GhostNet: More Features from Cheap Operations 【阅读记录】

code ：https: //github.com/huawei-noah/ghostnet

paper :https://arxiv.org/pdf/1911.11907.pdf

 

摘要：

核心思想：

1. 作者为了进一步压缩 CNN 网络结构，提出了一个 Ghost module，其核心是通过简单的线性变换，在内在特征图的基础上，生成更多可以完全揭示内在特征信息的 ghost feature map，从而以较小的计算代价生成更多特征；
2. 作者提出的 Ghost module 可以看作一个即插即用组件，用于升级现有的卷积神经网络，其核心是在输出通道数不变的情况下，减小卷积层的通道数并采用一个线性变换来升维，以此减小参数；
3. 同时作者也提出了一个 Ghost Bottleneck 模块用来堆叠 Ghost module，并以此构建了一个新的网络 GhostNet，实现了 75.7% top-1 准确率，在比 MobileNetV3 准确率还高的基础上，进一步压缩了模型；

如图1所示，卷积层的输出特征映射通常包含许多冗余，其中一些可能彼此相似。 我们指出，没有必要用大量的FLOP和参数逐个生成这些冗余特征映射。

分析：

 

 

 

 

# 2020.06.09-Changed for building GhostNet
#            Huawei Technologies Co., Ltd. <foss@huawei.com>
"""
Creates a GhostNet Model as defined in:
GhostNet: More Features from Cheap Operations By Kai Han, Yunhe Wang, Qi Tian, Jianyuan Guo, Chunjing Xu, Chang Xu.
https://arxiv.org/abs/1911.11907
Modified from https://github.com/d-li14/mobilenetv3.pytorch and https://github.com/rwightman/pytorch-image-models
"""
import torch
import torch.nn as nn
import torch.nn.functional as F
import math


__all__ = ['ghost_net']


def _make_divisible(v, divisor, min_value=None):
    """
    This function is taken from the original tf repo.
    It ensures that all layers have a channel number that is divisible by 8
    It can be seen here:
    https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet.py
    """
    if min_value is None:
        min_value = divisor
    new_v = max(min_value, int(v + divisor / 2) // divisor * divisor)
    # Make sure that round down does not go down by more than 10%.
    if new_v < 0.9 * v:
        new_v += divisor
    return new_v


def hard_sigmoid(x, inplace: bool = False):
    if inplace:
        return x.add_(3.).clamp_(0., 6.).div_(6.)
    else:
        return F.relu6(x + 3.) / 6.


class SqueezeExcite(nn.Module):
    def __init__(self, in_chs, se_ratio=0.25, reduced_base_chs=None,
                 act_layer=nn.ReLU, gate_fn=hard_sigmoid, divisor=4, **_):
        super(SqueezeExcite, self).__init__()
        self.gate_fn = gate_fn
        reduced_chs = _make_divisible((reduced_base_chs or in_chs) * se_ratio, divisor)
        self.avg_pool = nn.AdaptiveAvgPool2d(1)
        self.conv_reduce = nn.Conv2d(in_chs, reduced_chs, 1, bias=True)
        self.act1 = act_layer(inplace=True)
        self.conv_expand = nn.Conv2d(reduced_chs, in_chs, 1, bias=True)

    def forward(self, x):
        x_se = self.avg_pool(x)
        x_se = self.conv_reduce(x_se)
        x_se = self.act1(x_se)
        x_se = self.conv_expand(x_se)
        x = x * self.gate_fn(x_se)
        return x    

    
class ConvBnAct(nn.Module):
    def __init__(self, in_chs, out_chs, kernel_size,
                 stride=1, act_layer=nn.ReLU):
        super(ConvBnAct, self).__init__()
        self.conv = nn.Conv2d(in_chs, out_chs, kernel_size, stride, kernel_size//2, bias=False)
        self.bn1 = nn.BatchNorm2d(out_chs)
        self.act1 = act_layer(inplace=True)

    def forward(self, x):
        x = self.conv(x)
        x = self.bn1(x)
        x = self.act1(x)
        return x


class GhostModule(nn.Module):
    def __init__(self, inp, oup, kernel_size=1, ratio=2, dw_size=3, stride=1, relu=True):
        super(GhostModule, self).__init__()
        self.oup = oup
        init_channels = math.ceil(oup / ratio)
        new_channels = init_channels*(ratio-1)

        self.primary_conv = nn.Sequential(
            nn.Conv2d(inp, init_channels, kernel_size, stride, kernel_size//2, bias=False),
            nn.BatchNorm2d(init_channels),
            nn.ReLU(inplace=True) if relu else nn.Sequential(),
        )

