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='resnet50_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='resnet50_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_resnet50_v1b_best', save_frequency=10, use_gn=True, use_pretrained=False, use_rec=True, use_se=False, warmup_epochs=5, warmup_lr=0.0, wd=0.0001)
Namespace(batch_norm=False, batch_size=128, crop_ratio=0.875, data_dir='~/.mxnet/datasets/imagenet', dtype='float16', input_size=224, label_smoothing=True, last_gamma=True, log_interval=50, logging_file='resnet50_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='imperative', model='resnet50_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_resnet50_v1b_best', save_frequency=10, use_gn=True, use_pretrained=False, use_rec=True, use_se=False, warmup_epochs=5, warmup_lr=0.0, wd=0.0001)
Namespace(batch_norm=False, batch_size=128, crop_ratio=0.875, data_dir='~/.mxnet/datasets/imagenet', dtype='float16', input_size=224, label_smoothing=True, last_gamma=True, log_interval=50, logging_file='resnet50_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='imperative', model='resnet50_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_resnet50_v1b_best', save_frequency=10, use_gn=True, use_pretrained=False, use_rec=True, use_se=False, warmup_epochs=5, warmup_lr=0.0, wd=0.0001)
Namespace(batch_norm=False, batch_size=128, crop_ratio=0.875, data_dir='~/.mxnet/datasets/imagenet', dtype='float32', input_size=224, label_smoothing=True, last_gamma=True, log_interval=50, logging_file='resnet50_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='imperative', model='resnet50_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_resnet50_v1b_best', save_frequency=10, use_gn=True, use_pretrained=False, use_rec=True, use_se=False, warmup_epochs=5, warmup_lr=0.0, wd=0.0001)
Epoch[0] Batch [49]	Speed: 1153.909762 samples/sec	accuracy=0.001074	lr=0.003133
Epoch[0] Batch [99]	Speed: 1561.330237 samples/sec	accuracy=0.001045	lr=0.006331
Epoch[0] Batch [149]	Speed: 1562.328903 samples/sec	accuracy=0.001283	lr=0.009528
Epoch[0] Batch [199]	Speed: 1528.348942 samples/sec	accuracy=0.001660	lr=0.012726
Epoch[0] Batch [249]	Speed: 1481.066565 samples/sec	accuracy=0.002180	lr=0.015923
Epoch[0] Batch [299]	Speed: 1500.828920 samples/sec	accuracy=0.002796	lr=0.019121
Epoch[0] Batch [349]	Speed: 1439.030711 samples/sec	accuracy=0.003398	lr=0.022318
Epoch[0] Batch [399]	Speed: 1511.686824 samples/sec	accuracy=0.004028	lr=0.025516
Epoch[0] Batch [449]	Speed: 1511.376280 samples/sec	accuracy=0.004768	lr=0.028713
Epoch[0] Batch [499]	Speed: 1533.406388 samples/sec	accuracy=0.005582	lr=0.031910
Epoch[0] Batch [549]	Speed: 1543.822531 samples/sec	accuracy=0.006399	lr=0.035108
Epoch[0] Batch [599]	Speed: 1506.942291 samples/sec	accuracy=0.007280	lr=0.038305
Epoch[0] Batch [649]	Speed: 1477.929898 samples/sec	accuracy=0.008110	lr=0.041503
Epoch[0] Batch [699]	Speed: 1531.368585 samples/sec	accuracy=0.009025	lr=0.044700
Epoch[0] Batch [749]	Speed: 1511.202344 samples/sec	accuracy=0.009911	lr=0.047898
Epoch[0] Batch [799]	Speed: 1510.509805 samples/sec	accuracy=0.010670	lr=0.051095
Epoch[0] Batch [849]	Speed: 1470.104385 samples/sec	accuracy=0.011668	lr=0.054293
Epoch[0] Batch [899]	Speed: 1484.963777 samples/sec	accuracy=0.012662	lr=0.057490
Epoch[0] Batch [949]	Speed: 1501.460559 samples/sec	accuracy=0.013672	lr=0.060687
Epoch[0] Batch [999]	Speed: 1529.608367 samples/sec	accuracy=0.014633	lr=0.063885
Epoch[0] Batch [1049]	Speed: 1501.683985 samples/sec	accuracy=0.015656	lr=0.067082
Epoch[0] Batch [1099]	Speed: 1518.848418 samples/sec	accuracy=0.016731	lr=0.070280
Epoch[0] Batch [1149]	Speed: 1535.599623 samples/sec	accuracy=0.017820	lr=0.073477
Epoch[0] Batch [1199]	Speed: 1417.081666 samples/sec	accuracy=0.018829	lr=0.076675
Epoch[0] Batch [1249]	Speed: 1460.516089 samples/sec	accuracy=0.019927	lr=0.079872
[Epoch 0] training: accuracy=0.019957
[Epoch 0] speed: 1485 samples/sec	time cost: 886.523820
[Epoch 0] validation: err-top1=0.939852 err-top5=0.837771
Epoch[1] Batch [49]	Speed: 1529.455241 samples/sec	accuracy=0.050254	lr=0.083133
Epoch[1] Batch [99]	Speed: 1540.255847 samples/sec	accuracy=0.050107	lr=0.086331
Epoch[1] Batch [149]	Speed: 1490.342533 samples/sec	accuracy=0.050781	lr=0.089528
Epoch[1] Batch [199]	Speed: 1539.117393 samples/sec	accuracy=0.051484	lr=0.092726
Epoch[1] Batch [249]	Speed: 1517.715763 samples/sec	accuracy=0.052246	lr=0.095923
Epoch[1] Batch [299]	Speed: 1487.920095 samples/sec	accuracy=0.053047	lr=0.099121
Epoch[1] Batch [349]	Speed: 1542.273879 samples/sec	accuracy=0.053884	lr=0.102318
Epoch[1] Batch [399]	Speed: 1536.938678 samples/sec	accuracy=0.054915	lr=0.105516
Epoch[1] Batch [449]	Speed: 1542.342001 samples/sec	accuracy=0.055697	lr=0.108713
Epoch[1] Batch [499]	Speed: 1537.966401 samples/sec	accuracy=0.056590	lr=0.111910
Epoch[1] Batch [549]	Speed: 1470.871635 samples/sec	accuracy=0.057567	lr=0.115108
Epoch[1] Batch [599]	Speed: 1531.773765 samples/sec	accuracy=0.057956	lr=0.118305
Epoch[1] Batch [649]	Speed: 1459.545957 samples/sec	accuracy=0.058678	lr=0.121503
Epoch[1] Batch [699]	Speed: 1453.354749 samples/sec	accuracy=0.059611	lr=0.124700
Epoch[1] Batch [749]	Speed: 1531.731089 samples/sec	accuracy=0.060486	lr=0.127898
Epoch[1] Batch [799]	Speed: 1485.084543 samples/sec	accuracy=0.061528	lr=0.131095
Epoch[1] Batch [849]	Speed: 1509.689354 samples/sec	accuracy=0.062476	lr=0.134293
Epoch[1] Batch [899]	Speed: 1485.582314 samples/sec	accuracy=0.063465	lr=0.137490
Epoch[1] Batch [949]	Speed: 1529.638286 samples/sec	accuracy=0.064208	lr=0.140687
Epoch[1] Batch [999]	Speed: 1540.211030 samples/sec	accuracy=0.065157	lr=0.143885
Epoch[1] Batch [1049]	Speed: 1502.744970 samples/sec	accuracy=0.065996	lr=0.147082
Epoch[1] Batch [1099]	Speed: 1466.628590 samples/sec	accuracy=0.066770	lr=0.150280
Epoch[1] Batch [1149]	Speed: 1518.473699 samples/sec	accuracy=0.067618	lr=0.153477
Epoch[1] Batch [1199]	Speed: 1536.607590 samples/sec	accuracy=0.068387	lr=0.156675
Epoch[1] Batch [1249]	Speed: 1543.397919 samples/sec	accuracy=0.069493	lr=0.159872
[Epoch 1] training: accuracy=0.069508
[Epoch 1] speed: 1511 samples/sec	time cost: 870.596471
[Epoch 1] validation: err-top1=0.889888 err-top5=0.736109
Epoch[2] Batch [49]	Speed: 1472.719374 samples/sec	accuracy=0.095527	lr=0.163133
Epoch[2] Batch [99]	Speed: 1514.162276 samples/sec	accuracy=0.097910	lr=0.166331
Epoch[2] Batch [149]	Speed: 1543.942571 samples/sec	accuracy=0.099434	lr=0.169528
Epoch[2] Batch [199]	Speed: 1513.382112 samples/sec	accuracy=0.100469	lr=0.172726
Epoch[2] Batch [249]	Speed: 1522.584199 samples/sec	accuracy=0.101352	lr=0.175923
Epoch[2] Batch [299]	Speed: 1511.362664 samples/sec	accuracy=0.102611	lr=0.179121
Epoch[2] Batch [349]	Speed: 1485.344688 samples/sec	accuracy=0.104157	lr=0.182318
Epoch[2] Batch [399]	Speed: 1521.860538 samples/sec	accuracy=0.105022	lr=0.185516
Epoch[2] Batch [449]	Speed: 1489.273204 samples/sec	accuracy=0.106430	lr=0.188713
Epoch[2] Batch [499]	Speed: 1537.083206 samples/sec	accuracy=0.106689	lr=0.191910
Epoch[2] Batch [549]	Speed: 1509.054359 samples/sec	accuracy=0.107931	lr=0.195108
Epoch[2] Batch [599]	Speed: 1507.474306 samples/sec	accuracy=0.108997	lr=0.198305
Epoch[2] Batch [649]	Speed: 1494.430192 samples/sec	accuracy=0.110583	lr=0.201503
Epoch[2] Batch [699]	Speed: 1514.338005 samples/sec	accuracy=0.111634	lr=0.204700
Epoch[2] Batch [749]	Speed: 1539.966871 samples/sec	accuracy=0.111294	lr=0.207898
Epoch[2] Batch [799]	Speed: 1537.225913 samples/sec	accuracy=0.112219	lr=0.211095
Epoch[2] Batch [849]	Speed: 1484.157183 samples/sec	accuracy=0.113395	lr=0.214293
Epoch[2] Batch [899]	Speed: 1519.577453 samples/sec	accuracy=0.114259	lr=0.217490
Epoch[2] Batch [949]	Speed: 1514.432249 samples/sec	accuracy=0.115516	lr=0.220687
Epoch[2] Batch [999]	Speed: 1517.007931 samples/sec	accuracy=0.116687	lr=0.223885
Epoch[2] Batch [1049]	Speed: 1510.279284 samples/sec	accuracy=0.117886	lr=0.227082
Epoch[2] Batch [1099]	Speed: 1488.060490 samples/sec	accuracy=0.119037	lr=0.230280
Epoch[2] Batch [1149]	Speed: 1518.494658 samples/sec	accuracy=0.119966	lr=0.233477
Epoch[2] Batch [1199]	Speed: 1536.719429 samples/sec	accuracy=0.121366	lr=0.236675
Epoch[2] Batch [1249]	Speed: 1546.407668 samples/sec	accuracy=0.122627	lr=0.239872
[Epoch 2] training: accuracy=0.122627
[Epoch 2] speed: 1512 samples/sec	time cost: 869.857690
[Epoch 2] validation: err-top1=0.840123 err-top5=0.652224
Epoch[3] Batch [49]	Speed: 1509.897337 samples/sec	accuracy=0.150176	lr=0.243133
Epoch[3] Batch [99]	Speed: 1509.657939 samples/sec	accuracy=0.152598	lr=0.246331
Epoch[3] Batch [149]	Speed: 1536.789425 samples/sec	accuracy=0.150020	lr=0.249528
Epoch[3] Batch [199]	Speed: 1535.727888 samples/sec	accuracy=0.151914	lr=0.252726
Epoch[3] Batch [249]	Speed: 1543.590851 samples/sec	accuracy=0.153297	lr=0.255923
Epoch[3] Batch [299]	Speed: 1492.618982 samples/sec	accuracy=0.156064	lr=0.259121
Epoch[3] Batch [349]	Speed: 1466.943011 samples/sec	accuracy=0.155896	lr=0.262318
Epoch[3] Batch [399]	Speed: 1509.203861 samples/sec	accuracy=0.157600	lr=0.265516
Epoch[3] Batch [449]	Speed: 1539.092960 samples/sec	accuracy=0.159091	lr=0.268713
Epoch[3] Batch [499]	Speed: 1537.952402 samples/sec	accuracy=0.160156	lr=0.271910
Epoch[3] Batch [549]	Speed: 1546.618518 samples/sec	accuracy=0.161296	lr=0.275108
Epoch[3] Batch [599]	Speed: 1488.348776 samples/sec	accuracy=0.162806	lr=0.278305
Epoch[3] Batch [649]	Speed: 1512.922850 samples/sec	accuracy=0.164056	lr=0.281503
Epoch[3] Batch [699]	Speed: 1534.390119 samples/sec	accuracy=0.164727	lr=0.284700
Epoch[3] Batch [749]	Speed: 1475.608881 samples/sec	accuracy=0.166055	lr=0.287898
Epoch[3] Batch [799]	Speed: 1539.636520 samples/sec	accuracy=0.166984	lr=0.291095
Epoch[3] Batch [849]	Speed: 1545.608192 samples/sec	accuracy=0.167886	lr=0.294293
Epoch[3] Batch [899]	Speed: 1518.178316 samples/sec	accuracy=0.169003	lr=0.297490
Epoch[3] Batch [949]	Speed: 1511.987798 samples/sec	accuracy=0.169374	lr=0.300687
Epoch[3] Batch [999]	Speed: 1537.562375 samples/sec	accuracy=0.169859	lr=0.303885
Epoch[3] Batch [1049]	Speed: 1514.809389 samples/sec	accuracy=0.171022	lr=0.307082
Epoch[3] Batch [1099]	Speed: 1535.615117 samples/sec	accuracy=0.172105	lr=0.310280
Epoch[3] Batch [1149]	Speed: 1484.301290 samples/sec	accuracy=0.173197	lr=0.313477
Epoch[3] Batch [1199]	Speed: 1538.858088 samples/sec	accuracy=0.174462	lr=0.316675
Epoch[3] Batch [1249]	Speed: 1518.816546 samples/sec	accuracy=0.175110	lr=0.319872
[Epoch 3] training: accuracy=0.175111
[Epoch 3] speed: 1517 samples/sec	time cost: 866.770905
[Epoch 3] validation: err-top1=0.780712 err-top5=0.563098
Epoch[4] Batch [49]	Speed: 1525.690488 samples/sec	accuracy=0.204219	lr=0.323133
Epoch[4] Batch [99]	Speed: 1480.999070 samples/sec	accuracy=0.204238	lr=0.326331
Epoch[4] Batch [149]	Speed: 1510.249268 samples/sec	accuracy=0.208086	lr=0.329528
Epoch[4] Batch [199]	Speed: 1509.732519 samples/sec	accuracy=0.210869	lr=0.332726
Epoch[4] Batch [249]	Speed: 1535.953006 samples/sec	accuracy=0.211266	lr=0.335923
Epoch[4] Batch [299]	Speed: 1539.718761 samples/sec	accuracy=0.211836	lr=0.339121
Epoch[4] Batch [349]	Speed: 1480.889170 samples/sec	accuracy=0.212659	lr=0.342318
Epoch[4] Batch [399]	Speed: 1499.202630 samples/sec	accuracy=0.213655	lr=0.345516
Epoch[4] Batch [449]	Speed: 1481.395894 samples/sec	accuracy=0.215230	lr=0.348713
Epoch[4] Batch [499]	Speed: 1543.543321 samples/sec	accuracy=0.215332	lr=0.351910
Epoch[4] Batch [549]	Speed: 1482.046564 samples/sec	accuracy=0.216094	lr=0.355108
Epoch[4] Batch [599]	Speed: 1501.515495 samples/sec	accuracy=0.217378	lr=0.358305
Epoch[4] Batch [649]	Speed: 1463.185222 samples/sec	accuracy=0.217015	lr=0.361503
Epoch[4] Batch [699]	Speed: 1506.141829 samples/sec	accuracy=0.218492	lr=0.364700
Epoch[4] Batch [749]	Speed: 1533.243425 samples/sec	accuracy=0.218947	lr=0.367898
Epoch[4] Batch [799]	Speed: 1475.522843 samples/sec	accuracy=0.219435	lr=0.371095
Epoch[4] Batch [849]	Speed: 1542.155471 samples/sec	accuracy=0.220263	lr=0.374293
Epoch[4] Batch [899]	Speed: 1498.227476 samples/sec	accuracy=0.221687	lr=0.377490
Epoch[4] Batch [949]	Speed: 1485.735138 samples/sec	accuracy=0.223208	lr=0.380687
Epoch[4] Batch [999]	Speed: 1539.767712 samples/sec	accuracy=0.224021	lr=0.383885
Epoch[4] Batch [1049]	Speed: 1485.046802 samples/sec	accuracy=0.225025	lr=0.387082
Epoch[4] Batch [1099]	Speed: 1536.336644 samples/sec	accuracy=0.226372	lr=0.390280
Epoch[4] Batch [1149]	Speed: 1536.208509 samples/sec	accuracy=0.224606	lr=0.393477
Epoch[4] Batch [1199]	Speed: 1540.295409 samples/sec	accuracy=0.225032	lr=0.396675
Epoch[4] Batch [1249]	Speed: 1551.302012 samples/sec	accuracy=0.225995	lr=0.399872
[Epoch 4] training: accuracy=0.225972
[Epoch 4] speed: 1509 samples/sec	time cost: 871.370984
[Epoch 4] validation: err-top1=0.760643 err-top5=0.536093
Epoch[5] Batch [49]	Speed: 1427.565476 samples/sec	accuracy=0.261094	lr=0.400000
Epoch[5] Batch [99]	Speed: 1534.112348 samples/sec	accuracy=0.258672	lr=0.400000
Epoch[5] Batch [149]	Speed: 1506.193633 samples/sec	accuracy=0.261797	lr=0.399999
Epoch[5] Batch [199]	Speed: 1511.442784 samples/sec	accuracy=0.261514	lr=0.399998
Epoch[5] Batch [249]	Speed: 1526.907785 samples/sec	accuracy=0.263121	lr=0.399997
Epoch[5] Batch [299]	Speed: 1480.860382 samples/sec	accuracy=0.265570	lr=0.399996
Epoch[5] Batch [349]	Speed: 1506.216642 samples/sec	accuracy=0.267026	lr=0.399994
Epoch[5] Batch [399]	Speed: 1487.227148 samples/sec	accuracy=0.267080	lr=0.399992
Epoch[5] Batch [449]	Speed: 1517.785338 samples/sec	accuracy=0.268952	lr=0.399990
Epoch[5] Batch [499]	Speed: 1502.216896 samples/sec	accuracy=0.269551	lr=0.399988
Epoch[5] Batch [549]	Speed: 1516.567564 samples/sec	accuracy=0.269194	lr=0.399986
Epoch[5] Batch [599]	Speed: 1523.025864 samples/sec	accuracy=0.270221	lr=0.399983
Epoch[5] Batch [649]	Speed: 1535.535993 samples/sec	accuracy=0.271774	lr=0.399980
Epoch[5] Batch [699]	Speed: 1466.198243 samples/sec	accuracy=0.271120	lr=0.399977
Epoch[5] Batch [749]	Speed: 1489.293571 samples/sec	accuracy=0.272743	lr=0.399973
Epoch[5] Batch [799]	Speed: 1516.025345 samples/sec	accuracy=0.274078	lr=0.399970
Epoch[5] Batch [849]	Speed: 1509.924355 samples/sec	accuracy=0.275534	lr=0.399966
Epoch[5] Batch [899]	Speed: 1500.408357 samples/sec	accuracy=0.276729	lr=0.399961
Epoch[5] Batch [949]	Speed: 1506.479382 samples/sec	accuracy=0.277733	lr=0.399957
Epoch[5] Batch [999]	Speed: 1537.520004 samples/sec	accuracy=0.279190	lr=0.399952
Epoch[5] Batch [1049]	Speed: 1481.871306 samples/sec	accuracy=0.278429	lr=0.399948
Epoch[5] Batch [1099]	Speed: 1493.566340 samples/sec	accuracy=0.279277	lr=0.399942
Epoch[5] Batch [1149]	Speed: 1513.263941 samples/sec	accuracy=0.280599	lr=0.399937
Epoch[5] Batch [1199]	Speed: 1542.092670 samples/sec	accuracy=0.281629	lr=0.399931
Epoch[5] Batch [1249]	Speed: 1543.506676 samples/sec	accuracy=0.282705	lr=0.399926
[Epoch 5] training: accuracy=0.282750
[Epoch 5] speed: 1505 samples/sec	time cost: 873.358479
[Epoch 5] validation: err-top1=0.629110 err-top5=0.374410
Epoch[6] Batch [49]	Speed: 1505.279096 samples/sec	accuracy=0.317852	lr=0.399919
Epoch[6] Batch [99]	Speed: 1513.409192 samples/sec	accuracy=0.317627	lr=0.399913
Epoch[6] Batch [149]	Speed: 1488.028691 samples/sec	accuracy=0.319642	lr=0.399907
Epoch[6] Batch [199]	Speed: 1538.537703 samples/sec	accuracy=0.320352	lr=0.399900
Epoch[6] Batch [249]	Speed: 1514.359287 samples/sec	accuracy=0.321406	lr=0.399893
Epoch[6] Batch [299]	Speed: 1520.481709 samples/sec	accuracy=0.320290	lr=0.399885
Epoch[6] Batch [349]	Speed: 1497.563035 samples/sec	accuracy=0.320123	lr=0.399878
Epoch[6] Batch [399]	Speed: 1504.373173 samples/sec	accuracy=0.321729	lr=0.399870
Epoch[6] Batch [449]	Speed: 1507.672312 samples/sec	accuracy=0.322602	lr=0.399862
Epoch[6] Batch [499]	Speed: 1506.169157 samples/sec	accuracy=0.323613	lr=0.399854
Epoch[6] Batch [549]	Speed: 1489.013106 samples/sec	accuracy=0.324462	lr=0.399846
Epoch[6] Batch [599]	Speed: 1478.132254 samples/sec	accuracy=0.325036	lr=0.399837
Epoch[6] Batch [649]	Speed: 1450.259364 samples/sec	accuracy=0.326453	lr=0.399828
Epoch[6] Batch [699]	Speed: 1496.671891 samples/sec	accuracy=0.326744	lr=0.399819
Epoch[6] Batch [749]	Speed: 1529.793246 samples/sec	accuracy=0.325792	lr=0.399809
Epoch[6] Batch [799]	Speed: 1540.350970 samples/sec	accuracy=0.326580	lr=0.399800
Epoch[6] Batch [849]	Speed: 1512.169762 samples/sec	accuracy=0.327935	lr=0.399790
Epoch[6] Batch [899]	Speed: 1472.090943 samples/sec	accuracy=0.328765	lr=0.399780
Epoch[6] Batch [949]	Speed: 1533.262199 samples/sec	accuracy=0.329706	lr=0.399769
Epoch[6] Batch [999]	Speed: 1482.740557 samples/sec	accuracy=0.330414	lr=0.399759
Epoch[6] Batch [1049]	Speed: 1525.934380 samples/sec	accuracy=0.330766	lr=0.399748
Epoch[6] Batch [1099]	Speed: 1485.378561 samples/sec	accuracy=0.331578	lr=0.399737
Epoch[6] Batch [1149]	Speed: 1508.582193 samples/sec	accuracy=0.332315	lr=0.399725
Epoch[6] Batch [1199]	Speed: 1536.570197 samples/sec	accuracy=0.332875	lr=0.399714
Epoch[6] Batch [1249]	Speed: 1550.689369 samples/sec	accuracy=0.333445	lr=0.399702
[Epoch 6] training: accuracy=0.333475
[Epoch 6] speed: 1505 samples/sec	time cost: 873.102355
[Epoch 6] validation: err-top1=0.589904 err-top5=0.336177
Epoch[7] Batch [49]	Speed: 1458.493504 samples/sec	accuracy=0.367324	lr=0.399690
Epoch[7] Batch [99]	Speed: 1491.420743 samples/sec	accuracy=0.364561	lr=0.399677
Epoch[7] Batch [149]	Speed: 1503.487654 samples/sec	accuracy=0.362240	lr=0.399665
Epoch[7] Batch [199]	Speed: 1542.426392 samples/sec	accuracy=0.362798	lr=0.399652
Epoch[7] Batch [249]	Speed: 1484.233459 samples/sec	accuracy=0.362988	lr=0.399639
Epoch[7] Batch [299]	Speed: 1492.676054 samples/sec	accuracy=0.361647	lr=0.399626
Epoch[7] Batch [349]	Speed: 1530.317301 samples/sec	accuracy=0.363016	lr=0.399613
Epoch[7] Batch [399]	Speed: 1483.963439 samples/sec	accuracy=0.362646	lr=0.399599
Epoch[7] Batch [449]	Speed: 1500.501610 samples/sec	accuracy=0.362960	lr=0.399585
Epoch[7] Batch [499]	Speed: 1508.842049 samples/sec	accuracy=0.362910	lr=0.399571
Epoch[7] Batch [549]	Speed: 1478.512518 samples/sec	accuracy=0.363757	lr=0.399556
Epoch[7] Batch [599]	Speed: 1482.733534 samples/sec	accuracy=0.363880	lr=0.399542
Epoch[7] Batch [649]	Speed: 1488.251345 samples/sec	accuracy=0.364546	lr=0.399527
Epoch[7] Batch [699]	Speed: 1478.689782 samples/sec	accuracy=0.364586	lr=0.399512
Epoch[7] Batch [749]	Speed: 1531.539494 samples/sec	accuracy=0.365193	lr=0.399496
Epoch[7] Batch [799]	Speed: 1508.354315 samples/sec	accuracy=0.365800	lr=0.399481
Epoch[7] Batch [849]	Speed: 1513.962434 samples/sec	accuracy=0.365885	lr=0.399465
Epoch[7] Batch [899]	Speed: 1545.124196 samples/sec	accuracy=0.366533	lr=0.399449
Epoch[7] Batch [949]	Speed: 1487.331141 samples/sec	accuracy=0.367017	lr=0.399432
Epoch[7] Batch [999]	Speed: 1490.137513 samples/sec	accuracy=0.367787	lr=0.399416
Epoch[7] Batch [1049]	Speed: 1507.757355 samples/sec	accuracy=0.367817	lr=0.399399
Epoch[7] Batch [1099]	Speed: 1480.179932 samples/sec	accuracy=0.368440	lr=0.399382
Epoch[7] Batch [1149]	Speed: 1517.700843 samples/sec	accuracy=0.368840	lr=0.399365
Epoch[7] Batch [1199]	Speed: 1534.374727 samples/sec	accuracy=0.368298	lr=0.399347
Epoch[7] Batch [1249]	Speed: 1546.996916 samples/sec	accuracy=0.368830	lr=0.399329
[Epoch 7] training: accuracy=0.368672
[Epoch 7] speed: 1502 samples/sec	time cost: 876.169193
[Epoch 7] validation: err-top1=0.668427 err-top5=0.426220
Epoch[8] Batch [49]	Speed: 1501.303192 samples/sec	accuracy=0.360059	lr=0.399311
Epoch[8] Batch [99]	Speed: 1470.825183 samples/sec	accuracy=0.376875	lr=0.399293
Epoch[8] Batch [149]	Speed: 1513.150415 samples/sec	accuracy=0.381094	lr=0.399274
Epoch[8] Batch [199]	Speed: 1536.840753 samples/sec	accuracy=0.384785	lr=0.399256
Epoch[8] Batch [249]	Speed: 1499.498506 samples/sec	accuracy=0.386605	lr=0.399237
Epoch[8] Batch [299]	Speed: 1533.556659 samples/sec	accuracy=0.387633	lr=0.399218
Epoch[8] Batch [349]	Speed: 1544.404779 samples/sec	accuracy=0.388075	lr=0.399198
Epoch[8] Batch [399]	Speed: 1505.691267 samples/sec	accuracy=0.388538	lr=0.399179
Epoch[8] Batch [449]	Speed: 1548.729572 samples/sec	accuracy=0.389080	lr=0.399159
Epoch[8] Batch [499]	Speed: 1503.045223 samples/sec	accuracy=0.389609	lr=0.399138
Epoch[8] Batch [549]	Speed: 1411.062494 samples/sec	accuracy=0.389501	lr=0.399118
Epoch[8] Batch [599]	Speed: 1470.333838 samples/sec	accuracy=0.388968	lr=0.399098
Epoch[8] Batch [649]	Speed: 1508.259861 samples/sec	accuracy=0.389138	lr=0.399077
Epoch[8] Batch [699]	Speed: 1463.206866 samples/sec	accuracy=0.390061	lr=0.399056
Epoch[8] Batch [749]	Speed: 1504.417078 samples/sec	accuracy=0.390370	lr=0.399034
Epoch[8] Batch [799]	Speed: 1541.278389 samples/sec	accuracy=0.391154	lr=0.399013
Epoch[8] Batch [849]	Speed: 1510.469443 samples/sec	accuracy=0.391482	lr=0.398991
Epoch[8] Batch [899]	Speed: 1510.198225 samples/sec	accuracy=0.391921	lr=0.398969
Epoch[8] Batch [949]	Speed: 1522.604009 samples/sec	accuracy=0.392512	lr=0.398947
Epoch[8] Batch [999]	Speed: 1539.033287 samples/sec	accuracy=0.392651	lr=0.398924
Epoch[8] Batch [1049]	Speed: 1469.314391 samples/sec	accuracy=0.393056	lr=0.398901
Epoch[8] Batch [1099]	Speed: 1489.763328 samples/sec	accuracy=0.393397	lr=0.398878
Epoch[8] Batch [1149]	Speed: 1498.260360 samples/sec	accuracy=0.393865	lr=0.398855
Epoch[8] Batch [1199]	Speed: 1356.949173 samples/sec	accuracy=0.394335	lr=0.398832
Epoch[8] Batch [1249]	Speed: 1531.587926 samples/sec	accuracy=0.394666	lr=0.398808
[Epoch 8] training: accuracy=0.394696
[Epoch 8] speed: 1496 samples/sec	time cost: 878.464344
[Epoch 8] validation: err-top1=0.527563 err-top5=0.271185
Epoch[9] Batch [49]	Speed: 1467.501294 samples/sec	accuracy=0.420371	lr=0.398784
Epoch[9] Batch [99]	Speed: 1479.090045 samples/sec	accuracy=0.409180	lr=0.398760
Epoch[9] Batch [149]	Speed: 1495.689087 samples/sec	accuracy=0.408613	lr=0.398735
Epoch[9] Batch [199]	Speed: 1495.176307 samples/sec	accuracy=0.411108	lr=0.398710
Epoch[9] Batch [249]	Speed: 1537.043985 samples/sec	accuracy=0.411762	lr=0.398686
Epoch[9] Batch [299]	Speed: 1507.048182 samples/sec	accuracy=0.412301	lr=0.398660
Epoch[9] Batch [349]	Speed: 1533.624462 samples/sec	accuracy=0.413278	lr=0.398635
Epoch[9] Batch [399]	Speed: 1510.796503 samples/sec	accuracy=0.411353	lr=0.398610
Epoch[9] Batch [449]	Speed: 1527.601460 samples/sec	accuracy=0.411944	lr=0.398584
Epoch[9] Batch [499]	Speed: 1467.823665 samples/sec	accuracy=0.411455	lr=0.398558
Epoch[9] Batch [549]	Speed: 1475.483579 samples/sec	accuracy=0.410563	lr=0.398531
Epoch[9] Batch [599]	Speed: 1541.115209 samples/sec	accuracy=0.411527	lr=0.398505
Epoch[9] Batch [649]	Speed: 1542.407403 samples/sec	accuracy=0.409195	lr=0.398478
Epoch[9] Batch [699]	Speed: 1538.747018 samples/sec	accuracy=0.409739	lr=0.398451
Epoch[9] Batch [749]	Speed: 1463.819849 samples/sec	accuracy=0.410590	lr=0.398424
Epoch[9] Batch [799]	Speed: 1534.488949 samples/sec	accuracy=0.411345	lr=0.398396
Epoch[9] Batch [849]	Speed: 1490.944868 samples/sec	accuracy=0.410834	lr=0.398369
Epoch[9] Batch [899]	Speed: 1493.128191 samples/sec	accuracy=0.411399	lr=0.398341
Epoch[9] Batch [949]	Speed: 1502.534747 samples/sec	accuracy=0.411993	lr=0.398312
Epoch[9] Batch [999]	Speed: 1539.329691 samples/sec	accuracy=0.412378	lr=0.398284
Epoch[9] Batch [1049]	Speed: 1515.853430 samples/sec	accuracy=0.412840	lr=0.398255
Epoch[9] Batch [1099]	Speed: 1522.463518 samples/sec	accuracy=0.413433	lr=0.398226
Epoch[9] Batch [1149]	Speed: 1525.135271 samples/sec	accuracy=0.413871	lr=0.398197
Epoch[9] Batch [1199]	Speed: 1500.479258 samples/sec	accuracy=0.414263	lr=0.398168
Epoch[9] Batch [1249]	Speed: 1500.410905 samples/sec	accuracy=0.414692	lr=0.398138
[Epoch 9] training: accuracy=0.414696
[Epoch 9] speed: 1506 samples/sec	time cost: 872.895873
[Epoch 9] validation: err-top1=0.525869 err-top5=0.270627
Epoch[10] Batch [49]	Speed: 1513.738423 samples/sec	accuracy=0.433457	lr=0.398108
Epoch[10] Batch [99]	Speed: 1523.966208 samples/sec	accuracy=0.405879	lr=0.398078
Epoch[10] Batch [149]	Speed: 1486.498616 samples/sec	accuracy=0.414622	lr=0.398048
Epoch[10] Batch [199]	Speed: 1511.086841 samples/sec	accuracy=0.417861	lr=0.398017
Epoch[10] Batch [249]	Speed: 1538.802546 samples/sec	accuracy=0.420551	lr=0.397986
Epoch[10] Batch [299]	Speed: 1466.347274 samples/sec	accuracy=0.419261	lr=0.397955
Epoch[10] Batch [349]	Speed: 1526.441461 samples/sec	accuracy=0.421928	lr=0.397924
Epoch[10] Batch [399]	Speed: 1522.954351 samples/sec	accuracy=0.423606	lr=0.397892
Epoch[10] Batch [449]	Speed: 1512.891780 samples/sec	accuracy=0.424607	lr=0.397861
Epoch[10] Batch [499]	Speed: 1521.330253 samples/sec	accuracy=0.425270	lr=0.397829
Epoch[10] Batch [549]	Speed: 1539.304192 samples/sec	accuracy=0.426058	lr=0.397796
Epoch[10] Batch [599]	Speed: 1487.158720 samples/sec	accuracy=0.426333	lr=0.397764
Epoch[10] Batch [649]	Speed: 1533.664921 samples/sec	accuracy=0.427228	lr=0.397731
Epoch[10] Batch [699]	Speed: 1539.995761 samples/sec	accuracy=0.427741	lr=0.397698
Epoch[10] Batch [749]	Speed: 1542.945736 samples/sec	accuracy=0.426589	lr=0.397665
Epoch[10] Batch [799]	Speed: 1509.793699 samples/sec	accuracy=0.424790	lr=0.397632
Epoch[10] Batch [849]	Speed: 1540.797045 samples/sec	accuracy=0.425361	lr=0.397598
Epoch[10] Batch [899]	Speed: 1510.644676 samples/sec	accuracy=0.425897	lr=0.397564
Epoch[10] Batch [949]	Speed: 1457.951317 samples/sec	accuracy=0.426575	lr=0.397530
Epoch[10] Batch [999]	Speed: 1466.635992 samples/sec	accuracy=0.426504	lr=0.397496
Epoch[10] Batch [1049]	Speed: 1485.581194 samples/sec	accuracy=0.427153	lr=0.397461
Epoch[10] Batch [1099]	Speed: 1535.397245 samples/sec	accuracy=0.427675	lr=0.397427
Epoch[10] Batch [1149]	Speed: 1523.301417 samples/sec	accuracy=0.428163	lr=0.397392
Epoch[10] Batch [1199]	Speed: 1483.826994 samples/sec	accuracy=0.428515	lr=0.397356
Epoch[10] Batch [1249]	Speed: 1550.800098 samples/sec	accuracy=0.429116	lr=0.397321
[Epoch 10] training: accuracy=0.429117
[Epoch 10] speed: 1511 samples/sec	time cost: 869.737707
[Epoch 10] validation: err-top1=0.495436 err-top5=0.247907
Epoch[11] Batch [49]	Speed: 1437.226359 samples/sec	accuracy=0.452402	lr=0.397284
Epoch[11] Batch [99]	Speed: 1467.766199 samples/sec	accuracy=0.450732	lr=0.397248
Epoch[11] Batch [149]	Speed: 1514.480706 samples/sec	accuracy=0.447539	lr=0.397212
Epoch[11] Batch [199]	Speed: 1536.459710 samples/sec	accuracy=0.445659	lr=0.397176
Epoch[11] Batch [249]	Speed: 1473.648004 samples/sec	accuracy=0.445461	lr=0.397139
Epoch[11] Batch [299]	Speed: 1504.254107 samples/sec	accuracy=0.446146	lr=0.397102
Epoch[11] Batch [349]	Speed: 1530.590132 samples/sec	accuracy=0.446423	lr=0.397065
Epoch[11] Batch [399]	Speed: 1468.201192 samples/sec	accuracy=0.446758	lr=0.397028
Epoch[11] Batch [449]	Speed: 1498.341198 samples/sec	accuracy=0.446246	lr=0.396990
Epoch[11] Batch [499]	Speed: 1539.176621 samples/sec	accuracy=0.445059	lr=0.396952
Epoch[11] Batch [549]	Speed: 1470.272180 samples/sec	accuracy=0.445504	lr=0.396914
Epoch[11] Batch [599]	Speed: 1475.131601 samples/sec	accuracy=0.445070	lr=0.396876
Epoch[11] Batch [649]	Speed: 1519.169908 samples/sec	accuracy=0.445258	lr=0.396837
Epoch[11] Batch [699]	Speed: 1546.864768 samples/sec	accuracy=0.445399	lr=0.396798
Epoch[11] Batch [749]	Speed: 1494.805310 samples/sec	accuracy=0.445974	lr=0.396759
Epoch[11] Batch [799]	Speed: 1533.022800 samples/sec	accuracy=0.445753	lr=0.396720
Epoch[11] Batch [849]	Speed: 1475.871052 samples/sec	accuracy=0.445796	lr=0.396680
Epoch[11] Batch [899]	Speed: 1540.725876 samples/sec	accuracy=0.446174	lr=0.396641
Epoch[11] Batch [949]	Speed: 1496.700754 samples/sec	accuracy=0.446108	lr=0.396601
Epoch[11] Batch [999]	Speed: 1474.084436 samples/sec	accuracy=0.445789	lr=0.396561
Epoch[11] Batch [1049]	Speed: 1519.856751 samples/sec	accuracy=0.446075	lr=0.396520
Epoch[11] Batch [1099]	Speed: 1485.847229 samples/sec	accuracy=0.445141	lr=0.396479
Epoch[11] Batch [1149]	Speed: 1547.573364 samples/sec	accuracy=0.445296	lr=0.396439
Epoch[11] Batch [1199]	Speed: 1510.834066 samples/sec	accuracy=0.445462	lr=0.396397
Epoch[11] Batch [1249]	Speed: 1547.218038 samples/sec	accuracy=0.445517	lr=0.396356
[Epoch 11] training: accuracy=0.445512
[Epoch 11] speed: 1502 samples/sec	time cost: 874.966255
[Epoch 11] validation: err-top1=0.477051 err-top5=0.231567
Epoch[12] Batch [49]	Speed: 1471.431280 samples/sec	accuracy=0.466230	lr=0.396314
Epoch[12] Batch [99]	Speed: 1510.667386 samples/sec	accuracy=0.406348	lr=0.396272
Epoch[12] Batch [149]	Speed: 1479.438829 samples/sec	accuracy=0.413926	lr=0.396230
Epoch[12] Batch [199]	Speed: 1497.538556 samples/sec	accuracy=0.423145	lr=0.396187
Epoch[12] Batch [249]	Speed: 1528.816378 samples/sec	accuracy=0.429238	lr=0.396145
Epoch[12] Batch [299]	Speed: 1495.038199 samples/sec	accuracy=0.433695	lr=0.396102
Epoch[12] Batch [349]	Speed: 1465.401710 samples/sec	accuracy=0.435882	lr=0.396059
Epoch[12] Batch [399]	Speed: 1533.420041 samples/sec	accuracy=0.438564	lr=0.396016
Epoch[12] Batch [449]	Speed: 1550.057519 samples/sec	accuracy=0.439674	lr=0.395972
Epoch[12] Batch [499]	Speed: 1495.785921 samples/sec	accuracy=0.442129	lr=0.395928
Epoch[12] Batch [549]	Speed: 1488.110346 samples/sec	accuracy=0.443505	lr=0.395885
Epoch[12] Batch [599]	Speed: 1545.162607 samples/sec	accuracy=0.444640	lr=0.395840
Epoch[12] Batch [649]	Speed: 1504.878129 samples/sec	accuracy=0.445736	lr=0.395796
Epoch[12] Batch [699]	Speed: 1471.437974 samples/sec	accuracy=0.446805	lr=0.395751
Epoch[12] Batch [749]	Speed: 1481.925596 samples/sec	accuracy=0.447952	lr=0.395706
Epoch[12] Batch [799]	Speed: 1536.291713 samples/sec	accuracy=0.448800	lr=0.395661
Epoch[12] Batch [849]	Speed: 1536.780341 samples/sec	accuracy=0.449130	lr=0.395616
Epoch[12] Batch [899]	Speed: 1498.671530 samples/sec	accuracy=0.449308	lr=0.395570
Epoch[12] Batch [949]	Speed: 1529.728626 samples/sec	accuracy=0.449733	lr=0.395524
Epoch[12] Batch [999]	Speed: 1490.207684 samples/sec	accuracy=0.450161	lr=0.395478
Epoch[12] Batch [1049]	Speed: 1523.867529 samples/sec	accuracy=0.450494	lr=0.395432
Epoch[12] Batch [1099]	Speed: 1497.340479 samples/sec	accuracy=0.450710	lr=0.395386
Epoch[12] Batch [1149]	Speed: 1513.737676 samples/sec	accuracy=0.450954	lr=0.395339
Epoch[12] Batch [1199]	Speed: 1515.381973 samples/sec	accuracy=0.451411	lr=0.395292
Epoch[12] Batch [1249]	Speed: 1520.128103 samples/sec	accuracy=0.451827	lr=0.395245
[Epoch 12] training: accuracy=0.451834
[Epoch 12] speed: 1505 samples/sec	time cost: 873.402505
[Epoch 12] validation: err-top1=0.477918 err-top5=0.226543
Epoch[13] Batch [49]	Speed: 1530.454828 samples/sec	accuracy=0.467070	lr=0.395196
Epoch[13] Batch [99]	Speed: 1490.244075 samples/sec	accuracy=0.465625	lr=0.395149
Epoch[13] Batch [149]	Speed: 1519.995513 samples/sec	accuracy=0.465671	lr=0.395101
Epoch[13] Batch [199]	Speed: 1546.004482 samples/sec	accuracy=0.466519	lr=0.395052
Epoch[13] Batch [249]	Speed: 1512.084018 samples/sec	accuracy=0.462570	lr=0.395004
Epoch[13] Batch [299]	Speed: 1493.675658 samples/sec	accuracy=0.463841	lr=0.394955
Epoch[13] Batch [349]	Speed: 1480.817178 samples/sec	accuracy=0.464216	lr=0.394907
Epoch[13] Batch [399]	Speed: 1509.122059 samples/sec	accuracy=0.464072	lr=0.394858
Epoch[13] Batch [449]	Speed: 1528.632551 samples/sec	accuracy=0.464360	lr=0.394808
Epoch[13] Batch [499]	Speed: 1512.845056 samples/sec	accuracy=0.464461	lr=0.394759
Epoch[13] Batch [549]	Speed: 1519.441240 samples/sec	accuracy=0.464542	lr=0.394709
Epoch[13] Batch [599]	Speed: 1506.684569 samples/sec	accuracy=0.464421	lr=0.394659
Epoch[13] Batch [649]	Speed: 1543.701389 samples/sec	accuracy=0.464506	lr=0.394609
Epoch[13] Batch [699]	Speed: 1516.299859 samples/sec	accuracy=0.464844	lr=0.394558
Epoch[13] Batch [749]	Speed: 1519.773510 samples/sec	accuracy=0.462145	lr=0.394507
Epoch[13] Batch [799]	Speed: 1478.277880 samples/sec	accuracy=0.462335	lr=0.394457
Epoch[13] Batch [849]	Speed: 1510.654379 samples/sec	accuracy=0.462391	lr=0.394405
Epoch[13] Batch [899]	Speed: 1465.045390 samples/sec	accuracy=0.462413	lr=0.394354
Epoch[13] Batch [949]	Speed: 1505.412117 samples/sec	accuracy=0.462435	lr=0.394302
Epoch[13] Batch [999]	Speed: 1544.648247 samples/sec	accuracy=0.462245	lr=0.394250
Epoch[13] Batch [1049]	Speed: 1468.582278 samples/sec	accuracy=0.462599	lr=0.394198
Epoch[13] Batch [1099]	Speed: 1528.017608 samples/sec	accuracy=0.463082	lr=0.394146
Epoch[13] Batch [1149]	Speed: 1505.395275 samples/sec	accuracy=0.463229	lr=0.394093
Epoch[13] Batch [1199]	Speed: 1545.145430 samples/sec	accuracy=0.463433	lr=0.394041
Epoch[13] Batch [1249]	Speed: 1550.558426 samples/sec	accuracy=0.463512	lr=0.393988
[Epoch 13] training: accuracy=0.463508
[Epoch 13] speed: 1511 samples/sec	time cost: 870.192879
[Epoch 13] validation: err-top1=0.462891 err-top5=0.217016
Epoch[14] Batch [49]	Speed: 1527.943440 samples/sec	accuracy=0.475938	lr=0.393933
Epoch[14] Batch [99]	Speed: 1490.660573 samples/sec	accuracy=0.479375	lr=0.393880
Epoch[14] Batch [149]	Speed: 1489.611556 samples/sec	accuracy=0.478874	lr=0.393826
Epoch[14] Batch [199]	Speed: 1501.137978 samples/sec	accuracy=0.477842	lr=0.393772
Epoch[14] Batch [249]	Speed: 1535.239347 samples/sec	accuracy=0.448191	lr=0.393718
Epoch[14] Batch [299]	Speed: 1508.843480 samples/sec	accuracy=0.439922	lr=0.393664
Epoch[14] Batch [349]	Speed: 1527.258546 samples/sec	accuracy=0.442252	lr=0.393609
Epoch[14] Batch [399]	Speed: 1497.666223 samples/sec	accuracy=0.444951	lr=0.393554
Epoch[14] Batch [449]	Speed: 1549.046908 samples/sec	accuracy=0.447329	lr=0.393499
Epoch[14] Batch [499]	Speed: 1488.608962 samples/sec	accuracy=0.449475	lr=0.393444
Epoch[14] Batch [549]	Speed: 1511.118060 samples/sec	accuracy=0.451088	lr=0.393388
Epoch[14] Batch [599]	Speed: 1518.306939 samples/sec	accuracy=0.452493	lr=0.393332
