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_v1_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_v1', 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_v1_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: 1899.565520 samples/sec	accuracy=0.883496	lr=0.003133
Epoch[0] Batch [99]	Speed: 4810.490215 samples/sec	accuracy=0.920217	lr=0.006331
Epoch[0] Batch [149]	Speed: 4885.672011 samples/sec	accuracy=0.941568	lr=0.009528
Epoch[0] Batch [199]	Speed: 4977.229144 samples/sec	accuracy=0.951664	lr=0.012726
Epoch[0] Batch [249]	Speed: 4676.824599 samples/sec	accuracy=0.954637	lr=0.015923
Epoch[0] Batch [299]	Speed: 4857.797000 samples/sec	accuracy=0.952445	lr=0.019121
Epoch[0] Batch [349]	Speed: 5005.629886 samples/sec	accuracy=0.946694	lr=0.022318
Epoch[0] Batch [399]	Speed: 4668.898032 samples/sec	accuracy=0.937761	lr=0.025516
Epoch[0] Batch [449]	Speed: 4456.815666 samples/sec	accuracy=0.927260	lr=0.028713
Epoch[0] Batch [499]	Speed: 4593.681250 samples/sec	accuracy=0.915466	lr=0.031910
Epoch[0] Batch [549]	Speed: 4881.537865 samples/sec	accuracy=0.903045	lr=0.035108
Epoch[0] Batch [599]	Speed: 4749.025205 samples/sec	accuracy=0.890653	lr=0.038305
Epoch[0] Batch [649]	Speed: 4471.698106 samples/sec	accuracy=0.878325	lr=0.041503
Epoch[0] Batch [699]	Speed: 4546.444220 samples/sec	accuracy=0.866359	lr=0.044700
Epoch[0] Batch [749]	Speed: 4381.711781 samples/sec	accuracy=0.854754	lr=0.047898
Epoch[0] Batch [799]	Speed: 4707.918714 samples/sec	accuracy=0.843255	lr=0.051095
Epoch[0] Batch [849]	Speed: 4939.831470 samples/sec	accuracy=0.832109	lr=0.054293
Epoch[0] Batch [899]	Speed: 4817.178222 samples/sec	accuracy=0.821110	lr=0.057490
Epoch[0] Batch [949]	Speed: 4721.146618 samples/sec	accuracy=0.810316	lr=0.060687
Epoch[0] Batch [999]	Speed: 4863.024590 samples/sec	accuracy=0.799681	lr=0.063885
Epoch[0] Batch [1049]	Speed: 4636.798679 samples/sec	accuracy=0.789430	lr=0.067082
Epoch[0] Batch [1099]	Speed: 4939.497419 samples/sec	accuracy=0.779446	lr=0.070280
Epoch[0] Batch [1149]	Speed: 4821.105860 samples/sec	accuracy=0.769786	lr=0.073477
Epoch[0] Batch [1199]	Speed: 4448.397785 samples/sec	accuracy=0.760474	lr=0.076675
Epoch[0] Batch [1249]	Speed: 4834.809670 samples/sec	accuracy=0.751415	lr=0.079872
[Epoch 0] training: accuracy=0.751063
[Epoch 0] speed: 4457 samples/sec	time cost: 299.274206
[Epoch 0] validation: err-top1=0.947405 err-top5=0.853894
Epoch[1] Batch [49]	Speed: 4352.360657 samples/sec	accuracy=0.521088	lr=0.083133
Epoch[1] Batch [99]	Speed: 4765.489101 samples/sec	accuracy=0.517567	lr=0.086331
Epoch[1] Batch [149]	Speed: 4756.077262 samples/sec	accuracy=0.513748	lr=0.089528
Epoch[1] Batch [199]	Speed: 4640.218584 samples/sec	accuracy=0.510818	lr=0.092726
Epoch[1] Batch [249]	Speed: 4817.225768 samples/sec	accuracy=0.507785	lr=0.095923
Epoch[1] Batch [299]	Speed: 4524.594160 samples/sec	accuracy=0.504430	lr=0.099121
Epoch[1] Batch [349]	Speed: 4548.681949 samples/sec	accuracy=0.501443	lr=0.102318
Epoch[1] Batch [399]	Speed: 4807.577238 samples/sec	accuracy=0.498711	lr=0.105516
Epoch[1] Batch [449]	Speed: 5021.316611 samples/sec	accuracy=0.495866	lr=0.108713
Epoch[1] Batch [499]	Speed: 4858.392664 samples/sec	accuracy=0.493423	lr=0.111910
Epoch[1] Batch [549]	Speed: 4783.940132 samples/sec	accuracy=0.491049	lr=0.115108
Epoch[1] Batch [599]	Speed: 4907.233159 samples/sec	accuracy=0.488835	lr=0.118305
Epoch[1] Batch [649]	Speed: 4845.724710 samples/sec	accuracy=0.486889	lr=0.121503
Epoch[1] Batch [699]	Speed: 4993.445569 samples/sec	accuracy=0.485039	lr=0.124700
Epoch[1] Batch [749]	Speed: 4729.047721 samples/sec	accuracy=0.483292	lr=0.127898
Epoch[1] Batch [799]	Speed: 5188.448664 samples/sec	accuracy=0.481683	lr=0.131095
Epoch[1] Batch [849]	Speed: 4759.552868 samples/sec	accuracy=0.480163	lr=0.134293
Epoch[1] Batch [899]	Speed: 4962.101956 samples/sec	accuracy=0.478750	lr=0.137490
Epoch[1] Batch [949]	Speed: 4846.736665 samples/sec	accuracy=0.477317	lr=0.140687
Epoch[1] Batch [999]	Speed: 4870.602248 samples/sec	accuracy=0.476052	lr=0.143885
Epoch[1] Batch [1049]	Speed: 4844.217679 samples/sec	accuracy=0.474808	lr=0.147082
Epoch[1] Batch [1099]	Speed: 4884.162361 samples/sec	accuracy=0.473713	lr=0.150280
Epoch[1] Batch [1149]	Speed: 5024.125691 samples/sec	accuracy=0.472705	lr=0.153477
Epoch[1] Batch [1199]	Speed: 4878.888255 samples/sec	accuracy=0.471735	lr=0.156675
Epoch[1] Batch [1249]	Speed: 4982.761250 samples/sec	accuracy=0.470856	lr=0.159872
[Epoch 1] training: accuracy=0.470831
[Epoch 1] speed: 4813 samples/sec	time cost: 276.793966
[Epoch 1] validation: err-top1=0.876794 err-top5=0.711974
Epoch[2] Batch [49]	Speed: 4538.503676 samples/sec	accuracy=0.448103	lr=0.163133
Epoch[2] Batch [99]	Speed: 4673.894913 samples/sec	accuracy=0.448188	lr=0.166331
Epoch[2] Batch [149]	Speed: 4788.299683 samples/sec	accuracy=0.448293	lr=0.169528
Epoch[2] Batch [199]	Speed: 4827.387381 samples/sec	accuracy=0.448117	lr=0.172726
Epoch[2] Batch [249]	Speed: 4728.051208 samples/sec	accuracy=0.448151	lr=0.175923
Epoch[2] Batch [299]	Speed: 4719.918866 samples/sec	accuracy=0.448483	lr=0.179121
Epoch[2] Batch [349]	Speed: 4735.677077 samples/sec	accuracy=0.448569	lr=0.182318
Epoch[2] Batch [399]	Speed: 4853.938176 samples/sec	accuracy=0.448806	lr=0.185516
Epoch[2] Batch [449]	Speed: 4532.073125 samples/sec	accuracy=0.449113	lr=0.188713
Epoch[2] Batch [499]	Speed: 5034.727808 samples/sec	accuracy=0.449466	lr=0.191910
Epoch[2] Batch [549]	Speed: 4755.196940 samples/sec	accuracy=0.449954	lr=0.195108
Epoch[2] Batch [599]	Speed: 4443.787886 samples/sec	accuracy=0.450303	lr=0.198305
Epoch[2] Batch [649]	Speed: 4564.106834 samples/sec	accuracy=0.450601	lr=0.201503
Epoch[2] Batch [699]	Speed: 4629.456749 samples/sec	accuracy=0.451009	lr=0.204700
Epoch[2] Batch [749]	Speed: 4989.742308 samples/sec	accuracy=0.451475	lr=0.207898
Epoch[2] Batch [799]	Speed: 5048.186507 samples/sec	accuracy=0.451994	lr=0.211095
Epoch[2] Batch [849]	Speed: 4727.773079 samples/sec	accuracy=0.452469	lr=0.214293
Epoch[2] Batch [899]	Speed: 4804.648761 samples/sec	accuracy=0.452992	lr=0.217490
Epoch[2] Batch [949]	Speed: 4745.333099 samples/sec	accuracy=0.453483	lr=0.220687
Epoch[2] Batch [999]	Speed: 4821.589931 samples/sec	accuracy=0.453983	lr=0.223885
Epoch[2] Batch [1049]	Speed: 5011.722264 samples/sec	accuracy=0.454484	lr=0.227082
Epoch[2] Batch [1099]	Speed: 4969.585093 samples/sec	accuracy=0.455003	lr=0.230280
Epoch[2] Batch [1149]	Speed: 4492.091783 samples/sec	accuracy=0.455507	lr=0.233477
Epoch[2] Batch [1199]	Speed: 4626.693543 samples/sec	accuracy=0.456083	lr=0.236675
Epoch[2] Batch [1249]	Speed: 4474.152169 samples/sec	accuracy=0.456688	lr=0.239872
[Epoch 2] training: accuracy=0.456699
[Epoch 2] speed: 4728 samples/sec	time cost: 281.721205
[Epoch 2] validation: err-top1=0.804787 err-top5=0.590521
Epoch[3] Batch [49]	Speed: 4638.048773 samples/sec	accuracy=0.471394	lr=0.243133
Epoch[3] Batch [99]	Speed: 4644.698308 samples/sec	accuracy=0.473050	lr=0.246331
Epoch[3] Batch [149]	Speed: 4900.252570 samples/sec	accuracy=0.473077	lr=0.249528
Epoch[3] Batch [199]	Speed: 4988.019946 samples/sec	accuracy=0.474002	lr=0.252726
Epoch[3] Batch [249]	Speed: 4447.462329 samples/sec	accuracy=0.474507	lr=0.255923
Epoch[3] Batch [299]	Speed: 4343.333002 samples/sec	accuracy=0.475321	lr=0.259121
Epoch[3] Batch [349]	Speed: 4776.040999 samples/sec	accuracy=0.476285	lr=0.262318
Epoch[3] Batch [399]	Speed: 4672.831924 samples/sec	accuracy=0.476998	lr=0.265516
Epoch[3] Batch [449]	Speed: 4888.543096 samples/sec	accuracy=0.477839	lr=0.268713
Epoch[3] Batch [499]	Speed: 4417.423397 samples/sec	accuracy=0.478573	lr=0.271910
Epoch[3] Batch [549]	Speed: 5201.936845 samples/sec	accuracy=0.479478	lr=0.275108
Epoch[3] Batch [599]	Speed: 4843.719549 samples/sec	accuracy=0.480284	lr=0.278305
Epoch[3] Batch [649]	Speed: 4470.092745 samples/sec	accuracy=0.480899	lr=0.281503
Epoch[3] Batch [699]	Speed: 4670.827983 samples/sec	accuracy=0.481745	lr=0.284700
Epoch[3] Batch [749]	Speed: 4578.559713 samples/sec	accuracy=0.482597	lr=0.287898
Epoch[3] Batch [799]	Speed: 4600.070170 samples/sec	accuracy=0.483400	lr=0.291095
Epoch[3] Batch [849]	Speed: 4607.636832 samples/sec	accuracy=0.484225	lr=0.294293
Epoch[3] Batch [899]	Speed: 5006.044359 samples/sec	accuracy=0.484943	lr=0.297490
Epoch[3] Batch [949]	Speed: 4654.668133 samples/sec	accuracy=0.485821	lr=0.300687
Epoch[3] Batch [999]	Speed: 4540.827495 samples/sec	accuracy=0.486627	lr=0.303885
Epoch[3] Batch [1049]	Speed: 4491.684669 samples/sec	accuracy=0.487569	lr=0.307082
Epoch[3] Batch [1099]	Speed: 4516.712686 samples/sec	accuracy=0.488398	lr=0.310280
Epoch[3] Batch [1149]	Speed: 4535.489282 samples/sec	accuracy=0.489207	lr=0.313477
Epoch[3] Batch [1199]	Speed: 4514.675135 samples/sec	accuracy=0.490015	lr=0.316675
Epoch[3] Batch [1249]	Speed: 4980.289494 samples/sec	accuracy=0.490838	lr=0.319872
[Epoch 3] training: accuracy=0.490853
[Epoch 3] speed: 4661 samples/sec	time cost: 286.210159
[Epoch 3] validation: err-top1=0.663066 err-top5=0.401327
Epoch[4] Batch [49]	Speed: 4316.779812 samples/sec	accuracy=0.511344	lr=0.323133
Epoch[4] Batch [99]	Speed: 5021.271761 samples/sec	accuracy=0.512215	lr=0.326331
Epoch[4] Batch [149]	Speed: 4820.833992 samples/sec	accuracy=0.513152	lr=0.329528
Epoch[4] Batch [199]	Speed: 5096.617287 samples/sec	accuracy=0.514002	lr=0.332726
Epoch[4] Batch [249]	Speed: 4859.595647 samples/sec	accuracy=0.514839	lr=0.335923
Epoch[4] Batch [299]	Speed: 4364.306791 samples/sec	accuracy=0.515916	lr=0.339121
Epoch[4] Batch [349]	Speed: 4766.769986 samples/sec	accuracy=0.516737	lr=0.342318
Epoch[4] Batch [399]	Speed: 4821.505384 samples/sec	accuracy=0.517582	lr=0.345516
Epoch[4] Batch [449]	Speed: 4402.405989 samples/sec	accuracy=0.518271	lr=0.348713
Epoch[4] Batch [499]	Speed: 4708.839231 samples/sec	accuracy=0.518938	lr=0.351910
Epoch[4] Batch [549]	Speed: 5002.305266 samples/sec	accuracy=0.519764	lr=0.355108
Epoch[4] Batch [599]	Speed: 4593.966428 samples/sec	accuracy=0.520493	lr=0.358305
Epoch[4] Batch [649]	Speed: 4833.153321 samples/sec	accuracy=0.521411	lr=0.361503
Epoch[4] Batch [699]	Speed: 4903.294980 samples/sec	accuracy=0.522276	lr=0.364700
Epoch[4] Batch [749]	Speed: 4454.105908 samples/sec	accuracy=0.523220	lr=0.367898
Epoch[4] Batch [799]	Speed: 4905.323448 samples/sec	accuracy=0.524012	lr=0.371095
Epoch[4] Batch [849]	Speed: 4628.535979 samples/sec	accuracy=0.524892	lr=0.374293
Epoch[4] Batch [899]	Speed: 4788.146052 samples/sec	accuracy=0.525709	lr=0.377490
Epoch[4] Batch [949]	Speed: 4686.981048 samples/sec	accuracy=0.526447	lr=0.380687
Epoch[4] Batch [999]	Speed: 4362.409625 samples/sec	accuracy=0.527264	lr=0.383885
Epoch[4] Batch [1049]	Speed: 4623.918300 samples/sec	accuracy=0.528088	lr=0.387082
Epoch[4] Batch [1099]	Speed: 4952.851146 samples/sec	accuracy=0.528799	lr=0.390280
Epoch[4] Batch [1149]	Speed: 4718.077815 samples/sec	accuracy=0.529587	lr=0.393477
Epoch[4] Batch [1199]	Speed: 4810.533641 samples/sec	accuracy=0.530324	lr=0.396675
Epoch[4] Batch [1249]	Speed: 4699.739813 samples/sec	accuracy=0.531013	lr=0.399872
[Epoch 4] training: accuracy=0.531028
[Epoch 4] speed: 4712 samples/sec	time cost: 282.033370
[Epoch 4] validation: err-top1=0.646086 err-top5=0.382573
Epoch[5] Batch [49]	Speed: 4408.476303 samples/sec	accuracy=0.547623	lr=0.400000
Epoch[5] Batch [99]	Speed: 4958.002680 samples/sec	accuracy=0.549712	lr=0.400000
Epoch[5] Batch [149]	Speed: 4897.665936 samples/sec	accuracy=0.550599	lr=0.399999
Epoch[5] Batch [199]	Speed: 4971.254582 samples/sec	accuracy=0.551189	lr=0.399998
Epoch[5] Batch [249]	Speed: 4734.284152 samples/sec	accuracy=0.552238	lr=0.399997
Epoch[5] Batch [299]	Speed: 4340.085348 samples/sec	accuracy=0.552925	lr=0.399996
Epoch[5] Batch [349]	Speed: 4441.878427 samples/sec	accuracy=0.553669	lr=0.399994
Epoch[5] Batch [399]	Speed: 4456.983551 samples/sec	accuracy=0.554497	lr=0.399992
Epoch[5] Batch [449]	Speed: 4639.518843 samples/sec	accuracy=0.555192	lr=0.399990
Epoch[5] Batch [499]	Speed: 4533.639094 samples/sec	accuracy=0.555970	lr=0.399988
Epoch[5] Batch [549]	Speed: 5012.724125 samples/sec	accuracy=0.556621	lr=0.399986
Epoch[5] Batch [599]	Speed: 4819.650445 samples/sec	accuracy=0.557575	lr=0.399983
Epoch[5] Batch [649]	Speed: 4712.696263 samples/sec	accuracy=0.558412	lr=0.399980
Epoch[5] Batch [699]	Speed: 4981.757692 samples/sec	accuracy=0.559128	lr=0.399977
Epoch[5] Batch [749]	Speed: 4969.358892 samples/sec	accuracy=0.559863	lr=0.399973
Epoch[5] Batch [799]	Speed: 4411.734436 samples/sec	accuracy=0.560567	lr=0.399970
Epoch[5] Batch [849]	Speed: 4544.028112 samples/sec	accuracy=0.561416	lr=0.399966
Epoch[5] Batch [899]	Speed: 4902.859735 samples/sec	accuracy=0.562281	lr=0.399961
Epoch[5] Batch [949]	Speed: 4780.354341 samples/sec	accuracy=0.563030	lr=0.399957
Epoch[5] Batch [999]	Speed: 4965.417076 samples/sec	accuracy=0.563795	lr=0.399952
Epoch[5] Batch [1049]	Speed: 4663.031367 samples/sec	accuracy=0.564543	lr=0.399948
Epoch[5] Batch [1099]	Speed: 5025.165091 samples/sec	accuracy=0.565323	lr=0.399942
Epoch[5] Batch [1149]	Speed: 5100.580043 samples/sec	accuracy=0.566159	lr=0.399937
Epoch[5] Batch [1199]	Speed: 4700.315861 samples/sec	accuracy=0.566867	lr=0.399931
Epoch[5] Batch [1249]	Speed: 4600.054503 samples/sec	accuracy=0.567595	lr=0.399926
[Epoch 5] training: accuracy=0.567609
[Epoch 5] speed: 4724 samples/sec	time cost: 282.112344
[Epoch 5] validation: err-top1=0.562622 err-top5=0.299622
Epoch[6] Batch [49]	Speed: 4527.849635 samples/sec	accuracy=0.584044	lr=0.399919
Epoch[6] Batch [99]	Speed: 4455.320519 samples/sec	accuracy=0.585804	lr=0.399913
Epoch[6] Batch [149]	Speed: 4909.001726 samples/sec	accuracy=0.586889	lr=0.399907
Epoch[6] Batch [199]	Speed: 4539.027635 samples/sec	accuracy=0.587036	lr=0.399900
Epoch[6] Batch [249]	Speed: 4889.769075 samples/sec	accuracy=0.587704	lr=0.399893
Epoch[6] Batch [299]	Speed: 4792.451779 samples/sec	accuracy=0.588181	lr=0.399885
Epoch[6] Batch [349]	Speed: 4719.167195 samples/sec	accuracy=0.588926	lr=0.399878
Epoch[6] Batch [399]	Speed: 4952.021517 samples/sec	accuracy=0.589385	lr=0.399870
Epoch[6] Batch [449]	Speed: 4629.237000 samples/sec	accuracy=0.589948	lr=0.399862
Epoch[6] Batch [499]	Speed: 4809.655021 samples/sec	accuracy=0.590436	lr=0.399854
Epoch[6] Batch [549]	Speed: 4614.874752 samples/sec	accuracy=0.591032	lr=0.399846
Epoch[6] Batch [599]	Speed: 4580.513972 samples/sec	accuracy=0.591649	lr=0.399837
Epoch[6] Batch [649]	Speed: 4633.912710 samples/sec	accuracy=0.592365	lr=0.399828
Epoch[6] Batch [699]	Speed: 4787.199179 samples/sec	accuracy=0.593041	lr=0.399819
Epoch[6] Batch [749]	Speed: 4122.224331 samples/sec	accuracy=0.593681	lr=0.399809
Epoch[6] Batch [799]	Speed: 4606.343001 samples/sec	accuracy=0.594327	lr=0.399800
Epoch[6] Batch [849]	Speed: 4641.661035 samples/sec	accuracy=0.594882	lr=0.399790
Epoch[6] Batch [899]	Speed: 4716.884294 samples/sec	accuracy=0.595410	lr=0.399780
Epoch[6] Batch [949]	Speed: 4631.619725 samples/sec	accuracy=0.596036	lr=0.399769
Epoch[6] Batch [999]	Speed: 4752.529633 samples/sec	accuracy=0.596658	lr=0.399759
Epoch[6] Batch [1049]	Speed: 4815.099334 samples/sec	accuracy=0.597242	lr=0.399748
Epoch[6] Batch [1099]	Speed: 4988.027245 samples/sec	accuracy=0.597755	lr=0.399737
Epoch[6] Batch [1149]	Speed: 4556.731083 samples/sec	accuracy=0.598281	lr=0.399725
Epoch[6] Batch [1199]	Speed: 4354.178021 samples/sec	accuracy=0.598782	lr=0.399714
Epoch[6] Batch [1249]	Speed: 4627.956145 samples/sec	accuracy=0.599282	lr=0.399702
[Epoch 6] training: accuracy=0.599291
[Epoch 6] speed: 4654 samples/sec	time cost: 285.660642
[Epoch 6] validation: err-top1=0.556382 err-top5=0.296516
Epoch[7] Batch [49]	Speed: 4775.037431 samples/sec	accuracy=0.610089	lr=0.399690
Epoch[7] Batch [99]	Speed: 4934.716701 samples/sec	accuracy=0.610966	lr=0.399677
Epoch[7] Batch [149]	Speed: 4822.476816 samples/sec	accuracy=0.611286	lr=0.399665
Epoch[7] Batch [199]	Speed: 5026.823776 samples/sec	accuracy=0.611372	lr=0.399652
Epoch[7] Batch [249]	Speed: 4824.859953 samples/sec	accuracy=0.611228	lr=0.399639
Epoch[7] Batch [299]	Speed: 4815.725609 samples/sec	accuracy=0.611905	lr=0.399626
Epoch[7] Batch [349]	Speed: 5130.139792 samples/sec	accuracy=0.612513	lr=0.399613
Epoch[7] Batch [399]	Speed: 5110.061930 samples/sec	accuracy=0.612964	lr=0.399599
Epoch[7] Batch [449]	Speed: 4553.659123 samples/sec	accuracy=0.613546	lr=0.399585
Epoch[7] Batch [499]	Speed: 4864.633487 samples/sec	accuracy=0.614053	lr=0.399571
Epoch[7] Batch [549]	Speed: 4841.348420 samples/sec	accuracy=0.614597	lr=0.399556
Epoch[7] Batch [599]	Speed: 4506.445216 samples/sec	accuracy=0.615048	lr=0.399542
Epoch[7] Batch [649]	Speed: 5098.005657 samples/sec	accuracy=0.615438	lr=0.399527
Epoch[7] Batch [699]	Speed: 4609.788565 samples/sec	accuracy=0.615961	lr=0.399512
Epoch[7] Batch [749]	Speed: 4933.314061 samples/sec	accuracy=0.616447	lr=0.399496
Epoch[7] Batch [799]	Speed: 4869.786470 samples/sec	accuracy=0.616901	lr=0.399481
Epoch[7] Batch [849]	Speed: 4824.189468 samples/sec	accuracy=0.617477	lr=0.399465
Epoch[7] Batch [899]	Speed: 5040.836126 samples/sec	accuracy=0.617976	lr=0.399449
Epoch[7] Batch [949]	Speed: 4277.147502 samples/sec	accuracy=0.618376	lr=0.399432
Epoch[7] Batch [999]	Speed: 4736.456478 samples/sec	accuracy=0.618791	lr=0.399416
Epoch[7] Batch [1049]	Speed: 4936.681260 samples/sec	accuracy=0.619273	lr=0.399399
Epoch[7] Batch [1099]	Speed: 4819.255337 samples/sec	accuracy=0.619739	lr=0.399382
Epoch[7] Batch [1149]	Speed: 4977.874767 samples/sec	accuracy=0.620243	lr=0.399365
Epoch[7] Batch [1199]	Speed: 4303.200868 samples/sec	accuracy=0.620627	lr=0.399347
Epoch[7] Batch [1249]	Speed: 4253.342723 samples/sec	accuracy=0.621029	lr=0.399329
[Epoch 7] training: accuracy=0.621032
[Epoch 7] speed: 4776 samples/sec	time cost: 279.029361
[Epoch 7] validation: err-top1=0.510005 err-top5=0.250777
Epoch[8] Batch [49]	Speed: 4715.778571 samples/sec	accuracy=0.626968	lr=0.399311
Epoch[8] Batch [99]	Speed: 4660.870319 samples/sec	accuracy=0.627812	lr=0.399293
Epoch[8] Batch [149]	Speed: 4876.942390 samples/sec	accuracy=0.628223	lr=0.399274
Epoch[8] Batch [199]	Speed: 4916.206031 samples/sec	accuracy=0.629100	lr=0.399256
Epoch[8] Batch [249]	Speed: 4314.100938 samples/sec	accuracy=0.629256	lr=0.399237
Epoch[8] Batch [299]	Speed: 4718.498287 samples/sec	accuracy=0.629542	lr=0.399218
Epoch[8] Batch [349]	Speed: 4874.669528 samples/sec	accuracy=0.630133	lr=0.399198
Epoch[8] Batch [399]	Speed: 4666.447908 samples/sec	accuracy=0.630526	lr=0.399179
Epoch[8] Batch [449]	Speed: 4662.627909 samples/sec	accuracy=0.631214	lr=0.399159
Epoch[8] Batch [499]	Speed: 4767.519119 samples/sec	accuracy=0.631652	lr=0.399138
Epoch[8] Batch [549]	Speed: 4697.230507 samples/sec	accuracy=0.632097	lr=0.399118
Epoch[8] Batch [599]	Speed: 4901.633225 samples/sec	accuracy=0.632506	lr=0.399098
Epoch[8] Batch [649]	Speed: 5194.061151 samples/sec	accuracy=0.632866	lr=0.399077
Epoch[8] Batch [699]	Speed: 4977.811882 samples/sec	accuracy=0.633368	lr=0.399056
Epoch[8] Batch [749]	Speed: 4748.236413 samples/sec	accuracy=0.633858	lr=0.399034
Epoch[8] Batch [799]	Speed: 4717.027999 samples/sec	accuracy=0.634396	lr=0.399013
Epoch[8] Batch [849]	Speed: 5067.161694 samples/sec	accuracy=0.634695	lr=0.398991
Epoch[8] Batch [899]	Speed: 4940.912672 samples/sec	accuracy=0.635116	lr=0.398969
Epoch[8] Batch [949]	Speed: 5043.725850 samples/sec	accuracy=0.635465	lr=0.398947
Epoch[8] Batch [999]	Speed: 5029.982904 samples/sec	accuracy=0.635757	lr=0.398924
Epoch[8] Batch [1049]	Speed: 4883.678418 samples/sec	accuracy=0.636078	lr=0.398901
Epoch[8] Batch [1099]	Speed: 5106.499924 samples/sec	accuracy=0.636424	lr=0.398878
Epoch[8] Batch [1149]	Speed: 5049.484972 samples/sec	accuracy=0.636734	lr=0.398855
Epoch[8] Batch [1199]	Speed: 4973.696756 samples/sec	accuracy=0.637172	lr=0.398832
Epoch[8] Batch [1249]	Speed: 4748.241242 samples/sec	accuracy=0.637512	lr=0.398808
[Epoch 8] training: accuracy=0.637513
[Epoch 8] speed: 4836 samples/sec	time cost: 275.067825
[Epoch 8] validation: err-top1=0.522441 err-top5=0.265446
Epoch[9] Batch [49]	Speed: 4583.104510 samples/sec	accuracy=0.643725	lr=0.398784
Epoch[9] Batch [99]	Speed: 4712.274653 samples/sec	accuracy=0.643673	lr=0.398760
Epoch[9] Batch [149]	Speed: 4969.468827 samples/sec	accuracy=0.643141	lr=0.398735
Epoch[9] Batch [199]	Speed: 4799.729398 samples/sec	accuracy=0.643799	lr=0.398710
Epoch[9] Batch [249]	Speed: 4881.133877 samples/sec	accuracy=0.644370	lr=0.398686
Epoch[9] Batch [299]	Speed: 4649.736419 samples/sec	accuracy=0.644499	lr=0.398660
Epoch[9] Batch [349]	Speed: 4901.299063 samples/sec	accuracy=0.644940	lr=0.398635
Epoch[9] Batch [399]	Speed: 5104.443771 samples/sec	accuracy=0.645179	lr=0.398610
Epoch[9] Batch [449]	Speed: 4653.885864 samples/sec	accuracy=0.645488	lr=0.398584
Epoch[9] Batch [499]	Speed: 5185.792727 samples/sec	accuracy=0.645800	lr=0.398558
Epoch[9] Batch [549]	Speed: 5285.066989 samples/sec	accuracy=0.646251	lr=0.398531
Epoch[9] Batch [599]	Speed: 4577.589503 samples/sec	accuracy=0.646553	lr=0.398505
Epoch[9] Batch [649]	Speed: 4706.347843 samples/sec	accuracy=0.646888	lr=0.398478
Epoch[9] Batch [699]	Speed: 4700.793574 samples/sec	accuracy=0.647280	lr=0.398451
Epoch[9] Batch [749]	Speed: 4587.345819 samples/sec	accuracy=0.647547	lr=0.398424
Epoch[9] Batch [799]	Speed: 4607.029013 samples/sec	accuracy=0.647816	lr=0.398396
Epoch[9] Batch [849]	Speed: 4623.349577 samples/sec	accuracy=0.648067	lr=0.398369
Epoch[9] Batch [899]	Speed: 4967.063660 samples/sec	accuracy=0.648320	lr=0.398341
Epoch[9] Batch [949]	Speed: 4712.616009 samples/sec	accuracy=0.648627	lr=0.398312
Epoch[9] Batch [999]	Speed: 4535.232963 samples/sec	accuracy=0.648895	lr=0.398284
Epoch[9] Batch [1049]	Speed: 4772.342948 samples/sec	accuracy=0.649116	lr=0.398255
Epoch[9] Batch [1099]	Speed: 4127.685097 samples/sec	accuracy=0.649397	lr=0.398226
Epoch[9] Batch [1149]	Speed: 4743.434348 samples/sec	accuracy=0.649702	lr=0.398197
Epoch[9] Batch [1199]	Speed: 4744.149961 samples/sec	accuracy=0.649915	lr=0.398168
Epoch[9] Batch [1249]	Speed: 4969.713900 samples/sec	accuracy=0.650162	lr=0.398138
[Epoch 9] training: accuracy=0.650166
[Epoch 9] speed: 4746 samples/sec	time cost: 279.891733
[Epoch 9] validation: err-top1=0.502631 err-top5=0.246951
Epoch[10] Batch [49]	Speed: 4430.505223 samples/sec	accuracy=0.654137	lr=0.398108
Epoch[10] Batch [99]	Speed: 4569.984484 samples/sec	accuracy=0.653549	lr=0.398078
Epoch[10] Batch [149]	Speed: 4666.176878 samples/sec	accuracy=0.653742	lr=0.398048
Epoch[10] Batch [199]	Speed: 4623.071985 samples/sec	accuracy=0.653646	lr=0.398017
Epoch[10] Batch [249]	Speed: 4636.847036 samples/sec	accuracy=0.654002	lr=0.397986
Epoch[10] Batch [299]	Speed: 4476.813584 samples/sec	accuracy=0.654470	lr=0.397955
Epoch[10] Batch [349]	Speed: 4741.173238 samples/sec	accuracy=0.654997	lr=0.397924
Epoch[10] Batch [399]	Speed: 4884.555629 samples/sec	accuracy=0.655410	lr=0.397892
Epoch[10] Batch [449]	Speed: 4402.368535 samples/sec	accuracy=0.655913	lr=0.397861
Epoch[10] Batch [499]	Speed: 4355.057597 samples/sec	accuracy=0.656190	lr=0.397829
Epoch[10] Batch [549]	Speed: 4674.057068 samples/sec	accuracy=0.656300	lr=0.397796
Epoch[10] Batch [599]	Speed: 4820.285694 samples/sec	accuracy=0.656597	lr=0.397764
Epoch[10] Batch [649]	Speed: 4364.410479 samples/sec	accuracy=0.656951	lr=0.397731
Epoch[10] Batch [699]	Speed: 4755.916528 samples/sec	accuracy=0.657237	lr=0.397698
Epoch[10] Batch [749]	Speed: 4904.806064 samples/sec	accuracy=0.657550	lr=0.397665
Epoch[10] Batch [799]	Speed: 4439.643651 samples/sec	accuracy=0.657752	lr=0.397632
Epoch[10] Batch [849]	Speed: 4632.046708 samples/sec	accuracy=0.658078	lr=0.397598
Epoch[10] Batch [899]	Speed: 4804.339084 samples/sec	accuracy=0.658285	lr=0.397564
Epoch[10] Batch [949]	Speed: 4107.713873 samples/sec	accuracy=0.658576	lr=0.397530
Epoch[10] Batch [999]	Speed: 4581.587077 samples/sec	accuracy=0.658961	lr=0.397496
Epoch[10] Batch [1049]	Speed: 5018.464932 samples/sec	accuracy=0.659196	lr=0.397461
Epoch[10] Batch [1099]	Speed: 4928.227814 samples/sec	accuracy=0.659437	lr=0.397427
Epoch[10] Batch [1149]	Speed: 4661.661718 samples/sec	accuracy=0.659661	lr=0.397392
Epoch[10] Batch [1199]	Speed: 4859.598836 samples/sec	accuracy=0.659899	lr=0.397356
Epoch[10] Batch [1249]	Speed: 4810.022052 samples/sec	accuracy=0.660119	lr=0.397321
[Epoch 10] training: accuracy=0.660127
[Epoch 10] speed: 4628 samples/sec	time cost: 287.456416
[Epoch 10] validation: err-top1=0.498784 err-top5=0.247788
Epoch[11] Batch [49]	Speed: 4476.968139 samples/sec	accuracy=0.663577	lr=0.397284
Epoch[11] Batch [99]	Speed: 4366.438293 samples/sec	accuracy=0.663643	lr=0.397248
Epoch[11] Batch [149]	Speed: 4446.583982 samples/sec	accuracy=0.663444	lr=0.397212
Epoch[11] Batch [199]	Speed: 4901.690508 samples/sec	accuracy=0.664142	lr=0.397176
Epoch[11] Batch [249]	Speed: 4915.763877 samples/sec	accuracy=0.663982	lr=0.397139
Epoch[11] Batch [299]	Speed: 4750.686810 samples/sec	accuracy=0.664138	lr=0.397102
Epoch[11] Batch [349]	Speed: 4850.095994 samples/sec	accuracy=0.664262	lr=0.397065
Epoch[11] Batch [399]	Speed: 4841.295813 samples/sec	accuracy=0.664425	lr=0.397028
Epoch[11] Batch [449]	Speed: 4916.991165 samples/sec	accuracy=0.664776	lr=0.396990
Epoch[11] Batch [499]	Speed: 4739.262334 samples/sec	accuracy=0.664951	lr=0.396952
Epoch[11] Batch [549]	Speed: 4708.008716 samples/sec	accuracy=0.665134	lr=0.396914
Epoch[11] Batch [599]	Speed: 5039.402320 samples/sec	accuracy=0.665269	lr=0.396876
Epoch[11] Batch [649]	Speed: 4887.134881 samples/sec	accuracy=0.665533	lr=0.396837
Epoch[11] Batch [699]	Speed: 4785.445183 samples/sec	accuracy=0.665746	lr=0.396798
Epoch[11] Batch [749]	Speed: 4799.117572 samples/sec	accuracy=0.666059	lr=0.396759
Epoch[11] Batch [799]	Speed: 4700.319771 samples/sec	accuracy=0.666381	lr=0.396720
Epoch[11] Batch [849]	Speed: 4862.302392 samples/sec	accuracy=0.666646	lr=0.396680
Epoch[11] Batch [899]	Speed: 5069.968094 samples/sec	accuracy=0.666838	lr=0.396641
Epoch[11] Batch [949]	Speed: 4584.976309 samples/sec	accuracy=0.667031	lr=0.396601
Epoch[11] Batch [999]	Speed: 4760.839851 samples/sec	accuracy=0.667354	lr=0.396561
Epoch[11] Batch [1049]	Speed: 4928.299519 samples/sec	accuracy=0.667483	lr=0.396520
Epoch[11] Batch [1099]	Speed: 4742.981346 samples/sec	accuracy=0.667592	lr=0.396479
Epoch[11] Batch [1149]	Speed: 4552.285025 samples/sec	accuracy=0.667895	lr=0.396439
Epoch[11] Batch [1199]	Speed: 4581.343114 samples/sec	accuracy=0.668146	lr=0.396397
Epoch[11] Batch [1249]	Speed: 4914.856412 samples/sec	accuracy=0.668375	lr=0.396356
[Epoch 11] training: accuracy=0.668376
[Epoch 11] speed: 4751 samples/sec	time cost: 280.023701
[Epoch 11] validation: err-top1=0.479492 err-top5=0.224894
Epoch[12] Batch [49]	Speed: 4458.544980 samples/sec	accuracy=0.670644	lr=0.396314
Epoch[12] Batch [99]	Speed: 4761.355177 samples/sec	accuracy=0.670289	lr=0.396272
Epoch[12] Batch [149]	Speed: 5085.940974 samples/sec	accuracy=0.670593	lr=0.396230
Epoch[12] Batch [199]	Speed: 5201.049518 samples/sec	accuracy=0.670860	lr=0.396187
Epoch[12] Batch [249]	Speed: 4767.243524 samples/sec	accuracy=0.670860	lr=0.396145
Epoch[12] Batch [299]	Speed: 4662.631756 samples/sec	accuracy=0.671231	lr=0.396102
Epoch[12] Batch [349]	Speed: 4532.760537 samples/sec	accuracy=0.671353	lr=0.396059
Epoch[12] Batch [399]	Speed: 4958.653859 samples/sec	accuracy=0.671464	lr=0.396016
Epoch[12] Batch [449]	Speed: 4724.260155 samples/sec	accuracy=0.671456	lr=0.395972
Epoch[12] Batch [499]	Speed: 4733.129989 samples/sec	accuracy=0.671712	lr=0.395928
Epoch[12] Batch [549]	Speed: 4548.472595 samples/sec	accuracy=0.671773	lr=0.395885
Epoch[12] Batch [599]	Speed: 4628.935254 samples/sec	accuracy=0.671899	lr=0.395840
Epoch[12] Batch [649]	Speed: 4489.567322 samples/sec	accuracy=0.672053	lr=0.395796
Epoch[12] Batch [699]	Speed: 4703.052270 samples/sec	accuracy=0.672339	lr=0.395751
Epoch[12] Batch [749]	Speed: 4897.191932 samples/sec	accuracy=0.672557	lr=0.395706
Epoch[12] Batch [799]	Speed: 4661.295022 samples/sec	accuracy=0.672799	lr=0.395661
Epoch[12] Batch [849]	Speed: 4707.327902 samples/sec	accuracy=0.672982	lr=0.395616
Epoch[12] Batch [899]	Speed: 4824.600776 samples/sec	accuracy=0.673181	lr=0.395570
Epoch[12] Batch [949]	Speed: 4991.080825 samples/sec	accuracy=0.673396	lr=0.395524
Epoch[12] Batch [999]	Speed: 4741.224948 samples/sec	accuracy=0.673646	lr=0.395478
Epoch[12] Batch [1049]	Speed: 4587.218432 samples/sec	accuracy=0.673989	lr=0.395432
Epoch[12] Batch [1099]	Speed: 4894.844933 samples/sec	accuracy=0.674094	lr=0.395386
Epoch[12] Batch [1149]	Speed: 5021.304283 samples/sec	accuracy=0.674229	lr=0.395339
Epoch[12] Batch [1199]	Speed: 4851.488093 samples/sec	accuracy=0.674466	lr=0.395292
Epoch[12] Batch [1249]	Speed: 5015.250443 samples/sec	accuracy=0.674708	lr=0.395245
[Epoch 12] training: accuracy=0.674710
[Epoch 12] speed: 4766 samples/sec	time cost: 279.191423
[Epoch 12] validation: err-top1=0.489656 err-top5=0.234056
Epoch[13] Batch [49]	Speed: 4588.914432 samples/sec	accuracy=0.674597	lr=0.395196
Epoch[13] Batch [99]	Speed: 4894.690637 samples/sec	accuracy=0.675099	lr=0.395149
Epoch[13] Batch [149]	Speed: 4849.950311 samples/sec	accuracy=0.675576	lr=0.395101
Epoch[13] Batch [199]	Speed: 4957.866810 samples/sec	accuracy=0.676285	lr=0.395052
Epoch[13] Batch [249]	Speed: 4975.671614 samples/sec	accuracy=0.676744	lr=0.395004
Epoch[13] Batch [299]	Speed: 4921.083434 samples/sec	accuracy=0.677023	lr=0.394955
Epoch[13] Batch [349]	Speed: 5145.624679 samples/sec	accuracy=0.677200	lr=0.394907
Epoch[13] Batch [399]	Speed: 4916.071992 samples/sec	accuracy=0.677261	lr=0.394858
Epoch[13] Batch [449]	Speed: 4795.771810 samples/sec	accuracy=0.677455	lr=0.394808
Epoch[13] Batch [499]	Speed: 4329.009850 samples/sec	accuracy=0.677447	lr=0.394759
Epoch[13] Batch [549]	Speed: 4797.315824 samples/sec	accuracy=0.677641	lr=0.394709
Epoch[13] Batch [599]	Speed: 4817.054931 samples/sec	accuracy=0.677769	lr=0.394659
Epoch[13] Batch [649]	Speed: 4868.850528 samples/sec	accuracy=0.677985	lr=0.394609
Epoch[13] Batch [699]	Speed: 5074.891310 samples/sec	accuracy=0.678210	lr=0.394558
Epoch[13] Batch [749]	Speed: 4994.112828 samples/sec	accuracy=0.678460	lr=0.394507
Epoch[13] Batch [799]	Speed: 4991.380124 samples/sec	accuracy=0.678598	lr=0.394457
Epoch[13] Batch [849]	Speed: 4359.725702 samples/sec	accuracy=0.678760	lr=0.394405
Epoch[13] Batch [899]	Speed: 4807.732011 samples/sec	accuracy=0.679062	lr=0.394354
Epoch[13] Batch [949]	Speed: 4176.331906 samples/sec	accuracy=0.679188	lr=0.394302
Epoch[13] Batch [999]	Speed: 4814.297400 samples/sec	accuracy=0.679313	lr=0.394250
Epoch[13] Batch [1049]	Speed: 4865.955772 samples/sec	accuracy=0.679466	lr=0.394198
Epoch[13] Batch [1099]	Speed: 4953.113204 samples/sec	accuracy=0.679570	lr=0.394146
Epoch[13] Batch [1149]	Speed: 5069.885745 samples/sec	accuracy=0.679792	lr=0.394093
Epoch[13] Batch [1199]	Speed: 4664.085646 samples/sec	accuracy=0.679954	lr=0.394041
Epoch[13] Batch [1249]	Speed: 4726.159602 samples/sec	accuracy=0.680118	lr=0.393988
[Epoch 13] training: accuracy=0.680120
[Epoch 13] speed: 4796 samples/sec	time cost: 277.377055
[Epoch 13] validation: err-top1=0.499661 err-top5=0.245516
Epoch[14] Batch [49]	Speed: 4656.585023 samples/sec	accuracy=0.680825	lr=0.393933
Epoch[14] Batch [99]	Speed: 4691.763942 samples/sec	accuracy=0.681191	lr=0.393880
Epoch[14] Batch [149]	Speed: 4801.108402 samples/sec	accuracy=0.681600	lr=0.393826
Epoch[14] Batch [199]	Speed: 4935.129040 samples/sec	accuracy=0.681549	lr=0.393772
Epoch[14] Batch [249]	Speed: 4864.479106 samples/sec	accuracy=0.681402	lr=0.393718
Epoch[14] Batch [299]	Speed: 5058.759674 samples/sec	accuracy=0.681395	lr=0.393664
Epoch[14] Batch [349]	Speed: 5077.122127 samples/sec	accuracy=0.681618	lr=0.393609
Epoch[14] Batch [399]	Speed: 4685.143314 samples/sec	accuracy=0.681992	lr=0.393554
Epoch[14] Batch [449]	Speed: 5093.254217 samples/sec	accuracy=0.682187	lr=0.393499
Epoch[14] Batch [499]	Speed: 4973.520861 samples/sec	accuracy=0.682494	lr=0.393444
Epoch[14] Batch [549]	Speed: 5123.116398 samples/sec	accuracy=0.682731	lr=0.393388
Epoch[14] Batch [599]	Speed: 4678.845720 samples/sec	accuracy=0.682888	lr=0.393332
Epoch[14] Batch [649]	Speed: 4912.261870 samples/sec	accuracy=0.683031	lr=0.393276
Epoch[14] Batch [699]	Speed: 4599.288314 samples/sec	accuracy=0.683139	lr=0.393220
Epoch[14] Batch [749]	Speed: 4761.064356 samples/sec	accuracy=0.683218	lr=0.393163
Epoch[14] Batch [799]	Speed: 4788.623954 samples/sec	accuracy=0.683398	lr=0.393107
Epoch[14] Batch [849]	Speed: 4825.773634 samples/sec	accuracy=0.683581	lr=0.393050
Epoch[14] Batch [899]	Speed: 5085.356489 samples/sec	accuracy=0.683737	lr=0.392993
Epoch[14] Batch [949]	Speed: 4856.057662 samples/sec	accuracy=0.683967	lr=0.392935
Epoch[14] Batch [999]	Speed: 4888.955323 samples/sec	accuracy=0.684137	lr=0.392877
Epoch[14] Batch [1049]	Speed: 4698.821205 samples/sec	accuracy=0.684392	lr=0.392820
Epoch[14] Batch [1099]	Speed: 4808.375428 samples/sec	accuracy=0.684465	lr=0.392762
Epoch[14] Batch [1149]	Speed: 4855.405265 samples/sec	accuracy=0.684580	lr=0.392703
Epoch[14] Batch [1199]	Speed: 4600.340473 samples/sec	accuracy=0.684829	lr=0.392645
Epoch[14] Batch [1249]	Speed: 4907.163524 samples/sec	accuracy=0.685050	lr=0.392586
[Epoch 14] training: accuracy=0.685039
[Epoch 14] speed: 4840 samples/sec	time cost: 274.996906
[Epoch 14] validation: err-top1=0.455636 err-top5=0.204540
Epoch[15] Batch [49]	Speed: 4681.618824 samples/sec	accuracy=0.684455	lr=0.392526
Epoch[15] Batch [99]	Speed: 4888.670073 samples/sec	accuracy=0.686003	lr=0.392466
