[代码解读] A ConvNet for the 2020s

代码地址：https://github.com/facebookresearch/ConvNeXt
ConvNeXt/models/convnext.py整体的代码结构如下：

接下来我们一部分一部分来看。

1.Block类：

class Block(nn.Module):
    r""" ConvNeXt Block. There are two equivalent implementations:
    (1) DwConv -> LayerNorm (channels_first) -> 1x1 Conv -> GELU -> 1x1 Conv; all in (N, C, H, W)
    (2) DwConv -> Permute to (N, H, W, C); LayerNorm (channels_last) -> Linear -> GELU -> Linear; Permute back
    We use (2) as we find it slightly faster in PyTorch
    
    Args:
        dim (int): Number of input channels.
        drop_path (float): Stochastic depth rate. Default: 0.0
        layer_scale_init_value (float): Init value for Layer Scale. Default: 1e-6.
    """
    def __init__(self, dim, drop_path=0., layer_scale_init_value=1e-6):
        super().__init__()
        self.dwconv = nn.Conv2d(dim, dim, kernel_size=7, padding=3, groups=dim) # depthwise conv
        self.norm = LayerNorm(dim, eps=1e-6)
        self.pwconv1 = nn.Linear(dim, 4 * dim) # pointwise/1x1 convs, implemented with linear layers
        self.act = nn.GELU()
        self.pwconv2 = nn.Linear(4 * dim, dim)
        self.gamma = nn.Parameter(layer_scale_init_value * torch.ones((dim)), 
                                    requires_grad=True) if layer_scale_init_value > 0 else None
        self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()

    def forward(self, x):
        input = x
        x = self.dwconv(x)
        x = x.permute(0, 2, 3, 1) # (N, C, H, W) -> (N, H, W, C)
        x = self.norm(x)
        x = self.pwconv1(x)
        x = self.act(x)
        x = self.pwconv2(x)
        if self.gamma is not None:
            x = self.gamma * x
        x = x.permute(0, 3, 1, 2) # (N, H, W, C) -> (N, C, H, W)

        x = input + self.drop_path(x)
        return x

在这个类的注释中有写到两种等价的操作：
（1）DwConv -> LayerNorm (channels_first) -> 1x1 Conv -> GELU -> 1x1 Conv; all in (N, C, H, W)
（2）DwConv -> Permute to (N, H, W, C); LayerNorm (channels_last) -> Linear -> GELU -> Linear; Permute back
在这个类中作者用了第二种，因为他们发现在pytorch中第二种方式稍微快一点。

1.1 初始化方法

def __init__(self, dim, drop_path=0., layer_scale_init_value=1e-6):
        super().__init__()
        self.dwconv = nn.Conv2d(dim, dim, kernel_size=7, padding=3, groups=dim) # depthwise conv
        self.norm = LayerNorm(dim, eps=1e-6)
        self.pwconv1 = nn.Linear(dim, 4 * dim) # pointwise/1x1 convs, implemented with linear layers
        self.act = nn.GELU()
        self.pwconv2 = nn.Linear(4 * dim, dim)
        self.gamma = nn.Parameter(layer_scale_init_value * torch.ones((dim)), 
                                    requires_grad=True) if layer_scale_init_value > 0 else None
        self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()

在初始化函数中，定义了depthwise卷积self.dwconv,具体实现是用输入通道=输出通道=groups的逐通道卷积，其核大小为7，padding为3。然后是归一化层self.norm，具体实现是LayerNorm。之后定义了pointwise/1x1卷积，由于上面所提到的原因，这里具体使用线性层来实现的。接下来是激活函数，依照论文中的结论，这里的激活函数为GELU。DropPath提供了一种针对网络分支的网络正则化方法。

1.2前向传播

def forward(self, x):
        input = x
        x = self.dwconv(x)
        x = x.permute(0, 2, 3, 1) # (N, C, H, W) -> (N, H, W, C)
        x = self.norm(x)
        x = self.pwconv1(x)
        x = self.act(x)
        x = self.pwconv2(x)
        if self.gamma is not None:
            x = self.gamma * x
        x = x.permute(0, 3, 1, 2) # (N, H, W, C) -> (N, C, H, W)

        x = input + self.drop_path(x)
        return x

前向传播过程就如（2）DwConv -> Permute to (N, H, W, C); LayerNorm (channels_last) -> Linear -> GELU -> Linear; Permute back中所示，最后这个block的示意图如下：

图中主要画出的关键步骤是一个depthwise卷积和两个pointwise卷积以及一个残差连接结构，在经过这样一个block后，输入输出的通道维度不变。
**总结：**在block类的具体代码中我们可以看到文章中所提到的一些设计要点，包括：分组数等于通道数的分组卷积、两头小中间大的逆瓶颈结构（以及自然而然的上移的深度卷积层）、使用了7x7卷积核的深度卷积、GeLU激活函数、一个块中只有一个归一化层和一个激活层和使用LN。

2.ConvNeXt类

class ConvNeXt(nn.Module):
    r""" ConvNeXt
        A PyTorch impl of : `A ConvNet for the 2020s`  -
          https://arxiv.org/pdf/2201.03545.pdf

