Namespace(batch_norm=False, batch_size=128, data_dir='~/.mxnet/datasets/imagenet', dtype='float16', input_size=224, label_smoothing=True, last_gamma=True, log_interval=50, logging_file='resnet101_v1d.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_v1d', momentum=0.9, no_wd=True, num_epochs=120, num_gpus=8, num_workers=60, rec_train='/media/ramdisk/rec/train.rec', rec_train_idx='/media/ramdisk/rec/train.idx', rec_val='/media/ramdisk/rec/val.rec', rec_val_idx='/media/ramdisk/rec/val.idx', save_dir='params_resnet101_v1d', 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: 1135.463328 samples/sec	accuracy=0.814856	lr=0.003133
Epoch[0] Batch [99]	Speed: 2861.508895 samples/sec	accuracy=0.856625	lr=0.006331
Epoch[0] Batch [149]	Speed: 3179.444079 samples/sec	accuracy=0.883706	lr=0.009528
Epoch[0] Batch [199]	Speed: 3509.669598 samples/sec	accuracy=0.894473	lr=0.012726
Epoch[0] Batch [249]	Speed: 3497.024613 samples/sec	accuracy=0.894661	lr=0.015923
Epoch[0] Batch [299]	Speed: 3467.684620 samples/sec	accuracy=0.887755	lr=0.019121
Epoch[0] Batch [349]	Speed: 3325.274079 samples/sec	accuracy=0.876817	lr=0.022318
Epoch[0] Batch [399]	Speed: 3224.837532 samples/sec	accuracy=0.863683	lr=0.025516
Epoch[0] Batch [449]	Speed: 3354.837914 samples/sec	accuracy=0.849176	lr=0.028713
Epoch[0] Batch [499]	Speed: 3294.292282 samples/sec	accuracy=0.833664	lr=0.031910
Epoch[0] Batch [549]	Speed: 3179.015302 samples/sec	accuracy=0.818182	lr=0.035108
Epoch[0] Batch [599]	Speed: 3063.142274 samples/sec	accuracy=0.803004	lr=0.038305
Epoch[0] Batch [649]	Speed: 3143.882642 samples/sec	accuracy=0.788090	lr=0.041503
Epoch[0] Batch [699]	Speed: 2983.085475 samples/sec	accuracy=0.774046	lr=0.044700
Epoch[0] Batch [749]	Speed: 3064.312450 samples/sec	accuracy=0.760720	lr=0.047898
Epoch[0] Batch [799]	Speed: 2961.336466 samples/sec	accuracy=0.748173	lr=0.051095
Epoch[0] Batch [849]	Speed: 2971.845097 samples/sec	accuracy=0.736355	lr=0.054293
Epoch[0] Batch [899]	Speed: 2963.481554 samples/sec	accuracy=0.725191	lr=0.057490
Epoch[0] Batch [949]	Speed: 2954.575776 samples/sec	accuracy=0.714726	lr=0.060687
Epoch[0] Batch [999]	Speed: 2828.198019 samples/sec	accuracy=0.704777	lr=0.063885
Epoch[0] Batch [1049]	Speed: 2905.177698 samples/sec	accuracy=0.695514	lr=0.067082
Epoch[0] Batch [1099]	Speed: 2860.372441 samples/sec	accuracy=0.686641	lr=0.070280
Epoch[0] Batch [1149]	Speed: 2998.199947 samples/sec	accuracy=0.678233	lr=0.073477
Epoch[0] Batch [1199]	Speed: 3051.802360 samples/sec	accuracy=0.670397	lr=0.076675
Epoch[0] Batch [1249]	Speed: 3037.421814 samples/sec	accuracy=0.662986	lr=0.079872
[Epoch 0] training: accuracy=0.662697
[Epoch 0] speed: 2896 samples/sec	time cost: 452.900647
[Epoch 0] validation: err-top1=0.928870 err-top5=0.812520
Epoch[1] Batch [49]	Speed: 3126.844759 samples/sec	accuracy=0.474955	lr=0.083133
Epoch[1] Batch [99]	Speed: 2907.115206 samples/sec	accuracy=0.472382	lr=0.086331
Epoch[1] Batch [149]	Speed: 2879.185324 samples/sec	accuracy=0.471924	lr=0.089528
Epoch[1] Batch [199]	Speed: 2958.569296 samples/sec	accuracy=0.469336	lr=0.092726
Epoch[1] Batch [249]	Speed: 2741.765764 samples/sec	accuracy=0.466726	lr=0.095923
Epoch[1] Batch [299]	Speed: 3036.171931 samples/sec	accuracy=0.464465	lr=0.099121
Epoch[1] Batch [349]	Speed: 2802.309256 samples/sec	accuracy=0.463230	lr=0.102318
Epoch[1] Batch [399]	Speed: 2826.982506 samples/sec	accuracy=0.461163	lr=0.105516
Epoch[1] Batch [449]	Speed: 2872.073499 samples/sec	accuracy=0.459219	lr=0.108713
Epoch[1] Batch [499]	Speed: 2879.406376 samples/sec	accuracy=0.457377	lr=0.111910
Epoch[1] Batch [549]	Speed: 2866.469922 samples/sec	accuracy=0.455611	lr=0.115108
Epoch[1] Batch [599]	Speed: 2841.262688 samples/sec	accuracy=0.454024	lr=0.118305
Epoch[1] Batch [649]	Speed: 2871.097641 samples/sec	accuracy=0.452720	lr=0.121503
Epoch[1] Batch [699]	Speed: 2745.026017 samples/sec	accuracy=0.451469	lr=0.124700
Epoch[1] Batch [749]	Speed: 2913.215416 samples/sec	accuracy=0.450301	lr=0.127898
Epoch[1] Batch [799]	Speed: 2878.849565 samples/sec	accuracy=0.449300	lr=0.131095
Epoch[1] Batch [849]	Speed: 2870.308002 samples/sec	accuracy=0.448539	lr=0.134293
Epoch[1] Batch [899]	Speed: 2889.468113 samples/sec	accuracy=0.447825	lr=0.137490
Epoch[1] Batch [949]	Speed: 3008.750477 samples/sec	accuracy=0.447119	lr=0.140687
Epoch[1] Batch [999]	Speed: 2868.011207 samples/sec	accuracy=0.446631	lr=0.143885
Epoch[1] Batch [1049]	Speed: 2805.847218 samples/sec	accuracy=0.446084	lr=0.147082
Epoch[1] Batch [1099]	Speed: 2831.415372 samples/sec	accuracy=0.445708	lr=0.150280
Epoch[1] Batch [1149]	Speed: 2806.200303 samples/sec	accuracy=0.445462	lr=0.153477
Epoch[1] Batch [1199]	Speed: 2897.507958 samples/sec	accuracy=0.445264	lr=0.156675
Epoch[1] Batch [1249]	Speed: 2951.576011 samples/sec	accuracy=0.445168	lr=0.159872
[Epoch 1] training: accuracy=0.445165
[Epoch 1] speed: 2878 samples/sec	time cost: 454.974384
[Epoch 1] validation: err-top1=0.792471 err-top5=0.574259
Epoch[2] Batch [49]	Speed: 2923.752353 samples/sec	accuracy=0.444379	lr=0.163133
Epoch[2] Batch [99]	Speed: 2938.883509 samples/sec	accuracy=0.445510	lr=0.166331
Epoch[2] Batch [149]	Speed: 2772.240188 samples/sec	accuracy=0.446071	lr=0.169528
Epoch[2] Batch [199]	Speed: 2863.918758 samples/sec	accuracy=0.446788	lr=0.172726
Epoch[2] Batch [249]	Speed: 2977.357668 samples/sec	accuracy=0.447439	lr=0.175923
Epoch[2] Batch [299]	Speed: 2987.855008 samples/sec	accuracy=0.448176	lr=0.179121
Epoch[2] Batch [349]	Speed: 2876.530993 samples/sec	accuracy=0.449019	lr=0.182318
Epoch[2] Batch [399]	Speed: 2940.884214 samples/sec	accuracy=0.449968	lr=0.185516
Epoch[2] Batch [449]	Speed: 2873.661271 samples/sec	accuracy=0.451196	lr=0.188713
Epoch[2] Batch [499]	Speed: 2991.695517 samples/sec	accuracy=0.452150	lr=0.191910
Epoch[2] Batch [549]	Speed: 2830.621588 samples/sec	accuracy=0.453129	lr=0.195108
Epoch[2] Batch [599]	Speed: 3046.567091 samples/sec	accuracy=0.454105	lr=0.198305
Epoch[2] Batch [649]	Speed: 2887.902405 samples/sec	accuracy=0.455131	lr=0.201503
Epoch[2] Batch [699]	Speed: 2984.266477 samples/sec	accuracy=0.456168	lr=0.204700
Epoch[2] Batch [749]	Speed: 2954.919024 samples/sec	accuracy=0.457299	lr=0.207898
Epoch[2] Batch [799]	Speed: 2938.740135 samples/sec	accuracy=0.458326	lr=0.211095
Epoch[2] Batch [849]	Speed: 2834.977368 samples/sec	accuracy=0.459388	lr=0.214293
Epoch[2] Batch [899]	Speed: 2966.420183 samples/sec	accuracy=0.460466	lr=0.217490
Epoch[2] Batch [949]	Speed: 2943.338148 samples/sec	accuracy=0.461502	lr=0.220687
Epoch[2] Batch [999]	Speed: 2801.638648 samples/sec	accuracy=0.462580	lr=0.223885
Epoch[2] Batch [1049]	Speed: 2937.323426 samples/sec	accuracy=0.463707	lr=0.227082
Epoch[2] Batch [1099]	Speed: 2928.241778 samples/sec	accuracy=0.464800	lr=0.230280
Epoch[2] Batch [1149]	Speed: 3050.208805 samples/sec	accuracy=0.465863	lr=0.233477
Epoch[2] Batch [1199]	Speed: 2917.868870 samples/sec	accuracy=0.466900	lr=0.236675
Epoch[2] Batch [1249]	Speed: 3210.329053 samples/sec	accuracy=0.468019	lr=0.239872
[Epoch 2] training: accuracy=0.468039
[Epoch 2] speed: 2930 samples/sec	time cost: 447.150278
[Epoch 2] validation: err-top1=0.697883 err-top5=0.444097
Epoch[3] Batch [49]	Speed: 2880.487996 samples/sec	accuracy=0.494442	lr=0.243133
Epoch[3] Batch [99]	Speed: 2815.539336 samples/sec	accuracy=0.495632	lr=0.246331
Epoch[3] Batch [149]	Speed: 2950.782116 samples/sec	accuracy=0.496327	lr=0.249528
Epoch[3] Batch [199]	Speed: 2948.797484 samples/sec	accuracy=0.497139	lr=0.252726
Epoch[3] Batch [249]	Speed: 3030.683834 samples/sec	accuracy=0.497934	lr=0.255923
Epoch[3] Batch [299]	Speed: 2841.367423 samples/sec	accuracy=0.498852	lr=0.259121
Epoch[3] Batch [349]	Speed: 2954.584313 samples/sec	accuracy=0.499788	lr=0.262318
Epoch[3] Batch [399]	Speed: 2943.007386 samples/sec	accuracy=0.500572	lr=0.265516
Epoch[3] Batch [449]	Speed: 2817.745838 samples/sec	accuracy=0.501529	lr=0.268713
Epoch[3] Batch [499]	Speed: 2896.254729 samples/sec	accuracy=0.502546	lr=0.271910
Epoch[3] Batch [549]	Speed: 2854.399487 samples/sec	accuracy=0.503625	lr=0.275108
Epoch[3] Batch [599]	Speed: 2964.094375 samples/sec	accuracy=0.504530	lr=0.278305
Epoch[3] Batch [649]	Speed: 2856.708801 samples/sec	accuracy=0.505417	lr=0.281503
Epoch[3] Batch [699]	Speed: 2745.381157 samples/sec	accuracy=0.506235	lr=0.284700
Epoch[3] Batch [749]	Speed: 2841.849053 samples/sec	accuracy=0.507107	lr=0.287898
Epoch[3] Batch [799]	Speed: 2865.085783 samples/sec	accuracy=0.507991	lr=0.291095
Epoch[3] Batch [849]	Speed: 2796.116053 samples/sec	accuracy=0.508951	lr=0.294293
Epoch[3] Batch [899]	Speed: 2879.102911 samples/sec	accuracy=0.509826	lr=0.297490
Epoch[3] Batch [949]	Speed: 2943.114633 samples/sec	accuracy=0.510802	lr=0.300687
Epoch[3] Batch [999]	Speed: 2981.523284 samples/sec	accuracy=0.511721	lr=0.303885
Epoch[3] Batch [1049]	Speed: 2831.927357 samples/sec	accuracy=0.512640	lr=0.307082
Epoch[3] Batch [1099]	Speed: 2823.427159 samples/sec	accuracy=0.513593	lr=0.310280
Epoch[3] Batch [1149]	Speed: 2934.127520 samples/sec	accuracy=0.514454	lr=0.313477
Epoch[3] Batch [1199]	Speed: 2804.170637 samples/sec	accuracy=0.515285	lr=0.316675
Epoch[3] Batch [1249]	Speed: 2942.922973 samples/sec	accuracy=0.516196	lr=0.319872
[Epoch 3] training: accuracy=0.516213
[Epoch 3] speed: 2882 samples/sec	time cost: 454.288167
[Epoch 3] validation: err-top1=0.587233 err-top5=0.323322
Epoch[4] Batch [49]	Speed: 2950.748098 samples/sec	accuracy=0.537101	lr=0.323133
Epoch[4] Batch [99]	Speed: 2977.594547 samples/sec	accuracy=0.537991	lr=0.326331
Epoch[4] Batch [149]	Speed: 2801.759453 samples/sec	accuracy=0.538890	lr=0.329528
Epoch[4] Batch [199]	Speed: 2901.271653 samples/sec	accuracy=0.539732	lr=0.332726
Epoch[4] Batch [249]	Speed: 2820.703970 samples/sec	accuracy=0.540852	lr=0.335923
Epoch[4] Batch [299]	Speed: 2794.099088 samples/sec	accuracy=0.541474	lr=0.339121
Epoch[4] Batch [349]	Speed: 2949.364996 samples/sec	accuracy=0.542560	lr=0.342318
Epoch[4] Batch [399]	Speed: 2830.122011 samples/sec	accuracy=0.543363	lr=0.345516
Epoch[4] Batch [449]	Speed: 2924.068966 samples/sec	accuracy=0.544017	lr=0.348713
Epoch[4] Batch [499]	Speed: 2949.397118 samples/sec	accuracy=0.544727	lr=0.351910
Epoch[4] Batch [549]	Speed: 2945.579646 samples/sec	accuracy=0.545679	lr=0.355108
Epoch[4] Batch [599]	Speed: 2854.538241 samples/sec	accuracy=0.546432	lr=0.358305
Epoch[4] Batch [649]	Speed: 2950.939604 samples/sec	accuracy=0.547139	lr=0.361503
Epoch[4] Batch [699]	Speed: 2848.765090 samples/sec	accuracy=0.547807	lr=0.364700
Epoch[4] Batch [749]	Speed: 2978.895950 samples/sec	accuracy=0.548631	lr=0.367898
Epoch[4] Batch [799]	Speed: 2982.220868 samples/sec	accuracy=0.549302	lr=0.371095
Epoch[4] Batch [849]	Speed: 2828.268157 samples/sec	accuracy=0.549978	lr=0.374293
Epoch[4] Batch [899]	Speed: 2936.384436 samples/sec	accuracy=0.550712	lr=0.377490
Epoch[4] Batch [949]	Speed: 2912.627400 samples/sec	accuracy=0.551450	lr=0.380687
Epoch[4] Batch [999]	Speed: 2922.456676 samples/sec	accuracy=0.552204	lr=0.383885
Epoch[4] Batch [1049]	Speed: 2871.616667 samples/sec	accuracy=0.552898	lr=0.387082
Epoch[4] Batch [1099]	Speed: 2933.422319 samples/sec	accuracy=0.553594	lr=0.390280
Epoch[4] Batch [1149]	Speed: 2945.864836 samples/sec	accuracy=0.554362	lr=0.393477
Epoch[4] Batch [1199]	Speed: 2828.567407 samples/sec	accuracy=0.555117	lr=0.396675
Epoch[4] Batch [1249]	Speed: 2976.544812 samples/sec	accuracy=0.555806	lr=0.399872
[Epoch 4] training: accuracy=0.555827
[Epoch 4] speed: 2901 samples/sec	time cost: 451.295801
[Epoch 4] validation: err-top1=0.568339 err-top5=0.299007
Epoch[5] Batch [49]	Speed: 2975.897098 samples/sec	accuracy=0.571855	lr=0.400000
Epoch[5] Batch [99]	Speed: 2868.284333 samples/sec	accuracy=0.572990	lr=0.400000
Epoch[5] Batch [149]	Speed: 2846.201986 samples/sec	accuracy=0.573392	lr=0.399999
Epoch[5] Batch [199]	Speed: 2823.742020 samples/sec	accuracy=0.574395	lr=0.399998
Epoch[5] Batch [249]	Speed: 2777.665367 samples/sec	accuracy=0.575020	lr=0.399997
Epoch[5] Batch [299]	Speed: 2900.510030 samples/sec	accuracy=0.575863	lr=0.399996
Epoch[5] Batch [349]	Speed: 2869.007122 samples/sec	accuracy=0.576509	lr=0.399994
Epoch[5] Batch [399]	Speed: 2930.939761 samples/sec	accuracy=0.577416	lr=0.399992
Epoch[5] Batch [449]	Speed: 2849.231124 samples/sec	accuracy=0.577987	lr=0.399990
Epoch[5] Batch [499]	Speed: 2924.565024 samples/sec	accuracy=0.578694	lr=0.399988
Epoch[5] Batch [549]	Speed: 2878.446749 samples/sec	accuracy=0.579285	lr=0.399986
Epoch[5] Batch [599]	Speed: 2778.303698 samples/sec	accuracy=0.579902	lr=0.399983
Epoch[5] Batch [649]	Speed: 2853.978147 samples/sec	accuracy=0.580445	lr=0.399980
Epoch[5] Batch [699]	Speed: 2827.195428 samples/sec	accuracy=0.581015	lr=0.399977
Epoch[5] Batch [749]	Speed: 2840.622081 samples/sec	accuracy=0.581566	lr=0.399973
Epoch[5] Batch [799]	Speed: 2854.372739 samples/sec	accuracy=0.582198	lr=0.399970
Epoch[5] Batch [849]	Speed: 2820.338521 samples/sec	accuracy=0.582852	lr=0.399966
Epoch[5] Batch [899]	Speed: 2862.448445 samples/sec	accuracy=0.583473	lr=0.399961
Epoch[5] Batch [949]	Speed: 2841.713711 samples/sec	accuracy=0.584149	lr=0.399957
Epoch[5] Batch [999]	Speed: 2920.272735 samples/sec	accuracy=0.584815	lr=0.399952
Epoch[5] Batch [1049]	Speed: 2779.134187 samples/sec	accuracy=0.585457	lr=0.399948
Epoch[5] Batch [1099]	Speed: 2846.128240 samples/sec	accuracy=0.586116	lr=0.399942
Epoch[5] Batch [1149]	Speed: 2870.149528 samples/sec	accuracy=0.586819	lr=0.399937
Epoch[5] Batch [1199]	Speed: 2824.876078 samples/sec	accuracy=0.587489	lr=0.399931
Epoch[5] Batch [1249]	Speed: 3002.390883 samples/sec	accuracy=0.588095	lr=0.399926
[Epoch 5] training: accuracy=0.588106
[Epoch 5] speed: 2859 samples/sec	time cost: 457.468876
[Epoch 5] validation: err-top1=0.488525 err-top5=0.235311
Epoch[6] Batch [49]	Speed: 2937.137540 samples/sec	accuracy=0.601116	lr=0.399919
Epoch[6] Batch [99]	Speed: 2833.700157 samples/sec	accuracy=0.602522	lr=0.399913
Epoch[6] Batch [149]	Speed: 2849.206401 samples/sec	accuracy=0.602565	lr=0.399907
Epoch[6] Batch [199]	Speed: 2820.901829 samples/sec	accuracy=0.602636	lr=0.399900
Epoch[6] Batch [249]	Speed: 2817.282543 samples/sec	accuracy=0.602989	lr=0.399893
Epoch[6] Batch [299]	Speed: 2851.041939 samples/sec	accuracy=0.603520	lr=0.399885
Epoch[6] Batch [349]	Speed: 2859.739602 samples/sec	accuracy=0.604324	lr=0.399878
Epoch[6] Batch [399]	Speed: 2873.129128 samples/sec	accuracy=0.605083	lr=0.399870
Epoch[6] Batch [449]	Speed: 2860.724330 samples/sec	accuracy=0.605674	lr=0.399862
Epoch[6] Batch [499]	Speed: 2857.129312 samples/sec	accuracy=0.606152	lr=0.399854
Epoch[6] Batch [549]	Speed: 2855.521819 samples/sec	accuracy=0.606702	lr=0.399846
Epoch[6] Batch [599]	Speed: 2815.480349 samples/sec	accuracy=0.607319	lr=0.399837
Epoch[6] Batch [649]	Speed: 2861.465695 samples/sec	accuracy=0.607882	lr=0.399828
Epoch[6] Batch [699]	Speed: 2814.251141 samples/sec	accuracy=0.608384	lr=0.399819
Epoch[6] Batch [749]	Speed: 2878.539118 samples/sec	accuracy=0.609040	lr=0.399809
Epoch[6] Batch [799]	Speed: 2846.935691 samples/sec	accuracy=0.609576	lr=0.399800
Epoch[6] Batch [849]	Speed: 2935.804209 samples/sec	accuracy=0.610147	lr=0.399790
Epoch[6] Batch [899]	Speed: 2881.519545 samples/sec	accuracy=0.610844	lr=0.399780
Epoch[6] Batch [949]	Speed: 3022.785045 samples/sec	accuracy=0.611476	lr=0.399769
Epoch[6] Batch [999]	Speed: 2834.765143 samples/sec	accuracy=0.611989	lr=0.399759
Epoch[6] Batch [1049]	Speed: 2825.192155 samples/sec	accuracy=0.612499	lr=0.399748
Epoch[6] Batch [1099]	Speed: 2838.370654 samples/sec	accuracy=0.613044	lr=0.399737
Epoch[6] Batch [1149]	Speed: 2810.116218 samples/sec	accuracy=0.613679	lr=0.399725
Epoch[6] Batch [1199]	Speed: 2816.318811 samples/sec	accuracy=0.614219	lr=0.399714
Epoch[6] Batch [1249]	Speed: 2908.541405 samples/sec	accuracy=0.614742	lr=0.399702
[Epoch 6] training: accuracy=0.614751
[Epoch 6] speed: 2857 samples/sec	time cost: 458.636702
[Epoch 6] validation: err-top1=0.481206 err-top5=0.227220
Epoch[7] Batch [49]	Speed: 2948.441205 samples/sec	accuracy=0.626702	lr=0.399690
Epoch[7] Batch [99]	Speed: 2891.845642 samples/sec	accuracy=0.627123	lr=0.399677
Epoch[7] Batch [149]	Speed: 2900.319570 samples/sec	accuracy=0.627450	lr=0.399665
Epoch[7] Batch [199]	Speed: 2886.074462 samples/sec	accuracy=0.627430	lr=0.399652
Epoch[7] Batch [249]	Speed: 2832.153387 samples/sec	accuracy=0.627473	lr=0.399639
Epoch[7] Batch [299]	Speed: 2805.556018 samples/sec	accuracy=0.628007	lr=0.399626
Epoch[7] Batch [349]	Speed: 2912.229334 samples/sec	accuracy=0.628652	lr=0.399613
Epoch[7] Batch [399]	Speed: 2932.854036 samples/sec	accuracy=0.629126	lr=0.399599
Epoch[7] Batch [449]	Speed: 2923.275592 samples/sec	accuracy=0.629594	lr=0.399585
Epoch[7] Batch [499]	Speed: 2914.044422 samples/sec	accuracy=0.629951	lr=0.399571
Epoch[7] Batch [549]	Speed: 2939.468133 samples/sec	accuracy=0.630279	lr=0.399556
Epoch[7] Batch [599]	Speed: 2911.757152 samples/sec	accuracy=0.630769	lr=0.399542
Epoch[7] Batch [649]	Speed: 2862.205727 samples/sec	accuracy=0.631175	lr=0.399527
Epoch[7] Batch [699]	Speed: 2809.263253 samples/sec	accuracy=0.631759	lr=0.399512
Epoch[7] Batch [749]	Speed: 2864.060884 samples/sec	accuracy=0.632121	lr=0.399496
Epoch[7] Batch [799]	Speed: 2742.828615 samples/sec	accuracy=0.632557	lr=0.399481
Epoch[7] Batch [849]	Speed: 2849.438262 samples/sec	accuracy=0.633096	lr=0.399465
Epoch[7] Batch [899]	Speed: 2935.467594 samples/sec	accuracy=0.633630	lr=0.399449
Epoch[7] Batch [949]	Speed: 2920.693341 samples/sec	accuracy=0.634047	lr=0.399432
Epoch[7] Batch [999]	Speed: 2874.003090 samples/sec	accuracy=0.634411	lr=0.399416
Epoch[7] Batch [1049]	Speed: 2893.654430 samples/sec	accuracy=0.634881	lr=0.399399
Epoch[7] Batch [1099]	Speed: 2992.850605 samples/sec	accuracy=0.635209	lr=0.399382
Epoch[7] Batch [1149]	Speed: 2882.787102 samples/sec	accuracy=0.635611	lr=0.399365
Epoch[7] Batch [1199]	Speed: 2851.881683 samples/sec	accuracy=0.636017	lr=0.399347
Epoch[7] Batch [1249]	Speed: 2920.933288 samples/sec	accuracy=0.636330	lr=0.399329
[Epoch 7] training: accuracy=0.636340
[Epoch 7] speed: 2885 samples/sec	time cost: 454.504968
[Epoch 7] validation: err-top1=0.434670 err-top5=0.192323
Epoch[8] Batch [49]	Speed: 2908.870611 samples/sec	accuracy=0.643141	lr=0.399311
Epoch[8] Batch [99]	Speed: 2793.314823 samples/sec	accuracy=0.643059	lr=0.399293
Epoch[8] Batch [149]	Speed: 2846.003767 samples/sec	accuracy=0.643293	lr=0.399274
Epoch[8] Batch [199]	Speed: 2896.772732 samples/sec	accuracy=0.644290	lr=0.399256
Epoch[8] Batch [249]	Speed: 2800.323773 samples/sec	accuracy=0.644625	lr=0.399237
Epoch[8] Batch [299]	Speed: 2816.564521 samples/sec	accuracy=0.644827	lr=0.399218
Epoch[8] Batch [349]	Speed: 2784.275960 samples/sec	accuracy=0.645571	lr=0.399198
Epoch[8] Batch [399]	Speed: 2808.559487 samples/sec	accuracy=0.646008	lr=0.399179
Epoch[8] Batch [449]	Speed: 2963.509854 samples/sec	accuracy=0.646542	lr=0.399159
Epoch[8] Batch [499]	Speed: 2931.515666 samples/sec	accuracy=0.646956	lr=0.399138
Epoch[8] Batch [549]	Speed: 2884.638006 samples/sec	accuracy=0.647246	lr=0.399118
Epoch[8] Batch [599]	Speed: 2834.029843 samples/sec	accuracy=0.647579	lr=0.399098
Epoch[8] Batch [649]	Speed: 2887.444755 samples/sec	accuracy=0.647936	lr=0.399077
Epoch[8] Batch [699]	Speed: 2776.084010 samples/sec	accuracy=0.648373	lr=0.399056
Epoch[8] Batch [749]	Speed: 2833.016536 samples/sec	accuracy=0.648756	lr=0.399034
Epoch[8] Batch [799]	Speed: 2817.523949 samples/sec	accuracy=0.649189	lr=0.399013
Epoch[8] Batch [849]	Speed: 2858.585222 samples/sec	accuracy=0.649478	lr=0.398991
Epoch[8] Batch [899]	Speed: 2835.517448 samples/sec	accuracy=0.649861	lr=0.398969
Epoch[8] Batch [949]	Speed: 2922.434127 samples/sec	accuracy=0.650204	lr=0.398947
Epoch[8] Batch [999]	Speed: 2788.832342 samples/sec	accuracy=0.650474	lr=0.398924
Epoch[8] Batch [1049]	Speed: 2914.692587 samples/sec	accuracy=0.650834	lr=0.398901
Epoch[8] Batch [1099]	Speed: 2919.149846 samples/sec	accuracy=0.651141	lr=0.398878
Epoch[8] Batch [1149]	Speed: 2929.483125 samples/sec	accuracy=0.651555	lr=0.398855
Epoch[8] Batch [1199]	Speed: 2875.518105 samples/sec	accuracy=0.651948	lr=0.398832
Epoch[8] Batch [1249]	Speed: 3040.708276 samples/sec	accuracy=0.652301	lr=0.398808
[Epoch 8] training: accuracy=0.652305
[Epoch 8] speed: 2863 samples/sec	time cost: 457.631836
[Epoch 8] validation: err-top1=0.455955 err-top5=0.207370
Epoch[9] Batch [49]	Speed: 3004.302359 samples/sec	accuracy=0.657645	lr=0.398784
Epoch[9] Batch [99]	Speed: 2887.502410 samples/sec	accuracy=0.658160	lr=0.398760
Epoch[9] Batch [149]	Speed: 2910.560917 samples/sec	accuracy=0.658225	lr=0.398735
Epoch[9] Batch [199]	Speed: 2770.169744 samples/sec	accuracy=0.658653	lr=0.398710
Epoch[9] Batch [249]	Speed: 2832.490559 samples/sec	accuracy=0.659384	lr=0.398686
Epoch[9] Batch [299]	Speed: 2853.488871 samples/sec	accuracy=0.659474	lr=0.398660
Epoch[9] Batch [349]	Speed: 2943.526998 samples/sec	accuracy=0.659797	lr=0.398635
Epoch[9] Batch [399]	Speed: 2905.869501 samples/sec	accuracy=0.660376	lr=0.398610
Epoch[9] Batch [449]	Speed: 2781.882672 samples/sec	accuracy=0.660654	lr=0.398584
Epoch[9] Batch [499]	Speed: 2882.826227 samples/sec	accuracy=0.660888	lr=0.398558
Epoch[9] Batch [549]	Speed: 2759.688467 samples/sec	accuracy=0.661152	lr=0.398531
Epoch[9] Batch [599]	Speed: 2924.245476 samples/sec	accuracy=0.661416	lr=0.398505
Epoch[9] Batch [649]	Speed: 2827.386643 samples/sec	accuracy=0.661701	lr=0.398478
Epoch[9] Batch [699]	Speed: 2894.463834 samples/sec	accuracy=0.661917	lr=0.398451
Epoch[9] Batch [749]	Speed: 2960.007395 samples/sec	accuracy=0.662098	lr=0.398424
Epoch[9] Batch [799]	Speed: 2831.428550 samples/sec	accuracy=0.662226	lr=0.398396
Epoch[9] Batch [849]	Speed: 2868.107618 samples/sec	accuracy=0.662446	lr=0.398369
Epoch[9] Batch [899]	Speed: 2863.952445 samples/sec	accuracy=0.662762	lr=0.398341
Epoch[9] Batch [949]	Speed: 2906.609746 samples/sec	accuracy=0.663100	lr=0.398312
Epoch[9] Batch [999]	Speed: 2989.521266 samples/sec	accuracy=0.663327	lr=0.398284
Epoch[9] Batch [1049]	Speed: 2885.034300 samples/sec	accuracy=0.663695	lr=0.398255
Epoch[9] Batch [1099]	Speed: 2772.213813 samples/sec	accuracy=0.664079	lr=0.398226
Epoch[9] Batch [1149]	Speed: 2925.165680 samples/sec	accuracy=0.664370	lr=0.398197
Epoch[9] Batch [1199]	Speed: 2785.526086 samples/sec	accuracy=0.664640	lr=0.398168
Epoch[9] Batch [1249]	Speed: 3016.652735 samples/sec	accuracy=0.664920	lr=0.398138
[Epoch 9] training: accuracy=0.664931
[Epoch 9] speed: 2875 samples/sec	time cost: 455.474979
[Epoch 9] validation: err-top1=0.429448 err-top5=0.187998
Epoch[10] Batch [49]	Speed: 2957.355271 samples/sec	accuracy=0.667961	lr=0.398108
Epoch[10] Batch [99]	Speed: 2868.140562 samples/sec	accuracy=0.668164	lr=0.398078
Epoch[10] Batch [149]	Speed: 2856.045560 samples/sec	accuracy=0.667853	lr=0.398048
Epoch[10] Batch [199]	Speed: 2868.849903 samples/sec	accuracy=0.667704	lr=0.398017
Epoch[10] Batch [249]	Speed: 2934.659962 samples/sec	accuracy=0.668314	lr=0.397986
Epoch[10] Batch [299]	Speed: 2927.328297 samples/sec	accuracy=0.668875	lr=0.397955
Epoch[10] Batch [349]	Speed: 2995.295012 samples/sec	accuracy=0.669425	lr=0.397924
Epoch[10] Batch [399]	Speed: 2862.538226 samples/sec	accuracy=0.669911	lr=0.397892
Epoch[10] Batch [449]	Speed: 2921.159605 samples/sec	accuracy=0.670260	lr=0.397861
Epoch[10] Batch [499]	Speed: 2922.684860 samples/sec	accuracy=0.670653	lr=0.397829
Epoch[10] Batch [549]	Speed: 2857.397328 samples/sec	accuracy=0.671012	lr=0.397796
Epoch[10] Batch [599]	Speed: 2932.371259 samples/sec	accuracy=0.671324	lr=0.397764
Epoch[10] Batch [649]	Speed: 2864.166274 samples/sec	accuracy=0.671589	lr=0.397731
Epoch[10] Batch [699]	Speed: 2841.745110 samples/sec	accuracy=0.671897	lr=0.397698
Epoch[10] Batch [749]	Speed: 2794.974188 samples/sec	accuracy=0.672242	lr=0.397665
Epoch[10] Batch [799]	Speed: 2884.648584 samples/sec	accuracy=0.672409	lr=0.397632
Epoch[10] Batch [849]	Speed: 2860.519473 samples/sec	accuracy=0.672658	lr=0.397598
Epoch[10] Batch [899]	Speed: 2916.298972 samples/sec	accuracy=0.672919	lr=0.397564
Epoch[10] Batch [949]	Speed: 2894.925507 samples/sec	accuracy=0.673167	lr=0.397530
Epoch[10] Batch [999]	Speed: 2933.469362 samples/sec	accuracy=0.673374	lr=0.397496
Epoch[10] Batch [1049]	Speed: 2996.002400 samples/sec	accuracy=0.673589	lr=0.397461
Epoch[10] Batch [1099]	Speed: 2934.944446 samples/sec	accuracy=0.673809	lr=0.397427
Epoch[10] Batch [1149]	Speed: 2911.984695 samples/sec	accuracy=0.673995	lr=0.397392
Epoch[10] Batch [1199]	Speed: 2898.200454 samples/sec	accuracy=0.674216	lr=0.397356
Epoch[10] Batch [1249]	Speed: 2977.554501 samples/sec	accuracy=0.674415	lr=0.397321
[Epoch 10] training: accuracy=0.674421
[Epoch 10] speed: 2901 samples/sec	time cost: 451.478403
[Epoch 10] validation: err-top1=0.408721 err-top5=0.170819
Epoch[11] Batch [49]	Speed: 2992.075416 samples/sec	accuracy=0.677436	lr=0.397284
Epoch[11] Batch [99]	Speed: 2806.264587 samples/sec	accuracy=0.676317	lr=0.397248
Epoch[11] Batch [149]	Speed: 2807.815103 samples/sec	accuracy=0.677053	lr=0.397212
Epoch[11] Batch [199]	Speed: 2986.778378 samples/sec	accuracy=0.677904	lr=0.397176
Epoch[11] Batch [249]	Speed: 2877.185241 samples/sec	accuracy=0.677920	lr=0.397139
Epoch[11] Batch [299]	Speed: 2982.749904 samples/sec	accuracy=0.678289	lr=0.397102
Epoch[11] Batch [349]	Speed: 2668.718240 samples/sec	accuracy=0.678586	lr=0.397065
Epoch[11] Batch [399]	Speed: 3042.067084 samples/sec	accuracy=0.678813	lr=0.397028
Epoch[11] Batch [449]	Speed: 2929.927495 samples/sec	accuracy=0.679145	lr=0.396990
Epoch[11] Batch [499]	Speed: 2902.913973 samples/sec	accuracy=0.679325	lr=0.396952
Epoch[11] Batch [549]	Speed: 2804.558462 samples/sec	accuracy=0.679467	lr=0.396914
Epoch[11] Batch [599]	Speed: 2846.794165 samples/sec	accuracy=0.679741	lr=0.396876
Epoch[11] Batch [649]	Speed: 2863.422250 samples/sec	accuracy=0.679857	lr=0.396837
Epoch[11] Batch [699]	Speed: 2972.617078 samples/sec	accuracy=0.679919	lr=0.396798
Epoch[11] Batch [749]	Speed: 2758.519176 samples/sec	accuracy=0.680193	lr=0.396759
Epoch[11] Batch [799]	Speed: 2801.683460 samples/sec	accuracy=0.680490	lr=0.396720
Epoch[11] Batch [849]	Speed: 2949.274912 samples/sec	accuracy=0.680737	lr=0.396680
Epoch[11] Batch [899]	Speed: 2810.085550 samples/sec	accuracy=0.680958	lr=0.396641
Epoch[11] Batch [949]	Speed: 2859.968685 samples/sec	accuracy=0.681188	lr=0.396601
Epoch[11] Batch [999]	Speed: 2994.287035 samples/sec	accuracy=0.681506	lr=0.396561
Epoch[11] Batch [1049]	Speed: 2952.041068 samples/sec	accuracy=0.681664	lr=0.396520
Epoch[11] Batch [1099]	Speed: 2978.616062 samples/sec	accuracy=0.681890	lr=0.396479
Epoch[11] Batch [1149]	Speed: 3027.047176 samples/sec	accuracy=0.682165	lr=0.396439
Epoch[11] Batch [1199]	Speed: 2906.158144 samples/sec	accuracy=0.682392	lr=0.396397
Epoch[11] Batch [1249]	Speed: 2955.801272 samples/sec	accuracy=0.682573	lr=0.396356
[Epoch 11] training: accuracy=0.682575
[Epoch 11] speed: 2894 samples/sec	time cost: 452.936767
[Epoch 11] validation: err-top1=0.417440 err-top5=0.179362
Epoch[12] Batch [49]	Speed: 3107.648082 samples/sec	accuracy=0.685819	lr=0.396314
Epoch[12] Batch [99]	Speed: 3006.273504 samples/sec	accuracy=0.685721	lr=0.396272
Epoch[12] Batch [149]	Speed: 3006.643856 samples/sec	accuracy=0.685738	lr=0.396230
Epoch[12] Batch [199]	Speed: 2949.736287 samples/sec	accuracy=0.685423	lr=0.396187
Epoch[12] Batch [249]	Speed: 2953.861729 samples/sec	accuracy=0.685516	lr=0.396145
Epoch[12] Batch [299]	Speed: 2986.498751 samples/sec	accuracy=0.685827	lr=0.396102
Epoch[12] Batch [349]	Speed: 2841.575411 samples/sec	accuracy=0.686011	lr=0.396059
Epoch[12] Batch [399]	Speed: 2885.187290 samples/sec	accuracy=0.686239	lr=0.396016
Epoch[12] Batch [449]	Speed: 3106.085357 samples/sec	accuracy=0.686305	lr=0.395972
Epoch[12] Batch [499]	Speed: 2980.519424 samples/sec	accuracy=0.686363	lr=0.395928
Epoch[12] Batch [549]	Speed: 2884.487360 samples/sec	accuracy=0.686525	lr=0.395885
Epoch[12] Batch [599]	Speed: 2862.991219 samples/sec	accuracy=0.686662	lr=0.395840
Epoch[12] Batch [649]	Speed: 2906.738278 samples/sec	accuracy=0.686852	lr=0.395796
Epoch[12] Batch [699]	Speed: 2758.449478 samples/sec	accuracy=0.687174	lr=0.395751
Epoch[12] Batch [749]	Speed: 2988.506025 samples/sec	accuracy=0.687428	lr=0.395706
Epoch[12] Batch [799]	Speed: 2898.644930 samples/sec	accuracy=0.687595	lr=0.395661
Epoch[12] Batch [849]	Speed: 2931.706684 samples/sec	accuracy=0.687839	lr=0.395616
Epoch[12] Batch [899]	Speed: 2911.932692 samples/sec	accuracy=0.688030	lr=0.395570
Epoch[12] Batch [949]	Speed: 2984.108315 samples/sec	accuracy=0.688247	lr=0.395524
Epoch[12] Batch [999]	Speed: 2881.234306 samples/sec	accuracy=0.688518	lr=0.395478
Epoch[12] Batch [1049]	Speed: 2962.052417 samples/sec	accuracy=0.688758	lr=0.395432
Epoch[12] Batch [1099]	Speed: 2866.957230 samples/sec	accuracy=0.688904	lr=0.395386
Epoch[12] Batch [1149]	Speed: 2986.162910 samples/sec	accuracy=0.689145	lr=0.395339
Epoch[12] Batch [1199]	Speed: 2869.449628 samples/sec	accuracy=0.689342	lr=0.395292
Epoch[12] Batch [1249]	Speed: 2934.725092 samples/sec	accuracy=0.689569	lr=0.395245
[Epoch 12] training: accuracy=0.689573
[Epoch 12] speed: 2933 samples/sec	time cost: 447.122280
[Epoch 12] validation: err-top1=0.427356 err-top5=0.185647
Epoch[13] Batch [49]	Speed: 3006.017397 samples/sec	accuracy=0.690145	lr=0.395196
Epoch[13] Batch [99]	Speed: 3021.354687 samples/sec	accuracy=0.689789	lr=0.395149
Epoch[13] Batch [149]	Speed: 2876.964029 samples/sec	accuracy=0.690321	lr=0.395101
Epoch[13] Batch [199]	Speed: 2986.073221 samples/sec	accuracy=0.691100	lr=0.395052
Epoch[13] Batch [249]	Speed: 2883.652856 samples/sec	accuracy=0.691496	lr=0.395004
Epoch[13] Batch [299]	Speed: 3002.561317 samples/sec	accuracy=0.691729	lr=0.394955
Epoch[13] Batch [349]	Speed: 2905.059010 samples/sec	accuracy=0.691963	lr=0.394907
Epoch[13] Batch [399]	Speed: 2810.149718 samples/sec	accuracy=0.692159	lr=0.394858
Epoch[13] Batch [449]	Speed: 2928.530211 samples/sec	accuracy=0.692472	lr=0.394808
Epoch[13] Batch [499]	Speed: 2830.492312 samples/sec	accuracy=0.692860	lr=0.394759
Epoch[13] Batch [549]	Speed: 2814.617485 samples/sec	accuracy=0.693178	lr=0.394709
Epoch[13] Batch [599]	Speed: 2814.218502 samples/sec	accuracy=0.693359	lr=0.394659
Epoch[13] Batch [649]	Speed: 2914.846245 samples/sec	accuracy=0.693680	lr=0.394609
Epoch[13] Batch [699]	Speed: 3002.483654 samples/sec	accuracy=0.693817	lr=0.394558
Epoch[13] Batch [749]	Speed: 2817.379123 samples/sec	accuracy=0.694073	lr=0.394507
Epoch[13] Batch [799]	Speed: 2898.100014 samples/sec	accuracy=0.694292	lr=0.394457
Epoch[13] Batch [849]	Speed: 2962.541217 samples/sec	accuracy=0.694539	lr=0.394405
Epoch[13] Batch [899]	Speed: 2826.054709 samples/sec	accuracy=0.694740	lr=0.394354
Epoch[13] Batch [949]	Speed: 2843.570516 samples/sec	accuracy=0.694925	lr=0.394302
Epoch[13] Batch [999]	Speed: 2742.151679 samples/sec	accuracy=0.695085	lr=0.394250
Epoch[13] Batch [1049]	Speed: 2932.559225 samples/sec	accuracy=0.695257	lr=0.394198
Epoch[13] Batch [1099]	Speed: 2840.159422 samples/sec	accuracy=0.695411	lr=0.394146
Epoch[13] Batch [1149]	Speed: 2913.518130 samples/sec	accuracy=0.695580	lr=0.394093
Epoch[13] Batch [1199]	Speed: 2930.935681 samples/sec	accuracy=0.695768	lr=0.394041
Epoch[13] Batch [1249]	Speed: 3019.213108 samples/sec	accuracy=0.695931	lr=0.393988
[Epoch 13] training: accuracy=0.695932
[Epoch 13] speed: 2896 samples/sec	time cost: 452.359437
[Epoch 13] validation: err-top1=0.419384 err-top5=0.179588
Epoch[14] Batch [49]	Speed: 2959.134256 samples/sec	accuracy=0.696827	lr=0.393933
Epoch[14] Batch [99]	Speed: 2853.408112 samples/sec	accuracy=0.696660	lr=0.393880
Epoch[14] Batch [149]	Speed: 2935.662499 samples/sec	accuracy=0.696487	lr=0.393826
Epoch[14] Batch [199]	Speed: 2929.404841 samples/sec	accuracy=0.696851	lr=0.393772
Epoch[14] Batch [249]	Speed: 2884.389728 samples/sec	accuracy=0.697170	lr=0.393718
Epoch[14] Batch [299]	Speed: 3005.093232 samples/sec	accuracy=0.697327	lr=0.393664
Epoch[14] Batch [349]	Speed: 2882.542509 samples/sec	accuracy=0.697708	lr=0.393609
Epoch[14] Batch [399]	Speed: 2811.819358 samples/sec	accuracy=0.698000	lr=0.393554
Epoch[14] Batch [449]	Speed: 2853.621355 samples/sec	accuracy=0.698068	lr=0.393499
Epoch[14] Batch [499]	Speed: 2835.708892 samples/sec	accuracy=0.697980	lr=0.393444
Epoch[14] Batch [549]	Speed: 2822.804697 samples/sec	accuracy=0.698254	lr=0.393388
Epoch[14] Batch [599]	Speed: 2882.752235 samples/sec	accuracy=0.698343	lr=0.393332
Epoch[14] Batch [649]	Speed: 2874.552218 samples/sec	accuracy=0.698462	lr=0.393276
Epoch[14] Batch [699]	Speed: 2836.469378 samples/sec	accuracy=0.698755	lr=0.393220
Epoch[14] Batch [749]	Speed: 2958.322597 samples/sec	accuracy=0.699043	lr=0.393163
Epoch[14] Batch [799]	Speed: 2901.624346 samples/sec	accuracy=0.699182	lr=0.393107
Epoch[14] Batch [849]	Speed: 2929.779117 samples/sec	accuracy=0.699371	lr=0.393050
Epoch[14] Batch [899]	Speed: 2951.956502 samples/sec	accuracy=0.699628	lr=0.392993
Epoch[14] Batch [949]	Speed: 2958.435610 samples/sec	accuracy=0.699812	lr=0.392935
Epoch[14] Batch [999]	Speed: 2992.876674 samples/sec	accuracy=0.699964	lr=0.392877
Epoch[14] Batch [1049]	Speed: 2790.081562 samples/sec	accuracy=0.700281	lr=0.392820
Epoch[14] Batch [1099]	Speed: 2871.507272 samples/sec	accuracy=0.700384	lr=0.392762
Epoch[14] Batch [1149]	Speed: 2817.303980 samples/sec	accuracy=0.700471	lr=0.392703
Epoch[14] Batch [1199]	Speed: 2986.276814 samples/sec	accuracy=0.700709	lr=0.392645
Epoch[14] Batch [1249]	Speed: 3022.477067 samples/sec	accuracy=0.700863	lr=0.392586
