Namespace(batch_norm=False, batch_size=128, crop_ratio=0.875, data_dir='~/.mxnet/datasets/imagenet', dtype='float16', input_size=224, label_smoothing=True, last_gamma=True, log_interval=50, logging_file='resnet101_v1c_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='resnet101_v1c', momentum=0.9, no_wd=True, num_epochs=120, num_gpus=8, num_workers=60, rec_train='/media/ramdisk/rec/train.rec', rec_train_idx='/media/ramdisk/rec/train.idx', rec_val='/media/ramdisk/rec/val.rec', rec_val_idx='/media/ramdisk/rec/val.idx', resume_epoch=0, resume_params='', resume_states='', save_dir='params_resnet101_v1c_best', save_frequency=10, use_pretrained=False, use_rec=True, use_se=False, warmup_epochs=5, warmup_lr=0.0, wd=0.0001)
Namespace(batch_norm=False, batch_size=128, crop_ratio=0.875, data_dir='~/.mxnet/datasets/imagenet', dtype='float16', input_size=224, label_smoothing=True, last_gamma=True, log_interval=50, logging_file='resnet101_v1c_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='resnet101_v1c', momentum=0.9, no_wd=True, num_epochs=120, num_gpus=8, num_workers=60, rec_train='/media/ramdisk/rec/train.rec', rec_train_idx='/media/ramdisk/rec/train.idx', rec_val='/media/ramdisk/rec/val.rec', rec_val_idx='/media/ramdisk/rec/val.idx', resume_epoch=0, resume_params='', resume_states='', save_dir='params_resnet101_v1c_best', save_frequency=10, use_pretrained=False, use_rec=True, use_se=False, warmup_epochs=5, warmup_lr=0.0, wd=0.0001)
Namespace(batch_norm=False, batch_size=128, crop_ratio=0.875, data_dir='~/.mxnet/datasets/imagenet', dtype='float16', input_size=224, label_smoothing=True, last_gamma=True, log_interval=50, logging_file='resnet101_v1c_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='resnet101_v1c', momentum=0.9, no_wd=True, num_epochs=120, num_gpus=8, num_workers=60, rec_train='/media/ramdisk/rec/train.rec', rec_train_idx='/media/ramdisk/rec/train.idx', rec_val='/media/ramdisk/rec/val.rec', rec_val_idx='/media/ramdisk/rec/val.idx', resume_epoch=0, resume_params='', resume_states='', save_dir='params_resnet101_v1c_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: 1102.382504 samples/sec	accuracy=0.000996	lr=0.003133
Epoch[0] Batch [99]	Speed: 2717.174839 samples/sec	accuracy=0.001260	lr=0.006331
Epoch[0] Batch [149]	Speed: 2657.552249 samples/sec	accuracy=0.001745	lr=0.009528
Epoch[0] Batch [199]	Speed: 2710.932235 samples/sec	accuracy=0.002607	lr=0.012726
Epoch[0] Batch [249]	Speed: 2775.192293 samples/sec	accuracy=0.003586	lr=0.015923
Epoch[0] Batch [299]	Speed: 2734.622876 samples/sec	accuracy=0.004548	lr=0.019121
Epoch[0] Batch [349]	Speed: 2798.518878 samples/sec	accuracy=0.005508	lr=0.022318
Epoch[0] Batch [399]	Speed: 2842.093710 samples/sec	accuracy=0.006462	lr=0.025516
Epoch[0] Batch [449]	Speed: 2748.335783 samples/sec	accuracy=0.007326	lr=0.028713
Epoch[0] Batch [499]	Speed: 2799.132097 samples/sec	accuracy=0.008154	lr=0.031910
Epoch[0] Batch [549]	Speed: 2727.857788 samples/sec	accuracy=0.009013	lr=0.035108
Epoch[0] Batch [599]	Speed: 2822.741323 samples/sec	accuracy=0.009915	lr=0.038305
Epoch[0] Batch [649]	Speed: 2669.010121 samples/sec	accuracy=0.010892	lr=0.041503
Epoch[0] Batch [699]	Speed: 2756.691312 samples/sec	accuracy=0.011847	lr=0.044700
Epoch[0] Batch [749]	Speed: 2675.692107 samples/sec	accuracy=0.012741	lr=0.047898
Epoch[0] Batch [799]	Speed: 2802.950185 samples/sec	accuracy=0.013636	lr=0.051095
Epoch[0] Batch [849]	Speed: 2749.608299 samples/sec	accuracy=0.014497	lr=0.054293
Epoch[0] Batch [899]	Speed: 2781.330047 samples/sec	accuracy=0.015448	lr=0.057490
Epoch[0] Batch [949]	Speed: 2591.326677 samples/sec	accuracy=0.016377	lr=0.060687
Epoch[0] Batch [999]	Speed: 2643.820203 samples/sec	accuracy=0.017369	lr=0.063885
Epoch[0] Batch [1049]	Speed: 2606.951973 samples/sec	accuracy=0.018394	lr=0.067082
Epoch[0] Batch [1099]	Speed: 2676.270283 samples/sec	accuracy=0.019461	lr=0.070280
Epoch[0] Batch [1149]	Speed: 2732.936677 samples/sec	accuracy=0.020511	lr=0.073477
Epoch[0] Batch [1199]	Speed: 2750.985492 samples/sec	accuracy=0.021533	lr=0.076675
Epoch[0] Batch [1249]	Speed: 2899.152595 samples/sec	accuracy=0.022603	lr=0.079872
[Epoch 0] training: accuracy=0.022650
[Epoch 0] speed: 2579 samples/sec	time cost: 506.765036
[Epoch 0] validation: err-top1=0.943698 err-top5=0.844707
Epoch[1] Batch [49]	Speed: 2765.410540 samples/sec	accuracy=0.050547	lr=0.083133
Epoch[1] Batch [99]	Speed: 2684.118190 samples/sec	accuracy=0.051787	lr=0.086331
Epoch[1] Batch [149]	Speed: 2694.186962 samples/sec	accuracy=0.052988	lr=0.089528
Epoch[1] Batch [199]	Speed: 2702.302748 samples/sec	accuracy=0.054937	lr=0.092726
Epoch[1] Batch [249]	Speed: 2640.634272 samples/sec	accuracy=0.056539	lr=0.095923
Epoch[1] Batch [299]	Speed: 2786.835416 samples/sec	accuracy=0.058014	lr=0.099121
Epoch[1] Batch [349]	Speed: 2758.710604 samples/sec	accuracy=0.059665	lr=0.102318
Epoch[1] Batch [399]	Speed: 2751.179612 samples/sec	accuracy=0.061602	lr=0.105516
Epoch[1] Batch [449]	Speed: 2686.114267 samples/sec	accuracy=0.063299	lr=0.108713
Epoch[1] Batch [499]	Speed: 2752.795773 samples/sec	accuracy=0.065072	lr=0.111910
Epoch[1] Batch [549]	Speed: 2659.895853 samples/sec	accuracy=0.067129	lr=0.115108
Epoch[1] Batch [599]	Speed: 2643.299200 samples/sec	accuracy=0.069235	lr=0.118305
Epoch[1] Batch [649]	Speed: 2780.107653 samples/sec	accuracy=0.071143	lr=0.121503
Epoch[1] Batch [699]	Speed: 2744.778316 samples/sec	accuracy=0.073033	lr=0.124700
Epoch[1] Batch [749]	Speed: 2818.648948 samples/sec	accuracy=0.075052	lr=0.127898
Epoch[1] Batch [799]	Speed: 2703.751442 samples/sec	accuracy=0.077235	lr=0.131095
Epoch[1] Batch [849]	Speed: 2736.084129 samples/sec	accuracy=0.079418	lr=0.134293
Epoch[1] Batch [899]	Speed: 2697.776809 samples/sec	accuracy=0.081760	lr=0.137490
Epoch[1] Batch [949]	Speed: 2708.226937 samples/sec	accuracy=0.084227	lr=0.140687
Epoch[1] Batch [999]	Speed: 2687.609288 samples/sec	accuracy=0.086646	lr=0.143885
Epoch[1] Batch [1049]	Speed: 2685.869793 samples/sec	accuracy=0.089017	lr=0.147082
Epoch[1] Batch [1099]	Speed: 2662.468305 samples/sec	accuracy=0.091484	lr=0.150280
Epoch[1] Batch [1149]	Speed: 2672.372178 samples/sec	accuracy=0.093970	lr=0.153477
Epoch[1] Batch [1199]	Speed: 2669.338098 samples/sec	accuracy=0.096469	lr=0.156675
Epoch[1] Batch [1249]	Speed: 2892.174156 samples/sec	accuracy=0.099102	lr=0.159872
[Epoch 1] training: accuracy=0.099153
[Epoch 1] speed: 2716 samples/sec	time cost: 481.405550
[Epoch 1] validation: err-top1=0.864935 err-top5=0.684052
Epoch[2] Batch [49]	Speed: 2734.339412 samples/sec	accuracy=0.170684	lr=0.163133
Epoch[2] Batch [99]	Speed: 2697.075485 samples/sec	accuracy=0.171348	lr=0.166331
Epoch[2] Batch [149]	Speed: 2676.116837 samples/sec	accuracy=0.173757	lr=0.169528
Epoch[2] Batch [199]	Speed: 2670.920398 samples/sec	accuracy=0.176060	lr=0.172726
Epoch[2] Batch [249]	Speed: 2675.012280 samples/sec	accuracy=0.178477	lr=0.175923
Epoch[2] Batch [299]	Speed: 2740.669914 samples/sec	accuracy=0.181143	lr=0.179121
Epoch[2] Batch [349]	Speed: 2701.645632 samples/sec	accuracy=0.183758	lr=0.182318
Epoch[2] Batch [399]	Speed: 2696.101914 samples/sec	accuracy=0.186267	lr=0.185516
Epoch[2] Batch [449]	Speed: 2734.195005 samples/sec	accuracy=0.189076	lr=0.188713
Epoch[2] Batch [499]	Speed: 2646.366400 samples/sec	accuracy=0.191670	lr=0.191910
Epoch[2] Batch [549]	Speed: 2841.373212 samples/sec	accuracy=0.194574	lr=0.195108
Epoch[2] Batch [599]	Speed: 2711.315132 samples/sec	accuracy=0.197295	lr=0.198305
Epoch[2] Batch [649]	Speed: 2593.404324 samples/sec	accuracy=0.200093	lr=0.201503
Epoch[2] Batch [699]	Speed: 2734.061542 samples/sec	accuracy=0.202670	lr=0.204700
Epoch[2] Batch [749]	Speed: 2643.908413 samples/sec	accuracy=0.205434	lr=0.207898
Epoch[2] Batch [799]	Speed: 2723.858613 samples/sec	accuracy=0.208164	lr=0.211095
Epoch[2] Batch [849]	Speed: 2648.779106 samples/sec	accuracy=0.210695	lr=0.214293
Epoch[2] Batch [899]	Speed: 2684.781675 samples/sec	accuracy=0.213464	lr=0.217490
Epoch[2] Batch [949]	Speed: 2698.516883 samples/sec	accuracy=0.215899	lr=0.220687
Epoch[2] Batch [999]	Speed: 2635.831408 samples/sec	accuracy=0.218354	lr=0.223885
Epoch[2] Batch [1049]	Speed: 2668.976750 samples/sec	accuracy=0.220936	lr=0.227082
Epoch[2] Batch [1099]	Speed: 2660.843315 samples/sec	accuracy=0.223392	lr=0.230280
Epoch[2] Batch [1149]	Speed: 2597.005513 samples/sec	accuracy=0.225930	lr=0.233477
Epoch[2] Batch [1199]	Speed: 2660.570787 samples/sec	accuracy=0.228346	lr=0.236675
Epoch[2] Batch [1249]	Speed: 2809.304120 samples/sec	accuracy=0.230798	lr=0.239872
[Epoch 2] training: accuracy=0.230852
[Epoch 2] speed: 2688 samples/sec	time cost: 485.984897
[Epoch 2] validation: err-top1=0.720603 err-top5=0.469846
Epoch[3] Batch [49]	Speed: 2665.176648 samples/sec	accuracy=0.301211	lr=0.243133
Epoch[3] Batch [99]	Speed: 2760.486713 samples/sec	accuracy=0.300986	lr=0.246331
Epoch[3] Batch [149]	Speed: 2718.117796 samples/sec	accuracy=0.302311	lr=0.249528
Epoch[3] Batch [199]	Speed: 2726.995325 samples/sec	accuracy=0.304365	lr=0.252726
Epoch[3] Batch [249]	Speed: 2706.837689 samples/sec	accuracy=0.305547	lr=0.255923
Epoch[3] Batch [299]	Speed: 2709.540484 samples/sec	accuracy=0.307122	lr=0.259121
Epoch[3] Batch [349]	Speed: 2690.003129 samples/sec	accuracy=0.308747	lr=0.262318
Epoch[3] Batch [399]	Speed: 2663.584560 samples/sec	accuracy=0.310916	lr=0.265516
Epoch[3] Batch [449]	Speed: 2713.035277 samples/sec	accuracy=0.312281	lr=0.268713
Epoch[3] Batch [499]	Speed: 2679.142556 samples/sec	accuracy=0.314215	lr=0.271910
Epoch[3] Batch [549]	Speed: 2636.378502 samples/sec	accuracy=0.315565	lr=0.275108
Epoch[3] Batch [599]	Speed: 2722.677954 samples/sec	accuracy=0.317220	lr=0.278305
Epoch[3] Batch [649]	Speed: 2645.994552 samples/sec	accuracy=0.319226	lr=0.281503
Epoch[3] Batch [699]	Speed: 2679.756032 samples/sec	accuracy=0.321081	lr=0.284700
Epoch[3] Batch [749]	Speed: 2661.365156 samples/sec	accuracy=0.322477	lr=0.287898
Epoch[3] Batch [799]	Speed: 2680.372164 samples/sec	accuracy=0.324098	lr=0.291095
Epoch[3] Batch [849]	Speed: 2721.983050 samples/sec	accuracy=0.325795	lr=0.294293
Epoch[3] Batch [899]	Speed: 2671.759984 samples/sec	accuracy=0.327578	lr=0.297490
Epoch[3] Batch [949]	Speed: 2755.241340 samples/sec	accuracy=0.329021	lr=0.300687
Epoch[3] Batch [999]	Speed: 2685.596212 samples/sec	accuracy=0.330570	lr=0.303885
Epoch[3] Batch [1049]	Speed: 2722.030076 samples/sec	accuracy=0.332003	lr=0.307082
Epoch[3] Batch [1099]	Speed: 2660.509247 samples/sec	accuracy=0.333362	lr=0.310280
Epoch[3] Batch [1149]	Speed: 2679.719717 samples/sec	accuracy=0.334975	lr=0.313477
Epoch[3] Batch [1199]	Speed: 2751.643455 samples/sec	accuracy=0.336383	lr=0.316675
Epoch[3] Batch [1249]	Speed: 2870.530724 samples/sec	accuracy=0.337641	lr=0.319872
[Epoch 3] training: accuracy=0.337653
[Epoch 3] speed: 2700 samples/sec	time cost: 483.902936
[Epoch 3] validation: err-top1=0.658442 err-top5=0.392359
Epoch[4] Batch [49]	Speed: 2746.450782 samples/sec	accuracy=0.380449	lr=0.323133
Epoch[4] Batch [99]	Speed: 2593.027390 samples/sec	accuracy=0.380547	lr=0.326331
Epoch[4] Batch [149]	Speed: 2713.363914 samples/sec	accuracy=0.381406	lr=0.329528
Epoch[4] Batch [199]	Speed: 2679.137075 samples/sec	accuracy=0.382236	lr=0.332726
Epoch[4] Batch [249]	Speed: 2741.701041 samples/sec	accuracy=0.383582	lr=0.335923
Epoch[4] Batch [299]	Speed: 2639.370612 samples/sec	accuracy=0.383408	lr=0.339121
Epoch[4] Batch [349]	Speed: 2652.310767 samples/sec	accuracy=0.384782	lr=0.342318
Epoch[4] Batch [399]	Speed: 2755.247562 samples/sec	accuracy=0.385059	lr=0.345516
Epoch[4] Batch [449]	Speed: 2659.718848 samples/sec	accuracy=0.385859	lr=0.348713
Epoch[4] Batch [499]	Speed: 2821.653652 samples/sec	accuracy=0.386588	lr=0.351910
Epoch[4] Batch [549]	Speed: 2679.953808 samples/sec	accuracy=0.387438	lr=0.355108
Epoch[4] Batch [599]	Speed: 2695.605681 samples/sec	accuracy=0.388180	lr=0.358305
Epoch[4] Batch [649]	Speed: 2663.176779 samples/sec	accuracy=0.389120	lr=0.361503
Epoch[4] Batch [699]	Speed: 2815.377108 samples/sec	accuracy=0.389802	lr=0.364700
Epoch[4] Batch [749]	Speed: 2666.851249 samples/sec	accuracy=0.390493	lr=0.367898
Epoch[4] Batch [799]	Speed: 2774.566216 samples/sec	accuracy=0.391129	lr=0.371095
Epoch[4] Batch [849]	Speed: 2689.712299 samples/sec	accuracy=0.391930	lr=0.374293
Epoch[4] Batch [899]	Speed: 2612.677993 samples/sec	accuracy=0.392668	lr=0.377490
Epoch[4] Batch [949]	Speed: 2692.020593 samples/sec	accuracy=0.393489	lr=0.380687
Epoch[4] Batch [999]	Speed: 2703.197843 samples/sec	accuracy=0.394157	lr=0.383885
Epoch[4] Batch [1049]	Speed: 2652.014961 samples/sec	accuracy=0.395001	lr=0.387082
Epoch[4] Batch [1099]	Speed: 2643.793155 samples/sec	accuracy=0.395814	lr=0.390280
Epoch[4] Batch [1149]	Speed: 2652.152620 samples/sec	accuracy=0.396377	lr=0.393477
Epoch[4] Batch [1199]	Speed: 2730.201433 samples/sec	accuracy=0.397166	lr=0.396675
Epoch[4] Batch [1249]	Speed: 2874.813506 samples/sec	accuracy=0.397849	lr=0.399872
[Epoch 4] training: accuracy=0.397857
[Epoch 4] speed: 2698 samples/sec	time cost: 484.112248
[Epoch 4] validation: err-top1=0.588369 err-top5=0.318260
Epoch[5] Batch [49]	Speed: 2729.752390 samples/sec	accuracy=0.422090	lr=0.400000
Epoch[5] Batch [99]	Speed: 2715.624844 samples/sec	accuracy=0.422451	lr=0.400000
Epoch[5] Batch [149]	Speed: 2760.924167 samples/sec	accuracy=0.423359	lr=0.399999
Epoch[5] Batch [199]	Speed: 2757.602657 samples/sec	accuracy=0.424365	lr=0.399998
Epoch[5] Batch [249]	Speed: 2785.537612 samples/sec	accuracy=0.424852	lr=0.399997
Epoch[5] Batch [299]	Speed: 2680.679047 samples/sec	accuracy=0.426064	lr=0.399996
Epoch[5] Batch [349]	Speed: 2682.440538 samples/sec	accuracy=0.426186	lr=0.399994
Epoch[5] Batch [399]	Speed: 2646.753848 samples/sec	accuracy=0.426936	lr=0.399992
Epoch[5] Batch [449]	Speed: 2743.812251 samples/sec	accuracy=0.427624	lr=0.399990
Epoch[5] Batch [499]	Speed: 2745.147007 samples/sec	accuracy=0.428004	lr=0.399988
Epoch[5] Batch [549]	Speed: 2689.371684 samples/sec	accuracy=0.428722	lr=0.399986
Epoch[5] Batch [599]	Speed: 2703.572261 samples/sec	accuracy=0.429261	lr=0.399983
Epoch[5] Batch [649]	Speed: 2785.269541 samples/sec	accuracy=0.430116	lr=0.399980
Epoch[5] Batch [699]	Speed: 2729.228536 samples/sec	accuracy=0.430749	lr=0.399977
Epoch[5] Batch [749]	Speed: 2650.944475 samples/sec	accuracy=0.431639	lr=0.399973
Epoch[5] Batch [799]	Speed: 2688.707073 samples/sec	accuracy=0.432646	lr=0.399970
Epoch[5] Batch [849]	Speed: 2623.937898 samples/sec	accuracy=0.433202	lr=0.399966
Epoch[5] Batch [899]	Speed: 2746.173219 samples/sec	accuracy=0.434020	lr=0.399961
Epoch[5] Batch [949]	Speed: 2683.617035 samples/sec	accuracy=0.434641	lr=0.399957
Epoch[5] Batch [999]	Speed: 2717.370957 samples/sec	accuracy=0.435557	lr=0.399952
Epoch[5] Batch [1049]	Speed: 2704.867429 samples/sec	accuracy=0.436321	lr=0.399948
Epoch[5] Batch [1099]	Speed: 2804.330698 samples/sec	accuracy=0.436830	lr=0.399942
Epoch[5] Batch [1149]	Speed: 2781.513018 samples/sec	accuracy=0.437270	lr=0.399937
Epoch[5] Batch [1199]	Speed: 2677.273840 samples/sec	accuracy=0.437805	lr=0.399931
Epoch[5] Batch [1249]	Speed: 2812.224805 samples/sec	accuracy=0.438568	lr=0.399926
[Epoch 5] training: accuracy=0.438597
[Epoch 5] speed: 2719 samples/sec	time cost: 480.468877
[Epoch 5] validation: err-top1=0.527161 err-top5=0.266541
Epoch[6] Batch [49]	Speed: 2709.480111 samples/sec	accuracy=0.459355	lr=0.399919
Epoch[6] Batch [99]	Speed: 2694.962065 samples/sec	accuracy=0.462031	lr=0.399913
Epoch[6] Batch [149]	Speed: 2630.215712 samples/sec	accuracy=0.462611	lr=0.399907
Epoch[6] Batch [199]	Speed: 2680.435729 samples/sec	accuracy=0.463379	lr=0.399900
Epoch[6] Batch [249]	Speed: 2682.797665 samples/sec	accuracy=0.463254	lr=0.399893
Epoch[6] Batch [299]	Speed: 2649.711536 samples/sec	accuracy=0.463402	lr=0.399885
Epoch[6] Batch [349]	Speed: 2656.668324 samples/sec	accuracy=0.463825	lr=0.399878
Epoch[6] Batch [399]	Speed: 2704.426680 samples/sec	accuracy=0.464131	lr=0.399870
Epoch[6] Batch [449]	Speed: 2653.638004 samples/sec	accuracy=0.464679	lr=0.399862
Epoch[6] Batch [499]	Speed: 2720.633838 samples/sec	accuracy=0.465021	lr=0.399854
Epoch[6] Batch [549]	Speed: 2709.446575 samples/sec	accuracy=0.465137	lr=0.399846
Epoch[6] Batch [599]	Speed: 2684.828835 samples/sec	accuracy=0.465410	lr=0.399837
Epoch[6] Batch [649]	Speed: 2658.699440 samples/sec	accuracy=0.465772	lr=0.399828
Epoch[6] Batch [699]	Speed: 2693.783374 samples/sec	accuracy=0.465931	lr=0.399819
Epoch[6] Batch [749]	Speed: 2689.039910 samples/sec	accuracy=0.466250	lr=0.399809
Epoch[6] Batch [799]	Speed: 2721.622175 samples/sec	accuracy=0.466537	lr=0.399800
Epoch[6] Batch [849]	Speed: 2718.102349 samples/sec	accuracy=0.466695	lr=0.399790
Epoch[6] Batch [899]	Speed: 2669.623146 samples/sec	accuracy=0.467036	lr=0.399780
Epoch[6] Batch [949]	Speed: 2682.175628 samples/sec	accuracy=0.467392	lr=0.399769
Epoch[6] Batch [999]	Speed: 2723.281141 samples/sec	accuracy=0.467891	lr=0.399759
Epoch[6] Batch [1049]	Speed: 2661.458795 samples/sec	accuracy=0.468246	lr=0.399748
Epoch[6] Batch [1099]	Speed: 2658.712409 samples/sec	accuracy=0.468560	lr=0.399737
Epoch[6] Batch [1149]	Speed: 2682.503632 samples/sec	accuracy=0.469008	lr=0.399725
Epoch[6] Batch [1199]	Speed: 2660.902793 samples/sec	accuracy=0.469419	lr=0.399714
Epoch[6] Batch [1249]	Speed: 2838.819333 samples/sec	accuracy=0.469811	lr=0.399702
[Epoch 6] training: accuracy=0.469833
[Epoch 6] speed: 2687 samples/sec	time cost: 486.819553
[Epoch 6] validation: err-top1=0.527882 err-top5=0.266402
Epoch[7] Batch [49]	Speed: 2681.679049 samples/sec	accuracy=0.490215	lr=0.399690
Epoch[7] Batch [99]	Speed: 2756.378595 samples/sec	accuracy=0.488486	lr=0.399677
Epoch[7] Batch [149]	Speed: 2683.686121 samples/sec	accuracy=0.487461	lr=0.399665
Epoch[7] Batch [199]	Speed: 2799.533090 samples/sec	accuracy=0.487773	lr=0.399652
Epoch[7] Batch [249]	Speed: 2654.905321 samples/sec	accuracy=0.487398	lr=0.399639
Epoch[7] Batch [299]	Speed: 2679.477910 samples/sec	accuracy=0.487103	lr=0.399626
Epoch[7] Batch [349]	Speed: 2651.264067 samples/sec	accuracy=0.486596	lr=0.399613
Epoch[7] Batch [399]	Speed: 2691.148263 samples/sec	accuracy=0.486475	lr=0.399599
Epoch[7] Batch [449]	Speed: 2669.438571 samples/sec	accuracy=0.486936	lr=0.399585
Epoch[7] Batch [499]	Speed: 2690.717333 samples/sec	accuracy=0.486627	lr=0.399571
Epoch[7] Batch [549]	Speed: 2615.657679 samples/sec	accuracy=0.486559	lr=0.399556
Epoch[7] Batch [599]	Speed: 2743.003892 samples/sec	accuracy=0.486379	lr=0.399542
Epoch[7] Batch [649]	Speed: 2636.960890 samples/sec	accuracy=0.486457	lr=0.399527
Epoch[7] Batch [699]	Speed: 2717.011613 samples/sec	accuracy=0.486775	lr=0.399512
Epoch[7] Batch [749]	Speed: 2772.123597 samples/sec	accuracy=0.487030	lr=0.399496
Epoch[7] Batch [799]	Speed: 2724.218630 samples/sec	accuracy=0.487251	lr=0.399481
Epoch[7] Batch [849]	Speed: 2670.333374 samples/sec	accuracy=0.487606	lr=0.399465
Epoch[7] Batch [899]	Speed: 2619.085924 samples/sec	accuracy=0.487660	lr=0.399449
Epoch[7] Batch [949]	Speed: 2748.638797 samples/sec	accuracy=0.487810	lr=0.399432
Epoch[7] Batch [999]	Speed: 2661.366376 samples/sec	accuracy=0.487876	lr=0.399416
Epoch[7] Batch [1049]	Speed: 2651.095212 samples/sec	accuracy=0.488005	lr=0.399399
Epoch[7] Batch [1099]	Speed: 2745.587265 samples/sec	accuracy=0.488153	lr=0.399382
Epoch[7] Batch [1149]	Speed: 2664.579285 samples/sec	accuracy=0.488319	lr=0.399365
Epoch[7] Batch [1199]	Speed: 2658.348534 samples/sec	accuracy=0.488418	lr=0.399347
Epoch[7] Batch [1249]	Speed: 2939.260252 samples/sec	accuracy=0.488659	lr=0.399329
[Epoch 7] training: accuracy=0.488688
[Epoch 7] speed: 2698 samples/sec	time cost: 485.395642
[Epoch 7] validation: err-top1=0.497130 err-top5=0.238202
Epoch[8] Batch [49]	Speed: 2747.776575 samples/sec	accuracy=0.506895	lr=0.399311
Epoch[8] Batch [99]	Speed: 2773.887679 samples/sec	accuracy=0.506563	lr=0.399293
Epoch[8] Batch [149]	Speed: 2798.138082 samples/sec	accuracy=0.504635	lr=0.399274
Epoch[8] Batch [199]	Speed: 2689.013343 samples/sec	accuracy=0.504326	lr=0.399256
Epoch[8] Batch [249]	Speed: 2712.006209 samples/sec	accuracy=0.504383	lr=0.399237
Epoch[8] Batch [299]	Speed: 2780.243850 samples/sec	accuracy=0.503792	lr=0.399218
Epoch[8] Batch [349]	Speed: 2687.595632 samples/sec	accuracy=0.504188	lr=0.399198
Epoch[8] Batch [399]	Speed: 2792.779875 samples/sec	accuracy=0.503828	lr=0.399179
Epoch[8] Batch [449]	Speed: 2671.864994 samples/sec	accuracy=0.503713	lr=0.399159
Epoch[8] Batch [499]	Speed: 2766.953875 samples/sec	accuracy=0.503270	lr=0.399138
Epoch[8] Batch [549]	Speed: 2667.737092 samples/sec	accuracy=0.503471	lr=0.399118
Epoch[8] Batch [599]	Speed: 2711.005746 samples/sec	accuracy=0.503494	lr=0.399098
Epoch[8] Batch [649]	Speed: 2777.127346 samples/sec	accuracy=0.503879	lr=0.399077
Epoch[8] Batch [699]	Speed: 2749.689696 samples/sec	accuracy=0.503652	lr=0.399056
Epoch[8] Batch [749]	Speed: 2724.933694 samples/sec	accuracy=0.503740	lr=0.399034
Epoch[8] Batch [799]	Speed: 2698.296049 samples/sec	accuracy=0.503677	lr=0.399013
Epoch[8] Batch [849]	Speed: 2765.649549 samples/sec	accuracy=0.503695	lr=0.398991
Epoch[8] Batch [899]	Speed: 2831.151873 samples/sec	accuracy=0.503781	lr=0.398969
Epoch[8] Batch [949]	Speed: 2823.804324 samples/sec	accuracy=0.503749	lr=0.398947
Epoch[8] Batch [999]	Speed: 2642.912242 samples/sec	accuracy=0.503686	lr=0.398924
Epoch[8] Batch [1049]	Speed: 2707.444868 samples/sec	accuracy=0.503859	lr=0.398901
Epoch[8] Batch [1099]	Speed: 2739.753576 samples/sec	accuracy=0.504039	lr=0.398878
Epoch[8] Batch [1149]	Speed: 2699.952455 samples/sec	accuracy=0.503971	lr=0.398855
Epoch[8] Batch [1199]	Speed: 2769.760613 samples/sec	accuracy=0.504067	lr=0.398832
Epoch[8] Batch [1249]	Speed: 2940.401688 samples/sec	accuracy=0.504242	lr=0.398808
[Epoch 8] training: accuracy=0.504262
[Epoch 8] speed: 2743 samples/sec	time cost: 477.092371
[Epoch 8] validation: err-top1=0.471381 err-top5=0.216897
Epoch[9] Batch [49]	Speed: 2801.379638 samples/sec	accuracy=0.518496	lr=0.398784
Epoch[9] Batch [99]	Speed: 2730.135208 samples/sec	accuracy=0.518447	lr=0.398760
Epoch[9] Batch [149]	Speed: 2658.494981 samples/sec	accuracy=0.516764	lr=0.398735
Epoch[9] Batch [199]	Speed: 2676.377750 samples/sec	accuracy=0.516758	lr=0.398710
Epoch[9] Batch [249]	Speed: 2692.254881 samples/sec	accuracy=0.516328	lr=0.398686
Epoch[9] Batch [299]	Speed: 2692.025250 samples/sec	accuracy=0.515905	lr=0.398660
Epoch[9] Batch [349]	Speed: 2703.516885 samples/sec	accuracy=0.516094	lr=0.398635
Epoch[9] Batch [399]	Speed: 2664.990506 samples/sec	accuracy=0.516086	lr=0.398610
Epoch[9] Batch [449]	Speed: 2795.522168 samples/sec	accuracy=0.515241	lr=0.398584
Epoch[9] Batch [499]	Speed: 2629.210678 samples/sec	accuracy=0.515430	lr=0.398558
Epoch[9] Batch [549]	Speed: 2646.290777 samples/sec	accuracy=0.515394	lr=0.398531
Epoch[9] Batch [599]	Speed: 2673.556100 samples/sec	accuracy=0.515480	lr=0.398505
Epoch[9] Batch [649]	Speed: 2690.708567 samples/sec	accuracy=0.515475	lr=0.398478
Epoch[9] Batch [699]	Speed: 2687.146909 samples/sec	accuracy=0.515308	lr=0.398451
Epoch[9] Batch [749]	Speed: 2766.158476 samples/sec	accuracy=0.515121	lr=0.398424
Epoch[9] Batch [799]	Speed: 2765.776532 samples/sec	accuracy=0.515073	lr=0.398396
Epoch[9] Batch [849]	Speed: 2763.873836 samples/sec	accuracy=0.514959	lr=0.398369
Epoch[9] Batch [899]	Speed: 2697.235918 samples/sec	accuracy=0.515044	lr=0.398341
Epoch[9] Batch [949]	Speed: 2745.755452 samples/sec	accuracy=0.515042	lr=0.398312
Epoch[9] Batch [999]	Speed: 2713.232237 samples/sec	accuracy=0.515003	lr=0.398284
Epoch[9] Batch [1049]	Speed: 2823.854304 samples/sec	accuracy=0.515013	lr=0.398255
Epoch[9] Batch [1099]	Speed: 2696.184542 samples/sec	accuracy=0.515001	lr=0.398226
Epoch[9] Batch [1149]	Speed: 2683.923589 samples/sec	accuracy=0.515002	lr=0.398197
Epoch[9] Batch [1199]	Speed: 2657.570732 samples/sec	accuracy=0.514942	lr=0.398168
Epoch[9] Batch [1249]	Speed: 2859.028477 samples/sec	accuracy=0.514852	lr=0.398138
[Epoch 9] training: accuracy=0.514865
[Epoch 9] speed: 2713 samples/sec	time cost: 481.974912
[Epoch 9] validation: err-top1=0.439652 err-top5=0.193897
Epoch[10] Batch [49]	Speed: 2657.863600 samples/sec	accuracy=0.529766	lr=0.398108
Epoch[10] Batch [99]	Speed: 2632.637354 samples/sec	accuracy=0.528633	lr=0.398078
Epoch[10] Batch [149]	Speed: 2667.074253 samples/sec	accuracy=0.527051	lr=0.398048
Epoch[10] Batch [199]	Speed: 2708.244014 samples/sec	accuracy=0.525747	lr=0.398017
Epoch[10] Batch [249]	Speed: 2672.442748 samples/sec	accuracy=0.525469	lr=0.397986
Epoch[10] Batch [299]	Speed: 2657.359113 samples/sec	accuracy=0.525020	lr=0.397955
Epoch[10] Batch [349]	Speed: 2776.419987 samples/sec	accuracy=0.525045	lr=0.397924
Epoch[10] Batch [399]	Speed: 2782.975164 samples/sec	accuracy=0.524556	lr=0.397892
Epoch[10] Batch [449]	Speed: 2781.980804 samples/sec	accuracy=0.524071	lr=0.397861
Epoch[10] Batch [499]	Speed: 2813.073526 samples/sec	accuracy=0.523607	lr=0.397829
Epoch[10] Batch [549]	Speed: 2649.548632 samples/sec	accuracy=0.523608	lr=0.397796
Epoch[10] Batch [599]	Speed: 2753.704967 samples/sec	accuracy=0.523613	lr=0.397764
Epoch[10] Batch [649]	Speed: 2707.650918 samples/sec	accuracy=0.523208	lr=0.397731
Epoch[10] Batch [699]	Speed: 2664.701586 samples/sec	accuracy=0.523171	lr=0.397698
Epoch[10] Batch [749]	Speed: 2729.232906 samples/sec	accuracy=0.523212	lr=0.397665
Epoch[10] Batch [799]	Speed: 2660.829204 samples/sec	accuracy=0.523241	lr=0.397632
Epoch[10] Batch [849]	Speed: 2743.690958 samples/sec	accuracy=0.523248	lr=0.397598
Epoch[10] Batch [899]	Speed: 2743.634767 samples/sec	accuracy=0.523194	lr=0.397564
Epoch[10] Batch [949]	Speed: 2761.730481 samples/sec	accuracy=0.523097	lr=0.397530
Epoch[10] Batch [999]	Speed: 2708.629672 samples/sec	accuracy=0.523136	lr=0.397496
Epoch[10] Batch [1049]	Speed: 2712.525870 samples/sec	accuracy=0.523138	lr=0.397461
Epoch[10] Batch [1099]	Speed: 2692.097875 samples/sec	accuracy=0.523260	lr=0.397427
Epoch[10] Batch [1149]	Speed: 2636.721883 samples/sec	accuracy=0.523291	lr=0.397392
Epoch[10] Batch [1199]	Speed: 2690.024796 samples/sec	accuracy=0.523463	lr=0.397356
Epoch[10] Batch [1249]	Speed: 2869.251685 samples/sec	accuracy=0.523716	lr=0.397321
[Epoch 10] training: accuracy=0.523712
[Epoch 10] speed: 2711 samples/sec	time cost: 482.182981
[Epoch 10] validation: err-top1=0.444137 err-top5=0.195631
Epoch[11] Batch [49]	Speed: 2778.832503 samples/sec	accuracy=0.534922	lr=0.397284
Epoch[11] Batch [99]	Speed: 2801.232301 samples/sec	accuracy=0.536299	lr=0.397248
Epoch[11] Batch [149]	Speed: 2861.756147 samples/sec	accuracy=0.534727	lr=0.397212
Epoch[11] Batch [199]	Speed: 2780.970804 samples/sec	accuracy=0.534536	lr=0.397176
Epoch[11] Batch [249]	Speed: 2710.916185 samples/sec	accuracy=0.534156	lr=0.397139
Epoch[11] Batch [299]	Speed: 2647.596097 samples/sec	accuracy=0.533890	lr=0.397102
Epoch[11] Batch [349]	Speed: 2709.449720 samples/sec	accuracy=0.533468	lr=0.397065
Epoch[11] Batch [399]	Speed: 2710.420778 samples/sec	accuracy=0.533018	lr=0.397028
Epoch[11] Batch [449]	Speed: 2657.618716 samples/sec	accuracy=0.532528	lr=0.396990
Epoch[11] Batch [499]	Speed: 2739.473171 samples/sec	accuracy=0.532170	lr=0.396952
Epoch[11] Batch [549]	Speed: 2688.591141 samples/sec	accuracy=0.532033	lr=0.396914
Epoch[11] Batch [599]	Speed: 2686.650168 samples/sec	accuracy=0.531904	lr=0.396876
Epoch[11] Batch [649]	Speed: 2710.895617 samples/sec	accuracy=0.531513	lr=0.396837
Epoch[11] Batch [699]	Speed: 2758.178589 samples/sec	accuracy=0.531451	lr=0.396798
Epoch[11] Batch [749]	Speed: 2710.976142 samples/sec	accuracy=0.531490	lr=0.396759
Epoch[11] Batch [799]	Speed: 2733.202944 samples/sec	accuracy=0.531412	lr=0.396720
Epoch[11] Batch [849]	Speed: 2700.687440 samples/sec	accuracy=0.531434	lr=0.396680
Epoch[11] Batch [899]	Speed: 2694.002389 samples/sec	accuracy=0.531527	lr=0.396641
Epoch[11] Batch [949]	Speed: 2771.540897 samples/sec	accuracy=0.531164	lr=0.396601
Epoch[11] Batch [999]	Speed: 2697.108613 samples/sec	accuracy=0.531062	lr=0.396561
Epoch[11] Batch [1049]	Speed: 2714.288137 samples/sec	accuracy=0.530978	lr=0.396520
Epoch[11] Batch [1099]	Speed: 2755.613131 samples/sec	accuracy=0.531104	lr=0.396479
Epoch[11] Batch [1149]	Speed: 2687.939330 samples/sec	accuracy=0.531009	lr=0.396439
Epoch[11] Batch [1199]	Speed: 2711.466993 samples/sec	accuracy=0.530889	lr=0.396397
Epoch[11] Batch [1249]	Speed: 2827.533133 samples/sec	accuracy=0.530880	lr=0.396356
[Epoch 11] training: accuracy=0.530900
[Epoch 11] speed: 2727 samples/sec	time cost: 479.610466
[Epoch 11] validation: err-top1=0.439006 err-top5=0.194967
Epoch[12] Batch [49]	Speed: 2780.600169 samples/sec	accuracy=0.541777	lr=0.396314
Epoch[12] Batch [99]	Speed: 2735.164445 samples/sec	accuracy=0.540156	lr=0.396272
Epoch[12] Batch [149]	Speed: 2597.423817 samples/sec	accuracy=0.538685	lr=0.396230
Epoch[12] Batch [199]	Speed: 2712.348266 samples/sec	accuracy=0.538711	lr=0.396187
Epoch[12] Batch [249]	Speed: 2697.878146 samples/sec	accuracy=0.537668	lr=0.396145
Epoch[12] Batch [299]	Speed: 2695.927875 samples/sec	accuracy=0.537318	lr=0.396102
Epoch[12] Batch [349]	Speed: 2724.130785 samples/sec	accuracy=0.537327	lr=0.396059
Epoch[12] Batch [399]	Speed: 2668.234521 samples/sec	accuracy=0.536951	lr=0.396016
Epoch[12] Batch [449]	Speed: 2708.475805 samples/sec	accuracy=0.536871	lr=0.395972
Epoch[12] Batch [499]	Speed: 2655.233387 samples/sec	accuracy=0.536695	lr=0.395928
Epoch[12] Batch [549]	Speed: 2721.257913 samples/sec	accuracy=0.536752	lr=0.395885
Epoch[12] Batch [599]	Speed: 2702.243275 samples/sec	accuracy=0.536976	lr=0.395840
Epoch[12] Batch [649]	Speed: 2709.331275 samples/sec	accuracy=0.537213	lr=0.395796
Epoch[12] Batch [699]	Speed: 2675.504893 samples/sec	accuracy=0.537023	lr=0.395751
Epoch[12] Batch [749]	Speed: 2610.300910 samples/sec	accuracy=0.536816	lr=0.395706
Epoch[12] Batch [799]	Speed: 2804.559231 samples/sec	accuracy=0.536605	lr=0.395661
Epoch[12] Batch [849]	Speed: 2610.431829 samples/sec	accuracy=0.536519	lr=0.395616
Epoch[12] Batch [899]	Speed: 2697.893093 samples/sec	accuracy=0.536350	lr=0.395570
Epoch[12] Batch [949]	Speed: 2689.531606 samples/sec	accuracy=0.536160	lr=0.395524
Epoch[12] Batch [999]	Speed: 2639.776554 samples/sec	accuracy=0.536182	lr=0.395478
Epoch[12] Batch [1049]	Speed: 2607.871397 samples/sec	accuracy=0.536126	lr=0.395432
Epoch[12] Batch [1099]	Speed: 2719.846173 samples/sec	accuracy=0.536001	lr=0.395386
Epoch[12] Batch [1149]	Speed: 2731.996237 samples/sec	accuracy=0.535972	lr=0.395339
Epoch[12] Batch [1199]	Speed: 2708.045626 samples/sec	accuracy=0.536038	lr=0.395292
Epoch[12] Batch [1249]	Speed: 2811.252899 samples/sec	accuracy=0.536467	lr=0.395245
[Epoch 12] training: accuracy=0.536477
[Epoch 12] speed: 2693 samples/sec	time cost: 485.543663
[Epoch 12] validation: err-top1=0.442662 err-top5=0.191207
Epoch[13] Batch [49]	Speed: 2750.378144 samples/sec	accuracy=0.549219	lr=0.395196
Epoch[13] Batch [99]	Speed: 2681.991859 samples/sec	accuracy=0.548291	lr=0.395149
Epoch[13] Batch [149]	Speed: 2681.306183 samples/sec	accuracy=0.546719	lr=0.395101
Epoch[13] Batch [199]	Speed: 2672.832846 samples/sec	accuracy=0.546382	lr=0.395052
Epoch[13] Batch [249]	Speed: 2677.637873 samples/sec	accuracy=0.545996	lr=0.395004
Epoch[13] Batch [299]	Speed: 2725.670789 samples/sec	accuracy=0.545368	lr=0.394955
Epoch[13] Batch [349]	Speed: 2631.089018 samples/sec	accuracy=0.545106	lr=0.394907
Epoch[13] Batch [399]	Speed: 2678.573195 samples/sec	accuracy=0.544297	lr=0.394858
Epoch[13] Batch [449]	Speed: 2724.901641 samples/sec	accuracy=0.544384	lr=0.394808
Epoch[13] Batch [499]	Speed: 2751.974742 samples/sec	accuracy=0.544852	lr=0.394759
Epoch[13] Batch [549]	Speed: 2682.364815 samples/sec	accuracy=0.545036	lr=0.394709
Epoch[13] Batch [599]	Speed: 2723.916727 samples/sec	accuracy=0.544813	lr=0.394659
Epoch[13] Batch [649]	Speed: 2737.665462 samples/sec	accuracy=0.544443	lr=0.394609
Epoch[13] Batch [699]	Speed: 2748.775164 samples/sec	accuracy=0.544319	lr=0.394558
Epoch[13] Batch [749]	Speed: 2768.428016 samples/sec	accuracy=0.543996	lr=0.394507
Epoch[13] Batch [799]	Speed: 2680.398125 samples/sec	accuracy=0.543723	lr=0.394457
Epoch[13] Batch [849]	Speed: 2691.394407 samples/sec	accuracy=0.543366	lr=0.394405
Epoch[13] Batch [899]	Speed: 2728.658526 samples/sec	accuracy=0.543094	lr=0.394354
Epoch[13] Batch [949]	Speed: 2708.199717 samples/sec	accuracy=0.542943	lr=0.394302
Epoch[13] Batch [999]	Speed: 2869.035409 samples/sec	accuracy=0.542911	lr=0.394250
Epoch[13] Batch [1049]	Speed: 2730.679479 samples/sec	accuracy=0.542763	lr=0.394198
Epoch[13] Batch [1099]	Speed: 2730.275542 samples/sec	accuracy=0.542829	lr=0.394146
Epoch[13] Batch [1149]	Speed: 2701.066565 samples/sec	accuracy=0.542655	lr=0.394093
Epoch[13] Batch [1199]	Speed: 2746.717616 samples/sec	accuracy=0.542646	lr=0.394041
Epoch[13] Batch [1249]	Speed: 2867.225370 samples/sec	accuracy=0.542583	lr=0.393988
[Epoch 13] training: accuracy=0.542597
[Epoch 13] speed: 2720 samples/sec	time cost: 480.482871
[Epoch 13] validation: err-top1=0.454939 err-top5=0.207729
Epoch[14] Batch [49]	Speed: 2761.705264 samples/sec	accuracy=0.547852	lr=0.393933
Epoch[14] Batch [99]	Speed: 2729.297666 samples/sec	accuracy=0.553213	lr=0.393880
Epoch[14] Batch [149]	Speed: 2695.200552 samples/sec	accuracy=0.552240	lr=0.393826
Epoch[14] Batch [199]	Speed: 2693.068150 samples/sec	accuracy=0.551611	lr=0.393772
Epoch[14] Batch [249]	Speed: 2758.407279 samples/sec	accuracy=0.551297	lr=0.393718
Epoch[14] Batch [299]	Speed: 2727.993695 samples/sec	accuracy=0.550501	lr=0.393664
Epoch[14] Batch [349]	Speed: 2681.722449 samples/sec	accuracy=0.549621	lr=0.393609
Epoch[14] Batch [399]	Speed: 2694.401309 samples/sec	accuracy=0.548823	lr=0.393554
Epoch[14] Batch [449]	Speed: 2716.046236 samples/sec	accuracy=0.548325	lr=0.393499
