451 samples/sec	accuracy=0.801470	lr=0.046811
Epoch[94] Batch [599]	Speed: 4892.720445 samples/sec	accuracy=0.801375	lr=0.046670
Epoch[94] Batch [649]	Speed: 4879.690342 samples/sec	accuracy=0.801379	lr=0.046530
Epoch[94] Batch [699]	Speed: 5088.873976 samples/sec	accuracy=0.801464	lr=0.046390
Epoch[94] Batch [749]	Speed: 4975.471026 samples/sec	accuracy=0.801513	lr=0.046251
Epoch[94] Batch [799]	Speed: 4805.115125 samples/sec	accuracy=0.801587	lr=0.046111
Epoch[94] Batch [849]	Speed: 5218.060453 samples/sec	accuracy=0.801626	lr=0.045972
Epoch[94] Batch [899]	Speed: 5137.723125 samples/sec	accuracy=0.801571	lr=0.045833
Epoch[94] Batch [949]	Speed: 5045.710012 samples/sec	accuracy=0.801631	lr=0.045693
Epoch[94] Batch [999]	Speed: 5001.221025 samples/sec	accuracy=0.801679	lr=0.045555
Epoch[94] Batch [1049]	Speed: 4822.767932 samples/sec	accuracy=0.801752	lr=0.045416
Epoch[94] Batch [1099]	Speed: 5085.073147 samples/sec	accuracy=0.801829	lr=0.045278
Epoch[94] Batch [1149]	Speed: 5193.436733 samples/sec	accuracy=0.801885	lr=0.045139
Epoch[94] Batch [1199]	Speed: 5112.388651 samples/sec	accuracy=0.801950	lr=0.045001
Epoch[94] Batch [1249]	Speed: 5132.088165 samples/sec	accuracy=0.801956	lr=0.044863
[Epoch 94] training: accuracy=0.801955
[Epoch 94] speed: 4987 samples/sec	time cost: 266.602464
[Epoch 94] validation: err-top1=0.255301 err-top5=0.075753
Epoch[95] Batch [49]	Speed: 5119.911378 samples/sec	accuracy=0.801993	lr=0.044723
Epoch[95] Batch [99]	Speed: 4923.667775 samples/sec	accuracy=0.801430	lr=0.044585
Epoch[95] Batch [149]	Speed: 4992.372011 samples/sec	accuracy=0.800956	lr=0.044448
Epoch[95] Batch [199]	Speed: 4976.513684 samples/sec	accuracy=0.801136	lr=0.044311
Epoch[95] Batch [249]	Speed: 4952.069364 samples/sec	accuracy=0.801201	lr=0.044174
Epoch[95] Batch [299]	Speed: 4921.687953 samples/sec	accuracy=0.801129	lr=0.044037
Epoch[95] Batch [349]	Speed: 4995.822316 samples/sec	accuracy=0.801214	lr=0.043900
Epoch[95] Batch [399]	Speed: 4882.476804 samples/sec	accuracy=0.801265	lr=0.043764
Epoch[95] Batch [449]	Speed: 5074.025331 samples/sec	accuracy=0.801375	lr=0.043628
Epoch[95] Batch [499]	Speed: 5050.022406 samples/sec	accuracy=0.801619	lr=0.043492
Epoch[95] Batch [549]	Speed: 4980.965372 samples/sec	accuracy=0.801702	lr=0.043356
Epoch[95] Batch [599]	Speed: 4927.632995 samples/sec	accuracy=0.801779	lr=0.043220
Epoch[95] Batch [649]	Speed: 5008.873641 samples/sec	accuracy=0.801769	lr=0.043085
Epoch[95] Batch [699]	Speed: 4864.340050 samples/sec	accuracy=0.801842	lr=0.042949
Epoch[95] Batch [749]	Speed: 4945.572149 samples/sec	accuracy=0.801886	lr=0.042814
Epoch[95] Batch [799]	Speed: 5013.367638 samples/sec	accuracy=0.801975	lr=0.042679
Epoch[95] Batch [849]	Speed: 4993.110613 samples/sec	accuracy=0.802108	lr=0.042545
Epoch[95] Batch [899]	Speed: 5065.229682 samples/sec	accuracy=0.802185	lr=0.042410
Epoch[95] Batch [949]	Speed: 5031.444478 samples/sec	accuracy=0.802277	lr=0.042276
Epoch[95] Batch [999]	Speed: 4905.907626 samples/sec	accuracy=0.802299	lr=0.042141
Epoch[95] Batch [1049]	Speed: 5072.205127 samples/sec	accuracy=0.802383	lr=0.042007
Epoch[95] Batch [1099]	Speed: 5175.808173 samples/sec	accuracy=0.802462	lr=0.041874
Epoch[95] Batch [1149]	Speed: 5132.749810 samples/sec	accuracy=0.802550	lr=0.041740
Epoch[95] Batch [1199]	Speed: 4986.647521 samples/sec	accuracy=0.802633	lr=0.041607
Epoch[95] Batch [1249]	Speed: 5154.372404 samples/sec	accuracy=0.802663	lr=0.041473
[Epoch 95] training: accuracy=0.802665
[Epoch 95] speed: 5000 samples/sec	time cost: 265.448392
[Epoch 95] validation: err-top1=0.256358 err-top5=0.075833
Epoch[96] Batch [49]	Speed: 4872.846771 samples/sec	accuracy=0.803017	lr=0.041338
Epoch[96] Batch [99]	Speed: 4777.857095 samples/sec	accuracy=0.803083	lr=0.041205
Epoch[96] Batch [149]	Speed: 5096.024300 samples/sec	accuracy=0.803037	lr=0.041072
Epoch[96] Batch [199]	Speed: 4999.958435 samples/sec	accuracy=0.802893	lr=0.040940
Epoch[96] Batch [249]	Speed: 4997.199215 samples/sec	accuracy=0.802821	lr=0.040807
Epoch[96] Batch [299]	Speed: 5025.396871 samples/sec	accuracy=0.802932	lr=0.040675
Epoch[96] Batch [349]	Speed: 5145.406948 samples/sec	accuracy=0.802942	lr=0.040543
Epoch[96] Batch [399]	Speed: 5048.659451 samples/sec	accuracy=0.802940	lr=0.040412
Epoch[96] Batch [449]	Speed: 5047.336973 samples/sec	accuracy=0.802904	lr=0.040280
Epoch[96] Batch [499]	Speed: 4990.569200 samples/sec	accuracy=0.802991	lr=0.040149
Epoch[96] Batch [549]	Speed: 5062.929685 samples/sec	accuracy=0.803019	lr=0.040018
Epoch[96] Batch [599]	Speed: 5151.620230 samples/sec	accuracy=0.803072	lr=0.039887
Epoch[96] Batch [649]	Speed: 5057.891687 samples/sec	accuracy=0.803229	lr=0.039756
Epoch[96] Batch [699]	Speed: 4926.585625 samples/sec	accuracy=0.803310	lr=0.039625
Epoch[96] Batch [749]	Speed: 5022.243970 samples/sec	accuracy=0.803359	lr=0.039495
Epoch[96] Batch [799]	Speed: 5030.773218 samples/sec	accuracy=0.803390	lr=0.039365
Epoch[96] Batch [849]	Speed: 4971.416275 samples/sec	accuracy=0.803441	lr=0.039235
Epoch[96] Batch [899]	Speed: 5054.732752 samples/sec	accuracy=0.803474	lr=0.039105
Epoch[96] Batch [949]	Speed: 5084.761062 samples/sec	accuracy=0.803504	lr=0.038975
Epoch[96] Batch [999]	Speed: 5142.094077 samples/sec	accuracy=0.803595	lr=0.038846
Epoch[96] Batch [1049]	Speed: 4925.675742 samples/sec	accuracy=0.803671	lr=0.038717
Epoch[96] Batch [1099]	Speed: 4888.726163 samples/sec	accuracy=0.803761	lr=0.038588
Epoch[96] Batch [1149]	Speed: 5006.450731 samples/sec	accuracy=0.803856	lr=0.038459
Epoch[96] Batch [1199]	Speed: 4877.411707 samples/sec	accuracy=0.803911	lr=0.038330
Epoch[96] Batch [1249]	Speed: 5115.892946 samples/sec	accuracy=0.803990	lr=0.038202
[Epoch 96] training: accuracy=0.803990
[Epoch 96] speed: 5007 samples/sec	time cost: 265.201572
[Epoch 96] validation: err-top1=0.252212 err-top5=0.076391
Epoch[97] Batch [49]	Speed: 4757.728417 samples/sec	accuracy=0.803869	lr=0.038071
Epoch[97] Batch [99]	Speed: 4731.356903 samples/sec	accuracy=0.804217	lr=0.037943
Epoch[97] Batch [149]	Speed: 5105.476981 samples/sec	accuracy=0.804453	lr=0.037815
Epoch[97] Batch [199]	Speed: 4985.389038 samples/sec	accuracy=0.804275	lr=0.037687
Epoch[97] Batch [249]	Speed: 5055.938644 samples/sec	accuracy=0.804130	lr=0.037560
Epoch[97] Batch [299]	Speed: 4803.657097 samples/sec	accuracy=0.804054	lr=0.037432
Epoch[97] Batch [349]	Speed: 4995.232681 samples/sec	accuracy=0.804039	lr=0.037305
Epoch[97] Batch [399]	Speed: 4895.452397 samples/sec	accuracy=0.804107	lr=0.037178
Epoch[97] Batch [449]	Speed: 4923.482195 samples/sec	accuracy=0.804100	lr=0.037052
Epoch[97] Batch [499]	Speed: 4851.558349 samples/sec	accuracy=0.804106	lr=0.036925
Epoch[97] Batch [549]	Speed: 5079.737941 samples/sec	accuracy=0.804116	lr=0.036799
Epoch[97] Batch [599]	Speed: 5198.078301 samples/sec	accuracy=0.804168	lr=0.036673
Epoch[97] Batch [649]	Speed: 4943.554646 samples/sec	accuracy=0.804306	lr=0.036547
Epoch[97] Batch [699]	Speed: 4909.027648 samples/sec	accuracy=0.804465	lr=0.036421
Epoch[97] Batch [749]	Speed: 5175.325201 samples/sec	accuracy=0.804573	lr=0.036296
Epoch[97] Batch [799]	Speed: 5108.569035 samples/sec	accuracy=0.804627	lr=0.036170
Epoch[97] Batch [849]	Speed: 5117.422929 samples/sec	accuracy=0.804593	lr=0.036045