Epoch[14] Batch [649]	Speed: 1491.091435 samples/sec	accuracy=0.453995	lr=0.393276
Epoch[14] Batch [699]	Speed: 1513.901214 samples/sec	accuracy=0.454155	lr=0.393220
Epoch[14] Batch [749]	Speed: 1491.037352 samples/sec	accuracy=0.455505	lr=0.393163
Epoch[14] Batch [799]	Speed: 1538.462300 samples/sec	accuracy=0.456464	lr=0.393107
Epoch[14] Batch [849]	Speed: 1511.401085 samples/sec	accuracy=0.457420	lr=0.393050
Epoch[14] Batch [899]	Speed: 1480.414684 samples/sec	accuracy=0.458014	lr=0.392993
Epoch[14] Batch [949]	Speed: 1508.738662 samples/sec	accuracy=0.458755	lr=0.392935
Epoch[14] Batch [999]	Speed: 1470.826261 samples/sec	accuracy=0.459521	lr=0.392877
Epoch[14] Batch [1049]	Speed: 1518.248899 samples/sec	accuracy=0.460117	lr=0.392820
Epoch[14] Batch [1099]	Speed: 1537.846055 samples/sec	accuracy=0.460669	lr=0.392762
Epoch[14] Batch [1149]	Speed: 1471.794889 samples/sec	accuracy=0.461253	lr=0.392703
Epoch[14] Batch [1199]	Speed: 1542.455717 samples/sec	accuracy=0.461630	lr=0.392645
Epoch[14] Batch [1249]	Speed: 1525.019568 samples/sec	accuracy=0.462113	lr=0.392586
[Epoch 14] training: accuracy=0.462083
[Epoch 14] speed: 1508 samples/sec	time cost: 871.829741
[Epoch 14] validation: err-top1=0.473772 err-top5=0.223952
Epoch[15] Batch [49]	Speed: 1474.565342 samples/sec	accuracy=0.487305	lr=0.392526
Epoch[15] Batch [99]	Speed: 1534.934169 samples/sec	accuracy=0.483564	lr=0.392466
Epoch[15] Batch [149]	Speed: 1516.583683 samples/sec	accuracy=0.483327	lr=0.392407
Epoch[15] Batch [199]	Speed: 1491.001181 samples/sec	accuracy=0.482295	lr=0.392347
Epoch[15] Batch [249]	Speed: 1503.593818 samples/sec	accuracy=0.482828	lr=0.392287
Epoch[15] Batch [299]	Speed: 1464.646428 samples/sec	accuracy=0.482458	lr=0.392227
Epoch[15] Batch [349]	Speed: 1509.320593 samples/sec	accuracy=0.482193	lr=0.392167
Epoch[15] Batch [399]	Speed: 1444.864796 samples/sec	accuracy=0.476772	lr=0.392106
Epoch[15] Batch [449]	Speed: 1539.468766 samples/sec	accuracy=0.475234	lr=0.392045
Epoch[15] Batch [499]	Speed: 1538.339001 samples/sec	accuracy=0.475791	lr=0.391984
Epoch[15] Batch [549]	Speed: 1507.078609 samples/sec	accuracy=0.476481	lr=0.391923
Epoch[15] Batch [599]	Speed: 1524.375737 samples/sec	accuracy=0.476696	lr=0.391861
Epoch[15] Batch [649]	Speed: 1538.470544 samples/sec	accuracy=0.477272	lr=0.391799
Epoch[15] Batch [699]	Speed: 1538.012145 samples/sec	accuracy=0.477670	lr=0.391737
Epoch[15] Batch [749]	Speed: 1495.713922 samples/sec	accuracy=0.477302	lr=0.391675
Epoch[15] Batch [799]	Speed: 1531.950185 samples/sec	accuracy=0.477701	lr=0.391613
Epoch[15] Batch [849]	Speed: 1487.798486 samples/sec	accuracy=0.477213	lr=0.391550
Epoch[15] Batch [899]	Speed: 1521.137220 samples/sec	accuracy=0.477395	lr=0.391487
Epoch[15] Batch [949]	Speed: 1514.003666 samples/sec	accuracy=0.477772	lr=0.391424
Epoch[15] Batch [999]	Speed: 1532.392795 samples/sec	accuracy=0.478136	lr=0.391361
Epoch[15] Batch [1049]	Speed: 1542.655606 samples/sec	accuracy=0.478092	lr=0.391297
Epoch[15] Batch [1099]	Speed: 1502.290427 samples/sec	accuracy=0.477872	lr=0.391233
Epoch[15] Batch [1149]	Speed: 1448.233130 samples/sec	accuracy=0.478193	lr=0.391169
Epoch[15] Batch [1199]	Speed: 1484.467498 samples/sec	accuracy=0.478288	lr=0.391105
Epoch[15] Batch [1249]	Speed: 1518.909265 samples/sec	accuracy=0.478276	lr=0.391040
[Epoch 15] training: accuracy=0.478246
[Epoch 15] speed: 1506 samples/sec	time cost: 872.531242
[Epoch 15] validation: err-top1=0.515705 err-top5=0.258789
Epoch[16] Batch [49]	Speed: 1505.599427 samples/sec	accuracy=0.489531	lr=0.390974
Epoch[16] Batch [99]	Speed: 1486.242983 samples/sec	accuracy=0.489668	lr=0.390909
Epoch[16] Batch [149]	Speed: 1544.816799 samples/sec	accuracy=0.488984	lr=0.390844
Epoch[16] Batch [199]	Speed: 1508.032187 samples/sec	accuracy=0.489268	lr=0.390779
Epoch[16] Batch [249]	Speed: 1534.573557 samples/sec	accuracy=0.487781	lr=0.390713
Epoch[16] Batch [299]	Speed: 1500.883696 samples/sec	accuracy=0.487692	lr=0.390647
Epoch[16] Batch [349]	Speed: 1534.632787 samples/sec	accuracy=0.486267	lr=0.390581
Epoch[16] Batch [399]	Speed: 1481.960980 samples/sec	accuracy=0.486194	lr=0.390515
Epoch[16] Batch [449]	Speed: 1510.937231 samples/sec	accuracy=0.484334	lr=0.390448
Epoch[16] Batch [499]	Speed: 1546.104758 samples/sec	accuracy=0.484799	lr=0.390381
Epoch[16] Batch [549]	Speed: 1544.477388 samples/sec	accuracy=0.484886	lr=0.390314
Epoch[16] Batch [599]	Speed: 1538.959369 samples/sec	accuracy=0.484714	lr=0.390247
Epoch[16] Batch [649]	Speed: 1426.794477 samples/sec	accuracy=0.484435	lr=0.390180
Epoch[16] Batch [699]	Speed: 1532.329814 samples/sec	accuracy=0.484604	lr=0.390112
Epoch[16] Batch [749]	Speed: 1480.826327 samples/sec	accuracy=0.484702	lr=0.390044
Epoch[16] Batch [799]	Speed: 1516.466853 samples/sec	accuracy=0.484723	lr=0.389976
Epoch[16] Batch [849]	Speed: 1517.075972 samples/sec	accuracy=0.484585	lr=0.389907
Epoch[16] Batch [899]	Speed: 1545.316348 samples/sec	accuracy=0.484621	lr=0.389839
Epoch[16] Batch [949]	Speed: 1493.960585 samples/sec	accuracy=0.484398	lr=0.389770
Epoch[16] Batch [999]	Speed: 1484.046074 samples/sec	accuracy=0.484649	lr=0.389701
Epoch[16] Batch [1049]	Speed: 1483.191365 samples/sec	accuracy=0.484697	lr=0.389632
Epoch[16] Batch [1099]	Speed: 1543.006910 samples/sec	accuracy=0.484630	lr=0.389562
Epoch[16] Batch [1149]	Speed: 1513.172699 samples/sec	accuracy=0.484462	lr=0.389492
Epoch[16] Batch [1199]	Speed: 1542.450299 samples/sec	accuracy=0.484544	lr=0.389422
Epoch[16] Batch [1249]	Speed: 1525.239336 samples/sec	accuracy=0.484326	lr=0.389352
[Epoch 16] training: accuracy=0.484323
[Epoch 16] speed: 1511 samples/sec	time cost: 869.275952
[Epoch 16] validation: err-top1=0.476602 err-top5=0.228715
Epoch[17] Batch [49]	Speed: 1500.989523 samples/sec	accuracy=0.501895	lr=0.389280
Epoch[17] Batch [99]	Speed: 1461.442586 samples/sec	accuracy=0.499111	lr=0.389210
Epoch[17] Batch [149]	Speed: 1530.859208 samples/sec	accuracy=0.494948	lr=0.389139
Epoch[17] Batch [199]	Speed: 1543.112375 samples/sec	accuracy=0.493379	lr=0.389068
Epoch[17] Batch [249]	Speed: 1518.808619 samples/sec	accuracy=0.493145	lr=0.388996
Epoch[17] Batch [299]	Speed: 1515.879752 samples/sec	accuracy=0.492012	lr=0.388925
Epoch[17] Batch [349]	Speed: 1540.103509 samples/sec	accuracy=0.492031	lr=0.388853
Epoch[17] Batch [399]	Speed: 1471.805965 samples/sec	accuracy=0.490276	lr=0.388781
Epoch[17] Batch [449]	Speed: 1513.751676 samples/sec	accuracy=0.490447	lr=0.388709
Epoch[17] Batch [499]	Speed: 1518.861373 samples/sec	accuracy=0.489953	lr=0.388636
Epoch[17] Batch [549]	Speed: 1481.696039 samples/sec	accuracy=0.490531	lr=0.388564
Epoch[17] Batch [599]	Speed: 1515.055789 samples/sec	accuracy=0.491034	lr=0.388491
Epoch[17] Batch [649]	Speed: 1536.015242 samples/sec	accuracy=0.490386	lr=0.388418
Epoch[17] Batch [699]	Speed: 1494.202432 samples/sec	accuracy=0.490261	lr=0.388344
Epoch[17] Batch [749]	Speed: 1490.156932 samples/sec	accuracy=0.490147	lr=0.388271
Epoch[17] Batch [799]	Speed: 1544.657157 samples/sec	accuracy=0.490094	lr=0.388197
Epoch[17] Batch [849]	Speed: 1519.871982 samples/sec	accuracy=0.489676	lr=0.388123
Epoch[17] Batch [899]	Speed: 1534.290853 samples/sec	accuracy=0.489783	lr=0.388049
Epoch[17] Batch [949]	Speed: 1512.605607 samples/sec	accuracy=0.489585	lr=0.387974
Epoch[17] Batch [999]	Speed: 1522.048922 samples/sec	accuracy=0.489044	lr=0.387900
Epoch[17] Batch [1049]	Speed: 1520.392210 samples/sec	accuracy=0.488970	lr=0.387825
Epoch[17] Batch [1099]	Speed: 1496.845064 samples/sec	accuracy=0.489074	lr=0.387750
Epoch[17] Batch [1149]	Speed: 1493.999924 samples/sec	accuracy=0.488510	lr=0.387674
Epoch[17] Batch [1199]	Speed: 1507.346962 samples/sec	accuracy=0.488790	lr=0.387599
Epoch[17] Batch [1249]	Speed: 1539.119588 samples/sec	accuracy=0.488453	lr=0.387523
[Epoch 17] training: accuracy=0.488484
[Epoch 17] speed: 1511 samples/sec	time cost: 869.499760
[Epoch 17] validation: err-top1=0.428345 err-top5=0.191752
Epoch[18] Batch [49]	Speed: 1489.658861 samples/sec	accuracy=0.503965	lr=0.387445
Epoch[18] Batch [99]	Speed: 1495.670732 samples/sec	accuracy=0.505273	lr=0.387369
Epoch[18] Batch [149]	Speed: 1509.515286 samples/sec	accuracy=0.503945	lr=0.387293
Epoch[18] Batch [199]	Speed: 1488.946558 samples/sec	accuracy=0.502681	lr=0.387216
Epoch[18] Batch [249]	Speed: 1505.007778 samples/sec	accuracy=0.502211	lr=0.387139
Epoch[18] Batch [299]	Speed: 1527.398652 samples/sec	accuracy=0.496722	lr=0.387062
Epoch[18] Batch [349]	Speed: 1484.405839 samples/sec	accuracy=0.497207	lr=0.386984
Epoch[18] Batch [399]	Speed: 1492.776878 samples/sec	accuracy=0.497188	lr=0.386907
Epoch[18] Batch [449]	Speed: 1483.055739 samples/sec	accuracy=0.497741	lr=0.386829
Epoch[18] Batch [499]	Speed: 1531.653054 samples/sec	accuracy=0.497973	lr=0.386751
Epoch[18] Batch [549]	Speed: 1492.506852 samples/sec	accuracy=0.497260	lr=0.386673
Epoch[18] Batch [599]	Speed: 1529.872738 samples/sec	accuracy=0.497803	lr=0.386594
Epoch[18] Batch [649]	Speed: 1481.149461 samples/sec	accuracy=0.497228	lr=0.386515
Epoch[18] Batch [699]	Speed: 1541.365098 samples/sec	accuracy=0.497331	lr=0.386436
Epoch[18] Batch [749]	Speed: 1478.967115 samples/sec	accuracy=0.497030	lr=0.386357
Epoch[18] Batch [799]	Speed: 1513.032248 samples/sec	accuracy=0.496803	lr=0.386278
Epoch[18] Batch [849]	Speed: 1530.695684 samples/sec	accuracy=0.496174	lr=0.386198
Epoch[18] Batch [899]	Speed: 1511.253869 samples/sec	accuracy=0.495889	lr=0.386119
Epoch[18] Batch [949]	Speed: 1513.949509 samples/sec	accuracy=0.495943	lr=0.386038
Epoch[18] Batch [999]	Speed: 1499.788780 samples/sec	accuracy=0.495591	lr=0.385958
Epoch[18] Batch [1049]	Speed: 1513.044550 samples/sec	accuracy=0.495807	lr=0.385878
Epoch[18] Batch [1099]	Speed: 1534.333299 samples/sec	accuracy=0.495720	lr=0.385797
Epoch[18] Batch [1149]	Speed: 1509.335455 samples/sec	accuracy=0.495851	lr=0.385716
Epoch[18] Batch [1199]	Speed: 1526.701493 samples/sec	accuracy=0.495161	lr=0.385635
Epoch[18] Batch [1249]	Speed: 1541.967836 samples/sec	accuracy=0.495116	lr=0.385554
[Epoch 18] training: accuracy=0.495127
[Epoch 18] speed: 1507 samples/sec	time cost: 871.609697
[Epoch 18] validation: err-top1=0.433235 err-top5=0.192403
Epoch[19] Batch [49]	Speed: 1516.767430 samples/sec	accuracy=0.505137	lr=0.385470
Epoch[19] Batch [99]	Speed: 1511.825652 samples/sec	accuracy=0.504053	lr=0.385388
Epoch[19] Batch [149]	Speed: 1477.812907 samples/sec	accuracy=0.504980	lr=0.385306
Epoch[19] Batch [199]	Speed: 1511.116741 samples/sec	accuracy=0.503633	lr=0.385224
Epoch[19] Batch [249]	Speed: 1522.600716 samples/sec	accuracy=0.503180	lr=0.385142
Epoch[19] Batch [299]	Speed: 1511.481880 samples/sec	accuracy=0.503838	lr=0.385059
Epoch[19] Batch [349]	Speed: 1523.395863 samples/sec	accuracy=0.502938	lr=0.384976
Epoch[19] Batch [399]	Speed: 1537.475587 samples/sec	accuracy=0.502891	lr=0.384893
Epoch[19] Batch [449]	Speed: 1553.795867 samples/sec	accuracy=0.502214	lr=0.384810
Epoch[19] Batch [499]	Speed: 1471.578432 samples/sec	accuracy=0.502271	lr=0.384726
Epoch[19] Batch [549]	Speed: 1514.383294 samples/sec	accuracy=0.502248	lr=0.384642
Epoch[19] Batch [599]	Speed: 1539.348229 samples/sec	accuracy=0.501138	lr=0.384558
Epoch[19] Batch [649]	Speed: 1463.513378 samples/sec	accuracy=0.500945	lr=0.384474
Epoch[19] Batch [699]	Speed: 1504.946130 samples/sec	accuracy=0.500911	lr=0.384389
Epoch[19] Batch [749]	Speed: 1456.935935 samples/sec	accuracy=0.500635	lr=0.384305
Epoch[19] Batch [799]	Speed: 1508.108459 samples/sec	accuracy=0.500513	lr=0.384220
Epoch[19] Batch [849]	Speed: 1534.371668 samples/sec	accuracy=0.500360	lr=0.384135
Epoch[19] Batch [899]	Speed: 1522.838178 samples/sec	accuracy=0.500226	lr=0.384049
Epoch[19] Batch [949]	Speed: 1545.193693 samples/sec	accuracy=0.500453	lr=0.383964
Epoch[19] Batch [999]	Speed: 1436.705517 samples/sec	accuracy=0.500360	lr=0.383878
Epoch[19] Batch [1049]	Speed: 1505.948079 samples/sec	accuracy=0.500238	lr=0.383792
Epoch[19] Batch [1099]	Speed: 1540.973460 samples/sec	accuracy=0.500244	lr=0.383706
Epoch[19] Batch [1149]	Speed: 1459.076164 samples/sec	accuracy=0.499922	lr=0.383619
Epoch[19] Batch [1199]	Speed: 1512.105951 samples/sec	accuracy=0.499896	lr=0.383533
Epoch[19] Batch [1249]	Speed: 1545.452624 samples/sec	accuracy=0.499927	lr=0.383446
[Epoch 19] training: accuracy=0.499941
[Epoch 19] speed: 1507 samples/sec	time cost: 872.373753
[Epoch 19] validation: err-top1=0.433614 err-top5=0.192522
Epoch[20] Batch [49]	Speed: 1500.585039 samples/sec	accuracy=0.507793	lr=0.383357
Epoch[20] Batch [99]	Speed: 1539.373619 samples/sec	accuracy=0.506953	lr=0.383270
Epoch[20] Batch [149]	Speed: 1543.582130 samples/sec	accuracy=0.506348	lr=0.383182
Epoch[20] Batch [199]	Speed: 1528.224832 samples/sec	accuracy=0.506772	lr=0.383094
Epoch[20] Batch [249]	Speed: 1526.864805 samples/sec	accuracy=0.506766	lr=0.383006
Epoch[20] Batch [299]	Speed: 1488.817153 samples/sec	accuracy=0.506501	lr=0.382918
Epoch[20] Batch [349]	Speed: 1491.278802 samples/sec	accuracy=0.506722	lr=0.382830
Epoch[20] Batch [399]	Speed: 1488.136745 samples/sec	accuracy=0.505056	lr=0.382741
Epoch[20] Batch [449]	Speed: 1539.246455 samples/sec	accuracy=0.504987	lr=0.382652
Epoch[20] Batch [499]	Speed: 1540.897399 samples/sec	accuracy=0.505451	lr=0.382563
Epoch[20] Batch [549]	Speed: 1516.017704 samples/sec	accuracy=0.505094	lr=0.382474
Epoch[20] Batch [599]	Speed: 1487.342040 samples/sec	accuracy=0.504803	lr=0.382384
Epoch[20] Batch [649]	Speed: 1463.571827 samples/sec	accuracy=0.504573	lr=0.382295
Epoch[20] Batch [699]	Speed: 1542.649079 samples/sec	accuracy=0.504508	lr=0.382205
Epoch[20] Batch [749]	Speed: 1538.677450 samples/sec	accuracy=0.504430	lr=0.382115
Epoch[20] Batch [799]	Speed: 1521.976831 samples/sec	accuracy=0.504191	lr=0.382024
Epoch[20] Batch [849]	Speed: 1511.612243 samples/sec	accuracy=0.504208	lr=0.381934
Epoch[20] Batch [899]	Speed: 1517.447556 samples/sec	accuracy=0.504078	lr=0.381843
Epoch[20] Batch [949]	Speed: 1434.081715 samples/sec	accuracy=0.503814	lr=0.381752
Epoch[20] Batch [999]	Speed: 1545.992506 samples/sec	accuracy=0.503569	lr=0.381661
Epoch[20] Batch [1049]	Speed: 1533.973944 samples/sec	accuracy=0.503188	lr=0.381569
Epoch[20] Batch [1099]	Speed: 1511.137572 samples/sec	accuracy=0.503007	lr=0.381477
Epoch[20] Batch [1149]	Speed: 1535.319834 samples/sec	accuracy=0.503078	lr=0.381385
Epoch[20] Batch [1199]	Speed: 1488.592008 samples/sec	accuracy=0.502898	lr=0.381293
Epoch[20] Batch [1249]	Speed: 1521.625181 samples/sec	accuracy=0.502934	lr=0.381201
[Epoch 20] training: accuracy=0.502936
[Epoch 20] speed: 1512 samples/sec	time cost: 869.370164
[Epoch 20] validation: err-top1=0.430106 err-top5=0.191566
Epoch[21] Batch [49]	Speed: 1469.715438 samples/sec	accuracy=0.516211	lr=0.381107
Epoch[21] Batch [99]	Speed: 1536.235400 samples/sec	accuracy=0.512148	lr=0.381014
Epoch[21] Batch [149]	Speed: 1534.047099 samples/sec	accuracy=0.512240	lr=0.380921
Epoch[21] Batch [199]	Speed: 1523.469168 samples/sec	accuracy=0.512515	lr=0.380828
Epoch[21] Batch [249]	Speed: 1499.651776 samples/sec	accuracy=0.510664	lr=0.380734
Epoch[21] Batch [299]	Speed: 1520.955741 samples/sec	accuracy=0.510781	lr=0.380641
Epoch[21] Batch [349]	Speed: 1437.925716 samples/sec	accuracy=0.508901	lr=0.380547
Epoch[21] Batch [399]	Speed: 1513.233231 samples/sec	accuracy=0.508218	lr=0.380453
Epoch[21] Batch [449]	Speed: 1496.250099 samples/sec	accuracy=0.508609	lr=0.380359
Epoch[21] Batch [499]	Speed: 1524.630584 samples/sec	accuracy=0.507713	lr=0.380264
Epoch[21] Batch [549]	Speed: 1510.972058 samples/sec	accuracy=0.507820	lr=0.380169
Epoch[21] Batch [599]	Speed: 1494.826828 samples/sec	accuracy=0.507404	lr=0.380074
Epoch[21] Batch [649]	Speed: 1508.744396 samples/sec	accuracy=0.507682	lr=0.379979
Epoch[21] Batch [699]	Speed: 1538.093926 samples/sec	accuracy=0.507806	lr=0.379884
Epoch[21] Batch [749]	Speed: 1521.747240 samples/sec	accuracy=0.507663	lr=0.379788
Epoch[21] Batch [799]	Speed: 1540.967047 samples/sec	accuracy=0.501609	lr=0.379693
Epoch[21] Batch [849]	Speed: 1476.215913 samples/sec	accuracy=0.500022	lr=0.379597
Epoch[21] Batch [899]	Speed: 1495.318783 samples/sec	accuracy=0.500232	lr=0.379501
Epoch[21] Batch [949]	Speed: 1520.060745 samples/sec	accuracy=0.500297	lr=0.379404
Epoch[21] Batch [999]	Speed: 1538.966052 samples/sec	accuracy=0.500588	lr=0.379307
Epoch[21] Batch [1049]	Speed: 1539.465687 samples/sec	accuracy=0.501009	lr=0.379211
Epoch[21] Batch [1099]	Speed: 1525.199602 samples/sec	accuracy=0.500963	lr=0.379114
Epoch[21] Batch [1149]	Speed: 1535.795674 samples/sec	accuracy=0.500971	lr=0.379016
Epoch[21] Batch [1199]	Speed: 1542.248027 samples/sec	accuracy=0.500511	lr=0.378919
Epoch[21] Batch [1249]	Speed: 1547.067016 samples/sec	accuracy=0.500828	lr=0.378821
[Epoch 21] training: accuracy=0.500830
[Epoch 21] speed: 1514 samples/sec	time cost: 869.006766
[Epoch 21] validation: err-top1=0.426000 err-top5=0.188835
Epoch[22] Batch [49]	Speed: 1477.718680 samples/sec	accuracy=0.525488	lr=0.378721
Epoch[22] Batch [99]	Speed: 1477.454826 samples/sec	accuracy=0.522900	lr=0.378623
Epoch[22] Batch [149]	Speed: 1522.248573 samples/sec	accuracy=0.522005	lr=0.378525
Epoch[22] Batch [199]	Speed: 1489.823314 samples/sec	accuracy=0.520547	lr=0.378426
Epoch[22] Batch [249]	Speed: 1528.122796 samples/sec	accuracy=0.519484	lr=0.378328
Epoch[22] Batch [299]	Speed: 1543.339919 samples/sec	accuracy=0.518496	lr=0.378229
Epoch[22] Batch [349]	Speed: 1539.510328 samples/sec	accuracy=0.517319	lr=0.378129
Epoch[22] Batch [399]	Speed: 1509.510086 samples/sec	accuracy=0.516584	lr=0.378030
Epoch[22] Batch [449]	Speed: 1521.998966 samples/sec	accuracy=0.515881	lr=0.377930
Epoch[22] Batch [499]	Speed: 1439.024116 samples/sec	accuracy=0.515459	lr=0.377831
Epoch[22] Batch [549]	Speed: 1546.426387 samples/sec	accuracy=0.514707	lr=0.377731
Epoch[22] Batch [599]	Speed: 1484.773323 samples/sec	accuracy=0.514624	lr=0.377630
Epoch[22] Batch [649]	Speed: 1490.674583 samples/sec	accuracy=0.513809	lr=0.377530
Epoch[22] Batch [699]	Speed: 1524.263253 samples/sec	accuracy=0.513525	lr=0.377429
Epoch[22] Batch [749]	Speed: 1510.138180 samples/sec	accuracy=0.513082	lr=0.377328
Epoch[22] Batch [799]	Speed: 1504.506888 samples/sec	accuracy=0.513102	lr=0.377227
Epoch[22] Batch [849]	Speed: 1485.350287 samples/sec	accuracy=0.512738	lr=0.377126
Epoch[22] Batch [899]	Speed: 1534.259239 samples/sec	accuracy=0.512503	lr=0.377024
Epoch[22] Batch [949]	Speed: 1480.988867 samples/sec	accuracy=0.512499	lr=0.376923
Epoch[22] Batch [999]	Speed: 1517.425672 samples/sec	accuracy=0.512073	lr=0.376821
Epoch[22] Batch [1049]	Speed: 1539.730761 samples/sec	accuracy=0.511922	lr=0.376719
Epoch[22] Batch [1099]	Speed: 1497.402319 samples/sec	accuracy=0.511854	lr=0.376616
Epoch[22] Batch [1149]	Speed: 1520.966621 samples/sec	accuracy=0.511670	lr=0.376514
Epoch[22] Batch [1199]	Speed: 1523.974093 samples/sec	accuracy=0.511515	lr=0.376411
Epoch[22] Batch [1249]	Speed: 1550.503368 samples/sec	accuracy=0.511315	lr=0.376308
[Epoch 22] training: accuracy=0.511297
[Epoch 22] speed: 1508 samples/sec	time cost: 871.592606
[Epoch 22] validation: err-top1=0.426499 err-top5=0.188038
Epoch[23] Batch [49]	Speed: 1492.605298 samples/sec	accuracy=0.524473	lr=0.376203
Epoch[23] Batch [99]	Speed: 1540.682579 samples/sec	accuracy=0.522900	lr=0.376099
Epoch[23] Batch [149]	Speed: 1507.872519 samples/sec	accuracy=0.520534	lr=0.375996
Epoch[23] Batch [199]	Speed: 1540.091073 samples/sec	accuracy=0.519531	lr=0.375892
Epoch[23] Batch [249]	Speed: 1539.476182 samples/sec	accuracy=0.519281	lr=0.375788
Epoch[23] Batch [299]	Speed: 1533.066456 samples/sec	accuracy=0.517878	lr=0.375683
Epoch[23] Batch [349]	Speed: 1516.010683 samples/sec	accuracy=0.518072	lr=0.375579
Epoch[23] Batch [399]	Speed: 1541.443319 samples/sec	accuracy=0.517485	lr=0.375474
Epoch[23] Batch [449]	Speed: 1463.729624 samples/sec	accuracy=0.517040	lr=0.375369
Epoch[23] Batch [499]	Speed: 1523.216556 samples/sec	accuracy=0.516389	lr=0.375264
Epoch[23] Batch [549]	Speed: 1517.330089 samples/sec	accuracy=0.515607	lr=0.375159
Epoch[23] Batch [599]	Speed: 1514.728869 samples/sec	accuracy=0.515386	lr=0.375053
Epoch[23] Batch [649]	Speed: 1474.659409 samples/sec	accuracy=0.515153	lr=0.374948
Epoch[23] Batch [699]	Speed: 1517.593965 samples/sec	accuracy=0.515027	lr=0.374842
Epoch[23] Batch [749]	Speed: 1499.032573 samples/sec	accuracy=0.515027	lr=0.374736
Epoch[23] Batch [799]	Speed: 1538.634631 samples/sec	accuracy=0.514302	lr=0.374629
Epoch[23] Batch [849]	Speed: 1501.429496 samples/sec	accuracy=0.514212	lr=0.374523
Epoch[23] Batch [899]	Speed: 1501.099542 samples/sec	accuracy=0.513978	lr=0.374416
Epoch[23] Batch [949]	Speed: 1525.234061 samples/sec	accuracy=0.513733	lr=0.374309
Epoch[23] Batch [999]	Speed: 1524.236996 samples/sec	accuracy=0.513646	lr=0.374202
Epoch[23] Batch [1049]	Speed: 1487.447192 samples/sec	accuracy=0.513620	lr=0.374095
Epoch[23] Batch [1099]	Speed: 1540.857010 samples/sec	accuracy=0.513594	lr=0.373987
Epoch[23] Batch [1149]	Speed: 1541.268843 samples/sec	accuracy=0.512987	lr=0.373879
Epoch[23] Batch [1199]	Speed: 1506.967660 samples/sec	accuracy=0.513127	lr=0.373771
Epoch[23] Batch [1249]	Speed: 1560.460083 samples/sec	accuracy=0.513238	lr=0.373663
[Epoch 23] training: accuracy=0.513261
[Epoch 23] speed: 1516 samples/sec	time cost: 867.024625
[Epoch 23] validation: err-top1=0.422953 err-top5=0.181539
Epoch[24] Batch [49]	Speed: 1463.698028 samples/sec	accuracy=0.521230	lr=0.373552
Epoch[24] Batch [99]	Speed: 1498.022737 samples/sec	accuracy=0.521553	lr=0.373444
Epoch[24] Batch [149]	Speed: 1513.474617 samples/sec	accuracy=0.519297	lr=0.373335
Epoch[24] Batch [199]	Speed: 1513.588415 samples/sec	accuracy=0.519272	lr=0.373226
Epoch[24] Batch [249]	Speed: 1471.185157 samples/sec	accuracy=0.518938	lr=0.373117
Epoch[24] Batch [299]	Speed: 1524.900352 samples/sec	accuracy=0.518506	lr=0.373007
Epoch[24] Batch [349]	Speed: 1546.134424 samples/sec	accuracy=0.518736	lr=0.372897
Epoch[24] Batch [399]	Speed: 1536.670880 samples/sec	accuracy=0.518625	lr=0.372788
Epoch[24] Batch [449]	Speed: 1452.273125 samples/sec	accuracy=0.517326	lr=0.372678
Epoch[24] Batch [499]	Speed: 1536.958961 samples/sec	accuracy=0.516977	lr=0.372567
Epoch[24] Batch [549]	Speed: 1514.917606 samples/sec	accuracy=0.517372	lr=0.372457
Epoch[24] Batch [599]	Speed: 1542.258594 samples/sec	accuracy=0.516882	lr=0.372346
Epoch[24] Batch [649]	Speed: 1525.793209 samples/sec	accuracy=0.516714	lr=0.372235
Epoch[24] Batch [699]	Speed: 1501.989098 samples/sec	accuracy=0.516215	lr=0.372124
Epoch[24] Batch [749]	Speed: 1469.755915 samples/sec	accuracy=0.516484	lr=0.372013
Epoch[24] Batch [799]	Speed: 1513.815380 samples/sec	accuracy=0.516503	lr=0.371901
Epoch[24] Batch [849]	Speed: 1518.062882 samples/sec	accuracy=0.516087	lr=0.371790
Epoch[24] Batch [899]	Speed: 1515.273657 samples/sec	accuracy=0.515730	lr=0.371678
Epoch[24] Batch [949]	Speed: 1496.103746 samples/sec	accuracy=0.515597	lr=0.371565
Epoch[24] Batch [999]	Speed: 1517.221923 samples/sec	accuracy=0.515474	lr=0.371453
Epoch[24] Batch [1049]	Speed: 1510.634709 samples/sec	accuracy=0.515400	lr=0.371341
Epoch[24] Batch [1099]	Speed: 1458.053008 samples/sec	accuracy=0.515312	lr=0.371228
Epoch[24] Batch [1149]	Speed: 1513.290877 samples/sec	accuracy=0.515077	lr=0.371115
Epoch[24] Batch [1199]	Speed: 1504.659391 samples/sec	accuracy=0.515305	lr=0.371002
Epoch[24] Batch [1249]	Speed: 1507.404076 samples/sec	accuracy=0.514963	lr=0.370888
[Epoch 24] training: accuracy=0.514947
[Epoch 24] speed: 1505 samples/sec	time cost: 872.696759
[Epoch 24] validation: err-top1=0.420500 err-top5=0.179787
Epoch[25] Batch [49]	Speed: 1481.432203 samples/sec	accuracy=0.521758	lr=0.370773
Epoch[25] Batch [99]	Speed: 1505.292592 samples/sec	accuracy=0.522207	lr=0.370659
Epoch[25] Batch [149]	Speed: 1508.620897 samples/sec	accuracy=0.524505	lr=0.370545
Epoch[25] Batch [199]	Speed: 1538.485788 samples/sec	accuracy=0.523530	lr=0.370431
Epoch[25] Batch [249]	Speed: 1548.916064 samples/sec	accuracy=0.523191	lr=0.370316
Epoch[25] Batch [299]	Speed: 1469.111297 samples/sec	accuracy=0.522806	lr=0.370202
Epoch[25] Batch [349]	Speed: 1506.276830 samples/sec	accuracy=0.522176	lr=0.370087
Epoch[25] Batch [399]	Speed: 1499.059141 samples/sec	accuracy=0.520432	lr=0.369972
Epoch[25] Batch [449]	Speed: 1533.609874 samples/sec	accuracy=0.519612	lr=0.369857
Epoch[25] Batch [499]	Speed: 1523.535131 samples/sec	accuracy=0.520041	lr=0.369741
Epoch[25] Batch [549]	Speed: 1484.566425 samples/sec	accuracy=0.499995	lr=0.369626
Epoch[25] Batch [599]	Speed: 1485.965640 samples/sec	accuracy=0.498701	lr=0.369510
Epoch[25] Batch [649]	Speed: 1530.648781 samples/sec	accuracy=0.499223	lr=0.369394
Epoch[25] Batch [699]	Speed: 1517.733902 samples/sec	accuracy=0.499944	lr=0.369278
Epoch[25] Batch [749]	Speed: 1543.706238 samples/sec	accuracy=0.500915	lr=0.369161
Epoch[25] Batch [799]	Speed: 1467.315834 samples/sec	accuracy=0.501586	lr=0.369045
Epoch[25] Batch [849]	Speed: 1549.959593 samples/sec	accuracy=0.502299	lr=0.368928
Epoch[25] Batch [899]	Speed: 1490.704188 samples/sec	accuracy=0.503113	lr=0.368811
Epoch[25] Batch [949]	Speed: 1549.546257 samples/sec	accuracy=0.503460	lr=0.368694
Epoch[25] Batch [999]	Speed: 1518.166253 samples/sec	accuracy=0.504154	lr=0.368576
Epoch[25] Batch [1049]	Speed: 1471.849714 samples/sec	accuracy=0.504601	lr=0.368459
Epoch[25] Batch [1099]	Speed: 1536.926534 samples/sec	accuracy=0.505185	lr=0.368341
Epoch[25] Batch [1149]	Speed: 1503.495896 samples/sec	accuracy=0.505588	lr=0.368223
Epoch[25] Batch [1199]	Speed: 1487.809453 samples/sec	accuracy=0.506041	lr=0.368105
Epoch[25] Batch [1249]	Speed: 1544.561646 samples/sec	accuracy=0.506420	lr=0.367986
[Epoch 25] training: accuracy=0.506450
[Epoch 25] speed: 1510 samples/sec	time cost: 870.373327
[Epoch 25] validation: err-top1=0.416075 err-top5=0.181561
Epoch[26] Batch [49]	Speed: 1502.277269 samples/sec	accuracy=0.525137	lr=0.367865
Epoch[26] Batch [99]	Speed: 1490.010083 samples/sec	accuracy=0.528281	lr=0.367747
Epoch[26] Batch [149]	Speed: 1501.731230 samples/sec	accuracy=0.525703	lr=0.367628
Epoch[26] Batch [199]	Speed: 1537.812522 samples/sec	accuracy=0.521860	lr=0.367508
Epoch[26] Batch [249]	Speed: 1511.563438 samples/sec	accuracy=0.515645	lr=0.367389
Epoch[26] Batch [299]	Speed: 1490.851133 samples/sec	accuracy=0.516882	lr=0.367269
Epoch[26] Batch [349]	Speed: 1488.925467 samples/sec	accuracy=0.517740	lr=0.367150
Epoch[26] Batch [399]	Speed: 1513.035968 samples/sec	accuracy=0.518518	lr=0.367030
Epoch[26] Batch [449]	Speed: 1517.524934 samples/sec	accuracy=0.518696	lr=0.366909
Epoch[26] Batch [499]	Speed: 1438.758706 samples/sec	accuracy=0.519453	lr=0.366789
Epoch[26] Batch [549]	Speed: 1532.097220 samples/sec	accuracy=0.519640	lr=0.366668
Epoch[26] Batch [599]	Speed: 1540.568295 samples/sec	accuracy=0.519705	lr=0.366548
Epoch[26] Batch [649]	Speed: 1531.362753 samples/sec	accuracy=0.519243	lr=0.366426
Epoch[26] Batch [699]	Speed: 1493.643296 samples/sec	accuracy=0.519344	lr=0.366305
Epoch[26] Batch [749]	Speed: 1535.064857 samples/sec	accuracy=0.519454	lr=0.366184
Epoch[26] Batch [799]	Speed: 1508.571066 samples/sec	accuracy=0.519272	lr=0.366062
Epoch[26] Batch [849]	Speed: 1449.348210 samples/sec	accuracy=0.519343	lr=0.365940
Epoch[26] Batch [899]	Speed: 1533.976771 samples/sec	accuracy=0.519282	lr=0.365819
Epoch[26] Batch [949]	Speed: 1537.284566 samples/sec	accuracy=0.519071	lr=0.365696
Epoch[26] Batch [999]	Speed: 1533.071906 samples/sec	accuracy=0.519104	lr=0.365574
Epoch[26] Batch [1049]	Speed: 1497.804473 samples/sec	accuracy=0.519216	lr=0.365451
Epoch[26] Batch [1099]	Speed: 1480.014856 samples/sec	accuracy=0.518805	lr=0.365329
Epoch[26] Batch [1149]	Speed: 1486.408422 samples/sec	accuracy=0.518542	lr=0.365206
Epoch[26] Batch [1199]	Speed: 1512.623688 samples/sec	accuracy=0.518532	lr=0.365082
Epoch[26] Batch [1249]	Speed: 1522.896818 samples/sec	accuracy=0.518617	lr=0.364959
[Epoch 26] training: accuracy=0.518612
[Epoch 26] speed: 1505 samples/sec	time cost: 873.068653
[Epoch 26] validation: err-top1=0.417191 err-top5=0.178033
Epoch[27] Batch [49]	Speed: 1509.780866 samples/sec	accuracy=0.533809	lr=0.364833
Epoch[27] Batch [99]	Speed: 1483.148301 samples/sec	accuracy=0.530742	lr=0.364709
Epoch[27] Batch [149]	Speed: 1506.911371 samples/sec	accuracy=0.527513	lr=0.364585
Epoch[27] Batch [199]	Speed: 1545.188845 samples/sec	accuracy=0.526567	lr=0.364461
Epoch[27] Batch [249]	Speed: 1511.134318 samples/sec	accuracy=0.527043	lr=0.364337
Epoch[27] Batch [299]	Speed: 1487.286106 samples/sec	accuracy=0.526729	lr=0.364212
Epoch[27] Batch [349]	Speed: 1517.431097 samples/sec	accuracy=0.527015	lr=0.364087
Epoch[27] Batch [399]	Speed: 1511.781293 samples/sec	accuracy=0.526458	lr=0.363962
Epoch[27] Batch [449]	Speed: 1546.152246 samples/sec	accuracy=0.526441	lr=0.363837
Epoch[27] Batch [499]	Speed: 1511.674065 samples/sec	accuracy=0.526008	lr=0.363712
Epoch[27] Batch [549]	Speed: 1449.727811 samples/sec	accuracy=0.522646	lr=0.363586
Epoch[27] Batch [599]	Speed: 1515.764165 samples/sec	accuracy=0.521418	lr=0.363461
Epoch[27] Batch [649]	Speed: 1543.465275 samples/sec	accuracy=0.521558	lr=0.363335
Epoch[27] Batch [699]	Speed: 1531.864434 samples/sec	accuracy=0.521755	lr=0.363209
Epoch[27] Batch [749]	Speed: 1498.136094 samples/sec	accuracy=0.521027	lr=0.363082
Epoch[27] Batch [799]	Speed: 1520.649755 samples/sec	accuracy=0.521013	lr=0.362956
Epoch[27] Batch [849]	Speed: 1535.369559 samples/sec	accuracy=0.521068	lr=0.362829
Epoch[27] Batch [899]	Speed: 1517.733730 samples/sec	accuracy=0.521090	lr=0.362702
Epoch[27] Batch [949]	Speed: 1488.560434 samples/sec	accuracy=0.521171	lr=0.362575
Epoch[27] Batch [999]	Speed: 1534.099866 samples/sec	accuracy=0.520970	lr=0.362448
Epoch[27] Batch [1049]	Speed: 1482.229710 samples/sec	accuracy=0.521031	lr=0.362320
Epoch[27] Batch [1099]	Speed: 1514.762931 samples/sec	accuracy=0.521149	lr=0.362193
Epoch[27] Batch [1149]	Speed: 1540.992225 samples/sec	accuracy=0.521155	lr=0.362065
Epoch[27] Batch [1199]	Speed: 1540.442138 samples/sec	accuracy=0.521124	lr=0.361937
Epoch[27] Batch [1249]	Speed: 1558.170618 samples/sec	accuracy=0.516027	lr=0.361809
[Epoch 27] training: accuracy=0.515689
[Epoch 27] speed: 1514 samples/sec	time cost: 867.998760
[Epoch 27] validation: err-top1=0.860830 err-top5=0.697305
Epoch[28] Batch [49]	Speed: 1510.630894 samples/sec	accuracy=0.385000	lr=0.361678
Epoch[28] Batch [99]	Speed: 1515.797567 samples/sec	accuracy=0.443555	lr=0.361549
Epoch[28] Batch [149]	Speed: 1488.327393 samples/sec	accuracy=0.467760	lr=0.361420
Epoch[28] Batch [199]	Speed: 1499.911309 samples/sec	accuracy=0.479561	lr=0.361291
Epoch[28] Batch [249]	Speed: 1539.026581 samples/sec	accuracy=0.487652	lr=0.361162
Epoch[28] Batch [299]	Speed: 1508.292298 samples/sec	accuracy=0.494076	lr=0.361032
Epoch[28] Batch [349]	Speed: 1473.816760 samples/sec	accuracy=0.498491	lr=0.360903
Epoch[28] Batch [399]	Speed: 1509.064825 samples/sec	accuracy=0.502065	lr=0.360773
Epoch[28] Batch [449]	Speed: 1549.855640 samples/sec	accuracy=0.504542	lr=0.360643
Epoch[28] Batch [499]	Speed: 1476.074488 samples/sec	accuracy=0.506443	lr=0.360513
Epoch[28] Batch [549]	Speed: 1526.076694 samples/sec	accuracy=0.508127	lr=0.360382
Epoch[28] Batch [599]	Speed: 1491.238684 samples/sec	accuracy=0.509336	lr=0.360252
Epoch[28] Batch [649]	Speed: 1479.005801 samples/sec	accuracy=0.509947	lr=0.360121
Epoch[28] Batch [699]	Speed: 1514.040684 samples/sec	accuracy=0.511063	lr=0.359990
Epoch[28] Batch [749]	Speed: 1510.512992 samples/sec	accuracy=0.512031	lr=0.359859
Epoch[28] Batch [799]	Speed: 1523.192334 samples/sec	accuracy=0.512628	lr=0.359728
Epoch[28] Batch [849]	Speed: 1482.822964 samples/sec	accuracy=0.513131	lr=0.359596
Epoch[28] Batch [899]	Speed: 1519.387219 samples/sec	accuracy=0.513455	lr=0.359465
Epoch[28] Batch [949]	Speed: 1528.335030 samples/sec	accuracy=0.513994	lr=0.359333
Epoch[28] Batch [999]	Speed: 1490.429926 samples/sec	accuracy=0.514565	lr=0.359201
Epoch[28] Batch [1049]	Speed: 1542.700266 samples/sec	accuracy=0.514962	lr=0.359068
Epoch[28] Batch [1099]	Speed: 1499.580902 samples/sec	accuracy=0.515339	lr=0.358936
Epoch[28] Batch [1149]	Speed: 1471.058942 samples/sec	accuracy=0.515657	lr=0.358803
Epoch[28] Batch [1199]	Speed: 1509.259335 samples/sec	accuracy=0.514852	lr=0.358670
Epoch[28] Batch [1249]	Speed: 1524.200505 samples/sec	accuracy=0.515113	lr=0.358537
[Epoch 28] training: accuracy=0.515142
[Epoch 28] speed: 1505 samples/sec	time cost: 874.193413
[Epoch 28] validation: err-top1=0.412309 err-top5=0.176598
Epoch[29] Batch [49]	Speed: 1525.042527 samples/sec	accuracy=0.536992	lr=0.358401
Epoch[29] Batch [99]	Speed: 1445.424387 samples/sec	accuracy=0.536182	lr=0.358268
Epoch[29] Batch [149]	Speed: 1525.093550 samples/sec	accuracy=0.534264	lr=0.358134
Epoch[29] Batch [199]	Speed: 1518.700564 samples/sec	accuracy=0.533579	lr=0.358001
Epoch[29] Batch [249]	Speed: 1536.951525 samples/sec	accuracy=0.532805	lr=0.357867
Epoch[29] Batch [299]	Speed: 1471.924509 samples/sec	accuracy=0.531458	lr=0.357732
Epoch[29] Batch [349]	Speed: 1497.957606 samples/sec	accuracy=0.531828	lr=0.357598
Epoch[29] Batch [399]	Speed: 1513.187583 samples/sec	accuracy=0.531960	lr=0.357464
Epoch[29] Batch [449]	Speed: 1535.674756 samples/sec	accuracy=0.530410	lr=0.357329
Epoch[29] Batch [499]	Speed: 1482.720184 samples/sec	accuracy=0.530076	lr=0.357194
Epoch[29] Batch [549]	Speed: 1546.405430 samples/sec	accuracy=0.530629	lr=0.357059
Epoch[29] Batch [599]	Speed: 1521.375864 samples/sec	accuracy=0.530573	lr=0.356924
Epoch[29] Batch [649]	Speed: 1501.357551 samples/sec	accuracy=0.530285	lr=0.356788
Epoch[29] Batch [699]	Speed: 1499.426965 samples/sec	accuracy=0.529696	lr=0.356652
Epoch[29] Batch [749]	Speed: 1496.454635 samples/sec	accuracy=0.529376	lr=0.356517
