Namespace(accumulate=1, batch_norm=False, batch_size=8, 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=299, kvstore=None, label_smoothing=False, last_gamma=False, log_interval=50, logging_file='i3d_inceptionv3_seg1_kinetics400_b8_g8_video_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='i3d_inceptionv3_kinetics400', momentum=0.9, new_height=340, new_length=32, new_step=2, new_width=450, no_wd=False, num_classes=400, num_crop=1, num_epochs=100, num_gpus=8, num_segments=1, num_workers=32, partial_bn=False, prefetch_ratio=1.0, resume_epoch=0, resume_params='', resume_states='', save_dir='/home/ubuntu/yizhu/logs/mxnet/kinetics400/i3d/i3d_inceptionv3_seg1_kinetics400_b8_g8_video_cleanv3', save_frequency=20, 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 64 on 8 GPUs
I3D_InceptionV3(
  (features): HybridSequential(
    (0): HybridSequential(
      (0): Conv3D(3 -> 32, kernel_size=(3, 3, 3), stride=(2, 2, 2), padding=(1, 0, 0), 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): Conv3D(32 -> 32, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 0, 0), 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)
    )
    (2): HybridSequential(
      (0): Conv3D(32 -> 64, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 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): MaxPool3D(size=(3, 3, 3), stride=(1, 2, 2), padding=(1, 0, 0), ceil_mode=False, global_pool=False, pool_type=max, layout=NCDHW)
    (4): HybridSequential(
      (0): Conv3D(64 -> 80, kernel_size=(1, 1, 1), stride=(1, 1, 1), bias=False)
      (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=80)
      (2): Activation(relu)
    )
    (5): HybridSequential(
      (0): Conv3D(80 -> 192, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 0, 0), 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)
    )
    (6): MaxPool3D(size=(3, 3, 3), stride=(1, 2, 2), padding=(1, 0, 0), ceil_mode=False, global_pool=False, pool_type=max, layout=NCDHW)
    (7): HybridConcurrent(
      (0): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(192 -> 64, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(192 -> 48, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(48 -> 64, kernel_size=(5, 5, 5), stride=(1, 1, 1), padding=(2, 2, 2), 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)
        )
      )
      (2): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(192 -> 64, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(64 -> 96, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 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)
        )
        (2): HybridSequential(
          (0): Conv3D(96 -> 96, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 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): AvgPool3D(size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), ceil_mode=False, global_pool=False, pool_type=avg, layout=NCDHW)
        (1): HybridSequential(
          (0): Conv3D(192 -> 32, kernel_size=(1, 1, 1), stride=(1, 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)
        )
      )
    )
    (8): HybridConcurrent(
      (0): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(256 -> 64, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(256 -> 48, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(48 -> 64, kernel_size=(5, 5, 5), stride=(1, 1, 1), padding=(2, 2, 2), 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)
        )
      )
      (2): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(256 -> 64, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(64 -> 96, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 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)
        )
        (2): HybridSequential(
          (0): Conv3D(96 -> 96, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 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): AvgPool3D(size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), ceil_mode=False, global_pool=False, pool_type=avg, layout=NCDHW)
        (1): HybridSequential(
          (0): Conv3D(256 -> 64, kernel_size=(1, 1, 1), stride=(1, 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)
        )
      )
    )
    (9): HybridConcurrent(
      (0): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(288 -> 64, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(288 -> 48, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(48 -> 64, kernel_size=(5, 5, 5), stride=(1, 1, 1), padding=(2, 2, 2), 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)
        )
      )
      (2): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(288 -> 64, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(64 -> 96, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 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)
        )
        (2): HybridSequential(
          (0): Conv3D(96 -> 96, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 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): AvgPool3D(size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), ceil_mode=False, global_pool=False, pool_type=avg, layout=NCDHW)
        (1): HybridSequential(
          (0): Conv3D(288 -> 64, kernel_size=(1, 1, 1), stride=(1, 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)
        )
      )
    )
    (10): HybridConcurrent(
      (0): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(288 -> 384, kernel_size=(3, 3, 3), stride=(2, 2, 2), padding=(1, 0, 0), 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): Conv3D(288 -> 64, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(64 -> 96, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 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)
        )
        (2): HybridSequential(
          (0): Conv3D(96 -> 96, kernel_size=(3, 3, 3), stride=(2, 2, 2), padding=(1, 0, 0), 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)
        )
      )
      (2): HybridSequential(
        (0): MaxPool3D(size=(3, 3, 3), stride=(2, 2, 2), padding=(1, 0, 0), ceil_mode=False, global_pool=False, pool_type=max, layout=NCDHW)
      )
    )
    (11): HybridConcurrent(
      (0): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(768 -> 128, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(128 -> 128, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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): Conv3D(128 -> 192, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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): Conv3D(768 -> 128, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(128 -> 128, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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): Conv3D(128 -> 128, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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): Conv3D(128 -> 128, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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)
        )
        (4): HybridSequential(
          (0): Conv3D(128 -> 192, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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)
        )
      )
      (3): HybridSequential(
        (0): AvgPool3D(size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), ceil_mode=False, global_pool=False, pool_type=avg, layout=NCDHW)
        (1): HybridSequential(
          (0): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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)
        )
      )
    )
    (12): HybridConcurrent(
      (0): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(768 -> 160, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(160 -> 160, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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)
        )
        (2): HybridSequential(
          (0): Conv3D(160 -> 192, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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): Conv3D(768 -> 160, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(160 -> 160, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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)
        )
        (2): HybridSequential(
          (0): Conv3D(160 -> 160, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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)
        )
        (3): HybridSequential(
          (0): Conv3D(160 -> 160, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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)
        )
        (4): HybridSequential(
          (0): Conv3D(160 -> 192, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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)
        )
      )
      (3): HybridSequential(
        (0): AvgPool3D(size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), ceil_mode=False, global_pool=False, pool_type=avg, layout=NCDHW)
        (1): HybridSequential(
          (0): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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)
        )
      )
    )
    (13): HybridConcurrent(
      (0): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(768 -> 160, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(160 -> 160, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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)
        )
        (2): HybridSequential(
          (0): Conv3D(160 -> 192, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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): Conv3D(768 -> 160, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(160 -> 160, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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)
        )
        (2): HybridSequential(
          (0): Conv3D(160 -> 160, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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)
        )
        (3): HybridSequential(
          (0): Conv3D(160 -> 160, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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)
        )
        (4): HybridSequential(
          (0): Conv3D(160 -> 192, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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)
        )
      )
      (3): HybridSequential(
        (0): AvgPool3D(size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), ceil_mode=False, global_pool=False, pool_type=avg, layout=NCDHW)
        (1): HybridSequential(
          (0): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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)
        )
      )
    )
    (14): HybridConcurrent(
      (0): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(192 -> 192, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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): Conv3D(192 -> 192, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(192 -> 192, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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): Conv3D(192 -> 192, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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)
        )
        (3): HybridSequential(
          (0): Conv3D(192 -> 192, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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)
        )
        (4): HybridSequential(
          (0): Conv3D(192 -> 192, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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)
        )
      )
      (3): HybridSequential(
        (0): AvgPool3D(size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), ceil_mode=False, global_pool=False, pool_type=avg, layout=NCDHW)
        (1): HybridSequential(
          (0): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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)
        )
      )
    )
    (15): HybridConcurrent(
      (0): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(192 -> 320, kernel_size=(3, 3, 3), stride=(2, 2, 2), padding=(1, 0, 0), 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)
        )
      )
      (1): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(768 -> 192, kernel_size=(1, 1, 1), stride=(1, 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): Conv3D(192 -> 192, kernel_size=(7, 1, 7), stride=(1, 1, 1), padding=(3, 0, 3), 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): Conv3D(192 -> 192, kernel_size=(7, 7, 1), stride=(1, 1, 1), padding=(3, 3, 0), 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)
        )
        (3): HybridSequential(
          (0): Conv3D(192 -> 192, kernel_size=(3, 3, 3), stride=(2, 2, 2), padding=(1, 0, 0), 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): MaxPool3D(size=(3, 3, 3), stride=(2, 2, 2), padding=(1, 0, 0), ceil_mode=False, global_pool=False, pool_type=max, layout=NCDHW)
      )
    )
    (16): HybridConcurrent(
      (0): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(1280 -> 320, kernel_size=(1, 1, 1), stride=(1, 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)
        )
      )
      (1): HybridSequential(
        (0): HybridSequential(
          (0): HybridSequential(
            (0): Conv3D(1280 -> 384, kernel_size=(1, 1, 1), stride=(1, 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): HybridConcurrent(
          (0): HybridSequential(
            (0): HybridSequential(
              (0): Conv3D(384 -> 384, kernel_size=(3, 1, 3), stride=(1, 1, 1), padding=(1, 0, 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): Conv3D(384 -> 384, kernel_size=(3, 3, 1), stride=(1, 1, 1), padding=(1, 1, 0), 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): HybridSequential(
            (0): Conv3D(1280 -> 448, kernel_size=(1, 1, 1), stride=(1, 1, 1), bias=False)
            (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=448)
            (2): Activation(relu)
          )
          (1): HybridSequential(
            (0): Conv3D(448 -> 384, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 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): HybridConcurrent(
          (0): HybridSequential(
            (0): HybridSequential(
              (0): Conv3D(384 -> 384, kernel_size=(3, 1, 3), stride=(1, 1, 1), padding=(1, 0, 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): Conv3D(384 -> 384, kernel_size=(3, 3, 1), stride=(1, 1, 1), padding=(1, 1, 0), 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)
            )
          )
        )
      )
      (3): HybridSequential(
        (0): AvgPool3D(size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), ceil_mode=False, global_pool=False, pool_type=avg, layout=NCDHW)
        (1): HybridSequential(
          (0): Conv3D(1280 -> 192, kernel_size=(1, 1, 1), stride=(1, 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)
        )
      )
    )
    (17): HybridConcurrent(
      (0): HybridSequential(
        (0): HybridSequential(
          (0): Conv3D(2048 -> 320, kernel_size=(1, 1, 1), stride=(1, 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)
        )
      )
      (1): HybridSequential(
        (0): HybridSequential(
          (0): HybridSequential(
            (0): Conv3D(2048 -> 384, kernel_size=(1, 1, 1), stride=(1, 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): HybridConcurrent(
          (0): HybridSequential(
            (0): HybridSequential(
              (0): Conv3D(384 -> 384, kernel_size=(3, 1, 3), stride=(1, 1, 1), padding=(1, 0, 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): Conv3D(384 -> 384, kernel_size=(3, 3, 1), stride=(1, 1, 1), padding=(1, 1, 0), 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): HybridSequential(
            (0): Conv3D(2048 -> 448, kernel_size=(1, 1, 1), stride=(1, 1, 1), bias=False)
            (1): BatchNorm(axis=1, eps=0.001, momentum=0.9, fix_gamma=False, use_global_stats=False, in_channels=448)
            (2): Activation(relu)
          )
          (1): HybridSequential(
            (0): Conv3D(448 -> 384, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 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): HybridConcurrent(
          (0): HybridSequential(
            (0): HybridSequential(
              (0): Conv3D(384 -> 384, kernel_size=(3, 1, 3), stride=(1, 1, 1), padding=(1, 0, 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): Conv3D(384 -> 384, kernel_size=(3, 3, 1), stride=(1, 1, 1), padding=(1, 1, 0), 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)
            )
          )
        )
      )
      (3): HybridSequential(
        (0): AvgPool3D(size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), ceil_mode=False, global_pool=False, pool_type=avg, layout=NCDHW)
        (1): HybridSequential(
          (0): Conv3D(2048 -> 192, kernel_size=(1, 1, 1), stride=(1, 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)
        )
      )
    )
    (18): GlobalAvgPool3D(size=(1, 1, 1), stride=(1, 1, 1), padding=(0, 0, 0), ceil_mode=True, global_pool=True, pool_type=avg, layout=NCDHW)
  )
  (head): HybridSequential(
    (0): Dropout(p = 0.5, axes=())
    (1): Dense(2048 -> 400, linear)
  )
  (output): Dense(2048 -> 400, linear)
)
Load 240618 training samples and 19404 validation samples.
Epoch[000] Batch [0049]/[3759]	Speed: 20.016565 samples/sec	 accuracy=1.125000	 loss=5.929721	 lr=0.010000
Epoch[000] Batch [0099]/[3759]	Speed: 34.179887 samples/sec	 accuracy=1.359375	 loss=5.827388	 lr=0.010000
Epoch[000] Batch [0149]/[3759]	Speed: 34.852109 samples/sec	 accuracy=1.708333	 loss=5.731140	 lr=0.010000
Epoch[000] Batch [0199]/[3759]	Speed: 34.529270 samples/sec	 accuracy=2.000000	 loss=5.639669	 lr=0.010000
Epoch[000] Batch [0249]/[3759]	Speed: 34.747719 samples/sec	 accuracy=2.375000	 loss=5.569702	 lr=0.010000
Epoch[000] Batch [0299]/[3759]	Speed: 34.497962 samples/sec	 accuracy=2.765625	 loss=5.506411	 lr=0.010000
Epoch[000] Batch [0349]/[3759]	Speed: 34.552151 samples/sec	 accuracy=3.129464	 loss=5.434509	 lr=0.010000
Epoch[000] Batch [0399]/[3759]	Speed: 34.881526 samples/sec	 accuracy=3.421875	 loss=5.381165	 lr=0.010000
Epoch[000] Batch [0449]/[3759]	Speed: 34.713403 samples/sec	 accuracy=3.850694	 loss=5.324773	 lr=0.010000
Epoch[000] Batch [0499]/[3759]	Speed: 34.678624 samples/sec	 accuracy=4.193750	 loss=5.272362	 lr=0.010000
Epoch[000] Batch [0549]/[3759]	Speed: 34.696007 samples/sec	 accuracy=4.525568	 loss=5.225417	 lr=0.010000
Epoch[000] Batch [0599]/[3759]	Speed: 34.889778 samples/sec	 accuracy=4.828125	 loss=5.181585	 lr=0.010000
Epoch[000] Batch [0649]/[3759]	Speed: 34.767189 samples/sec	 accuracy=5.192308	 loss=5.142178	 lr=0.010000
Epoch[000] Batch [0699]/[3759]	Speed: 34.680802 samples/sec	 accuracy=5.479911	 loss=5.102238	 lr=0.010000
Epoch[000] Batch [0749]/[3759]	Speed: 34.689998 samples/sec	 accuracy=5.760417	 loss=5.066595	 lr=0.010000
Epoch[000] Batch [0799]/[3759]	Speed: 34.901050 samples/sec	 accuracy=6.062500	 loss=5.034108	 lr=0.010000
Epoch[000] Batch [0849]/[3759]	Speed: 34.736017 samples/sec	 accuracy=6.360294	 loss=4.998796	 lr=0.010000
Epoch[000] Batch [0899]/[3759]	Speed: 34.771863 samples/sec	 accuracy=6.651042	 loss=4.966479	 lr=0.010000
Epoch[000] Batch [0949]/[3759]	Speed: 34.669581 samples/sec	 accuracy=6.929276	 loss=4.939429	 lr=0.010000
Epoch[000] Batch [0999]/[3759]	Speed: 34.829667 samples/sec	 accuracy=7.195313	 loss=4.910995	 lr=0.010000
Epoch[000] Batch [1049]/[3759]	Speed: 34.606363 samples/sec	 accuracy=7.452381	 loss=4.884154	 lr=0.010000
Epoch[000] Batch [1099]/[3759]	Speed: 34.697252 samples/sec	 accuracy=7.738636	 loss=4.854211	 lr=0.010000
Epoch[000] Batch [1149]/[3759]	Speed: 34.711636 samples/sec	 accuracy=8.033967	 loss=4.828466	 lr=0.010000
Epoch[000] Batch [1199]/[3759]	Speed: 34.728519 samples/sec	 accuracy=8.261719	 loss=4.805078	 lr=0.010000
Epoch[000] Batch [1249]/[3759]	Speed: 34.818753 samples/sec	 accuracy=8.518750	 loss=4.780931	 lr=0.010000
Epoch[000] Batch [1299]/[3759]	Speed: 34.690800 samples/sec	 accuracy=8.719952	 loss=4.761305	 lr=0.010000
Epoch[000] Batch [1349]/[3759]	Speed: 34.749782 samples/sec	 accuracy=8.954861	 loss=4.739657	 lr=0.010000
Epoch[000] Batch [1399]/[3759]	Speed: 34.903695 samples/sec	 accuracy=9.141741	 loss=4.719020	 lr=0.010000
Epoch[000] Batch [1449]/[3759]	Speed: 34.679415 samples/sec	 accuracy=9.337284	 loss=4.701526	 lr=0.010000
Epoch[000] Batch [1499]/[3759]	Speed: 34.860648 samples/sec	 accuracy=9.532292	 loss=4.681656	 lr=0.010000
Epoch[000] Batch [1549]/[3759]	Speed: 34.674914 samples/sec	 accuracy=9.765121	 loss=4.661440	 lr=0.010000
Epoch[000] Batch [1599]/[3759]	Speed: 34.797984 samples/sec	 accuracy=9.988281	 loss=4.641744	 lr=0.010000
Epoch[000] Batch [1649]/[3759]	Speed: 34.853582 samples/sec	 accuracy=10.199811	 loss=4.622891	 lr=0.010000
Epoch[000] Batch [1699]/[3759]	Speed: 34.681273 samples/sec	 accuracy=10.418199	 loss=4.604846	 lr=0.010000
Epoch[000] Batch [1749]/[3759]	Speed: 34.936913 samples/sec	 accuracy=10.621429	 loss=4.586983	 lr=0.010000
Epoch[000] Batch [1799]/[3759]	Speed: 34.685657 samples/sec	 accuracy=10.825521	 loss=4.570967	 lr=0.010000
Epoch[000] Batch [1849]/[3759]	Speed: 35.010331 samples/sec	 accuracy=11.032095	 loss=4.554485	 lr=0.010000
Epoch[000] Batch [1899]/[3759]	Speed: 35.010412 samples/sec	 accuracy=11.216283	 loss=4.537243	 lr=0.010000
Epoch[000] Batch [1949]/[3759]	Speed: 35.066242 samples/sec	 accuracy=11.414263	 loss=4.519982	 lr=0.010000
Epoch[000] Batch [1999]/[3759]	Speed: 34.854634 samples/sec	 accuracy=11.553125	 loss=4.505902	 lr=0.010000
Epoch[000] Batch [2049]/[3759]	Speed: 34.856048 samples/sec	 accuracy=11.721037	 loss=4.491427	 lr=0.010000
Epoch[000] Batch [2099]/[3759]	Speed: 34.831929 samples/sec	 accuracy=11.901786	 loss=4.476111	 lr=0.010000
Epoch[000] Batch [2149]/[3759]	Speed: 34.959269 samples/sec	 accuracy=12.071948	 loss=4.461420	 lr=0.010000
Epoch[000] Batch [2199]/[3759]	Speed: 34.862897 samples/sec	 accuracy=12.268466	 loss=4.446696	 lr=0.010000
Epoch[000] Batch [2249]/[3759]	Speed: 34.744592 samples/sec	 accuracy=12.418750	 loss=4.432356	 lr=0.010000
Epoch[000] Batch [2299]/[3759]	Speed: 34.782643 samples/sec	 accuracy=12.565897	 loss=4.419788	 lr=0.010000
Epoch[000] Batch [2349]/[3759]	Speed: 35.085042 samples/sec	 accuracy=12.755319	 loss=4.405739	 lr=0.010000
Epoch[000] Batch [2399]/[3759]	Speed: 34.725036 samples/sec	 accuracy=12.916667	 loss=4.391509	 lr=0.010000
Epoch[000] Batch [2449]/[3759]	Speed: 34.917718 samples/sec	 accuracy=13.063776	 loss=4.378023	 lr=0.010000
Epoch[000] Batch [2499]/[3759]	Speed: 35.094841 samples/sec	 accuracy=13.214375	 loss=4.365331	 lr=0.010000
Epoch[000] Batch [2549]/[3759]	Speed: 34.717819 samples/sec	 accuracy=13.367034	 loss=4.354346	 lr=0.010000
Epoch[000] Batch [2599]/[3759]	Speed: 34.922004 samples/sec	 accuracy=13.504808	 loss=4.342394	 lr=0.010000
Epoch[000] Batch [2649]/[3759]	Speed: 34.846152 samples/sec	 accuracy=13.664505	 loss=4.329236	 lr=0.010000
Epoch[000] Batch [2699]/[3759]	Speed: 34.822405 samples/sec	 accuracy=13.804398	 loss=4.317859	 lr=0.010000
Epoch[000] Batch [2749]/[3759]	Speed: 34.650528 samples/sec	 accuracy=13.963636	 loss=4.305842	 lr=0.010000
Epoch[000] Batch [2799]/[3759]	Speed: 34.532859 samples/sec	 accuracy=14.101004	 loss=4.294300	 lr=0.010000
Epoch[000] Batch [2849]/[3759]	Speed: 34.766195 samples/sec	 accuracy=14.236294	 loss=4.282878	 lr=0.010000
Epoch[000] Batch [2899]/[3759]	Speed: 34.850417 samples/sec	 accuracy=14.379849	 loss=4.271593	 lr=0.010000
Epoch[000] Batch [2949]/[3759]	Speed: 34.783081 samples/sec	 accuracy=14.540254	 loss=4.259884	 lr=0.010000
Epoch[000] Batch [2999]/[3759]	Speed: 34.924436 samples/sec	 accuracy=14.689063	 loss=4.249163	 lr=0.010000
Epoch[000] Batch [3049]/[3759]	Speed: 34.801736 samples/sec	 accuracy=14.824795	 loss=4.238341	 lr=0.010000
Epoch[000] Batch [3099]/[3759]	Speed: 34.944483 samples/sec	 accuracy=14.949093	 loss=4.228641	 lr=0.010000
Epoch[000] Batch [3149]/[3759]	Speed: 34.851633 samples/sec	 accuracy=15.112599	 loss=4.217263	 lr=0.010000
Epoch[000] Batch [3199]/[3759]	Speed: 34.827080 samples/sec	 accuracy=15.251465	 loss=4.207290	 lr=0.010000
Epoch[000] Batch [3249]/[3759]	Speed: 35.094100 samples/sec	 accuracy=15.383173	 loss=4.197701	 lr=0.010000
Epoch[000] Batch [3299]/[3759]	Speed: 34.697941 samples/sec	 accuracy=15.507102	 loss=4.188553	 lr=0.010000
Epoch[000] Batch [3349]/[3759]	Speed: 34.946563 samples/sec	 accuracy=15.630597	 loss=4.179120	 lr=0.010000
Epoch[000] Batch [3399]/[3759]	Speed: 34.855621 samples/sec	 accuracy=15.755055	 loss=4.170026	 lr=0.010000
Epoch[000] Batch [3449]/[3759]	Speed: 34.943843 samples/sec	 accuracy=15.869565	 loss=4.161109	 lr=0.010000
Epoch[000] Batch [3499]/[3759]	Speed: 34.861828 samples/sec	 accuracy=15.987946	 loss=4.151575	 lr=0.010000
Epoch[000] Batch [3549]/[3759]	Speed: 34.850281 samples/sec	 accuracy=16.118838	 loss=4.142154	 lr=0.010000
Epoch[000] Batch [3599]/[3759]	Speed: 34.798682 samples/sec	 accuracy=16.228733	 loss=4.132711	 lr=0.010000
Epoch[000] Batch [3649]/[3759]	Speed: 34.766655 samples/sec	 accuracy=16.348459	 loss=4.123888	 lr=0.010000
Epoch[000] Batch [3699]/[3759]	Speed: 34.728921 samples/sec	 accuracy=16.478041	 loss=4.114273	 lr=0.010000
Epoch[000] Batch [3749]/[3759]	Speed: 36.344416 samples/sec	 accuracy=16.600833	 loss=4.106058	 lr=0.010000
Batch [0049]/[0303]: acc-top1=24.937500 acc-top5=51.312500
Batch [0099]/[0303]: acc-top1=25.265625 acc-top5=52.000000
Batch [0149]/[0303]: acc-top1=25.916667 acc-top5=52.302083
Batch [0199]/[0303]: acc-top1=25.593750 acc-top5=52.242188
Batch [0249]/[0303]: acc-top1=25.481250 acc-top5=52.037500
Batch [0299]/[0303]: acc-top1=25.479167 acc-top5=51.942708
[Epoch 000] training: accuracy=16.629672	 loss=4.104345
[Epoch 000] speed: 34 samples/sec	time cost: 7467.497211
[Epoch 000] validation: acc-top1=25.433168 acc-top5=51.918317 loss=3.831453
Epoch[001] Batch [0049]/[3760]	Speed: 22.678665 samples/sec	 accuracy=24.875000	 loss=3.444316	 lr=0.010000
Epoch[001] Batch [0099]/[3760]	Speed: 34.732525 samples/sec	 accuracy=25.062500	 loss=3.434450	 lr=0.010000
Epoch[001] Batch [0149]/[3760]	Speed: 34.975311 samples/sec	 accuracy=25.614583	 loss=3.414279	 lr=0.010000
Epoch[001] Batch [0199]/[3760]	Speed: 34.708811 samples/sec	 accuracy=25.835938	 loss=3.412627	 lr=0.010000
Epoch[001] Batch [0249]/[3760]	Speed: 35.081594 samples/sec	 accuracy=25.800000	 loss=3.411258	 lr=0.010000
Epoch[001] Batch [0299]/[3760]	Speed: 34.956471 samples/sec	 accuracy=25.744792	 loss=3.411499	 lr=0.010000
Epoch[001] Batch [0349]/[3760]	Speed: 34.955349 samples/sec	 accuracy=25.705357	 loss=3.405049	 lr=0.010000
Epoch[001] Batch [0399]/[3760]	Speed: 34.743027 samples/sec	 accuracy=25.972656	 loss=3.401123	 lr=0.010000
Epoch[001] Batch [0449]/[3760]	Speed: 35.043913 samples/sec	 accuracy=25.975694	 loss=3.406007	 lr=0.010000
Epoch[001] Batch [0499]/[3760]	Speed: 34.878099 samples/sec	 accuracy=26.034375	 loss=3.402132	 lr=0.010000
Epoch[001] Batch [0549]/[3760]	Speed: 34.890043 samples/sec	 accuracy=26.045455	 loss=3.401816	 lr=0.010000
Epoch[001] Batch [0599]/[3760]	Speed: 35.021644 samples/sec	 accuracy=26.049479	 loss=3.399029	 lr=0.010000
Epoch[001] Batch [0649]/[3760]	Speed: 34.935883 samples/sec	 accuracy=26.161058	 loss=3.394200	 lr=0.010000
Epoch[001] Batch [0699]/[3760]	Speed: 34.874856 samples/sec	 accuracy=26.308036	 loss=3.382827	 lr=0.010000
Epoch[001] Batch [0749]/[3760]	Speed: 35.016674 samples/sec	 accuracy=26.427083	 loss=3.377872	 lr=0.010000
Epoch[001] Batch [0799]/[3760]	Speed: 34.941201 samples/sec	 accuracy=26.550781	 loss=3.373640	 lr=0.010000
Epoch[001] Batch [0849]/[3760]	Speed: 34.952785 samples/sec	 accuracy=26.597426	 loss=3.368135	 lr=0.010000
Epoch[001] Batch [0899]/[3760]	Speed: 34.829814 samples/sec	 accuracy=26.791667	 loss=3.360610	 lr=0.010000
Epoch[001] Batch [0949]/[3760]	Speed: 35.043625 samples/sec	 accuracy=26.855263	 loss=3.355643	 lr=0.010000
Epoch[001] Batch [0999]/[3760]	Speed: 35.036843 samples/sec	 accuracy=26.878125	 loss=3.354513	 lr=0.010000
Epoch[001] Batch [1049]/[3760]	Speed: 34.794972 samples/sec	 accuracy=26.983631	 loss=3.350471	 lr=0.010000
Epoch[001] Batch [1099]/[3760]	Speed: 34.893655 samples/sec	 accuracy=27.012784	 loss=3.348506	 lr=0.010000
Epoch[001] Batch [1149]/[3760]	Speed: 34.927242 samples/sec	 accuracy=27.092391	 loss=3.344888	 lr=0.010000
Epoch[001] Batch [1199]/[3760]	Speed: 35.146778 samples/sec	 accuracy=27.182292	 loss=3.341635	 lr=0.010000
Epoch[001] Batch [1249]/[3760]	Speed: 34.845130 samples/sec	 accuracy=27.227500	 loss=3.338263	 lr=0.010000
Epoch[001] Batch [1299]/[3760]	Speed: 34.941855 samples/sec	 accuracy=27.316106	 loss=3.333544	 lr=0.010000
Epoch[001] Batch [1349]/[3760]	Speed: 34.957515 samples/sec	 accuracy=27.388889	 loss=3.329703	 lr=0.010000
Epoch[001] Batch [1399]/[3760]	Speed: 34.927254 samples/sec	 accuracy=27.487723	 loss=3.324675	 lr=0.010000
Epoch[001] Batch [1449]/[3760]	Speed: 34.980816 samples/sec	 accuracy=27.534483	 loss=3.321364	 lr=0.010000
Epoch[001] Batch [1499]/[3760]	Speed: 34.818821 samples/sec	 accuracy=27.596875	 loss=3.317231	 lr=0.010000
Epoch[001] Batch [1549]/[3760]	Speed: 34.932837 samples/sec	 accuracy=27.620968	 loss=3.314279	 lr=0.010000
Epoch[001] Batch [1599]/[3760]	Speed: 34.958049 samples/sec	 accuracy=27.685547	 loss=3.310644	 lr=0.010000
Epoch[001] Batch [1649]/[3760]	Speed: 35.062839 samples/sec	 accuracy=27.733902	 loss=3.306279	 lr=0.010000
Epoch[001] Batch [1699]/[3760]	Speed: 34.791718 samples/sec	 accuracy=27.789522	 loss=3.303031	 lr=0.010000
Epoch[001] Batch [1749]/[3760]	Speed: 34.744009 samples/sec	 accuracy=27.861607	 loss=3.299522	 lr=0.010000
Epoch[001] Batch [1799]/[3760]	Speed: 35.018849 samples/sec	 accuracy=27.966146	 loss=3.295261	 lr=0.010000
Epoch[001] Batch [1849]/[3760]	Speed: 34.954155 samples/sec	 accuracy=28.068412	 loss=3.290699	 lr=0.010000
Epoch[001] Batch [1899]/[3760]	Speed: 34.969558 samples/sec	 accuracy=28.138158	 loss=3.286648	 lr=0.010000
Epoch[001] Batch [1949]/[3760]	Speed: 34.903376 samples/sec	 accuracy=28.224359	 loss=3.284594	 lr=0.010000
Epoch[001] Batch [1999]/[3760]	Speed: 34.955422 samples/sec	 accuracy=28.245313	 loss=3.282959	 lr=0.010000
Epoch[001] Batch [2049]/[3760]	Speed: 34.961445 samples/sec	 accuracy=28.276677	 loss=3.279681	 lr=0.010000
Epoch[001] Batch [2099]/[3760]	Speed: 35.003437 samples/sec	 accuracy=28.337054	 loss=3.275315	 lr=0.010000
Epoch[001] Batch [2149]/[3760]	Speed: 34.969090 samples/sec	 accuracy=28.418605	 loss=3.271208	 lr=0.010000
Epoch[001] Batch [2199]/[3760]	Speed: 35.090569 samples/sec	 accuracy=28.485795	 loss=3.268110	 lr=0.010000
Epoch[001] Batch [2249]/[3760]	Speed: 34.717226 samples/sec	 accuracy=28.547917	 loss=3.265716	 lr=0.010000
Epoch[001] Batch [2299]/[3760]	Speed: 34.716079 samples/sec	 accuracy=28.600543	 loss=3.262650	 lr=0.010000
Epoch[001] Batch [2349]/[3760]	Speed: 34.942031 samples/sec	 accuracy=28.668218	 loss=3.259528	 lr=0.010000
Epoch[001] Batch [2399]/[3760]	Speed: 34.953660 samples/sec	 accuracy=28.740234	 loss=3.255630	 lr=0.010000
Epoch[001] Batch [2449]/[3760]	Speed: 34.900433 samples/sec	 accuracy=28.779337	 loss=3.253186	 lr=0.010000
Epoch[001] Batch [2499]/[3760]	Speed: 34.955514 samples/sec	 accuracy=28.818125	 loss=3.250723	 lr=0.010000
Epoch[001] Batch [2549]/[3760]	Speed: 34.759587 samples/sec	 accuracy=28.864583	 loss=3.247689	 lr=0.010000
Epoch[001] Batch [2599]/[3760]	Speed: 35.079615 samples/sec	 accuracy=28.902043	 loss=3.245193	 lr=0.010000
Epoch[001] Batch [2649]/[3760]	Speed: 34.915640 samples/sec	 accuracy=28.948703	 loss=3.243063	 lr=0.010000
Epoch[001] Batch [2699]/[3760]	Speed: 34.867151 samples/sec	 accuracy=29.005787	 loss=3.240731	 lr=0.010000
Epoch[001] Batch [2749]/[3760]	Speed: 34.917947 samples/sec	 accuracy=29.053409	 loss=3.237459	 lr=0.010000
Epoch[001] Batch [2799]/[3760]	Speed: 34.973982 samples/sec	 accuracy=29.128906	 loss=3.234386	 lr=0.010000
Epoch[001] Batch [2849]/[3760]	Speed: 34.914653 samples/sec	 accuracy=29.173246	 loss=3.232336	 lr=0.010000
Epoch[001] Batch [2899]/[3760]	Speed: 34.981926 samples/sec	 accuracy=29.221444	 loss=3.228957	 lr=0.010000
Epoch[001] Batch [2949]/[3760]	Speed: 34.983044 samples/sec	 accuracy=29.273835	 loss=3.225875	 lr=0.010000
Epoch[001] Batch [2999]/[3760]	Speed: 34.918718 samples/sec	 accuracy=29.330208	 loss=3.222851	 lr=0.010000
Epoch[001] Batch [3049]/[3760]	Speed: 35.029784 samples/sec	 accuracy=29.371926	 loss=3.220604	 lr=0.010000
Epoch[001] Batch [3099]/[3760]	Speed: 34.730294 samples/sec	 accuracy=29.396673	 loss=3.219501	 lr=0.010000
Epoch[001] Batch [3149]/[3760]	Speed: 34.996348 samples/sec	 accuracy=29.440972	 loss=3.216519	 lr=0.010000
Epoch[001] Batch [3199]/[3760]	Speed: 34.805540 samples/sec	 accuracy=29.495605	 loss=3.214048	 lr=0.010000
Epoch[001] Batch [3249]/[3760]	Speed: 35.001336 samples/sec	 accuracy=29.546635	 loss=3.211103	 lr=0.010000
Epoch[001] Batch [3299]/[3760]	Speed: 35.049981 samples/sec	 accuracy=29.606061	 loss=3.207546	 lr=0.010000
Epoch[001] Batch [3349]/[3760]	Speed: 35.042494 samples/sec	 accuracy=29.642257	 loss=3.205669	 lr=0.010000
Epoch[001] Batch [3399]/[3760]	Speed: 35.047371 samples/sec	 accuracy=29.697151	 loss=3.202572	 lr=0.010000
Epoch[001] Batch [3449]/[3760]	Speed: 34.825844 samples/sec	 accuracy=29.748641	 loss=3.199557	 lr=0.010000
Epoch[001] Batch [3499]/[3760]	Speed: 34.684390 samples/sec	 accuracy=29.783036	 loss=3.197860	 lr=0.010000
Epoch[001] Batch [3549]/[3760]	Speed: 34.999472 samples/sec	 accuracy=29.828785	 loss=3.195377	 lr=0.010000
Epoch[001] Batch [3599]/[3760]	Speed: 34.862775 samples/sec	 accuracy=29.877604	 loss=3.191880	 lr=0.010000
Epoch[001] Batch [3649]/[3760]	Speed: 34.956838 samples/sec	 accuracy=29.926798	 loss=3.189051	 lr=0.010000
Epoch[001] Batch [3699]/[3760]	Speed: 35.077918 samples/sec	 accuracy=29.965794	 loss=3.187387	 lr=0.010000
Epoch[001] Batch [3749]/[3760]	Speed: 36.253192 samples/sec	 accuracy=30.023750	 loss=3.184641	 lr=0.010000
Batch [0049]/[0303]: acc-top1=33.468750 acc-top5=62.718750
Batch [0099]/[0303]: acc-top1=34.578125 acc-top5=62.828125
Batch [0149]/[0303]: acc-top1=34.854167 acc-top5=63.302083
Batch [0199]/[0303]: acc-top1=34.968750 acc-top5=63.000000
Batch [0249]/[0303]: acc-top1=34.806250 acc-top5=62.912500
Batch [0299]/[0303]: acc-top1=34.953125 acc-top5=62.854167
[Epoch 001] training: accuracy=30.031998	 loss=3.184140
[Epoch 001] speed: 34 samples/sec	time cost: 7423.448955
[Epoch 001] validation: acc-top1=34.916460 acc-top5=62.819719 loss=3.086478
Epoch[002] Batch [0049]/[3759]	Speed: 23.512267 samples/sec	 accuracy=34.812500	 loss=2.971202	 lr=0.010000
Epoch[002] Batch [0099]/[3759]	Speed: 34.774597 samples/sec	 accuracy=34.781250	 loss=2.927978	 lr=0.010000
Epoch[002] Batch [0149]/[3759]	Speed: 35.108549 samples/sec	 accuracy=34.833333	 loss=2.923899	 lr=0.010000
Epoch[002] Batch [0199]/[3759]	Speed: 34.914359 samples/sec	 accuracy=34.851562	 loss=2.911677	 lr=0.010000
Epoch[002] Batch [0249]/[3759]	Speed: 34.877345 samples/sec	 accuracy=34.975000	 loss=2.907134	 lr=0.010000
Epoch[002] Batch [0299]/[3759]	Speed: 35.229735 samples/sec	 accuracy=34.989583	 loss=2.916665	 lr=0.010000
Epoch[002] Batch [0349]/[3759]	Speed: 35.097185 samples/sec	 accuracy=34.991071	 loss=2.907898	 lr=0.010000
Epoch[002] Batch [0399]/[3759]	Speed: 35.120391 samples/sec	 accuracy=35.054688	 loss=2.908739	 lr=0.010000
Epoch[002] Batch [0449]/[3759]	Speed: 35.027693 samples/sec	 accuracy=35.065972	 loss=2.905384	 lr=0.010000
Epoch[002] Batch [0499]/[3759]	Speed: 35.050590 samples/sec	 accuracy=35.153125	 loss=2.899952	 lr=0.010000
Epoch[002] Batch [0549]/[3759]	Speed: 35.019417 samples/sec	 accuracy=35.187500	 loss=2.899246	 lr=0.010000
Epoch[002] Batch [0599]/[3759]	Speed: 34.970538 samples/sec	 accuracy=35.101562	 loss=2.903092	 lr=0.010000
Epoch[002] Batch [0649]/[3759]	Speed: 35.033971 samples/sec	 accuracy=35.050481	 loss=2.900057	 lr=0.010000
Epoch[002] Batch [0699]/[3759]	Speed: 35.098556 samples/sec	 accuracy=34.968750	 loss=2.904053	 lr=0.010000
Epoch[002] Batch [0749]/[3759]	Speed: 35.021940 samples/sec	 accuracy=34.966667	 loss=2.901446	 lr=0.010000
Epoch[002] Batch [0799]/[3759]	Speed: 35.179035 samples/sec	 accuracy=34.978516	 loss=2.898931	 lr=0.010000
Epoch[002] Batch [0849]/[3759]	Speed: 34.955295 samples/sec	 accuracy=34.994485	 loss=2.900202	 lr=0.010000
Epoch[002] Batch [0899]/[3759]	Speed: 35.033931 samples/sec	 accuracy=35.017361	 loss=2.898992	 lr=0.010000
Epoch[002] Batch [0949]/[3759]	Speed: 35.262260 samples/sec	 accuracy=35.049342	 loss=2.896229	 lr=0.010000
Epoch[002] Batch [0999]/[3759]	Speed: 34.978220 samples/sec	 accuracy=35.092188	 loss=2.893954	 lr=0.010000
Epoch[002] Batch [1049]/[3759]	Speed: 34.941766 samples/sec	 accuracy=35.157738	 loss=2.891822	 lr=0.010000
Epoch[002] Batch [1099]/[3759]	Speed: 35.037782 samples/sec	 accuracy=35.190341	 loss=2.889022	 lr=0.010000
Epoch[002] Batch [1149]/[3759]	Speed: 35.137190 samples/sec	 accuracy=35.230978	 loss=2.888005	 lr=0.010000
Epoch[002] Batch [1199]/[3759]	Speed: 35.183662 samples/sec	 accuracy=35.226562	 loss=2.889256	 lr=0.010000
Epoch[002] Batch [1249]/[3759]	Speed: 34.836505 samples/sec	 accuracy=35.216250	 loss=2.891814	 lr=0.010000
Epoch[002] Batch [1299]/[3759]	Speed: 35.288943 samples/sec	 accuracy=35.187500	 loss=2.892378	 lr=0.010000
Epoch[002] Batch [1349]/[3759]	Speed: 35.062349 samples/sec	 accuracy=35.236111	 loss=2.890109	 lr=0.010000
Epoch[002] Batch [1399]/[3759]	Speed: 35.293737 samples/sec	 accuracy=35.314732	 loss=2.888151	 lr=0.010000
Epoch[002] Batch [1449]/[3759]	Speed: 35.000132 samples/sec	 accuracy=35.288793	 loss=2.889034	 lr=0.010000
Epoch[002] Batch [1499]/[3759]	Speed: 35.109908 samples/sec	 accuracy=35.305208	 loss=2.887023	 lr=0.010000
Epoch[002] Batch [1549]/[3759]	Speed: 35.150745 samples/sec	 accuracy=35.367944	 loss=2.884499	 lr=0.010000
Epoch[002] Batch [1599]/[3759]	Speed: 35.030530 samples/sec	 accuracy=35.371094	 loss=2.883876	 lr=0.010000
Epoch[002] Batch [1649]/[3759]	Speed: 34.936593 samples/sec	 accuracy=35.387311	 loss=2.883054	 lr=0.010000
Epoch[002] Batch [1699]/[3759]	Speed: 35.014571 samples/sec	 accuracy=35.448529	 loss=2.880046	 lr=0.010000
Epoch[002] Batch [1749]/[3759]	Speed: 35.164047 samples/sec	 accuracy=35.498214	 loss=2.877462	 lr=0.010000
Epoch[002] Batch [1799]/[3759]	Speed: 35.072818 samples/sec	 accuracy=35.530382	 loss=2.876102	 lr=0.010000
Epoch[002] Batch [1849]/[3759]	Speed: 34.974581 samples/sec	 accuracy=35.573480	 loss=2.873999	 lr=0.010000
Epoch[002] Batch [1899]/[3759]	Speed: 35.216592 samples/sec	 accuracy=35.617599	 loss=2.871781	 lr=0.010000
Epoch[002] Batch [1949]/[3759]	Speed: 35.163603 samples/sec	 accuracy=35.669071	 loss=2.869457	 lr=0.010000
Epoch[002] Batch [1999]/[3759]	Speed: 35.023340 samples/sec	 accuracy=35.657813	 loss=2.869742	 lr=0.010000
Epoch[002] Batch [2049]/[3759]	Speed: 35.030269 samples/sec	 accuracy=35.649390	 loss=2.868401	 lr=0.010000
Epoch[002] Batch [2099]/[3759]	Speed: 34.989872 samples/sec	 accuracy=35.686012	 loss=2.866508	 lr=0.010000
Epoch[002] Batch [2149]/[3759]	Speed: 35.155979 samples/sec	 accuracy=35.732558	 loss=2.864745	 lr=0.010000
Epoch[002] Batch [2199]/[3759]	Speed: 35.030711 samples/sec	 accuracy=35.750000	 loss=2.863799	 lr=0.010000
Epoch[002] Batch [2249]/[3759]	Speed: 34.915679 samples/sec	 accuracy=35.750000	 loss=2.863038	 lr=0.010000
Epoch[002] Batch [2299]/[3759]	Speed: 34.922223 samples/sec	 accuracy=35.798234	 loss=2.861269	 lr=0.010000
Epoch[002] Batch [2349]/[3759]	Speed: 35.010538 samples/sec	 accuracy=35.811835	 loss=2.859946	 lr=0.010000
Epoch[002] Batch [2399]/[3759]	Speed: 35.136469 samples/sec	 accuracy=35.809245	 loss=2.858755	 lr=0.010000
Epoch[002] Batch [2449]/[3759]	Speed: 34.840457 samples/sec	 accuracy=35.822066	 loss=2.857994	 lr=0.010000
Epoch[002] Batch [2499]/[3759]	Speed: 34.879107 samples/sec	 accuracy=35.864375	 loss=2.857051	 lr=0.010000
Epoch[002] Batch [2549]/[3759]	Speed: 34.940999 samples/sec	 accuracy=35.870098	 loss=2.856112	 lr=0.010000
Epoch[002] Batch [2599]/[3759]	Speed: 34.969537 samples/sec	 accuracy=35.903846	 loss=2.855310	 lr=0.010000
Epoch[002] Batch [2649]/[3759]	Speed: 35.140460 samples/sec	 accuracy=35.925708	 loss=2.854476	 lr=0.010000
Epoch[002] Batch [2699]/[3759]	Speed: 35.024997 samples/sec	 accuracy=35.920139	 loss=2.855254	 lr=0.010000
Epoch[002] Batch [2749]/[3759]	Speed: 34.851769 samples/sec	 accuracy=35.932386	 loss=2.855154	 lr=0.010000
Epoch[002] Batch [2799]/[3759]	Speed: 34.965573 samples/sec	 accuracy=35.964286	 loss=2.854399	 lr=0.010000
Epoch[002] Batch [2849]/[3759]	Speed: 34.908546 samples/sec	 accuracy=35.992873	 loss=2.853379	 lr=0.010000
Epoch[002] Batch [2899]/[3759]	Speed: 35.000815 samples/sec	 accuracy=36.035022	 loss=2.851426	 lr=0.010000
Epoch[002] Batch [2949]/[3759]	Speed: 34.794851 samples/sec	 accuracy=36.067797	 loss=2.849709	 lr=0.010000
Epoch[002] Batch [2999]/[3759]	Speed: 35.114484 samples/sec	 accuracy=36.082813	 loss=2.848641	 lr=0.010000
Epoch[002] Batch [3049]/[3759]	Speed: 34.959690 samples/sec	 accuracy=36.111680	 loss=2.846804	 lr=0.010000
Epoch[002] Batch [3099]/[3759]	Speed: 34.982382 samples/sec	 accuracy=36.154234	 loss=2.845133	 lr=0.010000
Epoch[002] Batch [3149]/[3759]	Speed: 34.875985 samples/sec	 accuracy=36.197917	 loss=2.843071	 lr=0.010000
Epoch[002] Batch [3199]/[3759]	Speed: 34.979813 samples/sec	 accuracy=36.219727	 loss=2.841479	 lr=0.010000
Epoch[002] Batch [3249]/[3759]	Speed: 34.931073 samples/sec	 accuracy=36.230769	 loss=2.840673	 lr=0.010000
Epoch[002] Batch [3299]/[3759]	Speed: 34.838154 samples/sec	 accuracy=36.235322	 loss=2.840793	 lr=0.010000
Epoch[002] Batch [3349]/[3759]	Speed: 34.804749 samples/sec	 accuracy=36.268657	 loss=2.838759	 lr=0.010000
Epoch[002] Batch [3399]/[3759]	Speed: 35.149278 samples/sec	 accuracy=36.311581	 loss=2.836642	 lr=0.010000
Epoch[002] Batch [3449]/[3759]	Speed: 34.757219 samples/sec	 accuracy=36.350543	 loss=2.834627	 lr=0.010000
Epoch[002] Batch [3499]/[3759]	Speed: 34.944818 samples/sec	 accuracy=36.364286	 loss=2.833367	 lr=0.010000
Epoch[002] Batch [3549]/[3759]	Speed: 34.853961 samples/sec	 accuracy=36.396127	 loss=2.831995	 lr=0.010000
Epoch[002] Batch [3599]/[3759]	Speed: 34.948118 samples/sec	 accuracy=36.437066	 loss=2.830204	 lr=0.010000
Epoch[002] Batch [3649]/[3759]	Speed: 34.862090 samples/sec	 accuracy=36.470462	 loss=2.828802	 lr=0.010000
Epoch[002] Batch [3699]/[3759]	Speed: 34.761919 samples/sec	 accuracy=36.489865	 loss=2.828028	 lr=0.010000
Epoch[002] Batch [3749]/[3759]	Speed: 36.305498 samples/sec	 accuracy=36.508750	 loss=2.826691	 lr=0.010000
Batch [0049]/[0303]: acc-top1=39.750000 acc-top5=67.156250
Batch [0099]/[0303]: acc-top1=40.156250 acc-top5=67.625000
Batch [0149]/[0303]: acc-top1=41.062500 acc-top5=68.395833
Batch [0199]/[0303]: acc-top1=40.992188 acc-top5=68.343750
Batch [0249]/[0303]: acc-top1=40.825000 acc-top5=67.943750
Batch [0299]/[0303]: acc-top1=40.812500 acc-top5=67.921875
[Epoch 002] training: accuracy=36.513617	 loss=2.826688
[Epoch 002] speed: 34 samples/sec	time cost: 7399.539416
[Epoch 002] validation: acc-top1=40.790017 acc-top5=67.863036 loss=2.862572
Epoch[003] Batch [0049]/[3760]	Speed: 23.767791 samples/sec	 accuracy=38.500000	 loss=2.666803	 lr=0.010000
Epoch[003] Batch [0099]/[3760]	Speed: 34.706919 samples/sec	 accuracy=39.390625	 loss=2.629270	 lr=0.010000
Epoch[003] Batch [0149]/[3760]	Speed: 35.229353 samples/sec	 accuracy=39.531250	 loss=2.632650	 lr=0.010000
Epoch[003] Batch [0199]/[3760]	Speed: 34.775461 samples/sec	 accuracy=39.585938	 loss=2.641015	 lr=0.010000
Epoch[003] Batch [0249]/[3760]	Speed: 34.947437 samples/sec	 accuracy=39.556250	 loss=2.641045	 lr=0.010000
Epoch[003] Batch [0299]/[3760]	Speed: 34.932623 samples/sec	 accuracy=39.942708	 loss=2.626787	 lr=0.010000
Epoch[003] Batch [0349]/[3760]	Speed: 35.189654 samples/sec	 accuracy=39.986607	 loss=2.625360	 lr=0.010000
Epoch[003] Batch [0399]/[3760]	Speed: 35.004415 samples/sec	 accuracy=40.136719	 loss=2.627465	 lr=0.010000
Epoch[003] Batch [0449]/[3760]	Speed: 34.917724 samples/sec	 accuracy=39.947917	 loss=2.633174	 lr=0.010000
Epoch[003] Batch [0499]/[3760]	Speed: 35.201988 samples/sec	 accuracy=39.968750	 loss=2.631885	 lr=0.010000
Epoch[003] Batch [0549]/[3760]	Speed: 34.890949 samples/sec	 accuracy=39.889205	 loss=2.636596	 lr=0.010000
Epoch[003] Batch [0599]/[3760]	Speed: 35.114522 samples/sec	 accuracy=39.864583	 loss=2.639368	 lr=0.010000
Epoch[003] Batch [0649]/[3760]	Speed: 34.825531 samples/sec	 accuracy=39.954327	 loss=2.634629	 lr=0.010000
Epoch[003] Batch [0699]/[3760]	Speed: 35.100868 samples/sec	 accuracy=39.924107	 loss=2.636474	 lr=0.010000
Epoch[003] Batch [0749]/[3760]	Speed: 35.012173 samples/sec	 accuracy=39.860417	 loss=2.639187	 lr=0.010000
Epoch[003] Batch [0799]/[3760]	Speed: 34.881434 samples/sec	 accuracy=39.828125	 loss=2.640522	 lr=0.010000
Epoch[003] Batch [0849]/[3760]	Speed: 35.084530 samples/sec	 accuracy=39.889706	 loss=2.637305	 lr=0.010000
Epoch[003] Batch [0899]/[3760]	Speed: 34.915247 samples/sec	 accuracy=39.866319	 loss=2.640540	 lr=0.010000
Epoch[003] Batch [0949]/[3760]	Speed: 34.940234 samples/sec	 accuracy=39.842105	 loss=2.643431	 lr=0.010000
Epoch[003] Batch [0999]/[3760]	Speed: 34.985482 samples/sec	 accuracy=39.829687	 loss=2.642594	 lr=0.010000
Epoch[003] Batch [1049]/[3760]	Speed: 34.819183 samples/sec	 accuracy=39.799107	 loss=2.643970	 lr=0.010000
Epoch[003] Batch [1099]/[3760]	Speed: 35.021804 samples/sec	 accuracy=39.801136	 loss=2.644757	 lr=0.010000
Epoch[003] Batch [1149]/[3760]	Speed: 34.762249 samples/sec	 accuracy=39.783967	 loss=2.642961	 lr=0.010000
Epoch[003] Batch [1199]/[3760]	Speed: 35.126369 samples/sec	 accuracy=39.763021	 loss=2.644400	 lr=0.010000
Epoch[003] Batch [1249]/[3760]	Speed: 35.115837 samples/sec	 accuracy=39.760000	 loss=2.642341	 lr=0.010000
Epoch[003] Batch [1299]/[3760]	Speed: 35.205276 samples/sec	 accuracy=39.783654	 loss=2.639570	 lr=0.010000
Epoch[003] Batch [1349]/[3760]	Speed: 34.969483 samples/sec	 accuracy=39.828704	 loss=2.637515	 lr=0.010000
Epoch[003] Batch [1399]/[3760]	Speed: 35.028194 samples/sec	 accuracy=39.813616	 loss=2.639256	 lr=0.010000
Epoch[003] Batch [1449]/[3760]	Speed: 35.172868 samples/sec	 accuracy=39.798491	 loss=2.641202	 lr=0.010000
Epoch[003] Batch [1499]/[3760]	Speed: 35.164757 samples/sec	 accuracy=39.771875	 loss=2.641566	 lr=0.010000
Epoch[003] Batch [1549]/[3760]	Speed: 35.007800 samples/sec	 accuracy=39.768145	 loss=2.641378	 lr=0.010000
Epoch[003] Batch [1599]/[3760]	Speed: 35.079789 samples/sec	 accuracy=39.756836	 loss=2.642487	 lr=0.010000
Epoch[003] Batch [1649]/[3760]	Speed: 35.128281 samples/sec	 accuracy=39.786932	 loss=2.642129	 lr=0.010000
Epoch[003] Batch [1699]/[3760]	Speed: 35.099823 samples/sec	 accuracy=39.835478	 loss=2.641622	 lr=0.010000
Epoch[003] Batch [1749]/[3760]	Speed: 35.099946 samples/sec	 accuracy=39.875893	 loss=2.639161	 lr=0.010000
Epoch[003] Batch [1799]/[3760]	Speed: 35.054088 samples/sec	 accuracy=39.913194	 loss=2.639329	 lr=0.010000
Epoch[003] Batch [1849]/[3760]	Speed: 34.873701 samples/sec	 accuracy=39.920608	 loss=2.638900	 lr=0.010000
Epoch[003] Batch [1899]/[3760]	Speed: 35.097728 samples/sec	 accuracy=39.939145	 loss=2.637695	 lr=0.010000
Epoch[003] Batch [1949]/[3760]	Speed: 35.106262 samples/sec	 accuracy=39.924679	 loss=2.638261	 lr=0.010000
Epoch[003] Batch [1999]/[3760]	Speed: 34.922656 samples/sec	 accuracy=39.941406	 loss=2.637178	 lr=0.010000
Epoch[003] Batch [2049]/[3760]	Speed: 35.028489 samples/sec	 accuracy=39.939024	 loss=2.636622	 lr=0.010000
Epoch[003] Batch [2099]/[3760]	Speed: 34.820615 samples/sec	 accuracy=39.921875	 loss=2.637765	 lr=0.010000
Epoch[003] Batch [2149]/[3760]	Speed: 35.037594 samples/sec	 accuracy=39.960029	 loss=2.636213	 lr=0.010000
Epoch[003] Batch [2199]/[3760]	Speed: 34.937564 samples/sec	 accuracy=40.006392	 loss=2.633912	 lr=0.010000
Epoch[003] Batch [2249]/[3760]	Speed: 34.947834 samples/sec	 accuracy=40.032639	 loss=2.632416	 lr=0.010000
Epoch[003] Batch [2299]/[3760]	Speed: 35.093645 samples/sec	 accuracy=40.055027	 loss=2.631378	 lr=0.010000
Epoch[003] Batch [2349]/[3760]	Speed: 34.912705 samples/sec	 accuracy=40.051197	 loss=2.630707	 lr=0.010000
Epoch[003] Batch [2399]/[3760]	Speed: 35.283366 samples/sec	 accuracy=40.072917	 loss=2.629982	 lr=0.010000
Epoch[003] Batch [2449]/[3760]	Speed: 34.769291 samples/sec	 accuracy=40.082908	 loss=2.628831	 lr=0.010000
Epoch[003] Batch [2499]/[3760]	Speed: 35.030532 samples/sec	 accuracy=40.118750	 loss=2.627446	 lr=0.010000
Epoch[003] Batch [2549]/[3760]	Speed: 35.107760 samples/sec	 accuracy=40.143382	 loss=2.626165	 lr=0.010000
Epoch[003] Batch [2599]/[3760]	Speed: 34.926022 samples/sec	 accuracy=40.169471	 loss=2.625588	 lr=0.010000
Epoch[003] Batch [2649]/[3760]	Speed: 35.153557 samples/sec	 accuracy=40.174528	 loss=2.625690	 lr=0.010000
Epoch[003] Batch [2699]/[3760]	Speed: 34.935082 samples/sec	 accuracy=40.226273	 loss=2.623990	 lr=0.010000
Epoch[003] Batch [2749]/[3760]	Speed: 34.925260 samples/sec	 accuracy=40.250000	 loss=2.623027	 lr=0.010000
Epoch[003] Batch [2799]/[3760]	Speed: 35.177440 samples/sec	 accuracy=40.247210	 loss=2.622384	 lr=0.010000
Epoch[003] Batch [2849]/[3760]	Speed: 34.790985 samples/sec	 accuracy=40.274123	 loss=2.620826	 lr=0.010000
Epoch[003] Batch [2899]/[3760]	Speed: 35.160450 samples/sec	 accuracy=40.284483	 loss=2.620395	 lr=0.010000
Epoch[003] Batch [2949]/[3760]	Speed: 34.755666 samples/sec	 accuracy=40.322034	 loss=2.618793	 lr=0.010000
Epoch[003] Batch [2999]/[3760]	Speed: 35.079406 samples/sec	 accuracy=40.336979	 loss=2.618030	 lr=0.010000
Epoch[003] Batch [3049]/[3760]	Speed: 35.173006 samples/sec	 accuracy=40.355020	 loss=2.617077	 lr=0.010000
Epoch[003] Batch [3099]/[3760]	Speed: 34.899309 samples/sec	 accuracy=40.338206	 loss=2.617375	 lr=0.010000
Epoch[003] Batch [3149]/[3760]	Speed: 35.038432 samples/sec	 accuracy=40.357639	 loss=2.616303	 lr=0.010000
Epoch[003] Batch [3199]/[3760]	Speed: 34.797775 samples/sec	 accuracy=40.379395	 loss=2.615663	 lr=0.010000
Epoch[003] Batch [3249]/[3760]	Speed: 35.144241 samples/sec	 accuracy=40.411538	 loss=2.613673	 lr=0.010000
Epoch[003] Batch [3299]/[3760]	Speed: 34.982608 samples/sec	 accuracy=40.435606	 loss=2.612804	 lr=0.010000
Epoch[003] Batch [3349]/[3760]	Speed: 34.861552 samples/sec	 accuracy=40.437500	 loss=2.612851	 lr=0.010000
Epoch[003] Batch [3399]/[3760]	Speed: 35.055236 samples/sec	 accuracy=40.456801	 loss=2.611686	 lr=0.010000
Epoch[003] Batch [3449]/[3760]	Speed: 34.832256 samples/sec	 accuracy=40.458786	 loss=2.612181	 lr=0.010000
Epoch[003] Batch [3499]/[3760]	Speed: 35.145983 samples/sec	 accuracy=40.468304	 loss=2.612634	 lr=0.010000
Epoch[003] Batch [3549]/[3760]	Speed: 34.689835 samples/sec	 accuracy=40.499560	 loss=2.612085	 lr=0.010000
Epoch[003] Batch [3599]/[3760]	Speed: 35.263757 samples/sec	 accuracy=40.517795	 loss=2.611328	 lr=0.010000
Epoch[003] Batch [3649]/[3760]	Speed: 34.864708 samples/sec	 accuracy=40.525685	 loss=2.610373	 lr=0.010000
Epoch[003] Batch [3699]/[3760]	Speed: 35.163911 samples/sec	 accuracy=40.549409	 loss=2.609276	 lr=0.010000
Epoch[003] Batch [3749]/[3760]	Speed: 36.263929 samples/sec	 accuracy=40.539167	 loss=2.608873	 lr=0.010000
Batch [0049]/[0303]: acc-top1=42.187500 acc-top5=70.468750
Batch [0099]/[0303]: acc-top1=42.812500 acc-top5=70.468750
Batch [0149]/[0303]: acc-top1=43.145833 acc-top5=70.802083
Batch [0199]/[0303]: acc-top1=43.171875 acc-top5=70.554688
Batch [0249]/[0303]: acc-top1=42.912500 acc-top5=70.256250
Batch [0299]/[0303]: acc-top1=42.880208 acc-top5=70.161458
[Epoch 003] training: accuracy=40.543135	 loss=2.608784
[Epoch 003] speed: 34 samples/sec	time cost: 7399.763301
[Epoch 003] validation: acc-top1=42.847566 acc-top5=70.147483 loss=2.748762
Epoch[004] Batch [0049]/[3760]	Speed: 24.063222 samples/sec	 accuracy=43.312500	 loss=2.428884	 lr=0.010000
Epoch[004] Batch [0099]/[3760]	Speed: 34.846363 samples/sec	 accuracy=43.093750	 loss=2.461940	 lr=0.010000
Epoch[004] Batch [0149]/[3760]	Speed: 35.133567 samples/sec	 accuracy=43.489583	 loss=2.456059	 lr=0.010000
Epoch[004] Batch [0199]/[3760]	Speed: 34.952187 samples/sec	 accuracy=43.328125	 loss=2.462652	 lr=0.010000
Epoch[004] Batch [0249]/[3760]	Speed: 35.131967 samples/sec	 accuracy=43.375000	 loss=2.466302	 lr=0.010000
Epoch[004] Batch [0299]/[3760]	Speed: 34.852535 samples/sec	 accuracy=43.468750	 loss=2.469413	 lr=0.010000
Epoch[004] Batch [0349]/[3760]	Speed: 35.089326 samples/sec	 accuracy=43.392857	 loss=2.476441	 lr=0.010000
Epoch[004] Batch [0399]/[3760]	Speed: 35.027156 samples/sec	 accuracy=43.304688	 loss=2.479271	 lr=0.010000
Epoch[004] Batch [0449]/[3760]	Speed: 34.700323 samples/sec	 accuracy=43.256944	 loss=2.476621	 lr=0.010000
Epoch[004] Batch [0499]/[3760]	Speed: 35.325538 samples/sec	 accuracy=43.300000	 loss=2.474773	 lr=0.010000
Epoch[004] Batch [0549]/[3760]	Speed: 34.813692 samples/sec	 accuracy=43.173295	 loss=2.477157	 lr=0.010000
Epoch[004] Batch [0599]/[3760]	Speed: 35.095739 samples/sec	 accuracy=43.135417	 loss=2.480614	 lr=0.010000
Epoch[004] Batch [0649]/[3760]	Speed: 34.962680 samples/sec	 accuracy=43.134615	 loss=2.478869	 lr=0.010000
Epoch[004] Batch [0699]/[3760]	Speed: 35.177150 samples/sec	 accuracy=43.120536	 loss=2.477989	 lr=0.010000
Epoch[004] Batch [0749]/[3760]	Speed: 34.967813 samples/sec	 accuracy=43.122917	 loss=2.478955	 lr=0.010000
Epoch[004] Batch [0799]/[3760]	Speed: 35.217452 samples/sec	 accuracy=43.099609	 loss=2.480010	 lr=0.010000
Epoch[004] Batch [0849]/[3760]	Speed: 35.233682 samples/sec	 accuracy=43.165441	 loss=2.478551	 lr=0.010000
Epoch[004] Batch [0899]/[3760]	Speed: 35.116494 samples/sec	 accuracy=43.135417	 loss=2.480671	 lr=0.010000
Epoch[004] Batch [0949]/[3760]	Speed: 35.099284 samples/sec	 accuracy=43.200658	 loss=2.477820	 lr=0.010000
Epoch[004] Batch [0999]/[3760]	Speed: 35.149223 samples/sec	 accuracy=43.171875	 loss=2.478624	 lr=0.010000
Epoch[004] Batch [1049]/[3760]	Speed: 35.145585 samples/sec	 accuracy=43.151786	 loss=2.480889	 lr=0.010000
Epoch[004] Batch [1099]/[3760]	Speed: 35.052392 samples/sec	 accuracy=43.093750	 loss=2.484895	 lr=0.010000
Epoch[004] Batch [1149]/[3760]	Speed: 34.957070 samples/sec	 accuracy=43.146739	 loss=2.482526	 lr=0.010000
Epoch[004] Batch [1199]/[3760]	Speed: 35.083144 samples/sec	 accuracy=43.130208	 loss=2.483482	 lr=0.010000
Epoch[004] Batch [1249]/[3760]	Speed: 34.931485 samples/sec	 accuracy=43.175000	 loss=2.481016	 lr=0.010000
Epoch[004] Batch [1299]/[3760]	Speed: 35.127985 samples/sec	 accuracy=43.134615	 loss=2.481007	 lr=0.010000
Epoch[004] Batch [1349]/[3760]	Speed: 35.190727 samples/sec	 accuracy=43.104167	 loss=2.483012	 lr=0.010000
Epoch[004] Batch [1399]/[3760]	Speed: 34.918592 samples/sec	 accuracy=43.094866	 loss=2.484436	 lr=0.010000
Epoch[004] Batch [1449]/[3760]	Speed: 35.242418 samples/sec	 accuracy=43.100216	 loss=2.481867	 lr=0.010000
Epoch[004] Batch [1499]/[3760]	Speed: 34.813656 samples/sec	 accuracy=43.081250	 loss=2.482302	 lr=0.010000
Epoch[004] Batch [1549]/[3760]	Speed: 35.427980 samples/sec	 accuracy=43.097782	 loss=2.482199	 lr=0.010000
Epoch[004] Batch [1599]/[3760]	Speed: 35.060719 samples/sec	 accuracy=43.089844	 loss=2.483052	 lr=0.010000
Epoch[004] Batch [1649]/[3760]	Speed: 35.259054 samples/sec	 accuracy=43.093750	 loss=2.482639	 lr=0.010000
Epoch[004] Batch [1699]/[3760]	Speed: 35.066690 samples/sec	 accuracy=43.102941	 loss=2.482642	 lr=0.010000
Epoch[004] Batch [1749]/[3760]	Speed: 35.230519 samples/sec	 accuracy=43.092857	 loss=2.483356	 lr=0.010000
Epoch[004] Batch [1799]/[3760]	Speed: 35.279077 samples/sec	 accuracy=43.130208	 loss=2.480119	 lr=0.010000
Epoch[004] Batch [1849]/[3760]	Speed: 35.033480 samples/sec	 accuracy=43.119932	 loss=2.480183	 lr=0.010000
Epoch[004] Batch [1899]/[3760]	Speed: 35.277322 samples/sec	 accuracy=43.080592	 loss=2.481236	 lr=0.010000
Epoch[004] Batch [1949]/[3760]	Speed: 35.228512 samples/sec	 accuracy=43.096154	 loss=2.481961	 lr=0.010000
Epoch[004] Batch [1999]/[3760]	Speed: 34.969563 samples/sec	 accuracy=43.113281	 loss=2.481315	 lr=0.010000
Epoch[004] Batch [2049]/[3760]	Speed: 34.880742 samples/sec	 accuracy=43.077744	 loss=2.482537	 lr=0.010000
Epoch[004] Batch [2099]/[3760]	Speed: 35.009320 samples/sec	 accuracy=43.092262	 loss=2.482030	 lr=0.010000
Epoch[004] Batch [2149]/[3760]	Speed: 35.177562 samples/sec	 accuracy=43.109012	 loss=2.480914	 lr=0.010000
Epoch[004] Batch [2199]/[3760]	Speed: 34.968496 samples/sec	 accuracy=43.139915	 loss=2.480194	 lr=0.010000
Epoch[004] Batch [2249]/[3760]	Speed: 34.785321 samples/sec	 accuracy=43.158333	 loss=2.479171	 lr=0.010000
Epoch[004] Batch [2299]/[3760]	Speed: 35.144088 samples/sec	 accuracy=43.137908	 loss=2.480071	 lr=0.010000
Epoch[004] Batch [2349]/[3760]	Speed: 34.823151 samples/sec	 accuracy=43.146941	 loss=2.479572	 lr=0.010000
Epoch[004] Batch [2399]/[3760]	Speed: 35.190048 samples/sec	 accuracy=43.143229	 loss=2.479438	 lr=0.010000
Epoch[004] Batch [2449]/[3760]	Speed: 34.906889 samples/sec	 accuracy=43.160714	 loss=2.478863	 lr=0.010000
Epoch[004] Batch [2499]/[3760]	Speed: 35.271980 samples/sec	 accuracy=43.173125	 loss=2.477703	 lr=0.010000
Epoch[004] Batch [2549]/[3760]	Speed: 34.929146 samples/sec	 accuracy=43.188113	 loss=2.476881	 lr=0.010000
Epoch[004] Batch [2599]/[3760]	Speed: 35.253738 samples/sec	 accuracy=43.214543	 loss=2.474839	 lr=0.010000
Epoch[004] Batch [2649]/[3760]	Speed: 35.128577 samples/sec	 accuracy=43.218750	 loss=2.473991	 lr=0.010000
Epoch[004] Batch [2699]/[3760]	Speed: 34.920145 samples/sec	 accuracy=43.216435	 loss=2.473630	 lr=0.010000
Epoch[004] Batch [2749]/[3760]	Speed: 35.223047 samples/sec	 accuracy=43.238068	 loss=2.472712	 lr=0.010000
Epoch[004] Batch [2799]/[3760]	Speed: 34.983331 samples/sec	 accuracy=43.269531	 loss=2.471868	 lr=0.010000
Epoch[004] Batch [2849]/[3760]	Speed: 35.164259 samples/sec	 accuracy=43.294956	 loss=2.471056	 lr=0.010000
Epoch[004] Batch [2899]/[3760]	Speed: 34.938263 samples/sec	 accuracy=43.296875	 loss=2.471679	 lr=0.010000
Epoch[004] Batch [2949]/[3760]	Speed: 35.026279 samples/sec	 accuracy=43.329979	 loss=2.469693	 lr=0.010000
Epoch[004] Batch [2999]/[3760]	Speed: 35.012780 samples/sec	 accuracy=43.345833	 loss=2.469120	 lr=0.010000
Epoch[004] Batch [3049]/[3760]	Speed: 35.056122 samples/sec	 accuracy=43.353484	 loss=2.468947	 lr=0.010000
Epoch[004] Batch [3099]/[3760]	Speed: 35.131615 samples/sec	 accuracy=43.359879	 loss=2.468631	 lr=0.010000
Epoch[004] Batch [3149]/[3760]	Speed: 35.054989 samples/sec	 accuracy=43.360119	 loss=2.468644	 lr=0.010000
Epoch[004] Batch [3199]/[3760]	Speed: 35.023462 samples/sec	 accuracy=43.364746	 loss=2.468137	 lr=0.010000
Epoch[004] Batch [3249]/[3760]	Speed: 35.174694 samples/sec	 accuracy=43.362019	 loss=2.467426	 lr=0.010000
Epoch[004] Batch [3299]/[3760]	Speed: 35.037526 samples/sec	 accuracy=43.371686	 loss=2.466912	 lr=0.010000
Epoch[004] Batch [3349]/[3760]	Speed: 35.109766 samples/sec	 accuracy=43.400187	 loss=2.466436	 lr=0.010000
Epoch[004] Batch [3399]/[3760]	Speed: 35.100556 samples/sec	 accuracy=43.423254	 loss=2.465003	 lr=0.010000
Epoch[004] Batch [3449]/[3760]	Speed: 35.348421 samples/sec	 accuracy=43.430254	 loss=2.463828	 lr=0.010000
Epoch[004] Batch [3499]/[3760]	Speed: 34.821146 samples/sec	 accuracy=43.446429	 loss=2.463371	 lr=0.010000
Epoch[004] Batch [3549]/[3760]	Speed: 35.400497 samples/sec	 accuracy=43.459947	 loss=2.462937	 lr=0.010000
Epoch[004] Batch [3599]/[3760]	Speed: 35.065063 samples/sec	 accuracy=43.478733	 loss=2.462236	 lr=0.010000
Epoch[004] Batch [3649]/[3760]	Speed: 35.249278 samples/sec	 accuracy=43.498716	 loss=2.461140	 lr=0.010000
Epoch[004] Batch [3699]/[3760]	Speed: 35.057671 samples/sec	 accuracy=43.497044	 loss=2.461175	 lr=0.010000
Epoch[004] Batch [3749]/[3760]	Speed: 36.546876 samples/sec	 accuracy=43.521667	 loss=2.460126	 lr=0.010000
Batch [0049]/[0303]: acc-top1=45.093750 acc-top5=71.312500
Batch [0099]/[0303]: acc-top1=45.296875 acc-top5=71.843750
Batch [0149]/[0303]: acc-top1=45.510417 acc-top5=71.916667
Batch [0199]/[0303]: acc-top1=45.531250 acc-top5=71.765625
Batch [0249]/[0303]: acc-top1=45.356250 acc-top5=71.537500
Batch [0299]/[0303]: acc-top1=45.406250 acc-top5=71.421875
[Epoch 004] training: accuracy=43.525598	 loss=2.460104
[Epoch 004] speed: 34 samples/sec	time cost: 7383.552210
[Epoch 004] validation: acc-top1=45.400165 acc-top5=71.400578 loss=2.607401
Epoch[005] Batch [0049]/[3759]	Speed: 24.059357 samples/sec	 accuracy=45.218750	 loss=2.351048	 lr=0.010000
Epoch[005] Batch [0099]/[3759]	Speed: 35.114628 samples/sec	 accuracy=45.546875	 loss=2.358415	 lr=0.010000
Epoch[005] Batch [0149]/[3759]	Speed: 35.064729 samples/sec	 accuracy=46.041667	 loss=2.323271	 lr=0.010000
Epoch[005] Batch [0199]/[3759]	Speed: 34.731522 samples/sec	 accuracy=46.250000	 loss=2.311294	 lr=0.010000
Epoch[005] Batch [0249]/[3759]	Speed: 35.247912 samples/sec	 accuracy=46.356250	 loss=2.304614	 lr=0.010000
Epoch[005] Batch [0299]/[3759]	Speed: 34.854018 samples/sec	 accuracy=46.218750	 loss=2.311918	 lr=0.010000
Epoch[005] Batch [0349]/[3759]	Speed: 35.215816 samples/sec	 accuracy=45.973214	 loss=2.320906	 lr=0.010000
Epoch[005] Batch [0399]/[3759]	Speed: 35.139783 samples/sec	 accuracy=45.781250	 loss=2.336126	 lr=0.010000
Epoch[005] Batch [0449]/[3759]	Speed: 35.126075 samples/sec	 accuracy=45.625000	 loss=2.342576	 lr=0.010000
Epoch[005] Batch [0499]/[3759]	Speed: 35.257209 samples/sec	 accuracy=45.725000	 loss=2.341557	 lr=0.010000
Epoch[005] Batch [0549]/[3759]	Speed: 35.004936 samples/sec	 accuracy=45.761364	 loss=2.341931	 lr=0.010000
Epoch[005] Batch [0599]/[3759]	Speed: 35.042095 samples/sec	 accuracy=45.664062	 loss=2.348171	 lr=0.010000
Epoch[005] Batch [0649]/[3759]	Speed: 35.125124 samples/sec	 accuracy=45.593750	 loss=2.347847	 lr=0.010000
Epoch[005] Batch [0699]/[3759]	Speed: 35.143430 samples/sec	 accuracy=45.535714	 loss=2.349449	 lr=0.010000
Epoch[005] Batch [0749]/[3759]	Speed: 35.238332 samples/sec	 accuracy=45.510417	 loss=2.348851	 lr=0.010000
Epoch[005] Batch [0799]/[3759]	Speed: 35.084277 samples/sec	 accuracy=45.539062	 loss=2.348008	 lr=0.010000
Epoch[005] Batch [0849]/[3759]	Speed: 35.138522 samples/sec	 accuracy=45.545956	 loss=2.346592	 lr=0.010000
Epoch[005] Batch [0899]/[3759]	Speed: 35.210264 samples/sec	 accuracy=45.501736	 loss=2.349328	 lr=0.010000
Epoch[005] Batch [0949]/[3759]	Speed: 35.231152 samples/sec	 accuracy=45.472039	 loss=2.351020	 lr=0.010000
Epoch[005] Batch [0999]/[3759]	Speed: 35.273737 samples/sec	 accuracy=45.464063	 loss=2.350513	 lr=0.010000
Epoch[005] Batch [1049]/[3759]	Speed: 35.097148 samples/sec	 accuracy=45.523810	 loss=2.349121	 lr=0.010000
Epoch[005] Batch [1099]/[3759]	Speed: 35.194439 samples/sec	 accuracy=45.491477	 loss=2.351324	 lr=0.010000
Epoch[005] Batch [1149]/[3759]	Speed: 35.105897 samples/sec	 accuracy=45.442935	 loss=2.352084	 lr=0.010000
Epoch[005] Batch [1199]/[3759]	Speed: 35.281433 samples/sec	 accuracy=45.440104	 loss=2.351313	 lr=0.010000
Epoch[005] Batch [1249]/[3759]	Speed: 34.930067 samples/sec	 accuracy=45.423750	 loss=2.351428	 lr=0.010000
Epoch[005] Batch [1299]/[3759]	Speed: 35.189539 samples/sec	 accuracy=45.411058	 loss=2.352346	 lr=0.010000
Epoch[005] Batch [1349]/[3759]	Speed: 35.124211 samples/sec	 accuracy=45.412037	 loss=2.353181	 lr=0.010000
Epoch[005] Batch [1399]/[3759]	Speed: 35.064326 samples/sec	 accuracy=45.447545	 loss=2.353238	 lr=0.010000
Epoch[005] Batch [1449]/[3759]	Speed: 35.168794 samples/sec	 accuracy=45.450431	 loss=2.353673	 lr=0.010000
Epoch[005] Batch [1499]/[3759]	Speed: 35.292536 samples/sec	 accuracy=45.440625	 loss=2.353426	 lr=0.010000
Epoch[005] Batch [1549]/[3759]	Speed: 34.860473 samples/sec	 accuracy=45.451613	 loss=2.353697	 lr=0.010000
Epoch[005] Batch [1599]/[3759]	Speed: 35.128432 samples/sec	 accuracy=45.476562	 loss=2.353195	 lr=0.010000
Epoch[005] Batch [1649]/[3759]	Speed: 35.032303 samples/sec	 accuracy=45.432765	 loss=2.353879	 lr=0.010000
Epoch[005] Batch [1699]/[3759]	Speed: 35.131868 samples/sec	 accuracy=45.430147	 loss=2.355618	 lr=0.010000
Epoch[005] Batch [1749]/[3759]	Speed: 34.955986 samples/sec	 accuracy=45.421429	 loss=2.356025	 lr=0.010000
Epoch[005] Batch [1799]/[3759]	Speed: 35.263735 samples/sec	 accuracy=45.419271	 loss=2.355210	 lr=0.010000
Epoch[005] Batch [1849]/[3759]	Speed: 35.253926 samples/sec	 accuracy=45.434122	 loss=2.355442	 lr=0.010000
Epoch[005] Batch [1899]/[3759]	Speed: 35.085743 samples/sec	 accuracy=45.462993	 loss=2.354851	 lr=0.010000
Epoch[005] Batch [1949]/[3759]	Speed: 35.061759 samples/sec	 accuracy=45.455929	 loss=2.355664	 lr=0.010000
Epoch[005] Batch [1999]/[3759]	Speed: 35.102586 samples/sec	 accuracy=45.427344	 loss=2.356518	 lr=0.010000
Epoch[005] Batch [2049]/[3759]	Speed: 35.189931 samples/sec	 accuracy=45.448171	 loss=2.355511	 lr=0.010000
Epoch[005] Batch [2099]/[3759]	Speed: 35.184071 samples/sec	 accuracy=45.479167	 loss=2.353655	 lr=0.010000
Epoch[005] Batch [2149]/[3759]	Speed: 35.102691 samples/sec	 accuracy=45.477471	 loss=2.354279	 lr=0.010000
Epoch[005] Batch [2199]/[3759]	Speed: 35.020670 samples/sec	 accuracy=45.485795	 loss=2.354048	 lr=0.010000
Epoch[005] Batch [2249]/[3759]	Speed: 35.154960 samples/sec	 accuracy=45.538194	 loss=2.351831	 lr=0.010000
Epoch[005] Batch [2299]/[3759]	Speed: 35.110090 samples/sec	 accuracy=45.538723	 loss=2.352479	 lr=0.010000
Epoch[005] Batch [2349]/[3759]	Speed: 34.972719 samples/sec	 accuracy=45.529255	 loss=2.352009	 lr=0.010000
Epoch[005] Batch [2399]/[3759]	Speed: 35.128557 samples/sec	 accuracy=45.538411	 loss=2.351674	 lr=0.010000
Epoch[005] Batch [2449]/[3759]	Speed: 35.129582 samples/sec	 accuracy=45.532526	 loss=2.351092	 lr=0.010000
Epoch[005] Batch [2499]/[3759]	Speed: 35.033090 samples/sec	 accuracy=45.547500	 loss=2.349979	 lr=0.010000
Epoch[005] Batch [2549]/[3759]	Speed: 35.074272 samples/sec	 accuracy=45.541667	 loss=2.349884	 lr=0.010000
Epoch[005] Batch [2599]/[3759]	Speed: 35.114148 samples/sec	 accuracy=45.552284	 loss=2.349619	 lr=0.010000
Epoch[005] Batch [2649]/[3759]	Speed: 35.022257 samples/sec	 accuracy=45.561910	 loss=2.348663	 lr=0.010000
Epoch[005] Batch [2699]/[3759]	Speed: 35.009803 samples/sec	 accuracy=45.565972	 loss=2.348791	 lr=0.010000
Epoch[005] Batch [2749]/[3759]	Speed: 35.274828 samples/sec	 accuracy=45.586364	 loss=2.347942	 lr=0.010000
Epoch[005] Batch [2799]/[3759]	Speed: 34.973513 samples/sec	 accuracy=45.614955	 loss=2.347629	 lr=0.010000
Epoch[005] Batch [2849]/[3759]	Speed: 35.009735 samples/sec	 accuracy=45.620614	 loss=2.347614	 lr=0.010000
Epoch[005] Batch [2899]/[3759]	Speed: 35.225612 samples/sec	 accuracy=45.629310	 loss=2.346682	 lr=0.010000
Epoch[005] Batch [2949]/[3759]	Speed: 34.979181 samples/sec	 accuracy=45.640890	 loss=2.345794	 lr=0.010000
Epoch[005] Batch [2999]/[3759]	Speed: 35.128303 samples/sec	 accuracy=45.651562	 loss=2.345210	 lr=0.010000
Epoch[005] Batch [3049]/[3759]	Speed: 35.078539 samples/sec	 accuracy=45.654713	 loss=2.345054	 lr=0.010000
Epoch[005] Batch [3099]/[3759]	Speed: 35.181375 samples/sec	 accuracy=45.655242	 loss=2.345655	 lr=0.010000
Epoch[005] Batch [3149]/[3759]	Speed: 35.072215 samples/sec	 accuracy=45.640377	 loss=2.346509	 lr=0.010000
Epoch[005] Batch [3199]/[3759]	Speed: 34.849986 samples/sec	 accuracy=45.642578	 loss=2.346627	 lr=0.010000
Epoch[005] Batch [3249]/[3759]	Speed: 35.180327 samples/sec	 accuracy=45.660096	 loss=2.346266	 lr=0.010000
Epoch[005] Batch [3299]/[3759]	Speed: 34.905797 samples/sec	 accuracy=45.661458	 loss=2.345839	 lr=0.010000
Epoch[005] Batch [3349]/[3759]	Speed: 35.311151 samples/sec	 accuracy=45.646455	 loss=2.346405	 lr=0.010000
Epoch[005] Batch [3399]/[3759]	Speed: 34.985760 samples/sec	 accuracy=45.650735	 loss=2.346068	 lr=0.010000
Epoch[005] Batch [3449]/[3759]	Speed: 34.979290 samples/sec	 accuracy=45.675725	 loss=2.344694	 lr=0.010000
Epoch[005] Batch [3499]/[3759]	Speed: 34.946438 samples/sec	 accuracy=45.677679	 loss=2.344502	 lr=0.010000
Epoch[005] Batch [3549]/[3759]	Speed: 35.103099 samples/sec	 accuracy=45.667254	 loss=2.344326	 lr=0.010000
Epoch[005] Batch [3599]/[3759]	Speed: 34.988763 samples/sec	 accuracy=45.675347	 loss=2.344338	 lr=0.010000
Epoch[005] Batch [3649]/[3759]	Speed: 34.966967 samples/sec	 accuracy=45.692209	 loss=2.343278	 lr=0.010000
Epoch[005] Batch [3699]/[3759]	Speed: 35.048106 samples/sec	 accuracy=45.686655	 loss=2.343365	 lr=0.010000
Epoch[005] Batch [3749]/[3759]	Speed: 36.319611 samples/sec	 accuracy=45.692500	 loss=2.343218	 lr=0.010000
Batch [0049]/[0303]: acc-top1=46.250000 acc-top5=73.187500
Batch [0099]/[0303]: acc-top1=47.437500 acc-top5=74.171875
Batch [0149]/[0303]: acc-top1=47.791667 acc-top5=74.208333
Batch [0199]/[0303]: acc-top1=47.820312 acc-top5=74.015625
Batch [0249]/[0303]: acc-top1=47.568750 acc-top5=73.887500
Batch [0299]/[0303]: acc-top1=47.729167 acc-top5=73.859375
[Epoch 005] training: accuracy=45.696578	 loss=2.343075
[Epoch 005] speed: 34 samples/sec	time cost: 7380.307971
[Epoch 005] validation: acc-top1=47.679455 acc-top5=73.839728 loss=2.447927
Epoch[006] Batch [0049]/[3760]	Speed: 24.069558 samples/sec	 accuracy=47.656250	 loss=2.235820	 lr=0.010000
Epoch[006] Batch [0099]/[3760]	Speed: 34.787511 samples/sec	 accuracy=47.625000	 loss=2.241828	 lr=0.010000
Epoch[006] Batch [0149]/[3760]	Speed: 35.241408 samples/sec	 accuracy=47.458333	 loss=2.251840	 lr=0.010000
Epoch[006] Batch [0199]/[3760]	Speed: 34.884302 samples/sec	 accuracy=47.304688	 loss=2.253137	 lr=0.010000
Epoch[006] Batch [0249]/[3760]	Speed: 35.035915 samples/sec	 accuracy=47.218750	 loss=2.247966	 lr=0.010000
Epoch[006] Batch [0299]/[3760]	Speed: 34.896152 samples/sec	 accuracy=47.145833	 loss=2.248777	 lr=0.010000
Epoch[006] Batch [0349]/[3760]	Speed: 35.135939 samples/sec	 accuracy=47.205357	 loss=2.254683	 lr=0.010000
Epoch[006] Batch [0399]/[3760]	Speed: 34.983903 samples/sec	 accuracy=47.304688	 loss=2.250725	 lr=0.010000
Epoch[006] Batch [0449]/[3760]	Speed: 35.089518 samples/sec	 accuracy=47.406250	 loss=2.247064	 lr=0.010000
Epoch[006] Batch [0499]/[3760]	Speed: 35.002388 samples/sec	 accuracy=47.521875	 loss=2.244732	 lr=0.010000
Epoch[006] Batch [0549]/[3760]	Speed: 35.109035 samples/sec	 accuracy=47.511364	 loss=2.247110	 lr=0.010000
Epoch[006] Batch [0599]/[3760]	Speed: 35.089122 samples/sec	 accuracy=47.447917	 loss=2.252307	 lr=0.010000
Epoch[006] Batch [0649]/[3760]	Speed: 34.933524 samples/sec	 accuracy=47.502404	 loss=2.251055	 lr=0.010000
Epoch[006] Batch [0699]/[3760]	Speed: 35.134222 samples/sec	 accuracy=47.508929	 loss=2.251152	 lr=0.010000
Epoch[006] Batch [0749]/[3760]	Speed: 34.953226 samples/sec	 accuracy=47.514583	 loss=2.251866	 lr=0.010000
Epoch[006] Batch [0799]/[3760]	Speed: 35.073318 samples/sec	 accuracy=47.509766	 loss=2.255025	 lr=0.010000
Epoch[006] Batch [0849]/[3760]	Speed: 35.064125 samples/sec	 accuracy=47.430147	 loss=2.258920	 lr=0.010000
Epoch[006] Batch [0899]/[3760]	Speed: 34.922812 samples/sec	 accuracy=47.500000	 loss=2.258052	 lr=0.010000
Epoch[006] Batch [0949]/[3760]	Speed: 35.143319 samples/sec	 accuracy=47.452303	 loss=2.259270	 lr=0.010000
Epoch[006] Batch [0999]/[3760]	Speed: 34.949621 samples/sec	 accuracy=47.478125	 loss=2.257785	 lr=0.010000
Epoch[006] Batch [1049]/[3760]	Speed: 35.214622 samples/sec	 accuracy=47.549107	 loss=2.254910	 lr=0.010000
Epoch[006] Batch [1099]/[3760]	Speed: 35.108096 samples/sec	 accuracy=47.558239	 loss=2.254298	 lr=0.010000
Epoch[006] Batch [1149]/[3760]	Speed: 34.963941 samples/sec	 accuracy=47.527174	 loss=2.254270	 lr=0.010000
Epoch[006] Batch [1199]/[3760]	Speed: 35.179622 samples/sec	 accuracy=47.515625	 loss=2.253060	 lr=0.010000
Epoch[006] Batch [1249]/[3760]	Speed: 34.879805 samples/sec	 accuracy=47.541250	 loss=2.252358	 lr=0.010000
Epoch[006] Batch [1299]/[3760]	Speed: 35.144687 samples/sec	 accuracy=47.526442	 loss=2.253681	 lr=0.010000
Epoch[006] Batch [1349]/[3760]	Speed: 35.091343 samples/sec	 accuracy=47.531250	 loss=2.252897	 lr=0.010000
Epoch[006] Batch [1399]/[3760]	Speed: 35.077960 samples/sec	 accuracy=47.556920	 loss=2.251766	 lr=0.010000
Epoch[006] Batch [1449]/[3760]	Speed: 35.191077 samples/sec	 accuracy=47.617457	 loss=2.251558	 lr=0.010000
Epoch[006] Batch [1499]/[3760]	Speed: 35.209951 samples/sec	 accuracy=47.600000	 loss=2.252770	 lr=0.010000
Epoch[006] Batch [1549]/[3760]	Speed: 35.067032 samples/sec	 accuracy=47.623992	 loss=2.252766	 lr=0.010000
Epoch[006] Batch [1599]/[3760]	Speed: 34.998329 samples/sec	 accuracy=47.666992	 loss=2.251477	 lr=0.010000
Epoch[006] Batch [1649]/[3760]	Speed: 35.119020 samples/sec	 accuracy=47.670455	 loss=2.252041	 lr=0.010000
Epoch[006] Batch [1699]/[3760]	Speed: 35.204496 samples/sec	 accuracy=47.625919	 loss=2.252822	 lr=0.010000
Epoch[006] Batch [1749]/[3760]	Speed: 34.976333 samples/sec	 accuracy=47.626786	 loss=2.252614	 lr=0.010000
Epoch[006] Batch [1799]/[3760]	Speed: 35.209880 samples/sec	 accuracy=47.600694	 loss=2.254536	 lr=0.010000
Epoch[006] Batch [1849]/[3760]	Speed: 35.161574 samples/sec	 accuracy=47.629223	 loss=2.253901	 lr=0.010000
Epoch[006] Batch [1899]/[3760]	Speed: 34.934655 samples/sec	 accuracy=47.630757	 loss=2.253513	 lr=0.010000
Epoch[006] Batch [1949]/[3760]	Speed: 35.113197 samples/sec	 accuracy=47.650641	 loss=2.253237	 lr=0.010000
Epoch[006] Batch [1999]/[3760]	Speed: 35.165834 samples/sec	 accuracy=47.693750	 loss=2.251915	 lr=0.010000
Epoch[006] Batch [2049]/[3760]	Speed: 34.885659 samples/sec	 accuracy=47.686738	 loss=2.251271	 lr=0.010000
Epoch[006] Batch [2099]/[3760]	Speed: 34.945679 samples/sec	 accuracy=47.671875	 loss=2.251808	 lr=0.010000
Epoch[006] Batch [2149]/[3760]	Speed: 34.939318 samples/sec	 accuracy=47.643169	 loss=2.253311	 lr=0.010000
Epoch[006] Batch [2199]/[3760]	Speed: 34.888845 samples/sec	 accuracy=47.640625	 loss=2.253473	 lr=0.010000
Epoch[006] Batch [2249]/[3760]	Speed: 35.197975 samples/sec	 accuracy=47.597222	 loss=2.255495	 lr=0.010000
Epoch[006] Batch [2299]/[3760]	Speed: 35.102998 samples/sec	 accuracy=47.576087	 loss=2.255122	 lr=0.010000
Epoch[006] Batch [2349]/[3760]	Speed: 35.265474 samples/sec	 accuracy=47.603723	 loss=2.254523	 lr=0.010000
Epoch[006] Batch [2399]/[3760]	Speed: 35.157778 samples/sec	 accuracy=47.602214	 loss=2.254721	 lr=0.010000
Epoch[006] Batch [2449]/[3760]	Speed: 34.971561 samples/sec	 accuracy=47.598214	 loss=2.254542	 lr=0.010000
Epoch[006] Batch [2499]/[3760]	Speed: 35.055595 samples/sec	 accuracy=47.591875	 loss=2.254627	 lr=0.010000
Epoch[006] Batch [2549]/[3760]	Speed: 35.090118 samples/sec	 accuracy=47.602328	 loss=2.254548	 lr=0.010000
Epoch[006] Batch [2599]/[3760]	Speed: 35.043112 samples/sec	 accuracy=47.597957	 loss=2.253973	 lr=0.010000
Epoch[006] Batch [2649]/[3760]	Speed: 34.884055 samples/sec	 accuracy=47.613797	 loss=2.254034	 lr=0.010000
Epoch[006] Batch [2699]/[3760]	Speed: 35.128990 samples/sec	 accuracy=47.608218	 loss=2.254246	 lr=0.010000
Epoch[006] Batch [2749]/[3760]	Speed: 35.136701 samples/sec	 accuracy=47.635227	 loss=2.253297	 lr=0.010000
Epoch[006] Batch [2799]/[3760]	Speed: 34.926007 samples/sec	 accuracy=47.629464	 loss=2.253634	 lr=0.010000
Epoch[006] Batch [2849]/[3760]	Speed: 35.005460 samples/sec	 accuracy=47.659539	 loss=2.252041	 lr=0.010000
Epoch[006] Batch [2899]/[3760]	Speed: 34.866058 samples/sec	 accuracy=47.647629	 loss=2.252661	 lr=0.010000
Epoch[006] Batch [2949]/[3760]	Speed: 35.235704 samples/sec	 accuracy=47.648835	 loss=2.252332	 lr=0.010000
Epoch[006] Batch [2999]/[3760]	Speed: 35.147020 samples/sec	 accuracy=47.640104	 loss=2.253074	 lr=0.010000
Epoch[006] Batch [3049]/[3760]	Speed: 34.936780 samples/sec	 accuracy=47.655738	 loss=2.252463	 lr=0.010000
Epoch[006] Batch [3099]/[3760]	Speed: 35.053274 samples/sec	 accuracy=47.649194	 loss=2.252871	 lr=0.010000
Epoch[006] Batch [3149]/[3760]	Speed: 35.204502 samples/sec	 accuracy=47.661210	 loss=2.252416	 lr=0.010000
Epoch[006] Batch [3199]/[3760]	Speed: 35.025997 samples/sec	 accuracy=47.666016	 loss=2.252168	 lr=0.010000
Epoch[006] Batch [3249]/[3760]	Speed: 35.175250 samples/sec	 accuracy=47.672115	 loss=2.252104	 lr=0.010000
Epoch[006] Batch [3299]/[3760]	Speed: 34.957835 samples/sec	 accuracy=47.669034	 loss=2.251637	 lr=0.010000
Epoch[006] Batch [3349]/[3760]	Speed: 35.118790 samples/sec	 accuracy=47.679571	 loss=2.251308	 lr=0.010000
Epoch[006] Batch [3399]/[3760]	Speed: 35.306509 samples/sec	 accuracy=47.681985	 loss=2.251428	 lr=0.010000
Epoch[006] Batch [3449]/[3760]	Speed: 35.002025 samples/sec	 accuracy=47.702446	 loss=2.251188	 lr=0.010000
Epoch[006] Batch [3499]/[3760]	Speed: 35.121672 samples/sec	 accuracy=47.737946	 loss=2.249726	 lr=0.010000
Epoch[006] Batch [3549]/[3760]	Speed: 35.086298 samples/sec	 accuracy=47.762324	 loss=2.248704	 lr=0.010000
Epoch[006] Batch [3599]/[3760]	Speed: 35.086758 samples/sec	 accuracy=47.774306	 loss=2.247900	 lr=0.010000
Epoch[006] Batch [3649]/[3760]	Speed: 34.876703 samples/sec	 accuracy=47.780394	 loss=2.247631	 lr=0.010000
Epoch[006] Batch [3699]/[3760]	Speed: 35.136598 samples/sec	 accuracy=47.773649	 loss=2.247754	 lr=0.010000
Epoch[006] Batch [3749]/[3760]	Speed: 36.350416 samples/sec	 accuracy=47.759583	 loss=2.247959	 lr=0.010000
Batch [0049]/[0303]: acc-top1=49.656250 acc-top5=75.156250
Batch [0099]/[0303]: acc-top1=49.703125 acc-top5=75.468750
Batch [0149]/[0303]: acc-top1=49.718750 acc-top5=75.552083
Batch [0199]/[0303]: acc-top1=49.570312 acc-top5=75.312500
Batch [0249]/[0303]: acc-top1=49.275000 acc-top5=75.043750
Batch [0299]/[0303]: acc-top1=49.395833 acc-top5=75.067708
[Epoch 006] training: accuracy=47.762217	 loss=2.248106
[Epoch 006] speed: 34 samples/sec	time cost: 7386.291556
[Epoch 006] validation: acc-top1=49.365718 acc-top5=75.051568 loss=2.399570
Epoch[007] Batch [0049]/[3760]	Speed: 23.949385 samples/sec	 accuracy=50.968750	 loss=2.125078	 lr=0.010000
Epoch[007] Batch [0099]/[3760]	Speed: 34.879980 samples/sec	 accuracy=49.781250	 loss=2.153803	 lr=0.010000
Epoch[007] Batch [0149]/[3760]	Speed: 35.132586 samples/sec	 accuracy=49.875000	 loss=2.144970	 lr=0.010000
Epoch[007] Batch [0199]/[3760]	Speed: 34.854784 samples/sec	 accuracy=50.062500	 loss=2.139704	 lr=0.010000
Epoch[007] Batch [0249]/[3760]	Speed: 35.367917 samples/sec	 accuracy=49.993750	 loss=2.140461	 lr=0.010000
Epoch[007] Batch [0299]/[3760]	Speed: 35.024609 samples/sec	 accuracy=49.781250	 loss=2.145110	 lr=0.010000
Epoch[007] Batch [0349]/[3760]	Speed: 35.435740 samples/sec	 accuracy=49.848214	 loss=2.137662	 lr=0.010000
Epoch[007] Batch [0399]/[3760]	Speed: 35.426217 samples/sec	 accuracy=49.796875	 loss=2.141423	 lr=0.010000
Epoch[007] Batch [0449]/[3760]	Speed: 35.179925 samples/sec	 accuracy=49.711806	 loss=2.148437	 lr=0.010000
Epoch[007] Batch [0499]/[3760]	Speed: 35.189342 samples/sec	 accuracy=49.737500	 loss=2.148038	 lr=0.010000
Epoch[007] Batch [0549]/[3760]	Speed: 35.156639 samples/sec	 accuracy=49.696023	 loss=2.150816	 lr=0.010000
Epoch[007] Batch [0599]/[3760]	Speed: 35.253277 samples/sec	 accuracy=49.583333	 loss=2.156892	 lr=0.010000
Epoch[007] Batch [0649]/[3760]	Speed: 35.271168 samples/sec	 accuracy=49.516827	 loss=2.160847	 lr=0.010000
Epoch[007] Batch [0699]/[3760]	Speed: 35.211142 samples/sec	 accuracy=49.453125	 loss=2.163162	 lr=0.010000
Epoch[007] Batch [0749]/[3760]	Speed: 34.918666 samples/sec	 accuracy=49.377083	 loss=2.166551	 lr=0.010000
Epoch[007] Batch [0799]/[3760]	Speed: 35.297381 samples/sec	 accuracy=49.402344	 loss=2.165504	 lr=0.010000
Epoch[007] Batch [0849]/[3760]	Speed: 35.354581 samples/sec	 accuracy=49.358456	 loss=2.170083	 lr=0.010000
Epoch[007] Batch [0899]/[3760]	Speed: 35.266824 samples/sec	 accuracy=49.390625	 loss=2.167279	 lr=0.010000
Epoch[007] Batch [0949]/[3760]	Speed: 35.297185 samples/sec	 accuracy=49.449013	 loss=2.166300	 lr=0.010000
Epoch[007] Batch [0999]/[3760]	Speed: 35.346628 samples/sec	 accuracy=49.489062	 loss=2.167173	 lr=0.010000
Epoch[007] Batch [1049]/[3760]	Speed: 35.114446 samples/sec	 accuracy=49.504464	 loss=2.166266	 lr=0.010000
Epoch[007] Batch [1099]/[3760]	Speed: 35.224669 samples/sec	 accuracy=49.444602	 loss=2.168223	 lr=0.010000
Epoch[007] Batch [1149]/[3760]	Speed: 35.080693 samples/sec	 accuracy=49.419837	 loss=2.168510	 lr=0.010000
Epoch[007] Batch [1199]/[3760]	Speed: 35.215147 samples/sec	 accuracy=49.417969	 loss=2.167930	 lr=0.010000
Epoch[007] Batch [1249]/[3760]	Speed: 35.268612 samples/sec	 accuracy=49.442500	 loss=2.166140	 lr=0.010000
Epoch[007] Batch [1299]/[3760]	Speed: 35.227591 samples/sec	 accuracy=49.420673	 loss=2.166388	 lr=0.010000
Epoch[007] Batch [1349]/[3760]	Speed: 35.094825 samples/sec	 accuracy=49.356481	 loss=2.166855	 lr=0.010000
Epoch[007] Batch [1399]/[3760]	Speed: 35.202406 samples/sec	 accuracy=49.400670	 loss=2.165181	 lr=0.010000
Epoch[007] Batch [1449]/[3760]	Speed: 35.123479 samples/sec	 accuracy=49.391164	 loss=2.166208	 lr=0.010000
Epoch[007] Batch [1499]/[3760]	Speed: 35.277178 samples/sec	 accuracy=49.363542	 loss=2.167386	 lr=0.010000
Epoch[007] Batch [1549]/[3760]	Speed: 35.310664 samples/sec	 accuracy=49.339718	 loss=2.167386	 lr=0.010000
Epoch[007] Batch [1599]/[3760]	Speed: 35.129152 samples/sec	 accuracy=49.328125	 loss=2.168246	 lr=0.010000
Epoch[007] Batch [1649]/[3760]	Speed: 35.203134 samples/sec	 accuracy=49.308712	 loss=2.169587	 lr=0.010000
Epoch[007] Batch [1699]/[3760]	Speed: 35.236347 samples/sec	 accuracy=49.310662	 loss=2.169864	 lr=0.010000
Epoch[007] Batch [1749]/[3760]	Speed: 35.074832 samples/sec	 accuracy=49.297321	 loss=2.169575	 lr=0.010000
Epoch[007] Batch [1799]/[3760]	Speed: 35.166547 samples/sec	 accuracy=49.317708	 loss=2.169774	 lr=0.010000
Epoch[007] Batch [1849]/[3760]	Speed: 35.145970 samples/sec	 accuracy=49.301520	 loss=2.169853	 lr=0.010000
Epoch[007] Batch [1899]/[3760]	Speed: 34.805613 samples/sec	 accuracy=49.314967	 loss=2.168740	 lr=0.010000
Epoch[007] Batch [1949]/[3760]	Speed: 35.422880 samples/sec	 accuracy=49.317308	 loss=2.168699	 lr=0.010000
Epoch[007] Batch [1999]/[3760]	Speed: 35.022694 samples/sec	 accuracy=49.310938	 loss=2.168598	 lr=0.010000
Epoch[007] Batch [2049]/[3760]	Speed: 35.329017 samples/sec	 accuracy=49.320122	 loss=2.168242	 lr=0.010000
Epoch[007] Batch [2099]/[3760]	Speed: 35.215828 samples/sec	 accuracy=49.333333	 loss=2.167964	 lr=0.010000
Epoch[007] Batch [2149]/[3760]	Speed: 35.225338 samples/sec	 accuracy=49.300145	 loss=2.168820	 lr=0.010000
Epoch[007] Batch [2199]/[3760]	Speed: 35.194102 samples/sec	 accuracy=49.288352	 loss=2.168864	 lr=0.010000
Epoch[007] Batch [2249]/[3760]	Speed: 35.207677 samples/sec	 accuracy=49.255556	 loss=2.170927	 lr=0.010000
Epoch[007] Batch [2299]/[3760]	Speed: 35.179423 samples/sec	 accuracy=49.234375	 loss=2.170875	 lr=0.010000
Epoch[007] Batch [2349]/[3760]	Speed: 35.142773 samples/sec	 accuracy=49.210771	 loss=2.171929	 lr=0.010000
Epoch[007] Batch [2399]/[3760]	Speed: 35.195891 samples/sec	 accuracy=49.195964	 loss=2.172407	 lr=0.010000
Epoch[007] Batch [2449]/[3760]	Speed: 34.932172 samples/sec	 accuracy=49.188776	 loss=2.173481	 lr=0.010000
Epoch[007] Batch [2499]/[3760]	Speed: 35.357175 samples/sec	 accuracy=49.195000	 loss=2.172746	 lr=0.010000
Epoch[007] Batch [2549]/[3760]	Speed: 34.883622 samples/sec	 accuracy=49.200980	 loss=2.171592	 lr=0.010000
Epoch[007] Batch [2599]/[3760]	Speed: 35.533284 samples/sec	 accuracy=49.195913	 loss=2.171713	 lr=0.010000
Epoch[007] Batch [2649]/[3760]	Speed: 35.142008 samples/sec	 accuracy=49.234080	 loss=2.169969	 lr=0.010000
Epoch[007] Batch [2699]/[3760]	Speed: 35.258188 samples/sec	 accuracy=49.210648	 loss=2.170871	 lr=0.010000
Epoch[007] Batch [2749]/[3760]	Speed: 35.129816 samples/sec	 accuracy=49.199432	 loss=2.171144	 lr=0.010000
Epoch[007] Batch [2799]/[3760]	Speed: 35.349546 samples/sec	 accuracy=49.177455	 loss=2.172000	 lr=0.010000
Epoch[007] Batch [2849]/[3760]	Speed: 34.998670 samples/sec	 accuracy=49.179276	 loss=2.172392	 lr=0.010000
Epoch[007] Batch [2899]/[3760]	Speed: 35.445150 samples/sec	 accuracy=49.192888	 loss=2.171356	 lr=0.010000
Epoch[007] Batch [2949]/[3760]	Speed: 35.224802 samples/sec	 accuracy=49.206568	 loss=2.170852	 lr=0.010000
Epoch[007] Batch [2999]/[3760]	Speed: 35.307163 samples/sec	 accuracy=49.182812	 loss=2.171683	 lr=0.010000
Epoch[007] Batch [3049]/[3760]	Speed: 35.121227 samples/sec	 accuracy=49.193648	 loss=2.171537	 lr=0.010000
Epoch[007] Batch [3099]/[3760]	Speed: 35.097450 samples/sec	 accuracy=49.188508	 loss=2.171268	 lr=0.010000
Epoch[007] Batch [3149]/[3760]	Speed: 35.257799 samples/sec	 accuracy=49.172123	 loss=2.172541	 lr=0.010000
Epoch[007] Batch [3199]/[3760]	Speed: 35.269794 samples/sec	 accuracy=49.173828	 loss=2.172141	 lr=0.010000
Epoch[007] Batch [3249]/[3760]	Speed: 35.066429 samples/sec	 accuracy=49.176923	 loss=2.171971	 lr=0.010000
Epoch[007] Batch [3299]/[3760]	Speed: 35.325357 samples/sec	 accuracy=49.192235	 loss=2.171040	 lr=0.010000
Epoch[007] Batch [3349]/[3760]	Speed: 35.100335 samples/sec	 accuracy=49.212687	 loss=2.170580	 lr=0.010000
Epoch[007] Batch [3399]/[3760]	Speed: 35.130849 samples/sec	 accuracy=49.231618	 loss=2.170340	 lr=0.010000
Epoch[007] Batch [3449]/[3760]	Speed: 35.430987 samples/sec	 accuracy=49.226902	 loss=2.170413	 lr=0.010000
Epoch[007] Batch [3499]/[3760]	Speed: 35.155602 samples/sec	 accuracy=49.213393	 loss=2.171118	 lr=0.010000
Epoch[007] Batch [3549]/[3760]	Speed: 35.320338 samples/sec	 accuracy=49.206426	 loss=2.171113	 lr=0.010000
Epoch[007] Batch [3599]/[3760]	Speed: 35.118141 samples/sec	 accuracy=49.209201	 loss=2.170901	 lr=0.010000
Epoch[007] Batch [3649]/[3760]	Speed: 35.173360 samples/sec	 accuracy=49.202483	 loss=2.170959	 lr=0.010000
Epoch[007] Batch [3699]/[3760]	Speed: 35.232175 samples/sec	 accuracy=49.205236	 loss=2.171061	 lr=0.010000
Epoch[007] Batch [3749]/[3760]	Speed: 36.415851 samples/sec	 accuracy=49.199583	 loss=2.171204	 lr=0.010000
Batch [0049]/[0303]: acc-top1=50.000000 acc-top5=76.406250
Batch [0099]/[0303]: acc-top1=50.453125 acc-top5=76.562500
Batch [0149]/[0303]: acc-top1=51.031250 acc-top5=76.531250
Batch [0199]/[0303]: acc-top1=50.960938 acc-top5=76.687500
Batch [0249]/[0303]: acc-top1=50.806250 acc-top5=76.231250
Batch [0299]/[0303]: acc-top1=50.937500 acc-top5=76.182292
[Epoch 007] training: accuracy=49.194648	 loss=2.171439
[Epoch 007] speed: 34 samples/sec	time cost: 7361.813933
[Epoch 007] validation: acc-top1=50.948845 acc-top5=76.165429 loss=2.346002
Epoch[008] Batch [0049]/[3759]	Speed: 24.565026 samples/sec	 accuracy=52.562500	 loss=1.996524	 lr=0.010000
Epoch[008] Batch [0099]/[3759]	Speed: 34.860123 samples/sec	 accuracy=51.687500	 loss=2.027475	 lr=0.010000
Epoch[008] Batch [0149]/[3759]	Speed: 35.130257 samples/sec	 accuracy=51.312500	 loss=2.038964	 lr=0.010000
Epoch[008] Batch [0199]/[3759]	Speed: 34.957819 samples/sec	 accuracy=51.078125	 loss=2.040867	 lr=0.010000
Epoch[008] Batch [0249]/[3759]	Speed: 35.338320 samples/sec	 accuracy=50.981250	 loss=2.057226	 lr=0.010000
Epoch[008] Batch [0299]/[3759]	Speed: 34.672232 samples/sec	 accuracy=50.890625	 loss=2.069612	 lr=0.010000
Epoch[008] Batch [0349]/[3759]	Speed: 35.451592 samples/sec	 accuracy=50.665179	 loss=2.085741	 lr=0.010000
Epoch[008] Batch [0399]/[3759]	Speed: 34.897075 samples/sec	 accuracy=50.593750	 loss=2.087432	 lr=0.010000
Epoch[008] Batch [0449]/[3759]	Speed: 35.202054 samples/sec	 accuracy=50.736111	 loss=2.082683	 lr=0.010000
Epoch[008] Batch [0499]/[3759]	Speed: 34.961647 samples/sec	 accuracy=50.668750	 loss=2.087047	 lr=0.010000
Epoch[008] Batch [0549]/[3759]	Speed: 35.314656 samples/sec	 accuracy=50.625000	 loss=2.090027	 lr=0.010000
Epoch[008] Batch [0599]/[3759]	Speed: 35.097234 samples/sec	 accuracy=50.645833	 loss=2.090838	 lr=0.010000
Epoch[008] Batch [0649]/[3759]	Speed: 35.179835 samples/sec	 accuracy=50.634615	 loss=2.087933	 lr=0.010000
Epoch[008] Batch [0699]/[3759]	Speed: 35.272493 samples/sec	 accuracy=50.696429	 loss=2.088244	 lr=0.010000
Epoch[008] Batch [0749]/[3759]	Speed: 35.354486 samples/sec	 accuracy=50.727083	 loss=2.087645	 lr=0.010000
Epoch[008] Batch [0799]/[3759]	Speed: 35.348798 samples/sec	 accuracy=50.771484	 loss=2.083924	 lr=0.010000
Epoch[008] Batch [0849]/[3759]	Speed: 35.176454 samples/sec	 accuracy=50.678309	 loss=2.089367	 lr=0.010000
Epoch[008] Batch [0899]/[3759]	Speed: 35.296164 samples/sec	 accuracy=50.661458	 loss=2.091374	 lr=0.010000
Epoch[008] Batch [0949]/[3759]	Speed: 35.194739 samples/sec	 accuracy=50.723684	 loss=2.093258	 lr=0.010000
Epoch[008] Batch [0999]/[3759]	Speed: 35.374949 samples/sec	 accuracy=50.671875	 loss=2.096103	 lr=0.010000
Epoch[008] Batch [1049]/[3759]	Speed: 35.297734 samples/sec	 accuracy=50.723214	 loss=2.095987	 lr=0.010000
Epoch[008] Batch [1099]/[3759]	Speed: 35.225902 samples/sec	 accuracy=50.691761	 loss=2.097877	 lr=0.010000
Epoch[008] Batch [1149]/[3759]	Speed: 35.041882 samples/sec	 accuracy=50.733696	 loss=2.097637	 lr=0.010000
Epoch[008] Batch [1199]/[3759]	Speed: 35.220897 samples/sec	 accuracy=50.787760	 loss=2.096085	 lr=0.010000
Epoch[008] Batch [1249]/[3759]	Speed: 35.279525 samples/sec	 accuracy=50.735000	 loss=2.097155	 lr=0.010000
Epoch[008] Batch [1299]/[3759]	Speed: 35.284779 samples/sec	 accuracy=50.717548	 loss=2.098974	 lr=0.010000
Epoch[008] Batch [1349]/[3759]	Speed: 35.066881 samples/sec	 accuracy=50.681713	 loss=2.100083	 lr=0.010000
Epoch[008] Batch [1399]/[3759]	Speed: 35.106927 samples/sec	 accuracy=50.700893	 loss=2.099118	 lr=0.010000
Epoch[008] Batch [1449]/[3759]	Speed: 35.148388 samples/sec	 accuracy=50.687500	 loss=2.099872	 lr=0.010000
Epoch[008] Batch [1499]/[3759]	Speed: 34.928232 samples/sec	 accuracy=50.696875	 loss=2.098755	 lr=0.010000
Epoch[008] Batch [1549]/[3759]	Speed: 35.244785 samples/sec	 accuracy=50.743952	 loss=2.096812	 lr=0.010000
Epoch[008] Batch [1599]/[3759]	Speed: 35.155262 samples/sec	 accuracy=50.735352	 loss=2.098817	 lr=0.010000
Epoch[008] Batch [1649]/[3759]	Speed: 35.225491 samples/sec	 accuracy=50.749053	 loss=2.099551	 lr=0.010000
Epoch[008] Batch [1699]/[3759]	Speed: 34.984606 samples/sec	 accuracy=50.788603	 loss=2.098733	 lr=0.010000
Epoch[008] Batch [1749]/[3759]	Speed: 35.380150 samples/sec	 accuracy=50.768750	 loss=2.098947	 lr=0.010000
Epoch[008] Batch [1799]/[3759]	Speed: 35.234608 samples/sec	 accuracy=50.730903	 loss=2.099418	 lr=0.010000
Epoch[008] Batch [1849]/[3759]	Speed: 35.420007 samples/sec	 accuracy=50.722128	 loss=2.097455	 lr=0.010000
Epoch[008] Batch [1899]/[3759]	Speed: 35.152124 samples/sec	 accuracy=50.743421	 loss=2.095952	 lr=0.010000
Epoch[008] Batch [1949]/[3759]	Speed: 35.238307 samples/sec	 accuracy=50.737179	 loss=2.096744	 lr=0.010000
Epoch[008] Batch [1999]/[3759]	Speed: 35.196743 samples/sec	 accuracy=50.746094	 loss=2.096353	 lr=0.010000
Epoch[008] Batch [2049]/[3759]	Speed: 35.298961 samples/sec	 accuracy=50.735518	 loss=2.096500	 lr=0.010000
Epoch[008] Batch [2099]/[3759]	Speed: 35.215455 samples/sec	 accuracy=50.716518	 loss=2.097680	 lr=0.010000
Epoch[008] Batch [2149]/[3759]	Speed: 35.228532 samples/sec	 accuracy=50.721657	 loss=2.097803	 lr=0.010000
Epoch[008] Batch [2199]/[3759]	Speed: 35.240801 samples/sec	 accuracy=50.741477	 loss=2.097512	 lr=0.010000
Epoch[008] Batch [2249]/[3759]	Speed: 34.987305 samples/sec	 accuracy=50.775000	 loss=2.097446	 lr=0.010000
Epoch[008] Batch [2299]/[3759]	Speed: 35.432715 samples/sec	 accuracy=50.760870	 loss=2.098120	 lr=0.010000
Epoch[008] Batch [2349]/[3759]	Speed: 34.996311 samples/sec	 accuracy=50.765293	 loss=2.097837	 lr=0.010000
Epoch[008] Batch [2399]/[3759]	Speed: 35.407945 samples/sec	 accuracy=50.753255	 loss=2.099218	 lr=0.010000
Epoch[008] Batch [2449]/[3759]	Speed: 35.008313 samples/sec	 accuracy=50.755102	 loss=2.099115	 lr=0.010000
Epoch[008] Batch [2499]/[3759]	Speed: 35.332105 samples/sec	 accuracy=50.746250	 loss=2.099730	 lr=0.010000
Epoch[008] Batch [2549]/[3759]	Speed: 35.041389 samples/sec	 accuracy=50.775735	 loss=2.099759	 lr=0.010000
Epoch[008] Batch [2599]/[3759]	Speed: 35.291236 samples/sec	 accuracy=50.758413	 loss=2.100277	 lr=0.010000
Epoch[008] Batch [2649]/[3759]	Speed: 35.030157 samples/sec	 accuracy=50.772995	 loss=2.099965	 lr=0.010000
Epoch[008] Batch [2699]/[3759]	Speed: 35.306404 samples/sec	 accuracy=50.758102	 loss=2.099726	 lr=0.010000
Epoch[008] Batch [2749]/[3759]	Speed: 35.216272 samples/sec	 accuracy=50.715341	 loss=2.101437	 lr=0.010000
Epoch[008] Batch [2799]/[3759]	Speed: 35.216259 samples/sec	 accuracy=50.689174	 loss=2.102372	 lr=0.010000
Epoch[008] Batch [2849]/[3759]	Speed: 35.138045 samples/sec	 accuracy=50.686404	 loss=2.103168	 lr=0.010000
Epoch[008] Batch [2899]/[3759]	Speed: 35.115229 samples/sec	 accuracy=50.696659	 loss=2.103819	 lr=0.010000
Epoch[008] Batch [2949]/[3759]	Speed: 35.179951 samples/sec	 accuracy=50.699153	 loss=2.103839	 lr=0.010000
Epoch[008] Batch [2999]/[3759]	Speed: 34.949514 samples/sec	 accuracy=50.705729	 loss=2.103617	 lr=0.010000
Epoch[008] Batch [3049]/[3759]	Speed: 35.269807 samples/sec	 accuracy=50.706967	 loss=2.103457	 lr=0.010000
Epoch[008] Batch [3099]/[3759]	Speed: 34.817178 samples/sec	 accuracy=50.706149	 loss=2.103391	 lr=0.010000
Epoch[008] Batch [3149]/[3759]	Speed: 35.471880 samples/sec	 accuracy=50.713294	 loss=2.103176	 lr=0.010000
Epoch[008] Batch [3199]/[3759]	Speed: 35.041764 samples/sec	 accuracy=50.712402	 loss=2.103899	 lr=0.010000
Epoch[008] Batch [3249]/[3759]	Speed: 35.371901 samples/sec	 accuracy=50.697596	 loss=2.104517	 lr=0.010000
Epoch[008] Batch [3299]/[3759]	Speed: 35.393770 samples/sec	 accuracy=50.696023	 loss=2.104711	 lr=0.010000
Epoch[008] Batch [3349]/[3759]	Speed: 35.435739 samples/sec	 accuracy=50.703825	 loss=2.105026	 lr=0.010000
Epoch[008] Batch [3399]/[3759]	Speed: 35.166659 samples/sec	 accuracy=50.686581	 loss=2.105447	 lr=0.010000
Epoch[008] Batch [3449]/[3759]	Speed: 35.304620 samples/sec	 accuracy=50.673913	 loss=2.106071	 lr=0.010000
Epoch[008] Batch [3499]/[3759]	Speed: 35.068943 samples/sec	 accuracy=50.669196	 loss=2.106281	 lr=0.010000
Epoch[008] Batch [3549]/[3759]	Speed: 35.159237 samples/sec	 accuracy=50.676496	 loss=2.106509	 lr=0.010000
Epoch[008] Batch [3599]/[3759]	Speed: 35.311893 samples/sec	 accuracy=50.666667	 loss=2.106675	 lr=0.010000
Epoch[008] Batch [3649]/[3759]	Speed: 35.171145 samples/sec	 accuracy=50.668236	 loss=2.106316	 lr=0.010000
Epoch[008] Batch [3699]/[3759]	Speed: 35.331456 samples/sec	 accuracy=50.677365	 loss=2.106055	 lr=0.010000
Epoch[008] Batch [3749]/[3759]	Speed: 36.331504 samples/sec	 accuracy=50.667083	 loss=2.106542	 lr=0.010000
Batch [0049]/[0303]: acc-top1=51.375000 acc-top5=76.750000
Batch [0099]/[0303]: acc-top1=51.546875 acc-top5=76.375000
Batch [0149]/[0303]: acc-top1=51.875000 acc-top5=76.739583
Batch [0199]/[0303]: acc-top1=51.742188 acc-top5=76.890625
Batch [0249]/[0303]: acc-top1=51.587500 acc-top5=76.556250
Batch [0299]/[0303]: acc-top1=51.713542 acc-top5=76.447917
[Epoch 008] training: accuracy=50.677125	 loss=2.106185
[Epoch 008] speed: 35 samples/sec	time cost: 7359.253261
[Epoch 008] validation: acc-top1=51.650165 acc-top5=76.418111 loss=2.294714
Epoch[009] Batch [0049]/[3760]	Speed: 23.975858 samples/sec	 accuracy=51.093750	 loss=2.086090	 lr=0.010000
Epoch[009] Batch [0099]/[3760]	Speed: 34.859701 samples/sec	 accuracy=51.890625	 loss=2.025543	 lr=0.010000
Epoch[009] Batch [0149]/[3760]	Speed: 35.165926 samples/sec	 accuracy=52.156250	 loss=2.008072	 lr=0.010000
Epoch[009] Batch [0199]/[3760]	Speed: 34.953119 samples/sec	 accuracy=52.039062	 loss=2.015579	 lr=0.010000
Epoch[009] Batch [0249]/[3760]	Speed: 35.435053 samples/sec	 accuracy=51.943750	 loss=2.016207	 lr=0.010000
Epoch[009] Batch [0299]/[3760]	Speed: 34.906336 samples/sec	 accuracy=51.843750	 loss=2.020276	 lr=0.010000
Epoch[009] Batch [0349]/[3760]	Speed: 35.380125 samples/sec	 accuracy=51.593750	 loss=2.035005	 lr=0.010000
Epoch[009] Batch [0399]/[3760]	Speed: 35.047929 samples/sec	 accuracy=51.519531	 loss=2.040648	 lr=0.010000
Epoch[009] Batch [0449]/[3760]	Speed: 35.357866 samples/sec	 accuracy=51.791667	 loss=2.032429	 lr=0.010000
Epoch[009] Batch [0499]/[3760]	Speed: 35.230949 samples/sec	 accuracy=51.871875	 loss=2.033672	 lr=0.010000
Epoch[009] Batch [0549]/[3760]	Speed: 35.198229 samples/sec	 accuracy=51.829545	 loss=2.034586	 lr=0.010000
Epoch[009] Batch [0599]/[3760]	Speed: 35.161829 samples/sec	 accuracy=51.906250	 loss=2.030856	 lr=0.010000
Epoch[009] Batch [0649]/[3760]	Speed: 35.039143 samples/sec	 accuracy=51.831731	 loss=2.033630	 lr=0.010000
Epoch[009] Batch [0699]/[3760]	Speed: 35.246463 samples/sec	 accuracy=51.698661	 loss=2.042059	 lr=0.010000
Epoch[009] Batch [0749]/[3760]	Speed: 34.949740 samples/sec	 accuracy=51.660417	 loss=2.044121	 lr=0.010000
Epoch[009] Batch [0799]/[3760]	Speed: 35.171557 samples/sec	 accuracy=51.716797	 loss=2.042835	 lr=0.010000
Epoch[009] Batch [0849]/[3760]	Speed: 34.997234 samples/sec	 accuracy=51.670956	 loss=2.047020	 lr=0.010000
Epoch[009] Batch [0899]/[3760]	Speed: 34.982417 samples/sec	 accuracy=51.765625	 loss=2.045034	 lr=0.010000
Epoch[009] Batch [0949]/[3760]	Speed: 35.290701 samples/sec	 accuracy=51.828947	 loss=2.043907	 lr=0.010000
Epoch[009] Batch [0999]/[3760]	Speed: 34.980367 samples/sec	 accuracy=51.798438	 loss=2.043294	 lr=0.010000
Epoch[009] Batch [1049]/[3760]	Speed: 35.179456 samples/sec	 accuracy=51.788690	 loss=2.043545	 lr=0.010000
Epoch[009] Batch [1099]/[3760]	Speed: 35.116737 samples/sec	 accuracy=51.833807	 loss=2.043728	 lr=0.010000
Epoch[009] Batch [1149]/[3760]	Speed: 35.167282 samples/sec	 accuracy=51.785326	 loss=2.045829	 lr=0.010000
Epoch[009] Batch [1199]/[3760]	Speed: 35.117435 samples/sec	 accuracy=51.777344	 loss=2.047459	 lr=0.010000
Epoch[009] Batch [1249]/[3760]	Speed: 35.246988 samples/sec	 accuracy=51.795000	 loss=2.046687	 lr=0.010000
Epoch[009] Batch [1299]/[3760]	Speed: 35.107589 samples/sec	 accuracy=51.816106	 loss=2.046602	 lr=0.010000
Epoch[009] Batch [1349]/[3760]	Speed: 35.149729 samples/sec	 accuracy=51.792824	 loss=2.046758	 lr=0.010000
Epoch[009] Batch [1399]/[3760]	Speed: 35.276113 samples/sec	 accuracy=51.830357	 loss=2.045256	 lr=0.010000
Epoch[009] Batch [1449]/[3760]	Speed: 34.789178 samples/sec	 accuracy=51.797414	 loss=2.045094	 lr=0.010000
Epoch[009] Batch [1499]/[3760]	Speed: 35.023919 samples/sec	 accuracy=51.810417	 loss=2.044736	 lr=0.010000
Epoch[009] Batch [1549]/[3760]	Speed: 35.039134 samples/sec	 accuracy=51.789315	 loss=2.044310	 lr=0.010000
Epoch[009] Batch [1599]/[3760]	Speed: 35.191605 samples/sec	 accuracy=51.753906	 loss=2.046582	 lr=0.010000
Epoch[009] Batch [1649]/[3760]	Speed: 35.304532 samples/sec	 accuracy=51.732955	 loss=2.047208	 lr=0.010000
Epoch[009] Batch [1699]/[3760]	Speed: 34.880035 samples/sec	 accuracy=51.776654	 loss=2.046346	 lr=0.010000
Epoch[009] Batch [1749]/[3760]	Speed: 35.347190 samples/sec	 accuracy=51.740179	 loss=2.047245	 lr=0.010000
Epoch[009] Batch [1799]/[3760]	Speed: 34.826582 samples/sec	 accuracy=51.742188	 loss=2.047963	 lr=0.010000
Epoch[009] Batch [1849]/[3760]	Speed: 35.215813 samples/sec	 accuracy=51.770270	 loss=2.048084	 lr=0.010000
Epoch[009] Batch [1899]/[3760]	Speed: 35.266084 samples/sec	 accuracy=51.747533	 loss=2.049159	 lr=0.010000
Epoch[009] Batch [1949]/[3760]	Speed: 35.097430 samples/sec	 accuracy=51.772436	 loss=2.049943	 lr=0.010000
Epoch[009] Batch [1999]/[3760]	Speed: 35.178416 samples/sec	 accuracy=51.749219	 loss=2.051920	 lr=0.010000
Epoch[009] Batch [2049]/[3760]	Speed: 35.196146 samples/sec	 accuracy=51.756098	 loss=2.052096	 lr=0.010000
Epoch[009] Batch [2099]/[3760]	Speed: 35.353926 samples/sec	 accuracy=51.744792	 loss=2.051889	 lr=0.010000
Epoch[009] Batch [2149]/[3760]	Speed: 35.045239 samples/sec	 accuracy=51.722384	 loss=2.052353	 lr=0.010000
Epoch[009] Batch [2199]/[3760]	Speed: 35.277068 samples/sec	 accuracy=51.730114	 loss=2.051390	 lr=0.010000
Epoch[009] Batch [2249]/[3760]	Speed: 35.240116 samples/sec	 accuracy=51.737500	 loss=2.051289	 lr=0.010000
Epoch[009] Batch [2299]/[3760]	Speed: 35.406107 samples/sec	 accuracy=51.729620	 loss=2.051380	 lr=0.010000
Epoch[009] Batch [2349]/[3760]	Speed: 35.148398 samples/sec	 accuracy=51.700133	 loss=2.052641	 lr=0.010000
Epoch[009] Batch [2399]/[3760]	Speed: 35.084267 samples/sec	 accuracy=51.677083	 loss=2.053402	 lr=0.010000
Epoch[009] Batch [2449]/[3760]	Speed: 35.168523 samples/sec	 accuracy=51.657526	 loss=2.054463	 lr=0.010000
Epoch[009] Batch [2499]/[3760]	Speed: 35.069707 samples/sec	 accuracy=51.666250	 loss=2.054688	 lr=0.010000
Epoch[009] Batch [2549]/[3760]	Speed: 35.084623 samples/sec	 accuracy=51.679534	 loss=2.054836	 lr=0.010000
Epoch[009] Batch [2599]/[3760]	Speed: 35.122493 samples/sec	 accuracy=51.688702	 loss=2.055075	 lr=0.010000
Epoch[009] Batch [2649]/[3760]	Speed: 35.163015 samples/sec	 accuracy=51.677476	 loss=2.055789	 lr=0.010000
Epoch[009] Batch [2699]/[3760]	Speed: 35.169893 samples/sec	 accuracy=51.652778	 loss=2.056891	 lr=0.010000
Epoch[009] Batch [2749]/[3760]	Speed: 35.268025 samples/sec	 accuracy=51.639773	 loss=2.057134	 lr=0.010000
Epoch[009] Batch [2799]/[3760]	Speed: 34.970474 samples/sec	 accuracy=51.628348	 loss=2.057719	 lr=0.010000
Epoch[009] Batch [2849]/[3760]	Speed: 35.177046 samples/sec	 accuracy=51.598136	 loss=2.059077	 lr=0.010000
Epoch[009] Batch [2899]/[3760]	Speed: 34.805241 samples/sec	 accuracy=51.598060	 loss=2.059815	 lr=0.010000
Epoch[009] Batch [2949]/[3760]	Speed: 34.633412 samples/sec	 accuracy=51.602754	 loss=2.059529	 lr=0.010000
Epoch[009] Batch [2999]/[3760]	Speed: 34.957569 samples/sec	 accuracy=51.597396	 loss=2.059763	 lr=0.010000
Epoch[009] Batch [3049]/[3760]	Speed: 34.373288 samples/sec	 accuracy=51.609631	 loss=2.059630	 lr=0.010000
Epoch[009] Batch [3099]/[3760]	Speed: 35.124587 samples/sec	 accuracy=51.598286	 loss=2.058979	 lr=0.010000
Epoch[009] Batch [3149]/[3760]	Speed: 34.707429 samples/sec	 accuracy=51.590278	 loss=2.058908	 lr=0.010000
Epoch[009] Batch [3199]/[3760]	Speed: 35.271329 samples/sec	 accuracy=51.606934	 loss=2.058677	 lr=0.010000
Epoch[009] Batch [3249]/[3760]	Speed: 35.045578 samples/sec	 accuracy=51.604327	 loss=2.058345	 lr=0.010000
Epoch[009] Batch [3299]/[3760]	Speed: 35.292287 samples/sec	 accuracy=51.607008	 loss=2.058374	 lr=0.010000
Epoch[009] Batch [3349]/[3760]	Speed: 35.278686 samples/sec	 accuracy=51.596082	 loss=2.058942	 lr=0.010000
Epoch[009] Batch [3399]/[3760]	Speed: 35.193271 samples/sec	 accuracy=51.590993	 loss=2.058730	 lr=0.010000
Epoch[009] Batch [3449]/[3760]	Speed: 35.206441 samples/sec	 accuracy=51.596467	 loss=2.058689	 lr=0.010000
Epoch[009] Batch [3499]/[3760]	Speed: 35.212628 samples/sec	 accuracy=51.589286	 loss=2.059036	 lr=0.010000
Epoch[009] Batch [3549]/[3760]	Speed: 35.295155 samples/sec	 accuracy=51.573504	 loss=2.059895	 lr=0.010000
Epoch[009] Batch [3599]/[3760]	Speed: 34.976543 samples/sec	 accuracy=51.589410	 loss=2.059745	 lr=0.010000
Epoch[009] Batch [3649]/[3760]	Speed: 35.198230 samples/sec	 accuracy=51.595034	 loss=2.058926	 lr=0.010000
Epoch[009] Batch [3699]/[3760]	Speed: 35.214321 samples/sec	 accuracy=51.613176	 loss=2.058436	 lr=0.010000
Epoch[009] Batch [3749]/[3760]	Speed: 36.404134 samples/sec	 accuracy=51.607500	 loss=2.058949	 lr=0.010000
Batch [0049]/[0303]: acc-top1=51.187500 acc-top5=77.312500
Batch [0099]/[0303]: acc-top1=51.750000 acc-top5=77.046875
Batch [0149]/[0303]: acc-top1=52.260417 acc-top5=77.187500
Batch [0199]/[0303]: acc-top1=52.125000 acc-top5=77.085938
Batch [0249]/[0303]: acc-top1=52.087500 acc-top5=76.775000
Batch [0299]/[0303]: acc-top1=52.046875 acc-top5=76.671875
[Epoch 009] training: accuracy=51.603640	 loss=2.059334
[Epoch 009] speed: 34 samples/sec	time cost: 7377.297972
[Epoch 009] validation: acc-top1=51.990512 acc-top5=76.665635 loss=2.268471
Epoch[010] Batch [0049]/[3760]	Speed: 23.894773 samples/sec	 accuracy=53.625000	 loss=1.948988	 lr=0.010000
Epoch[010] Batch [0099]/[3760]	Speed: 34.833967 samples/sec	 accuracy=52.812500	 loss=1.990885	 lr=0.010000
Epoch[010] Batch [0149]/[3760]	Speed: 35.209302 samples/sec	 accuracy=52.843750	 loss=1.986946	 lr=0.010000
Epoch[010] Batch [0199]/[3760]	Speed: 34.908539 samples/sec	 accuracy=52.664062	 loss=1.998705	 lr=0.010000
Epoch[010] Batch [0249]/[3760]	Speed: 35.097456 samples/sec	 accuracy=52.325000	 loss=2.008852	 lr=0.010000
Epoch[010] Batch [0299]/[3760]	Speed: 34.640235 samples/sec	 accuracy=52.630208	 loss=1.999037	 lr=0.010000
Epoch[010] Batch [0349]/[3760]	Speed: 35.505710 samples/sec	 accuracy=52.910714	 loss=1.983061	 lr=0.010000
Epoch[010] Batch [0399]/[3760]	Speed: 34.900309 samples/sec	 accuracy=53.039062	 loss=1.980974	 lr=0.010000
Epoch[010] Batch [0449]/[3760]	Speed: 35.238807 samples/sec	 accuracy=53.072917	 loss=1.982229	 lr=0.010000
Epoch[010] Batch [0499]/[3760]	Speed: 35.309880 samples/sec	 accuracy=52.962500	 loss=1.984539	 lr=0.010000
Epoch[010] Batch [0549]/[3760]	Speed: 35.189392 samples/sec	 accuracy=52.991477	 loss=1.985536	 lr=0.010000
Epoch[010] Batch [0599]/[3760]	Speed: 35.310511 samples/sec	 accuracy=52.950521	 loss=1.985381	 lr=0.010000
Epoch[010] Batch [0649]/[3760]	Speed: 35.215640 samples/sec	 accuracy=52.786058	 loss=1.991605	 lr=0.010000
Epoch[010] Batch [0699]/[3760]	Speed: 35.349347 samples/sec	 accuracy=52.743304	 loss=1.994217	 lr=0.010000
Epoch[010] Batch [0749]/[3760]	Speed: 34.996526 samples/sec	 accuracy=52.729167	 loss=1.995800	 lr=0.010000
Epoch[010] Batch [0799]/[3760]	Speed: 35.449478 samples/sec	 accuracy=52.705078	 loss=1.996477	 lr=0.010000
Epoch[010] Batch [0849]/[3760]	Speed: 35.290314 samples/sec	 accuracy=52.669118	 loss=1.997217	 lr=0.010000
Epoch[010] Batch [0899]/[3760]	Speed: 35.351758 samples/sec	 accuracy=52.675347	 loss=1.996871	 lr=0.010000
Epoch[010] Batch [0949]/[3760]	Speed: 35.277367 samples/sec	 accuracy=52.680921	 loss=1.998718	 lr=0.010000
Epoch[010] Batch [0999]/[3760]	Speed: 35.088420 samples/sec	 accuracy=52.700000	 loss=1.997734	 lr=0.010000
Epoch[010] Batch [1049]/[3760]	Speed: 35.511256 samples/sec	 accuracy=52.663690	 loss=1.999957	 lr=0.010000
Epoch[010] Batch [1099]/[3760]	Speed: 35.142845 samples/sec	 accuracy=52.634943	 loss=2.001409	 lr=0.010000
Epoch[010] Batch [1149]/[3760]	Speed: 35.235656 samples/sec	 accuracy=52.631793	 loss=2.000955	 lr=0.010000
Epoch[010] Batch [1199]/[3760]	Speed: 35.396725 samples/sec	 accuracy=52.544271	 loss=2.003962	 lr=0.010000
Epoch[010] Batch [1249]/[3760]	Speed: 35.421657 samples/sec	 accuracy=52.498750	 loss=2.006417	 lr=0.010000
Epoch[010] Batch [1299]/[3760]	Speed: 35.275167 samples/sec	 accuracy=52.542067	 loss=2.006065	 lr=0.010000
Epoch[010] Batch [1349]/[3760]	Speed: 35.153797 samples/sec	 accuracy=52.539352	 loss=2.007308	 lr=0.010000
Epoch[010] Batch [1399]/[3760]	Speed: 35.407811 samples/sec	 accuracy=52.496652	 loss=2.010391	 lr=0.010000
Epoch[010] Batch [1449]/[3760]	Speed: 34.970280 samples/sec	 accuracy=52.488147	 loss=2.011230	 lr=0.010000
Epoch[010] Batch [1499]/[3760]	Speed: 35.516058 samples/sec	 accuracy=52.492708	 loss=2.013208	 lr=0.010000
Epoch[010] Batch [1549]/[3760]	Speed: 35.043298 samples/sec	 accuracy=52.497984	 loss=2.012642	 lr=0.010000
Epoch[010] Batch [1599]/[3760]	Speed: 35.316027 samples/sec	 accuracy=52.500977	 loss=2.012234	 lr=0.010000
Epoch[010] Batch [1649]/[3760]	Speed: 35.119894 samples/sec	 accuracy=52.509470	 loss=2.011925	 lr=0.010000
Epoch[010] Batch [1699]/[3760]	Speed: 35.208992 samples/sec	 accuracy=52.465993	 loss=2.013383	 lr=0.010000
Epoch[010] Batch [1749]/[3760]	Speed: 35.083926 samples/sec	 accuracy=52.483036	 loss=2.013162	 lr=0.010000
Epoch[010] Batch [1799]/[3760]	Speed: 35.103848 samples/sec	 accuracy=52.506076	 loss=2.012782	 lr=0.010000
Epoch[010] Batch [1849]/[3760]	Speed: 34.902989 samples/sec	 accuracy=52.503378	 loss=2.014053	 lr=0.010000
Epoch[010] Batch [1899]/[3760]	Speed: 35.107532 samples/sec	 accuracy=52.475329	 loss=2.014858	 lr=0.010000
Epoch[010] Batch [1949]/[3760]	Speed: 35.141005 samples/sec	 accuracy=52.462340	 loss=2.016184	 lr=0.010000
Epoch[010] Batch [1999]/[3760]	Speed: 34.516146 samples/sec	 accuracy=52.432031	 loss=2.017196	 lr=0.010000
Epoch[010] Batch [2049]/[3760]	Speed: 35.185297 samples/sec	 accuracy=52.432165	 loss=2.016820	 lr=0.010000
Epoch[010] Batch [2099]/[3760]	Speed: 34.240098 samples/sec	 accuracy=52.436012	 loss=2.016646	 lr=0.010000
Epoch[010] Batch [2149]/[3760]	Speed: 35.194534 samples/sec	 accuracy=52.446948	 loss=2.016659	 lr=0.010000
Epoch[010] Batch [2199]/[3760]	Speed: 34.487358 samples/sec	 accuracy=52.442472	 loss=2.016157	 lr=0.010000
Epoch[010] Batch [2249]/[3760]	Speed: 34.872441 samples/sec	 accuracy=52.432639	 loss=2.016272	 lr=0.010000
Epoch[010] Batch [2299]/[3760]	Speed: 34.652966 samples/sec	 accuracy=52.447011	 loss=2.015603	 lr=0.010000
Epoch[010] Batch [2349]/[3760]	Speed: 34.714106 samples/sec	 accuracy=52.486037	 loss=2.014557	 lr=0.010000
Epoch[010] Batch [2399]/[3760]	Speed: 35.116020 samples/sec	 accuracy=52.480469	 loss=2.014785	 lr=0.010000
Epoch[010] Batch [2449]/[3760]	Speed: 34.395747 samples/sec	 accuracy=52.468112	 loss=2.015155	 lr=0.010000
Epoch[010] Batch [2499]/[3760]	Speed: 35.149665 samples/sec	 accuracy=52.486250	 loss=2.013417	 lr=0.010000
Epoch[010] Batch [2549]/[3760]	Speed: 34.364515 samples/sec	 accuracy=52.494485	 loss=2.013135	 lr=0.010000
Epoch[010] Batch [2599]/[3760]	Speed: 35.175898 samples/sec	 accuracy=52.480168	 loss=2.013323	 lr=0.010000
Epoch[010] Batch [2649]/[3760]	Speed: 34.477949 samples/sec	 accuracy=52.502948	 loss=2.012502	 lr=0.010000
Epoch[010] Batch [2699]/[3760]	Speed: 34.894889 samples/sec	 accuracy=52.471644	 loss=2.013750	 lr=0.010000
Epoch[010] Batch [2749]/[3760]	Speed: 35.342392 samples/sec	 accuracy=52.448864	 loss=2.014696	 lr=0.010000
Epoch[010] Batch [2799]/[3760]	Speed: 34.799572 samples/sec	 accuracy=52.441964	 loss=2.015078	 lr=0.010000
Epoch[010] Batch [2849]/[3760]	Speed: 35.358963 samples/sec	 accuracy=52.439693	 loss=2.015286	 lr=0.010000
Epoch[010] Batch [2899]/[3760]	Speed: 34.692103 samples/sec	 accuracy=52.432112	 loss=2.014935	 lr=0.010000
Epoch[010] Batch [2949]/[3760]	Speed: 35.589170 samples/sec	 accuracy=52.435911	 loss=2.015228	 lr=0.010000
Epoch[010] Batch [2999]/[3760]	Speed: 35.026816 samples/sec	 accuracy=52.458333	 loss=2.015043	 lr=0.010000
Epoch[010] Batch [3049]/[3760]	Speed: 35.276903 samples/sec	 accuracy=52.448258	 loss=2.015611	 lr=0.010000