Epoch[97] Batch [899]	Speed: 5024.862314 samples/sec	accuracy=0.804638	lr=0.035920
Epoch[97] Batch [949]	Speed: 5145.435427 samples/sec	accuracy=0.804763	lr=0.035795
Epoch[97] Batch [999]	Speed: 5162.834511 samples/sec	accuracy=0.804863	lr=0.035671
Epoch[97] Batch [1049]	Speed: 5010.359912 samples/sec	accuracy=0.804968	lr=0.035546
Epoch[97] Batch [1099]	Speed: 5089.133862 samples/sec	accuracy=0.805031	lr=0.035422
Epoch[97] Batch [1149]	Speed: 5017.214839 samples/sec	accuracy=0.805139	lr=0.035298
Epoch[97] Batch [1199]	Speed: 5125.551545 samples/sec	accuracy=0.805194	lr=0.035174
Epoch[97] Batch [1249]	Speed: 5267.325483 samples/sec	accuracy=0.805210	lr=0.035051
[Epoch 97] training: accuracy=0.805211
[Epoch 97] speed: 5011 samples/sec	time cost: 264.723829
[Epoch 97] validation: err-top1=0.249641 err-top5=0.072465
Epoch[98] Batch [49]	Speed: 5041.022375 samples/sec	accuracy=0.805635	lr=0.034925
Epoch[98] Batch [99]	Speed: 5055.516820 samples/sec	accuracy=0.805398	lr=0.034802
Epoch[98] Batch [149]	Speed: 4961.694154 samples/sec	accuracy=0.805523	lr=0.034679
Epoch[98] Batch [199]	Speed: 5028.043580 samples/sec	accuracy=0.805423	lr=0.034556
Epoch[98] Batch [249]	Speed: 4916.294831 samples/sec	accuracy=0.805089	lr=0.034433
Epoch[98] Batch [299]	Speed: 4906.320881 samples/sec	accuracy=0.805218	lr=0.034311
Epoch[98] Batch [349]	Speed: 5120.078859 samples/sec	accuracy=0.805169	lr=0.034189
Epoch[98] Batch [399]	Speed: 5092.266037 samples/sec	accuracy=0.805264	lr=0.034067
Epoch[98] Batch [449]	Speed: 4998.728830 samples/sec	accuracy=0.805267	lr=0.033945
Epoch[98] Batch [499]	Speed: 4972.298352 samples/sec	accuracy=0.805290	lr=0.033823
Epoch[98] Batch [549]	Speed: 5106.760156 samples/sec	accuracy=0.805378	lr=0.033702
Epoch[98] Batch [599]	Speed: 5058.396358 samples/sec	accuracy=0.805432	lr=0.033581
Epoch[98] Batch [649]	Speed: 5134.591024 samples/sec	accuracy=0.805449	lr=0.033460
Epoch[98] Batch [699]	Speed: 4966.164146 samples/sec	accuracy=0.805519	lr=0.033339
Epoch[98] Batch [749]	Speed: 5045.988391 samples/sec	accuracy=0.805470	lr=0.033218
Epoch[98] Batch [799]	Speed: 5054.002809 samples/sec	accuracy=0.805446	lr=0.033098
Epoch[98] Batch [849]	Speed: 5032.965998 samples/sec	accuracy=0.805432	lr=0.032978
Epoch[98] Batch [899]	Speed: 5025.991414 samples/sec	accuracy=0.805486	lr=0.032858
Epoch[98] Batch [949]	Speed: 5025.795570 samples/sec	accuracy=0.805601	lr=0.032738
Epoch[98] Batch [999]	Speed: 5090.314356 samples/sec	accuracy=0.805654	lr=0.032618
Epoch[98] Batch [1049]	Speed: 5020.952666 samples/sec	accuracy=0.805728	lr=0.032499
Epoch[98] Batch [1099]	Speed: 5031.269073 samples/sec	accuracy=0.805787	lr=0.032380
Epoch[98] Batch [1149]	Speed: 5056.417090 samples/sec	accuracy=0.805817	lr=0.032261
Epoch[98] Batch [1199]	Speed: 5065.763423 samples/sec	accuracy=0.805899	lr=0.032142
Epoch[98] Batch [1249]	Speed: 5229.075929 samples/sec	accuracy=0.805979	lr=0.032023
[Epoch 98] training: accuracy=0.805980
[Epoch 98] speed: 5036 samples/sec	time cost: 263.068170
[Epoch 98] validation: err-top1=0.251912 err-top5=0.075541
Epoch[99] Batch [49]	Speed: 4873.520786 samples/sec	accuracy=0.806419	lr=0.031902
Epoch[99] Batch [99]	Speed: 5061.224903 samples/sec	accuracy=0.806822	lr=0.031784
Epoch[99] Batch [149]	Speed: 5048.228873 samples/sec	accuracy=0.806837	lr=0.031666
Epoch[99] Batch [199]	Speed: 5100.312809 samples/sec	accuracy=0.806424	lr=0.031548
Epoch[99] Batch [249]	Speed: 5013.381331 samples/sec	accuracy=0.806497	lr=0.031431
Epoch[99] Batch [299]	Speed: 5009.133365 samples/sec	accuracy=0.806538	lr=0.031313
Epoch[99] Batch [349]	Speed: 5058.583669 samples/sec	accuracy=0.806393	lr=0.031196
Epoch[99] Batch [399]	Speed: 5109.471889 samples/sec	accuracy=0.806340	lr=0.031079
Epoch[99] Batch [449]	Speed: 4985.238007 samples/sec	accuracy=0.806402	lr=0.030962
Epoch[99] Batch [499]	Speed: 5004.212071 samples/sec	accuracy=0.806375	lr=0.030846
Epoch[99] Batch [549]	Speed: 4962.419348 samples/sec	accuracy=0.806324	lr=0.030729
Epoch[99] Batch [599]	Speed: 4997.780708 samples/sec	accuracy=0.806325	lr=0.030613
Epoch[99] Batch [649]	Speed: 5045.928753 samples/sec	accuracy=0.806317	lr=0.030497
Epoch[99] Batch [699]	Speed: 5015.426842 samples/sec	accuracy=0.806355	lr=0.030381
Epoch[99] Batch [749]	Speed: 5158.130505 samples/sec	accuracy=0.806449	lr=0.030266
Epoch[99] Batch [799]	Speed: 5103.259868 samples/sec	accuracy=0.806499	lr=0.030150
Epoch[99] Batch [849]	Speed: 5238.262630 samples/sec	accuracy=0.806587	lr=0.030035
Epoch[99] Batch [899]	Speed: 5086.565954 samples/sec	accuracy=0.806665	lr=0.029920
Epoch[99] Batch [949]	Speed: 4999.839813 samples/sec	accuracy=0.806730	lr=0.029805
Epoch[99] Batch [999]	Speed: 5054.018747 samples/sec	accuracy=0.806786	lr=0.029691
Epoch[99] Batch [1049]	Speed: 5087.823125 samples/sec	accuracy=0.806849	lr=0.029576
Epoch[99] Batch [1099]	Speed: 4984.664405 samples/sec	accuracy=0.806928	lr=0.029462
Epoch[99] Batch [1149]	Speed: 4999.975083 samples/sec	accuracy=0.806998	lr=0.029348
Epoch[99] Batch [1199]	Speed: 5016.878562 samples/sec	accuracy=0.807071	lr=0.029234
Epoch[99] Batch [1249]	Speed: 4969.839723 samples/sec	accuracy=0.807135	lr=0.029121
[Epoch 99] training: accuracy=0.807135
[Epoch 99] speed: 5034 samples/sec	time cost: 263.892481
[Epoch 99] validation: err-top1=0.245954 err-top5=0.072565
Epoch[100] Batch [49]	Speed: 4781.861609 samples/sec	accuracy=0.806602	lr=0.029005
Epoch[100] Batch [99]	Speed: 5008.117639 samples/sec	accuracy=0.806934	lr=0.028892
Epoch[100] Batch [149]	Speed: 5077.936450 samples/sec	accuracy=0.806989	lr=0.028779
Epoch[100] Batch [199]	Speed: 4927.273233 samples/sec	accuracy=0.807057	lr=0.028666
Epoch[100] Batch [249]	Speed: 4914.026193 samples/sec	accuracy=0.807079	lr=0.028554
Epoch[100] Batch [299]	Speed: 5037.463173 samples/sec	accuracy=0.807101	lr=0.028441
Epoch[100] Batch [349]	Speed: 4973.572580 samples/sec	accuracy=0.807206	lr=0.028329
Epoch[100] Batch [399]	Speed: 5108.228542 samples/sec	accuracy=0.807189	lr=0.028217
Epoch[100] Batch [449]	Speed: 5113.309292 samples/sec	accuracy=0.807192	lr=0.028105
Epoch[100] Batch [499]	Speed: 5103.129018 samples/sec	accuracy=0.807332	lr=0.027994
Epoch[100] Batch [549]	Speed: 5075.394461 samples/sec	accuracy=0.807330	lr=0.027883
Epoch[100] Batch [599]	Speed: 5152.028085 samples/sec	accuracy=0.807444	lr=0.027772
Epoch[100] Batch [649]	Speed: 5031.588419 samples/sec	accuracy=0.807648	lr=0.027661
Epoch[100] Batch [699]	Speed: 4922.268024 samples/sec	accuracy=0.807718	lr=0.027550
Epoch[100] Batch [749]	Speed: 5035.838083 samples/sec	accuracy=0.807827	lr=0.027439
Epoch[100] Batch [799]	Speed: 5154.034685 samples/sec	accuracy=0.807854	lr=0.027329
Epoch[100] Batch [849]	Speed: 4962.011608 samples/sec	accuracy=0.807948	lr=0.027219
Epoch[100] Batch [899]	Speed: 5038.769013 samples/sec	accuracy=0.808023	lr=0.027109
Epoch[100] Batch [949]	Speed: 5128.120892 samples/sec	accuracy=0.808053	lr=0.026999
Epoch[100] Batch [999]	Speed: 5057.406054 samples/sec	accuracy=0.808052	lr=0.026890
Epoch[100] Batch [1049]	Speed: 4972.901125 samples/sec	accuracy=0.808117	lr=0.026781
Epoch[100] Batch [1099]	Speed: 5119.878665 samples/sec	accuracy=0.808159	lr=0.026672
Epoch[100] Batch [1149]	Speed: 5100.139110 samples/sec	accuracy=0.808188	lr=0.026563