Epoch[29] Batch [799]	Speed: 1432.366429 samples/sec	accuracy=0.529169	lr=0.356380
Epoch[29] Batch [849]	Speed: 1540.581679 samples/sec	accuracy=0.528841	lr=0.356244
Epoch[29] Batch [899]	Speed: 1498.693505 samples/sec	accuracy=0.528193	lr=0.356108
Epoch[29] Batch [949]	Speed: 1474.926146 samples/sec	accuracy=0.527676	lr=0.355971
Epoch[29] Batch [999]	Speed: 1538.460316 samples/sec	accuracy=0.527312	lr=0.355834
Epoch[29] Batch [1049]	Speed: 1518.464659 samples/sec	accuracy=0.527057	lr=0.355698
Epoch[29] Batch [1099]	Speed: 1534.763237 samples/sec	accuracy=0.526756	lr=0.355560
Epoch[29] Batch [1149]	Speed: 1536.927700 samples/sec	accuracy=0.526527	lr=0.355423
Epoch[29] Batch [1199]	Speed: 1460.802475 samples/sec	accuracy=0.526540	lr=0.355285
Epoch[29] Batch [1249]	Speed: 1544.440955 samples/sec	accuracy=0.526031	lr=0.355148
[Epoch 29] training: accuracy=0.526044
[Epoch 29] speed: 1506 samples/sec	time cost: 871.962180
[Epoch 29] validation: err-top1=0.404704 err-top5=0.170654
Epoch[30] Batch [49]	Speed: 1442.867432 samples/sec	accuracy=0.539883	lr=0.355007
Epoch[30] Batch [99]	Speed: 1516.397796 samples/sec	accuracy=0.537695	lr=0.354869
Epoch[30] Batch [149]	Speed: 1495.562945 samples/sec	accuracy=0.535885	lr=0.354731
Epoch[30] Batch [199]	Speed: 1542.640280 samples/sec	accuracy=0.535210	lr=0.354592
Epoch[30] Batch [249]	Speed: 1501.527369 samples/sec	accuracy=0.534590	lr=0.354454
Epoch[30] Batch [299]	Speed: 1480.969666 samples/sec	accuracy=0.534580	lr=0.354315
Epoch[30] Batch [349]	Speed: 1471.724828 samples/sec	accuracy=0.533929	lr=0.354176
Epoch[30] Batch [399]	Speed: 1524.211745 samples/sec	accuracy=0.533669	lr=0.354037
Epoch[30] Batch [449]	Speed: 1498.032538 samples/sec	accuracy=0.533227	lr=0.353897
Epoch[30] Batch [499]	Speed: 1516.645666 samples/sec	accuracy=0.533158	lr=0.353758
Epoch[30] Batch [549]	Speed: 1494.737224 samples/sec	accuracy=0.531740	lr=0.353618
Epoch[30] Batch [599]	Speed: 1543.578369 samples/sec	accuracy=0.531281	lr=0.353478
Epoch[30] Batch [649]	Speed: 1518.277462 samples/sec	accuracy=0.530750	lr=0.353338
Epoch[30] Batch [699]	Speed: 1512.145166 samples/sec	accuracy=0.530792	lr=0.353198
Epoch[30] Batch [749]	Speed: 1498.458441 samples/sec	accuracy=0.530206	lr=0.353057
Epoch[30] Batch [799]	Speed: 1525.355723 samples/sec	accuracy=0.530261	lr=0.352917
Epoch[30] Batch [849]	Speed: 1515.216960 samples/sec	accuracy=0.529235	lr=0.352776
Epoch[30] Batch [899]	Speed: 1521.562920 samples/sec	accuracy=0.528835	lr=0.352635
Epoch[30] Batch [949]	Speed: 1509.943348 samples/sec	accuracy=0.528950	lr=0.352493
Epoch[30] Batch [999]	Speed: 1499.168709 samples/sec	accuracy=0.528824	lr=0.352352
Epoch[30] Batch [1049]	Speed: 1540.663545 samples/sec	accuracy=0.528524	lr=0.352211
Epoch[30] Batch [1099]	Speed: 1469.245159 samples/sec	accuracy=0.528579	lr=0.352069
Epoch[30] Batch [1149]	Speed: 1511.308313 samples/sec	accuracy=0.528473	lr=0.351927
Epoch[30] Batch [1199]	Speed: 1496.072269 samples/sec	accuracy=0.528057	lr=0.351785
Epoch[30] Batch [1249]	Speed: 1516.871428 samples/sec	accuracy=0.527896	lr=0.351642
[Epoch 30] training: accuracy=0.527900
[Epoch 30] speed: 1505 samples/sec	time cost: 872.783711
[Epoch 30] validation: err-top1=0.404297 err-top5=0.170879
Epoch[31] Batch [49]	Speed: 1469.352091 samples/sec	accuracy=0.543555	lr=0.351497
Epoch[31] Batch [99]	Speed: 1480.450792 samples/sec	accuracy=0.539814	lr=0.351354
Epoch[31] Batch [149]	Speed: 1516.391243 samples/sec	accuracy=0.539082	lr=0.351212
Epoch[31] Batch [199]	Speed: 1474.305112 samples/sec	accuracy=0.536992	lr=0.351069
Epoch[31] Batch [249]	Speed: 1512.006577 samples/sec	accuracy=0.535730	lr=0.350925
Epoch[31] Batch [299]	Speed: 1539.460709 samples/sec	accuracy=0.535365	lr=0.350782
Epoch[31] Batch [349]	Speed: 1488.821406 samples/sec	accuracy=0.535276	lr=0.350639
Epoch[31] Batch [399]	Speed: 1522.860187 samples/sec	accuracy=0.535017	lr=0.350495
Epoch[31] Batch [449]	Speed: 1514.559725 samples/sec	accuracy=0.534167	lr=0.350351
Epoch[31] Batch [499]	Speed: 1502.819572 samples/sec	accuracy=0.533803	lr=0.350207
Epoch[31] Batch [549]	Speed: 1514.762760 samples/sec	accuracy=0.532777	lr=0.350062
Epoch[31] Batch [599]	Speed: 1518.787189 samples/sec	accuracy=0.532770	lr=0.349918
Epoch[31] Batch [649]	Speed: 1546.208097 samples/sec	accuracy=0.533083	lr=0.349773
Epoch[31] Batch [699]	Speed: 1521.439986 samples/sec	accuracy=0.532868	lr=0.349629
Epoch[31] Batch [749]	Speed: 1546.448493 samples/sec	accuracy=0.532681	lr=0.349484
Epoch[31] Batch [799]	Speed: 1498.131683 samples/sec	accuracy=0.532428	lr=0.349338
Epoch[31] Batch [849]	Speed: 1482.059196 samples/sec	accuracy=0.517907	lr=0.349193
Epoch[31] Batch [899]	Speed: 1547.811786 samples/sec	accuracy=0.513541	lr=0.349048
Epoch[31] Batch [949]	Speed: 1523.518400 samples/sec	accuracy=0.513238	lr=0.348902
Epoch[31] Batch [999]	Speed: 1485.985115 samples/sec	accuracy=0.513490	lr=0.348756
Epoch[31] Batch [1049]	Speed: 1544.816699 samples/sec	accuracy=0.513718	lr=0.348610
Epoch[31] Batch [1099]	Speed: 1518.048363 samples/sec	accuracy=0.514382	lr=0.348464
Epoch[31] Batch [1149]	Speed: 1480.426165 samples/sec	accuracy=0.515005	lr=0.348317
Epoch[31] Batch [1199]	Speed: 1488.027886 samples/sec	accuracy=0.515452	lr=0.348171
Epoch[31] Batch [1249]	Speed: 1499.168437 samples/sec	accuracy=0.515988	lr=0.348024
[Epoch 31] training: accuracy=0.515993
[Epoch 31] speed: 1507 samples/sec	time cost: 871.564047
[Epoch 31] validation: err-top1=0.407087 err-top5=0.171596
Epoch[32] Batch [49]	Speed: 1480.298023 samples/sec	accuracy=0.538398	lr=0.347874
Epoch[32] Batch [99]	Speed: 1507.331980 samples/sec	accuracy=0.537207	lr=0.347727
Epoch[32] Batch [149]	Speed: 1548.218203 samples/sec	accuracy=0.537181	lr=0.347580
Epoch[32] Batch [199]	Speed: 1532.507641 samples/sec	accuracy=0.535679	lr=0.347432
Epoch[32] Batch [249]	Speed: 1513.032301 samples/sec	accuracy=0.536047	lr=0.347285
Epoch[32] Batch [299]	Speed: 1528.822092 samples/sec	accuracy=0.536025	lr=0.347137
Epoch[32] Batch [349]	Speed: 1508.302394 samples/sec	accuracy=0.535103	lr=0.346989
Epoch[32] Batch [399]	Speed: 1496.286160 samples/sec	accuracy=0.533960	lr=0.346841
Epoch[32] Batch [449]	Speed: 1474.835964 samples/sec	accuracy=0.533893	lr=0.346692
Epoch[32] Batch [499]	Speed: 1362.685916 samples/sec	accuracy=0.533361	lr=0.346544
Epoch[32] Batch [549]	Speed: 1495.183792 samples/sec	accuracy=0.533576	lr=0.346395
Epoch[32] Batch [599]	Speed: 1543.775342 samples/sec	accuracy=0.533377	lr=0.346246
Epoch[32] Batch [649]	Speed: 1494.158674 samples/sec	accuracy=0.533185	lr=0.346097
Epoch[32] Batch [699]	Speed: 1535.149211 samples/sec	accuracy=0.533027	lr=0.345948
Epoch[32] Batch [749]	Speed: 1540.956752 samples/sec	accuracy=0.532961	lr=0.345799
Epoch[32] Batch [799]	Speed: 1454.245757 samples/sec	accuracy=0.532715	lr=0.345649
Epoch[32] Batch [849]	Speed: 1544.735468 samples/sec	accuracy=0.532683	lr=0.345499
Epoch[32] Batch [899]	Speed: 1525.295637 samples/sec	accuracy=0.532559	lr=0.345349
Epoch[32] Batch [949]	Speed: 1541.488263 samples/sec	accuracy=0.532455	lr=0.345199
Epoch[32] Batch [999]	Speed: 1541.385078 samples/sec	accuracy=0.532423	lr=0.345049
Epoch[32] Batch [1049]	Speed: 1500.005234 samples/sec	accuracy=0.532167	lr=0.344899
Epoch[32] Batch [1099]	Speed: 1518.501208 samples/sec	accuracy=0.532077	lr=0.344748
Epoch[32] Batch [1149]	Speed: 1504.700983 samples/sec	accuracy=0.532051	lr=0.344597
Epoch[32] Batch [1199]	Speed: 1429.863022 samples/sec	accuracy=0.531821	lr=0.344446
Epoch[32] Batch [1249]	Speed: 1555.410904 samples/sec	accuracy=0.531788	lr=0.344295
[Epoch 32] training: accuracy=0.531796
[Epoch 32] speed: 1504 samples/sec	time cost: 873.848903
[Epoch 32] validation: err-top1=0.397501 err-top5=0.167012
Epoch[33] Batch [49]	Speed: 1482.222006 samples/sec	accuracy=0.539082	lr=0.344141
Epoch[33] Batch [99]	Speed: 1504.950802 samples/sec	accuracy=0.541191	lr=0.343989
Epoch[33] Batch [149]	Speed: 1541.012262 samples/sec	accuracy=0.540098	lr=0.343838
Epoch[33] Batch [199]	Speed: 1538.997573 samples/sec	accuracy=0.539170	lr=0.343686
Epoch[33] Batch [249]	Speed: 1484.407204 samples/sec	accuracy=0.538918	lr=0.343534
Epoch[33] Batch [299]	Speed: 1461.923249 samples/sec	accuracy=0.538994	lr=0.343382
Epoch[33] Batch [349]	Speed: 1519.812510 samples/sec	accuracy=0.538786	lr=0.343229
Epoch[33] Batch [399]	Speed: 1475.345344 samples/sec	accuracy=0.538040	lr=0.343077
Epoch[33] Batch [449]	Speed: 1510.418958 samples/sec	accuracy=0.537882	lr=0.342924
Epoch[33] Batch [499]	Speed: 1520.836675 samples/sec	accuracy=0.537822	lr=0.342771
Epoch[33] Batch [549]	Speed: 1538.666921 samples/sec	accuracy=0.536800	lr=0.342618
Epoch[33] Batch [599]	Speed: 1515.941421 samples/sec	accuracy=0.536667	lr=0.342465
Epoch[33] Batch [649]	Speed: 1494.758094 samples/sec	accuracy=0.536498	lr=0.342312
Epoch[33] Batch [699]	Speed: 1497.053072 samples/sec	accuracy=0.536431	lr=0.342158
Epoch[33] Batch [749]	Speed: 1538.526261 samples/sec	accuracy=0.536212	lr=0.342005
Epoch[33] Batch [799]	Speed: 1533.965551 samples/sec	accuracy=0.536246	lr=0.341851
Epoch[33] Batch [849]	Speed: 1464.872771 samples/sec	accuracy=0.536257	lr=0.341697
Epoch[33] Batch [899]	Speed: 1536.999586 samples/sec	accuracy=0.536163	lr=0.341543
Epoch[33] Batch [949]	Speed: 1528.258438 samples/sec	accuracy=0.535907	lr=0.341388
Epoch[33] Batch [999]	Speed: 1545.721634 samples/sec	accuracy=0.535773	lr=0.341234
Epoch[33] Batch [1049]	Speed: 1483.323339 samples/sec	accuracy=0.535709	lr=0.341079
Epoch[33] Batch [1099]	Speed: 1538.824323 samples/sec	accuracy=0.535404	lr=0.340924
Epoch[33] Batch [1149]	Speed: 1471.972181 samples/sec	accuracy=0.535121	lr=0.340769
Epoch[33] Batch [1199]	Speed: 1468.265688 samples/sec	accuracy=0.534986	lr=0.340614
Epoch[33] Batch [1249]	Speed: 1518.977444 samples/sec	accuracy=0.534905	lr=0.340458
[Epoch 33] training: accuracy=0.534928
[Epoch 33] speed: 1506 samples/sec	time cost: 871.697374
[Epoch 33] validation: err-top1=0.397760 err-top5=0.162548
Epoch[34] Batch [49]	Speed: 1519.818673 samples/sec	accuracy=0.549355	lr=0.340300
Epoch[34] Batch [99]	Speed: 1543.790002 samples/sec	accuracy=0.546348	lr=0.340144
Epoch[34] Batch [149]	Speed: 1514.157002 samples/sec	accuracy=0.544238	lr=0.339988
Epoch[34] Batch [199]	Speed: 1485.244588 samples/sec	accuracy=0.542842	lr=0.339832
Epoch[34] Batch [249]	Speed: 1513.882153 samples/sec	accuracy=0.542258	lr=0.339676
Epoch[34] Batch [299]	Speed: 1513.179448 samples/sec	accuracy=0.542074	lr=0.339520
Epoch[34] Batch [349]	Speed: 1486.629408 samples/sec	accuracy=0.541362	lr=0.339363
Epoch[34] Batch [399]	Speed: 1544.057433 samples/sec	accuracy=0.541382	lr=0.339206
Epoch[34] Batch [449]	Speed: 1516.901150 samples/sec	accuracy=0.540701	lr=0.339049
Epoch[34] Batch [499]	Speed: 1511.813413 samples/sec	accuracy=0.539697	lr=0.338892
Epoch[34] Batch [549]	Speed: 1552.945914 samples/sec	accuracy=0.539405	lr=0.338735
Epoch[34] Batch [599]	Speed: 1515.874755 samples/sec	accuracy=0.538923	lr=0.338578
Epoch[34] Batch [649]	Speed: 1506.370127 samples/sec	accuracy=0.538433	lr=0.338420
Epoch[34] Batch [699]	Speed: 1538.469652 samples/sec	accuracy=0.537903	lr=0.338263
Epoch[34] Batch [749]	Speed: 1467.894559 samples/sec	accuracy=0.537624	lr=0.338105
Epoch[34] Batch [799]	Speed: 1507.058715 samples/sec	accuracy=0.537581	lr=0.337947
Epoch[34] Batch [849]	Speed: 1519.332590 samples/sec	accuracy=0.537541	lr=0.337788
Epoch[34] Batch [899]	Speed: 1515.935963 samples/sec	accuracy=0.536213	lr=0.337630
Epoch[34] Batch [949]	Speed: 1539.543947 samples/sec	accuracy=0.536237	lr=0.337472
Epoch[34] Batch [999]	Speed: 1520.549728 samples/sec	accuracy=0.536228	lr=0.337313
Epoch[34] Batch [1049]	Speed: 1513.918749 samples/sec	accuracy=0.536072	lr=0.337154
Epoch[34] Batch [1099]	Speed: 1505.722897 samples/sec	accuracy=0.536031	lr=0.336995
Epoch[34] Batch [1149]	Speed: 1487.581130 samples/sec	accuracy=0.535927	lr=0.336836
Epoch[34] Batch [1199]	Speed: 1448.781307 samples/sec	accuracy=0.536082	lr=0.336677
Epoch[34] Batch [1249]	Speed: 1553.387146 samples/sec	accuracy=0.536131	lr=0.336517
[Epoch 34] training: accuracy=0.536153
[Epoch 34] speed: 1512 samples/sec	time cost: 869.055260
[Epoch 34] validation: err-top1=0.392110 err-top5=0.161092
Epoch[35] Batch [49]	Speed: 1518.282142 samples/sec	accuracy=0.552012	lr=0.336354
Epoch[35] Batch [99]	Speed: 1536.983481 samples/sec	accuracy=0.543486	lr=0.336194
Epoch[35] Batch [149]	Speed: 1542.718686 samples/sec	accuracy=0.541322	lr=0.336034
Epoch[35] Batch [199]	Speed: 1513.168946 samples/sec	accuracy=0.537695	lr=0.335874
Epoch[35] Batch [249]	Speed: 1540.601727 samples/sec	accuracy=0.537988	lr=0.335714
Epoch[35] Batch [299]	Speed: 1526.198914 samples/sec	accuracy=0.539619	lr=0.335553
Epoch[35] Batch [349]	Speed: 1546.199180 samples/sec	accuracy=0.538613	lr=0.335393
Epoch[35] Batch [399]	Speed: 1450.752583 samples/sec	accuracy=0.538616	lr=0.335232
Epoch[35] Batch [449]	Speed: 1516.383662 samples/sec	accuracy=0.538835	lr=0.335071
Epoch[35] Batch [499]	Speed: 1534.331512 samples/sec	accuracy=0.539477	lr=0.334910
Epoch[35] Batch [549]	Speed: 1398.024524 samples/sec	accuracy=0.539627	lr=0.334749
Epoch[35] Batch [599]	Speed: 1526.525618 samples/sec	accuracy=0.539159	lr=0.334587
Epoch[35] Batch [649]	Speed: 1537.559414 samples/sec	accuracy=0.539032	lr=0.334425
Epoch[35] Batch [699]	Speed: 1516.626097 samples/sec	accuracy=0.539110	lr=0.334264
Epoch[35] Batch [749]	Speed: 1539.170454 samples/sec	accuracy=0.539059	lr=0.334102
Epoch[35] Batch [799]	Speed: 1485.153520 samples/sec	accuracy=0.538848	lr=0.333940
Epoch[35] Batch [849]	Speed: 1515.569141 samples/sec	accuracy=0.538995	lr=0.333777
Epoch[35] Batch [899]	Speed: 1469.015775 samples/sec	accuracy=0.538993	lr=0.333615
Epoch[35] Batch [949]	Speed: 1480.720822 samples/sec	accuracy=0.539055	lr=0.333452
Epoch[35] Batch [999]	Speed: 1520.224759 samples/sec	accuracy=0.538901	lr=0.333290
Epoch[35] Batch [1049]	Speed: 1506.832492 samples/sec	accuracy=0.538952	lr=0.333127
Epoch[35] Batch [1099]	Speed: 1547.012084 samples/sec	accuracy=0.538918	lr=0.332964
Epoch[35] Batch [1149]	Speed: 1472.949977 samples/sec	accuracy=0.538785	lr=0.332801
Epoch[35] Batch [1199]	Speed: 1535.801342 samples/sec	accuracy=0.538743	lr=0.332637
Epoch[35] Batch [1249]	Speed: 1539.593979 samples/sec	accuracy=0.538627	lr=0.332474
[Epoch 35] training: accuracy=0.538655
[Epoch 35] speed: 1510 samples/sec	time cost: 870.793493
[Epoch 35] validation: err-top1=0.404257 err-top5=0.168367
Epoch[36] Batch [49]	Speed: 1464.571952 samples/sec	accuracy=0.551582	lr=0.332307
Epoch[36] Batch [99]	Speed: 1536.886518 samples/sec	accuracy=0.551191	lr=0.332143
Epoch[36] Batch [149]	Speed: 1543.489359 samples/sec	accuracy=0.549082	lr=0.331979
Epoch[36] Batch [199]	Speed: 1495.809937 samples/sec	accuracy=0.547222	lr=0.331815
Epoch[36] Batch [249]	Speed: 1511.003186 samples/sec	accuracy=0.546465	lr=0.331650
Epoch[36] Batch [299]	Speed: 1510.471748 samples/sec	accuracy=0.546097	lr=0.331486
Epoch[36] Batch [349]	Speed: 1504.876452 samples/sec	accuracy=0.546258	lr=0.331321
Epoch[36] Batch [399]	Speed: 1517.453464 samples/sec	accuracy=0.546421	lr=0.331157
Epoch[36] Batch [449]	Speed: 1509.060965 samples/sec	accuracy=0.540770	lr=0.330992
Epoch[36] Batch [499]	Speed: 1506.111175 samples/sec	accuracy=0.541063	lr=0.330827
Epoch[36] Batch [549]	Speed: 1509.809717 samples/sec	accuracy=0.541127	lr=0.330661
Epoch[36] Batch [599]	Speed: 1514.475035 samples/sec	accuracy=0.540807	lr=0.330496
Epoch[36] Batch [649]	Speed: 1525.598836 samples/sec	accuracy=0.540841	lr=0.330330
Epoch[36] Batch [699]	Speed: 1475.720940 samples/sec	accuracy=0.540445	lr=0.330165
Epoch[36] Batch [749]	Speed: 1522.484555 samples/sec	accuracy=0.540569	lr=0.329999
Epoch[36] Batch [799]	Speed: 1512.765587 samples/sec	accuracy=0.540574	lr=0.329833
Epoch[36] Batch [849]	Speed: 1539.812018 samples/sec	accuracy=0.540731	lr=0.329667
Epoch[36] Batch [899]	Speed: 1509.764329 samples/sec	accuracy=0.540475	lr=0.329500
Epoch[36] Batch [949]	Speed: 1542.174099 samples/sec	accuracy=0.540518	lr=0.329334
Epoch[36] Batch [999]	Speed: 1508.595143 samples/sec	accuracy=0.540339	lr=0.329167
Epoch[36] Batch [1049]	Speed: 1472.035283 samples/sec	accuracy=0.540213	lr=0.329000
Epoch[36] Batch [1099]	Speed: 1534.931470 samples/sec	accuracy=0.540173	lr=0.328833
Epoch[36] Batch [1149]	Speed: 1449.610163 samples/sec	accuracy=0.540163	lr=0.328666
Epoch[36] Batch [1199]	Speed: 1536.941416 samples/sec	accuracy=0.540309	lr=0.328499
Epoch[36] Batch [1249]	Speed: 1544.905529 samples/sec	accuracy=0.540198	lr=0.328332
[Epoch 36] training: accuracy=0.540208
[Epoch 36] speed: 1510 samples/sec	time cost: 869.805142
[Epoch 36] validation: err-top1=0.394392 err-top5=0.164023
Epoch[37] Batch [49]	Speed: 1455.635300 samples/sec	accuracy=0.554141	lr=0.328161
Epoch[37] Batch [99]	Speed: 1503.518212 samples/sec	accuracy=0.552002	lr=0.327993
Epoch[37] Batch [149]	Speed: 1518.656520 samples/sec	accuracy=0.551576	lr=0.327825
Epoch[37] Batch [199]	Speed: 1517.752663 samples/sec	accuracy=0.552456	lr=0.327657
Epoch[37] Batch [249]	Speed: 1540.055829 samples/sec	accuracy=0.551574	lr=0.327489
Epoch[37] Batch [299]	Speed: 1548.859558 samples/sec	accuracy=0.550036	lr=0.327321
Epoch[37] Batch [349]	Speed: 1489.501386 samples/sec	accuracy=0.549241	lr=0.327152
Epoch[37] Batch [399]	Speed: 1495.361994 samples/sec	accuracy=0.548865	lr=0.326983
Epoch[37] Batch [449]	Speed: 1490.131847 samples/sec	accuracy=0.548908	lr=0.326815
Epoch[37] Batch [499]	Speed: 1505.429562 samples/sec	accuracy=0.547906	lr=0.326646
Epoch[37] Batch [549]	Speed: 1541.880096 samples/sec	accuracy=0.546543	lr=0.326477
Epoch[37] Batch [599]	Speed: 1536.935752 samples/sec	accuracy=0.546421	lr=0.326307
Epoch[37] Batch [649]	Speed: 1524.193506 samples/sec	accuracy=0.545960	lr=0.326138
Epoch[37] Batch [699]	Speed: 1522.802889 samples/sec	accuracy=0.545823	lr=0.325968
Epoch[37] Batch [749]	Speed: 1539.447522 samples/sec	accuracy=0.545898	lr=0.325799
Epoch[37] Batch [799]	Speed: 1492.841631 samples/sec	accuracy=0.545769	lr=0.325629
Epoch[37] Batch [849]	Speed: 1508.762035 samples/sec	accuracy=0.545601	lr=0.325459
Epoch[37] Batch [899]	Speed: 1505.027259 samples/sec	accuracy=0.545033	lr=0.325289
Epoch[37] Batch [949]	Speed: 1547.590104 samples/sec	accuracy=0.544720	lr=0.325118
Epoch[37] Batch [999]	Speed: 1518.078700 samples/sec	accuracy=0.544604	lr=0.324948
Epoch[37] Batch [1049]	Speed: 1493.963859 samples/sec	accuracy=0.544203	lr=0.324777
Epoch[37] Batch [1099]	Speed: 1510.168319 samples/sec	accuracy=0.543826	lr=0.324607
Epoch[37] Batch [1149]	Speed: 1515.335255 samples/sec	accuracy=0.543877	lr=0.324436
Epoch[37] Batch [1199]	Speed: 1489.854756 samples/sec	accuracy=0.533645	lr=0.324265
Epoch[37] Batch [1249]	Speed: 1515.547407 samples/sec	accuracy=0.530261	lr=0.324094
[Epoch 37] training: accuracy=0.530230
[Epoch 37] speed: 1511 samples/sec	time cost: 869.435009
[Epoch 37] validation: err-top1=0.442682 err-top5=0.202009
Epoch[38] Batch [49]	Speed: 1498.773239 samples/sec	accuracy=0.517832	lr=0.323919
Epoch[38] Batch [99]	Speed: 1541.916464 samples/sec	accuracy=0.524453	lr=0.323747
Epoch[38] Batch [149]	Speed: 1498.515020 samples/sec	accuracy=0.526719	lr=0.323576
Epoch[38] Batch [199]	Speed: 1507.913568 samples/sec	accuracy=0.529707	lr=0.323404
Epoch[38] Batch [249]	Speed: 1462.193202 samples/sec	accuracy=0.531746	lr=0.323232
Epoch[38] Batch [299]	Speed: 1545.007841 samples/sec	accuracy=0.532806	lr=0.323060
Epoch[38] Batch [349]	Speed: 1505.137712 samples/sec	accuracy=0.532980	lr=0.322888
Epoch[38] Batch [399]	Speed: 1536.782497 samples/sec	accuracy=0.534431	lr=0.322715
Epoch[38] Batch [449]	Speed: 1495.608576 samples/sec	accuracy=0.535627	lr=0.322543
Epoch[38] Batch [499]	Speed: 1477.665135 samples/sec	accuracy=0.536309	lr=0.322370
Epoch[38] Batch [549]	Speed: 1509.077010 samples/sec	accuracy=0.536699	lr=0.322197
Epoch[38] Batch [599]	Speed: 1479.728555 samples/sec	accuracy=0.537096	lr=0.322025
Epoch[38] Batch [649]	Speed: 1481.719399 samples/sec	accuracy=0.537449	lr=0.321851
Epoch[38] Batch [699]	Speed: 1479.447798 samples/sec	accuracy=0.537937	lr=0.321678
Epoch[38] Batch [749]	Speed: 1537.209297 samples/sec	accuracy=0.538603	lr=0.321505
Epoch[38] Batch [799]	Speed: 1509.221702 samples/sec	accuracy=0.538586	lr=0.321331
Epoch[38] Batch [849]	Speed: 1499.820843 samples/sec	accuracy=0.538859	lr=0.321158
Epoch[38] Batch [899]	Speed: 1484.000192 samples/sec	accuracy=0.539110	lr=0.320984
Epoch[38] Batch [949]	Speed: 1490.559011 samples/sec	accuracy=0.539299	lr=0.320810
Epoch[38] Batch [999]	Speed: 1500.282319 samples/sec	accuracy=0.538972	lr=0.320636
Epoch[38] Batch [1049]	Speed: 1539.955110 samples/sec	accuracy=0.539203	lr=0.320462
Epoch[38] Batch [1099]	Speed: 1514.641627 samples/sec	accuracy=0.539478	lr=0.320287
Epoch[38] Batch [1149]	Speed: 1485.002860 samples/sec	accuracy=0.539744	lr=0.320113
Epoch[38] Batch [1199]	Speed: 1481.553224 samples/sec	accuracy=0.539655	lr=0.319938
Epoch[38] Batch [1249]	Speed: 1527.727598 samples/sec	accuracy=0.539662	lr=0.319763
[Epoch 38] training: accuracy=0.539665
[Epoch 38] speed: 1502 samples/sec	time cost: 874.478085
[Epoch 38] validation: err-top1=0.385882 err-top5=0.156350
Epoch[39] Batch [49]	Speed: 1517.943528 samples/sec	accuracy=0.555684	lr=0.319585
Epoch[39] Batch [99]	Speed: 1539.947623 samples/sec	accuracy=0.555088	lr=0.319410
Epoch[39] Batch [149]	Speed: 1479.609862 samples/sec	accuracy=0.553724	lr=0.319234
Epoch[39] Batch [199]	Speed: 1508.569190 samples/sec	accuracy=0.553247	lr=0.319059
Epoch[39] Batch [249]	Speed: 1508.936279 samples/sec	accuracy=0.553090	lr=0.318883
Epoch[39] Batch [299]	Speed: 1513.037802 samples/sec	accuracy=0.552520	lr=0.318708
Epoch[39] Batch [349]	Speed: 1504.775812 samples/sec	accuracy=0.551629	lr=0.318532
Epoch[39] Batch [399]	Speed: 1500.713592 samples/sec	accuracy=0.550786	lr=0.318356
Epoch[39] Batch [449]	Speed: 1527.151317 samples/sec	accuracy=0.550093	lr=0.318180
Epoch[39] Batch [499]	Speed: 1494.436796 samples/sec	accuracy=0.549834	lr=0.318004
Epoch[39] Batch [549]	Speed: 1483.300338 samples/sec	accuracy=0.549480	lr=0.317827
Epoch[39] Batch [599]	Speed: 1508.078794 samples/sec	accuracy=0.549186	lr=0.317651
Epoch[39] Batch [649]	Speed: 1505.803835 samples/sec	accuracy=0.548664	lr=0.317474
Epoch[39] Batch [699]	Speed: 1477.445414 samples/sec	accuracy=0.548637	lr=0.317297
Epoch[39] Batch [749]	Speed: 1539.599542 samples/sec	accuracy=0.548444	lr=0.317120
Epoch[39] Batch [799]	Speed: 1502.606616 samples/sec	accuracy=0.547946	lr=0.316943
Epoch[39] Batch [849]	Speed: 1495.011492 samples/sec	accuracy=0.547815	lr=0.316766
Epoch[39] Batch [899]	Speed: 1488.949944 samples/sec	accuracy=0.547689	lr=0.316589
Epoch[39] Batch [949]	Speed: 1501.470993 samples/sec	accuracy=0.547439	lr=0.316411
Epoch[39] Batch [999]	Speed: 1491.678792 samples/sec	accuracy=0.547216	lr=0.316233
Epoch[39] Batch [1049]	Speed: 1507.628756 samples/sec	accuracy=0.547094	lr=0.316056
Epoch[39] Batch [1099]	Speed: 1498.577281 samples/sec	accuracy=0.546990	lr=0.315878
Epoch[39] Batch [1149]	Speed: 1462.385884 samples/sec	accuracy=0.546641	lr=0.315700
Epoch[39] Batch [1199]	Speed: 1518.022620 samples/sec	accuracy=0.546393	lr=0.315522
Epoch[39] Batch [1249]	Speed: 1548.352603 samples/sec	accuracy=0.546391	lr=0.315343
[Epoch 39] training: accuracy=0.546393
[Epoch 39] speed: 1503 samples/sec	time cost: 874.100557
[Epoch 39] validation: err-top1=0.390585 err-top5=0.161870
Epoch[40] Batch [49]	Speed: 1506.233989 samples/sec	accuracy=0.556406	lr=0.315161
Epoch[40] Batch [99]	Speed: 1490.678070 samples/sec	accuracy=0.557412	lr=0.314983
Epoch[40] Batch [149]	Speed: 1438.783181 samples/sec	accuracy=0.553073	lr=0.314804
Epoch[40] Batch [199]	Speed: 1509.750605 samples/sec	accuracy=0.553281	lr=0.314625
Epoch[40] Batch [249]	Speed: 1487.142180 samples/sec	accuracy=0.552375	lr=0.314446
Epoch[40] Batch [299]	Speed: 1542.300950 samples/sec	accuracy=0.552520	lr=0.314267
Epoch[40] Batch [349]	Speed: 1521.913689 samples/sec	accuracy=0.552796	lr=0.314087
Epoch[40] Batch [399]	Speed: 1499.880803 samples/sec	accuracy=0.550291	lr=0.313908
Epoch[40] Batch [449]	Speed: 1537.767780 samples/sec	accuracy=0.550148	lr=0.313728
Epoch[40] Batch [499]	Speed: 1471.372262 samples/sec	accuracy=0.550074	lr=0.313549
Epoch[40] Batch [549]	Speed: 1462.130802 samples/sec	accuracy=0.549854	lr=0.313369
Epoch[40] Batch [599]	Speed: 1461.621421 samples/sec	accuracy=0.549403	lr=0.313189
Epoch[40] Batch [649]	Speed: 1501.939274 samples/sec	accuracy=0.549438	lr=0.313009
Epoch[40] Batch [699]	Speed: 1518.498008 samples/sec	accuracy=0.549388	lr=0.312829
Epoch[40] Batch [749]	Speed: 1517.472411 samples/sec	accuracy=0.549158	lr=0.312648
Epoch[40] Batch [799]	Speed: 1541.055877 samples/sec	accuracy=0.548934	lr=0.312468
Epoch[40] Batch [849]	Speed: 1517.775908 samples/sec	accuracy=0.548760	lr=0.312287
Epoch[40] Batch [899]	Speed: 1525.339840 samples/sec	accuracy=0.548602	lr=0.312106
Epoch[40] Batch [949]	Speed: 1435.018118 samples/sec	accuracy=0.548637	lr=0.311925
Epoch[40] Batch [999]	Speed: 1542.189892 samples/sec	accuracy=0.548790	lr=0.311744
Epoch[40] Batch [1049]	Speed: 1480.872187 samples/sec	accuracy=0.548661	lr=0.311563
Epoch[40] Batch [1099]	Speed: 1492.771907 samples/sec	accuracy=0.548703	lr=0.311382
Epoch[40] Batch [1149]	Speed: 1479.868643 samples/sec	accuracy=0.548412	lr=0.311201
Epoch[40] Batch [1199]	Speed: 1546.253020 samples/sec	accuracy=0.548125	lr=0.311019
Epoch[40] Batch [1249]	Speed: 1536.049070 samples/sec	accuracy=0.548090	lr=0.310837
[Epoch 40] training: accuracy=0.548090
[Epoch 40] speed: 1500 samples/sec	time cost: 875.386174
[Epoch 40] validation: err-top1=0.385478 err-top5=0.156759
Epoch[41] Batch [49]	Speed: 1490.432398 samples/sec	accuracy=0.560449	lr=0.310652
Epoch[41] Batch [99]	Speed: 1438.980068 samples/sec	accuracy=0.558047	lr=0.310470
Epoch[41] Batch [149]	Speed: 1509.493428 samples/sec	accuracy=0.558346	lr=0.310288
Epoch[41] Batch [199]	Speed: 1509.258433 samples/sec	accuracy=0.557134	lr=0.310105
Epoch[41] Batch [249]	Speed: 1487.696612 samples/sec	accuracy=0.556824	lr=0.309923
Epoch[41] Batch [299]	Speed: 1513.874822 samples/sec	accuracy=0.555856	lr=0.309741
Epoch[41] Batch [349]	Speed: 1502.715568 samples/sec	accuracy=0.555550	lr=0.309558
Epoch[41] Batch [399]	Speed: 1487.098238 samples/sec	accuracy=0.554617	lr=0.309375
Epoch[41] Batch [449]	Speed: 1512.877253 samples/sec	accuracy=0.554106	lr=0.309192
Epoch[41] Batch [499]	Speed: 1507.730869 samples/sec	accuracy=0.553975	lr=0.309009
Epoch[41] Batch [549]	Speed: 1499.169442 samples/sec	accuracy=0.553272	lr=0.308826
Epoch[41] Batch [599]	Speed: 1515.237808 samples/sec	accuracy=0.552899	lr=0.308643
Epoch[41] Batch [649]	Speed: 1520.566287 samples/sec	accuracy=0.552533	lr=0.308459
Epoch[41] Batch [699]	Speed: 1470.724311 samples/sec	accuracy=0.552072	lr=0.308276
Epoch[41] Batch [749]	Speed: 1489.000500 samples/sec	accuracy=0.552053	lr=0.308092
Epoch[41] Batch [799]	Speed: 1536.940174 samples/sec	accuracy=0.551416	lr=0.307908
Epoch[41] Batch [849]	Speed: 1495.265798 samples/sec	accuracy=0.551211	lr=0.307725
Epoch[41] Batch [899]	Speed: 1525.049773 samples/sec	accuracy=0.551189	lr=0.307540
Epoch[41] Batch [949]	Speed: 1471.861578 samples/sec	accuracy=0.550900	lr=0.307356
Epoch[41] Batch [999]	Speed: 1495.922438 samples/sec	accuracy=0.550883	lr=0.307172
Epoch[41] Batch [1049]	Speed: 1516.801659 samples/sec	accuracy=0.550947	lr=0.306988
Epoch[41] Batch [1099]	Speed: 1506.890921 samples/sec	accuracy=0.550834	lr=0.306803
Epoch[41] Batch [1149]	Speed: 1493.482153 samples/sec	accuracy=0.550594	lr=0.306618
Epoch[41] Batch [1199]	Speed: 1480.128391 samples/sec	accuracy=0.550325	lr=0.306433
Epoch[41] Batch [1249]	Speed: 1511.380885 samples/sec	accuracy=0.550177	lr=0.306249
[Epoch 41] training: accuracy=0.550187
[Epoch 41] speed: 1498 samples/sec	time cost: 877.050376
[Epoch 41] validation: err-top1=0.381019 err-top5=0.153659
Epoch[42] Batch [49]	Speed: 1523.540676 samples/sec	accuracy=0.562695	lr=0.306060
Epoch[42] Batch [99]	Speed: 1511.278394 samples/sec	accuracy=0.561855	lr=0.305875
Epoch[42] Batch [149]	Speed: 1528.994534 samples/sec	accuracy=0.559447	lr=0.305689
Epoch[42] Batch [199]	Speed: 1477.421059 samples/sec	accuracy=0.559297	lr=0.305504
Epoch[42] Batch [249]	Speed: 1488.050952 samples/sec	accuracy=0.559078	lr=0.305318
Epoch[42] Batch [299]	Speed: 1489.358797 samples/sec	accuracy=0.559215	lr=0.305133
Epoch[42] Batch [349]	Speed: 1546.301519 samples/sec	accuracy=0.558733	lr=0.304947
Epoch[42] Batch [399]	Speed: 1485.046093 samples/sec	accuracy=0.558447	lr=0.304761
Epoch[42] Batch [449]	Speed: 1522.044714 samples/sec	accuracy=0.558279	lr=0.304575
Epoch[42] Batch [499]	Speed: 1487.065934 samples/sec	accuracy=0.557937	lr=0.304388
Epoch[42] Batch [549]	Speed: 1472.803711 samples/sec	accuracy=0.557390	lr=0.304202
Epoch[42] Batch [599]	Speed: 1465.587527 samples/sec	accuracy=0.556696	lr=0.304016
Epoch[42] Batch [649]	Speed: 1547.654592 samples/sec	accuracy=0.556439	lr=0.303829
Epoch[42] Batch [699]	Speed: 1515.196680 samples/sec	accuracy=0.556006	lr=0.303642
Epoch[42] Batch [749]	Speed: 1522.882552 samples/sec	accuracy=0.555762	lr=0.303456
Epoch[42] Batch [799]	Speed: 1538.927287 samples/sec	accuracy=0.555603	lr=0.303269
Epoch[42] Batch [849]	Speed: 1468.333473 samples/sec	accuracy=0.555190	lr=0.303082
Epoch[42] Batch [899]	Speed: 1522.655225 samples/sec	accuracy=0.555036	lr=0.302894
Epoch[42] Batch [949]	Speed: 1540.902916 samples/sec	accuracy=0.554780	lr=0.302707
Epoch[42] Batch [999]	Speed: 1521.986313 samples/sec	accuracy=0.554561	lr=0.302520
Epoch[42] Batch [1049]	Speed: 1529.787133 samples/sec	accuracy=0.554149	lr=0.302332
Epoch[42] Batch [1099]	Speed: 1520.952650 samples/sec	accuracy=0.553853	lr=0.302144
Epoch[42] Batch [1149]	Speed: 1485.709122 samples/sec	accuracy=0.553803	lr=0.301957
Epoch[42] Batch [1199]	Speed: 1516.876399 samples/sec	accuracy=0.553748	lr=0.301769
Epoch[42] Batch [1249]	Speed: 1529.956653 samples/sec	accuracy=0.553743	lr=0.301581
[Epoch 42] training: accuracy=0.553753
[Epoch 42] speed: 1508 samples/sec	time cost: 871.744318
[Epoch 42] validation: err-top1=0.376116 err-top5=0.148916
Epoch[43] Batch [49]	Speed: 1477.780263 samples/sec	accuracy=0.560898	lr=0.301389
Epoch[43] Batch [99]	Speed: 1501.658941 samples/sec	accuracy=0.560937	lr=0.301200
Epoch[43] Batch [149]	Speed: 1532.881922 samples/sec	accuracy=0.561576	lr=0.301012
Epoch[43] Batch [199]	Speed: 1494.884840 samples/sec	accuracy=0.561689	lr=0.300823
Epoch[43] Batch [249]	Speed: 1544.749169 samples/sec	accuracy=0.560562	lr=0.300635
Epoch[43] Batch [299]	Speed: 1539.637182 samples/sec	accuracy=0.559906	lr=0.300446
Epoch[43] Batch [349]	Speed: 1495.548394 samples/sec	accuracy=0.560103	lr=0.300257
Epoch[43] Batch [399]	Speed: 1503.736850 samples/sec	accuracy=0.558457	lr=0.300068
Epoch[43] Batch [449]	Speed: 1508.156755 samples/sec	accuracy=0.558420	lr=0.299879
Epoch[43] Batch [499]	Speed: 1475.033988 samples/sec	accuracy=0.557951	lr=0.299690
Epoch[43] Batch [549]	Speed: 1532.762918 samples/sec	accuracy=0.557393	lr=0.299500
Epoch[43] Batch [599]	Speed: 1531.981758 samples/sec	accuracy=0.557583	lr=0.299311
Epoch[43] Batch [649]	Speed: 1521.715009 samples/sec	accuracy=0.557296	lr=0.299121
Epoch[43] Batch [699]	Speed: 1484.986162 samples/sec	accuracy=0.556786	lr=0.298932
Epoch[43] Batch [749]	Speed: 1512.003224 samples/sec	accuracy=0.556598	lr=0.298742
Epoch[43] Batch [799]	Speed: 1499.101637 samples/sec	accuracy=0.556395	lr=0.298552
Epoch[43] Batch [849]	Speed: 1547.419374 samples/sec	accuracy=0.556375	lr=0.298362
Epoch[43] Batch [899]	Speed: 1424.963273 samples/sec	accuracy=0.556374	lr=0.298171
Epoch[43] Batch [949]	Speed: 1541.889495 samples/sec	accuracy=0.556183	lr=0.297981
Epoch[43] Batch [999]	Speed: 1488.118926 samples/sec	accuracy=0.555932	lr=0.297791
Epoch[43] Batch [1049]	Speed: 1527.162405 samples/sec	accuracy=0.555546	lr=0.297600
Epoch[43] Batch [1099]	Speed: 1513.255314 samples/sec	accuracy=0.555458	lr=0.297409
Epoch[43] Batch [1149]	Speed: 1486.597186 samples/sec	accuracy=0.555404	lr=0.297219
Epoch[43] Batch [1199]	Speed: 1510.960672 samples/sec	accuracy=0.555108	lr=0.297028
Epoch[43] Batch [1249]	Speed: 1549.377141 samples/sec	accuracy=0.555028	lr=0.296837
[Epoch 43] training: accuracy=0.555044
[Epoch 43] speed: 1508 samples/sec	time cost: 871.549324
[Epoch 43] validation: err-top1=0.385682 err-top5=0.155732
Epoch[44] Batch [49]	Speed: 1503.654260 samples/sec	accuracy=0.568418	lr=0.296642
Epoch[44] Batch [99]	Speed: 1553.252466 samples/sec	accuracy=0.565254	lr=0.296451
Epoch[44] Batch [149]	Speed: 1523.961818 samples/sec	accuracy=0.564707	lr=0.296259
Epoch[44] Batch [199]	Speed: 1519.397389 samples/sec	accuracy=0.562905	lr=0.296068
Epoch[44] Batch [249]	Speed: 1517.586179 samples/sec	accuracy=0.561438	lr=0.295876
Epoch[44] Batch [299]	Speed: 1454.875678 samples/sec	accuracy=0.561689	lr=0.295684
Epoch[44] Batch [349]	Speed: 1504.049214 samples/sec	accuracy=0.561493	lr=0.295493
Epoch[44] Batch [399]	Speed: 1533.796598 samples/sec	accuracy=0.561201	lr=0.295301
Epoch[44] Batch [449]	Speed: 1513.865857 samples/sec	accuracy=0.560996	lr=0.295109
Epoch[44] Batch [499]	Speed: 1541.979284 samples/sec	accuracy=0.560594	lr=0.294917
Epoch[44] Batch [549]	Speed: 1495.276178 samples/sec	accuracy=0.560678	lr=0.294724
Epoch[44] Batch [599]	Speed: 1507.294337 samples/sec	accuracy=0.560820	lr=0.294532
Epoch[44] Batch [649]	Speed: 1539.613759 samples/sec	accuracy=0.559492	lr=0.294339
Epoch[44] Batch [699]	Speed: 1532.581969 samples/sec	accuracy=0.540375	lr=0.294147
Epoch[44] Batch [749]	Speed: 1509.891424 samples/sec	accuracy=0.538790	lr=0.293954
Epoch[44] Batch [799]	Speed: 1517.899474 samples/sec	accuracy=0.538485	lr=0.293761
Epoch[44] Batch [849]	Speed: 1475.167107 samples/sec	accuracy=0.539060	lr=0.293568
Epoch[44] Batch [899]	Speed: 1527.889651 samples/sec	accuracy=0.539467	lr=0.293375