Epoch[010] Batch [3099]/[3760]	Speed: 35.257243 samples/sec	 accuracy=52.457661	 loss=2.015538	 lr=0.010000
Epoch[010] Batch [3149]/[3760]	Speed: 35.289993 samples/sec	 accuracy=52.448909	 loss=2.016374	 lr=0.010000
Epoch[010] Batch [3199]/[3760]	Speed: 35.045089 samples/sec	 accuracy=52.444824	 loss=2.016708	 lr=0.010000
Epoch[010] Batch [3249]/[3760]	Speed: 35.315682 samples/sec	 accuracy=52.478365	 loss=2.015546	 lr=0.010000
Epoch[010] Batch [3299]/[3760]	Speed: 35.061630 samples/sec	 accuracy=52.506629	 loss=2.014961	 lr=0.010000
Epoch[010] Batch [3349]/[3760]	Speed: 35.184736 samples/sec	 accuracy=52.493470	 loss=2.015816	 lr=0.010000
Epoch[010] Batch [3399]/[3760]	Speed: 35.199510 samples/sec	 accuracy=52.495404	 loss=2.016379	 lr=0.010000
Epoch[010] Batch [3449]/[3760]	Speed: 35.264923 samples/sec	 accuracy=52.509964	 loss=2.015502	 lr=0.010000
Epoch[010] Batch [3499]/[3760]	Speed: 35.304674 samples/sec	 accuracy=52.512054	 loss=2.015676	 lr=0.010000
Epoch[010] Batch [3549]/[3760]	Speed: 35.242413 samples/sec	 accuracy=52.505282	 loss=2.015398	 lr=0.010000
Epoch[010] Batch [3599]/[3760]	Speed: 35.144352 samples/sec	 accuracy=52.489583	 loss=2.015742	 lr=0.010000
Epoch[010] Batch [3649]/[3760]	Speed: 35.113295 samples/sec	 accuracy=52.469606	 loss=2.016159	 lr=0.010000
Epoch[010] Batch [3699]/[3760]	Speed: 35.164742 samples/sec	 accuracy=52.492399	 loss=2.015638	 lr=0.010000
Epoch[010] Batch [3749]/[3760]	Speed: 36.315213 samples/sec	 accuracy=52.497500	 loss=2.015845	 lr=0.010000
Batch [0049]/[0303]: acc-top1=52.625000 acc-top5=77.656250
Batch [0099]/[0303]: acc-top1=52.828125 acc-top5=77.562500
Batch [0149]/[0303]: acc-top1=53.104167 acc-top5=77.750000
Batch [0199]/[0303]: acc-top1=52.726562 acc-top5=77.554688
Batch [0249]/[0303]: acc-top1=52.443750 acc-top5=77.381250
Batch [0299]/[0303]: acc-top1=52.468750 acc-top5=77.218750
[Epoch 010] training: accuracy=52.487118	 loss=2.016059
[Epoch 010] speed: 34 samples/sec	time cost: 7380.927936
[Epoch 010] validation: acc-top1=52.444307 acc-top5=77.196782 loss=2.197747
Epoch[011] Batch [0049]/[3759]	Speed: 24.496205 samples/sec	 accuracy=53.468750	 loss=1.931309	 lr=0.010000
Epoch[011] Batch [0099]/[3759]	Speed: 35.063691 samples/sec	 accuracy=54.734375	 loss=1.889061	 lr=0.010000
Epoch[011] Batch [0149]/[3759]	Speed: 35.231256 samples/sec	 accuracy=54.968750	 loss=1.889111	 lr=0.010000
Epoch[011] Batch [0199]/[3759]	Speed: 35.142337 samples/sec	 accuracy=54.960937	 loss=1.895827	 lr=0.010000
Epoch[011] Batch [0249]/[3759]	Speed: 35.422263 samples/sec	 accuracy=54.718750	 loss=1.908234	 lr=0.010000
Epoch[011] Batch [0299]/[3759]	Speed: 34.950186 samples/sec	 accuracy=54.723958	 loss=1.913139	 lr=0.010000
Epoch[011] Batch [0349]/[3759]	Speed: 35.403861 samples/sec	 accuracy=54.714286	 loss=1.914619	 lr=0.010000
Epoch[011] Batch [0399]/[3759]	Speed: 34.987172 samples/sec	 accuracy=54.359375	 loss=1.922640	 lr=0.010000
Epoch[011] Batch [0449]/[3759]	Speed: 35.408618 samples/sec	 accuracy=54.277778	 loss=1.922498	 lr=0.010000
Epoch[011] Batch [0499]/[3759]	Speed: 35.098542 samples/sec	 accuracy=54.209375	 loss=1.926133	 lr=0.010000
Epoch[011] Batch [0549]/[3759]	Speed: 35.349796 samples/sec	 accuracy=54.161932	 loss=1.930124	 lr=0.010000
Epoch[011] Batch [0599]/[3759]	Speed: 35.193282 samples/sec	 accuracy=53.992188	 loss=1.936401	 lr=0.010000
Epoch[011] Batch [0649]/[3759]	Speed: 35.175497 samples/sec	 accuracy=53.894231	 loss=1.939953	 lr=0.010000
Epoch[011] Batch [0699]/[3759]	Speed: 35.313806 samples/sec	 accuracy=53.933036	 loss=1.939391	 lr=0.010000
Epoch[011] Batch [0749]/[3759]	Speed: 35.101635 samples/sec	 accuracy=53.872917	 loss=1.942550	 lr=0.010000
Epoch[011] Batch [0799]/[3759]	Speed: 35.302448 samples/sec	 accuracy=53.837891	 loss=1.942166	 lr=0.010000
Epoch[011] Batch [0849]/[3759]	Speed: 35.036474 samples/sec	 accuracy=53.779412	 loss=1.943372	 lr=0.010000
Epoch[011] Batch [0899]/[3759]	Speed: 35.344619 samples/sec	 accuracy=53.755208	 loss=1.944794	 lr=0.010000
Epoch[011] Batch [0949]/[3759]	Speed: 34.912870 samples/sec	 accuracy=53.707237	 loss=1.943923	 lr=0.010000
Epoch[011] Batch [0999]/[3759]	Speed: 35.255666 samples/sec	 accuracy=53.664062	 loss=1.946118	 lr=0.010000
Epoch[011] Batch [1049]/[3759]	Speed: 35.210172 samples/sec	 accuracy=53.635417	 loss=1.947375	 lr=0.010000
Epoch[011] Batch [1099]/[3759]	Speed: 35.155177 samples/sec	 accuracy=53.634943	 loss=1.947074	 lr=0.010000
Epoch[011] Batch [1149]/[3759]	Speed: 35.145148 samples/sec	 accuracy=53.600543	 loss=1.946823	 lr=0.010000
Epoch[011] Batch [1199]/[3759]	Speed: 35.295426 samples/sec	 accuracy=53.678385	 loss=1.944786	 lr=0.010000
Epoch[011] Batch [1249]/[3759]	Speed: 35.123671 samples/sec	 accuracy=53.643750	 loss=1.947360	 lr=0.010000
Epoch[011] Batch [1299]/[3759]	Speed: 35.364221 samples/sec	 accuracy=53.680288	 loss=1.946884	 lr=0.010000
Epoch[011] Batch [1349]/[3759]	Speed: 35.071691 samples/sec	 accuracy=53.689815	 loss=1.946961	 lr=0.010000
Epoch[011] Batch [1399]/[3759]	Speed: 35.214231 samples/sec	 accuracy=53.677455	 loss=1.948937	 lr=0.010000
Epoch[011] Batch [1449]/[3759]	Speed: 35.142493 samples/sec	 accuracy=53.703664	 loss=1.948669	 lr=0.010000
Epoch[011] Batch [1499]/[3759]	Speed: 35.322742 samples/sec	 accuracy=53.710417	 loss=1.946356	 lr=0.010000
Epoch[011] Batch [1549]/[3759]	Speed: 35.041325 samples/sec	 accuracy=53.652218	 loss=1.948907	 lr=0.010000
Epoch[011] Batch [1599]/[3759]	Speed: 35.283327 samples/sec	 accuracy=53.671875	 loss=1.949027	 lr=0.010000
Epoch[011] Batch [1649]/[3759]	Speed: 35.249722 samples/sec	 accuracy=53.670455	 loss=1.950112	 lr=0.010000
Epoch[011] Batch [1699]/[3759]	Speed: 34.975040 samples/sec	 accuracy=53.642463	 loss=1.951558	 lr=0.010000
Epoch[011] Batch [1749]/[3759]	Speed: 35.275463 samples/sec	 accuracy=53.649107	 loss=1.951277	 lr=0.010000
Epoch[011] Batch [1799]/[3759]	Speed: 34.936775 samples/sec	 accuracy=53.638889	 loss=1.952776	 lr=0.010000
Epoch[011] Batch [1849]/[3759]	Speed: 35.481193 samples/sec	 accuracy=53.622466	 loss=1.953633	 lr=0.010000
Epoch[011] Batch [1899]/[3759]	Speed: 34.606598 samples/sec	 accuracy=53.628289	 loss=1.954411	 lr=0.010000
Epoch[011] Batch [1949]/[3759]	Speed: 35.257089 samples/sec	 accuracy=53.605769	 loss=1.955133	 lr=0.010000
Epoch[011] Batch [1999]/[3759]	Speed: 34.457323 samples/sec	 accuracy=53.566406	 loss=1.956928	 lr=0.010000
Epoch[011] Batch [2049]/[3759]	Speed: 35.289715 samples/sec	 accuracy=53.536585	 loss=1.958551	 lr=0.010000
Epoch[011] Batch [2099]/[3759]	Speed: 34.474630 samples/sec	 accuracy=53.502232	 loss=1.960111	 lr=0.010000
Epoch[011] Batch [2149]/[3759]	Speed: 34.740144 samples/sec	 accuracy=53.530523	 loss=1.960071	 lr=0.010000
Epoch[011] Batch [2199]/[3759]	Speed: 34.935071 samples/sec	 accuracy=53.533381	 loss=1.960515	 lr=0.010000
Epoch[011] Batch [2249]/[3759]	Speed: 34.476054 samples/sec	 accuracy=53.517361	 loss=1.960384	 lr=0.010000
Epoch[011] Batch [2299]/[3759]	Speed: 35.237308 samples/sec	 accuracy=53.507473	 loss=1.960949	 lr=0.010000
Epoch[011] Batch [2349]/[3759]	Speed: 34.564549 samples/sec	 accuracy=53.501330	 loss=1.961406	 lr=0.010000
Epoch[011] Batch [2399]/[3759]	Speed: 35.435285 samples/sec	 accuracy=53.457682	 loss=1.963624	 lr=0.010000
Epoch[011] Batch [2449]/[3759]	Speed: 34.791449 samples/sec	 accuracy=53.438138	 loss=1.964869	 lr=0.010000
Epoch[011] Batch [2499]/[3759]	Speed: 35.095881 samples/sec	 accuracy=53.416875	 loss=1.965541	 lr=0.010000
Epoch[011] Batch [2549]/[3759]	Speed: 34.697548 samples/sec	 accuracy=53.413603	 loss=1.966402	 lr=0.010000
Epoch[011] Batch [2599]/[3759]	Speed: 35.217587 samples/sec	 accuracy=53.387019	 loss=1.967770	 lr=0.010000
Epoch[011] Batch [2649]/[3759]	Speed: 35.039898 samples/sec	 accuracy=53.382665	 loss=1.967787	 lr=0.010000
Epoch[011] Batch [2699]/[3759]	Speed: 35.111855 samples/sec	 accuracy=53.357639	 loss=1.967952	 lr=0.010000
Epoch[011] Batch [2749]/[3759]	Speed: 35.396236 samples/sec	 accuracy=53.348295	 loss=1.968417	 lr=0.010000
Epoch[011] Batch [2799]/[3759]	Speed: 34.818157 samples/sec	 accuracy=53.373326	 loss=1.967617	 lr=0.010000
Epoch[011] Batch [2849]/[3759]	Speed: 35.436019 samples/sec	 accuracy=53.358004	 loss=1.968394	 lr=0.010000
Epoch[011] Batch [2899]/[3759]	Speed: 34.795179 samples/sec	 accuracy=53.351832	 loss=1.968864	 lr=0.010000
Epoch[011] Batch [2949]/[3759]	Speed: 35.248707 samples/sec	 accuracy=53.339513	 loss=1.969360	 lr=0.010000
Epoch[011] Batch [2999]/[3759]	Speed: 34.851035 samples/sec	 accuracy=53.341667	 loss=1.969144	 lr=0.010000
Epoch[011] Batch [3049]/[3759]	Speed: 35.509541 samples/sec	 accuracy=53.341189	 loss=1.968936	 lr=0.010000
Epoch[011] Batch [3099]/[3759]	Speed: 35.122933 samples/sec	 accuracy=53.351310	 loss=1.969032	 lr=0.010000
Epoch[011] Batch [3149]/[3759]	Speed: 35.248022 samples/sec	 accuracy=53.378968	 loss=1.967898	 lr=0.010000
Epoch[011] Batch [3199]/[3759]	Speed: 34.972668 samples/sec	 accuracy=53.365723	 loss=1.968391	 lr=0.010000
Epoch[011] Batch [3249]/[3759]	Speed: 35.571968 samples/sec	 accuracy=53.357692	 loss=1.968552	 lr=0.010000
Epoch[011] Batch [3299]/[3759]	Speed: 35.092637 samples/sec	 accuracy=53.344223	 loss=1.968918	 lr=0.010000
Epoch[011] Batch [3349]/[3759]	Speed: 35.354758 samples/sec	 accuracy=53.356810	 loss=1.968892	 lr=0.010000
Epoch[011] Batch [3399]/[3759]	Speed: 35.208781 samples/sec	 accuracy=53.329963	 loss=1.969983	 lr=0.010000
Epoch[011] Batch [3449]/[3759]	Speed: 35.340281 samples/sec	 accuracy=53.318388	 loss=1.970154	 lr=0.010000
Epoch[011] Batch [3499]/[3759]	Speed: 35.292407 samples/sec	 accuracy=53.296875	 loss=1.970745	 lr=0.010000
Epoch[011] Batch [3549]/[3759]	Speed: 35.290660 samples/sec	 accuracy=53.312500	 loss=1.970499	 lr=0.010000
Epoch[011] Batch [3599]/[3759]	Speed: 35.331182 samples/sec	 accuracy=53.314670	 loss=1.970801	 lr=0.010000
Epoch[011] Batch [3649]/[3759]	Speed: 35.302496 samples/sec	 accuracy=53.324914	 loss=1.970594	 lr=0.010000
Epoch[011] Batch [3699]/[3759]	Speed: 35.242291 samples/sec	 accuracy=53.334459	 loss=1.970591	 lr=0.010000
Epoch[011] Batch [3749]/[3759]	Speed: 36.465922 samples/sec	 accuracy=53.341250	 loss=1.970101	 lr=0.010000
Batch [0049]/[0303]: acc-top1=53.250000 acc-top5=78.468750
Batch [0099]/[0303]: acc-top1=53.656250 acc-top5=78.187500
Batch [0149]/[0303]: acc-top1=54.104167 acc-top5=78.385417
Batch [0199]/[0303]: acc-top1=53.562500 acc-top5=78.312500
Batch [0249]/[0303]: acc-top1=53.525000 acc-top5=78.075000
Batch [0299]/[0303]: acc-top1=53.552083 acc-top5=78.156250
[Epoch 011] training: accuracy=53.341564	 loss=1.970048
[Epoch 011] speed: 34 samples/sec	time cost: 7363.452621
[Epoch 011] validation: acc-top1=53.501444 acc-top5=78.150784 loss=2.210711
Epoch[012] Batch [0049]/[3760]	Speed: 24.038496 samples/sec	 accuracy=54.468750	 loss=1.902613	 lr=0.010000
Epoch[012] Batch [0099]/[3760]	Speed: 34.742907 samples/sec	 accuracy=54.421875	 loss=1.900471	 lr=0.010000
Epoch[012] Batch [0149]/[3760]	Speed: 35.174621 samples/sec	 accuracy=54.593750	 loss=1.903509	 lr=0.010000
Epoch[012] Batch [0199]/[3760]	Speed: 34.719440 samples/sec	 accuracy=54.554688	 loss=1.895440	 lr=0.010000
Epoch[012] Batch [0249]/[3760]	Speed: 35.145981 samples/sec	 accuracy=54.718750	 loss=1.889637	 lr=0.010000
Epoch[012] Batch [0299]/[3760]	Speed: 34.682510 samples/sec	 accuracy=54.578125	 loss=1.898552	 lr=0.010000
Epoch[012] Batch [0349]/[3760]	Speed: 35.150124 samples/sec	 accuracy=54.558036	 loss=1.897067	 lr=0.010000
Epoch[012] Batch [0399]/[3760]	Speed: 34.694639 samples/sec	 accuracy=54.824219	 loss=1.884286	 lr=0.010000
Epoch[012] Batch [0449]/[3760]	Speed: 35.350989 samples/sec	 accuracy=54.597222	 loss=1.893396	 lr=0.010000
Epoch[012] Batch [0499]/[3760]	Speed: 34.864498 samples/sec	 accuracy=54.553125	 loss=1.894841	 lr=0.010000
Epoch[012] Batch [0549]/[3760]	Speed: 35.167583 samples/sec	 accuracy=54.571023	 loss=1.897500	 lr=0.010000
Epoch[012] Batch [0599]/[3760]	Speed: 35.127023 samples/sec	 accuracy=54.604167	 loss=1.898093	 lr=0.010000
Epoch[012] Batch [0649]/[3760]	Speed: 35.171152 samples/sec	 accuracy=54.550481	 loss=1.900052	 lr=0.010000
Epoch[012] Batch [0699]/[3760]	Speed: 35.134293 samples/sec	 accuracy=54.566964	 loss=1.901177	 lr=0.010000
Epoch[012] Batch [0749]/[3760]	Speed: 35.164064 samples/sec	 accuracy=54.487500	 loss=1.905403	 lr=0.010000
Epoch[012] Batch [0799]/[3760]	Speed: 35.190914 samples/sec	 accuracy=54.414063	 loss=1.905891	 lr=0.010000
Epoch[012] Batch [0849]/[3760]	Speed: 35.020762 samples/sec	 accuracy=54.430147	 loss=1.906230	 lr=0.010000
Epoch[012] Batch [0899]/[3760]	Speed: 35.269211 samples/sec	 accuracy=54.503472	 loss=1.902334	 lr=0.010000
Epoch[012] Batch [0949]/[3760]	Speed: 35.042720 samples/sec	 accuracy=54.468750	 loss=1.904744	 lr=0.010000
Epoch[012] Batch [0999]/[3760]	Speed: 35.372071 samples/sec	 accuracy=54.526562	 loss=1.906393	 lr=0.010000
Epoch[012] Batch [1049]/[3760]	Speed: 35.271631 samples/sec	 accuracy=54.500000	 loss=1.909651	 lr=0.010000
Epoch[012] Batch [1099]/[3760]	Speed: 35.084011 samples/sec	 accuracy=54.518466	 loss=1.910076	 lr=0.010000
Epoch[012] Batch [1149]/[3760]	Speed: 35.467141 samples/sec	 accuracy=54.417120	 loss=1.912814	 lr=0.010000
Epoch[012] Batch [1199]/[3760]	Speed: 35.178298 samples/sec	 accuracy=54.388021	 loss=1.913794	 lr=0.010000
Epoch[012] Batch [1249]/[3760]	Speed: 35.222164 samples/sec	 accuracy=54.393750	 loss=1.913862	 lr=0.010000
Epoch[012] Batch [1299]/[3760]	Speed: 35.035392 samples/sec	 accuracy=54.378606	 loss=1.915138	 lr=0.010000
Epoch[012] Batch [1349]/[3760]	Speed: 35.312590 samples/sec	 accuracy=54.346065	 loss=1.915094	 lr=0.010000
Epoch[012] Batch [1399]/[3760]	Speed: 34.974775 samples/sec	 accuracy=54.348214	 loss=1.914490	 lr=0.010000
Epoch[012] Batch [1449]/[3760]	Speed: 35.454200 samples/sec	 accuracy=54.314655	 loss=1.915349	 lr=0.010000
Epoch[012] Batch [1499]/[3760]	Speed: 34.818412 samples/sec	 accuracy=54.258333	 loss=1.918111	 lr=0.010000
Epoch[012] Batch [1549]/[3760]	Speed: 35.499087 samples/sec	 accuracy=54.240927	 loss=1.920627	 lr=0.010000
Epoch[012] Batch [1599]/[3760]	Speed: 35.129328 samples/sec	 accuracy=54.188477	 loss=1.922718	 lr=0.010000
Epoch[012] Batch [1649]/[3760]	Speed: 35.415537 samples/sec	 accuracy=54.172348	 loss=1.923243	 lr=0.010000
Epoch[012] Batch [1699]/[3760]	Speed: 34.786642 samples/sec	 accuracy=54.182904	 loss=1.923265	 lr=0.010000
Epoch[012] Batch [1749]/[3760]	Speed: 35.405819 samples/sec	 accuracy=54.217857	 loss=1.922810	 lr=0.010000
Epoch[012] Batch [1799]/[3760]	Speed: 34.947653 samples/sec	 accuracy=54.202257	 loss=1.925024	 lr=0.010000
Epoch[012] Batch [1849]/[3760]	Speed: 35.305928 samples/sec	 accuracy=54.161318	 loss=1.925739	 lr=0.010000
Epoch[012] Batch [1899]/[3760]	Speed: 35.179484 samples/sec	 accuracy=54.147204	 loss=1.926582	 lr=0.010000
Epoch[012] Batch [1949]/[3760]	Speed: 35.362658 samples/sec	 accuracy=54.139423	 loss=1.927852	 lr=0.010000
Epoch[012] Batch [1999]/[3760]	Speed: 35.447866 samples/sec	 accuracy=54.122656	 loss=1.928602	 lr=0.010000
Epoch[012] Batch [2049]/[3760]	Speed: 35.385200 samples/sec	 accuracy=54.074695	 loss=1.930816	 lr=0.010000
Epoch[012] Batch [2099]/[3760]	Speed: 35.372142 samples/sec	 accuracy=54.029762	 loss=1.932124	 lr=0.010000
Epoch[012] Batch [2149]/[3760]	Speed: 35.356654 samples/sec	 accuracy=54.039971	 loss=1.931910	 lr=0.010000
Epoch[012] Batch [2199]/[3760]	Speed: 35.163355 samples/sec	 accuracy=54.032670	 loss=1.932680	 lr=0.010000
Epoch[012] Batch [2249]/[3760]	Speed: 35.259382 samples/sec	 accuracy=54.024306	 loss=1.933810	 lr=0.010000
Epoch[012] Batch [2299]/[3760]	Speed: 35.207247 samples/sec	 accuracy=54.009511	 loss=1.935229	 lr=0.010000
Epoch[012] Batch [2349]/[3760]	Speed: 35.270200 samples/sec	 accuracy=54.035239	 loss=1.934332	 lr=0.010000
Epoch[012] Batch [2399]/[3760]	Speed: 35.168885 samples/sec	 accuracy=54.034505	 loss=1.933903	 lr=0.010000
Epoch[012] Batch [2449]/[3760]	Speed: 35.349521 samples/sec	 accuracy=54.051658	 loss=1.933377	 lr=0.010000
Epoch[012] Batch [2499]/[3760]	Speed: 35.306765 samples/sec	 accuracy=54.055000	 loss=1.933761	 lr=0.010000
Epoch[012] Batch [2549]/[3760]	Speed: 34.871657 samples/sec	 accuracy=54.058824	 loss=1.933691	 lr=0.010000
Epoch[012] Batch [2599]/[3760]	Speed: 35.235034 samples/sec	 accuracy=54.054688	 loss=1.933662	 lr=0.010000
Epoch[012] Batch [2649]/[3760]	Speed: 34.896706 samples/sec	 accuracy=54.052476	 loss=1.933546	 lr=0.010000
Epoch[012] Batch [2699]/[3760]	Speed: 35.394627 samples/sec	 accuracy=54.079861	 loss=1.932362	 lr=0.010000
Epoch[012] Batch [2749]/[3760]	Speed: 34.823981 samples/sec	 accuracy=54.070455	 loss=1.932872	 lr=0.010000
Epoch[012] Batch [2799]/[3760]	Speed: 35.364220 samples/sec	 accuracy=54.088170	 loss=1.932251	 lr=0.010000
Epoch[012] Batch [2849]/[3760]	Speed: 34.924210 samples/sec	 accuracy=54.099232	 loss=1.932066	 lr=0.010000
Epoch[012] Batch [2899]/[3760]	Speed: 35.222636 samples/sec	 accuracy=54.078125	 loss=1.932603	 lr=0.010000
Epoch[012] Batch [2949]/[3760]	Speed: 35.131155 samples/sec	 accuracy=54.104343	 loss=1.931967	 lr=0.010000
Epoch[012] Batch [2999]/[3760]	Speed: 35.229711 samples/sec	 accuracy=54.097917	 loss=1.932006	 lr=0.010000
Epoch[012] Batch [3049]/[3760]	Speed: 35.301593 samples/sec	 accuracy=54.109119	 loss=1.931651	 lr=0.010000
Epoch[012] Batch [3099]/[3760]	Speed: 35.266411 samples/sec	 accuracy=54.090726	 loss=1.932801	 lr=0.010000
Epoch[012] Batch [3149]/[3760]	Speed: 35.365137 samples/sec	 accuracy=54.090774	 loss=1.932595	 lr=0.010000
Epoch[012] Batch [3199]/[3760]	Speed: 35.210486 samples/sec	 accuracy=54.077637	 loss=1.932975	 lr=0.010000
Epoch[012] Batch [3249]/[3760]	Speed: 35.130641 samples/sec	 accuracy=54.055769	 loss=1.934546	 lr=0.010000
Epoch[012] Batch [3299]/[3760]	Speed: 35.391628 samples/sec	 accuracy=54.058239	 loss=1.934360	 lr=0.010000
Epoch[012] Batch [3349]/[3760]	Speed: 35.128830 samples/sec	 accuracy=54.048507	 loss=1.934972	 lr=0.010000
Epoch[012] Batch [3399]/[3760]	Speed: 35.244745 samples/sec	 accuracy=54.048254	 loss=1.934750	 lr=0.010000
Epoch[012] Batch [3449]/[3760]	Speed: 35.099011 samples/sec	 accuracy=54.043931	 loss=1.935057	 lr=0.010000
Epoch[012] Batch [3499]/[3760]	Speed: 35.138323 samples/sec	 accuracy=54.048661	 loss=1.934977	 lr=0.010000
Epoch[012] Batch [3549]/[3760]	Speed: 34.854509 samples/sec	 accuracy=54.052377	 loss=1.935157	 lr=0.010000
Epoch[012] Batch [3599]/[3760]	Speed: 35.567302 samples/sec	 accuracy=54.037326	 loss=1.935630	 lr=0.010000
Epoch[012] Batch [3649]/[3760]	Speed: 34.762810 samples/sec	 accuracy=54.051370	 loss=1.935230	 lr=0.010000
Epoch[012] Batch [3699]/[3760]	Speed: 35.318781 samples/sec	 accuracy=54.046030	 loss=1.935445	 lr=0.010000
Epoch[012] Batch [3749]/[3760]	Speed: 36.174518 samples/sec	 accuracy=54.057917	 loss=1.935772	 lr=0.010000
Batch [0049]/[0303]: acc-top1=54.312500 acc-top5=78.562500
Batch [0099]/[0303]: acc-top1=55.171875 acc-top5=78.500000
Batch [0149]/[0303]: acc-top1=55.177083 acc-top5=78.760417
Batch [0199]/[0303]: acc-top1=55.234375 acc-top5=79.054688
Batch [0249]/[0303]: acc-top1=54.975000 acc-top5=78.918750
Batch [0299]/[0303]: acc-top1=54.916667 acc-top5=78.854167
[Epoch 012] training: accuracy=54.068318	 loss=1.935244
[Epoch 012] speed: 34 samples/sec	time cost: 7359.322441
[Epoch 012] validation: acc-top1=54.883457 acc-top5=78.846947 loss=2.176675
Epoch[013] Batch [0049]/[3760]	Speed: 24.457643 samples/sec	 accuracy=55.875000	 loss=1.837441	 lr=0.010000
Epoch[013] Batch [0099]/[3760]	Speed: 35.183322 samples/sec	 accuracy=55.906250	 loss=1.851465	 lr=0.010000
Epoch[013] Batch [0149]/[3760]	Speed: 35.170406 samples/sec	 accuracy=55.020833	 loss=1.879074	 lr=0.010000
Epoch[013] Batch [0199]/[3760]	Speed: 35.093438 samples/sec	 accuracy=55.742187	 loss=1.852830	 lr=0.010000
Epoch[013] Batch [0249]/[3760]	Speed: 35.129721 samples/sec	 accuracy=55.637500	 loss=1.854425	 lr=0.010000
Epoch[013] Batch [0299]/[3760]	Speed: 34.752528 samples/sec	 accuracy=55.744792	 loss=1.860884	 lr=0.010000
Epoch[013] Batch [0349]/[3760]	Speed: 35.326363 samples/sec	 accuracy=55.651786	 loss=1.864458	 lr=0.010000
Epoch[013] Batch [0399]/[3760]	Speed: 34.716127 samples/sec	 accuracy=55.589844	 loss=1.867661	 lr=0.010000
Epoch[013] Batch [0449]/[3760]	Speed: 35.247623 samples/sec	 accuracy=55.524306	 loss=1.869280	 lr=0.010000
Epoch[013] Batch [0499]/[3760]	Speed: 34.572532 samples/sec	 accuracy=55.384375	 loss=1.870357	 lr=0.010000
Epoch[013] Batch [0549]/[3760]	Speed: 35.130880 samples/sec	 accuracy=55.139205	 loss=1.882842	 lr=0.010000
Epoch[013] Batch [0599]/[3760]	Speed: 34.921905 samples/sec	 accuracy=55.205729	 loss=1.881659	 lr=0.010000
Epoch[013] Batch [0649]/[3760]	Speed: 35.187109 samples/sec	 accuracy=55.295673	 loss=1.880997	 lr=0.010000
Epoch[013] Batch [0699]/[3760]	Speed: 35.197531 samples/sec	 accuracy=55.183036	 loss=1.883676	 lr=0.010000
Epoch[013] Batch [0749]/[3760]	Speed: 35.220803 samples/sec	 accuracy=55.091667	 loss=1.885718	 lr=0.010000
Epoch[013] Batch [0799]/[3760]	Speed: 35.571805 samples/sec	 accuracy=55.119141	 loss=1.885181	 lr=0.010000
Epoch[013] Batch [0849]/[3760]	Speed: 35.115638 samples/sec	 accuracy=55.154412	 loss=1.883034	 lr=0.010000
Epoch[013] Batch [0899]/[3760]	Speed: 35.497110 samples/sec	 accuracy=55.131944	 loss=1.884802	 lr=0.010000
Epoch[013] Batch [0949]/[3760]	Speed: 34.805040 samples/sec	 accuracy=55.059211	 loss=1.887319	 lr=0.010000
Epoch[013] Batch [0999]/[3760]	Speed: 35.382900 samples/sec	 accuracy=55.064062	 loss=1.886992	 lr=0.010000
Epoch[013] Batch [1049]/[3760]	Speed: 35.085741 samples/sec	 accuracy=55.075893	 loss=1.887138	 lr=0.010000
Epoch[013] Batch [1099]/[3760]	Speed: 35.317164 samples/sec	 accuracy=55.066761	 loss=1.887889	 lr=0.010000
Epoch[013] Batch [1149]/[3760]	Speed: 35.400200 samples/sec	 accuracy=55.050272	 loss=1.886532	 lr=0.010000
Epoch[013] Batch [1199]/[3760]	Speed: 35.199758 samples/sec	 accuracy=55.024740	 loss=1.886994	 lr=0.010000
Epoch[013] Batch [1249]/[3760]	Speed: 35.389630 samples/sec	 accuracy=55.037500	 loss=1.885021	 lr=0.010000
Epoch[013] Batch [1299]/[3760]	Speed: 35.015354 samples/sec	 accuracy=55.025240	 loss=1.885528	 lr=0.010000
Epoch[013] Batch [1349]/[3760]	Speed: 35.627287 samples/sec	 accuracy=54.984954	 loss=1.886546	 lr=0.010000
Epoch[013] Batch [1399]/[3760]	Speed: 35.291741 samples/sec	 accuracy=54.963170	 loss=1.889595	 lr=0.010000
Epoch[013] Batch [1449]/[3760]	Speed: 35.378702 samples/sec	 accuracy=54.955819	 loss=1.889865	 lr=0.010000
Epoch[013] Batch [1499]/[3760]	Speed: 35.199345 samples/sec	 accuracy=54.892708	 loss=1.892053	 lr=0.010000
Epoch[013] Batch [1549]/[3760]	Speed: 35.537901 samples/sec	 accuracy=54.915323	 loss=1.890410	 lr=0.010000
Epoch[013] Batch [1599]/[3760]	Speed: 35.287105 samples/sec	 accuracy=54.921875	 loss=1.890440	 lr=0.010000
Epoch[013] Batch [1649]/[3760]	Speed: 35.377189 samples/sec	 accuracy=54.918561	 loss=1.891397	 lr=0.010000
Epoch[013] Batch [1699]/[3760]	Speed: 35.250088 samples/sec	 accuracy=54.924632	 loss=1.892043	 lr=0.010000
Epoch[013] Batch [1749]/[3760]	Speed: 35.422173 samples/sec	 accuracy=54.951786	 loss=1.891371	 lr=0.010000
Epoch[013] Batch [1799]/[3760]	Speed: 35.302638 samples/sec	 accuracy=54.921875	 loss=1.893128	 lr=0.010000
Epoch[013] Batch [1849]/[3760]	Speed: 35.503100 samples/sec	 accuracy=54.911318	 loss=1.893498	 lr=0.010000
Epoch[013] Batch [1899]/[3760]	Speed: 35.183362 samples/sec	 accuracy=54.895559	 loss=1.893890	 lr=0.010000
Epoch[013] Batch [1949]/[3760]	Speed: 35.305297 samples/sec	 accuracy=54.887019	 loss=1.894186	 lr=0.010000
Epoch[013] Batch [1999]/[3760]	Speed: 34.959466 samples/sec	 accuracy=54.882812	 loss=1.896043	 lr=0.010000
Epoch[013] Batch [2049]/[3760]	Speed: 35.383114 samples/sec	 accuracy=54.886433	 loss=1.896118	 lr=0.010000
Epoch[013] Batch [2099]/[3760]	Speed: 35.173596 samples/sec	 accuracy=54.872024	 loss=1.897424	 lr=0.010000
Epoch[013] Batch [2149]/[3760]	Speed: 35.098105 samples/sec	 accuracy=54.848837	 loss=1.898357	 lr=0.010000
Epoch[013] Batch [2199]/[3760]	Speed: 35.423748 samples/sec	 accuracy=54.830966	 loss=1.898312	 lr=0.010000
Epoch[013] Batch [2249]/[3760]	Speed: 35.199282 samples/sec	 accuracy=54.810417	 loss=1.899398	 lr=0.010000
Epoch[013] Batch [2299]/[3760]	Speed: 35.179159 samples/sec	 accuracy=54.827446	 loss=1.898814	 lr=0.010000
Epoch[013] Batch [2349]/[3760]	Speed: 35.255488 samples/sec	 accuracy=54.838431	 loss=1.899000	 lr=0.010000
Epoch[013] Batch [2399]/[3760]	Speed: 35.397507 samples/sec	 accuracy=54.845703	 loss=1.898498	 lr=0.010000
Epoch[013] Batch [2449]/[3760]	Speed: 35.379760 samples/sec	 accuracy=54.870536	 loss=1.898383	 lr=0.010000
Epoch[013] Batch [2499]/[3760]	Speed: 35.471449 samples/sec	 accuracy=54.843750	 loss=1.898913	 lr=0.010000
Epoch[013] Batch [2549]/[3760]	Speed: 35.084682 samples/sec	 accuracy=54.852328	 loss=1.899692	 lr=0.010000
Epoch[013] Batch [2599]/[3760]	Speed: 35.352681 samples/sec	 accuracy=54.831731	 loss=1.901052	 lr=0.010000
Epoch[013] Batch [2649]/[3760]	Speed: 34.993295 samples/sec	 accuracy=54.834316	 loss=1.901929	 lr=0.010000
Epoch[013] Batch [2699]/[3760]	Speed: 35.198158 samples/sec	 accuracy=54.844329	 loss=1.901736	 lr=0.010000
Epoch[013] Batch [2749]/[3760]	Speed: 34.807109 samples/sec	 accuracy=54.816477	 loss=1.902765	 lr=0.010000
Epoch[013] Batch [2799]/[3760]	Speed: 34.954844 samples/sec	 accuracy=54.811384	 loss=1.903698	 lr=0.010000
Epoch[013] Batch [2849]/[3760]	Speed: 34.884418 samples/sec	 accuracy=54.810855	 loss=1.904936	 lr=0.010000
Epoch[013] Batch [2899]/[3760]	Speed: 34.688966 samples/sec	 accuracy=54.795259	 loss=1.905459	 lr=0.010000
Epoch[013] Batch [2949]/[3760]	Speed: 35.325206 samples/sec	 accuracy=54.818326	 loss=1.904812	 lr=0.010000
Epoch[013] Batch [2999]/[3760]	Speed: 34.637314 samples/sec	 accuracy=54.813542	 loss=1.905539	 lr=0.010000
Epoch[013] Batch [3049]/[3760]	Speed: 35.222886 samples/sec	 accuracy=54.821209	 loss=1.905110	 lr=0.010000
Epoch[013] Batch [3099]/[3760]	Speed: 34.489335 samples/sec	 accuracy=54.832157	 loss=1.904361	 lr=0.010000
Epoch[013] Batch [3149]/[3760]	Speed: 35.481844 samples/sec	 accuracy=54.816468	 loss=1.905059	 lr=0.010000
Epoch[013] Batch [3199]/[3760]	Speed: 34.887045 samples/sec	 accuracy=54.817871	 loss=1.905276	 lr=0.010000
Epoch[013] Batch [3249]/[3760]	Speed: 35.254174 samples/sec	 accuracy=54.802404	 loss=1.906019	 lr=0.010000
Epoch[013] Batch [3299]/[3760]	Speed: 35.424807 samples/sec	 accuracy=54.790246	 loss=1.906169	 lr=0.010000
Epoch[013] Batch [3349]/[3760]	Speed: 35.168282 samples/sec	 accuracy=54.799907	 loss=1.906052	 lr=0.010000
Epoch[013] Batch [3399]/[3760]	Speed: 35.479054 samples/sec	 accuracy=54.797794	 loss=1.905893	 lr=0.010000
Epoch[013] Batch [3449]/[3760]	Speed: 35.292182 samples/sec	 accuracy=54.798007	 loss=1.906057	 lr=0.010000
Epoch[013] Batch [3499]/[3760]	Speed: 35.469427 samples/sec	 accuracy=54.788393	 loss=1.906285	 lr=0.010000
Epoch[013] Batch [3549]/[3760]	Speed: 34.982037 samples/sec	 accuracy=54.793134	 loss=1.905813	 lr=0.010000
Epoch[013] Batch [3599]/[3760]	Speed: 35.289247 samples/sec	 accuracy=54.792101	 loss=1.906470	 lr=0.010000
Epoch[013] Batch [3649]/[3760]	Speed: 34.906013 samples/sec	 accuracy=54.793664	 loss=1.906535	 lr=0.010000
Epoch[013] Batch [3699]/[3760]	Speed: 35.299803 samples/sec	 accuracy=54.805743	 loss=1.906202	 lr=0.010000
Epoch[013] Batch [3749]/[3760]	Speed: 36.303750 samples/sec	 accuracy=54.814583	 loss=1.906057	 lr=0.010000
Batch [0049]/[0303]: acc-top1=55.375000 acc-top5=79.125000
Batch [0099]/[0303]: acc-top1=55.187500 acc-top5=79.140625
Batch [0149]/[0303]: acc-top1=55.437500 acc-top5=79.250000
Batch [0199]/[0303]: acc-top1=54.960937 acc-top5=79.023438
Batch [0249]/[0303]: acc-top1=55.012500 acc-top5=78.912500
Batch [0299]/[0303]: acc-top1=54.848958 acc-top5=78.776042
[Epoch 013] training: accuracy=54.822141	 loss=1.905765
[Epoch 013] speed: 34 samples/sec	time cost: 7354.455866
[Epoch 013] validation: acc-top1=54.821576 acc-top5=78.759282 loss=2.134925
Epoch[014] Batch [0049]/[3759]	Speed: 24.088031 samples/sec	 accuracy=55.531250	 loss=1.818341	 lr=0.010000
Epoch[014] Batch [0099]/[3759]	Speed: 35.178340 samples/sec	 accuracy=55.703125	 loss=1.825418	 lr=0.010000
Epoch[014] Batch [0149]/[3759]	Speed: 35.164777 samples/sec	 accuracy=56.125000	 loss=1.831689	 lr=0.010000
Epoch[014] Batch [0199]/[3759]	Speed: 35.056640 samples/sec	 accuracy=56.492188	 loss=1.820151	 lr=0.010000
Epoch[014] Batch [0249]/[3759]	Speed: 35.528209 samples/sec	 accuracy=56.581250	 loss=1.814288	 lr=0.010000
Epoch[014] Batch [0299]/[3759]	Speed: 35.116990 samples/sec	 accuracy=56.447917	 loss=1.816963	 lr=0.010000
Epoch[014] Batch [0349]/[3759]	Speed: 35.382097 samples/sec	 accuracy=56.428571	 loss=1.825560	 lr=0.010000
Epoch[014] Batch [0399]/[3759]	Speed: 35.237334 samples/sec	 accuracy=56.425781	 loss=1.822027	 lr=0.010000
Epoch[014] Batch [0449]/[3759]	Speed: 35.358125 samples/sec	 accuracy=56.284722	 loss=1.824143	 lr=0.010000
Epoch[014] Batch [0499]/[3759]	Speed: 35.065061 samples/sec	 accuracy=56.281250	 loss=1.827082	 lr=0.010000
Epoch[014] Batch [0549]/[3759]	Speed: 35.359430 samples/sec	 accuracy=56.147727	 loss=1.831728	 lr=0.010000
Epoch[014] Batch [0599]/[3759]	Speed: 35.129894 samples/sec	 accuracy=56.203125	 loss=1.830755	 lr=0.010000
Epoch[014] Batch [0649]/[3759]	Speed: 35.113589 samples/sec	 accuracy=56.132212	 loss=1.833235	 lr=0.010000
Epoch[014] Batch [0699]/[3759]	Speed: 35.344600 samples/sec	 accuracy=56.111607	 loss=1.833776	 lr=0.010000
Epoch[014] Batch [0749]/[3759]	Speed: 35.150685 samples/sec	 accuracy=56.104167	 loss=1.835807	 lr=0.010000
Epoch[014] Batch [0799]/[3759]	Speed: 35.549797 samples/sec	 accuracy=56.162109	 loss=1.832925	 lr=0.010000
Epoch[014] Batch [0849]/[3759]	Speed: 35.163420 samples/sec	 accuracy=56.143382	 loss=1.832969	 lr=0.010000
Epoch[014] Batch [0899]/[3759]	Speed: 35.304596 samples/sec	 accuracy=56.154514	 loss=1.835430	 lr=0.010000
Epoch[014] Batch [0949]/[3759]	Speed: 35.142197 samples/sec	 accuracy=56.167763	 loss=1.838085	 lr=0.010000
Epoch[014] Batch [0999]/[3759]	Speed: 35.501578 samples/sec	 accuracy=56.090625	 loss=1.840197	 lr=0.010000
Epoch[014] Batch [1049]/[3759]	Speed: 35.030712 samples/sec	 accuracy=56.062500	 loss=1.841755	 lr=0.010000
Epoch[014] Batch [1099]/[3759]	Speed: 35.443032 samples/sec	 accuracy=55.984375	 loss=1.845546	 lr=0.010000
Epoch[014] Batch [1149]/[3759]	Speed: 34.971443 samples/sec	 accuracy=55.970109	 loss=1.845792	 lr=0.010000
Epoch[014] Batch [1199]/[3759]	Speed: 35.369991 samples/sec	 accuracy=55.957031	 loss=1.848276	 lr=0.010000
Epoch[014] Batch [1249]/[3759]	Speed: 34.785583 samples/sec	 accuracy=55.875000	 loss=1.851555	 lr=0.010000
Epoch[014] Batch [1299]/[3759]	Speed: 35.412150 samples/sec	 accuracy=55.847356	 loss=1.852097	 lr=0.010000
Epoch[014] Batch [1349]/[3759]	Speed: 35.177218 samples/sec	 accuracy=55.787037	 loss=1.854901	 lr=0.010000
Epoch[014] Batch [1399]/[3759]	Speed: 35.320664 samples/sec	 accuracy=55.754464	 loss=1.855988	 lr=0.010000
Epoch[014] Batch [1449]/[3759]	Speed: 35.080768 samples/sec	 accuracy=55.769397	 loss=1.856466	 lr=0.010000
Epoch[014] Batch [1499]/[3759]	Speed: 35.315452 samples/sec	 accuracy=55.755208	 loss=1.856448	 lr=0.010000
Epoch[014] Batch [1549]/[3759]	Speed: 34.861286 samples/sec	 accuracy=55.708669	 loss=1.858590	 lr=0.010000
Epoch[014] Batch [1599]/[3759]	Speed: 35.156930 samples/sec	 accuracy=55.749023	 loss=1.856146	 lr=0.010000
Epoch[014] Batch [1649]/[3759]	Speed: 35.192358 samples/sec	 accuracy=55.746212	 loss=1.855943	 lr=0.010000
Epoch[014] Batch [1699]/[3759]	Speed: 35.330210 samples/sec	 accuracy=55.765625	 loss=1.854853	 lr=0.010000
Epoch[014] Batch [1749]/[3759]	Speed: 35.214522 samples/sec	 accuracy=55.744643	 loss=1.856740	 lr=0.010000
Epoch[014] Batch [1799]/[3759]	Speed: 35.017063 samples/sec	 accuracy=55.717882	 loss=1.858324	 lr=0.010000
Epoch[014] Batch [1849]/[3759]	Speed: 35.330950 samples/sec	 accuracy=55.739020	 loss=1.858273	 lr=0.010000
Epoch[014] Batch [1899]/[3759]	Speed: 34.953802 samples/sec	 accuracy=55.741776	 loss=1.858054	 lr=0.010000
Epoch[014] Batch [1949]/[3759]	Speed: 35.648174 samples/sec	 accuracy=55.697115	 loss=1.859901	 lr=0.010000
Epoch[014] Batch [1999]/[3759]	Speed: 35.020581 samples/sec	 accuracy=55.732031	 loss=1.858788	 lr=0.010000
Epoch[014] Batch [2049]/[3759]	Speed: 35.399493 samples/sec	 accuracy=55.695122	 loss=1.860579	 lr=0.010000
Epoch[014] Batch [2099]/[3759]	Speed: 35.349559 samples/sec	 accuracy=55.706101	 loss=1.860049	 lr=0.010000
Epoch[014] Batch [2149]/[3759]	Speed: 35.486942 samples/sec	 accuracy=55.694041	 loss=1.860858	 lr=0.010000
Epoch[014] Batch [2199]/[3759]	Speed: 34.974144 samples/sec	 accuracy=55.697443	 loss=1.861388	 lr=0.010000
Epoch[014] Batch [2249]/[3759]	Speed: 35.333061 samples/sec	 accuracy=55.718056	 loss=1.860995	 lr=0.010000
Epoch[014] Batch [2299]/[3759]	Speed: 34.789384 samples/sec	 accuracy=55.686821	 loss=1.862761	 lr=0.010000
Epoch[014] Batch [2349]/[3759]	Speed: 34.973336 samples/sec	 accuracy=55.712766	 loss=1.862254	 lr=0.010000
Epoch[014] Batch [2399]/[3759]	Speed: 34.978385 samples/sec	 accuracy=55.707682	 loss=1.862815	 lr=0.010000
Epoch[014] Batch [2449]/[3759]	Speed: 34.931654 samples/sec	 accuracy=55.682398	 loss=1.863796	 lr=0.010000
Epoch[014] Batch [2499]/[3759]	Speed: 35.243314 samples/sec	 accuracy=55.683750	 loss=1.864272	 lr=0.010000
Epoch[014] Batch [2549]/[3759]	Speed: 34.817455 samples/sec	 accuracy=55.690564	 loss=1.863999	 lr=0.010000
Epoch[014] Batch [2599]/[3759]	Speed: 35.553026 samples/sec	 accuracy=55.657452	 loss=1.865008	 lr=0.010000
Epoch[014] Batch [2649]/[3759]	Speed: 35.085202 samples/sec	 accuracy=55.682193	 loss=1.864676	 lr=0.010000
Epoch[014] Batch [2699]/[3759]	Speed: 35.485296 samples/sec	 accuracy=55.688079	 loss=1.863875	 lr=0.010000
Epoch[014] Batch [2749]/[3759]	Speed: 35.329171 samples/sec	 accuracy=55.695455	 loss=1.864177	 lr=0.010000
Epoch[014] Batch [2799]/[3759]	Speed: 35.185806 samples/sec	 accuracy=55.662388	 loss=1.865606	 lr=0.010000
Epoch[014] Batch [2849]/[3759]	Speed: 35.427334 samples/sec	 accuracy=55.648026	 loss=1.866634	 lr=0.010000
Epoch[014] Batch [2899]/[3759]	Speed: 35.009247 samples/sec	 accuracy=55.595905	 loss=1.868904	 lr=0.010000
Epoch[014] Batch [2949]/[3759]	Speed: 35.484487 samples/sec	 accuracy=55.590042	 loss=1.869105	 lr=0.010000
Epoch[014] Batch [2999]/[3759]	Speed: 34.821805 samples/sec	 accuracy=55.588021	 loss=1.868997	 lr=0.010000
Epoch[014] Batch [3049]/[3759]	Speed: 35.429142 samples/sec	 accuracy=55.584016	 loss=1.869760	 lr=0.010000
Epoch[014] Batch [3099]/[3759]	Speed: 34.748522 samples/sec	 accuracy=55.566028	 loss=1.869801	 lr=0.010000
Epoch[014] Batch [3149]/[3759]	Speed: 35.394082 samples/sec	 accuracy=55.573413	 loss=1.869847	 lr=0.010000
Epoch[014] Batch [3199]/[3759]	Speed: 34.717252 samples/sec	 accuracy=55.539551	 loss=1.870990	 lr=0.010000
Epoch[014] Batch [3249]/[3759]	Speed: 35.437694 samples/sec	 accuracy=55.554327	 loss=1.870842	 lr=0.010000
Epoch[014] Batch [3299]/[3759]	Speed: 34.635581 samples/sec	 accuracy=55.534091	 loss=1.871661	 lr=0.010000
Epoch[014] Batch [3349]/[3759]	Speed: 35.045565 samples/sec	 accuracy=55.526586	 loss=1.871863	 lr=0.010000
Epoch[014] Batch [3399]/[3759]	Speed: 34.704062 samples/sec	 accuracy=55.545037	 loss=1.870820	 lr=0.010000
Epoch[014] Batch [3449]/[3759]	Speed: 34.749924 samples/sec	 accuracy=55.516304	 loss=1.871972	 lr=0.010000
Epoch[014] Batch [3499]/[3759]	Speed: 34.983739 samples/sec	 accuracy=55.504911	 loss=1.872469	 lr=0.010000
Epoch[014] Batch [3549]/[3759]	Speed: 34.666448 samples/sec	 accuracy=55.488556	 loss=1.873426	 lr=0.010000
Epoch[014] Batch [3599]/[3759]	Speed: 35.251659 samples/sec	 accuracy=55.498698	 loss=1.873543	 lr=0.010000
Epoch[014] Batch [3649]/[3759]	Speed: 34.273998 samples/sec	 accuracy=55.484161	 loss=1.874358	 lr=0.010000
Epoch[014] Batch [3699]/[3759]	Speed: 35.200159 samples/sec	 accuracy=55.478885	 loss=1.874788	 lr=0.010000
Epoch[014] Batch [3749]/[3759]	Speed: 36.036330 samples/sec	 accuracy=55.489167	 loss=1.874523	 lr=0.010000
Batch [0049]/[0303]: acc-top1=54.687500 acc-top5=79.125000
Batch [0099]/[0303]: acc-top1=55.015625 acc-top5=79.218750
Batch [0149]/[0303]: acc-top1=55.458333 acc-top5=79.468750
Batch [0199]/[0303]: acc-top1=55.140625 acc-top5=79.507812
Batch [0249]/[0303]: acc-top1=54.943750 acc-top5=79.231250
Batch [0299]/[0303]: acc-top1=55.104167 acc-top5=79.281250
[Epoch 014] training: accuracy=55.485585	 loss=1.874563
[Epoch 014] speed: 34 samples/sec	time cost: 7362.188261
[Epoch 014] validation: acc-top1=55.084571 acc-top5=79.285272 loss=2.151888
Epoch[015] Batch [0049]/[3760]	Speed: 24.201419 samples/sec	 accuracy=56.312500	 loss=1.819449	 lr=0.010000
Epoch[015] Batch [0099]/[3760]	Speed: 34.955571 samples/sec	 accuracy=57.750000	 loss=1.761497	 lr=0.010000
Epoch[015] Batch [0149]/[3760]	Speed: 34.963944 samples/sec	 accuracy=57.500000	 loss=1.782132	 lr=0.010000
Epoch[015] Batch [0199]/[3760]	Speed: 34.951979 samples/sec	 accuracy=57.585937	 loss=1.783198	 lr=0.010000
Epoch[015] Batch [0249]/[3760]	Speed: 35.209777 samples/sec	 accuracy=57.700000	 loss=1.782677	 lr=0.010000
Epoch[015] Batch [0299]/[3760]	Speed: 35.052528 samples/sec	 accuracy=57.312500	 loss=1.794211	 lr=0.010000
Epoch[015] Batch [0349]/[3760]	Speed: 35.422087 samples/sec	 accuracy=57.218750	 loss=1.800746	 lr=0.010000
Epoch[015] Batch [0399]/[3760]	Speed: 34.967634 samples/sec	 accuracy=57.050781	 loss=1.806477	 lr=0.010000
Epoch[015] Batch [0449]/[3760]	Speed: 35.324030 samples/sec	 accuracy=56.902778	 loss=1.811370	 lr=0.010000
Epoch[015] Batch [0499]/[3760]	Speed: 34.947083 samples/sec	 accuracy=56.928125	 loss=1.807348	 lr=0.010000
Epoch[015] Batch [0549]/[3760]	Speed: 35.242884 samples/sec	 accuracy=56.894886	 loss=1.811222	 lr=0.010000
Epoch[015] Batch [0599]/[3760]	Speed: 35.006975 samples/sec	 accuracy=56.828125	 loss=1.812703	 lr=0.010000
Epoch[015] Batch [0649]/[3760]	Speed: 35.458392 samples/sec	 accuracy=56.769231	 loss=1.812161	 lr=0.010000
Epoch[015] Batch [0699]/[3760]	Speed: 34.927028 samples/sec	 accuracy=56.859375	 loss=1.814335	 lr=0.010000
Epoch[015] Batch [0749]/[3760]	Speed: 35.061646 samples/sec	 accuracy=56.954167	 loss=1.811293	 lr=0.010000
Epoch[015] Batch [0799]/[3760]	Speed: 35.121704 samples/sec	 accuracy=56.882812	 loss=1.814330	 lr=0.010000
Epoch[015] Batch [0849]/[3760]	Speed: 34.865576 samples/sec	 accuracy=56.838235	 loss=1.815891	 lr=0.010000
Epoch[015] Batch [0899]/[3760]	Speed: 35.329619 samples/sec	 accuracy=56.748264	 loss=1.818994	 lr=0.010000
Epoch[015] Batch [0949]/[3760]	Speed: 34.873634 samples/sec	 accuracy=56.787829	 loss=1.817391	 lr=0.010000
Epoch[015] Batch [0999]/[3760]	Speed: 35.413872 samples/sec	 accuracy=56.739063	 loss=1.820091	 lr=0.010000
Epoch[015] Batch [1049]/[3760]	Speed: 34.745191 samples/sec	 accuracy=56.752976	 loss=1.821586	 lr=0.010000
Epoch[015] Batch [1099]/[3760]	Speed: 35.314536 samples/sec	 accuracy=56.765625	 loss=1.822469	 lr=0.010000
Epoch[015] Batch [1149]/[3760]	Speed: 34.893430 samples/sec	 accuracy=56.724185	 loss=1.822413	 lr=0.010000
Epoch[015] Batch [1199]/[3760]	Speed: 35.300534 samples/sec	 accuracy=56.673177	 loss=1.825256	 lr=0.010000
Epoch[015] Batch [1249]/[3760]	Speed: 34.922176 samples/sec	 accuracy=56.630000	 loss=1.828132	 lr=0.010000
Epoch[015] Batch [1299]/[3760]	Speed: 35.504101 samples/sec	 accuracy=56.568510	 loss=1.830426	 lr=0.010000
Epoch[015] Batch [1349]/[3760]	Speed: 35.126436 samples/sec	 accuracy=56.590278	 loss=1.828930	 lr=0.010000
Epoch[015] Batch [1399]/[3760]	Speed: 35.150090 samples/sec	 accuracy=56.599330	 loss=1.827295	 lr=0.010000
Epoch[015] Batch [1449]/[3760]	Speed: 34.984732 samples/sec	 accuracy=56.570043	 loss=1.825979	 lr=0.010000
Epoch[015] Batch [1499]/[3760]	Speed: 35.118002 samples/sec	 accuracy=56.531250	 loss=1.827223	 lr=0.010000
Epoch[015] Batch [1549]/[3760]	Speed: 35.268686 samples/sec	 accuracy=56.493952	 loss=1.827389	 lr=0.010000
Epoch[015] Batch [1599]/[3760]	Speed: 35.368382 samples/sec	 accuracy=56.472656	 loss=1.828534	 lr=0.010000
Epoch[015] Batch [1649]/[3760]	Speed: 35.236341 samples/sec	 accuracy=56.474432	 loss=1.827820	 lr=0.010000
Epoch[015] Batch [1699]/[3760]	Speed: 35.256041 samples/sec	 accuracy=56.476103	 loss=1.829289	 lr=0.010000
Epoch[015] Batch [1749]/[3760]	Speed: 35.311246 samples/sec	 accuracy=56.472321	 loss=1.830802	 lr=0.010000
Epoch[015] Batch [1799]/[3760]	Speed: 35.301977 samples/sec	 accuracy=56.512153	 loss=1.829904	 lr=0.010000
Epoch[015] Batch [1849]/[3760]	Speed: 35.364234 samples/sec	 accuracy=56.497466	 loss=1.830782	 lr=0.010000
Epoch[015] Batch [1899]/[3760]	Speed: 35.344371 samples/sec	 accuracy=56.468750	 loss=1.831288	 lr=0.010000
Epoch[015] Batch [1949]/[3760]	Speed: 35.320319 samples/sec	 accuracy=56.458333	 loss=1.831399	 lr=0.010000
Epoch[015] Batch [1999]/[3760]	Speed: 35.210719 samples/sec	 accuracy=56.475000	 loss=1.830253	 lr=0.010000
Epoch[015] Batch [2049]/[3760]	Speed: 35.400722 samples/sec	 accuracy=56.490091	 loss=1.830604	 lr=0.010000
Epoch[015] Batch [2099]/[3760]	Speed: 35.181307 samples/sec	 accuracy=56.479911	 loss=1.830532	 lr=0.010000
Epoch[015] Batch [2149]/[3760]	Speed: 35.194550 samples/sec	 accuracy=56.442587	 loss=1.833098	 lr=0.010000
Epoch[015] Batch [2199]/[3760]	Speed: 34.935477 samples/sec	 accuracy=56.433949	 loss=1.833948	 lr=0.010000
Epoch[015] Batch [2249]/[3760]	Speed: 35.149499 samples/sec	 accuracy=56.452778	 loss=1.832973	 lr=0.010000
Epoch[015] Batch [2299]/[3760]	Speed: 35.422345 samples/sec	 accuracy=56.451087	 loss=1.831694	 lr=0.010000
Epoch[015] Batch [2349]/[3760]	Speed: 35.013057 samples/sec	 accuracy=56.421543	 loss=1.832035	 lr=0.010000
Epoch[015] Batch [2399]/[3760]	Speed: 35.108083 samples/sec	 accuracy=56.415365	 loss=1.832365	 lr=0.010000
Epoch[015] Batch [2449]/[3760]	Speed: 34.966823 samples/sec	 accuracy=56.395408	 loss=1.832645	 lr=0.010000
Epoch[015] Batch [2499]/[3760]	Speed: 35.287291 samples/sec	 accuracy=56.387500	 loss=1.833475	 lr=0.010000
Epoch[015] Batch [2549]/[3760]	Speed: 34.719059 samples/sec	 accuracy=56.395221	 loss=1.833406	 lr=0.010000
Epoch[015] Batch [2599]/[3760]	Speed: 35.222100 samples/sec	 accuracy=56.379808	 loss=1.834562	 lr=0.010000
Epoch[015] Batch [2649]/[3760]	Speed: 34.649710 samples/sec	 accuracy=56.361439	 loss=1.835151	 lr=0.010000
Epoch[015] Batch [2699]/[3760]	Speed: 35.380301 samples/sec	 accuracy=56.326968	 loss=1.836702	 lr=0.010000
Epoch[015] Batch [2749]/[3760]	Speed: 34.860982 samples/sec	 accuracy=56.294318	 loss=1.837981	 lr=0.010000
Epoch[015] Batch [2799]/[3760]	Speed: 35.152118 samples/sec	 accuracy=56.267299	 loss=1.838504	 lr=0.010000
Epoch[015] Batch [2849]/[3760]	Speed: 34.834331 samples/sec	 accuracy=56.264254	 loss=1.838741	 lr=0.010000
Epoch[015] Batch [2899]/[3760]	Speed: 34.956605 samples/sec	 accuracy=56.237608	 loss=1.840324	 lr=0.010000
Epoch[015] Batch [2949]/[3760]	Speed: 35.459900 samples/sec	 accuracy=56.215042	 loss=1.840670	 lr=0.010000
Epoch[015] Batch [2999]/[3760]	Speed: 35.017290 samples/sec	 accuracy=56.176563	 loss=1.841972	 lr=0.010000
Epoch[015] Batch [3049]/[3760]	Speed: 35.365523 samples/sec	 accuracy=56.156762	 loss=1.843110	 lr=0.010000
Epoch[015] Batch [3099]/[3760]	Speed: 35.349469 samples/sec	 accuracy=56.180948	 loss=1.842820	 lr=0.010000
Epoch[015] Batch [3149]/[3760]	Speed: 35.440838 samples/sec	 accuracy=56.161706	 loss=1.843352	 lr=0.010000
Epoch[015] Batch [3199]/[3760]	Speed: 35.165445 samples/sec	 accuracy=56.141113	 loss=1.844248	 lr=0.010000
Epoch[015] Batch [3249]/[3760]	Speed: 35.265139 samples/sec	 accuracy=56.134615	 loss=1.844797	 lr=0.010000
Epoch[015] Batch [3299]/[3760]	Speed: 35.033289 samples/sec	 accuracy=56.154830	 loss=1.844059	 lr=0.010000
Epoch[015] Batch [3349]/[3760]	Speed: 35.362610 samples/sec	 accuracy=56.123601	 loss=1.845155	 lr=0.010000
Epoch[015] Batch [3399]/[3760]	Speed: 35.050067 samples/sec	 accuracy=56.109835	 loss=1.845484	 lr=0.010000
Epoch[015] Batch [3449]/[3760]	Speed: 35.402856 samples/sec	 accuracy=56.109601	 loss=1.845153	 lr=0.010000
Epoch[015] Batch [3499]/[3760]	Speed: 35.149186 samples/sec	 accuracy=56.102232	 loss=1.845408	 lr=0.010000
Epoch[015] Batch [3549]/[3760]	Speed: 35.504993 samples/sec	 accuracy=56.087148	 loss=1.845925	 lr=0.010000
Epoch[015] Batch [3599]/[3760]	Speed: 35.261728 samples/sec	 accuracy=56.082031	 loss=1.845970	 lr=0.010000
Epoch[015] Batch [3649]/[3760]	Speed: 35.385167 samples/sec	 accuracy=56.077911	 loss=1.846163	 lr=0.010000
Epoch[015] Batch [3699]/[3760]	Speed: 35.048441 samples/sec	 accuracy=56.059966	 loss=1.847229	 lr=0.010000
Epoch[015] Batch [3749]/[3760]	Speed: 36.528844 samples/sec	 accuracy=56.062917	 loss=1.847510	 lr=0.010000
Batch [0049]/[0303]: acc-top1=54.187500 acc-top5=79.031250
Batch [0099]/[0303]: acc-top1=55.187500 acc-top5=79.281250
Batch [0149]/[0303]: acc-top1=55.895833 acc-top5=79.760417
Batch [0199]/[0303]: acc-top1=55.984375 acc-top5=79.757812
Batch [0249]/[0303]: acc-top1=55.837500 acc-top5=79.587500
Batch [0299]/[0303]: acc-top1=55.916667 acc-top5=79.609375
[Epoch 015] training: accuracy=56.067570	 loss=1.847210
[Epoch 015] speed: 34 samples/sec	time cost: 7359.501570
[Epoch 015] validation: acc-top1=55.868399 acc-top5=79.599835 loss=2.101443
Epoch[016] Batch [0049]/[3760]	Speed: 25.455628 samples/sec	 accuracy=59.062500	 loss=1.716537	 lr=0.010000
Epoch[016] Batch [0099]/[3760]	Speed: 35.308827 samples/sec	 accuracy=57.984375	 loss=1.757090	 lr=0.010000
Epoch[016] Batch [0149]/[3760]	Speed: 35.054405 samples/sec	 accuracy=58.062500	 loss=1.746534	 lr=0.010000
Epoch[016] Batch [0199]/[3760]	Speed: 34.874509 samples/sec	 accuracy=58.054688	 loss=1.751295	 lr=0.010000
Epoch[016] Batch [0249]/[3760]	Speed: 35.360211 samples/sec	 accuracy=57.812500	 loss=1.759711	 lr=0.010000
Epoch[016] Batch [0299]/[3760]	Speed: 35.043144 samples/sec	 accuracy=57.703125	 loss=1.767920	 lr=0.010000
Epoch[016] Batch [0349]/[3760]	Speed: 35.458674 samples/sec	 accuracy=57.709821	 loss=1.766873	 lr=0.010000
Epoch[016] Batch [0399]/[3760]	Speed: 35.126773 samples/sec	 accuracy=57.781250	 loss=1.767331	 lr=0.010000
Epoch[016] Batch [0449]/[3760]	Speed: 35.458608 samples/sec	 accuracy=57.642361	 loss=1.769429	 lr=0.010000
Epoch[016] Batch [0499]/[3760]	Speed: 35.111077 samples/sec	 accuracy=57.418750	 loss=1.780110	 lr=0.010000
Epoch[016] Batch [0549]/[3760]	Speed: 35.432116 samples/sec	 accuracy=57.389205	 loss=1.783478	 lr=0.010000
Epoch[016] Batch [0599]/[3760]	Speed: 35.126235 samples/sec	 accuracy=57.296875	 loss=1.784834	 lr=0.010000
Epoch[016] Batch [0649]/[3760]	Speed: 35.529027 samples/sec	 accuracy=57.274038	 loss=1.784372	 lr=0.010000
Epoch[016] Batch [0699]/[3760]	Speed: 35.211164 samples/sec	 accuracy=57.263393	 loss=1.784120	 lr=0.010000
Epoch[016] Batch [0749]/[3760]	Speed: 35.284152 samples/sec	 accuracy=57.202083	 loss=1.787632	 lr=0.010000
Epoch[016] Batch [0799]/[3760]	Speed: 35.360465 samples/sec	 accuracy=57.097656	 loss=1.791881	 lr=0.010000
Epoch[016] Batch [0849]/[3760]	Speed: 35.345276 samples/sec	 accuracy=57.062500	 loss=1.793370	 lr=0.010000
Epoch[016] Batch [0899]/[3760]	Speed: 35.409523 samples/sec	 accuracy=57.111111	 loss=1.792538	 lr=0.010000
Epoch[016] Batch [0949]/[3760]	Speed: 35.255301 samples/sec	 accuracy=57.134868	 loss=1.792518	 lr=0.010000
Epoch[016] Batch [0999]/[3760]	Speed: 35.439299 samples/sec	 accuracy=57.135937	 loss=1.793380	 lr=0.010000
Epoch[016] Batch [1049]/[3760]	Speed: 35.310384 samples/sec	 accuracy=57.017857	 loss=1.796848	 lr=0.010000
Epoch[016] Batch [1099]/[3760]	Speed: 35.343731 samples/sec	 accuracy=57.019886	 loss=1.797127	 lr=0.010000
Epoch[016] Batch [1149]/[3760]	Speed: 35.271434 samples/sec	 accuracy=57.012228	 loss=1.797836	 lr=0.010000
Epoch[016] Batch [1199]/[3760]	Speed: 35.159022 samples/sec	 accuracy=56.901042	 loss=1.801326	 lr=0.010000
Epoch[016] Batch [1249]/[3760]	Speed: 35.462461 samples/sec	 accuracy=56.946250	 loss=1.800975	 lr=0.010000
Epoch[016] Batch [1299]/[3760]	Speed: 35.143235 samples/sec	 accuracy=56.913462	 loss=1.801007	 lr=0.010000
Epoch[016] Batch [1349]/[3760]	Speed: 35.448984 samples/sec	 accuracy=56.866898	 loss=1.803921	 lr=0.010000
Epoch[016] Batch [1399]/[3760]	Speed: 35.295647 samples/sec	 accuracy=56.911830	 loss=1.803008	 lr=0.010000
Epoch[016] Batch [1449]/[3760]	Speed: 35.311284 samples/sec	 accuracy=56.904095	 loss=1.802742	 lr=0.010000
Epoch[016] Batch [1499]/[3760]	Speed: 35.396345 samples/sec	 accuracy=56.858333	 loss=1.804790	 lr=0.010000
Epoch[016] Batch [1549]/[3760]	Speed: 35.208666 samples/sec	 accuracy=56.878024	 loss=1.803693	 lr=0.010000
Epoch[016] Batch [1599]/[3760]	Speed: 35.244615 samples/sec	 accuracy=56.854492	 loss=1.803693	 lr=0.010000
Epoch[016] Batch [1649]/[3760]	Speed: 35.218509 samples/sec	 accuracy=56.872159	 loss=1.802726	 lr=0.010000
Epoch[016] Batch [1699]/[3760]	Speed: 35.257172 samples/sec	 accuracy=56.863971	 loss=1.803466	 lr=0.010000
Epoch[016] Batch [1749]/[3760]	Speed: 35.430524 samples/sec	 accuracy=56.833036	 loss=1.805534	 lr=0.010000
Epoch[016] Batch [1799]/[3760]	Speed: 35.208860 samples/sec	 accuracy=56.815972	 loss=1.806750	 lr=0.010000
Epoch[016] Batch [1849]/[3760]	Speed: 35.377748 samples/sec	 accuracy=56.788007	 loss=1.807138	 lr=0.010000
Epoch[016] Batch [1899]/[3760]	Speed: 35.354945 samples/sec	 accuracy=56.787007	 loss=1.808319	 lr=0.010000
Epoch[016] Batch [1949]/[3760]	Speed: 35.449581 samples/sec	 accuracy=56.774038	 loss=1.808430	 lr=0.010000
Epoch[016] Batch [1999]/[3760]	Speed: 35.291277 samples/sec	 accuracy=56.743750	 loss=1.809777	 lr=0.010000
Epoch[016] Batch [2049]/[3760]	Speed: 35.284182 samples/sec	 accuracy=56.692835	 loss=1.811927	 lr=0.010000
Epoch[016] Batch [2099]/[3760]	Speed: 35.478155 samples/sec	 accuracy=56.691220	 loss=1.813351	 lr=0.010000
Epoch[016] Batch [2149]/[3760]	Speed: 35.392291 samples/sec	 accuracy=56.683140	 loss=1.813500	 lr=0.010000
Epoch[016] Batch [2199]/[3760]	Speed: 35.389260 samples/sec	 accuracy=56.686790	 loss=1.813116	 lr=0.010000
Epoch[016] Batch [2249]/[3760]	Speed: 35.188709 samples/sec	 accuracy=56.663889	 loss=1.814429	 lr=0.010000
Epoch[016] Batch [2299]/[3760]	Speed: 35.266849 samples/sec	 accuracy=56.663723	 loss=1.813913	 lr=0.010000
Epoch[016] Batch [2349]/[3760]	Speed: 35.038763 samples/sec	 accuracy=56.656915	 loss=1.814289	 lr=0.010000
Epoch[016] Batch [2399]/[3760]	Speed: 35.265384 samples/sec	 accuracy=56.636068	 loss=1.815310	 lr=0.010000
Epoch[016] Batch [2449]/[3760]	Speed: 34.704240 samples/sec	 accuracy=56.609694	 loss=1.816889	 lr=0.010000
Epoch[016] Batch [2499]/[3760]	Speed: 35.337439 samples/sec	 accuracy=56.614375	 loss=1.816587	 lr=0.010000
Epoch[016] Batch [2549]/[3760]	Speed: 34.829289 samples/sec	 accuracy=56.635417	 loss=1.815565	 lr=0.010000
Epoch[016] Batch [2599]/[3760]	Speed: 34.959159 samples/sec	 accuracy=56.641226	 loss=1.815592	 lr=0.010000
Epoch[016] Batch [2649]/[3760]	Speed: 35.089140 samples/sec	 accuracy=56.625590	 loss=1.816598	 lr=0.010000
Epoch[016] Batch [2699]/[3760]	Speed: 34.674035 samples/sec	 accuracy=56.630208	 loss=1.816939	 lr=0.010000
Epoch[016] Batch [2749]/[3760]	Speed: 35.307589 samples/sec	 accuracy=56.621591	 loss=1.817828	 lr=0.010000
Epoch[016] Batch [2799]/[3760]	Speed: 34.429440 samples/sec	 accuracy=56.603237	 loss=1.818813	 lr=0.010000
Epoch[016] Batch [2849]/[3760]	Speed: 35.465570 samples/sec	 accuracy=56.586075	 loss=1.819564	 lr=0.010000
Epoch[016] Batch [2899]/[3760]	Speed: 34.962308 samples/sec	 accuracy=56.583513	 loss=1.819278	 lr=0.010000
Epoch[016] Batch [2949]/[3760]	Speed: 35.415824 samples/sec	 accuracy=56.573093	 loss=1.819831	 lr=0.010000
Epoch[016] Batch [2999]/[3760]	Speed: 35.175098 samples/sec	 accuracy=56.581771	 loss=1.819752	 lr=0.010000
Epoch[016] Batch [3049]/[3760]	Speed: 35.077227 samples/sec	 accuracy=56.574283	 loss=1.819892	 lr=0.010000
Epoch[016] Batch [3099]/[3760]	Speed: 35.260278 samples/sec	 accuracy=56.554940	 loss=1.820626	 lr=0.010000
Epoch[016] Batch [3149]/[3760]	Speed: 35.069068 samples/sec	 accuracy=56.541171	 loss=1.821170	 lr=0.010000
Epoch[016] Batch [3199]/[3760]	Speed: 35.226830 samples/sec	 accuracy=56.537598	 loss=1.821274	 lr=0.010000
Epoch[016] Batch [3249]/[3760]	Speed: 34.910511 samples/sec	 accuracy=56.546635	 loss=1.821180	 lr=0.010000
Epoch[016] Batch [3299]/[3760]	Speed: 35.448812 samples/sec	 accuracy=56.543561	 loss=1.821645	 lr=0.010000
Epoch[016] Batch [3349]/[3760]	Speed: 35.066285 samples/sec	 accuracy=56.538246	 loss=1.822030	 lr=0.010000
Epoch[016] Batch [3399]/[3760]	Speed: 35.321441 samples/sec	 accuracy=56.523437	 loss=1.822234	 lr=0.010000
Epoch[016] Batch [3449]/[3760]	Speed: 34.826028 samples/sec	 accuracy=56.507246	 loss=1.822664	 lr=0.010000
Epoch[016] Batch [3499]/[3760]	Speed: 35.420758 samples/sec	 accuracy=56.502232	 loss=1.822920	 lr=0.010000
Epoch[016] Batch [3549]/[3760]	Speed: 34.928011 samples/sec	 accuracy=56.499560	 loss=1.822914	 lr=0.010000
Epoch[016] Batch [3599]/[3760]	Speed: 35.423902 samples/sec	 accuracy=56.521267	 loss=1.822418	 lr=0.010000
Epoch[016] Batch [3649]/[3760]	Speed: 34.970002 samples/sec	 accuracy=56.520548	 loss=1.822700	 lr=0.010000
Epoch[016] Batch [3699]/[3760]	Speed: 35.220001 samples/sec	 accuracy=56.532517	 loss=1.823100	 lr=0.010000
Epoch[016] Batch [3749]/[3760]	Speed: 36.239869 samples/sec	 accuracy=56.510833	 loss=1.824273	 lr=0.010000
Batch [0049]/[0303]: acc-top1=54.562500 acc-top5=79.750000
Batch [0099]/[0303]: acc-top1=54.859375 acc-top5=79.250000
Batch [0149]/[0303]: acc-top1=55.635417 acc-top5=79.604167
Batch [0199]/[0303]: acc-top1=55.468750 acc-top5=79.375000
Batch [0249]/[0303]: acc-top1=55.350000 acc-top5=79.168750
Batch [0299]/[0303]: acc-top1=55.135417 acc-top5=79.125000
[Epoch 016] training: accuracy=56.515542	 loss=1.824286
[Epoch 016] speed: 35 samples/sec	time cost: 7343.108438
[Epoch 016] validation: acc-top1=55.130982 acc-top5=79.109942 loss=2.156404
Epoch[017] Batch [0049]/[3759]	Speed: 24.148999 samples/sec	 accuracy=58.281250	 loss=1.764278	 lr=0.010000
Epoch[017] Batch [0099]/[3759]	Speed: 35.000133 samples/sec	 accuracy=58.000000	 loss=1.732302	 lr=0.010000
Epoch[017] Batch [0149]/[3759]	Speed: 34.972771 samples/sec	 accuracy=58.187500	 loss=1.734341	 lr=0.010000
Epoch[017] Batch [0199]/[3759]	Speed: 34.969092 samples/sec	 accuracy=57.984375	 loss=1.741649	 lr=0.010000
Epoch[017] Batch [0249]/[3759]	Speed: 35.297115 samples/sec	 accuracy=57.856250	 loss=1.745457	 lr=0.010000
Epoch[017] Batch [0299]/[3759]	Speed: 34.595685 samples/sec	 accuracy=58.000000	 loss=1.747655	 lr=0.010000
Epoch[017] Batch [0349]/[3759]	Speed: 35.390114 samples/sec	 accuracy=58.151786	 loss=1.737851	 lr=0.010000
Epoch[017] Batch [0399]/[3759]	Speed: 34.545198 samples/sec	 accuracy=58.101562	 loss=1.740977	 lr=0.010000
Epoch[017] Batch [0449]/[3759]	Speed: 35.305112 samples/sec	 accuracy=58.006944	 loss=1.747967	 lr=0.010000
Epoch[017] Batch [0499]/[3759]	Speed: 34.740560 samples/sec	 accuracy=58.071875	 loss=1.745152	 lr=0.010000
Epoch[017] Batch [0549]/[3759]	Speed: 35.004614 samples/sec	 accuracy=58.051136	 loss=1.746084	 lr=0.010000
Epoch[017] Batch [0599]/[3759]	Speed: 34.819895 samples/sec	 accuracy=58.075521	 loss=1.746586	 lr=0.010000
Epoch[017] Batch [0649]/[3759]	Speed: 34.887073 samples/sec	 accuracy=58.187500	 loss=1.744175	 lr=0.010000
Epoch[017] Batch [0699]/[3759]	Speed: 35.250866 samples/sec	 accuracy=58.100446	 loss=1.748467	 lr=0.010000
Epoch[017] Batch [0749]/[3759]	Speed: 34.764031 samples/sec	 accuracy=58.079167	 loss=1.750621	 lr=0.010000
Epoch[017] Batch [0799]/[3759]	Speed: 35.375138 samples/sec	 accuracy=58.015625	 loss=1.753568	 lr=0.010000
Epoch[017] Batch [0849]/[3759]	Speed: 34.704652 samples/sec	 accuracy=57.996324	 loss=1.754046	 lr=0.010000
Epoch[017] Batch [0899]/[3759]	Speed: 35.516361 samples/sec	 accuracy=57.939236	 loss=1.755600	 lr=0.010000
Epoch[017] Batch [0949]/[3759]	Speed: 35.057188 samples/sec	 accuracy=57.889803	 loss=1.759496	 lr=0.010000
Epoch[017] Batch [0999]/[3759]	Speed: 35.337861 samples/sec	 accuracy=57.910938	 loss=1.760086	 lr=0.010000
Epoch[017] Batch [1049]/[3759]	Speed: 34.965327 samples/sec	 accuracy=57.889881	 loss=1.761679	 lr=0.010000
Epoch[017] Batch [1099]/[3759]	Speed: 35.328123 samples/sec	 accuracy=57.913352	 loss=1.759891	 lr=0.010000
Epoch[017] Batch [1149]/[3759]	Speed: 35.265375 samples/sec	 accuracy=57.918478	 loss=1.760156	 lr=0.010000
Epoch[017] Batch [1199]/[3759]	Speed: 35.306931 samples/sec	 accuracy=57.950521	 loss=1.757397	 lr=0.010000
Epoch[017] Batch [1249]/[3759]	Speed: 35.208995 samples/sec	 accuracy=57.937500	 loss=1.758579	 lr=0.010000
Epoch[017] Batch [1299]/[3759]	Speed: 35.390648 samples/sec	 accuracy=57.917067	 loss=1.759376	 lr=0.010000
Epoch[017] Batch [1349]/[3759]	Speed: 35.471742 samples/sec	 accuracy=57.855324	 loss=1.761642	 lr=0.010000
Epoch[017] Batch [1399]/[3759]	Speed: 35.372008 samples/sec	 accuracy=57.800223	 loss=1.762502	 lr=0.010000
Epoch[017] Batch [1449]/[3759]	Speed: 35.443852 samples/sec	 accuracy=57.755388	 loss=1.765457	 lr=0.010000
Epoch[017] Batch [1499]/[3759]	Speed: 35.253479 samples/sec	 accuracy=57.728125	 loss=1.766642	 lr=0.010000
Epoch[017] Batch [1549]/[3759]	Speed: 35.300531 samples/sec	 accuracy=57.690524	 loss=1.767694	 lr=0.010000
Epoch[017] Batch [1599]/[3759]	Speed: 35.030414 samples/sec	 accuracy=57.639648	 loss=1.769343	 lr=0.010000
Epoch[017] Batch [1649]/[3759]	Speed: 35.467833 samples/sec	 accuracy=57.579545	 loss=1.772260	 lr=0.010000
Epoch[017] Batch [1699]/[3759]	Speed: 34.954540 samples/sec	 accuracy=57.592831	 loss=1.771726	 lr=0.010000
Epoch[017] Batch [1749]/[3759]	Speed: 35.567607 samples/sec	 accuracy=57.569643	 loss=1.773006	 lr=0.010000
Epoch[017] Batch [1799]/[3759]	Speed: 34.812093 samples/sec	 accuracy=57.579861	 loss=1.772586	 lr=0.010000
Epoch[017] Batch [1849]/[3759]	Speed: 35.543550 samples/sec	 accuracy=57.543919	 loss=1.774279	 lr=0.010000
Epoch[017] Batch [1899]/[3759]	Speed: 34.987229 samples/sec	 accuracy=57.550164	 loss=1.773768	 lr=0.010000
Epoch[017] Batch [1949]/[3759]	Speed: 35.253787 samples/sec	 accuracy=57.522436	 loss=1.775257	 lr=0.010000
Epoch[017] Batch [1999]/[3759]	Speed: 35.121632 samples/sec	 accuracy=57.475781	 loss=1.778096	 lr=0.010000
Epoch[017] Batch [2049]/[3759]	Speed: 35.200730 samples/sec	 accuracy=57.462652	 loss=1.778628	 lr=0.010000
Epoch[017] Batch [2099]/[3759]	Speed: 35.293202 samples/sec	 accuracy=57.430804	 loss=1.779155	 lr=0.010000
Epoch[017] Batch [2149]/[3759]	Speed: 35.272386 samples/sec	 accuracy=57.430959	 loss=1.779249	 lr=0.010000
Epoch[017] Batch [2199]/[3759]	Speed: 35.404643 samples/sec	 accuracy=57.415483	 loss=1.781293	 lr=0.010000
Epoch[017] Batch [2249]/[3759]	Speed: 35.276448 samples/sec	 accuracy=57.379167	 loss=1.782302	 lr=0.010000
Epoch[017] Batch [2299]/[3759]	Speed: 35.288289 samples/sec	 accuracy=57.362772	 loss=1.782807	 lr=0.010000
Epoch[017] Batch [2349]/[3759]	Speed: 35.079460 samples/sec	 accuracy=57.340426	 loss=1.783896	 lr=0.010000
Epoch[017] Batch [2399]/[3759]	Speed: 35.312326 samples/sec	 accuracy=57.346354	 loss=1.783416	 lr=0.010000
Epoch[017] Batch [2449]/[3759]	Speed: 34.963668 samples/sec	 accuracy=57.320791	 loss=1.784428	 lr=0.010000
Epoch[017] Batch [2499]/[3759]	Speed: 35.603228 samples/sec	 accuracy=57.305625	 loss=1.784772	 lr=0.010000
Epoch[017] Batch [2549]/[3759]	Speed: 34.981924 samples/sec	 accuracy=57.294118	 loss=1.785755	 lr=0.010000
Epoch[017] Batch [2599]/[3759]	Speed: 35.436734 samples/sec	 accuracy=57.262019	 loss=1.787507	 lr=0.010000
Epoch[017] Batch [2649]/[3759]	Speed: 35.010576 samples/sec	 accuracy=57.275354	 loss=1.786742	 lr=0.010000
Epoch[017] Batch [2699]/[3759]	Speed: 35.235978 samples/sec	 accuracy=57.301505	 loss=1.785778	 lr=0.010000
Epoch[017] Batch [2749]/[3759]	Speed: 35.050948 samples/sec	 accuracy=57.280682	 loss=1.787255	 lr=0.010000
Epoch[017] Batch [2799]/[3759]	Speed: 35.200997 samples/sec	 accuracy=57.257254	 loss=1.788123	 lr=0.010000
Epoch[017] Batch [2849]/[3759]	Speed: 35.169161 samples/sec	 accuracy=57.241228	 loss=1.788702	 lr=0.010000
Epoch[017] Batch [2899]/[3759]	Speed: 35.049341 samples/sec	 accuracy=57.254310	 loss=1.788552	 lr=0.010000
Epoch[017] Batch [2949]/[3759]	Speed: 35.281282 samples/sec	 accuracy=57.243644	 loss=1.789496	 lr=0.010000
Epoch[017] Batch [2999]/[3759]	Speed: 35.161767 samples/sec	 accuracy=57.246354	 loss=1.789481	 lr=0.010000
Epoch[017] Batch [3049]/[3759]	Speed: 35.328376 samples/sec	 accuracy=57.242316	 loss=1.789707	 lr=0.010000
Epoch[017] Batch [3099]/[3759]	Speed: 35.079779 samples/sec	 accuracy=57.215726	 loss=1.790402	 lr=0.010000
Epoch[017] Batch [3149]/[3759]	Speed: 35.346860 samples/sec	 accuracy=57.226687	 loss=1.790490	 lr=0.010000
Epoch[017] Batch [3199]/[3759]	Speed: 35.077889 samples/sec	 accuracy=57.214844	 loss=1.791500	 lr=0.010000
Epoch[017] Batch [3249]/[3759]	Speed: 35.333124 samples/sec	 accuracy=57.175000	 loss=1.793235	 lr=0.010000
Epoch[017] Batch [3299]/[3759]	Speed: 34.613490 samples/sec	 accuracy=57.159091	 loss=1.793618	 lr=0.010000
Epoch[017] Batch [3349]/[3759]	Speed: 35.564624 samples/sec	 accuracy=57.148321	 loss=1.793896	 lr=0.010000
Epoch[017] Batch [3399]/[3759]	Speed: 35.034705 samples/sec	 accuracy=57.146599	 loss=1.793978	 lr=0.010000
Epoch[017] Batch [3449]/[3759]	Speed: 35.451217 samples/sec	 accuracy=57.123188	 loss=1.795290	 lr=0.010000
Epoch[017] Batch [3499]/[3759]	Speed: 34.840952 samples/sec	 accuracy=57.099554	 loss=1.796320	 lr=0.010000
Epoch[017] Batch [3549]/[3759]	Speed: 35.191361 samples/sec	 accuracy=57.102993	 loss=1.796431	 lr=0.010000
Epoch[017] Batch [3599]/[3759]	Speed: 34.981968 samples/sec	 accuracy=57.130208	 loss=1.795407	 lr=0.010000
Epoch[017] Batch [3649]/[3759]	Speed: 34.604279 samples/sec	 accuracy=57.102312	 loss=1.796603	 lr=0.010000
Epoch[017] Batch [3699]/[3759]	Speed: 35.342101 samples/sec	 accuracy=57.081503	 loss=1.797786	 lr=0.010000
Epoch[017] Batch [3749]/[3759]	Speed: 36.387959 samples/sec	 accuracy=57.077083	 loss=1.798496	 lr=0.010000
Batch [0049]/[0303]: acc-top1=54.718750 acc-top5=80.375000
Batch [0099]/[0303]: acc-top1=55.812500 acc-top5=80.281250
Batch [0149]/[0303]: acc-top1=56.427083 acc-top5=80.291667
Batch [0199]/[0303]: acc-top1=56.281250 acc-top5=80.078125
Batch [0249]/[0303]: acc-top1=56.131250 acc-top5=79.818750
Batch [0299]/[0303]: acc-top1=56.078125 acc-top5=79.802083
[Epoch 017] training: accuracy=57.071778	 loss=1.798540
[Epoch 017] speed: 34 samples/sec	time cost: 7359.888093
[Epoch 017] validation: acc-top1=56.017946 acc-top5=79.764851 loss=2.094938
Epoch[018] Batch [0049]/[3760]	Speed: 24.095846 samples/sec	 accuracy=58.312500	 loss=1.737097	 lr=0.010000
Epoch[018] Batch [0099]/[3760]	Speed: 35.036356 samples/sec	 accuracy=58.812500	 loss=1.731065	 lr=0.010000
Epoch[018] Batch [0149]/[3760]	Speed: 34.900422 samples/sec	 accuracy=58.750000	 loss=1.739391	 lr=0.010000
Epoch[018] Batch [0199]/[3760]	Speed: 34.962476 samples/sec	 accuracy=58.851562	 loss=1.732544	 lr=0.010000
Epoch[018] Batch [0249]/[3760]	Speed: 35.349640 samples/sec	 accuracy=58.968750	 loss=1.720877	 lr=0.010000
Epoch[018] Batch [0299]/[3760]	Speed: 35.067018 samples/sec	 accuracy=58.614583	 loss=1.735846	 lr=0.010000
Epoch[018] Batch [0349]/[3760]	Speed: 35.468425 samples/sec	 accuracy=58.535714	 loss=1.741898	 lr=0.010000
Epoch[018] Batch [0399]/[3760]	Speed: 34.766934 samples/sec	 accuracy=58.609375	 loss=1.736456	 lr=0.010000
Epoch[018] Batch [0449]/[3760]	Speed: 35.308075 samples/sec	 accuracy=58.593750	 loss=1.735714	 lr=0.010000
Epoch[018] Batch [0499]/[3760]	Speed: 34.698005 samples/sec	 accuracy=58.606250	 loss=1.735342	 lr=0.010000
Epoch[018] Batch [0549]/[3760]	Speed: 35.359429 samples/sec	 accuracy=58.599432	 loss=1.737199	 lr=0.010000
Epoch[018] Batch [0599]/[3760]	Speed: 34.555729 samples/sec	 accuracy=58.518229	 loss=1.740624	 lr=0.010000
Epoch[018] Batch [0649]/[3760]	Speed: 35.243874 samples/sec	 accuracy=58.509615	 loss=1.742664	 lr=0.010000
Epoch[018] Batch [0699]/[3760]	Speed: 34.877529 samples/sec	 accuracy=58.372768	 loss=1.745555	 lr=0.010000
Epoch[018] Batch [0749]/[3760]	Speed: 35.063826 samples/sec	 accuracy=58.239583	 loss=1.750688	 lr=0.010000
Epoch[018] Batch [0799]/[3760]	Speed: 35.056592 samples/sec	 accuracy=58.232422	 loss=1.749949	 lr=0.010000
Epoch[018] Batch [0849]/[3760]	Speed: 35.001119 samples/sec	 accuracy=58.279412	 loss=1.748880	 lr=0.010000
Epoch[018] Batch [0899]/[3760]	Speed: 35.344139 samples/sec	 accuracy=58.272569	 loss=1.749480	 lr=0.010000
Epoch[018] Batch [0949]/[3760]	Speed: 35.099322 samples/sec	 accuracy=58.144737	 loss=1.753948	 lr=0.010000
Epoch[018] Batch [0999]/[3760]	Speed: 35.471852 samples/sec	 accuracy=58.200000	 loss=1.750816	 lr=0.010000
Epoch[018] Batch [1049]/[3760]	Speed: 35.026939 samples/sec	 accuracy=58.154762	 loss=1.751847	 lr=0.010000
Epoch[018] Batch [1099]/[3760]	Speed: 35.591044 samples/sec	 accuracy=58.157670	 loss=1.749495	 lr=0.010000
Epoch[018] Batch [1149]/[3760]	Speed: 35.150952 samples/sec	 accuracy=58.078804	 loss=1.752420	 lr=0.010000
Epoch[018] Batch [1199]/[3760]	Speed: 35.476053 samples/sec	 accuracy=58.109375	 loss=1.752378	 lr=0.010000
Epoch[018] Batch [1249]/[3760]	Speed: 35.169553 samples/sec	 accuracy=58.051250	 loss=1.754724	 lr=0.010000
Epoch[018] Batch [1299]/[3760]	Speed: 35.403234 samples/sec	 accuracy=58.036058	 loss=1.754181	 lr=0.010000
Epoch[018] Batch [1349]/[3760]	Speed: 35.299467 samples/sec	 accuracy=57.997685	 loss=1.755780	 lr=0.010000
Epoch[018] Batch [1399]/[3760]	Speed: 35.442581 samples/sec	 accuracy=57.996652	 loss=1.756961	 lr=0.010000
Epoch[018] Batch [1449]/[3760]	Speed: 35.426473 samples/sec	 accuracy=57.947198	 loss=1.759972	 lr=0.010000
Epoch[018] Batch [1499]/[3760]	Speed: 35.218564 samples/sec	 accuracy=57.898958	 loss=1.760595	 lr=0.010000
Epoch[018] Batch [1549]/[3760]	Speed: 35.234003 samples/sec	 accuracy=57.856855	 loss=1.760547	 lr=0.010000
Epoch[018] Batch [1599]/[3760]	Speed: 35.103635 samples/sec	 accuracy=57.828125	 loss=1.761095	 lr=0.010000
Epoch[018] Batch [1649]/[3760]	Speed: 35.322337 samples/sec	 accuracy=57.825758	 loss=1.761200	 lr=0.010000
Epoch[018] Batch [1699]/[3760]	Speed: 35.011538 samples/sec	 accuracy=57.794118	 loss=1.761944	 lr=0.010000
Epoch[018] Batch [1749]/[3760]	Speed: 35.342835 samples/sec	 accuracy=57.825000	 loss=1.760916	 lr=0.010000
Epoch[018] Batch [1799]/[3760]	Speed: 34.719570 samples/sec	 accuracy=57.817708	 loss=1.761466	 lr=0.010000
Epoch[018] Batch [1849]/[3760]	Speed: 35.224899 samples/sec	 accuracy=57.813345	 loss=1.761090	 lr=0.010000
Epoch[018] Batch [1899]/[3760]	Speed: 34.991031 samples/sec	 accuracy=57.784539	 loss=1.763148	 lr=0.010000
Epoch[018] Batch [1949]/[3760]	Speed: 35.021973 samples/sec	 accuracy=57.729968	 loss=1.766384	 lr=0.010000
Epoch[018] Batch [1999]/[3760]	Speed: 35.193259 samples/sec	 accuracy=57.729688	 loss=1.766800	 lr=0.010000
Epoch[018] Batch [2049]/[3760]	Speed: 34.986542 samples/sec	 accuracy=57.727134	 loss=1.766833	 lr=0.010000
Epoch[018] Batch [2099]/[3760]	Speed: 35.274850 samples/sec	 accuracy=57.744792	 loss=1.766305	 lr=0.010000
Epoch[018] Batch [2149]/[3760]	Speed: 34.872061 samples/sec	 accuracy=57.726744	 loss=1.767144	 lr=0.010000
Epoch[018] Batch [2199]/[3760]	Speed: 35.361194 samples/sec	 accuracy=57.740057	 loss=1.765710	 lr=0.010000
Epoch[018] Batch [2249]/[3760]	Speed: 34.793423 samples/sec	 accuracy=57.720833	 loss=1.766392	 lr=0.010000
Epoch[018] Batch [2299]/[3760]	Speed: 35.330400 samples/sec	 accuracy=57.702446	 loss=1.767367	 lr=0.010000
Epoch[018] Batch [2349]/[3760]	Speed: 34.985021 samples/sec	 accuracy=57.712766	 loss=1.767593	 lr=0.010000
Epoch[018] Batch [2399]/[3760]	Speed: 35.370559 samples/sec	 accuracy=57.693359	 loss=1.768162	 lr=0.010000
Epoch[018] Batch [2449]/[3760]	Speed: 35.168956 samples/sec	 accuracy=57.677934	 loss=1.768794	 lr=0.010000
Epoch[018] Batch [2499]/[3760]	Speed: 35.218433 samples/sec	 accuracy=57.683125	 loss=1.768829	 lr=0.010000
Epoch[018] Batch [2549]/[3760]	Speed: 35.132506 samples/sec	 accuracy=57.670343	 loss=1.769619	 lr=0.010000
Epoch[018] Batch [2599]/[3760]	Speed: 34.985361 samples/sec	 accuracy=57.662260	 loss=1.769686	 lr=0.010000
Epoch[018] Batch [2649]/[3760]	Speed: 35.348950 samples/sec	 accuracy=57.671580	 loss=1.769274	 lr=0.010000
Epoch[018] Batch [2699]/[3760]	Speed: 34.861619 samples/sec	 accuracy=57.637731	 loss=1.771146	 lr=0.010000
Epoch[018] Batch [2749]/[3760]	Speed: 35.378318 samples/sec	 accuracy=57.636932	 loss=1.771354	 lr=0.010000
Epoch[018] Batch [2799]/[3760]	Speed: 34.977712 samples/sec	 accuracy=57.639509	 loss=1.771388	 lr=0.010000
Epoch[018] Batch [2849]/[3760]	Speed: 35.586015 samples/sec	 accuracy=57.631579	 loss=1.771067	 lr=0.010000
Epoch[018] Batch [2899]/[3760]	Speed: 35.206787 samples/sec	 accuracy=57.661099	 loss=1.770524	 lr=0.010000
Epoch[018] Batch [2949]/[3760]	Speed: 35.279850 samples/sec	 accuracy=57.650424	 loss=1.771552	 lr=0.010000
Epoch[018] Batch [2999]/[3760]	Speed: 34.866447 samples/sec	 accuracy=57.632292	 loss=1.771977	 lr=0.010000
Epoch[018] Batch [3049]/[3760]	Speed: 35.572336 samples/sec	 accuracy=57.638832	 loss=1.772249	 lr=0.010000
Epoch[018] Batch [3099]/[3760]	Speed: 34.967409 samples/sec	 accuracy=57.633065	 loss=1.772232	 lr=0.010000
Epoch[018] Batch [3149]/[3760]	Speed: 35.332647 samples/sec	 accuracy=57.647817	 loss=1.772522	 lr=0.010000
Epoch[018] Batch [3199]/[3760]	Speed: 34.967733 samples/sec	 accuracy=57.639160	 loss=1.772588	 lr=0.010000
Epoch[018] Batch [3249]/[3760]	Speed: 35.332077 samples/sec	 accuracy=57.610096	 loss=1.774196	 lr=0.010000
Epoch[018] Batch [3299]/[3760]	Speed: 35.306245 samples/sec	 accuracy=57.588068	 loss=1.774794	 lr=0.010000
Epoch[018] Batch [3349]/[3760]	Speed: 35.393302 samples/sec	 accuracy=57.552705	 loss=1.775554	 lr=0.010000
Epoch[018] Batch [3399]/[3760]	Speed: 35.380452 samples/sec	 accuracy=57.532629	 loss=1.776576	 lr=0.010000
Epoch[018] Batch [3449]/[3760]	Speed: 35.291122 samples/sec	 accuracy=57.515851	 loss=1.777013	 lr=0.010000
Epoch[018] Batch [3499]/[3760]	Speed: 35.306124 samples/sec	 accuracy=57.487946	 loss=1.778255	 lr=0.010000
Epoch[018] Batch [3549]/[3760]	Speed: 35.582358 samples/sec	 accuracy=57.495599	 loss=1.778040	 lr=0.010000
Epoch[018] Batch [3599]/[3760]	Speed: 35.374045 samples/sec	 accuracy=57.496528	 loss=1.778535	 lr=0.010000
Epoch[018] Batch [3649]/[3760]	Speed: 35.281875 samples/sec	 accuracy=57.500428	 loss=1.778575	 lr=0.010000
Epoch[018] Batch [3699]/[3760]	Speed: 35.614824 samples/sec	 accuracy=57.511824	 loss=1.778423	 lr=0.010000
Epoch[018] Batch [3749]/[3760]	Speed: 36.509485 samples/sec	 accuracy=57.505417	 loss=1.778518	 lr=0.010000
Batch [0049]/[0303]: acc-top1=56.468750 acc-top5=80.218750
Batch [0099]/[0303]: acc-top1=56.687500 acc-top5=80.031250
Batch [0149]/[0303]: acc-top1=57.020833 acc-top5=80.437500
Batch [0199]/[0303]: acc-top1=56.757813 acc-top5=80.289062
Batch [0249]/[0303]: acc-top1=56.468750 acc-top5=79.981250
Batch [0299]/[0303]: acc-top1=56.609375 acc-top5=79.979167
[Epoch 018] training: accuracy=57.510805	 loss=1.778333
[Epoch 018] speed: 34 samples/sec	time cost: 7356.541609
[Epoch 018] validation: acc-top1=56.574876 acc-top5=79.965965 loss=2.061742
Epoch[019] Batch [0049]/[3760]	Speed: 24.490226 samples/sec	 accuracy=61.000000	 loss=1.616011	 lr=0.010000
Epoch[019] Batch [0099]/[3760]	Speed: 35.128548 samples/sec	 accuracy=60.046875	 loss=1.681340	 lr=0.010000
Epoch[019] Batch [0149]/[3760]	Speed: 35.146571 samples/sec	 accuracy=59.697917	 loss=1.693723	 lr=0.010000
Epoch[019] Batch [0199]/[3760]	Speed: 35.099852 samples/sec	 accuracy=59.257812	 loss=1.711800	 lr=0.010000
Epoch[019] Batch [0249]/[3760]	Speed: 35.284373 samples/sec	 accuracy=59.037500	 loss=1.729113	 lr=0.010000
Epoch[019] Batch [0299]/[3760]	Speed: 34.953546 samples/sec	 accuracy=58.854167	 loss=1.728818	 lr=0.010000
Epoch[019] Batch [0349]/[3760]	Speed: 35.258892 samples/sec	 accuracy=58.941964	 loss=1.726644	 lr=0.010000
Epoch[019] Batch [0399]/[3760]	Speed: 34.997109 samples/sec	 accuracy=58.890625	 loss=1.727792	 lr=0.010000
Epoch[019] Batch [0449]/[3760]	Speed: 35.597529 samples/sec	 accuracy=58.670139	 loss=1.733997	 lr=0.010000
Epoch[019] Batch [0499]/[3760]	Speed: 35.203104 samples/sec	 accuracy=58.756250	 loss=1.729973	 lr=0.010000
Epoch[019] Batch [0549]/[3760]	Speed: 35.570574 samples/sec	 accuracy=58.727273	 loss=1.728931	 lr=0.010000
Epoch[019] Batch [0599]/[3760]	Speed: 35.054342 samples/sec	 accuracy=58.466146	 loss=1.737305	 lr=0.010000
Epoch[019] Batch [0649]/[3760]	Speed: 35.422035 samples/sec	 accuracy=58.415865	 loss=1.736224	 lr=0.010000
Epoch[019] Batch [0699]/[3760]	Speed: 35.276441 samples/sec	 accuracy=58.339286	 loss=1.739676	 lr=0.010000
Epoch[019] Batch [0749]/[3760]	Speed: 35.336015 samples/sec	 accuracy=58.295833	 loss=1.743051	 lr=0.010000
Epoch[019] Batch [0799]/[3760]	Speed: 35.468396 samples/sec	 accuracy=58.318359	 loss=1.743570	 lr=0.010000
Epoch[019] Batch [0849]/[3760]	Speed: 35.421328 samples/sec	 accuracy=58.264706	 loss=1.744747	 lr=0.010000
Epoch[019] Batch [0899]/[3760]	Speed: 35.400320 samples/sec	 accuracy=58.288194	 loss=1.743592	 lr=0.010000
Epoch[019] Batch [0949]/[3760]	Speed: 35.341684 samples/sec	 accuracy=58.365132	 loss=1.740250	 lr=0.010000
Epoch[019] Batch [0999]/[3760]	Speed: 35.051131 samples/sec	 accuracy=58.398438	 loss=1.742098	 lr=0.010000
Epoch[019] Batch [1049]/[3760]	Speed: 35.391181 samples/sec	 accuracy=58.382440	 loss=1.740743	 lr=0.010000
Epoch[019] Batch [1099]/[3760]	Speed: 35.282838 samples/sec	 accuracy=58.275568	 loss=1.743385	 lr=0.010000
Epoch[019] Batch [1149]/[3760]	Speed: 35.376679 samples/sec	 accuracy=58.301630	 loss=1.741843	 lr=0.010000
Epoch[019] Batch [1199]/[3760]	Speed: 35.449706 samples/sec	 accuracy=58.313802	 loss=1.741886	 lr=0.010000
Epoch[019] Batch [1249]/[3760]	Speed: 35.256563 samples/sec	 accuracy=58.291250	 loss=1.743981	 lr=0.010000
Epoch[019] Batch [1299]/[3760]	Speed: 35.402885 samples/sec	 accuracy=58.296875	 loss=1.743280	 lr=0.010000
Epoch[019] Batch [1349]/[3760]	Speed: 35.370819 samples/sec	 accuracy=58.283565	 loss=1.743880	 lr=0.010000
Epoch[019] Batch [1399]/[3760]	Speed: 35.433084 samples/sec	 accuracy=58.286830	 loss=1.744031	 lr=0.010000
Epoch[019] Batch [1449]/[3760]	Speed: 35.134883 samples/sec	 accuracy=58.248922	 loss=1.745624	 lr=0.010000
Epoch[019] Batch [1499]/[3760]	Speed: 35.411566 samples/sec	 accuracy=58.193750	 loss=1.747078	 lr=0.010000
Epoch[019] Batch [1549]/[3760]	Speed: 35.116995 samples/sec	 accuracy=58.175403	 loss=1.747774	 lr=0.010000
Epoch[019] Batch [1599]/[3760]	Speed: 35.213669 samples/sec	 accuracy=58.134766	 loss=1.750146	 lr=0.010000
Epoch[019] Batch [1649]/[3760]	Speed: 35.405829 samples/sec	 accuracy=58.075758	 loss=1.752056	 lr=0.010000
Epoch[019] Batch [1699]/[3760]	Speed: 35.325973 samples/sec	 accuracy=58.047794	 loss=1.752577	 lr=0.010000
Epoch[019] Batch [1749]/[3760]	Speed: 35.402574 samples/sec	 accuracy=58.068750	 loss=1.751438	 lr=0.010000
Epoch[019] Batch [1799]/[3760]	Speed: 35.377756 samples/sec	 accuracy=58.064236	 loss=1.751687	 lr=0.010000
Epoch[019] Batch [1849]/[3760]	Speed: 35.354327 samples/sec	 accuracy=58.020270	 loss=1.752384	 lr=0.010000
Epoch[019] Batch [1899]/[3760]	Speed: 35.332011 samples/sec	 accuracy=58.014803	 loss=1.753107	 lr=0.010000
Epoch[019] Batch [1949]/[3760]	Speed: 35.516776 samples/sec	 accuracy=58.005609	 loss=1.752966	 lr=0.010000
Epoch[019] Batch [1999]/[3760]	Speed: 35.417291 samples/sec	 accuracy=57.971094	 loss=1.753225	 lr=0.010000
Epoch[019] Batch [2049]/[3760]	Speed: 35.210847 samples/sec	 accuracy=57.965701	 loss=1.753625	 lr=0.010000
Epoch[019] Batch [2099]/[3760]	Speed: 35.524010 samples/sec	 accuracy=57.968006	 loss=1.753516	 lr=0.010000
Epoch[019] Batch [2149]/[3760]	Speed: 34.972969 samples/sec	 accuracy=57.933140	 loss=1.754066	 lr=0.010000
Epoch[019] Batch [2199]/[3760]	Speed: 35.623169 samples/sec	 accuracy=57.937500	 loss=1.754236	 lr=0.010000
Epoch[019] Batch [2249]/[3760]	Speed: 35.424254 samples/sec	 accuracy=57.904167	 loss=1.754883	 lr=0.010000
Epoch[019] Batch [2299]/[3760]	Speed: 35.210207 samples/sec	 accuracy=57.914402	 loss=1.755026	 lr=0.010000
Epoch[019] Batch [2349]/[3760]	Speed: 35.370889 samples/sec	 accuracy=57.946144	 loss=1.755302	 lr=0.010000
Epoch[019] Batch [2399]/[3760]	Speed: 34.982730 samples/sec	 accuracy=57.907552	 loss=1.756703	 lr=0.010000
Epoch[019] Batch [2449]/[3760]	Speed: 35.488709 samples/sec	 accuracy=57.907526	 loss=1.758011	 lr=0.010000
Epoch[019] Batch [2499]/[3760]	Speed: 35.262932 samples/sec	 accuracy=57.888750	 loss=1.757993	 lr=0.010000
Epoch[019] Batch [2549]/[3760]	Speed: 35.325757 samples/sec	 accuracy=57.900123	 loss=1.758077	 lr=0.010000
Epoch[019] Batch [2599]/[3760]	Speed: 35.262579 samples/sec	 accuracy=57.891226	 loss=1.758445	 lr=0.010000
Epoch[019] Batch [2649]/[3760]	Speed: 35.121661 samples/sec	 accuracy=57.892099	 loss=1.758378	 lr=0.010000
Epoch[019] Batch [2699]/[3760]	Speed: 35.267466 samples/sec	 accuracy=57.878472	 loss=1.758956	 lr=0.010000
Epoch[019] Batch [2749]/[3760]	Speed: 34.755180 samples/sec	 accuracy=57.865909	 loss=1.759794	 lr=0.010000
Epoch[019] Batch [2799]/[3760]	Speed: 35.394267 samples/sec	 accuracy=57.852679	 loss=1.760636	 lr=0.010000
Epoch[019] Batch [2849]/[3760]	Speed: 35.261453 samples/sec	 accuracy=57.837171	 loss=1.761663	 lr=0.010000
Epoch[019] Batch [2899]/[3760]	Speed: 35.422618 samples/sec	 accuracy=57.859914	 loss=1.761284	 lr=0.010000
Epoch[019] Batch [2949]/[3760]	Speed: 35.141451 samples/sec	 accuracy=57.870763	 loss=1.761260	 lr=0.010000
Epoch[019] Batch [2999]/[3760]	Speed: 35.127930 samples/sec	 accuracy=57.880208	 loss=1.761245	 lr=0.010000
Epoch[019] Batch [3049]/[3760]	Speed: 35.349980 samples/sec	 accuracy=57.868852	 loss=1.761592	 lr=0.010000
Epoch[019] Batch [3099]/[3760]	Speed: 35.218399 samples/sec	 accuracy=57.872480	 loss=1.761272	 lr=0.010000
Epoch[019] Batch [3149]/[3760]	Speed: 35.144576 samples/sec	 accuracy=57.877976	 loss=1.761252	 lr=0.010000
Epoch[019] Batch [3199]/[3760]	Speed: 34.844690 samples/sec	 accuracy=57.871094	 loss=1.761427	 lr=0.010000
Epoch[019] Batch [3249]/[3760]	Speed: 35.281932 samples/sec	 accuracy=57.848558	 loss=1.762493	 lr=0.010000
Epoch[019] Batch [3299]/[3760]	Speed: 34.835764 samples/sec	 accuracy=57.840436	 loss=1.762769	 lr=0.010000
Epoch[019] Batch [3349]/[3760]	Speed: 35.030315 samples/sec	 accuracy=57.803638	 loss=1.764121	 lr=0.010000
Epoch[019] Batch [3399]/[3760]	Speed: 34.729449 samples/sec	 accuracy=57.819853	 loss=1.763647	 lr=0.010000
Epoch[019] Batch [3449]/[3760]	Speed: 35.066088 samples/sec	 accuracy=57.818388	 loss=1.763233	 lr=0.010000
Epoch[019] Batch [3499]/[3760]	Speed: 34.913718 samples/sec	 accuracy=57.835268	 loss=1.762394	 lr=0.010000
Epoch[019] Batch [3549]/[3760]	Speed: 34.796004 samples/sec	 accuracy=57.834507	 loss=1.762443	 lr=0.010000
Epoch[019] Batch [3599]/[3760]	Speed: 35.139431 samples/sec	 accuracy=57.824219	 loss=1.762937	 lr=0.010000
Epoch[019] Batch [3649]/[3760]	Speed: 34.542916 samples/sec	 accuracy=57.804795	 loss=1.763469	 lr=0.010000
Epoch[019] Batch [3699]/[3760]	Speed: 35.299497 samples/sec	 accuracy=57.788851	 loss=1.764175	 lr=0.010000
Epoch[019] Batch [3749]/[3760]	Speed: 36.072995 samples/sec	 accuracy=57.797917	 loss=1.763614	 lr=0.010000
Batch [0049]/[0303]: acc-top1=55.875000 acc-top5=80.281250
Batch [0099]/[0303]: acc-top1=56.484375 acc-top5=80.187500
Batch [0149]/[0303]: acc-top1=57.010417 acc-top5=80.260417
Batch [0199]/[0303]: acc-top1=56.789062 acc-top5=80.156250
Batch [0249]/[0303]: acc-top1=56.650000 acc-top5=80.075000
Batch [0299]/[0303]: acc-top1=56.661458 acc-top5=79.984375
[Epoch 019] training: accuracy=57.794215	 loss=1.763660
[Epoch 019] speed: 35 samples/sec	time cost: 7343.900899
[Epoch 019] validation: acc-top1=56.647071 acc-top5=79.976279 loss=2.067913
Epoch[020] Batch [0049]/[3759]	Speed: 23.861363 samples/sec	 accuracy=59.687500	 loss=1.677954	 lr=0.010000
Epoch[020] Batch [0099]/[3759]	Speed: 35.281373 samples/sec	 accuracy=59.656250	 loss=1.670663	 lr=0.010000
Epoch[020] Batch [0149]/[3759]	Speed: 34.947708 samples/sec	 accuracy=59.489583	 loss=1.682132	 lr=0.010000
Epoch[020] Batch [0199]/[3759]	Speed: 35.019097 samples/sec	 accuracy=59.554688	 loss=1.671838	 lr=0.010000
Epoch[020] Batch [0249]/[3759]	Speed: 35.283582 samples/sec	 accuracy=59.281250	 loss=1.684130	 lr=0.010000
Epoch[020] Batch [0299]/[3759]	Speed: 34.924789 samples/sec	 accuracy=59.083333	 loss=1.689785	 lr=0.010000
Epoch[020] Batch [0349]/[3759]	Speed: 35.428134 samples/sec	 accuracy=58.897321	 loss=1.701421	 lr=0.010000
Epoch[020] Batch [0399]/[3759]	Speed: 34.855827 samples/sec	 accuracy=59.003906	 loss=1.699628	 lr=0.010000
Epoch[020] Batch [0449]/[3759]	Speed: 35.542041 samples/sec	 accuracy=58.670139	 loss=1.709509	 lr=0.010000
Epoch[020] Batch [0499]/[3759]	Speed: 35.059511 samples/sec	 accuracy=58.609375	 loss=1.714267	 lr=0.010000
Epoch[020] Batch [0549]/[3759]	Speed: 35.399177 samples/sec	 accuracy=58.548295	 loss=1.715311	 lr=0.010000
Epoch[020] Batch [0599]/[3759]	Speed: 35.144608 samples/sec	 accuracy=58.481771	 loss=1.714739	 lr=0.010000
Epoch[020] Batch [0649]/[3759]	Speed: 35.494972 samples/sec	 accuracy=58.384615	 loss=1.716960	 lr=0.010000
Epoch[020] Batch [0699]/[3759]	Speed: 35.202670 samples/sec	 accuracy=58.392857	 loss=1.719385	 lr=0.010000
Epoch[020] Batch [0749]/[3759]	Speed: 35.336352 samples/sec	 accuracy=58.420833	 loss=1.719583	 lr=0.010000
Epoch[020] Batch [0799]/[3759]	Speed: 35.402004 samples/sec	 accuracy=58.351562	 loss=1.724897	 lr=0.010000
Epoch[020] Batch [0849]/[3759]	Speed: 35.276785 samples/sec	 accuracy=58.422794	 loss=1.722703	 lr=0.010000
Epoch[020] Batch [0899]/[3759]	Speed: 35.386210 samples/sec	 accuracy=58.414931	 loss=1.723111	 lr=0.010000
Epoch[020] Batch [0949]/[3759]	Speed: 35.309313 samples/sec	 accuracy=58.472039	 loss=1.720825	 lr=0.010000
Epoch[020] Batch [0999]/[3759]	Speed: 35.303919 samples/sec	 accuracy=58.371875	 loss=1.725412	 lr=0.010000
Epoch[020] Batch [1049]/[3759]	Speed: 35.447793 samples/sec	 accuracy=58.331845	 loss=1.727197	 lr=0.010000
Epoch[020] Batch [1099]/[3759]	Speed: 35.117740 samples/sec	 accuracy=58.400568	 loss=1.723759	 lr=0.010000
Epoch[020] Batch [1149]/[3759]	Speed: 35.518722 samples/sec	 accuracy=58.387228	 loss=1.725935	 lr=0.010000
Epoch[020] Batch [1199]/[3759]	Speed: 34.978576 samples/sec	 accuracy=58.355469	 loss=1.728111	 lr=0.010000
Epoch[020] Batch [1249]/[3759]	Speed: 35.455968 samples/sec	 accuracy=58.377500	 loss=1.727439	 lr=0.010000
Epoch[020] Batch [1299]/[3759]	Speed: 35.212484 samples/sec	 accuracy=58.405048	 loss=1.727197	 lr=0.010000
Epoch[020] Batch [1349]/[3759]	Speed: 35.522887 samples/sec	 accuracy=58.417824	 loss=1.728001	 lr=0.010000
Epoch[020] Batch [1399]/[3759]	Speed: 35.585001 samples/sec	 accuracy=58.373884	 loss=1.728451	 lr=0.010000
Epoch[020] Batch [1449]/[3759]	Speed: 35.367695 samples/sec	 accuracy=58.346983	 loss=1.730335	 lr=0.010000
Epoch[020] Batch [1499]/[3759]	Speed: 35.269637 samples/sec	 accuracy=58.309375	 loss=1.732233	 lr=0.010000
Epoch[020] Batch [1549]/[3759]	Speed: 35.343982 samples/sec	 accuracy=58.331653	 loss=1.732545	 lr=0.010000
Epoch[020] Batch [1599]/[3759]	Speed: 35.277551 samples/sec	 accuracy=58.355469	 loss=1.732817	 lr=0.010000
Epoch[020] Batch [1649]/[3759]	Speed: 35.259048 samples/sec	 accuracy=58.396780	 loss=1.732534	 lr=0.010000
Epoch[020] Batch [1699]/[3759]	Speed: 35.536453 samples/sec	 accuracy=58.409926	 loss=1.732794	 lr=0.010000
Epoch[020] Batch [1749]/[3759]	Speed: 35.002653 samples/sec	 accuracy=58.357143	 loss=1.734938	 lr=0.010000
Epoch[020] Batch [1799]/[3759]	Speed: 35.355148 samples/sec	 accuracy=58.391493	 loss=1.733972	 lr=0.010000
Epoch[020] Batch [1849]/[3759]	Speed: 35.031814 samples/sec	 accuracy=58.407095	 loss=1.734021	 lr=0.010000
Epoch[020] Batch [1899]/[3759]	Speed: 35.530234 samples/sec	 accuracy=58.377467	 loss=1.735319	 lr=0.010000
Epoch[020] Batch [1949]/[3759]	Speed: 35.089563 samples/sec	 accuracy=58.377404	 loss=1.736142	 lr=0.010000
Epoch[020] Batch [1999]/[3759]	Speed: 35.010690 samples/sec	 accuracy=58.358594	 loss=1.736391	 lr=0.010000
Epoch[020] Batch [2049]/[3759]	Speed: 35.244530 samples/sec	 accuracy=58.354421	 loss=1.736412	 lr=0.010000
Epoch[020] Batch [2099]/[3759]	Speed: 35.101213 samples/sec	 accuracy=58.331845	 loss=1.737564	 lr=0.010000
Epoch[020] Batch [2149]/[3759]	Speed: 35.420839 samples/sec	 accuracy=58.329215	 loss=1.738029	 lr=0.010000
Epoch[020] Batch [2199]/[3759]	Speed: 34.957223 samples/sec	 accuracy=58.343750	 loss=1.737777	 lr=0.010000
Epoch[020] Batch [2249]/[3759]	Speed: 35.532433 samples/sec	 accuracy=58.329167	 loss=1.738550	 lr=0.010000
Epoch[020] Batch [2299]/[3759]	Speed: 34.834702 samples/sec	 accuracy=58.315897	 loss=1.738033	 lr=0.010000
Epoch[020] Batch [2349]/[3759]	Speed: 35.400927 samples/sec	 accuracy=58.335771	 loss=1.736923	 lr=0.010000
Epoch[020] Batch [2399]/[3759]	Speed: 34.773761 samples/sec	 accuracy=58.327474	 loss=1.737457	 lr=0.010000
Epoch[020] Batch [2449]/[3759]	Speed: 35.435770 samples/sec	 accuracy=58.325893	 loss=1.737669	 lr=0.010000
Epoch[020] Batch [2499]/[3759]	Speed: 34.990712 samples/sec	 accuracy=58.308125	 loss=1.738682	 lr=0.010000
Epoch[020] Batch [2549]/[3759]	Speed: 35.463062 samples/sec	 accuracy=58.323529	 loss=1.738358	 lr=0.010000
Epoch[020] Batch [2599]/[3759]	Speed: 35.221205 samples/sec	 accuracy=58.312500	 loss=1.739491	 lr=0.010000
Epoch[020] Batch [2649]/[3759]	Speed: 35.062424 samples/sec	 accuracy=58.314269	 loss=1.739270	 lr=0.010000
Epoch[020] Batch [2699]/[3759]	Speed: 35.219076 samples/sec	 accuracy=58.313657	 loss=1.738717	 lr=0.010000
Epoch[020] Batch [2749]/[3759]	Speed: 34.806597 samples/sec	 accuracy=58.314773	 loss=1.739344	 lr=0.010000
Epoch[020] Batch [2799]/[3759]	Speed: 35.524934 samples/sec	 accuracy=58.301897	 loss=1.740083	 lr=0.010000
Epoch[020] Batch [2849]/[3759]	Speed: 34.756504 samples/sec	 accuracy=58.301535	 loss=1.739656	 lr=0.010000
Epoch[020] Batch [2899]/[3759]	Speed: 35.590513 samples/sec	 accuracy=58.294181	 loss=1.740040	 lr=0.010000
Epoch[020] Batch [2949]/[3759]	Speed: 34.970431 samples/sec	 accuracy=58.327331	 loss=1.739591	 lr=0.010000
Epoch[020] Batch [2999]/[3759]	Speed: 35.312249 samples/sec	 accuracy=58.336458	 loss=1.739818	 lr=0.010000
Epoch[020] Batch [3049]/[3759]	Speed: 35.123536 samples/sec	 accuracy=58.337602	 loss=1.740141	 lr=0.010000
Epoch[020] Batch [3099]/[3759]	Speed: 35.270346 samples/sec	 accuracy=58.328125	 loss=1.740826	 lr=0.010000
Epoch[020] Batch [3149]/[3759]	Speed: 35.220969 samples/sec	 accuracy=58.308036	 loss=1.741375	 lr=0.010000
Epoch[020] Batch [3199]/[3759]	Speed: 35.039245 samples/sec	 accuracy=58.284668	 loss=1.741808	 lr=0.010000
Epoch[020] Batch [3249]/[3759]	Speed: 35.307206 samples/sec	 accuracy=58.297115	 loss=1.741222	 lr=0.010000
Epoch[020] Batch [3299]/[3759]	Speed: 35.054356 samples/sec	 accuracy=58.264678	 loss=1.743100	 lr=0.010000
Epoch[020] Batch [3349]/[3759]	Speed: 35.466707 samples/sec	 accuracy=58.249067	 loss=1.743367	 lr=0.010000
Epoch[020] Batch [3399]/[3759]	Speed: 34.979995 samples/sec	 accuracy=58.250919	 loss=1.743913	 lr=0.010000
Epoch[020] Batch [3449]/[3759]	Speed: 35.599674 samples/sec	 accuracy=58.248641	 loss=1.744435	 lr=0.010000
Epoch[020] Batch [3499]/[3759]	Speed: 35.119437 samples/sec	 accuracy=58.234375	 loss=1.745324	 lr=0.010000
Epoch[020] Batch [3549]/[3759]	Speed: 35.345576 samples/sec	 accuracy=58.239437	 loss=1.745628	 lr=0.010000
Epoch[020] Batch [3599]/[3759]	Speed: 35.253136 samples/sec	 accuracy=58.228733	 loss=1.745688	 lr=0.010000
Epoch[020] Batch [3649]/[3759]	Speed: 35.392306 samples/sec	 accuracy=58.225171	 loss=1.746068	 lr=0.010000
Epoch[020] Batch [3699]/[3759]	Speed: 34.932639 samples/sec	 accuracy=58.210726	 loss=1.746348	 lr=0.010000
Epoch[020] Batch [3749]/[3759]	Speed: 36.635702 samples/sec	 accuracy=58.213750	 loss=1.746616	 lr=0.010000
Batch [0049]/[0303]: acc-top1=56.562500 acc-top5=80.437500
Batch [0099]/[0303]: acc-top1=57.187500 acc-top5=80.500000
Batch [0149]/[0303]: acc-top1=57.885417 acc-top5=80.666667
Batch [0199]/[0303]: acc-top1=57.898438 acc-top5=80.773438
Batch [0249]/[0303]: acc-top1=57.750000 acc-top5=80.556250
Batch [0299]/[0303]: acc-top1=57.557292 acc-top5=80.557292
[Epoch 020] training: accuracy=58.213205	 loss=1.746438
[Epoch 020] speed: 35 samples/sec	time cost: 7346.550678
[Epoch 020] validation: acc-top1=57.534035 acc-top5=80.548680 loss=2.019325
Epoch[021] Batch [0049]/[3760]	Speed: 24.429123 samples/sec	 accuracy=58.468750	 loss=1.721667	 lr=0.010000
Epoch[021] Batch [0099]/[3760]	Speed: 35.057741 samples/sec	 accuracy=58.921875	 loss=1.718535	 lr=0.010000
Epoch[021] Batch [0149]/[3760]	Speed: 34.877394 samples/sec	 accuracy=59.489583	 loss=1.694931	 lr=0.010000
Epoch[021] Batch [0199]/[3760]	Speed: 34.672392 samples/sec	 accuracy=59.531250	 loss=1.686017	 lr=0.010000
Epoch[021] Batch [0249]/[3760]	Speed: 35.121980 samples/sec	 accuracy=59.387500	 loss=1.680872	 lr=0.010000
Epoch[021] Batch [0299]/[3760]	Speed: 34.755813 samples/sec	 accuracy=59.406250	 loss=1.682724	 lr=0.010000
Epoch[021] Batch [0349]/[3760]	Speed: 35.202227 samples/sec	 accuracy=59.303571	 loss=1.684256	 lr=0.010000
Epoch[021] Batch [0399]/[3760]	Speed: 34.874309 samples/sec	 accuracy=59.402344	 loss=1.680950	 lr=0.010000
Epoch[021] Batch [0449]/[3760]	Speed: 35.387378 samples/sec	 accuracy=59.343750	 loss=1.685283	 lr=0.010000
Epoch[021] Batch [0499]/[3760]	Speed: 34.904729 samples/sec	 accuracy=59.296875	 loss=1.689490	 lr=0.010000
Epoch[021] Batch [0549]/[3760]	Speed: 35.300734 samples/sec	 accuracy=59.289773	 loss=1.692424	 lr=0.010000
Epoch[021] Batch [0599]/[3760]	Speed: 35.092473 samples/sec	 accuracy=59.182292	 loss=1.696919	 lr=0.010000
Epoch[021] Batch [0649]/[3760]	Speed: 35.060706 samples/sec	 accuracy=59.137019	 loss=1.701235	 lr=0.010000
Epoch[021] Batch [0699]/[3760]	Speed: 35.085865 samples/sec	 accuracy=59.136161	 loss=1.698043	 lr=0.010000
Epoch[021] Batch [0749]/[3760]	Speed: 35.019933 samples/sec	 accuracy=59.120833	 loss=1.701403	 lr=0.010000
Epoch[021] Batch [0799]/[3760]	Speed: 34.994697 samples/sec	 accuracy=59.128906	 loss=1.702363	 lr=0.010000
Epoch[021] Batch [0849]/[3760]	Speed: 34.687207 samples/sec	 accuracy=59.038603	 loss=1.704130	 lr=0.010000
Epoch[021] Batch [0899]/[3760]	Speed: 35.247529 samples/sec	 accuracy=59.057292	 loss=1.704971	 lr=0.010000
Epoch[021] Batch [0949]/[3760]	Speed: 34.551916 samples/sec	 accuracy=58.995066	 loss=1.708517	 lr=0.010000
Epoch[021] Batch [0999]/[3760]	Speed: 35.329603 samples/sec	 accuracy=58.937500	 loss=1.710707	 lr=0.010000
Epoch[021] Batch [1049]/[3760]	Speed: 34.785959 samples/sec	 accuracy=58.927083	 loss=1.709452	 lr=0.010000
Epoch[021] Batch [1099]/[3760]	Speed: 35.214251 samples/sec	 accuracy=58.872159	 loss=1.714062	 lr=0.010000
Epoch[021] Batch [1149]/[3760]	Speed: 34.737488 samples/sec	 accuracy=58.811141	 loss=1.715408	 lr=0.010000
Epoch[021] Batch [1199]/[3760]	Speed: 34.834321 samples/sec	 accuracy=58.759115	 loss=1.717363	 lr=0.010000
Epoch[021] Batch [1249]/[3760]	Speed: 34.977283 samples/sec	 accuracy=58.781250	 loss=1.716910	 lr=0.010000
Epoch[021] Batch [1299]/[3760]	Speed: 34.603301 samples/sec	 accuracy=58.775240	 loss=1.718300	 lr=0.010000
Epoch[021] Batch [1349]/[3760]	Speed: 35.147599 samples/sec	 accuracy=58.730324	 loss=1.719056	 lr=0.010000
Epoch[021] Batch [1399]/[3760]	Speed: 34.532094 samples/sec	 accuracy=58.686384	 loss=1.720266	 lr=0.010000
Epoch[021] Batch [1449]/[3760]	Speed: 35.331538 samples/sec	 accuracy=58.716595	 loss=1.719669	 lr=0.010000
Epoch[021] Batch [1499]/[3760]	Speed: 34.621982 samples/sec	 accuracy=58.702083	 loss=1.720843	 lr=0.010000
Epoch[021] Batch [1549]/[3760]	Speed: 35.404272 samples/sec	 accuracy=58.718750	 loss=1.718579	 lr=0.010000
Epoch[021] Batch [1599]/[3760]	Speed: 34.909557 samples/sec	 accuracy=58.759766	 loss=1.717035	 lr=0.010000
Epoch[021] Batch [1649]/[3760]	Speed: 35.255663 samples/sec	 accuracy=58.749053	 loss=1.716544	 lr=0.010000
Epoch[021] Batch [1699]/[3760]	Speed: 35.326637 samples/sec	 accuracy=58.727022	 loss=1.717557	 lr=0.010000
Epoch[021] Batch [1749]/[3760]	Speed: 35.144225 samples/sec	 accuracy=58.715179	 loss=1.717321	 lr=0.010000
Epoch[021] Batch [1799]/[3760]	Speed: 35.357343 samples/sec	 accuracy=58.707465	 loss=1.718107	 lr=0.010000
Epoch[021] Batch [1849]/[3760]	Speed: 35.076400 samples/sec	 accuracy=58.698480	 loss=1.719175	 lr=0.010000
Epoch[021] Batch [1899]/[3760]	Speed: 35.270886 samples/sec	 accuracy=58.699836	 loss=1.719450	 lr=0.010000
Epoch[021] Batch [1949]/[3760]	Speed: 35.024244 samples/sec	 accuracy=58.713942	 loss=1.719793	 lr=0.010000
Epoch[021] Batch [1999]/[3760]	Speed: 35.237978 samples/sec	 accuracy=58.739063	 loss=1.719089	 lr=0.010000
Epoch[021] Batch [2049]/[3760]	Speed: 34.642885 samples/sec	 accuracy=58.705030	 loss=1.719644	 lr=0.010000
Epoch[021] Batch [2099]/[3760]	Speed: 35.341812 samples/sec	 accuracy=58.707589	 loss=1.719355	 lr=0.010000
Epoch[021] Batch [2149]/[3760]	Speed: 34.700441 samples/sec	 accuracy=58.644622	 loss=1.722250	 lr=0.010000
Epoch[021] Batch [2199]/[3760]	Speed: 35.067215 samples/sec	 accuracy=58.603693	 loss=1.724033	 lr=0.010000
Epoch[021] Batch [2249]/[3760]	Speed: 35.004798 samples/sec	 accuracy=58.583333	 loss=1.725127	 lr=0.010000
Epoch[021] Batch [2299]/[3760]	Speed: 34.826345 samples/sec	 accuracy=58.546196	 loss=1.726556	 lr=0.010000
Epoch[021] Batch [2349]/[3760]	Speed: 35.029976 samples/sec	 accuracy=58.565824	 loss=1.725953	 lr=0.010000
Epoch[021] Batch [2399]/[3760]	Speed: 34.680170 samples/sec	 accuracy=58.561198	 loss=1.725553	 lr=0.010000
Epoch[021] Batch [2449]/[3760]	Speed: 35.243765 samples/sec	 accuracy=58.570153	 loss=1.724311	 lr=0.010000
Epoch[021] Batch [2499]/[3760]	Speed: 34.362978 samples/sec	 accuracy=58.561875	 loss=1.725135	 lr=0.010000
Epoch[021] Batch [2549]/[3760]	Speed: 35.422191 samples/sec	 accuracy=58.588235	 loss=1.723905	 lr=0.010000
Epoch[021] Batch [2599]/[3760]	Speed: 34.705120 samples/sec	 accuracy=58.569712	 loss=1.724312	 lr=0.010000
Epoch[021] Batch [2649]/[3760]	Speed: 35.178001 samples/sec	 accuracy=58.573703	 loss=1.724544	 lr=0.010000
Epoch[021] Batch [2699]/[3760]	Speed: 34.808994 samples/sec	 accuracy=58.545718	 loss=1.725868	 lr=0.010000
Epoch[021] Batch [2749]/[3760]	Speed: 35.240004 samples/sec	 accuracy=58.527841	 loss=1.727348	 lr=0.010000
Epoch[021] Batch [2799]/[3760]	Speed: 35.077702 samples/sec	 accuracy=58.560268	 loss=1.727128	 lr=0.010000
Epoch[021] Batch [2849]/[3760]	Speed: 35.129150 samples/sec	 accuracy=58.541118	 loss=1.727707	 lr=0.010000
Epoch[021] Batch [2899]/[3760]	Speed: 35.427674 samples/sec	 accuracy=58.530711	 loss=1.727112	 lr=0.010000
Epoch[021] Batch [2949]/[3760]	Speed: 35.040770 samples/sec	 accuracy=58.519068	 loss=1.727709	 lr=0.010000
Epoch[021] Batch [2999]/[3760]	Speed: 35.345219 samples/sec	 accuracy=58.493750	 loss=1.728696	 lr=0.010000
Epoch[021] Batch [3049]/[3760]	Speed: 35.213703 samples/sec	 accuracy=58.480020	 loss=1.729065	 lr=0.010000
Epoch[021] Batch [3099]/[3760]	Speed: 35.402895 samples/sec	 accuracy=58.488407	 loss=1.729132	 lr=0.010000
Epoch[021] Batch [3149]/[3760]	Speed: 35.367114 samples/sec	 accuracy=58.503472	 loss=1.729191	 lr=0.010000
Epoch[021] Batch [3199]/[3760]	Speed: 35.236069 samples/sec	 accuracy=58.458008	 loss=1.730948	 lr=0.010000
Epoch[021] Batch [3249]/[3760]	Speed: 35.183195 samples/sec	 accuracy=58.467308	 loss=1.731018	 lr=0.010000
Epoch[021] Batch [3299]/[3760]	Speed: 35.198017 samples/sec	 accuracy=58.464015	 loss=1.731067	 lr=0.010000
Epoch[021] Batch [3349]/[3760]	Speed: 35.320095 samples/sec	 accuracy=58.471082	 loss=1.730893	 lr=0.010000
Epoch[021] Batch [3399]/[3760]	Speed: 34.976213 samples/sec	 accuracy=58.484375	 loss=1.730407	 lr=0.010000
Epoch[021] Batch [3449]/[3760]	Speed: 35.264115 samples/sec	 accuracy=58.484149	 loss=1.730480	 lr=0.010000
Epoch[021] Batch [3499]/[3760]	Speed: 34.741604 samples/sec	 accuracy=58.486161	 loss=1.730840	 lr=0.010000
Epoch[021] Batch [3549]/[3760]	Speed: 35.484019 samples/sec	 accuracy=58.487676	 loss=1.730103	 lr=0.010000
Epoch[021] Batch [3599]/[3760]	Speed: 34.819128 samples/sec	 accuracy=58.467448	 loss=1.730628	 lr=0.010000
Epoch[021] Batch [3649]/[3760]	Speed: 35.326934 samples/sec	 accuracy=58.450342	 loss=1.731247	 lr=0.010000
Epoch[021] Batch [3699]/[3760]	Speed: 34.985687 samples/sec	 accuracy=58.461149	 loss=1.731107	 lr=0.010000
Epoch[021] Batch [3749]/[3760]	Speed: 36.551387 samples/sec	 accuracy=58.452500	 loss=1.731425	 lr=0.010000
Batch [0049]/[0303]: acc-top1=57.125000 acc-top5=80.343750
Batch [0099]/[0303]: acc-top1=56.765625 acc-top5=80.468750
Batch [0149]/[0303]: acc-top1=57.437500 acc-top5=80.833333
Batch [0199]/[0303]: acc-top1=57.273438 acc-top5=80.718750
Batch [0249]/[0303]: acc-top1=57.125000 acc-top5=80.443750
Batch [0299]/[0303]: acc-top1=57.041667 acc-top5=80.317708
[Epoch 021] training: accuracy=58.455785	 loss=1.731442
[Epoch 021] speed: 34 samples/sec	time cost: 7380.408786
[Epoch 021] validation: acc-top1=57.018358 acc-top5=80.321782 loss=2.104842
Epoch[022] Batch [0049]/[3760]	Speed: 24.848701 samples/sec	 accuracy=58.687500	 loss=1.664998	 lr=0.010000
Epoch[022] Batch [0099]/[3760]	Speed: 35.145982 samples/sec	 accuracy=59.109375	 loss=1.659685	 lr=0.010000
Epoch[022] Batch [0149]/[3760]	Speed: 35.094855 samples/sec	 accuracy=58.958333	 loss=1.681652	 lr=0.010000
Epoch[022] Batch [0199]/[3760]	Speed: 35.130798 samples/sec	 accuracy=59.304688	 loss=1.676049	 lr=0.010000
Epoch[022] Batch [0249]/[3760]	Speed: 35.275181 samples/sec	 accuracy=59.375000	 loss=1.671020	 lr=0.010000
Epoch[022] Batch [0299]/[3760]	Speed: 35.056037 samples/sec	 accuracy=59.364583	 loss=1.673167	 lr=0.010000
Epoch[022] Batch [0349]/[3760]	Speed: 35.269675 samples/sec	 accuracy=59.513393	 loss=1.672137	 lr=0.010000
Epoch[022] Batch [0399]/[3760]	Speed: 34.777373 samples/sec	 accuracy=59.488281	 loss=1.677113	 lr=0.010000
Epoch[022] Batch [0449]/[3760]	Speed: 35.405921 samples/sec	 accuracy=59.562500	 loss=1.675918	 lr=0.010000
Epoch[022] Batch [0499]/[3760]	Speed: 34.866220 samples/sec	 accuracy=59.631250	 loss=1.672216	 lr=0.010000
Epoch[022] Batch [0549]/[3760]	Speed: 35.406411 samples/sec	 accuracy=59.630682	 loss=1.676487	 lr=0.010000
Epoch[022] Batch [0599]/[3760]	Speed: 35.059981 samples/sec	 accuracy=59.697917	 loss=1.675685	 lr=0.010000
Epoch[022] Batch [0649]/[3760]	Speed: 35.352024 samples/sec	 accuracy=59.658654	 loss=1.675098	 lr=0.010000
Epoch[022] Batch [0699]/[3760]	Speed: 35.246828 samples/sec	 accuracy=59.698661	 loss=1.676698	 lr=0.010000
Epoch[022] Batch [0749]/[3760]	Speed: 35.195039 samples/sec	 accuracy=59.810417	 loss=1.672491	 lr=0.010000
Epoch[022] Batch [0799]/[3760]	Speed: 35.305514 samples/sec	 accuracy=59.730469	 loss=1.674893	 lr=0.010000
Epoch[022] Batch [0849]/[3760]	Speed: 35.211573 samples/sec	 accuracy=59.729779	 loss=1.676492	 lr=0.010000
Epoch[022] Batch [0899]/[3760]	Speed: 35.335938 samples/sec	 accuracy=59.701389	 loss=1.677905	 lr=0.010000
Epoch[022] Batch [0949]/[3760]	Speed: 35.458336 samples/sec	 accuracy=59.613487	 loss=1.680886	 lr=0.010000
Epoch[022] Batch [0999]/[3760]	Speed: 35.302946 samples/sec	 accuracy=59.623438	 loss=1.679315	 lr=0.010000
Epoch[022] Batch [1049]/[3760]	Speed: 35.483481 samples/sec	 accuracy=59.625000	 loss=1.677965	 lr=0.010000
Epoch[022] Batch [1099]/[3760]	Speed: 35.172988 samples/sec	 accuracy=59.600852	 loss=1.679055	 lr=0.010000
Epoch[022] Batch [1149]/[3760]	Speed: 35.525794 samples/sec	 accuracy=59.565217	 loss=1.680050	 lr=0.010000
Epoch[022] Batch [1199]/[3760]	Speed: 35.102178 samples/sec	 accuracy=59.531250	 loss=1.682062	 lr=0.010000
Epoch[022] Batch [1249]/[3760]	Speed: 35.535804 samples/sec	 accuracy=59.453750	 loss=1.684719	 lr=0.010000
Epoch[022] Batch [1299]/[3760]	Speed: 35.268115 samples/sec	 accuracy=59.408654	 loss=1.684747	 lr=0.010000
Epoch[022] Batch [1349]/[3760]	Speed: 35.319378 samples/sec	 accuracy=59.364583	 loss=1.687369	 lr=0.010000
Epoch[022] Batch [1399]/[3760]	Speed: 35.289971 samples/sec	 accuracy=59.339286	 loss=1.687935	 lr=0.010000
Epoch[022] Batch [1449]/[3760]	Speed: 35.376766 samples/sec	 accuracy=59.293103	 loss=1.689341	 lr=0.010000
Epoch[022] Batch [1499]/[3760]	Speed: 35.246868 samples/sec	 accuracy=59.261458	 loss=1.689901	 lr=0.010000
Epoch[022] Batch [1549]/[3760]	Speed: 35.198147 samples/sec	 accuracy=59.237903	 loss=1.691125	 lr=0.010000
Epoch[022] Batch [1599]/[3760]	Speed: 35.472881 samples/sec	 accuracy=59.241211	 loss=1.691889	 lr=0.010000
Epoch[022] Batch [1649]/[3760]	Speed: 35.177093 samples/sec	 accuracy=59.236742	 loss=1.693194	 lr=0.010000
Epoch[022] Batch [1699]/[3760]	Speed: 35.404996 samples/sec	 accuracy=59.258272	 loss=1.693740	 lr=0.010000
Epoch[022] Batch [1749]/[3760]	Speed: 35.318267 samples/sec	 accuracy=59.266964	 loss=1.693211	 lr=0.010000
Epoch[022] Batch [1799]/[3760]	Speed: 35.455945 samples/sec	 accuracy=59.237847	 loss=1.694793	 lr=0.010000
Epoch[022] Batch [1849]/[3760]	Speed: 35.350448 samples/sec	 accuracy=59.240709	 loss=1.695033	 lr=0.010000
Epoch[022] Batch [1899]/[3760]	Speed: 35.279832 samples/sec	 accuracy=59.230263	 loss=1.694810	 lr=0.010000
Epoch[022] Batch [1949]/[3760]	Speed: 35.135638 samples/sec	 accuracy=59.243590	 loss=1.695233	 lr=0.010000
Epoch[022] Batch [1999]/[3760]	Speed: 35.333139 samples/sec	 accuracy=59.239844	 loss=1.695141	 lr=0.010000
Epoch[022] Batch [2049]/[3760]	Speed: 35.153092 samples/sec	 accuracy=59.228659	 loss=1.695627	 lr=0.010000
Epoch[022] Batch [2099]/[3760]	Speed: 35.323325 samples/sec	 accuracy=59.209821	 loss=1.696417	 lr=0.010000
Epoch[022] Batch [2149]/[3760]	Speed: 35.213209 samples/sec	 accuracy=59.198401	 loss=1.696757	 lr=0.010000
Epoch[022] Batch [2199]/[3760]	Speed: 35.399254 samples/sec	 accuracy=59.208097	 loss=1.696991	 lr=0.010000
Epoch[022] Batch [2249]/[3760]	Speed: 34.915230 samples/sec	 accuracy=59.195139	 loss=1.697268	 lr=0.010000
Epoch[022] Batch [2299]/[3760]	Speed: 35.325342 samples/sec	 accuracy=59.208560	 loss=1.697534	 lr=0.010000
Epoch[022] Batch [2349]/[3760]	Speed: 34.678414 samples/sec	 accuracy=59.212101	 loss=1.697436	 lr=0.010000
Epoch[022] Batch [2399]/[3760]	Speed: 35.290800 samples/sec	 accuracy=59.207682	 loss=1.698017	 lr=0.010000
Epoch[022] Batch [2449]/[3760]	Speed: 34.403775 samples/sec	 accuracy=59.185587	 loss=1.698620	 lr=0.010000
Epoch[022] Batch [2499]/[3760]	Speed: 35.158223 samples/sec	 accuracy=59.173125	 loss=1.699446	 lr=0.010000
Epoch[022] Batch [2549]/[3760]	Speed: 34.776059 samples/sec	 accuracy=59.179534	 loss=1.699378	 lr=0.010000
Epoch[022] Batch [2599]/[3760]	Speed: 35.139489 samples/sec	 accuracy=59.131611	 loss=1.700361	 lr=0.010000
Epoch[022] Batch [2649]/[3760]	Speed: 34.885178 samples/sec	 accuracy=59.129127	 loss=1.700200	 lr=0.010000
Epoch[022] Batch [2699]/[3760]	Speed: 34.939426 samples/sec	 accuracy=59.106481	 loss=1.701277	 lr=0.010000
Epoch[022] Batch [2749]/[3760]	Speed: 35.109087 samples/sec	 accuracy=59.115341	 loss=1.701227	 lr=0.010000
Epoch[022] Batch [2799]/[3760]	Speed: 34.949573 samples/sec	 accuracy=59.106027	 loss=1.701859	 lr=0.010000
Epoch[022] Batch [2849]/[3760]	Speed: 35.567067 samples/sec	 accuracy=59.065241	 loss=1.702986	 lr=0.010000
Epoch[022] Batch [2899]/[3760]	Speed: 34.925024 samples/sec	 accuracy=59.037177	 loss=1.704115	 lr=0.010000
Epoch[022] Batch [2949]/[3760]	Speed: 35.287478 samples/sec	 accuracy=59.034428	 loss=1.703968	 lr=0.010000
Epoch[022] Batch [2999]/[3760]	Speed: 34.771428 samples/sec	 accuracy=58.991146	 loss=1.705604	 lr=0.010000
Epoch[022] Batch [3049]/[3760]	Speed: 35.368569 samples/sec	 accuracy=59.001025	 loss=1.705874	 lr=0.010000
Epoch[022] Batch [3099]/[3760]	Speed: 34.833049 samples/sec	 accuracy=58.997480	 loss=1.705891	 lr=0.010000
Epoch[022] Batch [3149]/[3760]	Speed: 35.241315 samples/sec	 accuracy=58.997520	 loss=1.706292	 lr=0.010000
Epoch[022] Batch [3199]/[3760]	Speed: 35.065505 samples/sec	 accuracy=58.985840	 loss=1.706982	 lr=0.010000
Epoch[022] Batch [3249]/[3760]	Speed: 35.304462 samples/sec	 accuracy=59.015865	 loss=1.705882	 lr=0.010000
Epoch[022] Batch [3299]/[3760]	Speed: 35.216238 samples/sec	 accuracy=58.996686	 loss=1.706587	 lr=0.010000
Epoch[022] Batch [3349]/[3760]	Speed: 35.048893 samples/sec	 accuracy=58.982276	 loss=1.707798	 lr=0.010000
Epoch[022] Batch [3399]/[3760]	Speed: 35.274346 samples/sec	 accuracy=59.006893	 loss=1.707350	 lr=0.010000
Epoch[022] Batch [3449]/[3760]	Speed: 34.907161 samples/sec	 accuracy=58.983243	 loss=1.708350	 lr=0.010000
Epoch[022] Batch [3499]/[3760]	Speed: 35.298285 samples/sec	 accuracy=58.981696	 loss=1.708874	 lr=0.010000
Epoch[022] Batch [3549]/[3760]	Speed: 34.836672 samples/sec	 accuracy=58.963028	 loss=1.709162	 lr=0.010000
Epoch[022] Batch [3599]/[3760]	Speed: 35.490282 samples/sec	 accuracy=58.964844	 loss=1.708988	 lr=0.010000
Epoch[022] Batch [3649]/[3760]	Speed: 34.699869 samples/sec	 accuracy=58.958476	 loss=1.709588	 lr=0.010000
Epoch[022] Batch [3699]/[3760]	Speed: 35.428399 samples/sec	 accuracy=58.937500	 loss=1.710156	 lr=0.010000
Epoch[022] Batch [3749]/[3760]	Speed: 36.402291 samples/sec	 accuracy=58.919583	 loss=1.710947	 lr=0.010000
Batch [0049]/[0303]: acc-top1=57.593750 acc-top5=80.531250
Batch [0099]/[0303]: acc-top1=57.859375 acc-top5=80.328125
Batch [0149]/[0303]: acc-top1=58.197917 acc-top5=80.593750
Batch [0199]/[0303]: acc-top1=58.093750 acc-top5=80.734375
Batch [0249]/[0303]: acc-top1=57.800000 acc-top5=80.625000
Batch [0299]/[0303]: acc-top1=57.609375 acc-top5=80.630208
[Epoch 022] training: accuracy=58.916223	 loss=1.711212
[Epoch 022] speed: 35 samples/sec	time cost: 7351.581604
[Epoch 022] validation: acc-top1=57.606229 acc-top5=80.620875 loss=2.029137
Epoch[023] Batch [0049]/[3759]	Speed: 24.140581 samples/sec	 accuracy=61.718750	 loss=1.612208	 lr=0.010000
Epoch[023] Batch [0099]/[3759]	Speed: 35.142339 samples/sec	 accuracy=60.750000	 loss=1.631152	 lr=0.010000
Epoch[023] Batch [0149]/[3759]	Speed: 34.959462 samples/sec	 accuracy=60.281250	 loss=1.658207	 lr=0.010000
Epoch[023] Batch [0199]/[3759]	Speed: 35.023702 samples/sec	 accuracy=60.218750	 loss=1.663236	 lr=0.010000
Epoch[023] Batch [0249]/[3759]	Speed: 35.333163 samples/sec	 accuracy=60.543750	 loss=1.649459	 lr=0.010000
Epoch[023] Batch [0299]/[3759]	Speed: 34.804259 samples/sec	 accuracy=60.453125	 loss=1.651093	 lr=0.010000
Epoch[023] Batch [0349]/[3759]	Speed: 35.237992 samples/sec	 accuracy=60.312500	 loss=1.649447	 lr=0.010000
Epoch[023] Batch [0399]/[3759]	Speed: 34.960033 samples/sec	 accuracy=60.234375	 loss=1.654070	 lr=0.010000
Epoch[023] Batch [0449]/[3759]	Speed: 35.551652 samples/sec	 accuracy=60.270833	 loss=1.649487	 lr=0.010000
Epoch[023] Batch [0499]/[3759]	Speed: 34.976893 samples/sec	 accuracy=60.068750	 loss=1.661850	 lr=0.010000
Epoch[023] Batch [0549]/[3759]	Speed: 35.534726 samples/sec	 accuracy=59.968750	 loss=1.667264	 lr=0.010000
Epoch[023] Batch [0599]/[3759]	Speed: 35.152531 samples/sec	 accuracy=60.221354	 loss=1.662087	 lr=0.010000
Epoch[023] Batch [0649]/[3759]	Speed: 35.417150 samples/sec	 accuracy=60.149038	 loss=1.662746	 lr=0.010000
Epoch[023] Batch [0699]/[3759]	Speed: 35.156892 samples/sec	 accuracy=60.087054	 loss=1.664971	 lr=0.010000
Epoch[023] Batch [0749]/[3759]	Speed: 35.386520 samples/sec	 accuracy=59.958333	 loss=1.670259	 lr=0.010000
Epoch[023] Batch [0799]/[3759]	Speed: 35.357085 samples/sec	 accuracy=59.900391	 loss=1.671565	 lr=0.010000
Epoch[023] Batch [0849]/[3759]	Speed: 35.316958 samples/sec	 accuracy=59.794118	 loss=1.672101	 lr=0.010000
Epoch[023] Batch [0899]/[3759]	Speed: 35.368689 samples/sec	 accuracy=59.784722	 loss=1.670708	 lr=0.010000
Epoch[023] Batch [0949]/[3759]	Speed: 35.122224 samples/sec	 accuracy=59.845395	 loss=1.669245	 lr=0.010000
Epoch[023] Batch [0999]/[3759]	Speed: 35.383506 samples/sec	 accuracy=59.754688	 loss=1.674145	 lr=0.010000
Epoch[023] Batch [1049]/[3759]	Speed: 35.311136 samples/sec	 accuracy=59.657738	 loss=1.679084	 lr=0.010000
Epoch[023] Batch [1099]/[3759]	Speed: 35.391561 samples/sec	 accuracy=59.684659	 loss=1.676480	 lr=0.010000
Epoch[023] Batch [1149]/[3759]	Speed: 35.144426 samples/sec	 accuracy=59.625000	 loss=1.678367	 lr=0.010000
Epoch[023] Batch [1199]/[3759]	Speed: 35.531743 samples/sec	 accuracy=59.630208	 loss=1.678893	 lr=0.010000
Epoch[023] Batch [1249]/[3759]	Speed: 35.082239 samples/sec	 accuracy=59.578750	 loss=1.682468	 lr=0.010000
Epoch[023] Batch [1299]/[3759]	Speed: 35.552154 samples/sec	 accuracy=59.617788	 loss=1.680870	 lr=0.010000
Epoch[023] Batch [1349]/[3759]	Speed: 35.040742 samples/sec	 accuracy=59.665509	 loss=1.678181	 lr=0.010000
Epoch[023] Batch [1399]/[3759]	Speed: 35.259127 samples/sec	 accuracy=59.617188	 loss=1.680806	 lr=0.010000
Epoch[023] Batch [1449]/[3759]	Speed: 34.975431 samples/sec	 accuracy=59.591595	 loss=1.682138	 lr=0.010000
Epoch[023] Batch [1499]/[3759]	Speed: 35.444094 samples/sec	 accuracy=59.554167	 loss=1.684286	 lr=0.010000
Epoch[023] Batch [1549]/[3759]	Speed: 35.267136 samples/sec	 accuracy=59.559476	 loss=1.684586	 lr=0.010000
Epoch[023] Batch [1599]/[3759]	Speed: 35.397791 samples/sec	 accuracy=59.542969	 loss=1.685953	 lr=0.010000
Epoch[023] Batch [1649]/[3759]	Speed: 35.214845 samples/sec	 accuracy=59.517045	 loss=1.688166	 lr=0.010000
Epoch[023] Batch [1699]/[3759]	Speed: 35.471593 samples/sec	 accuracy=59.491728	 loss=1.689168	 lr=0.010000
Epoch[023] Batch [1749]/[3759]	Speed: 35.060001 samples/sec	 accuracy=59.454464	 loss=1.690398	 lr=0.010000
Epoch[023] Batch [1799]/[3759]	Speed: 35.423883 samples/sec	 accuracy=59.432292	 loss=1.691449	 lr=0.010000
Epoch[023] Batch [1849]/[3759]	Speed: 35.342923 samples/sec	 accuracy=59.388514	 loss=1.692777	 lr=0.010000
Epoch[023] Batch [1899]/[3759]	Speed: 35.262870 samples/sec	 accuracy=59.398849	 loss=1.693510	 lr=0.010000
Epoch[023] Batch [1949]/[3759]	Speed: 35.290733 samples/sec	 accuracy=59.400641	 loss=1.694189	 lr=0.010000
Epoch[023] Batch [1999]/[3759]	Speed: 35.260964 samples/sec	 accuracy=59.432031	 loss=1.692323	 lr=0.010000
Epoch[023] Batch [2049]/[3759]	Speed: 35.333486 samples/sec	 accuracy=59.423018	 loss=1.692383	 lr=0.010000