[Epoch 14] training: accuracy=0.700864
[Epoch 14] speed: 2898 samples/sec	time cost: 452.266569
[Epoch 14] validation: err-top1=0.407506 err-top5=0.168288
Epoch[15] Batch [49]	Speed: 2915.624561 samples/sec	accuracy=0.701616	lr=0.392526
Epoch[15] Batch [99]	Speed: 2920.788680 samples/sec	accuracy=0.702857	lr=0.392466
Epoch[15] Batch [149]	Speed: 2833.031262 samples/sec	accuracy=0.703310	lr=0.392407
Epoch[15] Batch [199]	Speed: 2842.093371 samples/sec	accuracy=0.703096	lr=0.392347
Epoch[15] Batch [249]	Speed: 2766.968991 samples/sec	accuracy=0.702992	lr=0.392287
Epoch[15] Batch [299]	Speed: 2895.068268 samples/sec	accuracy=0.702715	lr=0.392227
Epoch[15] Batch [349]	Speed: 3000.345553 samples/sec	accuracy=0.702604	lr=0.392167
Epoch[15] Batch [399]	Speed: 2906.798988 samples/sec	accuracy=0.702670	lr=0.392106
Epoch[15] Batch [449]	Speed: 2795.813910 samples/sec	accuracy=0.702793	lr=0.392045
Epoch[15] Batch [499]	Speed: 2916.288002 samples/sec	accuracy=0.703065	lr=0.391984
Epoch[15] Batch [549]	Speed: 2817.003560 samples/sec	accuracy=0.703382	lr=0.391923
Epoch[15] Batch [599]	Speed: 2805.745086 samples/sec	accuracy=0.703380	lr=0.391861
Epoch[15] Batch [649]	Speed: 2884.506539 samples/sec	accuracy=0.703572	lr=0.391799
Epoch[15] Batch [699]	Speed: 2853.804443 samples/sec	accuracy=0.703526	lr=0.391737
Epoch[15] Batch [749]	Speed: 2903.434322 samples/sec	accuracy=0.703798	lr=0.391675
Epoch[15] Batch [799]	Speed: 3017.240649 samples/sec	accuracy=0.704028	lr=0.391613
Epoch[15] Batch [849]	Speed: 2906.778882 samples/sec	accuracy=0.704146	lr=0.391550
Epoch[15] Batch [899]	Speed: 2807.858350 samples/sec	accuracy=0.704319	lr=0.391487
Epoch[15] Batch [949]	Speed: 2894.393301 samples/sec	accuracy=0.704434	lr=0.391424
Epoch[15] Batch [999]	Speed: 2827.026495 samples/sec	accuracy=0.704509	lr=0.391361
Epoch[15] Batch [1049]	Speed: 2898.160402 samples/sec	accuracy=0.704662	lr=0.391297
Epoch[15] Batch [1099]	Speed: 2853.493193 samples/sec	accuracy=0.704818	lr=0.391233
Epoch[15] Batch [1149]	Speed: 2944.938673 samples/sec	accuracy=0.704971	lr=0.391169
Epoch[15] Batch [1199]	Speed: 2874.132948 samples/sec	accuracy=0.705088	lr=0.391105
Epoch[15] Batch [1249]	Speed: 3038.867687 samples/sec	accuracy=0.705269	lr=0.391040
[Epoch 15] training: accuracy=0.705275
[Epoch 15] speed: 2881 samples/sec	time cost: 454.418239
[Epoch 15] validation: err-top1=0.393674 err-top5=0.160116
Epoch[16] Batch [49]	Speed: 2834.823857 samples/sec	accuracy=0.705571	lr=0.390974
Epoch[16] Batch [99]	Speed: 2808.921118 samples/sec	accuracy=0.705713	lr=0.390909
Epoch[16] Batch [149]	Speed: 2948.581318 samples/sec	accuracy=0.706188	lr=0.390844
Epoch[16] Batch [199]	Speed: 2910.265285 samples/sec	accuracy=0.706302	lr=0.390779
Epoch[16] Batch [249]	Speed: 2861.301028 samples/sec	accuracy=0.706022	lr=0.390713
Epoch[16] Batch [299]	Speed: 2913.167518 samples/sec	accuracy=0.706119	lr=0.390647
Epoch[16] Batch [349]	Speed: 2882.371152 samples/sec	accuracy=0.706028	lr=0.390581
Epoch[16] Batch [399]	Speed: 2876.858003 samples/sec	accuracy=0.706120	lr=0.390515
Epoch[16] Batch [449]	Speed: 2913.528565 samples/sec	accuracy=0.706104	lr=0.390448
Epoch[16] Batch [499]	Speed: 3147.063812 samples/sec	accuracy=0.706284	lr=0.390381
Epoch[16] Batch [549]	Speed: 2754.974191 samples/sec	accuracy=0.706532	lr=0.390314
Epoch[16] Batch [599]	Speed: 2991.810760 samples/sec	accuracy=0.706799	lr=0.390247
Epoch[16] Batch [649]	Speed: 2788.203463 samples/sec	accuracy=0.706903	lr=0.390180
Epoch[16] Batch [699]	Speed: 2890.607304 samples/sec	accuracy=0.707100	lr=0.390112
Epoch[16] Batch [749]	Speed: 2930.712406 samples/sec	accuracy=0.707521	lr=0.390044
Epoch[16] Batch [799]	Speed: 2921.095592 samples/sec	accuracy=0.707729	lr=0.389976
Epoch[16] Batch [849]	Speed: 2964.069337 samples/sec	accuracy=0.707988	lr=0.389907
Epoch[16] Batch [899]	Speed: 2910.859252 samples/sec	accuracy=0.708133	lr=0.389839
Epoch[16] Batch [949]	Speed: 2950.086395 samples/sec	accuracy=0.708393	lr=0.389770
Epoch[16] Batch [999]	Speed: 2851.630870 samples/sec	accuracy=0.708440	lr=0.389701
Epoch[16] Batch [1049]	Speed: 2838.346757 samples/sec	accuracy=0.708564	lr=0.389632
Epoch[16] Batch [1099]	Speed: 2816.849181 samples/sec	accuracy=0.708743	lr=0.389562
Epoch[16] Batch [1149]	Speed: 2914.175986 samples/sec	accuracy=0.708863	lr=0.389492
Epoch[16] Batch [1199]	Speed: 2888.085334 samples/sec	accuracy=0.709009	lr=0.389422
Epoch[16] Batch [1249]	Speed: 3015.548522 samples/sec	accuracy=0.709128	lr=0.389352
[Epoch 16] training: accuracy=0.709130
[Epoch 16] speed: 2896 samples/sec	time cost: 452.234602
[Epoch 16] validation: err-top1=0.390485 err-top5=0.155951
Epoch[17] Batch [49]	Speed: 2964.670901 samples/sec	accuracy=0.709557	lr=0.389280
Epoch[17] Batch [99]	Speed: 2872.983064 samples/sec	accuracy=0.709938	lr=0.389210
Epoch[17] Batch [149]	Speed: 2980.749318 samples/sec	accuracy=0.710164	lr=0.389139
Epoch[17] Batch [199]	Speed: 2843.026386 samples/sec	accuracy=0.709885	lr=0.389068
Epoch[17] Batch [249]	Speed: 3004.089955 samples/sec	accuracy=0.710020	lr=0.388996
Epoch[17] Batch [299]	Speed: 2784.325308 samples/sec	accuracy=0.710226	lr=0.388925
Epoch[17] Batch [349]	Speed: 3035.660508 samples/sec	accuracy=0.710278	lr=0.388853
Epoch[17] Batch [399]	Speed: 2877.981099 samples/sec	accuracy=0.710548	lr=0.388781
Epoch[17] Batch [449]	Speed: 2896.634413 samples/sec	accuracy=0.710696	lr=0.388709
Epoch[17] Batch [499]	Speed: 2936.047689 samples/sec	accuracy=0.710978	lr=0.388636
Epoch[17] Batch [549]	Speed: 2882.095647 samples/sec	accuracy=0.711135	lr=0.388564
Epoch[17] Batch [599]	Speed: 2919.460820 samples/sec	accuracy=0.711226	lr=0.388491
Epoch[17] Batch [649]	Speed: 2865.673533 samples/sec	accuracy=0.711356	lr=0.388418
Epoch[17] Batch [699]	Speed: 2982.991496 samples/sec	accuracy=0.711536	lr=0.388344
Epoch[17] Batch [749]	Speed: 2864.278167 samples/sec	accuracy=0.711664	lr=0.388271
Epoch[17] Batch [799]	Speed: 2820.376933 samples/sec	accuracy=0.711717	lr=0.388197
Epoch[17] Batch [849]	Speed: 2842.193427 samples/sec	accuracy=0.711902	lr=0.388123
Epoch[17] Batch [899]	Speed: 2804.668346 samples/sec	accuracy=0.712004	lr=0.388049
Epoch[17] Batch [949]	Speed: 2850.527485 samples/sec	accuracy=0.712107	lr=0.387974
Epoch[17] Batch [999]	Speed: 2918.765626 samples/sec	accuracy=0.712291	lr=0.387900
Epoch[17] Batch [1049]	Speed: 2820.642913 samples/sec	accuracy=0.712373	lr=0.387825
Epoch[17] Batch [1099]	Speed: 2978.089523 samples/sec	accuracy=0.712479	lr=0.387750
Epoch[17] Batch [1149]	Speed: 2880.460757 samples/sec	accuracy=0.712536	lr=0.387674
Epoch[17] Batch [1199]	Speed: 2933.052119 samples/sec	accuracy=0.712602	lr=0.387599
Epoch[17] Batch [1249]	Speed: 2913.457811 samples/sec	accuracy=0.712706	lr=0.387523
[Epoch 17] training: accuracy=0.712709
[Epoch 17] speed: 2895 samples/sec	time cost: 452.606526
[Epoch 17] validation: err-top1=0.382772 err-top5=0.152893
Epoch[18] Batch [49]	Speed: 2822.910970 samples/sec	accuracy=0.712477	lr=0.387445
Epoch[18] Batch [99]	Speed: 2847.785216 samples/sec	accuracy=0.713569	lr=0.387369
Epoch[18] Batch [149]	Speed: 2779.757038 samples/sec	accuracy=0.713764	lr=0.387293
Epoch[18] Batch [199]	Speed: 2922.240497 samples/sec	accuracy=0.714131	lr=0.387216
Epoch[18] Batch [249]	Speed: 2778.674477 samples/sec	accuracy=0.713774	lr=0.387139
Epoch[18] Batch [299]	Speed: 2901.554757 samples/sec	accuracy=0.713725	lr=0.387062
Epoch[18] Batch [349]	Speed: 2937.684939 samples/sec	accuracy=0.713920	lr=0.386984
Epoch[18] Batch [399]	Speed: 2910.699543 samples/sec	accuracy=0.714042	lr=0.386907
Epoch[18] Batch [449]	Speed: 2855.222609 samples/sec	accuracy=0.714252	lr=0.386829
Epoch[18] Batch [499]	Speed: 2813.258228 samples/sec	accuracy=0.714319	lr=0.386751
Epoch[18] Batch [549]	Speed: 2907.140315 samples/sec	accuracy=0.714471	lr=0.386673
Epoch[18] Batch [599]	Speed: 2855.542058 samples/sec	accuracy=0.714602	lr=0.386594
Epoch[18] Batch [649]	Speed: 2888.082227 samples/sec	accuracy=0.714661	lr=0.386515
Epoch[18] Batch [699]	Speed: 2892.284937 samples/sec	accuracy=0.714775	lr=0.386436
Epoch[18] Batch [749]	Speed: 2836.701980 samples/sec	accuracy=0.714857	lr=0.386357
Epoch[18] Batch [799]	Speed: 2911.097901 samples/sec	accuracy=0.714959	lr=0.386278
Epoch[18] Batch [849]	Speed: 2958.920648 samples/sec	accuracy=0.715109	lr=0.386198
Epoch[18] Batch [899]	Speed: 2959.945299 samples/sec	accuracy=0.715245	lr=0.386119
Epoch[18] Batch [949]	Speed: 2929.101334 samples/sec	accuracy=0.715397	lr=0.386038
Epoch[18] Batch [999]	Speed: 2863.751861 samples/sec	accuracy=0.715507	lr=0.385958
Epoch[18] Batch [1049]	Speed: 2945.635888 samples/sec	accuracy=0.715730	lr=0.385878
Epoch[18] Batch [1099]	Speed: 2949.296865 samples/sec	accuracy=0.715761	lr=0.385797
Epoch[18] Batch [1149]	Speed: 3000.804722 samples/sec	accuracy=0.715869	lr=0.385716
Epoch[18] Batch [1199]	Speed: 2913.007596 samples/sec	accuracy=0.716008	lr=0.385635
Epoch[18] Batch [1249]	Speed: 2959.752460 samples/sec	accuracy=0.716094	lr=0.385554
[Epoch 18] training: accuracy=0.716095
[Epoch 18] speed: 2890 samples/sec	time cost: 453.536749
[Epoch 18] validation: err-top1=0.374482 err-top5=0.145428
Epoch[19] Batch [49]	Speed: 2820.724199 samples/sec	accuracy=0.715561	lr=0.385470
Epoch[19] Batch [99]	Speed: 2781.035882 samples/sec	accuracy=0.715720	lr=0.385388
Epoch[19] Batch [149]	Speed: 2959.201659 samples/sec	accuracy=0.716256	lr=0.385306
Epoch[19] Batch [199]	Speed: 2923.479904 samples/sec	accuracy=0.716279	lr=0.385224
Epoch[19] Batch [249]	Speed: 2859.262738 samples/sec	accuracy=0.716072	lr=0.385142
Epoch[19] Batch [299]	Speed: 2847.813086 samples/sec	accuracy=0.716389	lr=0.385059
Epoch[19] Batch [349]	Speed: 2874.122947 samples/sec	accuracy=0.716426	lr=0.384976
Epoch[19] Batch [399]	Speed: 2973.253362 samples/sec	accuracy=0.716554	lr=0.384893
Epoch[19] Batch [449]	Speed: 2949.378930 samples/sec	accuracy=0.716646	lr=0.384810
Epoch[19] Batch [499]	Speed: 2804.955809 samples/sec	accuracy=0.716818	lr=0.384726
Epoch[19] Batch [549]	Speed: 2924.594936 samples/sec	accuracy=0.717031	lr=0.384642
Epoch[19] Batch [599]	Speed: 2948.931840 samples/sec	accuracy=0.717083	lr=0.384558
Epoch[19] Batch [649]	Speed: 2857.974245 samples/sec	accuracy=0.717238	lr=0.384474
Epoch[19] Batch [699]	Speed: 2852.123146 samples/sec	accuracy=0.717380	lr=0.384389
Epoch[19] Batch [749]	Speed: 2813.679832 samples/sec	accuracy=0.717533	lr=0.384305
Epoch[19] Batch [799]	Speed: 2845.128391 samples/sec	accuracy=0.717661	lr=0.384220
Epoch[19] Batch [849]	Speed: 2913.294813 samples/sec	accuracy=0.717867	lr=0.384135
Epoch[19] Batch [899]	Speed: 2866.370751 samples/sec	accuracy=0.718083	lr=0.384049
Epoch[19] Batch [949]	Speed: 2860.208967 samples/sec	accuracy=0.718172	lr=0.383964
Epoch[19] Batch [999]	Speed: 2788.773237 samples/sec	accuracy=0.718296	lr=0.383878
Epoch[19] Batch [1049]	Speed: 2838.463808 samples/sec	accuracy=0.718391	lr=0.383792
Epoch[19] Batch [1099]	Speed: 2949.948368 samples/sec	accuracy=0.718509	lr=0.383706
Epoch[19] Batch [1149]	Speed: 2903.754285 samples/sec	accuracy=0.718646	lr=0.383619
Epoch[19] Batch [1199]	Speed: 2978.568923 samples/sec	accuracy=0.718828	lr=0.383533
Epoch[19] Batch [1249]	Speed: 3079.625316 samples/sec	accuracy=0.718991	lr=0.383446
[Epoch 19] training: accuracy=0.718990
[Epoch 19] speed: 2884 samples/sec	time cost: 453.967324
[Epoch 19] validation: err-top1=0.381796 err-top5=0.148996
Epoch[20] Batch [49]	Speed: 2875.997017 samples/sec	accuracy=0.718076	lr=0.383357
Epoch[20] Batch [99]	Speed: 2819.284346 samples/sec	accuracy=0.717641	lr=0.383270
Epoch[20] Batch [149]	Speed: 2775.915137 samples/sec	accuracy=0.718576	lr=0.383182
Epoch[20] Batch [199]	Speed: 2810.108790 samples/sec	accuracy=0.719075	lr=0.383094
Epoch[20] Batch [249]	Speed: 2923.006892 samples/sec	accuracy=0.719268	lr=0.383006
Epoch[20] Batch [299]	Speed: 3066.065101 samples/sec	accuracy=0.719383	lr=0.382918
Epoch[20] Batch [349]	Speed: 2915.323427 samples/sec	accuracy=0.719232	lr=0.382830
Epoch[20] Batch [399]	Speed: 2828.574710 samples/sec	accuracy=0.719328	lr=0.382741
Epoch[20] Batch [449]	Speed: 2888.860342 samples/sec	accuracy=0.719591	lr=0.382652
Epoch[20] Batch [499]	Speed: 2848.461589 samples/sec	accuracy=0.719719	lr=0.382563
Epoch[20] Batch [549]	Speed: 2941.779417 samples/sec	accuracy=0.719987	lr=0.382474
Epoch[20] Batch [599]	Speed: 2930.880839 samples/sec	accuracy=0.720119	lr=0.382384
Epoch[20] Batch [649]	Speed: 2954.085660 samples/sec	accuracy=0.720177	lr=0.382295
Epoch[20] Batch [699]	Speed: 2913.347853 samples/sec	accuracy=0.720251	lr=0.382205
Epoch[20] Batch [749]	Speed: 2889.427681 samples/sec	accuracy=0.720413	lr=0.382115
Epoch[20] Batch [799]	Speed: 2863.406099 samples/sec	accuracy=0.720574	lr=0.382024
Epoch[20] Batch [849]	Speed: 2937.702461 samples/sec	accuracy=0.720751	lr=0.381934
Epoch[20] Batch [899]	Speed: 2893.690731 samples/sec	accuracy=0.720887	lr=0.381843
Epoch[20] Batch [949]	Speed: 2782.185307 samples/sec	accuracy=0.720988	lr=0.381752
Epoch[20] Batch [999]	Speed: 2946.320699 samples/sec	accuracy=0.721134	lr=0.381661
Epoch[20] Batch [1049]	Speed: 2905.800573 samples/sec	accuracy=0.721220	lr=0.381569
Epoch[20] Batch [1099]	Speed: 3096.412788 samples/sec	accuracy=0.721411	lr=0.381477
Epoch[20] Batch [1149]	Speed: 2962.168412 samples/sec	accuracy=0.721537	lr=0.381385
Epoch[20] Batch [1199]	Speed: 2890.401686 samples/sec	accuracy=0.721657	lr=0.381293
Epoch[20] Batch [1249]	Speed: 2973.235784 samples/sec	accuracy=0.721760	lr=0.381201
[Epoch 20] training: accuracy=0.721754
[Epoch 20] speed: 2901 samples/sec	time cost: 451.392272
[Epoch 20] validation: err-top1=0.378966 err-top5=0.150012
Epoch[21] Batch [49]	Speed: 2973.135839 samples/sec	accuracy=0.720592	lr=0.381107
Epoch[21] Batch [99]	Speed: 2882.300588 samples/sec	accuracy=0.721382	lr=0.381014
Epoch[21] Batch [149]	Speed: 2844.676133 samples/sec	accuracy=0.721956	lr=0.380921
Epoch[21] Batch [199]	Speed: 2833.481693 samples/sec	accuracy=0.722059	lr=0.380828
Epoch[21] Batch [249]	Speed: 2824.547367 samples/sec	accuracy=0.721928	lr=0.380734
Epoch[21] Batch [299]	Speed: 2836.780897 samples/sec	accuracy=0.721893	lr=0.380641
Epoch[21] Batch [349]	Speed: 2907.861363 samples/sec	accuracy=0.722084	lr=0.380547
Epoch[21] Batch [399]	Speed: 2906.122709 samples/sec	accuracy=0.722049	lr=0.380453
Epoch[21] Batch [449]	Speed: 2844.960323 samples/sec	accuracy=0.722083	lr=0.380359
Epoch[21] Batch [499]	Speed: 2905.387547 samples/sec	accuracy=0.722277	lr=0.380264
Epoch[21] Batch [549]	Speed: 2853.382672 samples/sec	accuracy=0.722505	lr=0.380169
Epoch[21] Batch [599]	Speed: 2841.021144 samples/sec	accuracy=0.722546	lr=0.380074
Epoch[21] Batch [649]	Speed: 2888.208660 samples/sec	accuracy=0.722664	lr=0.379979
Epoch[21] Batch [699]	Speed: 2883.868514 samples/sec	accuracy=0.722783	lr=0.379884
Epoch[21] Batch [749]	Speed: 2829.087829 samples/sec	accuracy=0.722866	lr=0.379788
Epoch[21] Batch [799]	Speed: 2990.750437 samples/sec	accuracy=0.723018	lr=0.379693
Epoch[21] Batch [849]	Speed: 2916.665945 samples/sec	accuracy=0.723082	lr=0.379597
Epoch[21] Batch [899]	Speed: 2794.896671 samples/sec	accuracy=0.723168	lr=0.379501
Epoch[21] Batch [949]	Speed: 2867.544294 samples/sec	accuracy=0.723317	lr=0.379404
Epoch[21] Batch [999]	Speed: 2838.567210 samples/sec	accuracy=0.723551	lr=0.379307
Epoch[21] Batch [1049]	Speed: 2863.582731 samples/sec	accuracy=0.723731	lr=0.379211
Epoch[21] Batch [1099]	Speed: 3058.405418 samples/sec	accuracy=0.723864	lr=0.379114
Epoch[21] Batch [1149]	Speed: 2859.735298 samples/sec	accuracy=0.723976	lr=0.379016
Epoch[21] Batch [1199]	Speed: 2981.486981 samples/sec	accuracy=0.724088	lr=0.378919
Epoch[21] Batch [1249]	Speed: 3153.998313 samples/sec	accuracy=0.724236	lr=0.378821
[Epoch 21] training: accuracy=0.724250
[Epoch 21] speed: 2891 samples/sec	time cost: 453.123993
[Epoch 21] validation: err-top1=0.375399 err-top5=0.147840
Epoch[22] Batch [49]	Speed: 2904.139354 samples/sec	accuracy=0.724686	lr=0.378721
Epoch[22] Batch [99]	Speed: 2911.791737 samples/sec	accuracy=0.724931	lr=0.378623
Epoch[22] Batch [149]	Speed: 2983.018388 samples/sec	accuracy=0.724859	lr=0.378525
Epoch[22] Batch [199]	Speed: 2910.115224 samples/sec	accuracy=0.725074	lr=0.378426
Epoch[22] Batch [249]	Speed: 2859.006781 samples/sec	accuracy=0.724889	lr=0.378328
Epoch[22] Batch [299]	Speed: 2844.387743 samples/sec	accuracy=0.724681	lr=0.378229
Epoch[22] Batch [349]	Speed: 3024.355433 samples/sec	accuracy=0.724550	lr=0.378129
Epoch[22] Batch [399]	Speed: 2911.054691 samples/sec	accuracy=0.724656	lr=0.378030
Epoch[22] Batch [449]	Speed: 2860.087373 samples/sec	accuracy=0.724932	lr=0.377930
Epoch[22] Batch [499]	Speed: 2978.158330 samples/sec	accuracy=0.725131	lr=0.377831
Epoch[22] Batch [549]	Speed: 2863.102333 samples/sec	accuracy=0.725366	lr=0.377731
Epoch[22] Batch [599]	Speed: 2897.850391 samples/sec	accuracy=0.725455	lr=0.377630
Epoch[22] Batch [649]	Speed: 2928.849498 samples/sec	accuracy=0.725435	lr=0.377530
Epoch[22] Batch [699]	Speed: 2867.992247 samples/sec	accuracy=0.725510	lr=0.377429
Epoch[22] Batch [749]	Speed: 3094.360275 samples/sec	accuracy=0.725793	lr=0.377328
Epoch[22] Batch [799]	Speed: 2918.619526 samples/sec	accuracy=0.726029	lr=0.377227
Epoch[22] Batch [849]	Speed: 2921.029278 samples/sec	accuracy=0.726137	lr=0.377126
Epoch[22] Batch [899]	Speed: 2917.341037 samples/sec	accuracy=0.726351	lr=0.377024
Epoch[22] Batch [949]	Speed: 2821.251078 samples/sec	accuracy=0.726471	lr=0.376923
Epoch[22] Batch [999]	Speed: 2910.993488 samples/sec	accuracy=0.726583	lr=0.376821
Epoch[22] Batch [1049]	Speed: 2967.338957 samples/sec	accuracy=0.726748	lr=0.376719
Epoch[22] Batch [1099]	Speed: 2904.485321 samples/sec	accuracy=0.726858	lr=0.376616
Epoch[22] Batch [1149]	Speed: 2863.216014 samples/sec	accuracy=0.726939	lr=0.376514
Epoch[22] Batch [1199]	Speed: 2909.898225 samples/sec	accuracy=0.726955	lr=0.376411
Epoch[22] Batch [1249]	Speed: 2964.796556 samples/sec	accuracy=0.727060	lr=0.376308
[Epoch 22] training: accuracy=0.727065
[Epoch 22] speed: 2913 samples/sec	time cost: 449.117790
[Epoch 22] validation: err-top1=0.370855 err-top5=0.141004
Epoch[23] Batch [49]	Speed: 2957.545272 samples/sec	accuracy=0.726675	lr=0.376203
Epoch[23] Batch [99]	Speed: 2857.461735 samples/sec	accuracy=0.727485	lr=0.376099
Epoch[23] Batch [149]	Speed: 2873.089612 samples/sec	accuracy=0.727252	lr=0.375996
Epoch[23] Batch [199]	Speed: 2898.545477 samples/sec	accuracy=0.726880	lr=0.375892
Epoch[23] Batch [249]	Speed: 3000.977240 samples/sec	accuracy=0.727058	lr=0.375788
Epoch[23] Batch [299]	Speed: 2909.897752 samples/sec	accuracy=0.726995	lr=0.375683
Epoch[23] Batch [349]	Speed: 2815.940392 samples/sec	accuracy=0.726922	lr=0.375579
Epoch[23] Batch [399]	Speed: 2834.269714 samples/sec	accuracy=0.726981	lr=0.375474
Epoch[23] Batch [449]	Speed: 2950.053568 samples/sec	accuracy=0.727104	lr=0.375369
Epoch[23] Batch [499]	Speed: 2893.998721 samples/sec	accuracy=0.727134	lr=0.375264
Epoch[23] Batch [549]	Speed: 2971.745080 samples/sec	accuracy=0.727230	lr=0.375159
Epoch[23] Batch [599]	Speed: 2798.921082 samples/sec	accuracy=0.727243	lr=0.375053
Epoch[23] Batch [649]	Speed: 2933.347550 samples/sec	accuracy=0.727305	lr=0.374948
Epoch[23] Batch [699]	Speed: 2867.840769 samples/sec	accuracy=0.727472	lr=0.374842
Epoch[23] Batch [749]	Speed: 2902.693770 samples/sec	accuracy=0.727643	lr=0.374736
Epoch[23] Batch [799]	Speed: 2841.533149 samples/sec	accuracy=0.727670	lr=0.374629
Epoch[23] Batch [849]	Speed: 2932.209381 samples/sec	accuracy=0.727786	lr=0.374523
Epoch[23] Batch [899]	Speed: 2885.378482 samples/sec	accuracy=0.727935	lr=0.374416
Epoch[23] Batch [949]	Speed: 2924.236119 samples/sec	accuracy=0.728028	lr=0.374309
Epoch[23] Batch [999]	Speed: 3000.346978 samples/sec	accuracy=0.728120	lr=0.374202
Epoch[23] Batch [1049]	Speed: 2786.632037 samples/sec	accuracy=0.728291	lr=0.374095
Epoch[23] Batch [1099]	Speed: 2857.770542 samples/sec	accuracy=0.728390	lr=0.373987
Epoch[23] Batch [1149]	Speed: 2927.048001 samples/sec	accuracy=0.728429	lr=0.373879
Epoch[23] Batch [1199]	Speed: 2999.010100 samples/sec	accuracy=0.728549	lr=0.373771
Epoch[23] Batch [1249]	Speed: 3037.639688 samples/sec	accuracy=0.728648	lr=0.373663
[Epoch 23] training: accuracy=0.728654
[Epoch 23] speed: 2902 samples/sec	time cost: 451.243936
[Epoch 23] validation: err-top1=0.365540 err-top5=0.140991
Epoch[24] Batch [49]	Speed: 2899.435481 samples/sec	accuracy=0.728744	lr=0.373552
Epoch[24] Batch [99]	Speed: 2810.944865 samples/sec	accuracy=0.728559	lr=0.373444
Epoch[24] Batch [149]	Speed: 2825.213563 samples/sec	accuracy=0.728771	lr=0.373335
Epoch[24] Batch [199]	Speed: 2769.806769 samples/sec	accuracy=0.728817	lr=0.373226
Epoch[24] Batch [249]	Speed: 3014.275405 samples/sec	accuracy=0.728750	lr=0.373117
Epoch[24] Batch [299]	Speed: 2890.597344 samples/sec	accuracy=0.728862	lr=0.373007
Epoch[24] Batch [349]	Speed: 2872.530285 samples/sec	accuracy=0.728695	lr=0.372897
Epoch[24] Batch [399]	Speed: 2837.021683 samples/sec	accuracy=0.728941	lr=0.372788
Epoch[24] Batch [449]	Speed: 2853.588745 samples/sec	accuracy=0.729110	lr=0.372678
Epoch[24] Batch [499]	Speed: 2903.128440 samples/sec	accuracy=0.729370	lr=0.372567
Epoch[24] Batch [549]	Speed: 2914.653581 samples/sec	accuracy=0.729441	lr=0.372457
Epoch[24] Batch [599]	Speed: 2868.054260 samples/sec	accuracy=0.729661	lr=0.372346
Epoch[24] Batch [649]	Speed: 2907.659030 samples/sec	accuracy=0.729723	lr=0.372235
Epoch[24] Batch [699]	Speed: 2900.196540 samples/sec	accuracy=0.729829	lr=0.372124
Epoch[24] Batch [749]	Speed: 3061.057083 samples/sec	accuracy=0.729936	lr=0.372013
Epoch[24] Batch [799]	Speed: 2916.470426 samples/sec	accuracy=0.730069	lr=0.371901
Epoch[24] Batch [849]	Speed: 2978.677043 samples/sec	accuracy=0.730181	lr=0.371790
Epoch[24] Batch [899]	Speed: 2915.398586 samples/sec	accuracy=0.730265	lr=0.371678
Epoch[24] Batch [949]	Speed: 3002.604642 samples/sec	accuracy=0.730336	lr=0.371565
Epoch[24] Batch [999]	Speed: 2932.295182 samples/sec	accuracy=0.730471	lr=0.371453
Epoch[24] Batch [1049]	Speed: 2842.258279 samples/sec	accuracy=0.730577	lr=0.371341
Epoch[24] Batch [1099]	Speed: 2987.985255 samples/sec	accuracy=0.730746	lr=0.371228
Epoch[24] Batch [1149]	Speed: 2891.329789 samples/sec	accuracy=0.730816	lr=0.371115
Epoch[24] Batch [1199]	Speed: 3012.600465 samples/sec	accuracy=0.730822	lr=0.371002
Epoch[24] Batch [1249]	Speed: 2991.365715 samples/sec	accuracy=0.730889	lr=0.370888
[Epoch 24] training: accuracy=0.730892
[Epoch 24] speed: 2908 samples/sec	time cost: 450.611483
[Epoch 24] validation: err-top1=0.376834 err-top5=0.146185
Epoch[25] Batch [49]	Speed: 2992.923891 samples/sec	accuracy=0.730349	lr=0.370773
Epoch[25] Batch [99]	Speed: 2943.423916 samples/sec	accuracy=0.730535	lr=0.370659
Epoch[25] Batch [149]	Speed: 2944.286152 samples/sec	accuracy=0.730342	lr=0.370545
Epoch[25] Batch [199]	Speed: 2987.041854 samples/sec	accuracy=0.730698	lr=0.370431
Epoch[25] Batch [249]	Speed: 2887.671426 samples/sec	accuracy=0.730691	lr=0.370316
Epoch[25] Batch [299]	Speed: 2965.377983 samples/sec	accuracy=0.731038	lr=0.370202
Epoch[25] Batch [349]	Speed: 3034.537494 samples/sec	accuracy=0.731183	lr=0.370087
Epoch[25] Batch [399]	Speed: 2862.866985 samples/sec	accuracy=0.731320	lr=0.369972
Epoch[25] Batch [449]	Speed: 2816.342818 samples/sec	accuracy=0.731361	lr=0.369857
Epoch[25] Batch [499]	Speed: 2870.355805 samples/sec	accuracy=0.731509	lr=0.369741
Epoch[25] Batch [549]	Speed: 2826.726826 samples/sec	accuracy=0.731719	lr=0.369626
Epoch[25] Batch [599]	Speed: 2894.075086 samples/sec	accuracy=0.731961	lr=0.369510
Epoch[25] Batch [649]	Speed: 2907.399649 samples/sec	accuracy=0.732156	lr=0.369394
Epoch[25] Batch [699]	Speed: 2881.682835 samples/sec	accuracy=0.732311	lr=0.369278
Epoch[25] Batch [749]	Speed: 2870.341955 samples/sec	accuracy=0.732435	lr=0.369161
Epoch[25] Batch [799]	Speed: 2843.654673 samples/sec	accuracy=0.732486	lr=0.369045
Epoch[25] Batch [849]	Speed: 3004.093695 samples/sec	accuracy=0.732554	lr=0.368928
Epoch[25] Batch [899]	Speed: 2854.414587 samples/sec	accuracy=0.732628	lr=0.368811
Epoch[25] Batch [949]	Speed: 2877.114545 samples/sec	accuracy=0.732767	lr=0.368694
Epoch[25] Batch [999]	Speed: 2842.164425 samples/sec	accuracy=0.732892	lr=0.368576
Epoch[25] Batch [1049]	Speed: 2925.441472 samples/sec	accuracy=0.733031	lr=0.368459
Epoch[25] Batch [1099]	Speed: 2934.755853 samples/sec	accuracy=0.733089	lr=0.368341
Epoch[25] Batch [1149]	Speed: 2838.255637 samples/sec	accuracy=0.733231	lr=0.368223
Epoch[25] Batch [1199]	Speed: 2828.369403 samples/sec	accuracy=0.733368	lr=0.368105
Epoch[25] Batch [1249]	Speed: 3038.523190 samples/sec	accuracy=0.733404	lr=0.367986
[Epoch 25] training: accuracy=0.733405
[Epoch 25] speed: 2903 samples/sec	time cost: 450.839403
[Epoch 25] validation: err-top1=0.377830 err-top5=0.146265
Epoch[26] Batch [49]	Speed: 2890.547230 samples/sec	accuracy=0.733526	lr=0.367865
Epoch[26] Batch [99]	Speed: 2945.509629 samples/sec	accuracy=0.732452	lr=0.367747
Epoch[26] Batch [149]	Speed: 2988.421269 samples/sec	accuracy=0.732958	lr=0.367628
Epoch[26] Batch [199]	Speed: 2800.185310 samples/sec	accuracy=0.733265	lr=0.367508
Epoch[26] Batch [249]	Speed: 2809.351419 samples/sec	accuracy=0.733244	lr=0.367389
Epoch[26] Batch [299]	Speed: 2831.019600 samples/sec	accuracy=0.733263	lr=0.367269
Epoch[26] Batch [349]	Speed: 2778.287416 samples/sec	accuracy=0.733224	lr=0.367150
Epoch[26] Batch [399]	Speed: 2965.028371 samples/sec	accuracy=0.733476	lr=0.367030
Epoch[26] Batch [449]	Speed: 2865.661297 samples/sec	accuracy=0.733637	lr=0.366909
Epoch[26] Batch [499]	Speed: 2953.785021 samples/sec	accuracy=0.733624	lr=0.366789
Epoch[26] Batch [549]	Speed: 2924.649702 samples/sec	accuracy=0.733722	lr=0.366668
Epoch[26] Batch [599]	Speed: 2820.814012 samples/sec	accuracy=0.733994	lr=0.366548
Epoch[26] Batch [649]	Speed: 2937.636395 samples/sec	accuracy=0.734150	lr=0.366426
Epoch[26] Batch [699]	Speed: 2960.776830 samples/sec	accuracy=0.734271	lr=0.366305
Epoch[26] Batch [749]	Speed: 2987.842994 samples/sec	accuracy=0.734412	lr=0.366184
Epoch[26] Batch [799]	Speed: 2904.286639 samples/sec	accuracy=0.734541	lr=0.366062
Epoch[26] Batch [849]	Speed: 2849.643312 samples/sec	accuracy=0.734756	lr=0.365940
Epoch[26] Batch [899]	Speed: 2863.149400 samples/sec	accuracy=0.734827	lr=0.365819
Epoch[26] Batch [949]	Speed: 3022.527052 samples/sec	accuracy=0.735039	lr=0.365696
Epoch[26] Batch [999]	Speed: 2916.263607 samples/sec	accuracy=0.735139	lr=0.365574
Epoch[26] Batch [1049]	Speed: 2863.004502 samples/sec	accuracy=0.735242	lr=0.365451
Epoch[26] Batch [1099]	Speed: 2861.285664 samples/sec	accuracy=0.735365	lr=0.365329
Epoch[26] Batch [1149]	Speed: 2936.515816 samples/sec	accuracy=0.735468	lr=0.365206
Epoch[26] Batch [1199]	Speed: 2897.891060 samples/sec	accuracy=0.735568	lr=0.365082
Epoch[26] Batch [1249]	Speed: 3038.237960 samples/sec	accuracy=0.735665	lr=0.364959
[Epoch 26] training: accuracy=0.735667
[Epoch 26] speed: 2900 samples/sec	time cost: 451.356574
[Epoch 26] validation: err-top1=0.370875 err-top5=0.142060
Epoch[27] Batch [49]	Speed: 2907.429407 samples/sec	accuracy=0.733822	lr=0.364833
Epoch[27] Batch [99]	Speed: 2922.860408 samples/sec	accuracy=0.734022	lr=0.364709
Epoch[27] Batch [149]	Speed: 2946.559740 samples/sec	accuracy=0.734224	lr=0.364585
Epoch[27] Batch [199]	Speed: 2940.153984 samples/sec	accuracy=0.734677	lr=0.364461
Epoch[27] Batch [249]	Speed: 2875.299613 samples/sec	accuracy=0.734636	lr=0.364337
Epoch[27] Batch [299]	Speed: 2852.986117 samples/sec	accuracy=0.734768	lr=0.364212
Epoch[27] Batch [349]	Speed: 2941.949810 samples/sec	accuracy=0.734954	lr=0.364087
Epoch[27] Batch [399]	Speed: 2827.663851 samples/sec	accuracy=0.735121	lr=0.363962
Epoch[27] Batch [449]	Speed: 2842.771333 samples/sec	accuracy=0.735227	lr=0.363837
Epoch[27] Batch [499]	Speed: 2885.331496 samples/sec	accuracy=0.735303	lr=0.363712
Epoch[27] Batch [549]	Speed: 2953.398412 samples/sec	accuracy=0.735503	lr=0.363586
Epoch[27] Batch [599]	Speed: 2869.966754 samples/sec	accuracy=0.735745	lr=0.363461
Epoch[27] Batch [649]	Speed: 2857.719507 samples/sec	accuracy=0.735882	lr=0.363335
Epoch[27] Batch [699]	Speed: 2975.085740 samples/sec	accuracy=0.735901	lr=0.363209
Epoch[27] Batch [749]	Speed: 2901.573027 samples/sec	accuracy=0.736077	lr=0.363082
Epoch[27] Batch [799]	Speed: 2889.771162 samples/sec	accuracy=0.736158	lr=0.362956
Epoch[27] Batch [849]	Speed: 2818.718021 samples/sec	accuracy=0.736340	lr=0.362829
Epoch[27] Batch [899]	Speed: 2784.509106 samples/sec	accuracy=0.736457	lr=0.362702
Epoch[27] Batch [949]	Speed: 2966.404202 samples/sec	accuracy=0.736599	lr=0.362575
Epoch[27] Batch [999]	Speed: 2992.372726 samples/sec	accuracy=0.736717	lr=0.362448
Epoch[27] Batch [1049]	Speed: 2969.312092 samples/sec	accuracy=0.736830	lr=0.362320
Epoch[27] Batch [1099]	Speed: 2893.392122 samples/sec	accuracy=0.736958	lr=0.362193
Epoch[27] Batch [1149]	Speed: 2874.258316 samples/sec	accuracy=0.736996	lr=0.362065
Epoch[27] Batch [1199]	Speed: 2904.762649 samples/sec	accuracy=0.737141	lr=0.361937
Epoch[27] Batch [1249]	Speed: 2971.929491 samples/sec	accuracy=0.737258	lr=0.361809
[Epoch 27] training: accuracy=0.737266
[Epoch 27] speed: 2899 samples/sec	time cost: 451.575163
[Epoch 27] validation: err-top1=0.372369 err-top5=0.145568
Epoch[28] Batch [49]	Speed: 2946.420547 samples/sec	accuracy=0.736856	lr=0.361678
Epoch[28] Batch [99]	Speed: 2810.752740 samples/sec	accuracy=0.737332	lr=0.361549
Epoch[28] Batch [149]	Speed: 2878.293509 samples/sec	accuracy=0.736317	lr=0.361420
Epoch[28] Batch [199]	Speed: 2919.462487 samples/sec	accuracy=0.736548	lr=0.361291
Epoch[28] Batch [249]	Speed: 2961.550423 samples/sec	accuracy=0.736738	lr=0.361162
Epoch[28] Batch [299]	Speed: 2930.533235 samples/sec	accuracy=0.736649	lr=0.361032
Epoch[28] Batch [349]	Speed: 3025.514784 samples/sec	accuracy=0.736816	lr=0.360903
Epoch[28] Batch [399]	Speed: 2815.798202 samples/sec	accuracy=0.736768	lr=0.360773
Epoch[28] Batch [449]	Speed: 2875.085504 samples/sec	accuracy=0.736769	lr=0.360643
Epoch[28] Batch [499]	Speed: 2968.045383 samples/sec	accuracy=0.736914	lr=0.360513
Epoch[28] Batch [549]	Speed: 2946.992442 samples/sec	accuracy=0.737017	lr=0.360382
Epoch[28] Batch [599]	Speed: 2860.111028 samples/sec	accuracy=0.737182	lr=0.360252
Epoch[28] Batch [649]	Speed: 2885.820546 samples/sec	accuracy=0.737433	lr=0.360121
Epoch[28] Batch [699]	Speed: 2931.126823 samples/sec	accuracy=0.737558	lr=0.359990
Epoch[28] Batch [749]	Speed: 2867.560644 samples/sec	accuracy=0.737680	lr=0.359859
Epoch[28] Batch [799]	Speed: 2913.275961 samples/sec	accuracy=0.737822	lr=0.359728
Epoch[28] Batch [849]	Speed: 2943.147668 samples/sec	accuracy=0.737877	lr=0.359596
Epoch[28] Batch [899]	Speed: 2915.417861 samples/sec	accuracy=0.737926	lr=0.359465
Epoch[28] Batch [949]	Speed: 2978.834051 samples/sec	accuracy=0.737971	lr=0.359333
Epoch[28] Batch [999]	Speed: 2861.606624 samples/sec	accuracy=0.737994	lr=0.359201
Epoch[28] Batch [1049]	Speed: 2882.975074 samples/sec	accuracy=0.738088	lr=0.359068
Epoch[28] Batch [1099]	Speed: 2972.342153 samples/sec	accuracy=0.738170	lr=0.358936
Epoch[28] Batch [1149]	Speed: 2884.672841 samples/sec	accuracy=0.738348	lr=0.358803
Epoch[28] Batch [1199]	Speed: 2851.424739 samples/sec	accuracy=0.738471	lr=0.358670
Epoch[28] Batch [1249]	Speed: 3073.881513 samples/sec	accuracy=0.738614	lr=0.358537
[Epoch 28] training: accuracy=0.738620
[Epoch 28] speed: 2912 samples/sec	time cost: 449.581075
[Epoch 28] validation: err-top1=0.366171 err-top5=0.140346
Epoch[29] Batch [49]	Speed: 2833.244311 samples/sec	accuracy=0.739385	lr=0.358401
Epoch[29] Batch [99]	Speed: 3021.255561 samples/sec	accuracy=0.738969	lr=0.358268
Epoch[29] Batch [149]	Speed: 2997.454368 samples/sec	accuracy=0.738624	lr=0.358134
Epoch[29] Batch [199]	Speed: 2888.847479 samples/sec	accuracy=0.738609	lr=0.358001
Epoch[29] Batch [249]	Speed: 2919.141196 samples/sec	accuracy=0.738971	lr=0.357867
Epoch[29] Batch [299]	Speed: 2987.603775 samples/sec	accuracy=0.739109	lr=0.357732
Epoch[29] Batch [349]	Speed: 2936.350429 samples/sec	accuracy=0.739186	lr=0.357598
Epoch[29] Batch [399]	Speed: 3082.774228 samples/sec	accuracy=0.739257	lr=0.357464
Epoch[29] Batch [449]	Speed: 2811.691057 samples/sec	accuracy=0.739388	lr=0.357329
Epoch[29] Batch [499]	Speed: 2932.186200 samples/sec	accuracy=0.739367	lr=0.357194
Epoch[29] Batch [549]	Speed: 2828.201148 samples/sec	accuracy=0.739528	lr=0.357059
Epoch[29] Batch [599]	Speed: 3050.252043 samples/sec	accuracy=0.739591	lr=0.356924
Epoch[29] Batch [649]	Speed: 3058.978782 samples/sec	accuracy=0.739652	lr=0.356788
Epoch[29] Batch [699]	Speed: 2943.418187 samples/sec	accuracy=0.739766	lr=0.356652
Epoch[29] Batch [749]	Speed: 2949.411863 samples/sec	accuracy=0.739816	lr=0.356517
Epoch[29] Batch [799]	Speed: 2880.469527 samples/sec	accuracy=0.739874	lr=0.356380
Epoch[29] Batch [849]	Speed: 2857.871021 samples/sec	accuracy=0.740129	lr=0.356244
Epoch[29] Batch [899]	Speed: 2858.430094 samples/sec	accuracy=0.740350	lr=0.356108
Epoch[29] Batch [949]	Speed: 2797.201282 samples/sec	accuracy=0.740463	lr=0.355971
Epoch[29] Batch [999]	Speed: 3024.241459 samples/sec	accuracy=0.740547	lr=0.355834
Epoch[29] Batch [1049]	Speed: 2819.878262 samples/sec	accuracy=0.740668	lr=0.355698
Epoch[29] Batch [1099]	Speed: 2907.047557 samples/sec	accuracy=0.740737	lr=0.355560
Epoch[29] Batch [1149]	Speed: 2817.761329 samples/sec	accuracy=0.740834	lr=0.355423
Epoch[29] Batch [1199]	Speed: 2899.043244 samples/sec	accuracy=0.740958	lr=0.355285
Epoch[29] Batch [1249]	Speed: 3086.475167 samples/sec	accuracy=0.741056	lr=0.355148
[Epoch 29] training: accuracy=0.741062
[Epoch 29] speed: 2922 samples/sec	time cost: 447.833586
