Namespace(batch_norm=False, batch_size=128, crop_ratio=0.875, data_dir='~/.mxnet/datasets/imagenet', dtype='float16', input_size=224, label_smoothing=True, last_gamma=True, log_interval=50, logging_file='resnet101_v1b_best.log', lr=0.4, lr_decay=0.1, lr_decay_epoch='40,60', lr_decay_period=0, lr_mode='cosine', mixup=False, mixup_alpha=0.2, mixup_off_epoch=0, mode='hybrid', model='resnet101_v1b', momentum=0.9, no_wd=True, num_epochs=120, num_gpus=8, num_workers=60, rec_train='/media/ramdisk/rec/train.rec', rec_train_idx='/media/ramdisk/rec/train.idx', rec_val='/media/ramdisk/rec/val.rec', rec_val_idx='/media/ramdisk/rec/val.idx', resume_epoch=0, resume_params='', resume_states='', save_dir='params_resnet101_v1b_best', save_frequency=10, use_pretrained=False, use_rec=True, use_se=False, warmup_epochs=5, warmup_lr=0.0, wd=0.0001)
Epoch[0] Batch [49]	Speed: 1297.657788 samples/sec	accuracy=0.001094	lr=0.003133
Epoch[0] Batch [99]	Speed: 2847.734537 samples/sec	accuracy=0.001455	lr=0.006331
Epoch[0] Batch [149]	Speed: 2942.274247 samples/sec	accuracy=0.002077	lr=0.009528
Epoch[0] Batch [199]	Speed: 3087.663463 samples/sec	accuracy=0.002969	lr=0.012726
Epoch[0] Batch [249]	Speed: 3143.102006 samples/sec	accuracy=0.003805	lr=0.015923
Epoch[0] Batch [299]	Speed: 3015.249469 samples/sec	accuracy=0.004661	lr=0.019121
Epoch[0] Batch [349]	Speed: 3213.142555 samples/sec	accuracy=0.005592	lr=0.022318
Epoch[0] Batch [399]	Speed: 2922.396624 samples/sec	accuracy=0.006560	lr=0.025516
Epoch[0] Batch [449]	Speed: 2904.789838 samples/sec	accuracy=0.007500	lr=0.028713
Epoch[0] Batch [499]	Speed: 2928.456969 samples/sec	accuracy=0.008316	lr=0.031910
Epoch[0] Batch [549]	Speed: 3006.569517 samples/sec	accuracy=0.009206	lr=0.035108
Epoch[0] Batch [599]	Speed: 3157.501596 samples/sec	accuracy=0.010094	lr=0.038305
Epoch[0] Batch [649]	Speed: 2977.739674 samples/sec	accuracy=0.010998	lr=0.041503
Epoch[0] Batch [699]	Speed: 2958.086574 samples/sec	accuracy=0.011819	lr=0.044700
Epoch[0] Batch [749]	Speed: 3070.709327 samples/sec	accuracy=0.012689	lr=0.047898
Epoch[0] Batch [799]	Speed: 3147.466806 samples/sec	accuracy=0.013534	lr=0.051095
Epoch[0] Batch [849]	Speed: 3026.201294 samples/sec	accuracy=0.014354	lr=0.054293
Epoch[0] Batch [899]	Speed: 3011.862703 samples/sec	accuracy=0.015253	lr=0.057490
Epoch[0] Batch [949]	Speed: 2970.623112 samples/sec	accuracy=0.016094	lr=0.060687
Epoch[0] Batch [999]	Speed: 3015.149473 samples/sec	accuracy=0.016960	lr=0.063885
Epoch[0] Batch [1049]	Speed: 2902.416524 samples/sec	accuracy=0.017860	lr=0.067082
Epoch[0] Batch [1099]	Speed: 3065.874339 samples/sec	accuracy=0.018770	lr=0.070280
Epoch[0] Batch [1149]	Speed: 3135.511514 samples/sec	accuracy=0.019653	lr=0.073477
Epoch[0] Batch [1199]	Speed: 3167.448733 samples/sec	accuracy=0.020635	lr=0.076675
Epoch[0] Batch [1249]	Speed: 3337.536949 samples/sec	accuracy=0.021584	lr=0.079872
[Epoch 0] training: accuracy=0.021624
[Epoch 0] speed: 2879 samples/sec	time cost: 455.839574
[Epoch 0] validation: err-top1=0.943001 err-top5=0.840860
Epoch[1] Batch [49]	Speed: 3002.918283 samples/sec	accuracy=0.046367	lr=0.083133
Epoch[1] Batch [99]	Speed: 2882.642841 samples/sec	accuracy=0.047881	lr=0.086331
Epoch[1] Batch [149]	Speed: 2952.595418 samples/sec	accuracy=0.049473	lr=0.089528
Epoch[1] Batch [199]	Speed: 2977.001553 samples/sec	accuracy=0.050864	lr=0.092726
Epoch[1] Batch [249]	Speed: 3080.918620 samples/sec	accuracy=0.051996	lr=0.095923
Epoch[1] Batch [299]	Speed: 3067.736273 samples/sec	accuracy=0.053294	lr=0.099121
Epoch[1] Batch [349]	Speed: 3042.953939 samples/sec	accuracy=0.054688	lr=0.102318
Epoch[1] Batch [399]	Speed: 3029.973955 samples/sec	accuracy=0.056257	lr=0.105516
Epoch[1] Batch [449]	Speed: 2963.761182 samples/sec	accuracy=0.057513	lr=0.108713
Epoch[1] Batch [499]	Speed: 2923.906133 samples/sec	accuracy=0.059256	lr=0.111910
Epoch[1] Batch [549]	Speed: 2937.454006 samples/sec	accuracy=0.060850	lr=0.115108
Epoch[1] Batch [599]	Speed: 2939.365014 samples/sec	accuracy=0.062441	lr=0.118305
Epoch[1] Batch [649]	Speed: 3012.099740 samples/sec	accuracy=0.064041	lr=0.121503
Epoch[1] Batch [699]	Speed: 3061.971114 samples/sec	accuracy=0.065636	lr=0.124700
Epoch[1] Batch [749]	Speed: 2977.849633 samples/sec	accuracy=0.067353	lr=0.127898
Epoch[1] Batch [799]	Speed: 2993.060713 samples/sec	accuracy=0.069061	lr=0.131095
Epoch[1] Batch [849]	Speed: 3009.606665 samples/sec	accuracy=0.070852	lr=0.134293
Epoch[1] Batch [899]	Speed: 3011.204807 samples/sec	accuracy=0.072782	lr=0.137490
Epoch[1] Batch [949]	Speed: 2968.869939 samples/sec	accuracy=0.074705	lr=0.140687
Epoch[1] Batch [999]	Speed: 2986.442806 samples/sec	accuracy=0.076832	lr=0.143885
Epoch[1] Batch [1049]	Speed: 3013.115138 samples/sec	accuracy=0.079024	lr=0.147082
Epoch[1] Batch [1099]	Speed: 3054.772883 samples/sec	accuracy=0.081180	lr=0.150280
Epoch[1] Batch [1149]	Speed: 3005.507248 samples/sec	accuracy=0.083447	lr=0.153477
Epoch[1] Batch [1199]	Speed: 3000.305311 samples/sec	accuracy=0.085692	lr=0.156675
Epoch[1] Batch [1249]	Speed: 3034.256227 samples/sec	accuracy=0.087930	lr=0.159872
[Epoch 1] training: accuracy=0.087968
[Epoch 1] speed: 2993 samples/sec	time cost: 437.569154
[Epoch 1] validation: err-top1=0.859833 err-top5=0.679927
Epoch[2] Batch [49]	Speed: 3231.370538 samples/sec	accuracy=0.151855	lr=0.163133
Epoch[2] Batch [99]	Speed: 2913.888080 samples/sec	accuracy=0.154326	lr=0.166331
Epoch[2] Batch [149]	Speed: 2907.357217 samples/sec	accuracy=0.156732	lr=0.169528
Epoch[2] Batch [199]	Speed: 2870.169130 samples/sec	accuracy=0.159116	lr=0.172726
Epoch[2] Batch [249]	Speed: 2952.350201 samples/sec	accuracy=0.161785	lr=0.175923
Epoch[2] Batch [299]	Speed: 2895.010974 samples/sec	accuracy=0.164108	lr=0.179121
Epoch[2] Batch [349]	Speed: 2972.606915 samples/sec	accuracy=0.167171	lr=0.182318
Epoch[2] Batch [399]	Speed: 2954.593500 samples/sec	accuracy=0.169565	lr=0.185516
Epoch[2] Batch [449]	Speed: 2865.501423 samples/sec	accuracy=0.172159	lr=0.188713
Epoch[2] Batch [499]	Speed: 2913.503030 samples/sec	accuracy=0.174777	lr=0.191910
Epoch[2] Batch [549]	Speed: 2952.856429 samples/sec	accuracy=0.177479	lr=0.195108
Epoch[2] Batch [599]	Speed: 2900.004005 samples/sec	accuracy=0.180286	lr=0.198305
Epoch[2] Batch [649]	Speed: 2881.077329 samples/sec	accuracy=0.182945	lr=0.201503
Epoch[2] Batch [699]	Speed: 2847.839862 samples/sec	accuracy=0.185435	lr=0.204700
Epoch[2] Batch [749]	Speed: 3089.878477 samples/sec	accuracy=0.188034	lr=0.207898
Epoch[2] Batch [799]	Speed: 2919.080485 samples/sec	accuracy=0.190576	lr=0.211095
Epoch[2] Batch [849]	Speed: 2944.733571 samples/sec	accuracy=0.192927	lr=0.214293
Epoch[2] Batch [899]	Speed: 2988.100878 samples/sec	accuracy=0.195592	lr=0.217490
Epoch[2] Batch [949]	Speed: 2862.859313 samples/sec	accuracy=0.198030	lr=0.220687
Epoch[2] Batch [999]	Speed: 2956.424271 samples/sec	accuracy=0.200483	lr=0.223885
Epoch[2] Batch [1049]	Speed: 2938.236523 samples/sec	accuracy=0.202970	lr=0.227082
Epoch[2] Batch [1099]	Speed: 2893.589472 samples/sec	accuracy=0.205319	lr=0.230280
Epoch[2] Batch [1149]	Speed: 2892.979966 samples/sec	accuracy=0.207931	lr=0.233477
Epoch[2] Batch [1199]	Speed: 2934.622626 samples/sec	accuracy=0.210404	lr=0.236675
Epoch[2] Batch [1249]	Speed: 3149.496446 samples/sec	accuracy=0.212831	lr=0.239872
[Epoch 2] training: accuracy=0.212880
[Epoch 2] speed: 2940 samples/sec	time cost: 445.154385
[Epoch 2] validation: err-top1=0.720045 err-top5=0.466737
Epoch[3] Batch [49]	Speed: 2918.681765 samples/sec	accuracy=0.284766	lr=0.243133
Epoch[3] Batch [99]	Speed: 2858.202893 samples/sec	accuracy=0.285654	lr=0.246331
Epoch[3] Batch [149]	Speed: 3044.548597 samples/sec	accuracy=0.285872	lr=0.249528
Epoch[3] Batch [199]	Speed: 2961.103555 samples/sec	accuracy=0.287861	lr=0.252726
Epoch[3] Batch [249]	Speed: 2938.745363 samples/sec	accuracy=0.288906	lr=0.255923
Epoch[3] Batch [299]	Speed: 2971.853569 samples/sec	accuracy=0.290143	lr=0.259121
Epoch[3] Batch [349]	Speed: 2924.598441 samples/sec	accuracy=0.291939	lr=0.262318
Epoch[3] Batch [399]	Speed: 2908.062818 samples/sec	accuracy=0.293816	lr=0.265516
Epoch[3] Batch [449]	Speed: 2918.890951 samples/sec	accuracy=0.295688	lr=0.268713
Epoch[3] Batch [499]	Speed: 3006.213324 samples/sec	accuracy=0.297859	lr=0.271910
Epoch[3] Batch [549]	Speed: 2880.319472 samples/sec	accuracy=0.299272	lr=0.275108
Epoch[3] Batch [599]	Speed: 2999.200549 samples/sec	accuracy=0.301055	lr=0.278305
Epoch[3] Batch [649]	Speed: 2903.066236 samples/sec	accuracy=0.302999	lr=0.281503
Epoch[3] Batch [699]	Speed: 3073.393684 samples/sec	accuracy=0.304754	lr=0.284700
Epoch[3] Batch [749]	Speed: 2945.593504 samples/sec	accuracy=0.306392	lr=0.287898
Epoch[3] Batch [799]	Speed: 2927.623255 samples/sec	accuracy=0.308247	lr=0.291095
Epoch[3] Batch [849]	Speed: 2927.674982 samples/sec	accuracy=0.309841	lr=0.294293
Epoch[3] Batch [899]	Speed: 2944.087437 samples/sec	accuracy=0.311376	lr=0.297490
Epoch[3] Batch [949]	Speed: 3049.256016 samples/sec	accuracy=0.313171	lr=0.300687
Epoch[3] Batch [999]	Speed: 2979.600780 samples/sec	accuracy=0.314605	lr=0.303885
Epoch[3] Batch [1049]	Speed: 3024.517507 samples/sec	accuracy=0.315949	lr=0.307082
Epoch[3] Batch [1099]	Speed: 2904.006771 samples/sec	accuracy=0.317289	lr=0.310280
Epoch[3] Batch [1149]	Speed: 3064.320539 samples/sec	accuracy=0.318899	lr=0.313477
Epoch[3] Batch [1199]	Speed: 3071.825576 samples/sec	accuracy=0.320333	lr=0.316675
Epoch[3] Batch [1249]	Speed: 3047.955756 samples/sec	accuracy=0.321849	lr=0.319872
[Epoch 3] training: accuracy=0.321868
[Epoch 3] speed: 2964 samples/sec	time cost: 441.605841
[Epoch 3] validation: err-top1=0.645049 err-top5=0.380979
Epoch[4] Batch [49]	Speed: 3107.734474 samples/sec	accuracy=0.364180	lr=0.323133
Epoch[4] Batch [99]	Speed: 3157.750642 samples/sec	accuracy=0.366670	lr=0.326331
Epoch[4] Batch [149]	Speed: 2974.034354 samples/sec	accuracy=0.367363	lr=0.329528
Epoch[4] Batch [199]	Speed: 2963.418658 samples/sec	accuracy=0.369243	lr=0.332726
Epoch[4] Batch [249]	Speed: 3018.005775 samples/sec	accuracy=0.369121	lr=0.335923
Epoch[4] Batch [299]	Speed: 2929.118433 samples/sec	accuracy=0.369395	lr=0.339121
Epoch[4] Batch [349]	Speed: 2961.860692 samples/sec	accuracy=0.370753	lr=0.342318
Epoch[4] Batch [399]	Speed: 2882.512330 samples/sec	accuracy=0.371416	lr=0.345516
Epoch[4] Batch [449]	Speed: 2944.096196 samples/sec	accuracy=0.372411	lr=0.348713
Epoch[4] Batch [499]	Speed: 2934.258177 samples/sec	accuracy=0.373172	lr=0.351910
Epoch[4] Batch [549]	Speed: 2813.271348 samples/sec	accuracy=0.373908	lr=0.355108
Epoch[4] Batch [599]	Speed: 2971.636600 samples/sec	accuracy=0.374595	lr=0.358305
Epoch[4] Batch [649]	Speed: 2910.701042 samples/sec	accuracy=0.375712	lr=0.361503
Epoch[4] Batch [699]	Speed: 2891.763671 samples/sec	accuracy=0.376561	lr=0.364700
Epoch[4] Batch [749]	Speed: 3036.303935 samples/sec	accuracy=0.377324	lr=0.367898
Epoch[4] Batch [799]	Speed: 2953.102705 samples/sec	accuracy=0.377888	lr=0.371095
Epoch[4] Batch [849]	Speed: 2903.008782 samples/sec	accuracy=0.378821	lr=0.374293
Epoch[4] Batch [899]	Speed: 2922.223637 samples/sec	accuracy=0.379678	lr=0.377490
Epoch[4] Batch [949]	Speed: 2999.696407 samples/sec	accuracy=0.380626	lr=0.380687
Epoch[4] Batch [999]	Speed: 2967.465658 samples/sec	accuracy=0.381177	lr=0.383885
Epoch[4] Batch [1049]	Speed: 2945.298509 samples/sec	accuracy=0.382121	lr=0.387082
Epoch[4] Batch [1099]	Speed: 2899.410154 samples/sec	accuracy=0.383035	lr=0.390280
Epoch[4] Batch [1149]	Speed: 2919.848558 samples/sec	accuracy=0.383782	lr=0.393477
Epoch[4] Batch [1199]	Speed: 2933.734618 samples/sec	accuracy=0.384547	lr=0.396675
Epoch[4] Batch [1249]	Speed: 3057.932068 samples/sec	accuracy=0.385399	lr=0.399872
[Epoch 4] training: accuracy=0.385414
[Epoch 4] speed: 2956 samples/sec	time cost: 443.273284
[Epoch 4] validation: err-top1=0.623226 err-top5=0.360551
Epoch[5] Batch [49]	Speed: 3219.068953 samples/sec	accuracy=0.411836	lr=0.400000
Epoch[5] Batch [99]	Speed: 2961.111925 samples/sec	accuracy=0.412559	lr=0.400000
Epoch[5] Batch [149]	Speed: 3132.226452 samples/sec	accuracy=0.413490	lr=0.399999
Epoch[5] Batch [199]	Speed: 2955.173493 samples/sec	accuracy=0.414624	lr=0.399998
Epoch[5] Batch [249]	Speed: 3030.220008 samples/sec	accuracy=0.414633	lr=0.399997
Epoch[5] Batch [299]	Speed: 2902.242953 samples/sec	accuracy=0.415365	lr=0.399996
Epoch[5] Batch [349]	Speed: 3000.533363 samples/sec	accuracy=0.416071	lr=0.399994
Epoch[5] Batch [399]	Speed: 2973.719964 samples/sec	accuracy=0.416887	lr=0.399992
Epoch[5] Batch [449]	Speed: 3015.853650 samples/sec	accuracy=0.417413	lr=0.399990
Epoch[5] Batch [499]	Speed: 2874.332526 samples/sec	accuracy=0.417787	lr=0.399988
Epoch[5] Batch [549]	Speed: 2847.512808 samples/sec	accuracy=0.418540	lr=0.399986
Epoch[5] Batch [599]	Speed: 2972.754642 samples/sec	accuracy=0.419316	lr=0.399983
Epoch[5] Batch [649]	Speed: 3039.961634 samples/sec	accuracy=0.419835	lr=0.399980
Epoch[5] Batch [699]	Speed: 2871.917058 samples/sec	accuracy=0.420167	lr=0.399977
Epoch[5] Batch [749]	Speed: 2854.466301 samples/sec	accuracy=0.421104	lr=0.399973
Epoch[5] Batch [799]	Speed: 2891.730650 samples/sec	accuracy=0.421704	lr=0.399970
Epoch[5] Batch [849]	Speed: 2982.556030 samples/sec	accuracy=0.422400	lr=0.399966
Epoch[5] Batch [899]	Speed: 2928.401381 samples/sec	accuracy=0.423049	lr=0.399961
Epoch[5] Batch [949]	Speed: 2980.172933 samples/sec	accuracy=0.423788	lr=0.399957
Epoch[5] Batch [999]	Speed: 3024.322382 samples/sec	accuracy=0.424583	lr=0.399952
Epoch[5] Batch [1049]	Speed: 3206.044435 samples/sec	accuracy=0.425362	lr=0.399948
Epoch[5] Batch [1099]	Speed: 2909.841842 samples/sec	accuracy=0.425922	lr=0.399942
Epoch[5] Batch [1149]	Speed: 3042.747934 samples/sec	accuracy=0.426472	lr=0.399937
Epoch[5] Batch [1199]	Speed: 2975.184497 samples/sec	accuracy=0.427235	lr=0.399931
Epoch[5] Batch [1249]	Speed: 3317.945491 samples/sec	accuracy=0.427990	lr=0.399926
[Epoch 5] training: accuracy=0.428017
[Epoch 5] speed: 2989 samples/sec	time cost: 438.321939
[Epoch 5] validation: err-top1=0.528483 err-top5=0.266683
Epoch[6] Batch [49]	Speed: 3057.872022 samples/sec	accuracy=0.449395	lr=0.399919
Epoch[6] Batch [99]	Speed: 2919.487571 samples/sec	accuracy=0.452559	lr=0.399913
Epoch[6] Batch [149]	Speed: 2990.376496 samples/sec	accuracy=0.451335	lr=0.399907
Epoch[6] Batch [199]	Speed: 3110.295357 samples/sec	accuracy=0.451948	lr=0.399900
Epoch[6] Batch [249]	Speed: 3023.313973 samples/sec	accuracy=0.452605	lr=0.399893
Epoch[6] Batch [299]	Speed: 2998.119035 samples/sec	accuracy=0.452943	lr=0.399885
Epoch[6] Batch [349]	Speed: 2904.129300 samples/sec	accuracy=0.452969	lr=0.399878
Epoch[6] Batch [399]	Speed: 2941.096232 samples/sec	accuracy=0.453525	lr=0.399870
Epoch[6] Batch [449]	Speed: 2938.402163 samples/sec	accuracy=0.454332	lr=0.399862
Epoch[6] Batch [499]	Speed: 2976.427854 samples/sec	accuracy=0.454781	lr=0.399854
Epoch[6] Batch [549]	Speed: 2904.301761 samples/sec	accuracy=0.455044	lr=0.399846
Epoch[6] Batch [599]	Speed: 2992.686318 samples/sec	accuracy=0.455063	lr=0.399837
Epoch[6] Batch [649]	Speed: 2939.492275 samples/sec	accuracy=0.455464	lr=0.399828
Epoch[6] Batch [699]	Speed: 2950.380647 samples/sec	accuracy=0.455830	lr=0.399819
Epoch[6] Batch [749]	Speed: 3020.465803 samples/sec	accuracy=0.456362	lr=0.399809
Epoch[6] Batch [799]	Speed: 2942.251068 samples/sec	accuracy=0.456951	lr=0.399800
Epoch[6] Batch [849]	Speed: 3099.376626 samples/sec	accuracy=0.457283	lr=0.399790
Epoch[6] Batch [899]	Speed: 2959.695963 samples/sec	accuracy=0.457658	lr=0.399780
Epoch[6] Batch [949]	Speed: 3009.020626 samples/sec	accuracy=0.458010	lr=0.399769
Epoch[6] Batch [999]	Speed: 2956.704075 samples/sec	accuracy=0.458428	lr=0.399759
Epoch[6] Batch [1049]	Speed: 2958.992119 samples/sec	accuracy=0.458737	lr=0.399748
Epoch[6] Batch [1099]	Speed: 2919.644078 samples/sec	accuracy=0.459074	lr=0.399737
Epoch[6] Batch [1149]	Speed: 2935.454433 samples/sec	accuracy=0.459500	lr=0.399725
Epoch[6] Batch [1199]	Speed: 3031.906043 samples/sec	accuracy=0.459950	lr=0.399714
Epoch[6] Batch [1249]	Speed: 3304.822562 samples/sec	accuracy=0.460147	lr=0.399702
[Epoch 6] training: accuracy=0.460166
[Epoch 6] speed: 2986 samples/sec	time cost: 438.440140
[Epoch 6] validation: err-top1=0.493543 err-top5=0.240055
Epoch[7] Batch [49]	Speed: 3031.475350 samples/sec	accuracy=0.478066	lr=0.399690
Epoch[7] Batch [99]	Speed: 2907.285424 samples/sec	accuracy=0.479131	lr=0.399677
Epoch[7] Batch [149]	Speed: 2944.993073 samples/sec	accuracy=0.479245	lr=0.399665
Epoch[7] Batch [199]	Speed: 2909.469804 samples/sec	accuracy=0.478984	lr=0.399652
Epoch[7] Batch [249]	Speed: 3083.215283 samples/sec	accuracy=0.478770	lr=0.399639
Epoch[7] Batch [299]	Speed: 2991.416385 samples/sec	accuracy=0.478669	lr=0.399626
Epoch[7] Batch [349]	Speed: 2945.215984 samples/sec	accuracy=0.478211	lr=0.399613
Epoch[7] Batch [399]	Speed: 3031.460972 samples/sec	accuracy=0.478499	lr=0.399599
Epoch[7] Batch [449]	Speed: 3109.483493 samples/sec	accuracy=0.478331	lr=0.399585
Epoch[7] Batch [499]	Speed: 3059.177621 samples/sec	accuracy=0.478250	lr=0.399571
Epoch[7] Batch [549]	Speed: 2891.942533 samples/sec	accuracy=0.478592	lr=0.399556
Epoch[7] Batch [599]	Speed: 3075.720149 samples/sec	accuracy=0.478341	lr=0.399542
Epoch[7] Batch [649]	Speed: 2836.064813 samples/sec	accuracy=0.478221	lr=0.399527
Epoch[7] Batch [699]	Speed: 2999.220403 samples/sec	accuracy=0.478634	lr=0.399512
Epoch[7] Batch [749]	Speed: 3051.097684 samples/sec	accuracy=0.478533	lr=0.399496
Epoch[7] Batch [799]	Speed: 2949.012994 samples/sec	accuracy=0.478636	lr=0.399481
Epoch[7] Batch [849]	Speed: 2903.171416 samples/sec	accuracy=0.478783	lr=0.399465
Epoch[7] Batch [899]	Speed: 2955.796715 samples/sec	accuracy=0.478981	lr=0.399449
Epoch[7] Batch [949]	Speed: 3085.928078 samples/sec	accuracy=0.479281	lr=0.399432
Epoch[7] Batch [999]	Speed: 3039.784734 samples/sec	accuracy=0.479367	lr=0.399416
Epoch[7] Batch [1049]	Speed: 2876.389438 samples/sec	accuracy=0.479692	lr=0.399399
Epoch[7] Batch [1099]	Speed: 2919.426370 samples/sec	accuracy=0.479788	lr=0.399382
Epoch[7] Batch [1149]	Speed: 2904.851802 samples/sec	accuracy=0.480020	lr=0.399365
Epoch[7] Batch [1199]	Speed: 2913.809955 samples/sec	accuracy=0.480236	lr=0.399347
Epoch[7] Batch [1249]	Speed: 3138.361004 samples/sec	accuracy=0.480539	lr=0.399329
[Epoch 7] training: accuracy=0.480574
[Epoch 7] speed: 2978 samples/sec	time cost: 440.491884
[Epoch 7] validation: err-top1=0.483578 err-top5=0.229811
Epoch[8] Batch [49]	Speed: 3110.155761 samples/sec	accuracy=0.498418	lr=0.399311
Epoch[8] Batch [99]	Speed: 3014.403735 samples/sec	accuracy=0.498037	lr=0.399293
Epoch[8] Batch [149]	Speed: 2936.357495 samples/sec	accuracy=0.496452	lr=0.399274
Epoch[8] Batch [199]	Speed: 2837.729438 samples/sec	accuracy=0.496484	lr=0.399256
Epoch[8] Batch [249]	Speed: 2961.280277 samples/sec	accuracy=0.496375	lr=0.399237
Epoch[8] Batch [299]	Speed: 3090.175887 samples/sec	accuracy=0.495439	lr=0.399218
Epoch[8] Batch [349]	Speed: 3140.245282 samples/sec	accuracy=0.495226	lr=0.399198
Epoch[8] Batch [399]	Speed: 2945.463371 samples/sec	accuracy=0.494895	lr=0.399179
Epoch[8] Batch [449]	Speed: 2930.016881 samples/sec	accuracy=0.494787	lr=0.399159
Epoch[8] Batch [499]	Speed: 2911.962346 samples/sec	accuracy=0.494771	lr=0.399138
Epoch[8] Batch [549]	Speed: 3047.736443 samples/sec	accuracy=0.495105	lr=0.399118
Epoch[8] Batch [599]	Speed: 3058.280065 samples/sec	accuracy=0.494967	lr=0.399098
Epoch[8] Batch [649]	Speed: 2907.911527 samples/sec	accuracy=0.495107	lr=0.399077
Epoch[8] Batch [699]	Speed: 2928.199894 samples/sec	accuracy=0.495052	lr=0.399056
Epoch[8] Batch [749]	Speed: 3228.505477 samples/sec	accuracy=0.495254	lr=0.399034
Epoch[8] Batch [799]	Speed: 2877.001531 samples/sec	accuracy=0.495245	lr=0.399013
Epoch[8] Batch [849]	Speed: 2887.580567 samples/sec	accuracy=0.495248	lr=0.398991
Epoch[8] Batch [899]	Speed: 2953.701898 samples/sec	accuracy=0.495267	lr=0.398969
Epoch[8] Batch [949]	Speed: 2927.345735 samples/sec	accuracy=0.495432	lr=0.398947
Epoch[8] Batch [999]	Speed: 2919.267545 samples/sec	accuracy=0.495471	lr=0.398924
Epoch[8] Batch [1049]	Speed: 3154.641356 samples/sec	accuracy=0.495675	lr=0.398901
Epoch[8] Batch [1099]	Speed: 2919.440658 samples/sec	accuracy=0.495866	lr=0.398878
Epoch[8] Batch [1149]	Speed: 2859.652053 samples/sec	accuracy=0.495985	lr=0.398855
Epoch[8] Batch [1199]	Speed: 2907.628007 samples/sec	accuracy=0.496160	lr=0.398832
Epoch[8] Batch [1249]	Speed: 3184.376882 samples/sec	accuracy=0.496309	lr=0.398808
[Epoch 8] training: accuracy=0.496309
[Epoch 8] speed: 2980 samples/sec	time cost: 439.860767
[Epoch 8] validation: err-top1=0.473792 err-top5=0.221460
Epoch[9] Batch [49]	Speed: 3176.690620 samples/sec	accuracy=0.513848	lr=0.398784
Epoch[9] Batch [99]	Speed: 3016.218782 samples/sec	accuracy=0.512832	lr=0.398760
Epoch[9] Batch [149]	Speed: 3004.543713 samples/sec	accuracy=0.510632	lr=0.398735
Epoch[9] Batch [199]	Speed: 2951.825806 samples/sec	accuracy=0.510317	lr=0.398710
Epoch[9] Batch [249]	Speed: 2908.402787 samples/sec	accuracy=0.509273	lr=0.398686
Epoch[9] Batch [299]	Speed: 2990.984579 samples/sec	accuracy=0.508750	lr=0.398660
Epoch[9] Batch [349]	Speed: 3004.249991 samples/sec	accuracy=0.508541	lr=0.398635
Epoch[9] Batch [399]	Speed: 2997.651063 samples/sec	accuracy=0.508948	lr=0.398610
Epoch[9] Batch [449]	Speed: 2954.210744 samples/sec	accuracy=0.508214	lr=0.398584
Epoch[9] Batch [499]	Speed: 3074.318223 samples/sec	accuracy=0.507932	lr=0.398558
Epoch[9] Batch [549]	Speed: 3100.366818 samples/sec	accuracy=0.507830	lr=0.398531
Epoch[9] Batch [599]	Speed: 3099.704591 samples/sec	accuracy=0.508032	lr=0.398505
Epoch[9] Batch [649]	Speed: 3092.334214 samples/sec	accuracy=0.508102	lr=0.398478
Epoch[9] Batch [699]	Speed: 3058.089356 samples/sec	accuracy=0.507619	lr=0.398451
Epoch[9] Batch [749]	Speed: 2914.426926 samples/sec	accuracy=0.507622	lr=0.398424
Epoch[9] Batch [799]	Speed: 3010.396957 samples/sec	accuracy=0.507627	lr=0.398396
Epoch[9] Batch [849]	Speed: 3016.682441 samples/sec	accuracy=0.507410	lr=0.398369
Epoch[9] Batch [899]	Speed: 2989.975005 samples/sec	accuracy=0.507546	lr=0.398341
Epoch[9] Batch [949]	Speed: 2900.428547 samples/sec	accuracy=0.507513	lr=0.398312
Epoch[9] Batch [999]	Speed: 3211.399441 samples/sec	accuracy=0.507467	lr=0.398284
Epoch[9] Batch [1049]	Speed: 2949.199009 samples/sec	accuracy=0.507507	lr=0.398255
Epoch[9] Batch [1099]	Speed: 3146.304966 samples/sec	accuracy=0.507550	lr=0.398226
Epoch[9] Batch [1149]	Speed: 3064.447393 samples/sec	accuracy=0.507565	lr=0.398197
Epoch[9] Batch [1199]	Speed: 2892.812744 samples/sec	accuracy=0.507645	lr=0.398168
Epoch[9] Batch [1249]	Speed: 3089.345425 samples/sec	accuracy=0.507632	lr=0.398138
[Epoch 9] training: accuracy=0.507644
[Epoch 9] speed: 3020 samples/sec	time cost: 434.173691
[Epoch 9] validation: err-top1=0.499621 err-top5=0.246851
Epoch[10] Batch [49]	Speed: 2874.367075 samples/sec	accuracy=0.521309	lr=0.398108
Epoch[10] Batch [99]	Speed: 3006.706369 samples/sec	accuracy=0.519453	lr=0.398078
Epoch[10] Batch [149]	Speed: 2884.807733 samples/sec	accuracy=0.518079	lr=0.398048
Epoch[10] Batch [199]	Speed: 3003.538535 samples/sec	accuracy=0.517202	lr=0.398017
Epoch[10] Batch [249]	Speed: 2916.914858 samples/sec	accuracy=0.517309	lr=0.397986
Epoch[10] Batch [299]	Speed: 3024.407610 samples/sec	accuracy=0.517406	lr=0.397955
Epoch[10] Batch [349]	Speed: 3003.094614 samples/sec	accuracy=0.517640	lr=0.397924
Epoch[10] Batch [399]	Speed: 2929.368158 samples/sec	accuracy=0.517124	lr=0.397892
Epoch[10] Batch [449]	Speed: 3052.106930 samples/sec	accuracy=0.517003	lr=0.397861
Epoch[10] Batch [499]	Speed: 3070.366089 samples/sec	accuracy=0.516770	lr=0.397829
Epoch[10] Batch [549]	Speed: 3052.563118 samples/sec	accuracy=0.516305	lr=0.397796
Epoch[10] Batch [599]	Speed: 2975.582479 samples/sec	accuracy=0.516463	lr=0.397764
Epoch[10] Batch [649]	Speed: 2904.130321 samples/sec	accuracy=0.516097	lr=0.397731
Epoch[10] Batch [699]	Speed: 3061.300312 samples/sec	accuracy=0.516317	lr=0.397698
Epoch[10] Batch [749]	Speed: 3010.127531 samples/sec	accuracy=0.516397	lr=0.397665
Epoch[10] Batch [799]	Speed: 2946.841441 samples/sec	accuracy=0.516523	lr=0.397632
Epoch[10] Batch [849]	Speed: 2881.842895 samples/sec	accuracy=0.516461	lr=0.397598
Epoch[10] Batch [899]	Speed: 2910.223558 samples/sec	accuracy=0.516466	lr=0.397564
Epoch[10] Batch [949]	Speed: 2913.332952 samples/sec	accuracy=0.516449	lr=0.397530
Epoch[10] Batch [999]	Speed: 2912.489261 samples/sec	accuracy=0.516382	lr=0.397496
Epoch[10] Batch [1049]	Speed: 2963.862216 samples/sec	accuracy=0.516247	lr=0.397461
Epoch[10] Batch [1099]	Speed: 2992.370808 samples/sec	accuracy=0.516231	lr=0.397427
Epoch[10] Batch [1149]	Speed: 2973.853347 samples/sec	accuracy=0.516230	lr=0.397392
Epoch[10] Batch [1199]	Speed: 2991.548693 samples/sec	accuracy=0.516276	lr=0.397356
Epoch[10] Batch [1249]	Speed: 3220.401017 samples/sec	accuracy=0.516451	lr=0.397321
[Epoch 10] training: accuracy=0.516464
[Epoch 10] speed: 2975 samples/sec	time cost: 440.362168
[Epoch 10] validation: err-top1=0.456294 err-top5=0.210399
Epoch[11] Batch [49]	Speed: 3146.133310 samples/sec	accuracy=0.527793	lr=0.397284
Epoch[11] Batch [99]	Speed: 3098.613014 samples/sec	accuracy=0.528262	lr=0.397248
Epoch[11] Batch [149]	Speed: 3067.313216 samples/sec	accuracy=0.525944	lr=0.397212
Epoch[11] Batch [199]	Speed: 3086.441187 samples/sec	accuracy=0.525571	lr=0.397176
Epoch[11] Batch [249]	Speed: 2968.025980 samples/sec	accuracy=0.525133	lr=0.397139
Epoch[11] Batch [299]	Speed: 2850.517004 samples/sec	accuracy=0.525339	lr=0.397102
Epoch[11] Batch [349]	Speed: 2961.238667 samples/sec	accuracy=0.524916	lr=0.397065
Epoch[11] Batch [399]	Speed: 2939.160164 samples/sec	accuracy=0.524753	lr=0.397028
Epoch[11] Batch [449]	Speed: 2929.141047 samples/sec	accuracy=0.524627	lr=0.396990
Epoch[11] Batch [499]	Speed: 2968.207140 samples/sec	accuracy=0.524330	lr=0.396952
Epoch[11] Batch [549]	Speed: 2926.794723 samples/sec	accuracy=0.524267	lr=0.396914
Epoch[11] Batch [599]	Speed: 2994.679328 samples/sec	accuracy=0.524382	lr=0.396876
Epoch[11] Batch [649]	Speed: 3069.354824 samples/sec	accuracy=0.524129	lr=0.396837
Epoch[11] Batch [699]	Speed: 2904.791331 samples/sec	accuracy=0.524069	lr=0.396798
Epoch[11] Batch [749]	Speed: 2977.142660 samples/sec	accuracy=0.524443	lr=0.396759
Epoch[11] Batch [799]	Speed: 2950.122261 samples/sec	accuracy=0.524348	lr=0.396720
Epoch[11] Batch [849]	Speed: 3105.246682 samples/sec	accuracy=0.524210	lr=0.396680
Epoch[11] Batch [899]	Speed: 2992.359967 samples/sec	accuracy=0.524207	lr=0.396641
Epoch[11] Batch [949]	Speed: 3074.736566 samples/sec	accuracy=0.523984	lr=0.396601
Epoch[11] Batch [999]	Speed: 2950.019933 samples/sec	accuracy=0.523887	lr=0.396561
Epoch[11] Batch [1049]	Speed: 2925.201820 samples/sec	accuracy=0.523963	lr=0.396520
Epoch[11] Batch [1099]	Speed: 3030.148989 samples/sec	accuracy=0.524034	lr=0.396479
Epoch[11] Batch [1149]	Speed: 3132.573333 samples/sec	accuracy=0.524079	lr=0.396439
Epoch[11] Batch [1199]	Speed: 3009.821241 samples/sec	accuracy=0.524097	lr=0.396397
Epoch[11] Batch [1249]	Speed: 3073.275412 samples/sec	accuracy=0.524045	lr=0.396356
[Epoch 11] training: accuracy=0.524053
[Epoch 11] speed: 3001 samples/sec	time cost: 436.773556
[Epoch 11] validation: err-top1=0.443156 err-top5=0.200806
Epoch[12] Batch [49]	Speed: 2967.398452 samples/sec	accuracy=0.536172	lr=0.396314
Epoch[12] Batch [99]	Speed: 2949.626125 samples/sec	accuracy=0.534951	lr=0.396272
Epoch[12] Batch [149]	Speed: 2968.946406 samples/sec	accuracy=0.534564	lr=0.396230
Epoch[12] Batch [199]	Speed: 2870.674274 samples/sec	accuracy=0.534199	lr=0.396187
Epoch[12] Batch [249]	Speed: 3020.568786 samples/sec	accuracy=0.533652	lr=0.396145
Epoch[12] Batch [299]	Speed: 3126.881774 samples/sec	accuracy=0.532767	lr=0.396102
Epoch[12] Batch [349]	Speed: 2916.654655 samples/sec	accuracy=0.532871	lr=0.396059
Epoch[12] Batch [399]	Speed: 2963.238205 samples/sec	accuracy=0.532351	lr=0.396016
Epoch[12] Batch [449]	Speed: 2910.719743 samples/sec	accuracy=0.531964	lr=0.395972
Epoch[12] Batch [499]	Speed: 2933.186686 samples/sec	accuracy=0.531881	lr=0.395928
Epoch[12] Batch [549]	Speed: 3061.621621 samples/sec	accuracy=0.531648	lr=0.395885
Epoch[12] Batch [599]	Speed: 2944.053897 samples/sec	accuracy=0.531904	lr=0.395840
Epoch[12] Batch [649]	Speed: 3027.896093 samples/sec	accuracy=0.531923	lr=0.395796
Epoch[12] Batch [699]	Speed: 3108.319556 samples/sec	accuracy=0.531881	lr=0.395751
Epoch[12] Batch [749]	Speed: 2920.837344 samples/sec	accuracy=0.531655	lr=0.395706
Epoch[12] Batch [799]	Speed: 3012.328743 samples/sec	accuracy=0.531583	lr=0.395661
Epoch[12] Batch [849]	Speed: 2972.090437 samples/sec	accuracy=0.531310	lr=0.395616
Epoch[12] Batch [899]	Speed: 2954.360957 samples/sec	accuracy=0.531151	lr=0.395570
Epoch[12] Batch [949]	Speed: 3015.967924 samples/sec	accuracy=0.531130	lr=0.395524
Epoch[12] Batch [999]	Speed: 3039.705348 samples/sec	accuracy=0.531106	lr=0.395478
Epoch[12] Batch [1049]	Speed: 2923.366562 samples/sec	accuracy=0.530922	lr=0.395432
Epoch[12] Batch [1099]	Speed: 2975.562071 samples/sec	accuracy=0.530753	lr=0.395386
Epoch[12] Batch [1149]	Speed: 2974.898918 samples/sec	accuracy=0.530598	lr=0.395339
Epoch[12] Batch [1199]	Speed: 2893.287182 samples/sec	accuracy=0.530566	lr=0.395292
Epoch[12] Batch [1249]	Speed: 3139.995867 samples/sec	accuracy=0.530670	lr=0.395245
[Epoch 12] training: accuracy=0.530665
[Epoch 12] speed: 2979 samples/sec	time cost: 440.286441
[Epoch 12] validation: err-top1=0.478276 err-top5=0.230309
Epoch[13] Batch [49]	Speed: 3043.164457 samples/sec	accuracy=0.539746	lr=0.395196
Epoch[13] Batch [99]	Speed: 2865.378729 samples/sec	accuracy=0.540791	lr=0.395149
Epoch[13] Batch [149]	Speed: 2896.534902 samples/sec	accuracy=0.539616	lr=0.395101
Epoch[13] Batch [199]	Speed: 2975.951905 samples/sec	accuracy=0.538960	lr=0.395052
Epoch[13] Batch [249]	Speed: 2920.731198 samples/sec	accuracy=0.538457	lr=0.395004
Epoch[13] Batch [299]	Speed: 3046.171414 samples/sec	accuracy=0.537881	lr=0.394955
Epoch[13] Batch [349]	Speed: 2863.654787 samples/sec	accuracy=0.537355	lr=0.394907
Epoch[13] Batch [399]	Speed: 2868.818553 samples/sec	accuracy=0.537002	lr=0.394858
Epoch[13] Batch [449]	Speed: 3117.219128 samples/sec	accuracy=0.537192	lr=0.394808
Epoch[13] Batch [499]	Speed: 3020.949085 samples/sec	accuracy=0.537424	lr=0.394759
Epoch[13] Batch [549]	Speed: 2924.124907 samples/sec	accuracy=0.537383	lr=0.394709
Epoch[13] Batch [599]	Speed: 2938.279539 samples/sec	accuracy=0.537428	lr=0.394659
Epoch[13] Batch [649]	Speed: 2917.916248 samples/sec	accuracy=0.536840	lr=0.394609
Epoch[13] Batch [699]	Speed: 2848.229736 samples/sec	accuracy=0.536790	lr=0.394558
Epoch[13] Batch [749]	Speed: 3008.749044 samples/sec	accuracy=0.536505	lr=0.394507
Epoch[13] Batch [799]	Speed: 2911.438107 samples/sec	accuracy=0.536429	lr=0.394457
Epoch[13] Batch [849]	Speed: 3001.718825 samples/sec	accuracy=0.536333	lr=0.394405
Epoch[13] Batch [899]	Speed: 3028.668427 samples/sec	accuracy=0.536051	lr=0.394354
Epoch[13] Batch [949]	Speed: 2930.135258 samples/sec	accuracy=0.535876	lr=0.394302
Epoch[13] Batch [999]	Speed: 2945.604332 samples/sec	accuracy=0.535831	lr=0.394250
Epoch[13] Batch [1049]	Speed: 3010.934899 samples/sec	accuracy=0.535977	lr=0.394198
Epoch[13] Batch [1099]	Speed: 2965.179317 samples/sec	accuracy=0.535898	lr=0.394146
Epoch[13] Batch [1149]	Speed: 2883.403393 samples/sec	accuracy=0.535900	lr=0.394093
Epoch[13] Batch [1199]	Speed: 3000.153576 samples/sec	accuracy=0.535929	lr=0.394041
Epoch[13] Batch [1249]	Speed: 3170.812674 samples/sec	accuracy=0.535971	lr=0.393988
[Epoch 13] training: accuracy=0.535984
[Epoch 13] speed: 2960 samples/sec	time cost: 442.377903
[Epoch 13] validation: err-top1=0.449059 err-top5=0.202866
Epoch[14] Batch [49]	Speed: 3184.067531 samples/sec	accuracy=0.543574	lr=0.393933
Epoch[14] Batch [99]	Speed: 3002.886790 samples/sec	accuracy=0.546836	lr=0.393880
Epoch[14] Batch [149]	Speed: 3015.868600 samples/sec	accuracy=0.545853	lr=0.393826
Epoch[14] Batch [199]	Speed: 3082.550009 samples/sec	accuracy=0.544492	lr=0.393772
Epoch[14] Batch [249]	Speed: 3071.331371 samples/sec	accuracy=0.544734	lr=0.393718
Epoch[14] Batch [299]	Speed: 3011.594620 samples/sec	accuracy=0.544157	lr=0.393664
Epoch[14] Batch [349]	Speed: 3034.399340 samples/sec	accuracy=0.543192	lr=0.393609
Epoch[14] Batch [399]	Speed: 2891.032640 samples/sec	accuracy=0.542690	lr=0.393554
Epoch[14] Batch [449]	Speed: 3045.166045 samples/sec	accuracy=0.542292	lr=0.393499
Epoch[14] Batch [499]	Speed: 2873.603399 samples/sec	accuracy=0.542141	lr=0.393444
Epoch[14] Batch [549]	Speed: 3065.181219 samples/sec	accuracy=0.541816	lr=0.393388
Epoch[14] Batch [599]	Speed: 2848.639707 samples/sec	accuracy=0.541475	lr=0.393332
Epoch[14] Batch [649]	Speed: 2942.173994 samples/sec	accuracy=0.541149	lr=0.393276
Epoch[14] Batch [699]	Speed: 2915.280724 samples/sec	accuracy=0.540965	lr=0.393220
Epoch[14] Batch [749]	Speed: 2833.738111 samples/sec	accuracy=0.540745	lr=0.393163
Epoch[14] Batch [799]	Speed: 3177.206154 samples/sec	accuracy=0.540748	lr=0.393107
Epoch[14] Batch [849]	Speed: 3058.830029 samples/sec	accuracy=0.540656	lr=0.393050
Epoch[14] Batch [899]	Speed: 2917.959702 samples/sec	accuracy=0.540588	lr=0.392993
Epoch[14] Batch [949]	Speed: 3187.690111 samples/sec	accuracy=0.540447	lr=0.392935
Epoch[14] Batch [999]	Speed: 3108.224989 samples/sec	accuracy=0.540350	lr=0.392877
Epoch[14] Batch [1049]	Speed: 2319.231007 samples/sec	accuracy=0.540424	lr=0.392820
Epoch[14] Batch [1099]	Speed: 2817.041696 samples/sec	accuracy=0.540368	lr=0.392762