Epoch[023] Batch [2099]/[3759]	Speed: 35.177419 samples/sec	 accuracy=59.425595	 loss=1.692069	 lr=0.010000
Epoch[023] Batch [2149]/[3759]	Speed: 34.946761 samples/sec	 accuracy=59.428779	 loss=1.692071	 lr=0.010000
Epoch[023] Batch [2199]/[3759]	Speed: 35.438973 samples/sec	 accuracy=59.425426	 loss=1.692767	 lr=0.010000
Epoch[023] Batch [2249]/[3759]	Speed: 34.889630 samples/sec	 accuracy=59.418056	 loss=1.693641	 lr=0.010000
Epoch[023] Batch [2299]/[3759]	Speed: 35.422820 samples/sec	 accuracy=59.431386	 loss=1.693706	 lr=0.010000
Epoch[023] Batch [2349]/[3759]	Speed: 34.912956 samples/sec	 accuracy=59.426197	 loss=1.694318	 lr=0.010000
Epoch[023] Batch [2399]/[3759]	Speed: 35.406970 samples/sec	 accuracy=59.442708	 loss=1.693207	 lr=0.010000
Epoch[023] Batch [2449]/[3759]	Speed: 34.803845 samples/sec	 accuracy=59.414541	 loss=1.694622	 lr=0.010000
Epoch[023] Batch [2499]/[3759]	Speed: 35.500613 samples/sec	 accuracy=59.415625	 loss=1.694715	 lr=0.010000
Epoch[023] Batch [2549]/[3759]	Speed: 35.135678 samples/sec	 accuracy=59.402574	 loss=1.694762	 lr=0.010000
Epoch[023] Batch [2599]/[3759]	Speed: 35.294541 samples/sec	 accuracy=59.378606	 loss=1.695346	 lr=0.010000
Epoch[023] Batch [2649]/[3759]	Speed: 34.911969 samples/sec	 accuracy=59.382665	 loss=1.696092	 lr=0.010000
Epoch[023] Batch [2699]/[3759]	Speed: 35.174632 samples/sec	 accuracy=59.393519	 loss=1.696435	 lr=0.010000
Epoch[023] Batch [2749]/[3759]	Speed: 35.045457 samples/sec	 accuracy=59.398295	 loss=1.696930	 lr=0.010000
Epoch[023] Batch [2799]/[3759]	Speed: 34.774403 samples/sec	 accuracy=59.373884	 loss=1.697504	 lr=0.010000
Epoch[023] Batch [2849]/[3759]	Speed: 35.222880 samples/sec	 accuracy=59.341009	 loss=1.698762	 lr=0.010000
Epoch[023] Batch [2899]/[3759]	Speed: 34.659950 samples/sec	 accuracy=59.321121	 loss=1.699137	 lr=0.010000
Epoch[023] Batch [2949]/[3759]	Speed: 35.352369 samples/sec	 accuracy=59.327331	 loss=1.698715	 lr=0.010000
Epoch[023] Batch [2999]/[3759]	Speed: 34.822901 samples/sec	 accuracy=59.305729	 loss=1.699073	 lr=0.010000
Epoch[023] Batch [3049]/[3759]	Speed: 35.178722 samples/sec	 accuracy=59.303791	 loss=1.698906	 lr=0.010000
Epoch[023] Batch [3099]/[3759]	Speed: 35.001522 samples/sec	 accuracy=59.286794	 loss=1.699666	 lr=0.010000
Epoch[023] Batch [3149]/[3759]	Speed: 35.109833 samples/sec	 accuracy=59.263393	 loss=1.700767	 lr=0.010000
Epoch[023] Batch [3199]/[3759]	Speed: 34.887997 samples/sec	 accuracy=59.239746	 loss=1.702008	 lr=0.010000
Epoch[023] Batch [3249]/[3759]	Speed: 34.658382 samples/sec	 accuracy=59.258654	 loss=1.700845	 lr=0.010000
Epoch[023] Batch [3299]/[3759]	Speed: 35.355676 samples/sec	 accuracy=59.278409	 loss=1.700248	 lr=0.010000
Epoch[023] Batch [3349]/[3759]	Speed: 34.821074 samples/sec	 accuracy=59.276586	 loss=1.700247	 lr=0.010000
Epoch[023] Batch [3399]/[3759]	Speed: 35.478955 samples/sec	 accuracy=59.280790	 loss=1.699256	 lr=0.010000
Epoch[023] Batch [3449]/[3759]	Speed: 34.971266 samples/sec	 accuracy=59.259511	 loss=1.700799	 lr=0.010000
Epoch[023] Batch [3499]/[3759]	Speed: 35.437147 samples/sec	 accuracy=59.253571	 loss=1.700591	 lr=0.010000
Epoch[023] Batch [3549]/[3759]	Speed: 35.141591 samples/sec	 accuracy=59.230194	 loss=1.701691	 lr=0.010000
Epoch[023] Batch [3599]/[3759]	Speed: 35.468759 samples/sec	 accuracy=59.226997	 loss=1.702082	 lr=0.010000
Epoch[023] Batch [3649]/[3759]	Speed: 35.062116 samples/sec	 accuracy=59.227740	 loss=1.701926	 lr=0.010000
Epoch[023] Batch [3699]/[3759]	Speed: 35.361319 samples/sec	 accuracy=59.206081	 loss=1.703085	 lr=0.010000
Epoch[023] Batch [3749]/[3759]	Speed: 36.497967 samples/sec	 accuracy=59.177917	 loss=1.704831	 lr=0.010000
Batch [0049]/[0303]: acc-top1=56.031250 acc-top5=79.250000
Batch [0099]/[0303]: acc-top1=56.640625 acc-top5=79.703125
Batch [0149]/[0303]: acc-top1=56.895833 acc-top5=80.072917
Batch [0199]/[0303]: acc-top1=56.765625 acc-top5=80.179688
Batch [0249]/[0303]: acc-top1=56.737500 acc-top5=80.012500
Batch [0299]/[0303]: acc-top1=56.609375 acc-top5=79.932292
[Epoch 023] training: accuracy=59.170906	 loss=1.705050
[Epoch 023] speed: 34 samples/sec	time cost: 7353.273828
[Epoch 023] validation: acc-top1=56.580033 acc-top5=79.914398 loss=2.035718
Epoch[024] Batch [0049]/[3760]	Speed: 24.204983 samples/sec	 accuracy=60.968750	 loss=1.628828	 lr=0.010000
Epoch[024] Batch [0099]/[3760]	Speed: 35.110885 samples/sec	 accuracy=60.531250	 loss=1.656289	 lr=0.010000
Epoch[024] Batch [0149]/[3760]	Speed: 35.054314 samples/sec	 accuracy=59.822917	 loss=1.674697	 lr=0.010000
Epoch[024] Batch [0199]/[3760]	Speed: 35.039702 samples/sec	 accuracy=60.062500	 loss=1.662083	 lr=0.010000
Epoch[024] Batch [0249]/[3760]	Speed: 35.381897 samples/sec	 accuracy=59.887500	 loss=1.670475	 lr=0.010000
Epoch[024] Batch [0299]/[3760]	Speed: 34.754379 samples/sec	 accuracy=59.682292	 loss=1.680243	 lr=0.010000
Epoch[024] Batch [0349]/[3760]	Speed: 35.279066 samples/sec	 accuracy=59.598214	 loss=1.681141	 lr=0.010000
Epoch[024] Batch [0399]/[3760]	Speed: 35.005411 samples/sec	 accuracy=59.625000	 loss=1.681113	 lr=0.010000
Epoch[024] Batch [0449]/[3760]	Speed: 35.351724 samples/sec	 accuracy=59.614583	 loss=1.679374	 lr=0.010000
Epoch[024] Batch [0499]/[3760]	Speed: 34.879586 samples/sec	 accuracy=59.800000	 loss=1.671582	 lr=0.010000
Epoch[024] Batch [0549]/[3760]	Speed: 35.431354 samples/sec	 accuracy=59.784091	 loss=1.671101	 lr=0.010000
Epoch[024] Batch [0599]/[3760]	Speed: 35.173029 samples/sec	 accuracy=59.815104	 loss=1.669403	 lr=0.010000
Epoch[024] Batch [0649]/[3760]	Speed: 35.295695 samples/sec	 accuracy=59.701923	 loss=1.675115	 lr=0.010000
Epoch[024] Batch [0699]/[3760]	Speed: 35.282611 samples/sec	 accuracy=59.720982	 loss=1.674163	 lr=0.010000
Epoch[024] Batch [0749]/[3760]	Speed: 35.099788 samples/sec	 accuracy=59.702083	 loss=1.671862	 lr=0.010000
Epoch[024] Batch [0799]/[3760]	Speed: 35.196308 samples/sec	 accuracy=59.677734	 loss=1.672929	 lr=0.010000
Epoch[024] Batch [0849]/[3760]	Speed: 35.011540 samples/sec	 accuracy=59.740809	 loss=1.671401	 lr=0.010000
Epoch[024] Batch [0899]/[3760]	Speed: 35.328580 samples/sec	 accuracy=59.730903	 loss=1.670660	 lr=0.010000
Epoch[024] Batch [0949]/[3760]	Speed: 35.067192 samples/sec	 accuracy=59.717105	 loss=1.672275	 lr=0.010000
Epoch[024] Batch [0999]/[3760]	Speed: 35.147034 samples/sec	 accuracy=59.748437	 loss=1.671515	 lr=0.010000
Epoch[024] Batch [1049]/[3760]	Speed: 34.915151 samples/sec	 accuracy=59.803571	 loss=1.670419	 lr=0.010000
Epoch[024] Batch [1099]/[3760]	Speed: 35.434841 samples/sec	 accuracy=59.767045	 loss=1.674243	 lr=0.010000
Epoch[024] Batch [1149]/[3760]	Speed: 35.086222 samples/sec	 accuracy=59.722826	 loss=1.677136	 lr=0.010000
Epoch[024] Batch [1199]/[3760]	Speed: 35.317231 samples/sec	 accuracy=59.701823	 loss=1.676643	 lr=0.010000
Epoch[024] Batch [1249]/[3760]	Speed: 35.196003 samples/sec	 accuracy=59.663750	 loss=1.677961	 lr=0.010000
Epoch[024] Batch [1299]/[3760]	Speed: 35.281025 samples/sec	 accuracy=59.655048	 loss=1.678325	 lr=0.010000
Epoch[024] Batch [1349]/[3760]	Speed: 35.199226 samples/sec	 accuracy=59.670139	 loss=1.678569	 lr=0.010000
Epoch[024] Batch [1399]/[3760]	Speed: 35.041298 samples/sec	 accuracy=59.704241	 loss=1.677307	 lr=0.010000
Epoch[024] Batch [1449]/[3760]	Speed: 35.147962 samples/sec	 accuracy=59.693966	 loss=1.678385	 lr=0.010000
Epoch[024] Batch [1499]/[3760]	Speed: 34.936191 samples/sec	 accuracy=59.652083	 loss=1.680500	 lr=0.010000
Epoch[024] Batch [1549]/[3760]	Speed: 35.183223 samples/sec	 accuracy=59.640121	 loss=1.681661	 lr=0.010000
Epoch[024] Batch [1599]/[3760]	Speed: 35.038944 samples/sec	 accuracy=59.628906	 loss=1.681980	 lr=0.010000
Epoch[024] Batch [1649]/[3760]	Speed: 35.232775 samples/sec	 accuracy=59.608902	 loss=1.684757	 lr=0.010000
Epoch[024] Batch [1699]/[3760]	Speed: 35.017168 samples/sec	 accuracy=59.569853	 loss=1.685220	 lr=0.010000
Epoch[024] Batch [1749]/[3760]	Speed: 35.346068 samples/sec	 accuracy=59.525000	 loss=1.687069	 lr=0.010000
Epoch[024] Batch [1799]/[3760]	Speed: 34.850314 samples/sec	 accuracy=59.555556	 loss=1.687486	 lr=0.010000
Epoch[024] Batch [1849]/[3760]	Speed: 35.509672 samples/sec	 accuracy=59.529561	 loss=1.687995	 lr=0.010000
Epoch[024] Batch [1899]/[3760]	Speed: 35.026548 samples/sec	 accuracy=59.542763	 loss=1.687936	 lr=0.010000
Epoch[024] Batch [1949]/[3760]	Speed: 35.434818 samples/sec	 accuracy=59.556090	 loss=1.688715	 lr=0.010000
Epoch[024] Batch [1999]/[3760]	Speed: 35.319546 samples/sec	 accuracy=59.561719	 loss=1.688873	 lr=0.010000
Epoch[024] Batch [2049]/[3760]	Speed: 35.347385 samples/sec	 accuracy=59.527439	 loss=1.689362	 lr=0.010000
Epoch[024] Batch [2099]/[3760]	Speed: 35.185529 samples/sec	 accuracy=59.527530	 loss=1.689159	 lr=0.010000
Epoch[024] Batch [2149]/[3760]	Speed: 35.352978 samples/sec	 accuracy=59.521802	 loss=1.689084	 lr=0.010000
Epoch[024] Batch [2199]/[3760]	Speed: 35.134026 samples/sec	 accuracy=59.519886	 loss=1.689087	 lr=0.010000
Epoch[024] Batch [2249]/[3760]	Speed: 35.383702 samples/sec	 accuracy=59.536111	 loss=1.687469	 lr=0.010000
Epoch[024] Batch [2299]/[3760]	Speed: 35.160470 samples/sec	 accuracy=59.513587	 loss=1.687950	 lr=0.010000
Epoch[024] Batch [2349]/[3760]	Speed: 34.898669 samples/sec	 accuracy=59.474069	 loss=1.689328	 lr=0.010000
Epoch[024] Batch [2399]/[3760]	Speed: 35.245044 samples/sec	 accuracy=59.480469	 loss=1.689017	 lr=0.010000
Epoch[024] Batch [2449]/[3760]	Speed: 34.874098 samples/sec	 accuracy=59.465561	 loss=1.689895	 lr=0.010000
Epoch[024] Batch [2499]/[3760]	Speed: 35.373717 samples/sec	 accuracy=59.476250	 loss=1.689733	 lr=0.010000
Epoch[024] Batch [2549]/[3760]	Speed: 34.822669 samples/sec	 accuracy=59.441176	 loss=1.690947	 lr=0.010000
Epoch[024] Batch [2599]/[3760]	Speed: 35.323452 samples/sec	 accuracy=59.430889	 loss=1.691498	 lr=0.010000
Epoch[024] Batch [2649]/[3760]	Speed: 34.631698 samples/sec	 accuracy=59.432783	 loss=1.690797	 lr=0.010000
Epoch[024] Batch [2699]/[3760]	Speed: 35.124632 samples/sec	 accuracy=59.435185	 loss=1.690969	 lr=0.010000
Epoch[024] Batch [2749]/[3760]	Speed: 34.706125 samples/sec	 accuracy=59.431250	 loss=1.691996	 lr=0.010000
Epoch[024] Batch [2799]/[3760]	Speed: 35.001701 samples/sec	 accuracy=59.462054	 loss=1.691326	 lr=0.010000
Epoch[024] Batch [2849]/[3760]	Speed: 35.083069 samples/sec	 accuracy=59.478070	 loss=1.690814	 lr=0.010000
Epoch[024] Batch [2899]/[3760]	Speed: 34.810779 samples/sec	 accuracy=59.495690	 loss=1.689975	 lr=0.010000
Epoch[024] Batch [2949]/[3760]	Speed: 35.133007 samples/sec	 accuracy=59.488347	 loss=1.689865	 lr=0.010000
Epoch[024] Batch [2999]/[3760]	Speed: 34.806168 samples/sec	 accuracy=59.460417	 loss=1.690969	 lr=0.010000
Epoch[024] Batch [3049]/[3760]	Speed: 35.312054 samples/sec	 accuracy=59.486168	 loss=1.690220	 lr=0.010000
Epoch[024] Batch [3099]/[3760]	Speed: 34.757832 samples/sec	 accuracy=59.462198	 loss=1.690764	 lr=0.010000
Epoch[024] Batch [3149]/[3760]	Speed: 35.339944 samples/sec	 accuracy=59.475198	 loss=1.690574	 lr=0.010000
Epoch[024] Batch [3199]/[3760]	Speed: 34.729364 samples/sec	 accuracy=59.482910	 loss=1.690864	 lr=0.010000
Epoch[024] Batch [3249]/[3760]	Speed: 35.344467 samples/sec	 accuracy=59.471154	 loss=1.691969	 lr=0.010000
Epoch[024] Batch [3299]/[3760]	Speed: 35.104224 samples/sec	 accuracy=59.475379	 loss=1.692217	 lr=0.010000
Epoch[024] Batch [3349]/[3760]	Speed: 35.310755 samples/sec	 accuracy=59.434701	 loss=1.693990	 lr=0.010000
Epoch[024] Batch [3399]/[3760]	Speed: 34.935779 samples/sec	 accuracy=59.405331	 loss=1.694835	 lr=0.010000
Epoch[024] Batch [3449]/[3760]	Speed: 35.164931 samples/sec	 accuracy=59.380888	 loss=1.696525	 lr=0.010000
Epoch[024] Batch [3499]/[3760]	Speed: 35.348655 samples/sec	 accuracy=59.388393	 loss=1.696599	 lr=0.010000
Epoch[024] Batch [3549]/[3760]	Speed: 35.194531 samples/sec	 accuracy=59.393046	 loss=1.696435	 lr=0.010000
Epoch[024] Batch [3599]/[3760]	Speed: 35.337455 samples/sec	 accuracy=59.407552	 loss=1.695862	 lr=0.010000
Epoch[024] Batch [3649]/[3760]	Speed: 35.089087 samples/sec	 accuracy=59.412243	 loss=1.695473	 lr=0.010000
Epoch[024] Batch [3699]/[3760]	Speed: 35.443155 samples/sec	 accuracy=59.415963	 loss=1.695906	 lr=0.010000
Epoch[024] Batch [3749]/[3760]	Speed: 36.273840 samples/sec	 accuracy=59.410417	 loss=1.696194	 lr=0.010000
Batch [0049]/[0303]: acc-top1=57.593750 acc-top5=80.781250
Batch [0099]/[0303]: acc-top1=57.953125 acc-top5=80.703125
Batch [0149]/[0303]: acc-top1=58.250000 acc-top5=81.093750
Batch [0199]/[0303]: acc-top1=57.914062 acc-top5=80.921875
Batch [0249]/[0303]: acc-top1=57.693750 acc-top5=80.681250
Batch [0299]/[0303]: acc-top1=57.421875 acc-top5=80.645833
[Epoch 024] training: accuracy=59.416556	 loss=1.696063
[Epoch 024] speed: 34 samples/sec	time cost: 7364.165156
[Epoch 024] validation: acc-top1=57.363861 acc-top5=80.646658 loss=2.081856
Epoch[025] Batch [0049]/[3760]	Speed: 24.123616 samples/sec	 accuracy=60.218750	 loss=1.673349	 lr=0.010000
Epoch[025] Batch [0099]/[3760]	Speed: 35.031989 samples/sec	 accuracy=60.343750	 loss=1.635399	 lr=0.010000
Epoch[025] Batch [0149]/[3760]	Speed: 35.091047 samples/sec	 accuracy=60.864583	 loss=1.615462	 lr=0.010000
Epoch[025] Batch [0199]/[3760]	Speed: 35.233272 samples/sec	 accuracy=60.726562	 loss=1.621604	 lr=0.010000
Epoch[025] Batch [0249]/[3760]	Speed: 35.291954 samples/sec	 accuracy=60.750000	 loss=1.622849	 lr=0.010000
Epoch[025] Batch [0299]/[3760]	Speed: 35.207776 samples/sec	 accuracy=60.755208	 loss=1.630091	 lr=0.010000
Epoch[025] Batch [0349]/[3760]	Speed: 35.455362 samples/sec	 accuracy=60.558036	 loss=1.633631	 lr=0.010000
Epoch[025] Batch [0399]/[3760]	Speed: 34.949053 samples/sec	 accuracy=60.566406	 loss=1.635256	 lr=0.010000
Epoch[025] Batch [0449]/[3760]	Speed: 35.380360 samples/sec	 accuracy=60.659722	 loss=1.631390	 lr=0.010000
Epoch[025] Batch [0499]/[3760]	Speed: 35.061313 samples/sec	 accuracy=60.550000	 loss=1.634309	 lr=0.010000
Epoch[025] Batch [0549]/[3760]	Speed: 35.521192 samples/sec	 accuracy=60.457386	 loss=1.637300	 lr=0.010000
Epoch[025] Batch [0599]/[3760]	Speed: 35.138197 samples/sec	 accuracy=60.473958	 loss=1.638306	 lr=0.010000
Epoch[025] Batch [0649]/[3760]	Speed: 35.557308 samples/sec	 accuracy=60.480769	 loss=1.639842	 lr=0.010000
Epoch[025] Batch [0699]/[3760]	Speed: 35.264772 samples/sec	 accuracy=60.410714	 loss=1.642989	 lr=0.010000
Epoch[025] Batch [0749]/[3760]	Speed: 35.548931 samples/sec	 accuracy=60.402083	 loss=1.643339	 lr=0.010000
Epoch[025] Batch [0799]/[3760]	Speed: 35.294274 samples/sec	 accuracy=60.332031	 loss=1.646692	 lr=0.010000
Epoch[025] Batch [0849]/[3760]	Speed: 35.289233 samples/sec	 accuracy=60.303309	 loss=1.648479	 lr=0.010000
Epoch[025] Batch [0899]/[3760]	Speed: 35.260044 samples/sec	 accuracy=60.319444	 loss=1.649456	 lr=0.010000
Epoch[025] Batch [0949]/[3760]	Speed: 35.513654 samples/sec	 accuracy=60.226974	 loss=1.653816	 lr=0.010000
Epoch[025] Batch [0999]/[3760]	Speed: 34.980124 samples/sec	 accuracy=60.181250	 loss=1.656185	 lr=0.010000
Epoch[025] Batch [1049]/[3760]	Speed: 35.422079 samples/sec	 accuracy=60.181548	 loss=1.656683	 lr=0.010000
Epoch[025] Batch [1099]/[3760]	Speed: 34.817849 samples/sec	 accuracy=60.187500	 loss=1.653910	 lr=0.010000
Epoch[025] Batch [1149]/[3760]	Speed: 35.504031 samples/sec	 accuracy=60.167120	 loss=1.656519	 lr=0.010000
Epoch[025] Batch [1199]/[3760]	Speed: 34.690158 samples/sec	 accuracy=60.165365	 loss=1.656427	 lr=0.010000
Epoch[025] Batch [1249]/[3760]	Speed: 35.427853 samples/sec	 accuracy=60.156250	 loss=1.658277	 lr=0.010000
Epoch[025] Batch [1299]/[3760]	Speed: 34.545283 samples/sec	 accuracy=60.175481	 loss=1.658262	 lr=0.010000
Epoch[025] Batch [1349]/[3760]	Speed: 35.320427 samples/sec	 accuracy=60.145833	 loss=1.657925	 lr=0.010000
Epoch[025] Batch [1399]/[3760]	Speed: 35.047530 samples/sec	 accuracy=60.059152	 loss=1.661034	 lr=0.010000
Epoch[025] Batch [1449]/[3760]	Speed: 35.309551 samples/sec	 accuracy=60.024784	 loss=1.662576	 lr=0.010000
Epoch[025] Batch [1499]/[3760]	Speed: 35.116722 samples/sec	 accuracy=59.997917	 loss=1.663392	 lr=0.010000
Epoch[025] Batch [1549]/[3760]	Speed: 35.270412 samples/sec	 accuracy=59.951613	 loss=1.665756	 lr=0.010000
Epoch[025] Batch [1599]/[3760]	Speed: 35.247199 samples/sec	 accuracy=59.901367	 loss=1.668347	 lr=0.010000
Epoch[025] Batch [1649]/[3760]	Speed: 35.055509 samples/sec	 accuracy=59.908144	 loss=1.668268	 lr=0.010000
Epoch[025] Batch [1699]/[3760]	Speed: 35.471682 samples/sec	 accuracy=59.907169	 loss=1.668804	 lr=0.010000
Epoch[025] Batch [1749]/[3760]	Speed: 35.088604 samples/sec	 accuracy=59.902679	 loss=1.669581	 lr=0.010000
Epoch[025] Batch [1799]/[3760]	Speed: 35.719945 samples/sec	 accuracy=59.910590	 loss=1.668583	 lr=0.010000
Epoch[025] Batch [1849]/[3760]	Speed: 35.311726 samples/sec	 accuracy=59.917230	 loss=1.668607	 lr=0.010000
Epoch[025] Batch [1899]/[3760]	Speed: 35.515351 samples/sec	 accuracy=59.925164	 loss=1.668405	 lr=0.010000
Epoch[025] Batch [1949]/[3760]	Speed: 35.252988 samples/sec	 accuracy=59.921474	 loss=1.668278	 lr=0.010000
Epoch[025] Batch [1999]/[3760]	Speed: 35.203652 samples/sec	 accuracy=59.917969	 loss=1.668176	 lr=0.010000
Epoch[025] Batch [2049]/[3760]	Speed: 35.120176 samples/sec	 accuracy=59.904726	 loss=1.669463	 lr=0.010000
Epoch[025] Batch [2099]/[3760]	Speed: 35.131897 samples/sec	 accuracy=59.872024	 loss=1.669692	 lr=0.010000
Epoch[025] Batch [2149]/[3760]	Speed: 35.100977 samples/sec	 accuracy=59.883721	 loss=1.669639	 lr=0.010000
Epoch[025] Batch [2199]/[3760]	Speed: 34.917019 samples/sec	 accuracy=59.865767	 loss=1.670072	 lr=0.010000
Epoch[025] Batch [2249]/[3760]	Speed: 35.235537 samples/sec	 accuracy=59.849306	 loss=1.670830	 lr=0.010000
Epoch[025] Batch [2299]/[3760]	Speed: 34.893982 samples/sec	 accuracy=59.841033	 loss=1.670500	 lr=0.010000
Epoch[025] Batch [2349]/[3760]	Speed: 35.330080 samples/sec	 accuracy=59.813165	 loss=1.671228	 lr=0.010000
Epoch[025] Batch [2399]/[3760]	Speed: 34.719849 samples/sec	 accuracy=59.805990	 loss=1.671942	 lr=0.010000
Epoch[025] Batch [2449]/[3760]	Speed: 35.433685 samples/sec	 accuracy=59.811862	 loss=1.672807	 lr=0.010000
Epoch[025] Batch [2499]/[3760]	Speed: 34.953476 samples/sec	 accuracy=59.820625	 loss=1.672650	 lr=0.010000
Epoch[025] Batch [2549]/[3760]	Speed: 34.983057 samples/sec	 accuracy=59.857230	 loss=1.671968	 lr=0.010000
Epoch[025] Batch [2599]/[3760]	Speed: 34.542882 samples/sec	 accuracy=59.838942	 loss=1.672898	 lr=0.010000
Epoch[025] Batch [2649]/[3760]	Speed: 34.974452 samples/sec	 accuracy=59.826651	 loss=1.673060	 lr=0.010000
Epoch[025] Batch [2699]/[3760]	Speed: 35.025782 samples/sec	 accuracy=59.815972	 loss=1.673758	 lr=0.010000
Epoch[025] Batch [2749]/[3760]	Speed: 34.786626 samples/sec	 accuracy=59.795455	 loss=1.674886	 lr=0.010000
Epoch[025] Batch [2799]/[3760]	Speed: 35.331247 samples/sec	 accuracy=59.784040	 loss=1.675930	 lr=0.010000
Epoch[025] Batch [2849]/[3760]	Speed: 34.783587 samples/sec	 accuracy=59.808114	 loss=1.675388	 lr=0.010000
Epoch[025] Batch [2899]/[3760]	Speed: 35.390794 samples/sec	 accuracy=59.775862	 loss=1.676178	 lr=0.010000
Epoch[025] Batch [2949]/[3760]	Speed: 34.802015 samples/sec	 accuracy=59.765360	 loss=1.676823	 lr=0.010000
Epoch[025] Batch [2999]/[3760]	Speed: 35.277892 samples/sec	 accuracy=59.777083	 loss=1.676249	 lr=0.010000
Epoch[025] Batch [3049]/[3760]	Speed: 34.709168 samples/sec	 accuracy=59.786885	 loss=1.676031	 lr=0.010000
Epoch[025] Batch [3099]/[3760]	Speed: 35.179888 samples/sec	 accuracy=59.758569	 loss=1.676545	 lr=0.010000
Epoch[025] Batch [3149]/[3760]	Speed: 35.000948 samples/sec	 accuracy=59.757440	 loss=1.677249	 lr=0.010000
Epoch[025] Batch [3199]/[3760]	Speed: 35.062190 samples/sec	 accuracy=59.760742	 loss=1.676536	 lr=0.010000
Epoch[025] Batch [3249]/[3760]	Speed: 35.166698 samples/sec	 accuracy=59.752404	 loss=1.677187	 lr=0.010000
Epoch[025] Batch [3299]/[3760]	Speed: 35.176768 samples/sec	 accuracy=59.743371	 loss=1.677340	 lr=0.010000
Epoch[025] Batch [3349]/[3760]	Speed: 35.296271 samples/sec	 accuracy=59.754664	 loss=1.677897	 lr=0.010000
Epoch[025] Batch [3399]/[3760]	Speed: 35.118576 samples/sec	 accuracy=59.763787	 loss=1.677123	 lr=0.010000
Epoch[025] Batch [3449]/[3760]	Speed: 35.387909 samples/sec	 accuracy=59.772645	 loss=1.676717	 lr=0.010000
Epoch[025] Batch [3499]/[3760]	Speed: 35.005640 samples/sec	 accuracy=59.760714	 loss=1.676838	 lr=0.010000
Epoch[025] Batch [3549]/[3760]	Speed: 35.404254 samples/sec	 accuracy=59.757482	 loss=1.676908	 lr=0.010000
Epoch[025] Batch [3599]/[3760]	Speed: 34.841443 samples/sec	 accuracy=59.749566	 loss=1.677146	 lr=0.010000
Epoch[025] Batch [3649]/[3760]	Speed: 35.529397 samples/sec	 accuracy=59.735445	 loss=1.677627	 lr=0.010000
Epoch[025] Batch [3699]/[3760]	Speed: 35.156497 samples/sec	 accuracy=59.708193	 loss=1.678859	 lr=0.010000
Epoch[025] Batch [3749]/[3760]	Speed: 36.608539 samples/sec	 accuracy=59.722083	 loss=1.678534	 lr=0.010000
Batch [0049]/[0303]: acc-top1=57.812500 acc-top5=80.812500
Batch [0099]/[0303]: acc-top1=57.750000 acc-top5=80.343750
Batch [0149]/[0303]: acc-top1=58.010417 acc-top5=80.760417
Batch [0199]/[0303]: acc-top1=57.851562 acc-top5=80.898438
Batch [0249]/[0303]: acc-top1=57.562500 acc-top5=80.475000
Batch [0299]/[0303]: acc-top1=57.635417 acc-top5=80.385417
[Epoch 025] training: accuracy=59.724069	 loss=1.678356
[Epoch 025] speed: 34 samples/sec	time cost: 7359.314810
[Epoch 025] validation: acc-top1=57.616543 acc-top5=80.373350 loss=2.014491
Epoch[026] Batch [0049]/[3759]	Speed: 24.924136 samples/sec	 accuracy=60.218750	 loss=1.649137	 lr=0.010000
Epoch[026] Batch [0099]/[3759]	Speed: 34.965086 samples/sec	 accuracy=60.218750	 loss=1.653061	 lr=0.010000
Epoch[026] Batch [0149]/[3759]	Speed: 34.691589 samples/sec	 accuracy=60.395833	 loss=1.641636	 lr=0.010000
Epoch[026] Batch [0199]/[3759]	Speed: 34.851839 samples/sec	 accuracy=60.539062	 loss=1.636734	 lr=0.010000
Epoch[026] Batch [0249]/[3759]	Speed: 35.080374 samples/sec	 accuracy=60.606250	 loss=1.632509	 lr=0.010000
Epoch[026] Batch [0299]/[3759]	Speed: 34.667941 samples/sec	 accuracy=60.677083	 loss=1.634755	 lr=0.010000
Epoch[026] Batch [0349]/[3759]	Speed: 35.311424 samples/sec	 accuracy=60.812500	 loss=1.628100	 lr=0.010000
Epoch[026] Batch [0399]/[3759]	Speed: 34.892421 samples/sec	 accuracy=60.863281	 loss=1.627634	 lr=0.010000
Epoch[026] Batch [0449]/[3759]	Speed: 35.517092 samples/sec	 accuracy=60.868056	 loss=1.623534	 lr=0.010000
Epoch[026] Batch [0499]/[3759]	Speed: 34.915446 samples/sec	 accuracy=60.768750	 loss=1.627595	 lr=0.010000
Epoch[026] Batch [0549]/[3759]	Speed: 35.414944 samples/sec	 accuracy=60.502841	 loss=1.633357	 lr=0.010000
Epoch[026] Batch [0599]/[3759]	Speed: 35.007260 samples/sec	 accuracy=60.486979	 loss=1.635925	 lr=0.010000
Epoch[026] Batch [0649]/[3759]	Speed: 35.653455 samples/sec	 accuracy=60.557692	 loss=1.635845	 lr=0.010000
Epoch[026] Batch [0699]/[3759]	Speed: 35.118624 samples/sec	 accuracy=60.435268	 loss=1.638272	 lr=0.010000
Epoch[026] Batch [0749]/[3759]	Speed: 35.324225 samples/sec	 accuracy=60.420833	 loss=1.638564	 lr=0.010000
Epoch[026] Batch [0799]/[3759]	Speed: 35.401732 samples/sec	 accuracy=60.406250	 loss=1.638224	 lr=0.010000
Epoch[026] Batch [0849]/[3759]	Speed: 35.439678 samples/sec	 accuracy=60.360294	 loss=1.640229	 lr=0.010000
Epoch[026] Batch [0899]/[3759]	Speed: 35.294750 samples/sec	 accuracy=60.345486	 loss=1.639476	 lr=0.010000
Epoch[026] Batch [0949]/[3759]	Speed: 35.316399 samples/sec	 accuracy=60.309211	 loss=1.640374	 lr=0.010000
Epoch[026] Batch [0999]/[3759]	Speed: 35.145108 samples/sec	 accuracy=60.317188	 loss=1.640559	 lr=0.010000
Epoch[026] Batch [1049]/[3759]	Speed: 35.391094 samples/sec	 accuracy=60.315476	 loss=1.641379	 lr=0.010000
Epoch[026] Batch [1099]/[3759]	Speed: 35.177932 samples/sec	 accuracy=60.292614	 loss=1.641445	 lr=0.010000
Epoch[026] Batch [1149]/[3759]	Speed: 35.409750 samples/sec	 accuracy=60.274457	 loss=1.643010	 lr=0.010000
Epoch[026] Batch [1199]/[3759]	Speed: 35.345191 samples/sec	 accuracy=60.250000	 loss=1.643556	 lr=0.010000
Epoch[026] Batch [1249]/[3759]	Speed: 35.559285 samples/sec	 accuracy=60.280000	 loss=1.643023	 lr=0.010000
Epoch[026] Batch [1299]/[3759]	Speed: 35.292536 samples/sec	 accuracy=60.251202	 loss=1.644603	 lr=0.010000
Epoch[026] Batch [1349]/[3759]	Speed: 35.434727 samples/sec	 accuracy=60.293981	 loss=1.643125	 lr=0.010000
Epoch[026] Batch [1399]/[3759]	Speed: 35.144886 samples/sec	 accuracy=60.320312	 loss=1.641976	 lr=0.010000
Epoch[026] Batch [1449]/[3759]	Speed: 35.348739 samples/sec	 accuracy=60.290948	 loss=1.643001	 lr=0.010000
Epoch[026] Batch [1499]/[3759]	Speed: 35.252302 samples/sec	 accuracy=60.286458	 loss=1.643040	 lr=0.010000
Epoch[026] Batch [1549]/[3759]	Speed: 35.507568 samples/sec	 accuracy=60.283266	 loss=1.642847	 lr=0.010000
Epoch[026] Batch [1599]/[3759]	Speed: 35.254630 samples/sec	 accuracy=60.309570	 loss=1.641982	 lr=0.010000
Epoch[026] Batch [1649]/[3759]	Speed: 35.478049 samples/sec	 accuracy=60.306818	 loss=1.642014	 lr=0.010000
Epoch[026] Batch [1699]/[3759]	Speed: 35.229470 samples/sec	 accuracy=60.317096	 loss=1.642036	 lr=0.010000
Epoch[026] Batch [1749]/[3759]	Speed: 35.460129 samples/sec	 accuracy=60.325893	 loss=1.642304	 lr=0.010000
Epoch[026] Batch [1799]/[3759]	Speed: 35.104089 samples/sec	 accuracy=60.315104	 loss=1.642470	 lr=0.010000
Epoch[026] Batch [1849]/[3759]	Speed: 35.352399 samples/sec	 accuracy=60.299831	 loss=1.641882	 lr=0.010000
Epoch[026] Batch [1899]/[3759]	Speed: 35.241449 samples/sec	 accuracy=60.270559	 loss=1.643856	 lr=0.010000
Epoch[026] Batch [1949]/[3759]	Speed: 35.263309 samples/sec	 accuracy=60.290064	 loss=1.642349	 lr=0.010000
Epoch[026] Batch [1999]/[3759]	Speed: 35.027285 samples/sec	 accuracy=60.266406	 loss=1.644080	 lr=0.010000
Epoch[026] Batch [2049]/[3759]	Speed: 35.306147 samples/sec	 accuracy=60.267530	 loss=1.644071	 lr=0.010000
Epoch[026] Batch [2099]/[3759]	Speed: 35.054567 samples/sec	 accuracy=60.209077	 loss=1.646790	 lr=0.010000
Epoch[026] Batch [2149]/[3759]	Speed: 35.023609 samples/sec	 accuracy=60.194767	 loss=1.647173	 lr=0.010000
Epoch[026] Batch [2199]/[3759]	Speed: 35.538499 samples/sec	 accuracy=60.213068	 loss=1.646718	 lr=0.010000
Epoch[026] Batch [2249]/[3759]	Speed: 34.815003 samples/sec	 accuracy=60.232639	 loss=1.646449	 lr=0.010000
Epoch[026] Batch [2299]/[3759]	Speed: 35.405548 samples/sec	 accuracy=60.230299	 loss=1.647953	 lr=0.010000
Epoch[026] Batch [2349]/[3759]	Speed: 34.880032 samples/sec	 accuracy=60.202793	 loss=1.648387	 lr=0.010000
Epoch[026] Batch [2399]/[3759]	Speed: 35.574848 samples/sec	 accuracy=60.212891	 loss=1.649112	 lr=0.010000
Epoch[026] Batch [2449]/[3759]	Speed: 34.907973 samples/sec	 accuracy=60.179847	 loss=1.650269	 lr=0.010000
Epoch[026] Batch [2499]/[3759]	Speed: 35.483513 samples/sec	 accuracy=60.146250	 loss=1.651553	 lr=0.010000
Epoch[026] Batch [2549]/[3759]	Speed: 35.386009 samples/sec	 accuracy=60.142770	 loss=1.652169	 lr=0.010000
Epoch[026] Batch [2599]/[3759]	Speed: 35.379041 samples/sec	 accuracy=60.149038	 loss=1.652244	 lr=0.010000
Epoch[026] Batch [2649]/[3759]	Speed: 35.337099 samples/sec	 accuracy=60.154481	 loss=1.652095	 lr=0.010000
Epoch[026] Batch [2699]/[3759]	Speed: 35.293582 samples/sec	 accuracy=60.133102	 loss=1.652599	 lr=0.010000
Epoch[026] Batch [2749]/[3759]	Speed: 35.318777 samples/sec	 accuracy=60.123864	 loss=1.653359	 lr=0.010000
Epoch[026] Batch [2799]/[3759]	Speed: 35.289741 samples/sec	 accuracy=60.110491	 loss=1.653574	 lr=0.010000
Epoch[026] Batch [2849]/[3759]	Speed: 35.425157 samples/sec	 accuracy=60.095395	 loss=1.654721	 lr=0.010000
Epoch[026] Batch [2899]/[3759]	Speed: 35.237371 samples/sec	 accuracy=60.072737	 loss=1.655599	 lr=0.010000
Epoch[026] Batch [2949]/[3759]	Speed: 35.244416 samples/sec	 accuracy=60.042373	 loss=1.656834	 lr=0.010000
Epoch[026] Batch [2999]/[3759]	Speed: 35.226948 samples/sec	 accuracy=60.050521	 loss=1.657169	 lr=0.010000
Epoch[026] Batch [3049]/[3759]	Speed: 35.369025 samples/sec	 accuracy=60.045082	 loss=1.657113	 lr=0.010000
Epoch[026] Batch [3099]/[3759]	Speed: 35.380356 samples/sec	 accuracy=60.047883	 loss=1.657727	 lr=0.010000
Epoch[026] Batch [3149]/[3759]	Speed: 35.284022 samples/sec	 accuracy=60.025794	 loss=1.658581	 lr=0.010000
Epoch[026] Batch [3199]/[3759]	Speed: 35.033989 samples/sec	 accuracy=60.006836	 loss=1.659231	 lr=0.010000
Epoch[026] Batch [3249]/[3759]	Speed: 35.389507 samples/sec	 accuracy=59.993269	 loss=1.659430	 lr=0.010000
Epoch[026] Batch [3299]/[3759]	Speed: 35.271804 samples/sec	 accuracy=59.993371	 loss=1.660269	 lr=0.010000
Epoch[026] Batch [3349]/[3759]	Speed: 35.387318 samples/sec	 accuracy=59.980877	 loss=1.661160	 lr=0.010000
Epoch[026] Batch [3399]/[3759]	Speed: 35.334243 samples/sec	 accuracy=59.984835	 loss=1.661435	 lr=0.010000
Epoch[026] Batch [3449]/[3759]	Speed: 35.436533 samples/sec	 accuracy=59.984601	 loss=1.662209	 lr=0.010000
Epoch[026] Batch [3499]/[3759]	Speed: 35.288991 samples/sec	 accuracy=60.004911	 loss=1.661860	 lr=0.010000
Epoch[026] Batch [3549]/[3759]	Speed: 35.337651 samples/sec	 accuracy=59.991197	 loss=1.662576	 lr=0.010000
Epoch[026] Batch [3599]/[3759]	Speed: 35.385504 samples/sec	 accuracy=59.993056	 loss=1.663021	 lr=0.010000
Epoch[026] Batch [3649]/[3759]	Speed: 35.277246 samples/sec	 accuracy=59.982021	 loss=1.663469	 lr=0.010000
Epoch[026] Batch [3699]/[3759]	Speed: 35.146545 samples/sec	 accuracy=59.965794	 loss=1.663745	 lr=0.010000
Epoch[026] Batch [3749]/[3759]	Speed: 36.621733 samples/sec	 accuracy=59.957083	 loss=1.664459	 lr=0.010000
Batch [0049]/[0303]: acc-top1=57.718750 acc-top5=81.250000
Batch [0099]/[0303]: acc-top1=57.890625 acc-top5=80.906250
Batch [0149]/[0303]: acc-top1=58.541667 acc-top5=81.239583
Batch [0199]/[0303]: acc-top1=58.445312 acc-top5=81.421875
Batch [0249]/[0303]: acc-top1=58.406250 acc-top5=81.218750
Batch [0299]/[0303]: acc-top1=58.479167 acc-top5=81.125000
[Epoch 026] training: accuracy=59.961093	 loss=1.664545
[Epoch 026] speed: 35 samples/sec	time cost: 7335.686762
[Epoch 026] validation: acc-top1=58.467409 acc-top5=81.136551 loss=1.962915
Epoch[027] Batch [0049]/[3760]	Speed: 24.380987 samples/sec	 accuracy=61.468750	 loss=1.584449	 lr=0.010000
Epoch[027] Batch [0099]/[3760]	Speed: 35.226366 samples/sec	 accuracy=61.343750	 loss=1.595555	 lr=0.010000
Epoch[027] Batch [0149]/[3760]	Speed: 35.074803 samples/sec	 accuracy=61.635417	 loss=1.588151	 lr=0.010000
Epoch[027] Batch [0199]/[3760]	Speed: 35.085642 samples/sec	 accuracy=61.437500	 loss=1.593588	 lr=0.010000
Epoch[027] Batch [0249]/[3760]	Speed: 35.181790 samples/sec	 accuracy=61.312500	 loss=1.600142	 lr=0.010000
Epoch[027] Batch [0299]/[3760]	Speed: 34.928848 samples/sec	 accuracy=61.260417	 loss=1.598144	 lr=0.010000
Epoch[027] Batch [0349]/[3760]	Speed: 35.351654 samples/sec	 accuracy=61.062500	 loss=1.610542	 lr=0.010000
Epoch[027] Batch [0399]/[3760]	Speed: 34.729763 samples/sec	 accuracy=61.003906	 loss=1.616355	 lr=0.010000
Epoch[027] Batch [0449]/[3760]	Speed: 35.090892 samples/sec	 accuracy=61.031250	 loss=1.619263	 lr=0.010000
Epoch[027] Batch [0499]/[3760]	Speed: 34.831099 samples/sec	 accuracy=60.746875	 loss=1.627416	 lr=0.010000
Epoch[027] Batch [0549]/[3760]	Speed: 35.386637 samples/sec	 accuracy=60.772727	 loss=1.627250	 lr=0.010000
Epoch[027] Batch [0599]/[3760]	Speed: 34.624882 samples/sec	 accuracy=60.864583	 loss=1.623937	 lr=0.010000
Epoch[027] Batch [0649]/[3760]	Speed: 35.213481 samples/sec	 accuracy=60.812500	 loss=1.625668	 lr=0.010000
Epoch[027] Batch [0699]/[3760]	Speed: 34.898462 samples/sec	 accuracy=60.752232	 loss=1.629415	 lr=0.010000
Epoch[027] Batch [0749]/[3760]	Speed: 35.061208 samples/sec	 accuracy=60.808333	 loss=1.627936	 lr=0.010000
Epoch[027] Batch [0799]/[3760]	Speed: 34.996126 samples/sec	 accuracy=60.753906	 loss=1.628595	 lr=0.010000
Epoch[027] Batch [0849]/[3760]	Speed: 34.895269 samples/sec	 accuracy=60.726103	 loss=1.628293	 lr=0.010000
Epoch[027] Batch [0899]/[3760]	Speed: 35.288057 samples/sec	 accuracy=60.741319	 loss=1.628471	 lr=0.010000
Epoch[027] Batch [0949]/[3760]	Speed: 34.792545 samples/sec	 accuracy=60.761513	 loss=1.628308	 lr=0.010000
Epoch[027] Batch [0999]/[3760]	Speed: 35.426784 samples/sec	 accuracy=60.682813	 loss=1.629714	 lr=0.010000
Epoch[027] Batch [1049]/[3760]	Speed: 34.671225 samples/sec	 accuracy=60.732143	 loss=1.627834	 lr=0.010000
Epoch[027] Batch [1099]/[3760]	Speed: 35.369216 samples/sec	 accuracy=60.686080	 loss=1.628399	 lr=0.010000
Epoch[027] Batch [1149]/[3760]	Speed: 35.032890 samples/sec	 accuracy=60.677989	 loss=1.629103	 lr=0.010000
Epoch[027] Batch [1199]/[3760]	Speed: 35.381898 samples/sec	 accuracy=60.688802	 loss=1.630719	 lr=0.010000
Epoch[027] Batch [1249]/[3760]	Speed: 34.913057 samples/sec	 accuracy=60.713750	 loss=1.630817	 lr=0.010000
Epoch[027] Batch [1299]/[3760]	Speed: 35.333861 samples/sec	 accuracy=60.710337	 loss=1.629860	 lr=0.010000
Epoch[027] Batch [1349]/[3760]	Speed: 34.878020 samples/sec	 accuracy=60.728009	 loss=1.629324	 lr=0.010000
Epoch[027] Batch [1399]/[3760]	Speed: 35.163989 samples/sec	 accuracy=60.723214	 loss=1.630444	 lr=0.010000
Epoch[027] Batch [1449]/[3760]	Speed: 34.876825 samples/sec	 accuracy=60.761853	 loss=1.629069	 lr=0.010000
Epoch[027] Batch [1499]/[3760]	Speed: 35.189961 samples/sec	 accuracy=60.738542	 loss=1.631008	 lr=0.010000
Epoch[027] Batch [1549]/[3760]	Speed: 35.185365 samples/sec	 accuracy=60.770161	 loss=1.630364	 lr=0.010000
Epoch[027] Batch [1599]/[3760]	Speed: 35.048214 samples/sec	 accuracy=60.772461	 loss=1.631078	 lr=0.010000
Epoch[027] Batch [1649]/[3760]	Speed: 35.338572 samples/sec	 accuracy=60.738636	 loss=1.633461	 lr=0.010000
Epoch[027] Batch [1699]/[3760]	Speed: 35.180061 samples/sec	 accuracy=60.703125	 loss=1.636908	 lr=0.010000
Epoch[027] Batch [1749]/[3760]	Speed: 35.510504 samples/sec	 accuracy=60.670536	 loss=1.636932	 lr=0.010000
Epoch[027] Batch [1799]/[3760]	Speed: 35.234478 samples/sec	 accuracy=60.646701	 loss=1.638608	 lr=0.010000
Epoch[027] Batch [1849]/[3760]	Speed: 35.454475 samples/sec	 accuracy=60.634291	 loss=1.639460	 lr=0.010000
Epoch[027] Batch [1899]/[3760]	Speed: 35.362947 samples/sec	 accuracy=60.613487	 loss=1.640360	 lr=0.010000
Epoch[027] Batch [1949]/[3760]	Speed: 35.191596 samples/sec	 accuracy=60.563301	 loss=1.642879	 lr=0.010000
Epoch[027] Batch [1999]/[3760]	Speed: 35.434072 samples/sec	 accuracy=60.584375	 loss=1.642276	 lr=0.010000
Epoch[027] Batch [2049]/[3760]	Speed: 35.039644 samples/sec	 accuracy=60.576982	 loss=1.641930	 lr=0.010000
Epoch[027] Batch [2099]/[3760]	Speed: 35.384095 samples/sec	 accuracy=60.548363	 loss=1.644030	 lr=0.010000
Epoch[027] Batch [2149]/[3760]	Speed: 35.492211 samples/sec	 accuracy=60.565407	 loss=1.643510	 lr=0.010000
Epoch[027] Batch [2199]/[3760]	Speed: 35.447421 samples/sec	 accuracy=60.538352	 loss=1.645117	 lr=0.010000
Epoch[027] Batch [2249]/[3760]	Speed: 35.423364 samples/sec	 accuracy=60.554167	 loss=1.644349	 lr=0.010000
Epoch[027] Batch [2299]/[3760]	Speed: 35.466989 samples/sec	 accuracy=60.510190	 loss=1.645499	 lr=0.010000
Epoch[027] Batch [2349]/[3760]	Speed: 35.431214 samples/sec	 accuracy=60.511968	 loss=1.645782	 lr=0.010000
Epoch[027] Batch [2399]/[3760]	Speed: 35.420761 samples/sec	 accuracy=60.479818	 loss=1.646400	 lr=0.010000
Epoch[027] Batch [2449]/[3760]	Speed: 35.417143 samples/sec	 accuracy=60.458546	 loss=1.647149	 lr=0.010000
Epoch[027] Batch [2499]/[3760]	Speed: 35.093315 samples/sec	 accuracy=60.462500	 loss=1.646835	 lr=0.010000
Epoch[027] Batch [2549]/[3760]	Speed: 35.475041 samples/sec	 accuracy=60.443627	 loss=1.648018	 lr=0.010000
Epoch[027] Batch [2599]/[3760]	Speed: 35.152113 samples/sec	 accuracy=60.452524	 loss=1.648317	 lr=0.010000
Epoch[027] Batch [2649]/[3760]	Speed: 35.358790 samples/sec	 accuracy=60.439858	 loss=1.648277	 lr=0.010000
Epoch[027] Batch [2699]/[3760]	Speed: 35.141312 samples/sec	 accuracy=60.416088	 loss=1.648934	 lr=0.010000
Epoch[027] Batch [2749]/[3760]	Speed: 35.526606 samples/sec	 accuracy=60.413636	 loss=1.648239	 lr=0.010000
Epoch[027] Batch [2799]/[3760]	Speed: 35.131492 samples/sec	 accuracy=60.390625	 loss=1.649440	 lr=0.010000
Epoch[027] Batch [2849]/[3760]	Speed: 35.467736 samples/sec	 accuracy=60.361294	 loss=1.650335	 lr=0.010000
Epoch[027] Batch [2899]/[3760]	Speed: 35.326537 samples/sec	 accuracy=60.368534	 loss=1.649671	 lr=0.010000
Epoch[027] Batch [2949]/[3760]	Speed: 35.491164 samples/sec	 accuracy=60.360699	 loss=1.650082	 lr=0.010000
Epoch[027] Batch [2999]/[3760]	Speed: 35.420436 samples/sec	 accuracy=60.344792	 loss=1.651028	 lr=0.010000
Epoch[027] Batch [3049]/[3760]	Speed: 35.381233 samples/sec	 accuracy=60.340676	 loss=1.651344	 lr=0.010000
Epoch[027] Batch [3099]/[3760]	Speed: 35.485425 samples/sec	 accuracy=60.327621	 loss=1.651653	 lr=0.010000
Epoch[027] Batch [3149]/[3760]	Speed: 35.401404 samples/sec	 accuracy=60.319940	 loss=1.652241	 lr=0.010000
Epoch[027] Batch [3199]/[3760]	Speed: 35.301000 samples/sec	 accuracy=60.319336	 loss=1.651943	 lr=0.010000
Epoch[027] Batch [3249]/[3760]	Speed: 35.421676 samples/sec	 accuracy=60.294712	 loss=1.653553	 lr=0.010000
Epoch[027] Batch [3299]/[3760]	Speed: 35.319069 samples/sec	 accuracy=60.271307	 loss=1.654100	 lr=0.010000
Epoch[027] Batch [3349]/[3760]	Speed: 35.395058 samples/sec	 accuracy=60.278918	 loss=1.653853	 lr=0.010000
Epoch[027] Batch [3399]/[3760]	Speed: 35.454875 samples/sec	 accuracy=60.281250	 loss=1.653830	 lr=0.010000
Epoch[027] Batch [3449]/[3760]	Speed: 35.389316 samples/sec	 accuracy=60.269022	 loss=1.654201	 lr=0.010000
Epoch[027] Batch [3499]/[3760]	Speed: 35.231188 samples/sec	 accuracy=60.250000	 loss=1.655237	 lr=0.010000
Epoch[027] Batch [3549]/[3760]	Speed: 35.394774 samples/sec	 accuracy=60.216109	 loss=1.656687	 lr=0.010000
Epoch[027] Batch [3599]/[3760]	Speed: 35.275367 samples/sec	 accuracy=60.208333	 loss=1.656867	 lr=0.010000
Epoch[027] Batch [3649]/[3760]	Speed: 35.236909 samples/sec	 accuracy=60.198630	 loss=1.657338	 lr=0.010000
Epoch[027] Batch [3699]/[3760]	Speed: 35.264615 samples/sec	 accuracy=60.199747	 loss=1.656868	 lr=0.010000
Epoch[027] Batch [3749]/[3760]	Speed: 36.451056 samples/sec	 accuracy=60.195000	 loss=1.657309	 lr=0.010000
Batch [0049]/[0303]: acc-top1=57.750000 acc-top5=81.125000
Batch [0099]/[0303]: acc-top1=57.703125 acc-top5=80.890625
Batch [0149]/[0303]: acc-top1=58.208333 acc-top5=81.187500
Batch [0199]/[0303]: acc-top1=57.882812 acc-top5=81.281250
Batch [0249]/[0303]: acc-top1=57.962500 acc-top5=81.100000
Batch [0299]/[0303]: acc-top1=57.880208 acc-top5=80.973958
[Epoch 027] training: accuracy=60.185755	 loss=1.657602
[Epoch 027] speed: 35 samples/sec	time cost: 7341.390737
[Epoch 027] validation: acc-top1=57.889851 acc-top5=80.981848 loss=1.991599
Epoch[028] Batch [0049]/[3760]	Speed: 25.172660 samples/sec	 accuracy=61.062500	 loss=1.638288	 lr=0.010000
Epoch[028] Batch [0099]/[3760]	Speed: 35.203823 samples/sec	 accuracy=61.468750	 loss=1.590495	 lr=0.010000
Epoch[028] Batch [0149]/[3760]	Speed: 35.012370 samples/sec	 accuracy=61.395833	 loss=1.594062	 lr=0.010000
Epoch[028] Batch [0199]/[3760]	Speed: 35.142770 samples/sec	 accuracy=61.843750	 loss=1.583331	 lr=0.010000
Epoch[028] Batch [0249]/[3760]	Speed: 35.116086 samples/sec	 accuracy=61.912500	 loss=1.580880	 lr=0.010000
Epoch[028] Batch [0299]/[3760]	Speed: 34.822239 samples/sec	 accuracy=61.593750	 loss=1.599395	 lr=0.010000
Epoch[028] Batch [0349]/[3760]	Speed: 35.069503 samples/sec	 accuracy=61.562500	 loss=1.606726	 lr=0.010000
Epoch[028] Batch [0399]/[3760]	Speed: 34.492033 samples/sec	 accuracy=61.605469	 loss=1.610782	 lr=0.010000
Epoch[028] Batch [0449]/[3760]	Speed: 35.345528 samples/sec	 accuracy=61.458333	 loss=1.615568	 lr=0.010000
Epoch[028] Batch [0499]/[3760]	Speed: 34.698403 samples/sec	 accuracy=61.531250	 loss=1.613769	 lr=0.010000
Epoch[028] Batch [0549]/[3760]	Speed: 35.495543 samples/sec	 accuracy=61.346591	 loss=1.620387	 lr=0.010000
Epoch[028] Batch [0599]/[3760]	Speed: 35.019984 samples/sec	 accuracy=61.343750	 loss=1.622531	 lr=0.010000
Epoch[028] Batch [0649]/[3760]	Speed: 35.226309 samples/sec	 accuracy=61.161058	 loss=1.626579	 lr=0.010000
Epoch[028] Batch [0699]/[3760]	Speed: 35.254255 samples/sec	 accuracy=61.149554	 loss=1.624769	 lr=0.010000
Epoch[028] Batch [0749]/[3760]	Speed: 35.274696 samples/sec	 accuracy=61.362500	 loss=1.618277	 lr=0.010000
Epoch[028] Batch [0799]/[3760]	Speed: 35.277038 samples/sec	 accuracy=61.283203	 loss=1.618005	 lr=0.010000
Epoch[028] Batch [0849]/[3760]	Speed: 35.104886 samples/sec	 accuracy=61.257353	 loss=1.620111	 lr=0.010000
Epoch[028] Batch [0899]/[3760]	Speed: 35.476972 samples/sec	 accuracy=61.189236	 loss=1.622972	 lr=0.010000
Epoch[028] Batch [0949]/[3760]	Speed: 35.404773 samples/sec	 accuracy=61.047697	 loss=1.625779	 lr=0.010000
Epoch[028] Batch [0999]/[3760]	Speed: 35.380941 samples/sec	 accuracy=61.067187	 loss=1.625337	 lr=0.010000
Epoch[028] Batch [1049]/[3760]	Speed: 35.226393 samples/sec	 accuracy=61.034226	 loss=1.626150	 lr=0.010000
Epoch[028] Batch [1099]/[3760]	Speed: 35.271004 samples/sec	 accuracy=60.988636	 loss=1.627860	 lr=0.010000
Epoch[028] Batch [1149]/[3760]	Speed: 35.206736 samples/sec	 accuracy=60.979620	 loss=1.629162	 lr=0.010000
Epoch[028] Batch [1199]/[3760]	Speed: 35.367341 samples/sec	 accuracy=60.958333	 loss=1.629366	 lr=0.010000
Epoch[028] Batch [1249]/[3760]	Speed: 35.277927 samples/sec	 accuracy=60.991250	 loss=1.628632	 lr=0.010000
Epoch[028] Batch [1299]/[3760]	Speed: 35.278276 samples/sec	 accuracy=60.987981	 loss=1.627811	 lr=0.010000
Epoch[028] Batch [1349]/[3760]	Speed: 34.987872 samples/sec	 accuracy=60.953704	 loss=1.629688	 lr=0.010000
Epoch[028] Batch [1399]/[3760]	Speed: 35.317048 samples/sec	 accuracy=60.900670	 loss=1.631039	 lr=0.010000
Epoch[028] Batch [1449]/[3760]	Speed: 34.912732 samples/sec	 accuracy=60.898707	 loss=1.631246	 lr=0.010000
Epoch[028] Batch [1499]/[3760]	Speed: 35.422743 samples/sec	 accuracy=60.858333	 loss=1.633091	 lr=0.010000
Epoch[028] Batch [1549]/[3760]	Speed: 34.889662 samples/sec	 accuracy=60.853831	 loss=1.632906	 lr=0.010000
Epoch[028] Batch [1599]/[3760]	Speed: 35.339567 samples/sec	 accuracy=60.816406	 loss=1.633624	 lr=0.010000
Epoch[028] Batch [1649]/[3760]	Speed: 35.185736 samples/sec	 accuracy=60.796402	 loss=1.634379	 lr=0.010000
Epoch[028] Batch [1699]/[3760]	Speed: 35.373673 samples/sec	 accuracy=60.753676	 loss=1.635838	 lr=0.010000
Epoch[028] Batch [1749]/[3760]	Speed: 35.074355 samples/sec	 accuracy=60.750893	 loss=1.635867	 lr=0.010000
Epoch[028] Batch [1799]/[3760]	Speed: 35.477916 samples/sec	 accuracy=60.736979	 loss=1.636621	 lr=0.010000
Epoch[028] Batch [1849]/[3760]	Speed: 35.244955 samples/sec	 accuracy=60.690034	 loss=1.638845	 lr=0.010000
Epoch[028] Batch [1899]/[3760]	Speed: 35.363064 samples/sec	 accuracy=60.627467	 loss=1.641063	 lr=0.010000
Epoch[028] Batch [1949]/[3760]	Speed: 35.356341 samples/sec	 accuracy=60.597756	 loss=1.642069	 lr=0.010000
Epoch[028] Batch [1999]/[3760]	Speed: 35.332338 samples/sec	 accuracy=60.617969	 loss=1.641360	 lr=0.010000
Epoch[028] Batch [2049]/[3760]	Speed: 35.308935 samples/sec	 accuracy=60.639482	 loss=1.640274	 lr=0.010000
Epoch[028] Batch [2099]/[3760]	Speed: 35.380004 samples/sec	 accuracy=60.636905	 loss=1.640412	 lr=0.010000
Epoch[028] Batch [2149]/[3760]	Speed: 35.469934 samples/sec	 accuracy=60.616279	 loss=1.641833	 lr=0.010000
Epoch[028] Batch [2199]/[3760]	Speed: 35.375153 samples/sec	 accuracy=60.642756	 loss=1.640231	 lr=0.010000
Epoch[028] Batch [2249]/[3760]	Speed: 35.292062 samples/sec	 accuracy=60.618750	 loss=1.641328	 lr=0.010000
Epoch[028] Batch [2299]/[3760]	Speed: 35.382572 samples/sec	 accuracy=60.590353	 loss=1.642184	 lr=0.010000
Epoch[028] Batch [2349]/[3760]	Speed: 35.190037 samples/sec	 accuracy=60.595080	 loss=1.642134	 lr=0.010000
Epoch[028] Batch [2399]/[3760]	Speed: 35.223021 samples/sec	 accuracy=60.582682	 loss=1.642770	 lr=0.010000
Epoch[028] Batch [2449]/[3760]	Speed: 35.224750 samples/sec	 accuracy=60.575255	 loss=1.643295	 lr=0.010000
Epoch[028] Batch [2499]/[3760]	Speed: 35.310726 samples/sec	 accuracy=60.600000	 loss=1.642117	 lr=0.010000
Epoch[028] Batch [2549]/[3760]	Speed: 35.285263 samples/sec	 accuracy=60.594975	 loss=1.642275	 lr=0.010000
Epoch[028] Batch [2599]/[3760]	Speed: 35.376510 samples/sec	 accuracy=60.605769	 loss=1.642797	 lr=0.010000
Epoch[028] Batch [2649]/[3760]	Speed: 35.453954 samples/sec	 accuracy=60.599057	 loss=1.642877	 lr=0.010000
Epoch[028] Batch [2699]/[3760]	Speed: 35.396489 samples/sec	 accuracy=60.539352	 loss=1.644738	 lr=0.010000
Epoch[028] Batch [2749]/[3760]	Speed: 35.528506 samples/sec	 accuracy=60.536932	 loss=1.645331	 lr=0.010000
Epoch[028] Batch [2799]/[3760]	Speed: 35.312319 samples/sec	 accuracy=60.534598	 loss=1.644851	 lr=0.010000
Epoch[028] Batch [2849]/[3760]	Speed: 35.300078 samples/sec	 accuracy=60.514803	 loss=1.646023	 lr=0.010000
Epoch[028] Batch [2899]/[3760]	Speed: 35.250580 samples/sec	 accuracy=60.501078	 loss=1.646302	 lr=0.010000
Epoch[028] Batch [2949]/[3760]	Speed: 35.511158 samples/sec	 accuracy=60.514831	 loss=1.645866	 lr=0.010000
Epoch[028] Batch [2999]/[3760]	Speed: 35.226782 samples/sec	 accuracy=60.476562	 loss=1.647161	 lr=0.010000
Epoch[028] Batch [3049]/[3760]	Speed: 35.114058 samples/sec	 accuracy=60.459529	 loss=1.647936	 lr=0.010000
Epoch[028] Batch [3099]/[3760]	Speed: 35.390544 samples/sec	 accuracy=60.469758	 loss=1.647546	 lr=0.010000
Epoch[028] Batch [3149]/[3760]	Speed: 35.019444 samples/sec	 accuracy=60.468750	 loss=1.647658	 lr=0.010000
Epoch[028] Batch [3199]/[3760]	Speed: 35.207533 samples/sec	 accuracy=60.479492	 loss=1.647624	 lr=0.010000
Epoch[028] Batch [3249]/[3760]	Speed: 35.099449 samples/sec	 accuracy=60.471154	 loss=1.648192	 lr=0.010000
Epoch[028] Batch [3299]/[3760]	Speed: 35.325645 samples/sec	 accuracy=60.480587	 loss=1.647998	 lr=0.010000
Epoch[028] Batch [3349]/[3760]	Speed: 35.144772 samples/sec	 accuracy=60.490672	 loss=1.648203	 lr=0.010000
Epoch[028] Batch [3399]/[3760]	Speed: 35.037875 samples/sec	 accuracy=60.458640	 loss=1.649192	 lr=0.010000
Epoch[028] Batch [3449]/[3760]	Speed: 35.318909 samples/sec	 accuracy=60.451993	 loss=1.649059	 lr=0.010000
Epoch[028] Batch [3499]/[3760]	Speed: 35.028830 samples/sec	 accuracy=60.450000	 loss=1.648718	 lr=0.010000
Epoch[028] Batch [3549]/[3760]	Speed: 35.424307 samples/sec	 accuracy=60.451144	 loss=1.648497	 lr=0.010000
Epoch[028] Batch [3599]/[3760]	Speed: 35.187132 samples/sec	 accuracy=60.465278	 loss=1.647942	 lr=0.010000
Epoch[028] Batch [3649]/[3760]	Speed: 35.506923 samples/sec	 accuracy=60.444777	 loss=1.648059	 lr=0.010000
Epoch[028] Batch [3699]/[3760]	Speed: 34.832139 samples/sec	 accuracy=60.421875	 loss=1.648543	 lr=0.010000
Epoch[028] Batch [3749]/[3760]	Speed: 36.504607 samples/sec	 accuracy=60.428333	 loss=1.648401	 lr=0.010000
Batch [0049]/[0303]: acc-top1=58.406250 acc-top5=81.718750
Batch [0099]/[0303]: acc-top1=58.484375 acc-top5=81.171875
Batch [0149]/[0303]: acc-top1=58.729167 acc-top5=81.291667
Batch [0199]/[0303]: acc-top1=58.515625 acc-top5=81.367188
Batch [0249]/[0303]: acc-top1=58.118750 acc-top5=81.175000
Batch [0299]/[0303]: acc-top1=58.088542 acc-top5=81.244792
[Epoch 028] training: accuracy=60.434259	 loss=1.648210
[Epoch 028] speed: 35 samples/sec	time cost: 7340.255427
[Epoch 028] validation: acc-top1=58.054868 acc-top5=81.229373 loss=2.027786
Epoch[029] Batch [0049]/[3759]	Speed: 24.802554 samples/sec	 accuracy=60.812500	 loss=1.621960	 lr=0.010000
Epoch[029] Batch [0099]/[3759]	Speed: 35.316344 samples/sec	 accuracy=61.093750	 loss=1.590447	 lr=0.010000
Epoch[029] Batch [0149]/[3759]	Speed: 35.186540 samples/sec	 accuracy=61.708333	 loss=1.574124	 lr=0.010000
Epoch[029] Batch [0199]/[3759]	Speed: 35.100410 samples/sec	 accuracy=61.390625	 loss=1.587666	 lr=0.010000
Epoch[029] Batch [0249]/[3759]	Speed: 35.413625 samples/sec	 accuracy=61.418750	 loss=1.582216	 lr=0.010000
Epoch[029] Batch [0299]/[3759]	Speed: 35.239958 samples/sec	 accuracy=61.708333	 loss=1.578419	 lr=0.010000
Epoch[029] Batch [0349]/[3759]	Speed: 35.435478 samples/sec	 accuracy=61.683036	 loss=1.581030	 lr=0.010000
Epoch[029] Batch [0399]/[3759]	Speed: 35.122984 samples/sec	 accuracy=61.582031	 loss=1.587488	 lr=0.010000
Epoch[029] Batch [0449]/[3759]	Speed: 35.399777 samples/sec	 accuracy=61.722222	 loss=1.583826	 lr=0.010000
Epoch[029] Batch [0499]/[3759]	Speed: 35.274314 samples/sec	 accuracy=61.612500	 loss=1.589819	 lr=0.010000
Epoch[029] Batch [0549]/[3759]	Speed: 35.513634 samples/sec	 accuracy=61.428977	 loss=1.594170	 lr=0.010000
Epoch[029] Batch [0599]/[3759]	Speed: 35.374840 samples/sec	 accuracy=61.356771	 loss=1.599216	 lr=0.010000
Epoch[029] Batch [0649]/[3759]	Speed: 35.504978 samples/sec	 accuracy=61.382212	 loss=1.598072	 lr=0.010000
Epoch[029] Batch [0699]/[3759]	Speed: 35.482526 samples/sec	 accuracy=61.321429	 loss=1.599058	 lr=0.010000
Epoch[029] Batch [0749]/[3759]	Speed: 35.506781 samples/sec	 accuracy=61.172917	 loss=1.605055	 lr=0.010000
Epoch[029] Batch [0799]/[3759]	Speed: 35.206879 samples/sec	 accuracy=61.251953	 loss=1.604580	 lr=0.010000
Epoch[029] Batch [0849]/[3759]	Speed: 35.191000 samples/sec	 accuracy=61.189338	 loss=1.607312	 lr=0.010000
Epoch[029] Batch [0899]/[3759]	Speed: 35.582282 samples/sec	 accuracy=61.138889	 loss=1.608774	 lr=0.010000
Epoch[029] Batch [0949]/[3759]	Speed: 35.230556 samples/sec	 accuracy=61.098684	 loss=1.610498	 lr=0.010000
Epoch[029] Batch [0999]/[3759]	Speed: 35.268443 samples/sec	 accuracy=61.068750	 loss=1.611667	 lr=0.010000
Epoch[029] Batch [1049]/[3759]	Speed: 35.284568 samples/sec	 accuracy=61.063988	 loss=1.610210	 lr=0.010000
Epoch[029] Batch [1099]/[3759]	Speed: 35.378029 samples/sec	 accuracy=61.056818	 loss=1.611260	 lr=0.010000
Epoch[029] Batch [1149]/[3759]	Speed: 35.501901 samples/sec	 accuracy=60.995924	 loss=1.614651	 lr=0.010000
Epoch[029] Batch [1199]/[3759]	Speed: 35.450186 samples/sec	 accuracy=60.897135	 loss=1.617389	 lr=0.010000
Epoch[029] Batch [1249]/[3759]	Speed: 35.352102 samples/sec	 accuracy=60.907500	 loss=1.617375	 lr=0.010000
Epoch[029] Batch [1299]/[3759]	Speed: 35.247510 samples/sec	 accuracy=60.929087	 loss=1.617564	 lr=0.010000
Epoch[029] Batch [1349]/[3759]	Speed: 35.235663 samples/sec	 accuracy=60.958333	 loss=1.617954	 lr=0.010000
Epoch[029] Batch [1399]/[3759]	Speed: 35.019215 samples/sec	 accuracy=60.940848	 loss=1.620203	 lr=0.010000
Epoch[029] Batch [1449]/[3759]	Speed: 35.375570 samples/sec	 accuracy=60.967672	 loss=1.617748	 lr=0.010000
Epoch[029] Batch [1499]/[3759]	Speed: 34.812876 samples/sec	 accuracy=60.993750	 loss=1.618000	 lr=0.010000
Epoch[029] Batch [1549]/[3759]	Speed: 35.379151 samples/sec	 accuracy=61.015121	 loss=1.618352	 lr=0.010000
Epoch[029] Batch [1599]/[3759]	Speed: 34.810471 samples/sec	 accuracy=60.953125	 loss=1.619609	 lr=0.010000
Epoch[029] Batch [1649]/[3759]	Speed: 35.289614 samples/sec	 accuracy=60.914773	 loss=1.620546	 lr=0.010000
Epoch[029] Batch [1699]/[3759]	Speed: 34.821739 samples/sec	 accuracy=60.890625	 loss=1.622570	 lr=0.010000
Epoch[029] Batch [1749]/[3759]	Speed: 35.243469 samples/sec	 accuracy=60.858929	 loss=1.623622	 lr=0.010000
Epoch[029] Batch [1799]/[3759]	Speed: 35.326546 samples/sec	 accuracy=60.883681	 loss=1.622547	 lr=0.010000
Epoch[029] Batch [1849]/[3759]	Speed: 35.235261 samples/sec	 accuracy=60.896959	 loss=1.621974	 lr=0.010000
Epoch[029] Batch [1899]/[3759]	Speed: 35.457098 samples/sec	 accuracy=60.865954	 loss=1.623031	 lr=0.010000
Epoch[029] Batch [1949]/[3759]	Speed: 35.100739 samples/sec	 accuracy=60.875801	 loss=1.621697	 lr=0.010000
Epoch[029] Batch [1999]/[3759]	Speed: 35.603954 samples/sec	 accuracy=60.885156	 loss=1.621413	 lr=0.010000
Epoch[029] Batch [2049]/[3759]	Speed: 35.164007 samples/sec	 accuracy=60.845274	 loss=1.622479	 lr=0.010000
Epoch[029] Batch [2099]/[3759]	Speed: 35.638695 samples/sec	 accuracy=60.860863	 loss=1.622719	 lr=0.010000
Epoch[029] Batch [2149]/[3759]	Speed: 35.298598 samples/sec	 accuracy=60.825581	 loss=1.624620	 lr=0.010000
Epoch[029] Batch [2199]/[3759]	Speed: 35.545980 samples/sec	 accuracy=60.856534	 loss=1.624488	 lr=0.010000
Epoch[029] Batch [2249]/[3759]	Speed: 35.484296 samples/sec	 accuracy=60.807639	 loss=1.626508	 lr=0.010000
Epoch[029] Batch [2299]/[3759]	Speed: 35.443704 samples/sec	 accuracy=60.777853	 loss=1.626686	 lr=0.010000
Epoch[029] Batch [2349]/[3759]	Speed: 35.173675 samples/sec	 accuracy=60.763298	 loss=1.627198	 lr=0.010000
Epoch[029] Batch [2399]/[3759]	Speed: 35.461768 samples/sec	 accuracy=60.751953	 loss=1.628062	 lr=0.010000
Epoch[029] Batch [2449]/[3759]	Speed: 35.081458 samples/sec	 accuracy=60.737245	 loss=1.628759	 lr=0.010000
Epoch[029] Batch [2499]/[3759]	Speed: 35.258572 samples/sec	 accuracy=60.750625	 loss=1.628616	 lr=0.010000
Epoch[029] Batch [2549]/[3759]	Speed: 35.054978 samples/sec	 accuracy=60.738971	 loss=1.629677	 lr=0.010000
Epoch[029] Batch [2599]/[3759]	Speed: 35.278683 samples/sec	 accuracy=60.757812	 loss=1.628164	 lr=0.010000
Epoch[029] Batch [2649]/[3759]	Speed: 35.142490 samples/sec	 accuracy=60.790094	 loss=1.626544	 lr=0.010000
Epoch[029] Batch [2699]/[3759]	Speed: 35.041677 samples/sec	 accuracy=60.756944	 loss=1.627253	 lr=0.010000
Epoch[029] Batch [2749]/[3759]	Speed: 35.480781 samples/sec	 accuracy=60.746023	 loss=1.628547	 lr=0.010000
Epoch[029] Batch [2799]/[3759]	Speed: 34.669939 samples/sec	 accuracy=60.728237	 loss=1.628788	 lr=0.010000
Epoch[029] Batch [2849]/[3759]	Speed: 35.387775 samples/sec	 accuracy=60.720395	 loss=1.628967	 lr=0.010000
Epoch[029] Batch [2899]/[3759]	Speed: 34.675904 samples/sec	 accuracy=60.699353	 loss=1.629839	 lr=0.010000
Epoch[029] Batch [2949]/[3759]	Speed: 35.431339 samples/sec	 accuracy=60.670021	 loss=1.630516	 lr=0.010000
Epoch[029] Batch [2999]/[3759]	Speed: 35.138860 samples/sec	 accuracy=60.656250	 loss=1.630943	 lr=0.010000
Epoch[029] Batch [3049]/[3759]	Speed: 35.446869 samples/sec	 accuracy=60.634734	 loss=1.630847	 lr=0.010000
Epoch[029] Batch [3099]/[3759]	Speed: 35.259459 samples/sec	 accuracy=60.625504	 loss=1.631366	 lr=0.010000
Epoch[029] Batch [3149]/[3759]	Speed: 35.222396 samples/sec	 accuracy=60.626488	 loss=1.632008	 lr=0.010000
Epoch[029] Batch [3199]/[3759]	Speed: 35.297470 samples/sec	 accuracy=60.609863	 loss=1.633076	 lr=0.010000
Epoch[029] Batch [3249]/[3759]	Speed: 35.116159 samples/sec	 accuracy=60.617788	 loss=1.633335	 lr=0.010000
Epoch[029] Batch [3299]/[3759]	Speed: 35.362113 samples/sec	 accuracy=60.606061	 loss=1.633919	 lr=0.010000
Epoch[029] Batch [3349]/[3759]	Speed: 35.104775 samples/sec	 accuracy=60.604944	 loss=1.634363	 lr=0.010000
Epoch[029] Batch [3399]/[3759]	Speed: 35.371636 samples/sec	 accuracy=60.589614	 loss=1.634415	 lr=0.010000
Epoch[029] Batch [3449]/[3759]	Speed: 35.054882 samples/sec	 accuracy=60.575181	 loss=1.635134	 lr=0.010000
Epoch[029] Batch [3499]/[3759]	Speed: 35.516759 samples/sec	 accuracy=60.550893	 loss=1.636226	 lr=0.010000
Epoch[029] Batch [3549]/[3759]	Speed: 35.021293 samples/sec	 accuracy=60.527289	 loss=1.637485	 lr=0.010000
Epoch[029] Batch [3599]/[3759]	Speed: 35.365365 samples/sec	 accuracy=60.522569	 loss=1.637705	 lr=0.010000
Epoch[029] Batch [3649]/[3759]	Speed: 35.171779 samples/sec	 accuracy=60.531678	 loss=1.637967	 lr=0.010000
Epoch[029] Batch [3699]/[3759]	Speed: 35.210561 samples/sec	 accuracy=60.533361	 loss=1.638102	 lr=0.010000
Epoch[029] Batch [3749]/[3759]	Speed: 36.591366 samples/sec	 accuracy=60.519583	 loss=1.638349	 lr=0.010000
Batch [0049]/[0303]: acc-top1=57.031250 acc-top5=79.843750
Batch [0099]/[0303]: acc-top1=57.390625 acc-top5=80.046875
Batch [0149]/[0303]: acc-top1=57.708333 acc-top5=80.541667
Batch [0199]/[0303]: acc-top1=57.585937 acc-top5=80.656250
Batch [0249]/[0303]: acc-top1=57.325000 acc-top5=80.343750
Batch [0299]/[0303]: acc-top1=57.458333 acc-top5=80.401042
[Epoch 029] training: accuracy=60.516843	 loss=1.638534
[Epoch 029] speed: 35 samples/sec	time cost: 7337.738590
[Epoch 029] validation: acc-top1=57.415429 acc-top5=80.383663 loss=2.003578
Epoch[030] Batch [0049]/[3760]	Speed: 24.843492 samples/sec	 accuracy=60.468750	 loss=1.665686	 lr=0.010000
Epoch[030] Batch [0099]/[3760]	Speed: 35.114298 samples/sec	 accuracy=61.015625	 loss=1.627906	 lr=0.010000
Epoch[030] Batch [0149]/[3760]	Speed: 34.913266 samples/sec	 accuracy=61.135417	 loss=1.619983	 lr=0.010000
Epoch[030] Batch [0199]/[3760]	Speed: 34.916892 samples/sec	 accuracy=61.703125	 loss=1.595022	 lr=0.010000
Epoch[030] Batch [0249]/[3760]	Speed: 35.152690 samples/sec	 accuracy=61.812500	 loss=1.581819	 lr=0.010000
Epoch[030] Batch [0299]/[3760]	Speed: 34.880985 samples/sec	 accuracy=61.776042	 loss=1.587109	 lr=0.010000
Epoch[030] Batch [0349]/[3760]	Speed: 35.298244 samples/sec	 accuracy=61.741071	 loss=1.590255	 lr=0.010000
Epoch[030] Batch [0399]/[3760]	Speed: 34.729270 samples/sec	 accuracy=61.617188	 loss=1.588993	 lr=0.010000
Epoch[030] Batch [0449]/[3760]	Speed: 35.462584 samples/sec	 accuracy=61.572917	 loss=1.589530	 lr=0.010000
Epoch[030] Batch [0499]/[3760]	Speed: 35.109745 samples/sec	 accuracy=61.568750	 loss=1.589180	 lr=0.010000
Epoch[030] Batch [0549]/[3760]	Speed: 35.399902 samples/sec	 accuracy=61.593750	 loss=1.588831	 lr=0.010000
Epoch[030] Batch [0599]/[3760]	Speed: 35.109427 samples/sec	 accuracy=61.570312	 loss=1.590309	 lr=0.010000
Epoch[030] Batch [0649]/[3760]	Speed: 35.357484 samples/sec	 accuracy=61.507212	 loss=1.591870	 lr=0.010000
Epoch[030] Batch [0699]/[3760]	Speed: 34.881636 samples/sec	 accuracy=61.390625	 loss=1.593543	 lr=0.010000
Epoch[030] Batch [0749]/[3760]	Speed: 35.188322 samples/sec	 accuracy=61.466667	 loss=1.593041	 lr=0.010000
Epoch[030] Batch [0799]/[3760]	Speed: 35.113629 samples/sec	 accuracy=61.496094	 loss=1.592494	 lr=0.010000
Epoch[030] Batch [0849]/[3760]	Speed: 35.279454 samples/sec	 accuracy=61.511029	 loss=1.591614	 lr=0.010000
Epoch[030] Batch [0899]/[3760]	Speed: 35.144390 samples/sec	 accuracy=61.430556	 loss=1.593373	 lr=0.010000
Epoch[030] Batch [0949]/[3760]	Speed: 35.201854 samples/sec	 accuracy=61.442434	 loss=1.593473	 lr=0.010000
Epoch[030] Batch [0999]/[3760]	Speed: 35.299892 samples/sec	 accuracy=61.384375	 loss=1.595623	 lr=0.010000
Epoch[030] Batch [1049]/[3760]	Speed: 35.211132 samples/sec	 accuracy=61.415179	 loss=1.595188	 lr=0.010000
Epoch[030] Batch [1099]/[3760]	Speed: 35.432825 samples/sec	 accuracy=61.369318	 loss=1.597454	 lr=0.010000
Epoch[030] Batch [1149]/[3760]	Speed: 35.171243 samples/sec	 accuracy=61.320652	 loss=1.598032	 lr=0.010000
Epoch[030] Batch [1199]/[3760]	Speed: 35.467525 samples/sec	 accuracy=61.328125	 loss=1.598765	 lr=0.010000
Epoch[030] Batch [1249]/[3760]	Speed: 35.133030 samples/sec	 accuracy=61.256250	 loss=1.601356	 lr=0.010000
Epoch[030] Batch [1299]/[3760]	Speed: 35.389661 samples/sec	 accuracy=61.277644	 loss=1.601611	 lr=0.010000
Epoch[030] Batch [1349]/[3760]	Speed: 35.185326 samples/sec	 accuracy=61.265046	 loss=1.602889	 lr=0.010000
Epoch[030] Batch [1399]/[3760]	Speed: 35.318230 samples/sec	 accuracy=61.285714	 loss=1.602993	 lr=0.010000
Epoch[030] Batch [1449]/[3760]	Speed: 35.003581 samples/sec	 accuracy=61.270474	 loss=1.603979	 lr=0.010000
Epoch[030] Batch [1499]/[3760]	Speed: 35.244322 samples/sec	 accuracy=61.265625	 loss=1.605556	 lr=0.010000
Epoch[030] Batch [1549]/[3760]	Speed: 35.207445 samples/sec	 accuracy=61.227823	 loss=1.607840	 lr=0.010000
Epoch[030] Batch [1599]/[3760]	Speed: 35.018273 samples/sec	 accuracy=61.215820	 loss=1.607494	 lr=0.010000
Epoch[030] Batch [1649]/[3760]	Speed: 35.407956 samples/sec	 accuracy=61.266098	 loss=1.605779	 lr=0.010000
Epoch[030] Batch [1699]/[3760]	Speed: 34.772207 samples/sec	 accuracy=61.259191	 loss=1.605785	 lr=0.010000
Epoch[030] Batch [1749]/[3760]	Speed: 35.332700 samples/sec	 accuracy=61.193750	 loss=1.608613	 lr=0.010000
Epoch[030] Batch [1799]/[3760]	Speed: 34.824458 samples/sec	 accuracy=61.148438	 loss=1.610675	 lr=0.010000
Epoch[030] Batch [1849]/[3760]	Speed: 35.357602 samples/sec	 accuracy=61.130912	 loss=1.612162	 lr=0.010000
Epoch[030] Batch [1899]/[3760]	Speed: 34.881880 samples/sec	 accuracy=61.080592	 loss=1.613794	 lr=0.010000
Epoch[030] Batch [1949]/[3760]	Speed: 35.454812 samples/sec	 accuracy=61.048077	 loss=1.614382	 lr=0.010000
Epoch[030] Batch [1999]/[3760]	Speed: 34.791209 samples/sec	 accuracy=61.037500	 loss=1.614400	 lr=0.010000
Epoch[030] Batch [2049]/[3760]	Speed: 35.368023 samples/sec	 accuracy=61.033537	 loss=1.614805	 lr=0.010000
Epoch[030] Batch [2099]/[3760]	Speed: 35.167751 samples/sec	 accuracy=61.043155	 loss=1.614402	 lr=0.010000
Epoch[030] Batch [2149]/[3760]	Speed: 35.359273 samples/sec	 accuracy=60.977471	 loss=1.616667	 lr=0.010000
Epoch[030] Batch [2199]/[3760]	Speed: 35.136515 samples/sec	 accuracy=60.997159	 loss=1.616329	 lr=0.010000
Epoch[030] Batch [2249]/[3760]	Speed: 35.202882 samples/sec	 accuracy=60.961806	 loss=1.616421	 lr=0.010000
Epoch[030] Batch [2299]/[3760]	Speed: 35.316261 samples/sec	 accuracy=60.947690	 loss=1.617001	 lr=0.010000
Epoch[030] Batch [2349]/[3760]	Speed: 35.026040 samples/sec	 accuracy=60.942154	 loss=1.617060	 lr=0.010000
Epoch[030] Batch [2399]/[3760]	Speed: 35.137471 samples/sec	 accuracy=60.923177	 loss=1.618577	 lr=0.010000
Epoch[030] Batch [2449]/[3760]	Speed: 35.089440 samples/sec	 accuracy=60.892857	 loss=1.619807	 lr=0.010000
Epoch[030] Batch [2499]/[3760]	Speed: 35.066235 samples/sec	 accuracy=60.869375	 loss=1.620963	 lr=0.010000
Epoch[030] Batch [2549]/[3760]	Speed: 34.674101 samples/sec	 accuracy=60.844975	 loss=1.622186	 lr=0.010000
Epoch[030] Batch [2599]/[3760]	Speed: 35.239761 samples/sec	 accuracy=60.855168	 loss=1.621410	 lr=0.010000
Epoch[030] Batch [2649]/[3760]	Speed: 34.779720 samples/sec	 accuracy=60.816038	 loss=1.623238	 lr=0.010000
Epoch[030] Batch [2699]/[3760]	Speed: 35.245473 samples/sec	 accuracy=60.784722	 loss=1.624261	 lr=0.010000
Epoch[030] Batch [2749]/[3760]	Speed: 34.457740 samples/sec	 accuracy=60.760795	 loss=1.625196	 lr=0.010000
Epoch[030] Batch [2799]/[3760]	Speed: 35.269606 samples/sec	 accuracy=60.763951	 loss=1.624970	 lr=0.010000
Epoch[030] Batch [2849]/[3760]	Speed: 34.491251 samples/sec	 accuracy=60.769189	 loss=1.625213	 lr=0.010000
Epoch[030] Batch [2899]/[3760]	Speed: 35.004429 samples/sec	 accuracy=60.761315	 loss=1.625864	 lr=0.010000
Epoch[030] Batch [2949]/[3760]	Speed: 34.902363 samples/sec	 accuracy=60.755826	 loss=1.625815	 lr=0.010000
Epoch[030] Batch [2999]/[3760]	Speed: 34.869582 samples/sec	 accuracy=60.763542	 loss=1.626080	 lr=0.010000
Epoch[030] Batch [3049]/[3760]	Speed: 34.956624 samples/sec	 accuracy=60.758709	 loss=1.626533	 lr=0.010000
Epoch[030] Batch [3099]/[3760]	Speed: 34.599104 samples/sec	 accuracy=60.733871	 loss=1.627687	 lr=0.010000
Epoch[030] Batch [3149]/[3760]	Speed: 35.263807 samples/sec	 accuracy=60.712302	 loss=1.628562	 lr=0.010000
Epoch[030] Batch [3199]/[3760]	Speed: 34.723695 samples/sec	 accuracy=60.721680	 loss=1.629019	 lr=0.010000
Epoch[030] Batch [3249]/[3760]	Speed: 35.588064 samples/sec	 accuracy=60.710096	 loss=1.629988	 lr=0.010000
Epoch[030] Batch [3299]/[3760]	Speed: 35.139408 samples/sec	 accuracy=60.705492	 loss=1.630569	 lr=0.010000
Epoch[030] Batch [3349]/[3760]	Speed: 35.468874 samples/sec	 accuracy=60.698694	 loss=1.631098	 lr=0.010000
Epoch[030] Batch [3399]/[3760]	Speed: 34.971539 samples/sec	 accuracy=60.697610	 loss=1.631570	 lr=0.010000
Epoch[030] Batch [3449]/[3760]	Speed: 35.525066 samples/sec	 accuracy=60.696558	 loss=1.631934	 lr=0.010000
Epoch[030] Batch [3499]/[3760]	Speed: 35.217805 samples/sec	 accuracy=60.680357	 loss=1.633133	 lr=0.010000
Epoch[030] Batch [3549]/[3760]	Speed: 35.318432 samples/sec	 accuracy=60.665053	 loss=1.633700	 lr=0.010000
Epoch[030] Batch [3599]/[3760]	Speed: 35.045259 samples/sec	 accuracy=60.667969	 loss=1.633934	 lr=0.010000
Epoch[030] Batch [3649]/[3760]	Speed: 35.284258 samples/sec	 accuracy=60.666952	 loss=1.633250	 lr=0.010000
Epoch[030] Batch [3699]/[3760]	Speed: 35.222892 samples/sec	 accuracy=60.661740	 loss=1.634037	 lr=0.010000
Epoch[030] Batch [3749]/[3760]	Speed: 36.426631 samples/sec	 accuracy=60.647500	 loss=1.634594	 lr=0.010000
Batch [0049]/[0303]: acc-top1=57.531250 acc-top5=80.093750
Batch [0099]/[0303]: acc-top1=57.859375 acc-top5=80.140625
Batch [0149]/[0303]: acc-top1=58.343750 acc-top5=80.552083
Batch [0199]/[0303]: acc-top1=58.164062 acc-top5=80.601562
Batch [0249]/[0303]: acc-top1=57.856250 acc-top5=80.443750
Batch [0299]/[0303]: acc-top1=57.781250 acc-top5=80.505208
[Epoch 030] training: accuracy=60.648271	 loss=1.634610
[Epoch 030] speed: 34 samples/sec	time cost: 7364.217300
[Epoch 030] validation: acc-top1=57.766089 acc-top5=80.497112 loss=1.992782
Epoch[031] Batch [0049]/[3760]	Speed: 23.966269 samples/sec	 accuracy=62.281250	 loss=1.577971	 lr=0.010000
Epoch[031] Batch [0099]/[3760]	Speed: 35.154980 samples/sec	 accuracy=62.171875	 loss=1.586721	 lr=0.010000
Epoch[031] Batch [0149]/[3760]	Speed: 34.942196 samples/sec	 accuracy=61.989583	 loss=1.576481	 lr=0.010000
Epoch[031] Batch [0199]/[3760]	Speed: 35.209672 samples/sec	 accuracy=61.882812	 loss=1.581845	 lr=0.010000
Epoch[031] Batch [0249]/[3760]	Speed: 35.291880 samples/sec	 accuracy=62.000000	 loss=1.574131	 lr=0.010000
Epoch[031] Batch [0299]/[3760]	Speed: 35.295235 samples/sec	 accuracy=62.046875	 loss=1.570178	 lr=0.010000
Epoch[031] Batch [0349]/[3760]	Speed: 35.216671 samples/sec	 accuracy=61.727679	 loss=1.577264	 lr=0.010000
Epoch[031] Batch [0399]/[3760]	Speed: 35.136143 samples/sec	 accuracy=61.718750	 loss=1.579489	 lr=0.010000
Epoch[031] Batch [0449]/[3760]	Speed: 35.284015 samples/sec	 accuracy=61.670139	 loss=1.581170	 lr=0.010000
Epoch[031] Batch [0499]/[3760]	Speed: 35.187011 samples/sec	 accuracy=61.800000	 loss=1.582626	 lr=0.010000
Epoch[031] Batch [0549]/[3760]	Speed: 35.414357 samples/sec	 accuracy=61.801136	 loss=1.581012	 lr=0.010000
Epoch[031] Batch [0599]/[3760]	Speed: 35.111444 samples/sec	 accuracy=61.648438	 loss=1.586822	 lr=0.010000
Epoch[031] Batch [0649]/[3760]	Speed: 35.413462 samples/sec	 accuracy=61.608173	 loss=1.587834	 lr=0.010000
Epoch[031] Batch [0699]/[3760]	Speed: 35.441717 samples/sec	 accuracy=61.553571	 loss=1.589791	 lr=0.010000
Epoch[031] Batch [0749]/[3760]	Speed: 35.310193 samples/sec	 accuracy=61.489583	 loss=1.592984	 lr=0.010000
Epoch[031] Batch [0799]/[3760]	Speed: 35.183494 samples/sec	 accuracy=61.396484	 loss=1.594299	 lr=0.010000
Epoch[031] Batch [0849]/[3760]	Speed: 35.185604 samples/sec	 accuracy=61.360294	 loss=1.595423	 lr=0.010000
Epoch[031] Batch [0899]/[3760]	Speed: 35.278965 samples/sec	 accuracy=61.447917	 loss=1.593383	 lr=0.010000
Epoch[031] Batch [0949]/[3760]	Speed: 35.331205 samples/sec	 accuracy=61.416118	 loss=1.595758	 lr=0.010000
Epoch[031] Batch [0999]/[3760]	Speed: 35.257401 samples/sec	 accuracy=61.393750	 loss=1.596595	 lr=0.010000
Epoch[031] Batch [1049]/[3760]	Speed: 35.410546 samples/sec	 accuracy=61.395833	 loss=1.597342	 lr=0.010000
Epoch[031] Batch [1099]/[3760]	Speed: 35.387701 samples/sec	 accuracy=61.428977	 loss=1.595749	 lr=0.010000
Epoch[031] Batch [1149]/[3760]	Speed: 35.198744 samples/sec	 accuracy=61.388587	 loss=1.597256	 lr=0.010000
Epoch[031] Batch [1199]/[3760]	Speed: 35.337820 samples/sec	 accuracy=61.414062	 loss=1.597379	 lr=0.010000
Epoch[031] Batch [1249]/[3760]	Speed: 35.392745 samples/sec	 accuracy=61.385000	 loss=1.598173	 lr=0.010000
Epoch[031] Batch [1299]/[3760]	Speed: 35.279542 samples/sec	 accuracy=61.342548	 loss=1.599990	 lr=0.010000
Epoch[031] Batch [1349]/[3760]	Speed: 35.340298 samples/sec	 accuracy=61.359954	 loss=1.601589	 lr=0.010000
Epoch[031] Batch [1399]/[3760]	Speed: 35.309741 samples/sec	 accuracy=61.415179	 loss=1.600284	 lr=0.010000
Epoch[031] Batch [1449]/[3760]	Speed: 35.019165 samples/sec	 accuracy=61.390086	 loss=1.601676	 lr=0.010000
Epoch[031] Batch [1499]/[3760]	Speed: 35.311591 samples/sec	 accuracy=61.342708	 loss=1.602633	 lr=0.010000
Epoch[031] Batch [1549]/[3760]	Speed: 34.904969 samples/sec	 accuracy=61.335685	 loss=1.603540	 lr=0.010000
Epoch[031] Batch [1599]/[3760]	Speed: 35.153933 samples/sec	 accuracy=61.299805	 loss=1.604433	 lr=0.010000
Epoch[031] Batch [1649]/[3760]	Speed: 35.042121 samples/sec	 accuracy=61.288826	 loss=1.605326	 lr=0.010000
Epoch[031] Batch [1699]/[3760]	Speed: 35.037054 samples/sec	 accuracy=61.265625	 loss=1.605990	 lr=0.010000
Epoch[031] Batch [1749]/[3760]	Speed: 35.281189 samples/sec	 accuracy=61.278571	 loss=1.604942	 lr=0.010000
Epoch[031] Batch [1799]/[3760]	Speed: 34.883433 samples/sec	 accuracy=61.283854	 loss=1.604183	 lr=0.010000
Epoch[031] Batch [1849]/[3760]	Speed: 35.198431 samples/sec	 accuracy=61.280405	 loss=1.604650	 lr=0.010000
Epoch[031] Batch [1899]/[3760]	Speed: 34.694007 samples/sec	 accuracy=61.282072	 loss=1.605862	 lr=0.010000
Epoch[031] Batch [1949]/[3760]	Speed: 35.470562 samples/sec	 accuracy=61.268429	 loss=1.606213	 lr=0.010000
Epoch[031] Batch [1999]/[3760]	Speed: 35.097422 samples/sec	 accuracy=61.239062	 loss=1.607024	 lr=0.010000
Epoch[031] Batch [2049]/[3760]	Speed: 35.497082 samples/sec	 accuracy=61.201220	 loss=1.608724	 lr=0.010000
Epoch[031] Batch [2099]/[3760]	Speed: 35.128376 samples/sec	 accuracy=61.159970	 loss=1.608449	 lr=0.010000
Epoch[031] Batch [2149]/[3760]	Speed: 35.372148 samples/sec	 accuracy=61.154070	 loss=1.608781	 lr=0.010000
Epoch[031] Batch [2199]/[3760]	Speed: 35.077611 samples/sec	 accuracy=61.151278	 loss=1.608987	 lr=0.010000
Epoch[031] Batch [2249]/[3760]	Speed: 35.351639 samples/sec	 accuracy=61.165278	 loss=1.609181	 lr=0.010000
Epoch[031] Batch [2299]/[3760]	Speed: 35.182316 samples/sec	 accuracy=61.135190	 loss=1.610054	 lr=0.010000
Epoch[031] Batch [2349]/[3760]	Speed: 35.435139 samples/sec	 accuracy=61.157580	 loss=1.609505	 lr=0.010000
Epoch[031] Batch [2399]/[3760]	Speed: 35.166099 samples/sec	 accuracy=61.169922	 loss=1.609752	 lr=0.010000
Epoch[031] Batch [2449]/[3760]	Speed: 35.379409 samples/sec	 accuracy=61.185587	 loss=1.609081	 lr=0.010000
Epoch[031] Batch [2499]/[3760]	Speed: 34.955524 samples/sec	 accuracy=61.195625	 loss=1.609191	 lr=0.010000
Epoch[031] Batch [2549]/[3760]	Speed: 35.274233 samples/sec	 accuracy=61.175858	 loss=1.610622	 lr=0.010000
Epoch[031] Batch [2599]/[3760]	Speed: 35.106311 samples/sec	 accuracy=61.170673	 loss=1.611079	 lr=0.010000
Epoch[031] Batch [2649]/[3760]	Speed: 35.342195 samples/sec	 accuracy=61.139741	 loss=1.612057	 lr=0.010000
Epoch[031] Batch [2699]/[3760]	Speed: 35.056208 samples/sec	 accuracy=61.130787	 loss=1.612810	 lr=0.010000
Epoch[031] Batch [2749]/[3760]	Speed: 35.192811 samples/sec	 accuracy=61.107955	 loss=1.613368	 lr=0.010000
Epoch[031] Batch [2799]/[3760]	Speed: 35.353457 samples/sec	 accuracy=61.095424	 loss=1.613783	 lr=0.010000
Epoch[031] Batch [2849]/[3760]	Speed: 35.213897 samples/sec	 accuracy=61.067434	 loss=1.614443	 lr=0.010000
Epoch[031] Batch [2899]/[3760]	Speed: 35.378191 samples/sec	 accuracy=61.045797	 loss=1.615581	 lr=0.010000
Epoch[031] Batch [2949]/[3760]	Speed: 35.400914 samples/sec	 accuracy=61.024364	 loss=1.616479	 lr=0.010000
Epoch[031] Batch [2999]/[3760]	Speed: 35.145615 samples/sec	 accuracy=61.054167	 loss=1.616316	 lr=0.010000
Epoch[031] Batch [3049]/[3760]	Speed: 35.369183 samples/sec	 accuracy=61.027152	 loss=1.616862	 lr=0.010000
Epoch[031] Batch [3099]/[3760]	Speed: 35.103391 samples/sec	 accuracy=61.011089	 loss=1.617470	 lr=0.010000
Epoch[031] Batch [3149]/[3760]	Speed: 35.319774 samples/sec	 accuracy=61.029266	 loss=1.617041	 lr=0.010000
Epoch[031] Batch [3199]/[3760]	Speed: 35.050002 samples/sec	 accuracy=61.020508	 loss=1.617721	 lr=0.010000
Epoch[031] Batch [3249]/[3760]	Speed: 35.220575 samples/sec	 accuracy=61.025000	 loss=1.617735	 lr=0.010000
Epoch[031] Batch [3299]/[3760]	Speed: 34.988239 samples/sec	 accuracy=61.023201	 loss=1.618049	 lr=0.010000
Epoch[031] Batch [3349]/[3760]	Speed: 35.156416 samples/sec	 accuracy=61.021922	 loss=1.618581	 lr=0.010000
Epoch[031] Batch [3399]/[3760]	Speed: 35.227499 samples/sec	 accuracy=61.023438	 loss=1.618436	 lr=0.010000
Epoch[031] Batch [3449]/[3760]	Speed: 35.222351 samples/sec	 accuracy=61.012681	 loss=1.618906	 lr=0.010000
Epoch[031] Batch [3499]/[3760]	Speed: 35.129294 samples/sec	 accuracy=61.014286	 loss=1.619275	 lr=0.010000
Epoch[031] Batch [3549]/[3760]	Speed: 35.020481 samples/sec	 accuracy=60.997359	 loss=1.619910	 lr=0.010000
Epoch[031] Batch [3599]/[3760]	Speed: 35.221942 samples/sec	 accuracy=61.001302	 loss=1.619965	 lr=0.010000
Epoch[031] Batch [3649]/[3760]	Speed: 34.716516 samples/sec	 accuracy=60.993151	 loss=1.619888	 lr=0.010000
Epoch[031] Batch [3699]/[3760]	Speed: 35.208169 samples/sec	 accuracy=60.980997	 loss=1.620840	 lr=0.010000
Epoch[031] Batch [3749]/[3760]	Speed: 36.070613 samples/sec	 accuracy=60.984167	 loss=1.620692	 lr=0.010000
Batch [0049]/[0303]: acc-top1=58.500000 acc-top5=81.656250
Batch [0099]/[0303]: acc-top1=58.375000 acc-top5=80.921875
Batch [0149]/[0303]: acc-top1=59.010417 acc-top5=81.260417
Batch [0199]/[0303]: acc-top1=58.796875 acc-top5=81.296875
Batch [0249]/[0303]: acc-top1=58.462500 acc-top5=81.012500
Batch [0299]/[0303]: acc-top1=58.322917 acc-top5=81.062500
[Epoch 031] training: accuracy=60.982380	 loss=1.620880
[Epoch 031] speed: 35 samples/sec	time cost: 7350.455176
[Epoch 031] validation: acc-top1=58.302393 acc-top5=81.069513 loss=1.954365
Epoch[032] Batch [0049]/[3759]	Speed: 24.448602 samples/sec	 accuracy=61.656250	 loss=1.571884	 lr=0.010000
Epoch[032] Batch [0099]/[3759]	Speed: 34.967920 samples/sec	 accuracy=62.187500	 loss=1.540841	 lr=0.010000
Epoch[032] Batch [0149]/[3759]	Speed: 35.025397 samples/sec	 accuracy=61.760417	 loss=1.567066	 lr=0.010000
Epoch[032] Batch [0199]/[3759]	Speed: 35.056578 samples/sec	 accuracy=61.992187	 loss=1.557527	 lr=0.010000
Epoch[032] Batch [0249]/[3759]	Speed: 35.108621 samples/sec	 accuracy=62.087500	 loss=1.552569	 lr=0.010000
Epoch[032] Batch [0299]/[3759]	Speed: 35.136353 samples/sec	 accuracy=61.833333	 loss=1.565958	 lr=0.010000
Epoch[032] Batch [0349]/[3759]	Speed: 35.283774 samples/sec	 accuracy=61.857143	 loss=1.565228	 lr=0.010000
Epoch[032] Batch [0399]/[3759]	Speed: 34.710232 samples/sec	 accuracy=61.800781	 loss=1.570323	 lr=0.010000
Epoch[032] Batch [0449]/[3759]	Speed: 35.403627 samples/sec	 accuracy=61.621528	 loss=1.580644	 lr=0.010000
Epoch[032] Batch [0499]/[3759]	Speed: 34.885776 samples/sec	 accuracy=61.756250	 loss=1.576424	 lr=0.010000
Epoch[032] Batch [0549]/[3759]	Speed: 35.471260 samples/sec	 accuracy=61.934659	 loss=1.573140	 lr=0.010000
Epoch[032] Batch [0599]/[3759]	Speed: 35.237662 samples/sec	 accuracy=61.885417	 loss=1.572713	 lr=0.010000
Epoch[032] Batch [0649]/[3759]	Speed: 35.331616 samples/sec	 accuracy=61.843750	 loss=1.573741	 lr=0.010000
Epoch[032] Batch [0699]/[3759]	Speed: 35.183208 samples/sec	 accuracy=61.765625	 loss=1.574932	 lr=0.010000
Epoch[032] Batch [0749]/[3759]	Speed: 35.303404 samples/sec	 accuracy=61.668750	 loss=1.577505	 lr=0.010000
Epoch[032] Batch [0799]/[3759]	Speed: 35.123838 samples/sec	 accuracy=61.697266	 loss=1.577673	 lr=0.010000
Epoch[032] Batch [0849]/[3759]	Speed: 35.467265 samples/sec	 accuracy=61.604779	 loss=1.582338	 lr=0.010000
Epoch[032] Batch [0899]/[3759]	Speed: 35.123403 samples/sec	 accuracy=61.522569	 loss=1.582423	 lr=0.010000
Epoch[032] Batch [0949]/[3759]	Speed: 35.356612 samples/sec	 accuracy=61.490132	 loss=1.586361	 lr=0.010000
Epoch[032] Batch [0999]/[3759]	Speed: 35.033867 samples/sec	 accuracy=61.410937	 loss=1.588392	 lr=0.010000
Epoch[032] Batch [1049]/[3759]	Speed: 35.510482 samples/sec	 accuracy=61.383929	 loss=1.591405	 lr=0.010000
Epoch[032] Batch [1099]/[3759]	Speed: 35.272002 samples/sec	 accuracy=61.377841	 loss=1.591374	 lr=0.010000
Epoch[032] Batch [1149]/[3759]	Speed: 35.296375 samples/sec	 accuracy=61.354620	 loss=1.591861	 lr=0.010000
Epoch[032] Batch [1199]/[3759]	Speed: 35.339775 samples/sec	 accuracy=61.315104	 loss=1.593395	 lr=0.010000
Epoch[032] Batch [1249]/[3759]	Speed: 35.339026 samples/sec	 accuracy=61.400000	 loss=1.591654	 lr=0.010000
Epoch[032] Batch [1299]/[3759]	Speed: 35.379394 samples/sec	 accuracy=61.348558	 loss=1.593612	 lr=0.010000
Epoch[032] Batch [1349]/[3759]	Speed: 35.167317 samples/sec	 accuracy=61.347222	 loss=1.592921	 lr=0.010000
Epoch[032] Batch [1399]/[3759]	Speed: 35.449352 samples/sec	 accuracy=61.373884	 loss=1.590897	 lr=0.010000
Epoch[032] Batch [1449]/[3759]	Speed: 35.297527 samples/sec	 accuracy=61.356681	 loss=1.591016	 lr=0.010000
Epoch[032] Batch [1499]/[3759]	Speed: 35.275103 samples/sec	 accuracy=61.346875	 loss=1.591706	 lr=0.010000
Epoch[032] Batch [1549]/[3759]	Speed: 35.391926 samples/sec	 accuracy=61.368952	 loss=1.592007	 lr=0.010000
Epoch[032] Batch [1599]/[3759]	Speed: 35.400348 samples/sec	 accuracy=61.405273	 loss=1.592497	 lr=0.010000
Epoch[032] Batch [1649]/[3759]	Speed: 35.282802 samples/sec	 accuracy=61.433712	 loss=1.592501	 lr=0.010000
Epoch[032] Batch [1699]/[3759]	Speed: 35.237037 samples/sec	 accuracy=61.441176	 loss=1.592578	 lr=0.010000
Epoch[032] Batch [1749]/[3759]	Speed: 35.128403 samples/sec	 accuracy=61.435714	 loss=1.593452	 lr=0.010000
Epoch[032] Batch [1799]/[3759]	Speed: 34.802636 samples/sec	 accuracy=61.417535	 loss=1.593732	 lr=0.010000
Epoch[032] Batch [1849]/[3759]	Speed: 35.199507 samples/sec	 accuracy=61.417230	 loss=1.594161	 lr=0.010000
Epoch[032] Batch [1899]/[3759]	Speed: 34.710739 samples/sec	 accuracy=61.379112	 loss=1.596185	 lr=0.010000
Epoch[032] Batch [1949]/[3759]	Speed: 35.464644 samples/sec	 accuracy=61.339744	 loss=1.597337	 lr=0.010000
Epoch[032] Batch [1999]/[3759]	Speed: 34.837035 samples/sec	 accuracy=61.342969	 loss=1.598066	 lr=0.010000
Epoch[032] Batch [2049]/[3759]	Speed: 35.268607 samples/sec	 accuracy=61.328506	 loss=1.598442	 lr=0.010000
Epoch[032] Batch [2099]/[3759]	Speed: 35.012292 samples/sec	 accuracy=61.316964	 loss=1.599685	 lr=0.010000
Epoch[032] Batch [2149]/[3759]	Speed: 35.195795 samples/sec	 accuracy=61.281250	 loss=1.600019	 lr=0.010000
Epoch[032] Batch [2199]/[3759]	Speed: 35.106248 samples/sec	 accuracy=61.253551	 loss=1.599992	 lr=0.010000
Epoch[032] Batch [2249]/[3759]	Speed: 34.978517 samples/sec	 accuracy=61.268056	 loss=1.600414	 lr=0.010000
Epoch[032] Batch [2299]/[3759]	Speed: 35.239240 samples/sec	 accuracy=61.283288	 loss=1.599531	 lr=0.010000
Epoch[032] Batch [2349]/[3759]	Speed: 35.105761 samples/sec	 accuracy=61.301197	 loss=1.598779	 lr=0.010000
Epoch[032] Batch [2399]/[3759]	Speed: 35.382066 samples/sec	 accuracy=61.272135	 loss=1.600814	 lr=0.010000
Epoch[032] Batch [2449]/[3759]	Speed: 35.229635 samples/sec	 accuracy=61.259566	 loss=1.601663	 lr=0.010000
Epoch[032] Batch [2499]/[3759]	Speed: 35.482087 samples/sec	 accuracy=61.258125	 loss=1.601951	 lr=0.010000
Epoch[032] Batch [2549]/[3759]	Speed: 35.284851 samples/sec	 accuracy=61.242647	 loss=1.602623	 lr=0.010000
Epoch[032] Batch [2599]/[3759]	Speed: 35.428850 samples/sec	 accuracy=61.230168	 loss=1.602930	 lr=0.010000
Epoch[032] Batch [2649]/[3759]	Speed: 35.528824 samples/sec	 accuracy=61.198113	 loss=1.603742	 lr=0.010000
Epoch[032] Batch [2699]/[3759]	Speed: 35.207329 samples/sec	 accuracy=61.191551	 loss=1.603596	 lr=0.010000
Epoch[032] Batch [2749]/[3759]	Speed: 35.292109 samples/sec	 accuracy=61.199432	 loss=1.603800	 lr=0.010000
Epoch[032] Batch [2799]/[3759]	Speed: 35.182357 samples/sec	 accuracy=61.179129	 loss=1.605162	 lr=0.010000
Epoch[032] Batch [2849]/[3759]	Speed: 35.418651 samples/sec	 accuracy=61.166667	 loss=1.605104	 lr=0.010000
Epoch[032] Batch [2899]/[3759]	Speed: 34.839425 samples/sec	 accuracy=61.159483	 loss=1.605274	 lr=0.010000
Epoch[032] Batch [2949]/[3759]	Speed: 35.431191 samples/sec	 accuracy=61.141419	 loss=1.606040	 lr=0.010000
Epoch[032] Batch [2999]/[3759]	Speed: 34.851305 samples/sec	 accuracy=61.113542	 loss=1.607327	 lr=0.010000
Epoch[032] Batch [3049]/[3759]	Speed: 35.251638 samples/sec	 accuracy=61.110656	 loss=1.607384	 lr=0.010000
Epoch[032] Batch [3099]/[3759]	Speed: 35.060122 samples/sec	 accuracy=61.100806	 loss=1.607839	 lr=0.010000
Epoch[032] Batch [3149]/[3759]	Speed: 35.036456 samples/sec	 accuracy=61.071429	 loss=1.609136	 lr=0.010000
Epoch[032] Batch [3199]/[3759]	Speed: 35.305547 samples/sec	 accuracy=61.065918	 loss=1.609436	 lr=0.010000
Epoch[032] Batch [3249]/[3759]	Speed: 34.992498 samples/sec	 accuracy=61.036538	 loss=1.610156	 lr=0.010000
Epoch[032] Batch [3299]/[3759]	Speed: 35.440174 samples/sec	 accuracy=61.011837	 loss=1.611590	 lr=0.010000
Epoch[032] Batch [3349]/[3759]	Speed: 35.181856 samples/sec	 accuracy=61.005131	 loss=1.612305	 lr=0.010000
Epoch[032] Batch [3399]/[3759]	Speed: 35.602472 samples/sec	 accuracy=60.976103	 loss=1.613444	 lr=0.010000
Epoch[032] Batch [3449]/[3759]	Speed: 35.025702 samples/sec	 accuracy=60.954257	 loss=1.614605	 lr=0.010000
Epoch[032] Batch [3499]/[3759]	Speed: 35.388621 samples/sec	 accuracy=60.957143	 loss=1.615134	 lr=0.010000
Epoch[032] Batch [3549]/[3759]	Speed: 35.065207 samples/sec	 accuracy=60.975352	 loss=1.615002	 lr=0.010000
Epoch[032] Batch [3599]/[3759]	Speed: 35.418155 samples/sec	 accuracy=60.969184	 loss=1.614970	 lr=0.010000
Epoch[032] Batch [3649]/[3759]	Speed: 35.257163 samples/sec	 accuracy=60.958048	 loss=1.615768	 lr=0.010000
Epoch[032] Batch [3699]/[3759]	Speed: 35.342160 samples/sec	 accuracy=60.948480	 loss=1.616067	 lr=0.010000
Epoch[032] Batch [3749]/[3759]	Speed: 36.562306 samples/sec	 accuracy=60.951667	 loss=1.615720	 lr=0.010000
Batch [0049]/[0303]: acc-top1=57.218750 acc-top5=79.843750
Batch [0099]/[0303]: acc-top1=58.078125 acc-top5=80.078125
Batch [0149]/[0303]: acc-top1=58.020833 acc-top5=80.291667
Batch [0199]/[0303]: acc-top1=57.945312 acc-top5=80.257812
Batch [0249]/[0303]: acc-top1=57.800000 acc-top5=80.112500
Batch [0299]/[0303]: acc-top1=57.661458 acc-top5=80.192708
[Epoch 032] training: accuracy=60.944151	 loss=1.615939
[Epoch 032] speed: 35 samples/sec	time cost: 7342.919447
[Epoch 032] validation: acc-top1=57.616543 acc-top5=80.167079 loss=1.988521
Epoch[033] Batch [0049]/[3760]	Speed: 24.338604 samples/sec	 accuracy=63.406250	 loss=1.511218	 lr=0.010000
Epoch[033] Batch [0099]/[3760]	Speed: 34.959469 samples/sec	 accuracy=62.656250	 loss=1.548717	 lr=0.010000
Epoch[033] Batch [0149]/[3760]	Speed: 34.959402 samples/sec	 accuracy=62.250000	 loss=1.543449	 lr=0.010000
Epoch[033] Batch [0199]/[3760]	Speed: 35.028465 samples/sec	 accuracy=62.273437	 loss=1.539688	 lr=0.010000
Epoch[033] Batch [0249]/[3760]	Speed: 35.298317 samples/sec	 accuracy=62.125000	 loss=1.544072	 lr=0.010000
Epoch[033] Batch [0299]/[3760]	Speed: 34.997116 samples/sec	 accuracy=62.177083	 loss=1.549468	 lr=0.010000
Epoch[033] Batch [0349]/[3760]	Speed: 35.377843 samples/sec	 accuracy=62.008929	 loss=1.553971	 lr=0.010000
Epoch[033] Batch [0399]/[3760]	Speed: 34.964116 samples/sec	 accuracy=62.011719	 loss=1.557577	 lr=0.010000
Epoch[033] Batch [0449]/[3760]	Speed: 35.424233 samples/sec	 accuracy=62.083333	 loss=1.557872	 lr=0.010000
Epoch[033] Batch [0499]/[3760]	Speed: 35.164755 samples/sec	 accuracy=62.071875	 loss=1.559416	 lr=0.010000
Epoch[033] Batch [0549]/[3760]	Speed: 35.371527 samples/sec	 accuracy=61.963068	 loss=1.561112	 lr=0.010000
Epoch[033] Batch [0599]/[3760]	Speed: 34.952763 samples/sec	 accuracy=61.835938	 loss=1.564961	 lr=0.010000
Epoch[033] Batch [0649]/[3760]	Speed: 35.297262 samples/sec	 accuracy=61.826923	 loss=1.570457	 lr=0.010000
Epoch[033] Batch [0699]/[3760]	Speed: 35.026994 samples/sec	 accuracy=61.837054	 loss=1.570126	 lr=0.010000
Epoch[033] Batch [0749]/[3760]	Speed: 35.280918 samples/sec	 accuracy=61.820833	 loss=1.569721	 lr=0.010000
Epoch[033] Batch [0799]/[3760]	Speed: 35.359663 samples/sec	 accuracy=61.796875	 loss=1.572006	 lr=0.010000
Epoch[033] Batch [0849]/[3760]	Speed: 35.105457 samples/sec	 accuracy=61.750000	 loss=1.574548	 lr=0.010000
Epoch[033] Batch [0899]/[3760]	Speed: 35.369496 samples/sec	 accuracy=61.701389	 loss=1.576205	 lr=0.010000
Epoch[033] Batch [0949]/[3760]	Speed: 35.143254 samples/sec	 accuracy=61.768092	 loss=1.575621	 lr=0.010000
Epoch[033] Batch [0999]/[3760]	Speed: 35.198475 samples/sec	 accuracy=61.734375	 loss=1.578161	 lr=0.010000
Epoch[033] Batch [1049]/[3760]	Speed: 35.261947 samples/sec	 accuracy=61.690476	 loss=1.579995	 lr=0.010000
Epoch[033] Batch [1099]/[3760]	Speed: 35.352943 samples/sec	 accuracy=61.684659	 loss=1.582170	 lr=0.010000
Epoch[033] Batch [1149]/[3760]	Speed: 35.104040 samples/sec	 accuracy=61.672554	 loss=1.581734	 lr=0.010000
Epoch[033] Batch [1199]/[3760]	Speed: 35.212315 samples/sec	 accuracy=61.589844	 loss=1.585178	 lr=0.010000
Epoch[033] Batch [1249]/[3760]	Speed: 35.156061 samples/sec	 accuracy=61.610000	 loss=1.586145	 lr=0.010000
Epoch[033] Batch [1299]/[3760]	Speed: 35.196376 samples/sec	 accuracy=61.584135	 loss=1.589723	 lr=0.010000
Epoch[033] Batch [1349]/[3760]	Speed: 35.277190 samples/sec	 accuracy=61.553241	 loss=1.591721	 lr=0.010000
Epoch[033] Batch [1399]/[3760]	Speed: 35.362374 samples/sec	 accuracy=61.505580	 loss=1.593559	 lr=0.010000
Epoch[033] Batch [1449]/[3760]	Speed: 35.315539 samples/sec	 accuracy=61.529095	 loss=1.593272	 lr=0.010000
Epoch[033] Batch [1499]/[3760]	Speed: 35.296228 samples/sec	 accuracy=61.502083	 loss=1.594885	 lr=0.010000
Epoch[033] Batch [1549]/[3760]	Speed: 35.407518 samples/sec	 accuracy=61.465726	 loss=1.595324	 lr=0.010000
Epoch[033] Batch [1599]/[3760]	Speed: 35.283915 samples/sec	 accuracy=61.403320	 loss=1.598156	 lr=0.010000
Epoch[033] Batch [1649]/[3760]	Speed: 35.347818 samples/sec	 accuracy=61.411932	 loss=1.597601	 lr=0.010000
Epoch[033] Batch [1699]/[3760]	Speed: 35.332237 samples/sec	 accuracy=61.420956	 loss=1.596890	 lr=0.010000
Epoch[033] Batch [1749]/[3760]	Speed: 35.258991 samples/sec	 accuracy=61.450000	 loss=1.596592	 lr=0.010000
Epoch[033] Batch [1799]/[3760]	Speed: 35.148185 samples/sec	 accuracy=61.430556	 loss=1.597198	 lr=0.010000
Epoch[033] Batch [1849]/[3760]	Speed: 35.124766 samples/sec	 accuracy=61.432432	 loss=1.596818	 lr=0.010000
Epoch[033] Batch [1899]/[3760]	Speed: 35.129553 samples/sec	 accuracy=61.431743	 loss=1.597258	 lr=0.010000
Epoch[033] Batch [1949]/[3760]	Speed: 35.288929 samples/sec	 accuracy=61.438301	 loss=1.597246	 lr=0.010000
Epoch[033] Batch [1999]/[3760]	Speed: 35.256884 samples/sec	 accuracy=61.421094	 loss=1.598543	 lr=0.010000
Epoch[033] Batch [2049]/[3760]	Speed: 34.935839 samples/sec	 accuracy=61.419207	 loss=1.598293	 lr=0.010000
Epoch[033] Batch [2099]/[3760]	Speed: 35.447592 samples/sec	 accuracy=61.377976	 loss=1.599476	 lr=0.010000
Epoch[033] Batch [2149]/[3760]	Speed: 34.999832 samples/sec	 accuracy=61.332849	 loss=1.600916	 lr=0.010000
Epoch[033] Batch [2199]/[3760]	Speed: 35.462325 samples/sec	 accuracy=61.330256	 loss=1.600626	 lr=0.010000
Epoch[033] Batch [2249]/[3760]	Speed: 34.938703 samples/sec	 accuracy=61.354167	 loss=1.599994	 lr=0.010000
Epoch[033] Batch [2299]/[3760]	Speed: 35.481438 samples/sec	 accuracy=61.322011	 loss=1.601680	 lr=0.010000
Epoch[033] Batch [2349]/[3760]	Speed: 35.169883 samples/sec	 accuracy=61.315824	 loss=1.602067	 lr=0.010000
Epoch[033] Batch [2399]/[3760]	Speed: 35.395450 samples/sec	 accuracy=61.315755	 loss=1.602839	 lr=0.010000
Epoch[033] Batch [2449]/[3760]	Speed: 35.289729 samples/sec	 accuracy=61.262755	 loss=1.604492	 lr=0.010000
Epoch[033] Batch [2499]/[3760]	Speed: 35.543437 samples/sec	 accuracy=61.225625	 loss=1.605305	 lr=0.010000
Epoch[033] Batch [2549]/[3760]	Speed: 35.318121 samples/sec	 accuracy=61.258578	 loss=1.603930	 lr=0.010000
Epoch[033] Batch [2599]/[3760]	Speed: 35.136073 samples/sec	 accuracy=61.268630	 loss=1.603615	 lr=0.010000
Epoch[033] Batch [2649]/[3760]	Speed: 35.087253 samples/sec	 accuracy=61.219929	 loss=1.605737	 lr=0.010000
Epoch[033] Batch [2699]/[3760]	Speed: 35.104959 samples/sec	 accuracy=61.203125	 loss=1.606722	 lr=0.010000
Epoch[033] Batch [2749]/[3760]	Speed: 35.447010 samples/sec	 accuracy=61.193750	 loss=1.606748	 lr=0.010000
Epoch[033] Batch [2799]/[3760]	Speed: 35.196210 samples/sec	 accuracy=61.186384	 loss=1.607171	 lr=0.010000
Epoch[033] Batch [2849]/[3760]	Speed: 35.262715 samples/sec	 accuracy=61.190789	 loss=1.607247	 lr=0.010000
Epoch[033] Batch [2899]/[3760]	Speed: 35.249845 samples/sec	 accuracy=61.192349	 loss=1.606696	 lr=0.010000
Epoch[033] Batch [2949]/[3760]	Speed: 35.523538 samples/sec	 accuracy=61.181144	 loss=1.607178	 lr=0.010000
Epoch[033] Batch [2999]/[3760]	Speed: 35.163919 samples/sec	 accuracy=61.201562	 loss=1.606006	 lr=0.010000
Epoch[033] Batch [3049]/[3760]	Speed: 35.448480 samples/sec	 accuracy=61.201332	 loss=1.606228	 lr=0.010000
Epoch[033] Batch [3099]/[3760]	Speed: 35.246709 samples/sec	 accuracy=61.179940	 loss=1.606235	 lr=0.010000
Epoch[033] Batch [3149]/[3760]	Speed: 35.283916 samples/sec	 accuracy=61.209821	 loss=1.605236	 lr=0.010000
Epoch[033] Batch [3199]/[3760]	Speed: 35.208448 samples/sec	 accuracy=61.208008	 loss=1.605356	 lr=0.010000
Epoch[033] Batch [3249]/[3760]	Speed: 35.516051 samples/sec	 accuracy=61.189904	 loss=1.605691	 lr=0.010000
Epoch[033] Batch [3299]/[3760]	Speed: 35.366690 samples/sec	 accuracy=61.189394	 loss=1.606036	 lr=0.010000
Epoch[033] Batch [3349]/[3760]	Speed: 35.438892 samples/sec	 accuracy=61.169776	 loss=1.606446	 lr=0.010000
Epoch[033] Batch [3399]/[3760]	Speed: 35.429888 samples/sec	 accuracy=61.190717	 loss=1.605352	 lr=0.010000
Epoch[033] Batch [3449]/[3760]	Speed: 35.379797 samples/sec	 accuracy=61.188406	 loss=1.605340	 lr=0.010000
Epoch[033] Batch [3499]/[3760]	Speed: 35.425527 samples/sec	 accuracy=61.182589	 loss=1.605418	 lr=0.010000
Epoch[033] Batch [3549]/[3760]	Speed: 35.248571 samples/sec	 accuracy=61.182218	 loss=1.605601	 lr=0.010000
Epoch[033] Batch [3599]/[3760]	Speed: 35.480591 samples/sec	 accuracy=61.167101	 loss=1.606048	 lr=0.010000
Epoch[033] Batch [3649]/[3760]	Speed: 35.372921 samples/sec	 accuracy=61.181935	 loss=1.605355	 lr=0.010000
Epoch[033] Batch [3699]/[3760]	Speed: 35.332950 samples/sec	 accuracy=61.181588	 loss=1.605670	 lr=0.010000
Epoch[033] Batch [3749]/[3760]	Speed: 36.410392 samples/sec	 accuracy=61.159167	 loss=1.606911	 lr=0.010000
Batch [0049]/[0303]: acc-top1=57.468750 acc-top5=80.500000
Batch [0099]/[0303]: acc-top1=57.765625 acc-top5=79.968750
Batch [0149]/[0303]: acc-top1=58.114583 acc-top5=80.552083
Batch [0199]/[0303]: acc-top1=57.984375 acc-top5=80.554688
Batch [0249]/[0303]: acc-top1=57.581250 acc-top5=80.337500
Batch [0299]/[0303]: acc-top1=57.333333 acc-top5=80.291667
[Epoch 033] training: accuracy=61.153590	 loss=1.607113
[Epoch 033] speed: 35 samples/sec	time cost: 7336.611291
[Epoch 033] validation: acc-top1=57.322607 acc-top5=80.265058 loss=2.005023
Epoch[034] Batch [0049]/[3759]	Speed: 24.651762 samples/sec	 accuracy=61.093750	 loss=1.586212	 lr=0.010000
Epoch[034] Batch [0099]/[3759]	Speed: 35.185600 samples/sec	 accuracy=61.593750	 loss=1.569697	 lr=0.010000
Epoch[034] Batch [0149]/[3759]	Speed: 34.994248 samples/sec	 accuracy=61.916667	 loss=1.561493	 lr=0.010000
Epoch[034] Batch [0199]/[3759]	Speed: 34.984988 samples/sec	 accuracy=61.484375	 loss=1.572923	 lr=0.010000
Epoch[034] Batch [0249]/[3759]	Speed: 35.248874 samples/sec	 accuracy=61.650000	 loss=1.563178	 lr=0.010000
Epoch[034] Batch [0299]/[3759]	Speed: 35.188880 samples/sec	 accuracy=61.598958	 loss=1.563001	 lr=0.010000
Epoch[034] Batch [0349]/[3759]	Speed: 35.350829 samples/sec	 accuracy=61.544643	 loss=1.569381	 lr=0.010000
Epoch[034] Batch [0399]/[3759]	Speed: 35.115956 samples/sec	 accuracy=61.546875	 loss=1.567244	 lr=0.010000
Epoch[034] Batch [0449]/[3759]	Speed: 35.577910 samples/sec	 accuracy=61.538194	 loss=1.568823	 lr=0.010000
Epoch[034] Batch [0499]/[3759]	Speed: 35.155395 samples/sec	 accuracy=61.550000	 loss=1.571145	 lr=0.010000
Epoch[034] Batch [0549]/[3759]	Speed: 35.412208 samples/sec	 accuracy=61.573864	 loss=1.573246	 lr=0.010000
Epoch[034] Batch [0599]/[3759]	Speed: 35.190716 samples/sec	 accuracy=61.593750	 loss=1.575095	 lr=0.010000
Epoch[034] Batch [0649]/[3759]	Speed: 35.424698 samples/sec	 accuracy=61.634615	 loss=1.577058	 lr=0.010000
Epoch[034] Batch [0699]/[3759]	Speed: 35.182922 samples/sec	 accuracy=61.609375	 loss=1.578554	 lr=0.010000
Epoch[034] Batch [0749]/[3759]	Speed: 35.320378 samples/sec	 accuracy=61.647917	 loss=1.576211	 lr=0.010000
Epoch[034] Batch [0799]/[3759]	Speed: 35.121276 samples/sec	 accuracy=61.724609	 loss=1.577369	 lr=0.010000
Epoch[034] Batch [0849]/[3759]	Speed: 35.164382 samples/sec	 accuracy=61.704044	 loss=1.581467	 lr=0.010000
Epoch[034] Batch [0899]/[3759]	Speed: 35.325789 samples/sec	 accuracy=61.762153	 loss=1.578769	 lr=0.010000
Epoch[034] Batch [0949]/[3759]	Speed: 35.204858 samples/sec	 accuracy=61.677632	 loss=1.580985	 lr=0.010000
Epoch[034] Batch [0999]/[3759]	Speed: 35.270421 samples/sec	 accuracy=61.707813	 loss=1.582109	 lr=0.010000
Epoch[034] Batch [1049]/[3759]	Speed: 35.313087 samples/sec	 accuracy=61.671131	 loss=1.582610	 lr=0.010000
Epoch[034] Batch [1099]/[3759]	Speed: 35.298194 samples/sec	 accuracy=61.704545	 loss=1.583095	 lr=0.010000
Epoch[034] Batch [1149]/[3759]	Speed: 35.338011 samples/sec	 accuracy=61.737772	 loss=1.581692	 lr=0.010000
Epoch[034] Batch [1199]/[3759]	Speed: 35.528715 samples/sec	 accuracy=61.682292	 loss=1.582496	 lr=0.010000
Epoch[034] Batch [1249]/[3759]	Speed: 35.356958 samples/sec	 accuracy=61.692500	 loss=1.580328	 lr=0.010000
Epoch[034] Batch [1299]/[3759]	Speed: 35.103678 samples/sec	 accuracy=61.677885	 loss=1.580270	 lr=0.010000
Epoch[034] Batch [1349]/[3759]	Speed: 35.194309 samples/sec	 accuracy=61.680556	 loss=1.579208	 lr=0.010000
Epoch[034] Batch [1399]/[3759]	Speed: 34.991010 samples/sec	 accuracy=61.657366	 loss=1.580651	 lr=0.010000
Epoch[034] Batch [1449]/[3759]	Speed: 34.904742 samples/sec	 accuracy=61.698276	 loss=1.579950	 lr=0.010000
Epoch[034] Batch [1499]/[3759]	Speed: 35.197436 samples/sec	 accuracy=61.706250	 loss=1.579955	 lr=0.010000
Epoch[034] Batch [1549]/[3759]	Speed: 34.871362 samples/sec	 accuracy=61.688508	 loss=1.580863	 lr=0.010000
Epoch[034] Batch [1599]/[3759]	Speed: 35.222659 samples/sec	 accuracy=61.702148	 loss=1.580272	 lr=0.010000
Epoch[034] Batch [1649]/[3759]	Speed: 34.954606 samples/sec	 accuracy=61.721591	 loss=1.579415	 lr=0.010000
Epoch[034] Batch [1699]/[3759]	Speed: 35.564165 samples/sec	 accuracy=61.713235	 loss=1.581126	 lr=0.010000
Epoch[034] Batch [1749]/[3759]	Speed: 34.928494 samples/sec	 accuracy=61.737500	 loss=1.579750	 lr=0.010000
Epoch[034] Batch [1799]/[3759]	Speed: 35.336825 samples/sec	 accuracy=61.680556	 loss=1.582064	 lr=0.010000
Epoch[034] Batch [1849]/[3759]	Speed: 34.742185 samples/sec	 accuracy=61.746622	 loss=1.580324	 lr=0.010000
Epoch[034] Batch [1899]/[3759]	Speed: 35.169520 samples/sec	 accuracy=61.707237	 loss=1.582738	 lr=0.010000
Epoch[034] Batch [1949]/[3759]	Speed: 35.205213 samples/sec	 accuracy=61.694712	 loss=1.582656	 lr=0.010000
Epoch[034] Batch [1999]/[3759]	Speed: 35.208727 samples/sec	 accuracy=61.674219	 loss=1.582165	 lr=0.010000
Epoch[034] Batch [2049]/[3759]	Speed: 34.873770 samples/sec	 accuracy=61.646341	 loss=1.582182	 lr=0.010000
Epoch[034] Batch [2099]/[3759]	Speed: 34.943378 samples/sec	 accuracy=61.645089	 loss=1.583022	 lr=0.010000
Epoch[034] Batch [2149]/[3759]	Speed: 35.158810 samples/sec	 accuracy=61.650436	 loss=1.582902	 lr=0.010000
Epoch[034] Batch [2199]/[3759]	Speed: 34.723587 samples/sec	 accuracy=61.636364	 loss=1.584174	 lr=0.010000
Epoch[034] Batch [2249]/[3759]	Speed: 35.540061 samples/sec	 accuracy=61.654167	 loss=1.585038	 lr=0.010000
Epoch[034] Batch [2299]/[3759]	Speed: 34.866563 samples/sec	 accuracy=61.629076	 loss=1.585587	 lr=0.010000
Epoch[034] Batch [2349]/[3759]	Speed: 35.464288 samples/sec	 accuracy=61.619016	 loss=1.585783	 lr=0.010000
Epoch[034] Batch [2399]/[3759]	Speed: 35.044436 samples/sec	 accuracy=61.572266	 loss=1.587252	 lr=0.010000
Epoch[034] Batch [2449]/[3759]	Speed: 35.388080 samples/sec	 accuracy=61.556122	 loss=1.588116	 lr=0.010000
Epoch[034] Batch [2499]/[3759]	Speed: 35.159534 samples/sec	 accuracy=61.563750	 loss=1.587610	 lr=0.010000
Epoch[034] Batch [2549]/[3759]	Speed: 35.185255 samples/sec	 accuracy=61.526348	 loss=1.589105	 lr=0.010000
Epoch[034] Batch [2599]/[3759]	Speed: 35.305382 samples/sec	 accuracy=61.548077	 loss=1.588630	 lr=0.010000
Epoch[034] Batch [2649]/[3759]	Speed: 35.158608 samples/sec	 accuracy=61.530660	 loss=1.589248	 lr=0.010000
Epoch[034] Batch [2699]/[3759]	Speed: 35.255233 samples/sec	 accuracy=61.523148	 loss=1.589251	 lr=0.010000
Epoch[034] Batch [2749]/[3759]	Speed: 35.138917 samples/sec	 accuracy=61.488636	 loss=1.590327	 lr=0.010000
Epoch[034] Batch [2799]/[3759]	Speed: 35.280687 samples/sec	 accuracy=61.495536	 loss=1.590852	 lr=0.010000
Epoch[034] Batch [2849]/[3759]	Speed: 35.198607 samples/sec	 accuracy=61.490680	 loss=1.591533	 lr=0.010000
Epoch[034] Batch [2899]/[3759]	Speed: 35.414678 samples/sec	 accuracy=61.489224	 loss=1.591856	 lr=0.010000
Epoch[034] Batch [2949]/[3759]	Speed: 35.016471 samples/sec	 accuracy=61.459746	 loss=1.593212	 lr=0.010000
Epoch[034] Batch [2999]/[3759]	Speed: 35.373080 samples/sec	 accuracy=61.436458	 loss=1.593520	 lr=0.010000
Epoch[034] Batch [3049]/[3759]	Speed: 35.014285 samples/sec	 accuracy=61.428791	 loss=1.594067	 lr=0.010000
Epoch[034] Batch [3099]/[3759]	Speed: 35.615263 samples/sec	 accuracy=61.453125	 loss=1.594095	 lr=0.010000
Epoch[034] Batch [3149]/[3759]	Speed: 35.155103 samples/sec	 accuracy=61.430556	 loss=1.595454	 lr=0.010000
Epoch[034] Batch [3199]/[3759]	Speed: 35.201154 samples/sec	 accuracy=61.440918	 loss=1.594965	 lr=0.010000
Epoch[034] Batch [3249]/[3759]	Speed: 35.053093 samples/sec	 accuracy=61.449038	 loss=1.594990	 lr=0.010000
Epoch[034] Batch [3299]/[3759]	Speed: 35.453028 samples/sec	 accuracy=61.429451	 loss=1.595375	 lr=0.010000
Epoch[034] Batch [3349]/[3759]	Speed: 35.007593 samples/sec	 accuracy=61.432369	 loss=1.595274	 lr=0.010000
Epoch[034] Batch [3399]/[3759]	Speed: 35.285168 samples/sec	 accuracy=61.446232	 loss=1.594535	 lr=0.010000
Epoch[034] Batch [3449]/[3759]	Speed: 35.258085 samples/sec	 accuracy=61.435688	 loss=1.595000	 lr=0.010000
Epoch[034] Batch [3499]/[3759]	Speed: 35.234307 samples/sec	 accuracy=61.430804	 loss=1.595111	 lr=0.010000
Epoch[034] Batch [3549]/[3759]	Speed: 35.304098 samples/sec	 accuracy=61.416373	 loss=1.595574	 lr=0.010000
Epoch[034] Batch [3599]/[3759]	Speed: 35.219684 samples/sec	 accuracy=61.416667	 loss=1.595975	 lr=0.010000
Epoch[034] Batch [3649]/[3759]	Speed: 35.318082 samples/sec	 accuracy=61.388699	 loss=1.597416	 lr=0.010000
Epoch[034] Batch [3699]/[3759]	Speed: 35.104177 samples/sec	 accuracy=61.397804	 loss=1.597645	 lr=0.010000
Epoch[034] Batch [3749]/[3759]	Speed: 36.666302 samples/sec	 accuracy=61.373333	 loss=1.598508	 lr=0.010000
Batch [0049]/[0303]: acc-top1=58.531250 acc-top5=80.687500
Batch [0099]/[0303]: acc-top1=58.453125 acc-top5=80.937500
Batch [0149]/[0303]: acc-top1=58.708333 acc-top5=81.031250
Batch [0199]/[0303]: acc-top1=58.437500 acc-top5=81.046875
Batch [0249]/[0303]: acc-top1=58.231250 acc-top5=80.793750
Batch [0299]/[0303]: acc-top1=58.171875 acc-top5=80.822917
[Epoch 034] training: accuracy=61.366886	 loss=1.599005
[Epoch 034] speed: 35 samples/sec	time cost: 7344.257136
[Epoch 034] validation: acc-top1=58.106436 acc-top5=80.832302 loss=1.990331
Epoch[035] Batch [0049]/[3760]	Speed: 24.083720 samples/sec	 accuracy=60.781250	 loss=1.646813	 lr=0.010000
Epoch[035] Batch [0099]/[3760]	Speed: 35.128403 samples/sec	 accuracy=62.250000	 loss=1.570814	 lr=0.010000
Epoch[035] Batch [0149]/[3760]	Speed: 34.987105 samples/sec	 accuracy=62.791667	 loss=1.542551	 lr=0.010000
Epoch[035] Batch [0199]/[3760]	Speed: 35.203314 samples/sec	 accuracy=62.546875	 loss=1.554986	 lr=0.010000
Epoch[035] Batch [0249]/[3760]	Speed: 35.252927 samples/sec	 accuracy=62.456250	 loss=1.558403	 lr=0.010000
Epoch[035] Batch [0299]/[3760]	Speed: 35.057721 samples/sec	 accuracy=62.479167	 loss=1.562838	 lr=0.010000
Epoch[035] Batch [0349]/[3760]	Speed: 35.353452 samples/sec	 accuracy=62.571429	 loss=1.557947	 lr=0.010000
Epoch[035] Batch [0399]/[3760]	Speed: 34.919530 samples/sec	 accuracy=62.539062	 loss=1.559980	 lr=0.010000
Epoch[035] Batch [0449]/[3760]	Speed: 35.539094 samples/sec	 accuracy=62.357639	 loss=1.564621	 lr=0.010000
Epoch[035] Batch [0499]/[3760]	Speed: 35.015577 samples/sec	 accuracy=62.240625	 loss=1.567036	 lr=0.010000
Epoch[035] Batch [0549]/[3760]	Speed: 35.540919 samples/sec	 accuracy=62.159091	 loss=1.571330	 lr=0.010000
Epoch[035] Batch [0599]/[3760]	Speed: 34.849747 samples/sec	 accuracy=62.080729	 loss=1.575349	 lr=0.010000
Epoch[035] Batch [0649]/[3760]	Speed: 35.516171 samples/sec	 accuracy=62.122596	 loss=1.573927	 lr=0.010000
Epoch[035] Batch [0699]/[3760]	Speed: 34.920757 samples/sec	 accuracy=62.154018	 loss=1.570664	 lr=0.010000
Epoch[035] Batch [0749]/[3760]	Speed: 35.394286 samples/sec	 accuracy=62.187500	 loss=1.569812	 lr=0.010000
Epoch[035] Batch [0799]/[3760]	Speed: 35.176236 samples/sec	 accuracy=62.244141	 loss=1.566645	 lr=0.010000
Epoch[035] Batch [0849]/[3760]	Speed: 35.460973 samples/sec	 accuracy=62.222426	 loss=1.567423	 lr=0.010000
Epoch[035] Batch [0899]/[3760]	Speed: 35.195214 samples/sec	 accuracy=62.244792	 loss=1.567696	 lr=0.010000
Epoch[035] Batch [0949]/[3760]	Speed: 35.274258 samples/sec	 accuracy=62.167763	 loss=1.568916	 lr=0.010000
Epoch[035] Batch [0999]/[3760]	Speed: 35.169448 samples/sec	 accuracy=62.207813	 loss=1.567345	 lr=0.010000
Epoch[035] Batch [1049]/[3760]	Speed: 35.321854 samples/sec	 accuracy=62.254464	 loss=1.566498	 lr=0.010000
Epoch[035] Batch [1099]/[3760]	Speed: 35.274756 samples/sec	 accuracy=62.193182	 loss=1.569087	 lr=0.010000
Epoch[035] Batch [1149]/[3760]	Speed: 35.131156 samples/sec	 accuracy=62.191576	 loss=1.571342	 lr=0.010000
Epoch[035] Batch [1199]/[3760]	Speed: 35.234840 samples/sec	 accuracy=62.190104	 loss=1.571196	 lr=0.010000
Epoch[035] Batch [1249]/[3760]	Speed: 34.956719 samples/sec	 accuracy=62.130000	 loss=1.571709	 lr=0.010000
Epoch[035] Batch [1299]/[3760]	Speed: 35.279392 samples/sec	 accuracy=62.082933	 loss=1.572486	 lr=0.010000
Epoch[035] Batch [1349]/[3760]	Speed: 34.679401 samples/sec	 accuracy=62.057870	 loss=1.574534	 lr=0.010000
Epoch[035] Batch [1399]/[3760]	Speed: 35.328029 samples/sec	 accuracy=62.008929	 loss=1.576433	 lr=0.010000
Epoch[035] Batch [1449]/[3760]	Speed: 35.241450 samples/sec	 accuracy=61.968750	 loss=1.576985	 lr=0.010000
Epoch[035] Batch [1499]/[3760]	Speed: 35.337099 samples/sec	 accuracy=62.020833	 loss=1.575921	 lr=0.010000
Epoch[035] Batch [1549]/[3760]	Speed: 35.201124 samples/sec	 accuracy=62.014113	 loss=1.576262	 lr=0.010000
Epoch[035] Batch [1599]/[3760]	Speed: 35.254301 samples/sec	 accuracy=62.006836	 loss=1.576520	 lr=0.010000
Epoch[035] Batch [1649]/[3760]	Speed: 35.230181 samples/sec	 accuracy=61.995265	 loss=1.576891	 lr=0.010000
Epoch[035] Batch [1699]/[3760]	Speed: 35.266226 samples/sec	 accuracy=61.984375	 loss=1.577326	 lr=0.010000
Epoch[035] Batch [1749]/[3760]	Speed: 35.402306 samples/sec	 accuracy=61.989286	 loss=1.576823	 lr=0.010000
Epoch[035] Batch [1799]/[3760]	Speed: 35.267980 samples/sec	 accuracy=61.945313	 loss=1.578606	 lr=0.010000
Epoch[035] Batch [1849]/[3760]	Speed: 35.325813 samples/sec	 accuracy=61.985642	 loss=1.577243	 lr=0.010000
Epoch[035] Batch [1899]/[3760]	Speed: 35.390048 samples/sec	 accuracy=61.970395	 loss=1.577284	 lr=0.010000
Epoch[035] Batch [1949]/[3760]	Speed: 35.587437 samples/sec	 accuracy=61.963141	 loss=1.577073	 lr=0.010000
Epoch[035] Batch [1999]/[3760]	Speed: 35.306745 samples/sec	 accuracy=61.923437	 loss=1.578488	 lr=0.010000
Epoch[035] Batch [2049]/[3760]	Speed: 35.492527 samples/sec	 accuracy=61.892530	 loss=1.579678	 lr=0.010000
Epoch[035] Batch [2099]/[3760]	Speed: 35.166158 samples/sec	 accuracy=61.881696	 loss=1.578808	 lr=0.010000
Epoch[035] Batch [2149]/[3760]	Speed: 35.549347 samples/sec	 accuracy=61.888808	 loss=1.578302	 lr=0.010000
Epoch[035] Batch [2199]/[3760]	Speed: 35.536049 samples/sec	 accuracy=61.846591	 loss=1.579782	 lr=0.010000
Epoch[035] Batch [2249]/[3760]	Speed: 35.267832 samples/sec	 accuracy=61.856250	 loss=1.579236	 lr=0.010000
Epoch[035] Batch [2299]/[3760]	Speed: 35.389609 samples/sec	 accuracy=61.848505	 loss=1.579582	 lr=0.010000
Epoch[035] Batch [2349]/[3760]	Speed: 35.489938 samples/sec	 accuracy=61.827128	 loss=1.580074	 lr=0.010000
Epoch[035] Batch [2399]/[3760]	Speed: 35.138930 samples/sec	 accuracy=61.813802	 loss=1.580773	 lr=0.010000
Epoch[035] Batch [2449]/[3760]	Speed: 35.455433 samples/sec	 accuracy=61.779974	 loss=1.581797	 lr=0.010000
Epoch[035] Batch [2499]/[3760]	Speed: 34.889112 samples/sec	 accuracy=61.762500	 loss=1.583407	 lr=0.010000
Epoch[035] Batch [2549]/[3760]	Speed: 35.364388 samples/sec	 accuracy=61.749387	 loss=1.583389	 lr=0.010000
Epoch[035] Batch [2599]/[3760]	Speed: 35.125948 samples/sec	 accuracy=61.747596	 loss=1.583232	 lr=0.010000
Epoch[035] Batch [2649]/[3760]	Speed: 35.199160 samples/sec	 accuracy=61.745283	 loss=1.583916	 lr=0.010000
Epoch[035] Batch [2699]/[3760]	Speed: 35.296545 samples/sec	 accuracy=61.726852	 loss=1.584381	 lr=0.010000
Epoch[035] Batch [2749]/[3760]	Speed: 35.127164 samples/sec	 accuracy=61.731250	 loss=1.584416	 lr=0.010000
Epoch[035] Batch [2799]/[3760]	Speed: 35.414490 samples/sec	 accuracy=61.691406	 loss=1.585369	 lr=0.010000
Epoch[035] Batch [2849]/[3760]	Speed: 35.182965 samples/sec	 accuracy=61.660088	 loss=1.585987	 lr=0.010000
Epoch[035] Batch [2899]/[3760]	Speed: 35.478556 samples/sec	 accuracy=61.669181	 loss=1.585338	 lr=0.010000
Epoch[035] Batch [2949]/[3760]	Speed: 35.362761 samples/sec	 accuracy=61.685381	 loss=1.585488	 lr=0.010000
Epoch[035] Batch [2999]/[3760]	Speed: 35.420250 samples/sec	 accuracy=61.685417	 loss=1.585195	 lr=0.010000
Epoch[035] Batch [3049]/[3760]	Speed: 35.200194 samples/sec	 accuracy=61.695184	 loss=1.585145	 lr=0.010000
Epoch[035] Batch [3099]/[3760]	Speed: 35.403859 samples/sec	 accuracy=61.680948	 loss=1.585774	 lr=0.010000
Epoch[035] Batch [3149]/[3760]	Speed: 35.038550 samples/sec	 accuracy=61.680060	 loss=1.586041	 lr=0.010000
Epoch[035] Batch [3199]/[3760]	Speed: 35.348119 samples/sec	 accuracy=61.687012	 loss=1.585646	 lr=0.010000
Epoch[035] Batch [3249]/[3760]	Speed: 34.979077 samples/sec	 accuracy=61.677885	 loss=1.585865	 lr=0.010000
Epoch[035] Batch [3299]/[3760]	Speed: 35.615867 samples/sec	 accuracy=61.682292	 loss=1.586016	 lr=0.010000
Epoch[035] Batch [3349]/[3760]	Speed: 34.897887 samples/sec	 accuracy=61.693097	 loss=1.585521	 lr=0.010000
Epoch[035] Batch [3399]/[3760]	Speed: 35.340630 samples/sec	 accuracy=61.686121	 loss=1.586167	 lr=0.010000
Epoch[035] Batch [3449]/[3760]	Speed: 34.875655 samples/sec	 accuracy=61.666667	 loss=1.587309	 lr=0.010000
Epoch[035] Batch [3499]/[3760]	Speed: 35.385815 samples/sec	 accuracy=61.668750	 loss=1.588373	 lr=0.010000
Epoch[035] Batch [3549]/[3760]	Speed: 35.108706 samples/sec	 accuracy=61.662412	 loss=1.588660	 lr=0.010000
Epoch[035] Batch [3599]/[3760]	Speed: 35.280548 samples/sec	 accuracy=61.644097	 loss=1.589249	 lr=0.010000
Epoch[035] Batch [3649]/[3760]	Speed: 35.248772 samples/sec	 accuracy=61.639555	 loss=1.589846	 lr=0.010000
Epoch[035] Batch [3699]/[3760]	Speed: 35.155849 samples/sec	 accuracy=61.649916	 loss=1.589450	 lr=0.010000
Epoch[035] Batch [3749]/[3760]	Speed: 36.199864 samples/sec	 accuracy=61.641667	 loss=1.590128	 lr=0.010000
Batch [0049]/[0303]: acc-top1=58.062500 acc-top5=81.875000
Batch [0099]/[0303]: acc-top1=58.375000 acc-top5=81.328125
Batch [0149]/[0303]: acc-top1=58.833333 acc-top5=81.645833
Batch [0199]/[0303]: acc-top1=58.984375 acc-top5=81.671875
Batch [0249]/[0303]: acc-top1=58.718750 acc-top5=81.406250
Batch [0299]/[0303]: acc-top1=58.593750 acc-top5=81.390625
[Epoch 035] training: accuracy=61.644365	 loss=1.590329
[Epoch 035] speed: 35 samples/sec	time cost: 7334.872836
[Epoch 035] validation: acc-top1=58.611799 acc-top5=81.389233 loss=1.958987
Epoch[036] Batch [0049]/[3760]	Speed: 24.963751 samples/sec	 accuracy=61.875000	 loss=1.555490	 lr=0.010000
Epoch[036] Batch [0099]/[3760]	Speed: 35.049569 samples/sec	 accuracy=62.109375	 loss=1.548885	 lr=0.010000
Epoch[036] Batch [0149]/[3760]	Speed: 34.809227 samples/sec	 accuracy=62.354167	 loss=1.544825	 lr=0.010000
Epoch[036] Batch [0199]/[3760]	Speed: 34.868969 samples/sec	 accuracy=62.304688	 loss=1.545142	 lr=0.010000
Epoch[036] Batch [0249]/[3760]	Speed: 35.218345 samples/sec	 accuracy=62.343750	 loss=1.548337	 lr=0.010000
Epoch[036] Batch [0299]/[3760]	Speed: 34.743241 samples/sec	 accuracy=62.359375	 loss=1.556452	 lr=0.010000
Epoch[036] Batch [0349]/[3760]	Speed: 35.327182 samples/sec	 accuracy=62.191964	 loss=1.558471	 lr=0.010000
Epoch[036] Batch [0399]/[3760]	Speed: 34.636770 samples/sec	 accuracy=62.382812	 loss=1.556185	 lr=0.010000
Epoch[036] Batch [0449]/[3760]	Speed: 35.451876 samples/sec	 accuracy=62.513889	 loss=1.555416	 lr=0.010000
Epoch[036] Batch [0499]/[3760]	Speed: 35.067996 samples/sec	 accuracy=62.593750	 loss=1.554715	 lr=0.010000
Epoch[036] Batch [0549]/[3760]	Speed: 35.334470 samples/sec	 accuracy=62.511364	 loss=1.558025	 lr=0.010000
Epoch[036] Batch [0599]/[3760]	Speed: 35.103598 samples/sec	 accuracy=62.541667	 loss=1.553081	 lr=0.010000
Epoch[036] Batch [0649]/[3760]	Speed: 35.453056 samples/sec	 accuracy=62.540865	 loss=1.551971	 lr=0.010000
Epoch[036] Batch [0699]/[3760]	Speed: 34.908115 samples/sec	 accuracy=62.504464	 loss=1.551754	 lr=0.010000
Epoch[036] Batch [0749]/[3760]	Speed: 35.390311 samples/sec	 accuracy=62.400000	 loss=1.557037	 lr=0.010000
Epoch[036] Batch [0799]/[3760]	Speed: 35.225251 samples/sec	 accuracy=62.412109	 loss=1.558610	 lr=0.010000
Epoch[036] Batch [0849]/[3760]	Speed: 35.149850 samples/sec	 accuracy=62.347426	 loss=1.559928	 lr=0.010000
Epoch[036] Batch [0899]/[3760]	Speed: 35.140763 samples/sec	 accuracy=62.333333	 loss=1.561316	 lr=0.010000
Epoch[036] Batch [0949]/[3760]	Speed: 35.266995 samples/sec	 accuracy=62.330592	 loss=1.561351	 lr=0.010000
Epoch[036] Batch [0999]/[3760]	Speed: 35.120221 samples/sec	 accuracy=62.339062	 loss=1.560856	 lr=0.010000
Epoch[036] Batch [1049]/[3760]	Speed: 35.205327 samples/sec	 accuracy=62.360119	 loss=1.559555	 lr=0.010000
