Namespace(accumulate=1, batch_norm=False, batch_size=25, clip_grad=0, crop_ratio=0.875, data_dir='/home/ubuntu/yizhu/data/kinetics400/kinetics400/train_256', dataset='kinetics400', dtype='float32', eval=False, fast_temporal_stride=2, hard_weight=0.5, hashtag='', input_5d=False, input_size=224, kvstore=None, label_smoothing=False, last_gamma=False, log_interval=20, logging_file='tsn_2d_rgb_inceptionv1_seg7_kinetics400_b25_g8_cleanv3.txt', lr=0.01, lr_decay=0.1, lr_decay_epoch='40,80,100', lr_decay_period=0, lr_mode='step', mixup=False, mixup_alpha=0.2, mixup_off_epoch=0, mode='hybrid', model='inceptionv1_kinetics400', momentum=0.9, new_height=256, new_length=1, new_step=1, new_width=340, no_wd=False, num_classes=400, num_crop=1, num_epochs=100, num_gpus=8, num_segments=7, num_workers=32, partial_bn=False, prefetch_ratio=1.0, resume_epoch=0, resume_params='', resume_states='', save_dir='/home/ubuntu/yizhu/logs/kinetcis400/resnet2d/tsn_2d_rgb_inceptionv1_seg7_kinetics400_b25_g8_cleanv3', save_frequency=10, scale_ratios='1.0,0.8', slow_temporal_stride=16, slowfast=False, teacher=None, temperature=20, train_list='/home/ubuntu/yizhu/data/kinetics400/kinetics400/k400_train_240618.txt', use_amp=False, use_decord=True, use_gn=False, use_pretrained=False, use_se=False, use_tsn=False, val_data_dir='/home/ubuntu/yizhu/data/kinetics400/kinetics400/val_256', val_list='/home/ubuntu/yizhu/code/gluon-cv/extra/kinetics400/k400_val_19761_cleanv3.txt', video_loader=True, warmup_epochs=0, warmup_lr=0.0, wd=0.0001)
Total batch size is set to 200 on 8 GPUs
Model file not found. Downloading.
ActionRecInceptionV1(
  (conv1): HybridSequential(
    (0): Conv2D(3 -> 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
    (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=64)
    (2): Activation(relu)
  )
  (maxpool1): MaxPool2D(size=(3, 3), stride=(2, 2), padding=(0, 0), ceil_mode=True, global_pool=False, pool_type=max, layout=NCHW)
  (conv2): HybridSequential(
    (0): Conv2D(64 -> 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
    (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=64)
    (2): Activation(relu)
  )
  (conv3): HybridSequential(
    (0): Conv2D(64 -> 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
    (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=192)
    (2): Activation(relu)
  )
  (maxpool2): MaxPool2D(size=(3, 3), stride=(2, 2), padding=(0, 0), ceil_mode=True, global_pool=False, pool_type=max, layout=NCHW)
  (inception3a): HybridConcurrent(
    (0): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(192 -> 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=64)
        (2): Activation(relu)
      )
    )
    (1): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(192 -> 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=96)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(96 -> 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=128)
        (2): Activation(relu)
      )
    )
    (2): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(192 -> 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=16)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(16 -> 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=32)
        (2): Activation(relu)
      )
    )
    (3): HybridSequential(
      (0): MaxPool2D(size=(3, 3), stride=(1, 1), padding=(1, 1), ceil_mode=False, global_pool=False, pool_type=max, layout=NCHW)
      (1): HybridSequential(
        (0): Conv2D(192 -> 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=32)
        (2): Activation(relu)
      )
    )
  )
  (inception3b): HybridConcurrent(
    (0): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(256 -> 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=128)
        (2): Activation(relu)
      )
    )
    (1): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(256 -> 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=128)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(128 -> 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=192)
        (2): Activation(relu)
      )
    )
    (2): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(256 -> 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=32)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(32 -> 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=96)
        (2): Activation(relu)
      )
    )
    (3): HybridSequential(
      (0): MaxPool2D(size=(3, 3), stride=(1, 1), padding=(1, 1), ceil_mode=False, global_pool=False, pool_type=max, layout=NCHW)
      (1): HybridSequential(
        (0): Conv2D(256 -> 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=64)
        (2): Activation(relu)
      )
    )
  )
  (maxpool3): MaxPool2D(size=(3, 3), stride=(2, 2), padding=(0, 0), ceil_mode=True, global_pool=False, pool_type=max, layout=NCHW)
  (inception4a): HybridConcurrent(
    (0): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(480 -> 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=192)
        (2): Activation(relu)
      )
    )
    (1): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(480 -> 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=96)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(96 -> 208, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=208)
        (2): Activation(relu)
      )
    )
    (2): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(480 -> 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=16)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(16 -> 48, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=48)
        (2): Activation(relu)
      )
    )
    (3): HybridSequential(
      (0): MaxPool2D(size=(3, 3), stride=(1, 1), padding=(1, 1), ceil_mode=False, global_pool=False, pool_type=max, layout=NCHW)
      (1): HybridSequential(
        (0): Conv2D(480 -> 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=64)
        (2): Activation(relu)
      )
    )
  )
  (inception4b): HybridConcurrent(
    (0): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(512 -> 160, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=160)
        (2): Activation(relu)
      )
    )
    (1): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(512 -> 112, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=112)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(112 -> 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=224)
        (2): Activation(relu)
      )
    )
    (2): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(512 -> 24, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=24)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(24 -> 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=64)
        (2): Activation(relu)
      )
    )
    (3): HybridSequential(
      (0): MaxPool2D(size=(3, 3), stride=(1, 1), padding=(1, 1), ceil_mode=False, global_pool=False, pool_type=max, layout=NCHW)
      (1): HybridSequential(
        (0): Conv2D(512 -> 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=64)
        (2): Activation(relu)
      )
    )
  )
  (inception4c): HybridConcurrent(
    (0): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(512 -> 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=128)
        (2): Activation(relu)
      )
    )
    (1): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(512 -> 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=128)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(128 -> 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=256)
        (2): Activation(relu)
      )
    )
    (2): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(512 -> 24, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=24)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(24 -> 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=64)
        (2): Activation(relu)
      )
    )
    (3): HybridSequential(
      (0): MaxPool2D(size=(3, 3), stride=(1, 1), padding=(1, 1), ceil_mode=False, global_pool=False, pool_type=max, layout=NCHW)
      (1): HybridSequential(
        (0): Conv2D(512 -> 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=64)
        (2): Activation(relu)
      )
    )
  )
  (inception4d): HybridConcurrent(
    (0): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(512 -> 112, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=112)
        (2): Activation(relu)
      )
    )
    (1): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(512 -> 144, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=144)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(144 -> 288, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=288)
        (2): Activation(relu)
      )
    )
    (2): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(512 -> 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=32)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(32 -> 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=64)
        (2): Activation(relu)
      )
    )
    (3): HybridSequential(
      (0): MaxPool2D(size=(3, 3), stride=(1, 1), padding=(1, 1), ceil_mode=False, global_pool=False, pool_type=max, layout=NCHW)
      (1): HybridSequential(
        (0): Conv2D(512 -> 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=64)
        (2): Activation(relu)
      )
    )
  )
  (inception4e): HybridConcurrent(
    (0): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(528 -> 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=256)
        (2): Activation(relu)
      )
    )
    (1): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(528 -> 160, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=160)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(160 -> 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=320)
        (2): Activation(relu)
      )
    )
    (2): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(528 -> 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=32)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(32 -> 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=128)
        (2): Activation(relu)
      )
    )
    (3): HybridSequential(
      (0): MaxPool2D(size=(3, 3), stride=(1, 1), padding=(1, 1), ceil_mode=False, global_pool=False, pool_type=max, layout=NCHW)
      (1): HybridSequential(
        (0): Conv2D(528 -> 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=128)
        (2): Activation(relu)
      )
    )
  )
  (maxpool4): MaxPool2D(size=(2, 2), stride=(2, 2), padding=(0, 0), ceil_mode=False, global_pool=False, pool_type=max, layout=NCHW)
  (inception5a): HybridConcurrent(
    (0): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(832 -> 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=256)
        (2): Activation(relu)
      )
    )
    (1): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(832 -> 160, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=160)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(160 -> 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=320)
        (2): Activation(relu)
      )
    )
    (2): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(832 -> 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=32)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(32 -> 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=128)
        (2): Activation(relu)
      )
    )
    (3): HybridSequential(
      (0): MaxPool2D(size=(3, 3), stride=(1, 1), padding=(1, 1), ceil_mode=False, global_pool=False, pool_type=max, layout=NCHW)
      (1): HybridSequential(
        (0): Conv2D(832 -> 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=128)
        (2): Activation(relu)
      )
    )
  )
  (inception5b): HybridConcurrent(
    (0): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(832 -> 384, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=384)
        (2): Activation(relu)
      )
    )
    (1): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(832 -> 192, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=192)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(192 -> 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=384)
        (2): Activation(relu)
      )
    )
    (2): HybridSequential(
      (0): HybridSequential(
        (0): Conv2D(832 -> 48, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=48)
        (2): Activation(relu)
      )
      (1): HybridSequential(
        (0): Conv2D(48 -> 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=128)
        (2): Activation(relu)
      )
    )
    (3): HybridSequential(
      (0): MaxPool2D(size=(3, 3), stride=(1, 1), padding=(1, 1), ceil_mode=False, global_pool=False, pool_type=max, layout=NCHW)
      (1): HybridSequential(
        (0): Conv2D(832 -> 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=128)
        (2): Activation(relu)
      )
    )
  )
  (avgpool): AvgPool2D(size=(7, 7), stride=(7, 7), padding=(0, 0), ceil_mode=False, global_pool=False, pool_type=avg, layout=NCHW)
  (dropout): Dropout(p = 0.5, axes=())
  (output): Dense(1024 -> 400, linear)
)
Load 240618 training samples and 19404 validation samples.
Epoch[000] Batch [0019]/[1203]	Speed: 55.771397 samples/sec	 accuracy=0.600000	 loss=5.954186	 lr=0.010000
Epoch[000] Batch [0039]/[1203]	Speed: 95.189365 samples/sec	 accuracy=1.787500	 loss=5.865907	 lr=0.010000
Epoch[000] Batch [0059]/[1203]	Speed: 231.227261 samples/sec	 accuracy=3.441667	 loss=5.756378	 lr=0.010000
Epoch[000] Batch [0079]/[1203]	Speed: 89.468043 samples/sec	 accuracy=5.037500	 loss=5.621164	 lr=0.010000
Epoch[000] Batch [0099]/[1203]	Speed: 99.584691 samples/sec	 accuracy=6.460000	 loss=5.474860	 lr=0.010000
Epoch[000] Batch [0119]/[1203]	Speed: 189.817491 samples/sec	 accuracy=7.941667	 loss=5.335990	 lr=0.010000
Epoch[000] Batch [0139]/[1203]	Speed: 94.994093 samples/sec	 accuracy=9.082143	 loss=5.195342	 lr=0.010000
Epoch[000] Batch [0159]/[1203]	Speed: 188.122067 samples/sec	 accuracy=10.287500	 loss=5.067986	 lr=0.010000
Epoch[000] Batch [0179]/[1203]	Speed: 98.242572 samples/sec	 accuracy=11.363889	 loss=4.948880	 lr=0.010000
Epoch[000] Batch [0199]/[1203]	Speed: 101.881780 samples/sec	 accuracy=12.417500	 loss=4.842874	 lr=0.010000
Epoch[000] Batch [0219]/[1203]	Speed: 149.422258 samples/sec	 accuracy=13.384091	 loss=4.746271	 lr=0.010000
Epoch[000] Batch [0239]/[1203]	Speed: 95.558365 samples/sec	 accuracy=14.306250	 loss=4.652745	 lr=0.010000
Epoch[000] Batch [0259]/[1203]	Speed: 127.323188 samples/sec	 accuracy=15.205769	 loss=4.567517	 lr=0.010000
Epoch[000] Batch [0279]/[1203]	Speed: 134.904632 samples/sec	 accuracy=16.139286	 loss=4.487417	 lr=0.010000
Epoch[000] Batch [0299]/[1203]	Speed: 99.310837 samples/sec	 accuracy=16.888333	 loss=4.416412	 lr=0.010000
Epoch[000] Batch [0319]/[1203]	Speed: 152.988221 samples/sec	 accuracy=17.634375	 loss=4.349122	 lr=0.010000
Epoch[000] Batch [0339]/[1203]	Speed: 99.052378 samples/sec	 accuracy=18.410294	 loss=4.281581	 lr=0.010000
Epoch[000] Batch [0359]/[1203]	Speed: 124.119809 samples/sec	 accuracy=19.036111	 loss=4.225557	 lr=0.010000
Epoch[000] Batch [0379]/[1203]	Speed: 124.163922 samples/sec	 accuracy=19.688158	 loss=4.169489	 lr=0.010000
Epoch[000] Batch [0399]/[1203]	Speed: 109.239467 samples/sec	 accuracy=20.283750	 loss=4.116856	 lr=0.010000
Epoch[000] Batch [0419]/[1203]	Speed: 144.214587 samples/sec	 accuracy=20.834524	 loss=4.067684	 lr=0.010000
Epoch[000] Batch [0439]/[1203]	Speed: 97.606140 samples/sec	 accuracy=21.425000	 loss=4.019968	 lr=0.010000
Epoch[000] Batch [0459]/[1203]	Speed: 127.156501 samples/sec	 accuracy=21.945652	 loss=3.975890	 lr=0.010000
Epoch[000] Batch [0479]/[1203]	Speed: 111.933481 samples/sec	 accuracy=22.417708	 loss=3.934982	 lr=0.010000
Epoch[000] Batch [0499]/[1203]	Speed: 115.933700 samples/sec	 accuracy=22.956000	 loss=3.894140	 lr=0.010000
Epoch[000] Batch [0519]/[1203]	Speed: 151.094527 samples/sec	 accuracy=23.425962	 loss=3.856503	 lr=0.010000
Epoch[000] Batch [0539]/[1203]	Speed: 101.392111 samples/sec	 accuracy=23.887037	 loss=3.818897	 lr=0.010000
Epoch[000] Batch [0559]/[1203]	Speed: 109.982598 samples/sec	 accuracy=24.333929	 loss=3.785059	 lr=0.010000
Epoch[000] Batch [0579]/[1203]	Speed: 130.415573 samples/sec	 accuracy=24.789655	 loss=3.752153	 lr=0.010000
Epoch[000] Batch [0599]/[1203]	Speed: 124.377017 samples/sec	 accuracy=25.178333	 loss=3.720181	 lr=0.010000
Epoch[000] Batch [0619]/[1203]	Speed: 115.006407 samples/sec	 accuracy=25.567742	 loss=3.690631	 lr=0.010000
Epoch[000] Batch [0639]/[1203]	Speed: 114.758717 samples/sec	 accuracy=25.931250	 loss=3.662372	 lr=0.010000
Epoch[000] Batch [0659]/[1203]	Speed: 107.230556 samples/sec	 accuracy=26.285606	 loss=3.634578	 lr=0.010000
Epoch[000] Batch [0679]/[1203]	Speed: 135.204388 samples/sec	 accuracy=26.608824	 loss=3.608484	 lr=0.010000
Epoch[000] Batch [0699]/[1203]	Speed: 126.666658 samples/sec	 accuracy=26.913571	 loss=3.584134	 lr=0.010000
Epoch[000] Batch [0719]/[1203]	Speed: 106.601383 samples/sec	 accuracy=27.231250	 loss=3.559502	 lr=0.010000
Epoch[000] Batch [0739]/[1203]	Speed: 116.973284 samples/sec	 accuracy=27.542568	 loss=3.537196	 lr=0.010000
Epoch[000] Batch [0759]/[1203]	Speed: 124.253263 samples/sec	 accuracy=27.825000	 loss=3.515242	 lr=0.010000
Epoch[000] Batch [0779]/[1203]	Speed: 123.148679 samples/sec	 accuracy=28.116667	 loss=3.493215	 lr=0.010000
Epoch[000] Batch [0799]/[1203]	Speed: 99.464236 samples/sec	 accuracy=28.415625	 loss=3.472184	 lr=0.010000
Epoch[000] Batch [0819]/[1203]	Speed: 149.277835 samples/sec	 accuracy=28.717073	 loss=3.450056	 lr=0.010000
Epoch[000] Batch [0839]/[1203]	Speed: 104.679388 samples/sec	 accuracy=28.998810	 loss=3.429419	 lr=0.010000
Epoch[000] Batch [0859]/[1203]	Speed: 126.292812 samples/sec	 accuracy=29.265116	 loss=3.409474	 lr=0.010000
Epoch[000] Batch [0879]/[1203]	Speed: 121.293251 samples/sec	 accuracy=29.503409	 loss=3.391793	 lr=0.010000
Epoch[000] Batch [0899]/[1203]	Speed: 106.066568 samples/sec	 accuracy=29.792222	 loss=3.372708	 lr=0.010000
Epoch[000] Batch [0919]/[1203]	Speed: 130.156271 samples/sec	 accuracy=30.041304	 loss=3.354333	 lr=0.010000
Epoch[000] Batch [0939]/[1203]	Speed: 115.380011 samples/sec	 accuracy=30.286702	 loss=3.336835	 lr=0.010000
Epoch[000] Batch [0959]/[1203]	Speed: 108.921277 samples/sec	 accuracy=30.505729	 loss=3.320924	 lr=0.010000
Epoch[000] Batch [0979]/[1203]	Speed: 148.784856 samples/sec	 accuracy=30.729082	 loss=3.304867	 lr=0.010000
Epoch[000] Batch [0999]/[1203]	Speed: 93.147074 samples/sec	 accuracy=30.926000	 loss=3.289148	 lr=0.010000
Epoch[000] Batch [1019]/[1203]	Speed: 137.456021 samples/sec	 accuracy=31.106863	 loss=3.274965	 lr=0.010000
Epoch[000] Batch [1039]/[1203]	Speed: 106.328755 samples/sec	 accuracy=31.321635	 loss=3.260749	 lr=0.010000
Epoch[000] Batch [1059]/[1203]	Speed: 125.170007 samples/sec	 accuracy=31.508491	 loss=3.247159	 lr=0.010000
Epoch[000] Batch [1079]/[1203]	Speed: 137.543904 samples/sec	 accuracy=31.723148	 loss=3.233414	 lr=0.010000
Epoch[000] Batch [1099]/[1203]	Speed: 96.343168 samples/sec	 accuracy=31.927727	 loss=3.219897	 lr=0.010000
Epoch[000] Batch [1119]/[1203]	Speed: 134.386369 samples/sec	 accuracy=32.127679	 loss=3.206501	 lr=0.010000
Epoch[000] Batch [1139]/[1203]	Speed: 113.439870 samples/sec	 accuracy=32.302632	 loss=3.193298	 lr=0.010000
Epoch[000] Batch [1159]/[1203]	Speed: 120.344194 samples/sec	 accuracy=32.495690	 loss=3.180298	 lr=0.010000
Epoch[000] Batch [1179]/[1203]	Speed: 140.700893 samples/sec	 accuracy=32.663136	 loss=3.168343	 lr=0.010000
Epoch[000] Batch [1199]/[1203]	Speed: 187.889544 samples/sec	 accuracy=32.833750	 loss=3.156205	 lr=0.010000
Batch [0019]/[0097]: acc-top1=40.500000 acc-top5=68.375000
Batch [0039]/[0097]: acc-top1=39.100000 acc-top5=65.587500
Batch [0059]/[0097]: acc-top1=41.625000 acc-top5=67.575000
Batch [0079]/[0097]: acc-top1=40.950000 acc-top5=67.300000
[Epoch 000] training: accuracy=32.856608	 loss=3.154551
[Epoch 000] speed: 116 samples/sec	time cost: 2254.167413
[Epoch 000] validation: acc-top1=40.046392 acc-top5=67.556701 loss=2.587691
Epoch[001] Batch [0019]/[1203]	Speed: 67.871860 samples/sec	 accuracy=44.950000	 loss=2.404179	 lr=0.010000
Epoch[001] Batch [0039]/[1203]	Speed: 98.392353 samples/sec	 accuracy=44.887500	 loss=2.388485	 lr=0.010000
Epoch[001] Batch [0059]/[1203]	Speed: 228.889068 samples/sec	 accuracy=44.983333	 loss=2.371987	 lr=0.010000
Epoch[001] Batch [0079]/[1203]	Speed: 89.128180 samples/sec	 accuracy=44.731250	 loss=2.368131	 lr=0.010000
Epoch[001] Batch [0099]/[1203]	Speed: 109.522561 samples/sec	 accuracy=45.040000	 loss=2.356976	 lr=0.010000
Epoch[001] Batch [0119]/[1203]	Speed: 187.830877 samples/sec	 accuracy=45.054167	 loss=2.356432	 lr=0.010000
Epoch[001] Batch [0139]/[1203]	Speed: 99.678309 samples/sec	 accuracy=45.060714	 loss=2.351353	 lr=0.010000
Epoch[001] Batch [0159]/[1203]	Speed: 193.555867 samples/sec	 accuracy=45.300000	 loss=2.341850	 lr=0.010000
Epoch[001] Batch [0179]/[1203]	Speed: 96.107360 samples/sec	 accuracy=45.219444	 loss=2.341271	 lr=0.010000
Epoch[001] Batch [0199]/[1203]	Speed: 105.921931 samples/sec	 accuracy=45.317500	 loss=2.340713	 lr=0.010000
Epoch[001] Batch [0219]/[1203]	Speed: 161.418929 samples/sec	 accuracy=45.350000	 loss=2.341069	 lr=0.010000
Epoch[001] Batch [0239]/[1203]	Speed: 110.424259 samples/sec	 accuracy=45.345833	 loss=2.341842	 lr=0.010000
Epoch[001] Batch [0259]/[1203]	Speed: 129.250066 samples/sec	 accuracy=45.348077	 loss=2.340155	 lr=0.010000
Epoch[001] Batch [0279]/[1203]	Speed: 113.777061 samples/sec	 accuracy=45.360714	 loss=2.340929	 lr=0.010000
Epoch[001] Batch [0299]/[1203]	Speed: 108.936933 samples/sec	 accuracy=45.438333	 loss=2.335590	 lr=0.010000
Epoch[001] Batch [0319]/[1203]	Speed: 131.767297 samples/sec	 accuracy=45.473438	 loss=2.334471	 lr=0.010000
Epoch[001] Batch [0339]/[1203]	Speed: 106.890035 samples/sec	 accuracy=45.457353	 loss=2.334331	 lr=0.010000
Epoch[001] Batch [0359]/[1203]	Speed: 134.432015 samples/sec	 accuracy=45.509722	 loss=2.331786	 lr=0.010000
Epoch[001] Batch [0379]/[1203]	Speed: 122.852363 samples/sec	 accuracy=45.500000	 loss=2.330105	 lr=0.010000
Epoch[001] Batch [0399]/[1203]	Speed: 124.020478 samples/sec	 accuracy=45.561250	 loss=2.326349	 lr=0.010000
Epoch[001] Batch [0419]/[1203]	Speed: 119.552984 samples/sec	 accuracy=45.695238	 loss=2.321597	 lr=0.010000
Epoch[001] Batch [0439]/[1203]	Speed: 110.400170 samples/sec	 accuracy=45.684091	 loss=2.320236	 lr=0.010000
Epoch[001] Batch [0459]/[1203]	Speed: 137.352741 samples/sec	 accuracy=45.698913	 loss=2.318917	 lr=0.010000
Epoch[001] Batch [0479]/[1203]	Speed: 122.550793 samples/sec	 accuracy=45.702083	 loss=2.317899	 lr=0.010000
Epoch[001] Batch [0499]/[1203]	Speed: 107.595126 samples/sec	 accuracy=45.715000	 loss=2.316828	 lr=0.010000
Epoch[001] Batch [0519]/[1203]	Speed: 128.673995 samples/sec	 accuracy=45.741346	 loss=2.315470	 lr=0.010000
Epoch[001] Batch [0539]/[1203]	Speed: 110.813357 samples/sec	 accuracy=45.841667	 loss=2.311027	 lr=0.010000
Epoch[001] Batch [0559]/[1203]	Speed: 140.094622 samples/sec	 accuracy=45.883929	 loss=2.307178	 lr=0.010000
Epoch[001] Batch [0579]/[1203]	Speed: 115.364459 samples/sec	 accuracy=45.925000	 loss=2.306049	 lr=0.010000
Epoch[001] Batch [0599]/[1203]	Speed: 101.450170 samples/sec	 accuracy=45.933333	 loss=2.303987	 lr=0.010000
Epoch[001] Batch [0619]/[1203]	Speed: 141.063245 samples/sec	 accuracy=45.958871	 loss=2.302503	 lr=0.010000
Epoch[001] Batch [0639]/[1203]	Speed: 121.782022 samples/sec	 accuracy=46.051562	 loss=2.298296	 lr=0.010000
Epoch[001] Batch [0659]/[1203]	Speed: 109.724588 samples/sec	 accuracy=46.056061	 loss=2.297744	 lr=0.010000
Epoch[001] Batch [0679]/[1203]	Speed: 132.972748 samples/sec	 accuracy=46.083088	 loss=2.295787	 lr=0.010000
Epoch[001] Batch [0699]/[1203]	Speed: 106.586315 samples/sec	 accuracy=46.084286	 loss=2.295196	 lr=0.010000
Epoch[001] Batch [0719]/[1203]	Speed: 147.419061 samples/sec	 accuracy=46.138194	 loss=2.293495	 lr=0.010000
Epoch[001] Batch [0739]/[1203]	Speed: 102.815920 samples/sec	 accuracy=46.168919	 loss=2.291201	 lr=0.010000
Epoch[001] Batch [0759]/[1203]	Speed: 121.690880 samples/sec	 accuracy=46.208553	 loss=2.290691	 lr=0.010000
Epoch[001] Batch [0779]/[1203]	Speed: 119.558958 samples/sec	 accuracy=46.232692	 loss=2.290191	 lr=0.010000
Epoch[001] Batch [0799]/[1203]	Speed: 119.095178 samples/sec	 accuracy=46.230000	 loss=2.289624	 lr=0.010000
Epoch[001] Batch [0819]/[1203]	Speed: 144.712398 samples/sec	 accuracy=46.264634	 loss=2.287865	 lr=0.010000
Epoch[001] Batch [0839]/[1203]	Speed: 108.226175 samples/sec	 accuracy=46.319048	 loss=2.285854	 lr=0.010000
Epoch[001] Batch [0859]/[1203]	Speed: 120.634038 samples/sec	 accuracy=46.354651	 loss=2.283485	 lr=0.010000
Epoch[001] Batch [0879]/[1203]	Speed: 126.660885 samples/sec	 accuracy=46.413636	 loss=2.281828	 lr=0.010000
Epoch[001] Batch [0899]/[1203]	Speed: 118.096191 samples/sec	 accuracy=46.416667	 loss=2.280883	 lr=0.010000
Epoch[001] Batch [0919]/[1203]	Speed: 122.355697 samples/sec	 accuracy=46.433152	 loss=2.280142	 lr=0.010000
Epoch[001] Batch [0939]/[1203]	Speed: 124.404885 samples/sec	 accuracy=46.464894	 loss=2.278515	 lr=0.010000
Epoch[001] Batch [0959]/[1203]	Speed: 109.213867 samples/sec	 accuracy=46.484896	 loss=2.277395	 lr=0.010000
Epoch[001] Batch [0979]/[1203]	Speed: 126.080843 samples/sec	 accuracy=46.520408	 loss=2.274849	 lr=0.010000
Epoch[001] Batch [0999]/[1203]	Speed: 110.333638 samples/sec	 accuracy=46.583000	 loss=2.271473	 lr=0.010000
Epoch[001] Batch [1019]/[1203]	Speed: 117.445468 samples/sec	 accuracy=46.600980	 loss=2.270342	 lr=0.010000
Epoch[001] Batch [1039]/[1203]	Speed: 131.933620 samples/sec	 accuracy=46.629808	 loss=2.268576	 lr=0.010000
Epoch[001] Batch [1059]/[1203]	Speed: 116.550365 samples/sec	 accuracy=46.664623	 loss=2.267761	 lr=0.010000
Epoch[001] Batch [1079]/[1203]	Speed: 139.261615 samples/sec	 accuracy=46.684722	 loss=2.266462	 lr=0.010000
Epoch[001] Batch [1099]/[1203]	Speed: 107.422656 samples/sec	 accuracy=46.703182	 loss=2.264928	 lr=0.010000
Epoch[001] Batch [1119]/[1203]	Speed: 114.233875 samples/sec	 accuracy=46.754464	 loss=2.262669	 lr=0.010000
Epoch[001] Batch [1139]/[1203]	Speed: 140.469659 samples/sec	 accuracy=46.767105	 loss=2.262526	 lr=0.010000
Epoch[001] Batch [1159]/[1203]	Speed: 113.092070 samples/sec	 accuracy=46.800000	 loss=2.260767	 lr=0.010000
Epoch[001] Batch [1179]/[1203]	Speed: 115.194506 samples/sec	 accuracy=46.842373	 loss=2.258865	 lr=0.010000
Epoch[001] Batch [1199]/[1203]	Speed: 276.216811 samples/sec	 accuracy=46.873333	 loss=2.258055	 lr=0.010000
Batch [0019]/[0097]: acc-top1=47.200000 acc-top5=73.675000
Batch [0039]/[0097]: acc-top1=44.750000 acc-top5=71.000000
Batch [0059]/[0097]: acc-top1=47.725000 acc-top5=72.891667
Batch [0079]/[0097]: acc-top1=47.612500 acc-top5=73.275000
[Epoch 001] training: accuracy=46.875312	 loss=2.257778
[Epoch 001] speed: 120 samples/sec	time cost: 2197.009237
[Epoch 001] validation: acc-top1=46.716495 acc-top5=73.010309 loss=2.281428
Epoch[002] Batch [0019]/[1203]	Speed: 68.082447 samples/sec	 accuracy=49.875000	 loss=2.110576	 lr=0.010000
Epoch[002] Batch [0039]/[1203]	Speed: 98.237059 samples/sec	 accuracy=50.262500	 loss=2.072485	 lr=0.010000
Epoch[002] Batch [0059]/[1203]	Speed: 216.796248 samples/sec	 accuracy=50.141667	 loss=2.084848	 lr=0.010000
Epoch[002] Batch [0079]/[1203]	Speed: 94.126809 samples/sec	 accuracy=50.262500	 loss=2.086708	 lr=0.010000
Epoch[002] Batch [0099]/[1203]	Speed: 116.482408 samples/sec	 accuracy=50.490000	 loss=2.081253	 lr=0.010000
Epoch[002] Batch [0119]/[1203]	Speed: 146.114168 samples/sec	 accuracy=50.425000	 loss=2.090933	 lr=0.010000
Epoch[002] Batch [0139]/[1203]	Speed: 100.989594 samples/sec	 accuracy=50.614286	 loss=2.083770	 lr=0.010000
Epoch[002] Batch [0159]/[1203]	Speed: 188.765483 samples/sec	 accuracy=50.606250	 loss=2.084258	 lr=0.010000
Epoch[002] Batch [0179]/[1203]	Speed: 87.443734 samples/sec	 accuracy=50.655556	 loss=2.077603	 lr=0.010000
Epoch[002] Batch [0199]/[1203]	Speed: 132.770934 samples/sec	 accuracy=50.587500	 loss=2.077530	 lr=0.010000
Epoch[002] Batch [0219]/[1203]	Speed: 133.476649 samples/sec	 accuracy=50.675000	 loss=2.076933	 lr=0.010000
Epoch[002] Batch [0239]/[1203]	Speed: 101.651583 samples/sec	 accuracy=50.731250	 loss=2.071351	 lr=0.010000
Epoch[002] Batch [0259]/[1203]	Speed: 167.689280 samples/sec	 accuracy=50.757692	 loss=2.069756	 lr=0.010000
Epoch[002] Batch [0279]/[1203]	Speed: 93.275194 samples/sec	 accuracy=50.796429	 loss=2.069132	 lr=0.010000
Epoch[002] Batch [0299]/[1203]	Speed: 161.627689 samples/sec	 accuracy=50.868333	 loss=2.065354	 lr=0.010000
Epoch[002] Batch [0319]/[1203]	Speed: 106.843829 samples/sec	 accuracy=50.898438	 loss=2.063404	 lr=0.010000
Epoch[002] Batch [0339]/[1203]	Speed: 113.051523 samples/sec	 accuracy=50.875000	 loss=2.064558	 lr=0.010000
Epoch[002] Batch [0359]/[1203]	Speed: 147.900666 samples/sec	 accuracy=50.854167	 loss=2.067367	 lr=0.010000
Epoch[002] Batch [0379]/[1203]	Speed: 106.213480 samples/sec	 accuracy=50.848684	 loss=2.066251	 lr=0.010000
Epoch[002] Batch [0399]/[1203]	Speed: 115.047099 samples/sec	 accuracy=50.865000	 loss=2.066256	 lr=0.010000
Epoch[002] Batch [0419]/[1203]	Speed: 140.735220 samples/sec	 accuracy=50.919048	 loss=2.064707	 lr=0.010000
Epoch[002] Batch [0439]/[1203]	Speed: 105.267201 samples/sec	 accuracy=50.946591	 loss=2.064576	 lr=0.010000
Epoch[002] Batch [0459]/[1203]	Speed: 144.559320 samples/sec	 accuracy=51.027174	 loss=2.062181	 lr=0.010000
Epoch[002] Batch [0479]/[1203]	Speed: 116.491774 samples/sec	 accuracy=50.998958	 loss=2.062275	 lr=0.010000
Epoch[002] Batch [0499]/[1203]	Speed: 110.568102 samples/sec	 accuracy=51.011000	 loss=2.061011	 lr=0.010000
Epoch[002] Batch [0519]/[1203]	Speed: 126.511705 samples/sec	 accuracy=51.046154	 loss=2.059168	 lr=0.010000
Epoch[002] Batch [0539]/[1203]	Speed: 128.675981 samples/sec	 accuracy=51.053704	 loss=2.058356	 lr=0.010000
Epoch[002] Batch [0559]/[1203]	Speed: 117.774376 samples/sec	 accuracy=51.066071	 loss=2.056545	 lr=0.010000
Epoch[002] Batch [0579]/[1203]	Speed: 112.837162 samples/sec	 accuracy=51.056034	 loss=2.055893	 lr=0.010000
Epoch[002] Batch [0599]/[1203]	Speed: 124.624282 samples/sec	 accuracy=51.054167	 loss=2.055097	 lr=0.010000
Epoch[002] Batch [0619]/[1203]	Speed: 133.384241 samples/sec	 accuracy=51.102419	 loss=2.053834	 lr=0.010000
Epoch[002] Batch [0639]/[1203]	Speed: 119.127497 samples/sec	 accuracy=51.125000	 loss=2.053090	 lr=0.010000
Epoch[002] Batch [0659]/[1203]	Speed: 103.342659 samples/sec	 accuracy=51.146970	 loss=2.051897	 lr=0.010000
Epoch[002] Batch [0679]/[1203]	Speed: 128.225322 samples/sec	 accuracy=51.136765	 loss=2.052181	 lr=0.010000
Epoch[002] Batch [0699]/[1203]	Speed: 136.641303 samples/sec	 accuracy=51.132143	 loss=2.051879	 lr=0.010000
Epoch[002] Batch [0719]/[1203]	Speed: 115.997540 samples/sec	 accuracy=51.120833	 loss=2.052201	 lr=0.010000
Epoch[002] Batch [0739]/[1203]	Speed: 106.051897 samples/sec	 accuracy=51.134459	 loss=2.051634	 lr=0.010000
Epoch[002] Batch [0759]/[1203]	Speed: 122.464002 samples/sec	 accuracy=51.116447	 loss=2.051532	 lr=0.010000
Epoch[002] Batch [0779]/[1203]	Speed: 131.359672 samples/sec	 accuracy=51.100641	 loss=2.052328	 lr=0.010000
Epoch[002] Batch [0799]/[1203]	Speed: 122.110546 samples/sec	 accuracy=51.105000	 loss=2.052247	 lr=0.010000
Epoch[002] Batch [0819]/[1203]	Speed: 112.656841 samples/sec	 accuracy=51.133537	 loss=2.051531	 lr=0.010000
Epoch[002] Batch [0839]/[1203]	Speed: 125.039424 samples/sec	 accuracy=51.139881	 loss=2.050482	 lr=0.010000
Epoch[002] Batch [0859]/[1203]	Speed: 122.695866 samples/sec	 accuracy=51.134884	 loss=2.051037	 lr=0.010000
Epoch[002] Batch [0879]/[1203]	Speed: 125.036663 samples/sec	 accuracy=51.123295	 loss=2.051259	 lr=0.010000
Epoch[002] Batch [0899]/[1203]	Speed: 122.219287 samples/sec	 accuracy=51.138333	 loss=2.050609	 lr=0.010000
Epoch[002] Batch [0919]/[1203]	Speed: 106.163872 samples/sec	 accuracy=51.140217	 loss=2.050602	 lr=0.010000
Epoch[002] Batch [0939]/[1203]	Speed: 125.620558 samples/sec	 accuracy=51.155851	 loss=2.049558	 lr=0.010000
Epoch[002] Batch [0959]/[1203]	Speed: 125.602311 samples/sec	 accuracy=51.152083	 loss=2.049465	 lr=0.010000
Epoch[002] Batch [0979]/[1203]	Speed: 117.354213 samples/sec	 accuracy=51.193367	 loss=2.048289	 lr=0.010000
Epoch[002] Batch [0999]/[1203]	Speed: 135.756983 samples/sec	 accuracy=51.235000	 loss=2.046729	 lr=0.010000
Epoch[002] Batch [1019]/[1203]	Speed: 104.649805 samples/sec	 accuracy=51.244118	 loss=2.046214	 lr=0.010000
Epoch[002] Batch [1039]/[1203]	Speed: 130.320269 samples/sec	 accuracy=51.276442	 loss=2.044964	 lr=0.010000
Epoch[002] Batch [1059]/[1203]	Speed: 123.651478 samples/sec	 accuracy=51.273113	 loss=2.045199	 lr=0.010000
Epoch[002] Batch [1079]/[1203]	Speed: 116.278487 samples/sec	 accuracy=51.290278	 loss=2.044902	 lr=0.010000
Epoch[002] Batch [1099]/[1203]	Speed: 142.186041 samples/sec	 accuracy=51.305455	 loss=2.044411	 lr=0.010000
Epoch[002] Batch [1119]/[1203]	Speed: 105.561889 samples/sec	 accuracy=51.323661	 loss=2.043656	 lr=0.010000
Epoch[002] Batch [1139]/[1203]	Speed: 113.154095 samples/sec	 accuracy=51.347368	 loss=2.042767	 lr=0.010000
Epoch[002] Batch [1159]/[1203]	Speed: 138.270031 samples/sec	 accuracy=51.335776	 loss=2.042493	 lr=0.010000
Epoch[002] Batch [1179]/[1203]	Speed: 113.515096 samples/sec	 accuracy=51.304661	 loss=2.043675	 lr=0.010000
Epoch[002] Batch [1199]/[1203]	Speed: 269.858474 samples/sec	 accuracy=51.310417	 loss=2.043764	 lr=0.010000
Batch [0019]/[0097]: acc-top1=50.025000 acc-top5=74.525000
Batch [0039]/[0097]: acc-top1=48.137500 acc-top5=73.012500
Batch [0059]/[0097]: acc-top1=51.341667 acc-top5=75.391667
Batch [0079]/[0097]: acc-top1=50.881250 acc-top5=75.512500
[Epoch 002] training: accuracy=51.307564	 loss=2.043988
[Epoch 002] speed: 120 samples/sec	time cost: 2193.168125
[Epoch 002] validation: acc-top1=49.989691 acc-top5=75.453608 loss=2.143242
Epoch[003] Batch [0019]/[1203]	Speed: 66.400237 samples/sec	 accuracy=54.075000	 loss=1.880330	 lr=0.010000
Epoch[003] Batch [0039]/[1203]	Speed: 100.930823 samples/sec	 accuracy=53.750000	 loss=1.889128	 lr=0.010000
Epoch[003] Batch [0059]/[1203]	Speed: 225.374154 samples/sec	 accuracy=53.966667	 loss=1.885353	 lr=0.010000
Epoch[003] Batch [0079]/[1203]	Speed: 92.968746 samples/sec	 accuracy=53.856250	 loss=1.912593	 lr=0.010000
Epoch[003] Batch [0099]/[1203]	Speed: 100.979870 samples/sec	 accuracy=53.655000	 loss=1.913533	 lr=0.010000
Epoch[003] Batch [0119]/[1203]	Speed: 173.274864 samples/sec	 accuracy=54.033333	 loss=1.896454	 lr=0.010000
Epoch[003] Batch [0139]/[1203]	Speed: 95.726227 samples/sec	 accuracy=54.160714	 loss=1.892779	 lr=0.010000
Epoch[003] Batch [0159]/[1203]	Speed: 200.032232 samples/sec	 accuracy=54.075000	 loss=1.900814	 lr=0.010000
Epoch[003] Batch [0179]/[1203]	Speed: 102.082562 samples/sec	 accuracy=53.886111	 loss=1.906186	 lr=0.010000
Epoch[003] Batch [0199]/[1203]	Speed: 98.797267 samples/sec	 accuracy=53.817500	 loss=1.910534	 lr=0.010000
Epoch[003] Batch [0219]/[1203]	Speed: 165.652042 samples/sec	 accuracy=53.861364	 loss=1.910777	 lr=0.010000
Epoch[003] Batch [0239]/[1203]	Speed: 109.128851 samples/sec	 accuracy=53.835417	 loss=1.913273	 lr=0.010000
Epoch[003] Batch [0259]/[1203]	Speed: 112.265817 samples/sec	 accuracy=53.971154	 loss=1.908190	 lr=0.010000
Epoch[003] Batch [0279]/[1203]	Speed: 131.166939 samples/sec	 accuracy=53.912500	 loss=1.911123	 lr=0.010000
Epoch[003] Batch [0299]/[1203]	Speed: 111.950830 samples/sec	 accuracy=53.946667	 loss=1.909852	 lr=0.010000
Epoch[003] Batch [0319]/[1203]	Speed: 149.576395 samples/sec	 accuracy=53.996875	 loss=1.911801	 lr=0.010000
Epoch[003] Batch [0339]/[1203]	Speed: 107.256556 samples/sec	 accuracy=53.969118	 loss=1.911891	 lr=0.010000
Epoch[003] Batch [0359]/[1203]	Speed: 111.438218 samples/sec	 accuracy=53.940278	 loss=1.912411	 lr=0.010000
Epoch[003] Batch [0379]/[1203]	Speed: 133.740402 samples/sec	 accuracy=53.919737	 loss=1.912754	 lr=0.010000
Epoch[003] Batch [0399]/[1203]	Speed: 120.614043 samples/sec	 accuracy=53.905000	 loss=1.913707	 lr=0.010000
Epoch[003] Batch [0419]/[1203]	Speed: 139.154506 samples/sec	 accuracy=53.901190	 loss=1.913603	 lr=0.010000
Epoch[003] Batch [0439]/[1203]	Speed: 99.190890 samples/sec	 accuracy=53.942045	 loss=1.910740	 lr=0.010000
Epoch[003] Batch [0459]/[1203]	Speed: 137.831956 samples/sec	 accuracy=53.917391	 loss=1.912310	 lr=0.010000
Epoch[003] Batch [0479]/[1203]	Speed: 116.281554 samples/sec	 accuracy=53.900000	 loss=1.912212	 lr=0.010000
Epoch[003] Batch [0499]/[1203]	Speed: 124.265360 samples/sec	 accuracy=53.919000	 loss=1.911529	 lr=0.010000
Epoch[003] Batch [0519]/[1203]	Speed: 129.628302 samples/sec	 accuracy=53.930769	 loss=1.912476	 lr=0.010000
Epoch[003] Batch [0539]/[1203]	Speed: 100.935981 samples/sec	 accuracy=53.949074	 loss=1.911887	 lr=0.010000
Epoch[003] Batch [0559]/[1203]	Speed: 129.058576 samples/sec	 accuracy=53.964286	 loss=1.910409	 lr=0.010000
Epoch[003] Batch [0579]/[1203]	Speed: 137.125363 samples/sec	 accuracy=53.967241	 loss=1.909407	 lr=0.010000
Epoch[003] Batch [0599]/[1203]	Speed: 116.152032 samples/sec	 accuracy=53.970833	 loss=1.909368	 lr=0.010000
Epoch[003] Batch [0619]/[1203]	Speed: 122.735266 samples/sec	 accuracy=53.962903	 loss=1.910660	 lr=0.010000
Epoch[003] Batch [0639]/[1203]	Speed: 106.927605 samples/sec	 accuracy=53.953125	 loss=1.911148	 lr=0.010000
Epoch[003] Batch [0659]/[1203]	Speed: 127.219641 samples/sec	 accuracy=53.964394	 loss=1.910443	 lr=0.010000
Epoch[003] Batch [0679]/[1203]	Speed: 121.161745 samples/sec	 accuracy=53.911765	 loss=1.912495	 lr=0.010000
Epoch[003] Batch [0699]/[1203]	Speed: 117.337925 samples/sec	 accuracy=53.884286	 loss=1.912654	 lr=0.010000
Epoch[003] Batch [0719]/[1203]	Speed: 122.023135 samples/sec	 accuracy=53.865972	 loss=1.912860	 lr=0.010000
Epoch[003] Batch [0739]/[1203]	Speed: 129.935406 samples/sec	 accuracy=53.870946	 loss=1.912696	 lr=0.010000
Epoch[003] Batch [0759]/[1203]	Speed: 106.968165 samples/sec	 accuracy=53.842105	 loss=1.913905	 lr=0.010000
Epoch[003] Batch [0779]/[1203]	Speed: 133.012116 samples/sec	 accuracy=53.853205	 loss=1.913337	 lr=0.010000
Epoch[003] Batch [0799]/[1203]	Speed: 115.354201 samples/sec	 accuracy=53.861250	 loss=1.914015	 lr=0.010000
Epoch[003] Batch [0819]/[1203]	Speed: 119.110508 samples/sec	 accuracy=53.860976	 loss=1.914123	 lr=0.010000
Epoch[003] Batch [0839]/[1203]	Speed: 128.000560 samples/sec	 accuracy=53.822619	 loss=1.915084	 lr=0.010000
Epoch[003] Batch [0859]/[1203]	Speed: 116.724440 samples/sec	 accuracy=53.814535	 loss=1.914343	 lr=0.010000
Epoch[003] Batch [0879]/[1203]	Speed: 123.197641 samples/sec	 accuracy=53.844318	 loss=1.914033	 lr=0.010000
Epoch[003] Batch [0899]/[1203]	Speed: 115.520877 samples/sec	 accuracy=53.860000	 loss=1.913491	 lr=0.010000
Epoch[003] Batch [0919]/[1203]	Speed: 133.997705 samples/sec	 accuracy=53.865217	 loss=1.913891	 lr=0.010000
Epoch[003] Batch [0939]/[1203]	Speed: 107.872276 samples/sec	 accuracy=53.855851	 loss=1.913358	 lr=0.010000
Epoch[003] Batch [0959]/[1203]	Speed: 109.365114 samples/sec	 accuracy=53.845312	 loss=1.913914	 lr=0.010000
Epoch[003] Batch [0979]/[1203]	Speed: 144.819638 samples/sec	 accuracy=53.831633	 loss=1.914247	 lr=0.010000
Epoch[003] Batch [0999]/[1203]	Speed: 120.584930 samples/sec	 accuracy=53.831000	 loss=1.914239	 lr=0.010000
Epoch[003] Batch [1019]/[1203]	Speed: 117.705991 samples/sec	 accuracy=53.849020	 loss=1.914263	 lr=0.010000
Epoch[003] Batch [1039]/[1203]	Speed: 133.088134 samples/sec	 accuracy=53.859135	 loss=1.914443	 lr=0.010000
Epoch[003] Batch [1059]/[1203]	Speed: 98.608307 samples/sec	 accuracy=53.837264	 loss=1.915810	 lr=0.010000
Epoch[003] Batch [1079]/[1203]	Speed: 151.527526 samples/sec	 accuracy=53.838889	 loss=1.915754	 lr=0.010000
Epoch[003] Batch [1099]/[1203]	Speed: 119.580186 samples/sec	 accuracy=53.846818	 loss=1.915298	 lr=0.010000
Epoch[003] Batch [1119]/[1203]	Speed: 115.556346 samples/sec	 accuracy=53.854464	 loss=1.915280	 lr=0.010000
Epoch[003] Batch [1139]/[1203]	Speed: 116.735522 samples/sec	 accuracy=53.878947	 loss=1.914096	 lr=0.010000
Epoch[003] Batch [1159]/[1203]	Speed: 117.275545 samples/sec	 accuracy=53.900862	 loss=1.913283	 lr=0.010000
Epoch[003] Batch [1179]/[1203]	Speed: 139.055862 samples/sec	 accuracy=53.889407	 loss=1.914344	 lr=0.010000
Epoch[003] Batch [1199]/[1203]	Speed: 228.066317 samples/sec	 accuracy=53.885833	 loss=1.914493	 lr=0.010000
Batch [0019]/[0097]: acc-top1=52.675000 acc-top5=77.750000
Batch [0039]/[0097]: acc-top1=50.662500 acc-top5=76.537500
Batch [0059]/[0097]: acc-top1=53.791667 acc-top5=78.108333
Batch [0079]/[0097]: acc-top1=53.781250 acc-top5=78.037500
[Epoch 003] training: accuracy=53.881962	 loss=1.914526
[Epoch 003] speed: 120 samples/sec	time cost: 2195.595589
[Epoch 003] validation: acc-top1=52.871134 acc-top5=78.056701 loss=2.001868
Epoch[004] Batch [0019]/[1203]	Speed: 68.021963 samples/sec	 accuracy=56.550000	 loss=1.791372	 lr=0.010000
Epoch[004] Batch [0039]/[1203]	Speed: 96.662468 samples/sec	 accuracy=55.937500	 loss=1.801798	 lr=0.010000
Epoch[004] Batch [0059]/[1203]	Speed: 226.834223 samples/sec	 accuracy=56.275000	 loss=1.799840	 lr=0.010000
Epoch[004] Batch [0079]/[1203]	Speed: 93.225454 samples/sec	 accuracy=56.050000	 loss=1.805327	 lr=0.010000
Epoch[004] Batch [0099]/[1203]	Speed: 107.090715 samples/sec	 accuracy=55.940000	 loss=1.804909	 lr=0.010000
Epoch[004] Batch [0119]/[1203]	Speed: 177.044353 samples/sec	 accuracy=55.695833	 loss=1.816161	 lr=0.010000
Epoch[004] Batch [0139]/[1203]	Speed: 95.139568 samples/sec	 accuracy=55.796429	 loss=1.812887	 lr=0.010000
Epoch[004] Batch [0159]/[1203]	Speed: 185.007112 samples/sec	 accuracy=55.728125	 loss=1.816069	 lr=0.010000
Epoch[004] Batch [0179]/[1203]	Speed: 95.666410 samples/sec	 accuracy=55.772222	 loss=1.816474	 lr=0.010000
Epoch[004] Batch [0199]/[1203]	Speed: 120.479239 samples/sec	 accuracy=55.740000	 loss=1.819607	 lr=0.010000
Epoch[004] Batch [0219]/[1203]	Speed: 139.709964 samples/sec	 accuracy=55.640909	 loss=1.822211	 lr=0.010000
Epoch[004] Batch [0239]/[1203]	Speed: 102.903013 samples/sec	 accuracy=55.731250	 loss=1.822380	 lr=0.010000
Epoch[004] Batch [0259]/[1203]	Speed: 139.003633 samples/sec	 accuracy=55.617308	 loss=1.824914	 lr=0.010000
Epoch[004] Batch [0279]/[1203]	Speed: 107.539912 samples/sec	 accuracy=55.551786	 loss=1.827164	 lr=0.010000
Epoch[004] Batch [0299]/[1203]	Speed: 116.240619 samples/sec	 accuracy=55.510000	 loss=1.829790	 lr=0.010000
Epoch[004] Batch [0319]/[1203]	Speed: 144.648171 samples/sec	 accuracy=55.503125	 loss=1.830393	 lr=0.010000
Epoch[004] Batch [0339]/[1203]	Speed: 111.099697 samples/sec	 accuracy=55.508824	 loss=1.828565	 lr=0.010000
Epoch[004] Batch [0359]/[1203]	Speed: 126.270052 samples/sec	 accuracy=55.540278	 loss=1.828303	 lr=0.010000
Epoch[004] Batch [0379]/[1203]	Speed: 111.763195 samples/sec	 accuracy=55.597368	 loss=1.828073	 lr=0.010000
Epoch[004] Batch [0399]/[1203]	Speed: 132.270546 samples/sec	 accuracy=55.581250	 loss=1.830511	 lr=0.010000
Epoch[004] Batch [0419]/[1203]	Speed: 129.787061 samples/sec	 accuracy=55.576190	 loss=1.829131	 lr=0.010000
Epoch[004] Batch [0439]/[1203]	Speed: 110.933676 samples/sec	 accuracy=55.619318	 loss=1.828778	 lr=0.010000
Epoch[004] Batch [0459]/[1203]	Speed: 117.623352 samples/sec	 accuracy=55.605435	 loss=1.828427	 lr=0.010000
Epoch[004] Batch [0479]/[1203]	Speed: 112.587175 samples/sec	 accuracy=55.642708	 loss=1.826997	 lr=0.010000
Epoch[004] Batch [0499]/[1203]	Speed: 117.111204 samples/sec	 accuracy=55.604000	 loss=1.828610	 lr=0.010000
Epoch[004] Batch [0519]/[1203]	Speed: 149.888572 samples/sec	 accuracy=55.615385	 loss=1.828815	 lr=0.010000
Epoch[004] Batch [0539]/[1203]	Speed: 98.740992 samples/sec	 accuracy=55.639815	 loss=1.828577	 lr=0.010000
Epoch[004] Batch [0559]/[1203]	Speed: 144.457093 samples/sec	 accuracy=55.630357	 loss=1.830024	 lr=0.010000
Epoch[004] Batch [0579]/[1203]	Speed: 108.496609 samples/sec	 accuracy=55.611207	 loss=1.829670	 lr=0.010000
Epoch[004] Batch [0599]/[1203]	Speed: 106.536516 samples/sec	 accuracy=55.624167	 loss=1.827989	 lr=0.010000
Epoch[004] Batch [0619]/[1203]	Speed: 169.148383 samples/sec	 accuracy=55.666129	 loss=1.826566	 lr=0.010000
Epoch[004] Batch [0639]/[1203]	Speed: 104.971568 samples/sec	 accuracy=55.707031	 loss=1.824682	 lr=0.010000
Epoch[004] Batch [0659]/[1203]	Speed: 117.059180 samples/sec	 accuracy=55.746212	 loss=1.823981	 lr=0.010000
Epoch[004] Batch [0679]/[1203]	Speed: 130.864962 samples/sec	 accuracy=55.786029	 loss=1.822623	 lr=0.010000
Epoch[004] Batch [0699]/[1203]	Speed: 101.567771 samples/sec	 accuracy=55.782857	 loss=1.823606	 lr=0.010000
Epoch[004] Batch [0719]/[1203]	Speed: 162.911093 samples/sec	 accuracy=55.805556	 loss=1.823363	 lr=0.010000
Epoch[004] Batch [0739]/[1203]	Speed: 107.578327 samples/sec	 accuracy=55.791892	 loss=1.823881	 lr=0.010000
Epoch[004] Batch [0759]/[1203]	Speed: 110.361762 samples/sec	 accuracy=55.840132	 loss=1.822059	 lr=0.010000
Epoch[004] Batch [0779]/[1203]	Speed: 136.733320 samples/sec	 accuracy=55.832692	 loss=1.821359	 lr=0.010000
Epoch[004] Batch [0799]/[1203]	Speed: 111.129683 samples/sec	 accuracy=55.851875	 loss=1.820069	 lr=0.010000
Epoch[004] Batch [0819]/[1203]	Speed: 143.222875 samples/sec	 accuracy=55.843902	 loss=1.820630	 lr=0.010000
Epoch[004] Batch [0839]/[1203]	Speed: 100.614750 samples/sec	 accuracy=55.836905	 loss=1.820928	 lr=0.010000
Epoch[004] Batch [0859]/[1203]	Speed: 120.818663 samples/sec	 accuracy=55.837791	 loss=1.821221	 lr=0.010000
Epoch[004] Batch [0879]/[1203]	Speed: 133.046723 samples/sec	 accuracy=55.827841	 loss=1.822501	 lr=0.010000
Epoch[004] Batch [0899]/[1203]	Speed: 121.644022 samples/sec	 accuracy=55.847778	 loss=1.821965	 lr=0.010000
Epoch[004] Batch [0919]/[1203]	Speed: 131.064793 samples/sec	 accuracy=55.839130	 loss=1.822576	 lr=0.010000
Epoch[004] Batch [0939]/[1203]	Speed: 101.310909 samples/sec	 accuracy=55.839362	 loss=1.822201	 lr=0.010000
Epoch[004] Batch [0959]/[1203]	Speed: 128.238496 samples/sec	 accuracy=55.850000	 loss=1.821544	 lr=0.010000
Epoch[004] Batch [0979]/[1203]	Speed: 134.154561 samples/sec	 accuracy=55.828571	 loss=1.821633	 lr=0.010000
Epoch[004] Batch [0999]/[1203]	Speed: 105.505262 samples/sec	 accuracy=55.841500	 loss=1.821819	 lr=0.010000
Epoch[004] Batch [1019]/[1203]	Speed: 114.199405 samples/sec	 accuracy=55.836765	 loss=1.822203	 lr=0.010000
Epoch[004] Batch [1039]/[1203]	Speed: 137.716333 samples/sec	 accuracy=55.823558	 loss=1.823262	 lr=0.010000
Epoch[004] Batch [1059]/[1203]	Speed: 135.719360 samples/sec	 accuracy=55.823113	 loss=1.823653	 lr=0.010000
Epoch[004] Batch [1079]/[1203]	Speed: 108.632416 samples/sec	 accuracy=55.814815	 loss=1.824002	 lr=0.010000
Epoch[004] Batch [1099]/[1203]	Speed: 113.066413 samples/sec	 accuracy=55.816364	 loss=1.823889	 lr=0.010000
Epoch[004] Batch [1119]/[1203]	Speed: 131.985207 samples/sec	 accuracy=55.816518	 loss=1.824019	 lr=0.010000
Epoch[004] Batch [1139]/[1203]	Speed: 122.320459 samples/sec	 accuracy=55.812281	 loss=1.823784	 lr=0.010000
Epoch[004] Batch [1159]/[1203]	Speed: 99.433497 samples/sec	 accuracy=55.824569	 loss=1.822698	 lr=0.010000
Epoch[004] Batch [1179]/[1203]	Speed: 157.027403 samples/sec	 accuracy=55.832627	 loss=1.822643	 lr=0.010000
Epoch[004] Batch [1199]/[1203]	Speed: 222.228024 samples/sec	 accuracy=55.835000	 loss=1.822453	 lr=0.010000
Batch [0019]/[0097]: acc-top1=53.250000 acc-top5=78.100000
Batch [0039]/[0097]: acc-top1=51.750000 acc-top5=76.862500
Batch [0059]/[0097]: acc-top1=54.100000 acc-top5=78.475000
Batch [0079]/[0097]: acc-top1=53.750000 acc-top5=78.056250
[Epoch 004] training: accuracy=55.829177	 loss=1.822595
[Epoch 004] speed: 119 samples/sec	time cost: 2201.970821
[Epoch 004] validation: acc-top1=52.902062 acc-top5=77.932990 loss=2.017255
Epoch[005] Batch [0019]/[1203]	Speed: 66.975928 samples/sec	 accuracy=56.600000	 loss=1.744993	 lr=0.010000
Epoch[005] Batch [0039]/[1203]	Speed: 95.935478 samples/sec	 accuracy=56.937500	 loss=1.747677	 lr=0.010000
Epoch[005] Batch [0059]/[1203]	Speed: 238.187956 samples/sec	 accuracy=56.941667	 loss=1.744857	 lr=0.010000
Epoch[005] Batch [0079]/[1203]	Speed: 89.339920 samples/sec	 accuracy=57.112500	 loss=1.737143	 lr=0.010000
Epoch[005] Batch [0099]/[1203]	Speed: 107.597893 samples/sec	 accuracy=57.435000	 loss=1.730981	 lr=0.010000
Epoch[005] Batch [0119]/[1203]	Speed: 183.397450 samples/sec	 accuracy=57.475000	 loss=1.731669	 lr=0.010000
Epoch[005] Batch [0139]/[1203]	Speed: 92.731945 samples/sec	 accuracy=57.550000	 loss=1.730337	 lr=0.010000
Epoch[005] Batch [0159]/[1203]	Speed: 194.374906 samples/sec	 accuracy=57.340625	 loss=1.732352	 lr=0.010000
Epoch[005] Batch [0179]/[1203]	Speed: 100.226695 samples/sec	 accuracy=57.269444	 loss=1.736116	 lr=0.010000
Epoch[005] Batch [0199]/[1203]	Speed: 100.858442 samples/sec	 accuracy=57.412500	 loss=1.733917	 lr=0.010000
Epoch[005] Batch [0219]/[1203]	Speed: 201.975650 samples/sec	 accuracy=57.318182	 loss=1.738627	 lr=0.010000
Epoch[005] Batch [0239]/[1203]	Speed: 95.737784 samples/sec	 accuracy=57.237500	 loss=1.744042	 lr=0.010000
Epoch[005] Batch [0259]/[1203]	Speed: 119.074242 samples/sec	 accuracy=57.259615	 loss=1.741577	 lr=0.010000
Epoch[005] Batch [0279]/[1203]	Speed: 145.685304 samples/sec	 accuracy=57.257143	 loss=1.742931	 lr=0.010000
Epoch[005] Batch [0299]/[1203]	Speed: 95.460051 samples/sec	 accuracy=57.225000	 loss=1.745268	 lr=0.010000
Epoch[005] Batch [0319]/[1203]	Speed: 167.924994 samples/sec	 accuracy=57.229688	 loss=1.745593	 lr=0.010000
Epoch[005] Batch [0339]/[1203]	Speed: 101.562345 samples/sec	 accuracy=57.214706	 loss=1.745048	 lr=0.010000
Epoch[005] Batch [0359]/[1203]	Speed: 100.426401 samples/sec	 accuracy=57.240278	 loss=1.744718	 lr=0.010000
Epoch[005] Batch [0379]/[1203]	Speed: 195.315403 samples/sec	 accuracy=57.235526	 loss=1.745326	 lr=0.010000
Epoch[005] Batch [0399]/[1203]	Speed: 102.290424 samples/sec	 accuracy=57.232500	 loss=1.745801	 lr=0.010000
Epoch[005] Batch [0419]/[1203]	Speed: 129.584906 samples/sec	 accuracy=57.226190	 loss=1.746266	 lr=0.010000
Epoch[005] Batch [0439]/[1203]	Speed: 120.544986 samples/sec	 accuracy=57.231818	 loss=1.745519	 lr=0.010000
Epoch[005] Batch [0459]/[1203]	Speed: 99.376134 samples/sec	 accuracy=57.252174	 loss=1.745483	 lr=0.010000
Epoch[005] Batch [0479]/[1203]	Speed: 194.294477 samples/sec	 accuracy=57.201042	 loss=1.747701	 lr=0.010000
Epoch[005] Batch [0499]/[1203]	Speed: 94.239889 samples/sec	 accuracy=57.263000	 loss=1.746316	 lr=0.010000
Epoch[005] Batch [0519]/[1203]	Speed: 126.113829 samples/sec	 accuracy=57.250000	 loss=1.747442	 lr=0.010000
Epoch[005] Batch [0539]/[1203]	Speed: 124.262398 samples/sec	 accuracy=57.245370	 loss=1.746415	 lr=0.010000
Epoch[005] Batch [0559]/[1203]	Speed: 105.730273 samples/sec	 accuracy=57.216964	 loss=1.747590	 lr=0.010000
Epoch[005] Batch [0579]/[1203]	Speed: 144.749535 samples/sec	 accuracy=57.256897	 loss=1.746845	 lr=0.010000
Epoch[005] Batch [0599]/[1203]	Speed: 115.219590 samples/sec	 accuracy=57.257500	 loss=1.747553	 lr=0.010000
Epoch[005] Batch [0619]/[1203]	Speed: 109.197250 samples/sec	 accuracy=57.268548	 loss=1.747866	 lr=0.010000
Epoch[005] Batch [0639]/[1203]	Speed: 135.306495 samples/sec	 accuracy=57.259375	 loss=1.748451	 lr=0.010000
Epoch[005] Batch [0659]/[1203]	Speed: 117.990910 samples/sec	 accuracy=57.278030	 loss=1.748466	 lr=0.010000
Epoch[005] Batch [0679]/[1203]	Speed: 124.057851 samples/sec	 accuracy=57.246324	 loss=1.749401	 lr=0.010000
Epoch[005] Batch [0699]/[1203]	Speed: 113.476574 samples/sec	 accuracy=57.260714	 loss=1.748899	 lr=0.010000
Epoch[005] Batch [0719]/[1203]	Speed: 124.815247 samples/sec	 accuracy=57.264583	 loss=1.749308	 lr=0.010000
Epoch[005] Batch [0739]/[1203]	Speed: 126.226826 samples/sec	 accuracy=57.258784	 loss=1.749491	 lr=0.010000
Epoch[005] Batch [0759]/[1203]	Speed: 121.713961 samples/sec	 accuracy=57.265789	 loss=1.749487	 lr=0.010000
Epoch[005] Batch [0779]/[1203]	Speed: 107.956447 samples/sec	 accuracy=57.273718	 loss=1.749650	 lr=0.010000
Epoch[005] Batch [0799]/[1203]	Speed: 126.134979 samples/sec	 accuracy=57.264375	 loss=1.749581	 lr=0.010000
Epoch[005] Batch [0819]/[1203]	Speed: 126.555764 samples/sec	 accuracy=57.251220	 loss=1.749346	 lr=0.010000
Epoch[005] Batch [0839]/[1203]	Speed: 119.956137 samples/sec	 accuracy=57.257738	 loss=1.749148	 lr=0.010000
Epoch[005] Batch [0859]/[1203]	Speed: 105.025612 samples/sec	 accuracy=57.258721	 loss=1.749016	 lr=0.010000
Epoch[005] Batch [0879]/[1203]	Speed: 137.131878 samples/sec	 accuracy=57.272159	 loss=1.748400	 lr=0.010000
Epoch[005] Batch [0899]/[1203]	Speed: 116.835905 samples/sec	 accuracy=57.254444	 loss=1.749258	 lr=0.010000
Epoch[005] Batch [0919]/[1203]	Speed: 135.927012 samples/sec	 accuracy=57.255978	 loss=1.749238	 lr=0.010000
Epoch[005] Batch [0939]/[1203]	Speed: 111.530537 samples/sec	 accuracy=57.238830	 loss=1.750184	 lr=0.010000
Epoch[005] Batch [0959]/[1203]	Speed: 114.119572 samples/sec	 accuracy=57.228125	 loss=1.750950	 lr=0.010000
Epoch[005] Batch [0979]/[1203]	Speed: 135.193468 samples/sec	 accuracy=57.212755	 loss=1.751025	 lr=0.010000
Epoch[005] Batch [0999]/[1203]	Speed: 115.943435 samples/sec	 accuracy=57.212000	 loss=1.750439	 lr=0.010000
Epoch[005] Batch [1019]/[1203]	Speed: 108.544464 samples/sec	 accuracy=57.243137	 loss=1.749433	 lr=0.010000
Epoch[005] Batch [1039]/[1203]	Speed: 135.740254 samples/sec	 accuracy=57.252885	 loss=1.749378	 lr=0.010000
Epoch[005] Batch [1059]/[1203]	Speed: 112.810973 samples/sec	 accuracy=57.228774	 loss=1.750192	 lr=0.010000
Epoch[005] Batch [1079]/[1203]	Speed: 132.748053 samples/sec	 accuracy=57.222685	 loss=1.751159	 lr=0.010000
Epoch[005] Batch [1099]/[1203]	Speed: 122.226120 samples/sec	 accuracy=57.215909	 loss=1.751436	 lr=0.010000
Epoch[005] Batch [1119]/[1203]	Speed: 106.671376 samples/sec	 accuracy=57.229018	 loss=1.750811	 lr=0.010000
Epoch[005] Batch [1139]/[1203]	Speed: 143.297183 samples/sec	 accuracy=57.208333	 loss=1.751429	 lr=0.010000
Epoch[005] Batch [1159]/[1203]	Speed: 110.083734 samples/sec	 accuracy=57.206034	 loss=1.751613	 lr=0.010000
Epoch[005] Batch [1179]/[1203]	Speed: 153.540431 samples/sec	 accuracy=57.205932	 loss=1.751447	 lr=0.010000
Epoch[005] Batch [1199]/[1203]	Speed: 191.438499 samples/sec	 accuracy=57.209583	 loss=1.751304	 lr=0.010000
Batch [0019]/[0097]: acc-top1=52.975000 acc-top5=77.425000
Batch [0039]/[0097]: acc-top1=51.962500 acc-top5=76.512500
Batch [0059]/[0097]: acc-top1=54.283333 acc-top5=78.166667
Batch [0079]/[0097]: acc-top1=54.437500 acc-top5=78.437500
[Epoch 005] training: accuracy=57.213217	 loss=1.751449
[Epoch 005] speed: 120 samples/sec	time cost: 2198.405145
[Epoch 005] validation: acc-top1=54.097938 acc-top5=78.752577 loss=1.950310
Epoch[006] Batch [0019]/[1203]	Speed: 68.005732 samples/sec	 accuracy=59.625000	 loss=1.688434	 lr=0.010000
Epoch[006] Batch [0039]/[1203]	Speed: 96.055202 samples/sec	 accuracy=60.312500	 loss=1.641946	 lr=0.010000
Epoch[006] Batch [0059]/[1203]	Speed: 227.224430 samples/sec	 accuracy=60.183333	 loss=1.629282	 lr=0.010000
Epoch[006] Batch [0079]/[1203]	Speed: 95.338955 samples/sec	 accuracy=59.975000	 loss=1.642798	 lr=0.010000
Epoch[006] Batch [0099]/[1203]	Speed: 103.936706 samples/sec	 accuracy=59.785000	 loss=1.646703	 lr=0.010000
Epoch[006] Batch [0119]/[1203]	Speed: 181.928940 samples/sec	 accuracy=59.700000	 loss=1.653805	 lr=0.010000
Epoch[006] Batch [0139]/[1203]	Speed: 90.073749 samples/sec	 accuracy=59.635714	 loss=1.650803	 lr=0.010000
Epoch[006] Batch [0159]/[1203]	Speed: 221.826271 samples/sec	 accuracy=59.662500	 loss=1.653193	 lr=0.010000
Epoch[006] Batch [0179]/[1203]	Speed: 93.764596 samples/sec	 accuracy=59.641667	 loss=1.654576	 lr=0.010000
Epoch[006] Batch [0199]/[1203]	Speed: 110.180840 samples/sec	 accuracy=59.532500	 loss=1.659019	 lr=0.010000
Epoch[006] Batch [0219]/[1203]	Speed: 169.561317 samples/sec	 accuracy=59.440909	 loss=1.657829	 lr=0.010000
Epoch[006] Batch [0239]/[1203]	Speed: 98.183236 samples/sec	 accuracy=59.356250	 loss=1.658497	 lr=0.010000
Epoch[006] Batch [0259]/[1203]	Speed: 139.866954 samples/sec	 accuracy=59.284615	 loss=1.661971	 lr=0.010000
Epoch[006] Batch [0279]/[1203]	Speed: 110.717629 samples/sec	 accuracy=59.228571	 loss=1.662227	 lr=0.010000
Epoch[006] Batch [0299]/[1203]	Speed: 108.378648 samples/sec	 accuracy=59.255000	 loss=1.664605	 lr=0.010000
Epoch[006] Batch [0319]/[1203]	Speed: 150.825626 samples/sec	 accuracy=59.201563	 loss=1.665291	 lr=0.010000
Epoch[006] Batch [0339]/[1203]	Speed: 104.669757 samples/sec	 accuracy=59.163235	 loss=1.665880	 lr=0.010000
Epoch[006] Batch [0359]/[1203]	Speed: 139.193110 samples/sec	 accuracy=59.126389	 loss=1.665863	 lr=0.010000
Epoch[006] Batch [0379]/[1203]	Speed: 112.505373 samples/sec	 accuracy=59.111842	 loss=1.667675	 lr=0.010000
Epoch[006] Batch [0399]/[1203]	Speed: 103.203798 samples/sec	 accuracy=59.145000	 loss=1.666506	 lr=0.010000
Epoch[006] Batch [0419]/[1203]	Speed: 158.822197 samples/sec	 accuracy=59.096429	 loss=1.669242	 lr=0.010000
Epoch[006] Batch [0439]/[1203]	Speed: 106.902798 samples/sec	 accuracy=59.043182	 loss=1.670256	 lr=0.010000
Epoch[006] Batch [0459]/[1203]	Speed: 109.386110 samples/sec	 accuracy=58.984783	 loss=1.673500	 lr=0.010000
Epoch[006] Batch [0479]/[1203]	Speed: 142.560707 samples/sec	 accuracy=58.945833	 loss=1.675226	 lr=0.010000
Epoch[006] Batch [0499]/[1203]	Speed: 112.699702 samples/sec	 accuracy=58.977000	 loss=1.675343	 lr=0.010000
Epoch[006] Batch [0519]/[1203]	Speed: 150.908984 samples/sec	 accuracy=58.994231	 loss=1.673426	 lr=0.010000
Epoch[006] Batch [0539]/[1203]	Speed: 105.691259 samples/sec	 accuracy=58.997222	 loss=1.674135	 lr=0.010000
Epoch[006] Batch [0559]/[1203]	Speed: 109.270864 samples/sec	 accuracy=58.979464	 loss=1.672783	 lr=0.010000
Epoch[006] Batch [0579]/[1203]	Speed: 137.674032 samples/sec	 accuracy=58.981897	 loss=1.672449	 lr=0.010000
Epoch[006] Batch [0599]/[1203]	Speed: 111.497556 samples/sec	 accuracy=58.994167	 loss=1.672446	 lr=0.010000
Epoch[006] Batch [0619]/[1203]	Speed: 133.379741 samples/sec	 accuracy=59.025806	 loss=1.671358	 lr=0.010000
Epoch[006] Batch [0639]/[1203]	Speed: 120.324535 samples/sec	 accuracy=58.959375	 loss=1.674650	 lr=0.010000
Epoch[006] Batch [0659]/[1203]	Speed: 110.881992 samples/sec	 accuracy=58.946970	 loss=1.675063	 lr=0.010000
Epoch[006] Batch [0679]/[1203]	Speed: 135.144717 samples/sec	 accuracy=58.929412	 loss=1.675208	 lr=0.010000
Epoch[006] Batch [0699]/[1203]	Speed: 115.678073 samples/sec	 accuracy=58.917857	 loss=1.676102	 lr=0.010000
Epoch[006] Batch [0719]/[1203]	Speed: 113.970085 samples/sec	 accuracy=58.943056	 loss=1.675568	 lr=0.010000
Epoch[006] Batch [0739]/[1203]	Speed: 125.008842 samples/sec	 accuracy=58.926351	 loss=1.675792	 lr=0.010000
Epoch[006] Batch [0759]/[1203]	Speed: 125.110305 samples/sec	 accuracy=58.900658	 loss=1.676190	 lr=0.010000
Epoch[006] Batch [0779]/[1203]	Speed: 113.796923 samples/sec	 accuracy=58.897436	 loss=1.676427	 lr=0.010000
Epoch[006] Batch [0799]/[1203]	Speed: 116.002099 samples/sec	 accuracy=58.843125	 loss=1.678263	 lr=0.010000
Epoch[006] Batch [0819]/[1203]	Speed: 121.862251 samples/sec	 accuracy=58.859756	 loss=1.678556	 lr=0.010000
Epoch[006] Batch [0839]/[1203]	Speed: 141.534102 samples/sec	 accuracy=58.864286	 loss=1.678727	 lr=0.010000
Epoch[006] Batch [0859]/[1203]	Speed: 108.495379 samples/sec	 accuracy=58.842442	 loss=1.679512	 lr=0.010000
Epoch[006] Batch [0879]/[1203]	Speed: 117.460659 samples/sec	 accuracy=58.850000	 loss=1.679987	 lr=0.010000
Epoch[006] Batch [0899]/[1203]	Speed: 119.731145 samples/sec	 accuracy=58.825556	 loss=1.680927	 lr=0.010000
Epoch[006] Batch [0919]/[1203]	Speed: 131.944428 samples/sec	 accuracy=58.788043	 loss=1.683148	 lr=0.010000
Epoch[006] Batch [0939]/[1203]	Speed: 106.690512 samples/sec	 accuracy=58.764362	 loss=1.683693	 lr=0.010000
Epoch[006] Batch [0959]/[1203]	Speed: 125.328479 samples/sec	 accuracy=58.734375	 loss=1.684221	 lr=0.010000
Epoch[006] Batch [0979]/[1203]	Speed: 118.132998 samples/sec	 accuracy=58.744388	 loss=1.684086	 lr=0.010000
Epoch[006] Batch [0999]/[1203]	Speed: 127.205851 samples/sec	 accuracy=58.710500	 loss=1.684652	 lr=0.010000
Epoch[006] Batch [1019]/[1203]	Speed: 130.627708 samples/sec	 accuracy=58.701471	 loss=1.684780	 lr=0.010000
Epoch[006] Batch [1039]/[1203]	Speed: 108.681174 samples/sec	 accuracy=58.698077	 loss=1.685371	 lr=0.010000
Epoch[006] Batch [1059]/[1203]	Speed: 119.842780 samples/sec	 accuracy=58.699057	 loss=1.685288	 lr=0.010000
Epoch[006] Batch [1079]/[1203]	Speed: 114.460382 samples/sec	 accuracy=58.669907	 loss=1.686174	 lr=0.010000
Epoch[006] Batch [1099]/[1203]	Speed: 133.439709 samples/sec	 accuracy=58.658182	 loss=1.686648	 lr=0.010000
Epoch[006] Batch [1119]/[1203]	Speed: 120.502671 samples/sec	 accuracy=58.633482	 loss=1.687540	 lr=0.010000
Epoch[006] Batch [1139]/[1203]	Speed: 113.685815 samples/sec	 accuracy=58.611842	 loss=1.688445	 lr=0.010000
Epoch[006] Batch [1159]/[1203]	Speed: 129.100611 samples/sec	 accuracy=58.606897	 loss=1.688817	 lr=0.010000
Epoch[006] Batch [1179]/[1203]	Speed: 118.403938 samples/sec	 accuracy=58.602119	 loss=1.689113	 lr=0.010000
Epoch[006] Batch [1199]/[1203]	Speed: 259.079938 samples/sec	 accuracy=58.596667	 loss=1.689251	 lr=0.010000
Batch [0019]/[0097]: acc-top1=55.900000 acc-top5=79.625000
Batch [0039]/[0097]: acc-top1=53.650000 acc-top5=77.675000
Batch [0059]/[0097]: acc-top1=55.250000 acc-top5=78.608333
Batch [0079]/[0097]: acc-top1=55.293750 acc-top5=78.600000
[Epoch 006] training: accuracy=58.605569	 loss=1.688890
[Epoch 006] speed: 120 samples/sec	time cost: 2200.021900
[Epoch 006] validation: acc-top1=54.134021 acc-top5=78.335052 loss=1.952143
Epoch[007] Batch [0019]/[1203]	Speed: 67.682836 samples/sec	 accuracy=60.950000	 loss=1.586777	 lr=0.010000
Epoch[007] Batch [0039]/[1203]	Speed: 98.181587 samples/sec	 accuracy=61.037500	 loss=1.583138	 lr=0.010000
Epoch[007] Batch [0059]/[1203]	Speed: 208.258430 samples/sec	 accuracy=60.991667	 loss=1.570106	 lr=0.010000
Epoch[007] Batch [0079]/[1203]	Speed: 97.079828 samples/sec	 accuracy=60.906250	 loss=1.580592	 lr=0.010000
Epoch[007] Batch [0099]/[1203]	Speed: 118.822383 samples/sec	 accuracy=60.835000	 loss=1.587291	 lr=0.010000
Epoch[007] Batch [0119]/[1203]	Speed: 128.646808 samples/sec	 accuracy=61.045833	 loss=1.584397	 lr=0.010000
Epoch[007] Batch [0139]/[1203]	Speed: 106.084185 samples/sec	 accuracy=60.928571	 loss=1.587502	 lr=0.010000
Epoch[007] Batch [0159]/[1203]	Speed: 176.675927 samples/sec	 accuracy=60.793750	 loss=1.593507	 lr=0.010000
Epoch[007] Batch [0179]/[1203]	Speed: 98.134771 samples/sec	 accuracy=60.730556	 loss=1.598922	 lr=0.010000
Epoch[007] Batch [0199]/[1203]	Speed: 126.464990 samples/sec	 accuracy=60.682500	 loss=1.599170	 lr=0.010000
Epoch[007] Batch [0219]/[1203]	Speed: 120.885562 samples/sec	 accuracy=60.604545	 loss=1.599140	 lr=0.010000
Epoch[007] Batch [0239]/[1203]	Speed: 109.603512 samples/sec	 accuracy=60.560417	 loss=1.602798	 lr=0.010000
Epoch[007] Batch [0259]/[1203]	Speed: 155.413190 samples/sec	 accuracy=60.446154	 loss=1.606444	 lr=0.010000
Epoch[007] Batch [0279]/[1203]	Speed: 104.130297 samples/sec	 accuracy=60.432143	 loss=1.606510	 lr=0.010000
Epoch[007] Batch [0299]/[1203]	Speed: 114.632514 samples/sec	 accuracy=60.346667	 loss=1.611422	 lr=0.010000
Epoch[007] Batch [0319]/[1203]	Speed: 142.322966 samples/sec	 accuracy=60.253125	 loss=1.614786	 lr=0.010000
Epoch[007] Batch [0339]/[1203]	Speed: 105.829591 samples/sec	 accuracy=60.167647	 loss=1.619235	 lr=0.010000
Epoch[007] Batch [0359]/[1203]	Speed: 156.289886 samples/sec	 accuracy=60.136111	 loss=1.618975	 lr=0.010000
Epoch[007] Batch [0379]/[1203]	Speed: 100.102905 samples/sec	 accuracy=60.175000	 loss=1.618589	 lr=0.010000
Epoch[007] Batch [0399]/[1203]	Speed: 115.080915 samples/sec	 accuracy=60.191250	 loss=1.618386	 lr=0.010000
Epoch[007] Batch [0419]/[1203]	Speed: 146.883904 samples/sec	 accuracy=60.161905	 loss=1.618355	 lr=0.010000
Epoch[007] Batch [0439]/[1203]	Speed: 113.393662 samples/sec	 accuracy=60.167045	 loss=1.618470	 lr=0.010000
Epoch[007] Batch [0459]/[1203]	Speed: 140.853093 samples/sec	 accuracy=60.110870	 loss=1.620011	 lr=0.010000
Epoch[007] Batch [0479]/[1203]	Speed: 112.846999 samples/sec	 accuracy=60.047917	 loss=1.621957	 lr=0.010000
Epoch[007] Batch [0499]/[1203]	Speed: 100.484469 samples/sec	 accuracy=60.008000	 loss=1.621638	 lr=0.010000
Epoch[007] Batch [0519]/[1203]	Speed: 151.136461 samples/sec	 accuracy=59.944231	 loss=1.624322	 lr=0.010000
Epoch[007] Batch [0539]/[1203]	Speed: 114.245790 samples/sec	 accuracy=59.933333	 loss=1.624386	 lr=0.010000
Epoch[007] Batch [0559]/[1203]	Speed: 127.583096 samples/sec	 accuracy=59.916071	 loss=1.624338	 lr=0.010000
Epoch[007] Batch [0579]/[1203]	Speed: 112.451255 samples/sec	 accuracy=59.956897	 loss=1.622199	 lr=0.010000
Epoch[007] Batch [0599]/[1203]	Speed: 121.507963 samples/sec	 accuracy=59.938333	 loss=1.622727	 lr=0.010000
Epoch[007] Batch [0619]/[1203]	Speed: 137.408188 samples/sec	 accuracy=59.915323	 loss=1.623639	 lr=0.010000
Epoch[007] Batch [0639]/[1203]	Speed: 111.955497 samples/sec	 accuracy=59.875781	 loss=1.624774	 lr=0.010000
Epoch[007] Batch [0659]/[1203]	Speed: 117.687318 samples/sec	 accuracy=59.894697	 loss=1.624063	 lr=0.010000
Epoch[007] Batch [0679]/[1203]	Speed: 122.293856 samples/sec	 accuracy=59.875735	 loss=1.624286	 lr=0.010000
Epoch[007] Batch [0699]/[1203]	Speed: 122.409294 samples/sec	 accuracy=59.879286	 loss=1.624652	 lr=0.010000
Epoch[007] Batch [0719]/[1203]	Speed: 120.198294 samples/sec	 accuracy=59.881944	 loss=1.625180	 lr=0.010000
Epoch[007] Batch [0739]/[1203]	Speed: 127.247606 samples/sec	 accuracy=59.838514	 loss=1.627167	 lr=0.010000
Epoch[007] Batch [0759]/[1203]	Speed: 118.548824 samples/sec	 accuracy=59.824342	 loss=1.626918	 lr=0.010000
Epoch[007] Batch [0779]/[1203]	Speed: 121.288326 samples/sec	 accuracy=59.832051	 loss=1.627421	 lr=0.010000
Epoch[007] Batch [0799]/[1203]	Speed: 125.257467 samples/sec	 accuracy=59.824375	 loss=1.628481	 lr=0.010000
Epoch[007] Batch [0819]/[1203]	Speed: 117.713527 samples/sec	 accuracy=59.814024	 loss=1.629492	 lr=0.010000
Epoch[007] Batch [0839]/[1203]	Speed: 130.808606 samples/sec	 accuracy=59.829167	 loss=1.629646	 lr=0.010000
Epoch[007] Batch [0859]/[1203]	Speed: 116.697015 samples/sec	 accuracy=59.790116	 loss=1.631153	 lr=0.010000
Epoch[007] Batch [0879]/[1203]	Speed: 112.764784 samples/sec	 accuracy=59.784091	 loss=1.631787	 lr=0.010000
Epoch[007] Batch [0899]/[1203]	Speed: 116.863535 samples/sec	 accuracy=59.757778	 loss=1.632478	 lr=0.010000
Epoch[007] Batch [0919]/[1203]	Speed: 130.885016 samples/sec	 accuracy=59.773370	 loss=1.632530	 lr=0.010000
Epoch[007] Batch [0939]/[1203]	Speed: 123.252332 samples/sec	 accuracy=59.765957	 loss=1.632916	 lr=0.010000
Epoch[007] Batch [0959]/[1203]	Speed: 127.117592 samples/sec	 accuracy=59.747917	 loss=1.633275	 lr=0.010000
Epoch[007] Batch [0979]/[1203]	Speed: 107.440639 samples/sec	 accuracy=59.732653	 loss=1.633922	 lr=0.010000
Epoch[007] Batch [0999]/[1203]	Speed: 119.517927 samples/sec	 accuracy=59.744000	 loss=1.633667	 lr=0.010000
Epoch[007] Batch [1019]/[1203]	Speed: 146.701277 samples/sec	 accuracy=59.731863	 loss=1.633877	 lr=0.010000
Epoch[007] Batch [1039]/[1203]	Speed: 121.084733 samples/sec	 accuracy=59.721154	 loss=1.634583	 lr=0.010000
Epoch[007] Batch [1059]/[1203]	Speed: 112.634169 samples/sec	 accuracy=59.709434	 loss=1.634848	 lr=0.010000
Epoch[007] Batch [1079]/[1203]	Speed: 122.933109 samples/sec	 accuracy=59.685185	 loss=1.636058	 lr=0.010000
Epoch[007] Batch [1099]/[1203]	Speed: 109.301206 samples/sec	 accuracy=59.681818	 loss=1.635919	 lr=0.010000
Epoch[007] Batch [1119]/[1203]	Speed: 154.641985 samples/sec	 accuracy=59.651339	 loss=1.637305	 lr=0.010000
Epoch[007] Batch [1139]/[1203]	Speed: 105.549519 samples/sec	 accuracy=59.643421	 loss=1.637220	 lr=0.010000
Epoch[007] Batch [1159]/[1203]	Speed: 137.648155 samples/sec	 accuracy=59.617672	 loss=1.638384	 lr=0.010000
Epoch[007] Batch [1179]/[1203]	Speed: 120.514619 samples/sec	 accuracy=59.650000	 loss=1.637284	 lr=0.010000
Epoch[007] Batch [1199]/[1203]	Speed: 241.637307 samples/sec	 accuracy=59.645417	 loss=1.637389	 lr=0.010000
Batch [0019]/[0097]: acc-top1=55.250000 acc-top5=79.175000
Batch [0039]/[0097]: acc-top1=54.187500 acc-top5=78.537500
Batch [0059]/[0097]: acc-top1=56.516667 acc-top5=80.325000
Batch [0079]/[0097]: acc-top1=56.575000 acc-top5=80.237500
[Epoch 007] training: accuracy=59.640067	 loss=1.637652
[Epoch 007] speed: 120 samples/sec	time cost: 2186.108268
[Epoch 007] validation: acc-top1=55.927835 acc-top5=79.974227 loss=1.901246
Epoch[008] Batch [0019]/[1203]	Speed: 68.207727 samples/sec	 accuracy=60.725000	 loss=1.578834	 lr=0.010000
Epoch[008] Batch [0039]/[1203]	Speed: 101.230228 samples/sec	 accuracy=60.900000	 loss=1.573958	 lr=0.010000
Epoch[008] Batch [0059]/[1203]	Speed: 216.357076 samples/sec	 accuracy=61.416667	 loss=1.568272	 lr=0.010000
Epoch[008] Batch [0079]/[1203]	Speed: 93.974952 samples/sec	 accuracy=61.475000	 loss=1.555174	 lr=0.010000
Epoch[008] Batch [0099]/[1203]	Speed: 114.319837 samples/sec	 accuracy=61.465000	 loss=1.553128	 lr=0.010000
Epoch[008] Batch [0119]/[1203]	Speed: 152.193557 samples/sec	 accuracy=61.441667	 loss=1.551369	 lr=0.010000
Epoch[008] Batch [0139]/[1203]	Speed: 99.105035 samples/sec	 accuracy=61.328571	 loss=1.557849	 lr=0.010000
Epoch[008] Batch [0159]/[1203]	Speed: 187.456689 samples/sec	 accuracy=61.287500	 loss=1.561389	 lr=0.010000
Epoch[008] Batch [0179]/[1203]	Speed: 107.572030 samples/sec	 accuracy=61.158333	 loss=1.567895	 lr=0.010000
Epoch[008] Batch [0199]/[1203]	Speed: 100.727871 samples/sec	 accuracy=61.277500	 loss=1.566609	 lr=0.010000
Epoch[008] Batch [0219]/[1203]	Speed: 160.395341 samples/sec	 accuracy=61.218182	 loss=1.568357	 lr=0.010000
Epoch[008] Batch [0239]/[1203]	Speed: 108.440214 samples/sec	 accuracy=61.164583	 loss=1.572142	 lr=0.010000
Epoch[008] Batch [0259]/[1203]	Speed: 130.119315 samples/sec	 accuracy=61.128846	 loss=1.576434	 lr=0.010000
Epoch[008] Batch [0279]/[1203]	Speed: 110.710800 samples/sec	 accuracy=61.119643	 loss=1.575109	 lr=0.010000
Epoch[008] Batch [0299]/[1203]	Speed: 125.615277 samples/sec	 accuracy=61.086667	 loss=1.574798	 lr=0.010000
Epoch[008] Batch [0319]/[1203]	Speed: 138.914694 samples/sec	 accuracy=61.046875	 loss=1.575878	 lr=0.010000
Epoch[008] Batch [0339]/[1203]	Speed: 102.024403 samples/sec	 accuracy=61.033824	 loss=1.577021	 lr=0.010000
Epoch[008] Batch [0359]/[1203]	Speed: 135.842404 samples/sec	 accuracy=60.956944	 loss=1.581000	 lr=0.010000
Epoch[008] Batch [0379]/[1203]	Speed: 114.832386 samples/sec	 accuracy=61.011842	 loss=1.581153	 lr=0.010000
Epoch[008] Batch [0399]/[1203]	Speed: 124.446476 samples/sec	 accuracy=61.005000	 loss=1.581020	 lr=0.010000
Epoch[008] Batch [0419]/[1203]	Speed: 115.700091 samples/sec	 accuracy=61.033333	 loss=1.578392	 lr=0.010000
Epoch[008] Batch [0439]/[1203]	Speed: 118.592277 samples/sec	 accuracy=61.082955	 loss=1.576828	 lr=0.010000
Epoch[008] Batch [0459]/[1203]	Speed: 140.487870 samples/sec	 accuracy=61.063043	 loss=1.577304	 lr=0.010000
Epoch[008] Batch [0479]/[1203]	Speed: 104.797607 samples/sec	 accuracy=61.090625	 loss=1.577156	 lr=0.010000
Epoch[008] Batch [0499]/[1203]	Speed: 148.244983 samples/sec	 accuracy=61.067000	 loss=1.578798	 lr=0.010000
Epoch[008] Batch [0519]/[1203]	Speed: 118.040866 samples/sec	 accuracy=61.044231	 loss=1.579455	 lr=0.010000
Epoch[008] Batch [0539]/[1203]	Speed: 116.705844 samples/sec	 accuracy=61.032407	 loss=1.580216	 lr=0.010000
Epoch[008] Batch [0559]/[1203]	Speed: 127.130713 samples/sec	 accuracy=61.054464	 loss=1.578667	 lr=0.010000
Epoch[008] Batch [0579]/[1203]	Speed: 109.808861 samples/sec	 accuracy=61.056897	 loss=1.578522	 lr=0.010000
Epoch[008] Batch [0599]/[1203]	Speed: 115.536152 samples/sec	 accuracy=61.020000	 loss=1.580238	 lr=0.010000
Epoch[008] Batch [0619]/[1203]	Speed: 149.355642 samples/sec	 accuracy=61.004032	 loss=1.581998	 lr=0.010000
Epoch[008] Batch [0639]/[1203]	Speed: 105.900589 samples/sec	 accuracy=60.963281	 loss=1.583256	 lr=0.010000
Epoch[008] Batch [0659]/[1203]	Speed: 148.944184 samples/sec	 accuracy=60.946212	 loss=1.584338	 lr=0.010000
Epoch[008] Batch [0679]/[1203]	Speed: 104.234432 samples/sec	 accuracy=60.913971	 loss=1.585243	 lr=0.010000
Epoch[008] Batch [0699]/[1203]	Speed: 114.591919 samples/sec	 accuracy=60.942143	 loss=1.584686	 lr=0.010000
Epoch[008] Batch [0719]/[1203]	Speed: 142.723303 samples/sec	 accuracy=60.954861	 loss=1.584153	 lr=0.010000
Epoch[008] Batch [0739]/[1203]	Speed: 111.254836 samples/sec	 accuracy=60.925000	 loss=1.585169	 lr=0.010000
Epoch[008] Batch [0759]/[1203]	Speed: 123.353436 samples/sec	 accuracy=60.921711	 loss=1.584573	 lr=0.010000
Epoch[008] Batch [0779]/[1203]	Speed: 117.362742 samples/sec	 accuracy=60.915385	 loss=1.584913	 lr=0.010000
Epoch[008] Batch [0799]/[1203]	Speed: 125.461271 samples/sec	 accuracy=60.905625	 loss=1.584817	 lr=0.010000
Epoch[008] Batch [0819]/[1203]	Speed: 124.763121 samples/sec	 accuracy=60.905488	 loss=1.584965	 lr=0.010000
Epoch[008] Batch [0839]/[1203]	Speed: 118.393492 samples/sec	 accuracy=60.870833	 loss=1.584741	 lr=0.010000
Epoch[008] Batch [0859]/[1203]	Speed: 114.756576 samples/sec	 accuracy=60.858140	 loss=1.585769	 lr=0.010000
Epoch[008] Batch [0879]/[1203]	Speed: 121.206576 samples/sec	 accuracy=60.859659	 loss=1.586073	 lr=0.010000
Epoch[008] Batch [0899]/[1203]	Speed: 137.712872 samples/sec	 accuracy=60.855556	 loss=1.586734	 lr=0.010000
Epoch[008] Batch [0919]/[1203]	Speed: 123.250359 samples/sec	 accuracy=60.830978	 loss=1.587744	 lr=0.010000
Epoch[008] Batch [0939]/[1203]	Speed: 121.131489 samples/sec	 accuracy=60.829787	 loss=1.587932	 lr=0.010000
Epoch[008] Batch [0959]/[1203]	Speed: 113.158381 samples/sec	 accuracy=60.805729	 loss=1.588684	 lr=0.010000
Epoch[008] Batch [0979]/[1203]	Speed: 120.579632 samples/sec	 accuracy=60.812245	 loss=1.588323	 lr=0.010000
Epoch[008] Batch [0999]/[1203]	Speed: 142.926564 samples/sec	 accuracy=60.817000	 loss=1.587114	 lr=0.010000
Epoch[008] Batch [1019]/[1203]	Speed: 112.181662 samples/sec	 accuracy=60.823529	 loss=1.586917	 lr=0.010000
Epoch[008] Batch [1039]/[1203]	Speed: 114.704399 samples/sec	 accuracy=60.820192	 loss=1.587540	 lr=0.010000
Epoch[008] Batch [1059]/[1203]	Speed: 138.015354 samples/sec	 accuracy=60.789151	 loss=1.589176	 lr=0.010000
Epoch[008] Batch [1079]/[1203]	Speed: 104.612576 samples/sec	 accuracy=60.795370	 loss=1.589758	 lr=0.010000
Epoch[008] Batch [1099]/[1203]	Speed: 146.569906 samples/sec	 accuracy=60.775000	 loss=1.590613	 lr=0.010000
Epoch[008] Batch [1119]/[1203]	Speed: 111.059731 samples/sec	 accuracy=60.769643	 loss=1.590301	 lr=0.010000
Epoch[008] Batch [1139]/[1203]	Speed: 102.838117 samples/sec	 accuracy=60.740789	 loss=1.590911	 lr=0.010000
Epoch[008] Batch [1159]/[1203]	Speed: 147.179795 samples/sec	 accuracy=60.728448	 loss=1.591171	 lr=0.010000
Epoch[008] Batch [1179]/[1203]	Speed: 122.471084 samples/sec	 accuracy=60.710169	 loss=1.591536	 lr=0.010000
Epoch[008] Batch [1199]/[1203]	Speed: 286.586820 samples/sec	 accuracy=60.717500	 loss=1.591575	 lr=0.010000
Batch [0019]/[0097]: acc-top1=54.875000 acc-top5=79.375000
Batch [0039]/[0097]: acc-top1=54.612500 acc-top5=79.037500
Batch [0059]/[0097]: acc-top1=56.916667 acc-top5=80.550000
Batch [0079]/[0097]: acc-top1=56.850000 acc-top5=80.312500
[Epoch 008] training: accuracy=60.721945	 loss=1.591332
[Epoch 008] speed: 121 samples/sec	time cost: 2178.606521
[Epoch 008] validation: acc-top1=56.164948 acc-top5=80.201031 loss=1.866516
Epoch[009] Batch [0019]/[1203]	Speed: 69.272583 samples/sec	 accuracy=61.875000	 loss=1.525377	 lr=0.010000
Epoch[009] Batch [0039]/[1203]	Speed: 98.286571 samples/sec	 accuracy=61.500000	 loss=1.522171	 lr=0.010000
Epoch[009] Batch [0059]/[1203]	Speed: 234.251525 samples/sec	 accuracy=61.191667	 loss=1.546628	 lr=0.010000
Epoch[009] Batch [0079]/[1203]	Speed: 94.967764 samples/sec	 accuracy=61.362500	 loss=1.553795	 lr=0.010000
Epoch[009] Batch [0099]/[1203]	Speed: 99.079836 samples/sec	 accuracy=61.555000	 loss=1.547392	 lr=0.010000
Epoch[009] Batch [0119]/[1203]	Speed: 201.280481 samples/sec	 accuracy=61.625000	 loss=1.543869	 lr=0.010000
Epoch[009] Batch [0139]/[1203]	Speed: 98.314827 samples/sec	 accuracy=61.678571	 loss=1.539090	 lr=0.010000
Epoch[009] Batch [0159]/[1203]	Speed: 171.492277 samples/sec	 accuracy=61.762500	 loss=1.536867	 lr=0.010000
Epoch[009] Batch [0179]/[1203]	Speed: 104.068309 samples/sec	 accuracy=61.705556	 loss=1.535896	 lr=0.010000
Epoch[009] Batch [0199]/[1203]	Speed: 104.313217 samples/sec	 accuracy=61.692500	 loss=1.536838	 lr=0.010000
Epoch[009] Batch [0219]/[1203]	Speed: 152.228472 samples/sec	 accuracy=61.540909	 loss=1.541186	 lr=0.010000
Epoch[009] Batch [0239]/[1203]	Speed: 114.768283 samples/sec	 accuracy=61.570833	 loss=1.541116	 lr=0.010000
Epoch[009] Batch [0259]/[1203]	Speed: 117.635161 samples/sec	 accuracy=61.648077	 loss=1.538797	 lr=0.010000
Epoch[009] Batch [0279]/[1203]	Speed: 126.395244 samples/sec	 accuracy=61.707143	 loss=1.536959	 lr=0.010000
Epoch[009] Batch [0299]/[1203]	Speed: 108.766212 samples/sec	 accuracy=61.723333	 loss=1.534806	 lr=0.010000
Epoch[009] Batch [0319]/[1203]	Speed: 150.679405 samples/sec	 accuracy=61.759375	 loss=1.534841	 lr=0.010000
Epoch[009] Batch [0339]/[1203]	Speed: 120.234017 samples/sec	 accuracy=61.772059	 loss=1.533403	 lr=0.010000
Epoch[009] Batch [0359]/[1203]	Speed: 107.472080 samples/sec	 accuracy=61.750000	 loss=1.536984	 lr=0.010000
Epoch[009] Batch [0379]/[1203]	Speed: 127.972115 samples/sec	 accuracy=61.777632	 loss=1.535440	 lr=0.010000
Epoch[009] Batch [0399]/[1203]	Speed: 114.526766 samples/sec	 accuracy=61.700000	 loss=1.538679	 lr=0.010000
Epoch[009] Batch [0419]/[1203]	Speed: 138.913449 samples/sec	 accuracy=61.728571	 loss=1.540104	 lr=0.010000
Epoch[009] Batch [0439]/[1203]	Speed: 108.397877 samples/sec	 accuracy=61.790909	 loss=1.538429	 lr=0.010000
Epoch[009] Batch [0459]/[1203]	Speed: 123.776858 samples/sec	 accuracy=61.779348	 loss=1.538381	 lr=0.010000
Epoch[009] Batch [0479]/[1203]	Speed: 119.862651 samples/sec	 accuracy=61.739583	 loss=1.539168	 lr=0.010000
Epoch[009] Batch [0499]/[1203]	Speed: 140.271757 samples/sec	 accuracy=61.719000	 loss=1.539373	 lr=0.010000
Epoch[009] Batch [0519]/[1203]	Speed: 120.162442 samples/sec	 accuracy=61.693269	 loss=1.540014	 lr=0.010000
Epoch[009] Batch [0539]/[1203]	Speed: 110.321625 samples/sec	 accuracy=61.727778	 loss=1.539628	 lr=0.010000
Epoch[009] Batch [0559]/[1203]	Speed: 120.994028 samples/sec	 accuracy=61.719643	 loss=1.542159	 lr=0.010000
Epoch[009] Batch [0579]/[1203]	Speed: 120.451123 samples/sec	 accuracy=61.675862	 loss=1.543071	 lr=0.010000
Epoch[009] Batch [0599]/[1203]	Speed: 116.668849 samples/sec	 accuracy=61.705000	 loss=1.541749	 lr=0.010000
Epoch[009] Batch [0619]/[1203]	Speed: 139.407362 samples/sec	 accuracy=61.664516	 loss=1.543459	 lr=0.010000
Epoch[009] Batch [0639]/[1203]	Speed: 112.535950 samples/sec	 accuracy=61.654688	 loss=1.544727	 lr=0.010000
Epoch[009] Batch [0659]/[1203]	Speed: 127.600064 samples/sec	 accuracy=61.640152	 loss=1.545135	 lr=0.010000
Epoch[009] Batch [0679]/[1203]	Speed: 115.545367 samples/sec	 accuracy=61.626471	 loss=1.545710	 lr=0.010000
Epoch[009] Batch [0699]/[1203]	Speed: 116.405701 samples/sec	 accuracy=61.631429	 loss=1.545636	 lr=0.010000
Epoch[009] Batch [0719]/[1203]	Speed: 130.895623 samples/sec	 accuracy=61.638194	 loss=1.544768	 lr=0.010000
Epoch[009] Batch [0739]/[1203]	Speed: 117.015188 samples/sec	 accuracy=61.592568	 loss=1.547853	 lr=0.010000
Epoch[009] Batch [0759]/[1203]	Speed: 120.585519 samples/sec	 accuracy=61.571711	 loss=1.548952	 lr=0.010000
Epoch[009] Batch [0779]/[1203]	Speed: 122.209454 samples/sec	 accuracy=61.541026	 loss=1.549888	 lr=0.010000
Epoch[009] Batch [0799]/[1203]	Speed: 122.192609 samples/sec	 accuracy=61.533750	 loss=1.549981	 lr=0.010000
Epoch[009] Batch [0819]/[1203]	Speed: 131.969544 samples/sec	 accuracy=61.551829	 loss=1.549534	 lr=0.010000
Epoch[009] Batch [0839]/[1203]	Speed: 109.462212 samples/sec	 accuracy=61.572619	 loss=1.548768	 lr=0.010000
Epoch[009] Batch [0859]/[1203]	Speed: 112.698350 samples/sec	 accuracy=61.569186	 loss=1.549782	 lr=0.010000
Epoch[009] Batch [0879]/[1203]	Speed: 130.751760 samples/sec	 accuracy=61.570455	 loss=1.550363	 lr=0.010000
Epoch[009] Batch [0899]/[1203]	Speed: 125.167858 samples/sec	 accuracy=61.557778	 loss=1.550606	 lr=0.010000
Epoch[009] Batch [0919]/[1203]	Speed: 136.513835 samples/sec	 accuracy=61.572283	 loss=1.550787	 lr=0.010000
Epoch[009] Batch [0939]/[1203]	Speed: 110.156418 samples/sec	 accuracy=61.547340	 loss=1.551182	 lr=0.010000
Epoch[009] Batch [0959]/[1203]	Speed: 123.874722 samples/sec	 accuracy=61.554167	 loss=1.552012	 lr=0.010000
Epoch[009] Batch [0979]/[1203]	Speed: 117.355138 samples/sec	 accuracy=61.528571	 loss=1.552060	 lr=0.010000
Epoch[009] Batch [0999]/[1203]	Speed: 107.686555 samples/sec	 accuracy=61.522500	 loss=1.552488	 lr=0.010000
Epoch[009] Batch [1019]/[1203]	Speed: 148.127709 samples/sec	 accuracy=61.494608	 loss=1.553064	 lr=0.010000
Epoch[009] Batch [1039]/[1203]	Speed: 109.659870 samples/sec	 accuracy=61.466827	 loss=1.554091	 lr=0.010000
Epoch[009] Batch [1059]/[1203]	Speed: 129.508310 samples/sec	 accuracy=61.443868	 loss=1.554945	 lr=0.010000
Epoch[009] Batch [1079]/[1203]	Speed: 117.354015 samples/sec	 accuracy=61.450463	 loss=1.554938	 lr=0.010000
Epoch[009] Batch [1099]/[1203]	Speed: 116.333566 samples/sec	 accuracy=61.450455	 loss=1.554376	 lr=0.010000
Epoch[009] Batch [1119]/[1203]	Speed: 123.146742 samples/sec	 accuracy=61.458036	 loss=1.553675	 lr=0.010000
Epoch[009] Batch [1139]/[1203]	Speed: 113.863350 samples/sec	 accuracy=61.438596	 loss=1.554185	 lr=0.010000
Epoch[009] Batch [1159]/[1203]	Speed: 153.340408 samples/sec	 accuracy=61.445259	 loss=1.553671	 lr=0.010000
Epoch[009] Batch [1179]/[1203]	Speed: 117.872166 samples/sec	 accuracy=61.433898	 loss=1.553983	 lr=0.010000
Epoch[009] Batch [1199]/[1203]	Speed: 222.414138 samples/sec	 accuracy=61.412917	 loss=1.555209	 lr=0.010000
Batch [0019]/[0097]: acc-top1=55.275000 acc-top5=79.475000
Batch [0039]/[0097]: acc-top1=54.537500 acc-top5=78.325000
Batch [0059]/[0097]: acc-top1=57.000000 acc-top5=80.216667
Batch [0079]/[0097]: acc-top1=57.006250 acc-top5=80.325000
[Epoch 009] training: accuracy=61.412303	 loss=1.555266
[Epoch 009] speed: 120 samples/sec	time cost: 2180.190916
[Epoch 009] validation: acc-top1=56.288660 acc-top5=80.283505 loss=1.871801
Epoch[010] Batch [0019]/[1203]	Speed: 68.127749 samples/sec	 accuracy=63.925000	 loss=1.466839	 lr=0.010000
Epoch[010] Batch [0039]/[1203]	Speed: 96.842554 samples/sec	 accuracy=63.850000	 loss=1.483696	 lr=0.010000
Epoch[010] Batch [0059]/[1203]	Speed: 243.307742 samples/sec	 accuracy=63.783333	 loss=1.483707	 lr=0.010000
Epoch[010] Batch [0079]/[1203]	Speed: 91.524876 samples/sec	 accuracy=63.506250	 loss=1.480734	 lr=0.010000
Epoch[010] Batch [0099]/[1203]	Speed: 102.971551 samples/sec	 accuracy=63.445000	 loss=1.481339	 lr=0.010000
Epoch[010] Batch [0119]/[1203]	Speed: 201.273582 samples/sec	 accuracy=63.304167	 loss=1.483587	 lr=0.010000
Epoch[010] Batch [0139]/[1203]	Speed: 100.186299 samples/sec	 accuracy=63.214286	 loss=1.484485	 lr=0.010000
Epoch[010] Batch [0159]/[1203]	Speed: 174.292699 samples/sec	 accuracy=63.103125	 loss=1.487774	 lr=0.010000
Epoch[010] Batch [0179]/[1203]	Speed: 101.715413 samples/sec	 accuracy=63.144444	 loss=1.486122	 lr=0.010000
Epoch[010] Batch [0199]/[1203]	Speed: 113.986658 samples/sec	 accuracy=63.007500	 loss=1.487532	 lr=0.010000
Epoch[010] Batch [0219]/[1203]	Speed: 143.938323 samples/sec	 accuracy=62.909091	 loss=1.491596	 lr=0.010000
Epoch[010] Batch [0239]/[1203]	Speed: 110.760941 samples/sec	 accuracy=62.929167	 loss=1.491814	 lr=0.010000
Epoch[010] Batch [0259]/[1203]	Speed: 118.373279 samples/sec	 accuracy=62.819231	 loss=1.495307	 lr=0.010000
Epoch[010] Batch [0279]/[1203]	Speed: 126.807009 samples/sec	 accuracy=62.769643	 loss=1.494105	 lr=0.010000
Epoch[010] Batch [0299]/[1203]	Speed: 113.996106 samples/sec	 accuracy=62.675000	 loss=1.499271	 lr=0.010000
Epoch[010] Batch [0319]/[1203]	Speed: 136.451308 samples/sec	 accuracy=62.695312	 loss=1.499739	 lr=0.010000
Epoch[010] Batch [0339]/[1203]	Speed: 114.998974 samples/sec	 accuracy=62.573529	 loss=1.502341	 lr=0.010000
Epoch[010] Batch [0359]/[1203]	Speed: 135.849866 samples/sec	 accuracy=62.518056	 loss=1.503285	 lr=0.010000
Epoch[010] Batch [0379]/[1203]	Speed: 102.793245 samples/sec	 accuracy=62.505263	 loss=1.504327	 lr=0.010000
Epoch[010] Batch [0399]/[1203]	Speed: 116.284738 samples/sec	 accuracy=62.535000	 loss=1.503615	 lr=0.010000
Epoch[010] Batch [0419]/[1203]	Speed: 139.414909 samples/sec	 accuracy=62.523810	 loss=1.502754	 lr=0.010000
Epoch[010] Batch [0439]/[1203]	Speed: 122.216224 samples/sec	 accuracy=62.454545	 loss=1.504019	 lr=0.010000
Epoch[010] Batch [0459]/[1203]	Speed: 114.640603 samples/sec	 accuracy=62.483696	 loss=1.502757	 lr=0.010000
Epoch[010] Batch [0479]/[1203]	Speed: 123.731001 samples/sec	 accuracy=62.519792	 loss=1.500296	 lr=0.010000
Epoch[010] Batch [0499]/[1203]	Speed: 115.269037 samples/sec	 accuracy=62.571000	 loss=1.499107	 lr=0.010000
Epoch[010] Batch [0519]/[1203]	Speed: 151.378598 samples/sec	 accuracy=62.581731	 loss=1.499394	 lr=0.010000
Epoch[010] Batch [0539]/[1203]	Speed: 111.339875 samples/sec	 accuracy=62.576852	 loss=1.498911	 lr=0.010000
Epoch[010] Batch [0559]/[1203]	Speed: 110.622816 samples/sec	 accuracy=62.553571	 loss=1.499156	 lr=0.010000
Epoch[010] Batch [0579]/[1203]	Speed: 138.335218 samples/sec	 accuracy=62.509483	 loss=1.501472	 lr=0.010000
Epoch[010] Batch [0599]/[1203]	Speed: 113.862426 samples/sec	 accuracy=62.539167	 loss=1.500744	 lr=0.010000
Epoch[010] Batch [0619]/[1203]	Speed: 130.293888 samples/sec	 accuracy=62.527419	 loss=1.500777	 lr=0.010000
Epoch[010] Batch [0639]/[1203]	Speed: 118.986643 samples/sec	 accuracy=62.519531	 loss=1.502081	 lr=0.010000
Epoch[010] Batch [0659]/[1203]	Speed: 118.550524 samples/sec	 accuracy=62.500000	 loss=1.502645	 lr=0.010000
Epoch[010] Batch [0679]/[1203]	Speed: 126.744410 samples/sec	 accuracy=62.436765	 loss=1.504965	 lr=0.010000
Epoch[010] Batch [0699]/[1203]	Speed: 126.476034 samples/sec	 accuracy=62.467143	 loss=1.505372	 lr=0.010000
Epoch[010] Batch [0719]/[1203]	Speed: 109.715717 samples/sec	 accuracy=62.450000	 loss=1.505596	 lr=0.010000
Epoch[010] Batch [0739]/[1203]	Speed: 129.101955 samples/sec	 accuracy=62.425000	 loss=1.505895	 lr=0.010000
Epoch[010] Batch [0759]/[1203]	Speed: 117.740128 samples/sec	 accuracy=62.438158	 loss=1.505787	 lr=0.010000
Epoch[010] Batch [0779]/[1203]	Speed: 132.575393 samples/sec	 accuracy=62.426282	 loss=1.506395	 lr=0.010000
Epoch[010] Batch [0799]/[1203]	Speed: 111.951918 samples/sec	 accuracy=62.430000	 loss=1.506362	 lr=0.010000
Epoch[010] Batch [0819]/[1203]	Speed: 121.629988 samples/sec	 accuracy=62.440244	 loss=1.507140	 lr=0.010000
Epoch[010] Batch [0839]/[1203]	Speed: 134.677759 samples/sec	 accuracy=62.450000	 loss=1.506432	 lr=0.010000
Epoch[010] Batch [0859]/[1203]	Speed: 120.155324 samples/sec	 accuracy=62.448837	 loss=1.506839	 lr=0.010000
Epoch[010] Batch [0879]/[1203]	Speed: 117.709806 samples/sec	 accuracy=62.464773	 loss=1.506038	 lr=0.010000
Epoch[010] Batch [0899]/[1203]	Speed: 113.272724 samples/sec	 accuracy=62.436667	 loss=1.506877	 lr=0.010000
Epoch[010] Batch [0919]/[1203]	Speed: 132.223075 samples/sec	 accuracy=62.423370	 loss=1.507581	 lr=0.010000
Epoch[010] Batch [0939]/[1203]	Speed: 125.608869 samples/sec	 accuracy=62.415426	 loss=1.508179	 lr=0.010000
Epoch[010] Batch [0959]/[1203]	Speed: 120.507742 samples/sec	 accuracy=62.391667	 loss=1.508894	 lr=0.010000
Epoch[010] Batch [0979]/[1203]	Speed: 122.349895 samples/sec	 accuracy=62.395408	 loss=1.509108	 lr=0.010000
Epoch[010] Batch [0999]/[1203]	Speed: 108.380190 samples/sec	 accuracy=62.407500	 loss=1.508743	 lr=0.010000
Epoch[010] Batch [1019]/[1203]	Speed: 136.019776 samples/sec	 accuracy=62.367647	 loss=1.510742	 lr=0.010000
Epoch[010] Batch [1039]/[1203]	Speed: 119.809368 samples/sec	 accuracy=62.367308	 loss=1.511098	 lr=0.010000
Epoch[010] Batch [1059]/[1203]	Speed: 130.685600 samples/sec	 accuracy=62.366509	 loss=1.511753	 lr=0.010000
Epoch[010] Batch [1079]/[1203]	Speed: 133.851501 samples/sec	 accuracy=62.356944	 loss=1.512219	 lr=0.010000
Epoch[010] Batch [1099]/[1203]	Speed: 100.172055 samples/sec	 accuracy=62.336818	 loss=1.513114	 lr=0.010000
Epoch[010] Batch [1119]/[1203]	Speed: 141.489935 samples/sec	 accuracy=62.329464	 loss=1.513390	 lr=0.010000