        self.cheap_operation = nn.Sequential(
            nn.Conv2d(init_channels, new_channels, dw_size, 1, dw_size//2, groups=init_channels, bias=False),
            nn.BatchNorm2d(new_channels),
            nn.ReLU(inplace=True) if relu else nn.Sequential(),
        )

    def forward(self, x):
        x1 = self.primary_conv(x)
        x2 = self.cheap_operation(x1)
        out = torch.cat([x1,x2], dim=1)
        return out[:,:self.oup,:,:]


class GhostBottleneck(nn.Module):
    """ Ghost bottleneck w/ optional SE"""

    def __init__(self, in_chs, mid_chs, out_chs, dw_kernel_size=3,
                 stride=1, act_layer=nn.ReLU, se_ratio=0.):
        super(GhostBottleneck, self).__init__()
        has_se = se_ratio is not None and se_ratio > 0.
        self.stride = stride

        # Point-wise expansion
        self.ghost1 = GhostModule(in_chs, mid_chs, relu=True)

        # Depth-wise convolution
        if self.stride > 1:
            self.conv_dw = nn.Conv2d(mid_chs, mid_chs, dw_kernel_size, stride=stride,
                             padding=(dw_kernel_size-1)//2,
                             groups=mid_chs, bias=False)
            self.bn_dw = nn.BatchNorm2d(mid_chs)

        # Squeeze-and-excitation
        if has_se:
            self.se = SqueezeExcite(mid_chs, se_ratio=se_ratio)
        else:
            self.se = None

        # Point-wise linear projection
        self.ghost2 = GhostModule(mid_chs, out_chs, relu=False)
        
        # shortcut
        if (in_chs == out_chs and self.stride == 1):
            self.shortcut = nn.Sequential()
        else:
            self.shortcut = nn.Sequential(
                nn.Conv2d(in_chs, in_chs, dw_kernel_size, stride=stride,
                       padding=(dw_kernel_size-1)//2, groups=in_chs, bias=False),
                nn.BatchNorm2d(in_chs),
                nn.Conv2d(in_chs, out_chs, 1, stride=1, padding=0, bias=False),
                nn.BatchNorm2d(out_chs),
            )


    def forward(self, x):
        residual = x

        # 1st ghost bottleneck
        x = self.ghost1(x)

        # Depth-wise convolution
        if self.stride > 1:
            x = self.conv_dw(x)
            x = self.bn_dw(x)

        # Squeeze-and-excitation
        if self.se is not None:
            x = self.se(x)

        # 2nd ghost bottleneck
        x = self.ghost2(x)
        
        x += self.shortcut(residual)
        return x


class GhostNet(nn.Module):
    def __init__(self, cfgs, num_classes=1000, width=1.0, dropout=0.2):
        super(GhostNet, self).__init__()
        # setting of inverted residual blocks
        self.cfgs = cfgs
        self.dropout = dropout

        # building first layer
        output_channel = _make_divisible(16 * width, 4)
        self.conv_stem = nn.Conv2d(3, output_channel, 3, 2, 1, bias=False)
        self.bn1 = nn.BatchNorm2d(output_channel)
        self.act1 = nn.ReLU(inplace=True)
        input_channel = output_channel

        # building inverted residual blocks
        stages = []
        block = GhostBottleneck
        for cfg in self.cfgs:
            layers = []
            for k, exp_size, c, se_ratio, s in cfg:
                output_channel = _make_divisible(c * width, 4)
                hidden_channel = _make_divisible(exp_size * width, 4)
                layers.append(block(input_channel, hidden_channel, output_channel, k, s,
                              se_ratio=se_ratio))
                input_channel = output_channel
            stages.append(nn.Sequential(*layers))

        output_channel = _make_divisible(exp_size * width, 4)
        stages.append(nn.Sequential(ConvBnAct(input_channel, output_channel, 1)))
        input_channel = output_channel
        
        self.blocks = nn.Sequential(*stages)        

        # building last several layers
        output_channel = 1280
        self.global_pool = nn.AdaptiveAvgPool2d((1, 1))
        self.conv_head = nn.Conv2d(input_channel, output_channel, 1, 1, 0, bias=True)
        self.act2 = nn.ReLU(inplace=True)
        self.classifier = nn.Linear(output_channel, num_classes)

    def forward(self, x):
        x = self.conv_stem(x)
        x = self.bn1(x)
        x = self.act1(x)
        x = self.blocks(x)
        x = self.global_pool(x)
        x = self.conv_head(x)
        x = self.act2(x)
        x = x.view(x.size(0), -1)
        if self.dropout > 0.:
            x = F.dropout(x, p=self.dropout, training=self.training)
        x = self.classifier(x)
        return x