Epoch[14] Batch [499]	Speed: 2751.573048 samples/sec	accuracy=0.548096	lr=0.393444
Epoch[14] Batch [549]	Speed: 2724.892202 samples/sec	accuracy=0.548134	lr=0.393388
Epoch[14] Batch [599]	Speed: 2722.711576 samples/sec	accuracy=0.548073	lr=0.393332
Epoch[14] Batch [649]	Speed: 2660.598673 samples/sec	accuracy=0.547831	lr=0.393276
Epoch[14] Batch [699]	Speed: 2739.634983 samples/sec	accuracy=0.547709	lr=0.393220
Epoch[14] Batch [749]	Speed: 2663.738093 samples/sec	accuracy=0.547786	lr=0.393163
Epoch[14] Batch [799]	Speed: 2733.924612 samples/sec	accuracy=0.547771	lr=0.393107
Epoch[14] Batch [849]	Speed: 2772.144889 samples/sec	accuracy=0.547850	lr=0.393050
Epoch[14] Batch [899]	Speed: 2666.256610 samples/sec	accuracy=0.547714	lr=0.392993
Epoch[14] Batch [949]	Speed: 2649.591031 samples/sec	accuracy=0.547576	lr=0.392935
Epoch[14] Batch [999]	Speed: 2796.964402 samples/sec	accuracy=0.547547	lr=0.392877
Epoch[14] Batch [1049]	Speed: 2669.123374 samples/sec	accuracy=0.547460	lr=0.392820
Epoch[14] Batch [1099]	Speed: 2680.783087 samples/sec	accuracy=0.547338	lr=0.392762
Epoch[14] Batch [1149]	Speed: 2672.518776 samples/sec	accuracy=0.547325	lr=0.392703
Epoch[14] Batch [1199]	Speed: 2777.973086 samples/sec	accuracy=0.547314	lr=0.392645
Epoch[14] Batch [1249]	Speed: 2863.998088 samples/sec	accuracy=0.547121	lr=0.392586
[Epoch 14] training: accuracy=0.547132
[Epoch 14] speed: 2717 samples/sec	time cost: 481.562341
[Epoch 14] validation: err-top1=0.423250 err-top5=0.177037
Epoch[15] Batch [49]	Speed: 2657.492657 samples/sec	accuracy=0.559375	lr=0.392526
Epoch[15] Batch [99]	Speed: 2668.915783 samples/sec	accuracy=0.558799	lr=0.392466
Epoch[15] Batch [149]	Speed: 2650.756061 samples/sec	accuracy=0.556940	lr=0.392407
Epoch[15] Batch [199]	Speed: 2740.323580 samples/sec	accuracy=0.555376	lr=0.392347
Epoch[15] Batch [249]	Speed: 2693.560476 samples/sec	accuracy=0.555480	lr=0.392287
Epoch[15] Batch [299]	Speed: 2681.876540 samples/sec	accuracy=0.554616	lr=0.392227
Epoch[15] Batch [349]	Speed: 2707.714623 samples/sec	accuracy=0.553666	lr=0.392167
Epoch[15] Batch [399]	Speed: 2619.154922 samples/sec	accuracy=0.553071	lr=0.392106
Epoch[15] Batch [449]	Speed: 2756.038537 samples/sec	accuracy=0.552773	lr=0.392045
Epoch[15] Batch [499]	Speed: 2774.518576 samples/sec	accuracy=0.552416	lr=0.391984
Epoch[15] Batch [549]	Speed: 2712.000181 samples/sec	accuracy=0.552496	lr=0.391923
Epoch[15] Batch [599]	Speed: 2745.139498 samples/sec	accuracy=0.552181	lr=0.391861
Epoch[15] Batch [649]	Speed: 2720.007018 samples/sec	accuracy=0.552135	lr=0.391799
Epoch[15] Batch [699]	Speed: 2718.538792 samples/sec	accuracy=0.551787	lr=0.391737
Epoch[15] Batch [749]	Speed: 2773.657490 samples/sec	accuracy=0.551577	lr=0.391675
Epoch[15] Batch [799]	Speed: 2745.733932 samples/sec	accuracy=0.551459	lr=0.391613
Epoch[15] Batch [849]	Speed: 2696.649631 samples/sec	accuracy=0.551499	lr=0.391550
Epoch[15] Batch [899]	Speed: 2652.113872 samples/sec	accuracy=0.551458	lr=0.391487
Epoch[15] Batch [949]	Speed: 2612.163186 samples/sec	accuracy=0.551332	lr=0.391424
Epoch[15] Batch [999]	Speed: 2674.885399 samples/sec	accuracy=0.551182	lr=0.391361
Epoch[15] Batch [1049]	Speed: 2697.649013 samples/sec	accuracy=0.551037	lr=0.391297
Epoch[15] Batch [1099]	Speed: 2736.126520 samples/sec	accuracy=0.550883	lr=0.391233
Epoch[15] Batch [1149]	Speed: 2683.888838 samples/sec	accuracy=0.551146	lr=0.391169
Epoch[15] Batch [1199]	Speed: 2670.661245 samples/sec	accuracy=0.551055	lr=0.391105
Epoch[15] Batch [1249]	Speed: 2956.288580 samples/sec	accuracy=0.550873	lr=0.391040
[Epoch 15] training: accuracy=0.550880
[Epoch 15] speed: 2706 samples/sec	time cost: 483.220087
[Epoch 15] validation: err-top1=0.431182 err-top5=0.185427
Epoch[16] Batch [49]	Speed: 2876.352799 samples/sec	accuracy=0.559590	lr=0.390974
Epoch[16] Batch [99]	Speed: 2719.023606 samples/sec	accuracy=0.559756	lr=0.390909
Epoch[16] Batch [149]	Speed: 2730.624306 samples/sec	accuracy=0.559648	lr=0.390844
Epoch[16] Batch [199]	Speed: 2812.467297 samples/sec	accuracy=0.560186	lr=0.390779
Epoch[16] Batch [249]	Speed: 2666.156211 samples/sec	accuracy=0.559508	lr=0.390713
Epoch[16] Batch [299]	Speed: 2811.291211 samples/sec	accuracy=0.558392	lr=0.390647
Epoch[16] Batch [349]	Speed: 2658.880557 samples/sec	accuracy=0.558326	lr=0.390581
Epoch[16] Batch [399]	Speed: 2647.424869 samples/sec	accuracy=0.557869	lr=0.390515
Epoch[16] Batch [449]	Speed: 2717.105909 samples/sec	accuracy=0.557324	lr=0.390448
Epoch[16] Batch [499]	Speed: 2699.720797 samples/sec	accuracy=0.556936	lr=0.390381
Epoch[16] Batch [549]	Speed: 2669.049496 samples/sec	accuracy=0.556692	lr=0.390314
Epoch[16] Batch [599]	Speed: 2772.939402 samples/sec	accuracy=0.556590	lr=0.390247
Epoch[16] Batch [649]	Speed: 2681.030284 samples/sec	accuracy=0.556370	lr=0.390180
Epoch[16] Batch [699]	Speed: 2689.221648 samples/sec	accuracy=0.556300	lr=0.390112
Epoch[16] Batch [749]	Speed: 2675.786224 samples/sec	accuracy=0.556090	lr=0.390044
Epoch[16] Batch [799]	Speed: 2744.504808 samples/sec	accuracy=0.555844	lr=0.389976
Epoch[16] Batch [849]	Speed: 2652.794103 samples/sec	accuracy=0.555655	lr=0.389907
Epoch[16] Batch [899]	Speed: 2702.828698 samples/sec	accuracy=0.555403	lr=0.389839
Epoch[16] Batch [949]	Speed: 2709.130609 samples/sec	accuracy=0.555188	lr=0.389770
Epoch[16] Batch [999]	Speed: 2811.551137 samples/sec	accuracy=0.555099	lr=0.389701
Epoch[16] Batch [1049]	Speed: 2794.236223 samples/sec	accuracy=0.554715	lr=0.389632
Epoch[16] Batch [1099]	Speed: 2656.209924 samples/sec	accuracy=0.554672	lr=0.389562
Epoch[16] Batch [1149]	Speed: 2730.579586 samples/sec	accuracy=0.554642	lr=0.389492
Epoch[16] Batch [1199]	Speed: 2696.762973 samples/sec	accuracy=0.554644	lr=0.389422
Epoch[16] Batch [1249]	Speed: 2915.836238 samples/sec	accuracy=0.554424	lr=0.389352
[Epoch 16] training: accuracy=0.554432
[Epoch 16] speed: 2725 samples/sec	time cost: 479.819823
[Epoch 16] validation: err-top1=0.410096 err-top5=0.168427
Epoch[17] Batch [49]	Speed: 2809.051333 samples/sec	accuracy=0.564707	lr=0.389280
Epoch[17] Batch [99]	Speed: 2723.118492 samples/sec	accuracy=0.563232	lr=0.389210
Epoch[17] Batch [149]	Speed: 2752.036882 samples/sec	accuracy=0.563034	lr=0.389139
Epoch[17] Batch [199]	Speed: 2748.686327 samples/sec	accuracy=0.561577	lr=0.389068
Epoch[17] Batch [249]	Speed: 2737.445013 samples/sec	accuracy=0.561098	lr=0.388996
Epoch[17] Batch [299]	Speed: 2666.604176 samples/sec	accuracy=0.560918	lr=0.388925
Epoch[17] Batch [349]	Speed: 2741.930823 samples/sec	accuracy=0.560778	lr=0.388853
Epoch[17] Batch [399]	Speed: 2681.802356 samples/sec	accuracy=0.560640	lr=0.388781
Epoch[17] Batch [449]	Speed: 2678.610213 samples/sec	accuracy=0.560328	lr=0.388709
Epoch[17] Batch [499]	Speed: 2679.282444 samples/sec	accuracy=0.559930	lr=0.388636
Epoch[17] Batch [549]	Speed: 2682.556778 samples/sec	accuracy=0.559759	lr=0.388564
Epoch[17] Batch [599]	Speed: 2714.285118 samples/sec	accuracy=0.559753	lr=0.388491
Epoch[17] Batch [649]	Speed: 2652.267199 samples/sec	accuracy=0.559614	lr=0.388418
Epoch[17] Batch [699]	Speed: 2697.374449 samples/sec	accuracy=0.559008	lr=0.388344
Epoch[17] Batch [749]	Speed: 2641.887669 samples/sec	accuracy=0.558887	lr=0.388271
Epoch[17] Batch [799]	Speed: 2734.232532 samples/sec	accuracy=0.558656	lr=0.388197
Epoch[17] Batch [849]	Speed: 2741.671989 samples/sec	accuracy=0.558344	lr=0.388123
Epoch[17] Batch [899]	Speed: 2636.170018 samples/sec	accuracy=0.558136	lr=0.388049
Epoch[17] Batch [949]	Speed: 2781.082414 samples/sec	accuracy=0.558123	lr=0.387974
Epoch[17] Batch [999]	Speed: 2754.789677 samples/sec	accuracy=0.558065	lr=0.387900
Epoch[17] Batch [1049]	Speed: 2705.137250 samples/sec	accuracy=0.557899	lr=0.387825
Epoch[17] Batch [1099]	Speed: 2669.495481 samples/sec	accuracy=0.557683	lr=0.387750
Epoch[17] Batch [1149]	Speed: 2657.538699 samples/sec	accuracy=0.557737	lr=0.387674
Epoch[17] Batch [1199]	Speed: 2665.346905 samples/sec	accuracy=0.557758	lr=0.387599
Epoch[17] Batch [1249]	Speed: 2781.122897 samples/sec	accuracy=0.557499	lr=0.387523
[Epoch 17] training: accuracy=0.557509
[Epoch 17] speed: 2706 samples/sec	time cost: 483.392106
[Epoch 17] validation: err-top1=0.425863 err-top5=0.180583
Epoch[18] Batch [49]	Speed: 2825.988549 samples/sec	accuracy=0.561992	lr=0.387445
Epoch[18] Batch [99]	Speed: 2795.635313 samples/sec	accuracy=0.562783	lr=0.387369
Epoch[18] Batch [149]	Speed: 2761.975390 samples/sec	accuracy=0.564355	lr=0.387293
Epoch[18] Batch [199]	Speed: 2855.432631 samples/sec	accuracy=0.563491	lr=0.387216
Epoch[18] Batch [249]	Speed: 2617.760755 samples/sec	accuracy=0.563418	lr=0.387139
Epoch[18] Batch [299]	Speed: 2672.498555 samples/sec	accuracy=0.562923	lr=0.387062
Epoch[18] Batch [349]	Speed: 2653.778290 samples/sec	accuracy=0.562453	lr=0.386984
Epoch[18] Batch [399]	Speed: 2688.262655 samples/sec	accuracy=0.562332	lr=0.386907
Epoch[18] Batch [449]	Speed: 2653.804985 samples/sec	accuracy=0.562042	lr=0.386829
Epoch[18] Batch [499]	Speed: 2677.687453 samples/sec	accuracy=0.561623	lr=0.386751
Epoch[18] Batch [549]	Speed: 2723.540451 samples/sec	accuracy=0.561248	lr=0.386673
Epoch[18] Batch [599]	Speed: 2756.943504 samples/sec	accuracy=0.560993	lr=0.386594
Epoch[18] Batch [649]	Speed: 2701.000013 samples/sec	accuracy=0.560733	lr=0.386515
Epoch[18] Batch [699]	Speed: 2638.280422 samples/sec	accuracy=0.560430	lr=0.386436
Epoch[18] Batch [749]	Speed: 2705.501470 samples/sec	accuracy=0.560319	lr=0.386357
Epoch[18] Batch [799]	Speed: 2682.228894 samples/sec	accuracy=0.560183	lr=0.386278
Epoch[18] Batch [849]	Speed: 2720.844623 samples/sec	accuracy=0.560118	lr=0.386198
Epoch[18] Batch [899]	Speed: 2703.526857 samples/sec	accuracy=0.559778	lr=0.386119
Epoch[18] Batch [949]	Speed: 2728.712788 samples/sec	accuracy=0.559841	lr=0.386038
Epoch[18] Batch [999]	Speed: 2662.939401 samples/sec	accuracy=0.559925	lr=0.385958
Epoch[18] Batch [1049]	Speed: 2744.051048 samples/sec	accuracy=0.559874	lr=0.385878
Epoch[18] Batch [1099]	Speed: 2754.586743 samples/sec	accuracy=0.559788	lr=0.385797
Epoch[18] Batch [1149]	Speed: 2691.727201 samples/sec	accuracy=0.559689	lr=0.385716
Epoch[18] Batch [1199]	Speed: 2736.777499 samples/sec	accuracy=0.559693	lr=0.385635
Epoch[18] Batch [1249]	Speed: 2805.642705 samples/sec	accuracy=0.559590	lr=0.385554
[Epoch 18] training: accuracy=0.559579
[Epoch 18] speed: 2715 samples/sec	time cost: 482.166782
[Epoch 18] validation: err-top1=0.409578 err-top5=0.167550
Epoch[19] Batch [49]	Speed: 2782.565127 samples/sec	accuracy=0.570801	lr=0.385470
Epoch[19] Batch [99]	Speed: 2717.109519 samples/sec	accuracy=0.567939	lr=0.385388
Epoch[19] Batch [149]	Speed: 2667.052424 samples/sec	accuracy=0.568626	lr=0.385306
Epoch[19] Batch [199]	Speed: 2795.700787 samples/sec	accuracy=0.568779	lr=0.385224
Epoch[19] Batch [249]	Speed: 2783.641809 samples/sec	accuracy=0.568379	lr=0.385142
Epoch[19] Batch [299]	Speed: 2732.098007 samples/sec	accuracy=0.568405	lr=0.385059
Epoch[19] Batch [349]	Speed: 2769.398210 samples/sec	accuracy=0.568036	lr=0.384976
Epoch[19] Batch [399]	Speed: 2713.507741 samples/sec	accuracy=0.567344	lr=0.384893
Epoch[19] Batch [449]	Speed: 2723.039834 samples/sec	accuracy=0.567112	lr=0.384810
Epoch[19] Batch [499]	Speed: 2791.582126 samples/sec	accuracy=0.566957	lr=0.384726
Epoch[19] Batch [549]	Speed: 2685.887833 samples/sec	accuracy=0.566451	lr=0.384642
Epoch[19] Batch [599]	Speed: 2665.078877 samples/sec	accuracy=0.566252	lr=0.384558
Epoch[19] Batch [649]	Speed: 2681.046718 samples/sec	accuracy=0.565700	lr=0.384474
Epoch[19] Batch [699]	Speed: 2803.111790 samples/sec	accuracy=0.565604	lr=0.384389
Epoch[19] Batch [749]	Speed: 2720.727111 samples/sec	accuracy=0.565253	lr=0.384305
Epoch[19] Batch [799]	Speed: 2698.034539 samples/sec	accuracy=0.564908	lr=0.384220
Epoch[19] Batch [849]	Speed: 2735.145320 samples/sec	accuracy=0.564710	lr=0.384135
Epoch[19] Batch [899]	Speed: 2656.872896 samples/sec	accuracy=0.564430	lr=0.384049
Epoch[19] Batch [949]	Speed: 2669.191019 samples/sec	accuracy=0.564409	lr=0.383964
Epoch[19] Batch [999]	Speed: 2685.728478 samples/sec	accuracy=0.564083	lr=0.383878
Epoch[19] Batch [1049]	Speed: 2678.006480 samples/sec	accuracy=0.563973	lr=0.383792
Epoch[19] Batch [1099]	Speed: 2742.548701 samples/sec	accuracy=0.563921	lr=0.383706
Epoch[19] Batch [1149]	Speed: 2729.856733 samples/sec	accuracy=0.563858	lr=0.383619
Epoch[19] Batch [1199]	Speed: 2804.958264 samples/sec	accuracy=0.563612	lr=0.383533
Epoch[19] Batch [1249]	Speed: 2842.168187 samples/sec	accuracy=0.563536	lr=0.383446
[Epoch 19] training: accuracy=0.563570
[Epoch 19] speed: 2728 samples/sec	time cost: 479.549174
[Epoch 19] validation: err-top1=0.404257 err-top5=0.165836
Epoch[20] Batch [49]	Speed: 2743.597437 samples/sec	accuracy=0.568555	lr=0.383357
Epoch[20] Batch [99]	Speed: 2875.221619 samples/sec	accuracy=0.568965	lr=0.383270
Epoch[20] Batch [149]	Speed: 2716.623394 samples/sec	accuracy=0.569818	lr=0.383182
Epoch[20] Batch [199]	Speed: 2775.739037 samples/sec	accuracy=0.570039	lr=0.383094
Epoch[20] Batch [249]	Speed: 2739.743195 samples/sec	accuracy=0.570781	lr=0.383006
Epoch[20] Batch [299]	Speed: 2612.791062 samples/sec	accuracy=0.570527	lr=0.382918
Epoch[20] Batch [349]	Speed: 2754.880075 samples/sec	accuracy=0.570315	lr=0.382830
Epoch[20] Batch [399]	Speed: 2754.985183 samples/sec	accuracy=0.569353	lr=0.382741
Epoch[20] Batch [449]	Speed: 2746.555247 samples/sec	accuracy=0.569154	lr=0.382652
Epoch[20] Batch [499]	Speed: 2666.457630 samples/sec	accuracy=0.569248	lr=0.382563
Epoch[20] Batch [549]	Speed: 2699.668701 samples/sec	accuracy=0.568739	lr=0.382474
Epoch[20] Batch [599]	Speed: 2631.131183 samples/sec	accuracy=0.568218	lr=0.382384
Epoch[20] Batch [649]	Speed: 2780.669766 samples/sec	accuracy=0.567984	lr=0.382295
Epoch[20] Batch [699]	Speed: 2805.261470 samples/sec	accuracy=0.567393	lr=0.382205
Epoch[20] Batch [749]	Speed: 2683.721168 samples/sec	accuracy=0.567365	lr=0.382115
Epoch[20] Batch [799]	Speed: 2730.935235 samples/sec	accuracy=0.567170	lr=0.382024
Epoch[20] Batch [849]	Speed: 2641.178619 samples/sec	accuracy=0.567186	lr=0.381934
Epoch[20] Batch [899]	Speed: 2731.956998 samples/sec	accuracy=0.566974	lr=0.381843
Epoch[20] Batch [949]	Speed: 2620.560903 samples/sec	accuracy=0.566593	lr=0.381752
Epoch[20] Batch [999]	Speed: 2631.848395 samples/sec	accuracy=0.566561	lr=0.381661
Epoch[20] Batch [1049]	Speed: 2686.370345 samples/sec	accuracy=0.566386	lr=0.381569
Epoch[20] Batch [1099]	Speed: 2677.553574 samples/sec	accuracy=0.566160	lr=0.381477
Epoch[20] Batch [1149]	Speed: 2684.391202 samples/sec	accuracy=0.566040	lr=0.381385
Epoch[20] Batch [1199]	Speed: 2691.793736 samples/sec	accuracy=0.565704	lr=0.381293
Epoch[20] Batch [1249]	Speed: 2809.263878 samples/sec	accuracy=0.565623	lr=0.381201
[Epoch 20] training: accuracy=0.565619
[Epoch 20] speed: 2712 samples/sec	time cost: 482.443732
[Epoch 20] validation: err-top1=0.410316 err-top5=0.171177
Epoch[21] Batch [49]	Speed: 2785.544875 samples/sec	accuracy=0.574473	lr=0.381107
Epoch[21] Batch [99]	Speed: 2766.705872 samples/sec	accuracy=0.574463	lr=0.381014
Epoch[21] Batch [149]	Speed: 2735.933960 samples/sec	accuracy=0.573887	lr=0.380921
Epoch[21] Batch [199]	Speed: 2796.017722 samples/sec	accuracy=0.572583	lr=0.380828
Epoch[21] Batch [249]	Speed: 2727.094263 samples/sec	accuracy=0.572176	lr=0.380734
Epoch[21] Batch [299]	Speed: 2804.574138 samples/sec	accuracy=0.571611	lr=0.380641
Epoch[21] Batch [349]	Speed: 2762.836701 samples/sec	accuracy=0.571002	lr=0.380547
Epoch[21] Batch [399]	Speed: 2751.972520 samples/sec	accuracy=0.570566	lr=0.380453
Epoch[21] Batch [449]	Speed: 2803.350737 samples/sec	accuracy=0.570501	lr=0.380359
Epoch[21] Batch [499]	Speed: 2674.155565 samples/sec	accuracy=0.570541	lr=0.380264
Epoch[21] Batch [549]	Speed: 2721.786680 samples/sec	accuracy=0.570227	lr=0.380169
Epoch[21] Batch [599]	Speed: 2725.387275 samples/sec	accuracy=0.569993	lr=0.380074
Epoch[21] Batch [649]	Speed: 2720.024037 samples/sec	accuracy=0.569704	lr=0.379979
Epoch[21] Batch [699]	Speed: 2687.644438 samples/sec	accuracy=0.569662	lr=0.379884
Epoch[21] Batch [749]	Speed: 2714.874912 samples/sec	accuracy=0.569527	lr=0.379788
Epoch[21] Batch [799]	Speed: 2710.435967 samples/sec	accuracy=0.569425	lr=0.379693
Epoch[21] Batch [849]	Speed: 2809.516078 samples/sec	accuracy=0.569467	lr=0.379597
Epoch[21] Batch [899]	Speed: 2783.901375 samples/sec	accuracy=0.569227	lr=0.379501
Epoch[21] Batch [949]	Speed: 2690.633085 samples/sec	accuracy=0.568839	lr=0.379404
Epoch[21] Batch [999]	Speed: 2691.436571 samples/sec	accuracy=0.568463	lr=0.379307
Epoch[21] Batch [1049]	Speed: 2863.995988 samples/sec	accuracy=0.568129	lr=0.379211
Epoch[21] Batch [1099]	Speed: 2780.283696 samples/sec	accuracy=0.567952	lr=0.379114
Epoch[21] Batch [1149]	Speed: 2669.943175 samples/sec	accuracy=0.567955	lr=0.379016
Epoch[21] Batch [1199]	Speed: 2670.781482 samples/sec	accuracy=0.567891	lr=0.378919
Epoch[21] Batch [1249]	Speed: 2801.441362 samples/sec	accuracy=0.567991	lr=0.378821
[Epoch 21] training: accuracy=0.567980
[Epoch 21] speed: 2743 samples/sec	time cost: 476.806562
[Epoch 21] validation: err-top1=0.407924 err-top5=0.164561
Epoch[22] Batch [49]	Speed: 2884.259407 samples/sec	accuracy=0.579746	lr=0.378721
Epoch[22] Batch [99]	Speed: 2692.128620 samples/sec	accuracy=0.578389	lr=0.378623
Epoch[22] Batch [149]	Speed: 2727.997022 samples/sec	accuracy=0.577826	lr=0.378525
Epoch[22] Batch [199]	Speed: 2642.548453 samples/sec	accuracy=0.577261	lr=0.378426
Epoch[22] Batch [249]	Speed: 2735.303102 samples/sec	accuracy=0.577328	lr=0.378328
Epoch[22] Batch [299]	Speed: 2712.294927 samples/sec	accuracy=0.576296	lr=0.378229
Epoch[22] Batch [349]	Speed: 2666.315734 samples/sec	accuracy=0.575533	lr=0.378129
Epoch[22] Batch [399]	Speed: 2636.680347 samples/sec	accuracy=0.575247	lr=0.378030
Epoch[22] Batch [449]	Speed: 2827.253494 samples/sec	accuracy=0.574816	lr=0.377930
Epoch[22] Batch [499]	Speed: 2773.683893 samples/sec	accuracy=0.574562	lr=0.377831
Epoch[22] Batch [549]	Speed: 2723.670264 samples/sec	accuracy=0.573983	lr=0.377731
Epoch[22] Batch [599]	Speed: 2675.247782 samples/sec	accuracy=0.573682	lr=0.377630
Epoch[22] Batch [649]	Speed: 2826.050321 samples/sec	accuracy=0.573256	lr=0.377530
Epoch[22] Batch [699]	Speed: 2736.256628 samples/sec	accuracy=0.572871	lr=0.377429
Epoch[22] Batch [749]	Speed: 2673.257900 samples/sec	accuracy=0.572439	lr=0.377328
Epoch[22] Batch [799]	Speed: 2715.742982 samples/sec	accuracy=0.572032	lr=0.377227
Epoch[22] Batch [849]	Speed: 2836.638655 samples/sec	accuracy=0.572019	lr=0.377126
Epoch[22] Batch [899]	Speed: 2667.667599 samples/sec	accuracy=0.571782	lr=0.377024
Epoch[22] Batch [949]	Speed: 2670.997303 samples/sec	accuracy=0.571597	lr=0.376923
Epoch[22] Batch [999]	Speed: 2794.940795 samples/sec	accuracy=0.571512	lr=0.376821
Epoch[22] Batch [1049]	Speed: 2854.840040 samples/sec	accuracy=0.571348	lr=0.376719
Epoch[22] Batch [1099]	Speed: 2811.855365 samples/sec	accuracy=0.571162	lr=0.376616
Epoch[22] Batch [1149]	Speed: 2707.565435 samples/sec	accuracy=0.571072	lr=0.376514
Epoch[22] Batch [1199]	Speed: 2730.832406 samples/sec	accuracy=0.570962	lr=0.376411
Epoch[22] Batch [1249]	Speed: 2846.825677 samples/sec	accuracy=0.570855	lr=0.376308
[Epoch 22] training: accuracy=0.570837
[Epoch 22] speed: 2739 samples/sec	time cost: 477.223110
[Epoch 22] validation: err-top1=0.391083 err-top5=0.155533
Epoch[23] Batch [49]	Speed: 2784.304225 samples/sec	accuracy=0.583633	lr=0.376203
Epoch[23] Batch [99]	Speed: 2661.756117 samples/sec	accuracy=0.582695	lr=0.376099
Epoch[23] Batch [149]	Speed: 2658.992854 samples/sec	accuracy=0.581361	lr=0.375996
Epoch[23] Batch [199]	Speed: 2837.416773 samples/sec	accuracy=0.579102	lr=0.375892
Epoch[23] Batch [249]	Speed: 2914.614458 samples/sec	accuracy=0.578496	lr=0.375788
Epoch[23] Batch [299]	Speed: 2850.654435 samples/sec	accuracy=0.577842	lr=0.375683
Epoch[23] Batch [349]	Speed: 2842.255909 samples/sec	accuracy=0.577076	lr=0.375579
Epoch[23] Batch [399]	Speed: 2736.352021 samples/sec	accuracy=0.575896	lr=0.375474
Epoch[23] Batch [449]	Speed: 2715.653519 samples/sec	accuracy=0.575805	lr=0.375369
Epoch[23] Batch [499]	Speed: 2708.461185 samples/sec	accuracy=0.575150	lr=0.375264
Epoch[23] Batch [549]	Speed: 2717.903408 samples/sec	accuracy=0.574931	lr=0.375159
Epoch[23] Batch [599]	Speed: 2793.414344 samples/sec	accuracy=0.574515	lr=0.375053
Epoch[23] Batch [649]	Speed: 2738.950715 samples/sec	accuracy=0.574740	lr=0.374948
Epoch[23] Batch [699]	Speed: 2786.237948 samples/sec	accuracy=0.574283	lr=0.374842
Epoch[23] Batch [749]	Speed: 2682.096670 samples/sec	accuracy=0.574143	lr=0.374736
Epoch[23] Batch [799]	Speed: 2688.004366 samples/sec	accuracy=0.573812	lr=0.374629
Epoch[23] Batch [849]	Speed: 2755.396005 samples/sec	accuracy=0.573568	lr=0.374523
Epoch[23] Batch [899]	Speed: 2629.398649 samples/sec	accuracy=0.573383	lr=0.374416
Epoch[23] Batch [949]	Speed: 2697.816496 samples/sec	accuracy=0.572910	lr=0.374309
Epoch[23] Batch [999]	Speed: 2733.370000 samples/sec	accuracy=0.572652	lr=0.374202
Epoch[23] Batch [1049]	Speed: 2749.492723 samples/sec	accuracy=0.572528	lr=0.374095
Epoch[23] Batch [1099]	Speed: 2682.306183 samples/sec	accuracy=0.572317	lr=0.373987
Epoch[23] Batch [1149]	Speed: 2670.759759 samples/sec	accuracy=0.572210	lr=0.373879
Epoch[23] Batch [1199]	Speed: 2660.889638 samples/sec	accuracy=0.572171	lr=0.373771
Epoch[23] Batch [1249]	Speed: 2902.817679 samples/sec	accuracy=0.571929	lr=0.373663
[Epoch 23] training: accuracy=0.571935
[Epoch 23] speed: 2739 samples/sec	time cost: 477.353679
[Epoch 23] validation: err-top1=0.389648 err-top5=0.153544
Epoch[24] Batch [49]	Speed: 2671.878790 samples/sec	accuracy=0.579121	lr=0.373552
Epoch[24] Batch [99]	Speed: 2693.326806 samples/sec	accuracy=0.579648	lr=0.373444
Epoch[24] Batch [149]	Speed: 2698.458865 samples/sec	accuracy=0.578340	lr=0.373335
Epoch[24] Batch [199]	Speed: 2774.135216 samples/sec	accuracy=0.578354	lr=0.373226
Epoch[24] Batch [249]	Speed: 2760.030739 samples/sec	accuracy=0.577551	lr=0.373117
Epoch[24] Batch [299]	Speed: 2665.750587 samples/sec	accuracy=0.577588	lr=0.373007
Epoch[24] Batch [349]	Speed: 2766.608208 samples/sec	accuracy=0.577718	lr=0.372897
Epoch[24] Batch [399]	Speed: 2654.501571 samples/sec	accuracy=0.577373	lr=0.372788
Epoch[24] Batch [449]	Speed: 2711.981173 samples/sec	accuracy=0.576871	lr=0.372678
Epoch[24] Batch [499]	Speed: 2704.043444 samples/sec	accuracy=0.576656	lr=0.372567
Epoch[24] Batch [549]	Speed: 2668.247617 samples/sec	accuracy=0.576298	lr=0.372457
Epoch[24] Batch [599]	Speed: 2641.642763 samples/sec	accuracy=0.576203	lr=0.372346
Epoch[24] Batch [649]	Speed: 2731.959640 samples/sec	accuracy=0.575717	lr=0.372235
Epoch[24] Batch [699]	Speed: 2690.994252 samples/sec	accuracy=0.575737	lr=0.372124
Epoch[24] Batch [749]	Speed: 2687.582784 samples/sec	accuracy=0.575878	lr=0.372013
Epoch[24] Batch [799]	Speed: 2792.795710 samples/sec	accuracy=0.575857	lr=0.371901
Epoch[24] Batch [849]	Speed: 2701.988004 samples/sec	accuracy=0.575682	lr=0.371790
Epoch[24] Batch [899]	Speed: 2737.081074 samples/sec	accuracy=0.575660	lr=0.371678
Epoch[24] Batch [949]	Speed: 2661.352952 samples/sec	accuracy=0.575378	lr=0.371565
Epoch[24] Batch [999]	Speed: 2673.646972 samples/sec	accuracy=0.575078	lr=0.371453
Epoch[24] Batch [1049]	Speed: 2722.983795 samples/sec	accuracy=0.574944	lr=0.371341
Epoch[24] Batch [1099]	Speed: 2669.526375 samples/sec	accuracy=0.574738	lr=0.371228
Epoch[24] Batch [1149]	Speed: 2686.869705 samples/sec	accuracy=0.574564	lr=0.371115
Epoch[24] Batch [1199]	Speed: 2763.212891 samples/sec	accuracy=0.574548	lr=0.371002
Epoch[24] Batch [1249]	Speed: 2824.629659 samples/sec	accuracy=0.574466	lr=0.370888
[Epoch 24] training: accuracy=0.574439
[Epoch 24] speed: 2707 samples/sec	time cost: 483.224283
[Epoch 24] validation: err-top1=0.392977 err-top5=0.159220
Epoch[25] Batch [49]	Speed: 2852.888370 samples/sec	accuracy=0.583145	lr=0.370773
Epoch[25] Batch [99]	Speed: 2715.949266 samples/sec	accuracy=0.584424	lr=0.370659
Epoch[25] Batch [149]	Speed: 2776.332656 samples/sec	accuracy=0.582767	lr=0.370545
Epoch[25] Batch [199]	Speed: 2686.286369 samples/sec	accuracy=0.581787	lr=0.370431
Epoch[25] Batch [249]	Speed: 2690.860725 samples/sec	accuracy=0.580992	lr=0.370316
Epoch[25] Batch [299]	Speed: 2736.114388 samples/sec	accuracy=0.580820	lr=0.370202
Epoch[25] Batch [349]	Speed: 2702.957802 samples/sec	accuracy=0.580407	lr=0.370087
Epoch[25] Batch [399]	Speed: 2719.304454 samples/sec	accuracy=0.579895	lr=0.369972
Epoch[25] Batch [449]	Speed: 2711.215042 samples/sec	accuracy=0.579477	lr=0.369857
Epoch[25] Batch [499]	Speed: 2702.353348 samples/sec	accuracy=0.579133	lr=0.369741
Epoch[25] Batch [549]	Speed: 2690.352701 samples/sec	accuracy=0.579123	lr=0.369626
Epoch[25] Batch [599]	Speed: 2703.680468 samples/sec	accuracy=0.578760	lr=0.369510
Epoch[25] Batch [649]	Speed: 2735.083522 samples/sec	accuracy=0.578669	lr=0.369394
Epoch[25] Batch [699]	Speed: 2774.260291 samples/sec	accuracy=0.578322	lr=0.369278
Epoch[25] Batch [749]	Speed: 2794.554353 samples/sec	accuracy=0.578307	lr=0.369161
Epoch[25] Batch [799]	Speed: 2729.929505 samples/sec	accuracy=0.577947	lr=0.369045
Epoch[25] Batch [849]	Speed: 2682.878641 samples/sec	accuracy=0.577747	lr=0.368928
Epoch[25] Batch [899]	Speed: 2681.945101 samples/sec	accuracy=0.577716	lr=0.368811
Epoch[25] Batch [949]	Speed: 2682.084745 samples/sec	accuracy=0.577534	lr=0.368694
Epoch[25] Batch [999]	Speed: 2659.849961 samples/sec	accuracy=0.577529	lr=0.368576
Epoch[25] Batch [1049]	Speed: 2665.383824 samples/sec	accuracy=0.577272	lr=0.368459
Epoch[25] Batch [1099]	Speed: 2718.953688 samples/sec	accuracy=0.577093	lr=0.368341
Epoch[25] Batch [1149]	Speed: 2658.258141 samples/sec	accuracy=0.576904	lr=0.368223
Epoch[25] Batch [1199]	Speed: 2739.153378 samples/sec	accuracy=0.576733	lr=0.368105
Epoch[25] Batch [1249]	Speed: 2927.674982 samples/sec	accuracy=0.576490	lr=0.367986
[Epoch 25] training: accuracy=0.576493
[Epoch 25] speed: 2722 samples/sec	time cost: 480.558742
[Epoch 25] validation: err-top1=0.388154 err-top5=0.155951
Epoch[26] Batch [49]	Speed: 2693.997691 samples/sec	accuracy=0.584805	lr=0.367865
Epoch[26] Batch [99]	Speed: 2688.020752 samples/sec	accuracy=0.583252	lr=0.367747
Epoch[26] Batch [149]	Speed: 2639.664544 samples/sec	accuracy=0.583210	lr=0.367628
Epoch[26] Batch [199]	Speed: 2821.344854 samples/sec	accuracy=0.583208	lr=0.367508
Epoch[26] Batch [249]	Speed: 2659.867652 samples/sec	accuracy=0.582645	lr=0.367389
Epoch[26] Batch [299]	Speed: 2735.400310 samples/sec	accuracy=0.582100	lr=0.367269
Epoch[26] Batch [349]	Speed: 2737.686367 samples/sec	accuracy=0.582034	lr=0.367150
Epoch[26] Batch [399]	Speed: 2725.177376 samples/sec	accuracy=0.581089	lr=0.367030
Epoch[26] Batch [449]	Speed: 2585.597440 samples/sec	accuracy=0.580681	lr=0.366909
Epoch[26] Batch [499]	Speed: 2700.347299 samples/sec	accuracy=0.580494	lr=0.366789
Epoch[26] Batch [549]	Speed: 2692.533940 samples/sec	accuracy=0.580055	lr=0.366668
Epoch[26] Batch [599]	Speed: 2863.235674 samples/sec	accuracy=0.579954	lr=0.366548
Epoch[26] Batch [649]	Speed: 2705.035639 samples/sec	accuracy=0.579543	lr=0.366426
Epoch[26] Batch [699]	Speed: 2718.123851 samples/sec	accuracy=0.579499	lr=0.366305
Epoch[26] Batch [749]	Speed: 2709.144929 samples/sec	accuracy=0.579400	lr=0.366184
Epoch[26] Batch [799]	Speed: 2728.349537 samples/sec	accuracy=0.579264	lr=0.366062
Epoch[26] Batch [849]	Speed: 2669.979258 samples/sec	accuracy=0.579190	lr=0.365940
Epoch[26] Batch [899]	Speed: 2662.013413 samples/sec	accuracy=0.579030	lr=0.365819
Epoch[26] Batch [949]	Speed: 2745.556375 samples/sec	accuracy=0.578754	lr=0.365696
Epoch[26] Batch [999]	Speed: 2815.293804 samples/sec	accuracy=0.578580	lr=0.365574
Epoch[26] Batch [1049]	Speed: 2683.652952 samples/sec	accuracy=0.578364	lr=0.365451
Epoch[26] Batch [1099]	Speed: 2765.921766 samples/sec	accuracy=0.578252	lr=0.365329
Epoch[26] Batch [1149]	Speed: 2764.578516 samples/sec	accuracy=0.577986	lr=0.365206
Epoch[26] Batch [1199]	Speed: 2713.993335 samples/sec	accuracy=0.577771	lr=0.365082
Epoch[26] Batch [1249]	Speed: 2873.931628 samples/sec	accuracy=0.577766	lr=0.364959
[Epoch 26] training: accuracy=0.577775
[Epoch 26] speed: 2720 samples/sec	time cost: 481.139260
[Epoch 26] validation: err-top1=0.382175 err-top5=0.152404
Epoch[27] Batch [49]	Speed: 2758.998719 samples/sec	accuracy=0.594531	lr=0.364833
Epoch[27] Batch [99]	Speed: 2635.336769 samples/sec	accuracy=0.592080	lr=0.364709
Epoch[27] Batch [149]	Speed: 2699.839557 samples/sec	accuracy=0.589199	lr=0.364585
Epoch[27] Batch [199]	Speed: 2709.684590 samples/sec	accuracy=0.587524	lr=0.364461
Epoch[27] Batch [249]	Speed: 2766.473701 samples/sec	accuracy=0.586496	lr=0.364337
Epoch[27] Batch [299]	Speed: 2801.158931 samples/sec	accuracy=0.585413	lr=0.364212
Epoch[27] Batch [349]	Speed: 2733.263335 samples/sec	accuracy=0.584858	lr=0.364087
Epoch[27] Batch [399]	Speed: 2708.811983 samples/sec	accuracy=0.584480	lr=0.363962
Epoch[27] Batch [449]	Speed: 2667.183433 samples/sec	accuracy=0.583893	lr=0.363837
Epoch[27] Batch [499]	Speed: 2668.358982 samples/sec	accuracy=0.583930	lr=0.363712
Epoch[27] Batch [549]	Speed: 2638.347583 samples/sec	accuracy=0.583645	lr=0.363586
Epoch[27] Batch [599]	Speed: 2649.729256 samples/sec	accuracy=0.583218	lr=0.363461
Epoch[27] Batch [649]	Speed: 2760.140993 samples/sec	accuracy=0.582831	lr=0.363335
Epoch[27] Batch [699]	Speed: 2737.279064 samples/sec	accuracy=0.582637	lr=0.363209
Epoch[27] Batch [749]	Speed: 2688.025697 samples/sec	accuracy=0.582266	lr=0.363082
Epoch[27] Batch [799]	Speed: 2645.712213 samples/sec	accuracy=0.581919	lr=0.362956
Epoch[27] Batch [849]	Speed: 2699.168440 samples/sec	accuracy=0.581696	lr=0.362829
Epoch[27] Batch [899]	Speed: 2654.301628 samples/sec	accuracy=0.581400	lr=0.362702
Epoch[27] Batch [949]	Speed: 2805.329558 samples/sec	accuracy=0.581303	lr=0.362575
Epoch[27] Batch [999]	Speed: 2621.859779 samples/sec	accuracy=0.581140	lr=0.362448
Epoch[27] Batch [1049]	Speed: 2743.886645 samples/sec	accuracy=0.581120	lr=0.362320
Epoch[27] Batch [1099]	Speed: 2882.412355 samples/sec	accuracy=0.581104	lr=0.362193
Epoch[27] Batch [1149]	Speed: 2640.757794 samples/sec	accuracy=0.581103	lr=0.362065
Epoch[27] Batch [1199]	Speed: 2696.437025 samples/sec	accuracy=0.580859	lr=0.361937
Epoch[27] Batch [1249]	Speed: 2863.262130 samples/sec	accuracy=0.580803	lr=0.361809
[Epoch 27] training: accuracy=0.580819
[Epoch 27] speed: 2711 samples/sec	time cost: 482.167020
[Epoch 27] validation: err-top1=0.383410 err-top5=0.152683
Epoch[28] Batch [49]	Speed: 2982.555491 samples/sec	accuracy=0.592305	lr=0.361678
Epoch[28] Batch [99]	Speed: 2734.553232 samples/sec	accuracy=0.590625	lr=0.361549
Epoch[28] Batch [149]	Speed: 2660.667568 samples/sec	accuracy=0.589622	lr=0.361420
Epoch[28] Batch [199]	Speed: 2701.908556 samples/sec	accuracy=0.588662	lr=0.361291
Epoch[28] Batch [249]	Speed: 2698.772482 samples/sec	accuracy=0.587523	lr=0.361162
Epoch[28] Batch [299]	Speed: 2835.066032 samples/sec	accuracy=0.587650	lr=0.361032
Epoch[28] Batch [349]	Speed: 2772.877997 samples/sec	accuracy=0.587034	lr=0.360903
Epoch[28] Batch [399]	Speed: 2746.048346 samples/sec	accuracy=0.587097	lr=0.360773
Epoch[28] Batch [449]	Speed: 2680.265246 samples/sec	accuracy=0.585579	lr=0.360643
Epoch[28] Batch [499]	Speed: 2782.999977 samples/sec	accuracy=0.585186	lr=0.360513
Epoch[28] Batch [549]	Speed: 2707.386739 samples/sec	accuracy=0.584895	lr=0.360382
Epoch[28] Batch [599]	Speed: 2793.431568 samples/sec	accuracy=0.584710	lr=0.360252
Epoch[28] Batch [649]	Speed: 2685.137345 samples/sec	accuracy=0.584210	lr=0.360121
Epoch[28] Batch [699]	Speed: 2698.226649 samples/sec	accuracy=0.583938	lr=0.359990
Epoch[28] Batch [749]	Speed: 2665.213827 samples/sec	accuracy=0.583505	lr=0.359859
Epoch[28] Batch [799]	Speed: 2799.854253 samples/sec	accuracy=0.583060	lr=0.359728
Epoch[28] Batch [849]	Speed: 2693.467942 samples/sec	accuracy=0.583026	lr=0.359596
Epoch[28] Batch [899]	Speed: 2652.229562 samples/sec	accuracy=0.582811	lr=0.359465
Epoch[28] Batch [949]	Speed: 2741.142641 samples/sec	accuracy=0.582615	lr=0.359333
Epoch[28] Batch [999]	Speed: 2784.611829 samples/sec	accuracy=0.582540	lr=0.359201
Epoch[28] Batch [1049]	Speed: 2748.312428 samples/sec	accuracy=0.582607	lr=0.359068
Epoch[28] Batch [1099]	Speed: 2693.262492 samples/sec	accuracy=0.582548	lr=0.358936
Epoch[28] Batch [1149]	Speed: 2714.442252 samples/sec	accuracy=0.582354	lr=0.358803
Epoch[28] Batch [1199]	Speed: 2707.148991 samples/sec	accuracy=0.582287	lr=0.358670
Epoch[28] Batch [1249]	Speed: 2981.117752 samples/sec	accuracy=0.582262	lr=0.358537
[Epoch 28] training: accuracy=0.582280
[Epoch 28] speed: 2742 samples/sec	time cost: 477.208948
[Epoch 28] validation: err-top1=0.390705 err-top5=0.155991
Epoch[29] Batch [49]	Speed: 2681.312477 samples/sec	accuracy=0.591484	lr=0.358401
Epoch[29] Batch [99]	Speed: 2697.146519 samples/sec	accuracy=0.592793	lr=0.358268
Epoch[29] Batch [149]	Speed: 2698.988747 samples/sec	accuracy=0.591302	lr=0.358134
Epoch[29] Batch [199]	Speed: 2732.570111 samples/sec	accuracy=0.589800	lr=0.358001
Epoch[29] Batch [249]	Speed: 2718.278127 samples/sec	accuracy=0.589867	lr=0.357867
Epoch[29] Batch [299]	Speed: 2673.883066 samples/sec	accuracy=0.589756	lr=0.357732
Epoch[29] Batch [349]	Speed: 2658.265841 samples/sec	accuracy=0.589752	lr=0.357598
Epoch[29] Batch [399]	Speed: 2806.272288 samples/sec	accuracy=0.589773	lr=0.357464
Epoch[29] Batch [449]	Speed: 2673.038818 samples/sec	accuracy=0.588891	lr=0.357329
Epoch[29] Batch [499]	Speed: 2835.794493 samples/sec	accuracy=0.588436	lr=0.357194
Epoch[29] Batch [549]	Speed: 2651.249534 samples/sec	accuracy=0.587930	lr=0.357059
Epoch[29] Batch [599]	Speed: 2732.724431 samples/sec	accuracy=0.587834	lr=0.356924