Epoch[15] Batch [149]	Speed: 4740.395737 samples/sec	accuracy=0.686546	lr=0.392407
Epoch[15] Batch [199]	Speed: 4691.207716 samples/sec	accuracy=0.686560	lr=0.392347
Epoch[15] Batch [249]	Speed: 4976.837767 samples/sec	accuracy=0.686369	lr=0.392287
Epoch[15] Batch [299]	Speed: 4885.913446 samples/sec	accuracy=0.686499	lr=0.392227
Epoch[15] Batch [349]	Speed: 4868.438925 samples/sec	accuracy=0.686454	lr=0.392167
Epoch[15] Batch [399]	Speed: 4688.597362 samples/sec	accuracy=0.686781	lr=0.392106
Epoch[15] Batch [449]	Speed: 4797.923974 samples/sec	accuracy=0.686923	lr=0.392045
Epoch[15] Batch [499]	Speed: 4741.812049 samples/sec	accuracy=0.687317	lr=0.391984
Epoch[15] Batch [549]	Speed: 4898.426611 samples/sec	accuracy=0.687646	lr=0.391923
Epoch[15] Batch [599]	Speed: 4964.058551 samples/sec	accuracy=0.687748	lr=0.391861
Epoch[15] Batch [649]	Speed: 4874.142880 samples/sec	accuracy=0.688059	lr=0.391799
Epoch[15] Batch [699]	Speed: 4694.316972 samples/sec	accuracy=0.688089	lr=0.391737
Epoch[15] Batch [749]	Speed: 4909.366569 samples/sec	accuracy=0.688350	lr=0.391675
Epoch[15] Batch [799]	Speed: 4805.096094 samples/sec	accuracy=0.688479	lr=0.391613
Epoch[15] Batch [849]	Speed: 5040.532050 samples/sec	accuracy=0.688561	lr=0.391550
Epoch[15] Batch [899]	Speed: 4706.471514 samples/sec	accuracy=0.688600	lr=0.391487
Epoch[15] Batch [949]	Speed: 4947.478910 samples/sec	accuracy=0.688680	lr=0.391424
Epoch[15] Batch [999]	Speed: 4995.796050 samples/sec	accuracy=0.688763	lr=0.391361
Epoch[15] Batch [1049]	Speed: 4911.391079 samples/sec	accuracy=0.688881	lr=0.391297
Epoch[15] Batch [1099]	Speed: 4914.853599 samples/sec	accuracy=0.689079	lr=0.391233
Epoch[15] Batch [1149]	Speed: 4895.973502 samples/sec	accuracy=0.689250	lr=0.391169
Epoch[15] Batch [1199]	Speed: 4898.855817 samples/sec	accuracy=0.689409	lr=0.391105
Epoch[15] Batch [1249]	Speed: 4696.919832 samples/sec	accuracy=0.689595	lr=0.391040
[Epoch 15] training: accuracy=0.689597
[Epoch 15] speed: 4838 samples/sec	time cost: 274.785204
[Epoch 15] validation: err-top1=0.463628 err-top5=0.215023
Epoch[16] Batch [49]	Speed: 4666.507330 samples/sec	accuracy=0.691556	lr=0.390974
Epoch[16] Batch [99]	Speed: 5022.540675 samples/sec	accuracy=0.690210	lr=0.390909
Epoch[16] Batch [149]	Speed: 4905.500715 samples/sec	accuracy=0.690552	lr=0.390844
Epoch[16] Batch [199]	Speed: 5031.933038 samples/sec	accuracy=0.690376	lr=0.390779
Epoch[16] Batch [249]	Speed: 5045.110439 samples/sec	accuracy=0.690295	lr=0.390713
Epoch[16] Batch [299]	Speed: 5048.002457 samples/sec	accuracy=0.690168	lr=0.390647
Epoch[16] Batch [349]	Speed: 4887.697046 samples/sec	accuracy=0.690301	lr=0.390581
Epoch[16] Batch [399]	Speed: 4847.224365 samples/sec	accuracy=0.690305	lr=0.390515
Epoch[16] Batch [449]	Speed: 4765.797809 samples/sec	accuracy=0.690447	lr=0.390448
Epoch[16] Batch [499]	Speed: 5166.148070 samples/sec	accuracy=0.690625	lr=0.390381
Epoch[16] Batch [549]	Speed: 4861.326406 samples/sec	accuracy=0.690882	lr=0.390314
Epoch[16] Batch [599]	Speed: 4825.657060 samples/sec	accuracy=0.691103	lr=0.390247
Epoch[16] Batch [649]	Speed: 4521.226849 samples/sec	accuracy=0.691199	lr=0.390180
Epoch[16] Batch [699]	Speed: 5031.270959 samples/sec	accuracy=0.691450	lr=0.390112
Epoch[16] Batch [749]	Speed: 4698.773706 samples/sec	accuracy=0.691642	lr=0.390044
Epoch[16] Batch [799]	Speed: 5000.358464 samples/sec	accuracy=0.691764	lr=0.389976
Epoch[16] Batch [849]	Speed: 5129.670452 samples/sec	accuracy=0.692002	lr=0.389907
Epoch[16] Batch [899]	Speed: 5079.873603 samples/sec	accuracy=0.692158	lr=0.389839
Epoch[16] Batch [949]	Speed: 4911.987601 samples/sec	accuracy=0.692344	lr=0.389770
Epoch[16] Batch [999]	Speed: 5022.030683 samples/sec	accuracy=0.692380	lr=0.389701
Epoch[16] Batch [1049]	Speed: 4854.756445 samples/sec	accuracy=0.692540	lr=0.389632
Epoch[16] Batch [1099]	Speed: 4534.931280 samples/sec	accuracy=0.692678	lr=0.389562
Epoch[16] Batch [1149]	Speed: 4649.285031 samples/sec	accuracy=0.692786	lr=0.389492
Epoch[16] Batch [1199]	Speed: 4831.809990 samples/sec	accuracy=0.692977	lr=0.389422
Epoch[16] Batch [1249]	Speed: 4718.327850 samples/sec	accuracy=0.693090	lr=0.389352
[Epoch 16] training: accuracy=0.693087
[Epoch 16] speed: 4867 samples/sec	time cost: 273.513046
[Epoch 16] validation: err-top1=0.483917 err-top5=0.225426
Epoch[17] Batch [49]	Speed: 4358.918029 samples/sec	accuracy=0.693974	lr=0.389280
Epoch[17] Batch [99]	Speed: 4875.195405 samples/sec	accuracy=0.693710	lr=0.389210
Epoch[17] Batch [149]	Speed: 5034.235755 samples/sec	accuracy=0.693723	lr=0.389139
Epoch[17] Batch [199]	Speed: 4899.915801 samples/sec	accuracy=0.693790	lr=0.389068
Epoch[17] Batch [249]	Speed: 4662.036872 samples/sec	accuracy=0.693806	lr=0.388996
Epoch[17] Batch [299]	Speed: 4604.679316 samples/sec	accuracy=0.693995	lr=0.388925
Epoch[17] Batch [349]	Speed: 4789.404218 samples/sec	accuracy=0.693988	lr=0.388853
Epoch[17] Batch [399]	Speed: 4829.040214 samples/sec	accuracy=0.694091	lr=0.388781
Epoch[17] Batch [449]	Speed: 4700.065983 samples/sec	accuracy=0.694341	lr=0.388709
Epoch[17] Batch [499]	Speed: 4740.471707 samples/sec	accuracy=0.694561	lr=0.388636
Epoch[17] Batch [549]	Speed: 4786.539547 samples/sec	accuracy=0.694776	lr=0.388564
Epoch[17] Batch [599]	Speed: 4587.680488 samples/sec	accuracy=0.694891	lr=0.388491
Epoch[17] Batch [649]	Speed: 4585.967380 samples/sec	accuracy=0.694891	lr=0.388418
Epoch[17] Batch [699]	Speed: 4769.031647 samples/sec	accuracy=0.694982	lr=0.388344
Epoch[17] Batch [749]	Speed: 5025.851675 samples/sec	accuracy=0.695108	lr=0.388271
Epoch[17] Batch [799]	Speed: 4666.193608 samples/sec	accuracy=0.695197	lr=0.388197
Epoch[17] Batch [849]	Speed: 4762.392078 samples/sec	accuracy=0.695306	lr=0.388123
Epoch[17] Batch [899]	Speed: 4659.147006 samples/sec	accuracy=0.695427	lr=0.388049
Epoch[17] Batch [949]	Speed: 5037.064275 samples/sec	accuracy=0.695555	lr=0.387974
Epoch[17] Batch [999]	Speed: 4753.891852 samples/sec	accuracy=0.695628	lr=0.387900
Epoch[17] Batch [1049]	Speed: 4705.910972 samples/sec	accuracy=0.695743	lr=0.387825
Epoch[17] Batch [1099]	Speed: 4993.147067 samples/sec	accuracy=0.695849	lr=0.387750
Epoch[17] Batch [1149]	Speed: 4655.675535 samples/sec	accuracy=0.695977	lr=0.387674
Epoch[17] Batch [1199]	Speed: 4916.282113 samples/sec	accuracy=0.696106	lr=0.387599
Epoch[17] Batch [1249]	Speed: 4940.305809 samples/sec	accuracy=0.696277	lr=0.387523
[Epoch 17] training: accuracy=0.696283
[Epoch 17] speed: 4762 samples/sec	time cost: 279.216872
[Epoch 17] validation: err-top1=0.553630 err-top5=0.296733
Epoch[18] Batch [49]	Speed: 4739.034338 samples/sec	accuracy=0.695182	lr=0.387445
Epoch[18] Batch [99]	Speed: 4900.926135 samples/sec	accuracy=0.695457	lr=0.387369
Epoch[18] Batch [149]	Speed: 4720.521559 samples/sec	accuracy=0.695880	lr=0.387293
Epoch[18] Batch [199]	Speed: 4856.340546 samples/sec	accuracy=0.696273	lr=0.387216
Epoch[18] Batch [249]	Speed: 5108.463067 samples/sec	accuracy=0.696348	lr=0.387139
Epoch[18] Batch [299]	Speed: 4757.618374 samples/sec	accuracy=0.696660	lr=0.387062
Epoch[18] Batch [349]	Speed: 4691.861118 samples/sec	accuracy=0.696800	lr=0.386984
Epoch[18] Batch [399]	Speed: 4720.066802 samples/sec	accuracy=0.696848	lr=0.386907
Epoch[18] Batch [449]	Speed: 4824.793178 samples/sec	accuracy=0.697087	lr=0.386829
Epoch[18] Batch [499]	Speed: 4853.082455 samples/sec	accuracy=0.697116	lr=0.386751
Epoch[18] Batch [549]	Speed: 4363.522334 samples/sec	accuracy=0.697342	lr=0.386673
Epoch[18] Batch [599]	Speed: 4285.275460 samples/sec	accuracy=0.697370	lr=0.386594
Epoch[18] Batch [649]	Speed: 4452.837945 samples/sec	accuracy=0.697448	lr=0.386515
Epoch[18] Batch [699]	Speed: 4889.971275 samples/sec	accuracy=0.697579	lr=0.386436
Epoch[18] Batch [749]	Speed: 4758.457101 samples/sec	accuracy=0.697648	lr=0.386357
Epoch[18] Batch [799]	Speed: 4787.214760 samples/sec	accuracy=0.697833	lr=0.386278
Epoch[18] Batch [849]	Speed: 4906.811341 samples/sec	accuracy=0.698034	lr=0.386198
Epoch[18] Batch [899]	Speed: 4938.379701 samples/sec	accuracy=0.698126	lr=0.386119
Epoch[18] Batch [949]	Speed: 5076.204630 samples/sec	accuracy=0.698251	lr=0.386038
Epoch[18] Batch [999]	Speed: 4971.609861 samples/sec	accuracy=0.698448	lr=0.385958
Epoch[18] Batch [1049]	Speed: 4912.727107 samples/sec	accuracy=0.698573	lr=0.385878
Epoch[18] Batch [1099]	Speed: 4845.847614 samples/sec	accuracy=0.698690	lr=0.385797
Epoch[18] Batch [1149]	Speed: 4933.498004 samples/sec	accuracy=0.698820	lr=0.385716
Epoch[18] Batch [1199]	Speed: 4865.656552 samples/sec	accuracy=0.698991	lr=0.385635
Epoch[18] Batch [1249]	Speed: 4976.896706 samples/sec	accuracy=0.699150	lr=0.385554
[Epoch 18] training: accuracy=0.699147
[Epoch 18] speed: 4793 samples/sec	time cost: 277.764489
[Epoch 18] validation: err-top1=0.494918 err-top5=0.245137
Epoch[19] Batch [49]	Speed: 4897.365708 samples/sec	accuracy=0.698950	lr=0.385470
Epoch[19] Batch [99]	Speed: 4719.449497 samples/sec	accuracy=0.698443	lr=0.385388
Epoch[19] Batch [149]	Speed: 4731.058269 samples/sec	accuracy=0.698605	lr=0.385306
Epoch[19] Batch [199]	Speed: 5108.461973 samples/sec	accuracy=0.698932	lr=0.385224
Epoch[19] Batch [249]	Speed: 5093.296860 samples/sec	accuracy=0.698877	lr=0.385142
Epoch[19] Batch [299]	Speed: 4542.871811 samples/sec	accuracy=0.699304	lr=0.385059
Epoch[19] Batch [349]	Speed: 4926.193245 samples/sec	accuracy=0.699354	lr=0.384976
Epoch[19] Batch [399]	Speed: 4868.942704 samples/sec	accuracy=0.699554	lr=0.384893
Epoch[19] Batch [449]	Speed: 4888.130938 samples/sec	accuracy=0.699711	lr=0.384810
Epoch[19] Batch [499]	Speed: 5122.002736 samples/sec	accuracy=0.700085	lr=0.384726
Epoch[19] Batch [549]	Speed: 4453.738162 samples/sec	accuracy=0.700339	lr=0.384642
Epoch[19] Batch [599]	Speed: 4640.959255 samples/sec	accuracy=0.700464	lr=0.384558
Epoch[19] Batch [649]	Speed: 4743.516178 samples/sec	accuracy=0.700585	lr=0.384474
Epoch[19] Batch [699]	Speed: 4597.073349 samples/sec	accuracy=0.700689	lr=0.384389
Epoch[19] Batch [749]	Speed: 4732.589674 samples/sec	accuracy=0.700728	lr=0.384305
Epoch[19] Batch [799]	Speed: 4698.038106 samples/sec	accuracy=0.700897	lr=0.384220
Epoch[19] Batch [849]	Speed: 4778.980107 samples/sec	accuracy=0.701115	lr=0.384135
Epoch[19] Batch [899]	Speed: 5005.839331 samples/sec	accuracy=0.701215	lr=0.384049
Epoch[19] Batch [949]	Speed: 5158.213517 samples/sec	accuracy=0.701321	lr=0.383964
Epoch[19] Batch [999]	Speed: 5161.779447 samples/sec	accuracy=0.701486	lr=0.383878
Epoch[19] Batch [1049]	Speed: 4655.162042 samples/sec	accuracy=0.701685	lr=0.383792
Epoch[19] Batch [1099]	Speed: 4748.787763 samples/sec	accuracy=0.701864	lr=0.383706
Epoch[19] Batch [1149]	Speed: 4708.660818 samples/sec	accuracy=0.702004	lr=0.383619
Epoch[19] Batch [1199]	Speed: 4917.312833 samples/sec	accuracy=0.702154	lr=0.383533
Epoch[19] Batch [1249]	Speed: 4195.476767 samples/sec	accuracy=0.702344	lr=0.383446
[Epoch 19] training: accuracy=0.702341
[Epoch 19] speed: 4788 samples/sec	time cost: 278.170651
[Epoch 19] validation: err-top1=0.456354 err-top5=0.206852
Epoch[20] Batch [49]	Speed: 4266.915230 samples/sec	accuracy=0.699425	lr=0.383357
Epoch[20] Batch [99]	Speed: 4760.761960 samples/sec	accuracy=0.700263	lr=0.383270
Epoch[20] Batch [149]	Speed: 4309.236903 samples/sec	accuracy=0.700790	lr=0.383182
Epoch[20] Batch [199]	Speed: 4472.456649 samples/sec	accuracy=0.701484	lr=0.383094
Epoch[20] Batch [249]	Speed: 4641.417856 samples/sec	accuracy=0.701654	lr=0.383006
Epoch[20] Batch [299]	Speed: 4485.740174 samples/sec	accuracy=0.701636	lr=0.382918
Epoch[20] Batch [349]	Speed: 4954.422766 samples/sec	accuracy=0.701817	lr=0.382830
Epoch[20] Batch [399]	Speed: 4813.113931 samples/sec	accuracy=0.701986	lr=0.382741
Epoch[20] Batch [449]	Speed: 4749.858173 samples/sec	accuracy=0.702252	lr=0.382652
Epoch[20] Batch [499]	Speed: 4935.001112 samples/sec	accuracy=0.702249	lr=0.382563
Epoch[20] Batch [549]	Speed: 4889.165470 samples/sec	accuracy=0.702572	lr=0.382474
Epoch[20] Batch [599]	Speed: 4421.278355 samples/sec	accuracy=0.702704	lr=0.382384
Epoch[20] Batch [649]	Speed: 4898.458455 samples/sec	accuracy=0.702940	lr=0.382295
Epoch[20] Batch [699]	Speed: 4621.922759 samples/sec	accuracy=0.703087	lr=0.382205
Epoch[20] Batch [749]	Speed: 4904.799454 samples/sec	accuracy=0.703249	lr=0.382115
Epoch[20] Batch [799]	Speed: 4756.774569 samples/sec	accuracy=0.703376	lr=0.382024
Epoch[20] Batch [849]	Speed: 4888.045155 samples/sec	accuracy=0.703657	lr=0.381934
Epoch[20] Batch [899]	Speed: 4628.560321 samples/sec	accuracy=0.703742	lr=0.381843
Epoch[20] Batch [949]	Speed: 5029.347486 samples/sec	accuracy=0.703910	lr=0.381752
Epoch[20] Batch [999]	Speed: 5188.961295 samples/sec	accuracy=0.704073	lr=0.381661
Epoch[20] Batch [1049]	Speed: 4811.520061 samples/sec	accuracy=0.704192	lr=0.381569
Epoch[20] Batch [1099]	Speed: 4657.719529 samples/sec	accuracy=0.704381	lr=0.381477
Epoch[20] Batch [1149]	Speed: 4384.616838 samples/sec	accuracy=0.704503	lr=0.381385
Epoch[20] Batch [1199]	Speed: 4753.340579 samples/sec	accuracy=0.704638	lr=0.381293
Epoch[20] Batch [1249]	Speed: 4676.792923 samples/sec	accuracy=0.704705	lr=0.381201
[Epoch 20] training: accuracy=0.704706
[Epoch 20] speed: 4701 samples/sec	time cost: 283.670197
[Epoch 20] validation: err-top1=0.454181 err-top5=0.207270
Epoch[21] Batch [49]	Speed: 4748.619541 samples/sec	accuracy=0.704094	lr=0.381107
Epoch[21] Batch [99]	Speed: 4810.187003 samples/sec	accuracy=0.704759	lr=0.381014
Epoch[21] Batch [149]	Speed: 4766.936851 samples/sec	accuracy=0.704647	lr=0.380921
Epoch[21] Batch [199]	Speed: 4746.550923 samples/sec	accuracy=0.704568	lr=0.380828
Epoch[21] Batch [249]	Speed: 4709.931166 samples/sec	accuracy=0.704282	lr=0.380734
Epoch[21] Batch [299]	Speed: 4328.764558 samples/sec	accuracy=0.704016	lr=0.380641
Epoch[21] Batch [349]	Speed: 4847.933006 samples/sec	accuracy=0.704162	lr=0.380547
Epoch[21] Batch [399]	Speed: 4815.663514 samples/sec	accuracy=0.704344	lr=0.380453
Epoch[21] Batch [449]	Speed: 4616.700242 samples/sec	accuracy=0.704420	lr=0.380359
Epoch[21] Batch [499]	Speed: 4808.213723 samples/sec	accuracy=0.704579	lr=0.380264
Epoch[21] Batch [549]	Speed: 4455.785414 samples/sec	accuracy=0.704709	lr=0.380169
Epoch[21] Batch [599]	Speed: 4849.390443 samples/sec	accuracy=0.704741	lr=0.380074
Epoch[21] Batch [649]	Speed: 4843.638923 samples/sec	accuracy=0.704991	lr=0.379979
Epoch[21] Batch [699]	Speed: 4372.368271 samples/sec	accuracy=0.705186	lr=0.379884
Epoch[21] Batch [749]	Speed: 4887.679803 samples/sec	accuracy=0.705274	lr=0.379788
Epoch[21] Batch [799]	Speed: 4671.349105 samples/sec	accuracy=0.705391	lr=0.379693
Epoch[21] Batch [849]	Speed: 4680.886953 samples/sec	accuracy=0.705411	lr=0.379597
Epoch[21] Batch [899]	Speed: 4584.672964 samples/sec	accuracy=0.705621	lr=0.379501
Epoch[21] Batch [949]	Speed: 4592.544820 samples/sec	accuracy=0.705738	lr=0.379404
Epoch[21] Batch [999]	Speed: 4436.172897 samples/sec	accuracy=0.705868	lr=0.379307
Epoch[21] Batch [1049]	Speed: 4602.090769 samples/sec	accuracy=0.705962	lr=0.379211
Epoch[21] Batch [1099]	Speed: 4614.521527 samples/sec	accuracy=0.706039	lr=0.379114
Epoch[21] Batch [1149]	Speed: 4593.145580 samples/sec	accuracy=0.706190	lr=0.379016
Epoch[21] Batch [1199]	Speed: 4988.266852 samples/sec	accuracy=0.706360	lr=0.378919
Epoch[21] Batch [1249]	Speed: 4396.853111 samples/sec	accuracy=0.706496	lr=0.378821
[Epoch 21] training: accuracy=0.706507
[Epoch 21] speed: 4659 samples/sec	time cost: 285.363507
[Epoch 21] validation: err-top1=0.435786 err-top5=0.189194
Epoch[22] Batch [49]	Speed: 4623.089502 samples/sec	accuracy=0.706328	lr=0.378721
Epoch[22] Batch [99]	Speed: 4210.788827 samples/sec	accuracy=0.706746	lr=0.378623
Epoch[22] Batch [149]	Speed: 5026.746822 samples/sec	accuracy=0.706520	lr=0.378525
Epoch[22] Batch [199]	Speed: 4666.115741 samples/sec	accuracy=0.706527	lr=0.378426
Epoch[22] Batch [249]	Speed: 4476.345783 samples/sec	accuracy=0.706509	lr=0.378328
Epoch[22] Batch [299]	Speed: 4736.843871 samples/sec	accuracy=0.706432	lr=0.378229
Epoch[22] Batch [349]	Speed: 4806.950713 samples/sec	accuracy=0.706369	lr=0.378129
Epoch[22] Batch [399]	Speed: 4479.092019 samples/sec	accuracy=0.706395	lr=0.378030
Epoch[22] Batch [449]	Speed: 4745.625672 samples/sec	accuracy=0.706472	lr=0.377930
Epoch[22] Batch [499]	Speed: 4340.679250 samples/sec	accuracy=0.706693	lr=0.377831
Epoch[22] Batch [549]	Speed: 4663.412716 samples/sec	accuracy=0.706799	lr=0.377731
Epoch[22] Batch [599]	Speed: 4973.581795 samples/sec	accuracy=0.706934	lr=0.377630
Epoch[22] Batch [649]	Speed: 4918.773534 samples/sec	accuracy=0.707107	lr=0.377530
Epoch[22] Batch [699]	Speed: 4986.970492 samples/sec	accuracy=0.707173	lr=0.377429
Epoch[22] Batch [749]	Speed: 4862.762729 samples/sec	accuracy=0.707336	lr=0.377328
Epoch[22] Batch [799]	Speed: 4921.562977 samples/sec	accuracy=0.707439	lr=0.377227
Epoch[22] Batch [849]	Speed: 4624.852274 samples/sec	accuracy=0.707418	lr=0.377126
Epoch[22] Batch [899]	Speed: 4538.076118 samples/sec	accuracy=0.707611	lr=0.377024
Epoch[22] Batch [949]	Speed: 4677.647102 samples/sec	accuracy=0.707874	lr=0.376923
Epoch[22] Batch [999]	Speed: 4675.158273 samples/sec	accuracy=0.708013	lr=0.376821
Epoch[22] Batch [1049]	Speed: 4845.689393 samples/sec	accuracy=0.708226	lr=0.376719
Epoch[22] Batch [1099]	Speed: 4518.151130 samples/sec	accuracy=0.708406	lr=0.376616
Epoch[22] Batch [1149]	Speed: 4945.515316 samples/sec	accuracy=0.708538	lr=0.376514
Epoch[22] Batch [1199]	Speed: 4762.535928 samples/sec	accuracy=0.708677	lr=0.376411
Epoch[22] Batch [1249]	Speed: 4774.816172 samples/sec	accuracy=0.708740	lr=0.376308
[Epoch 22] training: accuracy=0.708743
[Epoch 22] speed: 4698 samples/sec	time cost: 282.890461
[Epoch 22] validation: err-top1=0.433673 err-top5=0.187460
Epoch[23] Batch [49]	Speed: 4869.177410 samples/sec	accuracy=0.708859	lr=0.376203
Epoch[23] Batch [99]	Speed: 4806.425471 samples/sec	accuracy=0.709357	lr=0.376099
Epoch[23] Batch [149]	Speed: 4712.285924 samples/sec	accuracy=0.709178	lr=0.375996
Epoch[23] Batch [199]	Speed: 4818.686206 samples/sec	accuracy=0.709267	lr=0.375892
Epoch[23] Batch [249]	Speed: 4463.673162 samples/sec	accuracy=0.709814	lr=0.375788
Epoch[23] Batch [299]	Speed: 4787.535361 samples/sec	accuracy=0.709925	lr=0.375683
Epoch[23] Batch [349]	Speed: 4264.204811 samples/sec	accuracy=0.709909	lr=0.375579
Epoch[23] Batch [399]	Speed: 4258.409616 samples/sec	accuracy=0.709843	lr=0.375474
Epoch[23] Batch [449]	Speed: 5046.366884 samples/sec	accuracy=0.709776	lr=0.375369
Epoch[23] Batch [499]	Speed: 4628.440211 samples/sec	accuracy=0.709737	lr=0.375264
Epoch[23] Batch [549]	Speed: 4455.224760 samples/sec	accuracy=0.709879	lr=0.375159
Epoch[23] Batch [599]	Speed: 4213.178629 samples/sec	accuracy=0.709930	lr=0.375053
Epoch[23] Batch [649]	Speed: 4520.676634 samples/sec	accuracy=0.710068	lr=0.374948
Epoch[23] Batch [699]	Speed: 4688.344122 samples/sec	accuracy=0.710061	lr=0.374842
Epoch[23] Batch [749]	Speed: 4771.425849 samples/sec	accuracy=0.710170	lr=0.374736
Epoch[23] Batch [799]	Speed: 4392.657540 samples/sec	accuracy=0.710230	lr=0.374629
Epoch[23] Batch [849]	Speed: 4990.617795 samples/sec	accuracy=0.710350	lr=0.374523
Epoch[23] Batch [899]	Speed: 4993.004504 samples/sec	accuracy=0.710510	lr=0.374416
Epoch[23] Batch [949]	Speed: 4622.843686 samples/sec	accuracy=0.710549	lr=0.374309
Epoch[23] Batch [999]	Speed: 4923.314574 samples/sec	accuracy=0.710688	lr=0.374202
Epoch[23] Batch [1049]	Speed: 5140.245880 samples/sec	accuracy=0.710808	lr=0.374095
Epoch[23] Batch [1099]	Speed: 4851.707308 samples/sec	accuracy=0.710851	lr=0.373987
Epoch[23] Batch [1149]	Speed: 4577.553107 samples/sec	accuracy=0.710991	lr=0.373879
Epoch[23] Batch [1199]	Speed: 4403.932837 samples/sec	accuracy=0.711055	lr=0.373771
Epoch[23] Batch [1249]	Speed: 4488.989502 samples/sec	accuracy=0.711115	lr=0.373663
[Epoch 23] training: accuracy=0.711117
[Epoch 23] speed: 4648 samples/sec	time cost: 285.827890
[Epoch 23] validation: err-top1=0.437459 err-top5=0.193359
Epoch[24] Batch [49]	Speed: 4714.271478 samples/sec	accuracy=0.710588	lr=0.373552
Epoch[24] Batch [99]	Speed: 4796.350431 samples/sec	accuracy=0.710785	lr=0.373444
Epoch[24] Batch [149]	Speed: 4675.732892 samples/sec	accuracy=0.710777	lr=0.373335
Epoch[24] Batch [199]	Speed: 4633.693437 samples/sec	accuracy=0.710865	lr=0.373226
Epoch[24] Batch [249]	Speed: 4880.743490 samples/sec	accuracy=0.710912	lr=0.373117
Epoch[24] Batch [299]	Speed: 4949.258931 samples/sec	accuracy=0.710706	lr=0.373007
Epoch[24] Batch [349]	Speed: 4689.556012 samples/sec	accuracy=0.710918	lr=0.372897
Epoch[24] Batch [399]	Speed: 4744.443646 samples/sec	accuracy=0.711107	lr=0.372788
Epoch[24] Batch [449]	Speed: 4696.719003 samples/sec	accuracy=0.711025	lr=0.372678
Epoch[24] Batch [499]	Speed: 4851.857262 samples/sec	accuracy=0.711238	lr=0.372567
Epoch[24] Batch [549]	Speed: 4928.526748 samples/sec	accuracy=0.711195	lr=0.372457
Epoch[24] Batch [599]	Speed: 4602.107338 samples/sec	accuracy=0.711218	lr=0.372346
Epoch[24] Batch [649]	Speed: 4636.308759 samples/sec	accuracy=0.711192	lr=0.372235
Epoch[24] Batch [699]	Speed: 4386.053253 samples/sec	accuracy=0.711275	lr=0.372124
Epoch[24] Batch [749]	Speed: 4448.393546 samples/sec	accuracy=0.711471	lr=0.372013
Epoch[24] Batch [799]	Speed: 4740.610364 samples/sec	accuracy=0.711645	lr=0.371901
Epoch[24] Batch [849]	Speed: 4913.823798 samples/sec	accuracy=0.711796	lr=0.371790
Epoch[24] Batch [899]	Speed: 4920.238710 samples/sec	accuracy=0.711844	lr=0.371678
Epoch[24] Batch [949]	Speed: 4986.834999 samples/sec	accuracy=0.711973	lr=0.371565
Epoch[24] Batch [999]	Speed: 4556.639810 samples/sec	accuracy=0.712176	lr=0.371453
Epoch[24] Batch [1049]	Speed: 4755.606782 samples/sec	accuracy=0.712344	lr=0.371341
Epoch[24] Batch [1099]	Speed: 4735.487435 samples/sec	accuracy=0.712484	lr=0.371228
Epoch[24] Batch [1149]	Speed: 4469.745055 samples/sec	accuracy=0.712531	lr=0.371115
Epoch[24] Batch [1199]	Speed: 4160.452059 samples/sec	accuracy=0.712561	lr=0.371002
Epoch[24] Batch [1249]	Speed: 4421.409710 samples/sec	accuracy=0.712621	lr=0.370888
[Epoch 24] training: accuracy=0.712621
[Epoch 24] speed: 4678 samples/sec	time cost: 284.602295
[Epoch 24] validation: err-top1=0.444157 err-top5=0.196249
Epoch[25] Batch [49]	Speed: 4840.884272 samples/sec	accuracy=0.714239	lr=0.370773
Epoch[25] Batch [99]	Speed: 4646.213511 samples/sec	accuracy=0.713783	lr=0.370659
Epoch[25] Batch [149]	Speed: 4666.760042 samples/sec	accuracy=0.713618	lr=0.370545
Epoch[25] Batch [199]	Speed: 4692.666051 samples/sec	accuracy=0.713552	lr=0.370431
Epoch[25] Batch [249]	Speed: 4875.513509 samples/sec	accuracy=0.713479	lr=0.370316
Epoch[25] Batch [299]	Speed: 4913.073620 samples/sec	accuracy=0.713425	lr=0.370202
Epoch[25] Batch [349]	Speed: 4970.689946 samples/sec	accuracy=0.713694	lr=0.370087
Epoch[25] Batch [399]	Speed: 4732.603232 samples/sec	accuracy=0.713623	lr=0.369972
Epoch[25] Batch [449]	Speed: 4063.805598 samples/sec	accuracy=0.713497	lr=0.369857
Epoch[25] Batch [499]	Speed: 4932.732743 samples/sec	accuracy=0.713590	lr=0.369741
Epoch[25] Batch [549]	Speed: 4808.750340 samples/sec	accuracy=0.713796	lr=0.369626
Epoch[25] Batch [599]	Speed: 4872.654277 samples/sec	accuracy=0.713991	lr=0.369510
Epoch[25] Batch [649]	Speed: 4186.671175 samples/sec	accuracy=0.714030	lr=0.369394
Epoch[25] Batch [699]	Speed: 4667.772685 samples/sec	accuracy=0.714211	lr=0.369278
Epoch[25] Batch [749]	Speed: 4710.137051 samples/sec	accuracy=0.714303	lr=0.369161
Epoch[25] Batch [799]	Speed: 4821.170801 samples/sec	accuracy=0.714404	lr=0.369045
Epoch[25] Batch [849]	Speed: 4791.235517 samples/sec	accuracy=0.714399	lr=0.368928
Epoch[25] Batch [899]	Speed: 4528.690284 samples/sec	accuracy=0.714312	lr=0.368811
Epoch[25] Batch [949]	Speed: 4349.294508 samples/sec	accuracy=0.714412	lr=0.368694
Epoch[25] Batch [999]	Speed: 4786.461132 samples/sec	accuracy=0.714487	lr=0.368576
Epoch[25] Batch [1049]	Speed: 5047.589312 samples/sec	accuracy=0.714635	lr=0.368459
Epoch[25] Batch [1099]	Speed: 4645.934071 samples/sec	accuracy=0.714774	lr=0.368341
Epoch[25] Batch [1149]	Speed: 4863.767051 samples/sec	accuracy=0.714882	lr=0.368223
Epoch[25] Batch [1199]	Speed: 4916.620010 samples/sec	accuracy=0.714976	lr=0.368105
Epoch[25] Batch [1249]	Speed: 4686.876402 samples/sec	accuracy=0.715022	lr=0.367986
[Epoch 25] training: accuracy=0.715018
[Epoch 25] speed: 4703 samples/sec	time cost: 282.416698
[Epoch 25] validation: err-top1=0.440230 err-top5=0.195851
Epoch[26] Batch [49]	Speed: 4552.599059 samples/sec	accuracy=0.714285	lr=0.367865
Epoch[26] Batch [99]	Speed: 4658.719257 samples/sec	accuracy=0.714022	lr=0.367747
Epoch[26] Batch [149]	Speed: 4626.961600 samples/sec	accuracy=0.713544	lr=0.367628
Epoch[26] Batch [199]	Speed: 4936.185378 samples/sec	accuracy=0.713664	lr=0.367508
Epoch[26] Batch [249]	Speed: 4874.097412 samples/sec	accuracy=0.714203	lr=0.367389
Epoch[26] Batch [299]	Speed: 4913.565319 samples/sec	accuracy=0.714279	lr=0.367269
Epoch[26] Batch [349]	Speed: 4736.678166 samples/sec	accuracy=0.714272	lr=0.367150
Epoch[26] Batch [399]	Speed: 5074.007228 samples/sec	accuracy=0.714666	lr=0.367030
Epoch[26] Batch [449]	Speed: 4867.779887 samples/sec	accuracy=0.714835	lr=0.366909
Epoch[26] Batch [499]	Speed: 4619.890377 samples/sec	accuracy=0.714970	lr=0.366789
Epoch[26] Batch [549]	Speed: 4794.173991 samples/sec	accuracy=0.715111	lr=0.366668
Epoch[26] Batch [599]	Speed: 4963.237667 samples/sec	accuracy=0.715197	lr=0.366548
Epoch[26] Batch [649]	Speed: 4922.187582 samples/sec	accuracy=0.715381	lr=0.366426
Epoch[26] Batch [699]	Speed: 4936.339919 samples/sec	accuracy=0.715498	lr=0.366305
Epoch[26] Batch [749]	Speed: 4453.646259 samples/sec	accuracy=0.715635	lr=0.366184
Epoch[26] Batch [799]	Speed: 4803.888345 samples/sec	accuracy=0.715756	lr=0.366062
Epoch[26] Batch [849]	Speed: 4591.357023 samples/sec	accuracy=0.715898	lr=0.365940
Epoch[26] Batch [899]	Speed: 4850.792218 samples/sec	accuracy=0.716004	lr=0.365819
Epoch[26] Batch [949]	Speed: 4782.074258 samples/sec	accuracy=0.716091	lr=0.365696
Epoch[26] Batch [999]	Speed: 4694.591178 samples/sec	accuracy=0.716202	lr=0.365574
Epoch[26] Batch [1049]	Speed: 4732.247712 samples/sec	accuracy=0.716421	lr=0.365451
Epoch[26] Batch [1099]	Speed: 4720.106433 samples/sec	accuracy=0.716557	lr=0.365329
Epoch[26] Batch [1149]	Speed: 4966.464254 samples/sec	accuracy=0.716720	lr=0.365206
Epoch[26] Batch [1199]	Speed: 4883.724842 samples/sec	accuracy=0.716751	lr=0.365082
Epoch[26] Batch [1249]	Speed: 4760.157812 samples/sec	accuracy=0.716872	lr=0.364959
[Epoch 26] training: accuracy=0.716866
[Epoch 26] speed: 4778 samples/sec	time cost: 278.245257
[Epoch 26] validation: err-top1=0.449597 err-top5=0.197525
Epoch[27] Batch [49]	Speed: 4589.359862 samples/sec	accuracy=0.716418	lr=0.364833
Epoch[27] Batch [99]	Speed: 4896.928165 samples/sec	accuracy=0.716455	lr=0.364709
Epoch[27] Batch [149]	Speed: 4617.887387 samples/sec	accuracy=0.716101	lr=0.364585
Epoch[27] Batch [199]	Speed: 4745.973241 samples/sec	accuracy=0.716205	lr=0.364461
Epoch[27] Batch [249]	Speed: 4883.540483 samples/sec	accuracy=0.716228	lr=0.364337
Epoch[27] Batch [299]	Speed: 4542.845959 samples/sec	accuracy=0.716220	lr=0.364212
Epoch[27] Batch [349]	Speed: 4970.505175 samples/sec	accuracy=0.716552	lr=0.364087
Epoch[27] Batch [399]	Speed: 4884.362209 samples/sec	accuracy=0.716517	lr=0.363962
Epoch[27] Batch [449]	Speed: 4863.186809 samples/sec	accuracy=0.716655	lr=0.363837
Epoch[27] Batch [499]	Speed: 4573.538118 samples/sec	accuracy=0.716866	lr=0.363712
Epoch[27] Batch [549]	Speed: 4947.761489 samples/sec	accuracy=0.717011	lr=0.363586
Epoch[27] Batch [599]	Speed: 5040.614868 samples/sec	accuracy=0.717196	lr=0.363461
Epoch[27] Batch [649]	Speed: 4949.346534 samples/sec	accuracy=0.717240	lr=0.363335
Epoch[27] Batch [699]	Speed: 5038.127474 samples/sec	accuracy=0.717314	lr=0.363209
Epoch[27] Batch [749]	Speed: 4803.248385 samples/sec	accuracy=0.717451	lr=0.363082
Epoch[27] Batch [799]	Speed: 4938.995633 samples/sec	accuracy=0.717566	lr=0.362956
Epoch[27] Batch [849]	Speed: 4811.345856 samples/sec	accuracy=0.717767	lr=0.362829
Epoch[27] Batch [899]	Speed: 4865.926664 samples/sec	accuracy=0.717864	lr=0.362702
Epoch[27] Batch [949]	Speed: 4405.974050 samples/sec	accuracy=0.717944	lr=0.362575
Epoch[27] Batch [999]	Speed: 4920.766459 samples/sec	accuracy=0.717965	lr=0.362448
Epoch[27] Batch [1049]	Speed: 4917.334564 samples/sec	accuracy=0.718034	lr=0.362320
Epoch[27] Batch [1099]	Speed: 4664.168915 samples/sec	accuracy=0.718138	lr=0.362193
Epoch[27] Batch [1149]	Speed: 4716.876420 samples/sec	accuracy=0.718217	lr=0.362065
Epoch[27] Batch [1199]	Speed: 4993.206973 samples/sec	accuracy=0.718329	lr=0.361937
Epoch[27] Batch [1249]	Speed: 4563.896155 samples/sec	accuracy=0.718475	lr=0.361809
[Epoch 27] training: accuracy=0.718481
[Epoch 27] speed: 4794 samples/sec	time cost: 277.116235
[Epoch 27] validation: err-top1=0.474091 err-top5=0.224470
Epoch[28] Batch [49]	Speed: 4362.312147 samples/sec	accuracy=0.718674	lr=0.361678
Epoch[28] Batch [99]	Speed: 4869.779512 samples/sec	accuracy=0.719347	lr=0.361549
Epoch[28] Batch [149]	Speed: 4924.229796 samples/sec	accuracy=0.718497	lr=0.361420
Epoch[28] Batch [199]	Speed: 5004.308860 samples/sec	accuracy=0.718368	lr=0.361291
Epoch[28] Batch [249]	Speed: 4738.209343 samples/sec	accuracy=0.718458	lr=0.361162
Epoch[28] Batch [299]	Speed: 4651.898942 samples/sec	accuracy=0.718406	lr=0.361032
Epoch[28] Batch [349]	Speed: 4523.649542 samples/sec	accuracy=0.718282	lr=0.360903
Epoch[28] Batch [399]	Speed: 4855.379906 samples/sec	accuracy=0.718357	lr=0.360773
Epoch[28] Batch [449]	Speed: 4407.607225 samples/sec	accuracy=0.718358	lr=0.360643
Epoch[28] Batch [499]	Speed: 4776.755329 samples/sec	accuracy=0.718362	lr=0.360513
Epoch[28] Batch [549]	Speed: 4665.579974 samples/sec	accuracy=0.718585	lr=0.360382
Epoch[28] Batch [599]	Speed: 5116.990174 samples/sec	accuracy=0.718808	lr=0.360252
Epoch[28] Batch [649]	Speed: 4534.990752 samples/sec	accuracy=0.719061	lr=0.360121
Epoch[28] Batch [699]	Speed: 4750.362403 samples/sec	accuracy=0.719106	lr=0.359990
Epoch[28] Batch [749]	Speed: 4632.719411 samples/sec	accuracy=0.719222	lr=0.359859
Epoch[28] Batch [799]	Speed: 4892.497397 samples/sec	accuracy=0.719310	lr=0.359728
Epoch[28] Batch [849]	Speed: 4412.277670 samples/sec	accuracy=0.719353	lr=0.359596
Epoch[28] Batch [899]	Speed: 4841.826958 samples/sec	accuracy=0.719467	lr=0.359465
Epoch[28] Batch [949]	Speed: 4690.031028 samples/sec	accuracy=0.719512	lr=0.359333
Epoch[28] Batch [999]	Speed: 4595.743348 samples/sec	accuracy=0.719575	lr=0.359201
Epoch[28] Batch [1049]	Speed: 4901.056105 samples/sec	accuracy=0.719623	lr=0.359068
Epoch[28] Batch [1099]	Speed: 4954.580394 samples/sec	accuracy=0.719723	lr=0.358936
Epoch[28] Batch [1149]	Speed: 4650.578527 samples/sec	accuracy=0.719831	lr=0.358803
Epoch[28] Batch [1199]	Speed: 5015.497592 samples/sec	accuracy=0.719927	lr=0.358670
Epoch[28] Batch [1249]	Speed: 5054.661961 samples/sec	accuracy=0.719979	lr=0.358537
[Epoch 28] training: accuracy=0.719978
[Epoch 28] speed: 4738 samples/sec	time cost: 280.974522
[Epoch 28] validation: err-top1=0.435288 err-top5=0.191267
Epoch[29] Batch [49]	Speed: 4083.194379 samples/sec	accuracy=0.718675	lr=0.358401
Epoch[29] Batch [99]	Speed: 4540.334990 samples/sec	accuracy=0.718567	lr=0.358268
Epoch[29] Batch [149]	Speed: 4885.555415 samples/sec	accuracy=0.718328	lr=0.358134
Epoch[29] Batch [199]	Speed: 5082.087155 samples/sec	accuracy=0.718397	lr=0.358001
Epoch[29] Batch [249]	Speed: 4599.468484 samples/sec	accuracy=0.718853	lr=0.357867
Epoch[29] Batch [299]	Speed: 5132.887705 samples/sec	accuracy=0.719023	lr=0.357732
Epoch[29] Batch [349]	Speed: 5014.027471 samples/sec	accuracy=0.719257	lr=0.357598
Epoch[29] Batch [399]	Speed: 4868.751512 samples/sec	accuracy=0.719178	lr=0.357464
Epoch[29] Batch [449]	Speed: 4804.333280 samples/sec	accuracy=0.719326	lr=0.357329
Epoch[29] Batch [499]	Speed: 4947.723187 samples/sec	accuracy=0.719327	lr=0.357194
Epoch[29] Batch [549]	Speed: 4915.143263 samples/sec	accuracy=0.719538	lr=0.357059
Epoch[29] Batch [599]	Speed: 5032.149642 samples/sec	accuracy=0.719550	lr=0.356924
Epoch[29] Batch [649]	Speed: 4857.286845 samples/sec	accuracy=0.719730	lr=0.356788
Epoch[29] Batch [699]	Speed: 5104.356173 samples/sec	accuracy=0.719932	lr=0.356652
Epoch[29] Batch [749]	Speed: 4721.713495 samples/sec	accuracy=0.720133	lr=0.356517
Epoch[29] Batch [799]	Speed: 4330.272615 samples/sec	accuracy=0.720273	lr=0.356380
Epoch[29] Batch [849]	Speed: 4751.295179 samples/sec	accuracy=0.720528	lr=0.356244
Epoch[29] Batch [899]	Speed: 4921.472181 samples/sec	accuracy=0.720682	lr=0.356108
Epoch[29] Batch [949]	Speed: 4491.842319 samples/sec	accuracy=0.720813	lr=0.355971
Epoch[29] Batch [999]	Speed: 4699.869720 samples/sec	accuracy=0.720885	lr=0.355834
Epoch[29] Batch [1049]	Speed: 4648.768317 samples/sec	accuracy=0.721019	lr=0.355698
Epoch[29] Batch [1099]	Speed: 4308.078070 samples/sec	accuracy=0.721132	lr=0.355560
Epoch[29] Batch [1149]	Speed: 4991.371307 samples/sec	accuracy=0.721192	lr=0.355423
Epoch[29] Batch [1199]	Speed: 4554.385359 samples/sec	accuracy=0.721332	lr=0.355285
Epoch[29] Batch [1249]	Speed: 4760.696842 samples/sec	accuracy=0.721467	lr=0.355148
[Epoch 29] training: accuracy=0.721472
[Epoch 29] speed: 4740 samples/sec	time cost: 281.201408
[Epoch 29] validation: err-top1=0.456217 err-top5=0.209127
Epoch[30] Batch [49]	Speed: 4845.733895 samples/sec	accuracy=0.721814	lr=0.355007
Epoch[30] Batch [99]	Speed: 4534.545663 samples/sec	accuracy=0.722518	lr=0.354869