Epoch[14] Batch [1149]	Speed: 2875.376303 samples/sec	accuracy=0.540274	lr=0.392703
Epoch[14] Batch [1199]	Speed: 2041.861113 samples/sec	accuracy=0.540311	lr=0.392645
Epoch[14] Batch [1249]	Speed: 2866.024165 samples/sec	accuracy=0.540272	lr=0.392586
[Epoch 14] training: accuracy=0.540244
[Epoch 14] speed: 2890 samples/sec	time cost: 469.617697
[Epoch 14] validation: err-top1=0.451889 err-top5=0.200255
Epoch[15] Batch [49]	Speed: 3008.469882 samples/sec	accuracy=0.553809	lr=0.392526
Epoch[15] Batch [99]	Speed: 2823.311790 samples/sec	accuracy=0.552373	lr=0.392466
Epoch[15] Batch [149]	Speed: 3065.960963 samples/sec	accuracy=0.550697	lr=0.392407
Epoch[15] Batch [199]	Speed: 3030.881364 samples/sec	accuracy=0.548745	lr=0.392347
Epoch[15] Batch [249]	Speed: 2419.504226 samples/sec	accuracy=0.548336	lr=0.392287
Epoch[15] Batch [299]	Speed: 2908.852920 samples/sec	accuracy=0.548245	lr=0.392227
Epoch[15] Batch [349]	Speed: 2443.708891 samples/sec	accuracy=0.547606	lr=0.392167
Epoch[15] Batch [399]	Speed: 2561.774147 samples/sec	accuracy=0.547297	lr=0.392106
Epoch[15] Batch [449]	Speed: 2525.996820 samples/sec	accuracy=0.547283	lr=0.392045
Epoch[15] Batch [499]	Speed: 2918.338079 samples/sec	accuracy=0.546746	lr=0.391984
Epoch[15] Batch [549]	Speed: 3012.667072 samples/sec	accuracy=0.546697	lr=0.391923
Epoch[15] Batch [599]	Speed: 1769.718436 samples/sec	accuracy=0.546468	lr=0.391861
Epoch[15] Batch [649]	Speed: 2414.020097 samples/sec	accuracy=0.546394	lr=0.391799
Epoch[15] Batch [699]	Speed: 2348.972110 samples/sec	accuracy=0.546044	lr=0.391737
Epoch[15] Batch [749]	Speed: 3220.723168 samples/sec	accuracy=0.545697	lr=0.391675
Epoch[15] Batch [799]	Speed: 2864.779788 samples/sec	accuracy=0.545796	lr=0.391613
Epoch[15] Batch [849]	Speed: 1910.060119 samples/sec	accuracy=0.545750	lr=0.391550
Epoch[15] Batch [899]	Speed: 2364.688439 samples/sec	accuracy=0.545524	lr=0.391487
Epoch[15] Batch [949]	Speed: 1364.120358 samples/sec	accuracy=0.545581	lr=0.391424
Epoch[15] Batch [999]	Speed: 2644.177115 samples/sec	accuracy=0.545349	lr=0.391361
Epoch[15] Batch [1049]	Speed: 3028.741727 samples/sec	accuracy=0.545152	lr=0.391297
Epoch[15] Batch [1099]	Speed: 2934.769129 samples/sec	accuracy=0.544896	lr=0.391233
Epoch[15] Batch [1149]	Speed: 2786.063861 samples/sec	accuracy=0.544948	lr=0.391169
Epoch[15] Batch [1199]	Speed: 2918.359535 samples/sec	accuracy=0.544772	lr=0.391105
Epoch[15] Batch [1249]	Speed: 3436.926663 samples/sec	accuracy=0.544666	lr=0.391040
[Epoch 15] training: accuracy=0.544674
[Epoch 15] speed: 2560 samples/sec	time cost: 530.343278
[Epoch 15] validation: err-top1=0.439174 err-top5=0.196967
Epoch[16] Batch [49]	Speed: 2184.100197 samples/sec	accuracy=0.557617	lr=0.390974
Epoch[16] Batch [99]	Speed: 2967.158026 samples/sec	accuracy=0.553838	lr=0.390909
Epoch[16] Batch [149]	Speed: 3121.570187 samples/sec	accuracy=0.555065	lr=0.390844
Epoch[16] Batch [199]	Speed: 3064.529388 samples/sec	accuracy=0.555034	lr=0.390779
Epoch[16] Batch [249]	Speed: 2772.227161 samples/sec	accuracy=0.554676	lr=0.390713
Epoch[16] Batch [299]	Speed: 2135.008792 samples/sec	accuracy=0.553721	lr=0.390647
Epoch[16] Batch [349]	Speed: 3389.033981 samples/sec	accuracy=0.553061	lr=0.390581
Epoch[16] Batch [399]	Speed: 3224.073344 samples/sec	accuracy=0.552400	lr=0.390515
Epoch[16] Batch [449]	Speed: 2398.018849 samples/sec	accuracy=0.552346	lr=0.390448
Epoch[16] Batch [499]	Speed: 2530.308613 samples/sec	accuracy=0.551850	lr=0.390381
Epoch[16] Batch [549]	Speed: 2361.897715 samples/sec	accuracy=0.551747	lr=0.390314
Epoch[16] Batch [599]	Speed: 2613.480020 samples/sec	accuracy=0.551527	lr=0.390247
Epoch[16] Batch [649]	Speed: 2480.411528 samples/sec	accuracy=0.551270	lr=0.390180
Epoch[16] Batch [699]	Speed: 3031.388097 samples/sec	accuracy=0.550981	lr=0.390112
Epoch[16] Batch [749]	Speed: 2789.370778 samples/sec	accuracy=0.550755	lr=0.390044
Epoch[16] Batch [799]	Speed: 2779.031437 samples/sec	accuracy=0.550703	lr=0.389976
Epoch[16] Batch [849]	Speed: 2279.482943 samples/sec	accuracy=0.550419	lr=0.389907
Epoch[16] Batch [899]	Speed: 3152.858818 samples/sec	accuracy=0.550303	lr=0.389839
Epoch[16] Batch [949]	Speed: 2703.358120 samples/sec	accuracy=0.550214	lr=0.389770
Epoch[16] Batch [999]	Speed: 2310.569836 samples/sec	accuracy=0.549962	lr=0.389701
Epoch[16] Batch [1049]	Speed: 2573.459062 samples/sec	accuracy=0.549656	lr=0.389632
Epoch[16] Batch [1099]	Speed: 2810.025357 samples/sec	accuracy=0.549555	lr=0.389562
Epoch[16] Batch [1149]	Speed: 2866.725074 samples/sec	accuracy=0.549390	lr=0.389492
Epoch[16] Batch [1199]	Speed: 3338.654774 samples/sec	accuracy=0.549329	lr=0.389422
Epoch[16] Batch [1249]	Speed: 2700.207037 samples/sec	accuracy=0.549132	lr=0.389352
[Epoch 16] training: accuracy=0.549133
[Epoch 16] speed: 2695 samples/sec	time cost: 485.070804
[Epoch 16] validation: err-top1=0.437121 err-top5=0.189951
Epoch[17] Batch [49]	Speed: 1926.223671 samples/sec	accuracy=0.558828	lr=0.389280
Epoch[17] Batch [99]	Speed: 3163.890120 samples/sec	accuracy=0.556436	lr=0.389210
Epoch[17] Batch [149]	Speed: 2643.976250 samples/sec	accuracy=0.555820	lr=0.389139
Epoch[17] Batch [199]	Speed: 3077.198974 samples/sec	accuracy=0.555391	lr=0.389068
Epoch[17] Batch [249]	Speed: 2648.087185 samples/sec	accuracy=0.555105	lr=0.388996
Epoch[17] Batch [299]	Speed: 2828.198019 samples/sec	accuracy=0.555280	lr=0.388925
Epoch[17] Batch [349]	Speed: 3379.015795 samples/sec	accuracy=0.555265	lr=0.388853
Epoch[17] Batch [399]	Speed: 2460.261855 samples/sec	accuracy=0.555073	lr=0.388781
Epoch[17] Batch [449]	Speed: 3019.501018 samples/sec	accuracy=0.554805	lr=0.388709
Epoch[17] Batch [499]	Speed: 3339.831471 samples/sec	accuracy=0.554629	lr=0.388636
Epoch[17] Batch [549]	Speed: 2720.686885 samples/sec	accuracy=0.554577	lr=0.388564
Epoch[17] Batch [599]	Speed: 2896.188249 samples/sec	accuracy=0.554331	lr=0.388491
Epoch[17] Batch [649]	Speed: 2633.702728 samples/sec	accuracy=0.554117	lr=0.388418
Epoch[17] Batch [699]	Speed: 2857.077502 samples/sec	accuracy=0.553788	lr=0.388344
Epoch[17] Batch [749]	Speed: 2871.906112 samples/sec	accuracy=0.553371	lr=0.388271
Epoch[17] Batch [799]	Speed: 2421.444471 samples/sec	accuracy=0.553173	lr=0.388197
Epoch[17] Batch [849]	Speed: 2473.479517 samples/sec	accuracy=0.552856	lr=0.388123
Epoch[17] Batch [899]	Speed: 2696.964284 samples/sec	accuracy=0.552693	lr=0.388049
Epoch[17] Batch [949]	Speed: 3037.121456 samples/sec	accuracy=0.552473	lr=0.387974
Epoch[17] Batch [999]	Speed: 2974.503467 samples/sec	accuracy=0.552461	lr=0.387900
Epoch[17] Batch [1049]	Speed: 2857.352693 samples/sec	accuracy=0.552388	lr=0.387825
Epoch[17] Batch [1099]	Speed: 3248.800216 samples/sec	accuracy=0.552328	lr=0.387750
Epoch[17] Batch [1149]	Speed: 3079.472164 samples/sec	accuracy=0.552358	lr=0.387674
Epoch[17] Batch [1199]	Speed: 2655.581960 samples/sec	accuracy=0.552416	lr=0.387599
Epoch[17] Batch [1249]	Speed: 3082.649127 samples/sec	accuracy=0.552205	lr=0.387523
[Epoch 17] training: accuracy=0.552224
[Epoch 17] speed: 2796 samples/sec	time cost: 472.720068
[Epoch 17] validation: err-top1=0.436239 err-top5=0.195089
Epoch[18] Batch [49]	Speed: 3379.633189 samples/sec	accuracy=0.559883	lr=0.387445
Epoch[18] Batch [99]	Speed: 3013.499948 samples/sec	accuracy=0.559502	lr=0.387369
Epoch[18] Batch [149]	Speed: 2987.731922 samples/sec	accuracy=0.558978	lr=0.387293
Epoch[18] Batch [199]	Speed: 3105.908269 samples/sec	accuracy=0.558813	lr=0.387216
Epoch[18] Batch [249]	Speed: 3145.981491 samples/sec	accuracy=0.558652	lr=0.387139
Epoch[18] Batch [299]	Speed: 2389.400164 samples/sec	accuracy=0.558060	lr=0.387062
Epoch[18] Batch [349]	Speed: 2605.065674 samples/sec	accuracy=0.557709	lr=0.386984
Epoch[18] Batch [399]	Speed: 1822.756776 samples/sec	accuracy=0.557480	lr=0.386907
Epoch[18] Batch [449]	Speed: 3156.641564 samples/sec	accuracy=0.557391	lr=0.386829
Epoch[18] Batch [499]	Speed: 3267.926097 samples/sec	accuracy=0.557275	lr=0.386751
Epoch[18] Batch [549]	Speed: 2946.688636 samples/sec	accuracy=0.556717	lr=0.386673
Epoch[18] Batch [599]	Speed: 3159.194483 samples/sec	accuracy=0.556606	lr=0.386594
Epoch[18] Batch [649]	Speed: 2896.066662 samples/sec	accuracy=0.556349	lr=0.386515
Epoch[18] Batch [699]	Speed: 3189.057560 samples/sec	accuracy=0.555951	lr=0.386436
Epoch[18] Batch [749]	Speed: 2892.269395 samples/sec	accuracy=0.555764	lr=0.386357
Epoch[18] Batch [799]	Speed: 3116.817599 samples/sec	accuracy=0.555721	lr=0.386278
Epoch[18] Batch [849]	Speed: 2421.570074 samples/sec	accuracy=0.555377	lr=0.386198
Epoch[18] Batch [899]	Speed: 2819.686025 samples/sec	accuracy=0.554825	lr=0.386119
Epoch[18] Batch [949]	Speed: 2817.912703 samples/sec	accuracy=0.554690	lr=0.386038
Epoch[18] Batch [999]	Speed: 2768.030674 samples/sec	accuracy=0.554904	lr=0.385958
Epoch[18] Batch [1049]	Speed: 3187.736956 samples/sec	accuracy=0.554753	lr=0.385878
Epoch[18] Batch [1099]	Speed: 2423.231140 samples/sec	accuracy=0.554545	lr=0.385797
Epoch[18] Batch [1149]	Speed: 3115.360160 samples/sec	accuracy=0.554419	lr=0.385716
Epoch[18] Batch [1199]	Speed: 2910.749174 samples/sec	accuracy=0.554268	lr=0.385635
Epoch[18] Batch [1249]	Speed: 2840.416749 samples/sec	accuracy=0.554184	lr=0.385554
[Epoch 18] training: accuracy=0.554182
[Epoch 18] speed: 2843 samples/sec	time cost: 460.554898
[Epoch 18] validation: err-top1=0.453085 err-top5=0.206413
Epoch[19] Batch [49]	Speed: 2911.574014 samples/sec	accuracy=0.563125	lr=0.385470
Epoch[19] Batch [99]	Speed: 2777.795108 samples/sec	accuracy=0.561914	lr=0.385388
Epoch[19] Batch [149]	Speed: 3086.166183 samples/sec	accuracy=0.563275	lr=0.385306
Epoch[19] Batch [199]	Speed: 3265.007740 samples/sec	accuracy=0.562695	lr=0.385224
Epoch[19] Batch [249]	Speed: 2945.655161 samples/sec	accuracy=0.562613	lr=0.385142
Epoch[19] Batch [299]	Speed: 3100.262439 samples/sec	accuracy=0.562695	lr=0.385059
Epoch[19] Batch [349]	Speed: 2913.390815 samples/sec	accuracy=0.562026	lr=0.384976
Epoch[19] Batch [399]	Speed: 2976.409166 samples/sec	accuracy=0.561709	lr=0.384893
Epoch[19] Batch [449]	Speed: 2974.080278 samples/sec	accuracy=0.561198	lr=0.384810
Epoch[19] Batch [499]	Speed: 3093.196180 samples/sec	accuracy=0.560701	lr=0.384726
Epoch[19] Batch [549]	Speed: 2957.897562 samples/sec	accuracy=0.560252	lr=0.384642
Epoch[19] Batch [599]	Speed: 3056.177783 samples/sec	accuracy=0.560236	lr=0.384558
Epoch[19] Batch [649]	Speed: 3051.989206 samples/sec	accuracy=0.559803	lr=0.384474
Epoch[19] Batch [699]	Speed: 2941.647768 samples/sec	accuracy=0.559876	lr=0.384389
Epoch[19] Batch [749]	Speed: 3034.585134 samples/sec	accuracy=0.559676	lr=0.384305
Epoch[19] Batch [799]	Speed: 3062.427898 samples/sec	accuracy=0.559304	lr=0.384220
Epoch[19] Batch [849]	Speed: 2933.374917 samples/sec	accuracy=0.559025	lr=0.384135
Epoch[19] Batch [899]	Speed: 2943.059335 samples/sec	accuracy=0.558627	lr=0.384049
Epoch[19] Batch [949]	Speed: 3285.802566 samples/sec	accuracy=0.558743	lr=0.383964
Epoch[19] Batch [999]	Speed: 2994.664127 samples/sec	accuracy=0.558593	lr=0.383878
Epoch[19] Batch [1049]	Speed: 2956.620096 samples/sec	accuracy=0.558339	lr=0.383792
Epoch[19] Batch [1099]	Speed: 2937.165379 samples/sec	accuracy=0.558127	lr=0.383706
Epoch[19] Batch [1149]	Speed: 2906.163453 samples/sec	accuracy=0.558057	lr=0.383619
Epoch[19] Batch [1199]	Speed: 2777.361416 samples/sec	accuracy=0.557909	lr=0.383533
Epoch[19] Batch [1249]	Speed: 2564.544491 samples/sec	accuracy=0.557766	lr=0.383446
[Epoch 19] training: accuracy=0.557787
[Epoch 19] speed: 2968 samples/sec	time cost: 441.283886
[Epoch 19] validation: err-top1=0.432159 err-top5=0.185985
Epoch[20] Batch [49]	Speed: 2944.236703 samples/sec	accuracy=0.565234	lr=0.383357
Epoch[20] Batch [99]	Speed: 3065.351724 samples/sec	accuracy=0.563115	lr=0.383270
Epoch[20] Batch [149]	Speed: 2769.244254 samples/sec	accuracy=0.563750	lr=0.383182
Epoch[20] Batch [199]	Speed: 3293.823988 samples/sec	accuracy=0.564370	lr=0.383094
Epoch[20] Batch [249]	Speed: 2859.897080 samples/sec	accuracy=0.564863	lr=0.383006
Epoch[20] Batch [299]	Speed: 2880.722811 samples/sec	accuracy=0.564375	lr=0.382918
Epoch[20] Batch [349]	Speed: 2966.742826 samples/sec	accuracy=0.563158	lr=0.382830
Epoch[20] Batch [399]	Speed: 2934.449392 samples/sec	accuracy=0.562988	lr=0.382741
Epoch[20] Batch [449]	Speed: 3190.306694 samples/sec	accuracy=0.562747	lr=0.382652
Epoch[20] Batch [499]	Speed: 2897.132306 samples/sec	accuracy=0.562498	lr=0.382563
Epoch[20] Batch [549]	Speed: 2947.886958 samples/sec	accuracy=0.562251	lr=0.382474
Epoch[20] Batch [599]	Speed: 2747.892955 samples/sec	accuracy=0.562031	lr=0.382384
Epoch[20] Batch [649]	Speed: 3038.089584 samples/sec	accuracy=0.562153	lr=0.382295
Epoch[20] Batch [699]	Speed: 3171.012786 samples/sec	accuracy=0.561844	lr=0.382205
Epoch[20] Batch [749]	Speed: 2950.267762 samples/sec	accuracy=0.561549	lr=0.382115
Epoch[20] Batch [799]	Speed: 2938.803033 samples/sec	accuracy=0.561526	lr=0.382024
Epoch[20] Batch [849]	Speed: 2998.898363 samples/sec	accuracy=0.561480	lr=0.381934
Epoch[20] Batch [899]	Speed: 3035.866971 samples/sec	accuracy=0.561529	lr=0.381843
Epoch[20] Batch [949]	Speed: 2882.494338 samples/sec	accuracy=0.561161	lr=0.381752
Epoch[20] Batch [999]	Speed: 2487.235777 samples/sec	accuracy=0.561180	lr=0.381661
Epoch[20] Batch [1049]	Speed: 2890.038335 samples/sec	accuracy=0.560884	lr=0.381569
Epoch[20] Batch [1099]	Speed: 2965.925801 samples/sec	accuracy=0.560735	lr=0.381477
Epoch[20] Batch [1149]	Speed: 3162.225631 samples/sec	accuracy=0.560800	lr=0.381385
Epoch[20] Batch [1199]	Speed: 2771.977711 samples/sec	accuracy=0.560630	lr=0.381293
Epoch[20] Batch [1249]	Speed: 3217.394842 samples/sec	accuracy=0.560634	lr=0.381201
[Epoch 20] training: accuracy=0.560619
[Epoch 20] speed: 2948 samples/sec	time cost: 444.412688
[Epoch 20] validation: err-top1=0.414541 err-top5=0.176559
Epoch[21] Batch [49]	Speed: 2748.120224 samples/sec	accuracy=0.569063	lr=0.381107
Epoch[21] Batch [99]	Speed: 3236.658003 samples/sec	accuracy=0.569590	lr=0.381014
Epoch[21] Batch [149]	Speed: 2947.775235 samples/sec	accuracy=0.568737	lr=0.380921
Epoch[21] Batch [199]	Speed: 3020.436872 samples/sec	accuracy=0.567773	lr=0.380828
Epoch[21] Batch [249]	Speed: 2897.233656 samples/sec	accuracy=0.567730	lr=0.380734
Epoch[21] Batch [299]	Speed: 3086.287667 samples/sec	accuracy=0.566556	lr=0.380641
Epoch[21] Batch [349]	Speed: 3134.762851 samples/sec	accuracy=0.565767	lr=0.380547
Epoch[21] Batch [399]	Speed: 2738.200311 samples/sec	accuracy=0.565483	lr=0.380453
Epoch[21] Batch [449]	Speed: 2933.410699 samples/sec	accuracy=0.565299	lr=0.380359
Epoch[21] Batch [499]	Speed: 3252.757656 samples/sec	accuracy=0.565500	lr=0.380264
Epoch[21] Batch [549]	Speed: 3245.410204 samples/sec	accuracy=0.564970	lr=0.380169
Epoch[21] Batch [599]	Speed: 3198.156847 samples/sec	accuracy=0.565034	lr=0.380074
Epoch[21] Batch [649]	Speed: 3002.502670 samples/sec	accuracy=0.564758	lr=0.379979
Epoch[21] Batch [699]	Speed: 3005.239453 samples/sec	accuracy=0.564563	lr=0.379884
Epoch[21] Batch [749]	Speed: 3150.486943 samples/sec	accuracy=0.564366	lr=0.379788
Epoch[21] Batch [799]	Speed: 3095.311750 samples/sec	accuracy=0.564178	lr=0.379693
Epoch[21] Batch [849]	Speed: 3194.895365 samples/sec	accuracy=0.564064	lr=0.379597
Epoch[21] Batch [899]	Speed: 2956.787571 samples/sec	accuracy=0.563768	lr=0.379501
Epoch[21] Batch [949]	Speed: 2794.935120 samples/sec	accuracy=0.563542	lr=0.379404
Epoch[21] Batch [999]	Speed: 3231.592664 samples/sec	accuracy=0.563281	lr=0.379307
Epoch[21] Batch [1049]	Speed: 2948.197932 samples/sec	accuracy=0.562862	lr=0.379211
Epoch[21] Batch [1099]	Speed: 2973.206063 samples/sec	accuracy=0.562821	lr=0.379114
Epoch[21] Batch [1149]	Speed: 3011.976591 samples/sec	accuracy=0.562796	lr=0.379016
Epoch[21] Batch [1199]	Speed: 2890.248142 samples/sec	accuracy=0.562589	lr=0.378919
Epoch[21] Batch [1249]	Speed: 3088.986634 samples/sec	accuracy=0.562680	lr=0.378821
[Epoch 21] training: accuracy=0.562681
[Epoch 21] speed: 3022 samples/sec	time cost: 435.024980
[Epoch 21] validation: err-top1=0.412787 err-top5=0.173131
Epoch[22] Batch [49]	Speed: 3241.847398 samples/sec	accuracy=0.573574	lr=0.378721
Epoch[22] Batch [99]	Speed: 2942.959351 samples/sec	accuracy=0.572197	lr=0.378623
Epoch[22] Batch [149]	Speed: 2953.540865 samples/sec	accuracy=0.570736	lr=0.378525
Epoch[22] Batch [199]	Speed: 3022.631452 samples/sec	accuracy=0.570874	lr=0.378426
Epoch[22] Batch [249]	Speed: 3020.173588 samples/sec	accuracy=0.570902	lr=0.378328
Epoch[22] Batch [299]	Speed: 2916.101483 samples/sec	accuracy=0.569805	lr=0.378229
Epoch[22] Batch [349]	Speed: 2986.689857 samples/sec	accuracy=0.569478	lr=0.378129
Epoch[22] Batch [399]	Speed: 3036.154117 samples/sec	accuracy=0.568997	lr=0.378030
Epoch[22] Batch [449]	Speed: 2976.038224 samples/sec	accuracy=0.568494	lr=0.377930
Epoch[22] Batch [499]	Speed: 2877.782903 samples/sec	accuracy=0.568492	lr=0.377831
Epoch[22] Batch [549]	Speed: 2960.871496 samples/sec	accuracy=0.568148	lr=0.377731
Epoch[22] Batch [599]	Speed: 2908.591987 samples/sec	accuracy=0.567975	lr=0.377630
Epoch[22] Batch [649]	Speed: 3048.462416 samples/sec	accuracy=0.567670	lr=0.377530
Epoch[22] Batch [699]	Speed: 2911.174224 samples/sec	accuracy=0.567349	lr=0.377429
Epoch[22] Batch [749]	Speed: 2951.717923 samples/sec	accuracy=0.566952	lr=0.377328
Epoch[22] Batch [799]	Speed: 3068.700472 samples/sec	accuracy=0.566661	lr=0.377227
Epoch[22] Batch [849]	Speed: 2963.092935 samples/sec	accuracy=0.566463	lr=0.377126
Epoch[22] Batch [899]	Speed: 2996.864311 samples/sec	accuracy=0.566287	lr=0.377024
Epoch[22] Batch [949]	Speed: 3054.654042 samples/sec	accuracy=0.566013	lr=0.376923
Epoch[22] Batch [999]	Speed: 3087.747238 samples/sec	accuracy=0.565930	lr=0.376821
Epoch[22] Batch [1049]	Speed: 2885.018564 samples/sec	accuracy=0.565516	lr=0.376719
Epoch[22] Batch [1099]	Speed: 2952.501605 samples/sec	accuracy=0.565366	lr=0.376616
Epoch[22] Batch [1149]	Speed: 3067.204436 samples/sec	accuracy=0.565261	lr=0.376514
Epoch[22] Batch [1199]	Speed: 2986.485003 samples/sec	accuracy=0.565117	lr=0.376411
Epoch[22] Batch [1249]	Speed: 3100.924322 samples/sec	accuracy=0.565086	lr=0.376308
[Epoch 22] training: accuracy=0.565082
[Epoch 22] speed: 2992 samples/sec	time cost: 437.927529
[Epoch 22] validation: err-top1=0.425343 err-top5=0.181023
Epoch[23] Batch [49]	Speed: 2986.913932 samples/sec	accuracy=0.575723	lr=0.376203
Epoch[23] Batch [99]	Speed: 3083.202534 samples/sec	accuracy=0.576143	lr=0.376099
Epoch[23] Batch [149]	Speed: 3065.643775 samples/sec	accuracy=0.574661	lr=0.375996
Epoch[23] Batch [199]	Speed: 3121.687213 samples/sec	accuracy=0.573374	lr=0.375892
Epoch[23] Batch [249]	Speed: 2970.312607 samples/sec	accuracy=0.572738	lr=0.375788
Epoch[23] Batch [299]	Speed: 2954.640574 samples/sec	accuracy=0.572246	lr=0.375683
Epoch[23] Batch [349]	Speed: 3102.088955 samples/sec	accuracy=0.571543	lr=0.375579
Epoch[23] Batch [399]	Speed: 2992.155877 samples/sec	accuracy=0.570935	lr=0.375474
Epoch[23] Batch [449]	Speed: 2896.254924 samples/sec	accuracy=0.570699	lr=0.375369
Epoch[23] Batch [499]	Speed: 2888.434168 samples/sec	accuracy=0.570506	lr=0.375264
Epoch[23] Batch [549]	Speed: 2893.743410 samples/sec	accuracy=0.570268	lr=0.375159
Epoch[23] Batch [599]	Speed: 3008.657108 samples/sec	accuracy=0.569723	lr=0.375053
Epoch[23] Batch [649]	Speed: 2925.730469 samples/sec	accuracy=0.569590	lr=0.374948
Epoch[23] Batch [699]	Speed: 2893.091783 samples/sec	accuracy=0.568955	lr=0.374842
Epoch[23] Batch [749]	Speed: 2984.986215 samples/sec	accuracy=0.568858	lr=0.374736
Epoch[23] Batch [799]	Speed: 2915.626540 samples/sec	accuracy=0.568568	lr=0.374629
Epoch[23] Batch [849]	Speed: 3109.079453 samples/sec	accuracy=0.568325	lr=0.374523
Epoch[23] Batch [899]	Speed: 3158.222239 samples/sec	accuracy=0.568155	lr=0.374416
Epoch[23] Batch [949]	Speed: 2988.128445 samples/sec	accuracy=0.568004	lr=0.374309
Epoch[23] Batch [999]	Speed: 2991.054316 samples/sec	accuracy=0.567765	lr=0.374202
Epoch[23] Batch [1049]	Speed: 2918.716514 samples/sec	accuracy=0.567698	lr=0.374095
Epoch[23] Batch [1099]	Speed: 3031.574078 samples/sec	accuracy=0.567557	lr=0.373987
Epoch[23] Batch [1149]	Speed: 3010.114620 samples/sec	accuracy=0.567361	lr=0.373879
Epoch[23] Batch [1199]	Speed: 3118.138715 samples/sec	accuracy=0.567207	lr=0.373771
Epoch[23] Batch [1249]	Speed: 2990.262986 samples/sec	accuracy=0.567157	lr=0.373663
[Epoch 23] training: accuracy=0.567175
[Epoch 23] speed: 2995 samples/sec	time cost: 439.407694
[Epoch 23] validation: err-top1=0.404948 err-top5=0.172221
Epoch[24] Batch [49]	Speed: 3194.619467 samples/sec	accuracy=0.580313	lr=0.373552
Epoch[24] Batch [99]	Speed: 3090.200344 samples/sec	accuracy=0.575391	lr=0.373444
Epoch[24] Batch [149]	Speed: 3039.721010 samples/sec	accuracy=0.573646	lr=0.373335
Epoch[24] Batch [199]	Speed: 3051.088017 samples/sec	accuracy=0.573374	lr=0.373226
Epoch[24] Batch [249]	Speed: 3078.565574 samples/sec	accuracy=0.572871	lr=0.373117
Epoch[24] Batch [299]	Speed: 2868.191318 samples/sec	accuracy=0.573076	lr=0.373007
Epoch[24] Batch [349]	Speed: 3119.070622 samples/sec	accuracy=0.572776	lr=0.372897
Epoch[24] Batch [399]	Speed: 3124.140797 samples/sec	accuracy=0.572432	lr=0.372788
Epoch[24] Batch [449]	Speed: 2955.070976 samples/sec	accuracy=0.571888	lr=0.372678
Epoch[24] Batch [499]	Speed: 2923.424665 samples/sec	accuracy=0.571898	lr=0.372567
Epoch[24] Batch [549]	Speed: 3003.712838 samples/sec	accuracy=0.571422	lr=0.372457
Epoch[24] Batch [599]	Speed: 2938.210392 samples/sec	accuracy=0.571141	lr=0.372346
Epoch[24] Batch [649]	Speed: 2875.556378 samples/sec	accuracy=0.571128	lr=0.372235
Epoch[24] Batch [699]	Speed: 3072.884024 samples/sec	accuracy=0.571037	lr=0.372124
Epoch[24] Batch [749]	Speed: 2887.413309 samples/sec	accuracy=0.570874	lr=0.372013
Epoch[24] Batch [799]	Speed: 2836.090245 samples/sec	accuracy=0.570848	lr=0.371901
Epoch[24] Batch [849]	Speed: 2938.260001 samples/sec	accuracy=0.570571	lr=0.371790
Epoch[24] Batch [899]	Speed: 2924.416870 samples/sec	accuracy=0.570409	lr=0.371678
Epoch[24] Batch [949]	Speed: 2872.568017 samples/sec	accuracy=0.570206	lr=0.371565
Epoch[24] Batch [999]	Speed: 3009.354544 samples/sec	accuracy=0.569999	lr=0.371453
Epoch[24] Batch [1049]	Speed: 2998.801961 samples/sec	accuracy=0.569847	lr=0.371341
Epoch[24] Batch [1099]	Speed: 2835.923430 samples/sec	accuracy=0.569706	lr=0.371228
Epoch[24] Batch [1149]	Speed: 3153.351504 samples/sec	accuracy=0.569595	lr=0.371115
Epoch[24] Batch [1199]	Speed: 3210.630616 samples/sec	accuracy=0.569545	lr=0.371002
Epoch[24] Batch [1249]	Speed: 3131.573928 samples/sec	accuracy=0.569406	lr=0.370888
[Epoch 24] training: accuracy=0.569396
[Epoch 24] speed: 2998 samples/sec	time cost: 438.610614
[Epoch 24] validation: err-top1=0.440868 err-top5=0.198063
Epoch[25] Batch [49]	Speed: 3139.167462 samples/sec	accuracy=0.577832	lr=0.370773
Epoch[25] Batch [99]	Speed: 2863.476391 samples/sec	accuracy=0.580752	lr=0.370659
Epoch[25] Batch [149]	Speed: 3078.738499 samples/sec	accuracy=0.579258	lr=0.370545
Epoch[25] Batch [199]	Speed: 3070.655891 samples/sec	accuracy=0.578232	lr=0.370431
Epoch[25] Batch [249]	Speed: 2971.944503 samples/sec	accuracy=0.577371	lr=0.370316
Epoch[25] Batch [299]	Speed: 2996.644009 samples/sec	accuracy=0.577018	lr=0.370202
Epoch[25] Batch [349]	Speed: 3085.042458 samples/sec	accuracy=0.576473	lr=0.370087
Epoch[25] Batch [399]	Speed: 2980.772404 samples/sec	accuracy=0.575630	lr=0.369972
Epoch[25] Batch [449]	Speed: 3018.077456 samples/sec	accuracy=0.575163	lr=0.369857
Epoch[25] Batch [499]	Speed: 3072.065729 samples/sec	accuracy=0.574750	lr=0.369741
Epoch[25] Batch [549]	Speed: 2918.616869 samples/sec	accuracy=0.574657	lr=0.369626
Epoch[25] Batch [599]	Speed: 2926.119755 samples/sec	accuracy=0.574111	lr=0.369510
Epoch[25] Batch [649]	Speed: 3019.279159 samples/sec	accuracy=0.573685	lr=0.369394
Epoch[25] Batch [699]	Speed: 3063.683324 samples/sec	accuracy=0.573701	lr=0.369278
Epoch[25] Batch [749]	Speed: 2949.139877 samples/sec	accuracy=0.573638	lr=0.369161
Epoch[25] Batch [799]	Speed: 2924.262519 samples/sec	accuracy=0.573470	lr=0.369045
Epoch[25] Batch [849]	Speed: 2901.284000 samples/sec	accuracy=0.573381	lr=0.368928
Epoch[25] Batch [899]	Speed: 3002.391093 samples/sec	accuracy=0.572974	lr=0.368811
Epoch[25] Batch [949]	Speed: 2959.682298 samples/sec	accuracy=0.572499	lr=0.368694
Epoch[25] Batch [999]	Speed: 2940.458417 samples/sec	accuracy=0.572468	lr=0.368576
Epoch[25] Batch [1049]	Speed: 2960.078224 samples/sec	accuracy=0.572323	lr=0.368459
Epoch[25] Batch [1099]	Speed: 2967.017044 samples/sec	accuracy=0.572006	lr=0.368341
Epoch[25] Batch [1149]	Speed: 2992.705044 samples/sec	accuracy=0.571642	lr=0.368223
Epoch[25] Batch [1199]	Speed: 3225.447341 samples/sec	accuracy=0.571365	lr=0.368105
Epoch[25] Batch [1249]	Speed: 3107.918562 samples/sec	accuracy=0.571253	lr=0.367986
[Epoch 25] training: accuracy=0.571259
[Epoch 25] speed: 3001 samples/sec	time cost: 437.288314
[Epoch 25] validation: err-top1=0.410156 err-top5=0.172732
Epoch[26] Batch [49]	Speed: 3131.134589 samples/sec	accuracy=0.581855	lr=0.367865
Epoch[26] Batch [99]	Speed: 2920.325195 samples/sec	accuracy=0.580781	lr=0.367747
Epoch[26] Batch [149]	Speed: 3106.920889 samples/sec	accuracy=0.580430	lr=0.367628
Epoch[26] Batch [199]	Speed: 3026.893833 samples/sec	accuracy=0.579595	lr=0.367508
Epoch[26] Batch [249]	Speed: 2899.694226 samples/sec	accuracy=0.579270	lr=0.367389
Epoch[26] Batch [299]	Speed: 2936.471847 samples/sec	accuracy=0.578750	lr=0.367269
Epoch[26] Batch [349]	Speed: 2996.755035 samples/sec	accuracy=0.578382	lr=0.367150
Epoch[26] Batch [399]	Speed: 2979.353207 samples/sec	accuracy=0.577607	lr=0.367030
Epoch[26] Batch [449]	Speed: 3045.213502 samples/sec	accuracy=0.577333	lr=0.366909
Epoch[26] Batch [499]	Speed: 2992.579639 samples/sec	accuracy=0.577262	lr=0.366789
Epoch[26] Batch [549]	Speed: 3031.146389 samples/sec	accuracy=0.576928	lr=0.366668
Epoch[26] Batch [599]	Speed: 2962.222387 samples/sec	accuracy=0.576777	lr=0.366548
Epoch[26] Batch [649]	Speed: 2901.081954 samples/sec	accuracy=0.575977	lr=0.366426
Epoch[26] Batch [699]	Speed: 2990.448620 samples/sec	accuracy=0.575891	lr=0.366305
Epoch[26] Batch [749]	Speed: 3016.383502 samples/sec	accuracy=0.575439	lr=0.366184
Epoch[26] Batch [799]	Speed: 2939.020181 samples/sec	accuracy=0.575144	lr=0.366062
Epoch[26] Batch [849]	Speed: 2972.922098 samples/sec	accuracy=0.575057	lr=0.365940
Epoch[26] Batch [899]	Speed: 3071.361197 samples/sec	accuracy=0.574865	lr=0.365819
Epoch[26] Batch [949]	Speed: 3019.516303 samples/sec	accuracy=0.574746	lr=0.365696
Epoch[26] Batch [999]	Speed: 3060.038815 samples/sec	accuracy=0.574432	lr=0.365574
Epoch[26] Batch [1049]	Speed: 2970.016584 samples/sec	accuracy=0.574310	lr=0.365451
Epoch[26] Batch [1099]	Speed: 3050.006408 samples/sec	accuracy=0.574142	lr=0.365329
Epoch[26] Batch [1149]	Speed: 2910.766100 samples/sec	accuracy=0.573956	lr=0.365206
Epoch[26] Batch [1199]	Speed: 2940.608886 samples/sec	accuracy=0.573689	lr=0.365082
Epoch[26] Batch [1249]	Speed: 3213.168564 samples/sec	accuracy=0.573520	lr=0.364959
[Epoch 26] training: accuracy=0.573512
[Epoch 26] speed: 2999 samples/sec	time cost: 437.453031
[Epoch 26] validation: err-top1=0.403938 err-top5=0.167471
Epoch[27] Batch [49]	Speed: 3024.626859 samples/sec	accuracy=0.591289	lr=0.364833
Epoch[27] Batch [99]	Speed: 3003.182682 samples/sec	accuracy=0.589189	lr=0.364709
Epoch[27] Batch [149]	Speed: 3132.049614 samples/sec	accuracy=0.585254	lr=0.364585
Epoch[27] Batch [199]	Speed: 2908.649149 samples/sec	accuracy=0.583306	lr=0.364461
Epoch[27] Batch [249]	Speed: 3143.401837 samples/sec	accuracy=0.582371	lr=0.364337
Epoch[27] Batch [299]	Speed: 2991.163677 samples/sec	accuracy=0.581784	lr=0.364212
Epoch[27] Batch [349]	Speed: 2929.096300 samples/sec	accuracy=0.580932	lr=0.364087
Epoch[27] Batch [399]	Speed: 2967.185207 samples/sec	accuracy=0.580073	lr=0.363962
Epoch[27] Batch [449]	Speed: 3009.620288 samples/sec	accuracy=0.579444	lr=0.363837
Epoch[27] Batch [499]	Speed: 2977.640582 samples/sec	accuracy=0.578957	lr=0.363712
Epoch[27] Batch [549]	Speed: 2908.719039 samples/sec	accuracy=0.578544	lr=0.363586
Epoch[27] Batch [599]	Speed: 2925.744380 samples/sec	accuracy=0.578003	lr=0.363461
Epoch[27] Batch [649]	Speed: 2907.354934 samples/sec	accuracy=0.577704	lr=0.363335
Epoch[27] Batch [699]	Speed: 2951.470012 samples/sec	accuracy=0.577425	lr=0.363209
Epoch[27] Batch [749]	Speed: 2972.391522 samples/sec	accuracy=0.577327	lr=0.363082
Epoch[27] Batch [799]	Speed: 2975.892768 samples/sec	accuracy=0.577092	lr=0.362956
Epoch[27] Batch [849]	Speed: 2906.178949 samples/sec	accuracy=0.576816	lr=0.362829
Epoch[27] Batch [899]	Speed: 2842.618067 samples/sec	accuracy=0.576414	lr=0.362702
Epoch[27] Batch [949]	Speed: 2898.392514 samples/sec	accuracy=0.576118	lr=0.362575
Epoch[27] Batch [999]	Speed: 2901.638970 samples/sec	accuracy=0.575891	lr=0.362448
Epoch[27] Batch [1049]	Speed: 2945.969625 samples/sec	accuracy=0.575805	lr=0.362320
Epoch[27] Batch [1099]	Speed: 2901.741420 samples/sec	accuracy=0.575721	lr=0.362193
Epoch[27] Batch [1149]	Speed: 2918.812398 samples/sec	accuracy=0.575449	lr=0.362065
Epoch[27] Batch [1199]	Speed: 3069.304418 samples/sec	accuracy=0.575208	lr=0.361937
Epoch[27] Batch [1249]	Speed: 3124.308743 samples/sec	accuracy=0.575053	lr=0.361809
[Epoch 27] training: accuracy=0.575055
[Epoch 27] speed: 2965 samples/sec	time cost: 442.506343
[Epoch 27] validation: err-top1=0.411810 err-top5=0.174267
Epoch[28] Batch [49]	Speed: 3054.392883 samples/sec	accuracy=0.582969	lr=0.361678
Epoch[28] Batch [99]	Speed: 2920.635585 samples/sec	accuracy=0.582959	lr=0.361549
Epoch[28] Batch [149]	Speed: 3060.133002 samples/sec	accuracy=0.582689	lr=0.361420
Epoch[28] Batch [199]	Speed: 2931.641368 samples/sec	accuracy=0.582378	lr=0.361291
Epoch[28] Batch [249]	Speed: 2911.981733 samples/sec	accuracy=0.582277	lr=0.361162
Epoch[28] Batch [299]	Speed: 3120.459893 samples/sec	accuracy=0.582435	lr=0.361032
Epoch[28] Batch [349]	Speed: 3035.291769 samples/sec	accuracy=0.581370	lr=0.360903
Epoch[28] Batch [399]	Speed: 2966.979741 samples/sec	accuracy=0.581113	lr=0.360773
Epoch[28] Batch [449]	Speed: 3053.301464 samples/sec	accuracy=0.580840	lr=0.360643
Epoch[28] Batch [499]	Speed: 2964.823162 samples/sec	accuracy=0.580639	lr=0.360513
Epoch[28] Batch [549]	Speed: 2998.699369 samples/sec	accuracy=0.580128	lr=0.360382
Epoch[28] Batch [599]	Speed: 2971.892064 samples/sec	accuracy=0.579845	lr=0.360252
Epoch[28] Batch [649]	Speed: 2965.417702 samples/sec	accuracy=0.579369	lr=0.360121
Epoch[28] Batch [699]	Speed: 3085.791236 samples/sec	accuracy=0.579078	lr=0.359990
Epoch[28] Batch [749]	Speed: 3008.873699 samples/sec	accuracy=0.578943	lr=0.359859
Epoch[28] Batch [799]	Speed: 2951.184019 samples/sec	accuracy=0.578558	lr=0.359728
Epoch[28] Batch [849]	Speed: 3044.400856 samples/sec	accuracy=0.578500	lr=0.359596
Epoch[28] Batch [899]	Speed: 2968.700395 samples/sec	accuracy=0.578123	lr=0.359465
Epoch[28] Batch [949]	Speed: 3008.849501 samples/sec	accuracy=0.577871	lr=0.359333
Epoch[28] Batch [999]	Speed: 3244.299285 samples/sec	accuracy=0.577660	lr=0.359201
Epoch[28] Batch [1049]	Speed: 2911.588581 samples/sec	accuracy=0.577526	lr=0.359068
Epoch[28] Batch [1099]	Speed: 2939.537179 samples/sec	accuracy=0.577338	lr=0.358936
Epoch[28] Batch [1149]	Speed: 3042.336051 samples/sec	accuracy=0.577217	lr=0.358803
Epoch[28] Batch [1199]	Speed: 2979.916746 samples/sec	accuracy=0.577075	lr=0.358670
Epoch[28] Batch [1249]	Speed: 3067.689952 samples/sec	accuracy=0.577048	lr=0.358537
[Epoch 28] training: accuracy=0.577065
[Epoch 28] speed: 3004 samples/sec	time cost: 436.631970
[Epoch 28] validation: err-top1=0.403958 err-top5=0.167191
Epoch[29] Batch [49]	Speed: 3101.389354 samples/sec	accuracy=0.585273	lr=0.358401
Epoch[29] Batch [99]	Speed: 3047.385608 samples/sec	accuracy=0.586631	lr=0.358268
Epoch[29] Batch [149]	Speed: 3012.700968 samples/sec	accuracy=0.586341	lr=0.358134
Epoch[29] Batch [199]	Speed: 2972.654194 samples/sec	accuracy=0.584976	lr=0.358001
Epoch[29] Batch [249]	Speed: 2929.547146 samples/sec	accuracy=0.584031	lr=0.357867
Epoch[29] Batch [299]	Speed: 2894.460557 samples/sec	accuracy=0.583356	lr=0.357732
Epoch[29] Batch [349]	Speed: 3089.917067 samples/sec	accuracy=0.583251	lr=0.357598
Epoch[29] Batch [399]	Speed: 2934.861497 samples/sec	accuracy=0.582966	lr=0.357464
Epoch[29] Batch [449]	Speed: 2958.771031 samples/sec	accuracy=0.582053	lr=0.357329
Epoch[29] Batch [499]	Speed: 2924.322967 samples/sec	accuracy=0.581551	lr=0.357194
Epoch[29] Batch [549]	Speed: 2955.508703 samples/sec	accuracy=0.581721	lr=0.357059
Epoch[29] Batch [599]	Speed: 2922.135760 samples/sec	accuracy=0.581629	lr=0.356924
Epoch[29] Batch [649]	Speed: 2948.719338 samples/sec	accuracy=0.581227	lr=0.356788
Epoch[29] Batch [699]	Speed: 2904.107229 samples/sec	accuracy=0.580735	lr=0.356652
Epoch[29] Batch [749]	Speed: 2985.904862 samples/sec	accuracy=0.580380	lr=0.356517
Epoch[29] Batch [799]	Speed: 2912.025959 samples/sec	accuracy=0.579932	lr=0.356380
Epoch[29] Batch [849]	Speed: 2965.525688 samples/sec	accuracy=0.579929	lr=0.356244
Epoch[29] Batch [899]	Speed: 2904.371482 samples/sec	accuracy=0.579693	lr=0.356108
Epoch[29] Batch [949]	Speed: 2935.604109 samples/sec	accuracy=0.579174	lr=0.355971
Epoch[29] Batch [999]	Speed: 3017.987282 samples/sec	accuracy=0.578921	lr=0.355834
Epoch[29] Batch [1049]	Speed: 2913.751601 samples/sec	accuracy=0.578634	lr=0.355698
Epoch[29] Batch [1099]	Speed: 2879.480389 samples/sec	accuracy=0.578508	lr=0.355560
Epoch[29] Batch [1149]	Speed: 2935.278934 samples/sec	accuracy=0.578235	lr=0.355423
Epoch[29] Batch [1199]	Speed: 2936.225487 samples/sec	accuracy=0.578083	lr=0.355285