Epoch[29] Batch [649]	Speed: 2749.745924 samples/sec	accuracy=0.587434	lr=0.356788
Epoch[29] Batch [699]	Speed: 2701.380381 samples/sec	accuracy=0.586948	lr=0.356652
Epoch[29] Batch [749]	Speed: 2731.430596 samples/sec	accuracy=0.586733	lr=0.356517
Epoch[29] Batch [799]	Speed: 2654.767213 samples/sec	accuracy=0.586342	lr=0.356380
Epoch[29] Batch [849]	Speed: 2782.770978 samples/sec	accuracy=0.585991	lr=0.356244
Epoch[29] Batch [899]	Speed: 2644.085632 samples/sec	accuracy=0.585773	lr=0.356108
Epoch[29] Batch [949]	Speed: 2687.816838 samples/sec	accuracy=0.585348	lr=0.355971
Epoch[29] Batch [999]	Speed: 2661.161935 samples/sec	accuracy=0.584900	lr=0.355834
Epoch[29] Batch [1049]	Speed: 2792.592223 samples/sec	accuracy=0.584569	lr=0.355698
Epoch[29] Batch [1099]	Speed: 2683.436623 samples/sec	accuracy=0.584313	lr=0.355560
Epoch[29] Batch [1149]	Speed: 2733.156365 samples/sec	accuracy=0.584034	lr=0.355423
Epoch[29] Batch [1199]	Speed: 2682.401604 samples/sec	accuracy=0.583951	lr=0.355285
Epoch[29] Batch [1249]	Speed: 2832.972622 samples/sec	accuracy=0.583702	lr=0.355148
[Epoch 29] training: accuracy=0.583706
[Epoch 29] speed: 2713 samples/sec	time cost: 482.145659
[Epoch 29] validation: err-top1=0.377299 err-top5=0.148560
Epoch[30] Batch [49]	Speed: 2732.489932 samples/sec	accuracy=0.589375	lr=0.355007
Epoch[30] Batch [99]	Speed: 2677.591800 samples/sec	accuracy=0.590986	lr=0.354869
Epoch[30] Batch [149]	Speed: 2687.912785 samples/sec	accuracy=0.591068	lr=0.354731
Epoch[30] Batch [199]	Speed: 2785.929985 samples/sec	accuracy=0.590435	lr=0.354592
Epoch[30] Batch [249]	Speed: 2728.482408 samples/sec	accuracy=0.590504	lr=0.354454
Epoch[30] Batch [299]	Speed: 2686.118198 samples/sec	accuracy=0.590550	lr=0.354315
Epoch[30] Batch [349]	Speed: 2779.731239 samples/sec	accuracy=0.589939	lr=0.354176
Epoch[30] Batch [399]	Speed: 2799.515792 samples/sec	accuracy=0.589329	lr=0.354037
Epoch[30] Batch [449]	Speed: 2732.493861 samples/sec	accuracy=0.588854	lr=0.353897
Epoch[30] Batch [499]	Speed: 2830.024817 samples/sec	accuracy=0.588631	lr=0.353758
Epoch[30] Batch [549]	Speed: 2670.453547 samples/sec	accuracy=0.588365	lr=0.353618
Epoch[30] Batch [599]	Speed: 2715.799581 samples/sec	accuracy=0.588092	lr=0.353478
Epoch[30] Batch [649]	Speed: 2753.549786 samples/sec	accuracy=0.587841	lr=0.353338
Epoch[30] Batch [699]	Speed: 2699.747168 samples/sec	accuracy=0.587570	lr=0.353198
Epoch[30] Batch [749]	Speed: 2763.931890 samples/sec	accuracy=0.586905	lr=0.353057
Epoch[30] Batch [799]	Speed: 2665.451246 samples/sec	accuracy=0.586680	lr=0.352917
Epoch[30] Batch [849]	Speed: 2683.930600 samples/sec	accuracy=0.586566	lr=0.352776
Epoch[30] Batch [899]	Speed: 2822.715351 samples/sec	accuracy=0.586720	lr=0.352635
Epoch[30] Batch [949]	Speed: 2731.994500 samples/sec	accuracy=0.586438	lr=0.352493
Epoch[30] Batch [999]	Speed: 2744.035024 samples/sec	accuracy=0.586281	lr=0.352352
Epoch[30] Batch [1049]	Speed: 2815.390875 samples/sec	accuracy=0.585993	lr=0.352211
Epoch[30] Batch [1099]	Speed: 2743.447738 samples/sec	accuracy=0.586126	lr=0.352069
Epoch[30] Batch [1149]	Speed: 2704.427974 samples/sec	accuracy=0.586072	lr=0.351927
Epoch[30] Batch [1199]	Speed: 2702.777026 samples/sec	accuracy=0.585962	lr=0.351785
Epoch[30] Batch [1249]	Speed: 2879.351978 samples/sec	accuracy=0.585670	lr=0.351642
[Epoch 30] training: accuracy=0.585675
[Epoch 30] speed: 2738 samples/sec	time cost: 477.810010
[Epoch 30] validation: err-top1=0.381975 err-top5=0.150211
Epoch[31] Batch [49]	Speed: 2859.355822 samples/sec	accuracy=0.598672	lr=0.351497
Epoch[31] Batch [99]	Speed: 2664.605503 samples/sec	accuracy=0.594111	lr=0.351354
Epoch[31] Batch [149]	Speed: 2783.021761 samples/sec	accuracy=0.592702	lr=0.351212
Epoch[31] Batch [199]	Speed: 2657.409195 samples/sec	accuracy=0.592119	lr=0.351069
Epoch[31] Batch [249]	Speed: 2692.135403 samples/sec	accuracy=0.591637	lr=0.350925
Epoch[31] Batch [299]	Speed: 2786.894005 samples/sec	accuracy=0.591383	lr=0.350782
Epoch[31] Batch [349]	Speed: 2729.227079 samples/sec	accuracy=0.591250	lr=0.350639
Epoch[31] Batch [399]	Speed: 2713.988293 samples/sec	accuracy=0.591067	lr=0.350495
Epoch[31] Batch [449]	Speed: 2706.669255 samples/sec	accuracy=0.590495	lr=0.350351
Epoch[31] Batch [499]	Speed: 2738.529939 samples/sec	accuracy=0.590033	lr=0.350207
Epoch[31] Batch [549]	Speed: 2667.383797 samples/sec	accuracy=0.589721	lr=0.350062
Epoch[31] Batch [599]	Speed: 2720.192691 samples/sec	accuracy=0.589575	lr=0.349918
Epoch[31] Batch [649]	Speed: 2663.854501 samples/sec	accuracy=0.589348	lr=0.349773
Epoch[31] Batch [699]	Speed: 2672.154039 samples/sec	accuracy=0.589185	lr=0.349629
Epoch[31] Batch [749]	Speed: 2790.604670 samples/sec	accuracy=0.589225	lr=0.349484
Epoch[31] Batch [799]	Speed: 2703.112233 samples/sec	accuracy=0.589177	lr=0.349338
Epoch[31] Batch [849]	Speed: 2734.696076 samples/sec	accuracy=0.588840	lr=0.349193
Epoch[31] Batch [899]	Speed: 2711.281654 samples/sec	accuracy=0.588511	lr=0.349048
Epoch[31] Batch [949]	Speed: 2706.976355 samples/sec	accuracy=0.588496	lr=0.348902
Epoch[31] Batch [999]	Speed: 2720.137355 samples/sec	accuracy=0.588209	lr=0.348756
Epoch[31] Batch [1049]	Speed: 2803.364130 samples/sec	accuracy=0.587954	lr=0.348610
Epoch[31] Batch [1099]	Speed: 2701.175455 samples/sec	accuracy=0.587892	lr=0.348464
Epoch[31] Batch [1149]	Speed: 2731.609110 samples/sec	accuracy=0.587876	lr=0.348317
Epoch[31] Batch [1199]	Speed: 2767.284686 samples/sec	accuracy=0.587739	lr=0.348171
Epoch[31] Batch [1249]	Speed: 2849.327714 samples/sec	accuracy=0.587451	lr=0.348024
[Epoch 31] training: accuracy=0.587478
[Epoch 31] speed: 2728 samples/sec	time cost: 479.782927
[Epoch 31] validation: err-top1=0.364876 err-top5=0.139449
Epoch[32] Batch [49]	Speed: 2712.448679 samples/sec	accuracy=0.598437	lr=0.347874
Epoch[32] Batch [99]	Speed: 2677.381654 samples/sec	accuracy=0.596543	lr=0.347727
Epoch[32] Batch [149]	Speed: 2669.165374 samples/sec	accuracy=0.596113	lr=0.347580
Epoch[32] Batch [199]	Speed: 2736.966933 samples/sec	accuracy=0.596021	lr=0.347432
Epoch[32] Batch [249]	Speed: 2650.827719 samples/sec	accuracy=0.595340	lr=0.347285
Epoch[32] Batch [299]	Speed: 2675.129276 samples/sec	accuracy=0.594795	lr=0.347137
Epoch[32] Batch [349]	Speed: 2745.628968 samples/sec	accuracy=0.594121	lr=0.346989
Epoch[32] Batch [399]	Speed: 2776.244290 samples/sec	accuracy=0.593540	lr=0.346841
Epoch[32] Batch [449]	Speed: 2747.362151 samples/sec	accuracy=0.592980	lr=0.346692
Epoch[32] Batch [499]	Speed: 2673.676431 samples/sec	accuracy=0.592777	lr=0.346544
Epoch[32] Batch [549]	Speed: 2761.762624 samples/sec	accuracy=0.592271	lr=0.346395
Epoch[32] Batch [599]	Speed: 2747.826607 samples/sec	accuracy=0.592026	lr=0.346246
Epoch[32] Batch [649]	Speed: 2631.264103 samples/sec	accuracy=0.591653	lr=0.346097
Epoch[32] Batch [699]	Speed: 2719.614464 samples/sec	accuracy=0.591253	lr=0.345948
Epoch[32] Batch [749]	Speed: 2665.767662 samples/sec	accuracy=0.590767	lr=0.345799
Epoch[32] Batch [799]	Speed: 2684.296781 samples/sec	accuracy=0.590463	lr=0.345649
Epoch[32] Batch [849]	Speed: 2626.354315 samples/sec	accuracy=0.590292	lr=0.345499
Epoch[32] Batch [899]	Speed: 2655.013771 samples/sec	accuracy=0.590268	lr=0.345349
Epoch[32] Batch [949]	Speed: 2681.369492 samples/sec	accuracy=0.590287	lr=0.345199
Epoch[32] Batch [999]	Speed: 2766.698386 samples/sec	accuracy=0.589968	lr=0.345049
Epoch[32] Batch [1049]	Speed: 2652.777161 samples/sec	accuracy=0.589793	lr=0.344899
Epoch[32] Batch [1099]	Speed: 2771.740612 samples/sec	accuracy=0.589788	lr=0.344748
Epoch[32] Batch [1149]	Speed: 2609.393185 samples/sec	accuracy=0.589661	lr=0.344597
Epoch[32] Batch [1199]	Speed: 2641.507427 samples/sec	accuracy=0.589383	lr=0.344446
Epoch[32] Batch [1249]	Speed: 2874.877085 samples/sec	accuracy=0.589226	lr=0.344295
[Epoch 32] training: accuracy=0.589242
[Epoch 32] speed: 2699 samples/sec	time cost: 484.535201
[Epoch 32] validation: err-top1=0.373824 err-top5=0.142319
Epoch[33] Batch [49]	Speed: 2890.629988 samples/sec	accuracy=0.603164	lr=0.344141
Epoch[33] Batch [99]	Speed: 2620.142339 samples/sec	accuracy=0.600527	lr=0.343989
Epoch[33] Batch [149]	Speed: 2749.456219 samples/sec	accuracy=0.599069	lr=0.343838
Epoch[33] Batch [199]	Speed: 2680.325294 samples/sec	accuracy=0.597788	lr=0.343686
Epoch[33] Batch [249]	Speed: 2704.352946 samples/sec	accuracy=0.596461	lr=0.343534
Epoch[33] Batch [299]	Speed: 2808.787828 samples/sec	accuracy=0.596309	lr=0.343382
Epoch[33] Batch [349]	Speed: 2696.935122 samples/sec	accuracy=0.595421	lr=0.343229
Epoch[33] Batch [399]	Speed: 2727.337259 samples/sec	accuracy=0.594775	lr=0.343077
Epoch[33] Batch [449]	Speed: 2721.581957 samples/sec	accuracy=0.593937	lr=0.342924
Epoch[33] Batch [499]	Speed: 2763.112808 samples/sec	accuracy=0.593434	lr=0.342771
Epoch[33] Batch [549]	Speed: 2716.998207 samples/sec	accuracy=0.592818	lr=0.342618
Epoch[33] Batch [599]	Speed: 2755.543262 samples/sec	accuracy=0.592562	lr=0.342465
Epoch[33] Batch [649]	Speed: 2665.790561 samples/sec	accuracy=0.592380	lr=0.342312
Epoch[33] Batch [699]	Speed: 2743.434105 samples/sec	accuracy=0.592221	lr=0.342158
Epoch[33] Batch [749]	Speed: 2666.682356 samples/sec	accuracy=0.591900	lr=0.342005
Epoch[33] Batch [799]	Speed: 2703.681353 samples/sec	accuracy=0.591771	lr=0.341851
Epoch[33] Batch [849]	Speed: 2739.661092 samples/sec	accuracy=0.591573	lr=0.341697
Epoch[33] Batch [899]	Speed: 2696.998934 samples/sec	accuracy=0.591392	lr=0.341543
Epoch[33] Batch [949]	Speed: 2772.014315 samples/sec	accuracy=0.591132	lr=0.341388
Epoch[33] Batch [999]	Speed: 2720.679233 samples/sec	accuracy=0.591209	lr=0.341234
Epoch[33] Batch [1049]	Speed: 2684.415832 samples/sec	accuracy=0.591101	lr=0.341079
Epoch[33] Batch [1099]	Speed: 2849.593171 samples/sec	accuracy=0.590790	lr=0.340924
Epoch[33] Batch [1149]	Speed: 2725.273970 samples/sec	accuracy=0.590603	lr=0.340769
Epoch[33] Batch [1199]	Speed: 2671.286061 samples/sec	accuracy=0.590529	lr=0.340614
Epoch[33] Batch [1249]	Speed: 2796.319945 samples/sec	accuracy=0.590495	lr=0.340458
[Epoch 33] training: accuracy=0.590503
[Epoch 33] speed: 2727 samples/sec	time cost: 479.552292
[Epoch 33] validation: err-top1=0.367825 err-top5=0.139150
Epoch[34] Batch [49]	Speed: 2804.960572 samples/sec	accuracy=0.602285	lr=0.340300
Epoch[34] Batch [99]	Speed: 2753.143643 samples/sec	accuracy=0.599619	lr=0.340144
Epoch[34] Batch [149]	Speed: 2706.481842 samples/sec	accuracy=0.598522	lr=0.339988
Epoch[34] Batch [199]	Speed: 2738.040972 samples/sec	accuracy=0.598306	lr=0.339832
Epoch[34] Batch [249]	Speed: 2751.136155 samples/sec	accuracy=0.597492	lr=0.339676
Epoch[34] Batch [299]	Speed: 2699.063986 samples/sec	accuracy=0.597402	lr=0.339520
Epoch[34] Batch [349]	Speed: 2672.075937 samples/sec	accuracy=0.596144	lr=0.339363
Epoch[34] Batch [399]	Speed: 2734.377397 samples/sec	accuracy=0.596091	lr=0.339206
Epoch[34] Batch [449]	Speed: 2637.651965 samples/sec	accuracy=0.595875	lr=0.339049
Epoch[34] Batch [499]	Speed: 2782.675242 samples/sec	accuracy=0.595918	lr=0.338892
Epoch[34] Batch [549]	Speed: 2685.805399 samples/sec	accuracy=0.595721	lr=0.338735
Epoch[34] Batch [599]	Speed: 2663.666462 samples/sec	accuracy=0.595366	lr=0.338578
Epoch[34] Batch [649]	Speed: 2761.115504 samples/sec	accuracy=0.595032	lr=0.338420
Epoch[34] Batch [699]	Speed: 2781.256599 samples/sec	accuracy=0.594646	lr=0.338263
Epoch[34] Batch [749]	Speed: 2688.218202 samples/sec	accuracy=0.594185	lr=0.338105
Epoch[34] Batch [799]	Speed: 2748.305394 samples/sec	accuracy=0.594059	lr=0.337947
Epoch[34] Batch [849]	Speed: 2682.981008 samples/sec	accuracy=0.593860	lr=0.337788
Epoch[34] Batch [899]	Speed: 2686.537540 samples/sec	accuracy=0.593638	lr=0.337630
Epoch[34] Batch [949]	Speed: 2774.334230 samples/sec	accuracy=0.593187	lr=0.337472
Epoch[34] Batch [999]	Speed: 2753.679191 samples/sec	accuracy=0.593034	lr=0.337313
Epoch[34] Batch [1049]	Speed: 2621.929980 samples/sec	accuracy=0.592901	lr=0.337154
Epoch[34] Batch [1099]	Speed: 2654.182608 samples/sec	accuracy=0.592684	lr=0.336995
Epoch[34] Batch [1149]	Speed: 2662.866955 samples/sec	accuracy=0.592695	lr=0.336836
Epoch[34] Batch [1199]	Speed: 2757.042504 samples/sec	accuracy=0.592433	lr=0.336677
Epoch[34] Batch [1249]	Speed: 2836.245654 samples/sec	accuracy=0.592140	lr=0.336517
[Epoch 34] training: accuracy=0.592147
[Epoch 34] speed: 2718 samples/sec	time cost: 480.681514
[Epoch 34] validation: err-top1=0.369405 err-top5=0.141744
Epoch[35] Batch [49]	Speed: 2817.227067 samples/sec	accuracy=0.601973	lr=0.336354
Epoch[35] Batch [99]	Speed: 2735.486060 samples/sec	accuracy=0.603047	lr=0.336194
Epoch[35] Batch [149]	Speed: 2683.254627 samples/sec	accuracy=0.599954	lr=0.336034
Epoch[35] Batch [199]	Speed: 2663.888570 samples/sec	accuracy=0.598828	lr=0.335874
Epoch[35] Batch [249]	Speed: 2704.146138 samples/sec	accuracy=0.598152	lr=0.335714
Epoch[35] Batch [299]	Speed: 2770.213948 samples/sec	accuracy=0.598057	lr=0.335553
Epoch[35] Batch [349]	Speed: 2735.598335 samples/sec	accuracy=0.598005	lr=0.335393
Epoch[35] Batch [399]	Speed: 2720.688953 samples/sec	accuracy=0.597930	lr=0.335232
Epoch[35] Batch [449]	Speed: 2752.285402 samples/sec	accuracy=0.597730	lr=0.335071
Epoch[35] Batch [499]	Speed: 2702.468021 samples/sec	accuracy=0.597510	lr=0.334910
Epoch[35] Batch [549]	Speed: 2694.591312 samples/sec	accuracy=0.597466	lr=0.334749
Epoch[35] Batch [599]	Speed: 2654.810436 samples/sec	accuracy=0.596785	lr=0.334587
Epoch[35] Batch [649]	Speed: 2711.715909 samples/sec	accuracy=0.596136	lr=0.334425
Epoch[35] Batch [699]	Speed: 2691.303303 samples/sec	accuracy=0.595579	lr=0.334264
Epoch[35] Batch [749]	Speed: 2701.642404 samples/sec	accuracy=0.595181	lr=0.334102
Epoch[35] Batch [799]	Speed: 2694.387279 samples/sec	accuracy=0.595026	lr=0.333940
Epoch[35] Batch [849]	Speed: 2749.823316 samples/sec	accuracy=0.594820	lr=0.333777
Epoch[35] Batch [899]	Speed: 2760.709043 samples/sec	accuracy=0.594696	lr=0.333615
Epoch[35] Batch [949]	Speed: 2771.211213 samples/sec	accuracy=0.594747	lr=0.333452
Epoch[35] Batch [999]	Speed: 2723.401776 samples/sec	accuracy=0.594346	lr=0.333290
Epoch[35] Batch [1049]	Speed: 2699.299773 samples/sec	accuracy=0.594164	lr=0.333127
Epoch[35] Batch [1099]	Speed: 2811.579444 samples/sec	accuracy=0.593945	lr=0.332964
Epoch[35] Batch [1149]	Speed: 2737.907727 samples/sec	accuracy=0.593726	lr=0.332801
Epoch[35] Batch [1199]	Speed: 2741.372189 samples/sec	accuracy=0.593424	lr=0.332637
Epoch[35] Batch [1249]	Speed: 2860.026352 samples/sec	accuracy=0.593273	lr=0.332474
[Epoch 35] training: accuracy=0.593248
[Epoch 35] speed: 2729 samples/sec	time cost: 479.854713
[Epoch 35] validation: err-top1=0.397700 err-top5=0.160395
Epoch[36] Batch [49]	Speed: 2987.807784 samples/sec	accuracy=0.603711	lr=0.332307
Epoch[36] Batch [99]	Speed: 2689.412639 samples/sec	accuracy=0.604395	lr=0.332143
Epoch[36] Batch [149]	Speed: 2750.418336 samples/sec	accuracy=0.602812	lr=0.331979
Epoch[36] Batch [199]	Speed: 2757.913668 samples/sec	accuracy=0.602080	lr=0.331815
Epoch[36] Batch [249]	Speed: 2738.076720 samples/sec	accuracy=0.601590	lr=0.331650
Epoch[36] Batch [299]	Speed: 2663.855460 samples/sec	accuracy=0.600527	lr=0.331486
Epoch[36] Batch [349]	Speed: 2803.022918 samples/sec	accuracy=0.600212	lr=0.331321
Epoch[36] Batch [399]	Speed: 2684.014495 samples/sec	accuracy=0.600015	lr=0.331157
Epoch[36] Batch [449]	Speed: 2789.934577 samples/sec	accuracy=0.599412	lr=0.330992
Epoch[36] Batch [499]	Speed: 2672.420599 samples/sec	accuracy=0.599309	lr=0.330827
Epoch[36] Batch [549]	Speed: 2816.182344 samples/sec	accuracy=0.598787	lr=0.330661
Epoch[36] Batch [599]	Speed: 2777.073871 samples/sec	accuracy=0.598301	lr=0.330496
Epoch[36] Batch [649]	Speed: 2670.291171 samples/sec	accuracy=0.598068	lr=0.330330
Epoch[36] Batch [699]	Speed: 2689.244278 samples/sec	accuracy=0.597919	lr=0.330165
Epoch[36] Batch [749]	Speed: 2783.010725 samples/sec	accuracy=0.597514	lr=0.329999
Epoch[36] Batch [799]	Speed: 2768.576812 samples/sec	accuracy=0.597284	lr=0.329833
Epoch[36] Batch [849]	Speed: 2687.374564 samples/sec	accuracy=0.596906	lr=0.329667
Epoch[36] Batch [899]	Speed: 2685.117469 samples/sec	accuracy=0.596628	lr=0.329500
Epoch[36] Batch [949]	Speed: 2742.984202 samples/sec	accuracy=0.596697	lr=0.329334
Epoch[36] Batch [999]	Speed: 2758.282000 samples/sec	accuracy=0.596321	lr=0.329167
Epoch[36] Batch [1049]	Speed: 2775.344687 samples/sec	accuracy=0.596172	lr=0.329000
Epoch[36] Batch [1099]	Speed: 2727.001870 samples/sec	accuracy=0.595908	lr=0.328833
Epoch[36] Batch [1149]	Speed: 2732.906315 samples/sec	accuracy=0.595782	lr=0.328666
Epoch[36] Batch [1199]	Speed: 2767.488567 samples/sec	accuracy=0.595477	lr=0.328499
Epoch[36] Batch [1249]	Speed: 2805.310832 samples/sec	accuracy=0.595323	lr=0.328332
[Epoch 36] training: accuracy=0.595303
[Epoch 36] speed: 2745 samples/sec	time cost: 476.507170
[Epoch 36] validation: err-top1=0.364198 err-top5=0.139509
Epoch[37] Batch [49]	Speed: 2687.451310 samples/sec	accuracy=0.606719	lr=0.328161
Epoch[37] Batch [99]	Speed: 2602.494053 samples/sec	accuracy=0.606357	lr=0.327993
Epoch[37] Batch [149]	Speed: 2657.861396 samples/sec	accuracy=0.604115	lr=0.327825
Epoch[37] Batch [199]	Speed: 2762.981164 samples/sec	accuracy=0.603462	lr=0.327657
Epoch[37] Batch [249]	Speed: 2725.247858 samples/sec	accuracy=0.603457	lr=0.327489
Epoch[37] Batch [299]	Speed: 2690.617241 samples/sec	accuracy=0.603177	lr=0.327321
Epoch[37] Batch [349]	Speed: 2707.519488 samples/sec	accuracy=0.602419	lr=0.327152
Epoch[37] Batch [399]	Speed: 2697.914277 samples/sec	accuracy=0.601489	lr=0.326983
Epoch[37] Batch [449]	Speed: 2738.611975 samples/sec	accuracy=0.600827	lr=0.326815
Epoch[37] Batch [499]	Speed: 2799.377737 samples/sec	accuracy=0.600418	lr=0.326646
Epoch[37] Batch [549]	Speed: 2718.136133 samples/sec	accuracy=0.600146	lr=0.326477
Epoch[37] Batch [599]	Speed: 2674.895194 samples/sec	accuracy=0.599795	lr=0.326307
Epoch[37] Batch [649]	Speed: 2728.851208 samples/sec	accuracy=0.599551	lr=0.326138
Epoch[37] Batch [699]	Speed: 2689.294020 samples/sec	accuracy=0.599210	lr=0.325968
Epoch[37] Batch [749]	Speed: 2737.823430 samples/sec	accuracy=0.598878	lr=0.325799
Epoch[37] Batch [799]	Speed: 2723.922255 samples/sec	accuracy=0.598662	lr=0.325629
Epoch[37] Batch [849]	Speed: 2623.902792 samples/sec	accuracy=0.598416	lr=0.325459
Epoch[37] Batch [899]	Speed: 2816.112952 samples/sec	accuracy=0.598304	lr=0.325289
Epoch[37] Batch [949]	Speed: 2707.374861 samples/sec	accuracy=0.598041	lr=0.325118
Epoch[37] Batch [999]	Speed: 2676.288928 samples/sec	accuracy=0.597719	lr=0.324948
Epoch[37] Batch [1049]	Speed: 2811.461618 samples/sec	accuracy=0.597666	lr=0.324777
Epoch[37] Batch [1099]	Speed: 2845.343415 samples/sec	accuracy=0.597540	lr=0.324607
Epoch[37] Batch [1149]	Speed: 2697.369198 samples/sec	accuracy=0.597565	lr=0.324436
Epoch[37] Batch [1199]	Speed: 2690.176034 samples/sec	accuracy=0.597277	lr=0.324265
Epoch[37] Batch [1249]	Speed: 2792.384481 samples/sec	accuracy=0.597115	lr=0.324094
[Epoch 37] training: accuracy=0.597132
[Epoch 37] speed: 2717 samples/sec	time cost: 481.109212
[Epoch 37] validation: err-top1=0.375777 err-top5=0.146664
Epoch[38] Batch [49]	Speed: 2859.674292 samples/sec	accuracy=0.609531	lr=0.323919
Epoch[38] Batch [99]	Speed: 2698.070946 samples/sec	accuracy=0.609062	lr=0.323747
Epoch[38] Batch [149]	Speed: 2721.272948 samples/sec	accuracy=0.606888	lr=0.323576
Epoch[38] Batch [199]	Speed: 2774.518396 samples/sec	accuracy=0.606187	lr=0.323404
Epoch[38] Batch [249]	Speed: 2806.616860 samples/sec	accuracy=0.605801	lr=0.323232
Epoch[38] Batch [299]	Speed: 2679.007262 samples/sec	accuracy=0.604346	lr=0.323060
Epoch[38] Batch [349]	Speed: 2668.484714 samples/sec	accuracy=0.603535	lr=0.322888
Epoch[38] Batch [399]	Speed: 2676.947813 samples/sec	accuracy=0.602927	lr=0.322715
Epoch[38] Batch [449]	Speed: 2694.879784 samples/sec	accuracy=0.602737	lr=0.322543
Epoch[38] Batch [499]	Speed: 2678.097253 samples/sec	accuracy=0.602686	lr=0.322370
Epoch[38] Batch [549]	Speed: 2651.136548 samples/sec	accuracy=0.602308	lr=0.322197
Epoch[38] Batch [599]	Speed: 2749.422742 samples/sec	accuracy=0.602207	lr=0.322025
Epoch[38] Batch [649]	Speed: 2826.673322 samples/sec	accuracy=0.601624	lr=0.321851
Epoch[38] Batch [699]	Speed: 2748.134995 samples/sec	accuracy=0.601422	lr=0.321678
Epoch[38] Batch [749]	Speed: 2724.236151 samples/sec	accuracy=0.601380	lr=0.321505
Epoch[38] Batch [799]	Speed: 2790.932492 samples/sec	accuracy=0.600918	lr=0.321331
Epoch[38] Batch [849]	Speed: 2747.206314 samples/sec	accuracy=0.600553	lr=0.321158
Epoch[38] Batch [899]	Speed: 2690.982787 samples/sec	accuracy=0.600270	lr=0.320984
Epoch[38] Batch [949]	Speed: 2689.662441 samples/sec	accuracy=0.599926	lr=0.320810
Epoch[38] Batch [999]	Speed: 2690.831020 samples/sec	accuracy=0.599752	lr=0.320636
Epoch[38] Batch [1049]	Speed: 2736.225669 samples/sec	accuracy=0.599515	lr=0.320462
Epoch[38] Batch [1099]	Speed: 2777.196050 samples/sec	accuracy=0.599479	lr=0.320287
Epoch[38] Batch [1149]	Speed: 2712.012648 samples/sec	accuracy=0.599355	lr=0.320113
Epoch[38] Batch [1199]	Speed: 2685.580561 samples/sec	accuracy=0.599121	lr=0.319938
Epoch[38] Batch [1249]	Speed: 2860.358687 samples/sec	accuracy=0.599066	lr=0.319763
[Epoch 38] training: accuracy=0.599075
[Epoch 38] speed: 2730 samples/sec	time cost: 479.131940
[Epoch 38] validation: err-top1=0.353217 err-top5=0.131218
Epoch[39] Batch [49]	Speed: 2792.578823 samples/sec	accuracy=0.610664	lr=0.319585
Epoch[39] Batch [99]	Speed: 2851.886417 samples/sec	accuracy=0.609365	lr=0.319410
Epoch[39] Batch [149]	Speed: 2659.071971 samples/sec	accuracy=0.608359	lr=0.319234
Epoch[39] Batch [199]	Speed: 2721.476520 samples/sec	accuracy=0.606182	lr=0.319059
Epoch[39] Batch [249]	Speed: 2718.999921 samples/sec	accuracy=0.606066	lr=0.318883
Epoch[39] Batch [299]	Speed: 2743.114542 samples/sec	accuracy=0.605840	lr=0.318708
Epoch[39] Batch [349]	Speed: 2694.756150 samples/sec	accuracy=0.604785	lr=0.318532
Epoch[39] Batch [399]	Speed: 2660.431000 samples/sec	accuracy=0.604607	lr=0.318356
Epoch[39] Batch [449]	Speed: 2668.407125 samples/sec	accuracy=0.603869	lr=0.318180
Epoch[39] Batch [499]	Speed: 2625.897679 samples/sec	accuracy=0.603588	lr=0.318004
Epoch[39] Batch [549]	Speed: 2738.544327 samples/sec	accuracy=0.603361	lr=0.317827
Epoch[39] Batch [599]	Speed: 2748.930547 samples/sec	accuracy=0.603223	lr=0.317651
Epoch[39] Batch [649]	Speed: 2652.646711 samples/sec	accuracy=0.602867	lr=0.317474
Epoch[39] Batch [699]	Speed: 2729.613566 samples/sec	accuracy=0.602563	lr=0.317297
Epoch[39] Batch [749]	Speed: 2775.164642 samples/sec	accuracy=0.602272	lr=0.317120
Epoch[39] Batch [799]	Speed: 2634.995366 samples/sec	accuracy=0.602050	lr=0.316943
Epoch[39] Batch [849]	Speed: 2660.023261 samples/sec	accuracy=0.601888	lr=0.316766
Epoch[39] Batch [899]	Speed: 2741.082461 samples/sec	accuracy=0.601678	lr=0.316589
Epoch[39] Batch [949]	Speed: 2714.103235 samples/sec	accuracy=0.601159	lr=0.316411
Epoch[39] Batch [999]	Speed: 2675.522226 samples/sec	accuracy=0.601012	lr=0.316233
Epoch[39] Batch [1049]	Speed: 2656.609758 samples/sec	accuracy=0.600723	lr=0.316056
Epoch[39] Batch [1099]	Speed: 2793.303559 samples/sec	accuracy=0.600610	lr=0.315878
Epoch[39] Batch [1149]	Speed: 2648.298767 samples/sec	accuracy=0.600481	lr=0.315700
Epoch[39] Batch [1199]	Speed: 2638.413514 samples/sec	accuracy=0.600383	lr=0.315522
Epoch[39] Batch [1249]	Speed: 2813.545021 samples/sec	accuracy=0.600175	lr=0.315343
[Epoch 39] training: accuracy=0.600173
[Epoch 39] speed: 2707 samples/sec	time cost: 482.599746
[Epoch 39] validation: err-top1=0.366291 err-top5=0.143953
Epoch[40] Batch [49]	Speed: 2763.482708 samples/sec	accuracy=0.611348	lr=0.315161
Epoch[40] Batch [99]	Speed: 2761.779779 samples/sec	accuracy=0.609688	lr=0.314983
Epoch[40] Batch [149]	Speed: 2808.892754 samples/sec	accuracy=0.609635	lr=0.314804
Epoch[40] Batch [199]	Speed: 2723.862103 samples/sec	accuracy=0.609268	lr=0.314625
Epoch[40] Batch [249]	Speed: 2682.096938 samples/sec	accuracy=0.608027	lr=0.314446
Epoch[40] Batch [299]	Speed: 2697.082734 samples/sec	accuracy=0.607764	lr=0.314267
Epoch[40] Batch [349]	Speed: 2662.195971 samples/sec	accuracy=0.607581	lr=0.314087
Epoch[40] Batch [399]	Speed: 2739.916924 samples/sec	accuracy=0.606787	lr=0.313908
Epoch[40] Batch [449]	Speed: 2659.937563 samples/sec	accuracy=0.606480	lr=0.313728
Epoch[40] Batch [499]	Speed: 2835.568180 samples/sec	accuracy=0.606527	lr=0.313549
Epoch[40] Batch [549]	Speed: 2809.787734 samples/sec	accuracy=0.605945	lr=0.313369
Epoch[40] Batch [599]	Speed: 2769.566755 samples/sec	accuracy=0.605314	lr=0.313189
Epoch[40] Batch [649]	Speed: 2745.354554 samples/sec	accuracy=0.605026	lr=0.313009
Epoch[40] Batch [699]	Speed: 2726.601201 samples/sec	accuracy=0.604854	lr=0.312829
Epoch[40] Batch [749]	Speed: 2801.795496 samples/sec	accuracy=0.604556	lr=0.312648
Epoch[40] Batch [799]	Speed: 2775.191827 samples/sec	accuracy=0.604250	lr=0.312468
Epoch[40] Batch [849]	Speed: 2700.434330 samples/sec	accuracy=0.604025	lr=0.312287
Epoch[40] Batch [899]	Speed: 2745.517377 samples/sec	accuracy=0.603777	lr=0.312106
Epoch[40] Batch [949]	Speed: 2723.788307 samples/sec	accuracy=0.603546	lr=0.311925
Epoch[40] Batch [999]	Speed: 2673.180998 samples/sec	accuracy=0.603406	lr=0.311744
Epoch[40] Batch [1049]	Speed: 2711.019196 samples/sec	accuracy=0.603136	lr=0.311563
Epoch[40] Batch [1099]	Speed: 2702.278503 samples/sec	accuracy=0.603026	lr=0.311382
Epoch[40] Batch [1149]	Speed: 2811.281164 samples/sec	accuracy=0.602833	lr=0.311201
Epoch[40] Batch [1199]	Speed: 2751.018936 samples/sec	accuracy=0.602701	lr=0.311019
Epoch[40] Batch [1249]	Speed: 2840.697909 samples/sec	accuracy=0.602475	lr=0.310837
[Epoch 40] training: accuracy=0.602480
[Epoch 40] speed: 2741 samples/sec	time cost: 476.969338
[Epoch 40] validation: err-top1=0.363566 err-top5=0.138855
Epoch[41] Batch [49]	Speed: 2773.909034 samples/sec	accuracy=0.608828	lr=0.310652
Epoch[41] Batch [99]	Speed: 2820.466242 samples/sec	accuracy=0.610371	lr=0.310470
Epoch[41] Batch [149]	Speed: 2816.625254 samples/sec	accuracy=0.611921	lr=0.310288
Epoch[41] Batch [199]	Speed: 2784.152218 samples/sec	accuracy=0.610142	lr=0.310105
Epoch[41] Batch [249]	Speed: 2643.491371 samples/sec	accuracy=0.610297	lr=0.309923
Epoch[41] Batch [299]	Speed: 2865.131195 samples/sec	accuracy=0.609251	lr=0.309741
Epoch[41] Batch [349]	Speed: 2768.464241 samples/sec	accuracy=0.608878	lr=0.309558
Epoch[41] Batch [399]	Speed: 2728.846423 samples/sec	accuracy=0.608604	lr=0.309375
Epoch[41] Batch [449]	Speed: 2817.844779 samples/sec	accuracy=0.607995	lr=0.309192
Epoch[41] Batch [499]	Speed: 2712.717308 samples/sec	accuracy=0.607441	lr=0.309009
Epoch[41] Batch [549]	Speed: 2690.195378 samples/sec	accuracy=0.606726	lr=0.308826
Epoch[41] Batch [599]	Speed: 2776.464391 samples/sec	accuracy=0.606499	lr=0.308643
Epoch[41] Batch [649]	Speed: 2770.079054 samples/sec	accuracy=0.605919	lr=0.308459
Epoch[41] Batch [699]	Speed: 2707.124659 samples/sec	accuracy=0.605544	lr=0.308276
Epoch[41] Batch [749]	Speed: 2862.314339 samples/sec	accuracy=0.605452	lr=0.308092
Epoch[41] Batch [799]	Speed: 2773.388083 samples/sec	accuracy=0.605045	lr=0.307908
Epoch[41] Batch [849]	Speed: 2719.627861 samples/sec	accuracy=0.604929	lr=0.307725
Epoch[41] Batch [899]	Speed: 2846.793184 samples/sec	accuracy=0.604722	lr=0.307540
Epoch[41] Batch [949]	Speed: 2804.397203 samples/sec	accuracy=0.604404	lr=0.307356
Epoch[41] Batch [999]	Speed: 2787.408321 samples/sec	accuracy=0.604459	lr=0.307172
Epoch[41] Batch [1049]	Speed: 2839.056074 samples/sec	accuracy=0.604330	lr=0.306988
Epoch[41] Batch [1099]	Speed: 2863.030267 samples/sec	accuracy=0.604152	lr=0.306803
Epoch[41] Batch [1149]	Speed: 2867.591621 samples/sec	accuracy=0.604059	lr=0.306618
Epoch[41] Batch [1199]	Speed: 2790.608043 samples/sec	accuracy=0.603604	lr=0.306433
Epoch[41] Batch [1249]	Speed: 2946.969674 samples/sec	accuracy=0.603540	lr=0.306249
[Epoch 41] training: accuracy=0.603545
[Epoch 41] speed: 2787 samples/sec	time cost: 469.654480
[Epoch 41] validation: err-top1=0.359275 err-top5=0.136539
Epoch[42] Batch [49]	Speed: 2660.656887 samples/sec	accuracy=0.610488	lr=0.306060
Epoch[42] Batch [99]	Speed: 2722.733566 samples/sec	accuracy=0.609854	lr=0.305875
Epoch[42] Batch [149]	Speed: 2726.315313 samples/sec	accuracy=0.610417	lr=0.305689
Epoch[42] Batch [199]	Speed: 2753.997971 samples/sec	accuracy=0.609907	lr=0.305504
Epoch[42] Batch [249]	Speed: 2841.673175 samples/sec	accuracy=0.609457	lr=0.305318
Epoch[42] Batch [299]	Speed: 2677.200344 samples/sec	accuracy=0.609342	lr=0.305133
Epoch[42] Batch [349]	Speed: 2778.743510 samples/sec	accuracy=0.608524	lr=0.304947
Epoch[42] Batch [399]	Speed: 2689.702428 samples/sec	accuracy=0.608416	lr=0.304761
Epoch[42] Batch [449]	Speed: 2790.684669 samples/sec	accuracy=0.608633	lr=0.304575
Epoch[42] Batch [499]	Speed: 2754.197638 samples/sec	accuracy=0.608172	lr=0.304388
Epoch[42] Batch [549]	Speed: 2775.785571 samples/sec	accuracy=0.607955	lr=0.304202
Epoch[42] Batch [599]	Speed: 2728.439838 samples/sec	accuracy=0.607152	lr=0.304016
Epoch[42] Batch [649]	Speed: 2713.122099 samples/sec	accuracy=0.606788	lr=0.303829
Epoch[42] Batch [699]	Speed: 2694.948605 samples/sec	accuracy=0.606893	lr=0.303642
Epoch[42] Batch [749]	Speed: 2741.508116 samples/sec	accuracy=0.606634	lr=0.303456
Epoch[42] Batch [799]	Speed: 2573.974367 samples/sec	accuracy=0.606241	lr=0.303269
Epoch[42] Batch [849]	Speed: 2632.524206 samples/sec	accuracy=0.606067	lr=0.303082
Epoch[42] Batch [899]	Speed: 2743.040926 samples/sec	accuracy=0.606081	lr=0.302894
Epoch[42] Batch [949]	Speed: 2726.745673 samples/sec	accuracy=0.605978	lr=0.302707
Epoch[42] Batch [999]	Speed: 2683.443799 samples/sec	accuracy=0.605799	lr=0.302520
Epoch[42] Batch [1049]	Speed: 2681.502480 samples/sec	accuracy=0.605525	lr=0.302332
Epoch[42] Batch [1099]	Speed: 2731.203161 samples/sec	accuracy=0.605325	lr=0.302144
Epoch[42] Batch [1149]	Speed: 2713.087136 samples/sec	accuracy=0.605252	lr=0.301957
Epoch[42] Batch [1199]	Speed: 2830.564280 samples/sec	accuracy=0.605193	lr=0.301769
Epoch[42] Batch [1249]	Speed: 2935.185376 samples/sec	accuracy=0.605027	lr=0.301581
[Epoch 42] training: accuracy=0.605019
[Epoch 42] speed: 2728 samples/sec	time cost: 479.912758
[Epoch 42] validation: err-top1=0.348932 err-top5=0.128687
Epoch[43] Batch [49]	Speed: 2729.769843 samples/sec	accuracy=0.616621	lr=0.301389
Epoch[43] Batch [99]	Speed: 2665.219947 samples/sec	accuracy=0.616162	lr=0.301200
Epoch[43] Batch [149]	Speed: 2723.058652 samples/sec	accuracy=0.614655	lr=0.301012
Epoch[43] Batch [199]	Speed: 2702.134266 samples/sec	accuracy=0.614053	lr=0.300823
Epoch[43] Batch [249]	Speed: 2695.494398 samples/sec	accuracy=0.613070	lr=0.300635
Epoch[43] Batch [299]	Speed: 2667.343079 samples/sec	accuracy=0.612282	lr=0.300446
Epoch[43] Batch [349]	Speed: 2658.645031 samples/sec	accuracy=0.611867	lr=0.300257
Epoch[43] Batch [399]	Speed: 2657.049819 samples/sec	accuracy=0.611255	lr=0.300068
Epoch[43] Batch [449]	Speed: 2679.238989 samples/sec	accuracy=0.611248	lr=0.299879
Epoch[43] Batch [499]	Speed: 2840.990136 samples/sec	accuracy=0.610656	lr=0.299690
Epoch[43] Batch [549]	Speed: 2749.860851 samples/sec	accuracy=0.610034	lr=0.299500
Epoch[43] Batch [599]	Speed: 2799.219920 samples/sec	accuracy=0.609712	lr=0.299311
Epoch[43] Batch [649]	Speed: 2724.575389 samples/sec	accuracy=0.609351	lr=0.299121
Epoch[43] Batch [699]	Speed: 2877.464127 samples/sec	accuracy=0.609036	lr=0.298932
Epoch[43] Batch [749]	Speed: 2714.021392 samples/sec	accuracy=0.608766	lr=0.298742
Epoch[43] Batch [799]	Speed: 2769.483176 samples/sec	accuracy=0.608442	lr=0.298552
Epoch[43] Batch [849]	Speed: 2681.554883 samples/sec	accuracy=0.608271	lr=0.298362
Epoch[43] Batch [899]	Speed: 2748.027737 samples/sec	accuracy=0.608166	lr=0.298171
Epoch[43] Batch [949]	Speed: 2731.845821 samples/sec	accuracy=0.607958	lr=0.297981
Epoch[43] Batch [999]	Speed: 2779.106566 samples/sec	accuracy=0.607867	lr=0.297791
Epoch[43] Batch [1049]	Speed: 2667.229977 samples/sec	accuracy=0.607737	lr=0.297600
Epoch[43] Batch [1099]	Speed: 2729.364441 samples/sec	accuracy=0.607645	lr=0.297409
Epoch[43] Batch [1149]	Speed: 2672.010938 samples/sec	accuracy=0.607536	lr=0.297219
Epoch[43] Batch [1199]	Speed: 2760.032229 samples/sec	accuracy=0.607360	lr=0.297028
Epoch[43] Batch [1249]	Speed: 2842.451386 samples/sec	accuracy=0.607198	lr=0.296837
[Epoch 43] training: accuracy=0.607213
[Epoch 43] speed: 2727 samples/sec	time cost: 479.762302
[Epoch 43] validation: err-top1=0.358279 err-top5=0.133669
Epoch[44] Batch [49]	Speed: 2642.934100 samples/sec	accuracy=0.619961	lr=0.296642
Epoch[44] Batch [99]	Speed: 2692.726651 samples/sec	accuracy=0.617227	lr=0.296451
Epoch[44] Batch [149]	Speed: 2674.179441 samples/sec	accuracy=0.616934	lr=0.296259
Epoch[44] Batch [199]	Speed: 2750.142647 samples/sec	accuracy=0.615303	lr=0.296068
Epoch[44] Batch [249]	Speed: 2684.593321 samples/sec	accuracy=0.615227	lr=0.295876
Epoch[44] Batch [299]	Speed: 2664.128397 samples/sec	accuracy=0.614860	lr=0.295684
Epoch[44] Batch [349]	Speed: 2746.973433 samples/sec	accuracy=0.614439	lr=0.295493
Epoch[44] Batch [399]	Speed: 2683.801764 samples/sec	accuracy=0.614067	lr=0.295301
Epoch[44] Batch [449]	Speed: 2837.339283 samples/sec	accuracy=0.613924	lr=0.295109
Epoch[44] Batch [499]	Speed: 2722.271204 samples/sec	accuracy=0.613141	lr=0.294917
Epoch[44] Batch [549]	Speed: 2669.445374 samples/sec	accuracy=0.612997	lr=0.294724
Epoch[44] Batch [599]	Speed: 2734.878361 samples/sec	accuracy=0.612544	lr=0.294532
Epoch[44] Batch [649]	Speed: 2799.756608 samples/sec	accuracy=0.612054	lr=0.294339
Epoch[44] Batch [699]	Speed: 2675.254515 samples/sec	accuracy=0.611733	lr=0.294147
Epoch[44] Batch [749]	Speed: 2799.979978 samples/sec	accuracy=0.611370	lr=0.293954
Epoch[44] Batch [799]	Speed: 2741.701881 samples/sec	accuracy=0.611016	lr=0.293761