[Epoch 29] validation: err-top1=0.399821 err-top5=0.162882
Epoch[30] Batch [49]	Speed: 2864.332952 samples/sec	accuracy=0.739722	lr=0.355007
Epoch[30] Batch [99]	Speed: 2796.303887 samples/sec	accuracy=0.739925	lr=0.354869
Epoch[30] Batch [149]	Speed: 2934.751642 samples/sec	accuracy=0.740034	lr=0.354731
Epoch[30] Batch [199]	Speed: 3058.352584 samples/sec	accuracy=0.740267	lr=0.354592
Epoch[30] Batch [249]	Speed: 2874.040977 samples/sec	accuracy=0.740703	lr=0.354454
Epoch[30] Batch [299]	Speed: 2960.824387 samples/sec	accuracy=0.740523	lr=0.354315
Epoch[30] Batch [349]	Speed: 2906.597157 samples/sec	accuracy=0.740780	lr=0.354176
Epoch[30] Batch [399]	Speed: 2883.853566 samples/sec	accuracy=0.740525	lr=0.354037
Epoch[30] Batch [449]	Speed: 2957.782227 samples/sec	accuracy=0.740509	lr=0.353897
Epoch[30] Batch [499]	Speed: 3091.519548 samples/sec	accuracy=0.740543	lr=0.353758
Epoch[30] Batch [549]	Speed: 3173.859132 samples/sec	accuracy=0.740714	lr=0.353618
Epoch[30] Batch [599]	Speed: 3078.685357 samples/sec	accuracy=0.740803	lr=0.353478
Epoch[30] Batch [649]	Speed: 2948.282608 samples/sec	accuracy=0.740925	lr=0.353338
Epoch[30] Batch [699]	Speed: 3103.230067 samples/sec	accuracy=0.741132	lr=0.353198
Epoch[30] Batch [749]	Speed: 3083.140341 samples/sec	accuracy=0.741267	lr=0.353057
Epoch[30] Batch [799]	Speed: 3108.418044 samples/sec	accuracy=0.741266	lr=0.352917
Epoch[30] Batch [849]	Speed: 3214.669315 samples/sec	accuracy=0.741359	lr=0.352776
Epoch[30] Batch [899]	Speed: 3155.064557 samples/sec	accuracy=0.741389	lr=0.352635
Epoch[30] Batch [949]	Speed: 3284.763708 samples/sec	accuracy=0.741452	lr=0.352493
Epoch[30] Batch [999]	Speed: 3160.652856 samples/sec	accuracy=0.741672	lr=0.352352
Epoch[30] Batch [1049]	Speed: 3145.841207 samples/sec	accuracy=0.741762	lr=0.352211
Epoch[30] Batch [1099]	Speed: 3089.215123 samples/sec	accuracy=0.741883	lr=0.352069
Epoch[30] Batch [1149]	Speed: 3093.154523 samples/sec	accuracy=0.742032	lr=0.351927
Epoch[30] Batch [1199]	Speed: 3141.256709 samples/sec	accuracy=0.742172	lr=0.351785
Epoch[30] Batch [1249]	Speed: 3104.716484 samples/sec	accuracy=0.742227	lr=0.351642
[Epoch 30] training: accuracy=0.742229
[Epoch 30] speed: 3041 samples/sec	time cost: 431.000410
[Epoch 30] validation: err-top1=0.367706 err-top5=0.138791
Epoch[31] Batch [49]	Speed: 3100.682636 samples/sec	accuracy=0.741440	lr=0.351497
Epoch[31] Batch [99]	Speed: 3202.358361 samples/sec	accuracy=0.741705	lr=0.351354
Epoch[31] Batch [149]	Speed: 3112.446885 samples/sec	accuracy=0.741517	lr=0.351212
Epoch[31] Batch [199]	Speed: 2988.597482 samples/sec	accuracy=0.741698	lr=0.351069
Epoch[31] Batch [249]	Speed: 2873.142966 samples/sec	accuracy=0.741883	lr=0.350925
Epoch[31] Batch [299]	Speed: 3012.485727 samples/sec	accuracy=0.741895	lr=0.350782
Epoch[31] Batch [349]	Speed: 2889.226661 samples/sec	accuracy=0.742011	lr=0.350639
Epoch[31] Batch [399]	Speed: 2882.600935 samples/sec	accuracy=0.741811	lr=0.350495
Epoch[31] Batch [449]	Speed: 2849.975583 samples/sec	accuracy=0.741761	lr=0.350351
Epoch[31] Batch [499]	Speed: 2969.868594 samples/sec	accuracy=0.741885	lr=0.350207
Epoch[31] Batch [549]	Speed: 2889.481021 samples/sec	accuracy=0.741897	lr=0.350062
Epoch[31] Batch [599]	Speed: 2831.444043 samples/sec	accuracy=0.742173	lr=0.349918
Epoch[31] Batch [649]	Speed: 2950.716474 samples/sec	accuracy=0.742280	lr=0.349773
Epoch[31] Batch [699]	Speed: 2837.405751 samples/sec	accuracy=0.742535	lr=0.349629
Epoch[31] Batch [749]	Speed: 2831.319246 samples/sec	accuracy=0.742770	lr=0.349484
Epoch[31] Batch [799]	Speed: 2936.265554 samples/sec	accuracy=0.742915	lr=0.349338
Epoch[31] Batch [849]	Speed: 2851.187029 samples/sec	accuracy=0.743115	lr=0.349193
Epoch[31] Batch [899]	Speed: 2926.297350 samples/sec	accuracy=0.743288	lr=0.349048
Epoch[31] Batch [949]	Speed: 2857.929136 samples/sec	accuracy=0.743510	lr=0.348902
Epoch[31] Batch [999]	Speed: 2844.085927 samples/sec	accuracy=0.743700	lr=0.348756
Epoch[31] Batch [1049]	Speed: 2787.724644 samples/sec	accuracy=0.743826	lr=0.348610
Epoch[31] Batch [1099]	Speed: 2793.987813 samples/sec	accuracy=0.743949	lr=0.348464
Epoch[31] Batch [1149]	Speed: 2927.680011 samples/sec	accuracy=0.744087	lr=0.348317
Epoch[31] Batch [1199]	Speed: 2952.229860 samples/sec	accuracy=0.744204	lr=0.348171
Epoch[31] Batch [1249]	Speed: 3091.957367 samples/sec	accuracy=0.744264	lr=0.348024
[Epoch 31] training: accuracy=0.744265
[Epoch 31] speed: 2921 samples/sec	time cost: 448.208812
[Epoch 31] validation: err-top1=0.347178 err-top5=0.128388
Epoch[32] Batch [49]	Speed: 2999.188066 samples/sec	accuracy=0.742583	lr=0.347874
Epoch[32] Batch [99]	Speed: 2828.501092 samples/sec	accuracy=0.743048	lr=0.347727
Epoch[32] Batch [149]	Speed: 2904.586479 samples/sec	accuracy=0.742608	lr=0.347580
Epoch[32] Batch [199]	Speed: 2862.797983 samples/sec	accuracy=0.742966	lr=0.347432
Epoch[32] Batch [249]	Speed: 2887.444755 samples/sec	accuracy=0.742950	lr=0.347285
Epoch[32] Batch [299]	Speed: 2884.576436 samples/sec	accuracy=0.743032	lr=0.347137
Epoch[32] Batch [349]	Speed: 3030.818911 samples/sec	accuracy=0.742948	lr=0.346989
Epoch[32] Batch [399]	Speed: 2876.329684 samples/sec	accuracy=0.743215	lr=0.346841
Epoch[32] Batch [449]	Speed: 2994.895499 samples/sec	accuracy=0.743563	lr=0.346692
Epoch[32] Batch [499]	Speed: 3127.525284 samples/sec	accuracy=0.743709	lr=0.346544
Epoch[32] Batch [549]	Speed: 2849.654580 samples/sec	accuracy=0.743842	lr=0.346395
Epoch[32] Batch [599]	Speed: 2998.180440 samples/sec	accuracy=0.744011	lr=0.346246
Epoch[32] Batch [649]	Speed: 2801.551916 samples/sec	accuracy=0.744100	lr=0.346097
Epoch[32] Batch [699]	Speed: 2877.549068 samples/sec	accuracy=0.744241	lr=0.345948
Epoch[32] Batch [749]	Speed: 2838.137478 samples/sec	accuracy=0.744183	lr=0.345799
Epoch[32] Batch [799]	Speed: 2901.502970 samples/sec	accuracy=0.744302	lr=0.345649
Epoch[32] Batch [849]	Speed: 2826.579824 samples/sec	accuracy=0.744362	lr=0.345499
Epoch[32] Batch [899]	Speed: 2893.642148 samples/sec	accuracy=0.744512	lr=0.345349
Epoch[32] Batch [949]	Speed: 2886.977820 samples/sec	accuracy=0.744632	lr=0.345199
Epoch[32] Batch [999]	Speed: 2879.825487 samples/sec	accuracy=0.744720	lr=0.345049
Epoch[32] Batch [1049]	Speed: 2910.369804 samples/sec	accuracy=0.744770	lr=0.344899
Epoch[32] Batch [1099]	Speed: 2881.627385 samples/sec	accuracy=0.744952	lr=0.344748
Epoch[32] Batch [1149]	Speed: 2919.654597 samples/sec	accuracy=0.745088	lr=0.344597
Epoch[32] Batch [1199]	Speed: 2905.780638 samples/sec	accuracy=0.745228	lr=0.344446
Epoch[32] Batch [1249]	Speed: 2985.059324 samples/sec	accuracy=0.745394	lr=0.344295
[Epoch 32] training: accuracy=0.745401
[Epoch 32] speed: 2906 samples/sec	time cost: 450.523412
[Epoch 32] validation: err-top1=0.353894 err-top5=0.129644
Epoch[33] Batch [49]	Speed: 2933.075794 samples/sec	accuracy=0.745738	lr=0.344141
Epoch[33] Batch [99]	Speed: 2886.981235 samples/sec	accuracy=0.746598	lr=0.343989
Epoch[33] Batch [149]	Speed: 2907.539155 samples/sec	accuracy=0.745872	lr=0.343838
Epoch[33] Batch [199]	Speed: 2821.050242 samples/sec	accuracy=0.745909	lr=0.343686
Epoch[33] Batch [249]	Speed: 2834.505060 samples/sec	accuracy=0.746167	lr=0.343534
Epoch[33] Batch [299]	Speed: 2926.065930 samples/sec	accuracy=0.746180	lr=0.343382
Epoch[33] Batch [349]	Speed: 2882.341634 samples/sec	accuracy=0.746315	lr=0.343229
Epoch[33] Batch [399]	Speed: 3021.685777 samples/sec	accuracy=0.746448	lr=0.343077
Epoch[33] Batch [449]	Speed: 2899.583582 samples/sec	accuracy=0.746596	lr=0.342924
Epoch[33] Batch [499]	Speed: 2980.511605 samples/sec	accuracy=0.746596	lr=0.342771
Epoch[33] Batch [549]	Speed: 2944.520867 samples/sec	accuracy=0.746659	lr=0.342618
Epoch[33] Batch [599]	Speed: 2965.321148 samples/sec	accuracy=0.746754	lr=0.342465
Epoch[33] Batch [649]	Speed: 2936.446431 samples/sec	accuracy=0.746863	lr=0.342312
Epoch[33] Batch [699]	Speed: 2900.286745 samples/sec	accuracy=0.747004	lr=0.342158
Epoch[33] Batch [749]	Speed: 2973.802735 samples/sec	accuracy=0.746964	lr=0.342005
Epoch[33] Batch [799]	Speed: 2953.889845 samples/sec	accuracy=0.747067	lr=0.341851
Epoch[33] Batch [849]	Speed: 2859.207158 samples/sec	accuracy=0.747131	lr=0.341697
Epoch[33] Batch [899]	Speed: 2901.520258 samples/sec	accuracy=0.747323	lr=0.341543
Epoch[33] Batch [949]	Speed: 3056.390874 samples/sec	accuracy=0.747404	lr=0.341388
Epoch[33] Batch [999]	Speed: 2910.515395 samples/sec	accuracy=0.747456	lr=0.341234
Epoch[33] Batch [1049]	Speed: 2929.099496 samples/sec	accuracy=0.747523	lr=0.341079
Epoch[33] Batch [1099]	Speed: 2970.623318 samples/sec	accuracy=0.747642	lr=0.340924
Epoch[33] Batch [1149]	Speed: 2851.265807 samples/sec	accuracy=0.747685	lr=0.340769
Epoch[33] Batch [1199]	Speed: 2926.912959 samples/sec	accuracy=0.747836	lr=0.340614
Epoch[33] Batch [1249]	Speed: 3005.669201 samples/sec	accuracy=0.747977	lr=0.340458
[Epoch 33] training: accuracy=0.747974
[Epoch 33] speed: 2923 samples/sec	time cost: 448.168036
[Epoch 33] validation: err-top1=0.356864 err-top5=0.134208
Epoch[34] Batch [49]	Speed: 2897.547484 samples/sec	accuracy=0.747998	lr=0.340300
Epoch[34] Batch [99]	Speed: 3012.142664 samples/sec	accuracy=0.746873	lr=0.340144
Epoch[34] Batch [149]	Speed: 2917.953042 samples/sec	accuracy=0.747106	lr=0.339988
Epoch[34] Batch [199]	Speed: 2927.706793 samples/sec	accuracy=0.747121	lr=0.339832
Epoch[34] Batch [249]	Speed: 2859.582901 samples/sec	accuracy=0.747232	lr=0.339676
Epoch[34] Batch [299]	Speed: 3026.365485 samples/sec	accuracy=0.747267	lr=0.339520
Epoch[34] Batch [349]	Speed: 2787.882326 samples/sec	accuracy=0.747417	lr=0.339363
Epoch[34] Batch [399]	Speed: 2894.752011 samples/sec	accuracy=0.747579	lr=0.339206
Epoch[34] Batch [449]	Speed: 2902.502238 samples/sec	accuracy=0.747769	lr=0.339049
Epoch[34] Batch [499]	Speed: 2820.618536 samples/sec	accuracy=0.747979	lr=0.338892
Epoch[34] Batch [549]	Speed: 2840.231131 samples/sec	accuracy=0.748047	lr=0.338735
Epoch[34] Batch [599]	Speed: 2941.705391 samples/sec	accuracy=0.748090	lr=0.338578
Epoch[34] Batch [649]	Speed: 2908.468057 samples/sec	accuracy=0.748285	lr=0.338420
Epoch[34] Batch [699]	Speed: 3000.415182 samples/sec	accuracy=0.748457	lr=0.338263
Epoch[34] Batch [749]	Speed: 2842.173867 samples/sec	accuracy=0.748608	lr=0.338105
Epoch[34] Batch [799]	Speed: 2908.620233 samples/sec	accuracy=0.748684	lr=0.337947
Epoch[34] Batch [849]	Speed: 2899.640547 samples/sec	accuracy=0.748818	lr=0.337788
Epoch[34] Batch [899]	Speed: 2867.568647 samples/sec	accuracy=0.748872	lr=0.337630
Epoch[34] Batch [949]	Speed: 3055.633511 samples/sec	accuracy=0.748982	lr=0.337472
Epoch[34] Batch [999]	Speed: 2959.647260 samples/sec	accuracy=0.749032	lr=0.337313
Epoch[34] Batch [1049]	Speed: 2901.859994 samples/sec	accuracy=0.749073	lr=0.337154
Epoch[34] Batch [1099]	Speed: 2868.350419 samples/sec	accuracy=0.749148	lr=0.336995
Epoch[34] Batch [1149]	Speed: 2928.954198 samples/sec	accuracy=0.749219	lr=0.336836
Epoch[34] Batch [1199]	Speed: 2859.683241 samples/sec	accuracy=0.749316	lr=0.336677
Epoch[34] Batch [1249]	Speed: 3099.466450 samples/sec	accuracy=0.749399	lr=0.336517
[Epoch 34] training: accuracy=0.749398
[Epoch 34] speed: 2913 samples/sec	time cost: 449.239500
[Epoch 34] validation: err-top1=0.353943 err-top5=0.134135
Epoch[35] Batch [49]	Speed: 3001.158083 samples/sec	accuracy=0.749594	lr=0.336354
Epoch[35] Batch [99]	Speed: 2924.495844 samples/sec	accuracy=0.748792	lr=0.336194
Epoch[35] Batch [149]	Speed: 2909.950510 samples/sec	accuracy=0.748245	lr=0.336034
Epoch[35] Batch [199]	Speed: 3031.921667 samples/sec	accuracy=0.748161	lr=0.335874
Epoch[35] Batch [249]	Speed: 2901.446597 samples/sec	accuracy=0.748387	lr=0.335714
Epoch[35] Batch [299]	Speed: 2906.237708 samples/sec	accuracy=0.748285	lr=0.335553
Epoch[35] Batch [349]	Speed: 3015.764667 samples/sec	accuracy=0.748459	lr=0.335393
Epoch[35] Batch [399]	Speed: 2946.192684 samples/sec	accuracy=0.748643	lr=0.335232
Epoch[35] Batch [449]	Speed: 2922.635975 samples/sec	accuracy=0.748754	lr=0.335071
Epoch[35] Batch [499]	Speed: 2833.712721 samples/sec	accuracy=0.748871	lr=0.334910
Epoch[35] Batch [549]	Speed: 2993.965386 samples/sec	accuracy=0.748964	lr=0.334749
Epoch[35] Batch [599]	Speed: 3006.897628 samples/sec	accuracy=0.749013	lr=0.334587
Epoch[35] Batch [649]	Speed: 2896.669148 samples/sec	accuracy=0.749154	lr=0.334425
Epoch[35] Batch [699]	Speed: 2873.254255 samples/sec	accuracy=0.749231	lr=0.334264
Epoch[35] Batch [749]	Speed: 2827.063674 samples/sec	accuracy=0.749366	lr=0.334102
Epoch[35] Batch [799]	Speed: 2878.571529 samples/sec	accuracy=0.749521	lr=0.333940
Epoch[35] Batch [849]	Speed: 2838.746307 samples/sec	accuracy=0.749625	lr=0.333777
Epoch[35] Batch [899]	Speed: 3044.321358 samples/sec	accuracy=0.749764	lr=0.333615
Epoch[35] Batch [949]	Speed: 2998.027204 samples/sec	accuracy=0.749859	lr=0.333452
Epoch[35] Batch [999]	Speed: 2972.680612 samples/sec	accuracy=0.749983	lr=0.333290
Epoch[35] Batch [1049]	Speed: 2921.504154 samples/sec	accuracy=0.750055	lr=0.333127
Epoch[35] Batch [1099]	Speed: 2944.926962 samples/sec	accuracy=0.750181	lr=0.332964
Epoch[35] Batch [1149]	Speed: 2881.705070 samples/sec	accuracy=0.750303	lr=0.332801
Epoch[35] Batch [1199]	Speed: 2864.251081 samples/sec	accuracy=0.750431	lr=0.332637
Epoch[35] Batch [1249]	Speed: 3134.763995 samples/sec	accuracy=0.750656	lr=0.332474
[Epoch 35] training: accuracy=0.750666
[Epoch 35] speed: 2935 samples/sec	time cost: 446.797441
[Epoch 35] validation: err-top1=0.352380 err-top5=0.132135
Epoch[36] Batch [49]	Speed: 2989.833386 samples/sec	accuracy=0.750712	lr=0.332307
Epoch[36] Batch [99]	Speed: 2929.058945 samples/sec	accuracy=0.750811	lr=0.332143
Epoch[36] Batch [149]	Speed: 2843.747533 samples/sec	accuracy=0.750802	lr=0.331979
Epoch[36] Batch [199]	Speed: 2876.379922 samples/sec	accuracy=0.751077	lr=0.331815
Epoch[36] Batch [249]	Speed: 2903.124006 samples/sec	accuracy=0.751351	lr=0.331650
Epoch[36] Batch [299]	Speed: 2838.620228 samples/sec	accuracy=0.751221	lr=0.331486
Epoch[36] Batch [349]	Speed: 2993.986966 samples/sec	accuracy=0.751314	lr=0.331321
Epoch[36] Batch [399]	Speed: 3033.176449 samples/sec	accuracy=0.751429	lr=0.331157
Epoch[36] Batch [449]	Speed: 2892.062059 samples/sec	accuracy=0.751537	lr=0.330992
Epoch[36] Batch [499]	Speed: 2885.846413 samples/sec	accuracy=0.751523	lr=0.330827
Epoch[36] Batch [549]	Speed: 2846.820733 samples/sec	accuracy=0.751532	lr=0.330661
Epoch[36] Batch [599]	Speed: 2913.288015 samples/sec	accuracy=0.751761	lr=0.330496
Epoch[36] Batch [649]	Speed: 3025.875054 samples/sec	accuracy=0.751826	lr=0.330330
Epoch[36] Batch [699]	Speed: 2791.656556 samples/sec	accuracy=0.751929	lr=0.330165
Epoch[36] Batch [749]	Speed: 2918.160101 samples/sec	accuracy=0.751953	lr=0.329999
Epoch[36] Batch [799]	Speed: 2789.320961 samples/sec	accuracy=0.751977	lr=0.329833
Epoch[36] Batch [849]	Speed: 2978.028401 samples/sec	accuracy=0.752116	lr=0.329667
Epoch[36] Batch [899]	Speed: 2887.017835 samples/sec	accuracy=0.752319	lr=0.329500
Epoch[36] Batch [949]	Speed: 2859.715343 samples/sec	accuracy=0.752450	lr=0.329334
Epoch[36] Batch [999]	Speed: 2946.961383 samples/sec	accuracy=0.752532	lr=0.329167
Epoch[36] Batch [1049]	Speed: 3006.938720 samples/sec	accuracy=0.752661	lr=0.329000
Epoch[36] Batch [1099]	Speed: 2939.105014 samples/sec	accuracy=0.752747	lr=0.328833
Epoch[36] Batch [1149]	Speed: 2882.283218 samples/sec	accuracy=0.752926	lr=0.328666
Epoch[36] Batch [1199]	Speed: 3039.314462 samples/sec	accuracy=0.753066	lr=0.328499
Epoch[36] Batch [1249]	Speed: 3074.714819 samples/sec	accuracy=0.753133	lr=0.328332
[Epoch 36] training: accuracy=0.753136
[Epoch 36] speed: 2919 samples/sec	time cost: 448.812059
[Epoch 36] validation: err-top1=0.364856 err-top5=0.141482
Epoch[37] Batch [49]	Speed: 2826.947376 samples/sec	accuracy=0.752385	lr=0.328161
Epoch[37] Batch [99]	Speed: 2905.169916 samples/sec	accuracy=0.752159	lr=0.327993
Epoch[37] Batch [149]	Speed: 2905.626282 samples/sec	accuracy=0.752431	lr=0.327825
Epoch[37] Batch [199]	Speed: 2842.652309 samples/sec	accuracy=0.752778	lr=0.327657
Epoch[37] Batch [249]	Speed: 2894.945332 samples/sec	accuracy=0.752641	lr=0.327489
Epoch[37] Batch [299]	Speed: 3023.596025 samples/sec	accuracy=0.752786	lr=0.327321
Epoch[37] Batch [349]	Speed: 3084.871041 samples/sec	accuracy=0.752646	lr=0.327152
Epoch[37] Batch [399]	Speed: 2962.133477 samples/sec	accuracy=0.752745	lr=0.326983
Epoch[37] Batch [449]	Speed: 2900.945536 samples/sec	accuracy=0.752634	lr=0.326815
Epoch[37] Batch [499]	Speed: 2974.435901 samples/sec	accuracy=0.752664	lr=0.326646
Epoch[37] Batch [549]	Speed: 2880.662529 samples/sec	accuracy=0.752877	lr=0.326477
Epoch[37] Batch [599]	Speed: 2872.774142 samples/sec	accuracy=0.753000	lr=0.326307
Epoch[37] Batch [649]	Speed: 2995.934730 samples/sec	accuracy=0.753106	lr=0.326138
Epoch[37] Batch [699]	Speed: 3030.699060 samples/sec	accuracy=0.753245	lr=0.325968
Epoch[37] Batch [749]	Speed: 3003.350223 samples/sec	accuracy=0.753422	lr=0.325799
Epoch[37] Batch [799]	Speed: 2944.827053 samples/sec	accuracy=0.753546	lr=0.325629
Epoch[37] Batch [849]	Speed: 3079.035490 samples/sec	accuracy=0.753728	lr=0.325459
Epoch[37] Batch [899]	Speed: 2865.404269 samples/sec	accuracy=0.753814	lr=0.325289
Epoch[37] Batch [949]	Speed: 2828.190980 samples/sec	accuracy=0.753818	lr=0.325118
Epoch[37] Batch [999]	Speed: 2907.522661 samples/sec	accuracy=0.753854	lr=0.324948
Epoch[37] Batch [1049]	Speed: 2831.350379 samples/sec	accuracy=0.754021	lr=0.324777
Epoch[37] Batch [1099]	Speed: 2905.254300 samples/sec	accuracy=0.754122	lr=0.324607
Epoch[37] Batch [1149]	Speed: 2917.734714 samples/sec	accuracy=0.754185	lr=0.324436
Epoch[37] Batch [1199]	Speed: 2838.685405 samples/sec	accuracy=0.754204	lr=0.324265
Epoch[37] Batch [1249]	Speed: 3076.801384 samples/sec	accuracy=0.754306	lr=0.324094
[Epoch 37] training: accuracy=0.754308
[Epoch 37] speed: 2927 samples/sec	time cost: 447.301064
[Epoch 37] validation: err-top1=0.344886 err-top5=0.127710
Epoch[38] Batch [49]	Speed: 2959.446751 samples/sec	accuracy=0.751997	lr=0.323919
Epoch[38] Batch [99]	Speed: 2934.524738 samples/sec	accuracy=0.752747	lr=0.323747
Epoch[38] Batch [149]	Speed: 2978.056112 samples/sec	accuracy=0.753211	lr=0.323576
Epoch[38] Batch [199]	Speed: 2979.868367 samples/sec	accuracy=0.753248	lr=0.323404
Epoch[38] Batch [249]	Speed: 2942.070523 samples/sec	accuracy=0.753484	lr=0.323232
Epoch[38] Batch [299]	Speed: 2896.797623 samples/sec	accuracy=0.753483	lr=0.323060
Epoch[38] Batch [349]	Speed: 2869.362634 samples/sec	accuracy=0.753631	lr=0.322888
Epoch[38] Batch [399]	Speed: 2975.616041 samples/sec	accuracy=0.753739	lr=0.322715
Epoch[38] Batch [449]	Speed: 2857.528274 samples/sec	accuracy=0.753993	lr=0.322543
Epoch[38] Batch [499]	Speed: 2889.671499 samples/sec	accuracy=0.754032	lr=0.322370
Epoch[38] Batch [549]	Speed: 2915.773528 samples/sec	accuracy=0.754206	lr=0.322197
Epoch[38] Batch [599]	Speed: 2909.136008 samples/sec	accuracy=0.754382	lr=0.322025
Epoch[38] Batch [649]	Speed: 2937.371317 samples/sec	accuracy=0.754436	lr=0.321851
Epoch[38] Batch [699]	Speed: 2956.622619 samples/sec	accuracy=0.754506	lr=0.321678
Epoch[38] Batch [749]	Speed: 2935.237691 samples/sec	accuracy=0.754614	lr=0.321505
Epoch[38] Batch [799]	Speed: 2894.249787 samples/sec	accuracy=0.754800	lr=0.321331
Epoch[38] Batch [849]	Speed: 2964.468525 samples/sec	accuracy=0.754937	lr=0.321158
Epoch[38] Batch [899]	Speed: 2923.507127 samples/sec	accuracy=0.755013	lr=0.320984
Epoch[38] Batch [949]	Speed: 2892.013414 samples/sec	accuracy=0.755112	lr=0.320810
Epoch[38] Batch [999]	Speed: 2948.273298 samples/sec	accuracy=0.755248	lr=0.320636
Epoch[38] Batch [1049]	Speed: 2940.356073 samples/sec	accuracy=0.755374	lr=0.320462
Epoch[38] Batch [1099]	Speed: 3062.036124 samples/sec	accuracy=0.755416	lr=0.320287
Epoch[38] Batch [1149]	Speed: 2879.545294 samples/sec	accuracy=0.755568	lr=0.320113
Epoch[38] Batch [1199]	Speed: 2881.929216 samples/sec	accuracy=0.755608	lr=0.319938
Epoch[38] Batch [1249]	Speed: 3092.007005 samples/sec	accuracy=0.755671	lr=0.319763
[Epoch 38] training: accuracy=0.755673
[Epoch 38] speed: 2933 samples/sec	time cost: 446.294023
[Epoch 38] validation: err-top1=0.348214 err-top5=0.130401
Epoch[39] Batch [49]	Speed: 3095.130580 samples/sec	accuracy=0.754597	lr=0.319585
Epoch[39] Batch [99]	Speed: 3052.214294 samples/sec	accuracy=0.755818	lr=0.319410
Epoch[39] Batch [149]	Speed: 2831.695686 samples/sec	accuracy=0.755290	lr=0.319234
Epoch[39] Batch [199]	Speed: 2933.422840 samples/sec	accuracy=0.755087	lr=0.319059
Epoch[39] Batch [249]	Speed: 2846.161623 samples/sec	accuracy=0.755314	lr=0.318883
Epoch[39] Batch [299]	Speed: 2936.957623 samples/sec	accuracy=0.755462	lr=0.318708
Epoch[39] Batch [349]	Speed: 2853.609790 samples/sec	accuracy=0.755594	lr=0.318532
Epoch[39] Batch [399]	Speed: 2861.140764 samples/sec	accuracy=0.755938	lr=0.318356
Epoch[39] Batch [449]	Speed: 2851.763637 samples/sec	accuracy=0.756063	lr=0.318180
Epoch[39] Batch [499]	Speed: 2974.244423 samples/sec	accuracy=0.756118	lr=0.318004
Epoch[39] Batch [549]	Speed: 2839.981236 samples/sec	accuracy=0.756088	lr=0.317827
Epoch[39] Batch [599]	Speed: 2939.422588 samples/sec	accuracy=0.756131	lr=0.317651
Epoch[39] Batch [649]	Speed: 3065.346211 samples/sec	accuracy=0.756079	lr=0.317474
Epoch[39] Batch [699]	Speed: 2863.198530 samples/sec	accuracy=0.756108	lr=0.317297
Epoch[39] Batch [749]	Speed: 2840.807262 samples/sec	accuracy=0.756211	lr=0.317120
Epoch[39] Batch [799]	Speed: 2964.973392 samples/sec	accuracy=0.756289	lr=0.316943
Epoch[39] Batch [849]	Speed: 2826.579080 samples/sec	accuracy=0.756417	lr=0.316766
Epoch[39] Batch [899]	Speed: 2851.872480 samples/sec	accuracy=0.756575	lr=0.316589
Epoch[39] Batch [949]	Speed: 3023.894437 samples/sec	accuracy=0.756705	lr=0.316411
Epoch[39] Batch [999]	Speed: 2953.468439 samples/sec	accuracy=0.756840	lr=0.316233
Epoch[39] Batch [1049]	Speed: 2843.456357 samples/sec	accuracy=0.757008	lr=0.316056
Epoch[39] Batch [1099]	Speed: 3097.809292 samples/sec	accuracy=0.757040	lr=0.315878
Epoch[39] Batch [1149]	Speed: 2981.085473 samples/sec	accuracy=0.757110	lr=0.315700
Epoch[39] Batch [1199]	Speed: 2991.475141 samples/sec	accuracy=0.757214	lr=0.315522
Epoch[39] Batch [1249]	Speed: 3060.659117 samples/sec	accuracy=0.757334	lr=0.315343
[Epoch 39] training: accuracy=0.757336
[Epoch 39] speed: 2930 samples/sec	time cost: 446.897073
[Epoch 39] validation: err-top1=0.353655 err-top5=0.134666
Epoch[40] Batch [49]	Speed: 2929.462864 samples/sec	accuracy=0.756752	lr=0.315161
Epoch[40] Batch [99]	Speed: 3121.749655 samples/sec	accuracy=0.757724	lr=0.314983
Epoch[40] Batch [149]	Speed: 2948.022566 samples/sec	accuracy=0.757366	lr=0.314804
Epoch[40] Batch [199]	Speed: 3076.191179 samples/sec	accuracy=0.757555	lr=0.314625
Epoch[40] Batch [249]	Speed: 2982.541573 samples/sec	accuracy=0.757402	lr=0.314446
Epoch[40] Batch [299]	Speed: 2938.727990 samples/sec	accuracy=0.757446	lr=0.314267
Epoch[40] Batch [349]	Speed: 3079.264938 samples/sec	accuracy=0.757380	lr=0.314087
Epoch[40] Batch [399]	Speed: 3131.449401 samples/sec	accuracy=0.757522	lr=0.313908
Epoch[40] Batch [449]	Speed: 3025.226578 samples/sec	accuracy=0.757419	lr=0.313728
Epoch[40] Batch [499]	Speed: 3089.908131 samples/sec	accuracy=0.757546	lr=0.313549
Epoch[40] Batch [549]	Speed: 3171.431961 samples/sec	accuracy=0.757652	lr=0.313369
Epoch[40] Batch [599]	Speed: 3209.167203 samples/sec	accuracy=0.757617	lr=0.313189
Epoch[40] Batch [649]	Speed: 3058.895472 samples/sec	accuracy=0.757653	lr=0.313009
Epoch[40] Batch [699]	Speed: 2961.672015 samples/sec	accuracy=0.757747	lr=0.312829
Epoch[40] Batch [749]	Speed: 2980.167598 samples/sec	accuracy=0.757900	lr=0.312648
Epoch[40] Batch [799]	Speed: 2994.731196 samples/sec	accuracy=0.757943	lr=0.312468
Epoch[40] Batch [849]	Speed: 2962.427767 samples/sec	accuracy=0.758019	lr=0.312287
Epoch[40] Batch [899]	Speed: 2948.982945 samples/sec	accuracy=0.758144	lr=0.312106
Epoch[40] Batch [949]	Speed: 2915.328414 samples/sec	accuracy=0.758267	lr=0.311925
Epoch[40] Batch [999]	Speed: 2871.427218 samples/sec	accuracy=0.758340	lr=0.311744
Epoch[40] Batch [1049]	Speed: 2899.462884 samples/sec	accuracy=0.758495	lr=0.311563
Epoch[40] Batch [1099]	Speed: 2874.393737 samples/sec	accuracy=0.758635	lr=0.311382
Epoch[40] Batch [1149]	Speed: 2995.505965 samples/sec	accuracy=0.758744	lr=0.311201
Epoch[40] Batch [1199]	Speed: 2881.332537 samples/sec	accuracy=0.758812	lr=0.311019
Epoch[40] Batch [1249]	Speed: 3166.633378 samples/sec	accuracy=0.758851	lr=0.310837
[Epoch 40] training: accuracy=0.758851
[Epoch 40] speed: 3003 samples/sec	time cost: 436.051588
[Epoch 40] validation: err-top1=0.341797 err-top5=0.124593
Epoch[41] Batch [49]	Speed: 3098.438387 samples/sec	accuracy=0.758026	lr=0.310652
Epoch[41] Batch [99]	Speed: 2805.085364 samples/sec	accuracy=0.757899	lr=0.310470
Epoch[41] Batch [149]	Speed: 2883.112439 samples/sec	accuracy=0.758464	lr=0.310288
Epoch[41] Batch [199]	Speed: 2928.646950 samples/sec	accuracy=0.758689	lr=0.310105
Epoch[41] Batch [249]	Speed: 2880.437073 samples/sec	accuracy=0.758857	lr=0.309923
Epoch[41] Batch [299]	Speed: 2979.190936 samples/sec	accuracy=0.758897	lr=0.309741
Epoch[41] Batch [349]	Speed: 3117.855815 samples/sec	accuracy=0.759014	lr=0.309558
Epoch[41] Batch [399]	Speed: 2943.859370 samples/sec	accuracy=0.758739	lr=0.309375
Epoch[41] Batch [449]	Speed: 2971.553456 samples/sec	accuracy=0.759060	lr=0.309192
Epoch[41] Batch [499]	Speed: 2970.046734 samples/sec	accuracy=0.759074	lr=0.309009
Epoch[41] Batch [549]	Speed: 2894.010148 samples/sec	accuracy=0.759062	lr=0.308826
Epoch[41] Batch [599]	Speed: 2996.320432 samples/sec	accuracy=0.759173	lr=0.308643
Epoch[41] Batch [649]	Speed: 2896.395473 samples/sec	accuracy=0.759338	lr=0.308459
Epoch[41] Batch [699]	Speed: 3021.234648 samples/sec	accuracy=0.759369	lr=0.308276
Epoch[41] Batch [749]	Speed: 2938.535533 samples/sec	accuracy=0.759514	lr=0.308092
Epoch[41] Batch [799]	Speed: 2924.745299 samples/sec	accuracy=0.759642	lr=0.307908
Epoch[41] Batch [849]	Speed: 2956.177644 samples/sec	accuracy=0.759706	lr=0.307725
Epoch[41] Batch [899]	Speed: 2934.397065 samples/sec	accuracy=0.759724	lr=0.307540
Epoch[41] Batch [949]	Speed: 2983.567729 samples/sec	accuracy=0.759778	lr=0.307356
Epoch[41] Batch [999]	Speed: 2904.257495 samples/sec	accuracy=0.759832	lr=0.307172
Epoch[41] Batch [1049]	Speed: 3042.308984 samples/sec	accuracy=0.759886	lr=0.306988
Epoch[41] Batch [1099]	Speed: 2949.314363 samples/sec	accuracy=0.759942	lr=0.306803
Epoch[41] Batch [1149]	Speed: 2788.690161 samples/sec	accuracy=0.760009	lr=0.306618
Epoch[41] Batch [1199]	Speed: 2956.731431 samples/sec	accuracy=0.760150	lr=0.306433
Epoch[41] Batch [1249]	Speed: 2985.844954 samples/sec	accuracy=0.760269	lr=0.306249
[Epoch 41] training: accuracy=0.760270
[Epoch 41] speed: 2945 samples/sec	time cost: 444.685325
[Epoch 41] validation: err-top1=0.337612 err-top5=0.122449
Epoch[42] Batch [49]	Speed: 2851.861194 samples/sec	accuracy=0.761045	lr=0.306060
Epoch[42] Batch [99]	Speed: 2759.075427 samples/sec	accuracy=0.760846	lr=0.305875
Epoch[42] Batch [149]	Speed: 2879.776905 samples/sec	accuracy=0.759999	lr=0.305689
Epoch[42] Batch [199]	Speed: 2885.963341 samples/sec	accuracy=0.759852	lr=0.305504
Epoch[42] Batch [249]	Speed: 2809.161975 samples/sec	accuracy=0.759539	lr=0.305318
Epoch[42] Batch [299]	Speed: 3110.757391 samples/sec	accuracy=0.760146	lr=0.305133
Epoch[42] Batch [349]	Speed: 2902.209379 samples/sec	accuracy=0.760302	lr=0.304947
Epoch[42] Batch [399]	Speed: 2896.182351 samples/sec	accuracy=0.760184	lr=0.304761
Epoch[42] Batch [449]	Speed: 2884.445478 samples/sec	accuracy=0.760245	lr=0.304575
Epoch[42] Batch [499]	Speed: 2970.247449 samples/sec	accuracy=0.760425	lr=0.304388
Epoch[42] Batch [549]	Speed: 2886.810476 samples/sec	accuracy=0.760541	lr=0.304202
Epoch[42] Batch [599]	Speed: 2960.247683 samples/sec	accuracy=0.760629	lr=0.304016
Epoch[42] Batch [649]	Speed: 2889.872464 samples/sec	accuracy=0.760713	lr=0.303829
Epoch[42] Batch [699]	Speed: 2900.949964 samples/sec	accuracy=0.760832	lr=0.303642
Epoch[42] Batch [749]	Speed: 2899.967546 samples/sec	accuracy=0.760929	lr=0.303456
Epoch[42] Batch [799]	Speed: 3048.239785 samples/sec	accuracy=0.761022	lr=0.303269
Epoch[42] Batch [849]	Speed: 2814.639730 samples/sec	accuracy=0.761059	lr=0.303082
Epoch[42] Batch [899]	Speed: 2952.866458 samples/sec	accuracy=0.761224	lr=0.302894
Epoch[42] Batch [949]	Speed: 2850.679032 samples/sec	accuracy=0.761389	lr=0.302707
Epoch[42] Batch [999]	Speed: 2867.441947 samples/sec	accuracy=0.761510	lr=0.302520
Epoch[42] Batch [1049]	Speed: 2903.459995 samples/sec	accuracy=0.761588	lr=0.302332
Epoch[42] Batch [1099]	Speed: 2848.000830 samples/sec	accuracy=0.761694	lr=0.302144
Epoch[42] Batch [1149]	Speed: 2875.303155 samples/sec	accuracy=0.761800	lr=0.301957
Epoch[42] Batch [1199]	Speed: 2866.638437 samples/sec	accuracy=0.761918	lr=0.301769
Epoch[42] Batch [1249]	Speed: 2997.253055 samples/sec	accuracy=0.761991	lr=0.301581
[Epoch 42] training: accuracy=0.761996
[Epoch 42] speed: 2897 samples/sec	time cost: 452.240693
[Epoch 42] validation: err-top1=0.343850 err-top5=0.125757
Epoch[43] Batch [49]	Speed: 2892.633968 samples/sec	accuracy=0.762301	lr=0.301389
Epoch[43] Batch [99]	Speed: 2953.359217 samples/sec	accuracy=0.761807	lr=0.301200
Epoch[43] Batch [149]	Speed: 2877.573977 samples/sec	accuracy=0.762062	lr=0.301012
Epoch[43] Batch [199]	Speed: 2856.091902 samples/sec	accuracy=0.762003	lr=0.300823
Epoch[43] Batch [249]	Speed: 2806.882562 samples/sec	accuracy=0.761716	lr=0.300635
Epoch[43] Batch [299]	Speed: 2823.836406 samples/sec	accuracy=0.761824	lr=0.300446
Epoch[43] Batch [349]	Speed: 2923.843552 samples/sec	accuracy=0.761844	lr=0.300257
Epoch[43] Batch [399]	Speed: 3045.642103 samples/sec	accuracy=0.761934	lr=0.300068
Epoch[43] Batch [449]	Speed: 2861.837074 samples/sec	accuracy=0.762017	lr=0.299879
Epoch[43] Batch [499]	Speed: 2887.296728 samples/sec	accuracy=0.762160	lr=0.299690
Epoch[43] Batch [549]	Speed: 2883.947792 samples/sec	accuracy=0.762349	lr=0.299500
Epoch[43] Batch [599]	Speed: 2860.055110 samples/sec	accuracy=0.762404	lr=0.299311
Epoch[43] Batch [649]	Speed: 2988.172519 samples/sec	accuracy=0.762470	lr=0.299121
Epoch[43] Batch [699]	Speed: 2882.647020 samples/sec	accuracy=0.762533	lr=0.298932
Epoch[43] Batch [749]	Speed: 2959.796598 samples/sec	accuracy=0.762469	lr=0.298742
Epoch[43] Batch [799]	Speed: 2870.965793 samples/sec	accuracy=0.762518	lr=0.298552
Epoch[43] Batch [849]	Speed: 2824.316011 samples/sec	accuracy=0.762597	lr=0.298362
Epoch[43] Batch [899]	Speed: 2972.147572 samples/sec	accuracy=0.762627	lr=0.298171
Epoch[43] Batch [949]	Speed: 2983.017021 samples/sec	accuracy=0.762704	lr=0.297981
Epoch[43] Batch [999]	Speed: 2914.280509 samples/sec	accuracy=0.762741	lr=0.297791
Epoch[43] Batch [1049]	Speed: 2840.039264 samples/sec	accuracy=0.762930	lr=0.297600
Epoch[43] Batch [1099]	Speed: 3049.558389 samples/sec	accuracy=0.763019	lr=0.297409
Epoch[43] Batch [1149]	Speed: 2908.558580 samples/sec	accuracy=0.763068	lr=0.297219
Epoch[43] Batch [1199]	Speed: 2870.669938 samples/sec	accuracy=0.763164	lr=0.297028
Epoch[43] Batch [1249]	Speed: 3079.007854 samples/sec	accuracy=0.763245	lr=0.296837
[Epoch 43] training: accuracy=0.763244
[Epoch 43] speed: 2908 samples/sec	time cost: 449.952784
[Epoch 43] validation: err-top1=0.344707 err-top5=0.126076
Epoch[44] Batch [49]	Speed: 3266.437723 samples/sec	accuracy=0.762797	lr=0.296642
Epoch[44] Batch [99]	Speed: 3053.871482 samples/sec	accuracy=0.763075	lr=0.296451
Epoch[44] Batch [149]	Speed: 3012.359716 samples/sec	accuracy=0.762827	lr=0.296259
Epoch[44] Batch [199]	Speed: 3008.276863 samples/sec	accuracy=0.763083	lr=0.296068
Epoch[44] Batch [249]	Speed: 3161.674588 samples/sec	accuracy=0.763331	lr=0.295876
Epoch[44] Batch [299]	Speed: 3008.834282 samples/sec	accuracy=0.763464	lr=0.295684
Epoch[44] Batch [349]	Speed: 3093.794475 samples/sec	accuracy=0.763283	lr=0.295493
Epoch[44] Batch [399]	Speed: 2975.400625 samples/sec	accuracy=0.763409	lr=0.295301
Epoch[44] Batch [449]	Speed: 3018.957805 samples/sec	accuracy=0.763457	lr=0.295109
Epoch[44] Batch [499]	Speed: 2899.168524 samples/sec	accuracy=0.763469	lr=0.294917
Epoch[44] Batch [549]	Speed: 2949.255753 samples/sec	accuracy=0.763531	lr=0.294724
Epoch[44] Batch [599]	Speed: 2817.100121 samples/sec	accuracy=0.763587	lr=0.294532
Epoch[44] Batch [649]	Speed: 2891.678550 samples/sec	accuracy=0.763635	lr=0.294339
Epoch[44] Batch [699]	Speed: 2948.699823 samples/sec	accuracy=0.763698	lr=0.294147
Epoch[44] Batch [749]	Speed: 2921.995485 samples/sec	accuracy=0.763807	lr=0.293954
Epoch[44] Batch [799]	Speed: 2864.296238 samples/sec	accuracy=0.763798	lr=0.293761
Epoch[44] Batch [849]	Speed: 3082.614966 samples/sec	accuracy=0.763906	lr=0.293568
Epoch[44] Batch [899]	Speed: 2959.588401 samples/sec	accuracy=0.764001	lr=0.293375
Epoch[44] Batch [949]	Speed: 2978.789881 samples/sec	accuracy=0.764143	lr=0.293182
Epoch[44] Batch [999]	Speed: 2932.879751 samples/sec	accuracy=0.764266	lr=0.292989
Epoch[44] Batch [1049]	Speed: 2854.844632 samples/sec	accuracy=0.764279	lr=0.292795
Epoch[44] Batch [1099]	Speed: 2957.507473 samples/sec	accuracy=0.764328	lr=0.292602
Epoch[44] Batch [1149]	Speed: 2845.683734 samples/sec	accuracy=0.764391	lr=0.292408
Epoch[44] Batch [1199]	Speed: 2938.466896 samples/sec	accuracy=0.764489	lr=0.292215
Epoch[44] Batch [1249]	Speed: 3030.884444 samples/sec	accuracy=0.764558	lr=0.292021
[Epoch 44] training: accuracy=0.764562
[Epoch 44] speed: 2973 samples/sec	time cost: 440.894573
[Epoch 44] validation: err-top1=0.334742 err-top5=0.121293