Epoch[29] Batch [1249]	Speed: 3134.409766 samples/sec	accuracy=0.577720	lr=0.355148
[Epoch 29] training: accuracy=0.577716
[Epoch 29] speed: 2961 samples/sec	time cost: 442.507495
[Epoch 29] validation: err-top1=0.451477 err-top5=0.204631
Epoch[30] Batch [49]	Speed: 3000.750211 samples/sec	accuracy=0.586309	lr=0.355007
Epoch[30] Batch [99]	Speed: 2900.469758 samples/sec	accuracy=0.587031	lr=0.354869
Epoch[30] Batch [149]	Speed: 2897.958127 samples/sec	accuracy=0.586341	lr=0.354731
Epoch[30] Batch [199]	Speed: 2840.915570 samples/sec	accuracy=0.585444	lr=0.354592
Epoch[30] Batch [249]	Speed: 2922.370615 samples/sec	accuracy=0.585504	lr=0.354454
Epoch[30] Batch [299]	Speed: 3013.797767 samples/sec	accuracy=0.585387	lr=0.354315
Epoch[30] Batch [349]	Speed: 3031.721604 samples/sec	accuracy=0.584936	lr=0.354176
Epoch[30] Batch [399]	Speed: 3251.996825 samples/sec	accuracy=0.584282	lr=0.354037
Epoch[30] Batch [449]	Speed: 2985.502495 samples/sec	accuracy=0.583770	lr=0.353897
Epoch[30] Batch [499]	Speed: 2912.207652 samples/sec	accuracy=0.583643	lr=0.353758
Epoch[30] Batch [549]	Speed: 2945.191789 samples/sec	accuracy=0.583281	lr=0.353618
Epoch[30] Batch [599]	Speed: 3012.549032 samples/sec	accuracy=0.582894	lr=0.353478
Epoch[30] Batch [649]	Speed: 2907.398665 samples/sec	accuracy=0.582784	lr=0.353338
Epoch[30] Batch [699]	Speed: 2978.251178 samples/sec	accuracy=0.582461	lr=0.353198
Epoch[30] Batch [749]	Speed: 2968.683856 samples/sec	accuracy=0.582207	lr=0.353057
Epoch[30] Batch [799]	Speed: 2918.274507 samples/sec	accuracy=0.581835	lr=0.352917
Epoch[30] Batch [849]	Speed: 2877.281577 samples/sec	accuracy=0.581790	lr=0.352776
Epoch[30] Batch [899]	Speed: 3068.688500 samples/sec	accuracy=0.581504	lr=0.352635
Epoch[30] Batch [949]	Speed: 2879.708707 samples/sec	accuracy=0.581394	lr=0.352493
Epoch[30] Batch [999]	Speed: 2966.424649 samples/sec	accuracy=0.581166	lr=0.352352
Epoch[30] Batch [1049]	Speed: 3011.617680 samples/sec	accuracy=0.580850	lr=0.352211
Epoch[30] Batch [1099]	Speed: 3030.086780 samples/sec	accuracy=0.580700	lr=0.352069
Epoch[30] Batch [1149]	Speed: 2969.748546 samples/sec	accuracy=0.580444	lr=0.351927
Epoch[30] Batch [1199]	Speed: 3070.770097 samples/sec	accuracy=0.580271	lr=0.351785
Epoch[30] Batch [1249]	Speed: 3143.723216 samples/sec	accuracy=0.579946	lr=0.351642
[Epoch 30] training: accuracy=0.579949
[Epoch 30] speed: 2975 samples/sec	time cost: 440.832985
[Epoch 30] validation: err-top1=0.428193 err-top5=0.186264
Epoch[31] Batch [49]	Speed: 3001.911380 samples/sec	accuracy=0.590996	lr=0.351497
Epoch[31] Batch [99]	Speed: 3076.899780 samples/sec	accuracy=0.588672	lr=0.351354
Epoch[31] Batch [149]	Speed: 3075.112530 samples/sec	accuracy=0.587585	lr=0.351212
Epoch[31] Batch [199]	Speed: 2971.658147 samples/sec	accuracy=0.585933	lr=0.351069
Epoch[31] Batch [249]	Speed: 2913.628892 samples/sec	accuracy=0.585852	lr=0.350925
Epoch[31] Batch [299]	Speed: 2986.517565 samples/sec	accuracy=0.585492	lr=0.350782
Epoch[31] Batch [349]	Speed: 2912.269894 samples/sec	accuracy=0.585343	lr=0.350639
Epoch[31] Batch [399]	Speed: 3087.426281 samples/sec	accuracy=0.584973	lr=0.350495
Epoch[31] Batch [449]	Speed: 3019.298389 samples/sec	accuracy=0.584694	lr=0.350351
Epoch[31] Batch [499]	Speed: 2913.053156 samples/sec	accuracy=0.584174	lr=0.350207
Epoch[31] Batch [549]	Speed: 2891.489793 samples/sec	accuracy=0.583979	lr=0.350062
Epoch[31] Batch [599]	Speed: 3043.667368 samples/sec	accuracy=0.583625	lr=0.349918
Epoch[31] Batch [649]	Speed: 2891.670257 samples/sec	accuracy=0.583438	lr=0.349773
Epoch[31] Batch [699]	Speed: 3008.281119 samples/sec	accuracy=0.583245	lr=0.349629
Epoch[31] Batch [749]	Speed: 2905.655768 samples/sec	accuracy=0.583227	lr=0.349484
Epoch[31] Batch [799]	Speed: 2898.694464 samples/sec	accuracy=0.583241	lr=0.349338
Epoch[31] Batch [849]	Speed: 3044.510916 samples/sec	accuracy=0.583061	lr=0.349193
Epoch[31] Batch [899]	Speed: 3011.133241 samples/sec	accuracy=0.583038	lr=0.349048
Epoch[31] Batch [949]	Speed: 2863.032519 samples/sec	accuracy=0.582964	lr=0.348902
Epoch[31] Batch [999]	Speed: 3012.865980 samples/sec	accuracy=0.582670	lr=0.348756
Epoch[31] Batch [1049]	Speed: 2885.681487 samples/sec	accuracy=0.582343	lr=0.348610
Epoch[31] Batch [1099]	Speed: 3047.187520 samples/sec	accuracy=0.582229	lr=0.348464
Epoch[31] Batch [1149]	Speed: 2869.853496 samples/sec	accuracy=0.582180	lr=0.348317
Epoch[31] Batch [1199]	Speed: 2981.216869 samples/sec	accuracy=0.582140	lr=0.348171
Epoch[31] Batch [1249]	Speed: 3083.003835 samples/sec	accuracy=0.581891	lr=0.348024
[Epoch 31] training: accuracy=0.581887
[Epoch 31] speed: 2971 samples/sec	time cost: 441.706746
[Epoch 31] validation: err-top1=0.383809 err-top5=0.154337
Epoch[32] Batch [49]	Speed: 3120.288689 samples/sec	accuracy=0.587676	lr=0.347874
Epoch[32] Batch [99]	Speed: 2981.569109 samples/sec	accuracy=0.588887	lr=0.347727
Epoch[32] Batch [149]	Speed: 2981.181981 samples/sec	accuracy=0.588359	lr=0.347580
Epoch[32] Batch [199]	Speed: 2977.059042 samples/sec	accuracy=0.588101	lr=0.347432
Epoch[32] Batch [249]	Speed: 3042.496782 samples/sec	accuracy=0.588332	lr=0.347285
Epoch[32] Batch [299]	Speed: 2890.811590 samples/sec	accuracy=0.588509	lr=0.347137
Epoch[32] Batch [349]	Speed: 2884.474226 samples/sec	accuracy=0.587919	lr=0.346989
Epoch[32] Batch [399]	Speed: 2949.002829 samples/sec	accuracy=0.587522	lr=0.346841
Epoch[32] Batch [449]	Speed: 2938.642977 samples/sec	accuracy=0.587207	lr=0.346692
Epoch[32] Batch [499]	Speed: 3111.909220 samples/sec	accuracy=0.586773	lr=0.346544
Epoch[32] Batch [549]	Speed: 2928.932266 samples/sec	accuracy=0.586681	lr=0.346395
Epoch[32] Batch [599]	Speed: 2934.786134 samples/sec	accuracy=0.586466	lr=0.346246
Epoch[32] Batch [649]	Speed: 3001.984230 samples/sec	accuracy=0.586077	lr=0.346097
Epoch[32] Batch [699]	Speed: 3002.299671 samples/sec	accuracy=0.585646	lr=0.345948
Epoch[32] Batch [749]	Speed: 3001.197676 samples/sec	accuracy=0.585473	lr=0.345799
Epoch[32] Batch [799]	Speed: 2978.446601 samples/sec	accuracy=0.584958	lr=0.345649
Epoch[32] Batch [849]	Speed: 3121.603447 samples/sec	accuracy=0.584691	lr=0.345499
Epoch[32] Batch [899]	Speed: 2906.982509 samples/sec	accuracy=0.584495	lr=0.345349
Epoch[32] Batch [949]	Speed: 2887.297077 samples/sec	accuracy=0.584437	lr=0.345199
Epoch[32] Batch [999]	Speed: 2971.215665 samples/sec	accuracy=0.584337	lr=0.345049
Epoch[32] Batch [1049]	Speed: 3046.296985 samples/sec	accuracy=0.584009	lr=0.344899
Epoch[32] Batch [1099]	Speed: 3145.098013 samples/sec	accuracy=0.583831	lr=0.344748
Epoch[32] Batch [1149]	Speed: 2872.035933 samples/sec	accuracy=0.583831	lr=0.344597
Epoch[32] Batch [1199]	Speed: 2980.630953 samples/sec	accuracy=0.583573	lr=0.344446
Epoch[32] Batch [1249]	Speed: 3107.145117 samples/sec	accuracy=0.583479	lr=0.344295
[Epoch 32] training: accuracy=0.583490
[Epoch 32] speed: 2986 samples/sec	time cost: 439.006368
[Epoch 32] validation: err-top1=0.390545 err-top5=0.156290
Epoch[33] Batch [49]	Speed: 2963.091912 samples/sec	accuracy=0.593066	lr=0.344141
Epoch[33] Batch [99]	Speed: 2975.334006 samples/sec	accuracy=0.592959	lr=0.343989
Epoch[33] Batch [149]	Speed: 2869.453654 samples/sec	accuracy=0.592292	lr=0.343838
Epoch[33] Batch [199]	Speed: 2950.348665 samples/sec	accuracy=0.590698	lr=0.343686
Epoch[33] Batch [249]	Speed: 3026.855350 samples/sec	accuracy=0.589621	lr=0.343534
Epoch[33] Batch [299]	Speed: 2917.758143 samples/sec	accuracy=0.589411	lr=0.343382
Epoch[33] Batch [349]	Speed: 2960.106500 samples/sec	accuracy=0.588956	lr=0.343229
Epoch[33] Batch [399]	Speed: 2939.239655 samples/sec	accuracy=0.588828	lr=0.343077
Epoch[33] Batch [449]	Speed: 2935.457322 samples/sec	accuracy=0.588696	lr=0.342924
Epoch[33] Batch [499]	Speed: 2973.569547 samples/sec	accuracy=0.588746	lr=0.342771
Epoch[33] Batch [549]	Speed: 2932.819910 samples/sec	accuracy=0.588075	lr=0.342618
Epoch[33] Batch [599]	Speed: 3058.390217 samples/sec	accuracy=0.587804	lr=0.342465
Epoch[33] Batch [649]	Speed: 2982.949232 samples/sec	accuracy=0.587748	lr=0.342312
Epoch[33] Batch [699]	Speed: 3122.231440 samples/sec	accuracy=0.587554	lr=0.342158
Epoch[33] Batch [749]	Speed: 2986.753454 samples/sec	accuracy=0.587406	lr=0.342005
Epoch[33] Batch [799]	Speed: 2987.014681 samples/sec	accuracy=0.587278	lr=0.341851
Epoch[33] Batch [849]	Speed: 3051.970816 samples/sec	accuracy=0.587053	lr=0.341697
Epoch[33] Batch [899]	Speed: 2978.745340 samples/sec	accuracy=0.586747	lr=0.341543
Epoch[33] Batch [949]	Speed: 3028.845061 samples/sec	accuracy=0.586292	lr=0.341388
Epoch[33] Batch [999]	Speed: 3120.051953 samples/sec	accuracy=0.586098	lr=0.341234
Epoch[33] Batch [1049]	Speed: 2903.306710 samples/sec	accuracy=0.586039	lr=0.341079
Epoch[33] Batch [1099]	Speed: 2969.348962 samples/sec	accuracy=0.585767	lr=0.340924
Epoch[33] Batch [1149]	Speed: 2901.549504 samples/sec	accuracy=0.585658	lr=0.340769
Epoch[33] Batch [1199]	Speed: 2908.208020 samples/sec	accuracy=0.585653	lr=0.340614
Epoch[33] Batch [1249]	Speed: 3132.138511 samples/sec	accuracy=0.585514	lr=0.340458
[Epoch 33] training: accuracy=0.585528
[Epoch 33] speed: 2978 samples/sec	time cost: 441.087254
[Epoch 33] validation: err-top1=0.429010 err-top5=0.189513
Epoch[34] Batch [49]	Speed: 2927.601144 samples/sec	accuracy=0.597187	lr=0.340300
Epoch[34] Batch [99]	Speed: 3040.701990 samples/sec	accuracy=0.596523	lr=0.340144
Epoch[34] Batch [149]	Speed: 3042.487385 samples/sec	accuracy=0.595247	lr=0.339988
Epoch[34] Batch [199]	Speed: 3020.733428 samples/sec	accuracy=0.595249	lr=0.339832
Epoch[34] Batch [249]	Speed: 3000.347146 samples/sec	accuracy=0.593875	lr=0.339676
Epoch[34] Batch [299]	Speed: 3018.709334 samples/sec	accuracy=0.592900	lr=0.339520
Epoch[34] Batch [349]	Speed: 3057.434879 samples/sec	accuracy=0.591978	lr=0.339363
Epoch[34] Batch [399]	Speed: 3043.389625 samples/sec	accuracy=0.591685	lr=0.339206
Epoch[34] Batch [449]	Speed: 2958.982008 samples/sec	accuracy=0.591701	lr=0.339049
Epoch[34] Batch [499]	Speed: 3070.589418 samples/sec	accuracy=0.591162	lr=0.338892
Epoch[34] Batch [549]	Speed: 2933.875460 samples/sec	accuracy=0.590849	lr=0.338735
Epoch[34] Batch [599]	Speed: 2955.361099 samples/sec	accuracy=0.590505	lr=0.338578
Epoch[34] Batch [649]	Speed: 3011.229381 samples/sec	accuracy=0.590048	lr=0.338420
Epoch[34] Batch [699]	Speed: 3006.969330 samples/sec	accuracy=0.589393	lr=0.338263
Epoch[34] Batch [749]	Speed: 3083.674708 samples/sec	accuracy=0.588794	lr=0.338105
Epoch[34] Batch [799]	Speed: 3006.146203 samples/sec	accuracy=0.588556	lr=0.337947
Epoch[34] Batch [849]	Speed: 3069.021977 samples/sec	accuracy=0.588321	lr=0.337788
Epoch[34] Batch [899]	Speed: 3020.929494 samples/sec	accuracy=0.588116	lr=0.337630
Epoch[34] Batch [949]	Speed: 3104.299277 samples/sec	accuracy=0.587921	lr=0.337472
Epoch[34] Batch [999]	Speed: 2912.342486 samples/sec	accuracy=0.587701	lr=0.337313
Epoch[34] Batch [1049]	Speed: 3068.810103 samples/sec	accuracy=0.587729	lr=0.337154
Epoch[34] Batch [1099]	Speed: 2991.261713 samples/sec	accuracy=0.587487	lr=0.336995
Epoch[34] Batch [1149]	Speed: 3088.895505 samples/sec	accuracy=0.587480	lr=0.336836
Epoch[34] Batch [1199]	Speed: 3011.335156 samples/sec	accuracy=0.587250	lr=0.336677
Epoch[34] Batch [1249]	Speed: 3175.604969 samples/sec	accuracy=0.587084	lr=0.336517
[Epoch 34] training: accuracy=0.587082
[Epoch 34] speed: 3021 samples/sec	time cost: 434.436146
[Epoch 34] validation: err-top1=0.419576 err-top5=0.179626
Epoch[35] Batch [49]	Speed: 2987.702451 samples/sec	accuracy=0.599766	lr=0.336354
Epoch[35] Batch [99]	Speed: 2938.076207 samples/sec	accuracy=0.597891	lr=0.336194
Epoch[35] Batch [149]	Speed: 2936.466306 samples/sec	accuracy=0.595827	lr=0.336034
Epoch[35] Batch [199]	Speed: 2919.433673 samples/sec	accuracy=0.595161	lr=0.335874
Epoch[35] Batch [249]	Speed: 2992.466797 samples/sec	accuracy=0.593840	lr=0.335714
Epoch[35] Batch [299]	Speed: 2886.867096 samples/sec	accuracy=0.593734	lr=0.335553
Epoch[35] Batch [349]	Speed: 2913.075720 samples/sec	accuracy=0.593541	lr=0.335393
Epoch[35] Batch [399]	Speed: 2906.727380 samples/sec	accuracy=0.592917	lr=0.335232
Epoch[35] Batch [449]	Speed: 3053.305501 samples/sec	accuracy=0.592227	lr=0.335071
Epoch[35] Batch [499]	Speed: 3055.038235 samples/sec	accuracy=0.591777	lr=0.334910
Epoch[35] Batch [549]	Speed: 2988.962617 samples/sec	accuracy=0.591561	lr=0.334749
Epoch[35] Batch [599]	Speed: 2970.026894 samples/sec	accuracy=0.591476	lr=0.334587
Epoch[35] Batch [649]	Speed: 2990.682505 samples/sec	accuracy=0.591169	lr=0.334425
Epoch[35] Batch [699]	Speed: 2937.457220 samples/sec	accuracy=0.590593	lr=0.334264
Epoch[35] Batch [749]	Speed: 2992.042733 samples/sec	accuracy=0.590289	lr=0.334102
Epoch[35] Batch [799]	Speed: 2968.197499 samples/sec	accuracy=0.590148	lr=0.333940
Epoch[35] Batch [849]	Speed: 2926.835171 samples/sec	accuracy=0.589875	lr=0.333777
Epoch[35] Batch [899]	Speed: 2961.411974 samples/sec	accuracy=0.589523	lr=0.333615
Epoch[35] Batch [949]	Speed: 3066.735234 samples/sec	accuracy=0.589440	lr=0.333452
Epoch[35] Batch [999]	Speed: 2955.342146 samples/sec	accuracy=0.589153	lr=0.333290
Epoch[35] Batch [1049]	Speed: 3025.226962 samples/sec	accuracy=0.588850	lr=0.333127
Epoch[35] Batch [1099]	Speed: 2925.492603 samples/sec	accuracy=0.588572	lr=0.332964
Epoch[35] Batch [1149]	Speed: 3043.661545 samples/sec	accuracy=0.588522	lr=0.332801
Epoch[35] Batch [1199]	Speed: 3171.999013 samples/sec	accuracy=0.588332	lr=0.332637
Epoch[35] Batch [1249]	Speed: 3242.100286 samples/sec	accuracy=0.588370	lr=0.332474
[Epoch 35] training: accuracy=0.588367
[Epoch 35] speed: 2986 samples/sec	time cost: 440.137127
[Epoch 35] validation: err-top1=0.380919 err-top5=0.149733
Epoch[36] Batch [49]	Speed: 3203.082572 samples/sec	accuracy=0.599434	lr=0.332307
Epoch[36] Batch [99]	Speed: 2898.359890 samples/sec	accuracy=0.600469	lr=0.332143
Epoch[36] Batch [149]	Speed: 2926.075778 samples/sec	accuracy=0.599342	lr=0.331979
Epoch[36] Batch [199]	Speed: 3057.028323 samples/sec	accuracy=0.598579	lr=0.331815
Epoch[36] Batch [249]	Speed: 3002.667869 samples/sec	accuracy=0.596957	lr=0.331650
Epoch[36] Batch [299]	Speed: 2909.966717 samples/sec	accuracy=0.595889	lr=0.331486
Epoch[36] Batch [349]	Speed: 3033.128467 samples/sec	accuracy=0.595446	lr=0.331321
Epoch[36] Batch [399]	Speed: 3040.617089 samples/sec	accuracy=0.595127	lr=0.331157
Epoch[36] Batch [449]	Speed: 3118.706253 samples/sec	accuracy=0.594208	lr=0.330992
Epoch[36] Batch [499]	Speed: 2966.735080 samples/sec	accuracy=0.594268	lr=0.330827
Epoch[36] Batch [549]	Speed: 2996.532616 samples/sec	accuracy=0.593999	lr=0.330661
Epoch[36] Batch [599]	Speed: 2893.695488 samples/sec	accuracy=0.593439	lr=0.330496
Epoch[36] Batch [649]	Speed: 2989.589520 samples/sec	accuracy=0.593111	lr=0.330330
Epoch[36] Batch [699]	Speed: 2962.483101 samples/sec	accuracy=0.592755	lr=0.330165
Epoch[36] Batch [749]	Speed: 2895.224900 samples/sec	accuracy=0.592497	lr=0.329999
Epoch[36] Batch [799]	Speed: 2903.015885 samples/sec	accuracy=0.592028	lr=0.329833
Epoch[36] Batch [849]	Speed: 3169.479488 samples/sec	accuracy=0.591842	lr=0.329667
Epoch[36] Batch [899]	Speed: 3185.441134 samples/sec	accuracy=0.591518	lr=0.329500
Epoch[36] Batch [949]	Speed: 3032.535974 samples/sec	accuracy=0.591504	lr=0.329334
Epoch[36] Batch [999]	Speed: 2938.928878 samples/sec	accuracy=0.591272	lr=0.329167
Epoch[36] Batch [1049]	Speed: 2876.067581 samples/sec	accuracy=0.591162	lr=0.329000
Epoch[36] Batch [1099]	Speed: 3001.955736 samples/sec	accuracy=0.590840	lr=0.328833
Epoch[36] Batch [1149]	Speed: 2835.691143 samples/sec	accuracy=0.590679	lr=0.328666
Epoch[36] Batch [1199]	Speed: 2971.277206 samples/sec	accuracy=0.590606	lr=0.328499
Epoch[36] Batch [1249]	Speed: 3119.587970 samples/sec	accuracy=0.590521	lr=0.328332
[Epoch 36] training: accuracy=0.590543
[Epoch 36] speed: 2991 samples/sec	time cost: 438.224613
[Epoch 36] validation: err-top1=0.399613 err-top5=0.165976
Epoch[37] Batch [49]	Speed: 2975.733307 samples/sec	accuracy=0.604141	lr=0.328161
Epoch[37] Batch [99]	Speed: 3034.268788 samples/sec	accuracy=0.601436	lr=0.327993
Epoch[37] Batch [149]	Speed: 2905.806392 samples/sec	accuracy=0.599818	lr=0.327825
Epoch[37] Batch [199]	Speed: 3111.202208 samples/sec	accuracy=0.598872	lr=0.327657
Epoch[37] Batch [249]	Speed: 2921.677373 samples/sec	accuracy=0.598762	lr=0.327489
Epoch[37] Batch [299]	Speed: 2940.441185 samples/sec	accuracy=0.597871	lr=0.327321
Epoch[37] Batch [349]	Speed: 2898.318034 samples/sec	accuracy=0.597207	lr=0.327152
Epoch[37] Batch [399]	Speed: 3065.449739 samples/sec	accuracy=0.596387	lr=0.326983
Epoch[37] Batch [449]	Speed: 3015.935903 samples/sec	accuracy=0.595836	lr=0.326815
Epoch[37] Batch [499]	Speed: 3029.619119 samples/sec	accuracy=0.595398	lr=0.326646
Epoch[37] Batch [549]	Speed: 3051.070721 samples/sec	accuracy=0.595233	lr=0.326477
Epoch[37] Batch [599]	Speed: 2871.220326 samples/sec	accuracy=0.594746	lr=0.326307
Epoch[37] Batch [649]	Speed: 2925.212140 samples/sec	accuracy=0.594156	lr=0.326138
Epoch[37] Batch [699]	Speed: 2909.782819 samples/sec	accuracy=0.593818	lr=0.325968
Epoch[37] Batch [749]	Speed: 2921.117128 samples/sec	accuracy=0.593596	lr=0.325799
Epoch[37] Batch [799]	Speed: 2921.207447 samples/sec	accuracy=0.593276	lr=0.325629
Epoch[37] Batch [849]	Speed: 2911.603819 samples/sec	accuracy=0.593073	lr=0.325459
Epoch[37] Batch [899]	Speed: 3169.169330 samples/sec	accuracy=0.593024	lr=0.325289
Epoch[37] Batch [949]	Speed: 3082.814721 samples/sec	accuracy=0.592758	lr=0.325118
Epoch[37] Batch [999]	Speed: 3030.318996 samples/sec	accuracy=0.592475	lr=0.324948
Epoch[37] Batch [1049]	Speed: 3122.005121 samples/sec	accuracy=0.592487	lr=0.324777
Epoch[37] Batch [1099]	Speed: 3043.381517 samples/sec	accuracy=0.592259	lr=0.324607
Epoch[37] Batch [1149]	Speed: 2939.955464 samples/sec	accuracy=0.592223	lr=0.324436
Epoch[37] Batch [1199]	Speed: 3004.168584 samples/sec	accuracy=0.592038	lr=0.324265
Epoch[37] Batch [1249]	Speed: 3374.626369 samples/sec	accuracy=0.591832	lr=0.324094
[Epoch 37] training: accuracy=0.591852
[Epoch 37] speed: 3001 samples/sec	time cost: 436.761311
[Epoch 37] validation: err-top1=0.383510 err-top5=0.151726
Epoch[38] Batch [49]	Speed: 2894.649742 samples/sec	accuracy=0.606230	lr=0.323919
Epoch[38] Batch [99]	Speed: 2919.430815 samples/sec	accuracy=0.603008	lr=0.323747
Epoch[38] Batch [149]	Speed: 3017.266254 samples/sec	accuracy=0.601725	lr=0.323576
Epoch[38] Batch [199]	Speed: 2993.639131 samples/sec	accuracy=0.601079	lr=0.323404
Epoch[38] Batch [249]	Speed: 2864.567242 samples/sec	accuracy=0.600359	lr=0.323232
Epoch[38] Batch [299]	Speed: 3048.347223 samples/sec	accuracy=0.599648	lr=0.323060
Epoch[38] Batch [349]	Speed: 2928.309099 samples/sec	accuracy=0.598987	lr=0.322888
Epoch[38] Batch [399]	Speed: 2961.301348 samples/sec	accuracy=0.598220	lr=0.322715
Epoch[38] Batch [449]	Speed: 2974.998982 samples/sec	accuracy=0.598327	lr=0.322543
Epoch[38] Batch [499]	Speed: 2913.926196 samples/sec	accuracy=0.598213	lr=0.322370
Epoch[38] Batch [549]	Speed: 2931.465324 samples/sec	accuracy=0.597795	lr=0.322197
Epoch[38] Batch [599]	Speed: 2954.480984 samples/sec	accuracy=0.597485	lr=0.322025
Epoch[38] Batch [649]	Speed: 2981.523491 samples/sec	accuracy=0.597033	lr=0.321851
Epoch[38] Batch [699]	Speed: 3029.764874 samples/sec	accuracy=0.596702	lr=0.321678
Epoch[38] Batch [749]	Speed: 2955.287404 samples/sec	accuracy=0.596676	lr=0.321505
Epoch[38] Batch [799]	Speed: 2976.300881 samples/sec	accuracy=0.596099	lr=0.321331
Epoch[38] Batch [849]	Speed: 3008.405736 samples/sec	accuracy=0.595736	lr=0.321158
Epoch[38] Batch [899]	Speed: 3004.483602 samples/sec	accuracy=0.595371	lr=0.320984
Epoch[38] Batch [949]	Speed: 2988.307368 samples/sec	accuracy=0.595025	lr=0.320810
Epoch[38] Batch [999]	Speed: 2960.949594 samples/sec	accuracy=0.594902	lr=0.320636
Epoch[38] Batch [1049]	Speed: 2974.838751 samples/sec	accuracy=0.594791	lr=0.320462
Epoch[38] Batch [1099]	Speed: 3001.192895 samples/sec	accuracy=0.594624	lr=0.320287
Epoch[38] Batch [1149]	Speed: 3055.677121 samples/sec	accuracy=0.594591	lr=0.320113
Epoch[38] Batch [1199]	Speed: 2864.827521 samples/sec	accuracy=0.594377	lr=0.319938
Epoch[38] Batch [1249]	Speed: 3290.983982 samples/sec	accuracy=0.594240	lr=0.319763
[Epoch 38] training: accuracy=0.594256
[Epoch 38] speed: 2975 samples/sec	time cost: 440.307689
[Epoch 38] validation: err-top1=0.371094 err-top5=0.147700
Epoch[39] Batch [49]	Speed: 2985.994333 samples/sec	accuracy=0.606289	lr=0.319585
Epoch[39] Batch [99]	Speed: 2925.314946 samples/sec	accuracy=0.604668	lr=0.319410
Epoch[39] Batch [149]	Speed: 2947.426201 samples/sec	accuracy=0.603086	lr=0.319234
Epoch[39] Batch [199]	Speed: 2959.949746 samples/sec	accuracy=0.603154	lr=0.319059
Epoch[39] Batch [249]	Speed: 2972.580334 samples/sec	accuracy=0.603187	lr=0.318883
Epoch[39] Batch [299]	Speed: 3087.781336 samples/sec	accuracy=0.602620	lr=0.318708
Epoch[39] Batch [349]	Speed: 2894.618138 samples/sec	accuracy=0.601855	lr=0.318532
Epoch[39] Batch [399]	Speed: 3104.567514 samples/sec	accuracy=0.600918	lr=0.318356
Epoch[39] Batch [449]	Speed: 2879.128812 samples/sec	accuracy=0.599985	lr=0.318180
Epoch[39] Batch [499]	Speed: 2981.390495 samples/sec	accuracy=0.599693	lr=0.318004
Epoch[39] Batch [549]	Speed: 2959.298101 samples/sec	accuracy=0.599561	lr=0.317827
Epoch[39] Batch [599]	Speed: 3014.240289 samples/sec	accuracy=0.599404	lr=0.317651
Epoch[39] Batch [649]	Speed: 3008.912991 samples/sec	accuracy=0.599135	lr=0.317474
Epoch[39] Batch [699]	Speed: 2927.211185 samples/sec	accuracy=0.599025	lr=0.317297
Epoch[39] Batch [749]	Speed: 3068.428094 samples/sec	accuracy=0.598335	lr=0.317120
Epoch[39] Batch [799]	Speed: 3000.659644 samples/sec	accuracy=0.597714	lr=0.316943
Epoch[39] Batch [849]	Speed: 2895.573860 samples/sec	accuracy=0.597437	lr=0.316766
Epoch[39] Batch [899]	Speed: 2939.007833 samples/sec	accuracy=0.597148	lr=0.316589
Epoch[39] Batch [949]	Speed: 2983.273036 samples/sec	accuracy=0.596797	lr=0.316411
Epoch[39] Batch [999]	Speed: 2943.714339 samples/sec	accuracy=0.596655	lr=0.316233
Epoch[39] Batch [1049]	Speed: 3057.177293 samples/sec	accuracy=0.596427	lr=0.316056
Epoch[39] Batch [1099]	Speed: 2933.694620 samples/sec	accuracy=0.596253	lr=0.315878
Epoch[39] Batch [1149]	Speed: 3060.116693 samples/sec	accuracy=0.596106	lr=0.315700
Epoch[39] Batch [1199]	Speed: 2941.642005 samples/sec	accuracy=0.595909	lr=0.315522
Epoch[39] Batch [1249]	Speed: 3037.307197 samples/sec	accuracy=0.595821	lr=0.315343
[Epoch 39] training: accuracy=0.595801
[Epoch 39] speed: 2977 samples/sec	time cost: 440.702425
[Epoch 39] validation: err-top1=0.386998 err-top5=0.155234
Epoch[40] Batch [49]	Speed: 3127.877001 samples/sec	accuracy=0.605039	lr=0.315161
Epoch[40] Batch [99]	Speed: 2965.748156 samples/sec	accuracy=0.604951	lr=0.314983
Epoch[40] Batch [149]	Speed: 3024.155388 samples/sec	accuracy=0.604212	lr=0.314804
Epoch[40] Batch [199]	Speed: 3034.731924 samples/sec	accuracy=0.603213	lr=0.314625
Epoch[40] Batch [249]	Speed: 2984.754630 samples/sec	accuracy=0.602215	lr=0.314446
Epoch[40] Batch [299]	Speed: 3006.582440 samples/sec	accuracy=0.602191	lr=0.314267
Epoch[40] Batch [349]	Speed: 2922.159101 samples/sec	accuracy=0.601917	lr=0.314087
Epoch[40] Batch [399]	Speed: 3009.178236 samples/sec	accuracy=0.601111	lr=0.313908
Epoch[40] Batch [449]	Speed: 3023.627443 samples/sec	accuracy=0.600974	lr=0.313728
Epoch[40] Batch [499]	Speed: 3039.469626 samples/sec	accuracy=0.600635	lr=0.313549
Epoch[40] Batch [549]	Speed: 2933.839827 samples/sec	accuracy=0.600256	lr=0.313369
Epoch[40] Batch [599]	Speed: 2953.807691 samples/sec	accuracy=0.599655	lr=0.313189
Epoch[40] Batch [649]	Speed: 3103.894519 samples/sec	accuracy=0.599594	lr=0.313009
Epoch[40] Batch [699]	Speed: 2918.977482 samples/sec	accuracy=0.599113	lr=0.312829
Epoch[40] Batch [749]	Speed: 3066.404444 samples/sec	accuracy=0.599012	lr=0.312648
Epoch[40] Batch [799]	Speed: 2887.746836 samples/sec	accuracy=0.598752	lr=0.312468
Epoch[40] Batch [849]	Speed: 3057.224907 samples/sec	accuracy=0.598439	lr=0.312287
Epoch[40] Batch [899]	Speed: 2994.569167 samples/sec	accuracy=0.598023	lr=0.312106
Epoch[40] Batch [949]	Speed: 3126.193701 samples/sec	accuracy=0.597840	lr=0.311925
Epoch[40] Batch [999]	Speed: 3013.885745 samples/sec	accuracy=0.597787	lr=0.311744
Epoch[40] Batch [1049]	Speed: 2929.286323 samples/sec	accuracy=0.597707	lr=0.311563
Epoch[40] Batch [1099]	Speed: 3088.862938 samples/sec	accuracy=0.597724	lr=0.311382
Epoch[40] Batch [1149]	Speed: 3025.813404 samples/sec	accuracy=0.597520	lr=0.311201
Epoch[40] Batch [1199]	Speed: 3108.239520 samples/sec	accuracy=0.597254	lr=0.311019
Epoch[40] Batch [1249]	Speed: 3168.938868 samples/sec	accuracy=0.597150	lr=0.310837
[Epoch 40] training: accuracy=0.597158
[Epoch 40] speed: 3016 samples/sec	time cost: 434.549566
[Epoch 40] validation: err-top1=0.384481 err-top5=0.156372
Epoch[41] Batch [49]	Speed: 3100.219249 samples/sec	accuracy=0.606992	lr=0.310652
Epoch[41] Batch [99]	Speed: 3111.897135 samples/sec	accuracy=0.605234	lr=0.310470
Epoch[41] Batch [149]	Speed: 2881.461936 samples/sec	accuracy=0.606185	lr=0.310288
Epoch[41] Batch [199]	Speed: 2790.642094 samples/sec	accuracy=0.605459	lr=0.310105
Epoch[41] Batch [249]	Speed: 2875.307505 samples/sec	accuracy=0.605105	lr=0.309923
Epoch[41] Batch [299]	Speed: 3166.091347 samples/sec	accuracy=0.604375	lr=0.309741
Epoch[41] Batch [349]	Speed: 2957.008240 samples/sec	accuracy=0.603959	lr=0.309558
Epoch[41] Batch [399]	Speed: 2894.455993 samples/sec	accuracy=0.604197	lr=0.309375
Epoch[41] Batch [449]	Speed: 2995.464808 samples/sec	accuracy=0.603639	lr=0.309192
Epoch[41] Batch [499]	Speed: 2885.835360 samples/sec	accuracy=0.603090	lr=0.309009
Epoch[41] Batch [549]	Speed: 3011.353398 samples/sec	accuracy=0.602614	lr=0.308826
Epoch[41] Batch [599]	Speed: 3003.429485 samples/sec	accuracy=0.602342	lr=0.308643
Epoch[41] Batch [649]	Speed: 3115.010664 samples/sec	accuracy=0.601845	lr=0.308459
Epoch[41] Batch [699]	Speed: 2971.625579 samples/sec	accuracy=0.601490	lr=0.308276
Epoch[41] Batch [749]	Speed: 2960.457687 samples/sec	accuracy=0.601398	lr=0.308092
Epoch[41] Batch [799]	Speed: 2903.615259 samples/sec	accuracy=0.600930	lr=0.307908
Epoch[41] Batch [849]	Speed: 2980.268554 samples/sec	accuracy=0.600670	lr=0.307725
Epoch[41] Batch [899]	Speed: 2989.473074 samples/sec	accuracy=0.600191	lr=0.307540
Epoch[41] Batch [949]	Speed: 2941.668721 samples/sec	accuracy=0.600045	lr=0.307356
Epoch[41] Batch [999]	Speed: 3048.782507 samples/sec	accuracy=0.599954	lr=0.307172
Epoch[41] Batch [1049]	Speed: 2991.904087 samples/sec	accuracy=0.599869	lr=0.306988
Epoch[41] Batch [1099]	Speed: 2997.515035 samples/sec	accuracy=0.599822	lr=0.306803
Epoch[41] Batch [1149]	Speed: 3013.520162 samples/sec	accuracy=0.599642	lr=0.306618
Epoch[41] Batch [1199]	Speed: 2937.524805 samples/sec	accuracy=0.599324	lr=0.306433
Epoch[41] Batch [1249]	Speed: 3129.275224 samples/sec	accuracy=0.599171	lr=0.306249
[Epoch 41] training: accuracy=0.599189
[Epoch 41] speed: 2981 samples/sec	time cost: 440.266049
[Epoch 41] validation: err-top1=0.401586 err-top5=0.168666
Epoch[42] Batch [49]	Speed: 2929.051914 samples/sec	accuracy=0.608223	lr=0.306060
Epoch[42] Batch [99]	Speed: 2961.883119 samples/sec	accuracy=0.606279	lr=0.305875
Epoch[42] Batch [149]	Speed: 2891.957371 samples/sec	accuracy=0.606673	lr=0.305689
Epoch[42] Batch [199]	Speed: 2972.886869 samples/sec	accuracy=0.606104	lr=0.305504
Epoch[42] Batch [249]	Speed: 2987.893711 samples/sec	accuracy=0.605945	lr=0.305318
Epoch[42] Batch [299]	Speed: 2876.833453 samples/sec	accuracy=0.605609	lr=0.305133
Epoch[42] Batch [349]	Speed: 3110.826021 samples/sec	accuracy=0.605061	lr=0.304947
Epoch[42] Batch [399]	Speed: 3034.616995 samples/sec	accuracy=0.604778	lr=0.304761
Epoch[42] Batch [449]	Speed: 3020.084521 samples/sec	accuracy=0.604944	lr=0.304575
Epoch[42] Batch [499]	Speed: 2971.710742 samples/sec	accuracy=0.604490	lr=0.304388
Epoch[42] Batch [549]	Speed: 3005.967409 samples/sec	accuracy=0.604208	lr=0.304202
Epoch[42] Batch [599]	Speed: 2924.768681 samples/sec	accuracy=0.603748	lr=0.304016
Epoch[42] Batch [649]	Speed: 3051.814547 samples/sec	accuracy=0.603447	lr=0.303829
Epoch[42] Batch [699]	Speed: 3080.486352 samples/sec	accuracy=0.603191	lr=0.303642
Epoch[42] Batch [749]	Speed: 3068.401437 samples/sec	accuracy=0.602759	lr=0.303456
Epoch[42] Batch [799]	Speed: 2874.268664 samples/sec	accuracy=0.602388	lr=0.303269
Epoch[42] Batch [849]	Speed: 2928.004462 samples/sec	accuracy=0.602050	lr=0.303082
Epoch[42] Batch [899]	Speed: 3144.925292 samples/sec	accuracy=0.602065	lr=0.302894
Epoch[42] Batch [949]	Speed: 3105.893940 samples/sec	accuracy=0.601645	lr=0.302707
Epoch[42] Batch [999]	Speed: 2981.050754 samples/sec	accuracy=0.601338	lr=0.302520
Epoch[42] Batch [1049]	Speed: 2915.667076 samples/sec	accuracy=0.601137	lr=0.302332
Epoch[42] Batch [1099]	Speed: 3176.261410 samples/sec	accuracy=0.600908	lr=0.302144
Epoch[42] Batch [1149]	Speed: 2933.551390 samples/sec	accuracy=0.600711	lr=0.301957
Epoch[42] Batch [1199]	Speed: 3129.519108 samples/sec	accuracy=0.600577	lr=0.301769
Epoch[42] Batch [1249]	Speed: 3135.516321 samples/sec	accuracy=0.600489	lr=0.301581
[Epoch 42] training: accuracy=0.600480
[Epoch 42] speed: 3004 samples/sec	time cost: 436.405694
[Epoch 42] validation: err-top1=0.383351 err-top5=0.156270
Epoch[43] Batch [49]	Speed: 3161.120015 samples/sec	accuracy=0.611895	lr=0.301389
Epoch[43] Batch [99]	Speed: 3117.475392 samples/sec	accuracy=0.611895	lr=0.301200
Epoch[43] Batch [149]	Speed: 3077.637950 samples/sec	accuracy=0.610820	lr=0.301012
Epoch[43] Batch [199]	Speed: 3150.735300 samples/sec	accuracy=0.610649	lr=0.300823
Epoch[43] Batch [249]	Speed: 2967.834999 samples/sec	accuracy=0.609043	lr=0.300635
Epoch[43] Batch [299]	Speed: 2941.525920 samples/sec	accuracy=0.608861	lr=0.300446
Epoch[43] Batch [349]	Speed: 2939.260614 samples/sec	accuracy=0.608248	lr=0.300257
Epoch[43] Batch [399]	Speed: 2994.000783 samples/sec	accuracy=0.607520	lr=0.300068
Epoch[43] Batch [449]	Speed: 2971.182202 samples/sec	accuracy=0.606923	lr=0.299879
Epoch[43] Batch [499]	Speed: 2917.473334 samples/sec	accuracy=0.606502	lr=0.299690
Epoch[43] Batch [549]	Speed: 2859.850006 samples/sec	accuracy=0.605961	lr=0.299500
Epoch[43] Batch [599]	Speed: 2969.268902 samples/sec	accuracy=0.605804	lr=0.299311
Epoch[43] Batch [649]	Speed: 3094.885291 samples/sec	accuracy=0.605584	lr=0.299121
Epoch[43] Batch [699]	Speed: 2981.771094 samples/sec	accuracy=0.605384	lr=0.298932
Epoch[43] Batch [749]	Speed: 3024.269526 samples/sec	accuracy=0.604970	lr=0.298742
Epoch[43] Batch [799]	Speed: 3000.587153 samples/sec	accuracy=0.604561	lr=0.298552
Epoch[43] Batch [849]	Speed: 3113.518529 samples/sec	accuracy=0.604311	lr=0.298362
Epoch[43] Batch [899]	Speed: 2973.286500 samples/sec	accuracy=0.604069	lr=0.298171
Epoch[43] Batch [949]	Speed: 3095.812766 samples/sec	accuracy=0.603677	lr=0.297981
Epoch[43] Batch [999]	Speed: 3224.168751 samples/sec	accuracy=0.603445	lr=0.297791
Epoch[43] Batch [1049]	Speed: 3006.600793 samples/sec	accuracy=0.603138	lr=0.297600
Epoch[43] Batch [1099]	Speed: 3099.248474 samples/sec	accuracy=0.603121	lr=0.297409
Epoch[43] Batch [1149]	Speed: 3073.394256 samples/sec	accuracy=0.602936	lr=0.297219
Epoch[43] Batch [1199]	Speed: 2963.254847 samples/sec	accuracy=0.602675	lr=0.297028
Epoch[43] Batch [1249]	Speed: 3070.228210 samples/sec	accuracy=0.602426	lr=0.296837
[Epoch 43] training: accuracy=0.602445
[Epoch 43] speed: 3026 samples/sec	time cost: 433.171928
[Epoch 43] validation: err-top1=0.398477 err-top5=0.167092
Epoch[44] Batch [49]	Speed: 2932.498596 samples/sec	accuracy=0.610742	lr=0.296642
Epoch[44] Batch [99]	Speed: 2875.001709 samples/sec	accuracy=0.611152	lr=0.296451
Epoch[44] Batch [149]	Speed: 2856.955832 samples/sec	accuracy=0.611230	lr=0.296259
Epoch[44] Batch [199]	Speed: 2913.168546 samples/sec	accuracy=0.609502	lr=0.296068
Epoch[44] Batch [249]	Speed: 2922.527232 samples/sec	accuracy=0.610293	lr=0.295876
Epoch[44] Batch [299]	Speed: 3015.021292 samples/sec	accuracy=0.609964	lr=0.295684
Epoch[44] Batch [349]	Speed: 3035.180616 samples/sec	accuracy=0.609595	lr=0.295493
Epoch[44] Batch [399]	Speed: 3014.065524 samples/sec	accuracy=0.608999	lr=0.295301
Epoch[44] Batch [449]	Speed: 2949.050738 samples/sec	accuracy=0.609065	lr=0.295109
Epoch[44] Batch [499]	Speed: 3069.754177 samples/sec	accuracy=0.608428	lr=0.294917
Epoch[44] Batch [549]	Speed: 2951.644734 samples/sec	accuracy=0.608434	lr=0.294724
Epoch[44] Batch [599]	Speed: 2910.102526 samples/sec	accuracy=0.608296	lr=0.294532
Epoch[44] Batch [649]	Speed: 2938.967691 samples/sec	accuracy=0.607882	lr=0.294339
Epoch[44] Batch [699]	Speed: 3242.194315 samples/sec	accuracy=0.607475	lr=0.294147
Epoch[44] Batch [749]	Speed: 3122.227082 samples/sec	accuracy=0.606932	lr=0.293954
Epoch[44] Batch [799]	Speed: 3087.468894 samples/sec	accuracy=0.606716	lr=0.293761
Epoch[44] Batch [849]	Speed: 3213.663497 samples/sec	accuracy=0.606591	lr=0.293568
Epoch[44] Batch [899]	Speed: 2946.245231 samples/sec	accuracy=0.606408	lr=0.293375
Epoch[44] Batch [949]	Speed: 2958.746776 samples/sec	accuracy=0.606066	lr=0.293182
Epoch[44] Batch [999]	Speed: 3012.601859 samples/sec	accuracy=0.605744	lr=0.292989
Epoch[44] Batch [1049]	Speed: 2965.309274 samples/sec	accuracy=0.605546	lr=0.292795
Epoch[44] Batch [1099]	Speed: 3101.108724 samples/sec	accuracy=0.605233	lr=0.292602
Epoch[44] Batch [1149]	Speed: 3009.695115 samples/sec	accuracy=0.604933	lr=0.292408
Epoch[44] Batch [1199]	Speed: 2928.368557 samples/sec	accuracy=0.604647	lr=0.292215
Epoch[44] Batch [1249]	Speed: 3237.698526 samples/sec	accuracy=0.604295	lr=0.292021