Epoch[44] Batch [849]	Speed: 2641.648904 samples/sec	accuracy=0.610898	lr=0.293568
Epoch[44] Batch [899]	Speed: 2732.869449 samples/sec	accuracy=0.610447	lr=0.293375
Epoch[44] Batch [949]	Speed: 2671.962698 samples/sec	accuracy=0.610323	lr=0.293182
Epoch[44] Batch [999]	Speed: 2694.902645 samples/sec	accuracy=0.609942	lr=0.292989
Epoch[44] Batch [1049]	Speed: 2648.908783 samples/sec	accuracy=0.609791	lr=0.292795
Epoch[44] Batch [1099]	Speed: 2676.975944 samples/sec	accuracy=0.609682	lr=0.292602
Epoch[44] Batch [1149]	Speed: 2727.164705 samples/sec	accuracy=0.609462	lr=0.292408
Epoch[44] Batch [1199]	Speed: 2666.398334 samples/sec	accuracy=0.609334	lr=0.292215
Epoch[44] Batch [1249]	Speed: 2851.498571 samples/sec	accuracy=0.609027	lr=0.292021
[Epoch 44] training: accuracy=0.609014
[Epoch 44] speed: 2710 samples/sec	time cost: 482.512086
[Epoch 44] validation: err-top1=0.353516 err-top5=0.133769
Epoch[45] Batch [49]	Speed: 2825.214158 samples/sec	accuracy=0.621270	lr=0.291823
Epoch[45] Batch [99]	Speed: 2680.816184 samples/sec	accuracy=0.620439	lr=0.291629
Epoch[45] Batch [149]	Speed: 2709.366005 samples/sec	accuracy=0.619876	lr=0.291435
Epoch[45] Batch [199]	Speed: 2696.305394 samples/sec	accuracy=0.618652	lr=0.291240
Epoch[45] Batch [249]	Speed: 2729.117095 samples/sec	accuracy=0.618543	lr=0.291046
Epoch[45] Batch [299]	Speed: 2670.703891 samples/sec	accuracy=0.617962	lr=0.290852
Epoch[45] Batch [349]	Speed: 2618.507668 samples/sec	accuracy=0.617090	lr=0.290657
Epoch[45] Batch [399]	Speed: 2728.152248 samples/sec	accuracy=0.616064	lr=0.290462
Epoch[45] Batch [449]	Speed: 2769.726319 samples/sec	accuracy=0.615356	lr=0.290267
Epoch[45] Batch [499]	Speed: 2779.749769 samples/sec	accuracy=0.615031	lr=0.290073
Epoch[45] Batch [549]	Speed: 2700.699124 samples/sec	accuracy=0.614686	lr=0.289878
Epoch[45] Batch [599]	Speed: 2666.880823 samples/sec	accuracy=0.614225	lr=0.289682
Epoch[45] Batch [649]	Speed: 2781.614259 samples/sec	accuracy=0.613720	lr=0.289487
Epoch[45] Batch [699]	Speed: 2789.834035 samples/sec	accuracy=0.613398	lr=0.289292
Epoch[45] Batch [749]	Speed: 2769.887366 samples/sec	accuracy=0.612911	lr=0.289096
Epoch[45] Batch [799]	Speed: 2685.700432 samples/sec	accuracy=0.612638	lr=0.288901
Epoch[45] Batch [849]	Speed: 2747.152509 samples/sec	accuracy=0.612469	lr=0.288705
Epoch[45] Batch [899]	Speed: 2704.055293 samples/sec	accuracy=0.611978	lr=0.288509
Epoch[45] Batch [949]	Speed: 2786.578051 samples/sec	accuracy=0.611738	lr=0.288314
Epoch[45] Batch [999]	Speed: 2795.485996 samples/sec	accuracy=0.611607	lr=0.288118
Epoch[45] Batch [1049]	Speed: 2639.758318 samples/sec	accuracy=0.611313	lr=0.287921
Epoch[45] Batch [1099]	Speed: 2735.364910 samples/sec	accuracy=0.611056	lr=0.287725
Epoch[45] Batch [1149]	Speed: 2801.274615 samples/sec	accuracy=0.610814	lr=0.287529
Epoch[45] Batch [1199]	Speed: 2719.910763 samples/sec	accuracy=0.610667	lr=0.287333
Epoch[45] Batch [1249]	Speed: 2869.999395 samples/sec	accuracy=0.610518	lr=0.287136
[Epoch 45] training: accuracy=0.610505
[Epoch 45] speed: 2732 samples/sec	time cost: 478.644157
[Epoch 45] validation: err-top1=0.343232 err-top5=0.125299
Epoch[46] Batch [49]	Speed: 2763.630795 samples/sec	accuracy=0.619277	lr=0.286936
Epoch[46] Batch [99]	Speed: 2779.551201 samples/sec	accuracy=0.617891	lr=0.286739
Epoch[46] Batch [149]	Speed: 2770.420942 samples/sec	accuracy=0.617682	lr=0.286542
Epoch[46] Batch [199]	Speed: 2755.492702 samples/sec	accuracy=0.617393	lr=0.286345
Epoch[46] Batch [249]	Speed: 2734.280714 samples/sec	accuracy=0.617039	lr=0.286148
Epoch[46] Batch [299]	Speed: 2708.886917 samples/sec	accuracy=0.616940	lr=0.285951
Epoch[46] Batch [349]	Speed: 2761.293730 samples/sec	accuracy=0.616225	lr=0.285754
Epoch[46] Batch [399]	Speed: 2781.872690 samples/sec	accuracy=0.616084	lr=0.285556
Epoch[46] Batch [449]	Speed: 2715.786736 samples/sec	accuracy=0.615703	lr=0.285359
Epoch[46] Batch [499]	Speed: 2650.261399 samples/sec	accuracy=0.615570	lr=0.285161
Epoch[46] Batch [549]	Speed: 2721.056476 samples/sec	accuracy=0.615426	lr=0.284964
Epoch[46] Batch [599]	Speed: 2742.658018 samples/sec	accuracy=0.615239	lr=0.284766
Epoch[46] Batch [649]	Speed: 2694.589656 samples/sec	accuracy=0.615008	lr=0.284568
Epoch[46] Batch [699]	Speed: 2669.443051 samples/sec	accuracy=0.614718	lr=0.284370
Epoch[46] Batch [749]	Speed: 2771.821644 samples/sec	accuracy=0.614452	lr=0.284172
Epoch[46] Batch [799]	Speed: 2646.485829 samples/sec	accuracy=0.614060	lr=0.283974
Epoch[46] Batch [849]	Speed: 2677.222039 samples/sec	accuracy=0.613812	lr=0.283776
Epoch[46] Batch [899]	Speed: 2672.561881 samples/sec	accuracy=0.613443	lr=0.283577
Epoch[46] Batch [949]	Speed: 2696.207289 samples/sec	accuracy=0.613315	lr=0.283379
Epoch[46] Batch [999]	Speed: 2685.365197 samples/sec	accuracy=0.613051	lr=0.283181
Epoch[46] Batch [1049]	Speed: 2780.646974 samples/sec	accuracy=0.612941	lr=0.282982
Epoch[46] Batch [1099]	Speed: 2672.477336 samples/sec	accuracy=0.612909	lr=0.282783
Epoch[46] Batch [1149]	Speed: 2706.880475 samples/sec	accuracy=0.612778	lr=0.282584
Epoch[46] Batch [1199]	Speed: 2763.159346 samples/sec	accuracy=0.612502	lr=0.282385
Epoch[46] Batch [1249]	Speed: 2883.597446 samples/sec	accuracy=0.612484	lr=0.282186
[Epoch 46] training: accuracy=0.612507
[Epoch 46] speed: 2725 samples/sec	time cost: 480.062795
[Epoch 46] validation: err-top1=0.343079 err-top5=0.124878
Epoch[47] Batch [49]	Speed: 2781.072906 samples/sec	accuracy=0.622227	lr=0.281983
Epoch[47] Batch [99]	Speed: 2689.376500 samples/sec	accuracy=0.623193	lr=0.281784
Epoch[47] Batch [149]	Speed: 2834.387102 samples/sec	accuracy=0.622643	lr=0.281585
Epoch[47] Batch [199]	Speed: 2781.799105 samples/sec	accuracy=0.620957	lr=0.281385
Epoch[47] Batch [249]	Speed: 2801.895184 samples/sec	accuracy=0.620559	lr=0.281186
Epoch[47] Batch [299]	Speed: 2708.947400 samples/sec	accuracy=0.620264	lr=0.280986
Epoch[47] Batch [349]	Speed: 2759.889280 samples/sec	accuracy=0.619911	lr=0.280786
Epoch[47] Batch [399]	Speed: 2662.086209 samples/sec	accuracy=0.619331	lr=0.280587
Epoch[47] Batch [449]	Speed: 2715.479110 samples/sec	accuracy=0.618817	lr=0.280387
Epoch[47] Batch [499]	Speed: 2691.452560 samples/sec	accuracy=0.618682	lr=0.280187
Epoch[47] Batch [549]	Speed: 2703.176031 samples/sec	accuracy=0.618237	lr=0.279987
Epoch[47] Batch [599]	Speed: 2724.007875 samples/sec	accuracy=0.618201	lr=0.279786
Epoch[47] Batch [649]	Speed: 2658.203618 samples/sec	accuracy=0.617713	lr=0.279586
Epoch[47] Batch [699]	Speed: 2704.266412 samples/sec	accuracy=0.617299	lr=0.279386
Epoch[47] Batch [749]	Speed: 2771.482844 samples/sec	accuracy=0.616844	lr=0.279185
Epoch[47] Batch [799]	Speed: 2737.443722 samples/sec	accuracy=0.616641	lr=0.278985
Epoch[47] Batch [849]	Speed: 2669.917415 samples/sec	accuracy=0.616271	lr=0.278784
Epoch[47] Batch [899]	Speed: 2702.027883 samples/sec	accuracy=0.616057	lr=0.278583
Epoch[47] Batch [949]	Speed: 2750.580739 samples/sec	accuracy=0.615960	lr=0.278382
Epoch[47] Batch [999]	Speed: 2687.510403 samples/sec	accuracy=0.615618	lr=0.278181
Epoch[47] Batch [1049]	Speed: 2695.657215 samples/sec	accuracy=0.615569	lr=0.277980
Epoch[47] Batch [1099]	Speed: 2762.817293 samples/sec	accuracy=0.615403	lr=0.277779
Epoch[47] Batch [1149]	Speed: 2668.882680 samples/sec	accuracy=0.615260	lr=0.277578
Epoch[47] Batch [1199]	Speed: 2696.323845 samples/sec	accuracy=0.615043	lr=0.277377
Epoch[47] Batch [1249]	Speed: 2867.800633 samples/sec	accuracy=0.614982	lr=0.277175
[Epoch 47] training: accuracy=0.614974
[Epoch 47] speed: 2726 samples/sec	time cost: 480.268903
[Epoch 47] validation: err-top1=0.351802 err-top5=0.132215
Epoch[48] Batch [49]	Speed: 2911.941458 samples/sec	accuracy=0.625625	lr=0.276970
Epoch[48] Batch [99]	Speed: 2707.586498 samples/sec	accuracy=0.626104	lr=0.276768
Epoch[48] Batch [149]	Speed: 2707.225847 samples/sec	accuracy=0.625169	lr=0.276566
Epoch[48] Batch [199]	Speed: 2677.644583 samples/sec	accuracy=0.623550	lr=0.276365
Epoch[48] Batch [249]	Speed: 2691.987185 samples/sec	accuracy=0.622535	lr=0.276163
Epoch[48] Batch [299]	Speed: 2736.165426 samples/sec	accuracy=0.621598	lr=0.275961
Epoch[48] Batch [349]	Speed: 2745.455671 samples/sec	accuracy=0.621002	lr=0.275759
Epoch[48] Batch [399]	Speed: 2705.065624 samples/sec	accuracy=0.620530	lr=0.275557
Epoch[48] Batch [449]	Speed: 2728.733106 samples/sec	accuracy=0.620326	lr=0.275354
Epoch[48] Batch [499]	Speed: 2668.132481 samples/sec	accuracy=0.619686	lr=0.275152
Epoch[48] Batch [549]	Speed: 2651.456678 samples/sec	accuracy=0.619531	lr=0.274950
Epoch[48] Batch [599]	Speed: 2714.012303 samples/sec	accuracy=0.619567	lr=0.274747
Epoch[48] Batch [649]	Speed: 2883.122039 samples/sec	accuracy=0.619312	lr=0.274545
Epoch[48] Batch [699]	Speed: 2728.074200 samples/sec	accuracy=0.618783	lr=0.274342
Epoch[48] Batch [749]	Speed: 2749.453579 samples/sec	accuracy=0.618598	lr=0.274139
Epoch[48] Batch [799]	Speed: 2722.662041 samples/sec	accuracy=0.618362	lr=0.273936
Epoch[48] Batch [849]	Speed: 2765.327177 samples/sec	accuracy=0.618095	lr=0.273733
Epoch[48] Batch [899]	Speed: 2713.528451 samples/sec	accuracy=0.618062	lr=0.273530
Epoch[48] Batch [949]	Speed: 2700.332087 samples/sec	accuracy=0.617862	lr=0.273327
Epoch[48] Batch [999]	Speed: 2704.695527 samples/sec	accuracy=0.617809	lr=0.273124
Epoch[48] Batch [1049]	Speed: 2828.644270 samples/sec	accuracy=0.617618	lr=0.272921
Epoch[48] Batch [1099]	Speed: 2704.840856 samples/sec	accuracy=0.617479	lr=0.272717
Epoch[48] Batch [1149]	Speed: 2767.247458 samples/sec	accuracy=0.617298	lr=0.272514
Epoch[48] Batch [1199]	Speed: 2652.366129 samples/sec	accuracy=0.617126	lr=0.272310
Epoch[48] Batch [1249]	Speed: 2863.814646 samples/sec	accuracy=0.617041	lr=0.272107
[Epoch 48] training: accuracy=0.617038
[Epoch 48] speed: 2733 samples/sec	time cost: 478.245958
[Epoch 48] validation: err-top1=0.343890 err-top5=0.126156
Epoch[49] Batch [49]	Speed: 2815.215009 samples/sec	accuracy=0.626289	lr=0.271899
Epoch[49] Batch [99]	Speed: 2731.515229 samples/sec	accuracy=0.624492	lr=0.271695
Epoch[49] Batch [149]	Speed: 2730.713404 samples/sec	accuracy=0.624466	lr=0.271491
Epoch[49] Batch [199]	Speed: 2704.237534 samples/sec	accuracy=0.624829	lr=0.271287
Epoch[49] Batch [249]	Speed: 2699.339064 samples/sec	accuracy=0.624551	lr=0.271083
Epoch[49] Batch [299]	Speed: 2703.997343 samples/sec	accuracy=0.623307	lr=0.270879
Epoch[49] Batch [349]	Speed: 2617.138271 samples/sec	accuracy=0.622628	lr=0.270675
Epoch[49] Batch [399]	Speed: 2829.174783 samples/sec	accuracy=0.622427	lr=0.270470
Epoch[49] Batch [449]	Speed: 2749.673712 samples/sec	accuracy=0.621934	lr=0.270266
Epoch[49] Batch [499]	Speed: 2711.171263 samples/sec	accuracy=0.621414	lr=0.270061
Epoch[49] Batch [549]	Speed: 2728.007488 samples/sec	accuracy=0.621202	lr=0.269857
Epoch[49] Batch [599]	Speed: 2727.406155 samples/sec	accuracy=0.621134	lr=0.269652
Epoch[49] Batch [649]	Speed: 2789.776917 samples/sec	accuracy=0.620753	lr=0.269447
Epoch[49] Batch [699]	Speed: 2647.280978 samples/sec	accuracy=0.620587	lr=0.269243
Epoch[49] Batch [749]	Speed: 2730.867515 samples/sec	accuracy=0.620527	lr=0.269038
Epoch[49] Batch [799]	Speed: 2759.058624 samples/sec	accuracy=0.620525	lr=0.268833
Epoch[49] Batch [849]	Speed: 2747.441905 samples/sec	accuracy=0.620056	lr=0.268628
Epoch[49] Batch [899]	Speed: 2732.713060 samples/sec	accuracy=0.619646	lr=0.268422
Epoch[49] Batch [949]	Speed: 2701.660655 samples/sec	accuracy=0.619328	lr=0.268217
Epoch[49] Batch [999]	Speed: 2720.167986 samples/sec	accuracy=0.618991	lr=0.268012
Epoch[49] Batch [1049]	Speed: 2708.874035 samples/sec	accuracy=0.618916	lr=0.267807
Epoch[49] Batch [1099]	Speed: 2619.811189 samples/sec	accuracy=0.618579	lr=0.267601
Epoch[49] Batch [1149]	Speed: 2659.404690 samples/sec	accuracy=0.618438	lr=0.267396
Epoch[49] Batch [1199]	Speed: 2665.128655 samples/sec	accuracy=0.618263	lr=0.267190
Epoch[49] Batch [1249]	Speed: 2934.603296 samples/sec	accuracy=0.617991	lr=0.266984
[Epoch 49] training: accuracy=0.617996
[Epoch 49] speed: 2723 samples/sec	time cost: 480.346302
[Epoch 49] validation: err-top1=0.349410 err-top5=0.129464
Epoch[50] Batch [49]	Speed: 2707.001981 samples/sec	accuracy=0.631875	lr=0.266774
Epoch[50] Batch [99]	Speed: 2731.024248 samples/sec	accuracy=0.629922	lr=0.266568
Epoch[50] Batch [149]	Speed: 2688.630120 samples/sec	accuracy=0.628861	lr=0.266362
Epoch[50] Batch [199]	Speed: 2728.911580 samples/sec	accuracy=0.627690	lr=0.266156
Epoch[50] Batch [249]	Speed: 2750.502389 samples/sec	accuracy=0.627563	lr=0.265950
Epoch[50] Batch [299]	Speed: 2742.414141 samples/sec	accuracy=0.626979	lr=0.265744
Epoch[50] Batch [349]	Speed: 2756.571390 samples/sec	accuracy=0.626624	lr=0.265538
Epoch[50] Batch [399]	Speed: 2669.507726 samples/sec	accuracy=0.626016	lr=0.265331
Epoch[50] Batch [449]	Speed: 2704.484818 samples/sec	accuracy=0.625391	lr=0.265125
Epoch[50] Batch [499]	Speed: 2605.342469 samples/sec	accuracy=0.624563	lr=0.264918
Epoch[50] Batch [549]	Speed: 2716.398556 samples/sec	accuracy=0.624059	lr=0.264712
Epoch[50] Batch [599]	Speed: 2686.629732 samples/sec	accuracy=0.624053	lr=0.264505
Epoch[50] Batch [649]	Speed: 2665.241348 samples/sec	accuracy=0.623567	lr=0.264298
Epoch[50] Batch [699]	Speed: 2703.960334 samples/sec	accuracy=0.623252	lr=0.264092
Epoch[50] Batch [749]	Speed: 2720.755135 samples/sec	accuracy=0.623124	lr=0.263885
Epoch[50] Batch [799]	Speed: 2732.472443 samples/sec	accuracy=0.622786	lr=0.263678
Epoch[50] Batch [849]	Speed: 2698.678776 samples/sec	accuracy=0.622495	lr=0.263471
Epoch[50] Batch [899]	Speed: 2767.419486 samples/sec	accuracy=0.622146	lr=0.263264
Epoch[50] Batch [949]	Speed: 2753.257620 samples/sec	accuracy=0.621892	lr=0.263056
Epoch[50] Batch [999]	Speed: 2757.308744 samples/sec	accuracy=0.621803	lr=0.262849
Epoch[50] Batch [1049]	Speed: 2732.755798 samples/sec	accuracy=0.621457	lr=0.262642
Epoch[50] Batch [1099]	Speed: 2715.556302 samples/sec	accuracy=0.621140	lr=0.262434
Epoch[50] Batch [1149]	Speed: 2764.028120 samples/sec	accuracy=0.620746	lr=0.262227
Epoch[50] Batch [1199]	Speed: 2851.310706 samples/sec	accuracy=0.620637	lr=0.262019
Epoch[50] Batch [1249]	Speed: 2849.460947 samples/sec	accuracy=0.620450	lr=0.261812
[Epoch 50] training: accuracy=0.620462
[Epoch 50] speed: 2725 samples/sec	time cost: 479.862651
[Epoch 50] validation: err-top1=0.339864 err-top5=0.123406
Epoch[51] Batch [49]	Speed: 2955.826862 samples/sec	accuracy=0.630176	lr=0.261600
Epoch[51] Batch [99]	Speed: 2780.063601 samples/sec	accuracy=0.629766	lr=0.261392
Epoch[51] Batch [149]	Speed: 2701.678737 samples/sec	accuracy=0.629167	lr=0.261184
Epoch[51] Batch [199]	Speed: 2734.310655 samples/sec	accuracy=0.628086	lr=0.260976
Epoch[51] Batch [249]	Speed: 2718.369483 samples/sec	accuracy=0.627387	lr=0.260768
Epoch[51] Batch [299]	Speed: 2728.937901 samples/sec	accuracy=0.626364	lr=0.260560
Epoch[51] Batch [349]	Speed: 2884.312053 samples/sec	accuracy=0.625756	lr=0.260352
Epoch[51] Batch [399]	Speed: 2662.518679 samples/sec	accuracy=0.625508	lr=0.260144
Epoch[51] Batch [449]	Speed: 2724.254710 samples/sec	accuracy=0.625063	lr=0.259935
Epoch[51] Batch [499]	Speed: 2674.340226 samples/sec	accuracy=0.624926	lr=0.259727
Epoch[51] Batch [549]	Speed: 2751.854172 samples/sec	accuracy=0.624419	lr=0.259519
Epoch[51] Batch [599]	Speed: 2726.661439 samples/sec	accuracy=0.624287	lr=0.259310
Epoch[51] Batch [649]	Speed: 2662.461373 samples/sec	accuracy=0.624283	lr=0.259102
Epoch[51] Batch [699]	Speed: 2814.455178 samples/sec	accuracy=0.624222	lr=0.258893
Epoch[51] Batch [749]	Speed: 2708.989398 samples/sec	accuracy=0.624001	lr=0.258684
Epoch[51] Batch [799]	Speed: 2680.401905 samples/sec	accuracy=0.623788	lr=0.258475
Epoch[51] Batch [849]	Speed: 2653.820104 samples/sec	accuracy=0.623705	lr=0.258267
Epoch[51] Batch [899]	Speed: 2706.533622 samples/sec	accuracy=0.623421	lr=0.258058
Epoch[51] Batch [949]	Speed: 2733.211710 samples/sec	accuracy=0.623313	lr=0.257849
Epoch[51] Batch [999]	Speed: 2774.742921 samples/sec	accuracy=0.623191	lr=0.257640
Epoch[51] Batch [1049]	Speed: 2750.322594 samples/sec	accuracy=0.623020	lr=0.257430
Epoch[51] Batch [1099]	Speed: 2735.296413 samples/sec	accuracy=0.622748	lr=0.257221
Epoch[51] Batch [1149]	Speed: 2769.499034 samples/sec	accuracy=0.622616	lr=0.257012
Epoch[51] Batch [1199]	Speed: 2776.495873 samples/sec	accuracy=0.622396	lr=0.256803
Epoch[51] Batch [1249]	Speed: 2976.275430 samples/sec	accuracy=0.622296	lr=0.256593
[Epoch 51] training: accuracy=0.622308
[Epoch 51] speed: 2747 samples/sec	time cost: 475.775579
[Epoch 51] validation: err-top1=0.331932 err-top5=0.118044
Epoch[52] Batch [49]	Speed: 2800.449687 samples/sec	accuracy=0.634590	lr=0.256379
Epoch[52] Batch [99]	Speed: 2708.066219 samples/sec	accuracy=0.635635	lr=0.256170
Epoch[52] Batch [149]	Speed: 2716.640405 samples/sec	accuracy=0.633561	lr=0.255960
Epoch[52] Batch [199]	Speed: 2753.838130 samples/sec	accuracy=0.632603	lr=0.255751
Epoch[52] Batch [249]	Speed: 2815.165002 samples/sec	accuracy=0.632492	lr=0.255541
Epoch[52] Batch [299]	Speed: 2709.720593 samples/sec	accuracy=0.631598	lr=0.255331
Epoch[52] Batch [349]	Speed: 2649.768196 samples/sec	accuracy=0.630670	lr=0.255121
Epoch[52] Batch [399]	Speed: 2663.691935 samples/sec	accuracy=0.630237	lr=0.254911
Epoch[52] Batch [449]	Speed: 2711.993537 samples/sec	accuracy=0.629707	lr=0.254701
Epoch[52] Batch [499]	Speed: 2649.070497 samples/sec	accuracy=0.629248	lr=0.254491
Epoch[52] Batch [549]	Speed: 2712.519566 samples/sec	accuracy=0.628754	lr=0.254281
Epoch[52] Batch [599]	Speed: 2720.729386 samples/sec	accuracy=0.628407	lr=0.254071
Epoch[52] Batch [649]	Speed: 2689.683057 samples/sec	accuracy=0.627868	lr=0.253861
Epoch[52] Batch [699]	Speed: 2774.036024 samples/sec	accuracy=0.627673	lr=0.253650
Epoch[52] Batch [749]	Speed: 2742.443665 samples/sec	accuracy=0.627380	lr=0.253440
Epoch[52] Batch [799]	Speed: 2641.727253 samples/sec	accuracy=0.627419	lr=0.253229
Epoch[52] Batch [849]	Speed: 2619.801281 samples/sec	accuracy=0.627048	lr=0.253019
Epoch[52] Batch [899]	Speed: 2814.125384 samples/sec	accuracy=0.626713	lr=0.252808
Epoch[52] Batch [949]	Speed: 2782.518869 samples/sec	accuracy=0.626416	lr=0.252598
Epoch[52] Batch [999]	Speed: 2674.562752 samples/sec	accuracy=0.626043	lr=0.252387
Epoch[52] Batch [1049]	Speed: 2684.641045 samples/sec	accuracy=0.625676	lr=0.252176
Epoch[52] Batch [1099]	Speed: 2889.150669 samples/sec	accuracy=0.625374	lr=0.251965
Epoch[52] Batch [1149]	Speed: 2840.485127 samples/sec	accuracy=0.625172	lr=0.251754
Epoch[52] Batch [1199]	Speed: 2611.569625 samples/sec	accuracy=0.625051	lr=0.251543
Epoch[52] Batch [1249]	Speed: 2886.853591 samples/sec	accuracy=0.624841	lr=0.251332
[Epoch 52] training: accuracy=0.624842
[Epoch 52] speed: 2726 samples/sec	time cost: 479.689363
[Epoch 52] validation: err-top1=0.327271 err-top5=0.117533
Epoch[53] Batch [49]	Speed: 2744.344349 samples/sec	accuracy=0.637422	lr=0.251117
Epoch[53] Batch [99]	Speed: 2671.222895 samples/sec	accuracy=0.635176	lr=0.250906
Epoch[53] Batch [149]	Speed: 2789.438859 samples/sec	accuracy=0.633750	lr=0.250695
Epoch[53] Batch [199]	Speed: 2761.458627 samples/sec	accuracy=0.633311	lr=0.250483
Epoch[53] Batch [249]	Speed: 2786.587633 samples/sec	accuracy=0.631727	lr=0.250272
Epoch[53] Batch [299]	Speed: 2759.005702 samples/sec	accuracy=0.631257	lr=0.250061
Epoch[53] Batch [349]	Speed: 2618.400137 samples/sec	accuracy=0.631256	lr=0.249849
Epoch[53] Batch [399]	Speed: 2737.877568 samples/sec	accuracy=0.630522	lr=0.249638
Epoch[53] Batch [449]	Speed: 2793.707755 samples/sec	accuracy=0.629881	lr=0.249426
Epoch[53] Batch [499]	Speed: 2699.328545 samples/sec	accuracy=0.629229	lr=0.249215
Epoch[53] Batch [549]	Speed: 2756.986119 samples/sec	accuracy=0.628814	lr=0.249003
Epoch[53] Batch [599]	Speed: 2755.163679 samples/sec	accuracy=0.628418	lr=0.248791
Epoch[53] Batch [649]	Speed: 2742.533745 samples/sec	accuracy=0.628344	lr=0.248579
Epoch[53] Batch [699]	Speed: 2641.929823 samples/sec	accuracy=0.628020	lr=0.248367
Epoch[53] Batch [749]	Speed: 2720.612434 samples/sec	accuracy=0.627445	lr=0.248156
Epoch[53] Batch [799]	Speed: 2787.886452 samples/sec	accuracy=0.627333	lr=0.247944
Epoch[53] Batch [849]	Speed: 2726.281636 samples/sec	accuracy=0.626920	lr=0.247732
Epoch[53] Batch [899]	Speed: 2860.031113 samples/sec	accuracy=0.626731	lr=0.247519
Epoch[53] Batch [949]	Speed: 2662.335812 samples/sec	accuracy=0.626552	lr=0.247307
Epoch[53] Batch [999]	Speed: 2695.150897 samples/sec	accuracy=0.626317	lr=0.247095
Epoch[53] Batch [1049]	Speed: 2785.985934 samples/sec	accuracy=0.626175	lr=0.246883
Epoch[53] Batch [1099]	Speed: 2757.813401 samples/sec	accuracy=0.626063	lr=0.246670
Epoch[53] Batch [1149]	Speed: 2696.372393 samples/sec	accuracy=0.625945	lr=0.246458
Epoch[53] Batch [1199]	Speed: 2679.717443 samples/sec	accuracy=0.625877	lr=0.246246
Epoch[53] Batch [1249]	Speed: 2918.602747 samples/sec	accuracy=0.625823	lr=0.246033
[Epoch 53] training: accuracy=0.625831
[Epoch 53] speed: 2738 samples/sec	time cost: 477.783526
[Epoch 53] validation: err-top1=0.338469 err-top5=0.123406
Epoch[54] Batch [49]	Speed: 2720.696019 samples/sec	accuracy=0.638691	lr=0.245816
Epoch[54] Batch [99]	Speed: 2684.281715 samples/sec	accuracy=0.636279	lr=0.245604
Epoch[54] Batch [149]	Speed: 2662.247193 samples/sec	accuracy=0.636549	lr=0.245391
Epoch[54] Batch [199]	Speed: 2742.455117 samples/sec	accuracy=0.634736	lr=0.245178
Epoch[54] Batch [249]	Speed: 2655.369017 samples/sec	accuracy=0.634785	lr=0.244966
Epoch[54] Batch [299]	Speed: 2747.637986 samples/sec	accuracy=0.634076	lr=0.244753
Epoch[54] Batch [349]	Speed: 2721.773192 samples/sec	accuracy=0.633496	lr=0.244540
Epoch[54] Batch [399]	Speed: 2793.376410 samples/sec	accuracy=0.633123	lr=0.244327
Epoch[54] Batch [449]	Speed: 2714.061798 samples/sec	accuracy=0.632587	lr=0.244114
Epoch[54] Batch [499]	Speed: 2694.122877 samples/sec	accuracy=0.632287	lr=0.243901
Epoch[54] Batch [549]	Speed: 2713.634164 samples/sec	accuracy=0.631742	lr=0.243688
Epoch[54] Batch [599]	Speed: 2765.239403 samples/sec	accuracy=0.631364	lr=0.243475
Epoch[54] Batch [649]	Speed: 2702.814717 samples/sec	accuracy=0.630959	lr=0.243262
Epoch[54] Batch [699]	Speed: 2683.283159 samples/sec	accuracy=0.630862	lr=0.243049
Epoch[54] Batch [749]	Speed: 2742.737079 samples/sec	accuracy=0.630617	lr=0.242835
Epoch[54] Batch [799]	Speed: 2685.707754 samples/sec	accuracy=0.630287	lr=0.242622
Epoch[54] Batch [849]	Speed: 2796.542184 samples/sec	accuracy=0.629876	lr=0.242409
Epoch[54] Batch [899]	Speed: 2694.663230 samples/sec	accuracy=0.629669	lr=0.242195
Epoch[54] Batch [949]	Speed: 2800.178884 samples/sec	accuracy=0.629288	lr=0.241982
Epoch[54] Batch [999]	Speed: 2684.720286 samples/sec	accuracy=0.629263	lr=0.241768
Epoch[54] Batch [1049]	Speed: 2675.534127 samples/sec	accuracy=0.629132	lr=0.241555
Epoch[54] Batch [1099]	Speed: 2634.056106 samples/sec	accuracy=0.628933	lr=0.241341
Epoch[54] Batch [1149]	Speed: 2733.226564 samples/sec	accuracy=0.628769	lr=0.241127
Epoch[54] Batch [1199]	Speed: 2741.152368 samples/sec	accuracy=0.628445	lr=0.240914
Epoch[54] Batch [1249]	Speed: 2849.338678 samples/sec	accuracy=0.628273	lr=0.240700
[Epoch 54] training: accuracy=0.628268
[Epoch 54] speed: 2718 samples/sec	time cost: 480.517683
[Epoch 54] validation: err-top1=0.335280 err-top5=0.120536
Epoch[55] Batch [49]	Speed: 2640.148572 samples/sec	accuracy=0.639473	lr=0.240482
Epoch[55] Batch [99]	Speed: 2616.730046 samples/sec	accuracy=0.640000	lr=0.240268
Epoch[55] Batch [149]	Speed: 2796.074368 samples/sec	accuracy=0.638060	lr=0.240054
Epoch[55] Batch [199]	Speed: 2720.745897 samples/sec	accuracy=0.636602	lr=0.239840
Epoch[55] Batch [249]	Speed: 2774.442655 samples/sec	accuracy=0.636695	lr=0.239626
Epoch[55] Batch [299]	Speed: 2686.827651 samples/sec	accuracy=0.635674	lr=0.239412
Epoch[55] Batch [349]	Speed: 2704.815577 samples/sec	accuracy=0.635067	lr=0.239198
Epoch[55] Batch [399]	Speed: 2634.565875 samples/sec	accuracy=0.634373	lr=0.238984
Epoch[55] Batch [449]	Speed: 2776.648266 samples/sec	accuracy=0.634405	lr=0.238769
Epoch[55] Batch [499]	Speed: 2788.126611 samples/sec	accuracy=0.634076	lr=0.238555
Epoch[55] Batch [549]	Speed: 2772.538974 samples/sec	accuracy=0.634263	lr=0.238341
Epoch[55] Batch [599]	Speed: 2651.573980 samples/sec	accuracy=0.633875	lr=0.238126
Epoch[55] Batch [649]	Speed: 2605.271289 samples/sec	accuracy=0.633419	lr=0.237912
Epoch[55] Batch [699]	Speed: 2656.941993 samples/sec	accuracy=0.632800	lr=0.237698
Epoch[55] Batch [749]	Speed: 2780.951465 samples/sec	accuracy=0.632370	lr=0.237483
Epoch[55] Batch [799]	Speed: 2665.107983 samples/sec	accuracy=0.631903	lr=0.237269
Epoch[55] Batch [849]	Speed: 2643.825541 samples/sec	accuracy=0.631743	lr=0.237054
Epoch[55] Batch [899]	Speed: 2797.210901 samples/sec	accuracy=0.631406	lr=0.236839
Epoch[55] Batch [949]	Speed: 2721.961348 samples/sec	accuracy=0.630944	lr=0.236625
Epoch[55] Batch [999]	Speed: 2752.643340 samples/sec	accuracy=0.630763	lr=0.236410
Epoch[55] Batch [1049]	Speed: 2885.814535 samples/sec	accuracy=0.630622	lr=0.236195
Epoch[55] Batch [1099]	Speed: 2683.538026 samples/sec	accuracy=0.630639	lr=0.235981
Epoch[55] Batch [1149]	Speed: 2727.413983 samples/sec	accuracy=0.630621	lr=0.235766
Epoch[55] Batch [1199]	Speed: 2688.275208 samples/sec	accuracy=0.630396	lr=0.235551
Epoch[55] Batch [1249]	Speed: 3078.348674 samples/sec	accuracy=0.630044	lr=0.235336
[Epoch 55] training: accuracy=0.630080
[Epoch 55] speed: 2724 samples/sec	time cost: 479.902702
[Epoch 55] validation: err-top1=0.332111 err-top5=0.116908
Epoch[56] Batch [49]	Speed: 2716.894534 samples/sec	accuracy=0.641660	lr=0.235117
Epoch[56] Batch [99]	Speed: 2793.868113 samples/sec	accuracy=0.640811	lr=0.234902
Epoch[56] Batch [149]	Speed: 2706.106890 samples/sec	accuracy=0.638444	lr=0.234687
Epoch[56] Batch [199]	Speed: 2660.818819 samples/sec	accuracy=0.638589	lr=0.234472
Epoch[56] Batch [249]	Speed: 2689.351611 samples/sec	accuracy=0.638004	lr=0.234256
Epoch[56] Batch [299]	Speed: 2743.180243 samples/sec	accuracy=0.637682	lr=0.234041
Epoch[56] Batch [349]	Speed: 2710.606000 samples/sec	accuracy=0.636705	lr=0.233826
Epoch[56] Batch [399]	Speed: 2632.415392 samples/sec	accuracy=0.636050	lr=0.233611
Epoch[56] Batch [449]	Speed: 2650.943002 samples/sec	accuracy=0.635888	lr=0.233396
Epoch[56] Batch [499]	Speed: 2733.140607 samples/sec	accuracy=0.635367	lr=0.233180
Epoch[56] Batch [549]	Speed: 2719.679147 samples/sec	accuracy=0.635094	lr=0.232965
Epoch[56] Batch [599]	Speed: 2825.951435 samples/sec	accuracy=0.634707	lr=0.232749
Epoch[56] Batch [649]	Speed: 2791.716981 samples/sec	accuracy=0.634581	lr=0.232534
Epoch[56] Batch [699]	Speed: 2759.623143 samples/sec	accuracy=0.634261	lr=0.232319
Epoch[56] Batch [749]	Speed: 2765.916850 samples/sec	accuracy=0.633958	lr=0.232103
Epoch[56] Batch [799]	Speed: 2697.767421 samples/sec	accuracy=0.633617	lr=0.231887
Epoch[56] Batch [849]	Speed: 2675.331670 samples/sec	accuracy=0.633282	lr=0.231672
Epoch[56] Batch [899]	Speed: 2729.421471 samples/sec	accuracy=0.632880	lr=0.231456
Epoch[56] Batch [949]	Speed: 2692.576679 samples/sec	accuracy=0.632589	lr=0.231241
Epoch[56] Batch [999]	Speed: 2723.771241 samples/sec	accuracy=0.632468	lr=0.231025
Epoch[56] Batch [1049]	Speed: 2805.153114 samples/sec	accuracy=0.632338	lr=0.230809
Epoch[56] Batch [1099]	Speed: 2680.288395 samples/sec	accuracy=0.632108	lr=0.230593
Epoch[56] Batch [1149]	Speed: 2721.772329 samples/sec	accuracy=0.631815	lr=0.230377
Epoch[56] Batch [1199]	Speed: 2650.298032 samples/sec	accuracy=0.631807	lr=0.230162
Epoch[56] Batch [1249]	Speed: 2914.771234 samples/sec	accuracy=0.631687	lr=0.229946
[Epoch 56] training: accuracy=0.631687
[Epoch 56] speed: 2724 samples/sec	time cost: 479.868569
[Epoch 56] validation: err-top1=0.327467 err-top5=0.116370
Epoch[57] Batch [49]	Speed: 2824.823833 samples/sec	accuracy=0.640469	lr=0.229725
Epoch[57] Batch [99]	Speed: 2711.579428 samples/sec	accuracy=0.640059	lr=0.229510
Epoch[57] Batch [149]	Speed: 2829.748301 samples/sec	accuracy=0.639453	lr=0.229294
Epoch[57] Batch [199]	Speed: 2662.529341 samples/sec	accuracy=0.638872	lr=0.229077
Epoch[57] Batch [249]	Speed: 2776.438366 samples/sec	accuracy=0.639102	lr=0.228861
Epoch[57] Batch [299]	Speed: 2682.467947 samples/sec	accuracy=0.638483	lr=0.228645
Epoch[57] Batch [349]	Speed: 2769.984797 samples/sec	accuracy=0.639222	lr=0.228429
Epoch[57] Batch [399]	Speed: 2752.700465 samples/sec	accuracy=0.638628	lr=0.228213
Epoch[57] Batch [449]	Speed: 2713.562396 samples/sec	accuracy=0.638084	lr=0.227997
Epoch[57] Batch [499]	Speed: 2699.731929 samples/sec	accuracy=0.637385	lr=0.227781
Epoch[57] Batch [549]	Speed: 2808.129539 samples/sec	accuracy=0.636942	lr=0.227564
Epoch[57] Batch [599]	Speed: 2775.282386 samples/sec	accuracy=0.636419	lr=0.227348
Epoch[57] Batch [649]	Speed: 2765.381731 samples/sec	accuracy=0.636367	lr=0.227132
Epoch[57] Batch [699]	Speed: 2729.180844 samples/sec	accuracy=0.636282	lr=0.226915
Epoch[57] Batch [749]	Speed: 2895.881080 samples/sec	accuracy=0.636132	lr=0.226699
Epoch[57] Batch [799]	Speed: 2705.958322 samples/sec	accuracy=0.635782	lr=0.226482
Epoch[57] Batch [849]	Speed: 2706.763414 samples/sec	accuracy=0.635584	lr=0.226266
Epoch[57] Batch [899]	Speed: 2718.348115 samples/sec	accuracy=0.635507	lr=0.226049
Epoch[57] Batch [949]	Speed: 2742.071918 samples/sec	accuracy=0.635516	lr=0.225833
Epoch[57] Batch [999]	Speed: 2739.379797 samples/sec	accuracy=0.635312	lr=0.225616
Epoch[57] Batch [1049]	Speed: 2697.747799 samples/sec	accuracy=0.635193	lr=0.225400
Epoch[57] Batch [1099]	Speed: 2768.331016 samples/sec	accuracy=0.634815	lr=0.225183
Epoch[57] Batch [1149]	Speed: 2740.569568 samples/sec	accuracy=0.634795	lr=0.224967
Epoch[57] Batch [1199]	Speed: 2688.974958 samples/sec	accuracy=0.634807	lr=0.224750
Epoch[57] Batch [1249]	Speed: 2903.404685 samples/sec	accuracy=0.634646	lr=0.224533
[Epoch 57] training: accuracy=0.634677
[Epoch 57] speed: 2749 samples/sec	time cost: 475.788633
[Epoch 57] validation: err-top1=0.332490 err-top5=0.120097
Epoch[58] Batch [49]	Speed: 2812.403834 samples/sec	accuracy=0.643125	lr=0.224312
Epoch[58] Batch [99]	Speed: 2635.609328 samples/sec	accuracy=0.642578	lr=0.224095
Epoch[58] Batch [149]	Speed: 2673.903608 samples/sec	accuracy=0.643704	lr=0.223879
Epoch[58] Batch [199]	Speed: 2591.084459 samples/sec	accuracy=0.642544	lr=0.223662
Epoch[58] Batch [249]	Speed: 2742.543132 samples/sec	accuracy=0.642586	lr=0.223445
Epoch[58] Batch [299]	Speed: 2761.250414 samples/sec	accuracy=0.641846	lr=0.223228
Epoch[58] Batch [349]	Speed: 2664.622828 samples/sec	accuracy=0.641688	lr=0.223011
Epoch[58] Batch [399]	Speed: 2626.225103 samples/sec	accuracy=0.641309	lr=0.222794
Epoch[58] Batch [449]	Speed: 2733.275024 samples/sec	accuracy=0.640818	lr=0.222577
Epoch[58] Batch [499]	Speed: 2753.850278 samples/sec	accuracy=0.640559	lr=0.222360
Epoch[58] Batch [549]	Speed: 2665.252132 samples/sec	accuracy=0.640085	lr=0.222143
Epoch[58] Batch [599]	Speed: 2705.929716 samples/sec	accuracy=0.639787	lr=0.221926
Epoch[58] Batch [649]	Speed: 2747.781884 samples/sec	accuracy=0.639204	lr=0.221709
Epoch[58] Batch [699]	Speed: 2740.601081 samples/sec	accuracy=0.638864	lr=0.221492
Epoch[58] Batch [749]	Speed: 2784.009322 samples/sec	accuracy=0.638858	lr=0.221275
Epoch[58] Batch [799]	Speed: 2661.582361 samples/sec	accuracy=0.638575	lr=0.221058
Epoch[58] Batch [849]	Speed: 2743.146884 samples/sec	accuracy=0.638296	lr=0.220841
Epoch[58] Batch [899]	Speed: 2793.729234 samples/sec	accuracy=0.637977	lr=0.220623
Epoch[58] Batch [949]	Speed: 2693.859777 samples/sec	accuracy=0.637563	lr=0.220406
Epoch[58] Batch [999]	Speed: 2697.170537 samples/sec	accuracy=0.637456	lr=0.220189
Epoch[58] Batch [1049]	Speed: 2778.048949 samples/sec	accuracy=0.637158	lr=0.219972
Epoch[58] Batch [1099]	Speed: 2773.335791 samples/sec	accuracy=0.637135	lr=0.219754
Epoch[58] Batch [1149]	Speed: 2699.611787 samples/sec	accuracy=0.636731	lr=0.219537
Epoch[58] Batch [1199]	Speed: 2805.885163 samples/sec	accuracy=0.636526	lr=0.219320
Epoch[58] Batch [1249]	Speed: 2899.877556 samples/sec	accuracy=0.636390	lr=0.219102
[Epoch 58] training: accuracy=0.636388
[Epoch 58] speed: 2723 samples/sec	time cost: 480.053402
[Epoch 58] validation: err-top1=0.324280 err-top5=0.114014
Epoch[59] Batch [49]	Speed: 2754.602608 samples/sec	accuracy=0.648945	lr=0.218881
Epoch[59] Batch [99]	Speed: 2747.527533 samples/sec	accuracy=0.647734	lr=0.218663
Epoch[59] Batch [149]	Speed: 2759.068585 samples/sec	accuracy=0.646536	lr=0.218446
Epoch[59] Batch [199]	Speed: 2766.748004 samples/sec	accuracy=0.645249	lr=0.218228
Epoch[59] Batch [249]	Speed: 2705.364386 samples/sec	accuracy=0.645047	lr=0.218011
Epoch[59] Batch [299]	Speed: 2732.892333 samples/sec	accuracy=0.644251	lr=0.217793
Epoch[59] Batch [349]	Speed: 2754.269064 samples/sec	accuracy=0.643114	lr=0.217576
Epoch[59] Batch [399]	Speed: 2663.655196 samples/sec	accuracy=0.642249	lr=0.217358
Epoch[59] Batch [449]	Speed: 2699.783451 samples/sec	accuracy=0.642174	lr=0.217141
Epoch[59] Batch [499]	Speed: 2776.501616 samples/sec	accuracy=0.641680	lr=0.216923
Epoch[59] Batch [549]	Speed: 2696.914936 samples/sec	accuracy=0.641438	lr=0.216706
Epoch[59] Batch [599]	Speed: 2684.526437 samples/sec	accuracy=0.640981	lr=0.216488
Epoch[59] Batch [649]	Speed: 2704.800144 samples/sec	accuracy=0.640536	lr=0.216270
Epoch[59] Batch [699]	Speed: 2676.211117 samples/sec	accuracy=0.640036	lr=0.216053
Epoch[59] Batch [749]	Speed: 2803.402337 samples/sec	accuracy=0.640078	lr=0.215835
Epoch[59] Batch [799]	Speed: 2764.976541 samples/sec	accuracy=0.639857	lr=0.215617
Epoch[59] Batch [849]	Speed: 2785.600627 samples/sec	accuracy=0.639816	lr=0.215400
Epoch[59] Batch [899]	Speed: 2649.041577 samples/sec	accuracy=0.639651	lr=0.215182
Epoch[59] Batch [949]	Speed: 2660.868174 samples/sec	accuracy=0.639441	lr=0.214964