Epoch[30] Batch [149]	Speed: 4581.185666 samples/sec	accuracy=0.722477	lr=0.354731
Epoch[30] Batch [199]	Speed: 4797.616986 samples/sec	accuracy=0.722022	lr=0.354592
Epoch[30] Batch [249]	Speed: 4876.396205 samples/sec	accuracy=0.721796	lr=0.354454
Epoch[30] Batch [299]	Speed: 4916.277724 samples/sec	accuracy=0.721778	lr=0.354315
Epoch[30] Batch [349]	Speed: 4517.758762 samples/sec	accuracy=0.721905	lr=0.354176
Epoch[30] Batch [399]	Speed: 4502.272828 samples/sec	accuracy=0.721932	lr=0.354037
Epoch[30] Batch [449]	Speed: 4693.840164 samples/sec	accuracy=0.721971	lr=0.353897
Epoch[30] Batch [499]	Speed: 4977.168697 samples/sec	accuracy=0.722128	lr=0.353758
Epoch[30] Batch [549]	Speed: 4781.196315 samples/sec	accuracy=0.722265	lr=0.353618
Epoch[30] Batch [599]	Speed: 4795.360798 samples/sec	accuracy=0.722310	lr=0.353478
Epoch[30] Batch [649]	Speed: 4720.992491 samples/sec	accuracy=0.722434	lr=0.353338
Epoch[30] Batch [699]	Speed: 4702.611273 samples/sec	accuracy=0.722637	lr=0.353198
Epoch[30] Batch [749]	Speed: 4928.977536 samples/sec	accuracy=0.722658	lr=0.353057
Epoch[30] Batch [799]	Speed: 5025.181318 samples/sec	accuracy=0.722729	lr=0.352917
Epoch[30] Batch [849]	Speed: 4635.345440 samples/sec	accuracy=0.722755	lr=0.352776
Epoch[30] Batch [899]	Speed: 4265.324485 samples/sec	accuracy=0.722762	lr=0.352635
Epoch[30] Batch [949]	Speed: 4864.188001 samples/sec	accuracy=0.722869	lr=0.352493
Epoch[30] Batch [999]	Speed: 4395.021813 samples/sec	accuracy=0.723093	lr=0.352352
Epoch[30] Batch [1049]	Speed: 4827.007821 samples/sec	accuracy=0.723170	lr=0.352211
Epoch[30] Batch [1099]	Speed: 4610.307042 samples/sec	accuracy=0.723280	lr=0.352069
Epoch[30] Batch [1149]	Speed: 4520.322648 samples/sec	accuracy=0.723375	lr=0.351927
Epoch[30] Batch [1199]	Speed: 4282.464982 samples/sec	accuracy=0.723491	lr=0.351785
Epoch[30] Batch [1249]	Speed: 4683.497612 samples/sec	accuracy=0.723580	lr=0.351642
[Epoch 30] training: accuracy=0.723580
[Epoch 30] speed: 4674 samples/sec	time cost: 284.985867
[Epoch 30] validation: err-top1=0.439453 err-top5=0.195532
Epoch[31] Batch [49]	Speed: 4742.981869 samples/sec	accuracy=0.723241	lr=0.351497
Epoch[31] Batch [99]	Speed: 4866.241134 samples/sec	accuracy=0.723029	lr=0.351354
Epoch[31] Batch [149]	Speed: 4604.714071 samples/sec	accuracy=0.722602	lr=0.351212
Epoch[31] Batch [199]	Speed: 4839.608187 samples/sec	accuracy=0.722364	lr=0.351069
Epoch[31] Batch [249]	Speed: 5011.094491 samples/sec	accuracy=0.722548	lr=0.350925
Epoch[31] Batch [299]	Speed: 4376.899618 samples/sec	accuracy=0.722782	lr=0.350782
Epoch[31] Batch [349]	Speed: 4667.035703 samples/sec	accuracy=0.722802	lr=0.350639
Epoch[31] Batch [399]	Speed: 5014.172408 samples/sec	accuracy=0.722676	lr=0.350495
Epoch[31] Batch [449]	Speed: 5188.653629 samples/sec	accuracy=0.722833	lr=0.350351
Epoch[31] Batch [499]	Speed: 4648.225157 samples/sec	accuracy=0.722873	lr=0.350207
Epoch[31] Batch [549]	Speed: 4751.813803 samples/sec	accuracy=0.722888	lr=0.350062
Epoch[31] Batch [599]	Speed: 4673.807024 samples/sec	accuracy=0.722934	lr=0.349918
Epoch[31] Batch [649]	Speed: 4604.425879 samples/sec	accuracy=0.723050	lr=0.349773
Epoch[31] Batch [699]	Speed: 4854.403953 samples/sec	accuracy=0.723186	lr=0.349629
Epoch[31] Batch [749]	Speed: 4969.667092 samples/sec	accuracy=0.723365	lr=0.349484
Epoch[31] Batch [799]	Speed: 4776.108237 samples/sec	accuracy=0.723505	lr=0.349338
Epoch[31] Batch [849]	Speed: 4855.603644 samples/sec	accuracy=0.723535	lr=0.349193
Epoch[31] Batch [899]	Speed: 5192.138006 samples/sec	accuracy=0.723659	lr=0.349048
Epoch[31] Batch [949]	Speed: 4834.451688 samples/sec	accuracy=0.723754	lr=0.348902
Epoch[31] Batch [999]	Speed: 4761.045250 samples/sec	accuracy=0.723876	lr=0.348756
Epoch[31] Batch [1049]	Speed: 4975.214320 samples/sec	accuracy=0.724012	lr=0.348610
Epoch[31] Batch [1099]	Speed: 4197.061929 samples/sec	accuracy=0.724109	lr=0.348464
Epoch[31] Batch [1149]	Speed: 4666.674145 samples/sec	accuracy=0.724271	lr=0.348317
Epoch[31] Batch [1199]	Speed: 4486.418194 samples/sec	accuracy=0.724404	lr=0.348171
Epoch[31] Batch [1249]	Speed: 5074.869243 samples/sec	accuracy=0.724517	lr=0.348024
[Epoch 31] training: accuracy=0.724524
[Epoch 31] speed: 4770 samples/sec	time cost: 279.343802
[Epoch 31] validation: err-top1=0.440091 err-top5=0.195213
Epoch[32] Batch [49]	Speed: 4624.677579 samples/sec	accuracy=0.725277	lr=0.347874
Epoch[32] Batch [99]	Speed: 4922.022532 samples/sec	accuracy=0.725172	lr=0.347727
Epoch[32] Batch [149]	Speed: 4750.443527 samples/sec	accuracy=0.724942	lr=0.347580
Epoch[32] Batch [199]	Speed: 4751.798346 samples/sec	accuracy=0.724534	lr=0.347432
Epoch[32] Batch [249]	Speed: 5063.484908 samples/sec	accuracy=0.724258	lr=0.347285
Epoch[32] Batch [299]	Speed: 4815.008753 samples/sec	accuracy=0.724593	lr=0.347137
Epoch[32] Batch [349]	Speed: 5083.834545 samples/sec	accuracy=0.724689	lr=0.346989
Epoch[32] Batch [399]	Speed: 4844.774167 samples/sec	accuracy=0.724867	lr=0.346841
Epoch[32] Batch [449]	Speed: 4813.988636 samples/sec	accuracy=0.725044	lr=0.346692
Epoch[32] Batch [499]	Speed: 4451.492639 samples/sec	accuracy=0.725136	lr=0.346544
Epoch[32] Batch [549]	Speed: 4525.338809 samples/sec	accuracy=0.725244	lr=0.346395
Epoch[32] Batch [599]	Speed: 5007.545532 samples/sec	accuracy=0.725385	lr=0.346246
Epoch[32] Batch [649]	Speed: 4597.519971 samples/sec	accuracy=0.725348	lr=0.346097
Epoch[32] Batch [699]	Speed: 4617.868718 samples/sec	accuracy=0.725441	lr=0.345948
Epoch[32] Batch [749]	Speed: 4887.212846 samples/sec	accuracy=0.725350	lr=0.345799
Epoch[32] Batch [799]	Speed: 4549.788290 samples/sec	accuracy=0.725425	lr=0.345649
Epoch[32] Batch [849]	Speed: 4836.211957 samples/sec	accuracy=0.725549	lr=0.345499
Epoch[32] Batch [899]	Speed: 5079.550021 samples/sec	accuracy=0.725637	lr=0.345349
Epoch[32] Batch [949]	Speed: 4863.471295 samples/sec	accuracy=0.725702	lr=0.345199
Epoch[32] Batch [999]	Speed: 4824.030274 samples/sec	accuracy=0.725661	lr=0.345049
Epoch[32] Batch [1049]	Speed: 4893.683318 samples/sec	accuracy=0.725731	lr=0.344899
Epoch[32] Batch [1099]	Speed: 4445.555418 samples/sec	accuracy=0.725941	lr=0.344748
Epoch[32] Batch [1149]	Speed: 4869.496937 samples/sec	accuracy=0.726104	lr=0.344597
Epoch[32] Batch [1199]	Speed: 4579.751836 samples/sec	accuracy=0.726221	lr=0.344446
Epoch[32] Batch [1249]	Speed: 4605.599310 samples/sec	accuracy=0.726301	lr=0.344295
[Epoch 32] training: accuracy=0.726301
[Epoch 32] speed: 4760 samples/sec	time cost: 279.544886
[Epoch 32] validation: err-top1=0.444176 err-top5=0.196987
Epoch[33] Batch [49]	Speed: 4864.793499 samples/sec	accuracy=0.724804	lr=0.344141
Epoch[33] Batch [99]	Speed: 4892.960793 samples/sec	accuracy=0.725308	lr=0.343989
Epoch[33] Batch [149]	Speed: 4851.098269 samples/sec	accuracy=0.725405	lr=0.343838
Epoch[33] Batch [199]	Speed: 5090.444309 samples/sec	accuracy=0.725334	lr=0.343686
Epoch[33] Batch [249]	Speed: 4633.203674 samples/sec	accuracy=0.725510	lr=0.343534
Epoch[33] Batch [299]	Speed: 4800.136224 samples/sec	accuracy=0.725509	lr=0.343382
Epoch[33] Batch [349]	Speed: 4495.983905 samples/sec	accuracy=0.725982	lr=0.343229
Epoch[33] Batch [399]	Speed: 4898.063728 samples/sec	accuracy=0.726174	lr=0.343077
Epoch[33] Batch [449]	Speed: 5010.462785 samples/sec	accuracy=0.726261	lr=0.342924
Epoch[33] Batch [499]	Speed: 4904.272324 samples/sec	accuracy=0.726092	lr=0.342771
Epoch[33] Batch [549]	Speed: 4714.540568 samples/sec	accuracy=0.726176	lr=0.342618
Epoch[33] Batch [599]	Speed: 4714.738576 samples/sec	accuracy=0.726244	lr=0.342465
Epoch[33] Batch [649]	Speed: 4480.016623 samples/sec	accuracy=0.726338	lr=0.342312
Epoch[33] Batch [699]	Speed: 4468.673854 samples/sec	accuracy=0.726506	lr=0.342158
Epoch[33] Batch [749]	Speed: 4676.425167 samples/sec	accuracy=0.726521	lr=0.342005
Epoch[33] Batch [799]	Speed: 5079.184794 samples/sec	accuracy=0.726658	lr=0.341851
Epoch[33] Batch [849]	Speed: 4799.369942 samples/sec	accuracy=0.726743	lr=0.341697
Epoch[33] Batch [899]	Speed: 4673.961951 samples/sec	accuracy=0.726853	lr=0.341543
Epoch[33] Batch [949]	Speed: 4460.215886 samples/sec	accuracy=0.727030	lr=0.341388
Epoch[33] Batch [999]	Speed: 4904.069388 samples/sec	accuracy=0.727188	lr=0.341234
Epoch[33] Batch [1049]	Speed: 4309.383390 samples/sec	accuracy=0.727287	lr=0.341079
Epoch[33] Batch [1099]	Speed: 4903.201499 samples/sec	accuracy=0.727351	lr=0.340924
Epoch[33] Batch [1149]	Speed: 4894.144485 samples/sec	accuracy=0.727370	lr=0.340769
Epoch[33] Batch [1199]	Speed: 4564.544647 samples/sec	accuracy=0.727481	lr=0.340614
Epoch[33] Batch [1249]	Speed: 5028.212992 samples/sec	accuracy=0.727679	lr=0.340458
[Epoch 33] training: accuracy=0.727678
[Epoch 33] speed: 4747 samples/sec	time cost: 280.152945
[Epoch 33] validation: err-top1=0.455098 err-top5=0.207211
Epoch[34] Batch [49]	Speed: 4438.028934 samples/sec	accuracy=0.726003	lr=0.340300
Epoch[34] Batch [99]	Speed: 4862.541963 samples/sec	accuracy=0.726745	lr=0.340144
Epoch[34] Batch [149]	Speed: 4797.710450 samples/sec	accuracy=0.726807	lr=0.339988
Epoch[34] Batch [199]	Speed: 4493.589995 samples/sec	accuracy=0.726603	lr=0.339832
Epoch[34] Batch [249]	Speed: 4831.844779 samples/sec	accuracy=0.726945	lr=0.339676
Epoch[34] Batch [299]	Speed: 4582.457090 samples/sec	accuracy=0.727413	lr=0.339520
Epoch[34] Batch [349]	Speed: 4643.768955 samples/sec	accuracy=0.727603	lr=0.339363
Epoch[34] Batch [399]	Speed: 4662.316632 samples/sec	accuracy=0.727787	lr=0.339206
Epoch[34] Batch [449]	Speed: 5029.611811 samples/sec	accuracy=0.727729	lr=0.339049
Epoch[34] Batch [499]	Speed: 4757.807578 samples/sec	accuracy=0.728027	lr=0.338892
Epoch[34] Batch [549]	Speed: 4502.970113 samples/sec	accuracy=0.728166	lr=0.338735
Epoch[34] Batch [599]	Speed: 4329.360166 samples/sec	accuracy=0.728165	lr=0.338578
Epoch[34] Batch [649]	Speed: 4976.826694 samples/sec	accuracy=0.728312	lr=0.338420
Epoch[34] Batch [699]	Speed: 5153.932512 samples/sec	accuracy=0.728404	lr=0.338263
Epoch[34] Batch [749]	Speed: 4461.446716 samples/sec	accuracy=0.728568	lr=0.338105
Epoch[34] Batch [799]	Speed: 4730.202812 samples/sec	accuracy=0.728782	lr=0.337947
Epoch[34] Batch [849]	Speed: 4737.999638 samples/sec	accuracy=0.729000	lr=0.337788
Epoch[34] Batch [899]	Speed: 4749.352475 samples/sec	accuracy=0.729050	lr=0.337630
Epoch[34] Batch [949]	Speed: 4412.479932 samples/sec	accuracy=0.729147	lr=0.337472
Epoch[34] Batch [999]	Speed: 4465.626667 samples/sec	accuracy=0.729203	lr=0.337313
Epoch[34] Batch [1049]	Speed: 4538.292093 samples/sec	accuracy=0.729325	lr=0.337154
Epoch[34] Batch [1099]	Speed: 4894.838462 samples/sec	accuracy=0.729381	lr=0.336995
Epoch[34] Batch [1149]	Speed: 4625.359401 samples/sec	accuracy=0.729420	lr=0.336836
Epoch[34] Batch [1199]	Speed: 4481.863421 samples/sec	accuracy=0.729457	lr=0.336677
Epoch[34] Batch [1249]	Speed: 4749.285568 samples/sec	accuracy=0.729548	lr=0.336517
[Epoch 34] training: accuracy=0.729557
[Epoch 34] speed: 4663 samples/sec	time cost: 284.966748
[Epoch 34] validation: err-top1=0.430888 err-top5=0.188497
Epoch[35] Batch [49]	Speed: 4784.426894 samples/sec	accuracy=0.730449	lr=0.336354
Epoch[35] Batch [99]	Speed: 4774.809483 samples/sec	accuracy=0.728660	lr=0.336194
Epoch[35] Batch [149]	Speed: 4859.217823 samples/sec	accuracy=0.729000	lr=0.336034
Epoch[35] Batch [199]	Speed: 4772.372325 samples/sec	accuracy=0.728632	lr=0.335874
Epoch[35] Batch [249]	Speed: 4895.543685 samples/sec	accuracy=0.728722	lr=0.335714
Epoch[35] Batch [299]	Speed: 4891.413546 samples/sec	accuracy=0.728665	lr=0.335553
Epoch[35] Batch [349]	Speed: 4866.596781 samples/sec	accuracy=0.728987	lr=0.335393
Epoch[35] Batch [399]	Speed: 4519.521436 samples/sec	accuracy=0.728870	lr=0.335232
Epoch[35] Batch [449]	Speed: 4642.758770 samples/sec	accuracy=0.728971	lr=0.335071
Epoch[35] Batch [499]	Speed: 4922.667001 samples/sec	accuracy=0.728964	lr=0.334910
Epoch[35] Batch [549]	Speed: 4495.849683 samples/sec	accuracy=0.729011	lr=0.334749
Epoch[35] Batch [599]	Speed: 4828.796983 samples/sec	accuracy=0.729103	lr=0.334587
Epoch[35] Batch [649]	Speed: 4658.783637 samples/sec	accuracy=0.729127	lr=0.334425
Epoch[35] Batch [699]	Speed: 5064.592376 samples/sec	accuracy=0.729262	lr=0.334264
Epoch[35] Batch [749]	Speed: 5031.751114 samples/sec	accuracy=0.729348	lr=0.334102
Epoch[35] Batch [799]	Speed: 4436.210837 samples/sec	accuracy=0.729568	lr=0.333940
Epoch[35] Batch [849]	Speed: 4667.725405 samples/sec	accuracy=0.729699	lr=0.333777
Epoch[35] Batch [899]	Speed: 4723.910459 samples/sec	accuracy=0.729871	lr=0.333615
Epoch[35] Batch [949]	Speed: 4874.356845 samples/sec	accuracy=0.729971	lr=0.333452
Epoch[35] Batch [999]	Speed: 4603.822559 samples/sec	accuracy=0.730056	lr=0.333290
Epoch[35] Batch [1049]	Speed: 4804.758409 samples/sec	accuracy=0.730093	lr=0.333127
Epoch[35] Batch [1099]	Speed: 5166.184112 samples/sec	accuracy=0.730211	lr=0.332964
Epoch[35] Batch [1149]	Speed: 5242.493468 samples/sec	accuracy=0.730304	lr=0.332801
Epoch[35] Batch [1199]	Speed: 4702.546397 samples/sec	accuracy=0.730422	lr=0.332637
Epoch[35] Batch [1249]	Speed: 4584.756358 samples/sec	accuracy=0.730523	lr=0.332474
[Epoch 35] training: accuracy=0.730530
[Epoch 35] speed: 4780 samples/sec	time cost: 278.449662
[Epoch 35] validation: err-top1=0.424067 err-top5=0.182577
Epoch[36] Batch [49]	Speed: 4447.839357 samples/sec	accuracy=0.729048	lr=0.332307
Epoch[36] Batch [99]	Speed: 5054.842571 samples/sec	accuracy=0.729517	lr=0.332143
Epoch[36] Batch [149]	Speed: 4537.722279 samples/sec	accuracy=0.729530	lr=0.331979
Epoch[36] Batch [199]	Speed: 4737.500851 samples/sec	accuracy=0.729405	lr=0.331815
Epoch[36] Batch [249]	Speed: 4928.044830 samples/sec	accuracy=0.729904	lr=0.331650
Epoch[36] Batch [299]	Speed: 4752.601049 samples/sec	accuracy=0.730055	lr=0.331486
Epoch[36] Batch [349]	Speed: 4925.852449 samples/sec	accuracy=0.730215	lr=0.331321
Epoch[36] Batch [399]	Speed: 4862.611329 samples/sec	accuracy=0.730357	lr=0.331157
Epoch[36] Batch [449]	Speed: 4914.666658 samples/sec	accuracy=0.730689	lr=0.330992
Epoch[36] Batch [499]	Speed: 4502.579338 samples/sec	accuracy=0.730755	lr=0.330827
Epoch[36] Batch [549]	Speed: 4395.531519 samples/sec	accuracy=0.730880	lr=0.330661
Epoch[36] Batch [599]	Speed: 4630.743428 samples/sec	accuracy=0.731047	lr=0.330496
Epoch[36] Batch [649]	Speed: 4431.165916 samples/sec	accuracy=0.731002	lr=0.330330
Epoch[36] Batch [699]	Speed: 4906.299808 samples/sec	accuracy=0.731234	lr=0.330165
Epoch[36] Batch [749]	Speed: 5055.075551 samples/sec	accuracy=0.731327	lr=0.329999
Epoch[36] Batch [799]	Speed: 4681.857762 samples/sec	accuracy=0.731383	lr=0.329833
Epoch[36] Batch [849]	Speed: 4685.307886 samples/sec	accuracy=0.731557	lr=0.329667
Epoch[36] Batch [899]	Speed: 4589.582020 samples/sec	accuracy=0.731710	lr=0.329500
Epoch[36] Batch [949]	Speed: 4546.765631 samples/sec	accuracy=0.731908	lr=0.329334
Epoch[36] Batch [999]	Speed: 4723.754489 samples/sec	accuracy=0.731982	lr=0.329167
Epoch[36] Batch [1049]	Speed: 4537.204277 samples/sec	accuracy=0.732077	lr=0.329000
Epoch[36] Batch [1099]	Speed: 4197.863657 samples/sec	accuracy=0.732081	lr=0.328833
Epoch[36] Batch [1149]	Speed: 4477.984586 samples/sec	accuracy=0.732107	lr=0.328666
Epoch[36] Batch [1199]	Speed: 4759.329877 samples/sec	accuracy=0.732171	lr=0.328499
Epoch[36] Batch [1249]	Speed: 4631.165156 samples/sec	accuracy=0.732207	lr=0.328332
[Epoch 36] training: accuracy=0.732212
[Epoch 36] speed: 4661 samples/sec	time cost: 286.546466
[Epoch 36] validation: err-top1=0.409558 err-top5=0.171317
Epoch[37] Batch [49]	Speed: 4750.571418 samples/sec	accuracy=0.731088	lr=0.328161
Epoch[37] Batch [99]	Speed: 4818.372231 samples/sec	accuracy=0.730608	lr=0.327993
Epoch[37] Batch [149]	Speed: 5010.675557 samples/sec	accuracy=0.731332	lr=0.327825
Epoch[37] Batch [199]	Speed: 4726.015236 samples/sec	accuracy=0.731930	lr=0.327657
Epoch[37] Batch [249]	Speed: 4775.828462 samples/sec	accuracy=0.731988	lr=0.327489
Epoch[37] Batch [299]	Speed: 4433.920054 samples/sec	accuracy=0.732145	lr=0.327321
Epoch[37] Batch [349]	Speed: 4737.306256 samples/sec	accuracy=0.732045	lr=0.327152
Epoch[37] Batch [399]	Speed: 4677.589842 samples/sec	accuracy=0.731974	lr=0.326983
Epoch[37] Batch [449]	Speed: 4524.258337 samples/sec	accuracy=0.731841	lr=0.326815
Epoch[37] Batch [499]	Speed: 4773.767274 samples/sec	accuracy=0.731951	lr=0.326646
Epoch[37] Batch [549]	Speed: 4746.774817 samples/sec	accuracy=0.732065	lr=0.326477
Epoch[37] Batch [599]	Speed: 4443.427358 samples/sec	accuracy=0.732216	lr=0.326307
Epoch[37] Batch [649]	Speed: 4755.530853 samples/sec	accuracy=0.732436	lr=0.326138
Epoch[37] Batch [699]	Speed: 4941.465901 samples/sec	accuracy=0.732460	lr=0.325968
Epoch[37] Batch [749]	Speed: 4902.118496 samples/sec	accuracy=0.732547	lr=0.325799
Epoch[37] Batch [799]	Speed: 4966.938322 samples/sec	accuracy=0.732714	lr=0.325629
Epoch[37] Batch [849]	Speed: 5070.230123 samples/sec	accuracy=0.732802	lr=0.325459
Epoch[37] Batch [899]	Speed: 4931.366335 samples/sec	accuracy=0.732967	lr=0.325289
Epoch[37] Batch [949]	Speed: 4858.679338 samples/sec	accuracy=0.733043	lr=0.325118
Epoch[37] Batch [999]	Speed: 4645.061091 samples/sec	accuracy=0.733171	lr=0.324948
Epoch[37] Batch [1049]	Speed: 5057.327565 samples/sec	accuracy=0.733302	lr=0.324777
Epoch[37] Batch [1099]	Speed: 5141.434575 samples/sec	accuracy=0.733389	lr=0.324607
Epoch[37] Batch [1149]	Speed: 4694.572294 samples/sec	accuracy=0.733504	lr=0.324436
Epoch[37] Batch [1199]	Speed: 5202.525498 samples/sec	accuracy=0.733574	lr=0.324265
Epoch[37] Batch [1249]	Speed: 5204.513485 samples/sec	accuracy=0.733653	lr=0.324094
[Epoch 37] training: accuracy=0.733654
[Epoch 37] speed: 4818 samples/sec	time cost: 276.006105
[Epoch 37] validation: err-top1=0.409399 err-top5=0.171018
Epoch[38] Batch [49]	Speed: 4769.640274 samples/sec	accuracy=0.732666	lr=0.323919
Epoch[38] Batch [99]	Speed: 4849.254001 samples/sec	accuracy=0.732602	lr=0.323747
Epoch[38] Batch [149]	Speed: 4853.300388 samples/sec	accuracy=0.732160	lr=0.323576
Epoch[38] Batch [199]	Speed: 4913.004381 samples/sec	accuracy=0.732180	lr=0.323404
Epoch[38] Batch [249]	Speed: 4767.860164 samples/sec	accuracy=0.732346	lr=0.323232
Epoch[38] Batch [299]	Speed: 4824.809438 samples/sec	accuracy=0.732534	lr=0.323060
Epoch[38] Batch [349]	Speed: 4805.285976 samples/sec	accuracy=0.732469	lr=0.322888
Epoch[38] Batch [399]	Speed: 4347.481305 samples/sec	accuracy=0.732459	lr=0.322715
Epoch[38] Batch [449]	Speed: 4683.567888 samples/sec	accuracy=0.732714	lr=0.322543
Epoch[38] Batch [499]	Speed: 5059.079063 samples/sec	accuracy=0.732623	lr=0.322370
Epoch[38] Batch [549]	Speed: 4845.021305 samples/sec	accuracy=0.732982	lr=0.322197
Epoch[38] Batch [599]	Speed: 4710.620689 samples/sec	accuracy=0.733058	lr=0.322025
Epoch[38] Batch [649]	Speed: 4830.438830 samples/sec	accuracy=0.733056	lr=0.321851
Epoch[38] Batch [699]	Speed: 4641.882467 samples/sec	accuracy=0.733146	lr=0.321678
Epoch[38] Batch [749]	Speed: 4536.270006 samples/sec	accuracy=0.733312	lr=0.321505
Epoch[38] Batch [799]	Speed: 4600.865501 samples/sec	accuracy=0.733507	lr=0.321331
Epoch[38] Batch [849]	Speed: 4774.981371 samples/sec	accuracy=0.733639	lr=0.321158
Epoch[38] Batch [899]	Speed: 4960.106004 samples/sec	accuracy=0.733719	lr=0.320984
Epoch[38] Batch [949]	Speed: 4982.319182 samples/sec	accuracy=0.733877	lr=0.320810
Epoch[38] Batch [999]	Speed: 4481.673454 samples/sec	accuracy=0.734007	lr=0.320636
Epoch[38] Batch [1049]	Speed: 4761.834820 samples/sec	accuracy=0.734101	lr=0.320462
Epoch[38] Batch [1099]	Speed: 4576.059337 samples/sec	accuracy=0.734194	lr=0.320287
Epoch[38] Batch [1149]	Speed: 4665.680326 samples/sec	accuracy=0.734336	lr=0.320113
Epoch[38] Batch [1199]	Speed: 4890.804189 samples/sec	accuracy=0.734492	lr=0.319938
Epoch[38] Batch [1249]	Speed: 4949.835480 samples/sec	accuracy=0.734626	lr=0.319763
[Epoch 38] training: accuracy=0.734634
[Epoch 38] speed: 4751 samples/sec	time cost: 279.560588
[Epoch 38] validation: err-top1=0.423968 err-top5=0.181023
Epoch[39] Batch [49]	Speed: 4742.422546 samples/sec	accuracy=0.733001	lr=0.319585
Epoch[39] Batch [99]	Speed: 4872.746597 samples/sec	accuracy=0.733602	lr=0.319410
Epoch[39] Batch [149]	Speed: 5007.710179 samples/sec	accuracy=0.733589	lr=0.319234
Epoch[39] Batch [199]	Speed: 4639.536585 samples/sec	accuracy=0.733937	lr=0.319059
Epoch[39] Batch [249]	Speed: 4989.629967 samples/sec	accuracy=0.733879	lr=0.318883
Epoch[39] Batch [299]	Speed: 4846.543056 samples/sec	accuracy=0.734001	lr=0.318708
Epoch[39] Batch [349]	Speed: 4851.360739 samples/sec	accuracy=0.734313	lr=0.318532
Epoch[39] Batch [399]	Speed: 3724.949714 samples/sec	accuracy=0.734499	lr=0.318356
Epoch[39] Batch [449]	Speed: 4971.927549 samples/sec	accuracy=0.734592	lr=0.318180
Epoch[39] Batch [499]	Speed: 4985.803175 samples/sec	accuracy=0.734622	lr=0.318004
Epoch[39] Batch [549]	Speed: 4596.655544 samples/sec	accuracy=0.734435	lr=0.317827
Epoch[39] Batch [599]	Speed: 5071.554926 samples/sec	accuracy=0.734530	lr=0.317651
Epoch[39] Batch [649]	Speed: 5100.821377 samples/sec	accuracy=0.734516	lr=0.317474
Epoch[39] Batch [699]	Speed: 4783.628643 samples/sec	accuracy=0.734639	lr=0.317297
Epoch[39] Batch [749]	Speed: 4692.950012 samples/sec	accuracy=0.734718	lr=0.317120
Epoch[39] Batch [799]	Speed: 4677.427849 samples/sec	accuracy=0.734787	lr=0.316943
Epoch[39] Batch [849]	Speed: 5016.897314 samples/sec	accuracy=0.734940	lr=0.316766
Epoch[39] Batch [899]	Speed: 5128.739869 samples/sec	accuracy=0.735097	lr=0.316589
Epoch[39] Batch [949]	Speed: 4848.724071 samples/sec	accuracy=0.735191	lr=0.316411
Epoch[39] Batch [999]	Speed: 4648.934269 samples/sec	accuracy=0.735257	lr=0.316233
Epoch[39] Batch [1049]	Speed: 4841.211010 samples/sec	accuracy=0.735339	lr=0.316056
Epoch[39] Batch [1099]	Speed: 4494.432926 samples/sec	accuracy=0.735396	lr=0.315878
Epoch[39] Batch [1149]	Speed: 4846.473273 samples/sec	accuracy=0.735484	lr=0.315700
Epoch[39] Batch [1199]	Speed: 4702.190127 samples/sec	accuracy=0.735570	lr=0.315522
Epoch[39] Batch [1249]	Speed: 5081.317909 samples/sec	accuracy=0.735662	lr=0.315343
[Epoch 39] training: accuracy=0.735664
[Epoch 39] speed: 4782 samples/sec	time cost: 278.622262
[Epoch 39] validation: err-top1=0.421696 err-top5=0.182816
Epoch[40] Batch [49]	Speed: 4622.553618 samples/sec	accuracy=0.735812	lr=0.315161
Epoch[40] Batch [99]	Speed: 5043.520092 samples/sec	accuracy=0.735968	lr=0.314983
Epoch[40] Batch [149]	Speed: 4816.697305 samples/sec	accuracy=0.735468	lr=0.314804
Epoch[40] Batch [199]	Speed: 4751.684583 samples/sec	accuracy=0.735666	lr=0.314625
Epoch[40] Batch [249]	Speed: 4451.953597 samples/sec	accuracy=0.735964	lr=0.314446
Epoch[40] Batch [299]	Speed: 4628.879180 samples/sec	accuracy=0.736229	lr=0.314267
Epoch[40] Batch [349]	Speed: 4835.532869 samples/sec	accuracy=0.735861	lr=0.314087
Epoch[40] Batch [399]	Speed: 4776.558346 samples/sec	accuracy=0.736026	lr=0.313908
Epoch[40] Batch [449]	Speed: 4741.413793 samples/sec	accuracy=0.736012	lr=0.313728
Epoch[40] Batch [499]	Speed: 4604.583151 samples/sec	accuracy=0.736137	lr=0.313549
Epoch[40] Batch [549]	Speed: 4257.680489 samples/sec	accuracy=0.736259	lr=0.313369
Epoch[40] Batch [599]	Speed: 4778.002093 samples/sec	accuracy=0.736238	lr=0.313189
Epoch[40] Batch [649]	Speed: 4379.449147 samples/sec	accuracy=0.736306	lr=0.313009
Epoch[40] Batch [699]	Speed: 4742.480986 samples/sec	accuracy=0.736332	lr=0.312829
Epoch[40] Batch [749]	Speed: 4912.159170 samples/sec	accuracy=0.736425	lr=0.312648
Epoch[40] Batch [799]	Speed: 4675.901997 samples/sec	accuracy=0.736480	lr=0.312468
Epoch[40] Batch [849]	Speed: 4717.256680 samples/sec	accuracy=0.736609	lr=0.312287
Epoch[40] Batch [899]	Speed: 4907.960915 samples/sec	accuracy=0.736641	lr=0.312106
Epoch[40] Batch [949]	Speed: 4721.721385 samples/sec	accuracy=0.736733	lr=0.311925
Epoch[40] Batch [999]	Speed: 4979.440835 samples/sec	accuracy=0.736861	lr=0.311744
Epoch[40] Batch [1049]	Speed: 4524.353083 samples/sec	accuracy=0.736999	lr=0.311563
Epoch[40] Batch [1099]	Speed: 4755.870185 samples/sec	accuracy=0.737134	lr=0.311382
Epoch[40] Batch [1149]	Speed: 4631.258240 samples/sec	accuracy=0.737294	lr=0.311201
Epoch[40] Batch [1199]	Speed: 4868.357474 samples/sec	accuracy=0.737393	lr=0.311019
Epoch[40] Batch [1249]	Speed: 4615.076874 samples/sec	accuracy=0.737454	lr=0.310837
[Epoch 40] training: accuracy=0.737453
[Epoch 40] speed: 4695 samples/sec	time cost: 282.906202
[Epoch 40] validation: err-top1=0.443075 err-top5=0.200216
Epoch[41] Batch [49]	Speed: 4571.120867 samples/sec	accuracy=0.735222	lr=0.310652
Epoch[41] Batch [99]	Speed: 4579.685520 samples/sec	accuracy=0.735356	lr=0.310470
Epoch[41] Batch [149]	Speed: 4828.730968 samples/sec	accuracy=0.736089	lr=0.310288
Epoch[41] Batch [199]	Speed: 4997.887601 samples/sec	accuracy=0.736052	lr=0.310105
Epoch[41] Batch [249]	Speed: 4482.588831 samples/sec	accuracy=0.735860	lr=0.309923
Epoch[41] Batch [299]	Speed: 4588.213609 samples/sec	accuracy=0.735974	lr=0.309741
Epoch[41] Batch [349]	Speed: 5224.516895 samples/sec	accuracy=0.736240	lr=0.309558
Epoch[41] Batch [399]	Speed: 5106.123162 samples/sec	accuracy=0.736250	lr=0.309375
Epoch[41] Batch [449]	Speed: 4907.010692 samples/sec	accuracy=0.736595	lr=0.309192
Epoch[41] Batch [499]	Speed: 4516.453641 samples/sec	accuracy=0.736751	lr=0.309009
Epoch[41] Batch [549]	Speed: 4601.783478 samples/sec	accuracy=0.736690	lr=0.308826
Epoch[41] Batch [599]	Speed: 4730.630345 samples/sec	accuracy=0.736816	lr=0.308643
Epoch[41] Batch [649]	Speed: 4557.901320 samples/sec	accuracy=0.736968	lr=0.308459
Epoch[41] Batch [699]	Speed: 5034.065937 samples/sec	accuracy=0.737098	lr=0.308276
Epoch[41] Batch [749]	Speed: 4949.493915 samples/sec	accuracy=0.737163	lr=0.308092
Epoch[41] Batch [799]	Speed: 4729.104582 samples/sec	accuracy=0.737251	lr=0.307908
Epoch[41] Batch [849]	Speed: 4962.682419 samples/sec	accuracy=0.737443	lr=0.307725
Epoch[41] Batch [899]	Speed: 4560.382661 samples/sec	accuracy=0.737530	lr=0.307540
Epoch[41] Batch [949]	Speed: 4423.409483 samples/sec	accuracy=0.737679	lr=0.307356
Epoch[41] Batch [999]	Speed: 4765.828904 samples/sec	accuracy=0.737721	lr=0.307172
Epoch[41] Batch [1049]	Speed: 4723.769348 samples/sec	accuracy=0.737815	lr=0.306988
Epoch[41] Batch [1099]	Speed: 4786.142169 samples/sec	accuracy=0.737983	lr=0.306803
Epoch[41] Batch [1149]	Speed: 4267.827922 samples/sec	accuracy=0.738027	lr=0.306618
Epoch[41] Batch [1199]	Speed: 4913.901156 samples/sec	accuracy=0.738147	lr=0.306433
Epoch[41] Batch [1249]	Speed: 4852.760911 samples/sec	accuracy=0.738260	lr=0.306249
[Epoch 41] training: accuracy=0.738265
[Epoch 41] speed: 4730 samples/sec	time cost: 282.009856
[Epoch 41] validation: err-top1=0.446110 err-top5=0.202348
Epoch[42] Batch [49]	Speed: 4509.739396 samples/sec	accuracy=0.738330	lr=0.306060
Epoch[42] Batch [99]	Speed: 4669.432430 samples/sec	accuracy=0.738660	lr=0.305875
Epoch[42] Batch [149]	Speed: 4762.248025 samples/sec	accuracy=0.738078	lr=0.305689
Epoch[42] Batch [199]	Speed: 5014.374138 samples/sec	accuracy=0.737799	lr=0.305504
Epoch[42] Batch [249]	Speed: 4808.963017 samples/sec	accuracy=0.737924	lr=0.305318
Epoch[42] Batch [299]	Speed: 4894.695657 samples/sec	accuracy=0.738418	lr=0.305133
Epoch[42] Batch [349]	Speed: 5098.802723 samples/sec	accuracy=0.738596	lr=0.304947
Epoch[42] Batch [399]	Speed: 5049.382865 samples/sec	accuracy=0.738611	lr=0.304761
Epoch[42] Batch [449]	Speed: 4604.470206 samples/sec	accuracy=0.738664	lr=0.304575
Epoch[42] Batch [499]	Speed: 4667.005580 samples/sec	accuracy=0.738865	lr=0.304388
Epoch[42] Batch [549]	Speed: 4822.162888 samples/sec	accuracy=0.738772	lr=0.304202
Epoch[42] Batch [599]	Speed: 4651.507988 samples/sec	accuracy=0.738865	lr=0.304016
Epoch[42] Batch [649]	Speed: 4435.899643 samples/sec	accuracy=0.738971	lr=0.303829
Epoch[42] Batch [699]	Speed: 4940.697144 samples/sec	accuracy=0.739140	lr=0.303642
Epoch[42] Batch [749]	Speed: 4644.181810 samples/sec	accuracy=0.739315	lr=0.303456
Epoch[42] Batch [799]	Speed: 4874.523914 samples/sec	accuracy=0.739364	lr=0.303269
Epoch[42] Batch [849]	Speed: 4531.094502 samples/sec	accuracy=0.739534	lr=0.303082
Epoch[42] Batch [899]	Speed: 4826.992414 samples/sec	accuracy=0.739704	lr=0.302894
Epoch[42] Batch [949]	Speed: 4714.605258 samples/sec	accuracy=0.739860	lr=0.302707
Epoch[42] Batch [999]	Speed: 4833.284726 samples/sec	accuracy=0.739957	lr=0.302520
Epoch[42] Batch [1049]	Speed: 4909.344010 samples/sec	accuracy=0.740037	lr=0.302332
Epoch[42] Batch [1099]	Speed: 4542.994248 samples/sec	accuracy=0.740075	lr=0.302144
Epoch[42] Batch [1149]	Speed: 4914.135606 samples/sec	accuracy=0.740140	lr=0.301957
Epoch[42] Batch [1199]	Speed: 5002.455235 samples/sec	accuracy=0.740215	lr=0.301769
Epoch[42] Batch [1249]	Speed: 4746.534977 samples/sec	accuracy=0.740214	lr=0.301581
[Epoch 42] training: accuracy=0.740223
[Epoch 42] speed: 4766 samples/sec	time cost: 279.303287
[Epoch 42] validation: err-top1=0.393395 err-top5=0.161133
Epoch[43] Batch [49]	Speed: 4726.031877 samples/sec	accuracy=0.740181	lr=0.301389
Epoch[43] Batch [99]	Speed: 4767.114310 samples/sec	accuracy=0.738734	lr=0.301200
Epoch[43] Batch [149]	Speed: 4629.067562 samples/sec	accuracy=0.738607	lr=0.301012
Epoch[43] Batch [199]	Speed: 4517.786705 samples/sec	accuracy=0.738706	lr=0.300823
Epoch[43] Batch [249]	Speed: 4598.761184 samples/sec	accuracy=0.738982	lr=0.300635
Epoch[43] Batch [299]	Speed: 4993.495729 samples/sec	accuracy=0.738693	lr=0.300446
Epoch[43] Batch [349]	Speed: 4700.911397 samples/sec	accuracy=0.739003	lr=0.300257
Epoch[43] Batch [399]	Speed: 4939.643760 samples/sec	accuracy=0.739117	lr=0.300068
Epoch[43] Batch [449]	Speed: 4668.207577 samples/sec	accuracy=0.739454	lr=0.299879
Epoch[43] Batch [499]	Speed: 4836.129293 samples/sec	accuracy=0.739618	lr=0.299690
Epoch[43] Batch [549]	Speed: 4762.945770 samples/sec	accuracy=0.739894	lr=0.299500
Epoch[43] Batch [599]	Speed: 5217.587186 samples/sec	accuracy=0.740045	lr=0.299311
Epoch[43] Batch [649]	Speed: 4986.955090 samples/sec	accuracy=0.740045	lr=0.299121
Epoch[43] Batch [699]	Speed: 5031.471710 samples/sec	accuracy=0.740208	lr=0.298932
Epoch[43] Batch [749]	Speed: 4873.405211 samples/sec	accuracy=0.740304	lr=0.298742
Epoch[43] Batch [799]	Speed: 4882.392107 samples/sec	accuracy=0.740460	lr=0.298552
Epoch[43] Batch [849]	Speed: 4672.894661 samples/sec	accuracy=0.740536	lr=0.298362
Epoch[43] Batch [899]	Speed: 4684.719772 samples/sec	accuracy=0.740636	lr=0.298171
Epoch[43] Batch [949]	Speed: 5048.448781 samples/sec	accuracy=0.740681	lr=0.297981
Epoch[43] Batch [999]	Speed: 4830.902390 samples/sec	accuracy=0.740825	lr=0.297791
Epoch[43] Batch [1049]	Speed: 4947.555394 samples/sec	accuracy=0.740991	lr=0.297600
Epoch[43] Batch [1099]	Speed: 4787.907351 samples/sec	accuracy=0.741093	lr=0.297409
Epoch[43] Batch [1149]	Speed: 4636.572827 samples/sec	accuracy=0.741162	lr=0.297219
Epoch[43] Batch [1199]	Speed: 4923.235001 samples/sec	accuracy=0.741247	lr=0.297028
Epoch[43] Batch [1249]	Speed: 4840.999727 samples/sec	accuracy=0.741264	lr=0.296837
[Epoch 43] training: accuracy=0.741261
[Epoch 43] speed: 4810 samples/sec	time cost: 276.561940
[Epoch 43] validation: err-top1=0.444894 err-top5=0.205238
Epoch[44] Batch [49]	Speed: 4695.394995 samples/sec	accuracy=0.739432	lr=0.296642
Epoch[44] Batch [99]	Speed: 4759.537888 samples/sec	accuracy=0.739417	lr=0.296451
Epoch[44] Batch [149]	Speed: 4856.228092 samples/sec	accuracy=0.739705	lr=0.296259
Epoch[44] Batch [199]	Speed: 4492.906610 samples/sec	accuracy=0.739881	lr=0.296068
Epoch[44] Batch [249]	Speed: 4892.193567 samples/sec	accuracy=0.740272	lr=0.295876
Epoch[44] Batch [299]	Speed: 4630.785767 samples/sec	accuracy=0.740201	lr=0.295684
Epoch[44] Batch [349]	Speed: 4972.814644 samples/sec	accuracy=0.739738	lr=0.295493
Epoch[44] Batch [399]	Speed: 4918.349730 samples/sec	accuracy=0.739977	lr=0.295301
Epoch[44] Batch [449]	Speed: 4771.538334 samples/sec	accuracy=0.740135	lr=0.295109
Epoch[44] Batch [499]	Speed: 4678.061316 samples/sec	accuracy=0.740219	lr=0.294917
Epoch[44] Batch [549]	Speed: 4683.665951 samples/sec	accuracy=0.740400	lr=0.294724
Epoch[44] Batch [599]	Speed: 4960.327582 samples/sec	accuracy=0.740488	lr=0.294532
Epoch[44] Batch [649]	Speed: 4429.737542 samples/sec	accuracy=0.740583	lr=0.294339
Epoch[44] Batch [699]	Speed: 4802.939641 samples/sec	accuracy=0.740784	lr=0.294147
Epoch[44] Batch [749]	Speed: 4549.630112 samples/sec	accuracy=0.740897	lr=0.293954
Epoch[44] Batch [799]	Speed: 4563.107445 samples/sec	accuracy=0.740957	lr=0.293761
Epoch[44] Batch [849]	Speed: 4590.922394 samples/sec	accuracy=0.741148	lr=0.293568
Epoch[44] Batch [899]	Speed: 4616.533309 samples/sec	accuracy=0.741341	lr=0.293375
Epoch[44] Batch [949]	Speed: 4774.338368 samples/sec	accuracy=0.741485	lr=0.293182
Epoch[44] Batch [999]	Speed: 4711.667655 samples/sec	accuracy=0.741594	lr=0.292989
Epoch[44] Batch [1049]	Speed: 4763.791237 samples/sec	accuracy=0.741625	lr=0.292795
Epoch[44] Batch [1099]	Speed: 4790.318731 samples/sec	accuracy=0.741649	lr=0.292602
Epoch[44] Batch [1149]	Speed: 4422.864143 samples/sec	accuracy=0.741796	lr=0.292408
Epoch[44] Batch [1199]	Speed: 4519.849896 samples/sec	accuracy=0.741927	lr=0.292215
Epoch[44] Batch [1249]	Speed: 4744.360526 samples/sec	accuracy=0.741967	lr=0.292021
[Epoch 44] training: accuracy=0.741975
[Epoch 44] speed: 4692 samples/sec	time cost: 283.588538
[Epoch 44] validation: err-top1=0.394471 err-top5=0.162428
Epoch[45] Batch [49]	Speed: 4697.043008 samples/sec	accuracy=0.740721	lr=0.291823
Epoch[45] Batch [99]	Speed: 4720.615157 samples/sec	accuracy=0.741111	lr=0.291629
Epoch[45] Batch [149]	Speed: 4517.862075 samples/sec	accuracy=0.741403	lr=0.291435