[Epoch 44] training: accuracy=0.604286
[Epoch 44] speed: 3002 samples/sec	time cost: 436.224114
[Epoch 44] validation: err-top1=0.399454 err-top5=0.168407
Epoch[45] Batch [49]	Speed: 3054.811514 samples/sec	accuracy=0.616699	lr=0.291823
Epoch[45] Batch [99]	Speed: 3014.463524 samples/sec	accuracy=0.615928	lr=0.291629
Epoch[45] Batch [149]	Speed: 2936.426876 samples/sec	accuracy=0.615130	lr=0.291435
Epoch[45] Batch [199]	Speed: 3008.658541 samples/sec	accuracy=0.614067	lr=0.291240
Epoch[45] Batch [249]	Speed: 2935.425182 samples/sec	accuracy=0.613363	lr=0.291046
Epoch[45] Batch [299]	Speed: 2939.201679 samples/sec	accuracy=0.613050	lr=0.290852
Epoch[45] Batch [349]	Speed: 2949.860517 samples/sec	accuracy=0.612101	lr=0.290657
Epoch[45] Batch [399]	Speed: 2925.148826 samples/sec	accuracy=0.611458	lr=0.290462
Epoch[45] Batch [449]	Speed: 2916.260914 samples/sec	accuracy=0.610688	lr=0.290267
Epoch[45] Batch [499]	Speed: 2860.473102 samples/sec	accuracy=0.610234	lr=0.290073
Epoch[45] Batch [549]	Speed: 3012.803591 samples/sec	accuracy=0.610302	lr=0.289878
Epoch[45] Batch [599]	Speed: 2930.478408 samples/sec	accuracy=0.609624	lr=0.289682
Epoch[45] Batch [649]	Speed: 3073.997411 samples/sec	accuracy=0.609147	lr=0.289487
Epoch[45] Batch [699]	Speed: 3050.475965 samples/sec	accuracy=0.608884	lr=0.289292
Epoch[45] Batch [749]	Speed: 2910.114593 samples/sec	accuracy=0.608444	lr=0.289096
Epoch[45] Batch [799]	Speed: 2977.628113 samples/sec	accuracy=0.608169	lr=0.288901
Epoch[45] Batch [849]	Speed: 2928.025222 samples/sec	accuracy=0.608100	lr=0.288705
Epoch[45] Batch [899]	Speed: 3020.826486 samples/sec	accuracy=0.607878	lr=0.288509
Epoch[45] Batch [949]	Speed: 2855.506859 samples/sec	accuracy=0.607644	lr=0.288314
Epoch[45] Batch [999]	Speed: 2932.573641 samples/sec	accuracy=0.607613	lr=0.288118
Epoch[45] Batch [1049]	Speed: 2927.909889 samples/sec	accuracy=0.607308	lr=0.287921
Epoch[45] Batch [1099]	Speed: 2872.257810 samples/sec	accuracy=0.607101	lr=0.287725
Epoch[45] Batch [1149]	Speed: 2902.696556 samples/sec	accuracy=0.606898	lr=0.287529
Epoch[45] Batch [1199]	Speed: 2934.237329 samples/sec	accuracy=0.606705	lr=0.287333
Epoch[45] Batch [1249]	Speed: 3169.038335 samples/sec	accuracy=0.606541	lr=0.287136
[Epoch 45] training: accuracy=0.606544
[Epoch 45] speed: 2957 samples/sec	time cost: 443.675597
[Epoch 45] validation: err-top1=0.376156 err-top5=0.148198
Epoch[46] Batch [49]	Speed: 3142.867178 samples/sec	accuracy=0.619180	lr=0.286936
Epoch[46] Batch [99]	Speed: 2987.701245 samples/sec	accuracy=0.617588	lr=0.286739
Epoch[46] Batch [149]	Speed: 3104.456659 samples/sec	accuracy=0.615098	lr=0.286542
Epoch[46] Batch [199]	Speed: 3010.443505 samples/sec	accuracy=0.614390	lr=0.286345
Epoch[46] Batch [249]	Speed: 3008.955403 samples/sec	accuracy=0.613711	lr=0.286148
Epoch[46] Batch [299]	Speed: 2963.290135 samples/sec	accuracy=0.613330	lr=0.285951
Epoch[46] Batch [349]	Speed: 2930.541953 samples/sec	accuracy=0.612567	lr=0.285754
Epoch[46] Batch [399]	Speed: 2992.778531 samples/sec	accuracy=0.612361	lr=0.285556
Epoch[46] Batch [449]	Speed: 3042.712496 samples/sec	accuracy=0.612383	lr=0.285359
Epoch[46] Batch [499]	Speed: 3027.173566 samples/sec	accuracy=0.612082	lr=0.285161
Epoch[46] Batch [549]	Speed: 2988.413950 samples/sec	accuracy=0.611591	lr=0.284964
Epoch[46] Batch [599]	Speed: 2926.303571 samples/sec	accuracy=0.611086	lr=0.284766
Epoch[46] Batch [649]	Speed: 2984.310976 samples/sec	accuracy=0.610864	lr=0.284568
Epoch[46] Batch [699]	Speed: 3054.594212 samples/sec	accuracy=0.610313	lr=0.284370
Epoch[46] Batch [749]	Speed: 2997.211516 samples/sec	accuracy=0.609930	lr=0.284172
Epoch[46] Batch [799]	Speed: 2926.165607 samples/sec	accuracy=0.609489	lr=0.283974
Epoch[46] Batch [849]	Speed: 3238.165839 samples/sec	accuracy=0.609293	lr=0.283776
Epoch[46] Batch [899]	Speed: 2995.378905 samples/sec	accuracy=0.609004	lr=0.283577
Epoch[46] Batch [949]	Speed: 2970.754861 samples/sec	accuracy=0.608935	lr=0.283379
Epoch[46] Batch [999]	Speed: 3157.643989 samples/sec	accuracy=0.608741	lr=0.283181
Epoch[46] Batch [1049]	Speed: 2988.107531 samples/sec	accuracy=0.608670	lr=0.282982
Epoch[46] Batch [1099]	Speed: 3158.123867 samples/sec	accuracy=0.608572	lr=0.282783
Epoch[46] Batch [1149]	Speed: 2950.049840 samples/sec	accuracy=0.608425	lr=0.282584
Epoch[46] Batch [1199]	Speed: 3098.246391 samples/sec	accuracy=0.608092	lr=0.282385
Epoch[46] Batch [1249]	Speed: 3156.034301 samples/sec	accuracy=0.607945	lr=0.282186
[Epoch 46] training: accuracy=0.607974
[Epoch 46] speed: 3027 samples/sec	time cost: 433.223377
[Epoch 46] validation: err-top1=0.368896 err-top5=0.143331
Epoch[47] Batch [49]	Speed: 3293.211031 samples/sec	accuracy=0.618691	lr=0.281983
Epoch[47] Batch [99]	Speed: 3006.809636 samples/sec	accuracy=0.618926	lr=0.281784
Epoch[47] Batch [149]	Speed: 3025.095237 samples/sec	accuracy=0.618053	lr=0.281585
Epoch[47] Batch [199]	Speed: 2957.899599 samples/sec	accuracy=0.616904	lr=0.281385
Epoch[47] Batch [249]	Speed: 2876.113649 samples/sec	accuracy=0.616688	lr=0.281186
Epoch[47] Batch [299]	Speed: 2947.209225 samples/sec	accuracy=0.616318	lr=0.280986
Epoch[47] Batch [349]	Speed: 2863.427709 samples/sec	accuracy=0.616032	lr=0.280786
Epoch[47] Batch [399]	Speed: 3138.039390 samples/sec	accuracy=0.615562	lr=0.280587
Epoch[47] Batch [449]	Speed: 3066.601141 samples/sec	accuracy=0.614818	lr=0.280387
Epoch[47] Batch [499]	Speed: 3074.890789 samples/sec	accuracy=0.614652	lr=0.280187
Epoch[47] Batch [549]	Speed: 3073.004640 samples/sec	accuracy=0.613984	lr=0.279987
Epoch[47] Batch [599]	Speed: 3091.656186 samples/sec	accuracy=0.613457	lr=0.279786
Epoch[47] Batch [649]	Speed: 2989.212331 samples/sec	accuracy=0.613202	lr=0.279586
Epoch[47] Batch [699]	Speed: 2994.350413 samples/sec	accuracy=0.612478	lr=0.279386
Epoch[47] Batch [749]	Speed: 2971.307916 samples/sec	accuracy=0.612016	lr=0.279185
Epoch[47] Batch [799]	Speed: 3053.642720 samples/sec	accuracy=0.611854	lr=0.278985
Epoch[47] Batch [849]	Speed: 3011.166005 samples/sec	accuracy=0.611557	lr=0.278784
Epoch[47] Batch [899]	Speed: 2889.934727 samples/sec	accuracy=0.611304	lr=0.278583
Epoch[47] Batch [949]	Speed: 2955.656770 samples/sec	accuracy=0.610920	lr=0.278382
Epoch[47] Batch [999]	Speed: 2893.006273 samples/sec	accuracy=0.610737	lr=0.278181
Epoch[47] Batch [1049]	Speed: 2958.236814 samples/sec	accuracy=0.610702	lr=0.277980
Epoch[47] Batch [1099]	Speed: 2944.869051 samples/sec	accuracy=0.610246	lr=0.277779
Epoch[47] Batch [1149]	Speed: 3157.349142 samples/sec	accuracy=0.610102	lr=0.277578
Epoch[47] Batch [1199]	Speed: 3108.827988 samples/sec	accuracy=0.609862	lr=0.277377
Epoch[47] Batch [1249]	Speed: 3151.792723 samples/sec	accuracy=0.609655	lr=0.277175
[Epoch 47] training: accuracy=0.609637
[Epoch 47] speed: 3014 samples/sec	time cost: 435.976130
[Epoch 47] validation: err-top1=0.377372 err-top5=0.151686
Epoch[48] Batch [49]	Speed: 3161.200052 samples/sec	accuracy=0.622344	lr=0.276970
Epoch[48] Batch [99]	Speed: 2955.945706 samples/sec	accuracy=0.621924	lr=0.276768
Epoch[48] Batch [149]	Speed: 3109.621363 samples/sec	accuracy=0.621790	lr=0.276566
Epoch[48] Batch [199]	Speed: 2907.994455 samples/sec	accuracy=0.620171	lr=0.276365
Epoch[48] Batch [249]	Speed: 3070.780943 samples/sec	accuracy=0.619086	lr=0.276163
Epoch[48] Batch [299]	Speed: 3170.127315 samples/sec	accuracy=0.618158	lr=0.275961
Epoch[48] Batch [349]	Speed: 3087.782845 samples/sec	accuracy=0.617299	lr=0.275759
Epoch[48] Batch [399]	Speed: 2981.761530 samples/sec	accuracy=0.616646	lr=0.275557
Epoch[48] Batch [449]	Speed: 3064.912046 samples/sec	accuracy=0.616202	lr=0.275354
Epoch[48] Batch [499]	Speed: 3118.932320 samples/sec	accuracy=0.615816	lr=0.275152
Epoch[48] Batch [549]	Speed: 3094.376549 samples/sec	accuracy=0.615520	lr=0.274950
Epoch[48] Batch [599]	Speed: 2869.568146 samples/sec	accuracy=0.615347	lr=0.274747
Epoch[48] Batch [649]	Speed: 3042.019553 samples/sec	accuracy=0.614884	lr=0.274545
Epoch[48] Batch [699]	Speed: 2996.482442 samples/sec	accuracy=0.614455	lr=0.274342
Epoch[48] Batch [749]	Speed: 2945.179591 samples/sec	accuracy=0.614254	lr=0.274139
Epoch[48] Batch [799]	Speed: 2947.664937 samples/sec	accuracy=0.614069	lr=0.273936
Epoch[48] Batch [849]	Speed: 2981.538227 samples/sec	accuracy=0.613968	lr=0.273733
Epoch[48] Batch [899]	Speed: 3158.203567 samples/sec	accuracy=0.613746	lr=0.273530
Epoch[48] Batch [949]	Speed: 2962.568640 samples/sec	accuracy=0.613165	lr=0.273327
Epoch[48] Batch [999]	Speed: 3139.779222 samples/sec	accuracy=0.612977	lr=0.273124
Epoch[48] Batch [1049]	Speed: 3070.542659 samples/sec	accuracy=0.612896	lr=0.272921
Epoch[48] Batch [1099]	Speed: 3041.819146 samples/sec	accuracy=0.612834	lr=0.272717
Epoch[48] Batch [1149]	Speed: 2924.322490 samples/sec	accuracy=0.612550	lr=0.272514
Epoch[48] Batch [1199]	Speed: 2927.296215 samples/sec	accuracy=0.612302	lr=0.272310
Epoch[48] Batch [1249]	Speed: 3106.886997 samples/sec	accuracy=0.612029	lr=0.272107
[Epoch 48] training: accuracy=0.612013
[Epoch 48] speed: 3028 samples/sec	time cost: 433.154852
[Epoch 48] validation: err-top1=0.363182 err-top5=0.140206
Epoch[49] Batch [49]	Speed: 3059.148206 samples/sec	accuracy=0.621621	lr=0.271899
Epoch[49] Batch [99]	Speed: 2956.184114 samples/sec	accuracy=0.621865	lr=0.271695
Epoch[49] Batch [149]	Speed: 2997.581186 samples/sec	accuracy=0.621569	lr=0.271491
Epoch[49] Batch [199]	Speed: 2986.189402 samples/sec	accuracy=0.620283	lr=0.271287
Epoch[49] Batch [249]	Speed: 2978.003953 samples/sec	accuracy=0.619914	lr=0.271083
Epoch[49] Batch [299]	Speed: 3037.294782 samples/sec	accuracy=0.619661	lr=0.270879
Epoch[49] Batch [349]	Speed: 3018.581529 samples/sec	accuracy=0.619127	lr=0.270675
Epoch[49] Batch [399]	Speed: 2944.506090 samples/sec	accuracy=0.618462	lr=0.270470
Epoch[49] Batch [449]	Speed: 2882.573695 samples/sec	accuracy=0.618038	lr=0.270266
Epoch[49] Batch [499]	Speed: 3021.435625 samples/sec	accuracy=0.617467	lr=0.270061
Epoch[49] Batch [549]	Speed: 3059.187906 samples/sec	accuracy=0.617111	lr=0.269857
Epoch[49] Batch [599]	Speed: 3120.748027 samples/sec	accuracy=0.617249	lr=0.269652
Epoch[49] Batch [649]	Speed: 3043.319626 samples/sec	accuracy=0.616961	lr=0.269447
Epoch[49] Batch [699]	Speed: 2975.208322 samples/sec	accuracy=0.616620	lr=0.269243
Epoch[49] Batch [749]	Speed: 3055.882836 samples/sec	accuracy=0.616608	lr=0.269038
Epoch[49] Batch [799]	Speed: 3264.431761 samples/sec	accuracy=0.616421	lr=0.268833
Epoch[49] Batch [849]	Speed: 2911.175210 samples/sec	accuracy=0.616085	lr=0.268628
Epoch[49] Batch [899]	Speed: 3064.277076 samples/sec	accuracy=0.615659	lr=0.268422
Epoch[49] Batch [949]	Speed: 2957.227354 samples/sec	accuracy=0.615168	lr=0.268217
Epoch[49] Batch [999]	Speed: 3052.861243 samples/sec	accuracy=0.614862	lr=0.268012
Epoch[49] Batch [1049]	Speed: 3011.248002 samples/sec	accuracy=0.614631	lr=0.267807
Epoch[49] Batch [1099]	Speed: 2933.022194 samples/sec	accuracy=0.614445	lr=0.267601
Epoch[49] Batch [1149]	Speed: 2953.636938 samples/sec	accuracy=0.614286	lr=0.267396
Epoch[49] Batch [1199]	Speed: 2885.734185 samples/sec	accuracy=0.614033	lr=0.267190
Epoch[49] Batch [1249]	Speed: 3130.910401 samples/sec	accuracy=0.613613	lr=0.266984
[Epoch 49] training: accuracy=0.613607
[Epoch 49] speed: 3007 samples/sec	time cost: 435.851857
[Epoch 49] validation: err-top1=0.371632 err-top5=0.145169
Epoch[50] Batch [49]	Speed: 3218.369427 samples/sec	accuracy=0.624219	lr=0.266774
Epoch[50] Batch [99]	Speed: 2905.761844 samples/sec	accuracy=0.624150	lr=0.266568
Epoch[50] Batch [149]	Speed: 2905.271633 samples/sec	accuracy=0.622487	lr=0.266362
Epoch[50] Batch [199]	Speed: 3009.898947 samples/sec	accuracy=0.621309	lr=0.266156
Epoch[50] Batch [249]	Speed: 3035.301723 samples/sec	accuracy=0.620816	lr=0.265950
Epoch[50] Batch [299]	Speed: 3004.636449 samples/sec	accuracy=0.620023	lr=0.265744
Epoch[50] Batch [349]	Speed: 3133.162412 samples/sec	accuracy=0.619763	lr=0.265538
Epoch[50] Batch [399]	Speed: 3011.913014 samples/sec	accuracy=0.619202	lr=0.265331
Epoch[50] Batch [449]	Speed: 3035.881048 samples/sec	accuracy=0.618980	lr=0.265125
Epoch[50] Batch [499]	Speed: 2980.595376 samples/sec	accuracy=0.618549	lr=0.264918
Epoch[50] Batch [549]	Speed: 2940.065146 samples/sec	accuracy=0.617978	lr=0.264712
Epoch[50] Batch [599]	Speed: 3000.311976 samples/sec	accuracy=0.617907	lr=0.264505
Epoch[50] Batch [649]	Speed: 2911.211637 samples/sec	accuracy=0.617634	lr=0.264298
Epoch[50] Batch [699]	Speed: 2925.187435 samples/sec	accuracy=0.617327	lr=0.264092
Epoch[50] Batch [749]	Speed: 2980.803435 samples/sec	accuracy=0.617121	lr=0.263885
Epoch[50] Batch [799]	Speed: 3086.565044 samples/sec	accuracy=0.616891	lr=0.263678
Epoch[50] Batch [849]	Speed: 3075.794070 samples/sec	accuracy=0.616790	lr=0.263471
Epoch[50] Batch [899]	Speed: 3012.855708 samples/sec	accuracy=0.616418	lr=0.263264
Epoch[50] Batch [949]	Speed: 2968.508465 samples/sec	accuracy=0.616055	lr=0.263056
Epoch[50] Batch [999]	Speed: 2958.659379 samples/sec	accuracy=0.616120	lr=0.262849
Epoch[50] Batch [1049]	Speed: 2914.777880 samples/sec	accuracy=0.615783	lr=0.262642
Epoch[50] Batch [1099]	Speed: 3047.769619 samples/sec	accuracy=0.615669	lr=0.262434
Epoch[50] Batch [1149]	Speed: 2987.894626 samples/sec	accuracy=0.615526	lr=0.262227
Epoch[50] Batch [1199]	Speed: 2941.547396 samples/sec	accuracy=0.615408	lr=0.262019
Epoch[50] Batch [1249]	Speed: 3130.948334 samples/sec	accuracy=0.615288	lr=0.261812
[Epoch 50] training: accuracy=0.615285
[Epoch 50] speed: 3001 samples/sec	time cost: 436.760926
[Epoch 50] validation: err-top1=0.383092 err-top5=0.156748
Epoch[51] Batch [49]	Speed: 2989.764455 samples/sec	accuracy=0.629004	lr=0.261600
Epoch[51] Batch [99]	Speed: 3077.467531 samples/sec	accuracy=0.626582	lr=0.261392
Epoch[51] Batch [149]	Speed: 2920.530764 samples/sec	accuracy=0.626048	lr=0.261184
Epoch[51] Batch [199]	Speed: 3123.909748 samples/sec	accuracy=0.624346	lr=0.260976
Epoch[51] Batch [249]	Speed: 2983.502361 samples/sec	accuracy=0.623613	lr=0.260768
Epoch[51] Batch [299]	Speed: 2959.733614 samples/sec	accuracy=0.622809	lr=0.260560
Epoch[51] Batch [349]	Speed: 2903.992752 samples/sec	accuracy=0.622550	lr=0.260352
Epoch[51] Batch [399]	Speed: 2937.994770 samples/sec	accuracy=0.622009	lr=0.260144
Epoch[51] Batch [449]	Speed: 2999.889751 samples/sec	accuracy=0.621580	lr=0.259935
Epoch[51] Batch [499]	Speed: 3185.453845 samples/sec	accuracy=0.621441	lr=0.259727
Epoch[51] Batch [549]	Speed: 2933.765358 samples/sec	accuracy=0.621186	lr=0.259519
Epoch[51] Batch [599]	Speed: 2863.578073 samples/sec	accuracy=0.621110	lr=0.259310
Epoch[51] Batch [649]	Speed: 3053.295994 samples/sec	accuracy=0.621061	lr=0.259102
Epoch[51] Batch [699]	Speed: 3045.512957 samples/sec	accuracy=0.620723	lr=0.258893
Epoch[51] Batch [749]	Speed: 3101.597731 samples/sec	accuracy=0.620263	lr=0.258684
Epoch[51] Batch [799]	Speed: 3057.641006 samples/sec	accuracy=0.619807	lr=0.258475
Epoch[51] Batch [849]	Speed: 2955.849930 samples/sec	accuracy=0.619619	lr=0.258267
Epoch[51] Batch [899]	Speed: 3103.030482 samples/sec	accuracy=0.619281	lr=0.258058
Epoch[51] Batch [949]	Speed: 2941.789734 samples/sec	accuracy=0.619231	lr=0.257849
Epoch[51] Batch [999]	Speed: 3022.941589 samples/sec	accuracy=0.618944	lr=0.257640
Epoch[51] Batch [1049]	Speed: 3116.156692 samples/sec	accuracy=0.618665	lr=0.257430
Epoch[51] Batch [1099]	Speed: 3031.441415 samples/sec	accuracy=0.618391	lr=0.257221
Epoch[51] Batch [1149]	Speed: 3109.601371 samples/sec	accuracy=0.618287	lr=0.257012
Epoch[51] Batch [1199]	Speed: 2996.705564 samples/sec	accuracy=0.618155	lr=0.256803
Epoch[51] Batch [1249]	Speed: 3052.121635 samples/sec	accuracy=0.617941	lr=0.256593
[Epoch 51] training: accuracy=0.617966
[Epoch 51] speed: 3013 samples/sec	time cost: 435.014652
[Epoch 51] validation: err-top1=0.350526 err-top5=0.132135
Epoch[52] Batch [49]	Speed: 2950.539956 samples/sec	accuracy=0.629941	lr=0.256379
Epoch[52] Batch [99]	Speed: 2982.179744 samples/sec	accuracy=0.628711	lr=0.256170
Epoch[52] Batch [149]	Speed: 2954.529233 samples/sec	accuracy=0.627480	lr=0.255960
Epoch[52] Batch [199]	Speed: 3020.395027 samples/sec	accuracy=0.626318	lr=0.255751
Epoch[52] Batch [249]	Speed: 2959.470202 samples/sec	accuracy=0.626004	lr=0.255541
Epoch[52] Batch [299]	Speed: 3045.376308 samples/sec	accuracy=0.625596	lr=0.255331
Epoch[52] Batch [349]	Speed: 3038.523878 samples/sec	accuracy=0.624780	lr=0.255121
Epoch[52] Batch [399]	Speed: 2884.949614 samples/sec	accuracy=0.624531	lr=0.254911
Epoch[52] Batch [449]	Speed: 2826.912097 samples/sec	accuracy=0.624351	lr=0.254701
Epoch[52] Batch [499]	Speed: 2868.683849 samples/sec	accuracy=0.623682	lr=0.254491
Epoch[52] Batch [549]	Speed: 2932.504763 samples/sec	accuracy=0.623290	lr=0.254281
Epoch[52] Batch [599]	Speed: 2904.390101 samples/sec	accuracy=0.623164	lr=0.254071
Epoch[52] Batch [649]	Speed: 3040.078776 samples/sec	accuracy=0.622849	lr=0.253861
Epoch[52] Batch [699]	Speed: 3041.545531 samples/sec	accuracy=0.622500	lr=0.253650
Epoch[52] Batch [749]	Speed: 2994.834019 samples/sec	accuracy=0.622212	lr=0.253440
Epoch[52] Batch [799]	Speed: 2943.986859 samples/sec	accuracy=0.621820	lr=0.253229
Epoch[52] Batch [849]	Speed: 2961.247732 samples/sec	accuracy=0.621466	lr=0.253019
Epoch[52] Batch [899]	Speed: 2935.494880 samples/sec	accuracy=0.621197	lr=0.252808
Epoch[52] Batch [949]	Speed: 2931.531633 samples/sec	accuracy=0.620787	lr=0.252598
Epoch[52] Batch [999]	Speed: 3134.983792 samples/sec	accuracy=0.620564	lr=0.252387
Epoch[52] Batch [1049]	Speed: 2942.346287 samples/sec	accuracy=0.620282	lr=0.252176
Epoch[52] Batch [1099]	Speed: 2901.346442 samples/sec	accuracy=0.620164	lr=0.251965
Epoch[52] Batch [1149]	Speed: 2962.474683 samples/sec	accuracy=0.619916	lr=0.251754
Epoch[52] Batch [1199]	Speed: 3013.240832 samples/sec	accuracy=0.619819	lr=0.251543
Epoch[52] Batch [1249]	Speed: 3084.098038 samples/sec	accuracy=0.619681	lr=0.251332
[Epoch 52] training: accuracy=0.619689
[Epoch 52] speed: 2966 samples/sec	time cost: 441.560250
[Epoch 52] validation: err-top1=0.397786 err-top5=0.166178
Epoch[53] Batch [49]	Speed: 3023.439242 samples/sec	accuracy=0.634961	lr=0.251117
Epoch[53] Batch [99]	Speed: 3037.958155 samples/sec	accuracy=0.631846	lr=0.250906
Epoch[53] Batch [149]	Speed: 2936.614479 samples/sec	accuracy=0.630592	lr=0.250695
Epoch[53] Batch [199]	Speed: 2931.798500 samples/sec	accuracy=0.629541	lr=0.250483
Epoch[53] Batch [249]	Speed: 3092.799702 samples/sec	accuracy=0.628230	lr=0.250272
Epoch[53] Batch [299]	Speed: 3065.802207 samples/sec	accuracy=0.627406	lr=0.250061
Epoch[53] Batch [349]	Speed: 2943.744118 samples/sec	accuracy=0.627274	lr=0.249849
Epoch[53] Batch [399]	Speed: 2966.308444 samples/sec	accuracy=0.626584	lr=0.249638
Epoch[53] Batch [449]	Speed: 2902.080070 samples/sec	accuracy=0.625890	lr=0.249426
Epoch[53] Batch [499]	Speed: 2991.110058 samples/sec	accuracy=0.625400	lr=0.249215
Epoch[53] Batch [549]	Speed: 2856.678666 samples/sec	accuracy=0.625138	lr=0.249003
Epoch[53] Batch [599]	Speed: 3069.987203 samples/sec	accuracy=0.624692	lr=0.248791
Epoch[53] Batch [649]	Speed: 3001.221584 samples/sec	accuracy=0.624488	lr=0.248579
Epoch[53] Batch [699]	Speed: 3033.807979 samples/sec	accuracy=0.624208	lr=0.248367
Epoch[53] Batch [749]	Speed: 2944.131432 samples/sec	accuracy=0.623740	lr=0.248156
Epoch[53] Batch [799]	Speed: 2954.141739 samples/sec	accuracy=0.623348	lr=0.247944
Epoch[53] Batch [849]	Speed: 3010.696273 samples/sec	accuracy=0.622879	lr=0.247732
Epoch[53] Batch [899]	Speed: 3148.111249 samples/sec	accuracy=0.622597	lr=0.247519
Epoch[53] Batch [949]	Speed: 2949.189167 samples/sec	accuracy=0.622286	lr=0.247307
Epoch[53] Batch [999]	Speed: 2982.283777 samples/sec	accuracy=0.622142	lr=0.247095
Epoch[53] Batch [1049]	Speed: 2911.659284 samples/sec	accuracy=0.621850	lr=0.246883
Epoch[53] Batch [1099]	Speed: 2949.479310 samples/sec	accuracy=0.621848	lr=0.246670
Epoch[53] Batch [1149]	Speed: 2983.435918 samples/sec	accuracy=0.621507	lr=0.246458
Epoch[53] Batch [1199]	Speed: 3083.077752 samples/sec	accuracy=0.621545	lr=0.246246
Epoch[53] Batch [1249]	Speed: 3094.275160 samples/sec	accuracy=0.621365	lr=0.246033
[Epoch 53] training: accuracy=0.621372
[Epoch 53] speed: 2990 samples/sec	time cost: 438.530436
[Epoch 53] validation: err-top1=0.350307 err-top5=0.132673
Epoch[54] Batch [49]	Speed: 3119.329818 samples/sec	accuracy=0.634219	lr=0.245816
Epoch[54] Batch [99]	Speed: 3090.103586 samples/sec	accuracy=0.632529	lr=0.245604
Epoch[54] Batch [149]	Speed: 2981.547417 samples/sec	accuracy=0.632428	lr=0.245391
Epoch[54] Batch [199]	Speed: 3065.880554 samples/sec	accuracy=0.629951	lr=0.245178
Epoch[54] Batch [249]	Speed: 3152.429082 samples/sec	accuracy=0.629840	lr=0.244966
Epoch[54] Batch [299]	Speed: 3174.584680 samples/sec	accuracy=0.629271	lr=0.244753
Epoch[54] Batch [349]	Speed: 2844.782966 samples/sec	accuracy=0.628761	lr=0.244540
Epoch[54] Batch [399]	Speed: 3068.274212 samples/sec	accuracy=0.628152	lr=0.244327
Epoch[54] Batch [449]	Speed: 3086.561894 samples/sec	accuracy=0.627823	lr=0.244114
Epoch[54] Batch [499]	Speed: 3111.910753 samples/sec	accuracy=0.627705	lr=0.243901
Epoch[54] Batch [549]	Speed: 2983.261018 samples/sec	accuracy=0.627037	lr=0.243688
Epoch[54] Batch [599]	Speed: 2967.491492 samples/sec	accuracy=0.626634	lr=0.243475
Epoch[54] Batch [649]	Speed: 2921.766376 samples/sec	accuracy=0.626110	lr=0.243262
Epoch[54] Batch [699]	Speed: 3145.819733 samples/sec	accuracy=0.625950	lr=0.243049
Epoch[54] Batch [749]	Speed: 3173.919503 samples/sec	accuracy=0.625870	lr=0.242835
Epoch[54] Batch [799]	Speed: 3110.994070 samples/sec	accuracy=0.625560	lr=0.242622
Epoch[54] Batch [849]	Speed: 3143.228197 samples/sec	accuracy=0.625284	lr=0.242409
Epoch[54] Batch [899]	Speed: 2954.328726 samples/sec	accuracy=0.624941	lr=0.242195
Epoch[54] Batch [949]	Speed: 3187.486138 samples/sec	accuracy=0.624648	lr=0.241982
Epoch[54] Batch [999]	Speed: 2965.795913 samples/sec	accuracy=0.624653	lr=0.241768
Epoch[54] Batch [1049]	Speed: 3044.148524 samples/sec	accuracy=0.624516	lr=0.241555
Epoch[54] Batch [1099]	Speed: 3023.267707 samples/sec	accuracy=0.624216	lr=0.241341
Epoch[54] Batch [1149]	Speed: 3212.464535 samples/sec	accuracy=0.623788	lr=0.241127
Epoch[54] Batch [1199]	Speed: 3016.365538 samples/sec	accuracy=0.623537	lr=0.240914
Epoch[54] Batch [1249]	Speed: 3244.980906 samples/sec	accuracy=0.623365	lr=0.240700
[Epoch 54] training: accuracy=0.623368
[Epoch 54] speed: 3066 samples/sec	time cost: 428.046509
[Epoch 54] validation: err-top1=0.364059 err-top5=0.140545
Epoch[55] Batch [49]	Speed: 3080.338859 samples/sec	accuracy=0.635195	lr=0.240482
Epoch[55] Batch [99]	Speed: 2939.167244 samples/sec	accuracy=0.635371	lr=0.240268
Epoch[55] Batch [149]	Speed: 2899.238782 samples/sec	accuracy=0.633913	lr=0.240054
Epoch[55] Batch [199]	Speed: 3069.137728 samples/sec	accuracy=0.632837	lr=0.239840
Epoch[55] Batch [249]	Speed: 2992.846225 samples/sec	accuracy=0.632996	lr=0.239626
Epoch[55] Batch [299]	Speed: 3093.753158 samples/sec	accuracy=0.631794	lr=0.239412
Epoch[55] Batch [349]	Speed: 3051.924449 samples/sec	accuracy=0.630901	lr=0.239198
Epoch[55] Batch [399]	Speed: 3152.401687 samples/sec	accuracy=0.630322	lr=0.238984
Epoch[55] Batch [449]	Speed: 2957.472282 samples/sec	accuracy=0.630239	lr=0.238769
Epoch[55] Batch [499]	Speed: 2979.819617 samples/sec	accuracy=0.629863	lr=0.238555
Epoch[55] Batch [549]	Speed: 3030.980053 samples/sec	accuracy=0.629771	lr=0.238341
Epoch[55] Batch [599]	Speed: 2998.137117 samples/sec	accuracy=0.629500	lr=0.238126
Epoch[55] Batch [649]	Speed: 3039.115659 samples/sec	accuracy=0.629259	lr=0.237912
Epoch[55] Batch [699]	Speed: 3023.741242 samples/sec	accuracy=0.628863	lr=0.237698
Epoch[55] Batch [749]	Speed: 3247.601910 samples/sec	accuracy=0.628427	lr=0.237483
Epoch[55] Batch [799]	Speed: 2999.357969 samples/sec	accuracy=0.628068	lr=0.237269
Epoch[55] Batch [849]	Speed: 3147.305218 samples/sec	accuracy=0.627646	lr=0.237054
Epoch[55] Batch [899]	Speed: 2904.268493 samples/sec	accuracy=0.627211	lr=0.236839
Epoch[55] Batch [949]	Speed: 2931.824357 samples/sec	accuracy=0.626891	lr=0.236625
Epoch[55] Batch [999]	Speed: 2966.253131 samples/sec	accuracy=0.626729	lr=0.236410
Epoch[55] Batch [1049]	Speed: 3012.094459 samples/sec	accuracy=0.626515	lr=0.236195
Epoch[55] Batch [1099]	Speed: 2955.495687 samples/sec	accuracy=0.626478	lr=0.235981
Epoch[55] Batch [1149]	Speed: 3073.839846 samples/sec	accuracy=0.626474	lr=0.235766
Epoch[55] Batch [1199]	Speed: 2942.122923 samples/sec	accuracy=0.626239	lr=0.235551
Epoch[55] Batch [1249]	Speed: 3041.078031 samples/sec	accuracy=0.625826	lr=0.235336
[Epoch 55] training: accuracy=0.625834
[Epoch 55] speed: 3016 samples/sec	time cost: 434.824327
[Epoch 55] validation: err-top1=0.394212 err-top5=0.164680
Epoch[56] Batch [49]	Speed: 3139.957531 samples/sec	accuracy=0.635957	lr=0.235117
Epoch[56] Batch [99]	Speed: 3042.759877 samples/sec	accuracy=0.635996	lr=0.234902
Epoch[56] Batch [149]	Speed: 2820.457536 samples/sec	accuracy=0.634479	lr=0.234687
Epoch[56] Batch [199]	Speed: 3033.910503 samples/sec	accuracy=0.633911	lr=0.234472
Epoch[56] Batch [249]	Speed: 3130.762055 samples/sec	accuracy=0.633070	lr=0.234256
Epoch[56] Batch [299]	Speed: 3064.347125 samples/sec	accuracy=0.632487	lr=0.234041
Epoch[56] Batch [349]	Speed: 3012.119427 samples/sec	accuracy=0.631867	lr=0.233826
Epoch[56] Batch [399]	Speed: 2987.688318 samples/sec	accuracy=0.631218	lr=0.233611
Epoch[56] Batch [449]	Speed: 2897.936814 samples/sec	accuracy=0.630907	lr=0.233396
Epoch[56] Batch [499]	Speed: 3018.874326 samples/sec	accuracy=0.630699	lr=0.233180
Epoch[56] Batch [549]	Speed: 2918.101890 samples/sec	accuracy=0.630423	lr=0.232965
Epoch[56] Batch [599]	Speed: 3090.372888 samples/sec	accuracy=0.630146	lr=0.232749
Epoch[56] Batch [649]	Speed: 3168.806676 samples/sec	accuracy=0.630015	lr=0.232534
Epoch[56] Batch [699]	Speed: 3054.054022 samples/sec	accuracy=0.629516	lr=0.232319
Epoch[56] Batch [749]	Speed: 2992.409461 samples/sec	accuracy=0.629445	lr=0.232103
Epoch[56] Batch [799]	Speed: 2939.799146 samples/sec	accuracy=0.629310	lr=0.231887
Epoch[56] Batch [849]	Speed: 3116.963902 samples/sec	accuracy=0.628788	lr=0.231672
Epoch[56] Batch [899]	Speed: 2973.221953 samples/sec	accuracy=0.628407	lr=0.231456
Epoch[56] Batch [949]	Speed: 3081.365335 samples/sec	accuracy=0.628154	lr=0.231241
Epoch[56] Batch [999]	Speed: 3050.498497 samples/sec	accuracy=0.628059	lr=0.231025
Epoch[56] Batch [1049]	Speed: 2989.990699 samples/sec	accuracy=0.627831	lr=0.230809
Epoch[56] Batch [1099]	Speed: 3104.707013 samples/sec	accuracy=0.627540	lr=0.230593
Epoch[56] Batch [1149]	Speed: 2996.941851 samples/sec	accuracy=0.627255	lr=0.230377
Epoch[56] Batch [1199]	Speed: 2939.918959 samples/sec	accuracy=0.627245	lr=0.230162
Epoch[56] Batch [1249]	Speed: 3068.364478 samples/sec	accuracy=0.627082	lr=0.229946
[Epoch 56] training: accuracy=0.627088
[Epoch 56] speed: 3021 samples/sec	time cost: 433.853629
[Epoch 56] validation: err-top1=0.352380 err-top5=0.131916
Epoch[57] Batch [49]	Speed: 3138.769296 samples/sec	accuracy=0.637227	lr=0.229725
Epoch[57] Batch [99]	Speed: 2914.037147 samples/sec	accuracy=0.635146	lr=0.229510
Epoch[57] Batch [149]	Speed: 2975.336439 samples/sec	accuracy=0.635299	lr=0.229294
Epoch[57] Batch [199]	Speed: 2919.297348 samples/sec	accuracy=0.634932	lr=0.229077
Epoch[57] Batch [249]	Speed: 3154.099763 samples/sec	accuracy=0.635148	lr=0.228861
Epoch[57] Batch [299]	Speed: 3040.394483 samples/sec	accuracy=0.634844	lr=0.228645
Epoch[57] Batch [349]	Speed: 3025.699448 samples/sec	accuracy=0.634690	lr=0.228429
Epoch[57] Batch [399]	Speed: 3085.453840 samples/sec	accuracy=0.634453	lr=0.228213
Epoch[57] Batch [449]	Speed: 3020.433346 samples/sec	accuracy=0.634431	lr=0.227997
Epoch[57] Batch [499]	Speed: 2938.494881 samples/sec	accuracy=0.633764	lr=0.227781
Epoch[57] Batch [549]	Speed: 2962.785309 samples/sec	accuracy=0.633308	lr=0.227564
Epoch[57] Batch [599]	Speed: 3005.030071 samples/sec	accuracy=0.632812	lr=0.227348
Epoch[57] Batch [649]	Speed: 3074.759107 samples/sec	accuracy=0.632724	lr=0.227132
Epoch[57] Batch [699]	Speed: 2983.009023 samples/sec	accuracy=0.632693	lr=0.226915
Epoch[57] Batch [749]	Speed: 2910.735129 samples/sec	accuracy=0.632366	lr=0.226699
Epoch[57] Batch [799]	Speed: 3075.742704 samples/sec	accuracy=0.632031	lr=0.226482
Epoch[57] Batch [849]	Speed: 3200.233620 samples/sec	accuracy=0.631735	lr=0.226266
Epoch[57] Batch [899]	Speed: 3132.777422 samples/sec	accuracy=0.631541	lr=0.226049
Epoch[57] Batch [949]	Speed: 3044.685431 samples/sec	accuracy=0.631122	lr=0.225833
Epoch[57] Batch [999]	Speed: 2880.102992 samples/sec	accuracy=0.630847	lr=0.225616
Epoch[57] Batch [1049]	Speed: 2971.693512 samples/sec	accuracy=0.630730	lr=0.225400
Epoch[57] Batch [1099]	Speed: 2964.379480 samples/sec	accuracy=0.630471	lr=0.225183
Epoch[57] Batch [1149]	Speed: 2946.424145 samples/sec	accuracy=0.630307	lr=0.224967
Epoch[57] Batch [1199]	Speed: 2908.476841 samples/sec	accuracy=0.630091	lr=0.224750
Epoch[57] Batch [1249]	Speed: 3137.555373 samples/sec	accuracy=0.629784	lr=0.224533
[Epoch 57] training: accuracy=0.629811
[Epoch 57] speed: 3012 samples/sec	time cost: 435.276411
[Epoch 57] validation: err-top1=0.357561 err-top5=0.138552
Epoch[58] Batch [49]	Speed: 3072.533091 samples/sec	accuracy=0.638145	lr=0.224312
Epoch[58] Batch [99]	Speed: 2964.146049 samples/sec	accuracy=0.639316	lr=0.224095
Epoch[58] Batch [149]	Speed: 3002.000764 samples/sec	accuracy=0.639460	lr=0.223879
Epoch[58] Batch [199]	Speed: 3043.077263 samples/sec	accuracy=0.638433	lr=0.223662
Epoch[58] Batch [249]	Speed: 3138.087859 samples/sec	accuracy=0.638035	lr=0.223445
Epoch[58] Batch [299]	Speed: 3061.716255 samples/sec	accuracy=0.637256	lr=0.223228
Epoch[58] Batch [349]	Speed: 3007.998505 samples/sec	accuracy=0.637123	lr=0.223011
Epoch[58] Batch [399]	Speed: 3018.016633 samples/sec	accuracy=0.636760	lr=0.222794
Epoch[58] Batch [449]	Speed: 3224.787750 samples/sec	accuracy=0.636771	lr=0.222577
Epoch[58] Batch [499]	Speed: 2977.162719 samples/sec	accuracy=0.636545	lr=0.222360
Epoch[58] Batch [549]	Speed: 2966.967895 samples/sec	accuracy=0.636415	lr=0.222143
Epoch[58] Batch [599]	Speed: 3054.511705 samples/sec	accuracy=0.635767	lr=0.221926
Epoch[58] Batch [649]	Speed: 2986.268093 samples/sec	accuracy=0.635403	lr=0.221709
Epoch[58] Batch [699]	Speed: 3081.638511 samples/sec	accuracy=0.635112	lr=0.221492
Epoch[58] Batch [749]	Speed: 2980.562942 samples/sec	accuracy=0.635121	lr=0.221275
Epoch[58] Batch [799]	Speed: 2937.816235 samples/sec	accuracy=0.634607	lr=0.221058
Epoch[58] Batch [849]	Speed: 2865.760610 samples/sec	accuracy=0.634383	lr=0.220841
Epoch[58] Batch [899]	Speed: 2882.959824 samples/sec	accuracy=0.634208	lr=0.220623
Epoch[58] Batch [949]	Speed: 3110.968697 samples/sec	accuracy=0.634142	lr=0.220406
Epoch[58] Batch [999]	Speed: 2986.031244 samples/sec	accuracy=0.633889	lr=0.220189
Epoch[58] Batch [1049]	Speed: 2998.584850 samples/sec	accuracy=0.633642	lr=0.219972
Epoch[58] Batch [1099]	Speed: 2992.830667 samples/sec	accuracy=0.633524	lr=0.219754
Epoch[58] Batch [1149]	Speed: 2922.626310 samples/sec	accuracy=0.633373	lr=0.219537
Epoch[58] Batch [1199]	Speed: 3088.566092 samples/sec	accuracy=0.633110	lr=0.219320
Epoch[58] Batch [1249]	Speed: 3402.643771 samples/sec	accuracy=0.632878	lr=0.219102
[Epoch 58] training: accuracy=0.632865
[Epoch 58] speed: 3024 samples/sec	time cost: 434.486354
[Epoch 58] validation: err-top1=0.354899 err-top5=0.134013
Epoch[59] Batch [49]	Speed: 3186.081984 samples/sec	accuracy=0.644902	lr=0.218881
Epoch[59] Batch [99]	Speed: 3123.533207 samples/sec	accuracy=0.644512	lr=0.218663
Epoch[59] Batch [149]	Speed: 3032.436113 samples/sec	accuracy=0.643444	lr=0.218446
Epoch[59] Batch [199]	Speed: 2979.131463 samples/sec	accuracy=0.641816	lr=0.218228
Epoch[59] Batch [249]	Speed: 3097.973838 samples/sec	accuracy=0.641258	lr=0.218011
Epoch[59] Batch [299]	Speed: 3052.689391 samples/sec	accuracy=0.640550	lr=0.217793
Epoch[59] Batch [349]	Speed: 3217.911908 samples/sec	accuracy=0.639478	lr=0.217576
Epoch[59] Batch [399]	Speed: 2935.214983 samples/sec	accuracy=0.638738	lr=0.217358
Epoch[59] Batch [449]	Speed: 3042.128751 samples/sec	accuracy=0.638505	lr=0.217141
Epoch[59] Batch [499]	Speed: 2943.154082 samples/sec	accuracy=0.637992	lr=0.216923
Epoch[59] Batch [549]	Speed: 2933.727243 samples/sec	accuracy=0.637612	lr=0.216706
Epoch[59] Batch [599]	Speed: 2938.688016 samples/sec	accuracy=0.637049	lr=0.216488
Epoch[59] Batch [649]	Speed: 3128.933723 samples/sec	accuracy=0.636621	lr=0.216270
Epoch[59] Batch [699]	Speed: 2914.395957 samples/sec	accuracy=0.635956	lr=0.216053
Epoch[59] Batch [749]	Speed: 2995.226832 samples/sec	accuracy=0.636056	lr=0.215835
Epoch[59] Batch [799]	Speed: 3113.625472 samples/sec	accuracy=0.635813	lr=0.215617
Epoch[59] Batch [849]	Speed: 3173.001700 samples/sec	accuracy=0.635717	lr=0.215400
Epoch[59] Batch [899]	Speed: 2940.316418 samples/sec	accuracy=0.635595	lr=0.215182
Epoch[59] Batch [949]	Speed: 3028.012819 samples/sec	accuracy=0.635405	lr=0.214964
Epoch[59] Batch [999]	Speed: 3011.321516 samples/sec	accuracy=0.635066	lr=0.214746
Epoch[59] Batch [1049]	Speed: 2861.738872 samples/sec	accuracy=0.634921	lr=0.214529
Epoch[59] Batch [1099]	Speed: 2880.722348 samples/sec	accuracy=0.634709	lr=0.214311
Epoch[59] Batch [1149]	Speed: 2931.099619 samples/sec	accuracy=0.634616	lr=0.214093
Epoch[59] Batch [1199]	Speed: 2910.078668 samples/sec	accuracy=0.634406	lr=0.213875
Epoch[59] Batch [1249]	Speed: 3075.958708 samples/sec	accuracy=0.634146	lr=0.213657
[Epoch 59] training: accuracy=0.634152
[Epoch 59] speed: 3012 samples/sec	time cost: 436.028603
[Epoch 59] validation: err-top1=0.358000 err-top5=0.136519