Epoch[45] Batch [49]	Speed: 2837.301833 samples/sec	accuracy=0.764784	lr=0.291823
Epoch[45] Batch [99]	Speed: 2886.205917 samples/sec	accuracy=0.764387	lr=0.291629
Epoch[45] Batch [149]	Speed: 2723.697934 samples/sec	accuracy=0.764539	lr=0.291435
Epoch[45] Batch [199]	Speed: 2875.154291 samples/sec	accuracy=0.764301	lr=0.291240
Epoch[45] Batch [249]	Speed: 2915.826974 samples/sec	accuracy=0.764244	lr=0.291046
Epoch[45] Batch [299]	Speed: 2835.251912 samples/sec	accuracy=0.764407	lr=0.290852
Epoch[45] Batch [349]	Speed: 3051.283837 samples/sec	accuracy=0.764307	lr=0.290657
Epoch[45] Batch [399]	Speed: 2917.282700 samples/sec	accuracy=0.764408	lr=0.290462
Epoch[45] Batch [449]	Speed: 3004.654442 samples/sec	accuracy=0.764477	lr=0.290267
Epoch[45] Batch [499]	Speed: 2915.677685 samples/sec	accuracy=0.764416	lr=0.290073
Epoch[45] Batch [549]	Speed: 2887.368236 samples/sec	accuracy=0.764555	lr=0.289878
Epoch[45] Batch [599]	Speed: 2909.509814 samples/sec	accuracy=0.764647	lr=0.289682
Epoch[45] Batch [649]	Speed: 2834.719604 samples/sec	accuracy=0.764825	lr=0.289487
Epoch[45] Batch [699]	Speed: 2837.720026 samples/sec	accuracy=0.765023	lr=0.289292
Epoch[45] Batch [749]	Speed: 2975.067935 samples/sec	accuracy=0.765087	lr=0.289096
Epoch[45] Batch [799]	Speed: 2907.141141 samples/sec	accuracy=0.765147	lr=0.288901
Epoch[45] Batch [849]	Speed: 2812.799834 samples/sec	accuracy=0.765259	lr=0.288705
Epoch[45] Batch [899]	Speed: 2868.305059 samples/sec	accuracy=0.765374	lr=0.288509
Epoch[45] Batch [949]	Speed: 2871.383603 samples/sec	accuracy=0.765400	lr=0.288314
Epoch[45] Batch [999]	Speed: 2814.414051 samples/sec	accuracy=0.765492	lr=0.288118
Epoch[45] Batch [1049]	Speed: 2872.592494 samples/sec	accuracy=0.765496	lr=0.287921
Epoch[45] Batch [1099]	Speed: 2994.998333 samples/sec	accuracy=0.765579	lr=0.287725
Epoch[45] Batch [1149]	Speed: 2863.848292 samples/sec	accuracy=0.765660	lr=0.287529
Epoch[45] Batch [1199]	Speed: 2874.262548 samples/sec	accuracy=0.765744	lr=0.287333
Epoch[45] Batch [1249]	Speed: 3008.793433 samples/sec	accuracy=0.765869	lr=0.287136
[Epoch 45] training: accuracy=0.765870
[Epoch 45] speed: 2887 samples/sec	time cost: 453.342272
[Epoch 45] validation: err-top1=0.331254 err-top5=0.116769
Epoch[46] Batch [49]	Speed: 3029.912993 samples/sec	accuracy=0.767077	lr=0.286936
Epoch[46] Batch [99]	Speed: 3032.372525 samples/sec	accuracy=0.766212	lr=0.286739
Epoch[46] Batch [149]	Speed: 2989.325969 samples/sec	accuracy=0.765727	lr=0.286542
Epoch[46] Batch [199]	Speed: 2932.144763 samples/sec	accuracy=0.765344	lr=0.286345
Epoch[46] Batch [249]	Speed: 2929.545788 samples/sec	accuracy=0.765261	lr=0.286148
Epoch[46] Batch [299]	Speed: 2915.784494 samples/sec	accuracy=0.765171	lr=0.285951
Epoch[46] Batch [349]	Speed: 2864.875332 samples/sec	accuracy=0.764958	lr=0.285754
Epoch[46] Batch [399]	Speed: 2956.876567 samples/sec	accuracy=0.765249	lr=0.285556
Epoch[46] Batch [449]	Speed: 2842.158068 samples/sec	accuracy=0.765347	lr=0.285359
Epoch[46] Batch [499]	Speed: 2847.697945 samples/sec	accuracy=0.765668	lr=0.285161
Epoch[46] Batch [549]	Speed: 2831.364639 samples/sec	accuracy=0.765923	lr=0.284964
Epoch[46] Batch [599]	Speed: 2897.705362 samples/sec	accuracy=0.766074	lr=0.284766
Epoch[46] Batch [649]	Speed: 2766.943714 samples/sec	accuracy=0.766264	lr=0.284568
Epoch[46] Batch [699]	Speed: 2903.457718 samples/sec	accuracy=0.766413	lr=0.284370
Epoch[46] Batch [749]	Speed: 2923.264768 samples/sec	accuracy=0.766557	lr=0.284172
Epoch[46] Batch [799]	Speed: 2784.948357 samples/sec	accuracy=0.766690	lr=0.283974
Epoch[46] Batch [849]	Speed: 2959.240602 samples/sec	accuracy=0.766849	lr=0.283776
Epoch[46] Batch [899]	Speed: 2766.261595 samples/sec	accuracy=0.766965	lr=0.283577
Epoch[46] Batch [949]	Speed: 2909.334566 samples/sec	accuracy=0.767077	lr=0.283379
Epoch[46] Batch [999]	Speed: 2911.571883 samples/sec	accuracy=0.767136	lr=0.283181
Epoch[46] Batch [1049]	Speed: 2878.153863 samples/sec	accuracy=0.767163	lr=0.282982
Epoch[46] Batch [1099]	Speed: 2847.475429 samples/sec	accuracy=0.767281	lr=0.282783
Epoch[46] Batch [1149]	Speed: 2806.259196 samples/sec	accuracy=0.767381	lr=0.282584
Epoch[46] Batch [1199]	Speed: 2936.243995 samples/sec	accuracy=0.767426	lr=0.282385
Epoch[46] Batch [1249]	Speed: 3041.887267 samples/sec	accuracy=0.767507	lr=0.282186
[Epoch 46] training: accuracy=0.767506
[Epoch 46] speed: 2896 samples/sec	time cost: 452.017814
[Epoch 46] validation: err-top1=0.331462 err-top5=0.119161
Epoch[47] Batch [49]	Speed: 2994.066570 samples/sec	accuracy=0.767346	lr=0.281983
Epoch[47] Batch [99]	Speed: 2974.451968 samples/sec	accuracy=0.766576	lr=0.281784
Epoch[47] Batch [149]	Speed: 2893.187745 samples/sec	accuracy=0.766928	lr=0.281585
Epoch[47] Batch [199]	Speed: 2943.472329 samples/sec	accuracy=0.766970	lr=0.281385
Epoch[47] Batch [249]	Speed: 2882.711642 samples/sec	accuracy=0.767207	lr=0.281186
Epoch[47] Batch [299]	Speed: 2934.756856 samples/sec	accuracy=0.767252	lr=0.280986
Epoch[47] Batch [349]	Speed: 2954.366891 samples/sec	accuracy=0.767442	lr=0.280786
Epoch[47] Batch [399]	Speed: 2906.015624 samples/sec	accuracy=0.767363	lr=0.280587
Epoch[47] Batch [449]	Speed: 2910.902773 samples/sec	accuracy=0.767261	lr=0.280387
Epoch[47] Batch [499]	Speed: 2895.195430 samples/sec	accuracy=0.767329	lr=0.280187
Epoch[47] Batch [549]	Speed: 2926.200216 samples/sec	accuracy=0.767463	lr=0.279987
Epoch[47] Batch [599]	Speed: 2855.617355 samples/sec	accuracy=0.767600	lr=0.279786
Epoch[47] Batch [649]	Speed: 2907.493767 samples/sec	accuracy=0.767590	lr=0.279586
Epoch[47] Batch [699]	Speed: 2880.955771 samples/sec	accuracy=0.767639	lr=0.279386
Epoch[47] Batch [749]	Speed: 2962.761152 samples/sec	accuracy=0.767760	lr=0.279185
Epoch[47] Batch [799]	Speed: 2909.107279 samples/sec	accuracy=0.767939	lr=0.278985
Epoch[47] Batch [849]	Speed: 3088.459708 samples/sec	accuracy=0.767958	lr=0.278784
Epoch[47] Batch [899]	Speed: 3025.196405 samples/sec	accuracy=0.768160	lr=0.278583
Epoch[47] Batch [949]	Speed: 2775.804192 samples/sec	accuracy=0.768262	lr=0.278382
Epoch[47] Batch [999]	Speed: 2967.782828 samples/sec	accuracy=0.768260	lr=0.278181
Epoch[47] Batch [1049]	Speed: 2960.204715 samples/sec	accuracy=0.768333	lr=0.277980
Epoch[47] Batch [1099]	Speed: 2851.431062 samples/sec	accuracy=0.768479	lr=0.277779
Epoch[47] Batch [1149]	Speed: 2964.263230 samples/sec	accuracy=0.768639	lr=0.277578
Epoch[47] Batch [1199]	Speed: 2943.523972 samples/sec	accuracy=0.768738	lr=0.277377
Epoch[47] Batch [1249]	Speed: 3024.113951 samples/sec	accuracy=0.768872	lr=0.277175
[Epoch 47] training: accuracy=0.768874
[Epoch 47] speed: 2929 samples/sec	time cost: 447.627157
[Epoch 47] validation: err-top1=0.338349 err-top5=0.122947
Epoch[48] Batch [49]	Speed: 3206.945249 samples/sec	accuracy=0.767032	lr=0.276970
Epoch[48] Batch [99]	Speed: 2923.602132 samples/sec	accuracy=0.768482	lr=0.276768
Epoch[48] Batch [149]	Speed: 2893.393448 samples/sec	accuracy=0.768778	lr=0.276566
Epoch[48] Batch [199]	Speed: 2967.500390 samples/sec	accuracy=0.768319	lr=0.276365
Epoch[48] Batch [249]	Speed: 2999.552484 samples/sec	accuracy=0.768263	lr=0.276163
Epoch[48] Batch [299]	Speed: 2897.884647 samples/sec	accuracy=0.768295	lr=0.275961
Epoch[48] Batch [349]	Speed: 2857.677220 samples/sec	accuracy=0.768510	lr=0.275759
Epoch[48] Batch [399]	Speed: 3045.854980 samples/sec	accuracy=0.768702	lr=0.275557
Epoch[48] Batch [449]	Speed: 2915.815176 samples/sec	accuracy=0.768729	lr=0.275354
Epoch[48] Batch [499]	Speed: 2958.774985 samples/sec	accuracy=0.768688	lr=0.275152
Epoch[48] Batch [549]	Speed: 2949.702780 samples/sec	accuracy=0.768737	lr=0.274950
Epoch[48] Batch [599]	Speed: 2907.573011 samples/sec	accuracy=0.768830	lr=0.274747
Epoch[48] Batch [649]	Speed: 3107.365150 samples/sec	accuracy=0.768957	lr=0.274545
Epoch[48] Batch [699]	Speed: 3043.092096 samples/sec	accuracy=0.769030	lr=0.274342
Epoch[48] Batch [749]	Speed: 2821.151787 samples/sec	accuracy=0.769044	lr=0.274139
Epoch[48] Batch [799]	Speed: 2952.312251 samples/sec	accuracy=0.769016	lr=0.273936
Epoch[48] Batch [849]	Speed: 2830.504997 samples/sec	accuracy=0.769191	lr=0.273733
Epoch[48] Batch [899]	Speed: 2923.604281 samples/sec	accuracy=0.769321	lr=0.273530
Epoch[48] Batch [949]	Speed: 3029.130154 samples/sec	accuracy=0.769517	lr=0.273327
Epoch[48] Batch [999]	Speed: 2865.955202 samples/sec	accuracy=0.769618	lr=0.273124
Epoch[48] Batch [1049]	Speed: 2856.387457 samples/sec	accuracy=0.769625	lr=0.272921
Epoch[48] Batch [1099]	Speed: 2933.635747 samples/sec	accuracy=0.769691	lr=0.272717
Epoch[48] Batch [1149]	Speed: 2841.291747 samples/sec	accuracy=0.769794	lr=0.272514
Epoch[48] Batch [1199]	Speed: 2981.623090 samples/sec	accuracy=0.769901	lr=0.272310
Epoch[48] Batch [1249]	Speed: 2961.972382 samples/sec	accuracy=0.770007	lr=0.272107
[Epoch 48] training: accuracy=0.770011
[Epoch 48] speed: 2942 samples/sec	time cost: 444.892121
[Epoch 48] validation: err-top1=0.329420 err-top5=0.119420
Epoch[49] Batch [49]	Speed: 2872.776064 samples/sec	accuracy=0.770163	lr=0.271899
Epoch[49] Batch [99]	Speed: 2938.744840 samples/sec	accuracy=0.769860	lr=0.271695
Epoch[49] Batch [149]	Speed: 2965.776663 samples/sec	accuracy=0.770042	lr=0.271491
Epoch[49] Batch [199]	Speed: 2860.597929 samples/sec	accuracy=0.769873	lr=0.271287
Epoch[49] Batch [249]	Speed: 3090.585748 samples/sec	accuracy=0.769771	lr=0.271083
Epoch[49] Batch [299]	Speed: 2885.857116 samples/sec	accuracy=0.770130	lr=0.270879
Epoch[49] Batch [349]	Speed: 2825.735875 samples/sec	accuracy=0.770172	lr=0.270675
Epoch[49] Batch [399]	Speed: 2897.826108 samples/sec	accuracy=0.770087	lr=0.270470
Epoch[49] Batch [449]	Speed: 2962.947024 samples/sec	accuracy=0.770177	lr=0.270266
Epoch[49] Batch [499]	Speed: 3090.191184 samples/sec	accuracy=0.770088	lr=0.270061
Epoch[49] Batch [549]	Speed: 2914.200188 samples/sec	accuracy=0.770069	lr=0.269857
Epoch[49] Batch [599]	Speed: 2889.837387 samples/sec	accuracy=0.770292	lr=0.269652
Epoch[49] Batch [649]	Speed: 3002.042352 samples/sec	accuracy=0.770475	lr=0.269447
Epoch[49] Batch [699]	Speed: 3031.294644 samples/sec	accuracy=0.770538	lr=0.269243
Epoch[49] Batch [749]	Speed: 2886.234389 samples/sec	accuracy=0.770624	lr=0.269038
Epoch[49] Batch [799]	Speed: 2919.037514 samples/sec	accuracy=0.770690	lr=0.268833
Epoch[49] Batch [849]	Speed: 2873.668346 samples/sec	accuracy=0.770715	lr=0.268628
Epoch[49] Batch [899]	Speed: 2938.505175 samples/sec	accuracy=0.770706	lr=0.268422
Epoch[49] Batch [949]	Speed: 2952.566433 samples/sec	accuracy=0.770814	lr=0.268217
Epoch[49] Batch [999]	Speed: 2945.465068 samples/sec	accuracy=0.770904	lr=0.268012
Epoch[49] Batch [1049]	Speed: 2820.259628 samples/sec	accuracy=0.770992	lr=0.267807
Epoch[49] Batch [1099]	Speed: 2811.671878 samples/sec	accuracy=0.771078	lr=0.267601
Epoch[49] Batch [1149]	Speed: 2947.951786 samples/sec	accuracy=0.771138	lr=0.267396
Epoch[49] Batch [1199]	Speed: 2809.515122 samples/sec	accuracy=0.771212	lr=0.267190
Epoch[49] Batch [1249]	Speed: 2965.145950 samples/sec	accuracy=0.771249	lr=0.266984
[Epoch 49] training: accuracy=0.771252
[Epoch 49] speed: 2919 samples/sec	time cost: 448.333469
[Epoch 49] validation: err-top1=0.329141 err-top5=0.114537
Epoch[50] Batch [49]	Speed: 2951.429164 samples/sec	accuracy=0.770786	lr=0.266774
Epoch[50] Batch [99]	Speed: 2979.224744 samples/sec	accuracy=0.770669	lr=0.266568
Epoch[50] Batch [149]	Speed: 2908.087627 samples/sec	accuracy=0.770863	lr=0.266362
Epoch[50] Batch [199]	Speed: 2913.159378 samples/sec	accuracy=0.771529	lr=0.266156
Epoch[50] Batch [249]	Speed: 3101.470157 samples/sec	accuracy=0.771443	lr=0.265950
Epoch[50] Batch [299]	Speed: 2971.960873 samples/sec	accuracy=0.771564	lr=0.265744
Epoch[50] Batch [349]	Speed: 2969.287130 samples/sec	accuracy=0.771414	lr=0.265538
Epoch[50] Batch [399]	Speed: 3033.477955 samples/sec	accuracy=0.771678	lr=0.265331
Epoch[50] Batch [449]	Speed: 2857.045307 samples/sec	accuracy=0.771704	lr=0.265125
Epoch[50] Batch [499]	Speed: 2979.134439 samples/sec	accuracy=0.771872	lr=0.264918
Epoch[50] Batch [549]	Speed: 3130.414206 samples/sec	accuracy=0.771868	lr=0.264712
Epoch[50] Batch [599]	Speed: 3142.548916 samples/sec	accuracy=0.771938	lr=0.264505
Epoch[50] Batch [649]	Speed: 2918.524568 samples/sec	accuracy=0.772041	lr=0.264298
Epoch[50] Batch [699]	Speed: 3095.997542 samples/sec	accuracy=0.772197	lr=0.264092
Epoch[50] Batch [749]	Speed: 3019.379599 samples/sec	accuracy=0.772293	lr=0.263885
Epoch[50] Batch [799]	Speed: 2968.770287 samples/sec	accuracy=0.772429	lr=0.263678
Epoch[50] Batch [849]	Speed: 2897.975686 samples/sec	accuracy=0.772463	lr=0.263471
Epoch[50] Batch [899]	Speed: 2865.804590 samples/sec	accuracy=0.772569	lr=0.263264
Epoch[50] Batch [949]	Speed: 2982.254248 samples/sec	accuracy=0.772595	lr=0.263056
Epoch[50] Batch [999]	Speed: 2886.474876 samples/sec	accuracy=0.772736	lr=0.262849
Epoch[50] Batch [1049]	Speed: 2862.831796 samples/sec	accuracy=0.772819	lr=0.262642
Epoch[50] Batch [1099]	Speed: 2868.439574 samples/sec	accuracy=0.772848	lr=0.262434
Epoch[50] Batch [1149]	Speed: 2986.117110 samples/sec	accuracy=0.772984	lr=0.262227
Epoch[50] Batch [1199]	Speed: 2919.956943 samples/sec	accuracy=0.773144	lr=0.262019
Epoch[50] Batch [1249]	Speed: 2986.762011 samples/sec	accuracy=0.773251	lr=0.261812
[Epoch 50] training: accuracy=0.773260
[Epoch 50] speed: 2963 samples/sec	time cost: 441.885443
[Epoch 50] validation: err-top1=0.322844 err-top5=0.114437
Epoch[51] Batch [49]	Speed: 2956.483858 samples/sec	accuracy=0.771742	lr=0.261600
Epoch[51] Batch [99]	Speed: 2943.852429 samples/sec	accuracy=0.772023	lr=0.261392
Epoch[51] Batch [149]	Speed: 2910.674571 samples/sec	accuracy=0.772026	lr=0.261184
Epoch[51] Batch [199]	Speed: 2962.901075 samples/sec	accuracy=0.772348	lr=0.260976
Epoch[51] Batch [249]	Speed: 2951.370632 samples/sec	accuracy=0.772362	lr=0.260768
Epoch[51] Batch [299]	Speed: 2939.438963 samples/sec	accuracy=0.772472	lr=0.260560
Epoch[51] Batch [349]	Speed: 2830.277811 samples/sec	accuracy=0.772337	lr=0.260352
Epoch[51] Batch [399]	Speed: 2898.608622 samples/sec	accuracy=0.772220	lr=0.260144
Epoch[51] Batch [449]	Speed: 3044.315446 samples/sec	accuracy=0.772404	lr=0.259935
Epoch[51] Batch [499]	Speed: 2799.871264 samples/sec	accuracy=0.772501	lr=0.259727
Epoch[51] Batch [549]	Speed: 2934.375573 samples/sec	accuracy=0.772577	lr=0.259519
Epoch[51] Batch [599]	Speed: 2838.067600 samples/sec	accuracy=0.772794	lr=0.259310
Epoch[51] Batch [649]	Speed: 2972.060451 samples/sec	accuracy=0.772964	lr=0.259102
Epoch[51] Batch [699]	Speed: 2888.513231 samples/sec	accuracy=0.773061	lr=0.258893
Epoch[51] Batch [749]	Speed: 2853.201724 samples/sec	accuracy=0.773303	lr=0.258684
Epoch[51] Batch [799]	Speed: 3180.549174 samples/sec	accuracy=0.773445	lr=0.258475
Epoch[51] Batch [849]	Speed: 3005.906863 samples/sec	accuracy=0.773416	lr=0.258267
Epoch[51] Batch [899]	Speed: 2898.899777 samples/sec	accuracy=0.773522	lr=0.258058
Epoch[51] Batch [949]	Speed: 2995.746451 samples/sec	accuracy=0.773665	lr=0.257849
Epoch[51] Batch [999]	Speed: 2838.808150 samples/sec	accuracy=0.773828	lr=0.257640
Epoch[51] Batch [1049]	Speed: 2840.131213 samples/sec	accuracy=0.773887	lr=0.257430
Epoch[51] Batch [1099]	Speed: 2956.841229 samples/sec	accuracy=0.774003	lr=0.257221
Epoch[51] Batch [1149]	Speed: 2967.166103 samples/sec	accuracy=0.774039	lr=0.257012
Epoch[51] Batch [1199]	Speed: 2890.880159 samples/sec	accuracy=0.774133	lr=0.256803
Epoch[51] Batch [1249]	Speed: 3015.645115 samples/sec	accuracy=0.774257	lr=0.256593
[Epoch 51] training: accuracy=0.774260
[Epoch 51] speed: 2928 samples/sec	time cost: 447.302624
[Epoch 51] validation: err-top1=0.331673 err-top5=0.120516
Epoch[52] Batch [49]	Speed: 2908.938503 samples/sec	accuracy=0.773338	lr=0.256379
Epoch[52] Batch [99]	Speed: 2847.704138 samples/sec	accuracy=0.774013	lr=0.256170
Epoch[52] Batch [149]	Speed: 2890.882805 samples/sec	accuracy=0.774026	lr=0.255960
Epoch[52] Batch [199]	Speed: 2810.889601 samples/sec	accuracy=0.774004	lr=0.255751
Epoch[52] Batch [249]	Speed: 2799.768544 samples/sec	accuracy=0.773948	lr=0.255541
Epoch[52] Batch [299]	Speed: 2858.863444 samples/sec	accuracy=0.774052	lr=0.255331
Epoch[52] Batch [349]	Speed: 2831.679854 samples/sec	accuracy=0.774029	lr=0.255121
Epoch[52] Batch [399]	Speed: 2843.747722 samples/sec	accuracy=0.774188	lr=0.254911
Epoch[52] Batch [449]	Speed: 2878.558565 samples/sec	accuracy=0.774223	lr=0.254701
Epoch[52] Batch [499]	Speed: 2864.931898 samples/sec	accuracy=0.774136	lr=0.254491
Epoch[52] Batch [549]	Speed: 2936.356813 samples/sec	accuracy=0.774368	lr=0.254281
Epoch[52] Batch [599]	Speed: 2819.185933 samples/sec	accuracy=0.774419	lr=0.254071
Epoch[52] Batch [649]	Speed: 2928.487959 samples/sec	accuracy=0.774574	lr=0.253861
Epoch[52] Batch [699]	Speed: 2851.735538 samples/sec	accuracy=0.774659	lr=0.253650
Epoch[52] Batch [749]	Speed: 2923.643686 samples/sec	accuracy=0.774727	lr=0.253440
Epoch[52] Batch [799]	Speed: 2984.349753 samples/sec	accuracy=0.774722	lr=0.253229
Epoch[52] Batch [849]	Speed: 2888.099356 samples/sec	accuracy=0.774840	lr=0.253019
Epoch[52] Batch [899]	Speed: 2900.426236 samples/sec	accuracy=0.774911	lr=0.252808
Epoch[52] Batch [949]	Speed: 2913.032924 samples/sec	accuracy=0.775043	lr=0.252598
Epoch[52] Batch [999]	Speed: 2826.852149 samples/sec	accuracy=0.775198	lr=0.252387
Epoch[52] Batch [1049]	Speed: 2848.578531 samples/sec	accuracy=0.775287	lr=0.252176
Epoch[52] Batch [1099]	Speed: 2875.377920 samples/sec	accuracy=0.775328	lr=0.251965
Epoch[52] Batch [1149]	Speed: 2943.966195 samples/sec	accuracy=0.775438	lr=0.251754
Epoch[52] Batch [1199]	Speed: 2888.080751 samples/sec	accuracy=0.775568	lr=0.251543
Epoch[52] Batch [1249]	Speed: 3018.908659 samples/sec	accuracy=0.775679	lr=0.251332
[Epoch 52] training: accuracy=0.775679
[Epoch 52] speed: 2880 samples/sec	time cost: 454.311697
[Epoch 52] validation: err-top1=0.324219 err-top5=0.115133
Epoch[53] Batch [49]	Speed: 3188.294566 samples/sec	accuracy=0.776007	lr=0.251117
Epoch[53] Batch [99]	Speed: 2918.387138 samples/sec	accuracy=0.776670	lr=0.250906
Epoch[53] Batch [149]	Speed: 2957.208744 samples/sec	accuracy=0.776355	lr=0.250695
Epoch[53] Batch [199]	Speed: 3059.847711 samples/sec	accuracy=0.775979	lr=0.250483
Epoch[53] Batch [249]	Speed: 2923.582589 samples/sec	accuracy=0.775857	lr=0.250272
Epoch[53] Batch [299]	Speed: 2972.165836 samples/sec	accuracy=0.775872	lr=0.250061
Epoch[53] Batch [349]	Speed: 2971.692648 samples/sec	accuracy=0.775882	lr=0.249849
Epoch[53] Batch [399]	Speed: 2838.883543 samples/sec	accuracy=0.775734	lr=0.249638
Epoch[53] Batch [449]	Speed: 2884.338396 samples/sec	accuracy=0.775702	lr=0.249426
Epoch[53] Batch [499]	Speed: 2947.055532 samples/sec	accuracy=0.775611	lr=0.249215
Epoch[53] Batch [549]	Speed: 2998.697987 samples/sec	accuracy=0.775696	lr=0.249003
Epoch[53] Batch [599]	Speed: 2937.385299 samples/sec	accuracy=0.775784	lr=0.248791
Epoch[53] Batch [649]	Speed: 2870.741392 samples/sec	accuracy=0.775941	lr=0.248579
Epoch[53] Batch [699]	Speed: 2839.109297 samples/sec	accuracy=0.776062	lr=0.248367
Epoch[53] Batch [749]	Speed: 2942.857035 samples/sec	accuracy=0.776303	lr=0.248156
Epoch[53] Batch [799]	Speed: 2897.166740 samples/sec	accuracy=0.776357	lr=0.247944
Epoch[53] Batch [849]	Speed: 2910.196267 samples/sec	accuracy=0.776420	lr=0.247732
Epoch[53] Batch [899]	Speed: 2969.036260 samples/sec	accuracy=0.776467	lr=0.247519
Epoch[53] Batch [949]	Speed: 2919.794131 samples/sec	accuracy=0.776510	lr=0.247307
Epoch[53] Batch [999]	Speed: 2890.326760 samples/sec	accuracy=0.776618	lr=0.247095
Epoch[53] Batch [1049]	Speed: 2961.105066 samples/sec	accuracy=0.776638	lr=0.246883
Epoch[53] Batch [1099]	Speed: 2931.915540 samples/sec	accuracy=0.776754	lr=0.246670
Epoch[53] Batch [1149]	Speed: 2906.492240 samples/sec	accuracy=0.776845	lr=0.246458
Epoch[53] Batch [1199]	Speed: 2876.920554 samples/sec	accuracy=0.776940	lr=0.246246
Epoch[53] Batch [1249]	Speed: 2975.201727 samples/sec	accuracy=0.777000	lr=0.246033
[Epoch 53] training: accuracy=0.777009
[Epoch 53] speed: 2935 samples/sec	time cost: 446.156651
[Epoch 53] validation: err-top1=0.326690 err-top5=0.118423
Epoch[54] Batch [49]	Speed: 2936.054914 samples/sec	accuracy=0.777347	lr=0.245816
Epoch[54] Batch [99]	Speed: 2875.948217 samples/sec	accuracy=0.776912	lr=0.245604
Epoch[54] Batch [149]	Speed: 2909.512928 samples/sec	accuracy=0.776952	lr=0.245391
Epoch[54] Batch [199]	Speed: 2873.986321 samples/sec	accuracy=0.776912	lr=0.245178
Epoch[54] Batch [249]	Speed: 2857.135052 samples/sec	accuracy=0.777053	lr=0.244966
Epoch[54] Batch [299]	Speed: 2993.719426 samples/sec	accuracy=0.777024	lr=0.244753
Epoch[54] Batch [349]	Speed: 2921.809587 samples/sec	accuracy=0.777042	lr=0.244540
Epoch[54] Batch [399]	Speed: 3018.066598 samples/sec	accuracy=0.777087	lr=0.244327
Epoch[54] Batch [449]	Speed: 2931.418506 samples/sec	accuracy=0.777083	lr=0.244114
Epoch[54] Batch [499]	Speed: 2887.921318 samples/sec	accuracy=0.776984	lr=0.243901
Epoch[54] Batch [549]	Speed: 2891.855260 samples/sec	accuracy=0.777025	lr=0.243688
Epoch[54] Batch [599]	Speed: 2823.640474 samples/sec	accuracy=0.777163	lr=0.243475
Epoch[54] Batch [649]	Speed: 2956.339453 samples/sec	accuracy=0.777220	lr=0.243262
Epoch[54] Batch [699]	Speed: 2812.408143 samples/sec	accuracy=0.777408	lr=0.243049
Epoch[54] Batch [749]	Speed: 2955.826211 samples/sec	accuracy=0.777423	lr=0.242835
Epoch[54] Batch [799]	Speed: 3012.267221 samples/sec	accuracy=0.777430	lr=0.242622
Epoch[54] Batch [849]	Speed: 2988.117468 samples/sec	accuracy=0.777557	lr=0.242409
Epoch[54] Batch [899]	Speed: 2904.632130 samples/sec	accuracy=0.777762	lr=0.242195
Epoch[54] Batch [949]	Speed: 2868.262994 samples/sec	accuracy=0.777838	lr=0.241982
Epoch[54] Batch [999]	Speed: 2867.942799 samples/sec	accuracy=0.777914	lr=0.241768
Epoch[54] Batch [1049]	Speed: 2958.643155 samples/sec	accuracy=0.778035	lr=0.241555
Epoch[54] Batch [1099]	Speed: 2869.877619 samples/sec	accuracy=0.778221	lr=0.241341
Epoch[54] Batch [1149]	Speed: 2983.865215 samples/sec	accuracy=0.778424	lr=0.241127
Epoch[54] Batch [1199]	Speed: 2867.116002 samples/sec	accuracy=0.778508	lr=0.240914
Epoch[54] Batch [1249]	Speed: 2997.330406 samples/sec	accuracy=0.778600	lr=0.240700
[Epoch 54] training: accuracy=0.778600
[Epoch 54] speed: 2915 samples/sec	time cost: 449.199250
[Epoch 54] validation: err-top1=0.314792 err-top5=0.107502
Epoch[55] Batch [49]	Speed: 2842.622470 samples/sec	accuracy=0.778781	lr=0.240482
Epoch[55] Batch [99]	Speed: 2890.076179 samples/sec	accuracy=0.778114	lr=0.240268
Epoch[55] Batch [149]	Speed: 2842.316363 samples/sec	accuracy=0.778019	lr=0.240054
Epoch[55] Batch [199]	Speed: 2858.047008 samples/sec	accuracy=0.778377	lr=0.239840
Epoch[55] Batch [249]	Speed: 2897.639440 samples/sec	accuracy=0.778346	lr=0.239626
Epoch[55] Batch [299]	Speed: 2818.399619 samples/sec	accuracy=0.778448	lr=0.239412
Epoch[55] Batch [349]	Speed: 3090.802330 samples/sec	accuracy=0.778731	lr=0.239198
Epoch[55] Batch [399]	Speed: 2875.530503 samples/sec	accuracy=0.778760	lr=0.238984
Epoch[55] Batch [449]	Speed: 2857.201044 samples/sec	accuracy=0.778845	lr=0.238769
Epoch[55] Batch [499]	Speed: 2929.227508 samples/sec	accuracy=0.778951	lr=0.238555
Epoch[55] Batch [549]	Speed: 2930.874799 samples/sec	accuracy=0.779069	lr=0.238341
Epoch[55] Batch [599]	Speed: 2902.843224 samples/sec	accuracy=0.779015	lr=0.238126
Epoch[55] Batch [649]	Speed: 2990.636525 samples/sec	accuracy=0.779077	lr=0.237912
Epoch[55] Batch [699]	Speed: 2846.450562 samples/sec	accuracy=0.779150	lr=0.237698
Epoch[55] Batch [749]	Speed: 2858.622399 samples/sec	accuracy=0.779180	lr=0.237483
Epoch[55] Batch [799]	Speed: 2888.579709 samples/sec	accuracy=0.779285	lr=0.237269
Epoch[55] Batch [849]	Speed: 2796.561267 samples/sec	accuracy=0.779312	lr=0.237054
Epoch[55] Batch [899]	Speed: 3110.143103 samples/sec	accuracy=0.779497	lr=0.236839
Epoch[55] Batch [949]	Speed: 2874.963605 samples/sec	accuracy=0.779660	lr=0.236625
Epoch[55] Batch [999]	Speed: 2936.383071 samples/sec	accuracy=0.779728	lr=0.236410
Epoch[55] Batch [1049]	Speed: 2996.484742 samples/sec	accuracy=0.779788	lr=0.236195
Epoch[55] Batch [1099]	Speed: 2870.476009 samples/sec	accuracy=0.779793	lr=0.235981
Epoch[55] Batch [1149]	Speed: 3097.902065 samples/sec	accuracy=0.779891	lr=0.235766
Epoch[55] Batch [1199]	Speed: 2918.262689 samples/sec	accuracy=0.779989	lr=0.235551
Epoch[55] Batch [1249]	Speed: 3021.177268 samples/sec	accuracy=0.779977	lr=0.235336
[Epoch 55] training: accuracy=0.779974
[Epoch 55] speed: 2912 samples/sec	time cost: 449.837943
[Epoch 55] validation: err-top1=0.316924 err-top5=0.111587
Epoch[56] Batch [49]	Speed: 3001.494117 samples/sec	accuracy=0.778827	lr=0.235117
Epoch[56] Batch [99]	Speed: 2943.414960 samples/sec	accuracy=0.778486	lr=0.234902
Epoch[56] Batch [149]	Speed: 2977.728402 samples/sec	accuracy=0.779102	lr=0.234687
Epoch[56] Batch [199]	Speed: 2834.895895 samples/sec	accuracy=0.779226	lr=0.234472
Epoch[56] Batch [249]	Speed: 2972.919341 samples/sec	accuracy=0.779393	lr=0.234256
Epoch[56] Batch [299]	Speed: 2986.987094 samples/sec	accuracy=0.779500	lr=0.234041
Epoch[56] Batch [349]	Speed: 2942.795616 samples/sec	accuracy=0.779680	lr=0.233826
Epoch[56] Batch [399]	Speed: 2795.991001 samples/sec	accuracy=0.779638	lr=0.233611
Epoch[56] Batch [449]	Speed: 3005.955418 samples/sec	accuracy=0.779703	lr=0.233396
Epoch[56] Batch [499]	Speed: 2855.807079 samples/sec	accuracy=0.779541	lr=0.233180
Epoch[56] Batch [549]	Speed: 2878.138819 samples/sec	accuracy=0.779778	lr=0.232965
Epoch[56] Batch [599]	Speed: 2840.607088 samples/sec	accuracy=0.780077	lr=0.232749
Epoch[56] Batch [649]	Speed: 2855.928612 samples/sec	accuracy=0.780227	lr=0.232534
Epoch[56] Batch [699]	Speed: 2891.202382 samples/sec	accuracy=0.780344	lr=0.232319
Epoch[56] Batch [749]	Speed: 2991.839770 samples/sec	accuracy=0.780441	lr=0.232103
Epoch[56] Batch [799]	Speed: 2843.384147 samples/sec	accuracy=0.780477	lr=0.231887
Epoch[56] Batch [849]	Speed: 2915.685167 samples/sec	accuracy=0.780707	lr=0.231672
Epoch[56] Batch [899]	Speed: 2855.371466 samples/sec	accuracy=0.780784	lr=0.231456
Epoch[56] Batch [949]	Speed: 2830.246030 samples/sec	accuracy=0.780815	lr=0.231241
Epoch[56] Batch [999]	Speed: 2862.364622 samples/sec	accuracy=0.780831	lr=0.231025
Epoch[56] Batch [1049]	Speed: 2966.875338 samples/sec	accuracy=0.780898	lr=0.230809
Epoch[56] Batch [1099]	Speed: 2903.448022 samples/sec	accuracy=0.780967	lr=0.230593
Epoch[56] Batch [1149]	Speed: 2922.767798 samples/sec	accuracy=0.781073	lr=0.230377
Epoch[56] Batch [1199]	Speed: 2865.703208 samples/sec	accuracy=0.781185	lr=0.230162
Epoch[56] Batch [1249]	Speed: 3104.151154 samples/sec	accuracy=0.781217	lr=0.229946
[Epoch 56] training: accuracy=0.781220
[Epoch 56] speed: 2910 samples/sec	time cost: 450.266384
[Epoch 56] validation: err-top1=0.330935 err-top5=0.118941
Epoch[57] Batch [49]	Speed: 3039.153207 samples/sec	accuracy=0.783338	lr=0.229725
Epoch[57] Batch [99]	Speed: 2966.516932 samples/sec	accuracy=0.782032	lr=0.229510
Epoch[57] Batch [149]	Speed: 3045.465103 samples/sec	accuracy=0.781461	lr=0.229294
Epoch[57] Batch [199]	Speed: 2985.560978 samples/sec	accuracy=0.781226	lr=0.229077
Epoch[57] Batch [249]	Speed: 2944.385378 samples/sec	accuracy=0.781176	lr=0.228861
Epoch[57] Batch [299]	Speed: 2846.388536 samples/sec	accuracy=0.781263	lr=0.228645
Epoch[57] Batch [349]	Speed: 2913.920067 samples/sec	accuracy=0.781200	lr=0.228429
Epoch[57] Batch [399]	Speed: 2851.449501 samples/sec	accuracy=0.781291	lr=0.228213
Epoch[57] Batch [449]	Speed: 2973.738289 samples/sec	accuracy=0.781462	lr=0.227997
Epoch[57] Batch [499]	Speed: 3032.447931 samples/sec	accuracy=0.781410	lr=0.227781
Epoch[57] Batch [549]	Speed: 2987.773240 samples/sec	accuracy=0.781367	lr=0.227564
Epoch[57] Batch [599]	Speed: 2954.936914 samples/sec	accuracy=0.781410	lr=0.227348
Epoch[57] Batch [649]	Speed: 2963.679705 samples/sec	accuracy=0.781612	lr=0.227132
Epoch[57] Batch [699]	Speed: 2956.828771 samples/sec	accuracy=0.781672	lr=0.226915
Epoch[57] Batch [749]	Speed: 2966.699833 samples/sec	accuracy=0.781782	lr=0.226699
Epoch[57] Batch [799]	Speed: 2919.880398 samples/sec	accuracy=0.781811	lr=0.226482
Epoch[57] Batch [849]	Speed: 3108.194263 samples/sec	accuracy=0.781841	lr=0.226266
Epoch[57] Batch [899]	Speed: 2900.195051 samples/sec	accuracy=0.781910	lr=0.226049
Epoch[57] Batch [949]	Speed: 2940.119688 samples/sec	accuracy=0.782066	lr=0.225833
Epoch[57] Batch [999]	Speed: 2837.102491 samples/sec	accuracy=0.782084	lr=0.225616
Epoch[57] Batch [1049]	Speed: 2829.956158 samples/sec	accuracy=0.782254	lr=0.225400
Epoch[57] Batch [1099]	Speed: 2889.459482 samples/sec	accuracy=0.782367	lr=0.225183
Epoch[57] Batch [1149]	Speed: 2832.315389 samples/sec	accuracy=0.782465	lr=0.224967
Epoch[57] Batch [1199]	Speed: 2851.970650 samples/sec	accuracy=0.782541	lr=0.224750
Epoch[57] Batch [1249]	Speed: 3022.576997 samples/sec	accuracy=0.782558	lr=0.224533
[Epoch 57] training: accuracy=0.782562
[Epoch 57] speed: 2938 samples/sec	time cost: 445.738102
[Epoch 57] validation: err-top1=0.322246 err-top5=0.112205
Epoch[58] Batch [49]	Speed: 3037.028852 samples/sec	accuracy=0.783550	lr=0.224312
Epoch[58] Batch [99]	Speed: 2879.486837 samples/sec	accuracy=0.783166	lr=0.224095
Epoch[58] Batch [149]	Speed: 2837.218202 samples/sec	accuracy=0.782368	lr=0.223879
Epoch[58] Batch [199]	Speed: 2852.237699 samples/sec	accuracy=0.782221	lr=0.223662
Epoch[58] Batch [249]	Speed: 2915.083372 samples/sec	accuracy=0.782201	lr=0.223445
Epoch[58] Batch [299]	Speed: 2913.667870 samples/sec	accuracy=0.782451	lr=0.223228
Epoch[58] Batch [349]	Speed: 2936.490037 samples/sec	accuracy=0.782314	lr=0.223011
Epoch[58] Batch [399]	Speed: 2923.519306 samples/sec	accuracy=0.782377	lr=0.222794
Epoch[58] Batch [449]	Speed: 2847.028124 samples/sec	accuracy=0.782420	lr=0.222577
Epoch[58] Batch [499]	Speed: 2931.251491 samples/sec	accuracy=0.782473	lr=0.222360
Epoch[58] Batch [549]	Speed: 2897.183664 samples/sec	accuracy=0.782620	lr=0.222143
Epoch[58] Batch [599]	Speed: 2952.974263 samples/sec	accuracy=0.782764	lr=0.221926
Epoch[58] Batch [649]	Speed: 2849.808908 samples/sec	accuracy=0.782938	lr=0.221709
Epoch[58] Batch [699]	Speed: 2875.773600 samples/sec	accuracy=0.783047	lr=0.221492
Epoch[58] Batch [749]	Speed: 2942.886112 samples/sec	accuracy=0.783159	lr=0.221275
Epoch[58] Batch [799]	Speed: 3058.264168 samples/sec	accuracy=0.783264	lr=0.221058
Epoch[58] Batch [849]	Speed: 2990.948920 samples/sec	accuracy=0.783337	lr=0.220841
Epoch[58] Batch [899]	Speed: 2836.495342 samples/sec	accuracy=0.783403	lr=0.220623
Epoch[58] Batch [949]	Speed: 2821.953655 samples/sec	accuracy=0.783398	lr=0.220406
Epoch[58] Batch [999]	Speed: 2884.669082 samples/sec	accuracy=0.783464	lr=0.220189
Epoch[58] Batch [1049]	Speed: 3112.909152 samples/sec	accuracy=0.783573	lr=0.219972
Epoch[58] Batch [1099]	Speed: 2942.241796 samples/sec	accuracy=0.783642	lr=0.219754
Epoch[58] Batch [1149]	Speed: 2881.960389 samples/sec	accuracy=0.783809	lr=0.219537
Epoch[58] Batch [1199]	Speed: 2932.765999 samples/sec	accuracy=0.783878	lr=0.219320
Epoch[58] Batch [1249]	Speed: 3036.711997 samples/sec	accuracy=0.783977	lr=0.219102
[Epoch 58] training: accuracy=0.783974
[Epoch 58] speed: 2919 samples/sec	time cost: 448.552757
[Epoch 58] validation: err-top1=0.318197 err-top5=0.110494
Epoch[59] Batch [49]	Speed: 3059.480004 samples/sec	accuracy=0.783822	lr=0.218881
Epoch[59] Batch [99]	Speed: 3120.584817 samples/sec	accuracy=0.784169	lr=0.218663
Epoch[59] Batch [149]	Speed: 3162.308238 samples/sec	accuracy=0.783994	lr=0.218446
Epoch[59] Batch [199]	Speed: 2875.365523 samples/sec	accuracy=0.784098	lr=0.218228
Epoch[59] Batch [249]	Speed: 2830.201121 samples/sec	accuracy=0.784162	lr=0.218011
Epoch[59] Batch [299]	Speed: 2845.005853 samples/sec	accuracy=0.784048	lr=0.217793
Epoch[59] Batch [349]	Speed: 2775.285398 samples/sec	accuracy=0.784122	lr=0.217576
Epoch[59] Batch [399]	Speed: 2836.940729 samples/sec	accuracy=0.784160	lr=0.217358
Epoch[59] Batch [449]	Speed: 2760.079231 samples/sec	accuracy=0.784318	lr=0.217141
Epoch[59] Batch [499]	Speed: 2853.136523 samples/sec	accuracy=0.784199	lr=0.216923
Epoch[59] Batch [549]	Speed: 2891.689842 samples/sec	accuracy=0.784308	lr=0.216706
Epoch[59] Batch [599]	Speed: 2907.270232 samples/sec	accuracy=0.784235	lr=0.216488
Epoch[59] Batch [649]	Speed: 2965.134322 samples/sec	accuracy=0.784233	lr=0.216270
Epoch[59] Batch [699]	Speed: 3010.123143 samples/sec	accuracy=0.784310	lr=0.216053
Epoch[59] Batch [749]	Speed: 2990.443706 samples/sec	accuracy=0.784354	lr=0.215835
Epoch[59] Batch [799]	Speed: 2925.323633 samples/sec	accuracy=0.784517	lr=0.215617
Epoch[59] Batch [849]	Speed: 2932.816746 samples/sec	accuracy=0.784650	lr=0.215400
Epoch[59] Batch [899]	Speed: 2749.578198 samples/sec	accuracy=0.784758	lr=0.215182
Epoch[59] Batch [949]	Speed: 3058.537097 samples/sec	accuracy=0.784828	lr=0.214964
Epoch[59] Batch [999]	Speed: 2869.746190 samples/sec	accuracy=0.784998	lr=0.214746
Epoch[59] Batch [1049]	Speed: 2918.416408 samples/sec	accuracy=0.785114	lr=0.214529
Epoch[59] Batch [1099]	Speed: 2924.532923 samples/sec	accuracy=0.785156	lr=0.214311
Epoch[59] Batch [1149]	Speed: 2877.581727 samples/sec	accuracy=0.785255	lr=0.214093
Epoch[59] Batch [1199]	Speed: 2814.936730 samples/sec	accuracy=0.785384	lr=0.213875
Epoch[59] Batch [1249]	Speed: 2999.192465 samples/sec	accuracy=0.785461	lr=0.213657
[Epoch 59] training: accuracy=0.785463
[Epoch 59] speed: 2912 samples/sec	time cost: 449.986126
[Epoch 59] validation: err-top1=0.320392 err-top5=0.112663
Epoch[60] Batch [49]	Speed: 2909.922909 samples/sec	accuracy=0.783585	lr=0.213435
Epoch[60] Batch [99]	Speed: 2947.514959 samples/sec	accuracy=0.784353	lr=0.213217