Epoch[45] Batch [199]	Speed: 4984.262025 samples/sec	accuracy=0.741272	lr=0.291240
Epoch[45] Batch [249]	Speed: 4626.274323 samples/sec	accuracy=0.741157	lr=0.291046
Epoch[45] Batch [299]	Speed: 4707.262174 samples/sec	accuracy=0.741556	lr=0.290852
Epoch[45] Batch [349]	Speed: 4791.735954 samples/sec	accuracy=0.741627	lr=0.290657
Epoch[45] Batch [399]	Speed: 4348.374818 samples/sec	accuracy=0.741747	lr=0.290462
Epoch[45] Batch [449]	Speed: 4912.598428 samples/sec	accuracy=0.741973	lr=0.290267
Epoch[45] Batch [499]	Speed: 5132.463125 samples/sec	accuracy=0.742142	lr=0.290073
Epoch[45] Batch [549]	Speed: 4570.274508 samples/sec	accuracy=0.742248	lr=0.289878
Epoch[45] Batch [599]	Speed: 4679.073160 samples/sec	accuracy=0.742431	lr=0.289682
Epoch[45] Batch [649]	Speed: 4491.838749 samples/sec	accuracy=0.742560	lr=0.289487
Epoch[45] Batch [699]	Speed: 4880.342961 samples/sec	accuracy=0.742684	lr=0.289292
Epoch[45] Batch [749]	Speed: 4856.264551 samples/sec	accuracy=0.742716	lr=0.289096
Epoch[45] Batch [799]	Speed: 4655.818158 samples/sec	accuracy=0.742761	lr=0.288901
Epoch[45] Batch [849]	Speed: 4666.855272 samples/sec	accuracy=0.742884	lr=0.288705
Epoch[45] Batch [899]	Speed: 4727.101313 samples/sec	accuracy=0.742981	lr=0.288509
Epoch[45] Batch [949]	Speed: 4623.523275 samples/sec	accuracy=0.743100	lr=0.288314
Epoch[45] Batch [999]	Speed: 4837.081023 samples/sec	accuracy=0.743194	lr=0.288118
Epoch[45] Batch [1049]	Speed: 4659.546019 samples/sec	accuracy=0.743291	lr=0.287921
Epoch[45] Batch [1099]	Speed: 4857.856010 samples/sec	accuracy=0.743362	lr=0.287725
Epoch[45] Batch [1149]	Speed: 4525.848764 samples/sec	accuracy=0.743527	lr=0.287529
Epoch[45] Batch [1199]	Speed: 4232.062845 samples/sec	accuracy=0.743568	lr=0.287333
Epoch[45] Batch [1249]	Speed: 4984.776986 samples/sec	accuracy=0.743731	lr=0.287136
[Epoch 45] training: accuracy=0.743730
[Epoch 45] speed: 4694 samples/sec	time cost: 283.606950
[Epoch 45] validation: err-top1=0.408861 err-top5=0.172294
Epoch[46] Batch [49]	Speed: 4895.392469 samples/sec	accuracy=0.742028	lr=0.286936
Epoch[46] Batch [99]	Speed: 4843.463805 samples/sec	accuracy=0.741972	lr=0.286739
Epoch[46] Batch [149]	Speed: 4837.006065 samples/sec	accuracy=0.741829	lr=0.286542
Epoch[46] Batch [199]	Speed: 4898.989030 samples/sec	accuracy=0.741613	lr=0.286345
Epoch[46] Batch [249]	Speed: 4752.506284 samples/sec	accuracy=0.741759	lr=0.286148
Epoch[46] Batch [299]	Speed: 4816.711890 samples/sec	accuracy=0.741759	lr=0.285951
Epoch[46] Batch [349]	Speed: 4613.368619 samples/sec	accuracy=0.742174	lr=0.285754
Epoch[46] Batch [399]	Speed: 4227.376367 samples/sec	accuracy=0.742342	lr=0.285556
Epoch[46] Batch [449]	Speed: 4688.566448 samples/sec	accuracy=0.742589	lr=0.285359
Epoch[46] Batch [499]	Speed: 4219.936902 samples/sec	accuracy=0.742699	lr=0.285161
Epoch[46] Batch [549]	Speed: 4588.078724 samples/sec	accuracy=0.742814	lr=0.284964
Epoch[46] Batch [599]	Speed: 4812.864213 samples/sec	accuracy=0.742947	lr=0.284766
Epoch[46] Batch [649]	Speed: 4910.561132 samples/sec	accuracy=0.743183	lr=0.284568
Epoch[46] Batch [699]	Speed: 4836.226661 samples/sec	accuracy=0.743455	lr=0.284370
Epoch[46] Batch [749]	Speed: 5077.439877 samples/sec	accuracy=0.743612	lr=0.284172
Epoch[46] Batch [799]	Speed: 4644.621158 samples/sec	accuracy=0.743798	lr=0.283974
Epoch[46] Batch [849]	Speed: 4676.707574 samples/sec	accuracy=0.743917	lr=0.283776
Epoch[46] Batch [899]	Speed: 4398.226482 samples/sec	accuracy=0.744036	lr=0.283577
Epoch[46] Batch [949]	Speed: 4740.277706 samples/sec	accuracy=0.744172	lr=0.283379
Epoch[46] Batch [999]	Speed: 4880.302922 samples/sec	accuracy=0.744260	lr=0.283181
Epoch[46] Batch [1049]	Speed: 4664.708614 samples/sec	accuracy=0.744310	lr=0.282982
Epoch[46] Batch [1099]	Speed: 4433.147987 samples/sec	accuracy=0.744468	lr=0.282783
Epoch[46] Batch [1149]	Speed: 4788.449267 samples/sec	accuracy=0.744594	lr=0.282584
Epoch[46] Batch [1199]	Speed: 4654.950037 samples/sec	accuracy=0.744656	lr=0.282385
Epoch[46] Batch [1249]	Speed: 4786.312633 samples/sec	accuracy=0.744770	lr=0.282186
[Epoch 46] training: accuracy=0.744774
[Epoch 46] speed: 4691 samples/sec	time cost: 283.305375
[Epoch 46] validation: err-top1=0.413900 err-top5=0.173238
Epoch[47] Batch [49]	Speed: 4705.090046 samples/sec	accuracy=0.744386	lr=0.281983
Epoch[47] Batch [99]	Speed: 4541.159347 samples/sec	accuracy=0.744485	lr=0.281784
Epoch[47] Batch [149]	Speed: 4267.810450 samples/sec	accuracy=0.744209	lr=0.281585
Epoch[47] Batch [199]	Speed: 4776.304332 samples/sec	accuracy=0.744183	lr=0.281385
Epoch[47] Batch [249]	Speed: 4560.746243 samples/sec	accuracy=0.744497	lr=0.281186
Epoch[47] Batch [299]	Speed: 5059.128048 samples/sec	accuracy=0.744403	lr=0.280986
Epoch[47] Batch [349]	Speed: 5049.461107 samples/sec	accuracy=0.744512	lr=0.280786
Epoch[47] Batch [399]	Speed: 4614.540268 samples/sec	accuracy=0.744478	lr=0.280587
Epoch[47] Batch [449]	Speed: 4200.065555 samples/sec	accuracy=0.744588	lr=0.280387
Epoch[47] Batch [499]	Speed: 4642.646353 samples/sec	accuracy=0.744725	lr=0.280187
Epoch[47] Batch [549]	Speed: 4713.879490 samples/sec	accuracy=0.744788	lr=0.279987
Epoch[47] Batch [599]	Speed: 4665.026494 samples/sec	accuracy=0.744904	lr=0.279786
Epoch[47] Batch [649]	Speed: 5192.165875 samples/sec	accuracy=0.744901	lr=0.279586
Epoch[47] Batch [699]	Speed: 4709.494558 samples/sec	accuracy=0.744965	lr=0.279386
Epoch[47] Batch [749]	Speed: 4755.623633 samples/sec	accuracy=0.745034	lr=0.279185
Epoch[47] Batch [799]	Speed: 4802.738774 samples/sec	accuracy=0.745199	lr=0.278985
Epoch[47] Batch [849]	Speed: 4877.803445 samples/sec	accuracy=0.745294	lr=0.278784
Epoch[47] Batch [899]	Speed: 4729.046472 samples/sec	accuracy=0.745475	lr=0.278583
Epoch[47] Batch [949]	Speed: 4549.797544 samples/sec	accuracy=0.745619	lr=0.278382
Epoch[47] Batch [999]	Speed: 4844.169599 samples/sec	accuracy=0.745682	lr=0.278181
Epoch[47] Batch [1049]	Speed: 4629.963789 samples/sec	accuracy=0.745774	lr=0.277980
Epoch[47] Batch [1099]	Speed: 5158.569132 samples/sec	accuracy=0.745913	lr=0.277779
Epoch[47] Batch [1149]	Speed: 4776.740560 samples/sec	accuracy=0.746033	lr=0.277578
Epoch[47] Batch [1199]	Speed: 4762.186457 samples/sec	accuracy=0.746172	lr=0.277377
Epoch[47] Batch [1249]	Speed: 4553.761670 samples/sec	accuracy=0.746255	lr=0.277175
[Epoch 47] training: accuracy=0.746256
[Epoch 47] speed: 4707 samples/sec	time cost: 283.363966
[Epoch 47] validation: err-top1=0.420460 err-top5=0.179668
Epoch[48] Batch [49]	Speed: 4465.343551 samples/sec	accuracy=0.747296	lr=0.276970
Epoch[48] Batch [99]	Speed: 4691.941076 samples/sec	accuracy=0.746863	lr=0.276768
Epoch[48] Batch [149]	Speed: 4982.365651 samples/sec	accuracy=0.746693	lr=0.276566
Epoch[48] Batch [199]	Speed: 5096.201226 samples/sec	accuracy=0.746301	lr=0.276365
Epoch[48] Batch [249]	Speed: 4223.770741 samples/sec	accuracy=0.746010	lr=0.276163
Epoch[48] Batch [299]	Speed: 4758.613684 samples/sec	accuracy=0.745893	lr=0.275961
Epoch[48] Batch [349]	Speed: 5081.231223 samples/sec	accuracy=0.745921	lr=0.275759
Epoch[48] Batch [399]	Speed: 4731.690500 samples/sec	accuracy=0.746067	lr=0.275557
Epoch[48] Batch [449]	Speed: 4874.696970 samples/sec	accuracy=0.746166	lr=0.275354
Epoch[48] Batch [499]	Speed: 4690.496407 samples/sec	accuracy=0.746194	lr=0.275152
Epoch[48] Batch [549]	Speed: 4255.861377 samples/sec	accuracy=0.746237	lr=0.274950
Epoch[48] Batch [599]	Speed: 4514.890122 samples/sec	accuracy=0.746398	lr=0.274747
Epoch[48] Batch [649]	Speed: 4679.067655 samples/sec	accuracy=0.746378	lr=0.274545
Epoch[48] Batch [699]	Speed: 4476.272165 samples/sec	accuracy=0.746369	lr=0.274342
Epoch[48] Batch [749]	Speed: 4304.156239 samples/sec	accuracy=0.746472	lr=0.274139
Epoch[48] Batch [799]	Speed: 4620.073059 samples/sec	accuracy=0.746597	lr=0.273936
Epoch[48] Batch [849]	Speed: 5013.989775 samples/sec	accuracy=0.746748	lr=0.273733
Epoch[48] Batch [899]	Speed: 4548.495620 samples/sec	accuracy=0.746851	lr=0.273530
Epoch[48] Batch [949]	Speed: 4959.692802 samples/sec	accuracy=0.747000	lr=0.273327
Epoch[48] Batch [999]	Speed: 4632.661246 samples/sec	accuracy=0.747192	lr=0.273124
Epoch[48] Batch [1049]	Speed: 4630.984292 samples/sec	accuracy=0.747259	lr=0.272921
Epoch[48] Batch [1099]	Speed: 4857.493290 samples/sec	accuracy=0.747282	lr=0.272717
Epoch[48] Batch [1149]	Speed: 4440.993376 samples/sec	accuracy=0.747388	lr=0.272514
Epoch[48] Batch [1199]	Speed: 4836.900496 samples/sec	accuracy=0.747472	lr=0.272310
Epoch[48] Batch [1249]	Speed: 4673.836930 samples/sec	accuracy=0.747473	lr=0.272107
[Epoch 48] training: accuracy=0.747476
[Epoch 48] speed: 4665 samples/sec	time cost: 284.362256
[Epoch 48] validation: err-top1=0.396883 err-top5=0.163803
Epoch[49] Batch [49]	Speed: 4328.375603 samples/sec	accuracy=0.746397	lr=0.271899
Epoch[49] Batch [99]	Speed: 4923.406905 samples/sec	accuracy=0.746530	lr=0.271695
Epoch[49] Batch [149]	Speed: 4717.229739 samples/sec	accuracy=0.746225	lr=0.271491
Epoch[49] Batch [199]	Speed: 4924.052754 samples/sec	accuracy=0.746339	lr=0.271287
Epoch[49] Batch [249]	Speed: 4885.878986 samples/sec	accuracy=0.746028	lr=0.271083
Epoch[49] Batch [299]	Speed: 4422.105209 samples/sec	accuracy=0.746373	lr=0.270879
Epoch[49] Batch [349]	Speed: 4429.449274 samples/sec	accuracy=0.746698	lr=0.270675
Epoch[49] Batch [399]	Speed: 4839.149716 samples/sec	accuracy=0.746800	lr=0.270470
Epoch[49] Batch [449]	Speed: 4540.917271 samples/sec	accuracy=0.746778	lr=0.270266
Epoch[49] Batch [499]	Speed: 4768.654537 samples/sec	accuracy=0.746799	lr=0.270061
Epoch[49] Batch [549]	Speed: 4752.102233 samples/sec	accuracy=0.747031	lr=0.269857
Epoch[49] Batch [599]	Speed: 4533.392555 samples/sec	accuracy=0.747072	lr=0.269652
Epoch[49] Batch [649]	Speed: 4556.797992 samples/sec	accuracy=0.747104	lr=0.269447
Epoch[49] Batch [699]	Speed: 4851.705116 samples/sec	accuracy=0.747203	lr=0.269243
Epoch[49] Batch [749]	Speed: 4796.129763 samples/sec	accuracy=0.747403	lr=0.269038
Epoch[49] Batch [799]	Speed: 4878.437827 samples/sec	accuracy=0.747533	lr=0.268833
Epoch[49] Batch [849]	Speed: 4458.513878 samples/sec	accuracy=0.747535	lr=0.268628
Epoch[49] Batch [899]	Speed: 4795.226950 samples/sec	accuracy=0.747552	lr=0.268422
Epoch[49] Batch [949]	Speed: 4910.444243 samples/sec	accuracy=0.747639	lr=0.268217
Epoch[49] Batch [999]	Speed: 4606.457816 samples/sec	accuracy=0.747738	lr=0.268012
Epoch[49] Batch [1049]	Speed: 4712.985653 samples/sec	accuracy=0.747779	lr=0.267807
Epoch[49] Batch [1099]	Speed: 4958.137641 samples/sec	accuracy=0.747868	lr=0.267601
Epoch[49] Batch [1149]	Speed: 5092.174872 samples/sec	accuracy=0.747976	lr=0.267396
Epoch[49] Batch [1199]	Speed: 5145.398195 samples/sec	accuracy=0.748120	lr=0.267190
Epoch[49] Batch [1249]	Speed: 4807.396216 samples/sec	accuracy=0.748223	lr=0.266984
[Epoch 49] training: accuracy=0.748225
[Epoch 49] speed: 4731 samples/sec	time cost: 280.951769
[Epoch 49] validation: err-top1=0.399454 err-top5=0.164680
Epoch[50] Batch [49]	Speed: 4899.913006 samples/sec	accuracy=0.749425	lr=0.266774
Epoch[50] Batch [99]	Speed: 4864.304902 samples/sec	accuracy=0.747577	lr=0.266568
Epoch[50] Batch [149]	Speed: 4945.062523 samples/sec	accuracy=0.747641	lr=0.266362
Epoch[50] Batch [199]	Speed: 4708.979967 samples/sec	accuracy=0.747851	lr=0.266156
Epoch[50] Batch [249]	Speed: 4631.793246 samples/sec	accuracy=0.748049	lr=0.265950
Epoch[50] Batch [299]	Speed: 4737.034457 samples/sec	accuracy=0.748089	lr=0.265744
Epoch[50] Batch [349]	Speed: 4492.294006 samples/sec	accuracy=0.748169	lr=0.265538
Epoch[50] Batch [399]	Speed: 4710.669668 samples/sec	accuracy=0.748017	lr=0.265331
Epoch[50] Batch [449]	Speed: 4795.278561 samples/sec	accuracy=0.748262	lr=0.265125
Epoch[50] Batch [499]	Speed: 4530.180903 samples/sec	accuracy=0.748441	lr=0.264918
Epoch[50] Batch [549]	Speed: 4899.593722 samples/sec	accuracy=0.748512	lr=0.264712
Epoch[50] Batch [599]	Speed: 4894.337342 samples/sec	accuracy=0.748494	lr=0.264505
Epoch[50] Batch [649]	Speed: 4627.297887 samples/sec	accuracy=0.748417	lr=0.264298
Epoch[50] Batch [699]	Speed: 4740.655155 samples/sec	accuracy=0.748516	lr=0.264092
Epoch[50] Batch [749]	Speed: 4701.200165 samples/sec	accuracy=0.748610	lr=0.263885
Epoch[50] Batch [799]	Speed: 4781.712011 samples/sec	accuracy=0.748637	lr=0.263678
Epoch[50] Batch [849]	Speed: 4823.762626 samples/sec	accuracy=0.748796	lr=0.263471
Epoch[50] Batch [899]	Speed: 5022.859384 samples/sec	accuracy=0.748932	lr=0.263264
Epoch[50] Batch [949]	Speed: 4759.836014 samples/sec	accuracy=0.749010	lr=0.263056
Epoch[50] Batch [999]	Speed: 4801.471339 samples/sec	accuracy=0.749196	lr=0.262849
Epoch[50] Batch [1049]	Speed: 4697.433845 samples/sec	accuracy=0.749323	lr=0.262642
Epoch[50] Batch [1099]	Speed: 4880.606275 samples/sec	accuracy=0.749435	lr=0.262434
Epoch[50] Batch [1149]	Speed: 4762.893057 samples/sec	accuracy=0.749577	lr=0.262227
Epoch[50] Batch [1199]	Speed: 4533.926534 samples/sec	accuracy=0.749697	lr=0.262019
Epoch[50] Batch [1249]	Speed: 4934.682456 samples/sec	accuracy=0.749760	lr=0.261812
[Epoch 50] training: accuracy=0.749770
[Epoch 50] speed: 4754 samples/sec	time cost: 279.821041
[Epoch 50] validation: err-top1=0.414820 err-top5=0.179488
Epoch[51] Batch [49]	Speed: 4741.847963 samples/sec	accuracy=0.748856	lr=0.261600
Epoch[51] Batch [99]	Speed: 4921.735779 samples/sec	accuracy=0.749327	lr=0.261392
Epoch[51] Batch [149]	Speed: 4629.391082 samples/sec	accuracy=0.749387	lr=0.261184
Epoch[51] Batch [199]	Speed: 4789.184723 samples/sec	accuracy=0.749440	lr=0.260976
Epoch[51] Batch [249]	Speed: 4993.135922 samples/sec	accuracy=0.749597	lr=0.260768
Epoch[51] Batch [299]	Speed: 4923.121007 samples/sec	accuracy=0.749544	lr=0.260560
Epoch[51] Batch [349]	Speed: 4502.762207 samples/sec	accuracy=0.749219	lr=0.260352
Epoch[51] Batch [399]	Speed: 4624.680966 samples/sec	accuracy=0.749268	lr=0.260144
Epoch[51] Batch [449]	Speed: 4982.238614 samples/sec	accuracy=0.749373	lr=0.259935
Epoch[51] Batch [499]	Speed: 4957.022113 samples/sec	accuracy=0.749471	lr=0.259727
Epoch[51] Batch [549]	Speed: 5040.034249 samples/sec	accuracy=0.749629	lr=0.259519
Epoch[51] Batch [599]	Speed: 5007.121470 samples/sec	accuracy=0.749756	lr=0.259310
Epoch[51] Batch [649]	Speed: 5080.268253 samples/sec	accuracy=0.749736	lr=0.259102
Epoch[51] Batch [699]	Speed: 5040.137515 samples/sec	accuracy=0.749798	lr=0.258893
Epoch[51] Batch [749]	Speed: 5057.959113 samples/sec	accuracy=0.749882	lr=0.258684
Epoch[51] Batch [799]	Speed: 4902.049006 samples/sec	accuracy=0.749882	lr=0.258475
Epoch[51] Batch [849]	Speed: 4601.629158 samples/sec	accuracy=0.749967	lr=0.258267
Epoch[51] Batch [899]	Speed: 4867.856243 samples/sec	accuracy=0.750117	lr=0.258058
Epoch[51] Batch [949]	Speed: 4698.305449 samples/sec	accuracy=0.750183	lr=0.257849
Epoch[51] Batch [999]	Speed: 4863.683002 samples/sec	accuracy=0.750278	lr=0.257640
Epoch[51] Batch [1049]	Speed: 4642.720026 samples/sec	accuracy=0.750383	lr=0.257430
Epoch[51] Batch [1099]	Speed: 4899.317512 samples/sec	accuracy=0.750516	lr=0.257221
Epoch[51] Batch [1149]	Speed: 4688.758902 samples/sec	accuracy=0.750605	lr=0.257012
Epoch[51] Batch [1199]	Speed: 4796.144652 samples/sec	accuracy=0.750792	lr=0.256803
Epoch[51] Batch [1249]	Speed: 4789.159197 samples/sec	accuracy=0.750946	lr=0.256593
[Epoch 51] training: accuracy=0.750951
[Epoch 51] speed: 4831 samples/sec	time cost: 275.927635
[Epoch 51] validation: err-top1=0.392120 err-top5=0.161432
Epoch[52] Batch [49]	Speed: 4859.895551 samples/sec	accuracy=0.750420	lr=0.256379
Epoch[52] Batch [99]	Speed: 4199.400366 samples/sec	accuracy=0.750651	lr=0.256170
Epoch[52] Batch [149]	Speed: 4720.997784 samples/sec	accuracy=0.750108	lr=0.255960
Epoch[52] Batch [199]	Speed: 4517.976799 samples/sec	accuracy=0.750205	lr=0.255751
Epoch[52] Batch [249]	Speed: 4559.185500 samples/sec	accuracy=0.750117	lr=0.255541
Epoch[52] Batch [299]	Speed: 4892.373007 samples/sec	accuracy=0.749992	lr=0.255331
Epoch[52] Batch [349]	Speed: 4862.682788 samples/sec	accuracy=0.750144	lr=0.255121
Epoch[52] Batch [399]	Speed: 4840.734231 samples/sec	accuracy=0.750482	lr=0.254911
Epoch[52] Batch [449]	Speed: 4800.030863 samples/sec	accuracy=0.750456	lr=0.254701
Epoch[52] Batch [499]	Speed: 4595.591990 samples/sec	accuracy=0.750369	lr=0.254491
Epoch[52] Batch [549]	Speed: 4776.245056 samples/sec	accuracy=0.750394	lr=0.254281
Epoch[52] Batch [599]	Speed: 4858.151963 samples/sec	accuracy=0.750542	lr=0.254071
Epoch[52] Batch [649]	Speed: 4555.732021 samples/sec	accuracy=0.750782	lr=0.253861
Epoch[52] Batch [699]	Speed: 4630.172226 samples/sec	accuracy=0.750903	lr=0.253650
Epoch[52] Batch [749]	Speed: 4827.979626 samples/sec	accuracy=0.751050	lr=0.253440
Epoch[52] Batch [799]	Speed: 4891.281747 samples/sec	accuracy=0.751141	lr=0.253229
Epoch[52] Batch [849]	Speed: 4876.359997 samples/sec	accuracy=0.751179	lr=0.253019
Epoch[52] Batch [899]	Speed: 4729.453173 samples/sec	accuracy=0.751243	lr=0.252808
Epoch[52] Batch [949]	Speed: 4811.129303 samples/sec	accuracy=0.751249	lr=0.252598
Epoch[52] Batch [999]	Speed: 4652.901315 samples/sec	accuracy=0.751355	lr=0.252387
Epoch[52] Batch [1049]	Speed: 4633.303837 samples/sec	accuracy=0.751429	lr=0.252176
Epoch[52] Batch [1099]	Speed: 4482.376535 samples/sec	accuracy=0.751487	lr=0.251965
Epoch[52] Batch [1149]	Speed: 5065.533878 samples/sec	accuracy=0.751549	lr=0.251754
Epoch[52] Batch [1199]	Speed: 4656.840903 samples/sec	accuracy=0.751696	lr=0.251543
Epoch[52] Batch [1249]	Speed: 5021.781481 samples/sec	accuracy=0.751900	lr=0.251332
[Epoch 52] training: accuracy=0.751907
[Epoch 52] speed: 4717 samples/sec	time cost: 281.533409
[Epoch 52] validation: err-top1=0.383911 err-top5=0.155945
Epoch[53] Batch [49]	Speed: 4821.482760 samples/sec	accuracy=0.752057	lr=0.251117
Epoch[53] Batch [99]	Speed: 4959.835416 samples/sec	accuracy=0.752414	lr=0.250906
Epoch[53] Batch [149]	Speed: 4722.073249 samples/sec	accuracy=0.751894	lr=0.250695
Epoch[53] Batch [199]	Speed: 4755.581297 samples/sec	accuracy=0.751857	lr=0.250483
Epoch[53] Batch [249]	Speed: 4931.447530 samples/sec	accuracy=0.751579	lr=0.250272
Epoch[53] Batch [299]	Speed: 4704.502520 samples/sec	accuracy=0.751791	lr=0.250061
Epoch[53] Batch [349]	Speed: 4783.822586 samples/sec	accuracy=0.751776	lr=0.249849
Epoch[53] Batch [399]	Speed: 4707.267230 samples/sec	accuracy=0.751770	lr=0.249638
Epoch[53] Batch [449]	Speed: 4809.823178 samples/sec	accuracy=0.751836	lr=0.249426
Epoch[53] Batch [499]	Speed: 4810.276864 samples/sec	accuracy=0.752012	lr=0.249215
Epoch[53] Batch [549]	Speed: 4956.728293 samples/sec	accuracy=0.751985	lr=0.249003
Epoch[53] Batch [599]	Speed: 4981.757808 samples/sec	accuracy=0.752099	lr=0.248791
Epoch[53] Batch [649]	Speed: 4675.003674 samples/sec	accuracy=0.752206	lr=0.248579
Epoch[53] Batch [699]	Speed: 4672.221321 samples/sec	accuracy=0.752449	lr=0.248367
Epoch[53] Batch [749]	Speed: 4846.920224 samples/sec	accuracy=0.752652	lr=0.248156
Epoch[53] Batch [799]	Speed: 5018.394098 samples/sec	accuracy=0.752727	lr=0.247944
Epoch[53] Batch [849]	Speed: 4986.530108 samples/sec	accuracy=0.752833	lr=0.247732
Epoch[53] Batch [899]	Speed: 4746.262642 samples/sec	accuracy=0.752956	lr=0.247519
Epoch[53] Batch [949]	Speed: 4812.446707 samples/sec	accuracy=0.752985	lr=0.247307
Epoch[53] Batch [999]	Speed: 4897.316120 samples/sec	accuracy=0.753069	lr=0.247095
Epoch[53] Batch [1049]	Speed: 4904.280724 samples/sec	accuracy=0.753162	lr=0.246883
Epoch[53] Batch [1099]	Speed: 4739.441714 samples/sec	accuracy=0.753264	lr=0.246670
Epoch[53] Batch [1149]	Speed: 4763.920694 samples/sec	accuracy=0.753438	lr=0.246458
Epoch[53] Batch [1199]	Speed: 4792.178962 samples/sec	accuracy=0.753548	lr=0.246246
Epoch[53] Batch [1249]	Speed: 5000.975863 samples/sec	accuracy=0.753649	lr=0.246033
[Epoch 53] training: accuracy=0.753654
[Epoch 53] speed: 4825 samples/sec	time cost: 275.826694
[Epoch 53] validation: err-top1=0.381138 err-top5=0.151507
Epoch[54] Batch [49]	Speed: 4749.359933 samples/sec	accuracy=0.751629	lr=0.245816
Epoch[54] Batch [99]	Speed: 4904.936128 samples/sec	accuracy=0.751680	lr=0.245604
Epoch[54] Batch [149]	Speed: 4249.773505 samples/sec	accuracy=0.751963	lr=0.245391
Epoch[54] Batch [199]	Speed: 4691.396800 samples/sec	accuracy=0.752249	lr=0.245178
Epoch[54] Batch [249]	Speed: 4827.625043 samples/sec	accuracy=0.752472	lr=0.244966
Epoch[54] Batch [299]	Speed: 4875.629073 samples/sec	accuracy=0.752790	lr=0.244753
Epoch[54] Batch [349]	Speed: 4863.559082 samples/sec	accuracy=0.752834	lr=0.244540
Epoch[54] Batch [399]	Speed: 4780.089284 samples/sec	accuracy=0.752906	lr=0.244327
Epoch[54] Batch [449]	Speed: 4667.831329 samples/sec	accuracy=0.753103	lr=0.244114
Epoch[54] Batch [499]	Speed: 4835.115123 samples/sec	accuracy=0.753064	lr=0.243901
Epoch[54] Batch [549]	Speed: 4924.921377 samples/sec	accuracy=0.753137	lr=0.243688
Epoch[54] Batch [599]	Speed: 4982.279765 samples/sec	accuracy=0.753396	lr=0.243475
Epoch[54] Batch [649]	Speed: 4790.831161 samples/sec	accuracy=0.753549	lr=0.243262
Epoch[54] Batch [699]	Speed: 4911.141504 samples/sec	accuracy=0.753727	lr=0.243049
Epoch[54] Batch [749]	Speed: 4898.839501 samples/sec	accuracy=0.753892	lr=0.242835
Epoch[54] Batch [799]	Speed: 4601.657359 samples/sec	accuracy=0.753988	lr=0.242622
Epoch[54] Batch [849]	Speed: 4825.451036 samples/sec	accuracy=0.754102	lr=0.242409
Epoch[54] Batch [899]	Speed: 4780.607083 samples/sec	accuracy=0.754217	lr=0.242195
Epoch[54] Batch [949]	Speed: 4830.518583 samples/sec	accuracy=0.754314	lr=0.241982
Epoch[54] Batch [999]	Speed: 5003.596904 samples/sec	accuracy=0.754336	lr=0.241768
Epoch[54] Batch [1049]	Speed: 4825.181388 samples/sec	accuracy=0.754404	lr=0.241555
Epoch[54] Batch [1099]	Speed: 4840.110489 samples/sec	accuracy=0.754577	lr=0.241341
Epoch[54] Batch [1149]	Speed: 4743.822465 samples/sec	accuracy=0.754733	lr=0.241127
Epoch[54] Batch [1199]	Speed: 4849.496997 samples/sec	accuracy=0.754768	lr=0.240914
Epoch[54] Batch [1249]	Speed: 5042.175088 samples/sec	accuracy=0.754826	lr=0.240700
[Epoch 54] training: accuracy=0.754822
[Epoch 54] speed: 4802 samples/sec	time cost: 278.231059
[Epoch 54] validation: err-top1=0.380122 err-top5=0.154895
Epoch[55] Batch [49]	Speed: 4706.497405 samples/sec	accuracy=0.755488	lr=0.240482
Epoch[55] Batch [99]	Speed: 4769.551078 samples/sec	accuracy=0.754393	lr=0.240268
Epoch[55] Batch [149]	Speed: 4728.930671 samples/sec	accuracy=0.754014	lr=0.240054
Epoch[55] Batch [199]	Speed: 4691.214480 samples/sec	accuracy=0.754129	lr=0.239840
Epoch[55] Batch [249]	Speed: 4663.182137 samples/sec	accuracy=0.754412	lr=0.239626
Epoch[55] Batch [299]	Speed: 4639.780770 samples/sec	accuracy=0.754431	lr=0.239412
Epoch[55] Batch [349]	Speed: 4960.502430 samples/sec	accuracy=0.754646	lr=0.239198
Epoch[55] Batch [399]	Speed: 4768.747406 samples/sec	accuracy=0.754419	lr=0.238984
Epoch[55] Batch [449]	Speed: 4830.422858 samples/sec	accuracy=0.754461	lr=0.238769
Epoch[55] Batch [499]	Speed: 4674.733482 samples/sec	accuracy=0.754579	lr=0.238555
Epoch[55] Batch [549]	Speed: 4856.184934 samples/sec	accuracy=0.754673	lr=0.238341
Epoch[55] Batch [599]	Speed: 4839.938898 samples/sec	accuracy=0.754727	lr=0.238126
Epoch[55] Batch [649]	Speed: 4815.592890 samples/sec	accuracy=0.754824	lr=0.237912
Epoch[55] Batch [699]	Speed: 4710.408355 samples/sec	accuracy=0.754873	lr=0.237698
Epoch[55] Batch [749]	Speed: 4730.976034 samples/sec	accuracy=0.754800	lr=0.237483
Epoch[55] Batch [799]	Speed: 5020.225522 samples/sec	accuracy=0.755029	lr=0.237269
Epoch[55] Batch [849]	Speed: 5029.370808 samples/sec	accuracy=0.755188	lr=0.237054
Epoch[55] Batch [899]	Speed: 4947.950843 samples/sec	accuracy=0.755257	lr=0.236839
Epoch[55] Batch [949]	Speed: 4850.598285 samples/sec	accuracy=0.755388	lr=0.236625
Epoch[55] Batch [999]	Speed: 4880.072244 samples/sec	accuracy=0.755470	lr=0.236410
Epoch[55] Batch [1049]	Speed: 4909.672087 samples/sec	accuracy=0.755621	lr=0.236195
Epoch[55] Batch [1099]	Speed: 4892.130933 samples/sec	accuracy=0.755755	lr=0.235981
Epoch[55] Batch [1149]	Speed: 4767.026902 samples/sec	accuracy=0.755871	lr=0.235766
Epoch[55] Batch [1199]	Speed: 4741.387831 samples/sec	accuracy=0.756028	lr=0.235551
Epoch[55] Batch [1249]	Speed: 4973.382182 samples/sec	accuracy=0.756086	lr=0.235336
[Epoch 55] training: accuracy=0.756082
[Epoch 55] speed: 4808 samples/sec	time cost: 276.255259
[Epoch 55] validation: err-top1=0.415856 err-top5=0.181342
Epoch[56] Batch [49]	Speed: 4517.213096 samples/sec	accuracy=0.755901	lr=0.235117
Epoch[56] Batch [99]	Speed: 4656.493139 samples/sec	accuracy=0.755305	lr=0.234902
Epoch[56] Batch [149]	Speed: 4880.908667 samples/sec	accuracy=0.755756	lr=0.234687
Epoch[56] Batch [199]	Speed: 4863.472287 samples/sec	accuracy=0.755738	lr=0.234472
Epoch[56] Batch [249]	Speed: 4704.406675 samples/sec	accuracy=0.755805	lr=0.234256
Epoch[56] Batch [299]	Speed: 4966.354911 samples/sec	accuracy=0.756079	lr=0.234041
Epoch[56] Batch [349]	Speed: 4754.498411 samples/sec	accuracy=0.756001	lr=0.233826
Epoch[56] Batch [399]	Speed: 5134.537866 samples/sec	accuracy=0.755904	lr=0.233611
Epoch[56] Batch [449]	Speed: 4736.051182 samples/sec	accuracy=0.755861	lr=0.233396
Epoch[56] Batch [499]	Speed: 5148.546091 samples/sec	accuracy=0.755895	lr=0.233180
Epoch[56] Batch [549]	Speed: 5143.538383 samples/sec	accuracy=0.756031	lr=0.232965
Epoch[56] Batch [599]	Speed: 4774.363737 samples/sec	accuracy=0.756169	lr=0.232749
Epoch[56] Batch [649]	Speed: 4078.923057 samples/sec	accuracy=0.756145	lr=0.232534
Epoch[56] Batch [699]	Speed: 4367.437319 samples/sec	accuracy=0.756151	lr=0.232319
Epoch[56] Batch [749]	Speed: 4732.474638 samples/sec	accuracy=0.756209	lr=0.232103
Epoch[56] Batch [799]	Speed: 4960.796698 samples/sec	accuracy=0.756346	lr=0.231887
Epoch[56] Batch [849]	Speed: 4852.392043 samples/sec	accuracy=0.756447	lr=0.231672
Epoch[56] Batch [899]	Speed: 4505.014971 samples/sec	accuracy=0.756495	lr=0.231456
Epoch[56] Batch [949]	Speed: 4782.598132 samples/sec	accuracy=0.756496	lr=0.231241
Epoch[56] Batch [999]	Speed: 4579.020319 samples/sec	accuracy=0.756646	lr=0.231025
Epoch[56] Batch [1049]	Speed: 4711.832338 samples/sec	accuracy=0.756732	lr=0.230809
Epoch[56] Batch [1099]	Speed: 3975.353487 samples/sec	accuracy=0.756887	lr=0.230593
Epoch[56] Batch [1149]	Speed: 4782.687497 samples/sec	accuracy=0.756975	lr=0.230377
Epoch[56] Batch [1199]	Speed: 4625.712991 samples/sec	accuracy=0.757073	lr=0.230162
Epoch[56] Batch [1249]	Speed: 4472.544952 samples/sec	accuracy=0.757088	lr=0.229946
[Epoch 56] training: accuracy=0.757088
[Epoch 56] speed: 4685 samples/sec	time cost: 283.430286
[Epoch 56] validation: err-top1=0.411133 err-top5=0.173828
Epoch[57] Batch [49]	Speed: 4282.792516 samples/sec	accuracy=0.758518	lr=0.229725
Epoch[57] Batch [99]	Speed: 4830.494461 samples/sec	accuracy=0.756456	lr=0.229510
Epoch[57] Batch [149]	Speed: 4480.174951 samples/sec	accuracy=0.756461	lr=0.229294
Epoch[57] Batch [199]	Speed: 4464.592060 samples/sec	accuracy=0.756082	lr=0.229077
Epoch[57] Batch [249]	Speed: 4774.597481 samples/sec	accuracy=0.755928	lr=0.228861
Epoch[57] Batch [299]	Speed: 4695.535339 samples/sec	accuracy=0.756211	lr=0.228645
Epoch[57] Batch [349]	Speed: 4402.339836 samples/sec	accuracy=0.756329	lr=0.228429
Epoch[57] Batch [399]	Speed: 4947.773458 samples/sec	accuracy=0.756319	lr=0.228213
Epoch[57] Batch [449]	Speed: 4332.424223 samples/sec	accuracy=0.756444	lr=0.227997
Epoch[57] Batch [499]	Speed: 4435.834953 samples/sec	accuracy=0.756553	lr=0.227781
Epoch[57] Batch [549]	Speed: 4207.336554 samples/sec	accuracy=0.756687	lr=0.227564
Epoch[57] Batch [599]	Speed: 4874.868268 samples/sec	accuracy=0.756771	lr=0.227348
Epoch[57] Batch [649]	Speed: 4631.766873 samples/sec	accuracy=0.756981	lr=0.227132
Epoch[57] Batch [699]	Speed: 4751.891401 samples/sec	accuracy=0.756883	lr=0.226915
Epoch[57] Batch [749]	Speed: 4959.972883 samples/sec	accuracy=0.756915	lr=0.226699
Epoch[57] Batch [799]	Speed: 4975.787248 samples/sec	accuracy=0.757048	lr=0.226482
Epoch[57] Batch [849]	Speed: 4890.001673 samples/sec	accuracy=0.757084	lr=0.226266
Epoch[57] Batch [899]	Speed: 4875.319808 samples/sec	accuracy=0.757285	lr=0.226049
Epoch[57] Batch [949]	Speed: 4850.280685 samples/sec	accuracy=0.757524	lr=0.225833
Epoch[57] Batch [999]	Speed: 4980.133460 samples/sec	accuracy=0.757642	lr=0.225616
Epoch[57] Batch [1049]	Speed: 4838.555708 samples/sec	accuracy=0.757819	lr=0.225400
Epoch[57] Batch [1099]	Speed: 5107.000132 samples/sec	accuracy=0.757974	lr=0.225183
Epoch[57] Batch [1149]	Speed: 4313.019006 samples/sec	accuracy=0.758094	lr=0.224967
Epoch[57] Batch [1199]	Speed: 4835.333296 samples/sec	accuracy=0.758200	lr=0.224750
Epoch[57] Batch [1249]	Speed: 4825.286880 samples/sec	accuracy=0.758266	lr=0.224533
[Epoch 57] training: accuracy=0.758266
[Epoch 57] speed: 4683 samples/sec	time cost: 283.627301
[Epoch 57] validation: err-top1=0.411571 err-top5=0.177894
Epoch[58] Batch [49]	Speed: 4701.030975 samples/sec	accuracy=0.758584	lr=0.224312
Epoch[58] Batch [99]	Speed: 4587.534952 samples/sec	accuracy=0.757877	lr=0.224095
Epoch[58] Batch [149]	Speed: 4390.734215 samples/sec	accuracy=0.757289	lr=0.223879
Epoch[58] Batch [199]	Speed: 4793.908362 samples/sec	accuracy=0.757529	lr=0.223662
Epoch[58] Batch [249]	Speed: 4907.718979 samples/sec	accuracy=0.757552	lr=0.223445
Epoch[58] Batch [299]	Speed: 4644.927063 samples/sec	accuracy=0.757680	lr=0.223228
Epoch[58] Batch [349]	Speed: 4958.671606 samples/sec	accuracy=0.757903	lr=0.223011
Epoch[58] Batch [399]	Speed: 4982.929012 samples/sec	accuracy=0.757882	lr=0.222794
Epoch[58] Batch [449]	Speed: 4813.926262 samples/sec	accuracy=0.757956	lr=0.222577
Epoch[58] Batch [499]	Speed: 4372.862050 samples/sec	accuracy=0.758022	lr=0.222360
Epoch[58] Batch [549]	Speed: 4756.014589 samples/sec	accuracy=0.758142	lr=0.222143
Epoch[58] Batch [599]	Speed: 4927.804302 samples/sec	accuracy=0.758307	lr=0.221926
Epoch[58] Batch [649]	Speed: 4597.880343 samples/sec	accuracy=0.758469	lr=0.221709
Epoch[58] Batch [699]	Speed: 4693.563891 samples/sec	accuracy=0.758586	lr=0.221492
Epoch[58] Batch [749]	Speed: 4994.721366 samples/sec	accuracy=0.758717	lr=0.221275
Epoch[58] Batch [799]	Speed: 5079.509771 samples/sec	accuracy=0.758840	lr=0.221058
Epoch[58] Batch [849]	Speed: 5014.371328 samples/sec	accuracy=0.758980	lr=0.220841
Epoch[58] Batch [899]	Speed: 4907.038948 samples/sec	accuracy=0.759110	lr=0.220623
Epoch[58] Batch [949]	Speed: 4431.713580 samples/sec	accuracy=0.759264	lr=0.220406
Epoch[58] Batch [999]	Speed: 4778.798574 samples/sec	accuracy=0.759372	lr=0.220189
Epoch[58] Batch [1049]	Speed: 4754.777482 samples/sec	accuracy=0.759493	lr=0.219972
Epoch[58] Batch [1099]	Speed: 4455.780052 samples/sec	accuracy=0.759568	lr=0.219754
Epoch[58] Batch [1149]	Speed: 4881.853357 samples/sec	accuracy=0.759697	lr=0.219537
Epoch[58] Batch [1199]	Speed: 4954.348927 samples/sec	accuracy=0.759793	lr=0.219320
Epoch[58] Batch [1249]	Speed: 4668.890825 samples/sec	accuracy=0.759879	lr=0.219102
[Epoch 58] training: accuracy=0.759880
[Epoch 58] speed: 4747 samples/sec	time cost: 279.947957
[Epoch 58] validation: err-top1=0.385376 err-top5=0.152140
Epoch[59] Batch [49]	Speed: 4536.288595 samples/sec	accuracy=0.758792	lr=0.218881
Epoch[59] Batch [99]	Speed: 4915.774680 samples/sec	accuracy=0.759469	lr=0.218663
Epoch[59] Batch [149]	Speed: 4692.074756 samples/sec	accuracy=0.759518	lr=0.218446
Epoch[59] Batch [199]	Speed: 4433.251585 samples/sec	accuracy=0.759681	lr=0.218228
Epoch[59] Batch [249]	Speed: 4671.886604 samples/sec	accuracy=0.759773	lr=0.218011
Epoch[59] Batch [299]	Speed: 4956.322403 samples/sec	accuracy=0.759694	lr=0.217793
Epoch[59] Batch [349]	Speed: 4807.283864 samples/sec	accuracy=0.759938	lr=0.217576
Epoch[59] Batch [399]	Speed: 5004.694306 samples/sec	accuracy=0.759797	lr=0.217358
Epoch[59] Batch [449]	Speed: 5080.800600 samples/sec	accuracy=0.759844	lr=0.217141
Epoch[59] Batch [499]	Speed: 4713.211666 samples/sec	accuracy=0.759822	lr=0.216923
Epoch[59] Batch [549]	Speed: 4921.872045 samples/sec	accuracy=0.759822	lr=0.216706
Epoch[59] Batch [599]	Speed: 4817.373598 samples/sec	accuracy=0.759789	lr=0.216488
Epoch[59] Batch [649]	Speed: 5010.888932 samples/sec	accuracy=0.759805	lr=0.216270
Epoch[59] Batch [699]	Speed: 4814.819398 samples/sec	accuracy=0.759912	lr=0.216053
Epoch[59] Batch [749]	Speed: 4693.530347 samples/sec	accuracy=0.759963	lr=0.215835
Epoch[59] Batch [799]	Speed: 4819.375604 samples/sec	accuracy=0.759989	lr=0.215617
Epoch[59] Batch [849]	Speed: 4871.621305 samples/sec	accuracy=0.760086	lr=0.215400
Epoch[59] Batch [899]	Speed: 4803.402020 samples/sec	accuracy=0.760141	lr=0.215182
Epoch[59] Batch [949]	Speed: 4656.421150 samples/sec	accuracy=0.760159	lr=0.214964
Epoch[59] Batch [999]	Speed: 4751.490188 samples/sec	accuracy=0.760316	lr=0.214746
Epoch[59] Batch [1049]	Speed: 4946.259143 samples/sec	accuracy=0.760434	lr=0.214529
Epoch[59] Batch [1099]	Speed: 4792.109880 samples/sec	accuracy=0.760571	lr=0.214311
Epoch[59] Batch [1149]	Speed: 5186.813409 samples/sec	accuracy=0.760698	lr=0.214093
Epoch[59] Batch [1199]	Speed: 4840.064672 samples/sec	accuracy=0.760768	lr=0.213875
Epoch[59] Batch [1249]	Speed: 4970.288438 samples/sec	accuracy=0.760838	lr=0.213657
[Epoch 59] training: accuracy=0.760837
[Epoch 59] speed: 4818 samples/sec	time cost: 276.438373
[Epoch 59] validation: err-top1=0.382932 err-top5=0.154197
Epoch[60] Batch [49]	Speed: 4720.351910 samples/sec	accuracy=0.760956	lr=0.213435
Epoch[60] Batch [99]	Speed: 4919.098139 samples/sec	accuracy=0.760546	lr=0.213217
Epoch[60] Batch [149]	Speed: 5169.324152 samples/sec	accuracy=0.760000	lr=0.212999
Epoch[60] Batch [199]	Speed: 4724.804599 samples/sec	accuracy=0.759705	lr=0.212781