Epoch[60] Batch [49]	Speed: 2968.929331 samples/sec	accuracy=0.650059	lr=0.213435
Epoch[60] Batch [99]	Speed: 3008.419517 samples/sec	accuracy=0.647529	lr=0.213217
Epoch[60] Batch [149]	Speed: 2932.937392 samples/sec	accuracy=0.646126	lr=0.212999
Epoch[60] Batch [199]	Speed: 2960.365944 samples/sec	accuracy=0.644824	lr=0.212781
Epoch[60] Batch [249]	Speed: 2925.762278 samples/sec	accuracy=0.644102	lr=0.212563
Epoch[60] Batch [299]	Speed: 3005.129187 samples/sec	accuracy=0.643148	lr=0.212345
Epoch[60] Batch [349]	Speed: 2960.625474 samples/sec	accuracy=0.642081	lr=0.212127
Epoch[60] Batch [399]	Speed: 2908.057422 samples/sec	accuracy=0.641951	lr=0.211909
Epoch[60] Batch [449]	Speed: 3011.994883 samples/sec	accuracy=0.641667	lr=0.211691
Epoch[60] Batch [499]	Speed: 3076.307472 samples/sec	accuracy=0.641129	lr=0.211473
Epoch[60] Batch [549]	Speed: 3075.303916 samples/sec	accuracy=0.640646	lr=0.211255
Epoch[60] Batch [599]	Speed: 2931.400859 samples/sec	accuracy=0.640199	lr=0.211037
Epoch[60] Batch [649]	Speed: 3003.926575 samples/sec	accuracy=0.639838	lr=0.210819
Epoch[60] Batch [699]	Speed: 2929.665645 samples/sec	accuracy=0.639446	lr=0.210601
Epoch[60] Batch [749]	Speed: 2911.832009 samples/sec	accuracy=0.638785	lr=0.210383
Epoch[60] Batch [799]	Speed: 2902.276489 samples/sec	accuracy=0.638354	lr=0.210165
Epoch[60] Batch [849]	Speed: 2952.443842 samples/sec	accuracy=0.638066	lr=0.209947
Epoch[60] Batch [899]	Speed: 2909.232919 samples/sec	accuracy=0.637862	lr=0.209729
Epoch[60] Batch [949]	Speed: 2876.606015 samples/sec	accuracy=0.637748	lr=0.209511
Epoch[60] Batch [999]	Speed: 2993.707907 samples/sec	accuracy=0.637663	lr=0.209293
Epoch[60] Batch [1049]	Speed: 3054.042165 samples/sec	accuracy=0.637513	lr=0.209075
Epoch[60] Batch [1099]	Speed: 3054.696146 samples/sec	accuracy=0.637275	lr=0.208856
Epoch[60] Batch [1149]	Speed: 3040.567666 samples/sec	accuracy=0.637036	lr=0.208638
Epoch[60] Batch [1199]	Speed: 2897.222633 samples/sec	accuracy=0.636898	lr=0.208420
Epoch[60] Batch [1249]	Speed: 3051.455705 samples/sec	accuracy=0.636730	lr=0.208202
[Epoch 60] training: accuracy=0.636749
[Epoch 60] speed: 2970 samples/sec	time cost: 441.825658
[Epoch 60] validation: err-top1=0.345623 err-top5=0.129205
Epoch[61] Batch [49]	Speed: 2964.178166 samples/sec	accuracy=0.647988	lr=0.207979
Epoch[61] Batch [99]	Speed: 3095.639971 samples/sec	accuracy=0.647236	lr=0.207761
Epoch[61] Batch [149]	Speed: 3094.659173 samples/sec	accuracy=0.647305	lr=0.207543
Epoch[61] Batch [199]	Speed: 3001.877936 samples/sec	accuracy=0.645854	lr=0.207325
Epoch[61] Batch [249]	Speed: 3041.626693 samples/sec	accuracy=0.644926	lr=0.207106
Epoch[61] Batch [299]	Speed: 2937.849633 samples/sec	accuracy=0.643971	lr=0.206888
Epoch[61] Batch [349]	Speed: 3005.648252 samples/sec	accuracy=0.643404	lr=0.206670
Epoch[61] Batch [399]	Speed: 2922.776629 samples/sec	accuracy=0.642974	lr=0.206452
Epoch[61] Batch [449]	Speed: 3092.959082 samples/sec	accuracy=0.642914	lr=0.206233
Epoch[61] Batch [499]	Speed: 2928.675107 samples/sec	accuracy=0.642838	lr=0.206015
Epoch[61] Batch [549]	Speed: 3012.747545 samples/sec	accuracy=0.642326	lr=0.205797
Epoch[61] Batch [599]	Speed: 2910.409208 samples/sec	accuracy=0.641906	lr=0.205579
Epoch[61] Batch [649]	Speed: 2934.795880 samples/sec	accuracy=0.641537	lr=0.205360
Epoch[61] Batch [699]	Speed: 3139.558981 samples/sec	accuracy=0.640889	lr=0.205142
Epoch[61] Batch [749]	Speed: 3043.392342 samples/sec	accuracy=0.640594	lr=0.204924
Epoch[61] Batch [799]	Speed: 2965.987984 samples/sec	accuracy=0.640262	lr=0.204705
Epoch[61] Batch [849]	Speed: 3046.895041 samples/sec	accuracy=0.640031	lr=0.204487
Epoch[61] Batch [899]	Speed: 2939.361353 samples/sec	accuracy=0.639936	lr=0.204269
Epoch[61] Batch [949]	Speed: 3017.349813 samples/sec	accuracy=0.639638	lr=0.204051
Epoch[61] Batch [999]	Speed: 3080.543666 samples/sec	accuracy=0.639346	lr=0.203832
Epoch[61] Batch [1049]	Speed: 3081.769456 samples/sec	accuracy=0.639085	lr=0.203614
Epoch[61] Batch [1099]	Speed: 2981.851830 samples/sec	accuracy=0.638886	lr=0.203396
Epoch[61] Batch [1149]	Speed: 2817.196983 samples/sec	accuracy=0.638640	lr=0.203177
Epoch[61] Batch [1199]	Speed: 2915.802230 samples/sec	accuracy=0.638603	lr=0.202959
Epoch[61] Batch [1249]	Speed: 3194.180220 samples/sec	accuracy=0.638655	lr=0.202740
[Epoch 61] training: accuracy=0.638669
[Epoch 61] speed: 3002 samples/sec	time cost: 436.768302
[Epoch 61] validation: err-top1=0.342714 err-top5=0.127531
Epoch[62] Batch [49]	Speed: 2981.967848 samples/sec	accuracy=0.647246	lr=0.202518
Epoch[62] Batch [99]	Speed: 2944.216924 samples/sec	accuracy=0.648760	lr=0.202299
Epoch[62] Batch [149]	Speed: 3120.705579 samples/sec	accuracy=0.648841	lr=0.202081
Epoch[62] Batch [199]	Speed: 2875.338843 samples/sec	accuracy=0.647393	lr=0.201863
Epoch[62] Batch [249]	Speed: 2898.408123 samples/sec	accuracy=0.646398	lr=0.201644
Epoch[62] Batch [299]	Speed: 2890.111574 samples/sec	accuracy=0.645254	lr=0.201426
Epoch[62] Batch [349]	Speed: 3013.899238 samples/sec	accuracy=0.645265	lr=0.201208
Epoch[62] Batch [399]	Speed: 2860.539592 samples/sec	accuracy=0.644397	lr=0.200989
Epoch[62] Batch [449]	Speed: 2970.617113 samples/sec	accuracy=0.643906	lr=0.200771
Epoch[62] Batch [499]	Speed: 3041.835777 samples/sec	accuracy=0.643695	lr=0.200552
Epoch[62] Batch [549]	Speed: 3032.676099 samples/sec	accuracy=0.643532	lr=0.200334
Epoch[62] Batch [599]	Speed: 3043.273177 samples/sec	accuracy=0.643564	lr=0.200116
Epoch[62] Batch [649]	Speed: 2873.665808 samples/sec	accuracy=0.643493	lr=0.199897
Epoch[62] Batch [699]	Speed: 3019.376924 samples/sec	accuracy=0.643217	lr=0.199679
Epoch[62] Batch [749]	Speed: 2932.576164 samples/sec	accuracy=0.642811	lr=0.199461
Epoch[62] Batch [799]	Speed: 3000.268927 samples/sec	accuracy=0.642692	lr=0.199242
Epoch[62] Batch [849]	Speed: 3077.381359 samples/sec	accuracy=0.642371	lr=0.199024
Epoch[62] Batch [899]	Speed: 2992.827331 samples/sec	accuracy=0.642378	lr=0.198806
Epoch[62] Batch [949]	Speed: 2933.520333 samples/sec	accuracy=0.642385	lr=0.198587
Epoch[62] Batch [999]	Speed: 3060.396408 samples/sec	accuracy=0.642240	lr=0.198369
Epoch[62] Batch [1049]	Speed: 2941.172242 samples/sec	accuracy=0.641944	lr=0.198150
Epoch[62] Batch [1099]	Speed: 3111.315844 samples/sec	accuracy=0.641725	lr=0.197932
Epoch[62] Batch [1149]	Speed: 2883.231121 samples/sec	accuracy=0.641489	lr=0.197714
Epoch[62] Batch [1199]	Speed: 2903.528694 samples/sec	accuracy=0.641274	lr=0.197495
Epoch[62] Batch [1249]	Speed: 3312.571128 samples/sec	accuracy=0.641080	lr=0.197277
[Epoch 62] training: accuracy=0.641079
[Epoch 62] speed: 2983 samples/sec	time cost: 439.711185
[Epoch 62] validation: err-top1=0.356346 err-top5=0.137018
Epoch[63] Batch [49]	Speed: 2951.920266 samples/sec	accuracy=0.650957	lr=0.197054
Epoch[63] Batch [99]	Speed: 2914.089739 samples/sec	accuracy=0.650352	lr=0.196836
Epoch[63] Batch [149]	Speed: 2981.218773 samples/sec	accuracy=0.649746	lr=0.196618
Epoch[63] Batch [199]	Speed: 2975.530448 samples/sec	accuracy=0.648164	lr=0.196399
Epoch[63] Batch [249]	Speed: 3091.488573 samples/sec	accuracy=0.648359	lr=0.196181
Epoch[63] Batch [299]	Speed: 2888.247621 samples/sec	accuracy=0.648021	lr=0.195963
Epoch[63] Batch [349]	Speed: 2982.363712 samples/sec	accuracy=0.647793	lr=0.195744
Epoch[63] Batch [399]	Speed: 2823.988879 samples/sec	accuracy=0.647471	lr=0.195526
Epoch[63] Batch [449]	Speed: 3042.835628 samples/sec	accuracy=0.647181	lr=0.195308
Epoch[63] Batch [499]	Speed: 3251.057878 samples/sec	accuracy=0.646787	lr=0.195089
Epoch[63] Batch [549]	Speed: 3106.959951 samples/sec	accuracy=0.646690	lr=0.194871
Epoch[63] Batch [599]	Speed: 2944.571819 samples/sec	accuracy=0.646484	lr=0.194653
Epoch[63] Batch [649]	Speed: 2991.230090 samples/sec	accuracy=0.645880	lr=0.194434
Epoch[63] Batch [699]	Speed: 2962.582168 samples/sec	accuracy=0.645727	lr=0.194216
Epoch[63] Batch [749]	Speed: 2937.949069 samples/sec	accuracy=0.645344	lr=0.193998
Epoch[63] Batch [799]	Speed: 2997.684413 samples/sec	accuracy=0.645222	lr=0.193780
Epoch[63] Batch [849]	Speed: 2972.140497 samples/sec	accuracy=0.645049	lr=0.193561
Epoch[63] Batch [899]	Speed: 2988.997521 samples/sec	accuracy=0.644814	lr=0.193343
Epoch[63] Batch [949]	Speed: 3076.121469 samples/sec	accuracy=0.644570	lr=0.193125
Epoch[63] Batch [999]	Speed: 2908.609832 samples/sec	accuracy=0.644411	lr=0.192907
Epoch[63] Batch [1049]	Speed: 2904.582197 samples/sec	accuracy=0.644212	lr=0.192688
Epoch[63] Batch [1099]	Speed: 3101.661971 samples/sec	accuracy=0.643918	lr=0.192470
Epoch[63] Batch [1149]	Speed: 3228.733084 samples/sec	accuracy=0.643640	lr=0.192252
Epoch[63] Batch [1199]	Speed: 3013.441973 samples/sec	accuracy=0.643577	lr=0.192034
Epoch[63] Batch [1249]	Speed: 3221.690442 samples/sec	accuracy=0.643407	lr=0.191816
[Epoch 63] training: accuracy=0.643406
[Epoch 63] speed: 3004 samples/sec	time cost: 436.529593
[Epoch 63] validation: err-top1=0.332540 err-top5=0.122579
Epoch[64] Batch [49]	Speed: 3111.362320 samples/sec	accuracy=0.659570	lr=0.191593
Epoch[64] Batch [99]	Speed: 2959.009936 samples/sec	accuracy=0.657158	lr=0.191375
Epoch[64] Batch [149]	Speed: 2939.509335 samples/sec	accuracy=0.656296	lr=0.191157
Epoch[64] Batch [199]	Speed: 3010.457094 samples/sec	accuracy=0.654341	lr=0.190939
Epoch[64] Batch [249]	Speed: 2959.588197 samples/sec	accuracy=0.652891	lr=0.190720
Epoch[64] Batch [299]	Speed: 2964.385127 samples/sec	accuracy=0.652207	lr=0.190502
Epoch[64] Batch [349]	Speed: 2952.010634 samples/sec	accuracy=0.651124	lr=0.190284
Epoch[64] Batch [399]	Speed: 3006.636994 samples/sec	accuracy=0.650972	lr=0.190066
Epoch[64] Batch [449]	Speed: 3141.828282 samples/sec	accuracy=0.650692	lr=0.189848
Epoch[64] Batch [499]	Speed: 3053.492619 samples/sec	accuracy=0.650682	lr=0.189630
Epoch[64] Batch [549]	Speed: 3045.045964 samples/sec	accuracy=0.650691	lr=0.189412
Epoch[64] Batch [599]	Speed: 3088.205794 samples/sec	accuracy=0.650081	lr=0.189194
Epoch[64] Batch [649]	Speed: 3142.529831 samples/sec	accuracy=0.649650	lr=0.188976
Epoch[64] Batch [699]	Speed: 3004.037426 samples/sec	accuracy=0.649489	lr=0.188758
Epoch[64] Batch [749]	Speed: 2962.023450 samples/sec	accuracy=0.649125	lr=0.188540
Epoch[64] Batch [799]	Speed: 2915.916808 samples/sec	accuracy=0.648560	lr=0.188322
Epoch[64] Batch [849]	Speed: 2931.720972 samples/sec	accuracy=0.648186	lr=0.188104
Epoch[64] Batch [899]	Speed: 2940.300717 samples/sec	accuracy=0.648023	lr=0.187886
Epoch[64] Batch [949]	Speed: 2981.716569 samples/sec	accuracy=0.647695	lr=0.187668
Epoch[64] Batch [999]	Speed: 3044.658840 samples/sec	accuracy=0.647469	lr=0.187450
Epoch[64] Batch [1049]	Speed: 2912.783014 samples/sec	accuracy=0.647363	lr=0.187232
Epoch[64] Batch [1099]	Speed: 3004.052218 samples/sec	accuracy=0.647052	lr=0.187014
Epoch[64] Batch [1149]	Speed: 2980.054324 samples/sec	accuracy=0.646867	lr=0.186796
Epoch[64] Batch [1199]	Speed: 2933.800508 samples/sec	accuracy=0.646745	lr=0.186578
Epoch[64] Batch [1249]	Speed: 3196.559270 samples/sec	accuracy=0.646645	lr=0.186360
[Epoch 64] training: accuracy=0.646633
[Epoch 64] speed: 3003 samples/sec	time cost: 437.622659
[Epoch 64] validation: err-top1=0.339007 err-top5=0.123545
Epoch[65] Batch [49]	Speed: 3044.058253 samples/sec	accuracy=0.660723	lr=0.186138
Epoch[65] Batch [99]	Speed: 3009.674573 samples/sec	accuracy=0.658096	lr=0.185920
Epoch[65] Batch [149]	Speed: 2991.299296 samples/sec	accuracy=0.657038	lr=0.185702
Epoch[65] Batch [199]	Speed: 3040.982429 samples/sec	accuracy=0.656553	lr=0.185485
Epoch[65] Batch [249]	Speed: 2900.645585 samples/sec	accuracy=0.655375	lr=0.185267
Epoch[65] Batch [299]	Speed: 3067.290215 samples/sec	accuracy=0.654570	lr=0.185049
Epoch[65] Batch [349]	Speed: 2956.802593 samples/sec	accuracy=0.654347	lr=0.184831
Epoch[65] Batch [399]	Speed: 3044.676927 samples/sec	accuracy=0.653618	lr=0.184614
Epoch[65] Batch [449]	Speed: 2888.851870 samples/sec	accuracy=0.653014	lr=0.184396
Epoch[65] Batch [499]	Speed: 2976.191325 samples/sec	accuracy=0.652379	lr=0.184178
Epoch[65] Batch [549]	Speed: 3097.802321 samples/sec	accuracy=0.651950	lr=0.183960
Epoch[65] Batch [599]	Speed: 2970.587814 samples/sec	accuracy=0.651419	lr=0.183743
Epoch[65] Batch [649]	Speed: 2975.483737 samples/sec	accuracy=0.650954	lr=0.183525
Epoch[65] Batch [699]	Speed: 2971.950014 samples/sec	accuracy=0.650677	lr=0.183308
Epoch[65] Batch [749]	Speed: 3028.917814 samples/sec	accuracy=0.650277	lr=0.183090
Epoch[65] Batch [799]	Speed: 3125.842134 samples/sec	accuracy=0.650042	lr=0.182872
Epoch[65] Batch [849]	Speed: 3037.280735 samples/sec	accuracy=0.649631	lr=0.182655
Epoch[65] Batch [899]	Speed: 2981.295671 samples/sec	accuracy=0.649403	lr=0.182437
Epoch[65] Batch [949]	Speed: 2824.445801 samples/sec	accuracy=0.649123	lr=0.182220
Epoch[65] Batch [999]	Speed: 3018.695288 samples/sec	accuracy=0.648783	lr=0.182002
Epoch[65] Batch [1049]	Speed: 3081.767245 samples/sec	accuracy=0.648728	lr=0.181785
Epoch[65] Batch [1099]	Speed: 3028.088990 samples/sec	accuracy=0.648533	lr=0.181567
Epoch[65] Batch [1149]	Speed: 2931.868867 samples/sec	accuracy=0.648478	lr=0.181350
Epoch[65] Batch [1199]	Speed: 3359.990317 samples/sec	accuracy=0.648337	lr=0.181132
Epoch[65] Batch [1249]	Speed: 3253.566312 samples/sec	accuracy=0.648091	lr=0.180915
[Epoch 65] training: accuracy=0.648097
[Epoch 65] speed: 3018 samples/sec	time cost: 435.149325
[Epoch 65] validation: err-top1=0.325973 err-top5=0.115175
Epoch[66] Batch [49]	Speed: 2996.080731 samples/sec	accuracy=0.658906	lr=0.180693
Epoch[66] Batch [99]	Speed: 3053.045485 samples/sec	accuracy=0.658984	lr=0.180476
Epoch[66] Batch [149]	Speed: 3010.609874 samples/sec	accuracy=0.658555	lr=0.180259
Epoch[66] Batch [199]	Speed: 2933.297426 samples/sec	accuracy=0.657939	lr=0.180041
Epoch[66] Batch [249]	Speed: 2885.915367 samples/sec	accuracy=0.657687	lr=0.179824
Epoch[66] Batch [299]	Speed: 3011.650623 samples/sec	accuracy=0.657288	lr=0.179607
Epoch[66] Batch [349]	Speed: 2936.760378 samples/sec	accuracy=0.656571	lr=0.179390
Epoch[66] Batch [399]	Speed: 3094.152333 samples/sec	accuracy=0.656252	lr=0.179172
Epoch[66] Batch [449]	Speed: 2971.817172 samples/sec	accuracy=0.656306	lr=0.178955
Epoch[66] Batch [499]	Speed: 3099.149806 samples/sec	accuracy=0.655533	lr=0.178738
Epoch[66] Batch [549]	Speed: 2990.481768 samples/sec	accuracy=0.655321	lr=0.178521
Epoch[66] Batch [599]	Speed: 3036.153216 samples/sec	accuracy=0.654880	lr=0.178304
Epoch[66] Batch [649]	Speed: 2983.175730 samples/sec	accuracy=0.654477	lr=0.178087
Epoch[66] Batch [699]	Speed: 2943.883583 samples/sec	accuracy=0.654097	lr=0.177870
Epoch[66] Batch [749]	Speed: 2926.739477 samples/sec	accuracy=0.653617	lr=0.177653
Epoch[66] Batch [799]	Speed: 3040.609899 samples/sec	accuracy=0.653214	lr=0.177436
Epoch[66] Batch [849]	Speed: 3058.755658 samples/sec	accuracy=0.652792	lr=0.177219
Epoch[66] Batch [899]	Speed: 2903.893245 samples/sec	accuracy=0.652368	lr=0.177002
Epoch[66] Batch [949]	Speed: 2979.495033 samples/sec	accuracy=0.652099	lr=0.176785
Epoch[66] Batch [999]	Speed: 3131.738472 samples/sec	accuracy=0.651864	lr=0.176568
Epoch[66] Batch [1049]	Speed: 2959.434068 samples/sec	accuracy=0.651687	lr=0.176351
Epoch[66] Batch [1099]	Speed: 2819.159434 samples/sec	accuracy=0.651435	lr=0.176134
Epoch[66] Batch [1149]	Speed: 3001.187988 samples/sec	accuracy=0.651211	lr=0.175918
Epoch[66] Batch [1199]	Speed: 3161.319278 samples/sec	accuracy=0.650984	lr=0.175701
Epoch[66] Batch [1249]	Speed: 3284.292593 samples/sec	accuracy=0.650772	lr=0.175484
[Epoch 66] training: accuracy=0.650779
[Epoch 66] speed: 3003 samples/sec	time cost: 437.061454
[Epoch 66] validation: err-top1=0.320711 err-top5=0.114557
Epoch[67] Batch [49]	Speed: 3325.639752 samples/sec	accuracy=0.656855	lr=0.175263
Epoch[67] Batch [99]	Speed: 3008.523493 samples/sec	accuracy=0.656787	lr=0.175046
Epoch[67] Batch [149]	Speed: 2923.827350 samples/sec	accuracy=0.658887	lr=0.174830
Epoch[67] Batch [199]	Speed: 2922.856231 samples/sec	accuracy=0.658896	lr=0.174613
Epoch[67] Batch [249]	Speed: 3009.080076 samples/sec	accuracy=0.658512	lr=0.174397
Epoch[67] Batch [299]	Speed: 3029.334661 samples/sec	accuracy=0.657887	lr=0.174180
Epoch[67] Batch [349]	Speed: 2842.637483 samples/sec	accuracy=0.657946	lr=0.173964
Epoch[67] Batch [399]	Speed: 2987.383295 samples/sec	accuracy=0.657773	lr=0.173747
Epoch[67] Batch [449]	Speed: 3178.415395 samples/sec	accuracy=0.657253	lr=0.173531
Epoch[67] Batch [499]	Speed: 2944.637228 samples/sec	accuracy=0.656936	lr=0.173314
Epoch[67] Batch [549]	Speed: 3023.837807 samples/sec	accuracy=0.656697	lr=0.173098
Epoch[67] Batch [599]	Speed: 3131.673529 samples/sec	accuracy=0.656141	lr=0.172881
Epoch[67] Batch [649]	Speed: 3069.907286 samples/sec	accuracy=0.655912	lr=0.172665
Epoch[67] Batch [699]	Speed: 3064.136771 samples/sec	accuracy=0.655598	lr=0.172449
Epoch[67] Batch [749]	Speed: 2866.598334 samples/sec	accuracy=0.655581	lr=0.172232
Epoch[67] Batch [799]	Speed: 2992.217789 samples/sec	accuracy=0.655486	lr=0.172016
Epoch[67] Batch [849]	Speed: 2927.358944 samples/sec	accuracy=0.655177	lr=0.171800
Epoch[67] Batch [899]	Speed: 2958.595954 samples/sec	accuracy=0.655004	lr=0.171584
Epoch[67] Batch [949]	Speed: 2973.571935 samples/sec	accuracy=0.654856	lr=0.171368
Epoch[67] Batch [999]	Speed: 3018.731951 samples/sec	accuracy=0.654612	lr=0.171152
Epoch[67] Batch [1049]	Speed: 2980.188028 samples/sec	accuracy=0.654587	lr=0.170935
Epoch[67] Batch [1099]	Speed: 2983.358910 samples/sec	accuracy=0.654381	lr=0.170719
Epoch[67] Batch [1149]	Speed: 3024.118040 samples/sec	accuracy=0.654085	lr=0.170503
Epoch[67] Batch [1199]	Speed: 2988.102791 samples/sec	accuracy=0.653743	lr=0.170287
Epoch[67] Batch [1249]	Speed: 3201.778971 samples/sec	accuracy=0.653475	lr=0.170072
[Epoch 67] training: accuracy=0.653488
[Epoch 67] speed: 3009 samples/sec	time cost: 436.224233
[Epoch 67] validation: err-top1=0.324737 err-top5=0.116968
Epoch[68] Batch [49]	Speed: 2931.482772 samples/sec	accuracy=0.665977	lr=0.169851
Epoch[68] Batch [99]	Speed: 3046.579409 samples/sec	accuracy=0.664277	lr=0.169635
Epoch[68] Batch [149]	Speed: 2963.781920 samples/sec	accuracy=0.662650	lr=0.169420
Epoch[68] Batch [199]	Speed: 2858.488803 samples/sec	accuracy=0.661450	lr=0.169204
Epoch[68] Batch [249]	Speed: 2976.864669 samples/sec	accuracy=0.661578	lr=0.168988
Epoch[68] Batch [299]	Speed: 2988.591534 samples/sec	accuracy=0.660732	lr=0.168772
Epoch[68] Batch [349]	Speed: 3018.568715 samples/sec	accuracy=0.660700	lr=0.168557
Epoch[68] Batch [399]	Speed: 3004.138620 samples/sec	accuracy=0.659773	lr=0.168341
Epoch[68] Batch [449]	Speed: 2986.840067 samples/sec	accuracy=0.658997	lr=0.168126
Epoch[68] Batch [499]	Speed: 3035.900833 samples/sec	accuracy=0.659033	lr=0.167910
Epoch[68] Batch [549]	Speed: 3044.795339 samples/sec	accuracy=0.658853	lr=0.167694
Epoch[68] Batch [599]	Speed: 3047.347510 samples/sec	accuracy=0.658631	lr=0.167479
Epoch[68] Batch [649]	Speed: 3287.090868 samples/sec	accuracy=0.658272	lr=0.167263
Epoch[68] Batch [699]	Speed: 2939.409552 samples/sec	accuracy=0.658207	lr=0.167048
Epoch[68] Batch [749]	Speed: 3011.960116 samples/sec	accuracy=0.658039	lr=0.166833
Epoch[68] Batch [799]	Speed: 3130.863842 samples/sec	accuracy=0.658131	lr=0.166617
Epoch[68] Batch [849]	Speed: 3050.396756 samples/sec	accuracy=0.657786	lr=0.166402
Epoch[68] Batch [899]	Speed: 2965.200239 samples/sec	accuracy=0.657473	lr=0.166187
Epoch[68] Batch [949]	Speed: 3100.946845 samples/sec	accuracy=0.657215	lr=0.165972
Epoch[68] Batch [999]	Speed: 3009.839676 samples/sec	accuracy=0.657021	lr=0.165756
Epoch[68] Batch [1049]	Speed: 3148.652957 samples/sec	accuracy=0.656987	lr=0.165541
Epoch[68] Batch [1099]	Speed: 2980.090055 samples/sec	accuracy=0.656806	lr=0.165326
Epoch[68] Batch [1149]	Speed: 2970.097752 samples/sec	accuracy=0.656650	lr=0.165111
Epoch[68] Batch [1199]	Speed: 2940.912245 samples/sec	accuracy=0.656356	lr=0.164896
Epoch[68] Batch [1249]	Speed: 3180.693984 samples/sec	accuracy=0.656167	lr=0.164681
[Epoch 68] training: accuracy=0.656169
[Epoch 68] speed: 3020 samples/sec	time cost: 434.876567
[Epoch 68] validation: err-top1=0.335160 err-top5=0.125937
Epoch[69] Batch [49]	Speed: 3103.146930 samples/sec	accuracy=0.667305	lr=0.164462
Epoch[69] Batch [99]	Speed: 2984.640303 samples/sec	accuracy=0.666748	lr=0.164247
Epoch[69] Batch [149]	Speed: 3042.139568 samples/sec	accuracy=0.664798	lr=0.164032
Epoch[69] Batch [199]	Speed: 2985.841633 samples/sec	accuracy=0.665088	lr=0.163818
Epoch[69] Batch [249]	Speed: 2972.517010 samples/sec	accuracy=0.664617	lr=0.163603
Epoch[69] Batch [299]	Speed: 3049.900325 samples/sec	accuracy=0.664235	lr=0.163388
Epoch[69] Batch [349]	Speed: 3015.240197 samples/sec	accuracy=0.663488	lr=0.163173
Epoch[69] Batch [399]	Speed: 3087.897529 samples/sec	accuracy=0.663269	lr=0.162959
Epoch[69] Batch [449]	Speed: 3037.583788 samples/sec	accuracy=0.662606	lr=0.162744
Epoch[69] Batch [499]	Speed: 3019.151178 samples/sec	accuracy=0.662355	lr=0.162530
Epoch[69] Batch [549]	Speed: 2941.363997 samples/sec	accuracy=0.661678	lr=0.162315
Epoch[69] Batch [599]	Speed: 3059.296074 samples/sec	accuracy=0.661724	lr=0.162101
Epoch[69] Batch [649]	Speed: 2951.515445 samples/sec	accuracy=0.661235	lr=0.161886
Epoch[69] Batch [699]	Speed: 3015.862205 samples/sec	accuracy=0.660968	lr=0.161672
Epoch[69] Batch [749]	Speed: 3010.960946 samples/sec	accuracy=0.660711	lr=0.161458
Epoch[69] Batch [799]	Speed: 3023.211994 samples/sec	accuracy=0.660466	lr=0.161244
Epoch[69] Batch [849]	Speed: 2911.529487 samples/sec	accuracy=0.660076	lr=0.161029
Epoch[69] Batch [899]	Speed: 2950.475663 samples/sec	accuracy=0.660043	lr=0.160815
Epoch[69] Batch [949]	Speed: 3160.157700 samples/sec	accuracy=0.659830	lr=0.160601
Epoch[69] Batch [999]	Speed: 2969.741975 samples/sec	accuracy=0.659656	lr=0.160387
Epoch[69] Batch [1049]	Speed: 3115.221418 samples/sec	accuracy=0.659440	lr=0.160173
Epoch[69] Batch [1099]	Speed: 2924.353551 samples/sec	accuracy=0.659237	lr=0.159959
Epoch[69] Batch [1149]	Speed: 2961.773479 samples/sec	accuracy=0.659006	lr=0.159745
Epoch[69] Batch [1199]	Speed: 3034.687838 samples/sec	accuracy=0.658939	lr=0.159531
Epoch[69] Batch [1249]	Speed: 3056.106846 samples/sec	accuracy=0.658724	lr=0.159317
[Epoch 69] training: accuracy=0.658725
[Epoch 69] speed: 3012 samples/sec	time cost: 435.583399
[Epoch 69] validation: err-top1=0.312642 err-top5=0.108256
Epoch[70] Batch [49]	Speed: 3123.610534 samples/sec	accuracy=0.671016	lr=0.159099
Epoch[70] Batch [99]	Speed: 3118.776547 samples/sec	accuracy=0.670742	lr=0.158886
Epoch[70] Batch [149]	Speed: 2940.777451 samples/sec	accuracy=0.668757	lr=0.158672
Epoch[70] Batch [199]	Speed: 3085.453131 samples/sec	accuracy=0.667852	lr=0.158458
Epoch[70] Batch [249]	Speed: 3032.949662 samples/sec	accuracy=0.667348	lr=0.158245
Epoch[70] Batch [299]	Speed: 3007.238654 samples/sec	accuracy=0.667012	lr=0.158031
Epoch[70] Batch [349]	Speed: 2847.505030 samples/sec	accuracy=0.666621	lr=0.157818
Epoch[70] Batch [399]	Speed: 2984.349753 samples/sec	accuracy=0.666582	lr=0.157604
Epoch[70] Batch [449]	Speed: 2963.431335 samples/sec	accuracy=0.665855	lr=0.157391
Epoch[70] Batch [499]	Speed: 3037.985575 samples/sec	accuracy=0.665518	lr=0.157177
Epoch[70] Batch [549]	Speed: 2960.003315 samples/sec	accuracy=0.664947	lr=0.156964
Epoch[70] Batch [599]	Speed: 3034.040124 samples/sec	accuracy=0.664328	lr=0.156751
Epoch[70] Batch [649]	Speed: 3132.627575 samples/sec	accuracy=0.663681	lr=0.156538
Epoch[70] Batch [699]	Speed: 2927.128194 samples/sec	accuracy=0.663117	lr=0.156325
Epoch[70] Batch [749]	Speed: 3028.213076 samples/sec	accuracy=0.663025	lr=0.156112
Epoch[70] Batch [799]	Speed: 3046.437477 samples/sec	accuracy=0.662758	lr=0.155899
Epoch[70] Batch [849]	Speed: 3045.752253 samples/sec	accuracy=0.662769	lr=0.155686
Epoch[70] Batch [899]	Speed: 2955.102531 samples/sec	accuracy=0.662667	lr=0.155473
Epoch[70] Batch [949]	Speed: 3032.588776 samples/sec	accuracy=0.662381	lr=0.155260
Epoch[70] Batch [999]	Speed: 2978.902686 samples/sec	accuracy=0.662076	lr=0.155047
Epoch[70] Batch [1049]	Speed: 2983.296659 samples/sec	accuracy=0.661827	lr=0.154834
Epoch[70] Batch [1099]	Speed: 2947.690063 samples/sec	accuracy=0.661531	lr=0.154622
Epoch[70] Batch [1149]	Speed: 3115.638990 samples/sec	accuracy=0.661260	lr=0.154409
Epoch[70] Batch [1199]	Speed: 3043.120126 samples/sec	accuracy=0.661104	lr=0.154196
Epoch[70] Batch [1249]	Speed: 3201.147395 samples/sec	accuracy=0.661044	lr=0.153984
[Epoch 70] training: accuracy=0.661035
[Epoch 70] speed: 3018 samples/sec	time cost: 435.744728
[Epoch 70] validation: err-top1=0.332549 err-top5=0.122608
Epoch[71] Batch [49]	Speed: 3139.427898 samples/sec	accuracy=0.672461	lr=0.153767
Epoch[71] Batch [99]	Speed: 3004.433540 samples/sec	accuracy=0.670693	lr=0.153555
Epoch[71] Batch [149]	Speed: 2882.959979 samples/sec	accuracy=0.670586	lr=0.153342
Epoch[71] Batch [199]	Speed: 3119.787153 samples/sec	accuracy=0.670532	lr=0.153130
Epoch[71] Batch [249]	Speed: 3167.105062 samples/sec	accuracy=0.670129	lr=0.152918
Epoch[71] Batch [299]	Speed: 3006.169179 samples/sec	accuracy=0.669837	lr=0.152705
Epoch[71] Batch [349]	Speed: 3062.462705 samples/sec	accuracy=0.669205	lr=0.152493
Epoch[71] Batch [399]	Speed: 2993.692466 samples/sec	accuracy=0.669395	lr=0.152281
Epoch[71] Batch [449]	Speed: 2968.327597 samples/sec	accuracy=0.668570	lr=0.152069
Epoch[71] Batch [499]	Speed: 2889.511035 samples/sec	accuracy=0.667891	lr=0.151857
Epoch[71] Batch [549]	Speed: 2934.287244 samples/sec	accuracy=0.667884	lr=0.151645
Epoch[71] Batch [599]	Speed: 3057.969908 samples/sec	accuracy=0.667404	lr=0.151433
Epoch[71] Batch [649]	Speed: 3019.961950 samples/sec	accuracy=0.666892	lr=0.151222
Epoch[71] Batch [699]	Speed: 2998.742750 samples/sec	accuracy=0.666636	lr=0.151010
Epoch[71] Batch [749]	Speed: 3007.031814 samples/sec	accuracy=0.666504	lr=0.150798
Epoch[71] Batch [799]	Speed: 3039.110799 samples/sec	accuracy=0.666262	lr=0.150587
Epoch[71] Batch [849]	Speed: 2975.880685 samples/sec	accuracy=0.665866	lr=0.150375
Epoch[71] Batch [899]	Speed: 3111.326798 samples/sec	accuracy=0.665503	lr=0.150163
Epoch[71] Batch [949]	Speed: 3032.078945 samples/sec	accuracy=0.665095	lr=0.149952
Epoch[71] Batch [999]	Speed: 2964.871259 samples/sec	accuracy=0.664929	lr=0.149741
Epoch[71] Batch [1049]	Speed: 3020.727565 samples/sec	accuracy=0.664627	lr=0.149529
Epoch[71] Batch [1099]	Speed: 3033.307807 samples/sec	accuracy=0.664410	lr=0.149318
Epoch[71] Batch [1149]	Speed: 2860.475045 samples/sec	accuracy=0.664286	lr=0.149107
Epoch[71] Batch [1199]	Speed: 2912.803400 samples/sec	accuracy=0.664167	lr=0.148896
Epoch[71] Batch [1249]	Speed: 3114.110127 samples/sec	accuracy=0.664219	lr=0.148685
[Epoch 71] training: accuracy=0.664234
[Epoch 71] speed: 3008 samples/sec	time cost: 436.297977
[Epoch 71] validation: err-top1=0.323820 err-top5=0.115015
Epoch[72] Batch [49]	Speed: 2878.820466 samples/sec	accuracy=0.673535	lr=0.148469
Epoch[72] Batch [99]	Speed: 2962.761152 samples/sec	accuracy=0.674424	lr=0.148258
Epoch[72] Batch [149]	Speed: 3002.042520 samples/sec	accuracy=0.673522	lr=0.148047
Epoch[72] Batch [199]	Speed: 2953.658632 samples/sec	accuracy=0.672671	lr=0.147837
Epoch[72] Batch [249]	Speed: 3019.746774 samples/sec	accuracy=0.672074	lr=0.147626
Epoch[72] Batch [299]	Speed: 2916.406935 samples/sec	accuracy=0.671348	lr=0.147415
Epoch[72] Batch [349]	Speed: 3014.860445 samples/sec	accuracy=0.670865	lr=0.147204
Epoch[72] Batch [399]	Speed: 3052.765897 samples/sec	accuracy=0.670562	lr=0.146994
Epoch[72] Batch [449]	Speed: 3023.562266 samples/sec	accuracy=0.670538	lr=0.146783
Epoch[72] Batch [499]	Speed: 3049.218998 samples/sec	accuracy=0.670291	lr=0.146573
Epoch[72] Batch [549]	Speed: 2920.384686 samples/sec	accuracy=0.670108	lr=0.146362
Epoch[72] Batch [599]	Speed: 2914.492150 samples/sec	accuracy=0.669598	lr=0.146152
Epoch[72] Batch [649]	Speed: 2938.647843 samples/sec	accuracy=0.669551	lr=0.145942
Epoch[72] Batch [699]	Speed: 2946.159783 samples/sec	accuracy=0.669361	lr=0.145732
Epoch[72] Batch [749]	Speed: 2921.613020 samples/sec	accuracy=0.669055	lr=0.145521
Epoch[72] Batch [799]	Speed: 2947.964857 samples/sec	accuracy=0.668867	lr=0.145311
Epoch[72] Batch [849]	Speed: 2988.233268 samples/sec	accuracy=0.668626	lr=0.145101
Epoch[72] Batch [899]	Speed: 2934.731790 samples/sec	accuracy=0.668278	lr=0.144891
Epoch[72] Batch [949]	Speed: 3008.213779 samples/sec	accuracy=0.668140	lr=0.144682
Epoch[72] Batch [999]	Speed: 2962.649566 samples/sec	accuracy=0.667931	lr=0.144472
Epoch[72] Batch [1049]	Speed: 2849.235131 samples/sec	accuracy=0.667713	lr=0.144262
Epoch[72] Batch [1099]	Speed: 2965.071275 samples/sec	accuracy=0.667569	lr=0.144052
Epoch[72] Batch [1149]	Speed: 2944.053897 samples/sec	accuracy=0.667307	lr=0.143843
Epoch[72] Batch [1199]	Speed: 3038.690527 samples/sec	accuracy=0.667059	lr=0.143633
Epoch[72] Batch [1249]	Speed: 3169.808562 samples/sec	accuracy=0.666862	lr=0.143424
[Epoch 72] training: accuracy=0.666881
[Epoch 72] speed: 2969 samples/sec	time cost: 441.688803
[Epoch 72] validation: err-top1=0.349330 err-top5=0.135862
Epoch[73] Batch [49]	Speed: 2901.988930 samples/sec	accuracy=0.676094	lr=0.143210
Epoch[73] Batch [99]	Speed: 2902.778363 samples/sec	accuracy=0.674990	lr=0.143001
Epoch[73] Batch [149]	Speed: 3100.849592 samples/sec	accuracy=0.676074	lr=0.142791
Epoch[73] Batch [199]	Speed: 2975.105112 samples/sec	accuracy=0.674644	lr=0.142582
Epoch[73] Batch [249]	Speed: 3163.702091 samples/sec	accuracy=0.674656	lr=0.142373
Epoch[73] Batch [299]	Speed: 3106.481115 samples/sec	accuracy=0.674502	lr=0.142164
Epoch[73] Batch [349]	Speed: 2888.349283 samples/sec	accuracy=0.674163	lr=0.141955
Epoch[73] Batch [399]	Speed: 2965.564880 samples/sec	accuracy=0.674084	lr=0.141746
Epoch[73] Batch [449]	Speed: 2987.372614 samples/sec	accuracy=0.673865	lr=0.141537
Epoch[73] Batch [499]	Speed: 3100.538037 samples/sec	accuracy=0.673264	lr=0.141328
Epoch[73] Batch [549]	Speed: 3016.446590 samples/sec	accuracy=0.672827	lr=0.141120
Epoch[73] Batch [599]	Speed: 2920.749511 samples/sec	accuracy=0.672282	lr=0.140911
Epoch[73] Batch [649]	Speed: 3036.034402 samples/sec	accuracy=0.671804	lr=0.140702
Epoch[73] Batch [699]	Speed: 3042.400273 samples/sec	accuracy=0.671655	lr=0.140494
Epoch[73] Batch [749]	Speed: 3083.043759 samples/sec	accuracy=0.671408	lr=0.140285
Epoch[73] Batch [799]	Speed: 3004.939960 samples/sec	accuracy=0.671057	lr=0.140077
Epoch[73] Batch [849]	Speed: 3053.568992 samples/sec	accuracy=0.670813	lr=0.139869
Epoch[73] Batch [899]	Speed: 3016.326602 samples/sec	accuracy=0.670590	lr=0.139661
Epoch[73] Batch [949]	Speed: 3044.133507 samples/sec	accuracy=0.670611	lr=0.139452
Epoch[73] Batch [999]	Speed: 3085.460224 samples/sec	accuracy=0.670325	lr=0.139244
Epoch[73] Batch [1049]	Speed: 2961.928874 samples/sec	accuracy=0.670041	lr=0.139036
Epoch[73] Batch [1099]	Speed: 2927.478981 samples/sec	accuracy=0.669878	lr=0.138828
Epoch[73] Batch [1149]	Speed: 3034.372586 samples/sec	accuracy=0.669597	lr=0.138620
Epoch[73] Batch [1199]	Speed: 3072.597978 samples/sec	accuracy=0.669333	lr=0.138413
Epoch[73] Batch [1249]	Speed: 2998.868671 samples/sec	accuracy=0.669067	lr=0.138205
[Epoch 73] training: accuracy=0.669081
[Epoch 73] speed: 3011 samples/sec	time cost: 436.262723
[Epoch 73] validation: err-top1=0.310288 err-top5=0.108159
Epoch[74] Batch [49]	Speed: 2977.431601 samples/sec	accuracy=0.679473	lr=0.137993
Epoch[74] Batch [99]	Speed: 2913.894209 samples/sec	accuracy=0.678076	lr=0.137786
Epoch[74] Batch [149]	Speed: 2925.802578 samples/sec	accuracy=0.677799	lr=0.137578
Epoch[74] Batch [199]	Speed: 2986.801890 samples/sec	accuracy=0.677690	lr=0.137371
Epoch[74] Batch [249]	Speed: 2936.430892 samples/sec	accuracy=0.677289	lr=0.137163
Epoch[74] Batch [299]	Speed: 2884.252899 samples/sec	accuracy=0.677269	lr=0.136956
Epoch[74] Batch [349]	Speed: 3033.666893 samples/sec	accuracy=0.676398	lr=0.136749
Epoch[74] Batch [399]	Speed: 3173.065462 samples/sec	accuracy=0.676287	lr=0.136542
Epoch[74] Batch [449]	Speed: 2950.756410 samples/sec	accuracy=0.675447	lr=0.136335
Epoch[74] Batch [499]	Speed: 3090.425277 samples/sec	accuracy=0.675629	lr=0.136128
Epoch[74] Batch [549]	Speed: 3018.186894 samples/sec	accuracy=0.675433	lr=0.135921
Epoch[74] Batch [599]	Speed: 2895.451701 samples/sec	accuracy=0.675412	lr=0.135714
Epoch[74] Batch [649]	Speed: 2922.620821 samples/sec	accuracy=0.675084	lr=0.135507
Epoch[74] Batch [699]	Speed: 2981.404692 samples/sec	accuracy=0.674841	lr=0.135301
Epoch[74] Batch [749]	Speed: 2926.607655 samples/sec	accuracy=0.674406	lr=0.135094
Epoch[74] Batch [799]	Speed: 2994.971893 samples/sec	accuracy=0.674340	lr=0.134887
Epoch[74] Batch [849]	Speed: 2989.915600 samples/sec	accuracy=0.674162	lr=0.134681
Epoch[74] Batch [899]	Speed: 2983.724714 samples/sec	accuracy=0.674034	lr=0.134475
Epoch[74] Batch [949]	Speed: 3148.431285 samples/sec	accuracy=0.674030	lr=0.134268
Epoch[74] Batch [999]	Speed: 2877.344647 samples/sec	accuracy=0.673880	lr=0.134062
Epoch[74] Batch [1049]	Speed: 2963.614020 samples/sec	accuracy=0.673603	lr=0.133856
Epoch[74] Batch [1099]	Speed: 2929.209248 samples/sec	accuracy=0.673471	lr=0.133650
Epoch[74] Batch [1149]	Speed: 3029.537913 samples/sec	accuracy=0.673061	lr=0.133444
Epoch[74] Batch [1199]	Speed: 3200.482107 samples/sec	accuracy=0.672846	lr=0.133238
Epoch[74] Batch [1249]	Speed: 3331.325186 samples/sec	accuracy=0.672677	lr=0.133032
[Epoch 74] training: accuracy=0.672695
[Epoch 74] speed: 2996 samples/sec	time cost: 438.137887
[Epoch 74] validation: err-top1=0.315888 err-top5=0.112404
Epoch[75] Batch [49]	Speed: 3102.004400 samples/sec	accuracy=0.686152	lr=0.132822
Epoch[75] Batch [99]	Speed: 2975.360967 samples/sec	accuracy=0.684619	lr=0.132617