Epoch[44] Batch [949]	Speed: 1517.829975 samples/sec	accuracy=0.539881	lr=0.293182
Epoch[44] Batch [999]	Speed: 1488.290373 samples/sec	accuracy=0.540276	lr=0.292989
Epoch[44] Batch [1049]	Speed: 1544.907440 samples/sec	accuracy=0.540863	lr=0.292795
Epoch[44] Batch [1099]	Speed: 1503.369781 samples/sec	accuracy=0.541324	lr=0.292602
Epoch[44] Batch [1149]	Speed: 1492.284914 samples/sec	accuracy=0.541982	lr=0.292408
Epoch[44] Batch [1199]	Speed: 1483.702926 samples/sec	accuracy=0.542314	lr=0.292215
Epoch[44] Batch [1249]	Speed: 1558.074491 samples/sec	accuracy=0.542601	lr=0.292021
[Epoch 44] training: accuracy=0.542620
[Epoch 44] speed: 1512 samples/sec	time cost: 868.500798
[Epoch 44] validation: err-top1=0.381736 err-top5=0.154157
Epoch[45] Batch [49]	Speed: 1497.467537 samples/sec	accuracy=0.565234	lr=0.291823
Epoch[45] Batch [99]	Speed: 1543.173097 samples/sec	accuracy=0.565664	lr=0.291629
Epoch[45] Batch [149]	Speed: 1519.609260 samples/sec	accuracy=0.566621	lr=0.291435
Epoch[45] Batch [199]	Speed: 1508.299788 samples/sec	accuracy=0.564946	lr=0.291240
Epoch[45] Batch [249]	Speed: 1500.810628 samples/sec	accuracy=0.565059	lr=0.291046
Epoch[45] Batch [299]	Speed: 1515.160920 samples/sec	accuracy=0.564639	lr=0.290852
Epoch[45] Batch [349]	Speed: 1485.787049 samples/sec	accuracy=0.563761	lr=0.290657
Epoch[45] Batch [399]	Speed: 1462.278171 samples/sec	accuracy=0.562847	lr=0.290462
Epoch[45] Batch [449]	Speed: 1532.563167 samples/sec	accuracy=0.562064	lr=0.290267
Epoch[45] Batch [499]	Speed: 1543.598163 samples/sec	accuracy=0.561730	lr=0.290073
Epoch[45] Batch [549]	Speed: 1506.965259 samples/sec	accuracy=0.561660	lr=0.289878
Epoch[45] Batch [599]	Speed: 1497.153444 samples/sec	accuracy=0.561242	lr=0.289682
Epoch[45] Batch [649]	Speed: 1505.582601 samples/sec	accuracy=0.561019	lr=0.289487
Epoch[45] Batch [699]	Speed: 1509.360544 samples/sec	accuracy=0.560908	lr=0.289292
Epoch[45] Batch [749]	Speed: 1543.206731 samples/sec	accuracy=0.560582	lr=0.289096
Epoch[45] Batch [799]	Speed: 1499.703303 samples/sec	accuracy=0.559750	lr=0.288901
Epoch[45] Batch [849]	Speed: 1489.626879 samples/sec	accuracy=0.559969	lr=0.288705
Epoch[45] Batch [899]	Speed: 1494.813083 samples/sec	accuracy=0.559595	lr=0.288509
Epoch[45] Batch [949]	Speed: 1488.192599 samples/sec	accuracy=0.559600	lr=0.288314
Epoch[45] Batch [999]	Speed: 1540.884142 samples/sec	accuracy=0.559718	lr=0.288118
Epoch[45] Batch [1049]	Speed: 1536.785037 samples/sec	accuracy=0.559511	lr=0.287921
Epoch[45] Batch [1099]	Speed: 1467.862753 samples/sec	accuracy=0.559504	lr=0.287725
Epoch[45] Batch [1149]	Speed: 1531.779916 samples/sec	accuracy=0.559270	lr=0.287529
Epoch[45] Batch [1199]	Speed: 1496.709380 samples/sec	accuracy=0.559075	lr=0.287333
Epoch[45] Batch [1249]	Speed: 1522.714736 samples/sec	accuracy=0.559054	lr=0.287136
[Epoch 45] training: accuracy=0.559054
[Epoch 45] speed: 1508 samples/sec	time cost: 870.959089
[Epoch 45] validation: err-top1=0.380959 err-top5=0.151905
Epoch[46] Batch [49]	Speed: 1473.731204 samples/sec	accuracy=0.565977	lr=0.286936
Epoch[46] Batch [99]	Speed: 1498.724820 samples/sec	accuracy=0.567715	lr=0.286739
Epoch[46] Batch [149]	Speed: 1480.174453 samples/sec	accuracy=0.566126	lr=0.286542
Epoch[46] Batch [199]	Speed: 1525.660789 samples/sec	accuracy=0.565425	lr=0.286345
Epoch[46] Batch [249]	Speed: 1544.367905 samples/sec	accuracy=0.565754	lr=0.286148
Epoch[46] Batch [299]	Speed: 1511.145611 samples/sec	accuracy=0.564993	lr=0.285951
Epoch[46] Batch [349]	Speed: 1529.354151 samples/sec	accuracy=0.563797	lr=0.285754
Epoch[46] Batch [399]	Speed: 1472.729534 samples/sec	accuracy=0.564219	lr=0.285556
Epoch[46] Batch [449]	Speed: 1513.699990 samples/sec	accuracy=0.564366	lr=0.285359
Epoch[46] Batch [499]	Speed: 1508.742435 samples/sec	accuracy=0.564072	lr=0.285161
Epoch[46] Batch [549]	Speed: 1528.000517 samples/sec	accuracy=0.563679	lr=0.284964
Epoch[46] Batch [599]	Speed: 1518.734869 samples/sec	accuracy=0.563511	lr=0.284766
Epoch[46] Batch [649]	Speed: 1504.917771 samples/sec	accuracy=0.563454	lr=0.284568
Epoch[46] Batch [699]	Speed: 1489.751754 samples/sec	accuracy=0.563535	lr=0.284370
Epoch[46] Batch [749]	Speed: 1486.200081 samples/sec	accuracy=0.563311	lr=0.284172
Epoch[46] Batch [799]	Speed: 1517.084985 samples/sec	accuracy=0.562885	lr=0.283974
Epoch[46] Batch [849]	Speed: 1537.081270 samples/sec	accuracy=0.562656	lr=0.283776
Epoch[46] Batch [899]	Speed: 1488.352500 samples/sec	accuracy=0.562420	lr=0.283577
Epoch[46] Batch [949]	Speed: 1476.181564 samples/sec	accuracy=0.562324	lr=0.283379
Epoch[46] Batch [999]	Speed: 1517.562242 samples/sec	accuracy=0.562021	lr=0.283181
Epoch[46] Batch [1049]	Speed: 1464.873580 samples/sec	accuracy=0.561853	lr=0.282982
Epoch[46] Batch [1099]	Speed: 1507.963218 samples/sec	accuracy=0.561738	lr=0.282783
Epoch[46] Batch [1149]	Speed: 1545.124296 samples/sec	accuracy=0.561557	lr=0.282584
Epoch[46] Batch [1199]	Speed: 1508.816161 samples/sec	accuracy=0.561345	lr=0.282385
Epoch[46] Batch [1249]	Speed: 1499.576975 samples/sec	accuracy=0.561220	lr=0.282186
[Epoch 46] training: accuracy=0.561241
[Epoch 46] speed: 1504 samples/sec	time cost: 872.836818
[Epoch 46] validation: err-top1=0.379985 err-top5=0.153870
Epoch[47] Batch [49]	Speed: 1535.151482 samples/sec	accuracy=0.572187	lr=0.281983
Epoch[47] Batch [99]	Speed: 1451.840363 samples/sec	accuracy=0.569541	lr=0.281784
Epoch[47] Batch [149]	Speed: 1523.881803 samples/sec	accuracy=0.568594	lr=0.281585
Epoch[47] Batch [199]	Speed: 1535.144766 samples/sec	accuracy=0.568389	lr=0.281385
Epoch[47] Batch [249]	Speed: 1498.897120 samples/sec	accuracy=0.568023	lr=0.281186
Epoch[47] Batch [299]	Speed: 1514.714531 samples/sec	accuracy=0.568584	lr=0.280986
Epoch[47] Batch [349]	Speed: 1518.655585 samples/sec	accuracy=0.568613	lr=0.280786
Epoch[47] Batch [399]	Speed: 1481.659778 samples/sec	accuracy=0.567524	lr=0.280587
Epoch[47] Batch [449]	Speed: 1512.109880 samples/sec	accuracy=0.567409	lr=0.280387
Epoch[47] Batch [499]	Speed: 1534.031166 samples/sec	accuracy=0.567004	lr=0.280187
Epoch[47] Batch [549]	Speed: 1502.760154 samples/sec	accuracy=0.566404	lr=0.279987
Epoch[47] Batch [599]	Speed: 1541.755772 samples/sec	accuracy=0.566453	lr=0.279786
Epoch[47] Batch [649]	Speed: 1538.660240 samples/sec	accuracy=0.565913	lr=0.279586
Epoch[47] Batch [699]	Speed: 1407.777697 samples/sec	accuracy=0.565241	lr=0.279386
Epoch[47] Batch [749]	Speed: 1497.695970 samples/sec	accuracy=0.565122	lr=0.279185
Epoch[47] Batch [799]	Speed: 1517.655720 samples/sec	accuracy=0.564900	lr=0.278985
Epoch[47] Batch [849]	Speed: 1511.339530 samples/sec	accuracy=0.564744	lr=0.278784
Epoch[47] Batch [899]	Speed: 1526.650254 samples/sec	accuracy=0.564690	lr=0.278583
Epoch[47] Batch [949]	Speed: 1511.589431 samples/sec	accuracy=0.564529	lr=0.278382
Epoch[47] Batch [999]	Speed: 1508.251631 samples/sec	accuracy=0.564276	lr=0.278181
Epoch[47] Batch [1049]	Speed: 1541.874893 samples/sec	accuracy=0.564292	lr=0.277980
Epoch[47] Batch [1099]	Speed: 1483.466946 samples/sec	accuracy=0.563929	lr=0.277779
Epoch[47] Batch [1149]	Speed: 1510.732192 samples/sec	accuracy=0.563885	lr=0.277578
Epoch[47] Batch [1199]	Speed: 1538.394664 samples/sec	accuracy=0.563584	lr=0.277377
Epoch[47] Batch [1249]	Speed: 1532.333203 samples/sec	accuracy=0.563144	lr=0.277175
[Epoch 47] training: accuracy=0.563135
[Epoch 47] speed: 1509 samples/sec	time cost: 870.625100
[Epoch 47] validation: err-top1=0.368782 err-top5=0.145747
Epoch[48] Batch [49]	Speed: 1514.988068 samples/sec	accuracy=0.576934	lr=0.276970
Epoch[48] Batch [99]	Speed: 1503.094647 samples/sec	accuracy=0.574150	lr=0.276768
Epoch[48] Batch [149]	Speed: 1497.376832 samples/sec	accuracy=0.572819	lr=0.276566
Epoch[48] Batch [199]	Speed: 1515.291256 samples/sec	accuracy=0.572134	lr=0.276365
Epoch[48] Batch [249]	Speed: 1474.836855 samples/sec	accuracy=0.571559	lr=0.276163
Epoch[48] Batch [299]	Speed: 1531.045841 samples/sec	accuracy=0.571305	lr=0.275961
Epoch[48] Batch [349]	Speed: 1524.380660 samples/sec	accuracy=0.570594	lr=0.275759
Epoch[48] Batch [399]	Speed: 1520.051976 samples/sec	accuracy=0.570071	lr=0.275557
Epoch[48] Batch [449]	Speed: 1521.455271 samples/sec	accuracy=0.569666	lr=0.275354
Epoch[48] Batch [499]	Speed: 1520.355247 samples/sec	accuracy=0.569285	lr=0.275152
Epoch[48] Batch [549]	Speed: 1477.969618 samples/sec	accuracy=0.568700	lr=0.274950
Epoch[48] Batch [599]	Speed: 1538.041644 samples/sec	accuracy=0.568446	lr=0.274747
Epoch[48] Batch [649]	Speed: 1517.391940 samples/sec	accuracy=0.567978	lr=0.274545
Epoch[48] Batch [699]	Speed: 1518.504193 samples/sec	accuracy=0.567704	lr=0.274342
Epoch[48] Batch [749]	Speed: 1494.555863 samples/sec	accuracy=0.567497	lr=0.274139
Epoch[48] Batch [799]	Speed: 1486.647027 samples/sec	accuracy=0.567291	lr=0.273936
Epoch[48] Batch [849]	Speed: 1541.700552 samples/sec	accuracy=0.567179	lr=0.273733
Epoch[48] Batch [899]	Speed: 1513.387391 samples/sec	accuracy=0.566998	lr=0.273530
Epoch[48] Batch [949]	Speed: 1464.958361 samples/sec	accuracy=0.566529	lr=0.273327
Epoch[48] Batch [999]	Speed: 1514.944986 samples/sec	accuracy=0.566479	lr=0.273124
Epoch[48] Batch [1049]	Speed: 1514.443239 samples/sec	accuracy=0.566303	lr=0.272921
Epoch[48] Batch [1099]	Speed: 1497.536750 samples/sec	accuracy=0.565900	lr=0.272717
Epoch[48] Batch [1149]	Speed: 1510.539884 samples/sec	accuracy=0.565830	lr=0.272514
Epoch[48] Batch [1199]	Speed: 1505.285786 samples/sec	accuracy=0.565695	lr=0.272310
Epoch[48] Batch [1249]	Speed: 1497.685379 samples/sec	accuracy=0.565327	lr=0.272107
[Epoch 48] training: accuracy=0.565343
[Epoch 48] speed: 1507 samples/sec	time cost: 871.896617
[Epoch 48] validation: err-top1=0.371612 err-top5=0.146883
Epoch[49] Batch [49]	Speed: 1444.187889 samples/sec	accuracy=0.576758	lr=0.271899
Epoch[49] Batch [99]	Speed: 1526.233418 samples/sec	accuracy=0.575957	lr=0.271695
Epoch[49] Batch [149]	Speed: 1548.676419 samples/sec	accuracy=0.575716	lr=0.271491
Epoch[49] Batch [199]	Speed: 1458.545292 samples/sec	accuracy=0.575103	lr=0.271287
Epoch[49] Batch [249]	Speed: 1537.557047 samples/sec	accuracy=0.574629	lr=0.271083
Epoch[49] Batch [299]	Speed: 1468.224530 samples/sec	accuracy=0.573750	lr=0.270879
Epoch[49] Batch [349]	Speed: 1484.253268 samples/sec	accuracy=0.573248	lr=0.270675
Epoch[49] Batch [399]	Speed: 1514.352154 samples/sec	accuracy=0.572671	lr=0.270470
Epoch[49] Batch [449]	Speed: 1523.735844 samples/sec	accuracy=0.572164	lr=0.270266
Epoch[49] Batch [499]	Speed: 1541.860734 samples/sec	accuracy=0.571898	lr=0.270061
Epoch[49] Batch [549]	Speed: 1448.326007 samples/sec	accuracy=0.571605	lr=0.269857
Epoch[49] Batch [599]	Speed: 1507.477819 samples/sec	accuracy=0.571717	lr=0.269652
Epoch[49] Batch [649]	Speed: 1534.662694 samples/sec	accuracy=0.571303	lr=0.269447
Epoch[49] Batch [699]	Speed: 1482.949885 samples/sec	accuracy=0.571327	lr=0.269243
Epoch[49] Batch [749]	Speed: 1519.890667 samples/sec	accuracy=0.571255	lr=0.269038
Epoch[49] Batch [799]	Speed: 1519.566582 samples/sec	accuracy=0.571239	lr=0.268833
Epoch[49] Batch [849]	Speed: 1487.489085 samples/sec	accuracy=0.570742	lr=0.268628
Epoch[49] Batch [899]	Speed: 1495.879913 samples/sec	accuracy=0.570280	lr=0.268422
Epoch[49] Batch [949]	Speed: 1471.207129 samples/sec	accuracy=0.570076	lr=0.268217
Epoch[49] Batch [999]	Speed: 1506.311939 samples/sec	accuracy=0.569777	lr=0.268012
Epoch[49] Batch [1049]	Speed: 1520.717229 samples/sec	accuracy=0.569687	lr=0.267807
Epoch[49] Batch [1099]	Speed: 1539.664702 samples/sec	accuracy=0.569387	lr=0.267601
Epoch[49] Batch [1149]	Speed: 1514.031088 samples/sec	accuracy=0.569352	lr=0.267396
Epoch[49] Batch [1199]	Speed: 1534.536461 samples/sec	accuracy=0.569239	lr=0.267190
Epoch[49] Batch [1249]	Speed: 1489.576436 samples/sec	accuracy=0.568909	lr=0.266984
[Epoch 49] training: accuracy=0.568899
[Epoch 49] speed: 1503 samples/sec	time cost: 873.916865
[Epoch 49] validation: err-top1=0.369161 err-top5=0.144790
Epoch[50] Batch [49]	Speed: 1484.806451 samples/sec	accuracy=0.579570	lr=0.266774
Epoch[50] Batch [99]	Speed: 1545.130310 samples/sec	accuracy=0.575791	lr=0.266568
Epoch[50] Batch [149]	Speed: 1539.717613 samples/sec	accuracy=0.575208	lr=0.266362
Epoch[50] Batch [199]	Speed: 1521.335749 samples/sec	accuracy=0.574634	lr=0.266156
Epoch[50] Batch [249]	Speed: 1508.177812 samples/sec	accuracy=0.573656	lr=0.265950
Epoch[50] Batch [299]	Speed: 1487.455641 samples/sec	accuracy=0.573275	lr=0.265744
Epoch[50] Batch [349]	Speed: 1521.922328 samples/sec	accuracy=0.573105	lr=0.265538
Epoch[50] Batch [399]	Speed: 1510.751758 samples/sec	accuracy=0.573054	lr=0.265331
Epoch[50] Batch [449]	Speed: 1525.611452 samples/sec	accuracy=0.573409	lr=0.265125
Epoch[50] Batch [499]	Speed: 1512.637741 samples/sec	accuracy=0.573316	lr=0.264918
Epoch[50] Batch [549]	Speed: 1497.269373 samples/sec	accuracy=0.572864	lr=0.264712
Epoch[50] Batch [599]	Speed: 1492.674684 samples/sec	accuracy=0.571823	lr=0.264505
Epoch[50] Batch [649]	Speed: 1481.060630 samples/sec	accuracy=0.571618	lr=0.264298
Epoch[50] Batch [699]	Speed: 1500.738679 samples/sec	accuracy=0.571776	lr=0.264092
Epoch[50] Batch [749]	Speed: 1537.010344 samples/sec	accuracy=0.571936	lr=0.263885
Epoch[50] Batch [799]	Speed: 1538.592289 samples/sec	accuracy=0.571730	lr=0.263678
Epoch[50] Batch [849]	Speed: 1507.620204 samples/sec	accuracy=0.571604	lr=0.263471
Epoch[50] Batch [899]	Speed: 1516.199997 samples/sec	accuracy=0.571383	lr=0.263264
Epoch[50] Batch [949]	Speed: 1475.973341 samples/sec	accuracy=0.571226	lr=0.263056
Epoch[50] Batch [999]	Speed: 1513.549488 samples/sec	accuracy=0.571296	lr=0.262849
Epoch[50] Batch [1049]	Speed: 1513.807067 samples/sec	accuracy=0.570990	lr=0.262642
Epoch[50] Batch [1099]	Speed: 1520.540846 samples/sec	accuracy=0.570742	lr=0.262434
Epoch[50] Batch [1149]	Speed: 1541.237162 samples/sec	accuracy=0.570461	lr=0.262227
Epoch[50] Batch [1199]	Speed: 1526.618325 samples/sec	accuracy=0.570413	lr=0.262019
Epoch[50] Batch [1249]	Speed: 1528.625902 samples/sec	accuracy=0.570292	lr=0.261812
[Epoch 50] training: accuracy=0.570314
[Epoch 50] speed: 1512 samples/sec	time cost: 869.429391
[Epoch 50] validation: err-top1=0.361507 err-top5=0.141621
Epoch[51] Batch [49]	Speed: 1445.366998 samples/sec	accuracy=0.583613	lr=0.261600
Epoch[51] Batch [99]	Speed: 1523.136382 samples/sec	accuracy=0.582178	lr=0.261392
Epoch[51] Batch [149]	Speed: 1498.761137 samples/sec	accuracy=0.581211	lr=0.261184
Epoch[51] Batch [199]	Speed: 1494.301372 samples/sec	accuracy=0.580825	lr=0.260976
Epoch[51] Batch [249]	Speed: 1504.267637 samples/sec	accuracy=0.579305	lr=0.260768
Epoch[51] Batch [299]	Speed: 1537.668770 samples/sec	accuracy=0.578760	lr=0.260560
Epoch[51] Batch [349]	Speed: 1534.967357 samples/sec	accuracy=0.577891	lr=0.260352
Epoch[51] Batch [399]	Speed: 1470.233315 samples/sec	accuracy=0.577473	lr=0.260144
Epoch[51] Batch [449]	Speed: 1507.640907 samples/sec	accuracy=0.577196	lr=0.259935
Epoch[51] Batch [499]	Speed: 1494.594952 samples/sec	accuracy=0.576979	lr=0.259727
Epoch[51] Batch [549]	Speed: 1518.969741 samples/sec	accuracy=0.575996	lr=0.259519
Epoch[51] Batch [599]	Speed: 1524.101417 samples/sec	accuracy=0.575513	lr=0.259310
Epoch[51] Batch [649]	Speed: 1505.117331 samples/sec	accuracy=0.575319	lr=0.259102
Epoch[51] Batch [699]	Speed: 1506.942671 samples/sec	accuracy=0.575321	lr=0.258893
Epoch[51] Batch [749]	Speed: 1540.420414 samples/sec	accuracy=0.575096	lr=0.258684
Epoch[51] Batch [799]	Speed: 1490.688118 samples/sec	accuracy=0.574807	lr=0.258475
Epoch[51] Batch [849]	Speed: 1495.299531 samples/sec	accuracy=0.574555	lr=0.258267
Epoch[51] Batch [899]	Speed: 1484.411236 samples/sec	accuracy=0.574287	lr=0.258058
Epoch[51] Batch [949]	Speed: 1506.805066 samples/sec	accuracy=0.573735	lr=0.257849
Epoch[51] Batch [999]	Speed: 1505.714958 samples/sec	accuracy=0.573413	lr=0.257640
Epoch[51] Batch [1049]	Speed: 1519.131564 samples/sec	accuracy=0.573465	lr=0.257430
Epoch[51] Batch [1099]	Speed: 1545.731714 samples/sec	accuracy=0.573405	lr=0.257221
Epoch[51] Batch [1149]	Speed: 1506.835675 samples/sec	accuracy=0.573123	lr=0.257012
Epoch[51] Batch [1199]	Speed: 1521.557842 samples/sec	accuracy=0.572986	lr=0.256803
Epoch[51] Batch [1249]	Speed: 1528.640918 samples/sec	accuracy=0.572720	lr=0.256593
[Epoch 51] training: accuracy=0.572733
[Epoch 51] speed: 1506 samples/sec	time cost: 872.413465
[Epoch 51] validation: err-top1=0.367168 err-top5=0.144173
Epoch[52] Batch [49]	Speed: 1451.102552 samples/sec	accuracy=0.587617	lr=0.256379
Epoch[52] Batch [99]	Speed: 1534.444751 samples/sec	accuracy=0.584980	lr=0.256170
Epoch[52] Batch [149]	Speed: 1488.240928 samples/sec	accuracy=0.583197	lr=0.255960
Epoch[52] Batch [199]	Speed: 1511.910558 samples/sec	accuracy=0.581641	lr=0.255751
Epoch[52] Batch [249]	Speed: 1490.622920 samples/sec	accuracy=0.581723	lr=0.255541
Epoch[52] Batch [299]	Speed: 1502.708491 samples/sec	accuracy=0.580446	lr=0.255331
Epoch[52] Batch [349]	Speed: 1512.541088 samples/sec	accuracy=0.579727	lr=0.255121
Epoch[52] Batch [399]	Speed: 1534.054617 samples/sec	accuracy=0.579111	lr=0.254911
Epoch[52] Batch [449]	Speed: 1519.932288 samples/sec	accuracy=0.579210	lr=0.254701
Epoch[52] Batch [499]	Speed: 1492.499757 samples/sec	accuracy=0.578172	lr=0.254491
Epoch[52] Batch [549]	Speed: 1542.938974 samples/sec	accuracy=0.577379	lr=0.254281
Epoch[52] Batch [599]	Speed: 1510.337014 samples/sec	accuracy=0.577300	lr=0.254071
Epoch[52] Batch [649]	Speed: 1476.835867 samples/sec	accuracy=0.577377	lr=0.253861
Epoch[52] Batch [699]	Speed: 1476.392089 samples/sec	accuracy=0.577181	lr=0.253650
Epoch[52] Batch [749]	Speed: 1514.638860 samples/sec	accuracy=0.576785	lr=0.253440
Epoch[52] Batch [799]	Speed: 1488.486466 samples/sec	accuracy=0.576643	lr=0.253229
Epoch[52] Batch [849]	Speed: 1510.299752 samples/sec	accuracy=0.576585	lr=0.253019
Epoch[52] Batch [899]	Speed: 1517.150471 samples/sec	accuracy=0.576211	lr=0.252808
Epoch[52] Batch [949]	Speed: 1449.313681 samples/sec	accuracy=0.575939	lr=0.252598
Epoch[52] Batch [999]	Speed: 1502.305203 samples/sec	accuracy=0.575705	lr=0.252387
Epoch[52] Batch [1049]	Speed: 1463.969945 samples/sec	accuracy=0.575237	lr=0.252176
Epoch[52] Batch [1099]	Speed: 1545.717339 samples/sec	accuracy=0.574992	lr=0.251965
Epoch[52] Batch [1149]	Speed: 1481.354150 samples/sec	accuracy=0.574832	lr=0.251754
Epoch[52] Batch [1199]	Speed: 1515.488337 samples/sec	accuracy=0.574618	lr=0.251543
Epoch[52] Batch [1249]	Speed: 1545.326801 samples/sec	accuracy=0.574482	lr=0.251332
[Epoch 52] training: accuracy=0.574493
[Epoch 52] speed: 1501 samples/sec	time cost: 874.760379
[Epoch 52] validation: err-top1=0.365560 err-top5=0.142782
Epoch[53] Batch [49]	Speed: 1502.839586 samples/sec	accuracy=0.592285	lr=0.251117
Epoch[53] Batch [99]	Speed: 1508.781582 samples/sec	accuracy=0.588838	lr=0.250906
Epoch[53] Batch [149]	Speed: 1513.872410 samples/sec	accuracy=0.586426	lr=0.250695
Epoch[53] Batch [199]	Speed: 1521.591533 samples/sec	accuracy=0.585361	lr=0.250483
Epoch[53] Batch [249]	Speed: 1517.143773 samples/sec	accuracy=0.583816	lr=0.250272
Epoch[53] Batch [299]	Speed: 1517.117267 samples/sec	accuracy=0.582926	lr=0.250061
Epoch[53] Batch [349]	Speed: 1499.442711 samples/sec	accuracy=0.582977	lr=0.249849
Epoch[53] Batch [399]	Speed: 1523.722362 samples/sec	accuracy=0.582388	lr=0.249638
Epoch[53] Batch [449]	Speed: 1510.012881 samples/sec	accuracy=0.582101	lr=0.249426
Epoch[53] Batch [499]	Speed: 1493.308583 samples/sec	accuracy=0.581330	lr=0.249215
Epoch[53] Batch [549]	Speed: 1478.928777 samples/sec	accuracy=0.581254	lr=0.249003
Epoch[53] Batch [599]	Speed: 1536.998596 samples/sec	accuracy=0.580534	lr=0.248791
Epoch[53] Batch [649]	Speed: 1511.509550 samples/sec	accuracy=0.580068	lr=0.248579
Epoch[53] Batch [699]	Speed: 1516.512227 samples/sec	accuracy=0.579855	lr=0.248367
Epoch[53] Batch [749]	Speed: 1514.489977 samples/sec	accuracy=0.579616	lr=0.248156
Epoch[53] Batch [799]	Speed: 1477.183049 samples/sec	accuracy=0.579354	lr=0.247944
Epoch[53] Batch [849]	Speed: 1490.395946 samples/sec	accuracy=0.578917	lr=0.247732
Epoch[53] Batch [899]	Speed: 1515.399906 samples/sec	accuracy=0.578894	lr=0.247519
Epoch[53] Batch [949]	Speed: 1515.631768 samples/sec	accuracy=0.578876	lr=0.247307
Epoch[53] Batch [999]	Speed: 1494.524097 samples/sec	accuracy=0.578682	lr=0.247095
Epoch[53] Batch [1049]	Speed: 1500.517662 samples/sec	accuracy=0.578379	lr=0.246883
Epoch[53] Batch [1099]	Speed: 1528.214675 samples/sec	accuracy=0.578224	lr=0.246670
Epoch[53] Batch [1149]	Speed: 1493.158392 samples/sec	accuracy=0.578118	lr=0.246458
Epoch[53] Batch [1199]	Speed: 1439.112999 samples/sec	accuracy=0.578289	lr=0.246246
Epoch[53] Batch [1249]	Speed: 1550.440008 samples/sec	accuracy=0.578113	lr=0.246033
[Epoch 53] training: accuracy=0.578109
[Epoch 53] speed: 1505 samples/sec	time cost: 873.326359
[Epoch 53] validation: err-top1=0.366390 err-top5=0.143515
Epoch[54] Batch [49]	Speed: 1497.809696 samples/sec	accuracy=0.590762	lr=0.245816
Epoch[54] Batch [99]	Speed: 1531.936524 samples/sec	accuracy=0.586895	lr=0.245604
Epoch[54] Batch [149]	Speed: 1536.187102 samples/sec	accuracy=0.589323	lr=0.245391
Epoch[54] Batch [199]	Speed: 1488.671538 samples/sec	accuracy=0.586704	lr=0.245178
Epoch[54] Batch [249]	Speed: 1495.537521 samples/sec	accuracy=0.586621	lr=0.244966
Epoch[54] Batch [299]	Speed: 1519.858332 samples/sec	accuracy=0.585804	lr=0.244753
Epoch[54] Batch [349]	Speed: 1494.548405 samples/sec	accuracy=0.584693	lr=0.244540
Epoch[54] Batch [399]	Speed: 1510.297234 samples/sec	accuracy=0.584021	lr=0.244327
Epoch[54] Batch [449]	Speed: 1544.511501 samples/sec	accuracy=0.583314	lr=0.244114
Epoch[54] Batch [499]	Speed: 1518.608461 samples/sec	accuracy=0.583186	lr=0.243901
Epoch[54] Batch [549]	Speed: 1478.575887 samples/sec	accuracy=0.582244	lr=0.243688
Epoch[54] Batch [599]	Speed: 1455.253289 samples/sec	accuracy=0.582041	lr=0.243475
Epoch[54] Batch [649]	Speed: 1513.334280 samples/sec	accuracy=0.581596	lr=0.243262
Epoch[54] Batch [699]	Speed: 1540.576971 samples/sec	accuracy=0.581237	lr=0.243049
Epoch[54] Batch [749]	Speed: 1524.880753 samples/sec	accuracy=0.581176	lr=0.242835
Epoch[54] Batch [799]	Speed: 1465.086670 samples/sec	accuracy=0.580760	lr=0.242622
Epoch[54] Batch [849]	Speed: 1534.824642 samples/sec	accuracy=0.580787	lr=0.242409
Epoch[54] Batch [899]	Speed: 1477.454918 samples/sec	accuracy=0.580418	lr=0.242195
Epoch[54] Batch [949]	Speed: 1510.099781 samples/sec	accuracy=0.580213	lr=0.241982
Epoch[54] Batch [999]	Speed: 1487.766904 samples/sec	accuracy=0.580098	lr=0.241768
Epoch[54] Batch [1049]	Speed: 1501.657628 samples/sec	accuracy=0.579953	lr=0.241555
Epoch[54] Batch [1099]	Speed: 1523.654620 samples/sec	accuracy=0.579888	lr=0.241341
Epoch[54] Batch [1149]	Speed: 1459.401260 samples/sec	accuracy=0.579772	lr=0.241127
Epoch[54] Batch [1199]	Speed: 1513.679835 samples/sec	accuracy=0.579670	lr=0.240914
Epoch[54] Batch [1249]	Speed: 1537.067297 samples/sec	accuracy=0.579258	lr=0.240700
[Epoch 54] training: accuracy=0.579269
[Epoch 54] speed: 1504 samples/sec	time cost: 873.598190
[Epoch 54] validation: err-top1=0.356266 err-top5=0.136181
Epoch[55] Batch [49]	Speed: 1506.840010 samples/sec	accuracy=0.593516	lr=0.240482
Epoch[55] Batch [99]	Speed: 1507.041878 samples/sec	accuracy=0.592031	lr=0.240268
Epoch[55] Batch [149]	Speed: 1539.178022 samples/sec	accuracy=0.590254	lr=0.240054
Epoch[55] Batch [199]	Speed: 1466.813656 samples/sec	accuracy=0.588643	lr=0.239840
Epoch[55] Batch [249]	Speed: 1520.067976 samples/sec	accuracy=0.587641	lr=0.239626
Epoch[55] Batch [299]	Speed: 1454.258934 samples/sec	accuracy=0.587174	lr=0.239412
Epoch[55] Batch [349]	Speed: 1520.545852 samples/sec	accuracy=0.586984	lr=0.239198
Epoch[55] Batch [399]	Speed: 1511.265696 samples/sec	accuracy=0.586721	lr=0.238984
Epoch[55] Batch [449]	Speed: 1532.954295 samples/sec	accuracy=0.586638	lr=0.238769
Epoch[55] Batch [499]	Speed: 1515.151994 samples/sec	accuracy=0.586352	lr=0.238555
Epoch[55] Batch [549]	Speed: 1486.592514 samples/sec	accuracy=0.585859	lr=0.238341
Epoch[55] Batch [599]	Speed: 1530.804449 samples/sec	accuracy=0.585417	lr=0.238126
Epoch[55] Batch [649]	Speed: 1451.507433 samples/sec	accuracy=0.585213	lr=0.237912
Epoch[55] Batch [699]	Speed: 1551.146865 samples/sec	accuracy=0.584897	lr=0.237698
Epoch[55] Batch [749]	Speed: 1517.524762 samples/sec	accuracy=0.584555	lr=0.237483
Epoch[55] Batch [799]	Speed: 1479.043073 samples/sec	accuracy=0.584031	lr=0.237269
Epoch[55] Batch [849]	Speed: 1533.779750 samples/sec	accuracy=0.583794	lr=0.237054
Epoch[55] Batch [899]	Speed: 1450.433836 samples/sec	accuracy=0.583248	lr=0.236839
Epoch[55] Batch [949]	Speed: 1538.102431 samples/sec	accuracy=0.583029	lr=0.236625
Epoch[55] Batch [999]	Speed: 1495.198106 samples/sec	accuracy=0.582881	lr=0.236410
Epoch[55] Batch [1049]	Speed: 1536.253159 samples/sec	accuracy=0.582993	lr=0.236195
Epoch[55] Batch [1099]	Speed: 1517.237638 samples/sec	accuracy=0.582847	lr=0.235981
Epoch[55] Batch [1149]	Speed: 1514.401342 samples/sec	accuracy=0.583016	lr=0.235766
Epoch[55] Batch [1199]	Speed: 1511.102695 samples/sec	accuracy=0.582687	lr=0.235551
Epoch[55] Batch [1249]	Speed: 1543.160389 samples/sec	accuracy=0.582277	lr=0.235336
[Epoch 55] training: accuracy=0.582261
[Epoch 55] speed: 1507 samples/sec	time cost: 871.698731
[Epoch 55] validation: err-top1=0.367188 err-top5=0.141781
Epoch[56] Batch [49]	Speed: 1507.373572 samples/sec	accuracy=0.590508	lr=0.235117
Epoch[56] Batch [99]	Speed: 1482.467292 samples/sec	accuracy=0.591260	lr=0.234902
Epoch[56] Batch [149]	Speed: 1521.041040 samples/sec	accuracy=0.589753	lr=0.234687
Epoch[56] Batch [199]	Speed: 1549.469828 samples/sec	accuracy=0.589741	lr=0.234472
Epoch[56] Batch [249]	Speed: 1475.958926 samples/sec	accuracy=0.588457	lr=0.234256
Epoch[56] Batch [299]	Speed: 1463.976561 samples/sec	accuracy=0.588118	lr=0.234041
Epoch[56] Batch [349]	Speed: 1511.604997 samples/sec	accuracy=0.588044	lr=0.233826
Epoch[56] Batch [399]	Speed: 1466.661985 samples/sec	accuracy=0.587935	lr=0.233611
Epoch[56] Batch [449]	Speed: 1534.387916 samples/sec	accuracy=0.587689	lr=0.233396
Epoch[56] Batch [499]	Speed: 1540.381001 samples/sec	accuracy=0.587773	lr=0.233180
Epoch[56] Batch [549]	Speed: 1523.458739 samples/sec	accuracy=0.587781	lr=0.232965
Epoch[56] Batch [599]	Speed: 1450.311822 samples/sec	accuracy=0.587370	lr=0.232749
Epoch[56] Batch [649]	Speed: 1514.431085 samples/sec	accuracy=0.587431	lr=0.232534
Epoch[56] Batch [699]	Speed: 1532.949797 samples/sec	accuracy=0.587028	lr=0.232319
Epoch[56] Batch [749]	Speed: 1475.432264 samples/sec	accuracy=0.586510	lr=0.232103
Epoch[56] Batch [799]	Speed: 1534.227211 samples/sec	accuracy=0.586165	lr=0.231887
Epoch[56] Batch [849]	Speed: 1505.437245 samples/sec	accuracy=0.585682	lr=0.231672
Epoch[56] Batch [899]	Speed: 1526.136354 samples/sec	accuracy=0.585414	lr=0.231456
Epoch[56] Batch [949]	Speed: 1535.803472 samples/sec	accuracy=0.585344	lr=0.231241
Epoch[56] Batch [999]	Speed: 1529.954451 samples/sec	accuracy=0.585271	lr=0.231025
Epoch[56] Batch [1049]	Speed: 1510.922698 samples/sec	accuracy=0.585222	lr=0.230809
Epoch[56] Batch [1099]	Speed: 1524.586228 samples/sec	accuracy=0.585089	lr=0.230593
Epoch[56] Batch [1149]	Speed: 1490.648767 samples/sec	accuracy=0.584931	lr=0.230377
Epoch[56] Batch [1199]	Speed: 1485.933272 samples/sec	accuracy=0.584943	lr=0.230162
Epoch[56] Batch [1249]	Speed: 1472.491266 samples/sec	accuracy=0.584902	lr=0.229946
[Epoch 56] training: accuracy=0.584884
[Epoch 56] speed: 1504 samples/sec	time cost: 873.349063
[Epoch 56] validation: err-top1=0.369858 err-top5=0.145647
Epoch[57] Batch [49]	Speed: 1516.427628 samples/sec	accuracy=0.596602	lr=0.229725
Epoch[57] Batch [99]	Speed: 1539.671314 samples/sec	accuracy=0.595732	lr=0.229510
Epoch[57] Batch [149]	Speed: 1518.022781 samples/sec	accuracy=0.592441	lr=0.229294
Epoch[57] Batch [199]	Speed: 1472.807782 samples/sec	accuracy=0.591792	lr=0.229077
Epoch[57] Batch [249]	Speed: 1536.883472 samples/sec	accuracy=0.591961	lr=0.228861
Epoch[57] Batch [299]	Speed: 1486.085489 samples/sec	accuracy=0.591081	lr=0.228645
Epoch[57] Batch [349]	Speed: 1482.354278 samples/sec	accuracy=0.590893	lr=0.228429
Epoch[57] Batch [399]	Speed: 1513.564647 samples/sec	accuracy=0.590073	lr=0.228213
Epoch[57] Batch [449]	Speed: 1516.417659 samples/sec	accuracy=0.590211	lr=0.227997
Epoch[57] Batch [499]	Speed: 1510.950636 samples/sec	accuracy=0.589854	lr=0.227781
Epoch[57] Batch [549]	Speed: 1495.889208 samples/sec	accuracy=0.589430	lr=0.227564
Epoch[57] Batch [599]	Speed: 1507.827057 samples/sec	accuracy=0.588963	lr=0.227348
Epoch[57] Batch [649]	Speed: 1516.380396 samples/sec	accuracy=0.588840	lr=0.227132
Epoch[57] Batch [699]	Speed: 1498.294731 samples/sec	accuracy=0.588781	lr=0.226915
Epoch[57] Batch [749]	Speed: 1517.100665 samples/sec	accuracy=0.588609	lr=0.226699
Epoch[57] Batch [799]	Speed: 1542.722010 samples/sec	accuracy=0.588536	lr=0.226482
Epoch[57] Batch [849]	Speed: 1536.923421 samples/sec	accuracy=0.588370	lr=0.226266
Epoch[57] Batch [899]	Speed: 1520.345183 samples/sec	accuracy=0.588270	lr=0.226049
Epoch[57] Batch [949]	Speed: 1495.676107 samples/sec	accuracy=0.588044	lr=0.225833
Epoch[57] Batch [999]	Speed: 1540.599650 samples/sec	accuracy=0.587735	lr=0.225616
Epoch[57] Batch [1049]	Speed: 1508.355216 samples/sec	accuracy=0.587689	lr=0.225400
Epoch[57] Batch [1099]	Speed: 1515.243111 samples/sec	accuracy=0.587488	lr=0.225183
Epoch[57] Batch [1149]	Speed: 1479.681961 samples/sec	accuracy=0.587352	lr=0.224967
Epoch[57] Batch [1199]	Speed: 1534.890451 samples/sec	accuracy=0.587266	lr=0.224750
Epoch[57] Batch [1249]	Speed: 1503.806286 samples/sec	accuracy=0.587199	lr=0.224533
[Epoch 57] training: accuracy=0.587226
[Epoch 57] speed: 1510 samples/sec	time cost: 870.582093
[Epoch 57] validation: err-top1=0.359833 err-top5=0.139090
Epoch[58] Batch [49]	Speed: 1494.109160 samples/sec	accuracy=0.599043	lr=0.224312
Epoch[58] Batch [99]	Speed: 1502.915439 samples/sec	accuracy=0.598086	lr=0.224095
Epoch[58] Batch [149]	Speed: 1519.788579 samples/sec	accuracy=0.594115	lr=0.223879
Epoch[58] Batch [199]	Speed: 1482.958139 samples/sec	accuracy=0.594043	lr=0.223662
Epoch[58] Batch [249]	Speed: 1535.402712 samples/sec	accuracy=0.594016	lr=0.223445
Epoch[58] Batch [299]	Speed: 1541.945371 samples/sec	accuracy=0.593763	lr=0.223228
Epoch[58] Batch [349]	Speed: 1495.585953 samples/sec	accuracy=0.593664	lr=0.223011
Epoch[58] Batch [399]	Speed: 1506.160980 samples/sec	accuracy=0.593784	lr=0.222794
Epoch[58] Batch [449]	Speed: 1529.303670 samples/sec	accuracy=0.593724	lr=0.222577
Epoch[58] Batch [499]	Speed: 1494.580421 samples/sec	accuracy=0.593432	lr=0.222360
Epoch[58] Batch [549]	Speed: 1536.401593 samples/sec	accuracy=0.593208	lr=0.222143
Epoch[58] Batch [599]	Speed: 1507.610499 samples/sec	accuracy=0.592720	lr=0.221926
Epoch[58] Batch [649]	Speed: 1494.809108 samples/sec	accuracy=0.591983	lr=0.221709
Epoch[58] Batch [699]	Speed: 1473.542589 samples/sec	accuracy=0.591611	lr=0.221492
Epoch[58] Batch [749]	Speed: 1537.605695 samples/sec	accuracy=0.591569	lr=0.221275
Epoch[58] Batch [799]	Speed: 1533.801319 samples/sec	accuracy=0.591370	lr=0.221058
Epoch[58] Batch [849]	Speed: 1460.088745 samples/sec	accuracy=0.591071	lr=0.220841
Epoch[58] Batch [899]	Speed: 1511.461816 samples/sec	accuracy=0.590809	lr=0.220623
Epoch[58] Batch [949]	Speed: 1540.000973 samples/sec	accuracy=0.590572	lr=0.220406
Epoch[58] Batch [999]	Speed: 1502.845907 samples/sec	accuracy=0.590397	lr=0.220189
Epoch[58] Batch [1049]	Speed: 1512.105099 samples/sec	accuracy=0.590187	lr=0.219972
Epoch[58] Batch [1099]	Speed: 1497.379390 samples/sec	accuracy=0.590052	lr=0.219754
Epoch[58] Batch [1149]	Speed: 1485.880343 samples/sec	accuracy=0.589939	lr=0.219537
Epoch[58] Batch [1199]	Speed: 1535.674943 samples/sec	accuracy=0.589717	lr=0.219320
Epoch[58] Batch [1249]	Speed: 1550.260893 samples/sec	accuracy=0.589570	lr=0.219102
[Epoch 58] training: accuracy=0.589576
[Epoch 58] speed: 1509 samples/sec	time cost: 870.418873
[Epoch 58] validation: err-top1=0.361003 err-top5=0.138163
Epoch[59] Batch [49]	Speed: 1495.940122 samples/sec	accuracy=0.605195	lr=0.218881
Epoch[59] Batch [99]	Speed: 1525.752947 samples/sec	accuracy=0.602002	lr=0.218663
Epoch[59] Batch [149]	Speed: 1535.098369 samples/sec	accuracy=0.599707	lr=0.218446
Epoch[59] Batch [199]	Speed: 1538.670129 samples/sec	accuracy=0.598149	lr=0.218228
Epoch[59] Batch [249]	Speed: 1488.860155 samples/sec	accuracy=0.597902	lr=0.218011
Epoch[59] Batch [299]	Speed: 1479.523030 samples/sec	accuracy=0.596875	lr=0.217793
Epoch[59] Batch [349]	Speed: 1509.129006 samples/sec	accuracy=0.595977	lr=0.217576
Epoch[59] Batch [399]	Speed: 1504.522066 samples/sec	accuracy=0.594600	lr=0.217358
Epoch[59] Batch [449]	Speed: 1512.079153 samples/sec	accuracy=0.594618	lr=0.217141
Epoch[59] Batch [499]	Speed: 1449.220188 samples/sec	accuracy=0.594182	lr=0.216923
Epoch[59] Batch [549]	Speed: 1497.492003 samples/sec	accuracy=0.593601	lr=0.216706
Epoch[59] Batch [599]	Speed: 1538.365118 samples/sec	accuracy=0.593488	lr=0.216488
Epoch[59] Batch [649]	Speed: 1537.070367 samples/sec	accuracy=0.593238	lr=0.216270
Epoch[59] Batch [699]	Speed: 1542.864580 samples/sec	accuracy=0.592902	lr=0.216053
Epoch[59] Batch [749]	Speed: 1498.095585 samples/sec	accuracy=0.592712	lr=0.215835
Epoch[59] Batch [799]	Speed: 1485.305731 samples/sec	accuracy=0.592631	lr=0.215617
Epoch[59] Batch [849]	Speed: 1537.126191 samples/sec	accuracy=0.592642	lr=0.215400
Epoch[59] Batch [899]	Speed: 1523.831954 samples/sec	accuracy=0.592319	lr=0.215182
Epoch[59] Batch [949]	Speed: 1516.852560 samples/sec	accuracy=0.592293	lr=0.214964
Epoch[59] Batch [999]	Speed: 1520.992551 samples/sec	accuracy=0.592188	lr=0.214746
Epoch[59] Batch [1049]	Speed: 1511.364558 samples/sec	accuracy=0.591834	lr=0.214529