Epoch[036] Batch [1099]/[3760]	Speed: 35.238062 samples/sec	 accuracy=62.292614	 loss=1.561668	 lr=0.010000
Epoch[036] Batch [1149]/[3760]	Speed: 35.020306 samples/sec	 accuracy=62.334239	 loss=1.557056	 lr=0.010000
Epoch[036] Batch [1199]/[3760]	Speed: 35.160976 samples/sec	 accuracy=62.385417	 loss=1.555870	 lr=0.010000
Epoch[036] Batch [1249]/[3760]	Speed: 35.151171 samples/sec	 accuracy=62.381250	 loss=1.553980	 lr=0.010000
Epoch[036] Batch [1299]/[3760]	Speed: 35.259073 samples/sec	 accuracy=62.390625	 loss=1.554007	 lr=0.010000
Epoch[036] Batch [1349]/[3760]	Speed: 35.271819 samples/sec	 accuracy=62.368056	 loss=1.553662	 lr=0.010000
Epoch[036] Batch [1399]/[3760]	Speed: 34.978862 samples/sec	 accuracy=62.395089	 loss=1.553295	 lr=0.010000
Epoch[036] Batch [1449]/[3760]	Speed: 35.411640 samples/sec	 accuracy=62.357759	 loss=1.555022	 lr=0.010000
Epoch[036] Batch [1499]/[3760]	Speed: 35.212153 samples/sec	 accuracy=62.401042	 loss=1.553140	 lr=0.010000
Epoch[036] Batch [1549]/[3760]	Speed: 35.408202 samples/sec	 accuracy=62.333669	 loss=1.554484	 lr=0.010000
Epoch[036] Batch [1599]/[3760]	Speed: 35.294951 samples/sec	 accuracy=62.333984	 loss=1.553894	 lr=0.010000
Epoch[036] Batch [1649]/[3760]	Speed: 35.229840 samples/sec	 accuracy=62.286932	 loss=1.555489	 lr=0.010000
Epoch[036] Batch [1699]/[3760]	Speed: 35.239950 samples/sec	 accuracy=62.289522	 loss=1.555519	 lr=0.010000
Epoch[036] Batch [1749]/[3760]	Speed: 35.489807 samples/sec	 accuracy=62.277679	 loss=1.555856	 lr=0.010000
Epoch[036] Batch [1799]/[3760]	Speed: 35.136895 samples/sec	 accuracy=62.235243	 loss=1.557791	 lr=0.010000
Epoch[036] Batch [1849]/[3760]	Speed: 35.301930 samples/sec	 accuracy=62.185811	 loss=1.560433	 lr=0.010000
Epoch[036] Batch [1899]/[3760]	Speed: 35.080197 samples/sec	 accuracy=62.164474	 loss=1.562615	 lr=0.010000
Epoch[036] Batch [1949]/[3760]	Speed: 35.131605 samples/sec	 accuracy=62.156250	 loss=1.562904	 lr=0.010000
Epoch[036] Batch [1999]/[3760]	Speed: 35.058262 samples/sec	 accuracy=62.125000	 loss=1.564467	 lr=0.010000
Epoch[036] Batch [2049]/[3760]	Speed: 35.213167 samples/sec	 accuracy=62.094512	 loss=1.564917	 lr=0.010000
Epoch[036] Batch [2099]/[3760]	Speed: 35.228101 samples/sec	 accuracy=62.071429	 loss=1.565728	 lr=0.010000
Epoch[036] Batch [2149]/[3760]	Speed: 34.910219 samples/sec	 accuracy=62.021076	 loss=1.567495	 lr=0.010000
Epoch[036] Batch [2199]/[3760]	Speed: 35.444052 samples/sec	 accuracy=62.024858	 loss=1.568666	 lr=0.010000
Epoch[036] Batch [2249]/[3760]	Speed: 34.889830 samples/sec	 accuracy=62.024306	 loss=1.569122	 lr=0.010000
Epoch[036] Batch [2299]/[3760]	Speed: 35.603264 samples/sec	 accuracy=61.996603	 loss=1.569750	 lr=0.010000
Epoch[036] Batch [2349]/[3760]	Speed: 35.126395 samples/sec	 accuracy=62.001995	 loss=1.569407	 lr=0.010000
Epoch[036] Batch [2399]/[3760]	Speed: 35.390604 samples/sec	 accuracy=61.998698	 loss=1.569609	 lr=0.010000
Epoch[036] Batch [2449]/[3760]	Speed: 35.110879 samples/sec	 accuracy=61.916454	 loss=1.572776	 lr=0.010000
Epoch[036] Batch [2499]/[3760]	Speed: 35.488028 samples/sec	 accuracy=61.861875	 loss=1.574029	 lr=0.010000
Epoch[036] Batch [2549]/[3760]	Speed: 34.940879 samples/sec	 accuracy=61.880515	 loss=1.574111	 lr=0.010000
Epoch[036] Batch [2599]/[3760]	Speed: 35.195805 samples/sec	 accuracy=61.852163	 loss=1.575401	 lr=0.010000
Epoch[036] Batch [2649]/[3760]	Speed: 34.891018 samples/sec	 accuracy=61.836675	 loss=1.576075	 lr=0.010000
Epoch[036] Batch [2699]/[3760]	Speed: 35.401400 samples/sec	 accuracy=61.849537	 loss=1.575568	 lr=0.010000
Epoch[036] Batch [2749]/[3760]	Speed: 34.531538 samples/sec	 accuracy=61.821591	 loss=1.576403	 lr=0.010000
Epoch[036] Batch [2799]/[3760]	Speed: 35.367561 samples/sec	 accuracy=61.818638	 loss=1.577128	 lr=0.010000
Epoch[036] Batch [2849]/[3760]	Speed: 35.076499 samples/sec	 accuracy=61.815241	 loss=1.577315	 lr=0.010000
Epoch[036] Batch [2899]/[3760]	Speed: 35.069148 samples/sec	 accuracy=61.804957	 loss=1.577249	 lr=0.010000
Epoch[036] Batch [2949]/[3760]	Speed: 35.119126 samples/sec	 accuracy=61.814089	 loss=1.576957	 lr=0.010000
Epoch[036] Batch [2999]/[3760]	Speed: 35.031626 samples/sec	 accuracy=61.781250	 loss=1.578450	 lr=0.010000
Epoch[036] Batch [3049]/[3760]	Speed: 35.364827 samples/sec	 accuracy=61.761783	 loss=1.579402	 lr=0.010000
Epoch[036] Batch [3099]/[3760]	Speed: 35.200183 samples/sec	 accuracy=61.742440	 loss=1.580156	 lr=0.010000
Epoch[036] Batch [3149]/[3760]	Speed: 35.377031 samples/sec	 accuracy=61.715774	 loss=1.581832	 lr=0.010000
Epoch[036] Batch [3199]/[3760]	Speed: 35.244630 samples/sec	 accuracy=61.689941	 loss=1.582583	 lr=0.010000
Epoch[036] Batch [3249]/[3760]	Speed: 35.511747 samples/sec	 accuracy=61.681250	 loss=1.583064	 lr=0.010000
Epoch[036] Batch [3299]/[3760]	Speed: 35.211215 samples/sec	 accuracy=61.700758	 loss=1.583430	 lr=0.010000
Epoch[036] Batch [3349]/[3760]	Speed: 35.432543 samples/sec	 accuracy=61.694030	 loss=1.583688	 lr=0.010000
Epoch[036] Batch [3399]/[3760]	Speed: 35.037244 samples/sec	 accuracy=61.715533	 loss=1.582875	 lr=0.010000
Epoch[036] Batch [3449]/[3760]	Speed: 35.481160 samples/sec	 accuracy=61.690217	 loss=1.583168	 lr=0.010000
Epoch[036] Batch [3499]/[3760]	Speed: 35.099605 samples/sec	 accuracy=61.684821	 loss=1.583506	 lr=0.010000
Epoch[036] Batch [3549]/[3760]	Speed: 35.403240 samples/sec	 accuracy=61.691461	 loss=1.583152	 lr=0.010000
Epoch[036] Batch [3599]/[3760]	Speed: 35.100071 samples/sec	 accuracy=61.674045	 loss=1.583520	 lr=0.010000
Epoch[036] Batch [3649]/[3760]	Speed: 35.467153 samples/sec	 accuracy=61.671233	 loss=1.583920	 lr=0.010000
Epoch[036] Batch [3699]/[3760]	Speed: 35.134566 samples/sec	 accuracy=61.663429	 loss=1.584560	 lr=0.010000
Epoch[036] Batch [3749]/[3760]	Speed: 36.683894 samples/sec	 accuracy=61.667917	 loss=1.584895	 lr=0.010000
Batch [0049]/[0303]: acc-top1=57.562500 acc-top5=80.437500
Batch [0099]/[0303]: acc-top1=58.000000 acc-top5=80.250000
Batch [0149]/[0303]: acc-top1=58.333333 acc-top5=80.802083
Batch [0199]/[0303]: acc-top1=58.046875 acc-top5=80.960938
Batch [0249]/[0303]: acc-top1=57.775000 acc-top5=80.800000
Batch [0299]/[0303]: acc-top1=57.755208 acc-top5=80.807292
[Epoch 036] training: accuracy=61.653923	 loss=1.585429
[Epoch 036] speed: 35 samples/sec	time cost: 7341.763641
[Epoch 036] validation: acc-top1=57.704208 acc-top5=80.806518 loss=1.997611
Epoch[037] Batch [0049]/[3759]	Speed: 24.301813 samples/sec	 accuracy=61.281250	 loss=1.560982	 lr=0.010000
Epoch[037] Batch [0099]/[3759]	Speed: 35.073038 samples/sec	 accuracy=61.921875	 loss=1.547196	 lr=0.010000
Epoch[037] Batch [0149]/[3759]	Speed: 34.946884 samples/sec	 accuracy=62.364583	 loss=1.528245	 lr=0.010000
Epoch[037] Batch [0199]/[3759]	Speed: 35.195686 samples/sec	 accuracy=61.953125	 loss=1.547721	 lr=0.010000
Epoch[037] Batch [0249]/[3759]	Speed: 35.168913 samples/sec	 accuracy=62.181250	 loss=1.554433	 lr=0.010000
Epoch[037] Batch [0299]/[3759]	Speed: 35.022905 samples/sec	 accuracy=62.057292	 loss=1.561607	 lr=0.010000
Epoch[037] Batch [0349]/[3759]	Speed: 35.225116 samples/sec	 accuracy=62.250000	 loss=1.553438	 lr=0.010000
Epoch[037] Batch [0399]/[3759]	Speed: 35.047219 samples/sec	 accuracy=62.222656	 loss=1.552385	 lr=0.010000
Epoch[037] Batch [0449]/[3759]	Speed: 35.508673 samples/sec	 accuracy=62.364583	 loss=1.545044	 lr=0.010000
Epoch[037] Batch [0499]/[3759]	Speed: 35.029748 samples/sec	 accuracy=62.337500	 loss=1.542543	 lr=0.010000
Epoch[037] Batch [0549]/[3759]	Speed: 35.596891 samples/sec	 accuracy=62.326705	 loss=1.544816	 lr=0.010000
Epoch[037] Batch [0599]/[3759]	Speed: 35.247930 samples/sec	 accuracy=62.302083	 loss=1.544485	 lr=0.010000
Epoch[037] Batch [0649]/[3759]	Speed: 35.344208 samples/sec	 accuracy=62.350962	 loss=1.543742	 lr=0.010000
Epoch[037] Batch [0699]/[3759]	Speed: 35.213846 samples/sec	 accuracy=62.274554	 loss=1.546586	 lr=0.010000
Epoch[037] Batch [0749]/[3759]	Speed: 35.241495 samples/sec	 accuracy=62.281250	 loss=1.547236	 lr=0.010000
Epoch[037] Batch [0799]/[3759]	Speed: 35.171904 samples/sec	 accuracy=62.263672	 loss=1.547540	 lr=0.010000
Epoch[037] Batch [0849]/[3759]	Speed: 35.507595 samples/sec	 accuracy=62.290441	 loss=1.547158	 lr=0.010000
Epoch[037] Batch [0899]/[3759]	Speed: 35.157502 samples/sec	 accuracy=62.331597	 loss=1.546599	 lr=0.010000
Epoch[037] Batch [0949]/[3759]	Speed: 35.518730 samples/sec	 accuracy=62.342105	 loss=1.548682	 lr=0.010000
Epoch[037] Batch [0999]/[3759]	Speed: 35.186386 samples/sec	 accuracy=62.337500	 loss=1.548875	 lr=0.010000
Epoch[037] Batch [1049]/[3759]	Speed: 35.411169 samples/sec	 accuracy=62.260417	 loss=1.550550	 lr=0.010000
Epoch[037] Batch [1099]/[3759]	Speed: 35.270031 samples/sec	 accuracy=62.201705	 loss=1.553583	 lr=0.010000
Epoch[037] Batch [1149]/[3759]	Speed: 35.331144 samples/sec	 accuracy=62.201087	 loss=1.554784	 lr=0.010000
Epoch[037] Batch [1199]/[3759]	Speed: 35.329323 samples/sec	 accuracy=62.182292	 loss=1.555426	 lr=0.010000
Epoch[037] Batch [1249]/[3759]	Speed: 35.301634 samples/sec	 accuracy=62.190000	 loss=1.555817	 lr=0.010000
Epoch[037] Batch [1299]/[3759]	Speed: 35.118534 samples/sec	 accuracy=62.189904	 loss=1.554307	 lr=0.010000
Epoch[037] Batch [1349]/[3759]	Speed: 35.357781 samples/sec	 accuracy=62.171296	 loss=1.556626	 lr=0.010000
Epoch[037] Batch [1399]/[3759]	Speed: 35.212720 samples/sec	 accuracy=62.175223	 loss=1.557136	 lr=0.010000
Epoch[037] Batch [1449]/[3759]	Speed: 35.138006 samples/sec	 accuracy=62.200431	 loss=1.557694	 lr=0.010000
Epoch[037] Batch [1499]/[3759]	Speed: 35.241760 samples/sec	 accuracy=62.225000	 loss=1.557417	 lr=0.010000
Epoch[037] Batch [1549]/[3759]	Speed: 35.279587 samples/sec	 accuracy=62.212702	 loss=1.557719	 lr=0.010000
Epoch[037] Batch [1599]/[3759]	Speed: 35.219425 samples/sec	 accuracy=62.203125	 loss=1.559218	 lr=0.010000
Epoch[037] Batch [1649]/[3759]	Speed: 34.806792 samples/sec	 accuracy=62.112689	 loss=1.561668	 lr=0.010000
Epoch[037] Batch [1699]/[3759]	Speed: 35.309805 samples/sec	 accuracy=62.108456	 loss=1.562499	 lr=0.010000
Epoch[037] Batch [1749]/[3759]	Speed: 34.989637 samples/sec	 accuracy=62.146429	 loss=1.560828	 lr=0.010000
Epoch[037] Batch [1799]/[3759]	Speed: 35.268980 samples/sec	 accuracy=62.133681	 loss=1.562442	 lr=0.010000
Epoch[037] Batch [1849]/[3759]	Speed: 34.871826 samples/sec	 accuracy=62.104730	 loss=1.562859	 lr=0.010000
Epoch[037] Batch [1899]/[3759]	Speed: 35.305118 samples/sec	 accuracy=62.097039	 loss=1.562933	 lr=0.010000
Epoch[037] Batch [1949]/[3759]	Speed: 34.998321 samples/sec	 accuracy=62.082532	 loss=1.562801	 lr=0.010000
Epoch[037] Batch [1999]/[3759]	Speed: 35.439652 samples/sec	 accuracy=62.054688	 loss=1.563897	 lr=0.010000
Epoch[037] Batch [2049]/[3759]	Speed: 34.817775 samples/sec	 accuracy=62.012957	 loss=1.565962	 lr=0.010000
Epoch[037] Batch [2099]/[3759]	Speed: 35.169650 samples/sec	 accuracy=62.002976	 loss=1.566754	 lr=0.010000
Epoch[037] Batch [2149]/[3759]	Speed: 35.112323 samples/sec	 accuracy=61.960756	 loss=1.567751	 lr=0.010000
Epoch[037] Batch [2199]/[3759]	Speed: 35.059697 samples/sec	 accuracy=61.970881	 loss=1.567565	 lr=0.010000
Epoch[037] Batch [2249]/[3759]	Speed: 35.281896 samples/sec	 accuracy=61.950000	 loss=1.568307	 lr=0.010000
Epoch[037] Batch [2299]/[3759]	Speed: 35.139683 samples/sec	 accuracy=61.960598	 loss=1.567894	 lr=0.010000
Epoch[037] Batch [2349]/[3759]	Speed: 35.418680 samples/sec	 accuracy=61.966755	 loss=1.567895	 lr=0.010000
Epoch[037] Batch [2399]/[3759]	Speed: 34.952049 samples/sec	 accuracy=61.970052	 loss=1.568498	 lr=0.010000
Epoch[037] Batch [2449]/[3759]	Speed: 35.533189 samples/sec	 accuracy=61.975765	 loss=1.569763	 lr=0.010000
Epoch[037] Batch [2499]/[3759]	Speed: 34.846138 samples/sec	 accuracy=61.944375	 loss=1.571137	 lr=0.010000
Epoch[037] Batch [2549]/[3759]	Speed: 35.507425 samples/sec	 accuracy=61.955882	 loss=1.570613	 lr=0.010000
Epoch[037] Batch [2599]/[3759]	Speed: 34.949750 samples/sec	 accuracy=61.953726	 loss=1.570778	 lr=0.010000
Epoch[037] Batch [2649]/[3759]	Speed: 35.300548 samples/sec	 accuracy=61.937500	 loss=1.571630	 lr=0.010000
Epoch[037] Batch [2699]/[3759]	Speed: 34.861653 samples/sec	 accuracy=61.910880	 loss=1.572957	 lr=0.010000
Epoch[037] Batch [2749]/[3759]	Speed: 35.413933 samples/sec	 accuracy=61.903409	 loss=1.573436	 lr=0.010000
Epoch[037] Batch [2799]/[3759]	Speed: 34.683386 samples/sec	 accuracy=61.928571	 loss=1.572155	 lr=0.010000
Epoch[037] Batch [2849]/[3759]	Speed: 34.865244 samples/sec	 accuracy=61.939693	 loss=1.571501	 lr=0.010000
Epoch[037] Batch [2899]/[3759]	Speed: 34.871582 samples/sec	 accuracy=61.909483	 loss=1.572789	 lr=0.010000
Epoch[037] Batch [2949]/[3759]	Speed: 34.759185 samples/sec	 accuracy=61.896716	 loss=1.572992	 lr=0.010000
Epoch[037] Batch [2999]/[3759]	Speed: 34.908175 samples/sec	 accuracy=61.881771	 loss=1.573969	 lr=0.010000
Epoch[037] Batch [3049]/[3759]	Speed: 34.549583 samples/sec	 accuracy=61.846311	 loss=1.575629	 lr=0.010000
Epoch[037] Batch [3099]/[3759]	Speed: 35.237004 samples/sec	 accuracy=61.861895	 loss=1.575422	 lr=0.010000
Epoch[037] Batch [3149]/[3759]	Speed: 34.758816 samples/sec	 accuracy=61.842758	 loss=1.576469	 lr=0.010000
Epoch[037] Batch [3199]/[3759]	Speed: 35.343277 samples/sec	 accuracy=61.839355	 loss=1.576732	 lr=0.010000
Epoch[037] Batch [3249]/[3759]	Speed: 34.971068 samples/sec	 accuracy=61.821154	 loss=1.577644	 lr=0.010000
Epoch[037] Batch [3299]/[3759]	Speed: 35.020226 samples/sec	 accuracy=61.814394	 loss=1.577583	 lr=0.010000
Epoch[037] Batch [3349]/[3759]	Speed: 34.944313 samples/sec	 accuracy=61.798507	 loss=1.578697	 lr=0.010000
Epoch[037] Batch [3399]/[3759]	Speed: 35.232482 samples/sec	 accuracy=61.785846	 loss=1.578958	 lr=0.010000
Epoch[037] Batch [3449]/[3759]	Speed: 35.091138 samples/sec	 accuracy=61.772645	 loss=1.579607	 lr=0.010000
Epoch[037] Batch [3499]/[3759]	Speed: 35.144808 samples/sec	 accuracy=61.756696	 loss=1.579825	 lr=0.010000
Epoch[037] Batch [3549]/[3759]	Speed: 35.085685 samples/sec	 accuracy=61.744718	 loss=1.580538	 lr=0.010000
Epoch[037] Batch [3599]/[3759]	Speed: 35.060380 samples/sec	 accuracy=61.733941	 loss=1.581380	 lr=0.010000
Epoch[037] Batch [3649]/[3759]	Speed: 35.105189 samples/sec	 accuracy=61.720034	 loss=1.582213	 lr=0.010000
Epoch[037] Batch [3699]/[3759]	Speed: 34.685816 samples/sec	 accuracy=61.734797	 loss=1.581316	 lr=0.010000
Epoch[037] Batch [3749]/[3759]	Speed: 36.432249 samples/sec	 accuracy=61.737083	 loss=1.581627	 lr=0.010000
Batch [0049]/[0303]: acc-top1=58.031250 acc-top5=81.812500
Batch [0099]/[0303]: acc-top1=58.296875 acc-top5=81.140625
Batch [0149]/[0303]: acc-top1=58.677083 acc-top5=81.322917
Batch [0199]/[0303]: acc-top1=58.492188 acc-top5=81.445312
Batch [0249]/[0303]: acc-top1=58.356250 acc-top5=80.912500
Batch [0299]/[0303]: acc-top1=58.234375 acc-top5=80.942708
[Epoch 037] training: accuracy=61.742235	 loss=1.581362
[Epoch 037] speed: 34 samples/sec	time cost: 7352.240981
[Epoch 037] validation: acc-top1=58.188944 acc-top5=80.935437 loss=2.003857
Epoch[038] Batch [0049]/[3760]	Speed: 25.115183 samples/sec	 accuracy=62.312500	 loss=1.578057	 lr=0.010000
Epoch[038] Batch [0099]/[3760]	Speed: 34.859842 samples/sec	 accuracy=61.671875	 loss=1.586392	 lr=0.010000
Epoch[038] Batch [0149]/[3760]	Speed: 34.828724 samples/sec	 accuracy=61.718750	 loss=1.575117	 lr=0.010000
Epoch[038] Batch [0199]/[3760]	Speed: 35.079658 samples/sec	 accuracy=61.796875	 loss=1.565742	 lr=0.010000
Epoch[038] Batch [0249]/[3760]	Speed: 35.203006 samples/sec	 accuracy=62.006250	 loss=1.567025	 lr=0.010000
Epoch[038] Batch [0299]/[3760]	Speed: 34.982276 samples/sec	 accuracy=62.015625	 loss=1.563411	 lr=0.010000
Epoch[038] Batch [0349]/[3760]	Speed: 35.379371 samples/sec	 accuracy=62.187500	 loss=1.558769	 lr=0.010000
Epoch[038] Batch [0399]/[3760]	Speed: 34.854331 samples/sec	 accuracy=62.089844	 loss=1.557459	 lr=0.010000
Epoch[038] Batch [0449]/[3760]	Speed: 35.452713 samples/sec	 accuracy=62.152778	 loss=1.557557	 lr=0.010000
Epoch[038] Batch [0499]/[3760]	Speed: 34.777001 samples/sec	 accuracy=62.096875	 loss=1.560805	 lr=0.010000
Epoch[038] Batch [0549]/[3760]	Speed: 35.430871 samples/sec	 accuracy=62.125000	 loss=1.561060	 lr=0.010000
Epoch[038] Batch [0599]/[3760]	Speed: 34.673007 samples/sec	 accuracy=62.109375	 loss=1.562036	 lr=0.010000
Epoch[038] Batch [0649]/[3760]	Speed: 35.306099 samples/sec	 accuracy=62.192308	 loss=1.556360	 lr=0.010000
Epoch[038] Batch [0699]/[3760]	Speed: 35.153622 samples/sec	 accuracy=62.243304	 loss=1.553228	 lr=0.010000
Epoch[038] Batch [0749]/[3760]	Speed: 35.290489 samples/sec	 accuracy=62.318750	 loss=1.551343	 lr=0.010000
Epoch[038] Batch [0799]/[3760]	Speed: 35.289811 samples/sec	 accuracy=62.324219	 loss=1.552016	 lr=0.010000
Epoch[038] Batch [0849]/[3760]	Speed: 35.204415 samples/sec	 accuracy=62.308824	 loss=1.551872	 lr=0.010000
Epoch[038] Batch [0899]/[3760]	Speed: 34.990165 samples/sec	 accuracy=62.218750	 loss=1.555472	 lr=0.010000
Epoch[038] Batch [0949]/[3760]	Speed: 35.014030 samples/sec	 accuracy=62.205592	 loss=1.555842	 lr=0.010000
Epoch[038] Batch [0999]/[3760]	Speed: 35.324163 samples/sec	 accuracy=62.178125	 loss=1.557436	 lr=0.010000
Epoch[038] Batch [1049]/[3760]	Speed: 34.915061 samples/sec	 accuracy=62.144345	 loss=1.558184	 lr=0.010000
Epoch[038] Batch [1099]/[3760]	Speed: 35.430731 samples/sec	 accuracy=62.119318	 loss=1.559469	 lr=0.010000
Epoch[038] Batch [1149]/[3760]	Speed: 35.111850 samples/sec	 accuracy=62.058424	 loss=1.561749	 lr=0.010000
Epoch[038] Batch [1199]/[3760]	Speed: 35.595384 samples/sec	 accuracy=62.016927	 loss=1.563716	 lr=0.010000
Epoch[038] Batch [1249]/[3760]	Speed: 35.054268 samples/sec	 accuracy=62.007500	 loss=1.564726	 lr=0.010000
Epoch[038] Batch [1299]/[3760]	Speed: 35.333589 samples/sec	 accuracy=62.022837	 loss=1.564843	 lr=0.010000
Epoch[038] Batch [1349]/[3760]	Speed: 35.155844 samples/sec	 accuracy=62.046296	 loss=1.563173	 lr=0.010000
Epoch[038] Batch [1399]/[3760]	Speed: 35.227504 samples/sec	 accuracy=62.047991	 loss=1.563432	 lr=0.010000
Epoch[038] Batch [1449]/[3760]	Speed: 35.015053 samples/sec	 accuracy=62.059267	 loss=1.563897	 lr=0.010000
Epoch[038] Batch [1499]/[3760]	Speed: 35.070584 samples/sec	 accuracy=62.129167	 loss=1.562470	 lr=0.010000
Epoch[038] Batch [1549]/[3760]	Speed: 35.258361 samples/sec	 accuracy=62.146169	 loss=1.562062	 lr=0.010000
Epoch[038] Batch [1599]/[3760]	Speed: 35.194133 samples/sec	 accuracy=62.156250	 loss=1.561681	 lr=0.010000
Epoch[038] Batch [1649]/[3760]	Speed: 35.390629 samples/sec	 accuracy=62.154356	 loss=1.561016	 lr=0.010000
Epoch[038] Batch [1699]/[3760]	Speed: 35.352366 samples/sec	 accuracy=62.159926	 loss=1.562050	 lr=0.010000
Epoch[038] Batch [1749]/[3760]	Speed: 35.240341 samples/sec	 accuracy=62.171429	 loss=1.562057	 lr=0.010000
Epoch[038] Batch [1799]/[3760]	Speed: 35.374648 samples/sec	 accuracy=62.150174	 loss=1.563119	 lr=0.010000
Epoch[038] Batch [1849]/[3760]	Speed: 35.150318 samples/sec	 accuracy=62.184966	 loss=1.561890	 lr=0.010000
Epoch[038] Batch [1899]/[3760]	Speed: 35.189612 samples/sec	 accuracy=62.217928	 loss=1.561929	 lr=0.010000
Epoch[038] Batch [1949]/[3760]	Speed: 35.221367 samples/sec	 accuracy=62.199519	 loss=1.562218	 lr=0.010000
Epoch[038] Batch [1999]/[3760]	Speed: 35.094111 samples/sec	 accuracy=62.210156	 loss=1.561531	 lr=0.010000
Epoch[038] Batch [2049]/[3760]	Speed: 35.267485 samples/sec	 accuracy=62.197409	 loss=1.562389	 lr=0.010000
Epoch[038] Batch [2099]/[3760]	Speed: 34.927876 samples/sec	 accuracy=62.174107	 loss=1.563088	 lr=0.010000
Epoch[038] Batch [2149]/[3760]	Speed: 35.162663 samples/sec	 accuracy=62.134448	 loss=1.564930	 lr=0.010000
Epoch[038] Batch [2199]/[3760]	Speed: 34.678033 samples/sec	 accuracy=62.145597	 loss=1.565863	 lr=0.010000
Epoch[038] Batch [2249]/[3760]	Speed: 35.161636 samples/sec	 accuracy=62.148611	 loss=1.565848	 lr=0.010000
Epoch[038] Batch [2299]/[3760]	Speed: 34.441243 samples/sec	 accuracy=62.133152	 loss=1.565849	 lr=0.010000
Epoch[038] Batch [2349]/[3760]	Speed: 35.255776 samples/sec	 accuracy=62.132314	 loss=1.567123	 lr=0.010000
Epoch[038] Batch [2399]/[3760]	Speed: 34.774405 samples/sec	 accuracy=62.149740	 loss=1.566994	 lr=0.010000
Epoch[038] Batch [2449]/[3760]	Speed: 34.860118 samples/sec	 accuracy=62.151786	 loss=1.566909	 lr=0.010000
Epoch[038] Batch [2499]/[3760]	Speed: 35.069308 samples/sec	 accuracy=62.120625	 loss=1.568589	 lr=0.010000
Epoch[038] Batch [2549]/[3760]	Speed: 34.751610 samples/sec	 accuracy=62.099877	 loss=1.569625	 lr=0.010000
Epoch[038] Batch [2599]/[3760]	Speed: 35.100392 samples/sec	 accuracy=62.111178	 loss=1.568657	 lr=0.010000
Epoch[038] Batch [2649]/[3760]	Speed: 34.747684 samples/sec	 accuracy=62.094929	 loss=1.569093	 lr=0.010000
Epoch[038] Batch [2699]/[3760]	Speed: 35.135820 samples/sec	 accuracy=62.080440	 loss=1.569796	 lr=0.010000
Epoch[038] Batch [2749]/[3760]	Speed: 34.548101 samples/sec	 accuracy=62.079545	 loss=1.570222	 lr=0.010000
Epoch[038] Batch [2799]/[3760]	Speed: 35.314241 samples/sec	 accuracy=62.079241	 loss=1.570610	 lr=0.010000
Epoch[038] Batch [2849]/[3760]	Speed: 34.759280 samples/sec	 accuracy=62.082785	 loss=1.570550	 lr=0.010000
Epoch[038] Batch [2899]/[3760]	Speed: 35.337863 samples/sec	 accuracy=62.063039	 loss=1.571649	 lr=0.010000
Epoch[038] Batch [2949]/[3760]	Speed: 35.099319 samples/sec	 accuracy=62.049788	 loss=1.572462	 lr=0.010000
Epoch[038] Batch [2999]/[3760]	Speed: 34.897857 samples/sec	 accuracy=62.057812	 loss=1.571817	 lr=0.010000
Epoch[038] Batch [3049]/[3760]	Speed: 35.295691 samples/sec	 accuracy=62.040471	 loss=1.572184	 lr=0.010000
Epoch[038] Batch [3099]/[3760]	Speed: 34.666437 samples/sec	 accuracy=62.036290	 loss=1.572476	 lr=0.010000
Epoch[038] Batch [3149]/[3760]	Speed: 35.350282 samples/sec	 accuracy=62.047619	 loss=1.572156	 lr=0.010000
Epoch[038] Batch [3199]/[3760]	Speed: 34.766544 samples/sec	 accuracy=62.051270	 loss=1.572000	 lr=0.010000
Epoch[038] Batch [3249]/[3760]	Speed: 35.436528 samples/sec	 accuracy=62.051442	 loss=1.572568	 lr=0.010000
Epoch[038] Batch [3299]/[3760]	Speed: 34.739395 samples/sec	 accuracy=62.032197	 loss=1.573035	 lr=0.010000
Epoch[038] Batch [3349]/[3760]	Speed: 35.299589 samples/sec	 accuracy=62.030784	 loss=1.573238	 lr=0.010000
Epoch[038] Batch [3399]/[3760]	Speed: 34.722412 samples/sec	 accuracy=62.036765	 loss=1.572608	 lr=0.010000
Epoch[038] Batch [3449]/[3760]	Speed: 35.189564 samples/sec	 accuracy=62.044384	 loss=1.572874	 lr=0.010000
Epoch[038] Batch [3499]/[3760]	Speed: 35.109735 samples/sec	 accuracy=62.020982	 loss=1.573982	 lr=0.010000
Epoch[038] Batch [3549]/[3760]	Speed: 35.077872 samples/sec	 accuracy=61.995158	 loss=1.575468	 lr=0.010000
Epoch[038] Batch [3599]/[3760]	Speed: 35.107631 samples/sec	 accuracy=62.004340	 loss=1.574648	 lr=0.010000
Epoch[038] Batch [3649]/[3760]	Speed: 34.992745 samples/sec	 accuracy=61.985445	 loss=1.575508	 lr=0.010000
Epoch[038] Batch [3699]/[3760]	Speed: 35.300326 samples/sec	 accuracy=61.983953	 loss=1.575672	 lr=0.010000
Epoch[038] Batch [3749]/[3760]	Speed: 36.319492 samples/sec	 accuracy=61.966667	 loss=1.576486	 lr=0.010000
Batch [0049]/[0303]: acc-top1=58.500000 acc-top5=81.093750
Batch [0099]/[0303]: acc-top1=58.703125 acc-top5=80.953125
Batch [0149]/[0303]: acc-top1=58.822917 acc-top5=81.218750
Batch [0199]/[0303]: acc-top1=58.585937 acc-top5=81.179688
Batch [0249]/[0303]: acc-top1=58.331250 acc-top5=80.825000
Batch [0299]/[0303]: acc-top1=58.286458 acc-top5=80.895833
[Epoch 038] training: accuracy=61.959358	 loss=1.576806
[Epoch 038] speed: 34 samples/sec	time cost: 7358.670676
[Epoch 038] validation: acc-top1=58.286922 acc-top5=80.883870 loss=1.957854
Epoch[039] Batch [0049]/[3760]	Speed: 24.125239 samples/sec	 accuracy=62.406250	 loss=1.533093	 lr=0.010000
Epoch[039] Batch [0099]/[3760]	Speed: 34.988193 samples/sec	 accuracy=62.359375	 loss=1.555401	 lr=0.010000
Epoch[039] Batch [0149]/[3760]	Speed: 34.943335 samples/sec	 accuracy=62.333333	 loss=1.554525	 lr=0.010000
Epoch[039] Batch [0199]/[3760]	Speed: 35.145382 samples/sec	 accuracy=62.742188	 loss=1.551657	 lr=0.010000
Epoch[039] Batch [0249]/[3760]	Speed: 35.159830 samples/sec	 accuracy=62.512500	 loss=1.560078	 lr=0.010000
Epoch[039] Batch [0299]/[3760]	Speed: 34.846658 samples/sec	 accuracy=62.463542	 loss=1.564250	 lr=0.010000
Epoch[039] Batch [0349]/[3760]	Speed: 35.193705 samples/sec	 accuracy=62.375000	 loss=1.563038	 lr=0.010000
Epoch[039] Batch [0399]/[3760]	Speed: 34.745325 samples/sec	 accuracy=62.386719	 loss=1.560912	 lr=0.010000
Epoch[039] Batch [0449]/[3760]	Speed: 35.324895 samples/sec	 accuracy=62.552083	 loss=1.554928	 lr=0.010000
Epoch[039] Batch [0499]/[3760]	Speed: 34.420953 samples/sec	 accuracy=62.500000	 loss=1.555525	 lr=0.010000
Epoch[039] Batch [0549]/[3760]	Speed: 35.373734 samples/sec	 accuracy=62.522727	 loss=1.554180	 lr=0.010000
Epoch[039] Batch [0599]/[3760]	Speed: 34.832548 samples/sec	 accuracy=62.528646	 loss=1.554296	 lr=0.010000
Epoch[039] Batch [0649]/[3760]	Speed: 35.292527 samples/sec	 accuracy=62.507212	 loss=1.555480	 lr=0.010000
Epoch[039] Batch [0699]/[3760]	Speed: 34.899168 samples/sec	 accuracy=62.537946	 loss=1.555591	 lr=0.010000
Epoch[039] Batch [0749]/[3760]	Speed: 35.110700 samples/sec	 accuracy=62.572917	 loss=1.552963	 lr=0.010000
Epoch[039] Batch [0799]/[3760]	Speed: 34.819303 samples/sec	 accuracy=62.486328	 loss=1.555668	 lr=0.010000
Epoch[039] Batch [0849]/[3760]	Speed: 35.025885 samples/sec	 accuracy=62.507353	 loss=1.555688	 lr=0.010000
Epoch[039] Batch [0899]/[3760]	Speed: 34.826984 samples/sec	 accuracy=62.479167	 loss=1.556027	 lr=0.010000
Epoch[039] Batch [0949]/[3760]	Speed: 34.774525 samples/sec	 accuracy=62.541118	 loss=1.553355	 lr=0.010000
Epoch[039] Batch [0999]/[3760]	Speed: 35.025291 samples/sec	 accuracy=62.567187	 loss=1.552717	 lr=0.010000
Epoch[039] Batch [1049]/[3760]	Speed: 34.543894 samples/sec	 accuracy=62.553571	 loss=1.554372	 lr=0.010000
Epoch[039] Batch [1099]/[3760]	Speed: 35.295799 samples/sec	 accuracy=62.556818	 loss=1.554728	 lr=0.010000
Epoch[039] Batch [1149]/[3760]	Speed: 34.323230 samples/sec	 accuracy=62.566576	 loss=1.555085	 lr=0.010000
Epoch[039] Batch [1199]/[3760]	Speed: 35.070304 samples/sec	 accuracy=62.578125	 loss=1.554442	 lr=0.010000
Epoch[039] Batch [1249]/[3760]	Speed: 34.547400 samples/sec	 accuracy=62.543750	 loss=1.555386	 lr=0.010000
Epoch[039] Batch [1299]/[3760]	Speed: 35.019018 samples/sec	 accuracy=62.536058	 loss=1.556301	 lr=0.010000
Epoch[039] Batch [1349]/[3760]	Speed: 35.035323 samples/sec	 accuracy=62.611111	 loss=1.553928	 lr=0.010000
Epoch[039] Batch [1399]/[3760]	Speed: 35.002414 samples/sec	 accuracy=62.635045	 loss=1.553769	 lr=0.010000
Epoch[039] Batch [1449]/[3760]	Speed: 35.313689 samples/sec	 accuracy=62.606681	 loss=1.554191	 lr=0.010000
Epoch[039] Batch [1499]/[3760]	Speed: 35.053742 samples/sec	 accuracy=62.632292	 loss=1.553434	 lr=0.010000
Epoch[039] Batch [1549]/[3760]	Speed: 35.266235 samples/sec	 accuracy=62.668347	 loss=1.553087	 lr=0.010000
Epoch[039] Batch [1599]/[3760]	Speed: 35.204304 samples/sec	 accuracy=62.657227	 loss=1.553232	 lr=0.010000
Epoch[039] Batch [1649]/[3760]	Speed: 35.214764 samples/sec	 accuracy=62.630682	 loss=1.554948	 lr=0.010000
Epoch[039] Batch [1699]/[3760]	Speed: 35.037212 samples/sec	 accuracy=62.612132	 loss=1.556003	 lr=0.010000
Epoch[039] Batch [1749]/[3760]	Speed: 35.269230 samples/sec	 accuracy=62.604464	 loss=1.557133	 lr=0.010000
Epoch[039] Batch [1799]/[3760]	Speed: 34.871486 samples/sec	 accuracy=62.581597	 loss=1.557526	 lr=0.010000
Epoch[039] Batch [1849]/[3760]	Speed: 35.227860 samples/sec	 accuracy=62.575169	 loss=1.557688	 lr=0.010000
Epoch[039] Batch [1899]/[3760]	Speed: 34.592043 samples/sec	 accuracy=62.569079	 loss=1.558143	 lr=0.010000
Epoch[039] Batch [1949]/[3760]	Speed: 35.212271 samples/sec	 accuracy=62.559295	 loss=1.558576	 lr=0.010000
Epoch[039] Batch [1999]/[3760]	Speed: 34.871659 samples/sec	 accuracy=62.548437	 loss=1.559451	 lr=0.010000
Epoch[039] Batch [2049]/[3760]	Speed: 35.122230 samples/sec	 accuracy=62.505335	 loss=1.560174	 lr=0.010000
Epoch[039] Batch [2099]/[3760]	Speed: 35.169613 samples/sec	 accuracy=62.491815	 loss=1.560522	 lr=0.010000
Epoch[039] Batch [2149]/[3760]	Speed: 35.330187 samples/sec	 accuracy=62.468023	 loss=1.561104	 lr=0.010000
Epoch[039] Batch [2199]/[3760]	Speed: 35.335818 samples/sec	 accuracy=62.458807	 loss=1.561734	 lr=0.010000
Epoch[039] Batch [2249]/[3760]	Speed: 34.964042 samples/sec	 accuracy=62.425000	 loss=1.563240	 lr=0.010000
Epoch[039] Batch [2299]/[3760]	Speed: 35.261001 samples/sec	 accuracy=62.395380	 loss=1.564407	 lr=0.010000
Epoch[039] Batch [2349]/[3760]	Speed: 34.781715 samples/sec	 accuracy=62.341755	 loss=1.565477	 lr=0.010000
Epoch[039] Batch [2399]/[3760]	Speed: 35.368194 samples/sec	 accuracy=62.314453	 loss=1.565980	 lr=0.010000
Epoch[039] Batch [2449]/[3760]	Speed: 34.908055 samples/sec	 accuracy=62.302934	 loss=1.566332	 lr=0.010000
Epoch[039] Batch [2499]/[3760]	Speed: 35.299698 samples/sec	 accuracy=62.284375	 loss=1.567226	 lr=0.010000
Epoch[039] Batch [2549]/[3760]	Speed: 35.065076 samples/sec	 accuracy=62.272672	 loss=1.567979	 lr=0.010000
Epoch[039] Batch [2599]/[3760]	Speed: 35.361620 samples/sec	 accuracy=62.269832	 loss=1.568477	 lr=0.010000
Epoch[039] Batch [2649]/[3760]	Speed: 34.899236 samples/sec	 accuracy=62.239387	 loss=1.569489	 lr=0.010000
Epoch[039] Batch [2699]/[3760]	Speed: 35.418034 samples/sec	 accuracy=62.248264	 loss=1.569094	 lr=0.010000
Epoch[039] Batch [2749]/[3760]	Speed: 34.953258 samples/sec	 accuracy=62.270455	 loss=1.568562	 lr=0.010000
Epoch[039] Batch [2799]/[3760]	Speed: 35.317453 samples/sec	 accuracy=62.259487	 loss=1.569201	 lr=0.010000
Epoch[039] Batch [2849]/[3760]	Speed: 34.985548 samples/sec	 accuracy=62.241776	 loss=1.569984	 lr=0.010000
Epoch[039] Batch [2899]/[3760]	Speed: 35.063604 samples/sec	 accuracy=62.248922	 loss=1.570062	 lr=0.010000
Epoch[039] Batch [2949]/[3760]	Speed: 34.810683 samples/sec	 accuracy=62.224576	 loss=1.570846	 lr=0.010000
Epoch[039] Batch [2999]/[3760]	Speed: 34.847079 samples/sec	 accuracy=62.226563	 loss=1.571239	 lr=0.010000
Epoch[039] Batch [3049]/[3760]	Speed: 34.953857 samples/sec	 accuracy=62.202869	 loss=1.571155	 lr=0.010000
Epoch[039] Batch [3099]/[3760]	Speed: 34.421290 samples/sec	 accuracy=62.182460	 loss=1.571527	 lr=0.010000
Epoch[039] Batch [3149]/[3760]	Speed: 34.829133 samples/sec	 accuracy=62.174603	 loss=1.571966	 lr=0.010000
Epoch[039] Batch [3199]/[3760]	Speed: 34.395879 samples/sec	 accuracy=62.177246	 loss=1.571663	 lr=0.010000
Epoch[039] Batch [3249]/[3760]	Speed: 35.270016 samples/sec	 accuracy=62.180288	 loss=1.571778	 lr=0.010000
Epoch[039] Batch [3299]/[3760]	Speed: 34.465478 samples/sec	 accuracy=62.173295	 loss=1.572044	 lr=0.010000
Epoch[039] Batch [3349]/[3760]	Speed: 35.212486 samples/sec	 accuracy=62.184701	 loss=1.571373	 lr=0.010000
Epoch[039] Batch [3399]/[3760]	Speed: 34.503395 samples/sec	 accuracy=62.153033	 loss=1.572472	 lr=0.010000
Epoch[039] Batch [3449]/[3760]	Speed: 34.901731 samples/sec	 accuracy=62.142663	 loss=1.573302	 lr=0.010000
Epoch[039] Batch [3499]/[3760]	Speed: 35.037707 samples/sec	 accuracy=62.112054	 loss=1.573677	 lr=0.010000
Epoch[039] Batch [3549]/[3760]	Speed: 34.751187 samples/sec	 accuracy=62.089789	 loss=1.575036	 lr=0.010000
Epoch[039] Batch [3599]/[3760]	Speed: 35.090463 samples/sec	 accuracy=62.071615	 loss=1.576134	 lr=0.010000
Epoch[039] Batch [3649]/[3760]	Speed: 34.897286 samples/sec	 accuracy=62.084332	 loss=1.575831	 lr=0.010000
Epoch[039] Batch [3699]/[3760]	Speed: 35.139050 samples/sec	 accuracy=62.087416	 loss=1.576389	 lr=0.010000
Epoch[039] Batch [3749]/[3760]	Speed: 36.080191 samples/sec	 accuracy=62.099167	 loss=1.575879	 lr=0.001000
Batch [0049]/[0303]: acc-top1=59.125000 acc-top5=81.562500
Batch [0099]/[0303]: acc-top1=59.109375 acc-top5=81.531250
Batch [0149]/[0303]: acc-top1=59.635417 acc-top5=81.927083
Batch [0199]/[0303]: acc-top1=59.718750 acc-top5=81.984375
Batch [0249]/[0303]: acc-top1=59.200000 acc-top5=81.781250
Batch [0299]/[0303]: acc-top1=59.197917 acc-top5=81.786458
[Epoch 039] training: accuracy=62.092337	 loss=1.576237
[Epoch 039] speed: 34 samples/sec	time cost: 7384.303889
[Epoch 039] validation: acc-top1=59.148102 acc-top5=81.796617 loss=1.936255
Epoch[040] Batch [0049]/[3759]	Speed: 24.176079 samples/sec	 accuracy=62.968750	 loss=1.504582	 lr=0.001000
Epoch[040] Batch [0099]/[3759]	Speed: 35.125800 samples/sec	 accuracy=64.093750	 loss=1.466913	 lr=0.001000
Epoch[040] Batch [0149]/[3759]	Speed: 34.765462 samples/sec	 accuracy=64.093750	 loss=1.461239	 lr=0.001000
Epoch[040] Batch [0199]/[3759]	Speed: 35.119705 samples/sec	 accuracy=64.398438	 loss=1.449055	 lr=0.001000
Epoch[040] Batch [0249]/[3759]	Speed: 35.211753 samples/sec	 accuracy=64.800000	 loss=1.440221	 lr=0.001000
Epoch[040] Batch [0299]/[3759]	Speed: 34.905206 samples/sec	 accuracy=64.927083	 loss=1.437050	 lr=0.001000
Epoch[040] Batch [0349]/[3759]	Speed: 35.127217 samples/sec	 accuracy=65.325893	 loss=1.424081	 lr=0.001000
Epoch[040] Batch [0399]/[3759]	Speed: 34.813162 samples/sec	 accuracy=65.609375	 loss=1.413950	 lr=0.001000
Epoch[040] Batch [0449]/[3759]	Speed: 35.413790 samples/sec	 accuracy=66.128472	 loss=1.393155	 lr=0.001000
Epoch[040] Batch [0499]/[3759]	Speed: 34.801128 samples/sec	 accuracy=66.200000	 loss=1.391372	 lr=0.001000
Epoch[040] Batch [0549]/[3759]	Speed: 35.394131 samples/sec	 accuracy=66.227273	 loss=1.392812	 lr=0.001000
Epoch[040] Batch [0599]/[3759]	Speed: 34.835756 samples/sec	 accuracy=66.312500	 loss=1.390308	 lr=0.001000
Epoch[040] Batch [0649]/[3759]	Speed: 35.413266 samples/sec	 accuracy=66.399038	 loss=1.386535	 lr=0.001000
Epoch[040] Batch [0699]/[3759]	Speed: 34.986432 samples/sec	 accuracy=66.598214	 loss=1.377995	 lr=0.001000
Epoch[040] Batch [0749]/[3759]	Speed: 35.187024 samples/sec	 accuracy=66.625000	 loss=1.374021	 lr=0.001000
Epoch[040] Batch [0799]/[3759]	Speed: 34.999297 samples/sec	 accuracy=66.628906	 loss=1.372266	 lr=0.001000
Epoch[040] Batch [0849]/[3759]	Speed: 35.420950 samples/sec	 accuracy=66.637868	 loss=1.370959	 lr=0.001000
Epoch[040] Batch [0899]/[3759]	Speed: 35.069446 samples/sec	 accuracy=66.741319	 loss=1.364868	 lr=0.001000
Epoch[040] Batch [0949]/[3759]	Speed: 35.346245 samples/sec	 accuracy=66.825658	 loss=1.360258	 lr=0.001000
Epoch[040] Batch [0999]/[3759]	Speed: 35.144459 samples/sec	 accuracy=66.878125	 loss=1.356568	 lr=0.001000
Epoch[040] Batch [1049]/[3759]	Speed: 35.035197 samples/sec	 accuracy=67.031250	 loss=1.352699	 lr=0.001000
Epoch[040] Batch [1099]/[3759]	Speed: 35.262692 samples/sec	 accuracy=67.115057	 loss=1.349164	 lr=0.001000
Epoch[040] Batch [1149]/[3759]	Speed: 35.107929 samples/sec	 accuracy=67.126359	 loss=1.349036	 lr=0.001000
Epoch[040] Batch [1199]/[3759]	Speed: 35.341377 samples/sec	 accuracy=67.213542	 loss=1.346346	 lr=0.001000
Epoch[040] Batch [1249]/[3759]	Speed: 35.010428 samples/sec	 accuracy=67.270000	 loss=1.344984	 lr=0.001000
Epoch[040] Batch [1299]/[3759]	Speed: 35.273218 samples/sec	 accuracy=67.317308	 loss=1.342073	 lr=0.001000
Epoch[040] Batch [1349]/[3759]	Speed: 34.835360 samples/sec	 accuracy=67.371528	 loss=1.339093	 lr=0.001000
Epoch[040] Batch [1399]/[3759]	Speed: 35.334895 samples/sec	 accuracy=67.425223	 loss=1.336567	 lr=0.001000
Epoch[040] Batch [1449]/[3759]	Speed: 35.159840 samples/sec	 accuracy=67.481681	 loss=1.333790	 lr=0.001000
Epoch[040] Batch [1499]/[3759]	Speed: 35.345513 samples/sec	 accuracy=67.535417	 loss=1.331288	 lr=0.001000
Epoch[040] Batch [1549]/[3759]	Speed: 35.366285 samples/sec	 accuracy=67.539315	 loss=1.330575	 lr=0.001000
Epoch[040] Batch [1599]/[3759]	Speed: 35.109626 samples/sec	 accuracy=67.555664	 loss=1.329153	 lr=0.001000
Epoch[040] Batch [1649]/[3759]	Speed: 35.290254 samples/sec	 accuracy=67.586174	 loss=1.327796	 lr=0.001000
Epoch[040] Batch [1699]/[3759]	Speed: 35.339290 samples/sec	 accuracy=67.646140	 loss=1.325459	 lr=0.001000
Epoch[040] Batch [1749]/[3759]	Speed: 35.274891 samples/sec	 accuracy=67.692857	 loss=1.322754	 lr=0.001000
Epoch[040] Batch [1799]/[3759]	Speed: 35.252398 samples/sec	 accuracy=67.765625	 loss=1.319627	 lr=0.001000
Epoch[040] Batch [1849]/[3759]	Speed: 35.280613 samples/sec	 accuracy=67.805743	 loss=1.318116	 lr=0.001000
Epoch[040] Batch [1899]/[3759]	Speed: 35.041571 samples/sec	 accuracy=67.855263	 loss=1.315658	 lr=0.001000
Epoch[040] Batch [1949]/[3759]	Speed: 35.380796 samples/sec	 accuracy=67.919071	 loss=1.314098	 lr=0.001000
Epoch[040] Batch [1999]/[3759]	Speed: 35.150792 samples/sec	 accuracy=67.957031	 loss=1.313226	 lr=0.001000
Epoch[040] Batch [2049]/[3759]	Speed: 35.380864 samples/sec	 accuracy=67.962652	 loss=1.312701	 lr=0.001000
Epoch[040] Batch [2099]/[3759]	Speed: 35.285078 samples/sec	 accuracy=67.968006	 loss=1.311361	 lr=0.001000
Epoch[040] Batch [2149]/[3759]	Speed: 35.167434 samples/sec	 accuracy=67.986919	 loss=1.310440	 lr=0.001000
Epoch[040] Batch [2199]/[3759]	Speed: 35.282922 samples/sec	 accuracy=68.026278	 loss=1.309345	 lr=0.001000
Epoch[040] Batch [2249]/[3759]	Speed: 35.141772 samples/sec	 accuracy=68.045139	 loss=1.307856	 lr=0.001000
Epoch[040] Batch [2299]/[3759]	Speed: 35.303401 samples/sec	 accuracy=68.066576	 loss=1.306805	 lr=0.001000
Epoch[040] Batch [2349]/[3759]	Speed: 35.136239 samples/sec	 accuracy=68.103059	 loss=1.305247	 lr=0.001000
Epoch[040] Batch [2399]/[3759]	Speed: 35.411797 samples/sec	 accuracy=68.123698	 loss=1.303744	 lr=0.001000
Epoch[040] Batch [2449]/[3759]	Speed: 35.333649 samples/sec	 accuracy=68.160714	 loss=1.303024	 lr=0.001000
Epoch[040] Batch [2499]/[3759]	Speed: 35.276876 samples/sec	 accuracy=68.205000	 loss=1.302392	 lr=0.001000
Epoch[040] Batch [2549]/[3759]	Speed: 34.998662 samples/sec	 accuracy=68.219363	 loss=1.301711	 lr=0.001000
Epoch[040] Batch [2599]/[3759]	Speed: 35.196925 samples/sec	 accuracy=68.229567	 loss=1.301935	 lr=0.001000
Epoch[040] Batch [2649]/[3759]	Speed: 35.100698 samples/sec	 accuracy=68.228774	 loss=1.301643	 lr=0.001000
Epoch[040] Batch [2699]/[3759]	Speed: 35.215673 samples/sec	 accuracy=68.236690	 loss=1.301238	 lr=0.001000
Epoch[040] Batch [2749]/[3759]	Speed: 35.073359 samples/sec	 accuracy=68.270455	 loss=1.299806	 lr=0.001000
Epoch[040] Batch [2799]/[3759]	Speed: 35.105198 samples/sec	 accuracy=68.297991	 loss=1.298631	 lr=0.001000
Epoch[040] Batch [2849]/[3759]	Speed: 34.866407 samples/sec	 accuracy=68.314693	 loss=1.298007	 lr=0.001000
Epoch[040] Batch [2899]/[3759]	Speed: 34.786967 samples/sec	 accuracy=68.341595	 loss=1.297394	 lr=0.001000
Epoch[040] Batch [2949]/[3759]	Speed: 35.005888 samples/sec	 accuracy=68.371822	 loss=1.296157	 lr=0.001000
Epoch[040] Batch [2999]/[3759]	Speed: 34.860117 samples/sec	 accuracy=68.380729	 loss=1.295253	 lr=0.001000
Epoch[040] Batch [3049]/[3759]	Speed: 35.249825 samples/sec	 accuracy=68.392418	 loss=1.294266	 lr=0.001000
Epoch[040] Batch [3099]/[3759]	Speed: 34.700069 samples/sec	 accuracy=68.424395	 loss=1.293168	 lr=0.001000
Epoch[040] Batch [3149]/[3759]	Speed: 35.285715 samples/sec	 accuracy=68.473214	 loss=1.290940	 lr=0.001000
Epoch[040] Batch [3199]/[3759]	Speed: 34.922789 samples/sec	 accuracy=68.489258	 loss=1.290092	 lr=0.001000
Epoch[040] Batch [3249]/[3759]	Speed: 35.540768 samples/sec	 accuracy=68.497596	 loss=1.289859	 lr=0.001000
Epoch[040] Batch [3299]/[3759]	Speed: 35.024542 samples/sec	 accuracy=68.529830	 loss=1.287897	 lr=0.001000
Epoch[040] Batch [3349]/[3759]	Speed: 35.334238 samples/sec	 accuracy=68.540112	 loss=1.287209	 lr=0.001000
Epoch[040] Batch [3399]/[3759]	Speed: 34.773390 samples/sec	 accuracy=68.559743	 loss=1.286041	 lr=0.001000
Epoch[040] Batch [3449]/[3759]	Speed: 35.232614 samples/sec	 accuracy=68.587862	 loss=1.284788	 lr=0.001000
Epoch[040] Batch [3499]/[3759]	Speed: 34.682245 samples/sec	 accuracy=68.605804	 loss=1.283698	 lr=0.001000
Epoch[040] Batch [3549]/[3759]	Speed: 35.125589 samples/sec	 accuracy=68.620158	 loss=1.282777	 lr=0.001000
Epoch[040] Batch [3599]/[3759]	Speed: 34.837398 samples/sec	 accuracy=68.646701	 loss=1.281725	 lr=0.001000
Epoch[040] Batch [3649]/[3759]	Speed: 34.901495 samples/sec	 accuracy=68.660959	 loss=1.281352	 lr=0.001000
Epoch[040] Batch [3699]/[3759]	Speed: 35.033952 samples/sec	 accuracy=68.658361	 loss=1.281394	 lr=0.001000
Epoch[040] Batch [3749]/[3759]	Speed: 36.314281 samples/sec	 accuracy=68.659583	 loss=1.281463	 lr=0.001000
Batch [0049]/[0303]: acc-top1=65.406250 acc-top5=85.062500
Batch [0099]/[0303]: acc-top1=64.718750 acc-top5=84.796875
Batch [0149]/[0303]: acc-top1=64.864583 acc-top5=85.250000
Batch [0199]/[0303]: acc-top1=64.796875 acc-top5=85.226562
Batch [0249]/[0303]: acc-top1=64.550000 acc-top5=84.893750
Batch [0299]/[0303]: acc-top1=64.427083 acc-top5=84.822917
[Epoch 040] training: accuracy=68.658553	 loss=1.281434
[Epoch 040] speed: 34 samples/sec	time cost: 7356.577198
[Epoch 040] validation: acc-top1=64.418317 acc-top5=84.839109 loss=1.656033
Epoch[041] Batch [0049]/[3760]	Speed: 24.102605 samples/sec	 accuracy=69.062500	 loss=1.247533	 lr=0.001000
Epoch[041] Batch [0099]/[3760]	Speed: 34.966660 samples/sec	 accuracy=69.578125	 loss=1.221175	 lr=0.001000
Epoch[041] Batch [0149]/[3760]	Speed: 34.903091 samples/sec	 accuracy=70.197917	 loss=1.208058	 lr=0.001000
Epoch[041] Batch [0199]/[3760]	Speed: 35.024444 samples/sec	 accuracy=69.750000	 loss=1.213377	 lr=0.001000
Epoch[041] Batch [0249]/[3760]	Speed: 35.113780 samples/sec	 accuracy=69.825000	 loss=1.218071	 lr=0.001000
Epoch[041] Batch [0299]/[3760]	Speed: 35.033191 samples/sec	 accuracy=69.848958	 loss=1.216834	 lr=0.001000
Epoch[041] Batch [0349]/[3760]	Speed: 35.190650 samples/sec	 accuracy=69.861607	 loss=1.221395	 lr=0.001000
Epoch[041] Batch [0399]/[3760]	Speed: 34.764597 samples/sec	 accuracy=70.097656	 loss=1.215312	 lr=0.001000
Epoch[041] Batch [0449]/[3760]	Speed: 35.388092 samples/sec	 accuracy=70.076389	 loss=1.213814	 lr=0.001000
Epoch[041] Batch [0499]/[3760]	Speed: 34.833092 samples/sec	 accuracy=70.115625	 loss=1.212288	 lr=0.001000
Epoch[041] Batch [0549]/[3760]	Speed: 35.388633 samples/sec	 accuracy=70.144886	 loss=1.211569	 lr=0.001000
Epoch[041] Batch [0599]/[3760]	Speed: 34.999473 samples/sec	 accuracy=70.197917	 loss=1.211034	 lr=0.001000
Epoch[041] Batch [0649]/[3760]	Speed: 35.414671 samples/sec	 accuracy=70.233173	 loss=1.208237	 lr=0.001000
Epoch[041] Batch [0699]/[3760]	Speed: 35.002292 samples/sec	 accuracy=70.189732	 loss=1.211330	 lr=0.001000
Epoch[041] Batch [0749]/[3760]	Speed: 35.287083 samples/sec	 accuracy=70.250000	 loss=1.209113	 lr=0.001000
Epoch[041] Batch [0799]/[3760]	Speed: 34.947363 samples/sec	 accuracy=70.218750	 loss=1.208455	 lr=0.001000
Epoch[041] Batch [0849]/[3760]	Speed: 35.244636 samples/sec	 accuracy=70.270221	 loss=1.206789	 lr=0.001000
Epoch[041] Batch [0899]/[3760]	Speed: 34.971214 samples/sec	 accuracy=70.274306	 loss=1.206407	 lr=0.001000
Epoch[041] Batch [0949]/[3760]	Speed: 35.067852 samples/sec	 accuracy=70.235197	 loss=1.208149	 lr=0.001000
Epoch[041] Batch [0999]/[3760]	Speed: 35.099518 samples/sec	 accuracy=70.264063	 loss=1.206312	 lr=0.001000
Epoch[041] Batch [1049]/[3760]	Speed: 35.147813 samples/sec	 accuracy=70.324405	 loss=1.204480	 lr=0.001000
Epoch[041] Batch [1099]/[3760]	Speed: 35.248066 samples/sec	 accuracy=70.400568	 loss=1.202042	 lr=0.001000
Epoch[041] Batch [1149]/[3760]	Speed: 35.264038 samples/sec	 accuracy=70.436141	 loss=1.200726	 lr=0.001000
Epoch[041] Batch [1199]/[3760]	Speed: 35.465888 samples/sec	 accuracy=70.417969	 loss=1.202466	 lr=0.001000
Epoch[041] Batch [1249]/[3760]	Speed: 35.235779 samples/sec	 accuracy=70.388750	 loss=1.204682	 lr=0.001000
Epoch[041] Batch [1299]/[3760]	Speed: 35.302333 samples/sec	 accuracy=70.397837	 loss=1.204655	 lr=0.001000
Epoch[041] Batch [1349]/[3760]	Speed: 35.182987 samples/sec	 accuracy=70.396991	 loss=1.204380	 lr=0.001000
Epoch[041] Batch [1399]/[3760]	Speed: 35.265004 samples/sec	 accuracy=70.431920	 loss=1.203600	 lr=0.001000
Epoch[041] Batch [1449]/[3760]	Speed: 35.407866 samples/sec	 accuracy=70.393319	 loss=1.204853	 lr=0.001000
Epoch[041] Batch [1499]/[3760]	Speed: 35.402210 samples/sec	 accuracy=70.434375	 loss=1.204395	 lr=0.001000
Epoch[041] Batch [1549]/[3760]	Speed: 35.127195 samples/sec	 accuracy=70.417339	 loss=1.204983	 lr=0.001000
Epoch[041] Batch [1599]/[3760]	Speed: 35.402825 samples/sec	 accuracy=70.410156	 loss=1.205483	 lr=0.001000
Epoch[041] Batch [1649]/[3760]	Speed: 35.269388 samples/sec	 accuracy=70.444129	 loss=1.203905	 lr=0.001000
Epoch[041] Batch [1699]/[3760]	Speed: 35.404848 samples/sec	 accuracy=70.431066	 loss=1.203858	 lr=0.001000
Epoch[041] Batch [1749]/[3760]	Speed: 35.057945 samples/sec	 accuracy=70.459821	 loss=1.202859	 lr=0.001000
Epoch[041] Batch [1799]/[3760]	Speed: 35.318682 samples/sec	 accuracy=70.467014	 loss=1.202869	 lr=0.001000
Epoch[041] Batch [1849]/[3760]	Speed: 35.159122 samples/sec	 accuracy=70.463682	 loss=1.204097	 lr=0.001000
Epoch[041] Batch [1899]/[3760]	Speed: 35.294060 samples/sec	 accuracy=70.469572	 loss=1.204054	 lr=0.001000
Epoch[041] Batch [1949]/[3760]	Speed: 35.139978 samples/sec	 accuracy=70.463942	 loss=1.204301	 lr=0.001000
Epoch[041] Batch [1999]/[3760]	Speed: 35.494097 samples/sec	 accuracy=70.478125	 loss=1.202686	 lr=0.001000
Epoch[041] Batch [2049]/[3760]	Speed: 34.913043 samples/sec	 accuracy=70.467988	 loss=1.203158	 lr=0.001000
Epoch[041] Batch [2099]/[3760]	Speed: 35.400576 samples/sec	 accuracy=70.477679	 loss=1.202522	 lr=0.001000
Epoch[041] Batch [2149]/[3760]	Speed: 35.146598 samples/sec	 accuracy=70.482558	 loss=1.202255	 lr=0.001000
Epoch[041] Batch [2199]/[3760]	Speed: 35.460333 samples/sec	 accuracy=70.473722	 loss=1.202669	 lr=0.001000
Epoch[041] Batch [2249]/[3760]	Speed: 34.951393 samples/sec	 accuracy=70.464583	 loss=1.203114	 lr=0.001000
Epoch[041] Batch [2299]/[3760]	Speed: 35.255854 samples/sec	 accuracy=70.483016	 loss=1.202870	 lr=0.001000
Epoch[041] Batch [2349]/[3760]	Speed: 35.100066 samples/sec	 accuracy=70.463431	 loss=1.203556	 lr=0.001000
Epoch[041] Batch [2399]/[3760]	Speed: 35.322842 samples/sec	 accuracy=70.447917	 loss=1.203442	 lr=0.001000
Epoch[041] Batch [2449]/[3760]	Speed: 35.122581 samples/sec	 accuracy=70.469388	 loss=1.202711	 lr=0.001000
Epoch[041] Batch [2499]/[3760]	Speed: 35.143998 samples/sec	 accuracy=70.462500	 loss=1.203349	 lr=0.001000
Epoch[041] Batch [2549]/[3760]	Speed: 35.093845 samples/sec	 accuracy=70.474265	 loss=1.202779	 lr=0.001000
Epoch[041] Batch [2599]/[3760]	Speed: 35.305042 samples/sec	 accuracy=70.472957	 loss=1.202996	 lr=0.001000
Epoch[041] Batch [2649]/[3760]	Speed: 35.414732 samples/sec	 accuracy=70.484670	 loss=1.202268	 lr=0.001000
Epoch[041] Batch [2699]/[3760]	Speed: 35.086141 samples/sec	 accuracy=70.491898	 loss=1.201550	 lr=0.001000
Epoch[041] Batch [2749]/[3760]	Speed: 35.316549 samples/sec	 accuracy=70.493182	 loss=1.201301	 lr=0.001000
Epoch[041] Batch [2799]/[3760]	Speed: 35.167458 samples/sec	 accuracy=70.489397	 loss=1.201411	 lr=0.001000
Epoch[041] Batch [2849]/[3760]	Speed: 35.255557 samples/sec	 accuracy=70.473136	 loss=1.201441	 lr=0.001000
Epoch[041] Batch [2899]/[3760]	Speed: 34.834662 samples/sec	 accuracy=70.474138	 loss=1.201146	 lr=0.001000
Epoch[041] Batch [2949]/[3760]	Speed: 35.420888 samples/sec	 accuracy=70.515890	 loss=1.200156	 lr=0.001000
Epoch[041] Batch [2999]/[3760]	Speed: 34.782888 samples/sec	 accuracy=70.520313	 loss=1.199549	 lr=0.001000
Epoch[041] Batch [3049]/[3760]	Speed: 35.181291 samples/sec	 accuracy=70.525102	 loss=1.198893	 lr=0.001000
Epoch[041] Batch [3099]/[3760]	Speed: 34.601294 samples/sec	 accuracy=70.551411	 loss=1.198040	 lr=0.001000
Epoch[041] Batch [3149]/[3760]	Speed: 35.177031 samples/sec	 accuracy=70.577381	 loss=1.196911	 lr=0.001000
Epoch[041] Batch [3199]/[3760]	Speed: 34.591878 samples/sec	 accuracy=70.592285	 loss=1.196165	 lr=0.001000
Epoch[041] Batch [3249]/[3760]	Speed: 35.120810 samples/sec	 accuracy=70.588942	 loss=1.196070	 lr=0.001000
Epoch[041] Batch [3299]/[3760]	Speed: 34.795676 samples/sec	 accuracy=70.618845	 loss=1.194838	 lr=0.001000
Epoch[041] Batch [3349]/[3760]	Speed: 35.011946 samples/sec	 accuracy=70.631530	 loss=1.194051	 lr=0.001000
Epoch[041] Batch [3399]/[3760]	Speed: 35.370923 samples/sec	 accuracy=70.653033	 loss=1.193009	 lr=0.001000
Epoch[041] Batch [3449]/[3760]	Speed: 34.882599 samples/sec	 accuracy=70.631793	 loss=1.193159	 lr=0.001000
Epoch[041] Batch [3499]/[3760]	Speed: 35.591504 samples/sec	 accuracy=70.647321	 loss=1.192203	 lr=0.001000
Epoch[041] Batch [3549]/[3760]	Speed: 35.339163 samples/sec	 accuracy=70.661532	 loss=1.191657	 lr=0.001000
Epoch[041] Batch [3599]/[3760]	Speed: 35.472633 samples/sec	 accuracy=70.667969	 loss=1.191380	 lr=0.001000
Epoch[041] Batch [3649]/[3760]	Speed: 35.261492 samples/sec	 accuracy=70.650257	 loss=1.191811	 lr=0.001000
Epoch[041] Batch [3699]/[3760]	Speed: 35.447481 samples/sec	 accuracy=70.665963	 loss=1.191317	 lr=0.001000
Epoch[041] Batch [3749]/[3760]	Speed: 36.466481 samples/sec	 accuracy=70.672917	 loss=1.191085	 lr=0.001000
Batch [0049]/[0303]: acc-top1=65.281250 acc-top5=85.625000
Batch [0099]/[0303]: acc-top1=65.250000 acc-top5=85.296875
Batch [0149]/[0303]: acc-top1=65.458333 acc-top5=85.458333
Batch [0199]/[0303]: acc-top1=65.343750 acc-top5=85.390625
Batch [0249]/[0303]: acc-top1=65.056250 acc-top5=85.206250
Batch [0299]/[0303]: acc-top1=65.005208 acc-top5=85.192708
[Epoch 041] training: accuracy=70.671958	 loss=1.190966
[Epoch 041] speed: 34 samples/sec	time cost: 7350.147787
[Epoch 041] validation: acc-top1=65.006188 acc-top5=85.194926 loss=1.636118
Epoch[042] Batch [0049]/[3760]	Speed: 24.255518 samples/sec	 accuracy=71.500000	 loss=1.115103	 lr=0.001000
Epoch[042] Batch [0099]/[3760]	Speed: 35.107611 samples/sec	 accuracy=71.093750	 loss=1.153348	 lr=0.001000
Epoch[042] Batch [0149]/[3760]	Speed: 34.949619 samples/sec	 accuracy=70.802083	 loss=1.162006	 lr=0.001000
Epoch[042] Batch [0199]/[3760]	Speed: 34.922655 samples/sec	 accuracy=71.328125	 loss=1.148038	 lr=0.001000
Epoch[042] Batch [0249]/[3760]	Speed: 35.015117 samples/sec	 accuracy=71.318750	 loss=1.151921	 lr=0.001000
Epoch[042] Batch [0299]/[3760]	Speed: 34.928640 samples/sec	 accuracy=71.296875	 loss=1.155605	 lr=0.001000
Epoch[042] Batch [0349]/[3760]	Speed: 35.233468 samples/sec	 accuracy=71.120536	 loss=1.160887	 lr=0.001000
Epoch[042] Batch [0399]/[3760]	Speed: 34.641608 samples/sec	 accuracy=71.167969	 loss=1.160085	 lr=0.001000
Epoch[042] Batch [0449]/[3760]	Speed: 35.062920 samples/sec	 accuracy=71.065972	 loss=1.164585	 lr=0.001000
Epoch[042] Batch [0499]/[3760]	Speed: 34.483361 samples/sec	 accuracy=71.059375	 loss=1.165248	 lr=0.001000
Epoch[042] Batch [0549]/[3760]	Speed: 35.269699 samples/sec	 accuracy=71.000000	 loss=1.167843	 lr=0.001000
Epoch[042] Batch [0599]/[3760]	Speed: 34.511404 samples/sec	 accuracy=71.114583	 loss=1.163993	 lr=0.001000
Epoch[042] Batch [0649]/[3760]	Speed: 35.225500 samples/sec	 accuracy=71.213942	 loss=1.162100	 lr=0.001000
Epoch[042] Batch [0699]/[3760]	Speed: 34.712362 samples/sec	 accuracy=71.207589	 loss=1.163763	 lr=0.001000
Epoch[042] Batch [0749]/[3760]	Speed: 34.572228 samples/sec	 accuracy=71.360417	 loss=1.158350	 lr=0.001000
Epoch[042] Batch [0799]/[3760]	Speed: 35.041472 samples/sec	 accuracy=71.378906	 loss=1.155581	 lr=0.001000
Epoch[042] Batch [0849]/[3760]	Speed: 34.855282 samples/sec	 accuracy=71.378676	 loss=1.156665	 lr=0.001000
Epoch[042] Batch [0899]/[3760]	Speed: 35.189135 samples/sec	 accuracy=71.376736	 loss=1.157582	 lr=0.001000
Epoch[042] Batch [0949]/[3760]	Speed: 34.811577 samples/sec	 accuracy=71.388158	 loss=1.158302	 lr=0.001000
Epoch[042] Batch [0999]/[3760]	Speed: 35.298439 samples/sec	 accuracy=71.382812	 loss=1.159094	 lr=0.001000
Epoch[042] Batch [1049]/[3760]	Speed: 34.638358 samples/sec	 accuracy=71.400298	 loss=1.158554	 lr=0.001000
Epoch[042] Batch [1099]/[3760]	Speed: 35.093894 samples/sec	 accuracy=71.367898	 loss=1.159343	 lr=0.001000
Epoch[042] Batch [1149]/[3760]	Speed: 34.781053 samples/sec	 accuracy=71.410326	 loss=1.158977	 lr=0.001000
Epoch[042] Batch [1199]/[3760]	Speed: 35.250403 samples/sec	 accuracy=71.446615	 loss=1.156845	 lr=0.001000
Epoch[042] Batch [1249]/[3760]	Speed: 34.822918 samples/sec	 accuracy=71.492500	 loss=1.154848	 lr=0.001000
Epoch[042] Batch [1299]/[3760]	Speed: 35.044082 samples/sec	 accuracy=71.507212	 loss=1.154676	 lr=0.001000
Epoch[042] Batch [1349]/[3760]	Speed: 34.841192 samples/sec	 accuracy=71.498843	 loss=1.155130	 lr=0.001000
Epoch[042] Batch [1399]/[3760]	Speed: 35.082620 samples/sec	 accuracy=71.534598	 loss=1.153728	 lr=0.001000
Epoch[042] Batch [1449]/[3760]	Speed: 34.873983 samples/sec	 accuracy=71.530172	 loss=1.153453	 lr=0.001000
Epoch[042] Batch [1499]/[3760]	Speed: 34.540690 samples/sec	 accuracy=71.563542	 loss=1.152516	 lr=0.001000
Epoch[042] Batch [1549]/[3760]	Speed: 34.992628 samples/sec	 accuracy=71.556452	 loss=1.153067	 lr=0.001000
Epoch[042] Batch [1599]/[3760]	Speed: 34.575982 samples/sec	 accuracy=71.540039	 loss=1.153339	 lr=0.001000
Epoch[042] Batch [1649]/[3760]	Speed: 35.258659 samples/sec	 accuracy=71.564394	 loss=1.152361	 lr=0.001000
Epoch[042] Batch [1699]/[3760]	Speed: 34.521742 samples/sec	 accuracy=71.568934	 loss=1.151989	 lr=0.001000
Epoch[042] Batch [1749]/[3760]	Speed: 35.344205 samples/sec	 accuracy=71.574107	 loss=1.151454	 lr=0.001000
Epoch[042] Batch [1799]/[3760]	Speed: 34.701643 samples/sec	 accuracy=71.557292	 loss=1.151934	 lr=0.001000
Epoch[042] Batch [1849]/[3760]	Speed: 34.954857 samples/sec	 accuracy=71.553209	 loss=1.152392	 lr=0.001000
Epoch[042] Batch [1899]/[3760]	Speed: 34.374082 samples/sec	 accuracy=71.527961	 loss=1.152763	 lr=0.001000
Epoch[042] Batch [1949]/[3760]	Speed: 34.878136 samples/sec	 accuracy=71.488782	 loss=1.153790	 lr=0.001000
Epoch[042] Batch [1999]/[3760]	Speed: 34.674558 samples/sec	 accuracy=71.495312	 loss=1.153436	 lr=0.001000
Epoch[042] Batch [2049]/[3760]	Speed: 34.598737 samples/sec	 accuracy=71.506860	 loss=1.153015	 lr=0.001000
Epoch[042] Batch [2099]/[3760]	Speed: 34.814096 samples/sec	 accuracy=71.466518	 loss=1.153843	 lr=0.001000
Epoch[042] Batch [2149]/[3760]	Speed: 34.397540 samples/sec	 accuracy=71.490552	 loss=1.152870	 lr=0.001000
Epoch[042] Batch [2199]/[3760]	Speed: 35.024518 samples/sec	 accuracy=71.489347	 loss=1.152686	 lr=0.001000
Epoch[042] Batch [2249]/[3760]	Speed: 34.343981 samples/sec	 accuracy=71.500000	 loss=1.152875	 lr=0.001000
Epoch[042] Batch [2299]/[3760]	Speed: 35.260960 samples/sec	 accuracy=71.508832	 loss=1.151818	 lr=0.001000
Epoch[042] Batch [2349]/[3760]	Speed: 34.519082 samples/sec	 accuracy=71.476064	 loss=1.153112	 lr=0.001000
Epoch[042] Batch [2399]/[3760]	Speed: 35.002031 samples/sec	 accuracy=71.481771	 loss=1.152550	 lr=0.001000
Epoch[042] Batch [2449]/[3760]	Speed: 34.905460 samples/sec	 accuracy=71.499362	 loss=1.152069	 lr=0.001000
Epoch[042] Batch [2499]/[3760]	Speed: 34.738347 samples/sec	 accuracy=71.479375	 loss=1.153178	 lr=0.001000
Epoch[042] Batch [2549]/[3760]	Speed: 35.163100 samples/sec	 accuracy=71.488358	 loss=1.152792	 lr=0.001000
Epoch[042] Batch [2599]/[3760]	Speed: 34.911017 samples/sec	 accuracy=71.473558	 loss=1.153482	 lr=0.001000
Epoch[042] Batch [2649]/[3760]	Speed: 35.383226 samples/sec	 accuracy=71.468750	 loss=1.153172	 lr=0.001000
Epoch[042] Batch [2699]/[3760]	Speed: 34.848907 samples/sec	 accuracy=71.472801	 loss=1.152758	 lr=0.001000
Epoch[042] Batch [2749]/[3760]	Speed: 35.287029 samples/sec	 accuracy=71.450568	 loss=1.153235	 lr=0.001000
Epoch[042] Batch [2799]/[3760]	Speed: 34.673333 samples/sec	 accuracy=71.427455	 loss=1.154127	 lr=0.001000
Epoch[042] Batch [2849]/[3760]	Speed: 35.429073 samples/sec	 accuracy=71.427632	 loss=1.154508	 lr=0.001000
Epoch[042] Batch [2899]/[3760]	Speed: 34.910684 samples/sec	 accuracy=71.431573	 loss=1.154104	 lr=0.001000
Epoch[042] Batch [2949]/[3760]	Speed: 35.472955 samples/sec	 accuracy=71.437500	 loss=1.154333	 lr=0.001000
Epoch[042] Batch [2999]/[3760]	Speed: 34.833741 samples/sec	 accuracy=71.420313	 loss=1.155302	 lr=0.001000
Epoch[042] Batch [3049]/[3760]	Speed: 35.464209 samples/sec	 accuracy=71.427254	 loss=1.155636	 lr=0.001000
Epoch[042] Batch [3099]/[3760]	Speed: 34.936494 samples/sec	 accuracy=71.427419	 loss=1.155812	 lr=0.001000
Epoch[042] Batch [3149]/[3760]	Speed: 35.295124 samples/sec	 accuracy=71.437500	 loss=1.155448	 lr=0.001000
Epoch[042] Batch [3199]/[3760]	Speed: 35.192493 samples/sec	 accuracy=71.437988	 loss=1.155329	 lr=0.001000
Epoch[042] Batch [3249]/[3760]	Speed: 35.411257 samples/sec	 accuracy=71.429808	 loss=1.155596	 lr=0.001000
Epoch[042] Batch [3299]/[3760]	Speed: 35.214640 samples/sec	 accuracy=71.437500	 loss=1.155450	 lr=0.001000
Epoch[042] Batch [3349]/[3760]	Speed: 35.341502 samples/sec	 accuracy=71.436101	 loss=1.155393	 lr=0.001000
Epoch[042] Batch [3399]/[3760]	Speed: 35.160670 samples/sec	 accuracy=71.424632	 loss=1.156004	 lr=0.001000
Epoch[042] Batch [3449]/[3760]	Speed: 35.295316 samples/sec	 accuracy=71.425725	 loss=1.156484	 lr=0.001000
Epoch[042] Batch [3499]/[3760]	Speed: 35.020332 samples/sec	 accuracy=71.437054	 loss=1.156484	 lr=0.001000
Epoch[042] Batch [3549]/[3760]	Speed: 35.252000 samples/sec	 accuracy=71.431778	 loss=1.157134	 lr=0.001000
Epoch[042] Batch [3599]/[3760]	Speed: 34.943547 samples/sec	 accuracy=71.453559	 loss=1.156476	 lr=0.001000
Epoch[042] Batch [3649]/[3760]	Speed: 35.235781 samples/sec	 accuracy=71.446062	 loss=1.156373	 lr=0.001000
Epoch[042] Batch [3699]/[3760]	Speed: 35.132728 samples/sec	 accuracy=71.444257	 loss=1.156552	 lr=0.001000
Epoch[042] Batch [3749]/[3760]	Speed: 36.394226 samples/sec	 accuracy=71.444583	 loss=1.156161	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.125000 acc-top5=85.656250
Batch [0099]/[0303]: acc-top1=65.609375 acc-top5=85.453125
Batch [0149]/[0303]: acc-top1=65.687500 acc-top5=85.625000
Batch [0199]/[0303]: acc-top1=65.578125 acc-top5=85.585938
Batch [0249]/[0303]: acc-top1=65.212500 acc-top5=85.368750
Batch [0299]/[0303]: acc-top1=65.145833 acc-top5=85.375000
[Epoch 042] training: accuracy=71.454870	 loss=1.155799
[Epoch 042] speed: 34 samples/sec	time cost: 7388.799814
[Epoch 042] validation: acc-top1=65.155734 acc-top5=85.380569 loss=1.627547
Epoch[043] Batch [0049]/[3759]	Speed: 24.262917 samples/sec	 accuracy=71.156250	 loss=1.148655	 lr=0.001000
Epoch[043] Batch [0099]/[3759]	Speed: 35.069283 samples/sec	 accuracy=72.109375	 loss=1.117640	 lr=0.001000
Epoch[043] Batch [0149]/[3759]	Speed: 34.823219 samples/sec	 accuracy=72.468750	 loss=1.105354	 lr=0.001000
Epoch[043] Batch [0199]/[3759]	Speed: 35.024643 samples/sec	 accuracy=72.437500	 loss=1.108535	 lr=0.001000
Epoch[043] Batch [0249]/[3759]	Speed: 35.278436 samples/sec	 accuracy=72.306250	 loss=1.106498	 lr=0.001000
Epoch[043] Batch [0299]/[3759]	Speed: 34.997715 samples/sec	 accuracy=72.270833	 loss=1.105186	 lr=0.001000
Epoch[043] Batch [0349]/[3759]	Speed: 35.304189 samples/sec	 accuracy=72.267857	 loss=1.105540	 lr=0.001000
Epoch[043] Batch [0399]/[3759]	Speed: 35.068029 samples/sec	 accuracy=72.304688	 loss=1.105411	 lr=0.001000
Epoch[043] Batch [0449]/[3759]	Speed: 35.293875 samples/sec	 accuracy=72.295139	 loss=1.106183	 lr=0.001000
Epoch[043] Batch [0499]/[3759]	Speed: 35.006616 samples/sec	 accuracy=72.228125	 loss=1.110579	 lr=0.001000
Epoch[043] Batch [0549]/[3759]	Speed: 35.393337 samples/sec	 accuracy=72.173295	 loss=1.113412	 lr=0.001000
Epoch[043] Batch [0599]/[3759]	Speed: 35.112307 samples/sec	 accuracy=72.179688	 loss=1.115042	 lr=0.001000
Epoch[043] Batch [0649]/[3759]	Speed: 35.462029 samples/sec	 accuracy=72.115385	 loss=1.116394	 lr=0.001000
Epoch[043] Batch [0699]/[3759]	Speed: 34.970511 samples/sec	 accuracy=72.120536	 loss=1.116646	 lr=0.001000
Epoch[043] Batch [0749]/[3759]	Speed: 35.492066 samples/sec	 accuracy=72.162500	 loss=1.116668	 lr=0.001000
Epoch[043] Batch [0799]/[3759]	Speed: 34.942249 samples/sec	 accuracy=72.058594	 loss=1.120816	 lr=0.001000
Epoch[043] Batch [0849]/[3759]	Speed: 35.462516 samples/sec	 accuracy=72.113971	 loss=1.120591	 lr=0.001000
Epoch[043] Batch [0899]/[3759]	Speed: 35.049542 samples/sec	 accuracy=72.076389	 loss=1.121710	 lr=0.001000
Epoch[043] Batch [0949]/[3759]	Speed: 35.506977 samples/sec	 accuracy=72.057566	 loss=1.122845	 lr=0.001000
Epoch[043] Batch [0999]/[3759]	Speed: 34.999820 samples/sec	 accuracy=72.104687	 loss=1.122214	 lr=0.001000
Epoch[043] Batch [1049]/[3759]	Speed: 35.350117 samples/sec	 accuracy=72.156250	 loss=1.121633	 lr=0.001000
Epoch[043] Batch [1099]/[3759]	Speed: 35.072472 samples/sec	 accuracy=72.173295	 loss=1.121512	 lr=0.001000
Epoch[043] Batch [1149]/[3759]	Speed: 35.241874 samples/sec	 accuracy=72.226902	 loss=1.120135	 lr=0.001000
Epoch[043] Batch [1199]/[3759]	Speed: 34.947458 samples/sec	 accuracy=72.255208	 loss=1.118911	 lr=0.001000
Epoch[043] Batch [1249]/[3759]	Speed: 34.974923 samples/sec	 accuracy=72.250000	 loss=1.118448	 lr=0.001000
Epoch[043] Batch [1299]/[3759]	Speed: 34.885079 samples/sec	 accuracy=72.266827	 loss=1.117404	 lr=0.001000
Epoch[043] Batch [1349]/[3759]	Speed: 34.739792 samples/sec	 accuracy=72.274306	 loss=1.118238	 lr=0.001000
Epoch[043] Batch [1399]/[3759]	Speed: 35.100806 samples/sec	 accuracy=72.238839	 loss=1.119687	 lr=0.001000
Epoch[043] Batch [1449]/[3759]	Speed: 34.739174 samples/sec	 accuracy=72.219828	 loss=1.119911	 lr=0.001000
Epoch[043] Batch [1499]/[3759]	Speed: 35.071215 samples/sec	 accuracy=72.189583	 loss=1.121205	 lr=0.001000
Epoch[043] Batch [1549]/[3759]	Speed: 34.557853 samples/sec	 accuracy=72.214718	 loss=1.120348	 lr=0.001000
Epoch[043] Batch [1599]/[3759]	Speed: 35.332656 samples/sec	 accuracy=72.210938	 loss=1.121396	 lr=0.001000
Epoch[043] Batch [1649]/[3759]	Speed: 34.833666 samples/sec	 accuracy=72.246212	 loss=1.120594	 lr=0.001000
Epoch[043] Batch [1699]/[3759]	Speed: 35.379394 samples/sec	 accuracy=72.247243	 loss=1.120911	 lr=0.001000
Epoch[043] Batch [1749]/[3759]	Speed: 35.118350 samples/sec	 accuracy=72.251786	 loss=1.120304	 lr=0.001000
Epoch[043] Batch [1799]/[3759]	Speed: 35.249238 samples/sec	 accuracy=72.219618	 loss=1.121148	 lr=0.001000
Epoch[043] Batch [1849]/[3759]	Speed: 34.907727 samples/sec	 accuracy=72.213682	 loss=1.122061	 lr=0.001000
Epoch[043] Batch [1899]/[3759]	Speed: 35.280417 samples/sec	 accuracy=72.200658	 loss=1.122293	 lr=0.001000
Epoch[043] Batch [1949]/[3759]	Speed: 35.027582 samples/sec	 accuracy=72.206731	 loss=1.122670	 lr=0.001000
Epoch[043] Batch [1999]/[3759]	Speed: 35.153310 samples/sec	 accuracy=72.211719	 loss=1.121889	 lr=0.001000
Epoch[043] Batch [2049]/[3759]	Speed: 35.135963 samples/sec	 accuracy=72.212652	 loss=1.120837	 lr=0.001000
Epoch[043] Batch [2099]/[3759]	Speed: 34.897958 samples/sec	 accuracy=72.206101	 loss=1.121986	 lr=0.001000
Epoch[043] Batch [2149]/[3759]	Speed: 35.348472 samples/sec	 accuracy=72.226017	 loss=1.122680	 lr=0.001000
Epoch[043] Batch [2199]/[3759]	Speed: 34.984356 samples/sec	 accuracy=72.234375	 loss=1.122669	 lr=0.001000
Epoch[043] Batch [2249]/[3759]	Speed: 35.486121 samples/sec	 accuracy=72.232639	 loss=1.123413	 lr=0.001000
Epoch[043] Batch [2299]/[3759]	Speed: 34.938156 samples/sec	 accuracy=72.240489	 loss=1.122940	 lr=0.001000
Epoch[043] Batch [2349]/[3759]	Speed: 35.352643 samples/sec	 accuracy=72.241356	 loss=1.122931	 lr=0.001000
Epoch[043] Batch [2399]/[3759]	Speed: 34.754184 samples/sec	 accuracy=72.261068	 loss=1.122172	 lr=0.001000
Epoch[043] Batch [2449]/[3759]	Speed: 35.214715 samples/sec	 accuracy=72.267219	 loss=1.121596	 lr=0.001000
Epoch[043] Batch [2499]/[3759]	Speed: 34.867801 samples/sec	 accuracy=72.261875	 loss=1.121636	 lr=0.001000
Epoch[043] Batch [2549]/[3759]	Speed: 35.163876 samples/sec	 accuracy=72.249387	 loss=1.122548	 lr=0.001000
Epoch[043] Batch [2599]/[3759]	Speed: 35.155863 samples/sec	 accuracy=72.231971	 loss=1.123752	 lr=0.001000
Epoch[043] Batch [2649]/[3759]	Speed: 35.111138 samples/sec	 accuracy=72.234080	 loss=1.124527	 lr=0.001000
Epoch[043] Batch [2699]/[3759]	Speed: 35.236678 samples/sec	 accuracy=72.230903	 loss=1.124744	 lr=0.001000
Epoch[043] Batch [2749]/[3759]	Speed: 35.032219 samples/sec	 accuracy=72.235227	 loss=1.124580	 lr=0.001000
Epoch[043] Batch [2799]/[3759]	Speed: 35.303248 samples/sec	 accuracy=72.232701	 loss=1.125360	 lr=0.001000
Epoch[043] Batch [2849]/[3759]	Speed: 34.978881 samples/sec	 accuracy=72.231360	 loss=1.125414	 lr=0.001000
Epoch[043] Batch [2899]/[3759]	Speed: 35.260838 samples/sec	 accuracy=72.198276	 loss=1.126106	 lr=0.001000
Epoch[043] Batch [2949]/[3759]	Speed: 34.856466 samples/sec	 accuracy=72.184322	 loss=1.126610	 lr=0.001000
Epoch[043] Batch [2999]/[3759]	Speed: 35.424602 samples/sec	 accuracy=72.186979	 loss=1.126369	 lr=0.001000
Epoch[043] Batch [3049]/[3759]	Speed: 34.563882 samples/sec	 accuracy=72.183402	 loss=1.126800	 lr=0.001000
Epoch[043] Batch [3099]/[3759]	Speed: 35.189638 samples/sec	 accuracy=72.182964	 loss=1.126586	 lr=0.001000
Epoch[043] Batch [3149]/[3759]	Speed: 34.629806 samples/sec	 accuracy=72.172123	 loss=1.126968	 lr=0.001000
Epoch[043] Batch [3199]/[3759]	Speed: 35.288417 samples/sec	 accuracy=72.169922	 loss=1.126647	 lr=0.001000
Epoch[043] Batch [3249]/[3759]	Speed: 35.088530 samples/sec	 accuracy=72.174519	 loss=1.126364	 lr=0.001000
Epoch[043] Batch [3299]/[3759]	Speed: 34.986289 samples/sec	 accuracy=72.178030	 loss=1.126345	 lr=0.001000
Epoch[043] Batch [3349]/[3759]	Speed: 35.257681 samples/sec	 accuracy=72.175373	 loss=1.125686	 lr=0.001000
Epoch[043] Batch [3399]/[3759]	Speed: 35.030537 samples/sec	 accuracy=72.185662	 loss=1.124917	 lr=0.001000
Epoch[043] Batch [3449]/[3759]	Speed: 35.125650 samples/sec	 accuracy=72.173913	 loss=1.124815	 lr=0.001000
Epoch[043] Batch [3499]/[3759]	Speed: 34.969305 samples/sec	 accuracy=72.163839	 loss=1.125124	 lr=0.001000
Epoch[043] Batch [3549]/[3759]	Speed: 35.189874 samples/sec	 accuracy=72.163732	 loss=1.125485	 lr=0.001000
Epoch[043] Batch [3599]/[3759]	Speed: 34.699368 samples/sec	 accuracy=72.189236	 loss=1.124827	 lr=0.001000
Epoch[043] Batch [3649]/[3759]	Speed: 35.222350 samples/sec	 accuracy=72.205051	 loss=1.124034	 lr=0.001000
Epoch[043] Batch [3699]/[3759]	Speed: 34.649198 samples/sec	 accuracy=72.214105	 loss=1.123813	 lr=0.001000
Epoch[043] Batch [3749]/[3759]	Speed: 36.458264 samples/sec	 accuracy=72.187500	 loss=1.124202	 lr=0.001000
Batch [0049]/[0303]: acc-top1=65.718750 acc-top5=85.781250
Batch [0099]/[0303]: acc-top1=65.234375 acc-top5=85.265625
Batch [0149]/[0303]: acc-top1=65.645833 acc-top5=85.479167
Batch [0199]/[0303]: acc-top1=65.554688 acc-top5=85.453125
Batch [0249]/[0303]: acc-top1=65.393750 acc-top5=85.281250
Batch [0299]/[0303]: acc-top1=65.276042 acc-top5=85.255208
[Epoch 043] training: accuracy=72.185089	 loss=1.124298
[Epoch 043] speed: 34 samples/sec	time cost: 7363.448868
[Epoch 043] validation: acc-top1=65.243399 acc-top5=85.272277 loss=1.629479
Epoch[044] Batch [0049]/[3760]	Speed: 24.787090 samples/sec	 accuracy=71.281250	 loss=1.130905	 lr=0.001000
Epoch[044] Batch [0099]/[3760]	Speed: 35.147550 samples/sec	 accuracy=71.609375	 loss=1.129974	 lr=0.001000
Epoch[044] Batch [0149]/[3760]	Speed: 35.014826 samples/sec	 accuracy=72.197917	 loss=1.115126	 lr=0.001000
Epoch[044] Batch [0199]/[3760]	Speed: 34.810128 samples/sec	 accuracy=72.585938	 loss=1.098491	 lr=0.001000
Epoch[044] Batch [0249]/[3760]	Speed: 35.153098 samples/sec	 accuracy=73.068750	 loss=1.086899	 lr=0.001000
Epoch[044] Batch [0299]/[3760]	Speed: 35.011275 samples/sec	 accuracy=73.177083	 loss=1.084890	 lr=0.001000
Epoch[044] Batch [0349]/[3760]	Speed: 35.308308 samples/sec	 accuracy=73.321429	 loss=1.080384	 lr=0.001000
Epoch[044] Batch [0399]/[3760]	Speed: 35.087181 samples/sec	 accuracy=73.281250	 loss=1.082543	 lr=0.001000
Epoch[044] Batch [0449]/[3760]	Speed: 35.157742 samples/sec	 accuracy=73.187500	 loss=1.086231	 lr=0.001000
Epoch[044] Batch [0499]/[3760]	Speed: 34.850652 samples/sec	 accuracy=73.128125	 loss=1.084255	 lr=0.001000
Epoch[044] Batch [0549]/[3760]	Speed: 35.250180 samples/sec	 accuracy=72.971591	 loss=1.092039	 lr=0.001000
Epoch[044] Batch [0599]/[3760]	Speed: 34.642980 samples/sec	 accuracy=72.880208	 loss=1.093191	 lr=0.001000
Epoch[044] Batch [0649]/[3760]	Speed: 35.456127 samples/sec	 accuracy=72.918269	 loss=1.091444	 lr=0.001000
Epoch[044] Batch [0699]/[3760]	Speed: 34.961633 samples/sec	 accuracy=72.924107	 loss=1.092282	 lr=0.001000
Epoch[044] Batch [0749]/[3760]	Speed: 35.320298 samples/sec	 accuracy=72.810417	 loss=1.097433	 lr=0.001000
Epoch[044] Batch [0799]/[3760]	Speed: 34.870772 samples/sec	 accuracy=72.794922	 loss=1.099571	 lr=0.001000
Epoch[044] Batch [0849]/[3760]	Speed: 35.329588 samples/sec	 accuracy=72.762868	 loss=1.101599	 lr=0.001000
Epoch[044] Batch [0899]/[3760]	Speed: 35.155067 samples/sec	 accuracy=72.763889	 loss=1.100740	 lr=0.001000
Epoch[044] Batch [0949]/[3760]	Speed: 35.023827 samples/sec	 accuracy=72.754934	 loss=1.100928	 lr=0.001000
Epoch[044] Batch [0999]/[3760]	Speed: 35.193256 samples/sec	 accuracy=72.685937	 loss=1.103448	 lr=0.001000
Epoch[044] Batch [1049]/[3760]	Speed: 34.971496 samples/sec	 accuracy=72.694940	 loss=1.101649	 lr=0.001000
Epoch[044] Batch [1099]/[3760]	Speed: 35.182253 samples/sec	 accuracy=72.656250	 loss=1.102730	 lr=0.001000
Epoch[044] Batch [1149]/[3760]	Speed: 34.731838 samples/sec	 accuracy=72.614130	 loss=1.103919	 lr=0.001000
Epoch[044] Batch [1199]/[3760]	Speed: 35.476774 samples/sec	 accuracy=72.660156	 loss=1.103236	 lr=0.001000
Epoch[044] Batch [1249]/[3760]	Speed: 34.937012 samples/sec	 accuracy=72.616250	 loss=1.103794	 lr=0.001000
Epoch[044] Batch [1299]/[3760]	Speed: 35.512445 samples/sec	 accuracy=72.676683	 loss=1.101498	 lr=0.001000
Epoch[044] Batch [1349]/[3760]	Speed: 34.874647 samples/sec	 accuracy=72.674769	 loss=1.102192	 lr=0.001000
Epoch[044] Batch [1399]/[3760]	Speed: 35.189408 samples/sec	 accuracy=72.650670	 loss=1.103325	 lr=0.001000
Epoch[044] Batch [1449]/[3760]	Speed: 34.790667 samples/sec	 accuracy=72.634698	 loss=1.104526	 lr=0.001000
Epoch[044] Batch [1499]/[3760]	Speed: 35.151655 samples/sec	 accuracy=72.597917	 loss=1.105649	 lr=0.001000
Epoch[044] Batch [1549]/[3760]	Speed: 35.008301 samples/sec	 accuracy=72.588710	 loss=1.107042	 lr=0.001000
Epoch[044] Batch [1599]/[3760]	Speed: 35.404074 samples/sec	 accuracy=72.543945	 loss=1.108785	 lr=0.001000
Epoch[044] Batch [1649]/[3760]	Speed: 35.279893 samples/sec	 accuracy=72.539773	 loss=1.108254	 lr=0.001000
Epoch[044] Batch [1699]/[3760]	Speed: 35.030943 samples/sec	 accuracy=72.498162	 loss=1.108972	 lr=0.001000
Epoch[044] Batch [1749]/[3760]	Speed: 35.372387 samples/sec	 accuracy=72.531250	 loss=1.107942	 lr=0.001000
Epoch[044] Batch [1799]/[3760]	Speed: 35.248918 samples/sec	 accuracy=72.533854	 loss=1.107303	 lr=0.001000
Epoch[044] Batch [1849]/[3760]	Speed: 35.357957 samples/sec	 accuracy=72.559966	 loss=1.105982	 lr=0.001000
Epoch[044] Batch [1899]/[3760]	Speed: 35.110304 samples/sec	 accuracy=72.549342	 loss=1.106174	 lr=0.001000
Epoch[044] Batch [1949]/[3760]	Speed: 35.324983 samples/sec	 accuracy=72.520032	 loss=1.106284	 lr=0.001000
Epoch[044] Batch [1999]/[3760]	Speed: 35.107926 samples/sec	 accuracy=72.500000	 loss=1.106744	 lr=0.001000
Epoch[044] Batch [2049]/[3760]	Speed: 35.460263 samples/sec	 accuracy=72.502287	 loss=1.106292	 lr=0.001000
Epoch[044] Batch [2099]/[3760]	Speed: 35.070678 samples/sec	 accuracy=72.473958	 loss=1.107904	 lr=0.001000
Epoch[044] Batch [2149]/[3760]	Speed: 35.238577 samples/sec	 accuracy=72.454942	 loss=1.108074	 lr=0.001000
Epoch[044] Batch [2199]/[3760]	Speed: 34.841047 samples/sec	 accuracy=72.473011	 loss=1.107185	 lr=0.001000
Epoch[044] Batch [2249]/[3760]	Speed: 35.199487 samples/sec	 accuracy=72.454167	 loss=1.107931	 lr=0.001000
Epoch[044] Batch [2299]/[3760]	Speed: 34.613587 samples/sec	 accuracy=72.438179	 loss=1.108662	 lr=0.001000
Epoch[044] Batch [2349]/[3760]	Speed: 35.179343 samples/sec	 accuracy=72.468085	 loss=1.107922	 lr=0.001000
Epoch[044] Batch [2399]/[3760]	Speed: 35.107577 samples/sec	 accuracy=72.451172	 loss=1.108869	 lr=0.001000
Epoch[044] Batch [2449]/[3760]	Speed: 34.916437 samples/sec	 accuracy=72.452168	 loss=1.108651	 lr=0.001000
Epoch[044] Batch [2499]/[3760]	Speed: 35.356271 samples/sec	 accuracy=72.462500	 loss=1.108382	 lr=0.001000
Epoch[044] Batch [2549]/[3760]	Speed: 34.973585 samples/sec	 accuracy=72.482843	 loss=1.108142	 lr=0.001000
Epoch[044] Batch [2599]/[3760]	Speed: 35.300654 samples/sec	 accuracy=72.487981	 loss=1.108542	 lr=0.001000
Epoch[044] Batch [2649]/[3760]	Speed: 34.862334 samples/sec	 accuracy=72.462264	 loss=1.109405	 lr=0.001000
Epoch[044] Batch [2699]/[3760]	Speed: 35.589827 samples/sec	 accuracy=72.465278	 loss=1.109319	 lr=0.001000
Epoch[044] Batch [2749]/[3760]	Speed: 35.205371 samples/sec	 accuracy=72.478409	 loss=1.108707	 lr=0.001000
Epoch[044] Batch [2799]/[3760]	Speed: 35.375578 samples/sec	 accuracy=72.477679	 loss=1.108950	 lr=0.001000
Epoch[044] Batch [2849]/[3760]	Speed: 35.215185 samples/sec	 accuracy=72.475877	 loss=1.109088	 lr=0.001000
Epoch[044] Batch [2899]/[3760]	Speed: 35.460516 samples/sec	 accuracy=72.480603	 loss=1.108876	 lr=0.001000
Epoch[044] Batch [2949]/[3760]	Speed: 34.951968 samples/sec	 accuracy=72.480403	 loss=1.108820	 lr=0.001000
Epoch[044] Batch [2999]/[3760]	Speed: 35.364246 samples/sec	 accuracy=72.496354	 loss=1.107919	 lr=0.001000
Epoch[044] Batch [3049]/[3760]	Speed: 34.891875 samples/sec	 accuracy=72.502049	 loss=1.107822	 lr=0.001000
Epoch[044] Batch [3099]/[3760]	Speed: 35.382501 samples/sec	 accuracy=72.490423	 loss=1.107834	 lr=0.001000
Epoch[044] Batch [3149]/[3760]	Speed: 34.827104 samples/sec	 accuracy=72.488591	 loss=1.108505	 lr=0.001000
Epoch[044] Batch [3199]/[3760]	Speed: 35.396636 samples/sec	 accuracy=72.482422	 loss=1.109104	 lr=0.001000
Epoch[044] Batch [3249]/[3760]	Speed: 35.256340 samples/sec	 accuracy=72.482692	 loss=1.109054	 lr=0.001000
Epoch[044] Batch [3299]/[3760]	Speed: 35.375186 samples/sec	 accuracy=72.472064	 loss=1.109198	 lr=0.001000
Epoch[044] Batch [3349]/[3760]	Speed: 35.214243 samples/sec	 accuracy=72.465485	 loss=1.109404	 lr=0.001000
Epoch[044] Batch [3399]/[3760]	Speed: 35.317114 samples/sec	 accuracy=72.477941	 loss=1.109716	 lr=0.001000
Epoch[044] Batch [3449]/[3760]	Speed: 35.115715 samples/sec	 accuracy=72.488678	 loss=1.109274	 lr=0.001000
Epoch[044] Batch [3499]/[3760]	Speed: 35.212255 samples/sec	 accuracy=72.485268	 loss=1.109721	 lr=0.001000
Epoch[044] Batch [3549]/[3760]	Speed: 35.269804 samples/sec	 accuracy=72.486796	 loss=1.109573	 lr=0.001000
Epoch[044] Batch [3599]/[3760]	Speed: 35.320003 samples/sec	 accuracy=72.483073	 loss=1.109789	 lr=0.001000
Epoch[044] Batch [3649]/[3760]	Speed: 35.217205 samples/sec	 accuracy=72.468750	 loss=1.110395	 lr=0.001000
Epoch[044] Batch [3699]/[3760]	Speed: 35.198945 samples/sec	 accuracy=72.473395	 loss=1.110453	 lr=0.001000
Epoch[044] Batch [3749]/[3760]	Speed: 36.462227 samples/sec	 accuracy=72.473333	 loss=1.110058	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.187500 acc-top5=85.375000
Batch [0099]/[0303]: acc-top1=65.609375 acc-top5=85.078125
Batch [0149]/[0303]: acc-top1=65.937500 acc-top5=85.406250
Batch [0199]/[0303]: acc-top1=65.804688 acc-top5=85.429688
Batch [0249]/[0303]: acc-top1=65.600000 acc-top5=85.250000
Batch [0299]/[0303]: acc-top1=65.520833 acc-top5=85.348958
[Epoch 044] training: accuracy=72.464678	 loss=1.110181
[Epoch 044] speed: 34 samples/sec	time cost: 7346.862277
[Epoch 044] validation: acc-top1=65.516708 acc-top5=85.375413 loss=1.605682
Epoch[045] Batch [0049]/[3760]	Speed: 24.863336 samples/sec	 accuracy=72.000000	 loss=1.093008	 lr=0.001000
Epoch[045] Batch [0099]/[3760]	Speed: 35.127171 samples/sec	 accuracy=71.890625	 loss=1.107295	 lr=0.001000
Epoch[045] Batch [0149]/[3760]	Speed: 35.027766 samples/sec	 accuracy=72.427083	 loss=1.092795	 lr=0.001000
Epoch[045] Batch [0199]/[3760]	Speed: 35.238403 samples/sec	 accuracy=72.906250	 loss=1.075650	 lr=0.001000
Epoch[045] Batch [0249]/[3760]	Speed: 35.049396 samples/sec	 accuracy=72.968750	 loss=1.076987	 lr=0.001000
Epoch[045] Batch [0299]/[3760]	Speed: 34.774181 samples/sec	 accuracy=72.921875	 loss=1.076097	 lr=0.001000
Epoch[045] Batch [0349]/[3760]	Speed: 35.106846 samples/sec	 accuracy=72.928571	 loss=1.076127	 lr=0.001000
Epoch[045] Batch [0399]/[3760]	Speed: 34.740131 samples/sec	 accuracy=73.039062	 loss=1.075719	 lr=0.001000
Epoch[045] Batch [0449]/[3760]	Speed: 35.185070 samples/sec	 accuracy=73.000000	 loss=1.077557	 lr=0.001000
Epoch[045] Batch [0499]/[3760]	Speed: 34.514358 samples/sec	 accuracy=73.228125	 loss=1.072019	 lr=0.001000
Epoch[045] Batch [0549]/[3760]	Speed: 35.364509 samples/sec	 accuracy=73.230114	 loss=1.075059	 lr=0.001000
Epoch[045] Batch [0599]/[3760]	Speed: 34.859400 samples/sec	 accuracy=73.174479	 loss=1.079772	 lr=0.001000
Epoch[045] Batch [0649]/[3760]	Speed: 35.222700 samples/sec	 accuracy=73.127404	 loss=1.080147	 lr=0.001000
Epoch[045] Batch [0699]/[3760]	Speed: 34.633967 samples/sec	 accuracy=73.136161	 loss=1.081072	 lr=0.001000
Epoch[045] Batch [0749]/[3760]	Speed: 34.991110 samples/sec	 accuracy=73.150000	 loss=1.080471	 lr=0.001000
Epoch[045] Batch [0799]/[3760]	Speed: 34.936875 samples/sec	 accuracy=73.126953	 loss=1.079845	 lr=0.001000
Epoch[045] Batch [0849]/[3760]	Speed: 34.916687 samples/sec	 accuracy=73.235294	 loss=1.078775	 lr=0.001000
Epoch[045] Batch [0899]/[3760]	Speed: 35.180946 samples/sec	 accuracy=73.199653	 loss=1.078657	 lr=0.001000
Epoch[045] Batch [0949]/[3760]	Speed: 35.025692 samples/sec	 accuracy=73.159539	 loss=1.077963	 lr=0.001000
Epoch[045] Batch [0999]/[3760]	Speed: 35.193548 samples/sec	 accuracy=73.210938	 loss=1.075990	 lr=0.001000
Epoch[045] Batch [1049]/[3760]	Speed: 34.839087 samples/sec	 accuracy=73.172619	 loss=1.076361	 lr=0.001000
Epoch[045] Batch [1099]/[3760]	Speed: 35.398375 samples/sec	 accuracy=73.214489	 loss=1.075558	 lr=0.001000
Epoch[045] Batch [1149]/[3760]	Speed: 34.847256 samples/sec	 accuracy=73.235054	 loss=1.076827	 lr=0.001000
Epoch[045] Batch [1199]/[3760]	Speed: 35.252878 samples/sec	 accuracy=73.157552	 loss=1.080295	 lr=0.001000
Epoch[045] Batch [1249]/[3760]	Speed: 34.545696 samples/sec	 accuracy=73.140000	 loss=1.082418	 lr=0.001000
Epoch[045] Batch [1299]/[3760]	Speed: 35.152406 samples/sec	 accuracy=73.112981	 loss=1.084394	 lr=0.001000
Epoch[045] Batch [1349]/[3760]	Speed: 35.114655 samples/sec	 accuracy=73.190972	 loss=1.080652	 lr=0.001000
Epoch[045] Batch [1399]/[3760]	Speed: 35.095998 samples/sec	 accuracy=73.204241	 loss=1.080315	 lr=0.001000
Epoch[045] Batch [1449]/[3760]	Speed: 35.136564 samples/sec	 accuracy=73.205819	 loss=1.080036	 lr=0.001000
Epoch[045] Batch [1499]/[3760]	Speed: 35.105684 samples/sec	 accuracy=73.206250	 loss=1.079903	 lr=0.001000
Epoch[045] Batch [1549]/[3760]	Speed: 35.128893 samples/sec	 accuracy=73.178427	 loss=1.081010	 lr=0.001000
Epoch[045] Batch [1599]/[3760]	Speed: 35.092034 samples/sec	 accuracy=73.153320	 loss=1.081508	 lr=0.001000
Epoch[045] Batch [1649]/[3760]	Speed: 35.462835 samples/sec	 accuracy=73.160038	 loss=1.081103	 lr=0.001000
Epoch[045] Batch [1699]/[3760]	Speed: 34.981739 samples/sec	 accuracy=73.117647	 loss=1.082781	 lr=0.001000
Epoch[045] Batch [1749]/[3760]	Speed: 35.407023 samples/sec	 accuracy=73.124107	 loss=1.082143	 lr=0.001000
Epoch[045] Batch [1799]/[3760]	Speed: 35.097074 samples/sec	 accuracy=73.131944	 loss=1.082517	 lr=0.001000
Epoch[045] Batch [1849]/[3760]	Speed: 35.332264 samples/sec	 accuracy=73.141892	 loss=1.081783	 lr=0.001000
Epoch[045] Batch [1899]/[3760]	Speed: 34.978836 samples/sec	 accuracy=73.171053	 loss=1.080726	 lr=0.001000
Epoch[045] Batch [1949]/[3760]	Speed: 35.357320 samples/sec	 accuracy=73.170673	 loss=1.080195	 lr=0.001000
Epoch[045] Batch [1999]/[3760]	Speed: 34.850006 samples/sec	 accuracy=73.142188	 loss=1.081395	 lr=0.001000
Epoch[045] Batch [2049]/[3760]	Speed: 35.372112 samples/sec	 accuracy=73.115854	 loss=1.082520	 lr=0.001000
Epoch[045] Batch [2099]/[3760]	Speed: 34.838770 samples/sec	 accuracy=73.104167	 loss=1.082309	 lr=0.001000
Epoch[045] Batch [2149]/[3760]	Speed: 35.259581 samples/sec	 accuracy=73.067587	 loss=1.084424	 lr=0.001000
Epoch[045] Batch [2199]/[3760]	Speed: 35.003138 samples/sec	 accuracy=73.073864	 loss=1.084581	 lr=0.001000
Epoch[045] Batch [2249]/[3760]	Speed: 35.139402 samples/sec	 accuracy=73.068750	 loss=1.085397	 lr=0.001000
Epoch[045] Batch [2299]/[3760]	Speed: 35.187140 samples/sec	 accuracy=73.072011	 loss=1.085140	 lr=0.001000
Epoch[045] Batch [2349]/[3760]	Speed: 34.814049 samples/sec	 accuracy=73.071809	 loss=1.085771	 lr=0.001000
Epoch[045] Batch [2399]/[3760]	Speed: 35.233961 samples/sec	 accuracy=73.083984	 loss=1.085566	 lr=0.001000
Epoch[045] Batch [2449]/[3760]	Speed: 34.731514 samples/sec	 accuracy=73.078444	 loss=1.086402	 lr=0.001000
Epoch[045] Batch [2499]/[3760]	Speed: 35.571767 samples/sec	 accuracy=73.044375	 loss=1.087063	 lr=0.001000
Epoch[045] Batch [2549]/[3760]	Speed: 35.218303 samples/sec	 accuracy=73.056985	 loss=1.086722	 lr=0.001000
Epoch[045] Batch [2599]/[3760]	Speed: 35.348179 samples/sec	 accuracy=73.051683	 loss=1.087091	 lr=0.001000
Epoch[045] Batch [2649]/[3760]	Speed: 35.099162 samples/sec	 accuracy=73.024175	 loss=1.088231	 lr=0.001000
Epoch[045] Batch [2699]/[3760]	Speed: 35.479962 samples/sec	 accuracy=73.024884	 loss=1.088473	 lr=0.001000
Epoch[045] Batch [2749]/[3760]	Speed: 35.226694 samples/sec	 accuracy=73.010795	 loss=1.088105	 lr=0.001000
Epoch[045] Batch [2799]/[3760]	Speed: 35.248260 samples/sec	 accuracy=73.006696	 loss=1.088303	 lr=0.001000
Epoch[045] Batch [2849]/[3760]	Speed: 35.426911 samples/sec	 accuracy=73.032895	 loss=1.087859	 lr=0.001000
Epoch[045] Batch [2899]/[3760]	Speed: 35.292604 samples/sec	 accuracy=73.006466	 loss=1.088585	 lr=0.001000
Epoch[045] Batch [2949]/[3760]	Speed: 35.379890 samples/sec	 accuracy=73.016419	 loss=1.088008	 lr=0.001000
Epoch[045] Batch [2999]/[3760]	Speed: 35.167968 samples/sec	 accuracy=73.023958	 loss=1.087946	 lr=0.001000
Epoch[045] Batch [3049]/[3760]	Speed: 35.341531 samples/sec	 accuracy=73.006660	 loss=1.088865	 lr=0.001000
Epoch[045] Batch [3099]/[3760]	Speed: 35.251901 samples/sec	 accuracy=73.017137	 loss=1.088000	 lr=0.001000
Epoch[045] Batch [3149]/[3760]	Speed: 35.025686 samples/sec	 accuracy=73.024802	 loss=1.087871	 lr=0.001000
Epoch[045] Batch [3199]/[3760]	Speed: 35.235751 samples/sec	 accuracy=73.015625	 loss=1.088523	 lr=0.001000
Epoch[045] Batch [3249]/[3760]	Speed: 34.715616 samples/sec	 accuracy=73.028365	 loss=1.088191	 lr=0.001000
Epoch[045] Batch [3299]/[3760]	Speed: 35.265417 samples/sec	 accuracy=73.032197	 loss=1.087716	 lr=0.001000
Epoch[045] Batch [3349]/[3760]	Speed: 34.584676 samples/sec	 accuracy=73.014459	 loss=1.088057	 lr=0.001000
Epoch[045] Batch [3399]/[3760]	Speed: 35.334678 samples/sec	 accuracy=73.009651	 loss=1.088258	 lr=0.001000
Epoch[045] Batch [3449]/[3760]	Speed: 34.740869 samples/sec	 accuracy=73.002717	 loss=1.088727	 lr=0.001000
Epoch[045] Batch [3499]/[3760]	Speed: 35.169299 samples/sec	 accuracy=73.005357	 loss=1.089044	 lr=0.001000
Epoch[045] Batch [3549]/[3760]	Speed: 34.831870 samples/sec	 accuracy=73.006162	 loss=1.089004	 lr=0.001000
Epoch[045] Batch [3599]/[3760]	Speed: 35.107184 samples/sec	 accuracy=72.999566	 loss=1.089133	 lr=0.001000
Epoch[045] Batch [3649]/[3760]	Speed: 34.812646 samples/sec	 accuracy=72.983305	 loss=1.089751	 lr=0.001000
Epoch[045] Batch [3699]/[3760]	Speed: 35.073123 samples/sec	 accuracy=72.990287	 loss=1.089572	 lr=0.001000
Epoch[045] Batch [3749]/[3760]	Speed: 36.304198 samples/sec	 accuracy=72.993333	 loss=1.089358	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.781250 acc-top5=86.125000
Batch [0099]/[0303]: acc-top1=66.000000 acc-top5=85.328125
Batch [0149]/[0303]: acc-top1=66.166667 acc-top5=85.572917
Batch [0199]/[0303]: acc-top1=65.984375 acc-top5=85.539062
Batch [0249]/[0303]: acc-top1=65.731250 acc-top5=85.412500
Batch [0299]/[0303]: acc-top1=65.713542 acc-top5=85.427083
[Epoch 045] training: accuracy=72.994099	 loss=1.089425
[Epoch 045] speed: 34 samples/sec	time cost: 7357.832250
[Epoch 045] validation: acc-top1=65.686881 acc-top5=85.432137 loss=1.610307
Epoch[046] Batch [0049]/[3759]	Speed: 24.650047 samples/sec	 accuracy=73.500000	 loss=1.062085	 lr=0.001000
Epoch[046] Batch [0099]/[3759]	Speed: 35.052852 samples/sec	 accuracy=73.531250	 loss=1.063567	 lr=0.001000
Epoch[046] Batch [0149]/[3759]	Speed: 34.838283 samples/sec	 accuracy=73.604167	 loss=1.063306	 lr=0.001000
Epoch[046] Batch [0199]/[3759]	Speed: 34.935508 samples/sec	 accuracy=73.914062	 loss=1.054249	 lr=0.001000
Epoch[046] Batch [0249]/[3759]	Speed: 35.042691 samples/sec	 accuracy=73.862500	 loss=1.055548	 lr=0.001000
Epoch[046] Batch [0299]/[3759]	Speed: 34.789155 samples/sec	 accuracy=73.817708	 loss=1.061101	 lr=0.001000
Epoch[046] Batch [0349]/[3759]	Speed: 35.101024 samples/sec	 accuracy=73.535714	 loss=1.067500	 lr=0.001000
Epoch[046] Batch [0399]/[3759]	Speed: 34.763072 samples/sec	 accuracy=73.464844	 loss=1.066378	 lr=0.001000
Epoch[046] Batch [0449]/[3759]	Speed: 35.063263 samples/sec	 accuracy=73.513889	 loss=1.066449	 lr=0.001000
Epoch[046] Batch [0499]/[3759]	Speed: 34.531996 samples/sec	 accuracy=73.581250	 loss=1.063747	 lr=0.001000
Epoch[046] Batch [0549]/[3759]	Speed: 35.362138 samples/sec	 accuracy=73.607955	 loss=1.061899	 lr=0.001000
Epoch[046] Batch [0599]/[3759]	Speed: 34.816036 samples/sec	 accuracy=73.486979	 loss=1.062595	 lr=0.001000
Epoch[046] Batch [0649]/[3759]	Speed: 35.256577 samples/sec	 accuracy=73.468750	 loss=1.063683	 lr=0.001000
Epoch[046] Batch [0699]/[3759]	Speed: 35.011108 samples/sec	 accuracy=73.520089	 loss=1.063623	 lr=0.001000
Epoch[046] Batch [0749]/[3759]	Speed: 35.224900 samples/sec	 accuracy=73.452083	 loss=1.067133	 lr=0.001000
Epoch[046] Batch [0799]/[3759]	Speed: 34.863931 samples/sec	 accuracy=73.460938	 loss=1.067883	 lr=0.001000
Epoch[046] Batch [0849]/[3759]	Speed: 35.242335 samples/sec	 accuracy=73.439338	 loss=1.068151	 lr=0.001000
Epoch[046] Batch [0899]/[3759]	Speed: 35.372072 samples/sec	 accuracy=73.454861	 loss=1.068217	 lr=0.001000
Epoch[046] Batch [0949]/[3759]	Speed: 35.112141 samples/sec	 accuracy=73.366776	 loss=1.070348	 lr=0.001000
Epoch[046] Batch [0999]/[3759]	Speed: 35.362515 samples/sec	 accuracy=73.353125	 loss=1.069499	 lr=0.001000
Epoch[046] Batch [1049]/[3759]	Speed: 35.005155 samples/sec	 accuracy=73.349702	 loss=1.070237	 lr=0.001000
Epoch[046] Batch [1099]/[3759]	Speed: 35.512798 samples/sec	 accuracy=73.434659	 loss=1.069025	 lr=0.001000
Epoch[046] Batch [1149]/[3759]	Speed: 35.091999 samples/sec	 accuracy=73.403533	 loss=1.070262	 lr=0.001000
Epoch[046] Batch [1199]/[3759]	Speed: 35.347513 samples/sec	 accuracy=73.430990	 loss=1.069308	 lr=0.001000
Epoch[046] Batch [1249]/[3759]	Speed: 35.142300 samples/sec	 accuracy=73.393750	 loss=1.071168	 lr=0.001000
Epoch[046] Batch [1299]/[3759]	Speed: 35.344182 samples/sec	 accuracy=73.382212	 loss=1.071924	 lr=0.001000
Epoch[046] Batch [1349]/[3759]	Speed: 35.423181 samples/sec	 accuracy=73.355324	 loss=1.074007	 lr=0.001000
Epoch[046] Batch [1399]/[3759]	Speed: 35.240176 samples/sec	 accuracy=73.342634	 loss=1.073625	 lr=0.001000
Epoch[046] Batch [1449]/[3759]	Speed: 35.300476 samples/sec	 accuracy=73.331897	 loss=1.073723	 lr=0.001000
Epoch[046] Batch [1499]/[3759]	Speed: 35.335401 samples/sec	 accuracy=73.343750	 loss=1.072071	 lr=0.001000
Epoch[046] Batch [1549]/[3759]	Speed: 35.426396 samples/sec	 accuracy=73.370968	 loss=1.071435	 lr=0.001000
Epoch[046] Batch [1599]/[3759]	Speed: 35.180885 samples/sec	 accuracy=73.398438	 loss=1.071390	 lr=0.001000
Epoch[046] Batch [1649]/[3759]	Speed: 35.260647 samples/sec	 accuracy=73.369318	 loss=1.072575	 lr=0.001000
Epoch[046] Batch [1699]/[3759]	Speed: 35.326186 samples/sec	 accuracy=73.363971	 loss=1.071952	 lr=0.001000
Epoch[046] Batch [1749]/[3759]	Speed: 35.248665 samples/sec	 accuracy=73.374107	 loss=1.071703	 lr=0.001000
Epoch[046] Batch [1799]/[3759]	Speed: 35.164547 samples/sec	 accuracy=73.365451	 loss=1.071915	 lr=0.001000
Epoch[046] Batch [1849]/[3759]	Speed: 35.189367 samples/sec	 accuracy=73.354730	 loss=1.072339	 lr=0.001000
Epoch[046] Batch [1899]/[3759]	Speed: 35.177096 samples/sec	 accuracy=73.333059	 loss=1.073535	 lr=0.001000
Epoch[046] Batch [1949]/[3759]	Speed: 35.027742 samples/sec	 accuracy=73.343750	 loss=1.072935	 lr=0.001000
Epoch[046] Batch [1999]/[3759]	Speed: 35.412102 samples/sec	 accuracy=73.327344	 loss=1.073876	 lr=0.001000
Epoch[046] Batch [2049]/[3759]	Speed: 34.689916 samples/sec	 accuracy=73.335366	 loss=1.074126	 lr=0.001000
Epoch[046] Batch [2099]/[3759]	Speed: 35.313377 samples/sec	 accuracy=73.307292	 loss=1.074624	 lr=0.001000
Epoch[046] Batch [2149]/[3759]	Speed: 34.466238 samples/sec	 accuracy=73.319041	 loss=1.073651	 lr=0.001000
Epoch[046] Batch [2199]/[3759]	Speed: 35.454622 samples/sec	 accuracy=73.311790	 loss=1.073093	 lr=0.001000
Epoch[046] Batch [2249]/[3759]	Speed: 34.797770 samples/sec	 accuracy=73.292361	 loss=1.073074	 lr=0.001000
Epoch[046] Batch [2299]/[3759]	Speed: 35.081696 samples/sec	 accuracy=73.287364	 loss=1.073040	 lr=0.001000
Epoch[046] Batch [2349]/[3759]	Speed: 34.821976 samples/sec	 accuracy=73.290559	 loss=1.072754	 lr=0.001000
Epoch[046] Batch [2399]/[3759]	Speed: 34.798250 samples/sec	 accuracy=73.297526	 loss=1.072736	 lr=0.001000
Epoch[046] Batch [2449]/[3759]	Speed: 35.086813 samples/sec	 accuracy=73.286352	 loss=1.072730	 lr=0.001000
Epoch[046] Batch [2499]/[3759]	Speed: 34.657045 samples/sec	 accuracy=73.266875	 loss=1.073712	 lr=0.001000
Epoch[046] Batch [2549]/[3759]	Speed: 35.176369 samples/sec	 accuracy=73.238358	 loss=1.074959	 lr=0.001000
Epoch[046] Batch [2599]/[3759]	Speed: 34.343111 samples/sec	 accuracy=73.252404	 loss=1.074913	 lr=0.001000
Epoch[046] Batch [2649]/[3759]	Speed: 35.249201 samples/sec	 accuracy=73.263561	 loss=1.074141	 lr=0.001000
Epoch[046] Batch [2699]/[3759]	Speed: 34.535797 samples/sec	 accuracy=73.262731	 loss=1.073468	 lr=0.001000
Epoch[046] Batch [2749]/[3759]	Speed: 35.201846 samples/sec	 accuracy=73.239205	 loss=1.074580	 lr=0.001000
Epoch[046] Batch [2799]/[3759]	Speed: 34.769360 samples/sec	 accuracy=73.230469	 loss=1.074769	 lr=0.001000
Epoch[046] Batch [2849]/[3759]	Speed: 35.116356 samples/sec	 accuracy=73.234649	 loss=1.075025	 lr=0.001000
Epoch[046] Batch [2899]/[3759]	Speed: 35.146314 samples/sec	 accuracy=73.254849	 loss=1.074646	 lr=0.001000
Epoch[046] Batch [2949]/[3759]	Speed: 35.060067 samples/sec	 accuracy=73.254237	 loss=1.074502	 lr=0.001000
Epoch[046] Batch [2999]/[3759]	Speed: 35.355494 samples/sec	 accuracy=73.276562	 loss=1.074487	 lr=0.001000
Epoch[046] Batch [3049]/[3759]	Speed: 34.965921 samples/sec	 accuracy=73.298156	 loss=1.073942	 lr=0.001000
Epoch[046] Batch [3099]/[3759]	Speed: 35.273680 samples/sec	 accuracy=73.273690	 loss=1.074439	 lr=0.001000
Epoch[046] Batch [3149]/[3759]	Speed: 34.686626 samples/sec	 accuracy=73.285218	 loss=1.074337	 lr=0.001000
Epoch[046] Batch [3199]/[3759]	Speed: 35.403716 samples/sec	 accuracy=73.270996	 loss=1.074953	 lr=0.001000
Epoch[046] Batch [3249]/[3759]	Speed: 34.580648 samples/sec	 accuracy=73.277404	 loss=1.075114	 lr=0.001000
Epoch[046] Batch [3299]/[3759]	Speed: 35.367800 samples/sec	 accuracy=73.304451	 loss=1.074109	 lr=0.001000
Epoch[046] Batch [3349]/[3759]	Speed: 34.592952 samples/sec	 accuracy=73.298507	 loss=1.074415	 lr=0.001000
Epoch[046] Batch [3399]/[3759]	Speed: 35.177747 samples/sec	 accuracy=73.295956	 loss=1.074416	 lr=0.001000
Epoch[046] Batch [3449]/[3759]	Speed: 35.028486 samples/sec	 accuracy=73.314764	 loss=1.073910	 lr=0.001000
Epoch[046] Batch [3499]/[3759]	Speed: 35.104077 samples/sec	 accuracy=73.313839	 loss=1.074041	 lr=0.001000
Epoch[046] Batch [3549]/[3759]	Speed: 35.113892 samples/sec	 accuracy=73.323504	 loss=1.073699	 lr=0.001000
Epoch[046] Batch [3599]/[3759]	Speed: 35.012907 samples/sec	 accuracy=73.322049	 loss=1.073707	 lr=0.001000
Epoch[046] Batch [3649]/[3759]	Speed: 35.330471 samples/sec	 accuracy=73.339897	 loss=1.072850	 lr=0.001000
Epoch[046] Batch [3699]/[3759]	Speed: 34.801665 samples/sec	 accuracy=73.345861	 loss=1.072319	 lr=0.001000
Epoch[046] Batch [3749]/[3759]	Speed: 36.858569 samples/sec	 accuracy=73.333750	 loss=1.072723	 lr=0.001000
Batch [0049]/[0303]: acc-top1=65.906250 acc-top5=86.093750
Batch [0099]/[0303]: acc-top1=65.843750 acc-top5=85.562500
Batch [0149]/[0303]: acc-top1=66.083333 acc-top5=85.541667
Batch [0199]/[0303]: acc-top1=65.906250 acc-top5=85.570312
Batch [0249]/[0303]: acc-top1=65.706250 acc-top5=85.431250
Batch [0299]/[0303]: acc-top1=65.682292 acc-top5=85.458333
[Epoch 046] training: accuracy=73.341065	 loss=1.072382
[Epoch 046] speed: 34 samples/sec	time cost: 7356.062986
[Epoch 046] validation: acc-top1=65.661097 acc-top5=85.463078 loss=1.588929
Epoch[047] Batch [0049]/[3760]	Speed: 25.295466 samples/sec	 accuracy=74.093750	 loss=1.022602	 lr=0.001000
Epoch[047] Batch [0099]/[3760]	Speed: 35.043236 samples/sec	 accuracy=73.781250	 loss=1.043173	 lr=0.001000
Epoch[047] Batch [0149]/[3760]	Speed: 34.855107 samples/sec	 accuracy=73.687500	 loss=1.054882	 lr=0.001000
Epoch[047] Batch [0199]/[3760]	Speed: 34.933253 samples/sec	 accuracy=73.742188	 loss=1.058161	 lr=0.001000
Epoch[047] Batch [0249]/[3760]	Speed: 35.164796 samples/sec	 accuracy=73.468750	 loss=1.068360	 lr=0.001000
Epoch[047] Batch [0299]/[3760]	Speed: 35.165974 samples/sec	 accuracy=73.427083	 loss=1.069998	 lr=0.001000
Epoch[047] Batch [0349]/[3760]	Speed: 35.322950 samples/sec	 accuracy=73.468750	 loss=1.072147	 lr=0.001000
Epoch[047] Batch [0399]/[3760]	Speed: 34.928827 samples/sec	 accuracy=73.496094	 loss=1.074007	 lr=0.001000
Epoch[047] Batch [0449]/[3760]	Speed: 35.423243 samples/sec	 accuracy=73.472222	 loss=1.074722	 lr=0.001000
Epoch[047] Batch [0499]/[3760]	Speed: 34.496625 samples/sec	 accuracy=73.331250	 loss=1.082326	 lr=0.001000
Epoch[047] Batch [0549]/[3760]	Speed: 35.325527 samples/sec	 accuracy=73.335227	 loss=1.081567	 lr=0.001000
Epoch[047] Batch [0599]/[3760]	Speed: 34.738247 samples/sec	 accuracy=73.351562	 loss=1.079727	 lr=0.001000
Epoch[047] Batch [0649]/[3760]	Speed: 35.133823 samples/sec	 accuracy=73.302885	 loss=1.080150	 lr=0.001000
Epoch[047] Batch [0699]/[3760]	Speed: 34.902567 samples/sec	 accuracy=73.229911	 loss=1.080760	 lr=0.001000
Epoch[047] Batch [0749]/[3760]	Speed: 35.179206 samples/sec	 accuracy=73.325000	 loss=1.075266	 lr=0.001000
Epoch[047] Batch [0799]/[3760]	Speed: 35.271730 samples/sec	 accuracy=73.386719	 loss=1.071569	 lr=0.001000
Epoch[047] Batch [0849]/[3760]	Speed: 35.064042 samples/sec	 accuracy=73.465074	 loss=1.067853	 lr=0.001000
Epoch[047] Batch [0899]/[3760]	Speed: 35.240377 samples/sec	 accuracy=73.519097	 loss=1.064436	 lr=0.001000
Epoch[047] Batch [0949]/[3760]	Speed: 35.031762 samples/sec	 accuracy=73.478618	 loss=1.066046	 lr=0.001000
Epoch[047] Batch [0999]/[3760]	Speed: 35.554312 samples/sec	 accuracy=73.503125	 loss=1.064228	 lr=0.001000
Epoch[047] Batch [1049]/[3760]	Speed: 35.189448 samples/sec	 accuracy=73.507440	 loss=1.065242	 lr=0.001000
Epoch[047] Batch [1099]/[3760]	Speed: 35.483883 samples/sec	 accuracy=73.612216	 loss=1.061821	 lr=0.001000
Epoch[047] Batch [1149]/[3760]	Speed: 34.991033 samples/sec	 accuracy=73.682065	 loss=1.060636	 lr=0.001000
Epoch[047] Batch [1199]/[3760]	Speed: 35.549397 samples/sec	 accuracy=73.730469	 loss=1.059473	 lr=0.001000
Epoch[047] Batch [1249]/[3760]	Speed: 34.989002 samples/sec	 accuracy=73.688750	 loss=1.060965	 lr=0.001000
Epoch[047] Batch [1299]/[3760]	Speed: 35.595652 samples/sec	 accuracy=73.698317	 loss=1.060446	 lr=0.001000
Epoch[047] Batch [1349]/[3760]	Speed: 34.982437 samples/sec	 accuracy=73.702546	 loss=1.059797	 lr=0.001000
Epoch[047] Batch [1399]/[3760]	Speed: 35.406384 samples/sec	 accuracy=73.713170	 loss=1.060016	 lr=0.001000
Epoch[047] Batch [1449]/[3760]	Speed: 35.216671 samples/sec	 accuracy=73.673491	 loss=1.061501	 lr=0.001000
Epoch[047] Batch [1499]/[3760]	Speed: 35.497562 samples/sec	 accuracy=73.686458	 loss=1.062159	 lr=0.001000
Epoch[047] Batch [1549]/[3760]	Speed: 34.985677 samples/sec	 accuracy=73.721774	 loss=1.060846	 lr=0.001000
Epoch[047] Batch [1599]/[3760]	Speed: 35.391123 samples/sec	 accuracy=73.690430	 loss=1.061511	 lr=0.001000
Epoch[047] Batch [1649]/[3760]	Speed: 34.945951 samples/sec	 accuracy=73.675189	 loss=1.061941	 lr=0.001000
Epoch[047] Batch [1699]/[3760]	Speed: 35.249105 samples/sec	 accuracy=73.690257	 loss=1.060558	 lr=0.001000
Epoch[047] Batch [1749]/[3760]	Speed: 34.976060 samples/sec	 accuracy=73.702679	 loss=1.060586	 lr=0.001000
Epoch[047] Batch [1799]/[3760]	Speed: 35.242575 samples/sec	 accuracy=73.669271	 loss=1.061437	 lr=0.001000
Epoch[047] Batch [1849]/[3760]	Speed: 34.917100 samples/sec	 accuracy=73.652872	 loss=1.061835	 lr=0.001000
Epoch[047] Batch [1899]/[3760]	Speed: 34.936615 samples/sec	 accuracy=73.640625	 loss=1.062622	 lr=0.001000
Epoch[047] Batch [1949]/[3760]	Speed: 34.941613 samples/sec	 accuracy=73.637019	 loss=1.062631	 lr=0.001000
Epoch[047] Batch [1999]/[3760]	Speed: 34.888309 samples/sec	 accuracy=73.674219	 loss=1.060505	 lr=0.001000
Epoch[047] Batch [2049]/[3760]	Speed: 35.298655 samples/sec	 accuracy=73.628811	 loss=1.061492	 lr=0.001000
Epoch[047] Batch [2099]/[3760]	Speed: 34.631993 samples/sec	 accuracy=73.621280	 loss=1.061679	 lr=0.001000
Epoch[047] Batch [2149]/[3760]	Speed: 35.461961 samples/sec	 accuracy=73.619913	 loss=1.061858	 lr=0.001000
Epoch[047] Batch [2199]/[3760]	Speed: 34.910342 samples/sec	 accuracy=73.612216	 loss=1.062771	 lr=0.001000
Epoch[047] Batch [2249]/[3760]	Speed: 35.495980 samples/sec	 accuracy=73.592361	 loss=1.063201	 lr=0.001000
Epoch[047] Batch [2299]/[3760]	Speed: 35.241826 samples/sec	 accuracy=73.606658	 loss=1.062337	 lr=0.001000
Epoch[047] Batch [2349]/[3760]	Speed: 35.442501 samples/sec	 accuracy=73.590426	 loss=1.062542	 lr=0.001000
Epoch[047] Batch [2399]/[3760]	Speed: 35.076718 samples/sec	 accuracy=73.589193	 loss=1.063378	 lr=0.001000
Epoch[047] Batch [2449]/[3760]	Speed: 35.205092 samples/sec	 accuracy=73.595026	 loss=1.063001	 lr=0.001000
Epoch[047] Batch [2499]/[3760]	Speed: 34.649114 samples/sec	 accuracy=73.608125	 loss=1.062603	 lr=0.001000
Epoch[047] Batch [2549]/[3760]	Speed: 34.816080 samples/sec	 accuracy=73.626838	 loss=1.061727	 lr=0.001000
Epoch[047] Batch [2599]/[3760]	Speed: 35.001713 samples/sec	 accuracy=73.636418	 loss=1.061808	 lr=0.001000
Epoch[047] Batch [2649]/[3760]	Speed: 34.521954 samples/sec	 accuracy=73.650354	 loss=1.061272	 lr=0.001000
Epoch[047] Batch [2699]/[3760]	Speed: 35.143414 samples/sec	 accuracy=73.641204	 loss=1.061011	 lr=0.001000
Epoch[047] Batch [2749]/[3760]	Speed: 34.303735 samples/sec	 accuracy=73.646591	 loss=1.060490	 lr=0.001000
Epoch[047] Batch [2799]/[3760]	Speed: 35.417459 samples/sec	 accuracy=73.650670	 loss=1.060655	 lr=0.001000
Epoch[047] Batch [2849]/[3760]	Speed: 34.243115 samples/sec	 accuracy=73.654605	 loss=1.060475	 lr=0.001000
Epoch[047] Batch [2899]/[3760]	Speed: 34.885711 samples/sec	 accuracy=73.668103	 loss=1.059942	 lr=0.001000
Epoch[047] Batch [2949]/[3760]	Speed: 34.924110 samples/sec	 accuracy=73.664195	 loss=1.060288	 lr=0.001000
Epoch[047] Batch [2999]/[3760]	Speed: 34.861615 samples/sec	 accuracy=73.646354	 loss=1.060736	 lr=0.001000
Epoch[047] Batch [3049]/[3760]	Speed: 35.130127 samples/sec	 accuracy=73.652664	 loss=1.060565	 lr=0.001000
Epoch[047] Batch [3099]/[3760]	Speed: 34.759355 samples/sec	 accuracy=73.661290	 loss=1.060463	 lr=0.001000
Epoch[047] Batch [3149]/[3760]	Speed: 35.326549 samples/sec	 accuracy=73.667163	 loss=1.060721	 lr=0.001000
Epoch[047] Batch [3199]/[3760]	Speed: 34.659747 samples/sec	 accuracy=73.661133	 loss=1.060949	 lr=0.001000
Epoch[047] Batch [3249]/[3760]	Speed: 35.369383 samples/sec	 accuracy=73.657692	 loss=1.061225	 lr=0.001000
Epoch[047] Batch [3299]/[3760]	Speed: 34.696313 samples/sec	 accuracy=73.666193	 loss=1.060727	 lr=0.001000
Epoch[047] Batch [3349]/[3760]	Speed: 35.291297 samples/sec	 accuracy=73.685634	 loss=1.060422	 lr=0.001000
Epoch[047] Batch [3399]/[3760]	Speed: 34.780318 samples/sec	 accuracy=73.674173	 loss=1.060617	 lr=0.001000
Epoch[047] Batch [3449]/[3760]	Speed: 34.921842 samples/sec	 accuracy=73.651268	 loss=1.061299	 lr=0.001000
Epoch[047] Batch [3499]/[3760]	Speed: 35.054815 samples/sec	 accuracy=73.650446	 loss=1.061321	 lr=0.001000
Epoch[047] Batch [3549]/[3760]	Speed: 34.737682 samples/sec	 accuracy=73.663732	 loss=1.060796	 lr=0.001000
Epoch[047] Batch [3599]/[3760]	Speed: 35.216211 samples/sec	 accuracy=73.657118	 loss=1.060857	 lr=0.001000
Epoch[047] Batch [3649]/[3760]	Speed: 34.591220 samples/sec	 accuracy=73.653682	 loss=1.061076	 lr=0.001000
Epoch[047] Batch [3699]/[3760]	Speed: 35.292795 samples/sec	 accuracy=73.686233	 loss=1.059865	 lr=0.001000
Epoch[047] Batch [3749]/[3760]	Speed: 36.088061 samples/sec	 accuracy=73.694583	 loss=1.059886	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.406250 acc-top5=86.062500
Batch [0099]/[0303]: acc-top1=66.062500 acc-top5=85.421875
Batch [0149]/[0303]: acc-top1=66.354167 acc-top5=85.625000
Batch [0199]/[0303]: acc-top1=66.156250 acc-top5=85.648438
Batch [0249]/[0303]: acc-top1=65.912500 acc-top5=85.506250
Batch [0299]/[0303]: acc-top1=65.848958 acc-top5=85.484375
[Epoch 047] training: accuracy=73.689744	 loss=1.060094
[Epoch 047] speed: 34 samples/sec	time cost: 7359.461948
[Epoch 047] validation: acc-top1=65.862211 acc-top5=85.483705 loss=1.599763
Epoch[048] Batch [0049]/[3760]	Speed: 24.005898 samples/sec	 accuracy=72.968750	 loss=1.075915	 lr=0.001000
Epoch[048] Batch [0099]/[3760]	Speed: 34.991151 samples/sec	 accuracy=73.234375	 loss=1.065720	 lr=0.001000
Epoch[048] Batch [0149]/[3760]	Speed: 35.013927 samples/sec	 accuracy=73.562500	 loss=1.055808	 lr=0.001000
Epoch[048] Batch [0199]/[3760]	Speed: 34.841215 samples/sec	 accuracy=73.671875	 loss=1.054284	 lr=0.001000
Epoch[048] Batch [0249]/[3760]	Speed: 35.093844 samples/sec	 accuracy=73.681250	 loss=1.056921	 lr=0.001000
Epoch[048] Batch [0299]/[3760]	Speed: 34.834608 samples/sec	 accuracy=73.875000	 loss=1.050712	 lr=0.001000
Epoch[048] Batch [0349]/[3760]	Speed: 35.370530 samples/sec	 accuracy=73.888393	 loss=1.048293	 lr=0.001000
Epoch[048] Batch [0399]/[3760]	Speed: 34.740131 samples/sec	 accuracy=73.960938	 loss=1.044592	 lr=0.001000
Epoch[048] Batch [0449]/[3760]	Speed: 35.377458 samples/sec	 accuracy=74.076389	 loss=1.040894	 lr=0.001000
Epoch[048] Batch [0499]/[3760]	Speed: 34.671285 samples/sec	 accuracy=74.146875	 loss=1.038427	 lr=0.001000
Epoch[048] Batch [0549]/[3760]	Speed: 35.323424 samples/sec	 accuracy=74.213068	 loss=1.039856	 lr=0.001000
Epoch[048] Batch [0599]/[3760]	Speed: 34.779350 samples/sec	 accuracy=74.260417	 loss=1.038010	 lr=0.001000
Epoch[048] Batch [0649]/[3760]	Speed: 35.214176 samples/sec	 accuracy=74.225962	 loss=1.037485	 lr=0.001000
Epoch[048] Batch [0699]/[3760]	Speed: 34.753895 samples/sec	 accuracy=74.180804	 loss=1.041236	 lr=0.001000
Epoch[048] Batch [0749]/[3760]	Speed: 35.247620 samples/sec	 accuracy=74.264583	 loss=1.036749	 lr=0.001000
Epoch[048] Batch [0799]/[3760]	Speed: 34.990000 samples/sec	 accuracy=74.289062	 loss=1.035929	 lr=0.001000
Epoch[048] Batch [0849]/[3760]	Speed: 34.944758 samples/sec	 accuracy=74.227941	 loss=1.037508	 lr=0.001000
Epoch[048] Batch [0899]/[3760]	Speed: 35.139568 samples/sec	 accuracy=74.230903	 loss=1.038330	 lr=0.001000
Epoch[048] Batch [0949]/[3760]	Speed: 34.838612 samples/sec	 accuracy=74.236842	 loss=1.036139	 lr=0.001000
Epoch[048] Batch [0999]/[3760]	Speed: 35.200071 samples/sec	 accuracy=74.262500	 loss=1.036953	 lr=0.001000
Epoch[048] Batch [1049]/[3760]	Speed: 34.744425 samples/sec	 accuracy=74.254464	 loss=1.037383	 lr=0.001000
Epoch[048] Batch [1099]/[3760]	Speed: 35.241523 samples/sec	 accuracy=74.211648	 loss=1.038531	 lr=0.001000
Epoch[048] Batch [1149]/[3760]	Speed: 34.764946 samples/sec	 accuracy=74.180707	 loss=1.039651	 lr=0.001000
Epoch[048] Batch [1199]/[3760]	Speed: 35.323265 samples/sec	 accuracy=74.183594	 loss=1.039594	 lr=0.001000
Epoch[048] Batch [1249]/[3760]	Speed: 34.914551 samples/sec	 accuracy=74.190000	 loss=1.038809	 lr=0.001000
Epoch[048] Batch [1299]/[3760]	Speed: 35.420123 samples/sec	 accuracy=74.157452	 loss=1.039758	 lr=0.001000
Epoch[048] Batch [1349]/[3760]	Speed: 34.903859 samples/sec	 accuracy=74.135417	 loss=1.040555	 lr=0.001000
Epoch[048] Batch [1399]/[3760]	Speed: 35.199926 samples/sec	 accuracy=74.125000	 loss=1.040577	 lr=0.001000
Epoch[048] Batch [1449]/[3760]	Speed: 35.014786 samples/sec	 accuracy=74.091595	 loss=1.042836	 lr=0.001000
Epoch[048] Batch [1499]/[3760]	Speed: 35.008214 samples/sec	 accuracy=74.048958	 loss=1.044053	 lr=0.001000
Epoch[048] Batch [1549]/[3760]	Speed: 34.913063 samples/sec	 accuracy=74.122984	 loss=1.041746	 lr=0.001000
Epoch[048] Batch [1599]/[3760]	Speed: 35.005864 samples/sec	 accuracy=74.083984	 loss=1.043087	 lr=0.001000
Epoch[048] Batch [1649]/[3760]	Speed: 35.224227 samples/sec	 accuracy=74.071970	 loss=1.043541	 lr=0.001000
Epoch[048] Batch [1699]/[3760]	Speed: 34.860337 samples/sec	 accuracy=74.068934	 loss=1.043136	 lr=0.001000
Epoch[048] Batch [1749]/[3760]	Speed: 35.264250 samples/sec	 accuracy=74.099107	 loss=1.042968	 lr=0.001000
Epoch[048] Batch [1799]/[3760]	Speed: 34.819799 samples/sec	 accuracy=74.105035	 loss=1.042699	 lr=0.001000
Epoch[048] Batch [1849]/[3760]	Speed: 35.247120 samples/sec	 accuracy=74.088682	 loss=1.043199	 lr=0.001000
Epoch[048] Batch [1899]/[3760]	Speed: 34.872911 samples/sec	 accuracy=74.039474	 loss=1.044751	 lr=0.001000
Epoch[048] Batch [1949]/[3760]	Speed: 35.435582 samples/sec	 accuracy=74.039263	 loss=1.045733	 lr=0.001000
Epoch[048] Batch [1999]/[3760]	Speed: 35.017300 samples/sec	 accuracy=74.029687	 loss=1.045665	 lr=0.001000
Epoch[048] Batch [2049]/[3760]	Speed: 35.291921 samples/sec	 accuracy=74.041159	 loss=1.046754	 lr=0.001000
Epoch[048] Batch [2099]/[3760]	Speed: 35.233283 samples/sec	 accuracy=74.023810	 loss=1.047872	 lr=0.001000
Epoch[048] Batch [2149]/[3760]	Speed: 35.395634 samples/sec	 accuracy=74.039244	 loss=1.047448	 lr=0.001000
Epoch[048] Batch [2199]/[3760]	Speed: 35.052281 samples/sec	 accuracy=74.028409	 loss=1.047835	 lr=0.001000
Epoch[048] Batch [2249]/[3760]	Speed: 35.246084 samples/sec	 accuracy=74.047222	 loss=1.047147	 lr=0.001000
Epoch[048] Batch [2299]/[3760]	Speed: 34.972286 samples/sec	 accuracy=74.057745	 loss=1.046713	 lr=0.001000
Epoch[048] Batch [2349]/[3760]	Speed: 35.476551 samples/sec	 accuracy=74.064495	 loss=1.046681	 lr=0.001000
Epoch[048] Batch [2399]/[3760]	Speed: 34.978565 samples/sec	 accuracy=74.049479	 loss=1.047275	 lr=0.001000
Epoch[048] Batch [2449]/[3760]	Speed: 35.051643 samples/sec	 accuracy=74.014668	 loss=1.047869	 lr=0.001000
Epoch[048] Batch [2499]/[3760]	Speed: 35.013523 samples/sec	 accuracy=73.994375	 loss=1.049058	 lr=0.001000
Epoch[048] Batch [2549]/[3760]	Speed: 34.870065 samples/sec	 accuracy=73.979779	 loss=1.049704	 lr=0.001000
Epoch[048] Batch [2599]/[3760]	Speed: 35.294792 samples/sec	 accuracy=73.966346	 loss=1.050096	 lr=0.001000
Epoch[048] Batch [2649]/[3760]	Speed: 34.895641 samples/sec	 accuracy=73.991745	 loss=1.048689	 lr=0.001000
Epoch[048] Batch [2699]/[3760]	Speed: 35.241035 samples/sec	 accuracy=73.996528	 loss=1.048270	 lr=0.001000
Epoch[048] Batch [2749]/[3760]	Speed: 34.869207 samples/sec	 accuracy=74.015341	 loss=1.047641	 lr=0.001000
Epoch[048] Batch [2799]/[3760]	Speed: 35.440843 samples/sec	 accuracy=74.007812	 loss=1.047539	 lr=0.001000
Epoch[048] Batch [2849]/[3760]	Speed: 35.033722 samples/sec	 accuracy=73.978618	 loss=1.048934	 lr=0.001000
Epoch[048] Batch [2899]/[3760]	Speed: 35.446015 samples/sec	 accuracy=73.979526	 loss=1.048605	 lr=0.001000
Epoch[048] Batch [2949]/[3760]	Speed: 34.873417 samples/sec	 accuracy=73.969809	 loss=1.049525	 lr=0.001000
Epoch[048] Batch [2999]/[3760]	Speed: 35.215483 samples/sec	 accuracy=73.976562	 loss=1.049241	 lr=0.001000
Epoch[048] Batch [3049]/[3760]	Speed: 35.030714 samples/sec	 accuracy=73.957992	 loss=1.049723	 lr=0.001000
Epoch[048] Batch [3099]/[3760]	Speed: 35.346698 samples/sec	 accuracy=73.959173	 loss=1.049787	 lr=0.001000
Epoch[048] Batch [3149]/[3760]	Speed: 34.996629 samples/sec	 accuracy=73.949405	 loss=1.050115	 lr=0.001000
Epoch[048] Batch [3199]/[3760]	Speed: 35.248433 samples/sec	 accuracy=73.941895	 loss=1.050718	 lr=0.001000
Epoch[048] Batch [3249]/[3760]	Speed: 35.170357 samples/sec	 accuracy=73.926923	 loss=1.051271	 lr=0.001000
Epoch[048] Batch [3299]/[3760]	Speed: 35.043574 samples/sec	 accuracy=73.933239	 loss=1.051243	 lr=0.001000
Epoch[048] Batch [3349]/[3760]	Speed: 35.490915 samples/sec	 accuracy=73.942631	 loss=1.051076	 lr=0.001000
Epoch[048] Batch [3399]/[3760]	Speed: 35.199027 samples/sec	 accuracy=73.943015	 loss=1.050843	 lr=0.001000
Epoch[048] Batch [3449]/[3760]	Speed: 35.392828 samples/sec	 accuracy=73.932518	 loss=1.051258	 lr=0.001000
Epoch[048] Batch [3499]/[3760]	Speed: 35.288984 samples/sec	 accuracy=73.941518	 loss=1.050825	 lr=0.001000
Epoch[048] Batch [3549]/[3760]	Speed: 35.347957 samples/sec	 accuracy=73.946303	 loss=1.050377	 lr=0.001000
Epoch[048] Batch [3599]/[3760]	Speed: 35.328310 samples/sec	 accuracy=73.975694	 loss=1.049705	 lr=0.001000
Epoch[048] Batch [3649]/[3760]	Speed: 35.424057 samples/sec	 accuracy=73.973887	 loss=1.049801	 lr=0.001000
Epoch[048] Batch [3699]/[3760]	Speed: 35.241351 samples/sec	 accuracy=73.972128	 loss=1.049703	 lr=0.001000
Epoch[048] Batch [3749]/[3760]	Speed: 36.418393 samples/sec	 accuracy=73.959167	 loss=1.049713	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.906250 acc-top5=85.656250
Batch [0099]/[0303]: acc-top1=66.046875 acc-top5=85.203125
Batch [0149]/[0303]: acc-top1=66.187500 acc-top5=85.385417
Batch [0199]/[0303]: acc-top1=65.937500 acc-top5=85.414062
Batch [0249]/[0303]: acc-top1=65.693750 acc-top5=85.300000
Batch [0299]/[0303]: acc-top1=65.682292 acc-top5=85.338542
[Epoch 048] training: accuracy=73.957779	 loss=1.049838
[Epoch 048] speed: 34 samples/sec	time cost: 7356.880799
[Epoch 048] validation: acc-top1=65.650784 acc-top5=85.349629 loss=1.596356
Epoch[049] Batch [0049]/[3759]	Speed: 24.543788 samples/sec	 accuracy=73.968750	 loss=1.022196	 lr=0.001000