Epoch[010] Batch [1139]/[1203]	Speed: 118.539578 samples/sec	 accuracy=62.327632	 loss=1.514160	 lr=0.010000
Epoch[010] Batch [1159]/[1203]	Speed: 107.225912 samples/sec	 accuracy=62.312931	 loss=1.514919	 lr=0.010000
Epoch[010] Batch [1179]/[1203]	Speed: 147.997226 samples/sec	 accuracy=62.334746	 loss=1.513846	 lr=0.010000
Epoch[010] Batch [1199]/[1203]	Speed: 225.614721 samples/sec	 accuracy=62.314167	 loss=1.514562	 lr=0.010000
Batch [0019]/[0097]: acc-top1=57.075000 acc-top5=79.825000
Batch [0039]/[0097]: acc-top1=55.725000 acc-top5=79.050000
Batch [0059]/[0097]: acc-top1=57.783333 acc-top5=80.508333
Batch [0079]/[0097]: acc-top1=57.468750 acc-top5=80.506250
[Epoch 010] training: accuracy=62.311305	 loss=1.514720
[Epoch 010] speed: 121 samples/sec	time cost: 2176.512620
[Epoch 010] validation: acc-top1=56.608247 acc-top5=80.087629 loss=1.874069
Epoch[011] Batch [0019]/[1204]	Speed: 67.687954 samples/sec	 accuracy=63.350000	 loss=1.464293	 lr=0.010000
Epoch[011] Batch [0039]/[1204]	Speed: 97.906630 samples/sec	 accuracy=63.725000	 loss=1.445887	 lr=0.010000
Epoch[011] Batch [0059]/[1204]	Speed: 226.738506 samples/sec	 accuracy=63.741667	 loss=1.440417	 lr=0.010000
Epoch[011] Batch [0079]/[1204]	Speed: 95.849330 samples/sec	 accuracy=64.068750	 loss=1.441999	 lr=0.010000
Epoch[011] Batch [0099]/[1204]	Speed: 98.253564 samples/sec	 accuracy=63.910000	 loss=1.447025	 lr=0.010000
Epoch[011] Batch [0119]/[1204]	Speed: 195.489491 samples/sec	 accuracy=64.033333	 loss=1.441124	 lr=0.010000
Epoch[011] Batch [0139]/[1204]	Speed: 93.178653 samples/sec	 accuracy=64.035714	 loss=1.440451	 lr=0.010000
Epoch[011] Batch [0159]/[1204]	Speed: 231.275991 samples/sec	 accuracy=64.175000	 loss=1.433398	 lr=0.010000
Epoch[011] Batch [0179]/[1204]	Speed: 93.495698 samples/sec	 accuracy=64.166667	 loss=1.437300	 lr=0.010000
Epoch[011] Batch [0199]/[1204]	Speed: 105.029560 samples/sec	 accuracy=64.160000	 loss=1.439310	 lr=0.010000
Epoch[011] Batch [0219]/[1204]	Speed: 175.586783 samples/sec	 accuracy=64.115909	 loss=1.440252	 lr=0.010000
Epoch[011] Batch [0239]/[1204]	Speed: 106.779608 samples/sec	 accuracy=64.020833	 loss=1.442145	 lr=0.010000
Epoch[011] Batch [0259]/[1204]	Speed: 118.659344 samples/sec	 accuracy=63.992308	 loss=1.443965	 lr=0.010000
Epoch[011] Batch [0279]/[1204]	Speed: 128.562942 samples/sec	 accuracy=64.017857	 loss=1.443625	 lr=0.010000
Epoch[011] Batch [0299]/[1204]	Speed: 118.944246 samples/sec	 accuracy=63.956667	 loss=1.444353	 lr=0.010000
Epoch[011] Batch [0319]/[1204]	Speed: 136.173094 samples/sec	 accuracy=63.909375	 loss=1.444384	 lr=0.010000
Epoch[011] Batch [0339]/[1204]	Speed: 101.649194 samples/sec	 accuracy=63.838235	 loss=1.446332	 lr=0.010000
Epoch[011] Batch [0359]/[1204]	Speed: 135.381954 samples/sec	 accuracy=63.858333	 loss=1.443815	 lr=0.010000
Epoch[011] Batch [0379]/[1204]	Speed: 131.845538 samples/sec	 accuracy=63.782895	 loss=1.446542	 lr=0.010000
Epoch[011] Batch [0399]/[1204]	Speed: 108.628081 samples/sec	 accuracy=63.746250	 loss=1.446378	 lr=0.010000
Epoch[011] Batch [0419]/[1204]	Speed: 130.015643 samples/sec	 accuracy=63.697619	 loss=1.448778	 lr=0.010000
Epoch[011] Batch [0439]/[1204]	Speed: 114.067742 samples/sec	 accuracy=63.639773	 loss=1.450903	 lr=0.010000
Epoch[011] Batch [0459]/[1204]	Speed: 128.977715 samples/sec	 accuracy=63.606522	 loss=1.452309	 lr=0.010000
Epoch[011] Batch [0479]/[1204]	Speed: 135.748063 samples/sec	 accuracy=63.611458	 loss=1.453796	 lr=0.010000
Epoch[011] Batch [0499]/[1204]	Speed: 100.563460 samples/sec	 accuracy=63.575000	 loss=1.455699	 lr=0.010000
Epoch[011] Batch [0519]/[1204]	Speed: 144.962641 samples/sec	 accuracy=63.506731	 loss=1.458195	 lr=0.010000
Epoch[011] Batch [0539]/[1204]	Speed: 124.532262 samples/sec	 accuracy=63.513889	 loss=1.458511	 lr=0.010000
Epoch[011] Batch [0559]/[1204]	Speed: 119.539811 samples/sec	 accuracy=63.466964	 loss=1.460465	 lr=0.010000
Epoch[011] Batch [0579]/[1204]	Speed: 129.276433 samples/sec	 accuracy=63.459483	 loss=1.460306	 lr=0.010000
Epoch[011] Batch [0599]/[1204]	Speed: 106.888303 samples/sec	 accuracy=63.440833	 loss=1.461232	 lr=0.010000
Epoch[011] Batch [0619]/[1204]	Speed: 128.615156 samples/sec	 accuracy=63.416129	 loss=1.462368	 lr=0.010000
Epoch[011] Batch [0639]/[1204]	Speed: 121.699241 samples/sec	 accuracy=63.362500	 loss=1.464683	 lr=0.010000
Epoch[011] Batch [0659]/[1204]	Speed: 115.695282 samples/sec	 accuracy=63.323485	 loss=1.466486	 lr=0.010000
Epoch[011] Batch [0679]/[1204]	Speed: 133.335707 samples/sec	 accuracy=63.301471	 loss=1.467164	 lr=0.010000
Epoch[011] Batch [0699]/[1204]	Speed: 119.988288 samples/sec	 accuracy=63.262857	 loss=1.468077	 lr=0.010000
Epoch[011] Batch [0719]/[1204]	Speed: 114.549071 samples/sec	 accuracy=63.234028	 loss=1.469391	 lr=0.010000
Epoch[011] Batch [0739]/[1204]	Speed: 142.261792 samples/sec	 accuracy=63.205405	 loss=1.470295	 lr=0.010000
Epoch[011] Batch [0759]/[1204]	Speed: 104.833317 samples/sec	 accuracy=63.182237	 loss=1.472046	 lr=0.010000
Epoch[011] Batch [0779]/[1204]	Speed: 118.967781 samples/sec	 accuracy=63.215385	 loss=1.471636	 lr=0.010000
Epoch[011] Batch [0799]/[1204]	Speed: 132.520230 samples/sec	 accuracy=63.164375	 loss=1.473501	 lr=0.010000
Epoch[011] Batch [0819]/[1204]	Speed: 109.055950 samples/sec	 accuracy=63.162195	 loss=1.474065	 lr=0.010000
Epoch[011] Batch [0839]/[1204]	Speed: 138.635112 samples/sec	 accuracy=63.161905	 loss=1.474275	 lr=0.010000
Epoch[011] Batch [0859]/[1204]	Speed: 110.905425 samples/sec	 accuracy=63.145349	 loss=1.475469	 lr=0.010000
Epoch[011] Batch [0879]/[1204]	Speed: 137.637609 samples/sec	 accuracy=63.168182	 loss=1.475225	 lr=0.010000
Epoch[011] Batch [0899]/[1204]	Speed: 111.093863 samples/sec	 accuracy=63.129444	 loss=1.475700	 lr=0.010000
Epoch[011] Batch [0919]/[1204]	Speed: 124.161755 samples/sec	 accuracy=63.151087	 loss=1.475269	 lr=0.010000
Epoch[011] Batch [0939]/[1204]	Speed: 135.930431 samples/sec	 accuracy=63.118085	 loss=1.476298	 lr=0.010000
Epoch[011] Batch [0959]/[1204]	Speed: 109.795286 samples/sec	 accuracy=63.083333	 loss=1.477710	 lr=0.010000
Epoch[011] Batch [0979]/[1204]	Speed: 129.689168 samples/sec	 accuracy=63.036224	 loss=1.479755	 lr=0.010000
Epoch[011] Batch [0999]/[1204]	Speed: 121.889766 samples/sec	 accuracy=63.042500	 loss=1.479518	 lr=0.010000
Epoch[011] Batch [1019]/[1204]	Speed: 115.514898 samples/sec	 accuracy=63.032353	 loss=1.480191	 lr=0.010000
Epoch[011] Batch [1039]/[1204]	Speed: 143.737646 samples/sec	 accuracy=63.019231	 loss=1.481197	 lr=0.010000
Epoch[011] Batch [1059]/[1204]	Speed: 111.201475 samples/sec	 accuracy=62.998585	 loss=1.482128	 lr=0.010000
Epoch[011] Batch [1079]/[1204]	Speed: 115.744442 samples/sec	 accuracy=62.995370	 loss=1.481809	 lr=0.010000
Epoch[011] Batch [1099]/[1204]	Speed: 131.349296 samples/sec	 accuracy=62.971364	 loss=1.482820	 lr=0.010000
Epoch[011] Batch [1119]/[1204]	Speed: 106.803086 samples/sec	 accuracy=62.956250	 loss=1.483679	 lr=0.010000
Epoch[011] Batch [1139]/[1204]	Speed: 131.642870 samples/sec	 accuracy=62.948684	 loss=1.484504	 lr=0.010000
Epoch[011] Batch [1159]/[1204]	Speed: 131.696737 samples/sec	 accuracy=62.945690	 loss=1.484882	 lr=0.010000
Epoch[011] Batch [1179]/[1204]	Speed: 121.371449 samples/sec	 accuracy=62.945763	 loss=1.485333	 lr=0.010000
Epoch[011] Batch [1199]/[1204]	Speed: 218.112150 samples/sec	 accuracy=62.962083	 loss=1.485554	 lr=0.010000
Batch [0019]/[0097]: acc-top1=57.750000 acc-top5=81.525000
Batch [0039]/[0097]: acc-top1=56.300000 acc-top5=79.812500
Batch [0059]/[0097]: acc-top1=58.341667 acc-top5=81.141667
Batch [0079]/[0097]: acc-top1=58.393750 acc-top5=81.175000
[Epoch 011] training: accuracy=62.963455	 loss=1.485527
[Epoch 011] speed: 121 samples/sec	time cost: 2179.377461
[Epoch 011] validation: acc-top1=57.747423 acc-top5=81.118557 loss=1.795732
Epoch[012] Batch [0019]/[1203]	Speed: 68.454602 samples/sec	 accuracy=65.025000	 loss=1.426093	 lr=0.010000
Epoch[012] Batch [0039]/[1203]	Speed: 97.813487 samples/sec	 accuracy=64.912500	 loss=1.417654	 lr=0.010000
Epoch[012] Batch [0059]/[1203]	Speed: 244.429788 samples/sec	 accuracy=64.741667	 loss=1.411016	 lr=0.010000
Epoch[012] Batch [0079]/[1203]	Speed: 89.279891 samples/sec	 accuracy=64.356250	 loss=1.419477	 lr=0.010000
Epoch[012] Batch [0099]/[1203]	Speed: 106.491210 samples/sec	 accuracy=64.230000	 loss=1.426395	 lr=0.010000
Epoch[012] Batch [0119]/[1203]	Speed: 169.035986 samples/sec	 accuracy=64.020833	 loss=1.427436	 lr=0.010000
Epoch[012] Batch [0139]/[1203]	Speed: 110.990012 samples/sec	 accuracy=64.189286	 loss=1.419643	 lr=0.010000
Epoch[012] Batch [0159]/[1203]	Speed: 164.559985 samples/sec	 accuracy=64.340625	 loss=1.415271	 lr=0.010000
Epoch[012] Batch [0179]/[1203]	Speed: 103.123379 samples/sec	 accuracy=64.313889	 loss=1.415564	 lr=0.010000
Epoch[012] Batch [0199]/[1203]	Speed: 106.418345 samples/sec	 accuracy=64.287500	 loss=1.417086	 lr=0.010000
Epoch[012] Batch [0219]/[1203]	Speed: 141.979130 samples/sec	 accuracy=64.295455	 loss=1.417210	 lr=0.010000
Epoch[012] Batch [0239]/[1203]	Speed: 126.809200 samples/sec	 accuracy=64.256250	 loss=1.418568	 lr=0.010000
Epoch[012] Batch [0259]/[1203]	Speed: 112.311523 samples/sec	 accuracy=64.217308	 loss=1.419679	 lr=0.010000
Epoch[012] Batch [0279]/[1203]	Speed: 116.896354 samples/sec	 accuracy=64.116071	 loss=1.423723	 lr=0.010000
Epoch[012] Batch [0299]/[1203]	Speed: 134.499639 samples/sec	 accuracy=64.115000	 loss=1.425713	 lr=0.010000
Epoch[012] Batch [0319]/[1203]	Speed: 132.486823 samples/sec	 accuracy=64.109375	 loss=1.426980	 lr=0.010000
Epoch[012] Batch [0339]/[1203]	Speed: 104.841697 samples/sec	 accuracy=64.142647	 loss=1.427123	 lr=0.010000
Epoch[012] Batch [0359]/[1203]	Speed: 129.680914 samples/sec	 accuracy=64.166667	 loss=1.428262	 lr=0.010000
Epoch[012] Batch [0379]/[1203]	Speed: 122.772537 samples/sec	 accuracy=64.161842	 loss=1.427886	 lr=0.010000
Epoch[012] Batch [0399]/[1203]	Speed: 127.959324 samples/sec	 accuracy=64.137500	 loss=1.429520	 lr=0.010000
Epoch[012] Batch [0419]/[1203]	Speed: 124.330099 samples/sec	 accuracy=64.105952	 loss=1.430322	 lr=0.010000
Epoch[012] Batch [0439]/[1203]	Speed: 96.892650 samples/sec	 accuracy=64.044318	 loss=1.432833	 lr=0.010000
Epoch[012] Batch [0459]/[1203]	Speed: 159.934315 samples/sec	 accuracy=64.011957	 loss=1.433214	 lr=0.010000
Epoch[012] Batch [0479]/[1203]	Speed: 111.216151 samples/sec	 accuracy=63.993750	 loss=1.433320	 lr=0.010000
Epoch[012] Batch [0499]/[1203]	Speed: 120.766990 samples/sec	 accuracy=63.984000	 loss=1.434128	 lr=0.010000
Epoch[012] Batch [0519]/[1203]	Speed: 138.029121 samples/sec	 accuracy=64.003846	 loss=1.434317	 lr=0.010000
Epoch[012] Batch [0539]/[1203]	Speed: 97.303962 samples/sec	 accuracy=63.964815	 loss=1.435343	 lr=0.010000
Epoch[012] Batch [0559]/[1203]	Speed: 184.324261 samples/sec	 accuracy=63.909821	 loss=1.436445	 lr=0.010000
Epoch[012] Batch [0579]/[1203]	Speed: 104.885305 samples/sec	 accuracy=63.873276	 loss=1.439031	 lr=0.010000
Epoch[012] Batch [0599]/[1203]	Speed: 107.565857 samples/sec	 accuracy=63.851667	 loss=1.438683	 lr=0.010000
Epoch[012] Batch [0619]/[1203]	Speed: 169.082509 samples/sec	 accuracy=63.826613	 loss=1.439422	 lr=0.010000
Epoch[012] Batch [0639]/[1203]	Speed: 101.256831 samples/sec	 accuracy=63.836719	 loss=1.439060	 lr=0.010000
Epoch[012] Batch [0659]/[1203]	Speed: 116.905863 samples/sec	 accuracy=63.835606	 loss=1.439466	 lr=0.010000
Epoch[012] Batch [0679]/[1203]	Speed: 143.828065 samples/sec	 accuracy=63.817647	 loss=1.441640	 lr=0.010000
Epoch[012] Batch [0699]/[1203]	Speed: 98.340671 samples/sec	 accuracy=63.793571	 loss=1.441981	 lr=0.010000
Epoch[012] Batch [0719]/[1203]	Speed: 180.010422 samples/sec	 accuracy=63.806250	 loss=1.442415	 lr=0.010000
Epoch[012] Batch [0739]/[1203]	Speed: 98.142139 samples/sec	 accuracy=63.775676	 loss=1.444193	 lr=0.010000
Epoch[012] Batch [0759]/[1203]	Speed: 125.856234 samples/sec	 accuracy=63.756579	 loss=1.444626	 lr=0.010000
Epoch[012] Batch [0779]/[1203]	Speed: 142.433200 samples/sec	 accuracy=63.748077	 loss=1.445217	 lr=0.010000
Epoch[012] Batch [0799]/[1203]	Speed: 99.059113 samples/sec	 accuracy=63.738125	 loss=1.445668	 lr=0.010000
Epoch[012] Batch [0819]/[1203]	Speed: 148.061229 samples/sec	 accuracy=63.726220	 loss=1.445961	 lr=0.010000
Epoch[012] Batch [0839]/[1203]	Speed: 109.081190 samples/sec	 accuracy=63.692262	 loss=1.446213	 lr=0.010000
Epoch[012] Batch [0859]/[1203]	Speed: 118.184385 samples/sec	 accuracy=63.655233	 loss=1.447804	 lr=0.010000
Epoch[012] Batch [0879]/[1203]	Speed: 151.392139 samples/sec	 accuracy=63.609091	 loss=1.449531	 lr=0.010000
Epoch[012] Batch [0899]/[1203]	Speed: 101.820994 samples/sec	 accuracy=63.586667	 loss=1.450278	 lr=0.010000
Epoch[012] Batch [0919]/[1203]	Speed: 174.301597 samples/sec	 accuracy=63.585326	 loss=1.449775	 lr=0.010000
Epoch[012] Batch [0939]/[1203]	Speed: 94.452916 samples/sec	 accuracy=63.593085	 loss=1.449166	 lr=0.010000
Epoch[012] Batch [0959]/[1203]	Speed: 111.116440 samples/sec	 accuracy=63.582292	 loss=1.449524	 lr=0.010000
Epoch[012] Batch [0979]/[1203]	Speed: 168.804421 samples/sec	 accuracy=63.561735	 loss=1.450840	 lr=0.010000
Epoch[012] Batch [0999]/[1203]	Speed: 112.351699 samples/sec	 accuracy=63.574500	 loss=1.450706	 lr=0.010000
Epoch[012] Batch [1019]/[1203]	Speed: 136.788152 samples/sec	 accuracy=63.577941	 loss=1.450111	 lr=0.010000
Epoch[012] Batch [1039]/[1203]	Speed: 105.918891 samples/sec	 accuracy=63.556250	 loss=1.451377	 lr=0.010000
Epoch[012] Batch [1059]/[1203]	Speed: 113.455750 samples/sec	 accuracy=63.563679	 loss=1.450836	 lr=0.010000
Epoch[012] Batch [1079]/[1203]	Speed: 138.598442 samples/sec	 accuracy=63.548611	 loss=1.451519	 lr=0.010000
Epoch[012] Batch [1099]/[1203]	Speed: 126.220684 samples/sec	 accuracy=63.540000	 loss=1.452432	 lr=0.010000
Epoch[012] Batch [1119]/[1203]	Speed: 117.706979 samples/sec	 accuracy=63.538839	 loss=1.453350	 lr=0.010000
Epoch[012] Batch [1139]/[1203]	Speed: 132.445113 samples/sec	 accuracy=63.537281	 loss=1.453966	 lr=0.010000
Epoch[012] Batch [1159]/[1203]	Speed: 104.078089 samples/sec	 accuracy=63.547414	 loss=1.453832	 lr=0.010000
Epoch[012] Batch [1179]/[1203]	Speed: 134.352324 samples/sec	 accuracy=63.533475	 loss=1.454293	 lr=0.010000
Epoch[012] Batch [1199]/[1203]	Speed: 255.561101 samples/sec	 accuracy=63.504167	 loss=1.455007	 lr=0.010000
Batch [0019]/[0097]: acc-top1=59.425000 acc-top5=81.875000
Batch [0039]/[0097]: acc-top1=57.825000 acc-top5=80.362500
Batch [0059]/[0097]: acc-top1=59.658333 acc-top5=81.525000
Batch [0079]/[0097]: acc-top1=59.700000 acc-top5=81.437500
[Epoch 012] training: accuracy=63.496259	 loss=1.455199
[Epoch 012] speed: 121 samples/sec	time cost: 2176.932960
[Epoch 012] validation: acc-top1=58.443299 acc-top5=81.453608 loss=1.781295
Epoch[013] Batch [0019]/[1203]	Speed: 68.206546 samples/sec	 accuracy=64.725000	 loss=1.398139	 lr=0.010000
Epoch[013] Batch [0039]/[1203]	Speed: 98.260690 samples/sec	 accuracy=64.312500	 loss=1.397419	 lr=0.010000
Epoch[013] Batch [0059]/[1203]	Speed: 236.688129 samples/sec	 accuracy=64.508333	 loss=1.386325	 lr=0.010000
Epoch[013] Batch [0079]/[1203]	Speed: 91.911811 samples/sec	 accuracy=64.437500	 loss=1.398582	 lr=0.010000
Epoch[013] Batch [0099]/[1203]	Speed: 113.700166 samples/sec	 accuracy=64.475000	 loss=1.403785	 lr=0.010000
Epoch[013] Batch [0119]/[1203]	Speed: 161.874797 samples/sec	 accuracy=64.508333	 loss=1.405219	 lr=0.010000
Epoch[013] Batch [0139]/[1203]	Speed: 103.653394 samples/sec	 accuracy=64.446429	 loss=1.405360	 lr=0.010000
Epoch[013] Batch [0159]/[1203]	Speed: 156.658167 samples/sec	 accuracy=64.531250	 loss=1.398656	 lr=0.010000
Epoch[013] Batch [0179]/[1203]	Speed: 109.224479 samples/sec	 accuracy=64.533333	 loss=1.399048	 lr=0.010000
Epoch[013] Batch [0199]/[1203]	Speed: 105.165670 samples/sec	 accuracy=64.470000	 loss=1.399968	 lr=0.010000
Epoch[013] Batch [0219]/[1203]	Speed: 162.556839 samples/sec	 accuracy=64.393182	 loss=1.404712	 lr=0.010000
Epoch[013] Batch [0239]/[1203]	Speed: 108.117235 samples/sec	 accuracy=64.327083	 loss=1.406985	 lr=0.010000
Epoch[013] Batch [0259]/[1203]	Speed: 117.643462 samples/sec	 accuracy=64.363462	 loss=1.405700	 lr=0.010000
Epoch[013] Batch [0279]/[1203]	Speed: 118.379212 samples/sec	 accuracy=64.408929	 loss=1.404266	 lr=0.010000
Epoch[013] Batch [0299]/[1203]	Speed: 134.392523 samples/sec	 accuracy=64.450000	 loss=1.402461	 lr=0.010000
Epoch[013] Batch [0319]/[1203]	Speed: 127.035623 samples/sec	 accuracy=64.465625	 loss=1.400187	 lr=0.010000
Epoch[013] Batch [0339]/[1203]	Speed: 102.570942 samples/sec	 accuracy=64.455882	 loss=1.401572	 lr=0.010000
Epoch[013] Batch [0359]/[1203]	Speed: 137.514443 samples/sec	 accuracy=64.423611	 loss=1.403729	 lr=0.010000
Epoch[013] Batch [0379]/[1203]	Speed: 118.900536 samples/sec	 accuracy=64.356579	 loss=1.406203	 lr=0.010000
Epoch[013] Batch [0399]/[1203]	Speed: 121.743205 samples/sec	 accuracy=64.417500	 loss=1.406434	 lr=0.010000
Epoch[013] Batch [0419]/[1203]	Speed: 127.124361 samples/sec	 accuracy=64.438095	 loss=1.405313	 lr=0.010000
Epoch[013] Batch [0439]/[1203]	Speed: 110.276166 samples/sec	 accuracy=64.397727	 loss=1.406089	 lr=0.010000
Epoch[013] Batch [0459]/[1203]	Speed: 129.043188 samples/sec	 accuracy=64.430435	 loss=1.406331	 lr=0.010000
Epoch[013] Batch [0479]/[1203]	Speed: 122.566574 samples/sec	 accuracy=64.458333	 loss=1.406050	 lr=0.010000
Epoch[013] Batch [0499]/[1203]	Speed: 124.030833 samples/sec	 accuracy=64.389000	 loss=1.407581	 lr=0.010000
Epoch[013] Batch [0519]/[1203]	Speed: 121.416515 samples/sec	 accuracy=64.318269	 loss=1.410980	 lr=0.010000
Epoch[013] Batch [0539]/[1203]	Speed: 117.972281 samples/sec	 accuracy=64.269444	 loss=1.413759	 lr=0.010000
Epoch[013] Batch [0559]/[1203]	Speed: 122.884871 samples/sec	 accuracy=64.279464	 loss=1.414048	 lr=0.010000
Epoch[013] Batch [0579]/[1203]	Speed: 118.867410 samples/sec	 accuracy=64.267241	 loss=1.415074	 lr=0.010000
Epoch[013] Batch [0599]/[1203]	Speed: 138.023091 samples/sec	 accuracy=64.270833	 loss=1.414675	 lr=0.010000
Epoch[013] Batch [0619]/[1203]	Speed: 112.717705 samples/sec	 accuracy=64.241129	 loss=1.415461	 lr=0.010000
Epoch[013] Batch [0639]/[1203]	Speed: 121.467132 samples/sec	 accuracy=64.206250	 loss=1.416861	 lr=0.010000
Epoch[013] Batch [0659]/[1203]	Speed: 135.923153 samples/sec	 accuracy=64.225000	 loss=1.416063	 lr=0.010000
Epoch[013] Batch [0679]/[1203]	Speed: 101.472870 samples/sec	 accuracy=64.252206	 loss=1.415306	 lr=0.010000
Epoch[013] Batch [0699]/[1203]	Speed: 118.365035 samples/sec	 accuracy=64.235000	 loss=1.415768	 lr=0.010000
Epoch[013] Batch [0719]/[1203]	Speed: 138.690784 samples/sec	 accuracy=64.256944	 loss=1.414852	 lr=0.010000
Epoch[013] Batch [0739]/[1203]	Speed: 109.610730 samples/sec	 accuracy=64.209459	 loss=1.416445	 lr=0.010000
Epoch[013] Batch [0759]/[1203]	Speed: 146.546499 samples/sec	 accuracy=64.192763	 loss=1.417542	 lr=0.010000
Epoch[013] Batch [0779]/[1203]	Speed: 102.964174 samples/sec	 accuracy=64.214744	 loss=1.417556	 lr=0.010000
Epoch[013] Batch [0799]/[1203]	Speed: 143.251222 samples/sec	 accuracy=64.210625	 loss=1.418448	 lr=0.010000
Epoch[013] Batch [0819]/[1203]	Speed: 122.659522 samples/sec	 accuracy=64.237195	 loss=1.418417	 lr=0.010000
Epoch[013] Batch [0839]/[1203]	Speed: 116.130760 samples/sec	 accuracy=64.238690	 loss=1.418735	 lr=0.010000
Epoch[013] Batch [0859]/[1203]	Speed: 126.820377 samples/sec	 accuracy=64.223256	 loss=1.419306	 lr=0.010000
Epoch[013] Batch [0879]/[1203]	Speed: 108.547526 samples/sec	 accuracy=64.227273	 loss=1.418606	 lr=0.010000
Epoch[013] Batch [0899]/[1203]	Speed: 140.045897 samples/sec	 accuracy=64.228333	 loss=1.418687	 lr=0.010000
Epoch[013] Batch [0919]/[1203]	Speed: 121.024884 samples/sec	 accuracy=64.183696	 loss=1.420434	 lr=0.010000
Epoch[013] Batch [0939]/[1203]	Speed: 103.741040 samples/sec	 accuracy=64.173404	 loss=1.421514	 lr=0.010000
Epoch[013] Batch [0959]/[1203]	Speed: 173.607107 samples/sec	 accuracy=64.142708	 loss=1.422845	 lr=0.010000
Epoch[013] Batch [0979]/[1203]	Speed: 101.884806 samples/sec	 accuracy=64.140816	 loss=1.423490	 lr=0.010000
Epoch[013] Batch [0999]/[1203]	Speed: 121.380712 samples/sec	 accuracy=64.129500	 loss=1.423867	 lr=0.010000
Epoch[013] Batch [1019]/[1203]	Speed: 135.039290 samples/sec	 accuracy=64.107843	 loss=1.424715	 lr=0.010000
Epoch[013] Batch [1039]/[1203]	Speed: 118.592632 samples/sec	 accuracy=64.111538	 loss=1.425590	 lr=0.010000
Epoch[013] Batch [1059]/[1203]	Speed: 130.173706 samples/sec	 accuracy=64.090094	 loss=1.426699	 lr=0.010000
Epoch[013] Batch [1079]/[1203]	Speed: 108.599361 samples/sec	 accuracy=64.074537	 loss=1.427512	 lr=0.010000
Epoch[013] Batch [1099]/[1203]	Speed: 113.982909 samples/sec	 accuracy=64.061818	 loss=1.428400	 lr=0.010000
Epoch[013] Batch [1119]/[1203]	Speed: 142.850238 samples/sec	 accuracy=64.078125	 loss=1.428072	 lr=0.010000
Epoch[013] Batch [1139]/[1203]	Speed: 118.049040 samples/sec	 accuracy=64.082895	 loss=1.427937	 lr=0.010000
Epoch[013] Batch [1159]/[1203]	Speed: 123.015169 samples/sec	 accuracy=64.053448	 loss=1.429295	 lr=0.010000
Epoch[013] Batch [1179]/[1203]	Speed: 120.999103 samples/sec	 accuracy=64.044068	 loss=1.430016	 lr=0.010000
Epoch[013] Batch [1199]/[1203]	Speed: 263.820285 samples/sec	 accuracy=64.019583	 loss=1.431000	 lr=0.010000
Batch [0019]/[0097]: acc-top1=58.600000 acc-top5=81.475000
Batch [0039]/[0097]: acc-top1=56.500000 acc-top5=79.700000
Batch [0059]/[0097]: acc-top1=58.683333 acc-top5=81.083333
Batch [0079]/[0097]: acc-top1=58.725000 acc-top5=81.118750
[Epoch 013] training: accuracy=64.018703	 loss=1.430886
[Epoch 013] speed: 121 samples/sec	time cost: 2175.214781
[Epoch 013] validation: acc-top1=58.335052 acc-top5=81.221649 loss=1.796236
Epoch[014] Batch [0019]/[1203]	Speed: 68.410384 samples/sec	 accuracy=65.675000	 loss=1.373936	 lr=0.010000
Epoch[014] Batch [0039]/[1203]	Speed: 100.801163 samples/sec	 accuracy=65.475000	 loss=1.365103	 lr=0.010000
Epoch[014] Batch [0059]/[1203]	Speed: 212.961297 samples/sec	 accuracy=65.775000	 loss=1.369146	 lr=0.010000
Epoch[014] Batch [0079]/[1203]	Speed: 90.643577 samples/sec	 accuracy=65.556250	 loss=1.376849	 lr=0.010000
Epoch[014] Batch [0099]/[1203]	Speed: 117.967201 samples/sec	 accuracy=65.565000	 loss=1.374813	 lr=0.010000
Epoch[014] Batch [0119]/[1203]	Speed: 162.153537 samples/sec	 accuracy=65.587500	 loss=1.376424	 lr=0.010000
Epoch[014] Batch [0139]/[1203]	Speed: 94.666957 samples/sec	 accuracy=65.707143	 loss=1.373945	 lr=0.010000
Epoch[014] Batch [0159]/[1203]	Speed: 204.394704 samples/sec	 accuracy=65.490625	 loss=1.375682	 lr=0.010000
Epoch[014] Batch [0179]/[1203]	Speed: 92.309780 samples/sec	 accuracy=65.336111	 loss=1.381406	 lr=0.010000
Epoch[014] Batch [0199]/[1203]	Speed: 132.910662 samples/sec	 accuracy=65.465000	 loss=1.373406	 lr=0.010000
Epoch[014] Batch [0219]/[1203]	Speed: 129.680981 samples/sec	 accuracy=65.350000	 loss=1.378005	 lr=0.010000
Epoch[014] Batch [0239]/[1203]	Speed: 96.692418 samples/sec	 accuracy=65.443750	 loss=1.372643	 lr=0.010000
Epoch[014] Batch [0259]/[1203]	Speed: 163.128600 samples/sec	 accuracy=65.455769	 loss=1.371460	 lr=0.010000
Epoch[014] Batch [0279]/[1203]	Speed: 112.848404 samples/sec	 accuracy=65.375000	 loss=1.375953	 lr=0.010000
Epoch[014] Batch [0299]/[1203]	Speed: 109.882386 samples/sec	 accuracy=65.341667	 loss=1.377957	 lr=0.010000
Epoch[014] Batch [0319]/[1203]	Speed: 155.125739 samples/sec	 accuracy=65.340625	 loss=1.378439	 lr=0.010000
Epoch[014] Batch [0339]/[1203]	Speed: 97.612885 samples/sec	 accuracy=65.263235	 loss=1.379887	 lr=0.010000
Epoch[014] Batch [0359]/[1203]	Speed: 174.979747 samples/sec	 accuracy=65.284722	 loss=1.379274	 lr=0.010000
Epoch[014] Batch [0379]/[1203]	Speed: 107.815162 samples/sec	 accuracy=65.301316	 loss=1.379971	 lr=0.010000
Epoch[014] Batch [0399]/[1203]	Speed: 103.182038 samples/sec	 accuracy=65.317500	 loss=1.379399	 lr=0.010000
Epoch[014] Batch [0419]/[1203]	Speed: 192.352224 samples/sec	 accuracy=65.290476	 loss=1.379157	 lr=0.010000
Epoch[014] Batch [0439]/[1203]	Speed: 94.649552 samples/sec	 accuracy=65.306818	 loss=1.379106	 lr=0.010000
Epoch[014] Batch [0459]/[1203]	Speed: 124.156983 samples/sec	 accuracy=65.250000	 loss=1.380936	 lr=0.010000
Epoch[014] Batch [0479]/[1203]	Speed: 139.908324 samples/sec	 accuracy=65.231250	 loss=1.382214	 lr=0.010000
Epoch[014] Batch [0499]/[1203]	Speed: 108.149864 samples/sec	 accuracy=65.193000	 loss=1.383896	 lr=0.010000
Epoch[014] Batch [0519]/[1203]	Speed: 182.390032 samples/sec	 accuracy=65.216346	 loss=1.383881	 lr=0.010000
Epoch[014] Batch [0539]/[1203]	Speed: 101.103070 samples/sec	 accuracy=65.171296	 loss=1.385071	 lr=0.010000
Epoch[014] Batch [0559]/[1203]	Speed: 115.364347 samples/sec	 accuracy=65.191964	 loss=1.383978	 lr=0.010000
Epoch[014] Batch [0579]/[1203]	Speed: 129.605504 samples/sec	 accuracy=65.181034	 loss=1.383565	 lr=0.010000
Epoch[014] Batch [0599]/[1203]	Speed: 111.568949 samples/sec	 accuracy=65.191667	 loss=1.383819	 lr=0.010000
Epoch[014] Batch [0619]/[1203]	Speed: 174.362135 samples/sec	 accuracy=65.166129	 loss=1.385211	 lr=0.010000
Epoch[014] Batch [0639]/[1203]	Speed: 98.915157 samples/sec	 accuracy=65.110156	 loss=1.386908	 lr=0.010000
Epoch[014] Batch [0659]/[1203]	Speed: 121.290225 samples/sec	 accuracy=65.093939	 loss=1.388423	 lr=0.010000
Epoch[014] Batch [0679]/[1203]	Speed: 143.691234 samples/sec	 accuracy=65.096324	 loss=1.388836	 lr=0.010000
Epoch[014] Batch [0699]/[1203]	Speed: 100.859643 samples/sec	 accuracy=65.034286	 loss=1.390528	 lr=0.010000
Epoch[014] Batch [0719]/[1203]	Speed: 133.245461 samples/sec	 accuracy=65.033333	 loss=1.390769	 lr=0.010000
Epoch[014] Batch [0739]/[1203]	Speed: 130.331635 samples/sec	 accuracy=64.999324	 loss=1.391918	 lr=0.010000
Epoch[014] Batch [0759]/[1203]	Speed: 112.411855 samples/sec	 accuracy=64.979605	 loss=1.392373	 lr=0.010000
Epoch[014] Batch [0779]/[1203]	Speed: 143.809628 samples/sec	 accuracy=64.916667	 loss=1.393966	 lr=0.010000
Epoch[014] Batch [0799]/[1203]	Speed: 112.129896 samples/sec	 accuracy=64.888125	 loss=1.394498	 lr=0.010000
Epoch[014] Batch [0819]/[1203]	Speed: 115.053748 samples/sec	 accuracy=64.886585	 loss=1.394758	 lr=0.010000
Epoch[014] Batch [0839]/[1203]	Speed: 140.374809 samples/sec	 accuracy=64.886310	 loss=1.395119	 lr=0.010000
Epoch[014] Batch [0859]/[1203]	Speed: 123.584931 samples/sec	 accuracy=64.856977	 loss=1.397021	 lr=0.010000
Epoch[014] Batch [0879]/[1203]	Speed: 129.399780 samples/sec	 accuracy=64.825000	 loss=1.398242	 lr=0.010000
Epoch[014] Batch [0899]/[1203]	Speed: 111.376978 samples/sec	 accuracy=64.807778	 loss=1.398451	 lr=0.010000
Epoch[014] Batch [0919]/[1203]	Speed: 115.078839 samples/sec	 accuracy=64.800543	 loss=1.398416	 lr=0.010000
Epoch[014] Batch [0939]/[1203]	Speed: 140.950478 samples/sec	 accuracy=64.776596	 loss=1.399008	 lr=0.010000
Epoch[014] Batch [0959]/[1203]	Speed: 105.779984 samples/sec	 accuracy=64.743229	 loss=1.400396	 lr=0.010000
Epoch[014] Batch [0979]/[1203]	Speed: 156.682443 samples/sec	 accuracy=64.704082	 loss=1.401939	 lr=0.010000
Epoch[014] Batch [0999]/[1203]	Speed: 112.260259 samples/sec	 accuracy=64.690000	 loss=1.402293	 lr=0.010000
Epoch[014] Batch [1019]/[1203]	Speed: 119.396429 samples/sec	 accuracy=64.652941	 loss=1.403516	 lr=0.010000
Epoch[014] Batch [1039]/[1203]	Speed: 124.859106 samples/sec	 accuracy=64.645192	 loss=1.404074	 lr=0.010000
Epoch[014] Batch [1059]/[1203]	Speed: 112.497532 samples/sec	 accuracy=64.623113	 loss=1.404919	 lr=0.010000
Epoch[014] Batch [1079]/[1203]	Speed: 153.944528 samples/sec	 accuracy=64.625463	 loss=1.405090	 lr=0.010000
Epoch[014] Batch [1099]/[1203]	Speed: 107.810648 samples/sec	 accuracy=64.633636	 loss=1.404981	 lr=0.010000
Epoch[014] Batch [1119]/[1203]	Speed: 124.612560 samples/sec	 accuracy=64.606250	 loss=1.406361	 lr=0.010000
Epoch[014] Batch [1139]/[1203]	Speed: 132.185225 samples/sec	 accuracy=64.594737	 loss=1.407256	 lr=0.010000
Epoch[014] Batch [1159]/[1203]	Speed: 106.390564 samples/sec	 accuracy=64.593534	 loss=1.407290	 lr=0.010000
Epoch[014] Batch [1179]/[1203]	Speed: 171.584689 samples/sec	 accuracy=64.583898	 loss=1.407806	 lr=0.010000
Epoch[014] Batch [1199]/[1203]	Speed: 197.251907 samples/sec	 accuracy=64.571250	 loss=1.408351	 lr=0.010000
Batch [0019]/[0097]: acc-top1=58.225000 acc-top5=81.350000
Batch [0039]/[0097]: acc-top1=56.587500 acc-top5=80.037500
Batch [0059]/[0097]: acc-top1=59.033333 acc-top5=81.491667
Batch [0079]/[0097]: acc-top1=58.843750 acc-top5=81.625000
[Epoch 014] training: accuracy=64.576475	 loss=1.408279
[Epoch 014] speed: 122 samples/sec	time cost: 2170.081110
[Epoch 014] validation: acc-top1=58.278351 acc-top5=81.613402 loss=1.806187
Epoch[015] Batch [0019]/[1203]	Speed: 68.686593 samples/sec	 accuracy=67.100000	 loss=1.307035	 lr=0.010000
Epoch[015] Batch [0039]/[1203]	Speed: 99.109315 samples/sec	 accuracy=66.575000	 loss=1.311555	 lr=0.010000
Epoch[015] Batch [0059]/[1203]	Speed: 231.221770 samples/sec	 accuracy=66.825000	 loss=1.311249	 lr=0.010000
Epoch[015] Batch [0079]/[1203]	Speed: 101.335983 samples/sec	 accuracy=66.950000	 loss=1.307971	 lr=0.010000
Epoch[015] Batch [0099]/[1203]	Speed: 102.651197 samples/sec	 accuracy=66.775000	 loss=1.316192	 lr=0.010000
Epoch[015] Batch [0119]/[1203]	Speed: 167.876811 samples/sec	 accuracy=66.550000	 loss=1.320804	 lr=0.010000
Epoch[015] Batch [0139]/[1203]	Speed: 113.318631 samples/sec	 accuracy=66.400000	 loss=1.326279	 lr=0.010000
Epoch[015] Batch [0159]/[1203]	Speed: 181.070827 samples/sec	 accuracy=66.356250	 loss=1.328576	 lr=0.010000
Epoch[015] Batch [0179]/[1203]	Speed: 102.442210 samples/sec	 accuracy=66.355556	 loss=1.329467	 lr=0.010000
Epoch[015] Batch [0199]/[1203]	Speed: 105.414393 samples/sec	 accuracy=66.355000	 loss=1.326747	 lr=0.010000
Epoch[015] Batch [0219]/[1203]	Speed: 155.450562 samples/sec	 accuracy=66.231818	 loss=1.329255	 lr=0.010000
Epoch[015] Batch [0239]/[1203]	Speed: 118.863587 samples/sec	 accuracy=66.108333	 loss=1.334567	 lr=0.010000
Epoch[015] Batch [0259]/[1203]	Speed: 123.397760 samples/sec	 accuracy=66.130769	 loss=1.334251	 lr=0.010000
Epoch[015] Batch [0279]/[1203]	Speed: 119.234513 samples/sec	 accuracy=66.051786	 loss=1.334584	 lr=0.010000
Epoch[015] Batch [0299]/[1203]	Speed: 117.111760 samples/sec	 accuracy=66.100000	 loss=1.334811	 lr=0.010000
Epoch[015] Batch [0319]/[1203]	Speed: 141.703696 samples/sec	 accuracy=66.065625	 loss=1.336770	 lr=0.010000
Epoch[015] Batch [0339]/[1203]	Speed: 117.417872 samples/sec	 accuracy=66.029412	 loss=1.339317	 lr=0.010000
Epoch[015] Batch [0359]/[1203]	Speed: 125.637167 samples/sec	 accuracy=65.926389	 loss=1.341525	 lr=0.010000
Epoch[015] Batch [0379]/[1203]	Speed: 119.376725 samples/sec	 accuracy=65.852632	 loss=1.344397	 lr=0.010000
Epoch[015] Batch [0399]/[1203]	Speed: 127.068006 samples/sec	 accuracy=65.805000	 loss=1.346219	 lr=0.010000
Epoch[015] Batch [0419]/[1203]	Speed: 126.096837 samples/sec	 accuracy=65.767857	 loss=1.348866	 lr=0.010000
Epoch[015] Batch [0439]/[1203]	Speed: 114.044039 samples/sec	 accuracy=65.761364	 loss=1.350120	 lr=0.010000
Epoch[015] Batch [0459]/[1203]	Speed: 127.822104 samples/sec	 accuracy=65.733696	 loss=1.351429	 lr=0.010000
Epoch[015] Batch [0479]/[1203]	Speed: 116.749643 samples/sec	 accuracy=65.696875	 loss=1.351178	 lr=0.010000
Epoch[015] Batch [0499]/[1203]	Speed: 145.372872 samples/sec	 accuracy=65.724000	 loss=1.351364	 lr=0.010000
Epoch[015] Batch [0519]/[1203]	Speed: 114.796173 samples/sec	 accuracy=65.728846	 loss=1.352936	 lr=0.010000
Epoch[015] Batch [0539]/[1203]	Speed: 118.157944 samples/sec	 accuracy=65.730556	 loss=1.353488	 lr=0.010000
Epoch[015] Batch [0559]/[1203]	Speed: 134.649262 samples/sec	 accuracy=65.679464	 loss=1.354742	 lr=0.010000
Epoch[015] Batch [0579]/[1203]	Speed: 116.694151 samples/sec	 accuracy=65.677586	 loss=1.355226	 lr=0.010000
Epoch[015] Batch [0599]/[1203]	Speed: 129.321814 samples/sec	 accuracy=65.675833	 loss=1.356231	 lr=0.010000
Epoch[015] Batch [0619]/[1203]	Speed: 118.547784 samples/sec	 accuracy=65.682258	 loss=1.356691	 lr=0.010000
Epoch[015] Batch [0639]/[1203]	Speed: 114.155261 samples/sec	 accuracy=65.670313	 loss=1.357434	 lr=0.010000
Epoch[015] Batch [0659]/[1203]	Speed: 153.584288 samples/sec	 accuracy=65.648485	 loss=1.358868	 lr=0.010000
Epoch[015] Batch [0679]/[1203]	Speed: 106.257981 samples/sec	 accuracy=65.602206	 loss=1.360167	 lr=0.010000
Epoch[015] Batch [0699]/[1203]	Speed: 125.858365 samples/sec	 accuracy=65.577143	 loss=1.361874	 lr=0.010000
Epoch[015] Batch [0719]/[1203]	Speed: 133.767501 samples/sec	 accuracy=65.529861	 loss=1.363257	 lr=0.010000
Epoch[015] Batch [0739]/[1203]	Speed: 118.953670 samples/sec	 accuracy=65.504054	 loss=1.363545	 lr=0.010000
Epoch[015] Batch [0759]/[1203]	Speed: 135.770195 samples/sec	 accuracy=65.540132	 loss=1.362987	 lr=0.010000
Epoch[015] Batch [0779]/[1203]	Speed: 111.550402 samples/sec	 accuracy=65.519872	 loss=1.363771	 lr=0.010000
Epoch[015] Batch [0799]/[1203]	Speed: 119.282646 samples/sec	 accuracy=65.506250	 loss=1.364616	 lr=0.010000
Epoch[015] Batch [0819]/[1203]	Speed: 139.603941 samples/sec	 accuracy=65.485366	 loss=1.365683	 lr=0.010000
Epoch[015] Batch [0839]/[1203]	Speed: 123.819855 samples/sec	 accuracy=65.447619	 loss=1.367351	 lr=0.010000
Epoch[015] Batch [0859]/[1203]	Speed: 131.358377 samples/sec	 accuracy=65.412209	 loss=1.368626	 lr=0.010000
Epoch[015] Batch [0879]/[1203]	Speed: 103.079343 samples/sec	 accuracy=65.414205	 loss=1.369385	 lr=0.010000
Epoch[015] Batch [0899]/[1203]	Speed: 145.116650 samples/sec	 accuracy=65.396111	 loss=1.369914	 lr=0.010000
Epoch[015] Batch [0919]/[1203]	Speed: 128.612887 samples/sec	 accuracy=65.380435	 loss=1.370570	 lr=0.010000
Epoch[015] Batch [0939]/[1203]	Speed: 98.991281 samples/sec	 accuracy=65.328723	 loss=1.372704	 lr=0.010000
Epoch[015] Batch [0959]/[1203]	Speed: 144.014437 samples/sec	 accuracy=65.302604	 loss=1.373668	 lr=0.010000
Epoch[015] Batch [0979]/[1203]	Speed: 116.160549 samples/sec	 accuracy=65.278571	 loss=1.374183	 lr=0.010000
Epoch[015] Batch [0999]/[1203]	Speed: 162.898757 samples/sec	 accuracy=65.257500	 loss=1.374706	 lr=0.010000
Epoch[015] Batch [1019]/[1203]	Speed: 117.152352 samples/sec	 accuracy=65.215686	 loss=1.375704	 lr=0.010000
Epoch[015] Batch [1039]/[1203]	Speed: 103.664774 samples/sec	 accuracy=65.198558	 loss=1.375920	 lr=0.010000
Epoch[015] Batch [1059]/[1203]	Speed: 154.713228 samples/sec	 accuracy=65.195755	 loss=1.376453	 lr=0.010000
Epoch[015] Batch [1079]/[1203]	Speed: 98.313432 samples/sec	 accuracy=65.181481	 loss=1.376951	 lr=0.010000
Epoch[015] Batch [1099]/[1203]	Speed: 204.261457 samples/sec	 accuracy=65.178182	 loss=1.377169	 lr=0.010000
Epoch[015] Batch [1119]/[1203]	Speed: 106.421295 samples/sec	 accuracy=65.168750	 loss=1.377780	 lr=0.010000
Epoch[015] Batch [1139]/[1203]	Speed: 115.311099 samples/sec	 accuracy=65.161404	 loss=1.377993	 lr=0.010000
Epoch[015] Batch [1159]/[1203]	Speed: 144.003465 samples/sec	 accuracy=65.155172	 loss=1.378170	 lr=0.010000
Epoch[015] Batch [1179]/[1203]	Speed: 102.771903 samples/sec	 accuracy=65.165678	 loss=1.378038	 lr=0.010000
Epoch[015] Batch [1199]/[1203]	Speed: 342.458950 samples/sec	 accuracy=65.152500	 loss=1.379044	 lr=0.010000
Batch [0019]/[0097]: acc-top1=58.950000 acc-top5=81.975000
Batch [0039]/[0097]: acc-top1=56.925000 acc-top5=80.400000
Batch [0059]/[0097]: acc-top1=59.733333 acc-top5=82.033333
Batch [0079]/[0097]: acc-top1=59.687500 acc-top5=82.125000
[Epoch 015] training: accuracy=65.158354	 loss=1.378987
[Epoch 015] speed: 123 samples/sec	time cost: 2142.162158
[Epoch 015] validation: acc-top1=58.824742 acc-top5=82.051546 loss=1.754666
Epoch[016] Batch [0019]/[1203]	Speed: 68.577812 samples/sec	 accuracy=65.900000	 loss=1.345318	 lr=0.010000
Epoch[016] Batch [0039]/[1203]	Speed: 97.224506 samples/sec	 accuracy=66.312500	 loss=1.324900	 lr=0.010000
Epoch[016] Batch [0059]/[1203]	Speed: 236.076592 samples/sec	 accuracy=66.441667	 loss=1.324770	 lr=0.010000
Epoch[016] Batch [0079]/[1203]	Speed: 96.801605 samples/sec	 accuracy=66.656250	 loss=1.319461	 lr=0.010000
Epoch[016] Batch [0099]/[1203]	Speed: 108.269242 samples/sec	 accuracy=66.555000	 loss=1.320886	 lr=0.010000
Epoch[016] Batch [0119]/[1203]	Speed: 182.250344 samples/sec	 accuracy=66.545833	 loss=1.317980	 lr=0.010000
Epoch[016] Batch [0139]/[1203]	Speed: 97.299265 samples/sec	 accuracy=66.514286	 loss=1.315115	 lr=0.010000
Epoch[016] Batch [0159]/[1203]	Speed: 214.838422 samples/sec	 accuracy=66.375000	 loss=1.323287	 lr=0.010000
Epoch[016] Batch [0179]/[1203]	Speed: 98.653988 samples/sec	 accuracy=66.522222	 loss=1.317856	 lr=0.010000
Epoch[016] Batch [0199]/[1203]	Speed: 111.572517 samples/sec	 accuracy=66.552500	 loss=1.317036	 lr=0.010000
Epoch[016] Batch [0219]/[1203]	Speed: 168.957417 samples/sec	 accuracy=66.477273	 loss=1.317049	 lr=0.010000
Epoch[016] Batch [0239]/[1203]	Speed: 101.708032 samples/sec	 accuracy=66.504167	 loss=1.316720	 lr=0.010000
Epoch[016] Batch [0259]/[1203]	Speed: 111.099441 samples/sec	 accuracy=66.548077	 loss=1.320044	 lr=0.010000
Epoch[016] Batch [0279]/[1203]	Speed: 160.503094 samples/sec	 accuracy=66.575000	 loss=1.319592	 lr=0.010000
Epoch[016] Batch [0299]/[1203]	Speed: 107.586823 samples/sec	 accuracy=66.546667	 loss=1.321867	 lr=0.010000
Epoch[016] Batch [0319]/[1203]	Speed: 143.366229 samples/sec	 accuracy=66.543750	 loss=1.321403	 lr=0.010000
Epoch[016] Batch [0339]/[1203]	Speed: 115.428608 samples/sec	 accuracy=66.520588	 loss=1.322681	 lr=0.010000
Epoch[016] Batch [0359]/[1203]	Speed: 114.185020 samples/sec	 accuracy=66.519444	 loss=1.324290	 lr=0.010000
Epoch[016] Batch [0379]/[1203]	Speed: 145.305844 samples/sec	 accuracy=66.503947	 loss=1.325239	 lr=0.010000
Epoch[016] Batch [0399]/[1203]	Speed: 119.545610 samples/sec	 accuracy=66.481250	 loss=1.326648	 lr=0.010000
Epoch[016] Batch [0419]/[1203]	Speed: 121.939358 samples/sec	 accuracy=66.436905	 loss=1.328370	 lr=0.010000
Epoch[016] Batch [0439]/[1203]	Speed: 117.587792 samples/sec	 accuracy=66.380682	 loss=1.329788	 lr=0.010000
Epoch[016] Batch [0459]/[1203]	Speed: 129.254039 samples/sec	 accuracy=66.364130	 loss=1.330510	 lr=0.010000
Epoch[016] Batch [0479]/[1203]	Speed: 133.420354 samples/sec	 accuracy=66.327083	 loss=1.332256	 lr=0.010000
Epoch[016] Batch [0499]/[1203]	Speed: 99.685723 samples/sec	 accuracy=66.252000	 loss=1.335224	 lr=0.010000
Epoch[016] Batch [0519]/[1203]	Speed: 160.516759 samples/sec	 accuracy=66.256731	 loss=1.335741	 lr=0.010000
Epoch[016] Batch [0539]/[1203]	Speed: 107.630930 samples/sec	 accuracy=66.251852	 loss=1.336425	 lr=0.010000
Epoch[016] Batch [0559]/[1203]	Speed: 144.504432 samples/sec	 accuracy=66.194643	 loss=1.338352	 lr=0.010000
Epoch[016] Batch [0579]/[1203]	Speed: 121.628571 samples/sec	 accuracy=66.173276	 loss=1.339434	 lr=0.010000
Epoch[016] Batch [0599]/[1203]	Speed: 105.869980 samples/sec	 accuracy=66.155000	 loss=1.340471	 lr=0.010000
Epoch[016] Batch [0619]/[1203]	Speed: 154.806337 samples/sec	 accuracy=66.090323	 loss=1.342525	 lr=0.010000
Epoch[016] Batch [0639]/[1203]	Speed: 116.725521 samples/sec	 accuracy=66.082813	 loss=1.343239	 lr=0.010000
Epoch[016] Batch [0659]/[1203]	Speed: 99.921550 samples/sec	 accuracy=66.039394	 loss=1.344475	 lr=0.010000
Epoch[016] Batch [0679]/[1203]	Speed: 192.444126 samples/sec	 accuracy=66.023529	 loss=1.344070	 lr=0.010000
Epoch[016] Batch [0699]/[1203]	Speed: 97.946090 samples/sec	 accuracy=66.002143	 loss=1.344133	 lr=0.010000
Epoch[016] Batch [0719]/[1203]	Speed: 148.865048 samples/sec	 accuracy=65.974306	 loss=1.345164	 lr=0.010000
Epoch[016] Batch [0739]/[1203]	Speed: 121.739056 samples/sec	 accuracy=65.980405	 loss=1.345164	 lr=0.010000
Epoch[016] Batch [0759]/[1203]	Speed: 101.629373 samples/sec	 accuracy=65.975658	 loss=1.345737	 lr=0.010000
Epoch[016] Batch [0779]/[1203]	Speed: 200.524189 samples/sec	 accuracy=65.962179	 loss=1.346597	 lr=0.010000
Epoch[016] Batch [0799]/[1203]	Speed: 90.515292 samples/sec	 accuracy=65.930000	 loss=1.347478	 lr=0.010000
Epoch[016] Batch [0819]/[1203]	Speed: 182.484447 samples/sec	 accuracy=65.868902	 loss=1.348997	 lr=0.010000
Epoch[016] Batch [0839]/[1203]	Speed: 106.873227 samples/sec	 accuracy=65.862500	 loss=1.349217	 lr=0.010000
Epoch[016] Batch [0859]/[1203]	Speed: 115.582676 samples/sec	 accuracy=65.837791	 loss=1.350102	 lr=0.010000
Epoch[016] Batch [0879]/[1203]	Speed: 169.213473 samples/sec	 accuracy=65.828977	 loss=1.351199	 lr=0.010000
Epoch[016] Batch [0899]/[1203]	Speed: 95.471438 samples/sec	 accuracy=65.810556	 loss=1.351816	 lr=0.010000
Epoch[016] Batch [0919]/[1203]	Speed: 118.565400 samples/sec	 accuracy=65.797283	 loss=1.352588	 lr=0.010000
Epoch[016] Batch [0939]/[1203]	Speed: 144.772640 samples/sec	 accuracy=65.783511	 loss=1.353534	 lr=0.010000
Epoch[016] Batch [0959]/[1203]	Speed: 112.945909 samples/sec	 accuracy=65.789583	 loss=1.353558	 lr=0.010000
Epoch[016] Batch [0979]/[1203]	Speed: 145.924589 samples/sec	 accuracy=65.769388	 loss=1.354472	 lr=0.010000
Epoch[016] Batch [0999]/[1203]	Speed: 108.109560 samples/sec	 accuracy=65.770500	 loss=1.354947	 lr=0.010000
Epoch[016] Batch [1019]/[1203]	Speed: 133.157819 samples/sec	 accuracy=65.737255	 loss=1.356050	 lr=0.010000
Epoch[016] Batch [1039]/[1203]	Speed: 128.805449 samples/sec	 accuracy=65.765385	 loss=1.355267	 lr=0.010000
Epoch[016] Batch [1059]/[1203]	Speed: 120.214276 samples/sec	 accuracy=65.758962	 loss=1.355408	 lr=0.010000
Epoch[016] Batch [1079]/[1203]	Speed: 130.087218 samples/sec	 accuracy=65.750926	 loss=1.355811	 lr=0.010000
Epoch[016] Batch [1099]/[1203]	Speed: 115.956800 samples/sec	 accuracy=65.720000	 loss=1.356748	 lr=0.010000
Epoch[016] Batch [1119]/[1203]	Speed: 118.186485 samples/sec	 accuracy=65.696429	 loss=1.357925	 lr=0.010000
Epoch[016] Batch [1139]/[1203]	Speed: 142.590592 samples/sec	 accuracy=65.700877	 loss=1.357867	 lr=0.010000
Epoch[016] Batch [1159]/[1203]	Speed: 122.650074 samples/sec	 accuracy=65.709914	 loss=1.357448	 lr=0.010000
Epoch[016] Batch [1179]/[1203]	Speed: 118.220341 samples/sec	 accuracy=65.704661	 loss=1.357610	 lr=0.010000
Epoch[016] Batch [1199]/[1203]	Speed: 283.050485 samples/sec	 accuracy=65.696667	 loss=1.357764	 lr=0.010000
Batch [0019]/[0097]: acc-top1=57.900000 acc-top5=80.950000
Batch [0039]/[0097]: acc-top1=57.575000 acc-top5=80.387500
Batch [0059]/[0097]: acc-top1=59.766667 acc-top5=82.158333
Batch [0079]/[0097]: acc-top1=59.418750 acc-top5=81.975000
[Epoch 016] training: accuracy=65.692020	 loss=1.357978
[Epoch 016] speed: 123 samples/sec	time cost: 2144.152531
[Epoch 016] validation: acc-top1=58.690722 acc-top5=81.690722 loss=1.772448
Epoch[017] Batch [0019]/[1203]	Speed: 68.540677 samples/sec	 accuracy=67.175000	 loss=1.280388	 lr=0.010000
Epoch[017] Batch [0039]/[1203]	Speed: 101.573636 samples/sec	 accuracy=67.462500	 loss=1.281386	 lr=0.010000
Epoch[017] Batch [0059]/[1203]	Speed: 212.929090 samples/sec	 accuracy=67.175000	 loss=1.286661	 lr=0.010000
Epoch[017] Batch [0079]/[1203]	Speed: 96.741721 samples/sec	 accuracy=67.456250	 loss=1.282138	 lr=0.010000
Epoch[017] Batch [0099]/[1203]	Speed: 109.405046 samples/sec	 accuracy=67.445000	 loss=1.281860	 lr=0.010000
Epoch[017] Batch [0119]/[1203]	Speed: 163.802812 samples/sec	 accuracy=67.387500	 loss=1.280143	 lr=0.010000
Epoch[017] Batch [0139]/[1203]	Speed: 107.914690 samples/sec	 accuracy=67.282143	 loss=1.286590	 lr=0.010000
Epoch[017] Batch [0159]/[1203]	Speed: 186.995995 samples/sec	 accuracy=67.137500	 loss=1.293460	 lr=0.010000
Epoch[017] Batch [0179]/[1203]	Speed: 100.018515 samples/sec	 accuracy=67.094444	 loss=1.291827	 lr=0.010000
Epoch[017] Batch [0199]/[1203]	Speed: 114.998637 samples/sec	 accuracy=67.025000	 loss=1.296403	 lr=0.010000
Epoch[017] Batch [0219]/[1203]	Speed: 141.392008 samples/sec	 accuracy=67.056818	 loss=1.296104	 lr=0.010000
Epoch[017] Batch [0239]/[1203]	Speed: 117.522436 samples/sec	 accuracy=67.020833	 loss=1.295931	 lr=0.010000
Epoch[017] Batch [0259]/[1203]	Speed: 129.742401 samples/sec	 accuracy=66.994231	 loss=1.298542	 lr=0.010000
Epoch[017] Batch [0279]/[1203]	Speed: 130.192810 samples/sec	 accuracy=66.841071	 loss=1.305113	 lr=0.010000
Epoch[017] Batch [0299]/[1203]	Speed: 114.645888 samples/sec	 accuracy=66.805000	 loss=1.307619	 lr=0.010000
Epoch[017] Batch [0319]/[1203]	Speed: 148.388635 samples/sec	 accuracy=66.753125	 loss=1.307905	 lr=0.010000
Epoch[017] Batch [0339]/[1203]	Speed: 104.829599 samples/sec	 accuracy=66.700000	 loss=1.310587	 lr=0.010000
Epoch[017] Batch [0359]/[1203]	Speed: 137.124012 samples/sec	 accuracy=66.701389	 loss=1.311198	 lr=0.010000
Epoch[017] Batch [0379]/[1203]	Speed: 122.153120 samples/sec	 accuracy=66.682895	 loss=1.310808	 lr=0.010000
Epoch[017] Batch [0399]/[1203]	Speed: 118.897858 samples/sec	 accuracy=66.626250	 loss=1.312472	 lr=0.010000
Epoch[017] Batch [0419]/[1203]	Speed: 140.124189 samples/sec	 accuracy=66.596429	 loss=1.313848	 lr=0.010000
Epoch[017] Batch [0439]/[1203]	Speed: 104.352343 samples/sec	 accuracy=66.561364	 loss=1.314763	 lr=0.010000
Epoch[017] Batch [0459]/[1203]	Speed: 139.575923 samples/sec	 accuracy=66.534783	 loss=1.317396	 lr=0.010000
Epoch[017] Batch [0479]/[1203]	Speed: 119.347701 samples/sec	 accuracy=66.482292	 loss=1.318605	 lr=0.010000
Epoch[017] Batch [0499]/[1203]	Speed: 110.802096 samples/sec	 accuracy=66.492000	 loss=1.318575	 lr=0.010000
Epoch[017] Batch [0519]/[1203]	Speed: 154.692601 samples/sec	 accuracy=66.520192	 loss=1.317233	 lr=0.010000
Epoch[017] Batch [0539]/[1203]	Speed: 106.765809 samples/sec	 accuracy=66.484259	 loss=1.318431	 lr=0.010000
Epoch[017] Batch [0559]/[1203]	Speed: 151.422398 samples/sec	 accuracy=66.479464	 loss=1.319195	 lr=0.010000
Epoch[017] Batch [0579]/[1203]	Speed: 114.515896 samples/sec	 accuracy=66.523276	 loss=1.319574	 lr=0.010000
Epoch[017] Batch [0599]/[1203]	Speed: 116.167429 samples/sec	 accuracy=66.498333	 loss=1.319452	 lr=0.010000
Epoch[017] Batch [0619]/[1203]	Speed: 144.742177 samples/sec	 accuracy=66.483065	 loss=1.320399	 lr=0.010000
Epoch[017] Batch [0639]/[1203]	Speed: 100.353278 samples/sec	 accuracy=66.511719	 loss=1.318793	 lr=0.010000
Epoch[017] Batch [0659]/[1203]	Speed: 151.205511 samples/sec	 accuracy=66.477273	 loss=1.320096	 lr=0.010000
Epoch[017] Batch [0679]/[1203]	Speed: 119.457807 samples/sec	 accuracy=66.488971	 loss=1.319409	 lr=0.010000
Epoch[017] Batch [0699]/[1203]	Speed: 116.357070 samples/sec	 accuracy=66.450000	 loss=1.321150	 lr=0.010000
Epoch[017] Batch [0719]/[1203]	Speed: 153.844225 samples/sec	 accuracy=66.435417	 loss=1.323030	 lr=0.010000
Epoch[017] Batch [0739]/[1203]	Speed: 101.085656 samples/sec	 accuracy=66.427027	 loss=1.323547	 lr=0.010000
Epoch[017] Batch [0759]/[1203]	Speed: 132.475272 samples/sec	 accuracy=66.400000	 loss=1.323819	 lr=0.010000
Epoch[017] Batch [0779]/[1203]	Speed: 124.529438 samples/sec	 accuracy=66.371795	 loss=1.325254	 lr=0.010000
Epoch[017] Batch [0799]/[1203]	Speed: 118.295306 samples/sec	 accuracy=66.338750	 loss=1.326889	 lr=0.010000
Epoch[017] Batch [0819]/[1203]	Speed: 152.269373 samples/sec	 accuracy=66.325610	 loss=1.327362	 lr=0.010000
Epoch[017] Batch [0839]/[1203]	Speed: 114.138441 samples/sec	 accuracy=66.301786	 loss=1.328372	 lr=0.010000
Epoch[017] Batch [0859]/[1203]	Speed: 119.610444 samples/sec	 accuracy=66.247093	 loss=1.329932	 lr=0.010000
Epoch[017] Batch [0879]/[1203]	Speed: 124.573709 samples/sec	 accuracy=66.272727	 loss=1.328763	 lr=0.010000
Epoch[017] Batch [0899]/[1203]	Speed: 113.057184 samples/sec	 accuracy=66.242222	 loss=1.330205	 lr=0.010000
Epoch[017] Batch [0919]/[1203]	Speed: 147.562518 samples/sec	 accuracy=66.225543	 loss=1.330450	 lr=0.010000
Epoch[017] Batch [0939]/[1203]	Speed: 114.480839 samples/sec	 accuracy=66.195213	 loss=1.332093	 lr=0.010000
Epoch[017] Batch [0959]/[1203]	Speed: 126.245412 samples/sec	 accuracy=66.181771	 loss=1.332842	 lr=0.010000
Epoch[017] Batch [0979]/[1203]	Speed: 130.534040 samples/sec	 accuracy=66.178061	 loss=1.332832	 lr=0.010000
Epoch[017] Batch [0999]/[1203]	Speed: 103.124941 samples/sec	 accuracy=66.174000	 loss=1.333068	 lr=0.010000
Epoch[017] Batch [1019]/[1203]	Speed: 174.024839 samples/sec	 accuracy=66.162255	 loss=1.333462	 lr=0.010000
Epoch[017] Batch [1039]/[1203]	Speed: 103.610720 samples/sec	 accuracy=66.177404	 loss=1.333174	 lr=0.010000
Epoch[017] Batch [1059]/[1203]	Speed: 128.935344 samples/sec	 accuracy=66.148113	 loss=1.333712	 lr=0.010000
Epoch[017] Batch [1079]/[1203]	Speed: 127.516983 samples/sec	 accuracy=66.135185	 loss=1.334647	 lr=0.010000
Epoch[017] Batch [1099]/[1203]	Speed: 108.750717 samples/sec	 accuracy=66.118636	 loss=1.335253	 lr=0.010000
Epoch[017] Batch [1119]/[1203]	Speed: 155.775362 samples/sec	 accuracy=66.102679	 loss=1.336020	 lr=0.010000
Epoch[017] Batch [1139]/[1203]	Speed: 118.287254 samples/sec	 accuracy=66.079386	 loss=1.336852	 lr=0.010000
Epoch[017] Batch [1159]/[1203]	Speed: 115.862346 samples/sec	 accuracy=66.068534	 loss=1.337170	 lr=0.010000
Epoch[017] Batch [1179]/[1203]	Speed: 162.234564 samples/sec	 accuracy=66.044068	 loss=1.338310	 lr=0.010000
Epoch[017] Batch [1199]/[1203]	Speed: 190.725937 samples/sec	 accuracy=66.019583	 loss=1.339574	 lr=0.010000
Batch [0019]/[0097]: acc-top1=57.175000 acc-top5=80.125000
Batch [0039]/[0097]: acc-top1=57.075000 acc-top5=80.150000
Batch [0059]/[0097]: acc-top1=59.550000 acc-top5=81.725000
Batch [0079]/[0097]: acc-top1=59.481250 acc-top5=81.643750
[Epoch 017] training: accuracy=66.013716	 loss=1.339654
[Epoch 017] speed: 123 samples/sec	time cost: 2143.144143
[Epoch 017] validation: acc-top1=58.737113 acc-top5=81.546392 loss=1.792526
Epoch[018] Batch [0019]/[1203]	Speed: 68.030578 samples/sec	 accuracy=66.775000	 loss=1.275566	 lr=0.010000
Epoch[018] Batch [0039]/[1203]	Speed: 101.343492 samples/sec	 accuracy=67.425000	 loss=1.272464	 lr=0.010000
Epoch[018] Batch [0059]/[1203]	Speed: 226.037278 samples/sec	 accuracy=67.483333	 loss=1.271439	 lr=0.010000
Epoch[018] Batch [0079]/[1203]	Speed: 90.557526 samples/sec	 accuracy=67.562500	 loss=1.265211	 lr=0.010000
Epoch[018] Batch [0099]/[1203]	Speed: 121.792361 samples/sec	 accuracy=67.595000	 loss=1.261048	 lr=0.010000
Epoch[018] Batch [0119]/[1203]	Speed: 161.532684 samples/sec	 accuracy=67.450000	 loss=1.268941	 lr=0.010000
Epoch[018] Batch [0139]/[1203]	Speed: 110.154892 samples/sec	 accuracy=67.335714	 loss=1.276574	 lr=0.010000
Epoch[018] Batch [0159]/[1203]	Speed: 162.395409 samples/sec	 accuracy=67.218750	 loss=1.283185	 lr=0.010000
Epoch[018] Batch [0179]/[1203]	Speed: 96.747968 samples/sec	 accuracy=67.172222	 loss=1.286592	 lr=0.010000
Epoch[018] Batch [0199]/[1203]	Speed: 124.059046 samples/sec	 accuracy=67.212500	 loss=1.288835	 lr=0.010000
Epoch[018] Batch [0219]/[1203]	Speed: 146.406503 samples/sec	 accuracy=67.145455	 loss=1.290431	 lr=0.010000
Epoch[018] Batch [0239]/[1203]	Speed: 107.549770 samples/sec	 accuracy=67.135417	 loss=1.290044	 lr=0.010000
Epoch[018] Batch [0259]/[1203]	Speed: 131.786546 samples/sec	 accuracy=67.178846	 loss=1.290520	 lr=0.010000
Epoch[018] Batch [0279]/[1203]	Speed: 131.328883 samples/sec	 accuracy=67.196429	 loss=1.292309	 lr=0.010000
Epoch[018] Batch [0299]/[1203]	Speed: 121.057937 samples/sec	 accuracy=67.143333	 loss=1.293679	 lr=0.010000
Epoch[018] Batch [0319]/[1203]	Speed: 128.234747 samples/sec	 accuracy=67.095313	 loss=1.296056	 lr=0.010000
Epoch[018] Batch [0339]/[1203]	Speed: 106.119766 samples/sec	 accuracy=67.139706	 loss=1.294542	 lr=0.010000
Epoch[018] Batch [0359]/[1203]	Speed: 161.244147 samples/sec	 accuracy=67.120833	 loss=1.295466	 lr=0.010000
Epoch[018] Batch [0379]/[1203]	Speed: 113.056466 samples/sec	 accuracy=67.063158	 loss=1.297823	 lr=0.010000
Epoch[018] Batch [0399]/[1203]	Speed: 114.278340 samples/sec	 accuracy=67.086250	 loss=1.297543	 lr=0.010000
Epoch[018] Batch [0419]/[1203]	Speed: 156.141593 samples/sec	 accuracy=67.116667	 loss=1.295881	 lr=0.010000
Epoch[018] Batch [0439]/[1203]	Speed: 113.938614 samples/sec	 accuracy=67.128409	 loss=1.296215	 lr=0.010000
Epoch[018] Batch [0459]/[1203]	Speed: 128.777633 samples/sec	 accuracy=67.105435	 loss=1.297330	 lr=0.010000
Epoch[018] Batch [0479]/[1203]	Speed: 117.120736 samples/sec	 accuracy=67.041667	 loss=1.299774	 lr=0.010000
Epoch[018] Batch [0499]/[1203]	Speed: 116.912679 samples/sec	 accuracy=67.037000	 loss=1.300031	 lr=0.010000
Epoch[018] Batch [0519]/[1203]	Speed: 133.846055 samples/sec	 accuracy=67.000000	 loss=1.300977	 lr=0.010000
Epoch[018] Batch [0539]/[1203]	Speed: 113.801276 samples/sec	 accuracy=66.943519	 loss=1.305420	 lr=0.010000
Epoch[018] Batch [0559]/[1203]	Speed: 125.440489 samples/sec	 accuracy=66.928571	 loss=1.306858	 lr=0.010000
Epoch[018] Batch [0579]/[1203]	Speed: 125.038065 samples/sec	 accuracy=66.875862	 loss=1.308207	 lr=0.010000
Epoch[018] Batch [0599]/[1203]	Speed: 136.762365 samples/sec	 accuracy=66.869167	 loss=1.308656	 lr=0.010000
Epoch[018] Batch [0619]/[1203]	Speed: 136.899544 samples/sec	 accuracy=66.837903	 loss=1.309829	 lr=0.010000
Epoch[018] Batch [0639]/[1203]	Speed: 100.030266 samples/sec	 accuracy=66.830469	 loss=1.310614	 lr=0.010000
Epoch[018] Batch [0659]/[1203]	Speed: 137.646022 samples/sec	 accuracy=66.783333	 loss=1.311753	 lr=0.010000
Epoch[018] Batch [0679]/[1203]	Speed: 113.755980 samples/sec	 accuracy=66.789706	 loss=1.311084	 lr=0.010000
Epoch[018] Batch [0699]/[1203]	Speed: 134.938965 samples/sec	 accuracy=66.769286	 loss=1.311333	 lr=0.010000
Epoch[018] Batch [0719]/[1203]	Speed: 140.216887 samples/sec	 accuracy=66.733333	 loss=1.312794	 lr=0.010000
Epoch[018] Batch [0739]/[1203]	Speed: 105.295120 samples/sec	 accuracy=66.726351	 loss=1.313516	 lr=0.010000
Epoch[018] Batch [0759]/[1203]	Speed: 141.278510 samples/sec	 accuracy=66.725000	 loss=1.312996	 lr=0.010000
Epoch[018] Batch [0779]/[1203]	Speed: 121.941982 samples/sec	 accuracy=66.694231	 loss=1.314096	 lr=0.010000
Epoch[018] Batch [0799]/[1203]	Speed: 101.948662 samples/sec	 accuracy=66.674375	 loss=1.314553	 lr=0.010000
Epoch[018] Batch [0819]/[1203]	Speed: 164.203461 samples/sec	 accuracy=66.641463	 loss=1.316066	 lr=0.010000
Epoch[018] Batch [0839]/[1203]	Speed: 116.391601 samples/sec	 accuracy=66.619048	 loss=1.317167	 lr=0.010000
Epoch[018] Batch [0859]/[1203]	Speed: 151.322314 samples/sec	 accuracy=66.595930	 loss=1.317535	 lr=0.010000
Epoch[018] Batch [0879]/[1203]	Speed: 106.931639 samples/sec	 accuracy=66.604545	 loss=1.317675	 lr=0.010000
Epoch[018] Batch [0899]/[1203]	Speed: 112.102747 samples/sec	 accuracy=66.591111	 loss=1.318110	 lr=0.010000
Epoch[018] Batch [0919]/[1203]	Speed: 156.899386 samples/sec	 accuracy=66.575543	 loss=1.318350	 lr=0.010000
Epoch[018] Batch [0939]/[1203]	Speed: 98.671611 samples/sec	 accuracy=66.580851	 loss=1.318413	 lr=0.010000
Epoch[018] Batch [0959]/[1203]	Speed: 135.856632 samples/sec	 accuracy=66.572917	 loss=1.318694	 lr=0.010000
Epoch[018] Batch [0979]/[1203]	Speed: 125.287860 samples/sec	 accuracy=66.571939	 loss=1.319260	 lr=0.010000
Epoch[018] Batch [0999]/[1203]	Speed: 124.305756 samples/sec	 accuracy=66.580000	 loss=1.318680	 lr=0.010000
Epoch[018] Batch [1019]/[1203]	Speed: 146.314922 samples/sec	 accuracy=66.559804	 loss=1.319036	 lr=0.010000
Epoch[018] Batch [1039]/[1203]	Speed: 101.023160 samples/sec	 accuracy=66.556250	 loss=1.319500	 lr=0.010000
Epoch[018] Batch [1059]/[1203]	Speed: 131.934301 samples/sec	 accuracy=66.528774	 loss=1.320389	 lr=0.010000
Epoch[018] Batch [1079]/[1203]	Speed: 129.408470 samples/sec	 accuracy=66.517593	 loss=1.320763	 lr=0.010000
Epoch[018] Batch [1099]/[1203]	Speed: 114.449325 samples/sec	 accuracy=66.513636	 loss=1.320764	 lr=0.010000
Epoch[018] Batch [1119]/[1203]	Speed: 158.572721 samples/sec	 accuracy=66.494643	 loss=1.321376	 lr=0.010000
Epoch[018] Batch [1139]/[1203]	Speed: 96.186240 samples/sec	 accuracy=66.492544	 loss=1.321689	 lr=0.010000
Epoch[018] Batch [1159]/[1203]	Speed: 136.902919 samples/sec	 accuracy=66.485345	 loss=1.322131	 lr=0.010000
Epoch[018] Batch [1179]/[1203]	Speed: 133.405473 samples/sec	 accuracy=66.452542	 loss=1.323372	 lr=0.010000
Epoch[018] Batch [1199]/[1203]	Speed: 246.847504 samples/sec	 accuracy=66.429583	 loss=1.324285	 lr=0.010000
Batch [0019]/[0097]: acc-top1=57.050000 acc-top5=80.975000
Batch [0039]/[0097]: acc-top1=56.937500 acc-top5=80.350000
Batch [0059]/[0097]: acc-top1=59.550000 acc-top5=81.825000
Batch [0079]/[0097]: acc-top1=59.362500 acc-top5=81.625000
[Epoch 018] training: accuracy=66.433084	 loss=1.324193
[Epoch 018] speed: 123 samples/sec	time cost: 2143.022117
[Epoch 018] validation: acc-top1=58.783505 acc-top5=81.659794 loss=1.790354
Epoch[019] Batch [0019]/[1203]	Speed: 67.899496 samples/sec	 accuracy=66.575000	 loss=1.281880	 lr=0.010000
Epoch[019] Batch [0039]/[1203]	Speed: 97.195044 samples/sec	 accuracy=67.625000	 loss=1.262197	 lr=0.010000
Epoch[019] Batch [0059]/[1203]	Speed: 239.665640 samples/sec	 accuracy=67.725000	 loss=1.266375	 lr=0.010000
Epoch[019] Batch [0079]/[1203]	Speed: 95.305191 samples/sec	 accuracy=67.843750	 loss=1.263490	 lr=0.010000
Epoch[019] Batch [0099]/[1203]	Speed: 112.699883 samples/sec	 accuracy=67.905000	 loss=1.267052	 lr=0.010000
Epoch[019] Batch [0119]/[1203]	Speed: 157.427597 samples/sec	 accuracy=67.783333	 loss=1.269442	 lr=0.010000
Epoch[019] Batch [0139]/[1203]	Speed: 107.599415 samples/sec	 accuracy=67.935714	 loss=1.263265	 lr=0.010000
Epoch[019] Batch [0159]/[1203]	Speed: 192.571034 samples/sec	 accuracy=68.000000	 loss=1.262504	 lr=0.010000
Epoch[019] Batch [0179]/[1203]	Speed: 99.029068 samples/sec	 accuracy=67.902778	 loss=1.264154	 lr=0.010000
Epoch[019] Batch [0199]/[1203]	Speed: 103.839072 samples/sec	 accuracy=67.945000	 loss=1.263212	 lr=0.010000
Epoch[019] Batch [0219]/[1203]	Speed: 178.018112 samples/sec	 accuracy=68.004545	 loss=1.259845	 lr=0.010000
Epoch[019] Batch [0239]/[1203]	Speed: 112.989669 samples/sec	 accuracy=67.970833	 loss=1.261134	 lr=0.010000
Epoch[019] Batch [0259]/[1203]	Speed: 135.109999 samples/sec	 accuracy=67.898077	 loss=1.264561	 lr=0.010000
Epoch[019] Batch [0279]/[1203]	Speed: 129.248528 samples/sec	 accuracy=67.883929	 loss=1.262098	 lr=0.010000
Epoch[019] Batch [0299]/[1203]	Speed: 99.530441 samples/sec	 accuracy=67.911667	 loss=1.262909	 lr=0.010000
Epoch[019] Batch [0319]/[1203]	Speed: 173.728954 samples/sec	 accuracy=67.817188	 loss=1.265906	 lr=0.010000
Epoch[019] Batch [0339]/[1203]	Speed: 110.281273 samples/sec	 accuracy=67.795588	 loss=1.266698	 lr=0.010000
Epoch[019] Batch [0359]/[1203]	Speed: 108.315944 samples/sec	 accuracy=67.718056	 loss=1.269951	 lr=0.010000
Epoch[019] Batch [0379]/[1203]	Speed: 142.303963 samples/sec	 accuracy=67.718421	 loss=1.269017	 lr=0.010000
Epoch[019] Batch [0399]/[1203]	Speed: 117.775615 samples/sec	 accuracy=67.675000	 loss=1.270268	 lr=0.010000
Epoch[019] Batch [0419]/[1203]	Speed: 164.853931 samples/sec	 accuracy=67.648810	 loss=1.272007	 lr=0.010000
Epoch[019] Batch [0439]/[1203]	Speed: 109.779516 samples/sec	 accuracy=67.643182	 loss=1.272777	 lr=0.010000
Epoch[019] Batch [0459]/[1203]	Speed: 105.450255 samples/sec	 accuracy=67.603261	 loss=1.274145	 lr=0.010000
Epoch[019] Batch [0479]/[1203]	Speed: 156.249155 samples/sec	 accuracy=67.654167	 loss=1.273306	 lr=0.010000
Epoch[019] Batch [0499]/[1203]	Speed: 110.733640 samples/sec	 accuracy=67.693000	 loss=1.272997	 lr=0.010000
Epoch[019] Batch [0519]/[1203]	Speed: 122.268635 samples/sec	 accuracy=67.661538	 loss=1.274098	 lr=0.010000
Epoch[019] Batch [0539]/[1203]	Speed: 152.697417 samples/sec	 accuracy=67.594444	 loss=1.275873	 lr=0.010000
Epoch[019] Batch [0559]/[1203]	Speed: 100.195586 samples/sec	 accuracy=67.574107	 loss=1.276473	 lr=0.010000
Epoch[019] Batch [0579]/[1203]	Speed: 150.840135 samples/sec	 accuracy=67.531034	 loss=1.277457	 lr=0.010000
Epoch[019] Batch [0599]/[1203]	Speed: 117.151224 samples/sec	 accuracy=67.481667	 loss=1.279587	 lr=0.010000
Epoch[019] Batch [0619]/[1203]	Speed: 110.509975 samples/sec	 accuracy=67.470968	 loss=1.280474	 lr=0.010000
Epoch[019] Batch [0639]/[1203]	Speed: 155.514981 samples/sec	 accuracy=67.461719	 loss=1.281153	 lr=0.010000
Epoch[019] Batch [0659]/[1203]	Speed: 109.298810 samples/sec	 accuracy=67.418939	 loss=1.282562	 lr=0.010000
Epoch[019] Batch [0679]/[1203]	Speed: 138.222461 samples/sec	 accuracy=67.363971	 loss=1.284502	 lr=0.010000
Epoch[019] Batch [0699]/[1203]	Speed: 120.172588 samples/sec	 accuracy=67.311429	 loss=1.286369	 lr=0.010000
Epoch[019] Batch [0719]/[1203]	Speed: 121.781041 samples/sec	 accuracy=67.313889	 loss=1.286711	 lr=0.010000
Epoch[019] Batch [0739]/[1203]	Speed: 135.845978 samples/sec	 accuracy=67.283784	 loss=1.287482	 lr=0.010000
Epoch[019] Batch [0759]/[1203]	Speed: 112.686784 samples/sec	 accuracy=67.248684	 loss=1.288135	 lr=0.010000
Epoch[019] Batch [0779]/[1203]	Speed: 137.272849 samples/sec	 accuracy=67.269231	 loss=1.287674	 lr=0.010000
Epoch[019] Batch [0799]/[1203]	Speed: 124.168289 samples/sec	 accuracy=67.291875	 loss=1.287349	 lr=0.010000
Epoch[019] Batch [0819]/[1203]	Speed: 108.807202 samples/sec	 accuracy=67.290244	 loss=1.288031	 lr=0.010000
Epoch[019] Batch [0839]/[1203]	Speed: 156.512835 samples/sec	 accuracy=67.296429	 loss=1.287827	 lr=0.010000
Epoch[019] Batch [0859]/[1203]	Speed: 102.680159 samples/sec	 accuracy=67.290116	 loss=1.288223	 lr=0.010000
Epoch[019] Batch [0879]/[1203]	Speed: 145.596165 samples/sec	 accuracy=67.258523	 loss=1.288802	 lr=0.010000
Epoch[019] Batch [0899]/[1203]	Speed: 123.322645 samples/sec	 accuracy=67.236667	 loss=1.289832	 lr=0.010000
Epoch[019] Batch [0919]/[1203]	Speed: 117.755567 samples/sec	 accuracy=67.209239	 loss=1.290832	 lr=0.010000
Epoch[019] Batch [0939]/[1203]	Speed: 136.969487 samples/sec	 accuracy=67.226064	 loss=1.290923	 lr=0.010000
Epoch[019] Batch [0959]/[1203]	Speed: 111.195932 samples/sec	 accuracy=67.237500	 loss=1.290620	 lr=0.010000
Epoch[019] Batch [0979]/[1203]	Speed: 128.858197 samples/sec	 accuracy=67.203061	 loss=1.291572	 lr=0.010000
Epoch[019] Batch [0999]/[1203]	Speed: 122.130586 samples/sec	 accuracy=67.183000	 loss=1.292462	 lr=0.010000
Epoch[019] Batch [1019]/[1203]	Speed: 126.358377 samples/sec	 accuracy=67.174020	 loss=1.293260	 lr=0.010000
Epoch[019] Batch [1039]/[1203]	Speed: 137.499848 samples/sec	 accuracy=67.162981	 loss=1.293685	 lr=0.010000
Epoch[019] Batch [1059]/[1203]	Speed: 107.776874 samples/sec	 accuracy=67.152830	 loss=1.294241	 lr=0.010000
Epoch[019] Batch [1079]/[1203]	Speed: 128.670142 samples/sec	 accuracy=67.150926	 loss=1.293854	 lr=0.010000
Epoch[019] Batch [1099]/[1203]	Speed: 131.525167 samples/sec	 accuracy=67.153636	 loss=1.294441	 lr=0.010000
Epoch[019] Batch [1119]/[1203]	Speed: 110.140250 samples/sec	 accuracy=67.127679	 loss=1.295974	 lr=0.010000
Epoch[019] Batch [1139]/[1203]	Speed: 157.087252 samples/sec	 accuracy=67.111842	 loss=1.296436	 lr=0.010000
Epoch[019] Batch [1159]/[1203]	Speed: 102.693764 samples/sec	 accuracy=67.092672	 loss=1.296932	 lr=0.010000
Epoch[019] Batch [1179]/[1203]	Speed: 153.757585 samples/sec	 accuracy=67.113136	 loss=1.296504	 lr=0.010000
Epoch[019] Batch [1199]/[1203]	Speed: 223.010767 samples/sec	 accuracy=67.110000	 loss=1.297072	 lr=0.010000
Batch [0019]/[0097]: acc-top1=58.525000 acc-top5=81.250000
Batch [0039]/[0097]: acc-top1=57.687500 acc-top5=80.362500
Batch [0059]/[0097]: acc-top1=60.166667 acc-top5=82.033333
Batch [0079]/[0097]: acc-top1=60.062500 acc-top5=81.912500
[Epoch 019] training: accuracy=67.106816	 loss=1.297096
[Epoch 019] speed: 123 samples/sec	time cost: 2137.827615
[Epoch 019] validation: acc-top1=59.685567 acc-top5=82.087629 loss=1.776081
Epoch[020] Batch [0019]/[1203]	Speed: 69.331896 samples/sec	 accuracy=67.975000	 loss=1.269996	 lr=0.010000
Epoch[020] Batch [0039]/[1203]	Speed: 97.370714 samples/sec	 accuracy=68.650000	 loss=1.250213	 lr=0.010000
Epoch[020] Batch [0059]/[1203]	Speed: 237.985990 samples/sec	 accuracy=68.750000	 loss=1.237289	 lr=0.010000
Epoch[020] Batch [0079]/[1203]	Speed: 95.581261 samples/sec	 accuracy=68.668750	 loss=1.238630	 lr=0.010000
Epoch[020] Batch [0099]/[1203]	Speed: 113.590242 samples/sec	 accuracy=68.450000	 loss=1.243922	 lr=0.010000
Epoch[020] Batch [0119]/[1203]	Speed: 167.667856 samples/sec	 accuracy=68.525000	 loss=1.241591	 lr=0.010000
Epoch[020] Batch [0139]/[1203]	Speed: 97.816492 samples/sec	 accuracy=68.585714	 loss=1.240912	 lr=0.010000
Epoch[020] Batch [0159]/[1203]	Speed: 212.941182 samples/sec	 accuracy=68.621875	 loss=1.238813	 lr=0.010000
Epoch[020] Batch [0179]/[1203]	Speed: 97.058455 samples/sec	 accuracy=68.722222	 loss=1.237020	 lr=0.010000
Epoch[020] Batch [0199]/[1203]	Speed: 109.058058 samples/sec	 accuracy=68.712500	 loss=1.237864	 lr=0.010000
Epoch[020] Batch [0219]/[1203]	Speed: 178.725253 samples/sec	 accuracy=68.779545	 loss=1.233985	 lr=0.010000
Epoch[020] Batch [0239]/[1203]	Speed: 102.930300 samples/sec	 accuracy=68.575000	 loss=1.240099	 lr=0.010000
Epoch[020] Batch [0259]/[1203]	Speed: 168.050492 samples/sec	 accuracy=68.546154	 loss=1.238967	 lr=0.010000
Epoch[020] Batch [0279]/[1203]	Speed: 108.357967 samples/sec	 accuracy=68.519643	 loss=1.238663	 lr=0.010000
Epoch[020] Batch [0299]/[1203]	Speed: 100.938836 samples/sec	 accuracy=68.450000	 loss=1.240691	 lr=0.010000
Epoch[020] Batch [0319]/[1203]	Speed: 180.669938 samples/sec	 accuracy=68.354687	 loss=1.244294	 lr=0.010000
Epoch[020] Batch [0339]/[1203]	Speed: 111.516702 samples/sec	 accuracy=68.294118	 loss=1.246519	 lr=0.010000
Epoch[020] Batch [0359]/[1203]	Speed: 116.277025 samples/sec	 accuracy=68.290278	 loss=1.245486	 lr=0.010000
Epoch[020] Batch [0379]/[1203]	Speed: 135.439187 samples/sec	 accuracy=68.218421	 loss=1.247819	 lr=0.010000
Epoch[020] Batch [0399]/[1203]	Speed: 113.648942 samples/sec	 accuracy=68.183750	 loss=1.248715	 lr=0.010000
Epoch[020] Batch [0419]/[1203]	Speed: 155.585651 samples/sec	 accuracy=68.107143	 loss=1.251832	 lr=0.010000
Epoch[020] Batch [0439]/[1203]	Speed: 119.365890 samples/sec	 accuracy=68.010227	 loss=1.255341	 lr=0.010000
Epoch[020] Batch [0459]/[1203]	Speed: 102.267846 samples/sec	 accuracy=68.002174	 loss=1.254372	 lr=0.010000
Epoch[020] Batch [0479]/[1203]	Speed: 138.879339 samples/sec	 accuracy=67.996875	 loss=1.255256	 lr=0.010000
Epoch[020] Batch [0499]/[1203]	Speed: 123.526279 samples/sec	 accuracy=67.974000	 loss=1.255439	 lr=0.010000
Epoch[020] Batch [0519]/[1203]	Speed: 128.293969 samples/sec	 accuracy=67.949038	 loss=1.256297	 lr=0.010000
Epoch[020] Batch [0539]/[1203]	Speed: 135.060834 samples/sec	 accuracy=67.892593	 loss=1.258455	 lr=0.010000
Epoch[020] Batch [0559]/[1203]	Speed: 115.107887 samples/sec	 accuracy=67.870536	 loss=1.259536	 lr=0.010000
Epoch[020] Batch [0579]/[1203]	Speed: 122.541596 samples/sec	 accuracy=67.834483	 loss=1.260713	 lr=0.010000
Epoch[020] Batch [0599]/[1203]	Speed: 138.042898 samples/sec	 accuracy=67.804167	 loss=1.262195	 lr=0.010000
Epoch[020] Batch [0619]/[1203]	Speed: 114.233793 samples/sec	 accuracy=67.774194	 loss=1.263503	 lr=0.010000
Epoch[020] Batch [0639]/[1203]	Speed: 124.501640 samples/sec	 accuracy=67.761719	 loss=1.262975	 lr=0.010000
Epoch[020] Batch [0659]/[1203]	Speed: 117.106678 samples/sec	 accuracy=67.768182	 loss=1.263284	 lr=0.010000
Epoch[020] Batch [0679]/[1203]	Speed: 136.124199 samples/sec	 accuracy=67.765441	 loss=1.264029	 lr=0.010000
Epoch[020] Batch [0699]/[1203]	Speed: 140.651563 samples/sec	 accuracy=67.685714	 loss=1.266471	 lr=0.010000
Epoch[020] Batch [0719]/[1203]	Speed: 108.794145 samples/sec	 accuracy=67.668056	 loss=1.267415	 lr=0.010000
Epoch[020] Batch [0739]/[1203]	Speed: 108.707961 samples/sec	 accuracy=67.635811	 loss=1.269512	 lr=0.010000
Epoch[020] Batch [0759]/[1203]	Speed: 147.565776 samples/sec	 accuracy=67.619737	 loss=1.269820	 lr=0.010000
Epoch[020] Batch [0779]/[1203]	Speed: 122.057412 samples/sec	 accuracy=67.652564	 loss=1.269592	 lr=0.010000
Epoch[020] Batch [0799]/[1203]	Speed: 130.784686 samples/sec	 accuracy=67.618750	 loss=1.270981	 lr=0.010000
Epoch[020] Batch [0819]/[1203]	Speed: 123.033862 samples/sec	 accuracy=67.603659	 loss=1.272229	 lr=0.010000
Epoch[020] Batch [0839]/[1203]	Speed: 113.099125 samples/sec	 accuracy=67.586905	 loss=1.272628	 lr=0.010000
Epoch[020] Batch [0859]/[1203]	Speed: 138.507497 samples/sec	 accuracy=67.568023	 loss=1.273044	 lr=0.010000
Epoch[020] Batch [0879]/[1203]	Speed: 118.895876 samples/sec	 accuracy=67.564773	 loss=1.273354	 lr=0.010000
Epoch[020] Batch [0899]/[1203]	Speed: 105.463008 samples/sec	 accuracy=67.520556	 loss=1.274834	 lr=0.010000
Epoch[020] Batch [0919]/[1203]	Speed: 168.706948 samples/sec	 accuracy=67.520109	 loss=1.274950	 lr=0.010000
Epoch[020] Batch [0939]/[1203]	Speed: 107.319117 samples/sec	 accuracy=67.483511	 loss=1.276090	 lr=0.010000
Epoch[020] Batch [0959]/[1203]	Speed: 127.610660 samples/sec	 accuracy=67.471354	 loss=1.277155	 lr=0.010000
Epoch[020] Batch [0979]/[1203]	Speed: 138.003775 samples/sec	 accuracy=67.420918	 loss=1.278612	 lr=0.010000
Epoch[020] Batch [0999]/[1203]	Speed: 97.511220 samples/sec	 accuracy=67.410000	 loss=1.279386	 lr=0.010000
Epoch[020] Batch [1019]/[1203]	Speed: 187.707914 samples/sec	 accuracy=67.379902	 loss=1.280417	 lr=0.010000
Epoch[020] Batch [1039]/[1203]	Speed: 103.384993 samples/sec	 accuracy=67.409615	 loss=1.280099	 lr=0.010000
Epoch[020] Batch [1059]/[1203]	Speed: 110.020407 samples/sec	 accuracy=67.433962	 loss=1.279769	 lr=0.010000
Epoch[020] Batch [1079]/[1203]	Speed: 174.166766 samples/sec	 accuracy=67.427315	 loss=1.279914	 lr=0.010000
Epoch[020] Batch [1099]/[1203]	Speed: 98.141808 samples/sec	 accuracy=67.407727	 loss=1.280731	 lr=0.010000
Epoch[020] Batch [1119]/[1203]	Speed: 178.156369 samples/sec	 accuracy=67.407589	 loss=1.280315	 lr=0.010000
Epoch[020] Batch [1139]/[1203]	Speed: 111.722872 samples/sec	 accuracy=67.383772	 loss=1.281508	 lr=0.010000
Epoch[020] Batch [1159]/[1203]	Speed: 105.424654 samples/sec	 accuracy=67.360776	 loss=1.282732	 lr=0.010000
Epoch[020] Batch [1179]/[1203]	Speed: 190.615756 samples/sec	 accuracy=67.341525	 loss=1.283737	 lr=0.010000
Epoch[020] Batch [1199]/[1203]	Speed: 189.049802 samples/sec	 accuracy=67.325417	 loss=1.284656	 lr=0.010000
Batch [0019]/[0097]: acc-top1=57.875000 acc-top5=80.900000
Batch [0039]/[0097]: acc-top1=57.287500 acc-top5=80.900000
Batch [0059]/[0097]: acc-top1=59.916667 acc-top5=82.391667
Batch [0079]/[0097]: acc-top1=60.262500 acc-top5=82.400000
[Epoch 020] training: accuracy=67.324190	 loss=1.284853
[Epoch 020] speed: 123 samples/sec	time cost: 2136.692352
[Epoch 020] validation: acc-top1=59.618557 acc-top5=82.391753 loss=1.752483
Epoch[021] Batch [0019]/[1203]	Speed: 69.253512 samples/sec	 accuracy=68.425000	 loss=1.223542	 lr=0.010000
Epoch[021] Batch [0039]/[1203]	Speed: 104.399070 samples/sec	 accuracy=67.887500	 loss=1.236853	 lr=0.010000
Epoch[021] Batch [0059]/[1203]	Speed: 212.399150 samples/sec	 accuracy=68.283333	 loss=1.229256	 lr=0.010000
Epoch[021] Batch [0079]/[1203]	Speed: 97.682270 samples/sec	 accuracy=68.306250	 loss=1.231127	 lr=0.010000
Epoch[021] Batch [0099]/[1203]	Speed: 105.427354 samples/sec	 accuracy=68.220000	 loss=1.238243	 lr=0.010000
Epoch[021] Batch [0119]/[1203]	Speed: 179.924186 samples/sec	 accuracy=68.304167	 loss=1.231960	 lr=0.010000
Epoch[021] Batch [0139]/[1203]	Speed: 94.454346 samples/sec	 accuracy=68.267857	 loss=1.232134	 lr=0.010000
Epoch[021] Batch [0159]/[1203]	Speed: 227.808166 samples/sec	 accuracy=68.268750	 loss=1.234536	 lr=0.010000
Epoch[021] Batch [0179]/[1203]	Speed: 101.507736 samples/sec	 accuracy=68.325000	 loss=1.231274	 lr=0.010000
Epoch[021] Batch [0199]/[1203]	Speed: 115.359233 samples/sec	 accuracy=68.397500	 loss=1.228296	 lr=0.010000
Epoch[021] Batch [0219]/[1203]	Speed: 162.839240 samples/sec	 accuracy=68.338636	 loss=1.232616	 lr=0.010000
Epoch[021] Batch [0239]/[1203]	Speed: 98.431915 samples/sec	 accuracy=68.170833	 loss=1.237508	 lr=0.010000
Epoch[021] Batch [0259]/[1203]	Speed: 126.978290 samples/sec	 accuracy=68.186538	 loss=1.238607	 lr=0.010000
Epoch[021] Batch [0279]/[1203]	Speed: 137.899777 samples/sec	 accuracy=68.237500	 loss=1.238989	 lr=0.010000
Epoch[021] Batch [0299]/[1203]	Speed: 97.101717 samples/sec	 accuracy=68.201667	 loss=1.241630	 lr=0.010000
Epoch[021] Batch [0319]/[1203]	Speed: 216.951403 samples/sec	 accuracy=68.210938	 loss=1.242235	 lr=0.010000
Epoch[021] Batch [0339]/[1203]	Speed: 100.520897 samples/sec	 accuracy=68.258824	 loss=1.240567	 lr=0.010000
Epoch[021] Batch [0359]/[1203]	Speed: 132.562419 samples/sec	 accuracy=68.261111	 loss=1.239402	 lr=0.010000
Epoch[021] Batch [0379]/[1203]	Speed: 133.691049 samples/sec	 accuracy=68.256579	 loss=1.238614	 lr=0.010000
Epoch[021] Batch [0399]/[1203]	Speed: 94.840954 samples/sec	 accuracy=68.158750	 loss=1.243332	 lr=0.010000
Epoch[021] Batch [0419]/[1203]	Speed: 215.468965 samples/sec	 accuracy=68.176190	 loss=1.243738	 lr=0.010000
Epoch[021] Batch [0439]/[1203]	Speed: 96.436552 samples/sec	 accuracy=68.194318	 loss=1.243596	 lr=0.010000
Epoch[021] Batch [0459]/[1203]	Speed: 132.028211 samples/sec	 accuracy=68.168478	 loss=1.244990	 lr=0.010000
Epoch[021] Batch [0479]/[1203]	Speed: 134.250460 samples/sec	 accuracy=68.193750	 loss=1.244668	 lr=0.010000
Epoch[021] Batch [0499]/[1203]	Speed: 101.909783 samples/sec	 accuracy=68.163000	 loss=1.246327	 lr=0.010000
Epoch[021] Batch [0519]/[1203]	Speed: 161.174185 samples/sec	 accuracy=68.145192	 loss=1.247092	 lr=0.010000
Epoch[021] Batch [0539]/[1203]	Speed: 106.747037 samples/sec	 accuracy=68.149074	 loss=1.247955	 lr=0.010000
Epoch[021] Batch [0559]/[1203]	Speed: 107.761413 samples/sec	 accuracy=68.136607	 loss=1.247942	 lr=0.010000
Epoch[021] Batch [0579]/[1203]	Speed: 171.374722 samples/sec	 accuracy=68.106897	 loss=1.249548	 lr=0.010000
Epoch[021] Batch [0599]/[1203]	Speed: 108.251736 samples/sec	 accuracy=68.138333	 loss=1.248460	 lr=0.010000
Epoch[021] Batch [0619]/[1203]	Speed: 143.031291 samples/sec	 accuracy=68.157258	 loss=1.248589	 lr=0.010000
Epoch[021] Batch [0639]/[1203]	Speed: 118.501696 samples/sec	 accuracy=68.135937	 loss=1.249404	 lr=0.010000
Epoch[021] Batch [0659]/[1203]	Speed: 108.710004 samples/sec	 accuracy=68.104545	 loss=1.251801	 lr=0.010000
Epoch[021] Batch [0679]/[1203]	Speed: 159.923944 samples/sec	 accuracy=68.079412	 loss=1.253155	 lr=0.010000
Epoch[021] Batch [0699]/[1203]	Speed: 109.508049 samples/sec	 accuracy=68.077143	 loss=1.253441	 lr=0.010000
Epoch[021] Batch [0719]/[1203]	Speed: 127.086657 samples/sec	 accuracy=68.001389	 loss=1.254997	 lr=0.010000
Epoch[021] Batch [0739]/[1203]	Speed: 123.976106 samples/sec	 accuracy=67.977027	 loss=1.255387	 lr=0.010000
Epoch[021] Batch [0759]/[1203]	Speed: 122.930868 samples/sec	 accuracy=67.978947	 loss=1.256353	 lr=0.010000
Epoch[021] Batch [0779]/[1203]	Speed: 131.277609 samples/sec	 accuracy=68.001282	 loss=1.255859	 lr=0.010000
Epoch[021] Batch [0799]/[1203]	Speed: 117.740994 samples/sec	 accuracy=67.994375	 loss=1.256461	 lr=0.010000
Epoch[021] Batch [0819]/[1203]	Speed: 117.363990 samples/sec	 accuracy=67.981098	 loss=1.256571	 lr=0.010000
Epoch[021] Batch [0839]/[1203]	Speed: 134.469626 samples/sec	 accuracy=67.961310	 loss=1.257374	 lr=0.010000
Epoch[021] Batch [0859]/[1203]	Speed: 124.058359 samples/sec	 accuracy=67.914535	 loss=1.259028	 lr=0.010000
Epoch[021] Batch [0879]/[1203]	Speed: 123.301956 samples/sec	 accuracy=67.885227	 loss=1.259964	 lr=0.010000
Epoch[021] Batch [0899]/[1203]	Speed: 122.286610 samples/sec	 accuracy=67.889444	 loss=1.259810	 lr=0.010000
Epoch[021] Batch [0919]/[1203]	Speed: 123.996243 samples/sec	 accuracy=67.866304	 loss=1.260113	 lr=0.010000
Epoch[021] Batch [0939]/[1203]	Speed: 129.464777 samples/sec	 accuracy=67.823404	 loss=1.261439	 lr=0.010000
Epoch[021] Batch [0959]/[1203]	Speed: 114.546796 samples/sec	 accuracy=67.790104	 loss=1.262404	 lr=0.010000
Epoch[021] Batch [0979]/[1203]	Speed: 137.993425 samples/sec	 accuracy=67.775510	 loss=1.263258	 lr=0.010000
Epoch[021] Batch [0999]/[1203]	Speed: 121.136177 samples/sec	 accuracy=67.757000	 loss=1.263962	 lr=0.010000
Epoch[021] Batch [1019]/[1203]	Speed: 119.097508 samples/sec	 accuracy=67.738725	 loss=1.264862	 lr=0.010000
Epoch[021] Batch [1039]/[1203]	Speed: 134.620538 samples/sec	 accuracy=67.755288	 loss=1.264544	 lr=0.010000
Epoch[021] Batch [1059]/[1203]	Speed: 113.507634 samples/sec	 accuracy=67.735377	 loss=1.265355	 lr=0.010000
Epoch[021] Batch [1079]/[1203]	Speed: 121.292742 samples/sec	 accuracy=67.702778	 loss=1.266611	 lr=0.010000
Epoch[021] Batch [1099]/[1203]	Speed: 124.144842 samples/sec	 accuracy=67.700909	 loss=1.267038	 lr=0.010000
Epoch[021] Batch [1119]/[1203]	Speed: 142.013882 samples/sec	 accuracy=67.689732	 loss=1.267153	 lr=0.010000
Epoch[021] Batch [1139]/[1203]	Speed: 121.369856 samples/sec	 accuracy=67.684211	 loss=1.267365	 lr=0.010000
Epoch[021] Batch [1159]/[1203]	Speed: 118.075549 samples/sec	 accuracy=67.678448	 loss=1.267552	 lr=0.010000
Epoch[021] Batch [1179]/[1203]	Speed: 136.666956 samples/sec	 accuracy=67.672034	 loss=1.267286	 lr=0.010000
Epoch[021] Batch [1199]/[1203]	Speed: 221.118461 samples/sec	 accuracy=67.654583	 loss=1.268008	 lr=0.010000
Batch [0019]/[0097]: acc-top1=59.050000 acc-top5=80.950000
Batch [0039]/[0097]: acc-top1=58.175000 acc-top5=81.025000
Batch [0059]/[0097]: acc-top1=59.891667 acc-top5=82.500000
Batch [0079]/[0097]: acc-top1=59.781250 acc-top5=82.250000
[Epoch 021] training: accuracy=67.647548	 loss=1.268232
[Epoch 021] speed: 123 samples/sec	time cost: 2138.784167
[Epoch 021] validation: acc-top1=59.469072 acc-top5=82.061856 loss=1.745282
Epoch[022] Batch [0019]/[1204]	Speed: 68.594620 samples/sec	 accuracy=68.875000	 loss=1.201337	 lr=0.010000
Epoch[022] Batch [0039]/[1204]	Speed: 97.850872 samples/sec	 accuracy=68.712500	 loss=1.200501	 lr=0.010000
Epoch[022] Batch [0059]/[1204]	Speed: 248.793020 samples/sec	 accuracy=69.208333	 loss=1.191132	 lr=0.010000
Epoch[022] Batch [0079]/[1204]	Speed: 96.115303 samples/sec	 accuracy=69.156250	 loss=1.191416	 lr=0.010000
Epoch[022] Batch [0099]/[1204]	Speed: 96.078068 samples/sec	 accuracy=68.970000	 loss=1.198535	 lr=0.010000
Epoch[022] Batch [0119]/[1204]	Speed: 220.335188 samples/sec	 accuracy=69.016667	 loss=1.194752	 lr=0.010000
Epoch[022] Batch [0139]/[1204]	Speed: 96.948855 samples/sec	 accuracy=68.928571	 loss=1.203056	 lr=0.010000
Epoch[022] Batch [0159]/[1204]	Speed: 236.710653 samples/sec	 accuracy=68.946875	 loss=1.207105	 lr=0.010000
Epoch[022] Batch [0179]/[1204]	Speed: 94.956360 samples/sec	 accuracy=68.861111	 loss=1.209695	 lr=0.010000
Epoch[022] Batch [0199]/[1204]	Speed: 100.189918 samples/sec	 accuracy=68.820000	 loss=1.209537	 lr=0.010000
Epoch[022] Batch [0219]/[1204]	Speed: 203.568295 samples/sec	 accuracy=68.715909	 loss=1.211002	 lr=0.010000
Epoch[022] Batch [0239]/[1204]	Speed: 96.779663 samples/sec	 accuracy=68.683333	 loss=1.212150	 lr=0.010000
Epoch[022] Batch [0259]/[1204]	Speed: 125.939844 samples/sec	 accuracy=68.569231	 loss=1.215541	 lr=0.010000
Epoch[022] Batch [0279]/[1204]	Speed: 148.415893 samples/sec	 accuracy=68.614286	 loss=1.214604	 lr=0.010000
Epoch[022] Batch [0299]/[1204]	Speed: 100.908247 samples/sec	 accuracy=68.575000	 loss=1.216730	 lr=0.010000
Epoch[022] Batch [0319]/[1204]	Speed: 169.947388 samples/sec	 accuracy=68.639063	 loss=1.215366	 lr=0.010000
Epoch[022] Batch [0339]/[1204]	Speed: 104.173498 samples/sec	 accuracy=68.627941	 loss=1.216425	 lr=0.010000
Epoch[022] Batch [0359]/[1204]	Speed: 114.050793 samples/sec	 accuracy=68.570833	 loss=1.218202	 lr=0.010000
Epoch[022] Batch [0379]/[1204]	Speed: 165.107412 samples/sec	 accuracy=68.490789	 loss=1.221969	 lr=0.010000
Epoch[022] Batch [0399]/[1204]	Speed: 98.262585 samples/sec	 accuracy=68.542500	 loss=1.221118	 lr=0.010000
Epoch[022] Batch [0419]/[1204]	Speed: 193.752910 samples/sec	 accuracy=68.480952	 loss=1.222509	 lr=0.010000
Epoch[022] Batch [0439]/[1204]	Speed: 102.478283 samples/sec	 accuracy=68.420455	 loss=1.225177	 lr=0.010000
Epoch[022] Batch [0459]/[1204]	Speed: 121.270289 samples/sec	 accuracy=68.328261	 loss=1.228138	 lr=0.010000
Epoch[022] Batch [0479]/[1204]	Speed: 147.684883 samples/sec	 accuracy=68.311458	 loss=1.229465	 lr=0.010000
Epoch[022] Batch [0499]/[1204]	Speed: 99.012653 samples/sec	 accuracy=68.301000	 loss=1.229945	 lr=0.010000
Epoch[022] Batch [0519]/[1204]	Speed: 135.132256 samples/sec	 accuracy=68.297115	 loss=1.230578	 lr=0.010000
Epoch[022] Batch [0539]/[1204]	Speed: 128.348403 samples/sec	 accuracy=68.264815	 loss=1.233083	 lr=0.010000
Epoch[022] Batch [0559]/[1204]	Speed: 114.989253 samples/sec	 accuracy=68.256250	 loss=1.233402	 lr=0.010000
Epoch[022] Batch [0579]/[1204]	Speed: 159.478586 samples/sec	 accuracy=68.248276	 loss=1.233915	 lr=0.010000
Epoch[022] Batch [0599]/[1204]	Speed: 111.687926 samples/sec	 accuracy=68.211667	 loss=1.236025	 lr=0.010000
Epoch[022] Batch [0619]/[1204]	Speed: 127.796955 samples/sec	 accuracy=68.199194	 loss=1.236380	 lr=0.010000
Epoch[022] Batch [0639]/[1204]	Speed: 130.089662 samples/sec	 accuracy=68.227344	 loss=1.235641	 lr=0.010000
Epoch[022] Batch [0659]/[1204]	Speed: 113.548020 samples/sec	 accuracy=68.170455	 loss=1.237534	 lr=0.010000
Epoch[022] Batch [0679]/[1204]	Speed: 132.394469 samples/sec	 accuracy=68.153676	 loss=1.238414	 lr=0.010000
Epoch[022] Batch [0699]/[1204]	Speed: 121.752549 samples/sec	 accuracy=68.122143	 loss=1.239734	 lr=0.010000
Epoch[022] Batch [0719]/[1204]	Speed: 112.418016 samples/sec	 accuracy=68.122222	 loss=1.239890	 lr=0.010000
Epoch[022] Batch [0739]/[1204]	Speed: 142.946004 samples/sec	 accuracy=68.108108	 loss=1.240428	 lr=0.010000
Epoch[022] Batch [0759]/[1204]	Speed: 117.596817 samples/sec	 accuracy=68.097368	 loss=1.241175	 lr=0.010000
Epoch[022] Batch [0779]/[1204]	Speed: 141.629184 samples/sec	 accuracy=68.060897	 loss=1.242228	 lr=0.010000
Epoch[022] Batch [0799]/[1204]	Speed: 104.914321 samples/sec	 accuracy=68.045625	 loss=1.243220	 lr=0.010000
Epoch[022] Batch [0819]/[1204]	Speed: 120.362811 samples/sec	 accuracy=68.016463	 loss=1.244861	 lr=0.010000
Epoch[022] Batch [0839]/[1204]	Speed: 138.507512 samples/sec	 accuracy=67.975000	 loss=1.246242	 lr=0.010000
Epoch[022] Batch [0859]/[1204]	Speed: 122.047906 samples/sec	 accuracy=67.948837	 loss=1.247089	 lr=0.010000
Epoch[022] Batch [0879]/[1204]	Speed: 136.550441 samples/sec	 accuracy=67.918182	 loss=1.248392	 lr=0.010000
Epoch[022] Batch [0899]/[1204]	Speed: 111.828019 samples/sec	 accuracy=67.912778	 loss=1.248248	 lr=0.010000
Epoch[022] Batch [0919]/[1204]	Speed: 124.181450 samples/sec	 accuracy=67.919565	 loss=1.248798	 lr=0.010000
Epoch[022] Batch [0939]/[1204]	Speed: 139.951223 samples/sec	 accuracy=67.889894	 loss=1.249303	 lr=0.010000
Epoch[022] Batch [0959]/[1204]	Speed: 113.546516 samples/sec	 accuracy=67.894271	 loss=1.248651	 lr=0.010000
Epoch[022] Batch [0979]/[1204]	Speed: 115.386440 samples/sec	 accuracy=67.889796	 loss=1.248687	 lr=0.010000
Epoch[022] Batch [0999]/[1204]	Speed: 142.557591 samples/sec	 accuracy=67.910500	 loss=1.248419	 lr=0.010000
Epoch[022] Batch [1019]/[1204]	Speed: 122.751527 samples/sec	 accuracy=67.898529	 loss=1.248852	 lr=0.010000
Epoch[022] Batch [1039]/[1204]	Speed: 133.640314 samples/sec	 accuracy=67.879327	 loss=1.249558	 lr=0.010000
Epoch[022] Batch [1059]/[1204]	Speed: 116.730597 samples/sec	 accuracy=67.867925	 loss=1.250117	 lr=0.010000
Epoch[022] Batch [1079]/[1204]	Speed: 114.710142 samples/sec	 accuracy=67.865278	 loss=1.250372	 lr=0.010000
Epoch[022] Batch [1099]/[1204]	Speed: 128.818169 samples/sec	 accuracy=67.860909	 loss=1.250971	 lr=0.010000
Epoch[022] Batch [1119]/[1204]	Speed: 127.710464 samples/sec	 accuracy=67.862054	 loss=1.251159	 lr=0.010000
Epoch[022] Batch [1139]/[1204]	Speed: 115.328295 samples/sec	 accuracy=67.834649	 loss=1.252202	 lr=0.010000
Epoch[022] Batch [1159]/[1204]	Speed: 125.339815 samples/sec	 accuracy=67.831466	 loss=1.252270	 lr=0.010000
Epoch[022] Batch [1179]/[1204]	Speed: 135.730947 samples/sec	 accuracy=67.807203	 loss=1.253127	 lr=0.010000
Epoch[022] Batch [1199]/[1204]	Speed: 240.865356 samples/sec	 accuracy=67.800417	 loss=1.253782	 lr=0.010000
Batch [0019]/[0097]: acc-top1=59.325000 acc-top5=81.625000
Batch [0039]/[0097]: acc-top1=57.475000 acc-top5=80.487500
Batch [0059]/[0097]: acc-top1=59.708333 acc-top5=82.108333
Batch [0079]/[0097]: acc-top1=59.837500 acc-top5=82.131250
[Epoch 022] training: accuracy=67.802326	 loss=1.253628
[Epoch 022] speed: 123 samples/sec	time cost: 2146.066702
[Epoch 022] validation: acc-top1=59.252577 acc-top5=82.046392 loss=1.754102
Epoch[023] Batch [0019]/[1203]	Speed: 69.107868 samples/sec	 accuracy=68.175000	 loss=1.231562	 lr=0.010000
Epoch[023] Batch [0039]/[1203]	Speed: 95.068032 samples/sec	 accuracy=69.062500	 loss=1.198012	 lr=0.010000
Epoch[023] Batch [0059]/[1203]	Speed: 265.409679 samples/sec	 accuracy=69.466667	 loss=1.184754	 lr=0.010000
Epoch[023] Batch [0079]/[1203]	Speed: 91.283412 samples/sec	 accuracy=69.400000	 loss=1.192624	 lr=0.010000
Epoch[023] Batch [0099]/[1203]	Speed: 136.747093 samples/sec	 accuracy=69.265000	 loss=1.196478	 lr=0.010000
Epoch[023] Batch [0119]/[1203]	Speed: 140.335598 samples/sec	 accuracy=69.112500	 loss=1.199603	 lr=0.010000
Epoch[023] Batch [0139]/[1203]	Speed: 97.038155 samples/sec	 accuracy=69.228571	 loss=1.195387	 lr=0.010000
Epoch[023] Batch [0159]/[1203]	Speed: 250.532432 samples/sec	 accuracy=69.184375	 loss=1.197824	 lr=0.010000
Epoch[023] Batch [0179]/[1203]	Speed: 89.799208 samples/sec	 accuracy=69.061111	 loss=1.198894	 lr=0.010000
Epoch[023] Batch [0199]/[1203]	Speed: 120.679846 samples/sec	 accuracy=69.152500	 loss=1.197056	 lr=0.010000
Epoch[023] Batch [0219]/[1203]	Speed: 152.665940 samples/sec	 accuracy=69.220455	 loss=1.194696	 lr=0.010000
Epoch[023] Batch [0239]/[1203]	Speed: 107.133769 samples/sec	 accuracy=69.204167	 loss=1.194897	 lr=0.010000
Epoch[023] Batch [0259]/[1203]	Speed: 160.752993 samples/sec	 accuracy=69.167308	 loss=1.196253	 lr=0.010000
Epoch[023] Batch [0279]/[1203]	Speed: 103.088404 samples/sec	 accuracy=69.200000	 loss=1.196821	 lr=0.010000
Epoch[023] Batch [0299]/[1203]	Speed: 110.799767 samples/sec	 accuracy=69.185000	 loss=1.197441	 lr=0.010000
Epoch[023] Batch [0319]/[1203]	Speed: 158.631421 samples/sec	 accuracy=69.189062	 loss=1.197312	 lr=0.010000
Epoch[023] Batch [0339]/[1203]	Speed: 113.911690 samples/sec	 accuracy=69.082353	 loss=1.202582	 lr=0.010000
Epoch[023] Batch [0359]/[1203]	Speed: 114.972125 samples/sec	 accuracy=69.051389	 loss=1.203465	 lr=0.010000
Epoch[023] Batch [0379]/[1203]	Speed: 142.246679 samples/sec	 accuracy=69.078947	 loss=1.204440	 lr=0.010000
Epoch[023] Batch [0399]/[1203]	Speed: 111.196390 samples/sec	 accuracy=69.045000	 loss=1.206076	 lr=0.010000
Epoch[023] Batch [0419]/[1203]	Speed: 170.103703 samples/sec	 accuracy=68.988095	 loss=1.208029	 lr=0.010000
Epoch[023] Batch [0439]/[1203]	Speed: 103.866778 samples/sec	 accuracy=68.950000	 loss=1.209327	 lr=0.010000
Epoch[023] Batch [0459]/[1203]	Speed: 119.525713 samples/sec	 accuracy=68.933696	 loss=1.210074	 lr=0.010000
Epoch[023] Batch [0479]/[1203]	Speed: 142.677412 samples/sec	 accuracy=68.909375	 loss=1.211103	 lr=0.010000
Epoch[023] Batch [0499]/[1203]	Speed: 109.003721 samples/sec	 accuracy=68.930000	 loss=1.210249	 lr=0.010000
Epoch[023] Batch [0519]/[1203]	Speed: 145.297153 samples/sec	 accuracy=68.912500	 loss=1.210793	 lr=0.010000
Epoch[023] Batch [0539]/[1203]	Speed: 120.765077 samples/sec	 accuracy=68.851852	 loss=1.213390	 lr=0.010000
Epoch[023] Batch [0559]/[1203]	Speed: 114.323746 samples/sec	 accuracy=68.837500	 loss=1.214602	 lr=0.010000
Epoch[023] Batch [0579]/[1203]	Speed: 144.420281 samples/sec	 accuracy=68.826724	 loss=1.214999	 lr=0.010000
Epoch[023] Batch [0599]/[1203]	Speed: 106.158688 samples/sec	 accuracy=68.795000	 loss=1.215207	 lr=0.010000
Epoch[023] Batch [0619]/[1203]	Speed: 130.999295 samples/sec	 accuracy=68.785484	 loss=1.215310	 lr=0.010000
Epoch[023] Batch [0639]/[1203]	Speed: 140.363299 samples/sec	 accuracy=68.796094	 loss=1.214728	 lr=0.010000
Epoch[023] Batch [0659]/[1203]	Speed: 115.233358 samples/sec	 accuracy=68.787879	 loss=1.215253	 lr=0.010000
Epoch[023] Batch [0679]/[1203]	Speed: 132.592497 samples/sec	 accuracy=68.745588	 loss=1.218043	 lr=0.010000
Epoch[023] Batch [0699]/[1203]	Speed: 111.209717 samples/sec	 accuracy=68.709286	 loss=1.219548	 lr=0.010000
Epoch[023] Batch [0719]/[1203]	Speed: 124.932063 samples/sec	 accuracy=68.692361	 loss=1.220224	 lr=0.010000
Epoch[023] Batch [0739]/[1203]	Speed: 136.169955 samples/sec	 accuracy=68.666892	 loss=1.220627	 lr=0.010000
Epoch[023] Batch [0759]/[1203]	Speed: 108.542190 samples/sec	 accuracy=68.648026	 loss=1.222056	 lr=0.010000
Epoch[023] Batch [0779]/[1203]	Speed: 131.192437 samples/sec	 accuracy=68.635256	 loss=1.222447	 lr=0.010000