Epoch[59] Batch [1099]	Speed: 1461.068944 samples/sec	accuracy=0.591694	lr=0.214311
Epoch[59] Batch [1149]	Speed: 1519.808821 samples/sec	accuracy=0.591677	lr=0.214093
Epoch[59] Batch [1199]	Speed: 1517.374925 samples/sec	accuracy=0.591633	lr=0.213875
Epoch[59] Batch [1249]	Speed: 1527.738629 samples/sec	accuracy=0.591572	lr=0.213657
[Epoch 59] training: accuracy=0.591563
[Epoch 59] speed: 1509 samples/sec	time cost: 870.088187
[Epoch 59] validation: err-top1=0.350785 err-top5=0.132095
Epoch[60] Batch [49]	Speed: 1419.188929 samples/sec	accuracy=0.610391	lr=0.213435
Epoch[60] Batch [99]	Speed: 1514.489218 samples/sec	accuracy=0.596670	lr=0.213217
Epoch[60] Batch [149]	Speed: 1511.478284 samples/sec	accuracy=0.586615	lr=0.212999
Epoch[60] Batch [199]	Speed: 1449.193216 samples/sec	accuracy=0.588076	lr=0.212781
Epoch[60] Batch [249]	Speed: 1521.667479 samples/sec	accuracy=0.589488	lr=0.212563
Epoch[60] Batch [299]	Speed: 1504.919290 samples/sec	accuracy=0.590495	lr=0.212345
Epoch[60] Batch [349]	Speed: 1489.472232 samples/sec	accuracy=0.591303	lr=0.212127
Epoch[60] Batch [399]	Speed: 1486.118028 samples/sec	accuracy=0.591567	lr=0.211909
Epoch[60] Batch [449]	Speed: 1493.402493 samples/sec	accuracy=0.592190	lr=0.211691
Epoch[60] Batch [499]	Speed: 1546.653495 samples/sec	accuracy=0.592479	lr=0.211473
Epoch[60] Batch [549]	Speed: 1513.105029 samples/sec	accuracy=0.592615	lr=0.211255
Epoch[60] Batch [599]	Speed: 1509.810481 samples/sec	accuracy=0.593040	lr=0.211037
Epoch[60] Batch [649]	Speed: 1522.067358 samples/sec	accuracy=0.593015	lr=0.210819
Epoch[60] Batch [699]	Speed: 1469.194457 samples/sec	accuracy=0.592990	lr=0.210601
Epoch[60] Batch [749]	Speed: 1487.184807 samples/sec	accuracy=0.592866	lr=0.210383
Epoch[60] Batch [799]	Speed: 1480.131625 samples/sec	accuracy=0.592732	lr=0.210165
Epoch[60] Batch [849]	Speed: 1545.037031 samples/sec	accuracy=0.592632	lr=0.209947
Epoch[60] Batch [899]	Speed: 1458.141120 samples/sec	accuracy=0.592517	lr=0.209729
Epoch[60] Batch [949]	Speed: 1537.001137 samples/sec	accuracy=0.592266	lr=0.209511
Epoch[60] Batch [999]	Speed: 1493.163573 samples/sec	accuracy=0.592527	lr=0.209293
Epoch[60] Batch [1049]	Speed: 1514.036308 samples/sec	accuracy=0.592657	lr=0.209075
Epoch[60] Batch [1099]	Speed: 1475.656315 samples/sec	accuracy=0.592562	lr=0.208856
Epoch[60] Batch [1149]	Speed: 1493.208809 samples/sec	accuracy=0.592430	lr=0.208638
Epoch[60] Batch [1199]	Speed: 1547.408927 samples/sec	accuracy=0.592423	lr=0.208420
Epoch[60] Batch [1249]	Speed: 1551.299759 samples/sec	accuracy=0.592154	lr=0.208202
[Epoch 60] training: accuracy=0.592172
[Epoch 60] speed: 1499 samples/sec	time cost: 875.821575
[Epoch 60] validation: err-top1=0.350626 err-top5=0.131836
Epoch[61] Batch [49]	Speed: 1496.813306 samples/sec	accuracy=0.606113	lr=0.207979
Epoch[61] Batch [99]	Speed: 1467.064381 samples/sec	accuracy=0.605205	lr=0.207761
Epoch[61] Batch [149]	Speed: 1417.653835 samples/sec	accuracy=0.604785	lr=0.207543
Epoch[61] Batch [199]	Speed: 1479.228830 samples/sec	accuracy=0.604028	lr=0.207325
Epoch[61] Batch [249]	Speed: 1532.960642 samples/sec	accuracy=0.602852	lr=0.207106
Epoch[61] Batch [299]	Speed: 1514.499483 samples/sec	accuracy=0.602142	lr=0.206888
Epoch[61] Batch [349]	Speed: 1546.866350 samples/sec	accuracy=0.601599	lr=0.206670
Epoch[61] Batch [399]	Speed: 1494.766636 samples/sec	accuracy=0.601211	lr=0.206452
Epoch[61] Batch [449]	Speed: 1507.547421 samples/sec	accuracy=0.600773	lr=0.206233
Epoch[61] Batch [499]	Speed: 1507.992846 samples/sec	accuracy=0.600664	lr=0.206015
Epoch[61] Batch [549]	Speed: 1483.264982 samples/sec	accuracy=0.600467	lr=0.205797
Epoch[61] Batch [599]	Speed: 1470.031758 samples/sec	accuracy=0.599552	lr=0.205579
Epoch[61] Batch [649]	Speed: 1529.130593 samples/sec	accuracy=0.599441	lr=0.205360
Epoch[61] Batch [699]	Speed: 1513.279563 samples/sec	accuracy=0.598782	lr=0.205142
Epoch[61] Batch [749]	Speed: 1517.918422 samples/sec	accuracy=0.598928	lr=0.204924
Epoch[61] Batch [799]	Speed: 1485.133163 samples/sec	accuracy=0.598887	lr=0.204705
Epoch[61] Batch [849]	Speed: 1488.793124 samples/sec	accuracy=0.599030	lr=0.204487
Epoch[61] Batch [899]	Speed: 1492.816321 samples/sec	accuracy=0.598987	lr=0.204269
Epoch[61] Batch [949]	Speed: 1491.457038 samples/sec	accuracy=0.598616	lr=0.204051
Epoch[61] Batch [999]	Speed: 1520.621640 samples/sec	accuracy=0.598554	lr=0.203832
Epoch[61] Batch [1049]	Speed: 1493.839815 samples/sec	accuracy=0.598172	lr=0.203614
Epoch[61] Batch [1099]	Speed: 1538.493305 samples/sec	accuracy=0.598069	lr=0.203396
Epoch[61] Batch [1149]	Speed: 1550.530505 samples/sec	accuracy=0.597733	lr=0.203177
Epoch[61] Batch [1199]	Speed: 1546.323475 samples/sec	accuracy=0.597695	lr=0.202959
Epoch[61] Batch [1249]	Speed: 1496.957388 samples/sec	accuracy=0.597629	lr=0.202740
[Epoch 61] training: accuracy=0.597612
[Epoch 61] speed: 1501 samples/sec	time cost: 874.820261
[Epoch 61] validation: err-top1=0.343690 err-top5=0.128049
Epoch[62] Batch [49]	Speed: 1533.632983 samples/sec	accuracy=0.607285	lr=0.202518
Epoch[62] Batch [99]	Speed: 1522.294251 samples/sec	accuracy=0.605244	lr=0.202299
Epoch[62] Batch [149]	Speed: 1499.084318 samples/sec	accuracy=0.604258	lr=0.202081
Epoch[62] Batch [199]	Speed: 1527.945125 samples/sec	accuracy=0.604116	lr=0.201863
Epoch[62] Batch [249]	Speed: 1503.263218 samples/sec	accuracy=0.602949	lr=0.201644
Epoch[62] Batch [299]	Speed: 1535.379779 samples/sec	accuracy=0.602699	lr=0.201426
Epoch[62] Batch [349]	Speed: 1474.991953 samples/sec	accuracy=0.602377	lr=0.201208
Epoch[62] Batch [399]	Speed: 1471.430554 samples/sec	accuracy=0.602319	lr=0.200989
Epoch[62] Batch [449]	Speed: 1501.250590 samples/sec	accuracy=0.602044	lr=0.200771
Epoch[62] Batch [499]	Speed: 1466.171095 samples/sec	accuracy=0.601877	lr=0.200552
Epoch[62] Batch [549]	Speed: 1536.716767 samples/sec	accuracy=0.601829	lr=0.200334
Epoch[62] Batch [599]	Speed: 1483.534758 samples/sec	accuracy=0.602030	lr=0.200116
Epoch[62] Batch [649]	Speed: 1533.074248 samples/sec	accuracy=0.602004	lr=0.199897
Epoch[62] Batch [699]	Speed: 1536.290482 samples/sec	accuracy=0.601569	lr=0.199679
Epoch[62] Batch [749]	Speed: 1518.298996 samples/sec	accuracy=0.601186	lr=0.199461
Epoch[62] Batch [799]	Speed: 1490.554769 samples/sec	accuracy=0.600864	lr=0.199242
Epoch[62] Batch [849]	Speed: 1445.295180 samples/sec	accuracy=0.600546	lr=0.199024
Epoch[62] Batch [899]	Speed: 1544.821633 samples/sec	accuracy=0.600472	lr=0.198806
Epoch[62] Batch [949]	Speed: 1519.052217 samples/sec	accuracy=0.600080	lr=0.198587
Epoch[62] Batch [999]	Speed: 1476.516135 samples/sec	accuracy=0.600029	lr=0.198369
Epoch[62] Batch [1049]	Speed: 1506.891207 samples/sec	accuracy=0.599726	lr=0.198150
Epoch[62] Batch [1099]	Speed: 1519.454173 samples/sec	accuracy=0.599597	lr=0.197932
Epoch[62] Batch [1149]	Speed: 1498.183827 samples/sec	accuracy=0.599379	lr=0.197714
Epoch[62] Batch [1199]	Speed: 1535.840389 samples/sec	accuracy=0.599105	lr=0.197495
Epoch[62] Batch [1249]	Speed: 1522.017660 samples/sec	accuracy=0.598900	lr=0.197277
[Epoch 62] training: accuracy=0.598882
[Epoch 62] speed: 1506 samples/sec	time cost: 872.731730
[Epoch 62] validation: err-top1=0.354114 err-top5=0.132553
Epoch[63] Batch [49]	Speed: 1512.036173 samples/sec	accuracy=0.609609	lr=0.197054
Epoch[63] Batch [99]	Speed: 1469.909886 samples/sec	accuracy=0.609521	lr=0.196836
Epoch[63] Batch [149]	Speed: 1544.908051 samples/sec	accuracy=0.609141	lr=0.196618
Epoch[63] Batch [199]	Speed: 1503.668410 samples/sec	accuracy=0.607021	lr=0.196399
Epoch[63] Batch [249]	Speed: 1444.284522 samples/sec	accuracy=0.606930	lr=0.196181
Epoch[63] Batch [299]	Speed: 1534.328738 samples/sec	accuracy=0.606615	lr=0.195963
Epoch[63] Batch [349]	Speed: 1543.018352 samples/sec	accuracy=0.606334	lr=0.195744
Epoch[63] Batch [399]	Speed: 1483.667305 samples/sec	accuracy=0.605833	lr=0.195526
Epoch[63] Batch [449]	Speed: 1538.456371 samples/sec	accuracy=0.605651	lr=0.195308
Epoch[63] Batch [499]	Speed: 1495.222686 samples/sec	accuracy=0.605426	lr=0.195089
Epoch[63] Batch [549]	Speed: 1486.374996 samples/sec	accuracy=0.605375	lr=0.194871
Epoch[63] Batch [599]	Speed: 1475.663301 samples/sec	accuracy=0.605096	lr=0.194653
Epoch[63] Batch [649]	Speed: 1540.921292 samples/sec	accuracy=0.604728	lr=0.194434
Epoch[63] Batch [699]	Speed: 1495.129973 samples/sec	accuracy=0.604424	lr=0.194216
Epoch[63] Batch [749]	Speed: 1530.755455 samples/sec	accuracy=0.604102	lr=0.193998
Epoch[63] Batch [799]	Speed: 1505.417109 samples/sec	accuracy=0.603993	lr=0.193780
Epoch[63] Batch [849]	Speed: 1510.800383 samples/sec	accuracy=0.603927	lr=0.193561
Epoch[63] Batch [899]	Speed: 1523.083309 samples/sec	accuracy=0.603624	lr=0.193343
Epoch[63] Batch [949]	Speed: 1522.715751 samples/sec	accuracy=0.603555	lr=0.193125
Epoch[63] Batch [999]	Speed: 1503.590123 samples/sec	accuracy=0.603327	lr=0.192907
Epoch[63] Batch [1049]	Speed: 1481.592096 samples/sec	accuracy=0.603271	lr=0.192688
Epoch[63] Batch [1099]	Speed: 1512.243610 samples/sec	accuracy=0.603167	lr=0.192470
Epoch[63] Batch [1149]	Speed: 1520.467660 samples/sec	accuracy=0.602945	lr=0.192252
Epoch[63] Batch [1199]	Speed: 1509.532284 samples/sec	accuracy=0.602845	lr=0.192034
Epoch[63] Batch [1249]	Speed: 1474.762260 samples/sec	accuracy=0.602833	lr=0.191816
[Epoch 63] training: accuracy=0.602822
[Epoch 63] speed: 1504 samples/sec	time cost: 873.011473
[Epoch 63] validation: err-top1=0.351990 err-top5=0.133280
Epoch[64] Batch [49]	Speed: 1520.555477 samples/sec	accuracy=0.609941	lr=0.191593
Epoch[64] Batch [99]	Speed: 1487.880972 samples/sec	accuracy=0.606777	lr=0.191375
Epoch[64] Batch [149]	Speed: 1394.702427 samples/sec	accuracy=0.608197	lr=0.191157
Epoch[64] Batch [199]	Speed: 1533.269589 samples/sec	accuracy=0.608555	lr=0.190939
Epoch[64] Batch [249]	Speed: 1546.554362 samples/sec	accuracy=0.607641	lr=0.190720
Epoch[64] Batch [299]	Speed: 1509.086978 samples/sec	accuracy=0.607624	lr=0.190502
Epoch[64] Batch [349]	Speed: 1491.799264 samples/sec	accuracy=0.607229	lr=0.190284
Epoch[64] Batch [399]	Speed: 1506.917367 samples/sec	accuracy=0.607385	lr=0.190066
Epoch[64] Batch [449]	Speed: 1540.237377 samples/sec	accuracy=0.607146	lr=0.189848
Epoch[64] Batch [499]	Speed: 1488.868155 samples/sec	accuracy=0.607195	lr=0.189630
Epoch[64] Batch [549]	Speed: 1534.503642 samples/sec	accuracy=0.607147	lr=0.189412
Epoch[64] Batch [599]	Speed: 1520.707537 samples/sec	accuracy=0.606623	lr=0.189194
Epoch[64] Batch [649]	Speed: 1496.286077 samples/sec	accuracy=0.606424	lr=0.188976
Epoch[64] Batch [699]	Speed: 1512.769935 samples/sec	accuracy=0.606172	lr=0.188758
Epoch[64] Batch [749]	Speed: 1509.959274 samples/sec	accuracy=0.605980	lr=0.188540
Epoch[64] Batch [799]	Speed: 1544.383665 samples/sec	accuracy=0.605574	lr=0.188322
Epoch[64] Batch [849]	Speed: 1466.434238 samples/sec	accuracy=0.605273	lr=0.188104
Epoch[64] Batch [899]	Speed: 1481.095575 samples/sec	accuracy=0.605122	lr=0.187886
Epoch[64] Batch [949]	Speed: 1477.983859 samples/sec	accuracy=0.605039	lr=0.187668
Epoch[64] Batch [999]	Speed: 1474.027451 samples/sec	accuracy=0.604870	lr=0.187450
Epoch[64] Batch [1049]	Speed: 1534.955717 samples/sec	accuracy=0.604665	lr=0.187232
Epoch[64] Batch [1099]	Speed: 1503.318076 samples/sec	accuracy=0.604605	lr=0.187014
Epoch[64] Batch [1149]	Speed: 1524.399921 samples/sec	accuracy=0.604519	lr=0.186796
Epoch[64] Batch [1199]	Speed: 1502.475383 samples/sec	accuracy=0.604386	lr=0.186578
Epoch[64] Batch [1249]	Speed: 1519.609378 samples/sec	accuracy=0.604265	lr=0.186360
[Epoch 64] training: accuracy=0.604272
[Epoch 64] speed: 1503 samples/sec	time cost: 875.071338
[Epoch 64] validation: err-top1=0.345524 err-top5=0.129684
Epoch[65] Batch [49]	Speed: 1462.635178 samples/sec	accuracy=0.618945	lr=0.186138
Epoch[65] Batch [99]	Speed: 1513.158881 samples/sec	accuracy=0.614971	lr=0.185920
Epoch[65] Batch [149]	Speed: 1517.859188 samples/sec	accuracy=0.614147	lr=0.185702
Epoch[65] Batch [199]	Speed: 1516.350609 samples/sec	accuracy=0.612603	lr=0.185485
Epoch[65] Batch [249]	Speed: 1539.767082 samples/sec	accuracy=0.611340	lr=0.185267
Epoch[65] Batch [299]	Speed: 1466.968404 samples/sec	accuracy=0.611276	lr=0.185049
Epoch[65] Batch [349]	Speed: 1512.510013 samples/sec	accuracy=0.611038	lr=0.184831
Epoch[65] Batch [399]	Speed: 1494.934312 samples/sec	accuracy=0.610637	lr=0.184614
Epoch[65] Batch [449]	Speed: 1475.836505 samples/sec	accuracy=0.610189	lr=0.184396
Epoch[65] Batch [499]	Speed: 1516.337717 samples/sec	accuracy=0.609822	lr=0.184178
Epoch[65] Batch [549]	Speed: 1483.249748 samples/sec	accuracy=0.609396	lr=0.183960
Epoch[65] Batch [599]	Speed: 1534.172079 samples/sec	accuracy=0.609051	lr=0.183743
Epoch[65] Batch [649]	Speed: 1497.305922 samples/sec	accuracy=0.608603	lr=0.183525
Epoch[65] Batch [699]	Speed: 1552.975338 samples/sec	accuracy=0.608401	lr=0.183308
Epoch[65] Batch [749]	Speed: 1525.534049 samples/sec	accuracy=0.608358	lr=0.183090
Epoch[65] Batch [799]	Speed: 1519.154250 samples/sec	accuracy=0.608296	lr=0.182872
Epoch[65] Batch [849]	Speed: 1478.359955 samples/sec	accuracy=0.608019	lr=0.182655
Epoch[65] Batch [899]	Speed: 1516.855453 samples/sec	accuracy=0.607643	lr=0.182437
Epoch[65] Batch [949]	Speed: 1490.398935 samples/sec	accuracy=0.607220	lr=0.182220
Epoch[65] Batch [999]	Speed: 1488.732035 samples/sec	accuracy=0.606917	lr=0.182002
Epoch[65] Batch [1049]	Speed: 1537.398099 samples/sec	accuracy=0.606958	lr=0.181785
Epoch[65] Batch [1099]	Speed: 1505.719001 samples/sec	accuracy=0.606656	lr=0.181567
Epoch[65] Batch [1149]	Speed: 1489.602205 samples/sec	accuracy=0.606641	lr=0.181350
Epoch[65] Batch [1199]	Speed: 1458.068828 samples/sec	accuracy=0.606610	lr=0.181132
Epoch[65] Batch [1249]	Speed: 1484.122063 samples/sec	accuracy=0.606678	lr=0.180915
[Epoch 65] training: accuracy=0.606697
[Epoch 65] speed: 1501 samples/sec	time cost: 874.813201
[Epoch 65] validation: err-top1=0.339047 err-top5=0.126395
Epoch[66] Batch [49]	Speed: 1501.127034 samples/sec	accuracy=0.622363	lr=0.180693
Epoch[66] Batch [99]	Speed: 1526.059278 samples/sec	accuracy=0.620898	lr=0.180476
Epoch[66] Batch [149]	Speed: 1486.819659 samples/sec	accuracy=0.618900	lr=0.180259
Epoch[66] Batch [199]	Speed: 1476.816286 samples/sec	accuracy=0.617383	lr=0.180041
Epoch[66] Batch [249]	Speed: 1538.461926 samples/sec	accuracy=0.616902	lr=0.179824
Epoch[66] Batch [299]	Speed: 1532.053979 samples/sec	accuracy=0.616243	lr=0.179607
Epoch[66] Batch [349]	Speed: 1485.731447 samples/sec	accuracy=0.615176	lr=0.179390
Epoch[66] Batch [399]	Speed: 1510.459743 samples/sec	accuracy=0.614646	lr=0.179172
Epoch[66] Batch [449]	Speed: 1532.944731 samples/sec	accuracy=0.614783	lr=0.178955
Epoch[66] Batch [499]	Speed: 1516.573016 samples/sec	accuracy=0.614182	lr=0.178738
Epoch[66] Batch [549]	Speed: 1498.667378 samples/sec	accuracy=0.614171	lr=0.178521
Epoch[66] Batch [599]	Speed: 1463.523631 samples/sec	accuracy=0.613433	lr=0.178304
Epoch[66] Batch [649]	Speed: 1491.361736 samples/sec	accuracy=0.613208	lr=0.178087
Epoch[66] Batch [699]	Speed: 1537.053348 samples/sec	accuracy=0.612987	lr=0.177870
Epoch[66] Batch [749]	Speed: 1508.496484 samples/sec	accuracy=0.612548	lr=0.177653
Epoch[66] Batch [799]	Speed: 1481.452827 samples/sec	accuracy=0.612130	lr=0.177436
Epoch[66] Batch [849]	Speed: 1451.596491 samples/sec	accuracy=0.611685	lr=0.177219
Epoch[66] Batch [899]	Speed: 1487.082791 samples/sec	accuracy=0.611376	lr=0.177002
Epoch[66] Batch [949]	Speed: 1509.347888 samples/sec	accuracy=0.611081	lr=0.176785
Epoch[66] Batch [999]	Speed: 1505.775728 samples/sec	accuracy=0.610691	lr=0.176568
Epoch[66] Batch [1049]	Speed: 1520.528734 samples/sec	accuracy=0.610546	lr=0.176351
Epoch[66] Batch [1099]	Speed: 1492.129320 samples/sec	accuracy=0.610501	lr=0.176134
Epoch[66] Batch [1149]	Speed: 1507.927608 samples/sec	accuracy=0.610422	lr=0.175918
Epoch[66] Batch [1199]	Speed: 1491.632488 samples/sec	accuracy=0.610274	lr=0.175701
Epoch[66] Batch [1249]	Speed: 1474.908388 samples/sec	accuracy=0.610018	lr=0.175484
[Epoch 66] training: accuracy=0.610043
[Epoch 66] speed: 1499 samples/sec	time cost: 876.559004
[Epoch 66] validation: err-top1=0.340701 err-top5=0.126654
Epoch[67] Batch [49]	Speed: 1509.116057 samples/sec	accuracy=0.618652	lr=0.175263
Epoch[67] Batch [99]	Speed: 1524.830849 samples/sec	accuracy=0.617559	lr=0.175046
Epoch[67] Batch [149]	Speed: 1541.481226 samples/sec	accuracy=0.617259	lr=0.174830
Epoch[67] Batch [199]	Speed: 1545.275995 samples/sec	accuracy=0.617900	lr=0.174613
Epoch[67] Batch [249]	Speed: 1523.857257 samples/sec	accuracy=0.618203	lr=0.174397
Epoch[67] Batch [299]	Speed: 1535.763340 samples/sec	accuracy=0.617165	lr=0.174180
Epoch[67] Batch [349]	Speed: 1494.861978 samples/sec	accuracy=0.616925	lr=0.173964
Epoch[67] Batch [399]	Speed: 1512.250181 samples/sec	accuracy=0.616943	lr=0.173747
Epoch[67] Batch [449]	Speed: 1489.429391 samples/sec	accuracy=0.616408	lr=0.173531
Epoch[67] Batch [499]	Speed: 1492.734531 samples/sec	accuracy=0.615826	lr=0.173314
Epoch[67] Batch [549]	Speed: 1481.904520 samples/sec	accuracy=0.615328	lr=0.173098
Epoch[67] Batch [599]	Speed: 1488.101189 samples/sec	accuracy=0.615145	lr=0.172881
Epoch[67] Batch [649]	Speed: 1442.894800 samples/sec	accuracy=0.614965	lr=0.172665
Epoch[67] Batch [699]	Speed: 1511.925460 samples/sec	accuracy=0.614358	lr=0.172449
Epoch[67] Batch [749]	Speed: 1487.724007 samples/sec	accuracy=0.614309	lr=0.172232
Epoch[67] Batch [799]	Speed: 1488.932910 samples/sec	accuracy=0.614030	lr=0.172016
Epoch[67] Batch [849]	Speed: 1503.076962 samples/sec	accuracy=0.613699	lr=0.171800
Epoch[67] Batch [899]	Speed: 1448.634007 samples/sec	accuracy=0.613650	lr=0.171584
Epoch[67] Batch [949]	Speed: 1523.315648 samples/sec	accuracy=0.613495	lr=0.171368
Epoch[67] Batch [999]	Speed: 1537.659169 samples/sec	accuracy=0.613255	lr=0.171152
Epoch[67] Batch [1049]	Speed: 1539.602931 samples/sec	accuracy=0.613228	lr=0.170935
Epoch[67] Batch [1099]	Speed: 1508.547571 samples/sec	accuracy=0.613073	lr=0.170719
Epoch[67] Batch [1149]	Speed: 1501.197813 samples/sec	accuracy=0.612989	lr=0.170503
Epoch[67] Batch [1199]	Speed: 1526.543729 samples/sec	accuracy=0.612826	lr=0.170287
Epoch[67] Batch [1249]	Speed: 1498.229441 samples/sec	accuracy=0.612691	lr=0.170072
[Epoch 67] training: accuracy=0.612700
[Epoch 67] speed: 1504 samples/sec	time cost: 872.608134
[Epoch 67] validation: err-top1=0.332171 err-top5=0.121173
Epoch[68] Batch [49]	Speed: 1524.011589 samples/sec	accuracy=0.624062	lr=0.169851
Epoch[68] Batch [99]	Speed: 1483.036955 samples/sec	accuracy=0.623271	lr=0.169635
Epoch[68] Batch [149]	Speed: 1541.336334 samples/sec	accuracy=0.620755	lr=0.169420
Epoch[68] Batch [199]	Speed: 1492.648477 samples/sec	accuracy=0.619985	lr=0.169204
Epoch[68] Batch [249]	Speed: 1529.352332 samples/sec	accuracy=0.620281	lr=0.168988
Epoch[68] Batch [299]	Speed: 1487.378909 samples/sec	accuracy=0.619824	lr=0.168772
Epoch[68] Batch [349]	Speed: 1548.171381 samples/sec	accuracy=0.619439	lr=0.168557
Epoch[68] Batch [399]	Speed: 1519.365827 samples/sec	accuracy=0.618936	lr=0.168341
Epoch[68] Batch [449]	Speed: 1534.930922 samples/sec	accuracy=0.618928	lr=0.168126
Epoch[68] Batch [499]	Speed: 1538.973430 samples/sec	accuracy=0.618627	lr=0.167910
Epoch[68] Batch [549]	Speed: 1520.827660 samples/sec	accuracy=0.618258	lr=0.167694
Epoch[68] Batch [599]	Speed: 1463.624595 samples/sec	accuracy=0.617578	lr=0.167479
Epoch[68] Batch [649]	Speed: 1538.554468 samples/sec	accuracy=0.617470	lr=0.167263
Epoch[68] Batch [699]	Speed: 1493.222940 samples/sec	accuracy=0.617446	lr=0.167048
Epoch[68] Batch [749]	Speed: 1520.795727 samples/sec	accuracy=0.617253	lr=0.166833
Epoch[68] Batch [799]	Speed: 1536.361198 samples/sec	accuracy=0.617321	lr=0.166617
Epoch[68] Batch [849]	Speed: 1538.988617 samples/sec	accuracy=0.617007	lr=0.166402
Epoch[68] Batch [899]	Speed: 1462.360660 samples/sec	accuracy=0.616814	lr=0.166187
Epoch[68] Batch [949]	Speed: 1480.976938 samples/sec	accuracy=0.616598	lr=0.165972
Epoch[68] Batch [999]	Speed: 1532.516773 samples/sec	accuracy=0.616468	lr=0.165756
Epoch[68] Batch [1049]	Speed: 1488.952607 samples/sec	accuracy=0.616227	lr=0.165541
Epoch[68] Batch [1099]	Speed: 1490.907646 samples/sec	accuracy=0.615979	lr=0.165326
Epoch[68] Batch [1149]	Speed: 1514.984391 samples/sec	accuracy=0.615695	lr=0.165111
Epoch[68] Batch [1199]	Speed: 1497.201563 samples/sec	accuracy=0.615553	lr=0.164896
Epoch[68] Batch [1249]	Speed: 1527.316538 samples/sec	accuracy=0.615369	lr=0.164681
[Epoch 68] training: accuracy=0.615373
[Epoch 68] speed: 1510 samples/sec	time cost: 869.885509
[Epoch 68] validation: err-top1=0.339146 err-top5=0.125100
Epoch[69] Batch [49]	Speed: 1483.217785 samples/sec	accuracy=0.626973	lr=0.164462
Epoch[69] Batch [99]	Speed: 1486.738865 samples/sec	accuracy=0.625479	lr=0.164247
Epoch[69] Batch [149]	Speed: 1531.543852 samples/sec	accuracy=0.623574	lr=0.164032
Epoch[69] Batch [199]	Speed: 1474.015462 samples/sec	accuracy=0.624121	lr=0.163818
Epoch[69] Batch [249]	Speed: 1487.041385 samples/sec	accuracy=0.624176	lr=0.163603
Epoch[69] Batch [299]	Speed: 1513.297244 samples/sec	accuracy=0.623841	lr=0.163388
Epoch[69] Batch [349]	Speed: 1543.145330 samples/sec	accuracy=0.623186	lr=0.163173
Epoch[69] Batch [399]	Speed: 1538.472010 samples/sec	accuracy=0.622524	lr=0.162959
Epoch[69] Batch [449]	Speed: 1494.066021 samples/sec	accuracy=0.622246	lr=0.162744
Epoch[69] Batch [499]	Speed: 1506.344736 samples/sec	accuracy=0.622209	lr=0.162530
Epoch[69] Batch [549]	Speed: 1517.691748 samples/sec	accuracy=0.621770	lr=0.162315
Epoch[69] Batch [599]	Speed: 1536.495966 samples/sec	accuracy=0.621533	lr=0.162101
Epoch[69] Batch [649]	Speed: 1546.554529 samples/sec	accuracy=0.621044	lr=0.161886
Epoch[69] Batch [699]	Speed: 1517.351177 samples/sec	accuracy=0.620545	lr=0.161672
Epoch[69] Batch [749]	Speed: 1521.162099 samples/sec	accuracy=0.620255	lr=0.161458
Epoch[69] Batch [799]	Speed: 1469.157539 samples/sec	accuracy=0.619974	lr=0.161244
Epoch[69] Batch [849]	Speed: 1540.709108 samples/sec	accuracy=0.619635	lr=0.161029
Epoch[69] Batch [899]	Speed: 1502.362136 samples/sec	accuracy=0.619417	lr=0.160815
Epoch[69] Batch [949]	Speed: 1542.846457 samples/sec	accuracy=0.619223	lr=0.160601
Epoch[69] Batch [999]	Speed: 1469.235770 samples/sec	accuracy=0.619260	lr=0.160387
Epoch[69] Batch [1049]	Speed: 1538.762499 samples/sec	accuracy=0.619139	lr=0.160173
Epoch[69] Batch [1099]	Speed: 1498.266956 samples/sec	accuracy=0.618881	lr=0.159959
Epoch[69] Batch [1149]	Speed: 1473.747325 samples/sec	accuracy=0.618713	lr=0.159745
Epoch[69] Batch [1199]	Speed: 1509.070095 samples/sec	accuracy=0.618534	lr=0.159531
Epoch[69] Batch [1249]	Speed: 1550.567886 samples/sec	accuracy=0.618382	lr=0.159317
[Epoch 69] training: accuracy=0.618387
[Epoch 69] speed: 1510 samples/sec	time cost: 869.651808
[Epoch 69] validation: err-top1=0.330933 err-top5=0.120829
Epoch[70] Batch [49]	Speed: 1458.954615 samples/sec	accuracy=0.627695	lr=0.159099
Epoch[70] Batch [99]	Speed: 1516.710097 samples/sec	accuracy=0.627549	lr=0.158886
Epoch[70] Batch [149]	Speed: 1512.718710 samples/sec	accuracy=0.627474	lr=0.158672
Epoch[70] Batch [199]	Speed: 1486.966457 samples/sec	accuracy=0.626519	lr=0.158458
Epoch[70] Batch [249]	Speed: 1494.312841 samples/sec	accuracy=0.625426	lr=0.158245
Epoch[70] Batch [299]	Speed: 1535.960981 samples/sec	accuracy=0.625472	lr=0.158031
Epoch[70] Batch [349]	Speed: 1474.273655 samples/sec	accuracy=0.625103	lr=0.157818
Epoch[70] Batch [399]	Speed: 1517.885291 samples/sec	accuracy=0.624954	lr=0.157604
Epoch[70] Batch [449]	Speed: 1502.402759 samples/sec	accuracy=0.624123	lr=0.157391
Epoch[70] Batch [499]	Speed: 1538.524784 samples/sec	accuracy=0.623768	lr=0.157177
Epoch[70] Batch [549]	Speed: 1467.240505 samples/sec	accuracy=0.623569	lr=0.156964
Epoch[70] Batch [599]	Speed: 1530.044088 samples/sec	accuracy=0.623237	lr=0.156751
Epoch[70] Batch [649]	Speed: 1547.983673 samples/sec	accuracy=0.622820	lr=0.156538
Epoch[70] Batch [699]	Speed: 1503.555541 samples/sec	accuracy=0.622663	lr=0.156325
Epoch[70] Batch [749]	Speed: 1504.414380 samples/sec	accuracy=0.622594	lr=0.156112
Epoch[70] Batch [799]	Speed: 1496.748270 samples/sec	accuracy=0.622350	lr=0.155899
Epoch[70] Batch [849]	Speed: 1504.735829 samples/sec	accuracy=0.622316	lr=0.155686
Epoch[70] Batch [899]	Speed: 1540.381797 samples/sec	accuracy=0.622230	lr=0.155473
Epoch[70] Batch [949]	Speed: 1523.687636 samples/sec	accuracy=0.622061	lr=0.155260
Epoch[70] Batch [999]	Speed: 1525.883561 samples/sec	accuracy=0.621811	lr=0.155047
Epoch[70] Batch [1049]	Speed: 1418.310653 samples/sec	accuracy=0.621696	lr=0.154834
Epoch[70] Batch [1099]	Speed: 1534.447492 samples/sec	accuracy=0.621498	lr=0.154622
Epoch[70] Batch [1149]	Speed: 1502.313653 samples/sec	accuracy=0.621305	lr=0.154409
Epoch[70] Batch [1199]	Speed: 1492.815625 samples/sec	accuracy=0.621147	lr=0.154196
Epoch[70] Batch [1249]	Speed: 1527.452875 samples/sec	accuracy=0.621133	lr=0.153984
[Epoch 70] training: accuracy=0.621117
[Epoch 70] speed: 1504 samples/sec	time cost: 872.963443
[Epoch 70] validation: err-top1=0.329042 err-top5=0.120037
Epoch[71] Batch [49]	Speed: 1485.682840 samples/sec	accuracy=0.627852	lr=0.153767
Epoch[71] Batch [99]	Speed: 1476.560967 samples/sec	accuracy=0.627412	lr=0.153555
Epoch[71] Batch [149]	Speed: 1495.093593 samples/sec	accuracy=0.627044	lr=0.153342
Epoch[71] Batch [199]	Speed: 1540.413818 samples/sec	accuracy=0.626807	lr=0.153130
Epoch[71] Batch [249]	Speed: 1531.662732 samples/sec	accuracy=0.627371	lr=0.152918
Epoch[71] Batch [299]	Speed: 1489.530562 samples/sec	accuracy=0.627568	lr=0.152705
Epoch[71] Batch [349]	Speed: 1482.919983 samples/sec	accuracy=0.627310	lr=0.152493
Epoch[71] Batch [399]	Speed: 1477.227759 samples/sec	accuracy=0.627578	lr=0.152281
Epoch[71] Batch [449]	Speed: 1542.362549 samples/sec	accuracy=0.627289	lr=0.152069
Epoch[71] Batch [499]	Speed: 1520.326982 samples/sec	accuracy=0.626980	lr=0.151857
Epoch[71] Batch [549]	Speed: 1533.198303 samples/sec	accuracy=0.627069	lr=0.151645
Epoch[71] Batch [599]	Speed: 1494.194364 samples/sec	accuracy=0.626520	lr=0.151433
Epoch[71] Batch [649]	Speed: 1493.309310 samples/sec	accuracy=0.626020	lr=0.151222
Epoch[71] Batch [699]	Speed: 1478.784122 samples/sec	accuracy=0.626021	lr=0.151010
Epoch[71] Batch [749]	Speed: 1512.168644 samples/sec	accuracy=0.625910	lr=0.150798
Epoch[71] Batch [799]	Speed: 1541.257637 samples/sec	accuracy=0.625314	lr=0.150587
Epoch[71] Batch [849]	Speed: 1483.875665 samples/sec	accuracy=0.625101	lr=0.150375
Epoch[71] Batch [899]	Speed: 1537.654578 samples/sec	accuracy=0.624855	lr=0.150163
Epoch[71] Batch [949]	Speed: 1459.414302 samples/sec	accuracy=0.624682	lr=0.149952
Epoch[71] Batch [999]	Speed: 1547.224994 samples/sec	accuracy=0.624295	lr=0.149741
Epoch[71] Batch [1049]	Speed: 1509.514702 samples/sec	accuracy=0.624046	lr=0.149529
Epoch[71] Batch [1099]	Speed: 1486.551053 samples/sec	accuracy=0.623875	lr=0.149318
Epoch[71] Batch [1149]	Speed: 1512.695843 samples/sec	accuracy=0.623689	lr=0.149107
Epoch[71] Batch [1199]	Speed: 1478.530973 samples/sec	accuracy=0.623584	lr=0.148896
Epoch[71] Batch [1249]	Speed: 1442.699049 samples/sec	accuracy=0.623611	lr=0.148685
[Epoch 71] training: accuracy=0.623617
[Epoch 71] speed: 1500 samples/sec	time cost: 876.269897
[Epoch 71] validation: err-top1=0.325554 err-top5=0.115932
Epoch[72] Batch [49]	Speed: 1528.187748 samples/sec	accuracy=0.634102	lr=0.148469
Epoch[72] Batch [99]	Speed: 1539.649126 samples/sec	accuracy=0.635225	lr=0.148258
Epoch[72] Batch [149]	Speed: 1551.039884 samples/sec	accuracy=0.633444	lr=0.148047
Epoch[72] Batch [199]	Speed: 1520.799259 samples/sec	accuracy=0.633091	lr=0.147837
Epoch[72] Batch [249]	Speed: 1505.472030 samples/sec	accuracy=0.632129	lr=0.147626
Epoch[72] Batch [299]	Speed: 1538.875610 samples/sec	accuracy=0.631963	lr=0.147415
Epoch[72] Batch [349]	Speed: 1461.282339 samples/sec	accuracy=0.631016	lr=0.147204
Epoch[72] Batch [399]	Speed: 1509.202037 samples/sec	accuracy=0.630120	lr=0.146994
Epoch[72] Batch [449]	Speed: 1517.389260 samples/sec	accuracy=0.629918	lr=0.146783
Epoch[72] Batch [499]	Speed: 1490.297615 samples/sec	accuracy=0.629598	lr=0.146573
Epoch[72] Batch [549]	Speed: 1489.259488 samples/sec	accuracy=0.629235	lr=0.146362
Epoch[72] Batch [599]	Speed: 1528.265671 samples/sec	accuracy=0.628983	lr=0.146152
Epoch[72] Batch [649]	Speed: 1521.627273 samples/sec	accuracy=0.629262	lr=0.145942
Epoch[72] Batch [699]	Speed: 1517.544837 samples/sec	accuracy=0.629088	lr=0.145732
Epoch[72] Batch [749]	Speed: 1540.458835 samples/sec	accuracy=0.628902	lr=0.145521
Epoch[72] Batch [799]	Speed: 1472.960858 samples/sec	accuracy=0.628872	lr=0.145311
Epoch[72] Batch [849]	Speed: 1466.858141 samples/sec	accuracy=0.628581	lr=0.145101
Epoch[72] Batch [899]	Speed: 1517.623469 samples/sec	accuracy=0.628212	lr=0.144891
Epoch[72] Batch [949]	Speed: 1537.912740 samples/sec	accuracy=0.628032	lr=0.144682
Epoch[72] Batch [999]	Speed: 1501.633435 samples/sec	accuracy=0.627779	lr=0.144472
Epoch[72] Batch [1049]	Speed: 1546.009112 samples/sec	accuracy=0.627759	lr=0.144262
Epoch[72] Batch [1099]	Speed: 1514.796160 samples/sec	accuracy=0.627600	lr=0.144052
Epoch[72] Batch [1149]	Speed: 1534.154291 samples/sec	accuracy=0.627444	lr=0.143843
Epoch[72] Batch [1199]	Speed: 1544.486419 samples/sec	accuracy=0.627127	lr=0.143633
Epoch[72] Batch [1249]	Speed: 1557.919377 samples/sec	accuracy=0.627074	lr=0.143424
[Epoch 72] training: accuracy=0.627062
[Epoch 72] speed: 1516 samples/sec	time cost: 866.664455
[Epoch 72] validation: err-top1=0.327308 err-top5=0.118503
Epoch[73] Batch [49]	Speed: 1440.948132 samples/sec	accuracy=0.634570	lr=0.143210
Epoch[73] Batch [99]	Speed: 1466.278781 samples/sec	accuracy=0.634414	lr=0.143001
Epoch[73] Batch [149]	Speed: 1541.005981 samples/sec	accuracy=0.633958	lr=0.142791
Epoch[73] Batch [199]	Speed: 1521.552808 samples/sec	accuracy=0.633325	lr=0.142582
Epoch[73] Batch [249]	Speed: 1492.749120 samples/sec	accuracy=0.632898	lr=0.142373
Epoch[73] Batch [299]	Speed: 1537.775632 samples/sec	accuracy=0.632412	lr=0.142164
Epoch[73] Batch [349]	Speed: 1544.474522 samples/sec	accuracy=0.632171	lr=0.141955
Epoch[73] Batch [399]	Speed: 1518.510947 samples/sec	accuracy=0.632183	lr=0.141746
Epoch[73] Batch [449]	Speed: 1503.778032 samples/sec	accuracy=0.631584	lr=0.141537
Epoch[73] Batch [499]	Speed: 1506.986832 samples/sec	accuracy=0.631717	lr=0.141328
Epoch[73] Batch [549]	Speed: 1491.612854 samples/sec	accuracy=0.631607	lr=0.141120
Epoch[73] Batch [599]	Speed: 1517.724570 samples/sec	accuracy=0.631426	lr=0.140911
Epoch[73] Batch [649]	Speed: 1516.120474 samples/sec	accuracy=0.631220	lr=0.140702
Epoch[73] Batch [699]	Speed: 1544.809398 samples/sec	accuracy=0.631101	lr=0.140494
Epoch[73] Batch [749]	Speed: 1418.065751 samples/sec	accuracy=0.630922	lr=0.140285
Epoch[73] Batch [799]	Speed: 1522.794261 samples/sec	accuracy=0.630664	lr=0.140077
Epoch[73] Batch [849]	Speed: 1511.642281 samples/sec	accuracy=0.630526	lr=0.139869
Epoch[73] Batch [899]	Speed: 1489.673875 samples/sec	accuracy=0.630416	lr=0.139661
Epoch[73] Batch [949]	Speed: 1481.893578 samples/sec	accuracy=0.630340	lr=0.139452
Epoch[73] Batch [999]	Speed: 1518.136009 samples/sec	accuracy=0.630234	lr=0.139244
Epoch[73] Batch [1049]	Speed: 1505.648407 samples/sec	accuracy=0.630068	lr=0.139036
Epoch[73] Batch [1099]	Speed: 1521.825456 samples/sec	accuracy=0.629839	lr=0.138828
Epoch[73] Batch [1149]	Speed: 1508.925061 samples/sec	accuracy=0.629757	lr=0.138620
Epoch[73] Batch [1199]	Speed: 1516.247154 samples/sec	accuracy=0.629418	lr=0.138413
Epoch[73] Batch [1249]	Speed: 1522.264349 samples/sec	accuracy=0.629239	lr=0.138205
[Epoch 73] training: accuracy=0.629252
[Epoch 73] speed: 1504 samples/sec	time cost: 873.017283
[Epoch 73] validation: err-top1=0.324757 err-top5=0.116889
Epoch[74] Batch [49]	Speed: 1499.660018 samples/sec	accuracy=0.639551	lr=0.137993
Epoch[74] Batch [99]	Speed: 1476.047074 samples/sec	accuracy=0.639385	lr=0.137786
Epoch[74] Batch [149]	Speed: 1514.762033 samples/sec	accuracy=0.637129	lr=0.137578
Epoch[74] Batch [199]	Speed: 1507.841879 samples/sec	accuracy=0.637070	lr=0.137371
Epoch[74] Batch [249]	Speed: 1539.530393 samples/sec	accuracy=0.636406	lr=0.137163
Epoch[74] Batch [299]	Speed: 1470.567582 samples/sec	accuracy=0.636045	lr=0.136956
Epoch[74] Batch [349]	Speed: 1479.632290 samples/sec	accuracy=0.635770	lr=0.136749
Epoch[74] Batch [399]	Speed: 1482.690793 samples/sec	accuracy=0.635732	lr=0.136542
Epoch[74] Batch [449]	Speed: 1536.351218 samples/sec	accuracy=0.635260	lr=0.136335
Epoch[74] Batch [499]	Speed: 1517.646989 samples/sec	accuracy=0.635154	lr=0.136128
Epoch[74] Batch [549]	Speed: 1509.926638 samples/sec	accuracy=0.634652	lr=0.135921
Epoch[74] Batch [599]	Speed: 1492.036690 samples/sec	accuracy=0.634769	lr=0.135714
Epoch[74] Batch [649]	Speed: 1515.870486 samples/sec	accuracy=0.634378	lr=0.135507
Epoch[74] Batch [699]	Speed: 1525.835251 samples/sec	accuracy=0.634208	lr=0.135301
Epoch[74] Batch [749]	Speed: 1541.141709 samples/sec	accuracy=0.633921	lr=0.135094
Epoch[74] Batch [799]	Speed: 1492.150563 samples/sec	accuracy=0.633872	lr=0.134887
Epoch[74] Batch [849]	Speed: 1534.109696 samples/sec	accuracy=0.633655	lr=0.134681
Epoch[74] Batch [899]	Speed: 1494.980186 samples/sec	accuracy=0.633457	lr=0.134475
Epoch[74] Batch [949]	Speed: 1514.989382 samples/sec	accuracy=0.633321	lr=0.134268
Epoch[74] Batch [999]	Speed: 1449.402325 samples/sec	accuracy=0.633073	lr=0.134062
Epoch[74] Batch [1049]	Speed: 1552.214574 samples/sec	accuracy=0.633026	lr=0.133856
Epoch[74] Batch [1099]	Speed: 1496.519896 samples/sec	accuracy=0.632865	lr=0.133650
Epoch[74] Batch [1149]	Speed: 1541.578271 samples/sec	accuracy=0.632478	lr=0.133444
Epoch[74] Batch [1199]	Speed: 1477.440707 samples/sec	accuracy=0.632390	lr=0.133238