Epoch[75] Batch [149]	Speed: 3002.778081 samples/sec	accuracy=0.684368	lr=0.132411
Epoch[75] Batch [199]	Speed: 2894.076763 samples/sec	accuracy=0.683936	lr=0.132206
Epoch[75] Batch [249]	Speed: 3031.817138 samples/sec	accuracy=0.683008	lr=0.132000
Epoch[75] Batch [299]	Speed: 2944.522885 samples/sec	accuracy=0.681680	lr=0.131795
Epoch[75] Batch [349]	Speed: 3006.566529 samples/sec	accuracy=0.680971	lr=0.131590
Epoch[75] Batch [399]	Speed: 3105.306223 samples/sec	accuracy=0.680281	lr=0.131385
Epoch[75] Batch [449]	Speed: 3084.102157 samples/sec	accuracy=0.679766	lr=0.131180
Epoch[75] Batch [499]	Speed: 2921.474465 samples/sec	accuracy=0.679146	lr=0.130975
Epoch[75] Batch [549]	Speed: 3051.094736 samples/sec	accuracy=0.678920	lr=0.130770
Epoch[75] Batch [599]	Speed: 2900.321058 samples/sec	accuracy=0.678429	lr=0.130565
Epoch[75] Batch [649]	Speed: 2923.707452 samples/sec	accuracy=0.678284	lr=0.130360
Epoch[75] Batch [699]	Speed: 3060.639531 samples/sec	accuracy=0.677931	lr=0.130155
Epoch[75] Batch [749]	Speed: 3072.990041 samples/sec	accuracy=0.677626	lr=0.129951
Epoch[75] Batch [799]	Speed: 2979.085878 samples/sec	accuracy=0.677324	lr=0.129746
Epoch[75] Batch [849]	Speed: 3066.083619 samples/sec	accuracy=0.677065	lr=0.129542
Epoch[75] Batch [899]	Speed: 2923.366761 samples/sec	accuracy=0.676885	lr=0.129338
Epoch[75] Batch [949]	Speed: 3082.653242 samples/sec	accuracy=0.676863	lr=0.129133
Epoch[75] Batch [999]	Speed: 2991.288796 samples/sec	accuracy=0.676579	lr=0.128929
Epoch[75] Batch [1049]	Speed: 3110.347749 samples/sec	accuracy=0.676411	lr=0.128725
Epoch[75] Batch [1099]	Speed: 2936.117898 samples/sec	accuracy=0.676231	lr=0.128521
Epoch[75] Batch [1149]	Speed: 3064.819402 samples/sec	accuracy=0.675864	lr=0.128317
Epoch[75] Batch [1199]	Speed: 2954.704805 samples/sec	accuracy=0.675625	lr=0.128113
Epoch[75] Batch [1249]	Speed: 3319.779689 samples/sec	accuracy=0.675470	lr=0.127910
[Epoch 75] training: accuracy=0.675489
[Epoch 75] speed: 3015 samples/sec	time cost: 434.778680
[Epoch 75] validation: err-top1=0.321757 err-top5=0.117920
Epoch[76] Batch [49]	Speed: 3190.003725 samples/sec	accuracy=0.684297	lr=0.127702
Epoch[76] Batch [99]	Speed: 3081.719350 samples/sec	accuracy=0.684092	lr=0.127498
Epoch[76] Batch [149]	Speed: 2879.325070 samples/sec	accuracy=0.683105	lr=0.127295
Epoch[76] Batch [199]	Speed: 3049.685583 samples/sec	accuracy=0.682695	lr=0.127092
Epoch[76] Batch [249]	Speed: 2904.821665 samples/sec	accuracy=0.681988	lr=0.126888
Epoch[76] Batch [299]	Speed: 3078.293516 samples/sec	accuracy=0.681839	lr=0.126685
Epoch[76] Batch [349]	Speed: 2922.119776 samples/sec	accuracy=0.681387	lr=0.126482
Epoch[76] Batch [399]	Speed: 3056.470089 samples/sec	accuracy=0.681362	lr=0.126279
Epoch[76] Batch [449]	Speed: 2992.589231 samples/sec	accuracy=0.680964	lr=0.126076
Epoch[76] Batch [499]	Speed: 2936.707165 samples/sec	accuracy=0.680561	lr=0.125873
Epoch[76] Batch [549]	Speed: 2865.713801 samples/sec	accuracy=0.680066	lr=0.125670
Epoch[76] Batch [599]	Speed: 2880.248197 samples/sec	accuracy=0.679792	lr=0.125468
Epoch[76] Batch [649]	Speed: 3079.494597 samples/sec	accuracy=0.679620	lr=0.125265
Epoch[76] Batch [699]	Speed: 2970.733122 samples/sec	accuracy=0.679671	lr=0.125062
Epoch[76] Batch [749]	Speed: 3090.673596 samples/sec	accuracy=0.679544	lr=0.124860
Epoch[76] Batch [799]	Speed: 3144.331966 samples/sec	accuracy=0.679421	lr=0.124658
Epoch[76] Batch [849]	Speed: 3031.484166 samples/sec	accuracy=0.679360	lr=0.124455
Epoch[76] Batch [899]	Speed: 3069.526671 samples/sec	accuracy=0.679094	lr=0.124253
Epoch[76] Batch [949]	Speed: 3042.273341 samples/sec	accuracy=0.678817	lr=0.124051
Epoch[76] Batch [999]	Speed: 3322.133054 samples/sec	accuracy=0.678638	lr=0.123849
Epoch[76] Batch [1049]	Speed: 2917.890834 samples/sec	accuracy=0.678417	lr=0.123647
Epoch[76] Batch [1099]	Speed: 2978.364273 samples/sec	accuracy=0.678380	lr=0.123446
Epoch[76] Batch [1149]	Speed: 2858.973096 samples/sec	accuracy=0.678244	lr=0.123244
Epoch[76] Batch [1199]	Speed: 2987.911171 samples/sec	accuracy=0.678193	lr=0.123042
Epoch[76] Batch [1249]	Speed: 3086.059211 samples/sec	accuracy=0.677976	lr=0.122841
[Epoch 76] training: accuracy=0.677983
[Epoch 76] speed: 3010 samples/sec	time cost: 435.599592
[Epoch 76] validation: err-top1=0.309809 err-top5=0.104731
Epoch[77] Batch [49]	Speed: 3164.900047 samples/sec	accuracy=0.691191	lr=0.122635
Epoch[77] Batch [99]	Speed: 3144.291176 samples/sec	accuracy=0.691113	lr=0.122434
Epoch[77] Batch [149]	Speed: 3006.758444 samples/sec	accuracy=0.689668	lr=0.122233
Epoch[77] Batch [199]	Speed: 2998.673408 samples/sec	accuracy=0.688081	lr=0.122032
Epoch[77] Batch [249]	Speed: 2957.049202 samples/sec	accuracy=0.687320	lr=0.121831
Epoch[77] Batch [299]	Speed: 3033.983456 samples/sec	accuracy=0.687038	lr=0.121630
Epoch[77] Batch [349]	Speed: 2971.945367 samples/sec	accuracy=0.686403	lr=0.121429
Epoch[77] Batch [399]	Speed: 2920.249901 samples/sec	accuracy=0.686433	lr=0.121228
Epoch[77] Batch [449]	Speed: 2910.738601 samples/sec	accuracy=0.686068	lr=0.121027
Epoch[77] Batch [499]	Speed: 3114.616705 samples/sec	accuracy=0.685822	lr=0.120827
Epoch[77] Batch [549]	Speed: 2992.011801 samples/sec	accuracy=0.685348	lr=0.120626
Epoch[77] Batch [599]	Speed: 3131.068758 samples/sec	accuracy=0.685146	lr=0.120426
Epoch[77] Batch [649]	Speed: 2930.805680 samples/sec	accuracy=0.684731	lr=0.120226
Epoch[77] Batch [699]	Speed: 2989.826060 samples/sec	accuracy=0.684323	lr=0.120025
Epoch[77] Batch [749]	Speed: 3078.112127 samples/sec	accuracy=0.683901	lr=0.119825
Epoch[77] Batch [799]	Speed: 3063.333614 samples/sec	accuracy=0.683569	lr=0.119625
Epoch[77] Batch [849]	Speed: 3071.598026 samples/sec	accuracy=0.683488	lr=0.119425
Epoch[77] Batch [899]	Speed: 3107.663912 samples/sec	accuracy=0.683220	lr=0.119226
Epoch[77] Batch [949]	Speed: 2985.308055 samples/sec	accuracy=0.682945	lr=0.119026
Epoch[77] Batch [999]	Speed: 2953.769704 samples/sec	accuracy=0.682775	lr=0.118826
Epoch[77] Batch [1049]	Speed: 2874.291708 samples/sec	accuracy=0.682478	lr=0.118627
Epoch[77] Batch [1099]	Speed: 2986.213238 samples/sec	accuracy=0.682254	lr=0.118427
Epoch[77] Batch [1149]	Speed: 2958.388782 samples/sec	accuracy=0.681974	lr=0.118228
Epoch[77] Batch [1199]	Speed: 2961.428596 samples/sec	accuracy=0.681973	lr=0.118029
Epoch[77] Batch [1249]	Speed: 3267.491071 samples/sec	accuracy=0.681796	lr=0.117830
[Epoch 77] training: accuracy=0.681785
[Epoch 77] speed: 3018 samples/sec	time cost: 434.875943
[Epoch 77] validation: err-top1=0.306999 err-top5=0.105030
Epoch[78] Batch [49]	Speed: 3297.819929 samples/sec	accuracy=0.694805	lr=0.117627
Epoch[78] Batch [99]	Speed: 3115.077177 samples/sec	accuracy=0.693623	lr=0.117428
Epoch[78] Batch [149]	Speed: 2974.215588 samples/sec	accuracy=0.692734	lr=0.117229
Epoch[78] Batch [199]	Speed: 3204.848759 samples/sec	accuracy=0.692720	lr=0.117030
Epoch[78] Batch [249]	Speed: 3035.958260 samples/sec	accuracy=0.691496	lr=0.116831
Epoch[78] Batch [299]	Speed: 3068.550421 samples/sec	accuracy=0.690088	lr=0.116633
Epoch[78] Batch [349]	Speed: 2901.489680 samples/sec	accuracy=0.689138	lr=0.116434
Epoch[78] Batch [399]	Speed: 2923.519982 samples/sec	accuracy=0.688601	lr=0.116236
Epoch[78] Batch [449]	Speed: 2943.526998 samples/sec	accuracy=0.688761	lr=0.116038
Epoch[78] Batch [499]	Speed: 2980.252920 samples/sec	accuracy=0.688412	lr=0.115840
Epoch[78] Batch [549]	Speed: 2978.626969 samples/sec	accuracy=0.687912	lr=0.115642
Epoch[78] Batch [599]	Speed: 3018.903184 samples/sec	accuracy=0.687708	lr=0.115444
Epoch[78] Batch [649]	Speed: 2992.512833 samples/sec	accuracy=0.687407	lr=0.115246
Epoch[78] Batch [699]	Speed: 3054.155095 samples/sec	accuracy=0.687021	lr=0.115048
Epoch[78] Batch [749]	Speed: 3021.321488 samples/sec	accuracy=0.686600	lr=0.114850
Epoch[78] Batch [799]	Speed: 3068.125825 samples/sec	accuracy=0.686218	lr=0.114653
Epoch[78] Batch [849]	Speed: 2967.443679 samples/sec	accuracy=0.685913	lr=0.114455
Epoch[78] Batch [899]	Speed: 2979.116915 samples/sec	accuracy=0.685522	lr=0.114258
Epoch[78] Batch [949]	Speed: 2947.991000 samples/sec	accuracy=0.685185	lr=0.114061
Epoch[78] Batch [999]	Speed: 3053.825580 samples/sec	accuracy=0.684858	lr=0.113864
Epoch[78] Batch [1049]	Speed: 3013.196143 samples/sec	accuracy=0.684914	lr=0.113667
Epoch[78] Batch [1099]	Speed: 2922.204789 samples/sec	accuracy=0.684722	lr=0.113470
Epoch[78] Batch [1149]	Speed: 2961.956776 samples/sec	accuracy=0.684552	lr=0.113273
Epoch[78] Batch [1199]	Speed: 3008.478564 samples/sec	accuracy=0.684513	lr=0.113076
Epoch[78] Batch [1249]	Speed: 3125.011220 samples/sec	accuracy=0.684533	lr=0.112880
[Epoch 78] training: accuracy=0.684538
[Epoch 78] speed: 3017 samples/sec	time cost: 435.104174
[Epoch 78] validation: err-top1=0.311384 err-top5=0.107980
Epoch[79] Batch [49]	Speed: 3143.521702 samples/sec	accuracy=0.696543	lr=0.112679
Epoch[79] Batch [99]	Speed: 3061.621752 samples/sec	accuracy=0.697451	lr=0.112483
Epoch[79] Batch [149]	Speed: 2857.950016 samples/sec	accuracy=0.695827	lr=0.112287
Epoch[79] Batch [199]	Speed: 3080.494527 samples/sec	accuracy=0.694351	lr=0.112090
Epoch[79] Batch [249]	Speed: 3026.069485 samples/sec	accuracy=0.694031	lr=0.111894
Epoch[79] Batch [299]	Speed: 2950.958135 samples/sec	accuracy=0.693704	lr=0.111698
Epoch[79] Batch [349]	Speed: 2973.776997 samples/sec	accuracy=0.693619	lr=0.111502
Epoch[79] Batch [399]	Speed: 3029.538127 samples/sec	accuracy=0.692935	lr=0.111307
Epoch[79] Batch [449]	Speed: 2954.800994 samples/sec	accuracy=0.692543	lr=0.111111
Epoch[79] Batch [499]	Speed: 2921.931834 samples/sec	accuracy=0.691588	lr=0.110915
Epoch[79] Batch [549]	Speed: 2859.317978 samples/sec	accuracy=0.691408	lr=0.110720
Epoch[79] Batch [599]	Speed: 3009.749277 samples/sec	accuracy=0.691253	lr=0.110525
Epoch[79] Batch [649]	Speed: 3200.247402 samples/sec	accuracy=0.690808	lr=0.110329
Epoch[79] Batch [699]	Speed: 2989.061549 samples/sec	accuracy=0.690347	lr=0.110134
Epoch[79] Batch [749]	Speed: 2983.663153 samples/sec	accuracy=0.690112	lr=0.109939
Epoch[79] Batch [799]	Speed: 3011.015322 samples/sec	accuracy=0.689625	lr=0.109744
Epoch[79] Batch [849]	Speed: 3026.366722 samples/sec	accuracy=0.689213	lr=0.109549
Epoch[79] Batch [899]	Speed: 2930.368041 samples/sec	accuracy=0.688981	lr=0.109355
Epoch[79] Batch [949]	Speed: 2937.342831 samples/sec	accuracy=0.688757	lr=0.109160
Epoch[79] Batch [999]	Speed: 3186.378441 samples/sec	accuracy=0.688567	lr=0.108966
Epoch[79] Batch [1049]	Speed: 3131.961089 samples/sec	accuracy=0.688432	lr=0.108771
Epoch[79] Batch [1099]	Speed: 3028.291008 samples/sec	accuracy=0.688286	lr=0.108577
Epoch[79] Batch [1149]	Speed: 2961.384246 samples/sec	accuracy=0.688236	lr=0.108383
Epoch[79] Batch [1199]	Speed: 3074.605294 samples/sec	accuracy=0.688071	lr=0.108189
Epoch[79] Batch [1249]	Speed: 3208.226460 samples/sec	accuracy=0.688004	lr=0.107995
[Epoch 79] training: accuracy=0.688014
[Epoch 79] speed: 3016 samples/sec	time cost: 434.461129
[Epoch 79] validation: err-top1=0.314015 err-top5=0.112145
Epoch[80] Batch [49]	Speed: 2955.613610 samples/sec	accuracy=0.702754	lr=0.107797
Epoch[80] Batch [99]	Speed: 3014.988952 samples/sec	accuracy=0.699053	lr=0.107603
Epoch[80] Batch [149]	Speed: 2932.239409 samples/sec	accuracy=0.697441	lr=0.107410
Epoch[80] Batch [199]	Speed: 2952.830728 samples/sec	accuracy=0.696577	lr=0.107216
Epoch[80] Batch [249]	Speed: 3001.315918 samples/sec	accuracy=0.696922	lr=0.107023
Epoch[80] Batch [299]	Speed: 3101.707844 samples/sec	accuracy=0.697018	lr=0.106830
Epoch[80] Batch [349]	Speed: 2919.740499 samples/sec	accuracy=0.696063	lr=0.106636
Epoch[80] Batch [399]	Speed: 3017.806908 samples/sec	accuracy=0.695491	lr=0.106443
Epoch[80] Batch [449]	Speed: 3000.199095 samples/sec	accuracy=0.695297	lr=0.106250
Epoch[80] Batch [499]	Speed: 3162.732335 samples/sec	accuracy=0.694541	lr=0.106057
Epoch[80] Batch [549]	Speed: 2960.772013 samples/sec	accuracy=0.694352	lr=0.105865
Epoch[80] Batch [599]	Speed: 3140.388190 samples/sec	accuracy=0.694028	lr=0.105672
Epoch[80] Batch [649]	Speed: 2982.721774 samples/sec	accuracy=0.693377	lr=0.105480
Epoch[80] Batch [699]	Speed: 2936.752426 samples/sec	accuracy=0.693248	lr=0.105287
Epoch[80] Batch [749]	Speed: 2884.721084 samples/sec	accuracy=0.692841	lr=0.105095
Epoch[80] Batch [799]	Speed: 2958.644663 samples/sec	accuracy=0.692612	lr=0.104903
Epoch[80] Batch [849]	Speed: 3117.798644 samples/sec	accuracy=0.692379	lr=0.104711
Epoch[80] Batch [899]	Speed: 3090.052452 samples/sec	accuracy=0.692142	lr=0.104519
Epoch[80] Batch [949]	Speed: 3102.670613 samples/sec	accuracy=0.691871	lr=0.104327
Epoch[80] Batch [999]	Speed: 3002.868818 samples/sec	accuracy=0.691578	lr=0.104135
Epoch[80] Batch [1049]	Speed: 3029.317995 samples/sec	accuracy=0.691230	lr=0.103944
Epoch[80] Batch [1099]	Speed: 2928.266574 samples/sec	accuracy=0.691121	lr=0.103752
Epoch[80] Batch [1149]	Speed: 2947.655955 samples/sec	accuracy=0.690878	lr=0.103561
Epoch[80] Batch [1199]	Speed: 2964.769132 samples/sec	accuracy=0.690755	lr=0.103370
Epoch[80] Batch [1249]	Speed: 3071.763490 samples/sec	accuracy=0.690586	lr=0.103178
[Epoch 80] training: accuracy=0.690587
[Epoch 80] speed: 3003 samples/sec	time cost: 437.091013
[Epoch 80] validation: err-top1=0.302914 err-top5=0.104153
Epoch[81] Batch [49]	Speed: 3018.232622 samples/sec	accuracy=0.700898	lr=0.102984
Epoch[81] Batch [99]	Speed: 2862.494384 samples/sec	accuracy=0.703340	lr=0.102793
Epoch[81] Batch [149]	Speed: 2990.710453 samples/sec	accuracy=0.701934	lr=0.102602
Epoch[81] Batch [199]	Speed: 3096.411270 samples/sec	accuracy=0.701455	lr=0.102411
Epoch[81] Batch [249]	Speed: 2980.757799 samples/sec	accuracy=0.700363	lr=0.102221
Epoch[81] Batch [299]	Speed: 3047.107402 samples/sec	accuracy=0.700189	lr=0.102030
Epoch[81] Batch [349]	Speed: 3118.378964 samples/sec	accuracy=0.699584	lr=0.101840
Epoch[81] Batch [399]	Speed: 2988.147820 samples/sec	accuracy=0.698804	lr=0.101650
Epoch[81] Batch [449]	Speed: 2973.912239 samples/sec	accuracy=0.698726	lr=0.101460
Epoch[81] Batch [499]	Speed: 2898.205813 samples/sec	accuracy=0.698691	lr=0.101270
Epoch[81] Batch [549]	Speed: 2942.135257 samples/sec	accuracy=0.698365	lr=0.101080
Epoch[81] Batch [599]	Speed: 2930.647294 samples/sec	accuracy=0.698358	lr=0.100890
Epoch[81] Batch [649]	Speed: 2973.611710 samples/sec	accuracy=0.697978	lr=0.100701
Epoch[81] Batch [699]	Speed: 3070.917423 samples/sec	accuracy=0.697443	lr=0.100511
Epoch[81] Batch [749]	Speed: 2950.215923 samples/sec	accuracy=0.696997	lr=0.100322
Epoch[81] Batch [799]	Speed: 3036.794876 samples/sec	accuracy=0.696680	lr=0.100132
Epoch[81] Batch [849]	Speed: 3125.085801 samples/sec	accuracy=0.696034	lr=0.099943
Epoch[81] Batch [899]	Speed: 3023.688365 samples/sec	accuracy=0.695760	lr=0.099754
Epoch[81] Batch [949]	Speed: 3065.734062 samples/sec	accuracy=0.695642	lr=0.099565
Epoch[81] Batch [999]	Speed: 2977.967777 samples/sec	accuracy=0.695384	lr=0.099377
Epoch[81] Batch [1049]	Speed: 3080.695069 samples/sec	accuracy=0.695206	lr=0.099188
Epoch[81] Batch [1099]	Speed: 2928.327507 samples/sec	accuracy=0.695026	lr=0.098999
Epoch[81] Batch [1149]	Speed: 2935.973589 samples/sec	accuracy=0.694963	lr=0.098811
Epoch[81] Batch [1199]	Speed: 3021.086356 samples/sec	accuracy=0.694764	lr=0.098623
Epoch[81] Batch [1249]	Speed: 3129.065436 samples/sec	accuracy=0.694510	lr=0.098434
[Epoch 81] training: accuracy=0.694514
[Epoch 81] speed: 3002 samples/sec	time cost: 437.390793
[Epoch 81] validation: err-top1=0.288635 err-top5=0.094442
Epoch[82] Batch [49]	Speed: 3085.656358 samples/sec	accuracy=0.705547	lr=0.098243
Epoch[82] Batch [99]	Speed: 3018.697962 samples/sec	accuracy=0.703887	lr=0.098055
Epoch[82] Batch [149]	Speed: 2951.393914 samples/sec	accuracy=0.703516	lr=0.097867
Epoch[82] Batch [199]	Speed: 3027.273422 samples/sec	accuracy=0.704248	lr=0.097679
Epoch[82] Batch [249]	Speed: 3039.706940 samples/sec	accuracy=0.703914	lr=0.097492
Epoch[82] Batch [299]	Speed: 3117.401898 samples/sec	accuracy=0.703415	lr=0.097304
Epoch[82] Batch [349]	Speed: 2967.225960 samples/sec	accuracy=0.702997	lr=0.097117
Epoch[82] Batch [399]	Speed: 2918.423705 samples/sec	accuracy=0.701965	lr=0.096930
Epoch[82] Batch [449]	Speed: 3066.133743 samples/sec	accuracy=0.701812	lr=0.096743
Epoch[82] Batch [499]	Speed: 2977.355934 samples/sec	accuracy=0.702031	lr=0.096556
Epoch[82] Batch [549]	Speed: 3056.205924 samples/sec	accuracy=0.701644	lr=0.096369
Epoch[82] Batch [599]	Speed: 2951.003390 samples/sec	accuracy=0.701304	lr=0.096182
Epoch[82] Batch [649]	Speed: 3096.553074 samples/sec	accuracy=0.700933	lr=0.095996
Epoch[82] Batch [699]	Speed: 3147.166753 samples/sec	accuracy=0.700557	lr=0.095809
Epoch[82] Batch [749]	Speed: 2968.281645 samples/sec	accuracy=0.700240	lr=0.095623
Epoch[82] Batch [799]	Speed: 2954.012353 samples/sec	accuracy=0.700082	lr=0.095437
Epoch[82] Batch [849]	Speed: 2996.734836 samples/sec	accuracy=0.699832	lr=0.095250
Epoch[82] Batch [899]	Speed: 2973.851453 samples/sec	accuracy=0.699442	lr=0.095064
Epoch[82] Batch [949]	Speed: 2922.235527 samples/sec	accuracy=0.699101	lr=0.094879
Epoch[82] Batch [999]	Speed: 3129.250966 samples/sec	accuracy=0.699115	lr=0.094693
Epoch[82] Batch [1049]	Speed: 2998.201370 samples/sec	accuracy=0.698689	lr=0.094507
Epoch[82] Batch [1099]	Speed: 3036.665148 samples/sec	accuracy=0.698324	lr=0.094322
Epoch[82] Batch [1149]	Speed: 3062.103013 samples/sec	accuracy=0.698080	lr=0.094137
Epoch[82] Batch [1199]	Speed: 2959.553324 samples/sec	accuracy=0.697795	lr=0.093951
Epoch[82] Batch [1249]	Speed: 3146.326032 samples/sec	accuracy=0.697631	lr=0.093766
[Epoch 82] training: accuracy=0.697636
[Epoch 82] speed: 3018 samples/sec	time cost: 435.194194
[Epoch 82] validation: err-top1=0.307199 err-top5=0.107761
Epoch[83] Batch [49]	Speed: 3231.986760 samples/sec	accuracy=0.708145	lr=0.093578
Epoch[83] Batch [99]	Speed: 3060.858916 samples/sec	accuracy=0.709014	lr=0.093393
Epoch[83] Batch [149]	Speed: 3043.677678 samples/sec	accuracy=0.708522	lr=0.093208
Epoch[83] Batch [199]	Speed: 2912.032395 samples/sec	accuracy=0.706758	lr=0.093024
Epoch[83] Batch [249]	Speed: 2913.447337 samples/sec	accuracy=0.707590	lr=0.092839
Epoch[83] Batch [299]	Speed: 3059.383068 samples/sec	accuracy=0.706722	lr=0.092655
Epoch[83] Batch [349]	Speed: 2999.140651 samples/sec	accuracy=0.706052	lr=0.092471
Epoch[83] Batch [399]	Speed: 2908.816868 samples/sec	accuracy=0.705532	lr=0.092287
Epoch[83] Batch [449]	Speed: 3016.316519 samples/sec	accuracy=0.705050	lr=0.092103
Epoch[83] Batch [499]	Speed: 3012.986919 samples/sec	accuracy=0.704934	lr=0.091919
Epoch[83] Batch [549]	Speed: 2934.712419 samples/sec	accuracy=0.704521	lr=0.091735
Epoch[83] Batch [599]	Speed: 2958.823008 samples/sec	accuracy=0.703989	lr=0.091552
Epoch[83] Batch [649]	Speed: 2935.618756 samples/sec	accuracy=0.703529	lr=0.091368
Epoch[83] Batch [699]	Speed: 2889.666950 samples/sec	accuracy=0.703256	lr=0.091185
Epoch[83] Batch [749]	Speed: 3004.192035 samples/sec	accuracy=0.702991	lr=0.091002
Epoch[83] Batch [799]	Speed: 2936.824557 samples/sec	accuracy=0.702882	lr=0.090819
Epoch[83] Batch [849]	Speed: 2903.324373 samples/sec	accuracy=0.702631	lr=0.090636
Epoch[83] Batch [899]	Speed: 2959.422281 samples/sec	accuracy=0.702374	lr=0.090453
Epoch[83] Batch [949]	Speed: 3048.762121 samples/sec	accuracy=0.702177	lr=0.090270
Epoch[83] Batch [999]	Speed: 2996.858833 samples/sec	accuracy=0.701908	lr=0.090088
Epoch[83] Batch [1049]	Speed: 3041.076782 samples/sec	accuracy=0.701827	lr=0.089905
Epoch[83] Batch [1099]	Speed: 2924.603778 samples/sec	accuracy=0.701594	lr=0.089723
Epoch[83] Batch [1149]	Speed: 2933.302314 samples/sec	accuracy=0.701432	lr=0.089541
Epoch[83] Batch [1199]	Speed: 3002.802392 samples/sec	accuracy=0.701255	lr=0.089359
Epoch[83] Batch [1249]	Speed: 3235.305879 samples/sec	accuracy=0.701113	lr=0.089177
[Epoch 83] training: accuracy=0.701118
[Epoch 83] speed: 2989 samples/sec	time cost: 438.828693
[Epoch 83] validation: err-top1=0.295061 err-top5=0.096799
Epoch[84] Batch [49]	Speed: 3187.690301 samples/sec	accuracy=0.710605	lr=0.088992
Epoch[84] Batch [99]	Speed: 3064.275415 samples/sec	accuracy=0.711816	lr=0.088810
Epoch[84] Batch [149]	Speed: 2820.890639 samples/sec	accuracy=0.711413	lr=0.088629
Epoch[84] Batch [199]	Speed: 3104.569803 samples/sec	accuracy=0.710459	lr=0.088448
Epoch[84] Batch [249]	Speed: 2945.183953 samples/sec	accuracy=0.709590	lr=0.088266
Epoch[84] Batch [299]	Speed: 2951.469160 samples/sec	accuracy=0.709131	lr=0.088085
Epoch[84] Batch [349]	Speed: 3008.359799 samples/sec	accuracy=0.708544	lr=0.087904
Epoch[84] Batch [399]	Speed: 2956.239704 samples/sec	accuracy=0.708806	lr=0.087724
Epoch[84] Batch [449]	Speed: 3013.619162 samples/sec	accuracy=0.708544	lr=0.087543
Epoch[84] Batch [499]	Speed: 3007.976764 samples/sec	accuracy=0.708166	lr=0.087363
Epoch[84] Batch [549]	Speed: 3112.521003 samples/sec	accuracy=0.708033	lr=0.087182
Epoch[84] Batch [599]	Speed: 3126.487674 samples/sec	accuracy=0.707733	lr=0.087002
Epoch[84] Batch [649]	Speed: 2934.646688 samples/sec	accuracy=0.707554	lr=0.086822
Epoch[84] Batch [699]	Speed: 2887.952077 samples/sec	accuracy=0.707536	lr=0.086642
Epoch[84] Batch [749]	Speed: 3129.889658 samples/sec	accuracy=0.707089	lr=0.086462
Epoch[84] Batch [799]	Speed: 3047.815902 samples/sec	accuracy=0.706876	lr=0.086282
Epoch[84] Batch [849]	Speed: 3080.607036 samples/sec	accuracy=0.706846	lr=0.086103
Epoch[84] Batch [899]	Speed: 2900.074342 samples/sec	accuracy=0.706637	lr=0.085923
Epoch[84] Batch [949]	Speed: 2936.539708 samples/sec	accuracy=0.706439	lr=0.085744
Epoch[84] Batch [999]	Speed: 2889.379279 samples/sec	accuracy=0.706333	lr=0.085565
Epoch[84] Batch [1049]	Speed: 2994.561818 samples/sec	accuracy=0.706143	lr=0.085386
Epoch[84] Batch [1099]	Speed: 2890.735242 samples/sec	accuracy=0.706021	lr=0.085207
Epoch[84] Batch [1149]	Speed: 2942.728676 samples/sec	accuracy=0.705782	lr=0.085028
Epoch[84] Batch [1199]	Speed: 3006.212062 samples/sec	accuracy=0.705542	lr=0.084850
Epoch[84] Batch [1249]	Speed: 3234.715189 samples/sec	accuracy=0.705305	lr=0.084671
[Epoch 84] training: accuracy=0.705297
[Epoch 84] speed: 3001 samples/sec	time cost: 437.014312
[Epoch 84] validation: err-top1=0.295540 err-top5=0.099669
Epoch[85] Batch [49]	Speed: 3119.946245 samples/sec	accuracy=0.719023	lr=0.084489
Epoch[85] Batch [99]	Speed: 3103.513144 samples/sec	accuracy=0.717773	lr=0.084311
Epoch[85] Batch [149]	Speed: 2944.954424 samples/sec	accuracy=0.716426	lr=0.084133
Epoch[85] Batch [199]	Speed: 3050.038594 samples/sec	accuracy=0.715581	lr=0.083955
Epoch[85] Batch [249]	Speed: 2978.605816 samples/sec	accuracy=0.714871	lr=0.083777
Epoch[85] Batch [299]	Speed: 3137.890964 samples/sec	accuracy=0.714899	lr=0.083600
Epoch[85] Batch [349]	Speed: 2969.002189 samples/sec	accuracy=0.714389	lr=0.083422
Epoch[85] Batch [399]	Speed: 2954.103377 samples/sec	accuracy=0.713352	lr=0.083245
Epoch[85] Batch [449]	Speed: 2911.004734 samples/sec	accuracy=0.712776	lr=0.083067
Epoch[85] Batch [499]	Speed: 2910.140897 samples/sec	accuracy=0.712029	lr=0.082890
Epoch[85] Batch [549]	Speed: 3118.877012 samples/sec	accuracy=0.711532	lr=0.082713
Epoch[85] Batch [599]	Speed: 2983.440975 samples/sec	accuracy=0.711226	lr=0.082537
Epoch[85] Batch [649]	Speed: 3037.887675 samples/sec	accuracy=0.710916	lr=0.082360
Epoch[85] Batch [699]	Speed: 3006.089225 samples/sec	accuracy=0.710843	lr=0.082183
Epoch[85] Batch [749]	Speed: 2924.629189 samples/sec	accuracy=0.710884	lr=0.082007
Epoch[85] Batch [799]	Speed: 2983.365334 samples/sec	accuracy=0.710756	lr=0.081831
Epoch[85] Batch [849]	Speed: 3033.584184 samples/sec	accuracy=0.710429	lr=0.081655
Epoch[85] Batch [899]	Speed: 2968.634487 samples/sec	accuracy=0.710101	lr=0.081479
Epoch[85] Batch [949]	Speed: 2952.585147 samples/sec	accuracy=0.709886	lr=0.081303
Epoch[85] Batch [999]	Speed: 2869.898215 samples/sec	accuracy=0.709588	lr=0.081127
Epoch[85] Batch [1049]	Speed: 3068.152652 samples/sec	accuracy=0.709469	lr=0.080952
Epoch[85] Batch [1099]	Speed: 3200.244588 samples/sec	accuracy=0.709240	lr=0.080776
Epoch[85] Batch [1149]	Speed: 3227.823434 samples/sec	accuracy=0.709243	lr=0.080601
Epoch[85] Batch [1199]	Speed: 3032.653786 samples/sec	accuracy=0.709214	lr=0.080426
Epoch[85] Batch [1249]	Speed: 3036.129821 samples/sec	accuracy=0.708823	lr=0.080251
[Epoch 85] training: accuracy=0.708830
[Epoch 85] speed: 3016 samples/sec	time cost: 434.579837
[Epoch 85] validation: err-top1=0.281429 err-top5=0.089346
Epoch[86] Batch [49]	Speed: 3108.731724 samples/sec	accuracy=0.722031	lr=0.080073
Epoch[86] Batch [99]	Speed: 2890.394488 samples/sec	accuracy=0.719307	lr=0.079898
Epoch[86] Batch [149]	Speed: 3032.443135 samples/sec	accuracy=0.717982	lr=0.079723
Epoch[86] Batch [199]	Speed: 3081.123726 samples/sec	accuracy=0.718354	lr=0.079549
Epoch[86] Batch [249]	Speed: 3001.037589 samples/sec	accuracy=0.717273	lr=0.079375
Epoch[86] Batch [299]	Speed: 3009.351043 samples/sec	accuracy=0.716117	lr=0.079201
Epoch[86] Batch [349]	Speed: 2962.907902 samples/sec	accuracy=0.716018	lr=0.079027
Epoch[86] Batch [399]	Speed: 3001.194363 samples/sec	accuracy=0.716011	lr=0.078853
Epoch[86] Batch [449]	Speed: 3018.163945 samples/sec	accuracy=0.716126	lr=0.078679
Epoch[86] Batch [499]	Speed: 3078.105112 samples/sec	accuracy=0.715818	lr=0.078506
Epoch[86] Batch [549]	Speed: 2972.332156 samples/sec	accuracy=0.715401	lr=0.078332
Epoch[86] Batch [599]	Speed: 2953.321607 samples/sec	accuracy=0.715509	lr=0.078159
Epoch[86] Batch [649]	Speed: 2988.204494 samples/sec	accuracy=0.715227	lr=0.077986
Epoch[86] Batch [699]	Speed: 2935.138117 samples/sec	accuracy=0.714990	lr=0.077813
Epoch[86] Batch [749]	Speed: 3031.421859 samples/sec	accuracy=0.714809	lr=0.077640
Epoch[86] Batch [799]	Speed: 2944.255877 samples/sec	accuracy=0.714434	lr=0.077467
Epoch[86] Batch [849]	Speed: 3239.801326 samples/sec	accuracy=0.714221	lr=0.077295
Epoch[86] Batch [899]	Speed: 2963.652261 samples/sec	accuracy=0.713972	lr=0.077123
Epoch[86] Batch [949]	Speed: 3165.440549 samples/sec	accuracy=0.713646	lr=0.076950
Epoch[86] Batch [999]	Speed: 3008.446364 samples/sec	accuracy=0.713543	lr=0.076778
Epoch[86] Batch [1049]	Speed: 3085.397718 samples/sec	accuracy=0.713432	lr=0.076606
Epoch[86] Batch [1099]	Speed: 2941.322300 samples/sec	accuracy=0.713299	lr=0.076435
Epoch[86] Batch [1149]	Speed: 2926.697676 samples/sec	accuracy=0.713122	lr=0.076263
Epoch[86] Batch [1199]	Speed: 2922.970011 samples/sec	accuracy=0.712809	lr=0.076091
Epoch[86] Batch [1249]	Speed: 3080.248416 samples/sec	accuracy=0.712766	lr=0.075920
[Epoch 86] training: accuracy=0.712784
[Epoch 86] speed: 3009 samples/sec	time cost: 435.761858
[Epoch 86] validation: err-top1=0.277702 err-top5=0.088429
Epoch[87] Batch [49]	Speed: 3209.385040 samples/sec	accuracy=0.725371	lr=0.075745
Epoch[87] Batch [99]	Speed: 2945.947196 samples/sec	accuracy=0.723975	lr=0.075574
Epoch[87] Batch [149]	Speed: 2908.712105 samples/sec	accuracy=0.722135	lr=0.075404
Epoch[87] Batch [199]	Speed: 2941.501222 samples/sec	accuracy=0.722080	lr=0.075233
Epoch[87] Batch [249]	Speed: 3046.005757 samples/sec	accuracy=0.721574	lr=0.075062
Epoch[87] Batch [299]	Speed: 3062.082492 samples/sec	accuracy=0.720791	lr=0.074892
Epoch[87] Batch [349]	Speed: 3043.405411 samples/sec	accuracy=0.720307	lr=0.074721
Epoch[87] Batch [399]	Speed: 3016.885440 samples/sec	accuracy=0.719832	lr=0.074551
Epoch[87] Batch [449]	Speed: 3070.565094 samples/sec	accuracy=0.719017	lr=0.074381
Epoch[87] Batch [499]	Speed: 2929.861618 samples/sec	accuracy=0.718906	lr=0.074211
Epoch[87] Batch [549]	Speed: 2920.145306 samples/sec	accuracy=0.719055	lr=0.074042
Epoch[87] Batch [599]	Speed: 3066.832024 samples/sec	accuracy=0.719082	lr=0.073872
Epoch[87] Batch [649]	Speed: 2981.481641 samples/sec	accuracy=0.718873	lr=0.073703
Epoch[87] Batch [699]	Speed: 2970.947904 samples/sec	accuracy=0.718471	lr=0.073534
Epoch[87] Batch [749]	Speed: 2888.098385 samples/sec	accuracy=0.718020	lr=0.073364
Epoch[87] Batch [799]	Speed: 2958.974710 samples/sec	accuracy=0.717671	lr=0.073196
Epoch[87] Batch [849]	Speed: 3017.133399 samples/sec	accuracy=0.717246	lr=0.073027
Epoch[87] Batch [899]	Speed: 3041.310514 samples/sec	accuracy=0.717318	lr=0.072858
Epoch[87] Batch [949]	Speed: 3037.832883 samples/sec	accuracy=0.717138	lr=0.072690
Epoch[87] Batch [999]	Speed: 3215.270903 samples/sec	accuracy=0.717034	lr=0.072521
Epoch[87] Batch [1049]	Speed: 3041.537605 samples/sec	accuracy=0.716815	lr=0.072353
Epoch[87] Batch [1099]	Speed: 2909.861517 samples/sec	accuracy=0.716696	lr=0.072185
Epoch[87] Batch [1149]	Speed: 2915.181986 samples/sec	accuracy=0.716565	lr=0.072017
Epoch[87] Batch [1199]	Speed: 2927.520646 samples/sec	accuracy=0.716391	lr=0.071849
Epoch[87] Batch [1249]	Speed: 3101.507380 samples/sec	accuracy=0.716175	lr=0.071682
[Epoch 87] training: accuracy=0.716163
[Epoch 87] speed: 3002 samples/sec	time cost: 436.787822
[Epoch 87] validation: err-top1=0.278280 err-top5=0.088582
Epoch[88] Batch [49]	Speed: 3146.693980 samples/sec	accuracy=0.727793	lr=0.071511
Epoch[88] Batch [99]	Speed: 2973.419680 samples/sec	accuracy=0.726816	lr=0.071344
Epoch[88] Batch [149]	Speed: 3038.005387 samples/sec	accuracy=0.727135	lr=0.071177
Epoch[88] Batch [199]	Speed: 3134.774702 samples/sec	accuracy=0.725928	lr=0.071010
Epoch[88] Batch [249]	Speed: 3195.402419 samples/sec	accuracy=0.724914	lr=0.070843
Epoch[88] Batch [299]	Speed: 3026.728859 samples/sec	accuracy=0.724626	lr=0.070676
Epoch[88] Batch [349]	Speed: 2979.371063 samples/sec	accuracy=0.724353	lr=0.070510
Epoch[88] Batch [399]	Speed: 2901.313241 samples/sec	accuracy=0.724075	lr=0.070343
Epoch[88] Batch [449]	Speed: 3063.980521 samples/sec	accuracy=0.723513	lr=0.070177
Epoch[88] Batch [499]	Speed: 2968.432061 samples/sec	accuracy=0.723174	lr=0.070011
Epoch[88] Batch [549]	Speed: 3087.389573 samples/sec	accuracy=0.722743	lr=0.069845
Epoch[88] Batch [599]	Speed: 3106.492979 samples/sec	accuracy=0.722627	lr=0.069680
Epoch[88] Batch [649]	Speed: 3042.919186 samples/sec	accuracy=0.722246	lr=0.069514
Epoch[88] Batch [699]	Speed: 3010.839832 samples/sec	accuracy=0.721924	lr=0.069349
Epoch[88] Batch [749]	Speed: 3129.192054 samples/sec	accuracy=0.721668	lr=0.069183
Epoch[88] Batch [799]	Speed: 3043.001629 samples/sec	accuracy=0.721624	lr=0.069018
Epoch[88] Batch [849]	Speed: 2959.300466 samples/sec	accuracy=0.721414	lr=0.068853
Epoch[88] Batch [899]	Speed: 3119.841133 samples/sec	accuracy=0.721329	lr=0.068689
Epoch[88] Batch [949]	Speed: 3078.192075 samples/sec	accuracy=0.721040	lr=0.068524
Epoch[88] Batch [999]	Speed: 3086.144584 samples/sec	accuracy=0.720716	lr=0.068359
Epoch[88] Batch [1049]	Speed: 2959.886429 samples/sec	accuracy=0.720476	lr=0.068195
Epoch[88] Batch [1099]	Speed: 3049.002144 samples/sec	accuracy=0.720360	lr=0.068031
Epoch[88] Batch [1149]	Speed: 2934.932493 samples/sec	accuracy=0.720240	lr=0.067867
Epoch[88] Batch [1199]	Speed: 3189.610231 samples/sec	accuracy=0.720144	lr=0.067703
Epoch[88] Batch [1249]	Speed: 3189.201772 samples/sec	accuracy=0.719929	lr=0.067539
[Epoch 88] training: accuracy=0.719932
[Epoch 88] speed: 3052 samples/sec	time cost: 430.317329
[Epoch 88] validation: err-top1=0.278061 err-top5=0.088249
Epoch[89] Batch [49]	Speed: 3091.661616 samples/sec	accuracy=0.731348	lr=0.067373
Epoch[89] Batch [99]	Speed: 2966.687989 samples/sec	accuracy=0.729727	lr=0.067209
Epoch[89] Batch [149]	Speed: 3046.192932 samples/sec	accuracy=0.729642	lr=0.067046
Epoch[89] Batch [199]	Speed: 3016.722954 samples/sec	accuracy=0.728677	lr=0.066883
Epoch[89] Batch [249]	Speed: 3063.010504 samples/sec	accuracy=0.728266	lr=0.066720
Epoch[89] Batch [299]	Speed: 3013.019892 samples/sec	accuracy=0.727526	lr=0.066557
Epoch[89] Batch [349]	Speed: 2994.292630 samples/sec	accuracy=0.727174	lr=0.066395
Epoch[89] Batch [399]	Speed: 3022.890739 samples/sec	accuracy=0.726753	lr=0.066232
Epoch[89] Batch [449]	Speed: 3068.890095 samples/sec	accuracy=0.726439	lr=0.066070
Epoch[89] Batch [499]	Speed: 3041.378784 samples/sec	accuracy=0.726092	lr=0.065908
Epoch[89] Batch [549]	Speed: 2938.027972 samples/sec	accuracy=0.725778	lr=0.065746
Epoch[89] Batch [599]	Speed: 3073.606411 samples/sec	accuracy=0.725409	lr=0.065584
Epoch[89] Batch [649]	Speed: 2933.598637 samples/sec	accuracy=0.725322	lr=0.065423
Epoch[89] Batch [699]	Speed: 2905.536767 samples/sec	accuracy=0.725028	lr=0.065261
Epoch[89] Batch [749]	Speed: 2967.988119 samples/sec	accuracy=0.724805	lr=0.065100
Epoch[89] Batch [799]	Speed: 3001.438868 samples/sec	accuracy=0.724585	lr=0.064939
Epoch[89] Batch [849]	Speed: 3032.862195 samples/sec	accuracy=0.724284	lr=0.064778
Epoch[89] Batch [899]	Speed: 2900.068233 samples/sec	accuracy=0.724022	lr=0.064617
Epoch[89] Batch [949]	Speed: 3072.355720 samples/sec	accuracy=0.723986	lr=0.064456
Epoch[89] Batch [999]	Speed: 2982.902495 samples/sec	accuracy=0.723743	lr=0.064296
Epoch[89] Batch [1049]	Speed: 3021.836297 samples/sec	accuracy=0.723595	lr=0.064135
Epoch[89] Batch [1099]	Speed: 2984.046572 samples/sec	accuracy=0.723385	lr=0.063975
Epoch[89] Batch [1149]	Speed: 2963.094325 samples/sec	accuracy=0.723205	lr=0.063815
Epoch[89] Batch [1199]	Speed: 2937.844408 samples/sec	accuracy=0.723018	lr=0.063655
Epoch[89] Batch [1249]	Speed: 3094.733025 samples/sec	accuracy=0.722980	lr=0.063496
[Epoch 89] training: accuracy=0.722983
[Epoch 89] speed: 3002 samples/sec	time cost: 436.671220
[Epoch 89] validation: err-top1=0.264330 err-top5=0.079998
Epoch[90] Batch [49]	Speed: 3041.115239 samples/sec	accuracy=0.733555	lr=0.063333
Epoch[90] Batch [99]	Speed: 3087.093010 samples/sec	accuracy=0.732744	lr=0.063174
Epoch[90] Batch [149]	Speed: 2919.844549 samples/sec	accuracy=0.733190	lr=0.063015