    Args:
        in_chans (int): Number of input image channels. Default: 3
        num_classes (int): Number of classes for classification head. Default: 1000
        depths (tuple(int)): Number of blocks at each stage. Default: [3, 3, 9, 3]
        dims (int): Feature dimension at each stage. Default: [96, 192, 384, 768]
        drop_path_rate (float): Stochastic depth rate. Default: 0.
        layer_scale_init_value (float): Init value for Layer Scale. Default: 1e-6.
        head_init_scale (float): Init scaling value for classifier weights and biases. Default: 1.
    """
    def __init__(self, in_chans=3, num_classes=1000, 
                 depths=[3, 3, 9, 3], dims=[96, 192, 384, 768], drop_path_rate=0., 
                 layer_scale_init_value=1e-6, head_init_scale=1.,
                 ):
        super().__init__()

        self.downsample_layers = nn.ModuleList() # stem and 3 intermediate downsampling conv layers
        stem = nn.Sequential(
            nn.Conv2d(in_chans, dims[0], kernel_size=4, stride=4),
            LayerNorm(dims[0], eps=1e-6, data_format="channels_first")
        )
        self.downsample_layers.append(stem)
        for i in range(3):
            downsample_layer = nn.Sequential(
                    LayerNorm(dims[i], eps=1e-6, data_format="channels_first"),
                    nn.Conv2d(dims[i], dims[i+1], kernel_size=2, stride=2),
            )
            self.downsample_layers.append(downsample_layer)

        self.stages = nn.ModuleList() # 4 feature resolution stages, each consisting of multiple residual blocks
        dp_rates=[x.item() for x in torch.linspace(0, drop_path_rate, sum(depths))] 
        cur = 0
        for i in range(4):
            stage = nn.Sequential(
                *[Block(dim=dims[i], drop_path=dp_rates[cur + j], 
                layer_scale_init_value=layer_scale_init_value) for j in range(depths[i])]
            )
            self.stages.append(stage)
            cur += depths[i]

        self.norm = nn.LayerNorm(dims[-1], eps=1e-6) # final norm layer
        self.head = nn.Linear(dims[-1], num_classes)

        self.apply(self._init_weights)
        self.head.weight.data.mul_(head_init_scale)
        self.head.bias.data.mul_(head_init_scale)

    def _init_weights(self, m):
        if isinstance(m, (nn.Conv2d, nn.Linear)):
            trunc_normal_(m.weight, std=.02)
            nn.init.constant_(m.bias, 0)

    def forward_features(self, x):
        for i in range(4):
            x = self.downsample_layers[i](x)
            x = self.stages[i](x)
        return self.norm(x.mean([-2, -1])) # global average pooling, (N, C, H, W) -> (N, C)

    def forward(self, x):
        x = self.forward_features(x)
        x = self.head(x)
        return x

在这个类中除了初始化方法外还包括初始化权重方法_init_weights，以及前向传播方法forward_features和forward。

2.1 初始化方法

def __init__(self, in_chans=3, num_classes=1000, 
                 depths=[3, 3, 9, 3], dims=[96, 192, 384, 768], drop_path_rate=0., 
                 layer_scale_init_value=1e-6, head_init_scale=1.,
                 ):
        super().__init__()

        self.downsample_layers = nn.ModuleList() # stem and 3 intermediate downsampling conv layers
        stem = nn.Sequential(
            nn.Conv2d(in_chans, dims[0], kernel_size=4, stride=4),
            LayerNorm(dims[0], eps=1e-6, data_format="channels_first")
        )
        self.downsample_layers.append(stem)
        for i in range(3):
            downsample_layer = nn.Sequential(
                    LayerNorm(dims[i], eps=1e-6, data_format="channels_first"),
                    nn.Conv2d(dims[i], dims[i+1], kernel_size=2, stride=2),
            )
            self.downsample_layers.append(downsample_layer)

        self.stages = nn.ModuleList() # 4 feature resolution stages, each consisting of multiple residual blocks
        dp_rates=[x.item() for x in torch.linspace(0, drop_path_rate, sum(depths))] 
        cur = 0
        for i in range(4):
            stage = nn.Sequential(
                *[Block(dim=dims[i], drop_path=dp_rates[cur + j], 
                layer_scale_init_value=layer_scale_init_value) for j in range(depths[i])]
            )
            self.stages.append(stage)
            cur += depths[i]

        self.norm = nn.LayerNorm(dims[-1], eps=1e-6) # final norm layer
        self.head = nn.Linear(dims[-1], num_classes)

        self.apply(self._init_weights)
        self.head.weight.data.mul_(head_init_scale)
        self.head.bias.data.mul_(head_init_scale)

首先定义了一个self.downsample_layers用来装载根块和三个中间的下采样层。根块由步长为4的4x4一个卷积层和一个LN层构成，下采样层由一个LN层加一个步长为2的2x2卷积层构成，下采样一共有三个。然后定义了self.stages用来装载4个stage块，每个stage块由depths[i]所指定的n个ConvNeXt Block构成。之后定义了最后的一层归一化层self.norm,最后是用Linear定义了一个self.head,用来最终输出。

2.2初始化权重

def _init_weights(self, m):
        if isinstance(m, (nn.Conv2d, nn.Linear)):
            trunc_normal_(m.weight, std=.02)
            nn.init.constant_(m.bias, 0)

用截断正态分布来对权重初始化，偏置初始化为0。

2.3前向传播

def forward_features(self, x):
        for i in range(4):
            x = self.downsample_layers[i](x)
            x = self.stages[i](x)
        return self.norm(x.mean([-2, -1])) # global average pooling, (N, C, H, W) -> (N, C)

    def forward(self, x):
        x = self.forward_features(x)
        x = self.head(x)
        return x

整个前向传播过程就是上面初始化过程中所定义的块的堆叠，整个网络结构图如下图所示：

总结
在b这个类的具体代码中我们可以看到文章中所提到的其他的设计要点，包括：使用3：3：9：3的阶段比、使用核大小为4×4，步长为4的卷积层来实现patchify层。

3.结束语

总的来看ConvNeXt的确是非常简单，无论从结构还是思路上。用的大部分的设计都在之前的研究中有过讨论，同时也借鉴了Transformers出现后所带来的一些新的思路，在当时确实在兼顾效率的情况下展现了不错的性能。重新认识卷积在计算机视觉领域的作用是十分必要的。