Swin-Transformer通用视觉骨干网络主体结构代码解释

Swin-Transformer由MSRA视觉计算组的team于2021年发表的工作,在多个视觉任务以及多个数据集上均取得了十分优秀的结果。在这里,我贴出我对于Swin-Transformer主体结构的一些代码的解释和tensor的shape的改变,由于时间的原因,可能会出现许多纰漏,希望大家多多指教

paper:https://arxiv.org/pdf/2111.09883v1.pdf

code:GitHub - microsoft/Swin-Transformer: This is an official implementation for "Swin Transformer: Hierarchical Vision Transformer using Shifted Windows".

import torch
import torch.nn as nn
import torch.utils.checkpoint as checkpoint
from timm.models.layers import DropPath, to_2tuple, trunc_normal_
#-------------------------------#
# 此为对于MLP模块的定义
#-------------------------------#
class Mlp(nn.Module):
    def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
        super(Mlp,self).__init__()
        #---------------------------------------#
        # 在这里使用Dropout的作用在于
        # 降低因为Linear层的使用所造成的过拟合现象的发生
        #---------------------------------------#
        out_features    = out_features or in_features
        hidden_features = hidden_features or in_features
        self.fc1        = nn.Linear(in_features, hidden_features)
        self.act        = act_layer()
        self.fc2        = nn.Linear(hidden_features, out_features)
        self.drop       = nn.Dropout(drop)

    def forward(self, x):
        x = self.fc1(x)
        x = self.act(x)
        x = self.drop(x)
        x = self.fc2(x)
        x = self.drop(x)
        return x

#--------------------------------------------------#
# 对于window_partition的定义
# 在这里x是一个tensor
# window size由自己定义
# view()的作用相当于numpy中的reshape,重新定义矩阵的形状
#--------------------------------------------------#
def window_partition(x, window_size):
    """
    Args:
        x: (B, H, W, C)
        window_size (int): window size

    Returns:
        windows: (num_windows**2*B, window_size, window_size, C)
    """
    B, H, W, C = x.shape
    x          = x.view(B, H // window_size, window_size, W // window_size, window_size, C)
    #-------------------------------------------------------------------------------------------#
    # 1. shape = B, H // window_size, W // window_size, window_size, window_size, C
    # 2. shape = num_windows**2*B, window_size, window_size, C(返回值window的shape)
    # 3. view(-1, window_size, window_size, C)的含义为 后三维度已经确定,第一维由整体矩阵根据后三个维度得到
    # 4. H // window_size = W // window_size即为 num_windows
    #-------------------------------------------------------------------------------------------#
    windows    = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
    return windows

#-----------------------------------#
# 此为对于window_reverse函数的定义
#-----------------------------------#
def window_reverse(windows, window_size, H, W):
    """
    Args:
        windows: (num_windows**2*B, window_size, window_size, C)
        window_size (int): Window size
        H (int): Height of image
        W (int): Width of image

    Returns:
        x: (B, H, W, C)
    """
    B = int(windows.shape[0] / (H * W / window_size / window_size))
    x = windows.view(B, H // window_size, W // window_size, window_size, window_size, -1)
    x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, H, W, -1)
    return x

#--------------------------------#
# 对于WindowAttention这个类的定义
# 该类支持滑动的以及未滑动的窗口图像
#--------------------------------#
class WindowAttention(nn.Module):
    r""" Window based multi-head self attention (W-MSA) module with relative position bias.
    It supports both of shifted and non-shifted window.

    Args:
        dim (int): Number of input channels.
        window_size (tuple[int]): The height and width of the window.
        num_heads (int): Number of attention heads.
        qkv_bias (bool, optional):  If True, add a learnable bias to query, key, value. Default: True
        qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set
        attn_drop (float, optional): Dropout ratio of attention weight. Default: 0.0
        proj_drop (float, optional): Dropout ratio of output. Default: 0.0
    """