Epoch[100] Batch [1199]	Speed: 5048.631559 samples/sec	accuracy=0.808223	lr=0.026454
Epoch[100] Batch [1249]	Speed: 5168.257478 samples/sec	accuracy=0.808268	lr=0.026346
[Epoch 100] training: accuracy=0.808269
[Epoch 100] speed: 5035 samples/sec	time cost: 263.986945
[Epoch 100] validation: err-top1=0.241311 err-top5=0.068977
Epoch[101] Batch [49]	Speed: 4868.511218 samples/sec	accuracy=0.809156	lr=0.026235
Epoch[101] Batch [99]	Speed: 5025.117232 samples/sec	accuracy=0.808837	lr=0.026127
Epoch[101] Batch [149]	Speed: 5109.030510 samples/sec	accuracy=0.808872	lr=0.026020
Epoch[101] Batch [199]	Speed: 5149.471405 samples/sec	accuracy=0.808837	lr=0.025912
Epoch[101] Batch [249]	Speed: 5065.388467 samples/sec	accuracy=0.808787	lr=0.025805
Epoch[101] Batch [299]	Speed: 5097.530078 samples/sec	accuracy=0.808785	lr=0.025697
Epoch[101] Batch [349]	Speed: 5073.903528 samples/sec	accuracy=0.808798	lr=0.025590
Epoch[101] Batch [399]	Speed: 4983.344707 samples/sec	accuracy=0.808652	lr=0.025484
Epoch[101] Batch [449]	Speed: 4891.508807 samples/sec	accuracy=0.808678	lr=0.025377
Epoch[101] Batch [499]	Speed: 5070.225574 samples/sec	accuracy=0.808698	lr=0.025271
Epoch[101] Batch [549]	Speed: 5000.333082 samples/sec	accuracy=0.808666	lr=0.025165
Epoch[101] Batch [599]	Speed: 5074.407923 samples/sec	accuracy=0.808683	lr=0.025059
Epoch[101] Batch [649]	Speed: 5093.324765 samples/sec	accuracy=0.808669	lr=0.024953
Epoch[101] Batch [699]	Speed: 5068.675106 samples/sec	accuracy=0.808656	lr=0.024847
Epoch[101] Batch [749]	Speed: 4996.051398 samples/sec	accuracy=0.808664	lr=0.024742
Epoch[101] Batch [799]	Speed: 5056.909916 samples/sec	accuracy=0.808677	lr=0.024637
Epoch[101] Batch [849]	Speed: 5026.959098 samples/sec	accuracy=0.808718	lr=0.024532
Epoch[101] Batch [899]	Speed: 5039.157303 samples/sec	accuracy=0.808731	lr=0.024427
Epoch[101] Batch [949]	Speed: 5139.291900 samples/sec	accuracy=0.808784	lr=0.024323
Epoch[101] Batch [999]	Speed: 5161.701904 samples/sec	accuracy=0.808838	lr=0.024219
Epoch[101] Batch [1049]	Speed: 5036.265959 samples/sec	accuracy=0.808941	lr=0.024115
Epoch[101] Batch [1099]	Speed: 5158.418330 samples/sec	accuracy=0.809000	lr=0.024011
Epoch[101] Batch [1149]	Speed: 5150.453626 samples/sec	accuracy=0.809054	lr=0.023907
Epoch[101] Batch [1199]	Speed: 5161.706495 samples/sec	accuracy=0.809128	lr=0.023804
Epoch[101] Batch [1249]	Speed: 5032.050830 samples/sec	accuracy=0.809181	lr=0.023700
[Epoch 101] training: accuracy=0.809180
[Epoch 101] speed: 5056 samples/sec	time cost: 262.656025
[Epoch 101] validation: err-top1=0.240513 err-top5=0.070332
Epoch[102] Batch [49]	Speed: 4771.638418 samples/sec	accuracy=0.810184	lr=0.023595
Epoch[102] Batch [99]	Speed: 5080.217296 samples/sec	accuracy=0.809830	lr=0.023493
Epoch[102] Batch [149]	Speed: 5032.449524 samples/sec	accuracy=0.809631	lr=0.023390
Epoch[102] Batch [199]	Speed: 5005.388141 samples/sec	accuracy=0.809685	lr=0.023288
Epoch[102] Batch [249]	Speed: 5109.306317 samples/sec	accuracy=0.809567	lr=0.023185
Epoch[102] Batch [299]	Speed: 5010.317128 samples/sec	accuracy=0.809408	lr=0.023083
Epoch[102] Batch [349]	Speed: 5110.867393 samples/sec	accuracy=0.809355	lr=0.022982
Epoch[102] Batch [399]	Speed: 5008.769198 samples/sec	accuracy=0.809460	lr=0.022880
Epoch[102] Batch [449]	Speed: 4918.627864 samples/sec	accuracy=0.809485	lr=0.022779
Epoch[102] Batch [499]	Speed: 5090.136631 samples/sec	accuracy=0.809612	lr=0.022678
Epoch[102] Batch [549]	Speed: 4949.726867 samples/sec	accuracy=0.809741	lr=0.022577
Epoch[102] Batch [599]	Speed: 5196.549387 samples/sec	accuracy=0.809820	lr=0.022476
Epoch[102] Batch [649]	Speed: 5101.200507 samples/sec	accuracy=0.809879	lr=0.022376
Epoch[102] Batch [699]	Speed: 5122.762720 samples/sec	accuracy=0.809896	lr=0.022276
Epoch[102] Batch [749]	Speed: 5058.642773 samples/sec	accuracy=0.809946	lr=0.022175
Epoch[102] Batch [799]	Speed: 4884.548186 samples/sec	accuracy=0.810013	lr=0.022076
Epoch[102] Batch [849]	Speed: 5052.570896 samples/sec	accuracy=0.810050	lr=0.021976
Epoch[102] Batch [899]	Speed: 4995.077684 samples/sec	accuracy=0.810142	lr=0.021877
Epoch[102] Batch [949]	Speed: 5111.900163 samples/sec	accuracy=0.810192	lr=0.021777
Epoch[102] Batch [999]	Speed: 5171.476878 samples/sec	accuracy=0.810181	lr=0.021678
Epoch[102] Batch [1049]	Speed: 5003.712207 samples/sec	accuracy=0.810247	lr=0.021580
Epoch[102] Batch [1099]	Speed: 5077.600028 samples/sec	accuracy=0.810313	lr=0.021481
Epoch[102] Batch [1149]	Speed: 5102.076393 samples/sec	accuracy=0.810368	lr=0.021383
Epoch[102] Batch [1199]	Speed: 5025.383582 samples/sec	accuracy=0.810409	lr=0.021285
Epoch[102] Batch [1249]	Speed: 5046.017321 samples/sec	accuracy=0.810461	lr=0.021187
[Epoch 102] training: accuracy=0.810461
[Epoch 102] speed: 5035 samples/sec	time cost: 263.700845
[Epoch 102] validation: err-top1=0.237883 err-top5=0.066765
Epoch[103] Batch [49]	Speed: 4826.809385 samples/sec	accuracy=0.811796	lr=0.021087
Epoch[103] Batch [99]	Speed: 4904.716333 samples/sec	accuracy=0.810922	lr=0.020990
Epoch[103] Batch [149]	Speed: 4953.791781 samples/sec	accuracy=0.810718	lr=0.020892
Epoch[103] Batch [199]	Speed: 5084.269293 samples/sec	accuracy=0.810798	lr=0.020795
Epoch[103] Batch [249]	Speed: 4995.420224 samples/sec	accuracy=0.810576	lr=0.020698
Epoch[103] Batch [299]	Speed: 5155.674708 samples/sec	accuracy=0.810620	lr=0.020602
Epoch[103] Batch [349]	Speed: 4963.604306 samples/sec	accuracy=0.810685	lr=0.020505
Epoch[103] Batch [399]	Speed: 5057.232168 samples/sec	accuracy=0.810774	lr=0.020409
Epoch[103] Batch [449]	Speed: 5034.188667 samples/sec	accuracy=0.810819	lr=0.020313
Epoch[103] Batch [499]	Speed: 5117.914912 samples/sec	accuracy=0.810779	lr=0.020217
Epoch[103] Batch [549]	Speed: 5039.477651 samples/sec	accuracy=0.810750	lr=0.020122
Epoch[103] Batch [599]	Speed: 5039.343784 samples/sec	accuracy=0.810834	lr=0.020026
Epoch[103] Batch [649]	Speed: 5073.231917 samples/sec	accuracy=0.810872	lr=0.019931
Epoch[103] Batch [699]	Speed: 4940.801041 samples/sec	accuracy=0.810876	lr=0.019836
Epoch[103] Batch [749]	Speed: 5089.421757 samples/sec	accuracy=0.810863	lr=0.019742
Epoch[103] Batch [799]	Speed: 5012.724125 samples/sec	accuracy=0.810963	lr=0.019647
Epoch[103] Batch [849]	Speed: 5024.289314 samples/sec	accuracy=0.811011	lr=0.019553
Epoch[103] Batch [899]	Speed: 5119.339197 samples/sec	accuracy=0.811050	lr=0.019459
Epoch[103] Batch [949]	Speed: 5005.187950 samples/sec	accuracy=0.811116	lr=0.019365
Epoch[103] Batch [999]	Speed: 5113.445535 samples/sec	accuracy=0.811174	lr=0.019271
Epoch[103] Batch [1049]	Speed: 5002.101942 samples/sec	accuracy=0.811196	lr=0.019178
Epoch[103] Batch [1099]	Speed: 5099.679845 samples/sec	accuracy=0.811238	lr=0.019085
Epoch[103] Batch [1149]	Speed: 5180.101120 samples/sec	accuracy=0.811278	lr=0.018992
Epoch[103] Batch [1199]	Speed: 5150.345419 samples/sec	accuracy=0.811331	lr=0.018899
Epoch[103] Batch [1249]	Speed: 4985.687539 samples/sec	accuracy=0.811340	lr=0.018806
[Epoch 103] training: accuracy=0.811338
[Epoch 103] speed: 5033 samples/sec	time cost: 263.612076
[Epoch 103] validation: err-top1=0.236228 err-top5=0.066526
Epoch[104] Batch [49]	Speed: 4841.517382 samples/sec	accuracy=0.811538	lr=0.018712