Epoch[049] Batch [0099]/[3759]	Speed: 35.039249 samples/sec	 accuracy=73.906250	 loss=1.031547	 lr=0.001000
Epoch[049] Batch [0149]/[3759]	Speed: 34.921740 samples/sec	 accuracy=73.302083	 loss=1.050717	 lr=0.001000
Epoch[049] Batch [0199]/[3759]	Speed: 35.115932 samples/sec	 accuracy=73.539062	 loss=1.044554	 lr=0.001000
Epoch[049] Batch [0249]/[3759]	Speed: 35.191665 samples/sec	 accuracy=73.381250	 loss=1.047532	 lr=0.001000
Epoch[049] Batch [0299]/[3759]	Speed: 34.945524 samples/sec	 accuracy=73.723958	 loss=1.039013	 lr=0.001000
Epoch[049] Batch [0349]/[3759]	Speed: 35.300226 samples/sec	 accuracy=73.812500	 loss=1.038976	 lr=0.001000
Epoch[049] Batch [0399]/[3759]	Speed: 34.767748 samples/sec	 accuracy=73.816406	 loss=1.042896	 lr=0.001000
Epoch[049] Batch [0449]/[3759]	Speed: 35.143320 samples/sec	 accuracy=73.760417	 loss=1.044473	 lr=0.001000
Epoch[049] Batch [0499]/[3759]	Speed: 34.747902 samples/sec	 accuracy=73.978125	 loss=1.037475	 lr=0.001000
Epoch[049] Batch [0549]/[3759]	Speed: 35.556750 samples/sec	 accuracy=73.934659	 loss=1.038470	 lr=0.001000
Epoch[049] Batch [0599]/[3759]	Speed: 35.024940 samples/sec	 accuracy=74.007812	 loss=1.038915	 lr=0.001000
Epoch[049] Batch [0649]/[3759]	Speed: 35.476467 samples/sec	 accuracy=74.062500	 loss=1.038502	 lr=0.001000
Epoch[049] Batch [0699]/[3759]	Speed: 35.044197 samples/sec	 accuracy=74.156250	 loss=1.035947	 lr=0.001000
Epoch[049] Batch [0749]/[3759]	Speed: 35.645951 samples/sec	 accuracy=74.177083	 loss=1.035920	 lr=0.001000
Epoch[049] Batch [0799]/[3759]	Speed: 35.031801 samples/sec	 accuracy=74.152344	 loss=1.036626	 lr=0.001000
Epoch[049] Batch [0849]/[3759]	Speed: 35.488776 samples/sec	 accuracy=74.112132	 loss=1.037676	 lr=0.001000
Epoch[049] Batch [0899]/[3759]	Speed: 35.225602 samples/sec	 accuracy=74.119792	 loss=1.036059	 lr=0.001000
Epoch[049] Batch [0949]/[3759]	Speed: 35.457156 samples/sec	 accuracy=74.152961	 loss=1.034684	 lr=0.001000
Epoch[049] Batch [0999]/[3759]	Speed: 35.274815 samples/sec	 accuracy=74.146875	 loss=1.036335	 lr=0.001000
Epoch[049] Batch [1049]/[3759]	Speed: 35.430207 samples/sec	 accuracy=74.221726	 loss=1.033907	 lr=0.001000
Epoch[049] Batch [1099]/[3759]	Speed: 35.128931 samples/sec	 accuracy=74.208807	 loss=1.034796	 lr=0.001000
Epoch[049] Batch [1149]/[3759]	Speed: 35.429407 samples/sec	 accuracy=74.205163	 loss=1.035111	 lr=0.001000
Epoch[049] Batch [1199]/[3759]	Speed: 35.353932 samples/sec	 accuracy=74.247396	 loss=1.033866	 lr=0.001000
Epoch[049] Batch [1249]/[3759]	Speed: 35.364734 samples/sec	 accuracy=74.253750	 loss=1.034745	 lr=0.001000
Epoch[049] Batch [1299]/[3759]	Speed: 35.209907 samples/sec	 accuracy=74.251202	 loss=1.035077	 lr=0.001000
Epoch[049] Batch [1349]/[3759]	Speed: 35.246170 samples/sec	 accuracy=74.266204	 loss=1.034496	 lr=0.001000
Epoch[049] Batch [1399]/[3759]	Speed: 35.010606 samples/sec	 accuracy=74.305804	 loss=1.032429	 lr=0.001000
Epoch[049] Batch [1449]/[3759]	Speed: 35.097025 samples/sec	 accuracy=74.322198	 loss=1.032054	 lr=0.001000
Epoch[049] Batch [1499]/[3759]	Speed: 35.366158 samples/sec	 accuracy=74.347917	 loss=1.031940	 lr=0.001000
Epoch[049] Batch [1549]/[3759]	Speed: 35.162047 samples/sec	 accuracy=74.355847	 loss=1.032445	 lr=0.001000
Epoch[049] Batch [1599]/[3759]	Speed: 35.471102 samples/sec	 accuracy=74.377930	 loss=1.031876	 lr=0.001000
Epoch[049] Batch [1649]/[3759]	Speed: 35.382314 samples/sec	 accuracy=74.333333	 loss=1.033542	 lr=0.001000
Epoch[049] Batch [1699]/[3759]	Speed: 35.524007 samples/sec	 accuracy=74.338235	 loss=1.033005	 lr=0.001000
Epoch[049] Batch [1749]/[3759]	Speed: 35.356980 samples/sec	 accuracy=74.336607	 loss=1.033657	 lr=0.001000
Epoch[049] Batch [1799]/[3759]	Speed: 35.341039 samples/sec	 accuracy=74.329861	 loss=1.034321	 lr=0.001000
Epoch[049] Batch [1849]/[3759]	Speed: 35.377988 samples/sec	 accuracy=74.316723	 loss=1.035049	 lr=0.001000
Epoch[049] Batch [1899]/[3759]	Speed: 35.382221 samples/sec	 accuracy=74.353618	 loss=1.033460	 lr=0.001000
Epoch[049] Batch [1949]/[3759]	Speed: 35.545909 samples/sec	 accuracy=74.361378	 loss=1.033040	 lr=0.001000
Epoch[049] Batch [1999]/[3759]	Speed: 34.994214 samples/sec	 accuracy=74.342187	 loss=1.033646	 lr=0.001000
Epoch[049] Batch [2049]/[3759]	Speed: 35.541711 samples/sec	 accuracy=74.350610	 loss=1.033948	 lr=0.001000
Epoch[049] Batch [2099]/[3759]	Speed: 34.772116 samples/sec	 accuracy=74.369048	 loss=1.033124	 lr=0.001000
Epoch[049] Batch [2149]/[3759]	Speed: 35.380099 samples/sec	 accuracy=74.352471	 loss=1.033524	 lr=0.001000
Epoch[049] Batch [2199]/[3759]	Speed: 34.320656 samples/sec	 accuracy=74.374290	 loss=1.032992	 lr=0.001000
Epoch[049] Batch [2249]/[3759]	Speed: 35.438115 samples/sec	 accuracy=74.356944	 loss=1.033602	 lr=0.001000
Epoch[049] Batch [2299]/[3759]	Speed: 34.736623 samples/sec	 accuracy=74.370245	 loss=1.033153	 lr=0.001000
Epoch[049] Batch [2349]/[3759]	Speed: 35.133046 samples/sec	 accuracy=74.352394	 loss=1.033800	 lr=0.001000
Epoch[049] Batch [2399]/[3759]	Speed: 35.144535 samples/sec	 accuracy=74.337240	 loss=1.034259	 lr=0.001000
Epoch[049] Batch [2449]/[3759]	Speed: 35.081904 samples/sec	 accuracy=74.359694	 loss=1.033456	 lr=0.001000
Epoch[049] Batch [2499]/[3759]	Speed: 35.229392 samples/sec	 accuracy=74.329375	 loss=1.034671	 lr=0.001000
Epoch[049] Batch [2549]/[3759]	Speed: 34.892267 samples/sec	 accuracy=74.326593	 loss=1.034584	 lr=0.001000
Epoch[049] Batch [2599]/[3759]	Speed: 35.288469 samples/sec	 accuracy=74.331731	 loss=1.034349	 lr=0.001000
Epoch[049] Batch [2649]/[3759]	Speed: 34.719276 samples/sec	 accuracy=74.329599	 loss=1.034359	 lr=0.001000
Epoch[049] Batch [2699]/[3759]	Speed: 35.375565 samples/sec	 accuracy=74.333912	 loss=1.033973	 lr=0.001000
Epoch[049] Batch [2749]/[3759]	Speed: 34.555311 samples/sec	 accuracy=74.349432	 loss=1.033459	 lr=0.001000
Epoch[049] Batch [2799]/[3759]	Speed: 35.138575 samples/sec	 accuracy=74.346540	 loss=1.033463	 lr=0.001000
Epoch[049] Batch [2849]/[3759]	Speed: 34.498029 samples/sec	 accuracy=74.353618	 loss=1.032595	 lr=0.001000
Epoch[049] Batch [2899]/[3759]	Speed: 34.964072 samples/sec	 accuracy=74.360453	 loss=1.032427	 lr=0.001000
Epoch[049] Batch [2949]/[3759]	Speed: 34.916456 samples/sec	 accuracy=74.367585	 loss=1.032388	 lr=0.001000
Epoch[049] Batch [2999]/[3759]	Speed: 34.849017 samples/sec	 accuracy=74.355208	 loss=1.033390	 lr=0.001000
Epoch[049] Batch [3049]/[3759]	Speed: 35.188521 samples/sec	 accuracy=74.349898	 loss=1.033021	 lr=0.001000
Epoch[049] Batch [3099]/[3759]	Speed: 34.350729 samples/sec	 accuracy=74.364919	 loss=1.032891	 lr=0.001000
Epoch[049] Batch [3149]/[3759]	Speed: 35.241455 samples/sec	 accuracy=74.368056	 loss=1.032906	 lr=0.001000
Epoch[049] Batch [3199]/[3759]	Speed: 34.469562 samples/sec	 accuracy=74.372559	 loss=1.033285	 lr=0.001000
Epoch[049] Batch [3249]/[3759]	Speed: 35.192590 samples/sec	 accuracy=74.372596	 loss=1.033243	 lr=0.001000
Epoch[049] Batch [3299]/[3759]	Speed: 34.701448 samples/sec	 accuracy=74.350852	 loss=1.033437	 lr=0.001000
Epoch[049] Batch [3349]/[3759]	Speed: 34.947784 samples/sec	 accuracy=74.365205	 loss=1.032829	 lr=0.001000
Epoch[049] Batch [3399]/[3759]	Speed: 35.054519 samples/sec	 accuracy=74.358915	 loss=1.032888	 lr=0.001000
Epoch[049] Batch [3449]/[3759]	Speed: 34.720140 samples/sec	 accuracy=74.355072	 loss=1.033247	 lr=0.001000
Epoch[049] Batch [3499]/[3759]	Speed: 35.182496 samples/sec	 accuracy=74.350893	 loss=1.033397	 lr=0.001000
Epoch[049] Batch [3549]/[3759]	Speed: 34.449366 samples/sec	 accuracy=74.351673	 loss=1.033170	 lr=0.001000
Epoch[049] Batch [3599]/[3759]	Speed: 35.272263 samples/sec	 accuracy=74.354601	 loss=1.033106	 lr=0.001000
Epoch[049] Batch [3649]/[3759]	Speed: 34.487807 samples/sec	 accuracy=74.366010	 loss=1.033031	 lr=0.001000
Epoch[049] Batch [3699]/[3759]	Speed: 35.140470 samples/sec	 accuracy=74.351351	 loss=1.033146	 lr=0.001000
Epoch[049] Batch [3749]/[3759]	Speed: 36.194274 samples/sec	 accuracy=74.337083	 loss=1.033814	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.937500 acc-top5=86.031250
Batch [0099]/[0303]: acc-top1=66.296875 acc-top5=85.562500
Batch [0149]/[0303]: acc-top1=66.510417 acc-top5=85.666667
Batch [0199]/[0303]: acc-top1=66.289062 acc-top5=85.679688
Batch [0249]/[0303]: acc-top1=66.075000 acc-top5=85.543750
Batch [0299]/[0303]: acc-top1=65.937500 acc-top5=85.572917
[Epoch 049] training: accuracy=74.332020	 loss=1.033812
[Epoch 049] speed: 34 samples/sec	time cost: 7350.927100
[Epoch 049] validation: acc-top1=65.934406 acc-top5=85.571370 loss=1.598309
Epoch[050] Batch [0049]/[3760]	Speed: 24.666834 samples/sec	 accuracy=74.156250	 loss=1.017564	 lr=0.001000
Epoch[050] Batch [0099]/[3760]	Speed: 35.071355 samples/sec	 accuracy=74.500000	 loss=1.005829	 lr=0.001000
Epoch[050] Batch [0149]/[3760]	Speed: 35.128323 samples/sec	 accuracy=74.416667	 loss=1.019794	 lr=0.001000
Epoch[050] Batch [0199]/[3760]	Speed: 35.312722 samples/sec	 accuracy=74.500000	 loss=1.018769	 lr=0.001000
Epoch[050] Batch [0249]/[3760]	Speed: 35.133223 samples/sec	 accuracy=74.462500	 loss=1.023173	 lr=0.001000
Epoch[050] Batch [0299]/[3760]	Speed: 34.859244 samples/sec	 accuracy=74.447917	 loss=1.024740	 lr=0.001000
Epoch[050] Batch [0349]/[3760]	Speed: 35.402694 samples/sec	 accuracy=74.272321	 loss=1.033886	 lr=0.001000
Epoch[050] Batch [0399]/[3760]	Speed: 34.770194 samples/sec	 accuracy=74.296875	 loss=1.032445	 lr=0.001000
Epoch[050] Batch [0449]/[3760]	Speed: 35.196549 samples/sec	 accuracy=74.340278	 loss=1.030169	 lr=0.001000
Epoch[050] Batch [0499]/[3760]	Speed: 34.544704 samples/sec	 accuracy=74.462500	 loss=1.025220	 lr=0.001000
Epoch[050] Batch [0549]/[3760]	Speed: 35.444962 samples/sec	 accuracy=74.369318	 loss=1.026199	 lr=0.001000
Epoch[050] Batch [0599]/[3760]	Speed: 34.698578 samples/sec	 accuracy=74.395833	 loss=1.027653	 lr=0.001000
Epoch[050] Batch [0649]/[3760]	Speed: 35.424624 samples/sec	 accuracy=74.490385	 loss=1.027059	 lr=0.001000
Epoch[050] Batch [0699]/[3760]	Speed: 34.999560 samples/sec	 accuracy=74.555804	 loss=1.024484	 lr=0.001000
Epoch[050] Batch [0749]/[3760]	Speed: 35.475569 samples/sec	 accuracy=74.566667	 loss=1.025391	 lr=0.001000
Epoch[050] Batch [0799]/[3760]	Speed: 35.077057 samples/sec	 accuracy=74.585938	 loss=1.025249	 lr=0.001000
Epoch[050] Batch [0849]/[3760]	Speed: 35.194003 samples/sec	 accuracy=74.566176	 loss=1.024404	 lr=0.001000
Epoch[050] Batch [0899]/[3760]	Speed: 35.305422 samples/sec	 accuracy=74.644097	 loss=1.022718	 lr=0.001000
Epoch[050] Batch [0949]/[3760]	Speed: 35.282726 samples/sec	 accuracy=74.638158	 loss=1.023466	 lr=0.001000
Epoch[050] Batch [0999]/[3760]	Speed: 35.414634 samples/sec	 accuracy=74.568750	 loss=1.026063	 lr=0.001000
Epoch[050] Batch [1049]/[3760]	Speed: 35.244775 samples/sec	 accuracy=74.552083	 loss=1.025136	 lr=0.001000
Epoch[050] Batch [1099]/[3760]	Speed: 35.326399 samples/sec	 accuracy=74.552557	 loss=1.024492	 lr=0.001000
Epoch[050] Batch [1149]/[3760]	Speed: 35.336979 samples/sec	 accuracy=74.623641	 loss=1.022196	 lr=0.001000
Epoch[050] Batch [1199]/[3760]	Speed: 35.264316 samples/sec	 accuracy=74.606771	 loss=1.023507	 lr=0.001000
Epoch[050] Batch [1249]/[3760]	Speed: 35.139225 samples/sec	 accuracy=74.597500	 loss=1.024563	 lr=0.001000
Epoch[050] Batch [1299]/[3760]	Speed: 35.539894 samples/sec	 accuracy=74.608173	 loss=1.024584	 lr=0.001000
Epoch[050] Batch [1349]/[3760]	Speed: 34.981140 samples/sec	 accuracy=74.587963	 loss=1.025631	 lr=0.001000
Epoch[050] Batch [1399]/[3760]	Speed: 35.488765 samples/sec	 accuracy=74.625000	 loss=1.024611	 lr=0.001000
Epoch[050] Batch [1449]/[3760]	Speed: 35.078216 samples/sec	 accuracy=74.580819	 loss=1.025346	 lr=0.001000
Epoch[050] Batch [1499]/[3760]	Speed: 35.567316 samples/sec	 accuracy=74.571875	 loss=1.024888	 lr=0.001000
Epoch[050] Batch [1549]/[3760]	Speed: 34.908495 samples/sec	 accuracy=74.603831	 loss=1.023174	 lr=0.001000
Epoch[050] Batch [1599]/[3760]	Speed: 35.589772 samples/sec	 accuracy=74.588867	 loss=1.024067	 lr=0.001000
Epoch[050] Batch [1649]/[3760]	Speed: 35.094844 samples/sec	 accuracy=74.647727	 loss=1.022455	 lr=0.001000
Epoch[050] Batch [1699]/[3760]	Speed: 35.601250 samples/sec	 accuracy=74.645221	 loss=1.022041	 lr=0.001000
Epoch[050] Batch [1749]/[3760]	Speed: 35.278384 samples/sec	 accuracy=74.651786	 loss=1.021625	 lr=0.001000
Epoch[050] Batch [1799]/[3760]	Speed: 35.467791 samples/sec	 accuracy=74.650174	 loss=1.021589	 lr=0.001000
Epoch[050] Batch [1849]/[3760]	Speed: 35.160869 samples/sec	 accuracy=74.646115	 loss=1.021447	 lr=0.001000
Epoch[050] Batch [1899]/[3760]	Speed: 35.314807 samples/sec	 accuracy=74.663651	 loss=1.019506	 lr=0.001000
Epoch[050] Batch [1949]/[3760]	Speed: 35.186277 samples/sec	 accuracy=74.695513	 loss=1.018464	 lr=0.001000
Epoch[050] Batch [1999]/[3760]	Speed: 35.394958 samples/sec	 accuracy=74.700781	 loss=1.018733	 lr=0.001000
Epoch[050] Batch [2049]/[3760]	Speed: 34.833329 samples/sec	 accuracy=74.732470	 loss=1.017292	 lr=0.001000
Epoch[050] Batch [2099]/[3760]	Speed: 35.055895 samples/sec	 accuracy=74.720238	 loss=1.017622	 lr=0.001000
Epoch[050] Batch [2149]/[3760]	Speed: 34.955130 samples/sec	 accuracy=74.760901	 loss=1.016227	 lr=0.001000
Epoch[050] Batch [2199]/[3760]	Speed: 34.997846 samples/sec	 accuracy=74.730114	 loss=1.017924	 lr=0.001000
Epoch[050] Batch [2249]/[3760]	Speed: 35.350966 samples/sec	 accuracy=74.727778	 loss=1.018801	 lr=0.001000
Epoch[050] Batch [2299]/[3760]	Speed: 34.676997 samples/sec	 accuracy=74.735734	 loss=1.018496	 lr=0.001000
Epoch[050] Batch [2349]/[3760]	Speed: 35.304716 samples/sec	 accuracy=74.721410	 loss=1.019600	 lr=0.001000
Epoch[050] Batch [2399]/[3760]	Speed: 34.598031 samples/sec	 accuracy=74.694661	 loss=1.020135	 lr=0.001000
Epoch[050] Batch [2449]/[3760]	Speed: 35.458638 samples/sec	 accuracy=74.697066	 loss=1.020179	 lr=0.001000
Epoch[050] Batch [2499]/[3760]	Speed: 34.873992 samples/sec	 accuracy=74.686250	 loss=1.020799	 lr=0.001000
Epoch[050] Batch [2549]/[3760]	Speed: 35.423355 samples/sec	 accuracy=74.683824	 loss=1.021045	 lr=0.001000
Epoch[050] Batch [2599]/[3760]	Speed: 34.888220 samples/sec	 accuracy=74.673077	 loss=1.021185	 lr=0.001000
Epoch[050] Batch [2649]/[3760]	Speed: 35.040084 samples/sec	 accuracy=74.670401	 loss=1.021322	 lr=0.001000
Epoch[050] Batch [2699]/[3760]	Speed: 35.037421 samples/sec	 accuracy=74.662037	 loss=1.020991	 lr=0.001000
Epoch[050] Batch [2749]/[3760]	Speed: 34.948870 samples/sec	 accuracy=74.655114	 loss=1.021328	 lr=0.001000
Epoch[050] Batch [2799]/[3760]	Speed: 35.388009 samples/sec	 accuracy=74.642857	 loss=1.020733	 lr=0.001000
Epoch[050] Batch [2849]/[3760]	Speed: 34.610207 samples/sec	 accuracy=74.632675	 loss=1.020845	 lr=0.001000
Epoch[050] Batch [2899]/[3760]	Speed: 35.429482 samples/sec	 accuracy=74.643858	 loss=1.020654	 lr=0.001000
Epoch[050] Batch [2949]/[3760]	Speed: 34.829158 samples/sec	 accuracy=74.629767	 loss=1.021379	 lr=0.001000
Epoch[050] Batch [2999]/[3760]	Speed: 35.419097 samples/sec	 accuracy=74.619792	 loss=1.021420	 lr=0.001000
Epoch[050] Batch [3049]/[3760]	Speed: 34.949335 samples/sec	 accuracy=74.634734	 loss=1.020969	 lr=0.001000
Epoch[050] Batch [3099]/[3760]	Speed: 35.148045 samples/sec	 accuracy=74.643649	 loss=1.020431	 lr=0.001000
Epoch[050] Batch [3149]/[3760]	Speed: 35.000033 samples/sec	 accuracy=74.608631	 loss=1.022083	 lr=0.001000
Epoch[050] Batch [3199]/[3760]	Speed: 34.845004 samples/sec	 accuracy=74.614258	 loss=1.021822	 lr=0.001000
Epoch[050] Batch [3249]/[3760]	Speed: 35.249907 samples/sec	 accuracy=74.603365	 loss=1.021916	 lr=0.001000
Epoch[050] Batch [3299]/[3760]	Speed: 34.950146 samples/sec	 accuracy=74.591383	 loss=1.022776	 lr=0.001000
Epoch[050] Batch [3349]/[3760]	Speed: 35.410855 samples/sec	 accuracy=74.562966	 loss=1.023359	 lr=0.001000
Epoch[050] Batch [3399]/[3760]	Speed: 35.003085 samples/sec	 accuracy=74.572610	 loss=1.022798	 lr=0.001000
Epoch[050] Batch [3449]/[3760]	Speed: 35.656991 samples/sec	 accuracy=74.575181	 loss=1.022865	 lr=0.001000
Epoch[050] Batch [3499]/[3760]	Speed: 35.164030 samples/sec	 accuracy=74.537500	 loss=1.023706	 lr=0.001000
Epoch[050] Batch [3549]/[3760]	Speed: 35.486493 samples/sec	 accuracy=74.551056	 loss=1.023391	 lr=0.001000
Epoch[050] Batch [3599]/[3760]	Speed: 35.234647 samples/sec	 accuracy=74.553385	 loss=1.023595	 lr=0.001000
Epoch[050] Batch [3649]/[3760]	Speed: 35.675347 samples/sec	 accuracy=74.559932	 loss=1.023326	 lr=0.001000
Epoch[050] Batch [3699]/[3760]	Speed: 35.202285 samples/sec	 accuracy=74.568834	 loss=1.023179	 lr=0.001000
Epoch[050] Batch [3749]/[3760]	Speed: 36.685450 samples/sec	 accuracy=74.558750	 loss=1.023286	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.562500 acc-top5=85.968750
Batch [0099]/[0303]: acc-top1=66.281250 acc-top5=85.468750
Batch [0149]/[0303]: acc-top1=66.427083 acc-top5=85.645833
Batch [0199]/[0303]: acc-top1=66.195312 acc-top5=85.687500
Batch [0249]/[0303]: acc-top1=65.912500 acc-top5=85.525000
Batch [0299]/[0303]: acc-top1=65.859375 acc-top5=85.546875
[Epoch 050] training: accuracy=74.564495	 loss=1.023063
[Epoch 050] speed: 34 samples/sec	time cost: 7339.621799
[Epoch 050] validation: acc-top1=65.826114 acc-top5=85.540429 loss=1.584776
Epoch[051] Batch [0049]/[3760]	Speed: 24.670803 samples/sec	 accuracy=74.968750	 loss=1.024763	 lr=0.001000
Epoch[051] Batch [0099]/[3760]	Speed: 35.037756 samples/sec	 accuracy=75.093750	 loss=1.012109	 lr=0.001000
Epoch[051] Batch [0149]/[3760]	Speed: 35.009794 samples/sec	 accuracy=74.697917	 loss=1.020683	 lr=0.001000
Epoch[051] Batch [0199]/[3760]	Speed: 35.175598 samples/sec	 accuracy=74.625000	 loss=1.017187	 lr=0.001000
Epoch[051] Batch [0249]/[3760]	Speed: 35.183256 samples/sec	 accuracy=74.975000	 loss=1.008779	 lr=0.001000
Epoch[051] Batch [0299]/[3760]	Speed: 35.027170 samples/sec	 accuracy=75.000000	 loss=1.012049	 lr=0.001000
Epoch[051] Batch [0349]/[3760]	Speed: 35.432498 samples/sec	 accuracy=74.808036	 loss=1.018059	 lr=0.001000
Epoch[051] Batch [0399]/[3760]	Speed: 35.104168 samples/sec	 accuracy=74.886719	 loss=1.011987	 lr=0.001000
Epoch[051] Batch [0449]/[3760]	Speed: 35.521365 samples/sec	 accuracy=74.958333	 loss=1.010320	 lr=0.001000
Epoch[051] Batch [0499]/[3760]	Speed: 35.153704 samples/sec	 accuracy=75.021875	 loss=1.005623	 lr=0.001000
Epoch[051] Batch [0549]/[3760]	Speed: 35.386467 samples/sec	 accuracy=75.017045	 loss=1.003950	 lr=0.001000
Epoch[051] Batch [0599]/[3760]	Speed: 35.241104 samples/sec	 accuracy=74.989583	 loss=1.005467	 lr=0.001000
Epoch[051] Batch [0649]/[3760]	Speed: 35.615975 samples/sec	 accuracy=74.963942	 loss=1.005887	 lr=0.001000
Epoch[051] Batch [0699]/[3760]	Speed: 35.330156 samples/sec	 accuracy=74.955357	 loss=1.007579	 lr=0.001000
Epoch[051] Batch [0749]/[3760]	Speed: 35.336900 samples/sec	 accuracy=74.931250	 loss=1.007194	 lr=0.001000
Epoch[051] Batch [0799]/[3760]	Speed: 35.338310 samples/sec	 accuracy=74.953125	 loss=1.006291	 lr=0.001000
Epoch[051] Batch [0849]/[3760]	Speed: 35.424787 samples/sec	 accuracy=74.966912	 loss=1.006631	 lr=0.001000
Epoch[051] Batch [0899]/[3760]	Speed: 35.236828 samples/sec	 accuracy=74.876736	 loss=1.009555	 lr=0.001000
Epoch[051] Batch [0949]/[3760]	Speed: 35.430705 samples/sec	 accuracy=74.837171	 loss=1.010684	 lr=0.001000
Epoch[051] Batch [0999]/[3760]	Speed: 35.265259 samples/sec	 accuracy=74.850000	 loss=1.009137	 lr=0.001000
Epoch[051] Batch [1049]/[3760]	Speed: 35.296706 samples/sec	 accuracy=74.782738	 loss=1.010641	 lr=0.001000
Epoch[051] Batch [1099]/[3760]	Speed: 35.469579 samples/sec	 accuracy=74.792614	 loss=1.010331	 lr=0.001000
Epoch[051] Batch [1149]/[3760]	Speed: 35.300529 samples/sec	 accuracy=74.894022	 loss=1.007461	 lr=0.001000
Epoch[051] Batch [1199]/[3760]	Speed: 35.505369 samples/sec	 accuracy=74.864583	 loss=1.009041	 lr=0.001000
Epoch[051] Batch [1249]/[3760]	Speed: 35.282486 samples/sec	 accuracy=74.891250	 loss=1.007950	 lr=0.001000
Epoch[051] Batch [1299]/[3760]	Speed: 35.390048 samples/sec	 accuracy=74.875000	 loss=1.008463	 lr=0.001000
Epoch[051] Batch [1349]/[3760]	Speed: 35.272128 samples/sec	 accuracy=74.910880	 loss=1.007359	 lr=0.001000
Epoch[051] Batch [1399]/[3760]	Speed: 35.376965 samples/sec	 accuracy=74.962054	 loss=1.005538	 lr=0.001000
Epoch[051] Batch [1449]/[3760]	Speed: 35.091752 samples/sec	 accuracy=74.988147	 loss=1.004765	 lr=0.001000
Epoch[051] Batch [1499]/[3760]	Speed: 35.372468 samples/sec	 accuracy=74.961458	 loss=1.004971	 lr=0.001000
Epoch[051] Batch [1549]/[3760]	Speed: 35.268716 samples/sec	 accuracy=74.975806	 loss=1.005638	 lr=0.001000
Epoch[051] Batch [1599]/[3760]	Speed: 35.277626 samples/sec	 accuracy=74.992188	 loss=1.005846	 lr=0.001000
Epoch[051] Batch [1649]/[3760]	Speed: 35.306861 samples/sec	 accuracy=75.000000	 loss=1.005551	 lr=0.001000
Epoch[051] Batch [1699]/[3760]	Speed: 35.209664 samples/sec	 accuracy=74.981618	 loss=1.006397	 lr=0.001000
Epoch[051] Batch [1749]/[3760]	Speed: 35.238661 samples/sec	 accuracy=74.985714	 loss=1.006052	 lr=0.001000
Epoch[051] Batch [1799]/[3760]	Speed: 35.180776 samples/sec	 accuracy=75.004340	 loss=1.006032	 lr=0.001000
Epoch[051] Batch [1849]/[3760]	Speed: 35.273541 samples/sec	 accuracy=75.005068	 loss=1.007025	 lr=0.001000
Epoch[051] Batch [1899]/[3760]	Speed: 35.040286 samples/sec	 accuracy=74.972862	 loss=1.007524	 lr=0.001000
Epoch[051] Batch [1949]/[3760]	Speed: 35.294105 samples/sec	 accuracy=74.992788	 loss=1.006556	 lr=0.001000
Epoch[051] Batch [1999]/[3760]	Speed: 34.896485 samples/sec	 accuracy=74.979687	 loss=1.006820	 lr=0.001000
Epoch[051] Batch [2049]/[3760]	Speed: 35.340808 samples/sec	 accuracy=74.963415	 loss=1.007428	 lr=0.001000
Epoch[051] Batch [2099]/[3760]	Speed: 35.024549 samples/sec	 accuracy=74.957589	 loss=1.007032	 lr=0.001000
Epoch[051] Batch [2149]/[3760]	Speed: 35.211086 samples/sec	 accuracy=74.969477	 loss=1.007213	 lr=0.001000
Epoch[051] Batch [2199]/[3760]	Speed: 34.999665 samples/sec	 accuracy=74.953835	 loss=1.007569	 lr=0.001000
Epoch[051] Batch [2249]/[3760]	Speed: 35.222708 samples/sec	 accuracy=74.957639	 loss=1.007925	 lr=0.001000
Epoch[051] Batch [2299]/[3760]	Speed: 35.063508 samples/sec	 accuracy=74.930027	 loss=1.008473	 lr=0.001000
Epoch[051] Batch [2349]/[3760]	Speed: 34.951147 samples/sec	 accuracy=74.928191	 loss=1.009095	 lr=0.001000
Epoch[051] Batch [2399]/[3760]	Speed: 35.097655 samples/sec	 accuracy=74.907552	 loss=1.009905	 lr=0.001000
Epoch[051] Batch [2449]/[3760]	Speed: 35.018717 samples/sec	 accuracy=74.912628	 loss=1.009994	 lr=0.001000
Epoch[051] Batch [2499]/[3760]	Speed: 35.164965 samples/sec	 accuracy=74.900625	 loss=1.010566	 lr=0.001000
Epoch[051] Batch [2549]/[3760]	Speed: 35.235213 samples/sec	 accuracy=74.917279	 loss=1.010281	 lr=0.001000
Epoch[051] Batch [2599]/[3760]	Speed: 35.140481 samples/sec	 accuracy=74.920673	 loss=1.010573	 lr=0.001000
Epoch[051] Batch [2649]/[3760]	Speed: 34.954843 samples/sec	 accuracy=74.943986	 loss=1.009813	 lr=0.001000
Epoch[051] Batch [2699]/[3760]	Speed: 35.065905 samples/sec	 accuracy=74.953704	 loss=1.009643	 lr=0.001000
Epoch[051] Batch [2749]/[3760]	Speed: 34.653703 samples/sec	 accuracy=74.913636	 loss=1.010997	 lr=0.001000
Epoch[051] Batch [2799]/[3760]	Speed: 35.218065 samples/sec	 accuracy=74.906808	 loss=1.011162	 lr=0.001000
Epoch[051] Batch [2849]/[3760]	Speed: 34.680717 samples/sec	 accuracy=74.903509	 loss=1.010563	 lr=0.001000
Epoch[051] Batch [2899]/[3760]	Speed: 35.075441 samples/sec	 accuracy=74.908405	 loss=1.010642	 lr=0.001000
Epoch[051] Batch [2949]/[3760]	Speed: 34.915410 samples/sec	 accuracy=74.905191	 loss=1.010379	 lr=0.001000
Epoch[051] Batch [2999]/[3760]	Speed: 35.051247 samples/sec	 accuracy=74.915625	 loss=1.010512	 lr=0.001000
Epoch[051] Batch [3049]/[3760]	Speed: 34.963359 samples/sec	 accuracy=74.910348	 loss=1.010404	 lr=0.001000
Epoch[051] Batch [3099]/[3760]	Speed: 34.757087 samples/sec	 accuracy=74.881552	 loss=1.010853	 lr=0.001000
Epoch[051] Batch [3149]/[3760]	Speed: 35.252912 samples/sec	 accuracy=74.882440	 loss=1.011234	 lr=0.001000
Epoch[051] Batch [3199]/[3760]	Speed: 34.621088 samples/sec	 accuracy=74.890137	 loss=1.011141	 lr=0.001000
Epoch[051] Batch [3249]/[3760]	Speed: 35.398286 samples/sec	 accuracy=74.894231	 loss=1.010868	 lr=0.001000
Epoch[051] Batch [3299]/[3760]	Speed: 34.696327 samples/sec	 accuracy=74.892992	 loss=1.011155	 lr=0.001000
Epoch[051] Batch [3349]/[3760]	Speed: 35.196215 samples/sec	 accuracy=74.901586	 loss=1.010813	 lr=0.001000
Epoch[051] Batch [3399]/[3760]	Speed: 34.553555 samples/sec	 accuracy=74.889246	 loss=1.011280	 lr=0.001000
Epoch[051] Batch [3449]/[3760]	Speed: 35.147823 samples/sec	 accuracy=74.882699	 loss=1.011416	 lr=0.001000
Epoch[051] Batch [3499]/[3760]	Speed: 34.945365 samples/sec	 accuracy=74.863839	 loss=1.011660	 lr=0.001000
Epoch[051] Batch [3549]/[3760]	Speed: 35.227926 samples/sec	 accuracy=74.852993	 loss=1.011917	 lr=0.001000
Epoch[051] Batch [3599]/[3760]	Speed: 34.947119 samples/sec	 accuracy=74.842014	 loss=1.012154	 lr=0.001000
Epoch[051] Batch [3649]/[3760]	Speed: 35.125508 samples/sec	 accuracy=74.833904	 loss=1.012692	 lr=0.001000
Epoch[051] Batch [3699]/[3760]	Speed: 35.051775 samples/sec	 accuracy=74.823902	 loss=1.012810	 lr=0.001000
Epoch[051] Batch [3749]/[3760]	Speed: 36.194553 samples/sec	 accuracy=74.841667	 loss=1.012008	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.250000 acc-top5=86.437500
Batch [0099]/[0303]: acc-top1=65.968750 acc-top5=85.765625
Batch [0149]/[0303]: acc-top1=66.416667 acc-top5=86.083333
Batch [0199]/[0303]: acc-top1=66.156250 acc-top5=86.015625
Batch [0249]/[0303]: acc-top1=65.937500 acc-top5=85.825000
Batch [0299]/[0303]: acc-top1=65.838542 acc-top5=85.744792
[Epoch 051] training: accuracy=74.841257	 loss=1.011952
[Epoch 051] speed: 34 samples/sec	time cost: 7343.452396
[Epoch 051] validation: acc-top1=65.831271 acc-top5=85.762170 loss=1.590731
Epoch[052] Batch [0049]/[3759]	Speed: 24.488293 samples/sec	 accuracy=75.718750	 loss=0.980160	 lr=0.001000
Epoch[052] Batch [0099]/[3759]	Speed: 35.062011 samples/sec	 accuracy=75.031250	 loss=0.995227	 lr=0.001000
Epoch[052] Batch [0149]/[3759]	Speed: 34.901904 samples/sec	 accuracy=74.718750	 loss=1.004093	 lr=0.001000
Epoch[052] Batch [0199]/[3759]	Speed: 35.047865 samples/sec	 accuracy=74.921875	 loss=1.003895	 lr=0.001000
Epoch[052] Batch [0249]/[3759]	Speed: 35.166150 samples/sec	 accuracy=74.762500	 loss=1.005874	 lr=0.001000
Epoch[052] Batch [0299]/[3759]	Speed: 34.858034 samples/sec	 accuracy=74.760417	 loss=1.003700	 lr=0.001000
Epoch[052] Batch [0349]/[3759]	Speed: 35.367619 samples/sec	 accuracy=74.803571	 loss=1.002549	 lr=0.001000
Epoch[052] Batch [0399]/[3759]	Speed: 34.990169 samples/sec	 accuracy=74.855469	 loss=1.001174	 lr=0.001000
Epoch[052] Batch [0449]/[3759]	Speed: 35.414929 samples/sec	 accuracy=74.934028	 loss=0.997321	 lr=0.001000
Epoch[052] Batch [0499]/[3759]	Speed: 35.005041 samples/sec	 accuracy=74.884375	 loss=1.003492	 lr=0.001000
Epoch[052] Batch [0549]/[3759]	Speed: 35.363452 samples/sec	 accuracy=74.875000	 loss=1.001776	 lr=0.001000
Epoch[052] Batch [0599]/[3759]	Speed: 34.876802 samples/sec	 accuracy=74.981771	 loss=0.999516	 lr=0.001000
Epoch[052] Batch [0649]/[3759]	Speed: 35.368994 samples/sec	 accuracy=75.048077	 loss=0.997855	 lr=0.001000
Epoch[052] Batch [0699]/[3759]	Speed: 35.035663 samples/sec	 accuracy=74.957589	 loss=1.000398	 lr=0.001000
Epoch[052] Batch [0749]/[3759]	Speed: 35.450195 samples/sec	 accuracy=75.008333	 loss=1.001406	 lr=0.001000
Epoch[052] Batch [0799]/[3759]	Speed: 35.274462 samples/sec	 accuracy=74.988281	 loss=0.999945	 lr=0.001000
Epoch[052] Batch [0849]/[3759]	Speed: 35.384442 samples/sec	 accuracy=75.029412	 loss=0.997974	 lr=0.001000
Epoch[052] Batch [0899]/[3759]	Speed: 35.224044 samples/sec	 accuracy=75.055556	 loss=0.997959	 lr=0.001000
Epoch[052] Batch [0949]/[3759]	Speed: 35.317104 samples/sec	 accuracy=75.092105	 loss=0.997952	 lr=0.001000
Epoch[052] Batch [0999]/[3759]	Speed: 35.191305 samples/sec	 accuracy=75.150000	 loss=0.995130	 lr=0.001000
Epoch[052] Batch [1049]/[3759]	Speed: 35.267709 samples/sec	 accuracy=75.065476	 loss=0.997175	 lr=0.001000
Epoch[052] Batch [1099]/[3759]	Speed: 35.157218 samples/sec	 accuracy=75.127841	 loss=0.996288	 lr=0.001000
Epoch[052] Batch [1149]/[3759]	Speed: 35.356379 samples/sec	 accuracy=75.108696	 loss=0.997038	 lr=0.001000
Epoch[052] Batch [1199]/[3759]	Speed: 34.899514 samples/sec	 accuracy=75.095052	 loss=0.997296	 lr=0.001000
Epoch[052] Batch [1249]/[3759]	Speed: 35.307054 samples/sec	 accuracy=75.097500	 loss=0.998534	 lr=0.001000
Epoch[052] Batch [1299]/[3759]	Speed: 35.237577 samples/sec	 accuracy=75.112981	 loss=0.997680	 lr=0.001000
Epoch[052] Batch [1349]/[3759]	Speed: 34.913960 samples/sec	 accuracy=75.153935	 loss=0.996723	 lr=0.001000
Epoch[052] Batch [1399]/[3759]	Speed: 35.141321 samples/sec	 accuracy=75.138393	 loss=0.996335	 lr=0.001000
Epoch[052] Batch [1449]/[3759]	Speed: 34.679033 samples/sec	 accuracy=75.129310	 loss=0.995930	 lr=0.001000
Epoch[052] Batch [1499]/[3759]	Speed: 35.363633 samples/sec	 accuracy=75.139583	 loss=0.996424	 lr=0.001000
Epoch[052] Batch [1549]/[3759]	Speed: 34.788967 samples/sec	 accuracy=75.165323	 loss=0.994258	 lr=0.001000
Epoch[052] Batch [1599]/[3759]	Speed: 35.319462 samples/sec	 accuracy=75.142578	 loss=0.995097	 lr=0.001000
Epoch[052] Batch [1649]/[3759]	Speed: 34.957799 samples/sec	 accuracy=75.171402	 loss=0.994870	 lr=0.001000
Epoch[052] Batch [1699]/[3759]	Speed: 35.370548 samples/sec	 accuracy=75.176471	 loss=0.994905	 lr=0.001000
Epoch[052] Batch [1749]/[3759]	Speed: 34.918136 samples/sec	 accuracy=75.192857	 loss=0.994946	 lr=0.001000
Epoch[052] Batch [1799]/[3759]	Speed: 35.316669 samples/sec	 accuracy=75.195312	 loss=0.994719	 lr=0.001000
Epoch[052] Batch [1849]/[3759]	Speed: 34.771353 samples/sec	 accuracy=75.193412	 loss=0.994352	 lr=0.001000
Epoch[052] Batch [1899]/[3759]	Speed: 34.880868 samples/sec	 accuracy=75.201480	 loss=0.995247	 lr=0.001000
Epoch[052] Batch [1949]/[3759]	Speed: 34.970758 samples/sec	 accuracy=75.226763	 loss=0.994935	 lr=0.001000
Epoch[052] Batch [1999]/[3759]	Speed: 34.573749 samples/sec	 accuracy=75.219531	 loss=0.994621	 lr=0.001000
Epoch[052] Batch [2049]/[3759]	Speed: 35.032279 samples/sec	 accuracy=75.204268	 loss=0.995415	 lr=0.001000
Epoch[052] Batch [2099]/[3759]	Speed: 34.677035 samples/sec	 accuracy=75.198661	 loss=0.996125	 lr=0.001000
Epoch[052] Batch [2149]/[3759]	Speed: 35.466375 samples/sec	 accuracy=75.226017	 loss=0.995408	 lr=0.001000
Epoch[052] Batch [2199]/[3759]	Speed: 34.914648 samples/sec	 accuracy=75.230824	 loss=0.995114	 lr=0.001000
Epoch[052] Batch [2249]/[3759]	Speed: 35.494312 samples/sec	 accuracy=75.221528	 loss=0.995340	 lr=0.001000
Epoch[052] Batch [2299]/[3759]	Speed: 35.034204 samples/sec	 accuracy=75.216033	 loss=0.995995	 lr=0.001000
Epoch[052] Batch [2349]/[3759]	Speed: 35.469751 samples/sec	 accuracy=75.206782	 loss=0.996325	 lr=0.001000
Epoch[052] Batch [2399]/[3759]	Speed: 35.115594 samples/sec	 accuracy=75.195964	 loss=0.996905	 lr=0.001000
Epoch[052] Batch [2449]/[3759]	Speed: 35.497969 samples/sec	 accuracy=75.183036	 loss=0.996407	 lr=0.001000
Epoch[052] Batch [2499]/[3759]	Speed: 35.059897 samples/sec	 accuracy=75.186875	 loss=0.996475	 lr=0.001000
Epoch[052] Batch [2549]/[3759]	Speed: 35.289456 samples/sec	 accuracy=75.176471	 loss=0.996789	 lr=0.001000
Epoch[052] Batch [2599]/[3759]	Speed: 34.999247 samples/sec	 accuracy=75.174880	 loss=0.997091	 lr=0.001000
Epoch[052] Batch [2649]/[3759]	Speed: 35.148826 samples/sec	 accuracy=75.158608	 loss=0.997641	 lr=0.001000
Epoch[052] Batch [2699]/[3759]	Speed: 35.032872 samples/sec	 accuracy=75.182292	 loss=0.996962	 lr=0.001000
Epoch[052] Batch [2749]/[3759]	Speed: 34.869665 samples/sec	 accuracy=75.157955	 loss=0.997688	 lr=0.001000
Epoch[052] Batch [2799]/[3759]	Speed: 35.380355 samples/sec	 accuracy=75.167411	 loss=0.997909	 lr=0.001000
Epoch[052] Batch [2849]/[3759]	Speed: 34.876552 samples/sec	 accuracy=75.169956	 loss=0.997854	 lr=0.001000
Epoch[052] Batch [2899]/[3759]	Speed: 35.468615 samples/sec	 accuracy=75.169720	 loss=0.998113	 lr=0.001000
Epoch[052] Batch [2949]/[3759]	Speed: 34.911787 samples/sec	 accuracy=75.174258	 loss=0.997464	 lr=0.001000
Epoch[052] Batch [2999]/[3759]	Speed: 35.500558 samples/sec	 accuracy=75.165104	 loss=0.997989	 lr=0.001000
Epoch[052] Batch [3049]/[3759]	Speed: 35.219409 samples/sec	 accuracy=75.150615	 loss=0.998773	 lr=0.001000
Epoch[052] Batch [3099]/[3759]	Speed: 35.437478 samples/sec	 accuracy=75.141129	 loss=0.999290	 lr=0.001000
Epoch[052] Batch [3149]/[3759]	Speed: 35.323100 samples/sec	 accuracy=75.135417	 loss=0.999154	 lr=0.001000
Epoch[052] Batch [3199]/[3759]	Speed: 35.298729 samples/sec	 accuracy=75.128906	 loss=0.999302	 lr=0.001000
Epoch[052] Batch [3249]/[3759]	Speed: 35.238283 samples/sec	 accuracy=75.137981	 loss=0.998799	 lr=0.001000
Epoch[052] Batch [3299]/[3759]	Speed: 35.282274 samples/sec	 accuracy=75.133996	 loss=0.999141	 lr=0.001000
Epoch[052] Batch [3349]/[3759]	Speed: 34.812064 samples/sec	 accuracy=75.135261	 loss=0.998959	 lr=0.001000
Epoch[052] Batch [3399]/[3759]	Speed: 35.097831 samples/sec	 accuracy=75.109835	 loss=0.999714	 lr=0.001000
Epoch[052] Batch [3449]/[3759]	Speed: 35.010018 samples/sec	 accuracy=75.090127	 loss=0.999820	 lr=0.001000
Epoch[052] Batch [3499]/[3759]	Speed: 34.971060 samples/sec	 accuracy=75.072321	 loss=0.999815	 lr=0.001000
Epoch[052] Batch [3549]/[3759]	Speed: 35.261317 samples/sec	 accuracy=75.077905	 loss=0.999802	 lr=0.001000
Epoch[052] Batch [3599]/[3759]	Speed: 34.939005 samples/sec	 accuracy=75.102431	 loss=0.998790	 lr=0.001000
Epoch[052] Batch [3649]/[3759]	Speed: 35.306278 samples/sec	 accuracy=75.114298	 loss=0.998403	 lr=0.001000
Epoch[052] Batch [3699]/[3759]	Speed: 34.877022 samples/sec	 accuracy=75.101774	 loss=0.998847	 lr=0.001000
Epoch[052] Batch [3749]/[3759]	Speed: 36.766089 samples/sec	 accuracy=75.094583	 loss=0.998920	 lr=0.001000
Batch [0049]/[0303]: acc-top1=67.187500 acc-top5=86.125000
Batch [0099]/[0303]: acc-top1=66.281250 acc-top5=85.406250
Batch [0149]/[0303]: acc-top1=66.479167 acc-top5=85.635417
Batch [0199]/[0303]: acc-top1=66.289062 acc-top5=85.765625
Batch [0249]/[0303]: acc-top1=66.062500 acc-top5=85.612500
Batch [0299]/[0303]: acc-top1=66.015625 acc-top5=85.640625
[Epoch 052] training: accuracy=75.098514	 loss=0.998824
[Epoch 052] speed: 34 samples/sec	time cost: 7346.148701
[Epoch 052] validation: acc-top1=66.001444 acc-top5=85.648721 loss=1.582827
Epoch[053] Batch [0049]/[3760]	Speed: 24.584310 samples/sec	 accuracy=74.687500	 loss=0.988981	 lr=0.001000
Epoch[053] Batch [0099]/[3760]	Speed: 34.992459 samples/sec	 accuracy=74.656250	 loss=1.004210	 lr=0.001000
Epoch[053] Batch [0149]/[3760]	Speed: 34.825169 samples/sec	 accuracy=75.000000	 loss=0.993905	 lr=0.001000
Epoch[053] Batch [0199]/[3760]	Speed: 34.976068 samples/sec	 accuracy=75.148438	 loss=0.992562	 lr=0.001000
Epoch[053] Batch [0249]/[3760]	Speed: 35.308215 samples/sec	 accuracy=75.087500	 loss=0.992686	 lr=0.001000
Epoch[053] Batch [0299]/[3760]	Speed: 34.907041 samples/sec	 accuracy=75.072917	 loss=0.994562	 lr=0.001000
Epoch[053] Batch [0349]/[3760]	Speed: 35.267372 samples/sec	 accuracy=75.142857	 loss=0.993702	 lr=0.001000
Epoch[053] Batch [0399]/[3760]	Speed: 34.996984 samples/sec	 accuracy=75.140625	 loss=0.990285	 lr=0.001000
Epoch[053] Batch [0449]/[3760]	Speed: 35.245480 samples/sec	 accuracy=75.017361	 loss=0.992806	 lr=0.001000
Epoch[053] Batch [0499]/[3760]	Speed: 34.798088 samples/sec	 accuracy=75.193750	 loss=0.988431	 lr=0.001000
Epoch[053] Batch [0549]/[3760]	Speed: 35.329419 samples/sec	 accuracy=75.238636	 loss=0.987953	 lr=0.001000
Epoch[053] Batch [0599]/[3760]	Speed: 34.793014 samples/sec	 accuracy=75.250000	 loss=0.988512	 lr=0.001000
Epoch[053] Batch [0649]/[3760]	Speed: 35.581716 samples/sec	 accuracy=75.281250	 loss=0.990657	 lr=0.001000
Epoch[053] Batch [0699]/[3760]	Speed: 34.865295 samples/sec	 accuracy=75.363839	 loss=0.987247	 lr=0.001000
Epoch[053] Batch [0749]/[3760]	Speed: 35.293897 samples/sec	 accuracy=75.406250	 loss=0.985709	 lr=0.001000
Epoch[053] Batch [0799]/[3760]	Speed: 35.200779 samples/sec	 accuracy=75.396484	 loss=0.984442	 lr=0.001000
Epoch[053] Batch [0849]/[3760]	Speed: 35.261704 samples/sec	 accuracy=75.516544	 loss=0.980598	 lr=0.001000
Epoch[053] Batch [0899]/[3760]	Speed: 35.060277 samples/sec	 accuracy=75.491319	 loss=0.982421	 lr=0.001000
Epoch[053] Batch [0949]/[3760]	Speed: 35.355154 samples/sec	 accuracy=75.424342	 loss=0.985945	 lr=0.001000
Epoch[053] Batch [0999]/[3760]	Speed: 35.285263 samples/sec	 accuracy=75.417188	 loss=0.987102	 lr=0.001000
Epoch[053] Batch [1049]/[3760]	Speed: 35.360101 samples/sec	 accuracy=75.433036	 loss=0.987335	 lr=0.001000
Epoch[053] Batch [1099]/[3760]	Speed: 35.142790 samples/sec	 accuracy=75.446023	 loss=0.987570	 lr=0.001000
Epoch[053] Batch [1149]/[3760]	Speed: 35.274168 samples/sec	 accuracy=75.456522	 loss=0.988254	 lr=0.001000
Epoch[053] Batch [1199]/[3760]	Speed: 35.142259 samples/sec	 accuracy=75.457031	 loss=0.988758	 lr=0.001000
Epoch[053] Batch [1249]/[3760]	Speed: 35.333338 samples/sec	 accuracy=75.432500	 loss=0.990368	 lr=0.001000
Epoch[053] Batch [1299]/[3760]	Speed: 35.252841 samples/sec	 accuracy=75.419471	 loss=0.991124	 lr=0.001000
Epoch[053] Batch [1349]/[3760]	Speed: 35.311970 samples/sec	 accuracy=75.439815	 loss=0.989666	 lr=0.001000
Epoch[053] Batch [1399]/[3760]	Speed: 35.269216 samples/sec	 accuracy=75.409598	 loss=0.989061	 lr=0.001000
Epoch[053] Batch [1449]/[3760]	Speed: 35.248838 samples/sec	 accuracy=75.377155	 loss=0.989748	 lr=0.001000
Epoch[053] Batch [1499]/[3760]	Speed: 35.307040 samples/sec	 accuracy=75.379167	 loss=0.989688	 lr=0.001000
Epoch[053] Batch [1549]/[3760]	Speed: 35.314140 samples/sec	 accuracy=75.365927	 loss=0.990662	 lr=0.001000
Epoch[053] Batch [1599]/[3760]	Speed: 35.381530 samples/sec	 accuracy=75.403320	 loss=0.989396	 lr=0.001000
Epoch[053] Batch [1649]/[3760]	Speed: 35.189506 samples/sec	 accuracy=75.410985	 loss=0.989038	 lr=0.001000
Epoch[053] Batch [1699]/[3760]	Speed: 35.527212 samples/sec	 accuracy=75.375000	 loss=0.991002	 lr=0.001000
Epoch[053] Batch [1749]/[3760]	Speed: 35.137717 samples/sec	 accuracy=75.400000	 loss=0.990089	 lr=0.001000
Epoch[053] Batch [1799]/[3760]	Speed: 35.281092 samples/sec	 accuracy=75.347222	 loss=0.991427	 lr=0.001000
Epoch[053] Batch [1849]/[3760]	Speed: 34.951140 samples/sec	 accuracy=75.336149	 loss=0.991934	 lr=0.001000
Epoch[053] Batch [1899]/[3760]	Speed: 35.607559 samples/sec	 accuracy=75.341283	 loss=0.992238	 lr=0.001000
Epoch[053] Batch [1949]/[3760]	Speed: 34.927804 samples/sec	 accuracy=75.322917	 loss=0.992408	 lr=0.001000
Epoch[053] Batch [1999]/[3760]	Speed: 35.316035 samples/sec	 accuracy=75.321875	 loss=0.992151	 lr=0.001000
Epoch[053] Batch [2049]/[3760]	Speed: 34.742627 samples/sec	 accuracy=75.310213	 loss=0.993148	 lr=0.001000
Epoch[053] Batch [2099]/[3760]	Speed: 35.211368 samples/sec	 accuracy=75.313244	 loss=0.993288	 lr=0.001000
Epoch[053] Batch [2149]/[3760]	Speed: 34.981267 samples/sec	 accuracy=75.326308	 loss=0.992911	 lr=0.001000
Epoch[053] Batch [2199]/[3760]	Speed: 34.804637 samples/sec	 accuracy=75.348722	 loss=0.992494	 lr=0.001000
Epoch[053] Batch [2249]/[3760]	Speed: 35.288070 samples/sec	 accuracy=75.341667	 loss=0.993024	 lr=0.001000
Epoch[053] Batch [2299]/[3760]	Speed: 34.626697 samples/sec	 accuracy=75.350543	 loss=0.992962	 lr=0.001000
Epoch[053] Batch [2349]/[3760]	Speed: 35.439464 samples/sec	 accuracy=75.359043	 loss=0.993751	 lr=0.001000
Epoch[053] Batch [2399]/[3760]	Speed: 34.869692 samples/sec	 accuracy=75.375651	 loss=0.993525	 lr=0.001000
Epoch[053] Batch [2449]/[3760]	Speed: 35.495383 samples/sec	 accuracy=75.355867	 loss=0.993979	 lr=0.001000
Epoch[053] Batch [2499]/[3760]	Speed: 35.039788 samples/sec	 accuracy=75.369375	 loss=0.992869	 lr=0.001000
Epoch[053] Batch [2549]/[3760]	Speed: 35.485535 samples/sec	 accuracy=75.335172	 loss=0.994010	 lr=0.001000
Epoch[053] Batch [2599]/[3760]	Speed: 35.048628 samples/sec	 accuracy=75.335938	 loss=0.993908	 lr=0.001000
Epoch[053] Batch [2649]/[3760]	Speed: 35.439039 samples/sec	 accuracy=75.330189	 loss=0.993963	 lr=0.001000
Epoch[053] Batch [2699]/[3760]	Speed: 35.007875 samples/sec	 accuracy=75.333912	 loss=0.993398	 lr=0.001000
Epoch[053] Batch [2749]/[3760]	Speed: 35.638303 samples/sec	 accuracy=75.349432	 loss=0.992731	 lr=0.001000
Epoch[053] Batch [2799]/[3760]	Speed: 35.148889 samples/sec	 accuracy=75.360491	 loss=0.991843	 lr=0.001000
Epoch[053] Batch [2849]/[3760]	Speed: 35.513369 samples/sec	 accuracy=75.357456	 loss=0.991658	 lr=0.001000
Epoch[053] Batch [2899]/[3760]	Speed: 35.276716 samples/sec	 accuracy=75.361530	 loss=0.991545	 lr=0.001000
Epoch[053] Batch [2949]/[3760]	Speed: 35.350633 samples/sec	 accuracy=75.369703	 loss=0.991328	 lr=0.001000
Epoch[053] Batch [2999]/[3760]	Speed: 35.226907 samples/sec	 accuracy=75.364062	 loss=0.991721	 lr=0.001000
Epoch[053] Batch [3049]/[3760]	Speed: 35.344732 samples/sec	 accuracy=75.337090	 loss=0.992816	 lr=0.001000
Epoch[053] Batch [3099]/[3760]	Speed: 35.397787 samples/sec	 accuracy=75.324597	 loss=0.992545	 lr=0.001000
Epoch[053] Batch [3149]/[3760]	Speed: 35.034983 samples/sec	 accuracy=75.320933	 loss=0.992277	 lr=0.001000
Epoch[053] Batch [3199]/[3760]	Speed: 35.488292 samples/sec	 accuracy=75.321777	 loss=0.992155	 lr=0.001000
Epoch[053] Batch [3249]/[3760]	Speed: 34.925546 samples/sec	 accuracy=75.308654	 loss=0.992152	 lr=0.001000
Epoch[053] Batch [3299]/[3760]	Speed: 35.595794 samples/sec	 accuracy=75.311553	 loss=0.992320	 lr=0.001000
Epoch[053] Batch [3349]/[3760]	Speed: 35.303774 samples/sec	 accuracy=75.305037	 loss=0.991829	 lr=0.001000
Epoch[053] Batch [3399]/[3760]	Speed: 35.533076 samples/sec	 accuracy=75.321232	 loss=0.991355	 lr=0.001000
Epoch[053] Batch [3449]/[3760]	Speed: 35.269407 samples/sec	 accuracy=75.327446	 loss=0.990677	 lr=0.001000
Epoch[053] Batch [3499]/[3760]	Speed: 35.457627 samples/sec	 accuracy=75.318304	 loss=0.990777	 lr=0.001000
Epoch[053] Batch [3549]/[3760]	Speed: 34.851961 samples/sec	 accuracy=75.316021	 loss=0.990713	 lr=0.001000
Epoch[053] Batch [3599]/[3760]	Speed: 35.358074 samples/sec	 accuracy=75.325087	 loss=0.990568	 lr=0.001000
Epoch[053] Batch [3649]/[3760]	Speed: 34.908444 samples/sec	 accuracy=75.321062	 loss=0.990336	 lr=0.001000
Epoch[053] Batch [3699]/[3760]	Speed: 35.542985 samples/sec	 accuracy=75.320524	 loss=0.990741	 lr=0.001000
Epoch[053] Batch [3749]/[3760]	Speed: 36.269005 samples/sec	 accuracy=75.315000	 loss=0.990926	 lr=0.001000
Batch [0049]/[0303]: acc-top1=67.312500 acc-top5=85.687500
Batch [0099]/[0303]: acc-top1=66.390625 acc-top5=85.281250
Batch [0149]/[0303]: acc-top1=66.614583 acc-top5=85.625000
Batch [0199]/[0303]: acc-top1=66.195312 acc-top5=85.687500
Batch [0249]/[0303]: acc-top1=65.956250 acc-top5=85.506250
Batch [0299]/[0303]: acc-top1=65.895833 acc-top5=85.526042
[Epoch 053] training: accuracy=75.314993	 loss=0.990775
[Epoch 053] speed: 35 samples/sec	time cost: 7332.857378
[Epoch 053] validation: acc-top1=65.908622 acc-top5=85.524959 loss=1.590397
Epoch[054] Batch [0049]/[3760]	Speed: 24.818456 samples/sec	 accuracy=75.843750	 loss=0.984465	 lr=0.001000
Epoch[054] Batch [0099]/[3760]	Speed: 35.013282 samples/sec	 accuracy=76.281250	 loss=0.958697	 lr=0.001000
Epoch[054] Batch [0149]/[3760]	Speed: 35.089234 samples/sec	 accuracy=75.968750	 loss=0.972782	 lr=0.001000
Epoch[054] Batch [0199]/[3760]	Speed: 34.997799 samples/sec	 accuracy=75.945312	 loss=0.965417	 lr=0.001000
Epoch[054] Batch [0249]/[3760]	Speed: 35.161182 samples/sec	 accuracy=75.675000	 loss=0.972543	 lr=0.001000
Epoch[054] Batch [0299]/[3760]	Speed: 35.173954 samples/sec	 accuracy=75.734375	 loss=0.974565	 lr=0.001000
Epoch[054] Batch [0349]/[3760]	Speed: 35.375153 samples/sec	 accuracy=75.816964	 loss=0.969134	 lr=0.001000
Epoch[054] Batch [0399]/[3760]	Speed: 35.132234 samples/sec	 accuracy=75.839844	 loss=0.970080	 lr=0.001000
Epoch[054] Batch [0449]/[3760]	Speed: 35.323565 samples/sec	 accuracy=75.881944	 loss=0.967385	 lr=0.001000
Epoch[054] Batch [0499]/[3760]	Speed: 35.002715 samples/sec	 accuracy=75.909375	 loss=0.965929	 lr=0.001000
Epoch[054] Batch [0549]/[3760]	Speed: 35.263096 samples/sec	 accuracy=75.815341	 loss=0.969694	 lr=0.001000
Epoch[054] Batch [0599]/[3760]	Speed: 34.808213 samples/sec	 accuracy=75.773438	 loss=0.970116	 lr=0.001000
Epoch[054] Batch [0649]/[3760]	Speed: 35.329201 samples/sec	 accuracy=75.817308	 loss=0.970131	 lr=0.001000
Epoch[054] Batch [0699]/[3760]	Speed: 34.917015 samples/sec	 accuracy=75.801339	 loss=0.970285	 lr=0.001000
Epoch[054] Batch [0749]/[3760]	Speed: 35.432271 samples/sec	 accuracy=75.787500	 loss=0.968936	 lr=0.001000
Epoch[054] Batch [0799]/[3760]	Speed: 34.809780 samples/sec	 accuracy=75.744141	 loss=0.970610	 lr=0.001000
Epoch[054] Batch [0849]/[3760]	Speed: 35.512756 samples/sec	 accuracy=75.761029	 loss=0.968284	 lr=0.001000
Epoch[054] Batch [0899]/[3760]	Speed: 34.985589 samples/sec	 accuracy=75.765625	 loss=0.968770	 lr=0.001000
Epoch[054] Batch [0949]/[3760]	Speed: 35.212934 samples/sec	 accuracy=75.802632	 loss=0.967918	 lr=0.001000
Epoch[054] Batch [0999]/[3760]	Speed: 35.185245 samples/sec	 accuracy=75.762500	 loss=0.968617	 lr=0.001000
Epoch[054] Batch [1049]/[3760]	Speed: 34.997145 samples/sec	 accuracy=75.748512	 loss=0.969144	 lr=0.001000
Epoch[054] Batch [1099]/[3760]	Speed: 35.220045 samples/sec	 accuracy=75.723011	 loss=0.970088	 lr=0.001000
Epoch[054] Batch [1149]/[3760]	Speed: 35.058189 samples/sec	 accuracy=75.736413	 loss=0.970215	 lr=0.001000
Epoch[054] Batch [1199]/[3760]	Speed: 35.301414 samples/sec	 accuracy=75.708333	 loss=0.970616	 lr=0.001000
Epoch[054] Batch [1249]/[3760]	Speed: 35.253446 samples/sec	 accuracy=75.696250	 loss=0.971296	 lr=0.001000
Epoch[054] Batch [1299]/[3760]	Speed: 35.354953 samples/sec	 accuracy=75.745192	 loss=0.970891	 lr=0.001000
Epoch[054] Batch [1349]/[3760]	Speed: 35.077634 samples/sec	 accuracy=75.782407	 loss=0.970143	 lr=0.001000
Epoch[054] Batch [1399]/[3760]	Speed: 35.337629 samples/sec	 accuracy=75.746652	 loss=0.970315	 lr=0.001000
Epoch[054] Batch [1449]/[3760]	Speed: 34.880644 samples/sec	 accuracy=75.758621	 loss=0.970278	 lr=0.001000
Epoch[054] Batch [1499]/[3760]	Speed: 35.372148 samples/sec	 accuracy=75.797917	 loss=0.969008	 lr=0.001000
Epoch[054] Batch [1549]/[3760]	Speed: 34.819709 samples/sec	 accuracy=75.802419	 loss=0.970278	 lr=0.001000
Epoch[054] Batch [1599]/[3760]	Speed: 35.316875 samples/sec	 accuracy=75.812500	 loss=0.969432	 lr=0.001000
Epoch[054] Batch [1649]/[3760]	Speed: 34.913240 samples/sec	 accuracy=75.785038	 loss=0.969502	 lr=0.001000
Epoch[054] Batch [1699]/[3760]	Speed: 35.297609 samples/sec	 accuracy=75.787684	 loss=0.969267	 lr=0.001000
Epoch[054] Batch [1749]/[3760]	Speed: 34.899281 samples/sec	 accuracy=75.787500	 loss=0.969557	 lr=0.001000
Epoch[054] Batch [1799]/[3760]	Speed: 35.127949 samples/sec	 accuracy=75.794271	 loss=0.970137	 lr=0.001000
Epoch[054] Batch [1849]/[3760]	Speed: 34.928603 samples/sec	 accuracy=75.783784	 loss=0.970443	 lr=0.001000
Epoch[054] Batch [1899]/[3760]	Speed: 35.174522 samples/sec	 accuracy=75.787007	 loss=0.970854	 lr=0.001000
Epoch[054] Batch [1949]/[3760]	Speed: 35.032540 samples/sec	 accuracy=75.756410	 loss=0.971353	 lr=0.001000
Epoch[054] Batch [1999]/[3760]	Speed: 34.947918 samples/sec	 accuracy=75.748437	 loss=0.971923	 lr=0.001000
Epoch[054] Batch [2049]/[3760]	Speed: 35.261030 samples/sec	 accuracy=75.703506	 loss=0.973099	 lr=0.001000
Epoch[054] Batch [2099]/[3760]	Speed: 34.716699 samples/sec	 accuracy=75.706845	 loss=0.973024	 lr=0.001000
Epoch[054] Batch [2149]/[3760]	Speed: 35.242315 samples/sec	 accuracy=75.704942	 loss=0.972262	 lr=0.001000
Epoch[054] Batch [2199]/[3760]	Speed: 34.624894 samples/sec	 accuracy=75.708807	 loss=0.971865	 lr=0.001000
Epoch[054] Batch [2249]/[3760]	Speed: 35.368756 samples/sec	 accuracy=75.731944	 loss=0.971339	 lr=0.001000
Epoch[054] Batch [2299]/[3760]	Speed: 34.809716 samples/sec	 accuracy=75.718071	 loss=0.972428	 lr=0.001000
Epoch[054] Batch [2349]/[3760]	Speed: 35.387522 samples/sec	 accuracy=75.694149	 loss=0.973269	 lr=0.001000
Epoch[054] Batch [2399]/[3760]	Speed: 34.913954 samples/sec	 accuracy=75.684245	 loss=0.973755	 lr=0.001000
Epoch[054] Batch [2449]/[3760]	Speed: 35.346072 samples/sec	 accuracy=75.681760	 loss=0.973858	 lr=0.001000
Epoch[054] Batch [2499]/[3760]	Speed: 34.771473 samples/sec	 accuracy=75.664375	 loss=0.974597	 lr=0.001000
Epoch[054] Batch [2549]/[3760]	Speed: 34.967608 samples/sec	 accuracy=75.664216	 loss=0.974615	 lr=0.001000
Epoch[054] Batch [2599]/[3760]	Speed: 34.956294 samples/sec	 accuracy=75.663462	 loss=0.974642	 lr=0.001000
Epoch[054] Batch [2649]/[3760]	Speed: 34.970993 samples/sec	 accuracy=75.661557	 loss=0.974951	 lr=0.001000
Epoch[054] Batch [2699]/[3760]	Speed: 35.090742 samples/sec	 accuracy=75.660880	 loss=0.975209	 lr=0.001000
Epoch[054] Batch [2749]/[3760]	Speed: 34.634231 samples/sec	 accuracy=75.652273	 loss=0.975343	 lr=0.001000
Epoch[054] Batch [2799]/[3760]	Speed: 35.138939 samples/sec	 accuracy=75.636161	 loss=0.975539	 lr=0.001000
Epoch[054] Batch [2849]/[3760]	Speed: 34.468820 samples/sec	 accuracy=75.627193	 loss=0.975744	 lr=0.001000
Epoch[054] Batch [2899]/[3760]	Speed: 35.167753 samples/sec	 accuracy=75.633082	 loss=0.975610	 lr=0.001000
Epoch[054] Batch [2949]/[3760]	Speed: 34.235999 samples/sec	 accuracy=75.653072	 loss=0.974742	 lr=0.001000
Epoch[054] Batch [2999]/[3760]	Speed: 35.054990 samples/sec	 accuracy=75.637500	 loss=0.975175	 lr=0.001000
Epoch[054] Batch [3049]/[3760]	Speed: 34.553954 samples/sec	 accuracy=75.636270	 loss=0.975331	 lr=0.001000
Epoch[054] Batch [3099]/[3760]	Speed: 34.993746 samples/sec	 accuracy=75.632056	 loss=0.975268	 lr=0.001000
Epoch[054] Batch [3149]/[3760]	Speed: 34.866995 samples/sec	 accuracy=75.612103	 loss=0.975861	 lr=0.001000
Epoch[054] Batch [3199]/[3760]	Speed: 34.673853 samples/sec	 accuracy=75.633789	 loss=0.975402	 lr=0.001000
Epoch[054] Batch [3249]/[3760]	Speed: 35.113049 samples/sec	 accuracy=75.636058	 loss=0.975445	 lr=0.001000
Epoch[054] Batch [3299]/[3760]	Speed: 34.493498 samples/sec	 accuracy=75.633523	 loss=0.975339	 lr=0.001000
Epoch[054] Batch [3349]/[3760]	Speed: 35.175337 samples/sec	 accuracy=75.640392	 loss=0.975150	 lr=0.001000
Epoch[054] Batch [3399]/[3760]	Speed: 34.570059 samples/sec	 accuracy=75.621783	 loss=0.975867	 lr=0.001000
Epoch[054] Batch [3449]/[3760]	Speed: 35.306052 samples/sec	 accuracy=75.610507	 loss=0.977072	 lr=0.001000
Epoch[054] Batch [3499]/[3760]	Speed: 34.614198 samples/sec	 accuracy=75.602679	 loss=0.977239	 lr=0.001000
Epoch[054] Batch [3549]/[3760]	Speed: 35.197825 samples/sec	 accuracy=75.601232	 loss=0.976833	 lr=0.001000
Epoch[054] Batch [3599]/[3760]	Speed: 34.773068 samples/sec	 accuracy=75.585938	 loss=0.977327	 lr=0.001000
Epoch[054] Batch [3649]/[3760]	Speed: 34.781876 samples/sec	 accuracy=75.574914	 loss=0.977708	 lr=0.001000
Epoch[054] Batch [3699]/[3760]	Speed: 35.052862 samples/sec	 accuracy=75.564611	 loss=0.978197	 lr=0.001000
Epoch[054] Batch [3749]/[3760]	Speed: 36.196618 samples/sec	 accuracy=75.555833	 loss=0.978692	 lr=0.001000
Batch [0049]/[0303]: acc-top1=67.156250 acc-top5=86.406250
Batch [0099]/[0303]: acc-top1=66.421875 acc-top5=85.593750
Batch [0149]/[0303]: acc-top1=66.666667 acc-top5=85.656250
Batch [0199]/[0303]: acc-top1=66.296875 acc-top5=85.742188
Batch [0249]/[0303]: acc-top1=66.100000 acc-top5=85.643750
Batch [0299]/[0303]: acc-top1=66.109375 acc-top5=85.625000
[Epoch 054] training: accuracy=75.547706	 loss=0.978954
[Epoch 054] speed: 34 samples/sec	time cost: 7365.327388
[Epoch 054] validation: acc-top1=66.099422 acc-top5=85.638408 loss=1.588307
Epoch[055] Batch [0049]/[3759]	Speed: 25.054835 samples/sec	 accuracy=75.000000	 loss=0.996908	 lr=0.001000
Epoch[055] Batch [0099]/[3759]	Speed: 35.398371 samples/sec	 accuracy=75.468750	 loss=0.966182	 lr=0.001000
Epoch[055] Batch [0149]/[3759]	Speed: 35.058327 samples/sec	 accuracy=75.802083	 loss=0.962707	 lr=0.001000
Epoch[055] Batch [0199]/[3759]	Speed: 35.390465 samples/sec	 accuracy=75.750000	 loss=0.963592	 lr=0.001000
Epoch[055] Batch [0249]/[3759]	Speed: 35.143607 samples/sec	 accuracy=75.806250	 loss=0.964406	 lr=0.001000
Epoch[055] Batch [0299]/[3759]	Speed: 35.228671 samples/sec	 accuracy=75.765625	 loss=0.969998	 lr=0.001000
Epoch[055] Batch [0349]/[3759]	Speed: 35.136359 samples/sec	 accuracy=75.732143	 loss=0.974853	 lr=0.001000
Epoch[055] Batch [0399]/[3759]	Speed: 35.236705 samples/sec	 accuracy=75.617188	 loss=0.978036	 lr=0.001000
Epoch[055] Batch [0449]/[3759]	Speed: 35.230511 samples/sec	 accuracy=75.604167	 loss=0.977670	 lr=0.001000
Epoch[055] Batch [0499]/[3759]	Speed: 34.981771 samples/sec	 accuracy=75.687500	 loss=0.971598	 lr=0.001000
Epoch[055] Batch [0549]/[3759]	Speed: 35.268556 samples/sec	 accuracy=75.698864	 loss=0.969724	 lr=0.001000
Epoch[055] Batch [0599]/[3759]	Speed: 35.098450 samples/sec	 accuracy=75.713542	 loss=0.967887	 lr=0.001000
Epoch[055] Batch [0649]/[3759]	Speed: 35.472694 samples/sec	 accuracy=75.795673	 loss=0.968047	 lr=0.001000
Epoch[055] Batch [0699]/[3759]	Speed: 35.095290 samples/sec	 accuracy=75.825893	 loss=0.967527	 lr=0.001000
Epoch[055] Batch [0749]/[3759]	Speed: 35.088887 samples/sec	 accuracy=75.841667	 loss=0.966472	 lr=0.001000
Epoch[055] Batch [0799]/[3759]	Speed: 34.987124 samples/sec	 accuracy=75.912109	 loss=0.964287	 lr=0.001000
Epoch[055] Batch [0849]/[3759]	Speed: 35.470606 samples/sec	 accuracy=75.946691	 loss=0.964620	 lr=0.001000
Epoch[055] Batch [0899]/[3759]	Speed: 34.940856 samples/sec	 accuracy=75.895833	 loss=0.967567	 lr=0.001000
Epoch[055] Batch [0949]/[3759]	Speed: 35.390945 samples/sec	 accuracy=75.939145	 loss=0.965136	 lr=0.001000
Epoch[055] Batch [0999]/[3759]	Speed: 35.238284 samples/sec	 accuracy=75.931250	 loss=0.964105	 lr=0.001000
Epoch[055] Batch [1049]/[3759]	Speed: 35.556980 samples/sec	 accuracy=75.953869	 loss=0.963161	 lr=0.001000
Epoch[055] Batch [1099]/[3759]	Speed: 34.950564 samples/sec	 accuracy=75.967330	 loss=0.963407	 lr=0.001000
Epoch[055] Batch [1149]/[3759]	Speed: 35.254176 samples/sec	 accuracy=75.974185	 loss=0.963783	 lr=0.001000
Epoch[055] Batch [1199]/[3759]	Speed: 35.034330 samples/sec	 accuracy=76.023438	 loss=0.962154	 lr=0.001000
Epoch[055] Batch [1249]/[3759]	Speed: 35.409952 samples/sec	 accuracy=76.036250	 loss=0.961986	 lr=0.001000
Epoch[055] Batch [1299]/[3759]	Speed: 35.227104 samples/sec	 accuracy=76.074519	 loss=0.960894	 lr=0.001000
Epoch[055] Batch [1349]/[3759]	Speed: 35.124905 samples/sec	 accuracy=76.064815	 loss=0.961129	 lr=0.001000
Epoch[055] Batch [1399]/[3759]	Speed: 35.226516 samples/sec	 accuracy=76.098214	 loss=0.960255	 lr=0.001000
Epoch[055] Batch [1449]/[3759]	Speed: 35.163679 samples/sec	 accuracy=76.112069	 loss=0.958953	 lr=0.001000
Epoch[055] Batch [1499]/[3759]	Speed: 35.137544 samples/sec	 accuracy=76.084375	 loss=0.959297	 lr=0.001000
Epoch[055] Batch [1549]/[3759]	Speed: 34.955432 samples/sec	 accuracy=76.052419	 loss=0.960468	 lr=0.001000
Epoch[055] Batch [1599]/[3759]	Speed: 35.425998 samples/sec	 accuracy=76.081055	 loss=0.959796	 lr=0.001000
Epoch[055] Batch [1649]/[3759]	Speed: 35.141275 samples/sec	 accuracy=76.053030	 loss=0.960331	 lr=0.001000
Epoch[055] Batch [1699]/[3759]	Speed: 35.395396 samples/sec	 accuracy=76.054228	 loss=0.960453	 lr=0.001000
Epoch[055] Batch [1749]/[3759]	Speed: 35.047104 samples/sec	 accuracy=76.041071	 loss=0.960463	 lr=0.001000
Epoch[055] Batch [1799]/[3759]	Speed: 35.358504 samples/sec	 accuracy=76.007812	 loss=0.961665	 lr=0.001000
Epoch[055] Batch [1849]/[3759]	Speed: 35.211835 samples/sec	 accuracy=76.013514	 loss=0.962275	 lr=0.001000
Epoch[055] Batch [1899]/[3759]	Speed: 35.496600 samples/sec	 accuracy=75.962993	 loss=0.963836	 lr=0.001000
Epoch[055] Batch [1949]/[3759]	Speed: 35.289549 samples/sec	 accuracy=75.947917	 loss=0.964327	 lr=0.001000
Epoch[055] Batch [1999]/[3759]	Speed: 35.313472 samples/sec	 accuracy=75.918750	 loss=0.965176	 lr=0.001000
Epoch[055] Batch [2049]/[3759]	Speed: 35.354365 samples/sec	 accuracy=75.887195	 loss=0.966314	 lr=0.001000
Epoch[055] Batch [2099]/[3759]	Speed: 35.099046 samples/sec	 accuracy=75.896577	 loss=0.966566	 lr=0.001000
Epoch[055] Batch [2149]/[3759]	Speed: 35.290953 samples/sec	 accuracy=75.864826	 loss=0.967637	 lr=0.001000
Epoch[055] Batch [2199]/[3759]	Speed: 35.269882 samples/sec	 accuracy=75.865767	 loss=0.967302	 lr=0.001000
Epoch[055] Batch [2249]/[3759]	Speed: 35.087024 samples/sec	 accuracy=75.838194	 loss=0.967886	 lr=0.001000
Epoch[055] Batch [2299]/[3759]	Speed: 35.242016 samples/sec	 accuracy=75.875679	 loss=0.966306	 lr=0.001000
Epoch[055] Batch [2349]/[3759]	Speed: 35.302486 samples/sec	 accuracy=75.863032	 loss=0.966596	 lr=0.001000
Epoch[055] Batch [2399]/[3759]	Speed: 35.401873 samples/sec	 accuracy=75.856120	 loss=0.966484	 lr=0.001000
Epoch[055] Batch [2449]/[3759]	Speed: 35.240781 samples/sec	 accuracy=75.811862	 loss=0.968080	 lr=0.001000
Epoch[055] Batch [2499]/[3759]	Speed: 35.332978 samples/sec	 accuracy=75.788750	 loss=0.968668	 lr=0.001000
Epoch[055] Batch [2549]/[3759]	Speed: 35.446620 samples/sec	 accuracy=75.813725	 loss=0.968150	 lr=0.001000
Epoch[055] Batch [2599]/[3759]	Speed: 35.297252 samples/sec	 accuracy=75.799880	 loss=0.968481	 lr=0.001000
Epoch[055] Batch [2649]/[3759]	Speed: 35.289804 samples/sec	 accuracy=75.789505	 loss=0.968957	 lr=0.001000
Epoch[055] Batch [2699]/[3759]	Speed: 35.195278 samples/sec	 accuracy=75.821181	 loss=0.967886	 lr=0.001000
Epoch[055] Batch [2749]/[3759]	Speed: 35.461128 samples/sec	 accuracy=75.817614	 loss=0.967822	 lr=0.001000
Epoch[055] Batch [2799]/[3759]	Speed: 35.258874 samples/sec	 accuracy=75.796875	 loss=0.967952	 lr=0.001000
Epoch[055] Batch [2849]/[3759]	Speed: 35.164430 samples/sec	 accuracy=75.793311	 loss=0.968353	 lr=0.001000
Epoch[055] Batch [2899]/[3759]	Speed: 35.320533 samples/sec	 accuracy=75.778556	 loss=0.968745	 lr=0.001000
Epoch[055] Batch [2949]/[3759]	Speed: 35.006747 samples/sec	 accuracy=75.758475	 loss=0.969537	 lr=0.001000
Epoch[055] Batch [2999]/[3759]	Speed: 35.183590 samples/sec	 accuracy=75.753125	 loss=0.969340	 lr=0.001000
Epoch[055] Batch [3049]/[3759]	Speed: 35.129317 samples/sec	 accuracy=75.758197	 loss=0.969221	 lr=0.001000
Epoch[055] Batch [3099]/[3759]	Speed: 35.301038 samples/sec	 accuracy=75.770161	 loss=0.969105	 lr=0.001000
Epoch[055] Batch [3149]/[3759]	Speed: 34.939798 samples/sec	 accuracy=75.782738	 loss=0.968672	 lr=0.001000
Epoch[055] Batch [3199]/[3759]	Speed: 35.349578 samples/sec	 accuracy=75.786621	 loss=0.968626	 lr=0.001000
Epoch[055] Batch [3249]/[3759]	Speed: 34.917172 samples/sec	 accuracy=75.792788	 loss=0.968272	 lr=0.001000
Epoch[055] Batch [3299]/[3759]	Speed: 35.442892 samples/sec	 accuracy=75.772254	 loss=0.969157	 lr=0.001000
Epoch[055] Batch [3349]/[3759]	Speed: 34.984420 samples/sec	 accuracy=75.778918	 loss=0.968808	 lr=0.001000
Epoch[055] Batch [3399]/[3759]	Speed: 35.250618 samples/sec	 accuracy=75.764246	 loss=0.969096	 lr=0.001000
Epoch[055] Batch [3449]/[3759]	Speed: 34.896037 samples/sec	 accuracy=75.777174	 loss=0.968448	 lr=0.001000
Epoch[055] Batch [3499]/[3759]	Speed: 35.257157 samples/sec	 accuracy=75.798214	 loss=0.967803	 lr=0.001000
Epoch[055] Batch [3549]/[3759]	Speed: 35.024128 samples/sec	 accuracy=75.774208	 loss=0.968675	 lr=0.001000
Epoch[055] Batch [3599]/[3759]	Speed: 35.133084 samples/sec	 accuracy=75.763455	 loss=0.969263	 lr=0.001000
Epoch[055] Batch [3649]/[3759]	Speed: 35.276451 samples/sec	 accuracy=75.744863	 loss=0.970174	 lr=0.001000
Epoch[055] Batch [3699]/[3759]	Speed: 35.003679 samples/sec	 accuracy=75.736064	 loss=0.970605	 lr=0.001000
Epoch[055] Batch [3749]/[3759]	Speed: 36.557486 samples/sec	 accuracy=75.730417	 loss=0.970639	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.500000 acc-top5=86.281250
Batch [0099]/[0303]: acc-top1=66.218750 acc-top5=85.328125
Batch [0149]/[0303]: acc-top1=66.468750 acc-top5=85.468750
Batch [0199]/[0303]: acc-top1=66.039062 acc-top5=85.578125
Batch [0249]/[0303]: acc-top1=66.000000 acc-top5=85.481250
Batch [0299]/[0303]: acc-top1=65.859375 acc-top5=85.567708
[Epoch 055] training: accuracy=75.727005	 loss=0.970827
[Epoch 055] speed: 35 samples/sec	time cost: 7328.437988
[Epoch 055] validation: acc-top1=65.851898 acc-top5=85.571370 loss=1.584064
Epoch[056] Batch [0049]/[3760]	Speed: 24.897281 samples/sec	 accuracy=78.062500	 loss=0.887118	 lr=0.001000
Epoch[056] Batch [0099]/[3760]	Speed: 35.019942 samples/sec	 accuracy=77.312500	 loss=0.909750	 lr=0.001000
Epoch[056] Batch [0149]/[3760]	Speed: 34.939665 samples/sec	 accuracy=77.375000	 loss=0.909072	 lr=0.001000
Epoch[056] Batch [0199]/[3760]	Speed: 34.777808 samples/sec	 accuracy=76.875000	 loss=0.926278	 lr=0.001000
Epoch[056] Batch [0249]/[3760]	Speed: 35.412851 samples/sec	 accuracy=76.762500	 loss=0.929173	 lr=0.001000
Epoch[056] Batch [0299]/[3760]	Speed: 34.985985 samples/sec	 accuracy=76.765625	 loss=0.931814	 lr=0.001000
Epoch[056] Batch [0349]/[3760]	Speed: 35.261779 samples/sec	 accuracy=76.830357	 loss=0.931323	 lr=0.001000
Epoch[056] Batch [0399]/[3760]	Speed: 34.982451 samples/sec	 accuracy=76.726562	 loss=0.930956	 lr=0.001000
Epoch[056] Batch [0449]/[3760]	Speed: 35.379680 samples/sec	 accuracy=76.805556	 loss=0.930008	 lr=0.001000
Epoch[056] Batch [0499]/[3760]	Speed: 34.783259 samples/sec	 accuracy=76.703125	 loss=0.932586	 lr=0.001000
Epoch[056] Batch [0549]/[3760]	Speed: 35.600319 samples/sec	 accuracy=76.659091	 loss=0.936087	 lr=0.001000
Epoch[056] Batch [0599]/[3760]	Speed: 34.993119 samples/sec	 accuracy=76.510417	 loss=0.939586	 lr=0.001000
Epoch[056] Batch [0649]/[3760]	Speed: 35.339932 samples/sec	 accuracy=76.526442	 loss=0.938963	 lr=0.001000
Epoch[056] Batch [0699]/[3760]	Speed: 34.943274 samples/sec	 accuracy=76.513393	 loss=0.938772	 lr=0.001000
Epoch[056] Batch [0749]/[3760]	Speed: 35.404970 samples/sec	 accuracy=76.475000	 loss=0.940738	 lr=0.001000
Epoch[056] Batch [0799]/[3760]	Speed: 35.089724 samples/sec	 accuracy=76.474609	 loss=0.940949	 lr=0.001000
Epoch[056] Batch [0849]/[3760]	Speed: 35.173047 samples/sec	 accuracy=76.514706	 loss=0.938789	 lr=0.001000
Epoch[056] Batch [0899]/[3760]	Speed: 35.380407 samples/sec	 accuracy=76.522569	 loss=0.936679	 lr=0.001000
Epoch[056] Batch [0949]/[3760]	Speed: 35.306952 samples/sec	 accuracy=76.480263	 loss=0.938541	 lr=0.001000
Epoch[056] Batch [0999]/[3760]	Speed: 35.424695 samples/sec	 accuracy=76.501562	 loss=0.937927	 lr=0.001000
Epoch[056] Batch [1049]/[3760]	Speed: 35.294649 samples/sec	 accuracy=76.479167	 loss=0.938928	 lr=0.001000
Epoch[056] Batch [1099]/[3760]	Speed: 35.345898 samples/sec	 accuracy=76.441761	 loss=0.940565	 lr=0.001000
Epoch[056] Batch [1149]/[3760]	Speed: 35.252833 samples/sec	 accuracy=76.463315	 loss=0.940766	 lr=0.001000
Epoch[056] Batch [1199]/[3760]	Speed: 35.259149 samples/sec	 accuracy=76.467448	 loss=0.939608	 lr=0.001000
Epoch[056] Batch [1249]/[3760]	Speed: 35.278591 samples/sec	 accuracy=76.456250	 loss=0.939902	 lr=0.001000
Epoch[056] Batch [1299]/[3760]	Speed: 35.467017 samples/sec	 accuracy=76.468750	 loss=0.939924	 lr=0.001000
Epoch[056] Batch [1349]/[3760]	Speed: 35.254093 samples/sec	 accuracy=76.438657	 loss=0.940185	 lr=0.001000
Epoch[056] Batch [1399]/[3760]	Speed: 35.460428 samples/sec	 accuracy=76.450893	 loss=0.941024	 lr=0.001000
Epoch[056] Batch [1449]/[3760]	Speed: 35.193049 samples/sec	 accuracy=76.395474	 loss=0.942830	 lr=0.001000
Epoch[056] Batch [1499]/[3760]	Speed: 35.325676 samples/sec	 accuracy=76.367708	 loss=0.943798	 lr=0.001000
Epoch[056] Batch [1549]/[3760]	Speed: 35.103752 samples/sec	 accuracy=76.370968	 loss=0.943058	 lr=0.001000
Epoch[056] Batch [1599]/[3760]	Speed: 35.369628 samples/sec	 accuracy=76.326172	 loss=0.944343	 lr=0.001000
Epoch[056] Batch [1649]/[3760]	Speed: 35.185956 samples/sec	 accuracy=76.318182	 loss=0.945101	 lr=0.001000
Epoch[056] Batch [1699]/[3760]	Speed: 35.289100 samples/sec	 accuracy=76.317096	 loss=0.944795	 lr=0.001000
Epoch[056] Batch [1749]/[3760]	Speed: 35.019161 samples/sec	 accuracy=76.269643	 loss=0.946271	 lr=0.001000
Epoch[056] Batch [1799]/[3760]	Speed: 35.478303 samples/sec	 accuracy=76.251736	 loss=0.947127	 lr=0.001000
Epoch[056] Batch [1849]/[3760]	Speed: 35.235222 samples/sec	 accuracy=76.253378	 loss=0.947376	 lr=0.001000
Epoch[056] Batch [1899]/[3760]	Speed: 35.405533 samples/sec	 accuracy=76.240132	 loss=0.947796	 lr=0.001000
Epoch[056] Batch [1949]/[3760]	Speed: 35.148521 samples/sec	 accuracy=76.242788	 loss=0.948198	 lr=0.001000
Epoch[056] Batch [1999]/[3760]	Speed: 35.412974 samples/sec	 accuracy=76.247656	 loss=0.948345	 lr=0.001000
Epoch[056] Batch [2049]/[3760]	Speed: 35.546855 samples/sec	 accuracy=76.223323	 loss=0.948801	 lr=0.001000
Epoch[056] Batch [2099]/[3760]	Speed: 35.371322 samples/sec	 accuracy=76.227679	 loss=0.948639	 lr=0.001000
Epoch[056] Batch [2149]/[3760]	Speed: 35.285006 samples/sec	 accuracy=76.208576	 loss=0.948698	 lr=0.001000
Epoch[056] Batch [2199]/[3760]	Speed: 35.558939 samples/sec	 accuracy=76.235085	 loss=0.947767	 lr=0.001000
Epoch[056] Batch [2249]/[3760]	Speed: 35.313262 samples/sec	 accuracy=76.227083	 loss=0.948553	 lr=0.001000
Epoch[056] Batch [2299]/[3760]	Speed: 35.435059 samples/sec	 accuracy=76.209239	 loss=0.949782	 lr=0.001000
Epoch[056] Batch [2349]/[3760]	Speed: 35.575933 samples/sec	 accuracy=76.193484	 loss=0.950653	 lr=0.001000
Epoch[056] Batch [2399]/[3760]	Speed: 35.444076 samples/sec	 accuracy=76.176432	 loss=0.950950	 lr=0.001000
Epoch[056] Batch [2449]/[3760]	Speed: 35.410723 samples/sec	 accuracy=76.172832	 loss=0.951148	 lr=0.001000
Epoch[056] Batch [2499]/[3760]	Speed: 35.413463 samples/sec	 accuracy=76.185625	 loss=0.951181	 lr=0.001000
Epoch[056] Batch [2549]/[3760]	Speed: 35.516833 samples/sec	 accuracy=76.163603	 loss=0.952198	 lr=0.001000
Epoch[056] Batch [2599]/[3760]	Speed: 35.277328 samples/sec	 accuracy=76.174279	 loss=0.952254	 lr=0.001000
Epoch[056] Batch [2649]/[3760]	Speed: 35.557487 samples/sec	 accuracy=76.178066	 loss=0.952954	 lr=0.001000
Epoch[056] Batch [2699]/[3760]	Speed: 35.319067 samples/sec	 accuracy=76.181134	 loss=0.952809	 lr=0.001000
Epoch[056] Batch [2749]/[3760]	Speed: 35.636951 samples/sec	 accuracy=76.181250	 loss=0.952870	 lr=0.001000
Epoch[056] Batch [2799]/[3760]	Speed: 35.336010 samples/sec	 accuracy=76.170759	 loss=0.953052	 lr=0.001000
Epoch[056] Batch [2849]/[3760]	Speed: 35.510822 samples/sec	 accuracy=76.180373	 loss=0.952943	 lr=0.001000
Epoch[056] Batch [2899]/[3760]	Speed: 35.141307 samples/sec	 accuracy=76.169181	 loss=0.953329	 lr=0.001000
Epoch[056] Batch [2949]/[3760]	Speed: 35.519284 samples/sec	 accuracy=76.145657	 loss=0.954311	 lr=0.001000
Epoch[056] Batch [2999]/[3760]	Speed: 35.409590 samples/sec	 accuracy=76.118229	 loss=0.955488	 lr=0.001000
Epoch[056] Batch [3049]/[3760]	Speed: 35.277574 samples/sec	 accuracy=76.135758	 loss=0.954884	 lr=0.001000
Epoch[056] Batch [3099]/[3760]	Speed: 35.340787 samples/sec	 accuracy=76.152218	 loss=0.954655	 lr=0.001000
Epoch[056] Batch [3149]/[3760]	Speed: 35.180007 samples/sec	 accuracy=76.145833	 loss=0.954852	 lr=0.001000
Epoch[056] Batch [3199]/[3760]	Speed: 35.529268 samples/sec	 accuracy=76.148438	 loss=0.955050	 lr=0.001000
Epoch[056] Batch [3249]/[3760]	Speed: 35.303846 samples/sec	 accuracy=76.156250	 loss=0.955103	 lr=0.001000
Epoch[056] Batch [3299]/[3760]	Speed: 35.511493 samples/sec	 accuracy=76.152462	 loss=0.955677	 lr=0.001000
Epoch[056] Batch [3349]/[3760]	Speed: 35.250936 samples/sec	 accuracy=76.152519	 loss=0.955588	 lr=0.001000
Epoch[056] Batch [3399]/[3760]	Speed: 35.602786 samples/sec	 accuracy=76.164522	 loss=0.955315	 lr=0.001000
Epoch[056] Batch [3449]/[3760]	Speed: 35.322113 samples/sec	 accuracy=76.179801	 loss=0.954832	 lr=0.001000
Epoch[056] Batch [3499]/[3760]	Speed: 35.645869 samples/sec	 accuracy=76.174554	 loss=0.954906	 lr=0.001000
Epoch[056] Batch [3549]/[3760]	Speed: 35.356020 samples/sec	 accuracy=76.174296	 loss=0.955191	 lr=0.001000
Epoch[056] Batch [3599]/[3760]	Speed: 35.273432 samples/sec	 accuracy=76.163628	 loss=0.955406	 lr=0.001000
Epoch[056] Batch [3649]/[3760]	Speed: 35.361393 samples/sec	 accuracy=76.157962	 loss=0.955593	 lr=0.001000
Epoch[056] Batch [3699]/[3760]	Speed: 35.355543 samples/sec	 accuracy=76.151182	 loss=0.955519	 lr=0.001000
Epoch[056] Batch [3749]/[3760]	Speed: 36.605966 samples/sec	 accuracy=76.154583	 loss=0.955109	 lr=0.001000
Batch [0049]/[0303]: acc-top1=67.343750 acc-top5=85.593750
Batch [0099]/[0303]: acc-top1=66.593750 acc-top5=85.296875
Batch [0149]/[0303]: acc-top1=66.864583 acc-top5=85.531250
Batch [0199]/[0303]: acc-top1=66.453125 acc-top5=85.765625
Batch [0249]/[0303]: acc-top1=66.331250 acc-top5=85.587500
Batch [0299]/[0303]: acc-top1=66.276042 acc-top5=85.583333
[Epoch 056] training: accuracy=76.148188	 loss=0.955427
[Epoch 056] speed: 35 samples/sec	time cost: 7309.253211
[Epoch 056] validation: acc-top1=66.264439 acc-top5=85.591997 loss=1.595150
Epoch[057] Batch [0049]/[3760]	Speed: 24.718411 samples/sec	 accuracy=77.312500	 loss=0.926548	 lr=0.001000
Epoch[057] Batch [0099]/[3760]	Speed: 35.262522 samples/sec	 accuracy=76.750000	 loss=0.949757	 lr=0.001000
Epoch[057] Batch [0149]/[3760]	Speed: 35.003512 samples/sec	 accuracy=77.166667	 loss=0.935888	 lr=0.001000
Epoch[057] Batch [0199]/[3760]	Speed: 35.214423 samples/sec	 accuracy=77.289062	 loss=0.923266	 lr=0.001000
Epoch[057] Batch [0249]/[3760]	Speed: 35.143888 samples/sec	 accuracy=77.218750	 loss=0.922425	 lr=0.001000
Epoch[057] Batch [0299]/[3760]	Speed: 35.211142 samples/sec	 accuracy=77.286458	 loss=0.917833	 lr=0.001000
Epoch[057] Batch [0349]/[3760]	Speed: 35.463539 samples/sec	 accuracy=77.125000	 loss=0.925388	 lr=0.001000
Epoch[057] Batch [0399]/[3760]	Speed: 35.005903 samples/sec	 accuracy=77.066406	 loss=0.928179	 lr=0.001000
Epoch[057] Batch [0449]/[3760]	Speed: 35.360932 samples/sec	 accuracy=77.000000	 loss=0.927973	 lr=0.001000
Epoch[057] Batch [0499]/[3760]	Speed: 35.049906 samples/sec	 accuracy=76.900000	 loss=0.930678	 lr=0.001000
Epoch[057] Batch [0549]/[3760]	Speed: 35.429120 samples/sec	 accuracy=76.894886	 loss=0.929972	 lr=0.001000
Epoch[057] Batch [0599]/[3760]	Speed: 35.230956 samples/sec	 accuracy=76.809896	 loss=0.933468	 lr=0.001000
Epoch[057] Batch [0649]/[3760]	Speed: 35.529963 samples/sec	 accuracy=76.790865	 loss=0.934403	 lr=0.001000
Epoch[057] Batch [0699]/[3760]	Speed: 35.425828 samples/sec	 accuracy=76.691964	 loss=0.938572	 lr=0.001000
Epoch[057] Batch [0749]/[3760]	Speed: 35.363866 samples/sec	 accuracy=76.662500	 loss=0.938986	 lr=0.001000
Epoch[057] Batch [0799]/[3760]	Speed: 35.141860 samples/sec	 accuracy=76.603516	 loss=0.940651	 lr=0.001000
Epoch[057] Batch [0849]/[3760]	Speed: 35.232171 samples/sec	 accuracy=76.584559	 loss=0.940940	 lr=0.001000
Epoch[057] Batch [0899]/[3760]	Speed: 35.129466 samples/sec	 accuracy=76.541667	 loss=0.942774	 lr=0.001000
Epoch[057] Batch [0949]/[3760]	Speed: 35.265289 samples/sec	 accuracy=76.508224	 loss=0.943729	 lr=0.001000
Epoch[057] Batch [0999]/[3760]	Speed: 35.118153 samples/sec	 accuracy=76.482812	 loss=0.945651	 lr=0.001000
Epoch[057] Batch [1049]/[3760]	Speed: 35.376995 samples/sec	 accuracy=76.517857	 loss=0.945477	 lr=0.001000
Epoch[057] Batch [1099]/[3760]	Speed: 35.117585 samples/sec	 accuracy=76.473011	 loss=0.947990	 lr=0.001000
Epoch[057] Batch [1149]/[3760]	Speed: 35.235837 samples/sec	 accuracy=76.388587	 loss=0.950867	 lr=0.001000
Epoch[057] Batch [1199]/[3760]	Speed: 35.225479 samples/sec	 accuracy=76.401042	 loss=0.951034	 lr=0.001000
Epoch[057] Batch [1249]/[3760]	Speed: 35.242203 samples/sec	 accuracy=76.408750	 loss=0.951226	 lr=0.001000
Epoch[057] Batch [1299]/[3760]	Speed: 35.403730 samples/sec	 accuracy=76.433894	 loss=0.950408	 lr=0.001000
Epoch[057] Batch [1349]/[3760]	Speed: 35.083902 samples/sec	 accuracy=76.480324	 loss=0.949052	 lr=0.001000
Epoch[057] Batch [1399]/[3760]	Speed: 35.474531 samples/sec	 accuracy=76.445312	 loss=0.950715	 lr=0.001000
Epoch[057] Batch [1449]/[3760]	Speed: 35.039060 samples/sec	 accuracy=76.438578	 loss=0.950566	 lr=0.001000
Epoch[057] Batch [1499]/[3760]	Speed: 35.401163 samples/sec	 accuracy=76.409375	 loss=0.951451	 lr=0.001000
Epoch[057] Batch [1549]/[3760]	Speed: 35.213188 samples/sec	 accuracy=76.386089	 loss=0.951400	 lr=0.001000
Epoch[057] Batch [1599]/[3760]	Speed: 35.234525 samples/sec	 accuracy=76.365234	 loss=0.951584	 lr=0.001000
Epoch[057] Batch [1649]/[3760]	Speed: 35.307492 samples/sec	 accuracy=76.366477	 loss=0.952126	 lr=0.001000
Epoch[057] Batch [1699]/[3760]	Speed: 35.152597 samples/sec	 accuracy=76.354779	 loss=0.951822	 lr=0.001000
Epoch[057] Batch [1749]/[3760]	Speed: 35.104338 samples/sec	 accuracy=76.358929	 loss=0.952189	 lr=0.001000
Epoch[057] Batch [1799]/[3760]	Speed: 35.122406 samples/sec	 accuracy=76.382812	 loss=0.951527	 lr=0.001000
Epoch[057] Batch [1849]/[3760]	Speed: 34.827425 samples/sec	 accuracy=76.359797	 loss=0.952142	 lr=0.001000
Epoch[057] Batch [1899]/[3760]	Speed: 35.119652 samples/sec	 accuracy=76.339638	 loss=0.952794	 lr=0.001000
Epoch[057] Batch [1949]/[3760]	Speed: 34.699963 samples/sec	 accuracy=76.361378	 loss=0.952039	 lr=0.001000
Epoch[057] Batch [1999]/[3760]	Speed: 35.249397 samples/sec	 accuracy=76.344531	 loss=0.953904	 lr=0.001000
Epoch[057] Batch [2049]/[3760]	Speed: 34.784287 samples/sec	 accuracy=76.361280	 loss=0.952714	 lr=0.001000
Epoch[057] Batch [2099]/[3760]	Speed: 34.950541 samples/sec	 accuracy=76.360863	 loss=0.953015	 lr=0.001000
Epoch[057] Batch [2149]/[3760]	Speed: 34.775058 samples/sec	 accuracy=76.376453	 loss=0.952414	 lr=0.001000
Epoch[057] Batch [2199]/[3760]	Speed: 34.741169 samples/sec	 accuracy=76.356534	 loss=0.952975	 lr=0.001000
Epoch[057] Batch [2249]/[3760]	Speed: 35.201863 samples/sec	 accuracy=76.365278	 loss=0.952737	 lr=0.001000
Epoch[057] Batch [2299]/[3760]	Speed: 34.497569 samples/sec	 accuracy=76.362092	 loss=0.953076	 lr=0.001000
Epoch[057] Batch [2349]/[3760]	Speed: 35.329596 samples/sec	 accuracy=76.320479	 loss=0.954265	 lr=0.001000
Epoch[057] Batch [2399]/[3760]	Speed: 34.476518 samples/sec	 accuracy=76.300781	 loss=0.955213	 lr=0.001000
Epoch[057] Batch [2449]/[3760]	Speed: 35.224131 samples/sec	 accuracy=76.294643	 loss=0.955010	 lr=0.001000
Epoch[057] Batch [2499]/[3760]	Speed: 34.681533 samples/sec	 accuracy=76.291250	 loss=0.955007	 lr=0.001000
Epoch[057] Batch [2549]/[3760]	Speed: 35.285444 samples/sec	 accuracy=76.245711	 loss=0.956487	 lr=0.001000
Epoch[057] Batch [2599]/[3760]	Speed: 34.689750 samples/sec	 accuracy=76.236178	 loss=0.956920	 lr=0.001000
Epoch[057] Batch [2649]/[3760]	Speed: 34.922183 samples/sec	 accuracy=76.236439	 loss=0.956698	 lr=0.001000
Epoch[057] Batch [2699]/[3760]	Speed: 35.214522 samples/sec	 accuracy=76.256944	 loss=0.955896	 lr=0.001000
Epoch[057] Batch [2749]/[3760]	Speed: 34.639323 samples/sec	 accuracy=76.248295	 loss=0.956330	 lr=0.001000
Epoch[057] Batch [2799]/[3760]	Speed: 35.085977 samples/sec	 accuracy=76.267299	 loss=0.955281	 lr=0.001000
Epoch[057] Batch [2849]/[3760]	Speed: 34.720688 samples/sec	 accuracy=76.271930	 loss=0.954821	 lr=0.001000
Epoch[057] Batch [2899]/[3760]	Speed: 35.412039 samples/sec	 accuracy=76.282328	 loss=0.954550	 lr=0.001000
Epoch[057] Batch [2949]/[3760]	Speed: 34.858420 samples/sec	 accuracy=76.261653	 loss=0.955191	 lr=0.001000
Epoch[057] Batch [2999]/[3760]	Speed: 35.497059 samples/sec	 accuracy=76.267188	 loss=0.954940	 lr=0.001000
Epoch[057] Batch [3049]/[3760]	Speed: 34.904211 samples/sec	 accuracy=76.248463	 loss=0.955442	 lr=0.001000
Epoch[057] Batch [3099]/[3760]	Speed: 35.281131 samples/sec	 accuracy=76.256048	 loss=0.955230	 lr=0.001000
Epoch[057] Batch [3149]/[3760]	Speed: 34.908521 samples/sec	 accuracy=76.266865	 loss=0.955273	 lr=0.001000
Epoch[057] Batch [3199]/[3760]	Speed: 35.221587 samples/sec	 accuracy=76.244141	 loss=0.955654	 lr=0.001000
Epoch[057] Batch [3249]/[3760]	Speed: 34.845749 samples/sec	 accuracy=76.245673	 loss=0.955584	 lr=0.001000
Epoch[057] Batch [3299]/[3760]	Speed: 35.255078 samples/sec	 accuracy=76.236742	 loss=0.955830	 lr=0.001000
Epoch[057] Batch [3349]/[3760]	Speed: 35.169684 samples/sec	 accuracy=76.228545	 loss=0.956246	 lr=0.001000
Epoch[057] Batch [3399]/[3760]	Speed: 35.280711 samples/sec	 accuracy=76.228860	 loss=0.956444	 lr=0.001000
Epoch[057] Batch [3449]/[3760]	Speed: 35.341164 samples/sec	 accuracy=76.220109	 loss=0.956656	 lr=0.001000
Epoch[057] Batch [3499]/[3760]	Speed: 35.430201 samples/sec	 accuracy=76.210268	 loss=0.956834	 lr=0.001000
Epoch[057] Batch [3549]/[3760]	Speed: 35.086498 samples/sec	 accuracy=76.188380	 loss=0.957206	 lr=0.001000
Epoch[057] Batch [3599]/[3760]	Speed: 35.475168 samples/sec	 accuracy=76.202691	 loss=0.956799	 lr=0.001000
Epoch[057] Batch [3649]/[3760]	Speed: 35.181735 samples/sec	 accuracy=76.225171	 loss=0.955639	 lr=0.001000
Epoch[057] Batch [3699]/[3760]	Speed: 35.147384 samples/sec	 accuracy=76.210304	 loss=0.956119	 lr=0.001000
Epoch[057] Batch [3749]/[3760]	Speed: 36.552644 samples/sec	 accuracy=76.213750	 loss=0.956225	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.437500 acc-top5=86.250000
Batch [0099]/[0303]: acc-top1=66.171875 acc-top5=85.250000
Batch [0149]/[0303]: acc-top1=66.770833 acc-top5=85.604167
Batch [0199]/[0303]: acc-top1=66.351562 acc-top5=85.679688
Batch [0249]/[0303]: acc-top1=66.243750 acc-top5=85.518750
Batch [0299]/[0303]: acc-top1=66.088542 acc-top5=85.619792
[Epoch 057] training: accuracy=76.212600	 loss=0.956334
[Epoch 057] speed: 34 samples/sec	time cost: 7345.899586
[Epoch 057] validation: acc-top1=66.099422 acc-top5=85.617781 loss=1.597277
Epoch[058] Batch [0049]/[3759]	Speed: 24.855564 samples/sec	 accuracy=76.937500	 loss=0.925939	 lr=0.001000
Epoch[058] Batch [0099]/[3759]	Speed: 35.345082 samples/sec	 accuracy=76.734375	 loss=0.930473	 lr=0.001000
Epoch[058] Batch [0149]/[3759]	Speed: 35.171432 samples/sec	 accuracy=76.635417	 loss=0.940027	 lr=0.001000
Epoch[058] Batch [0199]/[3759]	Speed: 35.438916 samples/sec	 accuracy=76.859375	 loss=0.931023	 lr=0.001000
Epoch[058] Batch [0249]/[3759]	Speed: 35.385893 samples/sec	 accuracy=76.618750	 loss=0.943161	 lr=0.001000
Epoch[058] Batch [0299]/[3759]	Speed: 35.407726 samples/sec	 accuracy=76.515625	 loss=0.943639	 lr=0.001000
Epoch[058] Batch [0349]/[3759]	Speed: 35.292866 samples/sec	 accuracy=76.366071	 loss=0.947695	 lr=0.001000
Epoch[058] Batch [0399]/[3759]	Speed: 35.339737 samples/sec	 accuracy=76.429688	 loss=0.945268	 lr=0.001000
Epoch[058] Batch [0449]/[3759]	Speed: 35.061860 samples/sec	 accuracy=76.375000	 loss=0.950802	 lr=0.001000
Epoch[058] Batch [0499]/[3759]	Speed: 35.480577 samples/sec	 accuracy=76.353125	 loss=0.949212	 lr=0.001000
Epoch[058] Batch [0549]/[3759]	Speed: 35.186103 samples/sec	 accuracy=76.434659	 loss=0.945410	 lr=0.001000
Epoch[058] Batch [0599]/[3759]	Speed: 35.249613 samples/sec	 accuracy=76.309896	 loss=0.949618	 lr=0.001000
Epoch[058] Batch [0649]/[3759]	Speed: 35.244949 samples/sec	 accuracy=76.242788	 loss=0.950089	 lr=0.001000
Epoch[058] Batch [0699]/[3759]	Speed: 35.323817 samples/sec	 accuracy=76.165179	 loss=0.953720	 lr=0.001000
Epoch[058] Batch [0749]/[3759]	Speed: 35.427689 samples/sec	 accuracy=76.116667	 loss=0.953646	 lr=0.001000
Epoch[058] Batch [0799]/[3759]	Speed: 35.213084 samples/sec	 accuracy=76.125000	 loss=0.951817	 lr=0.001000
Epoch[058] Batch [0849]/[3759]	Speed: 35.495602 samples/sec	 accuracy=76.147059	 loss=0.950369	 lr=0.001000
Epoch[058] Batch [0899]/[3759]	Speed: 35.444412 samples/sec	 accuracy=76.182292	 loss=0.949616	 lr=0.001000
Epoch[058] Batch [0949]/[3759]	Speed: 35.296684 samples/sec	 accuracy=76.098684	 loss=0.951787	 lr=0.001000
Epoch[058] Batch [0999]/[3759]	Speed: 35.372122 samples/sec	 accuracy=76.215625	 loss=0.948115	 lr=0.001000
Epoch[058] Batch [1049]/[3759]	Speed: 35.391194 samples/sec	 accuracy=76.227679	 loss=0.948029	 lr=0.001000
Epoch[058] Batch [1099]/[3759]	Speed: 35.428027 samples/sec	 accuracy=76.265625	 loss=0.948134	 lr=0.001000
Epoch[058] Batch [1149]/[3759]	Speed: 35.376044 samples/sec	 accuracy=76.254076	 loss=0.948043	 lr=0.001000
Epoch[058] Batch [1199]/[3759]	Speed: 35.329475 samples/sec	 accuracy=76.305990	 loss=0.947384	 lr=0.001000
Epoch[058] Batch [1249]/[3759]	Speed: 35.323152 samples/sec	 accuracy=76.288750	 loss=0.947657	 lr=0.001000
Epoch[058] Batch [1299]/[3759]	Speed: 35.479277 samples/sec	 accuracy=76.265625	 loss=0.948670	 lr=0.001000
Epoch[058] Batch [1349]/[3759]	Speed: 35.133810 samples/sec	 accuracy=76.251157	 loss=0.948523	 lr=0.001000
Epoch[058] Batch [1399]/[3759]	Speed: 35.428244 samples/sec	 accuracy=76.290179	 loss=0.946425	 lr=0.001000
Epoch[058] Batch [1449]/[3759]	Speed: 34.728802 samples/sec	 accuracy=76.324353	 loss=0.944854	 lr=0.001000
Epoch[058] Batch [1499]/[3759]	Speed: 35.276680 samples/sec	 accuracy=76.360417	 loss=0.942623	 lr=0.001000
Epoch[058] Batch [1549]/[3759]	Speed: 34.851077 samples/sec	 accuracy=76.380040	 loss=0.942103	 lr=0.001000
Epoch[058] Batch [1599]/[3759]	Speed: 35.259595 samples/sec	 accuracy=76.356445	 loss=0.942198	 lr=0.001000
Epoch[058] Batch [1649]/[3759]	Speed: 34.744002 samples/sec	 accuracy=76.361742	 loss=0.942586	 lr=0.001000
Epoch[058] Batch [1699]/[3759]	Speed: 35.142694 samples/sec	 accuracy=76.337316	 loss=0.942281	 lr=0.001000
Epoch[058] Batch [1749]/[3759]	Speed: 34.526849 samples/sec	 accuracy=76.302679	 loss=0.943090	 lr=0.001000
Epoch[058] Batch [1799]/[3759]	Speed: 34.922933 samples/sec	 accuracy=76.296007	 loss=0.943669	 lr=0.001000
Epoch[058] Batch [1849]/[3759]	Speed: 35.145277 samples/sec	 accuracy=76.306588	 loss=0.942528	 lr=0.001000
Epoch[058] Batch [1899]/[3759]	Speed: 34.942080 samples/sec	 accuracy=76.282072	 loss=0.943604	 lr=0.001000
Epoch[058] Batch [1949]/[3759]	Speed: 35.345679 samples/sec	 accuracy=76.273237	 loss=0.944182	 lr=0.001000
Epoch[058] Batch [1999]/[3759]	Speed: 35.049247 samples/sec	 accuracy=76.258594	 loss=0.944973	 lr=0.001000
Epoch[058] Batch [2049]/[3759]	Speed: 35.316146 samples/sec	 accuracy=76.223323	 loss=0.946230	 lr=0.001000
Epoch[058] Batch [2099]/[3759]	Speed: 35.127393 samples/sec	 accuracy=76.237351	 loss=0.945137	 lr=0.001000
Epoch[058] Batch [2149]/[3759]	Speed: 35.409446 samples/sec	 accuracy=76.247820	 loss=0.945266	 lr=0.001000
Epoch[058] Batch [2199]/[3759]	Speed: 34.998937 samples/sec	 accuracy=76.234375	 loss=0.945771	 lr=0.001000
Epoch[058] Batch [2249]/[3759]	Speed: 35.355716 samples/sec	 accuracy=76.211111	 loss=0.947171	 lr=0.001000
Epoch[058] Batch [2299]/[3759]	Speed: 34.956069 samples/sec	 accuracy=76.230299	 loss=0.946287	 lr=0.001000
Epoch[058] Batch [2349]/[3759]	Speed: 35.237832 samples/sec	 accuracy=76.250665	 loss=0.945612	 lr=0.001000
Epoch[058] Batch [2399]/[3759]	Speed: 35.193370 samples/sec	 accuracy=76.228516	 loss=0.947015	 lr=0.001000
Epoch[058] Batch [2449]/[3759]	Speed: 35.051629 samples/sec	 accuracy=76.237245	 loss=0.946832	 lr=0.001000
Epoch[058] Batch [2499]/[3759]	Speed: 35.302372 samples/sec	 accuracy=76.250625	 loss=0.946294	 lr=0.001000
Epoch[058] Batch [2549]/[3759]	Speed: 35.184704 samples/sec	 accuracy=76.238358	 loss=0.946867	 lr=0.001000
Epoch[058] Batch [2599]/[3759]	Speed: 35.393316 samples/sec	 accuracy=76.228966	 loss=0.947555	 lr=0.001000
Epoch[058] Batch [2649]/[3759]	Speed: 35.036849 samples/sec	 accuracy=76.245283	 loss=0.946724	 lr=0.001000
Epoch[058] Batch [2699]/[3759]	Speed: 35.356467 samples/sec	 accuracy=76.238426	 loss=0.946854	 lr=0.001000
Epoch[058] Batch [2749]/[3759]	Speed: 35.087547 samples/sec	 accuracy=76.241477	 loss=0.946666	 lr=0.001000
Epoch[058] Batch [2799]/[3759]	Speed: 35.352058 samples/sec	 accuracy=76.221540	 loss=0.948031	 lr=0.001000
Epoch[058] Batch [2849]/[3759]	Speed: 35.213623 samples/sec	 accuracy=76.237939	 loss=0.947892	 lr=0.001000
Epoch[058] Batch [2899]/[3759]	Speed: 35.343680 samples/sec	 accuracy=76.231681	 loss=0.948163	 lr=0.001000
Epoch[058] Batch [2949]/[3759]	Speed: 35.042661 samples/sec	 accuracy=76.223517	 loss=0.948437	 lr=0.001000
Epoch[058] Batch [2999]/[3759]	Speed: 35.290283 samples/sec	 accuracy=76.225521	 loss=0.948360	 lr=0.001000
Epoch[058] Batch [3049]/[3759]	Speed: 34.886969 samples/sec	 accuracy=76.221824	 loss=0.948405	 lr=0.001000
Epoch[058] Batch [3099]/[3759]	Speed: 35.045802 samples/sec	 accuracy=76.217742	 loss=0.948763	 lr=0.001000
Epoch[058] Batch [3149]/[3759]	Speed: 34.805095 samples/sec	 accuracy=76.206845	 loss=0.949555	 lr=0.001000
Epoch[058] Batch [3199]/[3759]	Speed: 34.890287 samples/sec	 accuracy=76.195312	 loss=0.950055	 lr=0.001000
Epoch[058] Batch [3249]/[3759]	Speed: 34.987758 samples/sec	 accuracy=76.183173	 loss=0.950301	 lr=0.001000
Epoch[058] Batch [3299]/[3759]	Speed: 34.416213 samples/sec	 accuracy=76.192235	 loss=0.950557	 lr=0.001000
Epoch[058] Batch [3349]/[3759]	Speed: 35.194500 samples/sec	 accuracy=76.198694	 loss=0.950674	 lr=0.001000
Epoch[058] Batch [3399]/[3759]	Speed: 34.304623 samples/sec	 accuracy=76.204044	 loss=0.950544	 lr=0.001000
Epoch[058] Batch [3449]/[3759]	Speed: 35.303606 samples/sec	 accuracy=76.212409	 loss=0.949741	 lr=0.001000
Epoch[058] Batch [3499]/[3759]	Speed: 34.703314 samples/sec	 accuracy=76.213839	 loss=0.949681	 lr=0.001000
Epoch[058] Batch [3549]/[3759]	Speed: 35.178610 samples/sec	 accuracy=76.218310	 loss=0.949494	 lr=0.001000
Epoch[058] Batch [3599]/[3759]	Speed: 35.033499 samples/sec	 accuracy=76.213108	 loss=0.949987	 lr=0.001000
Epoch[058] Batch [3649]/[3759]	Speed: 34.939481 samples/sec	 accuracy=76.211901	 loss=0.950058	 lr=0.001000
Epoch[058] Batch [3699]/[3759]	Speed: 35.179552 samples/sec	 accuracy=76.211571	 loss=0.950034	 lr=0.001000
Epoch[058] Batch [3749]/[3759]	Speed: 36.303097 samples/sec	 accuracy=76.218750	 loss=0.949896	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.750000 acc-top5=85.843750
Batch [0099]/[0303]: acc-top1=66.265625 acc-top5=85.453125
Batch [0149]/[0303]: acc-top1=66.656250 acc-top5=85.708333
Batch [0199]/[0303]: acc-top1=66.156250 acc-top5=85.765625
Batch [0249]/[0303]: acc-top1=65.943750 acc-top5=85.600000
Batch [0299]/[0303]: acc-top1=65.869792 acc-top5=85.609375
[Epoch 058] training: accuracy=76.219989	 loss=0.949841
[Epoch 058] speed: 34 samples/sec	time cost: 7339.148417
[Epoch 058] validation: acc-top1=65.846741 acc-top5=85.612624 loss=1.604402
Epoch[059] Batch [0049]/[3760]	Speed: 24.723206 samples/sec	 accuracy=75.781250	 loss=0.962573	 lr=0.001000
Epoch[059] Batch [0099]/[3760]	Speed: 35.146913 samples/sec	 accuracy=76.703125	 loss=0.942364	 lr=0.001000
Epoch[059] Batch [0149]/[3760]	Speed: 34.954088 samples/sec	 accuracy=76.645833	 loss=0.938896	 lr=0.001000
Epoch[059] Batch [0199]/[3760]	Speed: 35.052113 samples/sec	 accuracy=76.828125	 loss=0.937244	 lr=0.001000
Epoch[059] Batch [0249]/[3760]	Speed: 34.992855 samples/sec	 accuracy=76.825000	 loss=0.928505	 lr=0.001000
Epoch[059] Batch [0299]/[3760]	Speed: 35.069190 samples/sec	 accuracy=76.755208	 loss=0.928188	 lr=0.001000
Epoch[059] Batch [0349]/[3760]	Speed: 35.047344 samples/sec	 accuracy=76.732143	 loss=0.931911	 lr=0.001000
Epoch[059] Batch [0399]/[3760]	Speed: 34.895444 samples/sec	 accuracy=76.703125	 loss=0.930496	 lr=0.001000
Epoch[059] Batch [0449]/[3760]	Speed: 35.102820 samples/sec	 accuracy=76.701389	 loss=0.932233	 lr=0.001000
Epoch[059] Batch [0499]/[3760]	Speed: 34.769293 samples/sec	 accuracy=76.721875	 loss=0.932339	 lr=0.001000
Epoch[059] Batch [0549]/[3760]	Speed: 35.225550 samples/sec	 accuracy=76.755682	 loss=0.933443	 lr=0.001000
Epoch[059] Batch [0599]/[3760]	Speed: 34.695349 samples/sec	 accuracy=76.736979	 loss=0.933096	 lr=0.001000
Epoch[059] Batch [0649]/[3760]	Speed: 35.409863 samples/sec	 accuracy=76.723558	 loss=0.934045	 lr=0.001000
Epoch[059] Batch [0699]/[3760]	Speed: 34.907259 samples/sec	 accuracy=76.765625	 loss=0.932262	 lr=0.001000
Epoch[059] Batch [0749]/[3760]	Speed: 35.436486 samples/sec	 accuracy=76.797917	 loss=0.930021	 lr=0.001000
Epoch[059] Batch [0799]/[3760]	Speed: 35.238787 samples/sec	 accuracy=76.869141	 loss=0.927413	 lr=0.001000
Epoch[059] Batch [0849]/[3760]	Speed: 35.389436 samples/sec	 accuracy=76.810662	 loss=0.929184	 lr=0.001000
Epoch[059] Batch [0899]/[3760]	Speed: 35.350108 samples/sec	 accuracy=76.763889	 loss=0.931426	 lr=0.001000
Epoch[059] Batch [0949]/[3760]	Speed: 35.195059 samples/sec	 accuracy=76.822368	 loss=0.928694	 lr=0.001000
Epoch[059] Batch [0999]/[3760]	Speed: 35.296109 samples/sec	 accuracy=76.823437	 loss=0.929785	 lr=0.001000
Epoch[059] Batch [1049]/[3760]	Speed: 35.335813 samples/sec	 accuracy=76.816964	 loss=0.930232	 lr=0.001000
Epoch[059] Batch [1099]/[3760]	Speed: 35.301078 samples/sec	 accuracy=76.759943	 loss=0.931189	 lr=0.001000
Epoch[059] Batch [1149]/[3760]	Speed: 35.425455 samples/sec	 accuracy=76.740489	 loss=0.931921	 lr=0.001000
Epoch[059] Batch [1199]/[3760]	Speed: 35.260496 samples/sec	 accuracy=76.675781	 loss=0.934334	 lr=0.001000
Epoch[059] Batch [1249]/[3760]	Speed: 35.374416 samples/sec	 accuracy=76.670000	 loss=0.933424	 lr=0.001000
Epoch[059] Batch [1299]/[3760]	Speed: 35.127997 samples/sec	 accuracy=76.626202	 loss=0.936128	 lr=0.001000
Epoch[059] Batch [1349]/[3760]	Speed: 35.374736 samples/sec	 accuracy=76.664352	 loss=0.934901	 lr=0.001000
Epoch[059] Batch [1399]/[3760]	Speed: 35.337753 samples/sec	 accuracy=76.607143	 loss=0.936722	 lr=0.001000
Epoch[059] Batch [1449]/[3760]	Speed: 35.200780 samples/sec	 accuracy=76.625000	 loss=0.936475	 lr=0.001000
Epoch[059] Batch [1499]/[3760]	Speed: 35.244450 samples/sec	 accuracy=76.618750	 loss=0.937247	 lr=0.001000
Epoch[059] Batch [1549]/[3760]	Speed: 35.245632 samples/sec	 accuracy=76.652218	 loss=0.935513	 lr=0.001000
Epoch[059] Batch [1599]/[3760]	Speed: 35.420444 samples/sec	 accuracy=76.674805	 loss=0.934639	 lr=0.001000
Epoch[059] Batch [1649]/[3760]	Speed: 35.163806 samples/sec	 accuracy=76.672348	 loss=0.934390	 lr=0.001000
Epoch[059] Batch [1699]/[3760]	Speed: 35.546602 samples/sec	 accuracy=76.658088	 loss=0.935370	 lr=0.001000
Epoch[059] Batch [1749]/[3760]	Speed: 35.227708 samples/sec	 accuracy=76.642857	 loss=0.935411	 lr=0.001000
Epoch[059] Batch [1799]/[3760]	Speed: 35.392910 samples/sec	 accuracy=76.636285	 loss=0.935307	 lr=0.001000
Epoch[059] Batch [1849]/[3760]	Speed: 35.046649 samples/sec	 accuracy=76.625845	 loss=0.935552	 lr=0.001000
Epoch[059] Batch [1899]/[3760]	Speed: 35.441693 samples/sec	 accuracy=76.614309	 loss=0.935780	 lr=0.001000
Epoch[059] Batch [1949]/[3760]	Speed: 34.785872 samples/sec	 accuracy=76.618590	 loss=0.935426	 lr=0.001000
Epoch[059] Batch [1999]/[3760]	Speed: 35.540960 samples/sec	 accuracy=76.617969	 loss=0.935373	 lr=0.001000
Epoch[059] Batch [2049]/[3760]	Speed: 34.881575 samples/sec	 accuracy=76.632622	 loss=0.934937	 lr=0.001000
Epoch[059] Batch [2099]/[3760]	Speed: 35.412077 samples/sec	 accuracy=76.607143	 loss=0.935386	 lr=0.001000
Epoch[059] Batch [2149]/[3760]	Speed: 34.682571 samples/sec	 accuracy=76.624273	 loss=0.935071	 lr=0.001000
Epoch[059] Batch [2199]/[3760]	Speed: 35.079867 samples/sec	 accuracy=76.619318	 loss=0.935832	 lr=0.001000
Epoch[059] Batch [2249]/[3760]	Speed: 34.997628 samples/sec	 accuracy=76.617361	 loss=0.935645	 lr=0.001000
Epoch[059] Batch [2299]/[3760]	Speed: 35.006784 samples/sec	 accuracy=76.625000	 loss=0.935746	 lr=0.001000
Epoch[059] Batch [2349]/[3760]	Speed: 35.206287 samples/sec	 accuracy=76.623005	 loss=0.935612	 lr=0.001000
Epoch[059] Batch [2399]/[3760]	Speed: 34.609506 samples/sec	 accuracy=76.603516	 loss=0.936286	 lr=0.001000
Epoch[059] Batch [2449]/[3760]	Speed: 35.233881 samples/sec	 accuracy=76.626913	 loss=0.936011	 lr=0.001000
Epoch[059] Batch [2499]/[3760]	Speed: 34.438147 samples/sec	 accuracy=76.630000	 loss=0.935933	 lr=0.001000
Epoch[059] Batch [2549]/[3760]	Speed: 35.492717 samples/sec	 accuracy=76.639093	 loss=0.935697	 lr=0.001000
Epoch[059] Batch [2599]/[3760]	Speed: 34.767292 samples/sec	 accuracy=76.626803	 loss=0.935462	 lr=0.001000
Epoch[059] Batch [2649]/[3760]	Speed: 35.519034 samples/sec	 accuracy=76.593750	 loss=0.936341	 lr=0.001000
Epoch[059] Batch [2699]/[3760]	Speed: 35.058783 samples/sec	 accuracy=76.581019	 loss=0.936608	 lr=0.001000
Epoch[059] Batch [2749]/[3760]	Speed: 35.232583 samples/sec	 accuracy=76.574432	 loss=0.936792	 lr=0.001000
Epoch[059] Batch [2799]/[3760]	Speed: 34.831595 samples/sec	 accuracy=76.578683	 loss=0.936838	 lr=0.001000
Epoch[059] Batch [2849]/[3760]	Speed: 34.869956 samples/sec	 accuracy=76.577303	 loss=0.936982	 lr=0.001000
Epoch[059] Batch [2899]/[3760]	Speed: 35.059615 samples/sec	 accuracy=76.567349	 loss=0.937473	 lr=0.001000
Epoch[059] Batch [2949]/[3760]	Speed: 34.887233 samples/sec	 accuracy=76.577860	 loss=0.937065	 lr=0.001000
Epoch[059] Batch [2999]/[3760]	Speed: 35.451953 samples/sec	 accuracy=76.564583	 loss=0.936989	 lr=0.001000
Epoch[059] Batch [3049]/[3760]	Speed: 34.733802 samples/sec	 accuracy=76.553279	 loss=0.937505	 lr=0.001000
Epoch[059] Batch [3099]/[3760]	Speed: 35.530140 samples/sec	 accuracy=76.552419	 loss=0.937413	 lr=0.001000
Epoch[059] Batch [3149]/[3760]	Speed: 34.653087 samples/sec	 accuracy=76.560020	 loss=0.936890	 lr=0.001000
Epoch[059] Batch [3199]/[3760]	Speed: 35.453119 samples/sec	 accuracy=76.537598	 loss=0.937618	 lr=0.001000
Epoch[059] Batch [3249]/[3760]	Speed: 34.686244 samples/sec	 accuracy=76.521154	 loss=0.937960	 lr=0.001000
Epoch[059] Batch [3299]/[3760]	Speed: 35.362163 samples/sec	 accuracy=76.528883	 loss=0.937468	 lr=0.001000
Epoch[059] Batch [3349]/[3760]	Speed: 35.081408 samples/sec	 accuracy=76.517724	 loss=0.938215	 lr=0.001000
Epoch[059] Batch [3399]/[3760]	Speed: 35.194301 samples/sec	 accuracy=76.514706	 loss=0.938945	 lr=0.001000
Epoch[059] Batch [3449]/[3760]	Speed: 35.294474 samples/sec	 accuracy=76.514040	 loss=0.939161	 lr=0.001000
Epoch[059] Batch [3499]/[3760]	Speed: 35.206138 samples/sec	 accuracy=76.516964	 loss=0.938792	 lr=0.001000
Epoch[059] Batch [3549]/[3760]	Speed: 35.590001 samples/sec	 accuracy=76.513204	 loss=0.938764	 lr=0.001000
Epoch[059] Batch [3599]/[3760]	Speed: 35.185807 samples/sec	 accuracy=76.532986	 loss=0.938666	 lr=0.001000
Epoch[059] Batch [3649]/[3760]	Speed: 35.192604 samples/sec	 accuracy=76.535103	 loss=0.938642	 lr=0.001000
Epoch[059] Batch [3699]/[3760]	Speed: 35.099761 samples/sec	 accuracy=76.537584	 loss=0.938445	 lr=0.001000
Epoch[059] Batch [3749]/[3760]	Speed: 36.585538 samples/sec	 accuracy=76.532917	 loss=0.938755	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.937500 acc-top5=85.875000
Batch [0099]/[0303]: acc-top1=65.921875 acc-top5=85.468750
Batch [0149]/[0303]: acc-top1=66.489583 acc-top5=85.697917
Batch [0199]/[0303]: acc-top1=66.179688 acc-top5=85.804688
Batch [0249]/[0303]: acc-top1=66.012500 acc-top5=85.643750
Batch [0299]/[0303]: acc-top1=65.937500 acc-top5=85.687500
[Epoch 059] training: accuracy=76.527593	 loss=0.938825
[Epoch 059] speed: 34 samples/sec	time cost: 7341.777294
[Epoch 059] validation: acc-top1=65.944719 acc-top5=85.695132 loss=1.606369
Epoch[060] Batch [0049]/[3760]	Speed: 24.992904 samples/sec	 accuracy=75.593750	 loss=0.955203	 lr=0.001000
Epoch[060] Batch [0099]/[3760]	Speed: 35.207780 samples/sec	 accuracy=76.578125	 loss=0.941767	 lr=0.001000
Epoch[060] Batch [0149]/[3760]	Speed: 34.842223 samples/sec	 accuracy=76.604167	 loss=0.939834	 lr=0.001000
Epoch[060] Batch [0199]/[3760]	Speed: 35.179150 samples/sec	 accuracy=76.742188	 loss=0.930946	 lr=0.001000
Epoch[060] Batch [0249]/[3760]	Speed: 35.232535 samples/sec	 accuracy=76.668750	 loss=0.931198	 lr=0.001000
Epoch[060] Batch [0299]/[3760]	Speed: 35.352430 samples/sec	 accuracy=76.958333	 loss=0.923772	 lr=0.001000
Epoch[060] Batch [0349]/[3760]	Speed: 35.337836 samples/sec	 accuracy=76.816964	 loss=0.930571	 lr=0.001000
Epoch[060] Batch [0399]/[3760]	Speed: 35.290541 samples/sec	 accuracy=76.792969	 loss=0.929407	 lr=0.001000
Epoch[060] Batch [0449]/[3760]	Speed: 35.221093 samples/sec	 accuracy=76.888889	 loss=0.926378	 lr=0.001000
Epoch[060] Batch [0499]/[3760]	Speed: 35.355320 samples/sec	 accuracy=76.828125	 loss=0.928878	 lr=0.001000
Epoch[060] Batch [0549]/[3760]	Speed: 35.430741 samples/sec	 accuracy=76.795455	 loss=0.931711	 lr=0.001000
Epoch[060] Batch [0599]/[3760]	Speed: 35.156074 samples/sec	 accuracy=76.674479	 loss=0.934168	 lr=0.001000
Epoch[060] Batch [0649]/[3760]	Speed: 35.305491 samples/sec	 accuracy=76.694712	 loss=0.935264	 lr=0.001000
Epoch[060] Batch [0699]/[3760]	Speed: 35.302074 samples/sec	 accuracy=76.814732	 loss=0.932511	 lr=0.001000
Epoch[060] Batch [0749]/[3760]	Speed: 35.129075 samples/sec	 accuracy=76.835417	 loss=0.931960	 lr=0.001000
Epoch[060] Batch [0799]/[3760]	Speed: 35.372295 samples/sec	 accuracy=76.966797	 loss=0.926949	 lr=0.001000
Epoch[060] Batch [0849]/[3760]	Speed: 35.006647 samples/sec	 accuracy=76.930147	 loss=0.927243	 lr=0.001000
Epoch[060] Batch [0899]/[3760]	Speed: 35.277201 samples/sec	 accuracy=76.928819	 loss=0.927496	 lr=0.001000
Epoch[060] Batch [0949]/[3760]	Speed: 34.823823 samples/sec	 accuracy=76.860197	 loss=0.928835	 lr=0.001000
Epoch[060] Batch [0999]/[3760]	Speed: 35.062644 samples/sec	 accuracy=76.896875	 loss=0.927200	 lr=0.001000
Epoch[060] Batch [1049]/[3760]	Speed: 35.181779 samples/sec	 accuracy=76.852679	 loss=0.929552	 lr=0.001000
Epoch[060] Batch [1099]/[3760]	Speed: 34.595271 samples/sec	 accuracy=76.795455	 loss=0.930451	 lr=0.001000
Epoch[060] Batch [1149]/[3760]	Speed: 35.348901 samples/sec	 accuracy=76.786685	 loss=0.930844	 lr=0.001000
Epoch[060] Batch [1199]/[3760]	Speed: 34.667796 samples/sec	 accuracy=76.779948	 loss=0.931078	 lr=0.001000
Epoch[060] Batch [1249]/[3760]	Speed: 35.530494 samples/sec	 accuracy=76.820000	 loss=0.929283	 lr=0.001000
Epoch[060] Batch [1299]/[3760]	Speed: 35.162051 samples/sec	 accuracy=76.830529	 loss=0.929100	 lr=0.001000
Epoch[060] Batch [1349]/[3760]	Speed: 35.353956 samples/sec	 accuracy=76.814815	 loss=0.929441	 lr=0.001000
Epoch[060] Batch [1399]/[3760]	Speed: 34.971405 samples/sec	 accuracy=76.799107	 loss=0.929267	 lr=0.001000
Epoch[060] Batch [1449]/[3760]	Speed: 35.474710 samples/sec	 accuracy=76.789871	 loss=0.929417	 lr=0.001000
Epoch[060] Batch [1499]/[3760]	Speed: 34.951381 samples/sec	 accuracy=76.820833	 loss=0.928032	 lr=0.001000
Epoch[060] Batch [1549]/[3760]	Speed: 35.342508 samples/sec	 accuracy=76.817540	 loss=0.927818	 lr=0.001000
Epoch[060] Batch [1599]/[3760]	Speed: 34.887115 samples/sec	 accuracy=76.805664	 loss=0.928334	 lr=0.001000
Epoch[060] Batch [1649]/[3760]	Speed: 35.204666 samples/sec	 accuracy=76.784091	 loss=0.928696	 lr=0.001000
Epoch[060] Batch [1699]/[3760]	Speed: 35.093115 samples/sec	 accuracy=76.738051	 loss=0.930759	 lr=0.001000
Epoch[060] Batch [1749]/[3760]	Speed: 35.121082 samples/sec	 accuracy=76.743750	 loss=0.930938	 lr=0.001000
Epoch[060] Batch [1799]/[3760]	Speed: 35.244033 samples/sec	 accuracy=76.725694	 loss=0.930980	 lr=0.001000
Epoch[060] Batch [1849]/[3760]	Speed: 35.120620 samples/sec	 accuracy=76.736486	 loss=0.930885	 lr=0.001000
Epoch[060] Batch [1899]/[3760]	Speed: 35.189494 samples/sec	 accuracy=76.756579	 loss=0.929857	 lr=0.001000
Epoch[060] Batch [1949]/[3760]	Speed: 35.205356 samples/sec	 accuracy=76.746795	 loss=0.929286	 lr=0.001000
Epoch[060] Batch [1999]/[3760]	Speed: 35.388904 samples/sec	 accuracy=76.750781	 loss=0.929405	 lr=0.001000
Epoch[060] Batch [2049]/[3760]	Speed: 35.053526 samples/sec	 accuracy=76.724848	 loss=0.930799	 lr=0.001000
Epoch[060] Batch [2099]/[3760]	Speed: 35.243691 samples/sec	 accuracy=76.738095	 loss=0.930149	 lr=0.001000
Epoch[060] Batch [2149]/[3760]	Speed: 35.225414 samples/sec	 accuracy=76.763081	 loss=0.928943	 lr=0.001000
Epoch[060] Batch [2199]/[3760]	Speed: 35.345984 samples/sec	 accuracy=76.755682	 loss=0.929319	 lr=0.001000
Epoch[060] Batch [2249]/[3760]	Speed: 35.239865 samples/sec	 accuracy=76.771528	 loss=0.928705	 lr=0.001000
Epoch[060] Batch [2299]/[3760]	Speed: 35.354805 samples/sec	 accuracy=76.750679	 loss=0.929448	 lr=0.001000
Epoch[060] Batch [2349]/[3760]	Speed: 35.193827 samples/sec	 accuracy=76.759973	 loss=0.929067	 lr=0.001000
Epoch[060] Batch [2399]/[3760]	Speed: 35.366500 samples/sec	 accuracy=76.792969	 loss=0.927775	 lr=0.001000
Epoch[060] Batch [2449]/[3760]	Speed: 35.098159 samples/sec	 accuracy=76.790179	 loss=0.928361	 lr=0.001000
Epoch[060] Batch [2499]/[3760]	Speed: 35.237979 samples/sec	 accuracy=76.802500	 loss=0.927753	 lr=0.001000
Epoch[060] Batch [2549]/[3760]	Speed: 35.305526 samples/sec	 accuracy=76.800245	 loss=0.927571	 lr=0.001000
Epoch[060] Batch [2599]/[3760]	Speed: 35.062892 samples/sec	 accuracy=76.806490	 loss=0.927076	 lr=0.001000
Epoch[060] Batch [2649]/[3760]	Speed: 34.881687 samples/sec	 accuracy=76.829009	 loss=0.926345	 lr=0.001000
Epoch[060] Batch [2699]/[3760]	Speed: 35.405362 samples/sec	 accuracy=76.831597	 loss=0.925847	 lr=0.001000
Epoch[060] Batch [2749]/[3760]	Speed: 34.982292 samples/sec	 accuracy=76.810227	 loss=0.926458	 lr=0.001000
Epoch[060] Batch [2799]/[3760]	Speed: 35.362405 samples/sec	 accuracy=76.818080	 loss=0.926532	 lr=0.001000
Epoch[060] Batch [2849]/[3760]	Speed: 35.143514 samples/sec	 accuracy=76.813048	 loss=0.926548	 lr=0.001000
Epoch[060] Batch [2899]/[3760]	Speed: 35.403949 samples/sec	 accuracy=76.790948	 loss=0.927027	 lr=0.001000
Epoch[060] Batch [2949]/[3760]	Speed: 35.199770 samples/sec	 accuracy=76.790254	 loss=0.927015	 lr=0.001000
Epoch[060] Batch [2999]/[3760]	Speed: 35.401113 samples/sec	 accuracy=76.789062	 loss=0.927019	 lr=0.001000
Epoch[060] Batch [3049]/[3760]	Speed: 35.132827 samples/sec	 accuracy=76.780225	 loss=0.928171	 lr=0.001000
Epoch[060] Batch [3099]/[3760]	Speed: 35.321658 samples/sec	 accuracy=76.799899	 loss=0.927452	 lr=0.001000
Epoch[060] Batch [3149]/[3760]	Speed: 34.905582 samples/sec	 accuracy=76.773313	 loss=0.928099	 lr=0.001000
Epoch[060] Batch [3199]/[3760]	Speed: 35.309725 samples/sec	 accuracy=76.753418	 loss=0.928492	 lr=0.001000
Epoch[060] Batch [3249]/[3760]	Speed: 34.835253 samples/sec	 accuracy=76.737500	 loss=0.929207	 lr=0.001000
Epoch[060] Batch [3299]/[3760]	Speed: 35.189254 samples/sec	 accuracy=76.726326	 loss=0.929845	 lr=0.001000
Epoch[060] Batch [3349]/[3760]	Speed: 35.047778 samples/sec	 accuracy=76.739272	 loss=0.929854	 lr=0.001000
Epoch[060] Batch [3399]/[3760]	Speed: 35.007803 samples/sec	 accuracy=76.745404	 loss=0.929523	 lr=0.001000
Epoch[060] Batch [3449]/[3760]	Speed: 35.179482 samples/sec	 accuracy=76.732337	 loss=0.929838	 lr=0.001000
Epoch[060] Batch [3499]/[3760]	Speed: 35.123401 samples/sec	 accuracy=76.720536	 loss=0.930240	 lr=0.001000
Epoch[060] Batch [3549]/[3760]	Speed: 35.323458 samples/sec	 accuracy=76.717430	 loss=0.930591	 lr=0.001000
Epoch[060] Batch [3599]/[3760]	Speed: 34.747317 samples/sec	 accuracy=76.722656	 loss=0.930523	 lr=0.001000
Epoch[060] Batch [3649]/[3760]	Speed: 35.266698 samples/sec	 accuracy=76.715325	 loss=0.930638	 lr=0.001000
Epoch[060] Batch [3699]/[3760]	Speed: 34.816709 samples/sec	 accuracy=76.710726	 loss=0.931142	 lr=0.001000
Epoch[060] Batch [3749]/[3760]	Speed: 36.371222 samples/sec	 accuracy=76.702500	 loss=0.931668	 lr=0.001000
Batch [0049]/[0303]: acc-top1=67.000000 acc-top5=86.250000
Batch [0099]/[0303]: acc-top1=66.171875 acc-top5=85.343750
Batch [0149]/[0303]: acc-top1=66.333333 acc-top5=85.531250
Batch [0199]/[0303]: acc-top1=66.117188 acc-top5=85.609375
Batch [0249]/[0303]: acc-top1=65.893750 acc-top5=85.437500
Batch [0299]/[0303]: acc-top1=65.796875 acc-top5=85.427083
[Epoch 060] training: accuracy=76.709192	 loss=0.931424
[Epoch 060] speed: 34 samples/sec	time cost: 7336.250678
[Epoch 060] validation: acc-top1=65.774546 acc-top5=85.442450 loss=1.605006
Epoch[061] Batch [0049]/[3759]	Speed: 24.533479 samples/sec	 accuracy=77.593750	 loss=0.897086	 lr=0.001000
Epoch[061] Batch [0099]/[3759]	Speed: 35.136280 samples/sec	 accuracy=78.203125	 loss=0.875922	 lr=0.001000
Epoch[061] Batch [0149]/[3759]	Speed: 34.909691 samples/sec	 accuracy=77.500000	 loss=0.905771	 lr=0.001000
Epoch[061] Batch [0199]/[3759]	Speed: 35.129293 samples/sec	 accuracy=77.578125	 loss=0.901974	 lr=0.001000
Epoch[061] Batch [0249]/[3759]	Speed: 35.190282 samples/sec	 accuracy=77.706250	 loss=0.895880	 lr=0.001000
Epoch[061] Batch [0299]/[3759]	Speed: 35.213882 samples/sec	 accuracy=77.786458	 loss=0.890976	 lr=0.001000
Epoch[061] Batch [0349]/[3759]	Speed: 35.206333 samples/sec	 accuracy=77.607143	 loss=0.895942	 lr=0.001000
Epoch[061] Batch [0399]/[3759]	Speed: 35.199150 samples/sec	 accuracy=77.660156	 loss=0.898780	 lr=0.001000
Epoch[061] Batch [0449]/[3759]	Speed: 35.180886 samples/sec	 accuracy=77.579861	 loss=0.897581	 lr=0.001000
Epoch[061] Batch [0499]/[3759]	Speed: 34.955621 samples/sec	 accuracy=77.462500	 loss=0.903250	 lr=0.001000
Epoch[061] Batch [0549]/[3759]	Speed: 35.295081 samples/sec	 accuracy=77.400568	 loss=0.905776	 lr=0.001000
Epoch[061] Batch [0599]/[3759]	Speed: 34.924743 samples/sec	 accuracy=77.500000	 loss=0.903580	 lr=0.001000
Epoch[061] Batch [0649]/[3759]	Speed: 35.518172 samples/sec	 accuracy=77.399038	 loss=0.907167	 lr=0.001000
Epoch[061] Batch [0699]/[3759]	Speed: 35.081238 samples/sec	 accuracy=77.299107	 loss=0.909947	 lr=0.001000
Epoch[061] Batch [0749]/[3759]	Speed: 35.363348 samples/sec	 accuracy=77.247917	 loss=0.912522	 lr=0.001000
Epoch[061] Batch [0799]/[3759]	Speed: 35.204824 samples/sec	 accuracy=77.193359	 loss=0.910893	 lr=0.001000
Epoch[061] Batch [0849]/[3759]	Speed: 35.455821 samples/sec	 accuracy=77.185662	 loss=0.911924	 lr=0.001000
Epoch[061] Batch [0899]/[3759]	Speed: 35.401559 samples/sec	 accuracy=77.185764	 loss=0.910216	 lr=0.001000
Epoch[061] Batch [0949]/[3759]	Speed: 35.303337 samples/sec	 accuracy=77.223684	 loss=0.907671	 lr=0.001000
Epoch[061] Batch [0999]/[3759]	Speed: 35.209485 samples/sec	 accuracy=77.201562	 loss=0.908993	 lr=0.001000
Epoch[061] Batch [1049]/[3759]	Speed: 35.427466 samples/sec	 accuracy=77.113095	 loss=0.911317	 lr=0.001000
Epoch[061] Batch [1099]/[3759]	Speed: 35.352802 samples/sec	 accuracy=77.093750	 loss=0.911237	 lr=0.001000
Epoch[061] Batch [1149]/[3759]	Speed: 35.406501 samples/sec	 accuracy=77.063859	 loss=0.912351	 lr=0.001000
Epoch[061] Batch [1199]/[3759]	Speed: 35.204027 samples/sec	 accuracy=77.065104	 loss=0.912083	 lr=0.001000
Epoch[061] Batch [1249]/[3759]	Speed: 35.132568 samples/sec	 accuracy=77.096250	 loss=0.911992	 lr=0.001000
Epoch[061] Batch [1299]/[3759]	Speed: 35.254242 samples/sec	 accuracy=77.099760	 loss=0.912228	 lr=0.001000
Epoch[061] Batch [1349]/[3759]	Speed: 35.106440 samples/sec	 accuracy=77.108796	 loss=0.912268	 lr=0.001000
Epoch[061] Batch [1399]/[3759]	Speed: 35.353247 samples/sec	 accuracy=77.074777	 loss=0.913858	 lr=0.001000
Epoch[061] Batch [1449]/[3759]	Speed: 35.329787 samples/sec	 accuracy=77.065733	 loss=0.913715	 lr=0.001000
Epoch[061] Batch [1499]/[3759]	Speed: 35.316480 samples/sec	 accuracy=77.062500	 loss=0.914175	 lr=0.001000
Epoch[061] Batch [1549]/[3759]	Speed: 35.445479 samples/sec	 accuracy=77.089718	 loss=0.914567	 lr=0.001000
Epoch[061] Batch [1599]/[3759]	Speed: 35.233634 samples/sec	 accuracy=77.047852	 loss=0.915071	 lr=0.001000
Epoch[061] Batch [1649]/[3759]	Speed: 35.338074 samples/sec	 accuracy=77.070076	 loss=0.915008	 lr=0.001000
Epoch[061] Batch [1699]/[3759]	Speed: 35.396778 samples/sec	 accuracy=77.035846	 loss=0.916974	 lr=0.001000
Epoch[061] Batch [1749]/[3759]	Speed: 35.318543 samples/sec	 accuracy=77.048214	 loss=0.916291	 lr=0.001000
Epoch[061] Batch [1799]/[3759]	Speed: 35.183380 samples/sec	 accuracy=77.063368	 loss=0.915042	 lr=0.001000
Epoch[061] Batch [1849]/[3759]	Speed: 35.346533 samples/sec	 accuracy=77.059966	 loss=0.915186	 lr=0.001000
Epoch[061] Batch [1899]/[3759]	Speed: 34.791412 samples/sec	 accuracy=76.998355	 loss=0.918313	 lr=0.001000
Epoch[061] Batch [1949]/[3759]	Speed: 35.479121 samples/sec	 accuracy=76.996795	 loss=0.918091	 lr=0.001000
Epoch[061] Batch [1999]/[3759]	Speed: 34.870957 samples/sec	 accuracy=76.989063	 loss=0.918788	 lr=0.001000
Epoch[061] Batch [2049]/[3759]	Speed: 35.395507 samples/sec	 accuracy=76.951982	 loss=0.919754	 lr=0.001000
Epoch[061] Batch [2099]/[3759]	Speed: 34.826476 samples/sec	 accuracy=76.972470	 loss=0.918826	 lr=0.001000
Epoch[061] Batch [2149]/[3759]	Speed: 35.326125 samples/sec	 accuracy=76.941860	 loss=0.919931	 lr=0.001000
Epoch[061] Batch [2199]/[3759]	Speed: 34.955547 samples/sec	 accuracy=76.953125	 loss=0.919407	 lr=0.001000
Epoch[061] Batch [2249]/[3759]	Speed: 35.147132 samples/sec	 accuracy=76.978472	 loss=0.918728	 lr=0.001000
Epoch[061] Batch [2299]/[3759]	Speed: 35.184814 samples/sec	 accuracy=77.014266	 loss=0.918014	 lr=0.001000
Epoch[061] Batch [2349]/[3759]	Speed: 34.884936 samples/sec	 accuracy=77.013298	 loss=0.917677	 lr=0.001000
Epoch[061] Batch [2399]/[3759]	Speed: 35.319048 samples/sec	 accuracy=77.036458	 loss=0.916980	 lr=0.001000
Epoch[061] Batch [2449]/[3759]	Speed: 35.085808 samples/sec	 accuracy=77.029337	 loss=0.918282	 lr=0.001000
Epoch[061] Batch [2499]/[3759]	Speed: 35.368984 samples/sec	 accuracy=77.008125	 loss=0.918794	 lr=0.001000
Epoch[061] Batch [2549]/[3759]	Speed: 34.894290 samples/sec	 accuracy=77.018382	 loss=0.918352	 lr=0.001000
Epoch[061] Batch [2599]/[3759]	Speed: 35.401428 samples/sec	 accuracy=77.013221	 loss=0.918133	 lr=0.001000
Epoch[061] Batch [2649]/[3759]	Speed: 34.724124 samples/sec	 accuracy=77.022995	 loss=0.917862	 lr=0.001000
Epoch[061] Batch [2699]/[3759]	Speed: 35.205873 samples/sec	 accuracy=77.005208	 loss=0.918724	 lr=0.001000
Epoch[061] Batch [2749]/[3759]	Speed: 34.606905 samples/sec	 accuracy=77.003977	 loss=0.918673	 lr=0.001000
Epoch[061] Batch [2799]/[3759]	Speed: 35.222010 samples/sec	 accuracy=77.011161	 loss=0.918440	 lr=0.001000
Epoch[061] Batch [2849]/[3759]	Speed: 34.593448 samples/sec	 accuracy=77.007127	 loss=0.917855	 lr=0.001000
Epoch[061] Batch [2899]/[3759]	Speed: 35.212542 samples/sec	 accuracy=77.014547	 loss=0.917466	 lr=0.001000
Epoch[061] Batch [2949]/[3759]	Speed: 34.910833 samples/sec	 accuracy=77.013242	 loss=0.917292	 lr=0.001000
Epoch[061] Batch [2999]/[3759]	Speed: 35.004535 samples/sec	 accuracy=77.004688	 loss=0.917874	 lr=0.001000
Epoch[061] Batch [3049]/[3759]	Speed: 34.989577 samples/sec	 accuracy=76.979508	 loss=0.918676	 lr=0.001000