Epoch[90] Batch [199]	Speed: 2980.611716 samples/sec	accuracy=0.732300	lr=0.062855
Epoch[90] Batch [249]	Speed: 3087.538897 samples/sec	accuracy=0.731477	lr=0.062697
Epoch[90] Batch [299]	Speed: 3086.352427 samples/sec	accuracy=0.731387	lr=0.062538
Epoch[90] Batch [349]	Speed: 3126.434009 samples/sec	accuracy=0.731309	lr=0.062379
Epoch[90] Batch [399]	Speed: 2963.288090 samples/sec	accuracy=0.730991	lr=0.062221
Epoch[90] Batch [449]	Speed: 2923.613276 samples/sec	accuracy=0.730749	lr=0.062063
Epoch[90] Batch [499]	Speed: 3075.411069 samples/sec	accuracy=0.730168	lr=0.061905
Epoch[90] Batch [549]	Speed: 3006.746277 samples/sec	accuracy=0.729918	lr=0.061747
Epoch[90] Batch [599]	Speed: 2933.268378 samples/sec	accuracy=0.729702	lr=0.061589
Epoch[90] Batch [649]	Speed: 2878.518745 samples/sec	accuracy=0.729776	lr=0.061432
Epoch[90] Batch [699]	Speed: 3258.388089 samples/sec	accuracy=0.729562	lr=0.061274
Epoch[90] Batch [749]	Speed: 3034.527588 samples/sec	accuracy=0.729342	lr=0.061117
Epoch[90] Batch [799]	Speed: 2991.383882 samples/sec	accuracy=0.729066	lr=0.060960
Epoch[90] Batch [849]	Speed: 2948.273460 samples/sec	accuracy=0.728881	lr=0.060803
Epoch[90] Batch [899]	Speed: 2844.671385 samples/sec	accuracy=0.728857	lr=0.060646
Epoch[90] Batch [949]	Speed: 2975.331409 samples/sec	accuracy=0.728704	lr=0.060490
Epoch[90] Batch [999]	Speed: 2922.000813 samples/sec	accuracy=0.728597	lr=0.060333
Epoch[90] Batch [1049]	Speed: 3212.566129 samples/sec	accuracy=0.728570	lr=0.060177
Epoch[90] Batch [1099]	Speed: 2946.726967 samples/sec	accuracy=0.728271	lr=0.060021
Epoch[90] Batch [1149]	Speed: 2896.248206 samples/sec	accuracy=0.728267	lr=0.059865
Epoch[90] Batch [1199]	Speed: 2955.656770 samples/sec	accuracy=0.728077	lr=0.059710
Epoch[90] Batch [1249]	Speed: 3189.456839 samples/sec	accuracy=0.728080	lr=0.059554
[Epoch 90] training: accuracy=0.728085
[Epoch 90] speed: 3005 samples/sec	time cost: 437.370902
[Epoch 90] validation: err-top1=0.271445 err-top5=0.086017
Epoch[91] Batch [49]	Speed: 3231.367280 samples/sec	accuracy=0.737871	lr=0.059396
Epoch[91] Batch [99]	Speed: 2951.736139 samples/sec	accuracy=0.736553	lr=0.059240
Epoch[91] Batch [149]	Speed: 2979.172379 samples/sec	accuracy=0.737031	lr=0.059085
Epoch[91] Batch [199]	Speed: 3004.424755 samples/sec	accuracy=0.737764	lr=0.058930
Epoch[91] Batch [249]	Speed: 2940.100125 samples/sec	accuracy=0.737285	lr=0.058776
Epoch[91] Batch [299]	Speed: 2967.806617 samples/sec	accuracy=0.737038	lr=0.058621
Epoch[91] Batch [349]	Speed: 3003.117670 samples/sec	accuracy=0.736554	lr=0.058467
Epoch[91] Batch [399]	Speed: 3157.069943 samples/sec	accuracy=0.736187	lr=0.058313
Epoch[91] Batch [449]	Speed: 2966.392852 samples/sec	accuracy=0.735935	lr=0.058158
Epoch[91] Batch [499]	Speed: 2971.528127 samples/sec	accuracy=0.735480	lr=0.058005
Epoch[91] Batch [549]	Speed: 3229.055253 samples/sec	accuracy=0.735265	lr=0.057851
Epoch[91] Batch [599]	Speed: 2997.306225 samples/sec	accuracy=0.735251	lr=0.057697
Epoch[91] Batch [649]	Speed: 3014.074407 samples/sec	accuracy=0.734863	lr=0.057544
Epoch[91] Batch [699]	Speed: 3097.459746 samples/sec	accuracy=0.734628	lr=0.057391
Epoch[91] Batch [749]	Speed: 2903.017848 samples/sec	accuracy=0.734217	lr=0.057238
Epoch[91] Batch [799]	Speed: 3000.354566 samples/sec	accuracy=0.733916	lr=0.057085
Epoch[91] Batch [849]	Speed: 2968.370801 samples/sec	accuracy=0.733790	lr=0.056932
Epoch[91] Batch [899]	Speed: 3190.079261 samples/sec	accuracy=0.733506	lr=0.056780
Epoch[91] Batch [949]	Speed: 3183.734499 samples/sec	accuracy=0.733293	lr=0.056627
Epoch[91] Batch [999]	Speed: 2930.345089 samples/sec	accuracy=0.733069	lr=0.056475
Epoch[91] Batch [1049]	Speed: 3107.228243 samples/sec	accuracy=0.732892	lr=0.056323
Epoch[91] Batch [1099]	Speed: 3021.726595 samples/sec	accuracy=0.732712	lr=0.056171
Epoch[91] Batch [1149]	Speed: 2898.259986 samples/sec	accuracy=0.732517	lr=0.056020
Epoch[91] Batch [1199]	Speed: 3004.113405 samples/sec	accuracy=0.732419	lr=0.055868
Epoch[91] Batch [1249]	Speed: 3123.340723 samples/sec	accuracy=0.732310	lr=0.055717
[Epoch 91] training: accuracy=0.732314
[Epoch 91] speed: 3028 samples/sec	time cost: 433.216480
[Epoch 91] validation: err-top1=0.284778 err-top5=0.093909
Epoch[92] Batch [49]	Speed: 3072.634687 samples/sec	accuracy=0.746113	lr=0.055563
Epoch[92] Batch [99]	Speed: 3025.837493 samples/sec	accuracy=0.745654	lr=0.055412
Epoch[92] Batch [149]	Speed: 3353.785539 samples/sec	accuracy=0.743451	lr=0.055261
Epoch[92] Batch [199]	Speed: 3111.289519 samples/sec	accuracy=0.742954	lr=0.055111
Epoch[92] Batch [249]	Speed: 3034.112741 samples/sec	accuracy=0.742273	lr=0.054960
Epoch[92] Batch [299]	Speed: 3053.952174 samples/sec	accuracy=0.741442	lr=0.054810
Epoch[92] Batch [349]	Speed: 3083.312053 samples/sec	accuracy=0.741392	lr=0.054660
Epoch[92] Batch [399]	Speed: 2892.923106 samples/sec	accuracy=0.741309	lr=0.054510
Epoch[92] Batch [449]	Speed: 2994.685133 samples/sec	accuracy=0.741020	lr=0.054360
Epoch[92] Batch [499]	Speed: 3024.501065 samples/sec	accuracy=0.740531	lr=0.054210
Epoch[92] Batch [549]	Speed: 2969.759224 samples/sec	accuracy=0.740417	lr=0.054061
Epoch[92] Batch [599]	Speed: 3007.280135 samples/sec	accuracy=0.740073	lr=0.053912
Epoch[92] Batch [649]	Speed: 3021.861088 samples/sec	accuracy=0.739489	lr=0.053763
Epoch[92] Batch [699]	Speed: 2851.299122 samples/sec	accuracy=0.739291	lr=0.053614
Epoch[92] Batch [749]	Speed: 3126.349306 samples/sec	accuracy=0.738944	lr=0.053465
Epoch[92] Batch [799]	Speed: 2973.923441 samples/sec	accuracy=0.738680	lr=0.053317
Epoch[92] Batch [849]	Speed: 3067.810555 samples/sec	accuracy=0.738570	lr=0.053168
Epoch[92] Batch [899]	Speed: 2954.705821 samples/sec	accuracy=0.738296	lr=0.053020
Epoch[92] Batch [949]	Speed: 2999.648054 samples/sec	accuracy=0.738048	lr=0.052872
Epoch[92] Batch [999]	Speed: 3055.508082 samples/sec	accuracy=0.737814	lr=0.052724
Epoch[92] Batch [1049]	Speed: 3139.933290 samples/sec	accuracy=0.737597	lr=0.052577
Epoch[92] Batch [1099]	Speed: 3163.022272 samples/sec	accuracy=0.737504	lr=0.052429
Epoch[92] Batch [1149]	Speed: 2893.586041 samples/sec	accuracy=0.737167	lr=0.052282
Epoch[92] Batch [1199]	Speed: 3045.556581 samples/sec	accuracy=0.736969	lr=0.052135
Epoch[92] Batch [1249]	Speed: 3134.569026 samples/sec	accuracy=0.736766	lr=0.051988
[Epoch 92] training: accuracy=0.736758
[Epoch 92] speed: 3036 samples/sec	time cost: 432.126338
[Epoch 92] validation: err-top1=0.278221 err-top5=0.089385
Epoch[93] Batch [49]	Speed: 3165.895170 samples/sec	accuracy=0.747441	lr=0.051838
Epoch[93] Batch [99]	Speed: 3015.975506 samples/sec	accuracy=0.748047	lr=0.051691
Epoch[93] Batch [149]	Speed: 2928.773085 samples/sec	accuracy=0.748027	lr=0.051545
Epoch[93] Batch [199]	Speed: 3068.311344 samples/sec	accuracy=0.747056	lr=0.051399
Epoch[93] Batch [249]	Speed: 2978.864009 samples/sec	accuracy=0.746434	lr=0.051253
Epoch[93] Batch [299]	Speed: 2944.334755 samples/sec	accuracy=0.745527	lr=0.051107
Epoch[93] Batch [349]	Speed: 3015.566985 samples/sec	accuracy=0.744662	lr=0.050961
Epoch[93] Batch [399]	Speed: 3091.484478 samples/sec	accuracy=0.744343	lr=0.050816
Epoch[93] Batch [449]	Speed: 3058.497849 samples/sec	accuracy=0.744219	lr=0.050670
Epoch[93] Batch [499]	Speed: 2975.209847 samples/sec	accuracy=0.743527	lr=0.050525
Epoch[93] Batch [549]	Speed: 2984.919541 samples/sec	accuracy=0.743306	lr=0.050380
Epoch[93] Batch [599]	Speed: 2927.972365 samples/sec	accuracy=0.742788	lr=0.050235
Epoch[93] Batch [649]	Speed: 3040.538048 samples/sec	accuracy=0.742273	lr=0.050091
Epoch[93] Batch [699]	Speed: 2957.881673 samples/sec	accuracy=0.742139	lr=0.049946
Epoch[93] Batch [749]	Speed: 3003.218045 samples/sec	accuracy=0.741995	lr=0.049802
Epoch[93] Batch [799]	Speed: 2978.138175 samples/sec	accuracy=0.741910	lr=0.049658
Epoch[93] Batch [849]	Speed: 2955.850296 samples/sec	accuracy=0.741695	lr=0.049514
Epoch[93] Batch [899]	Speed: 3145.133112 samples/sec	accuracy=0.741484	lr=0.049370
Epoch[93] Batch [949]	Speed: 3042.079020 samples/sec	accuracy=0.741287	lr=0.049227
Epoch[93] Batch [999]	Speed: 3013.336939 samples/sec	accuracy=0.741104	lr=0.049083
Epoch[93] Batch [1049]	Speed: 3006.124699 samples/sec	accuracy=0.740865	lr=0.048940
Epoch[93] Batch [1099]	Speed: 2924.215731 samples/sec	accuracy=0.740743	lr=0.048797
Epoch[93] Batch [1149]	Speed: 2903.209448 samples/sec	accuracy=0.740639	lr=0.048654
Epoch[93] Batch [1199]	Speed: 2794.610176 samples/sec	accuracy=0.740550	lr=0.048511
Epoch[93] Batch [1249]	Speed: 3159.578370 samples/sec	accuracy=0.740456	lr=0.048369
[Epoch 93] training: accuracy=0.740461
[Epoch 93] speed: 2998 samples/sec	time cost: 437.878394
[Epoch 93] validation: err-top1=0.258870 err-top5=0.077494
Epoch[94] Batch [49]	Speed: 3113.165430 samples/sec	accuracy=0.751426	lr=0.048224
Epoch[94] Batch [99]	Speed: 3003.509130 samples/sec	accuracy=0.751826	lr=0.048082
Epoch[94] Batch [149]	Speed: 2944.403706 samples/sec	accuracy=0.751263	lr=0.047940
Epoch[94] Batch [199]	Speed: 2933.092259 samples/sec	accuracy=0.749526	lr=0.047798
Epoch[94] Batch [249]	Speed: 2953.272219 samples/sec	accuracy=0.749809	lr=0.047656
Epoch[94] Batch [299]	Speed: 3143.355733 samples/sec	accuracy=0.749476	lr=0.047515
Epoch[94] Batch [349]	Speed: 2880.023771 samples/sec	accuracy=0.748828	lr=0.047374
Epoch[94] Batch [399]	Speed: 2958.323657 samples/sec	accuracy=0.748855	lr=0.047233
Epoch[94] Batch [449]	Speed: 2986.634321 samples/sec	accuracy=0.748845	lr=0.047092
Epoch[94] Batch [499]	Speed: 3035.814956 samples/sec	accuracy=0.748459	lr=0.046951
Epoch[94] Batch [549]	Speed: 2959.238400 samples/sec	accuracy=0.748292	lr=0.046811
Epoch[94] Batch [599]	Speed: 2943.590464 samples/sec	accuracy=0.748130	lr=0.046670
Epoch[94] Batch [649]	Speed: 2915.620919 samples/sec	accuracy=0.747936	lr=0.046530
Epoch[94] Batch [699]	Speed: 2928.876980 samples/sec	accuracy=0.747702	lr=0.046390
Epoch[94] Batch [749]	Speed: 3004.402141 samples/sec	accuracy=0.747514	lr=0.046251
Epoch[94] Batch [799]	Speed: 3060.478709 samples/sec	accuracy=0.747278	lr=0.046111
Epoch[94] Batch [849]	Speed: 2994.610383 samples/sec	accuracy=0.747245	lr=0.045972
Epoch[94] Batch [899]	Speed: 3030.337640 samples/sec	accuracy=0.746839	lr=0.045833
Epoch[94] Batch [949]	Speed: 2944.179505 samples/sec	accuracy=0.746619	lr=0.045693
Epoch[94] Batch [999]	Speed: 3077.381535 samples/sec	accuracy=0.746216	lr=0.045555
Epoch[94] Batch [1049]	Speed: 3065.636028 samples/sec	accuracy=0.746164	lr=0.045416
Epoch[94] Batch [1099]	Speed: 3004.090291 samples/sec	accuracy=0.746025	lr=0.045278
Epoch[94] Batch [1149]	Speed: 2956.383286 samples/sec	accuracy=0.745813	lr=0.045139
Epoch[94] Batch [1199]	Speed: 2927.181220 samples/sec	accuracy=0.745794	lr=0.045001
Epoch[94] Batch [1249]	Speed: 3219.368540 samples/sec	accuracy=0.745552	lr=0.044863
[Epoch 94] training: accuracy=0.745564
[Epoch 94] speed: 2995 samples/sec	time cost: 437.709270
[Epoch 94] validation: err-top1=0.260443 err-top5=0.080018
Epoch[95] Batch [49]	Speed: 3003.726577 samples/sec	accuracy=0.756367	lr=0.044723
Epoch[95] Batch [99]	Speed: 3004.655493 samples/sec	accuracy=0.754951	lr=0.044585
Epoch[95] Batch [149]	Speed: 3075.436173 samples/sec	accuracy=0.754342	lr=0.044448
Epoch[95] Batch [199]	Speed: 2977.203044 samples/sec	accuracy=0.754570	lr=0.044311
Epoch[95] Batch [249]	Speed: 2994.032632 samples/sec	accuracy=0.755160	lr=0.044174
Epoch[95] Batch [299]	Speed: 3030.993529 samples/sec	accuracy=0.754880	lr=0.044037
Epoch[95] Batch [349]	Speed: 2897.870217 samples/sec	accuracy=0.754085	lr=0.043900
Epoch[95] Batch [399]	Speed: 3250.852260 samples/sec	accuracy=0.753545	lr=0.043764
Epoch[95] Batch [449]	Speed: 3014.228358 samples/sec	accuracy=0.753613	lr=0.043628
Epoch[95] Batch [499]	Speed: 3029.778339 samples/sec	accuracy=0.753547	lr=0.043492
Epoch[95] Batch [549]	Speed: 3021.305803 samples/sec	accuracy=0.753256	lr=0.043356
Epoch[95] Batch [599]	Speed: 2967.089235 samples/sec	accuracy=0.752845	lr=0.043220
Epoch[95] Batch [649]	Speed: 3023.109555 samples/sec	accuracy=0.752186	lr=0.043085
Epoch[95] Batch [699]	Speed: 3077.941523 samples/sec	accuracy=0.752367	lr=0.042949
Epoch[95] Batch [749]	Speed: 3005.792382 samples/sec	accuracy=0.752070	lr=0.042814
Epoch[95] Batch [799]	Speed: 3103.150114 samples/sec	accuracy=0.751934	lr=0.042679
Epoch[95] Batch [849]	Speed: 2894.628790 samples/sec	accuracy=0.751836	lr=0.042545
Epoch[95] Batch [899]	Speed: 2939.167284 samples/sec	accuracy=0.751706	lr=0.042410
Epoch[95] Batch [949]	Speed: 2958.004879 samples/sec	accuracy=0.751408	lr=0.042276
Epoch[95] Batch [999]	Speed: 3090.474555 samples/sec	accuracy=0.751092	lr=0.042141
Epoch[95] Batch [1049]	Speed: 2900.184202 samples/sec	accuracy=0.750717	lr=0.042007
Epoch[95] Batch [1099]	Speed: 2926.562985 samples/sec	accuracy=0.750643	lr=0.041874
Epoch[95] Batch [1149]	Speed: 3131.648229 samples/sec	accuracy=0.750412	lr=0.041740
Epoch[95] Batch [1199]	Speed: 2943.931487 samples/sec	accuracy=0.750325	lr=0.041607
Epoch[95] Batch [1249]	Speed: 3190.969133 samples/sec	accuracy=0.750082	lr=0.041473
[Epoch 95] training: accuracy=0.750077
[Epoch 95] speed: 3013 samples/sec	time cost: 435.406790
[Epoch 95] validation: err-top1=0.254026 err-top5=0.075674
Epoch[96] Batch [49]	Speed: 3025.695569 samples/sec	accuracy=0.760059	lr=0.041338
Epoch[96] Batch [99]	Speed: 2993.793506 samples/sec	accuracy=0.759844	lr=0.041205
Epoch[96] Batch [149]	Speed: 2974.668154 samples/sec	accuracy=0.759967	lr=0.041072
Epoch[96] Batch [199]	Speed: 3100.451283 samples/sec	accuracy=0.758755	lr=0.040940
Epoch[96] Batch [249]	Speed: 3000.513281 samples/sec	accuracy=0.758395	lr=0.040807
Epoch[96] Batch [299]	Speed: 3053.911915 samples/sec	accuracy=0.758477	lr=0.040675
Epoch[96] Batch [349]	Speed: 3239.637400 samples/sec	accuracy=0.758555	lr=0.040543
Epoch[96] Batch [399]	Speed: 3181.158793 samples/sec	accuracy=0.758782	lr=0.040412
Epoch[96] Batch [449]	Speed: 2896.882849 samples/sec	accuracy=0.758290	lr=0.040280
Epoch[96] Batch [499]	Speed: 2931.478570 samples/sec	accuracy=0.757953	lr=0.040149
Epoch[96] Batch [549]	Speed: 3074.240324 samples/sec	accuracy=0.757562	lr=0.040018
Epoch[96] Batch [599]	Speed: 2927.804426 samples/sec	accuracy=0.757347	lr=0.039887
Epoch[96] Batch [649]	Speed: 3010.523648 samples/sec	accuracy=0.756864	lr=0.039756
Epoch[96] Batch [699]	Speed: 2928.066582 samples/sec	accuracy=0.756719	lr=0.039625
Epoch[96] Batch [749]	Speed: 2804.609960 samples/sec	accuracy=0.756405	lr=0.039495
Epoch[96] Batch [799]	Speed: 2980.958434 samples/sec	accuracy=0.756287	lr=0.039365
Epoch[96] Batch [849]	Speed: 2953.672932 samples/sec	accuracy=0.756113	lr=0.039235
Epoch[96] Batch [899]	Speed: 3019.660789 samples/sec	accuracy=0.755923	lr=0.039105
Epoch[96] Batch [949]	Speed: 3026.663890 samples/sec	accuracy=0.755651	lr=0.038975
Epoch[96] Batch [999]	Speed: 2998.459120 samples/sec	accuracy=0.755331	lr=0.038846
Epoch[96] Batch [1049]	Speed: 2930.670971 samples/sec	accuracy=0.755212	lr=0.038717
Epoch[96] Batch [1099]	Speed: 3026.261211 samples/sec	accuracy=0.755022	lr=0.038588
Epoch[96] Batch [1149]	Speed: 2944.141927 samples/sec	accuracy=0.754822	lr=0.038459
Epoch[96] Batch [1199]	Speed: 2972.735136 samples/sec	accuracy=0.754688	lr=0.038330
Epoch[96] Batch [1249]	Speed: 3168.075914 samples/sec	accuracy=0.754473	lr=0.038202
[Epoch 96] training: accuracy=0.754477
[Epoch 96] speed: 3001 samples/sec	time cost: 436.822195
[Epoch 96] validation: err-top1=0.252870 err-top5=0.075574
Epoch[97] Batch [49]	Speed: 3185.599669 samples/sec	accuracy=0.770742	lr=0.038071
Epoch[97] Batch [99]	Speed: 2933.878186 samples/sec	accuracy=0.767510	lr=0.037943
Epoch[97] Batch [149]	Speed: 3030.801502 samples/sec	accuracy=0.766940	lr=0.037815
Epoch[97] Batch [199]	Speed: 2911.357311 samples/sec	accuracy=0.765591	lr=0.037687
Epoch[97] Batch [249]	Speed: 2963.058796 samples/sec	accuracy=0.765738	lr=0.037560
Epoch[97] Batch [299]	Speed: 2994.091992 samples/sec	accuracy=0.764564	lr=0.037432
Epoch[97] Batch [349]	Speed: 2921.081924 samples/sec	accuracy=0.763842	lr=0.037305
Epoch[97] Batch [399]	Speed: 3036.191248 samples/sec	accuracy=0.763477	lr=0.037178
Epoch[97] Batch [449]	Speed: 2895.007540 samples/sec	accuracy=0.763296	lr=0.037052
Epoch[97] Batch [499]	Speed: 3042.397902 samples/sec	accuracy=0.762781	lr=0.036925
Epoch[97] Batch [549]	Speed: 3059.243820 samples/sec	accuracy=0.762447	lr=0.036799
Epoch[97] Batch [599]	Speed: 3065.873420 samples/sec	accuracy=0.762238	lr=0.036673
Epoch[97] Batch [649]	Speed: 3085.628293 samples/sec	accuracy=0.762133	lr=0.036547
Epoch[97] Batch [699]	Speed: 2995.285152 samples/sec	accuracy=0.761766	lr=0.036421
Epoch[97] Batch [749]	Speed: 2981.936048 samples/sec	accuracy=0.761727	lr=0.036296
Epoch[97] Batch [799]	Speed: 2984.457961 samples/sec	accuracy=0.761401	lr=0.036170
Epoch[97] Batch [849]	Speed: 3200.947279 samples/sec	accuracy=0.761349	lr=0.036045
Epoch[97] Batch [899]	Speed: 2986.515987 samples/sec	accuracy=0.761013	lr=0.035920
Epoch[97] Batch [949]	Speed: 2974.022616 samples/sec	accuracy=0.760796	lr=0.035795
Epoch[97] Batch [999]	Speed: 3044.556539 samples/sec	accuracy=0.760653	lr=0.035671
Epoch[97] Batch [1049]	Speed: 3087.775741 samples/sec	accuracy=0.760494	lr=0.035546
Epoch[97] Batch [1099]	Speed: 2918.886110 samples/sec	accuracy=0.760337	lr=0.035422
Epoch[97] Batch [1149]	Speed: 2882.586503 samples/sec	accuracy=0.760273	lr=0.035298
Epoch[97] Batch [1199]	Speed: 3126.233022 samples/sec	accuracy=0.760020	lr=0.035174
Epoch[97] Batch [1249]	Speed: 3341.931874 samples/sec	accuracy=0.759834	lr=0.035051
[Epoch 97] training: accuracy=0.759833
[Epoch 97] speed: 3020 samples/sec	time cost: 433.918348
[Epoch 97] validation: err-top1=0.244360 err-top5=0.071429
Epoch[98] Batch [49]	Speed: 3116.945670 samples/sec	accuracy=0.774102	lr=0.034925
Epoch[98] Batch [99]	Speed: 2969.513406 samples/sec	accuracy=0.772852	lr=0.034802
Epoch[98] Batch [149]	Speed: 3134.447921 samples/sec	accuracy=0.771484	lr=0.034679
Epoch[98] Batch [199]	Speed: 3146.731697 samples/sec	accuracy=0.770156	lr=0.034556
Epoch[98] Batch [249]	Speed: 2905.753076 samples/sec	accuracy=0.769469	lr=0.034433
Epoch[98] Batch [299]	Speed: 3024.548944 samples/sec	accuracy=0.768708	lr=0.034311
Epoch[98] Batch [349]	Speed: 2905.864743 samples/sec	accuracy=0.767941	lr=0.034189
Epoch[98] Batch [399]	Speed: 2930.528556 samples/sec	accuracy=0.767798	lr=0.034067
Epoch[98] Batch [449]	Speed: 2983.817040 samples/sec	accuracy=0.767823	lr=0.033945
Epoch[98] Batch [499]	Speed: 2969.433830 samples/sec	accuracy=0.767102	lr=0.033823
Epoch[98] Batch [549]	Speed: 2823.754013 samples/sec	accuracy=0.766958	lr=0.033702
Epoch[98] Batch [599]	Speed: 2921.394978 samples/sec	accuracy=0.766956	lr=0.033581
Epoch[98] Batch [649]	Speed: 3118.487056 samples/sec	accuracy=0.766928	lr=0.033460
Epoch[98] Batch [699]	Speed: 3048.841850 samples/sec	accuracy=0.766671	lr=0.033339
Epoch[98] Batch [749]	Speed: 3084.614039 samples/sec	accuracy=0.766536	lr=0.033218
Epoch[98] Batch [799]	Speed: 2953.471526 samples/sec	accuracy=0.766417	lr=0.033098
Epoch[98] Batch [849]	Speed: 2957.319595 samples/sec	accuracy=0.766213	lr=0.032978
Epoch[98] Batch [899]	Speed: 3034.395782 samples/sec	accuracy=0.765955	lr=0.032858
Epoch[98] Batch [949]	Speed: 3064.239080 samples/sec	accuracy=0.765869	lr=0.032738
Epoch[98] Batch [999]	Speed: 3146.445337 samples/sec	accuracy=0.765666	lr=0.032618
Epoch[98] Batch [1049]	Speed: 2900.392860 samples/sec	accuracy=0.765645	lr=0.032499
Epoch[98] Batch [1099]	Speed: 2964.040536 samples/sec	accuracy=0.765439	lr=0.032380
Epoch[98] Batch [1149]	Speed: 2941.414236 samples/sec	accuracy=0.765262	lr=0.032261
Epoch[98] Batch [1199]	Speed: 2982.508725 samples/sec	accuracy=0.765074	lr=0.032142
Epoch[98] Batch [1249]	Speed: 3208.281148 samples/sec	accuracy=0.764787	lr=0.032023
[Epoch 98] training: accuracy=0.764794
[Epoch 98] speed: 3004 samples/sec	time cost: 435.812832
[Epoch 98] validation: err-top1=0.245076 err-top5=0.070455
Epoch[99] Batch [49]	Speed: 3360.646689 samples/sec	accuracy=0.776934	lr=0.031902
Epoch[99] Batch [99]	Speed: 3104.339754 samples/sec	accuracy=0.775645	lr=0.031784
Epoch[99] Batch [149]	Speed: 2826.249824 samples/sec	accuracy=0.775215	lr=0.031666
Epoch[99] Batch [199]	Speed: 2973.706869 samples/sec	accuracy=0.775269	lr=0.031548
Epoch[99] Batch [249]	Speed: 2872.988714 samples/sec	accuracy=0.775129	lr=0.031431
Epoch[99] Batch [299]	Speed: 3104.462493 samples/sec	accuracy=0.774352	lr=0.031313
Epoch[99] Batch [349]	Speed: 2917.977624 samples/sec	accuracy=0.774311	lr=0.031196
Epoch[99] Batch [399]	Speed: 3009.845666 samples/sec	accuracy=0.773828	lr=0.031079
Epoch[99] Batch [449]	Speed: 3035.382680 samples/sec	accuracy=0.773581	lr=0.030962
Epoch[99] Batch [499]	Speed: 2981.304031 samples/sec	accuracy=0.773059	lr=0.030846
Epoch[99] Batch [549]	Speed: 2983.595543 samples/sec	accuracy=0.772955	lr=0.030729
Epoch[99] Batch [599]	Speed: 2949.379092 samples/sec	accuracy=0.772855	lr=0.030613
Epoch[99] Batch [649]	Speed: 2994.394587 samples/sec	accuracy=0.772407	lr=0.030497
Epoch[99] Batch [699]	Speed: 3133.111398 samples/sec	accuracy=0.772312	lr=0.030381
Epoch[99] Batch [749]	Speed: 2902.054931 samples/sec	accuracy=0.772086	lr=0.030266
Epoch[99] Batch [799]	Speed: 2896.201021 samples/sec	accuracy=0.771642	lr=0.030150
Epoch[99] Batch [849]	Speed: 3028.296858 samples/sec	accuracy=0.771682	lr=0.030035
Epoch[99] Batch [899]	Speed: 2975.482005 samples/sec	accuracy=0.771437	lr=0.029920
Epoch[99] Batch [949]	Speed: 2944.215067 samples/sec	accuracy=0.771139	lr=0.029805
Epoch[99] Batch [999]	Speed: 3060.126766 samples/sec	accuracy=0.770868	lr=0.029691
Epoch[99] Batch [1049]	Speed: 2948.650387 samples/sec	accuracy=0.770857	lr=0.029576
Epoch[99] Batch [1099]	Speed: 3081.470655 samples/sec	accuracy=0.770732	lr=0.029462
Epoch[99] Batch [1149]	Speed: 3072.777575 samples/sec	accuracy=0.770633	lr=0.029348
Epoch[99] Batch [1199]	Speed: 2941.093533 samples/sec	accuracy=0.770460	lr=0.029234
Epoch[99] Batch [1249]	Speed: 3144.107035 samples/sec	accuracy=0.770260	lr=0.029121
[Epoch 99] training: accuracy=0.770246
[Epoch 99] speed: 3003 samples/sec	time cost: 437.008969
[Epoch 99] validation: err-top1=0.236846 err-top5=0.068439
Epoch[100] Batch [49]	Speed: 3223.748152 samples/sec	accuracy=0.781484	lr=0.029005
Epoch[100] Batch [99]	Speed: 3038.576029 samples/sec	accuracy=0.780176	lr=0.028892
Epoch[100] Batch [149]	Speed: 2851.006700 samples/sec	accuracy=0.780534	lr=0.028779
Epoch[100] Batch [199]	Speed: 3075.544833 samples/sec	accuracy=0.779429	lr=0.028666
Epoch[100] Batch [249]	Speed: 2960.755603 samples/sec	accuracy=0.778906	lr=0.028554
Epoch[100] Batch [299]	Speed: 3165.407655 samples/sec	accuracy=0.778258	lr=0.028441
Epoch[100] Batch [349]	Speed: 2981.827195 samples/sec	accuracy=0.778145	lr=0.028329
Epoch[100] Batch [399]	Speed: 3187.637921 samples/sec	accuracy=0.777690	lr=0.028217
Epoch[100] Batch [449]	Speed: 3210.480811 samples/sec	accuracy=0.777431	lr=0.028105
Epoch[100] Batch [499]	Speed: 3044.115988 samples/sec	accuracy=0.777000	lr=0.027994
Epoch[100] Batch [549]	Speed: 2994.118959 samples/sec	accuracy=0.776854	lr=0.027883
Epoch[100] Batch [599]	Speed: 3072.361434 samples/sec	accuracy=0.776860	lr=0.027772
Epoch[100] Batch [649]	Speed: 2831.557240 samples/sec	accuracy=0.776665	lr=0.027661
Epoch[100] Batch [699]	Speed: 3117.074114 samples/sec	accuracy=0.776582	lr=0.027550
Epoch[100] Batch [749]	Speed: 2962.236280 samples/sec	accuracy=0.776443	lr=0.027439
Epoch[100] Batch [799]	Speed: 2927.893842 samples/sec	accuracy=0.776157	lr=0.027329
Epoch[100] Batch [849]	Speed: 3118.471116 samples/sec	accuracy=0.776165	lr=0.027219
Epoch[100] Batch [899]	Speed: 3200.529663 samples/sec	accuracy=0.775905	lr=0.027109
Epoch[100] Batch [949]	Speed: 2901.939283 samples/sec	accuracy=0.775897	lr=0.026999
Epoch[100] Batch [999]	Speed: 3100.776700 samples/sec	accuracy=0.775719	lr=0.026890
Epoch[100] Batch [1049]	Speed: 3001.561660 samples/sec	accuracy=0.775682	lr=0.026781
Epoch[100] Batch [1099]	Speed: 3014.961649 samples/sec	accuracy=0.775485	lr=0.026672
Epoch[100] Batch [1149]	Speed: 2992.682815 samples/sec	accuracy=0.775459	lr=0.026563
Epoch[100] Batch [1199]	Speed: 2966.043656 samples/sec	accuracy=0.775252	lr=0.026454
Epoch[100] Batch [1249]	Speed: 3160.675092 samples/sec	accuracy=0.774962	lr=0.026346
[Epoch 100] training: accuracy=0.774968
[Epoch 100] speed: 3038 samples/sec	time cost: 433.120025
[Epoch 100] validation: err-top1=0.236727 err-top5=0.067702
Epoch[101] Batch [49]	Speed: 3225.000593 samples/sec	accuracy=0.786563	lr=0.026235
Epoch[101] Batch [99]	Speed: 2981.346207 samples/sec	accuracy=0.785479	lr=0.026127
Epoch[101] Batch [149]	Speed: 2944.943560 samples/sec	accuracy=0.784212	lr=0.026020
Epoch[101] Batch [199]	Speed: 3069.482709 samples/sec	accuracy=0.784136	lr=0.025912
Epoch[101] Batch [249]	Speed: 2994.521898 samples/sec	accuracy=0.783512	lr=0.025805
Epoch[101] Batch [299]	Speed: 3167.413648 samples/sec	accuracy=0.783835	lr=0.025697
Epoch[101] Batch [349]	Speed: 3010.060994 samples/sec	accuracy=0.783309	lr=0.025590
Epoch[101] Batch [399]	Speed: 3012.810354 samples/sec	accuracy=0.782813	lr=0.025484
Epoch[101] Batch [449]	Speed: 3099.855646 samples/sec	accuracy=0.782602	lr=0.025377
Epoch[101] Batch [499]	Speed: 2930.955362 samples/sec	accuracy=0.782586	lr=0.025271
Epoch[101] Batch [549]	Speed: 2981.871952 samples/sec	accuracy=0.782337	lr=0.025165
Epoch[101] Batch [599]	Speed: 3026.361775 samples/sec	accuracy=0.782264	lr=0.025059
Epoch[101] Batch [649]	Speed: 2949.517552 samples/sec	accuracy=0.781825	lr=0.024953
Epoch[101] Batch [699]	Speed: 3043.306256 samples/sec	accuracy=0.781670	lr=0.024847
Epoch[101] Batch [749]	Speed: 2929.439007 samples/sec	accuracy=0.781647	lr=0.024742
Epoch[101] Batch [799]	Speed: 3037.676984 samples/sec	accuracy=0.781494	lr=0.024637
Epoch[101] Batch [849]	Speed: 2880.292700 samples/sec	accuracy=0.781333	lr=0.024532
Epoch[101] Batch [899]	Speed: 3082.052922 samples/sec	accuracy=0.781249	lr=0.024427
Epoch[101] Batch [949]	Speed: 3082.356305 samples/sec	accuracy=0.780956	lr=0.024323
Epoch[101] Batch [999]	Speed: 2984.639764 samples/sec	accuracy=0.780773	lr=0.024219
Epoch[101] Batch [1049]	Speed: 2921.524345 samples/sec	accuracy=0.780622	lr=0.024115
Epoch[101] Batch [1099]	Speed: 2996.792881 samples/sec	accuracy=0.780558	lr=0.024011
Epoch[101] Batch [1149]	Speed: 3151.716677 samples/sec	accuracy=0.780526	lr=0.023907
Epoch[101] Batch [1199]	Speed: 2977.908683 samples/sec	accuracy=0.780516	lr=0.023804
Epoch[101] Batch [1249]	Speed: 3203.561976 samples/sec	accuracy=0.780409	lr=0.023700
[Epoch 101] training: accuracy=0.780430
[Epoch 101] speed: 3022 samples/sec	time cost: 434.022707
[Epoch 101] validation: err-top1=0.237663 err-top5=0.067383
Epoch[102] Batch [49]	Speed: 3147.011606 samples/sec	accuracy=0.790273	lr=0.023595
Epoch[102] Batch [99]	Speed: 2945.806363 samples/sec	accuracy=0.790137	lr=0.023493
Epoch[102] Batch [149]	Speed: 3071.362691 samples/sec	accuracy=0.790156	lr=0.023390
Epoch[102] Batch [199]	Speed: 2951.251547 samples/sec	accuracy=0.789795	lr=0.023288
Epoch[102] Batch [249]	Speed: 3044.981328 samples/sec	accuracy=0.789898	lr=0.023185
Epoch[102] Batch [299]	Speed: 2891.749458 samples/sec	accuracy=0.789329	lr=0.023083
Epoch[102] Batch [349]	Speed: 2993.850560 samples/sec	accuracy=0.788686	lr=0.022982
Epoch[102] Batch [399]	Speed: 2901.129298 samples/sec	accuracy=0.788630	lr=0.022880
Epoch[102] Batch [449]	Speed: 3121.402127 samples/sec	accuracy=0.788366	lr=0.022779
Epoch[102] Batch [499]	Speed: 3144.387996 samples/sec	accuracy=0.787883	lr=0.022678
Epoch[102] Batch [549]	Speed: 3049.562590 samples/sec	accuracy=0.787651	lr=0.022577
Epoch[102] Batch [599]	Speed: 3023.750311 samples/sec	accuracy=0.787440	lr=0.022476
Epoch[102] Batch [649]	Speed: 2948.016253 samples/sec	accuracy=0.787248	lr=0.022376
Epoch[102] Batch [699]	Speed: 2994.636608 samples/sec	accuracy=0.786872	lr=0.022276
Epoch[102] Batch [749]	Speed: 2964.139216 samples/sec	accuracy=0.786561	lr=0.022175
Epoch[102] Batch [799]	Speed: 3017.684861 samples/sec	accuracy=0.786238	lr=0.022076
Epoch[102] Batch [849]	Speed: 3066.304179 samples/sec	accuracy=0.786260	lr=0.021976
Epoch[102] Batch [899]	Speed: 2955.524730 samples/sec	accuracy=0.786314	lr=0.021877
Epoch[102] Batch [949]	Speed: 3030.277177 samples/sec	accuracy=0.786198	lr=0.021777
Epoch[102] Batch [999]	Speed: 2878.047248 samples/sec	accuracy=0.785939	lr=0.021678
Epoch[102] Batch [1049]	Speed: 2992.257856 samples/sec	accuracy=0.785778	lr=0.021580
Epoch[102] Batch [1099]	Speed: 2972.822338 samples/sec	accuracy=0.785869	lr=0.021481
Epoch[102] Batch [1149]	Speed: 3039.318463 samples/sec	accuracy=0.785701	lr=0.021383
Epoch[102] Batch [1199]	Speed: 2980.861941 samples/sec	accuracy=0.785426	lr=0.021285
Epoch[102] Batch [1249]	Speed: 3165.714743 samples/sec	accuracy=0.785382	lr=0.021187
[Epoch 102] training: accuracy=0.785397
[Epoch 102] speed: 3007 samples/sec	time cost: 435.756881
[Epoch 102] validation: err-top1=0.231645 err-top5=0.065749
Epoch[103] Batch [49]	Speed: 3193.595237 samples/sec	accuracy=0.796289	lr=0.021087
Epoch[103] Batch [99]	Speed: 2941.909265 samples/sec	accuracy=0.795420	lr=0.020990
Epoch[103] Batch [149]	Speed: 2998.982165 samples/sec	accuracy=0.795273	lr=0.020892
Epoch[103] Batch [199]	Speed: 3145.428585 samples/sec	accuracy=0.794263	lr=0.020795
Epoch[103] Batch [249]	Speed: 3138.159214 samples/sec	accuracy=0.794109	lr=0.020698
Epoch[103] Batch [299]	Speed: 3023.675678 samples/sec	accuracy=0.793675	lr=0.020602
Epoch[103] Batch [349]	Speed: 3155.318922 samples/sec	accuracy=0.793557	lr=0.020505
Epoch[103] Batch [399]	Speed: 2969.011261 samples/sec	accuracy=0.793381	lr=0.020409
Epoch[103] Batch [449]	Speed: 3271.138050 samples/sec	accuracy=0.792689	lr=0.020313
Epoch[103] Batch [499]	Speed: 3019.725416 samples/sec	accuracy=0.792654	lr=0.020217
Epoch[103] Batch [549]	Speed: 2918.120685 samples/sec	accuracy=0.792670	lr=0.020122
Epoch[103] Batch [599]	Speed: 2944.691375 samples/sec	accuracy=0.792329	lr=0.020026
Epoch[103] Batch [649]	Speed: 2939.919482 samples/sec	accuracy=0.792163	lr=0.019931
Epoch[103] Batch [699]	Speed: 3018.511435 samples/sec	accuracy=0.792118	lr=0.019836
Epoch[103] Batch [749]	Speed: 3076.004706 samples/sec	accuracy=0.792047	lr=0.019742
Epoch[103] Batch [799]	Speed: 3056.320841 samples/sec	accuracy=0.791862	lr=0.019647
Epoch[103] Batch [849]	Speed: 3013.174158 samples/sec	accuracy=0.791690	lr=0.019553
Epoch[103] Batch [899]	Speed: 2906.214267 samples/sec	accuracy=0.791695	lr=0.019459
Epoch[103] Batch [949]	Speed: 3181.012339 samples/sec	accuracy=0.791298	lr=0.019365
Epoch[103] Batch [999]	Speed: 3035.293786 samples/sec	accuracy=0.791137	lr=0.019271
Epoch[103] Batch [1049]	Speed: 3047.076575 samples/sec	accuracy=0.791155	lr=0.019178
Epoch[103] Batch [1099]	Speed: 3301.286320 samples/sec	accuracy=0.791032	lr=0.019085
Epoch[103] Batch [1149]	Speed: 3141.593277 samples/sec	accuracy=0.791000	lr=0.018992
Epoch[103] Batch [1199]	Speed: 3006.900617 samples/sec	accuracy=0.790785	lr=0.018899
Epoch[103] Batch [1249]	Speed: 3224.906784 samples/sec	accuracy=0.790825	lr=0.018806
[Epoch 103] training: accuracy=0.790819
[Epoch 103] speed: 3060 samples/sec	time cost: 430.274910
[Epoch 103] validation: err-top1=0.228635 err-top5=0.063576
Epoch[104] Batch [49]	Speed: 3080.342923 samples/sec	accuracy=0.798027	lr=0.018712
Epoch[104] Batch [99]	Speed: 2967.000319 samples/sec	accuracy=0.800068	lr=0.018620
Epoch[104] Batch [149]	Speed: 2962.866614 samples/sec	accuracy=0.799824	lr=0.018528
Epoch[104] Batch [199]	Speed: 2997.005509 samples/sec	accuracy=0.800166	lr=0.018436
Epoch[104] Batch [249]	Speed: 3119.273227 samples/sec	accuracy=0.800262	lr=0.018345
Epoch[104] Batch [299]	Speed: 3023.310908 samples/sec	accuracy=0.799368	lr=0.018254
Epoch[104] Batch [349]	Speed: 3119.970768 samples/sec	accuracy=0.799459	lr=0.018163
Epoch[104] Batch [399]	Speed: 3023.475169 samples/sec	accuracy=0.799375	lr=0.018072
Epoch[104] Batch [449]	Speed: 2951.324189 samples/sec	accuracy=0.799355	lr=0.017981
Epoch[104] Batch [499]	Speed: 2936.915850 samples/sec	accuracy=0.799072	lr=0.017891
Epoch[104] Batch [549]	Speed: 2982.896073 samples/sec	accuracy=0.798752	lr=0.017801
Epoch[104] Batch [599]	Speed: 3084.061808 samples/sec	accuracy=0.798371	lr=0.017711
Epoch[104] Batch [649]	Speed: 2973.152880 samples/sec	accuracy=0.798130	lr=0.017621
Epoch[104] Batch [699]	Speed: 3036.214086 samples/sec	accuracy=0.797800	lr=0.017532
Epoch[104] Batch [749]	Speed: 3106.384637 samples/sec	accuracy=0.797559	lr=0.017442
Epoch[104] Batch [799]	Speed: 3045.912827 samples/sec	accuracy=0.797402	lr=0.017353
Epoch[104] Batch [849]	Speed: 2923.573236 samples/sec	accuracy=0.797317	lr=0.017264
Epoch[104] Batch [899]	Speed: 3083.918621 samples/sec	accuracy=0.797168	lr=0.017176
Epoch[104] Batch [949]	Speed: 2932.201814 samples/sec	accuracy=0.797100	lr=0.017087
Epoch[104] Batch [999]	Speed: 2918.586762 samples/sec	accuracy=0.796961	lr=0.016999
Epoch[104] Batch [1049]	Speed: 2950.771250 samples/sec	accuracy=0.796838	lr=0.016911
Epoch[104] Batch [1099]	Speed: 2986.041914 samples/sec	accuracy=0.796719	lr=0.016823
Epoch[104] Batch [1149]	Speed: 2962.425520 samples/sec	accuracy=0.796604	lr=0.016736
Epoch[104] Batch [1199]	Speed: 2942.093861 samples/sec	accuracy=0.796592	lr=0.016648
Epoch[104] Batch [1249]	Speed: 3145.578554 samples/sec	accuracy=0.796514	lr=0.016561
[Epoch 104] training: accuracy=0.796530
[Epoch 104] speed: 3006 samples/sec	time cost: 436.528425
[Epoch 104] validation: err-top1=0.228292 err-top5=0.063761
Epoch[105] Batch [49]	Speed: 2963.699951 samples/sec	accuracy=0.807461	lr=0.016473
Epoch[105] Batch [99]	Speed: 2921.329763 samples/sec	accuracy=0.807705	lr=0.016386
Epoch[105] Batch [149]	Speed: 2989.754133 samples/sec	accuracy=0.806035	lr=0.016299