Epoch[023] Batch [0799]/[1203]	Speed: 135.937926 samples/sec	 accuracy=68.595000	 loss=1.223947	 lr=0.010000
Epoch[023] Batch [0819]/[1203]	Speed: 116.604036 samples/sec	 accuracy=68.581707	 loss=1.224892	 lr=0.010000
Epoch[023] Batch [0839]/[1203]	Speed: 136.969048 samples/sec	 accuracy=68.583929	 loss=1.224780	 lr=0.010000
Epoch[023] Batch [0859]/[1203]	Speed: 114.571703 samples/sec	 accuracy=68.593605	 loss=1.224698	 lr=0.010000
Epoch[023] Batch [0879]/[1203]	Speed: 119.773951 samples/sec	 accuracy=68.576136	 loss=1.225047	 lr=0.010000
Epoch[023] Batch [0899]/[1203]	Speed: 133.263326 samples/sec	 accuracy=68.569444	 loss=1.225954	 lr=0.010000
Epoch[023] Batch [0919]/[1203]	Speed: 126.387157 samples/sec	 accuracy=68.561957	 loss=1.226757	 lr=0.010000
Epoch[023] Batch [0939]/[1203]	Speed: 114.833055 samples/sec	 accuracy=68.565957	 loss=1.226651	 lr=0.010000
Epoch[023] Batch [0959]/[1203]	Speed: 141.250743 samples/sec	 accuracy=68.563542	 loss=1.226825	 lr=0.010000
Epoch[023] Batch [0979]/[1203]	Speed: 112.893609 samples/sec	 accuracy=68.525000	 loss=1.228428	 lr=0.010000
Epoch[023] Batch [0999]/[1203]	Speed: 132.981697 samples/sec	 accuracy=68.494000	 loss=1.229262	 lr=0.010000
Epoch[023] Batch [1019]/[1203]	Speed: 121.115618 samples/sec	 accuracy=68.495588	 loss=1.229693	 lr=0.010000
Epoch[023] Batch [1039]/[1203]	Speed: 124.016822 samples/sec	 accuracy=68.469231	 loss=1.230887	 lr=0.010000
Epoch[023] Batch [1059]/[1203]	Speed: 123.820182 samples/sec	 accuracy=68.430660	 loss=1.232339	 lr=0.010000
Epoch[023] Batch [1079]/[1203]	Speed: 115.588231 samples/sec	 accuracy=68.422222	 loss=1.233140	 lr=0.010000
Epoch[023] Batch [1099]/[1203]	Speed: 120.945150 samples/sec	 accuracy=68.418636	 loss=1.233832	 lr=0.010000
Epoch[023] Batch [1119]/[1203]	Speed: 136.114763 samples/sec	 accuracy=68.429018	 loss=1.233814	 lr=0.010000
Epoch[023] Batch [1139]/[1203]	Speed: 123.898777 samples/sec	 accuracy=68.432895	 loss=1.234016	 lr=0.010000
Epoch[023] Batch [1159]/[1203]	Speed: 118.242069 samples/sec	 accuracy=68.438362	 loss=1.234082	 lr=0.010000
Epoch[023] Batch [1179]/[1203]	Speed: 131.246489 samples/sec	 accuracy=68.432203	 loss=1.234299	 lr=0.010000
Epoch[023] Batch [1199]/[1203]	Speed: 241.792889 samples/sec	 accuracy=68.429167	 loss=1.233837	 lr=0.010000
Batch [0019]/[0097]: acc-top1=60.300000 acc-top5=82.250000
Batch [0039]/[0097]: acc-top1=58.012500 acc-top5=80.762500
Batch [0059]/[0097]: acc-top1=60.358333 acc-top5=82.408333
Batch [0079]/[0097]: acc-top1=59.643750 acc-top5=81.800000
[Epoch 023] training: accuracy=68.424771	 loss=1.233892
[Epoch 023] speed: 123 samples/sec	time cost: 2138.450844
[Epoch 023] validation: acc-top1=59.149485 acc-top5=81.788660 loss=1.787291
Epoch[024] Batch [0019]/[1203]	Speed: 69.200143 samples/sec	 accuracy=69.150000	 loss=1.207073	 lr=0.010000
Epoch[024] Batch [0039]/[1203]	Speed: 103.631103 samples/sec	 accuracy=69.662500	 loss=1.183696	 lr=0.010000
Epoch[024] Batch [0059]/[1203]	Speed: 212.371334 samples/sec	 accuracy=69.691667	 loss=1.180777	 lr=0.010000
Epoch[024] Batch [0079]/[1203]	Speed: 93.499031 samples/sec	 accuracy=69.775000	 loss=1.178849	 lr=0.010000
Epoch[024] Batch [0099]/[1203]	Speed: 115.456105 samples/sec	 accuracy=69.905000	 loss=1.177677	 lr=0.010000
Epoch[024] Batch [0119]/[1203]	Speed: 166.922320 samples/sec	 accuracy=69.875000	 loss=1.178285	 lr=0.010000
Epoch[024] Batch [0139]/[1203]	Speed: 97.556622 samples/sec	 accuracy=69.789286	 loss=1.181444	 lr=0.010000
Epoch[024] Batch [0159]/[1203]	Speed: 223.247334 samples/sec	 accuracy=69.671875	 loss=1.178483	 lr=0.010000
Epoch[024] Batch [0179]/[1203]	Speed: 97.927409 samples/sec	 accuracy=69.608333	 loss=1.179528	 lr=0.010000
Epoch[024] Batch [0199]/[1203]	Speed: 122.539604 samples/sec	 accuracy=69.527500	 loss=1.184303	 lr=0.010000
Epoch[024] Batch [0219]/[1203]	Speed: 131.646516 samples/sec	 accuracy=69.502273	 loss=1.185416	 lr=0.010000
Epoch[024] Batch [0239]/[1203]	Speed: 111.146849 samples/sec	 accuracy=69.502083	 loss=1.185665	 lr=0.010000
Epoch[024] Batch [0259]/[1203]	Speed: 161.781983 samples/sec	 accuracy=69.515385	 loss=1.185312	 lr=0.010000
Epoch[024] Batch [0279]/[1203]	Speed: 103.493680 samples/sec	 accuracy=69.482143	 loss=1.186025	 lr=0.010000
Epoch[024] Batch [0299]/[1203]	Speed: 133.118447 samples/sec	 accuracy=69.441667	 loss=1.187423	 lr=0.010000
Epoch[024] Batch [0319]/[1203]	Speed: 135.069644 samples/sec	 accuracy=69.439063	 loss=1.188523	 lr=0.010000
Epoch[024] Batch [0339]/[1203]	Speed: 111.361064 samples/sec	 accuracy=69.347059	 loss=1.190414	 lr=0.010000
Epoch[024] Batch [0359]/[1203]	Speed: 140.852555 samples/sec	 accuracy=69.400000	 loss=1.187939	 lr=0.010000
Epoch[024] Batch [0379]/[1203]	Speed: 117.060240 samples/sec	 accuracy=69.453947	 loss=1.187153	 lr=0.010000
Epoch[024] Batch [0399]/[1203]	Speed: 106.449936 samples/sec	 accuracy=69.521250	 loss=1.185416	 lr=0.010000
Epoch[024] Batch [0419]/[1203]	Speed: 161.135891 samples/sec	 accuracy=69.490476	 loss=1.185823	 lr=0.010000
Epoch[024] Batch [0439]/[1203]	Speed: 107.550968 samples/sec	 accuracy=69.448864	 loss=1.188102	 lr=0.010000
Epoch[024] Batch [0459]/[1203]	Speed: 131.857188 samples/sec	 accuracy=69.418478	 loss=1.188174	 lr=0.010000
Epoch[024] Batch [0479]/[1203]	Speed: 123.852275 samples/sec	 accuracy=69.388542	 loss=1.189569	 lr=0.010000
Epoch[024] Batch [0499]/[1203]	Speed: 119.562166 samples/sec	 accuracy=69.305000	 loss=1.192866	 lr=0.010000
Epoch[024] Batch [0519]/[1203]	Speed: 145.893950 samples/sec	 accuracy=69.297115	 loss=1.193217	 lr=0.010000
Epoch[024] Batch [0539]/[1203]	Speed: 113.245500 samples/sec	 accuracy=69.237037	 loss=1.194873	 lr=0.010000
Epoch[024] Batch [0559]/[1203]	Speed: 113.620849 samples/sec	 accuracy=69.153571	 loss=1.197409	 lr=0.010000
Epoch[024] Batch [0579]/[1203]	Speed: 143.111605 samples/sec	 accuracy=69.114655	 loss=1.198190	 lr=0.010000
Epoch[024] Batch [0599]/[1203]	Speed: 113.569991 samples/sec	 accuracy=69.078333	 loss=1.198405	 lr=0.010000
Epoch[024] Batch [0619]/[1203]	Speed: 137.732205 samples/sec	 accuracy=69.046774	 loss=1.200357	 lr=0.010000
Epoch[024] Batch [0639]/[1203]	Speed: 123.032102 samples/sec	 accuracy=69.029688	 loss=1.201246	 lr=0.010000
Epoch[024] Batch [0659]/[1203]	Speed: 124.851126 samples/sec	 accuracy=69.055303	 loss=1.200738	 lr=0.010000
Epoch[024] Batch [0679]/[1203]	Speed: 126.376698 samples/sec	 accuracy=69.064706	 loss=1.200665	 lr=0.010000
Epoch[024] Batch [0699]/[1203]	Speed: 114.655896 samples/sec	 accuracy=69.027143	 loss=1.202182	 lr=0.010000
Epoch[024] Batch [0719]/[1203]	Speed: 154.779970 samples/sec	 accuracy=69.011111	 loss=1.202866	 lr=0.010000
Epoch[024] Batch [0739]/[1203]	Speed: 104.902424 samples/sec	 accuracy=69.002027	 loss=1.203262	 lr=0.010000
Epoch[024] Batch [0759]/[1203]	Speed: 112.535239 samples/sec	 accuracy=68.969737	 loss=1.204304	 lr=0.010000
Epoch[024] Batch [0779]/[1203]	Speed: 171.701801 samples/sec	 accuracy=68.919231	 loss=1.205639	 lr=0.010000
Epoch[024] Batch [0799]/[1203]	Speed: 112.216119 samples/sec	 accuracy=68.910000	 loss=1.206543	 lr=0.010000
Epoch[024] Batch [0819]/[1203]	Speed: 122.664248 samples/sec	 accuracy=68.889024	 loss=1.207185	 lr=0.010000
Epoch[024] Batch [0839]/[1203]	Speed: 122.979834 samples/sec	 accuracy=68.872619	 loss=1.208372	 lr=0.010000
Epoch[024] Batch [0859]/[1203]	Speed: 117.648445 samples/sec	 accuracy=68.847093	 loss=1.209157	 lr=0.010000
Epoch[024] Batch [0879]/[1203]	Speed: 160.508909 samples/sec	 accuracy=68.831250	 loss=1.209977	 lr=0.010000
Epoch[024] Batch [0899]/[1203]	Speed: 111.973702 samples/sec	 accuracy=68.808889	 loss=1.211310	 lr=0.010000
Epoch[024] Batch [0919]/[1203]	Speed: 109.870249 samples/sec	 accuracy=68.783696	 loss=1.212304	 lr=0.010000
Epoch[024] Batch [0939]/[1203]	Speed: 129.507144 samples/sec	 accuracy=68.777660	 loss=1.212204	 lr=0.010000
Epoch[024] Batch [0959]/[1203]	Speed: 140.362016 samples/sec	 accuracy=68.787500	 loss=1.212393	 lr=0.010000
Epoch[024] Batch [0979]/[1203]	Speed: 124.179497 samples/sec	 accuracy=68.780612	 loss=1.212904	 lr=0.010000
Epoch[024] Batch [0999]/[1203]	Speed: 113.222328 samples/sec	 accuracy=68.759500	 loss=1.213929	 lr=0.010000
Epoch[024] Batch [1019]/[1203]	Speed: 127.202579 samples/sec	 accuracy=68.740686	 loss=1.215377	 lr=0.010000
Epoch[024] Batch [1039]/[1203]	Speed: 120.426820 samples/sec	 accuracy=68.709135	 loss=1.216829	 lr=0.010000
Epoch[024] Batch [1059]/[1203]	Speed: 138.511838 samples/sec	 accuracy=68.680189	 loss=1.217752	 lr=0.010000
Epoch[024] Batch [1079]/[1203]	Speed: 118.362937 samples/sec	 accuracy=68.652315	 loss=1.218575	 lr=0.010000
Epoch[024] Batch [1099]/[1203]	Speed: 115.265557 samples/sec	 accuracy=68.639091	 loss=1.219271	 lr=0.010000
Epoch[024] Batch [1119]/[1203]	Speed: 138.094324 samples/sec	 accuracy=68.613839	 loss=1.220351	 lr=0.010000
Epoch[024] Batch [1139]/[1203]	Speed: 135.585183 samples/sec	 accuracy=68.611404	 loss=1.220536	 lr=0.010000
Epoch[024] Batch [1159]/[1203]	Speed: 108.543317 samples/sec	 accuracy=68.599138	 loss=1.220768	 lr=0.010000
Epoch[024] Batch [1179]/[1203]	Speed: 143.755607 samples/sec	 accuracy=68.588136	 loss=1.220954	 lr=0.010000
Epoch[024] Batch [1199]/[1203]	Speed: 219.435286 samples/sec	 accuracy=68.594583	 loss=1.220895	 lr=0.010000
Batch [0019]/[0097]: acc-top1=58.375000 acc-top5=80.950000
Batch [0039]/[0097]: acc-top1=57.125000 acc-top5=80.462500
Batch [0059]/[0097]: acc-top1=59.925000 acc-top5=82.091667
Batch [0079]/[0097]: acc-top1=59.987500 acc-top5=82.275000
[Epoch 024] training: accuracy=68.593932	 loss=1.221103
[Epoch 024] speed: 123 samples/sec	time cost: 2135.729312
[Epoch 024] validation: acc-top1=59.262887 acc-top5=82.087629 loss=1.792807
Epoch[025] Batch [0019]/[1203]	Speed: 69.580129 samples/sec	 accuracy=70.250000	 loss=1.150951	 lr=0.010000
Epoch[025] Batch [0039]/[1203]	Speed: 101.308949 samples/sec	 accuracy=70.362500	 loss=1.155682	 lr=0.010000
Epoch[025] Batch [0059]/[1203]	Speed: 225.656734 samples/sec	 accuracy=70.266667	 loss=1.157864	 lr=0.010000
Epoch[025] Batch [0079]/[1203]	Speed: 93.857250 samples/sec	 accuracy=70.356250	 loss=1.162379	 lr=0.010000
Epoch[025] Batch [0099]/[1203]	Speed: 110.691126 samples/sec	 accuracy=70.335000	 loss=1.159609	 lr=0.010000
Epoch[025] Batch [0119]/[1203]	Speed: 179.823812 samples/sec	 accuracy=70.145833	 loss=1.168090	 lr=0.010000
Epoch[025] Batch [0139]/[1203]	Speed: 104.545966 samples/sec	 accuracy=70.000000	 loss=1.172258	 lr=0.010000
Epoch[025] Batch [0159]/[1203]	Speed: 183.753671 samples/sec	 accuracy=70.075000	 loss=1.169660	 lr=0.010000
Epoch[025] Batch [0179]/[1203]	Speed: 94.742764 samples/sec	 accuracy=70.011111	 loss=1.173307	 lr=0.010000
Epoch[025] Batch [0199]/[1203]	Speed: 124.739246 samples/sec	 accuracy=69.852500	 loss=1.179237	 lr=0.010000
Epoch[025] Batch [0219]/[1203]	Speed: 155.444457 samples/sec	 accuracy=69.820455	 loss=1.177388	 lr=0.010000
Epoch[025] Batch [0239]/[1203]	Speed: 96.420160 samples/sec	 accuracy=69.764583	 loss=1.178288	 lr=0.010000
Epoch[025] Batch [0259]/[1203]	Speed: 171.770955 samples/sec	 accuracy=69.786538	 loss=1.177092	 lr=0.010000
Epoch[025] Batch [0279]/[1203]	Speed: 115.147873 samples/sec	 accuracy=69.792857	 loss=1.175604	 lr=0.010000
Epoch[025] Batch [0299]/[1203]	Speed: 106.982867 samples/sec	 accuracy=69.785000	 loss=1.174796	 lr=0.010000
Epoch[025] Batch [0319]/[1203]	Speed: 172.799520 samples/sec	 accuracy=69.728125	 loss=1.177765	 lr=0.010000
Epoch[025] Batch [0339]/[1203]	Speed: 107.383556 samples/sec	 accuracy=69.698529	 loss=1.178494	 lr=0.010000
Epoch[025] Batch [0359]/[1203]	Speed: 124.988920 samples/sec	 accuracy=69.677778	 loss=1.180322	 lr=0.010000
Epoch[025] Batch [0379]/[1203]	Speed: 133.603345 samples/sec	 accuracy=69.677632	 loss=1.179433	 lr=0.010000
Epoch[025] Batch [0399]/[1203]	Speed: 101.662473 samples/sec	 accuracy=69.640000	 loss=1.180502	 lr=0.010000
Epoch[025] Batch [0419]/[1203]	Speed: 200.783353 samples/sec	 accuracy=69.623810	 loss=1.182136	 lr=0.010000
Epoch[025] Batch [0439]/[1203]	Speed: 99.562784 samples/sec	 accuracy=69.623864	 loss=1.182838	 lr=0.010000
Epoch[025] Batch [0459]/[1203]	Speed: 105.766566 samples/sec	 accuracy=69.604348	 loss=1.183485	 lr=0.010000
Epoch[025] Batch [0479]/[1203]	Speed: 173.815409 samples/sec	 accuracy=69.605208	 loss=1.183781	 lr=0.010000
Epoch[025] Batch [0499]/[1203]	Speed: 108.211219 samples/sec	 accuracy=69.553000	 loss=1.184689	 lr=0.010000
Epoch[025] Batch [0519]/[1203]	Speed: 175.962434 samples/sec	 accuracy=69.550000	 loss=1.185091	 lr=0.010000
Epoch[025] Batch [0539]/[1203]	Speed: 106.149124 samples/sec	 accuracy=69.537963	 loss=1.185050	 lr=0.010000
Epoch[025] Batch [0559]/[1203]	Speed: 107.076262 samples/sec	 accuracy=69.466964	 loss=1.188101	 lr=0.010000
Epoch[025] Batch [0579]/[1203]	Speed: 152.728445 samples/sec	 accuracy=69.413793	 loss=1.190298	 lr=0.010000
Epoch[025] Batch [0599]/[1203]	Speed: 112.730583 samples/sec	 accuracy=69.406667	 loss=1.191373	 lr=0.010000
Epoch[025] Batch [0619]/[1203]	Speed: 124.102036 samples/sec	 accuracy=69.394355	 loss=1.191704	 lr=0.010000
Epoch[025] Batch [0639]/[1203]	Speed: 137.322035 samples/sec	 accuracy=69.418750	 loss=1.191389	 lr=0.010000
Epoch[025] Batch [0659]/[1203]	Speed: 108.309744 samples/sec	 accuracy=69.414394	 loss=1.191484	 lr=0.010000
Epoch[025] Batch [0679]/[1203]	Speed: 148.834933 samples/sec	 accuracy=69.364706	 loss=1.192732	 lr=0.010000
Epoch[025] Batch [0699]/[1203]	Speed: 112.442805 samples/sec	 accuracy=69.362143	 loss=1.193294	 lr=0.010000
Epoch[025] Batch [0719]/[1203]	Speed: 136.953753 samples/sec	 accuracy=69.347917	 loss=1.193818	 lr=0.010000
Epoch[025] Batch [0739]/[1203]	Speed: 124.613293 samples/sec	 accuracy=69.345946	 loss=1.193999	 lr=0.010000
Epoch[025] Batch [0759]/[1203]	Speed: 118.463983 samples/sec	 accuracy=69.323026	 loss=1.195383	 lr=0.010000
Epoch[025] Batch [0779]/[1203]	Speed: 128.212493 samples/sec	 accuracy=69.290385	 loss=1.196128	 lr=0.010000
Epoch[025] Batch [0799]/[1203]	Speed: 122.480567 samples/sec	 accuracy=69.269375	 loss=1.196199	 lr=0.010000
Epoch[025] Batch [0819]/[1203]	Speed: 111.988131 samples/sec	 accuracy=69.246951	 loss=1.197123	 lr=0.010000
Epoch[025] Batch [0839]/[1203]	Speed: 143.736126 samples/sec	 accuracy=69.242262	 loss=1.197265	 lr=0.010000
Epoch[025] Batch [0859]/[1203]	Speed: 120.252405 samples/sec	 accuracy=69.227907	 loss=1.197672	 lr=0.010000
Epoch[025] Batch [0879]/[1203]	Speed: 125.834779 samples/sec	 accuracy=69.217045	 loss=1.197982	 lr=0.010000
Epoch[025] Batch [0899]/[1203]	Speed: 126.245542 samples/sec	 accuracy=69.203889	 loss=1.198718	 lr=0.010000
Epoch[025] Batch [0919]/[1203]	Speed: 118.283946 samples/sec	 accuracy=69.152174	 loss=1.199867	 lr=0.010000
Epoch[025] Batch [0939]/[1203]	Speed: 145.268399 samples/sec	 accuracy=69.154255	 loss=1.199731	 lr=0.010000
Epoch[025] Batch [0959]/[1203]	Speed: 123.469733 samples/sec	 accuracy=69.145313	 loss=1.200562	 lr=0.010000
Epoch[025] Batch [0979]/[1203]	Speed: 120.838024 samples/sec	 accuracy=69.123980	 loss=1.201132	 lr=0.010000
Epoch[025] Batch [0999]/[1203]	Speed: 118.053386 samples/sec	 accuracy=69.136500	 loss=1.201135	 lr=0.010000
Epoch[025] Batch [1019]/[1203]	Speed: 127.911798 samples/sec	 accuracy=69.125490	 loss=1.202180	 lr=0.010000
Epoch[025] Batch [1039]/[1203]	Speed: 141.888984 samples/sec	 accuracy=69.099519	 loss=1.202995	 lr=0.010000
Epoch[025] Batch [1059]/[1203]	Speed: 124.126049 samples/sec	 accuracy=69.096698	 loss=1.203792	 lr=0.010000
Epoch[025] Batch [1079]/[1203]	Speed: 114.080596 samples/sec	 accuracy=69.083796	 loss=1.204281	 lr=0.010000
Epoch[025] Batch [1099]/[1203]	Speed: 123.696579 samples/sec	 accuracy=69.073182	 loss=1.204821	 lr=0.010000
Epoch[025] Batch [1119]/[1203]	Speed: 130.764589 samples/sec	 accuracy=69.070982	 loss=1.205088	 lr=0.010000
Epoch[025] Batch [1139]/[1203]	Speed: 136.210602 samples/sec	 accuracy=69.046930	 loss=1.205597	 lr=0.010000
Epoch[025] Batch [1159]/[1203]	Speed: 112.144963 samples/sec	 accuracy=69.029741	 loss=1.206705	 lr=0.010000
Epoch[025] Batch [1179]/[1203]	Speed: 122.101788 samples/sec	 accuracy=69.003390	 loss=1.207420	 lr=0.010000
Epoch[025] Batch [1199]/[1203]	Speed: 254.417407 samples/sec	 accuracy=68.990000	 loss=1.208275	 lr=0.010000
Batch [0019]/[0097]: acc-top1=59.700000 acc-top5=81.875000
Batch [0039]/[0097]: acc-top1=58.337500 acc-top5=81.400000
Batch [0059]/[0097]: acc-top1=60.083333 acc-top5=82.683333
Batch [0079]/[0097]: acc-top1=60.206250 acc-top5=82.612500
[Epoch 025] training: accuracy=68.986284	 loss=1.208513
[Epoch 025] speed: 124 samples/sec	time cost: 2131.093673
[Epoch 025] validation: acc-top1=59.762887 acc-top5=82.587629 loss=1.765419
Epoch[026] Batch [0019]/[1203]	Speed: 69.182453 samples/sec	 accuracy=71.750000	 loss=1.091851	 lr=0.010000
Epoch[026] Batch [0039]/[1203]	Speed: 100.758341 samples/sec	 accuracy=71.062500	 loss=1.114587	 lr=0.010000
Epoch[026] Batch [0059]/[1203]	Speed: 233.230178 samples/sec	 accuracy=70.891667	 loss=1.123789	 lr=0.010000
Epoch[026] Batch [0079]/[1203]	Speed: 91.958586 samples/sec	 accuracy=70.500000	 loss=1.151605	 lr=0.010000
Epoch[026] Batch [0099]/[1203]	Speed: 115.225460 samples/sec	 accuracy=70.765000	 loss=1.138693	 lr=0.010000
Epoch[026] Batch [0119]/[1203]	Speed: 170.216645 samples/sec	 accuracy=70.625000	 loss=1.144969	 lr=0.010000
Epoch[026] Batch [0139]/[1203]	Speed: 97.146881 samples/sec	 accuracy=70.425000	 loss=1.149663	 lr=0.010000
Epoch[026] Batch [0159]/[1203]	Speed: 227.502022 samples/sec	 accuracy=70.415625	 loss=1.145675	 lr=0.010000
Epoch[026] Batch [0179]/[1203]	Speed: 90.585390 samples/sec	 accuracy=70.347222	 loss=1.149701	 lr=0.010000
Epoch[026] Batch [0199]/[1203]	Speed: 154.770597 samples/sec	 accuracy=70.245000	 loss=1.155362	 lr=0.010000
Epoch[026] Batch [0219]/[1203]	Speed: 119.109762 samples/sec	 accuracy=70.184091	 loss=1.156156	 lr=0.010000
Epoch[026] Batch [0239]/[1203]	Speed: 103.816171 samples/sec	 accuracy=70.104167	 loss=1.160653	 lr=0.010000
Epoch[026] Batch [0259]/[1203]	Speed: 165.667492 samples/sec	 accuracy=70.042308	 loss=1.163799	 lr=0.010000
Epoch[026] Batch [0279]/[1203]	Speed: 109.814469 samples/sec	 accuracy=69.985714	 loss=1.165283	 lr=0.010000
Epoch[026] Batch [0299]/[1203]	Speed: 112.793107 samples/sec	 accuracy=69.973333	 loss=1.165904	 lr=0.010000
Epoch[026] Batch [0319]/[1203]	Speed: 150.360569 samples/sec	 accuracy=69.901562	 loss=1.169369	 lr=0.010000
Epoch[026] Batch [0339]/[1203]	Speed: 103.857359 samples/sec	 accuracy=69.810294	 loss=1.173248	 lr=0.010000
Epoch[026] Batch [0359]/[1203]	Speed: 188.092886 samples/sec	 accuracy=69.791667	 loss=1.173881	 lr=0.010000
Epoch[026] Batch [0379]/[1203]	Speed: 109.735214 samples/sec	 accuracy=69.792105	 loss=1.174131	 lr=0.010000
Epoch[026] Batch [0399]/[1203]	Speed: 112.380494 samples/sec	 accuracy=69.803750	 loss=1.173702	 lr=0.010000
Epoch[026] Batch [0419]/[1203]	Speed: 146.723435 samples/sec	 accuracy=69.725000	 loss=1.176052	 lr=0.010000
Epoch[026] Batch [0439]/[1203]	Speed: 107.285926 samples/sec	 accuracy=69.702273	 loss=1.176693	 lr=0.010000
Epoch[026] Batch [0459]/[1203]	Speed: 127.776986 samples/sec	 accuracy=69.708696	 loss=1.174933	 lr=0.010000
Epoch[026] Batch [0479]/[1203]	Speed: 139.941901 samples/sec	 accuracy=69.678125	 loss=1.175838	 lr=0.010000
Epoch[026] Batch [0499]/[1203]	Speed: 112.651972 samples/sec	 accuracy=69.685000	 loss=1.176935	 lr=0.010000
Epoch[026] Batch [0519]/[1203]	Speed: 133.355477 samples/sec	 accuracy=69.670192	 loss=1.177335	 lr=0.010000
Epoch[026] Batch [0539]/[1203]	Speed: 121.255786 samples/sec	 accuracy=69.612037	 loss=1.180049	 lr=0.010000
Epoch[026] Batch [0559]/[1203]	Speed: 124.295689 samples/sec	 accuracy=69.575893	 loss=1.180462	 lr=0.010000
Epoch[026] Batch [0579]/[1203]	Speed: 141.515175 samples/sec	 accuracy=69.578448	 loss=1.180367	 lr=0.010000
Epoch[026] Batch [0599]/[1203]	Speed: 107.501353 samples/sec	 accuracy=69.593333	 loss=1.180076	 lr=0.010000
Epoch[026] Batch [0619]/[1203]	Speed: 141.275677 samples/sec	 accuracy=69.558065	 loss=1.180862	 lr=0.010000
Epoch[026] Batch [0639]/[1203]	Speed: 126.486895 samples/sec	 accuracy=69.539062	 loss=1.181098	 lr=0.010000
Epoch[026] Batch [0659]/[1203]	Speed: 112.422573 samples/sec	 accuracy=69.542424	 loss=1.181615	 lr=0.010000
Epoch[026] Batch [0679]/[1203]	Speed: 167.022150 samples/sec	 accuracy=69.546324	 loss=1.182065	 lr=0.010000
Epoch[026] Batch [0699]/[1203]	Speed: 100.091118 samples/sec	 accuracy=69.554286	 loss=1.181297	 lr=0.010000
Epoch[026] Batch [0719]/[1203]	Speed: 129.412624 samples/sec	 accuracy=69.496528	 loss=1.182086	 lr=0.010000
Epoch[026] Batch [0739]/[1203]	Speed: 136.127680 samples/sec	 accuracy=69.476351	 loss=1.182875	 lr=0.010000
Epoch[026] Batch [0759]/[1203]	Speed: 112.550129 samples/sec	 accuracy=69.472368	 loss=1.183756	 lr=0.010000
Epoch[026] Batch [0779]/[1203]	Speed: 119.407339 samples/sec	 accuracy=69.459615	 loss=1.184031	 lr=0.010000
Epoch[026] Batch [0799]/[1203]	Speed: 146.610234 samples/sec	 accuracy=69.441250	 loss=1.184256	 lr=0.010000
Epoch[026] Batch [0819]/[1203]	Speed: 114.553622 samples/sec	 accuracy=69.421951	 loss=1.185319	 lr=0.010000
Epoch[026] Batch [0839]/[1203]	Speed: 145.724178 samples/sec	 accuracy=69.392857	 loss=1.186936	 lr=0.010000
Epoch[026] Batch [0859]/[1203]	Speed: 118.308273 samples/sec	 accuracy=69.381977	 loss=1.187078	 lr=0.010000
Epoch[026] Batch [0879]/[1203]	Speed: 99.874667 samples/sec	 accuracy=69.378977	 loss=1.186756	 lr=0.010000
Epoch[026] Batch [0899]/[1203]	Speed: 152.838993 samples/sec	 accuracy=69.393333	 loss=1.187115	 lr=0.010000
Epoch[026] Batch [0919]/[1203]	Speed: 118.891675 samples/sec	 accuracy=69.399457	 loss=1.186933	 lr=0.010000
Epoch[026] Batch [0939]/[1203]	Speed: 123.328670 samples/sec	 accuracy=69.375000	 loss=1.188074	 lr=0.010000
Epoch[026] Batch [0959]/[1203]	Speed: 144.955711 samples/sec	 accuracy=69.364583	 loss=1.188075	 lr=0.010000
Epoch[026] Batch [0979]/[1203]	Speed: 109.645302 samples/sec	 accuracy=69.356122	 loss=1.188229	 lr=0.010000
Epoch[026] Batch [0999]/[1203]	Speed: 140.538297 samples/sec	 accuracy=69.362000	 loss=1.187947	 lr=0.010000
Epoch[026] Batch [1019]/[1203]	Speed: 118.003953 samples/sec	 accuracy=69.329412	 loss=1.189265	 lr=0.010000
Epoch[026] Batch [1039]/[1203]	Speed: 102.644989 samples/sec	 accuracy=69.298077	 loss=1.190340	 lr=0.010000
Epoch[026] Batch [1059]/[1203]	Speed: 174.181790 samples/sec	 accuracy=69.271698	 loss=1.190816	 lr=0.010000
Epoch[026] Batch [1079]/[1203]	Speed: 114.107738 samples/sec	 accuracy=69.249537	 loss=1.191521	 lr=0.010000
Epoch[026] Batch [1099]/[1203]	Speed: 140.790065 samples/sec	 accuracy=69.224091	 loss=1.192385	 lr=0.010000
Epoch[026] Batch [1119]/[1203]	Speed: 121.634906 samples/sec	 accuracy=69.236161	 loss=1.191617	 lr=0.010000
Epoch[026] Batch [1139]/[1203]	Speed: 107.102681 samples/sec	 accuracy=69.226316	 loss=1.192295	 lr=0.010000
Epoch[026] Batch [1159]/[1203]	Speed: 146.533003 samples/sec	 accuracy=69.219828	 loss=1.192732	 lr=0.010000
Epoch[026] Batch [1179]/[1203]	Speed: 121.183144 samples/sec	 accuracy=69.203814	 loss=1.193645	 lr=0.010000
Epoch[026] Batch [1199]/[1203]	Speed: 253.668066 samples/sec	 accuracy=69.186250	 loss=1.194404	 lr=0.010000
Batch [0019]/[0097]: acc-top1=60.575000 acc-top5=82.750000
Batch [0039]/[0097]: acc-top1=59.012500 acc-top5=81.500000
Batch [0059]/[0097]: acc-top1=61.325000 acc-top5=83.150000
Batch [0079]/[0097]: acc-top1=61.593750 acc-top5=83.168750
[Epoch 026] training: accuracy=69.182045	 loss=1.194568
[Epoch 026] speed: 124 samples/sec	time cost: 2133.576937
[Epoch 026] validation: acc-top1=60.639175 acc-top5=82.876289 loss=1.719018
Epoch[027] Batch [0019]/[1203]	Speed: 69.083674 samples/sec	 accuracy=71.225000	 loss=1.101360	 lr=0.010000
Epoch[027] Batch [0039]/[1203]	Speed: 103.225968 samples/sec	 accuracy=70.700000	 loss=1.126170	 lr=0.010000
Epoch[027] Batch [0059]/[1203]	Speed: 205.676067 samples/sec	 accuracy=70.808333	 loss=1.129093	 lr=0.010000
Epoch[027] Batch [0079]/[1203]	Speed: 103.171609 samples/sec	 accuracy=70.481250	 loss=1.139116	 lr=0.010000
Epoch[027] Batch [0099]/[1203]	Speed: 109.083499 samples/sec	 accuracy=70.670000	 loss=1.129941	 lr=0.010000
Epoch[027] Batch [0119]/[1203]	Speed: 170.576568 samples/sec	 accuracy=70.754167	 loss=1.123285	 lr=0.010000
Epoch[027] Batch [0139]/[1203]	Speed: 103.268237 samples/sec	 accuracy=70.650000	 loss=1.126854	 lr=0.010000
Epoch[027] Batch [0159]/[1203]	Speed: 188.842613 samples/sec	 accuracy=70.684375	 loss=1.129732	 lr=0.010000
Epoch[027] Batch [0179]/[1203]	Speed: 101.115513 samples/sec	 accuracy=70.563889	 loss=1.136094	 lr=0.010000
Epoch[027] Batch [0199]/[1203]	Speed: 99.102344 samples/sec	 accuracy=70.427500	 loss=1.143940	 lr=0.010000
Epoch[027] Batch [0219]/[1203]	Speed: 196.750902 samples/sec	 accuracy=70.347727	 loss=1.146343	 lr=0.010000
Epoch[027] Batch [0239]/[1203]	Speed: 100.642593 samples/sec	 accuracy=70.218750	 loss=1.148907	 lr=0.010000
Epoch[027] Batch [0259]/[1203]	Speed: 146.900985 samples/sec	 accuracy=70.215385	 loss=1.150385	 lr=0.010000
Epoch[027] Batch [0279]/[1203]	Speed: 120.520965 samples/sec	 accuracy=70.192857	 loss=1.150797	 lr=0.010000
Epoch[027] Batch [0299]/[1203]	Speed: 105.386589 samples/sec	 accuracy=70.175000	 loss=1.153920	 lr=0.010000
Epoch[027] Batch [0319]/[1203]	Speed: 199.787335 samples/sec	 accuracy=70.226562	 loss=1.154000	 lr=0.010000
Epoch[027] Batch [0339]/[1203]	Speed: 100.229116 samples/sec	 accuracy=70.208824	 loss=1.153776	 lr=0.010000
Epoch[027] Batch [0359]/[1203]	Speed: 111.947219 samples/sec	 accuracy=70.163889	 loss=1.154412	 lr=0.010000
Epoch[027] Batch [0379]/[1203]	Speed: 168.349753 samples/sec	 accuracy=70.151316	 loss=1.154439	 lr=0.010000
Epoch[027] Batch [0399]/[1203]	Speed: 103.134587 samples/sec	 accuracy=70.166250	 loss=1.154802	 lr=0.010000
Epoch[027] Batch [0419]/[1203]	Speed: 175.700324 samples/sec	 accuracy=70.195238	 loss=1.153788	 lr=0.010000
Epoch[027] Batch [0439]/[1203]	Speed: 107.624169 samples/sec	 accuracy=70.176136	 loss=1.153799	 lr=0.010000
Epoch[027] Batch [0459]/[1203]	Speed: 115.008409 samples/sec	 accuracy=70.143478	 loss=1.155036	 lr=0.010000
Epoch[027] Batch [0479]/[1203]	Speed: 150.013919 samples/sec	 accuracy=70.127083	 loss=1.156948	 lr=0.010000
Epoch[027] Batch [0499]/[1203]	Speed: 102.450837 samples/sec	 accuracy=70.092000	 loss=1.157356	 lr=0.010000
Epoch[027] Batch [0519]/[1203]	Speed: 127.037472 samples/sec	 accuracy=70.043269	 loss=1.158191	 lr=0.010000
Epoch[027] Batch [0539]/[1203]	Speed: 132.001814 samples/sec	 accuracy=70.013889	 loss=1.159567	 lr=0.010000
Epoch[027] Batch [0559]/[1203]	Speed: 121.373787 samples/sec	 accuracy=70.005357	 loss=1.159592	 lr=0.010000
Epoch[027] Batch [0579]/[1203]	Speed: 133.663458 samples/sec	 accuracy=69.976724	 loss=1.160705	 lr=0.010000
Epoch[027] Batch [0599]/[1203]	Speed: 112.866819 samples/sec	 accuracy=69.952500	 loss=1.162332	 lr=0.010000
Epoch[027] Batch [0619]/[1203]	Speed: 131.023018 samples/sec	 accuracy=69.933065	 loss=1.162900	 lr=0.010000
Epoch[027] Batch [0639]/[1203]	Speed: 126.511381 samples/sec	 accuracy=69.917187	 loss=1.162859	 lr=0.010000
Epoch[027] Batch [0659]/[1203]	Speed: 136.877630 samples/sec	 accuracy=69.900758	 loss=1.164385	 lr=0.010000
Epoch[027] Batch [0679]/[1203]	Speed: 117.366601 samples/sec	 accuracy=69.897794	 loss=1.164836	 lr=0.010000
Epoch[027] Batch [0699]/[1203]	Speed: 119.363886 samples/sec	 accuracy=69.865000	 loss=1.166349	 lr=0.010000
Epoch[027] Batch [0719]/[1203]	Speed: 136.890542 samples/sec	 accuracy=69.841667	 loss=1.167008	 lr=0.010000
Epoch[027] Batch [0739]/[1203]	Speed: 108.007147 samples/sec	 accuracy=69.809459	 loss=1.168253	 lr=0.010000
Epoch[027] Batch [0759]/[1203]	Speed: 134.498883 samples/sec	 accuracy=69.806579	 loss=1.169146	 lr=0.010000
Epoch[027] Batch [0779]/[1203]	Speed: 124.473438 samples/sec	 accuracy=69.778846	 loss=1.170036	 lr=0.010000
Epoch[027] Batch [0799]/[1203]	Speed: 111.302935 samples/sec	 accuracy=69.770625	 loss=1.169801	 lr=0.010000
Epoch[027] Batch [0819]/[1203]	Speed: 163.958812 samples/sec	 accuracy=69.751220	 loss=1.170910	 lr=0.010000
Epoch[027] Batch [0839]/[1203]	Speed: 107.009570 samples/sec	 accuracy=69.769048	 loss=1.170317	 lr=0.010000
Epoch[027] Batch [0859]/[1203]	Speed: 131.918167 samples/sec	 accuracy=69.758140	 loss=1.171434	 lr=0.010000
Epoch[027] Batch [0879]/[1203]	Speed: 125.949177 samples/sec	 accuracy=69.718750	 loss=1.172888	 lr=0.010000
Epoch[027] Batch [0899]/[1203]	Speed: 105.695432 samples/sec	 accuracy=69.702222	 loss=1.173282	 lr=0.010000
Epoch[027] Batch [0919]/[1203]	Speed: 169.070057 samples/sec	 accuracy=69.697283	 loss=1.173358	 lr=0.010000
Epoch[027] Batch [0939]/[1203]	Speed: 112.887180 samples/sec	 accuracy=69.675532	 loss=1.174071	 lr=0.010000
Epoch[027] Batch [0959]/[1203]	Speed: 111.397976 samples/sec	 accuracy=69.623958	 loss=1.175835	 lr=0.010000
Epoch[027] Batch [0979]/[1203]	Speed: 147.399862 samples/sec	 accuracy=69.613265	 loss=1.176471	 lr=0.010000
Epoch[027] Batch [0999]/[1203]	Speed: 109.430304 samples/sec	 accuracy=69.607000	 loss=1.176708	 lr=0.010000
Epoch[027] Batch [1019]/[1203]	Speed: 145.957577 samples/sec	 accuracy=69.591176	 loss=1.177771	 lr=0.010000
Epoch[027] Batch [1039]/[1203]	Speed: 119.633192 samples/sec	 accuracy=69.565865	 loss=1.178712	 lr=0.010000
Epoch[027] Batch [1059]/[1203]	Speed: 115.841773 samples/sec	 accuracy=69.536792	 loss=1.179808	 lr=0.010000
Epoch[027] Batch [1079]/[1203]	Speed: 139.664094 samples/sec	 accuracy=69.506019	 loss=1.180784	 lr=0.010000
Epoch[027] Batch [1099]/[1203]	Speed: 111.629915 samples/sec	 accuracy=69.490909	 loss=1.181305	 lr=0.010000
Epoch[027] Batch [1119]/[1203]	Speed: 121.210171 samples/sec	 accuracy=69.495536	 loss=1.181512	 lr=0.010000
Epoch[027] Batch [1139]/[1203]	Speed: 148.357849 samples/sec	 accuracy=69.482895	 loss=1.182067	 lr=0.010000
Epoch[027] Batch [1159]/[1203]	Speed: 98.881668 samples/sec	 accuracy=69.458621	 loss=1.183587	 lr=0.010000
Epoch[027] Batch [1179]/[1203]	Speed: 187.920963 samples/sec	 accuracy=69.425000	 loss=1.184775	 lr=0.010000
Epoch[027] Batch [1199]/[1203]	Speed: 189.992419 samples/sec	 accuracy=69.378333	 loss=1.186417	 lr=0.010000
Batch [0019]/[0097]: acc-top1=58.825000 acc-top5=82.000000
Batch [0039]/[0097]: acc-top1=58.712500 acc-top5=80.937500
Batch [0059]/[0097]: acc-top1=60.883333 acc-top5=82.433333
Batch [0079]/[0097]: acc-top1=60.881250 acc-top5=82.387500
[Epoch 027] training: accuracy=69.377805	 loss=1.186627
[Epoch 027] speed: 123 samples/sec	time cost: 2136.268463
[Epoch 027] validation: acc-top1=60.170103 acc-top5=82.314433 loss=1.745874
Epoch[028] Batch [0019]/[1203]	Speed: 68.471279 samples/sec	 accuracy=71.150000	 loss=1.104157	 lr=0.010000
Epoch[028] Batch [0039]/[1203]	Speed: 99.855494 samples/sec	 accuracy=71.512500	 loss=1.095298	 lr=0.010000
Epoch[028] Batch [0059]/[1203]	Speed: 241.584307 samples/sec	 accuracy=71.633333	 loss=1.080066	 lr=0.010000
Epoch[028] Batch [0079]/[1203]	Speed: 93.148030 samples/sec	 accuracy=71.168750	 loss=1.092970	 lr=0.010000
Epoch[028] Batch [0099]/[1203]	Speed: 105.197104 samples/sec	 accuracy=70.935000	 loss=1.113770	 lr=0.010000
Epoch[028] Batch [0119]/[1203]	Speed: 189.568705 samples/sec	 accuracy=70.745833	 loss=1.123890	 lr=0.010000
Epoch[028] Batch [0139]/[1203]	Speed: 95.629948 samples/sec	 accuracy=70.932143	 loss=1.118691	 lr=0.010000
Epoch[028] Batch [0159]/[1203]	Speed: 213.810807 samples/sec	 accuracy=70.881250	 loss=1.121787	 lr=0.010000
Epoch[028] Batch [0179]/[1203]	Speed: 101.425938 samples/sec	 accuracy=70.841667	 loss=1.123307	 lr=0.010000
Epoch[028] Batch [0199]/[1203]	Speed: 112.539401 samples/sec	 accuracy=70.907500	 loss=1.121432	 lr=0.010000
Epoch[028] Batch [0219]/[1203]	Speed: 167.528590 samples/sec	 accuracy=70.893182	 loss=1.121785	 lr=0.010000
Epoch[028] Batch [0239]/[1203]	Speed: 97.257378 samples/sec	 accuracy=70.856250	 loss=1.123760	 lr=0.010000
Epoch[028] Batch [0259]/[1203]	Speed: 126.247340 samples/sec	 accuracy=70.826923	 loss=1.123620	 lr=0.010000
Epoch[028] Batch [0279]/[1203]	Speed: 145.236398 samples/sec	 accuracy=70.764286	 loss=1.125530	 lr=0.010000
Epoch[028] Batch [0299]/[1203]	Speed: 113.605797 samples/sec	 accuracy=70.768333	 loss=1.126693	 lr=0.010000
Epoch[028] Batch [0319]/[1203]	Speed: 151.808930 samples/sec	 accuracy=70.723438	 loss=1.129073	 lr=0.010000
Epoch[028] Batch [0339]/[1203]	Speed: 107.327000 samples/sec	 accuracy=70.686765	 loss=1.130212	 lr=0.010000
Epoch[028] Batch [0359]/[1203]	Speed: 131.601280 samples/sec	 accuracy=70.700000	 loss=1.131433	 lr=0.010000
Epoch[028] Batch [0379]/[1203]	Speed: 125.094426 samples/sec	 accuracy=70.622368	 loss=1.133468	 lr=0.010000
Epoch[028] Batch [0399]/[1203]	Speed: 124.020054 samples/sec	 accuracy=70.606250	 loss=1.134688	 lr=0.010000
Epoch[028] Batch [0419]/[1203]	Speed: 122.701601 samples/sec	 accuracy=70.617857	 loss=1.134054	 lr=0.010000
Epoch[028] Batch [0439]/[1203]	Speed: 116.432884 samples/sec	 accuracy=70.593182	 loss=1.135621	 lr=0.010000
Epoch[028] Batch [0459]/[1203]	Speed: 136.991418 samples/sec	 accuracy=70.632609	 loss=1.134558	 lr=0.010000
Epoch[028] Batch [0479]/[1203]	Speed: 119.998655 samples/sec	 accuracy=70.545833	 loss=1.137952	 lr=0.010000
Epoch[028] Batch [0499]/[1203]	Speed: 107.914091 samples/sec	 accuracy=70.518000	 loss=1.140051	 lr=0.010000
Epoch[028] Batch [0519]/[1203]	Speed: 175.100284 samples/sec	 accuracy=70.550000	 loss=1.140126	 lr=0.010000
Epoch[028] Batch [0539]/[1203]	Speed: 104.951565 samples/sec	 accuracy=70.523148	 loss=1.141204	 lr=0.010000
Epoch[028] Batch [0559]/[1203]	Speed: 142.393531 samples/sec	 accuracy=70.476786	 loss=1.142963	 lr=0.010000
Epoch[028] Batch [0579]/[1203]	Speed: 118.824962 samples/sec	 accuracy=70.444828	 loss=1.143819	 lr=0.010000
Epoch[028] Batch [0599]/[1203]	Speed: 106.316361 samples/sec	 accuracy=70.430000	 loss=1.144666	 lr=0.010000
Epoch[028] Batch [0619]/[1203]	Speed: 169.241385 samples/sec	 accuracy=70.408871	 loss=1.146179	 lr=0.010000
Epoch[028] Batch [0639]/[1203]	Speed: 107.012888 samples/sec	 accuracy=70.411719	 loss=1.146362	 lr=0.010000
Epoch[028] Batch [0659]/[1203]	Speed: 115.591880 samples/sec	 accuracy=70.369697	 loss=1.148576	 lr=0.010000
Epoch[028] Batch [0679]/[1203]	Speed: 164.846120 samples/sec	 accuracy=70.353676	 loss=1.149323	 lr=0.010000
Epoch[028] Batch [0699]/[1203]	Speed: 95.548116 samples/sec	 accuracy=70.308571	 loss=1.150932	 lr=0.010000
Epoch[028] Batch [0719]/[1203]	Speed: 187.139302 samples/sec	 accuracy=70.300000	 loss=1.151459	 lr=0.010000
Epoch[028] Batch [0739]/[1203]	Speed: 104.728717 samples/sec	 accuracy=70.262838	 loss=1.152545	 lr=0.010000
Epoch[028] Batch [0759]/[1203]	Speed: 118.568882 samples/sec	 accuracy=70.227632	 loss=1.153302	 lr=0.010000
Epoch[028] Batch [0779]/[1203]	Speed: 149.103384 samples/sec	 accuracy=70.171154	 loss=1.155298	 lr=0.010000
Epoch[028] Batch [0799]/[1203]	Speed: 98.506849 samples/sec	 accuracy=70.151250	 loss=1.156593	 lr=0.010000
Epoch[028] Batch [0819]/[1203]	Speed: 142.283383 samples/sec	 accuracy=70.135976	 loss=1.157858	 lr=0.010000
Epoch[028] Batch [0839]/[1203]	Speed: 129.073597 samples/sec	 accuracy=70.106548	 loss=1.158641	 lr=0.010000
Epoch[028] Batch [0859]/[1203]	Speed: 113.774222 samples/sec	 accuracy=70.091860	 loss=1.159430	 lr=0.010000
Epoch[028] Batch [0879]/[1203]	Speed: 151.068879 samples/sec	 accuracy=70.056250	 loss=1.161416	 lr=0.010000
Epoch[028] Batch [0899]/[1203]	Speed: 114.683647 samples/sec	 accuracy=70.023333	 loss=1.162671	 lr=0.010000
Epoch[028] Batch [0919]/[1203]	Speed: 126.085367 samples/sec	 accuracy=70.014674	 loss=1.163177	 lr=0.010000
Epoch[028] Batch [0939]/[1203]	Speed: 134.243888 samples/sec	 accuracy=70.006383	 loss=1.164063	 lr=0.010000
Epoch[028] Batch [0959]/[1203]	Speed: 109.857797 samples/sec	 accuracy=69.979167	 loss=1.165392	 lr=0.010000
Epoch[028] Batch [0979]/[1203]	Speed: 152.721924 samples/sec	 accuracy=69.952551	 loss=1.166231	 lr=0.010000
Epoch[028] Batch [0999]/[1203]	Speed: 104.648269 samples/sec	 accuracy=69.940500	 loss=1.166647	 lr=0.010000
Epoch[028] Batch [1019]/[1203]	Speed: 143.561014 samples/sec	 accuracy=69.918137	 loss=1.167156	 lr=0.010000
Epoch[028] Batch [1039]/[1203]	Speed: 126.641121 samples/sec	 accuracy=69.902885	 loss=1.167902	 lr=0.010000
Epoch[028] Batch [1059]/[1203]	Speed: 113.684125 samples/sec	 accuracy=69.890094	 loss=1.168484	 lr=0.010000
Epoch[028] Batch [1079]/[1203]	Speed: 131.868168 samples/sec	 accuracy=69.883796	 loss=1.169322	 lr=0.010000
Epoch[028] Batch [1099]/[1203]	Speed: 127.304440 samples/sec	 accuracy=69.875909	 loss=1.169438	 lr=0.010000
Epoch[028] Batch [1119]/[1203]	Speed: 123.219566 samples/sec	 accuracy=69.860268	 loss=1.169892	 lr=0.010000
Epoch[028] Batch [1139]/[1203]	Speed: 124.529807 samples/sec	 accuracy=69.850439	 loss=1.170555	 lr=0.010000
Epoch[028] Batch [1159]/[1203]	Speed: 118.948222 samples/sec	 accuracy=69.848707	 loss=1.171120	 lr=0.010000
Epoch[028] Batch [1179]/[1203]	Speed: 152.387834 samples/sec	 accuracy=69.836017	 loss=1.171818	 lr=0.010000
Epoch[028] Batch [1199]/[1203]	Speed: 212.110417 samples/sec	 accuracy=69.791250	 loss=1.173106	 lr=0.010000
Batch [0019]/[0097]: acc-top1=59.575000 acc-top5=80.825000
Batch [0039]/[0097]: acc-top1=58.312500 acc-top5=80.612500
Batch [0059]/[0097]: acc-top1=60.791667 acc-top5=82.150000
Batch [0079]/[0097]: acc-top1=61.225000 acc-top5=82.343750
[Epoch 028] training: accuracy=69.795927	 loss=1.173045
[Epoch 028] speed: 124 samples/sec	time cost: 2133.044338
[Epoch 028] validation: acc-top1=60.164948 acc-top5=82.170103 loss=1.763064
Epoch[029] Batch [0019]/[1203]	Speed: 68.976422 samples/sec	 accuracy=71.200000	 loss=1.121129	 lr=0.010000
Epoch[029] Batch [0039]/[1203]	Speed: 101.672190 samples/sec	 accuracy=70.625000	 loss=1.132241	 lr=0.010000
Epoch[029] Batch [0059]/[1203]	Speed: 222.809834 samples/sec	 accuracy=70.916667	 loss=1.116111	 lr=0.010000
Epoch[029] Batch [0079]/[1203]	Speed: 94.797501 samples/sec	 accuracy=70.750000	 loss=1.120434	 lr=0.010000
Epoch[029] Batch [0099]/[1203]	Speed: 109.350723 samples/sec	 accuracy=70.775000	 loss=1.118375	 lr=0.010000
Epoch[029] Batch [0119]/[1203]	Speed: 173.192173 samples/sec	 accuracy=70.916667	 loss=1.120122	 lr=0.010000
Epoch[029] Batch [0139]/[1203]	Speed: 104.966689 samples/sec	 accuracy=70.953571	 loss=1.120890	 lr=0.010000
Epoch[029] Batch [0159]/[1203]	Speed: 214.201749 samples/sec	 accuracy=70.984375	 loss=1.119781	 lr=0.010000
Epoch[029] Batch [0179]/[1203]	Speed: 91.593418 samples/sec	 accuracy=70.944444	 loss=1.121501	 lr=0.010000
Epoch[029] Batch [0199]/[1203]	Speed: 114.652438 samples/sec	 accuracy=71.130000	 loss=1.119311	 lr=0.010000
Epoch[029] Batch [0219]/[1203]	Speed: 172.809746 samples/sec	 accuracy=71.015909	 loss=1.120544	 lr=0.010000
Epoch[029] Batch [0239]/[1203]	Speed: 106.188449 samples/sec	 accuracy=70.989583	 loss=1.122337	 lr=0.010000
Epoch[029] Batch [0259]/[1203]	Speed: 133.304379 samples/sec	 accuracy=71.038462	 loss=1.121110	 lr=0.010000
Epoch[029] Batch [0279]/[1203]	Speed: 135.008243 samples/sec	 accuracy=70.969643	 loss=1.123154	 lr=0.010000
Epoch[029] Batch [0299]/[1203]	Speed: 98.824437 samples/sec	 accuracy=70.965000	 loss=1.124351	 lr=0.010000
Epoch[029] Batch [0319]/[1203]	Speed: 168.359506 samples/sec	 accuracy=70.995312	 loss=1.124131	 lr=0.010000
Epoch[029] Batch [0339]/[1203]	Speed: 113.018373 samples/sec	 accuracy=70.907353	 loss=1.128344	 lr=0.010000
Epoch[029] Batch [0359]/[1203]	Speed: 109.517459 samples/sec	 accuracy=70.902778	 loss=1.128188	 lr=0.010000
Epoch[029] Batch [0379]/[1203]	Speed: 164.252127 samples/sec	 accuracy=70.927632	 loss=1.128130	 lr=0.010000
Epoch[029] Batch [0399]/[1203]	Speed: 101.111366 samples/sec	 accuracy=70.957500	 loss=1.128377	 lr=0.010000
Epoch[029] Batch [0419]/[1203]	Speed: 161.760902 samples/sec	 accuracy=70.950000	 loss=1.128318	 lr=0.010000
Epoch[029] Batch [0439]/[1203]	Speed: 115.617916 samples/sec	 accuracy=70.935227	 loss=1.128942	 lr=0.010000
Epoch[029] Batch [0459]/[1203]	Speed: 114.272701 samples/sec	 accuracy=70.923913	 loss=1.127963	 lr=0.010000
Epoch[029] Batch [0479]/[1203]	Speed: 154.682841 samples/sec	 accuracy=70.863542	 loss=1.129141	 lr=0.010000
Epoch[029] Batch [0499]/[1203]	Speed: 106.484437 samples/sec	 accuracy=70.826000	 loss=1.131043	 lr=0.010000
Epoch[029] Batch [0519]/[1203]	Speed: 112.802813 samples/sec	 accuracy=70.862500	 loss=1.130060	 lr=0.010000
Epoch[029] Batch [0539]/[1203]	Speed: 146.268209 samples/sec	 accuracy=70.848148	 loss=1.130569	 lr=0.010000
Epoch[029] Batch [0559]/[1203]	Speed: 123.234306 samples/sec	 accuracy=70.791071	 loss=1.133447	 lr=0.010000
Epoch[029] Batch [0579]/[1203]	Speed: 152.775972 samples/sec	 accuracy=70.792241	 loss=1.133973	 lr=0.010000
Epoch[029] Batch [0599]/[1203]	Speed: 107.414720 samples/sec	 accuracy=70.751667	 loss=1.135025	 lr=0.010000
Epoch[029] Batch [0619]/[1203]	Speed: 123.559909 samples/sec	 accuracy=70.709677	 loss=1.136665	 lr=0.010000
Epoch[029] Batch [0639]/[1203]	Speed: 132.841580 samples/sec	 accuracy=70.700781	 loss=1.136943	 lr=0.010000
Epoch[029] Batch [0659]/[1203]	Speed: 125.273567 samples/sec	 accuracy=70.661364	 loss=1.138846	 lr=0.010000
Epoch[029] Batch [0679]/[1203]	Speed: 108.128874 samples/sec	 accuracy=70.613971	 loss=1.140669	 lr=0.010000
Epoch[029] Batch [0699]/[1203]	Speed: 142.415152 samples/sec	 accuracy=70.600714	 loss=1.141410	 lr=0.010000
Epoch[029] Batch [0719]/[1203]	Speed: 125.989727 samples/sec	 accuracy=70.548611	 loss=1.143610	 lr=0.010000
Epoch[029] Batch [0739]/[1203]	Speed: 121.938387 samples/sec	 accuracy=70.531081	 loss=1.144295	 lr=0.010000
Epoch[029] Batch [0759]/[1203]	Speed: 119.248237 samples/sec	 accuracy=70.505921	 loss=1.145065	 lr=0.010000
Epoch[029] Batch [0779]/[1203]	Speed: 121.388929 samples/sec	 accuracy=70.458333	 loss=1.146567	 lr=0.010000
Epoch[029] Batch [0799]/[1203]	Speed: 139.915220 samples/sec	 accuracy=70.400625	 loss=1.148321	 lr=0.010000
Epoch[029] Batch [0819]/[1203]	Speed: 126.956652 samples/sec	 accuracy=70.378659	 loss=1.149104	 lr=0.010000
Epoch[029] Batch [0839]/[1203]	Speed: 122.681758 samples/sec	 accuracy=70.368452	 loss=1.149487	 lr=0.010000
Epoch[029] Batch [0859]/[1203]	Speed: 117.167323 samples/sec	 accuracy=70.356395	 loss=1.150384	 lr=0.010000
Epoch[029] Batch [0879]/[1203]	Speed: 145.220441 samples/sec	 accuracy=70.341477	 loss=1.151064	 lr=0.010000
Epoch[029] Batch [0899]/[1203]	Speed: 121.364356 samples/sec	 accuracy=70.307778	 loss=1.152037	 lr=0.010000
Epoch[029] Batch [0919]/[1203]	Speed: 104.477717 samples/sec	 accuracy=70.251087	 loss=1.153860	 lr=0.010000
Epoch[029] Batch [0939]/[1203]	Speed: 140.680794 samples/sec	 accuracy=70.196809	 loss=1.155298	 lr=0.010000
Epoch[029] Batch [0959]/[1203]	Speed: 121.624099 samples/sec	 accuracy=70.179688	 loss=1.155740	 lr=0.010000
Epoch[029] Batch [0979]/[1203]	Speed: 141.352556 samples/sec	 accuracy=70.147959	 loss=1.157328	 lr=0.010000
Epoch[029] Batch [0999]/[1203]	Speed: 122.639724 samples/sec	 accuracy=70.125000	 loss=1.158263	 lr=0.010000
Epoch[029] Batch [1019]/[1203]	Speed: 121.772988 samples/sec	 accuracy=70.132353	 loss=1.157749	 lr=0.010000
Epoch[029] Batch [1039]/[1203]	Speed: 125.409887 samples/sec	 accuracy=70.131250	 loss=1.157780	 lr=0.010000
Epoch[029] Batch [1059]/[1203]	Speed: 116.463424 samples/sec	 accuracy=70.115094	 loss=1.158630	 lr=0.010000
Epoch[029] Batch [1079]/[1203]	Speed: 150.793981 samples/sec	 accuracy=70.099537	 loss=1.159092	 lr=0.010000
Epoch[029] Batch [1099]/[1203]	Speed: 104.325221 samples/sec	 accuracy=70.082273	 loss=1.159647	 lr=0.010000
Epoch[029] Batch [1119]/[1203]	Speed: 133.591434 samples/sec	 accuracy=70.052679	 loss=1.160756	 lr=0.010000
Epoch[029] Batch [1139]/[1203]	Speed: 135.953763 samples/sec	 accuracy=70.041667	 loss=1.160897	 lr=0.010000
Epoch[029] Batch [1159]/[1203]	Speed: 114.336872 samples/sec	 accuracy=70.024569	 loss=1.161514	 lr=0.010000
Epoch[029] Batch [1179]/[1203]	Speed: 140.776889 samples/sec	 accuracy=70.011441	 loss=1.161956	 lr=0.010000
Epoch[029] Batch [1199]/[1203]	Speed: 224.028304 samples/sec	 accuracy=69.983750	 loss=1.163034	 lr=0.010000
Batch [0019]/[0097]: acc-top1=60.275000 acc-top5=82.125000
Batch [0039]/[0097]: acc-top1=58.850000 acc-top5=81.150000
Batch [0059]/[0097]: acc-top1=61.133333 acc-top5=82.550000
Batch [0079]/[0097]: acc-top1=60.962500 acc-top5=82.462500
[Epoch 029] training: accuracy=69.986284	 loss=1.162887
[Epoch 029] speed: 124 samples/sec	time cost: 2130.843069
[Epoch 029] validation: acc-top1=60.469072 acc-top5=82.438144 loss=1.747131
Epoch[030] Batch [0019]/[1203]	Speed: 69.280120 samples/sec	 accuracy=71.300000	 loss=1.080500	 lr=0.010000
Epoch[030] Batch [0039]/[1203]	Speed: 97.378173 samples/sec	 accuracy=71.650000	 loss=1.073926	 lr=0.010000
Epoch[030] Batch [0059]/[1203]	Speed: 244.124070 samples/sec	 accuracy=71.758333	 loss=1.077081	 lr=0.010000
Epoch[030] Batch [0079]/[1203]	Speed: 95.602921 samples/sec	 accuracy=71.681250	 loss=1.081081	 lr=0.010000
Epoch[030] Batch [0099]/[1203]	Speed: 122.333827 samples/sec	 accuracy=71.310000	 loss=1.096785	 lr=0.010000
Epoch[030] Batch [0119]/[1203]	Speed: 153.720259 samples/sec	 accuracy=71.166667	 loss=1.103675	 lr=0.010000
Epoch[030] Batch [0139]/[1203]	Speed: 99.581459 samples/sec	 accuracy=71.007143	 loss=1.110240	 lr=0.010000
Epoch[030] Batch [0159]/[1203]	Speed: 221.550991 samples/sec	 accuracy=71.181250	 loss=1.110989	 lr=0.010000
Epoch[030] Batch [0179]/[1203]	Speed: 90.970271 samples/sec	 accuracy=71.138889	 loss=1.112382	 lr=0.010000
Epoch[030] Batch [0199]/[1203]	Speed: 131.175102 samples/sec	 accuracy=71.002500	 loss=1.116443	 lr=0.010000
Epoch[030] Batch [0219]/[1203]	Speed: 143.787535 samples/sec	 accuracy=70.993182	 loss=1.116053	 lr=0.010000
Epoch[030] Batch [0239]/[1203]	Speed: 98.429227 samples/sec	 accuracy=70.933333	 loss=1.117978	 lr=0.010000
Epoch[030] Batch [0259]/[1203]	Speed: 156.802267 samples/sec	 accuracy=70.965385	 loss=1.118049	 lr=0.010000
Epoch[030] Batch [0279]/[1203]	Speed: 120.846408 samples/sec	 accuracy=70.864286	 loss=1.122793	 lr=0.010000
Epoch[030] Batch [0299]/[1203]	Speed: 101.600407 samples/sec	 accuracy=70.815000	 loss=1.124889	 lr=0.010000
Epoch[030] Batch [0319]/[1203]	Speed: 194.408065 samples/sec	 accuracy=70.815625	 loss=1.124734	 lr=0.010000
Epoch[030] Batch [0339]/[1203]	Speed: 98.831060 samples/sec	 accuracy=70.845588	 loss=1.125630	 lr=0.010000
Epoch[030] Batch [0359]/[1203]	Speed: 171.145903 samples/sec	 accuracy=70.829167	 loss=1.126605	 lr=0.010000
Epoch[030] Batch [0379]/[1203]	Speed: 117.779178 samples/sec	 accuracy=70.826316	 loss=1.126374	 lr=0.010000
Epoch[030] Batch [0399]/[1203]	Speed: 99.407916 samples/sec	 accuracy=70.813750	 loss=1.126167	 lr=0.010000
Epoch[030] Batch [0419]/[1203]	Speed: 177.865552 samples/sec	 accuracy=70.790476	 loss=1.127645	 lr=0.010000
Epoch[030] Batch [0439]/[1203]	Speed: 103.308898 samples/sec	 accuracy=70.730682	 loss=1.130185	 lr=0.010000
Epoch[030] Batch [0459]/[1203]	Speed: 116.737779 samples/sec	 accuracy=70.644565	 loss=1.132886	 lr=0.010000
Epoch[030] Batch [0479]/[1203]	Speed: 149.474274 samples/sec	 accuracy=70.630208	 loss=1.132776	 lr=0.010000
Epoch[030] Batch [0499]/[1203]	Speed: 101.696615 samples/sec	 accuracy=70.577000	 loss=1.134407	 lr=0.010000
Epoch[030] Batch [0519]/[1203]	Speed: 178.450960 samples/sec	 accuracy=70.535577	 loss=1.136187	 lr=0.010000
Epoch[030] Batch [0539]/[1203]	Speed: 103.527773 samples/sec	 accuracy=70.527778	 loss=1.137736	 lr=0.010000
Epoch[030] Batch [0559]/[1203]	Speed: 110.688103 samples/sec	 accuracy=70.508929	 loss=1.138138	 lr=0.010000
Epoch[030] Batch [0579]/[1203]	Speed: 186.763169 samples/sec	 accuracy=70.506897	 loss=1.138041	 lr=0.010000
Epoch[030] Batch [0599]/[1203]	Speed: 100.151860 samples/sec	 accuracy=70.486667	 loss=1.138053	 lr=0.010000
Epoch[030] Batch [0619]/[1203]	Speed: 183.629177 samples/sec	 accuracy=70.471774	 loss=1.138305	 lr=0.010000
Epoch[030] Batch [0639]/[1203]	Speed: 100.375824 samples/sec	 accuracy=70.428125	 loss=1.140109	 lr=0.010000
Epoch[030] Batch [0659]/[1203]	Speed: 111.852014 samples/sec	 accuracy=70.468939	 loss=1.139165	 lr=0.010000
Epoch[030] Batch [0679]/[1203]	Speed: 168.582363 samples/sec	 accuracy=70.438971	 loss=1.139389	 lr=0.010000
Epoch[030] Batch [0699]/[1203]	Speed: 107.688037 samples/sec	 accuracy=70.421429	 loss=1.140331	 lr=0.010000
Epoch[030] Batch [0719]/[1203]	Speed: 112.194062 samples/sec	 accuracy=70.407639	 loss=1.140401	 lr=0.010000
Epoch[030] Batch [0739]/[1203]	Speed: 151.403634 samples/sec	 accuracy=70.449324	 loss=1.139804	 lr=0.010000
Epoch[030] Batch [0759]/[1203]	Speed: 99.819914 samples/sec	 accuracy=70.448026	 loss=1.139857	 lr=0.010000
Epoch[030] Batch [0779]/[1203]	Speed: 152.171011 samples/sec	 accuracy=70.466667	 loss=1.139682	 lr=0.010000
Epoch[030] Batch [0799]/[1203]	Speed: 124.651966 samples/sec	 accuracy=70.453750	 loss=1.140145	 lr=0.010000
Epoch[030] Batch [0819]/[1203]	Speed: 109.566905 samples/sec	 accuracy=70.440854	 loss=1.140310	 lr=0.010000
Epoch[030] Batch [0839]/[1203]	Speed: 167.055319 samples/sec	 accuracy=70.433929	 loss=1.140911	 lr=0.010000
Epoch[030] Batch [0859]/[1203]	Speed: 98.922651 samples/sec	 accuracy=70.442442	 loss=1.140715	 lr=0.010000
Epoch[030] Batch [0879]/[1203]	Speed: 131.166230 samples/sec	 accuracy=70.450000	 loss=1.140696	 lr=0.010000
Epoch[030] Batch [0899]/[1203]	Speed: 139.031597 samples/sec	 accuracy=70.434444	 loss=1.140851	 lr=0.010000
Epoch[030] Batch [0919]/[1203]	Speed: 116.918743 samples/sec	 accuracy=70.408696	 loss=1.141617	 lr=0.010000
Epoch[030] Batch [0939]/[1203]	Speed: 139.395520 samples/sec	 accuracy=70.401064	 loss=1.142197	 lr=0.010000
Epoch[030] Batch [0959]/[1203]	Speed: 107.063042 samples/sec	 accuracy=70.370312	 loss=1.143078	 lr=0.010000
Epoch[030] Batch [0979]/[1203]	Speed: 142.593717 samples/sec	 accuracy=70.348469	 loss=1.144069	 lr=0.010000
Epoch[030] Batch [0999]/[1203]	Speed: 123.374063 samples/sec	 accuracy=70.364500	 loss=1.143413	 lr=0.010000
Epoch[030] Batch [1019]/[1203]	Speed: 109.870771 samples/sec	 accuracy=70.382843	 loss=1.142902	 lr=0.010000
Epoch[030] Batch [1039]/[1203]	Speed: 150.008446 samples/sec	 accuracy=70.384135	 loss=1.143070	 lr=0.010000
Epoch[030] Batch [1059]/[1203]	Speed: 110.723217 samples/sec	 accuracy=70.368396	 loss=1.143550	 lr=0.010000
Epoch[030] Batch [1079]/[1203]	Speed: 132.057820 samples/sec	 accuracy=70.354167	 loss=1.143524	 lr=0.010000
Epoch[030] Batch [1099]/[1203]	Speed: 131.907464 samples/sec	 accuracy=70.347727	 loss=1.143827	 lr=0.010000
Epoch[030] Batch [1119]/[1203]	Speed: 108.305775 samples/sec	 accuracy=70.366518	 loss=1.143606	 lr=0.010000
Epoch[030] Batch [1139]/[1203]	Speed: 139.664748 samples/sec	 accuracy=70.366667	 loss=1.143761	 lr=0.010000
Epoch[030] Batch [1159]/[1203]	Speed: 116.205662 samples/sec	 accuracy=70.337500	 loss=1.144733	 lr=0.010000
Epoch[030] Batch [1179]/[1203]	Speed: 136.006345 samples/sec	 accuracy=70.336864	 loss=1.144843	 lr=0.010000
Epoch[030] Batch [1199]/[1203]	Speed: 226.493774 samples/sec	 accuracy=70.316250	 loss=1.146109	 lr=0.010000
Batch [0019]/[0097]: acc-top1=61.025000 acc-top5=82.575000
Batch [0039]/[0097]: acc-top1=59.250000 acc-top5=81.612500
Batch [0059]/[0097]: acc-top1=61.825000 acc-top5=83.300000
Batch [0079]/[0097]: acc-top1=61.993750 acc-top5=82.950000
[Epoch 030] training: accuracy=70.316708	 loss=1.146215
[Epoch 030] speed: 123 samples/sec	time cost: 2136.988205
[Epoch 030] validation: acc-top1=61.036082 acc-top5=82.701031 loss=1.732920
Epoch[031] Batch [0019]/[1203]	Speed: 68.135612 samples/sec	 accuracy=70.800000	 loss=1.128961	 lr=0.010000
Epoch[031] Batch [0039]/[1203]	Speed: 97.465819 samples/sec	 accuracy=71.587500	 loss=1.100678	 lr=0.010000
Epoch[031] Batch [0059]/[1203]	Speed: 237.655589 samples/sec	 accuracy=71.250000	 loss=1.112794	 lr=0.010000
Epoch[031] Batch [0079]/[1203]	Speed: 95.256346 samples/sec	 accuracy=71.581250	 loss=1.099824	 lr=0.010000
Epoch[031] Batch [0099]/[1203]	Speed: 108.379046 samples/sec	 accuracy=71.625000	 loss=1.097394	 lr=0.010000
Epoch[031] Batch [0119]/[1203]	Speed: 169.547619 samples/sec	 accuracy=71.408333	 loss=1.104453	 lr=0.010000
Epoch[031] Batch [0139]/[1203]	Speed: 101.457404 samples/sec	 accuracy=71.528571	 loss=1.103204	 lr=0.010000
Epoch[031] Batch [0159]/[1203]	Speed: 205.760352 samples/sec	 accuracy=71.431250	 loss=1.104061	 lr=0.010000
Epoch[031] Batch [0179]/[1203]	Speed: 98.833185 samples/sec	 accuracy=71.483333	 loss=1.100681	 lr=0.010000
Epoch[031] Batch [0199]/[1203]	Speed: 113.430221 samples/sec	 accuracy=71.502500	 loss=1.100972	 lr=0.010000
Epoch[031] Batch [0219]/[1203]	Speed: 172.416867 samples/sec	 accuracy=71.404545	 loss=1.104211	 lr=0.010000
Epoch[031] Batch [0239]/[1203]	Speed: 100.233467 samples/sec	 accuracy=71.358333	 loss=1.107244	 lr=0.010000
Epoch[031] Batch [0259]/[1203]	Speed: 131.286621 samples/sec	 accuracy=71.209615	 loss=1.110776	 lr=0.010000
Epoch[031] Batch [0279]/[1203]	Speed: 131.695460 samples/sec	 accuracy=71.107143	 loss=1.113908	 lr=0.010000
Epoch[031] Batch [0299]/[1203]	Speed: 111.154611 samples/sec	 accuracy=71.093333	 loss=1.115449	 lr=0.010000
Epoch[031] Batch [0319]/[1203]	Speed: 145.487321 samples/sec	 accuracy=71.137500	 loss=1.113698	 lr=0.010000
Epoch[031] Batch [0339]/[1203]	Speed: 109.609015 samples/sec	 accuracy=71.133824	 loss=1.113263	 lr=0.010000
Epoch[031] Batch [0359]/[1203]	Speed: 145.683462 samples/sec	 accuracy=71.097222	 loss=1.115082	 lr=0.010000
Epoch[031] Batch [0379]/[1203]	Speed: 117.540731 samples/sec	 accuracy=71.102632	 loss=1.115678	 lr=0.010000
Epoch[031] Batch [0399]/[1203]	Speed: 108.623923 samples/sec	 accuracy=71.050000	 loss=1.116934	 lr=0.010000
Epoch[031] Batch [0419]/[1203]	Speed: 148.991566 samples/sec	 accuracy=70.978571	 loss=1.118709	 lr=0.010000
Epoch[031] Batch [0439]/[1203]	Speed: 114.115044 samples/sec	 accuracy=70.954545	 loss=1.119989	 lr=0.010000
Epoch[031] Batch [0459]/[1203]	Speed: 139.598401 samples/sec	 accuracy=70.923913	 loss=1.120811	 lr=0.010000
Epoch[031] Batch [0479]/[1203]	Speed: 125.007988 samples/sec	 accuracy=70.986458	 loss=1.119016	 lr=0.010000
Epoch[031] Batch [0499]/[1203]	Speed: 107.798824 samples/sec	 accuracy=70.997000	 loss=1.120224	 lr=0.010000
Epoch[031] Batch [0519]/[1203]	Speed: 144.238401 samples/sec	 accuracy=71.001923	 loss=1.121986	 lr=0.010000
Epoch[031] Batch [0539]/[1203]	Speed: 121.273664 samples/sec	 accuracy=71.041667	 loss=1.120595	 lr=0.010000
Epoch[031] Batch [0559]/[1203]	Speed: 113.315590 samples/sec	 accuracy=71.050893	 loss=1.121448	 lr=0.010000
Epoch[031] Batch [0579]/[1203]	Speed: 160.822366 samples/sec	 accuracy=71.037069	 loss=1.122173	 lr=0.010000
Epoch[031] Batch [0599]/[1203]	Speed: 99.964137 samples/sec	 accuracy=71.002500	 loss=1.122941	 lr=0.010000
Epoch[031] Batch [0619]/[1203]	Speed: 158.552798 samples/sec	 accuracy=71.007258	 loss=1.123943	 lr=0.010000
Epoch[031] Batch [0639]/[1203]	Speed: 108.296719 samples/sec	 accuracy=70.988281	 loss=1.124663	 lr=0.010000
Epoch[031] Batch [0659]/[1203]	Speed: 111.048575 samples/sec	 accuracy=70.972727	 loss=1.124673	 lr=0.010000
Epoch[031] Batch [0679]/[1203]	Speed: 175.779572 samples/sec	 accuracy=70.965441	 loss=1.124333	 lr=0.010000
Epoch[031] Batch [0699]/[1203]	Speed: 99.209726 samples/sec	 accuracy=70.965000	 loss=1.124121	 lr=0.010000
Epoch[031] Batch [0719]/[1203]	Speed: 125.209712 samples/sec	 accuracy=70.959028	 loss=1.124387	 lr=0.010000
Epoch[031] Batch [0739]/[1203]	Speed: 147.199330 samples/sec	 accuracy=70.952027	 loss=1.124840	 lr=0.010000
Epoch[031] Batch [0759]/[1203]	Speed: 111.313948 samples/sec	 accuracy=70.928947	 loss=1.125595	 lr=0.010000
Epoch[031] Batch [0779]/[1203]	Speed: 164.592525 samples/sec	 accuracy=70.932692	 loss=1.125764	 lr=0.010000
Epoch[031] Batch [0799]/[1203]	Speed: 101.145820 samples/sec	 accuracy=70.900000	 loss=1.126422	 lr=0.010000
Epoch[031] Batch [0819]/[1203]	Speed: 115.473046 samples/sec	 accuracy=70.885366	 loss=1.126981	 lr=0.010000
Epoch[031] Batch [0839]/[1203]	Speed: 159.293009 samples/sec	 accuracy=70.882738	 loss=1.127053	 lr=0.010000
Epoch[031] Batch [0859]/[1203]	Speed: 106.516123 samples/sec	 accuracy=70.887791	 loss=1.127733	 lr=0.010000
Epoch[031] Batch [0879]/[1203]	Speed: 131.813202 samples/sec	 accuracy=70.867045	 loss=1.128123	 lr=0.010000
Epoch[031] Batch [0899]/[1203]	Speed: 123.072501 samples/sec	 accuracy=70.865556	 loss=1.127968	 lr=0.010000
Epoch[031] Batch [0919]/[1203]	Speed: 128.900985 samples/sec	 accuracy=70.835870	 loss=1.129107	 lr=0.010000
Epoch[031] Batch [0939]/[1203]	Speed: 129.455593 samples/sec	 accuracy=70.828191	 loss=1.128814	 lr=0.010000
Epoch[031] Batch [0959]/[1203]	Speed: 102.803366 samples/sec	 accuracy=70.802604	 loss=1.130060	 lr=0.010000
Epoch[031] Batch [0979]/[1203]	Speed: 135.976011 samples/sec	 accuracy=70.789796	 loss=1.130799	 lr=0.010000
Epoch[031] Batch [0999]/[1203]	Speed: 135.938779 samples/sec	 accuracy=70.753000	 loss=1.131752	 lr=0.010000
Epoch[031] Batch [1019]/[1203]	Speed: 135.591356 samples/sec	 accuracy=70.722059	 loss=1.132424	 lr=0.010000
Epoch[031] Batch [1039]/[1203]	Speed: 119.598449 samples/sec	 accuracy=70.691827	 loss=1.133823	 lr=0.010000
Epoch[031] Batch [1059]/[1203]	Speed: 107.705087 samples/sec	 accuracy=70.694811	 loss=1.133667	 lr=0.010000
Epoch[031] Batch [1079]/[1203]	Speed: 138.677907 samples/sec	 accuracy=70.700000	 loss=1.134104	 lr=0.010000
Epoch[031] Batch [1099]/[1203]	Speed: 117.710431 samples/sec	 accuracy=70.684545	 loss=1.135063	 lr=0.010000
Epoch[031] Batch [1119]/[1203]	Speed: 137.259358 samples/sec	 accuracy=70.653571	 loss=1.136376	 lr=0.010000
Epoch[031] Batch [1139]/[1203]	Speed: 117.554398 samples/sec	 accuracy=70.642105	 loss=1.137027	 lr=0.010000
Epoch[031] Batch [1159]/[1203]	Speed: 121.972104 samples/sec	 accuracy=70.599569	 loss=1.137836	 lr=0.010000
Epoch[031] Batch [1179]/[1203]	Speed: 136.490961 samples/sec	 accuracy=70.596186	 loss=1.138326	 lr=0.010000
Epoch[031] Batch [1199]/[1203]	Speed: 230.783774 samples/sec	 accuracy=70.561667	 loss=1.139354	 lr=0.010000
Batch [0019]/[0097]: acc-top1=59.625000 acc-top5=81.475000
Batch [0039]/[0097]: acc-top1=58.800000 acc-top5=81.312500
Batch [0059]/[0097]: acc-top1=61.091667 acc-top5=82.875000
Batch [0079]/[0097]: acc-top1=60.856250 acc-top5=82.718750
[Epoch 031] training: accuracy=70.564838	 loss=1.139310
[Epoch 031] speed: 123 samples/sec	time cost: 2141.864850
[Epoch 031] validation: acc-top1=60.278351 acc-top5=82.489691 loss=1.743624
Epoch[032] Batch [0019]/[1203]	Speed: 69.840405 samples/sec	 accuracy=70.625000	 loss=1.110397	 lr=0.010000
Epoch[032] Batch [0039]/[1203]	Speed: 100.465375 samples/sec	 accuracy=70.850000	 loss=1.105018	 lr=0.010000
Epoch[032] Batch [0059]/[1203]	Speed: 210.681091 samples/sec	 accuracy=70.950000	 loss=1.098073	 lr=0.010000
Epoch[032] Batch [0079]/[1203]	Speed: 97.317619 samples/sec	 accuracy=71.412500	 loss=1.084419	 lr=0.010000
Epoch[032] Batch [0099]/[1203]	Speed: 115.362275 samples/sec	 accuracy=71.515000	 loss=1.087519	 lr=0.010000
Epoch[032] Batch [0119]/[1203]	Speed: 168.938057 samples/sec	 accuracy=71.495833	 loss=1.084720	 lr=0.010000
Epoch[032] Batch [0139]/[1203]	Speed: 102.602871 samples/sec	 accuracy=71.553571	 loss=1.083179	 lr=0.010000
Epoch[032] Batch [0159]/[1203]	Speed: 185.132853 samples/sec	 accuracy=71.753125	 loss=1.076960	 lr=0.010000
Epoch[032] Batch [0179]/[1203]	Speed: 94.946104 samples/sec	 accuracy=71.761111	 loss=1.080966	 lr=0.010000
Epoch[032] Batch [0199]/[1203]	Speed: 127.311347 samples/sec	 accuracy=71.710000	 loss=1.083694	 lr=0.010000
Epoch[032] Batch [0219]/[1203]	Speed: 144.912623 samples/sec	 accuracy=71.650000	 loss=1.082402	 lr=0.010000
Epoch[032] Batch [0239]/[1203]	Speed: 104.394181 samples/sec	 accuracy=71.702083	 loss=1.081039	 lr=0.010000
Epoch[032] Batch [0259]/[1203]	Speed: 131.254321 samples/sec	 accuracy=71.648077	 loss=1.083539	 lr=0.010000
Epoch[032] Batch [0279]/[1203]	Speed: 128.503532 samples/sec	 accuracy=71.628571	 loss=1.085376	 lr=0.010000
Epoch[032] Batch [0299]/[1203]	Speed: 130.074514 samples/sec	 accuracy=71.598333	 loss=1.087509	 lr=0.010000
Epoch[032] Batch [0319]/[1203]	Speed: 129.709822 samples/sec	 accuracy=71.515625	 loss=1.090640	 lr=0.010000
Epoch[032] Batch [0339]/[1203]	Speed: 103.763582 samples/sec	 accuracy=71.407353	 loss=1.092984	 lr=0.010000
Epoch[032] Batch [0359]/[1203]	Speed: 152.147055 samples/sec	 accuracy=71.466667	 loss=1.091885	 lr=0.010000
Epoch[032] Batch [0379]/[1203]	Speed: 110.436191 samples/sec	 accuracy=71.457895	 loss=1.092678	 lr=0.010000
Epoch[032] Batch [0399]/[1203]	Speed: 122.584746 samples/sec	 accuracy=71.381250	 loss=1.095342	 lr=0.010000
Epoch[032] Batch [0419]/[1203]	Speed: 139.152383 samples/sec	 accuracy=71.355952	 loss=1.095852	 lr=0.010000
Epoch[032] Batch [0439]/[1203]	Speed: 109.160437 samples/sec	 accuracy=71.313636	 loss=1.098214	 lr=0.010000
Epoch[032] Batch [0459]/[1203]	Speed: 152.811458 samples/sec	 accuracy=71.291304	 loss=1.099004	 lr=0.010000
Epoch[032] Batch [0479]/[1203]	Speed: 113.910438 samples/sec	 accuracy=71.258333	 loss=1.100807	 lr=0.010000
Epoch[032] Batch [0499]/[1203]	Speed: 111.007622 samples/sec	 accuracy=71.271000	 loss=1.101371	 lr=0.010000
Epoch[032] Batch [0519]/[1203]	Speed: 149.375164 samples/sec	 accuracy=71.243269	 loss=1.102854	 lr=0.010000
Epoch[032] Batch [0539]/[1203]	Speed: 115.626573 samples/sec	 accuracy=71.183333	 loss=1.105372	 lr=0.010000
Epoch[032] Batch [0559]/[1203]	Speed: 127.404983 samples/sec	 accuracy=71.204464	 loss=1.104537	 lr=0.010000
Epoch[032] Batch [0579]/[1203]	Speed: 116.192787 samples/sec	 accuracy=71.125862	 loss=1.107484	 lr=0.010000
Epoch[032] Batch [0599]/[1203]	Speed: 120.599084 samples/sec	 accuracy=71.093333	 loss=1.108201	 lr=0.010000
Epoch[032] Batch [0619]/[1203]	Speed: 154.880443 samples/sec	 accuracy=71.096774	 loss=1.108807	 lr=0.010000
Epoch[032] Batch [0639]/[1203]	Speed: 106.366991 samples/sec	 accuracy=71.069531	 loss=1.109115	 lr=0.010000
Epoch[032] Batch [0659]/[1203]	Speed: 118.681558 samples/sec	 accuracy=71.039394	 loss=1.109253	 lr=0.010000
Epoch[032] Batch [0679]/[1203]	Speed: 137.702497 samples/sec	 accuracy=71.036765	 loss=1.109005	 lr=0.010000
Epoch[032] Batch [0699]/[1203]	Speed: 113.497640 samples/sec	 accuracy=71.003571	 loss=1.110755	 lr=0.010000
Epoch[032] Batch [0719]/[1203]	Speed: 163.985294 samples/sec	 accuracy=70.991667	 loss=1.110734	 lr=0.010000
Epoch[032] Batch [0739]/[1203]	Speed: 113.399467 samples/sec	 accuracy=70.963514	 loss=1.111934	 lr=0.010000
Epoch[032] Batch [0759]/[1203]	Speed: 108.955036 samples/sec	 accuracy=70.925658	 loss=1.112876	 lr=0.010000
Epoch[032] Batch [0779]/[1203]	Speed: 140.255641 samples/sec	 accuracy=70.888462	 loss=1.114229	 lr=0.010000
Epoch[032] Batch [0799]/[1203]	Speed: 114.063178 samples/sec	 accuracy=70.876875	 loss=1.114807	 lr=0.010000
Epoch[032] Batch [0819]/[1203]	Speed: 161.518058 samples/sec	 accuracy=70.910976	 loss=1.113611	 lr=0.010000
Epoch[032] Batch [0839]/[1203]	Speed: 109.149747 samples/sec	 accuracy=70.901786	 loss=1.114252	 lr=0.010000
Epoch[032] Batch [0859]/[1203]	Speed: 113.518109 samples/sec	 accuracy=70.894767	 loss=1.114009	 lr=0.010000
Epoch[032] Batch [0879]/[1203]	Speed: 154.418059 samples/sec	 accuracy=70.872727	 loss=1.115068	 lr=0.010000
Epoch[032] Batch [0899]/[1203]	Speed: 102.700919 samples/sec	 accuracy=70.841111	 loss=1.116091	 lr=0.010000
Epoch[032] Batch [0919]/[1203]	Speed: 124.729559 samples/sec	 accuracy=70.854348	 loss=1.115813	 lr=0.010000
Epoch[032] Batch [0939]/[1203]	Speed: 148.272273 samples/sec	 accuracy=70.854787	 loss=1.116209	 lr=0.010000
Epoch[032] Batch [0959]/[1203]	Speed: 116.651004 samples/sec	 accuracy=70.845833	 loss=1.116858	 lr=0.010000
Epoch[032] Batch [0979]/[1203]	Speed: 140.738413 samples/sec	 accuracy=70.835714	 loss=1.117242	 lr=0.010000
Epoch[032] Batch [0999]/[1203]	Speed: 112.950537 samples/sec	 accuracy=70.819500	 loss=1.117919	 lr=0.010000
Epoch[032] Batch [1019]/[1203]	Speed: 119.951711 samples/sec	 accuracy=70.801961	 loss=1.118764	 lr=0.010000
Epoch[032] Batch [1039]/[1203]	Speed: 144.890858 samples/sec	 accuracy=70.781731	 loss=1.119330	 lr=0.010000
Epoch[032] Batch [1059]/[1203]	Speed: 104.960166 samples/sec	 accuracy=70.739623	 loss=1.120814	 lr=0.010000
Epoch[032] Batch [1079]/[1203]	Speed: 146.621274 samples/sec	 accuracy=70.720833	 loss=1.121512	 lr=0.010000
Epoch[032] Batch [1099]/[1203]	Speed: 125.431773 samples/sec	 accuracy=70.709545	 loss=1.122331	 lr=0.010000
Epoch[032] Batch [1119]/[1203]	Speed: 104.551821 samples/sec	 accuracy=70.708929	 loss=1.122760	 lr=0.010000
Epoch[032] Batch [1139]/[1203]	Speed: 167.278915 samples/sec	 accuracy=70.719298	 loss=1.122721	 lr=0.010000
Epoch[032] Batch [1159]/[1203]	Speed: 111.633853 samples/sec	 accuracy=70.683621	 loss=1.123765	 lr=0.010000
Epoch[032] Batch [1179]/[1203]	Speed: 132.895353 samples/sec	 accuracy=70.689407	 loss=1.123851	 lr=0.010000
Epoch[032] Batch [1199]/[1203]	Speed: 245.286548 samples/sec	 accuracy=70.658750	 loss=1.124909	 lr=0.010000
Batch [0019]/[0097]: acc-top1=60.675000 acc-top5=82.025000
Batch [0039]/[0097]: acc-top1=59.237500 acc-top5=81.650000
Batch [0059]/[0097]: acc-top1=61.358333 acc-top5=83.016667
Batch [0079]/[0097]: acc-top1=61.256250 acc-top5=82.887500
[Epoch 032] training: accuracy=70.661264	 loss=1.124799
[Epoch 032] speed: 123 samples/sec	time cost: 2133.339053
[Epoch 032] validation: acc-top1=60.932990 acc-top5=82.860825 loss=1.752270
Epoch[033] Batch [0019]/[1204]	Speed: 69.046556 samples/sec	 accuracy=71.250000	 loss=1.091187	 lr=0.010000
Epoch[033] Batch [0039]/[1204]	Speed: 97.869751 samples/sec	 accuracy=71.937500	 loss=1.083502	 lr=0.010000
Epoch[033] Batch [0059]/[1204]	Speed: 251.726126 samples/sec	 accuracy=72.308333	 loss=1.054857	 lr=0.010000
Epoch[033] Batch [0079]/[1204]	Speed: 88.232367 samples/sec	 accuracy=72.206250	 loss=1.065685	 lr=0.010000
Epoch[033] Batch [0099]/[1204]	Speed: 126.038647 samples/sec	 accuracy=71.960000	 loss=1.071078	 lr=0.010000
Epoch[033] Batch [0119]/[1204]	Speed: 159.796483 samples/sec	 accuracy=71.891667	 loss=1.074831	 lr=0.010000
Epoch[033] Batch [0139]/[1204]	Speed: 107.502918 samples/sec	 accuracy=71.960714	 loss=1.074798	 lr=0.010000
Epoch[033] Batch [0159]/[1204]	Speed: 180.822515 samples/sec	 accuracy=71.921875	 loss=1.077547	 lr=0.010000
Epoch[033] Batch [0179]/[1204]	Speed: 96.209027 samples/sec	 accuracy=71.902778	 loss=1.078494	 lr=0.010000
Epoch[033] Batch [0199]/[1204]	Speed: 117.893969 samples/sec	 accuracy=71.797500	 loss=1.081731	 lr=0.010000
Epoch[033] Batch [0219]/[1204]	Speed: 163.241756 samples/sec	 accuracy=71.850000	 loss=1.079579	 lr=0.010000
Epoch[033] Batch [0239]/[1204]	Speed: 115.607533 samples/sec	 accuracy=71.610417	 loss=1.086098	 lr=0.010000
Epoch[033] Batch [0259]/[1204]	Speed: 128.681591 samples/sec	 accuracy=71.625000	 loss=1.086586	 lr=0.010000
Epoch[033] Batch [0279]/[1204]	Speed: 122.475931 samples/sec	 accuracy=71.530357	 loss=1.089387	 lr=0.010000
Epoch[033] Batch [0299]/[1204]	Speed: 110.135933 samples/sec	 accuracy=71.550000	 loss=1.087322	 lr=0.010000
Epoch[033] Batch [0319]/[1204]	Speed: 152.430207 samples/sec	 accuracy=71.503125	 loss=1.090289	 lr=0.010000
Epoch[033] Batch [0339]/[1204]	Speed: 104.456811 samples/sec	 accuracy=71.455882	 loss=1.092273	 lr=0.010000
Epoch[033] Batch [0359]/[1204]	Speed: 117.618750 samples/sec	 accuracy=71.429167	 loss=1.093756	 lr=0.010000
Epoch[033] Batch [0379]/[1204]	Speed: 158.580639 samples/sec	 accuracy=71.409211	 loss=1.095086	 lr=0.010000
Epoch[033] Batch [0399]/[1204]	Speed: 111.173350 samples/sec	 accuracy=71.386250	 loss=1.097158	 lr=0.010000
Epoch[033] Batch [0419]/[1204]	Speed: 137.891982 samples/sec	 accuracy=71.396429	 loss=1.097004	 lr=0.010000
Epoch[033] Batch [0439]/[1204]	Speed: 109.830235 samples/sec	 accuracy=71.395455	 loss=1.096656	 lr=0.010000
Epoch[033] Batch [0459]/[1204]	Speed: 135.231096 samples/sec	 accuracy=71.379348	 loss=1.097699	 lr=0.010000
Epoch[033] Batch [0479]/[1204]	Speed: 126.872955 samples/sec	 accuracy=71.343750	 loss=1.099775	 lr=0.010000
Epoch[033] Batch [0499]/[1204]	Speed: 109.750809 samples/sec	 accuracy=71.382000	 loss=1.099408	 lr=0.010000
Epoch[033] Batch [0519]/[1204]	Speed: 159.619860 samples/sec	 accuracy=71.392308	 loss=1.100051	 lr=0.010000
Epoch[033] Batch [0539]/[1204]	Speed: 111.083870 samples/sec	 accuracy=71.421296	 loss=1.099728	 lr=0.010000
Epoch[033] Batch [0559]/[1204]	Speed: 117.872113 samples/sec	 accuracy=71.400000	 loss=1.101768	 lr=0.010000
Epoch[033] Batch [0579]/[1204]	Speed: 123.781243 samples/sec	 accuracy=71.363793	 loss=1.102792	 lr=0.010000
Epoch[033] Batch [0599]/[1204]	Speed: 125.549114 samples/sec	 accuracy=71.364167	 loss=1.102807	 lr=0.010000
Epoch[033] Batch [0619]/[1204]	Speed: 150.202829 samples/sec	 accuracy=71.356452	 loss=1.103242	 lr=0.010000
Epoch[033] Batch [0639]/[1204]	Speed: 116.037610 samples/sec	 accuracy=71.339844	 loss=1.103868	 lr=0.010000
Epoch[033] Batch [0659]/[1204]	Speed: 112.255640 samples/sec	 accuracy=71.278030	 loss=1.105905	 lr=0.010000
Epoch[033] Batch [0679]/[1204]	Speed: 128.574906 samples/sec	 accuracy=71.269853	 loss=1.106400	 lr=0.010000
Epoch[033] Batch [0699]/[1204]	Speed: 129.955915 samples/sec	 accuracy=71.268571	 loss=1.106478	 lr=0.010000
Epoch[033] Batch [0719]/[1204]	Speed: 123.814789 samples/sec	 accuracy=71.242361	 loss=1.107392	 lr=0.010000
Epoch[033] Batch [0739]/[1204]	Speed: 121.265866 samples/sec	 accuracy=71.246622	 loss=1.106864	 lr=0.010000
Epoch[033] Batch [0759]/[1204]	Speed: 129.420282 samples/sec	 accuracy=71.246711	 loss=1.107565	 lr=0.010000
Epoch[033] Batch [0779]/[1204]	Speed: 131.826083 samples/sec	 accuracy=71.237179	 loss=1.107857	 lr=0.010000
Epoch[033] Batch [0799]/[1204]	Speed: 115.745271 samples/sec	 accuracy=71.210000	 loss=1.108699	 lr=0.010000
Epoch[033] Batch [0819]/[1204]	Speed: 119.603075 samples/sec	 accuracy=71.200000	 loss=1.108633	 lr=0.010000
Epoch[033] Batch [0839]/[1204]	Speed: 141.124226 samples/sec	 accuracy=71.195238	 loss=1.108687	 lr=0.010000
Epoch[033] Batch [0859]/[1204]	Speed: 109.658314 samples/sec	 accuracy=71.168605	 loss=1.109403	 lr=0.010000
Epoch[033] Batch [0879]/[1204]	Speed: 123.945428 samples/sec	 accuracy=71.161364	 loss=1.109604	 lr=0.010000
Epoch[033] Batch [0899]/[1204]	Speed: 136.850477 samples/sec	 accuracy=71.120556	 loss=1.111636	 lr=0.010000
Epoch[033] Batch [0919]/[1204]	Speed: 127.153010 samples/sec	 accuracy=71.114674	 loss=1.112454	 lr=0.010000
Epoch[033] Batch [0939]/[1204]	Speed: 131.658482 samples/sec	 accuracy=71.095745	 loss=1.113468	 lr=0.010000
Epoch[033] Batch [0959]/[1204]	Speed: 114.942689 samples/sec	 accuracy=71.077083	 loss=1.114213	 lr=0.010000
Epoch[033] Batch [0979]/[1204]	Speed: 116.661139 samples/sec	 accuracy=71.044898	 loss=1.114539	 lr=0.010000
Epoch[033] Batch [0999]/[1204]	Speed: 143.422090 samples/sec	 accuracy=71.038000	 loss=1.114917	 lr=0.010000
Epoch[033] Batch [1019]/[1204]	Speed: 106.958905 samples/sec	 accuracy=71.025490	 loss=1.115613	 lr=0.010000
Epoch[033] Batch [1039]/[1204]	Speed: 160.569669 samples/sec	 accuracy=71.015865	 loss=1.115460	 lr=0.010000
Epoch[033] Batch [1059]/[1204]	Speed: 111.733391 samples/sec	 accuracy=71.001415	 loss=1.116258	 lr=0.010000
Epoch[033] Batch [1079]/[1204]	Speed: 117.908609 samples/sec	 accuracy=70.981481	 loss=1.116802	 lr=0.010000
Epoch[033] Batch [1099]/[1204]	Speed: 162.985675 samples/sec	 accuracy=70.946364	 loss=1.118033	 lr=0.010000
Epoch[033] Batch [1119]/[1204]	Speed: 107.827539 samples/sec	 accuracy=70.938393	 loss=1.118255	 lr=0.010000
Epoch[033] Batch [1139]/[1204]	Speed: 127.899455 samples/sec	 accuracy=70.960526	 loss=1.117445	 lr=0.010000
Epoch[033] Batch [1159]/[1204]	Speed: 124.067449 samples/sec	 accuracy=70.924569	 loss=1.118515	 lr=0.010000
Epoch[033] Batch [1179]/[1204]	Speed: 110.761372 samples/sec	 accuracy=70.922034	 loss=1.118912	 lr=0.010000
Epoch[033] Batch [1199]/[1204]	Speed: 327.333646 samples/sec	 accuracy=70.912917	 loss=1.119339	 lr=0.010000
Batch [0019]/[0097]: acc-top1=58.875000 acc-top5=81.850000
Batch [0039]/[0097]: acc-top1=58.225000 acc-top5=81.212500
Batch [0059]/[0097]: acc-top1=60.675000 acc-top5=82.483333
Batch [0079]/[0097]: acc-top1=61.025000 acc-top5=82.718750
[Epoch 033] training: accuracy=70.900332	 loss=1.119653
[Epoch 033] speed: 124 samples/sec	time cost: 2137.769336
[Epoch 033] validation: acc-top1=60.345361 acc-top5=82.711340 loss=1.755820
Epoch[034] Batch [0019]/[1203]	Speed: 69.547134 samples/sec	 accuracy=71.175000	 loss=1.111729	 lr=0.010000
Epoch[034] Batch [0039]/[1203]	Speed: 98.140540 samples/sec	 accuracy=71.587500	 loss=1.085881	 lr=0.010000
Epoch[034] Batch [0059]/[1203]	Speed: 241.815009 samples/sec	 accuracy=72.250000	 loss=1.065089	 lr=0.010000
Epoch[034] Batch [0079]/[1203]	Speed: 95.633314 samples/sec	 accuracy=72.662500	 loss=1.054130	 lr=0.010000
Epoch[034] Batch [0099]/[1203]	Speed: 122.312491 samples/sec	 accuracy=72.540000	 loss=1.057930	 lr=0.010000
Epoch[034] Batch [0119]/[1203]	Speed: 156.602531 samples/sec	 accuracy=72.445833	 loss=1.062925	 lr=0.010000
Epoch[034] Batch [0139]/[1203]	Speed: 101.169271 samples/sec	 accuracy=72.400000	 loss=1.061790	 lr=0.010000
Epoch[034] Batch [0159]/[1203]	Speed: 195.357578 samples/sec	 accuracy=72.409375	 loss=1.061116	 lr=0.010000
Epoch[034] Batch [0179]/[1203]	Speed: 99.562581 samples/sec	 accuracy=72.380556	 loss=1.062462	 lr=0.010000
Epoch[034] Batch [0199]/[1203]	Speed: 109.579992 samples/sec	 accuracy=72.377500	 loss=1.062932	 lr=0.010000
Epoch[034] Batch [0219]/[1203]	Speed: 175.663761 samples/sec	 accuracy=72.368182	 loss=1.062724	 lr=0.010000
Epoch[034] Batch [0239]/[1203]	Speed: 108.474712 samples/sec	 accuracy=72.308333	 loss=1.065109	 lr=0.010000
Epoch[034] Batch [0259]/[1203]	Speed: 148.057173 samples/sec	 accuracy=72.259615	 loss=1.066479	 lr=0.010000
Epoch[034] Batch [0279]/[1203]	Speed: 115.784028 samples/sec	 accuracy=72.217857	 loss=1.068285	 lr=0.010000
Epoch[034] Batch [0299]/[1203]	Speed: 107.400504 samples/sec	 accuracy=72.168333	 loss=1.069918	 lr=0.010000
Epoch[034] Batch [0319]/[1203]	Speed: 155.792949 samples/sec	 accuracy=72.154688	 loss=1.070561	 lr=0.010000
Epoch[034] Batch [0339]/[1203]	Speed: 112.799422 samples/sec	 accuracy=72.166176	 loss=1.070654	 lr=0.010000
Epoch[034] Batch [0359]/[1203]	Speed: 129.332730 samples/sec	 accuracy=72.140278	 loss=1.072690	 lr=0.010000
Epoch[034] Batch [0379]/[1203]	Speed: 115.119316 samples/sec	 accuracy=72.046053	 loss=1.076747	 lr=0.010000
Epoch[034] Batch [0399]/[1203]	Speed: 124.820628 samples/sec	 accuracy=71.968750	 loss=1.078529	 lr=0.010000
Epoch[034] Batch [0419]/[1203]	Speed: 153.626813 samples/sec	 accuracy=71.986905	 loss=1.078568	 lr=0.010000
Epoch[034] Batch [0439]/[1203]	Speed: 116.262227 samples/sec	 accuracy=71.962500	 loss=1.079675	 lr=0.010000
Epoch[034] Batch [0459]/[1203]	Speed: 110.963842 samples/sec	 accuracy=71.936957	 loss=1.080541	 lr=0.010000
Epoch[034] Batch [0479]/[1203]	Speed: 125.054125 samples/sec	 accuracy=71.920833	 loss=1.081223	 lr=0.010000
Epoch[034] Batch [0499]/[1203]	Speed: 131.788093 samples/sec	 accuracy=71.922000	 loss=1.081198	 lr=0.010000
Epoch[034] Batch [0519]/[1203]	Speed: 142.672270 samples/sec	 accuracy=71.908654	 loss=1.081777	 lr=0.010000
Epoch[034] Batch [0539]/[1203]	Speed: 121.459493 samples/sec	 accuracy=71.924074	 loss=1.081505	 lr=0.010000
Epoch[034] Batch [0559]/[1203]	Speed: 111.832567 samples/sec	 accuracy=71.893750	 loss=1.082878	 lr=0.010000
Epoch[034] Batch [0579]/[1203]	Speed: 131.231080 samples/sec	 accuracy=71.880172	 loss=1.083195	 lr=0.010000
Epoch[034] Batch [0599]/[1203]	Speed: 123.804398 samples/sec	 accuracy=71.832500	 loss=1.084442	 lr=0.010000
Epoch[034] Batch [0619]/[1203]	Speed: 136.518353 samples/sec	 accuracy=71.794355	 loss=1.085765	 lr=0.010000
Epoch[034] Batch [0639]/[1203]	Speed: 105.283914 samples/sec	 accuracy=71.722656	 loss=1.088574	 lr=0.010000
Epoch[034] Batch [0659]/[1203]	Speed: 137.108491 samples/sec	 accuracy=71.671970	 loss=1.089786	 lr=0.010000
Epoch[034] Batch [0679]/[1203]	Speed: 126.911418 samples/sec	 accuracy=71.653676	 loss=1.090366	 lr=0.010000
Epoch[034] Batch [0699]/[1203]	Speed: 112.324588 samples/sec	 accuracy=71.638571	 loss=1.091433	 lr=0.010000
Epoch[034] Batch [0719]/[1203]	Speed: 174.575253 samples/sec	 accuracy=71.636111	 loss=1.092030	 lr=0.010000
Epoch[034] Batch [0739]/[1203]	Speed: 104.732991 samples/sec	 accuracy=71.650676	 loss=1.092317	 lr=0.010000
Epoch[034] Batch [0759]/[1203]	Speed: 124.435829 samples/sec	 accuracy=71.635526	 loss=1.093558	 lr=0.010000
Epoch[034] Batch [0779]/[1203]	Speed: 133.289457 samples/sec	 accuracy=71.583974	 loss=1.095649	 lr=0.010000
Epoch[034] Batch [0799]/[1203]	Speed: 117.994302 samples/sec	 accuracy=71.561875	 loss=1.097192	 lr=0.010000
Epoch[034] Batch [0819]/[1203]	Speed: 134.447366 samples/sec	 accuracy=71.518293	 loss=1.098720	 lr=0.010000
Epoch[034] Batch [0839]/[1203]	Speed: 121.950048 samples/sec	 accuracy=71.482143	 loss=1.099850	 lr=0.010000
Epoch[034] Batch [0859]/[1203]	Speed: 128.864903 samples/sec	 accuracy=71.473256	 loss=1.100678	 lr=0.010000
Epoch[034] Batch [0879]/[1203]	Speed: 132.609266 samples/sec	 accuracy=71.441477	 loss=1.102081	 lr=0.010000
Epoch[034] Batch [0899]/[1203]	Speed: 113.808731 samples/sec	 accuracy=71.411111	 loss=1.103267	 lr=0.010000
Epoch[034] Batch [0919]/[1203]	Speed: 128.158478 samples/sec	 accuracy=71.386957	 loss=1.103881	 lr=0.010000
Epoch[034] Batch [0939]/[1203]	Speed: 126.871867 samples/sec	 accuracy=71.354787	 loss=1.104119	 lr=0.010000
Epoch[034] Batch [0959]/[1203]	Speed: 107.146669 samples/sec	 accuracy=71.325521	 loss=1.104554	 lr=0.010000
Epoch[034] Batch [0979]/[1203]	Speed: 145.741301 samples/sec	 accuracy=71.328571	 loss=1.104644	 lr=0.010000
Epoch[034] Batch [0999]/[1203]	Speed: 127.473346 samples/sec	 accuracy=71.301500	 loss=1.105406	 lr=0.010000
Epoch[034] Batch [1019]/[1203]	Speed: 117.545938 samples/sec	 accuracy=71.297549	 loss=1.105768	 lr=0.010000
Epoch[034] Batch [1039]/[1203]	Speed: 144.659235 samples/sec	 accuracy=71.283654	 loss=1.106024	 lr=0.010000
Epoch[034] Batch [1059]/[1203]	Speed: 105.846173 samples/sec	 accuracy=71.274057	 loss=1.106265	 lr=0.010000
Epoch[034] Batch [1079]/[1203]	Speed: 144.152774 samples/sec	 accuracy=71.257407	 loss=1.107072	 lr=0.010000
Epoch[034] Batch [1099]/[1203]	Speed: 115.076695 samples/sec	 accuracy=71.211818	 loss=1.108320	 lr=0.010000
Epoch[034] Batch [1119]/[1203]	Speed: 118.266912 samples/sec	 accuracy=71.190625	 loss=1.109456	 lr=0.010000
Epoch[034] Batch [1139]/[1203]	Speed: 159.726164 samples/sec	 accuracy=71.194298	 loss=1.109744	 lr=0.010000
Epoch[034] Batch [1159]/[1203]	Speed: 105.442395 samples/sec	 accuracy=71.182328	 loss=1.109870	 lr=0.010000
Epoch[034] Batch [1179]/[1203]	Speed: 141.224711 samples/sec	 accuracy=71.150847	 loss=1.110869	 lr=0.010000
Epoch[034] Batch [1199]/[1203]	Speed: 217.298907 samples/sec	 accuracy=71.134167	 loss=1.111796	 lr=0.010000
Batch [0019]/[0097]: acc-top1=60.550000 acc-top5=81.950000
Batch [0039]/[0097]: acc-top1=59.575000 acc-top5=82.137500
Batch [0059]/[0097]: acc-top1=61.333333 acc-top5=83.075000
Batch [0079]/[0097]: acc-top1=61.381250 acc-top5=82.856250
[Epoch 034] training: accuracy=71.135910	 loss=1.111874
[Epoch 034] speed: 124 samples/sec	time cost: 2129.460897
[Epoch 034] validation: acc-top1=60.396907 acc-top5=82.587629 loss=1.743429
Epoch[035] Batch [0019]/[1203]	Speed: 69.478000 samples/sec	 accuracy=71.900000	 loss=1.100015	 lr=0.010000
Epoch[035] Batch [0039]/[1203]	Speed: 102.744257 samples/sec	 accuracy=72.325000	 loss=1.078510	 lr=0.010000
Epoch[035] Batch [0059]/[1203]	Speed: 224.381063 samples/sec	 accuracy=72.666667	 loss=1.069674	 lr=0.010000
Epoch[035] Batch [0079]/[1203]	Speed: 91.616057 samples/sec	 accuracy=72.868750	 loss=1.058507	 lr=0.010000
Epoch[035] Batch [0099]/[1203]	Speed: 109.289772 samples/sec	 accuracy=72.900000	 loss=1.058898	 lr=0.010000
Epoch[035] Batch [0119]/[1203]	Speed: 168.463678 samples/sec	 accuracy=72.816667	 loss=1.054425	 lr=0.010000
Epoch[035] Batch [0139]/[1203]	Speed: 113.112639 samples/sec	 accuracy=72.628571	 loss=1.060336	 lr=0.010000
Epoch[035] Batch [0159]/[1203]	Speed: 175.453241 samples/sec	 accuracy=72.593750	 loss=1.059342	 lr=0.010000
Epoch[035] Batch [0179]/[1203]	Speed: 98.554340 samples/sec	 accuracy=72.497222	 loss=1.061270	 lr=0.010000
Epoch[035] Batch [0199]/[1203]	Speed: 121.497176 samples/sec	 accuracy=72.470000	 loss=1.061613	 lr=0.010000
Epoch[035] Batch [0219]/[1203]	Speed: 142.526061 samples/sec	 accuracy=72.525000	 loss=1.060875	 lr=0.010000
Epoch[035] Batch [0239]/[1203]	Speed: 110.855600 samples/sec	 accuracy=72.418750	 loss=1.063967	 lr=0.010000
Epoch[035] Batch [0259]/[1203]	Speed: 132.248776 samples/sec	 accuracy=72.388462	 loss=1.065873	 lr=0.010000
Epoch[035] Batch [0279]/[1203]	Speed: 129.083231 samples/sec	 accuracy=72.298214	 loss=1.068988	 lr=0.010000
Epoch[035] Batch [0299]/[1203]	Speed: 108.438625 samples/sec	 accuracy=72.348333	 loss=1.067080	 lr=0.010000
Epoch[035] Batch [0319]/[1203]	Speed: 152.932648 samples/sec	 accuracy=72.195312	 loss=1.070175	 lr=0.010000
Epoch[035] Batch [0339]/[1203]	Speed: 114.892699 samples/sec	 accuracy=72.122059	 loss=1.072753	 lr=0.010000
Epoch[035] Batch [0359]/[1203]	Speed: 122.061458 samples/sec	 accuracy=72.120833	 loss=1.073094	 lr=0.010000
Epoch[035] Batch [0379]/[1203]	Speed: 134.899440 samples/sec	 accuracy=72.059211	 loss=1.074624	 lr=0.010000
Epoch[035] Batch [0399]/[1203]	Speed: 116.972739 samples/sec	 accuracy=72.061250	 loss=1.075101	 lr=0.010000
Epoch[035] Batch [0419]/[1203]	Speed: 135.559082 samples/sec	 accuracy=72.046429	 loss=1.074983	 lr=0.010000
Epoch[035] Batch [0439]/[1203]	Speed: 109.059850 samples/sec	 accuracy=72.029545	 loss=1.075231	 lr=0.010000
Epoch[035] Batch [0459]/[1203]	Speed: 143.023446 samples/sec	 accuracy=72.028261	 loss=1.074362	 lr=0.010000
Epoch[035] Batch [0479]/[1203]	Speed: 122.455932 samples/sec	 accuracy=71.990625	 loss=1.074820	 lr=0.010000
Epoch[035] Batch [0499]/[1203]	Speed: 117.710862 samples/sec	 accuracy=71.950000	 loss=1.076458	 lr=0.010000
Epoch[035] Batch [0519]/[1203]	Speed: 140.791778 samples/sec	 accuracy=71.960577	 loss=1.075952	 lr=0.010000
Epoch[035] Batch [0539]/[1203]	Speed: 109.116815 samples/sec	 accuracy=71.912037	 loss=1.076879	 lr=0.010000
Epoch[035] Batch [0559]/[1203]	Speed: 146.422595 samples/sec	 accuracy=71.925000	 loss=1.076761	 lr=0.010000
Epoch[035] Batch [0579]/[1203]	Speed: 120.062533 samples/sec	 accuracy=71.889655	 loss=1.077765	 lr=0.010000
Epoch[035] Batch [0599]/[1203]	Speed: 109.307081 samples/sec	 accuracy=71.880000	 loss=1.077613	 lr=0.010000
Epoch[035] Batch [0619]/[1203]	Speed: 151.997788 samples/sec	 accuracy=71.858871	 loss=1.079062	 lr=0.010000
Epoch[035] Batch [0639]/[1203]	Speed: 108.422899 samples/sec	 accuracy=71.843750	 loss=1.080614	 lr=0.010000
Epoch[035] Batch [0659]/[1203]	Speed: 134.781355 samples/sec	 accuracy=71.819697	 loss=1.080974	 lr=0.010000
Epoch[035] Batch [0679]/[1203]	Speed: 120.366172 samples/sec	 accuracy=71.797059	 loss=1.081534	 lr=0.010000
Epoch[035] Batch [0699]/[1203]	Speed: 121.328294 samples/sec	 accuracy=71.758571	 loss=1.083841	 lr=0.010000
Epoch[035] Batch [0719]/[1203]	Speed: 136.191486 samples/sec	 accuracy=71.722917	 loss=1.084853	 lr=0.010000
Epoch[035] Batch [0739]/[1203]	Speed: 111.686551 samples/sec	 accuracy=71.698649	 loss=1.086108	 lr=0.010000
Epoch[035] Batch [0759]/[1203]	Speed: 132.981543 samples/sec	 accuracy=71.681579	 loss=1.087031	 lr=0.010000
Epoch[035] Batch [0779]/[1203]	Speed: 119.775140 samples/sec	 accuracy=71.658974	 loss=1.088138	 lr=0.010000
Epoch[035] Batch [0799]/[1203]	Speed: 124.864696 samples/sec	 accuracy=71.625625	 loss=1.089065	 lr=0.010000
Epoch[035] Batch [0819]/[1203]	Speed: 132.637500 samples/sec	 accuracy=71.626220	 loss=1.089574	 lr=0.010000
Epoch[035] Batch [0839]/[1203]	Speed: 113.451525 samples/sec	 accuracy=71.600000	 loss=1.090970	 lr=0.010000
Epoch[035] Batch [0859]/[1203]	Speed: 144.406321 samples/sec	 accuracy=71.605814	 loss=1.091132	 lr=0.010000
Epoch[035] Batch [0879]/[1203]	Speed: 111.882009 samples/sec	 accuracy=71.576705	 loss=1.092360	 lr=0.010000
Epoch[035] Batch [0899]/[1203]	Speed: 125.370861 samples/sec	 accuracy=71.547222	 loss=1.093333	 lr=0.010000
Epoch[035] Batch [0919]/[1203]	Speed: 150.166758 samples/sec	 accuracy=71.530435	 loss=1.094123	 lr=0.010000
Epoch[035] Batch [0939]/[1203]	Speed: 107.094731 samples/sec	 accuracy=71.534574	 loss=1.094015	 lr=0.010000
Epoch[035] Batch [0959]/[1203]	Speed: 125.363568 samples/sec	 accuracy=71.507292	 loss=1.094852	 lr=0.010000
Epoch[035] Batch [0979]/[1203]	Speed: 119.849835 samples/sec	 accuracy=71.503061	 loss=1.095197	 lr=0.010000
Epoch[035] Batch [0999]/[1203]	Speed: 115.960803 samples/sec	 accuracy=71.500000	 loss=1.096186	 lr=0.010000
Epoch[035] Batch [1019]/[1203]	Speed: 144.356291 samples/sec	 accuracy=71.502451	 loss=1.096955	 lr=0.010000
Epoch[035] Batch [1039]/[1203]	Speed: 123.957494 samples/sec	 accuracy=71.483173	 loss=1.097609	 lr=0.010000
Epoch[035] Batch [1059]/[1203]	Speed: 110.400251 samples/sec	 accuracy=71.474057	 loss=1.097557	 lr=0.010000
Epoch[035] Batch [1079]/[1203]	Speed: 124.209330 samples/sec	 accuracy=71.466667	 loss=1.097618	 lr=0.010000
Epoch[035] Batch [1099]/[1203]	Speed: 129.187387 samples/sec	 accuracy=71.425909	 loss=1.098983	 lr=0.010000
Epoch[035] Batch [1119]/[1203]	Speed: 147.588562 samples/sec	 accuracy=71.422321	 loss=1.099195	 lr=0.010000
Epoch[035] Batch [1139]/[1203]	Speed: 114.877770 samples/sec	 accuracy=71.400877	 loss=1.099973	 lr=0.010000
Epoch[035] Batch [1159]/[1203]	Speed: 126.626070 samples/sec	 accuracy=71.402586	 loss=1.099957	 lr=0.010000
Epoch[035] Batch [1179]/[1203]	Speed: 117.075170 samples/sec	 accuracy=71.364831	 loss=1.100822	 lr=0.010000
Epoch[035] Batch [1199]/[1203]	Speed: 291.355877 samples/sec	 accuracy=71.352500	 loss=1.101196	 lr=0.010000
Batch [0019]/[0097]: acc-top1=60.600000 acc-top5=82.650000
Batch [0039]/[0097]: acc-top1=58.800000 acc-top5=81.800000
Batch [0059]/[0097]: acc-top1=60.791667 acc-top5=82.666667
Batch [0079]/[0097]: acc-top1=60.881250 acc-top5=82.612500
[Epoch 035] training: accuracy=71.350374	 loss=1.101250
[Epoch 035] speed: 123 samples/sec	time cost: 2132.484007
[Epoch 035] validation: acc-top1=60.556701 acc-top5=82.680412 loss=1.730926
Epoch[036] Batch [0019]/[1203]	Speed: 69.692593 samples/sec	 accuracy=72.625000	 loss=1.057201	 lr=0.010000
Epoch[036] Batch [0039]/[1203]	Speed: 103.651824 samples/sec	 accuracy=72.512500	 loss=1.055405	 lr=0.010000
Epoch[036] Batch [0059]/[1203]	Speed: 221.304026 samples/sec	 accuracy=72.216667	 loss=1.059892	 lr=0.010000
Epoch[036] Batch [0079]/[1203]	Speed: 96.086255 samples/sec	 accuracy=72.043750	 loss=1.068709	 lr=0.010000
Epoch[036] Batch [0099]/[1203]	Speed: 106.964173 samples/sec	 accuracy=72.160000	 loss=1.068280	 lr=0.010000
Epoch[036] Batch [0119]/[1203]	Speed: 157.702440 samples/sec	 accuracy=72.275000	 loss=1.062194	 lr=0.010000
Epoch[036] Batch [0139]/[1203]	Speed: 109.301164 samples/sec	 accuracy=72.171429	 loss=1.064569	 lr=0.010000
Epoch[036] Batch [0159]/[1203]	Speed: 192.267296 samples/sec	 accuracy=72.206250	 loss=1.060792	 lr=0.010000
Epoch[036] Batch [0179]/[1203]	Speed: 102.613763 samples/sec	 accuracy=72.311111	 loss=1.058682	 lr=0.010000
Epoch[036] Batch [0199]/[1203]	Speed: 116.103357 samples/sec	 accuracy=72.252500	 loss=1.058450	 lr=0.010000
Epoch[036] Batch [0219]/[1203]	Speed: 156.539629 samples/sec	 accuracy=72.354545	 loss=1.054964	 lr=0.010000
Epoch[036] Batch [0239]/[1203]	Speed: 109.550928 samples/sec	 accuracy=72.529167	 loss=1.049522	 lr=0.010000
Epoch[036] Batch [0259]/[1203]	Speed: 146.266144 samples/sec	 accuracy=72.465385	 loss=1.050873	 lr=0.010000
Epoch[036] Batch [0279]/[1203]	Speed: 106.736971 samples/sec	 accuracy=72.448214	 loss=1.049909	 lr=0.010000
Epoch[036] Batch [0299]/[1203]	Speed: 124.840329 samples/sec	 accuracy=72.466667	 loss=1.049936	 lr=0.010000
Epoch[036] Batch [0319]/[1203]	Speed: 137.053465 samples/sec	 accuracy=72.432812	 loss=1.052302	 lr=0.010000
Epoch[036] Batch [0339]/[1203]	Speed: 105.486006 samples/sec	 accuracy=72.372059	 loss=1.054107	 lr=0.010000
Epoch[036] Batch [0359]/[1203]	Speed: 142.571661 samples/sec	 accuracy=72.329167	 loss=1.056627	 lr=0.010000
Epoch[036] Batch [0379]/[1203]	Speed: 127.776682 samples/sec	 accuracy=72.307895	 loss=1.056809	 lr=0.010000
Epoch[036] Batch [0399]/[1203]	Speed: 124.250010 samples/sec	 accuracy=72.311250	 loss=1.056767	 lr=0.010000
Epoch[036] Batch [0419]/[1203]	Speed: 129.710674 samples/sec	 accuracy=72.303571	 loss=1.057432	 lr=0.010000
Epoch[036] Batch [0439]/[1203]	Speed: 107.470711 samples/sec	 accuracy=72.321591	 loss=1.058377	 lr=0.010000
Epoch[036] Batch [0459]/[1203]	Speed: 139.872937 samples/sec	 accuracy=72.256522	 loss=1.060873	 lr=0.010000