Epoch[061] Batch [3099]/[3759]	Speed: 34.870808 samples/sec	 accuracy=76.973286	 loss=0.919269	 lr=0.001000
Epoch[061] Batch [3149]/[3759]	Speed: 35.060712 samples/sec	 accuracy=76.978671	 loss=0.919240	 lr=0.001000
Epoch[061] Batch [3199]/[3759]	Speed: 34.475514 samples/sec	 accuracy=76.996094	 loss=0.918781	 lr=0.001000
Epoch[061] Batch [3249]/[3759]	Speed: 35.499825 samples/sec	 accuracy=77.007212	 loss=0.918310	 lr=0.001000
Epoch[061] Batch [3299]/[3759]	Speed: 34.670786 samples/sec	 accuracy=76.992424	 loss=0.918941	 lr=0.001000
Epoch[061] Batch [3349]/[3759]	Speed: 35.420567 samples/sec	 accuracy=76.996269	 loss=0.918644	 lr=0.001000
Epoch[061] Batch [3399]/[3759]	Speed: 34.976406 samples/sec	 accuracy=77.000000	 loss=0.918755	 lr=0.001000
Epoch[061] Batch [3449]/[3759]	Speed: 35.424796 samples/sec	 accuracy=76.994565	 loss=0.918877	 lr=0.001000
Epoch[061] Batch [3499]/[3759]	Speed: 34.986551 samples/sec	 accuracy=76.993750	 loss=0.918709	 lr=0.001000
Epoch[061] Batch [3549]/[3759]	Speed: 35.317975 samples/sec	 accuracy=76.999560	 loss=0.918751	 lr=0.001000
Epoch[061] Batch [3599]/[3759]	Speed: 35.173979 samples/sec	 accuracy=76.977865	 loss=0.919623	 lr=0.001000
Epoch[061] Batch [3649]/[3759]	Speed: 35.105249 samples/sec	 accuracy=76.964897	 loss=0.920212	 lr=0.001000
Epoch[061] Batch [3699]/[3759]	Speed: 35.120302 samples/sec	 accuracy=76.956503	 loss=0.920939	 lr=0.001000
Epoch[061] Batch [3749]/[3759]	Speed: 36.282386 samples/sec	 accuracy=76.972917	 loss=0.920357	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.687500 acc-top5=85.937500
Batch [0099]/[0303]: acc-top1=66.187500 acc-top5=85.390625
Batch [0149]/[0303]: acc-top1=66.635417 acc-top5=85.593750
Batch [0199]/[0303]: acc-top1=66.382812 acc-top5=85.679688
Batch [0249]/[0303]: acc-top1=66.043750 acc-top5=85.512500
Batch [0299]/[0303]: acc-top1=65.937500 acc-top5=85.645833
[Epoch 061] training: accuracy=76.970271	 loss=0.920525
[Epoch 061] speed: 34 samples/sec	time cost: 7337.982986
[Epoch 061] validation: acc-top1=65.929249 acc-top5=85.664191 loss=1.619161
Epoch[062] Batch [0049]/[3760]	Speed: 24.687820 samples/sec	 accuracy=77.562500	 loss=0.893895	 lr=0.001000
Epoch[062] Batch [0099]/[3760]	Speed: 35.283215 samples/sec	 accuracy=77.000000	 loss=0.919601	 lr=0.001000
Epoch[062] Batch [0149]/[3760]	Speed: 35.030834 samples/sec	 accuracy=76.770833	 loss=0.924453	 lr=0.001000
Epoch[062] Batch [0199]/[3760]	Speed: 35.352689 samples/sec	 accuracy=77.187500	 loss=0.913193	 lr=0.001000
Epoch[062] Batch [0249]/[3760]	Speed: 35.389316 samples/sec	 accuracy=77.187500	 loss=0.912815	 lr=0.001000
Epoch[062] Batch [0299]/[3760]	Speed: 35.281852 samples/sec	 accuracy=77.239583	 loss=0.910362	 lr=0.001000
Epoch[062] Batch [0349]/[3760]	Speed: 35.294551 samples/sec	 accuracy=77.267857	 loss=0.908336	 lr=0.001000
Epoch[062] Batch [0399]/[3760]	Speed: 34.936260 samples/sec	 accuracy=77.191406	 loss=0.910102	 lr=0.001000
Epoch[062] Batch [0449]/[3760]	Speed: 34.978165 samples/sec	 accuracy=77.225694	 loss=0.907830	 lr=0.001000
Epoch[062] Batch [0499]/[3760]	Speed: 35.152541 samples/sec	 accuracy=77.309375	 loss=0.905938	 lr=0.001000
Epoch[062] Batch [0549]/[3760]	Speed: 35.229732 samples/sec	 accuracy=77.306818	 loss=0.903289	 lr=0.001000
Epoch[062] Batch [0599]/[3760]	Speed: 35.002821 samples/sec	 accuracy=77.252604	 loss=0.904790	 lr=0.001000
Epoch[062] Batch [0649]/[3760]	Speed: 35.422665 samples/sec	 accuracy=77.192308	 loss=0.907404	 lr=0.001000
Epoch[062] Batch [0699]/[3760]	Speed: 34.763592 samples/sec	 accuracy=77.176339	 loss=0.906067	 lr=0.001000
Epoch[062] Batch [0749]/[3760]	Speed: 35.475953 samples/sec	 accuracy=77.156250	 loss=0.908040	 lr=0.001000
Epoch[062] Batch [0799]/[3760]	Speed: 35.037259 samples/sec	 accuracy=77.072266	 loss=0.910402	 lr=0.001000
Epoch[062] Batch [0849]/[3760]	Speed: 35.372799 samples/sec	 accuracy=77.007353	 loss=0.911204	 lr=0.001000
Epoch[062] Batch [0899]/[3760]	Speed: 35.024651 samples/sec	 accuracy=77.026042	 loss=0.912416	 lr=0.001000
Epoch[062] Batch [0949]/[3760]	Speed: 35.477561 samples/sec	 accuracy=77.082237	 loss=0.910444	 lr=0.001000
Epoch[062] Batch [0999]/[3760]	Speed: 35.127037 samples/sec	 accuracy=77.082813	 loss=0.911634	 lr=0.001000
Epoch[062] Batch [1049]/[3760]	Speed: 35.498266 samples/sec	 accuracy=77.125000	 loss=0.911967	 lr=0.001000
Epoch[062] Batch [1099]/[3760]	Speed: 35.293899 samples/sec	 accuracy=77.166193	 loss=0.911877	 lr=0.001000
Epoch[062] Batch [1149]/[3760]	Speed: 35.410803 samples/sec	 accuracy=77.173913	 loss=0.912804	 lr=0.001000
Epoch[062] Batch [1199]/[3760]	Speed: 35.272819 samples/sec	 accuracy=77.165365	 loss=0.912655	 lr=0.001000
Epoch[062] Batch [1249]/[3760]	Speed: 35.546213 samples/sec	 accuracy=77.200000	 loss=0.912168	 lr=0.001000
Epoch[062] Batch [1299]/[3760]	Speed: 35.443968 samples/sec	 accuracy=77.199519	 loss=0.912170	 lr=0.001000
Epoch[062] Batch [1349]/[3760]	Speed: 35.246276 samples/sec	 accuracy=77.237269	 loss=0.912094	 lr=0.001000
Epoch[062] Batch [1399]/[3760]	Speed: 35.497956 samples/sec	 accuracy=77.202009	 loss=0.913223	 lr=0.001000
Epoch[062] Batch [1449]/[3760]	Speed: 35.261590 samples/sec	 accuracy=77.211207	 loss=0.912630	 lr=0.001000
Epoch[062] Batch [1499]/[3760]	Speed: 35.355150 samples/sec	 accuracy=77.256250	 loss=0.911638	 lr=0.001000
Epoch[062] Batch [1549]/[3760]	Speed: 35.413477 samples/sec	 accuracy=77.245968	 loss=0.911455	 lr=0.001000
Epoch[062] Batch [1599]/[3760]	Speed: 35.205723 samples/sec	 accuracy=77.253906	 loss=0.910616	 lr=0.001000
Epoch[062] Batch [1649]/[3760]	Speed: 35.554497 samples/sec	 accuracy=77.249053	 loss=0.910332	 lr=0.001000
Epoch[062] Batch [1699]/[3760]	Speed: 35.197801 samples/sec	 accuracy=77.252757	 loss=0.909731	 lr=0.001000
Epoch[062] Batch [1749]/[3760]	Speed: 35.270252 samples/sec	 accuracy=77.221429	 loss=0.910121	 lr=0.001000
Epoch[062] Batch [1799]/[3760]	Speed: 35.243666 samples/sec	 accuracy=77.260417	 loss=0.908190	 lr=0.001000
Epoch[062] Batch [1849]/[3760]	Speed: 35.486465 samples/sec	 accuracy=77.233953	 loss=0.908830	 lr=0.001000
Epoch[062] Batch [1899]/[3760]	Speed: 35.264681 samples/sec	 accuracy=77.238487	 loss=0.909376	 lr=0.001000
Epoch[062] Batch [1949]/[3760]	Speed: 35.413263 samples/sec	 accuracy=77.277244	 loss=0.907684	 lr=0.001000
Epoch[062] Batch [1999]/[3760]	Speed: 35.294396 samples/sec	 accuracy=77.300781	 loss=0.907248	 lr=0.001000
Epoch[062] Batch [2049]/[3760]	Speed: 35.512967 samples/sec	 accuracy=77.272104	 loss=0.908240	 lr=0.001000
Epoch[062] Batch [2099]/[3760]	Speed: 35.324874 samples/sec	 accuracy=77.264881	 loss=0.908977	 lr=0.001000
Epoch[062] Batch [2149]/[3760]	Speed: 35.565873 samples/sec	 accuracy=77.259448	 loss=0.909191	 lr=0.001000
Epoch[062] Batch [2199]/[3760]	Speed: 35.133367 samples/sec	 accuracy=77.262784	 loss=0.908838	 lr=0.001000
Epoch[062] Batch [2249]/[3760]	Speed: 35.462450 samples/sec	 accuracy=77.258333	 loss=0.908873	 lr=0.001000
Epoch[062] Batch [2299]/[3760]	Speed: 35.247212 samples/sec	 accuracy=77.236413	 loss=0.909359	 lr=0.001000
Epoch[062] Batch [2349]/[3760]	Speed: 35.357405 samples/sec	 accuracy=77.209441	 loss=0.910367	 lr=0.001000
Epoch[062] Batch [2399]/[3760]	Speed: 35.584515 samples/sec	 accuracy=77.213542	 loss=0.909844	 lr=0.001000
Epoch[062] Batch [2449]/[3760]	Speed: 35.448667 samples/sec	 accuracy=77.193240	 loss=0.910196	 lr=0.001000
Epoch[062] Batch [2499]/[3760]	Speed: 35.322751 samples/sec	 accuracy=77.205625	 loss=0.909920	 lr=0.001000
Epoch[062] Batch [2549]/[3760]	Speed: 35.331309 samples/sec	 accuracy=77.221201	 loss=0.910153	 lr=0.001000
Epoch[062] Batch [2599]/[3760]	Speed: 35.153660 samples/sec	 accuracy=77.210337	 loss=0.910104	 lr=0.001000
Epoch[062] Batch [2649]/[3760]	Speed: 35.237544 samples/sec	 accuracy=77.206958	 loss=0.910246	 lr=0.001000
Epoch[062] Batch [2699]/[3760]	Speed: 35.214846 samples/sec	 accuracy=77.208333	 loss=0.910126	 lr=0.001000
Epoch[062] Batch [2749]/[3760]	Speed: 34.823175 samples/sec	 accuracy=77.192045	 loss=0.910610	 lr=0.001000
Epoch[062] Batch [2799]/[3760]	Speed: 35.367230 samples/sec	 accuracy=77.190290	 loss=0.910446	 lr=0.001000
Epoch[062] Batch [2849]/[3760]	Speed: 34.715195 samples/sec	 accuracy=77.192982	 loss=0.910623	 lr=0.001000
Epoch[062] Batch [2899]/[3760]	Speed: 35.375613 samples/sec	 accuracy=77.186422	 loss=0.910776	 lr=0.001000
Epoch[062] Batch [2949]/[3760]	Speed: 34.621431 samples/sec	 accuracy=77.189089	 loss=0.910961	 lr=0.001000
Epoch[062] Batch [2999]/[3760]	Speed: 35.400193 samples/sec	 accuracy=77.229687	 loss=0.909489	 lr=0.001000
Epoch[062] Batch [3049]/[3760]	Speed: 34.757602 samples/sec	 accuracy=77.244877	 loss=0.909024	 lr=0.001000
Epoch[062] Batch [3099]/[3760]	Speed: 34.994281 samples/sec	 accuracy=77.238407	 loss=0.909092	 lr=0.001000
Epoch[062] Batch [3149]/[3760]	Speed: 34.932335 samples/sec	 accuracy=77.227679	 loss=0.909178	 lr=0.001000
Epoch[062] Batch [3199]/[3760]	Speed: 34.893304 samples/sec	 accuracy=77.203125	 loss=0.909960	 lr=0.001000
Epoch[062] Batch [3249]/[3760]	Speed: 34.933961 samples/sec	 accuracy=77.226442	 loss=0.909062	 lr=0.001000
Epoch[062] Batch [3299]/[3760]	Speed: 34.343368 samples/sec	 accuracy=77.212121	 loss=0.909363	 lr=0.001000
Epoch[062] Batch [3349]/[3760]	Speed: 35.226041 samples/sec	 accuracy=77.207556	 loss=0.909624	 lr=0.001000
Epoch[062] Batch [3399]/[3760]	Speed: 34.365418 samples/sec	 accuracy=77.209559	 loss=0.909536	 lr=0.001000
Epoch[062] Batch [3449]/[3760]	Speed: 35.194582 samples/sec	 accuracy=77.211504	 loss=0.910036	 lr=0.001000
Epoch[062] Batch [3499]/[3760]	Speed: 34.467629 samples/sec	 accuracy=77.206250	 loss=0.910156	 lr=0.001000
Epoch[062] Batch [3549]/[3760]	Speed: 34.940197 samples/sec	 accuracy=77.191461	 loss=0.910635	 lr=0.001000
Epoch[062] Batch [3599]/[3760]	Speed: 34.818938 samples/sec	 accuracy=77.191406	 loss=0.911148	 lr=0.001000
Epoch[062] Batch [3649]/[3760]	Speed: 34.652044 samples/sec	 accuracy=77.178938	 loss=0.911243	 lr=0.001000
Epoch[062] Batch [3699]/[3760]	Speed: 34.975300 samples/sec	 accuracy=77.189611	 loss=0.911186	 lr=0.001000
Epoch[062] Batch [3749]/[3760]	Speed: 36.042218 samples/sec	 accuracy=77.180417	 loss=0.911719	 lr=0.001000
Batch [0049]/[0303]: acc-top1=67.312500 acc-top5=85.843750
Batch [0099]/[0303]: acc-top1=66.593750 acc-top5=85.265625
Batch [0149]/[0303]: acc-top1=66.968750 acc-top5=85.447917
Batch [0199]/[0303]: acc-top1=66.539062 acc-top5=85.476562
Batch [0249]/[0303]: acc-top1=66.187500 acc-top5=85.381250
Batch [0299]/[0303]: acc-top1=66.067708 acc-top5=85.432292
[Epoch 062] training: accuracy=77.182929	 loss=0.911677
[Epoch 062] speed: 34 samples/sec	time cost: 7332.714885
[Epoch 062] validation: acc-top1=66.058168 acc-top5=85.442450 loss=1.607593
Epoch[063] Batch [0049]/[3760]	Speed: 25.355602 samples/sec	 accuracy=77.281250	 loss=0.915931	 lr=0.001000
Epoch[063] Batch [0099]/[3760]	Speed: 35.332564 samples/sec	 accuracy=77.562500	 loss=0.903206	 lr=0.001000
Epoch[063] Batch [0149]/[3760]	Speed: 35.290514 samples/sec	 accuracy=77.593750	 loss=0.902887	 lr=0.001000
Epoch[063] Batch [0199]/[3760]	Speed: 35.224946 samples/sec	 accuracy=77.843750	 loss=0.888401	 lr=0.001000
Epoch[063] Batch [0249]/[3760]	Speed: 35.233731 samples/sec	 accuracy=77.887500	 loss=0.890404	 lr=0.001000
Epoch[063] Batch [0299]/[3760]	Speed: 35.273468 samples/sec	 accuracy=77.984375	 loss=0.883623	 lr=0.001000
Epoch[063] Batch [0349]/[3760]	Speed: 35.287629 samples/sec	 accuracy=77.875000	 loss=0.887853	 lr=0.001000
Epoch[063] Batch [0399]/[3760]	Speed: 35.156946 samples/sec	 accuracy=77.777344	 loss=0.892078	 lr=0.001000
Epoch[063] Batch [0449]/[3760]	Speed: 34.940459 samples/sec	 accuracy=77.868056	 loss=0.891873	 lr=0.001000
Epoch[063] Batch [0499]/[3760]	Speed: 35.316930 samples/sec	 accuracy=77.775000	 loss=0.894582	 lr=0.001000
Epoch[063] Batch [0549]/[3760]	Speed: 34.672998 samples/sec	 accuracy=77.772727	 loss=0.891543	 lr=0.001000
Epoch[063] Batch [0599]/[3760]	Speed: 35.406115 samples/sec	 accuracy=77.796875	 loss=0.891947	 lr=0.001000
Epoch[063] Batch [0649]/[3760]	Speed: 34.591037 samples/sec	 accuracy=77.790865	 loss=0.893340	 lr=0.001000
Epoch[063] Batch [0699]/[3760]	Speed: 35.258301 samples/sec	 accuracy=77.723214	 loss=0.894683	 lr=0.001000
Epoch[063] Batch [0749]/[3760]	Speed: 34.836393 samples/sec	 accuracy=77.677083	 loss=0.895752	 lr=0.001000
Epoch[063] Batch [0799]/[3760]	Speed: 35.222872 samples/sec	 accuracy=77.644531	 loss=0.897044	 lr=0.001000
Epoch[063] Batch [0849]/[3760]	Speed: 34.870510 samples/sec	 accuracy=77.601103	 loss=0.899273	 lr=0.001000
Epoch[063] Batch [0899]/[3760]	Speed: 35.102187 samples/sec	 accuracy=77.543403	 loss=0.900360	 lr=0.001000
Epoch[063] Batch [0949]/[3760]	Speed: 34.813291 samples/sec	 accuracy=77.634868	 loss=0.896148	 lr=0.001000
Epoch[063] Batch [0999]/[3760]	Speed: 34.956917 samples/sec	 accuracy=77.615625	 loss=0.897068	 lr=0.001000
Epoch[063] Batch [1049]/[3760]	Speed: 35.041180 samples/sec	 accuracy=77.690476	 loss=0.894696	 lr=0.001000
Epoch[063] Batch [1099]/[3760]	Speed: 34.671112 samples/sec	 accuracy=77.666193	 loss=0.896265	 lr=0.001000
Epoch[063] Batch [1149]/[3760]	Speed: 35.084759 samples/sec	 accuracy=77.634511	 loss=0.896325	 lr=0.001000
Epoch[063] Batch [1199]/[3760]	Speed: 34.687479 samples/sec	 accuracy=77.574219	 loss=0.899089	 lr=0.001000
Epoch[063] Batch [1249]/[3760]	Speed: 35.213053 samples/sec	 accuracy=77.551250	 loss=0.899127	 lr=0.001000
Epoch[063] Batch [1299]/[3760]	Speed: 34.481505 samples/sec	 accuracy=77.524038	 loss=0.900514	 lr=0.001000
Epoch[063] Batch [1349]/[3760]	Speed: 35.179634 samples/sec	 accuracy=77.555556	 loss=0.899750	 lr=0.001000
Epoch[063] Batch [1399]/[3760]	Speed: 34.408084 samples/sec	 accuracy=77.545759	 loss=0.899511	 lr=0.001000
Epoch[063] Batch [1449]/[3760]	Speed: 35.162753 samples/sec	 accuracy=77.545259	 loss=0.898783	 lr=0.001000
Epoch[063] Batch [1499]/[3760]	Speed: 34.597816 samples/sec	 accuracy=77.511458	 loss=0.900742	 lr=0.001000
Epoch[063] Batch [1549]/[3760]	Speed: 35.044045 samples/sec	 accuracy=77.486895	 loss=0.901831	 lr=0.001000
Epoch[063] Batch [1599]/[3760]	Speed: 35.052681 samples/sec	 accuracy=77.475586	 loss=0.902372	 lr=0.001000
Epoch[063] Batch [1649]/[3760]	Speed: 34.886204 samples/sec	 accuracy=77.432765	 loss=0.904526	 lr=0.001000
Epoch[063] Batch [1699]/[3760]	Speed: 35.046374 samples/sec	 accuracy=77.441176	 loss=0.903746	 lr=0.001000
Epoch[063] Batch [1749]/[3760]	Speed: 34.803014 samples/sec	 accuracy=77.459821	 loss=0.903229	 lr=0.001000
Epoch[063] Batch [1799]/[3760]	Speed: 35.073404 samples/sec	 accuracy=77.446181	 loss=0.903320	 lr=0.001000
Epoch[063] Batch [1849]/[3760]	Speed: 34.581150 samples/sec	 accuracy=77.418074	 loss=0.904226	 lr=0.001000
Epoch[063] Batch [1899]/[3760]	Speed: 35.281257 samples/sec	 accuracy=77.431743	 loss=0.903766	 lr=0.001000
Epoch[063] Batch [1949]/[3760]	Speed: 34.534070 samples/sec	 accuracy=77.439904	 loss=0.903468	 lr=0.001000
Epoch[063] Batch [1999]/[3760]	Speed: 35.240031 samples/sec	 accuracy=77.444531	 loss=0.903134	 lr=0.001000
Epoch[063] Batch [2049]/[3760]	Speed: 34.528741 samples/sec	 accuracy=77.419970	 loss=0.904256	 lr=0.001000
Epoch[063] Batch [2099]/[3760]	Speed: 35.131170 samples/sec	 accuracy=77.437500	 loss=0.903369	 lr=0.001000
Epoch[063] Batch [2149]/[3760]	Speed: 34.764151 samples/sec	 accuracy=77.433140	 loss=0.903703	 lr=0.001000
Epoch[063] Batch [2199]/[3760]	Speed: 34.749871 samples/sec	 accuracy=77.438210	 loss=0.903328	 lr=0.001000
Epoch[063] Batch [2249]/[3760]	Speed: 34.810258 samples/sec	 accuracy=77.434722	 loss=0.903504	 lr=0.001000
Epoch[063] Batch [2299]/[3760]	Speed: 34.686809 samples/sec	 accuracy=77.415082	 loss=0.904762	 lr=0.001000
Epoch[063] Batch [2349]/[3760]	Speed: 34.755655 samples/sec	 accuracy=77.410239	 loss=0.905120	 lr=0.001000
Epoch[063] Batch [2399]/[3760]	Speed: 34.466112 samples/sec	 accuracy=77.417969	 loss=0.904886	 lr=0.001000
Epoch[063] Batch [2449]/[3760]	Speed: 35.024396 samples/sec	 accuracy=77.409439	 loss=0.905292	 lr=0.001000
Epoch[063] Batch [2499]/[3760]	Speed: 34.232717 samples/sec	 accuracy=77.418125	 loss=0.905202	 lr=0.001000
Epoch[063] Batch [2549]/[3760]	Speed: 35.228884 samples/sec	 accuracy=77.441789	 loss=0.904952	 lr=0.001000
Epoch[063] Batch [2599]/[3760]	Speed: 34.511113 samples/sec	 accuracy=77.433293	 loss=0.904517	 lr=0.001000
Epoch[063] Batch [2649]/[3760]	Speed: 35.058608 samples/sec	 accuracy=77.419222	 loss=0.904777	 lr=0.001000
Epoch[063] Batch [2699]/[3760]	Speed: 34.543256 samples/sec	 accuracy=77.413773	 loss=0.904783	 lr=0.001000
Epoch[063] Batch [2749]/[3760]	Speed: 34.691008 samples/sec	 accuracy=77.429545	 loss=0.904177	 lr=0.001000
Epoch[063] Batch [2799]/[3760]	Speed: 34.893562 samples/sec	 accuracy=77.449777	 loss=0.903762	 lr=0.001000
Epoch[063] Batch [2849]/[3760]	Speed: 34.435524 samples/sec	 accuracy=77.434759	 loss=0.904354	 lr=0.001000
Epoch[063] Batch [2899]/[3760]	Speed: 35.005702 samples/sec	 accuracy=77.432112	 loss=0.904618	 lr=0.001000
Epoch[063] Batch [2949]/[3760]	Speed: 34.455823 samples/sec	 accuracy=77.407309	 loss=0.905347	 lr=0.001000
Epoch[063] Batch [2999]/[3760]	Speed: 35.092959 samples/sec	 accuracy=77.395833	 loss=0.905937	 lr=0.001000
Epoch[063] Batch [3049]/[3760]	Speed: 34.270800 samples/sec	 accuracy=77.391906	 loss=0.905634	 lr=0.001000
Epoch[063] Batch [3099]/[3760]	Speed: 34.958504 samples/sec	 accuracy=77.373992	 loss=0.906171	 lr=0.001000
Epoch[063] Batch [3149]/[3760]	Speed: 34.718707 samples/sec	 accuracy=77.361111	 loss=0.906287	 lr=0.001000
Epoch[063] Batch [3199]/[3760]	Speed: 34.973332 samples/sec	 accuracy=77.352539	 loss=0.906616	 lr=0.001000
Epoch[063] Batch [3249]/[3760]	Speed: 35.032602 samples/sec	 accuracy=77.346154	 loss=0.907125	 lr=0.001000
Epoch[063] Batch [3299]/[3760]	Speed: 34.751322 samples/sec	 accuracy=77.332386	 loss=0.907427	 lr=0.001000
Epoch[063] Batch [3349]/[3760]	Speed: 35.003867 samples/sec	 accuracy=77.348881	 loss=0.906840	 lr=0.001000
Epoch[063] Batch [3399]/[3760]	Speed: 34.851112 samples/sec	 accuracy=77.348805	 loss=0.906951	 lr=0.001000
Epoch[063] Batch [3449]/[3760]	Speed: 35.242765 samples/sec	 accuracy=77.333786	 loss=0.907404	 lr=0.001000
Epoch[063] Batch [3499]/[3760]	Speed: 35.062137 samples/sec	 accuracy=77.341071	 loss=0.907311	 lr=0.001000
Epoch[063] Batch [3549]/[3760]	Speed: 35.215425 samples/sec	 accuracy=77.327905	 loss=0.907266	 lr=0.001000
Epoch[063] Batch [3599]/[3760]	Speed: 35.046953 samples/sec	 accuracy=77.332899	 loss=0.907011	 lr=0.001000
Epoch[063] Batch [3649]/[3760]	Speed: 35.155157 samples/sec	 accuracy=77.337329	 loss=0.906866	 lr=0.001000
Epoch[063] Batch [3699]/[3760]	Speed: 34.810121 samples/sec	 accuracy=77.335304	 loss=0.906884	 lr=0.001000
Epoch[063] Batch [3749]/[3760]	Speed: 36.568540 samples/sec	 accuracy=77.334167	 loss=0.906917	 lr=0.001000
Batch [0049]/[0303]: acc-top1=67.187500 acc-top5=86.187500
Batch [0099]/[0303]: acc-top1=66.750000 acc-top5=85.484375
Batch [0149]/[0303]: acc-top1=66.947917 acc-top5=85.718750
Batch [0199]/[0303]: acc-top1=66.640625 acc-top5=85.718750
Batch [0249]/[0303]: acc-top1=66.306250 acc-top5=85.537500
Batch [0299]/[0303]: acc-top1=66.151042 acc-top5=85.588542
[Epoch 063] training: accuracy=77.333361	 loss=0.906898
[Epoch 063] speed: 34 samples/sec	time cost: 7381.118358
[Epoch 063] validation: acc-top1=66.130363 acc-top5=85.597153 loss=1.617057
Epoch[064] Batch [0049]/[3759]	Speed: 25.063082 samples/sec	 accuracy=77.093750	 loss=0.921736	 lr=0.001000
Epoch[064] Batch [0099]/[3759]	Speed: 35.212296 samples/sec	 accuracy=77.359375	 loss=0.915548	 lr=0.001000
Epoch[064] Batch [0149]/[3759]	Speed: 35.371272 samples/sec	 accuracy=77.520833	 loss=0.905603	 lr=0.001000
Epoch[064] Batch [0199]/[3759]	Speed: 35.362196 samples/sec	 accuracy=77.531250	 loss=0.904396	 lr=0.001000
Epoch[064] Batch [0249]/[3759]	Speed: 35.283239 samples/sec	 accuracy=77.631250	 loss=0.897531	 lr=0.001000
Epoch[064] Batch [0299]/[3759]	Speed: 35.355045 samples/sec	 accuracy=77.609375	 loss=0.892950	 lr=0.001000
Epoch[064] Batch [0349]/[3759]	Speed: 35.261158 samples/sec	 accuracy=77.674107	 loss=0.890242	 lr=0.001000
Epoch[064] Batch [0399]/[3759]	Speed: 35.365102 samples/sec	 accuracy=77.824219	 loss=0.881215	 lr=0.001000
Epoch[064] Batch [0449]/[3759]	Speed: 35.304193 samples/sec	 accuracy=77.746528	 loss=0.884628	 lr=0.001000
Epoch[064] Batch [0499]/[3759]	Speed: 35.258291 samples/sec	 accuracy=77.781250	 loss=0.884569	 lr=0.001000
Epoch[064] Batch [0549]/[3759]	Speed: 35.209756 samples/sec	 accuracy=77.764205	 loss=0.885274	 lr=0.001000
Epoch[064] Batch [0599]/[3759]	Speed: 35.311188 samples/sec	 accuracy=77.653646	 loss=0.888535	 lr=0.001000
Epoch[064] Batch [0649]/[3759]	Speed: 35.084957 samples/sec	 accuracy=77.617788	 loss=0.887869	 lr=0.001000
Epoch[064] Batch [0699]/[3759]	Speed: 35.217123 samples/sec	 accuracy=77.694196	 loss=0.885479	 lr=0.001000
Epoch[064] Batch [0749]/[3759]	Speed: 35.241702 samples/sec	 accuracy=77.691667	 loss=0.886932	 lr=0.001000
Epoch[064] Batch [0799]/[3759]	Speed: 35.396710 samples/sec	 accuracy=77.683594	 loss=0.888745	 lr=0.001000
Epoch[064] Batch [0849]/[3759]	Speed: 35.195076 samples/sec	 accuracy=77.628676	 loss=0.890288	 lr=0.001000
Epoch[064] Batch [0899]/[3759]	Speed: 35.462425 samples/sec	 accuracy=77.642361	 loss=0.890073	 lr=0.001000
Epoch[064] Batch [0949]/[3759]	Speed: 35.187745 samples/sec	 accuracy=77.664474	 loss=0.889829	 lr=0.001000
Epoch[064] Batch [0999]/[3759]	Speed: 35.157905 samples/sec	 accuracy=77.681250	 loss=0.890013	 lr=0.001000
Epoch[064] Batch [1049]/[3759]	Speed: 35.192907 samples/sec	 accuracy=77.693452	 loss=0.889625	 lr=0.001000
Epoch[064] Batch [1099]/[3759]	Speed: 35.379285 samples/sec	 accuracy=77.707386	 loss=0.889860	 lr=0.001000
Epoch[064] Batch [1149]/[3759]	Speed: 35.309375 samples/sec	 accuracy=77.735054	 loss=0.888356	 lr=0.001000
Epoch[064] Batch [1199]/[3759]	Speed: 35.222917 samples/sec	 accuracy=77.727865	 loss=0.888614	 lr=0.001000
Epoch[064] Batch [1249]/[3759]	Speed: 35.192078 samples/sec	 accuracy=77.750000	 loss=0.888412	 lr=0.001000
Epoch[064] Batch [1299]/[3759]	Speed: 35.500407 samples/sec	 accuracy=77.724760	 loss=0.888307	 lr=0.001000
Epoch[064] Batch [1349]/[3759]	Speed: 35.199105 samples/sec	 accuracy=77.701389	 loss=0.889138	 lr=0.001000
Epoch[064] Batch [1399]/[3759]	Speed: 35.172210 samples/sec	 accuracy=77.714286	 loss=0.889037	 lr=0.001000
Epoch[064] Batch [1449]/[3759]	Speed: 35.170426 samples/sec	 accuracy=77.747845	 loss=0.888955	 lr=0.001000
Epoch[064] Batch [1499]/[3759]	Speed: 35.067692 samples/sec	 accuracy=77.711458	 loss=0.891081	 lr=0.001000
Epoch[064] Batch [1549]/[3759]	Speed: 35.162614 samples/sec	 accuracy=77.709677	 loss=0.891308	 lr=0.001000
Epoch[064] Batch [1599]/[3759]	Speed: 35.196698 samples/sec	 accuracy=77.708008	 loss=0.891766	 lr=0.001000
Epoch[064] Batch [1649]/[3759]	Speed: 35.245317 samples/sec	 accuracy=77.700758	 loss=0.892110	 lr=0.001000
Epoch[064] Batch [1699]/[3759]	Speed: 35.108489 samples/sec	 accuracy=77.690257	 loss=0.891989	 lr=0.001000
Epoch[064] Batch [1749]/[3759]	Speed: 35.210935 samples/sec	 accuracy=77.685714	 loss=0.892377	 lr=0.001000
Epoch[064] Batch [1799]/[3759]	Speed: 34.842834 samples/sec	 accuracy=77.680556	 loss=0.892186	 lr=0.001000
Epoch[064] Batch [1849]/[3759]	Speed: 35.342419 samples/sec	 accuracy=77.682432	 loss=0.891928	 lr=0.001000
Epoch[064] Batch [1899]/[3759]	Speed: 34.954868 samples/sec	 accuracy=77.683388	 loss=0.891909	 lr=0.001000
Epoch[064] Batch [1949]/[3759]	Speed: 35.267322 samples/sec	 accuracy=77.665064	 loss=0.892593	 lr=0.001000
Epoch[064] Batch [1999]/[3759]	Speed: 35.107330 samples/sec	 accuracy=77.645313	 loss=0.892817	 lr=0.001000
Epoch[064] Batch [2049]/[3759]	Speed: 35.317765 samples/sec	 accuracy=77.621951	 loss=0.893480	 lr=0.001000
Epoch[064] Batch [2099]/[3759]	Speed: 35.283961 samples/sec	 accuracy=77.655506	 loss=0.892368	 lr=0.001000
Epoch[064] Batch [2149]/[3759]	Speed: 35.453595 samples/sec	 accuracy=77.645349	 loss=0.893187	 lr=0.001000
Epoch[064] Batch [2199]/[3759]	Speed: 35.061865 samples/sec	 accuracy=77.656960	 loss=0.892597	 lr=0.001000
Epoch[064] Batch [2249]/[3759]	Speed: 35.343297 samples/sec	 accuracy=77.647917	 loss=0.893060	 lr=0.001000
Epoch[064] Batch [2299]/[3759]	Speed: 35.074378 samples/sec	 accuracy=77.634511	 loss=0.893082	 lr=0.001000
Epoch[064] Batch [2349]/[3759]	Speed: 35.290677 samples/sec	 accuracy=77.614362	 loss=0.893963	 lr=0.001000
Epoch[064] Batch [2399]/[3759]	Speed: 35.518452 samples/sec	 accuracy=77.606120	 loss=0.894245	 lr=0.001000
Epoch[064] Batch [2449]/[3759]	Speed: 35.072481 samples/sec	 accuracy=77.625000	 loss=0.893308	 lr=0.001000
Epoch[064] Batch [2499]/[3759]	Speed: 35.406274 samples/sec	 accuracy=77.635625	 loss=0.893414	 lr=0.001000
Epoch[064] Batch [2549]/[3759]	Speed: 35.223015 samples/sec	 accuracy=77.617647	 loss=0.893288	 lr=0.001000
Epoch[064] Batch [2599]/[3759]	Speed: 35.258773 samples/sec	 accuracy=77.621995	 loss=0.892865	 lr=0.001000
Epoch[064] Batch [2649]/[3759]	Speed: 35.091314 samples/sec	 accuracy=77.608491	 loss=0.893879	 lr=0.001000
Epoch[064] Batch [2699]/[3759]	Speed: 35.158411 samples/sec	 accuracy=77.607639	 loss=0.893741	 lr=0.001000
Epoch[064] Batch [2749]/[3759]	Speed: 35.015843 samples/sec	 accuracy=77.602841	 loss=0.893402	 lr=0.001000
Epoch[064] Batch [2799]/[3759]	Speed: 34.949336 samples/sec	 accuracy=77.604353	 loss=0.893379	 lr=0.001000
Epoch[064] Batch [2849]/[3759]	Speed: 34.874448 samples/sec	 accuracy=77.599781	 loss=0.893339	 lr=0.001000
Epoch[064] Batch [2899]/[3759]	Speed: 35.164015 samples/sec	 accuracy=77.566272	 loss=0.894098	 lr=0.001000
Epoch[064] Batch [2949]/[3759]	Speed: 34.705426 samples/sec	 accuracy=77.578919	 loss=0.893970	 lr=0.001000
Epoch[064] Batch [2999]/[3759]	Speed: 35.388845 samples/sec	 accuracy=77.553125	 loss=0.894521	 lr=0.001000
Epoch[064] Batch [3049]/[3759]	Speed: 34.724051 samples/sec	 accuracy=77.560963	 loss=0.894513	 lr=0.001000
Epoch[064] Batch [3099]/[3759]	Speed: 35.250623 samples/sec	 accuracy=77.553931	 loss=0.894552	 lr=0.001000
Epoch[064] Batch [3149]/[3759]	Speed: 34.797926 samples/sec	 accuracy=77.557044	 loss=0.894292	 lr=0.001000
Epoch[064] Batch [3199]/[3759]	Speed: 35.138746 samples/sec	 accuracy=77.571777	 loss=0.893834	 lr=0.001000
Epoch[064] Batch [3249]/[3759]	Speed: 35.108431 samples/sec	 accuracy=77.567788	 loss=0.894098	 lr=0.001000
Epoch[064] Batch [3299]/[3759]	Speed: 35.131003 samples/sec	 accuracy=77.566761	 loss=0.894090	 lr=0.001000
Epoch[064] Batch [3349]/[3759]	Speed: 35.122490 samples/sec	 accuracy=77.566231	 loss=0.894325	 lr=0.001000
Epoch[064] Batch [3399]/[3759]	Speed: 34.983025 samples/sec	 accuracy=77.553309	 loss=0.895127	 lr=0.001000
Epoch[064] Batch [3449]/[3759]	Speed: 35.373146 samples/sec	 accuracy=77.568841	 loss=0.894872	 lr=0.001000
Epoch[064] Batch [3499]/[3759]	Speed: 35.051220 samples/sec	 accuracy=77.563393	 loss=0.895332	 lr=0.001000
Epoch[064] Batch [3549]/[3759]	Speed: 35.402611 samples/sec	 accuracy=77.548415	 loss=0.895842	 lr=0.001000
Epoch[064] Batch [3599]/[3759]	Speed: 34.804368 samples/sec	 accuracy=77.539062	 loss=0.895791	 lr=0.001000
Epoch[064] Batch [3649]/[3759]	Speed: 35.376720 samples/sec	 accuracy=77.519692	 loss=0.896388	 lr=0.001000
Epoch[064] Batch [3699]/[3759]	Speed: 34.979838 samples/sec	 accuracy=77.527872	 loss=0.896223	 lr=0.001000
Epoch[064] Batch [3749]/[3759]	Speed: 36.404996 samples/sec	 accuracy=77.507083	 loss=0.897380	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.687500 acc-top5=85.343750
Batch [0099]/[0303]: acc-top1=66.015625 acc-top5=85.125000
Batch [0149]/[0303]: acc-top1=66.145833 acc-top5=85.250000
Batch [0199]/[0303]: acc-top1=66.000000 acc-top5=85.187500
Batch [0249]/[0303]: acc-top1=65.837500 acc-top5=85.075000
Batch [0299]/[0303]: acc-top1=65.786458 acc-top5=85.088542
[Epoch 064] training: accuracy=77.507316	 loss=0.897362
[Epoch 064] speed: 35 samples/sec	time cost: 7325.899477
[Epoch 064] validation: acc-top1=65.815800 acc-top5=85.096947 loss=1.620297
Epoch[065] Batch [0049]/[3760]	Speed: 24.983024 samples/sec	 accuracy=78.375000	 loss=0.869211	 lr=0.001000
Epoch[065] Batch [0099]/[3760]	Speed: 35.295087 samples/sec	 accuracy=78.078125	 loss=0.888280	 lr=0.001000
Epoch[065] Batch [0149]/[3760]	Speed: 35.349977 samples/sec	 accuracy=77.885417	 loss=0.893098	 lr=0.001000
Epoch[065] Batch [0199]/[3760]	Speed: 35.209185 samples/sec	 accuracy=78.375000	 loss=0.871140	 lr=0.001000
Epoch[065] Batch [0249]/[3760]	Speed: 35.400560 samples/sec	 accuracy=78.293750	 loss=0.869804	 lr=0.001000
Epoch[065] Batch [0299]/[3760]	Speed: 35.104880 samples/sec	 accuracy=78.239583	 loss=0.869126	 lr=0.001000
Epoch[065] Batch [0349]/[3760]	Speed: 35.260892 samples/sec	 accuracy=78.223214	 loss=0.865238	 lr=0.001000
Epoch[065] Batch [0399]/[3760]	Speed: 34.962660 samples/sec	 accuracy=78.292969	 loss=0.867529	 lr=0.001000
Epoch[065] Batch [0449]/[3760]	Speed: 35.193885 samples/sec	 accuracy=78.215278	 loss=0.868892	 lr=0.001000
Epoch[065] Batch [0499]/[3760]	Speed: 34.882367 samples/sec	 accuracy=78.146875	 loss=0.868370	 lr=0.001000
Epoch[065] Batch [0549]/[3760]	Speed: 35.224703 samples/sec	 accuracy=78.099432	 loss=0.868743	 lr=0.001000
Epoch[065] Batch [0599]/[3760]	Speed: 35.110304 samples/sec	 accuracy=78.067708	 loss=0.870102	 lr=0.001000
Epoch[065] Batch [0649]/[3760]	Speed: 35.205884 samples/sec	 accuracy=78.069712	 loss=0.870046	 lr=0.001000
Epoch[065] Batch [0699]/[3760]	Speed: 35.396953 samples/sec	 accuracy=77.957589	 loss=0.874286	 lr=0.001000
Epoch[065] Batch [0749]/[3760]	Speed: 35.291492 samples/sec	 accuracy=77.922917	 loss=0.874602	 lr=0.001000
Epoch[065] Batch [0799]/[3760]	Speed: 35.220982 samples/sec	 accuracy=77.925781	 loss=0.874419	 lr=0.001000
Epoch[065] Batch [0849]/[3760]	Speed: 35.157121 samples/sec	 accuracy=77.974265	 loss=0.873751	 lr=0.001000
Epoch[065] Batch [0899]/[3760]	Speed: 35.194189 samples/sec	 accuracy=77.946181	 loss=0.873899	 lr=0.001000
Epoch[065] Batch [0949]/[3760]	Speed: 35.139731 samples/sec	 accuracy=77.884868	 loss=0.876357	 lr=0.001000
Epoch[065] Batch [0999]/[3760]	Speed: 35.186272 samples/sec	 accuracy=77.910938	 loss=0.876413	 lr=0.001000
Epoch[065] Batch [1049]/[3760]	Speed: 35.263272 samples/sec	 accuracy=77.889881	 loss=0.877752	 lr=0.001000
Epoch[065] Batch [1099]/[3760]	Speed: 35.275932 samples/sec	 accuracy=77.880682	 loss=0.878315	 lr=0.001000
Epoch[065] Batch [1149]/[3760]	Speed: 35.214835 samples/sec	 accuracy=77.843750	 loss=0.879975	 lr=0.001000
Epoch[065] Batch [1199]/[3760]	Speed: 35.366824 samples/sec	 accuracy=77.869792	 loss=0.879308	 lr=0.001000
Epoch[065] Batch [1249]/[3760]	Speed: 35.065730 samples/sec	 accuracy=77.840000	 loss=0.880541	 lr=0.001000
Epoch[065] Batch [1299]/[3760]	Speed: 35.333342 samples/sec	 accuracy=77.793269	 loss=0.883134	 lr=0.001000
Epoch[065] Batch [1349]/[3760]	Speed: 34.951055 samples/sec	 accuracy=77.812500	 loss=0.883690	 lr=0.001000
Epoch[065] Batch [1399]/[3760]	Speed: 35.255512 samples/sec	 accuracy=77.824777	 loss=0.882976	 lr=0.001000
Epoch[065] Batch [1449]/[3760]	Speed: 35.226351 samples/sec	 accuracy=77.820043	 loss=0.883375	 lr=0.001000
Epoch[065] Batch [1499]/[3760]	Speed: 35.427176 samples/sec	 accuracy=77.753125	 loss=0.884683	 lr=0.001000
Epoch[065] Batch [1549]/[3760]	Speed: 34.970699 samples/sec	 accuracy=77.721774	 loss=0.884694	 lr=0.001000
Epoch[065] Batch [1599]/[3760]	Speed: 35.301798 samples/sec	 accuracy=77.681641	 loss=0.885971	 lr=0.001000
Epoch[065] Batch [1649]/[3760]	Speed: 35.329953 samples/sec	 accuracy=77.676136	 loss=0.886452	 lr=0.001000
Epoch[065] Batch [1699]/[3760]	Speed: 35.309762 samples/sec	 accuracy=77.660846	 loss=0.886749	 lr=0.001000
Epoch[065] Batch [1749]/[3760]	Speed: 35.163826 samples/sec	 accuracy=77.651786	 loss=0.886452	 lr=0.001000
Epoch[065] Batch [1799]/[3760]	Speed: 35.207145 samples/sec	 accuracy=77.640625	 loss=0.886242	 lr=0.001000
Epoch[065] Batch [1849]/[3760]	Speed: 35.097785 samples/sec	 accuracy=77.653716	 loss=0.885944	 lr=0.001000
Epoch[065] Batch [1899]/[3760]	Speed: 35.086062 samples/sec	 accuracy=77.666118	 loss=0.885646	 lr=0.001000
Epoch[065] Batch [1949]/[3760]	Speed: 35.247436 samples/sec	 accuracy=77.688301	 loss=0.884469	 lr=0.001000
Epoch[065] Batch [1999]/[3760]	Speed: 35.055693 samples/sec	 accuracy=77.657813	 loss=0.885413	 lr=0.001000
Epoch[065] Batch [2049]/[3760]	Speed: 35.051758 samples/sec	 accuracy=77.661585	 loss=0.885455	 lr=0.001000
Epoch[065] Batch [2099]/[3760]	Speed: 35.131342 samples/sec	 accuracy=77.662946	 loss=0.886140	 lr=0.001000
Epoch[065] Batch [2149]/[3760]	Speed: 35.326015 samples/sec	 accuracy=77.670058	 loss=0.885757	 lr=0.001000
Epoch[065] Batch [2199]/[3760]	Speed: 34.916381 samples/sec	 accuracy=77.656960	 loss=0.885469	 lr=0.001000
Epoch[065] Batch [2249]/[3760]	Speed: 35.378709 samples/sec	 accuracy=77.650000	 loss=0.886029	 lr=0.001000
Epoch[065] Batch [2299]/[3760]	Speed: 34.970045 samples/sec	 accuracy=77.622962	 loss=0.887210	 lr=0.001000
Epoch[065] Batch [2349]/[3760]	Speed: 35.348758 samples/sec	 accuracy=77.623670	 loss=0.887054	 lr=0.001000
Epoch[065] Batch [2399]/[3760]	Speed: 35.134146 samples/sec	 accuracy=77.640625	 loss=0.886566	 lr=0.001000
Epoch[065] Batch [2449]/[3760]	Speed: 35.379245 samples/sec	 accuracy=77.610969	 loss=0.887779	 lr=0.001000
Epoch[065] Batch [2499]/[3760]	Speed: 35.037436 samples/sec	 accuracy=77.625625	 loss=0.887350	 lr=0.001000
Epoch[065] Batch [2549]/[3760]	Speed: 35.337677 samples/sec	 accuracy=77.603554	 loss=0.887684	 lr=0.001000
Epoch[065] Batch [2599]/[3760]	Speed: 34.942564 samples/sec	 accuracy=77.597957	 loss=0.888195	 lr=0.001000
Epoch[065] Batch [2649]/[3760]	Speed: 35.392680 samples/sec	 accuracy=77.594340	 loss=0.889276	 lr=0.001000
Epoch[065] Batch [2699]/[3760]	Speed: 35.133870 samples/sec	 accuracy=77.597801	 loss=0.889687	 lr=0.001000
Epoch[065] Batch [2749]/[3760]	Speed: 35.383840 samples/sec	 accuracy=77.610795	 loss=0.888967	 lr=0.001000
Epoch[065] Batch [2799]/[3760]	Speed: 35.281115 samples/sec	 accuracy=77.637835	 loss=0.888781	 lr=0.001000
Epoch[065] Batch [2849]/[3760]	Speed: 35.420489 samples/sec	 accuracy=77.635417	 loss=0.889119	 lr=0.001000
Epoch[065] Batch [2899]/[3760]	Speed: 35.261242 samples/sec	 accuracy=77.636853	 loss=0.889293	 lr=0.001000
Epoch[065] Batch [2949]/[3760]	Speed: 35.509867 samples/sec	 accuracy=77.611229	 loss=0.889962	 lr=0.001000
Epoch[065] Batch [2999]/[3760]	Speed: 35.412918 samples/sec	 accuracy=77.609896	 loss=0.890311	 lr=0.001000
Epoch[065] Batch [3049]/[3760]	Speed: 35.325873 samples/sec	 accuracy=77.608094	 loss=0.890796	 lr=0.001000
Epoch[065] Batch [3099]/[3760]	Speed: 35.196269 samples/sec	 accuracy=77.612399	 loss=0.890658	 lr=0.001000
Epoch[065] Batch [3149]/[3760]	Speed: 35.314412 samples/sec	 accuracy=77.592758	 loss=0.891373	 lr=0.001000
Epoch[065] Batch [3199]/[3760]	Speed: 35.504954 samples/sec	 accuracy=77.575195	 loss=0.892293	 lr=0.001000
Epoch[065] Batch [3249]/[3760]	Speed: 35.367154 samples/sec	 accuracy=77.566827	 loss=0.892873	 lr=0.001000
Epoch[065] Batch [3299]/[3760]	Speed: 35.173290 samples/sec	 accuracy=77.577178	 loss=0.892524	 lr=0.001000
Epoch[065] Batch [3349]/[3760]	Speed: 35.294470 samples/sec	 accuracy=77.579291	 loss=0.892256	 lr=0.001000
Epoch[065] Batch [3399]/[3760]	Speed: 35.136271 samples/sec	 accuracy=77.566636	 loss=0.892643	 lr=0.001000
Epoch[065] Batch [3449]/[3760]	Speed: 35.331233 samples/sec	 accuracy=77.581975	 loss=0.892030	 lr=0.001000
Epoch[065] Batch [3499]/[3760]	Speed: 35.225890 samples/sec	 accuracy=77.587946	 loss=0.891681	 lr=0.001000
Epoch[065] Batch [3549]/[3760]	Speed: 35.273282 samples/sec	 accuracy=77.584067	 loss=0.891759	 lr=0.001000
Epoch[065] Batch [3599]/[3760]	Speed: 35.219927 samples/sec	 accuracy=77.583333	 loss=0.891954	 lr=0.001000
Epoch[065] Batch [3649]/[3760]	Speed: 35.041991 samples/sec	 accuracy=77.597175	 loss=0.891305	 lr=0.001000
Epoch[065] Batch [3699]/[3760]	Speed: 35.210788 samples/sec	 accuracy=77.584459	 loss=0.891643	 lr=0.001000
Epoch[065] Batch [3749]/[3760]	Speed: 36.163239 samples/sec	 accuracy=77.585417	 loss=0.891603	 lr=0.001000
Batch [0049]/[0303]: acc-top1=67.125000 acc-top5=85.375000
Batch [0099]/[0303]: acc-top1=66.359375 acc-top5=85.062500
Batch [0149]/[0303]: acc-top1=66.750000 acc-top5=85.416667
Batch [0199]/[0303]: acc-top1=66.406250 acc-top5=85.554688
Batch [0249]/[0303]: acc-top1=66.056250 acc-top5=85.456250
Batch [0299]/[0303]: acc-top1=66.000000 acc-top5=85.572917
[Epoch 065] training: accuracy=77.593501	 loss=0.891324
[Epoch 065] speed: 35 samples/sec	time cost: 7320.298661
[Epoch 065] validation: acc-top1=66.001444 acc-top5=85.581683 loss=1.615569
Epoch[066] Batch [0049]/[3759]	Speed: 25.318837 samples/sec	 accuracy=78.093750	 loss=0.864238	 lr=0.001000
Epoch[066] Batch [0099]/[3759]	Speed: 35.235559 samples/sec	 accuracy=78.234375	 loss=0.875618	 lr=0.001000
Epoch[066] Batch [0149]/[3759]	Speed: 35.243384 samples/sec	 accuracy=78.708333	 loss=0.869946	 lr=0.001000
Epoch[066] Batch [0199]/[3759]	Speed: 35.140192 samples/sec	 accuracy=78.320312	 loss=0.878625	 lr=0.001000
Epoch[066] Batch [0249]/[3759]	Speed: 35.304766 samples/sec	 accuracy=78.312500	 loss=0.875404	 lr=0.001000
Epoch[066] Batch [0299]/[3759]	Speed: 35.173710 samples/sec	 accuracy=78.255208	 loss=0.877957	 lr=0.001000
Epoch[066] Batch [0349]/[3759]	Speed: 35.135525 samples/sec	 accuracy=78.245536	 loss=0.876939	 lr=0.001000
Epoch[066] Batch [0399]/[3759]	Speed: 35.130848 samples/sec	 accuracy=78.312500	 loss=0.875770	 lr=0.001000
Epoch[066] Batch [0449]/[3759]	Speed: 34.925003 samples/sec	 accuracy=78.236111	 loss=0.882117	 lr=0.001000
Epoch[066] Batch [0499]/[3759]	Speed: 35.143011 samples/sec	 accuracy=78.321875	 loss=0.879563	 lr=0.001000
Epoch[066] Batch [0549]/[3759]	Speed: 34.872825 samples/sec	 accuracy=78.173295	 loss=0.880004	 lr=0.001000
Epoch[066] Batch [0599]/[3759]	Speed: 35.293810 samples/sec	 accuracy=78.096354	 loss=0.881929	 lr=0.001000
Epoch[066] Batch [0649]/[3759]	Speed: 35.000699 samples/sec	 accuracy=78.084135	 loss=0.882032	 lr=0.001000
Epoch[066] Batch [0699]/[3759]	Speed: 35.153646 samples/sec	 accuracy=78.133929	 loss=0.878161	 lr=0.001000
Epoch[066] Batch [0749]/[3759]	Speed: 34.822312 samples/sec	 accuracy=78.175000	 loss=0.875607	 lr=0.001000
Epoch[066] Batch [0799]/[3759]	Speed: 35.317446 samples/sec	 accuracy=78.189453	 loss=0.875641	 lr=0.001000
Epoch[066] Batch [0849]/[3759]	Speed: 34.951246 samples/sec	 accuracy=78.207721	 loss=0.874051	 lr=0.001000
Epoch[066] Batch [0899]/[3759]	Speed: 35.366598 samples/sec	 accuracy=78.239583	 loss=0.871802	 lr=0.001000
Epoch[066] Batch [0949]/[3759]	Speed: 35.017832 samples/sec	 accuracy=78.261513	 loss=0.871402	 lr=0.001000
Epoch[066] Batch [0999]/[3759]	Speed: 35.309510 samples/sec	 accuracy=78.176563	 loss=0.872976	 lr=0.001000
Epoch[066] Batch [1049]/[3759]	Speed: 34.947909 samples/sec	 accuracy=78.105655	 loss=0.876769	 lr=0.001000
Epoch[066] Batch [1099]/[3759]	Speed: 35.500414 samples/sec	 accuracy=78.076705	 loss=0.877340	 lr=0.001000
Epoch[066] Batch [1149]/[3759]	Speed: 34.933278 samples/sec	 accuracy=78.134511	 loss=0.875934	 lr=0.001000
Epoch[066] Batch [1199]/[3759]	Speed: 35.177817 samples/sec	 accuracy=78.138021	 loss=0.875939	 lr=0.001000
Epoch[066] Batch [1249]/[3759]	Speed: 34.999173 samples/sec	 accuracy=78.093750	 loss=0.877887	 lr=0.001000
Epoch[066] Batch [1299]/[3759]	Speed: 35.196825 samples/sec	 accuracy=78.078125	 loss=0.878562	 lr=0.001000
Epoch[066] Batch [1349]/[3759]	Speed: 34.932603 samples/sec	 accuracy=78.062500	 loss=0.878377	 lr=0.001000
Epoch[066] Batch [1399]/[3759]	Speed: 35.182931 samples/sec	 accuracy=78.092634	 loss=0.876971	 lr=0.001000
Epoch[066] Batch [1449]/[3759]	Speed: 35.042684 samples/sec	 accuracy=78.106681	 loss=0.876725	 lr=0.001000
Epoch[066] Batch [1499]/[3759]	Speed: 35.251913 samples/sec	 accuracy=78.087500	 loss=0.877396	 lr=0.001000
Epoch[066] Batch [1549]/[3759]	Speed: 35.049033 samples/sec	 accuracy=78.092742	 loss=0.877354	 lr=0.001000
Epoch[066] Batch [1599]/[3759]	Speed: 35.215093 samples/sec	 accuracy=78.089844	 loss=0.877452	 lr=0.001000
Epoch[066] Batch [1649]/[3759]	Speed: 35.201242 samples/sec	 accuracy=78.096591	 loss=0.877902	 lr=0.001000
Epoch[066] Batch [1699]/[3759]	Speed: 35.101963 samples/sec	 accuracy=78.111213	 loss=0.877403	 lr=0.001000
Epoch[066] Batch [1749]/[3759]	Speed: 35.237030 samples/sec	 accuracy=78.122321	 loss=0.876852	 lr=0.001000
Epoch[066] Batch [1799]/[3759]	Speed: 35.070183 samples/sec	 accuracy=78.169271	 loss=0.875647	 lr=0.001000
Epoch[066] Batch [1849]/[3759]	Speed: 35.132814 samples/sec	 accuracy=78.165541	 loss=0.875433	 lr=0.001000
Epoch[066] Batch [1899]/[3759]	Speed: 35.267856 samples/sec	 accuracy=78.168586	 loss=0.875519	 lr=0.001000
Epoch[066] Batch [1949]/[3759]	Speed: 35.336359 samples/sec	 accuracy=78.152244	 loss=0.875541	 lr=0.001000
Epoch[066] Batch [1999]/[3759]	Speed: 35.058362 samples/sec	 accuracy=78.139062	 loss=0.875841	 lr=0.001000
Epoch[066] Batch [2049]/[3759]	Speed: 35.263779 samples/sec	 accuracy=78.122713	 loss=0.875794	 lr=0.001000
Epoch[066] Batch [2099]/[3759]	Speed: 35.220708 samples/sec	 accuracy=78.104167	 loss=0.876443	 lr=0.001000
Epoch[066] Batch [2149]/[3759]	Speed: 35.110187 samples/sec	 accuracy=78.109738	 loss=0.876705	 lr=0.001000
Epoch[066] Batch [2199]/[3759]	Speed: 35.181425 samples/sec	 accuracy=78.117898	 loss=0.876411	 lr=0.001000
Epoch[066] Batch [2249]/[3759]	Speed: 35.193451 samples/sec	 accuracy=78.107639	 loss=0.876800	 lr=0.001000
Epoch[066] Batch [2299]/[3759]	Speed: 35.161104 samples/sec	 accuracy=78.105978	 loss=0.877132	 lr=0.001000
Epoch[066] Batch [2349]/[3759]	Speed: 35.200582 samples/sec	 accuracy=78.111702	 loss=0.877028	 lr=0.001000
Epoch[066] Batch [2399]/[3759]	Speed: 35.134945 samples/sec	 accuracy=78.127604	 loss=0.876754	 lr=0.001000
Epoch[066] Batch [2449]/[3759]	Speed: 35.324648 samples/sec	 accuracy=78.109056	 loss=0.876874	 lr=0.001000
Epoch[066] Batch [2499]/[3759]	Speed: 34.980749 samples/sec	 accuracy=78.083750	 loss=0.877589	 lr=0.001000
Epoch[066] Batch [2549]/[3759]	Speed: 35.233671 samples/sec	 accuracy=78.046569	 loss=0.878811	 lr=0.001000
Epoch[066] Batch [2599]/[3759]	Speed: 35.034189 samples/sec	 accuracy=78.020433	 loss=0.879652	 lr=0.001000
Epoch[066] Batch [2649]/[3759]	Speed: 35.316771 samples/sec	 accuracy=78.005896	 loss=0.880307	 lr=0.001000
Epoch[066] Batch [2699]/[3759]	Speed: 35.139919 samples/sec	 accuracy=78.012153	 loss=0.880173	 lr=0.001000
Epoch[066] Batch [2749]/[3759]	Speed: 35.162042 samples/sec	 accuracy=78.008523	 loss=0.880125	 lr=0.001000
Epoch[066] Batch [2799]/[3759]	Speed: 35.052140 samples/sec	 accuracy=77.978237	 loss=0.881502	 lr=0.001000
Epoch[066] Batch [2849]/[3759]	Speed: 35.306827 samples/sec	 accuracy=77.973136	 loss=0.881286	 lr=0.001000
Epoch[066] Batch [2899]/[3759]	Speed: 34.919001 samples/sec	 accuracy=77.957974	 loss=0.882705	 lr=0.001000
Epoch[066] Batch [2949]/[3759]	Speed: 35.038180 samples/sec	 accuracy=77.940148	 loss=0.883487	 lr=0.001000
Epoch[066] Batch [2999]/[3759]	Speed: 34.836961 samples/sec	 accuracy=77.928125	 loss=0.883778	 lr=0.001000
Epoch[066] Batch [3049]/[3759]	Speed: 34.879141 samples/sec	 accuracy=77.911885	 loss=0.884390	 lr=0.001000
Epoch[066] Batch [3099]/[3759]	Speed: 34.824522 samples/sec	 accuracy=77.914315	 loss=0.884065	 lr=0.001000
Epoch[066] Batch [3149]/[3759]	Speed: 34.866826 samples/sec	 accuracy=77.918651	 loss=0.884043	 lr=0.001000
Epoch[066] Batch [3199]/[3759]	Speed: 34.976997 samples/sec	 accuracy=77.900391	 loss=0.884092	 lr=0.001000
Epoch[066] Batch [3249]/[3759]	Speed: 34.724675 samples/sec	 accuracy=77.898077	 loss=0.884201	 lr=0.001000
Epoch[066] Batch [3299]/[3759]	Speed: 35.124174 samples/sec	 accuracy=77.871212	 loss=0.885091	 lr=0.001000
Epoch[066] Batch [3349]/[3759]	Speed: 34.611637 samples/sec	 accuracy=77.868004	 loss=0.885096	 lr=0.001000
Epoch[066] Batch [3399]/[3759]	Speed: 35.282767 samples/sec	 accuracy=77.873621	 loss=0.884677	 lr=0.001000
Epoch[066] Batch [3449]/[3759]	Speed: 34.974300 samples/sec	 accuracy=77.864583	 loss=0.884865	 lr=0.001000
Epoch[066] Batch [3499]/[3759]	Speed: 35.338520 samples/sec	 accuracy=77.848661	 loss=0.885099	 lr=0.001000
Epoch[066] Batch [3549]/[3759]	Speed: 35.110718 samples/sec	 accuracy=77.869278	 loss=0.884306	 lr=0.001000
Epoch[066] Batch [3599]/[3759]	Speed: 35.233916 samples/sec	 accuracy=77.871528	 loss=0.883839	 lr=0.001000
Epoch[066] Batch [3649]/[3759]	Speed: 35.123901 samples/sec	 accuracy=77.872432	 loss=0.883859	 lr=0.001000
Epoch[066] Batch [3699]/[3759]	Speed: 35.170667 samples/sec	 accuracy=77.883024	 loss=0.883555	 lr=0.001000
Epoch[066] Batch [3749]/[3759]	Speed: 36.378269 samples/sec	 accuracy=77.881667	 loss=0.883297	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.531250 acc-top5=85.937500
Batch [0099]/[0303]: acc-top1=65.765625 acc-top5=85.375000
Batch [0149]/[0303]: acc-top1=66.166667 acc-top5=85.468750
Batch [0199]/[0303]: acc-top1=65.992188 acc-top5=85.570312
Batch [0249]/[0303]: acc-top1=65.650000 acc-top5=85.525000
Batch [0299]/[0303]: acc-top1=65.661458 acc-top5=85.578125
[Epoch 066] training: accuracy=77.873936	 loss=0.883638
[Epoch 066] speed: 34 samples/sec	time cost: 7336.672817
[Epoch 066] validation: acc-top1=65.645627 acc-top5=85.612624 loss=1.629424
Epoch[067] Batch [0049]/[3760]	Speed: 25.268924 samples/sec	 accuracy=78.156250	 loss=0.863064	 lr=0.001000
Epoch[067] Batch [0099]/[3760]	Speed: 35.240624 samples/sec	 accuracy=78.625000	 loss=0.848744	 lr=0.001000
Epoch[067] Batch [0149]/[3760]	Speed: 34.951612 samples/sec	 accuracy=78.270833	 loss=0.856968	 lr=0.001000
Epoch[067] Batch [0199]/[3760]	Speed: 35.350259 samples/sec	 accuracy=78.179688	 loss=0.861901	 lr=0.001000
Epoch[067] Batch [0249]/[3760]	Speed: 34.909213 samples/sec	 accuracy=78.350000	 loss=0.857371	 lr=0.001000
Epoch[067] Batch [0299]/[3760]	Speed: 35.256747 samples/sec	 accuracy=78.427083	 loss=0.855080	 lr=0.001000
Epoch[067] Batch [0349]/[3760]	Speed: 35.105223 samples/sec	 accuracy=78.450893	 loss=0.854763	 lr=0.001000
Epoch[067] Batch [0399]/[3760]	Speed: 35.287903 samples/sec	 accuracy=78.492188	 loss=0.857228	 lr=0.001000
Epoch[067] Batch [0449]/[3760]	Speed: 35.017438 samples/sec	 accuracy=78.486111	 loss=0.858090	 lr=0.001000
Epoch[067] Batch [0499]/[3760]	Speed: 35.296659 samples/sec	 accuracy=78.396875	 loss=0.861507	 lr=0.001000
Epoch[067] Batch [0549]/[3760]	Speed: 34.769495 samples/sec	 accuracy=78.227273	 loss=0.866888	 lr=0.001000
Epoch[067] Batch [0599]/[3760]	Speed: 35.246911 samples/sec	 accuracy=78.140625	 loss=0.870372	 lr=0.001000
Epoch[067] Batch [0649]/[3760]	Speed: 34.928438 samples/sec	 accuracy=78.221154	 loss=0.866914	 lr=0.001000
Epoch[067] Batch [0699]/[3760]	Speed: 35.314126 samples/sec	 accuracy=78.212054	 loss=0.864673	 lr=0.001000
Epoch[067] Batch [0749]/[3760]	Speed: 34.727797 samples/sec	 accuracy=78.170833	 loss=0.865965	 lr=0.001000
Epoch[067] Batch [0799]/[3760]	Speed: 35.281659 samples/sec	 accuracy=78.185547	 loss=0.865996	 lr=0.001000
Epoch[067] Batch [0849]/[3760]	Speed: 34.773369 samples/sec	 accuracy=78.156250	 loss=0.867203	 lr=0.001000
Epoch[067] Batch [0899]/[3760]	Speed: 35.212646 samples/sec	 accuracy=78.157986	 loss=0.867303	 lr=0.001000
Epoch[067] Batch [0949]/[3760]	Speed: 34.844601 samples/sec	 accuracy=78.184211	 loss=0.867125	 lr=0.001000
Epoch[067] Batch [0999]/[3760]	Speed: 35.058872 samples/sec	 accuracy=78.206250	 loss=0.866459	 lr=0.001000
Epoch[067] Batch [1049]/[3760]	Speed: 35.139782 samples/sec	 accuracy=78.099702	 loss=0.870526	 lr=0.001000
Epoch[067] Batch [1099]/[3760]	Speed: 34.808413 samples/sec	 accuracy=78.113636	 loss=0.870616	 lr=0.001000
Epoch[067] Batch [1149]/[3760]	Speed: 35.188121 samples/sec	 accuracy=78.101902	 loss=0.871675	 lr=0.001000
Epoch[067] Batch [1199]/[3760]	Speed: 34.752321 samples/sec	 accuracy=78.122396	 loss=0.870968	 lr=0.001000
Epoch[067] Batch [1249]/[3760]	Speed: 35.040803 samples/sec	 accuracy=78.101250	 loss=0.870966	 lr=0.001000
Epoch[067] Batch [1299]/[3760]	Speed: 34.654347 samples/sec	 accuracy=78.074519	 loss=0.872145	 lr=0.001000
Epoch[067] Batch [1349]/[3760]	Speed: 35.180230 samples/sec	 accuracy=78.059028	 loss=0.873304	 lr=0.001000
Epoch[067] Batch [1399]/[3760]	Speed: 34.656075 samples/sec	 accuracy=78.060268	 loss=0.872608	 lr=0.001000
Epoch[067] Batch [1449]/[3760]	Speed: 35.048340 samples/sec	 accuracy=78.093750	 loss=0.871973	 lr=0.001000
Epoch[067] Batch [1499]/[3760]	Speed: 34.855887 samples/sec	 accuracy=78.098958	 loss=0.872088	 lr=0.001000
Epoch[067] Batch [1549]/[3760]	Speed: 35.077693 samples/sec	 accuracy=78.080645	 loss=0.873571	 lr=0.001000
Epoch[067] Batch [1599]/[3760]	Speed: 34.938817 samples/sec	 accuracy=78.104492	 loss=0.872884	 lr=0.001000
Epoch[067] Batch [1649]/[3760]	Speed: 34.936441 samples/sec	 accuracy=78.102273	 loss=0.872475	 lr=0.001000
Epoch[067] Batch [1699]/[3760]	Speed: 35.141658 samples/sec	 accuracy=78.089154	 loss=0.872392	 lr=0.001000
Epoch[067] Batch [1749]/[3760]	Speed: 34.953554 samples/sec	 accuracy=78.121429	 loss=0.872019	 lr=0.001000
Epoch[067] Batch [1799]/[3760]	Speed: 35.238643 samples/sec	 accuracy=78.109375	 loss=0.872101	 lr=0.001000
Epoch[067] Batch [1849]/[3760]	Speed: 34.905781 samples/sec	 accuracy=78.123311	 loss=0.871772	 lr=0.001000
Epoch[067] Batch [1899]/[3760]	Speed: 35.307831 samples/sec	 accuracy=78.119243	 loss=0.872664	 lr=0.001000
Epoch[067] Batch [1949]/[3760]	Speed: 34.879053 samples/sec	 accuracy=78.116186	 loss=0.873158	 lr=0.001000
Epoch[067] Batch [1999]/[3760]	Speed: 35.195691 samples/sec	 accuracy=78.136719	 loss=0.872288	 lr=0.001000
Epoch[067] Batch [2049]/[3760]	Speed: 34.654562 samples/sec	 accuracy=78.099848	 loss=0.873318	 lr=0.001000
Epoch[067] Batch [2099]/[3760]	Speed: 35.143668 samples/sec	 accuracy=78.093006	 loss=0.874517	 lr=0.001000
Epoch[067] Batch [2149]/[3760]	Speed: 34.482655 samples/sec	 accuracy=78.065407	 loss=0.875971	 lr=0.001000
Epoch[067] Batch [2199]/[3760]	Speed: 35.310315 samples/sec	 accuracy=78.071733	 loss=0.876209	 lr=0.001000
Epoch[067] Batch [2249]/[3760]	Speed: 34.727553 samples/sec	 accuracy=78.086111	 loss=0.875755	 lr=0.001000
Epoch[067] Batch [2299]/[3760]	Speed: 34.914610 samples/sec	 accuracy=78.067255	 loss=0.876481	 lr=0.001000
Epoch[067] Batch [2349]/[3760]	Speed: 34.890663 samples/sec	 accuracy=78.061170	 loss=0.876649	 lr=0.001000
Epoch[067] Batch [2399]/[3760]	Speed: 34.725549 samples/sec	 accuracy=78.070964	 loss=0.876953	 lr=0.001000
Epoch[067] Batch [2449]/[3760]	Speed: 35.203334 samples/sec	 accuracy=78.059311	 loss=0.877369	 lr=0.001000
Epoch[067] Batch [2499]/[3760]	Speed: 34.715994 samples/sec	 accuracy=78.047500	 loss=0.877685	 lr=0.001000
Epoch[067] Batch [2549]/[3760]	Speed: 35.141217 samples/sec	 accuracy=78.072304	 loss=0.876892	 lr=0.001000
Epoch[067] Batch [2599]/[3760]	Speed: 34.356393 samples/sec	 accuracy=78.060697	 loss=0.877136	 lr=0.001000
Epoch[067] Batch [2649]/[3760]	Speed: 35.337987 samples/sec	 accuracy=78.039505	 loss=0.878195	 lr=0.001000
Epoch[067] Batch [2699]/[3760]	Speed: 34.543128 samples/sec	 accuracy=78.034144	 loss=0.877728	 lr=0.001000
Epoch[067] Batch [2749]/[3760]	Speed: 35.095058 samples/sec	 accuracy=78.038636	 loss=0.877910	 lr=0.001000
Epoch[067] Batch [2799]/[3760]	Speed: 34.718199 samples/sec	 accuracy=78.051339	 loss=0.877491	 lr=0.001000
Epoch[067] Batch [2849]/[3760]	Speed: 34.893040 samples/sec	 accuracy=78.029057	 loss=0.878294	 lr=0.001000
Epoch[067] Batch [2899]/[3760]	Speed: 34.723119 samples/sec	 accuracy=78.009159	 loss=0.878332	 lr=0.001000
Epoch[067] Batch [2949]/[3760]	Speed: 34.504604 samples/sec	 accuracy=78.012182	 loss=0.878366	 lr=0.001000
Epoch[067] Batch [2999]/[3760]	Speed: 35.056433 samples/sec	 accuracy=78.021875	 loss=0.878088	 lr=0.001000
Epoch[067] Batch [3049]/[3760]	Speed: 34.228665 samples/sec	 accuracy=78.024078	 loss=0.878480	 lr=0.001000
Epoch[067] Batch [3099]/[3760]	Speed: 35.125825 samples/sec	 accuracy=78.018145	 loss=0.878657	 lr=0.001000
Epoch[067] Batch [3149]/[3760]	Speed: 34.417105 samples/sec	 accuracy=78.011905	 loss=0.878555	 lr=0.001000
Epoch[067] Batch [3199]/[3760]	Speed: 35.033498 samples/sec	 accuracy=78.017578	 loss=0.878389	 lr=0.001000
Epoch[067] Batch [3249]/[3760]	Speed: 34.508397 samples/sec	 accuracy=78.013942	 loss=0.878523	 lr=0.001000
Epoch[067] Batch [3299]/[3760]	Speed: 34.602007 samples/sec	 accuracy=78.026042	 loss=0.877814	 lr=0.001000
Epoch[067] Batch [3349]/[3760]	Speed: 34.903295 samples/sec	 accuracy=78.029384	 loss=0.877815	 lr=0.001000
Epoch[067] Batch [3399]/[3760]	Speed: 34.303348 samples/sec	 accuracy=78.033548	 loss=0.877601	 lr=0.001000
Epoch[067] Batch [3449]/[3760]	Speed: 35.318478 samples/sec	 accuracy=78.015399	 loss=0.878263	 lr=0.001000
Epoch[067] Batch [3499]/[3760]	Speed: 34.432011 samples/sec	 accuracy=78.012054	 loss=0.878094	 lr=0.001000
Epoch[067] Batch [3549]/[3760]	Speed: 35.149542 samples/sec	 accuracy=77.999560	 loss=0.878660	 lr=0.001000
Epoch[067] Batch [3599]/[3760]	Speed: 34.520872 samples/sec	 accuracy=77.993056	 loss=0.878856	 lr=0.001000
Epoch[067] Batch [3649]/[3760]	Speed: 35.136707 samples/sec	 accuracy=77.979452	 loss=0.879197	 lr=0.001000
Epoch[067] Batch [3699]/[3760]	Speed: 34.843963 samples/sec	 accuracy=77.986064	 loss=0.879179	 lr=0.001000
Epoch[067] Batch [3749]/[3760]	Speed: 36.226199 samples/sec	 accuracy=77.990833	 loss=0.878984	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.906250 acc-top5=85.562500
Batch [0099]/[0303]: acc-top1=66.093750 acc-top5=85.187500
Batch [0149]/[0303]: acc-top1=66.656250 acc-top5=85.489583
Batch [0199]/[0303]: acc-top1=66.218750 acc-top5=85.515625
Batch [0249]/[0303]: acc-top1=65.900000 acc-top5=85.337500
Batch [0299]/[0303]: acc-top1=65.760417 acc-top5=85.385417
[Epoch 067] training: accuracy=77.989943	 loss=0.879047
[Epoch 067] speed: 34 samples/sec	time cost: 7373.903295
[Epoch 067] validation: acc-top1=65.774546 acc-top5=85.385726 loss=1.620240
Epoch[068] Batch [0049]/[3760]	Speed: 25.393474 samples/sec	 accuracy=78.468750	 loss=0.856957	 lr=0.001000
Epoch[068] Batch [0099]/[3760]	Speed: 35.310499 samples/sec	 accuracy=78.140625	 loss=0.868865	 lr=0.001000
Epoch[068] Batch [0149]/[3760]	Speed: 35.456188 samples/sec	 accuracy=78.552083	 loss=0.855611	 lr=0.001000
Epoch[068] Batch [0199]/[3760]	Speed: 35.323749 samples/sec	 accuracy=78.515625	 loss=0.861414	 lr=0.001000
Epoch[068] Batch [0249]/[3760]	Speed: 35.289252 samples/sec	 accuracy=78.668750	 loss=0.859533	 lr=0.001000
Epoch[068] Batch [0299]/[3760]	Speed: 35.308929 samples/sec	 accuracy=78.505208	 loss=0.866321	 lr=0.001000
Epoch[068] Batch [0349]/[3760]	Speed: 35.248448 samples/sec	 accuracy=78.450893	 loss=0.865530	 lr=0.001000
Epoch[068] Batch [0399]/[3760]	Speed: 35.224196 samples/sec	 accuracy=78.671875	 loss=0.861584	 lr=0.001000
Epoch[068] Batch [0449]/[3760]	Speed: 35.251052 samples/sec	 accuracy=78.659722	 loss=0.860400	 lr=0.001000
Epoch[068] Batch [0499]/[3760]	Speed: 35.460139 samples/sec	 accuracy=78.721875	 loss=0.858836	 lr=0.001000
Epoch[068] Batch [0549]/[3760]	Speed: 35.374161 samples/sec	 accuracy=78.755682	 loss=0.861287	 lr=0.001000
Epoch[068] Batch [0599]/[3760]	Speed: 35.357635 samples/sec	 accuracy=78.838542	 loss=0.857646	 lr=0.001000
Epoch[068] Batch [0649]/[3760]	Speed: 35.342771 samples/sec	 accuracy=78.689904	 loss=0.862853	 lr=0.001000
Epoch[068] Batch [0699]/[3760]	Speed: 35.244529 samples/sec	 accuracy=78.669643	 loss=0.863919	 lr=0.001000
Epoch[068] Batch [0749]/[3760]	Speed: 35.333544 samples/sec	 accuracy=78.668750	 loss=0.862863	 lr=0.001000
Epoch[068] Batch [0799]/[3760]	Speed: 35.357811 samples/sec	 accuracy=78.591797	 loss=0.866655	 lr=0.001000
Epoch[068] Batch [0849]/[3760]	Speed: 35.063780 samples/sec	 accuracy=78.492647	 loss=0.868104	 lr=0.001000
Epoch[068] Batch [0899]/[3760]	Speed: 35.442813 samples/sec	 accuracy=78.503472	 loss=0.867172	 lr=0.001000
Epoch[068] Batch [0949]/[3760]	Speed: 35.130602 samples/sec	 accuracy=78.483553	 loss=0.866539	 lr=0.001000
Epoch[068] Batch [0999]/[3760]	Speed: 35.288053 samples/sec	 accuracy=78.453125	 loss=0.867933	 lr=0.001000
Epoch[068] Batch [1049]/[3760]	Speed: 35.403188 samples/sec	 accuracy=78.452381	 loss=0.867635	 lr=0.001000
Epoch[068] Batch [1099]/[3760]	Speed: 35.343788 samples/sec	 accuracy=78.427557	 loss=0.867409	 lr=0.001000
Epoch[068] Batch [1149]/[3760]	Speed: 35.213083 samples/sec	 accuracy=78.433424	 loss=0.866622	 lr=0.001000
Epoch[068] Batch [1199]/[3760]	Speed: 35.272021 samples/sec	 accuracy=78.446615	 loss=0.865095	 lr=0.001000
Epoch[068] Batch [1249]/[3760]	Speed: 35.320894 samples/sec	 accuracy=78.475000	 loss=0.864094	 lr=0.001000
Epoch[068] Batch [1299]/[3760]	Speed: 35.358631 samples/sec	 accuracy=78.411058	 loss=0.866672	 lr=0.001000
Epoch[068] Batch [1349]/[3760]	Speed: 35.356684 samples/sec	 accuracy=78.402778	 loss=0.867439	 lr=0.001000
Epoch[068] Batch [1399]/[3760]	Speed: 35.420552 samples/sec	 accuracy=78.405134	 loss=0.867609	 lr=0.001000
Epoch[068] Batch [1449]/[3760]	Speed: 35.345864 samples/sec	 accuracy=78.434267	 loss=0.866366	 lr=0.001000
Epoch[068] Batch [1499]/[3760]	Speed: 35.175177 samples/sec	 accuracy=78.419792	 loss=0.867222	 lr=0.001000
Epoch[068] Batch [1549]/[3760]	Speed: 35.095613 samples/sec	 accuracy=78.414315	 loss=0.866895	 lr=0.001000
Epoch[068] Batch [1599]/[3760]	Speed: 35.362604 samples/sec	 accuracy=78.393555	 loss=0.867509	 lr=0.001000
Epoch[068] Batch [1649]/[3760]	Speed: 35.030591 samples/sec	 accuracy=78.386364	 loss=0.867728	 lr=0.001000
Epoch[068] Batch [1699]/[3760]	Speed: 35.506453 samples/sec	 accuracy=78.354779	 loss=0.869097	 lr=0.001000
Epoch[068] Batch [1749]/[3760]	Speed: 35.044146 samples/sec	 accuracy=78.370536	 loss=0.867967	 lr=0.001000
Epoch[068] Batch [1799]/[3760]	Speed: 35.210972 samples/sec	 accuracy=78.372396	 loss=0.867906	 lr=0.001000
Epoch[068] Batch [1849]/[3760]	Speed: 35.151341 samples/sec	 accuracy=78.359797	 loss=0.868440	 lr=0.001000
Epoch[068] Batch [1899]/[3760]	Speed: 35.294601 samples/sec	 accuracy=78.384046	 loss=0.868225	 lr=0.001000
Epoch[068] Batch [1949]/[3760]	Speed: 35.266747 samples/sec	 accuracy=78.379808	 loss=0.868555	 lr=0.001000
Epoch[068] Batch [1999]/[3760]	Speed: 35.108564 samples/sec	 accuracy=78.385937	 loss=0.868046	 lr=0.001000
Epoch[068] Batch [2049]/[3760]	Speed: 35.309792 samples/sec	 accuracy=78.368902	 loss=0.868037	 lr=0.001000
Epoch[068] Batch [2099]/[3760]	Speed: 35.245773 samples/sec	 accuracy=78.353423	 loss=0.867527	 lr=0.001000
Epoch[068] Batch [2149]/[3760]	Speed: 35.378895 samples/sec	 accuracy=78.349564	 loss=0.868051	 lr=0.001000
Epoch[068] Batch [2199]/[3760]	Speed: 35.193164 samples/sec	 accuracy=78.330966	 loss=0.868882	 lr=0.001000
Epoch[068] Batch [2249]/[3760]	Speed: 35.436696 samples/sec	 accuracy=78.334722	 loss=0.869082	 lr=0.001000
Epoch[068] Batch [2299]/[3760]	Speed: 35.075927 samples/sec	 accuracy=78.327446	 loss=0.868739	 lr=0.001000
Epoch[068] Batch [2349]/[3760]	Speed: 35.433348 samples/sec	 accuracy=78.308511	 loss=0.869611	 lr=0.001000
Epoch[068] Batch [2399]/[3760]	Speed: 35.074374 samples/sec	 accuracy=78.291016	 loss=0.870012	 lr=0.001000
Epoch[068] Batch [2449]/[3760]	Speed: 35.442789 samples/sec	 accuracy=78.280612	 loss=0.870673	 lr=0.001000
Epoch[068] Batch [2499]/[3760]	Speed: 35.250085 samples/sec	 accuracy=78.280000	 loss=0.870236	 lr=0.001000
Epoch[068] Batch [2549]/[3760]	Speed: 35.420867 samples/sec	 accuracy=78.276348	 loss=0.870380	 lr=0.001000
Epoch[068] Batch [2599]/[3760]	Speed: 35.164711 samples/sec	 accuracy=78.254808	 loss=0.871017	 lr=0.001000
Epoch[068] Batch [2649]/[3760]	Speed: 35.528409 samples/sec	 accuracy=78.243514	 loss=0.871323	 lr=0.001000
Epoch[068] Batch [2699]/[3760]	Speed: 34.952871 samples/sec	 accuracy=78.212384	 loss=0.872643	 lr=0.001000
Epoch[068] Batch [2749]/[3760]	Speed: 35.394298 samples/sec	 accuracy=78.222727	 loss=0.872401	 lr=0.001000
Epoch[068] Batch [2799]/[3760]	Speed: 35.081452 samples/sec	 accuracy=78.225446	 loss=0.872322	 lr=0.001000
Epoch[068] Batch [2849]/[3760]	Speed: 35.334374 samples/sec	 accuracy=78.218202	 loss=0.872631	 lr=0.001000
Epoch[068] Batch [2899]/[3760]	Speed: 35.017196 samples/sec	 accuracy=78.214978	 loss=0.872500	 lr=0.001000
Epoch[068] Batch [2949]/[3760]	Speed: 35.290994 samples/sec	 accuracy=78.228284	 loss=0.872596	 lr=0.001000
Epoch[068] Batch [2999]/[3760]	Speed: 35.275558 samples/sec	 accuracy=78.225000	 loss=0.872443	 lr=0.001000
Epoch[068] Batch [3049]/[3760]	Speed: 34.860116 samples/sec	 accuracy=78.222848	 loss=0.872355	 lr=0.001000
Epoch[068] Batch [3099]/[3760]	Speed: 35.270562 samples/sec	 accuracy=78.230847	 loss=0.871954	 lr=0.001000
Epoch[068] Batch [3149]/[3760]	Speed: 34.718890 samples/sec	 accuracy=78.215774	 loss=0.872494	 lr=0.001000
Epoch[068] Batch [3199]/[3760]	Speed: 35.399365 samples/sec	 accuracy=78.204102	 loss=0.872342	 lr=0.001000
Epoch[068] Batch [3249]/[3760]	Speed: 34.998868 samples/sec	 accuracy=78.207212	 loss=0.872049	 lr=0.001000
Epoch[068] Batch [3299]/[3760]	Speed: 35.334673 samples/sec	 accuracy=78.205966	 loss=0.871692	 lr=0.001000
Epoch[068] Batch [3349]/[3760]	Speed: 35.323488 samples/sec	 accuracy=78.195896	 loss=0.871648	 lr=0.001000
Epoch[068] Batch [3399]/[3760]	Speed: 35.093637 samples/sec	 accuracy=78.192555	 loss=0.871440	 lr=0.001000
Epoch[068] Batch [3449]/[3760]	Speed: 35.086792 samples/sec	 accuracy=78.200181	 loss=0.871288	 lr=0.001000
Epoch[068] Batch [3499]/[3760]	Speed: 35.120626 samples/sec	 accuracy=78.182143	 loss=0.871717	 lr=0.001000
Epoch[068] Batch [3549]/[3760]	Speed: 35.366396 samples/sec	 accuracy=78.172535	 loss=0.872010	 lr=0.001000
Epoch[068] Batch [3599]/[3760]	Speed: 35.127179 samples/sec	 accuracy=78.180990	 loss=0.871983	 lr=0.001000
Epoch[068] Batch [3649]/[3760]	Speed: 35.468036 samples/sec	 accuracy=78.179795	 loss=0.871330	 lr=0.001000
Epoch[068] Batch [3699]/[3760]	Speed: 35.116420 samples/sec	 accuracy=78.178632	 loss=0.871518	 lr=0.001000
Epoch[068] Batch [3749]/[3760]	Speed: 36.399955 samples/sec	 accuracy=78.166667	 loss=0.871851	 lr=0.001000
Batch [0049]/[0303]: acc-top1=67.000000 acc-top5=85.750000
Batch [0099]/[0303]: acc-top1=66.234375 acc-top5=85.343750
Batch [0149]/[0303]: acc-top1=66.750000 acc-top5=85.531250
Batch [0199]/[0303]: acc-top1=66.578125 acc-top5=85.500000
Batch [0249]/[0303]: acc-top1=66.262500 acc-top5=85.343750
Batch [0299]/[0303]: acc-top1=66.166667 acc-top5=85.463542
[Epoch 068] training: accuracy=78.168634	 loss=0.871854
[Epoch 068] speed: 35 samples/sec	time cost: 7311.859176
[Epoch 068] validation: acc-top1=66.161304 acc-top5=85.483705 loss=1.604769
Epoch[069] Batch [0049]/[3759]	Speed: 25.170079 samples/sec	 accuracy=78.093750	 loss=0.848289	 lr=0.001000
Epoch[069] Batch [0099]/[3759]	Speed: 35.020197 samples/sec	 accuracy=78.375000	 loss=0.847546	 lr=0.001000
Epoch[069] Batch [0149]/[3759]	Speed: 35.331867 samples/sec	 accuracy=78.312500	 loss=0.852900	 lr=0.001000
Epoch[069] Batch [0199]/[3759]	Speed: 35.162719 samples/sec	 accuracy=78.492188	 loss=0.852591	 lr=0.001000
Epoch[069] Batch [0249]/[3759]	Speed: 35.212995 samples/sec	 accuracy=78.493750	 loss=0.848536	 lr=0.001000
Epoch[069] Batch [0299]/[3759]	Speed: 35.126637 samples/sec	 accuracy=78.682292	 loss=0.844219	 lr=0.001000
Epoch[069] Batch [0349]/[3759]	Speed: 35.159626 samples/sec	 accuracy=78.571429	 loss=0.849362	 lr=0.001000
Epoch[069] Batch [0399]/[3759]	Speed: 34.928866 samples/sec	 accuracy=78.511719	 loss=0.850596	 lr=0.001000
Epoch[069] Batch [0449]/[3759]	Speed: 35.083143 samples/sec	 accuracy=78.385417	 loss=0.853104	 lr=0.001000
Epoch[069] Batch [0499]/[3759]	Speed: 35.298254 samples/sec	 accuracy=78.290625	 loss=0.854484	 lr=0.001000
Epoch[069] Batch [0549]/[3759]	Speed: 35.213055 samples/sec	 accuracy=78.375000	 loss=0.853941	 lr=0.001000
Epoch[069] Batch [0599]/[3759]	Speed: 35.280675 samples/sec	 accuracy=78.382812	 loss=0.853072	 lr=0.001000
Epoch[069] Batch [0649]/[3759]	Speed: 35.044205 samples/sec	 accuracy=78.358173	 loss=0.853175	 lr=0.001000
Epoch[069] Batch [0699]/[3759]	Speed: 35.277908 samples/sec	 accuracy=78.386161	 loss=0.852697	 lr=0.001000
Epoch[069] Batch [0749]/[3759]	Speed: 35.032346 samples/sec	 accuracy=78.450000	 loss=0.852212	 lr=0.001000
Epoch[069] Batch [0799]/[3759]	Speed: 35.197640 samples/sec	 accuracy=78.507812	 loss=0.851197	 lr=0.001000
Epoch[069] Batch [0849]/[3759]	Speed: 34.972962 samples/sec	 accuracy=78.555147	 loss=0.850708	 lr=0.001000
Epoch[069] Batch [0899]/[3759]	Speed: 35.307687 samples/sec	 accuracy=78.506944	 loss=0.852897	 lr=0.001000
Epoch[069] Batch [0949]/[3759]	Speed: 34.932779 samples/sec	 accuracy=78.485197	 loss=0.853386	 lr=0.001000
Epoch[069] Batch [0999]/[3759]	Speed: 35.377774 samples/sec	 accuracy=78.515625	 loss=0.853126	 lr=0.001000
Epoch[069] Batch [1049]/[3759]	Speed: 34.892422 samples/sec	 accuracy=78.525298	 loss=0.854699	 lr=0.001000
Epoch[069] Batch [1099]/[3759]	Speed: 35.106532 samples/sec	 accuracy=78.535511	 loss=0.855511	 lr=0.001000
Epoch[069] Batch [1149]/[3759]	Speed: 34.878980 samples/sec	 accuracy=78.569293	 loss=0.854823	 lr=0.001000
Epoch[069] Batch [1199]/[3759]	Speed: 35.395262 samples/sec	 accuracy=78.513021	 loss=0.855227	 lr=0.001000
Epoch[069] Batch [1249]/[3759]	Speed: 35.096231 samples/sec	 accuracy=78.516250	 loss=0.854806	 lr=0.001000
Epoch[069] Batch [1299]/[3759]	Speed: 35.127440 samples/sec	 accuracy=78.519231	 loss=0.855468	 lr=0.001000
Epoch[069] Batch [1349]/[3759]	Speed: 35.098188 samples/sec	 accuracy=78.519676	 loss=0.854908	 lr=0.001000
Epoch[069] Batch [1399]/[3759]	Speed: 35.261141 samples/sec	 accuracy=78.537946	 loss=0.854926	 lr=0.001000
Epoch[069] Batch [1449]/[3759]	Speed: 35.011727 samples/sec	 accuracy=78.506466	 loss=0.856300	 lr=0.001000
Epoch[069] Batch [1499]/[3759]	Speed: 35.243062 samples/sec	 accuracy=78.486458	 loss=0.856590	 lr=0.001000
Epoch[069] Batch [1549]/[3759]	Speed: 34.959744 samples/sec	 accuracy=78.473790	 loss=0.858122	 lr=0.001000
Epoch[069] Batch [1599]/[3759]	Speed: 35.304577 samples/sec	 accuracy=78.468750	 loss=0.858027	 lr=0.001000
Epoch[069] Batch [1649]/[3759]	Speed: 34.913738 samples/sec	 accuracy=78.426136	 loss=0.859347	 lr=0.001000
Epoch[069] Batch [1699]/[3759]	Speed: 35.323812 samples/sec	 accuracy=78.401654	 loss=0.859522	 lr=0.001000
Epoch[069] Batch [1749]/[3759]	Speed: 35.041871 samples/sec	 accuracy=78.428571	 loss=0.858604	 lr=0.001000
Epoch[069] Batch [1799]/[3759]	Speed: 35.223792 samples/sec	 accuracy=78.393229	 loss=0.860110	 lr=0.001000
Epoch[069] Batch [1849]/[3759]	Speed: 35.080615 samples/sec	 accuracy=78.393581	 loss=0.859904	 lr=0.001000
Epoch[069] Batch [1899]/[3759]	Speed: 35.066558 samples/sec	 accuracy=78.370066	 loss=0.860909	 lr=0.001000
Epoch[069] Batch [1949]/[3759]	Speed: 34.862911 samples/sec	 accuracy=78.375000	 loss=0.860879	 lr=0.001000
Epoch[069] Batch [1999]/[3759]	Speed: 35.158936 samples/sec	 accuracy=78.357812	 loss=0.861339	 lr=0.001000
Epoch[069] Batch [2049]/[3759]	Speed: 34.977759 samples/sec	 accuracy=78.356707	 loss=0.861529	 lr=0.001000
Epoch[069] Batch [2099]/[3759]	Speed: 35.153305 samples/sec	 accuracy=78.368304	 loss=0.861436	 lr=0.001000
Epoch[069] Batch [2149]/[3759]	Speed: 34.849368 samples/sec	 accuracy=78.379360	 loss=0.860665	 lr=0.001000
Epoch[069] Batch [2199]/[3759]	Speed: 35.083244 samples/sec	 accuracy=78.357955	 loss=0.860928	 lr=0.001000
Epoch[069] Batch [2249]/[3759]	Speed: 34.641636 samples/sec	 accuracy=78.354861	 loss=0.861193	 lr=0.001000
Epoch[069] Batch [2299]/[3759]	Speed: 35.207542 samples/sec	 accuracy=78.354620	 loss=0.861270	 lr=0.001000
Epoch[069] Batch [2349]/[3759]	Speed: 34.773197 samples/sec	 accuracy=78.379654	 loss=0.860847	 lr=0.001000
Epoch[069] Batch [2399]/[3759]	Speed: 35.119398 samples/sec	 accuracy=78.388021	 loss=0.860822	 lr=0.001000
Epoch[069] Batch [2449]/[3759]	Speed: 35.087953 samples/sec	 accuracy=78.380102	 loss=0.861253	 lr=0.001000
Epoch[069] Batch [2499]/[3759]	Speed: 34.968434 samples/sec	 accuracy=78.368750	 loss=0.861562	 lr=0.001000
Epoch[069] Batch [2549]/[3759]	Speed: 35.106393 samples/sec	 accuracy=78.357230	 loss=0.861927	 lr=0.001000
Epoch[069] Batch [2599]/[3759]	Speed: 35.023272 samples/sec	 accuracy=78.366587	 loss=0.861618	 lr=0.001000
Epoch[069] Batch [2649]/[3759]	Speed: 35.092856 samples/sec	 accuracy=78.373231	 loss=0.861439	 lr=0.001000
Epoch[069] Batch [2699]/[3759]	Speed: 34.864753 samples/sec	 accuracy=78.370949	 loss=0.861758	 lr=0.001000
Epoch[069] Batch [2749]/[3759]	Speed: 35.130383 samples/sec	 accuracy=78.378409	 loss=0.861967	 lr=0.001000
Epoch[069] Batch [2799]/[3759]	Speed: 34.825304 samples/sec	 accuracy=78.362165	 loss=0.863213	 lr=0.001000
Epoch[069] Batch [2849]/[3759]	Speed: 35.152489 samples/sec	 accuracy=78.347588	 loss=0.863737	 lr=0.001000
Epoch[069] Batch [2899]/[3759]	Speed: 34.858231 samples/sec	 accuracy=78.341595	 loss=0.863952	 lr=0.001000
Epoch[069] Batch [2949]/[3759]	Speed: 35.139172 samples/sec	 accuracy=78.343750	 loss=0.863762	 lr=0.001000
Epoch[069] Batch [2999]/[3759]	Speed: 34.702102 samples/sec	 accuracy=78.336458	 loss=0.863878	 lr=0.001000
Epoch[069] Batch [3049]/[3759]	Speed: 35.346086 samples/sec	 accuracy=78.330430	 loss=0.864374	 lr=0.001000
Epoch[069] Batch [3099]/[3759]	Speed: 34.942017 samples/sec	 accuracy=78.320060	 loss=0.865248	 lr=0.001000
Epoch[069] Batch [3149]/[3759]	Speed: 35.220549 samples/sec	 accuracy=78.333333	 loss=0.864468	 lr=0.001000
Epoch[069] Batch [3199]/[3759]	Speed: 34.842447 samples/sec	 accuracy=78.334961	 loss=0.864923	 lr=0.001000
Epoch[069] Batch [3249]/[3759]	Speed: 35.122618 samples/sec	 accuracy=78.337019	 loss=0.865078	 lr=0.001000
Epoch[069] Batch [3299]/[3759]	Speed: 34.870708 samples/sec	 accuracy=78.342803	 loss=0.864653	 lr=0.001000
Epoch[069] Batch [3349]/[3759]	Speed: 35.119656 samples/sec	 accuracy=78.316698	 loss=0.865462	 lr=0.001000
Epoch[069] Batch [3399]/[3759]	Speed: 35.042416 samples/sec	 accuracy=78.315257	 loss=0.865718	 lr=0.001000
Epoch[069] Batch [3449]/[3759]	Speed: 35.060648 samples/sec	 accuracy=78.332428	 loss=0.865271	 lr=0.001000
Epoch[069] Batch [3499]/[3759]	Speed: 35.136662 samples/sec	 accuracy=78.333482	 loss=0.865474	 lr=0.001000
Epoch[069] Batch [3549]/[3759]	Speed: 35.093142 samples/sec	 accuracy=78.349032	 loss=0.865209	 lr=0.001000
Epoch[069] Batch [3599]/[3759]	Speed: 35.227387 samples/sec	 accuracy=78.337240	 loss=0.865350	 lr=0.001000
Epoch[069] Batch [3649]/[3759]	Speed: 35.210245 samples/sec	 accuracy=78.354024	 loss=0.865171	 lr=0.001000
Epoch[069] Batch [3699]/[3759]	Speed: 35.043132 samples/sec	 accuracy=78.372044	 loss=0.864333	 lr=0.001000
Epoch[069] Batch [3749]/[3759]	Speed: 36.335351 samples/sec	 accuracy=78.371667	 loss=0.864284	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.500000 acc-top5=86.312500
Batch [0099]/[0303]: acc-top1=65.984375 acc-top5=85.578125
Batch [0149]/[0303]: acc-top1=66.541667 acc-top5=85.625000
Batch [0199]/[0303]: acc-top1=66.359375 acc-top5=85.812500
Batch [0249]/[0303]: acc-top1=66.068750 acc-top5=85.687500
Batch [0299]/[0303]: acc-top1=66.010417 acc-top5=85.729167
[Epoch 069] training: accuracy=78.361931	 loss=0.864588
[Epoch 069] speed: 34 samples/sec	time cost: 7342.662124
[Epoch 069] validation: acc-top1=66.011757 acc-top5=85.746700 loss=1.630860
Epoch[070] Batch [0049]/[3760]	Speed: 25.093251 samples/sec	 accuracy=77.812500	 loss=0.872394	 lr=0.001000
Epoch[070] Batch [0099]/[3760]	Speed: 35.306724 samples/sec	 accuracy=77.921875	 loss=0.878946	 lr=0.001000
Epoch[070] Batch [0149]/[3760]	Speed: 35.381674 samples/sec	 accuracy=78.291667	 loss=0.864102	 lr=0.001000
Epoch[070] Batch [0199]/[3760]	Speed: 35.126308 samples/sec	 accuracy=78.531250	 loss=0.854114	 lr=0.001000
Epoch[070] Batch [0249]/[3760]	Speed: 35.139861 samples/sec	 accuracy=78.668750	 loss=0.849088	 lr=0.001000
Epoch[070] Batch [0299]/[3760]	Speed: 35.212599 samples/sec	 accuracy=78.802083	 loss=0.844703	 lr=0.001000
Epoch[070] Batch [0349]/[3760]	Speed: 34.928992 samples/sec	 accuracy=78.825893	 loss=0.848236	 lr=0.001000
Epoch[070] Batch [0399]/[3760]	Speed: 35.054758 samples/sec	 accuracy=78.898438	 loss=0.842616	 lr=0.001000
Epoch[070] Batch [0449]/[3760]	Speed: 35.101131 samples/sec	 accuracy=78.847222	 loss=0.843000	 lr=0.001000
Epoch[070] Batch [0499]/[3760]	Speed: 35.147141 samples/sec	 accuracy=78.834375	 loss=0.844783	 lr=0.001000
Epoch[070] Batch [0549]/[3760]	Speed: 34.954445 samples/sec	 accuracy=78.698864	 loss=0.847024	 lr=0.001000
Epoch[070] Batch [0599]/[3760]	Speed: 35.184027 samples/sec	 accuracy=78.710938	 loss=0.846716	 lr=0.001000
Epoch[070] Batch [0649]/[3760]	Speed: 34.648372 samples/sec	 accuracy=78.721154	 loss=0.847118	 lr=0.001000
Epoch[070] Batch [0699]/[3760]	Speed: 35.354226 samples/sec	 accuracy=78.741071	 loss=0.847331	 lr=0.001000
Epoch[070] Batch [0749]/[3760]	Speed: 34.819652 samples/sec	 accuracy=78.725000	 loss=0.846814	 lr=0.001000
Epoch[070] Batch [0799]/[3760]	Speed: 35.064982 samples/sec	 accuracy=78.695312	 loss=0.849477	 lr=0.001000
Epoch[070] Batch [0849]/[3760]	Speed: 34.995998 samples/sec	 accuracy=78.648897	 loss=0.852841	 lr=0.001000
Epoch[070] Batch [0899]/[3760]	Speed: 34.947708 samples/sec	 accuracy=78.666667	 loss=0.853887	 lr=0.001000
Epoch[070] Batch [0949]/[3760]	Speed: 35.217830 samples/sec	 accuracy=78.593750	 loss=0.856433	 lr=0.001000
Epoch[070] Batch [0999]/[3760]	Speed: 34.821557 samples/sec	 accuracy=78.560938	 loss=0.857730	 lr=0.001000
Epoch[070] Batch [1049]/[3760]	Speed: 35.389909 samples/sec	 accuracy=78.578869	 loss=0.856047	 lr=0.001000
Epoch[070] Batch [1099]/[3760]	Speed: 34.855724 samples/sec	 accuracy=78.572443	 loss=0.856114	 lr=0.001000
Epoch[070] Batch [1149]/[3760]	Speed: 35.443610 samples/sec	 accuracy=78.616848	 loss=0.855012	 lr=0.001000
Epoch[070] Batch [1199]/[3760]	Speed: 34.883019 samples/sec	 accuracy=78.582031	 loss=0.855553	 lr=0.001000
Epoch[070] Batch [1249]/[3760]	Speed: 35.198202 samples/sec	 accuracy=78.587500	 loss=0.855389	 lr=0.001000
Epoch[070] Batch [1299]/[3760]	Speed: 34.871837 samples/sec	 accuracy=78.582933	 loss=0.855291	 lr=0.001000
Epoch[070] Batch [1349]/[3760]	Speed: 35.349206 samples/sec	 accuracy=78.562500	 loss=0.855977	 lr=0.001000
Epoch[070] Batch [1399]/[3760]	Speed: 34.938663 samples/sec	 accuracy=78.574777	 loss=0.855173	 lr=0.001000
Epoch[070] Batch [1449]/[3760]	Speed: 34.977713 samples/sec	 accuracy=78.539871	 loss=0.856000	 lr=0.001000
Epoch[070] Batch [1499]/[3760]	Speed: 35.179585 samples/sec	 accuracy=78.546875	 loss=0.855760	 lr=0.001000
Epoch[070] Batch [1549]/[3760]	Speed: 35.022494 samples/sec	 accuracy=78.520161	 loss=0.856885	 lr=0.001000
Epoch[070] Batch [1599]/[3760]	Speed: 35.227389 samples/sec	 accuracy=78.528320	 loss=0.856361	 lr=0.001000
Epoch[070] Batch [1649]/[3760]	Speed: 35.026663 samples/sec	 accuracy=78.526515	 loss=0.856818	 lr=0.001000
Epoch[070] Batch [1699]/[3760]	Speed: 35.264010 samples/sec	 accuracy=78.545956	 loss=0.855938	 lr=0.001000
Epoch[070] Batch [1749]/[3760]	Speed: 34.864780 samples/sec	 accuracy=78.547321	 loss=0.855800	 lr=0.001000
Epoch[070] Batch [1799]/[3760]	Speed: 35.214282 samples/sec	 accuracy=78.540799	 loss=0.856199	 lr=0.001000
Epoch[070] Batch [1849]/[3760]	Speed: 34.799796 samples/sec	 accuracy=78.545608	 loss=0.856143	 lr=0.001000
Epoch[070] Batch [1899]/[3760]	Speed: 35.472916 samples/sec	 accuracy=78.532072	 loss=0.857001	 lr=0.001000
Epoch[070] Batch [1949]/[3760]	Speed: 34.775308 samples/sec	 accuracy=78.538462	 loss=0.856699	 lr=0.001000
Epoch[070] Batch [1999]/[3760]	Speed: 35.328251 samples/sec	 accuracy=78.510156	 loss=0.857087	 lr=0.001000
Epoch[070] Batch [2049]/[3760]	Speed: 34.847856 samples/sec	 accuracy=78.526677	 loss=0.856281	 lr=0.001000
Epoch[070] Batch [2099]/[3760]	Speed: 35.277676 samples/sec	 accuracy=78.512649	 loss=0.857094	 lr=0.001000
Epoch[070] Batch [2149]/[3760]	Speed: 34.921429 samples/sec	 accuracy=78.512355	 loss=0.857416	 lr=0.001000
Epoch[070] Batch [2199]/[3760]	Speed: 35.172754 samples/sec	 accuracy=78.535511	 loss=0.856649	 lr=0.001000
Epoch[070] Batch [2249]/[3760]	Speed: 35.042649 samples/sec	 accuracy=78.539583	 loss=0.856335	 lr=0.001000
Epoch[070] Batch [2299]/[3760]	Speed: 34.929003 samples/sec	 accuracy=78.540082	 loss=0.856246	 lr=0.001000
Epoch[070] Batch [2349]/[3760]	Speed: 34.913516 samples/sec	 accuracy=78.566489	 loss=0.855486	 lr=0.001000
Epoch[070] Batch [2399]/[3760]	Speed: 34.588566 samples/sec	 accuracy=78.571615	 loss=0.855521	 lr=0.001000
Epoch[070] Batch [2449]/[3760]	Speed: 34.970348 samples/sec	 accuracy=78.555485	 loss=0.856159	 lr=0.001000
Epoch[070] Batch [2499]/[3760]	Speed: 34.516980 samples/sec	 accuracy=78.515000	 loss=0.857648	 lr=0.001000
Epoch[070] Batch [2549]/[3760]	Speed: 35.152309 samples/sec	 accuracy=78.543505	 loss=0.856721	 lr=0.001000
Epoch[070] Batch [2599]/[3760]	Speed: 34.500723 samples/sec	 accuracy=78.539663	 loss=0.857373	 lr=0.001000
Epoch[070] Batch [2649]/[3760]	Speed: 35.058758 samples/sec	 accuracy=78.534788	 loss=0.857427	 lr=0.001000
Epoch[070] Batch [2699]/[3760]	Speed: 34.744498 samples/sec	 accuracy=78.556134	 loss=0.856754	 lr=0.001000
Epoch[070] Batch [2749]/[3760]	Speed: 34.996350 samples/sec	 accuracy=78.538636	 loss=0.857282	 lr=0.001000
Epoch[070] Batch [2799]/[3760]	Speed: 34.901180 samples/sec	 accuracy=78.552455	 loss=0.856491	 lr=0.001000
Epoch[070] Batch [2849]/[3760]	Speed: 34.927913 samples/sec	 accuracy=78.531250	 loss=0.857324	 lr=0.001000
Epoch[070] Batch [2899]/[3760]	Speed: 34.961954 samples/sec	 accuracy=78.543642	 loss=0.856868	 lr=0.001000
Epoch[070] Batch [2949]/[3760]	Speed: 34.696690 samples/sec	 accuracy=78.534958	 loss=0.857315	 lr=0.001000
Epoch[070] Batch [2999]/[3760]	Speed: 34.941325 samples/sec	 accuracy=78.517187	 loss=0.857597	 lr=0.001000
Epoch[070] Batch [3049]/[3760]	Speed: 34.445114 samples/sec	 accuracy=78.538422	 loss=0.856903	 lr=0.001000
Epoch[070] Batch [3099]/[3760]	Speed: 35.198762 samples/sec	 accuracy=78.509577	 loss=0.857204	 lr=0.001000
Epoch[070] Batch [3149]/[3760]	Speed: 34.232630 samples/sec	 accuracy=78.497024	 loss=0.857462	 lr=0.001000
Epoch[070] Batch [3199]/[3760]	Speed: 35.165428 samples/sec	 accuracy=78.470215	 loss=0.858360	 lr=0.001000
Epoch[070] Batch [3249]/[3760]	Speed: 34.612092 samples/sec	 accuracy=78.459135	 loss=0.858680	 lr=0.001000
Epoch[070] Batch [3299]/[3760]	Speed: 34.693676 samples/sec	 accuracy=78.461648	 loss=0.858947	 lr=0.001000
Epoch[070] Batch [3349]/[3760]	Speed: 34.808105 samples/sec	 accuracy=78.438899	 loss=0.859921	 lr=0.001000
Epoch[070] Batch [3399]/[3760]	Speed: 34.493293 samples/sec	 accuracy=78.454504	 loss=0.859647	 lr=0.001000
Epoch[070] Batch [3449]/[3760]	Speed: 34.976078 samples/sec	 accuracy=78.443841	 loss=0.860176	 lr=0.001000
Epoch[070] Batch [3499]/[3760]	Speed: 34.133462 samples/sec	 accuracy=78.440625	 loss=0.859914	 lr=0.001000
Epoch[070] Batch [3549]/[3760]	Speed: 35.089325 samples/sec	 accuracy=78.436180	 loss=0.860344	 lr=0.001000
Epoch[070] Batch [3599]/[3760]	Speed: 34.339015 samples/sec	 accuracy=78.427951	 loss=0.860440	 lr=0.001000
Epoch[070] Batch [3649]/[3760]	Speed: 34.863618 samples/sec	 accuracy=78.411815	 loss=0.860883	 lr=0.001000
Epoch[070] Batch [3699]/[3760]	Speed: 34.679687 samples/sec	 accuracy=78.430743	 loss=0.860247	 lr=0.001000
Epoch[070] Batch [3749]/[3760]	Speed: 35.985356 samples/sec	 accuracy=78.436250	 loss=0.860105	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.625000 acc-top5=85.750000
Batch [0099]/[0303]: acc-top1=66.140625 acc-top5=85.265625
Batch [0149]/[0303]: acc-top1=66.500000 acc-top5=85.437500
Batch [0199]/[0303]: acc-top1=66.226562 acc-top5=85.578125
Batch [0249]/[0303]: acc-top1=65.887500 acc-top5=85.450000
Batch [0299]/[0303]: acc-top1=65.744792 acc-top5=85.468750
[Epoch 070] training: accuracy=78.430851	 loss=0.860057
[Epoch 070] speed: 34 samples/sec	time cost: 7369.115739
[Epoch 070] validation: acc-top1=65.753919 acc-top5=85.483705 loss=1.639796
Epoch[071] Batch [0049]/[3760]	Speed: 25.278848 samples/sec	 accuracy=78.750000	 loss=0.816247	 lr=0.001000
Epoch[071] Batch [0099]/[3760]	Speed: 35.530655 samples/sec	 accuracy=79.265625	 loss=0.809985	 lr=0.001000
Epoch[071] Batch [0149]/[3760]	Speed: 35.098040 samples/sec	 accuracy=79.218750	 loss=0.820096	 lr=0.001000
Epoch[071] Batch [0199]/[3760]	Speed: 35.338186 samples/sec	 accuracy=79.171875	 loss=0.820001	 lr=0.001000
Epoch[071] Batch [0249]/[3760]	Speed: 35.182178 samples/sec	 accuracy=79.068750	 loss=0.825127	 lr=0.001000
Epoch[071] Batch [0299]/[3760]	Speed: 35.355194 samples/sec	 accuracy=79.187500	 loss=0.823654	 lr=0.001000
Epoch[071] Batch [0349]/[3760]	Speed: 35.235477 samples/sec	 accuracy=79.285714	 loss=0.822207	 lr=0.001000
Epoch[071] Batch [0399]/[3760]	Speed: 35.451275 samples/sec	 accuracy=79.300781	 loss=0.822443	 lr=0.001000
Epoch[071] Batch [0449]/[3760]	Speed: 35.155074 samples/sec	 accuracy=79.152778	 loss=0.826282	 lr=0.001000
Epoch[071] Batch [0499]/[3760]	Speed: 35.267396 samples/sec	 accuracy=79.140625	 loss=0.825992	 lr=0.001000
Epoch[071] Batch [0549]/[3760]	Speed: 35.154698 samples/sec	 accuracy=79.127841	 loss=0.826418	 lr=0.001000
Epoch[071] Batch [0599]/[3760]	Speed: 35.302292 samples/sec	 accuracy=79.119792	 loss=0.827695	 lr=0.001000
Epoch[071] Batch [0649]/[3760]	Speed: 35.159659 samples/sec	 accuracy=78.983173	 loss=0.833029	 lr=0.001000
Epoch[071] Batch [0699]/[3760]	Speed: 35.291466 samples/sec	 accuracy=79.017857	 loss=0.832366	 lr=0.001000
Epoch[071] Batch [0749]/[3760]	Speed: 35.277006 samples/sec	 accuracy=78.987500	 loss=0.833461	 lr=0.001000
Epoch[071] Batch [0799]/[3760]	Speed: 35.307316 samples/sec	 accuracy=78.912109	 loss=0.837435	 lr=0.001000
Epoch[071] Batch [0849]/[3760]	Speed: 35.271802 samples/sec	 accuracy=78.895221	 loss=0.836671	 lr=0.001000
Epoch[071] Batch [0899]/[3760]	Speed: 35.199658 samples/sec	 accuracy=78.854167	 loss=0.837581	 lr=0.001000
Epoch[071] Batch [0949]/[3760]	Speed: 35.363858 samples/sec	 accuracy=78.876645	 loss=0.838394	 lr=0.001000
Epoch[071] Batch [0999]/[3760]	Speed: 35.237680 samples/sec	 accuracy=78.857813	 loss=0.840091	 lr=0.001000
Epoch[071] Batch [1049]/[3760]	Speed: 35.243435 samples/sec	 accuracy=78.851190	 loss=0.841739	 lr=0.001000
Epoch[071] Batch [1099]/[3760]	Speed: 35.384486 samples/sec	 accuracy=78.897727	 loss=0.840671	 lr=0.001000
Epoch[071] Batch [1149]/[3760]	Speed: 35.080568 samples/sec	 accuracy=78.911685	 loss=0.840120	 lr=0.001000
Epoch[071] Batch [1199]/[3760]	Speed: 35.270863 samples/sec	 accuracy=78.934896	 loss=0.840710	 lr=0.001000
Epoch[071] Batch [1249]/[3760]	Speed: 35.142879 samples/sec	 accuracy=78.881250	 loss=0.842869	 lr=0.001000
Epoch[071] Batch [1299]/[3760]	Speed: 35.264311 samples/sec	 accuracy=78.846154	 loss=0.844670	 lr=0.001000
Epoch[071] Batch [1349]/[3760]	Speed: 35.105932 samples/sec	 accuracy=78.836806	 loss=0.844786	 lr=0.001000
Epoch[071] Batch [1399]/[3760]	Speed: 35.232257 samples/sec	 accuracy=78.842634	 loss=0.844571	 lr=0.001000
Epoch[071] Batch [1449]/[3760]	Speed: 35.126958 samples/sec	 accuracy=78.844828	 loss=0.844789	 lr=0.001000
Epoch[071] Batch [1499]/[3760]	Speed: 35.252216 samples/sec	 accuracy=78.806250	 loss=0.845417	 lr=0.001000
Epoch[071] Batch [1549]/[3760]	Speed: 34.958882 samples/sec	 accuracy=78.799395	 loss=0.845797	 lr=0.001000
Epoch[071] Batch [1599]/[3760]	Speed: 35.385212 samples/sec	 accuracy=78.783203	 loss=0.846187	 lr=0.001000
Epoch[071] Batch [1649]/[3760]	Speed: 35.165980 samples/sec	 accuracy=78.765152	 loss=0.845742	 lr=0.001000
Epoch[071] Batch [1699]/[3760]	Speed: 35.416800 samples/sec	 accuracy=78.760110	 loss=0.846173	 lr=0.001000
Epoch[071] Batch [1749]/[3760]	Speed: 35.121973 samples/sec	 accuracy=78.751786	 loss=0.846632	 lr=0.001000
Epoch[071] Batch [1799]/[3760]	Speed: 35.305486 samples/sec	 accuracy=78.757812	 loss=0.847055	 lr=0.001000
Epoch[071] Batch [1849]/[3760]	Speed: 35.274146 samples/sec	 accuracy=78.750000	 loss=0.847453	 lr=0.001000
Epoch[071] Batch [1899]/[3760]	Speed: 35.392095 samples/sec	 accuracy=78.773026	 loss=0.846727	 lr=0.001000
Epoch[071] Batch [1949]/[3760]	Speed: 35.195508 samples/sec	 accuracy=78.739583	 loss=0.847801	 lr=0.001000
Epoch[071] Batch [1999]/[3760]	Speed: 35.191539 samples/sec	 accuracy=78.738281	 loss=0.848301	 lr=0.001000
Epoch[071] Batch [2049]/[3760]	Speed: 35.150408 samples/sec	 accuracy=78.732470	 loss=0.848462	 lr=0.001000
Epoch[071] Batch [2099]/[3760]	Speed: 35.344820 samples/sec	 accuracy=78.722470	 loss=0.848645	 lr=0.001000
Epoch[071] Batch [2149]/[3760]	Speed: 35.132826 samples/sec	 accuracy=78.708576	 loss=0.848958	 lr=0.001000
Epoch[071] Batch [2199]/[3760]	Speed: 35.360947 samples/sec	 accuracy=78.713778	 loss=0.849126	 lr=0.001000
Epoch[071] Batch [2249]/[3760]	Speed: 34.798042 samples/sec	 accuracy=78.726389	 loss=0.848730	 lr=0.001000
Epoch[071] Batch [2299]/[3760]	Speed: 35.269643 samples/sec	 accuracy=78.711277	 loss=0.849766	 lr=0.001000
Epoch[071] Batch [2349]/[3760]	Speed: 34.929976 samples/sec	 accuracy=78.728723	 loss=0.849472	 lr=0.001000
Epoch[071] Batch [2399]/[3760]	Speed: 35.138111 samples/sec	 accuracy=78.714844	 loss=0.849962	 lr=0.001000
Epoch[071] Batch [2449]/[3760]	Speed: 34.918027 samples/sec	 accuracy=78.710459	 loss=0.850057	 lr=0.001000
Epoch[071] Batch [2499]/[3760]	Speed: 35.252427 samples/sec	 accuracy=78.726875	 loss=0.849454	 lr=0.001000
Epoch[071] Batch [2549]/[3760]	Speed: 35.095520 samples/sec	 accuracy=78.734681	 loss=0.849018	 lr=0.001000
Epoch[071] Batch [2599]/[3760]	Speed: 35.264510 samples/sec	 accuracy=78.710337	 loss=0.849874	 lr=0.001000
Epoch[071] Batch [2649]/[3760]	Speed: 35.384099 samples/sec	 accuracy=78.708137	 loss=0.849903	 lr=0.001000
Epoch[071] Batch [2699]/[3760]	Speed: 34.958770 samples/sec	 accuracy=78.687500	 loss=0.850789	 lr=0.001000
Epoch[071] Batch [2749]/[3760]	Speed: 35.273687 samples/sec	 accuracy=78.694886	 loss=0.850669	 lr=0.001000
Epoch[071] Batch [2799]/[3760]	Speed: 34.990611 samples/sec	 accuracy=78.683036	 loss=0.850978	 lr=0.001000
Epoch[071] Batch [2849]/[3760]	Speed: 35.453149 samples/sec	 accuracy=78.673246	 loss=0.851443	 lr=0.001000
Epoch[071] Batch [2899]/[3760]	Speed: 34.877163 samples/sec	 accuracy=78.679418	 loss=0.851955	 lr=0.001000
Epoch[071] Batch [2949]/[3760]	Speed: 35.356153 samples/sec	 accuracy=78.670551	 loss=0.852630	 lr=0.001000
Epoch[071] Batch [2999]/[3760]	Speed: 34.890038 samples/sec	 accuracy=78.657813	 loss=0.852832	 lr=0.001000
Epoch[071] Batch [3049]/[3760]	Speed: 35.247490 samples/sec	 accuracy=78.663422	 loss=0.852975	 lr=0.001000
Epoch[071] Batch [3099]/[3760]	Speed: 35.069285 samples/sec	 accuracy=78.662802	 loss=0.852796	 lr=0.001000
Epoch[071] Batch [3149]/[3760]	Speed: 35.089740 samples/sec	 accuracy=78.663194	 loss=0.853191	 lr=0.001000
Epoch[071] Batch [3199]/[3760]	Speed: 35.157308 samples/sec	 accuracy=78.685059	 loss=0.852481	 lr=0.001000
Epoch[071] Batch [3249]/[3760]	Speed: 35.027804 samples/sec	 accuracy=78.701442	 loss=0.852227	 lr=0.001000
Epoch[071] Batch [3299]/[3760]	Speed: 35.401357 samples/sec	 accuracy=78.713542	 loss=0.852111	 lr=0.001000
Epoch[071] Batch [3349]/[3760]	Speed: 35.241685 samples/sec	 accuracy=78.698228	 loss=0.852654	 lr=0.001000
Epoch[071] Batch [3399]/[3760]	Speed: 35.437466 samples/sec	 accuracy=78.680147	 loss=0.853467	 lr=0.001000
Epoch[071] Batch [3449]/[3760]	Speed: 35.325476 samples/sec	 accuracy=78.676630	 loss=0.853656	 lr=0.001000
Epoch[071] Batch [3499]/[3760]	Speed: 35.281955 samples/sec	 accuracy=78.677232	 loss=0.853858	 lr=0.001000
Epoch[071] Batch [3549]/[3760]	Speed: 35.174055 samples/sec	 accuracy=78.676056	 loss=0.853489	 lr=0.001000
Epoch[071] Batch [3599]/[3760]	Speed: 35.058762 samples/sec	 accuracy=78.664931	 loss=0.853343	 lr=0.001000
Epoch[071] Batch [3649]/[3760]	Speed: 34.939283 samples/sec	 accuracy=78.634418	 loss=0.853986	 lr=0.001000
Epoch[071] Batch [3699]/[3760]	Speed: 35.446755 samples/sec	 accuracy=78.626267	 loss=0.854509	 lr=0.001000
Epoch[071] Batch [3749]/[3760]	Speed: 36.479342 samples/sec	 accuracy=78.621250	 loss=0.854505	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.812500 acc-top5=85.656250
Batch [0099]/[0303]: acc-top1=66.453125 acc-top5=85.468750
Batch [0149]/[0303]: acc-top1=66.822917 acc-top5=85.479167
Batch [0199]/[0303]: acc-top1=66.398438 acc-top5=85.445312
Batch [0249]/[0303]: acc-top1=66.050000 acc-top5=85.412500
Batch [0299]/[0303]: acc-top1=65.817708 acc-top5=85.458333
[Epoch 071] training: accuracy=78.618684	 loss=0.854635
[Epoch 071] speed: 35 samples/sec	time cost: 7320.232545
[Epoch 071] validation: acc-top1=65.795173 acc-top5=85.478548 loss=1.623234
Epoch[072] Batch [0049]/[3759]	Speed: 25.012459 samples/sec	 accuracy=78.875000	 loss=0.817477	 lr=0.001000
Epoch[072] Batch [0099]/[3759]	Speed: 35.140637 samples/sec	 accuracy=79.593750	 loss=0.805580	 lr=0.001000
Epoch[072] Batch [0149]/[3759]	Speed: 35.334193 samples/sec	 accuracy=79.708333	 loss=0.809849	 lr=0.001000
Epoch[072] Batch [0199]/[3759]	Speed: 35.096927 samples/sec	 accuracy=79.906250	 loss=0.808459	 lr=0.001000
Epoch[072] Batch [0249]/[3759]	Speed: 35.237462 samples/sec	 accuracy=79.893750	 loss=0.813195	 lr=0.001000
Epoch[072] Batch [0299]/[3759]	Speed: 34.968665 samples/sec	 accuracy=79.619792	 loss=0.820926	 lr=0.001000
Epoch[072] Batch [0349]/[3759]	Speed: 35.388855 samples/sec	 accuracy=79.406250	 loss=0.826464	 lr=0.001000
Epoch[072] Batch [0399]/[3759]	Speed: 35.053381 samples/sec	 accuracy=79.335938	 loss=0.827328	 lr=0.001000
Epoch[072] Batch [0449]/[3759]	Speed: 35.366309 samples/sec	 accuracy=79.267361	 loss=0.828257	 lr=0.001000
Epoch[072] Batch [0499]/[3759]	Speed: 35.085057 samples/sec	 accuracy=79.246875	 loss=0.827828	 lr=0.001000
Epoch[072] Batch [0549]/[3759]	Speed: 35.242342 samples/sec	 accuracy=79.198864	 loss=0.829935	 lr=0.001000
Epoch[072] Batch [0599]/[3759]	Speed: 35.136990 samples/sec	 accuracy=79.265625	 loss=0.826011	 lr=0.001000
Epoch[072] Batch [0649]/[3759]	Speed: 35.181415 samples/sec	 accuracy=79.168269	 loss=0.829516	 lr=0.001000
Epoch[072] Batch [0699]/[3759]	Speed: 34.914996 samples/sec	 accuracy=79.107143	 loss=0.831169	 lr=0.001000
Epoch[072] Batch [0749]/[3759]	Speed: 35.258510 samples/sec	 accuracy=79.143750	 loss=0.829272	 lr=0.001000
Epoch[072] Batch [0799]/[3759]	Speed: 35.006232 samples/sec	 accuracy=79.191406	 loss=0.828582	 lr=0.001000
Epoch[072] Batch [0849]/[3759]	Speed: 35.191352 samples/sec	 accuracy=79.185662	 loss=0.828586	 lr=0.001000
Epoch[072] Batch [0899]/[3759]	Speed: 35.107688 samples/sec	 accuracy=79.222222	 loss=0.828642	 lr=0.001000
Epoch[072] Batch [0949]/[3759]	Speed: 35.281042 samples/sec	 accuracy=79.225329	 loss=0.828294	 lr=0.001000
Epoch[072] Batch [0999]/[3759]	Speed: 35.260979 samples/sec	 accuracy=79.231250	 loss=0.828104	 lr=0.001000
Epoch[072] Batch [1049]/[3759]	Speed: 35.414258 samples/sec	 accuracy=79.229167	 loss=0.828789	 lr=0.001000
Epoch[072] Batch [1099]/[3759]	Speed: 34.990250 samples/sec	 accuracy=79.208807	 loss=0.830751	 lr=0.001000
Epoch[072] Batch [1149]/[3759]	Speed: 35.343288 samples/sec	 accuracy=79.211957	 loss=0.830164	 lr=0.001000
Epoch[072] Batch [1199]/[3759]	Speed: 35.094263 samples/sec	 accuracy=79.212240	 loss=0.830225	 lr=0.001000
Epoch[072] Batch [1249]/[3759]	Speed: 35.212242 samples/sec	 accuracy=79.236250	 loss=0.830184	 lr=0.001000
Epoch[072] Batch [1299]/[3759]	Speed: 35.021477 samples/sec	 accuracy=79.265625	 loss=0.828922	 lr=0.001000
Epoch[072] Batch [1349]/[3759]	Speed: 35.204362 samples/sec	 accuracy=79.261574	 loss=0.828449	 lr=0.001000
Epoch[072] Batch [1399]/[3759]	Speed: 34.922501 samples/sec	 accuracy=79.227679	 loss=0.829876	 lr=0.001000
Epoch[072] Batch [1449]/[3759]	Speed: 35.093861 samples/sec	 accuracy=79.245690	 loss=0.830145	 lr=0.001000
Epoch[072] Batch [1499]/[3759]	Speed: 34.965546 samples/sec	 accuracy=79.267708	 loss=0.830196	 lr=0.001000
Epoch[072] Batch [1549]/[3759]	Speed: 35.083914 samples/sec	 accuracy=79.233871	 loss=0.830595	 lr=0.001000
Epoch[072] Batch [1599]/[3759]	Speed: 34.984923 samples/sec	 accuracy=79.230469	 loss=0.831570	 lr=0.001000
Epoch[072] Batch [1649]/[3759]	Speed: 35.179981 samples/sec	 accuracy=79.225379	 loss=0.831321	 lr=0.001000
Epoch[072] Batch [1699]/[3759]	Speed: 34.796404 samples/sec	 accuracy=79.196691	 loss=0.832954	 lr=0.001000
Epoch[072] Batch [1749]/[3759]	Speed: 34.981351 samples/sec	 accuracy=79.191964	 loss=0.833304	 lr=0.001000
Epoch[072] Batch [1799]/[3759]	Speed: 35.084142 samples/sec	 accuracy=79.198785	 loss=0.833010	 lr=0.001000
Epoch[072] Batch [1849]/[3759]	Speed: 34.962286 samples/sec	 accuracy=79.174831	 loss=0.832798	 lr=0.001000
Epoch[072] Batch [1899]/[3759]	Speed: 35.058465 samples/sec	 accuracy=79.152138	 loss=0.834022	 lr=0.001000
Epoch[072] Batch [1949]/[3759]	Speed: 34.777855 samples/sec	 accuracy=79.138622	 loss=0.834391	 lr=0.001000
Epoch[072] Batch [1999]/[3759]	Speed: 35.177787 samples/sec	 accuracy=79.130469	 loss=0.834831	 lr=0.001000
Epoch[072] Batch [2049]/[3759]	Speed: 34.735000 samples/sec	 accuracy=79.095274	 loss=0.835883	 lr=0.001000
Epoch[072] Batch [2099]/[3759]	Speed: 35.313075 samples/sec	 accuracy=79.085565	 loss=0.836270	 lr=0.001000
Epoch[072] Batch [2149]/[3759]	Speed: 34.756777 samples/sec	 accuracy=79.065407	 loss=0.837510	 lr=0.001000
Epoch[072] Batch [2199]/[3759]	Speed: 35.223185 samples/sec	 accuracy=79.054688	 loss=0.838038	 lr=0.001000
Epoch[072] Batch [2249]/[3759]	Speed: 34.819428 samples/sec	 accuracy=79.085417	 loss=0.837113	 lr=0.001000
Epoch[072] Batch [2299]/[3759]	Speed: 35.243832 samples/sec	 accuracy=79.042120	 loss=0.838534	 lr=0.001000
Epoch[072] Batch [2349]/[3759]	Speed: 34.737835 samples/sec	 accuracy=79.035239	 loss=0.838254	 lr=0.001000
Epoch[072] Batch [2399]/[3759]	Speed: 35.156264 samples/sec	 accuracy=79.022135	 loss=0.838507	 lr=0.001000
Epoch[072] Batch [2449]/[3759]	Speed: 34.925526 samples/sec	 accuracy=79.000638	 loss=0.839357	 lr=0.001000
Epoch[072] Batch [2499]/[3759]	Speed: 35.075480 samples/sec	 accuracy=78.981250	 loss=0.840194	 lr=0.001000
Epoch[072] Batch [2549]/[3759]	Speed: 34.869004 samples/sec	 accuracy=78.941176	 loss=0.841441	 lr=0.001000
Epoch[072] Batch [2599]/[3759]	Speed: 34.937612 samples/sec	 accuracy=78.903245	 loss=0.842835	 lr=0.001000
Epoch[072] Batch [2649]/[3759]	Speed: 35.030686 samples/sec	 accuracy=78.913325	 loss=0.842633	 lr=0.001000
Epoch[072] Batch [2699]/[3759]	Speed: 34.845617 samples/sec	 accuracy=78.909144	 loss=0.842841	 lr=0.001000
Epoch[072] Batch [2749]/[3759]	Speed: 35.292272 samples/sec	 accuracy=78.898864	 loss=0.842744	 lr=0.001000
Epoch[072] Batch [2799]/[3759]	Speed: 34.835520 samples/sec	 accuracy=78.882812	 loss=0.843554	 lr=0.001000
Epoch[072] Batch [2849]/[3759]	Speed: 35.310940 samples/sec	 accuracy=78.889803	 loss=0.843675	 lr=0.001000
Epoch[072] Batch [2899]/[3759]	Speed: 35.045646 samples/sec	 accuracy=78.892241	 loss=0.843598	 lr=0.001000
Epoch[072] Batch [2949]/[3759]	Speed: 35.070819 samples/sec	 accuracy=78.903602	 loss=0.843495	 lr=0.001000
Epoch[072] Batch [2999]/[3759]	Speed: 34.961775 samples/sec	 accuracy=78.901562	 loss=0.843755	 lr=0.001000
Epoch[072] Batch [3049]/[3759]	Speed: 35.285545 samples/sec	 accuracy=78.895492	 loss=0.844036	 lr=0.001000
Epoch[072] Batch [3099]/[3759]	Speed: 35.254518 samples/sec	 accuracy=78.891129	 loss=0.843891	 lr=0.001000
Epoch[072] Batch [3149]/[3759]	Speed: 35.297877 samples/sec	 accuracy=78.883929	 loss=0.843707	 lr=0.001000
Epoch[072] Batch [3199]/[3759]	Speed: 35.249247 samples/sec	 accuracy=78.862305	 loss=0.844268	 lr=0.001000
Epoch[072] Batch [3249]/[3759]	Speed: 35.296961 samples/sec	 accuracy=78.858654	 loss=0.844298	 lr=0.001000
Epoch[072] Batch [3299]/[3759]	Speed: 35.353009 samples/sec	 accuracy=78.853693	 loss=0.844100	 lr=0.001000
Epoch[072] Batch [3349]/[3759]	Speed: 35.355070 samples/sec	 accuracy=78.837220	 loss=0.844591	 lr=0.001000
Epoch[072] Batch [3399]/[3759]	Speed: 35.439621 samples/sec	 accuracy=78.824449	 loss=0.844977	 lr=0.001000
Epoch[072] Batch [3449]/[3759]	Speed: 35.180780 samples/sec	 accuracy=78.822464	 loss=0.845145	 lr=0.001000
Epoch[072] Batch [3499]/[3759]	Speed: 35.257677 samples/sec	 accuracy=78.839286	 loss=0.844711	 lr=0.001000
Epoch[072] Batch [3549]/[3759]	Speed: 35.120748 samples/sec	 accuracy=78.846831	 loss=0.844373	 lr=0.001000
Epoch[072] Batch [3599]/[3759]	Speed: 35.383519 samples/sec	 accuracy=78.847656	 loss=0.844220	 lr=0.001000
Epoch[072] Batch [3649]/[3759]	Speed: 35.052888 samples/sec	 accuracy=78.844606	 loss=0.844106	 lr=0.001000
Epoch[072] Batch [3699]/[3759]	Speed: 35.357107 samples/sec	 accuracy=78.831081	 loss=0.844377	 lr=0.001000
Epoch[072] Batch [3749]/[3759]	Speed: 36.310115 samples/sec	 accuracy=78.826250	 loss=0.844549	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.656250 acc-top5=86.031250
Batch [0099]/[0303]: acc-top1=65.968750 acc-top5=85.500000
Batch [0149]/[0303]: acc-top1=66.489583 acc-top5=85.583333
Batch [0199]/[0303]: acc-top1=66.203125 acc-top5=85.539062
Batch [0249]/[0303]: acc-top1=65.843750 acc-top5=85.437500
Batch [0299]/[0303]: acc-top1=65.739583 acc-top5=85.546875
[Epoch 072] training: accuracy=78.825402	 loss=0.844498
[Epoch 072] speed: 34 samples/sec	time cost: 7337.707845
[Epoch 072] validation: acc-top1=65.743606 acc-top5=85.555899 loss=1.624773
Epoch[073] Batch [0049]/[3760]	Speed: 25.198554 samples/sec	 accuracy=78.718750	 loss=0.820960	 lr=0.001000
Epoch[073] Batch [0099]/[3760]	Speed: 35.446370 samples/sec	 accuracy=78.734375	 loss=0.826770	 lr=0.001000
Epoch[073] Batch [0149]/[3760]	Speed: 34.923901 samples/sec	 accuracy=79.041667	 loss=0.815115	 lr=0.001000
Epoch[073] Batch [0199]/[3760]	Speed: 35.364477 samples/sec	 accuracy=78.992188	 loss=0.827262	 lr=0.001000
Epoch[073] Batch [0249]/[3760]	Speed: 35.149339 samples/sec	 accuracy=79.150000	 loss=0.826164	 lr=0.001000
Epoch[073] Batch [0299]/[3760]	Speed: 35.282445 samples/sec	 accuracy=79.109375	 loss=0.826672	 lr=0.001000
Epoch[073] Batch [0349]/[3760]	Speed: 35.026486 samples/sec	 accuracy=79.098214	 loss=0.829028	 lr=0.001000
Epoch[073] Batch [0399]/[3760]	Speed: 35.480429 samples/sec	 accuracy=79.179688	 loss=0.827786	 lr=0.001000
Epoch[073] Batch [0449]/[3760]	Speed: 35.152349 samples/sec	 accuracy=79.263889	 loss=0.826001	 lr=0.001000
Epoch[073] Batch [0499]/[3760]	Speed: 35.221589 samples/sec	 accuracy=79.209375	 loss=0.828737	 lr=0.001000
Epoch[073] Batch [0549]/[3760]	Speed: 35.067389 samples/sec	 accuracy=79.309659	 loss=0.827002	 lr=0.001000
Epoch[073] Batch [0599]/[3760]	Speed: 35.342802 samples/sec	 accuracy=79.494792	 loss=0.821155	 lr=0.001000
Epoch[073] Batch [0649]/[3760]	Speed: 35.171080 samples/sec	 accuracy=79.415865	 loss=0.823406	 lr=0.001000
Epoch[073] Batch [0699]/[3760]	Speed: 35.387358 samples/sec	 accuracy=79.330357	 loss=0.824486	 lr=0.001000
Epoch[073] Batch [0749]/[3760]	Speed: 34.990928 samples/sec	 accuracy=79.289583	 loss=0.825289	 lr=0.001000
Epoch[073] Batch [0799]/[3760]	Speed: 34.942968 samples/sec	 accuracy=79.294922	 loss=0.825379	 lr=0.001000
Epoch[073] Batch [0849]/[3760]	Speed: 34.858511 samples/sec	 accuracy=79.261029	 loss=0.828066	 lr=0.001000
Epoch[073] Batch [0899]/[3760]	Speed: 34.822428 samples/sec	 accuracy=79.302083	 loss=0.826215	 lr=0.001000
Epoch[073] Batch [0949]/[3760]	Speed: 35.168225 samples/sec	 accuracy=79.328947	 loss=0.825755	 lr=0.001000
Epoch[073] Batch [0999]/[3760]	Speed: 34.639890 samples/sec	 accuracy=79.295313	 loss=0.826387	 lr=0.001000
Epoch[073] Batch [1049]/[3760]	Speed: 35.130078 samples/sec	 accuracy=79.238095	 loss=0.827911	 lr=0.001000
Epoch[073] Batch [1099]/[3760]	Speed: 34.543739 samples/sec	 accuracy=79.204545	 loss=0.828357	 lr=0.001000
Epoch[073] Batch [1149]/[3760]	Speed: 35.143168 samples/sec	 accuracy=79.142663	 loss=0.830398	 lr=0.001000
Epoch[073] Batch [1199]/[3760]	Speed: 34.723416 samples/sec	 accuracy=79.177083	 loss=0.829789	 lr=0.001000
Epoch[073] Batch [1249]/[3760]	Speed: 35.036764 samples/sec	 accuracy=79.128750	 loss=0.831584	 lr=0.001000
Epoch[073] Batch [1299]/[3760]	Speed: 35.029178 samples/sec	 accuracy=79.134615	 loss=0.831903	 lr=0.001000
Epoch[073] Batch [1349]/[3760]	Speed: 34.827790 samples/sec	 accuracy=79.126157	 loss=0.832099	 lr=0.001000
Epoch[073] Batch [1399]/[3760]	Speed: 34.935892 samples/sec	 accuracy=79.132812	 loss=0.831873	 lr=0.001000
Epoch[073] Batch [1449]/[3760]	Speed: 34.617680 samples/sec	 accuracy=79.120690	 loss=0.831724	 lr=0.001000
Epoch[073] Batch [1499]/[3760]	Speed: 35.045715 samples/sec	 accuracy=79.116667	 loss=0.832404	 lr=0.001000
Epoch[073] Batch [1549]/[3760]	Speed: 34.435378 samples/sec	 accuracy=79.087702	 loss=0.832865	 lr=0.001000
Epoch[073] Batch [1599]/[3760]	Speed: 35.055896 samples/sec	 accuracy=79.063477	 loss=0.834526	 lr=0.001000
Epoch[073] Batch [1649]/[3760]	Speed: 34.355275 samples/sec	 accuracy=79.080492	 loss=0.833717	 lr=0.001000
Epoch[073] Batch [1699]/[3760]	Speed: 35.041067 samples/sec	 accuracy=79.049632	 loss=0.834687	 lr=0.001000
Epoch[073] Batch [1749]/[3760]	Speed: 34.615494 samples/sec	 accuracy=79.042857	 loss=0.834855	 lr=0.001000
Epoch[073] Batch [1799]/[3760]	Speed: 34.834137 samples/sec	 accuracy=79.036458	 loss=0.835134	 lr=0.001000
Epoch[073] Batch [1849]/[3760]	Speed: 34.811309 samples/sec	 accuracy=79.043919	 loss=0.835348	 lr=0.001000
Epoch[073] Batch [1899]/[3760]	Speed: 34.765576 samples/sec	 accuracy=79.050987	 loss=0.835304	 lr=0.001000
Epoch[073] Batch [1949]/[3760]	Speed: 34.957567 samples/sec	 accuracy=79.036058	 loss=0.835780	 lr=0.001000
Epoch[073] Batch [1999]/[3760]	Speed: 34.429654 samples/sec	 accuracy=79.044531	 loss=0.835851	 lr=0.001000
Epoch[073] Batch [2049]/[3760]	Speed: 35.340725 samples/sec	 accuracy=79.017530	 loss=0.836627	 lr=0.001000
Epoch[073] Batch [2099]/[3760]	Speed: 34.915856 samples/sec	 accuracy=79.015625	 loss=0.836603	 lr=0.001000
Epoch[073] Batch [2149]/[3760]	Speed: 35.294964 samples/sec	 accuracy=79.013808	 loss=0.836517	 lr=0.001000
Epoch[073] Batch [2199]/[3760]	Speed: 34.893053 samples/sec	 accuracy=78.992898	 loss=0.837087	 lr=0.001000
Epoch[073] Batch [2249]/[3760]	Speed: 35.278953 samples/sec	 accuracy=78.997917	 loss=0.837415	 lr=0.001000
Epoch[073] Batch [2299]/[3760]	Speed: 35.072295 samples/sec	 accuracy=78.980978	 loss=0.837680	 lr=0.001000
Epoch[073] Batch [2349]/[3760]	Speed: 34.979764 samples/sec	 accuracy=78.968085	 loss=0.837882	 lr=0.001000
Epoch[073] Batch [2399]/[3760]	Speed: 35.250497 samples/sec	 accuracy=78.971354	 loss=0.837071	 lr=0.001000
Epoch[073] Batch [2449]/[3760]	Speed: 35.050556 samples/sec	 accuracy=78.954082	 loss=0.837257	 lr=0.001000
Epoch[073] Batch [2499]/[3760]	Speed: 35.212344 samples/sec	 accuracy=78.953750	 loss=0.837852	 lr=0.001000
Epoch[073] Batch [2549]/[3760]	Speed: 35.056932 samples/sec	 accuracy=78.935662	 loss=0.838582	 lr=0.001000
Epoch[073] Batch [2599]/[3760]	Speed: 35.252867 samples/sec	 accuracy=78.939904	 loss=0.838277	 lr=0.001000
Epoch[073] Batch [2649]/[3760]	Speed: 35.106107 samples/sec	 accuracy=78.948703	 loss=0.838337	 lr=0.001000
Epoch[073] Batch [2699]/[3760]	Speed: 35.409856 samples/sec	 accuracy=78.938657	 loss=0.838518	 lr=0.001000
Epoch[073] Batch [2749]/[3760]	Speed: 34.984989 samples/sec	 accuracy=78.946591	 loss=0.838236	 lr=0.001000
Epoch[073] Batch [2799]/[3760]	Speed: 35.236592 samples/sec	 accuracy=78.948103	 loss=0.838210	 lr=0.001000
Epoch[073] Batch [2849]/[3760]	Speed: 35.095832 samples/sec	 accuracy=78.940789	 loss=0.838540	 lr=0.001000
Epoch[073] Batch [2899]/[3760]	Speed: 35.150177 samples/sec	 accuracy=78.925647	 loss=0.838665	 lr=0.001000
Epoch[073] Batch [2949]/[3760]	Speed: 35.036539 samples/sec	 accuracy=78.920021	 loss=0.838710	 lr=0.001000
Epoch[073] Batch [2999]/[3760]	Speed: 35.496747 samples/sec	 accuracy=78.922917	 loss=0.838793	 lr=0.001000
Epoch[073] Batch [3049]/[3760]	Speed: 35.130736 samples/sec	 accuracy=78.900102	 loss=0.839403	 lr=0.001000
Epoch[073] Batch [3099]/[3760]	Speed: 35.268579 samples/sec	 accuracy=78.897177	 loss=0.839199	 lr=0.001000
Epoch[073] Batch [3149]/[3760]	Speed: 34.923237 samples/sec	 accuracy=78.900794	 loss=0.839130	 lr=0.001000
Epoch[073] Batch [3199]/[3760]	Speed: 35.291716 samples/sec	 accuracy=78.917480	 loss=0.838650	 lr=0.001000
Epoch[073] Batch [3249]/[3760]	Speed: 35.324869 samples/sec	 accuracy=78.917788	 loss=0.838570	 lr=0.001000
Epoch[073] Batch [3299]/[3760]	Speed: 35.277330 samples/sec	 accuracy=78.913352	 loss=0.838718	 lr=0.001000
Epoch[073] Batch [3349]/[3760]	Speed: 35.310360 samples/sec	 accuracy=78.903451	 loss=0.839173	 lr=0.001000
Epoch[073] Batch [3399]/[3760]	Speed: 35.251451 samples/sec	 accuracy=78.893382	 loss=0.839467	 lr=0.001000
Epoch[073] Batch [3449]/[3760]	Speed: 35.294661 samples/sec	 accuracy=78.894475	 loss=0.839409	 lr=0.001000
Epoch[073] Batch [3499]/[3760]	Speed: 35.306221 samples/sec	 accuracy=78.899107	 loss=0.839613	 lr=0.001000
Epoch[073] Batch [3549]/[3760]	Speed: 35.116679 samples/sec	 accuracy=78.897887	 loss=0.839713	 lr=0.001000
Epoch[073] Batch [3599]/[3760]	Speed: 35.210241 samples/sec	 accuracy=78.898438	 loss=0.840011	 lr=0.001000
Epoch[073] Batch [3649]/[3760]	Speed: 34.989297 samples/sec	 accuracy=78.894264	 loss=0.839952	 lr=0.001000
Epoch[073] Batch [3699]/[3760]	Speed: 35.302825 samples/sec	 accuracy=78.889358	 loss=0.840422	 lr=0.001000
Epoch[073] Batch [3749]/[3760]	Speed: 36.409230 samples/sec	 accuracy=78.882500	 loss=0.840549	 lr=0.001000
Batch [0049]/[0303]: acc-top1=67.312500 acc-top5=86.031250
Batch [0099]/[0303]: acc-top1=65.968750 acc-top5=85.421875
Batch [0149]/[0303]: acc-top1=66.177083 acc-top5=85.364583
Batch [0199]/[0303]: acc-top1=66.023438 acc-top5=85.437500
Batch [0249]/[0303]: acc-top1=65.768750 acc-top5=85.256250
Batch [0299]/[0303]: acc-top1=65.791667 acc-top5=85.328125
[Epoch 073] training: accuracy=78.875499	 loss=0.840805
[Epoch 073] speed: 34 samples/sec	time cost: 7344.002203
[Epoch 073] validation: acc-top1=65.774546 acc-top5=85.334158 loss=1.642150
Epoch[074] Batch [0049]/[3760]	Speed: 25.442452 samples/sec	 accuracy=80.656250	 loss=0.767675	 lr=0.001000
Epoch[074] Batch [0099]/[3760]	Speed: 35.412052 samples/sec	 accuracy=79.921875	 loss=0.800657	 lr=0.001000
Epoch[074] Batch [0149]/[3760]	Speed: 35.543329 samples/sec	 accuracy=79.822917	 loss=0.800121	 lr=0.001000
Epoch[074] Batch [0199]/[3760]	Speed: 35.274018 samples/sec	 accuracy=79.750000	 loss=0.804643	 lr=0.001000
Epoch[074] Batch [0249]/[3760]	Speed: 35.391885 samples/sec	 accuracy=80.125000	 loss=0.793209	 lr=0.001000
Epoch[074] Batch [0299]/[3760]	Speed: 35.261569 samples/sec	 accuracy=80.052083	 loss=0.794986	 lr=0.001000
Epoch[074] Batch [0349]/[3760]	Speed: 35.339203 samples/sec	 accuracy=79.714286	 loss=0.805565	 lr=0.001000
Epoch[074] Batch [0399]/[3760]	Speed: 35.017333 samples/sec	 accuracy=79.621094	 loss=0.814547	 lr=0.001000
Epoch[074] Batch [0449]/[3760]	Speed: 35.288708 samples/sec	 accuracy=79.527778	 loss=0.817824	 lr=0.001000
Epoch[074] Batch [0499]/[3760]	Speed: 35.186076 samples/sec	 accuracy=79.484375	 loss=0.818175	 lr=0.001000
Epoch[074] Batch [0549]/[3760]	Speed: 35.047901 samples/sec	 accuracy=79.428977	 loss=0.820503	 lr=0.001000
Epoch[074] Batch [0599]/[3760]	Speed: 35.234233 samples/sec	 accuracy=79.414062	 loss=0.820186	 lr=0.001000
Epoch[074] Batch [0649]/[3760]	Speed: 35.122260 samples/sec	 accuracy=79.375000	 loss=0.823587	 lr=0.001000
Epoch[074] Batch [0699]/[3760]	Speed: 35.443530 samples/sec	 accuracy=79.348214	 loss=0.824136	 lr=0.001000
Epoch[074] Batch [0749]/[3760]	Speed: 35.174651 samples/sec	 accuracy=79.322917	 loss=0.825247	 lr=0.001000
Epoch[074] Batch [0799]/[3760]	Speed: 35.392792 samples/sec	 accuracy=79.263672	 loss=0.826915	 lr=0.001000
Epoch[074] Batch [0849]/[3760]	Speed: 35.363155 samples/sec	 accuracy=79.235294	 loss=0.829072	 lr=0.001000
Epoch[074] Batch [0899]/[3760]	Speed: 35.300924 samples/sec	 accuracy=79.230903	 loss=0.829037	 lr=0.001000
Epoch[074] Batch [0949]/[3760]	Speed: 35.228930 samples/sec	 accuracy=79.182566	 loss=0.833275	 lr=0.001000
Epoch[074] Batch [0999]/[3760]	Speed: 35.258930 samples/sec	 accuracy=79.179688	 loss=0.832962	 lr=0.001000
Epoch[074] Batch [1049]/[3760]	Speed: 35.336746 samples/sec	 accuracy=79.154762	 loss=0.834269	 lr=0.001000
Epoch[074] Batch [1099]/[3760]	Speed: 35.249091 samples/sec	 accuracy=79.156250	 loss=0.834193	 lr=0.001000
Epoch[074] Batch [1149]/[3760]	Speed: 35.303129 samples/sec	 accuracy=79.115489	 loss=0.834807	 lr=0.001000
Epoch[074] Batch [1199]/[3760]	Speed: 35.202081 samples/sec	 accuracy=79.132812	 loss=0.834422	 lr=0.001000
Epoch[074] Batch [1249]/[3760]	Speed: 35.210848 samples/sec	 accuracy=79.131250	 loss=0.833530	 lr=0.001000
Epoch[074] Batch [1299]/[3760]	Speed: 35.343667 samples/sec	 accuracy=79.173077	 loss=0.831798	 lr=0.001000
Epoch[074] Batch [1349]/[3760]	Speed: 35.013615 samples/sec	 accuracy=79.159722	 loss=0.831832	 lr=0.001000
Epoch[074] Batch [1399]/[3760]	Speed: 35.296139 samples/sec	 accuracy=79.139509	 loss=0.831541	 lr=0.001000
Epoch[074] Batch [1449]/[3760]	Speed: 34.916944 samples/sec	 accuracy=79.161638	 loss=0.831490	 lr=0.001000
Epoch[074] Batch [1499]/[3760]	Speed: 35.450412 samples/sec	 accuracy=79.146875	 loss=0.831556	 lr=0.001000
Epoch[074] Batch [1549]/[3760]	Speed: 35.154715 samples/sec	 accuracy=79.145161	 loss=0.831468	 lr=0.001000
Epoch[074] Batch [1599]/[3760]	Speed: 35.387519 samples/sec	 accuracy=79.131836	 loss=0.832344	 lr=0.001000
Epoch[074] Batch [1649]/[3760]	Speed: 35.130546 samples/sec	 accuracy=79.126894	 loss=0.833110	 lr=0.001000
Epoch[074] Batch [1699]/[3760]	Speed: 35.229628 samples/sec	 accuracy=79.121324	 loss=0.833557	 lr=0.001000
Epoch[074] Batch [1749]/[3760]	Speed: 35.162076 samples/sec	 accuracy=79.127679	 loss=0.833805	 lr=0.001000
Epoch[074] Batch [1799]/[3760]	Speed: 35.373325 samples/sec	 accuracy=79.104167	 loss=0.835458	 lr=0.001000
Epoch[074] Batch [1849]/[3760]	Speed: 35.498246 samples/sec	 accuracy=79.094595	 loss=0.835178	 lr=0.001000
Epoch[074] Batch [1899]/[3760]	Speed: 35.151836 samples/sec	 accuracy=79.081414	 loss=0.836110	 lr=0.001000
Epoch[074] Batch [1949]/[3760]	Speed: 35.427758 samples/sec	 accuracy=79.060897	 loss=0.836701	 lr=0.001000
Epoch[074] Batch [1999]/[3760]	Speed: 35.280557 samples/sec	 accuracy=79.037500	 loss=0.837648	 lr=0.001000
Epoch[074] Batch [2049]/[3760]	Speed: 35.407912 samples/sec	 accuracy=79.067835	 loss=0.837226	 lr=0.001000