Epoch[60] Batch [149]	Speed: 2838.843726 samples/sec	accuracy=0.784127	lr=0.212999
Epoch[60] Batch [199]	Speed: 3031.938832 samples/sec	accuracy=0.784223	lr=0.212781
Epoch[60] Batch [249]	Speed: 2901.283020 samples/sec	accuracy=0.784220	lr=0.212563
Epoch[60] Batch [299]	Speed: 2927.789337 samples/sec	accuracy=0.784440	lr=0.212345
Epoch[60] Batch [349]	Speed: 2868.389268 samples/sec	accuracy=0.784592	lr=0.212127
Epoch[60] Batch [399]	Speed: 3017.893890 samples/sec	accuracy=0.784550	lr=0.211909
Epoch[60] Batch [449]	Speed: 2913.091605 samples/sec	accuracy=0.784885	lr=0.211691
Epoch[60] Batch [499]	Speed: 2966.125263 samples/sec	accuracy=0.785002	lr=0.211473
Epoch[60] Batch [549]	Speed: 3075.763188 samples/sec	accuracy=0.785056	lr=0.211255
Epoch[60] Batch [599]	Speed: 2881.456562 samples/sec	accuracy=0.785227	lr=0.211037
Epoch[60] Batch [649]	Speed: 2784.913360 samples/sec	accuracy=0.785343	lr=0.210819
Epoch[60] Batch [699]	Speed: 2920.619061 samples/sec	accuracy=0.785467	lr=0.210601
Epoch[60] Batch [749]	Speed: 2859.649997 samples/sec	accuracy=0.785584	lr=0.210383
Epoch[60] Batch [799]	Speed: 2887.919687 samples/sec	accuracy=0.785699	lr=0.210165
Epoch[60] Batch [849]	Speed: 2897.933216 samples/sec	accuracy=0.785804	lr=0.209947
Epoch[60] Batch [899]	Speed: 2816.814006 samples/sec	accuracy=0.785952	lr=0.209729
Epoch[60] Batch [949]	Speed: 2940.505726 samples/sec	accuracy=0.786123	lr=0.209511
Epoch[60] Batch [999]	Speed: 2828.023193 samples/sec	accuracy=0.786294	lr=0.209293
Epoch[60] Batch [1049]	Speed: 2839.548748 samples/sec	accuracy=0.786343	lr=0.209075
Epoch[60] Batch [1099]	Speed: 2881.183241 samples/sec	accuracy=0.786452	lr=0.208856
Epoch[60] Batch [1149]	Speed: 2822.491084 samples/sec	accuracy=0.786541	lr=0.208638
Epoch[60] Batch [1199]	Speed: 2960.080672 samples/sec	accuracy=0.786677	lr=0.208420
Epoch[60] Batch [1249]	Speed: 3021.505131 samples/sec	accuracy=0.786749	lr=0.208202
[Epoch 60] training: accuracy=0.786751
[Epoch 60] speed: 2905 samples/sec	time cost: 450.440195
[Epoch 60] validation: err-top1=0.314134 err-top5=0.111488
Epoch[61] Batch [49]	Speed: 3007.161970 samples/sec	accuracy=0.787818	lr=0.207979
Epoch[61] Batch [99]	Speed: 2842.333217 samples/sec	accuracy=0.786216	lr=0.207761
Epoch[61] Batch [149]	Speed: 2900.471403 samples/sec	accuracy=0.785901	lr=0.207543
Epoch[61] Batch [199]	Speed: 3010.337118 samples/sec	accuracy=0.785972	lr=0.207325
Epoch[61] Batch [249]	Speed: 2887.710723 samples/sec	accuracy=0.786037	lr=0.207106
Epoch[61] Batch [299]	Speed: 2902.807164 samples/sec	accuracy=0.786193	lr=0.206888
Epoch[61] Batch [349]	Speed: 2848.223088 samples/sec	accuracy=0.786330	lr=0.206670
Epoch[61] Batch [399]	Speed: 2973.080559 samples/sec	accuracy=0.786574	lr=0.206452
Epoch[61] Batch [449]	Speed: 2935.756770 samples/sec	accuracy=0.786565	lr=0.206233
Epoch[61] Batch [499]	Speed: 2865.307695 samples/sec	accuracy=0.786688	lr=0.206015
Epoch[61] Batch [549]	Speed: 2990.874814 samples/sec	accuracy=0.786737	lr=0.205797
Epoch[61] Batch [599]	Speed: 3014.051986 samples/sec	accuracy=0.786889	lr=0.205579
Epoch[61] Batch [649]	Speed: 2833.815741 samples/sec	accuracy=0.787000	lr=0.205360
Epoch[61] Batch [699]	Speed: 2909.502206 samples/sec	accuracy=0.787104	lr=0.205142
Epoch[61] Batch [749]	Speed: 3003.944728 samples/sec	accuracy=0.787167	lr=0.204924
Epoch[61] Batch [799]	Speed: 2949.395579 samples/sec	accuracy=0.787231	lr=0.204705
Epoch[61] Batch [849]	Speed: 2992.330404 samples/sec	accuracy=0.787394	lr=0.204487
Epoch[61] Batch [899]	Speed: 2912.358561 samples/sec	accuracy=0.787466	lr=0.204269
Epoch[61] Batch [949]	Speed: 2811.171422 samples/sec	accuracy=0.787585	lr=0.204051
Epoch[61] Batch [999]	Speed: 2892.755072 samples/sec	accuracy=0.787598	lr=0.203832
Epoch[61] Batch [1049]	Speed: 2891.773483 samples/sec	accuracy=0.787660	lr=0.203614
Epoch[61] Batch [1099]	Speed: 2934.593351 samples/sec	accuracy=0.787724	lr=0.203396
Epoch[61] Batch [1149]	Speed: 2855.833322 samples/sec	accuracy=0.787770	lr=0.203177
Epoch[61] Batch [1199]	Speed: 3065.462210 samples/sec	accuracy=0.787840	lr=0.202959
Epoch[61] Batch [1249]	Speed: 3002.589151 samples/sec	accuracy=0.787941	lr=0.202740
[Epoch 61] training: accuracy=0.787943
[Epoch 61] speed: 2925 samples/sec	time cost: 447.644587
[Epoch 61] validation: err-top1=0.310328 err-top5=0.106346
Epoch[62] Batch [49]	Speed: 3093.807000 samples/sec	accuracy=0.788180	lr=0.202518
Epoch[62] Batch [99]	Speed: 2895.877800 samples/sec	accuracy=0.787993	lr=0.202299
Epoch[62] Batch [149]	Speed: 2891.984477 samples/sec	accuracy=0.787631	lr=0.202081
Epoch[62] Batch [199]	Speed: 2912.293788 samples/sec	accuracy=0.787623	lr=0.201863
Epoch[62] Batch [249]	Speed: 2882.655959 samples/sec	accuracy=0.787425	lr=0.201644
Epoch[62] Batch [299]	Speed: 2863.434429 samples/sec	accuracy=0.787510	lr=0.201426
Epoch[62] Batch [349]	Speed: 2900.478298 samples/sec	accuracy=0.787694	lr=0.201208
Epoch[62] Batch [399]	Speed: 2857.216060 samples/sec	accuracy=0.787698	lr=0.200989
Epoch[62] Batch [449]	Speed: 2956.328505 samples/sec	accuracy=0.787911	lr=0.200771
Epoch[62] Batch [499]	Speed: 2960.498948 samples/sec	accuracy=0.787925	lr=0.200552
Epoch[62] Batch [549]	Speed: 2876.352761 samples/sec	accuracy=0.788023	lr=0.200334
Epoch[62] Batch [599]	Speed: 2883.719924 samples/sec	accuracy=0.788059	lr=0.200116
Epoch[62] Batch [649]	Speed: 2921.414969 samples/sec	accuracy=0.788180	lr=0.199897
Epoch[62] Batch [699]	Speed: 2834.590477 samples/sec	accuracy=0.788274	lr=0.199679
Epoch[62] Batch [749]	Speed: 3046.130797 samples/sec	accuracy=0.788314	lr=0.199461
Epoch[62] Batch [799]	Speed: 2840.392141 samples/sec	accuracy=0.788319	lr=0.199242
Epoch[62] Batch [849]	Speed: 3019.429609 samples/sec	accuracy=0.788368	lr=0.199024
Epoch[62] Batch [899]	Speed: 2881.319238 samples/sec	accuracy=0.788492	lr=0.198806
Epoch[62] Batch [949]	Speed: 2885.627318 samples/sec	accuracy=0.788619	lr=0.198587
Epoch[62] Batch [999]	Speed: 2837.866088 samples/sec	accuracy=0.788730	lr=0.198369
Epoch[62] Batch [1049]	Speed: 2933.991303 samples/sec	accuracy=0.788826	lr=0.198150
Epoch[62] Batch [1099]	Speed: 3099.557309 samples/sec	accuracy=0.788909	lr=0.197932
Epoch[62] Batch [1149]	Speed: 3060.859658 samples/sec	accuracy=0.789078	lr=0.197714
Epoch[62] Batch [1199]	Speed: 3040.780007 samples/sec	accuracy=0.789148	lr=0.197495
Epoch[62] Batch [1249]	Speed: 3011.227312 samples/sec	accuracy=0.789268	lr=0.197277
[Epoch 62] training: accuracy=0.789271
[Epoch 62] speed: 2931 samples/sec	time cost: 446.556906
[Epoch 62] validation: err-top1=0.306581 err-top5=0.103097
Epoch[63] Batch [49]	Speed: 2881.331764 samples/sec	accuracy=0.790124	lr=0.197054
Epoch[63] Batch [99]	Speed: 2928.722837 samples/sec	accuracy=0.790272	lr=0.196836
Epoch[63] Batch [149]	Speed: 2907.073294 samples/sec	accuracy=0.790448	lr=0.196618
Epoch[63] Batch [199]	Speed: 2784.923798 samples/sec	accuracy=0.789860	lr=0.196399
Epoch[63] Batch [249]	Speed: 2924.809750 samples/sec	accuracy=0.789676	lr=0.196181
Epoch[63] Batch [299]	Speed: 2974.005523 samples/sec	accuracy=0.789766	lr=0.195963
Epoch[63] Batch [349]	Speed: 2824.980017 samples/sec	accuracy=0.789595	lr=0.195744
Epoch[63] Batch [399]	Speed: 2821.114429 samples/sec	accuracy=0.789558	lr=0.195526
Epoch[63] Batch [449]	Speed: 2883.423990 samples/sec	accuracy=0.789424	lr=0.195308
Epoch[63] Batch [499]	Speed: 2874.819972 samples/sec	accuracy=0.789466	lr=0.195089
Epoch[63] Batch [549]	Speed: 2903.048419 samples/sec	accuracy=0.789648	lr=0.194871
Epoch[63] Batch [599]	Speed: 2941.353885 samples/sec	accuracy=0.789668	lr=0.194653
Epoch[63] Batch [649]	Speed: 2865.061434 samples/sec	accuracy=0.789772	lr=0.194434
Epoch[63] Batch [699]	Speed: 2966.074791 samples/sec	accuracy=0.789854	lr=0.194216
Epoch[63] Batch [749]	Speed: 2896.486731 samples/sec	accuracy=0.790050	lr=0.193998
Epoch[63] Batch [799]	Speed: 2848.401025 samples/sec	accuracy=0.790054	lr=0.193780
Epoch[63] Batch [849]	Speed: 2888.404060 samples/sec	accuracy=0.790150	lr=0.193561
Epoch[63] Batch [899]	Speed: 2983.388793 samples/sec	accuracy=0.790190	lr=0.193343
Epoch[63] Batch [949]	Speed: 2949.852656 samples/sec	accuracy=0.790330	lr=0.193125
Epoch[63] Batch [999]	Speed: 2871.508270 samples/sec	accuracy=0.790481	lr=0.192907
Epoch[63] Batch [1049]	Speed: 2959.551081 samples/sec	accuracy=0.790538	lr=0.192688
Epoch[63] Batch [1099]	Speed: 2842.397135 samples/sec	accuracy=0.790598	lr=0.192470
Epoch[63] Batch [1149]	Speed: 2933.013782 samples/sec	accuracy=0.790687	lr=0.192252
Epoch[63] Batch [1199]	Speed: 2934.493260 samples/sec	accuracy=0.790778	lr=0.192034
Epoch[63] Batch [1249]	Speed: 3085.078313 samples/sec	accuracy=0.790899	lr=0.191816
[Epoch 63] training: accuracy=0.790905
[Epoch 63] speed: 2903 samples/sec	time cost: 450.524205
[Epoch 63] validation: err-top1=0.302226 err-top5=0.101705
Epoch[64] Batch [49]	Speed: 2907.807814 samples/sec	accuracy=0.791373	lr=0.191593
Epoch[64] Batch [99]	Speed: 2910.189839 samples/sec	accuracy=0.790372	lr=0.191375
Epoch[64] Batch [149]	Speed: 2788.294837 samples/sec	accuracy=0.790421	lr=0.191157
Epoch[64] Batch [199]	Speed: 3009.556052 samples/sec	accuracy=0.790445	lr=0.190939
Epoch[64] Batch [249]	Speed: 3005.310530 samples/sec	accuracy=0.790574	lr=0.190720
Epoch[64] Batch [299]	Speed: 2964.319901 samples/sec	accuracy=0.790520	lr=0.190502
Epoch[64] Batch [349]	Speed: 3052.426616 samples/sec	accuracy=0.790544	lr=0.190284
Epoch[64] Batch [399]	Speed: 2991.674928 samples/sec	accuracy=0.790748	lr=0.190066
Epoch[64] Batch [449]	Speed: 2973.787169 samples/sec	accuracy=0.790860	lr=0.189848
Epoch[64] Batch [499]	Speed: 2952.573090 samples/sec	accuracy=0.791138	lr=0.189630
Epoch[64] Batch [549]	Speed: 2865.464373 samples/sec	accuracy=0.791095	lr=0.189412
Epoch[64] Batch [599]	Speed: 2838.947607 samples/sec	accuracy=0.791287	lr=0.189194
Epoch[64] Batch [649]	Speed: 2966.893210 samples/sec	accuracy=0.791409	lr=0.188976
Epoch[64] Batch [699]	Speed: 2990.305916 samples/sec	accuracy=0.791469	lr=0.188758
Epoch[64] Batch [749]	Speed: 2880.346205 samples/sec	accuracy=0.791566	lr=0.188540
Epoch[64] Batch [799]	Speed: 2881.925967 samples/sec	accuracy=0.791666	lr=0.188322
Epoch[64] Batch [849]	Speed: 2880.012106 samples/sec	accuracy=0.791684	lr=0.188104
Epoch[64] Batch [899]	Speed: 2916.682583 samples/sec	accuracy=0.791671	lr=0.187886
Epoch[64] Batch [949]	Speed: 2939.815486 samples/sec	accuracy=0.791782	lr=0.187668
Epoch[64] Batch [999]	Speed: 3023.401784 samples/sec	accuracy=0.791824	lr=0.187450
Epoch[64] Batch [1049]	Speed: 2974.719990 samples/sec	accuracy=0.791875	lr=0.187232
Epoch[64] Batch [1099]	Speed: 2858.724726 samples/sec	accuracy=0.791915	lr=0.187014
Epoch[64] Batch [1149]	Speed: 3145.091841 samples/sec	accuracy=0.791972	lr=0.186796
Epoch[64] Batch [1199]	Speed: 2877.404135 samples/sec	accuracy=0.792056	lr=0.186578
Epoch[64] Batch [1249]	Speed: 3060.619378 samples/sec	accuracy=0.792150	lr=0.186360
[Epoch 64] training: accuracy=0.792154
[Epoch 64] speed: 2942 samples/sec	time cost: 446.193668
[Epoch 64] validation: err-top1=0.299526 err-top5=0.100108
Epoch[65] Batch [49]	Speed: 2844.969481 samples/sec	accuracy=0.793322	lr=0.186138
Epoch[65] Batch [99]	Speed: 3017.564860 samples/sec	accuracy=0.792440	lr=0.185920
Epoch[65] Batch [149]	Speed: 2868.286631 samples/sec	accuracy=0.792129	lr=0.185702
Epoch[65] Batch [199]	Speed: 2901.430290 samples/sec	accuracy=0.792168	lr=0.185485
Epoch[65] Batch [249]	Speed: 2903.008939 samples/sec	accuracy=0.792195	lr=0.185267
Epoch[65] Batch [299]	Speed: 2919.512258 samples/sec	accuracy=0.792274	lr=0.185049
Epoch[65] Batch [349]	Speed: 2911.199205 samples/sec	accuracy=0.792241	lr=0.184831
Epoch[65] Batch [399]	Speed: 2964.906094 samples/sec	accuracy=0.792340	lr=0.184614
Epoch[65] Batch [449]	Speed: 2842.356428 samples/sec	accuracy=0.792357	lr=0.184396
Epoch[65] Batch [499]	Speed: 3030.616556 samples/sec	accuracy=0.792526	lr=0.184178
Epoch[65] Batch [549]	Speed: 2973.835186 samples/sec	accuracy=0.792578	lr=0.183960
Epoch[65] Batch [599]	Speed: 3016.058359 samples/sec	accuracy=0.792725	lr=0.183743
Epoch[65] Batch [649]	Speed: 2861.042495 samples/sec	accuracy=0.792872	lr=0.183525
Epoch[65] Batch [699]	Speed: 3172.233670 samples/sec	accuracy=0.792931	lr=0.183308
Epoch[65] Batch [749]	Speed: 2870.393787 samples/sec	accuracy=0.793011	lr=0.183090
Epoch[65] Batch [799]	Speed: 2819.540977 samples/sec	accuracy=0.792965	lr=0.182872
Epoch[65] Batch [849]	Speed: 2894.116663 samples/sec	accuracy=0.793045	lr=0.182655
Epoch[65] Batch [899]	Speed: 2905.106445 samples/sec	accuracy=0.793108	lr=0.182437
Epoch[65] Batch [949]	Speed: 2877.670338 samples/sec	accuracy=0.793129	lr=0.182220
Epoch[65] Batch [999]	Speed: 2938.416476 samples/sec	accuracy=0.793199	lr=0.182002
Epoch[65] Batch [1049]	Speed: 2864.146448 samples/sec	accuracy=0.793303	lr=0.181785
Epoch[65] Batch [1099]	Speed: 2733.670941 samples/sec	accuracy=0.793404	lr=0.181567
Epoch[65] Batch [1149]	Speed: 2934.132090 samples/sec	accuracy=0.793478	lr=0.181350
Epoch[65] Batch [1199]	Speed: 3028.938064 samples/sec	accuracy=0.793587	lr=0.181132
Epoch[65] Batch [1249]	Speed: 3070.197089 samples/sec	accuracy=0.793685	lr=0.180915
[Epoch 65] training: accuracy=0.793689
[Epoch 65] speed: 2921 samples/sec	time cost: 448.288592
[Epoch 65] validation: err-top1=0.301917 err-top5=0.101921
Epoch[66] Batch [49]	Speed: 2938.720751 samples/sec	accuracy=0.795506	lr=0.180693
Epoch[66] Batch [99]	Speed: 2843.332194 samples/sec	accuracy=0.794459	lr=0.180476
Epoch[66] Batch [149]	Speed: 2831.977624 samples/sec	accuracy=0.793517	lr=0.180259
Epoch[66] Batch [199]	Speed: 2910.036355 samples/sec	accuracy=0.793261	lr=0.180041
Epoch[66] Batch [249]	Speed: 2939.059962 samples/sec	accuracy=0.793203	lr=0.179824
Epoch[66] Batch [299]	Speed: 2822.791896 samples/sec	accuracy=0.793039	lr=0.179607
Epoch[66] Batch [349]	Speed: 2954.162708 samples/sec	accuracy=0.793373	lr=0.179390
Epoch[66] Batch [399]	Speed: 2923.099000 samples/sec	accuracy=0.793313	lr=0.179172
Epoch[66] Batch [449]	Speed: 2856.880616 samples/sec	accuracy=0.793546	lr=0.178955
Epoch[66] Batch [499]	Speed: 2888.961075 samples/sec	accuracy=0.793593	lr=0.178738
Epoch[66] Batch [549]	Speed: 2885.901521 samples/sec	accuracy=0.793784	lr=0.178521
Epoch[66] Batch [599]	Speed: 2898.799876 samples/sec	accuracy=0.793773	lr=0.178304
Epoch[66] Batch [649]	Speed: 2972.700117 samples/sec	accuracy=0.793930	lr=0.178087
Epoch[66] Batch [699]	Speed: 2947.013108 samples/sec	accuracy=0.793942	lr=0.177870
Epoch[66] Batch [749]	Speed: 2889.788933 samples/sec	accuracy=0.794013	lr=0.177653
Epoch[66] Batch [799]	Speed: 2843.498978 samples/sec	accuracy=0.794077	lr=0.177436
Epoch[66] Batch [849]	Speed: 2881.701783 samples/sec	accuracy=0.794194	lr=0.177219
Epoch[66] Batch [899]	Speed: 2926.495944 samples/sec	accuracy=0.794339	lr=0.177002
Epoch[66] Batch [949]	Speed: 2940.015879 samples/sec	accuracy=0.794383	lr=0.176785
Epoch[66] Batch [999]	Speed: 2865.947476 samples/sec	accuracy=0.794536	lr=0.176568
Epoch[66] Batch [1049]	Speed: 2974.347409 samples/sec	accuracy=0.794609	lr=0.176351
Epoch[66] Batch [1099]	Speed: 2875.549601 samples/sec	accuracy=0.794696	lr=0.176134
Epoch[66] Batch [1149]	Speed: 2797.965130 samples/sec	accuracy=0.794769	lr=0.175918
Epoch[66] Batch [1199]	Speed: 2998.482314 samples/sec	accuracy=0.794819	lr=0.175701
Epoch[66] Batch [1249]	Speed: 3061.290406 samples/sec	accuracy=0.794948	lr=0.175484
[Epoch 66] training: accuracy=0.794949
[Epoch 66] speed: 2903 samples/sec	time cost: 451.159513
[Epoch 66] validation: err-top1=0.296576 err-top5=0.098075
Epoch[67] Batch [49]	Speed: 2988.834198 samples/sec	accuracy=0.793901	lr=0.175263
Epoch[67] Batch [99]	Speed: 2873.694649 samples/sec	accuracy=0.794013	lr=0.175046
Epoch[67] Batch [149]	Speed: 2998.009834 samples/sec	accuracy=0.793866	lr=0.174830
Epoch[67] Batch [199]	Speed: 3077.854750 samples/sec	accuracy=0.794044	lr=0.174613
Epoch[67] Batch [249]	Speed: 3027.895538 samples/sec	accuracy=0.794217	lr=0.174397
Epoch[67] Batch [299]	Speed: 3040.300603 samples/sec	accuracy=0.794330	lr=0.174180
Epoch[67] Batch [349]	Speed: 3179.226758 samples/sec	accuracy=0.794248	lr=0.173964
Epoch[67] Batch [399]	Speed: 3108.514468 samples/sec	accuracy=0.794248	lr=0.173747
Epoch[67] Batch [449]	Speed: 3107.979510 samples/sec	accuracy=0.794210	lr=0.173531
Epoch[67] Batch [499]	Speed: 2990.796755 samples/sec	accuracy=0.794342	lr=0.173314
Epoch[67] Batch [549]	Speed: 3095.569422 samples/sec	accuracy=0.794564	lr=0.173098
Epoch[67] Batch [599]	Speed: 2944.716450 samples/sec	accuracy=0.794532	lr=0.172881
Epoch[67] Batch [649]	Speed: 2925.589411 samples/sec	accuracy=0.794654	lr=0.172665
Epoch[67] Batch [699]	Speed: 2997.500684 samples/sec	accuracy=0.794715	lr=0.172449
Epoch[67] Batch [749]	Speed: 2930.615619 samples/sec	accuracy=0.794834	lr=0.172232
Epoch[67] Batch [799]	Speed: 2920.874807 samples/sec	accuracy=0.794876	lr=0.172016
Epoch[67] Batch [849]	Speed: 2795.771761 samples/sec	accuracy=0.794960	lr=0.171800
Epoch[67] Batch [899]	Speed: 2869.552233 samples/sec	accuracy=0.795092	lr=0.171584
Epoch[67] Batch [949]	Speed: 2875.352895 samples/sec	accuracy=0.795256	lr=0.171368
Epoch[67] Batch [999]	Speed: 2952.815746 samples/sec	accuracy=0.795318	lr=0.171152
Epoch[67] Batch [1049]	Speed: 2899.090364 samples/sec	accuracy=0.795399	lr=0.170935
Epoch[67] Batch [1099]	Speed: 2808.769313 samples/sec	accuracy=0.795524	lr=0.170719
Epoch[67] Batch [1149]	Speed: 2940.862104 samples/sec	accuracy=0.795642	lr=0.170503
Epoch[67] Batch [1199]	Speed: 2996.789034 samples/sec	accuracy=0.795724	lr=0.170287
Epoch[67] Batch [1249]	Speed: 3087.388463 samples/sec	accuracy=0.795824	lr=0.170072
[Epoch 67] training: accuracy=0.795827
[Epoch 67] speed: 2972 samples/sec	time cost: 440.887735
[Epoch 67] validation: err-top1=0.297015 err-top5=0.099390
Epoch[68] Batch [49]	Speed: 3062.026431 samples/sec	accuracy=0.795467	lr=0.169851
Epoch[68] Batch [99]	Speed: 2812.846182 samples/sec	accuracy=0.795764	lr=0.169635
Epoch[68] Batch [149]	Speed: 2902.105562 samples/sec	accuracy=0.795458	lr=0.169420
Epoch[68] Batch [199]	Speed: 3038.285931 samples/sec	accuracy=0.795806	lr=0.169204
Epoch[68] Batch [249]	Speed: 2899.975574 samples/sec	accuracy=0.795731	lr=0.168988
Epoch[68] Batch [299]	Speed: 3018.566975 samples/sec	accuracy=0.795644	lr=0.168772
Epoch[68] Batch [349]	Speed: 2891.393749 samples/sec	accuracy=0.795661	lr=0.168557
Epoch[68] Batch [399]	Speed: 2834.197109 samples/sec	accuracy=0.795924	lr=0.168341
Epoch[68] Batch [449]	Speed: 2841.144467 samples/sec	accuracy=0.796085	lr=0.168126
Epoch[68] Batch [499]	Speed: 2926.324984 samples/sec	accuracy=0.796279	lr=0.167910
Epoch[68] Batch [549]	Speed: 2978.327882 samples/sec	accuracy=0.796388	lr=0.167694
Epoch[68] Batch [599]	Speed: 2892.532472 samples/sec	accuracy=0.796569	lr=0.167479
Epoch[68] Batch [649]	Speed: 3080.370185 samples/sec	accuracy=0.796561	lr=0.167263
Epoch[68] Batch [699]	Speed: 3029.280348 samples/sec	accuracy=0.796727	lr=0.167048
Epoch[68] Batch [749]	Speed: 3014.615484 samples/sec	accuracy=0.796830	lr=0.166833
Epoch[68] Batch [799]	Speed: 2984.903567 samples/sec	accuracy=0.796919	lr=0.166617
Epoch[68] Batch [849]	Speed: 2867.612069 samples/sec	accuracy=0.797004	lr=0.166402
Epoch[68] Batch [899]	Speed: 2901.795687 samples/sec	accuracy=0.797113	lr=0.166187
Epoch[68] Batch [949]	Speed: 2952.556244 samples/sec	accuracy=0.797128	lr=0.165972
Epoch[68] Batch [999]	Speed: 2980.798594 samples/sec	accuracy=0.797201	lr=0.165756
Epoch[68] Batch [1049]	Speed: 2917.949037 samples/sec	accuracy=0.797288	lr=0.165541
Epoch[68] Batch [1099]	Speed: 2985.517437 samples/sec	accuracy=0.797425	lr=0.165326
Epoch[68] Batch [1149]	Speed: 3030.824301 samples/sec	accuracy=0.797523	lr=0.165111
Epoch[68] Batch [1199]	Speed: 3011.868702 samples/sec	accuracy=0.797601	lr=0.164896
Epoch[68] Batch [1249]	Speed: 3023.999996 samples/sec	accuracy=0.797718	lr=0.164681
[Epoch 68] training: accuracy=0.797720
[Epoch 68] speed: 2951 samples/sec	time cost: 443.879934
[Epoch 68] validation: err-top1=0.305026 err-top5=0.102978
Epoch[69] Batch [49]	Speed: 3154.298512 samples/sec	accuracy=0.797092	lr=0.164462
Epoch[69] Batch [99]	Speed: 3030.863227 samples/sec	accuracy=0.797610	lr=0.164247
Epoch[69] Batch [149]	Speed: 2889.251267 samples/sec	accuracy=0.797752	lr=0.164032
Epoch[69] Batch [199]	Speed: 3040.967271 samples/sec	accuracy=0.797695	lr=0.163818
Epoch[69] Batch [249]	Speed: 3195.935745 samples/sec	accuracy=0.797533	lr=0.163603
Epoch[69] Batch [299]	Speed: 2908.504100 samples/sec	accuracy=0.797500	lr=0.163388
Epoch[69] Batch [349]	Speed: 2950.761032 samples/sec	accuracy=0.797306	lr=0.163173
Epoch[69] Batch [399]	Speed: 2924.639346 samples/sec	accuracy=0.797202	lr=0.162959
Epoch[69] Batch [449]	Speed: 2952.250275 samples/sec	accuracy=0.797263	lr=0.162744
Epoch[69] Batch [499]	Speed: 3041.580253 samples/sec	accuracy=0.797331	lr=0.162530
Epoch[69] Batch [549]	Speed: 2843.731303 samples/sec	accuracy=0.797247	lr=0.162315
Epoch[69] Batch [599]	Speed: 2975.746625 samples/sec	accuracy=0.797392	lr=0.162101
Epoch[69] Batch [649]	Speed: 2848.683125 samples/sec	accuracy=0.797478	lr=0.161886
Epoch[69] Batch [699]	Speed: 2888.703698 samples/sec	accuracy=0.797631	lr=0.161672
Epoch[69] Batch [749]	Speed: 2823.752787 samples/sec	accuracy=0.797698	lr=0.161458
Epoch[69] Batch [799]	Speed: 2868.018599 samples/sec	accuracy=0.797833	lr=0.161244
Epoch[69] Batch [849]	Speed: 3091.869980 samples/sec	accuracy=0.797964	lr=0.161029
Epoch[69] Batch [899]	Speed: 2895.096057 samples/sec	accuracy=0.798088	lr=0.160815
Epoch[69] Batch [949]	Speed: 3024.046965 samples/sec	accuracy=0.798197	lr=0.160601
Epoch[69] Batch [999]	Speed: 3090.242322 samples/sec	accuracy=0.798333	lr=0.160387
Epoch[69] Batch [1049]	Speed: 3051.021131 samples/sec	accuracy=0.798436	lr=0.160173
Epoch[69] Batch [1099]	Speed: 3131.650923 samples/sec	accuracy=0.798429	lr=0.159959
Epoch[69] Batch [1149]	Speed: 2863.755871 samples/sec	accuracy=0.798469	lr=0.159745
Epoch[69] Batch [1199]	Speed: 2999.894654 samples/sec	accuracy=0.798546	lr=0.159531
Epoch[69] Batch [1249]	Speed: 3210.572823 samples/sec	accuracy=0.798678	lr=0.159317
[Epoch 69] training: accuracy=0.798682
[Epoch 69] speed: 2981 samples/sec	time cost: 439.057941
[Epoch 69] validation: err-top1=0.296611 err-top5=0.098267
Epoch[70] Batch [49]	Speed: 3023.916195 samples/sec	accuracy=0.798926	lr=0.159099
Epoch[70] Batch [99]	Speed: 2903.944804 samples/sec	accuracy=0.798478	lr=0.158886
Epoch[70] Batch [149]	Speed: 2938.277167 samples/sec	accuracy=0.799047	lr=0.158672
Epoch[70] Batch [199]	Speed: 2898.157782 samples/sec	accuracy=0.799145	lr=0.158458
Epoch[70] Batch [249]	Speed: 2885.239660 samples/sec	accuracy=0.799241	lr=0.158245
Epoch[70] Batch [299]	Speed: 2909.217509 samples/sec	accuracy=0.799240	lr=0.158031
Epoch[70] Batch [349]	Speed: 2850.156160 samples/sec	accuracy=0.799194	lr=0.157818
Epoch[70] Batch [399]	Speed: 2881.858326 samples/sec	accuracy=0.799218	lr=0.157604
Epoch[70] Batch [449]	Speed: 3077.589654 samples/sec	accuracy=0.799287	lr=0.157391
Epoch[70] Batch [499]	Speed: 3160.370562 samples/sec	accuracy=0.799374	lr=0.157177
Epoch[70] Batch [549]	Speed: 2886.890381 samples/sec	accuracy=0.799373	lr=0.156964
Epoch[70] Batch [599]	Speed: 2979.671559 samples/sec	accuracy=0.799364	lr=0.156751
Epoch[70] Batch [649]	Speed: 2908.021233 samples/sec	accuracy=0.799379	lr=0.156538
Epoch[70] Batch [699]	Speed: 2820.481615 samples/sec	accuracy=0.799516	lr=0.156325
Epoch[70] Batch [749]	Speed: 3007.295338 samples/sec	accuracy=0.799596	lr=0.156112
Epoch[70] Batch [799]	Speed: 2882.710984 samples/sec	accuracy=0.799672	lr=0.155899
Epoch[70] Batch [849]	Speed: 2829.600649 samples/sec	accuracy=0.799738	lr=0.155686
Epoch[70] Batch [899]	Speed: 2946.052798 samples/sec	accuracy=0.799747	lr=0.155473
Epoch[70] Batch [949]	Speed: 2861.220246 samples/sec	accuracy=0.799829	lr=0.155260
Epoch[70] Batch [999]	Speed: 2887.304414 samples/sec	accuracy=0.799875	lr=0.155047
Epoch[70] Batch [1049]	Speed: 3033.260335 samples/sec	accuracy=0.799981	lr=0.154834
Epoch[70] Batch [1099]	Speed: 2872.499008 samples/sec	accuracy=0.800041	lr=0.154622
Epoch[70] Batch [1149]	Speed: 2933.482986 samples/sec	accuracy=0.800101	lr=0.154409
Epoch[70] Batch [1199]	Speed: 2890.780459 samples/sec	accuracy=0.800208	lr=0.154196
Epoch[70] Batch [1249]	Speed: 3030.970171 samples/sec	accuracy=0.800332	lr=0.153984
[Epoch 70] training: accuracy=0.800334
[Epoch 70] speed: 2927 samples/sec	time cost: 447.362722
[Epoch 70] validation: err-top1=0.293347 err-top5=0.095623
Epoch[71] Batch [49]	Speed: 2867.435055 samples/sec	accuracy=0.800155	lr=0.153767
Epoch[71] Batch [99]	Speed: 3036.323941 samples/sec	accuracy=0.799094	lr=0.153555
Epoch[71] Batch [149]	Speed: 2866.753355 samples/sec	accuracy=0.799286	lr=0.153342
Epoch[71] Batch [199]	Speed: 2994.579356 samples/sec	accuracy=0.799179	lr=0.153130
Epoch[71] Batch [249]	Speed: 2965.886272 samples/sec	accuracy=0.799536	lr=0.152918
Epoch[71] Batch [299]	Speed: 2878.289767 samples/sec	accuracy=0.799833	lr=0.152705
Epoch[71] Batch [349]	Speed: 2914.761937 samples/sec	accuracy=0.799980	lr=0.152493
Epoch[71] Batch [399]	Speed: 2954.389896 samples/sec	accuracy=0.800174	lr=0.152281
Epoch[71] Batch [449]	Speed: 2919.126038 samples/sec	accuracy=0.800198	lr=0.152069
Epoch[71] Batch [499]	Speed: 2986.562713 samples/sec	accuracy=0.800313	lr=0.151857
Epoch[71] Batch [549]	Speed: 2904.949056 samples/sec	accuracy=0.800509	lr=0.151645
Epoch[71] Batch [599]	Speed: 2884.429206 samples/sec	accuracy=0.800556	lr=0.151433
Epoch[71] Batch [649]	Speed: 2942.937734 samples/sec	accuracy=0.800655	lr=0.151222
Epoch[71] Batch [699]	Speed: 2891.197828 samples/sec	accuracy=0.800691	lr=0.151010
Epoch[71] Batch [749]	Speed: 3003.361480 samples/sec	accuracy=0.800807	lr=0.150798
Epoch[71] Batch [799]	Speed: 3173.751106 samples/sec	accuracy=0.800897	lr=0.150587
Epoch[71] Batch [849]	Speed: 2856.904370 samples/sec	accuracy=0.800948	lr=0.150375
Epoch[71] Batch [899]	Speed: 2801.232009 samples/sec	accuracy=0.801013	lr=0.150163
Epoch[71] Batch [949]	Speed: 2961.220169 samples/sec	accuracy=0.801157	lr=0.149952
Epoch[71] Batch [999]	Speed: 2794.124755 samples/sec	accuracy=0.801247	lr=0.149741
Epoch[71] Batch [1049]	Speed: 2869.937451 samples/sec	accuracy=0.801395	lr=0.149529
Epoch[71] Batch [1099]	Speed: 2954.168316 samples/sec	accuracy=0.801403	lr=0.149318
Epoch[71] Batch [1149]	Speed: 2964.390447 samples/sec	accuracy=0.801546	lr=0.149107
Epoch[71] Batch [1199]	Speed: 2817.196355 samples/sec	accuracy=0.801674	lr=0.148896
Epoch[71] Batch [1249]	Speed: 3016.677992 samples/sec	accuracy=0.801795	lr=0.148685
[Epoch 71] training: accuracy=0.801797
[Epoch 71] speed: 2924 samples/sec	time cost: 447.909174
[Epoch 71] validation: err-top1=0.288644 err-top5=0.095942
Epoch[72] Batch [49]	Speed: 2948.637189 samples/sec	accuracy=0.802280	lr=0.148469
Epoch[72] Batch [99]	Speed: 2922.783989 samples/sec	accuracy=0.801657	lr=0.148258
Epoch[72] Batch [149]	Speed: 2920.381470 samples/sec	accuracy=0.801304	lr=0.148047
Epoch[72] Batch [199]	Speed: 2852.363058 samples/sec	accuracy=0.800962	lr=0.147837
Epoch[72] Batch [249]	Speed: 2868.947137 samples/sec	accuracy=0.801016	lr=0.147626
Epoch[72] Batch [299]	Speed: 2911.401004 samples/sec	accuracy=0.801309	lr=0.147415
Epoch[72] Batch [349]	Speed: 2977.230079 samples/sec	accuracy=0.801695	lr=0.147204
Epoch[72] Batch [399]	Speed: 2874.294978 samples/sec	accuracy=0.801910	lr=0.146994
Epoch[72] Batch [449]	Speed: 3032.534561 samples/sec	accuracy=0.801996	lr=0.146783
Epoch[72] Batch [499]	Speed: 2992.294046 samples/sec	accuracy=0.802168	lr=0.146573
Epoch[72] Batch [549]	Speed: 2918.206338 samples/sec	accuracy=0.802215	lr=0.146362
Epoch[72] Batch [599]	Speed: 2826.011941 samples/sec	accuracy=0.802199	lr=0.146152
Epoch[72] Batch [649]	Speed: 2802.890224 samples/sec	accuracy=0.802285	lr=0.145942
Epoch[72] Batch [699]	Speed: 2977.426730 samples/sec	accuracy=0.802270	lr=0.145732
Epoch[72] Batch [749]	Speed: 2861.693148 samples/sec	accuracy=0.802363	lr=0.145521
Epoch[72] Batch [799]	Speed: 2891.742604 samples/sec	accuracy=0.802418	lr=0.145311
Epoch[72] Batch [849]	Speed: 2900.717246 samples/sec	accuracy=0.802508	lr=0.145101
Epoch[72] Batch [899]	Speed: 2967.707364 samples/sec	accuracy=0.802622	lr=0.144891
Epoch[72] Batch [949]	Speed: 3030.157070 samples/sec	accuracy=0.802697	lr=0.144682
Epoch[72] Batch [999]	Speed: 2911.288754 samples/sec	accuracy=0.802788	lr=0.144472
Epoch[72] Batch [1049]	Speed: 2971.187875 samples/sec	accuracy=0.802837	lr=0.144262
Epoch[72] Batch [1099]	Speed: 2918.756462 samples/sec	accuracy=0.802861	lr=0.144052
Epoch[72] Batch [1149]	Speed: 2930.847439 samples/sec	accuracy=0.802892	lr=0.143843
Epoch[72] Batch [1199]	Speed: 2988.505360 samples/sec	accuracy=0.802954	lr=0.143633
Epoch[72] Batch [1249]	Speed: 2976.961233 samples/sec	accuracy=0.803004	lr=0.143424
[Epoch 72] training: accuracy=0.803007
[Epoch 72] speed: 2923 samples/sec	time cost: 447.677153
[Epoch 72] validation: err-top1=0.292471 err-top5=0.096142
Epoch[73] Batch [49]	Speed: 3074.666218 samples/sec	accuracy=0.802879	lr=0.143210
Epoch[73] Batch [99]	Speed: 2870.441861 samples/sec	accuracy=0.802562	lr=0.143001
Epoch[73] Batch [149]	Speed: 2984.564601 samples/sec	accuracy=0.802842	lr=0.142791
Epoch[73] Batch [199]	Speed: 2874.019399 samples/sec	accuracy=0.802522	lr=0.142582
Epoch[73] Batch [249]	Speed: 2921.031304 samples/sec	accuracy=0.802650	lr=0.142373
Epoch[73] Batch [299]	Speed: 2856.306534 samples/sec	accuracy=0.802816	lr=0.142164
Epoch[73] Batch [349]	Speed: 2951.604084 samples/sec	accuracy=0.802914	lr=0.141955
Epoch[73] Batch [399]	Speed: 2786.966666 samples/sec	accuracy=0.803048	lr=0.141746
Epoch[73] Batch [449]	Speed: 2830.292359 samples/sec	accuracy=0.803331	lr=0.141537
Epoch[73] Batch [499]	Speed: 2880.768218 samples/sec	accuracy=0.803461	lr=0.141328
Epoch[73] Batch [549]	Speed: 2824.511554 samples/sec	accuracy=0.803500	lr=0.141120
Epoch[73] Batch [599]	Speed: 3180.700108 samples/sec	accuracy=0.803597	lr=0.140911
Epoch[73] Batch [649]	Speed: 3112.421127 samples/sec	accuracy=0.803732	lr=0.140702
Epoch[73] Batch [699]	Speed: 2940.922233 samples/sec	accuracy=0.803834	lr=0.140494
Epoch[73] Batch [749]	Speed: 2792.899336 samples/sec	accuracy=0.803854	lr=0.140285
Epoch[73] Batch [799]	Speed: 2947.559421 samples/sec	accuracy=0.804025	lr=0.140077
Epoch[73] Batch [849]	Speed: 2968.857051 samples/sec	accuracy=0.804146	lr=0.139869
Epoch[73] Batch [899]	Speed: 2897.790993 samples/sec	accuracy=0.804231	lr=0.139661
Epoch[73] Batch [949]	Speed: 2859.242447 samples/sec	accuracy=0.804323	lr=0.139452
Epoch[73] Batch [999]	Speed: 2990.607371 samples/sec	accuracy=0.804425	lr=0.139244
Epoch[73] Batch [1049]	Speed: 2921.004286 samples/sec	accuracy=0.804471	lr=0.139036
Epoch[73] Batch [1099]	Speed: 2961.691376 samples/sec	accuracy=0.804569	lr=0.138828
Epoch[73] Batch [1149]	Speed: 2852.017844 samples/sec	accuracy=0.804674	lr=0.138620
Epoch[73] Batch [1199]	Speed: 2861.120180 samples/sec	accuracy=0.804754	lr=0.138413
Epoch[73] Batch [1249]	Speed: 3048.791597 samples/sec	accuracy=0.804834	lr=0.138205
[Epoch 73] training: accuracy=0.804838
[Epoch 73] speed: 2922 samples/sec	time cost: 448.089567
[Epoch 73] validation: err-top1=0.284857 err-top5=0.091737
Epoch[74] Batch [49]	Speed: 2957.830951 samples/sec	accuracy=0.804054	lr=0.137993
Epoch[74] Batch [99]	Speed: 2883.194424 samples/sec	accuracy=0.803663	lr=0.137786
Epoch[74] Batch [149]	Speed: 2918.370798 samples/sec	accuracy=0.804197	lr=0.137578
Epoch[74] Batch [199]	Speed: 2967.774954 samples/sec	accuracy=0.804069	lr=0.137371
Epoch[74] Batch [249]	Speed: 2828.023380 samples/sec	accuracy=0.804430	lr=0.137163
Epoch[74] Batch [299]	Speed: 2891.718929 samples/sec	accuracy=0.804346	lr=0.136956
Epoch[74] Batch [349]	Speed: 2783.709466 samples/sec	accuracy=0.804230	lr=0.136749
Epoch[74] Batch [399]	Speed: 2923.873050 samples/sec	accuracy=0.804119	lr=0.136542
Epoch[74] Batch [449]	Speed: 3057.477060 samples/sec	accuracy=0.804116	lr=0.136335
Epoch[74] Batch [499]	Speed: 2958.628970 samples/sec	accuracy=0.804125	lr=0.136128
Epoch[74] Batch [549]	Speed: 2941.775186 samples/sec	accuracy=0.804249	lr=0.135921
Epoch[74] Batch [599]	Speed: 2797.810388 samples/sec	accuracy=0.804363	lr=0.135714
Epoch[74] Batch [649]	Speed: 2901.981400 samples/sec	accuracy=0.804421	lr=0.135507
Epoch[74] Batch [699]	Speed: 2859.935434 samples/sec	accuracy=0.804636	lr=0.135301
Epoch[74] Batch [749]	Speed: 2959.872232 samples/sec	accuracy=0.804697	lr=0.135094
Epoch[74] Batch [799]	Speed: 2932.679970 samples/sec	accuracy=0.804838	lr=0.134887
Epoch[74] Batch [849]	Speed: 2965.074222 samples/sec	accuracy=0.804946	lr=0.134681
Epoch[74] Batch [899]	Speed: 2833.363782 samples/sec	accuracy=0.805067	lr=0.134475
Epoch[74] Batch [949]	Speed: 2858.151158 samples/sec	accuracy=0.805045	lr=0.134268
Epoch[74] Batch [999]	Speed: 2882.006453 samples/sec	accuracy=0.805151	lr=0.134062
Epoch[74] Batch [1049]	Speed: 2870.581259 samples/sec	accuracy=0.805270	lr=0.133856
Epoch[74] Batch [1099]	Speed: 2878.236222 samples/sec	accuracy=0.805395	lr=0.133650
Epoch[74] Batch [1149]	Speed: 2941.709622 samples/sec	accuracy=0.805556	lr=0.133444
Epoch[74] Batch [1199]	Speed: 2917.952011 samples/sec	accuracy=0.805717	lr=0.133238
Epoch[74] Batch [1249]	Speed: 3113.115471 samples/sec	accuracy=0.805837	lr=0.133032
[Epoch 74] training: accuracy=0.805839
[Epoch 74] speed: 2909 samples/sec	time cost: 450.035421
[Epoch 74] validation: err-top1=0.281609 err-top5=0.090222
Epoch[75] Batch [49]	Speed: 2932.596348 samples/sec	accuracy=0.807108	lr=0.132822
Epoch[75] Batch [99]	Speed: 2842.629431 samples/sec	accuracy=0.806569	lr=0.132617
Epoch[75] Batch [149]	Speed: 2892.982071 samples/sec	accuracy=0.806154	lr=0.132411