Epoch[036] Batch [0479]/[1203]	Speed: 126.378226 samples/sec	 accuracy=72.258333	 loss=1.061896	 lr=0.010000
Epoch[036] Batch [0499]/[1203]	Speed: 107.033718 samples/sec	 accuracy=72.174000	 loss=1.064566	 lr=0.010000
Epoch[036] Batch [0519]/[1203]	Speed: 162.654517 samples/sec	 accuracy=72.139423	 loss=1.066101	 lr=0.010000
Epoch[036] Batch [0539]/[1203]	Speed: 116.344549 samples/sec	 accuracy=72.106481	 loss=1.068068	 lr=0.010000
Epoch[036] Batch [0559]/[1203]	Speed: 136.339378 samples/sec	 accuracy=72.091964	 loss=1.069096	 lr=0.010000
Epoch[036] Batch [0579]/[1203]	Speed: 122.484230 samples/sec	 accuracy=72.049138	 loss=1.069744	 lr=0.010000
Epoch[036] Batch [0599]/[1203]	Speed: 104.643869 samples/sec	 accuracy=72.005000	 loss=1.071439	 lr=0.010000
Epoch[036] Batch [0619]/[1203]	Speed: 165.705070 samples/sec	 accuracy=71.980645	 loss=1.071930	 lr=0.010000
Epoch[036] Batch [0639]/[1203]	Speed: 116.085007 samples/sec	 accuracy=71.947656	 loss=1.073386	 lr=0.010000
Epoch[036] Batch [0659]/[1203]	Speed: 120.434599 samples/sec	 accuracy=71.963636	 loss=1.073567	 lr=0.010000
Epoch[036] Batch [0679]/[1203]	Speed: 135.649477 samples/sec	 accuracy=71.953676	 loss=1.073768	 lr=0.010000
Epoch[036] Batch [0699]/[1203]	Speed: 105.706819 samples/sec	 accuracy=71.963571	 loss=1.073216	 lr=0.010000
Epoch[036] Batch [0719]/[1203]	Speed: 154.721114 samples/sec	 accuracy=71.979861	 loss=1.072685	 lr=0.010000
Epoch[036] Batch [0739]/[1203]	Speed: 120.434289 samples/sec	 accuracy=71.981757	 loss=1.072604	 lr=0.010000
Epoch[036] Batch [0759]/[1203]	Speed: 103.737852 samples/sec	 accuracy=71.977632	 loss=1.073157	 lr=0.010000
Epoch[036] Batch [0779]/[1203]	Speed: 162.632341 samples/sec	 accuracy=71.956410	 loss=1.073245	 lr=0.010000
Epoch[036] Batch [0799]/[1203]	Speed: 99.555701 samples/sec	 accuracy=71.927500	 loss=1.074036	 lr=0.010000
Epoch[036] Batch [0819]/[1203]	Speed: 129.526164 samples/sec	 accuracy=71.928659	 loss=1.073890	 lr=0.010000
Epoch[036] Batch [0839]/[1203]	Speed: 142.893853 samples/sec	 accuracy=71.918452	 loss=1.074169	 lr=0.010000
Epoch[036] Batch [0859]/[1203]	Speed: 102.692254 samples/sec	 accuracy=71.879070	 loss=1.075025	 lr=0.010000
Epoch[036] Batch [0879]/[1203]	Speed: 174.974523 samples/sec	 accuracy=71.860227	 loss=1.075959	 lr=0.010000
Epoch[036] Batch [0899]/[1203]	Speed: 103.898132 samples/sec	 accuracy=71.861111	 loss=1.076951	 lr=0.010000
Epoch[036] Batch [0919]/[1203]	Speed: 121.729605 samples/sec	 accuracy=71.842391	 loss=1.078285	 lr=0.010000
Epoch[036] Batch [0939]/[1203]	Speed: 149.137799 samples/sec	 accuracy=71.834043	 loss=1.078394	 lr=0.010000
Epoch[036] Batch [0959]/[1203]	Speed: 94.956441 samples/sec	 accuracy=71.815104	 loss=1.079271	 lr=0.010000
Epoch[036] Batch [0979]/[1203]	Speed: 169.813985 samples/sec	 accuracy=71.803061	 loss=1.079821	 lr=0.010000
Epoch[036] Batch [0999]/[1203]	Speed: 118.494570 samples/sec	 accuracy=71.779500	 loss=1.080760	 lr=0.010000
Epoch[036] Batch [1019]/[1203]	Speed: 107.985989 samples/sec	 accuracy=71.756863	 loss=1.081544	 lr=0.010000
Epoch[036] Batch [1039]/[1203]	Speed: 170.565967 samples/sec	 accuracy=71.745192	 loss=1.081687	 lr=0.010000
Epoch[036] Batch [1059]/[1203]	Speed: 96.600735 samples/sec	 accuracy=71.743868	 loss=1.081897	 lr=0.010000
Epoch[036] Batch [1079]/[1203]	Speed: 155.383667 samples/sec	 accuracy=71.736574	 loss=1.082563	 lr=0.010000
Epoch[036] Batch [1099]/[1203]	Speed: 119.121852 samples/sec	 accuracy=71.711818	 loss=1.083157	 lr=0.010000
Epoch[036] Batch [1119]/[1203]	Speed: 109.927889 samples/sec	 accuracy=71.704464	 loss=1.083641	 lr=0.010000
Epoch[036] Batch [1139]/[1203]	Speed: 160.081349 samples/sec	 accuracy=71.689035	 loss=1.084612	 lr=0.010000
Epoch[036] Batch [1159]/[1203]	Speed: 108.561572 samples/sec	 accuracy=71.670259	 loss=1.085457	 lr=0.010000
Epoch[036] Batch [1179]/[1203]	Speed: 134.669168 samples/sec	 accuracy=71.656780	 loss=1.085826	 lr=0.010000
Epoch[036] Batch [1199]/[1203]	Speed: 251.905717 samples/sec	 accuracy=71.641250	 loss=1.086369	 lr=0.010000
Batch [0019]/[0097]: acc-top1=60.075000 acc-top5=82.275000
Batch [0039]/[0097]: acc-top1=59.337500 acc-top5=81.912500
Batch [0059]/[0097]: acc-top1=61.491667 acc-top5=83.075000
Batch [0079]/[0097]: acc-top1=61.350000 acc-top5=82.918750
[Epoch 036] training: accuracy=71.644638	 loss=1.086121
[Epoch 036] speed: 124 samples/sec	time cost: 2131.083775
[Epoch 036] validation: acc-top1=60.654639 acc-top5=82.798969 loss=1.766241
Epoch[037] Batch [0019]/[1203]	Speed: 69.502198 samples/sec	 accuracy=71.375000	 loss=1.098539	 lr=0.010000
Epoch[037] Batch [0039]/[1203]	Speed: 97.255023 samples/sec	 accuracy=72.000000	 loss=1.071899	 lr=0.010000
Epoch[037] Batch [0059]/[1203]	Speed: 246.009887 samples/sec	 accuracy=72.483333	 loss=1.055749	 lr=0.010000
Epoch[037] Batch [0079]/[1203]	Speed: 95.901963 samples/sec	 accuracy=72.487500	 loss=1.051737	 lr=0.010000
Epoch[037] Batch [0099]/[1203]	Speed: 105.283096 samples/sec	 accuracy=72.610000	 loss=1.047118	 lr=0.010000
Epoch[037] Batch [0119]/[1203]	Speed: 202.975517 samples/sec	 accuracy=72.558333	 loss=1.046687	 lr=0.010000
Epoch[037] Batch [0139]/[1203]	Speed: 93.508818 samples/sec	 accuracy=72.678571	 loss=1.043499	 lr=0.010000
Epoch[037] Batch [0159]/[1203]	Speed: 228.113767 samples/sec	 accuracy=72.646875	 loss=1.047035	 lr=0.010000
Epoch[037] Batch [0179]/[1203]	Speed: 98.177635 samples/sec	 accuracy=72.508333	 loss=1.049593	 lr=0.010000
Epoch[037] Batch [0199]/[1203]	Speed: 106.594459 samples/sec	 accuracy=72.540000	 loss=1.047006	 lr=0.010000
Epoch[037] Batch [0219]/[1203]	Speed: 185.128483 samples/sec	 accuracy=72.606818	 loss=1.042783	 lr=0.010000
Epoch[037] Batch [0239]/[1203]	Speed: 95.682203 samples/sec	 accuracy=72.489583	 loss=1.046212	 lr=0.010000
Epoch[037] Batch [0259]/[1203]	Speed: 128.162115 samples/sec	 accuracy=72.442308	 loss=1.048255	 lr=0.010000
Epoch[037] Batch [0279]/[1203]	Speed: 146.342559 samples/sec	 accuracy=72.466071	 loss=1.049082	 lr=0.010000
Epoch[037] Batch [0299]/[1203]	Speed: 104.621606 samples/sec	 accuracy=72.453333	 loss=1.051003	 lr=0.010000
Epoch[037] Batch [0319]/[1203]	Speed: 166.796020 samples/sec	 accuracy=72.376562	 loss=1.051614	 lr=0.010000
Epoch[037] Batch [0339]/[1203]	Speed: 106.288993 samples/sec	 accuracy=72.363235	 loss=1.051186	 lr=0.010000
Epoch[037] Batch [0359]/[1203]	Speed: 129.161629 samples/sec	 accuracy=72.337500	 loss=1.052500	 lr=0.010000
Epoch[037] Batch [0379]/[1203]	Speed: 128.710447 samples/sec	 accuracy=72.352632	 loss=1.053000	 lr=0.010000
Epoch[037] Batch [0399]/[1203]	Speed: 104.947270 samples/sec	 accuracy=72.332500	 loss=1.055073	 lr=0.010000
Epoch[037] Batch [0419]/[1203]	Speed: 140.818633 samples/sec	 accuracy=72.315476	 loss=1.055450	 lr=0.010000
Epoch[037] Batch [0439]/[1203]	Speed: 126.556129 samples/sec	 accuracy=72.261364	 loss=1.056081	 lr=0.010000
Epoch[037] Batch [0459]/[1203]	Speed: 116.371595 samples/sec	 accuracy=72.247826	 loss=1.057321	 lr=0.010000
Epoch[037] Batch [0479]/[1203]	Speed: 137.369265 samples/sec	 accuracy=72.235417	 loss=1.057973	 lr=0.010000
Epoch[037] Batch [0499]/[1203]	Speed: 103.035667 samples/sec	 accuracy=72.257000	 loss=1.057401	 lr=0.010000
Epoch[037] Batch [0519]/[1203]	Speed: 153.364370 samples/sec	 accuracy=72.230769	 loss=1.059258	 lr=0.010000
Epoch[037] Batch [0539]/[1203]	Speed: 113.245995 samples/sec	 accuracy=72.246296	 loss=1.058850	 lr=0.010000
Epoch[037] Batch [0559]/[1203]	Speed: 126.690654 samples/sec	 accuracy=72.242857	 loss=1.059940	 lr=0.010000
Epoch[037] Batch [0579]/[1203]	Speed: 152.429318 samples/sec	 accuracy=72.237931	 loss=1.059898	 lr=0.010000
Epoch[037] Batch [0599]/[1203]	Speed: 106.436909 samples/sec	 accuracy=72.209167	 loss=1.061219	 lr=0.010000
Epoch[037] Batch [0619]/[1203]	Speed: 136.532474 samples/sec	 accuracy=72.177419	 loss=1.062348	 lr=0.010000
Epoch[037] Batch [0639]/[1203]	Speed: 108.747413 samples/sec	 accuracy=72.139844	 loss=1.063093	 lr=0.010000
Epoch[037] Batch [0659]/[1203]	Speed: 121.214179 samples/sec	 accuracy=72.093939	 loss=1.064501	 lr=0.010000
Epoch[037] Batch [0679]/[1203]	Speed: 143.221145 samples/sec	 accuracy=72.102941	 loss=1.065613	 lr=0.010000
Epoch[037] Batch [0699]/[1203]	Speed: 115.104572 samples/sec	 accuracy=72.098571	 loss=1.066155	 lr=0.010000
Epoch[037] Batch [0719]/[1203]	Speed: 143.277285 samples/sec	 accuracy=72.091667	 loss=1.066453	 lr=0.010000
Epoch[037] Batch [0739]/[1203]	Speed: 102.024431 samples/sec	 accuracy=72.085811	 loss=1.066607	 lr=0.010000
Epoch[037] Batch [0759]/[1203]	Speed: 134.584088 samples/sec	 accuracy=72.073684	 loss=1.067996	 lr=0.010000
Epoch[037] Batch [0779]/[1203]	Speed: 130.768645 samples/sec	 accuracy=72.046154	 loss=1.069116	 lr=0.010000
Epoch[037] Batch [0799]/[1203]	Speed: 122.909164 samples/sec	 accuracy=72.063750	 loss=1.068545	 lr=0.010000
Epoch[037] Batch [0819]/[1203]	Speed: 115.268896 samples/sec	 accuracy=72.054878	 loss=1.069228	 lr=0.010000
Epoch[037] Batch [0839]/[1203]	Speed: 121.532213 samples/sec	 accuracy=72.055952	 loss=1.069027	 lr=0.010000
Epoch[037] Batch [0859]/[1203]	Speed: 151.098502 samples/sec	 accuracy=72.045349	 loss=1.069701	 lr=0.010000
Epoch[037] Batch [0879]/[1203]	Speed: 118.037578 samples/sec	 accuracy=72.047727	 loss=1.069485	 lr=0.010000
Epoch[037] Batch [0899]/[1203]	Speed: 108.404541 samples/sec	 accuracy=72.012778	 loss=1.071191	 lr=0.010000
Epoch[037] Batch [0919]/[1203]	Speed: 145.698019 samples/sec	 accuracy=72.001630	 loss=1.071413	 lr=0.010000
Epoch[037] Batch [0939]/[1203]	Speed: 108.412659 samples/sec	 accuracy=72.001596	 loss=1.071917	 lr=0.010000
Epoch[037] Batch [0959]/[1203]	Speed: 121.172249 samples/sec	 accuracy=71.996354	 loss=1.072561	 lr=0.010000
Epoch[037] Batch [0979]/[1203]	Speed: 160.117752 samples/sec	 accuracy=71.977551	 loss=1.073086	 lr=0.010000
Epoch[037] Batch [0999]/[1203]	Speed: 102.136608 samples/sec	 accuracy=71.964500	 loss=1.073670	 lr=0.010000
Epoch[037] Batch [1019]/[1203]	Speed: 181.569092 samples/sec	 accuracy=71.930392	 loss=1.074811	 lr=0.010000
Epoch[037] Batch [1039]/[1203]	Speed: 100.916948 samples/sec	 accuracy=71.905769	 loss=1.075424	 lr=0.010000
Epoch[037] Batch [1059]/[1203]	Speed: 115.454048 samples/sec	 accuracy=71.902358	 loss=1.075352	 lr=0.010000
Epoch[037] Batch [1079]/[1203]	Speed: 160.496702 samples/sec	 accuracy=71.900000	 loss=1.075821	 lr=0.010000
Epoch[037] Batch [1099]/[1203]	Speed: 110.530674 samples/sec	 accuracy=71.889091	 loss=1.076492	 lr=0.010000
Epoch[037] Batch [1119]/[1203]	Speed: 143.986972 samples/sec	 accuracy=71.878571	 loss=1.076969	 lr=0.010000
Epoch[037] Batch [1139]/[1203]	Speed: 109.155632 samples/sec	 accuracy=71.877632	 loss=1.077523	 lr=0.010000
Epoch[037] Batch [1159]/[1203]	Speed: 109.611716 samples/sec	 accuracy=71.858621	 loss=1.078347	 lr=0.010000
Epoch[037] Batch [1179]/[1203]	Speed: 181.927642 samples/sec	 accuracy=71.825000	 loss=1.079630	 lr=0.010000
Epoch[037] Batch [1199]/[1203]	Speed: 197.663696 samples/sec	 accuracy=71.823750	 loss=1.080001	 lr=0.010000
Batch [0019]/[0097]: acc-top1=60.525000 acc-top5=82.125000
Batch [0039]/[0097]: acc-top1=59.425000 acc-top5=82.100000
Batch [0059]/[0097]: acc-top1=61.433333 acc-top5=83.233333
Batch [0079]/[0097]: acc-top1=61.425000 acc-top5=83.218750
[Epoch 037] training: accuracy=71.821696	 loss=1.080059
[Epoch 037] speed: 123 samples/sec	time cost: 2138.254034
[Epoch 037] validation: acc-top1=60.984536 acc-top5=83.237113 loss=1.709789
Epoch[038] Batch [0019]/[1203]	Speed: 69.256289 samples/sec	 accuracy=72.575000	 loss=1.055907	 lr=0.010000
Epoch[038] Batch [0039]/[1203]	Speed: 97.883056 samples/sec	 accuracy=73.062500	 loss=1.032361	 lr=0.010000
Epoch[038] Batch [0059]/[1203]	Speed: 216.539044 samples/sec	 accuracy=73.241667	 loss=1.020981	 lr=0.010000
Epoch[038] Batch [0079]/[1203]	Speed: 96.774705 samples/sec	 accuracy=73.443750	 loss=1.015805	 lr=0.010000
Epoch[038] Batch [0099]/[1203]	Speed: 120.546641 samples/sec	 accuracy=73.140000	 loss=1.023281	 lr=0.010000
Epoch[038] Batch [0119]/[1203]	Speed: 142.622621 samples/sec	 accuracy=73.025000	 loss=1.025905	 lr=0.010000
Epoch[038] Batch [0139]/[1203]	Speed: 108.896240 samples/sec	 accuracy=73.125000	 loss=1.026752	 lr=0.010000
Epoch[038] Batch [0159]/[1203]	Speed: 185.605311 samples/sec	 accuracy=73.262500	 loss=1.022597	 lr=0.010000
Epoch[038] Batch [0179]/[1203]	Speed: 101.034482 samples/sec	 accuracy=72.950000	 loss=1.030176	 lr=0.010000
Epoch[038] Batch [0199]/[1203]	Speed: 111.508132 samples/sec	 accuracy=72.880000	 loss=1.032123	 lr=0.010000
Epoch[038] Batch [0219]/[1203]	Speed: 135.380982 samples/sec	 accuracy=72.822727	 loss=1.036094	 lr=0.010000
Epoch[038] Batch [0239]/[1203]	Speed: 125.404140 samples/sec	 accuracy=72.825000	 loss=1.036141	 lr=0.010000
Epoch[038] Batch [0259]/[1203]	Speed: 129.529659 samples/sec	 accuracy=72.801923	 loss=1.038743	 lr=0.010000
Epoch[038] Batch [0279]/[1203]	Speed: 124.323400 samples/sec	 accuracy=72.862500	 loss=1.038093	 lr=0.010000
Epoch[038] Batch [0299]/[1203]	Speed: 109.498655 samples/sec	 accuracy=72.863333	 loss=1.037457	 lr=0.010000
Epoch[038] Batch [0319]/[1203]	Speed: 144.991206 samples/sec	 accuracy=72.796875	 loss=1.040673	 lr=0.010000
Epoch[038] Batch [0339]/[1203]	Speed: 126.860556 samples/sec	 accuracy=72.761765	 loss=1.042577	 lr=0.010000
Epoch[038] Batch [0359]/[1203]	Speed: 125.637393 samples/sec	 accuracy=72.681944	 loss=1.045443	 lr=0.010000
Epoch[038] Batch [0379]/[1203]	Speed: 115.152092 samples/sec	 accuracy=72.693421	 loss=1.044974	 lr=0.010000
Epoch[038] Batch [0399]/[1203]	Speed: 132.381620 samples/sec	 accuracy=72.626250	 loss=1.046861	 lr=0.010000
Epoch[038] Batch [0419]/[1203]	Speed: 128.289839 samples/sec	 accuracy=72.623810	 loss=1.045515	 lr=0.010000
Epoch[038] Batch [0439]/[1203]	Speed: 126.821621 samples/sec	 accuracy=72.597727	 loss=1.046496	 lr=0.010000
Epoch[038] Batch [0459]/[1203]	Speed: 112.535449 samples/sec	 accuracy=72.521739	 loss=1.048820	 lr=0.010000
Epoch[038] Batch [0479]/[1203]	Speed: 129.412437 samples/sec	 accuracy=72.539583	 loss=1.048316	 lr=0.010000
Epoch[038] Batch [0499]/[1203]	Speed: 126.591510 samples/sec	 accuracy=72.558000	 loss=1.048058	 lr=0.010000
Epoch[038] Batch [0519]/[1203]	Speed: 132.891585 samples/sec	 accuracy=72.538462	 loss=1.048830	 lr=0.010000
Epoch[038] Batch [0539]/[1203]	Speed: 111.587771 samples/sec	 accuracy=72.507407	 loss=1.048996	 lr=0.010000
Epoch[038] Batch [0559]/[1203]	Speed: 139.708136 samples/sec	 accuracy=72.534821	 loss=1.048608	 lr=0.010000
Epoch[038] Batch [0579]/[1203]	Speed: 118.988439 samples/sec	 accuracy=72.513793	 loss=1.049051	 lr=0.010000
Epoch[038] Batch [0599]/[1203]	Speed: 119.595368 samples/sec	 accuracy=72.496667	 loss=1.048978	 lr=0.010000
Epoch[038] Batch [0619]/[1203]	Speed: 137.350312 samples/sec	 accuracy=72.429839	 loss=1.050917	 lr=0.010000
Epoch[038] Batch [0639]/[1203]	Speed: 118.293498 samples/sec	 accuracy=72.397656	 loss=1.052232	 lr=0.010000
Epoch[038] Batch [0659]/[1203]	Speed: 125.576690 samples/sec	 accuracy=72.393939	 loss=1.052044	 lr=0.010000
Epoch[038] Batch [0679]/[1203]	Speed: 131.326834 samples/sec	 accuracy=72.399265	 loss=1.051730	 lr=0.010000
Epoch[038] Batch [0699]/[1203]	Speed: 110.053284 samples/sec	 accuracy=72.386429	 loss=1.052729	 lr=0.010000
Epoch[038] Batch [0719]/[1203]	Speed: 163.546638 samples/sec	 accuracy=72.335417	 loss=1.054059	 lr=0.010000
Epoch[038] Batch [0739]/[1203]	Speed: 103.419173 samples/sec	 accuracy=72.298649	 loss=1.055352	 lr=0.010000
Epoch[038] Batch [0759]/[1203]	Speed: 110.120917 samples/sec	 accuracy=72.272368	 loss=1.056779	 lr=0.010000
Epoch[038] Batch [0779]/[1203]	Speed: 166.482049 samples/sec	 accuracy=72.271795	 loss=1.056942	 lr=0.010000
Epoch[038] Batch [0799]/[1203]	Speed: 111.230758 samples/sec	 accuracy=72.248125	 loss=1.058161	 lr=0.010000
Epoch[038] Batch [0819]/[1203]	Speed: 130.317381 samples/sec	 accuracy=72.222561	 loss=1.059109	 lr=0.010000
Epoch[038] Batch [0839]/[1203]	Speed: 128.609673 samples/sec	 accuracy=72.214286	 loss=1.059386	 lr=0.010000
Epoch[038] Batch [0859]/[1203]	Speed: 106.902140 samples/sec	 accuracy=72.242442	 loss=1.059328	 lr=0.010000
Epoch[038] Batch [0879]/[1203]	Speed: 163.215535 samples/sec	 accuracy=72.219886	 loss=1.060183	 lr=0.010000
Epoch[038] Batch [0899]/[1203]	Speed: 106.461190 samples/sec	 accuracy=72.216111	 loss=1.060603	 lr=0.010000
Epoch[038] Batch [0919]/[1203]	Speed: 124.747013 samples/sec	 accuracy=72.189674	 loss=1.061815	 lr=0.010000
Epoch[038] Batch [0939]/[1203]	Speed: 139.567558 samples/sec	 accuracy=72.181915	 loss=1.061769	 lr=0.010000
Epoch[038] Batch [0959]/[1203]	Speed: 111.293964 samples/sec	 accuracy=72.172396	 loss=1.062039	 lr=0.010000
Epoch[038] Batch [0979]/[1203]	Speed: 167.621982 samples/sec	 accuracy=72.156633	 loss=1.062889	 lr=0.010000
Epoch[038] Batch [0999]/[1203]	Speed: 104.112568 samples/sec	 accuracy=72.120500	 loss=1.064427	 lr=0.010000
Epoch[038] Batch [1019]/[1203]	Speed: 122.431648 samples/sec	 accuracy=72.119118	 loss=1.064513	 lr=0.010000
Epoch[038] Batch [1039]/[1203]	Speed: 140.342619 samples/sec	 accuracy=72.121635	 loss=1.064286	 lr=0.010000
Epoch[038] Batch [1059]/[1203]	Speed: 116.450789 samples/sec	 accuracy=72.133491	 loss=1.063713	 lr=0.010000
Epoch[038] Batch [1079]/[1203]	Speed: 142.960568 samples/sec	 accuracy=72.112963	 loss=1.064352	 lr=0.010000
Epoch[038] Batch [1099]/[1203]	Speed: 112.420937 samples/sec	 accuracy=72.095455	 loss=1.064510	 lr=0.010000
Epoch[038] Batch [1119]/[1203]	Speed: 121.302962 samples/sec	 accuracy=72.076339	 loss=1.065501	 lr=0.010000
Epoch[038] Batch [1139]/[1203]	Speed: 139.158626 samples/sec	 accuracy=72.056579	 loss=1.066271	 lr=0.010000
Epoch[038] Batch [1159]/[1203]	Speed: 116.197473 samples/sec	 accuracy=72.022845	 loss=1.068060	 lr=0.010000
Epoch[038] Batch [1179]/[1203]	Speed: 152.410017 samples/sec	 accuracy=72.019492	 loss=1.068478	 lr=0.010000
Epoch[038] Batch [1199]/[1203]	Speed: 202.181128 samples/sec	 accuracy=71.998333	 loss=1.069507	 lr=0.010000
Batch [0019]/[0097]: acc-top1=59.725000 acc-top5=81.275000
Batch [0039]/[0097]: acc-top1=58.612500 acc-top5=81.037500
Batch [0059]/[0097]: acc-top1=61.008333 acc-top5=82.408333
Batch [0079]/[0097]: acc-top1=61.331250 acc-top5=82.512500
[Epoch 038] training: accuracy=72.000000	 loss=1.069602
[Epoch 038] speed: 124 samples/sec	time cost: 2131.827854
[Epoch 038] validation: acc-top1=60.484536 acc-top5=82.360825 loss=1.746056
Epoch[039] Batch [0019]/[1203]	Speed: 68.900197 samples/sec	 accuracy=73.450000	 loss=1.010703	 lr=0.010000
Epoch[039] Batch [0039]/[1203]	Speed: 98.892962 samples/sec	 accuracy=73.587500	 loss=1.000266	 lr=0.010000
Epoch[039] Batch [0059]/[1203]	Speed: 251.653815 samples/sec	 accuracy=73.558333	 loss=1.008744	 lr=0.010000
Epoch[039] Batch [0079]/[1203]	Speed: 93.274007 samples/sec	 accuracy=73.406250	 loss=1.019852	 lr=0.010000
Epoch[039] Batch [0099]/[1203]	Speed: 100.267847 samples/sec	 accuracy=73.635000	 loss=1.019572	 lr=0.010000
Epoch[039] Batch [0119]/[1203]	Speed: 222.077696 samples/sec	 accuracy=73.558333	 loss=1.022994	 lr=0.010000
Epoch[039] Batch [0139]/[1203]	Speed: 103.153892 samples/sec	 accuracy=73.471429	 loss=1.026649	 lr=0.010000
Epoch[039] Batch [0159]/[1203]	Speed: 190.407304 samples/sec	 accuracy=73.368750	 loss=1.029533	 lr=0.010000
Epoch[039] Batch [0179]/[1203]	Speed: 94.741853 samples/sec	 accuracy=73.283333	 loss=1.030924	 lr=0.010000
Epoch[039] Batch [0199]/[1203]	Speed: 111.965284 samples/sec	 accuracy=73.405000	 loss=1.027511	 lr=0.010000
Epoch[039] Batch [0219]/[1203]	Speed: 153.065442 samples/sec	 accuracy=73.406818	 loss=1.027516	 lr=0.010000
Epoch[039] Batch [0239]/[1203]	Speed: 107.853139 samples/sec	 accuracy=73.445833	 loss=1.026081	 lr=0.010000
Epoch[039] Batch [0259]/[1203]	Speed: 140.082771 samples/sec	 accuracy=73.275000	 loss=1.028829	 lr=0.010000
Epoch[039] Batch [0279]/[1203]	Speed: 123.376229 samples/sec	 accuracy=73.262500	 loss=1.027231	 lr=0.010000
Epoch[039] Batch [0299]/[1203]	Speed: 116.086890 samples/sec	 accuracy=73.236667	 loss=1.028254	 lr=0.010000
Epoch[039] Batch [0319]/[1203]	Speed: 151.425838 samples/sec	 accuracy=73.195312	 loss=1.029831	 lr=0.010000
Epoch[039] Batch [0339]/[1203]	Speed: 105.520730 samples/sec	 accuracy=73.175000	 loss=1.030164	 lr=0.010000
Epoch[039] Batch [0359]/[1203]	Speed: 121.412866 samples/sec	 accuracy=73.087500	 loss=1.032801	 lr=0.010000
Epoch[039] Batch [0379]/[1203]	Speed: 147.425296 samples/sec	 accuracy=73.003947	 loss=1.036130	 lr=0.010000
Epoch[039] Batch [0399]/[1203]	Speed: 104.382310 samples/sec	 accuracy=72.992500	 loss=1.036512	 lr=0.010000
Epoch[039] Batch [0419]/[1203]	Speed: 170.968294 samples/sec	 accuracy=72.929762	 loss=1.038142	 lr=0.010000
Epoch[039] Batch [0439]/[1203]	Speed: 108.239538 samples/sec	 accuracy=72.942045	 loss=1.037578	 lr=0.010000
Epoch[039] Batch [0459]/[1203]	Speed: 107.561791 samples/sec	 accuracy=72.928261	 loss=1.039441	 lr=0.010000
Epoch[039] Batch [0479]/[1203]	Speed: 165.589198 samples/sec	 accuracy=72.905208	 loss=1.039213	 lr=0.010000
Epoch[039] Batch [0499]/[1203]	Speed: 109.870581 samples/sec	 accuracy=72.886000	 loss=1.040376	 lr=0.010000
Epoch[039] Batch [0519]/[1203]	Speed: 130.342869 samples/sec	 accuracy=72.871154	 loss=1.041418	 lr=0.010000
Epoch[039] Batch [0539]/[1203]	Speed: 129.568320 samples/sec	 accuracy=72.859259	 loss=1.040942	 lr=0.010000
Epoch[039] Batch [0559]/[1203]	Speed: 102.893704 samples/sec	 accuracy=72.825893	 loss=1.041548	 lr=0.010000
Epoch[039] Batch [0579]/[1203]	Speed: 173.262094 samples/sec	 accuracy=72.829310	 loss=1.042394	 lr=0.010000
Epoch[039] Batch [0599]/[1203]	Speed: 113.603840 samples/sec	 accuracy=72.783333	 loss=1.042934	 lr=0.010000
Epoch[039] Batch [0619]/[1203]	Speed: 141.687040 samples/sec	 accuracy=72.761290	 loss=1.043283	 lr=0.010000
Epoch[039] Batch [0639]/[1203]	Speed: 120.128679 samples/sec	 accuracy=72.718750	 loss=1.044712	 lr=0.010000
Epoch[039] Batch [0659]/[1203]	Speed: 106.778405 samples/sec	 accuracy=72.654545	 loss=1.046437	 lr=0.010000
Epoch[039] Batch [0679]/[1203]	Speed: 141.507817 samples/sec	 accuracy=72.642647	 loss=1.048022	 lr=0.010000
Epoch[039] Batch [0699]/[1203]	Speed: 129.983125 samples/sec	 accuracy=72.630000	 loss=1.048918	 lr=0.010000
Epoch[039] Batch [0719]/[1203]	Speed: 101.839258 samples/sec	 accuracy=72.606250	 loss=1.049294	 lr=0.010000
Epoch[039] Batch [0739]/[1203]	Speed: 149.598488 samples/sec	 accuracy=72.600676	 loss=1.050109	 lr=0.010000
Epoch[039] Batch [0759]/[1203]	Speed: 121.128719 samples/sec	 accuracy=72.603947	 loss=1.050008	 lr=0.010000
Epoch[039] Batch [0779]/[1203]	Speed: 130.373597 samples/sec	 accuracy=72.595513	 loss=1.049697	 lr=0.010000
Epoch[039] Batch [0799]/[1203]	Speed: 114.800985 samples/sec	 accuracy=72.563750	 loss=1.051244	 lr=0.010000
Epoch[039] Batch [0819]/[1203]	Speed: 110.965349 samples/sec	 accuracy=72.550610	 loss=1.051472	 lr=0.010000
Epoch[039] Batch [0839]/[1203]	Speed: 170.131919 samples/sec	 accuracy=72.542857	 loss=1.051934	 lr=0.010000
Epoch[039] Batch [0859]/[1203]	Speed: 108.116573 samples/sec	 accuracy=72.489535	 loss=1.053244	 lr=0.010000
Epoch[039] Batch [0879]/[1203]	Speed: 110.526333 samples/sec	 accuracy=72.468182	 loss=1.053741	 lr=0.010000
Epoch[039] Batch [0899]/[1203]	Speed: 147.039444 samples/sec	 accuracy=72.465000	 loss=1.054522	 lr=0.010000
Epoch[039] Batch [0919]/[1203]	Speed: 113.754873 samples/sec	 accuracy=72.457609	 loss=1.054915	 lr=0.010000
Epoch[039] Batch [0939]/[1203]	Speed: 156.339226 samples/sec	 accuracy=72.410106	 loss=1.056101	 lr=0.010000
Epoch[039] Batch [0959]/[1203]	Speed: 121.382396 samples/sec	 accuracy=72.385417	 loss=1.057337	 lr=0.010000
Epoch[039] Batch [0979]/[1203]	Speed: 102.638558 samples/sec	 accuracy=72.378061	 loss=1.057088	 lr=0.010000
Epoch[039] Batch [0999]/[1203]	Speed: 150.036859 samples/sec	 accuracy=72.387500	 loss=1.056679	 lr=0.010000
Epoch[039] Batch [1019]/[1203]	Speed: 121.246343 samples/sec	 accuracy=72.358824	 loss=1.057350	 lr=0.010000
Epoch[039] Batch [1039]/[1203]	Speed: 125.661329 samples/sec	 accuracy=72.344231	 loss=1.057905	 lr=0.010000
Epoch[039] Batch [1059]/[1203]	Speed: 128.187902 samples/sec	 accuracy=72.317925	 loss=1.058767	 lr=0.010000
Epoch[039] Batch [1079]/[1203]	Speed: 107.561872 samples/sec	 accuracy=72.287037	 loss=1.060162	 lr=0.010000
Epoch[039] Batch [1099]/[1203]	Speed: 162.140046 samples/sec	 accuracy=72.264091	 loss=1.060326	 lr=0.010000
Epoch[039] Batch [1119]/[1203]	Speed: 115.490354 samples/sec	 accuracy=72.244643	 loss=1.060597	 lr=0.010000
Epoch[039] Batch [1139]/[1203]	Speed: 130.408747 samples/sec	 accuracy=72.239035	 loss=1.061483	 lr=0.010000
Epoch[039] Batch [1159]/[1203]	Speed: 131.823032 samples/sec	 accuracy=72.224138	 loss=1.062033	 lr=0.010000
Epoch[039] Batch [1179]/[1203]	Speed: 104.432747 samples/sec	 accuracy=72.217373	 loss=1.062643	 lr=0.010000
Epoch[039] Batch [1199]/[1203]	Speed: 330.816731 samples/sec	 accuracy=72.208750	 loss=1.063026	 lr=0.001000
Batch [0019]/[0097]: acc-top1=60.250000 acc-top5=82.275000
Batch [0039]/[0097]: acc-top1=59.187500 acc-top5=81.562500
Batch [0059]/[0097]: acc-top1=61.591667 acc-top5=82.991667
Batch [0079]/[0097]: acc-top1=61.193750 acc-top5=82.793750
[Epoch 039] training: accuracy=72.216542	 loss=1.062885
[Epoch 039] speed: 124 samples/sec	time cost: 2131.137859
[Epoch 039] validation: acc-top1=60.587629 acc-top5=82.829897 loss=1.729005
Epoch[040] Batch [0019]/[1203]	Speed: 69.258061 samples/sec	 accuracy=73.100000	 loss=1.015715	 lr=0.001000
Epoch[040] Batch [0039]/[1203]	Speed: 101.042724 samples/sec	 accuracy=73.425000	 loss=1.010051	 lr=0.001000
Epoch[040] Batch [0059]/[1203]	Speed: 220.396988 samples/sec	 accuracy=73.441667	 loss=1.011640	 lr=0.001000
Epoch[040] Batch [0079]/[1203]	Speed: 94.832403 samples/sec	 accuracy=73.600000	 loss=1.005276	 lr=0.001000
Epoch[040] Batch [0099]/[1203]	Speed: 118.249352 samples/sec	 accuracy=74.055000	 loss=0.991778	 lr=0.001000
Epoch[040] Batch [0119]/[1203]	Speed: 157.457877 samples/sec	 accuracy=74.170833	 loss=0.986336	 lr=0.001000
Epoch[040] Batch [0139]/[1203]	Speed: 106.256131 samples/sec	 accuracy=74.471429	 loss=0.976203	 lr=0.001000
Epoch[040] Batch [0159]/[1203]	Speed: 183.711750 samples/sec	 accuracy=74.487500	 loss=0.976694	 lr=0.001000
Epoch[040] Batch [0179]/[1203]	Speed: 100.669749 samples/sec	 accuracy=74.661111	 loss=0.968724	 lr=0.001000
Epoch[040] Batch [0199]/[1203]	Speed: 121.653178 samples/sec	 accuracy=74.920000	 loss=0.960285	 lr=0.001000
Epoch[040] Batch [0219]/[1203]	Speed: 139.182544 samples/sec	 accuracy=75.056818	 loss=0.956020	 lr=0.001000
Epoch[040] Batch [0239]/[1203]	Speed: 113.724679 samples/sec	 accuracy=75.268750	 loss=0.949869	 lr=0.001000
Epoch[040] Batch [0259]/[1203]	Speed: 160.911026 samples/sec	 accuracy=75.248077	 loss=0.949684	 lr=0.001000
Epoch[040] Batch [0279]/[1203]	Speed: 103.191217 samples/sec	 accuracy=75.276786	 loss=0.948205	 lr=0.001000
Epoch[040] Batch [0299]/[1203]	Speed: 111.393128 samples/sec	 accuracy=75.425000	 loss=0.944206	 lr=0.001000
Epoch[040] Batch [0319]/[1203]	Speed: 159.873606 samples/sec	 accuracy=75.589062	 loss=0.938468	 lr=0.001000
Epoch[040] Batch [0339]/[1203]	Speed: 109.726675 samples/sec	 accuracy=75.666176	 loss=0.935636	 lr=0.001000
Epoch[040] Batch [0359]/[1203]	Speed: 146.795149 samples/sec	 accuracy=75.684722	 loss=0.932505	 lr=0.001000
Epoch[040] Batch [0379]/[1203]	Speed: 110.333924 samples/sec	 accuracy=75.710526	 loss=0.930138	 lr=0.001000
Epoch[040] Batch [0399]/[1203]	Speed: 119.000488 samples/sec	 accuracy=75.821250	 loss=0.926829	 lr=0.001000
Epoch[040] Batch [0419]/[1203]	Speed: 152.556102 samples/sec	 accuracy=75.873810	 loss=0.924508	 lr=0.001000
Epoch[040] Batch [0439]/[1203]	Speed: 103.763904 samples/sec	 accuracy=75.944318	 loss=0.923278	 lr=0.001000
Epoch[040] Batch [0459]/[1203]	Speed: 119.842064 samples/sec	 accuracy=75.936957	 loss=0.923272	 lr=0.001000
Epoch[040] Batch [0479]/[1203]	Speed: 140.832515 samples/sec	 accuracy=75.954167	 loss=0.921016	 lr=0.001000
Epoch[040] Batch [0499]/[1203]	Speed: 111.469414 samples/sec	 accuracy=75.996000	 loss=0.919632	 lr=0.001000
Epoch[040] Batch [0519]/[1203]	Speed: 154.297037 samples/sec	 accuracy=76.012500	 loss=0.918983	 lr=0.001000
Epoch[040] Batch [0539]/[1203]	Speed: 123.494596 samples/sec	 accuracy=76.025000	 loss=0.918482	 lr=0.001000
Epoch[040] Batch [0559]/[1203]	Speed: 115.990816 samples/sec	 accuracy=76.061607	 loss=0.917187	 lr=0.001000
Epoch[040] Batch [0579]/[1203]	Speed: 135.464277 samples/sec	 accuracy=76.106034	 loss=0.915494	 lr=0.001000
Epoch[040] Batch [0599]/[1203]	Speed: 114.406647 samples/sec	 accuracy=76.137500	 loss=0.913983	 lr=0.001000
Epoch[040] Batch [0619]/[1203]	Speed: 127.178615 samples/sec	 accuracy=76.166935	 loss=0.913186	 lr=0.001000
Epoch[040] Batch [0639]/[1203]	Speed: 124.875214 samples/sec	 accuracy=76.204688	 loss=0.911746	 lr=0.001000
Epoch[040] Batch [0659]/[1203]	Speed: 120.993041 samples/sec	 accuracy=76.268182	 loss=0.909666	 lr=0.001000
Epoch[040] Batch [0679]/[1203]	Speed: 138.110119 samples/sec	 accuracy=76.289706	 loss=0.908530	 lr=0.001000
Epoch[040] Batch [0699]/[1203]	Speed: 116.214532 samples/sec	 accuracy=76.307143	 loss=0.907517	 lr=0.001000
Epoch[040] Batch [0719]/[1203]	Speed: 125.860446 samples/sec	 accuracy=76.335417	 loss=0.906424	 lr=0.001000
Epoch[040] Batch [0739]/[1203]	Speed: 121.070614 samples/sec	 accuracy=76.342568	 loss=0.906434	 lr=0.001000
Epoch[040] Batch [0759]/[1203]	Speed: 125.912604 samples/sec	 accuracy=76.401974	 loss=0.904665	 lr=0.001000
Epoch[040] Batch [0779]/[1203]	Speed: 145.197724 samples/sec	 accuracy=76.419872	 loss=0.904227	 lr=0.001000
Epoch[040] Batch [0799]/[1203]	Speed: 117.736552 samples/sec	 accuracy=76.445625	 loss=0.903456	 lr=0.001000
Epoch[040] Batch [0819]/[1203]	Speed: 116.196674 samples/sec	 accuracy=76.471951	 loss=0.902425	 lr=0.001000
Epoch[040] Batch [0839]/[1203]	Speed: 116.501548 samples/sec	 accuracy=76.493452	 loss=0.901357	 lr=0.001000
Epoch[040] Batch [0859]/[1203]	Speed: 137.964339 samples/sec	 accuracy=76.522674	 loss=0.899890	 lr=0.001000
Epoch[040] Batch [0879]/[1203]	Speed: 131.386909 samples/sec	 accuracy=76.551705	 loss=0.898913	 lr=0.001000
Epoch[040] Batch [0899]/[1203]	Speed: 114.784556 samples/sec	 accuracy=76.575000	 loss=0.898064	 lr=0.001000
Epoch[040] Batch [0919]/[1203]	Speed: 115.979807 samples/sec	 accuracy=76.582609	 loss=0.897440	 lr=0.001000
Epoch[040] Batch [0939]/[1203]	Speed: 136.671532 samples/sec	 accuracy=76.590957	 loss=0.897237	 lr=0.001000
Epoch[040] Batch [0959]/[1203]	Speed: 120.231748 samples/sec	 accuracy=76.582813	 loss=0.897294	 lr=0.001000
Epoch[040] Batch [0979]/[1203]	Speed: 141.839531 samples/sec	 accuracy=76.620918	 loss=0.896004	 lr=0.001000
Epoch[040] Batch [0999]/[1203]	Speed: 117.963024 samples/sec	 accuracy=76.639000	 loss=0.895200	 lr=0.001000
Epoch[040] Batch [1019]/[1203]	Speed: 115.286082 samples/sec	 accuracy=76.675490	 loss=0.894150	 lr=0.001000
Epoch[040] Batch [1039]/[1203]	Speed: 131.513745 samples/sec	 accuracy=76.683654	 loss=0.893688	 lr=0.001000
Epoch[040] Batch [1059]/[1203]	Speed: 134.321890 samples/sec	 accuracy=76.708019	 loss=0.892913	 lr=0.001000
Epoch[040] Batch [1079]/[1203]	Speed: 110.707170 samples/sec	 accuracy=76.709722	 loss=0.892192	 lr=0.001000
Epoch[040] Batch [1099]/[1203]	Speed: 145.200178 samples/sec	 accuracy=76.734545	 loss=0.891279	 lr=0.001000
Epoch[040] Batch [1119]/[1203]	Speed: 109.320608 samples/sec	 accuracy=76.750000	 loss=0.890860	 lr=0.001000
Epoch[040] Batch [1139]/[1203]	Speed: 127.156663 samples/sec	 accuracy=76.778947	 loss=0.889972	 lr=0.001000
Epoch[040] Batch [1159]/[1203]	Speed: 140.908381 samples/sec	 accuracy=76.802155	 loss=0.889379	 lr=0.001000
Epoch[040] Batch [1179]/[1203]	Speed: 118.250283 samples/sec	 accuracy=76.802966	 loss=0.889436	 lr=0.001000
Epoch[040] Batch [1199]/[1203]	Speed: 252.306244 samples/sec	 accuracy=76.810000	 loss=0.889284	 lr=0.001000
Batch [0019]/[0097]: acc-top1=63.875000 acc-top5=84.875000
Batch [0039]/[0097]: acc-top1=63.087500 acc-top5=84.700000
Batch [0059]/[0097]: acc-top1=65.216667 acc-top5=85.883333
Batch [0079]/[0097]: acc-top1=65.318750 acc-top5=85.681250
[Epoch 040] training: accuracy=76.807149	 loss=0.889488
[Epoch 040] speed: 124 samples/sec	time cost: 2129.641500
[Epoch 040] validation: acc-top1=64.768041 acc-top5=85.577320 loss=1.531725
Epoch[041] Batch [0019]/[1203]	Speed: 68.441494 samples/sec	 accuracy=78.825000	 loss=0.828179	 lr=0.001000
Epoch[041] Batch [0039]/[1203]	Speed: 99.187250 samples/sec	 accuracy=78.275000	 loss=0.841240	 lr=0.001000
Epoch[041] Batch [0059]/[1203]	Speed: 222.942184 samples/sec	 accuracy=78.491667	 loss=0.830762	 lr=0.001000
Epoch[041] Batch [0079]/[1203]	Speed: 98.223455 samples/sec	 accuracy=78.456250	 loss=0.829650	 lr=0.001000
Epoch[041] Batch [0099]/[1203]	Speed: 111.480870 samples/sec	 accuracy=78.440000	 loss=0.828763	 lr=0.001000
Epoch[041] Batch [0119]/[1203]	Speed: 168.795545 samples/sec	 accuracy=78.341667	 loss=0.830765	 lr=0.001000
Epoch[041] Batch [0139]/[1203]	Speed: 105.005073 samples/sec	 accuracy=78.321429	 loss=0.832865	 lr=0.001000
Epoch[041] Batch [0159]/[1203]	Speed: 180.614618 samples/sec	 accuracy=78.215625	 loss=0.834750	 lr=0.001000
Epoch[041] Batch [0179]/[1203]	Speed: 105.718312 samples/sec	 accuracy=78.186111	 loss=0.835909	 lr=0.001000
Epoch[041] Batch [0199]/[1203]	Speed: 110.261706 samples/sec	 accuracy=78.067500	 loss=0.837695	 lr=0.001000
Epoch[041] Batch [0219]/[1203]	Speed: 141.994654 samples/sec	 accuracy=78.170455	 loss=0.836240	 lr=0.001000
Epoch[041] Batch [0239]/[1203]	Speed: 112.102025 samples/sec	 accuracy=78.195833	 loss=0.836547	 lr=0.001000
Epoch[041] Batch [0259]/[1203]	Speed: 148.005198 samples/sec	 accuracy=78.148077	 loss=0.838047	 lr=0.001000
Epoch[041] Batch [0279]/[1203]	Speed: 115.204003 samples/sec	 accuracy=78.189286	 loss=0.835634	 lr=0.001000
Epoch[041] Batch [0299]/[1203]	Speed: 123.214181 samples/sec	 accuracy=78.240000	 loss=0.835267	 lr=0.001000
Epoch[041] Batch [0319]/[1203]	Speed: 134.763607 samples/sec	 accuracy=78.181250	 loss=0.836364	 lr=0.001000
Epoch[041] Batch [0339]/[1203]	Speed: 121.292407 samples/sec	 accuracy=78.186765	 loss=0.835417	 lr=0.001000
Epoch[041] Batch [0359]/[1203]	Speed: 116.228104 samples/sec	 accuracy=78.265278	 loss=0.834328	 lr=0.001000
Epoch[041] Batch [0379]/[1203]	Speed: 124.633542 samples/sec	 accuracy=78.256579	 loss=0.833332	 lr=0.001000
Epoch[041] Batch [0399]/[1203]	Speed: 118.339883 samples/sec	 accuracy=78.253750	 loss=0.833470	 lr=0.001000
Epoch[041] Batch [0419]/[1203]	Speed: 147.732310 samples/sec	 accuracy=78.217857	 loss=0.834492	 lr=0.001000
Epoch[041] Batch [0439]/[1203]	Speed: 131.430256 samples/sec	 accuracy=78.180682	 loss=0.836248	 lr=0.001000
Epoch[041] Batch [0459]/[1203]	Speed: 119.707926 samples/sec	 accuracy=78.158696	 loss=0.835834	 lr=0.001000
Epoch[041] Batch [0479]/[1203]	Speed: 119.242245 samples/sec	 accuracy=78.165625	 loss=0.835558	 lr=0.001000
Epoch[041] Batch [0499]/[1203]	Speed: 113.127541 samples/sec	 accuracy=78.162000	 loss=0.835634	 lr=0.001000
Epoch[041] Batch [0519]/[1203]	Speed: 155.348622 samples/sec	 accuracy=78.173077	 loss=0.834776	 lr=0.001000
Epoch[041] Batch [0539]/[1203]	Speed: 119.535354 samples/sec	 accuracy=78.142593	 loss=0.834592	 lr=0.001000
Epoch[041] Batch [0559]/[1203]	Speed: 112.280430 samples/sec	 accuracy=78.155357	 loss=0.834131	 lr=0.001000
Epoch[041] Batch [0579]/[1203]	Speed: 137.334414 samples/sec	 accuracy=78.173276	 loss=0.833543	 lr=0.001000
Epoch[041] Batch [0599]/[1203]	Speed: 107.189703 samples/sec	 accuracy=78.180000	 loss=0.833398	 lr=0.001000
Epoch[041] Batch [0619]/[1203]	Speed: 160.210684 samples/sec	 accuracy=78.181452	 loss=0.834047	 lr=0.001000
Epoch[041] Batch [0639]/[1203]	Speed: 114.423715 samples/sec	 accuracy=78.214062	 loss=0.833341	 lr=0.001000
Epoch[041] Batch [0659]/[1203]	Speed: 111.774071 samples/sec	 accuracy=78.223485	 loss=0.833322	 lr=0.001000
Epoch[041] Batch [0679]/[1203]	Speed: 132.754221 samples/sec	 accuracy=78.224265	 loss=0.833743	 lr=0.001000
Epoch[041] Batch [0699]/[1203]	Speed: 120.749158 samples/sec	 accuracy=78.217143	 loss=0.833894	 lr=0.001000
Epoch[041] Batch [0719]/[1203]	Speed: 126.192259 samples/sec	 accuracy=78.209028	 loss=0.833946	 lr=0.001000
Epoch[041] Batch [0739]/[1203]	Speed: 143.144686 samples/sec	 accuracy=78.216216	 loss=0.833690	 lr=0.001000
Epoch[041] Batch [0759]/[1203]	Speed: 114.841919 samples/sec	 accuracy=78.216447	 loss=0.833765	 lr=0.001000
Epoch[041] Batch [0779]/[1203]	Speed: 132.154739 samples/sec	 accuracy=78.217308	 loss=0.833849	 lr=0.001000
Epoch[041] Batch [0799]/[1203]	Speed: 120.593437 samples/sec	 accuracy=78.240000	 loss=0.832804	 lr=0.001000
Epoch[041] Batch [0819]/[1203]	Speed: 130.873350 samples/sec	 accuracy=78.223780	 loss=0.832884	 lr=0.001000
Epoch[041] Batch [0839]/[1203]	Speed: 121.488007 samples/sec	 accuracy=78.263095	 loss=0.831918	 lr=0.001000
Epoch[041] Batch [0859]/[1203]	Speed: 134.294972 samples/sec	 accuracy=78.231395	 loss=0.832192	 lr=0.001000
Epoch[041] Batch [0879]/[1203]	Speed: 117.834833 samples/sec	 accuracy=78.261932	 loss=0.831779	 lr=0.001000
Epoch[041] Batch [0899]/[1203]	Speed: 122.396122 samples/sec	 accuracy=78.288333	 loss=0.830959	 lr=0.001000
Epoch[041] Batch [0919]/[1203]	Speed: 144.654772 samples/sec	 accuracy=78.268478	 loss=0.831582	 lr=0.001000
Epoch[041] Batch [0939]/[1203]	Speed: 113.900087 samples/sec	 accuracy=78.244681	 loss=0.832079	 lr=0.001000
Epoch[041] Batch [0959]/[1203]	Speed: 121.987216 samples/sec	 accuracy=78.263542	 loss=0.831455	 lr=0.001000
Epoch[041] Batch [0979]/[1203]	Speed: 119.932249 samples/sec	 accuracy=78.271429	 loss=0.831325	 lr=0.001000
Epoch[041] Batch [0999]/[1203]	Speed: 132.138363 samples/sec	 accuracy=78.301500	 loss=0.830467	 lr=0.001000
Epoch[041] Batch [1019]/[1203]	Speed: 136.105856 samples/sec	 accuracy=78.308333	 loss=0.830622	 lr=0.001000
Epoch[041] Batch [1039]/[1203]	Speed: 112.284119 samples/sec	 accuracy=78.329808	 loss=0.829804	 lr=0.001000
Epoch[041] Batch [1059]/[1203]	Speed: 132.351783 samples/sec	 accuracy=78.328774	 loss=0.830174	 lr=0.001000
Epoch[041] Batch [1079]/[1203]	Speed: 121.122138 samples/sec	 accuracy=78.350926	 loss=0.829434	 lr=0.001000
Epoch[041] Batch [1099]/[1203]	Speed: 123.297005 samples/sec	 accuracy=78.344545	 loss=0.829885	 lr=0.001000
Epoch[041] Batch [1119]/[1203]	Speed: 123.858194 samples/sec	 accuracy=78.319196	 loss=0.830348	 lr=0.001000
Epoch[041] Batch [1139]/[1203]	Speed: 117.346558 samples/sec	 accuracy=78.320175	 loss=0.830030	 lr=0.001000
Epoch[041] Batch [1159]/[1203]	Speed: 140.413819 samples/sec	 accuracy=78.319828	 loss=0.830033	 lr=0.001000
Epoch[041] Batch [1179]/[1203]	Speed: 138.253689 samples/sec	 accuracy=78.296186	 loss=0.830115	 lr=0.001000
Epoch[041] Batch [1199]/[1203]	Speed: 214.786720 samples/sec	 accuracy=78.289583	 loss=0.830134	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.400000 acc-top5=85.075000
Batch [0039]/[0097]: acc-top1=63.375000 acc-top5=84.437500
Batch [0059]/[0097]: acc-top1=65.475000 acc-top5=85.675000
Batch [0079]/[0097]: acc-top1=65.675000 acc-top5=85.518750
[Epoch 041] training: accuracy=78.285121	 loss=0.830373
[Epoch 041] speed: 124 samples/sec	time cost: 2127.239551
[Epoch 041] validation: acc-top1=65.020619 acc-top5=85.396907 loss=1.527571
Epoch[042] Batch [0019]/[1203]	Speed: 69.267602 samples/sec	 accuracy=79.075000	 loss=0.794907	 lr=0.001000
Epoch[042] Batch [0039]/[1203]	Speed: 103.085795 samples/sec	 accuracy=79.125000	 loss=0.797577	 lr=0.001000
Epoch[042] Batch [0059]/[1203]	Speed: 216.460225 samples/sec	 accuracy=79.033333	 loss=0.809647	 lr=0.001000
Epoch[042] Batch [0079]/[1203]	Speed: 91.397394 samples/sec	 accuracy=79.187500	 loss=0.806826	 lr=0.001000
Epoch[042] Batch [0099]/[1203]	Speed: 119.974028 samples/sec	 accuracy=79.075000	 loss=0.808934	 lr=0.001000
Epoch[042] Batch [0119]/[1203]	Speed: 148.713789 samples/sec	 accuracy=79.387500	 loss=0.799245	 lr=0.001000
Epoch[042] Batch [0139]/[1203]	Speed: 110.233804 samples/sec	 accuracy=79.557143	 loss=0.794351	 lr=0.001000
Epoch[042] Batch [0159]/[1203]	Speed: 193.150117 samples/sec	 accuracy=79.453125	 loss=0.795446	 lr=0.001000
Epoch[042] Batch [0179]/[1203]	Speed: 96.761991 samples/sec	 accuracy=79.430556	 loss=0.797118	 lr=0.001000
Epoch[042] Batch [0199]/[1203]	Speed: 108.300923 samples/sec	 accuracy=79.475000	 loss=0.794647	 lr=0.001000
Epoch[042] Batch [0219]/[1203]	Speed: 176.260571 samples/sec	 accuracy=79.500000	 loss=0.792399	 lr=0.001000
Epoch[042] Batch [0239]/[1203]	Speed: 106.462616 samples/sec	 accuracy=79.481250	 loss=0.791399	 lr=0.001000
Epoch[042] Batch [0259]/[1203]	Speed: 180.381864 samples/sec	 accuracy=79.411538	 loss=0.792950	 lr=0.001000
Epoch[042] Batch [0279]/[1203]	Speed: 103.782840 samples/sec	 accuracy=79.376786	 loss=0.793817	 lr=0.001000
Epoch[042] Batch [0299]/[1203]	Speed: 101.426559 samples/sec	 accuracy=79.345000	 loss=0.794742	 lr=0.001000
Epoch[042] Batch [0319]/[1203]	Speed: 177.658472 samples/sec	 accuracy=79.285938	 loss=0.795832	 lr=0.001000
Epoch[042] Batch [0339]/[1203]	Speed: 111.663181 samples/sec	 accuracy=79.244118	 loss=0.798086	 lr=0.001000
Epoch[042] Batch [0359]/[1203]	Speed: 115.856302 samples/sec	 accuracy=79.144444	 loss=0.801328	 lr=0.001000
Epoch[042] Batch [0379]/[1203]	Speed: 152.933986 samples/sec	 accuracy=79.057895	 loss=0.803025	 lr=0.001000
Epoch[042] Batch [0399]/[1203]	Speed: 109.482162 samples/sec	 accuracy=79.055000	 loss=0.803702	 lr=0.001000
Epoch[042] Batch [0419]/[1203]	Speed: 166.397197 samples/sec	 accuracy=79.029762	 loss=0.803882	 lr=0.001000
Epoch[042] Batch [0439]/[1203]	Speed: 102.200589 samples/sec	 accuracy=78.964773	 loss=0.805607	 lr=0.001000
Epoch[042] Batch [0459]/[1203]	Speed: 121.140747 samples/sec	 accuracy=78.891304	 loss=0.807264	 lr=0.001000
Epoch[042] Batch [0479]/[1203]	Speed: 151.376853 samples/sec	 accuracy=78.885417	 loss=0.807126	 lr=0.001000
Epoch[042] Batch [0499]/[1203]	Speed: 108.031958 samples/sec	 accuracy=78.912000	 loss=0.806678	 lr=0.001000
Epoch[042] Batch [0519]/[1203]	Speed: 169.533253 samples/sec	 accuracy=78.926923	 loss=0.805591	 lr=0.001000
Epoch[042] Batch [0539]/[1203]	Speed: 104.879514 samples/sec	 accuracy=78.939815	 loss=0.805868	 lr=0.001000
Epoch[042] Batch [0559]/[1203]	Speed: 109.381945 samples/sec	 accuracy=78.933036	 loss=0.806195	 lr=0.001000
Epoch[042] Batch [0579]/[1203]	Speed: 172.557852 samples/sec	 accuracy=78.888793	 loss=0.806937	 lr=0.001000
Epoch[042] Batch [0599]/[1203]	Speed: 105.811760 samples/sec	 accuracy=78.914167	 loss=0.805754	 lr=0.001000
Epoch[042] Batch [0619]/[1203]	Speed: 175.380210 samples/sec	 accuracy=78.879839	 loss=0.806210	 lr=0.001000
Epoch[042] Batch [0639]/[1203]	Speed: 103.893350 samples/sec	 accuracy=78.925781	 loss=0.805475	 lr=0.001000
Epoch[042] Batch [0659]/[1203]	Speed: 110.165360 samples/sec	 accuracy=78.911364	 loss=0.805287	 lr=0.001000
Epoch[042] Batch [0679]/[1203]	Speed: 199.132264 samples/sec	 accuracy=78.869853	 loss=0.806658	 lr=0.001000
Epoch[042] Batch [0699]/[1203]	Speed: 99.799020 samples/sec	 accuracy=78.872143	 loss=0.806378	 lr=0.001000
Epoch[042] Batch [0719]/[1203]	Speed: 125.801662 samples/sec	 accuracy=78.884722	 loss=0.806293	 lr=0.001000
Epoch[042] Batch [0739]/[1203]	Speed: 147.485395 samples/sec	 accuracy=78.856757	 loss=0.807252	 lr=0.001000
Epoch[042] Batch [0759]/[1203]	Speed: 101.866907 samples/sec	 accuracy=78.867105	 loss=0.807399	 lr=0.001000
Epoch[042] Batch [0779]/[1203]	Speed: 170.246091 samples/sec	 accuracy=78.871154	 loss=0.806626	 lr=0.001000
Epoch[042] Batch [0799]/[1203]	Speed: 104.052944 samples/sec	 accuracy=78.873750	 loss=0.806529	 lr=0.001000
Epoch[042] Batch [0819]/[1203]	Speed: 121.401330 samples/sec	 accuracy=78.890244	 loss=0.805608	 lr=0.001000
Epoch[042] Batch [0839]/[1203]	Speed: 149.043638 samples/sec	 accuracy=78.872024	 loss=0.806322	 lr=0.001000
Epoch[042] Batch [0859]/[1203]	Speed: 105.121618 samples/sec	 accuracy=78.896512	 loss=0.805809	 lr=0.001000
Epoch[042] Batch [0879]/[1203]	Speed: 145.863953 samples/sec	 accuracy=78.880682	 loss=0.806339	 lr=0.001000
Epoch[042] Batch [0899]/[1203]	Speed: 117.184511 samples/sec	 accuracy=78.870000	 loss=0.806616	 lr=0.001000
Epoch[042] Batch [0919]/[1203]	Speed: 115.586984 samples/sec	 accuracy=78.884239	 loss=0.806170	 lr=0.001000
Epoch[042] Batch [0939]/[1203]	Speed: 160.652733 samples/sec	 accuracy=78.888298	 loss=0.805891	 lr=0.001000
Epoch[042] Batch [0959]/[1203]	Speed: 105.981240 samples/sec	 accuracy=78.896875	 loss=0.805410	 lr=0.001000
Epoch[042] Batch [0979]/[1203]	Speed: 153.418407 samples/sec	 accuracy=78.887245	 loss=0.806102	 lr=0.001000
Epoch[042] Batch [0999]/[1203]	Speed: 113.874737 samples/sec	 accuracy=78.879500	 loss=0.806548	 lr=0.001000
Epoch[042] Batch [1019]/[1203]	Speed: 114.283209 samples/sec	 accuracy=78.874020	 loss=0.806593	 lr=0.001000
Epoch[042] Batch [1039]/[1203]	Speed: 148.841077 samples/sec	 accuracy=78.880288	 loss=0.806191	 lr=0.001000
Epoch[042] Batch [1059]/[1203]	Speed: 108.894526 samples/sec	 accuracy=78.895283	 loss=0.806193	 lr=0.001000
Epoch[042] Batch [1079]/[1203]	Speed: 136.007626 samples/sec	 accuracy=78.887037	 loss=0.806970	 lr=0.001000
Epoch[042] Batch [1099]/[1203]	Speed: 132.745575 samples/sec	 accuracy=78.880909	 loss=0.807131	 lr=0.001000
Epoch[042] Batch [1119]/[1203]	Speed: 103.821969 samples/sec	 accuracy=78.877232	 loss=0.807456	 lr=0.001000
Epoch[042] Batch [1139]/[1203]	Speed: 181.257784 samples/sec	 accuracy=78.879386	 loss=0.807802	 lr=0.001000
Epoch[042] Batch [1159]/[1203]	Speed: 106.281837 samples/sec	 accuracy=78.872414	 loss=0.807697	 lr=0.001000
Epoch[042] Batch [1179]/[1203]	Speed: 114.888435 samples/sec	 accuracy=78.888559	 loss=0.807526	 lr=0.001000
Epoch[042] Batch [1199]/[1203]	Speed: 362.129468 samples/sec	 accuracy=78.900000	 loss=0.807158	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.250000 acc-top5=84.700000
Batch [0039]/[0097]: acc-top1=62.925000 acc-top5=84.500000
Batch [0059]/[0097]: acc-top1=65.416667 acc-top5=85.766667
Batch [0079]/[0097]: acc-top1=65.550000 acc-top5=85.631250
[Epoch 042] training: accuracy=78.899002	 loss=0.807302
[Epoch 042] speed: 124 samples/sec	time cost: 2124.332321
[Epoch 042] validation: acc-top1=65.015464 acc-top5=85.530928 loss=1.527549
Epoch[043] Batch [0019]/[1203]	Speed: 68.941418 samples/sec	 accuracy=78.575000	 loss=0.801125	 lr=0.001000
Epoch[043] Batch [0039]/[1203]	Speed: 99.680110 samples/sec	 accuracy=78.312500	 loss=0.799352	 lr=0.001000
Epoch[043] Batch [0059]/[1203]	Speed: 237.111746 samples/sec	 accuracy=78.816667	 loss=0.790755	 lr=0.001000
Epoch[043] Batch [0079]/[1203]	Speed: 95.572034 samples/sec	 accuracy=79.025000	 loss=0.786111	 lr=0.001000
Epoch[043] Batch [0099]/[1203]	Speed: 113.744500 samples/sec	 accuracy=79.060000	 loss=0.791244	 lr=0.001000
Epoch[043] Batch [0119]/[1203]	Speed: 158.101633 samples/sec	 accuracy=79.162500	 loss=0.790890	 lr=0.001000
Epoch[043] Batch [0139]/[1203]	Speed: 107.941181 samples/sec	 accuracy=79.221429	 loss=0.788806	 lr=0.001000
Epoch[043] Batch [0159]/[1203]	Speed: 173.105572 samples/sec	 accuracy=79.171875	 loss=0.791696	 lr=0.001000
Epoch[043] Batch [0179]/[1203]	Speed: 110.809339 samples/sec	 accuracy=79.180556	 loss=0.792926	 lr=0.001000
Epoch[043] Batch [0199]/[1203]	Speed: 110.307290 samples/sec	 accuracy=79.145000	 loss=0.795188	 lr=0.001000
Epoch[043] Batch [0219]/[1203]	Speed: 150.298282 samples/sec	 accuracy=79.181818	 loss=0.795106	 lr=0.001000
Epoch[043] Batch [0239]/[1203]	Speed: 107.594295 samples/sec	 accuracy=79.189583	 loss=0.794167	 lr=0.001000
Epoch[043] Batch [0259]/[1203]	Speed: 168.555894 samples/sec	 accuracy=79.140385	 loss=0.795335	 lr=0.001000
Epoch[043] Batch [0279]/[1203]	Speed: 110.519138 samples/sec	 accuracy=79.137500	 loss=0.797021	 lr=0.001000
Epoch[043] Batch [0299]/[1203]	Speed: 106.973797 samples/sec	 accuracy=79.130000	 loss=0.798472	 lr=0.001000
Epoch[043] Batch [0319]/[1203]	Speed: 145.223562 samples/sec	 accuracy=79.146875	 loss=0.796739	 lr=0.001000
Epoch[043] Batch [0339]/[1203]	Speed: 122.747986 samples/sec	 accuracy=79.179412	 loss=0.795366	 lr=0.001000
Epoch[043] Batch [0359]/[1203]	Speed: 123.986748 samples/sec	 accuracy=79.198611	 loss=0.795654	 lr=0.001000
Epoch[043] Batch [0379]/[1203]	Speed: 129.714198 samples/sec	 accuracy=79.252632	 loss=0.795534	 lr=0.001000
Epoch[043] Batch [0399]/[1203]	Speed: 112.097791 samples/sec	 accuracy=79.240000	 loss=0.796355	 lr=0.001000
Epoch[043] Batch [0419]/[1203]	Speed: 133.064798 samples/sec	 accuracy=79.261905	 loss=0.795846	 lr=0.001000
Epoch[043] Batch [0439]/[1203]	Speed: 115.655859 samples/sec	 accuracy=79.279545	 loss=0.794752	 lr=0.001000
Epoch[043] Batch [0459]/[1203]	Speed: 155.845929 samples/sec	 accuracy=79.251087	 loss=0.795192	 lr=0.001000
Epoch[043] Batch [0479]/[1203]	Speed: 111.483885 samples/sec	 accuracy=79.277083	 loss=0.794151	 lr=0.001000
Epoch[043] Batch [0499]/[1203]	Speed: 117.585410 samples/sec	 accuracy=79.217000	 loss=0.795586	 lr=0.001000
Epoch[043] Batch [0519]/[1203]	Speed: 133.987970 samples/sec	 accuracy=79.231731	 loss=0.795321	 lr=0.001000
Epoch[043] Batch [0539]/[1203]	Speed: 104.597265 samples/sec	 accuracy=79.254630	 loss=0.794329	 lr=0.001000
Epoch[043] Batch [0559]/[1203]	Speed: 135.734882 samples/sec	 accuracy=79.233929	 loss=0.795425	 lr=0.001000
Epoch[043] Batch [0579]/[1203]	Speed: 125.803237 samples/sec	 accuracy=79.219828	 loss=0.795846	 lr=0.001000
Epoch[043] Batch [0599]/[1203]	Speed: 120.110116 samples/sec	 accuracy=79.201667	 loss=0.796417	 lr=0.001000
Epoch[043] Batch [0619]/[1203]	Speed: 147.540361 samples/sec	 accuracy=79.186290	 loss=0.796850	 lr=0.001000
Epoch[043] Batch [0639]/[1203]	Speed: 111.396490 samples/sec	 accuracy=79.171094	 loss=0.797048	 lr=0.001000
Epoch[043] Batch [0659]/[1203]	Speed: 128.133444 samples/sec	 accuracy=79.171212	 loss=0.797968	 lr=0.001000
Epoch[043] Batch [0679]/[1203]	Speed: 127.473911 samples/sec	 accuracy=79.147059	 loss=0.799168	 lr=0.001000
Epoch[043] Batch [0699]/[1203]	Speed: 120.437651 samples/sec	 accuracy=79.146429	 loss=0.798861	 lr=0.001000
Epoch[043] Batch [0719]/[1203]	Speed: 118.083496 samples/sec	 accuracy=79.152083	 loss=0.798234	 lr=0.001000
Epoch[043] Batch [0739]/[1203]	Speed: 131.453364 samples/sec	 accuracy=79.133784	 loss=0.798917	 lr=0.001000
Epoch[043] Batch [0759]/[1203]	Speed: 122.217010 samples/sec	 accuracy=79.113158	 loss=0.799195	 lr=0.001000
Epoch[043] Batch [0779]/[1203]	Speed: 141.688495 samples/sec	 accuracy=79.136538	 loss=0.799003	 lr=0.001000
Epoch[043] Batch [0799]/[1203]	Speed: 117.088038 samples/sec	 accuracy=79.126250	 loss=0.798768	 lr=0.001000
Epoch[043] Batch [0819]/[1203]	Speed: 121.568308 samples/sec	 accuracy=79.116463	 loss=0.799359	 lr=0.001000
Epoch[043] Batch [0839]/[1203]	Speed: 124.396384 samples/sec	 accuracy=79.141071	 loss=0.798941	 lr=0.001000
Epoch[043] Batch [0859]/[1203]	Speed: 136.163178 samples/sec	 accuracy=79.123837	 loss=0.799378	 lr=0.001000
Epoch[043] Batch [0879]/[1203]	Speed: 130.110809 samples/sec	 accuracy=79.110795	 loss=0.800322	 lr=0.001000
Epoch[043] Batch [0899]/[1203]	Speed: 98.213549 samples/sec	 accuracy=79.128889	 loss=0.799960	 lr=0.001000
Epoch[043] Batch [0919]/[1203]	Speed: 163.934504 samples/sec	 accuracy=79.110326	 loss=0.800130	 lr=0.001000
Epoch[043] Batch [0939]/[1203]	Speed: 111.517208 samples/sec	 accuracy=79.098936	 loss=0.799923	 lr=0.001000
Epoch[043] Batch [0959]/[1203]	Speed: 124.553883 samples/sec	 accuracy=79.120312	 loss=0.799540	 lr=0.001000
Epoch[043] Batch [0979]/[1203]	Speed: 135.713512 samples/sec	 accuracy=79.120918	 loss=0.799596	 lr=0.001000
Epoch[043] Batch [0999]/[1203]	Speed: 105.329070 samples/sec	 accuracy=79.138500	 loss=0.799334	 lr=0.001000
Epoch[043] Batch [1019]/[1203]	Speed: 146.298249 samples/sec	 accuracy=79.114216	 loss=0.799876	 lr=0.001000
Epoch[043] Batch [1039]/[1203]	Speed: 123.702233 samples/sec	 accuracy=79.113942	 loss=0.800124	 lr=0.001000
Epoch[043] Batch [1059]/[1203]	Speed: 116.810900 samples/sec	 accuracy=79.119340	 loss=0.799770	 lr=0.001000
Epoch[043] Batch [1079]/[1203]	Speed: 144.190978 samples/sec	 accuracy=79.115741	 loss=0.799628	 lr=0.001000
Epoch[043] Batch [1099]/[1203]	Speed: 107.828528 samples/sec	 accuracy=79.104091	 loss=0.799690	 lr=0.001000
Epoch[043] Batch [1119]/[1203]	Speed: 150.105569 samples/sec	 accuracy=79.102679	 loss=0.799763	 lr=0.001000
Epoch[043] Batch [1139]/[1203]	Speed: 111.477555 samples/sec	 accuracy=79.110088	 loss=0.799255	 lr=0.001000
Epoch[043] Batch [1159]/[1203]	Speed: 115.002671 samples/sec	 accuracy=79.110345	 loss=0.799367	 lr=0.001000
Epoch[043] Batch [1179]/[1203]	Speed: 170.076634 samples/sec	 accuracy=79.115678	 loss=0.799243	 lr=0.001000
Epoch[043] Batch [1199]/[1203]	Speed: 193.770732 samples/sec	 accuracy=79.123750	 loss=0.798869	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.500000 acc-top5=84.825000
Batch [0039]/[0097]: acc-top1=63.162500 acc-top5=84.762500
Batch [0059]/[0097]: acc-top1=65.341667 acc-top5=85.916667
Batch [0079]/[0097]: acc-top1=65.675000 acc-top5=85.787500
[Epoch 043] training: accuracy=79.127598	 loss=0.798921
[Epoch 043] speed: 124 samples/sec	time cost: 2136.119609
[Epoch 043] validation: acc-top1=65.108247 acc-top5=85.670103 loss=1.519631
Epoch[044] Batch [0019]/[1204]	Speed: 68.837069 samples/sec	 accuracy=79.600000	 loss=0.795722	 lr=0.001000
Epoch[044] Batch [0039]/[1204]	Speed: 102.362880 samples/sec	 accuracy=79.700000	 loss=0.789236	 lr=0.001000
Epoch[044] Batch [0059]/[1204]	Speed: 224.215368 samples/sec	 accuracy=79.291667	 loss=0.795585	 lr=0.001000
Epoch[044] Batch [0079]/[1204]	Speed: 92.218255 samples/sec	 accuracy=79.531250	 loss=0.784839	 lr=0.001000
Epoch[044] Batch [0099]/[1204]	Speed: 115.264264 samples/sec	 accuracy=79.290000	 loss=0.791001	 lr=0.001000
Epoch[044] Batch [0119]/[1204]	Speed: 172.072567 samples/sec	 accuracy=79.204167	 loss=0.796673	 lr=0.001000
Epoch[044] Batch [0139]/[1204]	Speed: 106.109674 samples/sec	 accuracy=79.210714	 loss=0.798150	 lr=0.001000
Epoch[044] Batch [0159]/[1204]	Speed: 201.712239 samples/sec	 accuracy=79.187500	 loss=0.798907	 lr=0.001000
Epoch[044] Batch [0179]/[1204]	Speed: 94.929305 samples/sec	 accuracy=79.166667	 loss=0.800586	 lr=0.001000
Epoch[044] Batch [0199]/[1204]	Speed: 110.920413 samples/sec	 accuracy=79.117500	 loss=0.800190	 lr=0.001000
Epoch[044] Batch [0219]/[1204]	Speed: 170.353984 samples/sec	 accuracy=79.215909	 loss=0.796222	 lr=0.001000
Epoch[044] Batch [0239]/[1204]	Speed: 107.987039 samples/sec	 accuracy=79.264583	 loss=0.793932	 lr=0.001000
Epoch[044] Batch [0259]/[1204]	Speed: 142.081917 samples/sec	 accuracy=79.309615	 loss=0.793922	 lr=0.001000
Epoch[044] Batch [0279]/[1204]	Speed: 113.752225 samples/sec	 accuracy=79.342857	 loss=0.792343	 lr=0.001000
Epoch[044] Batch [0299]/[1204]	Speed: 112.304099 samples/sec	 accuracy=79.355000	 loss=0.792464	 lr=0.001000
Epoch[044] Batch [0319]/[1204]	Speed: 170.175323 samples/sec	 accuracy=79.367188	 loss=0.792745	 lr=0.001000
Epoch[044] Batch [0339]/[1204]	Speed: 107.429169 samples/sec	 accuracy=79.379412	 loss=0.790634	 lr=0.001000
Epoch[044] Batch [0359]/[1204]	Speed: 107.687897 samples/sec	 accuracy=79.411111	 loss=0.790390	 lr=0.001000
Epoch[044] Batch [0379]/[1204]	Speed: 144.608254 samples/sec	 accuracy=79.405263	 loss=0.791115	 lr=0.001000
Epoch[044] Batch [0399]/[1204]	Speed: 117.263374 samples/sec	 accuracy=79.417500	 loss=0.790161	 lr=0.001000
Epoch[044] Batch [0419]/[1204]	Speed: 155.800693 samples/sec	 accuracy=79.398810	 loss=0.789973	 lr=0.001000
Epoch[044] Batch [0439]/[1204]	Speed: 110.726158 samples/sec	 accuracy=79.347727	 loss=0.790753	 lr=0.001000
Epoch[044] Batch [0459]/[1204]	Speed: 110.319257 samples/sec	 accuracy=79.379348	 loss=0.790033	 lr=0.001000
Epoch[044] Batch [0479]/[1204]	Speed: 150.309798 samples/sec	 accuracy=79.353125	 loss=0.789918	 lr=0.001000
Epoch[044] Batch [0499]/[1204]	Speed: 118.984557 samples/sec	 accuracy=79.356000	 loss=0.789892	 lr=0.001000
Epoch[044] Batch [0519]/[1204]	Speed: 137.503960 samples/sec	 accuracy=79.392308	 loss=0.789984	 lr=0.001000
Epoch[044] Batch [0539]/[1204]	Speed: 115.665187 samples/sec	 accuracy=79.436111	 loss=0.787958	 lr=0.001000
Epoch[044] Batch [0559]/[1204]	Speed: 112.587917 samples/sec	 accuracy=79.456250	 loss=0.786808	 lr=0.001000
Epoch[044] Batch [0579]/[1204]	Speed: 150.344742 samples/sec	 accuracy=79.462069	 loss=0.786842	 lr=0.001000
Epoch[044] Batch [0599]/[1204]	Speed: 118.784135 samples/sec	 accuracy=79.494167	 loss=0.786123	 lr=0.001000
Epoch[044] Batch [0619]/[1204]	Speed: 118.922983 samples/sec	 accuracy=79.499194	 loss=0.785385	 lr=0.001000
Epoch[044] Batch [0639]/[1204]	Speed: 123.315738 samples/sec	 accuracy=79.492969	 loss=0.785468	 lr=0.001000
Epoch[044] Batch [0659]/[1204]	Speed: 119.916242 samples/sec	 accuracy=79.484091	 loss=0.785574	 lr=0.001000
Epoch[044] Batch [0679]/[1204]	Speed: 134.210763 samples/sec	 accuracy=79.519118	 loss=0.784348	 lr=0.001000
Epoch[044] Batch [0699]/[1204]	Speed: 123.778025 samples/sec	 accuracy=79.525714	 loss=0.783392	 lr=0.001000
Epoch[044] Batch [0719]/[1204]	Speed: 109.094954 samples/sec	 accuracy=79.547917	 loss=0.782713	 lr=0.001000
Epoch[044] Batch [0739]/[1204]	Speed: 148.663322 samples/sec	 accuracy=79.539865	 loss=0.783002	 lr=0.001000
Epoch[044] Batch [0759]/[1204]	Speed: 120.492956 samples/sec	 accuracy=79.542105	 loss=0.782930	 lr=0.001000
Epoch[044] Batch [0779]/[1204]	Speed: 132.868810 samples/sec	 accuracy=79.558333	 loss=0.782251	 lr=0.001000
Epoch[044] Batch [0799]/[1204]	Speed: 126.140433 samples/sec	 accuracy=79.555000	 loss=0.782527	 lr=0.001000
Epoch[044] Batch [0819]/[1204]	Speed: 102.182743 samples/sec	 accuracy=79.557927	 loss=0.782775	 lr=0.001000
Epoch[044] Batch [0839]/[1204]	Speed: 151.499995 samples/sec	 accuracy=79.548810	 loss=0.782941	 lr=0.001000
Epoch[044] Batch [0859]/[1204]	Speed: 125.041681 samples/sec	 accuracy=79.568605	 loss=0.782966	 lr=0.001000
Epoch[044] Batch [0879]/[1204]	Speed: 104.919610 samples/sec	 accuracy=79.564205	 loss=0.783068	 lr=0.001000
Epoch[044] Batch [0899]/[1204]	Speed: 168.248775 samples/sec	 accuracy=79.552222	 loss=0.783918	 lr=0.001000
Epoch[044] Batch [0919]/[1204]	Speed: 109.629095 samples/sec	 accuracy=79.570109	 loss=0.783145	 lr=0.001000
Epoch[044] Batch [0939]/[1204]	Speed: 143.164129 samples/sec	 accuracy=79.590426	 loss=0.782463	 lr=0.001000
Epoch[044] Batch [0959]/[1204]	Speed: 114.838728 samples/sec	 accuracy=79.589062	 loss=0.782669	 lr=0.001000
Epoch[044] Batch [0979]/[1204]	Speed: 113.205156 samples/sec	 accuracy=79.577551	 loss=0.782862	 lr=0.001000
Epoch[044] Batch [0999]/[1204]	Speed: 146.425910 samples/sec	 accuracy=79.559000	 loss=0.783289	 lr=0.001000
Epoch[044] Batch [1019]/[1204]	Speed: 122.190171 samples/sec	 accuracy=79.566176	 loss=0.782883	 lr=0.001000
Epoch[044] Batch [1039]/[1204]	Speed: 120.288878 samples/sec	 accuracy=79.571154	 loss=0.782922	 lr=0.001000
Epoch[044] Batch [1059]/[1204]	Speed: 117.614809 samples/sec	 accuracy=79.571698	 loss=0.782772	 lr=0.001000
Epoch[044] Batch [1079]/[1204]	Speed: 132.215175 samples/sec	 accuracy=79.590278	 loss=0.782062	 lr=0.001000
Epoch[044] Batch [1099]/[1204]	Speed: 128.913076 samples/sec	 accuracy=79.601364	 loss=0.781866	 lr=0.001000
Epoch[044] Batch [1119]/[1204]	Speed: 128.859681 samples/sec	 accuracy=79.612946	 loss=0.780871	 lr=0.001000
Epoch[044] Batch [1139]/[1204]	Speed: 110.937699 samples/sec	 accuracy=79.602193	 loss=0.781332	 lr=0.001000
Epoch[044] Batch [1159]/[1204]	Speed: 132.250866 samples/sec	 accuracy=79.590086	 loss=0.781541	 lr=0.001000
Epoch[044] Batch [1179]/[1204]	Speed: 137.890769 samples/sec	 accuracy=79.603814	 loss=0.781195	 lr=0.001000
Epoch[044] Batch [1199]/[1204]	Speed: 222.227011 samples/sec	 accuracy=79.572917	 loss=0.781933	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.525000 acc-top5=84.975000
Batch [0039]/[0097]: acc-top1=63.150000 acc-top5=84.487500
Batch [0059]/[0097]: acc-top1=65.300000 acc-top5=85.775000
Batch [0079]/[0097]: acc-top1=65.625000 acc-top5=85.606250
[Epoch 044] training: accuracy=79.566860	 loss=0.782120
[Epoch 044] speed: 124 samples/sec	time cost: 2132.550831
[Epoch 044] validation: acc-top1=65.077320 acc-top5=85.536082 loss=1.522059
Epoch[045] Batch [0019]/[1203]	Speed: 69.461661 samples/sec	 accuracy=80.275000	 loss=0.766096	 lr=0.001000
Epoch[045] Batch [0039]/[1203]	Speed: 101.913372 samples/sec	 accuracy=79.900000	 loss=0.772164	 lr=0.001000
Epoch[045] Batch [0059]/[1203]	Speed: 217.220505 samples/sec	 accuracy=80.000000	 loss=0.766918	 lr=0.001000
Epoch[045] Batch [0079]/[1203]	Speed: 94.863970 samples/sec	 accuracy=80.137500	 loss=0.763004	 lr=0.001000
Epoch[045] Batch [0099]/[1203]	Speed: 110.177623 samples/sec	 accuracy=80.220000	 loss=0.759744	 lr=0.001000
Epoch[045] Batch [0119]/[1203]	Speed: 183.856113 samples/sec	 accuracy=79.958333	 loss=0.767330	 lr=0.001000
Epoch[045] Batch [0139]/[1203]	Speed: 105.083276 samples/sec	 accuracy=79.900000	 loss=0.770343	 lr=0.001000
Epoch[045] Batch [0159]/[1203]	Speed: 185.561635 samples/sec	 accuracy=80.062500	 loss=0.764249	 lr=0.001000
Epoch[045] Batch [0179]/[1203]	Speed: 94.334059 samples/sec	 accuracy=79.994444	 loss=0.766291	 lr=0.001000
Epoch[045] Batch [0199]/[1203]	Speed: 124.993441 samples/sec	 accuracy=79.970000	 loss=0.768005	 lr=0.001000
Epoch[045] Batch [0219]/[1203]	Speed: 152.902035 samples/sec	 accuracy=79.906818	 loss=0.771383	 lr=0.001000
Epoch[045] Batch [0239]/[1203]	Speed: 113.383803 samples/sec	 accuracy=79.895833	 loss=0.770440	 lr=0.001000
Epoch[045] Batch [0259]/[1203]	Speed: 109.565660 samples/sec	 accuracy=79.803846	 loss=0.772288	 lr=0.001000
Epoch[045] Batch [0279]/[1203]	Speed: 141.379810 samples/sec	 accuracy=79.792857	 loss=0.773827	 lr=0.001000
Epoch[045] Batch [0299]/[1203]	Speed: 118.055396 samples/sec	 accuracy=79.795000	 loss=0.773438	 lr=0.001000
Epoch[045] Batch [0319]/[1203]	Speed: 156.749813 samples/sec	 accuracy=79.771875	 loss=0.773685	 lr=0.001000
Epoch[045] Batch [0339]/[1203]	Speed: 116.023745 samples/sec	 accuracy=79.732353	 loss=0.774801	 lr=0.001000
Epoch[045] Batch [0359]/[1203]	Speed: 124.472317 samples/sec	 accuracy=79.704167	 loss=0.776442	 lr=0.001000
Epoch[045] Batch [0379]/[1203]	Speed: 123.337694 samples/sec	 accuracy=79.726316	 loss=0.775600	 lr=0.001000
Epoch[045] Batch [0399]/[1203]	Speed: 112.539409 samples/sec	 accuracy=79.715000	 loss=0.776229	 lr=0.001000
Epoch[045] Batch [0419]/[1203]	Speed: 137.297761 samples/sec	 accuracy=79.728571	 loss=0.775114	 lr=0.001000
Epoch[045] Batch [0439]/[1203]	Speed: 128.213344 samples/sec	 accuracy=79.718182	 loss=0.774734	 lr=0.001000
Epoch[045] Batch [0459]/[1203]	Speed: 105.123099 samples/sec	 accuracy=79.702174	 loss=0.775470	 lr=0.001000
Epoch[045] Batch [0479]/[1203]	Speed: 146.119680 samples/sec	 accuracy=79.716667	 loss=0.774881	 lr=0.001000
Epoch[045] Batch [0499]/[1203]	Speed: 116.356355 samples/sec	 accuracy=79.708000	 loss=0.774587	 lr=0.001000
Epoch[045] Batch [0519]/[1203]	Speed: 144.175175 samples/sec	 accuracy=79.763462	 loss=0.772735	 lr=0.001000
Epoch[045] Batch [0539]/[1203]	Speed: 117.088987 samples/sec	 accuracy=79.750926	 loss=0.772609	 lr=0.001000
Epoch[045] Batch [0559]/[1203]	Speed: 106.381350 samples/sec	 accuracy=79.755357	 loss=0.771492	 lr=0.001000
Epoch[045] Batch [0579]/[1203]	Speed: 148.712674 samples/sec	 accuracy=79.743966	 loss=0.772341	 lr=0.001000
Epoch[045] Batch [0599]/[1203]	Speed: 121.239825 samples/sec	 accuracy=79.796667	 loss=0.770400	 lr=0.001000
Epoch[045] Batch [0619]/[1203]	Speed: 134.440229 samples/sec	 accuracy=79.788710	 loss=0.770528	 lr=0.001000
Epoch[045] Batch [0639]/[1203]	Speed: 113.648778 samples/sec	 accuracy=79.779687	 loss=0.770889	 lr=0.001000
Epoch[045] Batch [0659]/[1203]	Speed: 120.902319 samples/sec	 accuracy=79.778030	 loss=0.770861	 lr=0.001000
Epoch[045] Batch [0679]/[1203]	Speed: 142.457705 samples/sec	 accuracy=79.755147	 loss=0.772074	 lr=0.001000
Epoch[045] Batch [0699]/[1203]	Speed: 123.850361 samples/sec	 accuracy=79.749286	 loss=0.773666	 lr=0.001000
Epoch[045] Batch [0719]/[1203]	Speed: 110.293582 samples/sec	 accuracy=79.767361	 loss=0.773410	 lr=0.001000
Epoch[045] Batch [0739]/[1203]	Speed: 141.090583 samples/sec	 accuracy=79.768243	 loss=0.773224	 lr=0.001000
Epoch[045] Batch [0759]/[1203]	Speed: 115.306856 samples/sec	 accuracy=79.790132	 loss=0.772830	 lr=0.001000
Epoch[045] Batch [0779]/[1203]	Speed: 158.394629 samples/sec	 accuracy=79.790385	 loss=0.773022	 lr=0.001000
Epoch[045] Batch [0799]/[1203]	Speed: 111.755207 samples/sec	 accuracy=79.793125	 loss=0.772760	 lr=0.001000
Epoch[045] Batch [0819]/[1203]	Speed: 110.856599 samples/sec	 accuracy=79.820122	 loss=0.772108	 lr=0.001000
Epoch[045] Batch [0839]/[1203]	Speed: 147.128551 samples/sec	 accuracy=79.807143	 loss=0.772303	 lr=0.001000
Epoch[045] Batch [0859]/[1203]	Speed: 118.967180 samples/sec	 accuracy=79.813372	 loss=0.772115	 lr=0.001000
Epoch[045] Batch [0879]/[1203]	Speed: 113.218166 samples/sec	 accuracy=79.814773	 loss=0.771905	 lr=0.001000
Epoch[045] Batch [0899]/[1203]	Speed: 137.992964 samples/sec	 accuracy=79.823333	 loss=0.771645	 lr=0.001000
Epoch[045] Batch [0919]/[1203]	Speed: 122.793949 samples/sec	 accuracy=79.816848	 loss=0.771912	 lr=0.001000
Epoch[045] Batch [0939]/[1203]	Speed: 143.613709 samples/sec	 accuracy=79.788830	 loss=0.772227	 lr=0.001000
Epoch[045] Batch [0959]/[1203]	Speed: 113.144980 samples/sec	 accuracy=79.794271	 loss=0.772026	 lr=0.001000
Epoch[045] Batch [0979]/[1203]	Speed: 116.383605 samples/sec	 accuracy=79.798980	 loss=0.772312	 lr=0.001000
Epoch[045] Batch [0999]/[1203]	Speed: 150.758849 samples/sec	 accuracy=79.791000	 loss=0.772046	 lr=0.001000
Epoch[045] Batch [1019]/[1203]	Speed: 107.754424 samples/sec	 accuracy=79.790686	 loss=0.772415	 lr=0.001000
Epoch[045] Batch [1039]/[1203]	Speed: 129.685192 samples/sec	 accuracy=79.779327	 loss=0.772844	 lr=0.001000
Epoch[045] Batch [1059]/[1203]	Speed: 128.948352 samples/sec	 accuracy=79.777830	 loss=0.773161	 lr=0.001000
Epoch[045] Batch [1079]/[1203]	Speed: 125.963144 samples/sec	 accuracy=79.780093	 loss=0.773333	 lr=0.001000
Epoch[045] Batch [1099]/[1203]	Speed: 139.404876 samples/sec	 accuracy=79.797727	 loss=0.773209	 lr=0.001000
Epoch[045] Batch [1119]/[1203]	Speed: 106.503607 samples/sec	 accuracy=79.782589	 loss=0.773775	 lr=0.001000
Epoch[045] Batch [1139]/[1203]	Speed: 131.263187 samples/sec	 accuracy=79.770175	 loss=0.774253	 lr=0.001000
Epoch[045] Batch [1159]/[1203]	Speed: 131.501668 samples/sec	 accuracy=79.773276	 loss=0.774237	 lr=0.001000
Epoch[045] Batch [1179]/[1203]	Speed: 124.298390 samples/sec	 accuracy=79.773729	 loss=0.774082	 lr=0.001000
Epoch[045] Batch [1199]/[1203]	Speed: 274.958289 samples/sec	 accuracy=79.761250	 loss=0.774595	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.625000 acc-top5=85.375000
Batch [0039]/[0097]: acc-top1=63.625000 acc-top5=84.825000
Batch [0059]/[0097]: acc-top1=65.800000 acc-top5=85.950000
Batch [0079]/[0097]: acc-top1=66.012500 acc-top5=85.818750
[Epoch 045] training: accuracy=79.760183	 loss=0.774637
[Epoch 045] speed: 124 samples/sec	time cost: 2125.831382
[Epoch 045] validation: acc-top1=65.407216 acc-top5=85.623711 loss=1.524650
Epoch[046] Batch [0019]/[1203]	Speed: 67.950401 samples/sec	 accuracy=79.800000	 loss=0.766538	 lr=0.001000
Epoch[046] Batch [0039]/[1203]	Speed: 105.097994 samples/sec	 accuracy=80.600000	 loss=0.750459	 lr=0.001000
Epoch[046] Batch [0059]/[1203]	Speed: 219.216721 samples/sec	 accuracy=80.375000	 loss=0.751489	 lr=0.001000
Epoch[046] Batch [0079]/[1203]	Speed: 94.556638 samples/sec	 accuracy=80.531250	 loss=0.749572	 lr=0.001000
Epoch[046] Batch [0099]/[1203]	Speed: 114.397054 samples/sec	 accuracy=80.250000	 loss=0.757813	 lr=0.001000
Epoch[046] Batch [0119]/[1203]	Speed: 161.184157 samples/sec	 accuracy=80.237500	 loss=0.761456	 lr=0.001000
Epoch[046] Batch [0139]/[1203]	Speed: 102.455471 samples/sec	 accuracy=80.207143	 loss=0.762378	 lr=0.001000
Epoch[046] Batch [0159]/[1203]	Speed: 209.440315 samples/sec	 accuracy=80.293750	 loss=0.759888	 lr=0.001000
Epoch[046] Batch [0179]/[1203]	Speed: 105.249489 samples/sec	 accuracy=80.255556	 loss=0.762397	 lr=0.001000
Epoch[046] Batch [0199]/[1203]	Speed: 112.575534 samples/sec	 accuracy=80.240000	 loss=0.762666	 lr=0.001000
Epoch[046] Batch [0219]/[1203]	Speed: 154.080527 samples/sec	 accuracy=80.270455	 loss=0.762209	 lr=0.001000
Epoch[046] Batch [0239]/[1203]	Speed: 102.978109 samples/sec	 accuracy=80.181250	 loss=0.763961	 lr=0.001000
Epoch[046] Batch [0259]/[1203]	Speed: 152.407214 samples/sec	 accuracy=80.213462	 loss=0.761724	 lr=0.001000
Epoch[046] Batch [0279]/[1203]	Speed: 125.098468 samples/sec	 accuracy=80.203571	 loss=0.760009	 lr=0.001000
Epoch[046] Batch [0299]/[1203]	Speed: 114.603496 samples/sec	 accuracy=80.200000	 loss=0.758960	 lr=0.001000
Epoch[046] Batch [0319]/[1203]	Speed: 161.843946 samples/sec	 accuracy=80.087500	 loss=0.761875	 lr=0.001000
Epoch[046] Batch [0339]/[1203]	Speed: 98.480698 samples/sec	 accuracy=80.114706	 loss=0.761129	 lr=0.001000
Epoch[046] Batch [0359]/[1203]	Speed: 135.168128 samples/sec	 accuracy=80.129167	 loss=0.762191	 lr=0.001000
Epoch[046] Batch [0379]/[1203]	Speed: 147.696746 samples/sec	 accuracy=80.113158	 loss=0.762861	 lr=0.001000
Epoch[046] Batch [0399]/[1203]	Speed: 95.186693 samples/sec	 accuracy=80.093750	 loss=0.764380	 lr=0.001000
Epoch[046] Batch [0419]/[1203]	Speed: 199.606085 samples/sec	 accuracy=80.089286	 loss=0.764879	 lr=0.001000
Epoch[046] Batch [0439]/[1203]	Speed: 103.919964 samples/sec	 accuracy=80.038636	 loss=0.765824	 lr=0.001000
Epoch[046] Batch [0459]/[1203]	Speed: 115.632982 samples/sec	 accuracy=80.044565	 loss=0.765631	 lr=0.001000
Epoch[046] Batch [0479]/[1203]	Speed: 160.872717 samples/sec	 accuracy=80.059375	 loss=0.764301	 lr=0.001000
Epoch[046] Batch [0499]/[1203]	Speed: 101.491267 samples/sec	 accuracy=80.063000	 loss=0.763892	 lr=0.001000
Epoch[046] Batch [0519]/[1203]	Speed: 147.263593 samples/sec	 accuracy=80.040385	 loss=0.764021	 lr=0.001000
Epoch[046] Batch [0539]/[1203]	Speed: 126.262639 samples/sec	 accuracy=80.055556	 loss=0.763754	 lr=0.001000
Epoch[046] Batch [0559]/[1203]	Speed: 121.319734 samples/sec	 accuracy=80.099107	 loss=0.762574	 lr=0.001000
Epoch[046] Batch [0579]/[1203]	Speed: 141.167530 samples/sec	 accuracy=80.109483	 loss=0.762749	 lr=0.001000
Epoch[046] Batch [0599]/[1203]	Speed: 112.825907 samples/sec	 accuracy=80.112500	 loss=0.763007	 lr=0.001000
Epoch[046] Batch [0619]/[1203]	Speed: 138.786759 samples/sec	 accuracy=80.102419	 loss=0.763162	 lr=0.001000
Epoch[046] Batch [0639]/[1203]	Speed: 127.669100 samples/sec	 accuracy=80.107031	 loss=0.762866	 lr=0.001000
Epoch[046] Batch [0659]/[1203]	Speed: 97.997693 samples/sec	 accuracy=80.075758	 loss=0.763061	 lr=0.001000
Epoch[046] Batch [0679]/[1203]	Speed: 175.697763 samples/sec	 accuracy=80.058088	 loss=0.763204	 lr=0.001000
Epoch[046] Batch [0699]/[1203]	Speed: 114.805624 samples/sec	 accuracy=80.077143	 loss=0.762645	 lr=0.001000
Epoch[046] Batch [0719]/[1203]	Speed: 126.163636 samples/sec	 accuracy=80.062500	 loss=0.763389	 lr=0.001000
Epoch[046] Batch [0739]/[1203]	Speed: 154.830648 samples/sec	 accuracy=80.054054	 loss=0.763435	 lr=0.001000
Epoch[046] Batch [0759]/[1203]	Speed: 95.916009 samples/sec	 accuracy=80.075658	 loss=0.762832	 lr=0.001000
Epoch[046] Batch [0779]/[1203]	Speed: 160.199418 samples/sec	 accuracy=80.094872	 loss=0.762397	 lr=0.001000
Epoch[046] Batch [0799]/[1203]	Speed: 117.731644 samples/sec	 accuracy=80.071875	 loss=0.762745	 lr=0.001000
Epoch[046] Batch [0819]/[1203]	Speed: 132.902266 samples/sec	 accuracy=80.055488	 loss=0.762820	 lr=0.001000
Epoch[046] Batch [0839]/[1203]	Speed: 115.803764 samples/sec	 accuracy=80.052976	 loss=0.763274	 lr=0.001000
Epoch[046] Batch [0859]/[1203]	Speed: 119.308040 samples/sec	 accuracy=80.045349	 loss=0.763967	 lr=0.001000
Epoch[046] Batch [0879]/[1203]	Speed: 152.657947 samples/sec	 accuracy=80.019318	 loss=0.764435	 lr=0.001000
Epoch[046] Batch [0899]/[1203]	Speed: 110.422889 samples/sec	 accuracy=80.027778	 loss=0.764391	 lr=0.001000
Epoch[046] Batch [0919]/[1203]	Speed: 124.441353 samples/sec	 accuracy=80.035870	 loss=0.764103	 lr=0.001000
Epoch[046] Batch [0939]/[1203]	Speed: 134.527951 samples/sec	 accuracy=80.053191	 loss=0.763484	 lr=0.001000
Epoch[046] Batch [0959]/[1203]	Speed: 124.530725 samples/sec	 accuracy=80.055208	 loss=0.763474	 lr=0.001000
Epoch[046] Batch [0979]/[1203]	Speed: 140.701154 samples/sec	 accuracy=80.047449	 loss=0.763510	 lr=0.001000
Epoch[046] Batch [0999]/[1203]	Speed: 106.909350 samples/sec	 accuracy=80.054500	 loss=0.763191	 lr=0.001000
Epoch[046] Batch [1019]/[1203]	Speed: 137.223654 samples/sec	 accuracy=80.043627	 loss=0.763219	 lr=0.001000
Epoch[046] Batch [1039]/[1203]	Speed: 138.607021 samples/sec	 accuracy=80.027404	 loss=0.763797	 lr=0.001000
Epoch[046] Batch [1059]/[1203]	Speed: 106.546873 samples/sec	 accuracy=80.022170	 loss=0.764213	 lr=0.001000
Epoch[046] Batch [1079]/[1203]	Speed: 128.195986 samples/sec	 accuracy=80.013426	 loss=0.764336	 lr=0.001000
Epoch[046] Batch [1099]/[1203]	Speed: 127.202959 samples/sec	 accuracy=80.000909	 loss=0.764458	 lr=0.001000
Epoch[046] Batch [1119]/[1203]	Speed: 119.411204 samples/sec	 accuracy=80.004464	 loss=0.764322	 lr=0.001000
Epoch[046] Batch [1139]/[1203]	Speed: 158.306905 samples/sec	 accuracy=80.009211	 loss=0.764644	 lr=0.001000
Epoch[046] Batch [1159]/[1203]	Speed: 97.851870 samples/sec	 accuracy=80.007759	 loss=0.764821	 lr=0.001000
Epoch[046] Batch [1179]/[1203]	Speed: 168.477696 samples/sec	 accuracy=79.999576	 loss=0.765198	 lr=0.001000
Epoch[046] Batch [1199]/[1203]	Speed: 221.974482 samples/sec	 accuracy=79.997917	 loss=0.765247	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.575000 acc-top5=85.150000
Batch [0039]/[0097]: acc-top1=63.200000 acc-top5=84.937500
Batch [0059]/[0097]: acc-top1=65.500000 acc-top5=86.091667
Batch [0079]/[0097]: acc-top1=65.787500 acc-top5=85.912500
[Epoch 046] training: accuracy=79.999169	 loss=0.765344
[Epoch 046] speed: 125 samples/sec	time cost: 2117.691920
[Epoch 046] validation: acc-top1=65.268041 acc-top5=85.773196 loss=1.520471
Epoch[047] Batch [0019]/[1203]	Speed: 69.390071 samples/sec	 accuracy=80.450000	 loss=0.749254	 lr=0.001000
Epoch[047] Batch [0039]/[1203]	Speed: 103.947839 samples/sec	 accuracy=80.150000	 loss=0.761523	 lr=0.001000
Epoch[047] Batch [0059]/[1203]	Speed: 218.994441 samples/sec	 accuracy=80.425000	 loss=0.754412	 lr=0.001000
Epoch[047] Batch [0079]/[1203]	Speed: 97.612350 samples/sec	 accuracy=80.512500	 loss=0.747024	 lr=0.001000
Epoch[047] Batch [0099]/[1203]	Speed: 102.437626 samples/sec	 accuracy=80.370000	 loss=0.751877	 lr=0.001000
Epoch[047] Batch [0119]/[1203]	Speed: 202.002368 samples/sec	 accuracy=80.379167	 loss=0.750177	 lr=0.001000
Epoch[047] Batch [0139]/[1203]	Speed: 102.901575 samples/sec	 accuracy=80.353571	 loss=0.752149	 lr=0.001000
Epoch[047] Batch [0159]/[1203]	Speed: 195.191980 samples/sec	 accuracy=80.306250	 loss=0.754756	 lr=0.001000
Epoch[047] Batch [0179]/[1203]	Speed: 108.252159 samples/sec	 accuracy=80.369444	 loss=0.751245	 lr=0.001000
Epoch[047] Batch [0199]/[1203]	Speed: 100.484463 samples/sec	 accuracy=80.260000	 loss=0.753427	 lr=0.001000
Epoch[047] Batch [0219]/[1203]	Speed: 166.846174 samples/sec	 accuracy=80.181818	 loss=0.756109	 lr=0.001000
Epoch[047] Batch [0239]/[1203]	Speed: 110.093297 samples/sec	 accuracy=80.214583	 loss=0.754969	 lr=0.001000
Epoch[047] Batch [0259]/[1203]	Speed: 121.405730 samples/sec	 accuracy=80.261538	 loss=0.754165	 lr=0.001000
Epoch[047] Batch [0279]/[1203]	Speed: 141.222070 samples/sec	 accuracy=80.291071	 loss=0.754890	 lr=0.001000
Epoch[047] Batch [0299]/[1203]	Speed: 117.310295 samples/sec	 accuracy=80.251667	 loss=0.754992	 lr=0.001000
Epoch[047] Batch [0319]/[1203]	Speed: 141.497706 samples/sec	 accuracy=80.251562	 loss=0.754398	 lr=0.001000
Epoch[047] Batch [0339]/[1203]	Speed: 113.029857 samples/sec	 accuracy=80.279412	 loss=0.753299	 lr=0.001000
Epoch[047] Batch [0359]/[1203]	Speed: 129.515740 samples/sec	 accuracy=80.262500	 loss=0.753860	 lr=0.001000
Epoch[047] Batch [0379]/[1203]	Speed: 133.481398 samples/sec	 accuracy=80.260526	 loss=0.754265	 lr=0.001000
Epoch[047] Batch [0399]/[1203]	Speed: 114.534278 samples/sec	 accuracy=80.227500	 loss=0.755046	 lr=0.001000
Epoch[047] Batch [0419]/[1203]	Speed: 152.659069 samples/sec	 accuracy=80.251190	 loss=0.755043	 lr=0.001000
Epoch[047] Batch [0439]/[1203]	Speed: 109.771170 samples/sec	 accuracy=80.272727	 loss=0.754764	 lr=0.001000
Epoch[047] Batch [0459]/[1203]	Speed: 127.722714 samples/sec	 accuracy=80.295652	 loss=0.753380	 lr=0.001000
Epoch[047] Batch [0479]/[1203]	Speed: 131.623685 samples/sec	 accuracy=80.329167	 loss=0.753007	 lr=0.001000
Epoch[047] Batch [0499]/[1203]	Speed: 113.575268 samples/sec	 accuracy=80.310000	 loss=0.753292	 lr=0.001000
Epoch[047] Batch [0519]/[1203]	Speed: 150.123390 samples/sec	 accuracy=80.284615	 loss=0.754688	 lr=0.001000
Epoch[047] Batch [0539]/[1203]	Speed: 110.313204 samples/sec	 accuracy=80.282407	 loss=0.754671	 lr=0.001000
Epoch[047] Batch [0559]/[1203]	Speed: 137.319940 samples/sec	 accuracy=80.321429	 loss=0.753638	 lr=0.001000
Epoch[047] Batch [0579]/[1203]	Speed: 127.939858 samples/sec	 accuracy=80.283621	 loss=0.754428	 lr=0.001000
Epoch[047] Batch [0599]/[1203]	Speed: 103.187798 samples/sec	 accuracy=80.268333	 loss=0.755086	 lr=0.001000
Epoch[047] Batch [0619]/[1203]	Speed: 166.115656 samples/sec	 accuracy=80.266935	 loss=0.754646	 lr=0.001000
Epoch[047] Batch [0639]/[1203]	Speed: 106.794693 samples/sec	 accuracy=80.247656	 loss=0.755435	 lr=0.001000
Epoch[047] Batch [0659]/[1203]	Speed: 143.693972 samples/sec	 accuracy=80.231061	 loss=0.756005	 lr=0.001000
Epoch[047] Batch [0679]/[1203]	Speed: 121.628936 samples/sec	 accuracy=80.239706	 loss=0.756897	 lr=0.001000
Epoch[047] Batch [0699]/[1203]	Speed: 106.992946 samples/sec	 accuracy=80.257143	 loss=0.756337	 lr=0.001000
Epoch[047] Batch [0719]/[1203]	Speed: 184.366778 samples/sec	 accuracy=80.273611	 loss=0.756068	 lr=0.001000
Epoch[047] Batch [0739]/[1203]	Speed: 99.739649 samples/sec	 accuracy=80.262838	 loss=0.756675	 lr=0.001000
Epoch[047] Batch [0759]/[1203]	Speed: 137.120860 samples/sec	 accuracy=80.257895	 loss=0.756237	 lr=0.001000
Epoch[047] Batch [0779]/[1203]	Speed: 130.021276 samples/sec	 accuracy=80.243590	 loss=0.756735	 lr=0.001000
Epoch[047] Batch [0799]/[1203]	Speed: 105.012250 samples/sec	 accuracy=80.213750	 loss=0.757374	 lr=0.001000
Epoch[047] Batch [0819]/[1203]	Speed: 210.038249 samples/sec	 accuracy=80.198780	 loss=0.757575	 lr=0.001000
Epoch[047] Batch [0839]/[1203]	Speed: 96.463784 samples/sec	 accuracy=80.213095	 loss=0.757103	 lr=0.001000
Epoch[047] Batch [0859]/[1203]	Speed: 119.695626 samples/sec	 accuracy=80.211047	 loss=0.757092	 lr=0.001000
Epoch[047] Batch [0879]/[1203]	Speed: 163.672661 samples/sec	 accuracy=80.225568	 loss=0.756710	 lr=0.001000
Epoch[047] Batch [0899]/[1203]	Speed: 97.073007 samples/sec	 accuracy=80.261111	 loss=0.755528	 lr=0.001000
Epoch[047] Batch [0919]/[1203]	Speed: 138.831542 samples/sec	 accuracy=80.267391	 loss=0.755384	 lr=0.001000
Epoch[047] Batch [0939]/[1203]	Speed: 131.136794 samples/sec	 accuracy=80.248404	 loss=0.755682	 lr=0.001000
Epoch[047] Batch [0959]/[1203]	Speed: 122.518568 samples/sec	 accuracy=80.240104	 loss=0.755854	 lr=0.001000
Epoch[047] Batch [0979]/[1203]	Speed: 159.704906 samples/sec	 accuracy=80.230102	 loss=0.756490	 lr=0.001000
Epoch[047] Batch [0999]/[1203]	Speed: 95.247764 samples/sec	 accuracy=80.215000	 loss=0.756654	 lr=0.001000
Epoch[047] Batch [1019]/[1203]	Speed: 143.116859 samples/sec	 accuracy=80.212745	 loss=0.756518	 lr=0.001000
Epoch[047] Batch [1039]/[1203]	Speed: 131.929748 samples/sec	 accuracy=80.202404	 loss=0.756733	 lr=0.001000
Epoch[047] Batch [1059]/[1203]	Speed: 118.379300 samples/sec	 accuracy=80.197642	 loss=0.757408	 lr=0.001000
Epoch[047] Batch [1079]/[1203]	Speed: 135.005794 samples/sec	 accuracy=80.186574	 loss=0.757915	 lr=0.001000
Epoch[047] Batch [1099]/[1203]	Speed: 111.279208 samples/sec	 accuracy=80.180909	 loss=0.758514	 lr=0.001000
Epoch[047] Batch [1119]/[1203]	Speed: 140.866373 samples/sec	 accuracy=80.168750	 loss=0.758724	 lr=0.001000
Epoch[047] Batch [1139]/[1203]	Speed: 132.742002 samples/sec	 accuracy=80.164474	 loss=0.759198	 lr=0.001000
Epoch[047] Batch [1159]/[1203]	Speed: 103.450184 samples/sec	 accuracy=80.178879	 loss=0.758890	 lr=0.001000
Epoch[047] Batch [1179]/[1203]	Speed: 157.775770 samples/sec	 accuracy=80.185169	 loss=0.758844	 lr=0.001000
Epoch[047] Batch [1199]/[1203]	Speed: 226.993343 samples/sec	 accuracy=80.187500	 loss=0.759275	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.825000 acc-top5=85.150000
Batch [0039]/[0097]: acc-top1=63.212500 acc-top5=84.662500
Batch [0059]/[0097]: acc-top1=65.466667 acc-top5=85.825000
Batch [0079]/[0097]: acc-top1=65.712500 acc-top5=85.668750
[Epoch 047] training: accuracy=80.189111	 loss=0.759174
[Epoch 047] speed: 124 samples/sec	time cost: 2118.992451
[Epoch 047] validation: acc-top1=65.195876 acc-top5=85.639175 loss=1.524099
Epoch[048] Batch [0019]/[1203]	Speed: 68.057859 samples/sec	 accuracy=80.075000	 loss=0.766736	 lr=0.001000
Epoch[048] Batch [0039]/[1203]	Speed: 107.608756 samples/sec	 accuracy=79.800000	 loss=0.757282	 lr=0.001000
Epoch[048] Batch [0059]/[1203]	Speed: 198.809752 samples/sec	 accuracy=79.983333	 loss=0.750794	 lr=0.001000
Epoch[048] Batch [0079]/[1203]	Speed: 98.745837 samples/sec	 accuracy=80.000000	 loss=0.751535	 lr=0.001000
Epoch[048] Batch [0099]/[1203]	Speed: 133.446730 samples/sec	 accuracy=80.210000	 loss=0.744909	 lr=0.001000
Epoch[048] Batch [0119]/[1203]	Speed: 141.115172 samples/sec	 accuracy=80.162500	 loss=0.744115	 lr=0.001000
Epoch[048] Batch [0139]/[1203]	Speed: 102.490554 samples/sec	 accuracy=80.167857	 loss=0.745931	 lr=0.001000
Epoch[048] Batch [0159]/[1203]	Speed: 213.350045 samples/sec	 accuracy=80.200000	 loss=0.745022	 lr=0.001000
Epoch[048] Batch [0179]/[1203]	Speed: 94.152775 samples/sec	 accuracy=80.188889	 loss=0.744554	 lr=0.001000
Epoch[048] Batch [0199]/[1203]	Speed: 121.250030 samples/sec	 accuracy=80.235000	 loss=0.743366	 lr=0.001000
Epoch[048] Batch [0219]/[1203]	Speed: 166.509231 samples/sec	 accuracy=80.254545	 loss=0.742248	 lr=0.001000
Epoch[048] Batch [0239]/[1203]	Speed: 114.398978 samples/sec	 accuracy=80.218750	 loss=0.743548	 lr=0.001000
Epoch[048] Batch [0259]/[1203]	Speed: 173.095949 samples/sec	 accuracy=80.278846	 loss=0.742049	 lr=0.001000
Epoch[048] Batch [0279]/[1203]	Speed: 96.683502 samples/sec	 accuracy=80.242857	 loss=0.742801	 lr=0.001000
Epoch[048] Batch [0299]/[1203]	Speed: 141.315422 samples/sec	 accuracy=80.221667	 loss=0.744515	 lr=0.001000
Epoch[048] Batch [0319]/[1203]	Speed: 126.025924 samples/sec	 accuracy=80.248438	 loss=0.745077	 lr=0.001000
Epoch[048] Batch [0339]/[1203]	Speed: 116.055144 samples/sec	 accuracy=80.230882	 loss=0.745423	 lr=0.001000
Epoch[048] Batch [0359]/[1203]	Speed: 157.210225 samples/sec	 accuracy=80.212500	 loss=0.746235	 lr=0.001000
Epoch[048] Batch [0379]/[1203]	Speed: 110.460509 samples/sec	 accuracy=80.194737	 loss=0.746355	 lr=0.001000
Epoch[048] Batch [0399]/[1203]	Speed: 119.388135 samples/sec	 accuracy=80.216250	 loss=0.746406	 lr=0.001000
Epoch[048] Batch [0419]/[1203]	Speed: 138.245803 samples/sec	 accuracy=80.230952	 loss=0.747081	 lr=0.001000
Epoch[048] Batch [0439]/[1203]	Speed: 120.138017 samples/sec	 accuracy=80.289773	 loss=0.745569	 lr=0.001000
Epoch[048] Batch [0459]/[1203]	Speed: 144.329041 samples/sec	 accuracy=80.264130	 loss=0.746299	 lr=0.001000
Epoch[048] Batch [0479]/[1203]	Speed: 120.364060 samples/sec	 accuracy=80.280208	 loss=0.746853	 lr=0.001000
Epoch[048] Batch [0499]/[1203]	Speed: 114.036640 samples/sec	 accuracy=80.304000	 loss=0.747782	 lr=0.001000
Epoch[048] Batch [0519]/[1203]	Speed: 141.104281 samples/sec	 accuracy=80.299038	 loss=0.748030	 lr=0.001000
Epoch[048] Batch [0539]/[1203]	Speed: 115.863351 samples/sec	 accuracy=80.306481	 loss=0.747620	 lr=0.001000
Epoch[048] Batch [0559]/[1203]	Speed: 132.168712 samples/sec	 accuracy=80.316071	 loss=0.748244	 lr=0.001000
Epoch[048] Batch [0579]/[1203]	Speed: 124.253825 samples/sec	 accuracy=80.324138	 loss=0.747986	 lr=0.001000
Epoch[048] Batch [0599]/[1203]	Speed: 126.344531 samples/sec	 accuracy=80.331667	 loss=0.748308	 lr=0.001000
Epoch[048] Batch [0619]/[1203]	Speed: 134.086050 samples/sec	 accuracy=80.322581	 loss=0.748590	 lr=0.001000
Epoch[048] Batch [0639]/[1203]	Speed: 126.880653 samples/sec	 accuracy=80.310156	 loss=0.749025	 lr=0.001000
Epoch[048] Batch [0659]/[1203]	Speed: 118.175527 samples/sec	 accuracy=80.295455	 loss=0.749565	 lr=0.001000
Epoch[048] Batch [0679]/[1203]	Speed: 133.458966 samples/sec	 accuracy=80.283824	 loss=0.749820	 lr=0.001000
Epoch[048] Batch [0699]/[1203]	Speed: 118.965094 samples/sec	 accuracy=80.282857	 loss=0.749514	 lr=0.001000
Epoch[048] Batch [0719]/[1203]	Speed: 133.244457 samples/sec	 accuracy=80.260417	 loss=0.750200	 lr=0.001000
Epoch[048] Batch [0739]/[1203]	Speed: 124.095987 samples/sec	 accuracy=80.276351	 loss=0.749846	 lr=0.001000
Epoch[048] Batch [0759]/[1203]	Speed: 126.481492 samples/sec	 accuracy=80.277632	 loss=0.749843	 lr=0.001000
Epoch[048] Batch [0779]/[1203]	Speed: 116.720439 samples/sec	 accuracy=80.299359	 loss=0.749428	 lr=0.001000
Epoch[048] Batch [0799]/[1203]	Speed: 131.842613 samples/sec	 accuracy=80.295625	 loss=0.749640	 lr=0.001000
Epoch[048] Batch [0819]/[1203]	Speed: 128.166082 samples/sec	 accuracy=80.297561	 loss=0.749392	 lr=0.001000
Epoch[048] Batch [0839]/[1203]	Speed: 122.565723 samples/sec	 accuracy=80.284524	 loss=0.749758	 lr=0.001000
Epoch[048] Batch [0859]/[1203]	Speed: 135.844141 samples/sec	 accuracy=80.301163	 loss=0.749810	 lr=0.001000
Epoch[048] Batch [0879]/[1203]	Speed: 120.713665 samples/sec	 accuracy=80.318750	 loss=0.749588	 lr=0.001000
Epoch[048] Batch [0899]/[1203]	Speed: 114.129096 samples/sec	 accuracy=80.329444	 loss=0.749453	 lr=0.001000
Epoch[048] Batch [0919]/[1203]	Speed: 159.438263 samples/sec	 accuracy=80.329891	 loss=0.749390	 lr=0.001000
Epoch[048] Batch [0939]/[1203]	Speed: 103.947653 samples/sec	 accuracy=80.345213	 loss=0.749374	 lr=0.001000
Epoch[048] Batch [0959]/[1203]	Speed: 134.073120 samples/sec	 accuracy=80.343750	 loss=0.749643	 lr=0.001000
Epoch[048] Batch [0979]/[1203]	Speed: 141.804003 samples/sec	 accuracy=80.331122	 loss=0.749458	 lr=0.001000
Epoch[048] Batch [0999]/[1203]	Speed: 113.273482 samples/sec	 accuracy=80.333000	 loss=0.749431	 lr=0.001000
Epoch[048] Batch [1019]/[1203]	Speed: 129.098203 samples/sec	 accuracy=80.331863	 loss=0.749663	 lr=0.001000
Epoch[048] Batch [1039]/[1203]	Speed: 118.271692 samples/sec	 accuracy=80.344712	 loss=0.749160	 lr=0.001000
Epoch[048] Batch [1059]/[1203]	Speed: 133.792155 samples/sec	 accuracy=80.345283	 loss=0.749155	 lr=0.001000
Epoch[048] Batch [1079]/[1203]	Speed: 148.071584 samples/sec	 accuracy=80.355556	 loss=0.748535	 lr=0.001000
Epoch[048] Batch [1099]/[1203]	Speed: 110.619254 samples/sec	 accuracy=80.365000	 loss=0.748562	 lr=0.001000
Epoch[048] Batch [1119]/[1203]	Speed: 131.553219 samples/sec	 accuracy=80.357143	 loss=0.748616	 lr=0.001000
Epoch[048] Batch [1139]/[1203]	Speed: 121.631438 samples/sec	 accuracy=80.347368	 loss=0.749055	 lr=0.001000
Epoch[048] Batch [1159]/[1203]	Speed: 131.725423 samples/sec	 accuracy=80.325000	 loss=0.749839	 lr=0.001000
Epoch[048] Batch [1179]/[1203]	Speed: 134.976233 samples/sec	 accuracy=80.329237	 loss=0.749547	 lr=0.001000
Epoch[048] Batch [1199]/[1203]	Speed: 239.339183 samples/sec	 accuracy=80.320833	 loss=0.750024	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.700000 acc-top5=85.400000
Batch [0039]/[0097]: acc-top1=63.300000 acc-top5=84.925000
Batch [0059]/[0097]: acc-top1=65.550000 acc-top5=86.041667
Batch [0079]/[0097]: acc-top1=65.743750 acc-top5=85.906250
[Epoch 048] training: accuracy=80.324190	 loss=0.750079
[Epoch 048] speed: 125 samples/sec	time cost: 2104.048466
[Epoch 048] validation: acc-top1=65.268041 acc-top5=85.773196 loss=1.523882
Epoch[049] Batch [0019]/[1203]	Speed: 69.608574 samples/sec	 accuracy=81.975000	 loss=0.712717	 lr=0.001000
Epoch[049] Batch [0039]/[1203]	Speed: 105.362344 samples/sec	 accuracy=81.375000	 loss=0.720559	 lr=0.001000
Epoch[049] Batch [0059]/[1203]	Speed: 208.856247 samples/sec	 accuracy=81.316667	 loss=0.720330	 lr=0.001000
Epoch[049] Batch [0079]/[1203]	Speed: 95.355678 samples/sec	 accuracy=81.193750	 loss=0.725197	 lr=0.001000
Epoch[049] Batch [0099]/[1203]	Speed: 116.148233 samples/sec	 accuracy=81.120000	 loss=0.728726	 lr=0.001000
Epoch[049] Batch [0119]/[1203]	Speed: 171.280058 samples/sec	 accuracy=80.962500	 loss=0.732869	 lr=0.001000