Epoch[60] Batch [249]	Speed: 5076.480384 samples/sec	accuracy=0.759964	lr=0.212563
Epoch[60] Batch [299]	Speed: 4666.688647 samples/sec	accuracy=0.760035	lr=0.212345
Epoch[60] Batch [349]	Speed: 4653.494172 samples/sec	accuracy=0.760044	lr=0.212127
Epoch[60] Batch [399]	Speed: 4739.096983 samples/sec	accuracy=0.760164	lr=0.211909
Epoch[60] Batch [449]	Speed: 4750.999909 samples/sec	accuracy=0.760406	lr=0.211691
Epoch[60] Batch [499]	Speed: 4920.381995 samples/sec	accuracy=0.760450	lr=0.211473
Epoch[60] Batch [549]	Speed: 4931.716842 samples/sec	accuracy=0.760295	lr=0.211255
Epoch[60] Batch [599]	Speed: 5006.049260 samples/sec	accuracy=0.760437	lr=0.211037
Epoch[60] Batch [649]	Speed: 4742.330281 samples/sec	accuracy=0.760649	lr=0.210819
Epoch[60] Batch [699]	Speed: 4789.018967 samples/sec	accuracy=0.760688	lr=0.210601
Epoch[60] Batch [749]	Speed: 4814.165731 samples/sec	accuracy=0.760693	lr=0.210383
Epoch[60] Batch [799]	Speed: 4873.418593 samples/sec	accuracy=0.760857	lr=0.210165
Epoch[60] Batch [849]	Speed: 4908.250215 samples/sec	accuracy=0.761009	lr=0.209947
Epoch[60] Batch [899]	Speed: 4824.933993 samples/sec	accuracy=0.761130	lr=0.209729
Epoch[60] Batch [949]	Speed: 4861.156609 samples/sec	accuracy=0.761299	lr=0.209511
Epoch[60] Batch [999]	Speed: 4818.181422 samples/sec	accuracy=0.761440	lr=0.209293
Epoch[60] Batch [1049]	Speed: 4926.606534 samples/sec	accuracy=0.761540	lr=0.209075
Epoch[60] Batch [1099]	Speed: 4867.405974 samples/sec	accuracy=0.761717	lr=0.208856
Epoch[60] Batch [1149]	Speed: 4917.787474 samples/sec	accuracy=0.761813	lr=0.208638
Epoch[60] Batch [1199]	Speed: 4781.644083 samples/sec	accuracy=0.761918	lr=0.208420
Epoch[60] Batch [1249]	Speed: 4958.718208 samples/sec	accuracy=0.761967	lr=0.208202
[Epoch 60] training: accuracy=0.761970
[Epoch 60] speed: 4847 samples/sec	time cost: 274.608788
[Epoch 60] validation: err-top1=0.389967 err-top5=0.159339
Epoch[61] Batch [49]	Speed: 4738.717168 samples/sec	accuracy=0.762423	lr=0.207979
Epoch[61] Batch [99]	Speed: 4879.795016 samples/sec	accuracy=0.761118	lr=0.207761
Epoch[61] Batch [149]	Speed: 4886.591414 samples/sec	accuracy=0.760131	lr=0.207543
Epoch[61] Batch [199]	Speed: 4735.295199 samples/sec	accuracy=0.760254	lr=0.207325
Epoch[61] Batch [249]	Speed: 4836.182442 samples/sec	accuracy=0.760444	lr=0.207106
Epoch[61] Batch [299]	Speed: 4862.967767 samples/sec	accuracy=0.760414	lr=0.206888
Epoch[61] Batch [349]	Speed: 4848.357678 samples/sec	accuracy=0.760417	lr=0.206670
Epoch[61] Batch [399]	Speed: 4716.331523 samples/sec	accuracy=0.760783	lr=0.206452
Epoch[61] Batch [449]	Speed: 4990.293307 samples/sec	accuracy=0.760938	lr=0.206233
Epoch[61] Batch [499]	Speed: 4797.197941 samples/sec	accuracy=0.761080	lr=0.206015
Epoch[61] Batch [549]	Speed: 4699.034654 samples/sec	accuracy=0.761308	lr=0.205797
Epoch[61] Batch [599]	Speed: 4825.644915 samples/sec	accuracy=0.761441	lr=0.205579
Epoch[61] Batch [649]	Speed: 4927.876899 samples/sec	accuracy=0.761697	lr=0.205360
Epoch[61] Batch [699]	Speed: 4991.138362 samples/sec	accuracy=0.761779	lr=0.205142
Epoch[61] Batch [749]	Speed: 4940.037377 samples/sec	accuracy=0.761846	lr=0.204924
Epoch[61] Batch [799]	Speed: 4919.482403 samples/sec	accuracy=0.761895	lr=0.204705
Epoch[61] Batch [849]	Speed: 4839.157567 samples/sec	accuracy=0.762081	lr=0.204487
Epoch[61] Batch [899]	Speed: 4901.226352 samples/sec	accuracy=0.762228	lr=0.204269
Epoch[61] Batch [949]	Speed: 4943.483065 samples/sec	accuracy=0.762351	lr=0.204051
Epoch[61] Batch [999]	Speed: 4874.728507 samples/sec	accuracy=0.762386	lr=0.203832
Epoch[61] Batch [1049]	Speed: 4949.873815 samples/sec	accuracy=0.762387	lr=0.203614
Epoch[61] Batch [1099]	Speed: 4841.678059 samples/sec	accuracy=0.762524	lr=0.203396
Epoch[61] Batch [1149]	Speed: 4999.495735 samples/sec	accuracy=0.762722	lr=0.203177
Epoch[61] Batch [1199]	Speed: 4968.904482 samples/sec	accuracy=0.762813	lr=0.202959
Epoch[61] Batch [1249]	Speed: 4980.864978 samples/sec	accuracy=0.762907	lr=0.202740
[Epoch 61] training: accuracy=0.762911
[Epoch 61] speed: 4869 samples/sec	time cost: 273.519789
[Epoch 61] validation: err-top1=0.363939 err-top5=0.142140
Epoch[62] Batch [49]	Speed: 4594.651708 samples/sec	accuracy=0.763729	lr=0.202518
Epoch[62] Batch [99]	Speed: 4955.585039 samples/sec	accuracy=0.762241	lr=0.202299
Epoch[62] Batch [149]	Speed: 4884.303997 samples/sec	accuracy=0.762180	lr=0.202081
Epoch[62] Batch [199]	Speed: 4702.183641 samples/sec	accuracy=0.762016	lr=0.201863
Epoch[62] Batch [249]	Speed: 4695.699924 samples/sec	accuracy=0.761899	lr=0.201644
Epoch[62] Batch [299]	Speed: 4815.932639 samples/sec	accuracy=0.762083	lr=0.201426
Epoch[62] Batch [349]	Speed: 4714.051261 samples/sec	accuracy=0.762393	lr=0.201208
Epoch[62] Batch [399]	Speed: 4839.700459 samples/sec	accuracy=0.762489	lr=0.200989
Epoch[62] Batch [449]	Speed: 4817.154125 samples/sec	accuracy=0.762518	lr=0.200771
Epoch[62] Batch [499]	Speed: 4820.232137 samples/sec	accuracy=0.762744	lr=0.200552
Epoch[62] Batch [549]	Speed: 4620.623379 samples/sec	accuracy=0.762799	lr=0.200334
Epoch[62] Batch [599]	Speed: 4941.236454 samples/sec	accuracy=0.762964	lr=0.200116
Epoch[62] Batch [649]	Speed: 4705.977487 samples/sec	accuracy=0.763155	lr=0.199897
Epoch[62] Batch [699]	Speed: 4609.477278 samples/sec	accuracy=0.763238	lr=0.199679
Epoch[62] Batch [749]	Speed: 4757.836356 samples/sec	accuracy=0.763335	lr=0.199461
Epoch[62] Batch [799]	Speed: 4706.240887 samples/sec	accuracy=0.763331	lr=0.199242
Epoch[62] Batch [849]	Speed: 4650.680349 samples/sec	accuracy=0.763407	lr=0.199024
Epoch[62] Batch [899]	Speed: 4834.484012 samples/sec	accuracy=0.763536	lr=0.198806
Epoch[62] Batch [949]	Speed: 4977.882960 samples/sec	accuracy=0.763629	lr=0.198587
Epoch[62] Batch [999]	Speed: 4715.693035 samples/sec	accuracy=0.763645	lr=0.198369
Epoch[62] Batch [1049]	Speed: 4733.380266 samples/sec	accuracy=0.763824	lr=0.198150
Epoch[62] Batch [1099]	Speed: 4708.978315 samples/sec	accuracy=0.763918	lr=0.197932
Epoch[62] Batch [1149]	Speed: 4524.461464 samples/sec	accuracy=0.764042	lr=0.197714
Epoch[62] Batch [1199]	Speed: 5071.369167 samples/sec	accuracy=0.764108	lr=0.197495
Epoch[62] Batch [1249]	Speed: 4760.726604 samples/sec	accuracy=0.764260	lr=0.197277
[Epoch 62] training: accuracy=0.764262
[Epoch 62] speed: 4757 samples/sec	time cost: 278.476013
[Epoch 62] validation: err-top1=0.371094 err-top5=0.146225
Epoch[63] Batch [49]	Speed: 4710.198624 samples/sec	accuracy=0.763977	lr=0.197054
Epoch[63] Batch [99]	Speed: 4982.459747 samples/sec	accuracy=0.763864	lr=0.196836
Epoch[63] Batch [149]	Speed: 4855.461802 samples/sec	accuracy=0.763716	lr=0.196618
Epoch[63] Batch [199]	Speed: 4967.121334 samples/sec	accuracy=0.763652	lr=0.196399
Epoch[63] Batch [249]	Speed: 4997.768146 samples/sec	accuracy=0.763574	lr=0.196181
Epoch[63] Batch [299]	Speed: 4809.628845 samples/sec	accuracy=0.763882	lr=0.195963
Epoch[63] Batch [349]	Speed: 4834.402387 samples/sec	accuracy=0.763810	lr=0.195744
Epoch[63] Batch [399]	Speed: 4797.264919 samples/sec	accuracy=0.764021	lr=0.195526
Epoch[63] Batch [449]	Speed: 4875.366738 samples/sec	accuracy=0.763887	lr=0.195308
Epoch[63] Batch [499]	Speed: 4904.676791 samples/sec	accuracy=0.764130	lr=0.195089
Epoch[63] Batch [549]	Speed: 4655.614168 samples/sec	accuracy=0.764279	lr=0.194871
Epoch[63] Batch [599]	Speed: 5012.893559 samples/sec	accuracy=0.764263	lr=0.194653
Epoch[63] Batch [649]	Speed: 4872.713206 samples/sec	accuracy=0.764301	lr=0.194434
Epoch[63] Batch [699]	Speed: 4834.599054 samples/sec	accuracy=0.764411	lr=0.194216
Epoch[63] Batch [749]	Speed: 4891.414437 samples/sec	accuracy=0.764549	lr=0.193998
Epoch[63] Batch [799]	Speed: 5077.911235 samples/sec	accuracy=0.764752	lr=0.193780
Epoch[63] Batch [849]	Speed: 4987.062331 samples/sec	accuracy=0.764854	lr=0.193561
Epoch[63] Batch [899]	Speed: 4841.162117 samples/sec	accuracy=0.764975	lr=0.193343
Epoch[63] Batch [949]	Speed: 5075.710436 samples/sec	accuracy=0.765093	lr=0.193125
Epoch[63] Batch [999]	Speed: 5099.158790 samples/sec	accuracy=0.765196	lr=0.192907
Epoch[63] Batch [1049]	Speed: 4862.778915 samples/sec	accuracy=0.765283	lr=0.192688
Epoch[63] Batch [1099]	Speed: 4847.975689 samples/sec	accuracy=0.765319	lr=0.192470
Epoch[63] Batch [1149]	Speed: 4782.950712 samples/sec	accuracy=0.765397	lr=0.192252
Epoch[63] Batch [1199]	Speed: 4819.101119 samples/sec	accuracy=0.765515	lr=0.192034
Epoch[63] Batch [1249]	Speed: 4742.572314 samples/sec	accuracy=0.765533	lr=0.191816
[Epoch 63] training: accuracy=0.765533
[Epoch 63] speed: 4877 samples/sec	time cost: 273.053281
[Epoch 63] validation: err-top1=0.372396 err-top5=0.145996
Epoch[64] Batch [49]	Speed: 4571.182848 samples/sec	accuracy=0.765793	lr=0.191593
Epoch[64] Batch [99]	Speed: 4749.802703 samples/sec	accuracy=0.764517	lr=0.191375
Epoch[64] Batch [149]	Speed: 4883.013693 samples/sec	accuracy=0.764279	lr=0.191157
Epoch[64] Batch [199]	Speed: 4684.627592 samples/sec	accuracy=0.764294	lr=0.190939
Epoch[64] Batch [249]	Speed: 4619.722815 samples/sec	accuracy=0.764651	lr=0.190720
Epoch[64] Batch [299]	Speed: 4675.489896 samples/sec	accuracy=0.764603	lr=0.190502
Epoch[64] Batch [349]	Speed: 4734.993040 samples/sec	accuracy=0.764581	lr=0.190284
Epoch[64] Batch [399]	Speed: 4736.447702 samples/sec	accuracy=0.764684	lr=0.190066
Epoch[64] Batch [449]	Speed: 4693.453309 samples/sec	accuracy=0.764805	lr=0.189848
Epoch[64] Batch [499]	Speed: 4791.749319 samples/sec	accuracy=0.765012	lr=0.189630
Epoch[64] Batch [549]	Speed: 4969.085454 samples/sec	accuracy=0.765149	lr=0.189412
Epoch[64] Batch [599]	Speed: 4579.198514 samples/sec	accuracy=0.765189	lr=0.189194
Epoch[64] Batch [649]	Speed: 4789.166246 samples/sec	accuracy=0.765424	lr=0.188976
Epoch[64] Batch [699]	Speed: 4794.479254 samples/sec	accuracy=0.765555	lr=0.188758
Epoch[64] Batch [749]	Speed: 4761.481439 samples/sec	accuracy=0.765777	lr=0.188540
Epoch[64] Batch [799]	Speed: 4937.218332 samples/sec	accuracy=0.765912	lr=0.188322
Epoch[64] Batch [849]	Speed: 4807.971725 samples/sec	accuracy=0.766013	lr=0.188104
Epoch[64] Batch [899]	Speed: 4912.683277 samples/sec	accuracy=0.766020	lr=0.187886
Epoch[64] Batch [949]	Speed: 5099.992674 samples/sec	accuracy=0.766110	lr=0.187668
Epoch[64] Batch [999]	Speed: 4780.545465 samples/sec	accuracy=0.766213	lr=0.187450
Epoch[64] Batch [1049]	Speed: 4822.969178 samples/sec	accuracy=0.766299	lr=0.187232
Epoch[64] Batch [1099]	Speed: 4844.469350 samples/sec	accuracy=0.766334	lr=0.187014
Epoch[64] Batch [1149]	Speed: 4795.120091 samples/sec	accuracy=0.766402	lr=0.186796
Epoch[64] Batch [1199]	Speed: 4873.286435 samples/sec	accuracy=0.766478	lr=0.186578
Epoch[64] Batch [1249]	Speed: 4941.302285 samples/sec	accuracy=0.766583	lr=0.186360
[Epoch 64] training: accuracy=0.766592
[Epoch 64] speed: 4787 samples/sec	time cost: 277.577134
[Epoch 64] validation: err-top1=0.415258 err-top5=0.179907
Epoch[65] Batch [49]	Speed: 4862.860621 samples/sec	accuracy=0.767193	lr=0.186138
Epoch[65] Batch [99]	Speed: 4761.735568 samples/sec	accuracy=0.767021	lr=0.185920
Epoch[65] Batch [149]	Speed: 4968.645119 samples/sec	accuracy=0.766798	lr=0.185702
Epoch[65] Batch [199]	Speed: 4884.661845 samples/sec	accuracy=0.766465	lr=0.185485
Epoch[65] Batch [249]	Speed: 4559.371834 samples/sec	accuracy=0.766541	lr=0.185267
Epoch[65] Batch [299]	Speed: 4841.499590 samples/sec	accuracy=0.766426	lr=0.185049
Epoch[65] Batch [349]	Speed: 5020.779400 samples/sec	accuracy=0.766522	lr=0.184831
Epoch[65] Batch [399]	Speed: 5081.833399 samples/sec	accuracy=0.766519	lr=0.184614
Epoch[65] Batch [449]	Speed: 4998.539526 samples/sec	accuracy=0.766653	lr=0.184396
Epoch[65] Batch [499]	Speed: 4917.625984 samples/sec	accuracy=0.766840	lr=0.184178
Epoch[65] Batch [549]	Speed: 5073.655264 samples/sec	accuracy=0.766910	lr=0.183960
Epoch[65] Batch [599]	Speed: 4862.148159 samples/sec	accuracy=0.766916	lr=0.183743
Epoch[65] Batch [649]	Speed: 4897.518833 samples/sec	accuracy=0.767093	lr=0.183525
Epoch[65] Batch [699]	Speed: 4813.160534 samples/sec	accuracy=0.767200	lr=0.183308
Epoch[65] Batch [749]	Speed: 4866.524654 samples/sec	accuracy=0.767320	lr=0.183090
Epoch[65] Batch [799]	Speed: 4494.364261 samples/sec	accuracy=0.767498	lr=0.182872
Epoch[65] Batch [849]	Speed: 4938.454994 samples/sec	accuracy=0.767585	lr=0.182655
Epoch[65] Batch [899]	Speed: 4838.348690 samples/sec	accuracy=0.767718	lr=0.182437
Epoch[65] Batch [949]	Speed: 4793.597286 samples/sec	accuracy=0.767802	lr=0.182220
Epoch[65] Batch [999]	Speed: 4813.062799 samples/sec	accuracy=0.767955	lr=0.182002
Epoch[65] Batch [1049]	Speed: 4865.119836 samples/sec	accuracy=0.768127	lr=0.181785
Epoch[65] Batch [1099]	Speed: 4763.010949 samples/sec	accuracy=0.768168	lr=0.181567
Epoch[65] Batch [1149]	Speed: 4758.579203 samples/sec	accuracy=0.768225	lr=0.181350
Epoch[65] Batch [1199]	Speed: 4861.087285 samples/sec	accuracy=0.768345	lr=0.181132
Epoch[65] Batch [1249]	Speed: 4943.551345 samples/sec	accuracy=0.768486	lr=0.180915
[Epoch 65] training: accuracy=0.768489
[Epoch 65] speed: 4851 samples/sec	time cost: 274.213162
[Epoch 65] validation: err-top1=0.375000 err-top5=0.149454
Epoch[66] Batch [49]	Speed: 4847.454355 samples/sec	accuracy=0.768332	lr=0.180693
Epoch[66] Batch [99]	Speed: 4907.128315 samples/sec	accuracy=0.768325	lr=0.180476
Epoch[66] Batch [149]	Speed: 4919.915870 samples/sec	accuracy=0.767953	lr=0.180259
Epoch[66] Batch [199]	Speed: 4665.270127 samples/sec	accuracy=0.768280	lr=0.180041
Epoch[66] Batch [249]	Speed: 4776.020498 samples/sec	accuracy=0.768064	lr=0.179824
Epoch[66] Batch [299]	Speed: 4750.837626 samples/sec	accuracy=0.768117	lr=0.179607
Epoch[66] Batch [349]	Speed: 4754.624310 samples/sec	accuracy=0.768196	lr=0.179390
Epoch[66] Batch [399]	Speed: 4897.970111 samples/sec	accuracy=0.768239	lr=0.179172
Epoch[66] Batch [449]	Speed: 4781.430302 samples/sec	accuracy=0.768200	lr=0.178955
Epoch[66] Batch [499]	Speed: 4798.019273 samples/sec	accuracy=0.768292	lr=0.178738
Epoch[66] Batch [549]	Speed: 4929.340489 samples/sec	accuracy=0.768423	lr=0.178521
Epoch[66] Batch [599]	Speed: 4810.594634 samples/sec	accuracy=0.768617	lr=0.178304
Epoch[66] Batch [649]	Speed: 4798.806785 samples/sec	accuracy=0.768770	lr=0.178087
Epoch[66] Batch [699]	Speed: 4745.663951 samples/sec	accuracy=0.768857	lr=0.177870
Epoch[66] Batch [749]	Speed: 5024.004862 samples/sec	accuracy=0.768854	lr=0.177653
Epoch[66] Batch [799]	Speed: 4883.380235 samples/sec	accuracy=0.768799	lr=0.177436
Epoch[66] Batch [849]	Speed: 4796.408386 samples/sec	accuracy=0.768785	lr=0.177219
Epoch[66] Batch [899]	Speed: 4900.224392 samples/sec	accuracy=0.768937	lr=0.177002
Epoch[66] Batch [949]	Speed: 5082.531107 samples/sec	accuracy=0.769051	lr=0.176785
Epoch[66] Batch [999]	Speed: 4960.862477 samples/sec	accuracy=0.769174	lr=0.176568
Epoch[66] Batch [1049]	Speed: 4837.651457 samples/sec	accuracy=0.769287	lr=0.176351
Epoch[66] Batch [1099]	Speed: 4850.762305 samples/sec	accuracy=0.769381	lr=0.176134
Epoch[66] Batch [1149]	Speed: 4923.008259 samples/sec	accuracy=0.769523	lr=0.175918
Epoch[66] Batch [1199]	Speed: 4875.002393 samples/sec	accuracy=0.769687	lr=0.175701
Epoch[66] Batch [1249]	Speed: 5008.969209 samples/sec	accuracy=0.769809	lr=0.175484
[Epoch 66] training: accuracy=0.769815
[Epoch 66] speed: 4854 samples/sec	time cost: 274.210877
[Epoch 66] validation: err-top1=0.367486 err-top5=0.143953
Epoch[67] Batch [49]	Speed: 4586.297636 samples/sec	accuracy=0.769154	lr=0.175263
Epoch[67] Batch [99]	Speed: 4845.378995 samples/sec	accuracy=0.769016	lr=0.175046
Epoch[67] Batch [149]	Speed: 4938.246835 samples/sec	accuracy=0.768656	lr=0.174830
Epoch[67] Batch [199]	Speed: 4842.163977 samples/sec	accuracy=0.768901	lr=0.174613
Epoch[67] Batch [249]	Speed: 4942.196113 samples/sec	accuracy=0.768964	lr=0.174397
Epoch[67] Batch [299]	Speed: 4682.001484 samples/sec	accuracy=0.769182	lr=0.174180
Epoch[67] Batch [349]	Speed: 4983.580510 samples/sec	accuracy=0.769018	lr=0.173964
Epoch[67] Batch [399]	Speed: 4896.862730 samples/sec	accuracy=0.769072	lr=0.173747
Epoch[67] Batch [449]	Speed: 4854.724728 samples/sec	accuracy=0.769285	lr=0.173531
Epoch[67] Batch [499]	Speed: 4809.187991 samples/sec	accuracy=0.769329	lr=0.173314
Epoch[67] Batch [549]	Speed: 4808.490199 samples/sec	accuracy=0.769208	lr=0.173098
Epoch[67] Batch [599]	Speed: 4764.328131 samples/sec	accuracy=0.769168	lr=0.172881
Epoch[67] Batch [649]	Speed: 4865.465177 samples/sec	accuracy=0.769196	lr=0.172665
Epoch[67] Batch [699]	Speed: 4704.277856 samples/sec	accuracy=0.769215	lr=0.172449
Epoch[67] Batch [749]	Speed: 4812.358922 samples/sec	accuracy=0.769377	lr=0.172232
Epoch[67] Batch [799]	Speed: 4980.269629 samples/sec	accuracy=0.769463	lr=0.172016
Epoch[67] Batch [849]	Speed: 4787.381886 samples/sec	accuracy=0.769558	lr=0.171800
Epoch[67] Batch [899]	Speed: 4717.474399 samples/sec	accuracy=0.769629	lr=0.171584
Epoch[67] Batch [949]	Speed: 4775.657894 samples/sec	accuracy=0.769807	lr=0.171368
Epoch[67] Batch [999]	Speed: 4985.816719 samples/sec	accuracy=0.769889	lr=0.171152
Epoch[67] Batch [1049]	Speed: 4774.119827 samples/sec	accuracy=0.769964	lr=0.170935
Epoch[67] Batch [1099]	Speed: 4745.362250 samples/sec	accuracy=0.770055	lr=0.170719
Epoch[67] Batch [1149]	Speed: 4647.693490 samples/sec	accuracy=0.770226	lr=0.170503
Epoch[67] Batch [1199]	Speed: 5046.715191 samples/sec	accuracy=0.770330	lr=0.170287
Epoch[67] Batch [1249]	Speed: 4912.881756 samples/sec	accuracy=0.770423	lr=0.170072
[Epoch 67] training: accuracy=0.770427
[Epoch 67] speed: 4821 samples/sec	time cost: 275.588186
[Epoch 67] validation: err-top1=0.348135 err-top5=0.130341
Epoch[68] Batch [49]	Speed: 4485.421055 samples/sec	accuracy=0.771489	lr=0.169851
Epoch[68] Batch [99]	Speed: 4574.239726 samples/sec	accuracy=0.770366	lr=0.169635
Epoch[68] Batch [149]	Speed: 4723.442062 samples/sec	accuracy=0.770359	lr=0.169420
Epoch[68] Batch [199]	Speed: 4930.559510 samples/sec	accuracy=0.770357	lr=0.169204
Epoch[68] Batch [249]	Speed: 4798.925497 samples/sec	accuracy=0.770213	lr=0.168988
Epoch[68] Batch [299]	Speed: 4932.878229 samples/sec	accuracy=0.770096	lr=0.168772
Epoch[68] Batch [349]	Speed: 5009.139792 samples/sec	accuracy=0.769934	lr=0.168557
Epoch[68] Batch [399]	Speed: 4788.692401 samples/sec	accuracy=0.769874	lr=0.168341
Epoch[68] Batch [449]	Speed: 4929.997402 samples/sec	accuracy=0.770025	lr=0.168126
Epoch[68] Batch [499]	Speed: 4604.199023 samples/sec	accuracy=0.770248	lr=0.167910
Epoch[68] Batch [549]	Speed: 4962.053342 samples/sec	accuracy=0.770396	lr=0.167694
Epoch[68] Batch [599]	Speed: 4718.679104 samples/sec	accuracy=0.770536	lr=0.167479
Epoch[68] Batch [649]	Speed: 4547.539745 samples/sec	accuracy=0.770553	lr=0.167263
Epoch[68] Batch [699]	Speed: 4811.875409 samples/sec	accuracy=0.770703	lr=0.167048
Epoch[68] Batch [749]	Speed: 5060.929812 samples/sec	accuracy=0.770864	lr=0.166833
Epoch[68] Batch [799]	Speed: 4705.264270 samples/sec	accuracy=0.771091	lr=0.166617
Epoch[68] Batch [849]	Speed: 4815.404461 samples/sec	accuracy=0.771179	lr=0.166402
Epoch[68] Batch [899]	Speed: 4918.552272 samples/sec	accuracy=0.771334	lr=0.166187
Epoch[68] Batch [949]	Speed: 5022.522350 samples/sec	accuracy=0.771368	lr=0.165972
Epoch[68] Batch [999]	Speed: 4886.984629 samples/sec	accuracy=0.771441	lr=0.165756
Epoch[68] Batch [1049]	Speed: 4840.399592 samples/sec	accuracy=0.771510	lr=0.165541
Epoch[68] Batch [1099]	Speed: 4907.526636 samples/sec	accuracy=0.771567	lr=0.165326
Epoch[68] Batch [1149]	Speed: 4867.444367 samples/sec	accuracy=0.771668	lr=0.165111
Epoch[68] Batch [1199]	Speed: 4943.149886 samples/sec	accuracy=0.771729	lr=0.164896
Epoch[68] Batch [1249]	Speed: 4893.926995 samples/sec	accuracy=0.771708	lr=0.164681
[Epoch 68] training: accuracy=0.771708
[Epoch 68] speed: 4818 samples/sec	time cost: 275.249686
[Epoch 68] validation: err-top1=0.379923 err-top5=0.154556
Epoch[69] Batch [49]	Speed: 4647.992254 samples/sec	accuracy=0.771183	lr=0.164462
Epoch[69] Batch [99]	Speed: 4812.907143 samples/sec	accuracy=0.770362	lr=0.164247
Epoch[69] Batch [149]	Speed: 4877.115729 samples/sec	accuracy=0.770760	lr=0.164032
Epoch[69] Batch [199]	Speed: 4797.218195 samples/sec	accuracy=0.771217	lr=0.163818
Epoch[69] Batch [249]	Speed: 4841.974773 samples/sec	accuracy=0.771152	lr=0.163603
Epoch[69] Batch [299]	Speed: 4789.930234 samples/sec	accuracy=0.770982	lr=0.163388
Epoch[69] Batch [349]	Speed: 4912.797350 samples/sec	accuracy=0.770980	lr=0.163173
Epoch[69] Batch [399]	Speed: 4735.870597 samples/sec	accuracy=0.770980	lr=0.162959
Epoch[69] Batch [449]	Speed: 4701.966508 samples/sec	accuracy=0.771217	lr=0.162744
Epoch[69] Batch [499]	Speed: 4766.769457 samples/sec	accuracy=0.771318	lr=0.162530
Epoch[69] Batch [549]	Speed: 4789.656956 samples/sec	accuracy=0.771420	lr=0.162315
Epoch[69] Batch [599]	Speed: 4990.781678 samples/sec	accuracy=0.771511	lr=0.162101
Epoch[69] Batch [649]	Speed: 5063.653254 samples/sec	accuracy=0.771585	lr=0.161886
Epoch[69] Batch [699]	Speed: 4898.335103 samples/sec	accuracy=0.771757	lr=0.161672
Epoch[69] Batch [749]	Speed: 4655.984513 samples/sec	accuracy=0.771896	lr=0.161458
Epoch[69] Batch [799]	Speed: 5107.272197 samples/sec	accuracy=0.771970	lr=0.161244
Epoch[69] Batch [849]	Speed: 5156.965038 samples/sec	accuracy=0.772091	lr=0.161029
Epoch[69] Batch [899]	Speed: 5097.187062 samples/sec	accuracy=0.772296	lr=0.160815
Epoch[69] Batch [949]	Speed: 5113.953804 samples/sec	accuracy=0.772358	lr=0.160601
Epoch[69] Batch [999]	Speed: 4827.173505 samples/sec	accuracy=0.772462	lr=0.160387
Epoch[69] Batch [1049]	Speed: 4849.322440 samples/sec	accuracy=0.772572	lr=0.160173
Epoch[69] Batch [1099]	Speed: 4689.790128 samples/sec	accuracy=0.772707	lr=0.159959
Epoch[69] Batch [1149]	Speed: 4748.491335 samples/sec	accuracy=0.772749	lr=0.159745
Epoch[69] Batch [1199]	Speed: 4985.878417 samples/sec	accuracy=0.772821	lr=0.159531
Epoch[69] Batch [1249]	Speed: 4915.666544 samples/sec	accuracy=0.772919	lr=0.159317
[Epoch 69] training: accuracy=0.772920
[Epoch 69] speed: 4862 samples/sec	time cost: 273.195330
[Epoch 69] validation: err-top1=0.354207 err-top5=0.134481
Epoch[70] Batch [49]	Speed: 4682.282420 samples/sec	accuracy=0.772580	lr=0.159099
Epoch[70] Batch [99]	Speed: 4858.960549 samples/sec	accuracy=0.772388	lr=0.158886
Epoch[70] Batch [149]	Speed: 4856.545043 samples/sec	accuracy=0.772116	lr=0.158672
Epoch[70] Batch [199]	Speed: 4932.425708 samples/sec	accuracy=0.771965	lr=0.158458
Epoch[70] Batch [249]	Speed: 4814.285636 samples/sec	accuracy=0.772381	lr=0.158245
Epoch[70] Batch [299]	Speed: 4708.135158 samples/sec	accuracy=0.772387	lr=0.158031
Epoch[70] Batch [349]	Speed: 4934.650592 samples/sec	accuracy=0.772622	lr=0.157818
Epoch[70] Batch [399]	Speed: 4845.854503 samples/sec	accuracy=0.772665	lr=0.157604
Epoch[70] Batch [449]	Speed: 5038.623716 samples/sec	accuracy=0.772571	lr=0.157391
Epoch[70] Batch [499]	Speed: 5011.407773 samples/sec	accuracy=0.772795	lr=0.157177
Epoch[70] Batch [549]	Speed: 4863.493985 samples/sec	accuracy=0.772791	lr=0.156964
Epoch[70] Batch [599]	Speed: 4839.775500 samples/sec	accuracy=0.772843	lr=0.156751
Epoch[70] Batch [649]	Speed: 4893.358267 samples/sec	accuracy=0.772980	lr=0.156538
Epoch[70] Batch [699]	Speed: 4736.559902 samples/sec	accuracy=0.772948	lr=0.156325
Epoch[70] Batch [749]	Speed: 4891.513486 samples/sec	accuracy=0.772960	lr=0.156112
Epoch[70] Batch [799]	Speed: 4818.234069 samples/sec	accuracy=0.773165	lr=0.155899
Epoch[70] Batch [849]	Speed: 4803.184785 samples/sec	accuracy=0.773218	lr=0.155686
Epoch[70] Batch [899]	Speed: 4699.730865 samples/sec	accuracy=0.773323	lr=0.155473
Epoch[70] Batch [949]	Speed: 5004.615113 samples/sec	accuracy=0.773438	lr=0.155260
Epoch[70] Batch [999]	Speed: 4994.889934 samples/sec	accuracy=0.773553	lr=0.155047
Epoch[70] Batch [1049]	Speed: 4789.553871 samples/sec	accuracy=0.773717	lr=0.154834
Epoch[70] Batch [1099]	Speed: 4947.321619 samples/sec	accuracy=0.773870	lr=0.154622
Epoch[70] Batch [1149]	Speed: 4803.985170 samples/sec	accuracy=0.773946	lr=0.154409
Epoch[70] Batch [1199]	Speed: 4886.172248 samples/sec	accuracy=0.774095	lr=0.154196
Epoch[70] Batch [1249]	Speed: 4942.927109 samples/sec	accuracy=0.774180	lr=0.153984
[Epoch 70] training: accuracy=0.774185
[Epoch 70] speed: 4858 samples/sec	time cost: 273.241991
[Epoch 70] validation: err-top1=0.340780 err-top5=0.127192
Epoch[71] Batch [49]	Speed: 4738.253984 samples/sec	accuracy=0.774718	lr=0.153767
Epoch[71] Batch [99]	Speed: 4843.852512 samples/sec	accuracy=0.774440	lr=0.153555
Epoch[71] Batch [149]	Speed: 4649.978759 samples/sec	accuracy=0.774450	lr=0.153342
Epoch[71] Batch [199]	Speed: 4808.896681 samples/sec	accuracy=0.774315	lr=0.153130
Epoch[71] Batch [249]	Speed: 4827.785777 samples/sec	accuracy=0.773934	lr=0.152918
Epoch[71] Batch [299]	Speed: 4818.643065 samples/sec	accuracy=0.773883	lr=0.152705
Epoch[71] Batch [349]	Speed: 4870.992451 samples/sec	accuracy=0.773963	lr=0.152493
Epoch[71] Batch [399]	Speed: 4784.072924 samples/sec	accuracy=0.774158	lr=0.152281
Epoch[71] Batch [449]	Speed: 5073.470671 samples/sec	accuracy=0.774162	lr=0.152069
Epoch[71] Batch [499]	Speed: 5042.540577 samples/sec	accuracy=0.774285	lr=0.151857
Epoch[71] Batch [549]	Speed: 4834.747082 samples/sec	accuracy=0.774543	lr=0.151645
Epoch[71] Batch [599]	Speed: 4833.561155 samples/sec	accuracy=0.774643	lr=0.151433
Epoch[71] Batch [649]	Speed: 4936.918343 samples/sec	accuracy=0.774698	lr=0.151222
Epoch[71] Batch [699]	Speed: 4881.117124 samples/sec	accuracy=0.774716	lr=0.151010
Epoch[71] Batch [749]	Speed: 4822.685294 samples/sec	accuracy=0.774837	lr=0.150798
Epoch[71] Batch [799]	Speed: 4973.326666 samples/sec	accuracy=0.774983	lr=0.150587
Epoch[71] Batch [849]	Speed: 4851.848054 samples/sec	accuracy=0.775052	lr=0.150375
Epoch[71] Batch [899]	Speed: 4781.285201 samples/sec	accuracy=0.775118	lr=0.150163
Epoch[71] Batch [949]	Speed: 4928.022551 samples/sec	accuracy=0.775229	lr=0.149952
Epoch[71] Batch [999]	Speed: 4924.542474 samples/sec	accuracy=0.775267	lr=0.149741
Epoch[71] Batch [1049]	Speed: 4784.359421 samples/sec	accuracy=0.775347	lr=0.149529
Epoch[71] Batch [1099]	Speed: 4953.014500 samples/sec	accuracy=0.775400	lr=0.149318
Epoch[71] Batch [1149]	Speed: 4821.166255 samples/sec	accuracy=0.775442	lr=0.149107
Epoch[71] Batch [1199]	Speed: 4853.369929 samples/sec	accuracy=0.775580	lr=0.148896
Epoch[71] Batch [1249]	Speed: 4940.263758 samples/sec	accuracy=0.775689	lr=0.148685
[Epoch 71] training: accuracy=0.775685
[Epoch 71] speed: 4857 samples/sec	time cost: 273.842321
[Epoch 71] validation: err-top1=0.362942 err-top5=0.141342
Epoch[72] Batch [49]	Speed: 4548.043928 samples/sec	accuracy=0.775816	lr=0.148469
Epoch[72] Batch [99]	Speed: 4778.064496 samples/sec	accuracy=0.775986	lr=0.148258
Epoch[72] Batch [149]	Speed: 4571.330559 samples/sec	accuracy=0.775837	lr=0.148047
Epoch[72] Batch [199]	Speed: 4673.506762 samples/sec	accuracy=0.775406	lr=0.147837
Epoch[72] Batch [249]	Speed: 4767.923679 samples/sec	accuracy=0.775217	lr=0.147626
Epoch[72] Batch [299]	Speed: 4808.790505 samples/sec	accuracy=0.775443	lr=0.147415
Epoch[72] Batch [349]	Speed: 4834.709095 samples/sec	accuracy=0.775750	lr=0.147204
Epoch[72] Batch [399]	Speed: 4924.625364 samples/sec	accuracy=0.775765	lr=0.146994
Epoch[72] Batch [449]	Speed: 4833.578562 samples/sec	accuracy=0.775817	lr=0.146783
Epoch[72] Batch [499]	Speed: 4815.526371 samples/sec	accuracy=0.775929	lr=0.146573
Epoch[72] Batch [549]	Speed: 4895.747814 samples/sec	accuracy=0.776054	lr=0.146362
Epoch[72] Batch [599]	Speed: 4996.633670 samples/sec	accuracy=0.776070	lr=0.146152
Epoch[72] Batch [649]	Speed: 4817.813685 samples/sec	accuracy=0.776140	lr=0.145942
Epoch[72] Batch [699]	Speed: 4949.188668 samples/sec	accuracy=0.776256	lr=0.145732
Epoch[72] Batch [749]	Speed: 4753.424961 samples/sec	accuracy=0.776276	lr=0.145521
Epoch[72] Batch [799]	Speed: 4874.889072 samples/sec	accuracy=0.776343	lr=0.145311
Epoch[72] Batch [849]	Speed: 4885.110756 samples/sec	accuracy=0.776473	lr=0.145101
Epoch[72] Batch [899]	Speed: 4868.199655 samples/sec	accuracy=0.776494	lr=0.144891
Epoch[72] Batch [949]	Speed: 4742.475645 samples/sec	accuracy=0.776556	lr=0.144682
Epoch[72] Batch [999]	Speed: 4733.033391 samples/sec	accuracy=0.776655	lr=0.144472
Epoch[72] Batch [1049]	Speed: 4813.790945 samples/sec	accuracy=0.776742	lr=0.144262
Epoch[72] Batch [1099]	Speed: 4963.006767 samples/sec	accuracy=0.776809	lr=0.144052
Epoch[72] Batch [1149]	Speed: 4826.080224 samples/sec	accuracy=0.776875	lr=0.143843
Epoch[72] Batch [1199]	Speed: 4843.261282 samples/sec	accuracy=0.776910	lr=0.143633
Epoch[72] Batch [1249]	Speed: 4854.736141 samples/sec	accuracy=0.776902	lr=0.143424
[Epoch 72] training: accuracy=0.776904
[Epoch 72] speed: 4808 samples/sec	time cost: 276.725585
[Epoch 72] validation: err-top1=0.341498 err-top5=0.125299
Epoch[73] Batch [49]	Speed: 4769.754899 samples/sec	accuracy=0.776997	lr=0.143210
Epoch[73] Batch [99]	Speed: 4944.475132 samples/sec	accuracy=0.776784	lr=0.143001
Epoch[73] Batch [149]	Speed: 4883.412773 samples/sec	accuracy=0.777089	lr=0.142791
Epoch[73] Batch [199]	Speed: 4788.778897 samples/sec	accuracy=0.776806	lr=0.142582
Epoch[73] Batch [249]	Speed: 4736.186134 samples/sec	accuracy=0.776904	lr=0.142373
Epoch[73] Batch [299]	Speed: 4774.195927 samples/sec	accuracy=0.776666	lr=0.142164
Epoch[73] Batch [349]	Speed: 4819.310278 samples/sec	accuracy=0.776738	lr=0.141955
Epoch[73] Batch [399]	Speed: 4848.358991 samples/sec	accuracy=0.776705	lr=0.141746
Epoch[73] Batch [449]	Speed: 5009.099014 samples/sec	accuracy=0.776862	lr=0.141537
Epoch[73] Batch [499]	Speed: 4959.198929 samples/sec	accuracy=0.777027	lr=0.141328
Epoch[73] Batch [549]	Speed: 4985.107468 samples/sec	accuracy=0.777017	lr=0.141120
Epoch[73] Batch [599]	Speed: 4797.328363 samples/sec	accuracy=0.777201	lr=0.140911
Epoch[73] Batch [649]	Speed: 4891.382685 samples/sec	accuracy=0.777294	lr=0.140702
Epoch[73] Batch [699]	Speed: 4839.633708 samples/sec	accuracy=0.777344	lr=0.140494
Epoch[73] Batch [749]	Speed: 4801.732116 samples/sec	accuracy=0.777415	lr=0.140285
Epoch[73] Batch [799]	Speed: 4993.539155 samples/sec	accuracy=0.777544	lr=0.140077
Epoch[73] Batch [849]	Speed: 4854.535967 samples/sec	accuracy=0.777694	lr=0.139869
Epoch[73] Batch [899]	Speed: 5033.118401 samples/sec	accuracy=0.777713	lr=0.139661
Epoch[73] Batch [949]	Speed: 4783.422569 samples/sec	accuracy=0.777761	lr=0.139452
Epoch[73] Batch [999]	Speed: 4902.648409 samples/sec	accuracy=0.777831	lr=0.139244
Epoch[73] Batch [1049]	Speed: 4794.895575 samples/sec	accuracy=0.777938	lr=0.139036
Epoch[73] Batch [1099]	Speed: 4872.594022 samples/sec	accuracy=0.778012	lr=0.138828
Epoch[73] Batch [1149]	Speed: 4911.047723 samples/sec	accuracy=0.778092	lr=0.138620
Epoch[73] Batch [1199]	Speed: 5008.407654 samples/sec	accuracy=0.778187	lr=0.138413
Epoch[73] Batch [1249]	Speed: 5014.767577 samples/sec	accuracy=0.778246	lr=0.138205
[Epoch 73] training: accuracy=0.778248
[Epoch 73] speed: 4875 samples/sec	time cost: 272.670386
[Epoch 73] validation: err-top1=0.336635 err-top5=0.123665
Epoch[74] Batch [49]	Speed: 4717.626637 samples/sec	accuracy=0.776774	lr=0.137993
Epoch[74] Batch [99]	Speed: 4783.770263 samples/sec	accuracy=0.777181	lr=0.137786
Epoch[74] Batch [149]	Speed: 4995.477745 samples/sec	accuracy=0.777303	lr=0.137578
Epoch[74] Batch [199]	Speed: 4917.374199 samples/sec	accuracy=0.777609	lr=0.137371
Epoch[74] Batch [249]	Speed: 4813.765479 samples/sec	accuracy=0.777540	lr=0.137163
Epoch[74] Batch [299]	Speed: 4737.470961 samples/sec	accuracy=0.777453	lr=0.136956
Epoch[74] Batch [349]	Speed: 4933.485650 samples/sec	accuracy=0.777354	lr=0.136749
Epoch[74] Batch [399]	Speed: 4921.186507 samples/sec	accuracy=0.777362	lr=0.136542
Epoch[74] Batch [449]	Speed: 4986.114345 samples/sec	accuracy=0.777473	lr=0.136335
Epoch[74] Batch [499]	Speed: 4787.596199 samples/sec	accuracy=0.777297	lr=0.136128
Epoch[74] Batch [549]	Speed: 4684.852121 samples/sec	accuracy=0.777378	lr=0.135921
Epoch[74] Batch [599]	Speed: 4976.509763 samples/sec	accuracy=0.777506	lr=0.135714
Epoch[74] Batch [649]	Speed: 5054.381911 samples/sec	accuracy=0.777585	lr=0.135507
Epoch[74] Batch [699]	Speed: 4916.217848 samples/sec	accuracy=0.777630	lr=0.135301
Epoch[74] Batch [749]	Speed: 4841.946388 samples/sec	accuracy=0.777702	lr=0.135094
Epoch[74] Batch [799]	Speed: 4922.512187 samples/sec	accuracy=0.777831	lr=0.134887
Epoch[74] Batch [849]	Speed: 4881.974549 samples/sec	accuracy=0.777927	lr=0.134681
Epoch[74] Batch [899]	Speed: 4847.854100 samples/sec	accuracy=0.777987	lr=0.134475
Epoch[74] Batch [949]	Speed: 4812.174412 samples/sec	accuracy=0.778084	lr=0.134268
Epoch[74] Batch [999]	Speed: 4899.119792 samples/sec	accuracy=0.778184	lr=0.134062
Epoch[74] Batch [1049]	Speed: 4767.444184 samples/sec	accuracy=0.778351	lr=0.133856
Epoch[74] Batch [1099]	Speed: 5035.770654 samples/sec	accuracy=0.778393	lr=0.133650
Epoch[74] Batch [1149]	Speed: 4923.121458 samples/sec	accuracy=0.778538	lr=0.133444
Epoch[74] Batch [1199]	Speed: 4734.562108 samples/sec	accuracy=0.778660	lr=0.133238
Epoch[74] Batch [1249]	Speed: 5008.135742 samples/sec	accuracy=0.778749	lr=0.133032
[Epoch 74] training: accuracy=0.778754
[Epoch 74] speed: 4869 samples/sec	time cost: 272.946017
[Epoch 74] validation: err-top1=0.341757 err-top5=0.128528
Epoch[75] Batch [49]	Speed: 4669.591737 samples/sec	accuracy=0.777972	lr=0.132822
Epoch[75] Batch [99]	Speed: 4898.415326 samples/sec	accuracy=0.778779	lr=0.132617
Epoch[75] Batch [149]	Speed: 4802.194585 samples/sec	accuracy=0.778371	lr=0.132411
Epoch[75] Batch [199]	Speed: 5023.900257 samples/sec	accuracy=0.778736	lr=0.132206
Epoch[75] Batch [249]	Speed: 4917.072000 samples/sec	accuracy=0.778807	lr=0.132000