Epoch[59] Batch [999]	Speed: 2705.648283 samples/sec	accuracy=0.639178	lr=0.214746
Epoch[59] Batch [1049]	Speed: 2725.919621 samples/sec	accuracy=0.638924	lr=0.214529
Epoch[59] Batch [1099]	Speed: 2668.772763 samples/sec	accuracy=0.638856	lr=0.214311
Epoch[59] Batch [1149]	Speed: 2685.757230 samples/sec	accuracy=0.638761	lr=0.214093
Epoch[59] Batch [1199]	Speed: 2807.800308 samples/sec	accuracy=0.638587	lr=0.213875
Epoch[59] Batch [1249]	Speed: 2892.251087 samples/sec	accuracy=0.638394	lr=0.213657
[Epoch 59] training: accuracy=0.638381
[Epoch 59] speed: 2727 samples/sec	time cost: 479.590546
[Epoch 59] validation: err-top1=0.323920 err-top5=0.114517
Epoch[60] Batch [49]	Speed: 2831.730935 samples/sec	accuracy=0.653340	lr=0.213435
Epoch[60] Batch [99]	Speed: 2735.371809 samples/sec	accuracy=0.651934	lr=0.213217
Epoch[60] Batch [149]	Speed: 2712.656930 samples/sec	accuracy=0.650488	lr=0.212999
Epoch[60] Batch [199]	Speed: 2782.854928 samples/sec	accuracy=0.649414	lr=0.212781
Epoch[60] Batch [249]	Speed: 2630.341484 samples/sec	accuracy=0.648551	lr=0.212563
Epoch[60] Batch [299]	Speed: 2822.006795 samples/sec	accuracy=0.647402	lr=0.212345
Epoch[60] Batch [349]	Speed: 2698.804092 samples/sec	accuracy=0.646950	lr=0.212127
Epoch[60] Batch [399]	Speed: 2680.364034 samples/sec	accuracy=0.646765	lr=0.211909
Epoch[60] Batch [449]	Speed: 2743.638308 samples/sec	accuracy=0.646374	lr=0.211691
Epoch[60] Batch [499]	Speed: 2725.968998 samples/sec	accuracy=0.646051	lr=0.211473
Epoch[60] Batch [549]	Speed: 2741.757538 samples/sec	accuracy=0.645856	lr=0.211255
Epoch[60] Batch [599]	Speed: 2767.937841 samples/sec	accuracy=0.645698	lr=0.211037
Epoch[60] Batch [649]	Speed: 2698.953605 samples/sec	accuracy=0.645329	lr=0.210819
Epoch[60] Batch [699]	Speed: 2791.123440 samples/sec	accuracy=0.645107	lr=0.210601
Epoch[60] Batch [749]	Speed: 2779.151846 samples/sec	accuracy=0.644549	lr=0.210383
Epoch[60] Batch [799]	Speed: 2699.532920 samples/sec	accuracy=0.644008	lr=0.210165
Epoch[60] Batch [849]	Speed: 2708.299585 samples/sec	accuracy=0.643579	lr=0.209947
Epoch[60] Batch [899]	Speed: 2674.145508 samples/sec	accuracy=0.643215	lr=0.209729
Epoch[60] Batch [949]	Speed: 2640.485306 samples/sec	accuracy=0.642944	lr=0.209511
Epoch[60] Batch [999]	Speed: 2626.562727 samples/sec	accuracy=0.642783	lr=0.209293
Epoch[60] Batch [1049]	Speed: 2619.002461 samples/sec	accuracy=0.642565	lr=0.209075
Epoch[60] Batch [1099]	Speed: 2684.343084 samples/sec	accuracy=0.642400	lr=0.208856
Epoch[60] Batch [1149]	Speed: 2622.608328 samples/sec	accuracy=0.642097	lr=0.208638
Epoch[60] Batch [1199]	Speed: 2754.682782 samples/sec	accuracy=0.641847	lr=0.208420
Epoch[60] Batch [1249]	Speed: 2866.935528 samples/sec	accuracy=0.641609	lr=0.208202
[Epoch 60] training: accuracy=0.641633
[Epoch 60] speed: 2718 samples/sec	time cost: 480.860723
[Epoch 60] validation: err-top1=0.321409 err-top5=0.112743
Epoch[61] Batch [49]	Speed: 2814.204709 samples/sec	accuracy=0.653613	lr=0.207979
Epoch[61] Batch [99]	Speed: 2756.542446 samples/sec	accuracy=0.651543	lr=0.207761
Epoch[61] Batch [149]	Speed: 2830.608119 samples/sec	accuracy=0.650996	lr=0.207543
Epoch[61] Batch [199]	Speed: 2676.473450 samples/sec	accuracy=0.650034	lr=0.207325
Epoch[61] Batch [249]	Speed: 2761.904488 samples/sec	accuracy=0.649750	lr=0.207106
Epoch[61] Batch [299]	Speed: 2688.359510 samples/sec	accuracy=0.648525	lr=0.206888
Epoch[61] Batch [349]	Speed: 2703.217817 samples/sec	accuracy=0.647877	lr=0.206670
Epoch[61] Batch [399]	Speed: 2733.696344 samples/sec	accuracy=0.647551	lr=0.206452
Epoch[61] Batch [449]	Speed: 2814.572922 samples/sec	accuracy=0.646955	lr=0.206233
Epoch[61] Batch [499]	Speed: 2715.375691 samples/sec	accuracy=0.646859	lr=0.206015
Epoch[61] Batch [549]	Speed: 2720.076199 samples/sec	accuracy=0.646335	lr=0.205797
Epoch[61] Batch [599]	Speed: 2898.077878 samples/sec	accuracy=0.645975	lr=0.205579
Epoch[61] Batch [649]	Speed: 2818.216090 samples/sec	accuracy=0.645509	lr=0.205360
Epoch[61] Batch [699]	Speed: 2698.447607 samples/sec	accuracy=0.644957	lr=0.205142
Epoch[61] Batch [749]	Speed: 2774.764862 samples/sec	accuracy=0.644828	lr=0.204924
Epoch[61] Batch [799]	Speed: 2733.082274 samples/sec	accuracy=0.644628	lr=0.204705
Epoch[61] Batch [849]	Speed: 2751.311614 samples/sec	accuracy=0.644509	lr=0.204487
Epoch[61] Batch [899]	Speed: 2700.954729 samples/sec	accuracy=0.644337	lr=0.204269
Epoch[61] Batch [949]	Speed: 2677.141805 samples/sec	accuracy=0.644054	lr=0.204051
Epoch[61] Batch [999]	Speed: 2665.363975 samples/sec	accuracy=0.643954	lr=0.203832
Epoch[61] Batch [1049]	Speed: 2723.452616 samples/sec	accuracy=0.643611	lr=0.203614
Epoch[61] Batch [1099]	Speed: 2721.553053 samples/sec	accuracy=0.643300	lr=0.203396
Epoch[61] Batch [1149]	Speed: 2599.992569 samples/sec	accuracy=0.643030	lr=0.203177
Epoch[61] Batch [1199]	Speed: 2698.612782 samples/sec	accuracy=0.643052	lr=0.202959
Epoch[61] Batch [1249]	Speed: 2894.017988 samples/sec	accuracy=0.642920	lr=0.202740
[Epoch 61] training: accuracy=0.642908
[Epoch 61] speed: 2739 samples/sec	time cost: 477.376658
[Epoch 61] validation: err-top1=0.314593 err-top5=0.110232
Epoch[62] Batch [49]	Speed: 2726.558793 samples/sec	accuracy=0.656367	lr=0.202518
Epoch[62] Batch [99]	Speed: 2693.482807 samples/sec	accuracy=0.653008	lr=0.202299
Epoch[62] Batch [149]	Speed: 2835.996648 samples/sec	accuracy=0.651973	lr=0.202081
Epoch[62] Batch [199]	Speed: 2692.015363 samples/sec	accuracy=0.651001	lr=0.201863
Epoch[62] Batch [249]	Speed: 2667.189429 samples/sec	accuracy=0.650598	lr=0.201644
Epoch[62] Batch [299]	Speed: 2747.442784 samples/sec	accuracy=0.649746	lr=0.201426
Epoch[62] Batch [349]	Speed: 2717.896632 samples/sec	accuracy=0.649428	lr=0.201208
Epoch[62] Batch [399]	Speed: 2668.929847 samples/sec	accuracy=0.648694	lr=0.200989
Epoch[62] Batch [449]	Speed: 2801.005662 samples/sec	accuracy=0.648418	lr=0.200771
Epoch[62] Batch [499]	Speed: 2673.512797 samples/sec	accuracy=0.648484	lr=0.200552
Epoch[62] Batch [549]	Speed: 2713.968571 samples/sec	accuracy=0.648184	lr=0.200334
Epoch[62] Batch [599]	Speed: 2691.148533 samples/sec	accuracy=0.648130	lr=0.200116
Epoch[62] Batch [649]	Speed: 2804.429614 samples/sec	accuracy=0.648123	lr=0.199897
Epoch[62] Batch [699]	Speed: 2748.758346 samples/sec	accuracy=0.647699	lr=0.199679
Epoch[62] Batch [749]	Speed: 2665.229374 samples/sec	accuracy=0.647346	lr=0.199461
Epoch[62] Batch [799]	Speed: 2791.537056 samples/sec	accuracy=0.647334	lr=0.199242
Epoch[62] Batch [849]	Speed: 2665.139768 samples/sec	accuracy=0.647005	lr=0.199024
Epoch[62] Batch [899]	Speed: 2765.539317 samples/sec	accuracy=0.646951	lr=0.198806
Epoch[62] Batch [949]	Speed: 2750.223619 samples/sec	accuracy=0.646745	lr=0.198587
Epoch[62] Batch [999]	Speed: 2706.673860 samples/sec	accuracy=0.646703	lr=0.198369
Epoch[62] Batch [1049]	Speed: 2765.097231 samples/sec	accuracy=0.646492	lr=0.198150
Epoch[62] Batch [1099]	Speed: 2840.350103 samples/sec	accuracy=0.646253	lr=0.197932
Epoch[62] Batch [1149]	Speed: 2755.540010 samples/sec	accuracy=0.645949	lr=0.197714
Epoch[62] Batch [1199]	Speed: 2695.591775 samples/sec	accuracy=0.645770	lr=0.197495
Epoch[62] Batch [1249]	Speed: 2966.221337 samples/sec	accuracy=0.645587	lr=0.197277
[Epoch 62] training: accuracy=0.645581
[Epoch 62] speed: 2738 samples/sec	time cost: 477.136127
[Epoch 62] validation: err-top1=0.319276 err-top5=0.108578
Epoch[63] Batch [49]	Speed: 2791.771166 samples/sec	accuracy=0.653418	lr=0.197054
Epoch[63] Batch [99]	Speed: 2678.780186 samples/sec	accuracy=0.653008	lr=0.196836
Epoch[63] Batch [149]	Speed: 2752.217183 samples/sec	accuracy=0.653743	lr=0.196618
Epoch[63] Batch [199]	Speed: 2696.110004 samples/sec	accuracy=0.652988	lr=0.196399
Epoch[63] Batch [249]	Speed: 2756.526276 samples/sec	accuracy=0.652066	lr=0.196181
Epoch[63] Batch [299]	Speed: 2707.321752 samples/sec	accuracy=0.652008	lr=0.195963
Epoch[63] Batch [349]	Speed: 2664.486087 samples/sec	accuracy=0.651286	lr=0.195744
Epoch[63] Batch [399]	Speed: 2643.615454 samples/sec	accuracy=0.651165	lr=0.195526
Epoch[63] Batch [449]	Speed: 2688.652977 samples/sec	accuracy=0.650922	lr=0.195308
Epoch[63] Batch [499]	Speed: 2604.421511 samples/sec	accuracy=0.650324	lr=0.195089
Epoch[63] Batch [549]	Speed: 2695.363570 samples/sec	accuracy=0.650117	lr=0.194871
Epoch[63] Batch [599]	Speed: 2812.299382 samples/sec	accuracy=0.649974	lr=0.194653
Epoch[63] Batch [649]	Speed: 2692.772402 samples/sec	accuracy=0.649590	lr=0.194434
Epoch[63] Batch [699]	Speed: 2745.499230 samples/sec	accuracy=0.649351	lr=0.194216
Epoch[63] Batch [749]	Speed: 2672.547512 samples/sec	accuracy=0.649250	lr=0.193998
Epoch[63] Batch [799]	Speed: 2728.860674 samples/sec	accuracy=0.649067	lr=0.193780
Epoch[63] Batch [849]	Speed: 2915.516180 samples/sec	accuracy=0.648943	lr=0.193561
Epoch[63] Batch [899]	Speed: 2603.669921 samples/sec	accuracy=0.648835	lr=0.193343
Epoch[63] Batch [949]	Speed: 2706.514588 samples/sec	accuracy=0.648588	lr=0.193125
Epoch[63] Batch [999]	Speed: 2785.378642 samples/sec	accuracy=0.648297	lr=0.192907
Epoch[63] Batch [1049]	Speed: 2808.646984 samples/sec	accuracy=0.648198	lr=0.192688
Epoch[63] Batch [1099]	Speed: 2776.656523 samples/sec	accuracy=0.648087	lr=0.192470
Epoch[63] Batch [1149]	Speed: 2786.462782 samples/sec	accuracy=0.648010	lr=0.192252
Epoch[63] Batch [1199]	Speed: 2733.122832 samples/sec	accuracy=0.647961	lr=0.192034
Epoch[63] Batch [1249]	Speed: 2781.879213 samples/sec	accuracy=0.647844	lr=0.191816
[Epoch 63] training: accuracy=0.647843
[Epoch 63] speed: 2725 samples/sec	time cost: 479.205948
[Epoch 63] validation: err-top1=0.310059 err-top5=0.106506
Epoch[64] Batch [49]	Speed: 2666.771270 samples/sec	accuracy=0.659023	lr=0.191593
Epoch[64] Batch [99]	Speed: 2613.982237 samples/sec	accuracy=0.656260	lr=0.191375
Epoch[64] Batch [149]	Speed: 2782.416337 samples/sec	accuracy=0.655339	lr=0.191157
Epoch[64] Batch [199]	Speed: 2689.664125 samples/sec	accuracy=0.655488	lr=0.190939
Epoch[64] Batch [249]	Speed: 2695.151472 samples/sec	accuracy=0.654609	lr=0.190720
Epoch[64] Batch [299]	Speed: 2779.973593 samples/sec	accuracy=0.653535	lr=0.190502
Epoch[64] Batch [349]	Speed: 2628.561406 samples/sec	accuracy=0.653027	lr=0.190284
Epoch[64] Batch [399]	Speed: 2885.734147 samples/sec	accuracy=0.653081	lr=0.190066
Epoch[64] Batch [449]	Speed: 2743.259894 samples/sec	accuracy=0.653008	lr=0.189848
Epoch[64] Batch [499]	Speed: 2650.621296 samples/sec	accuracy=0.653023	lr=0.189630
Epoch[64] Batch [549]	Speed: 2696.145952 samples/sec	accuracy=0.652914	lr=0.189412
Epoch[64] Batch [599]	Speed: 2725.569359 samples/sec	accuracy=0.652384	lr=0.189194
Epoch[64] Batch [649]	Speed: 2741.059439 samples/sec	accuracy=0.652163	lr=0.188976
Epoch[64] Batch [699]	Speed: 2724.103383 samples/sec	accuracy=0.652048	lr=0.188758
Epoch[64] Batch [749]	Speed: 2659.745498 samples/sec	accuracy=0.651781	lr=0.188540
Epoch[64] Batch [799]	Speed: 2668.922317 samples/sec	accuracy=0.651278	lr=0.188322
Epoch[64] Batch [849]	Speed: 2781.918205 samples/sec	accuracy=0.650994	lr=0.188104
Epoch[64] Batch [899]	Speed: 2727.831315 samples/sec	accuracy=0.650808	lr=0.187886
Epoch[64] Batch [949]	Speed: 2691.527143 samples/sec	accuracy=0.650493	lr=0.187668
Epoch[64] Batch [999]	Speed: 2722.334737 samples/sec	accuracy=0.650357	lr=0.187450
Epoch[64] Batch [1049]	Speed: 2715.194692 samples/sec	accuracy=0.650236	lr=0.187232
Epoch[64] Batch [1099]	Speed: 2744.127663 samples/sec	accuracy=0.650183	lr=0.187014
Epoch[64] Batch [1149]	Speed: 2729.015444 samples/sec	accuracy=0.650027	lr=0.186796
Epoch[64] Batch [1199]	Speed: 2664.140163 samples/sec	accuracy=0.649944	lr=0.186578
Epoch[64] Batch [1249]	Speed: 2923.187492 samples/sec	accuracy=0.649905	lr=0.186360
[Epoch 64] training: accuracy=0.649889
[Epoch 64] speed: 2718 samples/sec	time cost: 481.923213
[Epoch 64] validation: err-top1=0.312819 err-top5=0.109694
Epoch[65] Batch [49]	Speed: 2817.203192 samples/sec	accuracy=0.662891	lr=0.186138
Epoch[65] Batch [99]	Speed: 2683.595404 samples/sec	accuracy=0.661094	lr=0.185920
Epoch[65] Batch [149]	Speed: 2680.612392 samples/sec	accuracy=0.660469	lr=0.185702
Epoch[65] Batch [199]	Speed: 2812.036041 samples/sec	accuracy=0.659673	lr=0.185485
Epoch[65] Batch [249]	Speed: 2699.163894 samples/sec	accuracy=0.658488	lr=0.185267
Epoch[65] Batch [299]	Speed: 2691.435862 samples/sec	accuracy=0.657904	lr=0.185049
Epoch[65] Batch [349]	Speed: 2657.205493 samples/sec	accuracy=0.657288	lr=0.184831
Epoch[65] Batch [399]	Speed: 2731.040572 samples/sec	accuracy=0.656680	lr=0.184614
Epoch[65] Batch [449]	Speed: 2706.228327 samples/sec	accuracy=0.656235	lr=0.184396
Epoch[65] Batch [499]	Speed: 2818.293981 samples/sec	accuracy=0.655799	lr=0.184178
Epoch[65] Batch [549]	Speed: 2839.026198 samples/sec	accuracy=0.655298	lr=0.183960
Epoch[65] Batch [599]	Speed: 2753.537464 samples/sec	accuracy=0.654985	lr=0.183743
Epoch[65] Batch [649]	Speed: 2694.331839 samples/sec	accuracy=0.654627	lr=0.183525
Epoch[65] Batch [699]	Speed: 2728.927012 samples/sec	accuracy=0.654203	lr=0.183308
Epoch[65] Batch [749]	Speed: 2682.324744 samples/sec	accuracy=0.654009	lr=0.183090
Epoch[65] Batch [799]	Speed: 2703.354683 samples/sec	accuracy=0.653842	lr=0.182872
Epoch[65] Batch [849]	Speed: 2793.442069 samples/sec	accuracy=0.653683	lr=0.182655
Epoch[65] Batch [899]	Speed: 2730.737016 samples/sec	accuracy=0.653245	lr=0.182437
Epoch[65] Batch [949]	Speed: 2719.078312 samples/sec	accuracy=0.653058	lr=0.182220
Epoch[65] Batch [999]	Speed: 2733.660849 samples/sec	accuracy=0.652813	lr=0.182002
Epoch[65] Batch [1049]	Speed: 2806.740882 samples/sec	accuracy=0.652768	lr=0.181785
Epoch[65] Batch [1099]	Speed: 2763.275682 samples/sec	accuracy=0.652533	lr=0.181567
Epoch[65] Batch [1149]	Speed: 2773.187271 samples/sec	accuracy=0.652537	lr=0.181350
Epoch[65] Batch [1199]	Speed: 2664.972316 samples/sec	accuracy=0.652406	lr=0.181132
Epoch[65] Batch [1249]	Speed: 2895.452521 samples/sec	accuracy=0.652316	lr=0.180915
[Epoch 65] training: accuracy=0.652336
[Epoch 65] speed: 2740 samples/sec	time cost: 477.690147
[Epoch 65] validation: err-top1=0.310627 err-top5=0.103934
Epoch[66] Batch [49]	Speed: 2907.198916 samples/sec	accuracy=0.659785	lr=0.180693
Epoch[66] Batch [99]	Speed: 2669.300373 samples/sec	accuracy=0.661611	lr=0.180476
Epoch[66] Batch [149]	Speed: 2731.112990 samples/sec	accuracy=0.661465	lr=0.180259
Epoch[66] Batch [199]	Speed: 2738.639670 samples/sec	accuracy=0.661685	lr=0.180041
Epoch[66] Batch [249]	Speed: 2763.042736 samples/sec	accuracy=0.661672	lr=0.179824
Epoch[66] Batch [299]	Speed: 2725.085389 samples/sec	accuracy=0.660924	lr=0.179607
Epoch[66] Batch [349]	Speed: 2668.943878 samples/sec	accuracy=0.660338	lr=0.179390
Epoch[66] Batch [399]	Speed: 2828.645388 samples/sec	accuracy=0.660168	lr=0.179172
Epoch[66] Batch [449]	Speed: 2645.855217 samples/sec	accuracy=0.660369	lr=0.178955
Epoch[66] Batch [499]	Speed: 2649.132456 samples/sec	accuracy=0.660127	lr=0.178738
Epoch[66] Batch [549]	Speed: 2701.727784 samples/sec	accuracy=0.659824	lr=0.178521
Epoch[66] Batch [599]	Speed: 2684.052604 samples/sec	accuracy=0.659427	lr=0.178304
Epoch[66] Batch [649]	Speed: 2681.000929 samples/sec	accuracy=0.659043	lr=0.178087
Epoch[66] Batch [699]	Speed: 2691.291397 samples/sec	accuracy=0.658390	lr=0.177870
Epoch[66] Batch [749]	Speed: 2630.987736 samples/sec	accuracy=0.658094	lr=0.177653
Epoch[66] Batch [799]	Speed: 2780.399067 samples/sec	accuracy=0.658020	lr=0.177436
Epoch[66] Batch [849]	Speed: 2733.497689 samples/sec	accuracy=0.657713	lr=0.177219
Epoch[66] Batch [899]	Speed: 2788.934442 samples/sec	accuracy=0.657345	lr=0.177002
Epoch[66] Batch [949]	Speed: 2684.225415 samples/sec	accuracy=0.656884	lr=0.176785
Epoch[66] Batch [999]	Speed: 2762.732166 samples/sec	accuracy=0.656656	lr=0.176568
Epoch[66] Batch [1049]	Speed: 2685.728915 samples/sec	accuracy=0.656401	lr=0.176351
Epoch[66] Batch [1099]	Speed: 2687.400931 samples/sec	accuracy=0.656156	lr=0.176134
Epoch[66] Batch [1149]	Speed: 2679.003318 samples/sec	accuracy=0.655892	lr=0.175918
Epoch[66] Batch [1199]	Speed: 2644.977849 samples/sec	accuracy=0.655656	lr=0.175701
Epoch[66] Batch [1249]	Speed: 3048.283356 samples/sec	accuracy=0.655252	lr=0.175484
[Epoch 66] training: accuracy=0.655250
[Epoch 66] speed: 2723 samples/sec	time cost: 480.382236
[Epoch 66] validation: err-top1=0.311504 err-top5=0.106465
Epoch[67] Batch [49]	Speed: 2646.128717 samples/sec	accuracy=0.666621	lr=0.175263
Epoch[67] Batch [99]	Speed: 2704.482502 samples/sec	accuracy=0.665078	lr=0.175046
Epoch[67] Batch [149]	Speed: 2703.885230 samples/sec	accuracy=0.664720	lr=0.174830
Epoch[67] Batch [199]	Speed: 2842.758877 samples/sec	accuracy=0.663818	lr=0.174613
Epoch[67] Batch [249]	Speed: 2607.257056 samples/sec	accuracy=0.663145	lr=0.174397
Epoch[67] Batch [299]	Speed: 2711.531666 samples/sec	accuracy=0.661937	lr=0.174180
Epoch[67] Batch [349]	Speed: 2794.977535 samples/sec	accuracy=0.661560	lr=0.173964
Epoch[67] Batch [399]	Speed: 2610.614046 samples/sec	accuracy=0.661819	lr=0.173747
Epoch[67] Batch [449]	Speed: 2687.642622 samples/sec	accuracy=0.661545	lr=0.173531
Epoch[67] Batch [499]	Speed: 2797.207257 samples/sec	accuracy=0.660797	lr=0.173314
Epoch[67] Batch [549]	Speed: 2730.318655 samples/sec	accuracy=0.660659	lr=0.173098
Epoch[67] Batch [599]	Speed: 2752.636672 samples/sec	accuracy=0.660169	lr=0.172881
Epoch[67] Batch [649]	Speed: 2776.646148 samples/sec	accuracy=0.659947	lr=0.172665
Epoch[67] Batch [699]	Speed: 2783.133391 samples/sec	accuracy=0.659770	lr=0.172449
Epoch[67] Batch [749]	Speed: 2755.385611 samples/sec	accuracy=0.659590	lr=0.172232
Epoch[67] Batch [799]	Speed: 2862.106202 samples/sec	accuracy=0.659301	lr=0.172016
Epoch[67] Batch [849]	Speed: 2658.513608 samples/sec	accuracy=0.659088	lr=0.171800
Epoch[67] Batch [899]	Speed: 2679.515924 samples/sec	accuracy=0.658763	lr=0.171584
Epoch[67] Batch [949]	Speed: 2754.371581 samples/sec	accuracy=0.658441	lr=0.171368
Epoch[67] Batch [999]	Speed: 2753.077888 samples/sec	accuracy=0.658338	lr=0.171152
Epoch[67] Batch [1049]	Speed: 2710.443596 samples/sec	accuracy=0.658291	lr=0.170935
Epoch[67] Batch [1099]	Speed: 2726.592061 samples/sec	accuracy=0.657826	lr=0.170719
Epoch[67] Batch [1149]	Speed: 2733.437253 samples/sec	accuracy=0.657609	lr=0.170503
Epoch[67] Batch [1199]	Speed: 2656.558162 samples/sec	accuracy=0.657480	lr=0.170287
Epoch[67] Batch [1249]	Speed: 2848.866524 samples/sec	accuracy=0.657207	lr=0.170072
[Epoch 67] training: accuracy=0.657220
[Epoch 67] speed: 2727 samples/sec	time cost: 479.558480
[Epoch 67] validation: err-top1=0.310108 err-top5=0.106585
Epoch[68] Batch [49]	Speed: 2741.093797 samples/sec	accuracy=0.670137	lr=0.169851
Epoch[68] Batch [99]	Speed: 2720.537712 samples/sec	accuracy=0.666982	lr=0.169635
Epoch[68] Batch [149]	Speed: 2716.471195 samples/sec	accuracy=0.666452	lr=0.169420
Epoch[68] Batch [199]	Speed: 2727.232519 samples/sec	accuracy=0.666143	lr=0.169204
Epoch[68] Batch [249]	Speed: 2675.677838 samples/sec	accuracy=0.666465	lr=0.168988
Epoch[68] Batch [299]	Speed: 2721.800720 samples/sec	accuracy=0.666110	lr=0.168772
Epoch[68] Batch [349]	Speed: 2663.316160 samples/sec	accuracy=0.665653	lr=0.168557
Epoch[68] Batch [399]	Speed: 2706.773411 samples/sec	accuracy=0.664675	lr=0.168341
Epoch[68] Batch [449]	Speed: 2829.099569 samples/sec	accuracy=0.664316	lr=0.168126
Epoch[68] Batch [499]	Speed: 2794.624686 samples/sec	accuracy=0.664004	lr=0.167910
Epoch[68] Batch [549]	Speed: 2828.346568 samples/sec	accuracy=0.663507	lr=0.167694
Epoch[68] Batch [599]	Speed: 2670.199930 samples/sec	accuracy=0.663022	lr=0.167479
Epoch[68] Batch [649]	Speed: 2723.667051 samples/sec	accuracy=0.662754	lr=0.167263
Epoch[68] Batch [699]	Speed: 2676.763226 samples/sec	accuracy=0.662574	lr=0.167048
Epoch[68] Batch [749]	Speed: 2745.710095 samples/sec	accuracy=0.662087	lr=0.166833
Epoch[68] Batch [799]	Speed: 2731.290837 samples/sec	accuracy=0.661821	lr=0.166617
Epoch[68] Batch [849]	Speed: 2791.652600 samples/sec	accuracy=0.661466	lr=0.166402
Epoch[68] Batch [899]	Speed: 2664.798834 samples/sec	accuracy=0.661322	lr=0.166187
Epoch[68] Batch [949]	Speed: 2709.245140 samples/sec	accuracy=0.661103	lr=0.165972
Epoch[68] Batch [999]	Speed: 2685.353646 samples/sec	accuracy=0.660970	lr=0.165756
Epoch[68] Batch [1049]	Speed: 2736.973700 samples/sec	accuracy=0.660784	lr=0.165541
Epoch[68] Batch [1099]	Speed: 2712.086355 samples/sec	accuracy=0.660661	lr=0.165326
Epoch[68] Batch [1149]	Speed: 2710.925630 samples/sec	accuracy=0.660434	lr=0.165111
Epoch[68] Batch [1199]	Speed: 2792.247274 samples/sec	accuracy=0.660228	lr=0.164896
Epoch[68] Batch [1249]	Speed: 2860.866786 samples/sec	accuracy=0.659899	lr=0.164681
[Epoch 68] training: accuracy=0.659906
[Epoch 68] speed: 2730 samples/sec	time cost: 479.013104
[Epoch 68] validation: err-top1=0.301080 err-top5=0.102679
Epoch[69] Batch [49]	Speed: 2729.906531 samples/sec	accuracy=0.670430	lr=0.164462
Epoch[69] Batch [99]	Speed: 2808.395381 samples/sec	accuracy=0.669561	lr=0.164247
Epoch[69] Batch [149]	Speed: 2676.662067 samples/sec	accuracy=0.668724	lr=0.164032
Epoch[69] Batch [199]	Speed: 2725.006341 samples/sec	accuracy=0.668799	lr=0.163818
Epoch[69] Batch [249]	Speed: 2683.709463 samples/sec	accuracy=0.668023	lr=0.163603
Epoch[69] Batch [299]	Speed: 2696.698597 samples/sec	accuracy=0.667594	lr=0.163388
Epoch[69] Batch [349]	Speed: 2913.459116 samples/sec	accuracy=0.666928	lr=0.163173
Epoch[69] Batch [399]	Speed: 2778.745344 samples/sec	accuracy=0.666912	lr=0.162959
Epoch[69] Batch [449]	Speed: 2665.953681 samples/sec	accuracy=0.666808	lr=0.162744
Epoch[69] Batch [499]	Speed: 2709.393932 samples/sec	accuracy=0.666777	lr=0.162530
Epoch[69] Batch [549]	Speed: 2675.692374 samples/sec	accuracy=0.665998	lr=0.162315
Epoch[69] Batch [599]	Speed: 2765.998574 samples/sec	accuracy=0.665776	lr=0.162101
Epoch[69] Batch [649]	Speed: 2739.016356 samples/sec	accuracy=0.665518	lr=0.161886
Epoch[69] Batch [699]	Speed: 2847.525457 samples/sec	accuracy=0.665153	lr=0.161672
Epoch[69] Batch [749]	Speed: 2755.511830 samples/sec	accuracy=0.664741	lr=0.161458
Epoch[69] Batch [799]	Speed: 2773.012734 samples/sec	accuracy=0.664517	lr=0.161244
Epoch[69] Batch [849]	Speed: 2801.603560 samples/sec	accuracy=0.664151	lr=0.161029
Epoch[69] Batch [899]	Speed: 2765.087974 samples/sec	accuracy=0.664071	lr=0.160815
Epoch[69] Batch [949]	Speed: 2887.140603 samples/sec	accuracy=0.663869	lr=0.160601
Epoch[69] Batch [999]	Speed: 2725.118483 samples/sec	accuracy=0.663672	lr=0.160387
Epoch[69] Batch [1049]	Speed: 2730.037645 samples/sec	accuracy=0.663405	lr=0.160173
Epoch[69] Batch [1099]	Speed: 2785.958971 samples/sec	accuracy=0.663279	lr=0.159959
Epoch[69] Batch [1149]	Speed: 2682.259112 samples/sec	accuracy=0.663078	lr=0.159745
Epoch[69] Batch [1199]	Speed: 2721.863782 samples/sec	accuracy=0.662992	lr=0.159531
Epoch[69] Batch [1249]	Speed: 2888.023617 samples/sec	accuracy=0.662880	lr=0.159317
[Epoch 69] training: accuracy=0.662898
[Epoch 69] speed: 2753 samples/sec	time cost: 474.510246
[Epoch 69] validation: err-top1=0.297078 err-top5=0.097839
Epoch[70] Batch [49]	Speed: 2761.805707 samples/sec	accuracy=0.676797	lr=0.159099
Epoch[70] Batch [99]	Speed: 2719.670088 samples/sec	accuracy=0.676006	lr=0.158886
Epoch[70] Batch [149]	Speed: 2736.170376 samples/sec	accuracy=0.674030	lr=0.158672
Epoch[70] Batch [199]	Speed: 2804.259399 samples/sec	accuracy=0.673672	lr=0.158458
Epoch[70] Batch [249]	Speed: 2748.385905 samples/sec	accuracy=0.672559	lr=0.158245
Epoch[70] Batch [299]	Speed: 2665.764187 samples/sec	accuracy=0.671514	lr=0.158031
Epoch[70] Batch [349]	Speed: 2787.611415 samples/sec	accuracy=0.670949	lr=0.157818
Epoch[70] Batch [399]	Speed: 2713.132314 samples/sec	accuracy=0.670471	lr=0.157604
Epoch[70] Batch [449]	Speed: 2700.538720 samples/sec	accuracy=0.669907	lr=0.157391
Epoch[70] Batch [499]	Speed: 2637.266885 samples/sec	accuracy=0.669361	lr=0.157177
Epoch[70] Batch [549]	Speed: 2696.720236 samples/sec	accuracy=0.668855	lr=0.156964
Epoch[70] Batch [599]	Speed: 2766.040436 samples/sec	accuracy=0.668211	lr=0.156751
Epoch[70] Batch [649]	Speed: 2655.585835 samples/sec	accuracy=0.667545	lr=0.156538
Epoch[70] Batch [699]	Speed: 2690.653717 samples/sec	accuracy=0.667386	lr=0.156325
Epoch[70] Batch [749]	Speed: 2748.222038 samples/sec	accuracy=0.667225	lr=0.156112
Epoch[70] Batch [799]	Speed: 2800.291785 samples/sec	accuracy=0.666981	lr=0.155899
Epoch[70] Batch [849]	Speed: 2826.489644 samples/sec	accuracy=0.666698	lr=0.155686
Epoch[70] Batch [899]	Speed: 2711.567102 samples/sec	accuracy=0.666557	lr=0.155473
Epoch[70] Batch [949]	Speed: 2705.755701 samples/sec	accuracy=0.666414	lr=0.155260
Epoch[70] Batch [999]	Speed: 2841.419906 samples/sec	accuracy=0.666312	lr=0.155047
Epoch[70] Batch [1049]	Speed: 2812.933320 samples/sec	accuracy=0.666134	lr=0.154834
Epoch[70] Batch [1099]	Speed: 2793.843288 samples/sec	accuracy=0.665861	lr=0.154622
Epoch[70] Batch [1149]	Speed: 2698.389016 samples/sec	accuracy=0.665572	lr=0.154409
Epoch[70] Batch [1199]	Speed: 2741.341219 samples/sec	accuracy=0.665351	lr=0.154196
Epoch[70] Batch [1249]	Speed: 2852.232471 samples/sec	accuracy=0.665215	lr=0.153984
[Epoch 70] training: accuracy=0.665220
[Epoch 70] speed: 2741 samples/sec	time cost: 477.378429
[Epoch 70] validation: err-top1=0.301319 err-top5=0.100765
Epoch[71] Batch [49]	Speed: 2732.920783 samples/sec	accuracy=0.673008	lr=0.153767
Epoch[71] Batch [99]	Speed: 2708.989535 samples/sec	accuracy=0.673389	lr=0.153555
Epoch[71] Batch [149]	Speed: 2742.227734 samples/sec	accuracy=0.674082	lr=0.153342
Epoch[71] Batch [199]	Speed: 2738.595909 samples/sec	accuracy=0.673770	lr=0.153130
Epoch[71] Batch [249]	Speed: 2748.926148 samples/sec	accuracy=0.672965	lr=0.152918
Epoch[71] Batch [299]	Speed: 2730.730904 samples/sec	accuracy=0.672445	lr=0.152705
Epoch[71] Batch [349]	Speed: 2641.613322 samples/sec	accuracy=0.672174	lr=0.152493
Epoch[71] Batch [399]	Speed: 2639.948449 samples/sec	accuracy=0.671897	lr=0.152281
Epoch[71] Batch [449]	Speed: 2649.079516 samples/sec	accuracy=0.671313	lr=0.152069
Epoch[71] Batch [499]	Speed: 2723.060689 samples/sec	accuracy=0.671016	lr=0.151857
Epoch[71] Batch [549]	Speed: 2699.100997 samples/sec	accuracy=0.670916	lr=0.151645
Epoch[71] Batch [599]	Speed: 2756.687738 samples/sec	accuracy=0.670545	lr=0.151433
Epoch[71] Batch [649]	Speed: 2698.737074 samples/sec	accuracy=0.670180	lr=0.151222
Epoch[71] Batch [699]	Speed: 2719.740698 samples/sec	accuracy=0.669990	lr=0.151010
Epoch[71] Batch [749]	Speed: 2645.015998 samples/sec	accuracy=0.669918	lr=0.150798
Epoch[71] Batch [799]	Speed: 2729.614850 samples/sec	accuracy=0.669875	lr=0.150587
Epoch[71] Batch [849]	Speed: 2662.375948 samples/sec	accuracy=0.669708	lr=0.150375
Epoch[71] Batch [899]	Speed: 2716.037992 samples/sec	accuracy=0.669223	lr=0.150163
Epoch[71] Batch [949]	Speed: 2719.826538 samples/sec	accuracy=0.669003	lr=0.149952
Epoch[71] Batch [999]	Speed: 2663.505141 samples/sec	accuracy=0.668854	lr=0.149741
Epoch[71] Batch [1049]	Speed: 2689.036778 samples/sec	accuracy=0.668627	lr=0.149529
Epoch[71] Batch [1099]	Speed: 2683.923891 samples/sec	accuracy=0.668619	lr=0.149318
Epoch[71] Batch [1149]	Speed: 2649.337112 samples/sec	accuracy=0.668558	lr=0.149107
Epoch[71] Batch [1199]	Speed: 2716.598582 samples/sec	accuracy=0.668402	lr=0.148896
Epoch[71] Batch [1249]	Speed: 2872.491631 samples/sec	accuracy=0.668359	lr=0.148685
[Epoch 71] training: accuracy=0.668380
[Epoch 71] speed: 2704 samples/sec	time cost: 483.936571
[Epoch 71] validation: err-top1=0.295181 err-top5=0.096720
Epoch[72] Batch [49]	Speed: 2810.755021 samples/sec	accuracy=0.677363	lr=0.148469
Epoch[72] Batch [99]	Speed: 2769.245861 samples/sec	accuracy=0.677627	lr=0.148258
Epoch[72] Batch [149]	Speed: 2644.649707 samples/sec	accuracy=0.677188	lr=0.148047
Epoch[72] Batch [199]	Speed: 2766.236652 samples/sec	accuracy=0.677109	lr=0.147837
Epoch[72] Batch [249]	Speed: 2842.839335 samples/sec	accuracy=0.676453	lr=0.147626
Epoch[72] Batch [299]	Speed: 2702.249974 samples/sec	accuracy=0.676266	lr=0.147415
Epoch[72] Batch [349]	Speed: 2792.101840 samples/sec	accuracy=0.675778	lr=0.147204
Epoch[72] Batch [399]	Speed: 2745.613347 samples/sec	accuracy=0.675566	lr=0.146994
Epoch[72] Batch [449]	Speed: 2710.888020 samples/sec	accuracy=0.675356	lr=0.146783
Epoch[72] Batch [499]	Speed: 2700.395857 samples/sec	accuracy=0.675148	lr=0.146573
Epoch[72] Batch [549]	Speed: 2675.092087 samples/sec	accuracy=0.674616	lr=0.146362
Epoch[72] Batch [599]	Speed: 2685.125762 samples/sec	accuracy=0.673921	lr=0.146152
Epoch[72] Batch [649]	Speed: 2687.341708 samples/sec	accuracy=0.673941	lr=0.145942
Epoch[72] Batch [699]	Speed: 2721.085679 samples/sec	accuracy=0.673562	lr=0.145732
Epoch[72] Batch [749]	Speed: 2695.722760 samples/sec	accuracy=0.673038	lr=0.145521
Epoch[72] Batch [799]	Speed: 2703.222547 samples/sec	accuracy=0.672828	lr=0.145311
Epoch[72] Batch [849]	Speed: 2712.735264 samples/sec	accuracy=0.672431	lr=0.145101
Epoch[72] Batch [899]	Speed: 2720.831627 samples/sec	accuracy=0.672116	lr=0.144891
Epoch[72] Batch [949]	Speed: 2685.925457 samples/sec	accuracy=0.671846	lr=0.144682
Epoch[72] Batch [999]	Speed: 2626.866634 samples/sec	accuracy=0.671729	lr=0.144472
Epoch[72] Batch [1049]	Speed: 2699.918850 samples/sec	accuracy=0.671665	lr=0.144262
Epoch[72] Batch [1099]	Speed: 2751.200760 samples/sec	accuracy=0.671374	lr=0.144052
Epoch[72] Batch [1149]	Speed: 2676.665137 samples/sec	accuracy=0.671143	lr=0.143843
Epoch[72] Batch [1199]	Speed: 2728.975458 samples/sec	accuracy=0.670963	lr=0.143633
Epoch[72] Batch [1249]	Speed: 2937.001405 samples/sec	accuracy=0.670891	lr=0.143424
[Epoch 72] training: accuracy=0.670884
[Epoch 72] speed: 2724 samples/sec	time cost: 479.945084
[Epoch 72] validation: err-top1=0.291753 err-top5=0.095683
Epoch[73] Batch [49]	Speed: 2688.057998 samples/sec	accuracy=0.680703	lr=0.143210
Epoch[73] Batch [99]	Speed: 2804.594759 samples/sec	accuracy=0.681035	lr=0.143001
Epoch[73] Batch [149]	Speed: 2716.222470 samples/sec	accuracy=0.680430	lr=0.142791
Epoch[73] Batch [199]	Speed: 2686.954995 samples/sec	accuracy=0.679365	lr=0.142582
Epoch[73] Batch [249]	Speed: 2738.768900 samples/sec	accuracy=0.678625	lr=0.142373
Epoch[73] Batch [299]	Speed: 2679.820672 samples/sec	accuracy=0.678444	lr=0.142164
Epoch[73] Batch [349]	Speed: 2735.531394 samples/sec	accuracy=0.678203	lr=0.141955
Epoch[73] Batch [399]	Speed: 2671.425694 samples/sec	accuracy=0.677671	lr=0.141746
Epoch[73] Batch [449]	Speed: 2666.853633 samples/sec	accuracy=0.677214	lr=0.141537
Epoch[73] Batch [499]	Speed: 2720.461546 samples/sec	accuracy=0.676674	lr=0.141328
Epoch[73] Batch [549]	Speed: 2677.895142 samples/sec	accuracy=0.676199	lr=0.141120
Epoch[73] Batch [599]	Speed: 2698.516747 samples/sec	accuracy=0.675978	lr=0.140911
Epoch[73] Batch [649]	Speed: 2738.186066 samples/sec	accuracy=0.675563	lr=0.140702
Epoch[73] Batch [699]	Speed: 2729.136899 samples/sec	accuracy=0.675449	lr=0.140494
Epoch[73] Batch [749]	Speed: 2707.098314 samples/sec	accuracy=0.675126	lr=0.140285
Epoch[73] Batch [799]	Speed: 2709.874635 samples/sec	accuracy=0.674829	lr=0.140077
Epoch[73] Batch [849]	Speed: 2801.551770 samples/sec	accuracy=0.674759	lr=0.139869
Epoch[73] Batch [899]	Speed: 2710.074283 samples/sec	accuracy=0.674590	lr=0.139661
Epoch[73] Batch [949]	Speed: 2679.327037 samples/sec	accuracy=0.674453	lr=0.139452
Epoch[73] Batch [999]	Speed: 2662.758094 samples/sec	accuracy=0.674354	lr=0.139244
Epoch[73] Batch [1049]	Speed: 2814.448649 samples/sec	accuracy=0.674179	lr=0.139036
Epoch[73] Batch [1099]	Speed: 2731.591633 samples/sec	accuracy=0.673978	lr=0.138828
Epoch[73] Batch [1149]	Speed: 2830.845694 samples/sec	accuracy=0.673820	lr=0.138620
Epoch[73] Batch [1199]	Speed: 2759.483322 samples/sec	accuracy=0.673670	lr=0.138413
Epoch[73] Batch [1249]	Speed: 2890.439539 samples/sec	accuracy=0.673359	lr=0.138205
[Epoch 73] training: accuracy=0.673374
[Epoch 73] speed: 2726 samples/sec	time cost: 479.421905
[Epoch 73] validation: err-top1=0.293666 err-top5=0.096879
Epoch[74] Batch [49]	Speed: 2807.650973 samples/sec	accuracy=0.689336	lr=0.137993
Epoch[74] Batch [99]	Speed: 2725.109319 samples/sec	accuracy=0.685391	lr=0.137786
Epoch[74] Batch [149]	Speed: 2711.078645 samples/sec	accuracy=0.684414	lr=0.137578
Epoch[74] Batch [199]	Speed: 2757.604569 samples/sec	accuracy=0.683887	lr=0.137371
Epoch[74] Batch [249]	Speed: 2735.010440 samples/sec	accuracy=0.682586	lr=0.137163
Epoch[74] Batch [299]	Speed: 2713.494232 samples/sec	accuracy=0.681527	lr=0.136956
Epoch[74] Batch [349]	Speed: 2799.189343 samples/sec	accuracy=0.680686	lr=0.136749
Epoch[74] Batch [399]	Speed: 2725.258026 samples/sec	accuracy=0.680249	lr=0.136542
Epoch[74] Batch [449]	Speed: 2733.236618 samples/sec	accuracy=0.679362	lr=0.136335
Epoch[74] Batch [499]	Speed: 2608.437138 samples/sec	accuracy=0.679504	lr=0.136128
Epoch[74] Batch [549]	Speed: 2743.219945 samples/sec	accuracy=0.679332	lr=0.135921
Epoch[74] Batch [599]	Speed: 2784.698129 samples/sec	accuracy=0.679360	lr=0.135714
Epoch[74] Batch [649]	Speed: 2726.946118 samples/sec	accuracy=0.679133	lr=0.135507
Epoch[74] Batch [699]	Speed: 2739.480195 samples/sec	accuracy=0.678934	lr=0.135301
Epoch[74] Batch [749]	Speed: 2685.672016 samples/sec	accuracy=0.678440	lr=0.135094
Epoch[74] Batch [799]	Speed: 2698.401697 samples/sec	accuracy=0.678381	lr=0.134887
Epoch[74] Batch [849]	Speed: 2657.231402 samples/sec	accuracy=0.678127	lr=0.134681
Epoch[74] Batch [899]	Speed: 2804.568607 samples/sec	accuracy=0.677867	lr=0.134475
Epoch[74] Batch [949]	Speed: 2713.855835 samples/sec	accuracy=0.677801	lr=0.134268
Epoch[74] Batch [999]	Speed: 2671.762776 samples/sec	accuracy=0.677587	lr=0.134062
Epoch[74] Batch [1049]	Speed: 2732.876440 samples/sec	accuracy=0.677448	lr=0.133856
Epoch[74] Batch [1099]	Speed: 2723.619657 samples/sec	accuracy=0.677207	lr=0.133650
Epoch[74] Batch [1149]	Speed: 2716.709586 samples/sec	accuracy=0.676878	lr=0.133444