Epoch[75] Batch [199]	Speed: 2900.297243 samples/sec	accuracy=0.806167	lr=0.132206
Epoch[75] Batch [249]	Speed: 3048.680275 samples/sec	accuracy=0.806157	lr=0.132000
Epoch[75] Batch [299]	Speed: 2909.134077 samples/sec	accuracy=0.805928	lr=0.131795
Epoch[75] Batch [349]	Speed: 2851.697896 samples/sec	accuracy=0.805849	lr=0.131590
Epoch[75] Batch [399]	Speed: 2860.561959 samples/sec	accuracy=0.806050	lr=0.131385
Epoch[75] Batch [449]	Speed: 2989.057555 samples/sec	accuracy=0.806140	lr=0.131180
Epoch[75] Batch [499]	Speed: 2894.773980 samples/sec	accuracy=0.806100	lr=0.130975
Epoch[75] Batch [549]	Speed: 2787.201060 samples/sec	accuracy=0.806166	lr=0.130770
Epoch[75] Batch [599]	Speed: 3049.588963 samples/sec	accuracy=0.806255	lr=0.130565
Epoch[75] Batch [649]	Speed: 2877.823551 samples/sec	accuracy=0.806389	lr=0.130360
Epoch[75] Batch [699]	Speed: 2858.594507 samples/sec	accuracy=0.806456	lr=0.130155
Epoch[75] Batch [749]	Speed: 2921.415128 samples/sec	accuracy=0.806533	lr=0.129951
Epoch[75] Batch [799]	Speed: 2999.759135 samples/sec	accuracy=0.806661	lr=0.129746
Epoch[75] Batch [849]	Speed: 2843.575261 samples/sec	accuracy=0.806749	lr=0.129542
Epoch[75] Batch [899]	Speed: 2892.443411 samples/sec	accuracy=0.806790	lr=0.129338
Epoch[75] Batch [949]	Speed: 2835.795130 samples/sec	accuracy=0.807007	lr=0.129133
Epoch[75] Batch [999]	Speed: 2773.451911 samples/sec	accuracy=0.807095	lr=0.128929
Epoch[75] Batch [1049]	Speed: 2935.782697 samples/sec	accuracy=0.807189	lr=0.128725
Epoch[75] Batch [1099]	Speed: 2966.161725 samples/sec	accuracy=0.807235	lr=0.128521
Epoch[75] Batch [1149]	Speed: 2848.913615 samples/sec	accuracy=0.807268	lr=0.128317
Epoch[75] Batch [1199]	Speed: 2924.075775 samples/sec	accuracy=0.807325	lr=0.128113
Epoch[75] Batch [1249]	Speed: 3065.802907 samples/sec	accuracy=0.807471	lr=0.127910
[Epoch 75] training: accuracy=0.807473
[Epoch 75] speed: 2904 samples/sec	time cost: 450.963670
[Epoch 75] validation: err-top1=0.285868 err-top5=0.093669
Epoch[76] Batch [49]	Speed: 3019.132799 samples/sec	accuracy=0.807505	lr=0.127702
Epoch[76] Batch [99]	Speed: 2935.393724 samples/sec	accuracy=0.807124	lr=0.127498
Epoch[76] Batch [149]	Speed: 2913.967673 samples/sec	accuracy=0.806705	lr=0.127295
Epoch[76] Batch [199]	Speed: 2911.426660 samples/sec	accuracy=0.806580	lr=0.127092
Epoch[76] Batch [249]	Speed: 2941.003793 samples/sec	accuracy=0.806888	lr=0.126888
Epoch[76] Batch [299]	Speed: 2918.695093 samples/sec	accuracy=0.806747	lr=0.126685
Epoch[76] Batch [349]	Speed: 2937.643307 samples/sec	accuracy=0.806855	lr=0.126482
Epoch[76] Batch [399]	Speed: 2920.193433 samples/sec	accuracy=0.806878	lr=0.126279
Epoch[76] Batch [449]	Speed: 2967.899313 samples/sec	accuracy=0.807031	lr=0.126076
Epoch[76] Batch [499]	Speed: 2957.682952 samples/sec	accuracy=0.807180	lr=0.125873
Epoch[76] Batch [549]	Speed: 2882.865005 samples/sec	accuracy=0.807260	lr=0.125670
Epoch[76] Batch [599]	Speed: 2926.889223 samples/sec	accuracy=0.807316	lr=0.125468
Epoch[76] Batch [649]	Speed: 2878.502077 samples/sec	accuracy=0.807346	lr=0.125265
Epoch[76] Batch [699]	Speed: 2851.306882 samples/sec	accuracy=0.807376	lr=0.125062
Epoch[76] Batch [749]	Speed: 2987.920400 samples/sec	accuracy=0.807477	lr=0.124860
Epoch[76] Batch [799]	Speed: 2862.720856 samples/sec	accuracy=0.807484	lr=0.124658
Epoch[76] Batch [849]	Speed: 2855.995043 samples/sec	accuracy=0.807639	lr=0.124455
Epoch[76] Batch [899]	Speed: 2924.390427 samples/sec	accuracy=0.807798	lr=0.124253
Epoch[76] Batch [949]	Speed: 2851.762198 samples/sec	accuracy=0.807889	lr=0.124051
Epoch[76] Batch [999]	Speed: 2988.846761 samples/sec	accuracy=0.807944	lr=0.123849
Epoch[76] Batch [1049]	Speed: 2955.412305 samples/sec	accuracy=0.808080	lr=0.123647
Epoch[76] Batch [1099]	Speed: 2903.921282 samples/sec	accuracy=0.808165	lr=0.123446
Epoch[76] Batch [1149]	Speed: 2857.338512 samples/sec	accuracy=0.808282	lr=0.123244
Epoch[76] Batch [1199]	Speed: 2846.251215 samples/sec	accuracy=0.808389	lr=0.123042
Epoch[76] Batch [1249]	Speed: 3167.108939 samples/sec	accuracy=0.808493	lr=0.122841
[Epoch 76] training: accuracy=0.808494
[Epoch 76] speed: 2922 samples/sec	time cost: 448.567465
[Epoch 76] validation: err-top1=0.274235 err-top5=0.085120
Epoch[77] Batch [49]	Speed: 3073.833731 samples/sec	accuracy=0.809311	lr=0.122635
Epoch[77] Batch [99]	Speed: 2981.745549 samples/sec	accuracy=0.808341	lr=0.122434
Epoch[77] Batch [149]	Speed: 2912.936314 samples/sec	accuracy=0.808487	lr=0.122233
Epoch[77] Batch [199]	Speed: 2941.655182 samples/sec	accuracy=0.808323	lr=0.122032
Epoch[77] Batch [249]	Speed: 3006.734111 samples/sec	accuracy=0.808418	lr=0.121831
Epoch[77] Batch [299]	Speed: 2793.616316 samples/sec	accuracy=0.808345	lr=0.121630
Epoch[77] Batch [349]	Speed: 2904.238014 samples/sec	accuracy=0.808429	lr=0.121429
Epoch[77] Batch [399]	Speed: 2828.477696 samples/sec	accuracy=0.808504	lr=0.121228
Epoch[77] Batch [449]	Speed: 2877.557127 samples/sec	accuracy=0.808644	lr=0.121027
Epoch[77] Batch [499]	Speed: 2907.591671 samples/sec	accuracy=0.808755	lr=0.120827
Epoch[77] Batch [549]	Speed: 3070.953345 samples/sec	accuracy=0.808889	lr=0.120626
Epoch[77] Batch [599]	Speed: 2876.020589 samples/sec	accuracy=0.809027	lr=0.120426
Epoch[77] Batch [649]	Speed: 3151.065348 samples/sec	accuracy=0.809203	lr=0.120226
Epoch[77] Batch [699]	Speed: 2869.695225 samples/sec	accuracy=0.809263	lr=0.120025
Epoch[77] Batch [749]	Speed: 2840.692385 samples/sec	accuracy=0.809345	lr=0.119825
Epoch[77] Batch [799]	Speed: 2939.837621 samples/sec	accuracy=0.809464	lr=0.119625
Epoch[77] Batch [849]	Speed: 2846.465880 samples/sec	accuracy=0.809543	lr=0.119425
Epoch[77] Batch [899]	Speed: 3032.782443 samples/sec	accuracy=0.809668	lr=0.119226
Epoch[77] Batch [949]	Speed: 2866.701348 samples/sec	accuracy=0.809716	lr=0.119026
Epoch[77] Batch [999]	Speed: 2940.520543 samples/sec	accuracy=0.809835	lr=0.118826
Epoch[77] Batch [1049]	Speed: 3066.678434 samples/sec	accuracy=0.809988	lr=0.118627
Epoch[77] Batch [1099]	Speed: 2914.676525 samples/sec	accuracy=0.810063	lr=0.118427
Epoch[77] Batch [1149]	Speed: 2923.949248 samples/sec	accuracy=0.810200	lr=0.118228
Epoch[77] Batch [1199]	Speed: 2976.459702 samples/sec	accuracy=0.810271	lr=0.118029
Epoch[77] Batch [1249]	Speed: 3021.485533 samples/sec	accuracy=0.810384	lr=0.117830
[Epoch 77] training: accuracy=0.810383
[Epoch 77] speed: 2937 samples/sec	time cost: 445.880108
[Epoch 77] validation: err-top1=0.275411 err-top5=0.087173
Epoch[78] Batch [49]	Speed: 3026.401439 samples/sec	accuracy=0.811388	lr=0.117627
Epoch[78] Batch [99]	Speed: 2884.074676 samples/sec	accuracy=0.811222	lr=0.117428
Epoch[78] Batch [149]	Speed: 2858.795130 samples/sec	accuracy=0.810795	lr=0.117229
Epoch[78] Batch [199]	Speed: 3010.625828 samples/sec	accuracy=0.810690	lr=0.117030
Epoch[78] Batch [249]	Speed: 3131.601511 samples/sec	accuracy=0.810585	lr=0.116831
Epoch[78] Batch [299]	Speed: 3047.735145 samples/sec	accuracy=0.810441	lr=0.116633
Epoch[78] Batch [349]	Speed: 2920.338499 samples/sec	accuracy=0.810135	lr=0.116434
Epoch[78] Batch [399]	Speed: 2995.239615 samples/sec	accuracy=0.810138	lr=0.116236
Epoch[78] Batch [449]	Speed: 2883.375170 samples/sec	accuracy=0.810387	lr=0.116038
Epoch[78] Batch [499]	Speed: 2862.111962 samples/sec	accuracy=0.810465	lr=0.115840
Epoch[78] Batch [549]	Speed: 2922.847359 samples/sec	accuracy=0.810555	lr=0.115642
Epoch[78] Batch [599]	Speed: 2992.403832 samples/sec	accuracy=0.810661	lr=0.115444
Epoch[78] Batch [649]	Speed: 2784.297294 samples/sec	accuracy=0.810658	lr=0.115246
Epoch[78] Batch [699]	Speed: 2879.053697 samples/sec	accuracy=0.810734	lr=0.115048
Epoch[78] Batch [749]	Speed: 2765.900427 samples/sec	accuracy=0.810736	lr=0.114850
Epoch[78] Batch [799]	Speed: 2867.901473 samples/sec	accuracy=0.810767	lr=0.114653
Epoch[78] Batch [849]	Speed: 2837.835562 samples/sec	accuracy=0.810942	lr=0.114455
Epoch[78] Batch [899]	Speed: 2956.303760 samples/sec	accuracy=0.811015	lr=0.114258
Epoch[78] Batch [949]	Speed: 2967.612752 samples/sec	accuracy=0.811061	lr=0.114061
Epoch[78] Batch [999]	Speed: 2984.609524 samples/sec	accuracy=0.811115	lr=0.113864
Epoch[78] Batch [1049]	Speed: 2854.027304 samples/sec	accuracy=0.811162	lr=0.113667
Epoch[78] Batch [1099]	Speed: 2986.145013 samples/sec	accuracy=0.811256	lr=0.113470
Epoch[78] Batch [1149]	Speed: 2988.822343 samples/sec	accuracy=0.811313	lr=0.113273
Epoch[78] Batch [1199]	Speed: 3135.586826 samples/sec	accuracy=0.811383	lr=0.113076
Epoch[78] Batch [1249]	Speed: 3174.029463 samples/sec	accuracy=0.811441	lr=0.112880
[Epoch 78] training: accuracy=0.811444
[Epoch 78] speed: 2943 samples/sec	time cost: 445.209639
[Epoch 78] validation: err-top1=0.272421 err-top5=0.084104
Epoch[79] Batch [49]	Speed: 3174.948283 samples/sec	accuracy=0.810976	lr=0.112679
Epoch[79] Batch [99]	Speed: 2888.947861 samples/sec	accuracy=0.811560	lr=0.112483
Epoch[79] Batch [149]	Speed: 2851.757805 samples/sec	accuracy=0.811680	lr=0.112287
Epoch[79] Batch [199]	Speed: 3032.198520 samples/sec	accuracy=0.811412	lr=0.112090
Epoch[79] Batch [249]	Speed: 2908.848940 samples/sec	accuracy=0.811054	lr=0.111894
Epoch[79] Batch [299]	Speed: 2962.534718 samples/sec	accuracy=0.811187	lr=0.111698
Epoch[79] Batch [349]	Speed: 2877.208987 samples/sec	accuracy=0.811423	lr=0.111502
Epoch[79] Batch [399]	Speed: 3032.461249 samples/sec	accuracy=0.811478	lr=0.111307
Epoch[79] Batch [449]	Speed: 2945.952813 samples/sec	accuracy=0.811626	lr=0.111111
Epoch[79] Batch [499]	Speed: 2945.658918 samples/sec	accuracy=0.811574	lr=0.110915
Epoch[79] Batch [549]	Speed: 2921.118439 samples/sec	accuracy=0.811693	lr=0.110720
Epoch[79] Batch [599]	Speed: 2942.219625 samples/sec	accuracy=0.811724	lr=0.110525
Epoch[79] Batch [649]	Speed: 2969.228422 samples/sec	accuracy=0.811749	lr=0.110329
Epoch[79] Batch [699]	Speed: 3109.806486 samples/sec	accuracy=0.811798	lr=0.110134
Epoch[79] Batch [749]	Speed: 3129.522574 samples/sec	accuracy=0.811895	lr=0.109939
Epoch[79] Batch [799]	Speed: 2918.071952 samples/sec	accuracy=0.812030	lr=0.109744
Epoch[79] Batch [849]	Speed: 3207.751121 samples/sec	accuracy=0.812090	lr=0.109549
Epoch[79] Batch [899]	Speed: 3066.982126 samples/sec	accuracy=0.812190	lr=0.109355
Epoch[79] Batch [949]	Speed: 2884.442069 samples/sec	accuracy=0.812307	lr=0.109160
Epoch[79] Batch [999]	Speed: 3021.029321 samples/sec	accuracy=0.812417	lr=0.108966
Epoch[79] Batch [1049]	Speed: 2971.641082 samples/sec	accuracy=0.812496	lr=0.108771
Epoch[79] Batch [1099]	Speed: 2969.939733 samples/sec	accuracy=0.812606	lr=0.108577
Epoch[79] Batch [1149]	Speed: 2870.419033 samples/sec	accuracy=0.812758	lr=0.108383
Epoch[79] Batch [1199]	Speed: 3055.546383 samples/sec	accuracy=0.812863	lr=0.108189
Epoch[79] Batch [1249]	Speed: 2976.203988 samples/sec	accuracy=0.812974	lr=0.107995
[Epoch 79] training: accuracy=0.812975
[Epoch 79] speed: 2980 samples/sec	time cost: 439.547846
[Epoch 79] validation: err-top1=0.267937 err-top5=0.084642
Epoch[80] Batch [49]	Speed: 2910.100357 samples/sec	accuracy=0.812837	lr=0.107797
Epoch[80] Batch [99]	Speed: 3057.699257 samples/sec	accuracy=0.812092	lr=0.107603
Epoch[80] Batch [149]	Speed: 3047.114449 samples/sec	accuracy=0.811986	lr=0.107410
Epoch[80] Batch [199]	Speed: 3011.352004 samples/sec	accuracy=0.812513	lr=0.107216
Epoch[80] Batch [249]	Speed: 2867.360741 samples/sec	accuracy=0.812898	lr=0.107023
Epoch[80] Batch [299]	Speed: 3007.363522 samples/sec	accuracy=0.812918	lr=0.106830
Epoch[80] Batch [349]	Speed: 2845.909930 samples/sec	accuracy=0.812984	lr=0.106636
Epoch[80] Batch [399]	Speed: 2987.159315 samples/sec	accuracy=0.813020	lr=0.106443
Epoch[80] Batch [449]	Speed: 2953.307432 samples/sec	accuracy=0.813028	lr=0.106250
Epoch[80] Batch [499]	Speed: 3048.637081 samples/sec	accuracy=0.813091	lr=0.106057
Epoch[80] Batch [549]	Speed: 2895.931223 samples/sec	accuracy=0.813197	lr=0.105865
Epoch[80] Batch [599]	Speed: 2856.576637 samples/sec	accuracy=0.813383	lr=0.105672
Epoch[80] Batch [649]	Speed: 3028.251764 samples/sec	accuracy=0.813384	lr=0.105480
Epoch[80] Batch [699]	Speed: 2930.681730 samples/sec	accuracy=0.813487	lr=0.105287
Epoch[80] Batch [749]	Speed: 2909.403661 samples/sec	accuracy=0.813665	lr=0.105095
Epoch[80] Batch [799]	Speed: 2903.764297 samples/sec	accuracy=0.813717	lr=0.104903
Epoch[80] Batch [849]	Speed: 2868.721480 samples/sec	accuracy=0.813796	lr=0.104711
Epoch[80] Batch [899]	Speed: 2833.331932 samples/sec	accuracy=0.813801	lr=0.104519
Epoch[80] Batch [949]	Speed: 3053.867096 samples/sec	accuracy=0.813842	lr=0.104327
Epoch[80] Batch [999]	Speed: 2982.208734 samples/sec	accuracy=0.813905	lr=0.104135
Epoch[80] Batch [1049]	Speed: 2880.248969 samples/sec	accuracy=0.813939	lr=0.103944
Epoch[80] Batch [1099]	Speed: 2977.627783 samples/sec	accuracy=0.814058	lr=0.103752
Epoch[80] Batch [1149]	Speed: 2975.621401 samples/sec	accuracy=0.814118	lr=0.103561
Epoch[80] Batch [1199]	Speed: 2939.334840 samples/sec	accuracy=0.814237	lr=0.103370
Epoch[80] Batch [1249]	Speed: 3168.740561 samples/sec	accuracy=0.814286	lr=0.103178
[Epoch 80] training: accuracy=0.814288
[Epoch 80] speed: 2953 samples/sec	time cost: 443.663174
[Epoch 80] validation: err-top1=0.269252 err-top5=0.081732
Epoch[81] Batch [49]	Speed: 3025.270389 samples/sec	accuracy=0.814953	lr=0.102984
Epoch[81] Batch [99]	Speed: 2812.620902 samples/sec	accuracy=0.814629	lr=0.102793
Epoch[81] Batch [149]	Speed: 3014.713624 samples/sec	accuracy=0.814617	lr=0.102602
Epoch[81] Batch [199]	Speed: 2923.920146 samples/sec	accuracy=0.814659	lr=0.102411
Epoch[81] Batch [249]	Speed: 2988.332110 samples/sec	accuracy=0.814630	lr=0.102221
Epoch[81] Batch [299]	Speed: 2945.508942 samples/sec	accuracy=0.814634	lr=0.102030
Epoch[81] Batch [349]	Speed: 2793.169459 samples/sec	accuracy=0.814776	lr=0.101840
Epoch[81] Batch [399]	Speed: 2907.282118 samples/sec	accuracy=0.814686	lr=0.101650
Epoch[81] Batch [449]	Speed: 2859.134488 samples/sec	accuracy=0.814750	lr=0.101460
Epoch[81] Batch [499]	Speed: 2920.311018 samples/sec	accuracy=0.814752	lr=0.101270
Epoch[81] Batch [549]	Speed: 2946.680670 samples/sec	accuracy=0.814791	lr=0.101080
Epoch[81] Batch [599]	Speed: 3006.967688 samples/sec	accuracy=0.814844	lr=0.100890
Epoch[81] Batch [649]	Speed: 2986.607114 samples/sec	accuracy=0.815036	lr=0.100701
Epoch[81] Batch [699]	Speed: 2942.723030 samples/sec	accuracy=0.815033	lr=0.100511
Epoch[81] Batch [749]	Speed: 3023.544982 samples/sec	accuracy=0.815030	lr=0.100322
Epoch[81] Batch [799]	Speed: 2938.349172 samples/sec	accuracy=0.815090	lr=0.100132
Epoch[81] Batch [849]	Speed: 2942.243973 samples/sec	accuracy=0.815108	lr=0.099943
Epoch[81] Batch [899]	Speed: 2896.864678 samples/sec	accuracy=0.815095	lr=0.099754
Epoch[81] Batch [949]	Speed: 2832.585046 samples/sec	accuracy=0.815167	lr=0.099565
Epoch[81] Batch [999]	Speed: 2779.989608 samples/sec	accuracy=0.815240	lr=0.099377
Epoch[81] Batch [1049]	Speed: 2889.466830 samples/sec	accuracy=0.815420	lr=0.099188
Epoch[81] Batch [1099]	Speed: 2877.335279 samples/sec	accuracy=0.815517	lr=0.098999
Epoch[81] Batch [1149]	Speed: 2861.124983 samples/sec	accuracy=0.815554	lr=0.098811
Epoch[81] Batch [1199]	Speed: 2882.337727 samples/sec	accuracy=0.815720	lr=0.098623
Epoch[81] Batch [1249]	Speed: 3160.325075 samples/sec	accuracy=0.815797	lr=0.098434
[Epoch 81] training: accuracy=0.815802
[Epoch 81] speed: 2921 samples/sec	time cost: 448.328680
[Epoch 81] validation: err-top1=0.265808 err-top5=0.080017
Epoch[82] Batch [49]	Speed: 3048.314727 samples/sec	accuracy=0.816036	lr=0.098243
Epoch[82] Batch [99]	Speed: 2991.156511 samples/sec	accuracy=0.815620	lr=0.098055
Epoch[82] Batch [149]	Speed: 2874.080019 samples/sec	accuracy=0.815311	lr=0.097867
Epoch[82] Batch [199]	Speed: 2949.817485 samples/sec	accuracy=0.815383	lr=0.097679
Epoch[82] Batch [249]	Speed: 2853.292176 samples/sec	accuracy=0.815161	lr=0.097492
Epoch[82] Batch [299]	Speed: 2845.566050 samples/sec	accuracy=0.815387	lr=0.097304
Epoch[82] Batch [349]	Speed: 2954.682161 samples/sec	accuracy=0.815509	lr=0.097117
Epoch[82] Batch [399]	Speed: 2896.385472 samples/sec	accuracy=0.815674	lr=0.096930
Epoch[82] Batch [449]	Speed: 3043.228455 samples/sec	accuracy=0.815776	lr=0.096743
Epoch[82] Batch [499]	Speed: 2884.319375 samples/sec	accuracy=0.815936	lr=0.096556
Epoch[82] Batch [549]	Speed: 2979.844798 samples/sec	accuracy=0.815924	lr=0.096369
Epoch[82] Batch [599]	Speed: 2883.287756 samples/sec	accuracy=0.815968	lr=0.096182
Epoch[82] Batch [649]	Speed: 2878.600777 samples/sec	accuracy=0.815940	lr=0.095996
Epoch[82] Batch [699]	Speed: 2879.296386 samples/sec	accuracy=0.816046	lr=0.095809
Epoch[82] Batch [749]	Speed: 2934.442976 samples/sec	accuracy=0.816145	lr=0.095623
Epoch[82] Batch [799]	Speed: 2880.384762 samples/sec	accuracy=0.816199	lr=0.095437
Epoch[82] Batch [849]	Speed: 3039.737403 samples/sec	accuracy=0.816272	lr=0.095250
Epoch[82] Batch [899]	Speed: 2837.360015 samples/sec	accuracy=0.816402	lr=0.095064
Epoch[82] Batch [949]	Speed: 2922.572096 samples/sec	accuracy=0.816523	lr=0.094879
Epoch[82] Batch [999]	Speed: 2880.353159 samples/sec	accuracy=0.816618	lr=0.094693
Epoch[82] Batch [1049]	Speed: 2901.980577 samples/sec	accuracy=0.816726	lr=0.094507
Epoch[82] Batch [1099]	Speed: 2856.169659 samples/sec	accuracy=0.816808	lr=0.094322
Epoch[82] Batch [1149]	Speed: 2798.828500 samples/sec	accuracy=0.816865	lr=0.094137
Epoch[82] Batch [1199]	Speed: 2816.397853 samples/sec	accuracy=0.816973	lr=0.093951
Epoch[82] Batch [1249]	Speed: 3055.311893 samples/sec	accuracy=0.817018	lr=0.093766
[Epoch 82] training: accuracy=0.817018
[Epoch 82] speed: 2911 samples/sec	time cost: 450.118398
[Epoch 82] validation: err-top1=0.269192 err-top5=0.082669
Epoch[83] Batch [49]	Speed: 2946.168756 samples/sec	accuracy=0.816598	lr=0.093578
Epoch[83] Batch [99]	Speed: 2884.577405 samples/sec	accuracy=0.817151	lr=0.093393
Epoch[83] Batch [149]	Speed: 2965.381627 samples/sec	accuracy=0.816847	lr=0.093208
Epoch[83] Batch [199]	Speed: 2944.084370 samples/sec	accuracy=0.816857	lr=0.093024
Epoch[83] Batch [249]	Speed: 2861.807327 samples/sec	accuracy=0.816720	lr=0.092839
Epoch[83] Batch [299]	Speed: 2850.322611 samples/sec	accuracy=0.816700	lr=0.092655
Epoch[83] Batch [349]	Speed: 2809.517916 samples/sec	accuracy=0.816731	lr=0.092471
Epoch[83] Batch [399]	Speed: 3096.926934 samples/sec	accuracy=0.816807	lr=0.092287
Epoch[83] Batch [449]	Speed: 2966.976626 samples/sec	accuracy=0.816960	lr=0.092103
Epoch[83] Batch [499]	Speed: 2881.693701 samples/sec	accuracy=0.817068	lr=0.091919
Epoch[83] Batch [549]	Speed: 2943.713491 samples/sec	accuracy=0.817112	lr=0.091735
Epoch[83] Batch [599]	Speed: 2896.583074 samples/sec	accuracy=0.817253	lr=0.091552
Epoch[83] Batch [649]	Speed: 2784.934741 samples/sec	accuracy=0.817340	lr=0.091368
Epoch[83] Batch [699]	Speed: 2907.741275 samples/sec	accuracy=0.817367	lr=0.091185
Epoch[83] Batch [749]	Speed: 2859.408362 samples/sec	accuracy=0.817347	lr=0.091002
Epoch[83] Batch [799]	Speed: 2818.489283 samples/sec	accuracy=0.817414	lr=0.090819
Epoch[83] Batch [849]	Speed: 2886.654559 samples/sec	accuracy=0.817474	lr=0.090636
Epoch[83] Batch [899]	Speed: 2841.672234 samples/sec	accuracy=0.817489	lr=0.090453
Epoch[83] Batch [949]	Speed: 2899.840747 samples/sec	accuracy=0.817525	lr=0.090270
Epoch[83] Batch [999]	Speed: 2872.122475 samples/sec	accuracy=0.817632	lr=0.090088
Epoch[83] Batch [1049]	Speed: 2937.050451 samples/sec	accuracy=0.817831	lr=0.089905
Epoch[83] Batch [1099]	Speed: 2890.858950 samples/sec	accuracy=0.817947	lr=0.089723
Epoch[83] Batch [1149]	Speed: 2840.135946 samples/sec	accuracy=0.818045	lr=0.089541
Epoch[83] Batch [1199]	Speed: 2810.017488 samples/sec	accuracy=0.818129	lr=0.089359
Epoch[83] Batch [1249]	Speed: 3012.086305 samples/sec	accuracy=0.818222	lr=0.089177
[Epoch 83] training: accuracy=0.818225
[Epoch 83] speed: 2892 samples/sec	time cost: 452.624603
[Epoch 83] validation: err-top1=0.261201 err-top5=0.078803
Epoch[84] Batch [49]	Speed: 2870.668288 samples/sec	accuracy=0.819813	lr=0.088992
Epoch[84] Batch [99]	Speed: 2872.277594 samples/sec	accuracy=0.819279	lr=0.088810
Epoch[84] Batch [149]	Speed: 2794.536788 samples/sec	accuracy=0.818949	lr=0.088629
Epoch[84] Batch [199]	Speed: 2974.899042 samples/sec	accuracy=0.818819	lr=0.088448
Epoch[84] Batch [249]	Speed: 2954.921748 samples/sec	accuracy=0.818710	lr=0.088266
Epoch[84] Batch [299]	Speed: 2899.265202 samples/sec	accuracy=0.818605	lr=0.088085
Epoch[84] Batch [349]	Speed: 2877.349659 samples/sec	accuracy=0.818608	lr=0.087904
Epoch[84] Batch [399]	Speed: 2857.568124 samples/sec	accuracy=0.818575	lr=0.087724
Epoch[84] Batch [449]	Speed: 2821.798251 samples/sec	accuracy=0.818586	lr=0.087543
Epoch[84] Batch [499]	Speed: 3004.339388 samples/sec	accuracy=0.818600	lr=0.087363
Epoch[84] Batch [549]	Speed: 2899.753389 samples/sec	accuracy=0.818600	lr=0.087182
Epoch[84] Batch [599]	Speed: 2927.278299 samples/sec	accuracy=0.818702	lr=0.087002
Epoch[84] Batch [649]	Speed: 2898.806410 samples/sec	accuracy=0.818647	lr=0.086822
Epoch[84] Batch [699]	Speed: 3025.968088 samples/sec	accuracy=0.818823	lr=0.086642
Epoch[84] Batch [749]	Speed: 2815.273173 samples/sec	accuracy=0.818993	lr=0.086462
Epoch[84] Batch [799]	Speed: 2801.269097 samples/sec	accuracy=0.819095	lr=0.086282
Epoch[84] Batch [849]	Speed: 2910.154739 samples/sec	accuracy=0.819206	lr=0.086103
Epoch[84] Batch [899]	Speed: 2931.523910 samples/sec	accuracy=0.819300	lr=0.085923
Epoch[84] Batch [949]	Speed: 2982.793861 samples/sec	accuracy=0.819329	lr=0.085744
Epoch[84] Batch [999]	Speed: 2885.761330 samples/sec	accuracy=0.819384	lr=0.085565
Epoch[84] Batch [1049]	Speed: 3029.775475 samples/sec	accuracy=0.819476	lr=0.085386
Epoch[84] Batch [1099]	Speed: 2913.531727 samples/sec	accuracy=0.819516	lr=0.085207
Epoch[84] Batch [1149]	Speed: 2845.762171 samples/sec	accuracy=0.819569	lr=0.085028
Epoch[84] Batch [1199]	Speed: 2860.844072 samples/sec	accuracy=0.819674	lr=0.084850
Epoch[84] Batch [1249]	Speed: 3009.134552 samples/sec	accuracy=0.819754	lr=0.084671
[Epoch 84] training: accuracy=0.819756
[Epoch 84] speed: 2902 samples/sec	time cost: 451.124666
[Epoch 84] validation: err-top1=0.260563 err-top5=0.077706
Epoch[85] Batch [49]	Speed: 2957.575658 samples/sec	accuracy=0.819619	lr=0.084489
Epoch[85] Batch [99]	Speed: 2937.279794 samples/sec	accuracy=0.820254	lr=0.084311
Epoch[85] Batch [149]	Speed: 2889.941105 samples/sec	accuracy=0.820211	lr=0.084133
Epoch[85] Batch [199]	Speed: 3084.847111 samples/sec	accuracy=0.820035	lr=0.083955
Epoch[85] Batch [249]	Speed: 2962.270931 samples/sec	accuracy=0.819849	lr=0.083777
Epoch[85] Batch [299]	Speed: 2918.257058 samples/sec	accuracy=0.820002	lr=0.083600
Epoch[85] Batch [349]	Speed: 3015.144689 samples/sec	accuracy=0.820123	lr=0.083422
Epoch[85] Batch [399]	Speed: 3027.592750 samples/sec	accuracy=0.820228	lr=0.083245
Epoch[85] Batch [449]	Speed: 3030.865194 samples/sec	accuracy=0.820360	lr=0.083067
Epoch[85] Batch [499]	Speed: 2982.851036 samples/sec	accuracy=0.820379	lr=0.082890
Epoch[85] Batch [549]	Speed: 3076.490456 samples/sec	accuracy=0.820477	lr=0.082713
Epoch[85] Batch [599]	Speed: 2998.073202 samples/sec	accuracy=0.820519	lr=0.082537
Epoch[85] Batch [649]	Speed: 2883.774952 samples/sec	accuracy=0.820565	lr=0.082360
Epoch[85] Batch [699]	Speed: 2852.871353 samples/sec	accuracy=0.820727	lr=0.082183
Epoch[85] Batch [749]	Speed: 2904.474597 samples/sec	accuracy=0.820758	lr=0.082007
Epoch[85] Batch [799]	Speed: 2960.286408 samples/sec	accuracy=0.820844	lr=0.081831
Epoch[85] Batch [849]	Speed: 2966.104328 samples/sec	accuracy=0.820833	lr=0.081655
Epoch[85] Batch [899]	Speed: 2934.855601 samples/sec	accuracy=0.820909	lr=0.081479
Epoch[85] Batch [949]	Speed: 2869.174977 samples/sec	accuracy=0.821026	lr=0.081303
Epoch[85] Batch [999]	Speed: 2901.824311 samples/sec	accuracy=0.821125	lr=0.081127
Epoch[85] Batch [1049]	Speed: 2943.311765 samples/sec	accuracy=0.821165	lr=0.080952
Epoch[85] Batch [1099]	Speed: 2883.117936 samples/sec	accuracy=0.821173	lr=0.080776
Epoch[85] Batch [1149]	Speed: 2801.228355 samples/sec	accuracy=0.821232	lr=0.080601
Epoch[85] Batch [1199]	Speed: 3103.776714 samples/sec	accuracy=0.821353	lr=0.080426
Epoch[85] Batch [1249]	Speed: 3065.083483 samples/sec	accuracy=0.821448	lr=0.080251
[Epoch 85] training: accuracy=0.821458
[Epoch 85] speed: 2954 samples/sec	time cost: 443.790545
[Epoch 85] validation: err-top1=0.257454 err-top5=0.074777
Epoch[86] Batch [49]	Speed: 2976.476492 samples/sec	accuracy=0.823046	lr=0.080073
Epoch[86] Batch [99]	Speed: 2867.992094 samples/sec	accuracy=0.821954	lr=0.079898
Epoch[86] Batch [149]	Speed: 2908.971090 samples/sec	accuracy=0.821671	lr=0.079723
Epoch[86] Batch [199]	Speed: 2964.762952 samples/sec	accuracy=0.821899	lr=0.079549
Epoch[86] Batch [249]	Speed: 2916.490507 samples/sec	accuracy=0.821893	lr=0.079375
Epoch[86] Batch [299]	Speed: 2926.222266 samples/sec	accuracy=0.821876	lr=0.079201
Epoch[86] Batch [349]	Speed: 3053.442385 samples/sec	accuracy=0.821719	lr=0.079027
Epoch[86] Batch [399]	Speed: 2957.225318 samples/sec	accuracy=0.821683	lr=0.078853
Epoch[86] Batch [449]	Speed: 3079.331390 samples/sec	accuracy=0.821692	lr=0.078679
Epoch[86] Batch [499]	Speed: 2912.273567 samples/sec	accuracy=0.821673	lr=0.078506
Epoch[86] Batch [549]	Speed: 2889.116270 samples/sec	accuracy=0.821745	lr=0.078332
Epoch[86] Batch [599]	Speed: 2895.673266 samples/sec	accuracy=0.821840	lr=0.078159
Epoch[86] Batch [649]	Speed: 2837.498729 samples/sec	accuracy=0.821979	lr=0.077986
Epoch[86] Batch [699]	Speed: 2968.182894 samples/sec	accuracy=0.822066	lr=0.077813
Epoch[86] Batch [749]	Speed: 2900.595789 samples/sec	accuracy=0.822141	lr=0.077640
Epoch[86] Batch [799]	Speed: 2858.416549 samples/sec	accuracy=0.822229	lr=0.077467
Epoch[86] Batch [849]	Speed: 3035.412542 samples/sec	accuracy=0.822313	lr=0.077295
Epoch[86] Batch [899]	Speed: 2791.210942 samples/sec	accuracy=0.822319	lr=0.077123
Epoch[86] Batch [949]	Speed: 3035.171393 samples/sec	accuracy=0.822406	lr=0.076950
Epoch[86] Batch [999]	Speed: 2921.150108 samples/sec	accuracy=0.822474	lr=0.076778
Epoch[86] Batch [1049]	Speed: 2851.409330 samples/sec	accuracy=0.822484	lr=0.076606
Epoch[86] Batch [1099]	Speed: 3116.282810 samples/sec	accuracy=0.822524	lr=0.076435
Epoch[86] Batch [1149]	Speed: 2964.146908 samples/sec	accuracy=0.822543	lr=0.076263
Epoch[86] Batch [1199]	Speed: 2903.120552 samples/sec	accuracy=0.822593	lr=0.076091
Epoch[86] Batch [1249]	Speed: 3043.460706 samples/sec	accuracy=0.822658	lr=0.075920
[Epoch 86] training: accuracy=0.822660
[Epoch 86] speed: 2938 samples/sec	time cost: 445.778518
[Epoch 86] validation: err-top1=0.256099 err-top5=0.075295
Epoch[87] Batch [49]	Speed: 3127.109073 samples/sec	accuracy=0.822011	lr=0.075745
Epoch[87] Batch [99]	Speed: 2903.241043 samples/sec	accuracy=0.822256	lr=0.075574
Epoch[87] Batch [149]	Speed: 2912.137910 samples/sec	accuracy=0.822107	lr=0.075404
Epoch[87] Batch [199]	Speed: 2935.239015 samples/sec	accuracy=0.821915	lr=0.075233
Epoch[87] Batch [249]	Speed: 3000.464608 samples/sec	accuracy=0.821851	lr=0.075062
Epoch[87] Batch [299]	Speed: 2971.166334 samples/sec	accuracy=0.822073	lr=0.074892
Epoch[87] Batch [349]	Speed: 2905.795304 samples/sec	accuracy=0.822225	lr=0.074721
Epoch[87] Batch [399]	Speed: 2845.661599 samples/sec	accuracy=0.822464	lr=0.074551
Epoch[87] Batch [449]	Speed: 2896.717168 samples/sec	accuracy=0.822576	lr=0.074381
Epoch[87] Batch [499]	Speed: 2957.702097 samples/sec	accuracy=0.822650	lr=0.074211
Epoch[87] Batch [549]	Speed: 2921.752741 samples/sec	accuracy=0.822751	lr=0.074042
Epoch[87] Batch [599]	Speed: 2797.922150 samples/sec	accuracy=0.822803	lr=0.073872
Epoch[87] Batch [649]	Speed: 2850.303733 samples/sec	accuracy=0.822966	lr=0.073703
Epoch[87] Batch [699]	Speed: 2912.505970 samples/sec	accuracy=0.822962	lr=0.073534
Epoch[87] Batch [749]	Speed: 2949.231978 samples/sec	accuracy=0.822943	lr=0.073364
Epoch[87] Batch [799]	Speed: 2893.727695 samples/sec	accuracy=0.823029	lr=0.073196
Epoch[87] Batch [849]	Speed: 2886.334784 samples/sec	accuracy=0.823091	lr=0.073027
Epoch[87] Batch [899]	Speed: 2953.209674 samples/sec	accuracy=0.823201	lr=0.072858
Epoch[87] Batch [949]	Speed: 3102.620766 samples/sec	accuracy=0.823237	lr=0.072690
Epoch[87] Batch [999]	Speed: 2976.169340 samples/sec	accuracy=0.823269	lr=0.072521
Epoch[87] Batch [1049]	Speed: 2925.898848 samples/sec	accuracy=0.823375	lr=0.072353
Epoch[87] Batch [1099]	Speed: 3022.232313 samples/sec	accuracy=0.823475	lr=0.072185
Epoch[87] Batch [1149]	Speed: 2998.727466 samples/sec	accuracy=0.823562	lr=0.072017
Epoch[87] Batch [1199]	Speed: 3012.563021 samples/sec	accuracy=0.823611	lr=0.071849
Epoch[87] Batch [1249]	Speed: 3147.935890 samples/sec	accuracy=0.823742	lr=0.071682
[Epoch 87] training: accuracy=0.823746
[Epoch 87] speed: 2947 samples/sec	time cost: 444.408394
[Epoch 87] validation: err-top1=0.250631 err-top5=0.073242
Epoch[88] Batch [49]	Speed: 2865.290224 samples/sec	accuracy=0.824007	lr=0.071511
Epoch[88] Batch [99]	Speed: 2866.286354 samples/sec	accuracy=0.823113	lr=0.071344
Epoch[88] Batch [149]	Speed: 2874.129717 samples/sec	accuracy=0.823484	lr=0.071177
Epoch[88] Batch [199]	Speed: 2854.865051 samples/sec	accuracy=0.823532	lr=0.071010
Epoch[88] Batch [249]	Speed: 2863.516253 samples/sec	accuracy=0.823434	lr=0.070843
Epoch[88] Batch [299]	Speed: 2958.786074 samples/sec	accuracy=0.823632	lr=0.070676
Epoch[88] Batch [349]	Speed: 2921.957159 samples/sec	accuracy=0.823574	lr=0.070510
Epoch[88] Batch [399]	Speed: 2984.938377 samples/sec	accuracy=0.823880	lr=0.070343
Epoch[88] Batch [449]	Speed: 2976.742651 samples/sec	accuracy=0.824018	lr=0.070177
Epoch[88] Batch [499]	Speed: 2965.343300 samples/sec	accuracy=0.824024	lr=0.070011
Epoch[88] Batch [549]	Speed: 2882.998799 samples/sec	accuracy=0.824095	lr=0.069845
Epoch[88] Batch [599]	Speed: 2892.852141 samples/sec	accuracy=0.824213	lr=0.069680
Epoch[88] Batch [649]	Speed: 2838.553666 samples/sec	accuracy=0.824456	lr=0.069514
Epoch[88] Batch [699]	Speed: 2888.174283 samples/sec	accuracy=0.824466	lr=0.069349
Epoch[88] Batch [749]	Speed: 2976.046019 samples/sec	accuracy=0.824581	lr=0.069183
Epoch[88] Batch [799]	Speed: 2847.437107 samples/sec	accuracy=0.824759	lr=0.069018
Epoch[88] Batch [849]	Speed: 2773.527347 samples/sec	accuracy=0.824796	lr=0.068853
Epoch[88] Batch [899]	Speed: 2776.092730 samples/sec	accuracy=0.824825	lr=0.068689
Epoch[88] Batch [949]	Speed: 2861.600065 samples/sec	accuracy=0.824919	lr=0.068524
Epoch[88] Batch [999]	Speed: 2856.741786 samples/sec	accuracy=0.824971	lr=0.068359
Epoch[88] Batch [1049]	Speed: 2939.077178 samples/sec	accuracy=0.825024	lr=0.068195
Epoch[88] Batch [1099]	Speed: 2890.486692 samples/sec	accuracy=0.825156	lr=0.068031
Epoch[88] Batch [1149]	Speed: 2875.283136 samples/sec	accuracy=0.825254	lr=0.067867
Epoch[88] Batch [1199]	Speed: 2826.072152 samples/sec	accuracy=0.825354	lr=0.067703
Epoch[88] Batch [1249]	Speed: 3033.874713 samples/sec	accuracy=0.825425	lr=0.067539
[Epoch 88] training: accuracy=0.825427
[Epoch 88] speed: 2888 samples/sec	time cost: 453.713892
[Epoch 88] validation: err-top1=0.253627 err-top5=0.074677
Epoch[89] Batch [49]	Speed: 2958.961337 samples/sec	accuracy=0.824962	lr=0.067373
Epoch[89] Batch [99]	Speed: 2922.418298 samples/sec	accuracy=0.825235	lr=0.067209
Epoch[89] Batch [149]	Speed: 2975.326257 samples/sec	accuracy=0.825161	lr=0.067046
Epoch[89] Batch [199]	Speed: 2900.422906 samples/sec	accuracy=0.825323	lr=0.066883
Epoch[89] Batch [249]	Speed: 2882.371307 samples/sec	accuracy=0.825294	lr=0.066720
Epoch[89] Batch [299]	Speed: 2870.659961 samples/sec	accuracy=0.825463	lr=0.066557
Epoch[89] Batch [349]	Speed: 2908.496813 samples/sec	accuracy=0.825595	lr=0.066395
Epoch[89] Batch [399]	Speed: 2895.432845 samples/sec	accuracy=0.825740	lr=0.066232
Epoch[89] Batch [449]	Speed: 2917.311274 samples/sec	accuracy=0.825804	lr=0.066070
Epoch[89] Batch [499]	Speed: 2938.632683 samples/sec	accuracy=0.825857	lr=0.065908
Epoch[89] Batch [549]	Speed: 2892.995204 samples/sec	accuracy=0.825871	lr=0.065746
Epoch[89] Batch [599]	Speed: 2899.292446 samples/sec	accuracy=0.825888	lr=0.065584
Epoch[89] Batch [649]	Speed: 2891.674812 samples/sec	accuracy=0.825935	lr=0.065423
Epoch[89] Batch [699]	Speed: 2932.762114 samples/sec	accuracy=0.826072	lr=0.065261
Epoch[89] Batch [749]	Speed: 2956.853687 samples/sec	accuracy=0.826071	lr=0.065100
Epoch[89] Batch [799]	Speed: 2935.960945 samples/sec	accuracy=0.826122	lr=0.064939
Epoch[89] Batch [849]	Speed: 2833.112329 samples/sec	accuracy=0.826228	lr=0.064778
Epoch[89] Batch [899]	Speed: 2805.105626 samples/sec	accuracy=0.826248	lr=0.064617
Epoch[89] Batch [949]	Speed: 2943.384178 samples/sec	accuracy=0.826369	lr=0.064456
Epoch[89] Batch [999]	Speed: 2800.961055 samples/sec	accuracy=0.826479	lr=0.064296
Epoch[89] Batch [1049]	Speed: 2947.035311 samples/sec	accuracy=0.826503	lr=0.064135
Epoch[89] Batch [1099]	Speed: 2857.053251 samples/sec	accuracy=0.826584	lr=0.063975
Epoch[89] Batch [1149]	Speed: 2894.491144 samples/sec	accuracy=0.826635	lr=0.063815
Epoch[89] Batch [1199]	Speed: 2826.740705 samples/sec	accuracy=0.826638	lr=0.063655
Epoch[89] Batch [1249]	Speed: 3051.171119 samples/sec	accuracy=0.826751	lr=0.063496
[Epoch 89] training: accuracy=0.826754
[Epoch 89] speed: 2902 samples/sec	time cost: 451.299822
[Epoch 89] validation: err-top1=0.247509 err-top5=0.071767
Epoch[90] Batch [49]	Speed: 2911.817124 samples/sec	accuracy=0.827707	lr=0.063333
Epoch[90] Batch [99]	Speed: 2938.557447 samples/sec	accuracy=0.827065	lr=0.063174
Epoch[90] Batch [149]	Speed: 2857.836107 samples/sec	accuracy=0.827130	lr=0.063015
Epoch[90] Batch [199]	Speed: 2977.468094 samples/sec	accuracy=0.827180	lr=0.062855