Epoch[104] Batch [99]	Speed: 5018.436903 samples/sec	accuracy=0.811534	lr=0.018620
Epoch[104] Batch [149]	Speed: 5178.436532 samples/sec	accuracy=0.811578	lr=0.018528
Epoch[104] Batch [199]	Speed: 4911.053339 samples/sec	accuracy=0.811621	lr=0.018436
Epoch[104] Batch [249]	Speed: 4910.262465 samples/sec	accuracy=0.811706	lr=0.018345
Epoch[104] Batch [299]	Speed: 4800.654298 samples/sec	accuracy=0.811629	lr=0.018254
Epoch[104] Batch [349]	Speed: 4917.166912 samples/sec	accuracy=0.811637	lr=0.018163
Epoch[104] Batch [399]	Speed: 4966.448863 samples/sec	accuracy=0.811699	lr=0.018072
Epoch[104] Batch [449]	Speed: 5149.272116 samples/sec	accuracy=0.811728	lr=0.017981
Epoch[104] Batch [499]	Speed: 4950.333766 samples/sec	accuracy=0.811788	lr=0.017891
Epoch[104] Batch [549]	Speed: 4822.147079 samples/sec	accuracy=0.811779	lr=0.017801
Epoch[104] Batch [599]	Speed: 4963.675438 samples/sec	accuracy=0.811906	lr=0.017711
Epoch[104] Batch [649]	Speed: 4993.871846 samples/sec	accuracy=0.811991	lr=0.017621
Epoch[104] Batch [699]	Speed: 4865.334332 samples/sec	accuracy=0.812046	lr=0.017532
Epoch[104] Batch [749]	Speed: 4796.560191 samples/sec	accuracy=0.812147	lr=0.017442
Epoch[104] Batch [799]	Speed: 4951.634551 samples/sec	accuracy=0.812170	lr=0.017353
Epoch[104] Batch [849]	Speed: 4856.445099 samples/sec	accuracy=0.812235	lr=0.017264
Epoch[104] Batch [899]	Speed: 5003.471697 samples/sec	accuracy=0.812289	lr=0.017176
Epoch[104] Batch [949]	Speed: 5070.840232 samples/sec	accuracy=0.812350	lr=0.017087
Epoch[104] Batch [999]	Speed: 4806.711962 samples/sec	accuracy=0.812391	lr=0.016999
Epoch[104] Batch [1049]	Speed: 4993.705669 samples/sec	accuracy=0.812451	lr=0.016911
Epoch[104] Batch [1099]	Speed: 5090.063880 samples/sec	accuracy=0.812463	lr=0.016823
Epoch[104] Batch [1149]	Speed: 5008.522829 samples/sec	accuracy=0.812492	lr=0.016736
Epoch[104] Batch [1199]	Speed: 4974.722191 samples/sec	accuracy=0.812550	lr=0.016648
Epoch[104] Batch [1249]	Speed: 5154.857659 samples/sec	accuracy=0.812626	lr=0.016561
[Epoch 104] training: accuracy=0.812627
[Epoch 104] speed: 4952 samples/sec	time cost: 267.957423
[Epoch 104] validation: err-top1=0.231893 err-top5=0.065470
Epoch[105] Batch [49]	Speed: 4623.055763 samples/sec	accuracy=0.812809	lr=0.016473
Epoch[105] Batch [99]	Speed: 4948.161303 samples/sec	accuracy=0.812657	lr=0.016386
Epoch[105] Batch [149]	Speed: 4978.069779 samples/sec	accuracy=0.812879	lr=0.016299
Epoch[105] Batch [199]	Speed: 5095.313695 samples/sec	accuracy=0.812837	lr=0.016213
Epoch[105] Batch [249]	Speed: 4942.622558 samples/sec	accuracy=0.812988	lr=0.016127
Epoch[105] Batch [299]	Speed: 4795.125445 samples/sec	accuracy=0.812913	lr=0.016041
Epoch[105] Batch [349]	Speed: 4970.512423 samples/sec	accuracy=0.812934	lr=0.015956
Epoch[105] Batch [399]	Speed: 4900.872672 samples/sec	accuracy=0.812965	lr=0.015870
Epoch[105] Batch [449]	Speed: 4868.178356 samples/sec	accuracy=0.812958	lr=0.015785
Epoch[105] Batch [499]	Speed: 5236.788143 samples/sec	accuracy=0.812980	lr=0.015700
Epoch[105] Batch [549]	Speed: 4877.516725 samples/sec	accuracy=0.813034	lr=0.015616
Epoch[105] Batch [599]	Speed: 5068.504512 samples/sec	accuracy=0.813068	lr=0.015531
Epoch[105] Batch [649]	Speed: 4972.465525 samples/sec	accuracy=0.813121	lr=0.015447
Epoch[105] Batch [699]	Speed: 5059.165591 samples/sec	accuracy=0.813186	lr=0.015363
Epoch[105] Batch [749]	Speed: 4871.255199 samples/sec	accuracy=0.813262	lr=0.015279
Epoch[105] Batch [799]	Speed: 5000.620682 samples/sec	accuracy=0.813273	lr=0.015195
Epoch[105] Batch [849]	Speed: 5148.460182 samples/sec	accuracy=0.813291	lr=0.015112
Epoch[105] Batch [899]	Speed: 4889.675775 samples/sec	accuracy=0.813293	lr=0.015029
Epoch[105] Batch [949]	Speed: 4938.743699 samples/sec	accuracy=0.813350	lr=0.014946
Epoch[105] Batch [999]	Speed: 4928.497339 samples/sec	accuracy=0.813418	lr=0.014863
Epoch[105] Batch [1049]	Speed: 5006.412098 samples/sec	accuracy=0.813486	lr=0.014781
Epoch[105] Batch [1099]	Speed: 4971.249288 samples/sec	accuracy=0.813551	lr=0.014699
Epoch[105] Batch [1149]	Speed: 5063.463895 samples/sec	accuracy=0.813575	lr=0.014616
Epoch[105] Batch [1199]	Speed: 5218.132218 samples/sec	accuracy=0.813602	lr=0.014535
Epoch[105] Batch [1249]	Speed: 5308.266149 samples/sec	accuracy=0.813635	lr=0.014453
[Epoch 105] training: accuracy=0.813636
[Epoch 105] speed: 4977 samples/sec	time cost: 267.087842
[Epoch 105] validation: err-top1=0.231864 err-top5=0.065151
Epoch[106] Batch [49]	Speed: 4761.685521 samples/sec	accuracy=0.814406	lr=0.014370
Epoch[106] Batch [99]	Speed: 4978.693690 samples/sec	accuracy=0.814614	lr=0.014289
Epoch[106] Batch [149]	Speed: 5067.236901 samples/sec	accuracy=0.814230	lr=0.014208
Epoch[106] Batch [199]	Speed: 5042.288743 samples/sec	accuracy=0.814226	lr=0.014127
Epoch[106] Batch [249]	Speed: 4926.095499 samples/sec	accuracy=0.814134	lr=0.014047
Epoch[106] Batch [299]	Speed: 5109.373541 samples/sec	accuracy=0.813999	lr=0.013966
Epoch[106] Batch [349]	Speed: 4998.560003 samples/sec	accuracy=0.813999	lr=0.013886
Epoch[106] Batch [399]	Speed: 4785.838819 samples/sec	accuracy=0.813926	lr=0.013806
Epoch[106] Batch [449]	Speed: 5058.236344 samples/sec	accuracy=0.813985	lr=0.013727
Epoch[106] Batch [499]	Speed: 5084.484408 samples/sec	accuracy=0.813990	lr=0.013647
Epoch[106] Batch [549]	Speed: 5155.193506 samples/sec	accuracy=0.814062	lr=0.013568
Epoch[106] Batch [599]	Speed: 5130.492896 samples/sec	accuracy=0.814087	lr=0.013489
Epoch[106] Batch [649]	Speed: 5054.599144 samples/sec	accuracy=0.814162	lr=0.013411
Epoch[106] Batch [699]	Speed: 5015.081904 samples/sec	accuracy=0.814221	lr=0.013332
Epoch[106] Batch [749]	Speed: 5137.475091 samples/sec	accuracy=0.814211	lr=0.013254
Epoch[106] Batch [799]	Speed: 5047.501756 samples/sec	accuracy=0.814279	lr=0.013176
Epoch[106] Batch [849]	Speed: 5085.540383 samples/sec	accuracy=0.814328	lr=0.013098
Epoch[106] Batch [899]	Speed: 5032.475233 samples/sec	accuracy=0.814404	lr=0.013020
Epoch[106] Batch [949]	Speed: 5042.664786 samples/sec	accuracy=0.814450	lr=0.012943
Epoch[106] Batch [999]	Speed: 5019.088218 samples/sec	accuracy=0.814489	lr=0.012866
Epoch[106] Batch [1049]	Speed: 4990.666390 samples/sec	accuracy=0.814507	lr=0.012789
Epoch[106] Batch [1099]	Speed: 5104.024005 samples/sec	accuracy=0.814582	lr=0.012712
Epoch[106] Batch [1149]	Speed: 5018.696095 samples/sec	accuracy=0.814634	lr=0.012636
Epoch[106] Batch [1199]	Speed: 5032.452000 samples/sec	accuracy=0.814667	lr=0.012559
Epoch[106] Batch [1249]	Speed: 5317.483409 samples/sec	accuracy=0.814747	lr=0.012483
[Epoch 106] training: accuracy=0.814750
[Epoch 106] speed: 5033 samples/sec	time cost: 264.015512
[Epoch 106] validation: err-top1=0.228257 err-top5=0.064373
Epoch[107] Batch [49]	Speed: 4899.493451 samples/sec	accuracy=0.815670	lr=0.012406
Epoch[107] Batch [99]	Speed: 4919.893214 samples/sec	accuracy=0.815479	lr=0.012330
Epoch[107] Batch [149]	Speed: 5041.696848 samples/sec	accuracy=0.815263	lr=0.012255
Epoch[107] Batch [199]	Speed: 5078.122209 samples/sec	accuracy=0.815155	lr=0.012180
Epoch[107] Batch [249]	Speed: 5082.721774 samples/sec	accuracy=0.815208	lr=0.012105
Epoch[107] Batch [299]	Speed: 5175.773119 samples/sec	accuracy=0.815230	lr=0.012030