Epoch[74] Batch [1249]	Speed: 1506.706789 samples/sec	accuracy=0.632255	lr=0.133032
[Epoch 74] training: accuracy=0.632268
[Epoch 74] speed: 1505 samples/sec	time cost: 872.946714
[Epoch 74] validation: err-top1=0.320631 err-top5=0.114597
Epoch[75] Batch [49]	Speed: 1537.281716 samples/sec	accuracy=0.645117	lr=0.132822
Epoch[75] Batch [99]	Speed: 1505.734320 samples/sec	accuracy=0.642451	lr=0.132617
Epoch[75] Batch [149]	Speed: 1539.503563 samples/sec	accuracy=0.641302	lr=0.132411
Epoch[75] Batch [199]	Speed: 1518.722099 samples/sec	accuracy=0.641621	lr=0.132206
Epoch[75] Batch [249]	Speed: 1486.324999 samples/sec	accuracy=0.640352	lr=0.132000
Epoch[75] Batch [299]	Speed: 1505.168359 samples/sec	accuracy=0.639616	lr=0.131795
Epoch[75] Batch [349]	Speed: 1541.913929 samples/sec	accuracy=0.639414	lr=0.131590
Epoch[75] Batch [399]	Speed: 1502.603619 samples/sec	accuracy=0.639075	lr=0.131385
Epoch[75] Batch [449]	Speed: 1518.571134 samples/sec	accuracy=0.638778	lr=0.131180
Epoch[75] Batch [499]	Speed: 1506.497422 samples/sec	accuracy=0.638357	lr=0.130975
Epoch[75] Batch [549]	Speed: 1516.712979 samples/sec	accuracy=0.638217	lr=0.130770
Epoch[75] Batch [599]	Speed: 1516.814022 samples/sec	accuracy=0.637850	lr=0.130565
Epoch[75] Batch [649]	Speed: 1523.344986 samples/sec	accuracy=0.637554	lr=0.130360
Epoch[75] Batch [699]	Speed: 1526.184261 samples/sec	accuracy=0.637348	lr=0.130155
Epoch[75] Batch [749]	Speed: 1544.356721 samples/sec	accuracy=0.637135	lr=0.129951
Epoch[75] Batch [799]	Speed: 1458.700808 samples/sec	accuracy=0.636993	lr=0.129746
Epoch[75] Batch [849]	Speed: 1540.685110 samples/sec	accuracy=0.636614	lr=0.129542
Epoch[75] Batch [899]	Speed: 1496.425699 samples/sec	accuracy=0.636452	lr=0.129338
Epoch[75] Batch [949]	Speed: 1543.359307 samples/sec	accuracy=0.636265	lr=0.129133
Epoch[75] Batch [999]	Speed: 1496.686818 samples/sec	accuracy=0.635961	lr=0.128929
Epoch[75] Batch [1049]	Speed: 1519.861344 samples/sec	accuracy=0.635972	lr=0.128725
Epoch[75] Batch [1099]	Speed: 1493.836043 samples/sec	accuracy=0.635671	lr=0.128521
Epoch[75] Batch [1149]	Speed: 1537.162588 samples/sec	accuracy=0.635487	lr=0.128317
Epoch[75] Batch [1199]	Speed: 1420.264534 samples/sec	accuracy=0.635282	lr=0.128113
Epoch[75] Batch [1249]	Speed: 1515.658276 samples/sec	accuracy=0.634998	lr=0.127910
[Epoch 75] training: accuracy=0.635004
[Epoch 75] speed: 1510 samples/sec	time cost: 869.540569
[Epoch 75] validation: err-top1=0.318095 err-top5=0.113668
Epoch[76] Batch [49]	Speed: 1506.961251 samples/sec	accuracy=0.646055	lr=0.127702
Epoch[76] Batch [99]	Speed: 1537.878643 samples/sec	accuracy=0.644043	lr=0.127498
Epoch[76] Batch [149]	Speed: 1517.701380 samples/sec	accuracy=0.642780	lr=0.127295
Epoch[76] Batch [199]	Speed: 1496.360655 samples/sec	accuracy=0.642412	lr=0.127092
Epoch[76] Batch [249]	Speed: 1524.388657 samples/sec	accuracy=0.641949	lr=0.126888
Epoch[76] Batch [299]	Speed: 1533.891177 samples/sec	accuracy=0.641742	lr=0.126685
Epoch[76] Batch [349]	Speed: 1444.933763 samples/sec	accuracy=0.641236	lr=0.126482
Epoch[76] Batch [399]	Speed: 1467.030557 samples/sec	accuracy=0.641262	lr=0.126279
Epoch[76] Batch [449]	Speed: 1525.549449 samples/sec	accuracy=0.641339	lr=0.126076
Epoch[76] Batch [499]	Speed: 1465.820954 samples/sec	accuracy=0.641211	lr=0.125873
Epoch[76] Batch [549]	Speed: 1539.849326 samples/sec	accuracy=0.641014	lr=0.125670
Epoch[76] Batch [599]	Speed: 1464.570783 samples/sec	accuracy=0.641025	lr=0.125468
Epoch[76] Batch [649]	Speed: 1523.554522 samples/sec	accuracy=0.640859	lr=0.125265
Epoch[76] Batch [699]	Speed: 1507.255374 samples/sec	accuracy=0.640610	lr=0.125062
Epoch[76] Batch [749]	Speed: 1538.989400 samples/sec	accuracy=0.640611	lr=0.124860
Epoch[76] Batch [799]	Speed: 1521.459033 samples/sec	accuracy=0.640428	lr=0.124658
Epoch[76] Batch [849]	Speed: 1471.282766 samples/sec	accuracy=0.640253	lr=0.124455
Epoch[76] Batch [899]	Speed: 1538.435772 samples/sec	accuracy=0.640294	lr=0.124253
Epoch[76] Batch [949]	Speed: 1540.754197 samples/sec	accuracy=0.640000	lr=0.124051
Epoch[76] Batch [999]	Speed: 1474.387415 samples/sec	accuracy=0.639732	lr=0.123849
Epoch[76] Batch [1049]	Speed: 1505.923188 samples/sec	accuracy=0.639324	lr=0.123647
Epoch[76] Batch [1099]	Speed: 1505.053555 samples/sec	accuracy=0.639338	lr=0.123446
Epoch[76] Batch [1149]	Speed: 1476.167815 samples/sec	accuracy=0.639255	lr=0.123244
Epoch[76] Batch [1199]	Speed: 1502.204296 samples/sec	accuracy=0.639214	lr=0.123042
Epoch[76] Batch [1249]	Speed: 1507.551559 samples/sec	accuracy=0.638933	lr=0.122841
[Epoch 76] training: accuracy=0.638945
[Epoch 76] speed: 1503 samples/sec	time cost: 873.652626
[Epoch 76] validation: err-top1=0.313736 err-top5=0.109853
Epoch[77] Batch [49]	Speed: 1530.399476 samples/sec	accuracy=0.647012	lr=0.122635
Epoch[77] Batch [99]	Speed: 1524.571205 samples/sec	accuracy=0.646816	lr=0.122434
Epoch[77] Batch [149]	Speed: 1511.710395 samples/sec	accuracy=0.645312	lr=0.122233
Epoch[77] Batch [199]	Speed: 1511.430583 samples/sec	accuracy=0.645386	lr=0.122032
Epoch[77] Batch [249]	Speed: 1546.072812 samples/sec	accuracy=0.644738	lr=0.121831
Epoch[77] Batch [299]	Speed: 1501.054351 samples/sec	accuracy=0.644710	lr=0.121630
Epoch[77] Batch [349]	Speed: 1504.754101 samples/sec	accuracy=0.644746	lr=0.121429
Epoch[77] Batch [399]	Speed: 1517.968035 samples/sec	accuracy=0.644512	lr=0.121228
Epoch[77] Batch [449]	Speed: 1512.087191 samples/sec	accuracy=0.643919	lr=0.121027
Epoch[77] Batch [499]	Speed: 1533.903678 samples/sec	accuracy=0.643807	lr=0.120827
Epoch[77] Batch [549]	Speed: 1476.485974 samples/sec	accuracy=0.643475	lr=0.120626
Epoch[77] Batch [599]	Speed: 1509.516930 samples/sec	accuracy=0.643403	lr=0.120426
Epoch[77] Batch [649]	Speed: 1538.149010 samples/sec	accuracy=0.643290	lr=0.120226
Epoch[77] Batch [699]	Speed: 1544.998170 samples/sec	accuracy=0.643079	lr=0.120025
Epoch[77] Batch [749]	Speed: 1494.999481 samples/sec	accuracy=0.642867	lr=0.119825
Epoch[77] Batch [799]	Speed: 1536.521878 samples/sec	accuracy=0.642581	lr=0.119625
Epoch[77] Batch [849]	Speed: 1471.190065 samples/sec	accuracy=0.642397	lr=0.119425
Epoch[77] Batch [899]	Speed: 1538.720447 samples/sec	accuracy=0.642153	lr=0.119226
Epoch[77] Batch [949]	Speed: 1527.249629 samples/sec	accuracy=0.642026	lr=0.119026
Epoch[77] Batch [999]	Speed: 1504.865453 samples/sec	accuracy=0.641979	lr=0.118826
Epoch[77] Batch [1049]	Speed: 1517.871086 samples/sec	accuracy=0.641830	lr=0.118627
Epoch[77] Batch [1099]	Speed: 1536.999410 samples/sec	accuracy=0.641805	lr=0.118427
Epoch[77] Batch [1149]	Speed: 1504.561858 samples/sec	accuracy=0.641799	lr=0.118228
Epoch[77] Batch [1199]	Speed: 1438.527997 samples/sec	accuracy=0.641664	lr=0.118029
Epoch[77] Batch [1249]	Speed: 1515.030724 samples/sec	accuracy=0.641520	lr=0.117830
[Epoch 77] training: accuracy=0.641516
[Epoch 77] speed: 1512 samples/sec	time cost: 869.189171
[Epoch 77] validation: err-top1=0.316187 err-top5=0.110830
Epoch[78] Batch [49]	Speed: 1446.571743 samples/sec	accuracy=0.656113	lr=0.117627
Epoch[78] Batch [99]	Speed: 1534.411399 samples/sec	accuracy=0.654336	lr=0.117428
Epoch[78] Batch [149]	Speed: 1517.925900 samples/sec	accuracy=0.653522	lr=0.117229
Epoch[78] Batch [199]	Speed: 1548.777087 samples/sec	accuracy=0.652100	lr=0.117030
Epoch[78] Batch [249]	Speed: 1445.736214 samples/sec	accuracy=0.651031	lr=0.116831
Epoch[78] Batch [299]	Speed: 1505.705784 samples/sec	accuracy=0.649873	lr=0.116633
Epoch[78] Batch [349]	Speed: 1492.543760 samples/sec	accuracy=0.649216	lr=0.116434
Epoch[78] Batch [399]	Speed: 1511.829122 samples/sec	accuracy=0.648906	lr=0.116236
Epoch[78] Batch [449]	Speed: 1486.926118 samples/sec	accuracy=0.648340	lr=0.116038
Epoch[78] Batch [499]	Speed: 1467.367359 samples/sec	accuracy=0.647938	lr=0.115840
Epoch[78] Batch [549]	Speed: 1537.159717 samples/sec	accuracy=0.647477	lr=0.115642
Epoch[78] Batch [599]	Speed: 1481.566634 samples/sec	accuracy=0.647482	lr=0.115444
Epoch[78] Batch [649]	Speed: 1511.201823 samples/sec	accuracy=0.647215	lr=0.115246
Epoch[78] Batch [699]	Speed: 1500.581673 samples/sec	accuracy=0.646752	lr=0.115048
Epoch[78] Batch [749]	Speed: 1540.977253 samples/sec	accuracy=0.646492	lr=0.114850
Epoch[78] Batch [799]	Speed: 1493.876394 samples/sec	accuracy=0.646298	lr=0.114653
Epoch[78] Batch [849]	Speed: 1466.085053 samples/sec	accuracy=0.646087	lr=0.114455
Epoch[78] Batch [899]	Speed: 1541.044896 samples/sec	accuracy=0.645970	lr=0.114258
Epoch[78] Batch [949]	Speed: 1411.381709 samples/sec	accuracy=0.645768	lr=0.114061
Epoch[78] Batch [999]	Speed: 1506.810057 samples/sec	accuracy=0.645771	lr=0.113864
Epoch[78] Batch [1049]	Speed: 1467.083264 samples/sec	accuracy=0.645607	lr=0.113667
Epoch[78] Batch [1099]	Speed: 1533.771249 samples/sec	accuracy=0.645739	lr=0.113470
Epoch[78] Batch [1149]	Speed: 1498.722226 samples/sec	accuracy=0.645489	lr=0.113273
Epoch[78] Batch [1199]	Speed: 1515.360993 samples/sec	accuracy=0.645399	lr=0.113076
Epoch[78] Batch [1249]	Speed: 1492.677340 samples/sec	accuracy=0.645354	lr=0.112880
[Epoch 78] training: accuracy=0.645347
[Epoch 78] speed: 1496 samples/sec	time cost: 878.748135
[Epoch 78] validation: err-top1=0.314533 err-top5=0.110471
Epoch[79] Batch [49]	Speed: 1517.024606 samples/sec	accuracy=0.655762	lr=0.112679
Epoch[79] Batch [99]	Speed: 1476.254283 samples/sec	accuracy=0.656045	lr=0.112483
Epoch[79] Batch [149]	Speed: 1504.781843 samples/sec	accuracy=0.653893	lr=0.112287
Epoch[79] Batch [199]	Speed: 1513.648276 samples/sec	accuracy=0.652490	lr=0.112090
Epoch[79] Batch [249]	Speed: 1510.473352 samples/sec	accuracy=0.652496	lr=0.111894
Epoch[79] Batch [299]	Speed: 1540.232284 samples/sec	accuracy=0.652425	lr=0.111698
Epoch[79] Batch [349]	Speed: 1386.236485 samples/sec	accuracy=0.652012	lr=0.111502
Epoch[79] Batch [399]	Speed: 1465.251761 samples/sec	accuracy=0.651497	lr=0.111307
Epoch[79] Batch [449]	Speed: 1496.568767 samples/sec	accuracy=0.651539	lr=0.111111
Epoch[79] Batch [499]	Speed: 1516.121277 samples/sec	accuracy=0.650928	lr=0.110915
Epoch[79] Batch [549]	Speed: 1478.791464 samples/sec	accuracy=0.650730	lr=0.110720
Epoch[79] Batch [599]	Speed: 1463.245938 samples/sec	accuracy=0.650474	lr=0.110525
Epoch[79] Batch [649]	Speed: 1519.777468 samples/sec	accuracy=0.650308	lr=0.110329
Epoch[79] Batch [699]	Speed: 1478.921566 samples/sec	accuracy=0.649763	lr=0.110134
Epoch[79] Batch [749]	Speed: 1511.637333 samples/sec	accuracy=0.649271	lr=0.109939
Epoch[79] Batch [799]	Speed: 1540.608292 samples/sec	accuracy=0.648842	lr=0.109744
Epoch[79] Batch [849]	Speed: 1516.361262 samples/sec	accuracy=0.648674	lr=0.109549
Epoch[79] Batch [899]	Speed: 1519.460161 samples/sec	accuracy=0.648581	lr=0.109355
Epoch[79] Batch [949]	Speed: 1447.495516 samples/sec	accuracy=0.648506	lr=0.109160
Epoch[79] Batch [999]	Speed: 1449.621730 samples/sec	accuracy=0.648460	lr=0.108966
Epoch[79] Batch [1049]	Speed: 1410.022616 samples/sec	accuracy=0.648269	lr=0.108771
Epoch[79] Batch [1099]	Speed: 1515.152069 samples/sec	accuracy=0.648343	lr=0.108577
Epoch[79] Batch [1149]	Speed: 1521.514020 samples/sec	accuracy=0.648323	lr=0.108383
Epoch[79] Batch [1199]	Speed: 1528.146131 samples/sec	accuracy=0.648190	lr=0.108189
Epoch[79] Batch [1249]	Speed: 1518.838976 samples/sec	accuracy=0.648234	lr=0.107995
[Epoch 79] training: accuracy=0.648255
[Epoch 79] speed: 1491 samples/sec	time cost: 880.816744
[Epoch 79] validation: err-top1=0.309590 err-top5=0.108498
Epoch[80] Batch [49]	Speed: 1525.660702 samples/sec	accuracy=0.660859	lr=0.107797
Epoch[80] Batch [99]	Speed: 1484.426545 samples/sec	accuracy=0.659014	lr=0.107603
Epoch[80] Batch [149]	Speed: 1513.681340 samples/sec	accuracy=0.656966	lr=0.107410
Epoch[80] Batch [199]	Speed: 1468.091556 samples/sec	accuracy=0.657437	lr=0.107216
Epoch[80] Batch [249]	Speed: 1514.911707 samples/sec	accuracy=0.657371	lr=0.107023
Epoch[80] Batch [299]	Speed: 1490.365257 samples/sec	accuracy=0.657054	lr=0.106830
Epoch[80] Batch [349]	Speed: 1545.021757 samples/sec	accuracy=0.656540	lr=0.106636
Epoch[80] Batch [399]	Speed: 1506.380049 samples/sec	accuracy=0.656311	lr=0.106443
Epoch[80] Batch [449]	Speed: 1476.271190 samples/sec	accuracy=0.656074	lr=0.106250
Epoch[80] Batch [499]	Speed: 1541.960075 samples/sec	accuracy=0.655891	lr=0.106057
Epoch[80] Batch [549]	Speed: 1516.053664 samples/sec	accuracy=0.655559	lr=0.105865
Epoch[80] Batch [599]	Speed: 1516.754382 samples/sec	accuracy=0.655410	lr=0.105672
Epoch[80] Batch [649]	Speed: 1520.066684 samples/sec	accuracy=0.654955	lr=0.105480
Epoch[80] Batch [699]	Speed: 1520.941921 samples/sec	accuracy=0.654789	lr=0.105287
Epoch[80] Batch [749]	Speed: 1493.613596 samples/sec	accuracy=0.654378	lr=0.105095
Epoch[80] Batch [799]	Speed: 1509.742283 samples/sec	accuracy=0.654208	lr=0.104903
Epoch[80] Batch [849]	Speed: 1521.316942 samples/sec	accuracy=0.653942	lr=0.104711
Epoch[80] Batch [899]	Speed: 1511.992535 samples/sec	accuracy=0.653834	lr=0.104519
Epoch[80] Batch [949]	Speed: 1544.753892 samples/sec	accuracy=0.653584	lr=0.104327
Epoch[80] Batch [999]	Speed: 1485.903209 samples/sec	accuracy=0.653261	lr=0.104135
Epoch[80] Batch [1049]	Speed: 1464.341348 samples/sec	accuracy=0.652981	lr=0.103944
Epoch[80] Batch [1099]	Speed: 1492.351025 samples/sec	accuracy=0.652843	lr=0.103752
Epoch[80] Batch [1149]	Speed: 1516.084136 samples/sec	accuracy=0.652650	lr=0.103561
Epoch[80] Batch [1199]	Speed: 1491.803026 samples/sec	accuracy=0.652514	lr=0.103370
Epoch[80] Batch [1249]	Speed: 1523.525739 samples/sec	accuracy=0.652392	lr=0.103178
[Epoch 80] training: accuracy=0.652401
[Epoch 80] speed: 1506 samples/sec	time cost: 872.375962
[Epoch 80] validation: err-top1=0.308693 err-top5=0.107900
Epoch[81] Batch [49]	Speed: 1485.497175 samples/sec	accuracy=0.662109	lr=0.102984
Epoch[81] Batch [99]	Speed: 1547.098669 samples/sec	accuracy=0.661660	lr=0.102793
Epoch[81] Batch [149]	Speed: 1490.540761 samples/sec	accuracy=0.659870	lr=0.102602
Epoch[81] Batch [199]	Speed: 1520.434967 samples/sec	accuracy=0.660469	lr=0.102411
Epoch[81] Batch [249]	Speed: 1524.677964 samples/sec	accuracy=0.659762	lr=0.102221
Epoch[81] Batch [299]	Speed: 1516.505095 samples/sec	accuracy=0.659248	lr=0.102030
Epoch[81] Batch [349]	Speed: 1495.438657 samples/sec	accuracy=0.658954	lr=0.101840
Epoch[81] Batch [399]	Speed: 1529.935769 samples/sec	accuracy=0.658743	lr=0.101650
Epoch[81] Batch [449]	Speed: 1477.298843 samples/sec	accuracy=0.658249	lr=0.101460
Epoch[81] Batch [499]	Speed: 1517.638323 samples/sec	accuracy=0.658258	lr=0.101270
Epoch[81] Batch [549]	Speed: 1494.354819 samples/sec	accuracy=0.657962	lr=0.101080
Epoch[81] Batch [599]	Speed: 1524.283117 samples/sec	accuracy=0.657922	lr=0.100890
Epoch[81] Batch [649]	Speed: 1520.157522 samples/sec	accuracy=0.657635	lr=0.100701
Epoch[81] Batch [699]	Speed: 1486.795716 samples/sec	accuracy=0.657309	lr=0.100511
Epoch[81] Batch [749]	Speed: 1513.207896 samples/sec	accuracy=0.656895	lr=0.100322
Epoch[81] Batch [799]	Speed: 1435.619089 samples/sec	accuracy=0.656497	lr=0.100132
Epoch[81] Batch [849]	Speed: 1501.455320 samples/sec	accuracy=0.656097	lr=0.099943
Epoch[81] Batch [899]	Speed: 1511.854709 samples/sec	accuracy=0.655895	lr=0.099754
Epoch[81] Batch [949]	Speed: 1441.182655 samples/sec	accuracy=0.655714	lr=0.099565
Epoch[81] Batch [999]	Speed: 1546.322651 samples/sec	accuracy=0.655593	lr=0.099377
Epoch[81] Batch [1049]	Speed: 1464.962908 samples/sec	accuracy=0.655577	lr=0.099188
Epoch[81] Batch [1099]	Speed: 1495.839225 samples/sec	accuracy=0.655361	lr=0.098999
Epoch[81] Batch [1149]	Speed: 1475.221282 samples/sec	accuracy=0.655235	lr=0.098811
Epoch[81] Batch [1199]	Speed: 1544.937905 samples/sec	accuracy=0.655049	lr=0.098623
Epoch[81] Batch [1249]	Speed: 1524.025195 samples/sec	accuracy=0.654823	lr=0.098434
[Epoch 81] training: accuracy=0.654813
[Epoch 81] speed: 1501 samples/sec	time cost: 874.267366
[Epoch 81] validation: err-top1=0.303752 err-top5=0.106262
Epoch[82] Batch [49]	Speed: 1486.223296 samples/sec	accuracy=0.665039	lr=0.098243
Epoch[82] Batch [99]	Speed: 1501.088766 samples/sec	accuracy=0.664873	lr=0.098055
Epoch[82] Batch [149]	Speed: 1547.818781 samples/sec	accuracy=0.663848	lr=0.097867
Epoch[82] Batch [199]	Speed: 1521.806788 samples/sec	accuracy=0.663940	lr=0.097679
Epoch[82] Batch [249]	Speed: 1513.262715 samples/sec	accuracy=0.663687	lr=0.097492
Epoch[82] Batch [299]	Speed: 1490.633298 samples/sec	accuracy=0.663057	lr=0.097304
Epoch[82] Batch [349]	Speed: 1511.294486 samples/sec	accuracy=0.662885	lr=0.097117
Epoch[82] Batch [399]	Speed: 1551.294201 samples/sec	accuracy=0.662251	lr=0.096930
Epoch[82] Batch [449]	Speed: 1534.561966 samples/sec	accuracy=0.662079	lr=0.096743
Epoch[82] Batch [499]	Speed: 1440.049169 samples/sec	accuracy=0.661953	lr=0.096556
Epoch[82] Batch [549]	Speed: 1513.899805 samples/sec	accuracy=0.661790	lr=0.096369
Epoch[82] Batch [599]	Speed: 1497.325674 samples/sec	accuracy=0.661579	lr=0.096182
Epoch[82] Batch [649]	Speed: 1536.463041 samples/sec	accuracy=0.661508	lr=0.095996
Epoch[82] Batch [699]	Speed: 1474.857082 samples/sec	accuracy=0.661147	lr=0.095809
Epoch[82] Batch [749]	Speed: 1537.215404 samples/sec	accuracy=0.661086	lr=0.095623
Epoch[82] Batch [799]	Speed: 1512.751254 samples/sec	accuracy=0.660942	lr=0.095437
Epoch[82] Batch [849]	Speed: 1441.508843 samples/sec	accuracy=0.660535	lr=0.095250
Epoch[82] Batch [899]	Speed: 1518.354205 samples/sec	accuracy=0.660125	lr=0.095064
Epoch[82] Batch [949]	Speed: 1505.745923 samples/sec	accuracy=0.659857	lr=0.094879
Epoch[82] Batch [999]	Speed: 1543.180671 samples/sec	accuracy=0.659630	lr=0.094693
Epoch[82] Batch [1049]	Speed: 1539.436795 samples/sec	accuracy=0.659319	lr=0.094507
Epoch[82] Batch [1099]	Speed: 1489.879729 samples/sec	accuracy=0.659148	lr=0.094322
Epoch[82] Batch [1149]	Speed: 1541.152924 samples/sec	accuracy=0.659006	lr=0.094137
Epoch[82] Batch [1199]	Speed: 1526.781978 samples/sec	accuracy=0.658671	lr=0.093951
Epoch[82] Batch [1249]	Speed: 1518.279427 samples/sec	accuracy=0.658597	lr=0.093766
[Epoch 82] training: accuracy=0.658577
[Epoch 82] speed: 1510 samples/sec	time cost: 870.590450
[Epoch 82] validation: err-top1=0.305146 err-top5=0.103715
Epoch[83] Batch [49]	Speed: 1438.647091 samples/sec	accuracy=0.667363	lr=0.093578
Epoch[83] Batch [99]	Speed: 1535.034813 samples/sec	accuracy=0.667314	lr=0.093393
Epoch[83] Batch [149]	Speed: 1510.771633 samples/sec	accuracy=0.667331	lr=0.093208
Epoch[83] Batch [199]	Speed: 1518.329461 samples/sec	accuracy=0.665615	lr=0.093024
Epoch[83] Batch [249]	Speed: 1541.146763 samples/sec	accuracy=0.665824	lr=0.092839
Epoch[83] Batch [299]	Speed: 1474.967295 samples/sec	accuracy=0.665150	lr=0.092655
Epoch[83] Batch [349]	Speed: 1534.524618 samples/sec	accuracy=0.664805	lr=0.092471
Epoch[83] Batch [399]	Speed: 1504.246405 samples/sec	accuracy=0.664116	lr=0.092287
Epoch[83] Batch [449]	Speed: 1547.622102 samples/sec	accuracy=0.664097	lr=0.092103
Epoch[83] Batch [499]	Speed: 1518.891292 samples/sec	accuracy=0.663770	lr=0.091919
Epoch[83] Batch [549]	Speed: 1484.306163 samples/sec	accuracy=0.663413	lr=0.091735
Epoch[83] Batch [599]	Speed: 1510.896495 samples/sec	accuracy=0.663431	lr=0.091552
Epoch[83] Batch [649]	Speed: 1539.599189 samples/sec	accuracy=0.663622	lr=0.091368
Epoch[83] Batch [699]	Speed: 1538.783558 samples/sec	accuracy=0.663608	lr=0.091185
Epoch[83] Batch [749]	Speed: 1520.717670 samples/sec	accuracy=0.663335	lr=0.091002
Epoch[83] Batch [799]	Speed: 1539.517866 samples/sec	accuracy=0.663286	lr=0.090819
Epoch[83] Batch [849]	Speed: 1519.424254 samples/sec	accuracy=0.663165	lr=0.090636
Epoch[83] Batch [899]	Speed: 1513.824302 samples/sec	accuracy=0.662850	lr=0.090453
Epoch[83] Batch [949]	Speed: 1539.471580 samples/sec	accuracy=0.662653	lr=0.090270
Epoch[83] Batch [999]	Speed: 1516.732893 samples/sec	accuracy=0.662686	lr=0.090088
Epoch[83] Batch [1049]	Speed: 1489.814974 samples/sec	accuracy=0.662655	lr=0.089905
Epoch[83] Batch [1099]	Speed: 1534.739468 samples/sec	accuracy=0.662668	lr=0.089723
Epoch[83] Batch [1149]	Speed: 1485.299085 samples/sec	accuracy=0.662478	lr=0.089541
Epoch[83] Batch [1199]	Speed: 1470.787478 samples/sec	accuracy=0.662346	lr=0.089359
Epoch[83] Batch [1249]	Speed: 1547.238549 samples/sec	accuracy=0.662172	lr=0.089177
[Epoch 83] training: accuracy=0.662162
[Epoch 83] speed: 1513 samples/sec	time cost: 868.578231
[Epoch 83] validation: err-top1=0.302974 err-top5=0.103077
Epoch[84] Batch [49]	Speed: 1506.104932 samples/sec	accuracy=0.671621	lr=0.088992
Epoch[84] Batch [99]	Speed: 1544.862852 samples/sec	accuracy=0.672764	lr=0.088810
Epoch[84] Batch [149]	Speed: 1504.738465 samples/sec	accuracy=0.671803	lr=0.088629
Epoch[84] Batch [199]	Speed: 1537.615846 samples/sec	accuracy=0.671890	lr=0.088448
Epoch[84] Batch [249]	Speed: 1473.235601 samples/sec	accuracy=0.670813	lr=0.088266
Epoch[84] Batch [299]	Speed: 1543.321651 samples/sec	accuracy=0.670306	lr=0.088085
Epoch[84] Batch [349]	Speed: 1516.589927 samples/sec	accuracy=0.669950	lr=0.087904
Epoch[84] Batch [399]	Speed: 1523.904058 samples/sec	accuracy=0.669756	lr=0.087724
Epoch[84] Batch [449]	Speed: 1541.128061 samples/sec	accuracy=0.669221	lr=0.087543
Epoch[84] Batch [499]	Speed: 1486.233890 samples/sec	accuracy=0.668959	lr=0.087363
Epoch[84] Batch [549]	Speed: 1458.416958 samples/sec	accuracy=0.668624	lr=0.087182
Epoch[84] Batch [599]	Speed: 1492.447801 samples/sec	accuracy=0.668905	lr=0.087002
Epoch[84] Batch [649]	Speed: 1529.306154 samples/sec	accuracy=0.668471	lr=0.086822
Epoch[84] Batch [699]	Speed: 1499.560996 samples/sec	accuracy=0.668284	lr=0.086642
Epoch[84] Batch [749]	Speed: 1524.497457 samples/sec	accuracy=0.668061	lr=0.086462
Epoch[84] Batch [799]	Speed: 1432.574819 samples/sec	accuracy=0.667819	lr=0.086282
Epoch[84] Batch [849]	Speed: 1492.230557 samples/sec	accuracy=0.667594	lr=0.086103
Epoch[84] Batch [899]	Speed: 1534.008866 samples/sec	accuracy=0.667145	lr=0.085923
Epoch[84] Batch [949]	Speed: 1462.928256 samples/sec	accuracy=0.666856	lr=0.085744
Epoch[84] Batch [999]	Speed: 1489.308919 samples/sec	accuracy=0.666759	lr=0.085565
Epoch[84] Batch [1049]	Speed: 1479.414551 samples/sec	accuracy=0.666651	lr=0.085386
Epoch[84] Batch [1099]	Speed: 1512.111381 samples/sec	accuracy=0.666455	lr=0.085207
Epoch[84] Batch [1149]	Speed: 1493.636191 samples/sec	accuracy=0.666275	lr=0.085028
Epoch[84] Batch [1199]	Speed: 1541.732252 samples/sec	accuracy=0.666029	lr=0.084850
Epoch[84] Batch [1249]	Speed: 1548.390014 samples/sec	accuracy=0.666073	lr=0.084671
[Epoch 84] training: accuracy=0.666069
[Epoch 84] speed: 1504 samples/sec	time cost: 872.967762
[Epoch 84] validation: err-top1=0.296417 err-top5=0.100048
Epoch[85] Batch [49]	Speed: 1517.987070 samples/sec	accuracy=0.678730	lr=0.084489
Epoch[85] Batch [99]	Speed: 1489.233380 samples/sec	accuracy=0.677695	lr=0.084311
Epoch[85] Batch [149]	Speed: 1537.977724 samples/sec	accuracy=0.676250	lr=0.084133
Epoch[85] Batch [199]	Speed: 1548.308485 samples/sec	accuracy=0.674946	lr=0.083955
Epoch[85] Batch [249]	Speed: 1540.323603 samples/sec	accuracy=0.674316	lr=0.083777
Epoch[85] Batch [299]	Speed: 1519.067217 samples/sec	accuracy=0.673968	lr=0.083600
Epoch[85] Batch [349]	Speed: 1504.182597 samples/sec	accuracy=0.673571	lr=0.083422
Epoch[85] Batch [399]	Speed: 1460.756330 samples/sec	accuracy=0.673518	lr=0.083245
Epoch[85] Batch [449]	Speed: 1477.141888 samples/sec	accuracy=0.672921	lr=0.083067
Epoch[85] Batch [499]	Speed: 1514.750408 samples/sec	accuracy=0.672348	lr=0.082890
Epoch[85] Batch [549]	Speed: 1526.308397 samples/sec	accuracy=0.671946	lr=0.082713
Epoch[85] Batch [599]	Speed: 1518.658249 samples/sec	accuracy=0.671865	lr=0.082537
Epoch[85] Batch [649]	Speed: 1545.333039 samples/sec	accuracy=0.671618	lr=0.082360
Epoch[85] Batch [699]	Speed: 1518.591172 samples/sec	accuracy=0.671240	lr=0.082183
Epoch[85] Batch [749]	Speed: 1494.193740 samples/sec	accuracy=0.671258	lr=0.082007
Epoch[85] Batch [799]	Speed: 1490.229649 samples/sec	accuracy=0.670863	lr=0.081831
Epoch[85] Batch [849]	Speed: 1485.706018 samples/sec	accuracy=0.670326	lr=0.081655
Epoch[85] Batch [899]	Speed: 1502.629063 samples/sec	accuracy=0.670025	lr=0.081479
Epoch[85] Batch [949]	Speed: 1530.752771 samples/sec	accuracy=0.669976	lr=0.081303
Epoch[85] Batch [999]	Speed: 1515.637085 samples/sec	accuracy=0.669845	lr=0.081127
Epoch[85] Batch [1049]	Speed: 1514.710503 samples/sec	accuracy=0.669746	lr=0.080952
Epoch[85] Batch [1099]	Speed: 1446.492458 samples/sec	accuracy=0.669549	lr=0.080776
Epoch[85] Batch [1149]	Speed: 1505.282515 samples/sec	accuracy=0.669446	lr=0.080601
Epoch[85] Batch [1199]	Speed: 1532.310516 samples/sec	accuracy=0.669365	lr=0.080426
Epoch[85] Batch [1249]	Speed: 1547.560918 samples/sec	accuracy=0.669120	lr=0.080251
[Epoch 85] training: accuracy=0.669147
[Epoch 85] speed: 1509 samples/sec	time cost: 870.687451
[Epoch 85] validation: err-top1=0.296317 err-top5=0.100965
Epoch[86] Batch [49]	Speed: 1448.781590 samples/sec	accuracy=0.679355	lr=0.080073
Epoch[86] Batch [99]	Speed: 1486.654303 samples/sec	accuracy=0.678643	lr=0.079898
Epoch[86] Batch [149]	Speed: 1516.015017 samples/sec	accuracy=0.678607	lr=0.079723
Epoch[86] Batch [199]	Speed: 1484.262768 samples/sec	accuracy=0.677412	lr=0.079549
Epoch[86] Batch [249]	Speed: 1521.551525 samples/sec	accuracy=0.676398	lr=0.079375
Epoch[86] Batch [299]	Speed: 1519.418008 samples/sec	accuracy=0.676182	lr=0.079201
Epoch[86] Batch [349]	Speed: 1493.977983 samples/sec	accuracy=0.676122	lr=0.079027
Epoch[86] Batch [399]	Speed: 1517.082585 samples/sec	accuracy=0.675901	lr=0.078853
Epoch[86] Batch [449]	Speed: 1514.767130 samples/sec	accuracy=0.676050	lr=0.078679
Epoch[86] Batch [499]	Speed: 1477.484132 samples/sec	accuracy=0.675559	lr=0.078506
Epoch[86] Batch [549]	Speed: 1515.670139 samples/sec	accuracy=0.675330	lr=0.078332
Epoch[86] Batch [599]	Speed: 1492.284654 samples/sec	accuracy=0.675047	lr=0.078159
Epoch[86] Batch [649]	Speed: 1535.605663 samples/sec	accuracy=0.674627	lr=0.077986
Epoch[86] Batch [699]	Speed: 1494.354871 samples/sec	accuracy=0.674569	lr=0.077813
Epoch[86] Batch [749]	Speed: 1439.282909 samples/sec	accuracy=0.674374	lr=0.077640
Epoch[86] Batch [799]	Speed: 1489.821909 samples/sec	accuracy=0.674337	lr=0.077467
Epoch[86] Batch [849]	Speed: 1486.195802 samples/sec	accuracy=0.674256	lr=0.077295
Epoch[86] Batch [899]	Speed: 1515.376519 samples/sec	accuracy=0.673980	lr=0.077123
Epoch[86] Batch [949]	Speed: 1539.567401 samples/sec	accuracy=0.673763	lr=0.076950
Epoch[86] Batch [999]	Speed: 1542.131518 samples/sec	accuracy=0.673767	lr=0.076778
Epoch[86] Batch [1049]	Speed: 1516.291465 samples/sec	accuracy=0.673804	lr=0.076606
Epoch[86] Batch [1099]	Speed: 1516.091896 samples/sec	accuracy=0.673639	lr=0.076435
Epoch[86] Batch [1149]	Speed: 1513.776228 samples/sec	accuracy=0.673464	lr=0.076263
Epoch[86] Batch [1199]	Speed: 1459.682923 samples/sec	accuracy=0.673304	lr=0.076091
Epoch[86] Batch [1249]	Speed: 1544.175611 samples/sec	accuracy=0.673290	lr=0.075920
[Epoch 86] training: accuracy=0.673301
[Epoch 86] speed: 1501 samples/sec	time cost: 874.813997
[Epoch 86] validation: err-top1=0.296058 err-top5=0.099609
Epoch[87] Batch [49]	Speed: 1455.913300 samples/sec	accuracy=0.678535	lr=0.075745
Epoch[87] Batch [99]	Speed: 1492.248921 samples/sec	accuracy=0.680449	lr=0.075574
Epoch[87] Batch [149]	Speed: 1507.721840 samples/sec	accuracy=0.680605	lr=0.075404
Epoch[87] Batch [199]	Speed: 1483.995126 samples/sec	accuracy=0.681284	lr=0.075233
Epoch[87] Batch [249]	Speed: 1506.711176 samples/sec	accuracy=0.680383	lr=0.075062
Epoch[87] Batch [299]	Speed: 1414.482524 samples/sec	accuracy=0.680342	lr=0.074892
Epoch[87] Batch [349]	Speed: 1544.437523 samples/sec	accuracy=0.679766	lr=0.074721
Epoch[87] Batch [399]	Speed: 1482.925011 samples/sec	accuracy=0.679570	lr=0.074551
Epoch[87] Batch [449]	Speed: 1540.468416 samples/sec	accuracy=0.679403	lr=0.074381
Epoch[87] Batch [499]	Speed: 1546.132386 samples/sec	accuracy=0.679305	lr=0.074211
Epoch[87] Batch [549]	Speed: 1510.717632 samples/sec	accuracy=0.679435	lr=0.074042
Epoch[87] Batch [599]	Speed: 1474.888402 samples/sec	accuracy=0.679386	lr=0.073872
Epoch[87] Batch [649]	Speed: 1538.063235 samples/sec	accuracy=0.679072	lr=0.073703
Epoch[87] Batch [699]	Speed: 1511.465422 samples/sec	accuracy=0.678781	lr=0.073534
Epoch[87] Batch [749]	Speed: 1518.778886 samples/sec	accuracy=0.678354	lr=0.073364
Epoch[87] Batch [799]	Speed: 1541.574343 samples/sec	accuracy=0.678220	lr=0.073196
Epoch[87] Batch [849]	Speed: 1526.405016 samples/sec	accuracy=0.677961	lr=0.073027
Epoch[87] Batch [899]	Speed: 1535.154127 samples/sec	accuracy=0.678027	lr=0.072858
Epoch[87] Batch [949]	Speed: 1515.519010 samples/sec	accuracy=0.677787	lr=0.072690
Epoch[87] Batch [999]	Speed: 1510.792496 samples/sec	accuracy=0.677738	lr=0.072521
Epoch[87] Batch [1049]	Speed: 1514.116850 samples/sec	accuracy=0.677561	lr=0.072353
Epoch[87] Batch [1099]	Speed: 1484.226535 samples/sec	accuracy=0.677464	lr=0.072185
Epoch[87] Batch [1149]	Speed: 1494.786259 samples/sec	accuracy=0.677360	lr=0.072017
Epoch[87] Batch [1199]	Speed: 1481.768903 samples/sec	accuracy=0.677321	lr=0.071849
Epoch[87] Batch [1249]	Speed: 1520.656172 samples/sec	accuracy=0.677158	lr=0.071682
[Epoch 87] training: accuracy=0.677177
[Epoch 87] speed: 1504 samples/sec	time cost: 872.909597
[Epoch 87] validation: err-top1=0.291382 err-top5=0.096375
Epoch[88] Batch [49]	Speed: 1483.903226 samples/sec	accuracy=0.687754	lr=0.071511
Epoch[88] Batch [99]	Speed: 1504.057493 samples/sec	accuracy=0.688701	lr=0.071344
Epoch[88] Batch [149]	Speed: 1494.240692 samples/sec	accuracy=0.687604	lr=0.071177
Epoch[88] Batch [199]	Speed: 1518.950799 samples/sec	accuracy=0.686738	lr=0.071010
Epoch[88] Batch [249]	Speed: 1515.053437 samples/sec	accuracy=0.685453	lr=0.070843
Epoch[88] Batch [299]	Speed: 1495.014635 samples/sec	accuracy=0.684867	lr=0.070676
Epoch[88] Batch [349]	Speed: 1511.733370 samples/sec	accuracy=0.684372	lr=0.070510
Epoch[88] Batch [399]	Speed: 1535.120305 samples/sec	accuracy=0.684253	lr=0.070343
Epoch[88] Batch [449]	Speed: 1542.326116 samples/sec	accuracy=0.684119	lr=0.070177
Epoch[88] Batch [499]	Speed: 1484.337024 samples/sec	accuracy=0.684162	lr=0.070011
Epoch[88] Batch [549]	Speed: 1495.950355 samples/sec	accuracy=0.683883	lr=0.069845
Epoch[88] Batch [599]	Speed: 1506.644960 samples/sec	accuracy=0.683734	lr=0.069680
Epoch[88] Batch [649]	Speed: 1520.392565 samples/sec	accuracy=0.683338	lr=0.069514
Epoch[88] Batch [699]	Speed: 1538.187163 samples/sec	accuracy=0.683018	lr=0.069349
Epoch[88] Batch [749]	Speed: 1516.749047 samples/sec	accuracy=0.682686	lr=0.069183
Epoch[88] Batch [799]	Speed: 1492.609634 samples/sec	accuracy=0.682582	lr=0.069018
Epoch[88] Batch [849]	Speed: 1492.188408 samples/sec	accuracy=0.682277	lr=0.068853
Epoch[88] Batch [899]	Speed: 1517.388938 samples/sec	accuracy=0.682220	lr=0.068689
Epoch[88] Batch [949]	Speed: 1494.952055 samples/sec	accuracy=0.681889	lr=0.068524
Epoch[88] Batch [999]	Speed: 1521.828486 samples/sec	accuracy=0.681730	lr=0.068359
Epoch[88] Batch [1049]	Speed: 1492.320815 samples/sec	accuracy=0.681530	lr=0.068195
Epoch[88] Batch [1099]	Speed: 1541.742313 samples/sec	accuracy=0.681438	lr=0.068031
Epoch[88] Batch [1149]	Speed: 1474.815828 samples/sec	accuracy=0.681240	lr=0.067867
Epoch[88] Batch [1199]	Speed: 1537.821761 samples/sec	accuracy=0.681156	lr=0.067703
Epoch[88] Batch [1249]	Speed: 1529.926101 samples/sec	accuracy=0.680986	lr=0.067539
[Epoch 88] training: accuracy=0.680995
[Epoch 88] speed: 1508 samples/sec	time cost: 870.969353
[Epoch 88] validation: err-top1=0.291653 err-top5=0.095703
Epoch[89] Batch [49]	Speed: 1485.399993 samples/sec	accuracy=0.694043	lr=0.067373
Epoch[89] Batch [99]	Speed: 1539.357354 samples/sec	accuracy=0.693154	lr=0.067209
Epoch[89] Batch [149]	Speed: 1546.050728 samples/sec	accuracy=0.692155	lr=0.067046
Epoch[89] Batch [199]	Speed: 1498.709476 samples/sec	accuracy=0.691821	lr=0.066883
Epoch[89] Batch [249]	Speed: 1514.730557 samples/sec	accuracy=0.690723	lr=0.066720
Epoch[89] Batch [299]	Speed: 1496.743555 samples/sec	accuracy=0.690150	lr=0.066557
Epoch[89] Batch [349]	Speed: 1476.420947 samples/sec	accuracy=0.689523	lr=0.066395
Epoch[89] Batch [399]	Speed: 1528.494481 samples/sec	accuracy=0.688748	lr=0.066232
Epoch[89] Batch [449]	Speed: 1529.217660 samples/sec	accuracy=0.688199	lr=0.066070
Epoch[89] Batch [499]	Speed: 1524.940947 samples/sec	accuracy=0.687998	lr=0.065908
Epoch[89] Batch [549]	Speed: 1549.819333 samples/sec	accuracy=0.687615	lr=0.065746
Epoch[89] Batch [599]	Speed: 1487.301670 samples/sec	accuracy=0.687280	lr=0.065584
Epoch[89] Batch [649]	Speed: 1495.036242 samples/sec	accuracy=0.687073	lr=0.065423
Epoch[89] Batch [699]	Speed: 1516.515237 samples/sec	accuracy=0.686920	lr=0.065261
Epoch[89] Batch [749]	Speed: 1478.171232 samples/sec	accuracy=0.686661	lr=0.065100
Epoch[89] Batch [799]	Speed: 1471.497613 samples/sec	accuracy=0.686562	lr=0.064939
Epoch[89] Batch [849]	Speed: 1526.383328 samples/sec	accuracy=0.686375	lr=0.064778
Epoch[89] Batch [899]	Speed: 1495.685337 samples/sec	accuracy=0.686075	lr=0.064617
Epoch[89] Batch [949]	Speed: 1492.834564 samples/sec	accuracy=0.685986	lr=0.064456
Epoch[89] Batch [999]	Speed: 1456.006780 samples/sec	accuracy=0.685749	lr=0.064296
Epoch[89] Batch [1049]	Speed: 1534.875564 samples/sec	accuracy=0.685686	lr=0.064135
Epoch[89] Batch [1099]	Speed: 1489.471178 samples/sec	accuracy=0.685568	lr=0.063975
Epoch[89] Batch [1149]	Speed: 1496.838084 samples/sec	accuracy=0.685374	lr=0.063815
Epoch[89] Batch [1199]	Speed: 1539.874722 samples/sec	accuracy=0.685350	lr=0.063655
Epoch[89] Batch [1249]	Speed: 1557.761966 samples/sec	accuracy=0.685259	lr=0.063496
[Epoch 89] training: accuracy=0.685269
[Epoch 89] speed: 1507 samples/sec	time cost: 872.139628
[Epoch 89] validation: err-top1=0.285375 err-top5=0.094766