    def __init__(self, dim, window_size, num_heads, qkv_bias=True, qk_scale=None, attn_drop=0., proj_drop=0.):
        super(WindowAttention,self).__init__()
        head_dim = dim // num_heads
        self.dim         = dim
        self.window_size = window_size  # Wh,Ww
        self.num_heads   = num_heads
        self.scale       = qk_scale or head_dim ** -0.5
        #----------------------------------------------------------------#
        # define a parameter table of relative position bias
        # return: 2*Wh-1 * 2*Ww-1, nH
        # nn.Parameter作用为定义这些参数是可以学习的参数
        # torch.zeros():其形状由变量参数size定义,返回一个由标量值0填充的张量
        #----------------------------------------------------------------#
        self.relative_position_bias_table = nn.Parameter(torch.zeros((2 * window_size[0] - 1) * (2 * window_size[1] - 1), num_heads))
        #----------------------------------------------------------------------------#
        # get pair-wise relative position index for each token inside the window
        # torch.arange()所创建的张量是int类型
        # torch.meshgrid()的作用在于将两个类型相同的张量生成一个tensor矩阵
        # 这个矩阵的行数为第一个input_tensor的维度,列数为第二个input_tensor的维度
        # 之后又进行了stack操作,又增加了一个维度,所以此时的shape为2, Wh, Ww
        # 之后通过flatten将后两个维度压缩成一个维度,此时的shape为2, Wh*Ww
        #----------------------------------------------------------------------------#
        coords_h        = torch.arange(self.window_size[0])
        coords_w        = torch.arange(self.window_size[1])
        coords          = torch.stack(torch.meshgrid([coords_h, coords_w]))
        coords_flatten  = torch.flatten(coords, 1)
        #----------------------------------------------#
        # [:, :, None]其中的None代表增加一个维度,具体的值为1
        # relative_coords的shape为2, Wh*Ww, Wh*Ww
        # 之后对其进行转置,此时的shape为Wh*Ww, Wh*Ww, 2
        #----------------------------------------------#
        relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]
        relative_coords = relative_coords.permute(1, 2, 0).contiguous()
        relative_coords[:, :, 0] += self.window_size[0] - 1  # shift to start from 0
        relative_coords[:, :, 1] += self.window_size[1] - 1
        relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1
        #---------------------------------------------#
        # relative_position_index的shape为Wh*Ww, Wh*Ww
        # 之后将它作为一个模型的常数
        #----------------------------------------- ---#
        relative_position_index   = relative_coords.sum(-1)
        self.register_buffer("relative_position_index", relative_position_index)
        #--------------------------------------#
        # 定义qkv以及proj,在他们之后均有一个Dropout
        # 以降低过拟合发生的风险
        #--------------------------------------#
        self.qkv       = nn.Linear(dim, dim * 3, bias=qkv_bias)
        self.attn_drop = nn.Dropout(attn_drop)
        self.proj      = nn.Linear(dim, dim)
        self.proj_drop = nn.Dropout(proj_drop)
        #---------------------------------#
        # 利用正太分布来生成一个点
        # 之后又定义了一个softmax分类器
        #---------------------------------#
        trunc_normal_(self.relative_position_bias_table, std=.02)
        self.softmax   = nn.Softmax(dim=-1)
    #------------------------------------------------------#
    # 该类前向传播函数的定义
    # input-x即为shape:num_windows**2*B,N(windows_size), C
    #------------------------------------------------------#
    def forward(self, x, mask=None):
        """
        Args:
            x: input features with shape of (num_windows*B, N, C)
            mask: (0/-inf) mask with shape of (num_windows, Wh*Ww, Wh*Ww) or None
        """
        B_, N, C = x.shape
        #--------------------------------------------------#
        # 首先给qkv添加几个维度,之后再进行转置
        # shape= 3,B_,self.num_heads,N,C // self.num_heads
        # 之后便可以得知:q=3, k=B_, v=self.num_heads
        # 之后再求解q的值,最后进行注意力的计算
        #--------------------------------------------------#
        qkv      = self.qkv(x).reshape(B_, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
        q, k, v  = qkv[0], qkv[1], qkv[2]
        q        = q * self.scale
        attn     = (q @ k.transpose(-2, -1))
        #-----------------------------------------------#
        # relative_position_bias的shape为Wh*Ww,Wh*Ww,nH
        # 之后又进行转置,此时的shape为nH, Wh*Ww, Wh*Ww
        # 之后又再第一维增加了一个维度
        #-----------------------------------------------#
        relative_position_bias = self.relative_position_bias_table[self.relative_position_index.view(-1)].view(self.window_size[0] * self.window_size[1],self.window_size[0] * self.window_size[1], -1)
        relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()
        attn                   = attn + relative_position_bias.unsqueeze(0)
        #----------------------------#
        # 在这里我们的mask的值为None
        # 所以直接pass through softmax
        # 之后过dropout
        #----------------------------#
        if mask is not None:
            nW   = mask.shape[0]
            attn = attn.view(B_ // nW, nW, self.num_heads, N, N) + mask.unsqueeze(1).unsqueeze(0)
            attn = attn.view(-1, self.num_heads, N, N)
            attn = self.softmax(attn)
        else:
            attn = self.softmax(attn)