Epoch[107] Batch [349]	Speed: 5157.159349 samples/sec	accuracy=0.815171	lr=0.011956
Epoch[107] Batch [399]	Speed: 5004.914054 samples/sec	accuracy=0.815179	lr=0.011881
Epoch[107] Batch [449]	Speed: 5146.025543 samples/sec	accuracy=0.815138	lr=0.011807
Epoch[107] Batch [499]	Speed: 4992.390929 samples/sec	accuracy=0.815099	lr=0.011734
Epoch[107] Batch [549]	Speed: 5081.149108 samples/sec	accuracy=0.815144	lr=0.011660
Epoch[107] Batch [599]	Speed: 5130.216267 samples/sec	accuracy=0.815162	lr=0.011587
Epoch[107] Batch [649]	Speed: 4984.696917 samples/sec	accuracy=0.815212	lr=0.011513
Epoch[107] Batch [699]	Speed: 5018.837783 samples/sec	accuracy=0.815208	lr=0.011441
Epoch[107] Batch [749]	Speed: 5075.823568 samples/sec	accuracy=0.815239	lr=0.011368
Epoch[107] Batch [799]	Speed: 5013.414337 samples/sec	accuracy=0.815311	lr=0.011295
Epoch[107] Batch [849]	Speed: 5037.537028 samples/sec	accuracy=0.815344	lr=0.011223
Epoch[107] Batch [899]	Speed: 5063.413275 samples/sec	accuracy=0.815372	lr=0.011151
Epoch[107] Batch [949]	Speed: 4970.138321 samples/sec	accuracy=0.815426	lr=0.011079
Epoch[107] Batch [999]	Speed: 5055.754504 samples/sec	accuracy=0.815449	lr=0.011008
Epoch[107] Batch [1049]	Speed: 5159.316580 samples/sec	accuracy=0.815527	lr=0.010936
Epoch[107] Batch [1099]	Speed: 5134.979243 samples/sec	accuracy=0.815590	lr=0.010865
Epoch[107] Batch [1149]	Speed: 5125.337957 samples/sec	accuracy=0.815636	lr=0.010794
Epoch[107] Batch [1199]	Speed: 5039.446076 samples/sec	accuracy=0.815672	lr=0.010724
Epoch[107] Batch [1249]	Speed: 5224.949850 samples/sec	accuracy=0.815722	lr=0.010653
[Epoch 107] training: accuracy=0.815725
[Epoch 107] speed: 5059 samples/sec	time cost: 262.679726
[Epoch 107] validation: err-top1=0.227160 err-top5=0.061723
Epoch[108] Batch [49]	Speed: 4857.035708 samples/sec	accuracy=0.816356	lr=0.010582
Epoch[108] Batch [99]	Speed: 5016.298593 samples/sec	accuracy=0.816497	lr=0.010512
Epoch[108] Batch [149]	Speed: 5040.339921 samples/sec	accuracy=0.816417	lr=0.010442
Epoch[108] Batch [199]	Speed: 4909.822278 samples/sec	accuracy=0.816612	lr=0.010373
Epoch[108] Batch [249]	Speed: 5055.860796 samples/sec	accuracy=0.816632	lr=0.010303
Epoch[108] Batch [299]	Speed: 4900.243177 samples/sec	accuracy=0.816452	lr=0.010234
Epoch[108] Batch [349]	Speed: 5125.644399 samples/sec	accuracy=0.816502	lr=0.010165
Epoch[108] Batch [399]	Speed: 5135.048863 samples/sec	accuracy=0.816494	lr=0.010097
Epoch[108] Batch [449]	Speed: 5106.788209 samples/sec	accuracy=0.816546	lr=0.010028
Epoch[108] Batch [499]	Speed: 5104.790312 samples/sec	accuracy=0.816577	lr=0.009960
Epoch[108] Batch [549]	Speed: 4969.236772 samples/sec	accuracy=0.816636	lr=0.009892
Epoch[108] Batch [599]	Speed: 5003.156726 samples/sec	accuracy=0.816698	lr=0.009824
Epoch[108] Batch [649]	Speed: 5087.493467 samples/sec	accuracy=0.816715	lr=0.009757
Epoch[108] Batch [699]	Speed: 5036.119861 samples/sec	accuracy=0.816725	lr=0.009690
Epoch[108] Batch [749]	Speed: 5036.444903 samples/sec	accuracy=0.816711	lr=0.009623
Epoch[108] Batch [799]	Speed: 5065.083452 samples/sec	accuracy=0.816668	lr=0.009556
Epoch[108] Batch [849]	Speed: 4909.842147 samples/sec	accuracy=0.816663	lr=0.009489
Epoch[108] Batch [899]	Speed: 5119.885867 samples/sec	accuracy=0.816647	lr=0.009423
Epoch[108] Batch [949]	Speed: 5053.428614 samples/sec	accuracy=0.816653	lr=0.009357
Epoch[108] Batch [999]	Speed: 4940.361612 samples/sec	accuracy=0.816653	lr=0.009291
Epoch[108] Batch [1049]	Speed: 5049.498389 samples/sec	accuracy=0.816705	lr=0.009225
Epoch[108] Batch [1099]	Speed: 5198.595983 samples/sec	accuracy=0.816710	lr=0.009160
Epoch[108] Batch [1149]	Speed: 5179.519404 samples/sec	accuracy=0.816734	lr=0.009095
Epoch[108] Batch [1199]	Speed: 5094.895429 samples/sec	accuracy=0.816760	lr=0.009030
Epoch[108] Batch [1249]	Speed: 5265.339708 samples/sec	accuracy=0.816779	lr=0.008965
[Epoch 108] training: accuracy=0.816781
[Epoch 108] speed: 5044 samples/sec	time cost: 263.032140
[Epoch 108] validation: err-top1=0.222935 err-top5=0.062161
Epoch[109] Batch [49]	Speed: 4931.246755 samples/sec	accuracy=0.817594	lr=0.008899
Epoch[109] Batch [99]	Speed: 5071.285814 samples/sec	accuracy=0.817276	lr=0.008835
Epoch[109] Batch [149]	Speed: 5110.679595 samples/sec	accuracy=0.817009	lr=0.008771
Epoch[109] Batch [199]	Speed: 5144.201382 samples/sec	accuracy=0.817073	lr=0.008707
Epoch[109] Batch [249]	Speed: 5073.050470 samples/sec	accuracy=0.817093	lr=0.008643
Epoch[109] Batch [299]	Speed: 5019.691125 samples/sec	accuracy=0.817227	lr=0.008580
Epoch[109] Batch [349]	Speed: 5071.400306 samples/sec	accuracy=0.817293	lr=0.008517
Epoch[109] Batch [399]	Speed: 4995.524111 samples/sec	accuracy=0.817346	lr=0.008454
Epoch[109] Batch [449]	Speed: 5172.224459 samples/sec	accuracy=0.817385	lr=0.008391
Epoch[109] Batch [499]	Speed: 5231.463633 samples/sec	accuracy=0.817410	lr=0.008329
Epoch[109] Batch [549]	Speed: 5150.988429 samples/sec	accuracy=0.817425	lr=0.008266
Epoch[109] Batch [599]	Speed: 4847.426234 samples/sec	accuracy=0.817452	lr=0.008204
Epoch[109] Batch [649]	Speed: 5097.995612 samples/sec	accuracy=0.817514	lr=0.008143
Epoch[109] Batch [699]	Speed: 4922.328160 samples/sec	accuracy=0.817508	lr=0.008081
Epoch[109] Batch [749]	Speed: 4955.232160 samples/sec	accuracy=0.817563	lr=0.008020
Epoch[109] Batch [799]	Speed: 5214.136057 samples/sec	accuracy=0.817597	lr=0.007959
Epoch[109] Batch [849]	Speed: 5003.408047 samples/sec	accuracy=0.817582	lr=0.007898
Epoch[109] Batch [899]	Speed: 4894.719086 samples/sec	accuracy=0.817598	lr=0.007837
Epoch[109] Batch [949]	Speed: 4956.067669 samples/sec	accuracy=0.817620	lr=0.007777
Epoch[109] Batch [999]	Speed: 5143.587415 samples/sec	accuracy=0.817681	lr=0.007716
Epoch[109] Batch [1049]	Speed: 4967.388926 samples/sec	accuracy=0.817715	lr=0.007656
Epoch[109] Batch [1099]	Speed: 4931.526463 samples/sec	accuracy=0.817723	lr=0.007597
Epoch[109] Batch [1149]	Speed: 5041.489836 samples/sec	accuracy=0.817745	lr=0.007537
Epoch[109] Batch [1199]	Speed: 5034.053074 samples/sec	accuracy=0.817759	lr=0.007478
Epoch[109] Batch [1249]	Speed: 5260.285947 samples/sec	accuracy=0.817772	lr=0.007419
[Epoch 109] training: accuracy=0.817771
[Epoch 109] speed: 5043 samples/sec	time cost: 263.295903
[Epoch 109] validation: err-top1=0.222915 err-top5=0.060906
Epoch[110] Batch [49]	Speed: 4922.096200 samples/sec	accuracy=0.818208	lr=0.007359
Epoch[110] Batch [99]	Speed: 5080.533872 samples/sec	accuracy=0.818303	lr=0.007300
Epoch[110] Batch [149]	Speed: 4913.354643 samples/sec	accuracy=0.818270	lr=0.007242
Epoch[110] Batch [199]	Speed: 4896.368228 samples/sec	accuracy=0.818260	lr=0.007184
Epoch[110] Batch [249]	Speed: 5104.247831 samples/sec	accuracy=0.818291	lr=0.007126
Epoch[110] Batch [299]	Speed: 4832.994514 samples/sec	accuracy=0.818214	lr=0.007068
Epoch[110] Batch [349]	Speed: 5012.798310 samples/sec	accuracy=0.818204	lr=0.007011
Epoch[110] Batch [399]	Speed: 4945.091674 samples/sec	accuracy=0.818181	lr=0.006954
Epoch[110] Batch [449]	Speed: 5004.272243 samples/sec	accuracy=0.818187	lr=0.006897
Epoch[110] Batch [499]	Speed: 4970.357460 samples/sec	accuracy=0.818176	lr=0.006840
Epoch[110] Batch [549]	Speed: 4976.203136 samples/sec	accuracy=0.818163	lr=0.006784