Epoch[75] Batch [299]	Speed: 4792.861864 samples/sec	accuracy=0.778734	lr=0.131795
Epoch[75] Batch [349]	Speed: 5019.744630 samples/sec	accuracy=0.778880	lr=0.131590
Epoch[75] Batch [399]	Speed: 4670.202872 samples/sec	accuracy=0.778971	lr=0.131385
Epoch[75] Batch [449]	Speed: 4822.149786 samples/sec	accuracy=0.779107	lr=0.131180
Epoch[75] Batch [499]	Speed: 4864.358561 samples/sec	accuracy=0.779103	lr=0.130975
Epoch[75] Batch [549]	Speed: 5031.904504 samples/sec	accuracy=0.779105	lr=0.130770
Epoch[75] Batch [599]	Speed: 4845.919566 samples/sec	accuracy=0.779119	lr=0.130565
Epoch[75] Batch [649]	Speed: 4922.968985 samples/sec	accuracy=0.779162	lr=0.130360
Epoch[75] Batch [699]	Speed: 4936.996429 samples/sec	accuracy=0.779247	lr=0.130155
Epoch[75] Batch [749]	Speed: 5012.033752 samples/sec	accuracy=0.779385	lr=0.129951
Epoch[75] Batch [799]	Speed: 4732.068355 samples/sec	accuracy=0.779522	lr=0.129746
Epoch[75] Batch [849]	Speed: 4709.215510 samples/sec	accuracy=0.779617	lr=0.129542
Epoch[75] Batch [899]	Speed: 4629.419425 samples/sec	accuracy=0.779690	lr=0.129338
Epoch[75] Batch [949]	Speed: 4634.027703 samples/sec	accuracy=0.779780	lr=0.129133
Epoch[75] Batch [999]	Speed: 4840.360643 samples/sec	accuracy=0.779845	lr=0.128929
Epoch[75] Batch [1049]	Speed: 4907.402041 samples/sec	accuracy=0.779909	lr=0.128725
Epoch[75] Batch [1099]	Speed: 4727.347831 samples/sec	accuracy=0.779987	lr=0.128521
Epoch[75] Batch [1149]	Speed: 4928.804677 samples/sec	accuracy=0.780111	lr=0.128317
Epoch[75] Batch [1199]	Speed: 4926.322073 samples/sec	accuracy=0.780211	lr=0.128113
Epoch[75] Batch [1249]	Speed: 4948.732351 samples/sec	accuracy=0.780355	lr=0.127910
[Epoch 75] training: accuracy=0.780358
[Epoch 75] speed: 4841 samples/sec	time cost: 273.999996
[Epoch 75] validation: err-top1=0.341288 err-top5=0.130981
Epoch[76] Batch [49]	Speed: 4386.606757 samples/sec	accuracy=0.780753	lr=0.127702
Epoch[76] Batch [99]	Speed: 4916.912809 samples/sec	accuracy=0.779732	lr=0.127498
Epoch[76] Batch [149]	Speed: 4907.246279 samples/sec	accuracy=0.779590	lr=0.127295
Epoch[76] Batch [199]	Speed: 4893.647298 samples/sec	accuracy=0.779686	lr=0.127092
Epoch[76] Batch [249]	Speed: 4891.678502 samples/sec	accuracy=0.780305	lr=0.126888
Epoch[76] Batch [299]	Speed: 4776.233902 samples/sec	accuracy=0.780320	lr=0.126685
Epoch[76] Batch [349]	Speed: 4794.898252 samples/sec	accuracy=0.780374	lr=0.126482
Epoch[76] Batch [399]	Speed: 4859.695831 samples/sec	accuracy=0.780568	lr=0.126279
Epoch[76] Batch [449]	Speed: 4799.059015 samples/sec	accuracy=0.780561	lr=0.126076
Epoch[76] Batch [499]	Speed: 4827.078455 samples/sec	accuracy=0.780647	lr=0.125873
Epoch[76] Batch [549]	Speed: 4967.843863 samples/sec	accuracy=0.780692	lr=0.125670
Epoch[76] Batch [599]	Speed: 4831.949150 samples/sec	accuracy=0.780746	lr=0.125468
Epoch[76] Batch [649]	Speed: 4688.403796 samples/sec	accuracy=0.780786	lr=0.125265
Epoch[76] Batch [699]	Speed: 4655.974822 samples/sec	accuracy=0.780771	lr=0.125062
Epoch[76] Batch [749]	Speed: 4691.861118 samples/sec	accuracy=0.780794	lr=0.124860
Epoch[76] Batch [799]	Speed: 4831.002698 samples/sec	accuracy=0.780965	lr=0.124658
Epoch[76] Batch [849]	Speed: 4964.530323 samples/sec	accuracy=0.781001	lr=0.124455
Epoch[76] Batch [899]	Speed: 4810.235273 samples/sec	accuracy=0.781093	lr=0.124253
Epoch[76] Batch [949]	Speed: 4909.454561 samples/sec	accuracy=0.781086	lr=0.124051
Epoch[76] Batch [999]	Speed: 4858.648888 samples/sec	accuracy=0.781190	lr=0.123849
Epoch[76] Batch [1049]	Speed: 4948.830884 samples/sec	accuracy=0.781294	lr=0.123647
Epoch[76] Batch [1099]	Speed: 4999.144722 samples/sec	accuracy=0.781306	lr=0.123446
Epoch[76] Batch [1149]	Speed: 4764.562161 samples/sec	accuracy=0.781416	lr=0.123244
Epoch[76] Batch [1199]	Speed: 4708.331802 samples/sec	accuracy=0.781527	lr=0.123042
Epoch[76] Batch [1249]	Speed: 4874.914193 samples/sec	accuracy=0.781673	lr=0.122841
[Epoch 76] training: accuracy=0.781678
[Epoch 76] speed: 4813 samples/sec	time cost: 275.770888
[Epoch 76] validation: err-top1=0.337372 err-top5=0.124920
Epoch[77] Batch [49]	Speed: 4588.028634 samples/sec	accuracy=0.782583	lr=0.122635
Epoch[77] Batch [99]	Speed: 4954.213715 samples/sec	accuracy=0.781949	lr=0.122434
Epoch[77] Batch [149]	Speed: 4875.148589 samples/sec	accuracy=0.781833	lr=0.122233
Epoch[77] Batch [199]	Speed: 5081.073367 samples/sec	accuracy=0.781964	lr=0.122032
Epoch[77] Batch [249]	Speed: 5033.428072 samples/sec	accuracy=0.781735	lr=0.121831
Epoch[77] Batch [299]	Speed: 5035.239791 samples/sec	accuracy=0.781503	lr=0.121630
Epoch[77] Batch [349]	Speed: 4944.275002 samples/sec	accuracy=0.781633	lr=0.121429
Epoch[77] Batch [399]	Speed: 4782.905545 samples/sec	accuracy=0.781717	lr=0.121228
Epoch[77] Batch [449]	Speed: 4817.822764 samples/sec	accuracy=0.781808	lr=0.121027
Epoch[77] Batch [499]	Speed: 5010.835499 samples/sec	accuracy=0.781944	lr=0.120827
Epoch[77] Batch [549]	Speed: 4913.370156 samples/sec	accuracy=0.781951	lr=0.120626
Epoch[77] Batch [599]	Speed: 4903.711491 samples/sec	accuracy=0.781966	lr=0.120426
Epoch[77] Batch [649]	Speed: 4969.225734 samples/sec	accuracy=0.781983	lr=0.120226
Epoch[77] Batch [699]	Speed: 4995.806278 samples/sec	accuracy=0.782024	lr=0.120025
Epoch[77] Batch [749]	Speed: 4759.030444 samples/sec	accuracy=0.782183	lr=0.119825
Epoch[77] Batch [799]	Speed: 5033.201448 samples/sec	accuracy=0.782296	lr=0.119625
Epoch[77] Batch [849]	Speed: 4835.384467 samples/sec	accuracy=0.782279	lr=0.119425
Epoch[77] Batch [899]	Speed: 4970.755527 samples/sec	accuracy=0.782309	lr=0.119226
Epoch[77] Batch [949]	Speed: 4845.679771 samples/sec	accuracy=0.782471	lr=0.119026
Epoch[77] Batch [999]	Speed: 4692.759881 samples/sec	accuracy=0.782586	lr=0.118826
Epoch[77] Batch [1049]	Speed: 4747.989392 samples/sec	accuracy=0.782705	lr=0.118627
Epoch[77] Batch [1099]	Speed: 5052.470685 samples/sec	accuracy=0.782833	lr=0.118427
Epoch[77] Batch [1149]	Speed: 4729.855882 samples/sec	accuracy=0.782930	lr=0.118228
Epoch[77] Batch [1199]	Speed: 4881.090609 samples/sec	accuracy=0.783006	lr=0.118029
Epoch[77] Batch [1249]	Speed: 5018.572594 samples/sec	accuracy=0.783110	lr=0.117830
[Epoch 77] training: accuracy=0.783112
[Epoch 77] speed: 4891 samples/sec	time cost: 271.153045
[Epoch 77] validation: err-top1=0.326810 err-top5=0.118842
Epoch[78] Batch [49]	Speed: 4779.707338 samples/sec	accuracy=0.784827	lr=0.117627
Epoch[78] Batch [99]	Speed: 4750.761323 samples/sec	accuracy=0.784140	lr=0.117428
Epoch[78] Batch [149]	Speed: 5180.671968 samples/sec	accuracy=0.783858	lr=0.117229
Epoch[78] Batch [199]	Speed: 4891.390038 samples/sec	accuracy=0.783870	lr=0.117030
Epoch[78] Batch [249]	Speed: 4803.855892 samples/sec	accuracy=0.783535	lr=0.116831
Epoch[78] Batch [299]	Speed: 4855.733967 samples/sec	accuracy=0.783217	lr=0.116633
Epoch[78] Batch [349]	Speed: 4965.221676 samples/sec	accuracy=0.783112	lr=0.116434
Epoch[78] Batch [399]	Speed: 4951.244792 samples/sec	accuracy=0.783139	lr=0.116236
Epoch[78] Batch [449]	Speed: 4850.852373 samples/sec	accuracy=0.783250	lr=0.116038
Epoch[78] Batch [499]	Speed: 5096.004467 samples/sec	accuracy=0.783208	lr=0.115840
Epoch[78] Batch [549]	Speed: 4942.676480 samples/sec	accuracy=0.783307	lr=0.115642
Epoch[78] Batch [599]	Speed: 5139.352290 samples/sec	accuracy=0.783452	lr=0.115444
Epoch[78] Batch [649]	Speed: 4735.256566 samples/sec	accuracy=0.783506	lr=0.115246
Epoch[78] Batch [699]	Speed: 4852.586119 samples/sec	accuracy=0.783546	lr=0.115048
Epoch[78] Batch [749]	Speed: 4749.298277 samples/sec	accuracy=0.783582	lr=0.114850
Epoch[78] Batch [799]	Speed: 4836.416723 samples/sec	accuracy=0.783653	lr=0.114653
Epoch[78] Batch [849]	Speed: 4792.514988 samples/sec	accuracy=0.783731	lr=0.114455
Epoch[78] Batch [899]	Speed: 4996.014902 samples/sec	accuracy=0.783865	lr=0.114258
Epoch[78] Batch [949]	Speed: 4857.838098 samples/sec	accuracy=0.783968	lr=0.114061
Epoch[78] Batch [999]	Speed: 4820.213961 samples/sec	accuracy=0.784013	lr=0.113864
Epoch[78] Batch [1049]	Speed: 4841.056803 samples/sec	accuracy=0.784065	lr=0.113667
Epoch[78] Batch [1099]	Speed: 5011.307318 samples/sec	accuracy=0.784103	lr=0.113470
Epoch[78] Batch [1149]	Speed: 4884.487192 samples/sec	accuracy=0.784188	lr=0.113273
Epoch[78] Batch [1199]	Speed: 4931.697022 samples/sec	accuracy=0.784265	lr=0.113076
Epoch[78] Batch [1249]	Speed: 4890.108682 samples/sec	accuracy=0.784423	lr=0.112880
[Epoch 78] training: accuracy=0.784428
[Epoch 78] speed: 4890 samples/sec	time cost: 272.318719
[Epoch 78] validation: err-top1=0.337293 err-top5=0.124322
Epoch[79] Batch [49]	Speed: 4701.895267 samples/sec	accuracy=0.784134	lr=0.112679
Epoch[79] Batch [99]	Speed: 4992.249338 samples/sec	accuracy=0.783371	lr=0.112483
Epoch[79] Batch [149]	Speed: 4891.929446 samples/sec	accuracy=0.783325	lr=0.112287
Epoch[79] Batch [199]	Speed: 4749.550267 samples/sec	accuracy=0.783669	lr=0.112090
Epoch[79] Batch [249]	Speed: 4783.420332 samples/sec	accuracy=0.783492	lr=0.111894
Epoch[79] Batch [299]	Speed: 4849.621954 samples/sec	accuracy=0.783182	lr=0.111698
Epoch[79] Batch [349]	Speed: 4854.344588 samples/sec	accuracy=0.783223	lr=0.111502
Epoch[79] Batch [399]	Speed: 4934.189243 samples/sec	accuracy=0.783286	lr=0.111307
Epoch[79] Batch [449]	Speed: 4902.216188 samples/sec	accuracy=0.783450	lr=0.111111
Epoch[79] Batch [499]	Speed: 4972.893410 samples/sec	accuracy=0.783405	lr=0.110915
Epoch[79] Batch [549]	Speed: 4868.509342 samples/sec	accuracy=0.783564	lr=0.110720
Epoch[79] Batch [599]	Speed: 4919.621361 samples/sec	accuracy=0.783649	lr=0.110525
Epoch[79] Batch [649]	Speed: 4927.843201 samples/sec	accuracy=0.783821	lr=0.110329
Epoch[79] Batch [699]	Speed: 4961.993722 samples/sec	accuracy=0.783850	lr=0.110134
Epoch[79] Batch [749]	Speed: 4941.328435 samples/sec	accuracy=0.783851	lr=0.109939
Epoch[79] Batch [799]	Speed: 4899.317960 samples/sec	accuracy=0.784001	lr=0.109744
Epoch[79] Batch [849]	Speed: 4916.021462 samples/sec	accuracy=0.784125	lr=0.109549
Epoch[79] Batch [899]	Speed: 4905.935869 samples/sec	accuracy=0.784339	lr=0.109355
Epoch[79] Batch [949]	Speed: 4963.219887 samples/sec	accuracy=0.784484	lr=0.109160
Epoch[79] Batch [999]	Speed: 5047.015150 samples/sec	accuracy=0.784571	lr=0.108966
Epoch[79] Batch [1049]	Speed: 5013.077516 samples/sec	accuracy=0.784698	lr=0.108771
Epoch[79] Batch [1099]	Speed: 5068.960692 samples/sec	accuracy=0.784809	lr=0.108577
Epoch[79] Batch [1149]	Speed: 4807.069398 samples/sec	accuracy=0.784908	lr=0.108383
Epoch[79] Batch [1199]	Speed: 4750.965433 samples/sec	accuracy=0.785057	lr=0.108189
Epoch[79] Batch [1249]	Speed: 4912.830393 samples/sec	accuracy=0.785221	lr=0.107995
[Epoch 79] training: accuracy=0.785225
[Epoch 79] speed: 4895 samples/sec	time cost: 271.459265
[Epoch 79] validation: err-top1=0.324318 err-top5=0.118702
Epoch[80] Batch [49]	Speed: 4541.374367 samples/sec	accuracy=0.785933	lr=0.107797
Epoch[80] Batch [99]	Speed: 4914.105356 samples/sec	accuracy=0.785699	lr=0.107603
Epoch[80] Batch [149]	Speed: 5011.816888 samples/sec	accuracy=0.785265	lr=0.107410
Epoch[80] Batch [199]	Speed: 4785.803943 samples/sec	accuracy=0.785324	lr=0.107216
Epoch[80] Batch [249]	Speed: 4683.077330 samples/sec	accuracy=0.785660	lr=0.107023
Epoch[80] Batch [299]	Speed: 4871.169897 samples/sec	accuracy=0.785411	lr=0.106830
Epoch[80] Batch [349]	Speed: 4825.540708 samples/sec	accuracy=0.785343	lr=0.106636
Epoch[80] Batch [399]	Speed: 4929.647027 samples/sec	accuracy=0.785329	lr=0.106443
Epoch[80] Batch [449]	Speed: 4934.966409 samples/sec	accuracy=0.785337	lr=0.106250
Epoch[80] Batch [499]	Speed: 4860.112667 samples/sec	accuracy=0.785318	lr=0.106057
Epoch[80] Batch [549]	Speed: 4884.202907 samples/sec	accuracy=0.785461	lr=0.105865
Epoch[80] Batch [599]	Speed: 4842.947482 samples/sec	accuracy=0.785591	lr=0.105672
Epoch[80] Batch [649]	Speed: 5013.053994 samples/sec	accuracy=0.785701	lr=0.105480
Epoch[80] Batch [699]	Speed: 4741.509058 samples/sec	accuracy=0.785818	lr=0.105287
Epoch[80] Batch [749]	Speed: 4826.427204 samples/sec	accuracy=0.785991	lr=0.105095
Epoch[80] Batch [799]	Speed: 4721.090985 samples/sec	accuracy=0.786080	lr=0.104903
Epoch[80] Batch [849]	Speed: 4881.094270 samples/sec	accuracy=0.786194	lr=0.104711
Epoch[80] Batch [899]	Speed: 4847.731532 samples/sec	accuracy=0.786251	lr=0.104519
Epoch[80] Batch [949]	Speed: 4881.958123 samples/sec	accuracy=0.786291	lr=0.104327
Epoch[80] Batch [999]	Speed: 4614.149917 samples/sec	accuracy=0.786302	lr=0.104135
Epoch[80] Batch [1049]	Speed: 4883.886001 samples/sec	accuracy=0.786312	lr=0.103944
Epoch[80] Batch [1099]	Speed: 4855.851340 samples/sec	accuracy=0.786387	lr=0.103752
Epoch[80] Batch [1149]	Speed: 5091.747342 samples/sec	accuracy=0.786442	lr=0.103561
Epoch[80] Batch [1199]	Speed: 4641.358570 samples/sec	accuracy=0.786541	lr=0.103370
Epoch[80] Batch [1249]	Speed: 4960.689896 samples/sec	accuracy=0.786594	lr=0.103178
[Epoch 80] training: accuracy=0.786597
[Epoch 80] speed: 4834 samples/sec	time cost: 274.314534
[Epoch 80] validation: err-top1=0.326252 err-top5=0.117287
Epoch[81] Batch [49]	Speed: 4759.679351 samples/sec	accuracy=0.786177	lr=0.102984
Epoch[81] Batch [99]	Speed: 4657.201747 samples/sec	accuracy=0.786291	lr=0.102793
Epoch[81] Batch [149]	Speed: 4813.259568 samples/sec	accuracy=0.786385	lr=0.102602
Epoch[81] Batch [199]	Speed: 4833.771136 samples/sec	accuracy=0.786158	lr=0.102411
Epoch[81] Batch [249]	Speed: 4802.827067 samples/sec	accuracy=0.786343	lr=0.102221
Epoch[81] Batch [299]	Speed: 5034.196220 samples/sec	accuracy=0.786328	lr=0.102030
Epoch[81] Batch [349]	Speed: 4807.053150 samples/sec	accuracy=0.786398	lr=0.101840
Epoch[81] Batch [399]	Speed: 4703.735866 samples/sec	accuracy=0.786474	lr=0.101650
Epoch[81] Batch [449]	Speed: 4739.841417 samples/sec	accuracy=0.786500	lr=0.101460
Epoch[81] Batch [499]	Speed: 4985.606399 samples/sec	accuracy=0.786527	lr=0.101270
Epoch[81] Batch [549]	Speed: 4645.886831 samples/sec	accuracy=0.786597	lr=0.101080
Epoch[81] Batch [599]	Speed: 4669.185824 samples/sec	accuracy=0.786601	lr=0.100890
Epoch[81] Batch [649]	Speed: 4750.459815 samples/sec	accuracy=0.786716	lr=0.100701
Epoch[81] Batch [699]	Speed: 4944.649888 samples/sec	accuracy=0.786813	lr=0.100511
Epoch[81] Batch [749]	Speed: 4803.806245 samples/sec	accuracy=0.786837	lr=0.100322
Epoch[81] Batch [799]	Speed: 4619.096801 samples/sec	accuracy=0.786880	lr=0.100132
Epoch[81] Batch [849]	Speed: 4644.833529 samples/sec	accuracy=0.786955	lr=0.099943
Epoch[81] Batch [899]	Speed: 4869.706740 samples/sec	accuracy=0.786988	lr=0.099754
Epoch[81] Batch [949]	Speed: 5160.267717 samples/sec	accuracy=0.787022	lr=0.099565
Epoch[81] Batch [999]	Speed: 4763.165085 samples/sec	accuracy=0.787116	lr=0.099377
Epoch[81] Batch [1049]	Speed: 4916.443064 samples/sec	accuracy=0.787276	lr=0.099188
Epoch[81] Batch [1099]	Speed: 4769.079200 samples/sec	accuracy=0.787376	lr=0.098999
Epoch[81] Batch [1149]	Speed: 4708.868658 samples/sec	accuracy=0.787486	lr=0.098811
Epoch[81] Batch [1199]	Speed: 4668.139893 samples/sec	accuracy=0.787640	lr=0.098623
Epoch[81] Batch [1249]	Speed: 4998.862527 samples/sec	accuracy=0.787758	lr=0.098434
[Epoch 81] training: accuracy=0.787764
[Epoch 81] speed: 4793 samples/sec	time cost: 276.956312
[Epoch 81] validation: err-top1=0.335999 err-top5=0.124207
Epoch[82] Batch [49]	Speed: 4666.173533 samples/sec	accuracy=0.788605	lr=0.098243
Epoch[82] Batch [99]	Speed: 4815.980161 samples/sec	accuracy=0.787921	lr=0.098055
Epoch[82] Batch [149]	Speed: 4767.689212 samples/sec	accuracy=0.787518	lr=0.097867
Epoch[82] Batch [199]	Speed: 4582.721219 samples/sec	accuracy=0.787390	lr=0.097679
Epoch[82] Batch [249]	Speed: 4912.695077 samples/sec	accuracy=0.787387	lr=0.097492
Epoch[82] Batch [299]	Speed: 4787.473671 samples/sec	accuracy=0.787381	lr=0.097304
Epoch[82] Batch [349]	Speed: 4941.499217 samples/sec	accuracy=0.787677	lr=0.097117
Epoch[82] Batch [399]	Speed: 4933.632541 samples/sec	accuracy=0.787664	lr=0.096930
Epoch[82] Batch [449]	Speed: 4881.736499 samples/sec	accuracy=0.787606	lr=0.096743
Epoch[82] Batch [499]	Speed: 4970.520131 samples/sec	accuracy=0.787611	lr=0.096556
Epoch[82] Batch [549]	Speed: 4986.600972 samples/sec	accuracy=0.787647	lr=0.096369
Epoch[82] Batch [599]	Speed: 4857.499772 samples/sec	accuracy=0.787627	lr=0.096182
Epoch[82] Batch [649]	Speed: 4781.399429 samples/sec	accuracy=0.787605	lr=0.095996
Epoch[82] Batch [699]	Speed: 4627.608594 samples/sec	accuracy=0.787714	lr=0.095809
Epoch[82] Batch [749]	Speed: 4687.367757 samples/sec	accuracy=0.787806	lr=0.095623
Epoch[82] Batch [799]	Speed: 4648.323858 samples/sec	accuracy=0.787882	lr=0.095437
Epoch[82] Batch [849]	Speed: 4789.556114 samples/sec	accuracy=0.787999	lr=0.095250
Epoch[82] Batch [899]	Speed: 4982.027325 samples/sec	accuracy=0.788062	lr=0.095064
Epoch[82] Batch [949]	Speed: 4676.159087 samples/sec	accuracy=0.788169	lr=0.094879
Epoch[82] Batch [999]	Speed: 4949.214446 samples/sec	accuracy=0.788239	lr=0.094693
Epoch[82] Batch [1049]	Speed: 4943.568530 samples/sec	accuracy=0.788302	lr=0.094507
Epoch[82] Batch [1099]	Speed: 4823.529028 samples/sec	accuracy=0.788447	lr=0.094322
Epoch[82] Batch [1149]	Speed: 4996.185998 samples/sec	accuracy=0.788555	lr=0.094137
Epoch[82] Batch [1199]	Speed: 5106.207057 samples/sec	accuracy=0.788691	lr=0.093951
Epoch[82] Batch [1249]	Speed: 4894.186089 samples/sec	accuracy=0.788788	lr=0.093766
[Epoch 82] training: accuracy=0.788790
[Epoch 82] speed: 4832 samples/sec	time cost: 274.708243
[Epoch 82] validation: err-top1=0.328484 err-top5=0.120277
Epoch[83] Batch [49]	Speed: 4771.197716 samples/sec	accuracy=0.788791	lr=0.093578
Epoch[83] Batch [99]	Speed: 5020.848306 samples/sec	accuracy=0.789253	lr=0.093393
Epoch[83] Batch [149]	Speed: 4923.543263 samples/sec	accuracy=0.789123	lr=0.093208
Epoch[83] Batch [199]	Speed: 4986.034697 samples/sec	accuracy=0.789059	lr=0.093024
Epoch[83] Batch [249]	Speed: 4626.227083 samples/sec	accuracy=0.788933	lr=0.092839
Epoch[83] Batch [299]	Speed: 4706.827196 samples/sec	accuracy=0.788908	lr=0.092655
Epoch[83] Batch [349]	Speed: 4799.575569 samples/sec	accuracy=0.788904	lr=0.092471
Epoch[83] Batch [399]	Speed: 5013.124209 samples/sec	accuracy=0.789053	lr=0.092287
Epoch[83] Batch [449]	Speed: 5037.360134 samples/sec	accuracy=0.789142	lr=0.092103
Epoch[83] Batch [499]	Speed: 4969.129377 samples/sec	accuracy=0.789135	lr=0.091919
Epoch[83] Batch [549]	Speed: 4897.217730 samples/sec	accuracy=0.789132	lr=0.091735
Epoch[83] Batch [599]	Speed: 4866.761884 samples/sec	accuracy=0.789194	lr=0.091552
Epoch[83] Batch [649]	Speed: 4766.335472 samples/sec	accuracy=0.789262	lr=0.091368
Epoch[83] Batch [699]	Speed: 4985.742405 samples/sec	accuracy=0.789293	lr=0.091185
Epoch[83] Batch [749]	Speed: 4700.504445 samples/sec	accuracy=0.789350	lr=0.091002
Epoch[83] Batch [799]	Speed: 4831.695190 samples/sec	accuracy=0.789351	lr=0.090819
Epoch[83] Batch [849]	Speed: 4762.878479 samples/sec	accuracy=0.789395	lr=0.090636
Epoch[83] Batch [899]	Speed: 4792.397448 samples/sec	accuracy=0.789538	lr=0.090453
Epoch[83] Batch [949]	Speed: 4958.970352 samples/sec	accuracy=0.789584	lr=0.090270
Epoch[83] Batch [999]	Speed: 4842.624112 samples/sec	accuracy=0.789674	lr=0.090088
Epoch[83] Batch [1049]	Speed: 4718.192152 samples/sec	accuracy=0.789752	lr=0.089905
Epoch[83] Batch [1099]	Speed: 4735.184312 samples/sec	accuracy=0.789829	lr=0.089723
Epoch[83] Batch [1149]	Speed: 4995.221410 samples/sec	accuracy=0.789910	lr=0.089541
Epoch[83] Batch [1199]	Speed: 4836.412584 samples/sec	accuracy=0.789984	lr=0.089359
Epoch[83] Batch [1249]	Speed: 4985.883164 samples/sec	accuracy=0.790041	lr=0.089177
[Epoch 83] training: accuracy=0.790042
[Epoch 83] speed: 4853 samples/sec	time cost: 273.184303
[Epoch 83] validation: err-top1=0.312938 err-top5=0.108139
Epoch[84] Batch [49]	Speed: 4771.421397 samples/sec	accuracy=0.790351	lr=0.088992
Epoch[84] Batch [99]	Speed: 4784.949258 samples/sec	accuracy=0.789917	lr=0.088810
Epoch[84] Batch [149]	Speed: 4961.095126 samples/sec	accuracy=0.789884	lr=0.088629
Epoch[84] Batch [199]	Speed: 4812.617648 samples/sec	accuracy=0.789509	lr=0.088448
Epoch[84] Batch [249]	Speed: 5083.880520 samples/sec	accuracy=0.789534	lr=0.088266
Epoch[84] Batch [299]	Speed: 4937.834883 samples/sec	accuracy=0.789467	lr=0.088085
Epoch[84] Batch [349]	Speed: 5016.962598 samples/sec	accuracy=0.789643	lr=0.087904
Epoch[84] Batch [399]	Speed: 4844.576125 samples/sec	accuracy=0.789639	lr=0.087724
Epoch[84] Batch [449]	Speed: 4988.540899 samples/sec	accuracy=0.789788	lr=0.087543
Epoch[84] Batch [499]	Speed: 5137.339530 samples/sec	accuracy=0.789926	lr=0.087363
Epoch[84] Batch [549]	Speed: 4979.050957 samples/sec	accuracy=0.789943	lr=0.087182
Epoch[84] Batch [599]	Speed: 4857.991069 samples/sec	accuracy=0.790045	lr=0.087002
Epoch[84] Batch [649]	Speed: 4764.017817 samples/sec	accuracy=0.790055	lr=0.086822
Epoch[84] Batch [699]	Speed: 4941.095249 samples/sec	accuracy=0.790195	lr=0.086642
Epoch[84] Batch [749]	Speed: 4826.851262 samples/sec	accuracy=0.790324	lr=0.086462
Epoch[84] Batch [799]	Speed: 4927.468032 samples/sec	accuracy=0.790378	lr=0.086282
Epoch[84] Batch [849]	Speed: 4809.289123 samples/sec	accuracy=0.790457	lr=0.086103
Epoch[84] Batch [899]	Speed: 4806.061785 samples/sec	accuracy=0.790617	lr=0.085923
Epoch[84] Batch [949]	Speed: 5037.927137 samples/sec	accuracy=0.790728	lr=0.085744
Epoch[84] Batch [999]	Speed: 5055.287250 samples/sec	accuracy=0.790853	lr=0.085565
Epoch[84] Batch [1049]	Speed: 4894.109908 samples/sec	accuracy=0.791003	lr=0.085386
Epoch[84] Batch [1099]	Speed: 4750.689437 samples/sec	accuracy=0.791096	lr=0.085207
Epoch[84] Batch [1149]	Speed: 4912.167260 samples/sec	accuracy=0.791230	lr=0.085028
Epoch[84] Batch [1199]	Speed: 4937.754839 samples/sec	accuracy=0.791340	lr=0.084850
Epoch[84] Batch [1249]	Speed: 4974.132572 samples/sec	accuracy=0.791444	lr=0.084671
[Epoch 84] training: accuracy=0.791447
[Epoch 84] speed: 4905 samples/sec	time cost: 271.783212
[Epoch 84] validation: err-top1=0.305166 err-top5=0.104632
Epoch[85] Batch [49]	Speed: 4825.492998 samples/sec	accuracy=0.791528	lr=0.084489
Epoch[85] Batch [99]	Speed: 4997.289105 samples/sec	accuracy=0.791189	lr=0.084311
Epoch[85] Batch [149]	Speed: 4971.095776 samples/sec	accuracy=0.791726	lr=0.084133
Epoch[85] Batch [199]	Speed: 4949.329880 samples/sec	accuracy=0.791564	lr=0.083955
Epoch[85] Batch [249]	Speed: 4891.743910 samples/sec	accuracy=0.791213	lr=0.083777
Epoch[85] Batch [299]	Speed: 4715.480244 samples/sec	accuracy=0.791189	lr=0.083600
Epoch[85] Batch [349]	Speed: 5066.148833 samples/sec	accuracy=0.791160	lr=0.083422
Epoch[85] Batch [399]	Speed: 4929.750120 samples/sec	accuracy=0.791302	lr=0.083245
Epoch[85] Batch [449]	Speed: 4751.782470 samples/sec	accuracy=0.791452	lr=0.083067
Epoch[85] Batch [499]	Speed: 4744.240516 samples/sec	accuracy=0.791598	lr=0.082890
Epoch[85] Batch [549]	Speed: 4881.667697 samples/sec	accuracy=0.791631	lr=0.082713
Epoch[85] Batch [599]	Speed: 4730.846695 samples/sec	accuracy=0.791741	lr=0.082537
Epoch[85] Batch [649]	Speed: 4821.541757 samples/sec	accuracy=0.791724	lr=0.082360
Epoch[85] Batch [699]	Speed: 4996.876894 samples/sec	accuracy=0.791815	lr=0.082183
Epoch[85] Batch [749]	Speed: 4965.918045 samples/sec	accuracy=0.791805	lr=0.082007
Epoch[85] Batch [799]	Speed: 4910.178822 samples/sec	accuracy=0.791900	lr=0.081831
Epoch[85] Batch [849]	Speed: 4781.248049 samples/sec	accuracy=0.791971	lr=0.081655
Epoch[85] Batch [899]	Speed: 4714.145534 samples/sec	accuracy=0.792049	lr=0.081479
Epoch[85] Batch [949]	Speed: 4598.889607 samples/sec	accuracy=0.792182	lr=0.081303
Epoch[85] Batch [999]	Speed: 4727.605822 samples/sec	accuracy=0.792319	lr=0.081127
Epoch[85] Batch [1049]	Speed: 4763.890152 samples/sec	accuracy=0.792438	lr=0.080952
Epoch[85] Batch [1099]	Speed: 4816.324394 samples/sec	accuracy=0.792535	lr=0.080776
Epoch[85] Batch [1149]	Speed: 4873.920638 samples/sec	accuracy=0.792640	lr=0.080601
Epoch[85] Batch [1199]	Speed: 4732.958076 samples/sec	accuracy=0.792780	lr=0.080426
Epoch[85] Batch [1249]	Speed: 4972.761904 samples/sec	accuracy=0.792887	lr=0.080251
[Epoch 85] training: accuracy=0.792892
[Epoch 85] speed: 4838 samples/sec	time cost: 275.228450
[Epoch 85] validation: err-top1=0.320472 err-top5=0.114796
Epoch[86] Batch [49]	Speed: 4677.182231 samples/sec	accuracy=0.793321	lr=0.080073
Epoch[86] Batch [99]	Speed: 4588.860892 samples/sec	accuracy=0.792466	lr=0.079898
Epoch[86] Batch [149]	Speed: 4961.517273 samples/sec	accuracy=0.792188	lr=0.079723
Epoch[86] Batch [199]	Speed: 4800.510711 samples/sec	accuracy=0.792340	lr=0.079549
Epoch[86] Batch [249]	Speed: 4609.409702 samples/sec	accuracy=0.792419	lr=0.079375
Epoch[86] Batch [299]	Speed: 4589.181758 samples/sec	accuracy=0.792499	lr=0.079201
Epoch[86] Batch [349]	Speed: 4869.186242 samples/sec	accuracy=0.792725	lr=0.079027
Epoch[86] Batch [399]	Speed: 4878.558184 samples/sec	accuracy=0.792606	lr=0.078853
Epoch[86] Batch [449]	Speed: 4837.186165 samples/sec	accuracy=0.792585	lr=0.078679
Epoch[86] Batch [499]	Speed: 4739.901572 samples/sec	accuracy=0.792642	lr=0.078506
Epoch[86] Batch [549]	Speed: 4848.383839 samples/sec	accuracy=0.792681	lr=0.078332
Epoch[86] Batch [599]	Speed: 4856.123988 samples/sec	accuracy=0.792719	lr=0.078159
Epoch[86] Batch [649]	Speed: 4905.565036 samples/sec	accuracy=0.792797	lr=0.077986
Epoch[86] Batch [699]	Speed: 4787.151371 samples/sec	accuracy=0.792851	lr=0.077813
Epoch[86] Batch [749]	Speed: 4692.175840 samples/sec	accuracy=0.792898	lr=0.077640
Epoch[86] Batch [799]	Speed: 4915.994116 samples/sec	accuracy=0.792965	lr=0.077467
Epoch[86] Batch [849]	Speed: 4783.607118 samples/sec	accuracy=0.793142	lr=0.077295
Epoch[86] Batch [899]	Speed: 4780.401270 samples/sec	accuracy=0.793274	lr=0.077123
Epoch[86] Batch [949]	Speed: 4891.333998 samples/sec	accuracy=0.793355	lr=0.076950
Epoch[86] Batch [999]	Speed: 4804.528368 samples/sec	accuracy=0.793462	lr=0.076778
Epoch[86] Batch [1049]	Speed: 4687.038948 samples/sec	accuracy=0.793567	lr=0.076606
Epoch[86] Batch [1099]	Speed: 5038.699615 samples/sec	accuracy=0.793587	lr=0.076435
Epoch[86] Batch [1149]	Speed: 4661.137596 samples/sec	accuracy=0.793693	lr=0.076263
Epoch[86] Batch [1199]	Speed: 4849.747355 samples/sec	accuracy=0.793772	lr=0.076091
Epoch[86] Batch [1249]	Speed: 4884.920514 samples/sec	accuracy=0.793837	lr=0.075920
[Epoch 86] training: accuracy=0.793840
[Epoch 86] speed: 4790 samples/sec	time cost: 277.108090
[Epoch 86] validation: err-top1=0.324039 err-top5=0.115613
Epoch[87] Batch [49]	Speed: 4555.907538 samples/sec	accuracy=0.792782	lr=0.075745
Epoch[87] Batch [99]	Speed: 4668.793279 samples/sec	accuracy=0.792655	lr=0.075574
Epoch[87] Batch [149]	Speed: 5019.693003 samples/sec	accuracy=0.792563	lr=0.075404
Epoch[87] Batch [199]	Speed: 4690.926630 samples/sec	accuracy=0.792999	lr=0.075233
Epoch[87] Batch [249]	Speed: 4777.561172 samples/sec	accuracy=0.793051	lr=0.075062
Epoch[87] Batch [299]	Speed: 5075.360635 samples/sec	accuracy=0.793166	lr=0.074892
Epoch[87] Batch [349]	Speed: 4823.459473 samples/sec	accuracy=0.793197	lr=0.074721
Epoch[87] Batch [399]	Speed: 5020.373751 samples/sec	accuracy=0.793216	lr=0.074551
Epoch[87] Batch [449]	Speed: 4854.884417 samples/sec	accuracy=0.793311	lr=0.074381
Epoch[87] Batch [499]	Speed: 4776.613168 samples/sec	accuracy=0.793260	lr=0.074211
Epoch[87] Batch [549]	Speed: 4920.844938 samples/sec	accuracy=0.793407	lr=0.074042
Epoch[87] Batch [599]	Speed: 5060.629270 samples/sec	accuracy=0.793570	lr=0.073872
Epoch[87] Batch [649]	Speed: 4954.858526 samples/sec	accuracy=0.793675	lr=0.073703
Epoch[87] Batch [699]	Speed: 5075.140894 samples/sec	accuracy=0.793700	lr=0.073534
Epoch[87] Batch [749]	Speed: 4878.778522 samples/sec	accuracy=0.793726	lr=0.073364
Epoch[87] Batch [799]	Speed: 4923.469326 samples/sec	accuracy=0.793777	lr=0.073196
Epoch[87] Batch [849]	Speed: 4879.512829 samples/sec	accuracy=0.793775	lr=0.073027
Epoch[87] Batch [899]	Speed: 4905.942481 samples/sec	accuracy=0.793851	lr=0.072858
Epoch[87] Batch [949]	Speed: 4818.760490 samples/sec	accuracy=0.793968	lr=0.072690
Epoch[87] Batch [999]	Speed: 4980.203102 samples/sec	accuracy=0.794080	lr=0.072521
Epoch[87] Batch [1049]	Speed: 4887.287367 samples/sec	accuracy=0.794220	lr=0.072353
Epoch[87] Batch [1099]	Speed: 4801.941488 samples/sec	accuracy=0.794298	lr=0.072185
Epoch[87] Batch [1149]	Speed: 4962.935425 samples/sec	accuracy=0.794377	lr=0.072017
Epoch[87] Batch [1199]	Speed: 4782.367225 samples/sec	accuracy=0.794447	lr=0.071849
Epoch[87] Batch [1249]	Speed: 4971.975782 samples/sec	accuracy=0.794610	lr=0.071682
[Epoch 87] training: accuracy=0.794612
[Epoch 87] speed: 4874 samples/sec	time cost: 272.091917
[Epoch 87] validation: err-top1=0.310303 err-top5=0.109090
Epoch[88] Batch [49]	Speed: 4839.354730 samples/sec	accuracy=0.795990	lr=0.071511
Epoch[88] Batch [99]	Speed: 4870.877438 samples/sec	accuracy=0.795155	lr=0.071344
Epoch[88] Batch [149]	Speed: 4794.116517 samples/sec	accuracy=0.795253	lr=0.071177
Epoch[88] Batch [199]	Speed: 4939.966921 samples/sec	accuracy=0.795225	lr=0.071010
Epoch[88] Batch [249]	Speed: 4745.284340 samples/sec	accuracy=0.795105	lr=0.070843
Epoch[88] Batch [299]	Speed: 5027.389939 samples/sec	accuracy=0.795167	lr=0.070676
Epoch[88] Batch [349]	Speed: 4920.962774 samples/sec	accuracy=0.795026	lr=0.070510
Epoch[88] Batch [399]	Speed: 4915.473015 samples/sec	accuracy=0.795279	lr=0.070343
Epoch[88] Batch [449]	Speed: 5043.985173 samples/sec	accuracy=0.795351	lr=0.070177
Epoch[88] Batch [499]	Speed: 4857.837988 samples/sec	accuracy=0.795281	lr=0.070011
Epoch[88] Batch [549]	Speed: 4720.747881 samples/sec	accuracy=0.795338	lr=0.069845
Epoch[88] Batch [599]	Speed: 4726.284004 samples/sec	accuracy=0.795327	lr=0.069680
Epoch[88] Batch [649]	Speed: 4624.965026 samples/sec	accuracy=0.795407	lr=0.069514
Epoch[88] Batch [699]	Speed: 5026.724113 samples/sec	accuracy=0.795413	lr=0.069349
Epoch[88] Batch [749]	Speed: 4415.768776 samples/sec	accuracy=0.795372	lr=0.069183
Epoch[88] Batch [799]	Speed: 4955.351534 samples/sec	accuracy=0.795434	lr=0.069018
Epoch[88] Batch [849]	Speed: 4702.955969 samples/sec	accuracy=0.795491	lr=0.068853
Epoch[88] Batch [899]	Speed: 4842.120851 samples/sec	accuracy=0.795506	lr=0.068689
Epoch[88] Batch [949]	Speed: 4947.611361 samples/sec	accuracy=0.795576	lr=0.068524
Epoch[88] Batch [999]	Speed: 4934.862191 samples/sec	accuracy=0.795662	lr=0.068359
Epoch[88] Batch [1049]	Speed: 4688.427031 samples/sec	accuracy=0.795692	lr=0.068195
Epoch[88] Batch [1099]	Speed: 4883.256753 samples/sec	accuracy=0.795801	lr=0.068031
Epoch[88] Batch [1149]	Speed: 4894.782678 samples/sec	accuracy=0.795920	lr=0.067867
Epoch[88] Batch [1199]	Speed: 4729.947245 samples/sec	accuracy=0.796026	lr=0.067703
Epoch[88] Batch [1249]	Speed: 4978.627899 samples/sec	accuracy=0.796085	lr=0.067539
[Epoch 88] training: accuracy=0.796085
[Epoch 88] speed: 4831 samples/sec	time cost: 274.917370
[Epoch 88] validation: err-top1=0.323222 err-top5=0.117287
Epoch[89] Batch [49]	Speed: 4450.521016 samples/sec	accuracy=0.796041	lr=0.067373
Epoch[89] Batch [99]	Speed: 4772.834566 samples/sec	accuracy=0.795865	lr=0.067209
Epoch[89] Batch [149]	Speed: 4661.710898 samples/sec	accuracy=0.795867	lr=0.067046
Epoch[89] Batch [199]	Speed: 4728.184871 samples/sec	accuracy=0.795847	lr=0.066883
Epoch[89] Batch [249]	Speed: 4803.141061 samples/sec	accuracy=0.795873	lr=0.066720
Epoch[89] Batch [299]	Speed: 4942.691838 samples/sec	accuracy=0.796050	lr=0.066557
Epoch[89] Batch [349]	Speed: 4751.222540 samples/sec	accuracy=0.796027	lr=0.066395
Epoch[89] Batch [399]	Speed: 4930.461251 samples/sec	accuracy=0.796063	lr=0.066232
Epoch[89] Batch [449]	Speed: 5241.508071 samples/sec	accuracy=0.796191	lr=0.066070
Epoch[89] Batch [499]	Speed: 4770.227441 samples/sec	accuracy=0.796314	lr=0.065908
Epoch[89] Batch [549]	Speed: 5024.763200 samples/sec	accuracy=0.796416	lr=0.065746
Epoch[89] Batch [599]	Speed: 5052.655181 samples/sec	accuracy=0.796365	lr=0.065584
Epoch[89] Batch [649]	Speed: 4907.911561 samples/sec	accuracy=0.796401	lr=0.065423
Epoch[89] Batch [699]	Speed: 5068.186922 samples/sec	accuracy=0.796534	lr=0.065261
Epoch[89] Batch [749]	Speed: 5161.072341 samples/sec	accuracy=0.796556	lr=0.065100
Epoch[89] Batch [799]	Speed: 4893.914838 samples/sec	accuracy=0.796582	lr=0.064939
Epoch[89] Batch [849]	Speed: 4916.157862 samples/sec	accuracy=0.796689	lr=0.064778
Epoch[89] Batch [899]	Speed: 4869.512285 samples/sec	accuracy=0.796690	lr=0.064617
Epoch[89] Batch [949]	Speed: 4990.669870 samples/sec	accuracy=0.796722	lr=0.064456
Epoch[89] Batch [999]	Speed: 4889.967600 samples/sec	accuracy=0.796769	lr=0.064296
Epoch[89] Batch [1049]	Speed: 4973.331503 samples/sec	accuracy=0.796873	lr=0.064135
Epoch[89] Batch [1099]	Speed: 5030.119220 samples/sec	accuracy=0.796958	lr=0.063975
Epoch[89] Batch [1149]	Speed: 5007.613141 samples/sec	accuracy=0.797007	lr=0.063815
Epoch[89] Batch [1199]	Speed: 4714.239293 samples/sec	accuracy=0.797069	lr=0.063655
Epoch[89] Batch [1249]	Speed: 5035.033781 samples/sec	accuracy=0.797187	lr=0.063496
[Epoch 89] training: accuracy=0.797189
[Epoch 89] speed: 4892 samples/sec	time cost: 270.956270
[Epoch 89] validation: err-top1=0.295420 err-top5=0.100088
Epoch[90] Batch [49]	Speed: 4588.639685 samples/sec	accuracy=0.797466	lr=0.063333
Epoch[90] Batch [99]	Speed: 4671.589536 samples/sec	accuracy=0.797204	lr=0.063174
Epoch[90] Batch [149]	Speed: 4826.307127 samples/sec	accuracy=0.797154	lr=0.063015
Epoch[90] Batch [199]	Speed: 4822.216921 samples/sec	accuracy=0.797275	lr=0.062855
Epoch[90] Batch [249]	Speed: 4699.637271 samples/sec	accuracy=0.797152	lr=0.062697