Epoch[105] Batch [199]	Speed: 2904.281494 samples/sec	accuracy=0.805259	lr=0.016213
Epoch[105] Batch [249]	Speed: 3077.775217 samples/sec	accuracy=0.804875	lr=0.016127
Epoch[105] Batch [299]	Speed: 2920.974130 samples/sec	accuracy=0.804808	lr=0.016041
Epoch[105] Batch [349]	Speed: 2987.132141 samples/sec	accuracy=0.804428	lr=0.015956
Epoch[105] Batch [399]	Speed: 2932.664551 samples/sec	accuracy=0.803918	lr=0.015870
Epoch[105] Batch [449]	Speed: 3028.994714 samples/sec	accuracy=0.804008	lr=0.015785
Epoch[105] Batch [499]	Speed: 3126.157430 samples/sec	accuracy=0.803787	lr=0.015700
Epoch[105] Batch [549]	Speed: 3057.960981 samples/sec	accuracy=0.803642	lr=0.015616
Epoch[105] Batch [599]	Speed: 3014.559201 samples/sec	accuracy=0.803345	lr=0.015531
Epoch[105] Batch [649]	Speed: 3109.820806 samples/sec	accuracy=0.803098	lr=0.015447
Epoch[105] Batch [699]	Speed: 3077.272790 samples/sec	accuracy=0.802928	lr=0.015363
Epoch[105] Batch [749]	Speed: 3082.039033 samples/sec	accuracy=0.802719	lr=0.015279
Epoch[105] Batch [799]	Speed: 3098.142915 samples/sec	accuracy=0.802753	lr=0.015195
Epoch[105] Batch [849]	Speed: 2965.844369 samples/sec	accuracy=0.802611	lr=0.015112
Epoch[105] Batch [899]	Speed: 3019.408594 samples/sec	accuracy=0.802445	lr=0.015029
Epoch[105] Batch [949]	Speed: 2983.805224 samples/sec	accuracy=0.802308	lr=0.014946
Epoch[105] Batch [999]	Speed: 3134.319917 samples/sec	accuracy=0.802130	lr=0.014863
Epoch[105] Batch [1049]	Speed: 2865.522109 samples/sec	accuracy=0.802117	lr=0.014781
Epoch[105] Batch [1099]	Speed: 3066.593959 samples/sec	accuracy=0.801990	lr=0.014699
Epoch[105] Batch [1149]	Speed: 2984.732851 samples/sec	accuracy=0.801732	lr=0.014616
Epoch[105] Batch [1199]	Speed: 3024.504089 samples/sec	accuracy=0.801692	lr=0.014535
Epoch[105] Batch [1249]	Speed: 3109.061673 samples/sec	accuracy=0.801659	lr=0.014453
[Epoch 105] training: accuracy=0.801656
[Epoch 105] speed: 3014 samples/sec	time cost: 435.252683
[Epoch 105] validation: err-top1=0.228237 err-top5=0.063437
Epoch[106] Batch [49]	Speed: 3084.728887 samples/sec	accuracy=0.810586	lr=0.014370
Epoch[106] Batch [99]	Speed: 3054.366079 samples/sec	accuracy=0.810439	lr=0.014289
Epoch[106] Batch [149]	Speed: 2975.603960 samples/sec	accuracy=0.809837	lr=0.014208
Epoch[106] Batch [199]	Speed: 3117.563418 samples/sec	accuracy=0.809800	lr=0.014127
Epoch[106] Batch [249]	Speed: 2992.334949 samples/sec	accuracy=0.809695	lr=0.014047
Epoch[106] Batch [299]	Speed: 2887.071086 samples/sec	accuracy=0.809242	lr=0.013966
Epoch[106] Batch [349]	Speed: 2974.371756 samples/sec	accuracy=0.808970	lr=0.013886
Epoch[106] Batch [399]	Speed: 3011.614470 samples/sec	accuracy=0.808584	lr=0.013806
Epoch[106] Batch [449]	Speed: 3088.847966 samples/sec	accuracy=0.808650	lr=0.013727
Epoch[106] Batch [499]	Speed: 3021.286037 samples/sec	accuracy=0.808436	lr=0.013647
Epoch[106] Batch [549]	Speed: 3000.403780 samples/sec	accuracy=0.808533	lr=0.013568
Epoch[106] Batch [599]	Speed: 3013.479566 samples/sec	accuracy=0.808398	lr=0.013489
Epoch[106] Batch [649]	Speed: 2984.918711 samples/sec	accuracy=0.808245	lr=0.013411
Epoch[106] Batch [699]	Speed: 3115.852316 samples/sec	accuracy=0.808054	lr=0.013332
Epoch[106] Batch [749]	Speed: 3045.313645 samples/sec	accuracy=0.808001	lr=0.013254
Epoch[106] Batch [799]	Speed: 2966.167912 samples/sec	accuracy=0.807968	lr=0.013176
Epoch[106] Batch [849]	Speed: 3095.945454 samples/sec	accuracy=0.807807	lr=0.013098
Epoch[106] Batch [899]	Speed: 2982.488055 samples/sec	accuracy=0.807680	lr=0.013020
Epoch[106] Batch [949]	Speed: 3073.935413 samples/sec	accuracy=0.807504	lr=0.012943
Epoch[106] Batch [999]	Speed: 3087.642065 samples/sec	accuracy=0.807305	lr=0.012866
Epoch[106] Batch [1049]	Speed: 2898.094617 samples/sec	accuracy=0.807237	lr=0.012789
Epoch[106] Batch [1099]	Speed: 3054.491676 samples/sec	accuracy=0.807255	lr=0.012712
Epoch[106] Batch [1149]	Speed: 2986.940438 samples/sec	accuracy=0.807206	lr=0.012636
Epoch[106] Batch [1199]	Speed: 3057.864533 samples/sec	accuracy=0.807189	lr=0.012559
Epoch[106] Batch [1249]	Speed: 3124.687972 samples/sec	accuracy=0.807215	lr=0.012483
[Epoch 106] training: accuracy=0.807211
[Epoch 106] speed: 3024 samples/sec	time cost: 433.645083
[Epoch 106] validation: err-top1=0.221460 err-top5=0.060547
Epoch[107] Batch [49]	Speed: 3178.158328 samples/sec	accuracy=0.811367	lr=0.012406
Epoch[107] Batch [99]	Speed: 3055.053751 samples/sec	accuracy=0.812754	lr=0.012330
Epoch[107] Batch [149]	Speed: 2925.119062 samples/sec	accuracy=0.812878	lr=0.012255
Epoch[107] Batch [199]	Speed: 2923.130910 samples/sec	accuracy=0.813613	lr=0.012180
Epoch[107] Batch [249]	Speed: 2992.304136 samples/sec	accuracy=0.813234	lr=0.012105
Epoch[107] Batch [299]	Speed: 3012.678948 samples/sec	accuracy=0.813245	lr=0.012030
Epoch[107] Batch [349]	Speed: 2987.649454 samples/sec	accuracy=0.813507	lr=0.011956
Epoch[107] Batch [399]	Speed: 2993.616888 samples/sec	accuracy=0.813298	lr=0.011881
Epoch[107] Batch [449]	Speed: 3110.245130 samples/sec	accuracy=0.812914	lr=0.011807
Epoch[107] Batch [499]	Speed: 3018.140445 samples/sec	accuracy=0.813094	lr=0.011734
Epoch[107] Batch [549]	Speed: 2973.678539 samples/sec	accuracy=0.812882	lr=0.011660
Epoch[107] Batch [599]	Speed: 3186.113135 samples/sec	accuracy=0.812918	lr=0.011587
Epoch[107] Batch [649]	Speed: 3050.508117 samples/sec	accuracy=0.812796	lr=0.011513
Epoch[107] Batch [699]	Speed: 2954.909469 samples/sec	accuracy=0.812616	lr=0.011441
Epoch[107] Batch [749]	Speed: 3050.869120 samples/sec	accuracy=0.812512	lr=0.011368
Epoch[107] Batch [799]	Speed: 3014.162359 samples/sec	accuracy=0.812544	lr=0.011295
Epoch[107] Batch [849]	Speed: 3218.956864 samples/sec	accuracy=0.812501	lr=0.011223
Epoch[107] Batch [899]	Speed: 2959.963944 samples/sec	accuracy=0.812375	lr=0.011151
Epoch[107] Batch [949]	Speed: 3024.560190 samples/sec	accuracy=0.812207	lr=0.011079
Epoch[107] Batch [999]	Speed: 3013.703238 samples/sec	accuracy=0.812080	lr=0.011008
Epoch[107] Batch [1049]	Speed: 2974.312846 samples/sec	accuracy=0.812067	lr=0.010936
Epoch[107] Batch [1099]	Speed: 3222.307039 samples/sec	accuracy=0.812007	lr=0.010865
Epoch[107] Batch [1149]	Speed: 3066.968722 samples/sec	accuracy=0.811929	lr=0.010794
Epoch[107] Batch [1199]	Speed: 3014.120307 samples/sec	accuracy=0.811899	lr=0.010724
Epoch[107] Batch [1249]	Speed: 3251.451370 samples/sec	accuracy=0.811798	lr=0.010653
[Epoch 107] training: accuracy=0.811785
[Epoch 107] speed: 3042 samples/sec	time cost: 431.242632
[Epoch 107] validation: err-top1=0.223453 err-top5=0.059889
Epoch[108] Batch [49]	Speed: 3232.242830 samples/sec	accuracy=0.818848	lr=0.010582
Epoch[108] Batch [99]	Speed: 3031.148871 samples/sec	accuracy=0.819551	lr=0.010512
Epoch[108] Batch [149]	Speed: 2964.387664 samples/sec	accuracy=0.818952	lr=0.010442
Epoch[108] Batch [199]	Speed: 3019.117136 samples/sec	accuracy=0.818926	lr=0.010373
Epoch[108] Batch [249]	Speed: 3000.952665 samples/sec	accuracy=0.819301	lr=0.010303
Epoch[108] Batch [299]	Speed: 2848.913615 samples/sec	accuracy=0.818965	lr=0.010234
Epoch[108] Batch [349]	Speed: 3003.111664 samples/sec	accuracy=0.818708	lr=0.010165
Epoch[108] Batch [399]	Speed: 2833.710814 samples/sec	accuracy=0.818496	lr=0.010097
Epoch[108] Batch [449]	Speed: 3073.613581 samples/sec	accuracy=0.818587	lr=0.010028
Epoch[108] Batch [499]	Speed: 2872.426660 samples/sec	accuracy=0.818457	lr=0.009960
Epoch[108] Batch [549]	Speed: 3061.686398 samples/sec	accuracy=0.817963	lr=0.009892
Epoch[108] Batch [599]	Speed: 2835.591581 samples/sec	accuracy=0.817868	lr=0.009824
Epoch[108] Batch [649]	Speed: 3039.364103 samples/sec	accuracy=0.817700	lr=0.009757
Epoch[108] Batch [699]	Speed: 3061.110186 samples/sec	accuracy=0.817725	lr=0.009690
Epoch[108] Batch [749]	Speed: 3081.725276 samples/sec	accuracy=0.817684	lr=0.009623
Epoch[108] Batch [799]	Speed: 3111.415108 samples/sec	accuracy=0.817570	lr=0.009556
Epoch[108] Batch [849]	Speed: 3071.811208 samples/sec	accuracy=0.817536	lr=0.009489
Epoch[108] Batch [899]	Speed: 3115.755030 samples/sec	accuracy=0.817540	lr=0.009423
Epoch[108] Batch [949]	Speed: 3040.383851 samples/sec	accuracy=0.817499	lr=0.009357
Epoch[108] Batch [999]	Speed: 2951.692890 samples/sec	accuracy=0.817359	lr=0.009291
Epoch[108] Batch [1049]	Speed: 2969.816927 samples/sec	accuracy=0.817367	lr=0.009225
Epoch[108] Batch [1099]	Speed: 2932.526147 samples/sec	accuracy=0.817361	lr=0.009160
Epoch[108] Batch [1149]	Speed: 2973.474396 samples/sec	accuracy=0.817138	lr=0.009095
Epoch[108] Batch [1199]	Speed: 2955.780157 samples/sec	accuracy=0.817303	lr=0.009030
Epoch[108] Batch [1249]	Speed: 3109.605694 samples/sec	accuracy=0.817295	lr=0.008965
[Epoch 108] training: accuracy=0.817295
[Epoch 108] speed: 3002 samples/sec	time cost: 437.302861
[Epoch 108] validation: err-top1=0.219786 err-top5=0.058654
Epoch[109] Batch [49]	Speed: 3069.484947 samples/sec	accuracy=0.827012	lr=0.008899
Epoch[109] Batch [99]	Speed: 2960.253314 samples/sec	accuracy=0.824590	lr=0.008835
Epoch[109] Batch [149]	Speed: 2978.186539 samples/sec	accuracy=0.825033	lr=0.008771
Epoch[109] Batch [199]	Speed: 3047.087081 samples/sec	accuracy=0.824199	lr=0.008707
Epoch[109] Batch [249]	Speed: 3007.688057 samples/sec	accuracy=0.824320	lr=0.008643
Epoch[109] Batch [299]	Speed: 3051.360620 samples/sec	accuracy=0.824447	lr=0.008580
Epoch[109] Batch [349]	Speed: 2988.644523 samples/sec	accuracy=0.823806	lr=0.008517
Epoch[109] Batch [399]	Speed: 3047.947623 samples/sec	accuracy=0.823547	lr=0.008454
Epoch[109] Batch [449]	Speed: 2960.010822 samples/sec	accuracy=0.823481	lr=0.008391
Epoch[109] Batch [499]	Speed: 2902.071521 samples/sec	accuracy=0.823127	lr=0.008329
Epoch[109] Batch [549]	Speed: 3017.667602 samples/sec	accuracy=0.822994	lr=0.008266
Epoch[109] Batch [599]	Speed: 2931.960533 samples/sec	accuracy=0.822799	lr=0.008204
Epoch[109] Batch [649]	Speed: 3041.887999 samples/sec	accuracy=0.822697	lr=0.008143
Epoch[109] Batch [699]	Speed: 3075.304841 samples/sec	accuracy=0.822566	lr=0.008081
Epoch[109] Batch [749]	Speed: 3041.204647 samples/sec	accuracy=0.822503	lr=0.008020
Epoch[109] Batch [799]	Speed: 3077.136670 samples/sec	accuracy=0.822439	lr=0.007959
Epoch[109] Batch [849]	Speed: 3010.287915 samples/sec	accuracy=0.822400	lr=0.007898
Epoch[109] Batch [899]	Speed: 3048.971105 samples/sec	accuracy=0.822441	lr=0.007837
Epoch[109] Batch [949]	Speed: 3043.331012 samples/sec	accuracy=0.822429	lr=0.007777
Epoch[109] Batch [999]	Speed: 2886.407214 samples/sec	accuracy=0.822357	lr=0.007716
Epoch[109] Batch [1049]	Speed: 3128.019788 samples/sec	accuracy=0.822254	lr=0.007656
Epoch[109] Batch [1099]	Speed: 2928.062270 samples/sec	accuracy=0.822254	lr=0.007597
Epoch[109] Batch [1149]	Speed: 2924.138286 samples/sec	accuracy=0.822047	lr=0.007537
Epoch[109] Batch [1199]	Speed: 2883.083216 samples/sec	accuracy=0.821978	lr=0.007478
Epoch[109] Batch [1249]	Speed: 3106.914911 samples/sec	accuracy=0.821981	lr=0.007419
[Epoch 109] training: accuracy=0.821976
[Epoch 109] speed: 3002 samples/sec	time cost: 436.694445
[Epoch 109] validation: err-top1=0.215482 err-top5=0.057298
Epoch[110] Batch [49]	Speed: 3126.458452 samples/sec	accuracy=0.830977	lr=0.007359
Epoch[110] Batch [99]	Speed: 3000.245873 samples/sec	accuracy=0.830693	lr=0.007300
Epoch[110] Batch [149]	Speed: 2986.605744 samples/sec	accuracy=0.829368	lr=0.007242
Epoch[110] Batch [199]	Speed: 3003.628436 samples/sec	accuracy=0.828896	lr=0.007184
Epoch[110] Batch [249]	Speed: 3026.585871 samples/sec	accuracy=0.828633	lr=0.007126
Epoch[110] Batch [299]	Speed: 3067.455434 samples/sec	accuracy=0.828750	lr=0.007068
Epoch[110] Batch [349]	Speed: 3002.768046 samples/sec	accuracy=0.828301	lr=0.007011
Epoch[110] Batch [399]	Speed: 3036.066809 samples/sec	accuracy=0.828147	lr=0.006954
Epoch[110] Batch [449]	Speed: 3170.525800 samples/sec	accuracy=0.828140	lr=0.006897
Epoch[110] Batch [499]	Speed: 3014.433524 samples/sec	accuracy=0.827865	lr=0.006840
Epoch[110] Batch [549]	Speed: 3122.595544 samples/sec	accuracy=0.827349	lr=0.006784
Epoch[110] Batch [599]	Speed: 2950.241093 samples/sec	accuracy=0.827186	lr=0.006727
Epoch[110] Batch [649]	Speed: 2843.559446 samples/sec	accuracy=0.827048	lr=0.006671
Epoch[110] Batch [699]	Speed: 3079.903220 samples/sec	accuracy=0.826989	lr=0.006615
Epoch[110] Batch [749]	Speed: 2937.058726 samples/sec	accuracy=0.826928	lr=0.006560
Epoch[110] Batch [799]	Speed: 2988.567620 samples/sec	accuracy=0.826840	lr=0.006504
Epoch[110] Batch [849]	Speed: 3042.275496 samples/sec	accuracy=0.826824	lr=0.006449
Epoch[110] Batch [899]	Speed: 2971.942611 samples/sec	accuracy=0.826873	lr=0.006394
Epoch[110] Batch [949]	Speed: 2910.158407 samples/sec	accuracy=0.826822	lr=0.006340
Epoch[110] Batch [999]	Speed: 2919.071478 samples/sec	accuracy=0.826922	lr=0.006285
Epoch[110] Batch [1049]	Speed: 2970.522685 samples/sec	accuracy=0.826871	lr=0.006231
Epoch[110] Batch [1099]	Speed: 3007.120661 samples/sec	accuracy=0.826976	lr=0.006177
Epoch[110] Batch [1149]	Speed: 3051.330964 samples/sec	accuracy=0.826963	lr=0.006123
Epoch[110] Batch [1199]	Speed: 3086.045463 samples/sec	accuracy=0.826892	lr=0.006070
Epoch[110] Batch [1249]	Speed: 3066.392491 samples/sec	accuracy=0.826805	lr=0.006017
[Epoch 110] training: accuracy=0.826807
[Epoch 110] speed: 3011 samples/sec	time cost: 435.389111
[Epoch 110] validation: err-top1=0.214762 err-top5=0.057312
Epoch[111] Batch [49]	Speed: 3016.178833 samples/sec	accuracy=0.835195	lr=0.005963
Epoch[111] Batch [99]	Speed: 2971.951248 samples/sec	accuracy=0.834424	lr=0.005910
Epoch[111] Batch [149]	Speed: 3031.422415 samples/sec	accuracy=0.833294	lr=0.005857
Epoch[111] Batch [199]	Speed: 2940.105358 samples/sec	accuracy=0.832295	lr=0.005805
Epoch[111] Batch [249]	Speed: 2942.183104 samples/sec	accuracy=0.831965	lr=0.005753
Epoch[111] Batch [299]	Speed: 2889.231714 samples/sec	accuracy=0.832035	lr=0.005701
Epoch[111] Batch [349]	Speed: 3031.476249 samples/sec	accuracy=0.831861	lr=0.005649
Epoch[111] Batch [399]	Speed: 3109.649731 samples/sec	accuracy=0.831799	lr=0.005598
Epoch[111] Batch [449]	Speed: 2965.280407 samples/sec	accuracy=0.831567	lr=0.005547
Epoch[111] Batch [499]	Speed: 2846.021759 samples/sec	accuracy=0.831539	lr=0.005496
Epoch[111] Batch [549]	Speed: 2793.614790 samples/sec	accuracy=0.831500	lr=0.005445
Epoch[111] Batch [599]	Speed: 2927.303398 samples/sec	accuracy=0.831382	lr=0.005394
Epoch[111] Batch [649]	Speed: 2992.740911 samples/sec	accuracy=0.831492	lr=0.005344
Epoch[111] Batch [699]	Speed: 2932.921369 samples/sec	accuracy=0.831251	lr=0.005294
Epoch[111] Batch [749]	Speed: 2984.364020 samples/sec	accuracy=0.831223	lr=0.005244
Epoch[111] Batch [799]	Speed: 3024.660087 samples/sec	accuracy=0.831122	lr=0.005195
Epoch[111] Batch [849]	Speed: 2967.033360 samples/sec	accuracy=0.831103	lr=0.005145
Epoch[111] Batch [899]	Speed: 3061.205704 samples/sec	accuracy=0.831086	lr=0.005096
Epoch[111] Batch [949]	Speed: 3013.675917 samples/sec	accuracy=0.831033	lr=0.005047
Epoch[111] Batch [999]	Speed: 2917.165637 samples/sec	accuracy=0.831052	lr=0.004999
Epoch[111] Batch [1049]	Speed: 2976.222302 samples/sec	accuracy=0.831116	lr=0.004950
Epoch[111] Batch [1099]	Speed: 3001.543201 samples/sec	accuracy=0.831077	lr=0.004902
Epoch[111] Batch [1149]	Speed: 3128.914758 samples/sec	accuracy=0.831145	lr=0.004854
Epoch[111] Batch [1199]	Speed: 2927.046006 samples/sec	accuracy=0.831137	lr=0.004807
Epoch[111] Batch [1249]	Speed: 3084.117926 samples/sec	accuracy=0.831009	lr=0.004759
[Epoch 111] training: accuracy=0.831013
[Epoch 111] speed: 2974 samples/sec	time cost: 440.971328
[Epoch 111] validation: err-top1=0.212771 err-top5=0.056521
Epoch[112] Batch [49]	Speed: 3014.089298 samples/sec	accuracy=0.834980	lr=0.004711
Epoch[112] Batch [99]	Speed: 2972.701846 samples/sec	accuracy=0.835273	lr=0.004664
Epoch[112] Batch [149]	Speed: 2936.453819 samples/sec	accuracy=0.834499	lr=0.004617
Epoch[112] Batch [199]	Speed: 3061.356521 samples/sec	accuracy=0.834312	lr=0.004571
Epoch[112] Batch [249]	Speed: 2925.386476 samples/sec	accuracy=0.834621	lr=0.004524
Epoch[112] Batch [299]	Speed: 2794.539406 samples/sec	accuracy=0.834655	lr=0.004478
Epoch[112] Batch [349]	Speed: 2937.699608 samples/sec	accuracy=0.834637	lr=0.004432
Epoch[112] Batch [399]	Speed: 2979.304681 samples/sec	accuracy=0.834443	lr=0.004387
Epoch[112] Batch [449]	Speed: 3203.858636 samples/sec	accuracy=0.834737	lr=0.004341
Epoch[112] Batch [499]	Speed: 3096.328900 samples/sec	accuracy=0.834619	lr=0.004296
Epoch[112] Batch [549]	Speed: 2834.158208 samples/sec	accuracy=0.834980	lr=0.004251
Epoch[112] Batch [599]	Speed: 3047.518199 samples/sec	accuracy=0.835047	lr=0.004207
Epoch[112] Batch [649]	Speed: 3013.469459 samples/sec	accuracy=0.834904	lr=0.004162
Epoch[112] Batch [699]	Speed: 3188.458355 samples/sec	accuracy=0.834983	lr=0.004118
Epoch[112] Batch [749]	Speed: 3118.939704 samples/sec	accuracy=0.834974	lr=0.004074
Epoch[112] Batch [799]	Speed: 2918.779947 samples/sec	accuracy=0.834994	lr=0.004030
Epoch[112] Batch [849]	Speed: 3149.039504 samples/sec	accuracy=0.834855	lr=0.003987
Epoch[112] Batch [899]	Speed: 2925.312754 samples/sec	accuracy=0.834683	lr=0.003944
Epoch[112] Batch [949]	Speed: 3131.148787 samples/sec	accuracy=0.834569	lr=0.003901
Epoch[112] Batch [999]	Speed: 2979.167336 samples/sec	accuracy=0.834478	lr=0.003858
Epoch[112] Batch [1049]	Speed: 2913.705663 samples/sec	accuracy=0.834469	lr=0.003815
Epoch[112] Batch [1099]	Speed: 2992.780992 samples/sec	accuracy=0.834390	lr=0.003773
Epoch[112] Batch [1149]	Speed: 3053.112069 samples/sec	accuracy=0.834297	lr=0.003731
Epoch[112] Batch [1199]	Speed: 2965.024768 samples/sec	accuracy=0.834333	lr=0.003689
Epoch[112] Batch [1249]	Speed: 3142.904021 samples/sec	accuracy=0.834299	lr=0.003647
[Epoch 112] training: accuracy=0.834278
[Epoch 112] speed: 3006 samples/sec	time cost: 436.610477
[Epoch 112] validation: err-top1=0.211256 err-top5=0.055704
Epoch[113] Batch [49]	Speed: 2963.457548 samples/sec	accuracy=0.839609	lr=0.003605
Epoch[113] Batch [99]	Speed: 2952.792563 samples/sec	accuracy=0.839824	lr=0.003564
Epoch[113] Batch [149]	Speed: 3058.395313 samples/sec	accuracy=0.839792	lr=0.003523
Epoch[113] Batch [199]	Speed: 3072.562237 samples/sec	accuracy=0.838882	lr=0.003482
Epoch[113] Batch [249]	Speed: 2989.878426 samples/sec	accuracy=0.838996	lr=0.003442
Epoch[113] Batch [299]	Speed: 3025.394073 samples/sec	accuracy=0.838874	lr=0.003402
Epoch[113] Batch [349]	Speed: 3044.869248 samples/sec	accuracy=0.838945	lr=0.003362
Epoch[113] Batch [399]	Speed: 2990.542944 samples/sec	accuracy=0.838730	lr=0.003322
Epoch[113] Batch [449]	Speed: 3073.992087 samples/sec	accuracy=0.838507	lr=0.003282
Epoch[113] Batch [499]	Speed: 3004.853176 samples/sec	accuracy=0.838428	lr=0.003243
Epoch[113] Batch [549]	Speed: 2903.579023 samples/sec	accuracy=0.838370	lr=0.003204
Epoch[113] Batch [599]	Speed: 3118.829542 samples/sec	accuracy=0.838340	lr=0.003165
Epoch[113] Batch [649]	Speed: 2978.799054 samples/sec	accuracy=0.838299	lr=0.003127
Epoch[113] Batch [699]	Speed: 3008.401311 samples/sec	accuracy=0.838357	lr=0.003088
Epoch[113] Batch [749]	Speed: 2996.087586 samples/sec	accuracy=0.838316	lr=0.003050
Epoch[113] Batch [799]	Speed: 2971.046880 samples/sec	accuracy=0.838241	lr=0.003012
Epoch[113] Batch [849]	Speed: 2947.302864 samples/sec	accuracy=0.838423	lr=0.002975
Epoch[113] Batch [899]	Speed: 2995.630737 samples/sec	accuracy=0.838372	lr=0.002937
Epoch[113] Batch [949]	Speed: 2974.630328 samples/sec	accuracy=0.838440	lr=0.002900
Epoch[113] Batch [999]	Speed: 3086.698139 samples/sec	accuracy=0.838434	lr=0.002863
Epoch[113] Batch [1049]	Speed: 3009.967586 samples/sec	accuracy=0.838399	lr=0.002826
Epoch[113] Batch [1099]	Speed: 2981.822475 samples/sec	accuracy=0.838393	lr=0.002790
Epoch[113] Batch [1149]	Speed: 2941.938646 samples/sec	accuracy=0.838376	lr=0.002754
Epoch[113] Batch [1199]	Speed: 2968.890954 samples/sec	accuracy=0.838407	lr=0.002718
Epoch[113] Batch [1249]	Speed: 3126.463049 samples/sec	accuracy=0.838392	lr=0.002682
[Epoch 113] training: accuracy=0.838392
[Epoch 113] speed: 3004 samples/sec	time cost: 436.890056
[Epoch 113] validation: err-top1=0.210419 err-top5=0.054887
Epoch[114] Batch [49]	Speed: 2981.639277 samples/sec	accuracy=0.842363	lr=0.002646
Epoch[114] Batch [99]	Speed: 3081.467162 samples/sec	accuracy=0.841455	lr=0.002611
Epoch[114] Batch [149]	Speed: 3009.622735 samples/sec	accuracy=0.841302	lr=0.002575
Epoch[114] Batch [199]	Speed: 3111.476147 samples/sec	accuracy=0.841064	lr=0.002541
Epoch[114] Batch [249]	Speed: 2966.997450 samples/sec	accuracy=0.841148	lr=0.002506
Epoch[114] Batch [299]	Speed: 2978.415702 samples/sec	accuracy=0.841006	lr=0.002472
Epoch[114] Batch [349]	Speed: 3015.121067 samples/sec	accuracy=0.841378	lr=0.002438
Epoch[114] Batch [399]	Speed: 3013.884687 samples/sec	accuracy=0.841255	lr=0.002404
Epoch[114] Batch [449]	Speed: 3054.137807 samples/sec	accuracy=0.841209	lr=0.002370
Epoch[114] Batch [499]	Speed: 2965.271686 samples/sec	accuracy=0.840971	lr=0.002337
Epoch[114] Batch [549]	Speed: 2937.972341 samples/sec	accuracy=0.840998	lr=0.002304
Epoch[114] Batch [599]	Speed: 3040.038279 samples/sec	accuracy=0.840851	lr=0.002271
Epoch[114] Batch [649]	Speed: 3082.416785 samples/sec	accuracy=0.840962	lr=0.002238
Epoch[114] Batch [699]	Speed: 3046.319500 samples/sec	accuracy=0.840784	lr=0.002205
Epoch[114] Batch [749]	Speed: 3017.490489 samples/sec	accuracy=0.840837	lr=0.002173
Epoch[114] Batch [799]	Speed: 2981.190299 samples/sec	accuracy=0.840897	lr=0.002141
Epoch[114] Batch [849]	Speed: 2898.907330 samples/sec	accuracy=0.840979	lr=0.002110
Epoch[114] Batch [899]	Speed: 2914.147948 samples/sec	accuracy=0.840968	lr=0.002078
Epoch[114] Batch [949]	Speed: 3077.265734 samples/sec	accuracy=0.841013	lr=0.002047
Epoch[114] Batch [999]	Speed: 3072.494362 samples/sec	accuracy=0.841028	lr=0.002016
Epoch[114] Batch [1049]	Speed: 2893.337195 samples/sec	accuracy=0.841121	lr=0.001985
Epoch[114] Batch [1099]	Speed: 3049.634133 samples/sec	accuracy=0.841243	lr=0.001954
Epoch[114] Batch [1149]	Speed: 3087.944684 samples/sec	accuracy=0.841151	lr=0.001924
Epoch[114] Batch [1199]	Speed: 2965.176042 samples/sec	accuracy=0.841091	lr=0.001894
Epoch[114] Batch [1249]	Speed: 3102.483157 samples/sec	accuracy=0.841167	lr=0.001864
[Epoch 114] training: accuracy=0.841179
[Epoch 114] speed: 3010 samples/sec	time cost: 435.991748
[Epoch 114] validation: err-top1=0.209921 err-top5=0.054588
Epoch[115] Batch [49]	Speed: 3136.827405 samples/sec	accuracy=0.842852	lr=0.001834
Epoch[115] Batch [99]	Speed: 3057.571612 samples/sec	accuracy=0.843184	lr=0.001804
Epoch[115] Batch [149]	Speed: 3068.241289 samples/sec	accuracy=0.842884	lr=0.001775
Epoch[115] Batch [199]	Speed: 3063.765846 samples/sec	accuracy=0.843096	lr=0.001746
Epoch[115] Batch [249]	Speed: 2981.916794 samples/sec	accuracy=0.843512	lr=0.001718
Epoch[115] Batch [299]	Speed: 3098.367173 samples/sec	accuracy=0.843581	lr=0.001689
Epoch[115] Batch [349]	Speed: 2918.464478 samples/sec	accuracy=0.843577	lr=0.001661
Epoch[115] Batch [399]	Speed: 3026.493951 samples/sec	accuracy=0.843462	lr=0.001633
Epoch[115] Batch [449]	Speed: 2883.705945 samples/sec	accuracy=0.843698	lr=0.001605
Epoch[115] Batch [499]	Speed: 3041.533814 samples/sec	accuracy=0.843744	lr=0.001578
Epoch[115] Batch [549]	Speed: 2989.314609 samples/sec	accuracy=0.843798	lr=0.001551
Epoch[115] Batch [599]	Speed: 3017.138952 samples/sec	accuracy=0.843631	lr=0.001524
Epoch[115] Batch [649]	Speed: 2911.197390 samples/sec	accuracy=0.843625	lr=0.001497
Epoch[115] Batch [699]	Speed: 2915.929953 samples/sec	accuracy=0.843578	lr=0.001470
Epoch[115] Batch [749]	Speed: 2975.546857 samples/sec	accuracy=0.843596	lr=0.001444
Epoch[115] Batch [799]	Speed: 2930.551391 samples/sec	accuracy=0.843629	lr=0.001418
Epoch[115] Batch [849]	Speed: 3113.708811 samples/sec	accuracy=0.843564	lr=0.001392
Epoch[115] Batch [899]	Speed: 3014.499915 samples/sec	accuracy=0.843512	lr=0.001366
Epoch[115] Batch [949]	Speed: 2973.978917 samples/sec	accuracy=0.843454	lr=0.001341
Epoch[115] Batch [999]	Speed: 3059.827220 samples/sec	accuracy=0.843557	lr=0.001316
Epoch[115] Batch [1049]	Speed: 3164.872667 samples/sec	accuracy=0.843532	lr=0.001291
Epoch[115] Batch [1099]	Speed: 3071.027959 samples/sec	accuracy=0.843524	lr=0.001266
Epoch[115] Batch [1149]	Speed: 2949.273089 samples/sec	accuracy=0.843453	lr=0.001242
Epoch[115] Batch [1199]	Speed: 2974.704085 samples/sec	accuracy=0.843451	lr=0.001218
Epoch[115] Batch [1249]	Speed: 3116.352633 samples/sec	accuracy=0.843452	lr=0.001194
[Epoch 115] training: accuracy=0.843464
[Epoch 115] speed: 3014 samples/sec	time cost: 435.068940
[Epoch 115] validation: err-top1=0.208426 err-top5=0.054428
Epoch[116] Batch [49]	Speed: 3093.532241 samples/sec	accuracy=0.845352	lr=0.001170
Epoch[116] Batch [99]	Speed: 2991.276922 samples/sec	accuracy=0.845215	lr=0.001146
Epoch[116] Batch [149]	Speed: 3115.899877 samples/sec	accuracy=0.844674	lr=0.001123
Epoch[116] Batch [199]	Speed: 3113.309720 samples/sec	accuracy=0.844663	lr=0.001100
Epoch[116] Batch [249]	Speed: 2980.523188 samples/sec	accuracy=0.844555	lr=0.001077
Epoch[116] Batch [299]	Speed: 2922.013178 samples/sec	accuracy=0.844430	lr=0.001055
Epoch[116] Batch [349]	Speed: 2932.271479 samples/sec	accuracy=0.844417	lr=0.001032
Epoch[116] Batch [399]	Speed: 3048.906866 samples/sec	accuracy=0.844353	lr=0.001010
Epoch[116] Batch [449]	Speed: 3172.418027 samples/sec	accuracy=0.844301	lr=0.000989
Epoch[116] Batch [499]	Speed: 2925.118226 samples/sec	accuracy=0.844430	lr=0.000967
Epoch[116] Batch [549]	Speed: 2916.173553 samples/sec	accuracy=0.844386	lr=0.000946
Epoch[116] Batch [599]	Speed: 3108.911430 samples/sec	accuracy=0.844458	lr=0.000925
Epoch[116] Batch [649]	Speed: 2865.172862 samples/sec	accuracy=0.844698	lr=0.000904
Epoch[116] Batch [699]	Speed: 2863.260374 samples/sec	accuracy=0.844914	lr=0.000883
Epoch[116] Batch [749]	Speed: 3024.178982 samples/sec	accuracy=0.844926	lr=0.000863
Epoch[116] Batch [799]	Speed: 2974.061126 samples/sec	accuracy=0.844878	lr=0.000843
Epoch[116] Batch [849]	Speed: 2989.167478 samples/sec	accuracy=0.845036	lr=0.000823
Epoch[116] Batch [899]	Speed: 2993.294131 samples/sec	accuracy=0.844946	lr=0.000803
Epoch[116] Batch [949]	Speed: 3111.360832 samples/sec	accuracy=0.844912	lr=0.000784
Epoch[116] Batch [999]	Speed: 2916.615914 samples/sec	accuracy=0.844895	lr=0.000764
Epoch[116] Batch [1049]	Speed: 2898.069039 samples/sec	accuracy=0.844973	lr=0.000745
Epoch[116] Batch [1099]	Speed: 2949.677133 samples/sec	accuracy=0.845103	lr=0.000727
Epoch[116] Batch [1149]	Speed: 3008.925554 samples/sec	accuracy=0.845054	lr=0.000708
Epoch[116] Batch [1199]	Speed: 3007.154391 samples/sec	accuracy=0.845196	lr=0.000690
Epoch[116] Batch [1249]	Speed: 3130.903508 samples/sec	accuracy=0.845197	lr=0.000672
[Epoch 116] training: accuracy=0.845196
[Epoch 116] speed: 2997 samples/sec	time cost: 438.415123
[Epoch 116] validation: err-top1=0.208700 err-top5=0.053914
Epoch[117] Batch [49]	Speed: 3167.075496 samples/sec	accuracy=0.847305	lr=0.000654
Epoch[117] Batch [99]	Speed: 2839.355697 samples/sec	accuracy=0.846055	lr=0.000636
Epoch[117] Batch [149]	Speed: 2956.848027 samples/sec	accuracy=0.846784	lr=0.000619
Epoch[117] Batch [199]	Speed: 3008.023279 samples/sec	accuracy=0.846899	lr=0.000602
Epoch[117] Batch [249]	Speed: 2922.633827 samples/sec	accuracy=0.847406	lr=0.000585
Epoch[117] Batch [299]	Speed: 2955.916859 samples/sec	accuracy=0.847171	lr=0.000569
Epoch[117] Batch [349]	Speed: 2941.740530 samples/sec	accuracy=0.847355	lr=0.000552
Epoch[117] Batch [399]	Speed: 3035.308329 samples/sec	accuracy=0.847397	lr=0.000536
Epoch[117] Batch [449]	Speed: 3025.364451 samples/sec	accuracy=0.847617	lr=0.000520
Epoch[117] Batch [499]	Speed: 2969.290086 samples/sec	accuracy=0.847570	lr=0.000505
Epoch[117] Batch [549]	Speed: 3021.801898 samples/sec	accuracy=0.847473	lr=0.000489
Epoch[117] Batch [599]	Speed: 3016.406423 samples/sec	accuracy=0.847399	lr=0.000474
Epoch[117] Batch [649]	Speed: 2970.032398 samples/sec	accuracy=0.847257	lr=0.000459
Epoch[117] Batch [699]	Speed: 3204.332679 samples/sec	accuracy=0.847133	lr=0.000445
Epoch[117] Batch [749]	Speed: 3031.505135 samples/sec	accuracy=0.847070	lr=0.000430
Epoch[117] Batch [799]	Speed: 3113.538391 samples/sec	accuracy=0.847047	lr=0.000416
Epoch[117] Batch [849]	Speed: 2958.794961 samples/sec	accuracy=0.846937	lr=0.000402
Epoch[117] Batch [899]	Speed: 2931.714248 samples/sec	accuracy=0.846957	lr=0.000388
Epoch[117] Batch [949]	Speed: 2999.886566 samples/sec	accuracy=0.846740	lr=0.000375
Epoch[117] Batch [999]	Speed: 2893.936205 samples/sec	accuracy=0.846648	lr=0.000362
Epoch[117] Batch [1049]	Speed: 3053.194587 samples/sec	accuracy=0.846604	lr=0.000349
Epoch[117] Batch [1099]	Speed: 3122.340571 samples/sec	accuracy=0.846523	lr=0.000336
Epoch[117] Batch [1149]	Speed: 2963.407126 samples/sec	accuracy=0.846583	lr=0.000323
Epoch[117] Batch [1199]	Speed: 3055.475085 samples/sec	accuracy=0.846501	lr=0.000311
Epoch[117] Batch [1249]	Speed: 3043.057728 samples/sec	accuracy=0.846574	lr=0.000299
[Epoch 117] training: accuracy=0.846574
[Epoch 117] speed: 3003 samples/sec	time cost: 436.699541
[Epoch 117] validation: err-top1=0.208665 err-top5=0.053831
Epoch[118] Batch [49]	Speed: 3056.527904 samples/sec	accuracy=0.848359	lr=0.000287
Epoch[118] Batch [99]	Speed: 3078.624803 samples/sec	accuracy=0.847441	lr=0.000275
Epoch[118] Batch [149]	Speed: 2918.743133 samples/sec	accuracy=0.847611	lr=0.000264
Epoch[118] Batch [199]	Speed: 3033.815908 samples/sec	accuracy=0.847563	lr=0.000253
Epoch[118] Batch [249]	Speed: 2951.233174 samples/sec	accuracy=0.847547	lr=0.000242
Epoch[118] Batch [299]	Speed: 3061.243974 samples/sec	accuracy=0.847754	lr=0.000231
Epoch[118] Batch [349]	Speed: 3115.849107 samples/sec	accuracy=0.847790	lr=0.000221
Epoch[118] Batch [399]	Speed: 3065.112270 samples/sec	accuracy=0.847505	lr=0.000211
Epoch[118] Batch [449]	Speed: 2879.106964 samples/sec	accuracy=0.847470	lr=0.000201
Epoch[118] Batch [499]	Speed: 3083.430877 samples/sec	accuracy=0.847393	lr=0.000191
Epoch[118] Batch [549]	Speed: 2973.503957 samples/sec	accuracy=0.847599	lr=0.000182
Epoch[118] Batch [599]	Speed: 2955.913400 samples/sec	accuracy=0.847446	lr=0.000173
Epoch[118] Batch [649]	Speed: 2931.485813 samples/sec	accuracy=0.847577	lr=0.000164
Epoch[118] Batch [699]	Speed: 3078.957705 samples/sec	accuracy=0.847578	lr=0.000155
Epoch[118] Batch [749]	Speed: 2944.530516 samples/sec	accuracy=0.847595	lr=0.000147
Epoch[118] Batch [799]	Speed: 3109.444052 samples/sec	accuracy=0.847633	lr=0.000138
Epoch[118] Batch [849]	Speed: 3076.817122 samples/sec	accuracy=0.847779	lr=0.000130
Epoch[118] Batch [899]	Speed: 2942.464976 samples/sec	accuracy=0.847777	lr=0.000123
Epoch[118] Batch [949]	Speed: 2912.587186 samples/sec	accuracy=0.847878	lr=0.000115
Epoch[118] Batch [999]	Speed: 2911.422358 samples/sec	accuracy=0.847729	lr=0.000108
Epoch[118] Batch [1049]	Speed: 3004.560276 samples/sec	accuracy=0.847777	lr=0.000101
Epoch[118] Batch [1099]	Speed: 2993.800851 samples/sec	accuracy=0.847687	lr=0.000094
Epoch[118] Batch [1149]	Speed: 3079.379388 samples/sec	accuracy=0.847632	lr=0.000087
Epoch[118] Batch [1199]	Speed: 3038.763194 samples/sec	accuracy=0.847518	lr=0.000081
Epoch[118] Batch [1249]	Speed: 3189.141954 samples/sec	accuracy=0.847619	lr=0.000075
[Epoch 118] training: accuracy=0.847611
[Epoch 118] speed: 3010 samples/sec	time cost: 435.642193
[Epoch 118] validation: err-top1=0.208386 err-top5=0.053711
Epoch[119] Batch [49]	Speed: 3007.536837 samples/sec	accuracy=0.845684	lr=0.000069
Epoch[119] Batch [99]	Speed: 3050.598988 samples/sec	accuracy=0.846172	lr=0.000063
Epoch[119] Batch [149]	Speed: 3040.807089 samples/sec	accuracy=0.847064	lr=0.000058
Epoch[119] Batch [199]	Speed: 2883.798032 samples/sec	accuracy=0.847085	lr=0.000053
Epoch[119] Batch [249]	Speed: 2837.949157 samples/sec	accuracy=0.847762	lr=0.000048
Epoch[119] Batch [299]	Speed: 2967.972613 samples/sec	accuracy=0.847406	lr=0.000043
Epoch[119] Batch [349]	Speed: 2865.368636 samples/sec	accuracy=0.847285	lr=0.000039
Epoch[119] Batch [399]	Speed: 2915.926588 samples/sec	accuracy=0.847739	lr=0.000035
Epoch[119] Batch [449]	Speed: 3075.950778 samples/sec	accuracy=0.847706	lr=0.000031
Epoch[119] Batch [499]	Speed: 3044.315748 samples/sec	accuracy=0.847803	lr=0.000027
Epoch[119] Batch [549]	Speed: 2953.576125 samples/sec	accuracy=0.847653	lr=0.000023
Epoch[119] Batch [599]	Speed: 2995.650921 samples/sec	accuracy=0.847879	lr=0.000020
Epoch[119] Batch [649]	Speed: 3097.291324 samples/sec	accuracy=0.847656	lr=0.000017
Epoch[119] Batch [699]	Speed: 3096.885176 samples/sec	accuracy=0.847746	lr=0.000015
Epoch[119] Batch [749]	Speed: 3043.336489 samples/sec	accuracy=0.847730	lr=0.000012
Epoch[119] Batch [799]	Speed: 2973.563165 samples/sec	accuracy=0.847654	lr=0.000010
Epoch[119] Batch [849]	Speed: 2963.036351 samples/sec	accuracy=0.847557	lr=0.000008
Epoch[119] Batch [899]	Speed: 2912.803756 samples/sec	accuracy=0.847457	lr=0.000006
Epoch[119] Batch [949]	Speed: 2859.187705 samples/sec	accuracy=0.847557	lr=0.000004
Epoch[119] Batch [999]	Speed: 3084.424328 samples/sec	accuracy=0.847432	lr=0.000003
Epoch[119] Batch [1049]	Speed: 3099.567017 samples/sec	accuracy=0.847435	lr=0.000002
Epoch[119] Batch [1099]	Speed: 3050.698619 samples/sec	accuracy=0.847384	lr=0.000001
Epoch[119] Batch [1149]	Speed: 3080.638368 samples/sec	accuracy=0.847454	lr=0.000000
Epoch[119] Batch [1199]	Speed: 3044.294429 samples/sec	accuracy=0.847431	lr=0.000000
Epoch[119] Batch [1249]	Speed: 3371.721820 samples/sec	accuracy=0.847439	lr=0.000000
[Epoch 119] training: accuracy=0.847443
[Epoch 119] speed: 3006 samples/sec	time cost: 436.722144
[Epoch 119] validation: err-top1=0.207848 err-top5=0.054030