Epoch[110] Batch [599]	Speed: 5178.375844 samples/sec	accuracy=0.818140	lr=0.006727
Epoch[110] Batch [649]	Speed: 5129.214062 samples/sec	accuracy=0.818183	lr=0.006671
Epoch[110] Batch [699]	Speed: 5014.829994 samples/sec	accuracy=0.818233	lr=0.006615
Epoch[110] Batch [749]	Speed: 5159.865127 samples/sec	accuracy=0.818222	lr=0.006560
Epoch[110] Batch [799]	Speed: 4986.895796 samples/sec	accuracy=0.818247	lr=0.006504
Epoch[110] Batch [849]	Speed: 5072.119230 samples/sec	accuracy=0.818262	lr=0.006449
Epoch[110] Batch [899]	Speed: 4962.469804 samples/sec	accuracy=0.818309	lr=0.006394
Epoch[110] Batch [949]	Speed: 5171.717620 samples/sec	accuracy=0.818355	lr=0.006340
Epoch[110] Batch [999]	Speed: 4904.995841 samples/sec	accuracy=0.818384	lr=0.006285
Epoch[110] Batch [1049]	Speed: 4902.319368 samples/sec	accuracy=0.818414	lr=0.006231
Epoch[110] Batch [1099]	Speed: 5014.317352 samples/sec	accuracy=0.818396	lr=0.006177
Epoch[110] Batch [1149]	Speed: 5077.658977 samples/sec	accuracy=0.818383	lr=0.006123
Epoch[110] Batch [1199]	Speed: 5167.937214 samples/sec	accuracy=0.818424	lr=0.006070
Epoch[110] Batch [1249]	Speed: 5133.486235 samples/sec	accuracy=0.818475	lr=0.006017
[Epoch 110] training: accuracy=0.818475
[Epoch 110] speed: 5015 samples/sec	time cost: 264.709943
[Epoch 110] validation: err-top1=0.222127 err-top5=0.060099
Epoch[111] Batch [49]	Speed: 4787.008911 samples/sec	accuracy=0.819340	lr=0.005963
Epoch[111] Batch [99]	Speed: 4905.404685 samples/sec	accuracy=0.819427	lr=0.005910
Epoch[111] Batch [149]	Speed: 5013.570357 samples/sec	accuracy=0.819526	lr=0.005857
Epoch[111] Batch [199]	Speed: 4935.817446 samples/sec	accuracy=0.819194	lr=0.005805
Epoch[111] Batch [249]	Speed: 4951.411351 samples/sec	accuracy=0.819093	lr=0.005753
Epoch[111] Batch [299]	Speed: 5098.577195 samples/sec	accuracy=0.819011	lr=0.005701
Epoch[111] Batch [349]	Speed: 5015.494664 samples/sec	accuracy=0.818958	lr=0.005649
Epoch[111] Batch [399]	Speed: 5035.638990 samples/sec	accuracy=0.818937	lr=0.005598
Epoch[111] Batch [449]	Speed: 5055.379123 samples/sec	accuracy=0.818950	lr=0.005547
Epoch[111] Batch [499]	Speed: 4929.359951 samples/sec	accuracy=0.818968	lr=0.005496
Epoch[111] Batch [549]	Speed: 5004.022235 samples/sec	accuracy=0.818969	lr=0.005445
Epoch[111] Batch [599]	Speed: 4921.098094 samples/sec	accuracy=0.818976	lr=0.005394
Epoch[111] Batch [649]	Speed: 5117.808068 samples/sec	accuracy=0.819028	lr=0.005344
Epoch[111] Batch [699]	Speed: 4873.922961 samples/sec	accuracy=0.819066	lr=0.005294
Epoch[111] Batch [749]	Speed: 5167.466777 samples/sec	accuracy=0.819149	lr=0.005244
Epoch[111] Batch [799]	Speed: 4984.945809 samples/sec	accuracy=0.819194	lr=0.005195
Epoch[111] Batch [849]	Speed: 5086.092627 samples/sec	accuracy=0.819192	lr=0.005145
Epoch[111] Batch [899]	Speed: 4837.756078 samples/sec	accuracy=0.819200	lr=0.005096
Epoch[111] Batch [949]	Speed: 5047.072798 samples/sec	accuracy=0.819198	lr=0.005047
Epoch[111] Batch [999]	Speed: 5216.146612 samples/sec	accuracy=0.819227	lr=0.004999
Epoch[111] Batch [1049]	Speed: 5199.770651 samples/sec	accuracy=0.819280	lr=0.004950
Epoch[111] Batch [1099]	Speed: 5011.376432 samples/sec	accuracy=0.819300	lr=0.004902
Epoch[111] Batch [1149]	Speed: 4853.020271 samples/sec	accuracy=0.819322	lr=0.004854
Epoch[111] Batch [1199]	Speed: 5437.155293 samples/sec	accuracy=0.819342	lr=0.004807
Epoch[111] Batch [1249]	Speed: 5221.493294 samples/sec	accuracy=0.819370	lr=0.004759
[Epoch 111] training: accuracy=0.819371
[Epoch 111] speed: 5019 samples/sec	time cost: 264.656979
[Epoch 111] validation: err-top1=0.219527 err-top5=0.060049
Epoch[112] Batch [49]	Speed: 4856.647298 samples/sec	accuracy=0.819635	lr=0.004711
Epoch[112] Batch [99]	Speed: 5024.438136 samples/sec	accuracy=0.819897	lr=0.004664
Epoch[112] Batch [149]	Speed: 5056.245296 samples/sec	accuracy=0.819716	lr=0.004617
Epoch[112] Batch [199]	Speed: 4953.281488 samples/sec	accuracy=0.819675	lr=0.004571
Epoch[112] Batch [249]	Speed: 5051.658091 samples/sec	accuracy=0.819753	lr=0.004524
Epoch[112] Batch [299]	Speed: 4932.885821 samples/sec	accuracy=0.819860	lr=0.004478
Epoch[112] Batch [349]	Speed: 4960.553420 samples/sec	accuracy=0.819826	lr=0.004432
Epoch[112] Batch [399]	Speed: 5162.890118 samples/sec	accuracy=0.819814	lr=0.004387
Epoch[112] Batch [449]	Speed: 4974.105497 samples/sec	accuracy=0.819783	lr=0.004341
Epoch[112] Batch [499]	Speed: 5070.297879 samples/sec	accuracy=0.819818	lr=0.004296
Epoch[112] Batch [549]	Speed: 5002.451390 samples/sec	accuracy=0.819843	lr=0.004251
Epoch[112] Batch [599]	Speed: 5035.032718 samples/sec	accuracy=0.819891	lr=0.004207
Epoch[112] Batch [649]	Speed: 5072.759870 samples/sec	accuracy=0.819925	lr=0.004162
Epoch[112] Batch [699]	Speed: 5015.131094 samples/sec	accuracy=0.819920	lr=0.004118
Epoch[112] Batch [749]	Speed: 5062.266823 samples/sec	accuracy=0.819985	lr=0.004074
Epoch[112] Batch [799]	Speed: 5145.126244 samples/sec	accuracy=0.820011	lr=0.004030
Epoch[112] Batch [849]	Speed: 5185.175554 samples/sec	accuracy=0.820054	lr=0.003987
Epoch[112] Batch [899]	Speed: 5040.458580 samples/sec	accuracy=0.820076	lr=0.003944
Epoch[112] Batch [949]	Speed: 5027.482213 samples/sec	accuracy=0.820090	lr=0.003901
Epoch[112] Batch [999]	Speed: 5058.074434 samples/sec	accuracy=0.820083	lr=0.003858
Epoch[112] Batch [1049]	Speed: 4989.157817 samples/sec	accuracy=0.820089	lr=0.003815
Epoch[112] Batch [1099]	Speed: 5096.276813 samples/sec	accuracy=0.820095	lr=0.003773
Epoch[112] Batch [1149]	Speed: 5033.555254 samples/sec	accuracy=0.820123	lr=0.003731
Epoch[112] Batch [1199]	Speed: 4876.972184 samples/sec	accuracy=0.820146	lr=0.003689
Epoch[112] Batch [1249]	Speed: 5226.071469 samples/sec	accuracy=0.820182	lr=0.003647
[Epoch 112] training: accuracy=0.820182
[Epoch 112] speed: 5030 samples/sec	time cost: 263.874983
[Epoch 112] validation: err-top1=0.219308 err-top5=0.059411
Epoch[113] Batch [49]	Speed: 4589.099085 samples/sec	accuracy=0.820576	lr=0.003605
Epoch[113] Batch [99]	Speed: 4967.163958 samples/sec	accuracy=0.820536	lr=0.003564
Epoch[113] Batch [149]	Speed: 4894.928836 samples/sec	accuracy=0.820457	lr=0.003523
Epoch[113] Batch [199]	Speed: 5155.259468 samples/sec	accuracy=0.820404	lr=0.003482
Epoch[113] Batch [249]	Speed: 5010.681987 samples/sec	accuracy=0.820449	lr=0.003442
Epoch[113] Batch [299]	Speed: 5039.579358 samples/sec	accuracy=0.820378	lr=0.003402
Epoch[113] Batch [349]	Speed: 5082.320127 samples/sec	accuracy=0.820337	lr=0.003362
Epoch[113] Batch [399]	Speed: 5002.975828 samples/sec	accuracy=0.820334	lr=0.003322
Epoch[113] Batch [449]	Speed: 4986.998171 samples/sec	accuracy=0.820324	lr=0.003282
Epoch[113] Batch [499]	Speed: 5053.124444 samples/sec	accuracy=0.820253	lr=0.003243
Epoch[113] Batch [549]	Speed: 4955.820510 samples/sec	accuracy=0.820248	lr=0.003204
Epoch[113] Batch [599]	Speed: 4919.384585 samples/sec	accuracy=0.820290	lr=0.003165
Epoch[113] Batch [649]	Speed: 4994.529345 samples/sec	accuracy=0.820310	lr=0.003127
Epoch[113] Batch [699]	Speed: 5017.734053 samples/sec	accuracy=0.820332	lr=0.003088
Epoch[113] Batch [749]	Speed: 4848.462763 samples/sec	accuracy=0.820304	lr=0.003050
Epoch[113] Batch [799]	Speed: 5187.775088 samples/sec	accuracy=0.820317	lr=0.003012
Epoch[113] Batch [849]	Speed: 4857.829747 samples/sec	accuracy=0.820344	lr=0.002975