Epoch[90] Batch [49]	Speed: 1455.480053 samples/sec	accuracy=0.695020	lr=0.063333
Epoch[90] Batch [99]	Speed: 1544.530341 samples/sec	accuracy=0.694102	lr=0.063174
Epoch[90] Batch [149]	Speed: 1521.409924 samples/sec	accuracy=0.693926	lr=0.063015
Epoch[90] Batch [199]	Speed: 1488.797614 samples/sec	accuracy=0.693506	lr=0.062855
Epoch[90] Batch [249]	Speed: 1520.265214 samples/sec	accuracy=0.692332	lr=0.062697
Epoch[90] Batch [299]	Speed: 1452.714558 samples/sec	accuracy=0.692012	lr=0.062538
Epoch[90] Batch [349]	Speed: 1534.623805 samples/sec	accuracy=0.692112	lr=0.062379
Epoch[90] Batch [399]	Speed: 1522.396536 samples/sec	accuracy=0.691580	lr=0.062221
Epoch[90] Batch [449]	Speed: 1548.346630 samples/sec	accuracy=0.691170	lr=0.062063
Epoch[90] Batch [499]	Speed: 1483.364056 samples/sec	accuracy=0.691064	lr=0.061905
Epoch[90] Batch [549]	Speed: 1524.222910 samples/sec	accuracy=0.690922	lr=0.061747
Epoch[90] Batch [599]	Speed: 1520.073312 samples/sec	accuracy=0.690793	lr=0.061589
Epoch[90] Batch [649]	Speed: 1506.338861 samples/sec	accuracy=0.690781	lr=0.061432
Epoch[90] Batch [699]	Speed: 1543.957101 samples/sec	accuracy=0.690686	lr=0.061274
Epoch[90] Batch [749]	Speed: 1469.633807 samples/sec	accuracy=0.690406	lr=0.061117
Epoch[90] Batch [799]	Speed: 1492.880175 samples/sec	accuracy=0.690286	lr=0.060960
Epoch[90] Batch [849]	Speed: 1534.880511 samples/sec	accuracy=0.690030	lr=0.060803
Epoch[90] Batch [899]	Speed: 1494.021003 samples/sec	accuracy=0.689858	lr=0.060646
Epoch[90] Batch [949]	Speed: 1476.882374 samples/sec	accuracy=0.689765	lr=0.060490
Epoch[90] Batch [999]	Speed: 1520.647687 samples/sec	accuracy=0.689567	lr=0.060333
Epoch[90] Batch [1049]	Speed: 1548.159963 samples/sec	accuracy=0.689522	lr=0.060177
Epoch[90] Batch [1099]	Speed: 1478.816027 samples/sec	accuracy=0.689370	lr=0.060021
Epoch[90] Batch [1149]	Speed: 1498.191582 samples/sec	accuracy=0.689287	lr=0.059865
Epoch[90] Batch [1199]	Speed: 1542.724471 samples/sec	accuracy=0.689046	lr=0.059710
Epoch[90] Batch [1249]	Speed: 1517.376190 samples/sec	accuracy=0.689075	lr=0.059554
[Epoch 90] training: accuracy=0.689075
[Epoch 90] speed: 1507 samples/sec	time cost: 872.365952
[Epoch 90] validation: err-top1=0.287946 err-top5=0.095265
Epoch[91] Batch [49]	Speed: 1474.322299 samples/sec	accuracy=0.700820	lr=0.059396
Epoch[91] Batch [99]	Speed: 1536.202223 samples/sec	accuracy=0.700430	lr=0.059240
Epoch[91] Batch [149]	Speed: 1540.075301 samples/sec	accuracy=0.699342	lr=0.059085
Epoch[91] Batch [199]	Speed: 1528.768806 samples/sec	accuracy=0.699453	lr=0.058930
Epoch[91] Batch [249]	Speed: 1492.896064 samples/sec	accuracy=0.699727	lr=0.058776
Epoch[91] Batch [299]	Speed: 1432.395607 samples/sec	accuracy=0.698470	lr=0.058621
Epoch[91] Batch [349]	Speed: 1523.229338 samples/sec	accuracy=0.698089	lr=0.058467
Epoch[91] Batch [399]	Speed: 1535.589016 samples/sec	accuracy=0.697686	lr=0.058313
Epoch[91] Batch [449]	Speed: 1519.599453 samples/sec	accuracy=0.697072	lr=0.058158
Epoch[91] Batch [499]	Speed: 1535.381360 samples/sec	accuracy=0.696871	lr=0.058005
Epoch[91] Batch [549]	Speed: 1512.860275 samples/sec	accuracy=0.696674	lr=0.057851
Epoch[91] Batch [599]	Speed: 1541.899105 samples/sec	accuracy=0.696567	lr=0.057697
Epoch[91] Batch [649]	Speed: 1502.693781 samples/sec	accuracy=0.696384	lr=0.057544
Epoch[91] Batch [699]	Speed: 1473.868742 samples/sec	accuracy=0.695992	lr=0.057391
Epoch[91] Batch [749]	Speed: 1496.373260 samples/sec	accuracy=0.695604	lr=0.057238
Epoch[91] Batch [799]	Speed: 1516.753140 samples/sec	accuracy=0.695278	lr=0.057085
Epoch[91] Batch [849]	Speed: 1490.252493 samples/sec	accuracy=0.695046	lr=0.056932
Epoch[91] Batch [899]	Speed: 1537.800288 samples/sec	accuracy=0.694667	lr=0.056780
Epoch[91] Batch [949]	Speed: 1443.865603 samples/sec	accuracy=0.694502	lr=0.056627
Epoch[91] Batch [999]	Speed: 1519.098541 samples/sec	accuracy=0.694333	lr=0.056475
Epoch[91] Batch [1049]	Speed: 1516.815844 samples/sec	accuracy=0.694252	lr=0.056323
Epoch[91] Batch [1099]	Speed: 1507.259775 samples/sec	accuracy=0.694034	lr=0.056171
Epoch[91] Batch [1149]	Speed: 1485.395349 samples/sec	accuracy=0.693906	lr=0.056020
Epoch[91] Batch [1199]	Speed: 1470.826674 samples/sec	accuracy=0.693713	lr=0.055868
Epoch[91] Batch [1249]	Speed: 1541.065432 samples/sec	accuracy=0.693464	lr=0.055717
[Epoch 91] training: accuracy=0.693466
[Epoch 91] speed: 1505 samples/sec	time cost: 873.162043
[Epoch 91] validation: err-top1=0.279955 err-top5=0.090860
Epoch[92] Batch [49]	Speed: 1533.255752 samples/sec	accuracy=0.705820	lr=0.055563
Epoch[92] Batch [99]	Speed: 1530.836368 samples/sec	accuracy=0.704482	lr=0.055412
Epoch[92] Batch [149]	Speed: 1522.550875 samples/sec	accuracy=0.702923	lr=0.055261
Epoch[92] Batch [199]	Speed: 1540.070850 samples/sec	accuracy=0.702432	lr=0.055111
Epoch[92] Batch [249]	Speed: 1482.668907 samples/sec	accuracy=0.701832	lr=0.054960
Epoch[92] Batch [299]	Speed: 1515.364115 samples/sec	accuracy=0.701595	lr=0.054810
Epoch[92] Batch [349]	Speed: 1509.986921 samples/sec	accuracy=0.701518	lr=0.054660
Epoch[92] Batch [399]	Speed: 1475.496910 samples/sec	accuracy=0.701130	lr=0.054510
Epoch[92] Batch [449]	Speed: 1476.390739 samples/sec	accuracy=0.700723	lr=0.054360
Epoch[92] Batch [499]	Speed: 1460.820432 samples/sec	accuracy=0.700393	lr=0.054210
Epoch[92] Batch [549]	Speed: 1534.395349 samples/sec	accuracy=0.700000	lr=0.054061
Epoch[92] Batch [599]	Speed: 1453.584916 samples/sec	accuracy=0.699917	lr=0.053912
Epoch[92] Batch [649]	Speed: 1528.022990 samples/sec	accuracy=0.699659	lr=0.053763
Epoch[92] Batch [699]	Speed: 1528.805570 samples/sec	accuracy=0.699689	lr=0.053614
Epoch[92] Batch [749]	Speed: 1492.411364 samples/sec	accuracy=0.699362	lr=0.053465
Epoch[92] Batch [799]	Speed: 1539.549222 samples/sec	accuracy=0.698978	lr=0.053317
Epoch[92] Batch [849]	Speed: 1500.174380 samples/sec	accuracy=0.698901	lr=0.053168
Epoch[92] Batch [899]	Speed: 1533.448292 samples/sec	accuracy=0.698659	lr=0.053020
Epoch[92] Batch [949]	Speed: 1506.421144 samples/sec	accuracy=0.698554	lr=0.052872
Epoch[92] Batch [999]	Speed: 1544.460337 samples/sec	accuracy=0.698390	lr=0.052724
Epoch[92] Batch [1049]	Speed: 1487.783200 samples/sec	accuracy=0.698272	lr=0.052577
Epoch[92] Batch [1099]	Speed: 1504.644421 samples/sec	accuracy=0.698007	lr=0.052429
Epoch[92] Batch [1149]	Speed: 1490.292258 samples/sec	accuracy=0.697535	lr=0.052282
Epoch[92] Batch [1199]	Speed: 1492.690123 samples/sec	accuracy=0.697492	lr=0.052135
Epoch[92] Batch [1249]	Speed: 1519.248708 samples/sec	accuracy=0.697380	lr=0.051988
[Epoch 92] training: accuracy=0.697375
[Epoch 92] speed: 1506 samples/sec	time cost: 872.211187
[Epoch 92] validation: err-top1=0.278420 err-top5=0.092156
Epoch[93] Batch [49]	Speed: 1501.666543 samples/sec	accuracy=0.707344	lr=0.051838
Epoch[93] Batch [99]	Speed: 1540.886751 samples/sec	accuracy=0.707617	lr=0.051691
Epoch[93] Batch [149]	Speed: 1495.825253 samples/sec	accuracy=0.707695	lr=0.051545
Epoch[93] Batch [199]	Speed: 1511.955138 samples/sec	accuracy=0.707314	lr=0.051399
Epoch[93] Batch [249]	Speed: 1515.597240 samples/sec	accuracy=0.706559	lr=0.051253
Epoch[93] Batch [299]	Speed: 1475.470846 samples/sec	accuracy=0.705622	lr=0.051107
Epoch[93] Batch [349]	Speed: 1523.058161 samples/sec	accuracy=0.705405	lr=0.050961
Epoch[93] Batch [399]	Speed: 1536.450125 samples/sec	accuracy=0.705249	lr=0.050816
Epoch[93] Batch [449]	Speed: 1511.868961 samples/sec	accuracy=0.705111	lr=0.050670
Epoch[93] Batch [499]	Speed: 1471.454409 samples/sec	accuracy=0.704557	lr=0.050525
Epoch[93] Batch [549]	Speed: 1489.170695 samples/sec	accuracy=0.704389	lr=0.050380
Epoch[93] Batch [599]	Speed: 1492.861411 samples/sec	accuracy=0.704049	lr=0.050235
Epoch[93] Batch [649]	Speed: 1540.218785 samples/sec	accuracy=0.703837	lr=0.050091
Epoch[93] Batch [699]	Speed: 1523.986854 samples/sec	accuracy=0.703862	lr=0.049946
Epoch[93] Batch [749]	Speed: 1468.796638 samples/sec	accuracy=0.703776	lr=0.049802
Epoch[93] Batch [799]	Speed: 1502.189627 samples/sec	accuracy=0.703579	lr=0.049658
Epoch[93] Batch [849]	Speed: 1542.872173 samples/sec	accuracy=0.703421	lr=0.049514
Epoch[93] Batch [899]	Speed: 1501.424080 samples/sec	accuracy=0.703180	lr=0.049370
Epoch[93] Batch [949]	Speed: 1492.015200 samples/sec	accuracy=0.703093	lr=0.049227
Epoch[93] Batch [999]	Speed: 1521.950695 samples/sec	accuracy=0.703056	lr=0.049083
Epoch[93] Batch [1049]	Speed: 1517.095799 samples/sec	accuracy=0.702803	lr=0.048940
Epoch[93] Batch [1099]	Speed: 1490.221293 samples/sec	accuracy=0.702716	lr=0.048797
Epoch[93] Batch [1149]	Speed: 1516.422810 samples/sec	accuracy=0.702527	lr=0.048654
Epoch[93] Batch [1199]	Speed: 1543.623804 samples/sec	accuracy=0.702403	lr=0.048511
Epoch[93] Batch [1249]	Speed: 1520.880620 samples/sec	accuracy=0.702349	lr=0.048369
[Epoch 93] training: accuracy=0.702322
[Epoch 93] speed: 1508 samples/sec	time cost: 871.186039
[Epoch 93] validation: err-top1=0.278463 err-top5=0.090047
Epoch[94] Batch [49]	Speed: 1467.193019 samples/sec	accuracy=0.711055	lr=0.048224
Epoch[94] Batch [99]	Speed: 1532.947817 samples/sec	accuracy=0.711670	lr=0.048082
Epoch[94] Batch [149]	Speed: 1541.464651 samples/sec	accuracy=0.710410	lr=0.047940
Epoch[94] Batch [199]	Speed: 1533.589558 samples/sec	accuracy=0.710103	lr=0.047798
Epoch[94] Batch [249]	Speed: 1515.488957 samples/sec	accuracy=0.709719	lr=0.047656
Epoch[94] Batch [299]	Speed: 1516.989896 samples/sec	accuracy=0.709508	lr=0.047515
Epoch[94] Batch [349]	Speed: 1514.679542 samples/sec	accuracy=0.708655	lr=0.047374
Epoch[94] Batch [399]	Speed: 1511.695475 samples/sec	accuracy=0.708713	lr=0.047233
Epoch[94] Batch [449]	Speed: 1513.268739 samples/sec	accuracy=0.708529	lr=0.047092
Epoch[94] Batch [499]	Speed: 1509.464494 samples/sec	accuracy=0.707936	lr=0.046951
Epoch[94] Batch [549]	Speed: 1436.394988 samples/sec	accuracy=0.707713	lr=0.046811
Epoch[94] Batch [599]	Speed: 1498.722948 samples/sec	accuracy=0.707752	lr=0.046670
Epoch[94] Batch [649]	Speed: 1482.391944 samples/sec	accuracy=0.707550	lr=0.046530
Epoch[94] Batch [699]	Speed: 1519.682202 samples/sec	accuracy=0.707274	lr=0.046390
Epoch[94] Batch [749]	Speed: 1547.009409 samples/sec	accuracy=0.707277	lr=0.046251
Epoch[94] Batch [799]	Speed: 1511.380524 samples/sec	accuracy=0.707260	lr=0.046111
Epoch[94] Batch [849]	Speed: 1434.336195 samples/sec	accuracy=0.707089	lr=0.045972
Epoch[94] Batch [899]	Speed: 1515.816762 samples/sec	accuracy=0.706721	lr=0.045833
Epoch[94] Batch [949]	Speed: 1505.582242 samples/sec	accuracy=0.706623	lr=0.045693
Epoch[94] Batch [999]	Speed: 1495.758281 samples/sec	accuracy=0.706288	lr=0.045555
Epoch[94] Batch [1049]	Speed: 1470.417369 samples/sec	accuracy=0.706270	lr=0.045416
Epoch[94] Batch [1099]	Speed: 1509.362539 samples/sec	accuracy=0.706199	lr=0.045278
Epoch[94] Batch [1149]	Speed: 1493.113439 samples/sec	accuracy=0.706063	lr=0.045139
Epoch[94] Batch [1199]	Speed: 1491.993453 samples/sec	accuracy=0.706080	lr=0.045001
Epoch[94] Batch [1249]	Speed: 1549.281436 samples/sec	accuracy=0.705975	lr=0.044863
[Epoch 94] training: accuracy=0.705989
[Epoch 94] speed: 1502 samples/sec	time cost: 873.935442
[Epoch 94] validation: err-top1=0.271943 err-top5=0.087093
Epoch[95] Batch [49]	Speed: 1461.182494 samples/sec	accuracy=0.715527	lr=0.044723
Epoch[95] Batch [99]	Speed: 1494.099762 samples/sec	accuracy=0.714092	lr=0.044585
Epoch[95] Batch [149]	Speed: 1510.394865 samples/sec	accuracy=0.713477	lr=0.044448
Epoch[95] Batch [199]	Speed: 1510.563249 samples/sec	accuracy=0.714219	lr=0.044311
Epoch[95] Batch [249]	Speed: 1519.178720 samples/sec	accuracy=0.715187	lr=0.044174
Epoch[95] Batch [299]	Speed: 1463.846571 samples/sec	accuracy=0.714733	lr=0.044037
Epoch[95] Batch [349]	Speed: 1542.802774 samples/sec	accuracy=0.713786	lr=0.043900
Epoch[95] Batch [399]	Speed: 1514.797924 samples/sec	accuracy=0.713804	lr=0.043764
Epoch[95] Batch [449]	Speed: 1490.786262 samples/sec	accuracy=0.713709	lr=0.043628
Epoch[95] Batch [499]	Speed: 1474.003007 samples/sec	accuracy=0.713629	lr=0.043492
Epoch[95] Batch [549]	Speed: 1543.019827 samples/sec	accuracy=0.713020	lr=0.043356
Epoch[95] Batch [599]	Speed: 1491.255947 samples/sec	accuracy=0.712617	lr=0.043220
Epoch[95] Batch [649]	Speed: 1544.347403 samples/sec	accuracy=0.712371	lr=0.043085
Epoch[95] Batch [699]	Speed: 1537.395380 samples/sec	accuracy=0.712418	lr=0.042949
Epoch[95] Batch [749]	Speed: 1514.177148 samples/sec	accuracy=0.712319	lr=0.042814
Epoch[95] Batch [799]	Speed: 1492.691493 samples/sec	accuracy=0.712042	lr=0.042679
Epoch[95] Batch [849]	Speed: 1501.858614 samples/sec	accuracy=0.712005	lr=0.042545
Epoch[95] Batch [899]	Speed: 1512.044307 samples/sec	accuracy=0.711863	lr=0.042410
Epoch[95] Batch [949]	Speed: 1515.216982 samples/sec	accuracy=0.711604	lr=0.042276
Epoch[95] Batch [999]	Speed: 1467.988280 samples/sec	accuracy=0.711344	lr=0.042141
Epoch[95] Batch [1049]	Speed: 1494.066156 samples/sec	accuracy=0.711080	lr=0.042007
Epoch[95] Batch [1099]	Speed: 1527.664205 samples/sec	accuracy=0.710929	lr=0.041874
Epoch[95] Batch [1149]	Speed: 1487.597113 samples/sec	accuracy=0.710908	lr=0.041740
Epoch[95] Batch [1199]	Speed: 1486.034709 samples/sec	accuracy=0.710936	lr=0.041607
Epoch[95] Batch [1249]	Speed: 1545.862322 samples/sec	accuracy=0.710823	lr=0.041473
[Epoch 95] training: accuracy=0.710821
[Epoch 95] speed: 1504 samples/sec	time cost: 873.852981
[Epoch 95] validation: err-top1=0.269332 err-top5=0.087014
Epoch[96] Batch [49]	Speed: 1439.753341 samples/sec	accuracy=0.720625	lr=0.041338
Epoch[96] Batch [99]	Speed: 1505.104841 samples/sec	accuracy=0.719834	lr=0.041205
Epoch[96] Batch [149]	Speed: 1491.001595 samples/sec	accuracy=0.718392	lr=0.041072
Epoch[96] Batch [199]	Speed: 1513.353114 samples/sec	accuracy=0.718496	lr=0.040940
Epoch[96] Batch [249]	Speed: 1494.721743 samples/sec	accuracy=0.718262	lr=0.040807
Epoch[96] Batch [299]	Speed: 1510.230278 samples/sec	accuracy=0.718275	lr=0.040675
Epoch[96] Batch [349]	Speed: 1544.504403 samples/sec	accuracy=0.718136	lr=0.040543
Epoch[96] Batch [399]	Speed: 1522.715017 samples/sec	accuracy=0.718201	lr=0.040412
Epoch[96] Batch [449]	Speed: 1490.941172 samples/sec	accuracy=0.717962	lr=0.040280
Epoch[96] Batch [499]	Speed: 1477.443259 samples/sec	accuracy=0.717896	lr=0.040149
Epoch[96] Batch [549]	Speed: 1493.706525 samples/sec	accuracy=0.717386	lr=0.040018
Epoch[96] Batch [599]	Speed: 1508.876525 samples/sec	accuracy=0.717225	lr=0.039887
Epoch[96] Batch [649]	Speed: 1487.320047 samples/sec	accuracy=0.716831	lr=0.039756
Epoch[96] Batch [699]	Speed: 1538.535223 samples/sec	accuracy=0.716780	lr=0.039625
Epoch[96] Batch [749]	Speed: 1491.573527 samples/sec	accuracy=0.716518	lr=0.039495
Epoch[96] Batch [799]	Speed: 1443.504561 samples/sec	accuracy=0.716263	lr=0.039365
Epoch[96] Batch [849]	Speed: 1537.855581 samples/sec	accuracy=0.716066	lr=0.039235
Epoch[96] Batch [899]	Speed: 1457.891871 samples/sec	accuracy=0.716173	lr=0.039105
Epoch[96] Batch [949]	Speed: 1472.402563 samples/sec	accuracy=0.715875	lr=0.038975
Epoch[96] Batch [999]	Speed: 1515.851226 samples/sec	accuracy=0.715891	lr=0.038846
Epoch[96] Batch [1049]	Speed: 1540.656117 samples/sec	accuracy=0.715832	lr=0.038717
Epoch[96] Batch [1099]	Speed: 1456.981602 samples/sec	accuracy=0.715711	lr=0.038588
Epoch[96] Batch [1149]	Speed: 1492.194950 samples/sec	accuracy=0.715527	lr=0.038459
Epoch[96] Batch [1199]	Speed: 1523.431202 samples/sec	accuracy=0.715469	lr=0.038330
Epoch[96] Batch [1249]	Speed: 1547.110562 samples/sec	accuracy=0.715348	lr=0.038202
[Epoch 96] training: accuracy=0.715350
[Epoch 96] speed: 1498 samples/sec	time cost: 876.986868
[Epoch 96] validation: err-top1=0.268555 err-top5=0.085878
Epoch[97] Batch [49]	Speed: 1514.744254 samples/sec	accuracy=0.727598	lr=0.038071
Epoch[97] Batch [99]	Speed: 1541.443042 samples/sec	accuracy=0.726221	lr=0.037943
Epoch[97] Batch [149]	Speed: 1528.447125 samples/sec	accuracy=0.725957	lr=0.037815
Epoch[97] Batch [199]	Speed: 1513.060988 samples/sec	accuracy=0.724927	lr=0.037687
Epoch[97] Batch [249]	Speed: 1469.666183 samples/sec	accuracy=0.724684	lr=0.037560
Epoch[97] Batch [299]	Speed: 1539.851335 samples/sec	accuracy=0.723949	lr=0.037432
Epoch[97] Batch [349]	Speed: 1514.052255 samples/sec	accuracy=0.722955	lr=0.037305
Epoch[97] Batch [399]	Speed: 1534.486570 samples/sec	accuracy=0.722849	lr=0.037178
Epoch[97] Batch [449]	Speed: 1464.963508 samples/sec	accuracy=0.722652	lr=0.037052
Epoch[97] Batch [499]	Speed: 1531.324327 samples/sec	accuracy=0.722184	lr=0.036925
Epoch[97] Batch [549]	Speed: 1492.368500 samples/sec	accuracy=0.722085	lr=0.036799
Epoch[97] Batch [599]	Speed: 1498.607672 samples/sec	accuracy=0.721859	lr=0.036673
Epoch[97] Batch [649]	Speed: 1544.916120 samples/sec	accuracy=0.721373	lr=0.036547
Epoch[97] Batch [699]	Speed: 1487.508795 samples/sec	accuracy=0.721217	lr=0.036421
Epoch[97] Batch [749]	Speed: 1500.919645 samples/sec	accuracy=0.721176	lr=0.036296
Epoch[97] Batch [799]	Speed: 1481.333519 samples/sec	accuracy=0.720908	lr=0.036170
Epoch[97] Batch [849]	Speed: 1480.953529 samples/sec	accuracy=0.720790	lr=0.036045
Epoch[97] Batch [899]	Speed: 1471.851732 samples/sec	accuracy=0.720507	lr=0.035920
Epoch[97] Batch [949]	Speed: 1533.224214 samples/sec	accuracy=0.720270	lr=0.035795
Epoch[97] Batch [999]	Speed: 1540.536257 samples/sec	accuracy=0.720374	lr=0.035671
Epoch[97] Batch [1049]	Speed: 1432.170898 samples/sec	accuracy=0.720374	lr=0.035546
Epoch[97] Batch [1099]	Speed: 1541.327915 samples/sec	accuracy=0.720135	lr=0.035422
Epoch[97] Batch [1149]	Speed: 1514.054636 samples/sec	accuracy=0.720099	lr=0.035298
Epoch[97] Batch [1199]	Speed: 1504.035993 samples/sec	accuracy=0.719896	lr=0.035174
Epoch[97] Batch [1249]	Speed: 1483.191580 samples/sec	accuracy=0.719838	lr=0.035051
[Epoch 97] training: accuracy=0.719836
[Epoch 97] speed: 1504 samples/sec	time cost: 873.547931
[Epoch 97] validation: err-top1=0.266422 err-top5=0.082749
Epoch[98] Batch [49]	Speed: 1486.175520 samples/sec	accuracy=0.730215	lr=0.034925
Epoch[98] Batch [99]	Speed: 1491.212703 samples/sec	accuracy=0.730801	lr=0.034802
Epoch[98] Batch [149]	Speed: 1483.565300 samples/sec	accuracy=0.730462	lr=0.034679
Epoch[98] Batch [199]	Speed: 1493.627267 samples/sec	accuracy=0.728779	lr=0.034556
Epoch[98] Batch [249]	Speed: 1456.513637 samples/sec	accuracy=0.728500	lr=0.034433
Epoch[98] Batch [299]	Speed: 1541.677199 samples/sec	accuracy=0.727904	lr=0.034311
Epoch[98] Batch [349]	Speed: 1537.055493 samples/sec	accuracy=0.727313	lr=0.034189
Epoch[98] Batch [399]	Speed: 1479.719634 samples/sec	accuracy=0.727202	lr=0.034067
Epoch[98] Batch [449]	Speed: 1516.649555 samples/sec	accuracy=0.727042	lr=0.033945
Epoch[98] Batch [499]	Speed: 1479.659113 samples/sec	accuracy=0.726506	lr=0.033823
Epoch[98] Batch [549]	Speed: 1536.094975 samples/sec	accuracy=0.726119	lr=0.033702
Epoch[98] Batch [599]	Speed: 1484.864684 samples/sec	accuracy=0.726076	lr=0.033581
Epoch[98] Batch [649]	Speed: 1540.557940 samples/sec	accuracy=0.725849	lr=0.033460
Epoch[98] Batch [699]	Speed: 1520.638782 samples/sec	accuracy=0.725752	lr=0.033339
Epoch[98] Batch [749]	Speed: 1545.956469 samples/sec	accuracy=0.725566	lr=0.033218
Epoch[98] Batch [799]	Speed: 1486.957870 samples/sec	accuracy=0.725345	lr=0.033098
Epoch[98] Batch [849]	Speed: 1516.738388 samples/sec	accuracy=0.725269	lr=0.032978
Epoch[98] Batch [899]	Speed: 1518.307959 samples/sec	accuracy=0.725000	lr=0.032858
Epoch[98] Batch [949]	Speed: 1452.957517 samples/sec	accuracy=0.724875	lr=0.032738
Epoch[98] Batch [999]	Speed: 1509.750966 samples/sec	accuracy=0.724875	lr=0.032618
Epoch[98] Batch [1049]	Speed: 1503.117057 samples/sec	accuracy=0.724991	lr=0.032499
Epoch[98] Batch [1099]	Speed: 1510.039712 samples/sec	accuracy=0.725011	lr=0.032380
Epoch[98] Batch [1149]	Speed: 1518.298180 samples/sec	accuracy=0.725065	lr=0.032261
Epoch[98] Batch [1199]	Speed: 1471.602040 samples/sec	accuracy=0.725019	lr=0.032142
Epoch[98] Batch [1249]	Speed: 1526.440093 samples/sec	accuracy=0.724884	lr=0.032023
[Epoch 98] training: accuracy=0.724895
[Epoch 98] speed: 1502 samples/sec	time cost: 873.936895
[Epoch 98] validation: err-top1=0.263570 err-top5=0.081014
Epoch[99] Batch [49]	Speed: 1499.963744 samples/sec	accuracy=0.735176	lr=0.031902
Epoch[99] Batch [99]	Speed: 1509.038792 samples/sec	accuracy=0.734893	lr=0.031784
Epoch[99] Batch [149]	Speed: 1477.272238 samples/sec	accuracy=0.734440	lr=0.031666
Epoch[99] Batch [199]	Speed: 1487.728006 samples/sec	accuracy=0.734058	lr=0.031548
Epoch[99] Batch [249]	Speed: 1521.261560 samples/sec	accuracy=0.733977	lr=0.031431
Epoch[99] Batch [299]	Speed: 1536.057442 samples/sec	accuracy=0.732923	lr=0.031313
Epoch[99] Batch [349]	Speed: 1491.755419 samples/sec	accuracy=0.732645	lr=0.031196
Epoch[99] Batch [399]	Speed: 1501.202494 samples/sec	accuracy=0.732444	lr=0.031079
Epoch[99] Batch [449]	Speed: 1525.342527 samples/sec	accuracy=0.732268	lr=0.030962
Epoch[99] Batch [499]	Speed: 1478.578982 samples/sec	accuracy=0.731977	lr=0.030846
Epoch[99] Batch [549]	Speed: 1496.881144 samples/sec	accuracy=0.731712	lr=0.030729
Epoch[99] Batch [599]	Speed: 1512.811539 samples/sec	accuracy=0.731847	lr=0.030613
Epoch[99] Batch [649]	Speed: 1545.890053 samples/sec	accuracy=0.731349	lr=0.030497
Epoch[99] Batch [699]	Speed: 1467.644793 samples/sec	accuracy=0.731198	lr=0.030381
Epoch[99] Batch [749]	Speed: 1546.024316 samples/sec	accuracy=0.731242	lr=0.030266
Epoch[99] Batch [799]	Speed: 1526.546735 samples/sec	accuracy=0.730895	lr=0.030150
Epoch[99] Batch [849]	Speed: 1538.495200 samples/sec	accuracy=0.730874	lr=0.030035
Epoch[99] Batch [899]	Speed: 1503.986539 samples/sec	accuracy=0.730450	lr=0.029920
Epoch[99] Batch [949]	Speed: 1524.341739 samples/sec	accuracy=0.730359	lr=0.029805
Epoch[99] Batch [999]	Speed: 1486.620506 samples/sec	accuracy=0.730286	lr=0.029691
Epoch[99] Batch [1049]	Speed: 1469.897943 samples/sec	accuracy=0.730350	lr=0.029576
Epoch[99] Batch [1099]	Speed: 1487.028349 samples/sec	accuracy=0.730499	lr=0.029462
Epoch[99] Batch [1149]	Speed: 1499.361200 samples/sec	accuracy=0.730444	lr=0.029348
Epoch[99] Batch [1199]	Speed: 1509.937254 samples/sec	accuracy=0.730284	lr=0.029234
Epoch[99] Batch [1249]	Speed: 1525.596604 samples/sec	accuracy=0.730009	lr=0.029121
[Epoch 99] training: accuracy=0.729996
[Epoch 99] speed: 1505 samples/sec	time cost: 873.123253
[Epoch 99] validation: err-top1=0.262556 err-top5=0.080616
Epoch[100] Batch [49]	Speed: 1467.218892 samples/sec	accuracy=0.739336	lr=0.029005
Epoch[100] Batch [99]	Speed: 1457.394123 samples/sec	accuracy=0.740361	lr=0.028892
Epoch[100] Batch [149]	Speed: 1511.390905 samples/sec	accuracy=0.740859	lr=0.028779
Epoch[100] Batch [199]	Speed: 1493.476181 samples/sec	accuracy=0.738750	lr=0.028666
Epoch[100] Batch [249]	Speed: 1495.273294 samples/sec	accuracy=0.738078	lr=0.028554
Epoch[100] Batch [299]	Speed: 1501.662595 samples/sec	accuracy=0.737441	lr=0.028441
Epoch[100] Batch [349]	Speed: 1510.227856 samples/sec	accuracy=0.737360	lr=0.028329
Epoch[100] Batch [399]	Speed: 1458.969711 samples/sec	accuracy=0.737341	lr=0.028217
Epoch[100] Batch [449]	Speed: 1530.762875 samples/sec	accuracy=0.736994	lr=0.028105
Epoch[100] Batch [499]	Speed: 1526.139705 samples/sec	accuracy=0.736979	lr=0.027994
Epoch[100] Batch [549]	Speed: 1501.146499 samples/sec	accuracy=0.736754	lr=0.027883
Epoch[100] Batch [599]	Speed: 1540.908776 samples/sec	accuracy=0.736405	lr=0.027772
Epoch[100] Batch [649]	Speed: 1510.093771 samples/sec	accuracy=0.736426	lr=0.027661
Epoch[100] Batch [699]	Speed: 1515.777496 samples/sec	accuracy=0.736377	lr=0.027550
Epoch[100] Batch [749]	Speed: 1508.354019 samples/sec	accuracy=0.736172	lr=0.027439
Epoch[100] Batch [799]	Speed: 1488.410990 samples/sec	accuracy=0.735856	lr=0.027329
Epoch[100] Batch [849]	Speed: 1468.726008 samples/sec	accuracy=0.735634	lr=0.027219
Epoch[100] Batch [899]	Speed: 1511.249317 samples/sec	accuracy=0.735710	lr=0.027109
Epoch[100] Batch [949]	Speed: 1521.018093 samples/sec	accuracy=0.735508	lr=0.026999
Epoch[100] Batch [999]	Speed: 1484.641594 samples/sec	accuracy=0.735478	lr=0.026890
Epoch[100] Batch [1049]	Speed: 1532.983458 samples/sec	accuracy=0.735307	lr=0.026781
Epoch[100] Batch [1099]	Speed: 1545.340823 samples/sec	accuracy=0.735204	lr=0.026672
Epoch[100] Batch [1149]	Speed: 1527.236063 samples/sec	accuracy=0.735210	lr=0.026563
Epoch[100] Batch [1199]	Speed: 1494.926663 samples/sec	accuracy=0.735035	lr=0.026454
Epoch[100] Batch [1249]	Speed: 1486.288655 samples/sec	accuracy=0.735014	lr=0.026346
[Epoch 100] training: accuracy=0.735028
[Epoch 100] speed: 1502 samples/sec	time cost: 875.272974
[Epoch 100] validation: err-top1=0.258351 err-top5=0.079998
Epoch[101] Batch [49]	Speed: 1498.156537 samples/sec	accuracy=0.744648	lr=0.026235
Epoch[101] Batch [99]	Speed: 1540.517470 samples/sec	accuracy=0.742568	lr=0.026127
Epoch[101] Batch [149]	Speed: 1523.976385 samples/sec	accuracy=0.742168	lr=0.026020
Epoch[101] Batch [199]	Speed: 1534.883781 samples/sec	accuracy=0.742104	lr=0.025912
Epoch[101] Batch [249]	Speed: 1541.667614 samples/sec	accuracy=0.742035	lr=0.025805
Epoch[101] Batch [299]	Speed: 1522.878470 samples/sec	accuracy=0.741634	lr=0.025697
Epoch[101] Batch [349]	Speed: 1544.307577 samples/sec	accuracy=0.741688	lr=0.025590
Epoch[101] Batch [399]	Speed: 1467.012498 samples/sec	accuracy=0.741331	lr=0.025484
Epoch[101] Batch [449]	Speed: 1532.588816 samples/sec	accuracy=0.741007	lr=0.025377
Epoch[101] Batch [499]	Speed: 1539.878432 samples/sec	accuracy=0.741084	lr=0.025271
Epoch[101] Batch [549]	Speed: 1525.367219 samples/sec	accuracy=0.741184	lr=0.025165
Epoch[101] Batch [599]	Speed: 1448.318906 samples/sec	accuracy=0.741183	lr=0.025059
Epoch[101] Batch [649]	Speed: 1505.184405 samples/sec	accuracy=0.741118	lr=0.024953
Epoch[101] Batch [699]	Speed: 1506.570760 samples/sec	accuracy=0.741042	lr=0.024847
Epoch[101] Batch [749]	Speed: 1454.092697 samples/sec	accuracy=0.741038	lr=0.024742
Epoch[101] Batch [799]	Speed: 1534.549784 samples/sec	accuracy=0.740896	lr=0.024637
Epoch[101] Batch [849]	Speed: 1521.383797 samples/sec	accuracy=0.740909	lr=0.024532
Epoch[101] Batch [899]	Speed: 1473.322151 samples/sec	accuracy=0.740573	lr=0.024427
Epoch[101] Batch [949]	Speed: 1512.145975 samples/sec	accuracy=0.740435	lr=0.024323
Epoch[101] Batch [999]	Speed: 1493.444887 samples/sec	accuracy=0.740677	lr=0.024219
Epoch[101] Batch [1049]	Speed: 1461.357834 samples/sec	accuracy=0.740589	lr=0.024115
Epoch[101] Batch [1099]	Speed: 1491.783222 samples/sec	accuracy=0.740611	lr=0.024011
Epoch[101] Batch [1149]	Speed: 1537.709322 samples/sec	accuracy=0.740532	lr=0.023907
Epoch[101] Batch [1199]	Speed: 1538.041777 samples/sec	accuracy=0.740468	lr=0.023804
Epoch[101] Batch [1249]	Speed: 1514.953932 samples/sec	accuracy=0.740340	lr=0.023700
[Epoch 101] training: accuracy=0.740341
[Epoch 101] speed: 1508 samples/sec	time cost: 870.910479
[Epoch 101] validation: err-top1=0.253627 err-top5=0.078324
Epoch[102] Batch [49]	Speed: 1461.527418 samples/sec	accuracy=0.751563	lr=0.023595
Epoch[102] Batch [99]	Speed: 1532.722156 samples/sec	accuracy=0.750244	lr=0.023493
Epoch[102] Batch [149]	Speed: 1521.445882 samples/sec	accuracy=0.749740	lr=0.023390
Epoch[102] Batch [199]	Speed: 1476.394586 samples/sec	accuracy=0.748794	lr=0.023288
Epoch[102] Batch [249]	Speed: 1531.248145 samples/sec	accuracy=0.748551	lr=0.023185
Epoch[102] Batch [299]	Speed: 1539.143625 samples/sec	accuracy=0.748193	lr=0.023083
Epoch[102] Batch [349]	Speed: 1532.469583 samples/sec	accuracy=0.747715	lr=0.022982
Epoch[102] Batch [399]	Speed: 1544.157812 samples/sec	accuracy=0.747900	lr=0.022880
Epoch[102] Batch [449]	Speed: 1492.038950 samples/sec	accuracy=0.747715	lr=0.022779
Epoch[102] Batch [499]	Speed: 1508.306981 samples/sec	accuracy=0.747500	lr=0.022678
Epoch[102] Batch [549]	Speed: 1542.642973 samples/sec	accuracy=0.747195	lr=0.022577
Epoch[102] Batch [599]	Speed: 1517.268329 samples/sec	accuracy=0.746903	lr=0.022476
Epoch[102] Batch [649]	Speed: 1510.032863 samples/sec	accuracy=0.746714	lr=0.022376
Epoch[102] Batch [699]	Speed: 1492.760410 samples/sec	accuracy=0.746496	lr=0.022276
Epoch[102] Batch [749]	Speed: 1447.151741 samples/sec	accuracy=0.746467	lr=0.022175
Epoch[102] Batch [799]	Speed: 1477.679706 samples/sec	accuracy=0.746465	lr=0.022076
Epoch[102] Batch [849]	Speed: 1518.878497 samples/sec	accuracy=0.746286	lr=0.021976
Epoch[102] Batch [899]	Speed: 1544.755103 samples/sec	accuracy=0.746399	lr=0.021877
Epoch[102] Batch [949]	Speed: 1460.039240 samples/sec	accuracy=0.746421	lr=0.021777
Epoch[102] Batch [999]	Speed: 1449.561141 samples/sec	accuracy=0.746277	lr=0.021678
Epoch[102] Batch [1049]	Speed: 1487.048799 samples/sec	accuracy=0.746101	lr=0.021580
Epoch[102] Batch [1099]	Speed: 1489.129833 samples/sec	accuracy=0.746078	lr=0.021481
Epoch[102] Batch [1149]	Speed: 1486.263134 samples/sec	accuracy=0.745969	lr=0.021383
Epoch[102] Batch [1199]	Speed: 1512.779067 samples/sec	accuracy=0.745938	lr=0.021285
Epoch[102] Batch [1249]	Speed: 1505.086255 samples/sec	accuracy=0.745785	lr=0.021187
[Epoch 102] training: accuracy=0.745789
[Epoch 102] speed: 1501 samples/sec	time cost: 875.184857
[Epoch 102] validation: err-top1=0.253249 err-top5=0.075833
Epoch[103] Batch [49]	Speed: 1504.555143 samples/sec	accuracy=0.752188	lr=0.021087
Epoch[103] Batch [99]	Speed: 1505.388130 samples/sec	accuracy=0.754736	lr=0.020990
Epoch[103] Batch [149]	Speed: 1541.265336 samples/sec	accuracy=0.753717	lr=0.020892
Epoch[103] Batch [199]	Speed: 1505.026383 samples/sec	accuracy=0.753521	lr=0.020795
Epoch[103] Batch [249]	Speed: 1470.265073 samples/sec	accuracy=0.753629	lr=0.020698
Epoch[103] Batch [299]	Speed: 1513.859241 samples/sec	accuracy=0.753102	lr=0.020602
Epoch[103] Batch [349]	Speed: 1541.962344 samples/sec	accuracy=0.752754	lr=0.020505
Epoch[103] Batch [399]	Speed: 1505.967997 samples/sec	accuracy=0.752634	lr=0.020409
Epoch[103] Batch [449]	Speed: 1548.260028 samples/sec	accuracy=0.752344	lr=0.020313
Epoch[103] Batch [499]	Speed: 1484.407470 samples/sec	accuracy=0.752201	lr=0.020217
Epoch[103] Batch [549]	Speed: 1496.405313 samples/sec	accuracy=0.751951	lr=0.020122
Epoch[103] Batch [599]	Speed: 1458.088281 samples/sec	accuracy=0.751675	lr=0.020026
Epoch[103] Batch [649]	Speed: 1534.268984 samples/sec	accuracy=0.751343	lr=0.019931
Epoch[103] Batch [699]	Speed: 1516.777158 samples/sec	accuracy=0.751256	lr=0.019836
Epoch[103] Batch [749]	Speed: 1513.225361 samples/sec	accuracy=0.751186	lr=0.019742
Epoch[103] Batch [799]	Speed: 1485.698884 samples/sec	accuracy=0.751010	lr=0.019647
Epoch[103] Batch [849]	Speed: 1509.223420 samples/sec	accuracy=0.750923	lr=0.019553
Epoch[103] Batch [899]	Speed: 1510.697886 samples/sec	accuracy=0.750737	lr=0.019459
Epoch[103] Batch [949]	Speed: 1536.949314 samples/sec	accuracy=0.750729	lr=0.019365
Epoch[103] Batch [999]	Speed: 1443.436653 samples/sec	accuracy=0.750530	lr=0.019271
Epoch[103] Batch [1049]	Speed: 1539.730639 samples/sec	accuracy=0.750343	lr=0.019178
Epoch[103] Batch [1099]	Speed: 1485.344770 samples/sec	accuracy=0.750186	lr=0.019085
Epoch[103] Batch [1149]	Speed: 1419.213952 samples/sec	accuracy=0.750239	lr=0.018992
Epoch[103] Batch [1199]	Speed: 1514.974580 samples/sec	accuracy=0.750177	lr=0.018899
Epoch[103] Batch [1249]	Speed: 1547.379391 samples/sec	accuracy=0.750168	lr=0.018806
[Epoch 103] training: accuracy=0.750166
[Epoch 103] speed: 1503 samples/sec	time cost: 874.427155
[Epoch 103] validation: err-top1=0.251873 err-top5=0.076451
Epoch[104] Batch [49]	Speed: 1420.286542 samples/sec	accuracy=0.755137	lr=0.018712
Epoch[104] Batch [99]	Speed: 1508.505756 samples/sec	accuracy=0.758096	lr=0.018620
Epoch[104] Batch [149]	Speed: 1528.473539 samples/sec	accuracy=0.757135	lr=0.018528
Epoch[104] Batch [199]	Speed: 1538.149924 samples/sec	accuracy=0.757246	lr=0.018436
Epoch[104] Batch [249]	Speed: 1492.064369 samples/sec	accuracy=0.756957	lr=0.018345
Epoch[104] Batch [299]	Speed: 1540.921303 samples/sec	accuracy=0.756784	lr=0.018254
Epoch[104] Batch [349]	Speed: 1483.153628 samples/sec	accuracy=0.756708	lr=0.018163
Epoch[104] Batch [399]	Speed: 1517.400025 samples/sec	accuracy=0.756912	lr=0.018072
Epoch[104] Batch [449]	Speed: 1512.849575 samples/sec	accuracy=0.756840	lr=0.017981
Epoch[104] Batch [499]	Speed: 1505.337247 samples/sec	accuracy=0.756693	lr=0.017891
Epoch[104] Batch [549]	Speed: 1542.862807 samples/sec	accuracy=0.756607	lr=0.017801
Epoch[104] Batch [599]	Speed: 1446.061371 samples/sec	accuracy=0.756619	lr=0.017711
Epoch[104] Batch [649]	Speed: 1503.238521 samples/sec	accuracy=0.756280	lr=0.017621
Epoch[104] Batch [699]	Speed: 1476.232363 samples/sec	accuracy=0.756098	lr=0.017532
Epoch[104] Batch [749]	Speed: 1482.015737 samples/sec	accuracy=0.755934	lr=0.017442
Epoch[104] Batch [799]	Speed: 1527.825234 samples/sec	accuracy=0.756014	lr=0.017353
Epoch[104] Batch [849]	Speed: 1549.718877 samples/sec	accuracy=0.755988	lr=0.017264
Epoch[104] Batch [899]	Speed: 1507.793920 samples/sec	accuracy=0.755974	lr=0.017176
Epoch[104] Batch [949]	Speed: 1538.629659 samples/sec	accuracy=0.755821	lr=0.017087
Epoch[104] Batch [999]	Speed: 1544.197985 samples/sec	accuracy=0.755841	lr=0.016999
Epoch[104] Batch [1049]	Speed: 1479.230582 samples/sec	accuracy=0.755718	lr=0.016911
Epoch[104] Batch [1099]	Speed: 1491.469634 samples/sec	accuracy=0.755667	lr=0.016823
Epoch[104] Batch [1149]	Speed: 1540.948139 samples/sec	accuracy=0.755709	lr=0.016736
Epoch[104] Batch [1199]	Speed: 1515.400612 samples/sec	accuracy=0.755591	lr=0.016648
Epoch[104] Batch [1249]	Speed: 1521.203789 samples/sec	accuracy=0.755568	lr=0.016561
[Epoch 104] training: accuracy=0.755574
[Epoch 104] speed: 1506 samples/sec	time cost: 871.958997
[Epoch 104] validation: err-top1=0.249613 err-top5=0.075948
Epoch[105] Batch [49]	Speed: 1501.703590 samples/sec	accuracy=0.767266	lr=0.016473
Epoch[105] Batch [99]	Speed: 1513.305786 samples/sec	accuracy=0.765801	lr=0.016386