        attn     = self.attn_drop(attn)
        #------------------------#
        # 进行注意力加权运算
        # 之后过投影层
        # 最后过投影层的dropout
        #------------------------#
        x = (attn @ v).transpose(1, 2).reshape(B_, N, C)
        x = self.proj(x)
        x = self.proj_drop(x)
        return x
    #------------------------------#
    # extra_repr以及flops函数的定义
    #------------------------------#
    def extra_repr(self) -> str:
        return f'dim={self.dim}, window_size={self.window_size}, num_heads={self.num_heads}'

    def flops(self, N):
        # calculate flops for 1 window with token length of N
        flops  = 0
        # qkv  = self.qkv(x)
        flops += N * self.dim * 3 * self.dim
        # attn = (q @ k.transpose(-2, -1))
        flops += self.num_heads * N * (self.dim // self.num_heads) * N
        #  x   = (attn @ v)
        flops += self.num_heads * N * N * (self.dim // self.num_heads)
        # x    = self.proj(x)
        flops += N * self.dim * self.dim
        return flops


class SwinTransformerBlock(nn.Module):
    r""" Swin Transformer Block.

    Args:
        dim (int): Number of input channels. 输入图像的通道数
        input_resolution (tuple[int]): Input resulotion. 输入图像的分辨率
        num_heads (int): Number of attention heads.  多头注意力中注意力头的数目
        window_size (int): Window size.
        shift_size (int): Shift size for SW-MSA. 一次滑动操作所滑动的尺寸
        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
        qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
        qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
        drop (float, optional): Dropout rate. Default: 0.0
        attn_drop (float, optional): Attention dropout rate. Default: 0.0
        drop_path (float, optional): Stochastic depth rate. Default: 0.0
        act_layer (nn.Module, optional): Activation layer. Default: nn.GELU
        norm_layer (nn.Module, optional): Normalization layer.  Default: nn.LayerNorm
    """
    def __init__(self, dim, input_resolution, num_heads, window_size=7, shift_size=0,
                 mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0., drop_path=0.,
                 act_layer=nn.GELU, norm_layer=nn.LayerNorm):
        super(SwinTransformerBlock,self).__init__()
        self.dim              = dim
        self.input_resolution = input_resolution
        self.num_heads        = num_heads
        self.window_size      = window_size
        self.shift_size       = shift_size
        self.mlp_ratio        = mlp_ratio
        #--------------------------------#
        # 若input的尺寸小于窗口的大小
        #--------------------------------#
        if min(self.input_resolution) <= self.window_size:
            # if window size is larger than input resolution, we don't partition windows
            self.shift_size  = 0
            self.window_size = min(self.input_resolution)
        assert 0 <= self.shift_size < self.window_size, "shift_size must in 0-window_size"
        #-------------------------------------#
        # 定义使用LayerNorm并且定义了窗口注意力机制
        # nn.Identity()相当于pass
        #-------------------------------------#
        self.norm1     = norm_layer(dim)
        self.attn      = WindowAttention(dim, window_size=to_2tuple(self.window_size), num_heads=num_heads,qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop)
        self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
        self.norm2     = norm_layer(dim)
        mlp_hidden_dim = int(dim * mlp_ratio)
        self.mlp       = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)

        if self.shift_size > 0:
            # calculate attention mask for SW-MSA
            # slice用于对数组元素进行截取,返回值为截取元素组成的一个新数组
            H, W     = self.input_resolution
            img_mask = torch.zeros((1, H, W, 1))  # 1 H W 1
            h_slices = (slice(0, -self.window_size),
                        slice(-self.window_size, -self.shift_size),
                        slice(-self.shift_size, None))
            w_slices = (slice(0, -self.window_size),
                        slice(-self.window_size, -self.shift_size),
                        slice(-self.shift_size, None))
            cnt      = 0
            #---------------------------------------------#
            # 根据h_slices以及w_slices求取cnt的值并进行相关操作
            #---------------------------------------------#
            for h in h_slices:
                for w in w_slices:
                    img_mask[:, h, w, :] = cnt
                    cnt += 1