Epoch[74] Batch [1199]	Speed: 2751.952383 samples/sec	accuracy=0.676598	lr=0.133238
Epoch[74] Batch [1249]	Speed: 2840.174860 samples/sec	accuracy=0.676426	lr=0.133032
[Epoch 74] training: accuracy=0.676441
[Epoch 74] speed: 2729 samples/sec	time cost: 479.100374
[Epoch 74] validation: err-top1=0.286193 err-top5=0.093571
Epoch[75] Batch [49]	Speed: 2680.486819 samples/sec	accuracy=0.687480	lr=0.132822
Epoch[75] Batch [99]	Speed: 2717.771359 samples/sec	accuracy=0.686104	lr=0.132617
Epoch[75] Batch [149]	Speed: 2727.701245 samples/sec	accuracy=0.685853	lr=0.132411
Epoch[75] Batch [199]	Speed: 2915.437295 samples/sec	accuracy=0.685688	lr=0.132206
Epoch[75] Batch [249]	Speed: 2663.098672 samples/sec	accuracy=0.684863	lr=0.132000
Epoch[75] Batch [299]	Speed: 2796.172411 samples/sec	accuracy=0.684362	lr=0.131795
Epoch[75] Batch [349]	Speed: 2674.468721 samples/sec	accuracy=0.684425	lr=0.131590
Epoch[75] Batch [399]	Speed: 2714.092464 samples/sec	accuracy=0.684011	lr=0.131385
Epoch[75] Batch [449]	Speed: 2635.967909 samples/sec	accuracy=0.683626	lr=0.131180
Epoch[75] Batch [499]	Speed: 2736.526471 samples/sec	accuracy=0.683191	lr=0.130975
Epoch[75] Batch [549]	Speed: 2638.510441 samples/sec	accuracy=0.682836	lr=0.130770
Epoch[75] Batch [599]	Speed: 2697.581510 samples/sec	accuracy=0.682397	lr=0.130565
Epoch[75] Batch [649]	Speed: 2761.429864 samples/sec	accuracy=0.682339	lr=0.130360
Epoch[75] Batch [699]	Speed: 2715.699606 samples/sec	accuracy=0.682058	lr=0.130155
Epoch[75] Batch [749]	Speed: 2618.197935 samples/sec	accuracy=0.681572	lr=0.129951
Epoch[75] Batch [799]	Speed: 2678.735778 samples/sec	accuracy=0.681278	lr=0.129746
Epoch[75] Batch [849]	Speed: 2748.778119 samples/sec	accuracy=0.681060	lr=0.129542
Epoch[75] Batch [899]	Speed: 2714.677851 samples/sec	accuracy=0.681039	lr=0.129338
Epoch[75] Batch [949]	Speed: 2752.386325 samples/sec	accuracy=0.680701	lr=0.129133
Epoch[75] Batch [999]	Speed: 2695.957252 samples/sec	accuracy=0.680447	lr=0.128929
Epoch[75] Batch [1049]	Speed: 2775.660179 samples/sec	accuracy=0.680257	lr=0.128725
Epoch[75] Batch [1099]	Speed: 2767.293850 samples/sec	accuracy=0.680080	lr=0.128521
Epoch[75] Batch [1149]	Speed: 2738.478709 samples/sec	accuracy=0.679890	lr=0.128317
Epoch[75] Batch [1199]	Speed: 2728.061723 samples/sec	accuracy=0.679708	lr=0.128113
Epoch[75] Batch [1249]	Speed: 2846.396120 samples/sec	accuracy=0.679451	lr=0.127910
[Epoch 75] training: accuracy=0.679467
[Epoch 75] speed: 2722 samples/sec	time cost: 480.879295
[Epoch 75] validation: err-top1=0.287842 err-top5=0.092712
Epoch[76] Batch [49]	Speed: 2975.789180 samples/sec	accuracy=0.687754	lr=0.127702
Epoch[76] Batch [99]	Speed: 2796.102327 samples/sec	accuracy=0.689736	lr=0.127498
Epoch[76] Batch [149]	Speed: 2774.085941 samples/sec	accuracy=0.688939	lr=0.127295
Epoch[76] Batch [199]	Speed: 2695.887702 samples/sec	accuracy=0.688301	lr=0.127092
Epoch[76] Batch [249]	Speed: 2680.249256 samples/sec	accuracy=0.687480	lr=0.126888
Epoch[76] Batch [299]	Speed: 2665.980191 samples/sec	accuracy=0.686943	lr=0.126685
Epoch[76] Batch [349]	Speed: 2736.797693 samples/sec	accuracy=0.686342	lr=0.126482
Epoch[76] Batch [399]	Speed: 2679.848061 samples/sec	accuracy=0.686335	lr=0.126279
Epoch[76] Batch [449]	Speed: 2760.144718 samples/sec	accuracy=0.686411	lr=0.126076
Epoch[76] Batch [499]	Speed: 2634.969242 samples/sec	accuracy=0.686197	lr=0.125873
Epoch[76] Batch [549]	Speed: 2658.035951 samples/sec	accuracy=0.685861	lr=0.125670
Epoch[76] Batch [599]	Speed: 2733.638230 samples/sec	accuracy=0.685382	lr=0.125468
Epoch[76] Batch [649]	Speed: 2647.015593 samples/sec	accuracy=0.685179	lr=0.125265
Epoch[76] Batch [699]	Speed: 2716.029679 samples/sec	accuracy=0.684891	lr=0.125062
Epoch[76] Batch [749]	Speed: 2727.108046 samples/sec	accuracy=0.684665	lr=0.124860
Epoch[76] Batch [799]	Speed: 2686.756621 samples/sec	accuracy=0.684629	lr=0.124658
Epoch[76] Batch [849]	Speed: 2737.297486 samples/sec	accuracy=0.684247	lr=0.124455
Epoch[76] Batch [899]	Speed: 2726.378342 samples/sec	accuracy=0.684078	lr=0.124253
Epoch[76] Batch [949]	Speed: 2733.051073 samples/sec	accuracy=0.683620	lr=0.124051
Epoch[76] Batch [999]	Speed: 2743.156590 samples/sec	accuracy=0.683371	lr=0.123849
Epoch[76] Batch [1049]	Speed: 2654.172439 samples/sec	accuracy=0.683093	lr=0.123647
Epoch[76] Batch [1099]	Speed: 2762.042139 samples/sec	accuracy=0.683023	lr=0.123446
Epoch[76] Batch [1149]	Speed: 2701.338653 samples/sec	accuracy=0.682940	lr=0.123244
Epoch[76] Batch [1199]	Speed: 2701.762590 samples/sec	accuracy=0.682813	lr=0.123042
Epoch[76] Batch [1249]	Speed: 2876.636379 samples/sec	accuracy=0.682481	lr=0.122841
[Epoch 76] training: accuracy=0.682495
[Epoch 76] speed: 2724 samples/sec	time cost: 480.104830
[Epoch 76] validation: err-top1=0.288744 err-top5=0.092136
Epoch[77] Batch [49]	Speed: 2660.386770 samples/sec	accuracy=0.695781	lr=0.122635
Epoch[77] Batch [99]	Speed: 2657.052153 samples/sec	accuracy=0.694385	lr=0.122434
Epoch[77] Batch [149]	Speed: 2685.711449 samples/sec	accuracy=0.693711	lr=0.122233
Epoch[77] Batch [199]	Speed: 2762.230789 samples/sec	accuracy=0.692661	lr=0.122032
Epoch[77] Batch [249]	Speed: 2801.636821 samples/sec	accuracy=0.691688	lr=0.121831
Epoch[77] Batch [299]	Speed: 2689.888670 samples/sec	accuracy=0.691182	lr=0.121630
Epoch[77] Batch [349]	Speed: 2754.848057 samples/sec	accuracy=0.690698	lr=0.121429
Epoch[77] Batch [399]	Speed: 2781.454691 samples/sec	accuracy=0.690117	lr=0.121228
Epoch[77] Batch [449]	Speed: 2658.835160 samples/sec	accuracy=0.689412	lr=0.121027
Epoch[77] Batch [499]	Speed: 2734.328480 samples/sec	accuracy=0.689129	lr=0.120827
Epoch[77] Batch [549]	Speed: 2706.099183 samples/sec	accuracy=0.688507	lr=0.120626
Epoch[77] Batch [599]	Speed: 2681.511789 samples/sec	accuracy=0.688203	lr=0.120426
Epoch[77] Batch [649]	Speed: 2710.392659 samples/sec	accuracy=0.687867	lr=0.120226
Epoch[77] Batch [699]	Speed: 2667.238722 samples/sec	accuracy=0.687878	lr=0.120025
Epoch[77] Batch [749]	Speed: 2747.796370 samples/sec	accuracy=0.687250	lr=0.119825
Epoch[77] Batch [799]	Speed: 2732.482630 samples/sec	accuracy=0.686902	lr=0.119625
Epoch[77] Batch [849]	Speed: 2716.754987 samples/sec	accuracy=0.686772	lr=0.119425
Epoch[77] Batch [899]	Speed: 2656.917010 samples/sec	accuracy=0.686506	lr=0.119226
Epoch[77] Batch [949]	Speed: 2776.710915 samples/sec	accuracy=0.686344	lr=0.119026
Epoch[77] Batch [999]	Speed: 2827.083660 samples/sec	accuracy=0.686131	lr=0.118826
Epoch[77] Batch [1049]	Speed: 2737.552598 samples/sec	accuracy=0.685845	lr=0.118627
Epoch[77] Batch [1099]	Speed: 2649.720200 samples/sec	accuracy=0.685781	lr=0.118427
Epoch[77] Batch [1149]	Speed: 2717.233974 samples/sec	accuracy=0.685587	lr=0.118228
Epoch[77] Batch [1199]	Speed: 2678.344689 samples/sec	accuracy=0.685573	lr=0.118029
Epoch[77] Batch [1249]	Speed: 2951.652361 samples/sec	accuracy=0.685409	lr=0.117830
[Epoch 77] training: accuracy=0.685408
[Epoch 77] speed: 2722 samples/sec	time cost: 480.303812
[Epoch 77] validation: err-top1=0.285674 err-top5=0.093650
Epoch[78] Batch [49]	Speed: 2659.680439 samples/sec	accuracy=0.695312	lr=0.117627
Epoch[78] Batch [99]	Speed: 2651.405642 samples/sec	accuracy=0.695615	lr=0.117428
Epoch[78] Batch [149]	Speed: 2690.148872 samples/sec	accuracy=0.695814	lr=0.117229
Epoch[78] Batch [199]	Speed: 2690.933152 samples/sec	accuracy=0.694922	lr=0.117030
Epoch[78] Batch [249]	Speed: 2698.314255 samples/sec	accuracy=0.694730	lr=0.116831
Epoch[78] Batch [299]	Speed: 2676.905001 samples/sec	accuracy=0.693330	lr=0.116633
Epoch[78] Batch [349]	Speed: 2799.460210 samples/sec	accuracy=0.692651	lr=0.116434
Epoch[78] Batch [399]	Speed: 2706.221199 samples/sec	accuracy=0.691924	lr=0.116236
Epoch[78] Batch [449]	Speed: 2687.051318 samples/sec	accuracy=0.691936	lr=0.116038
Epoch[78] Batch [499]	Speed: 2715.909215 samples/sec	accuracy=0.691686	lr=0.115840
Epoch[78] Batch [549]	Speed: 2718.663516 samples/sec	accuracy=0.691161	lr=0.115642
Epoch[78] Batch [599]	Speed: 2867.306300 samples/sec	accuracy=0.691152	lr=0.115444
Epoch[78] Batch [649]	Speed: 2703.432650 samples/sec	accuracy=0.690841	lr=0.115246
Epoch[78] Batch [699]	Speed: 2618.779567 samples/sec	accuracy=0.690372	lr=0.115048
Epoch[78] Batch [749]	Speed: 2779.143826 samples/sec	accuracy=0.690108	lr=0.114850
Epoch[78] Batch [799]	Speed: 2711.507802 samples/sec	accuracy=0.689928	lr=0.114653
Epoch[78] Batch [849]	Speed: 2713.519673 samples/sec	accuracy=0.689839	lr=0.114455
Epoch[78] Batch [899]	Speed: 2744.417299 samples/sec	accuracy=0.689582	lr=0.114258
Epoch[78] Batch [949]	Speed: 2750.744606 samples/sec	accuracy=0.689344	lr=0.114061
Epoch[78] Batch [999]	Speed: 2699.001807 samples/sec	accuracy=0.689084	lr=0.113864
Epoch[78] Batch [1049]	Speed: 2688.777969 samples/sec	accuracy=0.688802	lr=0.113667
Epoch[78] Batch [1099]	Speed: 2659.853815 samples/sec	accuracy=0.688827	lr=0.113470
Epoch[78] Batch [1149]	Speed: 2723.709645 samples/sec	accuracy=0.688659	lr=0.113273
Epoch[78] Batch [1199]	Speed: 2609.481649 samples/sec	accuracy=0.688614	lr=0.113076
Epoch[78] Batch [1249]	Speed: 2871.303709 samples/sec	accuracy=0.688445	lr=0.112880
[Epoch 78] training: accuracy=0.688463
[Epoch 78] speed: 2710 samples/sec	time cost: 482.533491
[Epoch 78] validation: err-top1=0.282127 err-top5=0.087751
Epoch[79] Batch [49]	Speed: 2786.604773 samples/sec	accuracy=0.697949	lr=0.112679
Epoch[79] Batch [99]	Speed: 2800.049344 samples/sec	accuracy=0.699541	lr=0.112483
Epoch[79] Batch [149]	Speed: 2772.719251 samples/sec	accuracy=0.698737	lr=0.112287
Epoch[79] Batch [199]	Speed: 2731.001395 samples/sec	accuracy=0.697339	lr=0.112090
Epoch[79] Batch [249]	Speed: 2658.451506 samples/sec	accuracy=0.697055	lr=0.111894
Epoch[79] Batch [299]	Speed: 2758.530018 samples/sec	accuracy=0.696637	lr=0.111698
Epoch[79] Batch [349]	Speed: 2710.899929 samples/sec	accuracy=0.696708	lr=0.111502
Epoch[79] Batch [399]	Speed: 2749.201839 samples/sec	accuracy=0.695957	lr=0.111307
Epoch[79] Batch [449]	Speed: 2741.357246 samples/sec	accuracy=0.695651	lr=0.111111
Epoch[79] Batch [499]	Speed: 2702.416975 samples/sec	accuracy=0.695072	lr=0.110915
Epoch[79] Batch [549]	Speed: 2710.429536 samples/sec	accuracy=0.694957	lr=0.110720
Epoch[79] Batch [599]	Speed: 2691.120103 samples/sec	accuracy=0.694935	lr=0.110525
Epoch[79] Batch [649]	Speed: 2663.921186 samples/sec	accuracy=0.694422	lr=0.110329
Epoch[79] Batch [699]	Speed: 2644.655960 samples/sec	accuracy=0.694173	lr=0.110134
Epoch[79] Batch [749]	Speed: 2729.396598 samples/sec	accuracy=0.693911	lr=0.109939
Epoch[79] Batch [799]	Speed: 2785.279836 samples/sec	accuracy=0.693347	lr=0.109744
Epoch[79] Batch [849]	Speed: 2763.114621 samples/sec	accuracy=0.693142	lr=0.109549
Epoch[79] Batch [899]	Speed: 2679.230164 samples/sec	accuracy=0.692827	lr=0.109355
Epoch[79] Batch [949]	Speed: 2721.620726 samples/sec	accuracy=0.692657	lr=0.109160
Epoch[79] Batch [999]	Speed: 2734.410125 samples/sec	accuracy=0.692528	lr=0.108966
Epoch[79] Batch [1049]	Speed: 2860.866291 samples/sec	accuracy=0.692154	lr=0.108771
Epoch[79] Batch [1099]	Speed: 2666.224632 samples/sec	accuracy=0.691898	lr=0.108577
Epoch[79] Batch [1149]	Speed: 2680.064882 samples/sec	accuracy=0.691745	lr=0.108383
Epoch[79] Batch [1199]	Speed: 2757.063883 samples/sec	accuracy=0.691637	lr=0.108189
Epoch[79] Batch [1249]	Speed: 2912.809722 samples/sec	accuracy=0.691663	lr=0.107995
[Epoch 79] training: accuracy=0.691662
[Epoch 79] speed: 2733 samples/sec	time cost: 478.118989
[Epoch 79] validation: err-top1=0.279596 err-top5=0.089724
Epoch[80] Batch [49]	Speed: 2843.873767 samples/sec	accuracy=0.706934	lr=0.107797
Epoch[80] Batch [99]	Speed: 2670.207069 samples/sec	accuracy=0.704922	lr=0.107603
Epoch[80] Batch [149]	Speed: 2704.085699 samples/sec	accuracy=0.702129	lr=0.107410
Epoch[80] Batch [199]	Speed: 2681.799676 samples/sec	accuracy=0.701890	lr=0.107216
Epoch[80] Batch [249]	Speed: 2797.599718 samples/sec	accuracy=0.701633	lr=0.107023
Epoch[80] Batch [299]	Speed: 2654.759631 samples/sec	accuracy=0.700856	lr=0.106830
Epoch[80] Batch [349]	Speed: 2738.514608 samples/sec	accuracy=0.699852	lr=0.106636
Epoch[80] Batch [399]	Speed: 2759.637257 samples/sec	accuracy=0.699500	lr=0.106443
Epoch[80] Batch [449]	Speed: 2809.029581 samples/sec	accuracy=0.699342	lr=0.106250
Epoch[80] Batch [499]	Speed: 2779.064343 samples/sec	accuracy=0.699287	lr=0.106057
Epoch[80] Batch [549]	Speed: 2736.144648 samples/sec	accuracy=0.699078	lr=0.105865
Epoch[80] Batch [599]	Speed: 2766.456951 samples/sec	accuracy=0.698740	lr=0.105672
Epoch[80] Batch [649]	Speed: 2720.788537 samples/sec	accuracy=0.698101	lr=0.105480
Epoch[80] Batch [699]	Speed: 2761.213064 samples/sec	accuracy=0.697533	lr=0.105287
Epoch[80] Batch [749]	Speed: 2731.941706 samples/sec	accuracy=0.697034	lr=0.105095
Epoch[80] Batch [799]	Speed: 2802.083942 samples/sec	accuracy=0.696620	lr=0.104903
Epoch[80] Batch [849]	Speed: 2754.930932 samples/sec	accuracy=0.696153	lr=0.104711
Epoch[80] Batch [899]	Speed: 2772.385493 samples/sec	accuracy=0.696086	lr=0.104519
Epoch[80] Batch [949]	Speed: 2731.516862 samples/sec	accuracy=0.695689	lr=0.104327
Epoch[80] Batch [999]	Speed: 2728.209539 samples/sec	accuracy=0.695342	lr=0.104135
Epoch[80] Batch [1049]	Speed: 2736.261335 samples/sec	accuracy=0.695188	lr=0.103944
Epoch[80] Batch [1099]	Speed: 2703.682885 samples/sec	accuracy=0.695015	lr=0.103752
Epoch[80] Batch [1149]	Speed: 2712.282047 samples/sec	accuracy=0.694816	lr=0.103561
Epoch[80] Batch [1199]	Speed: 2712.729816 samples/sec	accuracy=0.694648	lr=0.103370
Epoch[80] Batch [1249]	Speed: 2863.146194 samples/sec	accuracy=0.694511	lr=0.103178
[Epoch 80] training: accuracy=0.694531
[Epoch 80] speed: 2744 samples/sec	time cost: 476.485146
[Epoch 80] validation: err-top1=0.275869 err-top5=0.085539
Epoch[81] Batch [49]	Speed: 2708.047914 samples/sec	accuracy=0.709395	lr=0.102984
Epoch[81] Batch [99]	Speed: 2719.289717 samples/sec	accuracy=0.708340	lr=0.102793
Epoch[81] Batch [149]	Speed: 2784.351481 samples/sec	accuracy=0.706120	lr=0.102602
Epoch[81] Batch [199]	Speed: 2743.029609 samples/sec	accuracy=0.705796	lr=0.102411
Epoch[81] Batch [249]	Speed: 2774.255309 samples/sec	accuracy=0.704840	lr=0.102221
Epoch[81] Batch [299]	Speed: 2811.409574 samples/sec	accuracy=0.704004	lr=0.102030
Epoch[81] Batch [349]	Speed: 2703.702321 samples/sec	accuracy=0.703387	lr=0.101840
Epoch[81] Batch [399]	Speed: 2749.130395 samples/sec	accuracy=0.702842	lr=0.101650
Epoch[81] Batch [449]	Speed: 2727.689915 samples/sec	accuracy=0.702582	lr=0.101460
Epoch[81] Batch [499]	Speed: 2635.847293 samples/sec	accuracy=0.702086	lr=0.101270
Epoch[81] Batch [549]	Speed: 2800.860556 samples/sec	accuracy=0.701937	lr=0.101080
Epoch[81] Batch [599]	Speed: 2652.983232 samples/sec	accuracy=0.701642	lr=0.100890
Epoch[81] Batch [649]	Speed: 2609.442521 samples/sec	accuracy=0.701166	lr=0.100701
Epoch[81] Batch [699]	Speed: 2683.431963 samples/sec	accuracy=0.700850	lr=0.100511
Epoch[81] Batch [749]	Speed: 2671.886868 samples/sec	accuracy=0.700676	lr=0.100322
Epoch[81] Batch [799]	Speed: 2724.916578 samples/sec	accuracy=0.700365	lr=0.100132
Epoch[81] Batch [849]	Speed: 2707.749448 samples/sec	accuracy=0.699851	lr=0.099943
Epoch[81] Batch [899]	Speed: 2625.713355 samples/sec	accuracy=0.699480	lr=0.099754
Epoch[81] Batch [949]	Speed: 2759.557114 samples/sec	accuracy=0.699269	lr=0.099565
Epoch[81] Batch [999]	Speed: 2703.416076 samples/sec	accuracy=0.698923	lr=0.099377
Epoch[81] Batch [1049]	Speed: 2668.802680 samples/sec	accuracy=0.698902	lr=0.099188
Epoch[81] Batch [1099]	Speed: 2674.488439 samples/sec	accuracy=0.698624	lr=0.098999
Epoch[81] Batch [1149]	Speed: 2682.103873 samples/sec	accuracy=0.698509	lr=0.098811
Epoch[81] Batch [1199]	Speed: 2714.216368 samples/sec	accuracy=0.698206	lr=0.098623
Epoch[81] Batch [1249]	Speed: 2851.396306 samples/sec	accuracy=0.698002	lr=0.098434
[Epoch 81] training: accuracy=0.697985
[Epoch 81] speed: 2712 samples/sec	time cost: 482.302104
[Epoch 81] validation: err-top1=0.276367 err-top5=0.087850
Epoch[82] Batch [49]	Speed: 2785.251659 samples/sec	accuracy=0.710625	lr=0.098243
Epoch[82] Batch [99]	Speed: 2619.047210 samples/sec	accuracy=0.708242	lr=0.098055
Epoch[82] Batch [149]	Speed: 2763.979453 samples/sec	accuracy=0.706784	lr=0.097867
Epoch[82] Batch [199]	Speed: 2738.857971 samples/sec	accuracy=0.707642	lr=0.097679
Epoch[82] Batch [249]	Speed: 2671.345408 samples/sec	accuracy=0.707301	lr=0.097492
Epoch[82] Batch [299]	Speed: 2681.025363 samples/sec	accuracy=0.707435	lr=0.097304
Epoch[82] Batch [349]	Speed: 2876.824397 samples/sec	accuracy=0.707430	lr=0.097117
Epoch[82] Batch [399]	Speed: 2713.786456 samples/sec	accuracy=0.706580	lr=0.096930
Epoch[82] Batch [449]	Speed: 2723.726192 samples/sec	accuracy=0.706131	lr=0.096743
Epoch[82] Batch [499]	Speed: 2692.529112 samples/sec	accuracy=0.706002	lr=0.096556
Epoch[82] Batch [549]	Speed: 2725.124293 samples/sec	accuracy=0.705511	lr=0.096369
Epoch[82] Batch [599]	Speed: 2758.781342 samples/sec	accuracy=0.705083	lr=0.096182
Epoch[82] Batch [649]	Speed: 2720.408198 samples/sec	accuracy=0.704675	lr=0.095996
Epoch[82] Batch [699]	Speed: 2663.251884 samples/sec	accuracy=0.704371	lr=0.095809
Epoch[82] Batch [749]	Speed: 2785.006541 samples/sec	accuracy=0.704107	lr=0.095623
Epoch[82] Batch [799]	Speed: 2744.632838 samples/sec	accuracy=0.703915	lr=0.095437
Epoch[82] Batch [849]	Speed: 2758.621407 samples/sec	accuracy=0.703591	lr=0.095250
Epoch[82] Batch [899]	Speed: 2707.892783 samples/sec	accuracy=0.703289	lr=0.095064
Epoch[82] Batch [949]	Speed: 2653.374653 samples/sec	accuracy=0.703075	lr=0.094879
Epoch[82] Batch [999]	Speed: 2713.979238 samples/sec	accuracy=0.703045	lr=0.094693
Epoch[82] Batch [1049]	Speed: 2823.088876 samples/sec	accuracy=0.702739	lr=0.094507
Epoch[82] Batch [1099]	Speed: 2657.196944 samples/sec	accuracy=0.702455	lr=0.094322
Epoch[82] Batch [1149]	Speed: 2656.592899 samples/sec	accuracy=0.702369	lr=0.094137
Epoch[82] Batch [1199]	Speed: 2659.485215 samples/sec	accuracy=0.702078	lr=0.093951
Epoch[82] Batch [1249]	Speed: 2832.951806 samples/sec	accuracy=0.701868	lr=0.093766
[Epoch 82] training: accuracy=0.701856
[Epoch 82] speed: 2721 samples/sec	time cost: 481.008181
[Epoch 82] validation: err-top1=0.269691 err-top5=0.082310
Epoch[83] Batch [49]	Speed: 2809.460393 samples/sec	accuracy=0.712871	lr=0.093578
Epoch[83] Batch [99]	Speed: 2747.350623 samples/sec	accuracy=0.713047	lr=0.093393
Epoch[83] Batch [149]	Speed: 2708.202757 samples/sec	accuracy=0.712311	lr=0.093208
Epoch[83] Batch [199]	Speed: 2689.977555 samples/sec	accuracy=0.711660	lr=0.093024
Epoch[83] Batch [249]	Speed: 2834.325750 samples/sec	accuracy=0.711375	lr=0.092839
Epoch[83] Batch [299]	Speed: 2636.071193 samples/sec	accuracy=0.710814	lr=0.092655
Epoch[83] Batch [349]	Speed: 2681.892985 samples/sec	accuracy=0.710025	lr=0.092471
Epoch[83] Batch [399]	Speed: 2769.499034 samples/sec	accuracy=0.709204	lr=0.092287
Epoch[83] Batch [449]	Speed: 2685.520579 samples/sec	accuracy=0.709128	lr=0.092103
Epoch[83] Batch [499]	Speed: 2638.804052 samples/sec	accuracy=0.708738	lr=0.091919
Epoch[83] Batch [549]	Speed: 2808.649445 samples/sec	accuracy=0.708468	lr=0.091735
Epoch[83] Batch [599]	Speed: 2756.769697 samples/sec	accuracy=0.708016	lr=0.091552
Epoch[83] Batch [649]	Speed: 2751.773563 samples/sec	accuracy=0.707719	lr=0.091368
Epoch[83] Batch [699]	Speed: 2764.967000 samples/sec	accuracy=0.707331	lr=0.091185
Epoch[83] Batch [749]	Speed: 2786.846664 samples/sec	accuracy=0.706971	lr=0.091002
Epoch[83] Batch [799]	Speed: 2687.707374 samples/sec	accuracy=0.706792	lr=0.090819
Epoch[83] Batch [849]	Speed: 2738.779413 samples/sec	accuracy=0.706756	lr=0.090636
Epoch[83] Batch [899]	Speed: 2679.241796 samples/sec	accuracy=0.706591	lr=0.090453
Epoch[83] Batch [949]	Speed: 2679.911568 samples/sec	accuracy=0.706228	lr=0.090270
Epoch[83] Batch [999]	Speed: 2705.938581 samples/sec	accuracy=0.706055	lr=0.090088
Epoch[83] Batch [1049]	Speed: 2736.721033 samples/sec	accuracy=0.705976	lr=0.089905
Epoch[83] Batch [1099]	Speed: 2751.806426 samples/sec	accuracy=0.705805	lr=0.089723
Epoch[83] Batch [1149]	Speed: 2721.822143 samples/sec	accuracy=0.705587	lr=0.089541
Epoch[83] Batch [1199]	Speed: 2718.328192 samples/sec	accuracy=0.705349	lr=0.089359
Epoch[83] Batch [1249]	Speed: 2953.912843 samples/sec	accuracy=0.705242	lr=0.089177
[Epoch 83] training: accuracy=0.705244
[Epoch 83] speed: 2734 samples/sec	time cost: 478.431368
[Epoch 83] validation: err-top1=0.269790 err-top5=0.082091
Epoch[84] Batch [49]	Speed: 2776.793100 samples/sec	accuracy=0.717500	lr=0.088992
Epoch[84] Batch [99]	Speed: 2678.339077 samples/sec	accuracy=0.717598	lr=0.088810
Epoch[84] Batch [149]	Speed: 2875.823510 samples/sec	accuracy=0.715872	lr=0.088629
Epoch[84] Batch [199]	Speed: 2768.667547 samples/sec	accuracy=0.714238	lr=0.088448
Epoch[84] Batch [249]	Speed: 2668.454142 samples/sec	accuracy=0.713004	lr=0.088266
Epoch[84] Batch [299]	Speed: 2676.123973 samples/sec	accuracy=0.712796	lr=0.088085
Epoch[84] Batch [349]	Speed: 2653.144068 samples/sec	accuracy=0.712614	lr=0.087904
Epoch[84] Batch [399]	Speed: 2735.296099 samples/sec	accuracy=0.712722	lr=0.087724
Epoch[84] Batch [449]	Speed: 2674.512754 samples/sec	accuracy=0.712478	lr=0.087543
Epoch[84] Batch [499]	Speed: 2654.691666 samples/sec	accuracy=0.712070	lr=0.087363
Epoch[84] Batch [549]	Speed: 2651.375263 samples/sec	accuracy=0.711628	lr=0.087182
Epoch[84] Batch [599]	Speed: 2682.659790 samples/sec	accuracy=0.711522	lr=0.087002
Epoch[84] Batch [649]	Speed: 2648.906594 samples/sec	accuracy=0.711046	lr=0.086822
Epoch[84] Batch [699]	Speed: 2661.870439 samples/sec	accuracy=0.710954	lr=0.086642
Epoch[84] Batch [749]	Speed: 2648.605242 samples/sec	accuracy=0.710668	lr=0.086462
Epoch[84] Batch [799]	Speed: 2653.505666 samples/sec	accuracy=0.710282	lr=0.086282
Epoch[84] Batch [849]	Speed: 2703.685472 samples/sec	accuracy=0.710239	lr=0.086103
Epoch[84] Batch [899]	Speed: 2711.578024 samples/sec	accuracy=0.709925	lr=0.085923
Epoch[84] Batch [949]	Speed: 2740.765545 samples/sec	accuracy=0.709566	lr=0.085744
Epoch[84] Batch [999]	Speed: 2673.775566 samples/sec	accuracy=0.709342	lr=0.085565
Epoch[84] Batch [1049]	Speed: 2710.080781 samples/sec	accuracy=0.709182	lr=0.085386
Epoch[84] Batch [1099]	Speed: 2776.439407 samples/sec	accuracy=0.709012	lr=0.085207
Epoch[84] Batch [1149]	Speed: 2652.550676 samples/sec	accuracy=0.708764	lr=0.085028
Epoch[84] Batch [1199]	Speed: 2689.069137 samples/sec	accuracy=0.708444	lr=0.084850
Epoch[84] Batch [1249]	Speed: 2934.913640 samples/sec	accuracy=0.708284	lr=0.084671
[Epoch 84] training: accuracy=0.708296
[Epoch 84] speed: 2704 samples/sec	time cost: 483.352709
[Epoch 84] validation: err-top1=0.267937 err-top5=0.081314
Epoch[85] Batch [49]	Speed: 2667.856668 samples/sec	accuracy=0.721797	lr=0.084489
Epoch[85] Batch [99]	Speed: 2675.965642 samples/sec	accuracy=0.720332	lr=0.084311
Epoch[85] Batch [149]	Speed: 2728.908945 samples/sec	accuracy=0.720020	lr=0.084133
Epoch[85] Batch [199]	Speed: 2803.752208 samples/sec	accuracy=0.719165	lr=0.083955
Epoch[85] Batch [249]	Speed: 2744.611476 samples/sec	accuracy=0.718160	lr=0.083777
Epoch[85] Batch [299]	Speed: 2785.967284 samples/sec	accuracy=0.718135	lr=0.083600
Epoch[85] Batch [349]	Speed: 2719.843279 samples/sec	accuracy=0.717796	lr=0.083422
Epoch[85] Batch [399]	Speed: 2782.389588 samples/sec	accuracy=0.717163	lr=0.083245
Epoch[85] Batch [449]	Speed: 2707.886193 samples/sec	accuracy=0.716621	lr=0.083067
Epoch[85] Batch [499]	Speed: 2719.334378 samples/sec	accuracy=0.715727	lr=0.082890
Epoch[85] Batch [549]	Speed: 2707.235301 samples/sec	accuracy=0.715309	lr=0.082713
Epoch[85] Batch [599]	Speed: 2700.709109 samples/sec	accuracy=0.714777	lr=0.082537
Epoch[85] Batch [649]	Speed: 2743.303418 samples/sec	accuracy=0.714443	lr=0.082360
Epoch[85] Batch [699]	Speed: 2870.833612 samples/sec	accuracy=0.713949	lr=0.082183
Epoch[85] Batch [749]	Speed: 2715.059473 samples/sec	accuracy=0.713674	lr=0.082007
Epoch[85] Batch [799]	Speed: 2883.336301 samples/sec	accuracy=0.713381	lr=0.081831
Epoch[85] Batch [849]	Speed: 2824.102260 samples/sec	accuracy=0.713225	lr=0.081655
Epoch[85] Batch [899]	Speed: 2783.336765 samples/sec	accuracy=0.712944	lr=0.081479
Epoch[85] Batch [949]	Speed: 2764.254792 samples/sec	accuracy=0.712755	lr=0.081303
Epoch[85] Batch [999]	Speed: 2669.520966 samples/sec	accuracy=0.712583	lr=0.081127
Epoch[85] Batch [1049]	Speed: 2718.876165 samples/sec	accuracy=0.712437	lr=0.080952
Epoch[85] Batch [1099]	Speed: 2714.723184 samples/sec	accuracy=0.712424	lr=0.080776
Epoch[85] Batch [1149]	Speed: 2712.420415 samples/sec	accuracy=0.712247	lr=0.080601
Epoch[85] Batch [1199]	Speed: 2703.755731 samples/sec	accuracy=0.712251	lr=0.080426
Epoch[85] Batch [1249]	Speed: 2849.213886 samples/sec	accuracy=0.711870	lr=0.080251
[Epoch 85] training: accuracy=0.711866
[Epoch 85] speed: 2745 samples/sec	time cost: 476.597501
[Epoch 85] validation: err-top1=0.264728 err-top5=0.080218
Epoch[86] Batch [49]	Speed: 2771.234136 samples/sec	accuracy=0.719805	lr=0.080073
Epoch[86] Batch [99]	Speed: 2858.444286 samples/sec	accuracy=0.721992	lr=0.079898
Epoch[86] Batch [149]	Speed: 2795.422824 samples/sec	accuracy=0.721016	lr=0.079723
Epoch[86] Batch [199]	Speed: 2755.750156 samples/sec	accuracy=0.720425	lr=0.079549
Epoch[86] Batch [249]	Speed: 2769.108990 samples/sec	accuracy=0.719613	lr=0.079375
Epoch[86] Batch [299]	Speed: 2720.505315 samples/sec	accuracy=0.719134	lr=0.079201
Epoch[86] Batch [349]	Speed: 2697.572531 samples/sec	accuracy=0.718825	lr=0.079027
Epoch[86] Batch [399]	Speed: 2634.361039 samples/sec	accuracy=0.719175	lr=0.078853
Epoch[86] Batch [449]	Speed: 2681.130534 samples/sec	accuracy=0.718759	lr=0.078679
Epoch[86] Batch [499]	Speed: 2732.471122 samples/sec	accuracy=0.718320	lr=0.078506
Epoch[86] Batch [549]	Speed: 2723.973667 samples/sec	accuracy=0.718166	lr=0.078332
Epoch[86] Batch [599]	Speed: 2724.329809 samples/sec	accuracy=0.718146	lr=0.078159
Epoch[86] Batch [649]	Speed: 2765.920412 samples/sec	accuracy=0.718194	lr=0.077986
Epoch[86] Batch [699]	Speed: 2719.083682 samples/sec	accuracy=0.717995	lr=0.077813
Epoch[86] Batch [749]	Speed: 2769.611225 samples/sec	accuracy=0.717589	lr=0.077640
Epoch[86] Batch [799]	Speed: 2765.228792 samples/sec	accuracy=0.717357	lr=0.077467
Epoch[86] Batch [849]	Speed: 2640.305161 samples/sec	accuracy=0.717183	lr=0.077295
Epoch[86] Batch [899]	Speed: 2675.943434 samples/sec	accuracy=0.717126	lr=0.077123
Epoch[86] Batch [949]	Speed: 2723.369518 samples/sec	accuracy=0.716892	lr=0.076950
Epoch[86] Batch [999]	Speed: 2692.993042 samples/sec	accuracy=0.716691	lr=0.076778
Epoch[86] Batch [1049]	Speed: 2731.005181 samples/sec	accuracy=0.716576	lr=0.076606
Epoch[86] Batch [1099]	Speed: 2775.056304 samples/sec	accuracy=0.716234	lr=0.076435
Epoch[86] Batch [1149]	Speed: 2663.943062 samples/sec	accuracy=0.716153	lr=0.076263
Epoch[86] Batch [1199]	Speed: 2815.589504 samples/sec	accuracy=0.715977	lr=0.076091
Epoch[86] Batch [1249]	Speed: 2883.655606 samples/sec	accuracy=0.715837	lr=0.075920
[Epoch 86] training: accuracy=0.715864
[Epoch 86] speed: 2736 samples/sec	time cost: 477.632709
[Epoch 86] validation: err-top1=0.261201 err-top5=0.078185
Epoch[87] Batch [49]	Speed: 2911.176710 samples/sec	accuracy=0.728359	lr=0.075745
Epoch[87] Batch [99]	Speed: 2780.963386 samples/sec	accuracy=0.728613	lr=0.075574
Epoch[87] Batch [149]	Speed: 2725.524217 samples/sec	accuracy=0.726126	lr=0.075404
Epoch[87] Batch [199]	Speed: 2751.114585 samples/sec	accuracy=0.726182	lr=0.075233
Epoch[87] Batch [249]	Speed: 2695.343272 samples/sec	accuracy=0.725453	lr=0.075062
Epoch[87] Batch [299]	Speed: 2716.296200 samples/sec	accuracy=0.725120	lr=0.074892
Epoch[87] Batch [349]	Speed: 2630.083961 samples/sec	accuracy=0.724428	lr=0.074721
Epoch[87] Batch [399]	Speed: 2611.564416 samples/sec	accuracy=0.724214	lr=0.074551
Epoch[87] Batch [449]	Speed: 2654.493335 samples/sec	accuracy=0.723987	lr=0.074381
Epoch[87] Batch [499]	Speed: 2853.610397 samples/sec	accuracy=0.723545	lr=0.074211
Epoch[87] Batch [549]	Speed: 2811.002117 samples/sec	accuracy=0.723216	lr=0.074042
Epoch[87] Batch [599]	Speed: 2675.887050 samples/sec	accuracy=0.723092	lr=0.073872
Epoch[87] Batch [649]	Speed: 2708.220380 samples/sec	accuracy=0.722550	lr=0.073703
Epoch[87] Batch [699]	Speed: 2707.827600 samples/sec	accuracy=0.722165	lr=0.073534
Epoch[87] Batch [749]	Speed: 2768.934253 samples/sec	accuracy=0.721570	lr=0.073364
Epoch[87] Batch [799]	Speed: 2786.554259 samples/sec	accuracy=0.721219	lr=0.073196
Epoch[87] Batch [849]	Speed: 2731.249916 samples/sec	accuracy=0.721012	lr=0.073027
Epoch[87] Batch [899]	Speed: 2642.462414 samples/sec	accuracy=0.721159	lr=0.072858
Epoch[87] Batch [949]	Speed: 2761.283611 samples/sec	accuracy=0.720799	lr=0.072690
Epoch[87] Batch [999]	Speed: 2723.571884 samples/sec	accuracy=0.720654	lr=0.072521
Epoch[87] Batch [1049]	Speed: 2763.174421 samples/sec	accuracy=0.720567	lr=0.072353
Epoch[87] Batch [1099]	Speed: 2677.857742 samples/sec	accuracy=0.720289	lr=0.072185
Epoch[87] Batch [1149]	Speed: 2773.753323 samples/sec	accuracy=0.719997	lr=0.072017
Epoch[87] Batch [1199]	Speed: 2714.987355 samples/sec	accuracy=0.719915	lr=0.071849
Epoch[87] Batch [1249]	Speed: 2842.739911 samples/sec	accuracy=0.719712	lr=0.071682
[Epoch 87] training: accuracy=0.719710
[Epoch 87] speed: 2733 samples/sec	time cost: 478.742935
[Epoch 87] validation: err-top1=0.261658 err-top5=0.080139
Epoch[88] Batch [49]	Speed: 2800.662612 samples/sec	accuracy=0.731484	lr=0.071511
Epoch[88] Batch [99]	Speed: 2720.371634 samples/sec	accuracy=0.732822	lr=0.071344
Epoch[88] Batch [149]	Speed: 2744.062654 samples/sec	accuracy=0.731738	lr=0.071177
Epoch[88] Batch [199]	Speed: 2833.153854 samples/sec	accuracy=0.731001	lr=0.071010
Epoch[88] Batch [249]	Speed: 2712.185311 samples/sec	accuracy=0.729863	lr=0.070843
Epoch[88] Batch [299]	Speed: 2749.566545 samples/sec	accuracy=0.729248	lr=0.070676
Epoch[88] Batch [349]	Speed: 2666.734876 samples/sec	accuracy=0.728931	lr=0.070510
Epoch[88] Batch [399]	Speed: 2721.136295 samples/sec	accuracy=0.728945	lr=0.070343
Epoch[88] Batch [449]	Speed: 2660.785060 samples/sec	accuracy=0.728203	lr=0.070177
Epoch[88] Batch [499]	Speed: 2771.973954 samples/sec	accuracy=0.727471	lr=0.070011
Epoch[88] Batch [549]	Speed: 2731.425767 samples/sec	accuracy=0.726999	lr=0.069845
Epoch[88] Batch [599]	Speed: 2727.539939 samples/sec	accuracy=0.726958	lr=0.069680
Epoch[88] Batch [649]	Speed: 2756.803494 samples/sec	accuracy=0.726639	lr=0.069514
Epoch[88] Batch [699]	Speed: 2789.750895 samples/sec	accuracy=0.726470	lr=0.069349
Epoch[88] Batch [749]	Speed: 2736.222427 samples/sec	accuracy=0.726263	lr=0.069183
Epoch[88] Batch [799]	Speed: 2745.877490 samples/sec	accuracy=0.725964	lr=0.069018
Epoch[88] Batch [849]	Speed: 2788.927886 samples/sec	accuracy=0.725471	lr=0.068853
Epoch[88] Batch [899]	Speed: 2761.092890 samples/sec	accuracy=0.725324	lr=0.068689
Epoch[88] Batch [949]	Speed: 2711.828775 samples/sec	accuracy=0.724840	lr=0.068524
Epoch[88] Batch [999]	Speed: 2751.425157 samples/sec	accuracy=0.724374	lr=0.068359
Epoch[88] Batch [1049]	Speed: 2631.182312 samples/sec	accuracy=0.724142	lr=0.068195
Epoch[88] Batch [1099]	Speed: 2806.529636 samples/sec	accuracy=0.723978	lr=0.068031
Epoch[88] Batch [1149]	Speed: 2662.401562 samples/sec	accuracy=0.723842	lr=0.067867
Epoch[88] Batch [1199]	Speed: 2733.089857 samples/sec	accuracy=0.723663	lr=0.067703
Epoch[88] Batch [1249]	Speed: 2864.130748 samples/sec	accuracy=0.723505	lr=0.067539
[Epoch 88] training: accuracy=0.723501
[Epoch 88] speed: 2740 samples/sec	time cost: 477.522246
[Epoch 88] validation: err-top1=0.256876 err-top5=0.075335
Epoch[89] Batch [49]	Speed: 2684.348688 samples/sec	accuracy=0.736602	lr=0.067373
Epoch[89] Batch [99]	Speed: 2744.239737 samples/sec	accuracy=0.734863	lr=0.067209
Epoch[89] Batch [149]	Speed: 2754.014817 samples/sec	accuracy=0.734017	lr=0.067046
Epoch[89] Batch [199]	Speed: 2811.118321 samples/sec	accuracy=0.733110	lr=0.066883
Epoch[89] Batch [249]	Speed: 2709.211713 samples/sec	accuracy=0.732340	lr=0.066720
Epoch[89] Batch [299]	Speed: 2733.826534 samples/sec	accuracy=0.732259	lr=0.066557
Epoch[89] Batch [349]	Speed: 2638.803663 samples/sec	accuracy=0.731655	lr=0.066395
Epoch[89] Batch [399]	Speed: 2686.484237 samples/sec	accuracy=0.731028	lr=0.066232
Epoch[89] Batch [449]	Speed: 2718.285078 samples/sec	accuracy=0.730762	lr=0.066070
Epoch[89] Batch [499]	Speed: 2708.419647 samples/sec	accuracy=0.730451	lr=0.065908
Epoch[89] Batch [549]	Speed: 2704.909267 samples/sec	accuracy=0.729920	lr=0.065746
Epoch[89] Batch [599]	Speed: 2745.912706 samples/sec	accuracy=0.729499	lr=0.065584
Epoch[89] Batch [649]	Speed: 2672.157530 samples/sec	accuracy=0.729416	lr=0.065423
Epoch[89] Batch [699]	Speed: 2691.759726 samples/sec	accuracy=0.729016	lr=0.065261
Epoch[89] Batch [749]	Speed: 2677.290195 samples/sec	accuracy=0.728829	lr=0.065100
Epoch[89] Batch [799]	Speed: 2716.087424 samples/sec	accuracy=0.728621	lr=0.064939
Epoch[89] Batch [849]	Speed: 2745.822263 samples/sec	accuracy=0.728610	lr=0.064778
Epoch[89] Batch [899]	Speed: 2844.809233 samples/sec	accuracy=0.728299	lr=0.064617
Epoch[89] Batch [949]	Speed: 2681.554179 samples/sec	accuracy=0.728168	lr=0.064456
Epoch[89] Batch [999]	Speed: 2736.928912 samples/sec	accuracy=0.727937	lr=0.064296
Epoch[89] Batch [1049]	Speed: 2663.472734 samples/sec	accuracy=0.727871	lr=0.064135
Epoch[89] Batch [1099]	Speed: 2725.125884 samples/sec	accuracy=0.727660	lr=0.063975
Epoch[89] Batch [1149]	Speed: 2812.378788 samples/sec	accuracy=0.727611	lr=0.063815
Epoch[89] Batch [1199]	Speed: 2814.946767 samples/sec	accuracy=0.727495	lr=0.063655
Epoch[89] Batch [1249]	Speed: 2947.025685 samples/sec	accuracy=0.727459	lr=0.063496
[Epoch 89] training: accuracy=0.727469
[Epoch 89] speed: 2731 samples/sec	time cost: 478.652080