Epoch[90] Batch [249]	Speed: 2990.593045 samples/sec	accuracy=0.827348	lr=0.062697
Epoch[90] Batch [299]	Speed: 2903.662372 samples/sec	accuracy=0.827535	lr=0.062538
Epoch[90] Batch [349]	Speed: 2886.144900 samples/sec	accuracy=0.827547	lr=0.062379
Epoch[90] Batch [399]	Speed: 2928.088381 samples/sec	accuracy=0.827562	lr=0.062221
Epoch[90] Batch [449]	Speed: 2947.431299 samples/sec	accuracy=0.827573	lr=0.062063
Epoch[90] Batch [499]	Speed: 2793.300217 samples/sec	accuracy=0.827668	lr=0.061905
Epoch[90] Batch [549]	Speed: 2993.711079 samples/sec	accuracy=0.827624	lr=0.061747
Epoch[90] Batch [599]	Speed: 2825.863452 samples/sec	accuracy=0.827622	lr=0.061589
Epoch[90] Batch [649]	Speed: 2883.212772 samples/sec	accuracy=0.827691	lr=0.061432
Epoch[90] Batch [699]	Speed: 2959.804308 samples/sec	accuracy=0.827792	lr=0.061274
Epoch[90] Batch [749]	Speed: 2869.152628 samples/sec	accuracy=0.827879	lr=0.061117
Epoch[90] Batch [799]	Speed: 2964.819233 samples/sec	accuracy=0.827912	lr=0.060960
Epoch[90] Batch [849]	Speed: 2959.033503 samples/sec	accuracy=0.827971	lr=0.060803
Epoch[90] Batch [899]	Speed: 3014.447022 samples/sec	accuracy=0.828046	lr=0.060646
Epoch[90] Batch [949]	Speed: 2994.274093 samples/sec	accuracy=0.828081	lr=0.060490
Epoch[90] Batch [999]	Speed: 3004.403024 samples/sec	accuracy=0.828127	lr=0.060333
Epoch[90] Batch [1049]	Speed: 2935.397737 samples/sec	accuracy=0.828165	lr=0.060177
Epoch[90] Batch [1099]	Speed: 2863.167952 samples/sec	accuracy=0.828258	lr=0.060021
Epoch[90] Batch [1149]	Speed: 2951.398701 samples/sec	accuracy=0.828330	lr=0.059865
Epoch[90] Batch [1199]	Speed: 2880.757513 samples/sec	accuracy=0.828393	lr=0.059710
Epoch[90] Batch [1249]	Speed: 3040.249727 samples/sec	accuracy=0.828448	lr=0.059554
[Epoch 90] training: accuracy=0.828450
[Epoch 90] speed: 2927 samples/sec	time cost: 447.466577
[Epoch 90] validation: err-top1=0.245396 err-top5=0.070133
Epoch[91] Batch [49]	Speed: 3175.017568 samples/sec	accuracy=0.829500	lr=0.059396
Epoch[91] Batch [99]	Speed: 2941.434300 samples/sec	accuracy=0.828764	lr=0.059240
Epoch[91] Batch [149]	Speed: 3224.621563 samples/sec	accuracy=0.828685	lr=0.059085
Epoch[91] Batch [199]	Speed: 2973.528620 samples/sec	accuracy=0.828782	lr=0.058930
Epoch[91] Batch [249]	Speed: 3135.310136 samples/sec	accuracy=0.828607	lr=0.058776
Epoch[91] Batch [299]	Speed: 2858.083867 samples/sec	accuracy=0.828726	lr=0.058621
Epoch[91] Batch [349]	Speed: 2849.090996 samples/sec	accuracy=0.828799	lr=0.058467
Epoch[91] Batch [399]	Speed: 2915.246373 samples/sec	accuracy=0.828883	lr=0.058313
Epoch[91] Batch [449]	Speed: 2903.097475 samples/sec	accuracy=0.828944	lr=0.058158
Epoch[91] Batch [499]	Speed: 2817.806621 samples/sec	accuracy=0.828939	lr=0.058005
Epoch[91] Batch [549]	Speed: 2926.579577 samples/sec	accuracy=0.828840	lr=0.057851
Epoch[91] Batch [599]	Speed: 2916.217272 samples/sec	accuracy=0.828838	lr=0.057697
Epoch[91] Batch [649]	Speed: 2890.348156 samples/sec	accuracy=0.828862	lr=0.057544
Epoch[91] Batch [699]	Speed: 3030.377494 samples/sec	accuracy=0.828941	lr=0.057391
Epoch[91] Batch [749]	Speed: 3047.803185 samples/sec	accuracy=0.828981	lr=0.057238
Epoch[91] Batch [799]	Speed: 3015.452275 samples/sec	accuracy=0.829081	lr=0.057085
Epoch[91] Batch [849]	Speed: 2949.677457 samples/sec	accuracy=0.829115	lr=0.056932
Epoch[91] Batch [899]	Speed: 2953.911461 samples/sec	accuracy=0.829141	lr=0.056780
Epoch[91] Batch [949]	Speed: 2985.942808 samples/sec	accuracy=0.829250	lr=0.056627
Epoch[91] Batch [999]	Speed: 3150.512734 samples/sec	accuracy=0.829279	lr=0.056475
Epoch[91] Batch [1049]	Speed: 3004.371121 samples/sec	accuracy=0.829326	lr=0.056323
Epoch[91] Batch [1099]	Speed: 2807.372940 samples/sec	accuracy=0.829386	lr=0.056171
Epoch[91] Batch [1149]	Speed: 2846.203721 samples/sec	accuracy=0.829446	lr=0.056020
Epoch[91] Batch [1199]	Speed: 2901.302031 samples/sec	accuracy=0.829525	lr=0.055868
Epoch[91] Batch [1249]	Speed: 3100.086910 samples/sec	accuracy=0.829572	lr=0.055717
[Epoch 91] training: accuracy=0.829575
[Epoch 91] speed: 2966 samples/sec	time cost: 441.453399
[Epoch 91] validation: err-top1=0.245556 err-top5=0.072704
Epoch[92] Batch [49]	Speed: 3108.850266 samples/sec	accuracy=0.830173	lr=0.055563
Epoch[92] Batch [99]	Speed: 3078.280014 samples/sec	accuracy=0.830257	lr=0.055412
Epoch[92] Batch [149]	Speed: 3014.330282 samples/sec	accuracy=0.829913	lr=0.055261
Epoch[92] Batch [199]	Speed: 3157.598814 samples/sec	accuracy=0.829906	lr=0.055111
Epoch[92] Batch [249]	Speed: 3045.347934 samples/sec	accuracy=0.829866	lr=0.054960
Epoch[92] Batch [299]	Speed: 2937.535775 samples/sec	accuracy=0.829728	lr=0.054810
Epoch[92] Batch [349]	Speed: 3034.176744 samples/sec	accuracy=0.829778	lr=0.054660
Epoch[92] Batch [399]	Speed: 2885.209579 samples/sec	accuracy=0.829841	lr=0.054510
Epoch[92] Batch [449]	Speed: 2919.959127 samples/sec	accuracy=0.829900	lr=0.054360
Epoch[92] Batch [499]	Speed: 2990.780510 samples/sec	accuracy=0.830090	lr=0.054210
Epoch[92] Batch [549]	Speed: 2912.643320 samples/sec	accuracy=0.830082	lr=0.054061
Epoch[92] Batch [599]	Speed: 3036.000236 samples/sec	accuracy=0.830162	lr=0.053912
Epoch[92] Batch [649]	Speed: 2932.689582 samples/sec	accuracy=0.830240	lr=0.053763
Epoch[92] Batch [699]	Speed: 2870.029006 samples/sec	accuracy=0.830278	lr=0.053614
Epoch[92] Batch [749]	Speed: 2815.231875 samples/sec	accuracy=0.830337	lr=0.053465
Epoch[92] Batch [799]	Speed: 2966.069424 samples/sec	accuracy=0.830451	lr=0.053317
Epoch[92] Batch [849]	Speed: 2875.548292 samples/sec	accuracy=0.830460	lr=0.053168
Epoch[92] Batch [899]	Speed: 2827.463329 samples/sec	accuracy=0.830523	lr=0.053020
Epoch[92] Batch [949]	Speed: 2853.697272 samples/sec	accuracy=0.830554	lr=0.052872
Epoch[92] Batch [999]	Speed: 3040.173675 samples/sec	accuracy=0.830574	lr=0.052724
Epoch[92] Batch [1049]	Speed: 2907.935665 samples/sec	accuracy=0.830664	lr=0.052577
Epoch[92] Batch [1099]	Speed: 2919.865074 samples/sec	accuracy=0.830737	lr=0.052429
Epoch[92] Batch [1149]	Speed: 2870.083243 samples/sec	accuracy=0.830770	lr=0.052282
Epoch[92] Batch [1199]	Speed: 2914.594204 samples/sec	accuracy=0.830757	lr=0.052135
Epoch[92] Batch [1249]	Speed: 3068.438572 samples/sec	accuracy=0.830771	lr=0.051988
[Epoch 92] training: accuracy=0.830770
[Epoch 92] speed: 2954 samples/sec	time cost: 443.523231
[Epoch 92] validation: err-top1=0.243084 err-top5=0.069874
Epoch[93] Batch [49]	Speed: 2819.697984 samples/sec	accuracy=0.831228	lr=0.051838
Epoch[93] Batch [99]	Speed: 3055.306937 samples/sec	accuracy=0.831359	lr=0.051691
Epoch[93] Batch [149]	Speed: 2828.471587 samples/sec	accuracy=0.831561	lr=0.051545
Epoch[93] Batch [199]	Speed: 2891.070043 samples/sec	accuracy=0.831473	lr=0.051399
Epoch[93] Batch [249]	Speed: 2890.687731 samples/sec	accuracy=0.831180	lr=0.051253
Epoch[93] Batch [299]	Speed: 2965.196800 samples/sec	accuracy=0.831133	lr=0.051107
Epoch[93] Batch [349]	Speed: 2873.920820 samples/sec	accuracy=0.831255	lr=0.050961
Epoch[93] Batch [399]	Speed: 2954.129141 samples/sec	accuracy=0.831398	lr=0.050816
Epoch[93] Batch [449]	Speed: 2979.559522 samples/sec	accuracy=0.831368	lr=0.050670
Epoch[93] Batch [499]	Speed: 2925.014597 samples/sec	accuracy=0.831311	lr=0.050525
Epoch[93] Batch [549]	Speed: 3033.616196 samples/sec	accuracy=0.831267	lr=0.050380
Epoch[93] Batch [599]	Speed: 3044.176056 samples/sec	accuracy=0.831332	lr=0.050235
Epoch[93] Batch [649]	Speed: 3025.102737 samples/sec	accuracy=0.831482	lr=0.050091
Epoch[93] Batch [699]	Speed: 2835.505842 samples/sec	accuracy=0.831530	lr=0.049946
Epoch[93] Batch [749]	Speed: 2988.647185 samples/sec	accuracy=0.831537	lr=0.049802
Epoch[93] Batch [799]	Speed: 2958.125691 samples/sec	accuracy=0.831542	lr=0.049658
Epoch[93] Batch [849]	Speed: 3014.390068 samples/sec	accuracy=0.831610	lr=0.049514
Epoch[93] Batch [899]	Speed: 2889.472701 samples/sec	accuracy=0.831686	lr=0.049370
Epoch[93] Batch [949]	Speed: 2869.620909 samples/sec	accuracy=0.831780	lr=0.049227
Epoch[93] Batch [999]	Speed: 2970.947246 samples/sec	accuracy=0.831855	lr=0.049083
Epoch[93] Batch [1049]	Speed: 2960.183823 samples/sec	accuracy=0.831906	lr=0.048940
Epoch[93] Batch [1099]	Speed: 2956.172842 samples/sec	accuracy=0.831921	lr=0.048797
Epoch[93] Batch [1149]	Speed: 3088.852098 samples/sec	accuracy=0.831937	lr=0.048654
Epoch[93] Batch [1199]	Speed: 2948.202870 samples/sec	accuracy=0.831982	lr=0.048511
Epoch[93] Batch [1249]	Speed: 3108.373141 samples/sec	accuracy=0.832045	lr=0.048369
[Epoch 93] training: accuracy=0.832042
[Epoch 93] speed: 2951 samples/sec	time cost: 444.157234
[Epoch 93] validation: err-top1=0.242228 err-top5=0.068848
Epoch[94] Batch [49]	Speed: 2945.209562 samples/sec	accuracy=0.832851	lr=0.048224
Epoch[94] Batch [99]	Speed: 2849.691033 samples/sec	accuracy=0.832300	lr=0.048082
Epoch[94] Batch [149]	Speed: 2929.672559 samples/sec	accuracy=0.832217	lr=0.047940
Epoch[94] Batch [199]	Speed: 2782.164149 samples/sec	accuracy=0.832027	lr=0.047798
Epoch[94] Batch [249]	Speed: 2828.361580 samples/sec	accuracy=0.831960	lr=0.047656
Epoch[94] Batch [299]	Speed: 2865.069117 samples/sec	accuracy=0.832214	lr=0.047515
Epoch[94] Batch [349]	Speed: 2790.326522 samples/sec	accuracy=0.832130	lr=0.047374
Epoch[94] Batch [399]	Speed: 2928.902945 samples/sec	accuracy=0.832239	lr=0.047233
Epoch[94] Batch [449]	Speed: 2981.680054 samples/sec	accuracy=0.832233	lr=0.047092
Epoch[94] Batch [499]	Speed: 2946.326277 samples/sec	accuracy=0.832306	lr=0.046951
Epoch[94] Batch [549]	Speed: 2965.441903 samples/sec	accuracy=0.832375	lr=0.046811
Epoch[94] Batch [599]	Speed: 2824.689030 samples/sec	accuracy=0.832391	lr=0.046670
Epoch[94] Batch [649]	Speed: 3150.328603 samples/sec	accuracy=0.832459	lr=0.046530
Epoch[94] Batch [699]	Speed: 2958.176137 samples/sec	accuracy=0.832547	lr=0.046390
Epoch[94] Batch [749]	Speed: 2929.732512 samples/sec	accuracy=0.832693	lr=0.046251
Epoch[94] Batch [799]	Speed: 2878.884685 samples/sec	accuracy=0.832708	lr=0.046111
Epoch[94] Batch [849]	Speed: 2964.857996 samples/sec	accuracy=0.832694	lr=0.045972
Epoch[94] Batch [899]	Speed: 2988.642443 samples/sec	accuracy=0.832720	lr=0.045833
Epoch[94] Batch [949]	Speed: 2926.730423 samples/sec	accuracy=0.832763	lr=0.045693
Epoch[94] Batch [999]	Speed: 2843.913425 samples/sec	accuracy=0.832778	lr=0.045555
Epoch[94] Batch [1049]	Speed: 2785.986404 samples/sec	accuracy=0.832853	lr=0.045416
Epoch[94] Batch [1099]	Speed: 2988.514759 samples/sec	accuracy=0.832945	lr=0.045278
Epoch[94] Batch [1149]	Speed: 2986.446171 samples/sec	accuracy=0.832996	lr=0.045139
Epoch[94] Batch [1199]	Speed: 3002.251990 samples/sec	accuracy=0.833095	lr=0.045001
Epoch[94] Batch [1249]	Speed: 3044.376255 samples/sec	accuracy=0.833121	lr=0.044863
[Epoch 94] training: accuracy=0.833123
[Epoch 94] speed: 2918 samples/sec	time cost: 448.869593
[Epoch 94] validation: err-top1=0.237643 err-top5=0.066346
Epoch[95] Batch [49]	Speed: 3050.671186 samples/sec	accuracy=0.834354	lr=0.044723
Epoch[95] Batch [99]	Speed: 2952.129780 samples/sec	accuracy=0.833378	lr=0.044585
Epoch[95] Batch [149]	Speed: 2985.519886 samples/sec	accuracy=0.833206	lr=0.044448
Epoch[95] Batch [199]	Speed: 2864.140031 samples/sec	accuracy=0.833333	lr=0.044311
Epoch[95] Batch [249]	Speed: 2966.916451 samples/sec	accuracy=0.833420	lr=0.044174
Epoch[95] Batch [299]	Speed: 2900.509325 samples/sec	accuracy=0.833446	lr=0.044037
Epoch[95] Batch [349]	Speed: 2841.075230 samples/sec	accuracy=0.833483	lr=0.043900
Epoch[95] Batch [399]	Speed: 2875.352549 samples/sec	accuracy=0.833548	lr=0.043764
Epoch[95] Batch [449]	Speed: 2868.063721 samples/sec	accuracy=0.833645	lr=0.043628
Epoch[95] Batch [499]	Speed: 2895.996207 samples/sec	accuracy=0.833723	lr=0.043492
Epoch[95] Batch [549]	Speed: 2990.656475 samples/sec	accuracy=0.833754	lr=0.043356
Epoch[95] Batch [599]	Speed: 2902.335247 samples/sec	accuracy=0.833800	lr=0.043220
Epoch[95] Batch [649]	Speed: 2922.678019 samples/sec	accuracy=0.833779	lr=0.043085
Epoch[95] Batch [699]	Speed: 3195.781982 samples/sec	accuracy=0.833912	lr=0.042949
Epoch[95] Batch [749]	Speed: 2838.042133 samples/sec	accuracy=0.834037	lr=0.042814
Epoch[95] Batch [799]	Speed: 2978.483491 samples/sec	accuracy=0.834133	lr=0.042679
Epoch[95] Batch [849]	Speed: 2960.731969 samples/sec	accuracy=0.834270	lr=0.042545
Epoch[95] Batch [899]	Speed: 3000.720315 samples/sec	accuracy=0.834303	lr=0.042410
Epoch[95] Batch [949]	Speed: 3044.190900 samples/sec	accuracy=0.834388	lr=0.042276
Epoch[95] Batch [999]	Speed: 2976.320062 samples/sec	accuracy=0.834469	lr=0.042141
Epoch[95] Batch [1049]	Speed: 2828.653361 samples/sec	accuracy=0.834534	lr=0.042007
Epoch[95] Batch [1099]	Speed: 2948.595042 samples/sec	accuracy=0.834604	lr=0.041874
Epoch[95] Batch [1149]	Speed: 2807.524998 samples/sec	accuracy=0.834670	lr=0.041740
Epoch[95] Batch [1199]	Speed: 2851.806204 samples/sec	accuracy=0.834728	lr=0.041607
Epoch[95] Batch [1249]	Speed: 3117.771711 samples/sec	accuracy=0.834715	lr=0.041473
[Epoch 95] training: accuracy=0.834715
[Epoch 95] speed: 2937 samples/sec	time cost: 446.006193
[Epoch 95] validation: err-top1=0.234594 err-top5=0.064493
Epoch[96] Batch [49]	Speed: 2969.171234 samples/sec	accuracy=0.835991	lr=0.041338
Epoch[96] Batch [99]	Speed: 2883.821268 samples/sec	accuracy=0.835635	lr=0.041205
Epoch[96] Batch [149]	Speed: 2877.693784 samples/sec	accuracy=0.835332	lr=0.041072
Epoch[96] Batch [199]	Speed: 2927.094042 samples/sec	accuracy=0.835035	lr=0.040940
Epoch[96] Batch [249]	Speed: 2903.067727 samples/sec	accuracy=0.834986	lr=0.040807
Epoch[96] Batch [299]	Speed: 2921.052799 samples/sec	accuracy=0.835171	lr=0.040675
Epoch[96] Batch [349]	Speed: 2874.002744 samples/sec	accuracy=0.835173	lr=0.040543
Epoch[96] Batch [399]	Speed: 2956.404124 samples/sec	accuracy=0.835281	lr=0.040412
Epoch[96] Batch [449]	Speed: 3015.187489 samples/sec	accuracy=0.835457	lr=0.040280
Epoch[96] Batch [499]	Speed: 3073.005344 samples/sec	accuracy=0.835429	lr=0.040149
Epoch[96] Batch [549]	Speed: 2844.162241 samples/sec	accuracy=0.835502	lr=0.040018
Epoch[96] Batch [599]	Speed: 2947.959920 samples/sec	accuracy=0.835539	lr=0.039887
Epoch[96] Batch [649]	Speed: 2917.828630 samples/sec	accuracy=0.835603	lr=0.039756
Epoch[96] Batch [699]	Speed: 3019.452577 samples/sec	accuracy=0.835763	lr=0.039625
Epoch[96] Batch [749]	Speed: 2882.162990 samples/sec	accuracy=0.835728	lr=0.039495
Epoch[96] Batch [799]	Speed: 2922.506232 samples/sec	accuracy=0.835691	lr=0.039365
Epoch[96] Batch [849]	Speed: 2860.896514 samples/sec	accuracy=0.835781	lr=0.039235
Epoch[96] Batch [899]	Speed: 2966.279435 samples/sec	accuracy=0.835803	lr=0.039105
Epoch[96] Batch [949]	Speed: 2878.490618 samples/sec	accuracy=0.835828	lr=0.038975
Epoch[96] Batch [999]	Speed: 2828.002450 samples/sec	accuracy=0.835932	lr=0.038846
Epoch[96] Batch [1049]	Speed: 2913.688941 samples/sec	accuracy=0.835969	lr=0.038717
Epoch[96] Batch [1099]	Speed: 2993.213984 samples/sec	accuracy=0.836036	lr=0.038588
Epoch[96] Batch [1149]	Speed: 2827.019796 samples/sec	accuracy=0.836073	lr=0.038459
Epoch[96] Batch [1199]	Speed: 2873.828209 samples/sec	accuracy=0.836106	lr=0.038330
Epoch[96] Batch [1249]	Speed: 3028.542126 samples/sec	accuracy=0.836176	lr=0.038202
[Epoch 96] training: accuracy=0.836179
[Epoch 96] speed: 2920 samples/sec	time cost: 448.808131
[Epoch 96] validation: err-top1=0.231346 err-top5=0.064912
Epoch[97] Batch [49]	Speed: 3011.906424 samples/sec	accuracy=0.837056	lr=0.038071
Epoch[97] Batch [99]	Speed: 2902.871319 samples/sec	accuracy=0.836962	lr=0.037943
Epoch[97] Batch [149]	Speed: 2913.431922 samples/sec	accuracy=0.836795	lr=0.037815
Epoch[97] Batch [199]	Speed: 2925.721819 samples/sec	accuracy=0.836723	lr=0.037687
Epoch[97] Batch [249]	Speed: 2868.933838 samples/sec	accuracy=0.836655	lr=0.037560
Epoch[97] Batch [299]	Speed: 2808.785808 samples/sec	accuracy=0.836518	lr=0.037432
Epoch[97] Batch [349]	Speed: 3003.634023 samples/sec	accuracy=0.836527	lr=0.037305
Epoch[97] Batch [399]	Speed: 2865.119115 samples/sec	accuracy=0.836674	lr=0.037178
Epoch[97] Batch [449]	Speed: 2797.858102 samples/sec	accuracy=0.836679	lr=0.037052
Epoch[97] Batch [499]	Speed: 2862.318802 samples/sec	accuracy=0.836656	lr=0.036925
Epoch[97] Batch [549]	Speed: 2850.317201 samples/sec	accuracy=0.836688	lr=0.036799
Epoch[97] Batch [599]	Speed: 3020.371535 samples/sec	accuracy=0.836671	lr=0.036673
Epoch[97] Batch [649]	Speed: 2906.061202 samples/sec	accuracy=0.836749	lr=0.036547
Epoch[97] Batch [699]	Speed: 2979.775210 samples/sec	accuracy=0.836768	lr=0.036421
Epoch[97] Batch [749]	Speed: 3055.670207 samples/sec	accuracy=0.836765	lr=0.036296
Epoch[97] Batch [799]	Speed: 2806.174158 samples/sec	accuracy=0.836863	lr=0.036170
Epoch[97] Batch [849]	Speed: 2866.123810 samples/sec	accuracy=0.836891	lr=0.036045
Epoch[97] Batch [899]	Speed: 2859.472898 samples/sec	accuracy=0.836942	lr=0.035920
Epoch[97] Batch [949]	Speed: 3089.398757 samples/sec	accuracy=0.837009	lr=0.035795
Epoch[97] Batch [999]	Speed: 2909.064128 samples/sec	accuracy=0.837099	lr=0.035671
Epoch[97] Batch [1049]	Speed: 2853.840547 samples/sec	accuracy=0.837160	lr=0.035546
Epoch[97] Batch [1099]	Speed: 2906.398186 samples/sec	accuracy=0.837218	lr=0.035422
Epoch[97] Batch [1149]	Speed: 2905.294705 samples/sec	accuracy=0.837276	lr=0.035298
Epoch[97] Batch [1199]	Speed: 2885.782620 samples/sec	accuracy=0.837336	lr=0.035174
Epoch[97] Batch [1249]	Speed: 3009.670693 samples/sec	accuracy=0.837385	lr=0.035051
[Epoch 97] training: accuracy=0.837389
[Epoch 97] speed: 2910 samples/sec	time cost: 450.104507
[Epoch 97] validation: err-top1=0.231146 err-top5=0.063178
Epoch[98] Batch [49]	Speed: 3028.708408 samples/sec	accuracy=0.838152	lr=0.034925
Epoch[98] Batch [99]	Speed: 2887.304065 samples/sec	accuracy=0.837794	lr=0.034802
Epoch[98] Batch [149]	Speed: 2885.287651 samples/sec	accuracy=0.837960	lr=0.034679
Epoch[98] Batch [199]	Speed: 2924.011515 samples/sec	accuracy=0.837900	lr=0.034556
Epoch[98] Batch [249]	Speed: 2891.508403 samples/sec	accuracy=0.837663	lr=0.034433
Epoch[98] Batch [299]	Speed: 2860.057357 samples/sec	accuracy=0.837721	lr=0.034311
Epoch[98] Batch [349]	Speed: 2853.153695 samples/sec	accuracy=0.837696	lr=0.034189
Epoch[98] Batch [399]	Speed: 2955.481532 samples/sec	accuracy=0.837724	lr=0.034067
Epoch[98] Batch [449]	Speed: 2934.579235 samples/sec	accuracy=0.837836	lr=0.033945
Epoch[98] Batch [499]	Speed: 2829.283400 samples/sec	accuracy=0.837909	lr=0.033823
Epoch[98] Batch [549]	Speed: 2964.077070 samples/sec	accuracy=0.837996	lr=0.033702
Epoch[98] Batch [599]	Speed: 2936.913480 samples/sec	accuracy=0.838075	lr=0.033581
Epoch[98] Batch [649]	Speed: 2931.358084 samples/sec	accuracy=0.838103	lr=0.033460
Epoch[98] Batch [699]	Speed: 2998.635807 samples/sec	accuracy=0.838187	lr=0.033339
Epoch[98] Batch [749]	Speed: 2897.194960 samples/sec	accuracy=0.838215	lr=0.033218
Epoch[98] Batch [799]	Speed: 2963.697702 samples/sec	accuracy=0.838218	lr=0.033098
Epoch[98] Batch [849]	Speed: 2989.723040 samples/sec	accuracy=0.838206	lr=0.032978
Epoch[98] Batch [899]	Speed: 3005.034991 samples/sec	accuracy=0.838292	lr=0.032858
Epoch[98] Batch [949]	Speed: 3083.446860 samples/sec	accuracy=0.838363	lr=0.032738
Epoch[98] Batch [999]	Speed: 2986.647696 samples/sec	accuracy=0.838419	lr=0.032618
Epoch[98] Batch [1049]	Speed: 2968.039887 samples/sec	accuracy=0.838467	lr=0.032499
Epoch[98] Batch [1099]	Speed: 2976.340151 samples/sec	accuracy=0.838491	lr=0.032380
Epoch[98] Batch [1149]	Speed: 2964.192568 samples/sec	accuracy=0.838519	lr=0.032261
Epoch[98] Batch [1199]	Speed: 3016.985000 samples/sec	accuracy=0.838585	lr=0.032142
Epoch[98] Batch [1249]	Speed: 3043.897874 samples/sec	accuracy=0.838606	lr=0.032023
[Epoch 98] training: accuracy=0.838607
[Epoch 98] speed: 2947 samples/sec	time cost: 444.078868
[Epoch 98] validation: err-top1=0.232361 err-top5=0.065776
Epoch[99] Batch [49]	Speed: 2978.034265 samples/sec	accuracy=0.840222	lr=0.031902
Epoch[99] Batch [99]	Speed: 2893.589316 samples/sec	accuracy=0.839518	lr=0.031784
Epoch[99] Batch [149]	Speed: 2951.418779 samples/sec	accuracy=0.839535	lr=0.031666
Epoch[99] Batch [199]	Speed: 2924.078801 samples/sec	accuracy=0.839135	lr=0.031548
Epoch[99] Batch [249]	Speed: 2875.820121 samples/sec	accuracy=0.839254	lr=0.031431
Epoch[99] Batch [299]	Speed: 2926.735289 samples/sec	accuracy=0.839366	lr=0.031313
Epoch[99] Batch [349]	Speed: 2911.661771 samples/sec	accuracy=0.839338	lr=0.031196
Epoch[99] Batch [399]	Speed: 3025.709467 samples/sec	accuracy=0.839282	lr=0.031079
Epoch[99] Batch [449]	Speed: 2777.847533 samples/sec	accuracy=0.839332	lr=0.030962
Epoch[99] Batch [499]	Speed: 2833.501957 samples/sec	accuracy=0.839290	lr=0.030846
Epoch[99] Batch [549]	Speed: 2884.373456 samples/sec	accuracy=0.839331	lr=0.030729
Epoch[99] Batch [599]	Speed: 2921.687549 samples/sec	accuracy=0.839363	lr=0.030613
Epoch[99] Batch [649]	Speed: 2935.325355 samples/sec	accuracy=0.839345	lr=0.030497
Epoch[99] Batch [699]	Speed: 2885.073525 samples/sec	accuracy=0.839458	lr=0.030381
Epoch[99] Batch [749]	Speed: 2885.564698 samples/sec	accuracy=0.839539	lr=0.030266
Epoch[99] Batch [799]	Speed: 2983.031979 samples/sec	accuracy=0.839514	lr=0.030150
Epoch[99] Batch [849]	Speed: 2860.142493 samples/sec	accuracy=0.839593	lr=0.030035
Epoch[99] Batch [899]	Speed: 2876.431472 samples/sec	accuracy=0.839648	lr=0.029920
Epoch[99] Batch [949]	Speed: 2886.071359 samples/sec	accuracy=0.839715	lr=0.029805
Epoch[99] Batch [999]	Speed: 2772.042690 samples/sec	accuracy=0.839755	lr=0.029691
Epoch[99] Batch [1049]	Speed: 3065.420771 samples/sec	accuracy=0.839821	lr=0.029576
Epoch[99] Batch [1099]	Speed: 2909.752066 samples/sec	accuracy=0.839866	lr=0.029462
Epoch[99] Batch [1149]	Speed: 2872.872296 samples/sec	accuracy=0.839896	lr=0.029348
Epoch[99] Batch [1199]	Speed: 2936.946015 samples/sec	accuracy=0.839955	lr=0.029234
Epoch[99] Batch [1249]	Speed: 3203.633137 samples/sec	accuracy=0.839988	lr=0.029121
[Epoch 99] training: accuracy=0.839987
[Epoch 99] speed: 2914 samples/sec	time cost: 449.797639
[Epoch 99] validation: err-top1=0.228815 err-top5=0.063257
Epoch[100] Batch [49]	Speed: 3038.959285 samples/sec	accuracy=0.840297	lr=0.029005
Epoch[100] Batch [99]	Speed: 2966.837178 samples/sec	accuracy=0.840432	lr=0.028892
Epoch[100] Batch [149]	Speed: 2906.693702 samples/sec	accuracy=0.840272	lr=0.028779
Epoch[100] Batch [199]	Speed: 2851.195622 samples/sec	accuracy=0.840193	lr=0.028666
Epoch[100] Batch [249]	Speed: 2830.443179 samples/sec	accuracy=0.840144	lr=0.028554
Epoch[100] Batch [299]	Speed: 2833.616851 samples/sec	accuracy=0.840144	lr=0.028441
Epoch[100] Batch [349]	Speed: 2920.841317 samples/sec	accuracy=0.840081	lr=0.028329
Epoch[100] Batch [399]	Speed: 2927.606852 samples/sec	accuracy=0.840167	lr=0.028217
Epoch[100] Batch [449]	Speed: 2915.991047 samples/sec	accuracy=0.840193	lr=0.028105
Epoch[100] Batch [499]	Speed: 2852.191142 samples/sec	accuracy=0.840313	lr=0.027994
Epoch[100] Batch [549]	Speed: 2960.675149 samples/sec	accuracy=0.840433	lr=0.027883
Epoch[100] Batch [599]	Speed: 2909.076502 samples/sec	accuracy=0.840496	lr=0.027772
Epoch[100] Batch [649]	Speed: 2848.854619 samples/sec	accuracy=0.840516	lr=0.027661
Epoch[100] Batch [699]	Speed: 2766.946566 samples/sec	accuracy=0.840567	lr=0.027550
Epoch[100] Batch [749]	Speed: 2991.059357 samples/sec	accuracy=0.840592	lr=0.027439
Epoch[100] Batch [799]	Speed: 2858.780250 samples/sec	accuracy=0.840697	lr=0.027329
Epoch[100] Batch [849]	Speed: 2785.920155 samples/sec	accuracy=0.840769	lr=0.027219
Epoch[100] Batch [899]	Speed: 3008.757896 samples/sec	accuracy=0.840873	lr=0.027109
Epoch[100] Batch [949]	Speed: 2988.598855 samples/sec	accuracy=0.840942	lr=0.026999
Epoch[100] Batch [999]	Speed: 2917.601045 samples/sec	accuracy=0.840967	lr=0.026890
Epoch[100] Batch [1049]	Speed: 2829.053578 samples/sec	accuracy=0.841022	lr=0.026781
Epoch[100] Batch [1099]	Speed: 2872.626193 samples/sec	accuracy=0.841059	lr=0.026672
Epoch[100] Batch [1149]	Speed: 3054.502798 samples/sec	accuracy=0.841103	lr=0.026563
Epoch[100] Batch [1199]	Speed: 2941.414075 samples/sec	accuracy=0.841178	lr=0.026454
Epoch[100] Batch [1249]	Speed: 3001.289492 samples/sec	accuracy=0.841245	lr=0.026346
[Epoch 100] training: accuracy=0.841247
[Epoch 100] speed: 2907 samples/sec	time cost: 450.886369
[Epoch 100] validation: err-top1=0.224649 err-top5=0.061962
Epoch[101] Batch [49]	Speed: 2969.490494 samples/sec	accuracy=0.842129	lr=0.026235
Epoch[101] Batch [99]	Speed: 2938.963629 samples/sec	accuracy=0.841794	lr=0.026127
Epoch[101] Batch [149]	Speed: 2887.549156 samples/sec	accuracy=0.841754	lr=0.026020
Epoch[101] Batch [199]	Speed: 3089.066792 samples/sec	accuracy=0.841855	lr=0.025912
Epoch[101] Batch [249]	Speed: 2854.710934 samples/sec	accuracy=0.841765	lr=0.025805
Epoch[101] Batch [299]	Speed: 3029.002405 samples/sec	accuracy=0.841823	lr=0.025697
Epoch[101] Batch [349]	Speed: 2925.399348 samples/sec	accuracy=0.841799	lr=0.025590
Epoch[101] Batch [399]	Speed: 2969.507575 samples/sec	accuracy=0.841706	lr=0.025484
Epoch[101] Batch [449]	Speed: 2846.739408 samples/sec	accuracy=0.841699	lr=0.025377
Epoch[101] Batch [499]	Speed: 2911.005049 samples/sec	accuracy=0.841769	lr=0.025271
Epoch[101] Batch [549]	Speed: 2875.114720 samples/sec	accuracy=0.841781	lr=0.025165
Epoch[101] Batch [599]	Speed: 2875.497429 samples/sec	accuracy=0.841878	lr=0.025059
Epoch[101] Batch [649]	Speed: 2984.165706 samples/sec	accuracy=0.841991	lr=0.024953
Epoch[101] Batch [699]	Speed: 2873.577405 samples/sec	accuracy=0.842033	lr=0.024847
Epoch[101] Batch [749]	Speed: 3086.347192 samples/sec	accuracy=0.842014	lr=0.024742
Epoch[101] Batch [799]	Speed: 2990.155564 samples/sec	accuracy=0.842041	lr=0.024637
Epoch[101] Batch [849]	Speed: 2981.512728 samples/sec	accuracy=0.842081	lr=0.024532
Epoch[101] Batch [899]	Speed: 2875.062756 samples/sec	accuracy=0.842140	lr=0.024427
Epoch[101] Batch [949]	Speed: 2870.256825 samples/sec	accuracy=0.842211	lr=0.024323
Epoch[101] Batch [999]	Speed: 2892.815043 samples/sec	accuracy=0.842313	lr=0.024219
Epoch[101] Batch [1049]	Speed: 2883.639188 samples/sec	accuracy=0.842378	lr=0.024115
Epoch[101] Batch [1099]	Speed: 2954.755826 samples/sec	accuracy=0.842391	lr=0.024011
Epoch[101] Batch [1149]	Speed: 3013.628550 samples/sec	accuracy=0.842487	lr=0.023907
Epoch[101] Batch [1199]	Speed: 2887.077568 samples/sec	accuracy=0.842549	lr=0.023804
Epoch[101] Batch [1249]	Speed: 3054.223812 samples/sec	accuracy=0.842609	lr=0.023700
[Epoch 101] training: accuracy=0.842610
[Epoch 101] speed: 2936 samples/sec	time cost: 445.790697
[Epoch 101] validation: err-top1=0.223075 err-top5=0.061145
Epoch[102] Batch [49]	Speed: 3048.024757 samples/sec	accuracy=0.843885	lr=0.023595
Epoch[102] Batch [99]	Speed: 2873.015466 samples/sec	accuracy=0.843438	lr=0.023493
Epoch[102] Batch [149]	Speed: 2912.712296 samples/sec	accuracy=0.843470	lr=0.023390
Epoch[102] Batch [199]	Speed: 2930.277154 samples/sec	accuracy=0.843535	lr=0.023288
Epoch[102] Batch [249]	Speed: 3056.072705 samples/sec	accuracy=0.843514	lr=0.023185
Epoch[102] Batch [299]	Speed: 2902.715977 samples/sec	accuracy=0.843528	lr=0.023083
Epoch[102] Batch [349]	Speed: 2828.149748 samples/sec	accuracy=0.843445	lr=0.022982
Epoch[102] Batch [399]	Speed: 3009.465669 samples/sec	accuracy=0.843509	lr=0.022880
Epoch[102] Batch [449]	Speed: 2939.909179 samples/sec	accuracy=0.843424	lr=0.022779
Epoch[102] Batch [499]	Speed: 2854.265185 samples/sec	accuracy=0.843512	lr=0.022678
Epoch[102] Batch [549]	Speed: 2890.388614 samples/sec	accuracy=0.843573	lr=0.022577
Epoch[102] Batch [599]	Speed: 3007.877840 samples/sec	accuracy=0.843675	lr=0.022476
Epoch[102] Batch [649]	Speed: 3049.788576 samples/sec	accuracy=0.843681	lr=0.022376
Epoch[102] Batch [699]	Speed: 2864.263879 samples/sec	accuracy=0.843673	lr=0.022276
Epoch[102] Batch [749]	Speed: 2926.640280 samples/sec	accuracy=0.843676	lr=0.022175
Epoch[102] Batch [799]	Speed: 2949.558874 samples/sec	accuracy=0.843744	lr=0.022076
Epoch[102] Batch [849]	Speed: 2869.002944 samples/sec	accuracy=0.843814	lr=0.021976
Epoch[102] Batch [899]	Speed: 2844.413814 samples/sec	accuracy=0.843876	lr=0.021877
Epoch[102] Batch [949]	Speed: 2802.222885 samples/sec	accuracy=0.843926	lr=0.021777
Epoch[102] Batch [999]	Speed: 3067.157212 samples/sec	accuracy=0.843962	lr=0.021678
Epoch[102] Batch [1049]	Speed: 2932.923732 samples/sec	accuracy=0.844028	lr=0.021580
Epoch[102] Batch [1099]	Speed: 2969.742879 samples/sec	accuracy=0.844091	lr=0.021481
Epoch[102] Batch [1149]	Speed: 2888.810988 samples/sec	accuracy=0.844139	lr=0.021383
Epoch[102] Batch [1199]	Speed: 2877.745882 samples/sec	accuracy=0.844159	lr=0.021285
Epoch[102] Batch [1249]	Speed: 3050.574287 samples/sec	accuracy=0.844197	lr=0.021187
[Epoch 102] training: accuracy=0.844200
[Epoch 102] speed: 2929 samples/sec	time cost: 447.198718
[Epoch 102] validation: err-top1=0.217435 err-top5=0.059371
Epoch[103] Batch [49]	Speed: 2904.950550 samples/sec	accuracy=0.845361	lr=0.021087
Epoch[103] Batch [99]	Speed: 3074.412939 samples/sec	accuracy=0.845485	lr=0.020990
Epoch[103] Batch [149]	Speed: 2916.918345 samples/sec	accuracy=0.845142	lr=0.020892
Epoch[103] Batch [199]	Speed: 2917.340363 samples/sec	accuracy=0.845176	lr=0.020795
Epoch[103] Batch [249]	Speed: 2924.690330 samples/sec	accuracy=0.845085	lr=0.020698
Epoch[103] Batch [299]	Speed: 2964.507484 samples/sec	accuracy=0.845108	lr=0.020602
Epoch[103] Batch [349]	Speed: 2878.184453 samples/sec	accuracy=0.845122	lr=0.020505
Epoch[103] Batch [399]	Speed: 2934.603818 samples/sec	accuracy=0.845096	lr=0.020409
Epoch[103] Batch [449]	Speed: 2929.065457 samples/sec	accuracy=0.845147	lr=0.020313
Epoch[103] Batch [499]	Speed: 2932.027344 samples/sec	accuracy=0.845067	lr=0.020217
Epoch[103] Batch [549]	Speed: 2905.485466 samples/sec	accuracy=0.845113	lr=0.020122
Epoch[103] Batch [599]	Speed: 2856.108159 samples/sec	accuracy=0.845151	lr=0.020026
Epoch[103] Batch [649]	Speed: 2777.495691 samples/sec	accuracy=0.845172	lr=0.019931
Epoch[103] Batch [699]	Speed: 2816.399072 samples/sec	accuracy=0.845123	lr=0.019836
Epoch[103] Batch [749]	Speed: 2896.907430 samples/sec	accuracy=0.845144	lr=0.019742
Epoch[103] Batch [799]	Speed: 2969.176776 samples/sec	accuracy=0.845215	lr=0.019647
Epoch[103] Batch [849]	Speed: 2863.855663 samples/sec	accuracy=0.845259	lr=0.019553
Epoch[103] Batch [899]	Speed: 3016.707147 samples/sec	accuracy=0.845310	lr=0.019459
Epoch[103] Batch [949]	Speed: 2812.740482 samples/sec	accuracy=0.845367	lr=0.019365
Epoch[103] Batch [999]	Speed: 2944.473630 samples/sec	accuracy=0.845397	lr=0.019271
Epoch[103] Batch [1049]	Speed: 3028.659286 samples/sec	accuracy=0.845414	lr=0.019178
Epoch[103] Batch [1099]	Speed: 3102.360350 samples/sec	accuracy=0.845464	lr=0.019085
Epoch[103] Batch [1149]	Speed: 2826.703720 samples/sec	accuracy=0.845513	lr=0.018992
Epoch[103] Batch [1199]	Speed: 2984.673364 samples/sec	accuracy=0.845569	lr=0.018899
Epoch[103] Batch [1249]	Speed: 2935.718884 samples/sec	accuracy=0.845600	lr=0.018806
[Epoch 103] training: accuracy=0.845601
[Epoch 103] speed: 2920 samples/sec	time cost: 448.622798
[Epoch 103] validation: err-top1=0.218112 err-top5=0.056999
Epoch[104] Batch [49]	Speed: 2964.610984 samples/sec	accuracy=0.845627	lr=0.018712
Epoch[104] Batch [99]	Speed: 2876.999450 samples/sec	accuracy=0.845946	lr=0.018620
Epoch[104] Batch [149]	Speed: 2856.932420 samples/sec	accuracy=0.846026	lr=0.018528
Epoch[104] Batch [199]	Speed: 2968.145561 samples/sec	accuracy=0.846084	lr=0.018436
Epoch[104] Batch [249]	Speed: 2897.315312 samples/sec	accuracy=0.846185	lr=0.018345
Epoch[104] Batch [299]	Speed: 2929.759891 samples/sec	accuracy=0.846160	lr=0.018254
Epoch[104] Batch [349]	Speed: 3008.891911 samples/sec	accuracy=0.846129	lr=0.018163
Epoch[104] Batch [399]	Speed: 2868.316820 samples/sec	accuracy=0.846151	lr=0.018072
Epoch[104] Batch [449]	Speed: 2910.136756 samples/sec	accuracy=0.846161	lr=0.017981
Epoch[104] Batch [499]	Speed: 2900.573458 samples/sec	accuracy=0.846183	lr=0.017891
Epoch[104] Batch [549]	Speed: 2890.918686 samples/sec	accuracy=0.846256	lr=0.017801
Epoch[104] Batch [599]	Speed: 2874.385003 samples/sec	accuracy=0.846295	lr=0.017711
Epoch[104] Batch [649]	Speed: 2913.137524 samples/sec	accuracy=0.846342	lr=0.017621
Epoch[104] Batch [699]	Speed: 2917.221829 samples/sec	accuracy=0.846317	lr=0.017532
Epoch[104] Batch [749]	Speed: 2952.374919 samples/sec	accuracy=0.846339	lr=0.017442
Epoch[104] Batch [799]	Speed: 2966.174262 samples/sec	accuracy=0.846340	lr=0.017353
Epoch[104] Batch [849]	Speed: 3029.675795 samples/sec	accuracy=0.846344	lr=0.017264
Epoch[104] Batch [899]	Speed: 2894.187300 samples/sec	accuracy=0.846345	lr=0.017176
Epoch[104] Batch [949]	Speed: 3009.199657 samples/sec	accuracy=0.846372	lr=0.017087
Epoch[104] Batch [999]	Speed: 2934.362783 samples/sec	accuracy=0.846417	lr=0.016999
Epoch[104] Batch [1049]	Speed: 2910.900958 samples/sec	accuracy=0.846447	lr=0.016911
Epoch[104] Batch [1099]	Speed: 3111.333424 samples/sec	accuracy=0.846466	lr=0.016823
Epoch[104] Batch [1149]	Speed: 2926.329171 samples/sec	accuracy=0.846546	lr=0.016736
Epoch[104] Batch [1199]	Speed: 2869.968825 samples/sec	accuracy=0.846577	lr=0.016648
Epoch[104] Batch [1249]	Speed: 3074.030105 samples/sec	accuracy=0.846619	lr=0.016561
[Epoch 104] training: accuracy=0.846621
[Epoch 104] speed: 2934 samples/sec	time cost: 446.182367
[Epoch 104] validation: err-top1=0.214986 err-top5=0.056620
Epoch[105] Batch [49]	Speed: 2993.591474 samples/sec	accuracy=0.848466	lr=0.016473
Epoch[105] Batch [99]	Speed: 2907.586199 samples/sec	accuracy=0.847617	lr=0.016386
Epoch[105] Batch [149]	Speed: 2875.278286 samples/sec	accuracy=0.847334	lr=0.016299
Epoch[105] Batch [199]	Speed: 2880.250244 samples/sec	accuracy=0.847309	lr=0.016213