def ghostnet(**kwargs):
    """
    Constructs a GhostNet model
    """
    cfgs = [
        # k, t, c, SE, s 
        # stage1
        [[3,  16,  16, 0, 1]],
        # stage2
        [[3,  48,  24, 0, 2]],
        [[3,  72,  24, 0, 1]],
        # stage3
        [[5,  72,  40, 0.25, 2]],
        [[5, 120,  40, 0.25, 1]],
        # stage4
        [[3, 240,  80, 0, 2]],
        [[3, 200,  80, 0, 1],
         [3, 184,  80, 0, 1],
         [3, 184,  80, 0, 1],
         [3, 480, 112, 0.25, 1],
         [3, 672, 112, 0.25, 1]
        ],
        # stage5
        [[5, 672, 160, 0.25, 2]],
        [[5, 960, 160, 0, 1],
         [5, 960, 160, 0.25, 1],
         [5, 960, 160, 0, 1],
         [5, 960, 160, 0.25, 1]
        ]
    ]
    return GhostNet(cfgs, **kwargs)


if __name__=='__main__':
    model = ghostnet()
    model.eval()
    print(model)
    input = torch.randn(32,3,320,256)
    y = model(input)
    print(y.size())

网络结构：

GhostNet(
  (conv_stem): Conv2d(3, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
  (bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (act1): ReLU(inplace=True)
  (blocks): Sequential(
    (0): Sequential(
      (0): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(16, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(8, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(8, 8, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=8, bias=False)
            (1): BatchNorm2d(8, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(16, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(8, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(8, 8, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=8, bias=False)
            (1): BatchNorm2d(8, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential()
      )
    )
    (1): Sequential(
      (0): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(16, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(24, 24, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=24, bias=False)
            (1): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (conv_dw): Conv2d(48, 48, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=48, bias=False)
        (bn_dw): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(48, 12, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(12, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(12, 12, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=12, bias=False)
            (1): BatchNorm2d(12, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential(
          (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=16, bias=False)
          (1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (2): Conv2d(16, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (3): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
    )
    (2): Sequential(
      (0): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(24, 36, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(36, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(36, 36, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=36, bias=False)
            (1): BatchNorm2d(36, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(72, 12, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(12, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(12, 12, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=12, bias=False)
            (1): BatchNorm2d(12, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential()
      )
    )
    (3): Sequential(
      (0): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(24, 36, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(36, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(36, 36, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=36, bias=False)
            (1): BatchNorm2d(36, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (conv_dw): Conv2d(72, 72, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=72, bias=False)
        (bn_dw): BatchNorm2d(72, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (se): SqueezeExcite(
          (avg_pool): AdaptiveAvgPool2d(output_size=1)
          (conv_reduce): Conv2d(72, 20, kernel_size=(1, 1), stride=(1, 1))
          (act1): ReLU(inplace=True)
          (conv_expand): Conv2d(20, 72, kernel_size=(1, 1), stride=(1, 1))
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(72, 20, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(20, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(20, 20, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=20, bias=False)
            (1): BatchNorm2d(20, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential(
          (0): Conv2d(24, 24, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=24, bias=False)
          (1): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (2): Conv2d(24, 40, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (3): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
    )
    (4): Sequential(
      (0): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(40, 60, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(60, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(60, 60, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=60, bias=False)
            (1): BatchNorm2d(60, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (se): SqueezeExcite(
          (avg_pool): AdaptiveAvgPool2d(output_size=1)
          (conv_reduce): Conv2d(120, 32, kernel_size=(1, 1), stride=(1, 1))
          (act1): ReLU(inplace=True)
          (conv_expand): Conv2d(32, 120, kernel_size=(1, 1), stride=(1, 1))
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(120, 20, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(20, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(20, 20, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=20, bias=False)
            (1): BatchNorm2d(20, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential()
      )
    )
    (5): Sequential(
      (0): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(40, 120, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(120, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(120, 120, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=120, bias=False)
            (1): BatchNorm2d(120, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (conv_dw): Conv2d(240, 240, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=240, bias=False)
        (bn_dw): BatchNorm2d(240, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(240, 40, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(40, 40, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=40, bias=False)
            (1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential(
          (0): Conv2d(40, 40, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=40, bias=False)
          (1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (2): Conv2d(40, 80, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (3): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
    )
    (6): Sequential(
      (0): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(80, 100, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(100, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(100, 100, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=100, bias=False)
            (1): BatchNorm2d(100, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(200, 40, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(40, 40, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=40, bias=False)
            (1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential()
      )
      (1): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(80, 92, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(92, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(92, 92, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=92, bias=False)
            (1): BatchNorm2d(92, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(184, 40, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(40, 40, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=40, bias=False)
            (1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential()