Epoch[90] Batch [299]	Speed: 4842.078271 samples/sec	accuracy=0.797357	lr=0.062538
Epoch[90] Batch [349]	Speed: 4974.358747 samples/sec	accuracy=0.797328	lr=0.062379
Epoch[90] Batch [399]	Speed: 4865.517319 samples/sec	accuracy=0.797486	lr=0.062221
Epoch[90] Batch [449]	Speed: 4885.736258 samples/sec	accuracy=0.797530	lr=0.062063
Epoch[90] Batch [499]	Speed: 4903.277515 samples/sec	accuracy=0.797660	lr=0.061905
Epoch[90] Batch [549]	Speed: 4943.075473 samples/sec	accuracy=0.797761	lr=0.061747
Epoch[90] Batch [599]	Speed: 5024.680784 samples/sec	accuracy=0.797811	lr=0.061589
Epoch[90] Batch [649]	Speed: 5036.888832 samples/sec	accuracy=0.797868	lr=0.061432
Epoch[90] Batch [699]	Speed: 4881.897638 samples/sec	accuracy=0.797956	lr=0.061274
Epoch[90] Batch [749]	Speed: 4856.304086 samples/sec	accuracy=0.798025	lr=0.061117
Epoch[90] Batch [799]	Speed: 4884.220126 samples/sec	accuracy=0.798035	lr=0.060960
Epoch[90] Batch [849]	Speed: 4951.759461 samples/sec	accuracy=0.798116	lr=0.060803
Epoch[90] Batch [899]	Speed: 4914.810068 samples/sec	accuracy=0.798141	lr=0.060646
Epoch[90] Batch [949]	Speed: 4905.099250 samples/sec	accuracy=0.798217	lr=0.060490
Epoch[90] Batch [999]	Speed: 4886.156016 samples/sec	accuracy=0.798289	lr=0.060333
Epoch[90] Batch [1049]	Speed: 4905.103059 samples/sec	accuracy=0.798338	lr=0.060177
Epoch[90] Batch [1099]	Speed: 4997.418306 samples/sec	accuracy=0.798412	lr=0.060021
Epoch[90] Batch [1149]	Speed: 4916.126800 samples/sec	accuracy=0.798488	lr=0.059865
Epoch[90] Batch [1199]	Speed: 4942.568637 samples/sec	accuracy=0.798552	lr=0.059710
Epoch[90] Batch [1249]	Speed: 4875.280406 samples/sec	accuracy=0.798647	lr=0.059554
[Epoch 90] training: accuracy=0.798648
[Epoch 90] speed: 4874 samples/sec	time cost: 273.059940
[Epoch 90] validation: err-top1=0.285375 err-top5=0.094448
Epoch[91] Batch [49]	Speed: 4409.777167 samples/sec	accuracy=0.799816	lr=0.059396
Epoch[91] Batch [99]	Speed: 4844.102179 samples/sec	accuracy=0.799520	lr=0.059240
Epoch[91] Batch [149]	Speed: 4614.934653 samples/sec	accuracy=0.799379	lr=0.059085
Epoch[91] Batch [199]	Speed: 4614.123248 samples/sec	accuracy=0.799102	lr=0.058930
Epoch[91] Batch [249]	Speed: 4843.714961 samples/sec	accuracy=0.798930	lr=0.058776
Epoch[91] Batch [299]	Speed: 4756.068941 samples/sec	accuracy=0.798873	lr=0.058621
Epoch[91] Batch [349]	Speed: 4812.609990 samples/sec	accuracy=0.798885	lr=0.058467
Epoch[91] Batch [399]	Speed: 4950.609481 samples/sec	accuracy=0.798699	lr=0.058313
Epoch[91] Batch [449]	Speed: 4954.848922 samples/sec	accuracy=0.798789	lr=0.058158
Epoch[91] Batch [499]	Speed: 4877.520492 samples/sec	accuracy=0.798735	lr=0.058005
Epoch[91] Batch [549]	Speed: 4839.249494 samples/sec	accuracy=0.798629	lr=0.057851
Epoch[91] Batch [599]	Speed: 4814.727101 samples/sec	accuracy=0.798693	lr=0.057697
Epoch[91] Batch [649]	Speed: 5057.513964 samples/sec	accuracy=0.798675	lr=0.057544
Epoch[91] Batch [699]	Speed: 4822.540819 samples/sec	accuracy=0.798696	lr=0.057391
Epoch[91] Batch [749]	Speed: 4878.480605 samples/sec	accuracy=0.798788	lr=0.057238
Epoch[91] Batch [799]	Speed: 4753.592682 samples/sec	accuracy=0.798778	lr=0.057085
Epoch[91] Batch [849]	Speed: 4973.797783 samples/sec	accuracy=0.798847	lr=0.056932
Epoch[91] Batch [899]	Speed: 4937.549123 samples/sec	accuracy=0.798916	lr=0.056780
Epoch[91] Batch [949]	Speed: 4880.781205 samples/sec	accuracy=0.799026	lr=0.056627
Epoch[91] Batch [999]	Speed: 4760.724388 samples/sec	accuracy=0.799071	lr=0.056475
Epoch[91] Batch [1049]	Speed: 4869.510408 samples/sec	accuracy=0.799127	lr=0.056323
Epoch[91] Batch [1099]	Speed: 4948.643287 samples/sec	accuracy=0.799181	lr=0.056171
Epoch[91] Batch [1149]	Speed: 4896.381290 samples/sec	accuracy=0.799260	lr=0.056020
Epoch[91] Batch [1199]	Speed: 4811.523079 samples/sec	accuracy=0.799342	lr=0.055868
Epoch[91] Batch [1249]	Speed: 5149.910659 samples/sec	accuracy=0.799369	lr=0.055717
[Epoch 91] training: accuracy=0.799367
[Epoch 91] speed: 4834 samples/sec	time cost: 274.939725
[Epoch 91] validation: err-top1=0.297553 err-top5=0.100207
Epoch[92] Batch [49]	Speed: 4615.810036 samples/sec	accuracy=0.800947	lr=0.055563
Epoch[92] Batch [99]	Speed: 4848.925190 samples/sec	accuracy=0.799677	lr=0.055412
Epoch[92] Batch [149]	Speed: 4879.404842 samples/sec	accuracy=0.799960	lr=0.055261
Epoch[92] Batch [199]	Speed: 4776.579276 samples/sec	accuracy=0.799931	lr=0.055111
Epoch[92] Batch [249]	Speed: 4830.401019 samples/sec	accuracy=0.799940	lr=0.054960
Epoch[92] Batch [299]	Speed: 4954.921861 samples/sec	accuracy=0.799793	lr=0.054810
Epoch[92] Batch [349]	Speed: 4811.987113 samples/sec	accuracy=0.799764	lr=0.054660
Epoch[92] Batch [399]	Speed: 4820.597972 samples/sec	accuracy=0.799699	lr=0.054510
Epoch[92] Batch [449]	Speed: 4774.200385 samples/sec	accuracy=0.799797	lr=0.054360
Epoch[92] Batch [499]	Speed: 4850.636413 samples/sec	accuracy=0.799816	lr=0.054210
Epoch[92] Batch [549]	Speed: 5033.085254 samples/sec	accuracy=0.799809	lr=0.054061
Epoch[92] Batch [599]	Speed: 4899.982547 samples/sec	accuracy=0.799891	lr=0.053912
Epoch[92] Batch [649]	Speed: 4811.229655 samples/sec	accuracy=0.799999	lr=0.053763
Epoch[92] Batch [699]	Speed: 4946.327955 samples/sec	accuracy=0.800097	lr=0.053614
Epoch[92] Batch [749]	Speed: 4929.037270 samples/sec	accuracy=0.800199	lr=0.053465
Epoch[92] Batch [799]	Speed: 4888.657943 samples/sec	accuracy=0.800271	lr=0.053317
Epoch[92] Batch [849]	Speed: 5061.747420 samples/sec	accuracy=0.800333	lr=0.053168
Epoch[92] Batch [899]	Speed: 4824.212985 samples/sec	accuracy=0.800383	lr=0.053020
Epoch[92] Batch [949]	Speed: 4978.944522 samples/sec	accuracy=0.800514	lr=0.052872
Epoch[92] Batch [999]	Speed: 4748.965659 samples/sec	accuracy=0.800616	lr=0.052724
Epoch[92] Batch [1049]	Speed: 4769.038849 samples/sec	accuracy=0.800648	lr=0.052577
Epoch[92] Batch [1099]	Speed: 4887.244990 samples/sec	accuracy=0.800704	lr=0.052429
Epoch[92] Batch [1149]	Speed: 4857.487356 samples/sec	accuracy=0.800806	lr=0.052282
Epoch[92] Batch [1199]	Speed: 4844.883359 samples/sec	accuracy=0.800851	lr=0.052135
Epoch[92] Batch [1249]	Speed: 4955.899767 samples/sec	accuracy=0.800898	lr=0.051988
[Epoch 92] training: accuracy=0.800898
[Epoch 92] speed: 4858 samples/sec	time cost: 273.833269
[Epoch 92] validation: err-top1=0.295918 err-top5=0.100187
Epoch[93] Batch [49]	Speed: 4558.540852 samples/sec	accuracy=0.800982	lr=0.051838
Epoch[93] Batch [99]	Speed: 4714.252747 samples/sec	accuracy=0.800802	lr=0.051691
Epoch[93] Batch [149]	Speed: 4655.813111 samples/sec	accuracy=0.800771	lr=0.051545
Epoch[93] Batch [199]	Speed: 4935.793282 samples/sec	accuracy=0.800656	lr=0.051399
Epoch[93] Batch [249]	Speed: 4933.902998 samples/sec	accuracy=0.800558	lr=0.051253
Epoch[93] Batch [299]	Speed: 4858.200321 samples/sec	accuracy=0.800733	lr=0.051107
Epoch[93] Batch [349]	Speed: 4725.464794 samples/sec	accuracy=0.800814	lr=0.050961
Epoch[93] Batch [399]	Speed: 4771.422351 samples/sec	accuracy=0.800877	lr=0.050816
Epoch[93] Batch [449]	Speed: 4883.984412 samples/sec	accuracy=0.801003	lr=0.050670
Epoch[93] Batch [499]	Speed: 4902.718812 samples/sec	accuracy=0.801095	lr=0.050525
Epoch[93] Batch [549]	Speed: 4799.225575 samples/sec	accuracy=0.801121	lr=0.050380
Epoch[93] Batch [599]	Speed: 4812.282032 samples/sec	accuracy=0.801201	lr=0.050235
Epoch[93] Batch [649]	Speed: 4965.139365 samples/sec	accuracy=0.801286	lr=0.050091
Epoch[93] Batch [699]	Speed: 4768.300250 samples/sec	accuracy=0.801418	lr=0.049946
Epoch[93] Batch [749]	Speed: 4851.657764 samples/sec	accuracy=0.801499	lr=0.049802
Epoch[93] Batch [799]	Speed: 4734.969341 samples/sec	accuracy=0.801541	lr=0.049658
Epoch[93] Batch [849]	Speed: 4782.717109 samples/sec	accuracy=0.801635	lr=0.049514
Epoch[93] Batch [899]	Speed: 4928.756374 samples/sec	accuracy=0.801699	lr=0.049370
Epoch[93] Batch [949]	Speed: 4747.117727 samples/sec	accuracy=0.801793	lr=0.049227
Epoch[93] Batch [999]	Speed: 4763.962438 samples/sec	accuracy=0.801839	lr=0.049083
Epoch[93] Batch [1049]	Speed: 4826.089226 samples/sec	accuracy=0.801933	lr=0.048940
Epoch[93] Batch [1099]	Speed: 4964.882691 samples/sec	accuracy=0.801948	lr=0.048797
Epoch[93] Batch [1149]	Speed: 4736.313990 samples/sec	accuracy=0.801959	lr=0.048654
Epoch[93] Batch [1199]	Speed: 4686.651577 samples/sec	accuracy=0.801942	lr=0.048511
Epoch[93] Batch [1249]	Speed: 5044.386708 samples/sec	accuracy=0.801974	lr=0.048369
[Epoch 93] training: accuracy=0.801972
[Epoch 93] speed: 4808 samples/sec	time cost: 275.859121
[Epoch 93] validation: err-top1=0.299866 err-top5=0.101949
Epoch[94] Batch [49]	Speed: 4799.806746 samples/sec	accuracy=0.802314	lr=0.048224
Epoch[94] Batch [99]	Speed: 4806.357699 samples/sec	accuracy=0.801878	lr=0.048082
Epoch[94] Batch [149]	Speed: 4806.591466 samples/sec	accuracy=0.801896	lr=0.047940
Epoch[94] Batch [199]	Speed: 4992.788239 samples/sec	accuracy=0.802119	lr=0.047798
Epoch[94] Batch [249]	Speed: 4919.266374 samples/sec	accuracy=0.802049	lr=0.047656
Epoch[94] Batch [299]	Speed: 4768.559977 samples/sec	accuracy=0.802318	lr=0.047515
Epoch[94] Batch [349]	Speed: 5050.248053 samples/sec	accuracy=0.802458	lr=0.047374
Epoch[94] Batch [399]	Speed: 4838.685553 samples/sec	accuracy=0.802509	lr=0.047233
Epoch[94] Batch [449]	Speed: 4945.523402 samples/sec	accuracy=0.802571	lr=0.047092
Epoch[94] Batch [499]	Speed: 4939.291557 samples/sec	accuracy=0.802578	lr=0.046951
Epoch[94] Batch [549]	Speed: 4991.325714 samples/sec	accuracy=0.802673	lr=0.046811
Epoch[94] Batch [599]	Speed: 4972.290178 samples/sec	accuracy=0.802702	lr=0.046670
Epoch[94] Batch [649]	Speed: 4857.848098 samples/sec	accuracy=0.802702	lr=0.046530
Epoch[94] Batch [699]	Speed: 4980.081142 samples/sec	accuracy=0.802785	lr=0.046390
Epoch[94] Batch [749]	Speed: 4912.814321 samples/sec	accuracy=0.802882	lr=0.046251
Epoch[94] Batch [799]	Speed: 4712.522832 samples/sec	accuracy=0.802914	lr=0.046111
Epoch[94] Batch [849]	Speed: 5036.418090 samples/sec	accuracy=0.802867	lr=0.045972
Epoch[94] Batch [899]	Speed: 4792.807525 samples/sec	accuracy=0.802946	lr=0.045833
Epoch[94] Batch [949]	Speed: 5126.085716 samples/sec	accuracy=0.802974	lr=0.045693
Epoch[94] Batch [999]	Speed: 4852.125734 samples/sec	accuracy=0.802994	lr=0.045555
Epoch[94] Batch [1049]	Speed: 4997.724181 samples/sec	accuracy=0.803084	lr=0.045416
Epoch[94] Batch [1099]	Speed: 4711.391760 samples/sec	accuracy=0.803211	lr=0.045278
Epoch[94] Batch [1149]	Speed: 4921.356351 samples/sec	accuracy=0.803317	lr=0.045139
Epoch[94] Batch [1199]	Speed: 4801.654706 samples/sec	accuracy=0.803405	lr=0.045001
Epoch[94] Batch [1249]	Speed: 4943.706007 samples/sec	accuracy=0.803467	lr=0.044863
[Epoch 94] training: accuracy=0.803468
[Epoch 94] speed: 4892 samples/sec	time cost: 271.373185
[Epoch 94] validation: err-top1=0.281509 err-top5=0.091000
Epoch[95] Batch [49]	Speed: 4776.411204 samples/sec	accuracy=0.803695	lr=0.044723
Epoch[95] Batch [99]	Speed: 4926.655361 samples/sec	accuracy=0.803100	lr=0.044585
Epoch[95] Batch [149]	Speed: 4792.666225 samples/sec	accuracy=0.802949	lr=0.044448
Epoch[95] Batch [199]	Speed: 4999.257492 samples/sec	accuracy=0.802804	lr=0.044311
Epoch[95] Batch [249]	Speed: 5094.188764 samples/sec	accuracy=0.802960	lr=0.044174
Epoch[95] Batch [299]	Speed: 4595.871603 samples/sec	accuracy=0.802882	lr=0.044037
Epoch[95] Batch [349]	Speed: 5095.954645 samples/sec	accuracy=0.802986	lr=0.043900
Epoch[95] Batch [399]	Speed: 5182.089135 samples/sec	accuracy=0.803047	lr=0.043764
Epoch[95] Batch [449]	Speed: 4982.954333 samples/sec	accuracy=0.803066	lr=0.043628
Epoch[95] Batch [499]	Speed: 4905.313588 samples/sec	accuracy=0.803153	lr=0.043492
Epoch[95] Batch [549]	Speed: 4950.058082 samples/sec	accuracy=0.803189	lr=0.043356
Epoch[95] Batch [599]	Speed: 5058.740726 samples/sec	accuracy=0.803230	lr=0.043220
Epoch[95] Batch [649]	Speed: 4815.362349 samples/sec	accuracy=0.803292	lr=0.043085
Epoch[95] Batch [699]	Speed: 4912.064452 samples/sec	accuracy=0.803373	lr=0.042949
Epoch[95] Batch [749]	Speed: 4618.968937 samples/sec	accuracy=0.803531	lr=0.042814
Epoch[95] Batch [799]	Speed: 4916.125337 samples/sec	accuracy=0.803596	lr=0.042679
Epoch[95] Batch [849]	Speed: 4855.330616 samples/sec	accuracy=0.803751	lr=0.042545
Epoch[95] Batch [899]	Speed: 5011.861569 samples/sec	accuracy=0.803817	lr=0.042410
Epoch[95] Batch [949]	Speed: 4893.457394 samples/sec	accuracy=0.803889	lr=0.042276
Epoch[95] Batch [999]	Speed: 4835.066679 samples/sec	accuracy=0.803972	lr=0.042141
Epoch[95] Batch [1049]	Speed: 4834.131629 samples/sec	accuracy=0.804028	lr=0.042007
Epoch[95] Batch [1099]	Speed: 4967.855355 samples/sec	accuracy=0.804107	lr=0.041874
Epoch[95] Batch [1149]	Speed: 5000.330520 samples/sec	accuracy=0.804152	lr=0.041740
Epoch[95] Batch [1199]	Speed: 4824.211576 samples/sec	accuracy=0.804234	lr=0.041607
Epoch[95] Batch [1249]	Speed: 5129.390850 samples/sec	accuracy=0.804309	lr=0.041473
[Epoch 95] training: accuracy=0.804309
[Epoch 95] speed: 4910 samples/sec	time cost: 270.319749
[Epoch 95] validation: err-top1=0.277982 err-top5=0.088369
Epoch[96] Batch [49]	Speed: 4791.676722 samples/sec	accuracy=0.803381	lr=0.041338
Epoch[96] Batch [99]	Speed: 4902.802872 samples/sec	accuracy=0.803773	lr=0.041205
Epoch[96] Batch [149]	Speed: 4913.748804 samples/sec	accuracy=0.803942	lr=0.041072
Epoch[96] Batch [199]	Speed: 4754.652943 samples/sec	accuracy=0.804126	lr=0.040940
Epoch[96] Batch [249]	Speed: 5079.460871 samples/sec	accuracy=0.804111	lr=0.040807
Epoch[96] Batch [299]	Speed: 4857.347272 samples/sec	accuracy=0.804160	lr=0.040675
Epoch[96] Batch [349]	Speed: 4846.758652 samples/sec	accuracy=0.804287	lr=0.040543
Epoch[96] Batch [399]	Speed: 4903.544206 samples/sec	accuracy=0.804484	lr=0.040412
Epoch[96] Batch [449]	Speed: 4848.736442 samples/sec	accuracy=0.804513	lr=0.040280
Epoch[96] Batch [499]	Speed: 5049.580552 samples/sec	accuracy=0.804462	lr=0.040149
Epoch[96] Batch [549]	Speed: 4811.356312 samples/sec	accuracy=0.804479	lr=0.040018
Epoch[96] Batch [599]	Speed: 5033.688815 samples/sec	accuracy=0.804590	lr=0.039887
Epoch[96] Batch [649]	Speed: 4976.085177 samples/sec	accuracy=0.804726	lr=0.039756
Epoch[96] Batch [699]	Speed: 4848.849426 samples/sec	accuracy=0.804842	lr=0.039625
Epoch[96] Batch [749]	Speed: 4697.471864 samples/sec	accuracy=0.804888	lr=0.039495
Epoch[96] Batch [799]	Speed: 4929.136944 samples/sec	accuracy=0.804983	lr=0.039365
Epoch[96] Batch [849]	Speed: 4858.714295 samples/sec	accuracy=0.805037	lr=0.039235
Epoch[96] Batch [899]	Speed: 4874.567952 samples/sec	accuracy=0.805011	lr=0.039105
Epoch[96] Batch [949]	Speed: 4856.060078 samples/sec	accuracy=0.805035	lr=0.038975
Epoch[96] Batch [999]	Speed: 4994.772481 samples/sec	accuracy=0.805087	lr=0.038846
Epoch[96] Batch [1049]	Speed: 4801.289810 samples/sec	accuracy=0.805163	lr=0.038717
Epoch[96] Batch [1099]	Speed: 4909.690944 samples/sec	accuracy=0.805223	lr=0.038588
Epoch[96] Batch [1149]	Speed: 4983.741040 samples/sec	accuracy=0.805313	lr=0.038459
Epoch[96] Batch [1199]	Speed: 4992.514654 samples/sec	accuracy=0.805414	lr=0.038330
Epoch[96] Batch [1249]	Speed: 4968.584766 samples/sec	accuracy=0.805493	lr=0.038202
[Epoch 96] training: accuracy=0.805496
[Epoch 96] speed: 4893 samples/sec	time cost: 271.972081
[Epoch 96] validation: err-top1=0.274633 err-top5=0.088827
Epoch[97] Batch [49]	Speed: 4696.923633 samples/sec	accuracy=0.806983	lr=0.038071
Epoch[97] Batch [99]	Speed: 4720.395385 samples/sec	accuracy=0.806735	lr=0.037943
Epoch[97] Batch [149]	Speed: 4761.855198 samples/sec	accuracy=0.806183	lr=0.037815
Epoch[97] Batch [199]	Speed: 4825.376546 samples/sec	accuracy=0.806247	lr=0.037687
Epoch[97] Batch [249]	Speed: 4810.988215 samples/sec	accuracy=0.806152	lr=0.037560
Epoch[97] Batch [299]	Speed: 4706.535983 samples/sec	accuracy=0.806032	lr=0.037432
Epoch[97] Batch [349]	Speed: 4784.871856 samples/sec	accuracy=0.805934	lr=0.037305
Epoch[97] Batch [399]	Speed: 4762.941755 samples/sec	accuracy=0.805991	lr=0.037178
Epoch[97] Batch [449]	Speed: 4924.467152 samples/sec	accuracy=0.805918	lr=0.037052
Epoch[97] Batch [499]	Speed: 4935.967880 samples/sec	accuracy=0.805902	lr=0.036925
Epoch[97] Batch [549]	Speed: 4964.336715 samples/sec	accuracy=0.805927	lr=0.036799
Epoch[97] Batch [599]	Speed: 4879.316482 samples/sec	accuracy=0.805922	lr=0.036673
Epoch[97] Batch [649]	Speed: 4781.067727 samples/sec	accuracy=0.805969	lr=0.036547
Epoch[97] Batch [699]	Speed: 4832.107888 samples/sec	accuracy=0.806019	lr=0.036421
Epoch[97] Batch [749]	Speed: 4961.354160 samples/sec	accuracy=0.806080	lr=0.036296
Epoch[97] Batch [799]	Speed: 4867.915057 samples/sec	accuracy=0.806147	lr=0.036170
Epoch[97] Batch [849]	Speed: 4713.150532 samples/sec	accuracy=0.806180	lr=0.036045
Epoch[97] Batch [899]	Speed: 4845.412887 samples/sec	accuracy=0.806237	lr=0.035920
Epoch[97] Batch [949]	Speed: 5067.787090 samples/sec	accuracy=0.806286	lr=0.035795
Epoch[97] Batch [999]	Speed: 5040.241141 samples/sec	accuracy=0.806356	lr=0.035671
Epoch[97] Batch [1049]	Speed: 4695.991441 samples/sec	accuracy=0.806426	lr=0.035546
Epoch[97] Batch [1099]	Speed: 4692.597758 samples/sec	accuracy=0.806489	lr=0.035422
Epoch[97] Batch [1149]	Speed: 4873.659925 samples/sec	accuracy=0.806559	lr=0.035298
Epoch[97] Batch [1199]	Speed: 4917.444124 samples/sec	accuracy=0.806620	lr=0.035174
Epoch[97] Batch [1249]	Speed: 5052.908051 samples/sec	accuracy=0.806671	lr=0.035051
[Epoch 97] training: accuracy=0.806672
[Epoch 97] speed: 4837 samples/sec	time cost: 274.989807
[Epoch 97] validation: err-top1=0.271983 err-top5=0.086575
Epoch[98] Batch [49]	Speed: 4797.901999 samples/sec	accuracy=0.806946	lr=0.034925
Epoch[98] Batch [99]	Speed: 4673.332644 samples/sec	accuracy=0.806640	lr=0.034802
Epoch[98] Batch [149]	Speed: 4971.454485 samples/sec	accuracy=0.806734	lr=0.034679
Epoch[98] Batch [199]	Speed: 4916.638471 samples/sec	accuracy=0.806588	lr=0.034556
Epoch[98] Batch [249]	Speed: 4936.669344 samples/sec	accuracy=0.806546	lr=0.034433
Epoch[98] Batch [299]	Speed: 4745.767358 samples/sec	accuracy=0.806501	lr=0.034311
Epoch[98] Batch [349]	Speed: 4770.108237 samples/sec	accuracy=0.806523	lr=0.034189
Epoch[98] Batch [399]	Speed: 5022.896861 samples/sec	accuracy=0.806535	lr=0.034067
Epoch[98] Batch [449]	Speed: 4915.750487 samples/sec	accuracy=0.806572	lr=0.033945
Epoch[98] Batch [499]	Speed: 4800.773101 samples/sec	accuracy=0.806629	lr=0.033823
Epoch[98] Batch [549]	Speed: 4867.738730 samples/sec	accuracy=0.806673	lr=0.033702
Epoch[98] Batch [599]	Speed: 4726.717696 samples/sec	accuracy=0.806751	lr=0.033581
Epoch[98] Batch [649]	Speed: 4856.844900 samples/sec	accuracy=0.806820	lr=0.033460
Epoch[98] Batch [699]	Speed: 4982.374783 samples/sec	accuracy=0.806891	lr=0.033339
Epoch[98] Batch [749]	Speed: 4769.436357 samples/sec	accuracy=0.806999	lr=0.033218
Epoch[98] Batch [799]	Speed: 5001.442331 samples/sec	accuracy=0.807015	lr=0.033098
Epoch[98] Batch [849]	Speed: 4950.580721 samples/sec	accuracy=0.807060	lr=0.032978
Epoch[98] Batch [899]	Speed: 4840.667780 samples/sec	accuracy=0.807160	lr=0.032858
Epoch[98] Batch [949]	Speed: 4810.452176 samples/sec	accuracy=0.807213	lr=0.032738
Epoch[98] Batch [999]	Speed: 4788.248543 samples/sec	accuracy=0.807274	lr=0.032618
Epoch[98] Batch [1049]	Speed: 4815.747855 samples/sec	accuracy=0.807385	lr=0.032499
Epoch[98] Batch [1099]	Speed: 4899.049046 samples/sec	accuracy=0.807414	lr=0.032380
Epoch[98] Batch [1149]	Speed: 4891.392489 samples/sec	accuracy=0.807459	lr=0.032261
Epoch[98] Batch [1199]	Speed: 4853.913710 samples/sec	accuracy=0.807507	lr=0.032142
Epoch[98] Batch [1249]	Speed: 4998.729296 samples/sec	accuracy=0.807580	lr=0.032023
[Epoch 98] training: accuracy=0.807582
[Epoch 98] speed: 4858 samples/sec	time cost: 272.902919
[Epoch 98] validation: err-top1=0.274068 err-top5=0.086304
Epoch[99] Batch [49]	Speed: 4329.296801 samples/sec	accuracy=0.809019	lr=0.031902
Epoch[99] Batch [99]	Speed: 4857.707553 samples/sec	accuracy=0.808309	lr=0.031784
Epoch[99] Batch [149]	Speed: 4970.731481 samples/sec	accuracy=0.808251	lr=0.031666
Epoch[99] Batch [199]	Speed: 4629.409445 samples/sec	accuracy=0.808120	lr=0.031548
Epoch[99] Batch [249]	Speed: 4932.136836 samples/sec	accuracy=0.808151	lr=0.031431
Epoch[99] Batch [299]	Speed: 4750.188395 samples/sec	accuracy=0.808138	lr=0.031313
Epoch[99] Batch [349]	Speed: 4757.928277 samples/sec	accuracy=0.807968	lr=0.031196
Epoch[99] Batch [399]	Speed: 5044.237058 samples/sec	accuracy=0.807975	lr=0.031079
Epoch[99] Batch [449]	Speed: 4831.368321 samples/sec	accuracy=0.808080	lr=0.030962
Epoch[99] Batch [499]	Speed: 4942.949977 samples/sec	accuracy=0.808113	lr=0.030846
Epoch[99] Batch [549]	Speed: 5023.034321 samples/sec	accuracy=0.808099	lr=0.030729
Epoch[99] Batch [599]	Speed: 4843.255493 samples/sec	accuracy=0.808128	lr=0.030613
Epoch[99] Batch [649]	Speed: 4950.118442 samples/sec	accuracy=0.808153	lr=0.030497
Epoch[99] Batch [699]	Speed: 4834.299217 samples/sec	accuracy=0.808191	lr=0.030381
Epoch[99] Batch [749]	Speed: 4870.139874 samples/sec	accuracy=0.808217	lr=0.030266
Epoch[99] Batch [799]	Speed: 4783.014523 samples/sec	accuracy=0.808235	lr=0.030150
Epoch[99] Batch [849]	Speed: 4888.616432 samples/sec	accuracy=0.808318	lr=0.030035
Epoch[99] Batch [899]	Speed: 4799.628132 samples/sec	accuracy=0.808355	lr=0.029920
Epoch[99] Batch [949]	Speed: 4700.692529 samples/sec	accuracy=0.808403	lr=0.029805
Epoch[99] Batch [999]	Speed: 5029.781329 samples/sec	accuracy=0.808463	lr=0.029691
Epoch[99] Batch [1049]	Speed: 4727.948987 samples/sec	accuracy=0.808484	lr=0.029576
Epoch[99] Batch [1099]	Speed: 4925.145471 samples/sec	accuracy=0.808611	lr=0.029462
Epoch[99] Batch [1149]	Speed: 4930.445063 samples/sec	accuracy=0.808686	lr=0.029348
Epoch[99] Batch [1199]	Speed: 4798.086274 samples/sec	accuracy=0.808755	lr=0.029234
Epoch[99] Batch [1249]	Speed: 4998.983198 samples/sec	accuracy=0.808818	lr=0.029121
[Epoch 99] training: accuracy=0.808819
[Epoch 99] speed: 4837 samples/sec	time cost: 275.790866
[Epoch 99] validation: err-top1=0.269810 err-top5=0.085619
Epoch[100] Batch [49]	Speed: 4794.610491 samples/sec	accuracy=0.809398	lr=0.029005
Epoch[100] Batch [99]	Speed: 4735.977965 samples/sec	accuracy=0.809804	lr=0.028892
Epoch[100] Batch [149]	Speed: 4849.090520 samples/sec	accuracy=0.809620	lr=0.028779
Epoch[100] Batch [199]	Speed: 4675.728209 samples/sec	accuracy=0.809707	lr=0.028666
Epoch[100] Batch [249]	Speed: 4867.065873 samples/sec	accuracy=0.809744	lr=0.028554
Epoch[100] Batch [299]	Speed: 4883.844572 samples/sec	accuracy=0.809549	lr=0.028441
Epoch[100] Batch [349]	Speed: 4990.372628 samples/sec	accuracy=0.809403	lr=0.028329
Epoch[100] Batch [399]	Speed: 4846.966937 samples/sec	accuracy=0.809346	lr=0.028217
Epoch[100] Batch [449]	Speed: 4940.506072 samples/sec	accuracy=0.809507	lr=0.028105
Epoch[100] Batch [499]	Speed: 4780.849847 samples/sec	accuracy=0.809653	lr=0.027994
Epoch[100] Batch [549]	Speed: 4740.544122 samples/sec	accuracy=0.809675	lr=0.027883
Epoch[100] Batch [599]	Speed: 4799.875298 samples/sec	accuracy=0.809714	lr=0.027772
Epoch[100] Batch [649]	Speed: 4727.391122 samples/sec	accuracy=0.809709	lr=0.027661
Epoch[100] Batch [699]	Speed: 4762.605110 samples/sec	accuracy=0.809724	lr=0.027550
Epoch[100] Batch [749]	Speed: 4982.504369 samples/sec	accuracy=0.809732	lr=0.027439
Epoch[100] Batch [799]	Speed: 4780.287835 samples/sec	accuracy=0.809809	lr=0.027329
Epoch[100] Batch [849]	Speed: 4992.469852 samples/sec	accuracy=0.809885	lr=0.027219
Epoch[100] Batch [899]	Speed: 4794.049734 samples/sec	accuracy=0.809988	lr=0.027109
Epoch[100] Batch [949]	Speed: 4792.730723 samples/sec	accuracy=0.810057	lr=0.026999
Epoch[100] Batch [999]	Speed: 4820.001153 samples/sec	accuracy=0.810128	lr=0.026890
Epoch[100] Batch [1049]	Speed: 4780.941808 samples/sec	accuracy=0.810148	lr=0.026781
Epoch[100] Batch [1099]	Speed: 4955.131886 samples/sec	accuracy=0.810161	lr=0.026672
Epoch[100] Batch [1149]	Speed: 4825.567058 samples/sec	accuracy=0.810198	lr=0.026563
Epoch[100] Batch [1199]	Speed: 4889.302164 samples/sec	accuracy=0.810258	lr=0.026454
Epoch[100] Batch [1249]	Speed: 4838.615669 samples/sec	accuracy=0.810294	lr=0.026346
[Epoch 100] training: accuracy=0.810297
[Epoch 100] speed: 4828 samples/sec	time cost: 275.643454
[Epoch 100] validation: err-top1=0.266203 err-top5=0.084782
Epoch[101] Batch [49]	Speed: 4700.415038 samples/sec	accuracy=0.811148	lr=0.026235
Epoch[101] Batch [99]	Speed: 4620.323948 samples/sec	accuracy=0.810901	lr=0.026127
Epoch[101] Batch [149]	Speed: 4631.093648 samples/sec	accuracy=0.810716	lr=0.026020
Epoch[101] Batch [199]	Speed: 4887.997091 samples/sec	accuracy=0.810705	lr=0.025912
Epoch[101] Batch [249]	Speed: 4729.269237 samples/sec	accuracy=0.810712	lr=0.025805
Epoch[101] Batch [299]	Speed: 4802.028356 samples/sec	accuracy=0.810880	lr=0.025697
Epoch[101] Batch [349]	Speed: 4670.722330 samples/sec	accuracy=0.810863	lr=0.025590
Epoch[101] Batch [399]	Speed: 4642.778945 samples/sec	accuracy=0.810872	lr=0.025484
Epoch[101] Batch [449]	Speed: 4988.886136 samples/sec	accuracy=0.810970	lr=0.025377
Epoch[101] Batch [499]	Speed: 4572.009881 samples/sec	accuracy=0.810982	lr=0.025271
Epoch[101] Batch [549]	Speed: 4956.590549 samples/sec	accuracy=0.810952	lr=0.025165
Epoch[101] Batch [599]	Speed: 4790.549551 samples/sec	accuracy=0.810966	lr=0.025059
Epoch[101] Batch [649]	Speed: 4960.037266 samples/sec	accuracy=0.810947	lr=0.024953
Epoch[101] Batch [699]	Speed: 4650.485571 samples/sec	accuracy=0.810891	lr=0.024847
Epoch[101] Batch [749]	Speed: 4815.874103 samples/sec	accuracy=0.810901	lr=0.024742
Epoch[101] Batch [799]	Speed: 4721.401857 samples/sec	accuracy=0.810886	lr=0.024637
Epoch[101] Batch [849]	Speed: 4626.814459 samples/sec	accuracy=0.810908	lr=0.024532
Epoch[101] Batch [899]	Speed: 4726.984359 samples/sec	accuracy=0.810950	lr=0.024427
Epoch[101] Batch [949]	Speed: 4871.655675 samples/sec	accuracy=0.810988	lr=0.024323
Epoch[101] Batch [999]	Speed: 4855.193839 samples/sec	accuracy=0.811020	lr=0.024219
Epoch[101] Batch [1049]	Speed: 4932.403050 samples/sec	accuracy=0.811041	lr=0.024115
Epoch[101] Batch [1099]	Speed: 4711.240027 samples/sec	accuracy=0.811087	lr=0.024011
Epoch[101] Batch [1149]	Speed: 4874.518271 samples/sec	accuracy=0.811139	lr=0.023907
Epoch[101] Batch [1199]	Speed: 4816.085251 samples/sec	accuracy=0.811193	lr=0.023804
Epoch[101] Batch [1249]	Speed: 5061.093933 samples/sec	accuracy=0.811259	lr=0.023700
[Epoch 101] training: accuracy=0.811259
[Epoch 101] speed: 4776 samples/sec	time cost: 278.945207
[Epoch 101] validation: err-top1=0.263353 err-top5=0.082848
Epoch[102] Batch [49]	Speed: 4761.610770 samples/sec	accuracy=0.811700	lr=0.023595
Epoch[102] Batch [99]	Speed: 4645.809239 samples/sec	accuracy=0.810927	lr=0.023493
Epoch[102] Batch [149]	Speed: 4761.191869 samples/sec	accuracy=0.811026	lr=0.023390
Epoch[102] Batch [199]	Speed: 4690.819554 samples/sec	accuracy=0.811122	lr=0.023288
Epoch[102] Batch [249]	Speed: 4910.711378 samples/sec	accuracy=0.811061	lr=0.023185
Epoch[102] Batch [299]	Speed: 4810.350564 samples/sec	accuracy=0.810983	lr=0.023083
Epoch[102] Batch [349]	Speed: 4687.326013 samples/sec	accuracy=0.810949	lr=0.022982
Epoch[102] Batch [399]	Speed: 4710.804934 samples/sec	accuracy=0.811030	lr=0.022880
Epoch[102] Batch [449]	Speed: 4905.873331 samples/sec	accuracy=0.811113	lr=0.022779
Epoch[102] Batch [499]	Speed: 4885.507400 samples/sec	accuracy=0.811156	lr=0.022678
Epoch[102] Batch [549]	Speed: 4851.962170 samples/sec	accuracy=0.811226	lr=0.022577
Epoch[102] Batch [599]	Speed: 4790.714559 samples/sec	accuracy=0.811339	lr=0.022476
Epoch[102] Batch [649]	Speed: 4735.903601 samples/sec	accuracy=0.811414	lr=0.022376
Epoch[102] Batch [699]	Speed: 4648.485451 samples/sec	accuracy=0.811486	lr=0.022276
Epoch[102] Batch [749]	Speed: 4654.881122 samples/sec	accuracy=0.811513	lr=0.022175
Epoch[102] Batch [799]	Speed: 4606.876812 samples/sec	accuracy=0.811542	lr=0.022076
Epoch[102] Batch [849]	Speed: 4978.656870 samples/sec	accuracy=0.811571	lr=0.021976
Epoch[102] Batch [899]	Speed: 4790.517385 samples/sec	accuracy=0.811696	lr=0.021877
Epoch[102] Batch [949]	Speed: 4591.413958 samples/sec	accuracy=0.811827	lr=0.021777
Epoch[102] Batch [999]	Speed: 5031.256107 samples/sec	accuracy=0.811865	lr=0.021678
Epoch[102] Batch [1049]	Speed: 4736.981376 samples/sec	accuracy=0.811936	lr=0.021580
Epoch[102] Batch [1099]	Speed: 4805.350922 samples/sec	accuracy=0.812016	lr=0.021481
Epoch[102] Batch [1149]	Speed: 4836.731420 samples/sec	accuracy=0.812070	lr=0.021383
Epoch[102] Batch [1199]	Speed: 4732.863988 samples/sec	accuracy=0.812111	lr=0.021285
Epoch[102] Batch [1249]	Speed: 4906.338816 samples/sec	accuracy=0.812148	lr=0.021187
[Epoch 102] training: accuracy=0.812148
[Epoch 102] speed: 4772 samples/sec	time cost: 277.591351
[Epoch 102] validation: err-top1=0.254624 err-top5=0.076272
Epoch[103] Batch [49]	Speed: 4662.926370 samples/sec	accuracy=0.813215	lr=0.021087
Epoch[103] Batch [99]	Speed: 4733.745765 samples/sec	accuracy=0.812993	lr=0.020990
Epoch[103] Batch [149]	Speed: 4773.097016 samples/sec	accuracy=0.812660	lr=0.020892
Epoch[103] Batch [199]	Speed: 4889.809714 samples/sec	accuracy=0.812604	lr=0.020795
Epoch[103] Batch [249]	Speed: 4644.356173 samples/sec	accuracy=0.812631	lr=0.020698
Epoch[103] Batch [299]	Speed: 4647.522195 samples/sec	accuracy=0.812575	lr=0.020602
Epoch[103] Batch [349]	Speed: 4683.427237 samples/sec	accuracy=0.812622	lr=0.020505
Epoch[103] Batch [399]	Speed: 4958.089563 samples/sec	accuracy=0.812660	lr=0.020409
Epoch[103] Batch [449]	Speed: 4945.544928 samples/sec	accuracy=0.812671	lr=0.020313
Epoch[103] Batch [499]	Speed: 4693.950764 samples/sec	accuracy=0.812606	lr=0.020217
Epoch[103] Batch [549]	Speed: 4705.764747 samples/sec	accuracy=0.812586	lr=0.020122
Epoch[103] Batch [599]	Speed: 4886.733747 samples/sec	accuracy=0.812644	lr=0.020026
Epoch[103] Batch [649]	Speed: 4877.525034 samples/sec	accuracy=0.812598	lr=0.019931
Epoch[103] Batch [699]	Speed: 4540.187643 samples/sec	accuracy=0.812636	lr=0.019836
Epoch[103] Batch [749]	Speed: 4897.108176 samples/sec	accuracy=0.812605	lr=0.019742
Epoch[103] Batch [799]	Speed: 4838.483102 samples/sec	accuracy=0.812667	lr=0.019647
Epoch[103] Batch [849]	Speed: 4967.973959 samples/sec	accuracy=0.812707	lr=0.019553
Epoch[103] Batch [899]	Speed: 5118.686616 samples/sec	accuracy=0.812772	lr=0.019459
Epoch[103] Batch [949]	Speed: 4936.710313 samples/sec	accuracy=0.812849	lr=0.019365
Epoch[103] Batch [999]	Speed: 5065.832494 samples/sec	accuracy=0.812905	lr=0.019271
Epoch[103] Batch [1049]	Speed: 4957.967195 samples/sec	accuracy=0.812973	lr=0.019178
Epoch[103] Batch [1099]	Speed: 4864.990773 samples/sec	accuracy=0.813061	lr=0.019085
Epoch[103] Batch [1149]	Speed: 4875.855124 samples/sec	accuracy=0.813120	lr=0.018992
Epoch[103] Batch [1199]	Speed: 4787.775733 samples/sec	accuracy=0.813191	lr=0.018899
Epoch[103] Batch [1249]	Speed: 4966.274859 samples/sec	accuracy=0.813242	lr=0.018806
[Epoch 103] training: accuracy=0.813242
[Epoch 103] speed: 4828 samples/sec	time cost: 274.613187
[Epoch 103] validation: err-top1=0.249123 err-top5=0.075415
Epoch[104] Batch [49]	Speed: 4757.850903 samples/sec	accuracy=0.814668	lr=0.018712
Epoch[104] Batch [99]	Speed: 4784.542764 samples/sec	accuracy=0.814144	lr=0.018620
Epoch[104] Batch [149]	Speed: 4754.321363 samples/sec	accuracy=0.813935	lr=0.018528
Epoch[104] Batch [199]	Speed: 4732.010589 samples/sec	accuracy=0.813939	lr=0.018436
Epoch[104] Batch [249]	Speed: 4914.966311 samples/sec	accuracy=0.813887	lr=0.018345
Epoch[104] Batch [299]	Speed: 4784.565149 samples/sec	accuracy=0.813975	lr=0.018254
Epoch[104] Batch [349]	Speed: 4774.746634 samples/sec	accuracy=0.813905	lr=0.018163
Epoch[104] Batch [399]	Speed: 4719.474701 samples/sec	accuracy=0.813934	lr=0.018072
Epoch[104] Batch [449]	Speed: 4999.815949 samples/sec	accuracy=0.814018	lr=0.017981
Epoch[104] Batch [499]	Speed: 4668.031825 samples/sec	accuracy=0.814049	lr=0.017891
Epoch[104] Batch [549]	Speed: 4705.460674 samples/sec	accuracy=0.814057	lr=0.017801
Epoch[104] Batch [599]	Speed: 4962.453979 samples/sec	accuracy=0.814186	lr=0.017711
Epoch[104] Batch [649]	Speed: 4639.142194 samples/sec	accuracy=0.814287	lr=0.017621
Epoch[104] Batch [699]	Speed: 4857.365180 samples/sec	accuracy=0.814276	lr=0.017532
Epoch[104] Batch [749]	Speed: 4808.074743 samples/sec	accuracy=0.814286	lr=0.017442
Epoch[104] Batch [799]	Speed: 4733.617943 samples/sec	accuracy=0.814305	lr=0.017353
Epoch[104] Batch [849]	Speed: 4790.243399 samples/sec	accuracy=0.814295	lr=0.017264
Epoch[104] Batch [899]	Speed: 4966.448863 samples/sec	accuracy=0.814309	lr=0.017176
Epoch[104] Batch [949]	Speed: 4882.700715 samples/sec	accuracy=0.814315	lr=0.017087
Epoch[104] Batch [999]	Speed: 4801.957165 samples/sec	accuracy=0.814339	lr=0.016999
Epoch[104] Batch [1049]	Speed: 4812.026469 samples/sec	accuracy=0.814345	lr=0.016911
Epoch[104] Batch [1099]	Speed: 5161.644710 samples/sec	accuracy=0.814350	lr=0.016823
Epoch[104] Batch [1149]	Speed: 5169.701462 samples/sec	accuracy=0.814411	lr=0.016736
Epoch[104] Batch [1199]	Speed: 5009.966580 samples/sec	accuracy=0.814451	lr=0.016648
Epoch[104] Batch [1249]	Speed: 5074.136949 samples/sec	accuracy=0.814517	lr=0.016561
[Epoch 104] training: accuracy=0.814518
[Epoch 104] speed: 4842 samples/sec	time cost: 274.159361
[Epoch 104] validation: err-top1=0.251180 err-top5=0.075439
Epoch[105] Batch [49]	Speed: 4886.076195 samples/sec	accuracy=0.815487	lr=0.016473
Epoch[105] Batch [99]	Speed: 5044.971294 samples/sec	accuracy=0.814621	lr=0.016386
Epoch[105] Batch [149]	Speed: 4868.285295 samples/sec	accuracy=0.814143	lr=0.016299
Epoch[105] Batch [199]	Speed: 5012.742846 samples/sec	accuracy=0.814277	lr=0.016213