Epoch[074] Batch [2099]/[3760]	Speed: 35.113534 samples/sec	 accuracy=79.078869	 loss=0.836234	 lr=0.001000
Epoch[074] Batch [2149]/[3760]	Speed: 35.375837 samples/sec	 accuracy=79.082122	 loss=0.836233	 lr=0.001000
Epoch[074] Batch [2199]/[3760]	Speed: 35.270548 samples/sec	 accuracy=79.102983	 loss=0.835266	 lr=0.001000
Epoch[074] Batch [2249]/[3760]	Speed: 35.281572 samples/sec	 accuracy=79.088889	 loss=0.835373	 lr=0.001000
Epoch[074] Batch [2299]/[3760]	Speed: 35.105274 samples/sec	 accuracy=79.078125	 loss=0.835584	 lr=0.001000
Epoch[074] Batch [2349]/[3760]	Speed: 35.373110 samples/sec	 accuracy=79.077128	 loss=0.835705	 lr=0.001000
Epoch[074] Batch [2399]/[3760]	Speed: 35.094091 samples/sec	 accuracy=79.093750	 loss=0.835570	 lr=0.001000
Epoch[074] Batch [2449]/[3760]	Speed: 35.219348 samples/sec	 accuracy=79.081633	 loss=0.835771	 lr=0.001000
Epoch[074] Batch [2499]/[3760]	Speed: 35.280569 samples/sec	 accuracy=79.088750	 loss=0.835725	 lr=0.001000
Epoch[074] Batch [2549]/[3760]	Speed: 35.328184 samples/sec	 accuracy=79.075368	 loss=0.835592	 lr=0.001000
Epoch[074] Batch [2599]/[3760]	Speed: 35.234266 samples/sec	 accuracy=79.082933	 loss=0.834830	 lr=0.001000
Epoch[074] Batch [2649]/[3760]	Speed: 35.165668 samples/sec	 accuracy=79.080778	 loss=0.835315	 lr=0.001000
Epoch[074] Batch [2699]/[3760]	Speed: 35.244310 samples/sec	 accuracy=79.070602	 loss=0.835795	 lr=0.001000
Epoch[074] Batch [2749]/[3760]	Speed: 35.331079 samples/sec	 accuracy=79.074432	 loss=0.836032	 lr=0.001000
Epoch[074] Batch [2799]/[3760]	Speed: 35.342994 samples/sec	 accuracy=79.074777	 loss=0.835896	 lr=0.001000
Epoch[074] Batch [2849]/[3760]	Speed: 35.110878 samples/sec	 accuracy=79.083882	 loss=0.835887	 lr=0.001000
Epoch[074] Batch [2899]/[3760]	Speed: 35.222472 samples/sec	 accuracy=79.068966	 loss=0.836502	 lr=0.001000
Epoch[074] Batch [2949]/[3760]	Speed: 34.998398 samples/sec	 accuracy=79.073623	 loss=0.836710	 lr=0.001000
Epoch[074] Batch [2999]/[3760]	Speed: 35.238565 samples/sec	 accuracy=79.071875	 loss=0.836548	 lr=0.001000
Epoch[074] Batch [3049]/[3760]	Speed: 35.199369 samples/sec	 accuracy=79.079406	 loss=0.836128	 lr=0.001000
Epoch[074] Batch [3099]/[3760]	Speed: 35.493354 samples/sec	 accuracy=79.058972	 loss=0.836732	 lr=0.001000
Epoch[074] Batch [3149]/[3760]	Speed: 35.097700 samples/sec	 accuracy=79.065476	 loss=0.836233	 lr=0.001000
Epoch[074] Batch [3199]/[3760]	Speed: 35.356286 samples/sec	 accuracy=79.066895	 loss=0.836350	 lr=0.001000
Epoch[074] Batch [3249]/[3760]	Speed: 34.972700 samples/sec	 accuracy=79.081250	 loss=0.835796	 lr=0.001000
Epoch[074] Batch [3299]/[3760]	Speed: 35.364571 samples/sec	 accuracy=79.079545	 loss=0.835647	 lr=0.001000
Epoch[074] Batch [3349]/[3760]	Speed: 34.934556 samples/sec	 accuracy=79.082090	 loss=0.835884	 lr=0.001000
Epoch[074] Batch [3399]/[3760]	Speed: 35.136418 samples/sec	 accuracy=79.093290	 loss=0.835268	 lr=0.001000
Epoch[074] Batch [3449]/[3760]	Speed: 35.274264 samples/sec	 accuracy=79.111413	 loss=0.834505	 lr=0.001000
Epoch[074] Batch [3499]/[3760]	Speed: 34.996818 samples/sec	 accuracy=79.111161	 loss=0.834675	 lr=0.001000
Epoch[074] Batch [3549]/[3760]	Speed: 35.299257 samples/sec	 accuracy=79.103873	 loss=0.834682	 lr=0.001000
Epoch[074] Batch [3599]/[3760]	Speed: 34.939168 samples/sec	 accuracy=79.092014	 loss=0.834950	 lr=0.001000
Epoch[074] Batch [3649]/[3760]	Speed: 35.467160 samples/sec	 accuracy=79.096747	 loss=0.834681	 lr=0.001000
Epoch[074] Batch [3699]/[3760]	Speed: 34.812792 samples/sec	 accuracy=79.096284	 loss=0.834759	 lr=0.001000
Epoch[074] Batch [3749]/[3760]	Speed: 36.565662 samples/sec	 accuracy=79.080417	 loss=0.835484	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.281250 acc-top5=85.937500
Batch [0099]/[0303]: acc-top1=66.046875 acc-top5=85.234375
Batch [0149]/[0303]: acc-top1=66.395833 acc-top5=85.541667
Batch [0199]/[0303]: acc-top1=66.023438 acc-top5=85.437500
Batch [0249]/[0303]: acc-top1=65.793750 acc-top5=85.318750
Batch [0299]/[0303]: acc-top1=65.630208 acc-top5=85.281250
[Epoch 074] training: accuracy=79.079953	 loss=0.835413
[Epoch 074] speed: 35 samples/sec	time cost: 7312.722118
[Epoch 074] validation: acc-top1=65.625000 acc-top5=85.282591 loss=1.640277
Epoch[075] Batch [0049]/[3759]	Speed: 25.082980 samples/sec	 accuracy=79.843750	 loss=0.816289	 lr=0.001000
Epoch[075] Batch [0099]/[3759]	Speed: 35.366520 samples/sec	 accuracy=80.203125	 loss=0.783122	 lr=0.001000
Epoch[075] Batch [0149]/[3759]	Speed: 35.316030 samples/sec	 accuracy=80.052083	 loss=0.789474	 lr=0.001000
Epoch[075] Batch [0199]/[3759]	Speed: 35.320188 samples/sec	 accuracy=79.656250	 loss=0.807880	 lr=0.001000
Epoch[075] Batch [0249]/[3759]	Speed: 35.350349 samples/sec	 accuracy=79.581250	 loss=0.809643	 lr=0.001000
Epoch[075] Batch [0299]/[3759]	Speed: 35.090687 samples/sec	 accuracy=79.588542	 loss=0.807711	 lr=0.001000
Epoch[075] Batch [0349]/[3759]	Speed: 35.100642 samples/sec	 accuracy=79.464286	 loss=0.810526	 lr=0.001000
Epoch[075] Batch [0399]/[3759]	Speed: 35.307493 samples/sec	 accuracy=79.535156	 loss=0.812542	 lr=0.001000
Epoch[075] Batch [0449]/[3759]	Speed: 35.304978 samples/sec	 accuracy=79.541667	 loss=0.813530	 lr=0.001000
Epoch[075] Batch [0499]/[3759]	Speed: 35.345968 samples/sec	 accuracy=79.553125	 loss=0.811609	 lr=0.001000
Epoch[075] Batch [0549]/[3759]	Speed: 35.360529 samples/sec	 accuracy=79.551136	 loss=0.812196	 lr=0.001000
Epoch[075] Batch [0599]/[3759]	Speed: 35.072388 samples/sec	 accuracy=79.546875	 loss=0.812716	 lr=0.001000
Epoch[075] Batch [0649]/[3759]	Speed: 35.235049 samples/sec	 accuracy=79.514423	 loss=0.813677	 lr=0.001000
Epoch[075] Batch [0699]/[3759]	Speed: 35.227239 samples/sec	 accuracy=79.506696	 loss=0.813608	 lr=0.001000
Epoch[075] Batch [0749]/[3759]	Speed: 35.224905 samples/sec	 accuracy=79.535417	 loss=0.812360	 lr=0.001000
Epoch[075] Batch [0799]/[3759]	Speed: 34.925340 samples/sec	 accuracy=79.527344	 loss=0.812858	 lr=0.001000
Epoch[075] Batch [0849]/[3759]	Speed: 35.275381 samples/sec	 accuracy=79.577206	 loss=0.812519	 lr=0.001000
Epoch[075] Batch [0899]/[3759]	Speed: 35.024090 samples/sec	 accuracy=79.593750	 loss=0.813413	 lr=0.001000
Epoch[075] Batch [0949]/[3759]	Speed: 35.068874 samples/sec	 accuracy=79.597039	 loss=0.813564	 lr=0.001000
Epoch[075] Batch [0999]/[3759]	Speed: 34.934195 samples/sec	 accuracy=79.585938	 loss=0.814402	 lr=0.001000
Epoch[075] Batch [1049]/[3759]	Speed: 34.947742 samples/sec	 accuracy=79.555060	 loss=0.815539	 lr=0.001000
Epoch[075] Batch [1099]/[3759]	Speed: 35.273903 samples/sec	 accuracy=79.509943	 loss=0.816373	 lr=0.001000
Epoch[075] Batch [1149]/[3759]	Speed: 35.014371 samples/sec	 accuracy=79.517663	 loss=0.817212	 lr=0.001000
Epoch[075] Batch [1199]/[3759]	Speed: 35.279263 samples/sec	 accuracy=79.497396	 loss=0.817859	 lr=0.001000
Epoch[075] Batch [1249]/[3759]	Speed: 35.026374 samples/sec	 accuracy=79.465000	 loss=0.819547	 lr=0.001000
Epoch[075] Batch [1299]/[3759]	Speed: 35.403441 samples/sec	 accuracy=79.472356	 loss=0.819226	 lr=0.001000
Epoch[075] Batch [1349]/[3759]	Speed: 35.043256 samples/sec	 accuracy=79.457176	 loss=0.819581	 lr=0.001000
Epoch[075] Batch [1399]/[3759]	Speed: 35.303507 samples/sec	 accuracy=79.386161	 loss=0.821673	 lr=0.001000
Epoch[075] Batch [1449]/[3759]	Speed: 34.907890 samples/sec	 accuracy=79.431034	 loss=0.821086	 lr=0.001000
Epoch[075] Batch [1499]/[3759]	Speed: 35.270542 samples/sec	 accuracy=79.396875	 loss=0.822181	 lr=0.001000
Epoch[075] Batch [1549]/[3759]	Speed: 34.840839 samples/sec	 accuracy=79.388105	 loss=0.822262	 lr=0.001000
Epoch[075] Batch [1599]/[3759]	Speed: 35.244900 samples/sec	 accuracy=79.357422	 loss=0.822864	 lr=0.001000
Epoch[075] Batch [1649]/[3759]	Speed: 34.872060 samples/sec	 accuracy=79.349432	 loss=0.822267	 lr=0.001000
Epoch[075] Batch [1699]/[3759]	Speed: 35.142949 samples/sec	 accuracy=79.331801	 loss=0.822767	 lr=0.001000
Epoch[075] Batch [1749]/[3759]	Speed: 35.023928 samples/sec	 accuracy=79.338393	 loss=0.822911	 lr=0.001000
Epoch[075] Batch [1799]/[3759]	Speed: 35.120053 samples/sec	 accuracy=79.358507	 loss=0.821264	 lr=0.001000
Epoch[075] Batch [1849]/[3759]	Speed: 34.968587 samples/sec	 accuracy=79.374155	 loss=0.820789	 lr=0.001000
Epoch[075] Batch [1899]/[3759]	Speed: 35.177855 samples/sec	 accuracy=79.375000	 loss=0.821143	 lr=0.001000
Epoch[075] Batch [1949]/[3759]	Speed: 35.333946 samples/sec	 accuracy=79.367788	 loss=0.821336	 lr=0.001000
Epoch[075] Batch [1999]/[3759]	Speed: 35.189710 samples/sec	 accuracy=79.378906	 loss=0.821244	 lr=0.001000
Epoch[075] Batch [2049]/[3759]	Speed: 34.994409 samples/sec	 accuracy=79.386433	 loss=0.820988	 lr=0.001000
Epoch[075] Batch [2099]/[3759]	Speed: 35.094848 samples/sec	 accuracy=79.380952	 loss=0.821656	 lr=0.001000
Epoch[075] Batch [2149]/[3759]	Speed: 34.774331 samples/sec	 accuracy=79.382994	 loss=0.822275	 lr=0.001000
Epoch[075] Batch [2199]/[3759]	Speed: 34.949825 samples/sec	 accuracy=79.351562	 loss=0.823106	 lr=0.001000
Epoch[075] Batch [2249]/[3759]	Speed: 34.642285 samples/sec	 accuracy=79.350000	 loss=0.822967	 lr=0.001000
Epoch[075] Batch [2299]/[3759]	Speed: 34.846521 samples/sec	 accuracy=79.360734	 loss=0.822490	 lr=0.001000
Epoch[075] Batch [2349]/[3759]	Speed: 34.765454 samples/sec	 accuracy=79.365691	 loss=0.822787	 lr=0.001000
Epoch[075] Batch [2399]/[3759]	Speed: 34.787684 samples/sec	 accuracy=79.350260	 loss=0.823257	 lr=0.001000
Epoch[075] Batch [2449]/[3759]	Speed: 35.034682 samples/sec	 accuracy=79.357781	 loss=0.823108	 lr=0.001000
Epoch[075] Batch [2499]/[3759]	Speed: 34.519138 samples/sec	 accuracy=79.346875	 loss=0.823758	 lr=0.001000
Epoch[075] Batch [2549]/[3759]	Speed: 35.186603 samples/sec	 accuracy=79.365809	 loss=0.823333	 lr=0.001000
Epoch[075] Batch [2599]/[3759]	Speed: 34.459980 samples/sec	 accuracy=79.356971	 loss=0.823646	 lr=0.001000
Epoch[075] Batch [2649]/[3759]	Speed: 35.159018 samples/sec	 accuracy=79.377948	 loss=0.823609	 lr=0.001000
Epoch[075] Batch [2699]/[3759]	Speed: 34.673305 samples/sec	 accuracy=79.377894	 loss=0.824073	 lr=0.001000
Epoch[075] Batch [2749]/[3759]	Speed: 35.106304 samples/sec	 accuracy=79.385795	 loss=0.823641	 lr=0.001000
Epoch[075] Batch [2799]/[3759]	Speed: 34.910485 samples/sec	 accuracy=79.376674	 loss=0.823895	 lr=0.001000
Epoch[075] Batch [2849]/[3759]	Speed: 34.973940 samples/sec	 accuracy=79.374452	 loss=0.823717	 lr=0.001000
Epoch[075] Batch [2899]/[3759]	Speed: 35.076870 samples/sec	 accuracy=79.384159	 loss=0.823460	 lr=0.001000
Epoch[075] Batch [2949]/[3759]	Speed: 34.881046 samples/sec	 accuracy=79.389301	 loss=0.823573	 lr=0.001000
Epoch[075] Batch [2999]/[3759]	Speed: 35.162220 samples/sec	 accuracy=79.386979	 loss=0.823467	 lr=0.001000
Epoch[075] Batch [3049]/[3759]	Speed: 35.066257 samples/sec	 accuracy=79.388320	 loss=0.823618	 lr=0.001000
Epoch[075] Batch [3099]/[3759]	Speed: 35.267181 samples/sec	 accuracy=79.377016	 loss=0.823685	 lr=0.001000
Epoch[075] Batch [3149]/[3759]	Speed: 35.041414 samples/sec	 accuracy=79.370040	 loss=0.823733	 lr=0.001000
Epoch[075] Batch [3199]/[3759]	Speed: 35.276019 samples/sec	 accuracy=79.394531	 loss=0.822922	 lr=0.001000
Epoch[075] Batch [3249]/[3759]	Speed: 34.965174 samples/sec	 accuracy=79.373077	 loss=0.823762	 lr=0.001000
Epoch[075] Batch [3299]/[3759]	Speed: 35.128124 samples/sec	 accuracy=79.365057	 loss=0.824027	 lr=0.001000
Epoch[075] Batch [3349]/[3759]	Speed: 34.545041 samples/sec	 accuracy=79.364272	 loss=0.824069	 lr=0.001000
Epoch[075] Batch [3399]/[3759]	Speed: 35.119059 samples/sec	 accuracy=79.368566	 loss=0.823975	 lr=0.001000
Epoch[075] Batch [3449]/[3759]	Speed: 34.266764 samples/sec	 accuracy=79.350091	 loss=0.824337	 lr=0.001000
Epoch[075] Batch [3499]/[3759]	Speed: 35.089475 samples/sec	 accuracy=79.334821	 loss=0.824596	 lr=0.001000
Epoch[075] Batch [3549]/[3759]	Speed: 34.382209 samples/sec	 accuracy=79.347711	 loss=0.823689	 lr=0.001000
Epoch[075] Batch [3599]/[3759]	Speed: 34.995626 samples/sec	 accuracy=79.327691	 loss=0.824319	 lr=0.001000
Epoch[075] Batch [3649]/[3759]	Speed: 34.804482 samples/sec	 accuracy=79.327911	 loss=0.824547	 lr=0.001000
Epoch[075] Batch [3699]/[3759]	Speed: 34.606998 samples/sec	 accuracy=79.339527	 loss=0.824514	 lr=0.001000
Epoch[075] Batch [3749]/[3759]	Speed: 36.438754 samples/sec	 accuracy=79.325000	 loss=0.824674	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.562500 acc-top5=85.187500
Batch [0099]/[0303]: acc-top1=65.906250 acc-top5=84.843750
Batch [0149]/[0303]: acc-top1=66.281250 acc-top5=85.239583
Batch [0199]/[0303]: acc-top1=65.914062 acc-top5=85.226562
Batch [0249]/[0303]: acc-top1=65.656250 acc-top5=85.137500
Batch [0299]/[0303]: acc-top1=65.630208 acc-top5=85.098958
[Epoch 075] training: accuracy=79.323790	 loss=0.824720
[Epoch 075] speed: 34 samples/sec	time cost: 7350.525818
[Epoch 075] validation: acc-top1=65.619843 acc-top5=85.107261 loss=1.660125
Epoch[076] Batch [0049]/[3760]	Speed: 24.849027 samples/sec	 accuracy=80.125000	 loss=0.821614	 lr=0.001000
Epoch[076] Batch [0099]/[3760]	Speed: 35.500456 samples/sec	 accuracy=80.156250	 loss=0.804582	 lr=0.001000
Epoch[076] Batch [0149]/[3760]	Speed: 35.024588 samples/sec	 accuracy=80.520833	 loss=0.794636	 lr=0.001000
Epoch[076] Batch [0199]/[3760]	Speed: 35.142059 samples/sec	 accuracy=80.414062	 loss=0.798649	 lr=0.001000
Epoch[076] Batch [0249]/[3760]	Speed: 34.730582 samples/sec	 accuracy=80.100000	 loss=0.806915	 lr=0.001000
Epoch[076] Batch [0299]/[3760]	Speed: 35.429102 samples/sec	 accuracy=79.739583	 loss=0.814498	 lr=0.001000
Epoch[076] Batch [0349]/[3760]	Speed: 34.822429 samples/sec	 accuracy=79.656250	 loss=0.815829	 lr=0.001000
Epoch[076] Batch [0399]/[3760]	Speed: 35.475604 samples/sec	 accuracy=79.648438	 loss=0.814534	 lr=0.001000
Epoch[076] Batch [0449]/[3760]	Speed: 35.010939 samples/sec	 accuracy=79.541667	 loss=0.817034	 lr=0.001000
Epoch[076] Batch [0499]/[3760]	Speed: 35.355129 samples/sec	 accuracy=79.559375	 loss=0.815710	 lr=0.001000
Epoch[076] Batch [0549]/[3760]	Speed: 34.924870 samples/sec	 accuracy=79.650568	 loss=0.813328	 lr=0.001000
Epoch[076] Batch [0599]/[3760]	Speed: 35.438734 samples/sec	 accuracy=79.638021	 loss=0.812764	 lr=0.001000
Epoch[076] Batch [0649]/[3760]	Speed: 35.196685 samples/sec	 accuracy=79.680288	 loss=0.810608	 lr=0.001000
Epoch[076] Batch [0699]/[3760]	Speed: 35.153848 samples/sec	 accuracy=79.698661	 loss=0.809576	 lr=0.001000
Epoch[076] Batch [0749]/[3760]	Speed: 35.002509 samples/sec	 accuracy=79.754167	 loss=0.809491	 lr=0.001000
Epoch[076] Batch [0799]/[3760]	Speed: 35.044320 samples/sec	 accuracy=79.753906	 loss=0.809979	 lr=0.001000
Epoch[076] Batch [0849]/[3760]	Speed: 35.062349 samples/sec	 accuracy=79.768382	 loss=0.808378	 lr=0.001000
Epoch[076] Batch [0899]/[3760]	Speed: 34.697744 samples/sec	 accuracy=79.732639	 loss=0.810108	 lr=0.001000
Epoch[076] Batch [0949]/[3760]	Speed: 35.186635 samples/sec	 accuracy=79.705592	 loss=0.811328	 lr=0.001000
Epoch[076] Batch [0999]/[3760]	Speed: 34.765591 samples/sec	 accuracy=79.693750	 loss=0.811490	 lr=0.001000
Epoch[076] Batch [1049]/[3760]	Speed: 35.356407 samples/sec	 accuracy=79.672619	 loss=0.811556	 lr=0.001000
Epoch[076] Batch [1099]/[3760]	Speed: 34.846547 samples/sec	 accuracy=79.724432	 loss=0.810028	 lr=0.001000
Epoch[076] Batch [1149]/[3760]	Speed: 35.313136 samples/sec	 accuracy=79.671196	 loss=0.810683	 lr=0.001000
Epoch[076] Batch [1199]/[3760]	Speed: 35.047169 samples/sec	 accuracy=79.703125	 loss=0.810565	 lr=0.001000
Epoch[076] Batch [1249]/[3760]	Speed: 35.157093 samples/sec	 accuracy=79.691250	 loss=0.810129	 lr=0.001000
Epoch[076] Batch [1299]/[3760]	Speed: 35.049641 samples/sec	 accuracy=79.693510	 loss=0.809773	 lr=0.001000
Epoch[076] Batch [1349]/[3760]	Speed: 35.213956 samples/sec	 accuracy=79.662037	 loss=0.810906	 lr=0.001000
Epoch[076] Batch [1399]/[3760]	Speed: 35.273691 samples/sec	 accuracy=79.672991	 loss=0.811452	 lr=0.001000
Epoch[076] Batch [1449]/[3760]	Speed: 35.169583 samples/sec	 accuracy=79.682112	 loss=0.811380	 lr=0.001000
Epoch[076] Batch [1499]/[3760]	Speed: 35.224682 samples/sec	 accuracy=79.700000	 loss=0.810830	 lr=0.001000
Epoch[076] Batch [1549]/[3760]	Speed: 35.386029 samples/sec	 accuracy=79.669355	 loss=0.812333	 lr=0.001000
Epoch[076] Batch [1599]/[3760]	Speed: 35.177094 samples/sec	 accuracy=79.663086	 loss=0.811944	 lr=0.001000
Epoch[076] Batch [1649]/[3760]	Speed: 35.186848 samples/sec	 accuracy=79.691288	 loss=0.810323	 lr=0.001000
Epoch[076] Batch [1699]/[3760]	Speed: 35.337713 samples/sec	 accuracy=79.669118	 loss=0.811141	 lr=0.001000
Epoch[076] Batch [1749]/[3760]	Speed: 35.240165 samples/sec	 accuracy=79.647321	 loss=0.812863	 lr=0.001000
Epoch[076] Batch [1799]/[3760]	Speed: 35.127893 samples/sec	 accuracy=79.610243	 loss=0.813778	 lr=0.001000
Epoch[076] Batch [1849]/[3760]	Speed: 35.199221 samples/sec	 accuracy=79.565878	 loss=0.815177	 lr=0.001000
Epoch[076] Batch [1899]/[3760]	Speed: 35.357079 samples/sec	 accuracy=79.540296	 loss=0.815981	 lr=0.001000
Epoch[076] Batch [1949]/[3760]	Speed: 35.131326 samples/sec	 accuracy=79.496795	 loss=0.817365	 lr=0.001000
Epoch[076] Batch [1999]/[3760]	Speed: 35.315881 samples/sec	 accuracy=79.495312	 loss=0.817468	 lr=0.001000
Epoch[076] Batch [2049]/[3760]	Speed: 35.146006 samples/sec	 accuracy=79.474085	 loss=0.817951	 lr=0.001000
Epoch[076] Batch [2099]/[3760]	Speed: 35.258053 samples/sec	 accuracy=79.453125	 loss=0.818868	 lr=0.001000
Epoch[076] Batch [2149]/[3760]	Speed: 35.083780 samples/sec	 accuracy=79.443314	 loss=0.819838	 lr=0.001000
Epoch[076] Batch [2199]/[3760]	Speed: 35.451732 samples/sec	 accuracy=79.424716	 loss=0.820267	 lr=0.001000
Epoch[076] Batch [2249]/[3760]	Speed: 35.192399 samples/sec	 accuracy=79.423611	 loss=0.820570	 lr=0.001000
Epoch[076] Batch [2299]/[3760]	Speed: 35.294292 samples/sec	 accuracy=79.427310	 loss=0.820361	 lr=0.001000
Epoch[076] Batch [2349]/[3760]	Speed: 35.259863 samples/sec	 accuracy=79.416223	 loss=0.820975	 lr=0.001000
Epoch[076] Batch [2399]/[3760]	Speed: 35.384040 samples/sec	 accuracy=79.405599	 loss=0.820810	 lr=0.001000
Epoch[076] Batch [2449]/[3760]	Speed: 35.206633 samples/sec	 accuracy=79.400510	 loss=0.821002	 lr=0.001000
Epoch[076] Batch [2499]/[3760]	Speed: 35.391938 samples/sec	 accuracy=79.416250	 loss=0.820723	 lr=0.001000
Epoch[076] Batch [2549]/[3760]	Speed: 35.102874 samples/sec	 accuracy=79.415441	 loss=0.821358	 lr=0.001000
Epoch[076] Batch [2599]/[3760]	Speed: 35.375593 samples/sec	 accuracy=79.411659	 loss=0.821476	 lr=0.001000
Epoch[076] Batch [2649]/[3760]	Speed: 35.037025 samples/sec	 accuracy=79.412736	 loss=0.820843	 lr=0.001000
Epoch[076] Batch [2699]/[3760]	Speed: 35.328042 samples/sec	 accuracy=79.415509	 loss=0.821003	 lr=0.001000
Epoch[076] Batch [2749]/[3760]	Speed: 34.980411 samples/sec	 accuracy=79.438068	 loss=0.820756	 lr=0.001000
Epoch[076] Batch [2799]/[3760]	Speed: 35.022443 samples/sec	 accuracy=79.426339	 loss=0.821611	 lr=0.001000
Epoch[076] Batch [2849]/[3760]	Speed: 34.929190 samples/sec	 accuracy=79.411184	 loss=0.822125	 lr=0.001000
Epoch[076] Batch [2899]/[3760]	Speed: 35.061127 samples/sec	 accuracy=79.406789	 loss=0.822178	 lr=0.001000
Epoch[076] Batch [2949]/[3760]	Speed: 35.136675 samples/sec	 accuracy=79.415254	 loss=0.821957	 lr=0.001000
Epoch[076] Batch [2999]/[3760]	Speed: 34.767054 samples/sec	 accuracy=79.413542	 loss=0.821660	 lr=0.001000
Epoch[076] Batch [3049]/[3760]	Speed: 35.272165 samples/sec	 accuracy=79.415471	 loss=0.821132	 lr=0.001000
Epoch[076] Batch [3099]/[3760]	Speed: 34.740945 samples/sec	 accuracy=79.402722	 loss=0.821692	 lr=0.001000
Epoch[076] Batch [3149]/[3760]	Speed: 35.328001 samples/sec	 accuracy=79.407242	 loss=0.821795	 lr=0.001000
Epoch[076] Batch [3199]/[3760]	Speed: 34.976661 samples/sec	 accuracy=79.408203	 loss=0.821584	 lr=0.001000
Epoch[076] Batch [3249]/[3760]	Speed: 35.351051 samples/sec	 accuracy=79.410096	 loss=0.821455	 lr=0.001000
Epoch[076] Batch [3299]/[3760]	Speed: 35.234206 samples/sec	 accuracy=79.407670	 loss=0.821683	 lr=0.001000
Epoch[076] Batch [3349]/[3760]	Speed: 35.193593 samples/sec	 accuracy=79.389925	 loss=0.822143	 lr=0.001000
Epoch[076] Batch [3399]/[3760]	Speed: 35.173689 samples/sec	 accuracy=79.396599	 loss=0.822392	 lr=0.001000
Epoch[076] Batch [3449]/[3760]	Speed: 35.481794 samples/sec	 accuracy=79.384964	 loss=0.822818	 lr=0.001000
Epoch[076] Batch [3499]/[3760]	Speed: 35.228272 samples/sec	 accuracy=79.375446	 loss=0.823395	 lr=0.001000
Epoch[076] Batch [3549]/[3760]	Speed: 35.470708 samples/sec	 accuracy=79.377201	 loss=0.822949	 lr=0.001000
Epoch[076] Batch [3599]/[3760]	Speed: 35.404620 samples/sec	 accuracy=79.384549	 loss=0.822489	 lr=0.001000
Epoch[076] Batch [3649]/[3760]	Speed: 35.294050 samples/sec	 accuracy=79.387414	 loss=0.822327	 lr=0.001000
Epoch[076] Batch [3699]/[3760]	Speed: 35.322397 samples/sec	 accuracy=79.389780	 loss=0.822477	 lr=0.001000
Epoch[076] Batch [3749]/[3760]	Speed: 36.426533 samples/sec	 accuracy=79.374583	 loss=0.823143	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.000000 acc-top5=85.593750
Batch [0099]/[0303]: acc-top1=65.609375 acc-top5=85.109375
Batch [0149]/[0303]: acc-top1=66.062500 acc-top5=85.343750
Batch [0199]/[0303]: acc-top1=65.703125 acc-top5=85.257812
Batch [0249]/[0303]: acc-top1=65.550000 acc-top5=85.218750
Batch [0299]/[0303]: acc-top1=65.442708 acc-top5=85.208333
[Epoch 076] training: accuracy=79.371676	 loss=0.823395
[Epoch 076] speed: 34 samples/sec	time cost: 7328.937367
[Epoch 076] validation: acc-top1=65.429043 acc-top5=85.225866 loss=1.661502
Epoch[077] Batch [0049]/[3760]	Speed: 25.361389 samples/sec	 accuracy=80.968750	 loss=0.767165	 lr=0.001000
Epoch[077] Batch [0099]/[3760]	Speed: 35.236507 samples/sec	 accuracy=80.593750	 loss=0.783183	 lr=0.001000
Epoch[077] Batch [0149]/[3760]	Speed: 35.414200 samples/sec	 accuracy=80.062500	 loss=0.796557	 lr=0.001000
Epoch[077] Batch [0199]/[3760]	Speed: 35.324915 samples/sec	 accuracy=80.125000	 loss=0.791311	 lr=0.001000
Epoch[077] Batch [0249]/[3760]	Speed: 35.319751 samples/sec	 accuracy=80.218750	 loss=0.790901	 lr=0.001000
Epoch[077] Batch [0299]/[3760]	Speed: 35.086697 samples/sec	 accuracy=80.036458	 loss=0.797916	 lr=0.001000
Epoch[077] Batch [0349]/[3760]	Speed: 35.194421 samples/sec	 accuracy=79.901786	 loss=0.800981	 lr=0.001000
Epoch[077] Batch [0399]/[3760]	Speed: 35.201984 samples/sec	 accuracy=80.046875	 loss=0.795136	 lr=0.001000
Epoch[077] Batch [0449]/[3760]	Speed: 35.614620 samples/sec	 accuracy=80.027778	 loss=0.798579	 lr=0.001000
Epoch[077] Batch [0499]/[3760]	Speed: 35.262708 samples/sec	 accuracy=79.943750	 loss=0.804672	 lr=0.001000
Epoch[077] Batch [0549]/[3760]	Speed: 35.502314 samples/sec	 accuracy=79.860795	 loss=0.806804	 lr=0.001000
Epoch[077] Batch [0599]/[3760]	Speed: 35.517329 samples/sec	 accuracy=79.807292	 loss=0.806855	 lr=0.001000
Epoch[077] Batch [0649]/[3760]	Speed: 35.380458 samples/sec	 accuracy=79.862981	 loss=0.804858	 lr=0.001000
Epoch[077] Batch [0699]/[3760]	Speed: 35.197241 samples/sec	 accuracy=79.883929	 loss=0.804281	 lr=0.001000
Epoch[077] Batch [0749]/[3760]	Speed: 35.325227 samples/sec	 accuracy=79.870833	 loss=0.803998	 lr=0.001000
Epoch[077] Batch [0799]/[3760]	Speed: 35.435140 samples/sec	 accuracy=79.832031	 loss=0.806374	 lr=0.001000
Epoch[077] Batch [0849]/[3760]	Speed: 35.443499 samples/sec	 accuracy=79.847426	 loss=0.804602	 lr=0.001000
Epoch[077] Batch [0899]/[3760]	Speed: 35.297197 samples/sec	 accuracy=79.848958	 loss=0.804327	 lr=0.001000
Epoch[077] Batch [0949]/[3760]	Speed: 35.432560 samples/sec	 accuracy=79.794408	 loss=0.808225	 lr=0.001000
Epoch[077] Batch [0999]/[3760]	Speed: 35.342753 samples/sec	 accuracy=79.721875	 loss=0.811077	 lr=0.001000
Epoch[077] Batch [1049]/[3760]	Speed: 35.057359 samples/sec	 accuracy=79.669643	 loss=0.813238	 lr=0.001000
Epoch[077] Batch [1099]/[3760]	Speed: 35.349687 samples/sec	 accuracy=79.714489	 loss=0.810532	 lr=0.001000
Epoch[077] Batch [1149]/[3760]	Speed: 35.301237 samples/sec	 accuracy=79.773098	 loss=0.808608	 lr=0.001000
Epoch[077] Batch [1199]/[3760]	Speed: 35.192249 samples/sec	 accuracy=79.753906	 loss=0.809281	 lr=0.001000
Epoch[077] Batch [1249]/[3760]	Speed: 35.428854 samples/sec	 accuracy=79.710000	 loss=0.811517	 lr=0.001000
Epoch[077] Batch [1299]/[3760]	Speed: 35.084850 samples/sec	 accuracy=79.719952	 loss=0.809829	 lr=0.001000
Epoch[077] Batch [1349]/[3760]	Speed: 35.257064 samples/sec	 accuracy=79.681713	 loss=0.810100	 lr=0.001000
Epoch[077] Batch [1399]/[3760]	Speed: 35.183650 samples/sec	 accuracy=79.699777	 loss=0.808916	 lr=0.001000
Epoch[077] Batch [1449]/[3760]	Speed: 35.277359 samples/sec	 accuracy=79.720905	 loss=0.807842	 lr=0.001000
Epoch[077] Batch [1499]/[3760]	Speed: 35.159866 samples/sec	 accuracy=79.722917	 loss=0.806982	 lr=0.001000
Epoch[077] Batch [1549]/[3760]	Speed: 35.171811 samples/sec	 accuracy=79.768145	 loss=0.806007	 lr=0.001000
Epoch[077] Batch [1599]/[3760]	Speed: 35.403460 samples/sec	 accuracy=79.763672	 loss=0.806017	 lr=0.001000
Epoch[077] Batch [1649]/[3760]	Speed: 35.251315 samples/sec	 accuracy=79.765152	 loss=0.805601	 lr=0.001000
Epoch[077] Batch [1699]/[3760]	Speed: 35.099831 samples/sec	 accuracy=79.746324	 loss=0.806944	 lr=0.001000
Epoch[077] Batch [1749]/[3760]	Speed: 35.566095 samples/sec	 accuracy=79.733036	 loss=0.807411	 lr=0.001000
Epoch[077] Batch [1799]/[3760]	Speed: 34.996230 samples/sec	 accuracy=79.706597	 loss=0.808752	 lr=0.001000
Epoch[077] Batch [1849]/[3760]	Speed: 35.300688 samples/sec	 accuracy=79.731419	 loss=0.807319	 lr=0.001000
Epoch[077] Batch [1899]/[3760]	Speed: 35.278422 samples/sec	 accuracy=79.691612	 loss=0.808779	 lr=0.001000
Epoch[077] Batch [1949]/[3760]	Speed: 35.290237 samples/sec	 accuracy=79.684295	 loss=0.809212	 lr=0.001000
Epoch[077] Batch [1999]/[3760]	Speed: 35.187452 samples/sec	 accuracy=79.664062	 loss=0.809612	 lr=0.001000
Epoch[077] Batch [2049]/[3760]	Speed: 35.227400 samples/sec	 accuracy=79.631860	 loss=0.810301	 lr=0.001000
Epoch[077] Batch [2099]/[3760]	Speed: 35.393336 samples/sec	 accuracy=79.639137	 loss=0.810193	 lr=0.001000
Epoch[077] Batch [2149]/[3760]	Speed: 35.109188 samples/sec	 accuracy=79.641715	 loss=0.809365	 lr=0.001000
Epoch[077] Batch [2199]/[3760]	Speed: 35.408482 samples/sec	 accuracy=79.654830	 loss=0.809165	 lr=0.001000
Epoch[077] Batch [2249]/[3760]	Speed: 35.251611 samples/sec	 accuracy=79.643056	 loss=0.809677	 lr=0.001000
Epoch[077] Batch [2299]/[3760]	Speed: 35.217631 samples/sec	 accuracy=79.645380	 loss=0.810264	 lr=0.001000
Epoch[077] Batch [2349]/[3760]	Speed: 35.372653 samples/sec	 accuracy=79.620346	 loss=0.810240	 lr=0.001000
Epoch[077] Batch [2399]/[3760]	Speed: 35.181573 samples/sec	 accuracy=79.625000	 loss=0.810244	 lr=0.001000
Epoch[077] Batch [2449]/[3760]	Speed: 35.432413 samples/sec	 accuracy=79.625638	 loss=0.810568	 lr=0.001000
Epoch[077] Batch [2499]/[3760]	Speed: 35.326339 samples/sec	 accuracy=79.635000	 loss=0.810111	 lr=0.001000
Epoch[077] Batch [2549]/[3760]	Speed: 35.500233 samples/sec	 accuracy=79.646446	 loss=0.810396	 lr=0.001000
Epoch[077] Batch [2599]/[3760]	Speed: 35.224201 samples/sec	 accuracy=79.665264	 loss=0.810174	 lr=0.001000
Epoch[077] Batch [2649]/[3760]	Speed: 35.390509 samples/sec	 accuracy=79.647406	 loss=0.810883	 lr=0.001000
Epoch[077] Batch [2699]/[3760]	Speed: 35.258004 samples/sec	 accuracy=79.637731	 loss=0.811146	 lr=0.001000
Epoch[077] Batch [2749]/[3760]	Speed: 35.004504 samples/sec	 accuracy=79.640341	 loss=0.811427	 lr=0.001000
Epoch[077] Batch [2799]/[3760]	Speed: 35.376725 samples/sec	 accuracy=79.642299	 loss=0.811808	 lr=0.001000
Epoch[077] Batch [2849]/[3760]	Speed: 34.981826 samples/sec	 accuracy=79.640351	 loss=0.811904	 lr=0.001000
Epoch[077] Batch [2899]/[3760]	Speed: 35.266723 samples/sec	 accuracy=79.609914	 loss=0.812955	 lr=0.001000
Epoch[077] Batch [2949]/[3760]	Speed: 35.015276 samples/sec	 accuracy=79.587924	 loss=0.813789	 lr=0.001000
Epoch[077] Batch [2999]/[3760]	Speed: 35.173972 samples/sec	 accuracy=79.565104	 loss=0.814369	 lr=0.001000
Epoch[077] Batch [3049]/[3760]	Speed: 35.386519 samples/sec	 accuracy=79.572746	 loss=0.814011	 lr=0.001000
Epoch[077] Batch [3099]/[3760]	Speed: 35.362127 samples/sec	 accuracy=79.580645	 loss=0.813669	 lr=0.001000
Epoch[077] Batch [3149]/[3760]	Speed: 35.363846 samples/sec	 accuracy=79.588294	 loss=0.813957	 lr=0.001000
Epoch[077] Batch [3199]/[3760]	Speed: 35.108828 samples/sec	 accuracy=79.593750	 loss=0.813521	 lr=0.001000
Epoch[077] Batch [3249]/[3760]	Speed: 35.097608 samples/sec	 accuracy=79.576442	 loss=0.813690	 lr=0.001000
Epoch[077] Batch [3299]/[3760]	Speed: 35.212888 samples/sec	 accuracy=79.565814	 loss=0.814136	 lr=0.001000
Epoch[077] Batch [3349]/[3760]	Speed: 35.486209 samples/sec	 accuracy=79.561101	 loss=0.814207	 lr=0.001000
Epoch[077] Batch [3399]/[3760]	Speed: 35.320292 samples/sec	 accuracy=79.559283	 loss=0.814103	 lr=0.001000
Epoch[077] Batch [3449]/[3760]	Speed: 35.291050 samples/sec	 accuracy=79.552536	 loss=0.814437	 lr=0.001000
Epoch[077] Batch [3499]/[3760]	Speed: 35.102316 samples/sec	 accuracy=79.533482	 loss=0.815348	 lr=0.001000
Epoch[077] Batch [3549]/[3760]	Speed: 35.392059 samples/sec	 accuracy=79.524208	 loss=0.815711	 lr=0.001000
Epoch[077] Batch [3599]/[3760]	Speed: 35.083344 samples/sec	 accuracy=79.521267	 loss=0.815446	 lr=0.001000
Epoch[077] Batch [3649]/[3760]	Speed: 35.409173 samples/sec	 accuracy=79.515839	 loss=0.815807	 lr=0.001000
Epoch[077] Batch [3699]/[3760]	Speed: 35.125087 samples/sec	 accuracy=79.518159	 loss=0.815447	 lr=0.001000
Epoch[077] Batch [3749]/[3760]	Speed: 36.415674 samples/sec	 accuracy=79.505833	 loss=0.816331	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.406250 acc-top5=85.562500
Batch [0099]/[0303]: acc-top1=66.406250 acc-top5=85.078125
Batch [0149]/[0303]: acc-top1=66.614583 acc-top5=85.145833
Batch [0199]/[0303]: acc-top1=66.132812 acc-top5=85.320312
Batch [0249]/[0303]: acc-top1=65.731250 acc-top5=85.200000
Batch [0299]/[0303]: acc-top1=65.505208 acc-top5=85.182292
[Epoch 077] training: accuracy=79.504654	 loss=0.816373
[Epoch 077] speed: 35 samples/sec	time cost: 7306.152020
[Epoch 077] validation: acc-top1=65.496081 acc-top5=85.205239 loss=1.650438
Epoch[078] Batch [0049]/[3759]	Speed: 25.172049 samples/sec	 accuracy=80.312500	 loss=0.806261	 lr=0.001000
Epoch[078] Batch [0099]/[3759]	Speed: 35.263794 samples/sec	 accuracy=80.000000	 loss=0.807597	 lr=0.001000
Epoch[078] Batch [0149]/[3759]	Speed: 35.467122 samples/sec	 accuracy=79.468750	 loss=0.815354	 lr=0.001000
Epoch[078] Batch [0199]/[3759]	Speed: 35.304210 samples/sec	 accuracy=79.804688	 loss=0.806914	 lr=0.001000
Epoch[078] Batch [0249]/[3759]	Speed: 35.309595 samples/sec	 accuracy=79.943750	 loss=0.801344	 lr=0.001000
Epoch[078] Batch [0299]/[3759]	Speed: 35.439888 samples/sec	 accuracy=80.072917	 loss=0.792780	 lr=0.001000
Epoch[078] Batch [0349]/[3759]	Speed: 35.182821 samples/sec	 accuracy=80.129464	 loss=0.790646	 lr=0.001000
Epoch[078] Batch [0399]/[3759]	Speed: 35.199566 samples/sec	 accuracy=80.109375	 loss=0.794032	 lr=0.001000
Epoch[078] Batch [0449]/[3759]	Speed: 35.339522 samples/sec	 accuracy=80.118056	 loss=0.791544	 lr=0.001000
Epoch[078] Batch [0499]/[3759]	Speed: 35.269727 samples/sec	 accuracy=80.021875	 loss=0.797237	 lr=0.001000
Epoch[078] Batch [0549]/[3759]	Speed: 35.280514 samples/sec	 accuracy=80.000000	 loss=0.798404	 lr=0.001000
Epoch[078] Batch [0599]/[3759]	Speed: 35.317654 samples/sec	 accuracy=79.958333	 loss=0.798530	 lr=0.001000
Epoch[078] Batch [0649]/[3759]	Speed: 35.417649 samples/sec	 accuracy=79.951923	 loss=0.797672	 lr=0.001000
Epoch[078] Batch [0699]/[3759]	Speed: 35.174964 samples/sec	 accuracy=79.917411	 loss=0.798708	 lr=0.001000
Epoch[078] Batch [0749]/[3759]	Speed: 35.445137 samples/sec	 accuracy=79.943750	 loss=0.796518	 lr=0.001000
Epoch[078] Batch [0799]/[3759]	Speed: 35.094602 samples/sec	 accuracy=79.955078	 loss=0.795138	 lr=0.001000
Epoch[078] Batch [0849]/[3759]	Speed: 35.302477 samples/sec	 accuracy=79.893382	 loss=0.795531	 lr=0.001000
Epoch[078] Batch [0899]/[3759]	Speed: 35.219568 samples/sec	 accuracy=79.916667	 loss=0.796214	 lr=0.001000
Epoch[078] Batch [0949]/[3759]	Speed: 35.351157 samples/sec	 accuracy=79.886513	 loss=0.796529	 lr=0.001000
Epoch[078] Batch [0999]/[3759]	Speed: 35.166275 samples/sec	 accuracy=79.910938	 loss=0.796912	 lr=0.001000
Epoch[078] Batch [1049]/[3759]	Speed: 35.273516 samples/sec	 accuracy=79.916667	 loss=0.796302	 lr=0.001000
Epoch[078] Batch [1099]/[3759]	Speed: 35.133124 samples/sec	 accuracy=79.883523	 loss=0.796234	 lr=0.001000
Epoch[078] Batch [1149]/[3759]	Speed: 35.267468 samples/sec	 accuracy=79.891304	 loss=0.796683	 lr=0.001000
Epoch[078] Batch [1199]/[3759]	Speed: 34.874327 samples/sec	 accuracy=79.851562	 loss=0.797266	 lr=0.001000
Epoch[078] Batch [1249]/[3759]	Speed: 35.047951 samples/sec	 accuracy=79.821250	 loss=0.800536	 lr=0.001000
Epoch[078] Batch [1299]/[3759]	Speed: 35.117679 samples/sec	 accuracy=79.788462	 loss=0.801733	 lr=0.001000
Epoch[078] Batch [1349]/[3759]	Speed: 35.173058 samples/sec	 accuracy=79.793981	 loss=0.800893	 lr=0.001000
Epoch[078] Batch [1399]/[3759]	Speed: 35.114104 samples/sec	 accuracy=79.786830	 loss=0.801888	 lr=0.001000
Epoch[078] Batch [1449]/[3759]	Speed: 35.026751 samples/sec	 accuracy=79.765086	 loss=0.801682	 lr=0.001000
Epoch[078] Batch [1499]/[3759]	Speed: 35.149308 samples/sec	 accuracy=79.805208	 loss=0.801837	 lr=0.001000
Epoch[078] Batch [1549]/[3759]	Speed: 34.832482 samples/sec	 accuracy=79.795363	 loss=0.801551	 lr=0.001000
Epoch[078] Batch [1599]/[3759]	Speed: 35.303424 samples/sec	 accuracy=79.814453	 loss=0.800942	 lr=0.001000
Epoch[078] Batch [1649]/[3759]	Speed: 34.871984 samples/sec	 accuracy=79.827652	 loss=0.800230	 lr=0.001000
Epoch[078] Batch [1699]/[3759]	Speed: 35.180285 samples/sec	 accuracy=79.847426	 loss=0.799466	 lr=0.001000
Epoch[078] Batch [1749]/[3759]	Speed: 34.859204 samples/sec	 accuracy=79.813393	 loss=0.801162	 lr=0.001000
Epoch[078] Batch [1799]/[3759]	Speed: 35.188217 samples/sec	 accuracy=79.799479	 loss=0.801975	 lr=0.001000
Epoch[078] Batch [1849]/[3759]	Speed: 34.867993 samples/sec	 accuracy=79.821791	 loss=0.800542	 lr=0.001000
Epoch[078] Batch [1899]/[3759]	Speed: 35.000789 samples/sec	 accuracy=79.844572	 loss=0.799624	 lr=0.001000
Epoch[078] Batch [1949]/[3759]	Speed: 34.567375 samples/sec	 accuracy=79.851763	 loss=0.798906	 lr=0.001000
Epoch[078] Batch [1999]/[3759]	Speed: 34.941243 samples/sec	 accuracy=79.853906	 loss=0.798894	 lr=0.001000
Epoch[078] Batch [2049]/[3759]	Speed: 34.633860 samples/sec	 accuracy=79.865854	 loss=0.798451	 lr=0.001000
Epoch[078] Batch [2099]/[3759]	Speed: 34.763336 samples/sec	 accuracy=79.856399	 loss=0.799257	 lr=0.001000
Epoch[078] Batch [2149]/[3759]	Speed: 35.026750 samples/sec	 accuracy=79.819041	 loss=0.800448	 lr=0.001000
Epoch[078] Batch [2199]/[3759]	Speed: 34.530926 samples/sec	 accuracy=79.798295	 loss=0.800872	 lr=0.001000
Epoch[078] Batch [2249]/[3759]	Speed: 35.208412 samples/sec	 accuracy=79.797917	 loss=0.801049	 lr=0.001000
Epoch[078] Batch [2299]/[3759]	Speed: 34.504101 samples/sec	 accuracy=79.788723	 loss=0.801284	 lr=0.001000
Epoch[078] Batch [2349]/[3759]	Speed: 35.219453 samples/sec	 accuracy=79.818484	 loss=0.800540	 lr=0.001000
Epoch[078] Batch [2399]/[3759]	Speed: 34.654606 samples/sec	 accuracy=79.800130	 loss=0.800679	 lr=0.001000
Epoch[078] Batch [2449]/[3759]	Speed: 35.402170 samples/sec	 accuracy=79.815051	 loss=0.799826	 lr=0.001000
Epoch[078] Batch [2499]/[3759]	Speed: 34.649409 samples/sec	 accuracy=79.795000	 loss=0.800693	 lr=0.001000
Epoch[078] Batch [2549]/[3759]	Speed: 35.134424 samples/sec	 accuracy=79.775735	 loss=0.801379	 lr=0.001000
Epoch[078] Batch [2599]/[3759]	Speed: 34.460786 samples/sec	 accuracy=79.759014	 loss=0.802066	 lr=0.001000
Epoch[078] Batch [2649]/[3759]	Speed: 34.919937 samples/sec	 accuracy=79.762382	 loss=0.802166	 lr=0.001000
Epoch[078] Batch [2699]/[3759]	Speed: 34.733115 samples/sec	 accuracy=79.756944	 loss=0.802767	 lr=0.001000
Epoch[078] Batch [2749]/[3759]	Speed: 34.619025 samples/sec	 accuracy=79.750568	 loss=0.803320	 lr=0.001000
Epoch[078] Batch [2799]/[3759]	Speed: 34.862409 samples/sec	 accuracy=79.773996	 loss=0.802581	 lr=0.001000
Epoch[078] Batch [2849]/[3759]	Speed: 34.361159 samples/sec	 accuracy=79.750000	 loss=0.803320	 lr=0.001000
Epoch[078] Batch [2899]/[3759]	Speed: 34.976856 samples/sec	 accuracy=79.728448	 loss=0.804544	 lr=0.001000
Epoch[078] Batch [2949]/[3759]	Speed: 34.178219 samples/sec	 accuracy=79.725636	 loss=0.804526	 lr=0.001000
Epoch[078] Batch [2999]/[3759]	Speed: 35.227195 samples/sec	 accuracy=79.721354	 loss=0.804753	 lr=0.001000
Epoch[078] Batch [3049]/[3759]	Speed: 34.479960 samples/sec	 accuracy=79.724385	 loss=0.804565	 lr=0.001000
Epoch[078] Batch [3099]/[3759]	Speed: 34.849283 samples/sec	 accuracy=79.724294	 loss=0.804480	 lr=0.001000
Epoch[078] Batch [3149]/[3759]	Speed: 34.656835 samples/sec	 accuracy=79.710813	 loss=0.804909	 lr=0.001000
Epoch[078] Batch [3199]/[3759]	Speed: 34.420296 samples/sec	 accuracy=79.721191	 loss=0.804732	 lr=0.001000
Epoch[078] Batch [3249]/[3759]	Speed: 35.031586 samples/sec	 accuracy=79.708654	 loss=0.804975	 lr=0.001000
Epoch[078] Batch [3299]/[3759]	Speed: 34.342083 samples/sec	 accuracy=79.686553	 loss=0.805433	 lr=0.001000
Epoch[078] Batch [3349]/[3759]	Speed: 35.103151 samples/sec	 accuracy=79.687966	 loss=0.805217	 lr=0.001000
Epoch[078] Batch [3399]/[3759]	Speed: 34.465828 samples/sec	 accuracy=79.689338	 loss=0.805454	 lr=0.001000
Epoch[078] Batch [3449]/[3759]	Speed: 35.015777 samples/sec	 accuracy=79.681159	 loss=0.806089	 lr=0.001000
Epoch[078] Batch [3499]/[3759]	Speed: 34.505220 samples/sec	 accuracy=79.670089	 loss=0.806728	 lr=0.001000
Epoch[078] Batch [3549]/[3759]	Speed: 34.968860 samples/sec	 accuracy=79.657570	 loss=0.807084	 lr=0.001000
Epoch[078] Batch [3599]/[3759]	Speed: 34.839116 samples/sec	 accuracy=79.651910	 loss=0.807499	 lr=0.001000
Epoch[078] Batch [3649]/[3759]	Speed: 34.529720 samples/sec	 accuracy=79.656250	 loss=0.807130	 lr=0.001000
Epoch[078] Batch [3699]/[3759]	Speed: 34.931691 samples/sec	 accuracy=79.660895	 loss=0.807010	 lr=0.001000
Epoch[078] Batch [3749]/[3759]	Speed: 35.996742 samples/sec	 accuracy=79.665000	 loss=0.807079	 lr=0.001000
Batch [0049]/[0303]: acc-top1=66.562500 acc-top5=85.718750
Batch [0099]/[0303]: acc-top1=66.437500 acc-top5=85.093750
Batch [0149]/[0303]: acc-top1=66.781250 acc-top5=85.364583
Batch [0199]/[0303]: acc-top1=66.375000 acc-top5=85.398438
Batch [0249]/[0303]: acc-top1=66.118750 acc-top5=85.200000
Batch [0299]/[0303]: acc-top1=65.968750 acc-top5=85.166667
[Epoch 078] training: accuracy=79.665885	 loss=0.806965
[Epoch 078] speed: 34 samples/sec	time cost: 7365.474059
[Epoch 078] validation: acc-top1=65.960190 acc-top5=85.169142 loss=1.648161
Epoch[079] Batch [0049]/[3760]	Speed: 25.008644 samples/sec	 accuracy=79.593750	 loss=0.793599	 lr=0.001000
Epoch[079] Batch [0099]/[3760]	Speed: 34.989347 samples/sec	 accuracy=80.015625	 loss=0.793314	 lr=0.001000
Epoch[079] Batch [0149]/[3760]	Speed: 35.426076 samples/sec	 accuracy=80.145833	 loss=0.791441	 lr=0.001000
Epoch[079] Batch [0199]/[3760]	Speed: 35.075541 samples/sec	 accuracy=80.171875	 loss=0.794800	 lr=0.001000
Epoch[079] Batch [0249]/[3760]	Speed: 35.237268 samples/sec	 accuracy=80.156250	 loss=0.794896	 lr=0.001000
Epoch[079] Batch [0299]/[3760]	Speed: 35.076617 samples/sec	 accuracy=80.166667	 loss=0.791646	 lr=0.001000
Epoch[079] Batch [0349]/[3760]	Speed: 35.270365 samples/sec	 accuracy=80.258929	 loss=0.791218	 lr=0.001000
Epoch[079] Batch [0399]/[3760]	Speed: 35.053231 samples/sec	 accuracy=80.140625	 loss=0.793616	 lr=0.001000
Epoch[079] Batch [0449]/[3760]	Speed: 35.213039 samples/sec	 accuracy=80.104167	 loss=0.793395	 lr=0.001000
Epoch[079] Batch [0499]/[3760]	Speed: 35.412147 samples/sec	 accuracy=80.028125	 loss=0.795694	 lr=0.001000
Epoch[079] Batch [0549]/[3760]	Speed: 35.044101 samples/sec	 accuracy=80.042614	 loss=0.795366	 lr=0.001000
Epoch[079] Batch [0599]/[3760]	Speed: 35.239950 samples/sec	 accuracy=80.054688	 loss=0.794632	 lr=0.001000
Epoch[079] Batch [0649]/[3760]	Speed: 35.181582 samples/sec	 accuracy=80.151442	 loss=0.793685	 lr=0.001000
Epoch[079] Batch [0699]/[3760]	Speed: 35.411926 samples/sec	 accuracy=80.133929	 loss=0.794442	 lr=0.001000
Epoch[079] Batch [0749]/[3760]	Speed: 35.163940 samples/sec	 accuracy=80.152083	 loss=0.793960	 lr=0.001000
Epoch[079] Batch [0799]/[3760]	Speed: 35.232062 samples/sec	 accuracy=80.162109	 loss=0.793181	 lr=0.001000
Epoch[079] Batch [0849]/[3760]	Speed: 35.202750 samples/sec	 accuracy=80.145221	 loss=0.792101	 lr=0.001000
Epoch[079] Batch [0899]/[3760]	Speed: 35.166561 samples/sec	 accuracy=80.154514	 loss=0.790383	 lr=0.001000
Epoch[079] Batch [0949]/[3760]	Speed: 35.356018 samples/sec	 accuracy=80.136513	 loss=0.788411	 lr=0.001000
Epoch[079] Batch [0999]/[3760]	Speed: 35.384887 samples/sec	 accuracy=80.140625	 loss=0.787908	 lr=0.001000
Epoch[079] Batch [1049]/[3760]	Speed: 35.235427 samples/sec	 accuracy=80.130952	 loss=0.787342	 lr=0.001000
Epoch[079] Batch [1099]/[3760]	Speed: 35.305321 samples/sec	 accuracy=80.143466	 loss=0.786938	 lr=0.001000
Epoch[079] Batch [1149]/[3760]	Speed: 35.456244 samples/sec	 accuracy=80.110054	 loss=0.789678	 lr=0.001000
Epoch[079] Batch [1199]/[3760]	Speed: 35.315180 samples/sec	 accuracy=80.071615	 loss=0.790265	 lr=0.001000
Epoch[079] Batch [1249]/[3760]	Speed: 35.161881 samples/sec	 accuracy=80.136250	 loss=0.788140	 lr=0.001000
Epoch[079] Batch [1299]/[3760]	Speed: 35.491760 samples/sec	 accuracy=80.110577	 loss=0.789106	 lr=0.001000
Epoch[079] Batch [1349]/[3760]	Speed: 35.286804 samples/sec	 accuracy=80.099537	 loss=0.789984	 lr=0.001000
Epoch[079] Batch [1399]/[3760]	Speed: 35.134284 samples/sec	 accuracy=80.084821	 loss=0.790910	 lr=0.001000
Epoch[079] Batch [1449]/[3760]	Speed: 35.225500 samples/sec	 accuracy=80.064655	 loss=0.791939	 lr=0.001000
Epoch[079] Batch [1499]/[3760]	Speed: 35.397657 samples/sec	 accuracy=80.035417	 loss=0.791983	 lr=0.001000
Epoch[079] Batch [1549]/[3760]	Speed: 35.398370 samples/sec	 accuracy=80.047379	 loss=0.791489	 lr=0.001000
Epoch[079] Batch [1599]/[3760]	Speed: 35.420505 samples/sec	 accuracy=80.031250	 loss=0.792464	 lr=0.001000
Epoch[079] Batch [1649]/[3760]	Speed: 35.365343 samples/sec	 accuracy=80.020833	 loss=0.792195	 lr=0.001000
Epoch[079] Batch [1699]/[3760]	Speed: 35.105286 samples/sec	 accuracy=79.987132	 loss=0.792848	 lr=0.001000
Epoch[079] Batch [1749]/[3760]	Speed: 35.222627 samples/sec	 accuracy=79.985714	 loss=0.792990	 lr=0.001000
Epoch[079] Batch [1799]/[3760]	Speed: 35.134630 samples/sec	 accuracy=79.993924	 loss=0.792268	 lr=0.001000
Epoch[079] Batch [1849]/[3760]	Speed: 35.180022 samples/sec	 accuracy=79.985642	 loss=0.792796	 lr=0.001000
Epoch[079] Batch [1899]/[3760]	Speed: 35.204980 samples/sec	 accuracy=79.958059	 loss=0.794392	 lr=0.001000
Epoch[079] Batch [1949]/[3760]	Speed: 35.445111 samples/sec	 accuracy=79.938301	 loss=0.795291	 lr=0.001000
Epoch[079] Batch [1999]/[3760]	Speed: 35.279034 samples/sec	 accuracy=79.918750	 loss=0.795978	 lr=0.001000
Epoch[079] Batch [2049]/[3760]	Speed: 35.355358 samples/sec	 accuracy=79.900915	 loss=0.796774	 lr=0.001000
Epoch[079] Batch [2099]/[3760]	Speed: 35.322247 samples/sec	 accuracy=79.864583	 loss=0.797439	 lr=0.001000
Epoch[079] Batch [2149]/[3760]	Speed: 35.278157 samples/sec	 accuracy=79.857558	 loss=0.798414	 lr=0.001000
Epoch[079] Batch [2199]/[3760]	Speed: 35.231119 samples/sec	 accuracy=79.882812	 loss=0.798006	 lr=0.001000
Epoch[079] Batch [2249]/[3760]	Speed: 35.297674 samples/sec	 accuracy=79.889583	 loss=0.798198	 lr=0.001000
Epoch[079] Batch [2299]/[3760]	Speed: 35.254335 samples/sec	 accuracy=79.904212	 loss=0.797829	 lr=0.001000
Epoch[079] Batch [2349]/[3760]	Speed: 35.390347 samples/sec	 accuracy=79.912234	 loss=0.798128	 lr=0.001000
Epoch[079] Batch [2399]/[3760]	Speed: 35.158095 samples/sec	 accuracy=79.902344	 loss=0.798067	 lr=0.001000
Epoch[079] Batch [2449]/[3760]	Speed: 35.471359 samples/sec	 accuracy=79.876276	 loss=0.799147	 lr=0.001000
Epoch[079] Batch [2499]/[3760]	Speed: 35.302131 samples/sec	 accuracy=79.870625	 loss=0.799019	 lr=0.001000
Epoch[079] Batch [2549]/[3760]	Speed: 35.205367 samples/sec	 accuracy=79.872549	 loss=0.798954	 lr=0.001000
Epoch[079] Batch [2599]/[3760]	Speed: 35.224048 samples/sec	 accuracy=79.846154	 loss=0.799656	 lr=0.001000
Epoch[079] Batch [2649]/[3760]	Speed: 35.286296 samples/sec	 accuracy=79.834906	 loss=0.800156	 lr=0.001000
Epoch[079] Batch [2699]/[3760]	Speed: 35.243496 samples/sec	 accuracy=79.832755	 loss=0.799592	 lr=0.001000
Epoch[079] Batch [2749]/[3760]	Speed: 35.240592 samples/sec	 accuracy=79.815909	 loss=0.800664	 lr=0.001000
Epoch[079] Batch [2799]/[3760]	Speed: 35.220810 samples/sec	 accuracy=79.814174	 loss=0.800595	 lr=0.001000
Epoch[079] Batch [2849]/[3760]	Speed: 34.966650 samples/sec	 accuracy=79.803728	 loss=0.801013	 lr=0.001000
Epoch[079] Batch [2899]/[3760]	Speed: 35.409212 samples/sec	 accuracy=79.799569	 loss=0.800944	 lr=0.001000
Epoch[079] Batch [2949]/[3760]	Speed: 35.109702 samples/sec	 accuracy=79.785487	 loss=0.801124	 lr=0.001000
Epoch[079] Batch [2999]/[3760]	Speed: 35.324411 samples/sec	 accuracy=79.779167	 loss=0.801673	 lr=0.001000
Epoch[079] Batch [3049]/[3760]	Speed: 35.298229 samples/sec	 accuracy=79.788422	 loss=0.801662	 lr=0.001000
Epoch[079] Batch [3099]/[3760]	Speed: 35.282161 samples/sec	 accuracy=79.827117	 loss=0.800050	 lr=0.001000
Epoch[079] Batch [3149]/[3760]	Speed: 35.218421 samples/sec	 accuracy=79.823909	 loss=0.800097	 lr=0.001000
Epoch[079] Batch [3199]/[3760]	Speed: 35.470413 samples/sec	 accuracy=79.826172	 loss=0.799301	 lr=0.001000
Epoch[079] Batch [3249]/[3760]	Speed: 35.046575 samples/sec	 accuracy=79.813942	 loss=0.799607	 lr=0.001000
Epoch[079] Batch [3299]/[3760]	Speed: 35.159687 samples/sec	 accuracy=79.819129	 loss=0.799347	 lr=0.001000
Epoch[079] Batch [3349]/[3760]	Speed: 35.316156 samples/sec	 accuracy=79.808769	 loss=0.799758	 lr=0.001000
Epoch[079] Batch [3399]/[3760]	Speed: 35.118011 samples/sec	 accuracy=79.806985	 loss=0.799668	 lr=0.001000
Epoch[079] Batch [3449]/[3760]	Speed: 35.173966 samples/sec	 accuracy=79.800725	 loss=0.800146	 lr=0.001000
Epoch[079] Batch [3499]/[3760]	Speed: 34.978819 samples/sec	 accuracy=79.804911	 loss=0.800410	 lr=0.001000
Epoch[079] Batch [3549]/[3760]	Speed: 35.558968 samples/sec	 accuracy=79.812500	 loss=0.800161	 lr=0.001000
Epoch[079] Batch [3599]/[3760]	Speed: 35.099717 samples/sec	 accuracy=79.802083	 loss=0.800355	 lr=0.001000
Epoch[079] Batch [3649]/[3760]	Speed: 35.235728 samples/sec	 accuracy=79.815068	 loss=0.800371	 lr=0.001000
Epoch[079] Batch [3699]/[3760]	Speed: 35.126321 samples/sec	 accuracy=79.794764	 loss=0.801591	 lr=0.001000
Epoch[079] Batch [3749]/[3760]	Speed: 36.585559 samples/sec	 accuracy=79.787917	 loss=0.801770	 lr=0.000100
Batch [0049]/[0303]: acc-top1=66.000000 acc-top5=86.218750
Batch [0099]/[0303]: acc-top1=65.703125 acc-top5=85.343750
Batch [0149]/[0303]: acc-top1=66.020833 acc-top5=85.489583
Batch [0199]/[0303]: acc-top1=65.804688 acc-top5=85.421875
Batch [0249]/[0303]: acc-top1=65.731250 acc-top5=85.306250
Batch [0299]/[0303]: acc-top1=65.588542 acc-top5=85.244792
[Epoch 079] training: accuracy=79.783078	 loss=0.802040
[Epoch 079] speed: 35 samples/sec	time cost: 7314.532427
[Epoch 079] validation: acc-top1=65.583746 acc-top5=85.267120 loss=1.656846
Epoch[080] Batch [0049]/[3760]	Speed: 25.309059 samples/sec	 accuracy=80.687500	 loss=0.767323	 lr=0.000100
Epoch[080] Batch [0099]/[3760]	Speed: 35.433277 samples/sec	 accuracy=80.515625	 loss=0.774744	 lr=0.000100
Epoch[080] Batch [0149]/[3760]	Speed: 35.230497 samples/sec	 accuracy=80.541667	 loss=0.775688	 lr=0.000100
Epoch[080] Batch [0199]/[3760]	Speed: 35.401938 samples/sec	 accuracy=80.609375	 loss=0.769016	 lr=0.000100
Epoch[080] Batch [0249]/[3760]	Speed: 35.402443 samples/sec	 accuracy=80.731250	 loss=0.764253	 lr=0.000100
Epoch[080] Batch [0299]/[3760]	Speed: 35.249603 samples/sec	 accuracy=80.666667	 loss=0.765266	 lr=0.000100
Epoch[080] Batch [0349]/[3760]	Speed: 35.396761 samples/sec	 accuracy=80.700893	 loss=0.764700	 lr=0.000100
Epoch[080] Batch [0399]/[3760]	Speed: 35.084235 samples/sec	 accuracy=80.843750	 loss=0.762057	 lr=0.000100
Epoch[080] Batch [0449]/[3760]	Speed: 35.199302 samples/sec	 accuracy=80.913194	 loss=0.759646	 lr=0.000100
Epoch[080] Batch [0499]/[3760]	Speed: 35.267436 samples/sec	 accuracy=80.834375	 loss=0.760388	 lr=0.000100
Epoch[080] Batch [0549]/[3760]	Speed: 35.513432 samples/sec	 accuracy=80.855114	 loss=0.758510	 lr=0.000100
Epoch[080] Batch [0599]/[3760]	Speed: 35.349783 samples/sec	 accuracy=80.908854	 loss=0.755378	 lr=0.000100
Epoch[080] Batch [0649]/[3760]	Speed: 35.350312 samples/sec	 accuracy=80.848558	 loss=0.757506	 lr=0.000100
Epoch[080] Batch [0699]/[3760]	Speed: 35.230984 samples/sec	 accuracy=80.899554	 loss=0.755707	 lr=0.000100
Epoch[080] Batch [0749]/[3760]	Speed: 35.427141 samples/sec	 accuracy=80.970833	 loss=0.754334	 lr=0.000100
Epoch[080] Batch [0799]/[3760]	Speed: 35.069612 samples/sec	 accuracy=80.947266	 loss=0.755678	 lr=0.000100
Epoch[080] Batch [0849]/[3760]	Speed: 35.380553 samples/sec	 accuracy=80.983456	 loss=0.754325	 lr=0.000100
Epoch[080] Batch [0899]/[3760]	Speed: 35.217160 samples/sec	 accuracy=81.001736	 loss=0.753675	 lr=0.000100
Epoch[080] Batch [0949]/[3760]	Speed: 35.134136 samples/sec	 accuracy=80.945724	 loss=0.756669	 lr=0.000100
Epoch[080] Batch [0999]/[3760]	Speed: 35.308963 samples/sec	 accuracy=80.962500	 loss=0.757131	 lr=0.000100
Epoch[080] Batch [1049]/[3760]	Speed: 35.346340 samples/sec	 accuracy=80.958333	 loss=0.758566	 lr=0.000100
Epoch[080] Batch [1099]/[3760]	Speed: 35.295405 samples/sec	 accuracy=80.955966	 loss=0.758472	 lr=0.000100
Epoch[080] Batch [1149]/[3760]	Speed: 35.197290 samples/sec	 accuracy=80.932065	 loss=0.759593	 lr=0.000100
Epoch[080] Batch [1199]/[3760]	Speed: 35.501492 samples/sec	 accuracy=80.940104	 loss=0.759389	 lr=0.000100
Epoch[080] Batch [1249]/[3760]	Speed: 35.072481 samples/sec	 accuracy=80.936250	 loss=0.760672	 lr=0.000100
Epoch[080] Batch [1299]/[3760]	Speed: 35.405594 samples/sec	 accuracy=80.953125	 loss=0.759487	 lr=0.000100
Epoch[080] Batch [1349]/[3760]	Speed: 35.034584 samples/sec	 accuracy=80.959491	 loss=0.758869	 lr=0.000100
Epoch[080] Batch [1399]/[3760]	Speed: 35.388240 samples/sec	 accuracy=80.991071	 loss=0.757598	 lr=0.000100
Epoch[080] Batch [1449]/[3760]	Speed: 35.300808 samples/sec	 accuracy=80.976293	 loss=0.758184	 lr=0.000100
Epoch[080] Batch [1499]/[3760]	Speed: 35.466598 samples/sec	 accuracy=80.933333	 loss=0.759551	 lr=0.000100
Epoch[080] Batch [1549]/[3760]	Speed: 35.206861 samples/sec	 accuracy=80.959677	 loss=0.758146	 lr=0.000100
Epoch[080] Batch [1599]/[3760]	Speed: 35.395352 samples/sec	 accuracy=80.963867	 loss=0.757328	 lr=0.000100
Epoch[080] Batch [1649]/[3760]	Speed: 35.117838 samples/sec	 accuracy=80.957386	 loss=0.757631	 lr=0.000100
Epoch[080] Batch [1699]/[3760]	Speed: 35.278423 samples/sec	 accuracy=80.971507	 loss=0.757344	 lr=0.000100
Epoch[080] Batch [1749]/[3760]	Speed: 35.297215 samples/sec	 accuracy=80.963393	 loss=0.757245	 lr=0.000100
Epoch[080] Batch [1799]/[3760]	Speed: 35.403154 samples/sec	 accuracy=80.992188	 loss=0.756845	 lr=0.000100
Epoch[080] Batch [1849]/[3760]	Speed: 35.448429 samples/sec	 accuracy=80.984797	 loss=0.756926	 lr=0.000100
Epoch[080] Batch [1899]/[3760]	Speed: 35.307195 samples/sec	 accuracy=80.986842	 loss=0.756420	 lr=0.000100
Epoch[080] Batch [1949]/[3760]	Speed: 35.235308 samples/sec	 accuracy=80.974359	 loss=0.757834	 lr=0.000100
Epoch[080] Batch [1999]/[3760]	Speed: 35.153533 samples/sec	 accuracy=80.964844	 loss=0.757932	 lr=0.000100
Epoch[080] Batch [2049]/[3760]	Speed: 35.222319 samples/sec	 accuracy=81.003811	 loss=0.756431	 lr=0.000100
Epoch[080] Batch [2099]/[3760]	Speed: 35.062639 samples/sec	 accuracy=81.019345	 loss=0.756087	 lr=0.000100
Epoch[080] Batch [2149]/[3760]	Speed: 35.179968 samples/sec	 accuracy=81.038517	 loss=0.755018	 lr=0.000100
Epoch[080] Batch [2199]/[3760]	Speed: 34.953681 samples/sec	 accuracy=81.027699	 loss=0.755838	 lr=0.000100
Epoch[080] Batch [2249]/[3760]	Speed: 35.195779 samples/sec	 accuracy=81.033333	 loss=0.755478	 lr=0.000100
Epoch[080] Batch [2299]/[3760]	Speed: 34.984214 samples/sec	 accuracy=81.074728	 loss=0.754159	 lr=0.000100
Epoch[080] Batch [2349]/[3760]	Speed: 35.233424 samples/sec	 accuracy=81.059840	 loss=0.754378	 lr=0.000100
Epoch[080] Batch [2399]/[3760]	Speed: 34.801405 samples/sec	 accuracy=81.079427	 loss=0.754111	 lr=0.000100
Epoch[080] Batch [2449]/[3760]	Speed: 35.033127 samples/sec	 accuracy=81.088648	 loss=0.753874	 lr=0.000100
Epoch[080] Batch [2499]/[3760]	Speed: 34.673379 samples/sec	 accuracy=81.093125	 loss=0.753359	 lr=0.000100
Epoch[080] Batch [2549]/[3760]	Speed: 35.049157 samples/sec	 accuracy=81.082721	 loss=0.753442	 lr=0.000100
Epoch[080] Batch [2599]/[3760]	Speed: 34.974368 samples/sec	 accuracy=81.094952	 loss=0.753134	 lr=0.000100
Epoch[080] Batch [2649]/[3760]	Speed: 35.083836 samples/sec	 accuracy=81.113208	 loss=0.752759	 lr=0.000100
Epoch[080] Batch [2699]/[3760]	Speed: 34.991885 samples/sec	 accuracy=81.126736	 loss=0.752476	 lr=0.000100
Epoch[080] Batch [2749]/[3760]	Speed: 34.783065 samples/sec	 accuracy=81.127273	 loss=0.752600	 lr=0.000100
Epoch[080] Batch [2799]/[3760]	Speed: 35.221651 samples/sec	 accuracy=81.108817	 loss=0.753424	 lr=0.000100
Epoch[080] Batch [2849]/[3760]	Speed: 34.351494 samples/sec	 accuracy=81.106360	 loss=0.753548	 lr=0.000100
Epoch[080] Batch [2899]/[3760]	Speed: 35.252578 samples/sec	 accuracy=81.126616	 loss=0.753226	 lr=0.000100
Epoch[080] Batch [2949]/[3760]	Speed: 34.549343 samples/sec	 accuracy=81.130826	 loss=0.752780	 lr=0.000100
Epoch[080] Batch [2999]/[3760]	Speed: 35.143328 samples/sec	 accuracy=81.120312	 loss=0.753730	 lr=0.000100
Epoch[080] Batch [3049]/[3760]	Speed: 34.977077 samples/sec	 accuracy=81.136270	 loss=0.753075	 lr=0.000100
Epoch[080] Batch [3099]/[3760]	Speed: 34.736725 samples/sec	 accuracy=81.143649	 loss=0.753170	 lr=0.000100
Epoch[080] Batch [3149]/[3760]	Speed: 35.258879 samples/sec	 accuracy=81.140377	 loss=0.753099	 lr=0.000100
Epoch[080] Batch [3199]/[3760]	Speed: 34.890628 samples/sec	 accuracy=81.149414	 loss=0.752681	 lr=0.000100
Epoch[080] Batch [3249]/[3760]	Speed: 34.953647 samples/sec	 accuracy=81.152404	 loss=0.752960	 lr=0.000100
Epoch[080] Batch [3299]/[3760]	Speed: 34.599619 samples/sec	 accuracy=81.170455	 loss=0.752211	 lr=0.000100
Epoch[080] Batch [3349]/[3760]	Speed: 35.271888 samples/sec	 accuracy=81.176772	 loss=0.751940	 lr=0.000100
Epoch[080] Batch [3399]/[3760]	Speed: 34.567860 samples/sec	 accuracy=81.182904	 loss=0.751537	 lr=0.000100
Epoch[080] Batch [3449]/[3760]	Speed: 35.203336 samples/sec	 accuracy=81.201087	 loss=0.750719	 lr=0.000100
Epoch[080] Batch [3499]/[3760]	Speed: 34.991364 samples/sec	 accuracy=81.204018	 loss=0.750517	 lr=0.000100
Epoch[080] Batch [3549]/[3760]	Speed: 35.077155 samples/sec	 accuracy=81.208627	 loss=0.750104	 lr=0.000100
Epoch[080] Batch [3599]/[3760]	Speed: 35.129855 samples/sec	 accuracy=81.217014	 loss=0.750015	 lr=0.000100
Epoch[080] Batch [3649]/[3760]	Speed: 35.304137 samples/sec	 accuracy=81.208476	 loss=0.750015	 lr=0.000100
Epoch[080] Batch [3699]/[3760]	Speed: 35.095747 samples/sec	 accuracy=81.223395	 loss=0.749895	 lr=0.000100
Epoch[080] Batch [3749]/[3760]	Speed: 36.282333 samples/sec	 accuracy=81.232500	 loss=0.749538	 lr=0.000100
Batch [0049]/[0303]: acc-top1=66.781250 acc-top5=86.125000
Batch [0099]/[0303]: acc-top1=66.234375 acc-top5=85.453125
Batch [0149]/[0303]: acc-top1=66.541667 acc-top5=85.593750
Batch [0199]/[0303]: acc-top1=66.218750 acc-top5=85.531250
Batch [0249]/[0303]: acc-top1=66.143750 acc-top5=85.406250
Batch [0299]/[0303]: acc-top1=65.932292 acc-top5=85.432292
[Epoch 080] training: accuracy=81.242520	 loss=0.749341
[Epoch 080] speed: 34 samples/sec	time cost: 7331.492391
[Epoch 080] validation: acc-top1=65.939563 acc-top5=85.442450 loss=1.637850
Epoch[081] Batch [0049]/[3759]	Speed: 24.979313 samples/sec	 accuracy=81.593750	 loss=0.728370	 lr=0.000100
Epoch[081] Batch [0099]/[3759]	Speed: 35.259490 samples/sec	 accuracy=81.515625	 loss=0.727507	 lr=0.000100
Epoch[081] Batch [0149]/[3759]	Speed: 35.297177 samples/sec	 accuracy=81.906250	 loss=0.717738	 lr=0.000100
Epoch[081] Batch [0199]/[3759]	Speed: 35.115140 samples/sec	 accuracy=82.132812	 loss=0.719124	 lr=0.000100
Epoch[081] Batch [0249]/[3759]	Speed: 35.354794 samples/sec	 accuracy=81.900000	 loss=0.724356	 lr=0.000100
Epoch[081] Batch [0299]/[3759]	Speed: 35.251618 samples/sec	 accuracy=81.776042	 loss=0.727597	 lr=0.000100
Epoch[081] Batch [0349]/[3759]	Speed: 35.035983 samples/sec	 accuracy=81.714286	 loss=0.729481	 lr=0.000100
Epoch[081] Batch [0399]/[3759]	Speed: 35.141597 samples/sec	 accuracy=81.531250	 loss=0.731990	 lr=0.000100
Epoch[081] Batch [0449]/[3759]	Speed: 35.043897 samples/sec	 accuracy=81.559028	 loss=0.733764	 lr=0.000100
Epoch[081] Batch [0499]/[3759]	Speed: 35.255448 samples/sec	 accuracy=81.565625	 loss=0.732194	 lr=0.000100
Epoch[081] Batch [0549]/[3759]	Speed: 35.066455 samples/sec	 accuracy=81.548295	 loss=0.731824	 lr=0.000100
Epoch[081] Batch [0599]/[3759]	Speed: 35.434592 samples/sec	 accuracy=81.518229	 loss=0.733799	 lr=0.000100
Epoch[081] Batch [0649]/[3759]	Speed: 35.163946 samples/sec	 accuracy=81.463942	 loss=0.734777	 lr=0.000100
Epoch[081] Batch [0699]/[3759]	Speed: 35.249831 samples/sec	 accuracy=81.453125	 loss=0.735390	 lr=0.000100
Epoch[081] Batch [0749]/[3759]	Speed: 35.006345 samples/sec	 accuracy=81.520833	 loss=0.733112	 lr=0.000100
Epoch[081] Batch [0799]/[3759]	Speed: 35.419502 samples/sec	 accuracy=81.503906	 loss=0.733516	 lr=0.000100
Epoch[081] Batch [0849]/[3759]	Speed: 34.919870 samples/sec	 accuracy=81.492647	 loss=0.735113	 lr=0.000100
Epoch[081] Batch [0899]/[3759]	Speed: 35.265186 samples/sec	 accuracy=81.484375	 loss=0.734303	 lr=0.000100
Epoch[081] Batch [0949]/[3759]	Speed: 35.050436 samples/sec	 accuracy=81.506579	 loss=0.734710	 lr=0.000100
Epoch[081] Batch [0999]/[3759]	Speed: 35.341945 samples/sec	 accuracy=81.489063	 loss=0.735483	 lr=0.000100
Epoch[081] Batch [1049]/[3759]	Speed: 35.081743 samples/sec	 accuracy=81.543155	 loss=0.734661	 lr=0.000100
Epoch[081] Batch [1099]/[3759]	Speed: 35.312081 samples/sec	 accuracy=81.535511	 loss=0.735049	 lr=0.000100
Epoch[081] Batch [1149]/[3759]	Speed: 34.865231 samples/sec	 accuracy=81.528533	 loss=0.736055	 lr=0.000100
Epoch[081] Batch [1199]/[3759]	Speed: 35.389418 samples/sec	 accuracy=81.493490	 loss=0.737506	 lr=0.000100
Epoch[081] Batch [1249]/[3759]	Speed: 35.255321 samples/sec	 accuracy=81.520000	 loss=0.736180	 lr=0.000100
Epoch[081] Batch [1299]/[3759]	Speed: 35.119739 samples/sec	 accuracy=81.574519	 loss=0.735145	 lr=0.000100
Epoch[081] Batch [1349]/[3759]	Speed: 35.146476 samples/sec	 accuracy=81.568287	 loss=0.735124	 lr=0.000100
Epoch[081] Batch [1399]/[3759]	Speed: 35.215879 samples/sec	 accuracy=81.573661	 loss=0.735769	 lr=0.000100
Epoch[081] Batch [1449]/[3759]	Speed: 35.201669 samples/sec	 accuracy=81.610991	 loss=0.735108	 lr=0.000100
Epoch[081] Batch [1499]/[3759]	Speed: 35.240743 samples/sec	 accuracy=81.618750	 loss=0.735278	 lr=0.000100
Epoch[081] Batch [1549]/[3759]	Speed: 35.034844 samples/sec	 accuracy=81.655242	 loss=0.734462	 lr=0.000100
Epoch[081] Batch [1599]/[3759]	Speed: 35.172027 samples/sec	 accuracy=81.658203	 loss=0.734129	 lr=0.000100
Epoch[081] Batch [1649]/[3759]	Speed: 35.044661 samples/sec	 accuracy=81.650568	 loss=0.733956	 lr=0.000100
Epoch[081] Batch [1699]/[3759]	Speed: 35.107197 samples/sec	 accuracy=81.642463	 loss=0.734045	 lr=0.000100
Epoch[081] Batch [1749]/[3759]	Speed: 35.088421 samples/sec	 accuracy=81.688393	 loss=0.732757	 lr=0.000100
Epoch[081] Batch [1799]/[3759]	Speed: 35.016227 samples/sec	 accuracy=81.680556	 loss=0.732672	 lr=0.000100
Epoch[081] Batch [1849]/[3759]	Speed: 35.133790 samples/sec	 accuracy=81.656250	 loss=0.733308	 lr=0.000100
Epoch[081] Batch [1899]/[3759]	Speed: 34.749476 samples/sec	 accuracy=81.666118	 loss=0.733194	 lr=0.000100
Epoch[081] Batch [1949]/[3759]	Speed: 35.429619 samples/sec	 accuracy=81.636218	 loss=0.734202	 lr=0.000100
Epoch[081] Batch [1999]/[3759]	Speed: 34.935172 samples/sec	 accuracy=81.639844	 loss=0.734301	 lr=0.000100
Epoch[081] Batch [2049]/[3759]	Speed: 35.187803 samples/sec	 accuracy=81.641768	 loss=0.734070	 lr=0.000100
Epoch[081] Batch [2099]/[3759]	Speed: 35.060769 samples/sec	 accuracy=81.661458	 loss=0.733066	 lr=0.000100
Epoch[081] Batch [2149]/[3759]	Speed: 35.380966 samples/sec	 accuracy=81.651890	 loss=0.733776	 lr=0.000100
Epoch[081] Batch [2199]/[3759]	Speed: 35.084467 samples/sec	 accuracy=81.645597	 loss=0.733551	 lr=0.000100
Epoch[081] Batch [2249]/[3759]	Speed: 35.038714 samples/sec	 accuracy=81.647222	 loss=0.733020	 lr=0.000100
Epoch[081] Batch [2299]/[3759]	Speed: 35.064651 samples/sec	 accuracy=81.648777	 loss=0.733589	 lr=0.000100
Epoch[081] Batch [2349]/[3759]	Speed: 35.281603 samples/sec	 accuracy=81.648936	 loss=0.732914	 lr=0.000100
Epoch[081] Batch [2399]/[3759]	Speed: 35.028892 samples/sec	 accuracy=81.657552	 loss=0.733157	 lr=0.000100
Epoch[081] Batch [2449]/[3759]	Speed: 35.274106 samples/sec	 accuracy=81.645408	 loss=0.733559	 lr=0.000100
Epoch[081] Batch [2499]/[3759]	Speed: 35.136789 samples/sec	 accuracy=81.600625	 loss=0.735283	 lr=0.000100
Epoch[081] Batch [2549]/[3759]	Speed: 35.433411 samples/sec	 accuracy=81.602328	 loss=0.735297	 lr=0.000100
Epoch[081] Batch [2599]/[3759]	Speed: 35.261513 samples/sec	 accuracy=81.597356	 loss=0.735508	 lr=0.000100
Epoch[081] Batch [2649]/[3759]	Speed: 35.255762 samples/sec	 accuracy=81.602005	 loss=0.735121	 lr=0.000100
Epoch[081] Batch [2699]/[3759]	Speed: 35.280094 samples/sec	 accuracy=81.626157	 loss=0.734720	 lr=0.000100
Epoch[081] Batch [2749]/[3759]	Speed: 35.293884 samples/sec	 accuracy=81.621591	 loss=0.734466	 lr=0.000100
Epoch[081] Batch [2799]/[3759]	Speed: 35.288553 samples/sec	 accuracy=81.628906	 loss=0.734007	 lr=0.000100
Epoch[081] Batch [2849]/[3759]	Speed: 35.312404 samples/sec	 accuracy=81.620066	 loss=0.734096	 lr=0.000100
Epoch[081] Batch [2899]/[3759]	Speed: 35.302722 samples/sec	 accuracy=81.623922	 loss=0.733974	 lr=0.000100
Epoch[081] Batch [2949]/[3759]	Speed: 35.249447 samples/sec	 accuracy=81.617585	 loss=0.733821	 lr=0.000100
Epoch[081] Batch [2999]/[3759]	Speed: 35.255024 samples/sec	 accuracy=81.608854	 loss=0.733976	 lr=0.000100
Epoch[081] Batch [3049]/[3759]	Speed: 35.098198 samples/sec	 accuracy=81.625512	 loss=0.733356	 lr=0.000100
Epoch[081] Batch [3099]/[3759]	Speed: 35.151983 samples/sec	 accuracy=81.605343	 loss=0.733858	 lr=0.000100
Epoch[081] Batch [3149]/[3759]	Speed: 34.947218 samples/sec	 accuracy=81.621032	 loss=0.733474	 lr=0.000100
Epoch[081] Batch [3199]/[3759]	Speed: 35.107543 samples/sec	 accuracy=81.609863	 loss=0.733912	 lr=0.000100
Epoch[081] Batch [3249]/[3759]	Speed: 34.965627 samples/sec	 accuracy=81.614904	 loss=0.733694	 lr=0.000100
Epoch[081] Batch [3299]/[3759]	Speed: 35.330778 samples/sec	 accuracy=81.607481	 loss=0.733976	 lr=0.000100
Epoch[081] Batch [3349]/[3759]	Speed: 34.849446 samples/sec	 accuracy=81.624067	 loss=0.733712	 lr=0.000100
Epoch[081] Batch [3399]/[3759]	Speed: 35.224867 samples/sec	 accuracy=81.615809	 loss=0.734055	 lr=0.000100
Epoch[081] Batch [3449]/[3759]	Speed: 34.769283 samples/sec	 accuracy=81.607790	 loss=0.734552	 lr=0.000100
Epoch[081] Batch [3499]/[3759]	Speed: 35.131472 samples/sec	 accuracy=81.604911	 loss=0.734682	 lr=0.000100
Epoch[081] Batch [3549]/[3759]	Speed: 34.332174 samples/sec	 accuracy=81.611356	 loss=0.734546	 lr=0.000100
Epoch[081] Batch [3599]/[3759]	Speed: 35.132652 samples/sec	 accuracy=81.611545	 loss=0.734307	 lr=0.000100
Epoch[081] Batch [3649]/[3759]	Speed: 34.341517 samples/sec	 accuracy=81.600599	 loss=0.734727	 lr=0.000100
Epoch[081] Batch [3699]/[3759]	Speed: 34.879359 samples/sec	 accuracy=81.599240	 loss=0.734472	 lr=0.000100
Epoch[081] Batch [3749]/[3759]	Speed: 36.035578 samples/sec	 accuracy=81.612500	 loss=0.733750	 lr=0.000100
Batch [0049]/[0303]: acc-top1=66.781250 acc-top5=86.093750
Batch [0099]/[0303]: acc-top1=66.296875 acc-top5=85.406250
Batch [0149]/[0303]: acc-top1=66.833333 acc-top5=85.645833
Batch [0199]/[0303]: acc-top1=66.578125 acc-top5=85.750000
Batch [0249]/[0303]: acc-top1=66.418750 acc-top5=85.668750
Batch [0299]/[0303]: acc-top1=66.234375 acc-top5=85.635417
[Epoch 081] training: accuracy=81.612048	 loss=0.733794
[Epoch 081] speed: 34 samples/sec	time cost: 7334.320215
[Epoch 081] validation: acc-top1=66.243812 acc-top5=85.648721 loss=1.645053
Epoch[082] Batch [0049]/[3760]	Speed: 24.906781 samples/sec	 accuracy=82.343750	 loss=0.721034	 lr=0.000100
Epoch[082] Batch [0099]/[3760]	Speed: 34.975316 samples/sec	 accuracy=82.281250	 loss=0.724227	 lr=0.000100
Epoch[082] Batch [0149]/[3760]	Speed: 35.360860 samples/sec	 accuracy=82.302083	 loss=0.719292	 lr=0.000100
Epoch[082] Batch [0199]/[3760]	Speed: 34.847682 samples/sec	 accuracy=82.156250	 loss=0.721962	 lr=0.000100
Epoch[082] Batch [0249]/[3760]	Speed: 35.246666 samples/sec	 accuracy=81.975000	 loss=0.727511	 lr=0.000100
Epoch[082] Batch [0299]/[3760]	Speed: 34.976110 samples/sec	 accuracy=81.947917	 loss=0.725564	 lr=0.000100
Epoch[082] Batch [0349]/[3760]	Speed: 35.318897 samples/sec	 accuracy=81.888393	 loss=0.726186	 lr=0.000100
Epoch[082] Batch [0399]/[3760]	Speed: 35.011113 samples/sec	 accuracy=81.875000	 loss=0.727170	 lr=0.000100
Epoch[082] Batch [0449]/[3760]	Speed: 35.205000 samples/sec	 accuracy=81.836806	 loss=0.725188	 lr=0.000100
Epoch[082] Batch [0499]/[3760]	Speed: 34.912781 samples/sec	 accuracy=81.856250	 loss=0.724649	 lr=0.000100
Epoch[082] Batch [0549]/[3760]	Speed: 35.148376 samples/sec	 accuracy=82.000000	 loss=0.720055	 lr=0.000100
Epoch[082] Batch [0599]/[3760]	Speed: 34.945876 samples/sec	 accuracy=81.958333	 loss=0.719909	 lr=0.000100
Epoch[082] Batch [0649]/[3760]	Speed: 35.149690 samples/sec	 accuracy=81.959135	 loss=0.719947	 lr=0.000100
Epoch[082] Batch [0699]/[3760]	Speed: 34.898491 samples/sec	 accuracy=81.948661	 loss=0.720241	 lr=0.000100
Epoch[082] Batch [0749]/[3760]	Speed: 35.216969 samples/sec	 accuracy=81.912500	 loss=0.722096	 lr=0.000100
Epoch[082] Batch [0799]/[3760]	Speed: 35.062907 samples/sec	 accuracy=81.927734	 loss=0.721926	 lr=0.000100
Epoch[082] Batch [0849]/[3760]	Speed: 35.002607 samples/sec	 accuracy=81.933824	 loss=0.721641	 lr=0.000100
Epoch[082] Batch [0899]/[3760]	Speed: 35.171716 samples/sec	 accuracy=81.920139	 loss=0.721547	 lr=0.000100
Epoch[082] Batch [0949]/[3760]	Speed: 35.255478 samples/sec	 accuracy=81.904605	 loss=0.721673	 lr=0.000100
Epoch[082] Batch [0999]/[3760]	Speed: 35.228645 samples/sec	 accuracy=81.851562	 loss=0.722318	 lr=0.000100
Epoch[082] Batch [1049]/[3760]	Speed: 35.266900 samples/sec	 accuracy=81.848214	 loss=0.722051	 lr=0.000100
Epoch[082] Batch [1099]/[3760]	Speed: 35.501512 samples/sec	 accuracy=81.870739	 loss=0.722426	 lr=0.000100
Epoch[082] Batch [1149]/[3760]	Speed: 35.268491 samples/sec	 accuracy=81.850543	 loss=0.722801	 lr=0.000100
Epoch[082] Batch [1199]/[3760]	Speed: 35.236349 samples/sec	 accuracy=81.833333	 loss=0.723573	 lr=0.000100
Epoch[082] Batch [1249]/[3760]	Speed: 35.051540 samples/sec	 accuracy=81.870000	 loss=0.722680	 lr=0.000100
Epoch[082] Batch [1299]/[3760]	Speed: 35.292462 samples/sec	 accuracy=81.832933	 loss=0.723538	 lr=0.000100
Epoch[082] Batch [1349]/[3760]	Speed: 35.228787 samples/sec	 accuracy=81.828704	 loss=0.723943	 lr=0.000100
Epoch[082] Batch [1399]/[3760]	Speed: 35.321035 samples/sec	 accuracy=81.821429	 loss=0.724370	 lr=0.000100
Epoch[082] Batch [1449]/[3760]	Speed: 34.934557 samples/sec	 accuracy=81.842672	 loss=0.724015	 lr=0.000100
Epoch[082] Batch [1499]/[3760]	Speed: 35.342004 samples/sec	 accuracy=81.792708	 loss=0.725694	 lr=0.000100
Epoch[082] Batch [1549]/[3760]	Speed: 34.942220 samples/sec	 accuracy=81.838710	 loss=0.724288	 lr=0.000100
Epoch[082] Batch [1599]/[3760]	Speed: 35.240217 samples/sec	 accuracy=81.821289	 loss=0.724845	 lr=0.000100
Epoch[082] Batch [1649]/[3760]	Speed: 35.264896 samples/sec	 accuracy=81.821023	 loss=0.724779	 lr=0.000100
Epoch[082] Batch [1699]/[3760]	Speed: 35.065909 samples/sec	 accuracy=81.847426	 loss=0.724547	 lr=0.000100
Epoch[082] Batch [1749]/[3760]	Speed: 35.122854 samples/sec	 accuracy=81.882143	 loss=0.723887	 lr=0.000100
Epoch[082] Batch [1799]/[3760]	Speed: 35.405814 samples/sec	 accuracy=81.863715	 loss=0.723839	 lr=0.000100
Epoch[082] Batch [1849]/[3760]	Speed: 34.971868 samples/sec	 accuracy=81.885980	 loss=0.723180	 lr=0.000100
Epoch[082] Batch [1899]/[3760]	Speed: 35.298521 samples/sec	 accuracy=81.893092	 loss=0.722704	 lr=0.000100
Epoch[082] Batch [1949]/[3760]	Speed: 35.024917 samples/sec	 accuracy=81.895833	 loss=0.722687	 lr=0.000100
Epoch[082] Batch [1999]/[3760]	Speed: 35.426820 samples/sec	 accuracy=81.878906	 loss=0.722564	 lr=0.000100
Epoch[082] Batch [2049]/[3760]	Speed: 34.911789 samples/sec	 accuracy=81.906250	 loss=0.721846	 lr=0.000100
Epoch[082] Batch [2099]/[3760]	Speed: 35.260212 samples/sec	 accuracy=81.929315	 loss=0.721406	 lr=0.000100
Epoch[082] Batch [2149]/[3760]	Speed: 34.962034 samples/sec	 accuracy=81.911337	 loss=0.721709	 lr=0.000100
Epoch[082] Batch [2199]/[3760]	Speed: 35.231495 samples/sec	 accuracy=81.909801	 loss=0.721356	 lr=0.000100
Epoch[082] Batch [2249]/[3760]	Speed: 35.242745 samples/sec	 accuracy=81.886806	 loss=0.722198	 lr=0.000100
Epoch[082] Batch [2299]/[3760]	Speed: 35.062314 samples/sec	 accuracy=81.908288	 loss=0.721907	 lr=0.000100
Epoch[082] Batch [2349]/[3760]	Speed: 35.234825 samples/sec	 accuracy=81.909574	 loss=0.722539	 lr=0.000100
Epoch[082] Batch [2399]/[3760]	Speed: 35.222293 samples/sec	 accuracy=81.891276	 loss=0.722838	 lr=0.000100
Epoch[082] Batch [2449]/[3760]	Speed: 35.045784 samples/sec	 accuracy=81.851403	 loss=0.724019	 lr=0.000100
Epoch[082] Batch [2499]/[3760]	Speed: 35.389838 samples/sec	 accuracy=81.854375	 loss=0.724473	 lr=0.000100
Epoch[082] Batch [2549]/[3760]	Speed: 35.151550 samples/sec	 accuracy=81.832108	 loss=0.725049	 lr=0.000100
Epoch[082] Batch [2599]/[3760]	Speed: 35.242440 samples/sec	 accuracy=81.820913	 loss=0.725082	 lr=0.000100
Epoch[082] Batch [2649]/[3760]	Speed: 35.361880 samples/sec	 accuracy=81.829599	 loss=0.724906	 lr=0.000100
Epoch[082] Batch [2699]/[3760]	Speed: 35.291539 samples/sec	 accuracy=81.833333	 loss=0.724747	 lr=0.000100
Epoch[082] Batch [2749]/[3760]	Speed: 35.325987 samples/sec	 accuracy=81.834091	 loss=0.724496	 lr=0.000100
Epoch[082] Batch [2799]/[3760]	Speed: 35.308879 samples/sec	 accuracy=81.821429	 loss=0.724746	 lr=0.000100
Epoch[082] Batch [2849]/[3760]	Speed: 35.433873 samples/sec	 accuracy=81.830044	 loss=0.724084	 lr=0.000100
Epoch[082] Batch [2899]/[3760]	Speed: 35.170213 samples/sec	 accuracy=81.831358	 loss=0.724116	 lr=0.000100
Epoch[082] Batch [2949]/[3760]	Speed: 35.182282 samples/sec	 accuracy=81.825742	 loss=0.723857	 lr=0.000100
Epoch[082] Batch [2999]/[3760]	Speed: 35.098407 samples/sec	 accuracy=81.835938	 loss=0.723633	 lr=0.000100
Epoch[082] Batch [3049]/[3760]	Speed: 35.318118 samples/sec	 accuracy=81.855533	 loss=0.723300	 lr=0.000100
Epoch[082] Batch [3099]/[3760]	Speed: 34.996629 samples/sec	 accuracy=81.843750	 loss=0.723752	 lr=0.000100
Epoch[082] Batch [3149]/[3760]	Speed: 35.281129 samples/sec	 accuracy=81.831349	 loss=0.724142	 lr=0.000100
Epoch[082] Batch [3199]/[3760]	Speed: 35.031569 samples/sec	 accuracy=81.830078	 loss=0.724481	 lr=0.000100
Epoch[082] Batch [3249]/[3760]	Speed: 35.384195 samples/sec	 accuracy=81.837981	 loss=0.724270	 lr=0.000100
Epoch[082] Batch [3299]/[3760]	Speed: 34.927193 samples/sec	 accuracy=81.833807	 loss=0.724405	 lr=0.000100
Epoch[082] Batch [3349]/[3760]	Speed: 35.309472 samples/sec	 accuracy=81.829757	 loss=0.724486	 lr=0.000100
Epoch[082] Batch [3399]/[3760]	Speed: 34.762314 samples/sec	 accuracy=81.835938	 loss=0.724375	 lr=0.000100
Epoch[082] Batch [3449]/[3760]	Speed: 35.228260 samples/sec	 accuracy=81.840127	 loss=0.724163	 lr=0.000100
Epoch[082] Batch [3499]/[3760]	Speed: 34.804937 samples/sec	 accuracy=81.853571	 loss=0.723474	 lr=0.000100
Epoch[082] Batch [3549]/[3760]	Speed: 35.209453 samples/sec	 accuracy=81.823504	 loss=0.724385	 lr=0.000100
Epoch[082] Batch [3599]/[3760]	Speed: 34.969898 samples/sec	 accuracy=81.823351	 loss=0.724582	 lr=0.000100
Epoch[082] Batch [3649]/[3760]	Speed: 35.100568 samples/sec	 accuracy=81.823202	 loss=0.724304	 lr=0.000100
Epoch[082] Batch [3699]/[3760]	Speed: 35.207731 samples/sec	 accuracy=81.815456	 loss=0.724670	 lr=0.000100
Epoch[082] Batch [3749]/[3760]	Speed: 36.498786 samples/sec	 accuracy=81.829583	 loss=0.724433	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.437500 acc-top5=86.187500
Batch [0099]/[0303]: acc-top1=66.843750 acc-top5=85.421875
Batch [0149]/[0303]: acc-top1=67.104167 acc-top5=85.635417
Batch [0199]/[0303]: acc-top1=66.750000 acc-top5=85.648438
Batch [0249]/[0303]: acc-top1=66.581250 acc-top5=85.525000
Batch [0299]/[0303]: acc-top1=66.401042 acc-top5=85.552083
[Epoch 082] training: accuracy=81.830535	 loss=0.724347
[Epoch 082] speed: 34 samples/sec	time cost: 7331.762691
[Epoch 082] validation: acc-top1=66.393358 acc-top5=85.571370 loss=1.627080
Epoch[083] Batch [0049]/[3760]	Speed: 25.050743 samples/sec	 accuracy=82.125000	 loss=0.746606	 lr=0.000100
Epoch[083] Batch [0099]/[3760]	Speed: 35.331381 samples/sec	 accuracy=81.937500	 loss=0.736341	 lr=0.000100
Epoch[083] Batch [0149]/[3760]	Speed: 35.149027 samples/sec	 accuracy=81.510417	 loss=0.737876	 lr=0.000100
Epoch[083] Batch [0199]/[3760]	Speed: 35.149073 samples/sec	 accuracy=81.539062	 loss=0.738787	 lr=0.000100
Epoch[083] Batch [0249]/[3760]	Speed: 35.085976 samples/sec	 accuracy=81.418750	 loss=0.739097	 lr=0.000100
Epoch[083] Batch [0299]/[3760]	Speed: 35.194670 samples/sec	 accuracy=81.390625	 loss=0.738346	 lr=0.000100
Epoch[083] Batch [0349]/[3760]	Speed: 35.163107 samples/sec	 accuracy=81.406250	 loss=0.735608	 lr=0.000100
Epoch[083] Batch [0399]/[3760]	Speed: 35.423561 samples/sec	 accuracy=81.457031	 loss=0.732483	 lr=0.000100
Epoch[083] Batch [0449]/[3760]	Speed: 35.132118 samples/sec	 accuracy=81.458333	 loss=0.734155	 lr=0.000100
Epoch[083] Batch [0499]/[3760]	Speed: 35.063448 samples/sec	 accuracy=81.531250	 loss=0.732088	 lr=0.000100
Epoch[083] Batch [0549]/[3760]	Speed: 35.165600 samples/sec	 accuracy=81.582386	 loss=0.732089	 lr=0.000100
Epoch[083] Batch [0599]/[3760]	Speed: 35.149513 samples/sec	 accuracy=81.552083	 loss=0.732743	 lr=0.000100
Epoch[083] Batch [0649]/[3760]	Speed: 35.429904 samples/sec	 accuracy=81.584135	 loss=0.731803	 lr=0.000100
Epoch[083] Batch [0699]/[3760]	Speed: 35.175943 samples/sec	 accuracy=81.622768	 loss=0.731042	 lr=0.000100
Epoch[083] Batch [0749]/[3760]	Speed: 35.347589 samples/sec	 accuracy=81.716667	 loss=0.727182	 lr=0.000100
Epoch[083] Batch [0799]/[3760]	Speed: 35.210288 samples/sec	 accuracy=81.746094	 loss=0.726257	 lr=0.000100
Epoch[083] Batch [0849]/[3760]	Speed: 35.427777 samples/sec	 accuracy=81.720588	 loss=0.727278	 lr=0.000100
Epoch[083] Batch [0899]/[3760]	Speed: 35.226022 samples/sec	 accuracy=81.722222	 loss=0.727766	 lr=0.000100
Epoch[083] Batch [0949]/[3760]	Speed: 35.474385 samples/sec	 accuracy=81.771382	 loss=0.726767	 lr=0.000100
Epoch[083] Batch [0999]/[3760]	Speed: 35.048155 samples/sec	 accuracy=81.781250	 loss=0.726464	 lr=0.000100
Epoch[083] Batch [1049]/[3760]	Speed: 35.313635 samples/sec	 accuracy=81.751488	 loss=0.726435	 lr=0.000100
Epoch[083] Batch [1099]/[3760]	Speed: 34.939224 samples/sec	 accuracy=81.826705	 loss=0.724281	 lr=0.000100
Epoch[083] Batch [1149]/[3760]	Speed: 35.229923 samples/sec	 accuracy=81.777174	 loss=0.726188	 lr=0.000100
Epoch[083] Batch [1199]/[3760]	Speed: 34.983151 samples/sec	 accuracy=81.812500	 loss=0.726060	 lr=0.000100
Epoch[083] Batch [1249]/[3760]	Speed: 35.371045 samples/sec	 accuracy=81.827500	 loss=0.724427	 lr=0.000100
Epoch[083] Batch [1299]/[3760]	Speed: 35.101299 samples/sec	 accuracy=81.828125	 loss=0.724619	 lr=0.000100
Epoch[083] Batch [1349]/[3760]	Speed: 34.845647 samples/sec	 accuracy=81.853009	 loss=0.723025	 lr=0.000100
Epoch[083] Batch [1399]/[3760]	Speed: 35.425827 samples/sec	 accuracy=81.852679	 loss=0.723482	 lr=0.000100
Epoch[083] Batch [1449]/[3760]	Speed: 34.801407 samples/sec	 accuracy=81.844828	 loss=0.722868	 lr=0.000100
Epoch[083] Batch [1499]/[3760]	Speed: 35.253007 samples/sec	 accuracy=81.863542	 loss=0.723039	 lr=0.000100
Epoch[083] Batch [1549]/[3760]	Speed: 35.060612 samples/sec	 accuracy=81.875000	 loss=0.723083	 lr=0.000100
Epoch[083] Batch [1599]/[3760]	Speed: 35.279326 samples/sec	 accuracy=81.871094	 loss=0.722393	 lr=0.000100
Epoch[083] Batch [1649]/[3760]	Speed: 35.036383 samples/sec	 accuracy=81.860795	 loss=0.721276	 lr=0.000100
Epoch[083] Batch [1699]/[3760]	Speed: 35.322218 samples/sec	 accuracy=81.860294	 loss=0.721383	 lr=0.000100
Epoch[083] Batch [1749]/[3760]	Speed: 35.217907 samples/sec	 accuracy=81.873214	 loss=0.721478	 lr=0.000100
Epoch[083] Batch [1799]/[3760]	Speed: 35.192864 samples/sec	 accuracy=81.880208	 loss=0.721477	 lr=0.000100
Epoch[083] Batch [1849]/[3760]	Speed: 35.277802 samples/sec	 accuracy=81.890203	 loss=0.721268	 lr=0.000100
Epoch[083] Batch [1899]/[3760]	Speed: 35.269208 samples/sec	 accuracy=81.906250	 loss=0.720828	 lr=0.000100
Epoch[083] Batch [1949]/[3760]	Speed: 35.293451 samples/sec	 accuracy=81.912660	 loss=0.721002	 lr=0.000100
Epoch[083] Batch [1999]/[3760]	Speed: 35.270777 samples/sec	 accuracy=81.896875	 loss=0.721330	 lr=0.000100
Epoch[083] Batch [2049]/[3760]	Speed: 35.394996 samples/sec	 accuracy=81.910061	 loss=0.720946	 lr=0.000100
Epoch[083] Batch [2099]/[3760]	Speed: 35.027243 samples/sec	 accuracy=81.901786	 loss=0.721029	 lr=0.000100
Epoch[083] Batch [2149]/[3760]	Speed: 35.319127 samples/sec	 accuracy=81.895349	 loss=0.721496	 lr=0.000100
Epoch[083] Batch [2199]/[3760]	Speed: 34.877035 samples/sec	 accuracy=81.872159	 loss=0.722546	 lr=0.000100
Epoch[083] Batch [2249]/[3760]	Speed: 35.288350 samples/sec	 accuracy=81.865972	 loss=0.723470	 lr=0.000100
Epoch[083] Batch [2299]/[3760]	Speed: 35.045960 samples/sec	 accuracy=81.873641	 loss=0.723943	 lr=0.000100
Epoch[083] Batch [2349]/[3760]	Speed: 35.332854 samples/sec	 accuracy=81.835106	 loss=0.725128	 lr=0.000100
Epoch[083] Batch [2399]/[3760]	Speed: 35.320062 samples/sec	 accuracy=81.833984	 loss=0.724566	 lr=0.000100
Epoch[083] Batch [2449]/[3760]	Speed: 35.436778 samples/sec	 accuracy=81.841837	 loss=0.724375	 lr=0.000100
Epoch[083] Batch [2499]/[3760]	Speed: 35.042729 samples/sec	 accuracy=81.852500	 loss=0.724182	 lr=0.000100
Epoch[083] Batch [2549]/[3760]	Speed: 35.325549 samples/sec	 accuracy=81.858456	 loss=0.724042	 lr=0.000100
Epoch[083] Batch [2599]/[3760]	Speed: 35.085594 samples/sec	 accuracy=81.866587	 loss=0.723381	 lr=0.000100
Epoch[083] Batch [2649]/[3760]	Speed: 35.464793 samples/sec	 accuracy=81.869104	 loss=0.723197	 lr=0.000100
Epoch[083] Batch [2699]/[3760]	Speed: 35.221163 samples/sec	 accuracy=81.883681	 loss=0.723006	 lr=0.000100
Epoch[083] Batch [2749]/[3760]	Speed: 35.296653 samples/sec	 accuracy=81.871023	 loss=0.723599	 lr=0.000100
Epoch[083] Batch [2799]/[3760]	Speed: 35.197480 samples/sec	 accuracy=81.862165	 loss=0.723328	 lr=0.000100
Epoch[083] Batch [2849]/[3760]	Speed: 35.282202 samples/sec	 accuracy=81.868969	 loss=0.723418	 lr=0.000100
Epoch[083] Batch [2899]/[3760]	Speed: 34.916297 samples/sec	 accuracy=81.848599	 loss=0.724035	 lr=0.000100
Epoch[083] Batch [2949]/[3760]	Speed: 35.200464 samples/sec	 accuracy=81.845339	 loss=0.724571	 lr=0.000100
Epoch[083] Batch [2999]/[3760]	Speed: 34.983725 samples/sec	 accuracy=81.843229	 loss=0.724302	 lr=0.000100
Epoch[083] Batch [3049]/[3760]	Speed: 35.136661 samples/sec	 accuracy=81.841701	 loss=0.724280	 lr=0.000100
Epoch[083] Batch [3099]/[3760]	Speed: 35.098678 samples/sec	 accuracy=81.845262	 loss=0.724161	 lr=0.000100
Epoch[083] Batch [3149]/[3760]	Speed: 34.903443 samples/sec	 accuracy=81.852183	 loss=0.723927	 lr=0.000100
Epoch[083] Batch [3199]/[3760]	Speed: 35.176206 samples/sec	 accuracy=81.871582	 loss=0.723309	 lr=0.000100
Epoch[083] Batch [3249]/[3760]	Speed: 34.688421 samples/sec	 accuracy=81.873077	 loss=0.723340	 lr=0.000100
Epoch[083] Batch [3299]/[3760]	Speed: 35.383727 samples/sec	 accuracy=81.877367	 loss=0.723236	 lr=0.000100
Epoch[083] Batch [3349]/[3760]	Speed: 34.896636 samples/sec	 accuracy=81.895056	 loss=0.722989	 lr=0.000100
Epoch[083] Batch [3399]/[3760]	Speed: 35.415493 samples/sec	 accuracy=81.893842	 loss=0.723083	 lr=0.000100
Epoch[083] Batch [3449]/[3760]	Speed: 35.013681 samples/sec	 accuracy=81.892663	 loss=0.723332	 lr=0.000100
Epoch[083] Batch [3499]/[3760]	Speed: 35.229188 samples/sec	 accuracy=81.912946	 loss=0.722648	 lr=0.000100
Epoch[083] Batch [3549]/[3760]	Speed: 34.748781 samples/sec	 accuracy=81.922535	 loss=0.722239	 lr=0.000100
Epoch[083] Batch [3599]/[3760]	Speed: 35.037432 samples/sec	 accuracy=81.903646	 loss=0.722913	 lr=0.000100
Epoch[083] Batch [3649]/[3760]	Speed: 35.088406 samples/sec	 accuracy=81.886558	 loss=0.723117	 lr=0.000100
Epoch[083] Batch [3699]/[3760]	Speed: 34.908098 samples/sec	 accuracy=81.874578	 loss=0.723568	 lr=0.000100
Epoch[083] Batch [3749]/[3760]	Speed: 35.871598 samples/sec	 accuracy=81.877083	 loss=0.723399	 lr=0.000100
Batch [0049]/[0303]: acc-top1=66.406250 acc-top5=86.281250
Batch [0099]/[0303]: acc-top1=66.046875 acc-top5=85.578125
Batch [0149]/[0303]: acc-top1=66.760417 acc-top5=85.770833
Batch [0199]/[0303]: acc-top1=66.359375 acc-top5=85.757812
Batch [0249]/[0303]: acc-top1=66.250000 acc-top5=85.668750
Batch [0299]/[0303]: acc-top1=66.166667 acc-top5=85.645833
[Epoch 083] training: accuracy=81.882480	 loss=0.723233
[Epoch 083] speed: 34 samples/sec	time cost: 7327.225539
[Epoch 083] validation: acc-top1=66.181931 acc-top5=85.659035 loss=1.625207
Epoch[084] Batch [0049]/[3759]	Speed: 25.139935 samples/sec	 accuracy=82.562500	 loss=0.692028	 lr=0.000100
Epoch[084] Batch [0099]/[3759]	Speed: 35.089735 samples/sec	 accuracy=82.015625	 loss=0.707792	 lr=0.000100
Epoch[084] Batch [0149]/[3759]	Speed: 35.409740 samples/sec	 accuracy=81.979167	 loss=0.717414	 lr=0.000100
Epoch[084] Batch [0199]/[3759]	Speed: 35.083474 samples/sec	 accuracy=81.546875	 loss=0.731312	 lr=0.000100
Epoch[084] Batch [0249]/[3759]	Speed: 35.330281 samples/sec	 accuracy=81.556250	 loss=0.730061	 lr=0.000100
Epoch[084] Batch [0299]/[3759]	Speed: 34.995742 samples/sec	 accuracy=81.661458	 loss=0.727246	 lr=0.000100
Epoch[084] Batch [0349]/[3759]	Speed: 35.248184 samples/sec	 accuracy=81.607143	 loss=0.728982	 lr=0.000100
Epoch[084] Batch [0399]/[3759]	Speed: 34.985781 samples/sec	 accuracy=81.687500	 loss=0.727680	 lr=0.000100
Epoch[084] Batch [0449]/[3759]	Speed: 35.436755 samples/sec	 accuracy=81.690972	 loss=0.725810	 lr=0.000100
Epoch[084] Batch [0499]/[3759]	Speed: 35.141854 samples/sec	 accuracy=81.737500	 loss=0.723007	 lr=0.000100
Epoch[084] Batch [0549]/[3759]	Speed: 35.250044 samples/sec	 accuracy=81.730114	 loss=0.726150	 lr=0.000100
Epoch[084] Batch [0599]/[3759]	Speed: 35.108609 samples/sec	 accuracy=81.765625	 loss=0.723845	 lr=0.000100
Epoch[084] Batch [0649]/[3759]	Speed: 35.401909 samples/sec	 accuracy=81.860577	 loss=0.721262	 lr=0.000100
Epoch[084] Batch [0699]/[3759]	Speed: 35.055767 samples/sec	 accuracy=81.966518	 loss=0.717888	 lr=0.000100
Epoch[084] Batch [0749]/[3759]	Speed: 35.329013 samples/sec	 accuracy=81.937500	 loss=0.718095	 lr=0.000100
Epoch[084] Batch [0799]/[3759]	Speed: 34.897519 samples/sec	 accuracy=81.937500	 loss=0.718026	 lr=0.000100
Epoch[084] Batch [0849]/[3759]	Speed: 35.294982 samples/sec	 accuracy=81.983456	 loss=0.718127	 lr=0.000100
Epoch[084] Batch [0899]/[3759]	Speed: 35.129978 samples/sec	 accuracy=82.005208	 loss=0.717992	 lr=0.000100
Epoch[084] Batch [0949]/[3759]	Speed: 34.920679 samples/sec	 accuracy=81.981908	 loss=0.718323	 lr=0.000100
Epoch[084] Batch [0999]/[3759]	Speed: 35.161362 samples/sec	 accuracy=82.040625	 loss=0.714998	 lr=0.000100
Epoch[084] Batch [1049]/[3759]	Speed: 35.002608 samples/sec	 accuracy=82.050595	 loss=0.715556	 lr=0.000100
Epoch[084] Batch [1099]/[3759]	Speed: 35.184266 samples/sec	 accuracy=82.029830	 loss=0.715396	 lr=0.000100
Epoch[084] Batch [1149]/[3759]	Speed: 34.931963 samples/sec	 accuracy=82.024457	 loss=0.715563	 lr=0.000100
Epoch[084] Batch [1199]/[3759]	Speed: 35.254388 samples/sec	 accuracy=82.040365	 loss=0.715538	 lr=0.000100
Epoch[084] Batch [1249]/[3759]	Speed: 34.767626 samples/sec	 accuracy=82.027500	 loss=0.716171	 lr=0.000100
Epoch[084] Batch [1299]/[3759]	Speed: 35.303077 samples/sec	 accuracy=82.001202	 loss=0.717791	 lr=0.000100
Epoch[084] Batch [1349]/[3759]	Speed: 34.896779 samples/sec	 accuracy=81.986111	 loss=0.718322	 lr=0.000100
Epoch[084] Batch [1399]/[3759]	Speed: 35.194300 samples/sec	 accuracy=81.962054	 loss=0.719172	 lr=0.000100
Epoch[084] Batch [1449]/[3759]	Speed: 34.681225 samples/sec	 accuracy=81.928879	 loss=0.720284	 lr=0.000100
Epoch[084] Batch [1499]/[3759]	Speed: 35.392651 samples/sec	 accuracy=81.942708	 loss=0.720834	 lr=0.000100
Epoch[084] Batch [1549]/[3759]	Speed: 34.901459 samples/sec	 accuracy=81.939516	 loss=0.721025	 lr=0.000100
Epoch[084] Batch [1599]/[3759]	Speed: 35.105527 samples/sec	 accuracy=81.945312	 loss=0.720031	 lr=0.000100
Epoch[084] Batch [1649]/[3759]	Speed: 35.358030 samples/sec	 accuracy=81.932765	 loss=0.720465	 lr=0.000100
Epoch[084] Batch [1699]/[3759]	Speed: 34.967031 samples/sec	 accuracy=81.944853	 loss=0.720540	 lr=0.000100
Epoch[084] Batch [1749]/[3759]	Speed: 35.219163 samples/sec	 accuracy=81.955357	 loss=0.720644	 lr=0.000100
Epoch[084] Batch [1799]/[3759]	Speed: 35.215920 samples/sec	 accuracy=81.947049	 loss=0.720707	 lr=0.000100
Epoch[084] Batch [1849]/[3759]	Speed: 35.336292 samples/sec	 accuracy=81.940878	 loss=0.720838	 lr=0.000100
Epoch[084] Batch [1899]/[3759]	Speed: 35.150469 samples/sec	 accuracy=81.966283	 loss=0.720290	 lr=0.000100
Epoch[084] Batch [1949]/[3759]	Speed: 35.245679 samples/sec	 accuracy=81.995192	 loss=0.719075	 lr=0.000100
Epoch[084] Batch [1999]/[3759]	Speed: 35.193414 samples/sec	 accuracy=82.003125	 loss=0.718253	 lr=0.000100
Epoch[084] Batch [2049]/[3759]	Speed: 35.289041 samples/sec	 accuracy=82.009909	 loss=0.718177	 lr=0.000100
Epoch[084] Batch [2099]/[3759]	Speed: 35.051676 samples/sec	 accuracy=82.013393	 loss=0.718357	 lr=0.000100
Epoch[084] Batch [2149]/[3759]	Speed: 35.403192 samples/sec	 accuracy=82.030523	 loss=0.717627	 lr=0.000100
Epoch[084] Batch [2199]/[3759]	Speed: 35.110297 samples/sec	 accuracy=82.026989	 loss=0.717210	 lr=0.000100
Epoch[084] Batch [2249]/[3759]	Speed: 35.241813 samples/sec	 accuracy=81.991667	 loss=0.718178	 lr=0.000100
Epoch[084] Batch [2299]/[3759]	Speed: 34.861504 samples/sec	 accuracy=81.995245	 loss=0.717683	 lr=0.000100
Epoch[084] Batch [2349]/[3759]	Speed: 35.358536 samples/sec	 accuracy=82.017287	 loss=0.717087	 lr=0.000100
Epoch[084] Batch [2399]/[3759]	Speed: 34.792815 samples/sec	 accuracy=82.034505	 loss=0.716963	 lr=0.000100
Epoch[084] Batch [2449]/[3759]	Speed: 35.239452 samples/sec	 accuracy=82.012755	 loss=0.717637	 lr=0.000100
Epoch[084] Batch [2499]/[3759]	Speed: 34.970555 samples/sec	 accuracy=82.015000	 loss=0.717573	 lr=0.000100
Epoch[084] Batch [2549]/[3759]	Speed: 35.021426 samples/sec	 accuracy=82.029412	 loss=0.716915	 lr=0.000100
Epoch[084] Batch [2599]/[3759]	Speed: 34.702110 samples/sec	 accuracy=82.033654	 loss=0.716873	 lr=0.000100
Epoch[084] Batch [2649]/[3759]	Speed: 35.069554 samples/sec	 accuracy=82.027712	 loss=0.716794	 lr=0.000100
Epoch[084] Batch [2699]/[3759]	Speed: 35.080612 samples/sec	 accuracy=82.010417	 loss=0.717016	 lr=0.000100
Epoch[084] Batch [2749]/[3759]	Speed: 34.854941 samples/sec	 accuracy=82.017045	 loss=0.716930	 lr=0.000100
Epoch[084] Batch [2799]/[3759]	Speed: 34.829741 samples/sec	 accuracy=82.017857	 loss=0.716916	 lr=0.000100
Epoch[084] Batch [2849]/[3759]	Speed: 34.633561 samples/sec	 accuracy=82.006579	 loss=0.716572	 lr=0.000100
Epoch[084] Batch [2899]/[3759]	Speed: 35.021924 samples/sec	 accuracy=82.010776	 loss=0.716307	 lr=0.000100
Epoch[084] Batch [2949]/[3759]	Speed: 34.696998 samples/sec	 accuracy=82.020657	 loss=0.716059	 lr=0.000100
Epoch[084] Batch [2999]/[3759]	Speed: 35.193494 samples/sec	 accuracy=82.043229	 loss=0.715309	 lr=0.000100
Epoch[084] Batch [3049]/[3759]	Speed: 34.673348 samples/sec	 accuracy=82.032787	 loss=0.715281	 lr=0.000100
Epoch[084] Batch [3099]/[3759]	Speed: 35.208064 samples/sec	 accuracy=82.025202	 loss=0.715438	 lr=0.000100
Epoch[084] Batch [3149]/[3759]	Speed: 34.745553 samples/sec	 accuracy=82.012401	 loss=0.715831	 lr=0.000100
Epoch[084] Batch [3199]/[3759]	Speed: 35.211360 samples/sec	 accuracy=82.020508	 loss=0.715781	 lr=0.000100
Epoch[084] Batch [3249]/[3759]	Speed: 35.048843 samples/sec	 accuracy=82.035577	 loss=0.715302	 lr=0.000100
Epoch[084] Batch [3299]/[3759]	Speed: 35.235364 samples/sec	 accuracy=82.044508	 loss=0.715045	 lr=0.000100
Epoch[084] Batch [3349]/[3759]	Speed: 34.911231 samples/sec	 accuracy=82.036847	 loss=0.715130	 lr=0.000100
Epoch[084] Batch [3399]/[3759]	Speed: 34.975677 samples/sec	 accuracy=82.052390	 loss=0.715017	 lr=0.000100
Epoch[084] Batch [3449]/[3759]	Speed: 34.986714 samples/sec	 accuracy=82.058877	 loss=0.714962	 lr=0.000100
Epoch[084] Batch [3499]/[3759]	Speed: 34.938384 samples/sec	 accuracy=82.066964	 loss=0.714718	 lr=0.000100
Epoch[084] Batch [3549]/[3759]	Speed: 35.019985 samples/sec	 accuracy=82.056338	 loss=0.715005	 lr=0.000100
Epoch[084] Batch [3599]/[3759]	Speed: 34.684621 samples/sec	 accuracy=82.060764	 loss=0.714761	 lr=0.000100
Epoch[084] Batch [3649]/[3759]	Speed: 35.222588 samples/sec	 accuracy=82.050942	 loss=0.715369	 lr=0.000100
Epoch[084] Batch [3699]/[3759]	Speed: 34.693083 samples/sec	 accuracy=82.059544	 loss=0.715354	 lr=0.000100
Epoch[084] Batch [3749]/[3759]	Speed: 36.566389 samples/sec	 accuracy=82.065417	 loss=0.715003	 lr=0.000100
Batch [0049]/[0303]: acc-top1=66.843750 acc-top5=86.062500
Batch [0099]/[0303]: acc-top1=66.312500 acc-top5=85.312500
Batch [0149]/[0303]: acc-top1=66.666667 acc-top5=85.458333
Batch [0199]/[0303]: acc-top1=66.398438 acc-top5=85.460938
Batch [0249]/[0303]: acc-top1=66.318750 acc-top5=85.456250
Batch [0299]/[0303]: acc-top1=66.307292 acc-top5=85.442708
[Epoch 084] training: accuracy=82.071362	 loss=0.714903
[Epoch 084] speed: 34 samples/sec	time cost: 7347.003619
[Epoch 084] validation: acc-top1=66.321163 acc-top5=85.447607 loss=1.624357
Epoch[085] Batch [0049]/[3760]	Speed: 25.367274 samples/sec	 accuracy=81.375000	 loss=0.730264	 lr=0.000100
Epoch[085] Batch [0099]/[3760]	Speed: 35.424555 samples/sec	 accuracy=81.750000	 loss=0.723533	 lr=0.000100
Epoch[085] Batch [0149]/[3760]	Speed: 35.261501 samples/sec	 accuracy=81.947917	 loss=0.724109	 lr=0.000100
Epoch[085] Batch [0199]/[3760]	Speed: 35.201560 samples/sec	 accuracy=82.023438	 loss=0.722063	 lr=0.000100
Epoch[085] Batch [0249]/[3760]	Speed: 35.076083 samples/sec	 accuracy=82.118750	 loss=0.718288	 lr=0.000100
Epoch[085] Batch [0299]/[3760]	Speed: 35.168770 samples/sec	 accuracy=82.135417	 loss=0.718744	 lr=0.000100
Epoch[085] Batch [0349]/[3760]	Speed: 35.125203 samples/sec	 accuracy=82.187500	 loss=0.716078	 lr=0.000100
Epoch[085] Batch [0399]/[3760]	Speed: 35.163944 samples/sec	 accuracy=82.210938	 loss=0.713350	 lr=0.000100
Epoch[085] Batch [0449]/[3760]	Speed: 34.846052 samples/sec	 accuracy=82.340278	 loss=0.711869	 lr=0.000100
Epoch[085] Batch [0499]/[3760]	Speed: 35.190831 samples/sec	 accuracy=82.275000	 loss=0.713147	 lr=0.000100
Epoch[085] Batch [0549]/[3760]	Speed: 34.760136 samples/sec	 accuracy=82.363636	 loss=0.707415	 lr=0.000100
Epoch[085] Batch [0599]/[3760]	Speed: 35.333266 samples/sec	 accuracy=82.333333	 loss=0.709029	 lr=0.000100
Epoch[085] Batch [0649]/[3760]	Speed: 34.695130 samples/sec	 accuracy=82.322115	 loss=0.709589	 lr=0.000100
Epoch[085] Batch [0699]/[3760]	Speed: 35.151452 samples/sec	 accuracy=82.352679	 loss=0.710597	 lr=0.000100
Epoch[085] Batch [0749]/[3760]	Speed: 34.639121 samples/sec	 accuracy=82.339583	 loss=0.710974	 lr=0.000100
Epoch[085] Batch [0799]/[3760]	Speed: 35.158290 samples/sec	 accuracy=82.289062	 loss=0.711896	 lr=0.000100
Epoch[085] Batch [0849]/[3760]	Speed: 34.675976 samples/sec	 accuracy=82.330882	 loss=0.711119	 lr=0.000100
Epoch[085] Batch [0899]/[3760]	Speed: 34.746831 samples/sec	 accuracy=82.321181	 loss=0.711495	 lr=0.000100
Epoch[085] Batch [0949]/[3760]	Speed: 35.073854 samples/sec	 accuracy=82.366776	 loss=0.710866	 lr=0.000100
Epoch[085] Batch [0999]/[3760]	Speed: 34.769078 samples/sec	 accuracy=82.376562	 loss=0.710012	 lr=0.000100
Epoch[085] Batch [1049]/[3760]	Speed: 35.349392 samples/sec	 accuracy=82.370536	 loss=0.709559	 lr=0.000100
Epoch[085] Batch [1099]/[3760]	Speed: 34.750375 samples/sec	 accuracy=82.350852	 loss=0.710530	 lr=0.000100
Epoch[085] Batch [1149]/[3760]	Speed: 35.417584 samples/sec	 accuracy=82.357337	 loss=0.710040	 lr=0.000100
Epoch[085] Batch [1199]/[3760]	Speed: 34.995330 samples/sec	 accuracy=82.325521	 loss=0.711182	 lr=0.000100
Epoch[085] Batch [1249]/[3760]	Speed: 35.453447 samples/sec	 accuracy=82.302500	 loss=0.712020	 lr=0.000100
Epoch[085] Batch [1299]/[3760]	Speed: 35.147322 samples/sec	 accuracy=82.320913	 loss=0.711926	 lr=0.000100
Epoch[085] Batch [1349]/[3760]	Speed: 35.383950 samples/sec	 accuracy=82.298611	 loss=0.713227	 lr=0.000100
Epoch[085] Batch [1399]/[3760]	Speed: 35.099792 samples/sec	 accuracy=82.311384	 loss=0.712252	 lr=0.000100
Epoch[085] Batch [1449]/[3760]	Speed: 35.022362 samples/sec	 accuracy=82.294181	 loss=0.711777	 lr=0.000100
Epoch[085] Batch [1499]/[3760]	Speed: 35.427985 samples/sec	 accuracy=82.308333	 loss=0.711265	 lr=0.000100
Epoch[085] Batch [1549]/[3760]	Speed: 35.092644 samples/sec	 accuracy=82.331653	 loss=0.711187	 lr=0.000100
Epoch[085] Batch [1599]/[3760]	Speed: 35.102513 samples/sec	 accuracy=82.302734	 loss=0.712025	 lr=0.000100
Epoch[085] Batch [1649]/[3760]	Speed: 34.972502 samples/sec	 accuracy=82.298295	 loss=0.712054	 lr=0.000100
Epoch[085] Batch [1699]/[3760]	Speed: 35.247846 samples/sec	 accuracy=82.286765	 loss=0.712884	 lr=0.000100
Epoch[085] Batch [1749]/[3760]	Speed: 35.006671 samples/sec	 accuracy=82.327679	 loss=0.711602	 lr=0.000100
Epoch[085] Batch [1799]/[3760]	Speed: 35.282539 samples/sec	 accuracy=82.320312	 loss=0.712424	 lr=0.000100
Epoch[085] Batch [1849]/[3760]	Speed: 34.694167 samples/sec	 accuracy=82.321791	 loss=0.712163	 lr=0.000100
Epoch[085] Batch [1899]/[3760]	Speed: 35.387448 samples/sec	 accuracy=82.298520	 loss=0.713477	 lr=0.000100
Epoch[085] Batch [1949]/[3760]	Speed: 34.831305 samples/sec	 accuracy=82.288462	 loss=0.714003	 lr=0.000100
Epoch[085] Batch [1999]/[3760]	Speed: 35.192661 samples/sec	 accuracy=82.303125	 loss=0.712514	 lr=0.000100
Epoch[085] Batch [2049]/[3760]	Speed: 35.142618 samples/sec	 accuracy=82.298780	 loss=0.713152	 lr=0.000100
Epoch[085] Batch [2099]/[3760]	Speed: 35.259052 samples/sec	 accuracy=82.307292	 loss=0.712034	 lr=0.000100
Epoch[085] Batch [2149]/[3760]	Speed: 35.133166 samples/sec	 accuracy=82.301599	 loss=0.712788	 lr=0.000100
Epoch[085] Batch [2199]/[3760]	Speed: 35.143318 samples/sec	 accuracy=82.292614	 loss=0.712313	 lr=0.000100
Epoch[085] Batch [2249]/[3760]	Speed: 35.304573 samples/sec	 accuracy=82.302778	 loss=0.711946	 lr=0.000100
Epoch[085] Batch [2299]/[3760]	Speed: 35.269443 samples/sec	 accuracy=82.322011	 loss=0.711859	 lr=0.000100
Epoch[085] Batch [2349]/[3760]	Speed: 35.331792 samples/sec	 accuracy=82.333777	 loss=0.711621	 lr=0.000100
Epoch[085] Batch [2399]/[3760]	Speed: 35.145602 samples/sec	 accuracy=82.324870	 loss=0.711871	 lr=0.000100
Epoch[085] Batch [2449]/[3760]	Speed: 34.944710 samples/sec	 accuracy=82.323342	 loss=0.712338	 lr=0.000100
Epoch[085] Batch [2499]/[3760]	Speed: 35.265420 samples/sec	 accuracy=82.318125	 loss=0.712660	 lr=0.000100
Epoch[085] Batch [2549]/[3760]	Speed: 35.009785 samples/sec	 accuracy=82.310662	 loss=0.712658	 lr=0.000100
Epoch[085] Batch [2599]/[3760]	Speed: 34.884515 samples/sec	 accuracy=82.331130	 loss=0.711756	 lr=0.000100
Epoch[085] Batch [2649]/[3760]	Speed: 35.000901 samples/sec	 accuracy=82.347877	 loss=0.710766	 lr=0.000100
Epoch[085] Batch [2699]/[3760]	Speed: 34.860645 samples/sec	 accuracy=82.354745	 loss=0.710326	 lr=0.000100
Epoch[085] Batch [2749]/[3760]	Speed: 35.073711 samples/sec	 accuracy=82.376705	 loss=0.709690	 lr=0.000100
Epoch[085] Batch [2799]/[3760]	Speed: 34.499872 samples/sec	 accuracy=82.375000	 loss=0.709565	 lr=0.000100
Epoch[085] Batch [2849]/[3760]	Speed: 35.193285 samples/sec	 accuracy=82.374452	 loss=0.709276	 lr=0.000100
Epoch[085] Batch [2899]/[3760]	Speed: 34.605896 samples/sec	 accuracy=82.365841	 loss=0.709879	 lr=0.000100
Epoch[085] Batch [2949]/[3760]	Speed: 35.213433 samples/sec	 accuracy=82.378178	 loss=0.709487	 lr=0.000100
Epoch[085] Batch [2999]/[3760]	Speed: 34.833945 samples/sec	 accuracy=82.391667	 loss=0.708657	 lr=0.000100
Epoch[085] Batch [3049]/[3760]	Speed: 34.895809 samples/sec	 accuracy=82.384734	 loss=0.708612	 lr=0.000100
Epoch[085] Batch [3099]/[3760]	Speed: 35.150936 samples/sec	 accuracy=82.367944	 loss=0.709214	 lr=0.000100
Epoch[085] Batch [3149]/[3760]	Speed: 34.752771 samples/sec	 accuracy=82.368552	 loss=0.709013	 lr=0.000100
Epoch[085] Batch [3199]/[3760]	Speed: 35.303572 samples/sec	 accuracy=82.376465	 loss=0.708429	 lr=0.000100
Epoch[085] Batch [3249]/[3760]	Speed: 34.802411 samples/sec	 accuracy=82.354808	 loss=0.709486	 lr=0.000100
Epoch[085] Batch [3299]/[3760]	Speed: 35.458097 samples/sec	 accuracy=82.357481	 loss=0.709553	 lr=0.000100
Epoch[085] Batch [3349]/[3760]	Speed: 35.087708 samples/sec	 accuracy=82.346082	 loss=0.709987	 lr=0.000100
Epoch[085] Batch [3399]/[3760]	Speed: 35.371719 samples/sec	 accuracy=82.340993	 loss=0.709960	 lr=0.000100
Epoch[085] Batch [3449]/[3760]	Speed: 35.222734 samples/sec	 accuracy=82.345562	 loss=0.709649	 lr=0.000100
Epoch[085] Batch [3499]/[3760]	Speed: 35.401658 samples/sec	 accuracy=82.328125	 loss=0.709682	 lr=0.000100
Epoch[085] Batch [3549]/[3760]	Speed: 35.068673 samples/sec	 accuracy=82.326585	 loss=0.710207	 lr=0.000100
Epoch[085] Batch [3599]/[3760]	Speed: 35.321293 samples/sec	 accuracy=82.316840	 loss=0.710858	 lr=0.000100
Epoch[085] Batch [3649]/[3760]	Speed: 35.119669 samples/sec	 accuracy=82.326199	 loss=0.710394	 lr=0.000100
Epoch[085] Batch [3699]/[3760]	Speed: 35.387555 samples/sec	 accuracy=82.310811	 loss=0.710795	 lr=0.000100
Epoch[085] Batch [3749]/[3760]	Speed: 36.420672 samples/sec	 accuracy=82.314167	 loss=0.710837	 lr=0.000100
Batch [0049]/[0303]: acc-top1=66.437500 acc-top5=86.125000
Batch [0099]/[0303]: acc-top1=66.328125 acc-top5=85.421875
Batch [0149]/[0303]: acc-top1=66.833333 acc-top5=85.500000
Batch [0199]/[0303]: acc-top1=66.523438 acc-top5=85.578125
Batch [0249]/[0303]: acc-top1=66.350000 acc-top5=85.537500
Batch [0299]/[0303]: acc-top1=66.234375 acc-top5=85.598958
[Epoch 085] training: accuracy=82.317570	 loss=0.710797
[Epoch 085] speed: 34 samples/sec	time cost: 7342.074641
[Epoch 085] validation: acc-top1=66.233498 acc-top5=85.591997 loss=1.631351
Epoch[086] Batch [0049]/[3760]	Speed: 25.007356 samples/sec	 accuracy=82.250000	 loss=0.700342	 lr=0.000100
Epoch[086] Batch [0099]/[3760]	Speed: 35.446540 samples/sec	 accuracy=82.109375	 loss=0.718966	 lr=0.000100
Epoch[086] Batch [0149]/[3760]	Speed: 35.224665 samples/sec	 accuracy=82.437500	 loss=0.704603	 lr=0.000100
Epoch[086] Batch [0199]/[3760]	Speed: 35.302242 samples/sec	 accuracy=82.445312	 loss=0.704694	 lr=0.000100
Epoch[086] Batch [0249]/[3760]	Speed: 35.269108 samples/sec	 accuracy=82.600000	 loss=0.695013	 lr=0.000100
Epoch[086] Batch [0299]/[3760]	Speed: 35.067969 samples/sec	 accuracy=82.437500	 loss=0.700050	 lr=0.000100
Epoch[086] Batch [0349]/[3760]	Speed: 34.939458 samples/sec	 accuracy=82.330357	 loss=0.708334	 lr=0.000100
Epoch[086] Batch [0399]/[3760]	Speed: 35.292939 samples/sec	 accuracy=82.292969	 loss=0.709053	 lr=0.000100
Epoch[086] Batch [0449]/[3760]	Speed: 34.949067 samples/sec	 accuracy=82.381944	 loss=0.705294	 lr=0.000100
Epoch[086] Batch [0499]/[3760]	Speed: 35.294621 samples/sec	 accuracy=82.281250	 loss=0.707293	 lr=0.000100
Epoch[086] Batch [0549]/[3760]	Speed: 34.809847 samples/sec	 accuracy=82.372159	 loss=0.702524	 lr=0.000100
Epoch[086] Batch [0599]/[3760]	Speed: 35.249704 samples/sec	 accuracy=82.333333	 loss=0.704932	 lr=0.000100
Epoch[086] Batch [0649]/[3760]	Speed: 34.725383 samples/sec	 accuracy=82.396635	 loss=0.702921	 lr=0.000100
Epoch[086] Batch [0699]/[3760]	Speed: 35.508636 samples/sec	 accuracy=82.386161	 loss=0.704038	 lr=0.000100
Epoch[086] Batch [0749]/[3760]	Speed: 34.975495 samples/sec	 accuracy=82.445833	 loss=0.703687	 lr=0.000100
Epoch[086] Batch [0799]/[3760]	Speed: 35.216991 samples/sec	 accuracy=82.439453	 loss=0.703996	 lr=0.000100
Epoch[086] Batch [0849]/[3760]	Speed: 35.026024 samples/sec	 accuracy=82.439338	 loss=0.702520	 lr=0.000100
Epoch[086] Batch [0899]/[3760]	Speed: 35.382327 samples/sec	 accuracy=82.484375	 loss=0.702879	 lr=0.000100
Epoch[086] Batch [0949]/[3760]	Speed: 34.828042 samples/sec	 accuracy=82.501645	 loss=0.702212	 lr=0.000100
Epoch[086] Batch [0999]/[3760]	Speed: 35.322744 samples/sec	 accuracy=82.560937	 loss=0.700194	 lr=0.000100
Epoch[086] Batch [1049]/[3760]	Speed: 35.033505 samples/sec	 accuracy=82.464286	 loss=0.703269	 lr=0.000100
Epoch[086] Batch [1099]/[3760]	Speed: 35.293274 samples/sec	 accuracy=82.407670	 loss=0.704270	 lr=0.000100
Epoch[086] Batch [1149]/[3760]	Speed: 35.247266 samples/sec	 accuracy=82.354620	 loss=0.705698	 lr=0.000100
Epoch[086] Batch [1199]/[3760]	Speed: 35.070191 samples/sec	 accuracy=82.325521	 loss=0.706876	 lr=0.000100
Epoch[086] Batch [1249]/[3760]	Speed: 35.363194 samples/sec	 accuracy=82.361250	 loss=0.704843	 lr=0.000100
Epoch[086] Batch [1299]/[3760]	Speed: 35.069662 samples/sec	 accuracy=82.355769	 loss=0.705041	 lr=0.000100
Epoch[086] Batch [1349]/[3760]	Speed: 35.364727 samples/sec	 accuracy=82.355324	 loss=0.704589	 lr=0.000100
Epoch[086] Batch [1399]/[3760]	Speed: 35.202726 samples/sec	 accuracy=82.370536	 loss=0.703408	 lr=0.000100
Epoch[086] Batch [1449]/[3760]	Speed: 35.277037 samples/sec	 accuracy=82.354526	 loss=0.703157	 lr=0.000100
Epoch[086] Batch [1499]/[3760]	Speed: 35.401382 samples/sec	 accuracy=82.311458	 loss=0.704488	 lr=0.000100
Epoch[086] Batch [1549]/[3760]	Speed: 35.345009 samples/sec	 accuracy=82.323589	 loss=0.704786	 lr=0.000100
Epoch[086] Batch [1599]/[3760]	Speed: 35.144746 samples/sec	 accuracy=82.300781	 loss=0.705884	 lr=0.000100
Epoch[086] Batch [1649]/[3760]	Speed: 35.435384 samples/sec	 accuracy=82.325758	 loss=0.705385	 lr=0.000100
Epoch[086] Batch [1699]/[3760]	Speed: 35.197474 samples/sec	 accuracy=82.345588	 loss=0.705353	 lr=0.000100
Epoch[086] Batch [1749]/[3760]	Speed: 35.352513 samples/sec	 accuracy=82.337500	 loss=0.706112	 lr=0.000100
Epoch[086] Batch [1799]/[3760]	Speed: 35.251037 samples/sec	 accuracy=82.306424	 loss=0.707289	 lr=0.000100
Epoch[086] Batch [1849]/[3760]	Speed: 35.378923 samples/sec	 accuracy=82.318412	 loss=0.706840	 lr=0.000100
Epoch[086] Batch [1899]/[3760]	Speed: 35.068362 samples/sec	 accuracy=82.305099	 loss=0.707299	 lr=0.000100
Epoch[086] Batch [1949]/[3760]	Speed: 35.295860 samples/sec	 accuracy=82.320513	 loss=0.706932	 lr=0.000100
Epoch[086] Batch [1999]/[3760]	Speed: 35.273618 samples/sec	 accuracy=82.303125	 loss=0.707958	 lr=0.000100
Epoch[086] Batch [2049]/[3760]	Speed: 35.069492 samples/sec	 accuracy=82.286585	 loss=0.707930	 lr=0.000100
Epoch[086] Batch [2099]/[3760]	Speed: 35.203100 samples/sec	 accuracy=82.320685	 loss=0.707404	 lr=0.000100
Epoch[086] Batch [2149]/[3760]	Speed: 35.083941 samples/sec	 accuracy=82.335756	 loss=0.707196	 lr=0.000100
Epoch[086] Batch [2199]/[3760]	Speed: 35.301896 samples/sec	 accuracy=82.330256	 loss=0.707080	 lr=0.000100
Epoch[086] Batch [2249]/[3760]	Speed: 35.065463 samples/sec	 accuracy=82.322222	 loss=0.707280	 lr=0.000100
Epoch[086] Batch [2299]/[3760]	Speed: 35.428362 samples/sec	 accuracy=82.326766	 loss=0.707294	 lr=0.000100
Epoch[086] Batch [2349]/[3760]	Speed: 35.195397 samples/sec	 accuracy=82.275931	 loss=0.708996	 lr=0.000100
Epoch[086] Batch [2399]/[3760]	Speed: 35.490943 samples/sec	 accuracy=82.259766	 loss=0.709456	 lr=0.000100
Epoch[086] Batch [2449]/[3760]	Speed: 35.344155 samples/sec	 accuracy=82.285077	 loss=0.708198	 lr=0.000100
Epoch[086] Batch [2499]/[3760]	Speed: 35.283696 samples/sec	 accuracy=82.290000	 loss=0.708072	 lr=0.000100
Epoch[086] Batch [2549]/[3760]	Speed: 35.380410 samples/sec	 accuracy=82.286765	 loss=0.707933	 lr=0.000100
Epoch[086] Batch [2599]/[3760]	Speed: 35.165135 samples/sec	 accuracy=82.272236	 loss=0.708580	 lr=0.000100
Epoch[086] Batch [2649]/[3760]	Speed: 35.170193 samples/sec	 accuracy=82.265920	 loss=0.708586	 lr=0.000100
Epoch[086] Batch [2699]/[3760]	Speed: 35.157550 samples/sec	 accuracy=82.250579	 loss=0.708874	 lr=0.000100
Epoch[086] Batch [2749]/[3760]	Speed: 35.310351 samples/sec	 accuracy=82.247727	 loss=0.708968	 lr=0.000100
Epoch[086] Batch [2799]/[3760]	Speed: 35.395429 samples/sec	 accuracy=82.260603	 loss=0.708464	 lr=0.000100
Epoch[086] Batch [2849]/[3760]	Speed: 35.235677 samples/sec	 accuracy=82.231908	 loss=0.709797	 lr=0.000100
Epoch[086] Batch [2899]/[3760]	Speed: 35.325672 samples/sec	 accuracy=82.220905	 loss=0.710269	 lr=0.000100
Epoch[086] Batch [2949]/[3760]	Speed: 35.022397 samples/sec	 accuracy=82.203390	 loss=0.711194	 lr=0.000100
Epoch[086] Batch [2999]/[3760]	Speed: 35.369158 samples/sec	 accuracy=82.206771	 loss=0.711457	 lr=0.000100
Epoch[086] Batch [3049]/[3760]	Speed: 35.070019 samples/sec	 accuracy=82.229508	 loss=0.710738	 lr=0.000100
Epoch[086] Batch [3099]/[3760]	Speed: 35.191882 samples/sec	 accuracy=82.253024	 loss=0.709937	 lr=0.000100
Epoch[086] Batch [3149]/[3760]	Speed: 35.088072 samples/sec	 accuracy=82.242560	 loss=0.710166	 lr=0.000100
Epoch[086] Batch [3199]/[3760]	Speed: 35.401944 samples/sec	 accuracy=82.249512	 loss=0.710249	 lr=0.000100
Epoch[086] Batch [3249]/[3760]	Speed: 35.235899 samples/sec	 accuracy=82.249519	 loss=0.709836	 lr=0.000100
Epoch[086] Batch [3299]/[3760]	Speed: 35.016031 samples/sec	 accuracy=82.240057	 loss=0.710157	 lr=0.000100
Epoch[086] Batch [3349]/[3760]	Speed: 35.046217 samples/sec	 accuracy=82.238340	 loss=0.710308	 lr=0.000100
Epoch[086] Batch [3399]/[3760]	Speed: 35.002887 samples/sec	 accuracy=82.244485	 loss=0.710213	 lr=0.000100
Epoch[086] Batch [3449]/[3760]	Speed: 35.143269 samples/sec	 accuracy=82.265399	 loss=0.709729	 lr=0.000100
Epoch[086] Batch [3499]/[3760]	Speed: 35.028199 samples/sec	 accuracy=82.274554	 loss=0.709384	 lr=0.000100
Epoch[086] Batch [3549]/[3760]	Speed: 35.221947 samples/sec	 accuracy=82.273768	 loss=0.709377	 lr=0.000100
Epoch[086] Batch [3599]/[3760]	Speed: 34.792716 samples/sec	 accuracy=82.263455	 loss=0.709791	 lr=0.000100
Epoch[086] Batch [3649]/[3760]	Speed: 35.217741 samples/sec	 accuracy=82.264983	 loss=0.709758	 lr=0.000100
Epoch[086] Batch [3699]/[3760]	Speed: 34.763382 samples/sec	 accuracy=82.261824	 loss=0.709692	 lr=0.000100
Epoch[086] Batch [3749]/[3760]	Speed: 36.307591 samples/sec	 accuracy=82.252500	 loss=0.710020	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.312500 acc-top5=85.906250
Batch [0099]/[0303]: acc-top1=66.609375 acc-top5=85.312500
Batch [0149]/[0303]: acc-top1=66.958333 acc-top5=85.531250
Batch [0199]/[0303]: acc-top1=66.593750 acc-top5=85.578125
Batch [0249]/[0303]: acc-top1=66.481250 acc-top5=85.475000
Batch [0299]/[0303]: acc-top1=66.338542 acc-top5=85.473958
[Epoch 086] training: accuracy=82.252743	 loss=0.710251
[Epoch 086] speed: 35 samples/sec	time cost: 7324.276762
[Epoch 086] validation: acc-top1=66.352104 acc-top5=85.483705 loss=1.639801
Epoch[087] Batch [0049]/[3759]	Speed: 25.067632 samples/sec	 accuracy=82.062500	 loss=0.687174	 lr=0.000100
Epoch[087] Batch [0099]/[3759]	Speed: 34.914242 samples/sec	 accuracy=82.062500	 loss=0.710613	 lr=0.000100
Epoch[087] Batch [0149]/[3759]	Speed: 35.230920 samples/sec	 accuracy=82.250000	 loss=0.700979	 lr=0.000100
Epoch[087] Batch [0199]/[3759]	Speed: 34.701723 samples/sec	 accuracy=82.015625	 loss=0.706415	 lr=0.000100
Epoch[087] Batch [0249]/[3759]	Speed: 35.382142 samples/sec	 accuracy=82.062500	 loss=0.707624	 lr=0.000100
Epoch[087] Batch [0299]/[3759]	Speed: 34.994582 samples/sec	 accuracy=82.276042	 loss=0.701777	 lr=0.000100
Epoch[087] Batch [0349]/[3759]	Speed: 35.249032 samples/sec	 accuracy=82.187500	 loss=0.705165	 lr=0.000100
Epoch[087] Batch [0399]/[3759]	Speed: 35.081962 samples/sec	 accuracy=82.242188	 loss=0.703550	 lr=0.000100
Epoch[087] Batch [0449]/[3759]	Speed: 35.022781 samples/sec	 accuracy=82.281250	 loss=0.703191	 lr=0.000100
Epoch[087] Batch [0499]/[3759]	Speed: 35.159970 samples/sec	 accuracy=82.303125	 loss=0.704406	 lr=0.000100
Epoch[087] Batch [0549]/[3759]	Speed: 35.123727 samples/sec	 accuracy=82.326705	 loss=0.701574	 lr=0.000100
Epoch[087] Batch [0599]/[3759]	Speed: 35.089230 samples/sec	 accuracy=82.395833	 loss=0.699120	 lr=0.000100
Epoch[087] Batch [0649]/[3759]	Speed: 35.107123 samples/sec	 accuracy=82.403846	 loss=0.698459	 lr=0.000100
Epoch[087] Batch [0699]/[3759]	Speed: 35.273729 samples/sec	 accuracy=82.341518	 loss=0.701259	 lr=0.000100
Epoch[087] Batch [0749]/[3759]	Speed: 34.936678 samples/sec	 accuracy=82.331250	 loss=0.702490	 lr=0.000100
Epoch[087] Batch [0799]/[3759]	Speed: 35.180431 samples/sec	 accuracy=82.345703	 loss=0.701452	 lr=0.000100
Epoch[087] Batch [0849]/[3759]	Speed: 35.162789 samples/sec	 accuracy=82.351103	 loss=0.701084	 lr=0.000100
Epoch[087] Batch [0899]/[3759]	Speed: 35.255278 samples/sec	 accuracy=82.338542	 loss=0.701269	 lr=0.000100
Epoch[087] Batch [0949]/[3759]	Speed: 34.748263 samples/sec	 accuracy=82.378289	 loss=0.700272	 lr=0.000100
Epoch[087] Batch [0999]/[3759]	Speed: 35.154748 samples/sec	 accuracy=82.403125	 loss=0.699064	 lr=0.000100
Epoch[087] Batch [1049]/[3759]	Speed: 34.352802 samples/sec	 accuracy=82.409226	 loss=0.698710	 lr=0.000100