      )
      (2): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(80, 92, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(92, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(92, 92, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=92, bias=False)
            (1): BatchNorm2d(92, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(184, 40, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(40, 40, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=40, bias=False)
            (1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential()
      )
      (3): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(80, 240, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(240, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(240, 240, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=240, bias=False)
            (1): BatchNorm2d(240, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (se): SqueezeExcite(
          (avg_pool): AdaptiveAvgPool2d(output_size=1)
          (conv_reduce): Conv2d(480, 120, kernel_size=(1, 1), stride=(1, 1))
          (act1): ReLU(inplace=True)
          (conv_expand): Conv2d(120, 480, kernel_size=(1, 1), stride=(1, 1))
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(480, 56, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(56, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(56, 56, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=56, bias=False)
            (1): BatchNorm2d(56, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential(
          (0): Conv2d(80, 80, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=80, bias=False)
          (1): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (2): Conv2d(80, 112, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (3): BatchNorm2d(112, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
      (4): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(112, 336, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(336, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(336, 336, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=336, bias=False)
            (1): BatchNorm2d(336, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (se): SqueezeExcite(
          (avg_pool): AdaptiveAvgPool2d(output_size=1)
          (conv_reduce): Conv2d(672, 168, kernel_size=(1, 1), stride=(1, 1))
          (act1): ReLU(inplace=True)
          (conv_expand): Conv2d(168, 672, kernel_size=(1, 1), stride=(1, 1))
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(672, 56, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(56, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(56, 56, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=56, bias=False)
            (1): BatchNorm2d(56, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential()
      )
    )
    (7): Sequential(
      (0): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(112, 336, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(336, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(336, 336, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=336, bias=False)
            (1): BatchNorm2d(336, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (conv_dw): Conv2d(672, 672, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=672, bias=False)
        (bn_dw): BatchNorm2d(672, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (se): SqueezeExcite(
          (avg_pool): AdaptiveAvgPool2d(output_size=1)
          (conv_reduce): Conv2d(672, 168, kernel_size=(1, 1), stride=(1, 1))
          (act1): ReLU(inplace=True)
          (conv_expand): Conv2d(168, 672, kernel_size=(1, 1), stride=(1, 1))
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(672, 80, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(80, 80, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=80, bias=False)
            (1): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential(
          (0): Conv2d(112, 112, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=112, bias=False)
          (1): BatchNorm2d(112, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (2): Conv2d(112, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (3): BatchNorm2d(160, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
    )
    (8): Sequential(
      (0): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(160, 480, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(480, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(480, 480, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=480, bias=False)
            (1): BatchNorm2d(480, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(960, 80, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(80, 80, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=80, bias=False)
            (1): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential()
      )
      (1): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(160, 480, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(480, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(480, 480, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=480, bias=False)
            (1): BatchNorm2d(480, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (se): SqueezeExcite(
          (avg_pool): AdaptiveAvgPool2d(output_size=1)
          (conv_reduce): Conv2d(960, 240, kernel_size=(1, 1), stride=(1, 1))
          (act1): ReLU(inplace=True)
          (conv_expand): Conv2d(240, 960, kernel_size=(1, 1), stride=(1, 1))
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(960, 80, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(80, 80, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=80, bias=False)
            (1): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential()
      )
      (2): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(160, 480, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(480, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(480, 480, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=480, bias=False)
            (1): BatchNorm2d(480, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(960, 80, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(80, 80, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=80, bias=False)
            (1): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential()
      )
      (3): GhostBottleneck(
        (ghost1): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(160, 480, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(480, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
          (cheap_operation): Sequential(
            (0): Conv2d(480, 480, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=480, bias=False)
            (1): BatchNorm2d(480, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): ReLU(inplace=True)
          )
        )
        (se): SqueezeExcite(
          (avg_pool): AdaptiveAvgPool2d(output_size=1)
          (conv_reduce): Conv2d(960, 240, kernel_size=(1, 1), stride=(1, 1))
          (act1): ReLU(inplace=True)
          (conv_expand): Conv2d(240, 960, kernel_size=(1, 1), stride=(1, 1))
        )
        (ghost2): GhostModule(
          (primary_conv): Sequential(
            (0): Conv2d(960, 80, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
          (cheap_operation): Sequential(
            (0): Conv2d(80, 80, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=80, bias=False)
            (1): BatchNorm2d(80, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
            (2): Sequential()
          )
        )
        (shortcut): Sequential()
      )
    )
    (9): Sequential(
      (0): ConvBnAct(
        (conv): Conv2d(160, 960, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (bn1): BatchNorm2d(960, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (act1): ReLU(inplace=True)
      )
    )
  )
  (global_pool): AdaptiveAvgPool2d(output_size=(1, 1))
  (conv_head): Conv2d(960, 1280, kernel_size=(1, 1), stride=(1, 1))
  (act2): ReLU(inplace=True)
  (classifier): Linear(in_features=1280, out_features=1000, bias=True)
)