Epoch[105] Batch [249]	Speed: 4882.486573 samples/sec	accuracy=0.814303	lr=0.016127
Epoch[105] Batch [299]	Speed: 4860.731563 samples/sec	accuracy=0.814323	lr=0.016041
Epoch[105] Batch [349]	Speed: 4812.914370 samples/sec	accuracy=0.814437	lr=0.015956
Epoch[105] Batch [399]	Speed: 5022.568162 samples/sec	accuracy=0.814600	lr=0.015870
Epoch[105] Batch [449]	Speed: 5030.557674 samples/sec	accuracy=0.814780	lr=0.015785
Epoch[105] Batch [499]	Speed: 4778.521142 samples/sec	accuracy=0.814779	lr=0.015700
Epoch[105] Batch [549]	Speed: 4897.757866 samples/sec	accuracy=0.814806	lr=0.015616
Epoch[105] Batch [599]	Speed: 4909.911746 samples/sec	accuracy=0.814865	lr=0.015531
Epoch[105] Batch [649]	Speed: 4812.319992 samples/sec	accuracy=0.814916	lr=0.015447
Epoch[105] Batch [699]	Speed: 4866.071875 samples/sec	accuracy=0.815007	lr=0.015363
Epoch[105] Batch [749]	Speed: 4907.400135 samples/sec	accuracy=0.815106	lr=0.015279
Epoch[105] Batch [799]	Speed: 4867.369788 samples/sec	accuracy=0.815138	lr=0.015195
Epoch[105] Batch [849]	Speed: 4889.297044 samples/sec	accuracy=0.815168	lr=0.015112
Epoch[105] Batch [899]	Speed: 5014.097714 samples/sec	accuracy=0.815219	lr=0.015029
Epoch[105] Batch [949]	Speed: 5019.432652 samples/sec	accuracy=0.815271	lr=0.014946
Epoch[105] Batch [999]	Speed: 4953.996111 samples/sec	accuracy=0.815325	lr=0.014863
Epoch[105] Batch [1049]	Speed: 4924.191518 samples/sec	accuracy=0.815428	lr=0.014781
Epoch[105] Batch [1099]	Speed: 4813.433372 samples/sec	accuracy=0.815511	lr=0.014699
Epoch[105] Batch [1149]	Speed: 4917.156554 samples/sec	accuracy=0.815550	lr=0.014616
Epoch[105] Batch [1199]	Speed: 4930.847292 samples/sec	accuracy=0.815599	lr=0.014535
Epoch[105] Batch [1249]	Speed: 4894.458485 samples/sec	accuracy=0.815635	lr=0.014453
[Epoch 105] training: accuracy=0.815636
[Epoch 105] speed: 4907 samples/sec	time cost: 270.905385
[Epoch 105] validation: err-top1=0.246472 err-top5=0.073123
Epoch[106] Batch [49]	Speed: 4583.221398 samples/sec	accuracy=0.816158	lr=0.014370
Epoch[106] Batch [99]	Speed: 4702.811267 samples/sec	accuracy=0.816027	lr=0.014289
Epoch[106] Batch [149]	Speed: 4859.508443 samples/sec	accuracy=0.815839	lr=0.014208
Epoch[106] Batch [199]	Speed: 4524.875305 samples/sec	accuracy=0.815726	lr=0.014127
Epoch[106] Batch [249]	Speed: 4878.451236 samples/sec	accuracy=0.815616	lr=0.014047
Epoch[106] Batch [299]	Speed: 4708.710995 samples/sec	accuracy=0.815508	lr=0.013966
Epoch[106] Batch [349]	Speed: 5078.209991 samples/sec	accuracy=0.815491	lr=0.013886
Epoch[106] Batch [399]	Speed: 4946.587956 samples/sec	accuracy=0.815605	lr=0.013806
Epoch[106] Batch [449]	Speed: 5032.650841 samples/sec	accuracy=0.815625	lr=0.013727
Epoch[106] Batch [499]	Speed: 4784.534662 samples/sec	accuracy=0.815590	lr=0.013647
Epoch[106] Batch [549]	Speed: 4834.017589 samples/sec	accuracy=0.815589	lr=0.013568
Epoch[106] Batch [599]	Speed: 4713.675967 samples/sec	accuracy=0.815634	lr=0.013489
Epoch[106] Batch [649]	Speed: 4936.171082 samples/sec	accuracy=0.815700	lr=0.013411
Epoch[106] Batch [699]	Speed: 4658.465293 samples/sec	accuracy=0.815792	lr=0.013332
Epoch[106] Batch [749]	Speed: 4919.688420 samples/sec	accuracy=0.815840	lr=0.013254
Epoch[106] Batch [799]	Speed: 4922.859179 samples/sec	accuracy=0.815907	lr=0.013176
Epoch[106] Batch [849]	Speed: 4722.706612 samples/sec	accuracy=0.815959	lr=0.013098
Epoch[106] Batch [899]	Speed: 4959.168123 samples/sec	accuracy=0.816043	lr=0.013020
Epoch[106] Batch [949]	Speed: 5069.674377 samples/sec	accuracy=0.816079	lr=0.012943
Epoch[106] Batch [999]	Speed: 4797.141360 samples/sec	accuracy=0.816119	lr=0.012866
Epoch[106] Batch [1049]	Speed: 4908.646026 samples/sec	accuracy=0.816176	lr=0.012789
Epoch[106] Batch [1099]	Speed: 4778.248632 samples/sec	accuracy=0.816246	lr=0.012712
Epoch[106] Batch [1149]	Speed: 4907.847515 samples/sec	accuracy=0.816311	lr=0.012636
Epoch[106] Batch [1199]	Speed: 4866.514508 samples/sec	accuracy=0.816365	lr=0.012559
Epoch[106] Batch [1249]	Speed: 5120.706152 samples/sec	accuracy=0.816417	lr=0.012483
[Epoch 106] training: accuracy=0.816418
[Epoch 106] speed: 4839 samples/sec	time cost: 274.351403
[Epoch 106] validation: err-top1=0.251674 err-top5=0.075973
Epoch[107] Batch [49]	Speed: 4676.811359 samples/sec	accuracy=0.817280	lr=0.012406
Epoch[107] Batch [99]	Speed: 4670.189262 samples/sec	accuracy=0.816934	lr=0.012330
Epoch[107] Batch [149]	Speed: 4946.644244 samples/sec	accuracy=0.816807	lr=0.012255
Epoch[107] Batch [199]	Speed: 5048.087420 samples/sec	accuracy=0.816758	lr=0.012180
Epoch[107] Batch [249]	Speed: 4862.454764 samples/sec	accuracy=0.816808	lr=0.012105
Epoch[107] Batch [299]	Speed: 4862.366136 samples/sec	accuracy=0.816866	lr=0.012030
Epoch[107] Batch [349]	Speed: 4887.280471 samples/sec	accuracy=0.816942	lr=0.011956
Epoch[107] Batch [399]	Speed: 4654.418041 samples/sec	accuracy=0.817004	lr=0.011881
Epoch[107] Batch [449]	Speed: 4927.630734 samples/sec	accuracy=0.817008	lr=0.011807
Epoch[107] Batch [499]	Speed: 4989.733961 samples/sec	accuracy=0.817009	lr=0.011734
Epoch[107] Batch [549]	Speed: 5083.925412 samples/sec	accuracy=0.817031	lr=0.011660
Epoch[107] Batch [599]	Speed: 4783.174111 samples/sec	accuracy=0.817045	lr=0.011587
Epoch[107] Batch [649]	Speed: 4732.405598 samples/sec	accuracy=0.817103	lr=0.011513
Epoch[107] Batch [699]	Speed: 4700.959351 samples/sec	accuracy=0.817102	lr=0.011441
Epoch[107] Batch [749]	Speed: 4816.363497 samples/sec	accuracy=0.817150	lr=0.011368
Epoch[107] Batch [799]	Speed: 4815.417634 samples/sec	accuracy=0.817143	lr=0.011295
Epoch[107] Batch [849]	Speed: 4995.019591 samples/sec	accuracy=0.817156	lr=0.011223
Epoch[107] Batch [899]	Speed: 4904.648226 samples/sec	accuracy=0.817228	lr=0.011151
Epoch[107] Batch [949]	Speed: 4709.856688 samples/sec	accuracy=0.817276	lr=0.011079
Epoch[107] Batch [999]	Speed: 4750.847716 samples/sec	accuracy=0.817313	lr=0.011008
Epoch[107] Batch [1049]	Speed: 4641.034980 samples/sec	accuracy=0.817375	lr=0.010936
Epoch[107] Batch [1099]	Speed: 4759.621014 samples/sec	accuracy=0.817435	lr=0.010865
Epoch[107] Batch [1149]	Speed: 4831.381473 samples/sec	accuracy=0.817505	lr=0.010794
Epoch[107] Batch [1199]	Speed: 4972.051182 samples/sec	accuracy=0.817535	lr=0.010724
Epoch[107] Batch [1249]	Speed: 4863.305093 samples/sec	accuracy=0.817590	lr=0.010653
[Epoch 107] training: accuracy=0.817593
[Epoch 107] speed: 4829 samples/sec	time cost: 275.681715
[Epoch 107] validation: err-top1=0.240653 err-top5=0.070592
Epoch[108] Batch [49]	Speed: 4554.000579 samples/sec	accuracy=0.818213	lr=0.010582
Epoch[108] Batch [99]	Speed: 4832.823101 samples/sec	accuracy=0.818324	lr=0.010512
Epoch[108] Batch [149]	Speed: 4918.279215 samples/sec	accuracy=0.818123	lr=0.010442
Epoch[108] Batch [199]	Speed: 4871.557871 samples/sec	accuracy=0.818039	lr=0.010373
Epoch[108] Batch [249]	Speed: 4732.029358 samples/sec	accuracy=0.818022	lr=0.010303
Epoch[108] Batch [299]	Speed: 4726.115500 samples/sec	accuracy=0.818020	lr=0.010234
Epoch[108] Batch [349]	Speed: 4992.953309 samples/sec	accuracy=0.818014	lr=0.010165
Epoch[108] Batch [399]	Speed: 4871.727511 samples/sec	accuracy=0.818044	lr=0.010097
Epoch[108] Batch [449]	Speed: 4951.925255 samples/sec	accuracy=0.818084	lr=0.010028
Epoch[108] Batch [499]	Speed: 4823.253203 samples/sec	accuracy=0.818095	lr=0.009960
Epoch[108] Batch [549]	Speed: 4781.384205 samples/sec	accuracy=0.818082	lr=0.009892
Epoch[108] Batch [599]	Speed: 5029.244308 samples/sec	accuracy=0.818084	lr=0.009824
Epoch[108] Batch [649]	Speed: 4698.311514 samples/sec	accuracy=0.818085	lr=0.009757
Epoch[108] Batch [699]	Speed: 4873.064271 samples/sec	accuracy=0.818049	lr=0.009690
Epoch[108] Batch [749]	Speed: 4727.749868 samples/sec	accuracy=0.818103	lr=0.009623
Epoch[108] Batch [799]	Speed: 4783.428856 samples/sec	accuracy=0.818124	lr=0.009556
Epoch[108] Batch [849]	Speed: 4863.881178 samples/sec	accuracy=0.818160	lr=0.009489
Epoch[108] Batch [899]	Speed: 5035.104613 samples/sec	accuracy=0.818179	lr=0.009423
Epoch[108] Batch [949]	Speed: 4708.227336 samples/sec	accuracy=0.818190	lr=0.009357
Epoch[108] Batch [999]	Speed: 4824.428983 samples/sec	accuracy=0.818189	lr=0.009291
Epoch[108] Batch [1049]	Speed: 4661.784772 samples/sec	accuracy=0.818229	lr=0.009225
Epoch[108] Batch [1099]	Speed: 4822.501832 samples/sec	accuracy=0.818239	lr=0.009160
Epoch[108] Batch [1149]	Speed: 4871.430897 samples/sec	accuracy=0.818275	lr=0.009095
Epoch[108] Batch [1199]	Speed: 4634.602258 samples/sec	accuracy=0.818320	lr=0.009030
Epoch[108] Batch [1249]	Speed: 4897.134307 samples/sec	accuracy=0.818349	lr=0.008965
[Epoch 108] training: accuracy=0.818350
[Epoch 108] speed: 4812 samples/sec	time cost: 275.410077
[Epoch 108] validation: err-top1=0.237643 err-top5=0.068838
Epoch[109] Batch [49]	Speed: 4746.307539 samples/sec	accuracy=0.818956	lr=0.008899
Epoch[109] Batch [99]	Speed: 4636.129095 samples/sec	accuracy=0.818832	lr=0.008835
Epoch[109] Batch [149]	Speed: 4508.637677 samples/sec	accuracy=0.818634	lr=0.008771
Epoch[109] Batch [199]	Speed: 4753.352153 samples/sec	accuracy=0.818607	lr=0.008707
Epoch[109] Batch [249]	Speed: 4670.324448 samples/sec	accuracy=0.818649	lr=0.008643
Epoch[109] Batch [299]	Speed: 4867.502177 samples/sec	accuracy=0.818691	lr=0.008580
Epoch[109] Batch [349]	Speed: 4561.868248 samples/sec	accuracy=0.818709	lr=0.008517
Epoch[109] Batch [399]	Speed: 4707.585880 samples/sec	accuracy=0.818802	lr=0.008454
Epoch[109] Batch [449]	Speed: 4717.489218 samples/sec	accuracy=0.818928	lr=0.008391
Epoch[109] Batch [499]	Speed: 4866.249845 samples/sec	accuracy=0.818997	lr=0.008329
Epoch[109] Batch [549]	Speed: 4947.152942 samples/sec	accuracy=0.818992	lr=0.008266
Epoch[109] Batch [599]	Speed: 4993.781731 samples/sec	accuracy=0.818994	lr=0.008204
Epoch[109] Batch [649]	Speed: 4578.356678 samples/sec	accuracy=0.819064	lr=0.008143
Epoch[109] Batch [699]	Speed: 4843.564198 samples/sec	accuracy=0.819096	lr=0.008081
Epoch[109] Batch [749]	Speed: 4913.653574 samples/sec	accuracy=0.819122	lr=0.008020
Epoch[109] Batch [799]	Speed: 4603.196407 samples/sec	accuracy=0.819161	lr=0.007959
Epoch[109] Batch [849]	Speed: 4634.161403 samples/sec	accuracy=0.819178	lr=0.007898
Epoch[109] Batch [899]	Speed: 4644.207020 samples/sec	accuracy=0.819243	lr=0.007837
Epoch[109] Batch [949]	Speed: 4742.147122 samples/sec	accuracy=0.819271	lr=0.007777
Epoch[109] Batch [999]	Speed: 4936.060119 samples/sec	accuracy=0.819303	lr=0.007716
Epoch[109] Batch [1049]	Speed: 4763.920694 samples/sec	accuracy=0.819313	lr=0.007656
Epoch[109] Batch [1099]	Speed: 5018.393160 samples/sec	accuracy=0.819343	lr=0.007597
Epoch[109] Batch [1149]	Speed: 4757.395143 samples/sec	accuracy=0.819365	lr=0.007537
Epoch[109] Batch [1199]	Speed: 4598.026142 samples/sec	accuracy=0.819387	lr=0.007478
Epoch[109] Batch [1249]	Speed: 5129.667879 samples/sec	accuracy=0.819399	lr=0.007419
[Epoch 109] training: accuracy=0.819401
[Epoch 109] speed: 4756 samples/sec	time cost: 279.022235
[Epoch 109] validation: err-top1=0.237285 err-top5=0.068100
Epoch[110] Batch [49]	Speed: 4378.246269 samples/sec	accuracy=0.820420	lr=0.007359
Epoch[110] Batch [99]	Speed: 4818.960103 samples/sec	accuracy=0.820476	lr=0.007300
Epoch[110] Batch [149]	Speed: 4908.610234 samples/sec	accuracy=0.820373	lr=0.007242
Epoch[110] Batch [199]	Speed: 4964.328567 samples/sec	accuracy=0.820189	lr=0.007184
Epoch[110] Batch [249]	Speed: 4915.698163 samples/sec	accuracy=0.819996	lr=0.007126
Epoch[110] Batch [299]	Speed: 4604.364967 samples/sec	accuracy=0.819897	lr=0.007068
Epoch[110] Batch [349]	Speed: 4947.046385 samples/sec	accuracy=0.819832	lr=0.007011
Epoch[110] Batch [399]	Speed: 4831.694755 samples/sec	accuracy=0.819808	lr=0.006954
Epoch[110] Batch [449]	Speed: 4945.160796 samples/sec	accuracy=0.819829	lr=0.006897
Epoch[110] Batch [499]	Speed: 4797.371123 samples/sec	accuracy=0.819806	lr=0.006840
Epoch[110] Batch [549]	Speed: 4954.047881 samples/sec	accuracy=0.819819	lr=0.006784
Epoch[110] Batch [599]	Speed: 4753.690647 samples/sec	accuracy=0.819792	lr=0.006727
Epoch[110] Batch [649]	Speed: 4757.785548 samples/sec	accuracy=0.819818	lr=0.006671
Epoch[110] Batch [699]	Speed: 4712.672786 samples/sec	accuracy=0.819874	lr=0.006615
Epoch[110] Batch [749]	Speed: 4522.769994 samples/sec	accuracy=0.819911	lr=0.006560
Epoch[110] Batch [799]	Speed: 4925.887024 samples/sec	accuracy=0.819900	lr=0.006504
Epoch[110] Batch [849]	Speed: 4977.511670 samples/sec	accuracy=0.819903	lr=0.006449
Epoch[110] Batch [899]	Speed: 4818.583598 samples/sec	accuracy=0.819902	lr=0.006394
Epoch[110] Batch [949]	Speed: 5036.529595 samples/sec	accuracy=0.819910	lr=0.006340
Epoch[110] Batch [999]	Speed: 4766.098094 samples/sec	accuracy=0.819921	lr=0.006285
Epoch[110] Batch [1049]	Speed: 4619.291444 samples/sec	accuracy=0.819965	lr=0.006231
Epoch[110] Batch [1099]	Speed: 4829.414663 samples/sec	accuracy=0.819983	lr=0.006177
Epoch[110] Batch [1149]	Speed: 4779.367787 samples/sec	accuracy=0.820003	lr=0.006123
Epoch[110] Batch [1199]	Speed: 4841.030829 samples/sec	accuracy=0.820028	lr=0.006070
Epoch[110] Batch [1249]	Speed: 4979.139618 samples/sec	accuracy=0.820053	lr=0.006017
[Epoch 110] training: accuracy=0.820055
[Epoch 110] speed: 4806 samples/sec	time cost: 276.022334
[Epoch 110] validation: err-top1=0.233114 err-top5=0.066162
Epoch[111] Batch [49]	Speed: 4513.062959 samples/sec	accuracy=0.820316	lr=0.005963
Epoch[111] Batch [99]	Speed: 4959.704371 samples/sec	accuracy=0.820678	lr=0.005910
Epoch[111] Batch [149]	Speed: 4506.405971 samples/sec	accuracy=0.820730	lr=0.005857
Epoch[111] Batch [199]	Speed: 4741.588519 samples/sec	accuracy=0.820487	lr=0.005805
Epoch[111] Batch [249]	Speed: 4781.786117 samples/sec	accuracy=0.820407	lr=0.005753
Epoch[111] Batch [299]	Speed: 4853.857648 samples/sec	accuracy=0.820298	lr=0.005701
Epoch[111] Batch [349]	Speed: 5025.519179 samples/sec	accuracy=0.820291	lr=0.005649
Epoch[111] Batch [399]	Speed: 4879.061732 samples/sec	accuracy=0.820360	lr=0.005598
Epoch[111] Batch [449]	Speed: 4919.159550 samples/sec	accuracy=0.820330	lr=0.005547
Epoch[111] Batch [499]	Speed: 4844.058690 samples/sec	accuracy=0.820377	lr=0.005496
Epoch[111] Batch [549]	Speed: 4988.581226 samples/sec	accuracy=0.820382	lr=0.005445
Epoch[111] Batch [599]	Speed: 5038.170498 samples/sec	accuracy=0.820402	lr=0.005394
Epoch[111] Batch [649]	Speed: 4948.726079 samples/sec	accuracy=0.820394	lr=0.005344
Epoch[111] Batch [699]	Speed: 5038.262341 samples/sec	accuracy=0.820441	lr=0.005294
Epoch[111] Batch [749]	Speed: 5041.843388 samples/sec	accuracy=0.820492	lr=0.005244
Epoch[111] Batch [799]	Speed: 5095.293505 samples/sec	accuracy=0.820516	lr=0.005195
Epoch[111] Batch [849]	Speed: 4968.293711 samples/sec	accuracy=0.820565	lr=0.005145
Epoch[111] Batch [899]	Speed: 4736.904786 samples/sec	accuracy=0.820574	lr=0.005096
Epoch[111] Batch [949]	Speed: 5033.910055 samples/sec	accuracy=0.820584	lr=0.005047
Epoch[111] Batch [999]	Speed: 4950.239509 samples/sec	accuracy=0.820603	lr=0.004999
Epoch[111] Batch [1049]	Speed: 4853.880687 samples/sec	accuracy=0.820643	lr=0.004950
Epoch[111] Batch [1099]	Speed: 4873.383424 samples/sec	accuracy=0.820686	lr=0.004902
Epoch[111] Batch [1149]	Speed: 4923.805954 samples/sec	accuracy=0.820696	lr=0.004854
Epoch[111] Batch [1199]	Speed: 4939.778064 samples/sec	accuracy=0.820716	lr=0.004807
Epoch[111] Batch [1249]	Speed: 5059.667536 samples/sec	accuracy=0.820761	lr=0.004759
[Epoch 111] training: accuracy=0.820761
[Epoch 111] speed: 4891 samples/sec	time cost: 271.399652
[Epoch 111] validation: err-top1=0.232561 err-top5=0.067104
Epoch[112] Batch [49]	Speed: 4681.630765 samples/sec	accuracy=0.821152	lr=0.004711
Epoch[112] Batch [99]	Speed: 4753.314066 samples/sec	accuracy=0.821050	lr=0.004664
Epoch[112] Batch [149]	Speed: 4804.407229 samples/sec	accuracy=0.820964	lr=0.004617
Epoch[112] Batch [199]	Speed: 4863.827088 samples/sec	accuracy=0.820871	lr=0.004571
Epoch[112] Batch [249]	Speed: 4719.988580 samples/sec	accuracy=0.820791	lr=0.004524
Epoch[112] Batch [299]	Speed: 5007.515640 samples/sec	accuracy=0.820800	lr=0.004478
Epoch[112] Batch [349]	Speed: 4869.085887 samples/sec	accuracy=0.820858	lr=0.004432
Epoch[112] Batch [399]	Speed: 4953.047172 samples/sec	accuracy=0.820905	lr=0.004387
Epoch[112] Batch [449]	Speed: 4850.736995 samples/sec	accuracy=0.820947	lr=0.004341
Epoch[112] Batch [499]	Speed: 4934.498878 samples/sec	accuracy=0.820995	lr=0.004296
Epoch[112] Batch [549]	Speed: 4955.261775 samples/sec	accuracy=0.821029	lr=0.004251
Epoch[112] Batch [599]	Speed: 4960.735160 samples/sec	accuracy=0.821058	lr=0.004207
Epoch[112] Batch [649]	Speed: 4818.625873 samples/sec	accuracy=0.821126	lr=0.004162
Epoch[112] Batch [699]	Speed: 4733.010129 samples/sec	accuracy=0.821158	lr=0.004118
Epoch[112] Batch [749]	Speed: 4968.486019 samples/sec	accuracy=0.821235	lr=0.004074
Epoch[112] Batch [799]	Speed: 5078.750314 samples/sec	accuracy=0.821285	lr=0.004030
Epoch[112] Batch [849]	Speed: 5024.956024 samples/sec	accuracy=0.821303	lr=0.003987
Epoch[112] Batch [899]	Speed: 4964.112483 samples/sec	accuracy=0.821331	lr=0.003944
Epoch[112] Batch [949]	Speed: 4903.603437 samples/sec	accuracy=0.821367	lr=0.003901
Epoch[112] Batch [999]	Speed: 4808.042664 samples/sec	accuracy=0.821397	lr=0.003858
Epoch[112] Batch [1049]	Speed: 5004.420931 samples/sec	accuracy=0.821432	lr=0.003815
Epoch[112] Batch [1099]	Speed: 4889.585595 samples/sec	accuracy=0.821447	lr=0.003773
Epoch[112] Batch [1149]	Speed: 5081.781449 samples/sec	accuracy=0.821480	lr=0.003731
Epoch[112] Batch [1199]	Speed: 4825.020069 samples/sec	accuracy=0.821516	lr=0.003689
Epoch[112] Batch [1249]	Speed: 4919.555206 samples/sec	accuracy=0.821553	lr=0.003647
[Epoch 112] training: accuracy=0.821554
[Epoch 112] speed: 4888 samples/sec	time cost: 271.940241
[Epoch 112] validation: err-top1=0.230329 err-top5=0.065729
Epoch[113] Batch [49]	Speed: 4708.683738 samples/sec	accuracy=0.822208	lr=0.003605
Epoch[113] Batch [99]	Speed: 4592.201388 samples/sec	accuracy=0.822053	lr=0.003564
Epoch[113] Batch [149]	Speed: 4859.862337 samples/sec	accuracy=0.821810	lr=0.003523
Epoch[113] Batch [199]	Speed: 4898.044177 samples/sec	accuracy=0.821774	lr=0.003482
Epoch[113] Batch [249]	Speed: 4909.739885 samples/sec	accuracy=0.821856	lr=0.003442
Epoch[113] Batch [299]	Speed: 4810.202303 samples/sec	accuracy=0.821863	lr=0.003402
Epoch[113] Batch [349]	Speed: 4890.070376 samples/sec	accuracy=0.821933	lr=0.003362
Epoch[113] Batch [399]	Speed: 4751.413864 samples/sec	accuracy=0.821962	lr=0.003322
Epoch[113] Batch [449]	Speed: 4717.264037 samples/sec	accuracy=0.822014	lr=0.003282
Epoch[113] Batch [499]	Speed: 4944.552433 samples/sec	accuracy=0.822012	lr=0.003243
Epoch[113] Batch [549]	Speed: 4815.619562 samples/sec	accuracy=0.821994	lr=0.003204
Epoch[113] Batch [599]	Speed: 4638.865709 samples/sec	accuracy=0.821989	lr=0.003165
Epoch[113] Batch [649]	Speed: 5008.156298 samples/sec	accuracy=0.821993	lr=0.003127
Epoch[113] Batch [699]	Speed: 4657.066209 samples/sec	accuracy=0.821968	lr=0.003088
Epoch[113] Batch [749]	Speed: 4901.277026 samples/sec	accuracy=0.821938	lr=0.003050
Epoch[113] Batch [799]	Speed: 4740.324687 samples/sec	accuracy=0.821909	lr=0.003012
Epoch[113] Batch [849]	Speed: 5042.797291 samples/sec	accuracy=0.821901	lr=0.002975
Epoch[113] Batch [899]	Speed: 4843.880373 samples/sec	accuracy=0.821904	lr=0.002937
Epoch[113] Batch [949]	Speed: 4801.292601 samples/sec	accuracy=0.821917	lr=0.002900
Epoch[113] Batch [999]	Speed: 5007.084695 samples/sec	accuracy=0.821911	lr=0.002863
Epoch[113] Batch [1049]	Speed: 4769.123683 samples/sec	accuracy=0.821929	lr=0.002826
Epoch[113] Batch [1099]	Speed: 4941.331391 samples/sec	accuracy=0.821956	lr=0.002790
Epoch[113] Batch [1149]	Speed: 4975.098367 samples/sec	accuracy=0.821983	lr=0.002754
Epoch[113] Batch [1199]	Speed: 4662.306105 samples/sec	accuracy=0.822006	lr=0.002718
Epoch[113] Batch [1249]	Speed: 4965.821127 samples/sec	accuracy=0.822019	lr=0.002682
[Epoch 113] training: accuracy=0.822019
[Epoch 113] speed: 4826 samples/sec	time cost: 274.864325
[Epoch 113] validation: err-top1=0.228835 err-top5=0.064493
Epoch[114] Batch [49]	Speed: 4491.519702 samples/sec	accuracy=0.822781	lr=0.002646
Epoch[114] Batch [99]	Speed: 4832.580794 samples/sec	accuracy=0.822740	lr=0.002611
Epoch[114] Batch [149]	Speed: 4935.953585 samples/sec	accuracy=0.822708	lr=0.002575
Epoch[114] Batch [199]	Speed: 4691.631510 samples/sec	accuracy=0.822761	lr=0.002541
Epoch[114] Batch [249]	Speed: 4757.236849 samples/sec	accuracy=0.822706	lr=0.002506
Epoch[114] Batch [299]	Speed: 4707.197170 samples/sec	accuracy=0.822653	lr=0.002472
Epoch[114] Batch [349]	Speed: 4730.276267 samples/sec	accuracy=0.822594	lr=0.002438
Epoch[114] Batch [399]	Speed: 4878.796699 samples/sec	accuracy=0.822559	lr=0.002404
Epoch[114] Batch [449]	Speed: 4624.097816 samples/sec	accuracy=0.822551	lr=0.002370
Epoch[114] Batch [499]	Speed: 4738.894728 samples/sec	accuracy=0.822545	lr=0.002337
Epoch[114] Batch [549]	Speed: 4718.888762 samples/sec	accuracy=0.822525	lr=0.002304
Epoch[114] Batch [599]	Speed: 5098.442711 samples/sec	accuracy=0.822485	lr=0.002271
Epoch[114] Batch [649]	Speed: 4790.554147 samples/sec	accuracy=0.822484	lr=0.002238
Epoch[114] Batch [699]	Speed: 4819.196828 samples/sec	accuracy=0.822489	lr=0.002205
Epoch[114] Batch [749]	Speed: 5017.451397 samples/sec	accuracy=0.822495	lr=0.002173
Epoch[114] Batch [799]	Speed: 4857.929748 samples/sec	accuracy=0.822485	lr=0.002141
Epoch[114] Batch [849]	Speed: 4927.265545 samples/sec	accuracy=0.822474	lr=0.002110
Epoch[114] Batch [899]	Speed: 4765.054504 samples/sec	accuracy=0.822494	lr=0.002078
Epoch[114] Batch [949]	Speed: 5020.498279 samples/sec	accuracy=0.822522	lr=0.002047
Epoch[114] Batch [999]	Speed: 4942.173706 samples/sec	accuracy=0.822525	lr=0.002016
Epoch[114] Batch [1049]	Speed: 5112.152306 samples/sec	accuracy=0.822547	lr=0.001985
Epoch[114] Batch [1099]	Speed: 4969.276328 samples/sec	accuracy=0.822555	lr=0.001954
Epoch[114] Batch [1149]	Speed: 4906.233001 samples/sec	accuracy=0.822573	lr=0.001924
Epoch[114] Batch [1199]	Speed: 4933.060213 samples/sec	accuracy=0.822573	lr=0.001894
Epoch[114] Batch [1249]	Speed: 5005.988579 samples/sec	accuracy=0.822582	lr=0.001864
[Epoch 114] training: accuracy=0.822583
[Epoch 114] speed: 4841 samples/sec	time cost: 274.044396
[Epoch 114] validation: err-top1=0.228715 err-top5=0.063696
Epoch[115] Batch [49]	Speed: 4687.161299 samples/sec	accuracy=0.822490	lr=0.001834
Epoch[115] Batch [99]	Speed: 4783.824398 samples/sec	accuracy=0.822572	lr=0.001804
Epoch[115] Batch [149]	Speed: 4745.355119 samples/sec	accuracy=0.822624	lr=0.001775
Epoch[115] Batch [199]	Speed: 4780.101414 samples/sec	accuracy=0.822626	lr=0.001746
Epoch[115] Batch [249]	Speed: 4818.396772 samples/sec	accuracy=0.822584	lr=0.001718
Epoch[115] Batch [299]	Speed: 4868.668615 samples/sec	accuracy=0.822570	lr=0.001689
Epoch[115] Batch [349]	Speed: 4760.743701 samples/sec	accuracy=0.822576	lr=0.001661
Epoch[115] Batch [399]	Speed: 4804.121764 samples/sec	accuracy=0.822570	lr=0.001633
Epoch[115] Batch [449]	Speed: 4823.376811 samples/sec	accuracy=0.822602	lr=0.001605
Epoch[115] Batch [499]	Speed: 4691.735651 samples/sec	accuracy=0.822635	lr=0.001578
Epoch[115] Batch [549]	Speed: 4904.207589 samples/sec	accuracy=0.822660	lr=0.001551
Epoch[115] Batch [599]	Speed: 4839.229320 samples/sec	accuracy=0.822685	lr=0.001524
Epoch[115] Batch [649]	Speed: 4968.277275 samples/sec	accuracy=0.822684	lr=0.001497
Epoch[115] Batch [699]	Speed: 4595.094122 samples/sec	accuracy=0.822684	lr=0.001470
Epoch[115] Batch [749]	Speed: 4960.438608 samples/sec	accuracy=0.822683	lr=0.001444
Epoch[115] Batch [799]	Speed: 4662.035556 samples/sec	accuracy=0.822699	lr=0.001418
Epoch[115] Batch [849]	Speed: 4854.850503 samples/sec	accuracy=0.822715	lr=0.001392
Epoch[115] Batch [899]	Speed: 4715.710742 samples/sec	accuracy=0.822736	lr=0.001366
Epoch[115] Batch [949]	Speed: 5060.117119 samples/sec	accuracy=0.822735	lr=0.001341
Epoch[115] Batch [999]	Speed: 5084.466591 samples/sec	accuracy=0.822744	lr=0.001316
Epoch[115] Batch [1049]	Speed: 5132.328320 samples/sec	accuracy=0.822767	lr=0.001291
Epoch[115] Batch [1099]	Speed: 4667.515704 samples/sec	accuracy=0.822785	lr=0.001266
Epoch[115] Batch [1149]	Speed: 5119.857182 samples/sec	accuracy=0.822789	lr=0.001242
Epoch[115] Batch [1199]	Speed: 4871.880916 samples/sec	accuracy=0.822784	lr=0.001218
Epoch[115] Batch [1249]	Speed: 5135.326872 samples/sec	accuracy=0.822798	lr=0.001194
[Epoch 115] training: accuracy=0.822799
[Epoch 115] speed: 4843 samples/sec	time cost: 274.170256
[Epoch 115] validation: err-top1=0.227419 err-top5=0.064214
Epoch[116] Batch [49]	Speed: 4785.014082 samples/sec	accuracy=0.823218	lr=0.001170
Epoch[116] Batch [99]	Speed: 4915.696813 samples/sec	accuracy=0.823131	lr=0.001146
Epoch[116] Batch [149]	Speed: 4864.832732 samples/sec	accuracy=0.823085	lr=0.001123
Epoch[116] Batch [199]	Speed: 4798.658377 samples/sec	accuracy=0.823072	lr=0.001100
Epoch[116] Batch [249]	Speed: 4798.700196 samples/sec	accuracy=0.823047	lr=0.001077
Epoch[116] Batch [299]	Speed: 4877.217966 samples/sec	accuracy=0.822972	lr=0.001055
Epoch[116] Batch [349]	Speed: 4817.581203 samples/sec	accuracy=0.822904	lr=0.001032
Epoch[116] Batch [399]	Speed: 4828.725431 samples/sec	accuracy=0.822855	lr=0.001010
Epoch[116] Batch [449]	Speed: 4977.639042 samples/sec	accuracy=0.822839	lr=0.000989
Epoch[116] Batch [499]	Speed: 4899.098221 samples/sec	accuracy=0.822860	lr=0.000967
Epoch[116] Batch [549]	Speed: 4969.080625 samples/sec	accuracy=0.822884	lr=0.000946
Epoch[116] Batch [599]	Speed: 4802.233352 samples/sec	accuracy=0.822878	lr=0.000925
Epoch[116] Batch [649]	Speed: 5021.345142 samples/sec	accuracy=0.822873	lr=0.000904
Epoch[116] Batch [699]	Speed: 4858.073933 samples/sec	accuracy=0.822875	lr=0.000883
Epoch[116] Batch [749]	Speed: 4984.695413 samples/sec	accuracy=0.822894	lr=0.000863
Epoch[116] Batch [799]	Speed: 4719.121565 samples/sec	accuracy=0.822885	lr=0.000843
Epoch[116] Batch [849]	Speed: 4918.154525 samples/sec	accuracy=0.822897	lr=0.000823
Epoch[116] Batch [899]	Speed: 4863.056086 samples/sec	accuracy=0.822927	lr=0.000803
Epoch[116] Batch [949]	Speed: 4811.911098 samples/sec	accuracy=0.822957	lr=0.000784
Epoch[116] Batch [999]	Speed: 4743.677018 samples/sec	accuracy=0.822968	lr=0.000764
Epoch[116] Batch [1049]	Speed: 4958.507420 samples/sec	accuracy=0.822989	lr=0.000745
Epoch[116] Batch [1099]	Speed: 4853.665335 samples/sec	accuracy=0.822987	lr=0.000727
Epoch[116] Batch [1149]	Speed: 4932.766961 samples/sec	accuracy=0.822996	lr=0.000708
Epoch[116] Batch [1199]	Speed: 4794.843009 samples/sec	accuracy=0.823004	lr=0.000690
Epoch[116] Batch [1249]	Speed: 4948.115812 samples/sec	accuracy=0.823015	lr=0.000672
[Epoch 116] training: accuracy=0.823014
[Epoch 116] speed: 4864 samples/sec	time cost: 273.281157
[Epoch 116] validation: err-top1=0.226562 err-top5=0.064229
Epoch[117] Batch [49]	Speed: 4382.035806 samples/sec	accuracy=0.823233	lr=0.000654
Epoch[117] Batch [99]	Speed: 4691.042627 samples/sec	accuracy=0.823367	lr=0.000636
Epoch[117] Batch [149]	Speed: 4759.694542 samples/sec	accuracy=0.823392	lr=0.000619
Epoch[117] Batch [199]	Speed: 4910.144917 samples/sec	accuracy=0.823362	lr=0.000602
Epoch[117] Batch [249]	Speed: 4749.296072 samples/sec	accuracy=0.823370	lr=0.000585
Epoch[117] Batch [299]	Speed: 4968.942653 samples/sec	accuracy=0.823375	lr=0.000569
Epoch[117] Batch [349]	Speed: 4955.462795 samples/sec	accuracy=0.823366	lr=0.000552
Epoch[117] Batch [399]	Speed: 4968.731455 samples/sec	accuracy=0.823377	lr=0.000536
Epoch[117] Batch [449]	Speed: 4725.563371 samples/sec	accuracy=0.823361	lr=0.000520
Epoch[117] Batch [499]	Speed: 4854.821857 samples/sec	accuracy=0.823369	lr=0.000505
Epoch[117] Batch [549]	Speed: 4987.141085 samples/sec	accuracy=0.823374	lr=0.000489
Epoch[117] Batch [599]	Speed: 4796.315830 samples/sec	accuracy=0.823383	lr=0.000474
Epoch[117] Batch [649]	Speed: 4794.734348 samples/sec	accuracy=0.823367	lr=0.000459
Epoch[117] Batch [699]	Speed: 4922.772511 samples/sec	accuracy=0.823356	lr=0.000445
Epoch[117] Batch [749]	Speed: 4951.551662 samples/sec	accuracy=0.823341	lr=0.000430
Epoch[117] Batch [799]	Speed: 5030.658078 samples/sec	accuracy=0.823347	lr=0.000416
Epoch[117] Batch [849]	Speed: 4611.069778 samples/sec	accuracy=0.823338	lr=0.000402
Epoch[117] Batch [899]	Speed: 4890.156342 samples/sec	accuracy=0.823335	lr=0.000388
Epoch[117] Batch [949]	Speed: 4916.995443 samples/sec	accuracy=0.823335	lr=0.000375
Epoch[117] Batch [999]	Speed: 4903.407162 samples/sec	accuracy=0.823330	lr=0.000362
Epoch[117] Batch [1049]	Speed: 4817.952471 samples/sec	accuracy=0.823329	lr=0.000349
Epoch[117] Batch [1099]	Speed: 5030.740101 samples/sec	accuracy=0.823330	lr=0.000336
Epoch[117] Batch [1149]	Speed: 4900.725377 samples/sec	accuracy=0.823333	lr=0.000323
Epoch[117] Batch [1199]	Speed: 5002.946456 samples/sec	accuracy=0.823331	lr=0.000311
Epoch[117] Batch [1249]	Speed: 5058.478335 samples/sec	accuracy=0.823328	lr=0.000299
[Epoch 117] training: accuracy=0.823327
[Epoch 117] speed: 4853 samples/sec	time cost: 274.322466
[Epoch 117] validation: err-top1=0.226921 err-top5=0.063636
Epoch[118] Batch [49]	Speed: 4570.252137 samples/sec	accuracy=0.823225	lr=0.000287
Epoch[118] Batch [99]	Speed: 4926.689495 samples/sec	accuracy=0.823344	lr=0.000275
Epoch[118] Batch [149]	Speed: 4842.248158 samples/sec	accuracy=0.823252	lr=0.000264
Epoch[118] Batch [199]	Speed: 4749.693867 samples/sec	accuracy=0.823286	lr=0.000253
Epoch[118] Batch [249]	Speed: 4749.024680 samples/sec	accuracy=0.823302	lr=0.000242
Epoch[118] Batch [299]	Speed: 4927.454803 samples/sec	accuracy=0.823281	lr=0.000231
Epoch[118] Batch [349]	Speed: 4907.301338 samples/sec	accuracy=0.823270	lr=0.000221
Epoch[118] Batch [399]	Speed: 5045.500181 samples/sec	accuracy=0.823284	lr=0.000211
Epoch[118] Batch [449]	Speed: 4949.591907 samples/sec	accuracy=0.823294	lr=0.000201
Epoch[118] Batch [499]	Speed: 4758.886594 samples/sec	accuracy=0.823306	lr=0.000191
Epoch[118] Batch [549]	Speed: 4975.631034 samples/sec	accuracy=0.823308	lr=0.000182
Epoch[118] Batch [599]	Speed: 4841.812766 samples/sec	accuracy=0.823323	lr=0.000173
Epoch[118] Batch [649]	Speed: 4905.421156 samples/sec	accuracy=0.823323	lr=0.000164
Epoch[118] Batch [699]	Speed: 5082.228956 samples/sec	accuracy=0.823335	lr=0.000155
Epoch[118] Batch [749]	Speed: 4905.321543 samples/sec	accuracy=0.823333	lr=0.000147
Epoch[118] Batch [799]	Speed: 4893.904020 samples/sec	accuracy=0.823350	lr=0.000138
Epoch[118] Batch [849]	Speed: 5062.470532 samples/sec	accuracy=0.823336	lr=0.000130
Epoch[118] Batch [899]	Speed: 4897.299703 samples/sec	accuracy=0.823335	lr=0.000123
Epoch[118] Batch [949]	Speed: 4809.488383 samples/sec	accuracy=0.823316	lr=0.000115
Epoch[118] Batch [999]	Speed: 4928.986021 samples/sec	accuracy=0.823316	lr=0.000108
Epoch[118] Batch [1049]	Speed: 4951.859142 samples/sec	accuracy=0.823314	lr=0.000101
Epoch[118] Batch [1099]	Speed: 4733.695783 samples/sec	accuracy=0.823303	lr=0.000094
Epoch[118] Batch [1149]	Speed: 4788.430048 samples/sec	accuracy=0.823311	lr=0.000087
Epoch[118] Batch [1199]	Speed: 4760.716367 samples/sec	accuracy=0.823317	lr=0.000081
Epoch[118] Batch [1249]	Speed: 5140.851789 samples/sec	accuracy=0.823321	lr=0.000075
[Epoch 118] training: accuracy=0.823320
[Epoch 118] speed: 4877 samples/sec	time cost: 272.428880
[Epoch 118] validation: err-top1=0.226762 err-top5=0.063815
Epoch[119] Batch [49]	Speed: 4697.164547 samples/sec	accuracy=0.823030	lr=0.000069
Epoch[119] Batch [99]	Speed: 4770.078464 samples/sec	accuracy=0.823213	lr=0.000063
Epoch[119] Batch [149]	Speed: 4773.454350 samples/sec	accuracy=0.823202	lr=0.000058
Epoch[119] Batch [199]	Speed: 4861.277436 samples/sec	accuracy=0.823223	lr=0.000053
Epoch[119] Batch [249]	Speed: 4697.701530 samples/sec	accuracy=0.823289	lr=0.000048
Epoch[119] Batch [299]	Speed: 4782.808183 samples/sec	accuracy=0.823319	lr=0.000043
Epoch[119] Batch [349]	Speed: 4782.746934 samples/sec	accuracy=0.823327	lr=0.000039
Epoch[119] Batch [399]	Speed: 4781.745124 samples/sec	accuracy=0.823340	lr=0.000035
Epoch[119] Batch [449]	Speed: 4554.504167 samples/sec	accuracy=0.823333	lr=0.000031
Epoch[119] Batch [499]	Speed: 4696.734104 samples/sec	accuracy=0.823342	lr=0.000027
Epoch[119] Batch [549]	Speed: 4956.925085 samples/sec	accuracy=0.823360	lr=0.000023
Epoch[119] Batch [599]	Speed: 5037.020088 samples/sec	accuracy=0.823350	lr=0.000020
Epoch[119] Batch [649]	Speed: 4857.071301 samples/sec	accuracy=0.823359	lr=0.000017
Epoch[119] Batch [699]	Speed: 4727.024106 samples/sec	accuracy=0.823359	lr=0.000015
Epoch[119] Batch [749]	Speed: 4760.935582 samples/sec	accuracy=0.823360	lr=0.000012
Epoch[119] Batch [799]	Speed: 4907.057673 samples/sec	accuracy=0.823347	lr=0.000010
Epoch[119] Batch [849]	Speed: 4851.138706 samples/sec	accuracy=0.823340	lr=0.000008
Epoch[119] Batch [899]	Speed: 4850.847771 samples/sec	accuracy=0.823356	lr=0.000006
Epoch[119] Batch [949]	Speed: 4839.411439 samples/sec	accuracy=0.823359	lr=0.000004
Epoch[119] Batch [999]	Speed: 4922.741817 samples/sec	accuracy=0.823339	lr=0.000003
Epoch[119] Batch [1049]	Speed: 4795.799120 samples/sec	accuracy=0.823333	lr=0.000002
Epoch[119] Batch [1099]	Speed: 4779.484689 samples/sec	accuracy=0.823328	lr=0.000001
Epoch[119] Batch [1149]	Speed: 4863.570207 samples/sec	accuracy=0.823322	lr=0.000000
Epoch[119] Batch [1199]	Speed: 4893.103385 samples/sec	accuracy=0.823327	lr=0.000000
Epoch[119] Batch [1249]	Speed: 4982.554772 samples/sec	accuracy=0.823325	lr=0.000000
[Epoch 119] training: accuracy=0.823323
[Epoch 119] speed: 4809 samples/sec	time cost: 275.770618
[Epoch 119] validation: err-top1=0.226662 err-top5=0.063795