Epoch[049] Batch [0139]/[1203]	Speed: 99.832901 samples/sec	 accuracy=80.882143	 loss=0.735836	 lr=0.001000
Epoch[049] Batch [0159]/[1203]	Speed: 239.445013 samples/sec	 accuracy=80.800000	 loss=0.738041	 lr=0.001000
Epoch[049] Batch [0179]/[1203]	Speed: 90.577917 samples/sec	 accuracy=80.863889	 loss=0.736994	 lr=0.001000
Epoch[049] Batch [0199]/[1203]	Speed: 124.559061 samples/sec	 accuracy=80.900000	 loss=0.735455	 lr=0.001000
Epoch[049] Batch [0219]/[1203]	Speed: 151.070710 samples/sec	 accuracy=80.981818	 loss=0.731005	 lr=0.001000
Epoch[049] Batch [0239]/[1203]	Speed: 111.781845 samples/sec	 accuracy=80.895833	 loss=0.732951	 lr=0.001000
Epoch[049] Batch [0259]/[1203]	Speed: 139.198312 samples/sec	 accuracy=80.873077	 loss=0.734507	 lr=0.001000
Epoch[049] Batch [0279]/[1203]	Speed: 113.888677 samples/sec	 accuracy=80.882143	 loss=0.735096	 lr=0.001000
Epoch[049] Batch [0299]/[1203]	Speed: 123.438322 samples/sec	 accuracy=80.858333	 loss=0.735091	 lr=0.001000
Epoch[049] Batch [0319]/[1203]	Speed: 147.911122 samples/sec	 accuracy=80.820312	 loss=0.735840	 lr=0.001000
Epoch[049] Batch [0339]/[1203]	Speed: 111.294517 samples/sec	 accuracy=80.842647	 loss=0.735079	 lr=0.001000
Epoch[049] Batch [0359]/[1203]	Speed: 146.267048 samples/sec	 accuracy=80.844444	 loss=0.734388	 lr=0.001000
Epoch[049] Batch [0379]/[1203]	Speed: 105.010729 samples/sec	 accuracy=80.865789	 loss=0.734600	 lr=0.001000
Epoch[049] Batch [0399]/[1203]	Speed: 127.502497 samples/sec	 accuracy=80.827500	 loss=0.734931	 lr=0.001000
Epoch[049] Batch [0419]/[1203]	Speed: 150.143319 samples/sec	 accuracy=80.811905	 loss=0.735599	 lr=0.001000
Epoch[049] Batch [0439]/[1203]	Speed: 107.007990 samples/sec	 accuracy=80.765909	 loss=0.736227	 lr=0.001000
Epoch[049] Batch [0459]/[1203]	Speed: 133.702825 samples/sec	 accuracy=80.753261	 loss=0.736459	 lr=0.001000
Epoch[049] Batch [0479]/[1203]	Speed: 130.394586 samples/sec	 accuracy=80.721875	 loss=0.737166	 lr=0.001000
Epoch[049] Batch [0499]/[1203]	Speed: 128.440011 samples/sec	 accuracy=80.671000	 loss=0.737959	 lr=0.001000
Epoch[049] Batch [0519]/[1203]	Speed: 138.072314 samples/sec	 accuracy=80.662500	 loss=0.738715	 lr=0.001000
Epoch[049] Batch [0539]/[1203]	Speed: 106.771197 samples/sec	 accuracy=80.661111	 loss=0.738866	 lr=0.001000
Epoch[049] Batch [0559]/[1203]	Speed: 135.430511 samples/sec	 accuracy=80.683929	 loss=0.737738	 lr=0.001000
Epoch[049] Batch [0579]/[1203]	Speed: 120.653096 samples/sec	 accuracy=80.685345	 loss=0.738137	 lr=0.001000
Epoch[049] Batch [0599]/[1203]	Speed: 134.065736 samples/sec	 accuracy=80.670833	 loss=0.739023	 lr=0.001000
Epoch[049] Batch [0619]/[1203]	Speed: 139.993465 samples/sec	 accuracy=80.631452	 loss=0.740462	 lr=0.001000
Epoch[049] Batch [0639]/[1203]	Speed: 112.403440 samples/sec	 accuracy=80.632031	 loss=0.740132	 lr=0.001000
Epoch[049] Batch [0659]/[1203]	Speed: 134.728301 samples/sec	 accuracy=80.610606	 loss=0.741127	 lr=0.001000
Epoch[049] Batch [0679]/[1203]	Speed: 124.086247 samples/sec	 accuracy=80.610294	 loss=0.741942	 lr=0.001000
Epoch[049] Batch [0699]/[1203]	Speed: 119.390849 samples/sec	 accuracy=80.588571	 loss=0.742063	 lr=0.001000
Epoch[049] Batch [0719]/[1203]	Speed: 144.990097 samples/sec	 accuracy=80.604861	 loss=0.742014	 lr=0.001000
Epoch[049] Batch [0739]/[1203]	Speed: 120.579050 samples/sec	 accuracy=80.604054	 loss=0.741996	 lr=0.001000
Epoch[049] Batch [0759]/[1203]	Speed: 127.739552 samples/sec	 accuracy=80.596711	 loss=0.742762	 lr=0.001000
Epoch[049] Batch [0779]/[1203]	Speed: 115.949266 samples/sec	 accuracy=80.597436	 loss=0.742291	 lr=0.001000
Epoch[049] Batch [0799]/[1203]	Speed: 129.885575 samples/sec	 accuracy=80.598125	 loss=0.742322	 lr=0.001000
Epoch[049] Batch [0819]/[1203]	Speed: 149.519493 samples/sec	 accuracy=80.643293	 loss=0.741065	 lr=0.001000
Epoch[049] Batch [0839]/[1203]	Speed: 121.260774 samples/sec	 accuracy=80.628571	 loss=0.741777	 lr=0.001000
Epoch[049] Batch [0859]/[1203]	Speed: 111.056120 samples/sec	 accuracy=80.640698	 loss=0.741447	 lr=0.001000
Epoch[049] Batch [0879]/[1203]	Speed: 118.813894 samples/sec	 accuracy=80.644886	 loss=0.741325	 lr=0.001000
Epoch[049] Batch [0899]/[1203]	Speed: 149.190857 samples/sec	 accuracy=80.649444	 loss=0.741290	 lr=0.001000
Epoch[049] Batch [0919]/[1203]	Speed: 130.884951 samples/sec	 accuracy=80.657065	 loss=0.740950	 lr=0.001000
Epoch[049] Batch [0939]/[1203]	Speed: 122.919763 samples/sec	 accuracy=80.659574	 loss=0.741334	 lr=0.001000
Epoch[049] Batch [0959]/[1203]	Speed: 116.196385 samples/sec	 accuracy=80.674479	 loss=0.740735	 lr=0.001000
Epoch[049] Batch [0979]/[1203]	Speed: 118.543284 samples/sec	 accuracy=80.675000	 loss=0.740811	 lr=0.001000
Epoch[049] Batch [0999]/[1203]	Speed: 139.546327 samples/sec	 accuracy=80.653000	 loss=0.741074	 lr=0.001000
Epoch[049] Batch [1019]/[1203]	Speed: 115.041146 samples/sec	 accuracy=80.644118	 loss=0.741102	 lr=0.001000
Epoch[049] Batch [1039]/[1203]	Speed: 134.358081 samples/sec	 accuracy=80.635577	 loss=0.741127	 lr=0.001000
Epoch[049] Batch [1059]/[1203]	Speed: 115.840185 samples/sec	 accuracy=80.599528	 loss=0.742197	 lr=0.001000
Epoch[049] Batch [1079]/[1203]	Speed: 138.045116 samples/sec	 accuracy=80.617130	 loss=0.742110	 lr=0.001000
Epoch[049] Batch [1099]/[1203]	Speed: 137.238179 samples/sec	 accuracy=80.612727	 loss=0.742102	 lr=0.001000
Epoch[049] Batch [1119]/[1203]	Speed: 117.189945 samples/sec	 accuracy=80.593750	 loss=0.742577	 lr=0.001000
Epoch[049] Batch [1139]/[1203]	Speed: 131.265797 samples/sec	 accuracy=80.609649	 loss=0.742281	 lr=0.001000
Epoch[049] Batch [1159]/[1203]	Speed: 122.360357 samples/sec	 accuracy=80.610776	 loss=0.742218	 lr=0.001000
Epoch[049] Batch [1179]/[1203]	Speed: 133.937100 samples/sec	 accuracy=80.612288	 loss=0.742236	 lr=0.001000
Epoch[049] Batch [1199]/[1203]	Speed: 252.383137 samples/sec	 accuracy=80.618750	 loss=0.741903	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.350000 acc-top5=85.075000
Batch [0039]/[0097]: acc-top1=63.325000 acc-top5=84.612500
Batch [0059]/[0097]: acc-top1=65.458333 acc-top5=85.850000
Batch [0079]/[0097]: acc-top1=65.650000 acc-top5=85.743750
[Epoch 049] training: accuracy=80.614298	 loss=0.742005
[Epoch 049] speed: 125 samples/sec	time cost: 2113.911220
[Epoch 049] validation: acc-top1=65.118557 acc-top5=85.711340 loss=1.527386
Epoch[050] Batch [0019]/[1203]	Speed: 69.578365 samples/sec	 accuracy=81.225000	 loss=0.736727	 lr=0.001000
Epoch[050] Batch [0039]/[1203]	Speed: 99.670444 samples/sec	 accuracy=80.775000	 loss=0.737301	 lr=0.001000
Epoch[050] Batch [0059]/[1203]	Speed: 216.602622 samples/sec	 accuracy=80.458333	 loss=0.741884	 lr=0.001000
Epoch[050] Batch [0079]/[1203]	Speed: 98.039668 samples/sec	 accuracy=80.762500	 loss=0.735968	 lr=0.001000
Epoch[050] Batch [0099]/[1203]	Speed: 108.791938 samples/sec	 accuracy=81.075000	 loss=0.726909	 lr=0.001000
Epoch[050] Batch [0119]/[1203]	Speed: 197.020839 samples/sec	 accuracy=80.775000	 loss=0.733770	 lr=0.001000
Epoch[050] Batch [0139]/[1203]	Speed: 108.934492 samples/sec	 accuracy=80.842857	 loss=0.728408	 lr=0.001000
Epoch[050] Batch [0159]/[1203]	Speed: 190.574024 samples/sec	 accuracy=80.865625	 loss=0.729069	 lr=0.001000
Epoch[050] Batch [0179]/[1203]	Speed: 98.326203 samples/sec	 accuracy=80.966667	 loss=0.724929	 lr=0.001000
Epoch[050] Batch [0199]/[1203]	Speed: 103.300797 samples/sec	 accuracy=80.952500	 loss=0.724888	 lr=0.001000
Epoch[050] Batch [0219]/[1203]	Speed: 188.655366 samples/sec	 accuracy=81.045455	 loss=0.724713	 lr=0.001000
Epoch[050] Batch [0239]/[1203]	Speed: 112.053229 samples/sec	 accuracy=80.987500	 loss=0.724950	 lr=0.001000
Epoch[050] Batch [0259]/[1203]	Speed: 106.265533 samples/sec	 accuracy=80.928846	 loss=0.725923	 lr=0.001000
Epoch[050] Batch [0279]/[1203]	Speed: 181.176269 samples/sec	 accuracy=80.957143	 loss=0.725537	 lr=0.001000
Epoch[050] Batch [0299]/[1203]	Speed: 98.550168 samples/sec	 accuracy=80.971667	 loss=0.726465	 lr=0.001000
Epoch[050] Batch [0319]/[1203]	Speed: 193.701147 samples/sec	 accuracy=80.907812	 loss=0.728043	 lr=0.001000
Epoch[050] Batch [0339]/[1203]	Speed: 100.693352 samples/sec	 accuracy=80.914706	 loss=0.728656	 lr=0.001000
Epoch[050] Batch [0359]/[1203]	Speed: 125.810870 samples/sec	 accuracy=80.995833	 loss=0.726445	 lr=0.001000
Epoch[050] Batch [0379]/[1203]	Speed: 152.769156 samples/sec	 accuracy=80.998684	 loss=0.726982	 lr=0.001000
Epoch[050] Batch [0399]/[1203]	Speed: 101.720382 samples/sec	 accuracy=80.961250	 loss=0.727993	 lr=0.001000
Epoch[050] Batch [0419]/[1203]	Speed: 128.265908 samples/sec	 accuracy=80.960714	 loss=0.727710	 lr=0.001000
Epoch[050] Batch [0439]/[1203]	Speed: 134.157196 samples/sec	 accuracy=80.987500	 loss=0.726400	 lr=0.001000
Epoch[050] Batch [0459]/[1203]	Speed: 119.455164 samples/sec	 accuracy=80.994565	 loss=0.725988	 lr=0.001000
Epoch[050] Batch [0479]/[1203]	Speed: 165.390704 samples/sec	 accuracy=80.966667	 loss=0.726281	 lr=0.001000
Epoch[050] Batch [0499]/[1203]	Speed: 92.016978 samples/sec	 accuracy=80.962000	 loss=0.726023	 lr=0.001000
Epoch[050] Batch [0519]/[1203]	Speed: 160.561158 samples/sec	 accuracy=80.928846	 loss=0.727204	 lr=0.001000
Epoch[050] Batch [0539]/[1203]	Speed: 125.152578 samples/sec	 accuracy=80.948148	 loss=0.726354	 lr=0.001000
Epoch[050] Batch [0559]/[1203]	Speed: 133.533314 samples/sec	 accuracy=80.944643	 loss=0.726316	 lr=0.001000
Epoch[050] Batch [0579]/[1203]	Speed: 131.506750 samples/sec	 accuracy=80.934483	 loss=0.726307	 lr=0.001000
Epoch[050] Batch [0599]/[1203]	Speed: 102.085890 samples/sec	 accuracy=80.910000	 loss=0.726704	 lr=0.001000
Epoch[050] Batch [0619]/[1203]	Speed: 154.239514 samples/sec	 accuracy=80.904032	 loss=0.726933	 lr=0.001000
Epoch[050] Batch [0639]/[1203]	Speed: 116.640960 samples/sec	 accuracy=80.885156	 loss=0.727845	 lr=0.001000
Epoch[050] Batch [0659]/[1203]	Speed: 110.297038 samples/sec	 accuracy=80.904545	 loss=0.727487	 lr=0.001000
Epoch[050] Batch [0679]/[1203]	Speed: 192.021884 samples/sec	 accuracy=80.877206	 loss=0.728574	 lr=0.001000
Epoch[050] Batch [0699]/[1203]	Speed: 93.290673 samples/sec	 accuracy=80.867857	 loss=0.728560	 lr=0.001000
Epoch[050] Batch [0719]/[1203]	Speed: 168.737456 samples/sec	 accuracy=80.867361	 loss=0.729388	 lr=0.001000
Epoch[050] Batch [0739]/[1203]	Speed: 119.799232 samples/sec	 accuracy=80.872973	 loss=0.729452	 lr=0.001000
Epoch[050] Batch [0759]/[1203]	Speed: 114.905978 samples/sec	 accuracy=80.858553	 loss=0.729882	 lr=0.001000
Epoch[050] Batch [0779]/[1203]	Speed: 160.034489 samples/sec	 accuracy=80.891667	 loss=0.729891	 lr=0.001000
Epoch[050] Batch [0799]/[1203]	Speed: 102.886025 samples/sec	 accuracy=80.872500	 loss=0.730403	 lr=0.001000
Epoch[050] Batch [0819]/[1203]	Speed: 127.895170 samples/sec	 accuracy=80.856098	 loss=0.730744	 lr=0.001000
Epoch[050] Batch [0839]/[1203]	Speed: 142.420076 samples/sec	 accuracy=80.848214	 loss=0.731052	 lr=0.001000
Epoch[050] Batch [0859]/[1203]	Speed: 105.676229 samples/sec	 accuracy=80.834302	 loss=0.731142	 lr=0.001000
Epoch[050] Batch [0879]/[1203]	Speed: 155.689724 samples/sec	 accuracy=80.823295	 loss=0.731585	 lr=0.001000
Epoch[050] Batch [0899]/[1203]	Speed: 107.715847 samples/sec	 accuracy=80.814444	 loss=0.731983	 lr=0.001000
Epoch[050] Batch [0919]/[1203]	Speed: 138.402480 samples/sec	 accuracy=80.829891	 loss=0.731923	 lr=0.001000
Epoch[050] Batch [0939]/[1203]	Speed: 145.735259 samples/sec	 accuracy=80.829255	 loss=0.732312	 lr=0.001000
Epoch[050] Batch [0959]/[1203]	Speed: 101.814837 samples/sec	 accuracy=80.814583	 loss=0.733024	 lr=0.001000
Epoch[050] Batch [0979]/[1203]	Speed: 156.318462 samples/sec	 accuracy=80.783163	 loss=0.734248	 lr=0.001000
Epoch[050] Batch [0999]/[1203]	Speed: 115.188217 samples/sec	 accuracy=80.773000	 loss=0.734769	 lr=0.001000
Epoch[050] Batch [1019]/[1203]	Speed: 125.547748 samples/sec	 accuracy=80.752451	 loss=0.735379	 lr=0.001000
Epoch[050] Batch [1039]/[1203]	Speed: 151.344508 samples/sec	 accuracy=80.755769	 loss=0.735584	 lr=0.001000
Epoch[050] Batch [1059]/[1203]	Speed: 104.450664 samples/sec	 accuracy=80.740094	 loss=0.735950	 lr=0.001000
Epoch[050] Batch [1079]/[1203]	Speed: 140.352117 samples/sec	 accuracy=80.733333	 loss=0.736317	 lr=0.001000
Epoch[050] Batch [1099]/[1203]	Speed: 126.871530 samples/sec	 accuracy=80.735909	 loss=0.736403	 lr=0.001000
Epoch[050] Batch [1119]/[1203]	Speed: 98.577361 samples/sec	 accuracy=80.743304	 loss=0.736312	 lr=0.001000
Epoch[050] Batch [1139]/[1203]	Speed: 194.230766 samples/sec	 accuracy=80.742982	 loss=0.736058	 lr=0.001000
Epoch[050] Batch [1159]/[1203]	Speed: 96.179617 samples/sec	 accuracy=80.735345	 loss=0.736171	 lr=0.001000
Epoch[050] Batch [1179]/[1203]	Speed: 162.591853 samples/sec	 accuracy=80.736017	 loss=0.736051	 lr=0.001000
Epoch[050] Batch [1199]/[1203]	Speed: 245.443786 samples/sec	 accuracy=80.734167	 loss=0.736240	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.400000 acc-top5=85.125000
Batch [0039]/[0097]: acc-top1=63.387500 acc-top5=84.637500
Batch [0059]/[0097]: acc-top1=65.875000 acc-top5=85.925000
Batch [0079]/[0097]: acc-top1=65.975000 acc-top5=85.762500
[Epoch 050] training: accuracy=80.743558	 loss=0.735969
[Epoch 050] speed: 125 samples/sec	time cost: 2111.954038
[Epoch 050] validation: acc-top1=65.226804 acc-top5=85.664948 loss=1.533448
Epoch[051] Batch [0019]/[1203]	Speed: 69.574466 samples/sec	 accuracy=79.825000	 loss=0.747277	 lr=0.001000
Epoch[051] Batch [0039]/[1203]	Speed: 107.193998 samples/sec	 accuracy=81.025000	 loss=0.724500	 lr=0.001000
Epoch[051] Batch [0059]/[1203]	Speed: 200.661988 samples/sec	 accuracy=80.850000	 loss=0.729750	 lr=0.001000
Epoch[051] Batch [0079]/[1203]	Speed: 100.111174 samples/sec	 accuracy=80.756250	 loss=0.731653	 lr=0.001000
Epoch[051] Batch [0099]/[1203]	Speed: 111.434117 samples/sec	 accuracy=80.980000	 loss=0.731045	 lr=0.001000
Epoch[051] Batch [0119]/[1203]	Speed: 169.184031 samples/sec	 accuracy=80.925000	 loss=0.731418	 lr=0.001000
Epoch[051] Batch [0139]/[1203]	Speed: 97.955188 samples/sec	 accuracy=81.053571	 loss=0.725962	 lr=0.001000
Epoch[051] Batch [0159]/[1203]	Speed: 212.926302 samples/sec	 accuracy=80.959375	 loss=0.729821	 lr=0.001000
Epoch[051] Batch [0179]/[1203]	Speed: 105.505439 samples/sec	 accuracy=80.855556	 loss=0.732288	 lr=0.001000
Epoch[051] Batch [0199]/[1203]	Speed: 124.279565 samples/sec	 accuracy=80.885000	 loss=0.730089	 lr=0.001000
Epoch[051] Batch [0219]/[1203]	Speed: 146.059866 samples/sec	 accuracy=80.813636	 loss=0.733274	 lr=0.001000
Epoch[051] Batch [0239]/[1203]	Speed: 103.153366 samples/sec	 accuracy=80.733333	 loss=0.734200	 lr=0.001000
Epoch[051] Batch [0259]/[1203]	Speed: 141.947925 samples/sec	 accuracy=80.851923	 loss=0.731344	 lr=0.001000
Epoch[051] Batch [0279]/[1203]	Speed: 123.534235 samples/sec	 accuracy=80.803571	 loss=0.733216	 lr=0.001000
Epoch[051] Batch [0299]/[1203]	Speed: 110.783866 samples/sec	 accuracy=80.815000	 loss=0.733181	 lr=0.001000
Epoch[051] Batch [0319]/[1203]	Speed: 154.835086 samples/sec	 accuracy=80.903125	 loss=0.732253	 lr=0.001000
Epoch[051] Batch [0339]/[1203]	Speed: 105.203673 samples/sec	 accuracy=80.877941	 loss=0.733349	 lr=0.001000
Epoch[051] Batch [0359]/[1203]	Speed: 169.603613 samples/sec	 accuracy=80.891667	 loss=0.733927	 lr=0.001000
Epoch[051] Batch [0379]/[1203]	Speed: 113.033445 samples/sec	 accuracy=80.882895	 loss=0.733760	 lr=0.001000
Epoch[051] Batch [0399]/[1203]	Speed: 109.000074 samples/sec	 accuracy=80.910000	 loss=0.731801	 lr=0.001000
Epoch[051] Batch [0419]/[1203]	Speed: 163.527217 samples/sec	 accuracy=80.921429	 loss=0.731375	 lr=0.001000
Epoch[051] Batch [0439]/[1203]	Speed: 107.423186 samples/sec	 accuracy=80.917045	 loss=0.731119	 lr=0.001000
Epoch[051] Batch [0459]/[1203]	Speed: 140.443910 samples/sec	 accuracy=80.900000	 loss=0.731996	 lr=0.001000
Epoch[051] Batch [0479]/[1203]	Speed: 128.354959 samples/sec	 accuracy=80.889583	 loss=0.732100	 lr=0.001000
Epoch[051] Batch [0499]/[1203]	Speed: 109.857783 samples/sec	 accuracy=80.899000	 loss=0.732707	 lr=0.001000
Epoch[051] Batch [0519]/[1203]	Speed: 156.706678 samples/sec	 accuracy=80.926923	 loss=0.731966	 lr=0.001000
Epoch[051] Batch [0539]/[1203]	Speed: 110.381518 samples/sec	 accuracy=80.912963	 loss=0.731839	 lr=0.001000
Epoch[051] Batch [0559]/[1203]	Speed: 133.124161 samples/sec	 accuracy=80.909821	 loss=0.731405	 lr=0.001000
Epoch[051] Batch [0579]/[1203]	Speed: 132.256148 samples/sec	 accuracy=80.904310	 loss=0.731967	 lr=0.001000
Epoch[051] Batch [0599]/[1203]	Speed: 105.584883 samples/sec	 accuracy=80.901667	 loss=0.732056	 lr=0.001000
Epoch[051] Batch [0619]/[1203]	Speed: 167.556215 samples/sec	 accuracy=80.944355	 loss=0.731008	 lr=0.001000
Epoch[051] Batch [0639]/[1203]	Speed: 108.906143 samples/sec	 accuracy=80.938281	 loss=0.731547	 lr=0.001000
Epoch[051] Batch [0659]/[1203]	Speed: 122.452173 samples/sec	 accuracy=80.925758	 loss=0.731535	 lr=0.001000
Epoch[051] Batch [0679]/[1203]	Speed: 158.141346 samples/sec	 accuracy=80.925735	 loss=0.731258	 lr=0.001000
Epoch[051] Batch [0699]/[1203]	Speed: 99.786677 samples/sec	 accuracy=80.915000	 loss=0.731339	 lr=0.001000
Epoch[051] Batch [0719]/[1203]	Speed: 182.629091 samples/sec	 accuracy=80.901389	 loss=0.731613	 lr=0.001000
Epoch[051] Batch [0739]/[1203]	Speed: 113.023940 samples/sec	 accuracy=80.893243	 loss=0.731271	 lr=0.001000
Epoch[051] Batch [0759]/[1203]	Speed: 101.293226 samples/sec	 accuracy=80.896053	 loss=0.731201	 lr=0.001000
Epoch[051] Batch [0779]/[1203]	Speed: 191.611007 samples/sec	 accuracy=80.908974	 loss=0.730584	 lr=0.001000
Epoch[051] Batch [0799]/[1203]	Speed: 105.830971 samples/sec	 accuracy=80.900000	 loss=0.730956	 lr=0.001000
Epoch[051] Batch [0819]/[1203]	Speed: 117.064837 samples/sec	 accuracy=80.893293	 loss=0.731079	 lr=0.001000
Epoch[051] Batch [0839]/[1203]	Speed: 164.882826 samples/sec	 accuracy=80.885714	 loss=0.731370	 lr=0.001000
Epoch[051] Batch [0859]/[1203]	Speed: 95.682447 samples/sec	 accuracy=80.873256	 loss=0.732008	 lr=0.001000
Epoch[051] Batch [0879]/[1203]	Speed: 195.300091 samples/sec	 accuracy=80.861364	 loss=0.731933	 lr=0.001000
Epoch[051] Batch [0899]/[1203]	Speed: 108.595351 samples/sec	 accuracy=80.851111	 loss=0.732123	 lr=0.001000
Epoch[051] Batch [0919]/[1203]	Speed: 110.469126 samples/sec	 accuracy=80.846739	 loss=0.732417	 lr=0.001000
Epoch[051] Batch [0939]/[1203]	Speed: 151.986722 samples/sec	 accuracy=80.857447	 loss=0.732269	 lr=0.001000
Epoch[051] Batch [0959]/[1203]	Speed: 115.740505 samples/sec	 accuracy=80.846875	 loss=0.732228	 lr=0.001000
Epoch[051] Batch [0979]/[1203]	Speed: 125.002165 samples/sec	 accuracy=80.823980	 loss=0.733370	 lr=0.001000
Epoch[051] Batch [0999]/[1203]	Speed: 133.766228 samples/sec	 accuracy=80.825000	 loss=0.733437	 lr=0.001000
Epoch[051] Batch [1019]/[1203]	Speed: 110.251234 samples/sec	 accuracy=80.817157	 loss=0.733556	 lr=0.001000
Epoch[051] Batch [1039]/[1203]	Speed: 163.799149 samples/sec	 accuracy=80.824038	 loss=0.733549	 lr=0.001000
Epoch[051] Batch [1059]/[1203]	Speed: 115.889559 samples/sec	 accuracy=80.827830	 loss=0.733279	 lr=0.001000
Epoch[051] Batch [1079]/[1203]	Speed: 106.522073 samples/sec	 accuracy=80.799074	 loss=0.734008	 lr=0.001000
Epoch[051] Batch [1099]/[1203]	Speed: 168.966828 samples/sec	 accuracy=80.801364	 loss=0.733771	 lr=0.001000
Epoch[051] Batch [1119]/[1203]	Speed: 106.138854 samples/sec	 accuracy=80.797768	 loss=0.733546	 lr=0.001000
Epoch[051] Batch [1139]/[1203]	Speed: 137.747068 samples/sec	 accuracy=80.793421	 loss=0.733601	 lr=0.001000
Epoch[051] Batch [1159]/[1203]	Speed: 125.780905 samples/sec	 accuracy=80.806034	 loss=0.732919	 lr=0.001000
Epoch[051] Batch [1179]/[1203]	Speed: 128.985102 samples/sec	 accuracy=80.805932	 loss=0.732849	 lr=0.001000
Epoch[051] Batch [1199]/[1203]	Speed: 248.838467 samples/sec	 accuracy=80.795417	 loss=0.733003	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.550000 acc-top5=84.975000
Batch [0039]/[0097]: acc-top1=63.225000 acc-top5=84.750000
Batch [0059]/[0097]: acc-top1=65.591667 acc-top5=86.025000
Batch [0079]/[0097]: acc-top1=65.750000 acc-top5=85.893750
[Epoch 051] training: accuracy=80.796758	 loss=0.732963
[Epoch 051] speed: 125 samples/sec	time cost: 2111.355898
[Epoch 051] validation: acc-top1=65.190722 acc-top5=85.824742 loss=1.527642
Epoch[052] Batch [0019]/[1203]	Speed: 69.466492 samples/sec	 accuracy=81.100000	 loss=0.719351	 lr=0.001000
Epoch[052] Batch [0039]/[1203]	Speed: 99.322516 samples/sec	 accuracy=81.187500	 loss=0.712116	 lr=0.001000
Epoch[052] Batch [0059]/[1203]	Speed: 263.338090 samples/sec	 accuracy=81.133333	 loss=0.714796	 lr=0.001000
Epoch[052] Batch [0079]/[1203]	Speed: 91.099041 samples/sec	 accuracy=80.837500	 loss=0.727830	 lr=0.001000
Epoch[052] Batch [0099]/[1203]	Speed: 111.879176 samples/sec	 accuracy=81.075000	 loss=0.728131	 lr=0.001000
Epoch[052] Batch [0119]/[1203]	Speed: 194.072952 samples/sec	 accuracy=80.891667	 loss=0.728967	 lr=0.001000
Epoch[052] Batch [0139]/[1203]	Speed: 101.982746 samples/sec	 accuracy=80.925000	 loss=0.726914	 lr=0.001000
Epoch[052] Batch [0159]/[1203]	Speed: 225.388251 samples/sec	 accuracy=80.893750	 loss=0.729829	 lr=0.001000
Epoch[052] Batch [0179]/[1203]	Speed: 95.619704 samples/sec	 accuracy=80.994444	 loss=0.727904	 lr=0.001000
Epoch[052] Batch [0199]/[1203]	Speed: 104.501456 samples/sec	 accuracy=80.962500	 loss=0.727345	 lr=0.001000
Epoch[052] Batch [0219]/[1203]	Speed: 194.211190 samples/sec	 accuracy=80.950000	 loss=0.727719	 lr=0.001000
Epoch[052] Batch [0239]/[1203]	Speed: 99.141832 samples/sec	 accuracy=81.033333	 loss=0.726314	 lr=0.001000
Epoch[052] Batch [0259]/[1203]	Speed: 119.746870 samples/sec	 accuracy=81.119231	 loss=0.724729	 lr=0.001000
Epoch[052] Batch [0279]/[1203]	Speed: 150.749337 samples/sec	 accuracy=81.108929	 loss=0.725686	 lr=0.001000
Epoch[052] Batch [0299]/[1203]	Speed: 112.597280 samples/sec	 accuracy=81.028333	 loss=0.727425	 lr=0.001000
Epoch[052] Batch [0319]/[1203]	Speed: 187.325687 samples/sec	 accuracy=81.096875	 loss=0.726521	 lr=0.001000
Epoch[052] Batch [0339]/[1203]	Speed: 92.514960 samples/sec	 accuracy=81.127941	 loss=0.727018	 lr=0.001000
Epoch[052] Batch [0359]/[1203]	Speed: 131.354273 samples/sec	 accuracy=81.104167	 loss=0.726979	 lr=0.001000
Epoch[052] Batch [0379]/[1203]	Speed: 149.556759 samples/sec	 accuracy=81.065789	 loss=0.728234	 lr=0.001000
Epoch[052] Batch [0399]/[1203]	Speed: 108.714147 samples/sec	 accuracy=81.093750	 loss=0.727474	 lr=0.001000
Epoch[052] Batch [0419]/[1203]	Speed: 165.840178 samples/sec	 accuracy=81.076190	 loss=0.726978	 lr=0.001000
Epoch[052] Batch [0439]/[1203]	Speed: 101.201803 samples/sec	 accuracy=81.048864	 loss=0.726865	 lr=0.001000
Epoch[052] Batch [0459]/[1203]	Speed: 121.885518 samples/sec	 accuracy=81.026087	 loss=0.727687	 lr=0.001000
Epoch[052] Batch [0479]/[1203]	Speed: 157.180231 samples/sec	 accuracy=81.045833	 loss=0.726756	 lr=0.001000
Epoch[052] Batch [0499]/[1203]	Speed: 106.063032 samples/sec	 accuracy=81.007000	 loss=0.727412	 lr=0.001000
Epoch[052] Batch [0519]/[1203]	Speed: 143.300525 samples/sec	 accuracy=80.982692	 loss=0.727753	 lr=0.001000
Epoch[052] Batch [0539]/[1203]	Speed: 122.353248 samples/sec	 accuracy=80.977778	 loss=0.727754	 lr=0.001000
Epoch[052] Batch [0559]/[1203]	Speed: 115.184056 samples/sec	 accuracy=80.984821	 loss=0.727763	 lr=0.001000
Epoch[052] Batch [0579]/[1203]	Speed: 140.914975 samples/sec	 accuracy=80.984483	 loss=0.727004	 lr=0.001000
Epoch[052] Batch [0599]/[1203]	Speed: 117.184456 samples/sec	 accuracy=80.987500	 loss=0.727496	 lr=0.001000
Epoch[052] Batch [0619]/[1203]	Speed: 148.746878 samples/sec	 accuracy=81.002419	 loss=0.727343	 lr=0.001000
Epoch[052] Batch [0639]/[1203]	Speed: 114.994003 samples/sec	 accuracy=81.032812	 loss=0.726224	 lr=0.001000
Epoch[052] Batch [0659]/[1203]	Speed: 110.872480 samples/sec	 accuracy=81.031061	 loss=0.725740	 lr=0.001000
Epoch[052] Batch [0679]/[1203]	Speed: 153.225705 samples/sec	 accuracy=81.027941	 loss=0.725174	 lr=0.001000
Epoch[052] Batch [0699]/[1203]	Speed: 114.848715 samples/sec	 accuracy=81.046429	 loss=0.724191	 lr=0.001000
Epoch[052] Batch [0719]/[1203]	Speed: 130.755803 samples/sec	 accuracy=81.070139	 loss=0.724002	 lr=0.001000
Epoch[052] Batch [0739]/[1203]	Speed: 129.493947 samples/sec	 accuracy=81.064865	 loss=0.723966	 lr=0.001000
Epoch[052] Batch [0759]/[1203]	Speed: 109.710349 samples/sec	 accuracy=81.055263	 loss=0.724049	 lr=0.001000
Epoch[052] Batch [0779]/[1203]	Speed: 178.898332 samples/sec	 accuracy=81.039744	 loss=0.724882	 lr=0.001000
Epoch[052] Batch [0799]/[1203]	Speed: 110.181238 samples/sec	 accuracy=81.035625	 loss=0.725190	 lr=0.001000
Epoch[052] Batch [0819]/[1203]	Speed: 106.513900 samples/sec	 accuracy=81.036585	 loss=0.725180	 lr=0.001000
Epoch[052] Batch [0839]/[1203]	Speed: 159.428683 samples/sec	 accuracy=81.030357	 loss=0.725314	 lr=0.001000
Epoch[052] Batch [0859]/[1203]	Speed: 107.824069 samples/sec	 accuracy=81.038372	 loss=0.725218	 lr=0.001000
Epoch[052] Batch [0879]/[1203]	Speed: 170.118297 samples/sec	 accuracy=81.055682	 loss=0.724526	 lr=0.001000
Epoch[052] Batch [0899]/[1203]	Speed: 114.683003 samples/sec	 accuracy=81.040556	 loss=0.724672	 lr=0.001000
Epoch[052] Batch [0919]/[1203]	Speed: 98.925283 samples/sec	 accuracy=81.046739	 loss=0.724196	 lr=0.001000
Epoch[052] Batch [0939]/[1203]	Speed: 177.315641 samples/sec	 accuracy=81.038830	 loss=0.724667	 lr=0.001000
Epoch[052] Batch [0959]/[1203]	Speed: 107.298778 samples/sec	 accuracy=81.057292	 loss=0.724188	 lr=0.001000
Epoch[052] Batch [0979]/[1203]	Speed: 127.467107 samples/sec	 accuracy=81.065306	 loss=0.724028	 lr=0.001000
Epoch[052] Batch [0999]/[1203]	Speed: 150.717263 samples/sec	 accuracy=81.028500	 loss=0.724490	 lr=0.001000
Epoch[052] Batch [1019]/[1203]	Speed: 102.352017 samples/sec	 accuracy=81.049020	 loss=0.723975	 lr=0.001000
Epoch[052] Batch [1039]/[1203]	Speed: 185.736740 samples/sec	 accuracy=81.049038	 loss=0.724101	 lr=0.001000
Epoch[052] Batch [1059]/[1203]	Speed: 100.164867 samples/sec	 accuracy=81.027830	 loss=0.724549	 lr=0.001000
Epoch[052] Batch [1079]/[1203]	Speed: 130.856650 samples/sec	 accuracy=81.002315	 loss=0.725083	 lr=0.001000
Epoch[052] Batch [1099]/[1203]	Speed: 137.420413 samples/sec	 accuracy=80.981364	 loss=0.725893	 lr=0.001000
Epoch[052] Batch [1119]/[1203]	Speed: 106.105162 samples/sec	 accuracy=80.985268	 loss=0.725831	 lr=0.001000
Epoch[052] Batch [1139]/[1203]	Speed: 176.984508 samples/sec	 accuracy=80.985526	 loss=0.726086	 lr=0.001000
Epoch[052] Batch [1159]/[1203]	Speed: 105.424357 samples/sec	 accuracy=80.984914	 loss=0.726051	 lr=0.001000
Epoch[052] Batch [1179]/[1203]	Speed: 126.031233 samples/sec	 accuracy=81.006356	 loss=0.726015	 lr=0.001000
Epoch[052] Batch [1199]/[1203]	Speed: 286.315800 samples/sec	 accuracy=81.007500	 loss=0.726077	 lr=0.001000
Batch [0019]/[0097]: acc-top1=65.175000 acc-top5=85.500000
Batch [0039]/[0097]: acc-top1=63.450000 acc-top5=84.925000
Batch [0059]/[0097]: acc-top1=65.775000 acc-top5=86.116667
Batch [0079]/[0097]: acc-top1=65.831250 acc-top5=85.912500
[Epoch 052] training: accuracy=81.000000	 loss=0.726413
[Epoch 052] speed: 125 samples/sec	time cost: 2112.957464
[Epoch 052] validation: acc-top1=65.262887 acc-top5=85.814433 loss=1.532040
Epoch[053] Batch [0019]/[1203]	Speed: 69.767243 samples/sec	 accuracy=81.600000	 loss=0.707498	 lr=0.001000
Epoch[053] Batch [0039]/[1203]	Speed: 98.189686 samples/sec	 accuracy=82.050000	 loss=0.688281	 lr=0.001000
Epoch[053] Batch [0059]/[1203]	Speed: 240.941163 samples/sec	 accuracy=81.675000	 loss=0.703291	 lr=0.001000
Epoch[053] Batch [0079]/[1203]	Speed: 99.324657 samples/sec	 accuracy=81.350000	 loss=0.714910	 lr=0.001000
Epoch[053] Batch [0099]/[1203]	Speed: 98.542939 samples/sec	 accuracy=81.285000	 loss=0.719223	 lr=0.001000
Epoch[053] Batch [0119]/[1203]	Speed: 240.417990 samples/sec	 accuracy=81.554167	 loss=0.706494	 lr=0.001000
Epoch[053] Batch [0139]/[1203]	Speed: 94.042768 samples/sec	 accuracy=81.439286	 loss=0.711587	 lr=0.001000
Epoch[053] Batch [0159]/[1203]	Speed: 239.568820 samples/sec	 accuracy=81.359375	 loss=0.712862	 lr=0.001000
Epoch[053] Batch [0179]/[1203]	Speed: 96.170153 samples/sec	 accuracy=81.372222	 loss=0.711475	 lr=0.001000
Epoch[053] Batch [0199]/[1203]	Speed: 114.455711 samples/sec	 accuracy=81.327500	 loss=0.713745	 lr=0.001000
Epoch[053] Batch [0219]/[1203]	Speed: 175.841369 samples/sec	 accuracy=81.284091	 loss=0.713155	 lr=0.001000
Epoch[053] Batch [0239]/[1203]	Speed: 102.068050 samples/sec	 accuracy=81.333333	 loss=0.711478	 lr=0.001000
Epoch[053] Batch [0259]/[1203]	Speed: 123.287200 samples/sec	 accuracy=81.321154	 loss=0.712379	 lr=0.001000
Epoch[053] Batch [0279]/[1203]	Speed: 151.601641 samples/sec	 accuracy=81.300000	 loss=0.713860	 lr=0.001000
Epoch[053] Batch [0299]/[1203]	Speed: 106.854667 samples/sec	 accuracy=81.396667	 loss=0.711769	 lr=0.001000
Epoch[053] Batch [0319]/[1203]	Speed: 195.508955 samples/sec	 accuracy=81.379688	 loss=0.711959	 lr=0.001000
Epoch[053] Batch [0339]/[1203]	Speed: 95.625944 samples/sec	 accuracy=81.382353	 loss=0.711421	 lr=0.001000
Epoch[053] Batch [0359]/[1203]	Speed: 146.158403 samples/sec	 accuracy=81.316667	 loss=0.712527	 lr=0.001000
Epoch[053] Batch [0379]/[1203]	Speed: 130.668893 samples/sec	 accuracy=81.313158	 loss=0.712606	 lr=0.001000
Epoch[053] Batch [0399]/[1203]	Speed: 108.516917 samples/sec	 accuracy=81.263750	 loss=0.714077	 lr=0.001000
Epoch[053] Batch [0419]/[1203]	Speed: 145.522287 samples/sec	 accuracy=81.276190	 loss=0.713447	 lr=0.001000
Epoch[053] Batch [0439]/[1203]	Speed: 115.766050 samples/sec	 accuracy=81.285227	 loss=0.713692	 lr=0.001000
Epoch[053] Batch [0459]/[1203]	Speed: 134.336394 samples/sec	 accuracy=81.269565	 loss=0.714177	 lr=0.001000
Epoch[053] Batch [0479]/[1203]	Speed: 138.704965 samples/sec	 accuracy=81.197917	 loss=0.716318	 lr=0.001000
Epoch[053] Batch [0499]/[1203]	Speed: 104.567946 samples/sec	 accuracy=81.208000	 loss=0.716633	 lr=0.001000
Epoch[053] Batch [0519]/[1203]	Speed: 160.626422 samples/sec	 accuracy=81.230769	 loss=0.717243	 lr=0.001000
Epoch[053] Batch [0539]/[1203]	Speed: 109.132693 samples/sec	 accuracy=81.225000	 loss=0.717543	 lr=0.001000
Epoch[053] Batch [0559]/[1203]	Speed: 120.900926 samples/sec	 accuracy=81.239286	 loss=0.717069	 lr=0.001000
Epoch[053] Batch [0579]/[1203]	Speed: 152.127343 samples/sec	 accuracy=81.244828	 loss=0.717072	 lr=0.001000
Epoch[053] Batch [0599]/[1203]	Speed: 104.366714 samples/sec	 accuracy=81.205000	 loss=0.717938	 lr=0.001000
Epoch[053] Batch [0619]/[1203]	Speed: 164.023155 samples/sec	 accuracy=81.210484	 loss=0.718107	 lr=0.001000
Epoch[053] Batch [0639]/[1203]	Speed: 108.253058 samples/sec	 accuracy=81.253906	 loss=0.717138	 lr=0.001000
Epoch[053] Batch [0659]/[1203]	Speed: 127.881451 samples/sec	 accuracy=81.218182	 loss=0.718157	 lr=0.001000
Epoch[053] Batch [0679]/[1203]	Speed: 156.888318 samples/sec	 accuracy=81.257353	 loss=0.717009	 lr=0.001000
Epoch[053] Batch [0699]/[1203]	Speed: 103.666599 samples/sec	 accuracy=81.245000	 loss=0.717324	 lr=0.001000
Epoch[053] Batch [0719]/[1203]	Speed: 138.717675 samples/sec	 accuracy=81.246528	 loss=0.717246	 lr=0.001000
Epoch[053] Batch [0739]/[1203]	Speed: 114.145907 samples/sec	 accuracy=81.239865	 loss=0.718045	 lr=0.001000
Epoch[053] Batch [0759]/[1203]	Speed: 119.367604 samples/sec	 accuracy=81.227632	 loss=0.717817	 lr=0.001000
Epoch[053] Batch [0779]/[1203]	Speed: 158.208571 samples/sec	 accuracy=81.246154	 loss=0.718035	 lr=0.001000
Epoch[053] Batch [0799]/[1203]	Speed: 114.448599 samples/sec	 accuracy=81.246250	 loss=0.718116	 lr=0.001000
Epoch[053] Batch [0819]/[1203]	Speed: 124.879956 samples/sec	 accuracy=81.245732	 loss=0.718575	 lr=0.001000
Epoch[053] Batch [0839]/[1203]	Speed: 126.539844 samples/sec	 accuracy=81.243452	 loss=0.719250	 lr=0.001000
Epoch[053] Batch [0859]/[1203]	Speed: 116.765289 samples/sec	 accuracy=81.247093	 loss=0.718680	 lr=0.001000
Epoch[053] Batch [0879]/[1203]	Speed: 144.733801 samples/sec	 accuracy=81.265909	 loss=0.718190	 lr=0.001000
Epoch[053] Batch [0899]/[1203]	Speed: 116.619422 samples/sec	 accuracy=81.255556	 loss=0.718164	 lr=0.001000
Epoch[053] Batch [0919]/[1203]	Speed: 133.255232 samples/sec	 accuracy=81.271196	 loss=0.717564	 lr=0.001000
Epoch[053] Batch [0939]/[1203]	Speed: 133.619665 samples/sec	 accuracy=81.250000	 loss=0.717956	 lr=0.001000
Epoch[053] Batch [0959]/[1203]	Speed: 113.775323 samples/sec	 accuracy=81.230729	 loss=0.718713	 lr=0.001000
Epoch[053] Batch [0979]/[1203]	Speed: 138.665353 samples/sec	 accuracy=81.206633	 loss=0.719149	 lr=0.001000
Epoch[053] Batch [0999]/[1203]	Speed: 120.350641 samples/sec	 accuracy=81.196500	 loss=0.719348	 lr=0.001000
Epoch[053] Batch [1019]/[1203]	Speed: 116.906866 samples/sec	 accuracy=81.192157	 loss=0.719305	 lr=0.001000
Epoch[053] Batch [1039]/[1203]	Speed: 156.303375 samples/sec	 accuracy=81.186058	 loss=0.719836	 lr=0.001000
Epoch[053] Batch [1059]/[1203]	Speed: 107.635796 samples/sec	 accuracy=81.195283	 loss=0.719668	 lr=0.001000
Epoch[053] Batch [1079]/[1203]	Speed: 139.089667 samples/sec	 accuracy=81.182870	 loss=0.719903	 lr=0.001000
Epoch[053] Batch [1099]/[1203]	Speed: 116.368122 samples/sec	 accuracy=81.152273	 loss=0.720684	 lr=0.001000
Epoch[053] Batch [1119]/[1203]	Speed: 121.445698 samples/sec	 accuracy=81.158482	 loss=0.720479	 lr=0.001000
Epoch[053] Batch [1139]/[1203]	Speed: 155.050048 samples/sec	 accuracy=81.169737	 loss=0.720439	 lr=0.001000
Epoch[053] Batch [1159]/[1203]	Speed: 113.254882 samples/sec	 accuracy=81.168966	 loss=0.720432	 lr=0.001000
Epoch[053] Batch [1179]/[1203]	Speed: 137.985015 samples/sec	 accuracy=81.159322	 loss=0.720980	 lr=0.001000
Epoch[053] Batch [1199]/[1203]	Speed: 226.651327 samples/sec	 accuracy=81.153750	 loss=0.721081	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.550000 acc-top5=85.350000
Batch [0039]/[0097]: acc-top1=63.362500 acc-top5=84.987500
Batch [0059]/[0097]: acc-top1=65.691667 acc-top5=86.125000
Batch [0079]/[0097]: acc-top1=65.831250 acc-top5=85.950000
[Epoch 053] training: accuracy=81.144223	 loss=0.721203
[Epoch 053] speed: 125 samples/sec	time cost: 2109.028389
[Epoch 053] validation: acc-top1=65.154639 acc-top5=85.783505 loss=1.528732
Epoch[054] Batch [0019]/[1203]	Speed: 69.856225 samples/sec	 accuracy=80.775000	 loss=0.729030	 lr=0.001000
Epoch[054] Batch [0039]/[1203]	Speed: 106.695208 samples/sec	 accuracy=81.525000	 loss=0.712519	 lr=0.001000
Epoch[054] Batch [0059]/[1203]	Speed: 197.804806 samples/sec	 accuracy=81.616667	 loss=0.712986	 lr=0.001000
Epoch[054] Batch [0079]/[1203]	Speed: 98.305242 samples/sec	 accuracy=81.850000	 loss=0.707034	 lr=0.001000
Epoch[054] Batch [0099]/[1203]	Speed: 129.665969 samples/sec	 accuracy=81.860000	 loss=0.706512	 lr=0.001000
Epoch[054] Batch [0119]/[1203]	Speed: 137.541920 samples/sec	 accuracy=81.787500	 loss=0.707845	 lr=0.001000
Epoch[054] Batch [0139]/[1203]	Speed: 107.218536 samples/sec	 accuracy=81.725000	 loss=0.710719	 lr=0.001000
Epoch[054] Batch [0159]/[1203]	Speed: 193.332657 samples/sec	 accuracy=81.784375	 loss=0.710162	 lr=0.001000
Epoch[054] Batch [0179]/[1203]	Speed: 98.164497 samples/sec	 accuracy=81.658333	 loss=0.712272	 lr=0.001000
Epoch[054] Batch [0199]/[1203]	Speed: 138.100588 samples/sec	 accuracy=81.572500	 loss=0.714362	 lr=0.001000
Epoch[054] Batch [0219]/[1203]	Speed: 129.359577 samples/sec	 accuracy=81.513636	 loss=0.714860	 lr=0.001000
Epoch[054] Batch [0239]/[1203]	Speed: 115.506942 samples/sec	 accuracy=81.516667	 loss=0.714805	 lr=0.001000
Epoch[054] Batch [0259]/[1203]	Speed: 150.010089 samples/sec	 accuracy=81.494231	 loss=0.715336	 lr=0.001000
Epoch[054] Batch [0279]/[1203]	Speed: 108.474607 samples/sec	 accuracy=81.530357	 loss=0.713452	 lr=0.001000
Epoch[054] Batch [0299]/[1203]	Speed: 134.974813 samples/sec	 accuracy=81.548333	 loss=0.713089	 lr=0.001000
Epoch[054] Batch [0319]/[1203]	Speed: 135.122501 samples/sec	 accuracy=81.498437	 loss=0.714931	 lr=0.001000
Epoch[054] Batch [0339]/[1203]	Speed: 109.977949 samples/sec	 accuracy=81.497059	 loss=0.714603	 lr=0.001000
Epoch[054] Batch [0359]/[1203]	Speed: 157.138718 samples/sec	 accuracy=81.451389	 loss=0.715974	 lr=0.001000
Epoch[054] Batch [0379]/[1203]	Speed: 115.115075 samples/sec	 accuracy=81.463158	 loss=0.715622	 lr=0.001000
Epoch[054] Batch [0399]/[1203]	Speed: 115.953465 samples/sec	 accuracy=81.482500	 loss=0.715871	 lr=0.001000
Epoch[054] Batch [0419]/[1203]	Speed: 154.611734 samples/sec	 accuracy=81.450000	 loss=0.716310	 lr=0.001000
Epoch[054] Batch [0439]/[1203]	Speed: 107.065691 samples/sec	 accuracy=81.413636	 loss=0.716732	 lr=0.001000
Epoch[054] Batch [0459]/[1203]	Speed: 142.071703 samples/sec	 accuracy=81.422826	 loss=0.716545	 lr=0.001000
Epoch[054] Batch [0479]/[1203]	Speed: 126.209482 samples/sec	 accuracy=81.393750	 loss=0.717234	 lr=0.001000
Epoch[054] Batch [0499]/[1203]	Speed: 124.080178 samples/sec	 accuracy=81.348000	 loss=0.717950	 lr=0.001000
Epoch[054] Batch [0519]/[1203]	Speed: 141.394865 samples/sec	 accuracy=81.341346	 loss=0.717741	 lr=0.001000
Epoch[054] Batch [0539]/[1203]	Speed: 109.637601 samples/sec	 accuracy=81.322222	 loss=0.718411	 lr=0.001000
Epoch[054] Batch [0559]/[1203]	Speed: 125.963365 samples/sec	 accuracy=81.292857	 loss=0.719284	 lr=0.001000
Epoch[054] Batch [0579]/[1203]	Speed: 137.602351 samples/sec	 accuracy=81.303448	 loss=0.718701	 lr=0.001000
Epoch[054] Batch [0599]/[1203]	Speed: 116.884918 samples/sec	 accuracy=81.296667	 loss=0.718918	 lr=0.001000
Epoch[054] Batch [0619]/[1203]	Speed: 143.779688 samples/sec	 accuracy=81.266935	 loss=0.719218	 lr=0.001000
Epoch[054] Batch [0639]/[1203]	Speed: 103.474791 samples/sec	 accuracy=81.256250	 loss=0.719031	 lr=0.001000
Epoch[054] Batch [0659]/[1203]	Speed: 137.957889 samples/sec	 accuracy=81.290152	 loss=0.717984	 lr=0.001000
Epoch[054] Batch [0679]/[1203]	Speed: 138.644742 samples/sec	 accuracy=81.291176	 loss=0.718662	 lr=0.001000
Epoch[054] Batch [0699]/[1203]	Speed: 119.327874 samples/sec	 accuracy=81.286429	 loss=0.718773	 lr=0.001000
Epoch[054] Batch [0719]/[1203]	Speed: 133.496270 samples/sec	 accuracy=81.272222	 loss=0.719478	 lr=0.001000
Epoch[054] Batch [0739]/[1203]	Speed: 103.835362 samples/sec	 accuracy=81.266216	 loss=0.719709	 lr=0.001000
Epoch[054] Batch [0759]/[1203]	Speed: 164.044779 samples/sec	 accuracy=81.265789	 loss=0.719689	 lr=0.001000
Epoch[054] Batch [0779]/[1203]	Speed: 123.401653 samples/sec	 accuracy=81.273077	 loss=0.719669	 lr=0.001000
Epoch[054] Batch [0799]/[1203]	Speed: 115.090441 samples/sec	 accuracy=81.280000	 loss=0.719241	 lr=0.001000
Epoch[054] Batch [0819]/[1203]	Speed: 144.309396 samples/sec	 accuracy=81.260366	 loss=0.719440	 lr=0.001000
Epoch[054] Batch [0839]/[1203]	Speed: 111.351033 samples/sec	 accuracy=81.283333	 loss=0.718721	 lr=0.001000
Epoch[054] Batch [0859]/[1203]	Speed: 146.272681 samples/sec	 accuracy=81.309302	 loss=0.717788	 lr=0.001000
Epoch[054] Batch [0879]/[1203]	Speed: 120.096769 samples/sec	 accuracy=81.325000	 loss=0.717124	 lr=0.001000
Epoch[054] Batch [0899]/[1203]	Speed: 110.782214 samples/sec	 accuracy=81.291667	 loss=0.717443	 lr=0.001000
Epoch[054] Batch [0919]/[1203]	Speed: 146.152168 samples/sec	 accuracy=81.298913	 loss=0.717296	 lr=0.001000
Epoch[054] Batch [0939]/[1203]	Speed: 126.507380 samples/sec	 accuracy=81.317021	 loss=0.717163	 lr=0.001000
Epoch[054] Batch [0959]/[1203]	Speed: 131.839745 samples/sec	 accuracy=81.291667	 loss=0.717871	 lr=0.001000
Epoch[054] Batch [0979]/[1203]	Speed: 117.013077 samples/sec	 accuracy=81.300000	 loss=0.717658	 lr=0.001000
Epoch[054] Batch [0999]/[1203]	Speed: 122.582128 samples/sec	 accuracy=81.308000	 loss=0.716910	 lr=0.001000
Epoch[054] Batch [1019]/[1203]	Speed: 132.883443 samples/sec	 accuracy=81.316667	 loss=0.716301	 lr=0.001000
Epoch[054] Batch [1039]/[1203]	Speed: 126.978438 samples/sec	 accuracy=81.312500	 loss=0.716590	 lr=0.001000
Epoch[054] Batch [1059]/[1203]	Speed: 113.834207 samples/sec	 accuracy=81.308962	 loss=0.716666	 lr=0.001000
Epoch[054] Batch [1079]/[1203]	Speed: 151.941996 samples/sec	 accuracy=81.303241	 loss=0.716598	 lr=0.001000
Epoch[054] Batch [1099]/[1203]	Speed: 118.870131 samples/sec	 accuracy=81.300455	 loss=0.716717	 lr=0.001000
Epoch[054] Batch [1119]/[1203]	Speed: 136.382893 samples/sec	 accuracy=81.282589	 loss=0.717338	 lr=0.001000
Epoch[054] Batch [1139]/[1203]	Speed: 116.270241 samples/sec	 accuracy=81.273246	 loss=0.717754	 lr=0.001000
Epoch[054] Batch [1159]/[1203]	Speed: 116.480640 samples/sec	 accuracy=81.273707	 loss=0.717706	 lr=0.001000
Epoch[054] Batch [1179]/[1203]	Speed: 157.248122 samples/sec	 accuracy=81.275424	 loss=0.717699	 lr=0.001000
Epoch[054] Batch [1199]/[1203]	Speed: 222.448069 samples/sec	 accuracy=81.261250	 loss=0.718262	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.650000 acc-top5=85.050000
Batch [0039]/[0097]: acc-top1=63.425000 acc-top5=84.775000
Batch [0059]/[0097]: acc-top1=65.691667 acc-top5=86.075000
Batch [0079]/[0097]: acc-top1=65.950000 acc-top5=86.006250
[Epoch 054] training: accuracy=81.259767	 loss=0.718336
[Epoch 054] speed: 125 samples/sec	time cost: 2105.784294
[Epoch 054] validation: acc-top1=65.350515 acc-top5=85.886598 loss=1.529259
Epoch[055] Batch [0019]/[1204]	Speed: 70.159467 samples/sec	 accuracy=81.250000	 loss=0.709435	 lr=0.001000
Epoch[055] Batch [0039]/[1204]	Speed: 103.908437 samples/sec	 accuracy=81.137500	 loss=0.719597	 lr=0.001000
Epoch[055] Batch [0059]/[1204]	Speed: 215.576909 samples/sec	 accuracy=81.266667	 loss=0.711729	 lr=0.001000
Epoch[055] Batch [0079]/[1204]	Speed: 93.855292 samples/sec	 accuracy=81.375000	 loss=0.706474	 lr=0.001000
Epoch[055] Batch [0099]/[1204]	Speed: 114.514977 samples/sec	 accuracy=81.345000	 loss=0.707474	 lr=0.001000
Epoch[055] Batch [0119]/[1204]	Speed: 182.028114 samples/sec	 accuracy=81.516667	 loss=0.704273	 lr=0.001000
Epoch[055] Batch [0139]/[1204]	Speed: 99.737292 samples/sec	 accuracy=81.371429	 loss=0.710175	 lr=0.001000
Epoch[055] Batch [0159]/[1204]	Speed: 202.413038 samples/sec	 accuracy=81.240625	 loss=0.713748	 lr=0.001000
Epoch[055] Batch [0179]/[1204]	Speed: 103.108014 samples/sec	 accuracy=81.388889	 loss=0.710686	 lr=0.001000
Epoch[055] Batch [0199]/[1204]	Speed: 118.632347 samples/sec	 accuracy=81.407500	 loss=0.708612	 lr=0.001000
Epoch[055] Batch [0219]/[1204]	Speed: 148.103453 samples/sec	 accuracy=81.447727	 loss=0.707649	 lr=0.001000
Epoch[055] Batch [0239]/[1204]	Speed: 108.249837 samples/sec	 accuracy=81.437500	 loss=0.707119	 lr=0.001000
Epoch[055] Batch [0259]/[1204]	Speed: 125.086167 samples/sec	 accuracy=81.459615	 loss=0.706661	 lr=0.001000
Epoch[055] Batch [0279]/[1204]	Speed: 145.135993 samples/sec	 accuracy=81.492857	 loss=0.706533	 lr=0.001000
Epoch[055] Batch [0299]/[1204]	Speed: 117.938150 samples/sec	 accuracy=81.546667	 loss=0.705043	 lr=0.001000
Epoch[055] Batch [0319]/[1204]	Speed: 140.531206 samples/sec	 accuracy=81.467188	 loss=0.707185	 lr=0.001000
Epoch[055] Batch [0339]/[1204]	Speed: 118.617611 samples/sec	 accuracy=81.438235	 loss=0.707856	 lr=0.001000
Epoch[055] Batch [0359]/[1204]	Speed: 115.124961 samples/sec	 accuracy=81.419444	 loss=0.708738	 lr=0.001000
Epoch[055] Batch [0379]/[1204]	Speed: 126.192665 samples/sec	 accuracy=81.423684	 loss=0.709088	 lr=0.001000
Epoch[055] Batch [0399]/[1204]	Speed: 136.130945 samples/sec	 accuracy=81.391250	 loss=0.710674	 lr=0.001000
Epoch[055] Batch [0419]/[1204]	Speed: 131.671656 samples/sec	 accuracy=81.401190	 loss=0.710357	 lr=0.001000
Epoch[055] Batch [0439]/[1204]	Speed: 121.878151 samples/sec	 accuracy=81.393182	 loss=0.711079	 lr=0.001000
Epoch[055] Batch [0459]/[1204]	Speed: 117.961158 samples/sec	 accuracy=81.455435	 loss=0.709486	 lr=0.001000
Epoch[055] Batch [0479]/[1204]	Speed: 128.243244 samples/sec	 accuracy=81.482292	 loss=0.708477	 lr=0.001000
Epoch[055] Batch [0499]/[1204]	Speed: 137.446121 samples/sec	 accuracy=81.485000	 loss=0.708369	 lr=0.001000
Epoch[055] Batch [0519]/[1204]	Speed: 132.672025 samples/sec	 accuracy=81.469231	 loss=0.709977	 lr=0.001000
Epoch[055] Batch [0539]/[1204]	Speed: 112.534758 samples/sec	 accuracy=81.437037	 loss=0.710287	 lr=0.001000
Epoch[055] Batch [0559]/[1204]	Speed: 120.260760 samples/sec	 accuracy=81.407143	 loss=0.712005	 lr=0.001000
Epoch[055] Batch [0579]/[1204]	Speed: 144.817975 samples/sec	 accuracy=81.424138	 loss=0.711487	 lr=0.001000
Epoch[055] Batch [0599]/[1204]	Speed: 126.403458 samples/sec	 accuracy=81.440000	 loss=0.710927	 lr=0.001000
Epoch[055] Batch [0619]/[1204]	Speed: 121.626893 samples/sec	 accuracy=81.424194	 loss=0.711069	 lr=0.001000
Epoch[055] Batch [0639]/[1204]	Speed: 106.897305 samples/sec	 accuracy=81.461719	 loss=0.710162	 lr=0.001000
Epoch[055] Batch [0659]/[1204]	Speed: 149.993588 samples/sec	 accuracy=81.452273	 loss=0.709968	 lr=0.001000
Epoch[055] Batch [0679]/[1204]	Speed: 122.319739 samples/sec	 accuracy=81.451471	 loss=0.709687	 lr=0.001000
Epoch[055] Batch [0699]/[1204]	Speed: 131.182446 samples/sec	 accuracy=81.436429	 loss=0.710160	 lr=0.001000
Epoch[055] Batch [0719]/[1204]	Speed: 126.967224 samples/sec	 accuracy=81.443056	 loss=0.710300	 lr=0.001000
Epoch[055] Batch [0739]/[1204]	Speed: 103.231762 samples/sec	 accuracy=81.431757	 loss=0.710406	 lr=0.001000
Epoch[055] Batch [0759]/[1204]	Speed: 185.883396 samples/sec	 accuracy=81.391447	 loss=0.711121	 lr=0.001000
Epoch[055] Batch [0779]/[1204]	Speed: 109.343199 samples/sec	 accuracy=81.407692	 loss=0.711172	 lr=0.001000
Epoch[055] Batch [0799]/[1204]	Speed: 117.610986 samples/sec	 accuracy=81.406250	 loss=0.711248	 lr=0.001000
Epoch[055] Batch [0819]/[1204]	Speed: 160.372619 samples/sec	 accuracy=81.401220	 loss=0.711673	 lr=0.001000
Epoch[055] Batch [0839]/[1204]	Speed: 92.060969 samples/sec	 accuracy=81.405952	 loss=0.711181	 lr=0.001000
Epoch[055] Batch [0859]/[1204]	Speed: 193.057064 samples/sec	 accuracy=81.401744	 loss=0.711287	 lr=0.001000
Epoch[055] Batch [0879]/[1204]	Speed: 109.135206 samples/sec	 accuracy=81.387500	 loss=0.711347	 lr=0.001000
Epoch[055] Batch [0899]/[1204]	Speed: 128.908986 samples/sec	 accuracy=81.418889	 loss=0.710527	 lr=0.001000
Epoch[055] Batch [0919]/[1204]	Speed: 147.131851 samples/sec	 accuracy=81.406522	 loss=0.710885	 lr=0.001000
Epoch[055] Batch [0939]/[1204]	Speed: 98.756301 samples/sec	 accuracy=81.403191	 loss=0.710895	 lr=0.001000
Epoch[055] Batch [0959]/[1204]	Speed: 190.743447 samples/sec	 accuracy=81.416667	 loss=0.710804	 lr=0.001000
Epoch[055] Batch [0979]/[1204]	Speed: 101.411620 samples/sec	 accuracy=81.415306	 loss=0.710712	 lr=0.001000
Epoch[055] Batch [0999]/[1204]	Speed: 116.590373 samples/sec	 accuracy=81.419000	 loss=0.710746	 lr=0.001000
Epoch[055] Batch [1019]/[1204]	Speed: 160.739953 samples/sec	 accuracy=81.418137	 loss=0.711083	 lr=0.001000
Epoch[055] Batch [1039]/[1204]	Speed: 117.002057 samples/sec	 accuracy=81.404327	 loss=0.711374	 lr=0.001000
Epoch[055] Batch [1059]/[1204]	Speed: 140.380340 samples/sec	 accuracy=81.412736	 loss=0.711150	 lr=0.001000
Epoch[055] Batch [1079]/[1204]	Speed: 104.321184 samples/sec	 accuracy=81.411111	 loss=0.711267	 lr=0.001000
Epoch[055] Batch [1099]/[1204]	Speed: 120.770476 samples/sec	 accuracy=81.384545	 loss=0.711837	 lr=0.001000
Epoch[055] Batch [1119]/[1204]	Speed: 157.512698 samples/sec	 accuracy=81.371429	 loss=0.712211	 lr=0.001000
Epoch[055] Batch [1139]/[1204]	Speed: 118.646240 samples/sec	 accuracy=81.348246	 loss=0.712903	 lr=0.001000
Epoch[055] Batch [1159]/[1204]	Speed: 120.137499 samples/sec	 accuracy=81.362931	 loss=0.712655	 lr=0.001000
Epoch[055] Batch [1179]/[1204]	Speed: 143.346884 samples/sec	 accuracy=81.355932	 loss=0.712850	 lr=0.001000
Epoch[055] Batch [1199]/[1204]	Speed: 224.770770 samples/sec	 accuracy=81.353333	 loss=0.712882	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.525000 acc-top5=85.350000
Batch [0039]/[0097]: acc-top1=63.437500 acc-top5=84.937500
Batch [0059]/[0097]: acc-top1=65.633333 acc-top5=86.133333
Batch [0079]/[0097]: acc-top1=65.775000 acc-top5=85.918750
[Epoch 055] training: accuracy=81.358389	 loss=0.712768
[Epoch 055] speed: 125 samples/sec	time cost: 2115.959648
[Epoch 055] validation: acc-top1=65.164948 acc-top5=85.804124 loss=1.532250
Epoch[056] Batch [0019]/[1203]	Speed: 70.525254 samples/sec	 accuracy=79.825000	 loss=0.760462	 lr=0.001000
Epoch[056] Batch [0039]/[1203]	Speed: 103.928539 samples/sec	 accuracy=80.500000	 loss=0.725746	 lr=0.001000
Epoch[056] Batch [0059]/[1203]	Speed: 220.194920 samples/sec	 accuracy=81.258333	 loss=0.709970	 lr=0.001000
Epoch[056] Batch [0079]/[1203]	Speed: 93.760169 samples/sec	 accuracy=81.256250	 loss=0.709197	 lr=0.001000
Epoch[056] Batch [0099]/[1203]	Speed: 120.252447 samples/sec	 accuracy=81.645000	 loss=0.699376	 lr=0.001000
Epoch[056] Batch [0119]/[1203]	Speed: 174.573660 samples/sec	 accuracy=81.491667	 loss=0.703967	 lr=0.001000
Epoch[056] Batch [0139]/[1203]	Speed: 107.725975 samples/sec	 accuracy=81.678571	 loss=0.699925	 lr=0.001000
Epoch[056] Batch [0159]/[1203]	Speed: 170.567785 samples/sec	 accuracy=81.687500	 loss=0.699118	 lr=0.001000
Epoch[056] Batch [0179]/[1203]	Speed: 104.686483 samples/sec	 accuracy=81.672222	 loss=0.699827	 lr=0.001000
Epoch[056] Batch [0199]/[1203]	Speed: 117.119004 samples/sec	 accuracy=81.557500	 loss=0.702607	 lr=0.001000
Epoch[056] Batch [0219]/[1203]	Speed: 149.922173 samples/sec	 accuracy=81.538636	 loss=0.702651	 lr=0.001000
Epoch[056] Batch [0239]/[1203]	Speed: 112.410013 samples/sec	 accuracy=81.516667	 loss=0.702277	 lr=0.001000
Epoch[056] Batch [0259]/[1203]	Speed: 175.463813 samples/sec	 accuracy=81.436538	 loss=0.705435	 lr=0.001000
Epoch[056] Batch [0279]/[1203]	Speed: 102.101555 samples/sec	 accuracy=81.508929	 loss=0.704136	 lr=0.001000
Epoch[056] Batch [0299]/[1203]	Speed: 129.410146 samples/sec	 accuracy=81.560000	 loss=0.702678	 lr=0.001000
Epoch[056] Batch [0319]/[1203]	Speed: 148.815740 samples/sec	 accuracy=81.545312	 loss=0.703302	 lr=0.001000
Epoch[056] Batch [0339]/[1203]	Speed: 101.535878 samples/sec	 accuracy=81.548529	 loss=0.704569	 lr=0.001000
Epoch[056] Batch [0359]/[1203]	Speed: 159.597574 samples/sec	 accuracy=81.572222	 loss=0.703373	 lr=0.001000
Epoch[056] Batch [0379]/[1203]	Speed: 113.815249 samples/sec	 accuracy=81.546053	 loss=0.703228	 lr=0.001000
Epoch[056] Batch [0399]/[1203]	Speed: 111.655347 samples/sec	 accuracy=81.618750	 loss=0.701473	 lr=0.001000
Epoch[056] Batch [0419]/[1203]	Speed: 178.771982 samples/sec	 accuracy=81.621429	 loss=0.700980	 lr=0.001000
Epoch[056] Batch [0439]/[1203]	Speed: 108.244843 samples/sec	 accuracy=81.622727	 loss=0.700861	 lr=0.001000
Epoch[056] Batch [0459]/[1203]	Speed: 110.481688 samples/sec	 accuracy=81.575000	 loss=0.701278	 lr=0.001000
Epoch[056] Batch [0479]/[1203]	Speed: 149.303681 samples/sec	 accuracy=81.556250	 loss=0.701884	 lr=0.001000
Epoch[056] Batch [0499]/[1203]	Speed: 128.397336 samples/sec	 accuracy=81.559000	 loss=0.701661	 lr=0.001000
Epoch[056] Batch [0519]/[1203]	Speed: 136.535474 samples/sec	 accuracy=81.519231	 loss=0.702295	 lr=0.001000
Epoch[056] Batch [0539]/[1203]	Speed: 114.053667 samples/sec	 accuracy=81.486111	 loss=0.703456	 lr=0.001000
Epoch[056] Batch [0559]/[1203]	Speed: 123.388027 samples/sec	 accuracy=81.475893	 loss=0.703833	 lr=0.001000
Epoch[056] Batch [0579]/[1203]	Speed: 134.076210 samples/sec	 accuracy=81.474138	 loss=0.704098	 lr=0.001000
Epoch[056] Batch [0599]/[1203]	Speed: 118.854872 samples/sec	 accuracy=81.453333	 loss=0.704869	 lr=0.001000
Epoch[056] Batch [0619]/[1203]	Speed: 145.271390 samples/sec	 accuracy=81.445968	 loss=0.704692	 lr=0.001000
Epoch[056] Batch [0639]/[1203]	Speed: 114.916812 samples/sec	 accuracy=81.454688	 loss=0.704019	 lr=0.001000
Epoch[056] Batch [0659]/[1203]	Speed: 127.795508 samples/sec	 accuracy=81.460606	 loss=0.703564	 lr=0.001000
Epoch[056] Batch [0679]/[1203]	Speed: 139.530124 samples/sec	 accuracy=81.468382	 loss=0.703528	 lr=0.001000
Epoch[056] Batch [0699]/[1203]	Speed: 107.056832 samples/sec	 accuracy=81.465000	 loss=0.703791	 lr=0.001000
Epoch[056] Batch [0719]/[1203]	Speed: 154.050025 samples/sec	 accuracy=81.440972	 loss=0.704427	 lr=0.001000
Epoch[056] Batch [0739]/[1203]	Speed: 112.676512 samples/sec	 accuracy=81.450000	 loss=0.704272	 lr=0.001000
Epoch[056] Batch [0759]/[1203]	Speed: 131.519040 samples/sec	 accuracy=81.451974	 loss=0.704923	 lr=0.001000
Epoch[056] Batch [0779]/[1203]	Speed: 141.897554 samples/sec	 accuracy=81.435897	 loss=0.705129	 lr=0.001000
Epoch[056] Batch [0799]/[1203]	Speed: 106.236167 samples/sec	 accuracy=81.440625	 loss=0.705200	 lr=0.001000
Epoch[056] Batch [0819]/[1203]	Speed: 138.120364 samples/sec	 accuracy=81.453659	 loss=0.704532	 lr=0.001000
Epoch[056] Batch [0839]/[1203]	Speed: 124.470549 samples/sec	 accuracy=81.425000	 loss=0.705308	 lr=0.001000
Epoch[056] Batch [0859]/[1203]	Speed: 111.299567 samples/sec	 accuracy=81.430233	 loss=0.705674	 lr=0.001000
Epoch[056] Batch [0879]/[1203]	Speed: 163.442549 samples/sec	 accuracy=81.400568	 loss=0.706591	 lr=0.001000
Epoch[056] Batch [0899]/[1203]	Speed: 108.866936 samples/sec	 accuracy=81.400000	 loss=0.707109	 lr=0.001000
Epoch[056] Batch [0919]/[1203]	Speed: 141.120225 samples/sec	 accuracy=81.411957	 loss=0.706550	 lr=0.001000
Epoch[056] Batch [0939]/[1203]	Speed: 114.975553 samples/sec	 accuracy=81.415426	 loss=0.706813	 lr=0.001000
Epoch[056] Batch [0959]/[1203]	Speed: 112.501928 samples/sec	 accuracy=81.406250	 loss=0.707178	 lr=0.001000
Epoch[056] Batch [0979]/[1203]	Speed: 179.472008 samples/sec	 accuracy=81.411735	 loss=0.706470	 lr=0.001000
Epoch[056] Batch [0999]/[1203]	Speed: 107.781733 samples/sec	 accuracy=81.414000	 loss=0.706867	 lr=0.001000
Epoch[056] Batch [1019]/[1203]	Speed: 137.916136 samples/sec	 accuracy=81.403431	 loss=0.707082	 lr=0.001000
Epoch[056] Batch [1039]/[1203]	Speed: 115.035988 samples/sec	 accuracy=81.403365	 loss=0.707239	 lr=0.001000
Epoch[056] Batch [1059]/[1203]	Speed: 139.452725 samples/sec	 accuracy=81.421226	 loss=0.706941	 lr=0.001000
Epoch[056] Batch [1079]/[1203]	Speed: 127.988708 samples/sec	 accuracy=81.434722	 loss=0.706868	 lr=0.001000
Epoch[056] Batch [1099]/[1203]	Speed: 114.710616 samples/sec	 accuracy=81.439091	 loss=0.706840	 lr=0.001000
Epoch[056] Batch [1119]/[1203]	Speed: 157.282987 samples/sec	 accuracy=81.442857	 loss=0.707045	 lr=0.001000
Epoch[056] Batch [1139]/[1203]	Speed: 112.042247 samples/sec	 accuracy=81.434211	 loss=0.707449	 lr=0.001000
Epoch[056] Batch [1159]/[1203]	Speed: 116.615142 samples/sec	 accuracy=81.438793	 loss=0.707440	 lr=0.001000
Epoch[056] Batch [1179]/[1203]	Speed: 160.204390 samples/sec	 accuracy=81.436017	 loss=0.707586	 lr=0.001000
Epoch[056] Batch [1199]/[1203]	Speed: 211.943644 samples/sec	 accuracy=81.435833	 loss=0.707908	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.400000 acc-top5=85.225000
Batch [0039]/[0097]: acc-top1=63.437500 acc-top5=84.637500
Batch [0059]/[0097]: acc-top1=65.516667 acc-top5=85.991667
Batch [0079]/[0097]: acc-top1=65.756250 acc-top5=85.818750
[Epoch 056] training: accuracy=81.429759	 loss=0.707973
[Epoch 056] speed: 125 samples/sec	time cost: 2101.370272
[Epoch 056] validation: acc-top1=65.190722 acc-top5=85.757732 loss=1.539385
Epoch[057] Batch [0019]/[1203]	Speed: 69.848697 samples/sec	 accuracy=81.600000	 loss=0.726902	 lr=0.001000
Epoch[057] Batch [0039]/[1203]	Speed: 103.634260 samples/sec	 accuracy=81.787500	 loss=0.716386	 lr=0.001000
Epoch[057] Batch [0059]/[1203]	Speed: 215.587277 samples/sec	 accuracy=81.616667	 loss=0.705201	 lr=0.001000
Epoch[057] Batch [0079]/[1203]	Speed: 96.548726 samples/sec	 accuracy=81.593750	 loss=0.702014	 lr=0.001000
Epoch[057] Batch [0099]/[1203]	Speed: 130.388591 samples/sec	 accuracy=81.620000	 loss=0.702289	 lr=0.001000
Epoch[057] Batch [0119]/[1203]	Speed: 140.065681 samples/sec	 accuracy=81.616667	 loss=0.701727	 lr=0.001000
Epoch[057] Batch [0139]/[1203]	Speed: 107.881856 samples/sec	 accuracy=81.610714	 loss=0.703410	 lr=0.001000
Epoch[057] Batch [0159]/[1203]	Speed: 194.869222 samples/sec	 accuracy=81.750000	 loss=0.698941	 lr=0.001000
Epoch[057] Batch [0179]/[1203]	Speed: 97.055634 samples/sec	 accuracy=81.738889	 loss=0.698899	 lr=0.001000
Epoch[057] Batch [0199]/[1203]	Speed: 125.377800 samples/sec	 accuracy=81.702500	 loss=0.700516	 lr=0.001000
Epoch[057] Batch [0219]/[1203]	Speed: 150.427199 samples/sec	 accuracy=81.806818	 loss=0.698509	 lr=0.001000
Epoch[057] Batch [0239]/[1203]	Speed: 107.925998 samples/sec	 accuracy=81.775000	 loss=0.699731	 lr=0.001000
Epoch[057] Batch [0259]/[1203]	Speed: 178.731506 samples/sec	 accuracy=81.790385	 loss=0.698453	 lr=0.001000
Epoch[057] Batch [0279]/[1203]	Speed: 94.156873 samples/sec	 accuracy=81.823214	 loss=0.696680	 lr=0.001000
Epoch[057] Batch [0299]/[1203]	Speed: 148.102375 samples/sec	 accuracy=81.701667	 loss=0.698592	 lr=0.001000
Epoch[057] Batch [0319]/[1203]	Speed: 128.031001 samples/sec	 accuracy=81.725000	 loss=0.698514	 lr=0.001000
Epoch[057] Batch [0339]/[1203]	Speed: 113.489750 samples/sec	 accuracy=81.689706	 loss=0.699028	 lr=0.001000
Epoch[057] Batch [0359]/[1203]	Speed: 121.012032 samples/sec	 accuracy=81.629167	 loss=0.700784	 lr=0.001000
Epoch[057] Batch [0379]/[1203]	Speed: 125.568281 samples/sec	 accuracy=81.631579	 loss=0.701816	 lr=0.001000
Epoch[057] Batch [0399]/[1203]	Speed: 131.217660 samples/sec	 accuracy=81.570000	 loss=0.703164	 lr=0.001000
Epoch[057] Batch [0419]/[1203]	Speed: 139.682110 samples/sec	 accuracy=81.559524	 loss=0.702383	 lr=0.001000
Epoch[057] Batch [0439]/[1203]	Speed: 110.890281 samples/sec	 accuracy=81.539773	 loss=0.702858	 lr=0.001000
Epoch[057] Batch [0459]/[1203]	Speed: 133.450729 samples/sec	 accuracy=81.558696	 loss=0.702619	 lr=0.001000
Epoch[057] Batch [0479]/[1203]	Speed: 127.859388 samples/sec	 accuracy=81.578125	 loss=0.701847	 lr=0.001000
Epoch[057] Batch [0499]/[1203]	Speed: 120.110929 samples/sec	 accuracy=81.548000	 loss=0.702232	 lr=0.001000
Epoch[057] Batch [0519]/[1203]	Speed: 140.780073 samples/sec	 accuracy=81.577885	 loss=0.701856	 lr=0.001000
Epoch[057] Batch [0539]/[1203]	Speed: 112.953929 samples/sec	 accuracy=81.619444	 loss=0.700565	 lr=0.001000
Epoch[057] Batch [0559]/[1203]	Speed: 140.418885 samples/sec	 accuracy=81.633036	 loss=0.700028	 lr=0.001000
Epoch[057] Batch [0579]/[1203]	Speed: 111.604739 samples/sec	 accuracy=81.625000	 loss=0.700244	 lr=0.001000
Epoch[057] Batch [0599]/[1203]	Speed: 130.464042 samples/sec	 accuracy=81.610000	 loss=0.700778	 lr=0.001000
Epoch[057] Batch [0619]/[1203]	Speed: 141.292608 samples/sec	 accuracy=81.591935	 loss=0.700963	 lr=0.001000
Epoch[057] Batch [0639]/[1203]	Speed: 109.499314 samples/sec	 accuracy=81.634375	 loss=0.700307	 lr=0.001000
Epoch[057] Batch [0659]/[1203]	Speed: 142.842843 samples/sec	 accuracy=81.601515	 loss=0.701405	 lr=0.001000
Epoch[057] Batch [0679]/[1203]	Speed: 125.328799 samples/sec	 accuracy=81.588971	 loss=0.701896	 lr=0.001000
Epoch[057] Batch [0699]/[1203]	Speed: 122.272993 samples/sec	 accuracy=81.573571	 loss=0.702282	 lr=0.001000
Epoch[057] Batch [0719]/[1203]	Speed: 116.090470 samples/sec	 accuracy=81.565278	 loss=0.702200	 lr=0.001000
Epoch[057] Batch [0739]/[1203]	Speed: 127.893579 samples/sec	 accuracy=81.560135	 loss=0.702697	 lr=0.001000
Epoch[057] Batch [0759]/[1203]	Speed: 134.405101 samples/sec	 accuracy=81.582237	 loss=0.702698	 lr=0.001000
Epoch[057] Batch [0779]/[1203]	Speed: 124.171497 samples/sec	 accuracy=81.580128	 loss=0.702597	 lr=0.001000
Epoch[057] Batch [0799]/[1203]	Speed: 133.690732 samples/sec	 accuracy=81.570000	 loss=0.702520	 lr=0.001000
Epoch[057] Batch [0819]/[1203]	Speed: 127.830024 samples/sec	 accuracy=81.555488	 loss=0.702761	 lr=0.001000
Epoch[057] Batch [0839]/[1203]	Speed: 114.032599 samples/sec	 accuracy=81.549405	 loss=0.702971	 lr=0.001000
Epoch[057] Batch [0859]/[1203]	Speed: 138.073252 samples/sec	 accuracy=81.527326	 loss=0.703661	 lr=0.001000
Epoch[057] Batch [0879]/[1203]	Speed: 120.436829 samples/sec	 accuracy=81.519886	 loss=0.703616	 lr=0.001000
Epoch[057] Batch [0899]/[1203]	Speed: 107.944175 samples/sec	 accuracy=81.520556	 loss=0.703491	 lr=0.001000
Epoch[057] Batch [0919]/[1203]	Speed: 158.771020 samples/sec	 accuracy=81.533152	 loss=0.703219	 lr=0.001000
Epoch[057] Batch [0939]/[1203]	Speed: 114.413086 samples/sec	 accuracy=81.532447	 loss=0.703105	 lr=0.001000
Epoch[057] Batch [0959]/[1203]	Speed: 123.237228 samples/sec	 accuracy=81.521875	 loss=0.703266	 lr=0.001000
Epoch[057] Batch [0979]/[1203]	Speed: 132.988396 samples/sec	 accuracy=81.508163	 loss=0.703263	 lr=0.001000
Epoch[057] Batch [0999]/[1203]	Speed: 109.178710 samples/sec	 accuracy=81.519500	 loss=0.703119	 lr=0.001000
Epoch[057] Batch [1019]/[1203]	Speed: 149.349169 samples/sec	 accuracy=81.517647	 loss=0.703317	 lr=0.001000
Epoch[057] Batch [1039]/[1203]	Speed: 125.241651 samples/sec	 accuracy=81.514904	 loss=0.703635	 lr=0.001000
Epoch[057] Batch [1059]/[1203]	Speed: 121.774990 samples/sec	 accuracy=81.509434	 loss=0.703776	 lr=0.001000
Epoch[057] Batch [1079]/[1203]	Speed: 143.131683 samples/sec	 accuracy=81.513426	 loss=0.704049	 lr=0.001000
Epoch[057] Batch [1099]/[1203]	Speed: 109.175742 samples/sec	 accuracy=81.520455	 loss=0.703745	 lr=0.001000
Epoch[057] Batch [1119]/[1203]	Speed: 141.087291 samples/sec	 accuracy=81.509821	 loss=0.703598	 lr=0.001000
Epoch[057] Batch [1139]/[1203]	Speed: 122.636245 samples/sec	 accuracy=81.514912	 loss=0.703596	 lr=0.001000
Epoch[057] Batch [1159]/[1203]	Speed: 126.491572 samples/sec	 accuracy=81.514224	 loss=0.704074	 lr=0.001000
Epoch[057] Batch [1179]/[1203]	Speed: 127.187508 samples/sec	 accuracy=81.511017	 loss=0.703994	 lr=0.001000
Epoch[057] Batch [1199]/[1203]	Speed: 244.916983 samples/sec	 accuracy=81.497083	 loss=0.704328	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.425000 acc-top5=85.025000
Batch [0039]/[0097]: acc-top1=63.475000 acc-top5=84.762500
Batch [0059]/[0097]: acc-top1=65.583333 acc-top5=86.075000
Batch [0079]/[0097]: acc-top1=65.625000 acc-top5=85.981250
[Epoch 057] training: accuracy=81.500416	 loss=0.704086
[Epoch 057] speed: 125 samples/sec	time cost: 2112.576003
[Epoch 057] validation: acc-top1=65.072165 acc-top5=85.840206 loss=1.540931
Epoch[058] Batch [0019]/[1203]	Speed: 70.707121 samples/sec	 accuracy=81.575000	 loss=0.701007	 lr=0.001000
Epoch[058] Batch [0039]/[1203]	Speed: 100.479627 samples/sec	 accuracy=81.912500	 loss=0.684362	 lr=0.001000
Epoch[058] Batch [0059]/[1203]	Speed: 241.770669 samples/sec	 accuracy=81.916667	 loss=0.690490	 lr=0.001000
Epoch[058] Batch [0079]/[1203]	Speed: 96.507174 samples/sec	 accuracy=81.850000	 loss=0.690279	 lr=0.001000
Epoch[058] Batch [0099]/[1203]	Speed: 122.087693 samples/sec	 accuracy=82.010000	 loss=0.690253	 lr=0.001000
Epoch[058] Batch [0119]/[1203]	Speed: 143.823015 samples/sec	 accuracy=81.970833	 loss=0.689352	 lr=0.001000
Epoch[058] Batch [0139]/[1203]	Speed: 104.369200 samples/sec	 accuracy=81.985714	 loss=0.687491	 lr=0.001000
Epoch[058] Batch [0159]/[1203]	Speed: 202.087966 samples/sec	 accuracy=82.106250	 loss=0.683810	 lr=0.001000
Epoch[058] Batch [0179]/[1203]	Speed: 102.522677 samples/sec	 accuracy=82.016667	 loss=0.686687	 lr=0.001000
Epoch[058] Batch [0199]/[1203]	Speed: 107.681534 samples/sec	 accuracy=82.020000	 loss=0.686474	 lr=0.001000
Epoch[058] Batch [0219]/[1203]	Speed: 175.761120 samples/sec	 accuracy=81.954545	 loss=0.687675	 lr=0.001000
Epoch[058] Batch [0239]/[1203]	Speed: 108.223239 samples/sec	 accuracy=82.002083	 loss=0.687659	 lr=0.001000
Epoch[058] Batch [0259]/[1203]	Speed: 178.899860 samples/sec	 accuracy=82.007692	 loss=0.685811	 lr=0.001000
Epoch[058] Batch [0279]/[1203]	Speed: 104.996451 samples/sec	 accuracy=81.994643	 loss=0.686007	 lr=0.001000
Epoch[058] Batch [0299]/[1203]	Speed: 117.607035 samples/sec	 accuracy=81.948333	 loss=0.687466	 lr=0.001000
Epoch[058] Batch [0319]/[1203]	Speed: 151.273818 samples/sec	 accuracy=81.907813	 loss=0.688920	 lr=0.001000
Epoch[058] Batch [0339]/[1203]	Speed: 120.910583 samples/sec	 accuracy=81.963235	 loss=0.687454	 lr=0.001000
Epoch[058] Batch [0359]/[1203]	Speed: 125.291150 samples/sec	 accuracy=81.959722	 loss=0.688476	 lr=0.001000
Epoch[058] Batch [0379]/[1203]	Speed: 122.677780 samples/sec	 accuracy=81.928947	 loss=0.688842	 lr=0.001000
Epoch[058] Batch [0399]/[1203]	Speed: 119.782483 samples/sec	 accuracy=81.886250	 loss=0.689937	 lr=0.001000
Epoch[058] Batch [0419]/[1203]	Speed: 141.130344 samples/sec	 accuracy=81.877381	 loss=0.690546	 lr=0.001000
Epoch[058] Batch [0439]/[1203]	Speed: 108.353036 samples/sec	 accuracy=81.834091	 loss=0.691602	 lr=0.001000
Epoch[058] Batch [0459]/[1203]	Speed: 139.484160 samples/sec	 accuracy=81.751087	 loss=0.694786	 lr=0.001000
Epoch[058] Batch [0479]/[1203]	Speed: 123.154847 samples/sec	 accuracy=81.736458	 loss=0.695107	 lr=0.001000
Epoch[058] Batch [0499]/[1203]	Speed: 128.596517 samples/sec	 accuracy=81.762000	 loss=0.694144	 lr=0.001000
Epoch[058] Batch [0519]/[1203]	Speed: 136.311120 samples/sec	 accuracy=81.758654	 loss=0.693621	 lr=0.001000
Epoch[058] Batch [0539]/[1203]	Speed: 106.996486 samples/sec	 accuracy=81.742593	 loss=0.694176	 lr=0.001000
Epoch[058] Batch [0559]/[1203]	Speed: 143.950381 samples/sec	 accuracy=81.743750	 loss=0.694509	 lr=0.001000
Epoch[058] Batch [0579]/[1203]	Speed: 115.914435 samples/sec	 accuracy=81.733621	 loss=0.695430	 lr=0.001000
Epoch[058] Batch [0599]/[1203]	Speed: 138.678098 samples/sec	 accuracy=81.760000	 loss=0.694910	 lr=0.001000
Epoch[058] Batch [0619]/[1203]	Speed: 133.036180 samples/sec	 accuracy=81.754839	 loss=0.694321	 lr=0.001000
Epoch[058] Batch [0639]/[1203]	Speed: 117.166176 samples/sec	 accuracy=81.758594	 loss=0.694363	 lr=0.001000
Epoch[058] Batch [0659]/[1203]	Speed: 134.998643 samples/sec	 accuracy=81.775758	 loss=0.693993	 lr=0.001000
Epoch[058] Batch [0679]/[1203]	Speed: 115.899417 samples/sec	 accuracy=81.770588	 loss=0.693902	 lr=0.001000
Epoch[058] Batch [0699]/[1203]	Speed: 128.544582 samples/sec	 accuracy=81.752143	 loss=0.694505	 lr=0.001000
Epoch[058] Batch [0719]/[1203]	Speed: 138.397469 samples/sec	 accuracy=81.765972	 loss=0.694134	 lr=0.001000
Epoch[058] Batch [0739]/[1203]	Speed: 108.002911 samples/sec	 accuracy=81.741216	 loss=0.694947	 lr=0.001000
Epoch[058] Batch [0759]/[1203]	Speed: 143.676878 samples/sec	 accuracy=81.755921	 loss=0.694490	 lr=0.001000
Epoch[058] Batch [0779]/[1203]	Speed: 110.886909 samples/sec	 accuracy=81.731410	 loss=0.694705	 lr=0.001000
Epoch[058] Batch [0799]/[1203]	Speed: 139.686160 samples/sec	 accuracy=81.706875	 loss=0.695769	 lr=0.001000
Epoch[058] Batch [0819]/[1203]	Speed: 129.897491 samples/sec	 accuracy=81.689024	 loss=0.695966	 lr=0.001000
Epoch[058] Batch [0839]/[1203]	Speed: 106.665755 samples/sec	 accuracy=81.702381	 loss=0.695642	 lr=0.001000
Epoch[058] Batch [0859]/[1203]	Speed: 165.640628 samples/sec	 accuracy=81.684302	 loss=0.696298	 lr=0.001000
Epoch[058] Batch [0879]/[1203]	Speed: 114.465711 samples/sec	 accuracy=81.685227	 loss=0.696578	 lr=0.001000
Epoch[058] Batch [0899]/[1203]	Speed: 111.804571 samples/sec	 accuracy=81.684444	 loss=0.696578	 lr=0.001000
Epoch[058] Batch [0919]/[1203]	Speed: 159.864606 samples/sec	 accuracy=81.691848	 loss=0.696458	 lr=0.001000
Epoch[058] Batch [0939]/[1203]	Speed: 106.959516 samples/sec	 accuracy=81.684574	 loss=0.696883	 lr=0.001000
Epoch[058] Batch [0959]/[1203]	Speed: 129.379903 samples/sec	 accuracy=81.662500	 loss=0.697752	 lr=0.001000
Epoch[058] Batch [0979]/[1203]	Speed: 132.808036 samples/sec	 accuracy=81.642857	 loss=0.698435	 lr=0.001000
Epoch[058] Batch [0999]/[1203]	Speed: 125.969560 samples/sec	 accuracy=81.634000	 loss=0.698642	 lr=0.001000
Epoch[058] Batch [1019]/[1203]	Speed: 121.659166 samples/sec	 accuracy=81.645098	 loss=0.698192	 lr=0.001000
Epoch[058] Batch [1039]/[1203]	Speed: 117.828915 samples/sec	 accuracy=81.637500	 loss=0.698452	 lr=0.001000
Epoch[058] Batch [1059]/[1203]	Speed: 136.603398 samples/sec	 accuracy=81.643868	 loss=0.698172	 lr=0.001000
Epoch[058] Batch [1079]/[1203]	Speed: 138.119613 samples/sec	 accuracy=81.627315	 loss=0.698915	 lr=0.001000
Epoch[058] Batch [1099]/[1203]	Speed: 116.670286 samples/sec	 accuracy=81.627273	 loss=0.698990	 lr=0.001000
Epoch[058] Batch [1119]/[1203]	Speed: 130.492748 samples/sec	 accuracy=81.607589	 loss=0.699584	 lr=0.001000
Epoch[058] Batch [1139]/[1203]	Speed: 111.868623 samples/sec	 accuracy=81.593421	 loss=0.700112	 lr=0.001000
Epoch[058] Batch [1159]/[1203]	Speed: 149.103239 samples/sec	 accuracy=81.594828	 loss=0.700037	 lr=0.001000
Epoch[058] Batch [1179]/[1203]	Speed: 131.454796 samples/sec	 accuracy=81.594915	 loss=0.700156	 lr=0.001000
Epoch[058] Batch [1199]/[1203]	Speed: 210.888049 samples/sec	 accuracy=81.584583	 loss=0.700103	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.475000 acc-top5=84.950000
Batch [0039]/[0097]: acc-top1=63.225000 acc-top5=84.725000
Batch [0059]/[0097]: acc-top1=65.516667 acc-top5=85.991667
Batch [0079]/[0097]: acc-top1=65.775000 acc-top5=85.800000
[Epoch 058] training: accuracy=81.586866	 loss=0.699957
[Epoch 058] speed: 125 samples/sec	time cost: 2113.460667
[Epoch 058] validation: acc-top1=65.247423 acc-top5=85.685567 loss=1.540835
Epoch[059] Batch [0019]/[1203]	Speed: 71.267516 samples/sec	 accuracy=82.200000	 loss=0.673171	 lr=0.001000
Epoch[059] Batch [0039]/[1203]	Speed: 101.093977 samples/sec	 accuracy=82.475000	 loss=0.658566	 lr=0.001000
Epoch[059] Batch [0059]/[1203]	Speed: 224.819254 samples/sec	 accuracy=82.166667	 loss=0.677060	 lr=0.001000
Epoch[059] Batch [0079]/[1203]	Speed: 100.700123 samples/sec	 accuracy=82.156250	 loss=0.678101	 lr=0.001000
Epoch[059] Batch [0099]/[1203]	Speed: 123.393481 samples/sec	 accuracy=82.205000	 loss=0.678245	 lr=0.001000
Epoch[059] Batch [0119]/[1203]	Speed: 151.662423 samples/sec	 accuracy=82.062500	 loss=0.681992	 lr=0.001000
Epoch[059] Batch [0139]/[1203]	Speed: 103.731824 samples/sec	 accuracy=82.046429	 loss=0.682331	 lr=0.001000
Epoch[059] Batch [0159]/[1203]	Speed: 196.631456 samples/sec	 accuracy=82.150000	 loss=0.679368	 lr=0.001000
Epoch[059] Batch [0179]/[1203]	Speed: 98.920918 samples/sec	 accuracy=82.258333	 loss=0.680316	 lr=0.001000
Epoch[059] Batch [0199]/[1203]	Speed: 112.682770 samples/sec	 accuracy=82.207500	 loss=0.682258	 lr=0.001000
Epoch[059] Batch [0219]/[1203]	Speed: 168.660844 samples/sec	 accuracy=82.170455	 loss=0.683620	 lr=0.001000
Epoch[059] Batch [0239]/[1203]	Speed: 107.358978 samples/sec	 accuracy=82.162500	 loss=0.683393	 lr=0.001000
Epoch[059] Batch [0259]/[1203]	Speed: 161.419429 samples/sec	 accuracy=82.121154	 loss=0.683519	 lr=0.001000
Epoch[059] Batch [0279]/[1203]	Speed: 101.912703 samples/sec	 accuracy=82.171429	 loss=0.683070	 lr=0.001000
Epoch[059] Batch [0299]/[1203]	Speed: 124.251982 samples/sec	 accuracy=82.170000	 loss=0.681307	 lr=0.001000
Epoch[059] Batch [0319]/[1203]	Speed: 146.350535 samples/sec	 accuracy=82.168750	 loss=0.682232	 lr=0.001000
Epoch[059] Batch [0339]/[1203]	Speed: 111.257978 samples/sec	 accuracy=82.161765	 loss=0.681539	 lr=0.001000
Epoch[059] Batch [0359]/[1203]	Speed: 140.283255 samples/sec	 accuracy=82.172222	 loss=0.680933	 lr=0.001000
Epoch[059] Batch [0379]/[1203]	Speed: 133.265547 samples/sec	 accuracy=82.121053	 loss=0.682549	 lr=0.001000
Epoch[059] Batch [0399]/[1203]	Speed: 99.703649 samples/sec	 accuracy=82.093750	 loss=0.683410	 lr=0.001000
Epoch[059] Batch [0419]/[1203]	Speed: 164.894303 samples/sec	 accuracy=82.033333	 loss=0.684521	 lr=0.001000
Epoch[059] Batch [0439]/[1203]	Speed: 113.316043 samples/sec	 accuracy=82.003409	 loss=0.685799	 lr=0.001000
Epoch[059] Batch [0459]/[1203]	Speed: 140.340408 samples/sec	 accuracy=82.052174	 loss=0.684747	 lr=0.001000
Epoch[059] Batch [0479]/[1203]	Speed: 127.537013 samples/sec	 accuracy=82.012500	 loss=0.685402	 lr=0.001000
Epoch[059] Batch [0499]/[1203]	Speed: 121.355698 samples/sec	 accuracy=82.000000	 loss=0.686748	 lr=0.001000
Epoch[059] Batch [0519]/[1203]	Speed: 138.913364 samples/sec	 accuracy=82.025000	 loss=0.686163	 lr=0.001000
Epoch[059] Batch [0539]/[1203]	Speed: 108.256735 samples/sec	 accuracy=82.012037	 loss=0.687445	 lr=0.001000
Epoch[059] Batch [0559]/[1203]	Speed: 131.689221 samples/sec	 accuracy=81.992857	 loss=0.687969	 lr=0.001000
Epoch[059] Batch [0579]/[1203]	Speed: 125.529697 samples/sec	 accuracy=81.973276	 loss=0.688916	 lr=0.001000
Epoch[059] Batch [0599]/[1203]	Speed: 122.596564 samples/sec	 accuracy=81.948333	 loss=0.689103	 lr=0.001000
Epoch[059] Batch [0619]/[1203]	Speed: 155.618859 samples/sec	 accuracy=81.969355	 loss=0.688452	 lr=0.001000
Epoch[059] Batch [0639]/[1203]	Speed: 102.234945 samples/sec	 accuracy=81.965625	 loss=0.689315	 lr=0.001000
Epoch[059] Batch [0659]/[1203]	Speed: 157.194796 samples/sec	 accuracy=81.961364	 loss=0.689589	 lr=0.001000
Epoch[059] Batch [0679]/[1203]	Speed: 112.949705 samples/sec	 accuracy=81.956618	 loss=0.689575	 lr=0.001000
Epoch[059] Batch [0699]/[1203]	Speed: 128.878912 samples/sec	 accuracy=81.928571	 loss=0.689993	 lr=0.001000
Epoch[059] Batch [0719]/[1203]	Speed: 140.735712 samples/sec	 accuracy=81.930556	 loss=0.689841	 lr=0.001000
Epoch[059] Batch [0739]/[1203]	Speed: 103.824909 samples/sec	 accuracy=81.924324	 loss=0.689693	 lr=0.001000
Epoch[059] Batch [0759]/[1203]	Speed: 126.847462 samples/sec	 accuracy=81.915789	 loss=0.690034	 lr=0.001000
Epoch[059] Batch [0779]/[1203]	Speed: 151.010076 samples/sec	 accuracy=81.893590	 loss=0.690924	 lr=0.001000
Epoch[059] Batch [0799]/[1203]	Speed: 112.234893 samples/sec	 accuracy=81.881250	 loss=0.691413	 lr=0.001000
Epoch[059] Batch [0819]/[1203]	Speed: 146.541934 samples/sec	 accuracy=81.899390	 loss=0.690865	 lr=0.001000
Epoch[059] Batch [0839]/[1203]	Speed: 109.842194 samples/sec	 accuracy=81.900595	 loss=0.690975	 lr=0.001000
Epoch[059] Batch [0859]/[1203]	Speed: 126.287359 samples/sec	 accuracy=81.911628	 loss=0.690716	 lr=0.001000
Epoch[059] Batch [0879]/[1203]	Speed: 158.943562 samples/sec	 accuracy=81.893182	 loss=0.691066	 lr=0.001000
Epoch[059] Batch [0899]/[1203]	Speed: 93.960731 samples/sec	 accuracy=81.895556	 loss=0.691079	 lr=0.001000
Epoch[059] Batch [0919]/[1203]	Speed: 137.737166 samples/sec	 accuracy=81.887500	 loss=0.691263	 lr=0.001000
Epoch[059] Batch [0939]/[1203]	Speed: 135.093978 samples/sec	 accuracy=81.882447	 loss=0.691514	 lr=0.001000
Epoch[059] Batch [0959]/[1203]	Speed: 137.848892 samples/sec	 accuracy=81.875521	 loss=0.691629	 lr=0.001000
Epoch[059] Batch [0979]/[1203]	Speed: 141.236253 samples/sec	 accuracy=81.848469	 loss=0.691863	 lr=0.001000
Epoch[059] Batch [0999]/[1203]	Speed: 90.697293 samples/sec	 accuracy=81.857000	 loss=0.691710	 lr=0.001000
Epoch[059] Batch [1019]/[1203]	Speed: 167.796355 samples/sec	 accuracy=81.846569	 loss=0.691854	 lr=0.001000
Epoch[059] Batch [1039]/[1203]	Speed: 118.752374 samples/sec	 accuracy=81.821635	 loss=0.692483	 lr=0.001000
Epoch[059] Batch [1059]/[1203]	Speed: 117.987852 samples/sec	 accuracy=81.802358	 loss=0.693289	 lr=0.001000
Epoch[059] Batch [1079]/[1203]	Speed: 155.982122 samples/sec	 accuracy=81.772685	 loss=0.693713	 lr=0.001000
Epoch[059] Batch [1099]/[1203]	Speed: 91.664700 samples/sec	 accuracy=81.767273	 loss=0.693964	 lr=0.001000
Epoch[059] Batch [1119]/[1203]	Speed: 188.711412 samples/sec	 accuracy=81.750000	 loss=0.694358	 lr=0.001000
Epoch[059] Batch [1139]/[1203]	Speed: 111.105470 samples/sec	 accuracy=81.735526	 loss=0.694868	 lr=0.001000
Epoch[059] Batch [1159]/[1203]	Speed: 125.142895 samples/sec	 accuracy=81.737931	 loss=0.694558	 lr=0.001000
Epoch[059] Batch [1179]/[1203]	Speed: 137.388816 samples/sec	 accuracy=81.740678	 loss=0.694658	 lr=0.001000
Epoch[059] Batch [1199]/[1203]	Speed: 229.909068 samples/sec	 accuracy=81.730417	 loss=0.695122	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.325000 acc-top5=85.250000
Batch [0039]/[0097]: acc-top1=63.000000 acc-top5=84.875000
Batch [0059]/[0097]: acc-top1=65.291667 acc-top5=86.066667
Batch [0079]/[0097]: acc-top1=65.618750 acc-top5=85.987500
[Epoch 059] training: accuracy=81.728180	 loss=0.695339
[Epoch 059] speed: 125 samples/sec	time cost: 2110.476942
[Epoch 059] validation: acc-top1=65.144330 acc-top5=85.917526 loss=1.540726
Epoch[060] Batch [0019]/[1203]	Speed: 69.116461 samples/sec	 accuracy=81.475000	 loss=0.670884	 lr=0.001000
Epoch[060] Batch [0039]/[1203]	Speed: 101.605273 samples/sec	 accuracy=82.075000	 loss=0.665305	 lr=0.001000
Epoch[060] Batch [0059]/[1203]	Speed: 239.077800 samples/sec	 accuracy=81.850000	 loss=0.673221	 lr=0.001000
Epoch[060] Batch [0079]/[1203]	Speed: 94.053569 samples/sec	 accuracy=81.956250	 loss=0.672877	 lr=0.001000
Epoch[060] Batch [0099]/[1203]	Speed: 110.528015 samples/sec	 accuracy=82.000000	 loss=0.675942	 lr=0.001000
Epoch[060] Batch [0119]/[1203]	Speed: 193.473541 samples/sec	 accuracy=81.854167	 loss=0.683064	 lr=0.001000
Epoch[060] Batch [0139]/[1203]	Speed: 108.931229 samples/sec	 accuracy=81.867857	 loss=0.683052	 lr=0.001000
Epoch[060] Batch [0159]/[1203]	Speed: 163.898850 samples/sec	 accuracy=81.946875	 loss=0.680734	 lr=0.001000
Epoch[060] Batch [0179]/[1203]	Speed: 112.174774 samples/sec	 accuracy=81.958333	 loss=0.684954	 lr=0.001000
Epoch[060] Batch [0199]/[1203]	Speed: 110.242015 samples/sec	 accuracy=81.930000	 loss=0.685300	 lr=0.001000
Epoch[060] Batch [0219]/[1203]	Speed: 146.295887 samples/sec	 accuracy=81.997727	 loss=0.685701	 lr=0.001000
Epoch[060] Batch [0239]/[1203]	Speed: 116.131554 samples/sec	 accuracy=81.993750	 loss=0.684354	 lr=0.001000
Epoch[060] Batch [0259]/[1203]	Speed: 124.386099 samples/sec	 accuracy=81.984615	 loss=0.684284	 lr=0.001000
Epoch[060] Batch [0279]/[1203]	Speed: 143.265654 samples/sec	 accuracy=81.953571	 loss=0.685923	 lr=0.001000
Epoch[060] Batch [0299]/[1203]	Speed: 110.578481 samples/sec	 accuracy=81.901667	 loss=0.687692	 lr=0.001000
Epoch[060] Batch [0319]/[1203]	Speed: 135.969504 samples/sec	 accuracy=81.928125	 loss=0.687263	 lr=0.001000
Epoch[060] Batch [0339]/[1203]	Speed: 120.816944 samples/sec	 accuracy=81.947059	 loss=0.686799	 lr=0.001000
Epoch[060] Batch [0359]/[1203]	Speed: 141.708205 samples/sec	 accuracy=81.919444	 loss=0.687757	 lr=0.001000
Epoch[060] Batch [0379]/[1203]	Speed: 121.554058 samples/sec	 accuracy=81.938158	 loss=0.687348	 lr=0.001000
Epoch[060] Batch [0399]/[1203]	Speed: 111.848145 samples/sec	 accuracy=81.906250	 loss=0.688712	 lr=0.001000
Epoch[060] Batch [0419]/[1203]	Speed: 149.848748 samples/sec	 accuracy=81.886905	 loss=0.689191	 lr=0.001000
Epoch[060] Batch [0439]/[1203]	Speed: 122.622801 samples/sec	 accuracy=81.870455	 loss=0.689828	 lr=0.001000
Epoch[060] Batch [0459]/[1203]	Speed: 143.264395 samples/sec	 accuracy=81.839130	 loss=0.690930	 lr=0.001000
Epoch[060] Batch [0479]/[1203]	Speed: 115.851228 samples/sec	 accuracy=81.879167	 loss=0.690948	 lr=0.001000
Epoch[060] Batch [0499]/[1203]	Speed: 108.712727 samples/sec	 accuracy=81.868000	 loss=0.691111	 lr=0.001000
Epoch[060] Batch [0519]/[1203]	Speed: 155.254624 samples/sec	 accuracy=81.872115	 loss=0.691118	 lr=0.001000
Epoch[060] Batch [0539]/[1203]	Speed: 113.991711 samples/sec	 accuracy=81.850000	 loss=0.691946	 lr=0.001000
Epoch[060] Batch [0559]/[1203]	Speed: 125.303050 samples/sec	 accuracy=81.836607	 loss=0.692660	 lr=0.001000
Epoch[060] Batch [0579]/[1203]	Speed: 144.975978 samples/sec	 accuracy=81.820690	 loss=0.692876	 lr=0.001000
Epoch[060] Batch [0599]/[1203]	Speed: 102.244096 samples/sec	 accuracy=81.850833	 loss=0.691781	 lr=0.001000
Epoch[060] Batch [0619]/[1203]	Speed: 157.247980 samples/sec	 accuracy=81.842742	 loss=0.691839	 lr=0.001000
Epoch[060] Batch [0639]/[1203]	Speed: 116.446421 samples/sec	 accuracy=81.867969	 loss=0.691034	 lr=0.001000
Epoch[060] Batch [0659]/[1203]	Speed: 121.396972 samples/sec	 accuracy=81.850758	 loss=0.690997	 lr=0.001000
Epoch[060] Batch [0679]/[1203]	Speed: 138.400237 samples/sec	 accuracy=81.889706	 loss=0.690230	 lr=0.001000
Epoch[060] Batch [0699]/[1203]	Speed: 112.238728 samples/sec	 accuracy=81.860000	 loss=0.691326	 lr=0.001000
Epoch[060] Batch [0719]/[1203]	Speed: 126.240345 samples/sec	 accuracy=81.865972	 loss=0.690785	 lr=0.001000
Epoch[060] Batch [0739]/[1203]	Speed: 142.636395 samples/sec	 accuracy=81.875000	 loss=0.690587	 lr=0.001000
Epoch[060] Batch [0759]/[1203]	Speed: 110.145168 samples/sec	 accuracy=81.878289	 loss=0.689879	 lr=0.001000
Epoch[060] Batch [0779]/[1203]	Speed: 148.748760 samples/sec	 accuracy=81.842308	 loss=0.691201	 lr=0.001000
Epoch[060] Batch [0799]/[1203]	Speed: 114.753939 samples/sec	 accuracy=81.840000	 loss=0.691811	 lr=0.001000
Epoch[060] Batch [0819]/[1203]	Speed: 128.635003 samples/sec	 accuracy=81.862805	 loss=0.691529	 lr=0.001000
Epoch[060] Batch [0839]/[1203]	Speed: 144.126759 samples/sec	 accuracy=81.847619	 loss=0.691801	 lr=0.001000
Epoch[060] Batch [0859]/[1203]	Speed: 104.316429 samples/sec	 accuracy=81.844767	 loss=0.692383	 lr=0.001000
Epoch[060] Batch [0879]/[1203]	Speed: 138.154753 samples/sec	 accuracy=81.848295	 loss=0.692204	 lr=0.001000
Epoch[060] Batch [0899]/[1203]	Speed: 126.307701 samples/sec	 accuracy=81.841111	 loss=0.692107	 lr=0.001000
Epoch[060] Batch [0919]/[1203]	Speed: 136.107416 samples/sec	 accuracy=81.837500	 loss=0.692408	 lr=0.001000
Epoch[060] Batch [0939]/[1203]	Speed: 122.031663 samples/sec	 accuracy=81.846277	 loss=0.692129	 lr=0.001000
Epoch[060] Batch [0959]/[1203]	Speed: 106.225838 samples/sec	 accuracy=81.832292	 loss=0.692820	 lr=0.001000
Epoch[060] Batch [0979]/[1203]	Speed: 135.531845 samples/sec	 accuracy=81.833673	 loss=0.692895	 lr=0.001000
Epoch[060] Batch [0999]/[1203]	Speed: 142.709702 samples/sec	 accuracy=81.850000	 loss=0.692480	 lr=0.001000
Epoch[060] Batch [1019]/[1203]	Speed: 112.861936 samples/sec	 accuracy=81.861275	 loss=0.691954	 lr=0.001000
Epoch[060] Batch [1039]/[1203]	Speed: 143.331647 samples/sec	 accuracy=81.854327	 loss=0.692236	 lr=0.001000
Epoch[060] Batch [1059]/[1203]	Speed: 115.684310 samples/sec	 accuracy=81.856604	 loss=0.692492	 lr=0.001000
Epoch[060] Batch [1079]/[1203]	Speed: 117.842188 samples/sec	 accuracy=81.867130	 loss=0.692161	 lr=0.001000
Epoch[060] Batch [1099]/[1203]	Speed: 142.340960 samples/sec	 accuracy=81.844545	 loss=0.692734	 lr=0.001000
Epoch[060] Batch [1119]/[1203]	Speed: 126.515962 samples/sec	 accuracy=81.849554	 loss=0.692667	 lr=0.001000
Epoch[060] Batch [1139]/[1203]	Speed: 123.518118 samples/sec	 accuracy=81.837281	 loss=0.693212	 lr=0.001000
Epoch[060] Batch [1159]/[1203]	Speed: 115.831673 samples/sec	 accuracy=81.846121	 loss=0.693134	 lr=0.001000
Epoch[060] Batch [1179]/[1203]	Speed: 137.877719 samples/sec	 accuracy=81.835169	 loss=0.693145	 lr=0.001000
Epoch[060] Batch [1199]/[1203]	Speed: 270.303330 samples/sec	 accuracy=81.851667	 loss=0.693031	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.525000 acc-top5=85.150000
Batch [0039]/[0097]: acc-top1=63.312500 acc-top5=84.675000
Batch [0059]/[0097]: acc-top1=65.333333 acc-top5=85.875000
Batch [0079]/[0097]: acc-top1=65.506250 acc-top5=85.843750
[Epoch 060] training: accuracy=81.856193	 loss=0.692859
[Epoch 060] speed: 125 samples/sec	time cost: 2111.005842
[Epoch 060] validation: acc-top1=65.020619 acc-top5=85.783505 loss=1.547285
Epoch[061] Batch [0019]/[1203]	Speed: 69.448063 samples/sec	 accuracy=82.650000	 loss=0.673295	 lr=0.001000
Epoch[061] Batch [0039]/[1203]	Speed: 104.909186 samples/sec	 accuracy=82.125000	 loss=0.673156	 lr=0.001000
Epoch[061] Batch [0059]/[1203]	Speed: 207.192379 samples/sec	 accuracy=82.108333	 loss=0.680983	 lr=0.001000
Epoch[061] Batch [0079]/[1203]	Speed: 94.100501 samples/sec	 accuracy=82.181250	 loss=0.678539	 lr=0.001000
Epoch[061] Batch [0099]/[1203]	Speed: 138.149647 samples/sec	 accuracy=82.075000	 loss=0.684492	 lr=0.001000
Epoch[061] Batch [0119]/[1203]	Speed: 136.279320 samples/sec	 accuracy=82.004167	 loss=0.684174	 lr=0.001000
Epoch[061] Batch [0139]/[1203]	Speed: 107.464413 samples/sec	 accuracy=82.057143	 loss=0.680851	 lr=0.001000
Epoch[061] Batch [0159]/[1203]	Speed: 187.127941 samples/sec	 accuracy=82.021875	 loss=0.682402	 lr=0.001000
Epoch[061] Batch [0179]/[1203]	Speed: 96.420968 samples/sec	 accuracy=82.052778	 loss=0.682280	 lr=0.001000
Epoch[061] Batch [0199]/[1203]	Speed: 136.212271 samples/sec	 accuracy=81.952500	 loss=0.681679	 lr=0.001000
Epoch[061] Batch [0219]/[1203]	Speed: 140.254530 samples/sec	 accuracy=81.936364	 loss=0.682940	 lr=0.001000
Epoch[061] Batch [0239]/[1203]	Speed: 116.965503 samples/sec	 accuracy=81.954167	 loss=0.682236	 lr=0.001000
Epoch[061] Batch [0259]/[1203]	Speed: 139.284329 samples/sec	 accuracy=81.940385	 loss=0.683486	 lr=0.001000
Epoch[061] Batch [0279]/[1203]	Speed: 110.322429 samples/sec	 accuracy=81.978571	 loss=0.684178	 lr=0.001000
Epoch[061] Batch [0299]/[1203]	Speed: 125.941742 samples/sec	 accuracy=81.943333	 loss=0.684800	 lr=0.001000
Epoch[061] Batch [0319]/[1203]	Speed: 132.188870 samples/sec	 accuracy=81.975000	 loss=0.682751	 lr=0.001000
Epoch[061] Batch [0339]/[1203]	Speed: 133.530179 samples/sec	 accuracy=81.907353	 loss=0.686260	 lr=0.001000
Epoch[061] Batch [0359]/[1203]	Speed: 139.147081 samples/sec	 accuracy=81.902778	 loss=0.686352	 lr=0.001000
Epoch[061] Batch [0379]/[1203]	Speed: 103.945216 samples/sec	 accuracy=81.921053	 loss=0.686344	 lr=0.001000
Epoch[061] Batch [0399]/[1203]	Speed: 141.247220 samples/sec	 accuracy=81.958750	 loss=0.685874	 lr=0.001000
Epoch[061] Batch [0419]/[1203]	Speed: 127.802111 samples/sec	 accuracy=81.957143	 loss=0.685406	 lr=0.001000
Epoch[061] Batch [0439]/[1203]	Speed: 116.951023 samples/sec	 accuracy=81.973864	 loss=0.685665	 lr=0.001000
Epoch[061] Batch [0459]/[1203]	Speed: 133.926271 samples/sec	 accuracy=82.006522	 loss=0.685466	 lr=0.001000
Epoch[061] Batch [0479]/[1203]	Speed: 119.244518 samples/sec	 accuracy=82.021875	 loss=0.684454	 lr=0.001000
Epoch[061] Batch [0499]/[1203]	Speed: 146.321351 samples/sec	 accuracy=82.020000	 loss=0.683835	 lr=0.001000
Epoch[061] Batch [0519]/[1203]	Speed: 117.109243 samples/sec	 accuracy=82.009615	 loss=0.684728	 lr=0.001000
Epoch[061] Batch [0539]/[1203]	Speed: 119.234092 samples/sec	 accuracy=81.995370	 loss=0.685441	 lr=0.001000
Epoch[061] Batch [0559]/[1203]	Speed: 138.012872 samples/sec	 accuracy=81.966964	 loss=0.686703	 lr=0.001000
Epoch[061] Batch [0579]/[1203]	Speed: 113.781552 samples/sec	 accuracy=81.936207	 loss=0.687561	 lr=0.001000
Epoch[061] Batch [0599]/[1203]	Speed: 150.857451 samples/sec	 accuracy=81.936667	 loss=0.687879	 lr=0.001000
Epoch[061] Batch [0619]/[1203]	Speed: 113.896697 samples/sec	 accuracy=81.979032	 loss=0.686612	 lr=0.001000
Epoch[061] Batch [0639]/[1203]	Speed: 115.252498 samples/sec	 accuracy=81.947656	 loss=0.687186	 lr=0.001000
Epoch[061] Batch [0659]/[1203]	Speed: 150.322687 samples/sec	 accuracy=81.985606	 loss=0.686331	 lr=0.001000
Epoch[061] Batch [0679]/[1203]	Speed: 112.157199 samples/sec	 accuracy=81.986029	 loss=0.686692	 lr=0.001000
Epoch[061] Batch [0699]/[1203]	Speed: 136.913131 samples/sec	 accuracy=82.020000	 loss=0.685956	 lr=0.001000
Epoch[061] Batch [0719]/[1203]	Speed: 127.002548 samples/sec	 accuracy=82.036111	 loss=0.685579	 lr=0.001000
Epoch[061] Batch [0739]/[1203]	Speed: 117.334654 samples/sec	 accuracy=82.020946	 loss=0.685532	 lr=0.001000
Epoch[061] Batch [0759]/[1203]	Speed: 144.512567 samples/sec	 accuracy=82.027632	 loss=0.685442	 lr=0.001000
Epoch[061] Batch [0779]/[1203]	Speed: 115.344303 samples/sec	 accuracy=82.008974	 loss=0.685850	 lr=0.001000
Epoch[061] Batch [0799]/[1203]	Speed: 131.525315 samples/sec	 accuracy=82.013125	 loss=0.685648	 lr=0.001000
Epoch[061] Batch [0819]/[1203]	Speed: 113.732294 samples/sec	 accuracy=82.020732	 loss=0.686063	 lr=0.001000
Epoch[061] Batch [0839]/[1203]	Speed: 136.857862 samples/sec	 accuracy=82.026786	 loss=0.685877	 lr=0.001000
Epoch[061] Batch [0859]/[1203]	Speed: 144.553018 samples/sec	 accuracy=82.020930	 loss=0.685744	 lr=0.001000
Epoch[061] Batch [0879]/[1203]	Speed: 107.735660 samples/sec	 accuracy=82.022159	 loss=0.685468	 lr=0.001000
Epoch[061] Batch [0899]/[1203]	Speed: 132.192742 samples/sec	 accuracy=82.016111	 loss=0.685613	 lr=0.001000
Epoch[061] Batch [0919]/[1203]	Speed: 130.588553 samples/sec	 accuracy=82.016848	 loss=0.685887	 lr=0.001000
Epoch[061] Batch [0939]/[1203]	Speed: 118.996377 samples/sec	 accuracy=82.020213	 loss=0.685740	 lr=0.001000
Epoch[061] Batch [0959]/[1203]	Speed: 130.725591 samples/sec	 accuracy=82.031771	 loss=0.685464	 lr=0.001000
Epoch[061] Batch [0979]/[1203]	Speed: 119.195501 samples/sec	 accuracy=82.021429	 loss=0.685774	 lr=0.001000
Epoch[061] Batch [0999]/[1203]	Speed: 136.957381 samples/sec	 accuracy=82.001500	 loss=0.685992	 lr=0.001000
Epoch[061] Batch [1019]/[1203]	Speed: 126.408169 samples/sec	 accuracy=81.993137	 loss=0.686433	 lr=0.001000
Epoch[061] Batch [1039]/[1203]	Speed: 112.771480 samples/sec	 accuracy=81.989904	 loss=0.686514	 lr=0.001000