Epoch[105] Batch [149]	Speed: 1483.340941 samples/sec	accuracy=0.764134	lr=0.016299
Epoch[105] Batch [199]	Speed: 1527.294781 samples/sec	accuracy=0.763457	lr=0.016213
Epoch[105] Batch [249]	Speed: 1458.067897 samples/sec	accuracy=0.763016	lr=0.016127
Epoch[105] Batch [299]	Speed: 1532.817544 samples/sec	accuracy=0.762425	lr=0.016041
Epoch[105] Batch [349]	Speed: 1469.791968 samples/sec	accuracy=0.762483	lr=0.015956
Epoch[105] Batch [399]	Speed: 1523.323958 samples/sec	accuracy=0.762112	lr=0.015870
Epoch[105] Batch [449]	Speed: 1545.028271 samples/sec	accuracy=0.762090	lr=0.015785
Epoch[105] Batch [499]	Speed: 1485.473705 samples/sec	accuracy=0.761898	lr=0.015700
Epoch[105] Batch [549]	Speed: 1495.422568 samples/sec	accuracy=0.761781	lr=0.015616
Epoch[105] Batch [599]	Speed: 1451.957627 samples/sec	accuracy=0.761789	lr=0.015531
Epoch[105] Batch [649]	Speed: 1469.108492 samples/sec	accuracy=0.761591	lr=0.015447
Epoch[105] Batch [699]	Speed: 1503.703693 samples/sec	accuracy=0.761660	lr=0.015363
Epoch[105] Batch [749]	Speed: 1479.977464 samples/sec	accuracy=0.761421	lr=0.015279
Epoch[105] Batch [799]	Speed: 1500.092652 samples/sec	accuracy=0.761466	lr=0.015195
Epoch[105] Batch [849]	Speed: 1513.048537 samples/sec	accuracy=0.761437	lr=0.015112
Epoch[105] Batch [899]	Speed: 1489.838560 samples/sec	accuracy=0.761427	lr=0.015029
Epoch[105] Batch [949]	Speed: 1514.015109 samples/sec	accuracy=0.761194	lr=0.014946
Epoch[105] Batch [999]	Speed: 1493.042465 samples/sec	accuracy=0.761222	lr=0.014863
Epoch[105] Batch [1049]	Speed: 1479.591308 samples/sec	accuracy=0.761076	lr=0.014781
Epoch[105] Batch [1099]	Speed: 1519.578077 samples/sec	accuracy=0.761035	lr=0.014699
Epoch[105] Batch [1149]	Speed: 1541.971169 samples/sec	accuracy=0.760912	lr=0.014616
Epoch[105] Batch [1199]	Speed: 1544.009430 samples/sec	accuracy=0.760868	lr=0.014535
Epoch[105] Batch [1249]	Speed: 1525.279181 samples/sec	accuracy=0.760835	lr=0.014453
[Epoch 105] training: accuracy=0.760841
[Epoch 105] speed: 1500 samples/sec	time cost: 875.859397
[Epoch 105] validation: err-top1=0.245356 err-top5=0.072804
Epoch[106] Batch [49]	Speed: 1510.368265 samples/sec	accuracy=0.768770	lr=0.014370
Epoch[106] Batch [99]	Speed: 1507.213006 samples/sec	accuracy=0.768984	lr=0.014289
Epoch[106] Batch [149]	Speed: 1489.929356 samples/sec	accuracy=0.768600	lr=0.014208
Epoch[106] Batch [199]	Speed: 1508.197577 samples/sec	accuracy=0.768877	lr=0.014127
Epoch[106] Batch [249]	Speed: 1541.496706 samples/sec	accuracy=0.768422	lr=0.014047
Epoch[106] Batch [299]	Speed: 1503.495202 samples/sec	accuracy=0.768356	lr=0.013966
Epoch[106] Batch [349]	Speed: 1546.967251 samples/sec	accuracy=0.767980	lr=0.013886
Epoch[106] Batch [399]	Speed: 1535.464706 samples/sec	accuracy=0.767959	lr=0.013806
Epoch[106] Batch [449]	Speed: 1507.734204 samples/sec	accuracy=0.767980	lr=0.013727
Epoch[106] Batch [499]	Speed: 1475.093857 samples/sec	accuracy=0.768104	lr=0.013647
Epoch[106] Batch [549]	Speed: 1441.174763 samples/sec	accuracy=0.767820	lr=0.013568
Epoch[106] Batch [599]	Speed: 1539.159224 samples/sec	accuracy=0.767660	lr=0.013489
Epoch[106] Batch [649]	Speed: 1511.792467 samples/sec	accuracy=0.767595	lr=0.013411
Epoch[106] Batch [699]	Speed: 1513.756648 samples/sec	accuracy=0.767210	lr=0.013332
Epoch[106] Batch [749]	Speed: 1498.632897 samples/sec	accuracy=0.767155	lr=0.013254
Epoch[106] Batch [799]	Speed: 1482.829015 samples/sec	accuracy=0.767007	lr=0.013176
Epoch[106] Batch [849]	Speed: 1515.505000 samples/sec	accuracy=0.767040	lr=0.013098
Epoch[106] Batch [899]	Speed: 1517.191770 samples/sec	accuracy=0.766878	lr=0.013020
Epoch[106] Batch [949]	Speed: 1515.628410 samples/sec	accuracy=0.766768	lr=0.012943
Epoch[106] Batch [999]	Speed: 1506.993738 samples/sec	accuracy=0.766649	lr=0.012866
Epoch[106] Batch [1049]	Speed: 1506.429820 samples/sec	accuracy=0.766643	lr=0.012789
Epoch[106] Batch [1099]	Speed: 1507.836458 samples/sec	accuracy=0.766565	lr=0.012712
Epoch[106] Batch [1149]	Speed: 1492.570928 samples/sec	accuracy=0.766553	lr=0.012636
Epoch[106] Batch [1199]	Speed: 1519.502253 samples/sec	accuracy=0.766573	lr=0.012559
Epoch[106] Batch [1249]	Speed: 1558.750832 samples/sec	accuracy=0.766608	lr=0.012483
[Epoch 106] training: accuracy=0.766605
[Epoch 106] speed: 1508 samples/sec	time cost: 871.103656
[Epoch 106] validation: err-top1=0.244320 err-top5=0.071887
Epoch[107] Batch [49]	Speed: 1465.464950 samples/sec	accuracy=0.773984	lr=0.012406
Epoch[107] Batch [99]	Speed: 1509.319352 samples/sec	accuracy=0.772861	lr=0.012330
Epoch[107] Batch [149]	Speed: 1481.811253 samples/sec	accuracy=0.772266	lr=0.012255
Epoch[107] Batch [199]	Speed: 1507.219301 samples/sec	accuracy=0.773218	lr=0.012180
Epoch[107] Batch [249]	Speed: 1540.353014 samples/sec	accuracy=0.773059	lr=0.012105
Epoch[107] Batch [299]	Speed: 1539.221665 samples/sec	accuracy=0.772933	lr=0.012030
Epoch[107] Batch [349]	Speed: 1517.775082 samples/sec	accuracy=0.772840	lr=0.011956
Epoch[107] Batch [399]	Speed: 1427.890380 samples/sec	accuracy=0.772620	lr=0.011881
Epoch[107] Batch [449]	Speed: 1530.524877 samples/sec	accuracy=0.772828	lr=0.011807
Epoch[107] Batch [499]	Speed: 1538.066584 samples/sec	accuracy=0.772811	lr=0.011734
Epoch[107] Batch [549]	Speed: 1530.352046 samples/sec	accuracy=0.772715	lr=0.011660
Epoch[107] Batch [599]	Speed: 1523.507505 samples/sec	accuracy=0.772498	lr=0.011587
Epoch[107] Batch [649]	Speed: 1513.638962 samples/sec	accuracy=0.772473	lr=0.011513
Epoch[107] Batch [699]	Speed: 1543.038220 samples/sec	accuracy=0.772492	lr=0.011441
Epoch[107] Batch [749]	Speed: 1510.692551 samples/sec	accuracy=0.772565	lr=0.011368
Epoch[107] Batch [799]	Speed: 1500.680243 samples/sec	accuracy=0.772417	lr=0.011295
Epoch[107] Batch [849]	Speed: 1537.576708 samples/sec	accuracy=0.772441	lr=0.011223
Epoch[107] Batch [899]	Speed: 1462.606578 samples/sec	accuracy=0.772397	lr=0.011151
Epoch[107] Batch [949]	Speed: 1516.108700 samples/sec	accuracy=0.772074	lr=0.011079
Epoch[107] Batch [999]	Speed: 1516.630617 samples/sec	accuracy=0.772047	lr=0.011008
Epoch[107] Batch [1049]	Speed: 1493.316029 samples/sec	accuracy=0.772003	lr=0.010936
Epoch[107] Batch [1099]	Speed: 1538.033570 samples/sec	accuracy=0.771923	lr=0.010865
Epoch[107] Batch [1149]	Speed: 1512.956564 samples/sec	accuracy=0.771793	lr=0.010794
Epoch[107] Batch [1199]	Speed: 1489.460444 samples/sec	accuracy=0.771784	lr=0.010724
Epoch[107] Batch [1249]	Speed: 1455.094624 samples/sec	accuracy=0.771695	lr=0.010653
[Epoch 107] training: accuracy=0.771716
[Epoch 107] speed: 1506 samples/sec	time cost: 872.886717
[Epoch 107] validation: err-top1=0.243104 err-top5=0.071747
Epoch[108] Batch [49]	Speed: 1501.156552 samples/sec	accuracy=0.780137	lr=0.010582
Epoch[108] Batch [99]	Speed: 1540.385266 samples/sec	accuracy=0.778906	lr=0.010512
Epoch[108] Batch [149]	Speed: 1481.557608 samples/sec	accuracy=0.778073	lr=0.010442
Epoch[108] Batch [199]	Speed: 1515.875044 samples/sec	accuracy=0.778237	lr=0.010373
Epoch[108] Batch [249]	Speed: 1538.878312 samples/sec	accuracy=0.777898	lr=0.010303
Epoch[108] Batch [299]	Speed: 1499.678262 samples/sec	accuracy=0.777731	lr=0.010234
Epoch[108] Batch [349]	Speed: 1485.584298 samples/sec	accuracy=0.777835	lr=0.010165
Epoch[108] Batch [399]	Speed: 1507.198133 samples/sec	accuracy=0.777178	lr=0.010097
Epoch[108] Batch [449]	Speed: 1538.538860 samples/sec	accuracy=0.777296	lr=0.010028
Epoch[108] Batch [499]	Speed: 1542.853518 samples/sec	accuracy=0.777113	lr=0.009960
Epoch[108] Batch [549]	Speed: 1544.633881 samples/sec	accuracy=0.776887	lr=0.009892
Epoch[108] Batch [599]	Speed: 1500.393199 samples/sec	accuracy=0.776873	lr=0.009824
Epoch[108] Batch [649]	Speed: 1535.673691 samples/sec	accuracy=0.776639	lr=0.009757
Epoch[108] Batch [699]	Speed: 1517.208932 samples/sec	accuracy=0.776684	lr=0.009690
Epoch[108] Batch [749]	Speed: 1507.304843 samples/sec	accuracy=0.776643	lr=0.009623
Epoch[108] Batch [799]	Speed: 1470.419030 samples/sec	accuracy=0.776469	lr=0.009556
Epoch[108] Batch [849]	Speed: 1536.537896 samples/sec	accuracy=0.776218	lr=0.009489
Epoch[108] Batch [899]	Speed: 1508.559515 samples/sec	accuracy=0.776314	lr=0.009423
Epoch[108] Batch [949]	Speed: 1514.185198 samples/sec	accuracy=0.776285	lr=0.009357
Epoch[108] Batch [999]	Speed: 1510.546365 samples/sec	accuracy=0.776322	lr=0.009291
Epoch[108] Batch [1049]	Speed: 1543.737743 samples/sec	accuracy=0.776409	lr=0.009225
Epoch[108] Batch [1099]	Speed: 1481.589489 samples/sec	accuracy=0.776441	lr=0.009160
Epoch[108] Batch [1149]	Speed: 1540.657123 samples/sec	accuracy=0.776467	lr=0.009095
Epoch[108] Batch [1199]	Speed: 1529.050992 samples/sec	accuracy=0.776597	lr=0.009030
Epoch[108] Batch [1249]	Speed: 1553.378370 samples/sec	accuracy=0.776638	lr=0.008965
[Epoch 108] training: accuracy=0.776658
[Epoch 108] speed: 1516 samples/sec	time cost: 866.421348
[Epoch 108] validation: err-top1=0.239118 err-top5=0.069456
Epoch[109] Batch [49]	Speed: 1494.542538 samples/sec	accuracy=0.786797	lr=0.008899
Epoch[109] Batch [99]	Speed: 1464.178828 samples/sec	accuracy=0.784268	lr=0.008835
Epoch[109] Batch [149]	Speed: 1498.276197 samples/sec	accuracy=0.783678	lr=0.008771
Epoch[109] Batch [199]	Speed: 1505.650561 samples/sec	accuracy=0.783184	lr=0.008707
Epoch[109] Batch [249]	Speed: 1464.061398 samples/sec	accuracy=0.783656	lr=0.008643
Epoch[109] Batch [299]	Speed: 1461.367709 samples/sec	accuracy=0.783343	lr=0.008580
Epoch[109] Batch [349]	Speed: 1531.962436 samples/sec	accuracy=0.782759	lr=0.008517
Epoch[109] Batch [399]	Speed: 1485.370670 samples/sec	accuracy=0.782397	lr=0.008454
Epoch[109] Batch [449]	Speed: 1520.317585 samples/sec	accuracy=0.782444	lr=0.008391
Epoch[109] Batch [499]	Speed: 1544.416119 samples/sec	accuracy=0.782025	lr=0.008329
Epoch[109] Batch [549]	Speed: 1525.619385 samples/sec	accuracy=0.781919	lr=0.008266
Epoch[109] Batch [599]	Speed: 1493.663264 samples/sec	accuracy=0.781872	lr=0.008204
Epoch[109] Batch [649]	Speed: 1502.498068 samples/sec	accuracy=0.781698	lr=0.008143
Epoch[109] Batch [699]	Speed: 1488.355553 samples/sec	accuracy=0.781536	lr=0.008081
Epoch[109] Batch [749]	Speed: 1543.281402 samples/sec	accuracy=0.781510	lr=0.008020
Epoch[109] Batch [799]	Speed: 1542.302777 samples/sec	accuracy=0.781722	lr=0.007959
Epoch[109] Batch [849]	Speed: 1496.598262 samples/sec	accuracy=0.781672	lr=0.007898
Epoch[109] Batch [899]	Speed: 1492.696379 samples/sec	accuracy=0.781581	lr=0.007837
Epoch[109] Batch [949]	Speed: 1520.701302 samples/sec	accuracy=0.781624	lr=0.007777
Epoch[109] Batch [999]	Speed: 1546.539771 samples/sec	accuracy=0.781566	lr=0.007716
Epoch[109] Batch [1049]	Speed: 1518.737297 samples/sec	accuracy=0.781652	lr=0.007656
Epoch[109] Batch [1099]	Speed: 1540.502275 samples/sec	accuracy=0.781603	lr=0.007597
Epoch[109] Batch [1149]	Speed: 1522.188246 samples/sec	accuracy=0.781485	lr=0.007537
Epoch[109] Batch [1199]	Speed: 1514.825919 samples/sec	accuracy=0.781497	lr=0.007478
Epoch[109] Batch [1249]	Speed: 1551.890219 samples/sec	accuracy=0.781392	lr=0.007419
[Epoch 109] training: accuracy=0.781412
[Epoch 109] speed: 1509 samples/sec	time cost: 870.938240
[Epoch 109] validation: err-top1=0.238122 err-top5=0.069097
Epoch[110] Batch [49]	Speed: 1467.570974 samples/sec	accuracy=0.787813	lr=0.007359
Epoch[110] Batch [99]	Speed: 1473.724519 samples/sec	accuracy=0.786689	lr=0.007300
Epoch[110] Batch [149]	Speed: 1520.535990 samples/sec	accuracy=0.786536	lr=0.007242
Epoch[110] Batch [199]	Speed: 1465.348014 samples/sec	accuracy=0.786987	lr=0.007184
Epoch[110] Batch [249]	Speed: 1537.291334 samples/sec	accuracy=0.786629	lr=0.007126
Epoch[110] Batch [299]	Speed: 1493.950659 samples/sec	accuracy=0.786188	lr=0.007068
Epoch[110] Batch [349]	Speed: 1497.176188 samples/sec	accuracy=0.786217	lr=0.007011
Epoch[110] Batch [399]	Speed: 1521.033111 samples/sec	accuracy=0.786240	lr=0.006954
Epoch[110] Batch [449]	Speed: 1513.552987 samples/sec	accuracy=0.786519	lr=0.006897
Epoch[110] Batch [499]	Speed: 1520.842232 samples/sec	accuracy=0.786541	lr=0.006840
Epoch[110] Batch [549]	Speed: 1542.843165 samples/sec	accuracy=0.786768	lr=0.006784
Epoch[110] Batch [599]	Speed: 1480.099062 samples/sec	accuracy=0.786776	lr=0.006727
Epoch[110] Batch [649]	Speed: 1521.723991 samples/sec	accuracy=0.786388	lr=0.006671
Epoch[110] Batch [699]	Speed: 1522.168469 samples/sec	accuracy=0.786480	lr=0.006615
Epoch[110] Batch [749]	Speed: 1539.501918 samples/sec	accuracy=0.786478	lr=0.006560
Epoch[110] Batch [799]	Speed: 1548.796154 samples/sec	accuracy=0.786547	lr=0.006504
Epoch[110] Batch [849]	Speed: 1503.915412 samples/sec	accuracy=0.786560	lr=0.006449
Epoch[110] Batch [899]	Speed: 1492.596002 samples/sec	accuracy=0.786568	lr=0.006394
Epoch[110] Batch [949]	Speed: 1513.946680 samples/sec	accuracy=0.786434	lr=0.006340
Epoch[110] Batch [999]	Speed: 1538.506167 samples/sec	accuracy=0.786325	lr=0.006285
Epoch[110] Batch [1049]	Speed: 1497.631014 samples/sec	accuracy=0.786363	lr=0.006231
Epoch[110] Batch [1099]	Speed: 1512.663739 samples/sec	accuracy=0.786416	lr=0.006177
Epoch[110] Batch [1149]	Speed: 1514.533353 samples/sec	accuracy=0.786365	lr=0.006123
Epoch[110] Batch [1199]	Speed: 1535.276587 samples/sec	accuracy=0.786272	lr=0.006070
Epoch[110] Batch [1249]	Speed: 1452.401383 samples/sec	accuracy=0.786233	lr=0.006017
[Epoch 110] training: accuracy=0.786220
[Epoch 110] speed: 1507 samples/sec	time cost: 871.158911
[Epoch 110] validation: err-top1=0.235697 err-top5=0.067891
Epoch[111] Batch [49]	Speed: 1479.140698 samples/sec	accuracy=0.796816	lr=0.005963
Epoch[111] Batch [99]	Speed: 1510.110751 samples/sec	accuracy=0.794990	lr=0.005910
Epoch[111] Batch [149]	Speed: 1524.120314 samples/sec	accuracy=0.793835	lr=0.005857
Epoch[111] Batch [199]	Speed: 1493.416004 samples/sec	accuracy=0.792271	lr=0.005805
Epoch[111] Batch [249]	Speed: 1436.836653 samples/sec	accuracy=0.792316	lr=0.005753
Epoch[111] Batch [299]	Speed: 1531.080728 samples/sec	accuracy=0.792106	lr=0.005701
Epoch[111] Batch [349]	Speed: 1518.057785 samples/sec	accuracy=0.791922	lr=0.005649
Epoch[111] Batch [399]	Speed: 1520.054343 samples/sec	accuracy=0.791982	lr=0.005598
Epoch[111] Batch [449]	Speed: 1536.649317 samples/sec	accuracy=0.791953	lr=0.005547
Epoch[111] Batch [499]	Speed: 1545.427689 samples/sec	accuracy=0.791912	lr=0.005496
Epoch[111] Batch [549]	Speed: 1500.023832 samples/sec	accuracy=0.791664	lr=0.005445
Epoch[111] Batch [599]	Speed: 1486.847361 samples/sec	accuracy=0.791541	lr=0.005394
Epoch[111] Batch [649]	Speed: 1539.044217 samples/sec	accuracy=0.791591	lr=0.005344
Epoch[111] Batch [699]	Speed: 1491.623640 samples/sec	accuracy=0.791589	lr=0.005294
Epoch[111] Batch [749]	Speed: 1514.723901 samples/sec	accuracy=0.791565	lr=0.005244
Epoch[111] Batch [799]	Speed: 1487.524065 samples/sec	accuracy=0.791411	lr=0.005195
Epoch[111] Batch [849]	Speed: 1488.254408 samples/sec	accuracy=0.791309	lr=0.005145
Epoch[111] Batch [899]	Speed: 1509.629827 samples/sec	accuracy=0.791415	lr=0.005096
Epoch[111] Batch [949]	Speed: 1529.895844 samples/sec	accuracy=0.791196	lr=0.005047
Epoch[111] Batch [999]	Speed: 1517.663807 samples/sec	accuracy=0.791168	lr=0.004999
Epoch[111] Batch [1049]	Speed: 1542.488832 samples/sec	accuracy=0.791213	lr=0.004950
Epoch[111] Batch [1099]	Speed: 1542.108893 samples/sec	accuracy=0.791199	lr=0.004902
Epoch[111] Batch [1149]	Speed: 1509.699054 samples/sec	accuracy=0.791254	lr=0.004854
Epoch[111] Batch [1199]	Speed: 1524.328463 samples/sec	accuracy=0.791282	lr=0.004807
Epoch[111] Batch [1249]	Speed: 1536.426820 samples/sec	accuracy=0.791143	lr=0.004759
[Epoch 111] training: accuracy=0.791126
[Epoch 111] speed: 1511 samples/sec	time cost: 869.392996
[Epoch 111] validation: err-top1=0.233956 err-top5=0.066685
Epoch[112] Batch [49]	Speed: 1474.480438 samples/sec	accuracy=0.796328	lr=0.004711
Epoch[112] Batch [99]	Speed: 1515.658565 samples/sec	accuracy=0.794658	lr=0.004664
Epoch[112] Batch [149]	Speed: 1517.351328 samples/sec	accuracy=0.794779	lr=0.004617
Epoch[112] Batch [199]	Speed: 1476.747979 samples/sec	accuracy=0.795298	lr=0.004571
Epoch[112] Batch [249]	Speed: 1536.966463 samples/sec	accuracy=0.795734	lr=0.004524
Epoch[112] Batch [299]	Speed: 1523.385110 samples/sec	accuracy=0.795560	lr=0.004478
Epoch[112] Batch [349]	Speed: 1546.403336 samples/sec	accuracy=0.795142	lr=0.004432
Epoch[112] Batch [399]	Speed: 1550.265974 samples/sec	accuracy=0.795283	lr=0.004387
Epoch[112] Batch [449]	Speed: 1531.563076 samples/sec	accuracy=0.795647	lr=0.004341
Epoch[112] Batch [499]	Speed: 1474.770838 samples/sec	accuracy=0.795355	lr=0.004296
Epoch[112] Batch [549]	Speed: 1483.377295 samples/sec	accuracy=0.795446	lr=0.004251
Epoch[112] Batch [599]	Speed: 1541.071249 samples/sec	accuracy=0.795272	lr=0.004207
Epoch[112] Batch [649]	Speed: 1543.763045 samples/sec	accuracy=0.795162	lr=0.004162
Epoch[112] Batch [699]	Speed: 1454.372836 samples/sec	accuracy=0.795134	lr=0.004118
Epoch[112] Batch [749]	Speed: 1533.616894 samples/sec	accuracy=0.795211	lr=0.004074
Epoch[112] Batch [799]	Speed: 1534.455990 samples/sec	accuracy=0.795352	lr=0.004030
Epoch[112] Batch [849]	Speed: 1513.721308 samples/sec	accuracy=0.795205	lr=0.003987
Epoch[112] Batch [899]	Speed: 1497.857460 samples/sec	accuracy=0.795029	lr=0.003944
Epoch[112] Batch [949]	Speed: 1507.822642 samples/sec	accuracy=0.795052	lr=0.003901
Epoch[112] Batch [999]	Speed: 1538.205761 samples/sec	accuracy=0.794989	lr=0.003858
Epoch[112] Batch [1049]	Speed: 1481.074951 samples/sec	accuracy=0.794963	lr=0.003815
Epoch[112] Batch [1099]	Speed: 1501.138944 samples/sec	accuracy=0.794986	lr=0.003773
Epoch[112] Batch [1149]	Speed: 1541.018355 samples/sec	accuracy=0.794986	lr=0.003731
Epoch[112] Batch [1199]	Speed: 1512.259840 samples/sec	accuracy=0.795072	lr=0.003689
Epoch[112] Batch [1249]	Speed: 1557.650241 samples/sec	accuracy=0.795049	lr=0.003647
[Epoch 112] training: accuracy=0.795048
[Epoch 112] speed: 1513 samples/sec	time cost: 868.816067
[Epoch 112] validation: err-top1=0.231824 err-top5=0.065350
Epoch[113] Batch [49]	Speed: 1487.069198 samples/sec	accuracy=0.801641	lr=0.003605
Epoch[113] Batch [99]	Speed: 1512.840058 samples/sec	accuracy=0.800098	lr=0.003564
Epoch[113] Batch [149]	Speed: 1532.029245 samples/sec	accuracy=0.799876	lr=0.003523
Epoch[113] Batch [199]	Speed: 1542.187134 samples/sec	accuracy=0.799009	lr=0.003482
Epoch[113] Batch [249]	Speed: 1508.454534 samples/sec	accuracy=0.799461	lr=0.003442
Epoch[113] Batch [299]	Speed: 1496.395835 samples/sec	accuracy=0.799951	lr=0.003402
Epoch[113] Batch [349]	Speed: 1539.682121 samples/sec	accuracy=0.799788	lr=0.003362
Epoch[113] Batch [399]	Speed: 1544.198085 samples/sec	accuracy=0.799895	lr=0.003322
Epoch[113] Batch [449]	Speed: 1489.874323 samples/sec	accuracy=0.799898	lr=0.003282
Epoch[113] Batch [499]	Speed: 1543.198813 samples/sec	accuracy=0.799635	lr=0.003243
Epoch[113] Batch [549]	Speed: 1538.853115 samples/sec	accuracy=0.799741	lr=0.003204
Epoch[113] Batch [599]	Speed: 1515.586896 samples/sec	accuracy=0.799883	lr=0.003165
Epoch[113] Batch [649]	Speed: 1518.349181 samples/sec	accuracy=0.799784	lr=0.003127
Epoch[113] Batch [699]	Speed: 1496.109801 samples/sec	accuracy=0.799795	lr=0.003088
Epoch[113] Batch [749]	Speed: 1538.272907 samples/sec	accuracy=0.799398	lr=0.003050
Epoch[113] Batch [799]	Speed: 1510.939431 samples/sec	accuracy=0.799397	lr=0.003012
Epoch[113] Batch [849]	Speed: 1484.329893 samples/sec	accuracy=0.799386	lr=0.002975
Epoch[113] Batch [899]	Speed: 1495.528679 samples/sec	accuracy=0.799327	lr=0.002937
Epoch[113] Batch [949]	Speed: 1542.136612 samples/sec	accuracy=0.799397	lr=0.002900
Epoch[113] Batch [999]	Speed: 1461.088576 samples/sec	accuracy=0.799257	lr=0.002863
Epoch[113] Batch [1049]	Speed: 1463.716774 samples/sec	accuracy=0.799189	lr=0.002826
Epoch[113] Batch [1099]	Speed: 1404.571185 samples/sec	accuracy=0.799174	lr=0.002790
Epoch[113] Batch [1149]	Speed: 1487.738240 samples/sec	accuracy=0.799196	lr=0.002754
Epoch[113] Batch [1199]	Speed: 1456.506149 samples/sec	accuracy=0.799141	lr=0.002718
Epoch[113] Batch [1249]	Speed: 1546.068393 samples/sec	accuracy=0.799106	lr=0.002682
[Epoch 113] training: accuracy=0.799106
[Epoch 113] speed: 1504 samples/sec	time cost: 873.629024
[Epoch 113] validation: err-top1=0.229472 err-top5=0.065450
Epoch[114] Batch [49]	Speed: 1503.708505 samples/sec	accuracy=0.804297	lr=0.002646
Epoch[114] Batch [99]	Speed: 1506.896039 samples/sec	accuracy=0.802080	lr=0.002611
Epoch[114] Batch [149]	Speed: 1470.766496 samples/sec	accuracy=0.801934	lr=0.002575
Epoch[114] Batch [199]	Speed: 1477.938402 samples/sec	accuracy=0.802432	lr=0.002541
Epoch[114] Batch [249]	Speed: 1472.169790 samples/sec	accuracy=0.801801	lr=0.002506
Epoch[114] Batch [299]	Speed: 1536.551913 samples/sec	accuracy=0.801761	lr=0.002472
Epoch[114] Batch [349]	Speed: 1519.470934 samples/sec	accuracy=0.802037	lr=0.002438
Epoch[114] Batch [399]	Speed: 1491.128097 samples/sec	accuracy=0.802068	lr=0.002404
Epoch[114] Batch [449]	Speed: 1506.153058 samples/sec	accuracy=0.802281	lr=0.002370
Epoch[114] Batch [499]	Speed: 1530.253214 samples/sec	accuracy=0.802385	lr=0.002337
Epoch[114] Batch [549]	Speed: 1471.792125 samples/sec	accuracy=0.802228	lr=0.002304
Epoch[114] Batch [599]	Speed: 1476.489913 samples/sec	accuracy=0.802240	lr=0.002271
Epoch[114] Batch [649]	Speed: 1507.877569 samples/sec	accuracy=0.802273	lr=0.002238
Epoch[114] Batch [699]	Speed: 1490.436835 samples/sec	accuracy=0.802204	lr=0.002205
Epoch[114] Batch [749]	Speed: 1534.992735 samples/sec	accuracy=0.802329	lr=0.002173
Epoch[114] Batch [799]	Speed: 1512.707287 samples/sec	accuracy=0.802336	lr=0.002141
Epoch[114] Batch [849]	Speed: 1515.833699 samples/sec	accuracy=0.802251	lr=0.002110
Epoch[114] Batch [899]	Speed: 1490.042753 samples/sec	accuracy=0.802316	lr=0.002078
Epoch[114] Batch [949]	Speed: 1505.552519 samples/sec	accuracy=0.802294	lr=0.002047
Epoch[114] Batch [999]	Speed: 1549.461309 samples/sec	accuracy=0.802187	lr=0.002016
Epoch[114] Batch [1049]	Speed: 1501.619554 samples/sec	accuracy=0.802135	lr=0.001985
Epoch[114] Batch [1099]	Speed: 1494.124306 samples/sec	accuracy=0.802067	lr=0.001954
Epoch[114] Batch [1149]	Speed: 1478.980112 samples/sec	accuracy=0.802121	lr=0.001924
Epoch[114] Batch [1199]	Speed: 1507.985752 samples/sec	accuracy=0.801960	lr=0.001894
Epoch[114] Batch [1249]	Speed: 1525.271977 samples/sec	accuracy=0.802119	lr=0.001864
[Epoch 114] training: accuracy=0.802121
[Epoch 114] speed: 1501 samples/sec	time cost: 875.823070
[Epoch 114] validation: err-top1=0.229373 err-top5=0.064951
Epoch[115] Batch [49]	Speed: 1495.453882 samples/sec	accuracy=0.805293	lr=0.001834
Epoch[115] Batch [99]	Speed: 1470.078864 samples/sec	accuracy=0.805049	lr=0.001804
Epoch[115] Batch [149]	Speed: 1455.675192 samples/sec	accuracy=0.805690	lr=0.001775
Epoch[115] Batch [199]	Speed: 1488.732148 samples/sec	accuracy=0.806001	lr=0.001746
Epoch[115] Batch [249]	Speed: 1519.206824 samples/sec	accuracy=0.805352	lr=0.001718
Epoch[115] Batch [299]	Speed: 1535.418421 samples/sec	accuracy=0.805299	lr=0.001689
Epoch[115] Batch [349]	Speed: 1463.703624 samples/sec	accuracy=0.804911	lr=0.001661
Epoch[115] Batch [399]	Speed: 1544.093427 samples/sec	accuracy=0.804666	lr=0.001633
Epoch[115] Batch [449]	Speed: 1493.655992 samples/sec	accuracy=0.804861	lr=0.001605
Epoch[115] Batch [499]	Speed: 1535.899507 samples/sec	accuracy=0.804771	lr=0.001578
Epoch[115] Batch [549]	Speed: 1452.760895 samples/sec	accuracy=0.804640	lr=0.001551
Epoch[115] Batch [599]	Speed: 1491.807554 samples/sec	accuracy=0.804378	lr=0.001524
Epoch[115] Batch [649]	Speed: 1508.689597 samples/sec	accuracy=0.804337	lr=0.001497
Epoch[115] Batch [699]	Speed: 1525.632251 samples/sec	accuracy=0.804329	lr=0.001470
Epoch[115] Batch [749]	Speed: 1535.521028 samples/sec	accuracy=0.804176	lr=0.001444
Epoch[115] Batch [799]	Speed: 1537.657826 samples/sec	accuracy=0.804221	lr=0.001418
Epoch[115] Batch [849]	Speed: 1473.420205 samples/sec	accuracy=0.804301	lr=0.001392
Epoch[115] Batch [899]	Speed: 1508.569444 samples/sec	accuracy=0.804166	lr=0.001366
Epoch[115] Batch [949]	Speed: 1482.840995 samples/sec	accuracy=0.804300	lr=0.001341
Epoch[115] Batch [999]	Speed: 1528.267172 samples/sec	accuracy=0.804551	lr=0.001316
Epoch[115] Batch [1049]	Speed: 1468.933046 samples/sec	accuracy=0.804736	lr=0.001291
Epoch[115] Batch [1099]	Speed: 1477.594920 samples/sec	accuracy=0.804816	lr=0.001266
Epoch[115] Batch [1149]	Speed: 1532.916182 samples/sec	accuracy=0.804851	lr=0.001242
Epoch[115] Batch [1199]	Speed: 1535.961541 samples/sec	accuracy=0.804917	lr=0.001218
Epoch[115] Batch [1249]	Speed: 1542.015092 samples/sec	accuracy=0.804948	lr=0.001194
[Epoch 115] training: accuracy=0.804946
[Epoch 115] speed: 1502 samples/sec	time cost: 874.910565
[Epoch 115] validation: err-top1=0.228436 err-top5=0.064553
Epoch[116] Batch [49]	Speed: 1454.727826 samples/sec	accuracy=0.808340	lr=0.001170
Epoch[116] Batch [99]	Speed: 1489.686193 samples/sec	accuracy=0.808018	lr=0.001146
Epoch[116] Batch [149]	Speed: 1514.736198 samples/sec	accuracy=0.807591	lr=0.001123
Epoch[116] Batch [199]	Speed: 1539.893250 samples/sec	accuracy=0.807051	lr=0.001100
Epoch[116] Batch [249]	Speed: 1533.010007 samples/sec	accuracy=0.807156	lr=0.001077
Epoch[116] Batch [299]	Speed: 1542.764879 samples/sec	accuracy=0.807321	lr=0.001055
Epoch[116] Batch [349]	Speed: 1514.138031 samples/sec	accuracy=0.807316	lr=0.001032
Epoch[116] Batch [399]	Speed: 1543.520067 samples/sec	accuracy=0.807417	lr=0.001010
Epoch[116] Batch [449]	Speed: 1478.616019 samples/sec	accuracy=0.807426	lr=0.000989
Epoch[116] Batch [499]	Speed: 1538.286030 samples/sec	accuracy=0.807635	lr=0.000967
Epoch[116] Batch [549]	Speed: 1515.138300 samples/sec	accuracy=0.807433	lr=0.000946
Epoch[116] Batch [599]	Speed: 1521.990800 samples/sec	accuracy=0.807202	lr=0.000925
Epoch[116] Batch [649]	Speed: 1507.363097 samples/sec	accuracy=0.807144	lr=0.000904
Epoch[116] Batch [699]	Speed: 1499.721893 samples/sec	accuracy=0.807347	lr=0.000883
Epoch[116] Batch [749]	Speed: 1523.356192 samples/sec	accuracy=0.807474	lr=0.000863
Epoch[116] Batch [799]	Speed: 1495.455308 samples/sec	accuracy=0.807601	lr=0.000843
Epoch[116] Batch [849]	Speed: 1507.189745 samples/sec	accuracy=0.807575	lr=0.000823
Epoch[116] Batch [899]	Speed: 1511.220890 samples/sec	accuracy=0.807555	lr=0.000803
Epoch[116] Batch [949]	Speed: 1539.522005 samples/sec	accuracy=0.807510	lr=0.000784
Epoch[116] Batch [999]	Speed: 1541.641728 samples/sec	accuracy=0.807653	lr=0.000764
Epoch[116] Batch [1049]	Speed: 1472.227136 samples/sec	accuracy=0.807824	lr=0.000745
Epoch[116] Batch [1099]	Speed: 1499.973048 samples/sec	accuracy=0.807958	lr=0.000727
Epoch[116] Batch [1149]	Speed: 1511.590271 samples/sec	accuracy=0.807893	lr=0.000708
Epoch[116] Batch [1199]	Speed: 1500.560419 samples/sec	accuracy=0.807863	lr=0.000690
Epoch[116] Batch [1249]	Speed: 1519.632003 samples/sec	accuracy=0.807899	lr=0.000672
[Epoch 116] training: accuracy=0.807914
[Epoch 116] speed: 1511 samples/sec	time cost: 869.395842
[Epoch 116] validation: err-top1=0.227478 err-top5=0.064779
Epoch[117] Batch [49]	Speed: 1512.969931 samples/sec	accuracy=0.811484	lr=0.000654
Epoch[117] Batch [99]	Speed: 1534.230335 samples/sec	accuracy=0.807715	lr=0.000636
Epoch[117] Batch [149]	Speed: 1506.194785 samples/sec	accuracy=0.807695	lr=0.000619
Epoch[117] Batch [199]	Speed: 1507.833811 samples/sec	accuracy=0.807319	lr=0.000602
Epoch[117] Batch [249]	Speed: 1510.942631 samples/sec	accuracy=0.808547	lr=0.000585
Epoch[117] Batch [299]	Speed: 1542.522902 samples/sec	accuracy=0.808659	lr=0.000569
Epoch[117] Batch [349]	Speed: 1489.566072 samples/sec	accuracy=0.808783	lr=0.000552
Epoch[117] Batch [399]	Speed: 1492.183918 samples/sec	accuracy=0.808923	lr=0.000536
Epoch[117] Batch [449]	Speed: 1487.535925 samples/sec	accuracy=0.808832	lr=0.000520
Epoch[117] Batch [499]	Speed: 1517.263376 samples/sec	accuracy=0.808820	lr=0.000505
Epoch[117] Batch [549]	Speed: 1507.091555 samples/sec	accuracy=0.808825	lr=0.000489
Epoch[117] Batch [599]	Speed: 1504.967002 samples/sec	accuracy=0.808737	lr=0.000474
Epoch[117] Batch [649]	Speed: 1521.228218 samples/sec	accuracy=0.808729	lr=0.000459
Epoch[117] Batch [699]	Speed: 1514.367606 samples/sec	accuracy=0.808781	lr=0.000445
Epoch[117] Batch [749]	Speed: 1543.459673 samples/sec	accuracy=0.808595	lr=0.000430
Epoch[117] Batch [799]	Speed: 1491.518434 samples/sec	accuracy=0.808446	lr=0.000416
Epoch[117] Batch [849]	Speed: 1514.218360 samples/sec	accuracy=0.808549	lr=0.000402
Epoch[117] Batch [899]	Speed: 1515.131822 samples/sec	accuracy=0.808557	lr=0.000388
Epoch[117] Batch [949]	Speed: 1527.318602 samples/sec	accuracy=0.808602	lr=0.000375
Epoch[117] Batch [999]	Speed: 1484.504676 samples/sec	accuracy=0.808558	lr=0.000362
Epoch[117] Batch [1049]	Speed: 1486.075000 samples/sec	accuracy=0.808568	lr=0.000349
Epoch[117] Batch [1099]	Speed: 1514.862849 samples/sec	accuracy=0.808679	lr=0.000336
Epoch[117] Batch [1149]	Speed: 1508.537812 samples/sec	accuracy=0.808654	lr=0.000323
Epoch[117] Batch [1199]	Speed: 1507.785271 samples/sec	accuracy=0.808566	lr=0.000311
Epoch[117] Batch [1249]	Speed: 1545.788803 samples/sec	accuracy=0.808629	lr=0.000299
[Epoch 117] training: accuracy=0.808626
[Epoch 117] speed: 1510 samples/sec	time cost: 870.172962
[Epoch 117] validation: err-top1=0.226901 err-top5=0.064553
Epoch[118] Batch [49]	Speed: 1506.922094 samples/sec	accuracy=0.810391	lr=0.000287
Epoch[118] Batch [99]	Speed: 1502.461623 samples/sec	accuracy=0.810264	lr=0.000275
Epoch[118] Batch [149]	Speed: 1488.403335 samples/sec	accuracy=0.809811	lr=0.000264
Epoch[118] Batch [199]	Speed: 1542.203448 samples/sec	accuracy=0.809829	lr=0.000253
Epoch[118] Batch [249]	Speed: 1511.220646 samples/sec	accuracy=0.809449	lr=0.000242
Epoch[118] Batch [299]	Speed: 1503.200345 samples/sec	accuracy=0.809596	lr=0.000231
Epoch[118] Batch [349]	Speed: 1518.905849 samples/sec	accuracy=0.809205	lr=0.000221
Epoch[118] Batch [399]	Speed: 1523.905680 samples/sec	accuracy=0.809495	lr=0.000211
Epoch[118] Batch [449]	Speed: 1529.472408 samples/sec	accuracy=0.809486	lr=0.000201
Epoch[118] Batch [499]	Speed: 1503.360141 samples/sec	accuracy=0.809324	lr=0.000191
Epoch[118] Batch [549]	Speed: 1516.304441 samples/sec	accuracy=0.809412	lr=0.000182
Epoch[118] Batch [599]	Speed: 1463.642112 samples/sec	accuracy=0.809237	lr=0.000173
Epoch[118] Batch [649]	Speed: 1524.303728 samples/sec	accuracy=0.809516	lr=0.000164
Epoch[118] Batch [699]	Speed: 1513.442159 samples/sec	accuracy=0.809495	lr=0.000155
Epoch[118] Batch [749]	Speed: 1514.948054 samples/sec	accuracy=0.809281	lr=0.000147
Epoch[118] Batch [799]	Speed: 1540.482836 samples/sec	accuracy=0.809362	lr=0.000138
Epoch[118] Batch [849]	Speed: 1517.333949 samples/sec	accuracy=0.809331	lr=0.000130
Epoch[118] Batch [899]	Speed: 1514.087696 samples/sec	accuracy=0.809433	lr=0.000123
Epoch[118] Batch [949]	Speed: 1481.722057 samples/sec	accuracy=0.809497	lr=0.000115
Epoch[118] Batch [999]	Speed: 1492.267513 samples/sec	accuracy=0.809514	lr=0.000108
Epoch[118] Batch [1049]	Speed: 1504.266204 samples/sec	accuracy=0.809600	lr=0.000101
Epoch[118] Batch [1099]	Speed: 1498.567347 samples/sec	accuracy=0.809561	lr=0.000094
Epoch[118] Batch [1149]	Speed: 1482.776788 samples/sec	accuracy=0.809652	lr=0.000087
Epoch[118] Batch [1199]	Speed: 1545.101717 samples/sec	accuracy=0.809542	lr=0.000081
Epoch[118] Batch [1249]	Speed: 1528.558214 samples/sec	accuracy=0.809521	lr=0.000075
[Epoch 118] training: accuracy=0.809517
[Epoch 118] speed: 1509 samples/sec	time cost: 870.973769
[Epoch 118] validation: err-top1=0.226543 err-top5=0.064174
Epoch[119] Batch [49]	Speed: 1498.762068 samples/sec	accuracy=0.809863	lr=0.000069
Epoch[119] Batch [99]	Speed: 1506.933863 samples/sec	accuracy=0.809873	lr=0.000063
Epoch[119] Batch [149]	Speed: 1496.611456 samples/sec	accuracy=0.809238	lr=0.000058
Epoch[119] Batch [199]	Speed: 1448.326320 samples/sec	accuracy=0.809380	lr=0.000053
Epoch[119] Batch [249]	Speed: 1476.752153 samples/sec	accuracy=0.809574	lr=0.000048
Epoch[119] Batch [299]	Speed: 1546.302610 samples/sec	accuracy=0.809554	lr=0.000043
Epoch[119] Batch [349]	Speed: 1515.280842 samples/sec	accuracy=0.809489	lr=0.000039
Epoch[119] Batch [399]	Speed: 1530.438522 samples/sec	accuracy=0.809961	lr=0.000035
Epoch[119] Batch [449]	Speed: 1490.282940 samples/sec	accuracy=0.809883	lr=0.000031
Epoch[119] Batch [499]	Speed: 1518.857366 samples/sec	accuracy=0.809787	lr=0.000027
Epoch[119] Batch [549]	Speed: 1499.875283 samples/sec	accuracy=0.809719	lr=0.000023
Epoch[119] Batch [599]	Speed: 1477.836745 samples/sec	accuracy=0.809946	lr=0.000020
Epoch[119] Batch [649]	Speed: 1493.421166 samples/sec	accuracy=0.809835	lr=0.000017
Epoch[119] Batch [699]	Speed: 1495.372366 samples/sec	accuracy=0.809922	lr=0.000015
Epoch[119] Batch [749]	Speed: 1530.797814 samples/sec	accuracy=0.810005	lr=0.000012
Epoch[119] Batch [799]	Speed: 1498.651429 samples/sec	accuracy=0.809890	lr=0.000010
Epoch[119] Batch [849]	Speed: 1464.214627 samples/sec	accuracy=0.809823	lr=0.000008
Epoch[119] Batch [899]	Speed: 1463.019609 samples/sec	accuracy=0.809803	lr=0.000006
Epoch[119] Batch [949]	Speed: 1500.664786 samples/sec	accuracy=0.809815	lr=0.000004
Epoch[119] Batch [999]	Speed: 1541.006711 samples/sec	accuracy=0.809631	lr=0.000003
Epoch[119] Batch [1049]	Speed: 1517.602105 samples/sec	accuracy=0.809796	lr=0.000002
Epoch[119] Batch [1099]	Speed: 1542.505607 samples/sec	accuracy=0.809691	lr=0.000001
Epoch[119] Batch [1149]	Speed: 1488.202087 samples/sec	accuracy=0.809812	lr=0.000000
Epoch[119] Batch [1199]	Speed: 1521.607489 samples/sec	accuracy=0.809749	lr=0.000000
Epoch[119] Batch [1249]	Speed: 1516.640782 samples/sec	accuracy=0.809788	lr=0.000000
[Epoch 119] training: accuracy=0.809771
[Epoch 119] speed: 1501 samples/sec	time cost: 874.980576
[Epoch 119] validation: err-top1=0.226822 err-top5=0.064134