Epoch[113] Batch [899]	Speed: 5030.499814 samples/sec	accuracy=0.820364	lr=0.002937
Epoch[113] Batch [949]	Speed: 5002.672340 samples/sec	accuracy=0.820412	lr=0.002900
Epoch[113] Batch [999]	Speed: 4906.824010 samples/sec	accuracy=0.820430	lr=0.002863
Epoch[113] Batch [1049]	Speed: 4848.513774 samples/sec	accuracy=0.820469	lr=0.002826
Epoch[113] Batch [1099]	Speed: 5090.193338 samples/sec	accuracy=0.820494	lr=0.002790
Epoch[113] Batch [1149]	Speed: 5105.735531 samples/sec	accuracy=0.820517	lr=0.002754
Epoch[113] Batch [1199]	Speed: 4907.600991 samples/sec	accuracy=0.820542	lr=0.002718
Epoch[113] Batch [1249]	Speed: 4915.497093 samples/sec	accuracy=0.820554	lr=0.002682
[Epoch 113] training: accuracy=0.820554
[Epoch 113] speed: 4967 samples/sec	time cost: 267.212430
[Epoch 113] validation: err-top1=0.217176 err-top5=0.058315
Epoch[114] Batch [49]	Speed: 4872.895532 samples/sec	accuracy=0.821422	lr=0.002646
Epoch[114] Batch [99]	Speed: 5016.927202 samples/sec	accuracy=0.821295	lr=0.002611
Epoch[114] Batch [149]	Speed: 4858.601401 samples/sec	accuracy=0.821187	lr=0.002575
Epoch[114] Batch [199]	Speed: 5090.965635 samples/sec	accuracy=0.821225	lr=0.002541
Epoch[114] Batch [249]	Speed: 4936.379521 samples/sec	accuracy=0.821240	lr=0.002506
Epoch[114] Batch [299]	Speed: 5009.277434 samples/sec	accuracy=0.821150	lr=0.002472
Epoch[114] Batch [349]	Speed: 5256.474969 samples/sec	accuracy=0.821179	lr=0.002438
Epoch[114] Batch [399]	Speed: 4944.346719 samples/sec	accuracy=0.821214	lr=0.002404
Epoch[114] Batch [449]	Speed: 4925.683312 samples/sec	accuracy=0.821255	lr=0.002370
Epoch[114] Batch [499]	Speed: 5197.562232 samples/sec	accuracy=0.821246	lr=0.002337
Epoch[114] Batch [549]	Speed: 5210.833122 samples/sec	accuracy=0.821195	lr=0.002304
Epoch[114] Batch [599]	Speed: 5093.566500 samples/sec	accuracy=0.821166	lr=0.002271
Epoch[114] Batch [649]	Speed: 5131.821175 samples/sec	accuracy=0.821187	lr=0.002238
Epoch[114] Batch [699]	Speed: 5141.319608 samples/sec	accuracy=0.821169	lr=0.002205
Epoch[114] Batch [749]	Speed: 5040.243744 samples/sec	accuracy=0.821156	lr=0.002173
Epoch[114] Batch [799]	Speed: 5109.627501 samples/sec	accuracy=0.821162	lr=0.002141
Epoch[114] Batch [849]	Speed: 4949.685226 samples/sec	accuracy=0.821186	lr=0.002110
Epoch[114] Batch [899]	Speed: 4988.426294 samples/sec	accuracy=0.821205	lr=0.002078
Epoch[114] Batch [949]	Speed: 5077.409145 samples/sec	accuracy=0.821213	lr=0.002047
Epoch[114] Batch [999]	Speed: 5157.454002 samples/sec	accuracy=0.821206	lr=0.002016
Epoch[114] Batch [1049]	Speed: 5131.082163 samples/sec	accuracy=0.821213	lr=0.001985
Epoch[114] Batch [1099]	Speed: 5038.204303 samples/sec	accuracy=0.821218	lr=0.001954
Epoch[114] Batch [1149]	Speed: 5079.301084 samples/sec	accuracy=0.821242	lr=0.001924
Epoch[114] Batch [1199]	Speed: 5082.271293 samples/sec	accuracy=0.821263	lr=0.001894
Epoch[114] Batch [1249]	Speed: 5140.063668 samples/sec	accuracy=0.821273	lr=0.001864
[Epoch 114] training: accuracy=0.821273
[Epoch 114] speed: 5053 samples/sec	time cost: 262.965863
[Epoch 114] validation: err-top1=0.217096 err-top5=0.058693
Epoch[115] Batch [49]	Speed: 4675.607370 samples/sec	accuracy=0.821660	lr=0.001834
Epoch[115] Batch [99]	Speed: 5069.735176 samples/sec	accuracy=0.821745	lr=0.001804
Epoch[115] Batch [149]	Speed: 5044.684376 samples/sec	accuracy=0.821686	lr=0.001775
Epoch[115] Batch [199]	Speed: 5012.473856 samples/sec	accuracy=0.821527	lr=0.001746
Epoch[115] Batch [249]	Speed: 5026.149277 samples/sec	accuracy=0.821438	lr=0.001718
Epoch[115] Batch [299]	Speed: 5075.295262 samples/sec	accuracy=0.821441	lr=0.001689
Epoch[115] Batch [349]	Speed: 5013.741370 samples/sec	accuracy=0.821417	lr=0.001661
Epoch[115] Batch [399]	Speed: 5091.989532 samples/sec	accuracy=0.821386	lr=0.001633
Epoch[115] Batch [449]	Speed: 5192.214834 samples/sec	accuracy=0.821414	lr=0.001605
Epoch[115] Batch [499]	Speed: 5070.580654 samples/sec	accuracy=0.821459	lr=0.001578
Epoch[115] Batch [549]	Speed: 5050.771988 samples/sec	accuracy=0.821491	lr=0.001551
Epoch[115] Batch [599]	Speed: 5088.249393 samples/sec	accuracy=0.821527	lr=0.001524
Epoch[115] Batch [649]	Speed: 5095.763467 samples/sec	accuracy=0.821568	lr=0.001497
Epoch[115] Batch [699]	Speed: 5065.657192 samples/sec	accuracy=0.821609	lr=0.001470
Epoch[115] Batch [749]	Speed: 5071.573131 samples/sec	accuracy=0.821627	lr=0.001444
Epoch[115] Batch [799]	Speed: 5178.889735 samples/sec	accuracy=0.821627	lr=0.001418
Epoch[115] Batch [849]	Speed: 5162.705925 samples/sec	accuracy=0.821660	lr=0.001392
Epoch[115] Batch [899]	Speed: 5114.858074 samples/sec	accuracy=0.821660	lr=0.001366
Epoch[115] Batch [949]	Speed: 5200.744195 samples/sec	accuracy=0.821658	lr=0.001341
Epoch[115] Batch [999]	Speed: 5094.320607 samples/sec	accuracy=0.821677	lr=0.001316
Epoch[115] Batch [1049]	Speed: 5057.409984 samples/sec	accuracy=0.821686	lr=0.001291
Epoch[115] Batch [1099]	Speed: 5072.936383 samples/sec	accuracy=0.821698	lr=0.001266
Epoch[115] Batch [1149]	Speed: 5014.929185 samples/sec	accuracy=0.821697	lr=0.001242
Epoch[115] Batch [1199]	Speed: 5076.575549 samples/sec	accuracy=0.821680	lr=0.001218
Epoch[115] Batch [1249]	Speed: 5048.398579 samples/sec	accuracy=0.821672	lr=0.001194
[Epoch 115] training: accuracy=0.821672
[Epoch 115] speed: 5060 samples/sec	time cost: 262.361233
[Epoch 115] validation: err-top1=0.216199 err-top5=0.058235
Epoch[116] Batch [49]	Speed: 4915.431048 samples/sec	accuracy=0.821665	lr=0.001170
Epoch[116] Batch [99]	Speed: 5015.904446 samples/sec	accuracy=0.821707	lr=0.001146
Epoch[116] Batch [149]	Speed: 5053.891599 samples/sec	accuracy=0.821706	lr=0.001123
Epoch[116] Batch [199]	Speed: 5093.449918 samples/sec	accuracy=0.821706	lr=0.001100
Epoch[116] Batch [249]	Speed: 5058.392307 samples/sec	accuracy=0.821767	lr=0.001077
Epoch[116] Batch [299]	Speed: 5055.609177 samples/sec	accuracy=0.821714	lr=0.001055
Epoch[116] Batch [349]	Speed: 4956.540327 samples/sec	accuracy=0.821669	lr=0.001032
Epoch[116] Batch [399]	Speed: 4877.651550 samples/sec	accuracy=0.821654	lr=0.001010
Epoch[116] Batch [449]	Speed: 5031.991403 samples/sec	accuracy=0.821679	lr=0.000989
Epoch[116] Batch [499]	Speed: 4931.587732 samples/sec	accuracy=0.821696	lr=0.000967
Epoch[116] Batch [549]	Speed: 5026.106693 samples/sec	accuracy=0.821735	lr=0.000946
Epoch[116] Batch [599]	Speed: 4896.940001 samples/sec	accuracy=0.821745	lr=0.000925
Epoch[116] Batch [649]	Speed: 5076.690760 samples/sec	accuracy=0.821736	lr=0.000904
Epoch[116] Batch [699]	Speed: 4930.589962 samples/sec	accuracy=0.821746	lr=0.000883
Epoch[116] Batch [749]	Speed: 4960.419817 samples/sec	accuracy=0.821762	lr=0.000863
Epoch[116] Batch [799]	Speed: 4906.512233 samples/sec	accuracy=0.821766	lr=0.000843
Epoch[116] Batch [849]	Speed: 5020.329387 samples/sec	accuracy=0.821780	lr=0.000823
Epoch[116] Batch [899]	Speed: 4988.287245 samples/sec	accuracy=0.821794	lr=0.000803
Epoch[116] Batch [949]	Speed: 4882.325950 samples/sec	accuracy=0.821800	lr=0.000784
Epoch[116] Batch [999]	Speed: 5083.443431 samples/sec	accuracy=0.821807	lr=0.000764
Epoch[116] Batch [1049]	Speed: 5061.109439 samples/sec	accuracy=0.821830	lr=0.000745
Epoch[116] Batch [1099]	Speed: 4999.006123 samples/sec	accuracy=0.821832	lr=0.000727