            mask_windows = window_partition(img_mask, self.window_size)  # nW, window_size, window_size, 1
            mask_windows = mask_windows.view(-1, self.window_size * self.window_size)
            attn_mask    = mask_windows.unsqueeze(1) - mask_windows.unsqueeze(2)
            attn_mask    = attn_mask.masked_fill(attn_mask != 0, float(-100.0)).masked_fill(attn_mask == 0, float(0.0))
        else:
            attn_mask = None

        self.register_buffer("attn_mask", attn_mask)

    def forward(self, x):
        H, W     = self.input_resolution
        B, L, C  = x.shape
        assert L == H * W, "input feature has wrong size"

        shortcut = x
        x        = self.norm1(x)
        x        = x.view(B, H, W, C)

        # cyclic shift
        if self.shift_size > 0:
            shifted_x = torch.roll(x, shifts=(-self.shift_size, -self.shift_size), dims=(1, 2))
        else:
            shifted_x = x

        # partition windows
        x_windows = window_partition(shifted_x, self.window_size)  # nW*B, window_size, window_size, C
        x_windows = x_windows.view(-1, self.window_size * self.window_size, C)  # nW*B, window_size*window_size, C

        # W-MSA/SW-MSA
        attn_windows = self.attn(x_windows, mask=self.attn_mask)  # nW*B, window_size*window_size, C

        # merge windows
        attn_windows = attn_windows.view(-1, self.window_size, self.window_size, C)
        shifted_x = window_reverse(attn_windows, self.window_size, H, W)  # B H' W' C

        # reverse cyclic shift
        if self.shift_size > 0:
            x = torch.roll(shifted_x, shifts=(self.shift_size, self.shift_size), dims=(1, 2))
        else:
            x = shifted_x
        x = x.view(B, H * W, C)

        # FFN
        x = shortcut + self.drop_path(x)
        x = x + self.drop_path(self.mlp(self.norm2(x)))

        return x

    def extra_repr(self) -> str:
        return f"dim={self.dim}, input_resolution={self.input_resolution}, num_heads={self.num_heads}, " \
               f"window_size={self.window_size}, shift_size={self.shift_size}, mlp_ratio={self.mlp_ratio}"

    def flops(self):
        flops  = 0
        H, W   = self.input_resolution
        # norm1
        flops += self.dim * H * W
        # W-MSA/SW-MSA
        nW     = H * W / self.window_size / self.window_size
        flops += nW * self.attn.flops(self.window_size * self.window_size)
        # mlp
        flops += 2 * H * W * self.dim * self.dim * self.mlp_ratio
        # norm2
        flops += self.dim * H * W
        return flops

#--------------------------#
# PatchMerging这个类的定义
#--------------------------#
class PatchMerging(nn.Module):
    r""" Patch Merging Layer.

    Args:
        input_resolution (tuple[int]): Resolution of input feature.
        dim (int): Number of input channels.
        norm_layer (nn.Module, optional): Normalization layer.  Default: nn.LayerNorm
    """

    def __init__(self, input_resolution, dim, norm_layer=nn.LayerNorm):
        super(PatchMerging,self).__init__()
        self.input_resolution = input_resolution
        self.dim              = dim
        self.reduction        = nn.Linear(4 * dim, 2 * dim, bias=False)
        self.norm             = norm_layer(4 * dim)

    def forward(self, x):
        #-------------------#
        # x的shape:B,H*W,C
        #-------------------#
        H, W    = self.input_resolution
        B, L, C = x.shape
        assert L == H * W, "input feature has wrong size"
        assert H % 2 == 0 and W % 2 == 0, f"x size ({H}*{W}) are not even."

        x  = x.view(B, H, W, C)
        x0 = x[:, 0::2, 0::2, :]  # B H/2 W/2 C
        x1 = x[:, 1::2, 0::2, :]  # B H/2 W/2 C
        x2 = x[:, 0::2, 1::2, :]  # B H/2 W/2 C
        x3 = x[:, 1::2, 1::2, :]  # B H/2 W/2 C
        #-------------------------------#
        # 在最后一个维度将四个tensor进行拼接
        # tensor的shape为 B H/2 W/2 4*C
        # 之后进行view操作即为reshape
        # 此时张量的shape为B H/2*W/2 4*C
        # 之后过LayerNorm
        # 最后通过全连接层来降低通道数
        # 此时的shape为B H/2*W/2 2*C
        #-------------------------------#
        x  = torch.cat([x0, x1, x2, x3], -1)
        x  = x.view(B, -1, 4 * C)
        x  = self.norm(x)
        x  = self.reduction(x)
        return x
    #-------------------------------#
    # 定义extra_repr以及flops这两个函数
    #-------------------------------#
    def extra_repr(self) -> str:
        return f"input_resolution={self.input_resolution}, dim={self.dim}"

    def flops(self):
        H, W   = self.input_resolution
        flops  = H * W * self.dim
        flops += (H // 2) * (W // 2) * 4 * self.dim * 2 * self.dim
        return flops