Epoch[061] Batch [1059]/[1203]	Speed: 143.474540 samples/sec	 accuracy=81.987264	 loss=0.686828	 lr=0.001000
Epoch[061] Batch [1079]/[1203]	Speed: 122.109135 samples/sec	 accuracy=81.968519	 loss=0.687383	 lr=0.001000
Epoch[061] Batch [1099]/[1203]	Speed: 130.314030 samples/sec	 accuracy=81.958182	 loss=0.687938	 lr=0.001000
Epoch[061] Batch [1119]/[1203]	Speed: 122.892548 samples/sec	 accuracy=81.951339	 loss=0.688468	 lr=0.001000
Epoch[061] Batch [1139]/[1203]	Speed: 118.645733 samples/sec	 accuracy=81.940789	 loss=0.688760	 lr=0.001000
Epoch[061] Batch [1159]/[1203]	Speed: 132.454832 samples/sec	 accuracy=81.946983	 loss=0.689047	 lr=0.001000
Epoch[061] Batch [1179]/[1203]	Speed: 128.445783 samples/sec	 accuracy=81.950000	 loss=0.688999	 lr=0.001000
Epoch[061] Batch [1199]/[1203]	Speed: 265.799870 samples/sec	 accuracy=81.949167	 loss=0.688999	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.675000 acc-top5=85.275000
Batch [0039]/[0097]: acc-top1=63.575000 acc-top5=84.675000
Batch [0059]/[0097]: acc-top1=65.650000 acc-top5=85.908333
Batch [0079]/[0097]: acc-top1=65.812500 acc-top5=85.787500
[Epoch 061] training: accuracy=81.953450	 loss=0.688820
[Epoch 061] speed: 125 samples/sec	time cost: 2110.603609
[Epoch 061] validation: acc-top1=65.175258 acc-top5=85.695876 loss=1.553017
Epoch[062] Batch [0019]/[1203]	Speed: 69.639970 samples/sec	 accuracy=81.725000	 loss=0.686455	 lr=0.001000
Epoch[062] Batch [0039]/[1203]	Speed: 92.598778 samples/sec	 accuracy=82.050000	 loss=0.680445	 lr=0.001000
Epoch[062] Batch [0059]/[1203]	Speed: 295.059849 samples/sec	 accuracy=82.433333	 loss=0.669504	 lr=0.001000
Epoch[062] Batch [0079]/[1203]	Speed: 92.283106 samples/sec	 accuracy=82.612500	 loss=0.667345	 lr=0.001000
Epoch[062] Batch [0099]/[1203]	Speed: 115.207990 samples/sec	 accuracy=82.440000	 loss=0.674608	 lr=0.001000
Epoch[062] Batch [0119]/[1203]	Speed: 190.319158 samples/sec	 accuracy=82.554167	 loss=0.669066	 lr=0.001000
Epoch[062] Batch [0139]/[1203]	Speed: 95.036635 samples/sec	 accuracy=82.385714	 loss=0.672637	 lr=0.001000
Epoch[062] Batch [0159]/[1203]	Speed: 238.953853 samples/sec	 accuracy=82.243750	 loss=0.677948	 lr=0.001000
Epoch[062] Batch [0179]/[1203]	Speed: 95.977800 samples/sec	 accuracy=82.400000	 loss=0.676031	 lr=0.001000
Epoch[062] Batch [0199]/[1203]	Speed: 115.385147 samples/sec	 accuracy=82.335000	 loss=0.678626	 lr=0.001000
Epoch[062] Batch [0219]/[1203]	Speed: 182.277253 samples/sec	 accuracy=82.272727	 loss=0.680617	 lr=0.001000
Epoch[062] Batch [0239]/[1203]	Speed: 99.227763 samples/sec	 accuracy=82.304167	 loss=0.679552	 lr=0.001000
Epoch[062] Batch [0259]/[1203]	Speed: 129.952698 samples/sec	 accuracy=82.315385	 loss=0.678496	 lr=0.001000
Epoch[062] Batch [0279]/[1203]	Speed: 134.568073 samples/sec	 accuracy=82.371429	 loss=0.675301	 lr=0.001000
Epoch[062] Batch [0299]/[1203]	Speed: 112.314760 samples/sec	 accuracy=82.340000	 loss=0.676820	 lr=0.001000
Epoch[062] Batch [0319]/[1203]	Speed: 177.467044 samples/sec	 accuracy=82.295312	 loss=0.678385	 lr=0.001000
Epoch[062] Batch [0339]/[1203]	Speed: 109.077980 samples/sec	 accuracy=82.319118	 loss=0.677124	 lr=0.001000
Epoch[062] Batch [0359]/[1203]	Speed: 120.716996 samples/sec	 accuracy=82.336111	 loss=0.677114	 lr=0.001000
Epoch[062] Batch [0379]/[1203]	Speed: 137.888156 samples/sec	 accuracy=82.303947	 loss=0.677390	 lr=0.001000
Epoch[062] Batch [0399]/[1203]	Speed: 104.631677 samples/sec	 accuracy=82.322500	 loss=0.677388	 lr=0.001000
Epoch[062] Batch [0419]/[1203]	Speed: 184.722646 samples/sec	 accuracy=82.292857	 loss=0.679502	 lr=0.001000
Epoch[062] Batch [0439]/[1203]	Speed: 109.250373 samples/sec	 accuracy=82.280682	 loss=0.679526	 lr=0.001000
Epoch[062] Batch [0459]/[1203]	Speed: 119.161168 samples/sec	 accuracy=82.270652	 loss=0.679140	 lr=0.001000
Epoch[062] Batch [0479]/[1203]	Speed: 136.268884 samples/sec	 accuracy=82.266667	 loss=0.679290	 lr=0.001000
Epoch[062] Batch [0499]/[1203]	Speed: 110.678885 samples/sec	 accuracy=82.261000	 loss=0.679485	 lr=0.001000
Epoch[062] Batch [0519]/[1203]	Speed: 166.351601 samples/sec	 accuracy=82.234615	 loss=0.679687	 lr=0.001000
Epoch[062] Batch [0539]/[1203]	Speed: 114.059513 samples/sec	 accuracy=82.217593	 loss=0.680832	 lr=0.001000
Epoch[062] Batch [0559]/[1203]	Speed: 101.435538 samples/sec	 accuracy=82.228571	 loss=0.680626	 lr=0.001000
Epoch[062] Batch [0579]/[1203]	Speed: 188.908093 samples/sec	 accuracy=82.246552	 loss=0.681402	 lr=0.001000
Epoch[062] Batch [0599]/[1203]	Speed: 105.186088 samples/sec	 accuracy=82.271667	 loss=0.681017	 lr=0.001000
Epoch[062] Batch [0619]/[1203]	Speed: 143.054175 samples/sec	 accuracy=82.237097	 loss=0.682219	 lr=0.001000
Epoch[062] Batch [0639]/[1203]	Speed: 126.989377 samples/sec	 accuracy=82.205469	 loss=0.682700	 lr=0.001000
Epoch[062] Batch [0659]/[1203]	Speed: 107.566260 samples/sec	 accuracy=82.188636	 loss=0.683665	 lr=0.001000
Epoch[062] Batch [0679]/[1203]	Speed: 147.320348 samples/sec	 accuracy=82.177941	 loss=0.683386	 lr=0.001000
Epoch[062] Batch [0699]/[1203]	Speed: 111.464359 samples/sec	 accuracy=82.181429	 loss=0.682847	 lr=0.001000
Epoch[062] Batch [0719]/[1203]	Speed: 135.231315 samples/sec	 accuracy=82.178472	 loss=0.683025	 lr=0.001000
Epoch[062] Batch [0739]/[1203]	Speed: 143.893043 samples/sec	 accuracy=82.164189	 loss=0.683388	 lr=0.001000
Epoch[062] Batch [0759]/[1203]	Speed: 105.527666 samples/sec	 accuracy=82.184211	 loss=0.682394	 lr=0.001000
Epoch[062] Batch [0779]/[1203]	Speed: 147.319675 samples/sec	 accuracy=82.186538	 loss=0.682145	 lr=0.001000
Epoch[062] Batch [0799]/[1203]	Speed: 111.192969 samples/sec	 accuracy=82.186250	 loss=0.681631	 lr=0.001000
Epoch[062] Batch [0819]/[1203]	Speed: 115.004537 samples/sec	 accuracy=82.184146	 loss=0.681836	 lr=0.001000
Epoch[062] Batch [0839]/[1203]	Speed: 166.869856 samples/sec	 accuracy=82.161310	 loss=0.682876	 lr=0.001000
Epoch[062] Batch [0859]/[1203]	Speed: 117.118593 samples/sec	 accuracy=82.152907	 loss=0.683217	 lr=0.001000
Epoch[062] Batch [0879]/[1203]	Speed: 127.651887 samples/sec	 accuracy=82.152841	 loss=0.682954	 lr=0.001000
Epoch[062] Batch [0899]/[1203]	Speed: 119.349568 samples/sec	 accuracy=82.136667	 loss=0.683261	 lr=0.001000
Epoch[062] Batch [0919]/[1203]	Speed: 135.927205 samples/sec	 accuracy=82.121196	 loss=0.683838	 lr=0.001000
Epoch[062] Batch [0939]/[1203]	Speed: 130.453779 samples/sec	 accuracy=82.102128	 loss=0.683947	 lr=0.001000
Epoch[062] Batch [0959]/[1203]	Speed: 112.425074 samples/sec	 accuracy=82.088542	 loss=0.684057	 lr=0.001000
Epoch[062] Batch [0979]/[1203]	Speed: 124.139344 samples/sec	 accuracy=82.098980	 loss=0.683646	 lr=0.001000
Epoch[062] Batch [0999]/[1203]	Speed: 129.340446 samples/sec	 accuracy=82.094000	 loss=0.684029	 lr=0.001000
Epoch[062] Batch [1019]/[1203]	Speed: 132.808338 samples/sec	 accuracy=82.086765	 loss=0.684115	 lr=0.001000
Epoch[062] Batch [1039]/[1203]	Speed: 144.596714 samples/sec	 accuracy=82.071154	 loss=0.684567	 lr=0.001000
Epoch[062] Batch [1059]/[1203]	Speed: 105.393092 samples/sec	 accuracy=82.075943	 loss=0.684211	 lr=0.001000
Epoch[062] Batch [1079]/[1203]	Speed: 137.709367 samples/sec	 accuracy=82.069907	 loss=0.684404	 lr=0.001000
Epoch[062] Batch [1099]/[1203]	Speed: 121.795565 samples/sec	 accuracy=82.094091	 loss=0.684196	 lr=0.001000
Epoch[062] Batch [1119]/[1203]	Speed: 123.756595 samples/sec	 accuracy=82.076339	 loss=0.684649	 lr=0.001000
Epoch[062] Batch [1139]/[1203]	Speed: 131.964691 samples/sec	 accuracy=82.082456	 loss=0.684685	 lr=0.001000
Epoch[062] Batch [1159]/[1203]	Speed: 116.598600 samples/sec	 accuracy=82.073707	 loss=0.684868	 lr=0.001000
Epoch[062] Batch [1179]/[1203]	Speed: 152.218788 samples/sec	 accuracy=82.084746	 loss=0.684596	 lr=0.001000
Epoch[062] Batch [1199]/[1203]	Speed: 216.580448 samples/sec	 accuracy=82.075000	 loss=0.684911	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.250000 acc-top5=85.350000
Batch [0039]/[0097]: acc-top1=63.200000 acc-top5=84.937500
Batch [0059]/[0097]: acc-top1=65.191667 acc-top5=86.100000
Batch [0079]/[0097]: acc-top1=65.350000 acc-top5=85.887500
[Epoch 062] training: accuracy=82.072319	 loss=0.685012
[Epoch 062] speed: 125 samples/sec	time cost: 2112.954102
[Epoch 062] validation: acc-top1=64.850515 acc-top5=85.814433 loss=1.558826
Epoch[063] Batch [0019]/[1203]	Speed: 69.450976 samples/sec	 accuracy=82.675000	 loss=0.651763	 lr=0.001000
Epoch[063] Batch [0039]/[1203]	Speed: 104.401309 samples/sec	 accuracy=82.362500	 loss=0.669451	 lr=0.001000
Epoch[063] Batch [0059]/[1203]	Speed: 205.207248 samples/sec	 accuracy=82.650000	 loss=0.661006	 lr=0.001000
Epoch[063] Batch [0079]/[1203]	Speed: 97.959116 samples/sec	 accuracy=82.375000	 loss=0.668383	 lr=0.001000
Epoch[063] Batch [0099]/[1203]	Speed: 108.031267 samples/sec	 accuracy=82.330000	 loss=0.672156	 lr=0.001000
Epoch[063] Batch [0119]/[1203]	Speed: 175.311003 samples/sec	 accuracy=82.404167	 loss=0.671358	 lr=0.001000
Epoch[063] Batch [0139]/[1203]	Speed: 114.435762 samples/sec	 accuracy=82.332143	 loss=0.676760	 lr=0.001000
Epoch[063] Batch [0159]/[1203]	Speed: 172.635239 samples/sec	 accuracy=82.443750	 loss=0.673664	 lr=0.001000
Epoch[063] Batch [0179]/[1203]	Speed: 102.771188 samples/sec	 accuracy=82.469444	 loss=0.671937	 lr=0.001000
Epoch[063] Batch [0199]/[1203]	Speed: 112.575521 samples/sec	 accuracy=82.392500	 loss=0.673975	 lr=0.001000
Epoch[063] Batch [0219]/[1203]	Speed: 152.156853 samples/sec	 accuracy=82.386364	 loss=0.673759	 lr=0.001000
Epoch[063] Batch [0239]/[1203]	Speed: 115.757107 samples/sec	 accuracy=82.397917	 loss=0.673604	 lr=0.001000
Epoch[063] Batch [0259]/[1203]	Speed: 133.333346 samples/sec	 accuracy=82.336538	 loss=0.674328	 lr=0.001000
Epoch[063] Batch [0279]/[1203]	Speed: 124.582984 samples/sec	 accuracy=82.319643	 loss=0.675113	 lr=0.001000
Epoch[063] Batch [0299]/[1203]	Speed: 110.797539 samples/sec	 accuracy=82.265000	 loss=0.677347	 lr=0.001000
Epoch[063] Batch [0319]/[1203]	Speed: 168.564384 samples/sec	 accuracy=82.196875	 loss=0.679397	 lr=0.001000
Epoch[063] Batch [0339]/[1203]	Speed: 111.451746 samples/sec	 accuracy=82.202941	 loss=0.679223	 lr=0.001000
Epoch[063] Batch [0359]/[1203]	Speed: 116.451512 samples/sec	 accuracy=82.195833	 loss=0.679605	 lr=0.001000
Epoch[063] Batch [0379]/[1203]	Speed: 129.482741 samples/sec	 accuracy=82.201316	 loss=0.679066	 lr=0.001000
Epoch[063] Batch [0399]/[1203]	Speed: 124.071218 samples/sec	 accuracy=82.205000	 loss=0.678982	 lr=0.001000
Epoch[063] Batch [0419]/[1203]	Speed: 140.051614 samples/sec	 accuracy=82.182143	 loss=0.679597	 lr=0.001000
Epoch[063] Batch [0439]/[1203]	Speed: 109.226015 samples/sec	 accuracy=82.222727	 loss=0.678242	 lr=0.001000
Epoch[063] Batch [0459]/[1203]	Speed: 127.219134 samples/sec	 accuracy=82.195652	 loss=0.678515	 lr=0.001000
Epoch[063] Batch [0479]/[1203]	Speed: 141.241622 samples/sec	 accuracy=82.191667	 loss=0.679775	 lr=0.001000
Epoch[063] Batch [0499]/[1203]	Speed: 115.822774 samples/sec	 accuracy=82.219000	 loss=0.679185	 lr=0.001000
Epoch[063] Batch [0519]/[1203]	Speed: 150.461826 samples/sec	 accuracy=82.189423	 loss=0.681145	 lr=0.001000
Epoch[063] Batch [0539]/[1203]	Speed: 108.438940 samples/sec	 accuracy=82.215741	 loss=0.680164	 lr=0.001000
Epoch[063] Batch [0559]/[1203]	Speed: 125.694481 samples/sec	 accuracy=82.252679	 loss=0.679148	 lr=0.001000
Epoch[063] Batch [0579]/[1203]	Speed: 142.811951 samples/sec	 accuracy=82.256034	 loss=0.679872	 lr=0.001000
Epoch[063] Batch [0599]/[1203]	Speed: 106.510793 samples/sec	 accuracy=82.279167	 loss=0.678626	 lr=0.001000
Epoch[063] Batch [0619]/[1203]	Speed: 140.138694 samples/sec	 accuracy=82.266129	 loss=0.678641	 lr=0.001000
Epoch[063] Batch [0639]/[1203]	Speed: 118.759781 samples/sec	 accuracy=82.260937	 loss=0.678704	 lr=0.001000
Epoch[063] Batch [0659]/[1203]	Speed: 133.849190 samples/sec	 accuracy=82.271970	 loss=0.678474	 lr=0.001000
Epoch[063] Batch [0679]/[1203]	Speed: 141.404549 samples/sec	 accuracy=82.300000	 loss=0.678051	 lr=0.001000
Epoch[063] Batch [0699]/[1203]	Speed: 105.192201 samples/sec	 accuracy=82.292857	 loss=0.678320	 lr=0.001000
Epoch[063] Batch [0719]/[1203]	Speed: 131.701949 samples/sec	 accuracy=82.293750	 loss=0.678329	 lr=0.001000
Epoch[063] Batch [0739]/[1203]	Speed: 124.816047 samples/sec	 accuracy=82.293919	 loss=0.678079	 lr=0.001000
Epoch[063] Batch [0759]/[1203]	Speed: 135.745218 samples/sec	 accuracy=82.274342	 loss=0.678770	 lr=0.001000
Epoch[063] Batch [0779]/[1203]	Speed: 131.829236 samples/sec	 accuracy=82.289744	 loss=0.678156	 lr=0.001000
Epoch[063] Batch [0799]/[1203]	Speed: 108.057741 samples/sec	 accuracy=82.295625	 loss=0.677651	 lr=0.001000
Epoch[063] Batch [0819]/[1203]	Speed: 138.723700 samples/sec	 accuracy=82.284756	 loss=0.678174	 lr=0.001000
Epoch[063] Batch [0839]/[1203]	Speed: 125.908282 samples/sec	 accuracy=82.290476	 loss=0.677916	 lr=0.001000
Epoch[063] Batch [0859]/[1203]	Speed: 112.577888 samples/sec	 accuracy=82.287791	 loss=0.678310	 lr=0.001000
Epoch[063] Batch [0879]/[1203]	Speed: 142.918668 samples/sec	 accuracy=82.278409	 loss=0.678471	 lr=0.001000
Epoch[063] Batch [0899]/[1203]	Speed: 121.520955 samples/sec	 accuracy=82.286667	 loss=0.678808	 lr=0.001000
Epoch[063] Batch [0919]/[1203]	Speed: 138.678272 samples/sec	 accuracy=82.292935	 loss=0.678537	 lr=0.001000
Epoch[063] Batch [0939]/[1203]	Speed: 115.042947 samples/sec	 accuracy=82.289894	 loss=0.678464	 lr=0.001000
Epoch[063] Batch [0959]/[1203]	Speed: 121.233456 samples/sec	 accuracy=82.263021	 loss=0.679102	 lr=0.001000
Epoch[063] Batch [0979]/[1203]	Speed: 133.640896 samples/sec	 accuracy=82.266837	 loss=0.678685	 lr=0.001000
Epoch[063] Batch [0999]/[1203]	Speed: 123.612773 samples/sec	 accuracy=82.283000	 loss=0.678582	 lr=0.001000
Epoch[063] Batch [1019]/[1203]	Speed: 148.359004 samples/sec	 accuracy=82.285294	 loss=0.678386	 lr=0.001000
Epoch[063] Batch [1039]/[1203]	Speed: 116.642851 samples/sec	 accuracy=82.279808	 loss=0.678083	 lr=0.001000
Epoch[063] Batch [1059]/[1203]	Speed: 111.117388 samples/sec	 accuracy=82.261321	 loss=0.678664	 lr=0.001000
Epoch[063] Batch [1079]/[1203]	Speed: 153.279110 samples/sec	 accuracy=82.252315	 loss=0.678854	 lr=0.001000
Epoch[063] Batch [1099]/[1203]	Speed: 114.252836 samples/sec	 accuracy=82.248636	 loss=0.679148	 lr=0.001000
Epoch[063] Batch [1119]/[1203]	Speed: 122.013616 samples/sec	 accuracy=82.238839	 loss=0.679274	 lr=0.001000
Epoch[063] Batch [1139]/[1203]	Speed: 133.623627 samples/sec	 accuracy=82.234211	 loss=0.679520	 lr=0.001000
Epoch[063] Batch [1159]/[1203]	Speed: 114.066572 samples/sec	 accuracy=82.232759	 loss=0.679595	 lr=0.001000
Epoch[063] Batch [1179]/[1203]	Speed: 168.613321 samples/sec	 accuracy=82.224576	 loss=0.679787	 lr=0.001000
Epoch[063] Batch [1199]/[1203]	Speed: 197.586917 samples/sec	 accuracy=82.215000	 loss=0.679943	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.500000 acc-top5=85.150000
Batch [0039]/[0097]: acc-top1=63.250000 acc-top5=84.600000
Batch [0059]/[0097]: acc-top1=65.475000 acc-top5=85.966667
Batch [0079]/[0097]: acc-top1=65.706250 acc-top5=85.806250
[Epoch 063] training: accuracy=82.211139	 loss=0.680104
[Epoch 063] speed: 125 samples/sec	time cost: 2113.757918
[Epoch 063] validation: acc-top1=65.067010 acc-top5=85.778351 loss=1.539339
Epoch[064] Batch [0019]/[1203]	Speed: 68.226805 samples/sec	 accuracy=83.025000	 loss=0.635603	 lr=0.001000
Epoch[064] Batch [0039]/[1203]	Speed: 103.689767 samples/sec	 accuracy=82.487500	 loss=0.662072	 lr=0.001000
Epoch[064] Batch [0059]/[1203]	Speed: 176.743942 samples/sec	 accuracy=82.091667	 loss=0.671950	 lr=0.001000
Epoch[064] Batch [0079]/[1203]	Speed: 107.507510 samples/sec	 accuracy=82.237500	 loss=0.670395	 lr=0.001000
Epoch[064] Batch [0099]/[1203]	Speed: 94.732257 samples/sec	 accuracy=82.310000	 loss=0.667322	 lr=0.001000
Epoch[064] Batch [0119]/[1203]	Speed: 251.994209 samples/sec	 accuracy=82.354167	 loss=0.664435	 lr=0.001000
Epoch[064] Batch [0139]/[1203]	Speed: 97.428673 samples/sec	 accuracy=82.303571	 loss=0.668970	 lr=0.001000
Epoch[064] Batch [0159]/[1203]	Speed: 242.506221 samples/sec	 accuracy=82.243750	 loss=0.670598	 lr=0.001000
Epoch[064] Batch [0179]/[1203]	Speed: 97.809902 samples/sec	 accuracy=82.263889	 loss=0.669594	 lr=0.001000
Epoch[064] Batch [0199]/[1203]	Speed: 100.578624 samples/sec	 accuracy=82.322500	 loss=0.667973	 lr=0.001000
Epoch[064] Batch [0219]/[1203]	Speed: 196.701156 samples/sec	 accuracy=82.356818	 loss=0.666196	 lr=0.001000
Epoch[064] Batch [0239]/[1203]	Speed: 100.348161 samples/sec	 accuracy=82.343750	 loss=0.667464	 lr=0.001000
Epoch[064] Batch [0259]/[1203]	Speed: 139.870377 samples/sec	 accuracy=82.353846	 loss=0.667398	 lr=0.001000
Epoch[064] Batch [0279]/[1203]	Speed: 132.421801 samples/sec	 accuracy=82.373214	 loss=0.667094	 lr=0.001000
Epoch[064] Batch [0299]/[1203]	Speed: 109.068151 samples/sec	 accuracy=82.413333	 loss=0.666745	 lr=0.001000
Epoch[064] Batch [0319]/[1203]	Speed: 173.990800 samples/sec	 accuracy=82.439063	 loss=0.666631	 lr=0.001000
Epoch[064] Batch [0339]/[1203]	Speed: 109.206938 samples/sec	 accuracy=82.395588	 loss=0.667296	 lr=0.001000
Epoch[064] Batch [0359]/[1203]	Speed: 132.170972 samples/sec	 accuracy=82.376389	 loss=0.668840	 lr=0.001000
Epoch[064] Batch [0379]/[1203]	Speed: 126.933694 samples/sec	 accuracy=82.415789	 loss=0.666703	 lr=0.001000
Epoch[064] Batch [0399]/[1203]	Speed: 106.518993 samples/sec	 accuracy=82.396250	 loss=0.666896	 lr=0.001000
Epoch[064] Batch [0419]/[1203]	Speed: 152.968488 samples/sec	 accuracy=82.358333	 loss=0.668353	 lr=0.001000
Epoch[064] Batch [0439]/[1203]	Speed: 112.042664 samples/sec	 accuracy=82.319318	 loss=0.670859	 lr=0.001000
Epoch[064] Batch [0459]/[1203]	Speed: 137.931012 samples/sec	 accuracy=82.342391	 loss=0.671002	 lr=0.001000
Epoch[064] Batch [0479]/[1203]	Speed: 128.229410 samples/sec	 accuracy=82.298958	 loss=0.672055	 lr=0.001000
Epoch[064] Batch [0499]/[1203]	Speed: 110.236348 samples/sec	 accuracy=82.291000	 loss=0.671769	 lr=0.001000
Epoch[064] Batch [0519]/[1203]	Speed: 146.803226 samples/sec	 accuracy=82.283654	 loss=0.672879	 lr=0.001000
Epoch[064] Batch [0539]/[1203]	Speed: 117.443291 samples/sec	 accuracy=82.291667	 loss=0.672530	 lr=0.001000
Epoch[064] Batch [0559]/[1203]	Speed: 114.212294 samples/sec	 accuracy=82.280357	 loss=0.673114	 lr=0.001000
Epoch[064] Batch [0579]/[1203]	Speed: 150.509395 samples/sec	 accuracy=82.225862	 loss=0.674752	 lr=0.001000
Epoch[064] Batch [0599]/[1203]	Speed: 129.962037 samples/sec	 accuracy=82.233333	 loss=0.674759	 lr=0.001000
Epoch[064] Batch [0619]/[1203]	Speed: 141.381679 samples/sec	 accuracy=82.250000	 loss=0.674255	 lr=0.001000
Epoch[064] Batch [0639]/[1203]	Speed: 108.361185 samples/sec	 accuracy=82.254688	 loss=0.674481	 lr=0.001000
Epoch[064] Batch [0659]/[1203]	Speed: 120.110456 samples/sec	 accuracy=82.237879	 loss=0.674632	 lr=0.001000
Epoch[064] Batch [0679]/[1203]	Speed: 130.267768 samples/sec	 accuracy=82.211765	 loss=0.675541	 lr=0.001000
Epoch[064] Batch [0699]/[1203]	Speed: 124.033539 samples/sec	 accuracy=82.191429	 loss=0.675849	 lr=0.001000
Epoch[064] Batch [0719]/[1203]	Speed: 152.974187 samples/sec	 accuracy=82.200000	 loss=0.675741	 lr=0.001000
Epoch[064] Batch [0739]/[1203]	Speed: 116.368486 samples/sec	 accuracy=82.163514	 loss=0.676231	 lr=0.001000
Epoch[064] Batch [0759]/[1203]	Speed: 116.803323 samples/sec	 accuracy=82.153947	 loss=0.676728	 lr=0.001000
Epoch[064] Batch [0779]/[1203]	Speed: 124.632411 samples/sec	 accuracy=82.139103	 loss=0.677514	 lr=0.001000
Epoch[064] Batch [0799]/[1203]	Speed: 126.827639 samples/sec	 accuracy=82.153750	 loss=0.676962	 lr=0.001000
Epoch[064] Batch [0819]/[1203]	Speed: 148.409025 samples/sec	 accuracy=82.153049	 loss=0.677578	 lr=0.001000
Epoch[064] Batch [0839]/[1203]	Speed: 109.057104 samples/sec	 accuracy=82.148214	 loss=0.677741	 lr=0.001000
Epoch[064] Batch [0859]/[1203]	Speed: 131.544720 samples/sec	 accuracy=82.147674	 loss=0.678008	 lr=0.001000
Epoch[064] Batch [0879]/[1203]	Speed: 121.549302 samples/sec	 accuracy=82.159091	 loss=0.677847	 lr=0.001000
Epoch[064] Batch [0899]/[1203]	Speed: 125.566915 samples/sec	 accuracy=82.150556	 loss=0.678202	 lr=0.001000
Epoch[064] Batch [0919]/[1203]	Speed: 125.396445 samples/sec	 accuracy=82.166304	 loss=0.677901	 lr=0.001000
Epoch[064] Batch [0939]/[1203]	Speed: 124.529394 samples/sec	 accuracy=82.146809	 loss=0.678151	 lr=0.001000
Epoch[064] Batch [0959]/[1203]	Speed: 130.223835 samples/sec	 accuracy=82.155729	 loss=0.678127	 lr=0.001000
Epoch[064] Batch [0979]/[1203]	Speed: 124.481232 samples/sec	 accuracy=82.147959	 loss=0.678461	 lr=0.001000
Epoch[064] Batch [0999]/[1203]	Speed: 126.405290 samples/sec	 accuracy=82.145500	 loss=0.678356	 lr=0.001000
Epoch[064] Batch [1019]/[1203]	Speed: 127.124239 samples/sec	 accuracy=82.148529	 loss=0.678405	 lr=0.001000
Epoch[064] Batch [1039]/[1203]	Speed: 126.634165 samples/sec	 accuracy=82.137981	 loss=0.678705	 lr=0.001000
Epoch[064] Batch [1059]/[1203]	Speed: 126.840779 samples/sec	 accuracy=82.146226	 loss=0.678563	 lr=0.001000
Epoch[064] Batch [1079]/[1203]	Speed: 122.362658 samples/sec	 accuracy=82.169907	 loss=0.677876	 lr=0.001000
Epoch[064] Batch [1099]/[1203]	Speed: 128.601969 samples/sec	 accuracy=82.155909	 loss=0.678195	 lr=0.001000
Epoch[064] Batch [1119]/[1203]	Speed: 122.391992 samples/sec	 accuracy=82.168750	 loss=0.677923	 lr=0.001000
Epoch[064] Batch [1139]/[1203]	Speed: 127.851166 samples/sec	 accuracy=82.193860	 loss=0.677218	 lr=0.001000
Epoch[064] Batch [1159]/[1203]	Speed: 118.898654 samples/sec	 accuracy=82.196552	 loss=0.677219	 lr=0.001000
Epoch[064] Batch [1179]/[1203]	Speed: 143.450707 samples/sec	 accuracy=82.182203	 loss=0.677569	 lr=0.001000
Epoch[064] Batch [1199]/[1203]	Speed: 235.360702 samples/sec	 accuracy=82.183750	 loss=0.677645	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.850000 acc-top5=85.100000
Batch [0039]/[0097]: acc-top1=63.275000 acc-top5=84.812500
Batch [0059]/[0097]: acc-top1=65.350000 acc-top5=86.058333
Batch [0079]/[0097]: acc-top1=65.550000 acc-top5=85.887500
[Epoch 064] training: accuracy=82.178304	 loss=0.677876
[Epoch 064] speed: 125 samples/sec	time cost: 2114.292632
[Epoch 064] validation: acc-top1=64.953608 acc-top5=85.773196 loss=1.558499
Epoch[065] Batch [0019]/[1203]	Speed: 70.149695 samples/sec	 accuracy=82.225000	 loss=0.697240	 lr=0.001000
Epoch[065] Batch [0039]/[1203]	Speed: 98.865888 samples/sec	 accuracy=82.337500	 loss=0.680519	 lr=0.001000
Epoch[065] Batch [0059]/[1203]	Speed: 242.455328 samples/sec	 accuracy=82.391667	 loss=0.674416	 lr=0.001000
Epoch[065] Batch [0079]/[1203]	Speed: 94.175560 samples/sec	 accuracy=82.531250	 loss=0.670674	 lr=0.001000
Epoch[065] Batch [0099]/[1203]	Speed: 110.256109 samples/sec	 accuracy=82.575000	 loss=0.667085	 lr=0.001000
Epoch[065] Batch [0119]/[1203]	Speed: 204.256335 samples/sec	 accuracy=82.616667	 loss=0.666771	 lr=0.001000
Epoch[065] Batch [0139]/[1203]	Speed: 96.532454 samples/sec	 accuracy=82.646429	 loss=0.666318	 lr=0.001000
Epoch[065] Batch [0159]/[1203]	Speed: 231.536780 samples/sec	 accuracy=82.787500	 loss=0.661163	 lr=0.001000
Epoch[065] Batch [0179]/[1203]	Speed: 95.432639 samples/sec	 accuracy=82.713889	 loss=0.663103	 lr=0.001000
Epoch[065] Batch [0199]/[1203]	Speed: 100.917192 samples/sec	 accuracy=82.635000	 loss=0.666937	 lr=0.001000
Epoch[065] Batch [0219]/[1203]	Speed: 217.232630 samples/sec	 accuracy=82.697727	 loss=0.663855	 lr=0.001000
Epoch[065] Batch [0239]/[1203]	Speed: 100.690622 samples/sec	 accuracy=82.714583	 loss=0.664111	 lr=0.001000
Epoch[065] Batch [0259]/[1203]	Speed: 112.511538 samples/sec	 accuracy=82.680769	 loss=0.666440	 lr=0.001000
Epoch[065] Batch [0279]/[1203]	Speed: 167.888223 samples/sec	 accuracy=82.664286	 loss=0.666185	 lr=0.001000
Epoch[065] Batch [0299]/[1203]	Speed: 103.498350 samples/sec	 accuracy=82.625000	 loss=0.667013	 lr=0.001000
Epoch[065] Batch [0319]/[1203]	Speed: 188.896573 samples/sec	 accuracy=82.632812	 loss=0.665896	 lr=0.001000
Epoch[065] Batch [0339]/[1203]	Speed: 102.306421 samples/sec	 accuracy=82.591176	 loss=0.667828	 lr=0.001000
Epoch[065] Batch [0359]/[1203]	Speed: 123.263764 samples/sec	 accuracy=82.601389	 loss=0.667097	 lr=0.001000
Epoch[065] Batch [0379]/[1203]	Speed: 148.477689 samples/sec	 accuracy=82.572368	 loss=0.666697	 lr=0.001000
Epoch[065] Batch [0399]/[1203]	Speed: 114.870201 samples/sec	 accuracy=82.553750	 loss=0.667089	 lr=0.001000
Epoch[065] Batch [0419]/[1203]	Speed: 136.268695 samples/sec	 accuracy=82.520238	 loss=0.667425	 lr=0.001000
Epoch[065] Batch [0439]/[1203]	Speed: 109.958903 samples/sec	 accuracy=82.511364	 loss=0.666999	 lr=0.001000
Epoch[065] Batch [0459]/[1203]	Speed: 132.376481 samples/sec	 accuracy=82.513043	 loss=0.666919	 lr=0.001000
Epoch[065] Batch [0479]/[1203]	Speed: 142.002205 samples/sec	 accuracy=82.535417	 loss=0.666678	 lr=0.001000
Epoch[065] Batch [0499]/[1203]	Speed: 110.421959 samples/sec	 accuracy=82.502000	 loss=0.667554	 lr=0.001000
Epoch[065] Batch [0519]/[1203]	Speed: 133.836630 samples/sec	 accuracy=82.525962	 loss=0.666633	 lr=0.001000
Epoch[065] Batch [0539]/[1203]	Speed: 128.748965 samples/sec	 accuracy=82.526852	 loss=0.666537	 lr=0.001000
Epoch[065] Batch [0559]/[1203]	Speed: 124.472743 samples/sec	 accuracy=82.541071	 loss=0.665793	 lr=0.001000
Epoch[065] Batch [0579]/[1203]	Speed: 153.309944 samples/sec	 accuracy=82.559483	 loss=0.665587	 lr=0.001000
Epoch[065] Batch [0599]/[1203]	Speed: 102.183433 samples/sec	 accuracy=82.533333	 loss=0.666231	 lr=0.001000
Epoch[065] Batch [0619]/[1203]	Speed: 145.689652 samples/sec	 accuracy=82.555645	 loss=0.666407	 lr=0.001000
Epoch[065] Batch [0639]/[1203]	Speed: 116.641609 samples/sec	 accuracy=82.551563	 loss=0.666790	 lr=0.001000
Epoch[065] Batch [0659]/[1203]	Speed: 128.022803 samples/sec	 accuracy=82.537879	 loss=0.667195	 lr=0.001000
Epoch[065] Batch [0679]/[1203]	Speed: 118.189277 samples/sec	 accuracy=82.536029	 loss=0.667280	 lr=0.001000
Epoch[065] Batch [0699]/[1203]	Speed: 119.960740 samples/sec	 accuracy=82.525714	 loss=0.667920	 lr=0.001000
Epoch[065] Batch [0719]/[1203]	Speed: 176.803320 samples/sec	 accuracy=82.552083	 loss=0.666868	 lr=0.001000
Epoch[065] Batch [0739]/[1203]	Speed: 109.408081 samples/sec	 accuracy=82.534459	 loss=0.667363	 lr=0.001000
Epoch[065] Batch [0759]/[1203]	Speed: 121.546462 samples/sec	 accuracy=82.513158	 loss=0.667852	 lr=0.001000
Epoch[065] Batch [0779]/[1203]	Speed: 141.478191 samples/sec	 accuracy=82.501923	 loss=0.669145	 lr=0.001000
Epoch[065] Batch [0799]/[1203]	Speed: 110.620242 samples/sec	 accuracy=82.478125	 loss=0.670035	 lr=0.001000
Epoch[065] Batch [0819]/[1203]	Speed: 155.277827 samples/sec	 accuracy=82.478049	 loss=0.670055	 lr=0.001000
Epoch[065] Batch [0839]/[1203]	Speed: 114.235607 samples/sec	 accuracy=82.467857	 loss=0.670158	 lr=0.001000
Epoch[065] Batch [0859]/[1203]	Speed: 121.617549 samples/sec	 accuracy=82.456395	 loss=0.670392	 lr=0.001000
Epoch[065] Batch [0879]/[1203]	Speed: 159.049793 samples/sec	 accuracy=82.461364	 loss=0.670562	 lr=0.001000
Epoch[065] Batch [0899]/[1203]	Speed: 106.249638 samples/sec	 accuracy=82.450000	 loss=0.670558	 lr=0.001000
Epoch[065] Batch [0919]/[1203]	Speed: 129.449554 samples/sec	 accuracy=82.454891	 loss=0.670353	 lr=0.001000
Epoch[065] Batch [0939]/[1203]	Speed: 136.912888 samples/sec	 accuracy=82.457979	 loss=0.670302	 lr=0.001000
Epoch[065] Batch [0959]/[1203]	Speed: 116.659101 samples/sec	 accuracy=82.453646	 loss=0.670276	 lr=0.001000
Epoch[065] Batch [0979]/[1203]	Speed: 159.732516 samples/sec	 accuracy=82.452041	 loss=0.670583	 lr=0.001000
Epoch[065] Batch [0999]/[1203]	Speed: 100.640262 samples/sec	 accuracy=82.438500	 loss=0.670856	 lr=0.001000
Epoch[065] Batch [1019]/[1203]	Speed: 129.190407 samples/sec	 accuracy=82.423039	 loss=0.671248	 lr=0.001000
Epoch[065] Batch [1039]/[1203]	Speed: 153.789483 samples/sec	 accuracy=82.413462	 loss=0.671360	 lr=0.001000
Epoch[065] Batch [1059]/[1203]	Speed: 113.458984 samples/sec	 accuracy=82.392453	 loss=0.671728	 lr=0.001000
Epoch[065] Batch [1079]/[1203]	Speed: 131.662100 samples/sec	 accuracy=82.408796	 loss=0.671314	 lr=0.001000
Epoch[065] Batch [1099]/[1203]	Speed: 128.317431 samples/sec	 accuracy=82.410909	 loss=0.671079	 lr=0.001000
Epoch[065] Batch [1119]/[1203]	Speed: 111.750774 samples/sec	 accuracy=82.399107	 loss=0.671158	 lr=0.001000
Epoch[065] Batch [1139]/[1203]	Speed: 162.557578 samples/sec	 accuracy=82.407895	 loss=0.671182	 lr=0.001000
Epoch[065] Batch [1159]/[1203]	Speed: 114.856566 samples/sec	 accuracy=82.409052	 loss=0.670964	 lr=0.001000
Epoch[065] Batch [1179]/[1203]	Speed: 125.522474 samples/sec	 accuracy=82.410593	 loss=0.671016	 lr=0.001000
Epoch[065] Batch [1199]/[1203]	Speed: 264.222506 samples/sec	 accuracy=82.396250	 loss=0.671177	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.500000 acc-top5=85.025000
Batch [0039]/[0097]: acc-top1=63.450000 acc-top5=84.812500
Batch [0059]/[0097]: acc-top1=65.491667 acc-top5=86.083333
Batch [0079]/[0097]: acc-top1=65.643750 acc-top5=85.900000
[Epoch 065] training: accuracy=82.398587	 loss=0.671146
[Epoch 065] speed: 125 samples/sec	time cost: 2108.786652
[Epoch 065] validation: acc-top1=65.113402 acc-top5=85.829897 loss=1.553272
Epoch[066] Batch [0019]/[1204]	Speed: 68.536719 samples/sec	 accuracy=82.425000	 loss=0.679582	 lr=0.001000
Epoch[066] Batch [0039]/[1204]	Speed: 97.246358 samples/sec	 accuracy=82.787500	 loss=0.671146	 lr=0.001000
Epoch[066] Batch [0059]/[1204]	Speed: 255.978002 samples/sec	 accuracy=82.916667	 loss=0.663268	 lr=0.001000
Epoch[066] Batch [0079]/[1204]	Speed: 98.930330 samples/sec	 accuracy=82.718750	 loss=0.665473	 lr=0.001000
Epoch[066] Batch [0099]/[1204]	Speed: 104.806790 samples/sec	 accuracy=82.815000	 loss=0.659770	 lr=0.001000
Epoch[066] Batch [0119]/[1204]	Speed: 208.248581 samples/sec	 accuracy=82.762500	 loss=0.660699	 lr=0.001000
Epoch[066] Batch [0139]/[1204]	Speed: 96.868224 samples/sec	 accuracy=82.728571	 loss=0.660114	 lr=0.001000
Epoch[066] Batch [0159]/[1204]	Speed: 248.068896 samples/sec	 accuracy=82.693750	 loss=0.659802	 lr=0.001000
Epoch[066] Batch [0179]/[1204]	Speed: 95.282820 samples/sec	 accuracy=82.686111	 loss=0.659879	 lr=0.001000
Epoch[066] Batch [0199]/[1204]	Speed: 103.518049 samples/sec	 accuracy=82.757500	 loss=0.656951	 lr=0.001000
Epoch[066] Batch [0219]/[1204]	Speed: 193.077902 samples/sec	 accuracy=82.734091	 loss=0.657631	 lr=0.001000
Epoch[066] Batch [0239]/[1204]	Speed: 101.736545 samples/sec	 accuracy=82.672917	 loss=0.657948	 lr=0.001000
Epoch[066] Batch [0259]/[1204]	Speed: 122.401785 samples/sec	 accuracy=82.659615	 loss=0.659884	 lr=0.001000
Epoch[066] Batch [0279]/[1204]	Speed: 161.956962 samples/sec	 accuracy=82.687500	 loss=0.658510	 lr=0.001000
Epoch[066] Batch [0299]/[1204]	Speed: 104.744097 samples/sec	 accuracy=82.681667	 loss=0.659635	 lr=0.001000
Epoch[066] Batch [0319]/[1204]	Speed: 162.141144 samples/sec	 accuracy=82.664062	 loss=0.661104	 lr=0.001000
Epoch[066] Batch [0339]/[1204]	Speed: 102.447531 samples/sec	 accuracy=82.650000	 loss=0.661579	 lr=0.001000
Epoch[066] Batch [0359]/[1204]	Speed: 138.409408 samples/sec	 accuracy=82.648611	 loss=0.661864	 lr=0.001000
Epoch[066] Batch [0379]/[1204]	Speed: 139.557519 samples/sec	 accuracy=82.550000	 loss=0.665389	 lr=0.001000
Epoch[066] Batch [0399]/[1204]	Speed: 112.282672 samples/sec	 accuracy=82.533750	 loss=0.665352	 lr=0.001000
Epoch[066] Batch [0419]/[1204]	Speed: 149.415269 samples/sec	 accuracy=82.525000	 loss=0.664944	 lr=0.001000
Epoch[066] Batch [0439]/[1204]	Speed: 106.433640 samples/sec	 accuracy=82.546591	 loss=0.664861	 lr=0.001000
Epoch[066] Batch [0459]/[1204]	Speed: 125.974020 samples/sec	 accuracy=82.553261	 loss=0.664953	 lr=0.001000
Epoch[066] Batch [0479]/[1204]	Speed: 153.452325 samples/sec	 accuracy=82.547917	 loss=0.666060	 lr=0.001000
Epoch[066] Batch [0499]/[1204]	Speed: 104.115149 samples/sec	 accuracy=82.575000	 loss=0.664962	 lr=0.001000
Epoch[066] Batch [0519]/[1204]	Speed: 144.631212 samples/sec	 accuracy=82.589423	 loss=0.664556	 lr=0.001000
Epoch[066] Batch [0539]/[1204]	Speed: 123.653135 samples/sec	 accuracy=82.557407	 loss=0.664813	 lr=0.001000
Epoch[066] Batch [0559]/[1204]	Speed: 111.386440 samples/sec	 accuracy=82.569643	 loss=0.664991	 lr=0.001000
Epoch[066] Batch [0579]/[1204]	Speed: 152.475990 samples/sec	 accuracy=82.569828	 loss=0.664952	 lr=0.001000
Epoch[066] Batch [0599]/[1204]	Speed: 114.345283 samples/sec	 accuracy=82.541667	 loss=0.665903	 lr=0.001000
Epoch[066] Batch [0619]/[1204]	Speed: 153.358865 samples/sec	 accuracy=82.537903	 loss=0.666572	 lr=0.001000
Epoch[066] Batch [0639]/[1204]	Speed: 112.515212 samples/sec	 accuracy=82.570312	 loss=0.665836	 lr=0.001000
Epoch[066] Batch [0659]/[1204]	Speed: 116.617658 samples/sec	 accuracy=82.562879	 loss=0.667053	 lr=0.001000
Epoch[066] Batch [0679]/[1204]	Speed: 156.205019 samples/sec	 accuracy=82.538971	 loss=0.668163	 lr=0.001000
Epoch[066] Batch [0699]/[1204]	Speed: 116.643085 samples/sec	 accuracy=82.516429	 loss=0.668738	 lr=0.001000
Epoch[066] Batch [0719]/[1204]	Speed: 126.207725 samples/sec	 accuracy=82.520833	 loss=0.668924	 lr=0.001000
Epoch[066] Batch [0739]/[1204]	Speed: 126.448938 samples/sec	 accuracy=82.521622	 loss=0.668967	 lr=0.001000
Epoch[066] Batch [0759]/[1204]	Speed: 114.330302 samples/sec	 accuracy=82.528947	 loss=0.668640	 lr=0.001000
Epoch[066] Batch [0779]/[1204]	Speed: 153.568974 samples/sec	 accuracy=82.512179	 loss=0.669148	 lr=0.001000
Epoch[066] Batch [0799]/[1204]	Speed: 115.421383 samples/sec	 accuracy=82.501250	 loss=0.668956	 lr=0.001000
Epoch[066] Batch [0819]/[1204]	Speed: 124.695934 samples/sec	 accuracy=82.538415	 loss=0.668058	 lr=0.001000
Epoch[066] Batch [0839]/[1204]	Speed: 140.304567 samples/sec	 accuracy=82.516071	 loss=0.668588	 lr=0.001000
Epoch[066] Batch [0859]/[1204]	Speed: 113.119028 samples/sec	 accuracy=82.523256	 loss=0.668001	 lr=0.001000
Epoch[066] Batch [0879]/[1204]	Speed: 131.066212 samples/sec	 accuracy=82.523295	 loss=0.668237	 lr=0.001000
Epoch[066] Batch [0899]/[1204]	Speed: 129.820832 samples/sec	 accuracy=82.518333	 loss=0.668590	 lr=0.001000
Epoch[066] Batch [0919]/[1204]	Speed: 115.088662 samples/sec	 accuracy=82.502717	 loss=0.668866	 lr=0.001000
Epoch[066] Batch [0939]/[1204]	Speed: 145.294872 samples/sec	 accuracy=82.494681	 loss=0.668933	 lr=0.001000
Epoch[066] Batch [0959]/[1204]	Speed: 120.737628 samples/sec	 accuracy=82.490104	 loss=0.668894	 lr=0.001000
Epoch[066] Batch [0979]/[1204]	Speed: 127.224070 samples/sec	 accuracy=82.478571	 loss=0.669191	 lr=0.001000
Epoch[066] Batch [0999]/[1204]	Speed: 127.525131 samples/sec	 accuracy=82.495000	 loss=0.668741	 lr=0.001000
Epoch[066] Batch [1019]/[1204]	Speed: 112.508996 samples/sec	 accuracy=82.488235	 loss=0.668456	 lr=0.001000
Epoch[066] Batch [1039]/[1204]	Speed: 171.616161 samples/sec	 accuracy=82.480288	 loss=0.668465	 lr=0.001000
Epoch[066] Batch [1059]/[1204]	Speed: 105.283099 samples/sec	 accuracy=82.479245	 loss=0.668281	 lr=0.001000
Epoch[066] Batch [1079]/[1204]	Speed: 124.726944 samples/sec	 accuracy=82.464815	 loss=0.668448	 lr=0.001000
Epoch[066] Batch [1099]/[1204]	Speed: 123.281426 samples/sec	 accuracy=82.477273	 loss=0.667919	 lr=0.001000
Epoch[066] Batch [1119]/[1204]	Speed: 126.534502 samples/sec	 accuracy=82.481250	 loss=0.667494	 lr=0.001000
Epoch[066] Batch [1139]/[1204]	Speed: 139.359323 samples/sec	 accuracy=82.468421	 loss=0.667880	 lr=0.001000
Epoch[066] Batch [1159]/[1204]	Speed: 120.811782 samples/sec	 accuracy=82.455172	 loss=0.668468	 lr=0.001000
Epoch[066] Batch [1179]/[1204]	Speed: 124.011821 samples/sec	 accuracy=82.447458	 loss=0.669075	 lr=0.001000
Epoch[066] Batch [1199]/[1204]	Speed: 236.727150 samples/sec	 accuracy=82.432083	 loss=0.669293	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.525000 acc-top5=85.000000
Batch [0039]/[0097]: acc-top1=62.987500 acc-top5=84.775000
Batch [0059]/[0097]: acc-top1=65.191667 acc-top5=85.966667
Batch [0079]/[0097]: acc-top1=65.450000 acc-top5=85.887500
[Epoch 066] training: accuracy=82.430233	 loss=0.669346
[Epoch 066] speed: 125 samples/sec	time cost: 2110.587331
[Epoch 066] validation: acc-top1=64.953608 acc-top5=85.835052 loss=1.551376
Epoch[067] Batch [0019]/[1203]	Speed: 69.554733 samples/sec	 accuracy=82.725000	 loss=0.650933	 lr=0.001000
Epoch[067] Batch [0039]/[1203]	Speed: 95.108423 samples/sec	 accuracy=83.425000	 loss=0.636368	 lr=0.001000
Epoch[067] Batch [0059]/[1203]	Speed: 260.025773 samples/sec	 accuracy=83.066667	 loss=0.645857	 lr=0.001000
Epoch[067] Batch [0079]/[1203]	Speed: 93.958925 samples/sec	 accuracy=82.768750	 loss=0.653718	 lr=0.001000
Epoch[067] Batch [0099]/[1203]	Speed: 114.960905 samples/sec	 accuracy=82.820000	 loss=0.653750	 lr=0.001000
Epoch[067] Batch [0119]/[1203]	Speed: 184.447729 samples/sec	 accuracy=82.745833	 loss=0.654763	 lr=0.001000
Epoch[067] Batch [0139]/[1203]	Speed: 94.236439 samples/sec	 accuracy=82.778571	 loss=0.653964	 lr=0.001000
Epoch[067] Batch [0159]/[1203]	Speed: 214.552882 samples/sec	 accuracy=82.693750	 loss=0.655164	 lr=0.001000
Epoch[067] Batch [0179]/[1203]	Speed: 102.774049 samples/sec	 accuracy=82.777778	 loss=0.654842	 lr=0.001000
Epoch[067] Batch [0199]/[1203]	Speed: 116.309918 samples/sec	 accuracy=82.817500	 loss=0.653597	 lr=0.001000
Epoch[067] Batch [0219]/[1203]	Speed: 169.389741 samples/sec	 accuracy=82.788636	 loss=0.653940	 lr=0.001000
Epoch[067] Batch [0239]/[1203]	Speed: 108.153269 samples/sec	 accuracy=82.741667	 loss=0.655606	 lr=0.001000
Epoch[067] Batch [0259]/[1203]	Speed: 146.091611 samples/sec	 accuracy=82.734615	 loss=0.655046	 lr=0.001000
Epoch[067] Batch [0279]/[1203]	Speed: 116.642853 samples/sec	 accuracy=82.694643	 loss=0.657430	 lr=0.001000
Epoch[067] Batch [0299]/[1203]	Speed: 119.200850 samples/sec	 accuracy=82.678333	 loss=0.658283	 lr=0.001000
Epoch[067] Batch [0319]/[1203]	Speed: 155.494120 samples/sec	 accuracy=82.642187	 loss=0.660019	 lr=0.001000
Epoch[067] Batch [0339]/[1203]	Speed: 112.704268 samples/sec	 accuracy=82.647059	 loss=0.659978	 lr=0.001000
Epoch[067] Batch [0359]/[1203]	Speed: 130.587155 samples/sec	 accuracy=82.648611	 loss=0.660931	 lr=0.001000
Epoch[067] Batch [0379]/[1203]	Speed: 122.445254 samples/sec	 accuracy=82.646053	 loss=0.661400	 lr=0.001000
Epoch[067] Batch [0399]/[1203]	Speed: 109.507595 samples/sec	 accuracy=82.625000	 loss=0.661229	 lr=0.001000
Epoch[067] Batch [0419]/[1203]	Speed: 158.911778 samples/sec	 accuracy=82.636905	 loss=0.659808	 lr=0.001000
Epoch[067] Batch [0439]/[1203]	Speed: 113.076282 samples/sec	 accuracy=82.630682	 loss=0.660642	 lr=0.001000
Epoch[067] Batch [0459]/[1203]	Speed: 116.777397 samples/sec	 accuracy=82.644565	 loss=0.659931	 lr=0.001000
Epoch[067] Batch [0479]/[1203]	Speed: 151.507486 samples/sec	 accuracy=82.636458	 loss=0.660207	 lr=0.001000
Epoch[067] Batch [0499]/[1203]	Speed: 106.234489 samples/sec	 accuracy=82.580000	 loss=0.662008	 lr=0.001000
Epoch[067] Batch [0519]/[1203]	Speed: 171.073026 samples/sec	 accuracy=82.629808	 loss=0.660414	 lr=0.001000
Epoch[067] Batch [0539]/[1203]	Speed: 117.197719 samples/sec	 accuracy=82.590741	 loss=0.661789	 lr=0.001000
Epoch[067] Batch [0559]/[1203]	Speed: 106.397804 samples/sec	 accuracy=82.561607	 loss=0.662214	 lr=0.001000
Epoch[067] Batch [0579]/[1203]	Speed: 154.426264 samples/sec	 accuracy=82.526724	 loss=0.662955	 lr=0.001000
Epoch[067] Batch [0599]/[1203]	Speed: 108.929435 samples/sec	 accuracy=82.489167	 loss=0.663862	 lr=0.001000
Epoch[067] Batch [0619]/[1203]	Speed: 122.489616 samples/sec	 accuracy=82.473387	 loss=0.663873	 lr=0.001000
Epoch[067] Batch [0639]/[1203]	Speed: 155.718632 samples/sec	 accuracy=82.496094	 loss=0.663750	 lr=0.001000
Epoch[067] Batch [0659]/[1203]	Speed: 112.016052 samples/sec	 accuracy=82.481818	 loss=0.663660	 lr=0.001000
Epoch[067] Batch [0679]/[1203]	Speed: 157.263745 samples/sec	 accuracy=82.494853	 loss=0.663263	 lr=0.001000
Epoch[067] Batch [0699]/[1203]	Speed: 111.932295 samples/sec	 accuracy=82.507857	 loss=0.662640	 lr=0.001000
Epoch[067] Batch [0719]/[1203]	Speed: 113.654344 samples/sec	 accuracy=82.493750	 loss=0.663652	 lr=0.001000
Epoch[067] Batch [0739]/[1203]	Speed: 138.621543 samples/sec	 accuracy=82.472297	 loss=0.664211	 lr=0.001000
Epoch[067] Batch [0759]/[1203]	Speed: 120.860642 samples/sec	 accuracy=82.453289	 loss=0.664558	 lr=0.001000
Epoch[067] Batch [0779]/[1203]	Speed: 149.458645 samples/sec	 accuracy=82.418590	 loss=0.665335	 lr=0.001000
Epoch[067] Batch [0799]/[1203]	Speed: 114.971345 samples/sec	 accuracy=82.431875	 loss=0.665215	 lr=0.001000
Epoch[067] Batch [0819]/[1203]	Speed: 108.556692 samples/sec	 accuracy=82.435366	 loss=0.664858	 lr=0.001000
Epoch[067] Batch [0839]/[1203]	Speed: 141.633576 samples/sec	 accuracy=82.435714	 loss=0.664800	 lr=0.001000
Epoch[067] Batch [0859]/[1203]	Speed: 119.867687 samples/sec	 accuracy=82.406977	 loss=0.665697	 lr=0.001000
Epoch[067] Batch [0879]/[1203]	Speed: 128.079102 samples/sec	 accuracy=82.406250	 loss=0.665338	 lr=0.001000
Epoch[067] Batch [0899]/[1203]	Speed: 132.393173 samples/sec	 accuracy=82.396111	 loss=0.665973	 lr=0.001000
Epoch[067] Batch [0919]/[1203]	Speed: 112.620374 samples/sec	 accuracy=82.385326	 loss=0.666174	 lr=0.001000
Epoch[067] Batch [0939]/[1203]	Speed: 136.060543 samples/sec	 accuracy=82.379255	 loss=0.666017	 lr=0.001000
Epoch[067] Batch [0959]/[1203]	Speed: 125.972321 samples/sec	 accuracy=82.355729	 loss=0.666452	 lr=0.001000
Epoch[067] Batch [0979]/[1203]	Speed: 131.907840 samples/sec	 accuracy=82.347959	 loss=0.666838	 lr=0.001000
Epoch[067] Batch [0999]/[1203]	Speed: 121.961577 samples/sec	 accuracy=82.333500	 loss=0.667489	 lr=0.001000
Epoch[067] Batch [1019]/[1203]	Speed: 120.990320 samples/sec	 accuracy=82.325980	 loss=0.667555	 lr=0.001000
Epoch[067] Batch [1039]/[1203]	Speed: 148.556480 samples/sec	 accuracy=82.313942	 loss=0.667601	 lr=0.001000
Epoch[067] Batch [1059]/[1203]	Speed: 118.587893 samples/sec	 accuracy=82.304245	 loss=0.667977	 lr=0.001000
Epoch[067] Batch [1079]/[1203]	Speed: 106.636896 samples/sec	 accuracy=82.281019	 loss=0.668694	 lr=0.001000
Epoch[067] Batch [1099]/[1203]	Speed: 145.813146 samples/sec	 accuracy=82.285000	 loss=0.668685	 lr=0.001000
Epoch[067] Batch [1119]/[1203]	Speed: 123.860090 samples/sec	 accuracy=82.286607	 loss=0.668880	 lr=0.001000
Epoch[067] Batch [1139]/[1203]	Speed: 127.023263 samples/sec	 accuracy=82.301316	 loss=0.668345	 lr=0.001000
Epoch[067] Batch [1159]/[1203]	Speed: 112.586273 samples/sec	 accuracy=82.284914	 loss=0.668990	 lr=0.001000
Epoch[067] Batch [1179]/[1203]	Speed: 152.043550 samples/sec	 accuracy=82.272881	 loss=0.669290	 lr=0.001000
Epoch[067] Batch [1199]/[1203]	Speed: 231.420799 samples/sec	 accuracy=82.277500	 loss=0.669137	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.100000 acc-top5=85.350000
Batch [0039]/[0097]: acc-top1=62.987500 acc-top5=84.962500
Batch [0059]/[0097]: acc-top1=65.166667 acc-top5=86.191667
Batch [0079]/[0097]: acc-top1=65.306250 acc-top5=85.956250
[Epoch 067] training: accuracy=82.273067	 loss=0.669304
[Epoch 067] speed: 125 samples/sec	time cost: 2111.433243
[Epoch 067] validation: acc-top1=64.840206 acc-top5=85.814433 loss=1.559764
Epoch[068] Batch [0019]/[1203]	Speed: 69.980705 samples/sec	 accuracy=82.475000	 loss=0.658835	 lr=0.001000
Epoch[068] Batch [0039]/[1203]	Speed: 104.610901 samples/sec	 accuracy=82.325000	 loss=0.668972	 lr=0.001000
Epoch[068] Batch [0059]/[1203]	Speed: 211.820607 samples/sec	 accuracy=82.500000	 loss=0.665491	 lr=0.001000
Epoch[068] Batch [0079]/[1203]	Speed: 96.410231 samples/sec	 accuracy=82.593750	 loss=0.661402	 lr=0.001000
Epoch[068] Batch [0099]/[1203]	Speed: 132.757433 samples/sec	 accuracy=82.595000	 loss=0.661063	 lr=0.001000
Epoch[068] Batch [0119]/[1203]	Speed: 141.206437 samples/sec	 accuracy=82.491667	 loss=0.661853	 lr=0.001000
Epoch[068] Batch [0139]/[1203]	Speed: 107.603499 samples/sec	 accuracy=82.517857	 loss=0.660841	 lr=0.001000
Epoch[068] Batch [0159]/[1203]	Speed: 194.856000 samples/sec	 accuracy=82.643750	 loss=0.657589	 lr=0.001000
Epoch[068] Batch [0179]/[1203]	Speed: 95.396178 samples/sec	 accuracy=82.600000	 loss=0.661666	 lr=0.001000
Epoch[068] Batch [0199]/[1203]	Speed: 133.839171 samples/sec	 accuracy=82.575000	 loss=0.663922	 lr=0.001000
Epoch[068] Batch [0219]/[1203]	Speed: 121.171075 samples/sec	 accuracy=82.670455	 loss=0.659940	 lr=0.001000
Epoch[068] Batch [0239]/[1203]	Speed: 115.773143 samples/sec	 accuracy=82.706250	 loss=0.659452	 lr=0.001000
Epoch[068] Batch [0259]/[1203]	Speed: 182.973839 samples/sec	 accuracy=82.750000	 loss=0.656944	 lr=0.001000
Epoch[068] Batch [0279]/[1203]	Speed: 98.483855 samples/sec	 accuracy=82.735714	 loss=0.656552	 lr=0.001000
Epoch[068] Batch [0299]/[1203]	Speed: 176.223886 samples/sec	 accuracy=82.703333	 loss=0.658381	 lr=0.001000
Epoch[068] Batch [0319]/[1203]	Speed: 117.053766 samples/sec	 accuracy=82.656250	 loss=0.660031	 lr=0.001000
Epoch[068] Batch [0339]/[1203]	Speed: 102.173841 samples/sec	 accuracy=82.625000	 loss=0.660733	 lr=0.001000
Epoch[068] Batch [0359]/[1203]	Speed: 169.785222 samples/sec	 accuracy=82.656944	 loss=0.660377	 lr=0.001000
Epoch[068] Batch [0379]/[1203]	Speed: 103.930804 samples/sec	 accuracy=82.661842	 loss=0.660161	 lr=0.001000
Epoch[068] Batch [0399]/[1203]	Speed: 140.611358 samples/sec	 accuracy=82.645000	 loss=0.660338	 lr=0.001000
Epoch[068] Batch [0419]/[1203]	Speed: 139.041974 samples/sec	 accuracy=82.658333	 loss=0.660743	 lr=0.001000
Epoch[068] Batch [0439]/[1203]	Speed: 97.207596 samples/sec	 accuracy=82.680682	 loss=0.659836	 lr=0.001000
Epoch[068] Batch [0459]/[1203]	Speed: 174.228578 samples/sec	 accuracy=82.682609	 loss=0.659859	 lr=0.001000
Epoch[068] Batch [0479]/[1203]	Speed: 108.646627 samples/sec	 accuracy=82.708333	 loss=0.659011	 lr=0.001000
Epoch[068] Batch [0499]/[1203]	Speed: 134.741717 samples/sec	 accuracy=82.671000	 loss=0.659360	 lr=0.001000
Epoch[068] Batch [0519]/[1203]	Speed: 141.297893 samples/sec	 accuracy=82.657692	 loss=0.659879	 lr=0.001000
Epoch[068] Batch [0539]/[1203]	Speed: 98.488006 samples/sec	 accuracy=82.706481	 loss=0.659036	 lr=0.001000
Epoch[068] Batch [0559]/[1203]	Speed: 179.078884 samples/sec	 accuracy=82.705357	 loss=0.658943	 lr=0.001000
Epoch[068] Batch [0579]/[1203]	Speed: 109.249282 samples/sec	 accuracy=82.735345	 loss=0.658097	 lr=0.001000
Epoch[068] Batch [0599]/[1203]	Speed: 149.029678 samples/sec	 accuracy=82.727500	 loss=0.658626	 lr=0.001000
Epoch[068] Batch [0619]/[1203]	Speed: 120.673749 samples/sec	 accuracy=82.696774	 loss=0.659356	 lr=0.001000
Epoch[068] Batch [0639]/[1203]	Speed: 105.613161 samples/sec	 accuracy=82.671094	 loss=0.659980	 lr=0.001000
Epoch[068] Batch [0659]/[1203]	Speed: 147.465147 samples/sec	 accuracy=82.673485	 loss=0.660607	 lr=0.001000
Epoch[068] Batch [0679]/[1203]	Speed: 122.120941 samples/sec	 accuracy=82.644118	 loss=0.660750	 lr=0.001000
Epoch[068] Batch [0699]/[1203]	Speed: 118.744599 samples/sec	 accuracy=82.635000	 loss=0.660919	 lr=0.001000
Epoch[068] Batch [0719]/[1203]	Speed: 164.677978 samples/sec	 accuracy=82.634028	 loss=0.660854	 lr=0.001000
Epoch[068] Batch [0739]/[1203]	Speed: 101.505239 samples/sec	 accuracy=82.627027	 loss=0.660970	 lr=0.001000
Epoch[068] Batch [0759]/[1203]	Speed: 141.158802 samples/sec	 accuracy=82.623684	 loss=0.661124	 lr=0.001000
Epoch[068] Batch [0779]/[1203]	Speed: 125.939664 samples/sec	 accuracy=82.624359	 loss=0.660830	 lr=0.001000
Epoch[068] Batch [0799]/[1203]	Speed: 120.963551 samples/sec	 accuracy=82.645000	 loss=0.660253	 lr=0.001000
Epoch[068] Batch [0819]/[1203]	Speed: 138.792690 samples/sec	 accuracy=82.657317	 loss=0.660014	 lr=0.001000
Epoch[068] Batch [0839]/[1203]	Speed: 112.124706 samples/sec	 accuracy=82.631548	 loss=0.660955	 lr=0.001000
Epoch[068] Batch [0859]/[1203]	Speed: 124.925127 samples/sec	 accuracy=82.623837	 loss=0.661520	 lr=0.001000
Epoch[068] Batch [0879]/[1203]	Speed: 129.277723 samples/sec	 accuracy=82.625568	 loss=0.661308	 lr=0.001000
Epoch[068] Batch [0899]/[1203]	Speed: 125.588627 samples/sec	 accuracy=82.630000	 loss=0.661522	 lr=0.001000
Epoch[068] Batch [0919]/[1203]	Speed: 137.300139 samples/sec	 accuracy=82.612500	 loss=0.662219	 lr=0.001000
Epoch[068] Batch [0939]/[1203]	Speed: 120.614470 samples/sec	 accuracy=82.614894	 loss=0.661943	 lr=0.001000
Epoch[068] Batch [0959]/[1203]	Speed: 118.029953 samples/sec	 accuracy=82.608854	 loss=0.662502	 lr=0.001000
Epoch[068] Batch [0979]/[1203]	Speed: 129.996829 samples/sec	 accuracy=82.591837	 loss=0.663041	 lr=0.001000
Epoch[068] Batch [0999]/[1203]	Speed: 131.274450 samples/sec	 accuracy=82.578500	 loss=0.663414	 lr=0.001000
Epoch[068] Batch [1019]/[1203]	Speed: 130.729655 samples/sec	 accuracy=82.575490	 loss=0.663720	 lr=0.001000
Epoch[068] Batch [1039]/[1203]	Speed: 121.277150 samples/sec	 accuracy=82.576442	 loss=0.663624	 lr=0.001000
Epoch[068] Batch [1059]/[1203]	Speed: 121.154991 samples/sec	 accuracy=82.593396	 loss=0.663420	 lr=0.001000
Epoch[068] Batch [1079]/[1203]	Speed: 139.007094 samples/sec	 accuracy=82.578704	 loss=0.664352	 lr=0.001000
Epoch[068] Batch [1099]/[1203]	Speed: 119.229965 samples/sec	 accuracy=82.560909	 loss=0.664852	 lr=0.001000
Epoch[068] Batch [1119]/[1203]	Speed: 121.247136 samples/sec	 accuracy=82.576786	 loss=0.664405	 lr=0.001000
Epoch[068] Batch [1139]/[1203]	Speed: 135.641812 samples/sec	 accuracy=82.558772	 loss=0.665263	 lr=0.001000
Epoch[068] Batch [1159]/[1203]	Speed: 128.265356 samples/sec	 accuracy=82.565948	 loss=0.665091	 lr=0.001000
Epoch[068] Batch [1179]/[1203]	Speed: 128.050136 samples/sec	 accuracy=82.570339	 loss=0.665175	 lr=0.001000
Epoch[068] Batch [1199]/[1203]	Speed: 233.808985 samples/sec	 accuracy=82.572917	 loss=0.665112	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.325000 acc-top5=85.200000
Batch [0039]/[0097]: acc-top1=63.050000 acc-top5=84.837500
Batch [0059]/[0097]: acc-top1=65.075000 acc-top5=85.975000
Batch [0079]/[0097]: acc-top1=65.406250 acc-top5=85.868750
[Epoch 068] training: accuracy=82.572735	 loss=0.665237
[Epoch 068] speed: 125 samples/sec	time cost: 2107.952018
[Epoch 068] validation: acc-top1=64.881443 acc-top5=85.680412 loss=1.562208
Epoch[069] Batch [0019]/[1203]	Speed: 68.496977 samples/sec	 accuracy=82.375000	 loss=0.660030	 lr=0.001000
Epoch[069] Batch [0039]/[1203]	Speed: 100.983023 samples/sec	 accuracy=83.012500	 loss=0.641603	 lr=0.001000
Epoch[069] Batch [0059]/[1203]	Speed: 227.150611 samples/sec	 accuracy=82.483333	 loss=0.651344	 lr=0.001000
Epoch[069] Batch [0079]/[1203]	Speed: 94.924554 samples/sec	 accuracy=82.325000	 loss=0.655965	 lr=0.001000
Epoch[069] Batch [0099]/[1203]	Speed: 114.688548 samples/sec	 accuracy=82.485000	 loss=0.656508	 lr=0.001000
Epoch[069] Batch [0119]/[1203]	Speed: 177.360374 samples/sec	 accuracy=82.654167	 loss=0.652080	 lr=0.001000
Epoch[069] Batch [0139]/[1203]	Speed: 100.276762 samples/sec	 accuracy=82.725000	 loss=0.651806	 lr=0.001000
Epoch[069] Batch [0159]/[1203]	Speed: 205.104532 samples/sec	 accuracy=82.615625	 loss=0.655143	 lr=0.001000
Epoch[069] Batch [0179]/[1203]	Speed: 96.168273 samples/sec	 accuracy=82.702778	 loss=0.654397	 lr=0.001000
Epoch[069] Batch [0199]/[1203]	Speed: 111.949094 samples/sec	 accuracy=82.712500	 loss=0.653193	 lr=0.001000
Epoch[069] Batch [0219]/[1203]	Speed: 175.902447 samples/sec	 accuracy=82.722727	 loss=0.652492	 lr=0.001000
Epoch[069] Batch [0239]/[1203]	Speed: 107.232355 samples/sec	 accuracy=82.720833	 loss=0.654219	 lr=0.001000
Epoch[069] Batch [0259]/[1203]	Speed: 134.611569 samples/sec	 accuracy=82.719231	 loss=0.654112	 lr=0.001000
Epoch[069] Batch [0279]/[1203]	Speed: 121.543168 samples/sec	 accuracy=82.796429	 loss=0.652517	 lr=0.001000
Epoch[069] Batch [0299]/[1203]	Speed: 114.563735 samples/sec	 accuracy=82.831667	 loss=0.651561	 lr=0.001000
Epoch[069] Batch [0319]/[1203]	Speed: 170.794888 samples/sec	 accuracy=82.887500	 loss=0.651871	 lr=0.001000
Epoch[069] Batch [0339]/[1203]	Speed: 106.455106 samples/sec	 accuracy=82.894118	 loss=0.652053	 lr=0.001000
Epoch[069] Batch [0359]/[1203]	Speed: 109.915110 samples/sec	 accuracy=82.947222	 loss=0.650886	 lr=0.001000
Epoch[069] Batch [0379]/[1203]	Speed: 152.864705 samples/sec	 accuracy=82.972368	 loss=0.650772	 lr=0.001000
Epoch[069] Batch [0399]/[1203]	Speed: 111.066161 samples/sec	 accuracy=82.967500	 loss=0.651355	 lr=0.001000
Epoch[069] Batch [0419]/[1203]	Speed: 180.388645 samples/sec	 accuracy=82.932143	 loss=0.652664	 lr=0.001000
Epoch[069] Batch [0439]/[1203]	Speed: 102.602138 samples/sec	 accuracy=82.886364	 loss=0.653504	 lr=0.001000
Epoch[069] Batch [0459]/[1203]	Speed: 126.055496 samples/sec	 accuracy=82.823913	 loss=0.654736	 lr=0.001000
Epoch[069] Batch [0479]/[1203]	Speed: 143.967482 samples/sec	 accuracy=82.832292	 loss=0.655352	 lr=0.001000
Epoch[069] Batch [0499]/[1203]	Speed: 108.322189 samples/sec	 accuracy=82.830000	 loss=0.655683	 lr=0.001000
Epoch[069] Batch [0519]/[1203]	Speed: 154.538823 samples/sec	 accuracy=82.799038	 loss=0.656823	 lr=0.001000
Epoch[069] Batch [0539]/[1203]	Speed: 115.088879 samples/sec	 accuracy=82.809259	 loss=0.655816	 lr=0.001000
Epoch[069] Batch [0559]/[1203]	Speed: 124.388880 samples/sec	 accuracy=82.810714	 loss=0.655570	 lr=0.001000
Epoch[069] Batch [0579]/[1203]	Speed: 134.285042 samples/sec	 accuracy=82.817241	 loss=0.655191	 lr=0.001000
Epoch[069] Batch [0599]/[1203]	Speed: 111.490431 samples/sec	 accuracy=82.820833	 loss=0.654640	 lr=0.001000
Epoch[069] Batch [0619]/[1203]	Speed: 138.530036 samples/sec	 accuracy=82.807258	 loss=0.655684	 lr=0.001000
Epoch[069] Batch [0639]/[1203]	Speed: 129.542469 samples/sec	 accuracy=82.802344	 loss=0.656052	 lr=0.001000
Epoch[069] Batch [0659]/[1203]	Speed: 125.165158 samples/sec	 accuracy=82.790909	 loss=0.656800	 lr=0.001000
Epoch[069] Batch [0679]/[1203]	Speed: 128.242665 samples/sec	 accuracy=82.752941	 loss=0.657709	 lr=0.001000
Epoch[069] Batch [0699]/[1203]	Speed: 117.029429 samples/sec	 accuracy=82.798571	 loss=0.656890	 lr=0.001000
Epoch[069] Batch [0719]/[1203]	Speed: 128.209040 samples/sec	 accuracy=82.800694	 loss=0.657070	 lr=0.001000
Epoch[069] Batch [0739]/[1203]	Speed: 125.909429 samples/sec	 accuracy=82.777703	 loss=0.658209	 lr=0.001000
Epoch[069] Batch [0759]/[1203]	Speed: 121.281491 samples/sec	 accuracy=82.773684	 loss=0.658178	 lr=0.001000
Epoch[069] Batch [0779]/[1203]	Speed: 128.712555 samples/sec	 accuracy=82.777564	 loss=0.658446	 lr=0.001000
Epoch[069] Batch [0799]/[1203]	Speed: 127.990117 samples/sec	 accuracy=82.750625	 loss=0.659470	 lr=0.001000
Epoch[069] Batch [0819]/[1203]	Speed: 116.617959 samples/sec	 accuracy=82.760976	 loss=0.659099	 lr=0.001000
Epoch[069] Batch [0839]/[1203]	Speed: 142.243157 samples/sec	 accuracy=82.758333	 loss=0.658881	 lr=0.001000
Epoch[069] Batch [0859]/[1203]	Speed: 121.264505 samples/sec	 accuracy=82.779651	 loss=0.658607	 lr=0.001000
Epoch[069] Batch [0879]/[1203]	Speed: 119.822790 samples/sec	 accuracy=82.757955	 loss=0.659235	 lr=0.001000
Epoch[069] Batch [0899]/[1203]	Speed: 141.002656 samples/sec	 accuracy=82.730000	 loss=0.659749	 lr=0.001000
Epoch[069] Batch [0919]/[1203]	Speed: 120.272404 samples/sec	 accuracy=82.684239	 loss=0.660999	 lr=0.001000
Epoch[069] Batch [0939]/[1203]	Speed: 126.069094 samples/sec	 accuracy=82.688298	 loss=0.660226	 lr=0.001000
Epoch[069] Batch [0959]/[1203]	Speed: 135.572633 samples/sec	 accuracy=82.675521	 loss=0.660341	 lr=0.001000
Epoch[069] Batch [0979]/[1203]	Speed: 115.988523 samples/sec	 accuracy=82.651531	 loss=0.660735	 lr=0.001000
Epoch[069] Batch [0999]/[1203]	Speed: 113.404200 samples/sec	 accuracy=82.648000	 loss=0.660690	 lr=0.001000
Epoch[069] Batch [1019]/[1203]	Speed: 141.093310 samples/sec	 accuracy=82.664706	 loss=0.660421	 lr=0.001000
Epoch[069] Batch [1039]/[1203]	Speed: 135.100954 samples/sec	 accuracy=82.684615	 loss=0.659829	 lr=0.001000
Epoch[069] Batch [1059]/[1203]	Speed: 133.146060 samples/sec	 accuracy=82.696698	 loss=0.659783	 lr=0.001000
Epoch[069] Batch [1079]/[1203]	Speed: 114.225385 samples/sec	 accuracy=82.694907	 loss=0.660021	 lr=0.001000
Epoch[069] Batch [1099]/[1203]	Speed: 120.079333 samples/sec	 accuracy=82.680000	 loss=0.660220	 lr=0.001000
Epoch[069] Batch [1119]/[1203]	Speed: 139.815978 samples/sec	 accuracy=82.666518	 loss=0.660881	 lr=0.001000
Epoch[069] Batch [1139]/[1203]	Speed: 127.738504 samples/sec	 accuracy=82.656140	 loss=0.661360	 lr=0.001000
Epoch[069] Batch [1159]/[1203]	Speed: 126.432588 samples/sec	 accuracy=82.651724	 loss=0.661285	 lr=0.001000
Epoch[069] Batch [1179]/[1203]	Speed: 129.534398 samples/sec	 accuracy=82.633898	 loss=0.661756	 lr=0.001000
Epoch[069] Batch [1199]/[1203]	Speed: 235.303642 samples/sec	 accuracy=82.643333	 loss=0.661620	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.350000 acc-top5=85.000000
Batch [0039]/[0097]: acc-top1=63.012500 acc-top5=84.800000
Batch [0059]/[0097]: acc-top1=65.275000 acc-top5=85.958333
Batch [0079]/[0097]: acc-top1=65.493750 acc-top5=85.868750
[Epoch 069] training: accuracy=82.644638	 loss=0.661468
[Epoch 069] speed: 125 samples/sec	time cost: 2115.102926
[Epoch 069] validation: acc-top1=64.963918 acc-top5=85.793814 loss=1.554221
Epoch[070] Batch [0019]/[1203]	Speed: 70.232771 samples/sec	 accuracy=83.000000	 loss=0.668939	 lr=0.001000
Epoch[070] Batch [0039]/[1203]	Speed: 102.460262 samples/sec	 accuracy=83.075000	 loss=0.661322	 lr=0.001000
Epoch[070] Batch [0059]/[1203]	Speed: 217.789738 samples/sec	 accuracy=83.183333	 loss=0.653759	 lr=0.001000
Epoch[070] Batch [0079]/[1203]	Speed: 101.677055 samples/sec	 accuracy=83.243750	 loss=0.651274	 lr=0.001000
Epoch[070] Batch [0099]/[1203]	Speed: 106.085047 samples/sec	 accuracy=83.430000	 loss=0.642789	 lr=0.001000
Epoch[070] Batch [0119]/[1203]	Speed: 184.951641 samples/sec	 accuracy=83.291667	 loss=0.643725	 lr=0.001000
Epoch[070] Batch [0139]/[1203]	Speed: 96.704657 samples/sec	 accuracy=83.367857	 loss=0.643174	 lr=0.001000
Epoch[070] Batch [0159]/[1203]	Speed: 222.247459 samples/sec	 accuracy=83.478125	 loss=0.643340	 lr=0.001000
Epoch[070] Batch [0179]/[1203]	Speed: 101.953290 samples/sec	 accuracy=83.413889	 loss=0.644461	 lr=0.001000
Epoch[070] Batch [0199]/[1203]	Speed: 118.766403 samples/sec	 accuracy=83.380000	 loss=0.641232	 lr=0.001000
Epoch[070] Batch [0219]/[1203]	Speed: 160.794573 samples/sec	 accuracy=83.288636	 loss=0.642257	 lr=0.001000
Epoch[070] Batch [0239]/[1203]	Speed: 100.244196 samples/sec	 accuracy=83.225000	 loss=0.644791	 lr=0.001000
Epoch[070] Batch [0259]/[1203]	Speed: 158.234279 samples/sec	 accuracy=83.155769	 loss=0.645895	 lr=0.001000
Epoch[070] Batch [0279]/[1203]	Speed: 115.866595 samples/sec	 accuracy=83.123214	 loss=0.647474	 lr=0.001000
Epoch[070] Batch [0299]/[1203]	Speed: 107.428698 samples/sec	 accuracy=83.028333	 loss=0.650033	 lr=0.001000
Epoch[070] Batch [0319]/[1203]	Speed: 191.254540 samples/sec	 accuracy=83.035938	 loss=0.650534	 lr=0.001000
Epoch[070] Batch [0339]/[1203]	Speed: 101.164225 samples/sec	 accuracy=83.044118	 loss=0.651711	 lr=0.001000
Epoch[070] Batch [0359]/[1203]	Speed: 126.428843 samples/sec	 accuracy=83.044444	 loss=0.651303	 lr=0.001000
Epoch[070] Batch [0379]/[1203]	Speed: 139.001602 samples/sec	 accuracy=83.027632	 loss=0.652323	 lr=0.001000
Epoch[070] Batch [0399]/[1203]	Speed: 120.321394 samples/sec	 accuracy=83.015000	 loss=0.652554	 lr=0.001000
Epoch[070] Batch [0419]/[1203]	Speed: 159.593567 samples/sec	 accuracy=82.994048	 loss=0.652946	 lr=0.001000
Epoch[070] Batch [0439]/[1203]	Speed: 99.758419 samples/sec	 accuracy=82.985227	 loss=0.653357	 lr=0.001000
Epoch[070] Batch [0459]/[1203]	Speed: 135.438266 samples/sec	 accuracy=82.960870	 loss=0.654195	 lr=0.001000
Epoch[070] Batch [0479]/[1203]	Speed: 132.128548 samples/sec	 accuracy=82.972917	 loss=0.653674	 lr=0.001000
Epoch[070] Batch [0499]/[1203]	Speed: 111.061383 samples/sec	 accuracy=82.989000	 loss=0.652551	 lr=0.001000
Epoch[070] Batch [0519]/[1203]	Speed: 152.767949 samples/sec	 accuracy=82.972115	 loss=0.652894	 lr=0.001000
Epoch[070] Batch [0539]/[1203]	Speed: 113.393638 samples/sec	 accuracy=82.951852	 loss=0.653719	 lr=0.001000
Epoch[070] Batch [0559]/[1203]	Speed: 134.997343 samples/sec	 accuracy=82.953571	 loss=0.653287	 lr=0.001000
Epoch[070] Batch [0579]/[1203]	Speed: 133.788147 samples/sec	 accuracy=82.966379	 loss=0.652639	 lr=0.001000
Epoch[070] Batch [0599]/[1203]	Speed: 105.354285 samples/sec	 accuracy=82.935000	 loss=0.652896	 lr=0.001000
Epoch[070] Batch [0619]/[1203]	Speed: 160.076769 samples/sec	 accuracy=82.941129	 loss=0.652927	 lr=0.001000
Epoch[070] Batch [0639]/[1203]	Speed: 112.599460 samples/sec	 accuracy=82.920313	 loss=0.653668	 lr=0.001000
Epoch[070] Batch [0659]/[1203]	Speed: 128.156735 samples/sec	 accuracy=82.887121	 loss=0.654199	 lr=0.001000
Epoch[070] Batch [0679]/[1203]	Speed: 141.789956 samples/sec	 accuracy=82.840441	 loss=0.655561	 lr=0.001000
Epoch[070] Batch [0699]/[1203]	Speed: 113.948788 samples/sec	 accuracy=82.843571	 loss=0.655376	 lr=0.001000
Epoch[070] Batch [0719]/[1203]	Speed: 125.970659 samples/sec	 accuracy=82.828472	 loss=0.656291	 lr=0.001000
Epoch[070] Batch [0739]/[1203]	Speed: 129.786833 samples/sec	 accuracy=82.829054	 loss=0.655886	 lr=0.001000
Epoch[070] Batch [0759]/[1203]	Speed: 128.896782 samples/sec	 accuracy=82.836842	 loss=0.655299	 lr=0.001000
Epoch[070] Batch [0779]/[1203]	Speed: 136.721991 samples/sec	 accuracy=82.824359	 loss=0.655292	 lr=0.001000
Epoch[070] Batch [0799]/[1203]	Speed: 123.631117 samples/sec	 accuracy=82.812500	 loss=0.655465	 lr=0.001000
Epoch[070] Batch [0819]/[1203]	Speed: 116.596103 samples/sec	 accuracy=82.807927	 loss=0.655490	 lr=0.001000
Epoch[070] Batch [0839]/[1203]	Speed: 127.553510 samples/sec	 accuracy=82.795833	 loss=0.655827	 lr=0.001000
Epoch[070] Batch [0859]/[1203]	Speed: 133.879357 samples/sec	 accuracy=82.801744	 loss=0.655941	 lr=0.001000
Epoch[070] Batch [0879]/[1203]	Speed: 125.741610 samples/sec	 accuracy=82.818750	 loss=0.655595	 lr=0.001000
Epoch[070] Batch [0899]/[1203]	Speed: 106.604668 samples/sec	 accuracy=82.835000	 loss=0.655514	 lr=0.001000
Epoch[070] Batch [0919]/[1203]	Speed: 152.805358 samples/sec	 accuracy=82.809239	 loss=0.656233	 lr=0.001000
Epoch[070] Batch [0939]/[1203]	Speed: 126.282400 samples/sec	 accuracy=82.797872	 loss=0.656494	 lr=0.001000
Epoch[070] Batch [0959]/[1203]	Speed: 119.985576 samples/sec	 accuracy=82.792708	 loss=0.656717	 lr=0.001000
Epoch[070] Batch [0979]/[1203]	Speed: 145.075445 samples/sec	 accuracy=82.790306	 loss=0.656624	 lr=0.001000
Epoch[070] Batch [0999]/[1203]	Speed: 109.491278 samples/sec	 accuracy=82.783000	 loss=0.656518	 lr=0.001000
Epoch[070] Batch [1019]/[1203]	Speed: 144.623545 samples/sec	 accuracy=82.798529	 loss=0.656463	 lr=0.001000
Epoch[070] Batch [1039]/[1203]	Speed: 111.252501 samples/sec	 accuracy=82.776923	 loss=0.656905	 lr=0.001000
Epoch[070] Batch [1059]/[1203]	Speed: 118.763268 samples/sec	 accuracy=82.775943	 loss=0.657034	 lr=0.001000
Epoch[070] Batch [1079]/[1203]	Speed: 151.282071 samples/sec	 accuracy=82.782407	 loss=0.656623	 lr=0.001000
Epoch[070] Batch [1099]/[1203]	Speed: 123.270145 samples/sec	 accuracy=82.787727	 loss=0.656656	 lr=0.001000
Epoch[070] Batch [1119]/[1203]	Speed: 130.739050 samples/sec	 accuracy=82.774107	 loss=0.657075	 lr=0.001000
Epoch[070] Batch [1139]/[1203]	Speed: 118.049779 samples/sec	 accuracy=82.778070	 loss=0.656841	 lr=0.001000
Epoch[070] Batch [1159]/[1203]	Speed: 118.179017 samples/sec	 accuracy=82.780172	 loss=0.656731	 lr=0.001000
Epoch[070] Batch [1179]/[1203]	Speed: 164.123344 samples/sec	 accuracy=82.768644	 loss=0.656812	 lr=0.001000
Epoch[070] Batch [1199]/[1203]	Speed: 207.583692 samples/sec	 accuracy=82.763750	 loss=0.656933	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.775000 acc-top5=85.150000
Batch [0039]/[0097]: acc-top1=63.137500 acc-top5=84.637500
Batch [0059]/[0097]: acc-top1=65.216667 acc-top5=85.816667
Batch [0079]/[0097]: acc-top1=65.462500 acc-top5=85.743750
[Epoch 070] training: accuracy=82.763508	 loss=0.656930
[Epoch 070] speed: 125 samples/sec	time cost: 2105.415833
[Epoch 070] validation: acc-top1=64.912371 acc-top5=85.664948 loss=1.564159
Epoch[071] Batch [0019]/[1203]	Speed: 70.464966 samples/sec	 accuracy=82.900000	 loss=0.640366	 lr=0.001000
Epoch[071] Batch [0039]/[1203]	Speed: 101.724107 samples/sec	 accuracy=83.587500	 loss=0.635822	 lr=0.001000
Epoch[071] Batch [0059]/[1203]	Speed: 232.127094 samples/sec	 accuracy=83.458333	 loss=0.640207	 lr=0.001000
Epoch[071] Batch [0079]/[1203]	Speed: 96.370561 samples/sec	 accuracy=83.456250	 loss=0.642196	 lr=0.001000
Epoch[071] Batch [0099]/[1203]	Speed: 102.405418 samples/sec	 accuracy=83.370000	 loss=0.645144	 lr=0.001000
Epoch[071] Batch [0119]/[1203]	Speed: 212.568731 samples/sec	 accuracy=83.208333	 loss=0.651002	 lr=0.001000
Epoch[071] Batch [0139]/[1203]	Speed: 97.797326 samples/sec	 accuracy=83.082143	 loss=0.655333	 lr=0.001000
Epoch[071] Batch [0159]/[1203]	Speed: 247.622967 samples/sec	 accuracy=83.143750	 loss=0.653444	 lr=0.001000
Epoch[071] Batch [0179]/[1203]	Speed: 97.656707 samples/sec	 accuracy=83.180556	 loss=0.651296	 lr=0.001000
Epoch[071] Batch [0199]/[1203]	Speed: 94.450150 samples/sec	 accuracy=83.162500	 loss=0.653093	 lr=0.001000
Epoch[071] Batch [0219]/[1203]	Speed: 227.411237 samples/sec	 accuracy=83.152273	 loss=0.653944	 lr=0.001000
Epoch[071] Batch [0239]/[1203]	Speed: 98.490181 samples/sec	 accuracy=83.141667	 loss=0.653126	 lr=0.001000
Epoch[071] Batch [0259]/[1203]	Speed: 143.993033 samples/sec	 accuracy=83.107692	 loss=0.651433	 lr=0.001000
Epoch[071] Batch [0279]/[1203]	Speed: 130.946086 samples/sec	 accuracy=83.041071	 loss=0.652824	 lr=0.001000
Epoch[071] Batch [0299]/[1203]	Speed: 100.704348 samples/sec	 accuracy=83.048333	 loss=0.651034	 lr=0.001000
Epoch[071] Batch [0319]/[1203]	Speed: 196.013224 samples/sec	 accuracy=83.004688	 loss=0.651055	 lr=0.001000
Epoch[071] Batch [0339]/[1203]	Speed: 103.638765 samples/sec	 accuracy=83.017647	 loss=0.650315	 lr=0.001000
Epoch[071] Batch [0359]/[1203]	Speed: 116.765369 samples/sec	 accuracy=83.004167	 loss=0.651665	 lr=0.001000
Epoch[071] Batch [0379]/[1203]	Speed: 150.914075 samples/sec	 accuracy=82.990789	 loss=0.652492	 lr=0.001000
Epoch[071] Batch [0399]/[1203]	Speed: 106.007093 samples/sec	 accuracy=82.992500	 loss=0.652680	 lr=0.001000
Epoch[071] Batch [0419]/[1203]	Speed: 183.282492 samples/sec	 accuracy=83.001190	 loss=0.652355	 lr=0.001000
Epoch[071] Batch [0439]/[1203]	Speed: 96.320579 samples/sec	 accuracy=83.029545	 loss=0.652130	 lr=0.001000
Epoch[071] Batch [0459]/[1203]	Speed: 134.376743 samples/sec	 accuracy=82.994565	 loss=0.653597	 lr=0.001000
Epoch[071] Batch [0479]/[1203]	Speed: 143.903849 samples/sec	 accuracy=82.991667	 loss=0.653698	 lr=0.001000
Epoch[071] Batch [0499]/[1203]	Speed: 109.521886 samples/sec	 accuracy=82.983000	 loss=0.654157	 lr=0.001000
Epoch[071] Batch [0519]/[1203]	Speed: 161.878244 samples/sec	 accuracy=83.005769	 loss=0.652489	 lr=0.001000
Epoch[071] Batch [0539]/[1203]	Speed: 105.312142 samples/sec	 accuracy=83.010185	 loss=0.652428	 lr=0.001000
Epoch[071] Batch [0559]/[1203]	Speed: 132.973957 samples/sec	 accuracy=83.005357	 loss=0.652473	 lr=0.001000
Epoch[071] Batch [0579]/[1203]	Speed: 136.268423 samples/sec	 accuracy=82.986207	 loss=0.653479	 lr=0.001000
Epoch[071] Batch [0599]/[1203]	Speed: 118.156375 samples/sec	 accuracy=83.006667	 loss=0.652409	 lr=0.001000
Epoch[071] Batch [0619]/[1203]	Speed: 121.504109 samples/sec	 accuracy=83.011290	 loss=0.652370	 lr=0.001000
Epoch[071] Batch [0639]/[1203]	Speed: 125.807759 samples/sec	 accuracy=82.995312	 loss=0.652920	 lr=0.001000
Epoch[071] Batch [0659]/[1203]	Speed: 127.713516 samples/sec	 accuracy=83.011364	 loss=0.652307	 lr=0.001000
Epoch[071] Batch [0679]/[1203]	Speed: 121.217795 samples/sec	 accuracy=83.005882	 loss=0.652211	 lr=0.001000
Epoch[071] Batch [0699]/[1203]	Speed: 134.447991 samples/sec	 accuracy=82.986429	 loss=0.652805	 lr=0.001000
Epoch[071] Batch [0719]/[1203]	Speed: 116.248408 samples/sec	 accuracy=82.994444	 loss=0.653088	 lr=0.001000
Epoch[071] Batch [0739]/[1203]	Speed: 140.139805 samples/sec	 accuracy=82.995946	 loss=0.652966	 lr=0.001000
Epoch[071] Batch [0759]/[1203]	Speed: 116.188953 samples/sec	 accuracy=82.997368	 loss=0.652669	 lr=0.001000
Epoch[071] Batch [0779]/[1203]	Speed: 133.767410 samples/sec	 accuracy=82.994872	 loss=0.652869	 lr=0.001000
Epoch[071] Batch [0799]/[1203]	Speed: 122.462072 samples/sec	 accuracy=83.010625	 loss=0.653018	 lr=0.001000
Epoch[071] Batch [0819]/[1203]	Speed: 125.179557 samples/sec	 accuracy=83.005488	 loss=0.652679	 lr=0.001000
Epoch[071] Batch [0839]/[1203]	Speed: 127.075775 samples/sec	 accuracy=82.994643	 loss=0.653070	 lr=0.001000
Epoch[071] Batch [0859]/[1203]	Speed: 116.202554 samples/sec	 accuracy=83.004070	 loss=0.652800	 lr=0.001000
Epoch[071] Batch [0879]/[1203]	Speed: 146.436390 samples/sec	 accuracy=83.003409	 loss=0.652721	 lr=0.001000
Epoch[071] Batch [0899]/[1203]	Speed: 111.932492 samples/sec	 accuracy=83.018889	 loss=0.652397	 lr=0.001000
Epoch[071] Batch [0919]/[1203]	Speed: 128.498597 samples/sec	 accuracy=82.997283	 loss=0.653043	 lr=0.001000
Epoch[071] Batch [0939]/[1203]	Speed: 141.892125 samples/sec	 accuracy=82.999468	 loss=0.652907	 lr=0.001000
Epoch[071] Batch [0959]/[1203]	Speed: 111.737085 samples/sec	 accuracy=82.971354	 loss=0.653492	 lr=0.001000
Epoch[071] Batch [0979]/[1203]	Speed: 141.215685 samples/sec	 accuracy=82.984184	 loss=0.653207	 lr=0.001000
Epoch[071] Batch [0999]/[1203]	Speed: 111.777556 samples/sec	 accuracy=82.977500	 loss=0.653709	 lr=0.001000
Epoch[071] Batch [1019]/[1203]	Speed: 133.001961 samples/sec	 accuracy=82.970588	 loss=0.653681	 lr=0.001000
Epoch[071] Batch [1039]/[1203]	Speed: 131.285215 samples/sec	 accuracy=82.958654	 loss=0.653746	 lr=0.001000
Epoch[071] Batch [1059]/[1203]	Speed: 125.924176 samples/sec	 accuracy=82.957075	 loss=0.653513	 lr=0.001000
Epoch[071] Batch [1079]/[1203]	Speed: 133.657206 samples/sec	 accuracy=82.972685	 loss=0.653041	 lr=0.001000
Epoch[071] Batch [1099]/[1203]	Speed: 106.851501 samples/sec	 accuracy=82.954091	 loss=0.653499	 lr=0.001000
Epoch[071] Batch [1119]/[1203]	Speed: 130.678149 samples/sec	 accuracy=82.951339	 loss=0.653803	 lr=0.001000
Epoch[071] Batch [1139]/[1203]	Speed: 133.660930 samples/sec	 accuracy=82.942982	 loss=0.654425	 lr=0.001000
Epoch[071] Batch [1159]/[1203]	Speed: 131.387775 samples/sec	 accuracy=82.939655	 loss=0.654754	 lr=0.001000
Epoch[071] Batch [1179]/[1203]	Speed: 132.384123 samples/sec	 accuracy=82.936017	 loss=0.654337	 lr=0.001000
Epoch[071] Batch [1199]/[1203]	Speed: 226.359755 samples/sec	 accuracy=82.931250	 loss=0.654660	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.800000 acc-top5=85.125000
Batch [0039]/[0097]: acc-top1=63.325000 acc-top5=84.887500
Batch [0059]/[0097]: acc-top1=65.500000 acc-top5=86.091667
Batch [0079]/[0097]: acc-top1=65.631250 acc-top5=85.856250
[Epoch 071] training: accuracy=82.931837	 loss=0.654682
[Epoch 071] speed: 125 samples/sec	time cost: 2110.822800
[Epoch 071] validation: acc-top1=65.103093 acc-top5=85.716495 loss=1.562781
Epoch[072] Batch [0019]/[1203]	Speed: 70.482256 samples/sec	 accuracy=82.350000	 loss=0.657160	 lr=0.001000
Epoch[072] Batch [0039]/[1203]	Speed: 104.915665 samples/sec	 accuracy=82.812500	 loss=0.639600	 lr=0.001000
Epoch[072] Batch [0059]/[1203]	Speed: 223.090092 samples/sec	 accuracy=83.316667	 loss=0.633525	 lr=0.001000
Epoch[072] Batch [0079]/[1203]	Speed: 96.929620 samples/sec	 accuracy=83.125000	 loss=0.638214	 lr=0.001000
Epoch[072] Batch [0099]/[1203]	Speed: 108.174199 samples/sec	 accuracy=83.230000	 loss=0.635906	 lr=0.001000
Epoch[072] Batch [0119]/[1203]	Speed: 187.892288 samples/sec	 accuracy=83.175000	 loss=0.638105	 lr=0.001000
Epoch[072] Batch [0139]/[1203]	Speed: 93.721246 samples/sec	 accuracy=83.214286	 loss=0.637131	 lr=0.001000
Epoch[072] Batch [0159]/[1203]	Speed: 263.268052 samples/sec	 accuracy=83.243750	 loss=0.635749	 lr=0.001000
Epoch[072] Batch [0179]/[1203]	Speed: 94.008710 samples/sec	 accuracy=83.261111	 loss=0.635493	 lr=0.001000
Epoch[072] Batch [0199]/[1203]	Speed: 104.721298 samples/sec	 accuracy=83.252500	 loss=0.636484	 lr=0.001000
Epoch[072] Batch [0219]/[1203]	Speed: 203.653027 samples/sec	 accuracy=83.306818	 loss=0.636933	 lr=0.001000
Epoch[072] Batch [0239]/[1203]	Speed: 100.856425 samples/sec	 accuracy=83.347917	 loss=0.635742	 lr=0.001000
Epoch[072] Batch [0259]/[1203]	Speed: 125.863897 samples/sec	 accuracy=83.332692	 loss=0.636936	 lr=0.001000
Epoch[072] Batch [0279]/[1203]	Speed: 148.789784 samples/sec	 accuracy=83.298214	 loss=0.637729	 lr=0.001000
Epoch[072] Batch [0299]/[1203]	Speed: 105.110870 samples/sec	 accuracy=83.271667	 loss=0.639426	 lr=0.001000
Epoch[072] Batch [0319]/[1203]	Speed: 170.484135 samples/sec	 accuracy=83.314062	 loss=0.638727	 lr=0.001000
Epoch[072] Batch [0339]/[1203]	Speed: 113.159817 samples/sec	 accuracy=83.305882	 loss=0.638850	 lr=0.001000
Epoch[072] Batch [0359]/[1203]	Speed: 110.689550 samples/sec	 accuracy=83.305556	 loss=0.639170	 lr=0.001000
Epoch[072] Batch [0379]/[1203]	Speed: 149.540405 samples/sec	 accuracy=83.298684	 loss=0.639959	 lr=0.001000
Epoch[072] Batch [0399]/[1203]	Speed: 106.786405 samples/sec	 accuracy=83.318750	 loss=0.639747	 lr=0.001000
Epoch[072] Batch [0419]/[1203]	Speed: 169.301206 samples/sec	 accuracy=83.353571	 loss=0.639078	 lr=0.001000
Epoch[072] Batch [0439]/[1203]	Speed: 113.498771 samples/sec	 accuracy=83.367045	 loss=0.638888	 lr=0.001000
Epoch[072] Batch [0459]/[1203]	Speed: 106.354531 samples/sec	 accuracy=83.346739	 loss=0.639669	 lr=0.001000
Epoch[072] Batch [0479]/[1203]	Speed: 158.651720 samples/sec	 accuracy=83.325000	 loss=0.640214	 lr=0.001000
Epoch[072] Batch [0499]/[1203]	Speed: 110.025852 samples/sec	 accuracy=83.315000	 loss=0.640259	 lr=0.001000
Epoch[072] Batch [0519]/[1203]	Speed: 168.759345 samples/sec	 accuracy=83.298077	 loss=0.640639	 lr=0.001000
Epoch[072] Batch [0539]/[1203]	Speed: 111.318649 samples/sec	 accuracy=83.289815	 loss=0.641060	 lr=0.001000
Epoch[072] Batch [0559]/[1203]	Speed: 109.405689 samples/sec	 accuracy=83.291071	 loss=0.641170	 lr=0.001000
Epoch[072] Batch [0579]/[1203]	Speed: 157.324815 samples/sec	 accuracy=83.250000	 loss=0.642664	 lr=0.001000
Epoch[072] Batch [0599]/[1203]	Speed: 110.222415 samples/sec	 accuracy=83.199167	 loss=0.643856	 lr=0.001000
Epoch[072] Batch [0619]/[1203]	Speed: 119.955857 samples/sec	 accuracy=83.170968	 loss=0.643937	 lr=0.001000
Epoch[072] Batch [0639]/[1203]	Speed: 163.595838 samples/sec	 accuracy=83.144531	 loss=0.644814	 lr=0.001000
Epoch[072] Batch [0659]/[1203]	Speed: 106.906106 samples/sec	 accuracy=83.131061	 loss=0.644887	 lr=0.001000
Epoch[072] Batch [0679]/[1203]	Speed: 139.209048 samples/sec	 accuracy=83.107353	 loss=0.646201	 lr=0.001000
Epoch[072] Batch [0699]/[1203]	Speed: 117.844776 samples/sec	 accuracy=83.082143	 loss=0.647183	 lr=0.001000
Epoch[072] Batch [0719]/[1203]	Speed: 120.327457 samples/sec	 accuracy=83.079861	 loss=0.647870	 lr=0.001000
Epoch[072] Batch [0739]/[1203]	Speed: 157.083204 samples/sec	 accuracy=83.025000	 loss=0.649011	 lr=0.001000
Epoch[072] Batch [0759]/[1203]	Speed: 102.824066 samples/sec	 accuracy=82.988816	 loss=0.649627	 lr=0.001000
Epoch[072] Batch [0779]/[1203]	Speed: 151.586664 samples/sec	 accuracy=82.976923	 loss=0.650172	 lr=0.001000
Epoch[072] Batch [0799]/[1203]	Speed: 130.412264 samples/sec	 accuracy=82.982500	 loss=0.649780	 lr=0.001000
Epoch[072] Batch [0819]/[1203]	Speed: 113.048525 samples/sec	 accuracy=82.956707	 loss=0.650452	 lr=0.001000
Epoch[072] Batch [0839]/[1203]	Speed: 132.700238 samples/sec	 accuracy=82.954762	 loss=0.650165	 lr=0.001000
Epoch[072] Batch [0859]/[1203]	Speed: 119.718509 samples/sec	 accuracy=82.927907	 loss=0.650893	 lr=0.001000
Epoch[072] Batch [0879]/[1203]	Speed: 120.592389 samples/sec	 accuracy=82.906818	 loss=0.651666	 lr=0.001000
Epoch[072] Batch [0899]/[1203]	Speed: 145.032506 samples/sec	 accuracy=82.908333	 loss=0.652055	 lr=0.001000
Epoch[072] Batch [0919]/[1203]	Speed: 117.333256 samples/sec	 accuracy=82.911957	 loss=0.651671	 lr=0.001000
Epoch[072] Batch [0939]/[1203]	Speed: 125.289246 samples/sec	 accuracy=82.916489	 loss=0.651135	 lr=0.001000
Epoch[072] Batch [0959]/[1203]	Speed: 134.988771 samples/sec	 accuracy=82.908854	 loss=0.651183	 lr=0.001000
Epoch[072] Batch [0979]/[1203]	Speed: 122.250350 samples/sec	 accuracy=82.917347	 loss=0.651295	 lr=0.001000
Epoch[072] Batch [0999]/[1203]	Speed: 122.872573 samples/sec	 accuracy=82.895000	 loss=0.651686	 lr=0.001000
Epoch[072] Batch [1019]/[1203]	Speed: 121.538182 samples/sec	 accuracy=82.909804	 loss=0.651383	 lr=0.001000
Epoch[072] Batch [1039]/[1203]	Speed: 140.080955 samples/sec	 accuracy=82.912019	 loss=0.651150	 lr=0.001000
Epoch[072] Batch [1059]/[1203]	Speed: 131.944223 samples/sec	 accuracy=82.893396	 loss=0.651753	 lr=0.001000
Epoch[072] Batch [1079]/[1203]	Speed: 122.718898 samples/sec	 accuracy=82.866667	 loss=0.652254	 lr=0.001000
Epoch[072] Batch [1099]/[1203]	Speed: 112.484154 samples/sec	 accuracy=82.868182	 loss=0.652367	 lr=0.001000
Epoch[072] Batch [1119]/[1203]	Speed: 120.994790 samples/sec	 accuracy=82.874107	 loss=0.652341	 lr=0.001000
Epoch[072] Batch [1139]/[1203]	Speed: 147.358138 samples/sec	 accuracy=82.887281	 loss=0.652037	 lr=0.001000
Epoch[072] Batch [1159]/[1203]	Speed: 116.322037 samples/sec	 accuracy=82.894397	 loss=0.651829	 lr=0.001000
Epoch[072] Batch [1179]/[1203]	Speed: 139.452489 samples/sec	 accuracy=82.888136	 loss=0.652007	 lr=0.001000
Epoch[072] Batch [1199]/[1203]	Speed: 236.266023 samples/sec	 accuracy=82.886667	 loss=0.652134	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.875000 acc-top5=84.775000
Batch [0039]/[0097]: acc-top1=63.575000 acc-top5=84.787500
Batch [0059]/[0097]: acc-top1=65.450000 acc-top5=85.983333
Batch [0079]/[0097]: acc-top1=65.612500 acc-top5=85.862500
[Epoch 072] training: accuracy=82.895677	 loss=0.651985
[Epoch 072] speed: 125 samples/sec	time cost: 2110.607820
[Epoch 072] validation: acc-top1=65.046392 acc-top5=85.762887 loss=1.563845
Epoch[073] Batch [0019]/[1203]	Speed: 69.154566 samples/sec	 accuracy=83.225000	 loss=0.648164	 lr=0.001000
Epoch[073] Batch [0039]/[1203]	Speed: 99.380623 samples/sec	 accuracy=83.350000	 loss=0.648355	 lr=0.001000
Epoch[073] Batch [0059]/[1203]	Speed: 240.163253 samples/sec	 accuracy=83.016667	 loss=0.647991	 lr=0.001000
Epoch[073] Batch [0079]/[1203]	Speed: 95.207220 samples/sec	 accuracy=82.918750	 loss=0.649726	 lr=0.001000
Epoch[073] Batch [0099]/[1203]	Speed: 110.130012 samples/sec	 accuracy=83.070000	 loss=0.645019	 lr=0.001000
Epoch[073] Batch [0119]/[1203]	Speed: 175.712041 samples/sec	 accuracy=83.195833	 loss=0.643741	 lr=0.001000
Epoch[073] Batch [0139]/[1203]	Speed: 93.938300 samples/sec	 accuracy=83.267857	 loss=0.642763	 lr=0.001000
Epoch[073] Batch [0159]/[1203]	Speed: 270.428196 samples/sec	 accuracy=83.268750	 loss=0.642851	 lr=0.001000
Epoch[073] Batch [0179]/[1203]	Speed: 88.244042 samples/sec	 accuracy=83.283333	 loss=0.641225	 lr=0.001000
Epoch[073] Batch [0199]/[1203]	Speed: 130.543195 samples/sec	 accuracy=83.232500	 loss=0.640511	 lr=0.001000
Epoch[073] Batch [0219]/[1203]	Speed: 159.021206 samples/sec	 accuracy=83.311364	 loss=0.638376	 lr=0.001000
Epoch[073] Batch [0239]/[1203]	Speed: 107.555661 samples/sec	 accuracy=83.270833	 loss=0.640418	 lr=0.001000
Epoch[073] Batch [0259]/[1203]	Speed: 138.278013 samples/sec	 accuracy=83.276923	 loss=0.641057	 lr=0.001000
Epoch[073] Batch [0279]/[1203]	Speed: 113.030657 samples/sec	 accuracy=83.239286	 loss=0.642303	 lr=0.001000
Epoch[073] Batch [0299]/[1203]	Speed: 144.641842 samples/sec	 accuracy=83.276667	 loss=0.640196	 lr=0.001000
Epoch[073] Batch [0319]/[1203]	Speed: 127.040645 samples/sec	 accuracy=83.287500	 loss=0.640969	 lr=0.001000
Epoch[073] Batch [0339]/[1203]	Speed: 111.958181 samples/sec	 accuracy=83.283824	 loss=0.640622	 lr=0.001000
Epoch[073] Batch [0359]/[1203]	Speed: 144.677815 samples/sec	 accuracy=83.275000	 loss=0.640950	 lr=0.001000
Epoch[073] Batch [0379]/[1203]	Speed: 118.566845 samples/sec	 accuracy=83.247368	 loss=0.641315	 lr=0.001000
Epoch[073] Batch [0399]/[1203]	Speed: 132.716028 samples/sec	 accuracy=83.240000	 loss=0.641450	 lr=0.001000
Epoch[073] Batch [0419]/[1203]	Speed: 136.868653 samples/sec	 accuracy=83.245238	 loss=0.640684	 lr=0.001000
Epoch[073] Batch [0439]/[1203]	Speed: 107.851509 samples/sec	 accuracy=83.210227	 loss=0.640994	 lr=0.001000
Epoch[073] Batch [0459]/[1203]	Speed: 148.413803 samples/sec	 accuracy=83.207609	 loss=0.641039	 lr=0.001000
Epoch[073] Batch [0479]/[1203]	Speed: 120.443576 samples/sec	 accuracy=83.200000	 loss=0.641293	 lr=0.001000
Epoch[073] Batch [0499]/[1203]	Speed: 114.911017 samples/sec	 accuracy=83.208000	 loss=0.640886	 lr=0.001000
Epoch[073] Batch [0519]/[1203]	Speed: 164.559882 samples/sec	 accuracy=83.238462	 loss=0.640881	 lr=0.001000
Epoch[073] Batch [0539]/[1203]	Speed: 102.953572 samples/sec	 accuracy=83.262963	 loss=0.640353	 lr=0.001000
Epoch[073] Batch [0559]/[1203]	Speed: 131.111377 samples/sec	 accuracy=83.255357	 loss=0.641029	 lr=0.001000
Epoch[073] Batch [0579]/[1203]	Speed: 133.896539 samples/sec	 accuracy=83.231034	 loss=0.641690	 lr=0.001000
Epoch[073] Batch [0599]/[1203]	Speed: 120.947322 samples/sec	 accuracy=83.237500	 loss=0.641399	 lr=0.001000
Epoch[073] Batch [0619]/[1203]	Speed: 162.249150 samples/sec	 accuracy=83.197581	 loss=0.642502	 lr=0.001000
Epoch[073] Batch [0639]/[1203]	Speed: 99.778855 samples/sec	 accuracy=83.182812	 loss=0.642666	 lr=0.001000
Epoch[073] Batch [0659]/[1203]	Speed: 148.010208 samples/sec	 accuracy=83.196970	 loss=0.642239	 lr=0.001000
Epoch[073] Batch [0679]/[1203]	Speed: 128.181704 samples/sec	 accuracy=83.188971	 loss=0.642523	 lr=0.001000
Epoch[073] Batch [0699]/[1203]	Speed: 111.348045 samples/sec	 accuracy=83.180714	 loss=0.642695	 lr=0.001000
Epoch[073] Batch [0719]/[1203]	Speed: 132.045262 samples/sec	 accuracy=83.159028	 loss=0.642708	 lr=0.001000
Epoch[073] Batch [0739]/[1203]	Speed: 118.304367 samples/sec	 accuracy=83.157432	 loss=0.642647	 lr=0.001000
Epoch[073] Batch [0759]/[1203]	Speed: 140.260316 samples/sec	 accuracy=83.134868	 loss=0.643065	 lr=0.001000
Epoch[073] Batch [0779]/[1203]	Speed: 130.084548 samples/sec	 accuracy=83.097436	 loss=0.643838	 lr=0.001000
Epoch[073] Batch [0799]/[1203]	Speed: 112.016926 samples/sec	 accuracy=83.091250	 loss=0.643671	 lr=0.001000
Epoch[073] Batch [0819]/[1203]	Speed: 135.146053 samples/sec	 accuracy=83.087805	 loss=0.643751	 lr=0.001000
Epoch[073] Batch [0839]/[1203]	Speed: 121.420773 samples/sec	 accuracy=83.097619	 loss=0.643584	 lr=0.001000
Epoch[073] Batch [0859]/[1203]	Speed: 113.936969 samples/sec	 accuracy=83.097093	 loss=0.643591	 lr=0.001000
Epoch[073] Batch [0879]/[1203]	Speed: 143.231537 samples/sec	 accuracy=83.061364	 loss=0.644478	 lr=0.001000
Epoch[073] Batch [0899]/[1203]	Speed: 119.848258 samples/sec	 accuracy=83.069444	 loss=0.644302	 lr=0.001000
Epoch[073] Batch [0919]/[1203]	Speed: 132.964665 samples/sec	 accuracy=83.071196	 loss=0.643988	 lr=0.001000
Epoch[073] Batch [0939]/[1203]	Speed: 130.316312 samples/sec	 accuracy=83.052660	 loss=0.644960	 lr=0.001000
Epoch[073] Batch [0959]/[1203]	Speed: 105.015315 samples/sec	 accuracy=83.055729	 loss=0.645221	 lr=0.001000
Epoch[073] Batch [0979]/[1203]	Speed: 159.778192 samples/sec	 accuracy=83.058163	 loss=0.645246	 lr=0.001000
Epoch[073] Batch [0999]/[1203]	Speed: 121.587070 samples/sec	 accuracy=83.051500	 loss=0.645595	 lr=0.001000
Epoch[073] Batch [1019]/[1203]	Speed: 129.408762 samples/sec	 accuracy=83.033333	 loss=0.646487	 lr=0.001000
Epoch[073] Batch [1039]/[1203]	Speed: 134.526762 samples/sec	 accuracy=83.016827	 loss=0.647077	 lr=0.001000
Epoch[073] Batch [1059]/[1203]	Speed: 110.898240 samples/sec	 accuracy=82.989151	 loss=0.647902	 lr=0.001000
Epoch[073] Batch [1079]/[1203]	Speed: 125.283362 samples/sec	 accuracy=82.978704	 loss=0.648586	 lr=0.001000
Epoch[073] Batch [1099]/[1203]	Speed: 128.232339 samples/sec	 accuracy=82.973182	 loss=0.649040	 lr=0.001000
Epoch[073] Batch [1119]/[1203]	Speed: 110.930190 samples/sec	 accuracy=82.959375	 loss=0.649279	 lr=0.001000
Epoch[073] Batch [1139]/[1203]	Speed: 169.349700 samples/sec	 accuracy=82.939912	 loss=0.649922	 lr=0.001000
Epoch[073] Batch [1159]/[1203]	Speed: 112.549485 samples/sec	 accuracy=82.942241	 loss=0.649878	 lr=0.001000
Epoch[073] Batch [1179]/[1203]	Speed: 138.409432 samples/sec	 accuracy=82.936017	 loss=0.650214	 lr=0.001000
Epoch[073] Batch [1199]/[1203]	Speed: 234.188060 samples/sec	 accuracy=82.926250	 loss=0.650618	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.175000 acc-top5=84.650000
Batch [0039]/[0097]: acc-top1=63.012500 acc-top5=84.675000
Batch [0059]/[0097]: acc-top1=65.225000 acc-top5=85.875000
Batch [0079]/[0097]: acc-top1=65.468750 acc-top5=85.637500
[Epoch 073] training: accuracy=82.916459	 loss=0.650937
[Epoch 073] speed: 125 samples/sec	time cost: 2111.367270
[Epoch 073] validation: acc-top1=64.860825 acc-top5=85.500000 loss=1.570346
Epoch[074] Batch [0019]/[1203]	Speed: 69.202151 samples/sec	 accuracy=82.875000	 loss=0.658554	 lr=0.001000
Epoch[074] Batch [0039]/[1203]	Speed: 105.008287 samples/sec	 accuracy=83.637500	 loss=0.640607	 lr=0.001000
Epoch[074] Batch [0059]/[1203]	Speed: 199.498607 samples/sec	 accuracy=83.575000	 loss=0.630925	 lr=0.001000
Epoch[074] Batch [0079]/[1203]	Speed: 101.476002 samples/sec	 accuracy=83.412500	 loss=0.634402	 lr=0.001000
Epoch[074] Batch [0099]/[1203]	Speed: 125.948901 samples/sec	 accuracy=83.325000	 loss=0.634413	 lr=0.001000
Epoch[074] Batch [0119]/[1203]	Speed: 132.888234 samples/sec	 accuracy=83.220833	 loss=0.635885	 lr=0.001000
Epoch[074] Batch [0139]/[1203]	Speed: 116.523142 samples/sec	 accuracy=83.189286	 loss=0.636790	 lr=0.001000
Epoch[074] Batch [0159]/[1203]	Speed: 181.917108 samples/sec	 accuracy=83.296875	 loss=0.632419	 lr=0.001000
Epoch[074] Batch [0179]/[1203]	Speed: 109.741122 samples/sec	 accuracy=83.222222	 loss=0.636154	 lr=0.001000
Epoch[074] Batch [0199]/[1203]	Speed: 109.562657 samples/sec	 accuracy=83.265000	 loss=0.635421	 lr=0.001000
Epoch[074] Batch [0219]/[1203]	Speed: 148.717002 samples/sec	 accuracy=83.329545	 loss=0.635959	 lr=0.001000
Epoch[074] Batch [0239]/[1203]	Speed: 121.619283 samples/sec	 accuracy=83.287500	 loss=0.637622	 lr=0.001000
Epoch[074] Batch [0259]/[1203]	Speed: 136.187839 samples/sec	 accuracy=83.290385	 loss=0.637814	 lr=0.001000
Epoch[074] Batch [0279]/[1203]	Speed: 128.363415 samples/sec	 accuracy=83.312500	 loss=0.637806	 lr=0.001000
Epoch[074] Batch [0299]/[1203]	Speed: 113.708387 samples/sec	 accuracy=83.396667	 loss=0.635497	 lr=0.001000
Epoch[074] Batch [0319]/[1203]	Speed: 135.041577 samples/sec	 accuracy=83.401562	 loss=0.635399	 lr=0.001000
Epoch[074] Batch [0339]/[1203]	Speed: 121.692905 samples/sec	 accuracy=83.363235	 loss=0.635759	 lr=0.001000
Epoch[074] Batch [0359]/[1203]	Speed: 140.586958 samples/sec	 accuracy=83.375000	 loss=0.634449	 lr=0.001000
Epoch[074] Batch [0379]/[1203]	Speed: 121.113371 samples/sec	 accuracy=83.311842	 loss=0.636027	 lr=0.001000
Epoch[074] Batch [0399]/[1203]	Speed: 123.796858 samples/sec	 accuracy=83.337500	 loss=0.634807	 lr=0.001000
Epoch[074] Batch [0419]/[1203]	Speed: 135.179754 samples/sec	 accuracy=83.290476	 loss=0.635446	 lr=0.001000
Epoch[074] Batch [0439]/[1203]	Speed: 112.314326 samples/sec	 accuracy=83.317045	 loss=0.633599	 lr=0.001000
Epoch[074] Batch [0459]/[1203]	Speed: 141.093743 samples/sec	 accuracy=83.298913	 loss=0.634558	 lr=0.001000
Epoch[074] Batch [0479]/[1203]	Speed: 117.535100 samples/sec	 accuracy=83.310417	 loss=0.634631	 lr=0.001000
Epoch[074] Batch [0499]/[1203]	Speed: 120.134791 samples/sec	 accuracy=83.263000	 loss=0.636169	 lr=0.001000
Epoch[074] Batch [0519]/[1203]	Speed: 156.324331 samples/sec	 accuracy=83.224038	 loss=0.638209	 lr=0.001000
Epoch[074] Batch [0539]/[1203]	Speed: 113.738105 samples/sec	 accuracy=83.204630	 loss=0.639405	 lr=0.001000
Epoch[074] Batch [0559]/[1203]	Speed: 147.719839 samples/sec	 accuracy=83.198214	 loss=0.639940	 lr=0.001000
Epoch[074] Batch [0579]/[1203]	Speed: 109.465705 samples/sec	 accuracy=83.189655	 loss=0.640490	 lr=0.001000
Epoch[074] Batch [0599]/[1203]	Speed: 126.358573 samples/sec	 accuracy=83.162500	 loss=0.641346	 lr=0.001000
Epoch[074] Batch [0619]/[1203]	Speed: 141.971238 samples/sec	 accuracy=83.129032	 loss=0.641762	 lr=0.001000
Epoch[074] Batch [0639]/[1203]	Speed: 104.094709 samples/sec	 accuracy=83.129688	 loss=0.641981	 lr=0.001000
Epoch[074] Batch [0659]/[1203]	Speed: 147.466757 samples/sec	 accuracy=83.128788	 loss=0.641728	 lr=0.001000
Epoch[074] Batch [0679]/[1203]	Speed: 129.943864 samples/sec	 accuracy=83.135294	 loss=0.641434	 lr=0.001000
Epoch[074] Batch [0699]/[1203]	Speed: 118.114531 samples/sec	 accuracy=83.133571	 loss=0.641608	 lr=0.001000