Epoch[116] Batch [1149]	Speed: 4928.857506 samples/sec	accuracy=0.821844	lr=0.000708
Epoch[116] Batch [1199]	Speed: 4934.870356 samples/sec	accuracy=0.821854	lr=0.000690
Epoch[116] Batch [1249]	Speed: 5201.695802 samples/sec	accuracy=0.821861	lr=0.000672
[Epoch 116] training: accuracy=0.821860
[Epoch 116] speed: 4989 samples/sec	time cost: 266.134380
[Epoch 116] validation: err-top1=0.214783 err-top5=0.057556
Epoch[117] Batch [49]	Speed: 4821.954022 samples/sec	accuracy=0.822065	lr=0.000654
Epoch[117] Batch [99]	Speed: 4718.195573 samples/sec	accuracy=0.822289	lr=0.000636
Epoch[117] Batch [149]	Speed: 5105.219914 samples/sec	accuracy=0.822286	lr=0.000619
Epoch[117] Batch [199]	Speed: 4717.143632 samples/sec	accuracy=0.822211	lr=0.000602
Epoch[117] Batch [249]	Speed: 5080.585797 samples/sec	accuracy=0.822196	lr=0.000585
Epoch[117] Batch [299]	Speed: 4959.995565 samples/sec	accuracy=0.822173	lr=0.000569
Epoch[117] Batch [349]	Speed: 4904.394967 samples/sec	accuracy=0.822201	lr=0.000552
Epoch[117] Batch [399]	Speed: 4947.356496 samples/sec	accuracy=0.822198	lr=0.000536
Epoch[117] Batch [449]	Speed: 4995.197010 samples/sec	accuracy=0.822172	lr=0.000520
Epoch[117] Batch [499]	Speed: 4777.622713 samples/sec	accuracy=0.822169	lr=0.000505
Epoch[117] Batch [549]	Speed: 5021.253445 samples/sec	accuracy=0.822158	lr=0.000489
Epoch[117] Batch [599]	Speed: 4939.044251 samples/sec	accuracy=0.822164	lr=0.000474
Epoch[117] Batch [649]	Speed: 5038.568153 samples/sec	accuracy=0.822156	lr=0.000459
Epoch[117] Batch [699]	Speed: 4956.407397 samples/sec	accuracy=0.822128	lr=0.000445
Epoch[117] Batch [749]	Speed: 4987.759047 samples/sec	accuracy=0.822109	lr=0.000430
Epoch[117] Batch [799]	Speed: 4902.779814 samples/sec	accuracy=0.822104	lr=0.000416
Epoch[117] Batch [849]	Speed: 5049.127973 samples/sec	accuracy=0.822105	lr=0.000402
Epoch[117] Batch [899]	Speed: 5219.973416 samples/sec	accuracy=0.822113	lr=0.000388
Epoch[117] Batch [949]	Speed: 4991.872653 samples/sec	accuracy=0.822106	lr=0.000375
Epoch[117] Batch [999]	Speed: 5078.525115 samples/sec	accuracy=0.822096	lr=0.000362
Epoch[117] Batch [1049]	Speed: 4947.452125 samples/sec	accuracy=0.822112	lr=0.000349
Epoch[117] Batch [1099]	Speed: 4882.992708 samples/sec	accuracy=0.822120	lr=0.000336
Epoch[117] Batch [1149]	Speed: 5115.902086 samples/sec	accuracy=0.822109	lr=0.000323
Epoch[117] Batch [1199]	Speed: 5034.040684 samples/sec	accuracy=0.822100	lr=0.000311
Epoch[117] Batch [1249]	Speed: 4991.506583 samples/sec	accuracy=0.822101	lr=0.000299
[Epoch 117] training: accuracy=0.822101
[Epoch 117] speed: 4960 samples/sec	time cost: 268.026159
[Epoch 117] validation: err-top1=0.215103 err-top5=0.057976
Epoch[118] Batch [49]	Speed: 4661.250505 samples/sec	accuracy=0.821982	lr=0.000287
Epoch[118] Batch [99]	Speed: 4998.348608 samples/sec	accuracy=0.821836	lr=0.000275
Epoch[118] Batch [149]	Speed: 4764.510364 samples/sec	accuracy=0.821959	lr=0.000264
Epoch[118] Batch [199]	Speed: 4870.811262 samples/sec	accuracy=0.822010	lr=0.000253
Epoch[118] Batch [249]	Speed: 5137.296762 samples/sec	accuracy=0.822038	lr=0.000242
Epoch[118] Batch [299]	Speed: 5013.600321 samples/sec	accuracy=0.822064	lr=0.000231
Epoch[118] Batch [349]	Speed: 4903.715746 samples/sec	accuracy=0.822058	lr=0.000221
Epoch[118] Batch [399]	Speed: 5172.983720 samples/sec	accuracy=0.822083	lr=0.000211
Epoch[118] Batch [449]	Speed: 4929.245107 samples/sec	accuracy=0.822061	lr=0.000201
Epoch[118] Batch [499]	Speed: 5245.314282 samples/sec	accuracy=0.822075	lr=0.000191
Epoch[118] Batch [549]	Speed: 5104.112927 samples/sec	accuracy=0.822078	lr=0.000182
Epoch[118] Batch [599]	Speed: 5209.414346 samples/sec	accuracy=0.822080	lr=0.000173
Epoch[118] Batch [649]	Speed: 5164.086827 samples/sec	accuracy=0.822075	lr=0.000164
Epoch[118] Batch [699]	Speed: 5177.678291 samples/sec	accuracy=0.822083	lr=0.000155
Epoch[118] Batch [749]	Speed: 5030.294310 samples/sec	accuracy=0.822073	lr=0.000147
Epoch[118] Batch [799]	Speed: 5045.969539 samples/sec	accuracy=0.822081	lr=0.000138
Epoch[118] Batch [849]	Speed: 4984.191806 samples/sec	accuracy=0.822079	lr=0.000130
Epoch[118] Batch [899]	Speed: 5210.877756 samples/sec	accuracy=0.822077	lr=0.000123
Epoch[118] Batch [949]	Speed: 5119.230584 samples/sec	accuracy=0.822065	lr=0.000115
Epoch[118] Batch [999]	Speed: 4777.043607 samples/sec	accuracy=0.822072	lr=0.000108
Epoch[118] Batch [1049]	Speed: 4882.445500 samples/sec	accuracy=0.822071	lr=0.000101
Epoch[118] Batch [1099]	Speed: 4968.410842 samples/sec	accuracy=0.822070	lr=0.000094
Epoch[118] Batch [1149]	Speed: 4875.126344 samples/sec	accuracy=0.822072	lr=0.000087
Epoch[118] Batch [1199]	Speed: 5001.948033 samples/sec	accuracy=0.822081	lr=0.000081
Epoch[118] Batch [1249]	Speed: 4974.243295 samples/sec	accuracy=0.822090	lr=0.000075
[Epoch 118] training: accuracy=0.822091
[Epoch 118] speed: 5000 samples/sec	time cost: 265.325530
[Epoch 118] validation: err-top1=0.214804 err-top5=0.057677
Epoch[119] Batch [49]	Speed: 4763.257740 samples/sec	accuracy=0.822178	lr=0.000069
Epoch[119] Batch [99]	Speed: 5001.378267 samples/sec	accuracy=0.822149	lr=0.000063
Epoch[119] Batch [149]	Speed: 4747.551683 samples/sec	accuracy=0.822097	lr=0.000058
Epoch[119] Batch [199]	Speed: 4922.260916 samples/sec	accuracy=0.822080	lr=0.000053
Epoch[119] Batch [249]	Speed: 4992.180751 samples/sec	accuracy=0.822159	lr=0.000048
Epoch[119] Batch [299]	Speed: 5075.783377 samples/sec	accuracy=0.822211	lr=0.000043
Epoch[119] Batch [349]	Speed: 5265.079328 samples/sec	accuracy=0.822213	lr=0.000039
Epoch[119] Batch [399]	Speed: 4987.564665 samples/sec	accuracy=0.822178	lr=0.000035
Epoch[119] Batch [449]	Speed: 5023.636649 samples/sec	accuracy=0.822155	lr=0.000031
Epoch[119] Batch [499]	Speed: 5090.805726 samples/sec	accuracy=0.822161	lr=0.000027
Epoch[119] Batch [549]	Speed: 5333.285571 samples/sec	accuracy=0.822199	lr=0.000023
Epoch[119] Batch [599]	Speed: 5157.607225 samples/sec	accuracy=0.822168	lr=0.000020
Epoch[119] Batch [649]	Speed: 5141.288466 samples/sec	accuracy=0.822194	lr=0.000017
Epoch[119] Batch [699]	Speed: 5237.014825 samples/sec	accuracy=0.822208	lr=0.000015
Epoch[119] Batch [749]	Speed: 5138.928731 samples/sec	accuracy=0.822208	lr=0.000012
Epoch[119] Batch [799]	Speed: 5033.031700 samples/sec	accuracy=0.822198	lr=0.000010
Epoch[119] Batch [849]	Speed: 4981.661311 samples/sec	accuracy=0.822174	lr=0.000008
Epoch[119] Batch [899]	Speed: 5120.535578 samples/sec	accuracy=0.822170	lr=0.000006
Epoch[119] Batch [949]	Speed: 5079.974903 samples/sec	accuracy=0.822174	lr=0.000004
Epoch[119] Batch [999]	Speed: 4960.230883 samples/sec	accuracy=0.822165	lr=0.000003
Epoch[119] Batch [1049]	Speed: 5118.760310 samples/sec	accuracy=0.822152	lr=0.000002
Epoch[119] Batch [1099]	Speed: 5033.837840 samples/sec	accuracy=0.822135	lr=0.000001
Epoch[119] Batch [1149]	Speed: 4889.997553 samples/sec	accuracy=0.822142	lr=0.000000
Epoch[119] Batch [1199]	Speed: 5083.515151 samples/sec	accuracy=0.822151	lr=0.000000
Epoch[119] Batch [1249]	Speed: 5314.441629 samples/sec	accuracy=0.822141	lr=0.000000
[Epoch 119] training: accuracy=0.822140
[Epoch 119] speed: 5051 samples/sec	time cost: 262.886739
[Epoch 119] validation: err-top1=0.215302 err-top5=0.057418