#----------------------------------------------------#
# SwinTransformer整体结构中共有四个stage
# 四个stage中的layer的数目分别为2 2 6 2
# 这个类即为对于一个stage中所用到的layer的定义
#----------------------------------------------------#
class BasicLayer(nn.Module):
    r""" A basic Swin Transformer layer for one stage.
    Args:
        dim (int): Number of input channels.
        input_resolution (tuple[int]): Input resolution.
        depth (int): Number of blocks.
        num_heads (int): Number of attention heads.
        window_size (int): Local window size.
        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
        qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
        qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
        drop (float, optional): Dropout rate. Default: 0.0
        attn_drop (float, optional): Attention dropout rate. Default: 0.0
        drop_path (float | tuple[float], optional): Stochastic depth rate. Default: 0.0
        norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
        downsample (nn.Module | None, optional): Downsample layer at the end of the layer. Default: None
        use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False. 是否使用checkpointing来节省内存,默认值为False
    """
    def __init__(self, dim, input_resolution, depth, num_heads, window_size,
                 mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0.,
                 drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False):
        super(BasicLayer,self).__init__()
        self.dim              = dim
        self.input_resolution = input_resolution
        self.depth            = depth
        self.use_checkpoint   = use_checkpoint
        #-----------------------------------#
        # build blocks
        # 根据depth的值来确定一个layer中所用到的
        # SwinTransformerBlock的块数
        #-----------------------------------#
        self.blocks = nn.ModuleList([
            SwinTransformerBlock(dim=dim, input_resolution=input_resolution,
                                 num_heads=num_heads, window_size=window_size,
                                 shift_size=0 if (i % 2 == 0) else window_size // 2,
                                 mlp_ratio=mlp_ratio,
                                 qkv_bias=qkv_bias, qk_scale=qk_scale,
                                 drop=drop, attn_drop=attn_drop,
                                 drop_path=drop_path[i] if isinstance(drop_path, list) else drop_path,
                                 norm_layer=norm_layer)
            for i in range(depth)])
        #-----------------------------------------------------#
        # patch merging layer
        # 在这里我们定义的downsample的值为None所以无downsample操作
        #-----------------------------------------------------#
        if downsample is not None:
            self.downsample = downsample(input_resolution, dim=dim, norm_layer=norm_layer)
        else:
            self.downsample = None

    def forward(self, x):
        for blk in self.blocks:
            if self.use_checkpoint:
                x = checkpoint.checkpoint(blk, x)
            else:
                x = blk(x)
        if self.downsample is not None:
            x = self.downsample(x)
        return x

    def extra_repr(self) -> str:
        return f"dim={self.dim}, input_resolution={self.input_resolution}, depth={self.depth}"

    def flops(self):
        flops = 0
        for blk in self.blocks:
            flops += blk.flops()
        if self.downsample is not None:
            flops += self.downsample.flops()
        return flops

#------------------------------#
# PatchEmbed这个类的定义
#------------------------------#
class PatchEmbed(nn.Module):
    r""" Image to Patch Embedding
    Args:
        img_size (int): Image size.  Default: 224.
        patch_size (int): Patch token size. Default: 4.
        in_chans (int): Number of input image channels. Default: 3.
        embed_dim (int): Number of linear projection output channels. Default: 96.
        norm_layer (nn.Module, optional): Normalization layer. Default: None
    """
    def __init__(self, img_size=224, patch_size=4, in_chans=3, embed_dim=96, norm_layer=None):
        super(PatchEmbed,self).__init__()
        #--------------------------------#
        # to_2tuple()的作用在于生成一个元组
        # 并且该元组中有两个值相同的元素
        #--------------------------------#
        img_size                = to_2tuple(img_size)
        patch_size              = to_2tuple(patch_size)
        patches_resolution      = [img_size[0] // patch_size[0], img_size[1] // patch_size[1]]
        self.img_size           = img_size
        self.patch_size         = patch_size
        self.patches_resolution = patches_resolution
        self.num_patches        = patches_resolution[0] * patches_resolution[1]
        self.in_chans           = in_chans
        self.embed_dim          = embed_dim
        self.proj               = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)
        if norm_layer is not None:
            self.norm = norm_layer(embed_dim)
        else:
            self.norm = None