Epoch[105] Batch [249]	Speed: 2925.878477 samples/sec	accuracy=0.847531	lr=0.016127
Epoch[105] Batch [299]	Speed: 2801.807486 samples/sec	accuracy=0.847522	lr=0.016041
Epoch[105] Batch [349]	Speed: 2829.639910 samples/sec	accuracy=0.847596	lr=0.015956
Epoch[105] Batch [399]	Speed: 3013.113701 samples/sec	accuracy=0.847680	lr=0.015870
Epoch[105] Batch [449]	Speed: 2877.238246 samples/sec	accuracy=0.847680	lr=0.015785
Epoch[105] Batch [499]	Speed: 2925.374760 samples/sec	accuracy=0.847682	lr=0.015700
Epoch[105] Batch [549]	Speed: 2918.553166 samples/sec	accuracy=0.847725	lr=0.015616
Epoch[105] Batch [599]	Speed: 2975.379311 samples/sec	accuracy=0.847758	lr=0.015531
Epoch[105] Batch [649]	Speed: 2847.120261 samples/sec	accuracy=0.847755	lr=0.015447
Epoch[105] Batch [699]	Speed: 2868.321647 samples/sec	accuracy=0.847771	lr=0.015363
Epoch[105] Batch [749]	Speed: 2900.533693 samples/sec	accuracy=0.847832	lr=0.015279
Epoch[105] Batch [799]	Speed: 2952.723949 samples/sec	accuracy=0.847877	lr=0.015195
Epoch[105] Batch [849]	Speed: 2992.641402 samples/sec	accuracy=0.847913	lr=0.015112
Epoch[105] Batch [899]	Speed: 2926.130799 samples/sec	accuracy=0.847929	lr=0.015029
Epoch[105] Batch [949]	Speed: 2935.709774 samples/sec	accuracy=0.847932	lr=0.014946
Epoch[105] Batch [999]	Speed: 2908.378523 samples/sec	accuracy=0.848008	lr=0.014863
Epoch[105] Batch [1049]	Speed: 2895.041728 samples/sec	accuracy=0.848073	lr=0.014781
Epoch[105] Batch [1099]	Speed: 2927.204960 samples/sec	accuracy=0.848104	lr=0.014699
Epoch[105] Batch [1149]	Speed: 2843.697374 samples/sec	accuracy=0.848116	lr=0.014616
Epoch[105] Batch [1199]	Speed: 2912.992857 samples/sec	accuracy=0.848133	lr=0.014535
Epoch[105] Batch [1249]	Speed: 3046.478232 samples/sec	accuracy=0.848174	lr=0.014453
[Epoch 105] training: accuracy=0.848174
[Epoch 105] speed: 2912 samples/sec	time cost: 449.977182
[Epoch 105] validation: err-top1=0.213927 err-top5=0.057717
Epoch[106] Batch [49]	Speed: 2988.472713 samples/sec	accuracy=0.848293	lr=0.014370
Epoch[106] Batch [99]	Speed: 2944.382714 samples/sec	accuracy=0.848603	lr=0.014289
Epoch[106] Batch [149]	Speed: 2895.193127 samples/sec	accuracy=0.848760	lr=0.014208
Epoch[106] Batch [199]	Speed: 2938.999587 samples/sec	accuracy=0.848714	lr=0.014127
Epoch[106] Batch [249]	Speed: 2997.819578 samples/sec	accuracy=0.848684	lr=0.014047
Epoch[106] Batch [299]	Speed: 2882.408796 samples/sec	accuracy=0.848639	lr=0.013966
Epoch[106] Batch [349]	Speed: 2881.851364 samples/sec	accuracy=0.848660	lr=0.013886
Epoch[106] Batch [399]	Speed: 2894.735662 samples/sec	accuracy=0.848674	lr=0.013806
Epoch[106] Batch [449]	Speed: 2913.546669 samples/sec	accuracy=0.848770	lr=0.013727
Epoch[106] Batch [499]	Speed: 2886.821381 samples/sec	accuracy=0.848773	lr=0.013647
Epoch[106] Batch [549]	Speed: 2891.731312 samples/sec	accuracy=0.848886	lr=0.013568
Epoch[106] Batch [599]	Speed: 2849.514675 samples/sec	accuracy=0.848864	lr=0.013489
Epoch[106] Batch [649]	Speed: 2958.460023 samples/sec	accuracy=0.848834	lr=0.013411
Epoch[106] Batch [699]	Speed: 3019.003047 samples/sec	accuracy=0.848856	lr=0.013332
Epoch[106] Batch [749]	Speed: 2948.933865 samples/sec	accuracy=0.848895	lr=0.013254
Epoch[106] Batch [799]	Speed: 2961.897826 samples/sec	accuracy=0.848918	lr=0.013176
Epoch[106] Batch [849]	Speed: 2892.459423 samples/sec	accuracy=0.848954	lr=0.013098
Epoch[106] Batch [899]	Speed: 2909.444340 samples/sec	accuracy=0.849026	lr=0.013020
Epoch[106] Batch [949]	Speed: 2964.833846 samples/sec	accuracy=0.849091	lr=0.012943
Epoch[106] Batch [999]	Speed: 2914.152100 samples/sec	accuracy=0.849141	lr=0.012866
Epoch[106] Batch [1049]	Speed: 2976.374390 samples/sec	accuracy=0.849160	lr=0.012789
Epoch[106] Batch [1099]	Speed: 2847.509146 samples/sec	accuracy=0.849192	lr=0.012712
Epoch[106] Batch [1149]	Speed: 2844.261614 samples/sec	accuracy=0.849232	lr=0.012636
Epoch[106] Batch [1199]	Speed: 3032.781158 samples/sec	accuracy=0.849288	lr=0.012559
Epoch[106] Batch [1249]	Speed: 3020.534032 samples/sec	accuracy=0.849346	lr=0.012483
[Epoch 106] training: accuracy=0.849347
[Epoch 106] speed: 2926 samples/sec	time cost: 447.381387
[Epoch 106] validation: err-top1=0.213070 err-top5=0.056362
Epoch[107] Batch [49]	Speed: 3026.424897 samples/sec	accuracy=0.849779	lr=0.012406
Epoch[107] Batch [99]	Speed: 2864.134759 samples/sec	accuracy=0.850203	lr=0.012330
Epoch[107] Batch [149]	Speed: 2916.413272 samples/sec	accuracy=0.850142	lr=0.012255
Epoch[107] Batch [199]	Speed: 2889.109118 samples/sec	accuracy=0.850232	lr=0.012180
Epoch[107] Batch [249]	Speed: 2943.922568 samples/sec	accuracy=0.850200	lr=0.012105
Epoch[107] Batch [299]	Speed: 2991.618790 samples/sec	accuracy=0.850151	lr=0.012030
Epoch[107] Batch [349]	Speed: 2856.324238 samples/sec	accuracy=0.850170	lr=0.011956
Epoch[107] Batch [399]	Speed: 2931.473528 samples/sec	accuracy=0.850178	lr=0.011881
Epoch[107] Batch [449]	Speed: 2882.819725 samples/sec	accuracy=0.850252	lr=0.011807
Epoch[107] Batch [499]	Speed: 2940.759208 samples/sec	accuracy=0.850199	lr=0.011734
Epoch[107] Batch [549]	Speed: 2819.934360 samples/sec	accuracy=0.850227	lr=0.011660
Epoch[107] Batch [599]	Speed: 2846.471313 samples/sec	accuracy=0.850231	lr=0.011587
Epoch[107] Batch [649]	Speed: 2913.331806 samples/sec	accuracy=0.850300	lr=0.011513
Epoch[107] Batch [699]	Speed: 2939.119013 samples/sec	accuracy=0.850312	lr=0.011441
Epoch[107] Batch [749]	Speed: 2903.692367 samples/sec	accuracy=0.850332	lr=0.011368
Epoch[107] Batch [799]	Speed: 2958.892273 samples/sec	accuracy=0.850391	lr=0.011295
Epoch[107] Batch [849]	Speed: 2886.380329 samples/sec	accuracy=0.850429	lr=0.011223
Epoch[107] Batch [899]	Speed: 2910.928617 samples/sec	accuracy=0.850519	lr=0.011151
Epoch[107] Batch [949]	Speed: 3052.576396 samples/sec	accuracy=0.850562	lr=0.011079
Epoch[107] Batch [999]	Speed: 2883.415859 samples/sec	accuracy=0.850603	lr=0.011008
Epoch[107] Batch [1049]	Speed: 2858.753800 samples/sec	accuracy=0.850640	lr=0.010936
Epoch[107] Batch [1099]	Speed: 2923.556678 samples/sec	accuracy=0.850680	lr=0.010865
Epoch[107] Batch [1149]	Speed: 2850.197128 samples/sec	accuracy=0.850691	lr=0.010794
Epoch[107] Batch [1199]	Speed: 2885.934758 samples/sec	accuracy=0.850717	lr=0.010724
Epoch[107] Batch [1249]	Speed: 2983.801410 samples/sec	accuracy=0.850786	lr=0.010653
[Epoch 107] training: accuracy=0.850789
[Epoch 107] speed: 2911 samples/sec	time cost: 450.288887
[Epoch 107] validation: err-top1=0.211296 err-top5=0.055186
Epoch[108] Batch [49]	Speed: 2827.790635 samples/sec	accuracy=0.851822	lr=0.010582
Epoch[108] Batch [99]	Speed: 2834.003027 samples/sec	accuracy=0.852153	lr=0.010512
Epoch[108] Batch [149]	Speed: 2961.089673 samples/sec	accuracy=0.851948	lr=0.010442
Epoch[108] Batch [199]	Speed: 2983.623524 samples/sec	accuracy=0.851816	lr=0.010373
Epoch[108] Batch [249]	Speed: 2944.542023 samples/sec	accuracy=0.851857	lr=0.010303
Epoch[108] Batch [299]	Speed: 2902.591096 samples/sec	accuracy=0.851719	lr=0.010234
Epoch[108] Batch [349]	Speed: 2982.187737 samples/sec	accuracy=0.851710	lr=0.010165
Epoch[108] Batch [399]	Speed: 2883.550983 samples/sec	accuracy=0.851704	lr=0.010097
Epoch[108] Batch [449]	Speed: 2908.239763 samples/sec	accuracy=0.851644	lr=0.010028
Epoch[108] Batch [499]	Speed: 3027.300990 samples/sec	accuracy=0.851675	lr=0.009960
Epoch[108] Batch [549]	Speed: 2929.141206 samples/sec	accuracy=0.851686	lr=0.009892
Epoch[108] Batch [599]	Speed: 2997.888714 samples/sec	accuracy=0.851698	lr=0.009824
Epoch[108] Batch [649]	Speed: 2913.768008 samples/sec	accuracy=0.851693	lr=0.009757
Epoch[108] Batch [699]	Speed: 2868.833845 samples/sec	accuracy=0.851662	lr=0.009690
Epoch[108] Batch [749]	Speed: 2898.002240 samples/sec	accuracy=0.851635	lr=0.009623
Epoch[108] Batch [799]	Speed: 2917.770947 samples/sec	accuracy=0.851626	lr=0.009556
Epoch[108] Batch [849]	Speed: 3055.116989 samples/sec	accuracy=0.851643	lr=0.009489
Epoch[108] Batch [899]	Speed: 2827.612582 samples/sec	accuracy=0.851671	lr=0.009423
Epoch[108] Batch [949]	Speed: 2971.533842 samples/sec	accuracy=0.851685	lr=0.009357
Epoch[108] Batch [999]	Speed: 2973.252333 samples/sec	accuracy=0.851714	lr=0.009291
Epoch[108] Batch [1049]	Speed: 2889.225534 samples/sec	accuracy=0.851761	lr=0.009225
Epoch[108] Batch [1099]	Speed: 2916.982016 samples/sec	accuracy=0.851810	lr=0.009160
Epoch[108] Batch [1149]	Speed: 2965.716373 samples/sec	accuracy=0.851855	lr=0.009095
Epoch[108] Batch [1199]	Speed: 2994.846841 samples/sec	accuracy=0.851867	lr=0.009030
Epoch[108] Batch [1249]	Speed: 3026.176432 samples/sec	accuracy=0.851882	lr=0.008965
[Epoch 108] training: accuracy=0.851884
[Epoch 108] speed: 2932 samples/sec	time cost: 446.715869
[Epoch 108] validation: err-top1=0.209363 err-top5=0.054986
Epoch[109] Batch [49]	Speed: 2933.913619 samples/sec	accuracy=0.852412	lr=0.008899
Epoch[109] Batch [99]	Speed: 2932.176111 samples/sec	accuracy=0.852367	lr=0.008835
Epoch[109] Batch [149]	Speed: 2873.433718 samples/sec	accuracy=0.852320	lr=0.008771
Epoch[109] Batch [199]	Speed: 2994.905857 samples/sec	accuracy=0.852396	lr=0.008707
Epoch[109] Batch [249]	Speed: 2947.898126 samples/sec	accuracy=0.852367	lr=0.008643
Epoch[109] Batch [299]	Speed: 2891.697863 samples/sec	accuracy=0.852460	lr=0.008580
Epoch[109] Batch [349]	Speed: 2893.745515 samples/sec	accuracy=0.852528	lr=0.008517
Epoch[109] Batch [399]	Speed: 2865.984080 samples/sec	accuracy=0.852613	lr=0.008454
Epoch[109] Batch [449]	Speed: 2967.270854 samples/sec	accuracy=0.852720	lr=0.008391
Epoch[109] Batch [499]	Speed: 2932.507646 samples/sec	accuracy=0.852702	lr=0.008329
Epoch[109] Batch [549]	Speed: 2995.205108 samples/sec	accuracy=0.852718	lr=0.008266
Epoch[109] Batch [599]	Speed: 2869.215841 samples/sec	accuracy=0.852739	lr=0.008204
Epoch[109] Batch [649]	Speed: 2879.918137 samples/sec	accuracy=0.852795	lr=0.008143
Epoch[109] Batch [699]	Speed: 2867.607704 samples/sec	accuracy=0.852791	lr=0.008081
Epoch[109] Batch [749]	Speed: 2977.885641 samples/sec	accuracy=0.852816	lr=0.008020
Epoch[109] Batch [799]	Speed: 2840.000466 samples/sec	accuracy=0.852872	lr=0.007959
Epoch[109] Batch [849]	Speed: 3016.746050 samples/sec	accuracy=0.852927	lr=0.007898
Epoch[109] Batch [899]	Speed: 2991.305588 samples/sec	accuracy=0.852976	lr=0.007837
Epoch[109] Batch [949]	Speed: 3104.193512 samples/sec	accuracy=0.853025	lr=0.007777
Epoch[109] Batch [999]	Speed: 2885.271835 samples/sec	accuracy=0.853052	lr=0.007716
Epoch[109] Batch [1049]	Speed: 3034.485138 samples/sec	accuracy=0.853078	lr=0.007656
Epoch[109] Batch [1099]	Speed: 2908.293643 samples/sec	accuracy=0.853121	lr=0.007597
Epoch[109] Batch [1149]	Speed: 3011.747853 samples/sec	accuracy=0.853135	lr=0.007537
Epoch[109] Batch [1199]	Speed: 2783.527108 samples/sec	accuracy=0.853178	lr=0.007478
Epoch[109] Batch [1249]	Speed: 3026.859275 samples/sec	accuracy=0.853222	lr=0.007419
[Epoch 109] training: accuracy=0.853222
[Epoch 109] speed: 2933 samples/sec	time cost: 446.557114
[Epoch 109] validation: err-top1=0.209323 err-top5=0.054169
Epoch[110] Batch [49]	Speed: 2893.093576 samples/sec	accuracy=0.854017	lr=0.007359
Epoch[110] Batch [99]	Speed: 2941.548726 samples/sec	accuracy=0.853930	lr=0.007300
Epoch[110] Batch [149]	Speed: 2895.897482 samples/sec	accuracy=0.853909	lr=0.007242
Epoch[110] Batch [199]	Speed: 2906.101433 samples/sec	accuracy=0.853948	lr=0.007184
Epoch[110] Batch [249]	Speed: 2846.287543 samples/sec	accuracy=0.853890	lr=0.007126
Epoch[110] Batch [299]	Speed: 2813.123568 samples/sec	accuracy=0.853851	lr=0.007068
Epoch[110] Batch [349]	Speed: 2989.055974 samples/sec	accuracy=0.853873	lr=0.007011
Epoch[110] Batch [399]	Speed: 3010.724384 samples/sec	accuracy=0.853852	lr=0.006954
Epoch[110] Batch [449]	Speed: 2921.819406 samples/sec	accuracy=0.853814	lr=0.006897
Epoch[110] Batch [499]	Speed: 2833.256534 samples/sec	accuracy=0.853842	lr=0.006840
Epoch[110] Batch [549]	Speed: 2842.198167 samples/sec	accuracy=0.853860	lr=0.006784
Epoch[110] Batch [599]	Speed: 2815.466248 samples/sec	accuracy=0.853857	lr=0.006727
Epoch[110] Batch [649]	Speed: 2942.165368 samples/sec	accuracy=0.853894	lr=0.006671
Epoch[110] Batch [699]	Speed: 3020.575541 samples/sec	accuracy=0.853941	lr=0.006615
Epoch[110] Batch [749]	Speed: 2956.653190 samples/sec	accuracy=0.853959	lr=0.006560
Epoch[110] Batch [799]	Speed: 2983.134000 samples/sec	accuracy=0.853985	lr=0.006504
Epoch[110] Batch [849]	Speed: 2952.018303 samples/sec	accuracy=0.854002	lr=0.006449
Epoch[110] Batch [899]	Speed: 2902.438687 samples/sec	accuracy=0.854003	lr=0.006394
Epoch[110] Batch [949]	Speed: 2908.942325 samples/sec	accuracy=0.854014	lr=0.006340
Epoch[110] Batch [999]	Speed: 2888.582118 samples/sec	accuracy=0.854039	lr=0.006285
Epoch[110] Batch [1049]	Speed: 2897.387434 samples/sec	accuracy=0.854071	lr=0.006231
Epoch[110] Batch [1099]	Speed: 2903.722639 samples/sec	accuracy=0.854079	lr=0.006177
Epoch[110] Batch [1149]	Speed: 2995.912077 samples/sec	accuracy=0.854121	lr=0.006123
Epoch[110] Batch [1199]	Speed: 2988.243372 samples/sec	accuracy=0.854139	lr=0.006070
Epoch[110] Batch [1249]	Speed: 3082.046243 samples/sec	accuracy=0.854178	lr=0.006017
[Epoch 110] training: accuracy=0.854180
[Epoch 110] speed: 2921 samples/sec	time cost: 448.093205
[Epoch 110] validation: err-top1=0.206421 err-top5=0.053975
Epoch[111] Batch [49]	Speed: 2858.664562 samples/sec	accuracy=0.854914	lr=0.005963
Epoch[111] Batch [99]	Speed: 2875.872922 samples/sec	accuracy=0.855008	lr=0.005910
Epoch[111] Batch [149]	Speed: 3003.769011 samples/sec	accuracy=0.854907	lr=0.005857
Epoch[111] Batch [199]	Speed: 2887.494722 samples/sec	accuracy=0.854726	lr=0.005805
Epoch[111] Batch [249]	Speed: 2872.467271 samples/sec	accuracy=0.854661	lr=0.005753
Epoch[111] Batch [299]	Speed: 2808.066124 samples/sec	accuracy=0.854634	lr=0.005701
Epoch[111] Batch [349]	Speed: 2873.004012 samples/sec	accuracy=0.854551	lr=0.005649
Epoch[111] Batch [399]	Speed: 2812.262443 samples/sec	accuracy=0.854599	lr=0.005598
Epoch[111] Batch [449]	Speed: 2901.564637 samples/sec	accuracy=0.854653	lr=0.005547
Epoch[111] Batch [499]	Speed: 2981.577471 samples/sec	accuracy=0.854729	lr=0.005496
Epoch[111] Batch [549]	Speed: 2932.549293 samples/sec	accuracy=0.854730	lr=0.005445
Epoch[111] Batch [599]	Speed: 2853.856931 samples/sec	accuracy=0.854739	lr=0.005394
Epoch[111] Batch [649]	Speed: 2910.329968 samples/sec	accuracy=0.854779	lr=0.005344
Epoch[111] Batch [699]	Speed: 2932.908791 samples/sec	accuracy=0.854807	lr=0.005294
Epoch[111] Batch [749]	Speed: 2926.424160 samples/sec	accuracy=0.854840	lr=0.005244
Epoch[111] Batch [799]	Speed: 2959.506746 samples/sec	accuracy=0.854854	lr=0.005195
Epoch[111] Batch [849]	Speed: 2879.411858 samples/sec	accuracy=0.854902	lr=0.005145
Epoch[111] Batch [899]	Speed: 2966.122190 samples/sec	accuracy=0.854948	lr=0.005096
Epoch[111] Batch [949]	Speed: 2895.224939 samples/sec	accuracy=0.854943	lr=0.005047
Epoch[111] Batch [999]	Speed: 2823.004967 samples/sec	accuracy=0.854949	lr=0.004999
Epoch[111] Batch [1049]	Speed: 2871.315495 samples/sec	accuracy=0.854949	lr=0.004950
Epoch[111] Batch [1099]	Speed: 2953.352596 samples/sec	accuracy=0.854950	lr=0.004902
Epoch[111] Batch [1149]	Speed: 2943.264204 samples/sec	accuracy=0.854974	lr=0.004854
Epoch[111] Batch [1199]	Speed: 2822.980994 samples/sec	accuracy=0.854996	lr=0.004807
Epoch[111] Batch [1249]	Speed: 3054.307781 samples/sec	accuracy=0.855020	lr=0.004759
[Epoch 111] training: accuracy=0.855020
[Epoch 111] speed: 2900 samples/sec	time cost: 451.662456
[Epoch 111] validation: err-top1=0.206075 err-top5=0.053093
Epoch[112] Batch [49]	Speed: 2980.552021 samples/sec	accuracy=0.855228	lr=0.004711
Epoch[112] Batch [99]	Speed: 2874.444715 samples/sec	accuracy=0.855391	lr=0.004664
Epoch[112] Batch [149]	Speed: 2980.220288 samples/sec	accuracy=0.855344	lr=0.004617
Epoch[112] Batch [199]	Speed: 3004.700644 samples/sec	accuracy=0.855365	lr=0.004571
Epoch[112] Batch [249]	Speed: 2916.814820 samples/sec	accuracy=0.855437	lr=0.004524
Epoch[112] Batch [299]	Speed: 2906.000799 samples/sec	accuracy=0.855464	lr=0.004478
Epoch[112] Batch [349]	Speed: 2963.413137 samples/sec	accuracy=0.855482	lr=0.004432
Epoch[112] Batch [399]	Speed: 2912.076267 samples/sec	accuracy=0.855498	lr=0.004387
Epoch[112] Batch [449]	Speed: 2908.512491 samples/sec	accuracy=0.855504	lr=0.004341
Epoch[112] Batch [499]	Speed: 2886.951429 samples/sec	accuracy=0.855537	lr=0.004296
Epoch[112] Batch [549]	Speed: 3099.734076 samples/sec	accuracy=0.855533	lr=0.004251
Epoch[112] Batch [599]	Speed: 2909.745285 samples/sec	accuracy=0.855584	lr=0.004207
Epoch[112] Batch [649]	Speed: 2949.452331 samples/sec	accuracy=0.855644	lr=0.004162
Epoch[112] Batch [699]	Speed: 2877.062046 samples/sec	accuracy=0.855647	lr=0.004118
Epoch[112] Batch [749]	Speed: 2893.306283 samples/sec	accuracy=0.855658	lr=0.004074
Epoch[112] Batch [799]	Speed: 3072.619915 samples/sec	accuracy=0.855679	lr=0.004030
Epoch[112] Batch [849]	Speed: 2857.113233 samples/sec	accuracy=0.855704	lr=0.003987
Epoch[112] Batch [899]	Speed: 2824.534142 samples/sec	accuracy=0.855703	lr=0.003944
Epoch[112] Batch [949]	Speed: 2929.139528 samples/sec	accuracy=0.855715	lr=0.003901
Epoch[112] Batch [999]	Speed: 2804.159103 samples/sec	accuracy=0.855734	lr=0.003858
Epoch[112] Batch [1049]	Speed: 2922.761115 samples/sec	accuracy=0.855766	lr=0.003815
Epoch[112] Batch [1099]	Speed: 2969.387515 samples/sec	accuracy=0.855782	lr=0.003773
Epoch[112] Batch [1149]	Speed: 2971.371888 samples/sec	accuracy=0.855805	lr=0.003731
Epoch[112] Batch [1199]	Speed: 2888.666551 samples/sec	accuracy=0.855832	lr=0.003689
Epoch[112] Batch [1249]	Speed: 3060.388339 samples/sec	accuracy=0.855837	lr=0.003647
[Epoch 112] training: accuracy=0.855837
[Epoch 112] speed: 2930 samples/sec	time cost: 446.666856
[Epoch 112] validation: err-top1=0.203563 err-top5=0.053093
Epoch[113] Batch [49]	Speed: 3091.504594 samples/sec	accuracy=0.856036	lr=0.003605
Epoch[113] Batch [99]	Speed: 3022.232483 samples/sec	accuracy=0.855907	lr=0.003564
Epoch[113] Batch [149]	Speed: 2865.359040 samples/sec	accuracy=0.855915	lr=0.003523
Epoch[113] Batch [199]	Speed: 2967.213742 samples/sec	accuracy=0.855965	lr=0.003482
Epoch[113] Batch [249]	Speed: 2953.143072 samples/sec	accuracy=0.855976	lr=0.003442
Epoch[113] Batch [299]	Speed: 3181.879433 samples/sec	accuracy=0.855996	lr=0.003402
Epoch[113] Batch [349]	Speed: 2887.856657 samples/sec	accuracy=0.856031	lr=0.003362
Epoch[113] Batch [399]	Speed: 2844.602956 samples/sec	accuracy=0.856038	lr=0.003322
Epoch[113] Batch [449]	Speed: 2937.955500 samples/sec	accuracy=0.856038	lr=0.003282
Epoch[113] Batch [499]	Speed: 2807.744250 samples/sec	accuracy=0.856065	lr=0.003243
Epoch[113] Batch [549]	Speed: 2901.086540 samples/sec	accuracy=0.856114	lr=0.003204
Epoch[113] Batch [599]	Speed: 2888.010062 samples/sec	accuracy=0.856132	lr=0.003165
Epoch[113] Batch [649]	Speed: 2785.833743 samples/sec	accuracy=0.856161	lr=0.003127
Epoch[113] Batch [699]	Speed: 3049.940089 samples/sec	accuracy=0.856174	lr=0.003088
Epoch[113] Batch [749]	Speed: 2968.606581 samples/sec	accuracy=0.856199	lr=0.003050
Epoch[113] Batch [799]	Speed: 2817.012392 samples/sec	accuracy=0.856218	lr=0.003012
Epoch[113] Batch [849]	Speed: 2885.593972 samples/sec	accuracy=0.856249	lr=0.002975
Epoch[113] Batch [899]	Speed: 2929.040168 samples/sec	accuracy=0.856261	lr=0.002937
Epoch[113] Batch [949]	Speed: 2809.898139 samples/sec	accuracy=0.856278	lr=0.002900
Epoch[113] Batch [999]	Speed: 2907.943619 samples/sec	accuracy=0.856275	lr=0.002863
Epoch[113] Batch [1049]	Speed: 3017.964337 samples/sec	accuracy=0.856296	lr=0.002826
Epoch[113] Batch [1099]	Speed: 2924.237473 samples/sec	accuracy=0.856322	lr=0.002790
Epoch[113] Batch [1149]	Speed: 2931.277179 samples/sec	accuracy=0.856328	lr=0.002754
Epoch[113] Batch [1199]	Speed: 3020.370601 samples/sec	accuracy=0.856331	lr=0.002718
Epoch[113] Batch [1249]	Speed: 2990.388155 samples/sec	accuracy=0.856349	lr=0.002682
[Epoch 113] training: accuracy=0.856350
[Epoch 113] speed: 2930 samples/sec	time cost: 446.758700
[Epoch 113] validation: err-top1=0.204102 err-top5=0.052515
Epoch[114] Batch [49]	Speed: 2938.234191 samples/sec	accuracy=0.857050	lr=0.002646
Epoch[114] Batch [99]	Speed: 2963.562038 samples/sec	accuracy=0.856815	lr=0.002611
Epoch[114] Batch [149]	Speed: 2931.426189 samples/sec	accuracy=0.856790	lr=0.002575
Epoch[114] Batch [199]	Speed: 2944.077427 samples/sec	accuracy=0.856768	lr=0.002541
Epoch[114] Batch [249]	Speed: 2870.999109 samples/sec	accuracy=0.856769	lr=0.002506
Epoch[114] Batch [299]	Speed: 3075.856848 samples/sec	accuracy=0.856752	lr=0.002472
Epoch[114] Batch [349]	Speed: 2812.781744 samples/sec	accuracy=0.856784	lr=0.002438
Epoch[114] Batch [399]	Speed: 2993.199090 samples/sec	accuracy=0.856784	lr=0.002404
Epoch[114] Batch [449]	Speed: 2932.709527 samples/sec	accuracy=0.856792	lr=0.002370
Epoch[114] Batch [499]	Speed: 2940.462121 samples/sec	accuracy=0.856797	lr=0.002337
Epoch[114] Batch [549]	Speed: 3016.841151 samples/sec	accuracy=0.856764	lr=0.002304
Epoch[114] Batch [599]	Speed: 2915.357583 samples/sec	accuracy=0.856738	lr=0.002271
Epoch[114] Batch [649]	Speed: 2938.785659 samples/sec	accuracy=0.856786	lr=0.002238
Epoch[114] Batch [699]	Speed: 3026.784147 samples/sec	accuracy=0.856791	lr=0.002205
Epoch[114] Batch [749]	Speed: 2941.289669 samples/sec	accuracy=0.856811	lr=0.002173
Epoch[114] Batch [799]	Speed: 2936.556573 samples/sec	accuracy=0.856824	lr=0.002141
Epoch[114] Batch [849]	Speed: 2903.690208 samples/sec	accuracy=0.856823	lr=0.002110
Epoch[114] Batch [899]	Speed: 2891.642144 samples/sec	accuracy=0.856838	lr=0.002078
Epoch[114] Batch [949]	Speed: 2913.210634 samples/sec	accuracy=0.856852	lr=0.002047
Epoch[114] Batch [999]	Speed: 2895.883736 samples/sec	accuracy=0.856879	lr=0.002016
Epoch[114] Batch [1049]	Speed: 2924.886116 samples/sec	accuracy=0.856908	lr=0.001985
Epoch[114] Batch [1099]	Speed: 3039.502106 samples/sec	accuracy=0.856920	lr=0.001954
Epoch[114] Batch [1149]	Speed: 2868.680821 samples/sec	accuracy=0.856949	lr=0.001924
Epoch[114] Batch [1199]	Speed: 2950.146213 samples/sec	accuracy=0.856962	lr=0.001894
Epoch[114] Batch [1249]	Speed: 3010.930846 samples/sec	accuracy=0.856994	lr=0.001864
[Epoch 114] training: accuracy=0.856995
[Epoch 114] speed: 2940 samples/sec	time cost: 445.790346
[Epoch 114] validation: err-top1=0.204181 err-top5=0.052874
Epoch[115] Batch [49]	Speed: 2953.307107 samples/sec	accuracy=0.857319	lr=0.001834
Epoch[115] Batch [99]	Speed: 3001.995686 samples/sec	accuracy=0.857374	lr=0.001804
Epoch[115] Batch [149]	Speed: 2897.372696 samples/sec	accuracy=0.857352	lr=0.001775
Epoch[115] Batch [199]	Speed: 2879.430660 samples/sec	accuracy=0.857308	lr=0.001746
Epoch[115] Batch [249]	Speed: 3022.178041 samples/sec	accuracy=0.857256	lr=0.001718
Epoch[115] Batch [299]	Speed: 3030.892657 samples/sec	accuracy=0.857280	lr=0.001689
Epoch[115] Batch [349]	Speed: 3258.626158 samples/sec	accuracy=0.857298	lr=0.001661
Epoch[115] Batch [399]	Speed: 2984.929830 samples/sec	accuracy=0.857294	lr=0.001633
Epoch[115] Batch [449]	Speed: 3055.149457 samples/sec	accuracy=0.857294	lr=0.001605
Epoch[115] Batch [499]	Speed: 3046.842086 samples/sec	accuracy=0.857294	lr=0.001578
Epoch[115] Batch [549]	Speed: 3059.563564 samples/sec	accuracy=0.857290	lr=0.001551
Epoch[115] Batch [599]	Speed: 2915.214515 samples/sec	accuracy=0.857295	lr=0.001524
Epoch[115] Batch [649]	Speed: 2959.418692 samples/sec	accuracy=0.857279	lr=0.001497
Epoch[115] Batch [699]	Speed: 3053.996821 samples/sec	accuracy=0.857295	lr=0.001470
Epoch[115] Batch [749]	Speed: 2883.792184 samples/sec	accuracy=0.857319	lr=0.001444
Epoch[115] Batch [799]	Speed: 2839.452670 samples/sec	accuracy=0.857338	lr=0.001418
Epoch[115] Batch [849]	Speed: 2866.918382 samples/sec	accuracy=0.857356	lr=0.001392
Epoch[115] Batch [899]	Speed: 3011.704981 samples/sec	accuracy=0.857359	lr=0.001366
Epoch[115] Batch [949]	Speed: 2953.409501 samples/sec	accuracy=0.857356	lr=0.001341
Epoch[115] Batch [999]	Speed: 2860.206071 samples/sec	accuracy=0.857375	lr=0.001316
Epoch[115] Batch [1049]	Speed: 2893.785094 samples/sec	accuracy=0.857391	lr=0.001291
Epoch[115] Batch [1099]	Speed: 2880.188707 samples/sec	accuracy=0.857404	lr=0.001266
Epoch[115] Batch [1149]	Speed: 2832.164966 samples/sec	accuracy=0.857423	lr=0.001242
Epoch[115] Batch [1199]	Speed: 2946.057244 samples/sec	accuracy=0.857419	lr=0.001218
Epoch[115] Batch [1249]	Speed: 3035.899031 samples/sec	accuracy=0.857422	lr=0.001194
[Epoch 115] training: accuracy=0.857422
[Epoch 115] speed: 2959 samples/sec	time cost: 442.486139
[Epoch 115] validation: err-top1=0.202966 err-top5=0.053053
Epoch[116] Batch [49]	Speed: 3149.635347 samples/sec	accuracy=0.857422	lr=0.001170
Epoch[116] Batch [99]	Speed: 2792.730808 samples/sec	accuracy=0.857518	lr=0.001146
Epoch[116] Batch [149]	Speed: 2945.630473 samples/sec	accuracy=0.857520	lr=0.001123
Epoch[116] Batch [199]	Speed: 2888.005868 samples/sec	accuracy=0.857602	lr=0.001100
Epoch[116] Batch [249]	Speed: 2881.108445 samples/sec	accuracy=0.857594	lr=0.001077
Epoch[116] Batch [299]	Speed: 2965.787353 samples/sec	accuracy=0.857653	lr=0.001055
Epoch[116] Batch [349]	Speed: 2871.289389 samples/sec	accuracy=0.857623	lr=0.001032
Epoch[116] Batch [399]	Speed: 2998.433665 samples/sec	accuracy=0.857612	lr=0.001010
Epoch[116] Batch [449]	Speed: 2802.943161 samples/sec	accuracy=0.857603	lr=0.000989
Epoch[116] Batch [499]	Speed: 2936.412381 samples/sec	accuracy=0.857614	lr=0.000967
Epoch[116] Batch [549]	Speed: 3056.661510 samples/sec	accuracy=0.857635	lr=0.000946
Epoch[116] Batch [599]	Speed: 2850.782077 samples/sec	accuracy=0.857638	lr=0.000925
Epoch[116] Batch [649]	Speed: 3015.243415 samples/sec	accuracy=0.857650	lr=0.000904
Epoch[116] Batch [699]	Speed: 2866.350703 samples/sec	accuracy=0.857690	lr=0.000883
Epoch[116] Batch [749]	Speed: 2800.717913 samples/sec	accuracy=0.857720	lr=0.000863
Epoch[116] Batch [799]	Speed: 2874.152335 samples/sec	accuracy=0.857722	lr=0.000843
Epoch[116] Batch [849]	Speed: 3043.085369 samples/sec	accuracy=0.857742	lr=0.000823
Epoch[116] Batch [899]	Speed: 2867.107313 samples/sec	accuracy=0.857743	lr=0.000803
Epoch[116] Batch [949]	Speed: 2834.178257 samples/sec	accuracy=0.857742	lr=0.000784
Epoch[116] Batch [999]	Speed: 2878.802636 samples/sec	accuracy=0.857753	lr=0.000764
Epoch[116] Batch [1049]	Speed: 2851.080661 samples/sec	accuracy=0.857758	lr=0.000745
Epoch[116] Batch [1099]	Speed: 2852.890113 samples/sec	accuracy=0.857769	lr=0.000727
Epoch[116] Batch [1149]	Speed: 2932.001844 samples/sec	accuracy=0.857780	lr=0.000708
Epoch[116] Batch [1199]	Speed: 2879.965449 samples/sec	accuracy=0.857798	lr=0.000690
Epoch[116] Batch [1249]	Speed: 3025.901915 samples/sec	accuracy=0.857807	lr=0.000672
[Epoch 116] training: accuracy=0.857806
[Epoch 116] speed: 2909 samples/sec	time cost: 449.962434
[Epoch 116] validation: err-top1=0.202616 err-top5=0.052429
Epoch[117] Batch [49]	Speed: 2978.957439 samples/sec	accuracy=0.858360	lr=0.000654
Epoch[117] Batch [99]	Speed: 3066.528230 samples/sec	accuracy=0.858231	lr=0.000636
Epoch[117] Batch [149]	Speed: 2845.609110 samples/sec	accuracy=0.858230	lr=0.000619
Epoch[117] Batch [199]	Speed: 2966.647087 samples/sec	accuracy=0.858196	lr=0.000602
Epoch[117] Batch [249]	Speed: 2847.775850 samples/sec	accuracy=0.858122	lr=0.000585
Epoch[117] Batch [299]	Speed: 2902.655085 samples/sec	accuracy=0.858025	lr=0.000569
Epoch[117] Batch [349]	Speed: 3039.065339 samples/sec	accuracy=0.857999	lr=0.000552
Epoch[117] Batch [399]	Speed: 2885.699053 samples/sec	accuracy=0.857950	lr=0.000536
Epoch[117] Batch [449]	Speed: 3001.499780 samples/sec	accuracy=0.857937	lr=0.000520
Epoch[117] Batch [499]	Speed: 2835.260709 samples/sec	accuracy=0.857917	lr=0.000505
Epoch[117] Batch [549]	Speed: 2981.988138 samples/sec	accuracy=0.857949	lr=0.000489
Epoch[117] Batch [599]	Speed: 2822.216373 samples/sec	accuracy=0.857978	lr=0.000474
Epoch[117] Batch [649]	Speed: 2839.363544 samples/sec	accuracy=0.857967	lr=0.000459
Epoch[117] Batch [699]	Speed: 2953.526201 samples/sec	accuracy=0.857975	lr=0.000445
Epoch[117] Batch [749]	Speed: 2872.864571 samples/sec	accuracy=0.857970	lr=0.000430
Epoch[117] Batch [799]	Speed: 2981.116717 samples/sec	accuracy=0.857976	lr=0.000416
Epoch[117] Batch [849]	Speed: 2828.829756 samples/sec	accuracy=0.857977	lr=0.000402
Epoch[117] Batch [899]	Speed: 2952.726994 samples/sec	accuracy=0.857985	lr=0.000388
Epoch[117] Batch [949]	Speed: 2902.292963 samples/sec	accuracy=0.857994	lr=0.000375
Epoch[117] Batch [999]	Speed: 2948.372631 samples/sec	accuracy=0.857991	lr=0.000362
Epoch[117] Batch [1049]	Speed: 2773.356672 samples/sec	accuracy=0.857989	lr=0.000349
Epoch[117] Batch [1099]	Speed: 2902.828548 samples/sec	accuracy=0.857994	lr=0.000336
Epoch[117] Batch [1149]	Speed: 2841.196002 samples/sec	accuracy=0.858009	lr=0.000323
Epoch[117] Batch [1199]	Speed: 2871.393585 samples/sec	accuracy=0.857998	lr=0.000311
Epoch[117] Batch [1249]	Speed: 2995.689324 samples/sec	accuracy=0.858006	lr=0.000299
[Epoch 117] training: accuracy=0.858005
[Epoch 117] speed: 2909 samples/sec	time cost: 450.413648
[Epoch 117] validation: err-top1=0.202447 err-top5=0.052575
Epoch[118] Batch [49]	Speed: 3037.207021 samples/sec	accuracy=0.858250	lr=0.000287
Epoch[118] Batch [99]	Speed: 2949.087065 samples/sec	accuracy=0.858148	lr=0.000275
Epoch[118] Batch [149]	Speed: 2904.589779 samples/sec	accuracy=0.858124	lr=0.000264
Epoch[118] Batch [199]	Speed: 2828.200850 samples/sec	accuracy=0.858121	lr=0.000253
Epoch[118] Batch [249]	Speed: 2969.287623 samples/sec	accuracy=0.858160	lr=0.000242
Epoch[118] Batch [299]	Speed: 2889.229265 samples/sec	accuracy=0.858151	lr=0.000231
Epoch[118] Batch [349]	Speed: 2847.046166 samples/sec	accuracy=0.858142	lr=0.000221
Epoch[118] Batch [399]	Speed: 2845.892053 samples/sec	accuracy=0.858153	lr=0.000211
Epoch[118] Batch [449]	Speed: 2851.598949 samples/sec	accuracy=0.858159	lr=0.000201
Epoch[118] Batch [499]	Speed: 3020.167175 samples/sec	accuracy=0.858165	lr=0.000191
Epoch[118] Batch [549]	Speed: 2797.095989 samples/sec	accuracy=0.858163	lr=0.000182
Epoch[118] Batch [599]	Speed: 2908.532502 samples/sec	accuracy=0.858169	lr=0.000173
Epoch[118] Batch [649]	Speed: 2882.145313 samples/sec	accuracy=0.858154	lr=0.000164
Epoch[118] Batch [699]	Speed: 2857.582345 samples/sec	accuracy=0.858158	lr=0.000155
Epoch[118] Batch [749]	Speed: 2911.974784 samples/sec	accuracy=0.858152	lr=0.000147
Epoch[118] Batch [799]	Speed: 2921.041118 samples/sec	accuracy=0.858158	lr=0.000138
Epoch[118] Batch [849]	Speed: 3010.083145 samples/sec	accuracy=0.858148	lr=0.000130
Epoch[118] Batch [899]	Speed: 2913.179967 samples/sec	accuracy=0.858139	lr=0.000123
Epoch[118] Batch [949]	Speed: 2959.055928 samples/sec	accuracy=0.858115	lr=0.000115
Epoch[118] Batch [999]	Speed: 3002.870414 samples/sec	accuracy=0.858106	lr=0.000108
Epoch[118] Batch [1049]	Speed: 2843.227540 samples/sec	accuracy=0.858100	lr=0.000101
Epoch[118] Batch [1099]	Speed: 2903.388708 samples/sec	accuracy=0.858091	lr=0.000094
Epoch[118] Batch [1149]	Speed: 2814.677211 samples/sec	accuracy=0.858096	lr=0.000087
Epoch[118] Batch [1199]	Speed: 2974.997952 samples/sec	accuracy=0.858098	lr=0.000081
Epoch[118] Batch [1249]	Speed: 3055.786215 samples/sec	accuracy=0.858097	lr=0.000075
[Epoch 118] training: accuracy=0.858098
[Epoch 118] speed: 2912 samples/sec	time cost: 449.644324
[Epoch 118] validation: err-top1=0.202009 err-top5=0.051937
Epoch[119] Batch [49]	Speed: 2958.944091 samples/sec	accuracy=0.857935	lr=0.000069
Epoch[119] Batch [99]	Speed: 2947.399057 samples/sec	accuracy=0.858043	lr=0.000063
Epoch[119] Batch [149]	Speed: 3020.527065 samples/sec	accuracy=0.858052	lr=0.000058
Epoch[119] Batch [199]	Speed: 2875.452034 samples/sec	accuracy=0.858113	lr=0.000053
Epoch[119] Batch [249]	Speed: 2991.630834 samples/sec	accuracy=0.858129	lr=0.000048
Epoch[119] Batch [299]	Speed: 2861.784482 samples/sec	accuracy=0.858123	lr=0.000043
Epoch[119] Batch [349]	Speed: 2851.141642 samples/sec	accuracy=0.858126	lr=0.000039
Epoch[119] Batch [399]	Speed: 2947.963157 samples/sec	accuracy=0.858098	lr=0.000035
Epoch[119] Batch [449]	Speed: 2895.834415 samples/sec	accuracy=0.858068	lr=0.000031
Epoch[119] Batch [499]	Speed: 2896.872220 samples/sec	accuracy=0.858046	lr=0.000027
Epoch[119] Batch [549]	Speed: 3113.815993 samples/sec	accuracy=0.858045	lr=0.000023
Epoch[119] Batch [599]	Speed: 2868.130908 samples/sec	accuracy=0.858051	lr=0.000020
Epoch[119] Batch [649]	Speed: 2901.781218 samples/sec	accuracy=0.858071	lr=0.000017
Epoch[119] Batch [699]	Speed: 2902.598825 samples/sec	accuracy=0.858070	lr=0.000015
Epoch[119] Batch [749]	Speed: 2809.377476 samples/sec	accuracy=0.858088	lr=0.000012
Epoch[119] Batch [799]	Speed: 2891.096160 samples/sec	accuracy=0.858076	lr=0.000010
Epoch[119] Batch [849]	Speed: 2945.179227 samples/sec	accuracy=0.858073	lr=0.000008
Epoch[119] Batch [899]	Speed: 2854.734007 samples/sec	accuracy=0.858075	lr=0.000006
Epoch[119] Batch [949]	Speed: 2987.993279 samples/sec	accuracy=0.858080	lr=0.000004
Epoch[119] Batch [999]	Speed: 2932.048001 samples/sec	accuracy=0.858077	lr=0.000003
Epoch[119] Batch [1049]	Speed: 2996.459822 samples/sec	accuracy=0.858086	lr=0.000002
Epoch[119] Batch [1099]	Speed: 2858.630885 samples/sec	accuracy=0.858083	lr=0.000001
Epoch[119] Batch [1149]	Speed: 2914.916434 samples/sec	accuracy=0.858081	lr=0.000000
Epoch[119] Batch [1199]	Speed: 2851.847030 samples/sec	accuracy=0.858092	lr=0.000000
Epoch[119] Batch [1249]	Speed: 3005.018255 samples/sec	accuracy=0.858092	lr=0.000000
[Epoch 119] training: accuracy=0.858090
[Epoch 119] speed: 2919 samples/sec	time cost: 448.586103
[Epoch 119] validation: err-top1=0.202129 err-top5=0.052276