Epoch[074] Batch [0719]/[1203]	Speed: 153.571398 samples/sec	 accuracy=83.127083	 loss=0.642465	 lr=0.001000
Epoch[074] Batch [0739]/[1203]	Speed: 106.807653 samples/sec	 accuracy=83.093919	 loss=0.642911	 lr=0.001000
Epoch[074] Batch [0759]/[1203]	Speed: 120.760933 samples/sec	 accuracy=83.067105	 loss=0.643398	 lr=0.001000
Epoch[074] Batch [0779]/[1203]	Speed: 146.378971 samples/sec	 accuracy=83.062179	 loss=0.643982	 lr=0.001000
Epoch[074] Batch [0799]/[1203]	Speed: 120.157110 samples/sec	 accuracy=83.065000	 loss=0.643949	 lr=0.001000
Epoch[074] Batch [0819]/[1203]	Speed: 152.957039 samples/sec	 accuracy=83.021951	 loss=0.644987	 lr=0.001000
Epoch[074] Batch [0839]/[1203]	Speed: 111.690267 samples/sec	 accuracy=83.027381	 loss=0.644888	 lr=0.001000
Epoch[074] Batch [0859]/[1203]	Speed: 114.549847 samples/sec	 accuracy=83.033721	 loss=0.644577	 lr=0.001000
Epoch[074] Batch [0879]/[1203]	Speed: 147.413594 samples/sec	 accuracy=83.025000	 loss=0.644793	 lr=0.001000
Epoch[074] Batch [0899]/[1203]	Speed: 131.950609 samples/sec	 accuracy=83.022222	 loss=0.645092	 lr=0.001000
Epoch[074] Batch [0919]/[1203]	Speed: 105.212496 samples/sec	 accuracy=83.007609	 loss=0.645327	 lr=0.001000
Epoch[074] Batch [0939]/[1203]	Speed: 167.051307 samples/sec	 accuracy=82.992021	 loss=0.645821	 lr=0.001000
Epoch[074] Batch [0959]/[1203]	Speed: 102.332631 samples/sec	 accuracy=82.976562	 loss=0.646605	 lr=0.001000
Epoch[074] Batch [0979]/[1203]	Speed: 162.947854 samples/sec	 accuracy=82.979082	 loss=0.646792	 lr=0.001000
Epoch[074] Batch [0999]/[1203]	Speed: 121.219416 samples/sec	 accuracy=82.972500	 loss=0.647004	 lr=0.001000
Epoch[074] Batch [1019]/[1203]	Speed: 109.385230 samples/sec	 accuracy=82.970588	 loss=0.647170	 lr=0.001000
Epoch[074] Batch [1039]/[1203]	Speed: 161.292988 samples/sec	 accuracy=82.982212	 loss=0.647255	 lr=0.001000
Epoch[074] Batch [1059]/[1203]	Speed: 109.046726 samples/sec	 accuracy=82.953774	 loss=0.647857	 lr=0.001000
Epoch[074] Batch [1079]/[1203]	Speed: 139.855767 samples/sec	 accuracy=82.954630	 loss=0.647493	 lr=0.001000
Epoch[074] Batch [1099]/[1203]	Speed: 122.413688 samples/sec	 accuracy=82.946818	 loss=0.647595	 lr=0.001000
Epoch[074] Batch [1119]/[1203]	Speed: 127.845616 samples/sec	 accuracy=82.943304	 loss=0.647859	 lr=0.001000
Epoch[074] Batch [1139]/[1203]	Speed: 128.728119 samples/sec	 accuracy=82.948246	 loss=0.648021	 lr=0.001000
Epoch[074] Batch [1159]/[1203]	Speed: 135.004993 samples/sec	 accuracy=82.953017	 loss=0.647964	 lr=0.001000
Epoch[074] Batch [1179]/[1203]	Speed: 128.393618 samples/sec	 accuracy=82.958051	 loss=0.648322	 lr=0.001000
Epoch[074] Batch [1199]/[1203]	Speed: 239.936789 samples/sec	 accuracy=82.966667	 loss=0.648101	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.575000 acc-top5=85.200000
Batch [0039]/[0097]: acc-top1=63.150000 acc-top5=84.925000
Batch [0059]/[0097]: acc-top1=65.175000 acc-top5=86.066667
Batch [0079]/[0097]: acc-top1=65.356250 acc-top5=85.862500
[Epoch 074] training: accuracy=82.960515	 loss=0.648177
[Epoch 074] speed: 126 samples/sec	time cost: 2100.078313
[Epoch 074] validation: acc-top1=64.737113 acc-top5=85.685567 loss=1.573604
Epoch[075] Batch [0019]/[1203]	Speed: 68.867497 samples/sec	 accuracy=83.825000	 loss=0.643238	 lr=0.001000
Epoch[075] Batch [0039]/[1203]	Speed: 104.169856 samples/sec	 accuracy=83.962500	 loss=0.621700	 lr=0.001000
Epoch[075] Batch [0059]/[1203]	Speed: 221.960338 samples/sec	 accuracy=83.725000	 loss=0.624796	 lr=0.001000
Epoch[075] Batch [0079]/[1203]	Speed: 94.943218 samples/sec	 accuracy=83.837500	 loss=0.625084	 lr=0.001000
Epoch[075] Batch [0099]/[1203]	Speed: 112.961858 samples/sec	 accuracy=83.810000	 loss=0.630151	 lr=0.001000
Epoch[075] Batch [0119]/[1203]	Speed: 187.070125 samples/sec	 accuracy=83.887500	 loss=0.626663	 lr=0.001000
Epoch[075] Batch [0139]/[1203]	Speed: 98.250296 samples/sec	 accuracy=83.760714	 loss=0.626132	 lr=0.001000
Epoch[075] Batch [0159]/[1203]	Speed: 243.987321 samples/sec	 accuracy=83.756250	 loss=0.628399	 lr=0.001000
Epoch[075] Batch [0179]/[1203]	Speed: 98.367156 samples/sec	 accuracy=83.700000	 loss=0.630345	 lr=0.001000
Epoch[075] Batch [0199]/[1203]	Speed: 109.075265 samples/sec	 accuracy=83.542500	 loss=0.634746	 lr=0.001000
Epoch[075] Batch [0219]/[1203]	Speed: 161.789064 samples/sec	 accuracy=83.581818	 loss=0.632569	 lr=0.001000
Epoch[075] Batch [0239]/[1203]	Speed: 109.617972 samples/sec	 accuracy=83.620833	 loss=0.631664	 lr=0.001000
Epoch[075] Batch [0259]/[1203]	Speed: 143.360025 samples/sec	 accuracy=83.588462	 loss=0.633278	 lr=0.001000
Epoch[075] Batch [0279]/[1203]	Speed: 130.358688 samples/sec	 accuracy=83.533929	 loss=0.635421	 lr=0.001000
Epoch[075] Batch [0299]/[1203]	Speed: 107.130605 samples/sec	 accuracy=83.486667	 loss=0.636894	 lr=0.001000
Epoch[075] Batch [0319]/[1203]	Speed: 167.447600 samples/sec	 accuracy=83.440625	 loss=0.636854	 lr=0.001000
Epoch[075] Batch [0339]/[1203]	Speed: 109.791964 samples/sec	 accuracy=83.420588	 loss=0.637233	 lr=0.001000
Epoch[075] Batch [0359]/[1203]	Speed: 142.113959 samples/sec	 accuracy=83.388889	 loss=0.637241	 lr=0.001000
Epoch[075] Batch [0379]/[1203]	Speed: 125.307059 samples/sec	 accuracy=83.392105	 loss=0.636742	 lr=0.001000
Epoch[075] Batch [0399]/[1203]	Speed: 113.682727 samples/sec	 accuracy=83.366250	 loss=0.637453	 lr=0.001000
Epoch[075] Batch [0419]/[1203]	Speed: 155.466226 samples/sec	 accuracy=83.347619	 loss=0.638083	 lr=0.001000
Epoch[075] Batch [0439]/[1203]	Speed: 109.184832 samples/sec	 accuracy=83.313636	 loss=0.638998	 lr=0.001000
Epoch[075] Batch [0459]/[1203]	Speed: 140.090308 samples/sec	 accuracy=83.309783	 loss=0.639517	 lr=0.001000
Epoch[075] Batch [0479]/[1203]	Speed: 142.380105 samples/sec	 accuracy=83.296875	 loss=0.639192	 lr=0.001000
Epoch[075] Batch [0499]/[1203]	Speed: 110.034202 samples/sec	 accuracy=83.321000	 loss=0.638206	 lr=0.001000
Epoch[075] Batch [0519]/[1203]	Speed: 154.979583 samples/sec	 accuracy=83.323077	 loss=0.638496	 lr=0.001000
Epoch[075] Batch [0539]/[1203]	Speed: 113.673791 samples/sec	 accuracy=83.292593	 loss=0.639199	 lr=0.001000
Epoch[075] Batch [0559]/[1203]	Speed: 117.092252 samples/sec	 accuracy=83.307143	 loss=0.638717	 lr=0.001000
Epoch[075] Batch [0579]/[1203]	Speed: 163.025758 samples/sec	 accuracy=83.298276	 loss=0.638144	 lr=0.001000
Epoch[075] Batch [0599]/[1203]	Speed: 102.264849 samples/sec	 accuracy=83.298333	 loss=0.637643	 lr=0.001000
Epoch[075] Batch [0619]/[1203]	Speed: 156.763325 samples/sec	 accuracy=83.237903	 loss=0.639507	 lr=0.001000
Epoch[075] Batch [0639]/[1203]	Speed: 111.270415 samples/sec	 accuracy=83.219531	 loss=0.640006	 lr=0.001000
Epoch[075] Batch [0659]/[1203]	Speed: 118.392172 samples/sec	 accuracy=83.195455	 loss=0.640602	 lr=0.001000
Epoch[075] Batch [0679]/[1203]	Speed: 162.284961 samples/sec	 accuracy=83.185294	 loss=0.640520	 lr=0.001000
Epoch[075] Batch [0699]/[1203]	Speed: 104.343309 samples/sec	 accuracy=83.193571	 loss=0.640081	 lr=0.001000
Epoch[075] Batch [0719]/[1203]	Speed: 164.522113 samples/sec	 accuracy=83.226389	 loss=0.639271	 lr=0.001000
Epoch[075] Batch [0739]/[1203]	Speed: 114.607170 samples/sec	 accuracy=83.212162	 loss=0.639788	 lr=0.001000
Epoch[075] Batch [0759]/[1203]	Speed: 123.699731 samples/sec	 accuracy=83.225000	 loss=0.639722	 lr=0.001000
Epoch[075] Batch [0779]/[1203]	Speed: 141.939659 samples/sec	 accuracy=83.226923	 loss=0.639542	 lr=0.001000
Epoch[075] Batch [0799]/[1203]	Speed: 110.416198 samples/sec	 accuracy=83.198750	 loss=0.640217	 lr=0.001000
Epoch[075] Batch [0819]/[1203]	Speed: 140.362926 samples/sec	 accuracy=83.197561	 loss=0.640818	 lr=0.001000
Epoch[075] Batch [0839]/[1203]	Speed: 116.785642 samples/sec	 accuracy=83.208333	 loss=0.640516	 lr=0.001000
Epoch[075] Batch [0859]/[1203]	Speed: 124.225113 samples/sec	 accuracy=83.185465	 loss=0.640898	 lr=0.001000
Epoch[075] Batch [0879]/[1203]	Speed: 146.024142 samples/sec	 accuracy=83.186364	 loss=0.640873	 lr=0.001000
Epoch[075] Batch [0899]/[1203]	Speed: 130.716124 samples/sec	 accuracy=83.181111	 loss=0.641037	 lr=0.001000
Epoch[075] Batch [0919]/[1203]	Speed: 103.768803 samples/sec	 accuracy=83.155435	 loss=0.641747	 lr=0.001000
Epoch[075] Batch [0939]/[1203]	Speed: 130.260146 samples/sec	 accuracy=83.137234	 loss=0.642550	 lr=0.001000
Epoch[075] Batch [0959]/[1203]	Speed: 133.612996 samples/sec	 accuracy=83.144792	 loss=0.642354	 lr=0.001000
Epoch[075] Batch [0979]/[1203]	Speed: 142.061904 samples/sec	 accuracy=83.157143	 loss=0.642095	 lr=0.001000
Epoch[075] Batch [0999]/[1203]	Speed: 114.986672 samples/sec	 accuracy=83.152500	 loss=0.642515	 lr=0.001000
Epoch[075] Batch [1019]/[1203]	Speed: 114.798379 samples/sec	 accuracy=83.148529	 loss=0.642789	 lr=0.001000
Epoch[075] Batch [1039]/[1203]	Speed: 135.459468 samples/sec	 accuracy=83.154327	 loss=0.642677	 lr=0.001000
Epoch[075] Batch [1059]/[1203]	Speed: 135.510073 samples/sec	 accuracy=83.155660	 loss=0.642509	 lr=0.001000
Epoch[075] Batch [1079]/[1203]	Speed: 124.207670 samples/sec	 accuracy=83.143519	 loss=0.643009	 lr=0.001000
Epoch[075] Batch [1099]/[1203]	Speed: 116.519230 samples/sec	 accuracy=83.132273	 loss=0.643361	 lr=0.001000
Epoch[075] Batch [1119]/[1203]	Speed: 126.634740 samples/sec	 accuracy=83.145982	 loss=0.643141	 lr=0.001000
Epoch[075] Batch [1139]/[1203]	Speed: 143.174493 samples/sec	 accuracy=83.157895	 loss=0.642758	 lr=0.001000
Epoch[075] Batch [1159]/[1203]	Speed: 128.702371 samples/sec	 accuracy=83.142672	 loss=0.643320	 lr=0.001000
Epoch[075] Batch [1179]/[1203]	Speed: 111.695564 samples/sec	 accuracy=83.150847	 loss=0.642679	 lr=0.001000
Epoch[075] Batch [1199]/[1203]	Speed: 309.887523 samples/sec	 accuracy=83.142083	 loss=0.642931	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.400000 acc-top5=84.900000
Batch [0039]/[0097]: acc-top1=63.112500 acc-top5=84.587500
Batch [0059]/[0097]: acc-top1=65.225000 acc-top5=85.725000
Batch [0079]/[0097]: acc-top1=65.381250 acc-top5=85.656250
[Epoch 075] training: accuracy=83.143392	 loss=0.642823
[Epoch 075] speed: 126 samples/sec	time cost: 2098.228396
[Epoch 075] validation: acc-top1=64.835052 acc-top5=85.520619 loss=1.580368
Epoch[076] Batch [0019]/[1203]	Speed: 69.401361 samples/sec	 accuracy=83.275000	 loss=0.637133	 lr=0.001000
Epoch[076] Batch [0039]/[1203]	Speed: 101.475229 samples/sec	 accuracy=83.675000	 loss=0.625061	 lr=0.001000
Epoch[076] Batch [0059]/[1203]	Speed: 218.822828 samples/sec	 accuracy=83.600000	 loss=0.628223	 lr=0.001000
Epoch[076] Batch [0079]/[1203]	Speed: 101.659307 samples/sec	 accuracy=83.493750	 loss=0.628522	 lr=0.001000
Epoch[076] Batch [0099]/[1203]	Speed: 101.016752 samples/sec	 accuracy=83.450000	 loss=0.634324	 lr=0.001000
Epoch[076] Batch [0119]/[1203]	Speed: 207.964462 samples/sec	 accuracy=83.295833	 loss=0.637435	 lr=0.001000
Epoch[076] Batch [0139]/[1203]	Speed: 100.070324 samples/sec	 accuracy=83.225000	 loss=0.642096	 lr=0.001000
Epoch[076] Batch [0159]/[1203]	Speed: 241.386594 samples/sec	 accuracy=83.293750	 loss=0.638959	 lr=0.001000
Epoch[076] Batch [0179]/[1203]	Speed: 95.424937 samples/sec	 accuracy=83.302778	 loss=0.639480	 lr=0.001000
Epoch[076] Batch [0199]/[1203]	Speed: 112.946381 samples/sec	 accuracy=83.305000	 loss=0.638588	 lr=0.001000
Epoch[076] Batch [0219]/[1203]	Speed: 179.778186 samples/sec	 accuracy=83.354545	 loss=0.635721	 lr=0.001000
Epoch[076] Batch [0239]/[1203]	Speed: 103.580380 samples/sec	 accuracy=83.304167	 loss=0.637232	 lr=0.001000
Epoch[076] Batch [0259]/[1203]	Speed: 132.596906 samples/sec	 accuracy=83.409615	 loss=0.635530	 lr=0.001000
Epoch[076] Batch [0279]/[1203]	Speed: 129.683280 samples/sec	 accuracy=83.448214	 loss=0.634623	 lr=0.001000
Epoch[076] Batch [0299]/[1203]	Speed: 124.719811 samples/sec	 accuracy=83.381667	 loss=0.636537	 lr=0.001000
Epoch[076] Batch [0319]/[1203]	Speed: 151.214855 samples/sec	 accuracy=83.384375	 loss=0.635670	 lr=0.001000
Epoch[076] Batch [0339]/[1203]	Speed: 108.715821 samples/sec	 accuracy=83.414706	 loss=0.636495	 lr=0.001000
Epoch[076] Batch [0359]/[1203]	Speed: 117.895824 samples/sec	 accuracy=83.390278	 loss=0.637079	 lr=0.001000
Epoch[076] Batch [0379]/[1203]	Speed: 138.615446 samples/sec	 accuracy=83.398684	 loss=0.636941	 lr=0.001000
Epoch[076] Batch [0399]/[1203]	Speed: 131.776086 samples/sec	 accuracy=83.423750	 loss=0.635783	 lr=0.001000
Epoch[076] Batch [0419]/[1203]	Speed: 146.532222 samples/sec	 accuracy=83.386905	 loss=0.636361	 lr=0.001000
Epoch[076] Batch [0439]/[1203]	Speed: 110.758558 samples/sec	 accuracy=83.361364	 loss=0.636296	 lr=0.001000
Epoch[076] Batch [0459]/[1203]	Speed: 126.047448 samples/sec	 accuracy=83.358696	 loss=0.637037	 lr=0.001000
Epoch[076] Batch [0479]/[1203]	Speed: 125.782995 samples/sec	 accuracy=83.333333	 loss=0.638073	 lr=0.001000
Epoch[076] Batch [0499]/[1203]	Speed: 122.295086 samples/sec	 accuracy=83.339000	 loss=0.637207	 lr=0.001000
Epoch[076] Batch [0519]/[1203]	Speed: 148.447695 samples/sec	 accuracy=83.353846	 loss=0.636976	 lr=0.001000
Epoch[076] Batch [0539]/[1203]	Speed: 111.690663 samples/sec	 accuracy=83.353704	 loss=0.637143	 lr=0.001000
Epoch[076] Batch [0559]/[1203]	Speed: 141.431994 samples/sec	 accuracy=83.348214	 loss=0.636933	 lr=0.001000
Epoch[076] Batch [0579]/[1203]	Speed: 122.377133 samples/sec	 accuracy=83.331034	 loss=0.637287	 lr=0.001000
Epoch[076] Batch [0599]/[1203]	Speed: 106.627114 samples/sec	 accuracy=83.334167	 loss=0.636722	 lr=0.001000
Epoch[076] Batch [0619]/[1203]	Speed: 157.474723 samples/sec	 accuracy=83.318548	 loss=0.637126	 lr=0.001000
Epoch[076] Batch [0639]/[1203]	Speed: 113.343301 samples/sec	 accuracy=83.300781	 loss=0.637897	 lr=0.001000
Epoch[076] Batch [0659]/[1203]	Speed: 169.337781 samples/sec	 accuracy=83.296970	 loss=0.637569	 lr=0.001000
Epoch[076] Batch [0679]/[1203]	Speed: 120.370371 samples/sec	 accuracy=83.278676	 loss=0.637568	 lr=0.001000
Epoch[076] Batch [0699]/[1203]	Speed: 112.330260 samples/sec	 accuracy=83.255000	 loss=0.638443	 lr=0.001000
Epoch[076] Batch [0719]/[1203]	Speed: 152.123272 samples/sec	 accuracy=83.254861	 loss=0.638080	 lr=0.001000
Epoch[076] Batch [0739]/[1203]	Speed: 110.031569 samples/sec	 accuracy=83.250676	 loss=0.637948	 lr=0.001000
Epoch[076] Batch [0759]/[1203]	Speed: 133.816509 samples/sec	 accuracy=83.241447	 loss=0.638352	 lr=0.001000
Epoch[076] Batch [0779]/[1203]	Speed: 124.494972 samples/sec	 accuracy=83.254487	 loss=0.638007	 lr=0.001000
Epoch[076] Batch [0799]/[1203]	Speed: 131.020280 samples/sec	 accuracy=83.231250	 loss=0.638254	 lr=0.001000
Epoch[076] Batch [0819]/[1203]	Speed: 143.987296 samples/sec	 accuracy=83.240854	 loss=0.638292	 lr=0.001000
Epoch[076] Batch [0839]/[1203]	Speed: 113.969789 samples/sec	 accuracy=83.240476	 loss=0.638464	 lr=0.001000
Epoch[076] Batch [0859]/[1203]	Speed: 120.140844 samples/sec	 accuracy=83.233140	 loss=0.638630	 lr=0.001000
Epoch[076] Batch [0879]/[1203]	Speed: 137.050691 samples/sec	 accuracy=83.241477	 loss=0.638577	 lr=0.001000
Epoch[076] Batch [0899]/[1203]	Speed: 126.983916 samples/sec	 accuracy=83.235556	 loss=0.638506	 lr=0.001000
Epoch[076] Batch [0919]/[1203]	Speed: 117.177643 samples/sec	 accuracy=83.233696	 loss=0.638532	 lr=0.001000
Epoch[076] Batch [0939]/[1203]	Speed: 119.425813 samples/sec	 accuracy=83.226064	 loss=0.639289	 lr=0.001000
Epoch[076] Batch [0959]/[1203]	Speed: 161.970497 samples/sec	 accuracy=83.217187	 loss=0.639352	 lr=0.001000
Epoch[076] Batch [0979]/[1203]	Speed: 122.968024 samples/sec	 accuracy=83.210714	 loss=0.639299	 lr=0.001000
Epoch[076] Batch [0999]/[1203]	Speed: 128.166223 samples/sec	 accuracy=83.186000	 loss=0.639624	 lr=0.001000
Epoch[076] Batch [1019]/[1203]	Speed: 119.884996 samples/sec	 accuracy=83.168627	 loss=0.640268	 lr=0.001000
Epoch[076] Batch [1039]/[1203]	Speed: 114.242305 samples/sec	 accuracy=83.153846	 loss=0.640543	 lr=0.001000
Epoch[076] Batch [1059]/[1203]	Speed: 168.737456 samples/sec	 accuracy=83.160849	 loss=0.640572	 lr=0.001000
Epoch[076] Batch [1079]/[1203]	Speed: 115.840806 samples/sec	 accuracy=83.166204	 loss=0.640960	 lr=0.001000
Epoch[076] Batch [1099]/[1203]	Speed: 109.209756 samples/sec	 accuracy=83.158182	 loss=0.640898	 lr=0.001000
Epoch[076] Batch [1119]/[1203]	Speed: 156.880462 samples/sec	 accuracy=83.147321	 loss=0.640906	 lr=0.001000
Epoch[076] Batch [1139]/[1203]	Speed: 109.867092 samples/sec	 accuracy=83.148684	 loss=0.640692	 lr=0.001000
Epoch[076] Batch [1159]/[1203]	Speed: 166.841445 samples/sec	 accuracy=83.131897	 loss=0.641103	 lr=0.001000
Epoch[076] Batch [1179]/[1203]	Speed: 116.854931 samples/sec	 accuracy=83.130508	 loss=0.640948	 lr=0.001000
Epoch[076] Batch [1199]/[1203]	Speed: 229.980312 samples/sec	 accuracy=83.111250	 loss=0.641521	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.425000 acc-top5=84.525000
Batch [0039]/[0097]: acc-top1=63.112500 acc-top5=84.512500
Batch [0059]/[0097]: acc-top1=65.108333 acc-top5=85.708333
Batch [0079]/[0097]: acc-top1=65.300000 acc-top5=85.606250
[Epoch 076] training: accuracy=83.115960	 loss=0.641382
[Epoch 076] speed: 126 samples/sec	time cost: 2095.454993
[Epoch 076] validation: acc-top1=64.845361 acc-top5=85.494845 loss=1.570244
Epoch[077] Batch [0019]/[1204]	Speed: 70.540241 samples/sec	 accuracy=83.375000	 loss=0.647728	 lr=0.001000
Epoch[077] Batch [0039]/[1204]	Speed: 99.000208 samples/sec	 accuracy=83.187500	 loss=0.648128	 lr=0.001000
Epoch[077] Batch [0059]/[1204]	Speed: 264.946145 samples/sec	 accuracy=83.275000	 loss=0.635083	 lr=0.001000
Epoch[077] Batch [0079]/[1204]	Speed: 93.298124 samples/sec	 accuracy=83.412500	 loss=0.629307	 lr=0.001000
Epoch[077] Batch [0099]/[1204]	Speed: 111.311627 samples/sec	 accuracy=83.520000	 loss=0.626156	 lr=0.001000
Epoch[077] Batch [0119]/[1204]	Speed: 179.112939 samples/sec	 accuracy=83.454167	 loss=0.625323	 lr=0.001000
Epoch[077] Batch [0139]/[1204]	Speed: 106.853857 samples/sec	 accuracy=83.578571	 loss=0.621607	 lr=0.001000
Epoch[077] Batch [0159]/[1204]	Speed: 221.359253 samples/sec	 accuracy=83.450000	 loss=0.628022	 lr=0.001000
Epoch[077] Batch [0179]/[1204]	Speed: 99.027722 samples/sec	 accuracy=83.472222	 loss=0.629183	 lr=0.001000
Epoch[077] Batch [0199]/[1204]	Speed: 111.217783 samples/sec	 accuracy=83.410000	 loss=0.630766	 lr=0.001000
Epoch[077] Batch [0219]/[1204]	Speed: 174.298675 samples/sec	 accuracy=83.477273	 loss=0.630758	 lr=0.001000
Epoch[077] Batch [0239]/[1204]	Speed: 101.748088 samples/sec	 accuracy=83.489583	 loss=0.629992	 lr=0.001000
Epoch[077] Batch [0259]/[1204]	Speed: 151.832353 samples/sec	 accuracy=83.461538	 loss=0.630110	 lr=0.001000
Epoch[077] Batch [0279]/[1204]	Speed: 129.183365 samples/sec	 accuracy=83.432143	 loss=0.631094	 lr=0.001000
Epoch[077] Batch [0299]/[1204]	Speed: 108.010091 samples/sec	 accuracy=83.428333	 loss=0.630223	 lr=0.001000
Epoch[077] Batch [0319]/[1204]	Speed: 153.928483 samples/sec	 accuracy=83.489063	 loss=0.629011	 lr=0.001000
Epoch[077] Batch [0339]/[1204]	Speed: 120.407265 samples/sec	 accuracy=83.441176	 loss=0.630269	 lr=0.001000
Epoch[077] Batch [0359]/[1204]	Speed: 110.724291 samples/sec	 accuracy=83.405556	 loss=0.631089	 lr=0.001000
Epoch[077] Batch [0379]/[1204]	Speed: 151.610283 samples/sec	 accuracy=83.375000	 loss=0.632238	 lr=0.001000
Epoch[077] Batch [0399]/[1204]	Speed: 115.089354 samples/sec	 accuracy=83.375000	 loss=0.632857	 lr=0.001000
Epoch[077] Batch [0419]/[1204]	Speed: 145.284260 samples/sec	 accuracy=83.366667	 loss=0.632989	 lr=0.001000
Epoch[077] Batch [0439]/[1204]	Speed: 103.410627 samples/sec	 accuracy=83.320455	 loss=0.633671	 lr=0.001000
Epoch[077] Batch [0459]/[1204]	Speed: 139.891697 samples/sec	 accuracy=83.329348	 loss=0.633802	 lr=0.001000
Epoch[077] Batch [0479]/[1204]	Speed: 126.995872 samples/sec	 accuracy=83.339583	 loss=0.634283	 lr=0.001000
Epoch[077] Batch [0499]/[1204]	Speed: 133.519675 samples/sec	 accuracy=83.323000	 loss=0.635405	 lr=0.001000
Epoch[077] Batch [0519]/[1204]	Speed: 124.263233 samples/sec	 accuracy=83.293269	 loss=0.636047	 lr=0.001000
Epoch[077] Batch [0539]/[1204]	Speed: 114.830474 samples/sec	 accuracy=83.284259	 loss=0.636015	 lr=0.001000
Epoch[077] Batch [0559]/[1204]	Speed: 134.299112 samples/sec	 accuracy=83.283036	 loss=0.636511	 lr=0.001000
Epoch[077] Batch [0579]/[1204]	Speed: 140.754890 samples/sec	 accuracy=83.306897	 loss=0.635652	 lr=0.001000
Epoch[077] Batch [0599]/[1204]	Speed: 126.014531 samples/sec	 accuracy=83.280833	 loss=0.636948	 lr=0.001000
Epoch[077] Batch [0619]/[1204]	Speed: 131.653737 samples/sec	 accuracy=83.262097	 loss=0.636832	 lr=0.001000
Epoch[077] Batch [0639]/[1204]	Speed: 108.250972 samples/sec	 accuracy=83.279687	 loss=0.636334	 lr=0.001000
Epoch[077] Batch [0659]/[1204]	Speed: 137.666091 samples/sec	 accuracy=83.271212	 loss=0.636822	 lr=0.001000
Epoch[077] Batch [0679]/[1204]	Speed: 130.416086 samples/sec	 accuracy=83.269853	 loss=0.636801	 lr=0.001000
Epoch[077] Batch [0699]/[1204]	Speed: 130.399150 samples/sec	 accuracy=83.272143	 loss=0.636845	 lr=0.001000
Epoch[077] Batch [0719]/[1204]	Speed: 118.064002 samples/sec	 accuracy=83.230556	 loss=0.637842	 lr=0.001000
Epoch[077] Batch [0739]/[1204]	Speed: 130.871918 samples/sec	 accuracy=83.212838	 loss=0.638904	 lr=0.001000
Epoch[077] Batch [0759]/[1204]	Speed: 125.111552 samples/sec	 accuracy=83.219079	 loss=0.638604	 lr=0.001000
Epoch[077] Batch [0779]/[1204]	Speed: 119.627168 samples/sec	 accuracy=83.211538	 loss=0.638736	 lr=0.001000
Epoch[077] Batch [0799]/[1204]	Speed: 136.573651 samples/sec	 accuracy=83.243750	 loss=0.637968	 lr=0.001000
Epoch[077] Batch [0819]/[1204]	Speed: 121.743091 samples/sec	 accuracy=83.224390	 loss=0.638610	 lr=0.001000
Epoch[077] Batch [0839]/[1204]	Speed: 137.006921 samples/sec	 accuracy=83.232143	 loss=0.638577	 lr=0.001000
Epoch[077] Batch [0859]/[1204]	Speed: 125.795605 samples/sec	 accuracy=83.260465	 loss=0.637757	 lr=0.001000
Epoch[077] Batch [0879]/[1204]	Speed: 119.123287 samples/sec	 accuracy=83.254545	 loss=0.638082	 lr=0.001000
Epoch[077] Batch [0899]/[1204]	Speed: 126.691643 samples/sec	 accuracy=83.250556	 loss=0.638319	 lr=0.001000
Epoch[077] Batch [0919]/[1204]	Speed: 133.732285 samples/sec	 accuracy=83.228261	 loss=0.638757	 lr=0.001000
Epoch[077] Batch [0939]/[1204]	Speed: 125.550890 samples/sec	 accuracy=83.214362	 loss=0.639386	 lr=0.001000
Epoch[077] Batch [0959]/[1204]	Speed: 112.470025 samples/sec	 accuracy=83.205208	 loss=0.639609	 lr=0.001000
Epoch[077] Batch [0979]/[1204]	Speed: 155.468250 samples/sec	 accuracy=83.198980	 loss=0.639674	 lr=0.001000
Epoch[077] Batch [0999]/[1204]	Speed: 116.352866 samples/sec	 accuracy=83.192500	 loss=0.639775	 lr=0.001000
Epoch[077] Batch [1019]/[1204]	Speed: 129.763955 samples/sec	 accuracy=83.182353	 loss=0.640232	 lr=0.001000
Epoch[077] Batch [1039]/[1204]	Speed: 140.701520 samples/sec	 accuracy=83.173077	 loss=0.640442	 lr=0.001000
Epoch[077] Batch [1059]/[1204]	Speed: 103.719649 samples/sec	 accuracy=83.178774	 loss=0.640247	 lr=0.001000
Epoch[077] Batch [1079]/[1204]	Speed: 159.584855 samples/sec	 accuracy=83.186574	 loss=0.639838	 lr=0.001000
Epoch[077] Batch [1099]/[1204]	Speed: 110.657368 samples/sec	 accuracy=83.188636	 loss=0.639598	 lr=0.001000
Epoch[077] Batch [1119]/[1204]	Speed: 131.655087 samples/sec	 accuracy=83.207589	 loss=0.639097	 lr=0.001000
Epoch[077] Batch [1139]/[1204]	Speed: 140.462673 samples/sec	 accuracy=83.203947	 loss=0.639249	 lr=0.001000
Epoch[077] Batch [1159]/[1204]	Speed: 111.043235 samples/sec	 accuracy=83.217672	 loss=0.638830	 lr=0.001000
Epoch[077] Batch [1179]/[1204]	Speed: 138.423996 samples/sec	 accuracy=83.210169	 loss=0.639122	 lr=0.001000
Epoch[077] Batch [1199]/[1204]	Speed: 242.762470 samples/sec	 accuracy=83.213750	 loss=0.639117	 lr=0.001000
Batch [0019]/[0097]: acc-top1=63.750000 acc-top5=85.050000
Batch [0039]/[0097]: acc-top1=63.000000 acc-top5=84.662500
Batch [0059]/[0097]: acc-top1=65.000000 acc-top5=85.950000
Batch [0079]/[0097]: acc-top1=65.068750 acc-top5=85.781250
[Epoch 077] training: accuracy=83.212625	 loss=0.639079
[Epoch 077] speed: 126 samples/sec	time cost: 2098.975682
[Epoch 077] validation: acc-top1=64.582474 acc-top5=85.644330 loss=1.581318
Epoch[078] Batch [0019]/[1203]	Speed: 69.164642 samples/sec	 accuracy=83.900000	 loss=0.608642	 lr=0.001000
Epoch[078] Batch [0039]/[1203]	Speed: 102.621167 samples/sec	 accuracy=83.387500	 loss=0.609939	 lr=0.001000
Epoch[078] Batch [0059]/[1203]	Speed: 233.794053 samples/sec	 accuracy=83.241667	 loss=0.613594	 lr=0.001000
Epoch[078] Batch [0079]/[1203]	Speed: 93.534001 samples/sec	 accuracy=83.262500	 loss=0.618107	 lr=0.001000
Epoch[078] Batch [0099]/[1203]	Speed: 108.752717 samples/sec	 accuracy=83.285000	 loss=0.617349	 lr=0.001000
Epoch[078] Batch [0119]/[1203]	Speed: 207.184808 samples/sec	 accuracy=83.395833	 loss=0.620998	 lr=0.001000
Epoch[078] Batch [0139]/[1203]	Speed: 107.050744 samples/sec	 accuracy=83.389286	 loss=0.624172	 lr=0.001000
Epoch[078] Batch [0159]/[1203]	Speed: 195.629075 samples/sec	 accuracy=83.396875	 loss=0.626138	 lr=0.001000
Epoch[078] Batch [0179]/[1203]	Speed: 94.213321 samples/sec	 accuracy=83.427778	 loss=0.625658	 lr=0.001000
Epoch[078] Batch [0199]/[1203]	Speed: 108.079964 samples/sec	 accuracy=83.605000	 loss=0.621637	 lr=0.001000
Epoch[078] Batch [0219]/[1203]	Speed: 194.173906 samples/sec	 accuracy=83.584091	 loss=0.622766	 lr=0.001000
Epoch[078] Batch [0239]/[1203]	Speed: 103.915214 samples/sec	 accuracy=83.560417	 loss=0.623379	 lr=0.001000
Epoch[078] Batch [0259]/[1203]	Speed: 115.321681 samples/sec	 accuracy=83.517308	 loss=0.623748	 lr=0.001000
Epoch[078] Batch [0279]/[1203]	Speed: 165.124724 samples/sec	 accuracy=83.471429	 loss=0.625818	 lr=0.001000
Epoch[078] Batch [0299]/[1203]	Speed: 106.425783 samples/sec	 accuracy=83.411667	 loss=0.626898	 lr=0.001000
Epoch[078] Batch [0319]/[1203]	Speed: 193.098811 samples/sec	 accuracy=83.421875	 loss=0.625867	 lr=0.001000
Epoch[078] Batch [0339]/[1203]	Speed: 101.239338 samples/sec	 accuracy=83.372059	 loss=0.627141	 lr=0.001000
Epoch[078] Batch [0359]/[1203]	Speed: 110.414502 samples/sec	 accuracy=83.416667	 loss=0.626725	 lr=0.001000
Epoch[078] Batch [0379]/[1203]	Speed: 179.512871 samples/sec	 accuracy=83.375000	 loss=0.628243	 lr=0.001000
Epoch[078] Batch [0399]/[1203]	Speed: 101.536290 samples/sec	 accuracy=83.398750	 loss=0.628113	 lr=0.001000
Epoch[078] Batch [0419]/[1203]	Speed: 138.805853 samples/sec	 accuracy=83.389286	 loss=0.630132	 lr=0.001000
Epoch[078] Batch [0439]/[1203]	Speed: 127.406560 samples/sec	 accuracy=83.386364	 loss=0.630506	 lr=0.001000
Epoch[078] Batch [0459]/[1203]	Speed: 123.650924 samples/sec	 accuracy=83.417391	 loss=0.629425	 lr=0.001000
Epoch[078] Batch [0479]/[1203]	Speed: 156.432100 samples/sec	 accuracy=83.416667	 loss=0.629777	 lr=0.001000
Epoch[078] Batch [0499]/[1203]	Speed: 106.292640 samples/sec	 accuracy=83.430000	 loss=0.629371	 lr=0.001000
Epoch[078] Batch [0519]/[1203]	Speed: 131.141460 samples/sec	 accuracy=83.430769	 loss=0.628966	 lr=0.001000
Epoch[078] Batch [0539]/[1203]	Speed: 133.722633 samples/sec	 accuracy=83.426852	 loss=0.628491	 lr=0.001000
Epoch[078] Batch [0559]/[1203]	Speed: 117.747541 samples/sec	 accuracy=83.424107	 loss=0.629057	 lr=0.001000
Epoch[078] Batch [0579]/[1203]	Speed: 162.141727 samples/sec	 accuracy=83.390517	 loss=0.630002	 lr=0.001000
Epoch[078] Batch [0599]/[1203]	Speed: 103.071312 samples/sec	 accuracy=83.399167	 loss=0.629733	 lr=0.001000
Epoch[078] Batch [0619]/[1203]	Speed: 141.856776 samples/sec	 accuracy=83.391935	 loss=0.630314	 lr=0.001000
Epoch[078] Batch [0639]/[1203]	Speed: 127.498857 samples/sec	 accuracy=83.389062	 loss=0.630674	 lr=0.001000
Epoch[078] Batch [0659]/[1203]	Speed: 106.112129 samples/sec	 accuracy=83.384091	 loss=0.630537	 lr=0.001000
Epoch[078] Batch [0679]/[1203]	Speed: 170.736390 samples/sec	 accuracy=83.353676	 loss=0.632106	 lr=0.001000
Epoch[078] Batch [0699]/[1203]	Speed: 104.150362 samples/sec	 accuracy=83.343571	 loss=0.632785	 lr=0.001000
Epoch[078] Batch [0719]/[1203]	Speed: 159.355940 samples/sec	 accuracy=83.350000	 loss=0.632734	 lr=0.001000
Epoch[078] Batch [0739]/[1203]	Speed: 117.314261 samples/sec	 accuracy=83.364189	 loss=0.632349	 lr=0.001000
Epoch[078] Batch [0759]/[1203]	Speed: 107.329282 samples/sec	 accuracy=83.384211	 loss=0.631918	 lr=0.001000
Epoch[078] Batch [0779]/[1203]	Speed: 177.182728 samples/sec	 accuracy=83.362821	 loss=0.633013	 lr=0.001000
Epoch[078] Batch [0799]/[1203]	Speed: 115.908314 samples/sec	 accuracy=83.321875	 loss=0.634377	 lr=0.001000
Epoch[078] Batch [0819]/[1203]	Speed: 132.167730 samples/sec	 accuracy=83.331098	 loss=0.633805	 lr=0.001000
Epoch[078] Batch [0839]/[1203]	Speed: 126.689319 samples/sec	 accuracy=83.327381	 loss=0.633478	 lr=0.001000
Epoch[078] Batch [0859]/[1203]	Speed: 107.022158 samples/sec	 accuracy=83.325581	 loss=0.633425	 lr=0.001000
Epoch[078] Batch [0879]/[1203]	Speed: 153.157699 samples/sec	 accuracy=83.309659	 loss=0.633889	 lr=0.001000
Epoch[078] Batch [0899]/[1203]	Speed: 122.871631 samples/sec	 accuracy=83.310000	 loss=0.633696	 lr=0.001000
Epoch[078] Batch [0919]/[1203]	Speed: 98.684786 samples/sec	 accuracy=83.295652	 loss=0.634115	 lr=0.001000
Epoch[078] Batch [0939]/[1203]	Speed: 172.209241 samples/sec	 accuracy=83.285638	 loss=0.634337	 lr=0.001000
Epoch[078] Batch [0959]/[1203]	Speed: 119.222040 samples/sec	 accuracy=83.300521	 loss=0.633855	 lr=0.001000
Epoch[078] Batch [0979]/[1203]	Speed: 161.745307 samples/sec	 accuracy=83.300510	 loss=0.634205	 lr=0.001000
Epoch[078] Batch [0999]/[1203]	Speed: 106.993180 samples/sec	 accuracy=83.285500	 loss=0.634620	 lr=0.001000
Epoch[078] Batch [1019]/[1203]	Speed: 117.976627 samples/sec	 accuracy=83.287745	 loss=0.634511	 lr=0.001000
Epoch[078] Batch [1039]/[1203]	Speed: 150.493677 samples/sec	 accuracy=83.292788	 loss=0.634489	 lr=0.001000
Epoch[078] Batch [1059]/[1203]	Speed: 114.390153 samples/sec	 accuracy=83.308019	 loss=0.634249	 lr=0.001000
Epoch[078] Batch [1079]/[1203]	Speed: 120.786035 samples/sec	 accuracy=83.291204	 loss=0.634693	 lr=0.001000
Epoch[078] Batch [1099]/[1203]	Speed: 147.037930 samples/sec	 accuracy=83.291364	 loss=0.634992	 lr=0.001000
Epoch[078] Batch [1119]/[1203]	Speed: 122.275034 samples/sec	 accuracy=83.308036	 loss=0.634592	 lr=0.001000
Epoch[078] Batch [1139]/[1203]	Speed: 139.313747 samples/sec	 accuracy=83.289912	 loss=0.634885	 lr=0.001000
Epoch[078] Batch [1159]/[1203]	Speed: 104.074937 samples/sec	 accuracy=83.305172	 loss=0.634920	 lr=0.001000
Epoch[078] Batch [1179]/[1203]	Speed: 142.030009 samples/sec	 accuracy=83.290678	 loss=0.635426	 lr=0.001000
Epoch[078] Batch [1199]/[1203]	Speed: 270.153155 samples/sec	 accuracy=83.296250	 loss=0.635556	 lr=0.001000
Batch [0019]/[0097]: acc-top1=64.200000 acc-top5=84.825000
Batch [0039]/[0097]: acc-top1=62.975000 acc-top5=84.387500
Batch [0059]/[0097]: acc-top1=64.941667 acc-top5=85.733333
Batch [0079]/[0097]: acc-top1=65.068750 acc-top5=85.693750
[Epoch 078] training: accuracy=83.295927	 loss=0.635530
[Epoch 078] speed: 126 samples/sec	time cost: 2100.265148
[Epoch 078] validation: acc-top1=64.474227 acc-top5=85.541237 loss=1.587488
Epoch[079] Batch [0019]/[1203]	Speed: 69.831327 samples/sec	 accuracy=82.925000	 loss=0.631861	 lr=0.001000
Epoch[079] Batch [0039]/[1203]	Speed: 100.142211 samples/sec	 accuracy=83.837500	 loss=0.614954	 lr=0.001000
Epoch[079] Batch [0059]/[1203]	Speed: 249.349459 samples/sec	 accuracy=83.600000	 loss=0.628345	 lr=0.001000
Epoch[079] Batch [0079]/[1203]	Speed: 93.380882 samples/sec	 accuracy=83.662500	 loss=0.625334	 lr=0.001000
Epoch[079] Batch [0099]/[1203]	Speed: 114.688083 samples/sec	 accuracy=83.580000	 loss=0.627469	 lr=0.001000
Epoch[079] Batch [0119]/[1203]	Speed: 183.964659 samples/sec	 accuracy=83.591667	 loss=0.628093	 lr=0.001000
Epoch[079] Batch [0139]/[1203]	Speed: 107.558000 samples/sec	 accuracy=83.700000	 loss=0.625583	 lr=0.001000
Epoch[079] Batch [0159]/[1203]	Speed: 173.677013 samples/sec	 accuracy=83.671875	 loss=0.625215	 lr=0.001000
Epoch[079] Batch [0179]/[1203]	Speed: 107.105906 samples/sec	 accuracy=83.622222	 loss=0.624810	 lr=0.001000
Epoch[079] Batch [0199]/[1203]	Speed: 113.934443 samples/sec	 accuracy=83.640000	 loss=0.624296	 lr=0.001000
Epoch[079] Batch [0219]/[1203]	Speed: 165.780945 samples/sec	 accuracy=83.563636	 loss=0.626844	 lr=0.001000
Epoch[079] Batch [0239]/[1203]	Speed: 112.953236 samples/sec	 accuracy=83.570833	 loss=0.627053	 lr=0.001000
Epoch[079] Batch [0259]/[1203]	Speed: 119.367436 samples/sec	 accuracy=83.553846	 loss=0.627085	 lr=0.001000
Epoch[079] Batch [0279]/[1203]	Speed: 134.346149 samples/sec	 accuracy=83.576786	 loss=0.627073	 lr=0.001000
Epoch[079] Batch [0299]/[1203]	Speed: 118.507243 samples/sec	 accuracy=83.550000	 loss=0.627286	 lr=0.001000
Epoch[079] Batch [0319]/[1203]	Speed: 136.490697 samples/sec	 accuracy=83.529688	 loss=0.627245	 lr=0.001000
Epoch[079] Batch [0339]/[1203]	Speed: 130.921588 samples/sec	 accuracy=83.526471	 loss=0.628185	 lr=0.001000
Epoch[079] Batch [0359]/[1203]	Speed: 117.242907 samples/sec	 accuracy=83.547222	 loss=0.627414	 lr=0.001000
Epoch[079] Batch [0379]/[1203]	Speed: 140.841545 samples/sec	 accuracy=83.576316	 loss=0.627351	 lr=0.001000
Epoch[079] Batch [0399]/[1203]	Speed: 113.774739 samples/sec	 accuracy=83.548750	 loss=0.627895	 lr=0.001000
Epoch[079] Batch [0419]/[1203]	Speed: 144.375611 samples/sec	 accuracy=83.520238	 loss=0.628158	 lr=0.001000
Epoch[079] Batch [0439]/[1203]	Speed: 124.334166 samples/sec	 accuracy=83.514773	 loss=0.628597	 lr=0.001000
Epoch[079] Batch [0459]/[1203]	Speed: 116.109846 samples/sec	 accuracy=83.515217	 loss=0.628921	 lr=0.001000
Epoch[079] Batch [0479]/[1203]	Speed: 164.976091 samples/sec	 accuracy=83.468750	 loss=0.629344	 lr=0.001000
Epoch[079] Batch [0499]/[1203]	Speed: 107.872084 samples/sec	 accuracy=83.463000	 loss=0.629334	 lr=0.001000
Epoch[079] Batch [0519]/[1203]	Speed: 134.393418 samples/sec	 accuracy=83.500962	 loss=0.628587	 lr=0.001000
Epoch[079] Batch [0539]/[1203]	Speed: 137.713261 samples/sec	 accuracy=83.500000	 loss=0.629014	 lr=0.001000
Epoch[079] Batch [0559]/[1203]	Speed: 104.561981 samples/sec	 accuracy=83.491071	 loss=0.629351	 lr=0.001000
Epoch[079] Batch [0579]/[1203]	Speed: 147.133580 samples/sec	 accuracy=83.511207	 loss=0.629391	 lr=0.001000
Epoch[079] Batch [0599]/[1203]	Speed: 129.363476 samples/sec	 accuracy=83.500833	 loss=0.629745	 lr=0.001000
Epoch[079] Batch [0619]/[1203]	Speed: 103.257305 samples/sec	 accuracy=83.473387	 loss=0.630548	 lr=0.001000
Epoch[079] Batch [0639]/[1203]	Speed: 160.583136 samples/sec	 accuracy=83.446094	 loss=0.631421	 lr=0.001000
Epoch[079] Batch [0659]/[1203]	Speed: 111.832089 samples/sec	 accuracy=83.460606	 loss=0.631217	 lr=0.001000
Epoch[079] Batch [0679]/[1203]	Speed: 170.248059 samples/sec	 accuracy=83.483088	 loss=0.630637	 lr=0.001000
Epoch[079] Batch [0699]/[1203]	Speed: 119.456124 samples/sec	 accuracy=83.458571	 loss=0.631067	 lr=0.001000
Epoch[079] Batch [0719]/[1203]	Speed: 106.290611 samples/sec	 accuracy=83.476389	 loss=0.630496	 lr=0.001000
Epoch[079] Batch [0739]/[1203]	Speed: 159.601644 samples/sec	 accuracy=83.440541	 loss=0.631754	 lr=0.001000
Epoch[079] Batch [0759]/[1203]	Speed: 112.757773 samples/sec	 accuracy=83.467105	 loss=0.631198	 lr=0.001000
Epoch[079] Batch [0779]/[1203]	Speed: 121.605979 samples/sec	 accuracy=83.448718	 loss=0.631744	 lr=0.001000
Epoch[079] Batch [0799]/[1203]	Speed: 143.949492 samples/sec	 accuracy=83.446875	 loss=0.631790	 lr=0.001000
Epoch[079] Batch [0819]/[1203]	Speed: 113.962959 samples/sec	 accuracy=83.455488	 loss=0.631388	 lr=0.001000
Epoch[079] Batch [0839]/[1203]	Speed: 158.391917 samples/sec	 accuracy=83.456548	 loss=0.631347	 lr=0.001000
Epoch[079] Batch [0859]/[1203]	Speed: 116.371612 samples/sec	 accuracy=83.444186	 loss=0.631684	 lr=0.001000
Epoch[079] Batch [0879]/[1203]	Speed: 112.866260 samples/sec	 accuracy=83.454545	 loss=0.631259	 lr=0.001000
Epoch[079] Batch [0899]/[1203]	Speed: 155.686603 samples/sec	 accuracy=83.441111	 loss=0.631659	 lr=0.001000
Epoch[079] Batch [0919]/[1203]	Speed: 117.622443 samples/sec	 accuracy=83.434783	 loss=0.631654	 lr=0.001000
Epoch[079] Batch [0939]/[1203]	Speed: 126.643160 samples/sec	 accuracy=83.438830	 loss=0.631643	 lr=0.001000
Epoch[079] Batch [0959]/[1203]	Speed: 131.280273 samples/sec	 accuracy=83.426563	 loss=0.632053	 lr=0.001000
Epoch[079] Batch [0979]/[1203]	Speed: 113.645671 samples/sec	 accuracy=83.436735	 loss=0.632256	 lr=0.001000
Epoch[079] Batch [0999]/[1203]	Speed: 148.933058 samples/sec	 accuracy=83.443500	 loss=0.631781	 lr=0.001000
Epoch[079] Batch [1019]/[1203]	Speed: 127.014973 samples/sec	 accuracy=83.451961	 loss=0.631723	 lr=0.001000
Epoch[079] Batch [1039]/[1203]	Speed: 110.971472 samples/sec	 accuracy=83.440865	 loss=0.631706	 lr=0.001000
Epoch[079] Batch [1059]/[1203]	Speed: 148.310258 samples/sec	 accuracy=83.412264	 loss=0.632779	 lr=0.001000
Epoch[079] Batch [1079]/[1203]	Speed: 108.179166 samples/sec	 accuracy=83.418519	 loss=0.632770	 lr=0.001000
Epoch[079] Batch [1099]/[1203]	Speed: 156.207586 samples/sec	 accuracy=83.400000	 loss=0.633195	 lr=0.001000
Epoch[079] Batch [1119]/[1203]	Speed: 128.069159 samples/sec	 accuracy=83.398214	 loss=0.633264	 lr=0.001000
Epoch[079] Batch [1139]/[1203]	Speed: 115.578917 samples/sec	 accuracy=83.398246	 loss=0.633428	 lr=0.001000
Epoch[079] Batch [1159]/[1203]	Speed: 136.812754 samples/sec	 accuracy=83.385776	 loss=0.633769	 lr=0.001000
Epoch[079] Batch [1179]/[1203]	Speed: 121.191625 samples/sec	 accuracy=83.373729	 loss=0.634040	 lr=0.001000
Epoch[079] Batch [1199]/[1203]	Speed: 293.622261 samples/sec	 accuracy=83.386667	 loss=0.633812	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.150000 acc-top5=84.500000
Batch [0039]/[0097]: acc-top1=62.937500 acc-top5=84.637500
Batch [0059]/[0097]: acc-top1=65.291667 acc-top5=85.991667
Batch [0079]/[0097]: acc-top1=65.325000 acc-top5=85.868750
[Epoch 079] training: accuracy=83.384456	 loss=0.633769
[Epoch 079] speed: 126 samples/sec	time cost: 2098.406562
[Epoch 079] validation: acc-top1=64.752577 acc-top5=85.639175 loss=1.581203
Epoch[080] Batch [0019]/[1203]	Speed: 69.911206 samples/sec	 accuracy=83.825000	 loss=0.606802	 lr=0.000100
Epoch[080] Batch [0039]/[1203]	Speed: 101.941038 samples/sec	 accuracy=84.037500	 loss=0.602645	 lr=0.000100
Epoch[080] Batch [0059]/[1203]	Speed: 218.493945 samples/sec	 accuracy=83.741667	 loss=0.609493	 lr=0.000100
Epoch[080] Batch [0079]/[1203]	Speed: 102.669112 samples/sec	 accuracy=83.618750	 loss=0.614141	 lr=0.000100
Epoch[080] Batch [0099]/[1203]	Speed: 96.301376 samples/sec	 accuracy=83.615000	 loss=0.612258	 lr=0.000100
Epoch[080] Batch [0119]/[1203]	Speed: 248.499649 samples/sec	 accuracy=83.525000	 loss=0.617293	 lr=0.000100
Epoch[080] Batch [0139]/[1203]	Speed: 101.007228 samples/sec	 accuracy=83.582143	 loss=0.618118	 lr=0.000100
Epoch[080] Batch [0159]/[1203]	Speed: 195.387906 samples/sec	 accuracy=83.528125	 loss=0.618906	 lr=0.000100
Epoch[080] Batch [0179]/[1203]	Speed: 110.579277 samples/sec	 accuracy=83.569444	 loss=0.617634	 lr=0.000100
Epoch[080] Batch [0199]/[1203]	Speed: 96.721354 samples/sec	 accuracy=83.547500	 loss=0.618798	 lr=0.000100
Epoch[080] Batch [0219]/[1203]	Speed: 216.576967 samples/sec	 accuracy=83.593182	 loss=0.616328	 lr=0.000100
Epoch[080] Batch [0239]/[1203]	Speed: 100.109131 samples/sec	 accuracy=83.589583	 loss=0.616776	 lr=0.000100
Epoch[080] Batch [0259]/[1203]	Speed: 138.706891 samples/sec	 accuracy=83.582692	 loss=0.617862	 lr=0.000100
Epoch[080] Batch [0279]/[1203]	Speed: 134.902246 samples/sec	 accuracy=83.587500	 loss=0.617043	 lr=0.000100
Epoch[080] Batch [0299]/[1203]	Speed: 110.280278 samples/sec	 accuracy=83.610000	 loss=0.616634	 lr=0.000100
Epoch[080] Batch [0319]/[1203]	Speed: 158.942832 samples/sec	 accuracy=83.590625	 loss=0.617584	 lr=0.000100
Epoch[080] Batch [0339]/[1203]	Speed: 107.677244 samples/sec	 accuracy=83.639706	 loss=0.616510	 lr=0.000100
Epoch[080] Batch [0359]/[1203]	Speed: 125.370767 samples/sec	 accuracy=83.647222	 loss=0.616616	 lr=0.000100
Epoch[080] Batch [0379]/[1203]	Speed: 147.197462 samples/sec	 accuracy=83.631579	 loss=0.616969	 lr=0.000100
Epoch[080] Batch [0399]/[1203]	Speed: 109.885016 samples/sec	 accuracy=83.606250	 loss=0.617416	 lr=0.000100
Epoch[080] Batch [0419]/[1203]	Speed: 181.613689 samples/sec	 accuracy=83.640476	 loss=0.616905	 lr=0.000100
Epoch[080] Batch [0439]/[1203]	Speed: 93.203983 samples/sec	 accuracy=83.675000	 loss=0.616562	 lr=0.000100
Epoch[080] Batch [0459]/[1203]	Speed: 139.923788 samples/sec	 accuracy=83.738043	 loss=0.614851	 lr=0.000100
Epoch[080] Batch [0479]/[1203]	Speed: 142.246139 samples/sec	 accuracy=83.698958	 loss=0.616097	 lr=0.000100
Epoch[080] Batch [0499]/[1203]	Speed: 112.509825 samples/sec	 accuracy=83.707000	 loss=0.616041	 lr=0.000100
Epoch[080] Batch [0519]/[1203]	Speed: 166.793876 samples/sec	 accuracy=83.704808	 loss=0.616655	 lr=0.000100
Epoch[080] Batch [0539]/[1203]	Speed: 107.789907 samples/sec	 accuracy=83.705556	 loss=0.617027	 lr=0.000100
Epoch[080] Batch [0559]/[1203]	Speed: 124.475445 samples/sec	 accuracy=83.712500	 loss=0.616919	 lr=0.000100
Epoch[080] Batch [0579]/[1203]	Speed: 146.240287 samples/sec	 accuracy=83.725000	 loss=0.616353	 lr=0.000100
Epoch[080] Batch [0599]/[1203]	Speed: 108.292090 samples/sec	 accuracy=83.690000	 loss=0.617741	 lr=0.000100
Epoch[080] Batch [0619]/[1203]	Speed: 171.800836 samples/sec	 accuracy=83.728226	 loss=0.616816	 lr=0.000100
Epoch[080] Batch [0639]/[1203]	Speed: 100.502913 samples/sec	 accuracy=83.704687	 loss=0.616759	 lr=0.000100
Epoch[080] Batch [0659]/[1203]	Speed: 143.455098 samples/sec	 accuracy=83.685606	 loss=0.617001	 lr=0.000100
Epoch[080] Batch [0679]/[1203]	Speed: 133.881181 samples/sec	 accuracy=83.694118	 loss=0.616838	 lr=0.000100
Epoch[080] Batch [0699]/[1203]	Speed: 117.997547 samples/sec	 accuracy=83.697857	 loss=0.616847	 lr=0.000100
Epoch[080] Batch [0719]/[1203]	Speed: 133.933267 samples/sec	 accuracy=83.698611	 loss=0.617194	 lr=0.000100
Epoch[080] Batch [0739]/[1203]	Speed: 115.454847 samples/sec	 accuracy=83.703378	 loss=0.616956	 lr=0.000100
Epoch[080] Batch [0759]/[1203]	Speed: 123.191134 samples/sec	 accuracy=83.692763	 loss=0.617284	 lr=0.000100
Epoch[080] Batch [0779]/[1203]	Speed: 139.416026 samples/sec	 accuracy=83.698077	 loss=0.617265	 lr=0.000100
Epoch[080] Batch [0799]/[1203]	Speed: 112.329157 samples/sec	 accuracy=83.691875	 loss=0.617717	 lr=0.000100
Epoch[080] Batch [0819]/[1203]	Speed: 167.272789 samples/sec	 accuracy=83.673780	 loss=0.618043	 lr=0.000100
Epoch[080] Batch [0839]/[1203]	Speed: 105.602290 samples/sec	 accuracy=83.672024	 loss=0.618105	 lr=0.000100
Epoch[080] Batch [0859]/[1203]	Speed: 133.441051 samples/sec	 accuracy=83.680233	 loss=0.617952	 lr=0.000100
Epoch[080] Batch [0879]/[1203]	Speed: 133.186449 samples/sec	 accuracy=83.678977	 loss=0.617840	 lr=0.000100
Epoch[080] Batch [0899]/[1203]	Speed: 116.932873 samples/sec	 accuracy=83.677778	 loss=0.618182	 lr=0.000100
Epoch[080] Batch [0919]/[1203]	Speed: 136.090014 samples/sec	 accuracy=83.672283	 loss=0.618402	 lr=0.000100
Epoch[080] Batch [0939]/[1203]	Speed: 119.782701 samples/sec	 accuracy=83.667021	 loss=0.618669	 lr=0.000100
Epoch[080] Batch [0959]/[1203]	Speed: 129.542154 samples/sec	 accuracy=83.665625	 loss=0.618844	 lr=0.000100
Epoch[080] Batch [0979]/[1203]	Speed: 145.542032 samples/sec	 accuracy=83.680612	 loss=0.618397	 lr=0.000100
Epoch[080] Batch [0999]/[1203]	Speed: 107.262753 samples/sec	 accuracy=83.700000	 loss=0.617586	 lr=0.000100
Epoch[080] Batch [1019]/[1203]	Speed: 142.480436 samples/sec	 accuracy=83.705882	 loss=0.617565	 lr=0.000100
Epoch[080] Batch [1039]/[1203]	Speed: 125.622975 samples/sec	 accuracy=83.719712	 loss=0.617108	 lr=0.000100
Epoch[080] Batch [1059]/[1203]	Speed: 116.306942 samples/sec	 accuracy=83.727830	 loss=0.616938	 lr=0.000100
Epoch[080] Batch [1079]/[1203]	Speed: 148.597262 samples/sec	 accuracy=83.743056	 loss=0.616360	 lr=0.000100
Epoch[080] Batch [1099]/[1203]	Speed: 110.685987 samples/sec	 accuracy=83.734091	 loss=0.616757	 lr=0.000100
Epoch[080] Batch [1119]/[1203]	Speed: 149.749050 samples/sec	 accuracy=83.755357	 loss=0.616113	 lr=0.000100
Epoch[080] Batch [1139]/[1203]	Speed: 116.236087 samples/sec	 accuracy=83.761404	 loss=0.615815	 lr=0.000100
Epoch[080] Batch [1159]/[1203]	Speed: 108.407344 samples/sec	 accuracy=83.760776	 loss=0.616112	 lr=0.000100
Epoch[080] Batch [1179]/[1203]	Speed: 171.113174 samples/sec	 accuracy=83.760169	 loss=0.616015	 lr=0.000100
Epoch[080] Batch [1199]/[1203]	Speed: 226.800069 samples/sec	 accuracy=83.767917	 loss=0.615735	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.600000 acc-top5=84.900000
Batch [0039]/[0097]: acc-top1=63.212500 acc-top5=84.750000
Batch [0059]/[0097]: acc-top1=65.500000 acc-top5=86.033333
Batch [0079]/[0097]: acc-top1=65.500000 acc-top5=85.931250
[Epoch 080] training: accuracy=83.773899	 loss=0.615563
[Epoch 080] speed: 126 samples/sec	time cost: 2097.830710
[Epoch 080] validation: acc-top1=64.896907 acc-top5=85.768041 loss=1.574343
Epoch[081] Batch [0019]/[1203]	Speed: 70.508624 samples/sec	 accuracy=84.200000	 loss=0.584926	 lr=0.000100
Epoch[081] Batch [0039]/[1203]	Speed: 98.439839 samples/sec	 accuracy=84.087500	 loss=0.595320	 lr=0.000100
Epoch[081] Batch [0059]/[1203]	Speed: 268.886760 samples/sec	 accuracy=84.491667	 loss=0.593933	 lr=0.000100
Epoch[081] Batch [0079]/[1203]	Speed: 95.158354 samples/sec	 accuracy=84.293750	 loss=0.601618	 lr=0.000100
Epoch[081] Batch [0099]/[1203]	Speed: 99.682170 samples/sec	 accuracy=84.090000	 loss=0.606467	 lr=0.000100
Epoch[081] Batch [0119]/[1203]	Speed: 230.227491 samples/sec	 accuracy=84.150000	 loss=0.605596	 lr=0.000100
Epoch[081] Batch [0139]/[1203]	Speed: 99.222932 samples/sec	 accuracy=84.192857	 loss=0.605968	 lr=0.000100
Epoch[081] Batch [0159]/[1203]	Speed: 244.016974 samples/sec	 accuracy=84.190625	 loss=0.605779	 lr=0.000100
Epoch[081] Batch [0179]/[1203]	Speed: 92.889157 samples/sec	 accuracy=84.155556	 loss=0.607231	 lr=0.000100
Epoch[081] Batch [0199]/[1203]	Speed: 115.444512 samples/sec	 accuracy=84.172500	 loss=0.608297	 lr=0.000100
Epoch[081] Batch [0219]/[1203]	Speed: 185.031213 samples/sec	 accuracy=84.190909	 loss=0.608056	 lr=0.000100
Epoch[081] Batch [0239]/[1203]	Speed: 100.000155 samples/sec	 accuracy=84.208333	 loss=0.605909	 lr=0.000100
Epoch[081] Batch [0259]/[1203]	Speed: 131.195682 samples/sec	 accuracy=84.225000	 loss=0.603879	 lr=0.000100
Epoch[081] Batch [0279]/[1203]	Speed: 139.308241 samples/sec	 accuracy=84.208929	 loss=0.603859	 lr=0.000100
Epoch[081] Batch [0299]/[1203]	Speed: 101.624432 samples/sec	 accuracy=84.118333	 loss=0.606724	 lr=0.000100
Epoch[081] Batch [0319]/[1203]	Speed: 177.647556 samples/sec	 accuracy=84.067187	 loss=0.608470	 lr=0.000100
Epoch[081] Batch [0339]/[1203]	Speed: 106.833770 samples/sec	 accuracy=84.038235	 loss=0.608630	 lr=0.000100
Epoch[081] Batch [0359]/[1203]	Speed: 140.264599 samples/sec	 accuracy=84.061111	 loss=0.607225	 lr=0.000100
Epoch[081] Batch [0379]/[1203]	Speed: 130.454874 samples/sec	 accuracy=84.086842	 loss=0.607025	 lr=0.000100
Epoch[081] Batch [0399]/[1203]	Speed: 97.995676 samples/sec	 accuracy=84.106250	 loss=0.606740	 lr=0.000100
Epoch[081] Batch [0419]/[1203]	Speed: 220.269172 samples/sec	 accuracy=84.125000	 loss=0.606861	 lr=0.000100
Epoch[081] Batch [0439]/[1203]	Speed: 103.596720 samples/sec	 accuracy=84.143182	 loss=0.606728	 lr=0.000100
Epoch[081] Batch [0459]/[1203]	Speed: 116.860806 samples/sec	 accuracy=84.159783	 loss=0.605739	 lr=0.000100
Epoch[081] Batch [0479]/[1203]	Speed: 164.639326 samples/sec	 accuracy=84.173958	 loss=0.605543	 lr=0.000100
Epoch[081] Batch [0499]/[1203]	Speed: 101.644331 samples/sec	 accuracy=84.160000	 loss=0.606118	 lr=0.000100
Epoch[081] Batch [0519]/[1203]	Speed: 180.654748 samples/sec	 accuracy=84.116346	 loss=0.607010	 lr=0.000100
Epoch[081] Batch [0539]/[1203]	Speed: 104.657689 samples/sec	 accuracy=84.100000	 loss=0.607329	 lr=0.000100
Epoch[081] Batch [0559]/[1203]	Speed: 106.596856 samples/sec	 accuracy=84.111607	 loss=0.607379	 lr=0.000100
Epoch[081] Batch [0579]/[1203]	Speed: 194.618147 samples/sec	 accuracy=84.102586	 loss=0.607676	 lr=0.000100
Epoch[081] Batch [0599]/[1203]	Speed: 99.798241 samples/sec	 accuracy=84.087500	 loss=0.607979	 lr=0.000100
Epoch[081] Batch [0619]/[1203]	Speed: 149.907807 samples/sec	 accuracy=84.071774	 loss=0.608195	 lr=0.000100
Epoch[081] Batch [0639]/[1203]	Speed: 119.621269 samples/sec	 accuracy=84.090625	 loss=0.607407	 lr=0.000100
Epoch[081] Batch [0659]/[1203]	Speed: 116.976909 samples/sec	 accuracy=84.084848	 loss=0.607653	 lr=0.000100
Epoch[081] Batch [0679]/[1203]	Speed: 160.489111 samples/sec	 accuracy=84.047794	 loss=0.608553	 lr=0.000100
Epoch[081] Batch [0699]/[1203]	Speed: 107.770609 samples/sec	 accuracy=84.014286	 loss=0.609028	 lr=0.000100
Epoch[081] Batch [0719]/[1203]	Speed: 122.572195 samples/sec	 accuracy=84.002083	 loss=0.609285	 lr=0.000100
Epoch[081] Batch [0739]/[1203]	Speed: 137.042247 samples/sec	 accuracy=84.006757	 loss=0.609406	 lr=0.000100
Epoch[081] Batch [0759]/[1203]	Speed: 122.741498 samples/sec	 accuracy=84.009211	 loss=0.609487	 lr=0.000100
Epoch[081] Batch [0779]/[1203]	Speed: 166.351465 samples/sec	 accuracy=84.010256	 loss=0.609605	 lr=0.000100
Epoch[081] Batch [0799]/[1203]	Speed: 105.302034 samples/sec	 accuracy=84.011875	 loss=0.609474	 lr=0.000100
Epoch[081] Batch [0819]/[1203]	Speed: 121.319104 samples/sec	 accuracy=84.004268	 loss=0.609545	 lr=0.000100
Epoch[081] Batch [0839]/[1203]	Speed: 130.495958 samples/sec	 accuracy=84.025595	 loss=0.609332	 lr=0.000100
Epoch[081] Batch [0859]/[1203]	Speed: 127.233898 samples/sec	 accuracy=84.004651	 loss=0.610000	 lr=0.000100
Epoch[081] Batch [0879]/[1203]	Speed: 148.593358 samples/sec	 accuracy=84.006818	 loss=0.610216	 lr=0.000100
Epoch[081] Batch [0899]/[1203]	Speed: 111.350380 samples/sec	 accuracy=84.023889	 loss=0.609750	 lr=0.000100
Epoch[081] Batch [0919]/[1203]	Speed: 107.102203 samples/sec	 accuracy=84.048913	 loss=0.609260	 lr=0.000100
Epoch[081] Batch [0939]/[1203]	Speed: 175.765047 samples/sec	 accuracy=84.064362	 loss=0.608791	 lr=0.000100
Epoch[081] Batch [0959]/[1203]	Speed: 110.295231 samples/sec	 accuracy=84.070833	 loss=0.608845	 lr=0.000100
Epoch[081] Batch [0979]/[1203]	Speed: 142.623914 samples/sec	 accuracy=84.098980	 loss=0.608291	 lr=0.000100
Epoch[081] Batch [0999]/[1203]	Speed: 117.625031 samples/sec	 accuracy=84.101000	 loss=0.607759	 lr=0.000100
Epoch[081] Batch [1019]/[1203]	Speed: 113.864303 samples/sec	 accuracy=84.090686	 loss=0.607975	 lr=0.000100
Epoch[081] Batch [1039]/[1203]	Speed: 152.829265 samples/sec	 accuracy=84.105288	 loss=0.607641	 lr=0.000100
Epoch[081] Batch [1059]/[1203]	Speed: 119.025723 samples/sec	 accuracy=84.091509	 loss=0.607876	 lr=0.000100
Epoch[081] Batch [1079]/[1203]	Speed: 118.121705 samples/sec	 accuracy=84.114352	 loss=0.607476	 lr=0.000100
Epoch[081] Batch [1099]/[1203]	Speed: 139.891824 samples/sec	 accuracy=84.108636	 loss=0.607575	 lr=0.000100
Epoch[081] Batch [1119]/[1203]	Speed: 118.382334 samples/sec	 accuracy=84.095089	 loss=0.607977	 lr=0.000100
Epoch[081] Batch [1139]/[1203]	Speed: 150.447656 samples/sec	 accuracy=84.090789	 loss=0.608057	 lr=0.000100
Epoch[081] Batch [1159]/[1203]	Speed: 108.977639 samples/sec	 accuracy=84.091379	 loss=0.608009	 lr=0.000100
Epoch[081] Batch [1179]/[1203]	Speed: 139.094992 samples/sec	 accuracy=84.089407	 loss=0.607976	 lr=0.000100
Epoch[081] Batch [1199]/[1203]	Speed: 237.352999 samples/sec	 accuracy=84.074583	 loss=0.608316	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.900000 acc-top5=85.000000
Batch [0039]/[0097]: acc-top1=63.500000 acc-top5=84.837500
Batch [0059]/[0097]: acc-top1=65.558333 acc-top5=86.091667
Batch [0079]/[0097]: acc-top1=65.693750 acc-top5=85.956250
[Epoch 081] training: accuracy=84.072319	 loss=0.608351
[Epoch 081] speed: 125 samples/sec	time cost: 2105.224039
[Epoch 081] validation: acc-top1=65.103093 acc-top5=85.721649 loss=1.569823
Epoch[082] Batch [0019]/[1203]	Speed: 70.664229 samples/sec	 accuracy=83.975000	 loss=0.618871	 lr=0.000100
Epoch[082] Batch [0039]/[1203]	Speed: 102.507002 samples/sec	 accuracy=84.362500	 loss=0.596739	 lr=0.000100
Epoch[082] Batch [0059]/[1203]	Speed: 251.033388 samples/sec	 accuracy=84.516667	 loss=0.596392	 lr=0.000100
Epoch[082] Batch [0079]/[1203]	Speed: 95.370500 samples/sec	 accuracy=84.362500	 loss=0.600668	 lr=0.000100
Epoch[082] Batch [0099]/[1203]	Speed: 98.739742 samples/sec	 accuracy=84.325000	 loss=0.600564	 lr=0.000100
Epoch[082] Batch [0119]/[1203]	Speed: 245.387536 samples/sec	 accuracy=84.308333	 loss=0.601914	 lr=0.000100
Epoch[082] Batch [0139]/[1203]	Speed: 95.788722 samples/sec	 accuracy=84.367857	 loss=0.598604	 lr=0.000100
Epoch[082] Batch [0159]/[1203]	Speed: 240.943661 samples/sec	 accuracy=84.350000	 loss=0.598744	 lr=0.000100
Epoch[082] Batch [0179]/[1203]	Speed: 98.385417 samples/sec	 accuracy=84.333333	 loss=0.600854	 lr=0.000100
Epoch[082] Batch [0199]/[1203]	Speed: 104.954912 samples/sec	 accuracy=84.290000	 loss=0.600789	 lr=0.000100
Epoch[082] Batch [0219]/[1203]	Speed: 213.745407 samples/sec	 accuracy=84.295455	 loss=0.599692	 lr=0.000100
Epoch[082] Batch [0239]/[1203]	Speed: 101.707564 samples/sec	 accuracy=84.327083	 loss=0.598138	 lr=0.000100
Epoch[082] Batch [0259]/[1203]	Speed: 109.228571 samples/sec	 accuracy=84.301923	 loss=0.598912	 lr=0.000100
Epoch[082] Batch [0279]/[1203]	Speed: 181.518576 samples/sec	 accuracy=84.246429	 loss=0.601040	 lr=0.000100
Epoch[082] Batch [0299]/[1203]	Speed: 98.413614 samples/sec	 accuracy=84.290000	 loss=0.600639	 lr=0.000100
Epoch[082] Batch [0319]/[1203]	Speed: 213.228128 samples/sec	 accuracy=84.231250	 loss=0.600759	 lr=0.000100
Epoch[082] Batch [0339]/[1203]	Speed: 104.694873 samples/sec	 accuracy=84.264706	 loss=0.599235	 lr=0.000100
Epoch[082] Batch [0359]/[1203]	Speed: 117.334733 samples/sec	 accuracy=84.313889	 loss=0.598750	 lr=0.000100
Epoch[082] Batch [0379]/[1203]	Speed: 159.932774 samples/sec	 accuracy=84.359211	 loss=0.599123	 lr=0.000100
Epoch[082] Batch [0399]/[1203]	Speed: 103.633049 samples/sec	 accuracy=84.373750	 loss=0.598338	 lr=0.000100
Epoch[082] Batch [0419]/[1203]	Speed: 131.442575 samples/sec	 accuracy=84.321429	 loss=0.599744	 lr=0.000100
Epoch[082] Batch [0439]/[1203]	Speed: 133.674060 samples/sec	 accuracy=84.306818	 loss=0.600570	 lr=0.000100
Epoch[082] Batch [0459]/[1203]	Speed: 119.451699 samples/sec	 accuracy=84.227174	 loss=0.602029	 lr=0.000100
Epoch[082] Batch [0479]/[1203]	Speed: 145.504141 samples/sec	 accuracy=84.258333	 loss=0.601512	 lr=0.000100
Epoch[082] Batch [0499]/[1203]	Speed: 107.985594 samples/sec	 accuracy=84.308000	 loss=0.600871	 lr=0.000100
Epoch[082] Batch [0519]/[1203]	Speed: 161.329488 samples/sec	 accuracy=84.278846	 loss=0.602346	 lr=0.000100
Epoch[082] Batch [0539]/[1203]	Speed: 112.564559 samples/sec	 accuracy=84.291667	 loss=0.602296	 lr=0.000100
Epoch[082] Batch [0559]/[1203]	Speed: 108.756066 samples/sec	 accuracy=84.292857	 loss=0.602521	 lr=0.000100
Epoch[082] Batch [0579]/[1203]	Speed: 178.308335 samples/sec	 accuracy=84.264655	 loss=0.603577	 lr=0.000100
Epoch[082] Batch [0599]/[1203]	Speed: 107.515931 samples/sec	 accuracy=84.269167	 loss=0.604050	 lr=0.000100
Epoch[082] Batch [0619]/[1203]	Speed: 136.427814 samples/sec	 accuracy=84.262097	 loss=0.604003	 lr=0.000100
Epoch[082] Batch [0639]/[1203]	Speed: 122.775860 samples/sec	 accuracy=84.243750	 loss=0.604009	 lr=0.000100
Epoch[082] Batch [0659]/[1203]	Speed: 113.633444 samples/sec	 accuracy=84.258333	 loss=0.604184	 lr=0.000100
Epoch[082] Batch [0679]/[1203]	Speed: 185.796201 samples/sec	 accuracy=84.211765	 loss=0.605266	 lr=0.000100
Epoch[082] Batch [0699]/[1203]	Speed: 108.061213 samples/sec	 accuracy=84.194286	 loss=0.605457	 lr=0.000100
Epoch[082] Batch [0719]/[1203]	Speed: 120.258495 samples/sec	 accuracy=84.209722	 loss=0.605098	 lr=0.000100
Epoch[082] Batch [0739]/[1203]	Speed: 143.802117 samples/sec	 accuracy=84.204730	 loss=0.605573	 lr=0.000100
Epoch[082] Batch [0759]/[1203]	Speed: 109.078459 samples/sec	 accuracy=84.211184	 loss=0.605262	 lr=0.000100
Epoch[082] Batch [0779]/[1203]	Speed: 150.130764 samples/sec	 accuracy=84.196795	 loss=0.605471	 lr=0.000100
Epoch[082] Batch [0799]/[1203]	Speed: 121.877183 samples/sec	 accuracy=84.189375	 loss=0.605539	 lr=0.000100
Epoch[082] Batch [0819]/[1203]	Speed: 117.049663 samples/sec	 accuracy=84.181707	 loss=0.605615	 lr=0.000100
Epoch[082] Batch [0839]/[1203]	Speed: 138.700340 samples/sec	 accuracy=84.183333	 loss=0.605757	 lr=0.000100
Epoch[082] Batch [0859]/[1203]	Speed: 120.690217 samples/sec	 accuracy=84.203488	 loss=0.605120	 lr=0.000100
Epoch[082] Batch [0879]/[1203]	Speed: 138.116856 samples/sec	 accuracy=84.193750	 loss=0.605250	 lr=0.000100
Epoch[082] Batch [0899]/[1203]	Speed: 134.341057 samples/sec	 accuracy=84.181111	 loss=0.605627	 lr=0.000100
Epoch[082] Batch [0919]/[1203]	Speed: 107.931895 samples/sec	 accuracy=84.179348	 loss=0.605771	 lr=0.000100
Epoch[082] Batch [0939]/[1203]	Speed: 158.065287 samples/sec	 accuracy=84.187234	 loss=0.605621	 lr=0.000100
Epoch[082] Batch [0959]/[1203]	Speed: 118.011872 samples/sec	 accuracy=84.180729	 loss=0.605577	 lr=0.000100
Epoch[082] Batch [0979]/[1203]	Speed: 116.496813 samples/sec	 accuracy=84.167347	 loss=0.605733	 lr=0.000100
Epoch[082] Batch [0999]/[1203]	Speed: 149.708057 samples/sec	 accuracy=84.160500	 loss=0.606208	 lr=0.000100
Epoch[082] Batch [1019]/[1203]	Speed: 104.253838 samples/sec	 accuracy=84.152941	 loss=0.606213	 lr=0.000100
Epoch[082] Batch [1039]/[1203]	Speed: 158.606261 samples/sec	 accuracy=84.138942	 loss=0.606456	 lr=0.000100
Epoch[082] Batch [1059]/[1203]	Speed: 116.722179 samples/sec	 accuracy=84.124057	 loss=0.606834	 lr=0.000100
Epoch[082] Batch [1079]/[1203]	Speed: 117.228266 samples/sec	 accuracy=84.111111	 loss=0.607059	 lr=0.000100
Epoch[082] Batch [1099]/[1203]	Speed: 158.977906 samples/sec	 accuracy=84.114545	 loss=0.607028	 lr=0.000100
Epoch[082] Batch [1119]/[1203]	Speed: 110.342105 samples/sec	 accuracy=84.108929	 loss=0.606997	 lr=0.000100
Epoch[082] Batch [1139]/[1203]	Speed: 135.948408 samples/sec	 accuracy=84.100877	 loss=0.607044	 lr=0.000100
Epoch[082] Batch [1159]/[1203]	Speed: 131.752778 samples/sec	 accuracy=84.094397	 loss=0.606869	 lr=0.000100
Epoch[082] Batch [1179]/[1203]	Speed: 124.431879 samples/sec	 accuracy=84.080932	 loss=0.607181	 lr=0.000100
Epoch[082] Batch [1199]/[1203]	Speed: 262.597983 samples/sec	 accuracy=84.064583	 loss=0.607566	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.525000 acc-top5=84.850000
Batch [0039]/[0097]: acc-top1=63.175000 acc-top5=84.762500
Batch [0059]/[0097]: acc-top1=65.300000 acc-top5=86.083333
Batch [0079]/[0097]: acc-top1=65.437500 acc-top5=85.937500
[Epoch 082] training: accuracy=84.061513	 loss=0.607629
[Epoch 082] speed: 126 samples/sec	time cost: 2094.405712
[Epoch 082] validation: acc-top1=64.938144 acc-top5=85.793814 loss=1.570771
Epoch[083] Batch [0019]/[1203]	Speed: 69.054910 samples/sec	 accuracy=83.875000	 loss=0.621050	 lr=0.000100
Epoch[083] Batch [0039]/[1203]	Speed: 103.844462 samples/sec	 accuracy=83.375000	 loss=0.626275	 lr=0.000100
Epoch[083] Batch [0059]/[1203]	Speed: 218.738354 samples/sec	 accuracy=83.675000	 loss=0.619877	 lr=0.000100
Epoch[083] Batch [0079]/[1203]	Speed: 98.206953 samples/sec	 accuracy=83.668750	 loss=0.619614	 lr=0.000100
Epoch[083] Batch [0099]/[1203]	Speed: 99.855354 samples/sec	 accuracy=83.765000	 loss=0.615814	 lr=0.000100
Epoch[083] Batch [0119]/[1203]	Speed: 254.806485 samples/sec	 accuracy=83.795833	 loss=0.613011	 lr=0.000100
Epoch[083] Batch [0139]/[1203]	Speed: 96.501200 samples/sec	 accuracy=83.900000	 loss=0.608423	 lr=0.000100
Epoch[083] Batch [0159]/[1203]	Speed: 210.607703 samples/sec	 accuracy=83.912500	 loss=0.607282	 lr=0.000100
Epoch[083] Batch [0179]/[1203]	Speed: 103.036034 samples/sec	 accuracy=83.966667	 loss=0.608075	 lr=0.000100
Epoch[083] Batch [0199]/[1203]	Speed: 99.769656 samples/sec	 accuracy=84.045000	 loss=0.603964	 lr=0.000100
Epoch[083] Batch [0219]/[1203]	Speed: 221.528346 samples/sec	 accuracy=84.018182	 loss=0.603002	 lr=0.000100
Epoch[083] Batch [0239]/[1203]	Speed: 102.065579 samples/sec	 accuracy=84.010417	 loss=0.604847	 lr=0.000100
Epoch[083] Batch [0259]/[1203]	Speed: 119.059212 samples/sec	 accuracy=84.013462	 loss=0.604953	 lr=0.000100
Epoch[083] Batch [0279]/[1203]	Speed: 165.150823 samples/sec	 accuracy=84.085714	 loss=0.603593	 lr=0.000100
Epoch[083] Batch [0299]/[1203]	Speed: 103.817858 samples/sec	 accuracy=84.120000	 loss=0.602460	 lr=0.000100
Epoch[083] Batch [0319]/[1203]	Speed: 146.122518 samples/sec	 accuracy=84.092188	 loss=0.603015	 lr=0.000100
Epoch[083] Batch [0339]/[1203]	Speed: 118.485796 samples/sec	 accuracy=84.127941	 loss=0.601783	 lr=0.000100
Epoch[083] Batch [0359]/[1203]	Speed: 115.480348 samples/sec	 accuracy=84.059722	 loss=0.604065	 lr=0.000100
Epoch[083] Batch [0379]/[1203]	Speed: 145.828713 samples/sec	 accuracy=84.053947	 loss=0.604618	 lr=0.000100
Epoch[083] Batch [0399]/[1203]	Speed: 115.958482 samples/sec	 accuracy=84.055000	 loss=0.605103	 lr=0.000100
Epoch[083] Batch [0419]/[1203]	Speed: 127.880686 samples/sec	 accuracy=84.044048	 loss=0.605724	 lr=0.000100
Epoch[083] Batch [0439]/[1203]	Speed: 148.846634 samples/sec	 accuracy=84.011364	 loss=0.606998	 lr=0.000100
Epoch[083] Batch [0459]/[1203]	Speed: 115.087618 samples/sec	 accuracy=84.018478	 loss=0.606898	 lr=0.000100
Epoch[083] Batch [0479]/[1203]	Speed: 139.537288 samples/sec	 accuracy=84.052083	 loss=0.606453	 lr=0.000100
Epoch[083] Batch [0499]/[1203]	Speed: 112.479535 samples/sec	 accuracy=84.045000	 loss=0.606874	 lr=0.000100
Epoch[083] Batch [0519]/[1203]	Speed: 123.898963 samples/sec	 accuracy=84.037500	 loss=0.607467	 lr=0.000100
Epoch[083] Batch [0539]/[1203]	Speed: 142.712973 samples/sec	 accuracy=84.041667	 loss=0.607484	 lr=0.000100
Epoch[083] Batch [0559]/[1203]	Speed: 128.116465 samples/sec	 accuracy=84.008036	 loss=0.608150	 lr=0.000100
Epoch[083] Batch [0579]/[1203]	Speed: 114.364616 samples/sec	 accuracy=84.006034	 loss=0.608061	 lr=0.000100
Epoch[083] Batch [0599]/[1203]	Speed: 132.052996 samples/sec	 accuracy=84.060000	 loss=0.607240	 lr=0.000100
Epoch[083] Batch [0619]/[1203]	Speed: 123.615701 samples/sec	 accuracy=84.033871	 loss=0.608005	 lr=0.000100
Epoch[083] Batch [0639]/[1203]	Speed: 140.103975 samples/sec	 accuracy=84.039844	 loss=0.608380	 lr=0.000100
Epoch[083] Batch [0659]/[1203]	Speed: 108.810369 samples/sec	 accuracy=84.043939	 loss=0.608583	 lr=0.000100
Epoch[083] Batch [0679]/[1203]	Speed: 141.965167 samples/sec	 accuracy=84.071324	 loss=0.607877	 lr=0.000100
Epoch[083] Batch [0699]/[1203]	Speed: 132.756292 samples/sec	 accuracy=84.094286	 loss=0.607477	 lr=0.000100
Epoch[083] Batch [0719]/[1203]	Speed: 125.645664 samples/sec	 accuracy=84.094444	 loss=0.607343	 lr=0.000100
Epoch[083] Batch [0739]/[1203]	Speed: 121.986419 samples/sec	 accuracy=84.085811	 loss=0.607788	 lr=0.000100
Epoch[083] Batch [0759]/[1203]	Speed: 119.642989 samples/sec	 accuracy=84.082895	 loss=0.607128	 lr=0.000100
Epoch[083] Batch [0779]/[1203]	Speed: 127.884158 samples/sec	 accuracy=84.097436	 loss=0.607111	 lr=0.000100
Epoch[083] Batch [0799]/[1203]	Speed: 148.155699 samples/sec	 accuracy=84.103125	 loss=0.606688	 lr=0.000100
Epoch[083] Batch [0819]/[1203]	Speed: 115.626757 samples/sec	 accuracy=84.100610	 loss=0.606525	 lr=0.000100
Epoch[083] Batch [0839]/[1203]	Speed: 118.842644 samples/sec	 accuracy=84.078571	 loss=0.607260	 lr=0.000100
Epoch[083] Batch [0859]/[1203]	Speed: 130.825458 samples/sec	 accuracy=84.083140	 loss=0.607129	 lr=0.000100
Epoch[083] Batch [0879]/[1203]	Speed: 136.500560 samples/sec	 accuracy=84.090909	 loss=0.607057	 lr=0.000100
Epoch[083] Batch [0899]/[1203]	Speed: 128.640420 samples/sec	 accuracy=84.109444	 loss=0.606834	 lr=0.000100
Epoch[083] Batch [0919]/[1203]	Speed: 120.841694 samples/sec	 accuracy=84.116848	 loss=0.606615	 lr=0.000100
Epoch[083] Batch [0939]/[1203]	Speed: 119.087171 samples/sec	 accuracy=84.135638	 loss=0.606183	 lr=0.000100
Epoch[083] Batch [0959]/[1203]	Speed: 129.295578 samples/sec	 accuracy=84.133333	 loss=0.606576	 lr=0.000100
Epoch[083] Batch [0979]/[1203]	Speed: 132.784362 samples/sec	 accuracy=84.138776	 loss=0.606726	 lr=0.000100
Epoch[083] Batch [0999]/[1203]	Speed: 136.483114 samples/sec	 accuracy=84.123000	 loss=0.607161	 lr=0.000100
Epoch[083] Batch [1019]/[1203]	Speed: 119.826747 samples/sec	 accuracy=84.120098	 loss=0.607068	 lr=0.000100
Epoch[083] Batch [1039]/[1203]	Speed: 120.320031 samples/sec	 accuracy=84.113942	 loss=0.607184	 lr=0.000100
Epoch[083] Batch [1059]/[1203]	Speed: 128.765249 samples/sec	 accuracy=84.121226	 loss=0.606899	 lr=0.000100
Epoch[083] Batch [1079]/[1203]	Speed: 117.279221 samples/sec	 accuracy=84.106481	 loss=0.607208	 lr=0.000100
Epoch[083] Batch [1099]/[1203]	Speed: 142.900037 samples/sec	 accuracy=84.120909	 loss=0.606699	 lr=0.000100
Epoch[083] Batch [1119]/[1203]	Speed: 127.932509 samples/sec	 accuracy=84.124554	 loss=0.606697	 lr=0.000100
Epoch[083] Batch [1139]/[1203]	Speed: 113.281279 samples/sec	 accuracy=84.117544	 loss=0.606775	 lr=0.000100
Epoch[083] Batch [1159]/[1203]	Speed: 125.641260 samples/sec	 accuracy=84.119828	 loss=0.606897	 lr=0.000100
Epoch[083] Batch [1179]/[1203]	Speed: 134.753268 samples/sec	 accuracy=84.111441	 loss=0.607367	 lr=0.000100
Epoch[083] Batch [1199]/[1203]	Speed: 282.742364 samples/sec	 accuracy=84.126667	 loss=0.606852	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.500000 acc-top5=84.875000
Batch [0039]/[0097]: acc-top1=63.137500 acc-top5=84.787500
Batch [0059]/[0097]: acc-top1=65.291667 acc-top5=85.958333
Batch [0079]/[0097]: acc-top1=65.512500 acc-top5=85.862500
[Epoch 083] training: accuracy=84.130507	 loss=0.606712
[Epoch 083] speed: 126 samples/sec	time cost: 2102.479611
[Epoch 083] validation: acc-top1=64.984536 acc-top5=85.716495 loss=1.569822
Epoch[084] Batch [0019]/[1203]	Speed: 70.696431 samples/sec	 accuracy=83.700000	 loss=0.613207	 lr=0.000100
Epoch[084] Batch [0039]/[1203]	Speed: 106.433639 samples/sec	 accuracy=83.925000	 loss=0.612806	 lr=0.000100
Epoch[084] Batch [0059]/[1203]	Speed: 208.056545 samples/sec	 accuracy=84.200000	 loss=0.608481	 lr=0.000100
Epoch[084] Batch [0079]/[1203]	Speed: 93.205008 samples/sec	 accuracy=84.237500	 loss=0.603682	 lr=0.000100
Epoch[084] Batch [0099]/[1203]	Speed: 134.023794 samples/sec	 accuracy=84.345000	 loss=0.603209	 lr=0.000100
Epoch[084] Batch [0119]/[1203]	Speed: 153.001750 samples/sec	 accuracy=84.287500	 loss=0.603249	 lr=0.000100
Epoch[084] Batch [0139]/[1203]	Speed: 112.630329 samples/sec	 accuracy=84.214286	 loss=0.603553	 lr=0.000100
Epoch[084] Batch [0159]/[1203]	Speed: 172.352748 samples/sec	 accuracy=84.140625	 loss=0.606212	 lr=0.000100
Epoch[084] Batch [0179]/[1203]	Speed: 96.599241 samples/sec	 accuracy=84.166667	 loss=0.607870	 lr=0.000100
Epoch[084] Batch [0199]/[1203]	Speed: 126.147495 samples/sec	 accuracy=84.115000	 loss=0.607816	 lr=0.000100
Epoch[084] Batch [0219]/[1203]	Speed: 146.448355 samples/sec	 accuracy=84.122727	 loss=0.608673	 lr=0.000100
Epoch[084] Batch [0239]/[1203]	Speed: 106.807240 samples/sec	 accuracy=84.114583	 loss=0.608256	 lr=0.000100
Epoch[084] Batch [0259]/[1203]	Speed: 209.580995 samples/sec	 accuracy=84.100000	 loss=0.609894	 lr=0.000100
Epoch[084] Batch [0279]/[1203]	Speed: 100.583150 samples/sec	 accuracy=84.107143	 loss=0.609657	 lr=0.000100
Epoch[084] Batch [0299]/[1203]	Speed: 130.270320 samples/sec	 accuracy=84.100000	 loss=0.610557	 lr=0.000100
Epoch[084] Batch [0319]/[1203]	Speed: 158.727137 samples/sec	 accuracy=84.154688	 loss=0.608758	 lr=0.000100
Epoch[084] Batch [0339]/[1203]	Speed: 95.964475 samples/sec	 accuracy=84.173529	 loss=0.608400	 lr=0.000100
Epoch[084] Batch [0359]/[1203]	Speed: 184.732818 samples/sec	 accuracy=84.188889	 loss=0.607295	 lr=0.000100
Epoch[084] Batch [0379]/[1203]	Speed: 107.373684 samples/sec	 accuracy=84.180263	 loss=0.607213	 lr=0.000100
Epoch[084] Batch [0399]/[1203]	Speed: 118.294047 samples/sec	 accuracy=84.222500	 loss=0.606018	 lr=0.000100
Epoch[084] Batch [0419]/[1203]	Speed: 143.037967 samples/sec	 accuracy=84.208333	 loss=0.606707	 lr=0.000100
Epoch[084] Batch [0439]/[1203]	Speed: 118.259449 samples/sec	 accuracy=84.239773	 loss=0.605527	 lr=0.000100
Epoch[084] Batch [0459]/[1203]	Speed: 164.308206 samples/sec	 accuracy=84.246739	 loss=0.604472	 lr=0.000100
Epoch[084] Batch [0479]/[1203]	Speed: 103.714861 samples/sec	 accuracy=84.293750	 loss=0.602878	 lr=0.000100
Epoch[084] Batch [0499]/[1203]	Speed: 125.515500 samples/sec	 accuracy=84.252000	 loss=0.603504	 lr=0.000100
Epoch[084] Batch [0519]/[1203]	Speed: 148.991681 samples/sec	 accuracy=84.270192	 loss=0.602684	 lr=0.000100
Epoch[084] Batch [0539]/[1203]	Speed: 113.594457 samples/sec	 accuracy=84.254630	 loss=0.602923	 lr=0.000100
Epoch[084] Batch [0559]/[1203]	Speed: 144.338578 samples/sec	 accuracy=84.233929	 loss=0.602798	 lr=0.000100
Epoch[084] Batch [0579]/[1203]	Speed: 118.050432 samples/sec	 accuracy=84.231034	 loss=0.603198	 lr=0.000100
Epoch[084] Batch [0599]/[1203]	Speed: 123.273386 samples/sec	 accuracy=84.222500	 loss=0.602893	 lr=0.000100
Epoch[084] Batch [0619]/[1203]	Speed: 146.076377 samples/sec	 accuracy=84.232258	 loss=0.603403	 lr=0.000100
Epoch[084] Batch [0639]/[1203]	Speed: 115.203104 samples/sec	 accuracy=84.192969	 loss=0.604196	 lr=0.000100
Epoch[084] Batch [0659]/[1203]	Speed: 135.383665 samples/sec	 accuracy=84.194697	 loss=0.603971	 lr=0.000100
Epoch[084] Batch [0679]/[1203]	Speed: 127.261423 samples/sec	 accuracy=84.197059	 loss=0.604242	 lr=0.000100
Epoch[084] Batch [0699]/[1203]	Speed: 110.344188 samples/sec	 accuracy=84.177143	 loss=0.604358	 lr=0.000100
Epoch[084] Batch [0719]/[1203]	Speed: 160.593989 samples/sec	 accuracy=84.177778	 loss=0.604670	 lr=0.000100
Epoch[084] Batch [0739]/[1203]	Speed: 108.872824 samples/sec	 accuracy=84.178378	 loss=0.604359	 lr=0.000100
Epoch[084] Batch [0759]/[1203]	Speed: 120.692326 samples/sec	 accuracy=84.161184	 loss=0.605186	 lr=0.000100
Epoch[084] Batch [0779]/[1203]	Speed: 147.574573 samples/sec	 accuracy=84.143590	 loss=0.605821	 lr=0.000100
Epoch[084] Batch [0799]/[1203]	Speed: 116.191771 samples/sec	 accuracy=84.148750	 loss=0.605385	 lr=0.000100
Epoch[084] Batch [0819]/[1203]	Speed: 150.180157 samples/sec	 accuracy=84.138415	 loss=0.605414	 lr=0.000100
Epoch[084] Batch [0839]/[1203]	Speed: 107.464474 samples/sec	 accuracy=84.129762	 loss=0.605373	 lr=0.000100
Epoch[084] Batch [0859]/[1203]	Speed: 125.777542 samples/sec	 accuracy=84.123837	 loss=0.605782	 lr=0.000100
Epoch[084] Batch [0879]/[1203]	Speed: 146.594474 samples/sec	 accuracy=84.104545	 loss=0.607042	 lr=0.000100
Epoch[084] Batch [0899]/[1203]	Speed: 123.330924 samples/sec	 accuracy=84.110556	 loss=0.606709	 lr=0.000100
Epoch[084] Batch [0919]/[1203]	Speed: 130.783674 samples/sec	 accuracy=84.106522	 loss=0.606843	 lr=0.000100
Epoch[084] Batch [0939]/[1203]	Speed: 116.042909 samples/sec	 accuracy=84.118085	 loss=0.606667	 lr=0.000100
Epoch[084] Batch [0959]/[1203]	Speed: 132.333920 samples/sec	 accuracy=84.122396	 loss=0.606504	 lr=0.000100
Epoch[084] Batch [0979]/[1203]	Speed: 136.549846 samples/sec	 accuracy=84.113265	 loss=0.606588	 lr=0.000100
Epoch[084] Batch [0999]/[1203]	Speed: 111.260569 samples/sec	 accuracy=84.125500	 loss=0.606542	 lr=0.000100
Epoch[084] Batch [1019]/[1203]	Speed: 149.504490 samples/sec	 accuracy=84.136275	 loss=0.605786	 lr=0.000100
Epoch[084] Batch [1039]/[1203]	Speed: 118.159564 samples/sec	 accuracy=84.150962	 loss=0.604999	 lr=0.000100
Epoch[084] Batch [1059]/[1203]	Speed: 133.818474 samples/sec	 accuracy=84.133019	 loss=0.605505	 lr=0.000100
Epoch[084] Batch [1079]/[1203]	Speed: 131.363561 samples/sec	 accuracy=84.142593	 loss=0.605199	 lr=0.000100
Epoch[084] Batch [1099]/[1203]	Speed: 111.517116 samples/sec	 accuracy=84.150909	 loss=0.605127	 lr=0.000100
Epoch[084] Batch [1119]/[1203]	Speed: 133.603923 samples/sec	 accuracy=84.139286	 loss=0.605461	 lr=0.000100
Epoch[084] Batch [1139]/[1203]	Speed: 126.804426 samples/sec	 accuracy=84.142544	 loss=0.605205	 lr=0.000100
Epoch[084] Batch [1159]/[1203]	Speed: 123.233343 samples/sec	 accuracy=84.128879	 loss=0.605805	 lr=0.000100
Epoch[084] Batch [1179]/[1203]	Speed: 148.167435 samples/sec	 accuracy=84.152119	 loss=0.605143	 lr=0.000100
Epoch[084] Batch [1199]/[1203]	Speed: 224.636674 samples/sec	 accuracy=84.143333	 loss=0.605671	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.700000 acc-top5=85.125000
Batch [0039]/[0097]: acc-top1=63.537500 acc-top5=84.912500
Batch [0059]/[0097]: acc-top1=65.475000 acc-top5=86.133333
Batch [0079]/[0097]: acc-top1=65.500000 acc-top5=85.918750
[Epoch 084] training: accuracy=84.142976	 loss=0.605581
[Epoch 084] speed: 126 samples/sec	time cost: 2096.842495
[Epoch 084] validation: acc-top1=65.097938 acc-top5=85.757732 loss=1.569952
Epoch[085] Batch [0019]/[1203]	Speed: 69.173619 samples/sec	 accuracy=83.350000	 loss=0.625533	 lr=0.000100
Epoch[085] Batch [0039]/[1203]	Speed: 102.375261 samples/sec	 accuracy=83.725000	 loss=0.609766	 lr=0.000100
Epoch[085] Batch [0059]/[1203]	Speed: 235.807103 samples/sec	 accuracy=84.141667	 loss=0.602718	 lr=0.000100
Epoch[085] Batch [0079]/[1203]	Speed: 100.186130 samples/sec	 accuracy=84.025000	 loss=0.603318	 lr=0.000100
Epoch[085] Batch [0099]/[1203]	Speed: 104.717063 samples/sec	 accuracy=84.130000	 loss=0.604814	 lr=0.000100
Epoch[085] Batch [0119]/[1203]	Speed: 200.564596 samples/sec	 accuracy=84.037500	 loss=0.607402	 lr=0.000100
Epoch[085] Batch [0139]/[1203]	Speed: 105.019737 samples/sec	 accuracy=84.089286	 loss=0.605239	 lr=0.000100
Epoch[085] Batch [0159]/[1203]	Speed: 193.029556 samples/sec	 accuracy=84.100000	 loss=0.604158	 lr=0.000100
Epoch[085] Batch [0179]/[1203]	Speed: 106.286127 samples/sec	 accuracy=84.172222	 loss=0.601384	 lr=0.000100
Epoch[085] Batch [0199]/[1203]	Speed: 108.246838 samples/sec	 accuracy=84.217500	 loss=0.601101	 lr=0.000100
Epoch[085] Batch [0219]/[1203]	Speed: 177.877909 samples/sec	 accuracy=84.225000	 loss=0.602283	 lr=0.000100
Epoch[085] Batch [0239]/[1203]	Speed: 100.803031 samples/sec	 accuracy=84.260417	 loss=0.602093	 lr=0.000100
Epoch[085] Batch [0259]/[1203]	Speed: 136.371320 samples/sec	 accuracy=84.242308	 loss=0.602613	 lr=0.000100
Epoch[085] Batch [0279]/[1203]	Speed: 136.540571 samples/sec	 accuracy=84.219643	 loss=0.603809	 lr=0.000100
Epoch[085] Batch [0299]/[1203]	Speed: 101.332486 samples/sec	 accuracy=84.221667	 loss=0.604355	 lr=0.000100
Epoch[085] Batch [0319]/[1203]	Speed: 186.076254 samples/sec	 accuracy=84.273438	 loss=0.602617	 lr=0.000100
Epoch[085] Batch [0339]/[1203]	Speed: 108.961772 samples/sec	 accuracy=84.242647	 loss=0.603504	 lr=0.000100
Epoch[085] Batch [0359]/[1203]	Speed: 134.588755 samples/sec	 accuracy=84.236111	 loss=0.604376	 lr=0.000100
Epoch[085] Batch [0379]/[1203]	Speed: 116.182655 samples/sec	 accuracy=84.218421	 loss=0.604875	 lr=0.000100
Epoch[085] Batch [0399]/[1203]	Speed: 126.262815 samples/sec	 accuracy=84.268750	 loss=0.602789	 lr=0.000100
Epoch[085] Batch [0419]/[1203]	Speed: 152.495028 samples/sec	 accuracy=84.250000	 loss=0.603458	 lr=0.000100
Epoch[085] Batch [0439]/[1203]	Speed: 111.745731 samples/sec	 accuracy=84.246591	 loss=0.603801	 lr=0.000100
Epoch[085] Batch [0459]/[1203]	Speed: 148.224976 samples/sec	 accuracy=84.251087	 loss=0.603487	 lr=0.000100
Epoch[085] Batch [0479]/[1203]	Speed: 109.292265 samples/sec	 accuracy=84.246875	 loss=0.603173	 lr=0.000100
Epoch[085] Batch [0499]/[1203]	Speed: 118.083997 samples/sec	 accuracy=84.279000	 loss=0.602806	 lr=0.000100
Epoch[085] Batch [0519]/[1203]	Speed: 153.992983 samples/sec	 accuracy=84.274038	 loss=0.603400	 lr=0.000100
Epoch[085] Batch [0539]/[1203]	Speed: 132.135329 samples/sec	 accuracy=84.306481	 loss=0.602635	 lr=0.000100
Epoch[085] Batch [0559]/[1203]	Speed: 107.684403 samples/sec	 accuracy=84.284821	 loss=0.603095	 lr=0.000100
Epoch[085] Batch [0579]/[1203]	Speed: 132.872558 samples/sec	 accuracy=84.226724	 loss=0.604844	 lr=0.000100
Epoch[085] Batch [0599]/[1203]	Speed: 124.336183 samples/sec	 accuracy=84.209167	 loss=0.605370	 lr=0.000100
Epoch[085] Batch [0619]/[1203]	Speed: 151.136914 samples/sec	 accuracy=84.202419	 loss=0.605804	 lr=0.000100
Epoch[085] Batch [0639]/[1203]	Speed: 118.025359 samples/sec	 accuracy=84.221094	 loss=0.605296	 lr=0.000100
Epoch[085] Batch [0659]/[1203]	Speed: 112.998426 samples/sec	 accuracy=84.256061	 loss=0.604424	 lr=0.000100
Epoch[085] Batch [0679]/[1203]	Speed: 151.192667 samples/sec	 accuracy=84.270588	 loss=0.604132	 lr=0.000100
Epoch[085] Batch [0699]/[1203]	Speed: 113.737677 samples/sec	 accuracy=84.279286	 loss=0.604113	 lr=0.000100
Epoch[085] Batch [0719]/[1203]	Speed: 148.720713 samples/sec	 accuracy=84.290972	 loss=0.604113	 lr=0.000100
Epoch[085] Batch [0739]/[1203]	Speed: 126.660745 samples/sec	 accuracy=84.271622	 loss=0.604876	 lr=0.000100
Epoch[085] Batch [0759]/[1203]	Speed: 102.234816 samples/sec	 accuracy=84.250000	 loss=0.605116	 lr=0.000100
Epoch[085] Batch [0779]/[1203]	Speed: 156.398896 samples/sec	 accuracy=84.269231	 loss=0.604251	 lr=0.000100
Epoch[085] Batch [0799]/[1203]	Speed: 114.999472 samples/sec	 accuracy=84.249375	 loss=0.604600	 lr=0.000100
Epoch[085] Batch [0819]/[1203]	Speed: 151.769739 samples/sec	 accuracy=84.215854	 loss=0.605458	 lr=0.000100
Epoch[085] Batch [0839]/[1203]	Speed: 127.624001 samples/sec	 accuracy=84.214881	 loss=0.605701	 lr=0.000100
Epoch[085] Batch [0859]/[1203]	Speed: 109.218432 samples/sec	 accuracy=84.215116	 loss=0.605742	 lr=0.000100
Epoch[085] Batch [0879]/[1203]	Speed: 143.926243 samples/sec	 accuracy=84.218750	 loss=0.605539	 lr=0.000100
Epoch[085] Batch [0899]/[1203]	Speed: 119.664365 samples/sec	 accuracy=84.202222	 loss=0.605885	 lr=0.000100
Epoch[085] Batch [0919]/[1203]	Speed: 113.953879 samples/sec	 accuracy=84.197826	 loss=0.606315	 lr=0.000100
Epoch[085] Batch [0939]/[1203]	Speed: 154.104160 samples/sec	 accuracy=84.225532	 loss=0.605487	 lr=0.000100
Epoch[085] Batch [0959]/[1203]	Speed: 116.391953 samples/sec	 accuracy=84.236458	 loss=0.605122	 lr=0.000100
Epoch[085] Batch [0979]/[1203]	Speed: 129.952717 samples/sec	 accuracy=84.236224	 loss=0.604776	 lr=0.000100
Epoch[085] Batch [0999]/[1203]	Speed: 122.415639 samples/sec	 accuracy=84.213000	 loss=0.605147	 lr=0.000100
Epoch[085] Batch [1019]/[1203]	Speed: 131.228031 samples/sec	 accuracy=84.219118	 loss=0.605115	 lr=0.000100
Epoch[085] Batch [1039]/[1203]	Speed: 140.614601 samples/sec	 accuracy=84.218269	 loss=0.605120	 lr=0.000100
Epoch[085] Batch [1059]/[1203]	Speed: 112.671425 samples/sec	 accuracy=84.204717	 loss=0.605412	 lr=0.000100
Epoch[085] Batch [1079]/[1203]	Speed: 114.709285 samples/sec	 accuracy=84.195370	 loss=0.605538	 lr=0.000100
Epoch[085] Batch [1099]/[1203]	Speed: 146.919906 samples/sec	 accuracy=84.191818	 loss=0.605822	 lr=0.000100
Epoch[085] Batch [1119]/[1203]	Speed: 119.426639 samples/sec	 accuracy=84.180804	 loss=0.606193	 lr=0.000100
Epoch[085] Batch [1139]/[1203]	Speed: 148.466229 samples/sec	 accuracy=84.178947	 loss=0.606125	 lr=0.000100
Epoch[085] Batch [1159]/[1203]	Speed: 112.612105 samples/sec	 accuracy=84.179741	 loss=0.606118	 lr=0.000100
Epoch[085] Batch [1179]/[1203]	Speed: 133.774346 samples/sec	 accuracy=84.179661	 loss=0.605966	 lr=0.000100
Epoch[085] Batch [1199]/[1203]	Speed: 263.032561 samples/sec	 accuracy=84.186667	 loss=0.605629	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.650000 acc-top5=84.900000
Batch [0039]/[0097]: acc-top1=63.287500 acc-top5=84.875000
Batch [0059]/[0097]: acc-top1=65.375000 acc-top5=86.141667
Batch [0079]/[0097]: acc-top1=65.518750 acc-top5=85.956250
[Epoch 085] training: accuracy=84.186201	 loss=0.605618
[Epoch 085] speed: 126 samples/sec	time cost: 2100.054430
[Epoch 085] validation: acc-top1=65.056701 acc-top5=85.835052 loss=1.567483
Epoch[086] Batch [0019]/[1203]	Speed: 70.179009 samples/sec	 accuracy=84.400000	 loss=0.585981	 lr=0.000100
Epoch[086] Batch [0039]/[1203]	Speed: 107.040698 samples/sec	 accuracy=84.450000	 loss=0.592194	 lr=0.000100
Epoch[086] Batch [0059]/[1203]	Speed: 204.856918 samples/sec	 accuracy=84.441667	 loss=0.602836	 lr=0.000100
Epoch[086] Batch [0079]/[1203]	Speed: 94.479181 samples/sec	 accuracy=84.368750	 loss=0.604252	 lr=0.000100
Epoch[086] Batch [0099]/[1203]	Speed: 131.570142 samples/sec	 accuracy=84.365000	 loss=0.608994	 lr=0.000100
Epoch[086] Batch [0119]/[1203]	Speed: 147.144275 samples/sec	 accuracy=84.391667	 loss=0.607570	 lr=0.000100
Epoch[086] Batch [0139]/[1203]	Speed: 109.173879 samples/sec	 accuracy=84.271429	 loss=0.608908	 lr=0.000100
Epoch[086] Batch [0159]/[1203]	Speed: 195.812153 samples/sec	 accuracy=84.393750	 loss=0.606420	 lr=0.000100
Epoch[086] Batch [0179]/[1203]	Speed: 101.629027 samples/sec	 accuracy=84.400000	 loss=0.605419	 lr=0.000100
Epoch[086] Batch [0199]/[1203]	Speed: 119.524234 samples/sec	 accuracy=84.327500	 loss=0.606287	 lr=0.000100
Epoch[086] Batch [0219]/[1203]	Speed: 159.593812 samples/sec	 accuracy=84.375000	 loss=0.602582	 lr=0.000100
Epoch[086] Batch [0239]/[1203]	Speed: 116.424085 samples/sec	 accuracy=84.254167	 loss=0.606155	 lr=0.000100
Epoch[086] Batch [0259]/[1203]	Speed: 127.404335 samples/sec	 accuracy=84.259615	 loss=0.604731	 lr=0.000100
Epoch[086] Batch [0279]/[1203]	Speed: 116.890562 samples/sec	 accuracy=84.253571	 loss=0.605062	 lr=0.000100
Epoch[086] Batch [0299]/[1203]	Speed: 118.934777 samples/sec	 accuracy=84.268333	 loss=0.604373	 lr=0.000100
Epoch[086] Batch [0319]/[1203]	Speed: 155.287184 samples/sec	 accuracy=84.310938	 loss=0.604383	 lr=0.000100
Epoch[086] Batch [0339]/[1203]	Speed: 115.832447 samples/sec	 accuracy=84.316176	 loss=0.604139	 lr=0.000100
Epoch[086] Batch [0359]/[1203]	Speed: 122.236911 samples/sec	 accuracy=84.277778	 loss=0.605597	 lr=0.000100
Epoch[086] Batch [0379]/[1203]	Speed: 126.677997 samples/sec	 accuracy=84.313158	 loss=0.605058	 lr=0.000100
Epoch[086] Batch [0399]/[1203]	Speed: 126.309295 samples/sec	 accuracy=84.267500	 loss=0.605978	 lr=0.000100
Epoch[086] Batch [0419]/[1203]	Speed: 136.050453 samples/sec	 accuracy=84.279762	 loss=0.605869	 lr=0.000100
Epoch[086] Batch [0439]/[1203]	Speed: 125.119881 samples/sec	 accuracy=84.269318	 loss=0.606377	 lr=0.000100
Epoch[086] Batch [0459]/[1203]	Speed: 119.738503 samples/sec	 accuracy=84.281522	 loss=0.605458	 lr=0.000100
Epoch[086] Batch [0479]/[1203]	Speed: 125.764140 samples/sec	 accuracy=84.264583	 loss=0.605097	 lr=0.000100
Epoch[086] Batch [0499]/[1203]	Speed: 141.295932 samples/sec	 accuracy=84.227000	 loss=0.605363	 lr=0.000100
Epoch[086] Batch [0519]/[1203]	Speed: 142.647420 samples/sec	 accuracy=84.234615	 loss=0.604922	 lr=0.000100
Epoch[086] Batch [0539]/[1203]	Speed: 105.711141 samples/sec	 accuracy=84.249074	 loss=0.603912	 lr=0.000100
Epoch[086] Batch [0559]/[1203]	Speed: 135.791674 samples/sec	 accuracy=84.240179	 loss=0.604852	 lr=0.000100
Epoch[086] Batch [0579]/[1203]	Speed: 126.152559 samples/sec	 accuracy=84.268966	 loss=0.604085	 lr=0.000100
Epoch[086] Batch [0599]/[1203]	Speed: 133.590680 samples/sec	 accuracy=84.263333	 loss=0.603949	 lr=0.000100
Epoch[086] Batch [0619]/[1203]	Speed: 127.091721 samples/sec	 accuracy=84.260484	 loss=0.603784	 lr=0.000100
Epoch[086] Batch [0639]/[1203]	Speed: 115.975837 samples/sec	 accuracy=84.235938	 loss=0.604717	 lr=0.000100
Epoch[086] Batch [0659]/[1203]	Speed: 129.846529 samples/sec	 accuracy=84.227273	 loss=0.604579	 lr=0.000100
Epoch[086] Batch [0679]/[1203]	Speed: 138.581175 samples/sec	 accuracy=84.229412	 loss=0.604381	 lr=0.000100
Epoch[086] Batch [0699]/[1203]	Speed: 121.057366 samples/sec	 accuracy=84.242857	 loss=0.604081	 lr=0.000100
Epoch[086] Batch [0719]/[1203]	Speed: 119.301660 samples/sec	 accuracy=84.236806	 loss=0.603985	 lr=0.000100
Epoch[086] Batch [0739]/[1203]	Speed: 130.096870 samples/sec	 accuracy=84.242568	 loss=0.603811	 lr=0.000100
Epoch[086] Batch [0759]/[1203]	Speed: 126.503747 samples/sec	 accuracy=84.252632	 loss=0.603487	 lr=0.000100
Epoch[086] Batch [0779]/[1203]	Speed: 137.340173 samples/sec	 accuracy=84.233974	 loss=0.603653	 lr=0.000100
Epoch[086] Batch [0799]/[1203]	Speed: 130.331635 samples/sec	 accuracy=84.240625	 loss=0.603738	 lr=0.000100
Epoch[086] Batch [0819]/[1203]	Speed: 120.153107 samples/sec	 accuracy=84.221951	 loss=0.604178	 lr=0.000100
Epoch[086] Batch [0839]/[1203]	Speed: 120.337071 samples/sec	 accuracy=84.203571	 loss=0.604457	 lr=0.000100
Epoch[086] Batch [0859]/[1203]	Speed: 133.904838 samples/sec	 accuracy=84.219186	 loss=0.603855	 lr=0.000100
Epoch[086] Batch [0879]/[1203]	Speed: 135.443885 samples/sec	 accuracy=84.222159	 loss=0.604205	 lr=0.000100
Epoch[086] Batch [0899]/[1203]	Speed: 107.238777 samples/sec	 accuracy=84.223333	 loss=0.604059	 lr=0.000100
Epoch[086] Batch [0919]/[1203]	Speed: 159.470005 samples/sec	 accuracy=84.210326	 loss=0.604670	 lr=0.000100
Epoch[086] Batch [0939]/[1203]	Speed: 112.566474 samples/sec	 accuracy=84.212234	 loss=0.604778	 lr=0.000100
Epoch[086] Batch [0959]/[1203]	Speed: 120.862239 samples/sec	 accuracy=84.235417	 loss=0.604077	 lr=0.000100
Epoch[086] Batch [0979]/[1203]	Speed: 158.273366 samples/sec	 accuracy=84.241327	 loss=0.604131	 lr=0.000100
Epoch[086] Batch [0999]/[1203]	Speed: 100.881698 samples/sec	 accuracy=84.257500	 loss=0.603692	 lr=0.000100
Epoch[086] Batch [1019]/[1203]	Speed: 168.690404 samples/sec	 accuracy=84.274510	 loss=0.603311	 lr=0.000100
Epoch[086] Batch [1039]/[1203]	Speed: 118.450832 samples/sec	 accuracy=84.276442	 loss=0.603010	 lr=0.000100
Epoch[086] Batch [1059]/[1203]	Speed: 121.461017 samples/sec	 accuracy=84.265094	 loss=0.603109	 lr=0.000100
Epoch[086] Batch [1079]/[1203]	Speed: 147.937190 samples/sec	 accuracy=84.275463	 loss=0.602624	 lr=0.000100
Epoch[086] Batch [1099]/[1203]	Speed: 101.123222 samples/sec	 accuracy=84.262273	 loss=0.602665	 lr=0.000100
Epoch[086] Batch [1119]/[1203]	Speed: 165.668071 samples/sec	 accuracy=84.270982	 loss=0.602615	 lr=0.000100
Epoch[086] Batch [1139]/[1203]	Speed: 113.380260 samples/sec	 accuracy=84.257895	 loss=0.602769	 lr=0.000100
Epoch[086] Batch [1159]/[1203]	Speed: 111.310120 samples/sec	 accuracy=84.266810	 loss=0.602437	 lr=0.000100
Epoch[086] Batch [1179]/[1203]	Speed: 179.434928 samples/sec	 accuracy=84.263136	 loss=0.602513	 lr=0.000100
Epoch[086] Batch [1199]/[1203]	Speed: 213.315412 samples/sec	 accuracy=84.251667	 loss=0.602858	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.425000 acc-top5=84.525000
Batch [0039]/[0097]: acc-top1=63.150000 acc-top5=84.700000
Batch [0059]/[0097]: acc-top1=65.258333 acc-top5=86.008333
Batch [0079]/[0097]: acc-top1=65.343750 acc-top5=85.800000
[Epoch 086] training: accuracy=84.264339	 loss=0.602546
[Epoch 086] speed: 126 samples/sec	time cost: 2096.688882
[Epoch 086] validation: acc-top1=64.845361 acc-top5=85.685567 loss=1.575820
Epoch[087] Batch [0019]/[1203]	Speed: 69.025744 samples/sec	 accuracy=83.325000	 loss=0.614123	 lr=0.000100
Epoch[087] Batch [0039]/[1203]	Speed: 103.451206 samples/sec	 accuracy=84.375000	 loss=0.595123	 lr=0.000100
Epoch[087] Batch [0059]/[1203]	Speed: 211.846904 samples/sec	 accuracy=84.275000	 loss=0.605307	 lr=0.000100
Epoch[087] Batch [0079]/[1203]	Speed: 99.993332 samples/sec	 accuracy=84.618750	 loss=0.591557	 lr=0.000100
Epoch[087] Batch [0099]/[1203]	Speed: 128.928710 samples/sec	 accuracy=84.370000	 loss=0.600437	 lr=0.000100
Epoch[087] Batch [0119]/[1203]	Speed: 152.446330 samples/sec	 accuracy=84.250000	 loss=0.601773	 lr=0.000100
Epoch[087] Batch [0139]/[1203]	Speed: 106.540953 samples/sec	 accuracy=84.207143	 loss=0.604358	 lr=0.000100
Epoch[087] Batch [0159]/[1203]	Speed: 183.779671 samples/sec	 accuracy=84.156250	 loss=0.605271	 lr=0.000100
Epoch[087] Batch [0179]/[1203]	Speed: 97.046633 samples/sec	 accuracy=84.177778	 loss=0.606633	 lr=0.000100
Epoch[087] Batch [0199]/[1203]	Speed: 125.800593 samples/sec	 accuracy=84.152500	 loss=0.608289	 lr=0.000100
Epoch[087] Batch [0219]/[1203]	Speed: 144.575477 samples/sec	 accuracy=84.286364	 loss=0.604302	 lr=0.000100
Epoch[087] Batch [0239]/[1203]	Speed: 106.155141 samples/sec	 accuracy=84.352083	 loss=0.601871	 lr=0.000100
Epoch[087] Batch [0259]/[1203]	Speed: 196.721866 samples/sec	 accuracy=84.355769	 loss=0.601338	 lr=0.000100
Epoch[087] Batch [0279]/[1203]	Speed: 101.581599 samples/sec	 accuracy=84.353571	 loss=0.601185	 lr=0.000100
Epoch[087] Batch [0299]/[1203]	Speed: 148.613701 samples/sec	 accuracy=84.258333	 loss=0.602376	 lr=0.000100
Epoch[087] Batch [0319]/[1203]	Speed: 121.477404 samples/sec	 accuracy=84.257812	 loss=0.602009	 lr=0.000100
Epoch[087] Batch [0339]/[1203]	Speed: 107.581968 samples/sec	 accuracy=84.276471	 loss=0.601122	 lr=0.000100