Epoch[087] Batch [1099]/[3759]	Speed: 35.205518 samples/sec	 accuracy=82.348011	 loss=0.700996	 lr=0.000100
Epoch[087] Batch [1149]/[3759]	Speed: 34.633105 samples/sec	 accuracy=82.358696	 loss=0.701637	 lr=0.000100
Epoch[087] Batch [1199]/[3759]	Speed: 35.001130 samples/sec	 accuracy=82.339844	 loss=0.703040	 lr=0.000100
Epoch[087] Batch [1249]/[3759]	Speed: 35.066354 samples/sec	 accuracy=82.340000	 loss=0.703164	 lr=0.000100
Epoch[087] Batch [1299]/[3759]	Speed: 34.983584 samples/sec	 accuracy=82.358173	 loss=0.703754	 lr=0.000100
Epoch[087] Batch [1349]/[3759]	Speed: 34.972719 samples/sec	 accuracy=82.318287	 loss=0.705457	 lr=0.000100
Epoch[087] Batch [1399]/[3759]	Speed: 34.880237 samples/sec	 accuracy=82.353795	 loss=0.704665	 lr=0.000100
Epoch[087] Batch [1449]/[3759]	Speed: 34.849132 samples/sec	 accuracy=82.337284	 loss=0.704836	 lr=0.000100
Epoch[087] Batch [1499]/[3759]	Speed: 34.859526 samples/sec	 accuracy=82.323958	 loss=0.704930	 lr=0.000100
Epoch[087] Batch [1549]/[3759]	Speed: 35.096511 samples/sec	 accuracy=82.320565	 loss=0.705576	 lr=0.000100
Epoch[087] Batch [1599]/[3759]	Speed: 34.815405 samples/sec	 accuracy=82.334961	 loss=0.705261	 lr=0.000100
Epoch[087] Batch [1649]/[3759]	Speed: 35.228694 samples/sec	 accuracy=82.345644	 loss=0.704188	 lr=0.000100
Epoch[087] Batch [1699]/[3759]	Speed: 34.594184 samples/sec	 accuracy=82.331801	 loss=0.703579	 lr=0.000100
Epoch[087] Batch [1749]/[3759]	Speed: 35.144363 samples/sec	 accuracy=82.308929	 loss=0.704123	 lr=0.000100
Epoch[087] Batch [1799]/[3759]	Speed: 34.511140 samples/sec	 accuracy=82.287326	 loss=0.704713	 lr=0.000100
Epoch[087] Batch [1849]/[3759]	Speed: 35.226684 samples/sec	 accuracy=82.288007	 loss=0.704785	 lr=0.000100
Epoch[087] Batch [1899]/[3759]	Speed: 34.674910 samples/sec	 accuracy=82.258224	 loss=0.706272	 lr=0.000100
Epoch[087] Batch [1949]/[3759]	Speed: 35.090607 samples/sec	 accuracy=82.230769	 loss=0.706746	 lr=0.000100
Epoch[087] Batch [1999]/[3759]	Speed: 34.915091 samples/sec	 accuracy=82.214844	 loss=0.707290	 lr=0.000100
Epoch[087] Batch [2049]/[3759]	Speed: 34.815330 samples/sec	 accuracy=82.188262	 loss=0.707655	 lr=0.000100
Epoch[087] Batch [2099]/[3759]	Speed: 34.985202 samples/sec	 accuracy=82.210565	 loss=0.707378	 lr=0.000100
Epoch[087] Batch [2149]/[3759]	Speed: 34.704712 samples/sec	 accuracy=82.232558	 loss=0.706570	 lr=0.000100
Epoch[087] Batch [2199]/[3759]	Speed: 35.264248 samples/sec	 accuracy=82.237926	 loss=0.705940	 lr=0.000100
Epoch[087] Batch [2249]/[3759]	Speed: 34.809221 samples/sec	 accuracy=82.255556	 loss=0.705317	 lr=0.000100
Epoch[087] Batch [2299]/[3759]	Speed: 35.306063 samples/sec	 accuracy=82.267663	 loss=0.705493	 lr=0.000100
Epoch[087] Batch [2349]/[3759]	Speed: 34.789596 samples/sec	 accuracy=82.257979	 loss=0.705874	 lr=0.000100
Epoch[087] Batch [2399]/[3759]	Speed: 35.388208 samples/sec	 accuracy=82.273438	 loss=0.705655	 lr=0.000100
Epoch[087] Batch [2449]/[3759]	Speed: 34.864634 samples/sec	 accuracy=82.246173	 loss=0.706391	 lr=0.000100
Epoch[087] Batch [2499]/[3759]	Speed: 35.266164 samples/sec	 accuracy=82.258750	 loss=0.706416	 lr=0.000100
Epoch[087] Batch [2549]/[3759]	Speed: 35.304025 samples/sec	 accuracy=82.258578	 loss=0.706525	 lr=0.000100
Epoch[087] Batch [2599]/[3759]	Speed: 35.106841 samples/sec	 accuracy=82.245192	 loss=0.706907	 lr=0.000100
Epoch[087] Batch [2649]/[3759]	Speed: 35.290917 samples/sec	 accuracy=82.243514	 loss=0.707678	 lr=0.000100
Epoch[087] Batch [2699]/[3759]	Speed: 35.068301 samples/sec	 accuracy=82.217014	 loss=0.708699	 lr=0.000100
Epoch[087] Batch [2749]/[3759]	Speed: 35.342176 samples/sec	 accuracy=82.233523	 loss=0.708325	 lr=0.000100
Epoch[087] Batch [2799]/[3759]	Speed: 35.269714 samples/sec	 accuracy=82.239397	 loss=0.707852	 lr=0.000100
Epoch[087] Batch [2849]/[3759]	Speed: 35.315703 samples/sec	 accuracy=82.229167	 loss=0.708615	 lr=0.000100
Epoch[087] Batch [2899]/[3759]	Speed: 35.221448 samples/sec	 accuracy=82.224138	 loss=0.708954	 lr=0.000100
Epoch[087] Batch [2949]/[3759]	Speed: 35.174982 samples/sec	 accuracy=82.228814	 loss=0.708427	 lr=0.000100
Epoch[087] Batch [2999]/[3759]	Speed: 35.225687 samples/sec	 accuracy=82.231250	 loss=0.708346	 lr=0.000100
Epoch[087] Batch [3049]/[3759]	Speed: 35.367220 samples/sec	 accuracy=82.245389	 loss=0.708094	 lr=0.000100
Epoch[087] Batch [3099]/[3759]	Speed: 35.246782 samples/sec	 accuracy=82.254536	 loss=0.708128	 lr=0.000100
Epoch[087] Batch [3149]/[3759]	Speed: 35.430888 samples/sec	 accuracy=82.268849	 loss=0.707598	 lr=0.000100
Epoch[087] Batch [3199]/[3759]	Speed: 35.138912 samples/sec	 accuracy=82.278320	 loss=0.707214	 lr=0.000100
Epoch[087] Batch [3249]/[3759]	Speed: 35.292893 samples/sec	 accuracy=82.267308	 loss=0.707286	 lr=0.000100
Epoch[087] Batch [3299]/[3759]	Speed: 35.185781 samples/sec	 accuracy=82.265152	 loss=0.707491	 lr=0.000100
Epoch[087] Batch [3349]/[3759]	Speed: 35.379111 samples/sec	 accuracy=82.263993	 loss=0.707748	 lr=0.000100
Epoch[087] Batch [3399]/[3759]	Speed: 35.020186 samples/sec	 accuracy=82.262868	 loss=0.707932	 lr=0.000100
Epoch[087] Batch [3449]/[3759]	Speed: 35.414905 samples/sec	 accuracy=82.265851	 loss=0.707730	 lr=0.000100
Epoch[087] Batch [3499]/[3759]	Speed: 34.882488 samples/sec	 accuracy=82.265625	 loss=0.707516	 lr=0.000100
Epoch[087] Batch [3549]/[3759]	Speed: 35.339203 samples/sec	 accuracy=82.271127	 loss=0.707496	 lr=0.000100
Epoch[087] Batch [3599]/[3759]	Speed: 35.045881 samples/sec	 accuracy=82.283854	 loss=0.707389	 lr=0.000100
Epoch[087] Batch [3649]/[3759]	Speed: 35.250943 samples/sec	 accuracy=82.296661	 loss=0.706949	 lr=0.000100
Epoch[087] Batch [3699]/[3759]	Speed: 35.054475 samples/sec	 accuracy=82.298142	 loss=0.706997	 lr=0.000100
Epoch[087] Batch [3749]/[3759]	Speed: 36.512576 samples/sec	 accuracy=82.307500	 loss=0.707355	 lr=0.000100
Batch [0049]/[0303]: acc-top1=66.906250 acc-top5=86.125000
Batch [0099]/[0303]: acc-top1=66.437500 acc-top5=85.296875
Batch [0149]/[0303]: acc-top1=66.875000 acc-top5=85.572917
Batch [0199]/[0303]: acc-top1=66.562500 acc-top5=85.570312
Batch [0249]/[0303]: acc-top1=66.412500 acc-top5=85.525000
Batch [0299]/[0303]: acc-top1=66.291667 acc-top5=85.578125
[Epoch 087] training: accuracy=82.306631	 loss=0.707435
[Epoch 087] speed: 34 samples/sec	time cost: 7343.305966
[Epoch 087] validation: acc-top1=66.305693 acc-top5=85.586840 loss=1.629973
Epoch[088] Batch [0049]/[3760]	Speed: 25.322904 samples/sec	 accuracy=81.843750	 loss=0.736815	 lr=0.000100
Epoch[088] Batch [0099]/[3760]	Speed: 35.389084 samples/sec	 accuracy=81.843750	 loss=0.727437	 lr=0.000100
Epoch[088] Batch [0149]/[3760]	Speed: 35.234432 samples/sec	 accuracy=82.114583	 loss=0.712483	 lr=0.000100
Epoch[088] Batch [0199]/[3760]	Speed: 35.012742 samples/sec	 accuracy=81.984375	 loss=0.715815	 lr=0.000100
Epoch[088] Batch [0249]/[3760]	Speed: 35.173707 samples/sec	 accuracy=82.081250	 loss=0.720061	 lr=0.000100
Epoch[088] Batch [0299]/[3760]	Speed: 35.259782 samples/sec	 accuracy=82.000000	 loss=0.723212	 lr=0.000100
Epoch[088] Batch [0349]/[3760]	Speed: 34.911518 samples/sec	 accuracy=81.986607	 loss=0.720109	 lr=0.000100
Epoch[088] Batch [0399]/[3760]	Speed: 35.387672 samples/sec	 accuracy=82.070312	 loss=0.719697	 lr=0.000100
Epoch[088] Batch [0449]/[3760]	Speed: 34.649904 samples/sec	 accuracy=82.045139	 loss=0.722305	 lr=0.000100
Epoch[088] Batch [0499]/[3760]	Speed: 35.439763 samples/sec	 accuracy=82.146875	 loss=0.715923	 lr=0.000100
Epoch[088] Batch [0549]/[3760]	Speed: 34.910592 samples/sec	 accuracy=82.116477	 loss=0.714496	 lr=0.000100
Epoch[088] Batch [0599]/[3760]	Speed: 35.432177 samples/sec	 accuracy=82.164062	 loss=0.713997	 lr=0.000100
Epoch[088] Batch [0649]/[3760]	Speed: 34.862184 samples/sec	 accuracy=82.177885	 loss=0.712935	 lr=0.000100
Epoch[088] Batch [0699]/[3760]	Speed: 35.326728 samples/sec	 accuracy=82.234375	 loss=0.709675	 lr=0.000100
Epoch[088] Batch [0749]/[3760]	Speed: 34.958589 samples/sec	 accuracy=82.302083	 loss=0.706205	 lr=0.000100
Epoch[088] Batch [0799]/[3760]	Speed: 35.431331 samples/sec	 accuracy=82.351562	 loss=0.706430	 lr=0.000100
Epoch[088] Batch [0849]/[3760]	Speed: 34.961213 samples/sec	 accuracy=82.369485	 loss=0.705639	 lr=0.000100
Epoch[088] Batch [0899]/[3760]	Speed: 35.162925 samples/sec	 accuracy=82.364583	 loss=0.707236	 lr=0.000100
Epoch[088] Batch [0949]/[3760]	Speed: 35.286956 samples/sec	 accuracy=82.417763	 loss=0.705205	 lr=0.000100
Epoch[088] Batch [0999]/[3760]	Speed: 35.043758 samples/sec	 accuracy=82.464062	 loss=0.702609	 lr=0.000100
Epoch[088] Batch [1049]/[3760]	Speed: 35.079909 samples/sec	 accuracy=82.461310	 loss=0.702857	 lr=0.000100
Epoch[088] Batch [1099]/[3760]	Speed: 34.842802 samples/sec	 accuracy=82.490057	 loss=0.701029	 lr=0.000100
Epoch[088] Batch [1149]/[3760]	Speed: 35.281049 samples/sec	 accuracy=82.468750	 loss=0.701749	 lr=0.000100
Epoch[088] Batch [1199]/[3760]	Speed: 35.063694 samples/sec	 accuracy=82.440104	 loss=0.701603	 lr=0.000100
Epoch[088] Batch [1249]/[3760]	Speed: 35.299339 samples/sec	 accuracy=82.426250	 loss=0.701638	 lr=0.000100
Epoch[088] Batch [1299]/[3760]	Speed: 34.868581 samples/sec	 accuracy=82.381010	 loss=0.703901	 lr=0.000100
Epoch[088] Batch [1349]/[3760]	Speed: 35.556511 samples/sec	 accuracy=82.370370	 loss=0.705588	 lr=0.000100
Epoch[088] Batch [1399]/[3760]	Speed: 35.148027 samples/sec	 accuracy=82.366071	 loss=0.705885	 lr=0.000100
Epoch[088] Batch [1449]/[3760]	Speed: 35.347378 samples/sec	 accuracy=82.366379	 loss=0.705003	 lr=0.000100
Epoch[088] Batch [1499]/[3760]	Speed: 35.331512 samples/sec	 accuracy=82.347917	 loss=0.704884	 lr=0.000100
Epoch[088] Batch [1549]/[3760]	Speed: 35.197148 samples/sec	 accuracy=82.346774	 loss=0.705028	 lr=0.000100
Epoch[088] Batch [1599]/[3760]	Speed: 35.254358 samples/sec	 accuracy=82.380859	 loss=0.703862	 lr=0.000100
Epoch[088] Batch [1649]/[3760]	Speed: 35.288352 samples/sec	 accuracy=82.432765	 loss=0.702145	 lr=0.000100
Epoch[088] Batch [1699]/[3760]	Speed: 35.287740 samples/sec	 accuracy=82.414522	 loss=0.702495	 lr=0.000100
Epoch[088] Batch [1749]/[3760]	Speed: 35.290243 samples/sec	 accuracy=82.414286	 loss=0.702628	 lr=0.000100
Epoch[088] Batch [1799]/[3760]	Speed: 35.130126 samples/sec	 accuracy=82.428819	 loss=0.701558	 lr=0.000100
Epoch[088] Batch [1849]/[3760]	Speed: 35.063253 samples/sec	 accuracy=82.416385	 loss=0.702399	 lr=0.000100
Epoch[088] Batch [1899]/[3760]	Speed: 35.100038 samples/sec	 accuracy=82.435033	 loss=0.702100	 lr=0.000100
Epoch[088] Batch [1949]/[3760]	Speed: 34.951088 samples/sec	 accuracy=82.443910	 loss=0.702200	 lr=0.000100
Epoch[088] Batch [1999]/[3760]	Speed: 35.273671 samples/sec	 accuracy=82.432031	 loss=0.702741	 lr=0.000100
Epoch[088] Batch [2049]/[3760]	Speed: 34.784318 samples/sec	 accuracy=82.414634	 loss=0.703564	 lr=0.000100
Epoch[088] Batch [2099]/[3760]	Speed: 35.425324 samples/sec	 accuracy=82.413690	 loss=0.703670	 lr=0.000100
Epoch[088] Batch [2149]/[3760]	Speed: 35.003304 samples/sec	 accuracy=82.439680	 loss=0.702842	 lr=0.000100
Epoch[088] Batch [2199]/[3760]	Speed: 35.249173 samples/sec	 accuracy=82.448153	 loss=0.702441	 lr=0.000100
Epoch[088] Batch [2249]/[3760]	Speed: 34.898990 samples/sec	 accuracy=82.439583	 loss=0.702950	 lr=0.000100
Epoch[088] Batch [2299]/[3760]	Speed: 35.214792 samples/sec	 accuracy=82.430707	 loss=0.702682	 lr=0.000100
Epoch[088] Batch [2349]/[3760]	Speed: 34.879337 samples/sec	 accuracy=82.422872	 loss=0.703045	 lr=0.000100
Epoch[088] Batch [2399]/[3760]	Speed: 35.288273 samples/sec	 accuracy=82.415365	 loss=0.703326	 lr=0.000100
Epoch[088] Batch [2449]/[3760]	Speed: 35.284074 samples/sec	 accuracy=82.434311	 loss=0.703091	 lr=0.000100
Epoch[088] Batch [2499]/[3760]	Speed: 35.093278 samples/sec	 accuracy=82.445625	 loss=0.703187	 lr=0.000100
Epoch[088] Batch [2549]/[3760]	Speed: 35.321217 samples/sec	 accuracy=82.434436	 loss=0.703376	 lr=0.000100
Epoch[088] Batch [2599]/[3760]	Speed: 34.867093 samples/sec	 accuracy=82.432091	 loss=0.703287	 lr=0.000100
Epoch[088] Batch [2649]/[3760]	Speed: 35.398365 samples/sec	 accuracy=82.443396	 loss=0.702936	 lr=0.000100
Epoch[088] Batch [2699]/[3760]	Speed: 35.037579 samples/sec	 accuracy=82.445023	 loss=0.702861	 lr=0.000100
Epoch[088] Batch [2749]/[3760]	Speed: 35.308808 samples/sec	 accuracy=82.459091	 loss=0.702708	 lr=0.000100
Epoch[088] Batch [2799]/[3760]	Speed: 35.015245 samples/sec	 accuracy=82.455357	 loss=0.702457	 lr=0.000100
Epoch[088] Batch [2849]/[3760]	Speed: 35.289291 samples/sec	 accuracy=82.459978	 loss=0.702128	 lr=0.000100
Epoch[088] Batch [2899]/[3760]	Speed: 35.253970 samples/sec	 accuracy=82.481142	 loss=0.701308	 lr=0.000100
Epoch[088] Batch [2949]/[3760]	Speed: 35.140708 samples/sec	 accuracy=82.489407	 loss=0.701230	 lr=0.000100
Epoch[088] Batch [2999]/[3760]	Speed: 35.025216 samples/sec	 accuracy=82.502604	 loss=0.700594	 lr=0.000100
Epoch[088] Batch [3049]/[3760]	Speed: 35.158351 samples/sec	 accuracy=82.503586	 loss=0.700514	 lr=0.000100
Epoch[088] Batch [3099]/[3760]	Speed: 35.266536 samples/sec	 accuracy=82.489919	 loss=0.700825	 lr=0.000100
Epoch[088] Batch [3149]/[3760]	Speed: 34.938221 samples/sec	 accuracy=82.492063	 loss=0.700826	 lr=0.000100
Epoch[088] Batch [3199]/[3760]	Speed: 35.437153 samples/sec	 accuracy=82.500488	 loss=0.700770	 lr=0.000100
Epoch[088] Batch [3249]/[3760]	Speed: 34.962332 samples/sec	 accuracy=82.480288	 loss=0.701332	 lr=0.000100
Epoch[088] Batch [3299]/[3760]	Speed: 35.546973 samples/sec	 accuracy=82.465909	 loss=0.701756	 lr=0.000100
Epoch[088] Batch [3349]/[3760]	Speed: 34.761375 samples/sec	 accuracy=82.470149	 loss=0.701172	 lr=0.000100
Epoch[088] Batch [3399]/[3760]	Speed: 35.431893 samples/sec	 accuracy=82.477941	 loss=0.700961	 lr=0.000100
Epoch[088] Batch [3449]/[3760]	Speed: 34.935425 samples/sec	 accuracy=82.469203	 loss=0.701556	 lr=0.000100
Epoch[088] Batch [3499]/[3760]	Speed: 35.341671 samples/sec	 accuracy=82.486161	 loss=0.701432	 lr=0.000100
Epoch[088] Batch [3549]/[3760]	Speed: 35.142749 samples/sec	 accuracy=82.484155	 loss=0.701563	 lr=0.000100
Epoch[088] Batch [3599]/[3760]	Speed: 35.329809 samples/sec	 accuracy=82.470920	 loss=0.702207	 lr=0.000100
Epoch[088] Batch [3649]/[3760]	Speed: 35.110372 samples/sec	 accuracy=82.473887	 loss=0.702153	 lr=0.000100
Epoch[088] Batch [3699]/[3760]	Speed: 35.128945 samples/sec	 accuracy=82.494932	 loss=0.701645	 lr=0.000100
Epoch[088] Batch [3749]/[3760]	Speed: 36.475915 samples/sec	 accuracy=82.473750	 loss=0.702133	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.281250 acc-top5=85.906250
Batch [0099]/[0303]: acc-top1=66.390625 acc-top5=85.156250
Batch [0149]/[0303]: acc-top1=66.937500 acc-top5=85.489583
Batch [0199]/[0303]: acc-top1=66.554688 acc-top5=85.523438
Batch [0249]/[0303]: acc-top1=66.412500 acc-top5=85.418750
Batch [0299]/[0303]: acc-top1=66.286458 acc-top5=85.390625
[Epoch 088] training: accuracy=82.474651	 loss=0.702352
[Epoch 088] speed: 34 samples/sec	time cost: 7326.342015
[Epoch 088] validation: acc-top1=66.295380 acc-top5=85.396040 loss=1.637782
Epoch[089] Batch [0049]/[3760]	Speed: 25.233651 samples/sec	 accuracy=82.656250	 loss=0.710728	 lr=0.000100
Epoch[089] Batch [0099]/[3760]	Speed: 35.034848 samples/sec	 accuracy=82.734375	 loss=0.698166	 lr=0.000100
Epoch[089] Batch [0149]/[3760]	Speed: 35.445591 samples/sec	 accuracy=82.937500	 loss=0.691518	 lr=0.000100
Epoch[089] Batch [0199]/[3760]	Speed: 35.101471 samples/sec	 accuracy=82.507812	 loss=0.705497	 lr=0.000100
Epoch[089] Batch [0249]/[3760]	Speed: 35.320078 samples/sec	 accuracy=82.587500	 loss=0.702280	 lr=0.000100
Epoch[089] Batch [0299]/[3760]	Speed: 35.257723 samples/sec	 accuracy=82.604167	 loss=0.700345	 lr=0.000100
Epoch[089] Batch [0349]/[3760]	Speed: 35.427057 samples/sec	 accuracy=82.732143	 loss=0.694973	 lr=0.000100
Epoch[089] Batch [0399]/[3760]	Speed: 35.003826 samples/sec	 accuracy=82.738281	 loss=0.698069	 lr=0.000100
Epoch[089] Batch [0449]/[3760]	Speed: 35.201200 samples/sec	 accuracy=82.791667	 loss=0.693078	 lr=0.000100
Epoch[089] Batch [0499]/[3760]	Speed: 35.466432 samples/sec	 accuracy=82.746875	 loss=0.693858	 lr=0.000100
Epoch[089] Batch [0549]/[3760]	Speed: 35.131520 samples/sec	 accuracy=82.664773	 loss=0.697532	 lr=0.000100
Epoch[089] Batch [0599]/[3760]	Speed: 35.528499 samples/sec	 accuracy=82.648438	 loss=0.697509	 lr=0.000100
Epoch[089] Batch [0649]/[3760]	Speed: 34.942012 samples/sec	 accuracy=82.584135	 loss=0.699312	 lr=0.000100
Epoch[089] Batch [0699]/[3760]	Speed: 35.347910 samples/sec	 accuracy=82.629464	 loss=0.698223	 lr=0.000100
Epoch[089] Batch [0749]/[3760]	Speed: 35.031215 samples/sec	 accuracy=82.635417	 loss=0.697022	 lr=0.000100
Epoch[089] Batch [0799]/[3760]	Speed: 35.434166 samples/sec	 accuracy=82.568359	 loss=0.698985	 lr=0.000100
Epoch[089] Batch [0849]/[3760]	Speed: 35.516427 samples/sec	 accuracy=82.586397	 loss=0.698072	 lr=0.000100
Epoch[089] Batch [0899]/[3760]	Speed: 35.296092 samples/sec	 accuracy=82.520833	 loss=0.698575	 lr=0.000100
Epoch[089] Batch [0949]/[3760]	Speed: 35.490233 samples/sec	 accuracy=82.513158	 loss=0.698418	 lr=0.000100
Epoch[089] Batch [0999]/[3760]	Speed: 35.269604 samples/sec	 accuracy=82.514062	 loss=0.697708	 lr=0.000100
Epoch[089] Batch [1049]/[3760]	Speed: 35.516023 samples/sec	 accuracy=82.568452	 loss=0.694828	 lr=0.000100
Epoch[089] Batch [1099]/[3760]	Speed: 35.096750 samples/sec	 accuracy=82.612216	 loss=0.693198	 lr=0.000100
Epoch[089] Batch [1149]/[3760]	Speed: 35.350000 samples/sec	 accuracy=82.550272	 loss=0.693679	 lr=0.000100
Epoch[089] Batch [1199]/[3760]	Speed: 35.009108 samples/sec	 accuracy=82.571615	 loss=0.694277	 lr=0.000100
Epoch[089] Batch [1249]/[3760]	Speed: 35.476110 samples/sec	 accuracy=82.580000	 loss=0.694413	 lr=0.000100
Epoch[089] Batch [1299]/[3760]	Speed: 34.759636 samples/sec	 accuracy=82.569712	 loss=0.694599	 lr=0.000100
Epoch[089] Batch [1349]/[3760]	Speed: 35.524699 samples/sec	 accuracy=82.605324	 loss=0.694072	 lr=0.000100
Epoch[089] Batch [1399]/[3760]	Speed: 34.872304 samples/sec	 accuracy=82.587054	 loss=0.695562	 lr=0.000100
Epoch[089] Batch [1449]/[3760]	Speed: 35.427183 samples/sec	 accuracy=82.567888	 loss=0.696079	 lr=0.000100
Epoch[089] Batch [1499]/[3760]	Speed: 35.102588 samples/sec	 accuracy=82.550000	 loss=0.696793	 lr=0.000100
Epoch[089] Batch [1549]/[3760]	Speed: 35.305924 samples/sec	 accuracy=82.573589	 loss=0.695823	 lr=0.000100
Epoch[089] Batch [1599]/[3760]	Speed: 35.319711 samples/sec	 accuracy=82.592773	 loss=0.694848	 lr=0.000100
Epoch[089] Batch [1649]/[3760]	Speed: 35.190470 samples/sec	 accuracy=82.589015	 loss=0.695314	 lr=0.000100
Epoch[089] Batch [1699]/[3760]	Speed: 35.466529 samples/sec	 accuracy=82.601103	 loss=0.694248	 lr=0.000100
Epoch[089] Batch [1749]/[3760]	Speed: 35.201328 samples/sec	 accuracy=82.630357	 loss=0.693394	 lr=0.000100
Epoch[089] Batch [1799]/[3760]	Speed: 35.451848 samples/sec	 accuracy=82.625868	 loss=0.693713	 lr=0.000100
Epoch[089] Batch [1849]/[3760]	Speed: 35.149809 samples/sec	 accuracy=82.598818	 loss=0.694699	 lr=0.000100
Epoch[089] Batch [1899]/[3760]	Speed: 35.307672 samples/sec	 accuracy=82.587171	 loss=0.694569	 lr=0.000100
Epoch[089] Batch [1949]/[3760]	Speed: 35.048870 samples/sec	 accuracy=82.567308	 loss=0.695035	 lr=0.000100
Epoch[089] Batch [1999]/[3760]	Speed: 35.529203 samples/sec	 accuracy=82.591406	 loss=0.694883	 lr=0.000100
Epoch[089] Batch [2049]/[3760]	Speed: 34.879554 samples/sec	 accuracy=82.597561	 loss=0.694650	 lr=0.000100
Epoch[089] Batch [2099]/[3760]	Speed: 35.551678 samples/sec	 accuracy=82.582589	 loss=0.695208	 lr=0.000100
Epoch[089] Batch [2149]/[3760]	Speed: 35.096559 samples/sec	 accuracy=82.593023	 loss=0.695275	 lr=0.000100
Epoch[089] Batch [2199]/[3760]	Speed: 35.398051 samples/sec	 accuracy=82.578125	 loss=0.695609	 lr=0.000100
Epoch[089] Batch [2249]/[3760]	Speed: 35.292955 samples/sec	 accuracy=82.588194	 loss=0.694600	 lr=0.000100
Epoch[089] Batch [2299]/[3760]	Speed: 35.117877 samples/sec	 accuracy=82.595109	 loss=0.694506	 lr=0.000100
Epoch[089] Batch [2349]/[3760]	Speed: 35.209750 samples/sec	 accuracy=82.600399	 loss=0.694674	 lr=0.000100
Epoch[089] Batch [2399]/[3760]	Speed: 35.261957 samples/sec	 accuracy=82.598307	 loss=0.695232	 lr=0.000100
Epoch[089] Batch [2449]/[3760]	Speed: 35.337320 samples/sec	 accuracy=82.586735	 loss=0.695686	 lr=0.000100
Epoch[089] Batch [2499]/[3760]	Speed: 35.068774 samples/sec	 accuracy=82.608750	 loss=0.695018	 lr=0.000100
Epoch[089] Batch [2549]/[3760]	Speed: 35.436185 samples/sec	 accuracy=82.612132	 loss=0.694950	 lr=0.000100
Epoch[089] Batch [2599]/[3760]	Speed: 34.839297 samples/sec	 accuracy=82.624399	 loss=0.694736	 lr=0.000100
Epoch[089] Batch [2649]/[3760]	Speed: 35.305062 samples/sec	 accuracy=82.600825	 loss=0.695367	 lr=0.000100
Epoch[089] Batch [2699]/[3760]	Speed: 34.645435 samples/sec	 accuracy=82.589699	 loss=0.695590	 lr=0.000100
Epoch[089] Batch [2749]/[3760]	Speed: 35.485467 samples/sec	 accuracy=82.616477	 loss=0.695152	 lr=0.000100
Epoch[089] Batch [2799]/[3760]	Speed: 34.766061 samples/sec	 accuracy=82.594308	 loss=0.695937	 lr=0.000100
Epoch[089] Batch [2849]/[3760]	Speed: 35.240211 samples/sec	 accuracy=82.587171	 loss=0.696125	 lr=0.000100
Epoch[089] Batch [2899]/[3760]	Speed: 35.108456 samples/sec	 accuracy=82.584052	 loss=0.696209	 lr=0.000100
Epoch[089] Batch [2949]/[3760]	Speed: 34.941309 samples/sec	 accuracy=82.597987	 loss=0.696367	 lr=0.000100
Epoch[089] Batch [2999]/[3760]	Speed: 35.002077 samples/sec	 accuracy=82.590625	 loss=0.696943	 lr=0.000100
Epoch[089] Batch [3049]/[3760]	Speed: 34.723185 samples/sec	 accuracy=82.585041	 loss=0.697485	 lr=0.000100
Epoch[089] Batch [3099]/[3760]	Speed: 35.318595 samples/sec	 accuracy=82.587198	 loss=0.697451	 lr=0.000100
Epoch[089] Batch [3149]/[3760]	Speed: 34.724872 samples/sec	 accuracy=82.574901	 loss=0.698241	 lr=0.000100
Epoch[089] Batch [3199]/[3760]	Speed: 35.401694 samples/sec	 accuracy=82.546875	 loss=0.699099	 lr=0.000100
Epoch[089] Batch [3249]/[3760]	Speed: 34.740513 samples/sec	 accuracy=82.540385	 loss=0.699094	 lr=0.000100
Epoch[089] Batch [3299]/[3760]	Speed: 35.345471 samples/sec	 accuracy=82.552083	 loss=0.698615	 lr=0.000100
Epoch[089] Batch [3349]/[3760]	Speed: 34.957380 samples/sec	 accuracy=82.535914	 loss=0.698987	 lr=0.000100
Epoch[089] Batch [3399]/[3760]	Speed: 35.256944 samples/sec	 accuracy=82.538143	 loss=0.698839	 lr=0.000100
Epoch[089] Batch [3449]/[3760]	Speed: 35.064801 samples/sec	 accuracy=82.542120	 loss=0.698860	 lr=0.000100
Epoch[089] Batch [3499]/[3760]	Speed: 35.248763 samples/sec	 accuracy=82.528125	 loss=0.699072	 lr=0.000100
Epoch[089] Batch [3549]/[3760]	Speed: 35.122044 samples/sec	 accuracy=82.528609	 loss=0.699238	 lr=0.000100
Epoch[089] Batch [3599]/[3760]	Speed: 35.234278 samples/sec	 accuracy=82.536892	 loss=0.698784	 lr=0.000100
Epoch[089] Batch [3649]/[3760]	Speed: 35.430357 samples/sec	 accuracy=82.535103	 loss=0.698875	 lr=0.000100
Epoch[089] Batch [3699]/[3760]	Speed: 35.130847 samples/sec	 accuracy=82.534628	 loss=0.699509	 lr=0.000100
Epoch[089] Batch [3749]/[3760]	Speed: 36.465141 samples/sec	 accuracy=82.531667	 loss=0.699762	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.062500 acc-top5=86.000000
Batch [0099]/[0303]: acc-top1=66.765625 acc-top5=85.421875
Batch [0149]/[0303]: acc-top1=67.145833 acc-top5=85.687500
Batch [0199]/[0303]: acc-top1=66.703125 acc-top5=85.664062
Batch [0249]/[0303]: acc-top1=66.543750 acc-top5=85.568750
Batch [0299]/[0303]: acc-top1=66.505208 acc-top5=85.552083
[Epoch 089] training: accuracy=82.533660	 loss=0.699712
[Epoch 089] speed: 35 samples/sec	time cost: 7318.848115
[Epoch 089] validation: acc-top1=66.506807 acc-top5=85.561056 loss=1.628078
Epoch[090] Batch [0049]/[3759]	Speed: 24.706588 samples/sec	 accuracy=83.000000	 loss=0.686650	 lr=0.000100
Epoch[090] Batch [0099]/[3759]	Speed: 34.988544 samples/sec	 accuracy=82.437500	 loss=0.686040	 lr=0.000100
Epoch[090] Batch [0149]/[3759]	Speed: 35.173949 samples/sec	 accuracy=82.177083	 loss=0.699857	 lr=0.000100
Epoch[090] Batch [0199]/[3759]	Speed: 34.951224 samples/sec	 accuracy=82.296875	 loss=0.699988	 lr=0.000100
Epoch[090] Batch [0249]/[3759]	Speed: 34.917557 samples/sec	 accuracy=82.425000	 loss=0.700498	 lr=0.000100
Epoch[090] Batch [0299]/[3759]	Speed: 35.063714 samples/sec	 accuracy=82.229167	 loss=0.704773	 lr=0.000100
Epoch[090] Batch [0349]/[3759]	Speed: 34.899889 samples/sec	 accuracy=82.236607	 loss=0.704447	 lr=0.000100
Epoch[090] Batch [0399]/[3759]	Speed: 35.311310 samples/sec	 accuracy=82.289062	 loss=0.703892	 lr=0.000100
Epoch[090] Batch [0449]/[3759]	Speed: 34.968925 samples/sec	 accuracy=82.090278	 loss=0.709246	 lr=0.000100
Epoch[090] Batch [0499]/[3759]	Speed: 35.348823 samples/sec	 accuracy=82.146875	 loss=0.707508	 lr=0.000100
Epoch[090] Batch [0549]/[3759]	Speed: 34.729504 samples/sec	 accuracy=82.198864	 loss=0.706386	 lr=0.000100
Epoch[090] Batch [0599]/[3759]	Speed: 35.386195 samples/sec	 accuracy=82.195312	 loss=0.706949	 lr=0.000100
Epoch[090] Batch [0649]/[3759]	Speed: 34.873121 samples/sec	 accuracy=82.259615	 loss=0.704621	 lr=0.000100
Epoch[090] Batch [0699]/[3759]	Speed: 35.146171 samples/sec	 accuracy=82.200893	 loss=0.706769	 lr=0.000100
Epoch[090] Batch [0749]/[3759]	Speed: 34.829186 samples/sec	 accuracy=82.204167	 loss=0.706220	 lr=0.000100
Epoch[090] Batch [0799]/[3759]	Speed: 35.308011 samples/sec	 accuracy=82.138672	 loss=0.707439	 lr=0.000100
Epoch[090] Batch [0849]/[3759]	Speed: 34.924550 samples/sec	 accuracy=82.178309	 loss=0.706508	 lr=0.000100
Epoch[090] Batch [0899]/[3759]	Speed: 35.107249 samples/sec	 accuracy=82.140625	 loss=0.709766	 lr=0.000100
Epoch[090] Batch [0949]/[3759]	Speed: 34.944333 samples/sec	 accuracy=82.164474	 loss=0.707936	 lr=0.000100
Epoch[090] Batch [0999]/[3759]	Speed: 35.251798 samples/sec	 accuracy=82.134375	 loss=0.709080	 lr=0.000100
Epoch[090] Batch [1049]/[3759]	Speed: 35.250605 samples/sec	 accuracy=82.129464	 loss=0.708048	 lr=0.000100
Epoch[090] Batch [1099]/[3759]	Speed: 35.112782 samples/sec	 accuracy=82.125000	 loss=0.708657	 lr=0.000100
Epoch[090] Batch [1149]/[3759]	Speed: 35.091471 samples/sec	 accuracy=82.138587	 loss=0.708322	 lr=0.000100
Epoch[090] Batch [1199]/[3759]	Speed: 34.963228 samples/sec	 accuracy=82.178385	 loss=0.707614	 lr=0.000100
Epoch[090] Batch [1249]/[3759]	Speed: 35.185423 samples/sec	 accuracy=82.272500	 loss=0.704899	 lr=0.000100
Epoch[090] Batch [1299]/[3759]	Speed: 34.869359 samples/sec	 accuracy=82.253606	 loss=0.705484	 lr=0.000100
Epoch[090] Batch [1349]/[3759]	Speed: 35.468297 samples/sec	 accuracy=82.252315	 loss=0.704725	 lr=0.000100
Epoch[090] Batch [1399]/[3759]	Speed: 35.065621 samples/sec	 accuracy=82.253348	 loss=0.704572	 lr=0.000100
Epoch[090] Batch [1449]/[3759]	Speed: 35.319163 samples/sec	 accuracy=82.276940	 loss=0.704889	 lr=0.000100
Epoch[090] Batch [1499]/[3759]	Speed: 34.901240 samples/sec	 accuracy=82.300000	 loss=0.704555	 lr=0.000100
Epoch[090] Batch [1549]/[3759]	Speed: 35.416962 samples/sec	 accuracy=82.305444	 loss=0.704765	 lr=0.000100
Epoch[090] Batch [1599]/[3759]	Speed: 34.904453 samples/sec	 accuracy=82.292969	 loss=0.704745	 lr=0.000100
Epoch[090] Batch [1649]/[3759]	Speed: 35.177878 samples/sec	 accuracy=82.299242	 loss=0.705019	 lr=0.000100
Epoch[090] Batch [1699]/[3759]	Speed: 35.107979 samples/sec	 accuracy=82.316176	 loss=0.705254	 lr=0.000100
Epoch[090] Batch [1749]/[3759]	Speed: 35.174035 samples/sec	 accuracy=82.325893	 loss=0.704946	 lr=0.000100
Epoch[090] Batch [1799]/[3759]	Speed: 34.874010 samples/sec	 accuracy=82.308160	 loss=0.705407	 lr=0.000100
Epoch[090] Batch [1849]/[3759]	Speed: 35.201697 samples/sec	 accuracy=82.293074	 loss=0.706060	 lr=0.000100
Epoch[090] Batch [1899]/[3759]	Speed: 34.659488 samples/sec	 accuracy=82.339638	 loss=0.705051	 lr=0.000100
Epoch[090] Batch [1949]/[3759]	Speed: 35.023483 samples/sec	 accuracy=82.342949	 loss=0.705477	 lr=0.000100
Epoch[090] Batch [1999]/[3759]	Speed: 34.706824 samples/sec	 accuracy=82.373438	 loss=0.704922	 lr=0.000100
Epoch[090] Batch [2049]/[3759]	Speed: 34.829839 samples/sec	 accuracy=82.382622	 loss=0.704280	 lr=0.000100
Epoch[090] Batch [2099]/[3759]	Speed: 34.863865 samples/sec	 accuracy=82.389881	 loss=0.704664	 lr=0.000100
Epoch[090] Batch [2149]/[3759]	Speed: 34.713383 samples/sec	 accuracy=82.404797	 loss=0.703662	 lr=0.000100
Epoch[090] Batch [2199]/[3759]	Speed: 35.103123 samples/sec	 accuracy=82.424716	 loss=0.703320	 lr=0.000100
Epoch[090] Batch [2249]/[3759]	Speed: 34.669687 samples/sec	 accuracy=82.446528	 loss=0.702607	 lr=0.000100
Epoch[090] Batch [2299]/[3759]	Speed: 35.206042 samples/sec	 accuracy=82.455842	 loss=0.702809	 lr=0.000100
Epoch[090] Batch [2349]/[3759]	Speed: 34.476436 samples/sec	 accuracy=82.456782	 loss=0.703071	 lr=0.000100
Epoch[090] Batch [2399]/[3759]	Speed: 35.087726 samples/sec	 accuracy=82.445964	 loss=0.703150	 lr=0.000100
Epoch[090] Batch [2449]/[3759]	Speed: 34.780484 samples/sec	 accuracy=82.476403	 loss=0.702063	 lr=0.000100
Epoch[090] Batch [2499]/[3759]	Speed: 35.123879 samples/sec	 accuracy=82.465625	 loss=0.702284	 lr=0.000100
Epoch[090] Batch [2549]/[3759]	Speed: 34.979995 samples/sec	 accuracy=82.463235	 loss=0.701933	 lr=0.000100
Epoch[090] Batch [2599]/[3759]	Speed: 35.146206 samples/sec	 accuracy=82.492188	 loss=0.700788	 lr=0.000100
Epoch[090] Batch [2649]/[3759]	Speed: 34.922280 samples/sec	 accuracy=82.506486	 loss=0.700323	 lr=0.000100
Epoch[090] Batch [2699]/[3759]	Speed: 35.160291 samples/sec	 accuracy=82.493634	 loss=0.700811	 lr=0.000100
Epoch[090] Batch [2749]/[3759]	Speed: 35.019440 samples/sec	 accuracy=82.495455	 loss=0.700734	 lr=0.000100
Epoch[090] Batch [2799]/[3759]	Speed: 34.998716 samples/sec	 accuracy=82.502790	 loss=0.700530	 lr=0.000100
Epoch[090] Batch [2849]/[3759]	Speed: 35.064010 samples/sec	 accuracy=82.483553	 loss=0.701544	 lr=0.000100
Epoch[090] Batch [2899]/[3759]	Speed: 35.005481 samples/sec	 accuracy=82.482759	 loss=0.701980	 lr=0.000100
Epoch[090] Batch [2949]/[3759]	Speed: 35.114812 samples/sec	 accuracy=82.497352	 loss=0.701610	 lr=0.000100
Epoch[090] Batch [2999]/[3759]	Speed: 34.823097 samples/sec	 accuracy=82.484375	 loss=0.702188	 lr=0.000100
Epoch[090] Batch [3049]/[3759]	Speed: 35.113718 samples/sec	 accuracy=82.471311	 loss=0.702827	 lr=0.000100
Epoch[090] Batch [3099]/[3759]	Speed: 34.689928 samples/sec	 accuracy=82.468750	 loss=0.702977	 lr=0.000100
Epoch[090] Batch [3149]/[3759]	Speed: 35.177006 samples/sec	 accuracy=82.466766	 loss=0.702871	 lr=0.000100
Epoch[090] Batch [3199]/[3759]	Speed: 34.754201 samples/sec	 accuracy=82.481445	 loss=0.702448	 lr=0.000100
Epoch[090] Batch [3249]/[3759]	Speed: 35.283604 samples/sec	 accuracy=82.494231	 loss=0.701973	 lr=0.000100
Epoch[090] Batch [3299]/[3759]	Speed: 34.804293 samples/sec	 accuracy=82.490057	 loss=0.701934	 lr=0.000100
Epoch[090] Batch [3349]/[3759]	Speed: 35.292244 samples/sec	 accuracy=82.511194	 loss=0.701300	 lr=0.000100
Epoch[090] Batch [3399]/[3759]	Speed: 35.000785 samples/sec	 accuracy=82.493107	 loss=0.701804	 lr=0.000100
Epoch[090] Batch [3449]/[3759]	Speed: 35.208373 samples/sec	 accuracy=82.493659	 loss=0.701537	 lr=0.000100
Epoch[090] Batch [3499]/[3759]	Speed: 35.072602 samples/sec	 accuracy=82.490179	 loss=0.701646	 lr=0.000100
Epoch[090] Batch [3549]/[3759]	Speed: 35.074136 samples/sec	 accuracy=82.484155	 loss=0.701913	 lr=0.000100
Epoch[090] Batch [3599]/[3759]	Speed: 35.194953 samples/sec	 accuracy=82.484375	 loss=0.701677	 lr=0.000100
Epoch[090] Batch [3649]/[3759]	Speed: 35.176650 samples/sec	 accuracy=82.493151	 loss=0.701319	 lr=0.000100
Epoch[090] Batch [3699]/[3759]	Speed: 35.136592 samples/sec	 accuracy=82.494088	 loss=0.701356	 lr=0.000100
Epoch[090] Batch [3749]/[3759]	Speed: 36.358719 samples/sec	 accuracy=82.502083	 loss=0.701375	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.375000 acc-top5=86.000000
Batch [0099]/[0303]: acc-top1=66.718750 acc-top5=85.234375
Batch [0149]/[0303]: acc-top1=67.187500 acc-top5=85.531250
Batch [0199]/[0303]: acc-top1=66.812500 acc-top5=85.617188
Batch [0249]/[0303]: acc-top1=66.625000 acc-top5=85.531250
Batch [0299]/[0303]: acc-top1=66.468750 acc-top5=85.578125
[Epoch 090] training: accuracy=82.500333	 loss=0.701215
[Epoch 090] speed: 34 samples/sec	time cost: 7352.593488
[Epoch 090] validation: acc-top1=66.481023 acc-top5=85.586840 loss=1.621881
Epoch[091] Batch [0049]/[3760]	Speed: 25.188099 samples/sec	 accuracy=83.062500	 loss=0.698541	 lr=0.000100
Epoch[091] Batch [0099]/[3760]	Speed: 35.264623 samples/sec	 accuracy=82.609375	 loss=0.709297	 lr=0.000100
Epoch[091] Batch [0149]/[3760]	Speed: 35.209594 samples/sec	 accuracy=82.427083	 loss=0.708521	 lr=0.000100
Epoch[091] Batch [0199]/[3760]	Speed: 35.371931 samples/sec	 accuracy=82.507812	 loss=0.704409	 lr=0.000100
Epoch[091] Batch [0249]/[3760]	Speed: 34.963501 samples/sec	 accuracy=82.531250	 loss=0.702156	 lr=0.000100
Epoch[091] Batch [0299]/[3760]	Speed: 35.221802 samples/sec	 accuracy=82.546875	 loss=0.700936	 lr=0.000100
Epoch[091] Batch [0349]/[3760]	Speed: 35.105254 samples/sec	 accuracy=82.593750	 loss=0.699422	 lr=0.000100
Epoch[091] Batch [0399]/[3760]	Speed: 35.146277 samples/sec	 accuracy=82.593750	 loss=0.703201	 lr=0.000100
Epoch[091] Batch [0449]/[3760]	Speed: 35.030449 samples/sec	 accuracy=82.628472	 loss=0.703279	 lr=0.000100
Epoch[091] Batch [0499]/[3760]	Speed: 35.288513 samples/sec	 accuracy=82.675000	 loss=0.700418	 lr=0.000100
Epoch[091] Batch [0549]/[3760]	Speed: 35.090421 samples/sec	 accuracy=82.616477	 loss=0.702451	 lr=0.000100
Epoch[091] Batch [0599]/[3760]	Speed: 35.259526 samples/sec	 accuracy=82.705729	 loss=0.696813	 lr=0.000100
Epoch[091] Batch [0649]/[3760]	Speed: 35.016270 samples/sec	 accuracy=82.668269	 loss=0.697135	 lr=0.000100
Epoch[091] Batch [0699]/[3760]	Speed: 34.854736 samples/sec	 accuracy=82.651786	 loss=0.697408	 lr=0.000100
Epoch[091] Batch [0749]/[3760]	Speed: 35.038864 samples/sec	 accuracy=82.683333	 loss=0.697571	 lr=0.000100
Epoch[091] Batch [0799]/[3760]	Speed: 34.732742 samples/sec	 accuracy=82.654297	 loss=0.697583	 lr=0.000100
Epoch[091] Batch [0849]/[3760]	Speed: 35.016024 samples/sec	 accuracy=82.726103	 loss=0.695456	 lr=0.000100
Epoch[091] Batch [0899]/[3760]	Speed: 34.427410 samples/sec	 accuracy=82.769097	 loss=0.693567	 lr=0.000100
Epoch[091] Batch [0949]/[3760]	Speed: 35.096109 samples/sec	 accuracy=82.746711	 loss=0.694666	 lr=0.000100
Epoch[091] Batch [0999]/[3760]	Speed: 34.514369 samples/sec	 accuracy=82.742188	 loss=0.695745	 lr=0.000100
Epoch[091] Batch [1049]/[3760]	Speed: 35.201694 samples/sec	 accuracy=82.700893	 loss=0.696285	 lr=0.000100
Epoch[091] Batch [1099]/[3760]	Speed: 34.717174 samples/sec	 accuracy=82.696023	 loss=0.696234	 lr=0.000100
Epoch[091] Batch [1149]/[3760]	Speed: 34.914280 samples/sec	 accuracy=82.683424	 loss=0.696728	 lr=0.000100
Epoch[091] Batch [1199]/[3760]	Speed: 34.706827 samples/sec	 accuracy=82.666667	 loss=0.696885	 lr=0.000100
Epoch[091] Batch [1249]/[3760]	Speed: 34.753448 samples/sec	 accuracy=82.643750	 loss=0.696880	 lr=0.000100
Epoch[091] Batch [1299]/[3760]	Speed: 34.948647 samples/sec	 accuracy=82.635817	 loss=0.696425	 lr=0.000100
Epoch[091] Batch [1349]/[3760]	Speed: 34.659277 samples/sec	 accuracy=82.623843	 loss=0.697182	 lr=0.000100
Epoch[091] Batch [1399]/[3760]	Speed: 35.130319 samples/sec	 accuracy=82.622768	 loss=0.696822	 lr=0.000100
Epoch[091] Batch [1449]/[3760]	Speed: 34.448954 samples/sec	 accuracy=82.665948	 loss=0.695213	 lr=0.000100
Epoch[091] Batch [1499]/[3760]	Speed: 34.996659 samples/sec	 accuracy=82.640625	 loss=0.695914	 lr=0.000100
Epoch[091] Batch [1549]/[3760]	Speed: 34.404087 samples/sec	 accuracy=82.630040	 loss=0.695585	 lr=0.000100
Epoch[091] Batch [1599]/[3760]	Speed: 35.015821 samples/sec	 accuracy=82.617188	 loss=0.695985	 lr=0.000100
Epoch[091] Batch [1649]/[3760]	Speed: 34.651836 samples/sec	 accuracy=82.587121	 loss=0.696202	 lr=0.000100
Epoch[091] Batch [1699]/[3760]	Speed: 34.954585 samples/sec	 accuracy=82.616728	 loss=0.695423	 lr=0.000100
Epoch[091] Batch [1749]/[3760]	Speed: 35.140971 samples/sec	 accuracy=82.643750	 loss=0.694852	 lr=0.000100
Epoch[091] Batch [1799]/[3760]	Speed: 34.838470 samples/sec	 accuracy=82.652778	 loss=0.694986	 lr=0.000100
Epoch[091] Batch [1849]/[3760]	Speed: 34.890743 samples/sec	 accuracy=82.611486	 loss=0.696252	 lr=0.000100
Epoch[091] Batch [1899]/[3760]	Speed: 34.559394 samples/sec	 accuracy=82.600329	 loss=0.696430	 lr=0.000100
Epoch[091] Batch [1949]/[3760]	Speed: 35.215990 samples/sec	 accuracy=82.613782	 loss=0.695944	 lr=0.000100
Epoch[091] Batch [1999]/[3760]	Speed: 34.459753 samples/sec	 accuracy=82.639063	 loss=0.695657	 lr=0.000100
Epoch[091] Batch [2049]/[3760]	Speed: 35.461695 samples/sec	 accuracy=82.642530	 loss=0.695523	 lr=0.000100
Epoch[091] Batch [2099]/[3760]	Speed: 34.653823 samples/sec	 accuracy=82.659226	 loss=0.694998	 lr=0.000100
Epoch[091] Batch [2149]/[3760]	Speed: 35.117692 samples/sec	 accuracy=82.651163	 loss=0.695394	 lr=0.000100
Epoch[091] Batch [2199]/[3760]	Speed: 34.874987 samples/sec	 accuracy=82.655540	 loss=0.695426	 lr=0.000100
Epoch[091] Batch [2249]/[3760]	Speed: 35.059860 samples/sec	 accuracy=82.661806	 loss=0.695516	 lr=0.000100
Epoch[091] Batch [2299]/[3760]	Speed: 34.841821 samples/sec	 accuracy=82.652853	 loss=0.695877	 lr=0.000100
Epoch[091] Batch [2349]/[3760]	Speed: 34.781042 samples/sec	 accuracy=82.640957	 loss=0.695849	 lr=0.000100
Epoch[091] Batch [2399]/[3760]	Speed: 35.058215 samples/sec	 accuracy=82.654297	 loss=0.695554	 lr=0.000100
Epoch[091] Batch [2449]/[3760]	Speed: 34.819182 samples/sec	 accuracy=82.673469	 loss=0.695179	 lr=0.000100
Epoch[091] Batch [2499]/[3760]	Speed: 35.180860 samples/sec	 accuracy=82.671875	 loss=0.694980	 lr=0.000100
Epoch[091] Batch [2549]/[3760]	Speed: 34.573864 samples/sec	 accuracy=82.664828	 loss=0.695637	 lr=0.000100
Epoch[091] Batch [2599]/[3760]	Speed: 35.152378 samples/sec	 accuracy=82.673678	 loss=0.695109	 lr=0.000100
Epoch[091] Batch [2649]/[3760]	Speed: 34.427940 samples/sec	 accuracy=82.676887	 loss=0.694593	 lr=0.000100
Epoch[091] Batch [2699]/[3760]	Speed: 35.155232 samples/sec	 accuracy=82.665509	 loss=0.695002	 lr=0.000100
Epoch[091] Batch [2749]/[3760]	Speed: 34.563011 samples/sec	 accuracy=82.671591	 loss=0.694480	 lr=0.000100
Epoch[091] Batch [2799]/[3760]	Speed: 34.799060 samples/sec	 accuracy=82.667969	 loss=0.694313	 lr=0.000100
Epoch[091] Batch [2849]/[3760]	Speed: 34.547047 samples/sec	 accuracy=82.668311	 loss=0.694052	 lr=0.000100
Epoch[091] Batch [2899]/[3760]	Speed: 34.665702 samples/sec	 accuracy=82.668642	 loss=0.693781	 lr=0.000100
Epoch[091] Batch [2949]/[3760]	Speed: 34.841531 samples/sec	 accuracy=82.658369	 loss=0.693929	 lr=0.000100
Epoch[091] Batch [2999]/[3760]	Speed: 34.465033 samples/sec	 accuracy=82.631771	 loss=0.695107	 lr=0.000100
Epoch[091] Batch [3049]/[3760]	Speed: 35.073342 samples/sec	 accuracy=82.626537	 loss=0.695256	 lr=0.000100
Epoch[091] Batch [3099]/[3760]	Speed: 34.212143 samples/sec	 accuracy=82.626008	 loss=0.695641	 lr=0.000100
Epoch[091] Batch [3149]/[3760]	Speed: 35.208943 samples/sec	 accuracy=82.629464	 loss=0.695621	 lr=0.000100
Epoch[091] Batch [3199]/[3760]	Speed: 34.459034 samples/sec	 accuracy=82.599121	 loss=0.696324	 lr=0.000100
Epoch[091] Batch [3249]/[3760]	Speed: 35.046017 samples/sec	 accuracy=82.604327	 loss=0.696431	 lr=0.000100
Epoch[091] Batch [3299]/[3760]	Speed: 34.616791 samples/sec	 accuracy=82.611742	 loss=0.696017	 lr=0.000100
Epoch[091] Batch [3349]/[3760]	Speed: 34.761070 samples/sec	 accuracy=82.619869	 loss=0.695932	 lr=0.000100
Epoch[091] Batch [3399]/[3760]	Speed: 34.614474 samples/sec	 accuracy=82.624540	 loss=0.695594	 lr=0.000100
Epoch[091] Batch [3449]/[3760]	Speed: 34.461362 samples/sec	 accuracy=82.619565	 loss=0.695780	 lr=0.000100
Epoch[091] Batch [3499]/[3760]	Speed: 34.967789 samples/sec	 accuracy=82.619643	 loss=0.695821	 lr=0.000100
Epoch[091] Batch [3549]/[3760]	Speed: 34.209728 samples/sec	 accuracy=82.607835	 loss=0.696106	 lr=0.000100
Epoch[091] Batch [3599]/[3760]	Speed: 35.273723 samples/sec	 accuracy=82.605035	 loss=0.696168	 lr=0.000100
Epoch[091] Batch [3649]/[3760]	Speed: 34.452597 samples/sec	 accuracy=82.607021	 loss=0.696089	 lr=0.000100
Epoch[091] Batch [3699]/[3760]	Speed: 34.827345 samples/sec	 accuracy=82.614865	 loss=0.695472	 lr=0.000100
Epoch[091] Batch [3749]/[3760]	Speed: 35.961649 samples/sec	 accuracy=82.607500	 loss=0.696143	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.218750 acc-top5=85.906250
Batch [0099]/[0303]: acc-top1=66.625000 acc-top5=85.406250
Batch [0149]/[0303]: acc-top1=67.041667 acc-top5=85.583333
Batch [0199]/[0303]: acc-top1=66.703125 acc-top5=85.671875
Batch [0249]/[0303]: acc-top1=66.543750 acc-top5=85.606250
Batch [0299]/[0303]: acc-top1=66.437500 acc-top5=85.598958
[Epoch 091] training: accuracy=82.610954	 loss=0.696047
[Epoch 091] speed: 34 samples/sec	time cost: 7383.699720
[Epoch 091] validation: acc-top1=66.444926 acc-top5=85.617781 loss=1.634328
Epoch[092] Batch [0049]/[3760]	Speed: 25.044427 samples/sec	 accuracy=82.937500	 loss=0.717612	 lr=0.000100
Epoch[092] Batch [0099]/[3760]	Speed: 35.331727 samples/sec	 accuracy=82.765625	 loss=0.715426	 lr=0.000100
Epoch[092] Batch [0149]/[3760]	Speed: 35.453533 samples/sec	 accuracy=82.614583	 loss=0.706795	 lr=0.000100
Epoch[092] Batch [0199]/[3760]	Speed: 35.361596 samples/sec	 accuracy=82.562500	 loss=0.712191	 lr=0.000100
Epoch[092] Batch [0249]/[3760]	Speed: 35.546777 samples/sec	 accuracy=82.756250	 loss=0.703162	 lr=0.000100
Epoch[092] Batch [0299]/[3760]	Speed: 35.230829 samples/sec	 accuracy=82.687500	 loss=0.703817	 lr=0.000100
Epoch[092] Batch [0349]/[3760]	Speed: 35.217462 samples/sec	 accuracy=82.723214	 loss=0.701764	 lr=0.000100
Epoch[092] Batch [0399]/[3760]	Speed: 35.276357 samples/sec	 accuracy=82.738281	 loss=0.699372	 lr=0.000100
Epoch[092] Batch [0449]/[3760]	Speed: 35.168190 samples/sec	 accuracy=82.680556	 loss=0.699545	 lr=0.000100
Epoch[092] Batch [0499]/[3760]	Speed: 35.145756 samples/sec	 accuracy=82.831250	 loss=0.693998	 lr=0.000100
Epoch[092] Batch [0549]/[3760]	Speed: 35.216261 samples/sec	 accuracy=82.801136	 loss=0.694783	 lr=0.000100
Epoch[092] Batch [0599]/[3760]	Speed: 35.117613 samples/sec	 accuracy=82.802083	 loss=0.696218	 lr=0.000100
Epoch[092] Batch [0649]/[3760]	Speed: 35.006422 samples/sec	 accuracy=82.805288	 loss=0.695202	 lr=0.000100
Epoch[092] Batch [0699]/[3760]	Speed: 35.099455 samples/sec	 accuracy=82.819196	 loss=0.692696	 lr=0.000100
Epoch[092] Batch [0749]/[3760]	Speed: 35.160531 samples/sec	 accuracy=82.789583	 loss=0.693536	 lr=0.000100
Epoch[092] Batch [0799]/[3760]	Speed: 34.822142 samples/sec	 accuracy=82.734375	 loss=0.694715	 lr=0.000100
Epoch[092] Batch [0849]/[3760]	Speed: 35.377410 samples/sec	 accuracy=82.693015	 loss=0.695745	 lr=0.000100
Epoch[092] Batch [0899]/[3760]	Speed: 34.714926 samples/sec	 accuracy=82.715278	 loss=0.694537	 lr=0.000100
Epoch[092] Batch [0949]/[3760]	Speed: 35.356564 samples/sec	 accuracy=82.748355	 loss=0.693147	 lr=0.000100
Epoch[092] Batch [0999]/[3760]	Speed: 34.895797 samples/sec	 accuracy=82.768750	 loss=0.693295	 lr=0.000100
Epoch[092] Batch [1049]/[3760]	Speed: 35.304849 samples/sec	 accuracy=82.808036	 loss=0.692474	 lr=0.000100
Epoch[092] Batch [1099]/[3760]	Speed: 35.141141 samples/sec	 accuracy=82.795455	 loss=0.692708	 lr=0.000100
Epoch[092] Batch [1149]/[3760]	Speed: 35.299346 samples/sec	 accuracy=82.801630	 loss=0.692171	 lr=0.000100
Epoch[092] Batch [1199]/[3760]	Speed: 35.215432 samples/sec	 accuracy=82.813802	 loss=0.691580	 lr=0.000100
Epoch[092] Batch [1249]/[3760]	Speed: 35.328135 samples/sec	 accuracy=82.812500	 loss=0.690806	 lr=0.000100
Epoch[092] Batch [1299]/[3760]	Speed: 35.344075 samples/sec	 accuracy=82.795673	 loss=0.691317	 lr=0.000100
Epoch[092] Batch [1349]/[3760]	Speed: 35.111675 samples/sec	 accuracy=82.785880	 loss=0.691645	 lr=0.000100
Epoch[092] Batch [1399]/[3760]	Speed: 35.230578 samples/sec	 accuracy=82.818080	 loss=0.690885	 lr=0.000100
Epoch[092] Batch [1449]/[3760]	Speed: 34.986702 samples/sec	 accuracy=82.860991	 loss=0.689500	 lr=0.000100
Epoch[092] Batch [1499]/[3760]	Speed: 35.335382 samples/sec	 accuracy=82.871875	 loss=0.689213	 lr=0.000100
Epoch[092] Batch [1549]/[3760]	Speed: 34.678232 samples/sec	 accuracy=82.855847	 loss=0.689849	 lr=0.000100
Epoch[092] Batch [1599]/[3760]	Speed: 35.512641 samples/sec	 accuracy=82.815430	 loss=0.691024	 lr=0.000100
Epoch[092] Batch [1649]/[3760]	Speed: 34.867101 samples/sec	 accuracy=82.820076	 loss=0.690738	 lr=0.000100
Epoch[092] Batch [1699]/[3760]	Speed: 35.089301 samples/sec	 accuracy=82.780331	 loss=0.690941	 lr=0.000100
Epoch[092] Batch [1749]/[3760]	Speed: 35.221350 samples/sec	 accuracy=82.784821	 loss=0.691065	 lr=0.000100
Epoch[092] Batch [1799]/[3760]	Speed: 34.998627 samples/sec	 accuracy=82.756944	 loss=0.691709	 lr=0.000100
Epoch[092] Batch [1849]/[3760]	Speed: 35.343326 samples/sec	 accuracy=82.754223	 loss=0.691540	 lr=0.000100
Epoch[092] Batch [1899]/[3760]	Speed: 34.894766 samples/sec	 accuracy=82.755757	 loss=0.691395	 lr=0.000100
Epoch[092] Batch [1949]/[3760]	Speed: 35.365522 samples/sec	 accuracy=82.750801	 loss=0.691568	 lr=0.000100
Epoch[092] Batch [1999]/[3760]	Speed: 34.886377 samples/sec	 accuracy=82.731250	 loss=0.692434	 lr=0.000100
Epoch[092] Batch [2049]/[3760]	Speed: 35.442185 samples/sec	 accuracy=82.720274	 loss=0.692786	 lr=0.000100
Epoch[092] Batch [2099]/[3760]	Speed: 35.059846 samples/sec	 accuracy=82.700149	 loss=0.693689	 lr=0.000100
Epoch[092] Batch [2149]/[3760]	Speed: 35.241537 samples/sec	 accuracy=82.696948	 loss=0.693503	 lr=0.000100
Epoch[092] Batch [2199]/[3760]	Speed: 35.438264 samples/sec	 accuracy=82.668324	 loss=0.694052	 lr=0.000100
Epoch[092] Batch [2249]/[3760]	Speed: 35.288722 samples/sec	 accuracy=82.639583	 loss=0.694865	 lr=0.000100
Epoch[092] Batch [2299]/[3760]	Speed: 35.333421 samples/sec	 accuracy=82.660326	 loss=0.694291	 lr=0.000100
Epoch[092] Batch [2349]/[3760]	Speed: 35.277331 samples/sec	 accuracy=82.654920	 loss=0.694568	 lr=0.000100
Epoch[092] Batch [2399]/[3760]	Speed: 35.366117 samples/sec	 accuracy=82.667969	 loss=0.694199	 lr=0.000100
Epoch[092] Batch [2449]/[3760]	Speed: 35.530211 samples/sec	 accuracy=82.660077	 loss=0.694379	 lr=0.000100
Epoch[092] Batch [2499]/[3760]	Speed: 35.262733 samples/sec	 accuracy=82.636250	 loss=0.694571	 lr=0.000100
Epoch[092] Batch [2549]/[3760]	Speed: 35.230518 samples/sec	 accuracy=82.614583	 loss=0.695565	 lr=0.000100
Epoch[092] Batch [2599]/[3760]	Speed: 34.961100 samples/sec	 accuracy=82.603365	 loss=0.695781	 lr=0.000100
Epoch[092] Batch [2649]/[3760]	Speed: 35.342182 samples/sec	 accuracy=82.594340	 loss=0.695942	 lr=0.000100
Epoch[092] Batch [2699]/[3760]	Speed: 34.857461 samples/sec	 accuracy=82.607639	 loss=0.695556	 lr=0.000100
Epoch[092] Batch [2749]/[3760]	Speed: 35.179203 samples/sec	 accuracy=82.606818	 loss=0.696002	 lr=0.000100
Epoch[092] Batch [2799]/[3760]	Speed: 34.967531 samples/sec	 accuracy=82.595424	 loss=0.696510	 lr=0.000100
Epoch[092] Batch [2849]/[3760]	Speed: 34.725648 samples/sec	 accuracy=82.608004	 loss=0.696049	 lr=0.000100
Epoch[092] Batch [2899]/[3760]	Speed: 35.168331 samples/sec	 accuracy=82.599677	 loss=0.696231	 lr=0.000100
Epoch[092] Batch [2949]/[3760]	Speed: 34.706995 samples/sec	 accuracy=82.605932	 loss=0.696125	 lr=0.000100
Epoch[092] Batch [2999]/[3760]	Speed: 35.195447 samples/sec	 accuracy=82.597396	 loss=0.696553	 lr=0.000100
Epoch[092] Batch [3049]/[3760]	Speed: 34.700372 samples/sec	 accuracy=82.594262	 loss=0.696792	 lr=0.000100
Epoch[092] Batch [3099]/[3760]	Speed: 35.312372 samples/sec	 accuracy=82.588206	 loss=0.697208	 lr=0.000100
Epoch[092] Batch [3149]/[3760]	Speed: 34.600140 samples/sec	 accuracy=82.595238	 loss=0.697219	 lr=0.000100
Epoch[092] Batch [3199]/[3760]	Speed: 35.140066 samples/sec	 accuracy=82.588379	 loss=0.697614	 lr=0.000100
Epoch[092] Batch [3249]/[3760]	Speed: 34.651192 samples/sec	 accuracy=82.594231	 loss=0.697153	 lr=0.000100
Epoch[092] Batch [3299]/[3760]	Speed: 34.992045 samples/sec	 accuracy=82.589489	 loss=0.697412	 lr=0.000100
Epoch[092] Batch [3349]/[3760]	Speed: 34.912218 samples/sec	 accuracy=82.589086	 loss=0.697365	 lr=0.000100
Epoch[092] Batch [3399]/[3760]	Speed: 34.861137 samples/sec	 accuracy=82.594669	 loss=0.696741	 lr=0.000100
Epoch[092] Batch [3449]/[3760]	Speed: 35.090035 samples/sec	 accuracy=82.588768	 loss=0.697058	 lr=0.000100
Epoch[092] Batch [3499]/[3760]	Speed: 34.469531 samples/sec	 accuracy=82.579464	 loss=0.697133	 lr=0.000100
Epoch[092] Batch [3549]/[3760]	Speed: 35.174926 samples/sec	 accuracy=82.589789	 loss=0.696990	 lr=0.000100
Epoch[092] Batch [3599]/[3760]	Speed: 34.299292 samples/sec	 accuracy=82.592448	 loss=0.696754	 lr=0.000100
Epoch[092] Batch [3649]/[3760]	Speed: 35.057314 samples/sec	 accuracy=82.583048	 loss=0.696922	 lr=0.000100
Epoch[092] Batch [3699]/[3760]	Speed: 34.341154 samples/sec	 accuracy=82.581081	 loss=0.696895	 lr=0.000100
Epoch[092] Batch [3749]/[3760]	Speed: 36.136684 samples/sec	 accuracy=82.583333	 loss=0.697014	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.031250 acc-top5=86.437500
Batch [0099]/[0303]: acc-top1=66.562500 acc-top5=85.562500
Batch [0149]/[0303]: acc-top1=66.989583 acc-top5=85.635417
Batch [0199]/[0303]: acc-top1=66.742188 acc-top5=85.656250
Batch [0249]/[0303]: acc-top1=66.562500 acc-top5=85.512500
Batch [0299]/[0303]: acc-top1=66.473958 acc-top5=85.515625
[Epoch 092] training: accuracy=82.592670	 loss=0.696754
[Epoch 092] speed: 34 samples/sec	time cost: 7336.764330
[Epoch 092] validation: acc-top1=66.465553 acc-top5=85.519802 loss=1.632878
Epoch[093] Batch [0049]/[3759]	Speed: 25.356503 samples/sec	 accuracy=82.718750	 loss=0.702983	 lr=0.000100
Epoch[093] Batch [0099]/[3759]	Speed: 35.079204 samples/sec	 accuracy=82.312500	 loss=0.709950	 lr=0.000100
Epoch[093] Batch [0149]/[3759]	Speed: 35.421907 samples/sec	 accuracy=82.531250	 loss=0.701556	 lr=0.000100
Epoch[093] Batch [0199]/[3759]	Speed: 35.136205 samples/sec	 accuracy=82.617188	 loss=0.699378	 lr=0.000100
Epoch[093] Batch [0249]/[3759]	Speed: 35.230778 samples/sec	 accuracy=82.643750	 loss=0.698894	 lr=0.000100
Epoch[093] Batch [0299]/[3759]	Speed: 35.000628 samples/sec	 accuracy=82.750000	 loss=0.694588	 lr=0.000100
Epoch[093] Batch [0349]/[3759]	Speed: 35.088175 samples/sec	 accuracy=82.772321	 loss=0.692345	 lr=0.000100
Epoch[093] Batch [0399]/[3759]	Speed: 35.198391 samples/sec	 accuracy=82.679688	 loss=0.693280	 lr=0.000100
Epoch[093] Batch [0449]/[3759]	Speed: 34.931440 samples/sec	 accuracy=82.687500	 loss=0.692836	 lr=0.000100
Epoch[093] Batch [0499]/[3759]	Speed: 35.336298 samples/sec	 accuracy=82.628125	 loss=0.691873	 lr=0.000100
Epoch[093] Batch [0549]/[3759]	Speed: 34.788629 samples/sec	 accuracy=82.670455	 loss=0.688626	 lr=0.000100
Epoch[093] Batch [0599]/[3759]	Speed: 35.265652 samples/sec	 accuracy=82.617188	 loss=0.692223	 lr=0.000100
Epoch[093] Batch [0649]/[3759]	Speed: 34.734425 samples/sec	 accuracy=82.622596	 loss=0.692978	 lr=0.000100
Epoch[093] Batch [0699]/[3759]	Speed: 35.343187 samples/sec	 accuracy=82.680804	 loss=0.690319	 lr=0.000100
Epoch[093] Batch [0749]/[3759]	Speed: 34.808927 samples/sec	 accuracy=82.818750	 loss=0.686535	 lr=0.000100
Epoch[093] Batch [0799]/[3759]	Speed: 35.213522 samples/sec	 accuracy=82.794922	 loss=0.686188	 lr=0.000100
Epoch[093] Batch [0849]/[3759]	Speed: 35.217284 samples/sec	 accuracy=82.836397	 loss=0.684169	 lr=0.000100
Epoch[093] Batch [0899]/[3759]	Speed: 35.124888 samples/sec	 accuracy=82.781250	 loss=0.686235	 lr=0.000100
Epoch[093] Batch [0949]/[3759]	Speed: 35.163241 samples/sec	 accuracy=82.774671	 loss=0.687447	 lr=0.000100
Epoch[093] Batch [0999]/[3759]	Speed: 35.430565 samples/sec	 accuracy=82.804688	 loss=0.686852	 lr=0.000100
Epoch[093] Batch [1049]/[3759]	Speed: 34.880121 samples/sec	 accuracy=82.828869	 loss=0.685857	 lr=0.000100
Epoch[093] Batch [1099]/[3759]	Speed: 35.260537 samples/sec	 accuracy=82.845170	 loss=0.684756	 lr=0.000100
Epoch[093] Batch [1149]/[3759]	Speed: 35.191293 samples/sec	 accuracy=82.858696	 loss=0.684536	 lr=0.000100
Epoch[093] Batch [1199]/[3759]	Speed: 35.284484 samples/sec	 accuracy=82.855469	 loss=0.686270	 lr=0.000100
Epoch[093] Batch [1249]/[3759]	Speed: 35.201441 samples/sec	 accuracy=82.840000	 loss=0.687978	 lr=0.000100
Epoch[093] Batch [1299]/[3759]	Speed: 35.223758 samples/sec	 accuracy=82.860577	 loss=0.687545	 lr=0.000100
Epoch[093] Batch [1349]/[3759]	Speed: 35.080002 samples/sec	 accuracy=82.839120	 loss=0.688316	 lr=0.000100
Epoch[093] Batch [1399]/[3759]	Speed: 35.094608 samples/sec	 accuracy=82.883929	 loss=0.686572	 lr=0.000100
Epoch[093] Batch [1449]/[3759]	Speed: 35.116895 samples/sec	 accuracy=82.870690	 loss=0.686973	 lr=0.000100
Epoch[093] Batch [1499]/[3759]	Speed: 35.065204 samples/sec	 accuracy=82.854167	 loss=0.687022	 lr=0.000100
Epoch[093] Batch [1549]/[3759]	Speed: 35.096788 samples/sec	 accuracy=82.892137	 loss=0.686388	 lr=0.000100
Epoch[093] Batch [1599]/[3759]	Speed: 35.157123 samples/sec	 accuracy=82.907227	 loss=0.685428	 lr=0.000100
Epoch[093] Batch [1649]/[3759]	Speed: 35.393609 samples/sec	 accuracy=82.900568	 loss=0.685087	 lr=0.000100
Epoch[093] Batch [1699]/[3759]	Speed: 35.227997 samples/sec	 accuracy=82.890625	 loss=0.685890	 lr=0.000100
Epoch[093] Batch [1749]/[3759]	Speed: 35.359881 samples/sec	 accuracy=82.882143	 loss=0.685579	 lr=0.000100
Epoch[093] Batch [1799]/[3759]	Speed: 34.961797 samples/sec	 accuracy=82.908854	 loss=0.685107	 lr=0.000100
Epoch[093] Batch [1849]/[3759]	Speed: 35.340844 samples/sec	 accuracy=82.934122	 loss=0.684601	 lr=0.000100
Epoch[093] Batch [1899]/[3759]	Speed: 34.945289 samples/sec	 accuracy=82.919408	 loss=0.684522	 lr=0.000100
Epoch[093] Batch [1949]/[3759]	Speed: 35.459142 samples/sec	 accuracy=82.935897	 loss=0.683698	 lr=0.000100
Epoch[093] Batch [1999]/[3759]	Speed: 35.042009 samples/sec	 accuracy=82.917969	 loss=0.684147	 lr=0.000100
Epoch[093] Batch [2049]/[3759]	Speed: 35.361199 samples/sec	 accuracy=82.929878	 loss=0.683507	 lr=0.000100
Epoch[093] Batch [2099]/[3759]	Speed: 35.174477 samples/sec	 accuracy=82.929315	 loss=0.683424	 lr=0.000100
Epoch[093] Batch [2149]/[3759]	Speed: 35.296856 samples/sec	 accuracy=82.914244	 loss=0.684175	 lr=0.000100
Epoch[093] Batch [2199]/[3759]	Speed: 35.054595 samples/sec	 accuracy=82.895597	 loss=0.685006	 lr=0.000100
Epoch[093] Batch [2249]/[3759]	Speed: 35.130046 samples/sec	 accuracy=82.887500	 loss=0.685223	 lr=0.000100
Epoch[093] Batch [2299]/[3759]	Speed: 34.939605 samples/sec	 accuracy=82.900136	 loss=0.685256	 lr=0.000100
Epoch[093] Batch [2349]/[3759]	Speed: 35.138676 samples/sec	 accuracy=82.925532	 loss=0.684984	 lr=0.000100
Epoch[093] Batch [2399]/[3759]	Speed: 35.188402 samples/sec	 accuracy=82.894531	 loss=0.685640	 lr=0.000100
Epoch[093] Batch [2449]/[3759]	Speed: 35.023578 samples/sec	 accuracy=82.873087	 loss=0.685936	 lr=0.000100
Epoch[093] Batch [2499]/[3759]	Speed: 35.280790 samples/sec	 accuracy=82.870625	 loss=0.686011	 lr=0.000100
Epoch[093] Batch [2549]/[3759]	Speed: 35.157153 samples/sec	 accuracy=82.866422	 loss=0.685856	 lr=0.000100
Epoch[093] Batch [2599]/[3759]	Speed: 35.415316 samples/sec	 accuracy=82.866587	 loss=0.686219	 lr=0.000100
Epoch[093] Batch [2649]/[3759]	Speed: 35.266308 samples/sec	 accuracy=82.871462	 loss=0.685830	 lr=0.000100
Epoch[093] Batch [2699]/[3759]	Speed: 35.267667 samples/sec	 accuracy=82.839699	 loss=0.686686	 lr=0.000100
Epoch[093] Batch [2749]/[3759]	Speed: 35.393608 samples/sec	 accuracy=82.835227	 loss=0.687082	 lr=0.000100
Epoch[093] Batch [2799]/[3759]	Speed: 35.198364 samples/sec	 accuracy=82.825893	 loss=0.687298	 lr=0.000100
Epoch[093] Batch [2849]/[3759]	Speed: 35.386927 samples/sec	 accuracy=82.817982	 loss=0.687532	 lr=0.000100
Epoch[093] Batch [2899]/[3759]	Speed: 35.248912 samples/sec	 accuracy=82.825970	 loss=0.687274	 lr=0.000100
Epoch[093] Batch [2949]/[3759]	Speed: 35.057663 samples/sec	 accuracy=82.805085	 loss=0.687824	 lr=0.000100
Epoch[093] Batch [2999]/[3759]	Speed: 35.309279 samples/sec	 accuracy=82.831771	 loss=0.687360	 lr=0.000100
Epoch[093] Batch [3049]/[3759]	Speed: 35.077095 samples/sec	 accuracy=82.818648	 loss=0.687605	 lr=0.000100
Epoch[093] Batch [3099]/[3759]	Speed: 35.306414 samples/sec	 accuracy=82.829133	 loss=0.686854	 lr=0.000100
Epoch[093] Batch [3149]/[3759]	Speed: 35.028731 samples/sec	 accuracy=82.815972	 loss=0.687677	 lr=0.000100
Epoch[093] Batch [3199]/[3759]	Speed: 35.407043 samples/sec	 accuracy=82.829102	 loss=0.687408	 lr=0.000100
Epoch[093] Batch [3249]/[3759]	Speed: 34.996884 samples/sec	 accuracy=82.830288	 loss=0.687636	 lr=0.000100
Epoch[093] Batch [3299]/[3759]	Speed: 35.462293 samples/sec	 accuracy=82.808712	 loss=0.687928	 lr=0.000100
Epoch[093] Batch [3349]/[3759]	Speed: 34.939963 samples/sec	 accuracy=82.811567	 loss=0.687750	 lr=0.000100
Epoch[093] Batch [3399]/[3759]	Speed: 35.445154 samples/sec	 accuracy=82.800092	 loss=0.688119	 lr=0.000100
Epoch[093] Batch [3449]/[3759]	Speed: 34.887257 samples/sec	 accuracy=82.805254	 loss=0.688083	 lr=0.000100
Epoch[093] Batch [3499]/[3759]	Speed: 35.405385 samples/sec	 accuracy=82.807143	 loss=0.687930	 lr=0.000100
Epoch[093] Batch [3549]/[3759]	Speed: 35.115339 samples/sec	 accuracy=82.798856	 loss=0.688071	 lr=0.000100
Epoch[093] Batch [3599]/[3759]	Speed: 35.449419 samples/sec	 accuracy=82.790799	 loss=0.688379	 lr=0.000100
Epoch[093] Batch [3649]/[3759]	Speed: 35.035575 samples/sec	 accuracy=82.778682	 loss=0.688801	 lr=0.000100
Epoch[093] Batch [3699]/[3759]	Speed: 35.379087 samples/sec	 accuracy=82.775760	 loss=0.688763	 lr=0.000100
Epoch[093] Batch [3749]/[3759]	Speed: 36.343633 samples/sec	 accuracy=82.782500	 loss=0.688427	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.250000 acc-top5=86.093750
Batch [0099]/[0303]: acc-top1=66.734375 acc-top5=85.359375
Batch [0149]/[0303]: acc-top1=67.218750 acc-top5=85.572917
Batch [0199]/[0303]: acc-top1=66.812500 acc-top5=85.648438
Batch [0249]/[0303]: acc-top1=66.556250 acc-top5=85.531250
Batch [0299]/[0303]: acc-top1=66.432292 acc-top5=85.505208
[Epoch 093] training: accuracy=82.779662	 loss=0.688496
[Epoch 093] speed: 35 samples/sec	time cost: 7321.284561
[Epoch 093] validation: acc-top1=66.434612 acc-top5=85.514645 loss=1.626500
Epoch[094] Batch [0049]/[3760]	Speed: 25.074986 samples/sec	 accuracy=83.281250	 loss=0.658579	 lr=0.000100
Epoch[094] Batch [0099]/[3760]	Speed: 35.292253 samples/sec	 accuracy=83.093750	 loss=0.663729	 lr=0.000100
Epoch[094] Batch [0149]/[3760]	Speed: 35.376777 samples/sec	 accuracy=82.906250	 loss=0.667539	 lr=0.000100
Epoch[094] Batch [0199]/[3760]	Speed: 35.085523 samples/sec	 accuracy=82.585938	 loss=0.689769	 lr=0.000100
Epoch[094] Batch [0249]/[3760]	Speed: 35.365092 samples/sec	 accuracy=82.768750	 loss=0.681744	 lr=0.000100
Epoch[094] Batch [0299]/[3760]	Speed: 34.854100 samples/sec	 accuracy=82.645833	 loss=0.689189	 lr=0.000100
Epoch[094] Batch [0349]/[3760]	Speed: 35.322762 samples/sec	 accuracy=82.464286	 loss=0.695155	 lr=0.000100
Epoch[094] Batch [0399]/[3760]	Speed: 35.102232 samples/sec	 accuracy=82.585938	 loss=0.690334	 lr=0.000100
Epoch[094] Batch [0449]/[3760]	Speed: 35.036393 samples/sec	 accuracy=82.725694	 loss=0.685447	 lr=0.000100
Epoch[094] Batch [0499]/[3760]	Speed: 35.171362 samples/sec	 accuracy=82.700000	 loss=0.686728	 lr=0.000100
Epoch[094] Batch [0549]/[3760]	Speed: 35.132389 samples/sec	 accuracy=82.769886	 loss=0.687129	 lr=0.000100
Epoch[094] Batch [0599]/[3760]	Speed: 35.197696 samples/sec	 accuracy=82.768229	 loss=0.687257	 lr=0.000100
Epoch[094] Batch [0649]/[3760]	Speed: 34.955077 samples/sec	 accuracy=82.750000	 loss=0.687183	 lr=0.000100
Epoch[094] Batch [0699]/[3760]	Speed: 35.441322 samples/sec	 accuracy=82.808036	 loss=0.684380	 lr=0.000100
Epoch[094] Batch [0749]/[3760]	Speed: 34.963724 samples/sec	 accuracy=82.768750	 loss=0.685870	 lr=0.000100
Epoch[094] Batch [0799]/[3760]	Speed: 35.495406 samples/sec	 accuracy=82.746094	 loss=0.686960	 lr=0.000100
Epoch[094] Batch [0849]/[3760]	Speed: 34.972445 samples/sec	 accuracy=82.704044	 loss=0.687690	 lr=0.000100
Epoch[094] Batch [0899]/[3760]	Speed: 35.313655 samples/sec	 accuracy=82.661458	 loss=0.688840	 lr=0.000100
Epoch[094] Batch [0949]/[3760]	Speed: 35.127831 samples/sec	 accuracy=82.649671	 loss=0.690068	 lr=0.000100
Epoch[094] Batch [0999]/[3760]	Speed: 35.405338 samples/sec	 accuracy=82.645313	 loss=0.689709	 lr=0.000100
Epoch[094] Batch [1049]/[3760]	Speed: 35.103944 samples/sec	 accuracy=82.641369	 loss=0.688868	 lr=0.000100
Epoch[094] Batch [1099]/[3760]	Speed: 35.318580 samples/sec	 accuracy=82.686080	 loss=0.686922	 lr=0.000100
Epoch[094] Batch [1149]/[3760]	Speed: 34.975252 samples/sec	 accuracy=82.733696	 loss=0.685334	 lr=0.000100
Epoch[094] Batch [1199]/[3760]	Speed: 35.355371 samples/sec	 accuracy=82.731771	 loss=0.684610	 lr=0.000100
Epoch[094] Batch [1249]/[3760]	Speed: 35.126378 samples/sec	 accuracy=82.715000	 loss=0.686093	 lr=0.000100
Epoch[094] Batch [1299]/[3760]	Speed: 35.117272 samples/sec	 accuracy=82.736779	 loss=0.686192	 lr=0.000100
Epoch[094] Batch [1349]/[3760]	Speed: 35.086488 samples/sec	 accuracy=82.734954	 loss=0.686053	 lr=0.000100
Epoch[094] Batch [1399]/[3760]	Speed: 34.922098 samples/sec	 accuracy=82.707589	 loss=0.687145	 lr=0.000100
Epoch[094] Batch [1449]/[3760]	Speed: 35.227070 samples/sec	 accuracy=82.701509	 loss=0.688843	 lr=0.000100
Epoch[094] Batch [1499]/[3760]	Speed: 34.822307 samples/sec	 accuracy=82.694792	 loss=0.688767	 lr=0.000100
Epoch[094] Batch [1549]/[3760]	Speed: 35.397837 samples/sec	 accuracy=82.711694	 loss=0.688489	 lr=0.000100
Epoch[094] Batch [1599]/[3760]	Speed: 34.924312 samples/sec	 accuracy=82.731445	 loss=0.687479	 lr=0.000100
Epoch[094] Batch [1649]/[3760]	Speed: 35.363629 samples/sec	 accuracy=82.746212	 loss=0.687326	 lr=0.000100
Epoch[094] Batch [1699]/[3760]	Speed: 34.997788 samples/sec	 accuracy=82.729779	 loss=0.688199	 lr=0.000100
Epoch[094] Batch [1749]/[3760]	Speed: 35.476398 samples/sec	 accuracy=82.699107	 loss=0.689169	 lr=0.000100
Epoch[094] Batch [1799]/[3760]	Speed: 34.930930 samples/sec	 accuracy=82.687500	 loss=0.689928	 lr=0.000100
Epoch[094] Batch [1849]/[3760]	Speed: 35.335356 samples/sec	 accuracy=82.692568	 loss=0.690064	 lr=0.000100
Epoch[094] Batch [1899]/[3760]	Speed: 35.312663 samples/sec	 accuracy=82.689967	 loss=0.690174	 lr=0.000100
Epoch[094] Batch [1949]/[3760]	Speed: 35.235763 samples/sec	 accuracy=82.666667	 loss=0.690670	 lr=0.000100
Epoch[094] Batch [1999]/[3760]	Speed: 35.097020 samples/sec	 accuracy=82.673438	 loss=0.690907	 lr=0.000100
Epoch[094] Batch [2049]/[3760]	Speed: 35.179169 samples/sec	 accuracy=82.689024	 loss=0.690354	 lr=0.000100
Epoch[094] Batch [2099]/[3760]	Speed: 35.414623 samples/sec	 accuracy=82.705357	 loss=0.689631	 lr=0.000100
Epoch[094] Batch [2149]/[3760]	Speed: 35.049993 samples/sec	 accuracy=82.681686	 loss=0.690108	 lr=0.000100
Epoch[094] Batch [2199]/[3760]	Speed: 35.395131 samples/sec	 accuracy=82.688210	 loss=0.690141	 lr=0.000100
Epoch[094] Batch [2249]/[3760]	Speed: 35.128114 samples/sec	 accuracy=82.709722	 loss=0.689816	 lr=0.000100
Epoch[094] Batch [2299]/[3760]	Speed: 35.350956 samples/sec	 accuracy=82.694293	 loss=0.690679	 lr=0.000100
Epoch[094] Batch [2349]/[3760]	Speed: 34.840793 samples/sec	 accuracy=82.678856	 loss=0.691396	 lr=0.000100
Epoch[094] Batch [2399]/[3760]	Speed: 35.324855 samples/sec	 accuracy=82.661458	 loss=0.691807	 lr=0.000100
Epoch[094] Batch [2449]/[3760]	Speed: 35.013224 samples/sec	 accuracy=82.667730	 loss=0.691836	 lr=0.000100
Epoch[094] Batch [2499]/[3760]	Speed: 35.343520 samples/sec	 accuracy=82.657500	 loss=0.692139	 lr=0.000100
Epoch[094] Batch [2549]/[3760]	Speed: 35.096603 samples/sec	 accuracy=82.656250	 loss=0.692317	 lr=0.000100
Epoch[094] Batch [2599]/[3760]	Speed: 35.203719 samples/sec	 accuracy=82.665865	 loss=0.692229	 lr=0.000100
Epoch[094] Batch [2649]/[3760]	Speed: 34.981903 samples/sec	 accuracy=82.662736	 loss=0.692439	 lr=0.000100
Epoch[094] Batch [2699]/[3760]	Speed: 35.210512 samples/sec	 accuracy=82.675347	 loss=0.692606	 lr=0.000100
Epoch[094] Batch [2749]/[3760]	Speed: 35.208291 samples/sec	 accuracy=82.671023	 loss=0.692818	 lr=0.000100
Epoch[094] Batch [2799]/[3760]	Speed: 35.094325 samples/sec	 accuracy=82.675781	 loss=0.692997	 lr=0.000100
Epoch[094] Batch [2849]/[3760]	Speed: 35.348291 samples/sec	 accuracy=82.685855	 loss=0.692923	 lr=0.000100
Epoch[094] Batch [2899]/[3760]	Speed: 35.184222 samples/sec	 accuracy=82.678879	 loss=0.693062	 lr=0.000100
Epoch[094] Batch [2949]/[3760]	Speed: 35.318634 samples/sec	 accuracy=82.689089	 loss=0.692715	 lr=0.000100
Epoch[094] Batch [2999]/[3760]	Speed: 35.239145 samples/sec	 accuracy=82.690104	 loss=0.692306	 lr=0.000100
Epoch[094] Batch [3049]/[3760]	Speed: 35.321638 samples/sec	 accuracy=82.691086	 loss=0.692443	 lr=0.000100
Epoch[094] Batch [3099]/[3760]	Speed: 35.283740 samples/sec	 accuracy=82.681956	 loss=0.692606	 lr=0.000100
Epoch[094] Batch [3149]/[3760]	Speed: 35.525256 samples/sec	 accuracy=82.688988	 loss=0.692280	 lr=0.000100
Epoch[094] Batch [3199]/[3760]	Speed: 35.191857 samples/sec	 accuracy=82.691895	 loss=0.692295	 lr=0.000100
Epoch[094] Batch [3249]/[3760]	Speed: 35.438929 samples/sec	 accuracy=82.702404	 loss=0.691771	 lr=0.000100
Epoch[094] Batch [3299]/[3760]	Speed: 35.185014 samples/sec	 accuracy=82.681818	 loss=0.692607	 lr=0.000100
Epoch[094] Batch [3349]/[3760]	Speed: 35.411378 samples/sec	 accuracy=82.684701	 loss=0.692741	 lr=0.000100
Epoch[094] Batch [3399]/[3760]	Speed: 35.178802 samples/sec	 accuracy=82.675551	 loss=0.693030	 lr=0.000100
Epoch[094] Batch [3449]/[3760]	Speed: 35.258952 samples/sec	 accuracy=82.677536	 loss=0.692715	 lr=0.000100
Epoch[094] Batch [3499]/[3760]	Speed: 35.161664 samples/sec	 accuracy=82.670089	 loss=0.692791	 lr=0.000100
Epoch[094] Batch [3549]/[3760]	Speed: 35.189718 samples/sec	 accuracy=82.669014	 loss=0.692679	 lr=0.000100
Epoch[094] Batch [3599]/[3760]	Speed: 35.370816 samples/sec	 accuracy=82.677517	 loss=0.692566	 lr=0.000100
Epoch[094] Batch [3649]/[3760]	Speed: 35.270115 samples/sec	 accuracy=82.657106	 loss=0.693384	 lr=0.000100
Epoch[094] Batch [3699]/[3760]	Speed: 35.479196 samples/sec	 accuracy=82.652027	 loss=0.693551	 lr=0.000100
Epoch[094] Batch [3749]/[3760]	Speed: 36.405723 samples/sec	 accuracy=82.663750	 loss=0.693288	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.031250 acc-top5=86.125000
Batch [0099]/[0303]: acc-top1=66.656250 acc-top5=85.343750
Batch [0149]/[0303]: acc-top1=67.104167 acc-top5=85.583333
Batch [0199]/[0303]: acc-top1=66.726562 acc-top5=85.664062
Batch [0249]/[0303]: acc-top1=66.450000 acc-top5=85.568750
Batch [0299]/[0303]: acc-top1=66.291667 acc-top5=85.630208
[Epoch 094] training: accuracy=82.661237	 loss=0.693394
[Epoch 094] speed: 35 samples/sec	time cost: 7317.363612
[Epoch 094] validation: acc-top1=66.300536 acc-top5=85.643564 loss=1.635985
Epoch[095] Batch [0049]/[3760]	Speed: 25.344295 samples/sec	 accuracy=83.437500	 loss=0.646433	 lr=0.000100
Epoch[095] Batch [0099]/[3760]	Speed: 35.132710 samples/sec	 accuracy=83.515625	 loss=0.648671	 lr=0.000100
Epoch[095] Batch [0149]/[3760]	Speed: 35.441348 samples/sec	 accuracy=82.989583	 loss=0.666010	 lr=0.000100
Epoch[095] Batch [0199]/[3760]	Speed: 35.192708 samples/sec	 accuracy=83.125000	 loss=0.662381	 lr=0.000100
Epoch[095] Batch [0249]/[3760]	Speed: 35.538344 samples/sec	 accuracy=82.987500	 loss=0.669864	 lr=0.000100
Epoch[095] Batch [0299]/[3760]	Speed: 35.048769 samples/sec	 accuracy=83.015625	 loss=0.671428	 lr=0.000100
Epoch[095] Batch [0349]/[3760]	Speed: 35.416019 samples/sec	 accuracy=83.120536	 loss=0.668196	 lr=0.000100
Epoch[095] Batch [0399]/[3760]	Speed: 35.206529 samples/sec	 accuracy=83.101562	 loss=0.670810	 lr=0.000100
Epoch[095] Batch [0449]/[3760]	Speed: 35.406809 samples/sec	 accuracy=83.104167	 loss=0.670940	 lr=0.000100
Epoch[095] Batch [0499]/[3760]	Speed: 35.521555 samples/sec	 accuracy=82.971875	 loss=0.678579	 lr=0.000100
Epoch[095] Batch [0549]/[3760]	Speed: 35.273964 samples/sec	 accuracy=82.951705	 loss=0.679878	 lr=0.000100
Epoch[095] Batch [0599]/[3760]	Speed: 35.329221 samples/sec	 accuracy=82.908854	 loss=0.680806	 lr=0.000100
Epoch[095] Batch [0649]/[3760]	Speed: 35.372641 samples/sec	 accuracy=82.884615	 loss=0.682093	 lr=0.000100
Epoch[095] Batch [0699]/[3760]	Speed: 35.430231 samples/sec	 accuracy=82.908482	 loss=0.681313	 lr=0.000100
Epoch[095] Batch [0749]/[3760]	Speed: 35.308979 samples/sec	 accuracy=82.812500	 loss=0.684370	 lr=0.000100
Epoch[095] Batch [0799]/[3760]	Speed: 35.319079 samples/sec	 accuracy=82.808594	 loss=0.686154	 lr=0.000100
Epoch[095] Batch [0849]/[3760]	Speed: 35.389545 samples/sec	 accuracy=82.836397	 loss=0.685176	 lr=0.000100
Epoch[095] Batch [0899]/[3760]	Speed: 35.480898 samples/sec	 accuracy=82.789931	 loss=0.687051	 lr=0.000100
Epoch[095] Batch [0949]/[3760]	Speed: 35.464021 samples/sec	 accuracy=82.810855	 loss=0.687238	 lr=0.000100
Epoch[095] Batch [0999]/[3760]	Speed: 35.173362 samples/sec	 accuracy=82.848438	 loss=0.685307	 lr=0.000100
Epoch[095] Batch [1049]/[3760]	Speed: 35.197083 samples/sec	 accuracy=82.891369	 loss=0.684435	 lr=0.000100
Epoch[095] Batch [1099]/[3760]	Speed: 35.204632 samples/sec	 accuracy=82.869318	 loss=0.685917	 lr=0.000100
Epoch[095] Batch [1149]/[3760]	Speed: 35.172789 samples/sec	 accuracy=82.883152	 loss=0.685420	 lr=0.000100
Epoch[095] Batch [1199]/[3760]	Speed: 35.155187 samples/sec	 accuracy=82.864583	 loss=0.685340	 lr=0.000100
Epoch[095] Batch [1249]/[3760]	Speed: 35.002724 samples/sec	 accuracy=82.841250	 loss=0.685505	 lr=0.000100
Epoch[095] Batch [1299]/[3760]	Speed: 35.402298 samples/sec	 accuracy=82.799279	 loss=0.687417	 lr=0.000100
Epoch[095] Batch [1349]/[3760]	Speed: 35.037317 samples/sec	 accuracy=82.767361	 loss=0.688780	 lr=0.000100
Epoch[095] Batch [1399]/[3760]	Speed: 35.532773 samples/sec	 accuracy=82.794643	 loss=0.687485	 lr=0.000100
Epoch[095] Batch [1449]/[3760]	Speed: 35.083564 samples/sec	 accuracy=82.775862	 loss=0.688315	 lr=0.000100
Epoch[095] Batch [1499]/[3760]	Speed: 35.381471 samples/sec	 accuracy=82.800000	 loss=0.687651	 lr=0.000100
Epoch[095] Batch [1549]/[3760]	Speed: 35.239882 samples/sec	 accuracy=82.806452	 loss=0.686688	 lr=0.000100
Epoch[095] Batch [1599]/[3760]	Speed: 35.293595 samples/sec	 accuracy=82.825195	 loss=0.685978	 lr=0.000100
Epoch[095] Batch [1649]/[3760]	Speed: 35.061225 samples/sec	 accuracy=82.791667	 loss=0.686695	 lr=0.000100
Epoch[095] Batch [1699]/[3760]	Speed: 35.346732 samples/sec	 accuracy=82.777574	 loss=0.686022	 lr=0.000100
Epoch[095] Batch [1749]/[3760]	Speed: 35.411440 samples/sec	 accuracy=82.707143	 loss=0.688734	 lr=0.000100
Epoch[095] Batch [1799]/[3760]	Speed: 35.227982 samples/sec	 accuracy=82.722222	 loss=0.687918	 lr=0.000100
Epoch[095] Batch [1849]/[3760]	Speed: 35.280843 samples/sec	 accuracy=82.755068	 loss=0.687195	 lr=0.000100
Epoch[095] Batch [1899]/[3760]	Speed: 35.318341 samples/sec	 accuracy=82.775493	 loss=0.686453	 lr=0.000100
Epoch[095] Batch [1949]/[3760]	Speed: 35.406170 samples/sec	 accuracy=82.762821	 loss=0.686641	 lr=0.000100
Epoch[095] Batch [1999]/[3760]	Speed: 35.107043 samples/sec	 accuracy=82.773438	 loss=0.686199	 lr=0.000100
Epoch[095] Batch [2049]/[3760]	Speed: 35.436706 samples/sec	 accuracy=82.748476	 loss=0.687532	 lr=0.000100
Epoch[095] Batch [2099]/[3760]	Speed: 35.167258 samples/sec	 accuracy=82.770089	 loss=0.686711	 lr=0.000100
Epoch[095] Batch [2149]/[3760]	Speed: 35.593337 samples/sec	 accuracy=82.759448	 loss=0.686520	 lr=0.000100
Epoch[095] Batch [2199]/[3760]	Speed: 35.120024 samples/sec	 accuracy=82.756392	 loss=0.686602	 lr=0.000100
Epoch[095] Batch [2249]/[3760]	Speed: 35.481336 samples/sec	 accuracy=82.772917	 loss=0.686088	 lr=0.000100
Epoch[095] Batch [2299]/[3760]	Speed: 35.237434 samples/sec	 accuracy=82.751359	 loss=0.687096	 lr=0.000100
Epoch[095] Batch [2349]/[3760]	Speed: 35.395592 samples/sec	 accuracy=82.742686	 loss=0.687802	 lr=0.000100
Epoch[095] Batch [2399]/[3760]	Speed: 35.024286 samples/sec	 accuracy=82.759766	 loss=0.687457	 lr=0.000100
Epoch[095] Batch [2449]/[3760]	Speed: 35.485761 samples/sec	 accuracy=82.774235	 loss=0.686812	 lr=0.000100
Epoch[095] Batch [2499]/[3760]	Speed: 35.176159 samples/sec	 accuracy=82.792500	 loss=0.686540	 lr=0.000100
Epoch[095] Batch [2549]/[3760]	Speed: 35.303818 samples/sec	 accuracy=82.783701	 loss=0.687003	 lr=0.000100
Epoch[095] Batch [2599]/[3760]	Speed: 34.744278 samples/sec	 accuracy=82.789663	 loss=0.687385	 lr=0.000100
Epoch[095] Batch [2649]/[3760]	Speed: 35.374313 samples/sec	 accuracy=82.772995	 loss=0.688273	 lr=0.000100
Epoch[095] Batch [2699]/[3760]	Speed: 34.981851 samples/sec	 accuracy=82.789931	 loss=0.687394	 lr=0.000100
Epoch[095] Batch [2749]/[3760]	Speed: 35.350537 samples/sec	 accuracy=82.764773	 loss=0.688001	 lr=0.000100
Epoch[095] Batch [2799]/[3760]	Speed: 34.845871 samples/sec	 accuracy=82.760603	 loss=0.688271	 lr=0.000100
Epoch[095] Batch [2849]/[3760]	Speed: 35.251597 samples/sec	 accuracy=82.761513	 loss=0.688253	 lr=0.000100
Epoch[095] Batch [2899]/[3760]	Speed: 35.059081 samples/sec	 accuracy=82.768319	 loss=0.688117	 lr=0.000100
Epoch[095] Batch [2949]/[3760]	Speed: 34.931920 samples/sec	 accuracy=82.764831	 loss=0.688047	 lr=0.000100
Epoch[095] Batch [2999]/[3760]	Speed: 35.260454 samples/sec	 accuracy=82.739583	 loss=0.688732	 lr=0.000100
Epoch[095] Batch [3049]/[3760]	Speed: 34.774634 samples/sec	 accuracy=82.773053	 loss=0.687924	 lr=0.000100
Epoch[095] Batch [3099]/[3760]	Speed: 35.242837 samples/sec	 accuracy=82.773185	 loss=0.688208	 lr=0.000100
Epoch[095] Batch [3149]/[3760]	Speed: 34.837564 samples/sec	 accuracy=82.776290	 loss=0.687786	 lr=0.000100
Epoch[095] Batch [3199]/[3760]	Speed: 35.386501 samples/sec	 accuracy=82.771484	 loss=0.688126	 lr=0.000100
Epoch[095] Batch [3249]/[3760]	Speed: 34.801819 samples/sec	 accuracy=82.764423	 loss=0.688132	 lr=0.000100
Epoch[095] Batch [3299]/[3760]	Speed: 35.352614 samples/sec	 accuracy=82.762784	 loss=0.687924	 lr=0.000100
Epoch[095] Batch [3349]/[3760]	Speed: 35.070844 samples/sec	 accuracy=82.770056	 loss=0.688066	 lr=0.000100
Epoch[095] Batch [3399]/[3760]	Speed: 35.341985 samples/sec	 accuracy=82.776654	 loss=0.687967	 lr=0.000100
Epoch[095] Batch [3449]/[3760]	Speed: 35.213638 samples/sec	 accuracy=82.786232	 loss=0.688058	 lr=0.000100
Epoch[095] Batch [3499]/[3760]	Speed: 34.914053 samples/sec	 accuracy=82.783929	 loss=0.688272	 lr=0.000100
Epoch[095] Batch [3549]/[3760]	Speed: 35.427953 samples/sec	 accuracy=82.790493	 loss=0.688191	 lr=0.000100
Epoch[095] Batch [3599]/[3760]	Speed: 35.180633 samples/sec	 accuracy=82.803819	 loss=0.687995	 lr=0.000100
Epoch[095] Batch [3649]/[3760]	Speed: 35.428689 samples/sec	 accuracy=82.811644	 loss=0.687710	 lr=0.000100
Epoch[095] Batch [3699]/[3760]	Speed: 35.084259 samples/sec	 accuracy=82.797297	 loss=0.688188	 lr=0.000100
Epoch[095] Batch [3749]/[3760]	Speed: 36.440299 samples/sec	 accuracy=82.783750	 loss=0.688591	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.406250 acc-top5=86.093750
Batch [0099]/[0303]: acc-top1=66.640625 acc-top5=85.281250
Batch [0149]/[0303]: acc-top1=67.104167 acc-top5=85.593750
Batch [0199]/[0303]: acc-top1=66.703125 acc-top5=85.601562
Batch [0249]/[0303]: acc-top1=66.406250 acc-top5=85.556250
Batch [0299]/[0303]: acc-top1=66.244792 acc-top5=85.598958
[Epoch 095] training: accuracy=82.780918	 loss=0.688558
[Epoch 095] speed: 35 samples/sec	time cost: 7313.681340
[Epoch 095] validation: acc-top1=66.243812 acc-top5=85.617781 loss=1.632296
Epoch[096] Batch [0049]/[3759]	Speed: 24.986747 samples/sec	 accuracy=82.593750	 loss=0.710596	 lr=0.000100
Epoch[096] Batch [0099]/[3759]	Speed: 35.536751 samples/sec	 accuracy=82.593750	 loss=0.694401	 lr=0.000100
Epoch[096] Batch [0149]/[3759]	Speed: 35.197810 samples/sec	 accuracy=82.927083	 loss=0.688573	 lr=0.000100
Epoch[096] Batch [0199]/[3759]	Speed: 35.344639 samples/sec	 accuracy=82.828125	 loss=0.689373	 lr=0.000100
Epoch[096] Batch [0249]/[3759]	Speed: 35.296402 samples/sec	 accuracy=82.918750	 loss=0.683942	 lr=0.000100
Epoch[096] Batch [0299]/[3759]	Speed: 35.357762 samples/sec	 accuracy=82.911458	 loss=0.684063	 lr=0.000100
Epoch[096] Batch [0349]/[3759]	Speed: 35.040236 samples/sec	 accuracy=82.754464	 loss=0.686104	 lr=0.000100
Epoch[096] Batch [0399]/[3759]	Speed: 35.198864 samples/sec	 accuracy=82.773438	 loss=0.689951	 lr=0.000100
Epoch[096] Batch [0449]/[3759]	Speed: 34.999666 samples/sec	 accuracy=82.715278	 loss=0.691449	 lr=0.000100
Epoch[096] Batch [0499]/[3759]	Speed: 35.240583 samples/sec	 accuracy=82.840625	 loss=0.688376	 lr=0.000100
Epoch[096] Batch [0549]/[3759]	Speed: 34.982594 samples/sec	 accuracy=82.835227	 loss=0.688185	 lr=0.000100
Epoch[096] Batch [0599]/[3759]	Speed: 35.181344 samples/sec	 accuracy=82.791667	 loss=0.687922	 lr=0.000100
Epoch[096] Batch [0649]/[3759]	Speed: 35.220741 samples/sec	 accuracy=82.915865	 loss=0.685334	 lr=0.000100
Epoch[096] Batch [0699]/[3759]	Speed: 34.928591 samples/sec	 accuracy=82.928571	 loss=0.685733	 lr=0.000100
Epoch[096] Batch [0749]/[3759]	Speed: 35.311331 samples/sec	 accuracy=82.850000	 loss=0.687343	 lr=0.000100
Epoch[096] Batch [0799]/[3759]	Speed: 35.017434 samples/sec	 accuracy=82.833984	 loss=0.685021	 lr=0.000100
Epoch[096] Batch [0849]/[3759]	Speed: 35.374025 samples/sec	 accuracy=82.806985	 loss=0.685109	 lr=0.000100
Epoch[096] Batch [0899]/[3759]	Speed: 34.869722 samples/sec	 accuracy=82.789931	 loss=0.685702	 lr=0.000100
Epoch[096] Batch [0949]/[3759]	Speed: 35.255305 samples/sec	 accuracy=82.771382	 loss=0.685375	 lr=0.000100
Epoch[096] Batch [0999]/[3759]	Speed: 34.962626 samples/sec	 accuracy=82.806250	 loss=0.684022	 lr=0.000100
Epoch[096] Batch [1049]/[3759]	Speed: 35.156066 samples/sec	 accuracy=82.827381	 loss=0.684613	 lr=0.000100
Epoch[096] Batch [1099]/[3759]	Speed: 35.093790 samples/sec	 accuracy=82.840909	 loss=0.684656	 lr=0.000100
Epoch[096] Batch [1149]/[3759]	Speed: 35.259553 samples/sec	 accuracy=82.824728	 loss=0.684120	 lr=0.000100
Epoch[096] Batch [1199]/[3759]	Speed: 34.941933 samples/sec	 accuracy=82.802083	 loss=0.686082	 lr=0.000100
Epoch[096] Batch [1249]/[3759]	Speed: 35.218647 samples/sec	 accuracy=82.741250	 loss=0.688863	 lr=0.000100
Epoch[096] Batch [1299]/[3759]	Speed: 34.923055 samples/sec	 accuracy=82.763221	 loss=0.688780	 lr=0.000100
Epoch[096] Batch [1349]/[3759]	Speed: 35.156498 samples/sec	 accuracy=82.771991	 loss=0.688372	 lr=0.000100
Epoch[096] Batch [1399]/[3759]	Speed: 35.062669 samples/sec	 accuracy=82.755580	 loss=0.688645	 lr=0.000100
Epoch[096] Batch [1449]/[3759]	Speed: 35.153504 samples/sec	 accuracy=82.750000	 loss=0.689179	 lr=0.000100
Epoch[096] Batch [1499]/[3759]	Speed: 35.025739 samples/sec	 accuracy=82.736458	 loss=0.689916	 lr=0.000100
Epoch[096] Batch [1549]/[3759]	Speed: 34.967805 samples/sec	 accuracy=82.764113	 loss=0.689498	 lr=0.000100
Epoch[096] Batch [1599]/[3759]	Speed: 35.370103 samples/sec	 accuracy=82.791016	 loss=0.688926	 lr=0.000100
Epoch[096] Batch [1649]/[3759]	Speed: 34.837488 samples/sec	 accuracy=82.789773	 loss=0.688745	 lr=0.000100
Epoch[096] Batch [1699]/[3759]	Speed: 35.305589 samples/sec	 accuracy=82.789522	 loss=0.688519	 lr=0.000100
Epoch[096] Batch [1749]/[3759]	Speed: 34.692122 samples/sec	 accuracy=82.808036	 loss=0.687874	 lr=0.000100
Epoch[096] Batch [1799]/[3759]	Speed: 35.173721 samples/sec	 accuracy=82.812500	 loss=0.687586	 lr=0.000100
Epoch[096] Batch [1849]/[3759]	Speed: 34.933466 samples/sec	 accuracy=82.820946	 loss=0.686896	 lr=0.000100
Epoch[096] Batch [1899]/[3759]	Speed: 35.279861 samples/sec	 accuracy=82.812500	 loss=0.687041	 lr=0.000100
Epoch[096] Batch [1949]/[3759]	Speed: 34.833878 samples/sec	 accuracy=82.781250	 loss=0.687707	 lr=0.000100
Epoch[096] Batch [1999]/[3759]	Speed: 35.207407 samples/sec	 accuracy=82.784375	 loss=0.687618	 lr=0.000100
Epoch[096] Batch [2049]/[3759]	Speed: 35.297040 samples/sec	 accuracy=82.793445	 loss=0.686743	 lr=0.000100
Epoch[096] Batch [2099]/[3759]	Speed: 35.119997 samples/sec	 accuracy=82.800595	 loss=0.686688	 lr=0.000100
Epoch[096] Batch [2149]/[3759]	Speed: 35.106506 samples/sec	 accuracy=82.792878	 loss=0.687075	 lr=0.000100
Epoch[096] Batch [2199]/[3759]	Speed: 35.116967 samples/sec	 accuracy=82.784801	 loss=0.687363	 lr=0.000100
Epoch[096] Batch [2249]/[3759]	Speed: 35.268636 samples/sec	 accuracy=82.777083	 loss=0.687647	 lr=0.000100
Epoch[096] Batch [2299]/[3759]	Speed: 35.209995 samples/sec	 accuracy=82.793478	 loss=0.687129	 lr=0.000100
Epoch[096] Batch [2349]/[3759]	Speed: 35.390281 samples/sec	 accuracy=82.786569	 loss=0.687461	 lr=0.000100
Epoch[096] Batch [2399]/[3759]	Speed: 35.401576 samples/sec	 accuracy=82.803385	 loss=0.687195	 lr=0.000100
Epoch[096] Batch [2449]/[3759]	Speed: 35.333963 samples/sec	 accuracy=82.781888	 loss=0.687572	 lr=0.000100
Epoch[096] Batch [2499]/[3759]	Speed: 35.082042 samples/sec	 accuracy=82.786250	 loss=0.687253	 lr=0.000100
Epoch[096] Batch [2549]/[3759]	Speed: 35.106525 samples/sec	 accuracy=82.797181	 loss=0.686646	 lr=0.000100
Epoch[096] Batch [2599]/[3759]	Speed: 35.318123 samples/sec	 accuracy=82.796875	 loss=0.686595	 lr=0.000100
Epoch[096] Batch [2649]/[3759]	Speed: 35.085240 samples/sec	 accuracy=82.797759	 loss=0.686784	 lr=0.000100
Epoch[096] Batch [2699]/[3759]	Speed: 35.319412 samples/sec	 accuracy=82.792245	 loss=0.686904	 lr=0.000100
Epoch[096] Batch [2749]/[3759]	Speed: 35.169281 samples/sec	 accuracy=82.781818	 loss=0.687032	 lr=0.000100
Epoch[096] Batch [2799]/[3759]	Speed: 35.194016 samples/sec	 accuracy=82.786830	 loss=0.687068	 lr=0.000100
Epoch[096] Batch [2849]/[3759]	Speed: 35.393566 samples/sec	 accuracy=82.805921	 loss=0.686346	 lr=0.000100
Epoch[096] Batch [2899]/[3759]	Speed: 35.195108 samples/sec	 accuracy=82.790409	 loss=0.686724	 lr=0.000100
Epoch[096] Batch [2949]/[3759]	Speed: 35.140392 samples/sec	 accuracy=82.796610	 loss=0.686671	 lr=0.000100
Epoch[096] Batch [2999]/[3759]	Speed: 35.183509 samples/sec	 accuracy=82.798958	 loss=0.686664	 lr=0.000100
Epoch[096] Batch [3049]/[3759]	Speed: 35.143728 samples/sec	 accuracy=82.803791	 loss=0.686667	 lr=0.000100
Epoch[096] Batch [3099]/[3759]	Speed: 35.067734 samples/sec	 accuracy=82.806452	 loss=0.686561	 lr=0.000100
Epoch[096] Batch [3149]/[3759]	Speed: 35.099997 samples/sec	 accuracy=82.807540	 loss=0.686875	 lr=0.000100
Epoch[096] Batch [3199]/[3759]	Speed: 35.437929 samples/sec	 accuracy=82.803711	 loss=0.686810	 lr=0.000100
Epoch[096] Batch [3249]/[3759]	Speed: 34.910221 samples/sec	 accuracy=82.821635	 loss=0.686531	 lr=0.000100
Epoch[096] Batch [3299]/[3759]	Speed: 35.469250 samples/sec	 accuracy=82.839015	 loss=0.685871	 lr=0.000100
Epoch[096] Batch [3349]/[3759]	Speed: 35.307782 samples/sec	 accuracy=82.812034	 loss=0.686545	 lr=0.000100
Epoch[096] Batch [3399]/[3759]	Speed: 35.304428 samples/sec	 accuracy=82.811581	 loss=0.686867	 lr=0.000100
Epoch[096] Batch [3449]/[3759]	Speed: 35.429881 samples/sec	 accuracy=82.827899	 loss=0.686579	 lr=0.000100
Epoch[096] Batch [3499]/[3759]	Speed: 35.240042 samples/sec	 accuracy=82.840179	 loss=0.686337	 lr=0.000100
Epoch[096] Batch [3549]/[3759]	Speed: 35.284607 samples/sec	 accuracy=82.843750	 loss=0.686248	 lr=0.000100
Epoch[096] Batch [3599]/[3759]	Speed: 35.202659 samples/sec	 accuracy=82.845920	 loss=0.686137	 lr=0.000100
Epoch[096] Batch [3649]/[3759]	Speed: 35.134592 samples/sec	 accuracy=82.847175	 loss=0.686229	 lr=0.000100
Epoch[096] Batch [3699]/[3759]	Speed: 35.143528 samples/sec	 accuracy=82.847973	 loss=0.685915	 lr=0.000100
Epoch[096] Batch [3749]/[3759]	Speed: 36.372234 samples/sec	 accuracy=82.847083	 loss=0.685914	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.031250 acc-top5=86.531250
Batch [0099]/[0303]: acc-top1=66.375000 acc-top5=85.578125
Batch [0149]/[0303]: acc-top1=67.000000 acc-top5=85.645833
Batch [0199]/[0303]: acc-top1=66.718750 acc-top5=85.601562
Batch [0249]/[0303]: acc-top1=66.531250 acc-top5=85.556250
Batch [0299]/[0303]: acc-top1=66.432292 acc-top5=85.531250
[Epoch 096] training: accuracy=82.841597	 loss=0.685997
[Epoch 096] speed: 34 samples/sec	time cost: 7325.845905
[Epoch 096] validation: acc-top1=66.424299 acc-top5=85.545586 loss=1.639184
Epoch[097] Batch [0049]/[3760]	Speed: 25.472834 samples/sec	 accuracy=82.937500	 loss=0.659798	 lr=0.000100
Epoch[097] Batch [0099]/[3760]	Speed: 35.095436 samples/sec	 accuracy=83.062500	 loss=0.664250	 lr=0.000100
Epoch[097] Batch [0149]/[3760]	Speed: 35.374104 samples/sec	 accuracy=82.968750	 loss=0.675044	 lr=0.000100
Epoch[097] Batch [0199]/[3760]	Speed: 34.877947 samples/sec	 accuracy=82.890625	 loss=0.681757	 lr=0.000100
Epoch[097] Batch [0249]/[3760]	Speed: 35.347458 samples/sec	 accuracy=82.862500	 loss=0.683855	 lr=0.000100
Epoch[097] Batch [0299]/[3760]	Speed: 35.060841 samples/sec	 accuracy=82.828125	 loss=0.684546	 lr=0.000100
Epoch[097] Batch [0349]/[3760]	Speed: 35.176449 samples/sec	 accuracy=82.843750	 loss=0.684963	 lr=0.000100
Epoch[097] Batch [0399]/[3760]	Speed: 34.950595 samples/sec	 accuracy=82.746094	 loss=0.688901	 lr=0.000100
Epoch[097] Batch [0449]/[3760]	Speed: 35.340790 samples/sec	 accuracy=82.656250	 loss=0.693053	 lr=0.000100
Epoch[097] Batch [0499]/[3760]	Speed: 34.910825 samples/sec	 accuracy=82.678125	 loss=0.691432	 lr=0.000100
Epoch[097] Batch [0549]/[3760]	Speed: 35.376569 samples/sec	 accuracy=82.627841	 loss=0.691813	 lr=0.000100
Epoch[097] Batch [0599]/[3760]	Speed: 34.926471 samples/sec	 accuracy=82.690104	 loss=0.689389	 lr=0.000100
Epoch[097] Batch [0649]/[3760]	Speed: 35.483971 samples/sec	 accuracy=82.709135	 loss=0.688390	 lr=0.000100
Epoch[097] Batch [0699]/[3760]	Speed: 35.055728 samples/sec	 accuracy=82.736607	 loss=0.687135	 lr=0.000100
Epoch[097] Batch [0749]/[3760]	Speed: 35.221651 samples/sec	 accuracy=82.814583	 loss=0.685071	 lr=0.000100
Epoch[097] Batch [0799]/[3760]	Speed: 34.848741 samples/sec	 accuracy=82.818359	 loss=0.684739	 lr=0.000100
Epoch[097] Batch [0849]/[3760]	Speed: 35.313358 samples/sec	 accuracy=82.812500	 loss=0.685978	 lr=0.000100
Epoch[097] Batch [0899]/[3760]	Speed: 34.975738 samples/sec	 accuracy=82.803819	 loss=0.686863	 lr=0.000100
Epoch[097] Batch [0949]/[3760]	Speed: 35.273589 samples/sec	 accuracy=82.827303	 loss=0.686499	 lr=0.000100
Epoch[097] Batch [0999]/[3760]	Speed: 35.145364 samples/sec	 accuracy=82.848438	 loss=0.686154	 lr=0.000100
Epoch[097] Batch [1049]/[3760]	Speed: 35.120484 samples/sec	 accuracy=82.816964	 loss=0.686516	 lr=0.000100
Epoch[097] Batch [1099]/[3760]	Speed: 34.958199 samples/sec	 accuracy=82.821023	 loss=0.686695	 lr=0.000100
Epoch[097] Batch [1149]/[3760]	Speed: 35.031436 samples/sec	 accuracy=82.843750	 loss=0.686480	 lr=0.000100
Epoch[097] Batch [1199]/[3760]	Speed: 35.122964 samples/sec	 accuracy=82.851562	 loss=0.686843	 lr=0.000100
Epoch[097] Batch [1249]/[3760]	Speed: 34.782002 samples/sec	 accuracy=82.820000	 loss=0.687297	 lr=0.000100
Epoch[097] Batch [1299]/[3760]	Speed: 35.002282 samples/sec	 accuracy=82.807692	 loss=0.688083	 lr=0.000100
Epoch[097] Batch [1349]/[3760]	Speed: 34.633422 samples/sec	 accuracy=82.835648	 loss=0.686634	 lr=0.000100
Epoch[097] Batch [1399]/[3760]	Speed: 35.145110 samples/sec	 accuracy=82.825893	 loss=0.686797	 lr=0.000100
Epoch[097] Batch [1449]/[3760]	Speed: 34.514070 samples/sec	 accuracy=82.837284	 loss=0.686541	 lr=0.000100
Epoch[097] Batch [1499]/[3760]	Speed: 35.171552 samples/sec	 accuracy=82.841667	 loss=0.686782	 lr=0.000100
Epoch[097] Batch [1549]/[3760]	Speed: 34.509919 samples/sec	 accuracy=82.853831	 loss=0.686380	 lr=0.000100
Epoch[097] Batch [1599]/[3760]	Speed: 35.004358 samples/sec	 accuracy=82.846680	 loss=0.686688	 lr=0.000100
Epoch[097] Batch [1649]/[3760]	Speed: 34.639817 samples/sec	 accuracy=82.848485	 loss=0.687005	 lr=0.000100
Epoch[097] Batch [1699]/[3760]	Speed: 34.859585 samples/sec	 accuracy=82.853860	 loss=0.686532	 lr=0.000100
Epoch[097] Batch [1749]/[3760]	Speed: 34.760394 samples/sec	 accuracy=82.858929	 loss=0.686639	 lr=0.000100
Epoch[097] Batch [1799]/[3760]	Speed: 34.727893 samples/sec	 accuracy=82.869792	 loss=0.686470	 lr=0.000100
Epoch[097] Batch [1849]/[3760]	Speed: 35.005066 samples/sec	 accuracy=82.873311	 loss=0.685801	 lr=0.000100
Epoch[097] Batch [1899]/[3760]	Speed: 34.500301 samples/sec	 accuracy=82.833882	 loss=0.686609	 lr=0.000100
Epoch[097] Batch [1949]/[3760]	Speed: 35.245171 samples/sec	 accuracy=82.825321	 loss=0.687246	 lr=0.000100
Epoch[097] Batch [1999]/[3760]	Speed: 34.425855 samples/sec	 accuracy=82.839063	 loss=0.686371	 lr=0.000100
Epoch[097] Batch [2049]/[3760]	Speed: 35.217771 samples/sec	 accuracy=82.862805	 loss=0.685754	 lr=0.000100
Epoch[097] Batch [2099]/[3760]	Speed: 34.567109 samples/sec	 accuracy=82.854167	 loss=0.685895	 lr=0.000100
Epoch[097] Batch [2149]/[3760]	Speed: 35.149553 samples/sec	 accuracy=82.856105	 loss=0.685670	 lr=0.000100
Epoch[097] Batch [2199]/[3760]	Speed: 35.128254 samples/sec	 accuracy=82.848722	 loss=0.686350	 lr=0.000100
Epoch[097] Batch [2249]/[3760]	Speed: 35.187971 samples/sec	 accuracy=82.820139	 loss=0.687829	 lr=0.000100
Epoch[097] Batch [2299]/[3760]	Speed: 35.072217 samples/sec	 accuracy=82.840353	 loss=0.687346	 lr=0.000100
Epoch[097] Batch [2349]/[3760]	Speed: 35.183505 samples/sec	 accuracy=82.859707	 loss=0.687204	 lr=0.000100
Epoch[097] Batch [2399]/[3760]	Speed: 35.303149 samples/sec	 accuracy=82.843750	 loss=0.687663	 lr=0.000100
Epoch[097] Batch [2449]/[3760]	Speed: 35.111646 samples/sec	 accuracy=82.841199	 loss=0.687906	 lr=0.000100
Epoch[097] Batch [2499]/[3760]	Speed: 35.353308 samples/sec	 accuracy=82.836875	 loss=0.688010	 lr=0.000100
Epoch[097] Batch [2549]/[3760]	Speed: 35.015266 samples/sec	 accuracy=82.847426	 loss=0.687456	 lr=0.000100
Epoch[097] Batch [2599]/[3760]	Speed: 35.381605 samples/sec	 accuracy=82.861178	 loss=0.687043	 lr=0.000100
Epoch[097] Batch [2649]/[3760]	Speed: 35.061910 samples/sec	 accuracy=82.850825	 loss=0.687489	 lr=0.000100
Epoch[097] Batch [2699]/[3760]	Speed: 35.323130 samples/sec	 accuracy=82.855903	 loss=0.687327	 lr=0.000100
Epoch[097] Batch [2749]/[3760]	Speed: 34.983309 samples/sec	 accuracy=82.860795	 loss=0.687470	 lr=0.000100
Epoch[097] Batch [2799]/[3760]	Speed: 35.193299 samples/sec	 accuracy=82.853237	 loss=0.687923	 lr=0.000100
Epoch[097] Batch [2849]/[3760]	Speed: 34.809046 samples/sec	 accuracy=82.873904	 loss=0.687741	 lr=0.000100
Epoch[097] Batch [2899]/[3760]	Speed: 35.461891 samples/sec	 accuracy=82.849138	 loss=0.688476	 lr=0.000100
Epoch[097] Batch [2949]/[3760]	Speed: 34.880944 samples/sec	 accuracy=82.854873	 loss=0.688500	 lr=0.000100
Epoch[097] Batch [2999]/[3760]	Speed: 35.351872 samples/sec	 accuracy=82.846875	 loss=0.688577	 lr=0.000100
Epoch[097] Batch [3049]/[3760]	Speed: 34.942247 samples/sec	 accuracy=82.836066	 loss=0.689121	 lr=0.000100
Epoch[097] Batch [3099]/[3760]	Speed: 35.355056 samples/sec	 accuracy=82.831149	 loss=0.688568	 lr=0.000100
Epoch[097] Batch [3149]/[3760]	Speed: 35.047887 samples/sec	 accuracy=82.830853	 loss=0.688778	 lr=0.000100
Epoch[097] Batch [3199]/[3760]	Speed: 35.184623 samples/sec	 accuracy=82.813477	 loss=0.689128	 lr=0.000100
Epoch[097] Batch [3249]/[3760]	Speed: 35.381490 samples/sec	 accuracy=82.810577	 loss=0.689017	 lr=0.000100
Epoch[097] Batch [3299]/[3760]	Speed: 35.109546 samples/sec	 accuracy=82.803030	 loss=0.689217	 lr=0.000100
Epoch[097] Batch [3349]/[3760]	Speed: 35.218872 samples/sec	 accuracy=82.791978	 loss=0.689632	 lr=0.000100
Epoch[097] Batch [3399]/[3760]	Speed: 35.179570 samples/sec	 accuracy=82.830882	 loss=0.688131	 lr=0.000100
Epoch[097] Batch [3449]/[3760]	Speed: 34.994554 samples/sec	 accuracy=82.840127	 loss=0.687572	 lr=0.000100
Epoch[097] Batch [3499]/[3760]	Speed: 35.277487 samples/sec	 accuracy=82.833929	 loss=0.687645	 lr=0.000100
Epoch[097] Batch [3549]/[3760]	Speed: 34.980088 samples/sec	 accuracy=82.822623	 loss=0.687924	 lr=0.000100
Epoch[097] Batch [3599]/[3760]	Speed: 34.853196 samples/sec	 accuracy=82.833767	 loss=0.687815	 lr=0.000100
Epoch[097] Batch [3649]/[3760]	Speed: 34.936682 samples/sec	 accuracy=82.821062	 loss=0.688240	 lr=0.000100
Epoch[097] Batch [3699]/[3760]	Speed: 34.615540 samples/sec	 accuracy=82.828125	 loss=0.687886	 lr=0.000100
Epoch[097] Batch [3749]/[3760]	Speed: 36.338951 samples/sec	 accuracy=82.842500	 loss=0.687582	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.343750 acc-top5=86.031250
Batch [0099]/[0303]: acc-top1=66.640625 acc-top5=85.218750
Batch [0149]/[0303]: acc-top1=67.166667 acc-top5=85.416667
Batch [0199]/[0303]: acc-top1=66.742188 acc-top5=85.500000
Batch [0249]/[0303]: acc-top1=66.562500 acc-top5=85.412500
Batch [0299]/[0303]: acc-top1=66.239583 acc-top5=85.364583
[Epoch 097] training: accuracy=82.840342	 loss=0.687511
[Epoch 097] speed: 34 samples/sec	time cost: 7345.968766
[Epoch 097] validation: acc-top1=66.233498 acc-top5=85.375413 loss=1.641600
Epoch[098] Batch [0049]/[3759]	Speed: 24.944074 samples/sec	 accuracy=81.968750	 loss=0.711876	 lr=0.000100
Epoch[098] Batch [0099]/[3759]	Speed: 35.567634 samples/sec	 accuracy=82.843750	 loss=0.699806	 lr=0.000100
Epoch[098] Batch [0149]/[3759]	Speed: 35.252144 samples/sec	 accuracy=82.541667	 loss=0.707285	 lr=0.000100
Epoch[098] Batch [0199]/[3759]	Speed: 35.292458 samples/sec	 accuracy=82.734375	 loss=0.696739	 lr=0.000100
Epoch[098] Batch [0249]/[3759]	Speed: 35.019028 samples/sec	 accuracy=82.981250	 loss=0.686696	 lr=0.000100
Epoch[098] Batch [0299]/[3759]	Speed: 35.364430 samples/sec	 accuracy=83.098958	 loss=0.681775	 lr=0.000100
Epoch[098] Batch [0349]/[3759]	Speed: 35.101017 samples/sec	 accuracy=83.044643	 loss=0.683525	 lr=0.000100
Epoch[098] Batch [0399]/[3759]	Speed: 35.302308 samples/sec	 accuracy=82.960938	 loss=0.684658	 lr=0.000100
Epoch[098] Batch [0449]/[3759]	Speed: 35.260912 samples/sec	 accuracy=82.937500	 loss=0.686113	 lr=0.000100
Epoch[098] Batch [0499]/[3759]	Speed: 35.364942 samples/sec	 accuracy=82.853125	 loss=0.690278	 lr=0.000100
Epoch[098] Batch [0549]/[3759]	Speed: 35.043774 samples/sec	 accuracy=82.923295	 loss=0.687920	 lr=0.000100
Epoch[098] Batch [0599]/[3759]	Speed: 35.164455 samples/sec	 accuracy=82.908854	 loss=0.688574	 lr=0.000100
Epoch[098] Batch [0649]/[3759]	Speed: 35.155459 samples/sec	 accuracy=82.932692	 loss=0.686693	 lr=0.000100
Epoch[098] Batch [0699]/[3759]	Speed: 35.210694 samples/sec	 accuracy=82.966518	 loss=0.686208	 lr=0.000100
Epoch[098] Batch [0749]/[3759]	Speed: 35.341482 samples/sec	 accuracy=82.897917	 loss=0.686629	 lr=0.000100
Epoch[098] Batch [0799]/[3759]	Speed: 34.925828 samples/sec	 accuracy=82.951172	 loss=0.685605	 lr=0.000100
Epoch[098] Batch [0849]/[3759]	Speed: 35.312407 samples/sec	 accuracy=82.955882	 loss=0.685144	 lr=0.000100
Epoch[098] Batch [0899]/[3759]	Speed: 35.103780 samples/sec	 accuracy=82.920139	 loss=0.685898	 lr=0.000100
Epoch[098] Batch [0949]/[3759]	Speed: 35.238231 samples/sec	 accuracy=82.901316	 loss=0.686760	 lr=0.000100
Epoch[098] Batch [0999]/[3759]	Speed: 35.099840 samples/sec	 accuracy=82.934375	 loss=0.685791	 lr=0.000100
Epoch[098] Batch [1049]/[3759]	Speed: 35.309160 samples/sec	 accuracy=82.934524	 loss=0.684965	 lr=0.000100
Epoch[098] Batch [1099]/[3759]	Speed: 34.765321 samples/sec	 accuracy=82.853693	 loss=0.687033	 lr=0.000100
Epoch[098] Batch [1149]/[3759]	Speed: 35.361735 samples/sec	 accuracy=82.802989	 loss=0.688156	 lr=0.000100
Epoch[098] Batch [1199]/[3759]	Speed: 35.008292 samples/sec	 accuracy=82.747396	 loss=0.689087	 lr=0.000100
Epoch[098] Batch [1249]/[3759]	Speed: 35.375475 samples/sec	 accuracy=82.738750	 loss=0.688324	 lr=0.000100
Epoch[098] Batch [1299]/[3759]	Speed: 34.902562 samples/sec	 accuracy=82.770433	 loss=0.686835	 lr=0.000100
Epoch[098] Batch [1349]/[3759]	Speed: 35.418473 samples/sec	 accuracy=82.752315	 loss=0.686967	 lr=0.000100
Epoch[098] Batch [1399]/[3759]	Speed: 34.992742 samples/sec	 accuracy=82.753348	 loss=0.686918	 lr=0.000100
Epoch[098] Batch [1449]/[3759]	Speed: 35.318041 samples/sec	 accuracy=82.759698	 loss=0.687252	 lr=0.000100
Epoch[098] Batch [1499]/[3759]	Speed: 35.130680 samples/sec	 accuracy=82.782292	 loss=0.687119	 lr=0.000100
Epoch[098] Batch [1549]/[3759]	Speed: 35.135895 samples/sec	 accuracy=82.814516	 loss=0.686411	 lr=0.000100
Epoch[098] Batch [1599]/[3759]	Speed: 35.514603 samples/sec	 accuracy=82.843750	 loss=0.686274	 lr=0.000100
Epoch[098] Batch [1649]/[3759]	Speed: 35.243916 samples/sec	 accuracy=82.857955	 loss=0.686476	 lr=0.000100
Epoch[098] Batch [1699]/[3759]	Speed: 35.407368 samples/sec	 accuracy=82.834559	 loss=0.686856	 lr=0.000100
Epoch[098] Batch [1749]/[3759]	Speed: 35.209216 samples/sec	 accuracy=82.849107	 loss=0.686621	 lr=0.000100
Epoch[098] Batch [1799]/[3759]	Speed: 35.252527 samples/sec	 accuracy=82.850694	 loss=0.685735	 lr=0.000100
Epoch[098] Batch [1849]/[3759]	Speed: 35.166299 samples/sec	 accuracy=82.867399	 loss=0.684690	 lr=0.000100
Epoch[098] Batch [1899]/[3759]	Speed: 34.935792 samples/sec	 accuracy=82.874178	 loss=0.684737	 lr=0.000100
Epoch[098] Batch [1949]/[3759]	Speed: 34.953985 samples/sec	 accuracy=82.866987	 loss=0.684871	 lr=0.000100
Epoch[098] Batch [1999]/[3759]	Speed: 35.246778 samples/sec	 accuracy=82.857813	 loss=0.685392	 lr=0.000100
Epoch[098] Batch [2049]/[3759]	Speed: 34.776641 samples/sec	 accuracy=82.853659	 loss=0.685228	 lr=0.000100
Epoch[098] Batch [2099]/[3759]	Speed: 35.110280 samples/sec	 accuracy=82.863839	 loss=0.684766	 lr=0.000100
Epoch[098] Batch [2149]/[3759]	Speed: 34.627376 samples/sec	 accuracy=82.867006	 loss=0.684490	 lr=0.000100
Epoch[098] Batch [2199]/[3759]	Speed: 35.251198 samples/sec	 accuracy=82.877131	 loss=0.684051	 lr=0.000100
Epoch[098] Batch [2249]/[3759]	Speed: 34.774384 samples/sec	 accuracy=82.879861	 loss=0.683248	 lr=0.000100
Epoch[098] Batch [2299]/[3759]	Speed: 35.282604 samples/sec	 accuracy=82.902174	 loss=0.682462	 lr=0.000100
Epoch[098] Batch [2349]/[3759]	Speed: 35.019903 samples/sec	 accuracy=82.908245	 loss=0.682532	 lr=0.000100
Epoch[098] Batch [2399]/[3759]	Speed: 34.946335 samples/sec	 accuracy=82.894531	 loss=0.682939	 lr=0.000100
Epoch[098] Batch [2449]/[3759]	Speed: 35.099608 samples/sec	 accuracy=82.894133	 loss=0.682566	 lr=0.000100
Epoch[098] Batch [2499]/[3759]	Speed: 35.030520 samples/sec	 accuracy=82.896875	 loss=0.682789	 lr=0.000100
Epoch[098] Batch [2549]/[3759]	Speed: 35.205951 samples/sec	 accuracy=82.909926	 loss=0.682147	 lr=0.000100
Epoch[098] Batch [2599]/[3759]	Speed: 34.645742 samples/sec	 accuracy=82.918269	 loss=0.682262	 lr=0.000100
Epoch[098] Batch [2649]/[3759]	Speed: 35.307261 samples/sec	 accuracy=82.927476	 loss=0.681949	 lr=0.000100
Epoch[098] Batch [2699]/[3759]	Speed: 34.760763 samples/sec	 accuracy=82.908565	 loss=0.682472	 lr=0.000100
Epoch[098] Batch [2749]/[3759]	Speed: 35.428612 samples/sec	 accuracy=82.909091	 loss=0.682605	 lr=0.000100
Epoch[098] Batch [2799]/[3759]	Speed: 34.832357 samples/sec	 accuracy=82.909040	 loss=0.682518	 lr=0.000100
Epoch[098] Batch [2849]/[3759]	Speed: 35.309328 samples/sec	 accuracy=82.915022	 loss=0.682091	 lr=0.000100
Epoch[098] Batch [2899]/[3759]	Speed: 34.753311 samples/sec	 accuracy=82.910022	 loss=0.682397	 lr=0.000100
Epoch[098] Batch [2949]/[3759]	Speed: 35.286546 samples/sec	 accuracy=82.890360	 loss=0.682932	 lr=0.000100
Epoch[098] Batch [2999]/[3759]	Speed: 34.572634 samples/sec	 accuracy=82.889583	 loss=0.682863	 lr=0.000100
Epoch[098] Batch [3049]/[3759]	Speed: 35.061692 samples/sec	 accuracy=82.909836	 loss=0.682066	 lr=0.000100
Epoch[098] Batch [3099]/[3759]	Speed: 34.534934 samples/sec	 accuracy=82.909274	 loss=0.682025	 lr=0.000100
Epoch[098] Batch [3149]/[3759]	Speed: 34.814671 samples/sec	 accuracy=82.904266	 loss=0.682635	 lr=0.000100
Epoch[098] Batch [3199]/[3759]	Speed: 34.832605 samples/sec	 accuracy=82.875977	 loss=0.683142	 lr=0.000100
Epoch[098] Batch [3249]/[3759]	Speed: 34.559163 samples/sec	 accuracy=82.869712	 loss=0.683431	 lr=0.000100
Epoch[098] Batch [3299]/[3759]	Speed: 35.066000 samples/sec	 accuracy=82.870265	 loss=0.683177	 lr=0.000100
Epoch[098] Batch [3349]/[3759]	Speed: 34.285578 samples/sec	 accuracy=82.875466	 loss=0.683125	 lr=0.000100
Epoch[098] Batch [3399]/[3759]	Speed: 34.935420 samples/sec	 accuracy=82.880974	 loss=0.682946	 lr=0.000100
Epoch[098] Batch [3449]/[3759]	Speed: 34.454066 samples/sec	 accuracy=82.867301	 loss=0.683218	 lr=0.000100
Epoch[098] Batch [3499]/[3759]	Speed: 35.005764 samples/sec	 accuracy=82.872321	 loss=0.683039	 lr=0.000100
Epoch[098] Batch [3549]/[3759]	Speed: 34.757817 samples/sec	 accuracy=82.863996	 loss=0.683246	 lr=0.000100
Epoch[098] Batch [3599]/[3759]	Speed: 35.073694 samples/sec	 accuracy=82.880208	 loss=0.683017	 lr=0.000100
Epoch[098] Batch [3649]/[3759]	Speed: 34.969455 samples/sec	 accuracy=82.882277	 loss=0.683069	 lr=0.000100
Epoch[098] Batch [3699]/[3759]	Speed: 35.049612 samples/sec	 accuracy=82.879645	 loss=0.682927	 lr=0.000100
Epoch[098] Batch [3749]/[3759]	Speed: 36.247289 samples/sec	 accuracy=82.880417	 loss=0.682798	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.562500 acc-top5=85.937500
Batch [0099]/[0303]: acc-top1=66.765625 acc-top5=85.296875
Batch [0149]/[0303]: acc-top1=67.031250 acc-top5=85.437500
Batch [0199]/[0303]: acc-top1=66.742188 acc-top5=85.554688
Batch [0249]/[0303]: acc-top1=66.568750 acc-top5=85.475000
Batch [0299]/[0303]: acc-top1=66.401042 acc-top5=85.473958
[Epoch 098] training: accuracy=82.881085	 loss=0.682925
[Epoch 098] speed: 34 samples/sec	time cost: 7344.280421
[Epoch 098] validation: acc-top1=66.408828 acc-top5=85.483705 loss=1.644588
Epoch[099] Batch [0049]/[3760]	Speed: 24.960580 samples/sec	 accuracy=83.093750	 loss=0.709836	 lr=0.000100
Epoch[099] Batch [0099]/[3760]	Speed: 34.791147 samples/sec	 accuracy=83.187500	 loss=0.676613	 lr=0.000100
Epoch[099] Batch [0149]/[3760]	Speed: 35.264750 samples/sec	 accuracy=83.479167	 loss=0.665646	 lr=0.000100
Epoch[099] Batch [0199]/[3760]	Speed: 34.703969 samples/sec	 accuracy=83.437500	 loss=0.667147	 lr=0.000100
Epoch[099] Batch [0249]/[3760]	Speed: 35.252445 samples/sec	 accuracy=83.287500	 loss=0.671528	 lr=0.000100
Epoch[099] Batch [0299]/[3760]	Speed: 34.703813 samples/sec	 accuracy=83.380208	 loss=0.671482	 lr=0.000100
Epoch[099] Batch [0349]/[3760]	Speed: 35.072405 samples/sec	 accuracy=83.218750	 loss=0.673505	 lr=0.000100
Epoch[099] Batch [0399]/[3760]	Speed: 34.568773 samples/sec	 accuracy=83.101562	 loss=0.674463	 lr=0.000100
Epoch[099] Batch [0449]/[3760]	Speed: 35.018552 samples/sec	 accuracy=83.097222	 loss=0.674096	 lr=0.000100
Epoch[099] Batch [0499]/[3760]	Speed: 34.960265 samples/sec	 accuracy=83.171875	 loss=0.670576	 lr=0.000100
Epoch[099] Batch [0549]/[3760]	Speed: 34.836120 samples/sec	 accuracy=83.215909	 loss=0.671122	 lr=0.000100
Epoch[099] Batch [0599]/[3760]	Speed: 35.173421 samples/sec	 accuracy=83.145833	 loss=0.673465	 lr=0.000100
Epoch[099] Batch [0649]/[3760]	Speed: 35.025878 samples/sec	 accuracy=83.120192	 loss=0.673012	 lr=0.000100
Epoch[099] Batch [0699]/[3760]	Speed: 35.185984 samples/sec	 accuracy=83.147321	 loss=0.671150	 lr=0.000100
Epoch[099] Batch [0749]/[3760]	Speed: 34.938408 samples/sec	 accuracy=83.097917	 loss=0.673786	 lr=0.000100
Epoch[099] Batch [0799]/[3760]	Speed: 35.352953 samples/sec	 accuracy=83.021484	 loss=0.678038	 lr=0.000100
Epoch[099] Batch [0849]/[3760]	Speed: 35.145744 samples/sec	 accuracy=83.068015	 loss=0.678521	 lr=0.000100
Epoch[099] Batch [0899]/[3760]	Speed: 35.190348 samples/sec	 accuracy=83.092014	 loss=0.677786	 lr=0.000100
Epoch[099] Batch [0949]/[3760]	Speed: 34.975389 samples/sec	 accuracy=83.023026	 loss=0.679171	 lr=0.000100
Epoch[099] Batch [0999]/[3760]	Speed: 35.252864 samples/sec	 accuracy=82.973438	 loss=0.680331	 lr=0.000100
Epoch[099] Batch [1049]/[3760]	Speed: 35.124264 samples/sec	 accuracy=83.043155	 loss=0.678808	 lr=0.000100
Epoch[099] Batch [1099]/[3760]	Speed: 35.379696 samples/sec	 accuracy=83.009943	 loss=0.680027	 lr=0.000100
Epoch[099] Batch [1149]/[3760]	Speed: 35.138737 samples/sec	 accuracy=83.042120	 loss=0.679713	 lr=0.000100
Epoch[099] Batch [1199]/[3760]	Speed: 35.135980 samples/sec	 accuracy=83.058594	 loss=0.678816	 lr=0.000100
Epoch[099] Batch [1249]/[3760]	Speed: 35.007752 samples/sec	 accuracy=83.061250	 loss=0.678948	 lr=0.000100
Epoch[099] Batch [1299]/[3760]	Speed: 35.170141 samples/sec	 accuracy=83.069712	 loss=0.678645	 lr=0.000100
Epoch[099] Batch [1349]/[3760]	Speed: 34.842648 samples/sec	 accuracy=83.078704	 loss=0.679061	 lr=0.000100
Epoch[099] Batch [1399]/[3760]	Speed: 35.124550 samples/sec	 accuracy=83.063616	 loss=0.679787	 lr=0.000100
Epoch[099] Batch [1449]/[3760]	Speed: 34.718931 samples/sec	 accuracy=83.059267	 loss=0.681245	 lr=0.000100
Epoch[099] Batch [1499]/[3760]	Speed: 35.238913 samples/sec	 accuracy=83.086458	 loss=0.679726	 lr=0.000100
Epoch[099] Batch [1549]/[3760]	Speed: 34.876775 samples/sec	 accuracy=83.142137	 loss=0.678074	 lr=0.000100
Epoch[099] Batch [1599]/[3760]	Speed: 35.017620 samples/sec	 accuracy=83.112305	 loss=0.679174	 lr=0.000100
Epoch[099] Batch [1649]/[3760]	Speed: 34.983196 samples/sec	 accuracy=83.108902	 loss=0.678793	 lr=0.000100
Epoch[099] Batch [1699]/[3760]	Speed: 34.903419 samples/sec	 accuracy=83.133272	 loss=0.678386	 lr=0.000100
Epoch[099] Batch [1749]/[3760]	Speed: 35.107023 samples/sec	 accuracy=83.116964	 loss=0.679097	 lr=0.000100
Epoch[099] Batch [1799]/[3760]	Speed: 35.014784 samples/sec	 accuracy=83.140625	 loss=0.677739	 lr=0.000100
Epoch[099] Batch [1849]/[3760]	Speed: 35.128242 samples/sec	 accuracy=83.149493	 loss=0.677358	 lr=0.000100
Epoch[099] Batch [1899]/[3760]	Speed: 35.035870 samples/sec	 accuracy=83.156250	 loss=0.677640	 lr=0.000100
Epoch[099] Batch [1949]/[3760]	Speed: 35.147837 samples/sec	 accuracy=83.141827	 loss=0.677965	 lr=0.000100
Epoch[099] Batch [1999]/[3760]	Speed: 35.069569 samples/sec	 accuracy=83.155469	 loss=0.677224	 lr=0.000100
Epoch[099] Batch [2049]/[3760]	Speed: 35.246932 samples/sec	 accuracy=83.148628	 loss=0.677277	 lr=0.000100
Epoch[099] Batch [2099]/[3760]	Speed: 35.083027 samples/sec	 accuracy=83.150298	 loss=0.677291	 lr=0.000100
Epoch[099] Batch [2149]/[3760]	Speed: 35.463185 samples/sec	 accuracy=83.111919	 loss=0.678565	 lr=0.000100
Epoch[099] Batch [2199]/[3760]	Speed: 35.166918 samples/sec	 accuracy=83.115767	 loss=0.679015	 lr=0.000100
Epoch[099] Batch [2249]/[3760]	Speed: 35.297555 samples/sec	 accuracy=83.106250	 loss=0.679204	 lr=0.000100
Epoch[099] Batch [2299]/[3760]	Speed: 35.146108 samples/sec	 accuracy=83.104620	 loss=0.678962	 lr=0.000100
Epoch[099] Batch [2349]/[3760]	Speed: 35.119345 samples/sec	 accuracy=83.075133	 loss=0.679174	 lr=0.000100
Epoch[099] Batch [2399]/[3760]	Speed: 34.948560 samples/sec	 accuracy=83.096354	 loss=0.678351	 lr=0.000100
Epoch[099] Batch [2449]/[3760]	Speed: 35.271689 samples/sec	 accuracy=83.084184	 loss=0.678754	 lr=0.000100
Epoch[099] Batch [2499]/[3760]	Speed: 34.903266 samples/sec	 accuracy=83.111250	 loss=0.677735	 lr=0.000100
Epoch[099] Batch [2549]/[3760]	Speed: 35.240404 samples/sec	 accuracy=83.116422	 loss=0.677484	 lr=0.000100
Epoch[099] Batch [2599]/[3760]	Speed: 35.044090 samples/sec	 accuracy=83.117188	 loss=0.677466	 lr=0.000100
Epoch[099] Batch [2649]/[3760]	Speed: 35.201035 samples/sec	 accuracy=83.114387	 loss=0.677488	 lr=0.000100
Epoch[099] Batch [2699]/[3760]	Speed: 35.028647 samples/sec	 accuracy=83.115741	 loss=0.677353	 lr=0.000100
Epoch[099] Batch [2749]/[3760]	Speed: 34.901896 samples/sec	 accuracy=83.108523	 loss=0.677636	 lr=0.000100
Epoch[099] Batch [2799]/[3760]	Speed: 35.272880 samples/sec	 accuracy=83.096540	 loss=0.677980	 lr=0.000100
Epoch[099] Batch [2849]/[3760]	Speed: 34.794225 samples/sec	 accuracy=83.107456	 loss=0.677675	 lr=0.000100
Epoch[099] Batch [2899]/[3760]	Speed: 35.336455 samples/sec	 accuracy=83.077047	 loss=0.678699	 lr=0.000100
Epoch[099] Batch [2949]/[3760]	Speed: 34.850358 samples/sec	 accuracy=83.073623	 loss=0.678740	 lr=0.000100
Epoch[099] Batch [2999]/[3760]	Speed: 35.224757 samples/sec	 accuracy=83.058854	 loss=0.679215	 lr=0.000100
Epoch[099] Batch [3049]/[3760]	Speed: 34.842552 samples/sec	 accuracy=83.072746	 loss=0.678607	 lr=0.000100
Epoch[099] Batch [3099]/[3760]	Speed: 35.190910 samples/sec	 accuracy=83.100806	 loss=0.677646	 lr=0.000100
Epoch[099] Batch [3149]/[3760]	Speed: 35.068107 samples/sec	 accuracy=83.106151	 loss=0.677187	 lr=0.000100
Epoch[099] Batch [3199]/[3760]	Speed: 35.135807 samples/sec	 accuracy=83.103516	 loss=0.677235	 lr=0.000100
Epoch[099] Batch [3249]/[3760]	Speed: 35.046643 samples/sec	 accuracy=83.092788	 loss=0.677424	 lr=0.000100
Epoch[099] Batch [3299]/[3760]	Speed: 35.153126 samples/sec	 accuracy=83.101326	 loss=0.676998	 lr=0.000100
Epoch[099] Batch [3349]/[3760]	Speed: 35.134797 samples/sec	 accuracy=83.083489	 loss=0.677539	 lr=0.000100
Epoch[099] Batch [3399]/[3760]	Speed: 35.166700 samples/sec	 accuracy=83.092831	 loss=0.677350	 lr=0.000100
Epoch[099] Batch [3449]/[3760]	Speed: 35.100130 samples/sec	 accuracy=83.096920	 loss=0.677043	 lr=0.000100
Epoch[099] Batch [3499]/[3760]	Speed: 35.160124 samples/sec	 accuracy=83.104911	 loss=0.676488	 lr=0.000100
Epoch[099] Batch [3549]/[3760]	Speed: 35.266913 samples/sec	 accuracy=83.087588	 loss=0.677105	 lr=0.000100
Epoch[099] Batch [3599]/[3760]	Speed: 35.023182 samples/sec	 accuracy=83.091580	 loss=0.676855	 lr=0.000100
Epoch[099] Batch [3649]/[3760]	Speed: 35.076129 samples/sec	 accuracy=83.095890	 loss=0.676567	 lr=0.000100
Epoch[099] Batch [3699]/[3760]	Speed: 34.825437 samples/sec	 accuracy=83.105574	 loss=0.676182	 lr=0.000100
Epoch[099] Batch [3749]/[3760]	Speed: 36.404900 samples/sec	 accuracy=83.094583	 loss=0.676621	 lr=0.000100
Batch [0049]/[0303]: acc-top1=67.312500 acc-top5=85.875000
Batch [0099]/[0303]: acc-top1=66.640625 acc-top5=85.187500
Batch [0149]/[0303]: acc-top1=67.062500 acc-top5=85.395833
Batch [0199]/[0303]: acc-top1=66.765625 acc-top5=85.421875
Batch [0249]/[0303]: acc-top1=66.612500 acc-top5=85.337500
Batch [0299]/[0303]: acc-top1=66.505208 acc-top5=85.395833
[Epoch 099] training: accuracy=83.087600	 loss=0.676853
[Epoch 099] speed: 34 samples/sec	time cost: 7344.960073
[Epoch 099] validation: acc-top1=66.496493 acc-top5=85.406353 loss=1.645211