[Epoch 89] validation: err-top1=0.258610 err-top5=0.078025
Epoch[90] Batch [49]	Speed: 2758.314098 samples/sec	accuracy=0.736230	lr=0.063333
Epoch[90] Batch [99]	Speed: 2787.291934 samples/sec	accuracy=0.736494	lr=0.063174
Epoch[90] Batch [149]	Speed: 2690.452504 samples/sec	accuracy=0.735957	lr=0.063015
Epoch[90] Batch [199]	Speed: 2674.901325 samples/sec	accuracy=0.735542	lr=0.062855
Epoch[90] Batch [249]	Speed: 2687.916890 samples/sec	accuracy=0.735023	lr=0.062697
Epoch[90] Batch [299]	Speed: 2675.875580 samples/sec	accuracy=0.735046	lr=0.062538
Epoch[90] Batch [349]	Speed: 2775.645327 samples/sec	accuracy=0.734852	lr=0.062379
Epoch[90] Batch [399]	Speed: 2642.081682 samples/sec	accuracy=0.734465	lr=0.062221
Epoch[90] Batch [449]	Speed: 2722.791907 samples/sec	accuracy=0.734668	lr=0.062063
Epoch[90] Batch [499]	Speed: 2777.039359 samples/sec	accuracy=0.734324	lr=0.061905
Epoch[90] Batch [549]	Speed: 2741.283409 samples/sec	accuracy=0.733871	lr=0.061747
Epoch[90] Batch [599]	Speed: 2706.694739 samples/sec	accuracy=0.733784	lr=0.061589
Epoch[90] Batch [649]	Speed: 2655.746559 samples/sec	accuracy=0.733523	lr=0.061432
Epoch[90] Batch [699]	Speed: 2735.049593 samples/sec	accuracy=0.733228	lr=0.061274
Epoch[90] Batch [749]	Speed: 2744.766739 samples/sec	accuracy=0.732914	lr=0.061117
Epoch[90] Batch [799]	Speed: 2725.560780 samples/sec	accuracy=0.732570	lr=0.060960
Epoch[90] Batch [849]	Speed: 2743.050176 samples/sec	accuracy=0.732430	lr=0.060803
Epoch[90] Batch [899]	Speed: 2727.944695 samples/sec	accuracy=0.732332	lr=0.060646
Epoch[90] Batch [949]	Speed: 2722.928346 samples/sec	accuracy=0.732195	lr=0.060490
Epoch[90] Batch [999]	Speed: 2752.373872 samples/sec	accuracy=0.731843	lr=0.060333
Epoch[90] Batch [1049]	Speed: 2654.256748 samples/sec	accuracy=0.731758	lr=0.060177
Epoch[90] Batch [1099]	Speed: 2794.317339 samples/sec	accuracy=0.731545	lr=0.060021
Epoch[90] Batch [1149]	Speed: 2769.972792 samples/sec	accuracy=0.731458	lr=0.059865
Epoch[90] Batch [1199]	Speed: 2726.427390 samples/sec	accuracy=0.731371	lr=0.059710
Epoch[90] Batch [1249]	Speed: 2909.108619 samples/sec	accuracy=0.731193	lr=0.059554
[Epoch 90] training: accuracy=0.731196
[Epoch 90] speed: 2729 samples/sec	time cost: 479.566841
[Epoch 90] validation: err-top1=0.253209 err-top5=0.074936
Epoch[91] Batch [49]	Speed: 2827.186198 samples/sec	accuracy=0.741387	lr=0.059396
Epoch[91] Batch [99]	Speed: 2890.982512 samples/sec	accuracy=0.740869	lr=0.059240
Epoch[91] Batch [149]	Speed: 2718.248262 samples/sec	accuracy=0.740846	lr=0.059085
Epoch[91] Batch [199]	Speed: 2664.567614 samples/sec	accuracy=0.741304	lr=0.058930
Epoch[91] Batch [249]	Speed: 2793.903808 samples/sec	accuracy=0.741230	lr=0.058776
Epoch[91] Batch [299]	Speed: 2687.116479 samples/sec	accuracy=0.740771	lr=0.058621
Epoch[91] Batch [349]	Speed: 2754.106401 samples/sec	accuracy=0.740301	lr=0.058467
Epoch[91] Batch [399]	Speed: 2704.668378 samples/sec	accuracy=0.739888	lr=0.058313
Epoch[91] Batch [449]	Speed: 2717.348744 samples/sec	accuracy=0.739399	lr=0.058158
Epoch[91] Batch [499]	Speed: 2789.473063 samples/sec	accuracy=0.739049	lr=0.058005
Epoch[91] Batch [549]	Speed: 2708.460604 samples/sec	accuracy=0.738789	lr=0.057851
Epoch[91] Batch [599]	Speed: 2752.835577 samples/sec	accuracy=0.738488	lr=0.057697
Epoch[91] Batch [649]	Speed: 2705.644601 samples/sec	accuracy=0.738133	lr=0.057544
Epoch[91] Batch [699]	Speed: 2676.365976 samples/sec	accuracy=0.738128	lr=0.057391
Epoch[91] Batch [749]	Speed: 2712.509493 samples/sec	accuracy=0.737810	lr=0.057238
Epoch[91] Batch [799]	Speed: 2735.179494 samples/sec	accuracy=0.737627	lr=0.057085
Epoch[91] Batch [849]	Speed: 2688.434865 samples/sec	accuracy=0.737405	lr=0.056932
Epoch[91] Batch [899]	Speed: 2825.669283 samples/sec	accuracy=0.736987	lr=0.056780
Epoch[91] Batch [949]	Speed: 2673.072557 samples/sec	accuracy=0.736843	lr=0.056627
Epoch[91] Batch [999]	Speed: 2806.815059 samples/sec	accuracy=0.736527	lr=0.056475
Epoch[91] Batch [1049]	Speed: 2724.467232 samples/sec	accuracy=0.736408	lr=0.056323
Epoch[91] Batch [1099]	Speed: 2657.123298 samples/sec	accuracy=0.736196	lr=0.056171
Epoch[91] Batch [1149]	Speed: 2799.564185 samples/sec	accuracy=0.735928	lr=0.056020
Epoch[91] Batch [1199]	Speed: 2710.791144 samples/sec	accuracy=0.735806	lr=0.055868
Epoch[91] Batch [1249]	Speed: 2902.455006 samples/sec	accuracy=0.735586	lr=0.055717
[Epoch 91] training: accuracy=0.735604
[Epoch 91] speed: 2741 samples/sec	time cost: 476.869048
[Epoch 91] validation: err-top1=0.246193 err-top5=0.070492
Epoch[92] Batch [49]	Speed: 2824.811683 samples/sec	accuracy=0.751016	lr=0.055563
Epoch[92] Batch [99]	Speed: 2717.204578 samples/sec	accuracy=0.749355	lr=0.055412
Epoch[92] Batch [149]	Speed: 2699.463186 samples/sec	accuracy=0.747695	lr=0.055261
Epoch[92] Batch [199]	Speed: 2674.509090 samples/sec	accuracy=0.746558	lr=0.055111
Epoch[92] Batch [249]	Speed: 2772.445266 samples/sec	accuracy=0.745293	lr=0.054960
Epoch[92] Batch [299]	Speed: 2666.695569 samples/sec	accuracy=0.744349	lr=0.054810
Epoch[92] Batch [349]	Speed: 2721.708374 samples/sec	accuracy=0.744180	lr=0.054660
Epoch[92] Batch [399]	Speed: 2703.085115 samples/sec	accuracy=0.743928	lr=0.054510
Epoch[92] Batch [449]	Speed: 2744.184751 samples/sec	accuracy=0.743344	lr=0.054360
Epoch[92] Batch [499]	Speed: 2679.025610 samples/sec	accuracy=0.743234	lr=0.054210
Epoch[92] Batch [549]	Speed: 2772.806355 samples/sec	accuracy=0.742965	lr=0.054061
Epoch[92] Batch [599]	Speed: 2708.673095 samples/sec	accuracy=0.742858	lr=0.053912
Epoch[92] Batch [649]	Speed: 2735.851806 samples/sec	accuracy=0.742509	lr=0.053763
Epoch[92] Batch [699]	Speed: 2747.645193 samples/sec	accuracy=0.742352	lr=0.053614
Epoch[92] Batch [749]	Speed: 2727.974185 samples/sec	accuracy=0.742008	lr=0.053465
Epoch[92] Batch [799]	Speed: 2720.187212 samples/sec	accuracy=0.741776	lr=0.053317
Epoch[92] Batch [849]	Speed: 2703.250824 samples/sec	accuracy=0.741839	lr=0.053168
Epoch[92] Batch [899]	Speed: 2778.984074 samples/sec	accuracy=0.741718	lr=0.053020
Epoch[92] Batch [949]	Speed: 2690.733313 samples/sec	accuracy=0.741349	lr=0.052872
Epoch[92] Batch [999]	Speed: 2831.697702 samples/sec	accuracy=0.741054	lr=0.052724
Epoch[92] Batch [1049]	Speed: 2753.725412 samples/sec	accuracy=0.740765	lr=0.052577
Epoch[92] Batch [1099]	Speed: 2695.947336 samples/sec	accuracy=0.740621	lr=0.052429
Epoch[92] Batch [1149]	Speed: 2649.604173 samples/sec	accuracy=0.740256	lr=0.052282
Epoch[92] Batch [1199]	Speed: 2679.204927 samples/sec	accuracy=0.740128	lr=0.052135
Epoch[92] Batch [1249]	Speed: 2974.226999 samples/sec	accuracy=0.739975	lr=0.051988
[Epoch 92] training: accuracy=0.739975
[Epoch 92] speed: 2731 samples/sec	time cost: 478.773646
[Epoch 92] validation: err-top1=0.245516 err-top5=0.070452
Epoch[93] Batch [49]	Speed: 2826.116447 samples/sec	accuracy=0.750723	lr=0.051838
Epoch[93] Batch [99]	Speed: 2737.706296 samples/sec	accuracy=0.751709	lr=0.051691
Epoch[93] Batch [149]	Speed: 2671.430015 samples/sec	accuracy=0.751178	lr=0.051545
Epoch[93] Batch [199]	Speed: 2705.249604 samples/sec	accuracy=0.750034	lr=0.051399
Epoch[93] Batch [249]	Speed: 2701.263524 samples/sec	accuracy=0.749426	lr=0.051253
Epoch[93] Batch [299]	Speed: 2736.122581 samples/sec	accuracy=0.748477	lr=0.051107
Epoch[93] Batch [349]	Speed: 2761.989848 samples/sec	accuracy=0.748005	lr=0.050961
Epoch[93] Batch [399]	Speed: 2855.354913 samples/sec	accuracy=0.748000	lr=0.050816
Epoch[93] Batch [449]	Speed: 2697.442822 samples/sec	accuracy=0.747376	lr=0.050670
Epoch[93] Batch [499]	Speed: 2680.056486 samples/sec	accuracy=0.746941	lr=0.050525
Epoch[93] Batch [549]	Speed: 2728.686021 samples/sec	accuracy=0.746710	lr=0.050380
Epoch[93] Batch [599]	Speed: 2718.403240 samples/sec	accuracy=0.746201	lr=0.050235
Epoch[93] Batch [649]	Speed: 2712.947432 samples/sec	accuracy=0.745989	lr=0.050091
Epoch[93] Batch [699]	Speed: 2701.494394 samples/sec	accuracy=0.745873	lr=0.049946
Epoch[93] Batch [749]	Speed: 2783.590826 samples/sec	accuracy=0.745624	lr=0.049802
Epoch[93] Batch [799]	Speed: 2677.853535 samples/sec	accuracy=0.745428	lr=0.049658
Epoch[93] Batch [849]	Speed: 2753.890678 samples/sec	accuracy=0.745101	lr=0.049514
Epoch[93] Batch [899]	Speed: 2743.067660 samples/sec	accuracy=0.744889	lr=0.049370
Epoch[93] Batch [949]	Speed: 2703.810981 samples/sec	accuracy=0.744732	lr=0.049227
Epoch[93] Batch [999]	Speed: 2646.068691 samples/sec	accuracy=0.744520	lr=0.049083
Epoch[93] Batch [1049]	Speed: 2686.502150 samples/sec	accuracy=0.744355	lr=0.048940
Epoch[93] Batch [1099]	Speed: 2743.329877 samples/sec	accuracy=0.744419	lr=0.048797
Epoch[93] Batch [1149]	Speed: 2747.099480 samples/sec	accuracy=0.744213	lr=0.048654
Epoch[93] Batch [1199]	Speed: 2705.405966 samples/sec	accuracy=0.744172	lr=0.048511
Epoch[93] Batch [1249]	Speed: 2839.275699 samples/sec	accuracy=0.743992	lr=0.048369
[Epoch 93] training: accuracy=0.744026
[Epoch 93] speed: 2728 samples/sec	time cost: 479.981196
[Epoch 93] validation: err-top1=0.245443 err-top5=0.071493
Epoch[94] Batch [49]	Speed: 2844.484118 samples/sec	accuracy=0.754824	lr=0.048224
Epoch[94] Batch [99]	Speed: 2766.447827 samples/sec	accuracy=0.755020	lr=0.048082
Epoch[94] Batch [149]	Speed: 2749.863633 samples/sec	accuracy=0.754395	lr=0.047940
Epoch[94] Batch [199]	Speed: 2687.551470 samples/sec	accuracy=0.753486	lr=0.047798
Epoch[94] Batch [249]	Speed: 2737.109890 samples/sec	accuracy=0.753641	lr=0.047656
Epoch[94] Batch [299]	Speed: 2788.888552 samples/sec	accuracy=0.754072	lr=0.047515
Epoch[94] Batch [349]	Speed: 2654.168699 samples/sec	accuracy=0.753457	lr=0.047374
Epoch[94] Batch [399]	Speed: 2789.632501 samples/sec	accuracy=0.753262	lr=0.047233
Epoch[94] Batch [449]	Speed: 2661.739127 samples/sec	accuracy=0.752993	lr=0.047092
Epoch[94] Batch [499]	Speed: 2739.240691 samples/sec	accuracy=0.752453	lr=0.046951
Epoch[94] Batch [549]	Speed: 2763.033635 samples/sec	accuracy=0.752147	lr=0.046811
Epoch[94] Batch [599]	Speed: 2799.000975 samples/sec	accuracy=0.752025	lr=0.046670
Epoch[94] Batch [649]	Speed: 2732.035616 samples/sec	accuracy=0.751716	lr=0.046530
Epoch[94] Batch [699]	Speed: 2694.780802 samples/sec	accuracy=0.751076	lr=0.046390
Epoch[94] Batch [749]	Speed: 2651.197622 samples/sec	accuracy=0.750895	lr=0.046251
Epoch[94] Batch [799]	Speed: 2772.612213 samples/sec	accuracy=0.750645	lr=0.046111
Epoch[94] Batch [849]	Speed: 2695.569781 samples/sec	accuracy=0.750438	lr=0.045972
Epoch[94] Batch [899]	Speed: 2692.307265 samples/sec	accuracy=0.750082	lr=0.045833
Epoch[94] Batch [949]	Speed: 2773.898643 samples/sec	accuracy=0.749948	lr=0.045693
Epoch[94] Batch [999]	Speed: 2744.477310 samples/sec	accuracy=0.749507	lr=0.045555
Epoch[94] Batch [1049]	Speed: 2755.345733 samples/sec	accuracy=0.749207	lr=0.045416
Epoch[94] Batch [1099]	Speed: 2729.595490 samples/sec	accuracy=0.748951	lr=0.045278
Epoch[94] Batch [1149]	Speed: 2715.578348 samples/sec	accuracy=0.748791	lr=0.045139
Epoch[94] Batch [1199]	Speed: 2875.500162 samples/sec	accuracy=0.748707	lr=0.045001
Epoch[94] Batch [1249]	Speed: 2937.955822 samples/sec	accuracy=0.748465	lr=0.044863
[Epoch 94] training: accuracy=0.748465
[Epoch 94] speed: 2746 samples/sec	time cost: 476.540098
[Epoch 94] validation: err-top1=0.241430 err-top5=0.069376
Epoch[95] Batch [49]	Speed: 2702.404562 samples/sec	accuracy=0.762422	lr=0.044723
Epoch[95] Batch [99]	Speed: 2746.176345 samples/sec	accuracy=0.760332	lr=0.044585
Epoch[95] Batch [149]	Speed: 2815.480976 samples/sec	accuracy=0.760026	lr=0.044448
Epoch[95] Batch [199]	Speed: 2760.736619 samples/sec	accuracy=0.759365	lr=0.044311
Epoch[95] Batch [249]	Speed: 2728.827767 samples/sec	accuracy=0.759410	lr=0.044174
Epoch[95] Batch [299]	Speed: 2742.777434 samples/sec	accuracy=0.758574	lr=0.044037
Epoch[95] Batch [349]	Speed: 2699.078811 samples/sec	accuracy=0.758192	lr=0.043900
Epoch[95] Batch [399]	Speed: 2690.216980 samples/sec	accuracy=0.757874	lr=0.043764
Epoch[95] Batch [449]	Speed: 2755.676320 samples/sec	accuracy=0.757865	lr=0.043628
Epoch[95] Batch [499]	Speed: 2696.044136 samples/sec	accuracy=0.757484	lr=0.043492
Epoch[95] Batch [549]	Speed: 2741.532510 samples/sec	accuracy=0.757090	lr=0.043356
Epoch[95] Batch [599]	Speed: 2742.854889 samples/sec	accuracy=0.756673	lr=0.043220
Epoch[95] Batch [649]	Speed: 2754.785507 samples/sec	accuracy=0.756367	lr=0.043085
Epoch[95] Batch [699]	Speed: 2752.616419 samples/sec	accuracy=0.756067	lr=0.042949
Epoch[95] Batch [749]	Speed: 2700.597302 samples/sec	accuracy=0.755848	lr=0.042814
Epoch[95] Batch [799]	Speed: 2712.885194 samples/sec	accuracy=0.755605	lr=0.042679
Epoch[95] Batch [849]	Speed: 2836.335522 samples/sec	accuracy=0.755442	lr=0.042545
Epoch[95] Batch [899]	Speed: 2705.997909 samples/sec	accuracy=0.755203	lr=0.042410
Epoch[95] Batch [949]	Speed: 2748.174383 samples/sec	accuracy=0.754879	lr=0.042276
Epoch[95] Batch [999]	Speed: 2720.424809 samples/sec	accuracy=0.754572	lr=0.042141
Epoch[95] Batch [1049]	Speed: 2887.249213 samples/sec	accuracy=0.754186	lr=0.042007
Epoch[95] Batch [1099]	Speed: 2770.181715 samples/sec	accuracy=0.754107	lr=0.041874
Epoch[95] Batch [1149]	Speed: 2618.974992 samples/sec	accuracy=0.754003	lr=0.041740
Epoch[95] Batch [1199]	Speed: 2707.681268 samples/sec	accuracy=0.753949	lr=0.041607
Epoch[95] Batch [1249]	Speed: 2807.870061 samples/sec	accuracy=0.753680	lr=0.041473
[Epoch 95] training: accuracy=0.753670
[Epoch 95] speed: 2739 samples/sec	time cost: 477.748560
[Epoch 95] validation: err-top1=0.238500 err-top5=0.067801
Epoch[96] Batch [49]	Speed: 2632.904060 samples/sec	accuracy=0.764355	lr=0.041338
Epoch[96] Batch [99]	Speed: 2696.726467 samples/sec	accuracy=0.763193	lr=0.041205
Epoch[96] Batch [149]	Speed: 2735.005494 samples/sec	accuracy=0.763314	lr=0.041072
Epoch[96] Batch [199]	Speed: 2683.510328 samples/sec	accuracy=0.762256	lr=0.040940
Epoch[96] Batch [249]	Speed: 2731.643648 samples/sec	accuracy=0.761813	lr=0.040807
Epoch[96] Batch [299]	Speed: 2617.379228 samples/sec	accuracy=0.761650	lr=0.040675
Epoch[96] Batch [349]	Speed: 2702.101320 samples/sec	accuracy=0.761568	lr=0.040543
Epoch[96] Batch [399]	Speed: 2746.314364 samples/sec	accuracy=0.761025	lr=0.040412
Epoch[96] Batch [449]	Speed: 2821.844266 samples/sec	accuracy=0.760803	lr=0.040280
Epoch[96] Batch [499]	Speed: 2780.642150 samples/sec	accuracy=0.760611	lr=0.040149
Epoch[96] Batch [549]	Speed: 2657.628945 samples/sec	accuracy=0.760408	lr=0.040018
Epoch[96] Batch [599]	Speed: 2646.887112 samples/sec	accuracy=0.760073	lr=0.039887
Epoch[96] Batch [649]	Speed: 2782.008771 samples/sec	accuracy=0.759808	lr=0.039756
Epoch[96] Batch [699]	Speed: 2755.646794 samples/sec	accuracy=0.759537	lr=0.039625
Epoch[96] Batch [749]	Speed: 2730.772851 samples/sec	accuracy=0.759247	lr=0.039495
Epoch[96] Batch [799]	Speed: 2737.930870 samples/sec	accuracy=0.758950	lr=0.039365
Epoch[96] Batch [849]	Speed: 2671.083150 samples/sec	accuracy=0.758904	lr=0.039235
Epoch[96] Batch [899]	Speed: 2706.656257 samples/sec	accuracy=0.758754	lr=0.039105
Epoch[96] Batch [949]	Speed: 2909.478081 samples/sec	accuracy=0.758486	lr=0.038975
Epoch[96] Batch [999]	Speed: 2818.931254 samples/sec	accuracy=0.758341	lr=0.038846
Epoch[96] Batch [1049]	Speed: 2797.513200 samples/sec	accuracy=0.758298	lr=0.038717
Epoch[96] Batch [1099]	Speed: 2851.196228 samples/sec	accuracy=0.758063	lr=0.038588
Epoch[96] Batch [1149]	Speed: 2757.553756 samples/sec	accuracy=0.757858	lr=0.038459
Epoch[96] Batch [1199]	Speed: 2687.806409 samples/sec	accuracy=0.757707	lr=0.038330
Epoch[96] Batch [1249]	Speed: 2915.685682 samples/sec	accuracy=0.757484	lr=0.038202
[Epoch 96] training: accuracy=0.757492
[Epoch 96] speed: 2738 samples/sec	time cost: 477.531596
[Epoch 96] validation: err-top1=0.238660 err-top5=0.067562
Epoch[97] Batch [49]	Speed: 2830.047231 samples/sec	accuracy=0.768965	lr=0.038071
Epoch[97] Batch [99]	Speed: 2738.973562 samples/sec	accuracy=0.769229	lr=0.037943
Epoch[97] Batch [149]	Speed: 2715.565642 samples/sec	accuracy=0.768724	lr=0.037815
Epoch[97] Batch [199]	Speed: 2780.973829 samples/sec	accuracy=0.768130	lr=0.037687
Epoch[97] Batch [249]	Speed: 2742.504325 samples/sec	accuracy=0.767699	lr=0.037560
Epoch[97] Batch [299]	Speed: 2684.713238 samples/sec	accuracy=0.767230	lr=0.037432
Epoch[97] Batch [349]	Speed: 2722.447073 samples/sec	accuracy=0.767015	lr=0.037305
Epoch[97] Batch [399]	Speed: 2764.441502 samples/sec	accuracy=0.767017	lr=0.037178
Epoch[97] Batch [449]	Speed: 2697.211628 samples/sec	accuracy=0.766565	lr=0.037052
Epoch[97] Batch [499]	Speed: 2765.255817 samples/sec	accuracy=0.766002	lr=0.036925
Epoch[97] Batch [549]	Speed: 2683.938650 samples/sec	accuracy=0.765842	lr=0.036799
Epoch[97] Batch [599]	Speed: 2697.503338 samples/sec	accuracy=0.765498	lr=0.036673
Epoch[97] Batch [649]	Speed: 2679.659395 samples/sec	accuracy=0.765300	lr=0.036547
Epoch[97] Batch [699]	Speed: 2753.686924 samples/sec	accuracy=0.765174	lr=0.036421
Epoch[97] Batch [749]	Speed: 2665.237577 samples/sec	accuracy=0.765049	lr=0.036296
Epoch[97] Batch [799]	Speed: 2745.920079 samples/sec	accuracy=0.764849	lr=0.036170
Epoch[97] Batch [849]	Speed: 2794.528860 samples/sec	accuracy=0.764660	lr=0.036045
Epoch[97] Batch [899]	Speed: 2696.848621 samples/sec	accuracy=0.764453	lr=0.035920
Epoch[97] Batch [949]	Speed: 2704.239271 samples/sec	accuracy=0.764277	lr=0.035795
Epoch[97] Batch [999]	Speed: 2728.492011 samples/sec	accuracy=0.764172	lr=0.035671
Epoch[97] Batch [1049]	Speed: 2782.846129 samples/sec	accuracy=0.763863	lr=0.035546
Epoch[97] Batch [1099]	Speed: 2681.219276 samples/sec	accuracy=0.763622	lr=0.035422
Epoch[97] Batch [1149]	Speed: 2730.732640 samples/sec	accuracy=0.763370	lr=0.035298
Epoch[97] Batch [1199]	Speed: 2742.362135 samples/sec	accuracy=0.763228	lr=0.035174
Epoch[97] Batch [1249]	Speed: 2864.585966 samples/sec	accuracy=0.763001	lr=0.035051
[Epoch 97] training: accuracy=0.763018
[Epoch 97] speed: 2733 samples/sec	time cost: 478.581484
[Epoch 97] validation: err-top1=0.236567 err-top5=0.065569
Epoch[98] Batch [49]	Speed: 2659.786017 samples/sec	accuracy=0.773574	lr=0.034925
Epoch[98] Batch [99]	Speed: 2701.488718 samples/sec	accuracy=0.774404	lr=0.034802
Epoch[98] Batch [149]	Speed: 2755.180716 samples/sec	accuracy=0.773854	lr=0.034679
Epoch[98] Batch [199]	Speed: 2697.486701 samples/sec	accuracy=0.772554	lr=0.034556
Epoch[98] Batch [249]	Speed: 2767.593747 samples/sec	accuracy=0.772016	lr=0.034433
Epoch[98] Batch [299]	Speed: 2708.103033 samples/sec	accuracy=0.771416	lr=0.034311
Epoch[98] Batch [349]	Speed: 2739.365435 samples/sec	accuracy=0.770798	lr=0.034189
Epoch[98] Batch [399]	Speed: 2764.206757 samples/sec	accuracy=0.770479	lr=0.034067
Epoch[98] Batch [449]	Speed: 2735.473516 samples/sec	accuracy=0.770618	lr=0.033945
Epoch[98] Batch [499]	Speed: 2756.133050 samples/sec	accuracy=0.770119	lr=0.033823
Epoch[98] Batch [549]	Speed: 2629.759214 samples/sec	accuracy=0.769656	lr=0.033702
Epoch[98] Batch [599]	Speed: 2824.353268 samples/sec	accuracy=0.769546	lr=0.033581
Epoch[98] Batch [649]	Speed: 2728.045851 samples/sec	accuracy=0.769372	lr=0.033460
Epoch[98] Batch [699]	Speed: 2768.827150 samples/sec	accuracy=0.769012	lr=0.033339
Epoch[98] Batch [749]	Speed: 2694.125581 samples/sec	accuracy=0.768901	lr=0.033218
Epoch[98] Batch [799]	Speed: 2674.448470 samples/sec	accuracy=0.768678	lr=0.033098
Epoch[98] Batch [849]	Speed: 2763.783273 samples/sec	accuracy=0.768547	lr=0.032978
Epoch[98] Batch [899]	Speed: 2747.539802 samples/sec	accuracy=0.768213	lr=0.032858
Epoch[98] Batch [949]	Speed: 2710.923782 samples/sec	accuracy=0.768025	lr=0.032738
Epoch[98] Batch [999]	Speed: 2703.659330 samples/sec	accuracy=0.767955	lr=0.032618
Epoch[98] Batch [1049]	Speed: 2729.855276 samples/sec	accuracy=0.767902	lr=0.032499
Epoch[98] Batch [1099]	Speed: 2811.456833 samples/sec	accuracy=0.767884	lr=0.032380
Epoch[98] Batch [1149]	Speed: 2765.431623 samples/sec	accuracy=0.767742	lr=0.032261
Epoch[98] Batch [1199]	Speed: 2792.001323 samples/sec	accuracy=0.767616	lr=0.032142
Epoch[98] Batch [1249]	Speed: 2885.309089 samples/sec	accuracy=0.767373	lr=0.032023
[Epoch 98] training: accuracy=0.767395
[Epoch 98] speed: 2737 samples/sec	time cost: 477.249060
[Epoch 98] validation: err-top1=0.233988 err-top5=0.064575
Epoch[99] Batch [49]	Speed: 2850.428619 samples/sec	accuracy=0.780098	lr=0.031902
Epoch[99] Batch [99]	Speed: 2817.053004 samples/sec	accuracy=0.779160	lr=0.031784
Epoch[99] Batch [149]	Speed: 2797.305344 samples/sec	accuracy=0.779349	lr=0.031666
Epoch[99] Batch [199]	Speed: 2649.207981 samples/sec	accuracy=0.778726	lr=0.031548
Epoch[99] Batch [249]	Speed: 2679.542838 samples/sec	accuracy=0.778172	lr=0.031431
Epoch[99] Batch [299]	Speed: 2894.608813 samples/sec	accuracy=0.777386	lr=0.031313
Epoch[99] Batch [349]	Speed: 2762.997588 samples/sec	accuracy=0.777148	lr=0.031196
Epoch[99] Batch [399]	Speed: 2675.318371 samples/sec	accuracy=0.776384	lr=0.031079
Epoch[99] Batch [449]	Speed: 2978.072549 samples/sec	accuracy=0.776009	lr=0.030962
Epoch[99] Batch [499]	Speed: 2725.256435 samples/sec	accuracy=0.775516	lr=0.030846
Epoch[99] Batch [549]	Speed: 2794.997215 samples/sec	accuracy=0.775140	lr=0.030729
Epoch[99] Batch [599]	Speed: 2674.810768 samples/sec	accuracy=0.774842	lr=0.030613
Epoch[99] Batch [649]	Speed: 2749.475580 samples/sec	accuracy=0.774698	lr=0.030497
Epoch[99] Batch [699]	Speed: 2788.968634 samples/sec	accuracy=0.774669	lr=0.030381
Epoch[99] Batch [749]	Speed: 2767.193471 samples/sec	accuracy=0.774535	lr=0.030266
Epoch[99] Batch [799]	Speed: 2787.569947 samples/sec	accuracy=0.774109	lr=0.030150
Epoch[99] Batch [849]	Speed: 2809.808285 samples/sec	accuracy=0.773858	lr=0.030035
Epoch[99] Batch [899]	Speed: 2721.577163 samples/sec	accuracy=0.773560	lr=0.029920
Epoch[99] Batch [949]	Speed: 2703.521446 samples/sec	accuracy=0.773412	lr=0.029805
Epoch[99] Batch [999]	Speed: 2714.168857 samples/sec	accuracy=0.773208	lr=0.029691
Epoch[99] Batch [1049]	Speed: 2766.524344 samples/sec	accuracy=0.773151	lr=0.029576
Epoch[99] Batch [1099]	Speed: 2719.794158 samples/sec	accuracy=0.772947	lr=0.029462
Epoch[99] Batch [1149]	Speed: 2645.148500 samples/sec	accuracy=0.772950	lr=0.029348
Epoch[99] Batch [1199]	Speed: 2759.524525 samples/sec	accuracy=0.772605	lr=0.029234
Epoch[99] Batch [1249]	Speed: 2866.368800 samples/sec	accuracy=0.772359	lr=0.029121
[Epoch 99] training: accuracy=0.772345
[Epoch 99] speed: 2759 samples/sec	time cost: 474.371100
[Epoch 99] validation: err-top1=0.227798 err-top5=0.062759
Epoch[100] Batch [49]	Speed: 2694.316086 samples/sec	accuracy=0.781660	lr=0.029005
Epoch[100] Batch [99]	Speed: 2658.634136 samples/sec	accuracy=0.782803	lr=0.028892
Epoch[100] Batch [149]	Speed: 2744.942123 samples/sec	accuracy=0.783040	lr=0.028779
Epoch[100] Batch [199]	Speed: 2648.661006 samples/sec	accuracy=0.781553	lr=0.028666
Epoch[100] Batch [249]	Speed: 2659.412298 samples/sec	accuracy=0.781379	lr=0.028554
Epoch[100] Batch [299]	Speed: 2787.434154 samples/sec	accuracy=0.780872	lr=0.028441
Epoch[100] Batch [349]	Speed: 2699.742756 samples/sec	accuracy=0.780622	lr=0.028329
Epoch[100] Batch [399]	Speed: 2757.742252 samples/sec	accuracy=0.780449	lr=0.028217
Epoch[100] Batch [449]	Speed: 2759.831537 samples/sec	accuracy=0.779976	lr=0.028105
Epoch[100] Batch [499]	Speed: 2756.171147 samples/sec	accuracy=0.779854	lr=0.027994
Epoch[100] Batch [549]	Speed: 2662.027140 samples/sec	accuracy=0.779739	lr=0.027883
Epoch[100] Batch [599]	Speed: 2758.516376 samples/sec	accuracy=0.779455	lr=0.027772
Epoch[100] Batch [649]	Speed: 2647.196263 samples/sec	accuracy=0.779363	lr=0.027661
Epoch[100] Batch [699]	Speed: 2689.869734 samples/sec	accuracy=0.779215	lr=0.027550
Epoch[100] Batch [749]	Speed: 2691.062403 samples/sec	accuracy=0.779017	lr=0.027439
Epoch[100] Batch [799]	Speed: 2735.957802 samples/sec	accuracy=0.778921	lr=0.027329
Epoch[100] Batch [849]	Speed: 2775.036617 samples/sec	accuracy=0.778935	lr=0.027219
Epoch[100] Batch [899]	Speed: 2719.942560 samples/sec	accuracy=0.778834	lr=0.027109
Epoch[100] Batch [949]	Speed: 2683.770672 samples/sec	accuracy=0.778610	lr=0.026999
Epoch[100] Batch [999]	Speed: 2701.854811 samples/sec	accuracy=0.778405	lr=0.026890
Epoch[100] Batch [1049]	Speed: 2718.586285 samples/sec	accuracy=0.778285	lr=0.026781
Epoch[100] Batch [1099]	Speed: 2757.052308 samples/sec	accuracy=0.778207	lr=0.026672
Epoch[100] Batch [1149]	Speed: 2663.959684 samples/sec	accuracy=0.778168	lr=0.026563
Epoch[100] Batch [1199]	Speed: 2783.441024 samples/sec	accuracy=0.778052	lr=0.026454
Epoch[100] Batch [1249]	Speed: 2829.474822 samples/sec	accuracy=0.777824	lr=0.026346
[Epoch 100] training: accuracy=0.777828
[Epoch 100] speed: 2716 samples/sec	time cost: 481.769071
[Epoch 100] validation: err-top1=0.227918 err-top5=0.062301
Epoch[101] Batch [49]	Speed: 2871.000568 samples/sec	accuracy=0.789492	lr=0.026235
Epoch[101] Batch [99]	Speed: 2679.856889 samples/sec	accuracy=0.787471	lr=0.026127
Epoch[101] Batch [149]	Speed: 2730.725348 samples/sec	accuracy=0.786927	lr=0.026020
Epoch[101] Batch [199]	Speed: 2703.252117 samples/sec	accuracy=0.786455	lr=0.025912
Epoch[101] Batch [249]	Speed: 2660.304081 samples/sec	accuracy=0.785680	lr=0.025805
Epoch[101] Batch [299]	Speed: 2679.861638 samples/sec	accuracy=0.785404	lr=0.025697
Epoch[101] Batch [349]	Speed: 2662.853780 samples/sec	accuracy=0.785329	lr=0.025590
Epoch[101] Batch [399]	Speed: 2713.881146 samples/sec	accuracy=0.785051	lr=0.025484
Epoch[101] Batch [449]	Speed: 2664.730585 samples/sec	accuracy=0.784935	lr=0.025377
Epoch[101] Batch [499]	Speed: 2697.418935 samples/sec	accuracy=0.784832	lr=0.025271
Epoch[101] Batch [549]	Speed: 2692.835274 samples/sec	accuracy=0.784590	lr=0.025165
Epoch[101] Batch [599]	Speed: 2817.054223 samples/sec	accuracy=0.784756	lr=0.025059
Epoch[101] Batch [649]	Speed: 2721.764809 samples/sec	accuracy=0.784473	lr=0.024953
Epoch[101] Batch [699]	Speed: 2661.662754 samples/sec	accuracy=0.784335	lr=0.024847
Epoch[101] Batch [749]	Speed: 2676.815577 samples/sec	accuracy=0.784241	lr=0.024742
Epoch[101] Batch [799]	Speed: 2716.325061 samples/sec	accuracy=0.784264	lr=0.024637
Epoch[101] Batch [849]	Speed: 2738.452029 samples/sec	accuracy=0.784303	lr=0.024532
Epoch[101] Batch [899]	Speed: 2680.405284 samples/sec	accuracy=0.783989	lr=0.024427
Epoch[101] Batch [949]	Speed: 2674.366603 samples/sec	accuracy=0.783802	lr=0.024323
Epoch[101] Batch [999]	Speed: 2823.316133 samples/sec	accuracy=0.783647	lr=0.024219
Epoch[101] Batch [1049]	Speed: 2688.368967 samples/sec	accuracy=0.783645	lr=0.024115
Epoch[101] Batch [1099]	Speed: 2737.698967 samples/sec	accuracy=0.783525	lr=0.024011
Epoch[101] Batch [1149]	Speed: 2684.820108 samples/sec	accuracy=0.783296	lr=0.023907
Epoch[101] Batch [1199]	Speed: 2728.049317 samples/sec	accuracy=0.783163	lr=0.023804
Epoch[101] Batch [1249]	Speed: 2828.457207 samples/sec	accuracy=0.783066	lr=0.023700
[Epoch 101] training: accuracy=0.783078
[Epoch 101] speed: 2714 samples/sec	time cost: 481.834150
[Epoch 101] validation: err-top1=0.225028 err-top5=0.062241
Epoch[102] Batch [49]	Speed: 2621.106250 samples/sec	accuracy=0.792578	lr=0.023595
Epoch[102] Batch [99]	Speed: 2772.033673 samples/sec	accuracy=0.793330	lr=0.023493
Epoch[102] Batch [149]	Speed: 2741.414603 samples/sec	accuracy=0.792943	lr=0.023390
Epoch[102] Batch [199]	Speed: 2817.199755 samples/sec	accuracy=0.792559	lr=0.023288
Epoch[102] Batch [249]	Speed: 2815.308346 samples/sec	accuracy=0.791836	lr=0.023185
Epoch[102] Batch [299]	Speed: 2724.449396 samples/sec	accuracy=0.791243	lr=0.023083
Epoch[102] Batch [349]	Speed: 2742.742929 samples/sec	accuracy=0.790594	lr=0.022982
Epoch[102] Batch [399]	Speed: 2710.093093 samples/sec	accuracy=0.790605	lr=0.022880
Epoch[102] Batch [449]	Speed: 2727.441037 samples/sec	accuracy=0.790423	lr=0.022779
Epoch[102] Batch [499]	Speed: 2812.108951 samples/sec	accuracy=0.790350	lr=0.022678
Epoch[102] Batch [549]	Speed: 2656.439235 samples/sec	accuracy=0.790002	lr=0.022577
Epoch[102] Batch [599]	Speed: 2704.805697 samples/sec	accuracy=0.789883	lr=0.022476
Epoch[102] Batch [649]	Speed: 2702.260651 samples/sec	accuracy=0.790005	lr=0.022376
Epoch[102] Batch [699]	Speed: 2690.141930 samples/sec	accuracy=0.789576	lr=0.022276
Epoch[102] Batch [749]	Speed: 2796.169062 samples/sec	accuracy=0.789371	lr=0.022175
Epoch[102] Batch [799]	Speed: 2729.154796 samples/sec	accuracy=0.789165	lr=0.022076
Epoch[102] Batch [849]	Speed: 2705.344755 samples/sec	accuracy=0.788998	lr=0.021976
Epoch[102] Batch [899]	Speed: 2717.556132 samples/sec	accuracy=0.789012	lr=0.021877
Epoch[102] Batch [949]	Speed: 2693.195614 samples/sec	accuracy=0.788926	lr=0.021777
Epoch[102] Batch [999]	Speed: 2706.146242 samples/sec	accuracy=0.788750	lr=0.021678
Epoch[102] Batch [1049]	Speed: 2697.205768 samples/sec	accuracy=0.788557	lr=0.021580
Epoch[102] Batch [1099]	Speed: 2777.350748 samples/sec	accuracy=0.788581	lr=0.021481
Epoch[102] Batch [1149]	Speed: 2681.526220 samples/sec	accuracy=0.788477	lr=0.021383
Epoch[102] Batch [1199]	Speed: 2775.466857 samples/sec	accuracy=0.788415	lr=0.021285
Epoch[102] Batch [1249]	Speed: 2874.714065 samples/sec	accuracy=0.788397	lr=0.021187
[Epoch 102] training: accuracy=0.788405
[Epoch 102] speed: 2732 samples/sec	time cost: 478.278833
[Epoch 102] validation: err-top1=0.225785 err-top5=0.060427
Epoch[103] Batch [49]	Speed: 2760.855271 samples/sec	accuracy=0.797305	lr=0.021087
Epoch[103] Batch [99]	Speed: 2671.017369 samples/sec	accuracy=0.796650	lr=0.020990
Epoch[103] Batch [149]	Speed: 2682.546792 samples/sec	accuracy=0.796504	lr=0.020892
Epoch[103] Batch [199]	Speed: 2744.351679 samples/sec	accuracy=0.796226	lr=0.020795
Epoch[103] Batch [249]	Speed: 2707.243287 samples/sec	accuracy=0.796117	lr=0.020698
Epoch[103] Batch [299]	Speed: 2683.392866 samples/sec	accuracy=0.795316	lr=0.020602
Epoch[103] Batch [349]	Speed: 2833.550074 samples/sec	accuracy=0.795552	lr=0.020505
Epoch[103] Batch [399]	Speed: 2744.432766 samples/sec	accuracy=0.795315	lr=0.020409
Epoch[103] Batch [449]	Speed: 2697.172366 samples/sec	accuracy=0.794933	lr=0.020313
Epoch[103] Batch [499]	Speed: 2617.346594 samples/sec	accuracy=0.795207	lr=0.020217
Epoch[103] Batch [549]	Speed: 2698.576768 samples/sec	accuracy=0.795004	lr=0.020122
Epoch[103] Batch [599]	Speed: 2681.743648 samples/sec	accuracy=0.794958	lr=0.020026
Epoch[103] Batch [649]	Speed: 2678.113785 samples/sec	accuracy=0.794769	lr=0.019931
Epoch[103] Batch [699]	Speed: 2773.436759 samples/sec	accuracy=0.794549	lr=0.019836
Epoch[103] Batch [749]	Speed: 2677.457096 samples/sec	accuracy=0.794699	lr=0.019742
Epoch[103] Batch [799]	Speed: 2785.706863 samples/sec	accuracy=0.794705	lr=0.019647
Epoch[103] Batch [849]	Speed: 2740.453808 samples/sec	accuracy=0.794467	lr=0.019553
Epoch[103] Batch [899]	Speed: 2657.013953 samples/sec	accuracy=0.794362	lr=0.019459
Epoch[103] Batch [949]	Speed: 2716.416870 samples/sec	accuracy=0.794305	lr=0.019365
Epoch[103] Batch [999]	Speed: 2676.494265 samples/sec	accuracy=0.794034	lr=0.019271
Epoch[103] Batch [1049]	Speed: 2752.647892 samples/sec	accuracy=0.793971	lr=0.019178
Epoch[103] Batch [1099]	Speed: 2644.887418 samples/sec	accuracy=0.793839	lr=0.019085
Epoch[103] Batch [1149]	Speed: 2700.472771 samples/sec	accuracy=0.793749	lr=0.018992
Epoch[103] Batch [1199]	Speed: 2685.492840 samples/sec	accuracy=0.793684	lr=0.018899
Epoch[103] Batch [1249]	Speed: 2820.032529 samples/sec	accuracy=0.793642	lr=0.018806
[Epoch 103] training: accuracy=0.793635
[Epoch 103] speed: 2710 samples/sec	time cost: 482.362473
[Epoch 103] validation: err-top1=0.221640 err-top5=0.059032
Epoch[104] Batch [49]	Speed: 2814.117788 samples/sec	accuracy=0.799160	lr=0.018712
Epoch[104] Batch [99]	Speed: 2714.453197 samples/sec	accuracy=0.802129	lr=0.018620
Epoch[104] Batch [149]	Speed: 2666.143434 samples/sec	accuracy=0.802721	lr=0.018528
Epoch[104] Batch [199]	Speed: 2829.370813 samples/sec	accuracy=0.801963	lr=0.018436
Epoch[104] Batch [249]	Speed: 2764.348233 samples/sec	accuracy=0.801961	lr=0.018345
Epoch[104] Batch [299]	Speed: 2818.029960 samples/sec	accuracy=0.802116	lr=0.018254
Epoch[104] Batch [349]	Speed: 2807.026312 samples/sec	accuracy=0.802360	lr=0.018163
Epoch[104] Batch [399]	Speed: 2819.864932 samples/sec	accuracy=0.802378	lr=0.018072
Epoch[104] Batch [449]	Speed: 2746.266319 samples/sec	accuracy=0.801916	lr=0.017981
Epoch[104] Batch [499]	Speed: 2762.119159 samples/sec	accuracy=0.801699	lr=0.017891
Epoch[104] Batch [549]	Speed: 2718.643175 samples/sec	accuracy=0.801433	lr=0.017801
Epoch[104] Batch [599]	Speed: 2760.139078 samples/sec	accuracy=0.801326	lr=0.017711
Epoch[104] Batch [649]	Speed: 2766.364116 samples/sec	accuracy=0.801041	lr=0.017621
Epoch[104] Batch [699]	Speed: 2585.498696 samples/sec	accuracy=0.800854	lr=0.017532
Epoch[104] Batch [749]	Speed: 2691.165665 samples/sec	accuracy=0.800460	lr=0.017442
Epoch[104] Batch [799]	Speed: 2692.919964 samples/sec	accuracy=0.800254	lr=0.017353
Epoch[104] Batch [849]	Speed: 2724.538679 samples/sec	accuracy=0.800140	lr=0.017264
Epoch[104] Batch [899]	Speed: 2654.024796 samples/sec	accuracy=0.799991	lr=0.017176
Epoch[104] Batch [949]	Speed: 2638.345249 samples/sec	accuracy=0.799876	lr=0.017087
Epoch[104] Batch [999]	Speed: 2668.053620 samples/sec	accuracy=0.799672	lr=0.016999
Epoch[104] Batch [1049]	Speed: 2744.866936 samples/sec	accuracy=0.799443	lr=0.016911
Epoch[104] Batch [1099]	Speed: 2839.898086 samples/sec	accuracy=0.799306	lr=0.016823
Epoch[104] Batch [1149]	Speed: 2653.591113 samples/sec	accuracy=0.799249	lr=0.016736
Epoch[104] Batch [1199]	Speed: 2773.589856 samples/sec	accuracy=0.799150	lr=0.016648
Epoch[104] Batch [1249]	Speed: 2909.879220 samples/sec	accuracy=0.799018	lr=0.016561
[Epoch 104] training: accuracy=0.799012
[Epoch 104] speed: 2738 samples/sec	time cost: 477.558076
[Epoch 104] validation: err-top1=0.218689 err-top5=0.058411
Epoch[105] Batch [49]	Speed: 2725.672968 samples/sec	accuracy=0.810840	lr=0.016473