Epoch[087] Batch [0359]/[1203]	Speed: 196.371692 samples/sec	 accuracy=84.245833	 loss=0.601724	 lr=0.000100
Epoch[087] Batch [0379]/[1203]	Speed: 99.467165 samples/sec	 accuracy=84.277632	 loss=0.601180	 lr=0.000100
Epoch[087] Batch [0399]/[1203]	Speed: 122.635389 samples/sec	 accuracy=84.256250	 loss=0.601610	 lr=0.000100
Epoch[087] Batch [0419]/[1203]	Speed: 149.712816 samples/sec	 accuracy=84.252381	 loss=0.601569	 lr=0.000100
Epoch[087] Batch [0439]/[1203]	Speed: 115.631149 samples/sec	 accuracy=84.228409	 loss=0.601934	 lr=0.000100
Epoch[087] Batch [0459]/[1203]	Speed: 160.588277 samples/sec	 accuracy=84.245652	 loss=0.601783	 lr=0.000100
Epoch[087] Batch [0479]/[1203]	Speed: 115.104607 samples/sec	 accuracy=84.255208	 loss=0.601662	 lr=0.000100
Epoch[087] Batch [0499]/[1203]	Speed: 109.848323 samples/sec	 accuracy=84.283000	 loss=0.600746	 lr=0.000100
Epoch[087] Batch [0519]/[1203]	Speed: 151.505681 samples/sec	 accuracy=84.231731	 loss=0.601655	 lr=0.000100
Epoch[087] Batch [0539]/[1203]	Speed: 112.897414 samples/sec	 accuracy=84.239815	 loss=0.601363	 lr=0.000100
Epoch[087] Batch [0559]/[1203]	Speed: 137.782375 samples/sec	 accuracy=84.261607	 loss=0.601074	 lr=0.000100
Epoch[087] Batch [0579]/[1203]	Speed: 129.732983 samples/sec	 accuracy=84.243103	 loss=0.602161	 lr=0.000100
Epoch[087] Batch [0599]/[1203]	Speed: 117.891015 samples/sec	 accuracy=84.237500	 loss=0.602370	 lr=0.000100
Epoch[087] Batch [0619]/[1203]	Speed: 151.571727 samples/sec	 accuracy=84.237097	 loss=0.602365	 lr=0.000100
Epoch[087] Batch [0639]/[1203]	Speed: 110.213713 samples/sec	 accuracy=84.220313	 loss=0.602988	 lr=0.000100
Epoch[087] Batch [0659]/[1203]	Speed: 122.931116 samples/sec	 accuracy=84.209848	 loss=0.603119	 lr=0.000100
Epoch[087] Batch [0679]/[1203]	Speed: 140.557045 samples/sec	 accuracy=84.207353	 loss=0.602766	 lr=0.000100
Epoch[087] Batch [0699]/[1203]	Speed: 121.933410 samples/sec	 accuracy=84.230000	 loss=0.602337	 lr=0.000100
Epoch[087] Batch [0719]/[1203]	Speed: 154.478270 samples/sec	 accuracy=84.225694	 loss=0.601845	 lr=0.000100
Epoch[087] Batch [0739]/[1203]	Speed: 112.774341 samples/sec	 accuracy=84.235135	 loss=0.601351	 lr=0.000100
Epoch[087] Batch [0759]/[1203]	Speed: 129.017860 samples/sec	 accuracy=84.215789	 loss=0.602276	 lr=0.000100
Epoch[087] Batch [0779]/[1203]	Speed: 120.338138 samples/sec	 accuracy=84.230128	 loss=0.601954	 lr=0.000100
Epoch[087] Batch [0799]/[1203]	Speed: 146.361426 samples/sec	 accuracy=84.193750	 loss=0.602790	 lr=0.000100
Epoch[087] Batch [0819]/[1203]	Speed: 118.841528 samples/sec	 accuracy=84.201220	 loss=0.602355	 lr=0.000100
Epoch[087] Batch [0839]/[1203]	Speed: 151.157916 samples/sec	 accuracy=84.226190	 loss=0.601985	 lr=0.000100
Epoch[087] Batch [0859]/[1203]	Speed: 100.646597 samples/sec	 accuracy=84.259302	 loss=0.600988	 lr=0.000100
Epoch[087] Batch [0879]/[1203]	Speed: 136.780107 samples/sec	 accuracy=84.278409	 loss=0.600650	 lr=0.000100
Epoch[087] Batch [0899]/[1203]	Speed: 143.063921 samples/sec	 accuracy=84.276667	 loss=0.600613	 lr=0.000100
Epoch[087] Batch [0919]/[1203]	Speed: 119.982082 samples/sec	 accuracy=84.274457	 loss=0.600868	 lr=0.000100
Epoch[087] Batch [0939]/[1203]	Speed: 113.664030 samples/sec	 accuracy=84.269149	 loss=0.600982	 lr=0.000100
Epoch[087] Batch [0959]/[1203]	Speed: 146.175406 samples/sec	 accuracy=84.246875	 loss=0.601511	 lr=0.000100
Epoch[087] Batch [0979]/[1203]	Speed: 109.003529 samples/sec	 accuracy=84.243878	 loss=0.601520	 lr=0.000100
Epoch[087] Batch [0999]/[1203]	Speed: 150.851188 samples/sec	 accuracy=84.230000	 loss=0.601531	 lr=0.000100
Epoch[087] Batch [1019]/[1203]	Speed: 127.520184 samples/sec	 accuracy=84.220588	 loss=0.601717	 lr=0.000100
Epoch[087] Batch [1039]/[1203]	Speed: 117.029718 samples/sec	 accuracy=84.220673	 loss=0.601522	 lr=0.000100
Epoch[087] Batch [1059]/[1203]	Speed: 151.713087 samples/sec	 accuracy=84.218868	 loss=0.601657	 lr=0.000100
Epoch[087] Batch [1079]/[1203]	Speed: 107.498777 samples/sec	 accuracy=84.219444	 loss=0.601641	 lr=0.000100
Epoch[087] Batch [1099]/[1203]	Speed: 147.042483 samples/sec	 accuracy=84.195909	 loss=0.602281	 lr=0.000100
Epoch[087] Batch [1119]/[1203]	Speed: 131.474747 samples/sec	 accuracy=84.181250	 loss=0.602778	 lr=0.000100
Epoch[087] Batch [1139]/[1203]	Speed: 108.946364 samples/sec	 accuracy=84.180702	 loss=0.602730	 lr=0.000100
Epoch[087] Batch [1159]/[1203]	Speed: 150.642588 samples/sec	 accuracy=84.190517	 loss=0.602768	 lr=0.000100
Epoch[087] Batch [1179]/[1203]	Speed: 124.446691 samples/sec	 accuracy=84.178390	 loss=0.603495	 lr=0.000100
Epoch[087] Batch [1199]/[1203]	Speed: 257.695563 samples/sec	 accuracy=84.170833	 loss=0.603738	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.775000 acc-top5=84.800000
Batch [0039]/[0097]: acc-top1=63.300000 acc-top5=84.675000
Batch [0059]/[0097]: acc-top1=65.333333 acc-top5=86.058333
Batch [0079]/[0097]: acc-top1=65.500000 acc-top5=85.956250
[Epoch 087] training: accuracy=84.171654	 loss=0.603684
[Epoch 087] speed: 126 samples/sec	time cost: 2095.797609
[Epoch 087] validation: acc-top1=65.010309 acc-top5=85.809278 loss=1.570297
Epoch[088] Batch [0019]/[1204]	Speed: 69.708508 samples/sec	 accuracy=85.075000	 loss=0.577897	 lr=0.000100
Epoch[088] Batch [0039]/[1204]	Speed: 97.673674 samples/sec	 accuracy=84.725000	 loss=0.591765	 lr=0.000100
Epoch[088] Batch [0059]/[1204]	Speed: 283.591910 samples/sec	 accuracy=84.358333	 loss=0.600830	 lr=0.000100
Epoch[088] Batch [0079]/[1204]	Speed: 95.817665 samples/sec	 accuracy=84.168750	 loss=0.606902	 lr=0.000100
Epoch[088] Batch [0099]/[1204]	Speed: 114.575141 samples/sec	 accuracy=84.225000	 loss=0.607693	 lr=0.000100
Epoch[088] Batch [0119]/[1204]	Speed: 171.561946 samples/sec	 accuracy=84.125000	 loss=0.612466	 lr=0.000100
Epoch[088] Batch [0139]/[1204]	Speed: 100.686909 samples/sec	 accuracy=84.000000	 loss=0.614824	 lr=0.000100
Epoch[088] Batch [0159]/[1204]	Speed: 221.933939 samples/sec	 accuracy=84.056250	 loss=0.613650	 lr=0.000100
Epoch[088] Batch [0179]/[1204]	Speed: 102.171741 samples/sec	 accuracy=84.097222	 loss=0.611504	 lr=0.000100
Epoch[088] Batch [0199]/[1204]	Speed: 107.554656 samples/sec	 accuracy=84.095000	 loss=0.609173	 lr=0.000100
Epoch[088] Batch [0219]/[1204]	Speed: 196.110438 samples/sec	 accuracy=84.088636	 loss=0.609401	 lr=0.000100
Epoch[088] Batch [0239]/[1204]	Speed: 102.723479 samples/sec	 accuracy=84.062500	 loss=0.609309	 lr=0.000100
Epoch[088] Batch [0259]/[1204]	Speed: 146.948044 samples/sec	 accuracy=84.069231	 loss=0.608380	 lr=0.000100
Epoch[088] Batch [0279]/[1204]	Speed: 120.730620 samples/sec	 accuracy=84.160714	 loss=0.605915	 lr=0.000100
Epoch[088] Batch [0299]/[1204]	Speed: 99.628328 samples/sec	 accuracy=84.145000	 loss=0.605078	 lr=0.000100
Epoch[088] Batch [0319]/[1204]	Speed: 237.823109 samples/sec	 accuracy=84.137500	 loss=0.605324	 lr=0.000100
Epoch[088] Batch [0339]/[1204]	Speed: 92.162029 samples/sec	 accuracy=84.135294	 loss=0.605390	 lr=0.000100
Epoch[088] Batch [0359]/[1204]	Speed: 118.305201 samples/sec	 accuracy=84.154167	 loss=0.604392	 lr=0.000100
Epoch[088] Batch [0379]/[1204]	Speed: 166.719619 samples/sec	 accuracy=84.142105	 loss=0.605161	 lr=0.000100
Epoch[088] Batch [0399]/[1204]	Speed: 105.608410 samples/sec	 accuracy=84.198750	 loss=0.603167	 lr=0.000100
Epoch[088] Batch [0419]/[1204]	Speed: 202.003504 samples/sec	 accuracy=84.202381	 loss=0.604035	 lr=0.000100
Epoch[088] Batch [0439]/[1204]	Speed: 100.818949 samples/sec	 accuracy=84.203409	 loss=0.604844	 lr=0.000100
Epoch[088] Batch [0459]/[1204]	Speed: 116.681718 samples/sec	 accuracy=84.200000	 loss=0.604264	 lr=0.000100
Epoch[088] Batch [0479]/[1204]	Speed: 158.549219 samples/sec	 accuracy=84.207292	 loss=0.604038	 lr=0.000100
Epoch[088] Batch [0499]/[1204]	Speed: 99.500223 samples/sec	 accuracy=84.244000	 loss=0.603635	 lr=0.000100
Epoch[088] Batch [0519]/[1204]	Speed: 146.408341 samples/sec	 accuracy=84.266346	 loss=0.603982	 lr=0.000100
Epoch[088] Batch [0539]/[1204]	Speed: 133.832439 samples/sec	 accuracy=84.273148	 loss=0.603691	 lr=0.000100
Epoch[088] Batch [0559]/[1204]	Speed: 114.871311 samples/sec	 accuracy=84.261607	 loss=0.603831	 lr=0.000100
Epoch[088] Batch [0579]/[1204]	Speed: 158.486890 samples/sec	 accuracy=84.275000	 loss=0.602600	 lr=0.000100
Epoch[088] Batch [0599]/[1204]	Speed: 101.808111 samples/sec	 accuracy=84.301667	 loss=0.601945	 lr=0.000100
Epoch[088] Batch [0619]/[1204]	Speed: 122.936209 samples/sec	 accuracy=84.281452	 loss=0.602834	 lr=0.000100
Epoch[088] Batch [0639]/[1204]	Speed: 154.342305 samples/sec	 accuracy=84.286719	 loss=0.602664	 lr=0.000100
Epoch[088] Batch [0659]/[1204]	Speed: 118.407115 samples/sec	 accuracy=84.270455	 loss=0.603748	 lr=0.000100
Epoch[088] Batch [0679]/[1204]	Speed: 139.867521 samples/sec	 accuracy=84.286029	 loss=0.603517	 lr=0.000100
Epoch[088] Batch [0699]/[1204]	Speed: 113.308185 samples/sec	 accuracy=84.290000	 loss=0.602602	 lr=0.000100
Epoch[088] Batch [0719]/[1204]	Speed: 117.415002 samples/sec	 accuracy=84.316667	 loss=0.602130	 lr=0.000100
Epoch[088] Batch [0739]/[1204]	Speed: 169.354717 samples/sec	 accuracy=84.338514	 loss=0.601387	 lr=0.000100
Epoch[088] Batch [0759]/[1204]	Speed: 98.812945 samples/sec	 accuracy=84.325658	 loss=0.601510	 lr=0.000100
Epoch[088] Batch [0779]/[1204]	Speed: 147.358901 samples/sec	 accuracy=84.335897	 loss=0.601323	 lr=0.000100
Epoch[088] Batch [0799]/[1204]	Speed: 131.378681 samples/sec	 accuracy=84.355625	 loss=0.600625	 lr=0.000100
Epoch[088] Batch [0819]/[1204]	Speed: 116.494377 samples/sec	 accuracy=84.368293	 loss=0.600305	 lr=0.000100
Epoch[088] Batch [0839]/[1204]	Speed: 151.135963 samples/sec	 accuracy=84.348214	 loss=0.600436	 lr=0.000100
Epoch[088] Batch [0859]/[1204]	Speed: 112.686043 samples/sec	 accuracy=84.363372	 loss=0.599842	 lr=0.000100
Epoch[088] Batch [0879]/[1204]	Speed: 122.449859 samples/sec	 accuracy=84.350568	 loss=0.600156	 lr=0.000100
Epoch[088] Batch [0899]/[1204]	Speed: 141.644657 samples/sec	 accuracy=84.337778	 loss=0.600374	 lr=0.000100
Epoch[088] Batch [0919]/[1204]	Speed: 117.542085 samples/sec	 accuracy=84.346739	 loss=0.599765	 lr=0.000100
Epoch[088] Batch [0939]/[1204]	Speed: 126.295299 samples/sec	 accuracy=84.338830	 loss=0.600034	 lr=0.000100
Epoch[088] Batch [0959]/[1204]	Speed: 137.689714 samples/sec	 accuracy=84.339062	 loss=0.600008	 lr=0.000100
Epoch[088] Batch [0979]/[1204]	Speed: 116.753824 samples/sec	 accuracy=84.339796	 loss=0.600036	 lr=0.000100
Epoch[088] Batch [0999]/[1204]	Speed: 129.593053 samples/sec	 accuracy=84.341000	 loss=0.600226	 lr=0.000100
Epoch[088] Batch [1019]/[1204]	Speed: 135.610594 samples/sec	 accuracy=84.333824	 loss=0.600686	 lr=0.000100
Epoch[088] Batch [1039]/[1204]	Speed: 114.255001 samples/sec	 accuracy=84.335577	 loss=0.600578	 lr=0.000100
Epoch[088] Batch [1059]/[1204]	Speed: 149.974224 samples/sec	 accuracy=84.332547	 loss=0.600714	 lr=0.000100
Epoch[088] Batch [1079]/[1204]	Speed: 116.504199 samples/sec	 accuracy=84.335648	 loss=0.600333	 lr=0.000100
Epoch[088] Batch [1099]/[1204]	Speed: 114.959884 samples/sec	 accuracy=84.300455	 loss=0.601333	 lr=0.000100
Epoch[088] Batch [1119]/[1204]	Speed: 142.816869 samples/sec	 accuracy=84.296875	 loss=0.601450	 lr=0.000100
Epoch[088] Batch [1139]/[1204]	Speed: 117.645197 samples/sec	 accuracy=84.307018	 loss=0.601206	 lr=0.000100
Epoch[088] Batch [1159]/[1204]	Speed: 124.096513 samples/sec	 accuracy=84.325862	 loss=0.600683	 lr=0.000100
Epoch[088] Batch [1179]/[1204]	Speed: 148.570270 samples/sec	 accuracy=84.327119	 loss=0.600743	 lr=0.000100
Epoch[088] Batch [1199]/[1204]	Speed: 225.464941 samples/sec	 accuracy=84.318333	 loss=0.600939	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.525000 acc-top5=84.925000
Batch [0039]/[0097]: acc-top1=63.150000 acc-top5=84.800000
Batch [0059]/[0097]: acc-top1=65.400000 acc-top5=86.116667
Batch [0079]/[0097]: acc-top1=65.531250 acc-top5=85.962500
[Epoch 088] training: accuracy=84.323505	 loss=0.601021
[Epoch 088] speed: 126 samples/sec	time cost: 2103.144475
[Epoch 088] validation: acc-top1=65.015464 acc-top5=85.850515 loss=1.571930
Epoch[089] Batch [0019]/[1203]	Speed: 69.813071 samples/sec	 accuracy=84.700000	 loss=0.602327	 lr=0.000100
Epoch[089] Batch [0039]/[1203]	Speed: 100.789787 samples/sec	 accuracy=84.425000	 loss=0.601098	 lr=0.000100
Epoch[089] Batch [0059]/[1203]	Speed: 255.192508 samples/sec	 accuracy=84.108333	 loss=0.604847	 lr=0.000100
Epoch[089] Batch [0079]/[1203]	Speed: 90.784892 samples/sec	 accuracy=84.156250	 loss=0.601620	 lr=0.000100
Epoch[089] Batch [0099]/[1203]	Speed: 110.481855 samples/sec	 accuracy=84.150000	 loss=0.604496	 lr=0.000100
Epoch[089] Batch [0119]/[1203]	Speed: 193.628592 samples/sec	 accuracy=84.075000	 loss=0.602785	 lr=0.000100
Epoch[089] Batch [0139]/[1203]	Speed: 103.621425 samples/sec	 accuracy=83.975000	 loss=0.608360	 lr=0.000100
Epoch[089] Batch [0159]/[1203]	Speed: 215.784800 samples/sec	 accuracy=84.065625	 loss=0.606989	 lr=0.000100
Epoch[089] Batch [0179]/[1203]	Speed: 103.993870 samples/sec	 accuracy=84.069444	 loss=0.606340	 lr=0.000100
Epoch[089] Batch [0199]/[1203]	Speed: 108.597622 samples/sec	 accuracy=84.057500	 loss=0.608290	 lr=0.000100
Epoch[089] Batch [0219]/[1203]	Speed: 158.849166 samples/sec	 accuracy=84.150000	 loss=0.606312	 lr=0.000100
Epoch[089] Batch [0239]/[1203]	Speed: 110.442805 samples/sec	 accuracy=84.154167	 loss=0.606382	 lr=0.000100
Epoch[089] Batch [0259]/[1203]	Speed: 167.019586 samples/sec	 accuracy=84.188462	 loss=0.605414	 lr=0.000100
Epoch[089] Batch [0279]/[1203]	Speed: 115.595367 samples/sec	 accuracy=84.225000	 loss=0.605512	 lr=0.000100
Epoch[089] Batch [0299]/[1203]	Speed: 104.318156 samples/sec	 accuracy=84.226667	 loss=0.604742	 lr=0.000100
Epoch[089] Batch [0319]/[1203]	Speed: 161.941996 samples/sec	 accuracy=84.201563	 loss=0.605612	 lr=0.000100
Epoch[089] Batch [0339]/[1203]	Speed: 118.345997 samples/sec	 accuracy=84.150000	 loss=0.607489	 lr=0.000100
Epoch[089] Batch [0359]/[1203]	Speed: 128.578276 samples/sec	 accuracy=84.252778	 loss=0.605346	 lr=0.000100
Epoch[089] Batch [0379]/[1203]	Speed: 120.918524 samples/sec	 accuracy=84.227632	 loss=0.604853	 lr=0.000100
Epoch[089] Batch [0399]/[1203]	Speed: 119.718349 samples/sec	 accuracy=84.242500	 loss=0.604329	 lr=0.000100
Epoch[089] Batch [0419]/[1203]	Speed: 153.334464 samples/sec	 accuracy=84.266667	 loss=0.603903	 lr=0.000100
Epoch[089] Batch [0439]/[1203]	Speed: 131.406460 samples/sec	 accuracy=84.242045	 loss=0.604284	 lr=0.000100
Epoch[089] Batch [0459]/[1203]	Speed: 104.035868 samples/sec	 accuracy=84.254348	 loss=0.603420	 lr=0.000100
Epoch[089] Batch [0479]/[1203]	Speed: 124.587585 samples/sec	 accuracy=84.285417	 loss=0.602795	 lr=0.000100
Epoch[089] Batch [0499]/[1203]	Speed: 146.340575 samples/sec	 accuracy=84.294000	 loss=0.603299	 lr=0.000100
Epoch[089] Batch [0519]/[1203]	Speed: 137.321872 samples/sec	 accuracy=84.304808	 loss=0.603875	 lr=0.000100
Epoch[089] Batch [0539]/[1203]	Speed: 128.686099 samples/sec	 accuracy=84.294444	 loss=0.604396	 lr=0.000100
Epoch[089] Batch [0559]/[1203]	Speed: 104.609841 samples/sec	 accuracy=84.293750	 loss=0.603314	 lr=0.000100
Epoch[089] Batch [0579]/[1203]	Speed: 147.897934 samples/sec	 accuracy=84.298276	 loss=0.603370	 lr=0.000100
Epoch[089] Batch [0599]/[1203]	Speed: 130.787869 samples/sec	 accuracy=84.291667	 loss=0.603200	 lr=0.000100
Epoch[089] Batch [0619]/[1203]	Speed: 135.104694 samples/sec	 accuracy=84.283065	 loss=0.603412	 lr=0.000100
Epoch[089] Batch [0639]/[1203]	Speed: 131.273073 samples/sec	 accuracy=84.292969	 loss=0.603129	 lr=0.000100
Epoch[089] Batch [0659]/[1203]	Speed: 118.828127 samples/sec	 accuracy=84.263636	 loss=0.603867	 lr=0.000100
Epoch[089] Batch [0679]/[1203]	Speed: 110.328776 samples/sec	 accuracy=84.290441	 loss=0.602950	 lr=0.000100
Epoch[089] Batch [0699]/[1203]	Speed: 150.696652 samples/sec	 accuracy=84.290714	 loss=0.602519	 lr=0.000100
Epoch[089] Batch [0719]/[1203]	Speed: 118.481070 samples/sec	 accuracy=84.300694	 loss=0.602501	 lr=0.000100
Epoch[089] Batch [0739]/[1203]	Speed: 126.248807 samples/sec	 accuracy=84.318919	 loss=0.602051	 lr=0.000100
Epoch[089] Batch [0759]/[1203]	Speed: 142.906146 samples/sec	 accuracy=84.314474	 loss=0.601743	 lr=0.000100
Epoch[089] Batch [0779]/[1203]	Speed: 103.076187 samples/sec	 accuracy=84.326282	 loss=0.601241	 lr=0.000100
Epoch[089] Batch [0799]/[1203]	Speed: 150.507036 samples/sec	 accuracy=84.321875	 loss=0.601506	 lr=0.000100
Epoch[089] Batch [0819]/[1203]	Speed: 125.737181 samples/sec	 accuracy=84.318902	 loss=0.601499	 lr=0.000100
Epoch[089] Batch [0839]/[1203]	Speed: 112.104488 samples/sec	 accuracy=84.329762	 loss=0.601324	 lr=0.000100
Epoch[089] Batch [0859]/[1203]	Speed: 145.687510 samples/sec	 accuracy=84.316860	 loss=0.601394	 lr=0.000100
Epoch[089] Batch [0879]/[1203]	Speed: 124.722867 samples/sec	 accuracy=84.305682	 loss=0.601278	 lr=0.000100
Epoch[089] Batch [0899]/[1203]	Speed: 139.681325 samples/sec	 accuracy=84.324444	 loss=0.600333	 lr=0.000100
Epoch[089] Batch [0919]/[1203]	Speed: 116.412264 samples/sec	 accuracy=84.316848	 loss=0.600282	 lr=0.000100
Epoch[089] Batch [0939]/[1203]	Speed: 121.571216 samples/sec	 accuracy=84.329787	 loss=0.599550	 lr=0.000100
Epoch[089] Batch [0959]/[1203]	Speed: 144.016992 samples/sec	 accuracy=84.321354	 loss=0.599515	 lr=0.000100
Epoch[089] Batch [0979]/[1203]	Speed: 123.918200 samples/sec	 accuracy=84.322959	 loss=0.599541	 lr=0.000100
Epoch[089] Batch [0999]/[1203]	Speed: 129.814534 samples/sec	 accuracy=84.324000	 loss=0.599549	 lr=0.000100
Epoch[089] Batch [1019]/[1203]	Speed: 128.792396 samples/sec	 accuracy=84.333824	 loss=0.599187	 lr=0.000100
Epoch[089] Batch [1039]/[1203]	Speed: 111.622687 samples/sec	 accuracy=84.335096	 loss=0.598919	 lr=0.000100
Epoch[089] Batch [1059]/[1203]	Speed: 145.644008 samples/sec	 accuracy=84.342453	 loss=0.598976	 lr=0.000100
Epoch[089] Batch [1079]/[1203]	Speed: 121.200066 samples/sec	 accuracy=84.335648	 loss=0.599768	 lr=0.000100
Epoch[089] Batch [1099]/[1203]	Speed: 127.605492 samples/sec	 accuracy=84.331818	 loss=0.600013	 lr=0.000100
Epoch[089] Batch [1119]/[1203]	Speed: 134.355655 samples/sec	 accuracy=84.325446	 loss=0.600027	 lr=0.000100
Epoch[089] Batch [1139]/[1203]	Speed: 110.930285 samples/sec	 accuracy=84.332456	 loss=0.600087	 lr=0.000100
Epoch[089] Batch [1159]/[1203]	Speed: 142.943493 samples/sec	 accuracy=84.328017	 loss=0.600220	 lr=0.000100
Epoch[089] Batch [1179]/[1203]	Speed: 134.927395 samples/sec	 accuracy=84.320763	 loss=0.600268	 lr=0.000100
Epoch[089] Batch [1199]/[1203]	Speed: 240.192283 samples/sec	 accuracy=84.309167	 loss=0.600631	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.775000 acc-top5=84.950000
Batch [0039]/[0097]: acc-top1=63.250000 acc-top5=84.737500
Batch [0059]/[0097]: acc-top1=65.375000 acc-top5=86.083333
Batch [0079]/[0097]: acc-top1=65.568750 acc-top5=85.943750
[Epoch 089] training: accuracy=84.309643	 loss=0.600679
[Epoch 089] speed: 126 samples/sec	time cost: 2096.512575
[Epoch 089] validation: acc-top1=65.108247 acc-top5=85.860825 loss=1.569323
Epoch[090] Batch [0019]/[1203]	Speed: 69.094951 samples/sec	 accuracy=84.625000	 loss=0.586901	 lr=0.000100
Epoch[090] Batch [0039]/[1203]	Speed: 104.846799 samples/sec	 accuracy=85.000000	 loss=0.581346	 lr=0.000100
Epoch[090] Batch [0059]/[1203]	Speed: 218.298640 samples/sec	 accuracy=84.641667	 loss=0.590936	 lr=0.000100
Epoch[090] Batch [0079]/[1203]	Speed: 95.011928 samples/sec	 accuracy=84.668750	 loss=0.593349	 lr=0.000100
Epoch[090] Batch [0099]/[1203]	Speed: 116.926152 samples/sec	 accuracy=84.655000	 loss=0.590302	 lr=0.000100
Epoch[090] Batch [0119]/[1203]	Speed: 180.743690 samples/sec	 accuracy=84.570833	 loss=0.594810	 lr=0.000100
Epoch[090] Batch [0139]/[1203]	Speed: 103.736916 samples/sec	 accuracy=84.464286	 loss=0.597043	 lr=0.000100
Epoch[090] Batch [0159]/[1203]	Speed: 210.044363 samples/sec	 accuracy=84.378125	 loss=0.598279	 lr=0.000100
Epoch[090] Batch [0179]/[1203]	Speed: 98.562790 samples/sec	 accuracy=84.355556	 loss=0.601092	 lr=0.000100
Epoch[090] Batch [0199]/[1203]	Speed: 108.391472 samples/sec	 accuracy=84.272500	 loss=0.602870	 lr=0.000100
Epoch[090] Batch [0219]/[1203]	Speed: 177.590174 samples/sec	 accuracy=84.215909	 loss=0.602858	 lr=0.000100
Epoch[090] Batch [0239]/[1203]	Speed: 110.235802 samples/sec	 accuracy=84.154167	 loss=0.602070	 lr=0.000100
Epoch[090] Batch [0259]/[1203]	Speed: 116.851000 samples/sec	 accuracy=84.167308	 loss=0.602510	 lr=0.000100
Epoch[090] Batch [0279]/[1203]	Speed: 141.493375 samples/sec	 accuracy=84.110714	 loss=0.603258	 lr=0.000100
Epoch[090] Batch [0299]/[1203]	Speed: 114.262908 samples/sec	 accuracy=84.078333	 loss=0.604364	 lr=0.000100
Epoch[090] Batch [0319]/[1203]	Speed: 169.826789 samples/sec	 accuracy=84.134375	 loss=0.603751	 lr=0.000100
Epoch[090] Batch [0339]/[1203]	Speed: 113.901797 samples/sec	 accuracy=84.094118	 loss=0.605172	 lr=0.000100
Epoch[090] Batch [0359]/[1203]	Speed: 123.907396 samples/sec	 accuracy=84.108333	 loss=0.604717	 lr=0.000100
Epoch[090] Batch [0379]/[1203]	Speed: 127.111563 samples/sec	 accuracy=84.142105	 loss=0.604667	 lr=0.000100
Epoch[090] Batch [0399]/[1203]	Speed: 120.800052 samples/sec	 accuracy=84.116250	 loss=0.604919	 lr=0.000100
Epoch[090] Batch [0419]/[1203]	Speed: 125.053990 samples/sec	 accuracy=84.108333	 loss=0.605483	 lr=0.000100
Epoch[090] Batch [0439]/[1203]	Speed: 137.562136 samples/sec	 accuracy=84.110227	 loss=0.604888	 lr=0.000100
Epoch[090] Batch [0459]/[1203]	Speed: 116.091683 samples/sec	 accuracy=84.108696	 loss=0.604734	 lr=0.000100
Epoch[090] Batch [0479]/[1203]	Speed: 136.966640 samples/sec	 accuracy=84.070833	 loss=0.605449	 lr=0.000100
Epoch[090] Batch [0499]/[1203]	Speed: 133.088558 samples/sec	 accuracy=84.078000	 loss=0.604530	 lr=0.000100
Epoch[090] Batch [0519]/[1203]	Speed: 121.355265 samples/sec	 accuracy=84.048077	 loss=0.604625	 lr=0.000100
Epoch[090] Batch [0539]/[1203]	Speed: 108.791568 samples/sec	 accuracy=84.079630	 loss=0.603838	 lr=0.000100
Epoch[090] Batch [0559]/[1203]	Speed: 162.780645 samples/sec	 accuracy=84.067857	 loss=0.604492	 lr=0.000100
Epoch[090] Batch [0579]/[1203]	Speed: 118.227908 samples/sec	 accuracy=84.106034	 loss=0.603256	 lr=0.000100
Epoch[090] Batch [0599]/[1203]	Speed: 139.918456 samples/sec	 accuracy=84.110000	 loss=0.603149	 lr=0.000100
Epoch[090] Batch [0619]/[1203]	Speed: 117.875113 samples/sec	 accuracy=84.126613	 loss=0.603203	 lr=0.000100
Epoch[090] Batch [0639]/[1203]	Speed: 114.917092 samples/sec	 accuracy=84.143750	 loss=0.602693	 lr=0.000100
Epoch[090] Batch [0659]/[1203]	Speed: 153.556942 samples/sec	 accuracy=84.137879	 loss=0.603441	 lr=0.000100
Epoch[090] Batch [0679]/[1203]	Speed: 116.359172 samples/sec	 accuracy=84.127206	 loss=0.603719	 lr=0.000100
Epoch[090] Batch [0699]/[1203]	Speed: 122.969220 samples/sec	 accuracy=84.134286	 loss=0.603165	 lr=0.000100
Epoch[090] Batch [0719]/[1203]	Speed: 134.086601 samples/sec	 accuracy=84.138889	 loss=0.603195	 lr=0.000100
Epoch[090] Batch [0739]/[1203]	Speed: 115.184837 samples/sec	 accuracy=84.161486	 loss=0.602839	 lr=0.000100
Epoch[090] Batch [0759]/[1203]	Speed: 156.409776 samples/sec	 accuracy=84.157895	 loss=0.603143	 lr=0.000100
Epoch[090] Batch [0779]/[1203]	Speed: 112.010941 samples/sec	 accuracy=84.198077	 loss=0.602041	 lr=0.000100
Epoch[090] Batch [0799]/[1203]	Speed: 124.082453 samples/sec	 accuracy=84.216875	 loss=0.601553	 lr=0.000100
Epoch[090] Batch [0819]/[1203]	Speed: 146.591966 samples/sec	 accuracy=84.209146	 loss=0.601718	 lr=0.000100
Epoch[090] Batch [0839]/[1203]	Speed: 101.439167 samples/sec	 accuracy=84.207143	 loss=0.602041	 lr=0.000100
Epoch[090] Batch [0859]/[1203]	Speed: 138.929028 samples/sec	 accuracy=84.198256	 loss=0.602008	 lr=0.000100
Epoch[090] Batch [0879]/[1203]	Speed: 141.846328 samples/sec	 accuracy=84.175568	 loss=0.603157	 lr=0.000100
Epoch[090] Batch [0899]/[1203]	Speed: 107.609108 samples/sec	 accuracy=84.165556	 loss=0.603080	 lr=0.000100
Epoch[090] Batch [0919]/[1203]	Speed: 160.510110 samples/sec	 accuracy=84.168478	 loss=0.602786	 lr=0.000100
Epoch[090] Batch [0939]/[1203]	Speed: 107.384128 samples/sec	 accuracy=84.174468	 loss=0.602671	 lr=0.000100
Epoch[090] Batch [0959]/[1203]	Speed: 125.956059 samples/sec	 accuracy=84.160417	 loss=0.602991	 lr=0.000100
Epoch[090] Batch [0979]/[1203]	Speed: 144.376475 samples/sec	 accuracy=84.185204	 loss=0.602205	 lr=0.000100
Epoch[090] Batch [0999]/[1203]	Speed: 113.667359 samples/sec	 accuracy=84.205000	 loss=0.602243	 lr=0.000100
Epoch[090] Batch [1019]/[1203]	Speed: 158.864928 samples/sec	 accuracy=84.190196	 loss=0.602575	 lr=0.000100
Epoch[090] Batch [1039]/[1203]	Speed: 109.873105 samples/sec	 accuracy=84.192788	 loss=0.602256	 lr=0.000100
Epoch[090] Batch [1059]/[1203]	Speed: 123.591089 samples/sec	 accuracy=84.202830	 loss=0.602213	 lr=0.000100
Epoch[090] Batch [1079]/[1203]	Speed: 149.767018 samples/sec	 accuracy=84.212037	 loss=0.602195	 lr=0.000100
Epoch[090] Batch [1099]/[1203]	Speed: 99.262123 samples/sec	 accuracy=84.220909	 loss=0.602027	 lr=0.000100
Epoch[090] Batch [1119]/[1203]	Speed: 185.651533 samples/sec	 accuracy=84.216071	 loss=0.602010	 lr=0.000100
Epoch[090] Batch [1139]/[1203]	Speed: 100.565600 samples/sec	 accuracy=84.214035	 loss=0.602281	 lr=0.000100
Epoch[090] Batch [1159]/[1203]	Speed: 135.996132 samples/sec	 accuracy=84.216379	 loss=0.602487	 lr=0.000100
Epoch[090] Batch [1179]/[1203]	Speed: 145.967276 samples/sec	 accuracy=84.221610	 loss=0.602776	 lr=0.000100
Epoch[090] Batch [1199]/[1203]	Speed: 216.284378 samples/sec	 accuracy=84.230833	 loss=0.602682	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.750000 acc-top5=85.025000
Batch [0039]/[0097]: acc-top1=63.525000 acc-top5=85.112500
Batch [0059]/[0097]: acc-top1=65.500000 acc-top5=86.300000
Batch [0079]/[0097]: acc-top1=65.568750 acc-top5=86.075000
[Epoch 090] training: accuracy=84.228595	 loss=0.602640
[Epoch 090] speed: 126 samples/sec	time cost: 2106.810573
[Epoch 090] validation: acc-top1=65.103093 acc-top5=85.948454 loss=1.567756
Epoch[091] Batch [0019]/[1203]	Speed: 69.253991 samples/sec	 accuracy=83.375000	 loss=0.619032	 lr=0.000100
Epoch[091] Batch [0039]/[1203]	Speed: 99.794669 samples/sec	 accuracy=83.550000	 loss=0.616616	 lr=0.000100
Epoch[091] Batch [0059]/[1203]	Speed: 242.912299 samples/sec	 accuracy=83.625000	 loss=0.616361	 lr=0.000100
Epoch[091] Batch [0079]/[1203]	Speed: 97.891846 samples/sec	 accuracy=83.950000	 loss=0.606390	 lr=0.000100
Epoch[091] Batch [0099]/[1203]	Speed: 137.804286 samples/sec	 accuracy=84.110000	 loss=0.606122	 lr=0.000100
Epoch[091] Batch [0119]/[1203]	Speed: 134.151136 samples/sec	 accuracy=84.070833	 loss=0.605067	 lr=0.000100
Epoch[091] Batch [0139]/[1203]	Speed: 102.904199 samples/sec	 accuracy=84.017857	 loss=0.605010	 lr=0.000100
Epoch[091] Batch [0159]/[1203]	Speed: 200.940156 samples/sec	 accuracy=84.096875	 loss=0.603471	 lr=0.000100
Epoch[091] Batch [0179]/[1203]	Speed: 101.447619 samples/sec	 accuracy=84.166667	 loss=0.601847	 lr=0.000100
Epoch[091] Batch [0199]/[1203]	Speed: 136.849409 samples/sec	 accuracy=84.160000	 loss=0.603530	 lr=0.000100
Epoch[091] Batch [0219]/[1203]	Speed: 134.378040 samples/sec	 accuracy=84.163636	 loss=0.602698	 lr=0.000100
Epoch[091] Batch [0239]/[1203]	Speed: 106.515375 samples/sec	 accuracy=84.185417	 loss=0.603695	 lr=0.000100
Epoch[091] Batch [0259]/[1203]	Speed: 187.192862 samples/sec	 accuracy=84.186538	 loss=0.603338	 lr=0.000100
Epoch[091] Batch [0279]/[1203]	Speed: 102.258923 samples/sec	 accuracy=84.216071	 loss=0.604454	 lr=0.000100
Epoch[091] Batch [0299]/[1203]	Speed: 131.637300 samples/sec	 accuracy=84.230000	 loss=0.603560	 lr=0.000100
Epoch[091] Batch [0319]/[1203]	Speed: 133.491437 samples/sec	 accuracy=84.245312	 loss=0.602969	 lr=0.000100
Epoch[091] Batch [0339]/[1203]	Speed: 115.104005 samples/sec	 accuracy=84.283824	 loss=0.601680	 lr=0.000100
Epoch[091] Batch [0359]/[1203]	Speed: 161.863042 samples/sec	 accuracy=84.350000	 loss=0.600104	 lr=0.000100
Epoch[091] Batch [0379]/[1203]	Speed: 107.078811 samples/sec	 accuracy=84.314474	 loss=0.600895	 lr=0.000100
Epoch[091] Batch [0399]/[1203]	Speed: 139.319517 samples/sec	 accuracy=84.298750	 loss=0.601727	 lr=0.000100
Epoch[091] Batch [0419]/[1203]	Speed: 128.736991 samples/sec	 accuracy=84.279762	 loss=0.602997	 lr=0.000100
Epoch[091] Batch [0439]/[1203]	Speed: 114.789222 samples/sec	 accuracy=84.269318	 loss=0.603093	 lr=0.000100
Epoch[091] Batch [0459]/[1203]	Speed: 175.022665 samples/sec	 accuracy=84.265217	 loss=0.603246	 lr=0.000100
Epoch[091] Batch [0479]/[1203]	Speed: 97.266474 samples/sec	 accuracy=84.237500	 loss=0.603965	 lr=0.000100
Epoch[091] Batch [0499]/[1203]	Speed: 148.790644 samples/sec	 accuracy=84.241000	 loss=0.603885	 lr=0.000100
Epoch[091] Batch [0519]/[1203]	Speed: 120.433318 samples/sec	 accuracy=84.281731	 loss=0.602878	 lr=0.000100
Epoch[091] Batch [0539]/[1203]	Speed: 112.428412 samples/sec	 accuracy=84.262037	 loss=0.603345	 lr=0.000100
Epoch[091] Batch [0559]/[1203]	Speed: 190.379879 samples/sec	 accuracy=84.226786	 loss=0.603805	 lr=0.000100
Epoch[091] Batch [0579]/[1203]	Speed: 96.522727 samples/sec	 accuracy=84.225000	 loss=0.603353	 lr=0.000100
Epoch[091] Batch [0599]/[1203]	Speed: 141.001598 samples/sec	 accuracy=84.234167	 loss=0.602622	 lr=0.000100
Epoch[091] Batch [0619]/[1203]	Speed: 131.543860 samples/sec	 accuracy=84.241129	 loss=0.602182	 lr=0.000100
Epoch[091] Batch [0639]/[1203]	Speed: 112.970474 samples/sec	 accuracy=84.257031	 loss=0.601593	 lr=0.000100
Epoch[091] Batch [0659]/[1203]	Speed: 175.058046 samples/sec	 accuracy=84.295455	 loss=0.600555	 lr=0.000100
Epoch[091] Batch [0679]/[1203]	Speed: 103.826975 samples/sec	 accuracy=84.316176	 loss=0.600435	 lr=0.000100
Epoch[091] Batch [0699]/[1203]	Speed: 119.283997 samples/sec	 accuracy=84.287143	 loss=0.601460	 lr=0.000100
Epoch[091] Batch [0719]/[1203]	Speed: 140.868602 samples/sec	 accuracy=84.284028	 loss=0.601734	 lr=0.000100
Epoch[091] Batch [0739]/[1203]	Speed: 117.228724 samples/sec	 accuracy=84.269595	 loss=0.601989	 lr=0.000100
Epoch[091] Batch [0759]/[1203]	Speed: 134.859983 samples/sec	 accuracy=84.264474	 loss=0.602061	 lr=0.000100
Epoch[091] Batch [0779]/[1203]	Speed: 122.411489 samples/sec	 accuracy=84.289744	 loss=0.601408	 lr=0.000100
Epoch[091] Batch [0799]/[1203]	Speed: 132.693580 samples/sec	 accuracy=84.280625	 loss=0.601554	 lr=0.000100
Epoch[091] Batch [0819]/[1203]	Speed: 131.857937 samples/sec	 accuracy=84.246951	 loss=0.602427	 lr=0.000100
Epoch[091] Batch [0839]/[1203]	Speed: 126.136349 samples/sec	 accuracy=84.250595	 loss=0.602089	 lr=0.000100
Epoch[091] Batch [0859]/[1203]	Speed: 119.266852 samples/sec	 accuracy=84.241279	 loss=0.602682	 lr=0.000100
Epoch[091] Batch [0879]/[1203]	Speed: 125.685795 samples/sec	 accuracy=84.233523	 loss=0.602759	 lr=0.000100
Epoch[091] Batch [0899]/[1203]	Speed: 130.251325 samples/sec	 accuracy=84.237778	 loss=0.602601	 lr=0.000100
Epoch[091] Batch [0919]/[1203]	Speed: 138.774797 samples/sec	 accuracy=84.235870	 loss=0.602977	 lr=0.000100
Epoch[091] Batch [0939]/[1203]	Speed: 115.510635 samples/sec	 accuracy=84.238298	 loss=0.602724	 lr=0.000100
Epoch[091] Batch [0959]/[1203]	Speed: 123.582767 samples/sec	 accuracy=84.236979	 loss=0.602824	 lr=0.000100
Epoch[091] Batch [0979]/[1203]	Speed: 131.732315 samples/sec	 accuracy=84.223980	 loss=0.603184	 lr=0.000100
Epoch[091] Batch [0999]/[1203]	Speed: 122.236116 samples/sec	 accuracy=84.225000	 loss=0.603158	 lr=0.000100
Epoch[091] Batch [1019]/[1203]	Speed: 140.896085 samples/sec	 accuracy=84.230882	 loss=0.602903	 lr=0.000100
Epoch[091] Batch [1039]/[1203]	Speed: 113.381809 samples/sec	 accuracy=84.233173	 loss=0.602773	 lr=0.000100
Epoch[091] Batch [1059]/[1203]	Speed: 128.982278 samples/sec	 accuracy=84.206604	 loss=0.603368	 lr=0.000100
Epoch[091] Batch [1079]/[1203]	Speed: 125.964110 samples/sec	 accuracy=84.200926	 loss=0.603515	 lr=0.000100
Epoch[091] Batch [1099]/[1203]	Speed: 125.676230 samples/sec	 accuracy=84.218636	 loss=0.603046	 lr=0.000100
Epoch[091] Batch [1119]/[1203]	Speed: 139.511565 samples/sec	 accuracy=84.226786	 loss=0.602772	 lr=0.000100
Epoch[091] Batch [1139]/[1203]	Speed: 109.047995 samples/sec	 accuracy=84.216228	 loss=0.603437	 lr=0.000100
Epoch[091] Batch [1159]/[1203]	Speed: 146.180648 samples/sec	 accuracy=84.220690	 loss=0.603279	 lr=0.000100
Epoch[091] Batch [1179]/[1203]	Speed: 121.269593 samples/sec	 accuracy=84.224576	 loss=0.603308	 lr=0.000100
Epoch[091] Batch [1199]/[1203]	Speed: 303.255122 samples/sec	 accuracy=84.220833	 loss=0.603348	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.600000 acc-top5=85.125000
Batch [0039]/[0097]: acc-top1=63.275000 acc-top5=84.950000
Batch [0059]/[0097]: acc-top1=65.375000 acc-top5=86.208333
Batch [0079]/[0097]: acc-top1=65.562500 acc-top5=86.050000
[Epoch 091] training: accuracy=84.211554	 loss=0.603534
[Epoch 091] speed: 126 samples/sec	time cost: 2100.613563
[Epoch 091] validation: acc-top1=65.103093 acc-top5=85.902062 loss=1.569140
Epoch[092] Batch [0019]/[1203]	Speed: 70.878963 samples/sec	 accuracy=83.850000	 loss=0.615093	 lr=0.000100
Epoch[092] Batch [0039]/[1203]	Speed: 98.286004 samples/sec	 accuracy=83.925000	 loss=0.614131	 lr=0.000100
Epoch[092] Batch [0059]/[1203]	Speed: 266.970760 samples/sec	 accuracy=84.325000	 loss=0.607319	 lr=0.000100
Epoch[092] Batch [0079]/[1203]	Speed: 95.441122 samples/sec	 accuracy=84.281250	 loss=0.607204	 lr=0.000100
Epoch[092] Batch [0099]/[1203]	Speed: 110.172810 samples/sec	 accuracy=84.505000	 loss=0.599489	 lr=0.000100
Epoch[092] Batch [0119]/[1203]	Speed: 183.163965 samples/sec	 accuracy=84.395833	 loss=0.599847	 lr=0.000100
Epoch[092] Batch [0139]/[1203]	Speed: 104.799155 samples/sec	 accuracy=84.364286	 loss=0.599479	 lr=0.000100
Epoch[092] Batch [0159]/[1203]	Speed: 201.225475 samples/sec	 accuracy=84.309375	 loss=0.601834	 lr=0.000100
Epoch[092] Batch [0179]/[1203]	Speed: 96.723325 samples/sec	 accuracy=84.266667	 loss=0.603877	 lr=0.000100
Epoch[092] Batch [0199]/[1203]	Speed: 118.918808 samples/sec	 accuracy=84.310000	 loss=0.601599	 lr=0.000100
Epoch[092] Batch [0219]/[1203]	Speed: 174.005307 samples/sec	 accuracy=84.234091	 loss=0.603023	 lr=0.000100
Epoch[092] Batch [0239]/[1203]	Speed: 111.460496 samples/sec	 accuracy=84.181250	 loss=0.603921	 lr=0.000100
Epoch[092] Batch [0259]/[1203]	Speed: 150.454263 samples/sec	 accuracy=84.188462	 loss=0.603781	 lr=0.000100
Epoch[092] Batch [0279]/[1203]	Speed: 117.088865 samples/sec	 accuracy=84.248214	 loss=0.602334	 lr=0.000100
Epoch[092] Batch [0299]/[1203]	Speed: 109.199718 samples/sec	 accuracy=84.256667	 loss=0.601935	 lr=0.000100
Epoch[092] Batch [0319]/[1203]	Speed: 167.516431 samples/sec	 accuracy=84.240625	 loss=0.602971	 lr=0.000100
Epoch[092] Batch [0339]/[1203]	Speed: 107.616490 samples/sec	 accuracy=84.210294	 loss=0.603364	 lr=0.000100
Epoch[092] Batch [0359]/[1203]	Speed: 123.448124 samples/sec	 accuracy=84.234722	 loss=0.602391	 lr=0.000100
Epoch[092] Batch [0379]/[1203]	Speed: 142.789392 samples/sec	 accuracy=84.209211	 loss=0.603144	 lr=0.000100
Epoch[092] Batch [0399]/[1203]	Speed: 119.247486 samples/sec	 accuracy=84.211250	 loss=0.603416	 lr=0.000100
Epoch[092] Batch [0419]/[1203]	Speed: 165.862358 samples/sec	 accuracy=84.260714	 loss=0.603296	 lr=0.000100
Epoch[092] Batch [0439]/[1203]	Speed: 106.718031 samples/sec	 accuracy=84.251136	 loss=0.603815	 lr=0.000100
Epoch[092] Batch [0459]/[1203]	Speed: 118.390966 samples/sec	 accuracy=84.236957	 loss=0.603500	 lr=0.000100
Epoch[092] Batch [0479]/[1203]	Speed: 142.229372 samples/sec	 accuracy=84.259375	 loss=0.603183	 lr=0.000100
Epoch[092] Batch [0499]/[1203]	Speed: 119.076489 samples/sec	 accuracy=84.264000	 loss=0.602509	 lr=0.000100
Epoch[092] Batch [0519]/[1203]	Speed: 120.443788 samples/sec	 accuracy=84.310577	 loss=0.600953	 lr=0.000100
Epoch[092] Batch [0539]/[1203]	Speed: 123.422539 samples/sec	 accuracy=84.328704	 loss=0.600456	 lr=0.000100
Epoch[092] Batch [0559]/[1203]	Speed: 137.322831 samples/sec	 accuracy=84.325000	 loss=0.600888	 lr=0.000100
Epoch[092] Batch [0579]/[1203]	Speed: 140.857735 samples/sec	 accuracy=84.317241	 loss=0.600779	 lr=0.000100
Epoch[092] Batch [0599]/[1203]	Speed: 109.173181 samples/sec	 accuracy=84.329167	 loss=0.600557	 lr=0.000100
Epoch[092] Batch [0619]/[1203]	Speed: 143.130285 samples/sec	 accuracy=84.331452	 loss=0.600011	 lr=0.000100
Epoch[092] Batch [0639]/[1203]	Speed: 118.258066 samples/sec	 accuracy=84.354687	 loss=0.599450	 lr=0.000100
Epoch[092] Batch [0659]/[1203]	Speed: 129.369855 samples/sec	 accuracy=84.369697	 loss=0.598837	 lr=0.000100
Epoch[092] Batch [0679]/[1203]	Speed: 128.419909 samples/sec	 accuracy=84.369118	 loss=0.598841	 lr=0.000100
Epoch[092] Batch [0699]/[1203]	Speed: 137.067519 samples/sec	 accuracy=84.380000	 loss=0.598678	 lr=0.000100
Epoch[092] Batch [0719]/[1203]	Speed: 117.791569 samples/sec	 accuracy=84.369444	 loss=0.598516	 lr=0.000100
Epoch[092] Batch [0739]/[1203]	Speed: 109.354103 samples/sec	 accuracy=84.337162	 loss=0.599148	 lr=0.000100
Epoch[092] Batch [0759]/[1203]	Speed: 143.962099 samples/sec	 accuracy=84.324342	 loss=0.599527	 lr=0.000100
Epoch[092] Batch [0779]/[1203]	Speed: 135.947935 samples/sec	 accuracy=84.312821	 loss=0.600034	 lr=0.000100
Epoch[092] Batch [0799]/[1203]	Speed: 130.985782 samples/sec	 accuracy=84.315000	 loss=0.599953	 lr=0.000100
Epoch[092] Batch [0819]/[1203]	Speed: 107.371643 samples/sec	 accuracy=84.283537	 loss=0.600462	 lr=0.000100
Epoch[092] Batch [0839]/[1203]	Speed: 128.122913 samples/sec	 accuracy=84.295238	 loss=0.599880	 lr=0.000100
Epoch[092] Batch [0859]/[1203]	Speed: 150.004820 samples/sec	 accuracy=84.328488	 loss=0.598781	 lr=0.000100
Epoch[092] Batch [0879]/[1203]	Speed: 109.064920 samples/sec	 accuracy=84.342045	 loss=0.598497	 lr=0.000100
Epoch[092] Batch [0899]/[1203]	Speed: 136.118809 samples/sec	 accuracy=84.326667	 loss=0.598825	 lr=0.000100
Epoch[092] Batch [0919]/[1203]	Speed: 130.537097 samples/sec	 accuracy=84.338587	 loss=0.598523	 lr=0.000100
Epoch[092] Batch [0939]/[1203]	Speed: 121.262515 samples/sec	 accuracy=84.328723	 loss=0.598401	 lr=0.000100
Epoch[092] Batch [0959]/[1203]	Speed: 149.280062 samples/sec	 accuracy=84.314583	 loss=0.598715	 lr=0.000100
Epoch[092] Batch [0979]/[1203]	Speed: 117.261973 samples/sec	 accuracy=84.319388	 loss=0.598853	 lr=0.000100
Epoch[092] Batch [0999]/[1203]	Speed: 125.997425 samples/sec	 accuracy=84.315000	 loss=0.599020	 lr=0.000100
Epoch[092] Batch [1019]/[1203]	Speed: 124.006129 samples/sec	 accuracy=84.313235	 loss=0.598990	 lr=0.000100
Epoch[092] Batch [1039]/[1203]	Speed: 130.846285 samples/sec	 accuracy=84.307692	 loss=0.598977	 lr=0.000100
Epoch[092] Batch [1059]/[1203]	Speed: 128.522565 samples/sec	 accuracy=84.316509	 loss=0.599006	 lr=0.000100
Epoch[092] Batch [1079]/[1203]	Speed: 133.298999 samples/sec	 accuracy=84.288426	 loss=0.599660	 lr=0.000100
Epoch[092] Batch [1099]/[1203]	Speed: 113.378642 samples/sec	 accuracy=84.278182	 loss=0.600102	 lr=0.000100
Epoch[092] Batch [1119]/[1203]	Speed: 134.230224 samples/sec	 accuracy=84.261607	 loss=0.600274	 lr=0.000100
Epoch[092] Batch [1139]/[1203]	Speed: 130.413184 samples/sec	 accuracy=84.260526	 loss=0.600538	 lr=0.000100
Epoch[092] Batch [1159]/[1203]	Speed: 124.373823 samples/sec	 accuracy=84.252155	 loss=0.600580	 lr=0.000100
Epoch[092] Batch [1179]/[1203]	Speed: 133.747260 samples/sec	 accuracy=84.258475	 loss=0.600617	 lr=0.000100
Epoch[092] Batch [1199]/[1203]	Speed: 250.358174 samples/sec	 accuracy=84.250000	 loss=0.600582	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.750000 acc-top5=84.725000
Batch [0039]/[0097]: acc-top1=63.400000 acc-top5=84.725000
Batch [0059]/[0097]: acc-top1=65.508333 acc-top5=86.058333
Batch [0079]/[0097]: acc-top1=65.612500 acc-top5=85.906250
[Epoch 092] training: accuracy=84.249792	 loss=0.600637
[Epoch 092] speed: 126 samples/sec	time cost: 2096.584971
[Epoch 092] validation: acc-top1=65.097938 acc-top5=85.793814 loss=1.565290
Epoch[093] Batch [0019]/[1203]	Speed: 69.667880 samples/sec	 accuracy=83.850000	 loss=0.603394	 lr=0.000100
Epoch[093] Batch [0039]/[1203]	Speed: 97.772471 samples/sec	 accuracy=83.825000	 loss=0.608624	 lr=0.000100
Epoch[093] Batch [0059]/[1203]	Speed: 244.825649 samples/sec	 accuracy=84.358333	 loss=0.595535	 lr=0.000100
Epoch[093] Batch [0079]/[1203]	Speed: 97.346720 samples/sec	 accuracy=84.375000	 loss=0.591186	 lr=0.000100
Epoch[093] Batch [0099]/[1203]	Speed: 114.075824 samples/sec	 accuracy=84.255000	 loss=0.593610	 lr=0.000100
Epoch[093] Batch [0119]/[1203]	Speed: 185.888607 samples/sec	 accuracy=84.266667	 loss=0.594189	 lr=0.000100
Epoch[093] Batch [0139]/[1203]	Speed: 102.216340 samples/sec	 accuracy=84.285714	 loss=0.595817	 lr=0.000100
Epoch[093] Batch [0159]/[1203]	Speed: 228.875217 samples/sec	 accuracy=84.371875	 loss=0.593152	 lr=0.000100
Epoch[093] Batch [0179]/[1203]	Speed: 87.563184 samples/sec	 accuracy=84.297222	 loss=0.595902	 lr=0.000100
Epoch[093] Batch [0199]/[1203]	Speed: 121.376748 samples/sec	 accuracy=84.262500	 loss=0.598629	 lr=0.000100
Epoch[093] Batch [0219]/[1203]	Speed: 152.115473 samples/sec	 accuracy=84.356818	 loss=0.595205	 lr=0.000100
Epoch[093] Batch [0239]/[1203]	Speed: 123.363901 samples/sec	 accuracy=84.302083	 loss=0.597271	 lr=0.000100
Epoch[093] Batch [0259]/[1203]	Speed: 154.034157 samples/sec	 accuracy=84.251923	 loss=0.599272	 lr=0.000100
Epoch[093] Batch [0279]/[1203]	Speed: 110.124338 samples/sec	 accuracy=84.258929	 loss=0.599326	 lr=0.000100
Epoch[093] Batch [0299]/[1203]	Speed: 121.437502 samples/sec	 accuracy=84.278333	 loss=0.600332	 lr=0.000100
Epoch[093] Batch [0319]/[1203]	Speed: 142.105173 samples/sec	 accuracy=84.243750	 loss=0.599666	 lr=0.000100
Epoch[093] Batch [0339]/[1203]	Speed: 116.061041 samples/sec	 accuracy=84.266176	 loss=0.599503	 lr=0.000100
Epoch[093] Batch [0359]/[1203]	Speed: 116.431358 samples/sec	 accuracy=84.277778	 loss=0.598862	 lr=0.000100
Epoch[093] Batch [0379]/[1203]	Speed: 150.659300 samples/sec	 accuracy=84.334211	 loss=0.596820	 lr=0.000100
Epoch[093] Batch [0399]/[1203]	Speed: 112.822829 samples/sec	 accuracy=84.407500	 loss=0.595342	 lr=0.000100
Epoch[093] Batch [0419]/[1203]	Speed: 165.563020 samples/sec	 accuracy=84.367857	 loss=0.596818	 lr=0.000100
Epoch[093] Batch [0439]/[1203]	Speed: 114.874642 samples/sec	 accuracy=84.329545	 loss=0.597583	 lr=0.000100
Epoch[093] Batch [0459]/[1203]	Speed: 108.086532 samples/sec	 accuracy=84.319565	 loss=0.596981	 lr=0.000100
Epoch[093] Batch [0479]/[1203]	Speed: 140.046611 samples/sec	 accuracy=84.323958	 loss=0.597731	 lr=0.000100
Epoch[093] Batch [0499]/[1203]	Speed: 130.966474 samples/sec	 accuracy=84.341000	 loss=0.597696	 lr=0.000100
Epoch[093] Batch [0519]/[1203]	Speed: 121.809945 samples/sec	 accuracy=84.350000	 loss=0.597915	 lr=0.000100
Epoch[093] Batch [0539]/[1203]	Speed: 141.111804 samples/sec	 accuracy=84.350926	 loss=0.598058	 lr=0.000100
Epoch[093] Batch [0559]/[1203]	Speed: 109.210316 samples/sec	 accuracy=84.375893	 loss=0.597636	 lr=0.000100
Epoch[093] Batch [0579]/[1203]	Speed: 130.431452 samples/sec	 accuracy=84.381034	 loss=0.597811	 lr=0.000100
Epoch[093] Batch [0599]/[1203]	Speed: 141.333941 samples/sec	 accuracy=84.385833	 loss=0.597274	 lr=0.000100
Epoch[093] Batch [0619]/[1203]	Speed: 131.506596 samples/sec	 accuracy=84.408871	 loss=0.596820	 lr=0.000100
Epoch[093] Batch [0639]/[1203]	Speed: 138.666065 samples/sec	 accuracy=84.367969	 loss=0.598640	 lr=0.000100
Epoch[093] Batch [0659]/[1203]	Speed: 109.224756 samples/sec	 accuracy=84.365152	 loss=0.598636	 lr=0.000100
Epoch[093] Batch [0679]/[1203]	Speed: 128.561265 samples/sec	 accuracy=84.391912	 loss=0.597563	 lr=0.000100
Epoch[093] Batch [0699]/[1203]	Speed: 136.820869 samples/sec	 accuracy=84.387143	 loss=0.597566	 lr=0.000100
Epoch[093] Batch [0719]/[1203]	Speed: 127.221882 samples/sec	 accuracy=84.388889	 loss=0.597457	 lr=0.000100
Epoch[093] Batch [0739]/[1203]	Speed: 114.011784 samples/sec	 accuracy=84.376351	 loss=0.597752	 lr=0.000100
Epoch[093] Batch [0759]/[1203]	Speed: 129.774780 samples/sec	 accuracy=84.355263	 loss=0.598296	 lr=0.000100
Epoch[093] Batch [0779]/[1203]	Speed: 125.043326 samples/sec	 accuracy=84.343590	 loss=0.598446	 lr=0.000100
Epoch[093] Batch [0799]/[1203]	Speed: 147.355900 samples/sec	 accuracy=84.350000	 loss=0.598604	 lr=0.000100
Epoch[093] Batch [0819]/[1203]	Speed: 119.857308 samples/sec	 accuracy=84.337195	 loss=0.599124	 lr=0.000100
Epoch[093] Batch [0839]/[1203]	Speed: 117.774379 samples/sec	 accuracy=84.341667	 loss=0.599275	 lr=0.000100
Epoch[093] Batch [0859]/[1203]	Speed: 128.496512 samples/sec	 accuracy=84.310465	 loss=0.600224	 lr=0.000100
Epoch[093] Batch [0879]/[1203]	Speed: 128.217496 samples/sec	 accuracy=84.295455	 loss=0.600625	 lr=0.000100
Epoch[093] Batch [0899]/[1203]	Speed: 146.854967 samples/sec	 accuracy=84.279444	 loss=0.600960	 lr=0.000100
Epoch[093] Batch [0919]/[1203]	Speed: 107.226867 samples/sec	 accuracy=84.286413	 loss=0.600668	 lr=0.000100
Epoch[093] Batch [0939]/[1203]	Speed: 120.404625 samples/sec	 accuracy=84.284043	 loss=0.600503	 lr=0.000100
Epoch[093] Batch [0959]/[1203]	Speed: 155.521599 samples/sec	 accuracy=84.309375	 loss=0.599551	 lr=0.000100
Epoch[093] Batch [0979]/[1203]	Speed: 120.812559 samples/sec	 accuracy=84.301020	 loss=0.599881	 lr=0.000100
Epoch[093] Batch [0999]/[1203]	Speed: 137.514298 samples/sec	 accuracy=84.306000	 loss=0.599676	 lr=0.000100
Epoch[093] Batch [1019]/[1203]	Speed: 118.590864 samples/sec	 accuracy=84.296078	 loss=0.599825	 lr=0.000100
Epoch[093] Batch [1039]/[1203]	Speed: 111.534857 samples/sec	 accuracy=84.306250	 loss=0.599711	 lr=0.000100
Epoch[093] Batch [1059]/[1203]	Speed: 144.202815 samples/sec	 accuracy=84.287736	 loss=0.600225	 lr=0.000100
Epoch[093] Batch [1079]/[1203]	Speed: 118.159585 samples/sec	 accuracy=84.284259	 loss=0.600398	 lr=0.000100
Epoch[093] Batch [1099]/[1203]	Speed: 133.077848 samples/sec	 accuracy=84.277727	 loss=0.600097	 lr=0.000100
Epoch[093] Batch [1119]/[1203]	Speed: 123.995526 samples/sec	 accuracy=84.247321	 loss=0.600910	 lr=0.000100
Epoch[093] Batch [1139]/[1203]	Speed: 134.457410 samples/sec	 accuracy=84.231140	 loss=0.601105	 lr=0.000100
Epoch[093] Batch [1159]/[1203]	Speed: 132.187398 samples/sec	 accuracy=84.235776	 loss=0.601224	 lr=0.000100
Epoch[093] Batch [1179]/[1203]	Speed: 124.035676 samples/sec	 accuracy=84.241949	 loss=0.601068	 lr=0.000100
Epoch[093] Batch [1199]/[1203]	Speed: 247.724817 samples/sec	 accuracy=84.251250	 loss=0.600547	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.800000 acc-top5=85.025000
Batch [0039]/[0097]: acc-top1=63.475000 acc-top5=84.862500
Batch [0059]/[0097]: acc-top1=65.500000 acc-top5=86.116667
Batch [0079]/[0097]: acc-top1=65.618750 acc-top5=86.025000
[Epoch 093] training: accuracy=84.246883	 loss=0.600829
[Epoch 093] speed: 126 samples/sec	time cost: 2096.300542
[Epoch 093] validation: acc-top1=65.123711 acc-top5=85.865979 loss=1.567565
Epoch[094] Batch [0019]/[1203]	Speed: 70.214601 samples/sec	 accuracy=84.000000	 loss=0.593424	 lr=0.000100
Epoch[094] Batch [0039]/[1203]	Speed: 100.680591 samples/sec	 accuracy=84.137500	 loss=0.585696	 lr=0.000100
Epoch[094] Batch [0059]/[1203]	Speed: 271.532693 samples/sec	 accuracy=84.200000	 loss=0.585303	 lr=0.000100
Epoch[094] Batch [0079]/[1203]	Speed: 94.846143 samples/sec	 accuracy=84.325000	 loss=0.588241	 lr=0.000100
Epoch[094] Batch [0099]/[1203]	Speed: 105.790883 samples/sec	 accuracy=84.310000	 loss=0.590670	 lr=0.000100
Epoch[094] Batch [0119]/[1203]	Speed: 199.312304 samples/sec	 accuracy=84.337500	 loss=0.590116	 lr=0.000100
Epoch[094] Batch [0139]/[1203]	Speed: 100.490555 samples/sec	 accuracy=84.410714	 loss=0.590478	 lr=0.000100
Epoch[094] Batch [0159]/[1203]	Speed: 233.045994 samples/sec	 accuracy=84.290625	 loss=0.596016	 lr=0.000100
Epoch[094] Batch [0179]/[1203]	Speed: 94.682924 samples/sec	 accuracy=84.350000	 loss=0.596095	 lr=0.000100
Epoch[094] Batch [0199]/[1203]	Speed: 106.648271 samples/sec	 accuracy=84.362500	 loss=0.595855	 lr=0.000100
Epoch[094] Batch [0219]/[1203]	Speed: 213.567284 samples/sec	 accuracy=84.370455	 loss=0.595784	 lr=0.000100
Epoch[094] Batch [0239]/[1203]	Speed: 96.934608 samples/sec	 accuracy=84.352083	 loss=0.596463	 lr=0.000100
Epoch[094] Batch [0259]/[1203]	Speed: 132.377164 samples/sec	 accuracy=84.282692	 loss=0.597140	 lr=0.000100
Epoch[094] Batch [0279]/[1203]	Speed: 141.112099 samples/sec	 accuracy=84.278571	 loss=0.597847	 lr=0.000100
Epoch[094] Batch [0299]/[1203]	Speed: 117.648041 samples/sec	 accuracy=84.288333	 loss=0.597631	 lr=0.000100
Epoch[094] Batch [0319]/[1203]	Speed: 153.867531 samples/sec	 accuracy=84.337500	 loss=0.596480	 lr=0.000100
Epoch[094] Batch [0339]/[1203]	Speed: 111.959379 samples/sec	 accuracy=84.316176	 loss=0.597466	 lr=0.000100
Epoch[094] Batch [0359]/[1203]	Speed: 118.727534 samples/sec	 accuracy=84.412500	 loss=0.595459	 lr=0.000100
Epoch[094] Batch [0379]/[1203]	Speed: 145.544060 samples/sec	 accuracy=84.451316	 loss=0.593931	 lr=0.000100
Epoch[094] Batch [0399]/[1203]	Speed: 112.573294 samples/sec	 accuracy=84.418750	 loss=0.594703	 lr=0.000100
Epoch[094] Batch [0419]/[1203]	Speed: 142.816543 samples/sec	 accuracy=84.405952	 loss=0.595131	 lr=0.000100
Epoch[094] Batch [0439]/[1203]	Speed: 130.874219 samples/sec	 accuracy=84.385227	 loss=0.596722	 lr=0.000100
Epoch[094] Batch [0459]/[1203]	Speed: 112.408427 samples/sec	 accuracy=84.360870	 loss=0.597316	 lr=0.000100
Epoch[094] Batch [0479]/[1203]	Speed: 146.001242 samples/sec	 accuracy=84.425000	 loss=0.595961	 lr=0.000100
Epoch[094] Batch [0499]/[1203]	Speed: 113.023038 samples/sec	 accuracy=84.414000	 loss=0.596052	 lr=0.000100
Epoch[094] Batch [0519]/[1203]	Speed: 136.262247 samples/sec	 accuracy=84.363462	 loss=0.597299	 lr=0.000100
Epoch[094] Batch [0539]/[1203]	Speed: 124.863467 samples/sec	 accuracy=84.346296	 loss=0.597604	 lr=0.000100
Epoch[094] Batch [0559]/[1203]	Speed: 121.895054 samples/sec	 accuracy=84.363393	 loss=0.596503	 lr=0.000100
Epoch[094] Batch [0579]/[1203]	Speed: 153.369332 samples/sec	 accuracy=84.331034	 loss=0.597324	 lr=0.000100
Epoch[094] Batch [0599]/[1203]	Speed: 103.801919 samples/sec	 accuracy=84.332500	 loss=0.597102	 lr=0.000100
Epoch[094] Batch [0619]/[1203]	Speed: 145.147217 samples/sec	 accuracy=84.312097	 loss=0.597093	 lr=0.000100
Epoch[094] Batch [0639]/[1203]	Speed: 125.162919 samples/sec	 accuracy=84.296094	 loss=0.598129	 lr=0.000100
Epoch[094] Batch [0659]/[1203]	Speed: 120.635833 samples/sec	 accuracy=84.316667	 loss=0.597233	 lr=0.000100
Epoch[094] Batch [0679]/[1203]	Speed: 158.100720 samples/sec	 accuracy=84.333824	 loss=0.597290	 lr=0.000100
Epoch[094] Batch [0699]/[1203]	Speed: 102.479713 samples/sec	 accuracy=84.310000	 loss=0.597898	 lr=0.000100
Epoch[094] Batch [0719]/[1203]	Speed: 133.941086 samples/sec	 accuracy=84.317361	 loss=0.597687	 lr=0.000100
Epoch[094] Batch [0739]/[1203]	Speed: 142.471695 samples/sec	 accuracy=84.312838	 loss=0.597842	 lr=0.000100
Epoch[094] Batch [0759]/[1203]	Speed: 114.358980 samples/sec	 accuracy=84.313158	 loss=0.597271	 lr=0.000100
Epoch[094] Batch [0779]/[1203]	Speed: 120.135708 samples/sec	 accuracy=84.308333	 loss=0.597554	 lr=0.000100
Epoch[094] Batch [0799]/[1203]	Speed: 125.477687 samples/sec	 accuracy=84.305625	 loss=0.597397	 lr=0.000100
Epoch[094] Batch [0819]/[1203]	Speed: 147.704510 samples/sec	 accuracy=84.298171	 loss=0.597550	 lr=0.000100
Epoch[094] Batch [0839]/[1203]	Speed: 128.578445 samples/sec	 accuracy=84.279762	 loss=0.597771	 lr=0.000100
Epoch[094] Batch [0859]/[1203]	Speed: 108.641295 samples/sec	 accuracy=84.290116	 loss=0.597737	 lr=0.000100
Epoch[094] Batch [0879]/[1203]	Speed: 142.941039 samples/sec	 accuracy=84.288068	 loss=0.597769	 lr=0.000100
Epoch[094] Batch [0899]/[1203]	Speed: 120.433549 samples/sec	 accuracy=84.281111	 loss=0.598167	 lr=0.000100
Epoch[094] Batch [0919]/[1203]	Speed: 142.770993 samples/sec	 accuracy=84.288587	 loss=0.598014	 lr=0.000100
Epoch[094] Batch [0939]/[1203]	Speed: 119.066803 samples/sec	 accuracy=84.316489	 loss=0.597142	 lr=0.000100
Epoch[094] Batch [0959]/[1203]	Speed: 121.479415 samples/sec	 accuracy=84.312500	 loss=0.597326	 lr=0.000100
Epoch[094] Batch [0979]/[1203]	Speed: 123.596692 samples/sec	 accuracy=84.294898	 loss=0.597700	 lr=0.000100
Epoch[094] Batch [0999]/[1203]	Speed: 124.658740 samples/sec	 accuracy=84.292000	 loss=0.597792	 lr=0.000100
Epoch[094] Batch [1019]/[1203]	Speed: 127.406567 samples/sec	 accuracy=84.288235	 loss=0.597612	 lr=0.000100
Epoch[094] Batch [1039]/[1203]	Speed: 148.080551 samples/sec	 accuracy=84.290385	 loss=0.597435	 lr=0.000100
Epoch[094] Batch [1059]/[1203]	Speed: 119.089307 samples/sec	 accuracy=84.312736	 loss=0.597084	 lr=0.000100
Epoch[094] Batch [1079]/[1203]	Speed: 115.486319 samples/sec	 accuracy=84.301389	 loss=0.597374	 lr=0.000100
Epoch[094] Batch [1099]/[1203]	Speed: 138.609601 samples/sec	 accuracy=84.296818	 loss=0.597560	 lr=0.000100
Epoch[094] Batch [1119]/[1203]	Speed: 117.832471 samples/sec	 accuracy=84.304464	 loss=0.597489	 lr=0.000100
Epoch[094] Batch [1139]/[1203]	Speed: 143.285794 samples/sec	 accuracy=84.300439	 loss=0.597612	 lr=0.000100
Epoch[094] Batch [1159]/[1203]	Speed: 112.341430 samples/sec	 accuracy=84.293534	 loss=0.597752	 lr=0.000100
Epoch[094] Batch [1179]/[1203]	Speed: 132.808268 samples/sec	 accuracy=84.288559	 loss=0.598168	 lr=0.000100
Epoch[094] Batch [1199]/[1203]	Speed: 283.584005 samples/sec	 accuracy=84.278750	 loss=0.598425	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.650000 acc-top5=84.975000
Batch [0039]/[0097]: acc-top1=63.337500 acc-top5=84.637500
Batch [0059]/[0097]: acc-top1=65.350000 acc-top5=86.041667
Batch [0079]/[0097]: acc-top1=65.406250 acc-top5=85.831250
[Epoch 094] training: accuracy=84.276808	 loss=0.598549
[Epoch 094] speed: 126 samples/sec	time cost: 2096.639373
[Epoch 094] validation: acc-top1=64.865979 acc-top5=85.711340 loss=1.574309
Epoch[095] Batch [0019]/[1203]	Speed: 70.623378 samples/sec	 accuracy=84.925000	 loss=0.578457	 lr=0.000100
Epoch[095] Batch [0039]/[1203]	Speed: 104.984959 samples/sec	 accuracy=85.075000	 loss=0.573989	 lr=0.000100
Epoch[095] Batch [0059]/[1203]	Speed: 210.434681 samples/sec	 accuracy=84.691667	 loss=0.583563	 lr=0.000100
Epoch[095] Batch [0079]/[1203]	Speed: 96.240647 samples/sec	 accuracy=84.600000	 loss=0.582021	 lr=0.000100
Epoch[095] Batch [0099]/[1203]	Speed: 117.827451 samples/sec	 accuracy=84.580000	 loss=0.587878	 lr=0.000100
Epoch[095] Batch [0119]/[1203]	Speed: 162.060166 samples/sec	 accuracy=84.529167	 loss=0.587099	 lr=0.000100
Epoch[095] Batch [0139]/[1203]	Speed: 117.561754 samples/sec	 accuracy=84.517857	 loss=0.588915	 lr=0.000100
Epoch[095] Batch [0159]/[1203]	Speed: 176.413771 samples/sec	 accuracy=84.506250	 loss=0.586664	 lr=0.000100
Epoch[095] Batch [0179]/[1203]	Speed: 95.113020 samples/sec	 accuracy=84.430556	 loss=0.587291	 lr=0.000100
Epoch[095] Batch [0199]/[1203]	Speed: 145.932794 samples/sec	 accuracy=84.417500	 loss=0.588311	 lr=0.000100
Epoch[095] Batch [0219]/[1203]	Speed: 121.788059 samples/sec	 accuracy=84.425000	 loss=0.587704	 lr=0.000100
Epoch[095] Batch [0239]/[1203]	Speed: 104.921848 samples/sec	 accuracy=84.408333	 loss=0.588287	 lr=0.000100
Epoch[095] Batch [0259]/[1203]	Speed: 187.090083 samples/sec	 accuracy=84.398077	 loss=0.590039	 lr=0.000100
Epoch[095] Batch [0279]/[1203]	Speed: 106.354943 samples/sec	 accuracy=84.382143	 loss=0.591214	 lr=0.000100
Epoch[095] Batch [0299]/[1203]	Speed: 133.901982 samples/sec	 accuracy=84.431667	 loss=0.590372	 lr=0.000100
Epoch[095] Batch [0319]/[1203]	Speed: 136.169383 samples/sec	 accuracy=84.431250	 loss=0.590098	 lr=0.000100
Epoch[095] Batch [0339]/[1203]	Speed: 109.731311 samples/sec	 accuracy=84.423529	 loss=0.590095	 lr=0.000100
Epoch[095] Batch [0359]/[1203]	Speed: 172.793660 samples/sec	 accuracy=84.434722	 loss=0.589622	 lr=0.000100
Epoch[095] Batch [0379]/[1203]	Speed: 111.200160 samples/sec	 accuracy=84.442105	 loss=0.590173	 lr=0.000100
Epoch[095] Batch [0399]/[1203]	Speed: 102.661391 samples/sec	 accuracy=84.425000	 loss=0.590171	 lr=0.000100
Epoch[095] Batch [0419]/[1203]	Speed: 171.115059 samples/sec	 accuracy=84.421429	 loss=0.590638	 lr=0.000100
Epoch[095] Batch [0439]/[1203]	Speed: 109.900955 samples/sec	 accuracy=84.412500	 loss=0.590677	 lr=0.000100
Epoch[095] Batch [0459]/[1203]	Speed: 149.306343 samples/sec	 accuracy=84.389130	 loss=0.591139	 lr=0.000100
Epoch[095] Batch [0479]/[1203]	Speed: 123.233475 samples/sec	 accuracy=84.394792	 loss=0.591540	 lr=0.000100
Epoch[095] Batch [0499]/[1203]	Speed: 111.336569 samples/sec	 accuracy=84.397000	 loss=0.591122	 lr=0.000100
Epoch[095] Batch [0519]/[1203]	Speed: 158.568998 samples/sec	 accuracy=84.387500	 loss=0.591689	 lr=0.000100
Epoch[095] Batch [0539]/[1203]	Speed: 108.573136 samples/sec	 accuracy=84.397222	 loss=0.591343	 lr=0.000100
Epoch[095] Batch [0559]/[1203]	Speed: 150.902202 samples/sec	 accuracy=84.437500	 loss=0.590902	 lr=0.000100
Epoch[095] Batch [0579]/[1203]	Speed: 117.244948 samples/sec	 accuracy=84.449138	 loss=0.590420	 lr=0.000100
Epoch[095] Batch [0599]/[1203]	Speed: 122.918356 samples/sec	 accuracy=84.454167	 loss=0.590077	 lr=0.000100
Epoch[095] Batch [0619]/[1203]	Speed: 142.368420 samples/sec	 accuracy=84.479032	 loss=0.589271	 lr=0.000100
Epoch[095] Batch [0639]/[1203]	Speed: 119.164364 samples/sec	 accuracy=84.440625	 loss=0.590856	 lr=0.000100
Epoch[095] Batch [0659]/[1203]	Speed: 118.537231 samples/sec	 accuracy=84.449242	 loss=0.590260	 lr=0.000100
Epoch[095] Batch [0679]/[1203]	Speed: 148.691832 samples/sec	 accuracy=84.474265	 loss=0.590035	 lr=0.000100
Epoch[095] Batch [0699]/[1203]	Speed: 102.313743 samples/sec	 accuracy=84.457857	 loss=0.590530	 lr=0.000100
Epoch[095] Batch [0719]/[1203]	Speed: 152.706710 samples/sec	 accuracy=84.432639	 loss=0.591097	 lr=0.000100
Epoch[095] Batch [0739]/[1203]	Speed: 131.379299 samples/sec	 accuracy=84.441216	 loss=0.590778	 lr=0.000100
Epoch[095] Batch [0759]/[1203]	Speed: 108.474785 samples/sec	 accuracy=84.425000	 loss=0.591380	 lr=0.000100
Epoch[095] Batch [0779]/[1203]	Speed: 178.218685 samples/sec	 accuracy=84.410256	 loss=0.592461	 lr=0.000100
Epoch[095] Batch [0799]/[1203]	Speed: 100.258829 samples/sec	 accuracy=84.428125	 loss=0.592312	 lr=0.000100
Epoch[095] Batch [0819]/[1203]	Speed: 138.708520 samples/sec	 accuracy=84.396341	 loss=0.593106	 lr=0.000100
Epoch[095] Batch [0839]/[1203]	Speed: 135.016968 samples/sec	 accuracy=84.401190	 loss=0.592976	 lr=0.000100
Epoch[095] Batch [0859]/[1203]	Speed: 104.723128 samples/sec	 accuracy=84.411047	 loss=0.592919	 lr=0.000100
Epoch[095] Batch [0879]/[1203]	Speed: 178.337606 samples/sec	 accuracy=84.408523	 loss=0.592844	 lr=0.000100
Epoch[095] Batch [0899]/[1203]	Speed: 103.370982 samples/sec	 accuracy=84.397778	 loss=0.593118	 lr=0.000100
Epoch[095] Batch [0919]/[1203]	Speed: 134.611985 samples/sec	 accuracy=84.403261	 loss=0.593221	 lr=0.000100
Epoch[095] Batch [0939]/[1203]	Speed: 134.285379 samples/sec	 accuracy=84.391489	 loss=0.593544	 lr=0.000100
Epoch[095] Batch [0959]/[1203]	Speed: 121.696524 samples/sec	 accuracy=84.375000	 loss=0.593891	 lr=0.000100
Epoch[095] Batch [0979]/[1203]	Speed: 120.862641 samples/sec	 accuracy=84.345918	 loss=0.594535	 lr=0.000100
Epoch[095] Batch [0999]/[1203]	Speed: 133.190720 samples/sec	 accuracy=84.349000	 loss=0.594822	 lr=0.000100
Epoch[095] Batch [1019]/[1203]	Speed: 120.945202 samples/sec	 accuracy=84.335784	 loss=0.594905	 lr=0.000100
Epoch[095] Batch [1039]/[1203]	Speed: 146.609640 samples/sec	 accuracy=84.336058	 loss=0.595226	 lr=0.000100
Epoch[095] Batch [1059]/[1203]	Speed: 117.281754 samples/sec	 accuracy=84.327358	 loss=0.595484	 lr=0.000100
Epoch[095] Batch [1079]/[1203]	Speed: 129.598619 samples/sec	 accuracy=84.311111	 loss=0.595691	 lr=0.000100
Epoch[095] Batch [1099]/[1203]	Speed: 119.683308 samples/sec	 accuracy=84.304545	 loss=0.595996	 lr=0.000100
Epoch[095] Batch [1119]/[1203]	Speed: 140.522872 samples/sec	 accuracy=84.299107	 loss=0.596322	 lr=0.000100
Epoch[095] Batch [1139]/[1203]	Speed: 134.644207 samples/sec	 accuracy=84.286842	 loss=0.597115	 lr=0.000100
Epoch[095] Batch [1159]/[1203]	Speed: 110.944413 samples/sec	 accuracy=84.294397	 loss=0.597043	 lr=0.000100
Epoch[095] Batch [1179]/[1203]	Speed: 144.142877 samples/sec	 accuracy=84.303390	 loss=0.596918	 lr=0.000100
Epoch[095] Batch [1199]/[1203]	Speed: 235.136811 samples/sec	 accuracy=84.310417	 loss=0.597112	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.500000 acc-top5=84.800000
Batch [0039]/[0097]: acc-top1=63.175000 acc-top5=84.812500
Batch [0059]/[0097]: acc-top1=65.358333 acc-top5=86.050000
Batch [0079]/[0097]: acc-top1=65.462500 acc-top5=85.937500
[Epoch 095] training: accuracy=84.314630	 loss=0.597078
[Epoch 095] speed: 126 samples/sec	time cost: 2096.131224
[Epoch 095] validation: acc-top1=64.917526 acc-top5=85.824742 loss=1.570635
Epoch[096] Batch [0019]/[1203]	Speed: 68.864358 samples/sec	 accuracy=84.425000	 loss=0.588730	 lr=0.000100
Epoch[096] Batch [0039]/[1203]	Speed: 104.111483 samples/sec	 accuracy=84.162500	 loss=0.593953	 lr=0.000100
Epoch[096] Batch [0059]/[1203]	Speed: 218.128560 samples/sec	 accuracy=84.691667	 loss=0.584133	 lr=0.000100
Epoch[096] Batch [0079]/[1203]	Speed: 103.121670 samples/sec	 accuracy=84.631250	 loss=0.585559	 lr=0.000100
Epoch[096] Batch [0099]/[1203]	Speed: 102.627027 samples/sec	 accuracy=84.600000	 loss=0.585383	 lr=0.000100
Epoch[096] Batch [0119]/[1203]	Speed: 194.236388 samples/sec	 accuracy=84.658333	 loss=0.586071	 lr=0.000100
Epoch[096] Batch [0139]/[1203]	Speed: 99.181840 samples/sec	 accuracy=84.578571	 loss=0.587407	 lr=0.000100
Epoch[096] Batch [0159]/[1203]	Speed: 197.585386 samples/sec	 accuracy=84.534375	 loss=0.590563	 lr=0.000100
Epoch[096] Batch [0179]/[1203]	Speed: 109.741745 samples/sec	 accuracy=84.519444	 loss=0.589551	 lr=0.000100
Epoch[096] Batch [0199]/[1203]	Speed: 113.170044 samples/sec	 accuracy=84.425000	 loss=0.591882	 lr=0.000100
Epoch[096] Batch [0219]/[1203]	Speed: 147.669413 samples/sec	 accuracy=84.354545	 loss=0.594499	 lr=0.000100
Epoch[096] Batch [0239]/[1203]	Speed: 107.713218 samples/sec	 accuracy=84.429167	 loss=0.592130	 lr=0.000100
Epoch[096] Batch [0259]/[1203]	Speed: 143.697287 samples/sec	 accuracy=84.450000	 loss=0.592311	 lr=0.000100
Epoch[096] Batch [0279]/[1203]	Speed: 141.323316 samples/sec	 accuracy=84.425000	 loss=0.592824	 lr=0.000100
Epoch[096] Batch [0299]/[1203]	Speed: 105.017865 samples/sec	 accuracy=84.390000	 loss=0.592255	 lr=0.000100
Epoch[096] Batch [0319]/[1203]	Speed: 144.896111 samples/sec	 accuracy=84.364063	 loss=0.591851	 lr=0.000100
Epoch[096] Batch [0339]/[1203]	Speed: 126.734328 samples/sec	 accuracy=84.307353	 loss=0.594149	 lr=0.000100
Epoch[096] Batch [0359]/[1203]	Speed: 139.788305 samples/sec	 accuracy=84.325000	 loss=0.594870	 lr=0.000100
Epoch[096] Batch [0379]/[1203]	Speed: 115.578455 samples/sec	 accuracy=84.321053	 loss=0.595852	 lr=0.000100
Epoch[096] Batch [0399]/[1203]	Speed: 116.862803 samples/sec	 accuracy=84.361250	 loss=0.594873	 lr=0.000100
Epoch[096] Batch [0419]/[1203]	Speed: 158.191099 samples/sec	 accuracy=84.361905	 loss=0.593901	 lr=0.000100
Epoch[096] Batch [0439]/[1203]	Speed: 121.223060 samples/sec	 accuracy=84.377273	 loss=0.593804	 lr=0.000100
Epoch[096] Batch [0459]/[1203]	Speed: 115.205366 samples/sec	 accuracy=84.408696	 loss=0.593181	 lr=0.000100
Epoch[096] Batch [0479]/[1203]	Speed: 143.601633 samples/sec	 accuracy=84.384375	 loss=0.593877	 lr=0.000100
Epoch[096] Batch [0499]/[1203]	Speed: 115.965490 samples/sec	 accuracy=84.403000	 loss=0.593623	 lr=0.000100
Epoch[096] Batch [0519]/[1203]	Speed: 143.881938 samples/sec	 accuracy=84.405769	 loss=0.593774	 lr=0.000100
Epoch[096] Batch [0539]/[1203]	Speed: 115.768264 samples/sec	 accuracy=84.406481	 loss=0.594341	 lr=0.000100
Epoch[096] Batch [0559]/[1203]	Speed: 122.400256 samples/sec	 accuracy=84.414286	 loss=0.594358	 lr=0.000100
Epoch[096] Batch [0579]/[1203]	Speed: 145.101179 samples/sec	 accuracy=84.424138	 loss=0.593596	 lr=0.000100
Epoch[096] Batch [0599]/[1203]	Speed: 120.723659 samples/sec	 accuracy=84.431667	 loss=0.593527	 lr=0.000100
Epoch[096] Batch [0619]/[1203]	Speed: 144.169162 samples/sec	 accuracy=84.426613	 loss=0.593884	 lr=0.000100
Epoch[096] Batch [0639]/[1203]	Speed: 125.113242 samples/sec	 accuracy=84.425000	 loss=0.593910	 lr=0.000100
Epoch[096] Batch [0659]/[1203]	Speed: 110.290133 samples/sec	 accuracy=84.439394	 loss=0.593653	 lr=0.000100
Epoch[096] Batch [0679]/[1203]	Speed: 135.632094 samples/sec	 accuracy=84.441176	 loss=0.593834	 lr=0.000100
Epoch[096] Batch [0699]/[1203]	Speed: 134.041714 samples/sec	 accuracy=84.456429	 loss=0.593308	 lr=0.000100
Epoch[096] Batch [0719]/[1203]	Speed: 130.052172 samples/sec	 accuracy=84.459028	 loss=0.593129	 lr=0.000100
Epoch[096] Batch [0739]/[1203]	Speed: 120.192196 samples/sec	 accuracy=84.443243	 loss=0.593657	 lr=0.000100
Epoch[096] Batch [0759]/[1203]	Speed: 116.398463 samples/sec	 accuracy=84.446711	 loss=0.593509	 lr=0.000100
Epoch[096] Batch [0779]/[1203]	Speed: 162.177966 samples/sec	 accuracy=84.448718	 loss=0.593633	 lr=0.000100
Epoch[096] Batch [0799]/[1203]	Speed: 120.133842 samples/sec	 accuracy=84.468750	 loss=0.593030	 lr=0.000100
Epoch[096] Batch [0819]/[1203]	Speed: 115.199103 samples/sec	 accuracy=84.454268	 loss=0.593843	 lr=0.000100
Epoch[096] Batch [0839]/[1203]	Speed: 135.730749 samples/sec	 accuracy=84.461310	 loss=0.593295	 lr=0.000100
Epoch[096] Batch [0859]/[1203]	Speed: 121.970374 samples/sec	 accuracy=84.458140	 loss=0.593385	 lr=0.000100
Epoch[096] Batch [0879]/[1203]	Speed: 150.516345 samples/sec	 accuracy=84.465909	 loss=0.592990	 lr=0.000100
Epoch[096] Batch [0899]/[1203]	Speed: 120.613064 samples/sec	 accuracy=84.451667	 loss=0.593516	 lr=0.000100
Epoch[096] Batch [0919]/[1203]	Speed: 117.971008 samples/sec	 accuracy=84.426087	 loss=0.594156	 lr=0.000100
Epoch[096] Batch [0939]/[1203]	Speed: 125.614806 samples/sec	 accuracy=84.427660	 loss=0.594293	 lr=0.000100
Epoch[096] Batch [0959]/[1203]	Speed: 127.923872 samples/sec	 accuracy=84.447917	 loss=0.593730	 lr=0.000100
Epoch[096] Batch [0979]/[1203]	Speed: 130.836550 samples/sec	 accuracy=84.442347	 loss=0.594031	 lr=0.000100
Epoch[096] Batch [0999]/[1203]	Speed: 134.603180 samples/sec	 accuracy=84.427500	 loss=0.594636	 lr=0.000100
Epoch[096] Batch [1019]/[1203]	Speed: 124.821269 samples/sec	 accuracy=84.406373	 loss=0.595317	 lr=0.000100
Epoch[096] Batch [1039]/[1203]	Speed: 124.579028 samples/sec	 accuracy=84.397115	 loss=0.595770	 lr=0.000100
Epoch[096] Batch [1059]/[1203]	Speed: 132.028597 samples/sec	 accuracy=84.434434	 loss=0.594703	 lr=0.000100
Epoch[096] Batch [1079]/[1203]	Speed: 123.672035 samples/sec	 accuracy=84.428241	 loss=0.595187	 lr=0.000100
Epoch[096] Batch [1099]/[1203]	Speed: 132.767235 samples/sec	 accuracy=84.421818	 loss=0.595314	 lr=0.000100
Epoch[096] Batch [1119]/[1203]	Speed: 119.240948 samples/sec	 accuracy=84.418750	 loss=0.595413	 lr=0.000100
Epoch[096] Batch [1139]/[1203]	Speed: 142.615381 samples/sec	 accuracy=84.423246	 loss=0.595367	 lr=0.000100
Epoch[096] Batch [1159]/[1203]	Speed: 124.478172 samples/sec	 accuracy=84.426293	 loss=0.595328	 lr=0.000100
Epoch[096] Batch [1179]/[1203]	Speed: 124.880824 samples/sec	 accuracy=84.431780	 loss=0.595585	 lr=0.000100
Epoch[096] Batch [1199]/[1203]	Speed: 247.231986 samples/sec	 accuracy=84.437500	 loss=0.595369	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.500000 acc-top5=84.750000
Batch [0039]/[0097]: acc-top1=63.325000 acc-top5=84.837500
Batch [0059]/[0097]: acc-top1=65.400000 acc-top5=86.141667
Batch [0079]/[0097]: acc-top1=65.475000 acc-top5=85.956250
[Epoch 096] training: accuracy=84.435578	 loss=0.595479
[Epoch 096] speed: 126 samples/sec	time cost: 2090.313995
[Epoch 096] validation: acc-top1=64.953608 acc-top5=85.850515 loss=1.569177
Epoch[097] Batch [0019]/[1203]	Speed: 70.019401 samples/sec	 accuracy=84.375000	 loss=0.595299	 lr=0.000100
Epoch[097] Batch [0039]/[1203]	Speed: 100.996902 samples/sec	 accuracy=84.250000	 loss=0.590054	 lr=0.000100
Epoch[097] Batch [0059]/[1203]	Speed: 250.935537 samples/sec	 accuracy=84.200000	 loss=0.593121	 lr=0.000100
Epoch[097] Batch [0079]/[1203]	Speed: 95.775936 samples/sec	 accuracy=84.187500	 loss=0.591523	 lr=0.000100
Epoch[097] Batch [0099]/[1203]	Speed: 116.347090 samples/sec	 accuracy=84.215000	 loss=0.590900	 lr=0.000100
Epoch[097] Batch [0119]/[1203]	Speed: 185.612934 samples/sec	 accuracy=84.079167	 loss=0.594561	 lr=0.000100
Epoch[097] Batch [0139]/[1203]	Speed: 97.197095 samples/sec	 accuracy=84.300000	 loss=0.586820	 lr=0.000100
Epoch[097] Batch [0159]/[1203]	Speed: 233.329885 samples/sec	 accuracy=84.334375	 loss=0.588507	 lr=0.000100
Epoch[097] Batch [0179]/[1203]	Speed: 101.554018 samples/sec	 accuracy=84.358333	 loss=0.588839	 lr=0.000100
Epoch[097] Batch [0199]/[1203]	Speed: 108.734472 samples/sec	 accuracy=84.305000	 loss=0.589252	 lr=0.000100
Epoch[097] Batch [0219]/[1203]	Speed: 179.621188 samples/sec	 accuracy=84.286364	 loss=0.589483	 lr=0.000100
Epoch[097] Batch [0239]/[1203]	Speed: 99.652327 samples/sec	 accuracy=84.325000	 loss=0.586738	 lr=0.000100
Epoch[097] Batch [0259]/[1203]	Speed: 138.983901 samples/sec	 accuracy=84.280769	 loss=0.588180	 lr=0.000100
Epoch[097] Batch [0279]/[1203]	Speed: 142.851808 samples/sec	 accuracy=84.255357	 loss=0.590727	 lr=0.000100
Epoch[097] Batch [0299]/[1203]	Speed: 101.011356 samples/sec	 accuracy=84.326667	 loss=0.590951	 lr=0.000100
Epoch[097] Batch [0319]/[1203]	Speed: 179.895330 samples/sec	 accuracy=84.340625	 loss=0.591582	 lr=0.000100
Epoch[097] Batch [0339]/[1203]	Speed: 104.538004 samples/sec	 accuracy=84.423529	 loss=0.589915	 lr=0.000100
Epoch[097] Batch [0359]/[1203]	Speed: 171.502616 samples/sec	 accuracy=84.383333	 loss=0.591714	 lr=0.000100
Epoch[097] Batch [0379]/[1203]	Speed: 117.674113 samples/sec	 accuracy=84.327632	 loss=0.593807	 lr=0.000100
Epoch[097] Batch [0399]/[1203]	Speed: 101.435658 samples/sec	 accuracy=84.295000	 loss=0.594925	 lr=0.000100
Epoch[097] Batch [0419]/[1203]	Speed: 191.937588 samples/sec	 accuracy=84.311905	 loss=0.594794	 lr=0.000100
Epoch[097] Batch [0439]/[1203]	Speed: 100.086007 samples/sec	 accuracy=84.307955	 loss=0.595294	 lr=0.000100
Epoch[097] Batch [0459]/[1203]	Speed: 141.310280 samples/sec	 accuracy=84.352174	 loss=0.594023	 lr=0.000100
Epoch[097] Batch [0479]/[1203]	Speed: 128.894576 samples/sec	 accuracy=84.388542	 loss=0.594133	 lr=0.000100
Epoch[097] Batch [0499]/[1203]	Speed: 117.902995 samples/sec	 accuracy=84.407000	 loss=0.593833	 lr=0.000100
Epoch[097] Batch [0519]/[1203]	Speed: 156.184899 samples/sec	 accuracy=84.392308	 loss=0.593602	 lr=0.000100
Epoch[097] Batch [0539]/[1203]	Speed: 96.952498 samples/sec	 accuracy=84.404630	 loss=0.593657	 lr=0.000100
Epoch[097] Batch [0559]/[1203]	Speed: 157.457991 samples/sec	 accuracy=84.412500	 loss=0.593560	 lr=0.000100
Epoch[097] Batch [0579]/[1203]	Speed: 121.501525 samples/sec	 accuracy=84.427586	 loss=0.593676	 lr=0.000100
Epoch[097] Batch [0599]/[1203]	Speed: 124.376849 samples/sec	 accuracy=84.425833	 loss=0.594007	 lr=0.000100
Epoch[097] Batch [0619]/[1203]	Speed: 164.508646 samples/sec	 accuracy=84.432258	 loss=0.593807	 lr=0.000100
Epoch[097] Batch [0639]/[1203]	Speed: 98.137983 samples/sec	 accuracy=84.428906	 loss=0.594356	 lr=0.000100
Epoch[097] Batch [0659]/[1203]	Speed: 144.367351 samples/sec	 accuracy=84.415152	 loss=0.595016	 lr=0.000100
Epoch[097] Batch [0679]/[1203]	Speed: 125.999644 samples/sec	 accuracy=84.411765	 loss=0.595296	 lr=0.000100
Epoch[097] Batch [0699]/[1203]	Speed: 116.761857 samples/sec	 accuracy=84.393571	 loss=0.596088	 lr=0.000100
Epoch[097] Batch [0719]/[1203]	Speed: 168.992660 samples/sec	 accuracy=84.381250	 loss=0.596112	 lr=0.000100
Epoch[097] Batch [0739]/[1203]	Speed: 106.111676 samples/sec	 accuracy=84.405405	 loss=0.595666	 lr=0.000100
Epoch[097] Batch [0759]/[1203]	Speed: 125.501162 samples/sec	 accuracy=84.409868	 loss=0.595540	 lr=0.000100
Epoch[097] Batch [0779]/[1203]	Speed: 148.876888 samples/sec	 accuracy=84.427564	 loss=0.595327	 lr=0.000100
Epoch[097] Batch [0799]/[1203]	Speed: 97.263727 samples/sec	 accuracy=84.434375	 loss=0.594841	 lr=0.000100
Epoch[097] Batch [0819]/[1203]	Speed: 208.014177 samples/sec	 accuracy=84.412195	 loss=0.595294	 lr=0.000100
Epoch[097] Batch [0839]/[1203]	Speed: 104.642005 samples/sec	 accuracy=84.420238	 loss=0.594904	 lr=0.000100
Epoch[097] Batch [0859]/[1203]	Speed: 125.496982 samples/sec	 accuracy=84.415116	 loss=0.595516	 lr=0.000100
Epoch[097] Batch [0879]/[1203]	Speed: 156.627078 samples/sec	 accuracy=84.399432	 loss=0.595813	 lr=0.000100
Epoch[097] Batch [0899]/[1203]	Speed: 96.355560 samples/sec	 accuracy=84.405556	 loss=0.595875	 lr=0.000100
Epoch[097] Batch [0919]/[1203]	Speed: 141.657885 samples/sec	 accuracy=84.406522	 loss=0.595926	 lr=0.000100
Epoch[097] Batch [0939]/[1203]	Speed: 134.214690 samples/sec	 accuracy=84.412766	 loss=0.595623	 lr=0.000100
Epoch[097] Batch [0959]/[1203]	Speed: 114.434738 samples/sec	 accuracy=84.419271	 loss=0.595374	 lr=0.000100
Epoch[097] Batch [0979]/[1203]	Speed: 179.536791 samples/sec	 accuracy=84.422449	 loss=0.595218	 lr=0.000100
Epoch[097] Batch [0999]/[1203]	Speed: 100.232316 samples/sec	 accuracy=84.417500	 loss=0.595331	 lr=0.000100
Epoch[097] Batch [1019]/[1203]	Speed: 141.256150 samples/sec	 accuracy=84.414706	 loss=0.595262	 lr=0.000100
Epoch[097] Batch [1039]/[1203]	Speed: 130.202692 samples/sec	 accuracy=84.415865	 loss=0.595008	 lr=0.000100
Epoch[097] Batch [1059]/[1203]	Speed: 108.544909 samples/sec	 accuracy=84.419340	 loss=0.595231	 lr=0.000100
Epoch[097] Batch [1079]/[1203]	Speed: 164.607366 samples/sec	 accuracy=84.427315	 loss=0.594846	 lr=0.000100
Epoch[097] Batch [1099]/[1203]	Speed: 110.255181 samples/sec	 accuracy=84.434545	 loss=0.594637	 lr=0.000100
Epoch[097] Batch [1119]/[1203]	Speed: 113.962390 samples/sec	 accuracy=84.437500	 loss=0.594673	 lr=0.000100
Epoch[097] Batch [1139]/[1203]	Speed: 147.229609 samples/sec	 accuracy=84.430702	 loss=0.595142	 lr=0.000100
Epoch[097] Batch [1159]/[1203]	Speed: 122.594010 samples/sec	 accuracy=84.427155	 loss=0.594982	 lr=0.000100
Epoch[097] Batch [1179]/[1203]	Speed: 152.130770 samples/sec	 accuracy=84.415678	 loss=0.595306	 lr=0.000100
Epoch[097] Batch [1199]/[1203]	Speed: 217.143157 samples/sec	 accuracy=84.412917	 loss=0.595369	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.650000 acc-top5=85.100000
Batch [0039]/[0097]: acc-top1=63.350000 acc-top5=84.937500
Batch [0059]/[0097]: acc-top1=65.516667 acc-top5=86.175000
Batch [0079]/[0097]: acc-top1=65.556250 acc-top5=85.987500
[Epoch 097] training: accuracy=84.410224	 loss=0.595325
[Epoch 097] speed: 126 samples/sec	time cost: 2093.015252
[Epoch 097] validation: acc-top1=64.963918 acc-top5=85.835052 loss=1.570514
Epoch[098] Batch [0019]/[1203]	Speed: 71.113354 samples/sec	 accuracy=83.650000	 loss=0.608221	 lr=0.000100
Epoch[098] Batch [0039]/[1203]	Speed: 103.563576 samples/sec	 accuracy=83.662500	 loss=0.614101	 lr=0.000100
Epoch[098] Batch [0059]/[1203]	Speed: 223.655353 samples/sec	 accuracy=83.833333	 loss=0.606314	 lr=0.000100
Epoch[098] Batch [0079]/[1203]	Speed: 91.835799 samples/sec	 accuracy=83.918750	 loss=0.608244	 lr=0.000100
Epoch[098] Batch [0099]/[1203]	Speed: 126.311092 samples/sec	 accuracy=83.785000	 loss=0.612232	 lr=0.000100
Epoch[098] Batch [0119]/[1203]	Speed: 161.581467 samples/sec	 accuracy=83.883333	 loss=0.610918	 lr=0.000100
Epoch[098] Batch [0139]/[1203]	Speed: 109.514228 samples/sec	 accuracy=83.942857	 loss=0.606978	 lr=0.000100
Epoch[098] Batch [0159]/[1203]	Speed: 191.839475 samples/sec	 accuracy=84.003125	 loss=0.606234	 lr=0.000100
Epoch[098] Batch [0179]/[1203]	Speed: 96.865941 samples/sec	 accuracy=83.972222	 loss=0.608998	 lr=0.000100
Epoch[098] Batch [0199]/[1203]	Speed: 136.702587 samples/sec	 accuracy=84.025000	 loss=0.607358	 lr=0.000100
Epoch[098] Batch [0219]/[1203]	Speed: 141.535078 samples/sec	 accuracy=84.145455	 loss=0.603368	 lr=0.000100
Epoch[098] Batch [0239]/[1203]	Speed: 110.242922 samples/sec	 accuracy=84.191667	 loss=0.602066	 lr=0.000100
Epoch[098] Batch [0259]/[1203]	Speed: 139.618164 samples/sec	 accuracy=84.246154	 loss=0.601113	 lr=0.000100
Epoch[098] Batch [0279]/[1203]	Speed: 118.148146 samples/sec	 accuracy=84.273214	 loss=0.598906	 lr=0.000100
Epoch[098] Batch [0299]/[1203]	Speed: 120.428105 samples/sec	 accuracy=84.238333	 loss=0.598956	 lr=0.000100
Epoch[098] Batch [0319]/[1203]	Speed: 164.257799 samples/sec	 accuracy=84.229687	 loss=0.599918	 lr=0.000100
Epoch[098] Batch [0339]/[1203]	Speed: 106.224887 samples/sec	 accuracy=84.177941	 loss=0.601715	 lr=0.000100
Epoch[098] Batch [0359]/[1203]	Speed: 164.340045 samples/sec	 accuracy=84.169444	 loss=0.602617	 lr=0.000100
Epoch[098] Batch [0379]/[1203]	Speed: 108.315282 samples/sec	 accuracy=84.200000	 loss=0.602647	 lr=0.000100
Epoch[098] Batch [0399]/[1203]	Speed: 120.148883 samples/sec	 accuracy=84.195000	 loss=0.603115	 lr=0.000100
Epoch[098] Batch [0419]/[1203]	Speed: 139.543509 samples/sec	 accuracy=84.203571	 loss=0.602525	 lr=0.000100
Epoch[098] Batch [0439]/[1203]	Speed: 123.022407 samples/sec	 accuracy=84.202273	 loss=0.601528	 lr=0.000100
Epoch[098] Batch [0459]/[1203]	Speed: 140.794509 samples/sec	 accuracy=84.220652	 loss=0.601617	 lr=0.000100
Epoch[098] Batch [0479]/[1203]	Speed: 118.612729 samples/sec	 accuracy=84.205208	 loss=0.602799	 lr=0.000100
Epoch[098] Batch [0499]/[1203]	Speed: 123.455155 samples/sec	 accuracy=84.246000	 loss=0.601957	 lr=0.000100
Epoch[098] Batch [0519]/[1203]	Speed: 152.409852 samples/sec	 accuracy=84.255769	 loss=0.601268	 lr=0.000100
Epoch[098] Batch [0539]/[1203]	Speed: 108.744992 samples/sec	 accuracy=84.242593	 loss=0.600847	 lr=0.000100
Epoch[098] Batch [0559]/[1203]	Speed: 134.882783 samples/sec	 accuracy=84.247321	 loss=0.600076	 lr=0.000100
Epoch[098] Batch [0579]/[1203]	Speed: 124.341334 samples/sec	 accuracy=84.249138	 loss=0.599865	 lr=0.000100
Epoch[098] Batch [0599]/[1203]	Speed: 146.845824 samples/sec	 accuracy=84.256667	 loss=0.599967	 lr=0.000100
Epoch[098] Batch [0619]/[1203]	Speed: 138.794410 samples/sec	 accuracy=84.274194	 loss=0.600100	 lr=0.000100
Epoch[098] Batch [0639]/[1203]	Speed: 101.865492 samples/sec	 accuracy=84.286719	 loss=0.600026	 lr=0.000100
Epoch[098] Batch [0659]/[1203]	Speed: 137.514820 samples/sec	 accuracy=84.305303	 loss=0.599320	 lr=0.000100
Epoch[098] Batch [0679]/[1203]	Speed: 134.136563 samples/sec	 accuracy=84.307353	 loss=0.599123	 lr=0.000100
Epoch[098] Batch [0699]/[1203]	Speed: 113.811153 samples/sec	 accuracy=84.330000	 loss=0.598556	 lr=0.000100
Epoch[098] Batch [0719]/[1203]	Speed: 163.566536 samples/sec	 accuracy=84.356250	 loss=0.597459	 lr=0.000100
Epoch[098] Batch [0739]/[1203]	Speed: 105.195924 samples/sec	 accuracy=84.382432	 loss=0.596780	 lr=0.000100
Epoch[098] Batch [0759]/[1203]	Speed: 141.320379 samples/sec	 accuracy=84.398026	 loss=0.596333	 lr=0.000100
Epoch[098] Batch [0779]/[1203]	Speed: 138.815866 samples/sec	 accuracy=84.401923	 loss=0.595643	 lr=0.000100
Epoch[098] Batch [0799]/[1203]	Speed: 107.738560 samples/sec	 accuracy=84.416875	 loss=0.595279	 lr=0.000100
Epoch[098] Batch [0819]/[1203]	Speed: 135.762095 samples/sec	 accuracy=84.434756	 loss=0.595235	 lr=0.000100
Epoch[098] Batch [0839]/[1203]	Speed: 120.564799 samples/sec	 accuracy=84.442262	 loss=0.595091	 lr=0.000100
Epoch[098] Batch [0859]/[1203]	Speed: 131.459459 samples/sec	 accuracy=84.425000	 loss=0.595955	 lr=0.000100
Epoch[098] Batch [0879]/[1203]	Speed: 153.315105 samples/sec	 accuracy=84.442614	 loss=0.595530	 lr=0.000100
Epoch[098] Batch [0899]/[1203]	Speed: 104.580000 samples/sec	 accuracy=84.416111	 loss=0.596227	 lr=0.000100
Epoch[098] Batch [0919]/[1203]	Speed: 120.336477 samples/sec	 accuracy=84.419565	 loss=0.596018	 lr=0.000100
Epoch[098] Batch [0939]/[1203]	Speed: 163.354359 samples/sec	 accuracy=84.437766	 loss=0.595556	 lr=0.000100
Epoch[098] Batch [0959]/[1203]	Speed: 107.544107 samples/sec	 accuracy=84.430729	 loss=0.596213	 lr=0.000100
Epoch[098] Batch [0979]/[1203]	Speed: 161.529911 samples/sec	 accuracy=84.440306	 loss=0.596002	 lr=0.000100
Epoch[098] Batch [0999]/[1203]	Speed: 104.149707 samples/sec	 accuracy=84.440000	 loss=0.595987	 lr=0.000100
Epoch[098] Batch [1019]/[1203]	Speed: 152.639861 samples/sec	 accuracy=84.410294	 loss=0.596918	 lr=0.000100
Epoch[098] Batch [1039]/[1203]	Speed: 127.051042 samples/sec	 accuracy=84.423558	 loss=0.596658	 lr=0.000100
Epoch[098] Batch [1059]/[1203]	Speed: 109.932916 samples/sec	 accuracy=84.391981	 loss=0.597590	 lr=0.000100
Epoch[098] Batch [1079]/[1203]	Speed: 141.859437 samples/sec	 accuracy=84.406019	 loss=0.597032	 lr=0.000100
Epoch[098] Batch [1099]/[1203]	Speed: 130.235003 samples/sec	 accuracy=84.400909	 loss=0.596952	 lr=0.000100
Epoch[098] Batch [1119]/[1203]	Speed: 119.179461 samples/sec	 accuracy=84.404911	 loss=0.596779	 lr=0.000100
Epoch[098] Batch [1139]/[1203]	Speed: 143.436525 samples/sec	 accuracy=84.394737	 loss=0.596998	 lr=0.000100
Epoch[098] Batch [1159]/[1203]	Speed: 111.305014 samples/sec	 accuracy=84.399138	 loss=0.596538	 lr=0.000100
Epoch[098] Batch [1179]/[1203]	Speed: 154.359458 samples/sec	 accuracy=84.386441	 loss=0.596532	 lr=0.000100
Epoch[098] Batch [1199]/[1203]	Speed: 236.938293 samples/sec	 accuracy=84.381250	 loss=0.596688	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.525000 acc-top5=85.075000
Batch [0039]/[0097]: acc-top1=63.187500 acc-top5=84.900000
Batch [0059]/[0097]: acc-top1=65.350000 acc-top5=86.200000
Batch [0079]/[0097]: acc-top1=65.462500 acc-top5=86.150000
[Epoch 098] training: accuracy=84.386118	 loss=0.596583
[Epoch 098] speed: 126 samples/sec	time cost: 2088.852056
[Epoch 098] validation: acc-top1=64.865979 acc-top5=85.974227 loss=1.568927
Epoch[099] Batch [0019]/[1204]	Speed: 69.060116 samples/sec	 accuracy=84.175000	 loss=0.588093	 lr=0.000100
Epoch[099] Batch [0039]/[1204]	Speed: 100.111749 samples/sec	 accuracy=84.550000	 loss=0.586671	 lr=0.000100
Epoch[099] Batch [0059]/[1204]	Speed: 252.148859 samples/sec	 accuracy=84.550000	 loss=0.588663	 lr=0.000100
Epoch[099] Batch [0079]/[1204]	Speed: 99.271684 samples/sec	 accuracy=84.500000	 loss=0.592917	 lr=0.000100
Epoch[099] Batch [0099]/[1204]	Speed: 113.979557 samples/sec	 accuracy=84.565000	 loss=0.593709	 lr=0.000100
Epoch[099] Batch [0119]/[1204]	Speed: 183.655421 samples/sec	 accuracy=84.537500	 loss=0.594178	 lr=0.000100
Epoch[099] Batch [0139]/[1204]	Speed: 93.035972 samples/sec	 accuracy=84.453571	 loss=0.598601	 lr=0.000100
Epoch[099] Batch [0159]/[1204]	Speed: 245.364503 samples/sec	 accuracy=84.478125	 loss=0.598277	 lr=0.000100
Epoch[099] Batch [0179]/[1204]	Speed: 91.236064 samples/sec	 accuracy=84.466667	 loss=0.597225	 lr=0.000100
Epoch[099] Batch [0199]/[1204]	Speed: 123.741981 samples/sec	 accuracy=84.485000	 loss=0.595571	 lr=0.000100
Epoch[099] Batch [0219]/[1204]	Speed: 162.942557 samples/sec	 accuracy=84.406818	 loss=0.595428	 lr=0.000100
Epoch[099] Batch [0239]/[1204]	Speed: 100.603695 samples/sec	 accuracy=84.437500	 loss=0.595873	 lr=0.000100
Epoch[099] Batch [0259]/[1204]	Speed: 173.775513 samples/sec	 accuracy=84.442308	 loss=0.596300	 lr=0.000100
Epoch[099] Batch [0279]/[1204]	Speed: 111.754066 samples/sec	 accuracy=84.476786	 loss=0.595445	 lr=0.000100
Epoch[099] Batch [0299]/[1204]	Speed: 117.439359 samples/sec	 accuracy=84.458333	 loss=0.596615	 lr=0.000100
Epoch[099] Batch [0319]/[1204]	Speed: 153.959911 samples/sec	 accuracy=84.517188	 loss=0.594160	 lr=0.000100
Epoch[099] Batch [0339]/[1204]	Speed: 108.443064 samples/sec	 accuracy=84.560294	 loss=0.593851	 lr=0.000100
Epoch[099] Batch [0359]/[1204]	Speed: 137.171852 samples/sec	 accuracy=84.523611	 loss=0.594278	 lr=0.000100
Epoch[099] Batch [0379]/[1204]	Speed: 126.609751 samples/sec	 accuracy=84.494737	 loss=0.594564	 lr=0.000100
Epoch[099] Batch [0399]/[1204]	Speed: 129.685114 samples/sec	 accuracy=84.467500	 loss=0.594703	 lr=0.000100
Epoch[099] Batch [0419]/[1204]	Speed: 143.615407 samples/sec	 accuracy=84.455952	 loss=0.594700	 lr=0.000100
Epoch[099] Batch [0439]/[1204]	Speed: 113.076199 samples/sec	 accuracy=84.437500	 loss=0.594700	 lr=0.000100
Epoch[099] Batch [0459]/[1204]	Speed: 142.878150 samples/sec	 accuracy=84.439130	 loss=0.594195	 lr=0.000100
Epoch[099] Batch [0479]/[1204]	Speed: 114.861904 samples/sec	 accuracy=84.439583	 loss=0.593664	 lr=0.000100
Epoch[099] Batch [0499]/[1204]	Speed: 135.356145 samples/sec	 accuracy=84.459000	 loss=0.593097	 lr=0.000100
Epoch[099] Batch [0519]/[1204]	Speed: 120.103313 samples/sec	 accuracy=84.425962	 loss=0.594573	 lr=0.000100
Epoch[099] Batch [0539]/[1204]	Speed: 137.311596 samples/sec	 accuracy=84.437963	 loss=0.594224	 lr=0.000100
Epoch[099] Batch [0559]/[1204]	Speed: 124.434837 samples/sec	 accuracy=84.420536	 loss=0.593646	 lr=0.000100
Epoch[099] Batch [0579]/[1204]	Speed: 126.291109 samples/sec	 accuracy=84.437069	 loss=0.593463	 lr=0.000100
Epoch[099] Batch [0599]/[1204]	Speed: 136.477984 samples/sec	 accuracy=84.448333	 loss=0.593272	 lr=0.000100
Epoch[099] Batch [0619]/[1204]	Speed: 122.192825 samples/sec	 accuracy=84.438710	 loss=0.593393	 lr=0.000100
Epoch[099] Batch [0639]/[1204]	Speed: 122.664923 samples/sec	 accuracy=84.449219	 loss=0.592373	 lr=0.000100
Epoch[099] Batch [0659]/[1204]	Speed: 128.562016 samples/sec	 accuracy=84.440909	 loss=0.592981	 lr=0.000100
Epoch[099] Batch [0679]/[1204]	Speed: 128.412299 samples/sec	 accuracy=84.435294	 loss=0.593316	 lr=0.000100
Epoch[099] Batch [0699]/[1204]	Speed: 134.321793 samples/sec	 accuracy=84.443571	 loss=0.592996	 lr=0.000100
Epoch[099] Batch [0719]/[1204]	Speed: 121.905877 samples/sec	 accuracy=84.445833	 loss=0.593894	 lr=0.000100
Epoch[099] Batch [0739]/[1204]	Speed: 115.360054 samples/sec	 accuracy=84.450676	 loss=0.593622	 lr=0.000100
Epoch[099] Batch [0759]/[1204]	Speed: 143.301560 samples/sec	 accuracy=84.455921	 loss=0.593886	 lr=0.000100
Epoch[099] Batch [0779]/[1204]	Speed: 111.050046 samples/sec	 accuracy=84.459615	 loss=0.593493	 lr=0.000100
Epoch[099] Batch [0799]/[1204]	Speed: 158.558155 samples/sec	 accuracy=84.459375	 loss=0.593324	 lr=0.000100
Epoch[099] Batch [0819]/[1204]	Speed: 115.061228 samples/sec	 accuracy=84.463415	 loss=0.593520	 lr=0.000100
Epoch[099] Batch [0839]/[1204]	Speed: 128.871960 samples/sec	 accuracy=84.443452	 loss=0.593938	 lr=0.000100
Epoch[099] Batch [0859]/[1204]	Speed: 145.196291 samples/sec	 accuracy=84.450000	 loss=0.594205	 lr=0.000100
Epoch[099] Batch [0879]/[1204]	Speed: 115.689725 samples/sec	 accuracy=84.496591	 loss=0.593361	 lr=0.000100
Epoch[099] Batch [0899]/[1204]	Speed: 151.036503 samples/sec	 accuracy=84.496111	 loss=0.593610	 lr=0.000100
Epoch[099] Batch [0919]/[1204]	Speed: 106.548086 samples/sec	 accuracy=84.500000	 loss=0.593611	 lr=0.000100
Epoch[099] Batch [0939]/[1204]	Speed: 118.030817 samples/sec	 accuracy=84.502660	 loss=0.593699	 lr=0.000100
Epoch[099] Batch [0959]/[1204]	Speed: 173.755502 samples/sec	 accuracy=84.513021	 loss=0.593323	 lr=0.000100
Epoch[099] Batch [0979]/[1204]	Speed: 108.845003 samples/sec	 accuracy=84.517857	 loss=0.593237	 lr=0.000100
Epoch[099] Batch [0999]/[1204]	Speed: 128.993869 samples/sec	 accuracy=84.512000	 loss=0.593470	 lr=0.000100
Epoch[099] Batch [1019]/[1204]	Speed: 128.207362 samples/sec	 accuracy=84.482843	 loss=0.594317	 lr=0.000100
Epoch[099] Batch [1039]/[1204]	Speed: 115.477819 samples/sec	 accuracy=84.477885	 loss=0.594408	 lr=0.000100
Epoch[099] Batch [1059]/[1204]	Speed: 165.649939 samples/sec	 accuracy=84.484434	 loss=0.594433	 lr=0.000100
Epoch[099] Batch [1079]/[1204]	Speed: 109.592317 samples/sec	 accuracy=84.490741	 loss=0.593881	 lr=0.000100
Epoch[099] Batch [1099]/[1204]	Speed: 115.911178 samples/sec	 accuracy=84.484545	 loss=0.593907	 lr=0.000100
Epoch[099] Batch [1119]/[1204]	Speed: 150.367312 samples/sec	 accuracy=84.480804	 loss=0.594067	 lr=0.000100
Epoch[099] Batch [1139]/[1204]	Speed: 124.349540 samples/sec	 accuracy=84.484649	 loss=0.593805	 lr=0.000100
Epoch[099] Batch [1159]/[1204]	Speed: 151.123905 samples/sec	 accuracy=84.484914	 loss=0.593920	 lr=0.000100
Epoch[099] Batch [1179]/[1204]	Speed: 112.136068 samples/sec	 accuracy=84.469492	 loss=0.594310	 lr=0.000100
Epoch[099] Batch [1199]/[1204]	Speed: 238.978381 samples/sec	 accuracy=84.479583	 loss=0.594009	 lr=0.000100
Batch [0019]/[0097]: acc-top1=64.675000 acc-top5=84.775000
Batch [0039]/[0097]: acc-top1=63.375000 acc-top5=84.800000
Batch [0059]/[0097]: acc-top1=65.391667 acc-top5=86.050000
Batch [0079]/[0097]: acc-top1=65.543750 acc-top5=85.918750
[Epoch 099] training: accuracy=84.477575	 loss=0.594005
[Epoch 099] speed: 126 samples/sec	time cost: 2093.171963
[Epoch 099] validation: acc-top1=64.958763 acc-top5=85.783505 loss=1.569759