    def forward(self, x):
        B, C, H, W = x.shape
        # FIXME look at relaxing size constraints
        assert H == self.img_size[0] and W == self.img_size[1], \
            f"Input image size ({H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})."
        #-----------------------------------------------#
        # 1.先进行一次卷积,此时shape=batch_size, C, Ph, Pw
        # 2.将后两个维度进行flatten,使其成为Ph*Pw
        # 3.进行转置,此时的shape为batch_size,Ph*Pw C
        #-----------------------------------------------#
        x = self.proj(x).flatten(2).transpose(1, 2)
        if self.norm is not None:
            x = self.norm(x)
        return x

    def flops(self):
        Ho, Wo = self.patches_resolution
        flops  = Ho * Wo * self.embed_dim * self.in_chans * (self.patch_size[0] * self.patch_size[1])
        if self.norm is not None:
            flops += Ho * Wo * self.embed_dim
        return flops

#----------------------------------------------#
# The build of SwinTransformer
# 若在一个类中想要调用另外一个类,则我们并不
# 需要定义self方法,我们使用该类的类名直接进行调用即可
#----------------------------------------------#
class SwinTransformer(nn.Module):
    r""" Swin Transformer
        A PyTorch impl of : `Swin Transformer: Hierarchical Vision Transformer using Shifted Windows`  -
          https://arxiv.org/pdf/2103.14030
    Args:
        img_size (int | tuple(int)): Input image size. Default 224
        patch_size (int | tuple(int)): Patch size. Default: 4
        in_chans (int): Number of input image channels. Default: 3
        num_classes (int): Number of classes for classification head. Default: 1000 -> ImageNet
        embed_dim (int): Patch embedding dimension. Default: 96
        depths (tuple(int)): Depth of each Swin Transformer layer.
        num_heads (tuple(int)): Number of attention heads in different layers.
        window_size (int): Window size. Default: 7
        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim. Default: 4
        qkv_bias (bool): If True, add a learnable bias to query, key, value. Default: True
        qk_scale (float): Override default qk scale of head_dim ** -0.5 if set. Default: None
        drop_rate (float): Dropout rate. Default: 0
        attn_drop_rate (float): Attention dropout rate. Default: 0
        drop_path_rate (float): Stochastic depth rate. Default: 0.1
        norm_layer (nn.Module): Normalization layer. Default: nn.LayerNorm.
        ape (bool): If True, add absolute position embedding to the patch embedding. Default: False
        patch_norm (bool): If True, add normalization after patch embedding. Default: True
        use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False
    """
    def __init__(self, img_size=224, patch_size=4, in_chans=3, num_classes=1000,
                 embed_dim=96, depths=[2, 2, 6, 2], num_heads=[3, 6, 12, 24],
                 window_size=7, mlp_ratio=4., qkv_bias=True, qk_scale=None,
                 drop_rate=0., attn_drop_rate=0., drop_path_rate=0.1,
                 norm_layer=nn.LayerNorm, ape=False, patch_norm=True,
                 use_checkpoint=False, **kwargs):
        super(SwinTransformer,self).__init__()
        self.num_classes  = num_classes
        self.num_layers   = len(depths)
        self.embed_dim    = embed_dim
        self.ape          = ape
        self.patch_norm   = patch_norm
        self.num_features = int(embed_dim * 2 ** (self.num_layers - 1))
        self.mlp_ratio    = mlp_ratio
        #------------------------------------------------#
        # split image into non-overlapping patches
        # 将输入的图片split成多个patch
        #------------------------------------------------#
        self.patch_embed = PatchEmbed(
            img_size   = img_size, patch_size=patch_size, in_chans=in_chans, embed_dim=embed_dim,
            norm_layer = norm_layer if self.patch_norm else None)
        num_patches             = self.patch_embed.num_patches
        patches_resolution      = self.patch_embed.patches_resolution
        self.patches_resolution = patches_resolution
        #---------------------------------------#
        # absolute position embedding
        # 类中定义的ape=False,所以我们直接pass即可
        #---------------------------------------#
        if self.ape:
            self.absolute_pos_embed = nn.Parameter(torch.zeros(1, num_patches, embed_dim))
            trunc_normal_(self.absolute_pos_embed, std=.02)
        #----------------------------#
        # 定义一个Dropout层
        #----------------------------#
        self.pos_drop = nn.Dropout(p=drop_rate)