Epoch[105] Batch [99]	Speed: 2872.478567 samples/sec	accuracy=0.809775	lr=0.016386
Epoch[105] Batch [149]	Speed: 2682.901702 samples/sec	accuracy=0.807910	lr=0.016299
Epoch[105] Batch [199]	Speed: 2641.404790 samples/sec	accuracy=0.807378	lr=0.016213
Epoch[105] Batch [249]	Speed: 2690.773029 samples/sec	accuracy=0.807109	lr=0.016127
Epoch[105] Batch [299]	Speed: 2641.835765 samples/sec	accuracy=0.806559	lr=0.016041
Epoch[105] Batch [349]	Speed: 2791.231766 samples/sec	accuracy=0.806281	lr=0.015956
Epoch[105] Batch [399]	Speed: 2699.932088 samples/sec	accuracy=0.805920	lr=0.015870
Epoch[105] Batch [449]	Speed: 2695.459381 samples/sec	accuracy=0.805979	lr=0.015785
Epoch[105] Batch [499]	Speed: 2702.879488 samples/sec	accuracy=0.805820	lr=0.015700
Epoch[105] Batch [549]	Speed: 2725.068445 samples/sec	accuracy=0.805666	lr=0.015616
Epoch[105] Batch [599]	Speed: 2641.185408 samples/sec	accuracy=0.805422	lr=0.015531
Epoch[105] Batch [649]	Speed: 2664.880645 samples/sec	accuracy=0.805225	lr=0.015447
Epoch[105] Batch [699]	Speed: 2790.854619 samples/sec	accuracy=0.805177	lr=0.015363
Epoch[105] Batch [749]	Speed: 2720.498973 samples/sec	accuracy=0.805217	lr=0.015279
Epoch[105] Batch [799]	Speed: 2810.804577 samples/sec	accuracy=0.805156	lr=0.015195
Epoch[105] Batch [849]	Speed: 2857.742324 samples/sec	accuracy=0.805067	lr=0.015112
Epoch[105] Batch [899]	Speed: 2828.484551 samples/sec	accuracy=0.805132	lr=0.015029
Epoch[105] Batch [949]	Speed: 2665.685366 samples/sec	accuracy=0.804949	lr=0.014946
Epoch[105] Batch [999]	Speed: 2669.934710 samples/sec	accuracy=0.804756	lr=0.014863
Epoch[105] Batch [1049]	Speed: 2685.898381 samples/sec	accuracy=0.804668	lr=0.014781
Epoch[105] Batch [1099]	Speed: 2758.695755 samples/sec	accuracy=0.804548	lr=0.014699
Epoch[105] Batch [1149]	Speed: 2744.045297 samples/sec	accuracy=0.804381	lr=0.014616
Epoch[105] Batch [1199]	Speed: 2670.990460 samples/sec	accuracy=0.804312	lr=0.014535
Epoch[105] Batch [1249]	Speed: 2843.923217 samples/sec	accuracy=0.804252	lr=0.014453
[Epoch 105] training: accuracy=0.804261
[Epoch 105] speed: 2725 samples/sec	time cost: 479.889403
[Epoch 105] validation: err-top1=0.217973 err-top5=0.057478
Epoch[106] Batch [49]	Speed: 2742.157387 samples/sec	accuracy=0.813984	lr=0.014370
Epoch[106] Batch [99]	Speed: 2707.492247 samples/sec	accuracy=0.814834	lr=0.014289
Epoch[106] Batch [149]	Speed: 2671.615031 samples/sec	accuracy=0.813066	lr=0.014208
Epoch[106] Batch [199]	Speed: 2775.712272 samples/sec	accuracy=0.812573	lr=0.014127
Epoch[106] Batch [249]	Speed: 2712.168389 samples/sec	accuracy=0.812961	lr=0.014047
Epoch[106] Batch [299]	Speed: 2709.784157 samples/sec	accuracy=0.812526	lr=0.013966
Epoch[106] Batch [349]	Speed: 2755.843907 samples/sec	accuracy=0.812400	lr=0.013886
Epoch[106] Batch [399]	Speed: 2819.779363 samples/sec	accuracy=0.812065	lr=0.013806
Epoch[106] Batch [449]	Speed: 2755.109102 samples/sec	accuracy=0.811981	lr=0.013727
Epoch[106] Batch [499]	Speed: 2665.353290 samples/sec	accuracy=0.811676	lr=0.013647
Epoch[106] Batch [549]	Speed: 2752.885335 samples/sec	accuracy=0.811364	lr=0.013568
Epoch[106] Batch [599]	Speed: 2813.496917 samples/sec	accuracy=0.811029	lr=0.013489
Epoch[106] Batch [649]	Speed: 2663.127338 samples/sec	accuracy=0.810992	lr=0.013411
Epoch[106] Batch [699]	Speed: 2833.504574 samples/sec	accuracy=0.810723	lr=0.013332
Epoch[106] Batch [749]	Speed: 2761.772036 samples/sec	accuracy=0.810652	lr=0.013254
Epoch[106] Batch [799]	Speed: 2801.517708 samples/sec	accuracy=0.810515	lr=0.013176
Epoch[106] Batch [849]	Speed: 2682.098781 samples/sec	accuracy=0.810516	lr=0.013098
Epoch[106] Batch [899]	Speed: 2672.800178 samples/sec	accuracy=0.810489	lr=0.013020
Epoch[106] Batch [949]	Speed: 2691.151804 samples/sec	accuracy=0.810301	lr=0.012943
Epoch[106] Batch [999]	Speed: 2779.646254 samples/sec	accuracy=0.810263	lr=0.012866
Epoch[106] Batch [1049]	Speed: 2776.679572 samples/sec	accuracy=0.810264	lr=0.012789
Epoch[106] Batch [1099]	Speed: 2737.185246 samples/sec	accuracy=0.810150	lr=0.012712
Epoch[106] Batch [1149]	Speed: 2756.957591 samples/sec	accuracy=0.809997	lr=0.012636
Epoch[106] Batch [1199]	Speed: 2644.777547 samples/sec	accuracy=0.810067	lr=0.012559
Epoch[106] Batch [1249]	Speed: 2841.105977 samples/sec	accuracy=0.810052	lr=0.012483
[Epoch 106] training: accuracy=0.810033
[Epoch 106] speed: 2738 samples/sec	time cost: 477.733985
[Epoch 106] validation: err-top1=0.216239 err-top5=0.057239
Epoch[107] Batch [49]	Speed: 2943.669993 samples/sec	accuracy=0.818086	lr=0.012406
Epoch[107] Batch [99]	Speed: 2778.670702 samples/sec	accuracy=0.818184	lr=0.012330
Epoch[107] Batch [149]	Speed: 2717.625257 samples/sec	accuracy=0.817786	lr=0.012255
Epoch[107] Batch [199]	Speed: 2759.661053 samples/sec	accuracy=0.817607	lr=0.012180
Epoch[107] Batch [249]	Speed: 2649.312436 samples/sec	accuracy=0.817937	lr=0.012105
Epoch[107] Batch [299]	Speed: 2711.916548 samples/sec	accuracy=0.817731	lr=0.012030
Epoch[107] Batch [349]	Speed: 2778.697236 samples/sec	accuracy=0.817559	lr=0.011956
Epoch[107] Batch [399]	Speed: 2721.516425 samples/sec	accuracy=0.817012	lr=0.011881
Epoch[107] Batch [449]	Speed: 2678.888389 samples/sec	accuracy=0.816641	lr=0.011807
Epoch[107] Batch [499]	Speed: 2698.919583 samples/sec	accuracy=0.816539	lr=0.011734
Epoch[107] Batch [549]	Speed: 2777.619380 samples/sec	accuracy=0.816186	lr=0.011660
Epoch[107] Batch [599]	Speed: 2719.001057 samples/sec	accuracy=0.815881	lr=0.011587
Epoch[107] Batch [649]	Speed: 2755.304229 samples/sec	accuracy=0.815786	lr=0.011513
Epoch[107] Batch [699]	Speed: 2862.412619 samples/sec	accuracy=0.815931	lr=0.011441
Epoch[107] Batch [749]	Speed: 2664.773703 samples/sec	accuracy=0.815810	lr=0.011368
Epoch[107] Batch [799]	Speed: 2828.352714 samples/sec	accuracy=0.815618	lr=0.011295
Epoch[107] Batch [849]	Speed: 2716.503187 samples/sec	accuracy=0.815641	lr=0.011223
Epoch[107] Batch [899]	Speed: 2826.535850 samples/sec	accuracy=0.815522	lr=0.011151
Epoch[107] Batch [949]	Speed: 2797.683326 samples/sec	accuracy=0.815426	lr=0.011079
Epoch[107] Batch [999]	Speed: 2687.329837 samples/sec	accuracy=0.815332	lr=0.011008
Epoch[107] Batch [1049]	Speed: 2802.833741 samples/sec	accuracy=0.815165	lr=0.010936
Epoch[107] Batch [1099]	Speed: 2765.464030 samples/sec	accuracy=0.814983	lr=0.010865
Epoch[107] Batch [1149]	Speed: 2723.851393 samples/sec	accuracy=0.815062	lr=0.010794
Epoch[107] Batch [1199]	Speed: 2773.083312 samples/sec	accuracy=0.815039	lr=0.010724
Epoch[107] Batch [1249]	Speed: 2910.100199 samples/sec	accuracy=0.814878	lr=0.010653
[Epoch 107] training: accuracy=0.814884
[Epoch 107] speed: 2758 samples/sec	time cost: 474.474311
[Epoch 107] validation: err-top1=0.213429 err-top5=0.056621
Epoch[108] Batch [49]	Speed: 2790.711324 samples/sec	accuracy=0.822168	lr=0.010582
Epoch[108] Batch [99]	Speed: 2853.162148 samples/sec	accuracy=0.822979	lr=0.010512
Epoch[108] Batch [149]	Speed: 2815.536863 samples/sec	accuracy=0.821719	lr=0.010442
Epoch[108] Batch [199]	Speed: 2763.230633 samples/sec	accuracy=0.821997	lr=0.010373
Epoch[108] Batch [249]	Speed: 2800.428835 samples/sec	accuracy=0.822309	lr=0.010303
Epoch[108] Batch [299]	Speed: 2769.796587 samples/sec	accuracy=0.821898	lr=0.010234
Epoch[108] Batch [349]	Speed: 2639.211442 samples/sec	accuracy=0.821554	lr=0.010165
Epoch[108] Batch [399]	Speed: 2735.809843 samples/sec	accuracy=0.821343	lr=0.010097
Epoch[108] Batch [449]	Speed: 2690.326985 samples/sec	accuracy=0.821545	lr=0.010028
Epoch[108] Batch [499]	Speed: 2678.311352 samples/sec	accuracy=0.821365	lr=0.009960
Epoch[108] Batch [549]	Speed: 2699.662999 samples/sec	accuracy=0.821101	lr=0.009892
Epoch[108] Batch [599]	Speed: 2717.319037 samples/sec	accuracy=0.821006	lr=0.009824
Epoch[108] Batch [649]	Speed: 2761.976882 samples/sec	accuracy=0.820913	lr=0.009757
Epoch[108] Batch [699]	Speed: 2692.076074 samples/sec	accuracy=0.820898	lr=0.009690
Epoch[108] Batch [749]	Speed: 2768.297507 samples/sec	accuracy=0.820754	lr=0.009623
Epoch[108] Batch [799]	Speed: 2630.426509 samples/sec	accuracy=0.820544	lr=0.009556
Epoch[108] Batch [849]	Speed: 2704.265118 samples/sec	accuracy=0.820461	lr=0.009489
Epoch[108] Batch [899]	Speed: 2745.714939 samples/sec	accuracy=0.820309	lr=0.009423
Epoch[108] Batch [949]	Speed: 2741.148239 samples/sec	accuracy=0.820216	lr=0.009357
Epoch[108] Batch [999]	Speed: 2672.821669 samples/sec	accuracy=0.820164	lr=0.009291
Epoch[108] Batch [1049]	Speed: 2736.550916 samples/sec	accuracy=0.820135	lr=0.009225
Epoch[108] Batch [1099]	Speed: 2773.516816 samples/sec	accuracy=0.820082	lr=0.009160
Epoch[108] Batch [1149]	Speed: 2639.928588 samples/sec	accuracy=0.819868	lr=0.009095
Epoch[108] Batch [1199]	Speed: 2792.891853 samples/sec	accuracy=0.819878	lr=0.009030
Epoch[108] Batch [1249]	Speed: 2874.301249 samples/sec	accuracy=0.819760	lr=0.008965
[Epoch 108] training: accuracy=0.819765
[Epoch 108] speed: 2736 samples/sec	time cost: 477.738264
[Epoch 108] validation: err-top1=0.210419 err-top5=0.055086
Epoch[109] Batch [49]	Speed: 2693.533111 samples/sec	accuracy=0.827500	lr=0.008899
Epoch[109] Batch [99]	Speed: 2703.649356 samples/sec	accuracy=0.825146	lr=0.008835
Epoch[109] Batch [149]	Speed: 2709.429483 samples/sec	accuracy=0.826510	lr=0.008771
Epoch[109] Batch [199]	Speed: 2677.327880 samples/sec	accuracy=0.825977	lr=0.008707
Epoch[109] Batch [249]	Speed: 2795.615296 samples/sec	accuracy=0.825891	lr=0.008643
Epoch[109] Batch [299]	Speed: 2638.434034 samples/sec	accuracy=0.825775	lr=0.008580
Epoch[109] Batch [349]	Speed: 2771.574664 samples/sec	accuracy=0.825064	lr=0.008517
Epoch[109] Batch [399]	Speed: 2677.326177 samples/sec	accuracy=0.825068	lr=0.008454
Epoch[109] Batch [449]	Speed: 2669.949448 samples/sec	accuracy=0.825022	lr=0.008391
Epoch[109] Batch [499]	Speed: 2728.468611 samples/sec	accuracy=0.824629	lr=0.008329
Epoch[109] Batch [549]	Speed: 2752.285402 samples/sec	accuracy=0.824547	lr=0.008266
Epoch[109] Batch [599]	Speed: 2705.220841 samples/sec	accuracy=0.824419	lr=0.008204
Epoch[109] Batch [649]	Speed: 2681.157180 samples/sec	accuracy=0.824527	lr=0.008143
Epoch[109] Batch [699]	Speed: 2647.719880 samples/sec	accuracy=0.824457	lr=0.008081
Epoch[109] Batch [749]	Speed: 2679.772451 samples/sec	accuracy=0.824514	lr=0.008020
Epoch[109] Batch [799]	Speed: 2646.714182 samples/sec	accuracy=0.824680	lr=0.007959
Epoch[109] Batch [849]	Speed: 2747.082506 samples/sec	accuracy=0.824663	lr=0.007898
Epoch[109] Batch [899]	Speed: 2775.092272 samples/sec	accuracy=0.824671	lr=0.007837
Epoch[109] Batch [949]	Speed: 2695.904895 samples/sec	accuracy=0.824585	lr=0.007777
Epoch[109] Batch [999]	Speed: 2622.694872 samples/sec	accuracy=0.824548	lr=0.007716
Epoch[109] Batch [1049]	Speed: 2726.378481 samples/sec	accuracy=0.824422	lr=0.007656
Epoch[109] Batch [1099]	Speed: 2756.664736 samples/sec	accuracy=0.824410	lr=0.007597
Epoch[109] Batch [1149]	Speed: 2689.110386 samples/sec	accuracy=0.824304	lr=0.007537
Epoch[109] Batch [1199]	Speed: 2740.248330 samples/sec	accuracy=0.824288	lr=0.007478
Epoch[109] Batch [1249]	Speed: 2874.278474 samples/sec	accuracy=0.824143	lr=0.007419
[Epoch 109] training: accuracy=0.824159
[Epoch 109] speed: 2709 samples/sec	time cost: 482.699944
[Epoch 109] validation: err-top1=0.211077 err-top5=0.055146
Epoch[110] Batch [49]	Speed: 2814.426850 samples/sec	accuracy=0.829668	lr=0.007359
Epoch[110] Batch [99]	Speed: 2751.461538 samples/sec	accuracy=0.831025	lr=0.007300
Epoch[110] Batch [149]	Speed: 2728.953783 samples/sec	accuracy=0.830964	lr=0.007242
Epoch[110] Batch [199]	Speed: 2791.763545 samples/sec	accuracy=0.830518	lr=0.007184
Epoch[110] Batch [249]	Speed: 2717.531062 samples/sec	accuracy=0.830348	lr=0.007126
Epoch[110] Batch [299]	Speed: 2684.485932 samples/sec	accuracy=0.830426	lr=0.007068
Epoch[110] Batch [349]	Speed: 2749.673290 samples/sec	accuracy=0.830229	lr=0.007011
Epoch[110] Batch [399]	Speed: 2878.696127 samples/sec	accuracy=0.830264	lr=0.006954
Epoch[110] Batch [449]	Speed: 2783.749592 samples/sec	accuracy=0.830289	lr=0.006897
Epoch[110] Batch [499]	Speed: 2664.607586 samples/sec	accuracy=0.829971	lr=0.006840
Epoch[110] Batch [549]	Speed: 2747.145726 samples/sec	accuracy=0.829636	lr=0.006784
Epoch[110] Batch [599]	Speed: 2685.275341 samples/sec	accuracy=0.829557	lr=0.006727
Epoch[110] Batch [649]	Speed: 2768.485941 samples/sec	accuracy=0.829390	lr=0.006671
Epoch[110] Batch [699]	Speed: 2804.681386 samples/sec	accuracy=0.829290	lr=0.006615
Epoch[110] Batch [749]	Speed: 2816.064169 samples/sec	accuracy=0.829115	lr=0.006560
Epoch[110] Batch [799]	Speed: 2834.000970 samples/sec	accuracy=0.829200	lr=0.006504
Epoch[110] Batch [849]	Speed: 2748.876252 samples/sec	accuracy=0.829067	lr=0.006449
Epoch[110] Batch [899]	Speed: 2692.016780 samples/sec	accuracy=0.828929	lr=0.006394
Epoch[110] Batch [949]	Speed: 2839.098412 samples/sec	accuracy=0.829066	lr=0.006340
Epoch[110] Batch [999]	Speed: 2823.628668 samples/sec	accuracy=0.828990	lr=0.006285
Epoch[110] Batch [1049]	Speed: 2749.814901 samples/sec	accuracy=0.828975	lr=0.006231
Epoch[110] Batch [1099]	Speed: 2674.442208 samples/sec	accuracy=0.828954	lr=0.006177
Epoch[110] Batch [1149]	Speed: 2740.333091 samples/sec	accuracy=0.828956	lr=0.006123
Epoch[110] Batch [1199]	Speed: 2693.027050 samples/sec	accuracy=0.828905	lr=0.006070
Epoch[110] Batch [1249]	Speed: 2889.024348 samples/sec	accuracy=0.828845	lr=0.006017
[Epoch 110] training: accuracy=0.828834
[Epoch 110] speed: 2759 samples/sec	time cost: 474.101925
[Epoch 110] validation: err-top1=0.208883 err-top5=0.054179
Epoch[111] Batch [49]	Speed: 2780.834105 samples/sec	accuracy=0.834336	lr=0.005963
Epoch[111] Batch [99]	Speed: 2726.025714 samples/sec	accuracy=0.833818	lr=0.005910
Epoch[111] Batch [149]	Speed: 2704.696413 samples/sec	accuracy=0.834316	lr=0.005857
Epoch[111] Batch [199]	Speed: 2734.407339 samples/sec	accuracy=0.833765	lr=0.005805
Epoch[111] Batch [249]	Speed: 2643.494398 samples/sec	accuracy=0.833969	lr=0.005753
Epoch[111] Batch [299]	Speed: 2718.365320 samples/sec	accuracy=0.834144	lr=0.005701
Epoch[111] Batch [349]	Speed: 2757.002436 samples/sec	accuracy=0.833982	lr=0.005649
Epoch[111] Batch [399]	Speed: 2649.233767 samples/sec	accuracy=0.833804	lr=0.005598
Epoch[111] Batch [449]	Speed: 2711.126151 samples/sec	accuracy=0.833644	lr=0.005547
Epoch[111] Batch [499]	Speed: 2733.539025 samples/sec	accuracy=0.833654	lr=0.005496
Epoch[111] Batch [549]	Speed: 2747.709634 samples/sec	accuracy=0.833414	lr=0.005445
Epoch[111] Batch [599]	Speed: 2699.629435 samples/sec	accuracy=0.833201	lr=0.005394
Epoch[111] Batch [649]	Speed: 2748.860911 samples/sec	accuracy=0.833188	lr=0.005344
Epoch[111] Batch [699]	Speed: 2756.383477 samples/sec	accuracy=0.833429	lr=0.005294
Epoch[111] Batch [749]	Speed: 2653.408749 samples/sec	accuracy=0.833452	lr=0.005244
Epoch[111] Batch [799]	Speed: 2685.072985 samples/sec	accuracy=0.833705	lr=0.005195
Epoch[111] Batch [849]	Speed: 2763.606468 samples/sec	accuracy=0.833514	lr=0.005145
Epoch[111] Batch [899]	Speed: 2659.037564 samples/sec	accuracy=0.833478	lr=0.005096
Epoch[111] Batch [949]	Speed: 2750.121974 samples/sec	accuracy=0.833438	lr=0.005047
Epoch[111] Batch [999]	Speed: 2714.457246 samples/sec	accuracy=0.833357	lr=0.004999
Epoch[111] Batch [1049]	Speed: 2736.751306 samples/sec	accuracy=0.833388	lr=0.004950
Epoch[111] Batch [1099]	Speed: 2627.829263 samples/sec	accuracy=0.833285	lr=0.004902
Epoch[111] Batch [1149]	Speed: 2680.006250 samples/sec	accuracy=0.833210	lr=0.004854
Epoch[111] Batch [1199]	Speed: 2652.691700 samples/sec	accuracy=0.833177	lr=0.004807
Epoch[111] Batch [1249]	Speed: 2871.579728 samples/sec	accuracy=0.833189	lr=0.004759
[Epoch 111] training: accuracy=0.833180
[Epoch 111] speed: 2713 samples/sec	time cost: 482.147088
[Epoch 111] validation: err-top1=0.207928 err-top5=0.053292
Epoch[112] Batch [49]	Speed: 2882.999264 samples/sec	accuracy=0.836855	lr=0.004711
Epoch[112] Batch [99]	Speed: 2644.574800 samples/sec	accuracy=0.836475	lr=0.004664
Epoch[112] Batch [149]	Speed: 2644.059653 samples/sec	accuracy=0.836510	lr=0.004617
Epoch[112] Batch [199]	Speed: 2747.237276 samples/sec	accuracy=0.836646	lr=0.004571
Epoch[112] Batch [249]	Speed: 2738.697089 samples/sec	accuracy=0.836812	lr=0.004524
Epoch[112] Batch [299]	Speed: 2746.661301 samples/sec	accuracy=0.837044	lr=0.004478
Epoch[112] Batch [349]	Speed: 2687.614233 samples/sec	accuracy=0.837126	lr=0.004432
Epoch[112] Batch [399]	Speed: 2682.984662 samples/sec	accuracy=0.836958	lr=0.004387
Epoch[112] Batch [449]	Speed: 2686.923696 samples/sec	accuracy=0.837196	lr=0.004341
Epoch[112] Batch [499]	Speed: 2735.488987 samples/sec	accuracy=0.837205	lr=0.004296
Epoch[112] Batch [549]	Speed: 2699.457044 samples/sec	accuracy=0.837372	lr=0.004251
Epoch[112] Batch [599]	Speed: 2753.339481 samples/sec	accuracy=0.837459	lr=0.004207
Epoch[112] Batch [649]	Speed: 2688.804060 samples/sec	accuracy=0.837542	lr=0.004162
Epoch[112] Batch [699]	Speed: 2777.791335 samples/sec	accuracy=0.837307	lr=0.004118
Epoch[112] Batch [749]	Speed: 2701.810451 samples/sec	accuracy=0.837348	lr=0.004074
Epoch[112] Batch [799]	Speed: 2779.860381 samples/sec	accuracy=0.837350	lr=0.004030
Epoch[112] Batch [849]	Speed: 2743.426009 samples/sec	accuracy=0.837325	lr=0.003987
Epoch[112] Batch [899]	Speed: 2737.250384 samples/sec	accuracy=0.837307	lr=0.003944
Epoch[112] Batch [949]	Speed: 2750.590780 samples/sec	accuracy=0.837363	lr=0.003901
Epoch[112] Batch [999]	Speed: 2704.235389 samples/sec	accuracy=0.837330	lr=0.003858
Epoch[112] Batch [1049]	Speed: 2678.680412 samples/sec	accuracy=0.837299	lr=0.003815
Epoch[112] Batch [1099]	Speed: 2700.542999 samples/sec	accuracy=0.837279	lr=0.003773
Epoch[112] Batch [1149]	Speed: 2741.952704 samples/sec	accuracy=0.837238	lr=0.003731
Epoch[112] Batch [1199]	Speed: 2770.603981 samples/sec	accuracy=0.837323	lr=0.003689
Epoch[112] Batch [1249]	Speed: 2828.221262 samples/sec	accuracy=0.837253	lr=0.003647
[Epoch 112] training: accuracy=0.837241
[Epoch 112] speed: 2727 samples/sec	time cost: 479.254082
[Epoch 112] validation: err-top1=0.206613 err-top5=0.053850
Epoch[113] Batch [49]	Speed: 2674.112009 samples/sec	accuracy=0.841504	lr=0.003605
Epoch[113] Batch [99]	Speed: 2750.818672 samples/sec	accuracy=0.841123	lr=0.003564
Epoch[113] Batch [149]	Speed: 2788.105652 samples/sec	accuracy=0.841823	lr=0.003523
Epoch[113] Batch [199]	Speed: 2617.833321 samples/sec	accuracy=0.841821	lr=0.003482
Epoch[113] Batch [249]	Speed: 2720.075923 samples/sec	accuracy=0.841930	lr=0.003442
Epoch[113] Batch [299]	Speed: 2747.562651 samples/sec	accuracy=0.841904	lr=0.003402
Epoch[113] Batch [349]	Speed: 2754.666210 samples/sec	accuracy=0.841641	lr=0.003362
Epoch[113] Batch [399]	Speed: 2830.138384 samples/sec	accuracy=0.841511	lr=0.003322
Epoch[113] Batch [449]	Speed: 2710.062108 samples/sec	accuracy=0.841482	lr=0.003282
Epoch[113] Batch [499]	Speed: 2781.367367 samples/sec	accuracy=0.841404	lr=0.003243
Epoch[113] Batch [549]	Speed: 2754.894176 samples/sec	accuracy=0.841286	lr=0.003204
Epoch[113] Batch [599]	Speed: 2764.803962 samples/sec	accuracy=0.841424	lr=0.003165
Epoch[113] Batch [649]	Speed: 2684.278326 samples/sec	accuracy=0.841209	lr=0.003127
Epoch[113] Batch [699]	Speed: 2764.930582 samples/sec	accuracy=0.841177	lr=0.003088
Epoch[113] Batch [749]	Speed: 2686.251826 samples/sec	accuracy=0.841064	lr=0.003050
Epoch[113] Batch [799]	Speed: 2681.365173 samples/sec	accuracy=0.841051	lr=0.003012
Epoch[113] Batch [849]	Speed: 2659.366850 samples/sec	accuracy=0.840967	lr=0.002975
Epoch[113] Batch [899]	Speed: 2709.301606 samples/sec	accuracy=0.840753	lr=0.002937
Epoch[113] Batch [949]	Speed: 2698.278249 samples/sec	accuracy=0.840839	lr=0.002900
Epoch[113] Batch [999]	Speed: 2704.456685 samples/sec	accuracy=0.840788	lr=0.002863
Epoch[113] Batch [1049]	Speed: 2738.445744 samples/sec	accuracy=0.840688	lr=0.002826
Epoch[113] Batch [1099]	Speed: 2772.922466 samples/sec	accuracy=0.840601	lr=0.002790
Epoch[113] Batch [1149]	Speed: 2683.455032 samples/sec	accuracy=0.840591	lr=0.002754
Epoch[113] Batch [1199]	Speed: 2678.883911 samples/sec	accuracy=0.840663	lr=0.002718
Epoch[113] Batch [1249]	Speed: 2856.091902 samples/sec	accuracy=0.840633	lr=0.002682
[Epoch 113] training: accuracy=0.840645
[Epoch 113] speed: 2725 samples/sec	time cost: 479.486323
[Epoch 113] validation: err-top1=0.205198 err-top5=0.053133
Epoch[114] Batch [49]	Speed: 2933.427007 samples/sec	accuracy=0.843340	lr=0.002646
Epoch[114] Batch [99]	Speed: 2707.081968 samples/sec	accuracy=0.843584	lr=0.002611
Epoch[114] Batch [149]	Speed: 2717.630588 samples/sec	accuracy=0.844603	lr=0.002575
Epoch[114] Batch [199]	Speed: 2796.684075 samples/sec	accuracy=0.844697	lr=0.002541
Epoch[114] Batch [249]	Speed: 2684.724247 samples/sec	accuracy=0.844461	lr=0.002506
Epoch[114] Batch [299]	Speed: 2785.015463 samples/sec	accuracy=0.844404	lr=0.002472
Epoch[114] Batch [349]	Speed: 2730.049480 samples/sec	accuracy=0.844085	lr=0.002438
Epoch[114] Batch [399]	Speed: 2591.720602 samples/sec	accuracy=0.843979	lr=0.002404
Epoch[114] Batch [449]	Speed: 2674.045513 samples/sec	accuracy=0.843989	lr=0.002370
Epoch[114] Batch [499]	Speed: 2701.920522 samples/sec	accuracy=0.843707	lr=0.002337
Epoch[114] Batch [549]	Speed: 2790.880479 samples/sec	accuracy=0.843386	lr=0.002304
Epoch[114] Batch [599]	Speed: 2641.400014 samples/sec	accuracy=0.843350	lr=0.002271
Epoch[114] Batch [649]	Speed: 2635.429006 samples/sec	accuracy=0.843559	lr=0.002238
Epoch[114] Batch [699]	Speed: 2789.802141 samples/sec	accuracy=0.843464	lr=0.002205
Epoch[114] Batch [749]	Speed: 2746.554053 samples/sec	accuracy=0.843408	lr=0.002173
Epoch[114] Batch [799]	Speed: 2735.855013 samples/sec	accuracy=0.843452	lr=0.002141
Epoch[114] Batch [849]	Speed: 2715.320688 samples/sec	accuracy=0.843362	lr=0.002110
Epoch[114] Batch [899]	Speed: 2741.256990 samples/sec	accuracy=0.843314	lr=0.002078
Epoch[114] Batch [949]	Speed: 2798.805884 samples/sec	accuracy=0.843237	lr=0.002047
Epoch[114] Batch [999]	Speed: 2728.412626 samples/sec	accuracy=0.843201	lr=0.002016
Epoch[114] Batch [1049]	Speed: 2756.956140 samples/sec	accuracy=0.842985	lr=0.001985
Epoch[114] Batch [1099]	Speed: 2701.075976 samples/sec	accuracy=0.843074	lr=0.001954
Epoch[114] Batch [1149]	Speed: 2680.281939 samples/sec	accuracy=0.843004	lr=0.001924
Epoch[114] Batch [1199]	Speed: 2788.947047 samples/sec	accuracy=0.842890	lr=0.001894
Epoch[114] Batch [1249]	Speed: 2777.112262 samples/sec	accuracy=0.842928	lr=0.001864
[Epoch 114] training: accuracy=0.842923
[Epoch 114] speed: 2730 samples/sec	time cost: 478.912932
[Epoch 114] validation: err-top1=0.205058 err-top5=0.052954
Epoch[115] Batch [49]	Speed: 2760.085794 samples/sec	accuracy=0.846309	lr=0.001834
Epoch[115] Batch [99]	Speed: 2742.867816 samples/sec	accuracy=0.844805	lr=0.001804
Epoch[115] Batch [149]	Speed: 2721.664429 samples/sec	accuracy=0.844915	lr=0.001775
Epoch[115] Batch [199]	Speed: 2683.127398 samples/sec	accuracy=0.844795	lr=0.001746
Epoch[115] Batch [249]	Speed: 2699.968104 samples/sec	accuracy=0.845063	lr=0.001718
Epoch[115] Batch [299]	Speed: 2717.122342 samples/sec	accuracy=0.845456	lr=0.001689
Epoch[115] Batch [349]	Speed: 2735.459753 samples/sec	accuracy=0.845583	lr=0.001661
Epoch[115] Batch [399]	Speed: 2773.287906 samples/sec	accuracy=0.845408	lr=0.001633
Epoch[115] Batch [449]	Speed: 2679.963875 samples/sec	accuracy=0.845927	lr=0.001605
Epoch[115] Batch [499]	Speed: 2839.768149 samples/sec	accuracy=0.845471	lr=0.001578
Epoch[115] Batch [549]	Speed: 2717.250856 samples/sec	accuracy=0.845517	lr=0.001551
Epoch[115] Batch [599]	Speed: 2664.016129 samples/sec	accuracy=0.845324	lr=0.001524
Epoch[115] Batch [649]	Speed: 2647.866268 samples/sec	accuracy=0.845225	lr=0.001497
Epoch[115] Batch [699]	Speed: 2633.365559 samples/sec	accuracy=0.845035	lr=0.001470
Epoch[115] Batch [749]	Speed: 2610.283841 samples/sec	accuracy=0.845044	lr=0.001444
Epoch[115] Batch [799]	Speed: 2750.705003 samples/sec	accuracy=0.845046	lr=0.001418
Epoch[115] Batch [849]	Speed: 2701.899853 samples/sec	accuracy=0.844969	lr=0.001392
Epoch[115] Batch [899]	Speed: 2802.275504 samples/sec	accuracy=0.844914	lr=0.001366
Epoch[115] Batch [949]	Speed: 2678.947807 samples/sec	accuracy=0.844981	lr=0.001341
Epoch[115] Batch [999]	Speed: 2771.372433 samples/sec	accuracy=0.844921	lr=0.001316
Epoch[115] Batch [1049]	Speed: 2803.677278 samples/sec	accuracy=0.844993	lr=0.001291
Epoch[115] Batch [1099]	Speed: 2884.867530 samples/sec	accuracy=0.845029	lr=0.001266
Epoch[115] Batch [1149]	Speed: 2953.985291 samples/sec	accuracy=0.845118	lr=0.001242
Epoch[115] Batch [1199]	Speed: 2980.878574 samples/sec	accuracy=0.845236	lr=0.001218
Epoch[115] Batch [1249]	Speed: 2909.031262 samples/sec	accuracy=0.845235	lr=0.001194
[Epoch 115] training: accuracy=0.845234
[Epoch 115] speed: 2749 samples/sec	time cost: 475.553302
[Epoch 115] validation: err-top1=0.203663 err-top5=0.052615
Epoch[116] Batch [49]	Speed: 2830.202091 samples/sec	accuracy=0.846172	lr=0.001170
Epoch[116] Batch [99]	Speed: 2743.551239 samples/sec	accuracy=0.846748	lr=0.001146
Epoch[116] Batch [149]	Speed: 2805.633797 samples/sec	accuracy=0.846667	lr=0.001123
Epoch[116] Batch [199]	Speed: 2774.031043 samples/sec	accuracy=0.846909	lr=0.001100
Epoch[116] Batch [249]	Speed: 2804.369846 samples/sec	accuracy=0.846543	lr=0.001077
Epoch[116] Batch [299]	Speed: 2757.407027 samples/sec	accuracy=0.846891	lr=0.001055
Epoch[116] Batch [349]	Speed: 2739.677484 samples/sec	accuracy=0.846883	lr=0.001032
Epoch[116] Batch [399]	Speed: 2713.160764 samples/sec	accuracy=0.846804	lr=0.001010
Epoch[116] Batch [449]	Speed: 2761.992654 samples/sec	accuracy=0.846682	lr=0.000989
Epoch[116] Batch [499]	Speed: 2734.202037 samples/sec	accuracy=0.846732	lr=0.000967
Epoch[116] Batch [549]	Speed: 2669.965050 samples/sec	accuracy=0.846914	lr=0.000946
Epoch[116] Batch [599]	Speed: 2724.237914 samples/sec	accuracy=0.847012	lr=0.000925
Epoch[116] Batch [649]	Speed: 2810.378281 samples/sec	accuracy=0.846940	lr=0.000904
Epoch[116] Batch [699]	Speed: 2748.515458 samples/sec	accuracy=0.847200	lr=0.000883
Epoch[116] Batch [749]	Speed: 2720.152825 samples/sec	accuracy=0.847362	lr=0.000863
Epoch[116] Batch [799]	Speed: 2741.699851 samples/sec	accuracy=0.847389	lr=0.000843
Epoch[116] Batch [849]	Speed: 2690.128652 samples/sec	accuracy=0.847437	lr=0.000823
Epoch[116] Batch [899]	Speed: 2721.188086 samples/sec	accuracy=0.847426	lr=0.000803
Epoch[116] Batch [949]	Speed: 2752.919213 samples/sec	accuracy=0.847275	lr=0.000784
Epoch[116] Batch [999]	Speed: 2753.392081 samples/sec	accuracy=0.847293	lr=0.000764
Epoch[116] Batch [1049]	Speed: 2687.014234 samples/sec	accuracy=0.847269	lr=0.000745
Epoch[116] Batch [1099]	Speed: 2639.595824 samples/sec	accuracy=0.847297	lr=0.000727
Epoch[116] Batch [1149]	Speed: 2712.804829 samples/sec	accuracy=0.847350	lr=0.000708
Epoch[116] Batch [1199]	Speed: 2748.824880 samples/sec	accuracy=0.847380	lr=0.000690
Epoch[116] Batch [1249]	Speed: 2868.966608 samples/sec	accuracy=0.847343	lr=0.000672
[Epoch 116] training: accuracy=0.847356
[Epoch 116] speed: 2743 samples/sec	time cost: 476.899930
[Epoch 116] validation: err-top1=0.204000 err-top5=0.052572
Epoch[117] Batch [49]	Speed: 2966.834104 samples/sec	accuracy=0.850918	lr=0.000654
Epoch[117] Batch [99]	Speed: 2786.136009 samples/sec	accuracy=0.849502	lr=0.000636
Epoch[117] Batch [149]	Speed: 2879.522320 samples/sec	accuracy=0.848607	lr=0.000619
Epoch[117] Batch [199]	Speed: 2736.708582 samples/sec	accuracy=0.848564	lr=0.000602
Epoch[117] Batch [249]	Speed: 2880.754267 samples/sec	accuracy=0.848680	lr=0.000585
Epoch[117] Batch [299]	Speed: 2838.193180 samples/sec	accuracy=0.848584	lr=0.000569
Epoch[117] Batch [349]	Speed: 2802.268922 samples/sec	accuracy=0.848418	lr=0.000552
Epoch[117] Batch [399]	Speed: 2733.875607 samples/sec	accuracy=0.848689	lr=0.000536
Epoch[117] Batch [449]	Speed: 2664.214762 samples/sec	accuracy=0.848770	lr=0.000520
Epoch[117] Batch [499]	Speed: 2637.201561 samples/sec	accuracy=0.848773	lr=0.000505
Epoch[117] Batch [549]	Speed: 2756.775288 samples/sec	accuracy=0.848667	lr=0.000489
Epoch[117] Batch [599]	Speed: 2787.626432 samples/sec	accuracy=0.848462	lr=0.000474
Epoch[117] Batch [649]	Speed: 2720.333762 samples/sec	accuracy=0.848480	lr=0.000459
Epoch[117] Batch [699]	Speed: 2736.618047 samples/sec	accuracy=0.848631	lr=0.000445
Epoch[117] Batch [749]	Speed: 2816.481443 samples/sec	accuracy=0.848514	lr=0.000430
Epoch[117] Batch [799]	Speed: 2712.117969 samples/sec	accuracy=0.848396	lr=0.000416
Epoch[117] Batch [849]	Speed: 2742.308975 samples/sec	accuracy=0.848465	lr=0.000402
Epoch[117] Batch [899]	Speed: 2838.253124 samples/sec	accuracy=0.848519	lr=0.000388
Epoch[117] Batch [949]	Speed: 2637.006126 samples/sec	accuracy=0.848483	lr=0.000375
Epoch[117] Batch [999]	Speed: 2730.135069 samples/sec	accuracy=0.848493	lr=0.000362
Epoch[117] Batch [1049]	Speed: 2743.477319 samples/sec	accuracy=0.848511	lr=0.000349
Epoch[117] Batch [1099]	Speed: 2697.541424 samples/sec	accuracy=0.848580	lr=0.000336
Epoch[117] Batch [1149]	Speed: 2741.178750 samples/sec	accuracy=0.848572	lr=0.000323
Epoch[117] Batch [1199]	Speed: 2714.374353 samples/sec	accuracy=0.848467	lr=0.000311
Epoch[117] Batch [1249]	Speed: 2864.165816 samples/sec	accuracy=0.848429	lr=0.000299
[Epoch 117] training: accuracy=0.848434
[Epoch 117] speed: 2762 samples/sec	time cost: 473.917979
[Epoch 117] validation: err-top1=0.204042 err-top5=0.052196
Epoch[118] Batch [49]	Speed: 2720.059317 samples/sec	accuracy=0.851172	lr=0.000287
Epoch[118] Batch [99]	Speed: 2724.704747 samples/sec	accuracy=0.850225	lr=0.000275
Epoch[118] Batch [149]	Speed: 2800.251363 samples/sec	accuracy=0.849466	lr=0.000264
Epoch[118] Batch [199]	Speed: 2739.307395 samples/sec	accuracy=0.849570	lr=0.000253
Epoch[118] Batch [249]	Speed: 2627.645053 samples/sec	accuracy=0.849379	lr=0.000242
Epoch[118] Batch [299]	Speed: 2644.587013 samples/sec	accuracy=0.849619	lr=0.000231
Epoch[118] Batch [349]	Speed: 2708.893512 samples/sec	accuracy=0.849713	lr=0.000221
Epoch[118] Batch [399]	Speed: 2737.392323 samples/sec	accuracy=0.849624	lr=0.000211
Epoch[118] Batch [449]	Speed: 2682.433502 samples/sec	accuracy=0.849405	lr=0.000201
Epoch[118] Batch [499]	Speed: 2890.852256 samples/sec	accuracy=0.849338	lr=0.000191
Epoch[118] Batch [549]	Speed: 2706.996692 samples/sec	accuracy=0.849606	lr=0.000182
Epoch[118] Batch [599]	Speed: 2773.012125 samples/sec	accuracy=0.849432	lr=0.000173
Epoch[118] Batch [649]	Speed: 2753.568993 samples/sec	accuracy=0.849509	lr=0.000164
Epoch[118] Batch [699]	Speed: 2736.937388 samples/sec	accuracy=0.849262	lr=0.000155
Epoch[118] Batch [749]	Speed: 2780.684888 samples/sec	accuracy=0.849227	lr=0.000147
Epoch[118] Batch [799]	Speed: 2741.188127 samples/sec	accuracy=0.849371	lr=0.000138
Epoch[118] Batch [849]	Speed: 2839.713511 samples/sec	accuracy=0.849413	lr=0.000130
Epoch[118] Batch [899]	Speed: 2787.870636 samples/sec	accuracy=0.849311	lr=0.000123
Epoch[118] Batch [949]	Speed: 2604.093438 samples/sec	accuracy=0.849455	lr=0.000115
Epoch[118] Batch [999]	Speed: 2771.389564 samples/sec	accuracy=0.849448	lr=0.000108
Epoch[118] Batch [1049]	Speed: 2759.102616 samples/sec	accuracy=0.849594	lr=0.000101
Epoch[118] Batch [1099]	Speed: 2739.070891 samples/sec	accuracy=0.849680	lr=0.000094
Epoch[118] Batch [1149]	Speed: 2785.088784 samples/sec	accuracy=0.849693	lr=0.000087
Epoch[118] Batch [1199]	Speed: 2799.596777 samples/sec	accuracy=0.849583	lr=0.000081
Epoch[118] Batch [1249]	Speed: 2786.967137 samples/sec	accuracy=0.849498	lr=0.000075
[Epoch 118] training: accuracy=0.849492
[Epoch 118] speed: 2742 samples/sec	time cost: 477.049644
[Epoch 118] validation: err-top1=0.203723 err-top5=0.052156
Epoch[119] Batch [49]	Speed: 2798.522088 samples/sec	accuracy=0.848887	lr=0.000069
Epoch[119] Batch [99]	Speed: 2689.413784 samples/sec	accuracy=0.849727	lr=0.000063
Epoch[119] Batch [149]	Speed: 2734.431712 samples/sec	accuracy=0.849928	lr=0.000058
Epoch[119] Batch [199]	Speed: 2720.343411 samples/sec	accuracy=0.849146	lr=0.000053
Epoch[119] Batch [249]	Speed: 2649.366006 samples/sec	accuracy=0.849582	lr=0.000048
Epoch[119] Batch [299]	Speed: 2811.572634 samples/sec	accuracy=0.849395	lr=0.000043
Epoch[119] Batch [349]	Speed: 2752.581913 samples/sec	accuracy=0.849509	lr=0.000039
Epoch[119] Batch [399]	Speed: 2717.380757 samples/sec	accuracy=0.850139	lr=0.000035
Epoch[119] Batch [449]	Speed: 2691.059435 samples/sec	accuracy=0.850206	lr=0.000031
Epoch[119] Batch [499]	Speed: 2679.895984 samples/sec	accuracy=0.850219	lr=0.000027
Epoch[119] Batch [549]	Speed: 2678.577972 samples/sec	accuracy=0.850307	lr=0.000023
Epoch[119] Batch [599]	Speed: 2731.992449 samples/sec	accuracy=0.850321	lr=0.000020
Epoch[119] Batch [649]	Speed: 2826.183059 samples/sec	accuracy=0.849913	lr=0.000017
Epoch[119] Batch [699]	Speed: 2671.502064 samples/sec	accuracy=0.849907	lr=0.000015
Epoch[119] Batch [749]	Speed: 2775.056161 samples/sec	accuracy=0.849771	lr=0.000012
Epoch[119] Batch [799]	Speed: 2779.958659 samples/sec	accuracy=0.849572	lr=0.000010
Epoch[119] Batch [849]	Speed: 2771.609934 samples/sec	accuracy=0.849513	lr=0.000008
Epoch[119] Batch [899]	Speed: 2740.838444 samples/sec	accuracy=0.849601	lr=0.000006
Epoch[119] Batch [949]	Speed: 2697.740546 samples/sec	accuracy=0.849561	lr=0.000004
Epoch[119] Batch [999]	Speed: 2700.522351 samples/sec	accuracy=0.849540	lr=0.000003
Epoch[119] Batch [1049]	Speed: 2724.876263 samples/sec	accuracy=0.849575	lr=0.000002
Epoch[119] Batch [1099]	Speed: 2742.045063 samples/sec	accuracy=0.849672	lr=0.000001
Epoch[119] Batch [1149]	Speed: 2788.163969 samples/sec	accuracy=0.849700	lr=0.000000
Epoch[119] Batch [1199]	Speed: 2664.970927 samples/sec	accuracy=0.849678	lr=0.000000
Epoch[119] Batch [1249]	Speed: 2919.515592 samples/sec	accuracy=0.849695	lr=0.000000
[Epoch 119] training: accuracy=0.849710
[Epoch 119] speed: 2735 samples/sec	time cost: 478.384993
[Epoch 119] validation: err-top1=0.203783 err-top5=0.052117