        # stochastic depth
        dpr = [x.item() for x in torch.linspace(0, drop_path_rate, sum(depths))]  # stochastic depth decay rule
        #--------------------------------------------------#
        # build layers
        # 首先定义一个空的ModuleList
        # 之后根据layer的数目将BasicLayer添加至ModuleList中
        #--------------------------------------------------#
        self.layers = nn.ModuleList()
        for i_layer in range(self.num_layers):
            layer = BasicLayer(dim=int(embed_dim * 2 ** i_layer),
                               input_resolution=(patches_resolution[0] // (2 ** i_layer),patches_resolution[1] // (2 ** i_layer)),
                               depth          = depths[i_layer],
                               num_heads      = num_heads[i_layer],
                               window_size    = window_size,
                               mlp_ratio      = self.mlp_ratio,
                               qkv_bias       = qkv_bias, qk_scale=qk_scale,
                               drop           = drop_rate, attn_drop=attn_drop_rate,
                               drop_path      = dpr[sum(depths[:i_layer]):sum(depths[:i_layer + 1])],
                               norm_layer     = norm_layer,
                               downsample     = PatchMerging if (i_layer < self.num_layers - 1) else None,
                               use_checkpoint = use_checkpoint)
            self.layers.append(layer)
        #-------------------------------------#
        # 定义LayerNorm层
        # 定义自适应二维平均池化层
        # 定义head,使用Linear来实现
        # 之后进行了一个网络权重初始化的定义
        #-------------------------------------#
        self.norm    = norm_layer(self.num_features)
        self.avgpool = nn.AdaptiveAvgPool1d(1)
        self.head    = nn.Linear(self.num_features, num_classes) if num_classes > 0 else nn.Identity()
        self.apply(self._init_weights)
    #-------------------------------#
    # 网络的权重初始化
    #-------------------------------#
    def _init_weights(self, m):
        if isinstance(m, nn.Linear):
            trunc_normal_(m.weight, std=.02)
            if isinstance(m, nn.Linear) and m.bias is not None:
                nn.init.constant_(m.bias, 0)
        elif isinstance(m, nn.LayerNorm):
            nn.init.constant_(m.bias, 0)
            nn.init.constant_(m.weight, 1.0)

    @torch.jit.ignore
    def no_weight_decay(self):
        return {'absolute_pos_embed'}

    @torch.jit.ignore
    def no_weight_decay_keywords(self):
        return {'relative_position_bias_table'}
    #--------------------------------------------------------------------------------------#
    # forward_features函数的定义
    # 首先将输入的图片打成多个小的patch
    # 之后过Dropout来降低发生过拟合的风险,因为Linear层的存在所以说会定义比较多层的Dropout
    # 过SwinTransformer的四个stage,过LayerNorm,再过二维平均池化层,最后将后面的两个维度进行flatten
    #--------------------------------------------------------------------------------------#
    def forward_features(self, x):
        x = self.patch_embed(x)
        if self.ape:
            x = x + self.absolute_pos_embed
        x = self.pos_drop(x)

        for layer in self.layers:
            x = layer(x)
        #-----------------------------#
        # shape的变化
        #  B L C -> B C 1 -> B C
        #-----------------------------#
        x = self.norm(x)
        x = self.avgpool(x.transpose(1, 2))
        x = torch.flatten(x, 1)
        return x
    #---------------------------------#
    # 主要为运用Linear层将特征图像的
    # 通道数转变为ImageNet上所规定的类别数
    #---------------------------------#
    def forward(self, x):
        x = self.forward_features(x)
        x = self.head(x)
        return x

    def flops(self):
        flops = 0
        flops += self.patch_embed.flops()
        for i, layer in enumerate(self.layers):
            flops += layer.flops()
        flops += self.num_features * self.patches_resolution[0] * self.patches_resolution[1] // (2 ** self.num_layers)
        flops += self.num_features * self.num_classes
        return flops

你可能感兴趣的:(transformer,深度学习,计算机视觉)