目录
摘要
Abstract
一、Swin Transformer
1.1 输入
1.2 Patch Partition
1.3 Linear Embedding
1.4 Patch Merging
1.5 Swin Transformer Block
1.6 代码
二、MLP-Mixer
2.1 网络模型整体结构
2.2 Mixer Layer
2.3 MLP
总结
摘要
本篇博客介绍了采用类似于卷积核的移动窗口进行图像特征提取的Swin Transformer网络模型，详细学习了该模型每一个组成模块的网络结构和参数传递过程。博客第二章介绍了不使用卷积和自注意力模块的MLP-Mixer模型，阐述了在全使用全连接的情况下会拥有哪些优势。在每一章的最后都会附上复现的PyTorch代码。
Abstract
This blog introduces the Swin Transformer network model that uses a moving window similar to a convolutional kernel for image feature extraction, and provides a detailed study of the network structure and parameter transfer process of each component module in the model. Chapter 2 of the blog introduced the MLP Mixer model without using convolution and self attention modules, and explained the advantages it would have when fully connected. At the end of each chapter, a reproduced PyTorch code will be attached.
一、Swin Transformer
在上一篇博客中ViT通过将图片打成patch传入Transformer之中，虽然简单直接，但也存在一些问题。ViT是按照语言模型的逻辑去处理图片，对所有patch都做了多头自注意力，但是图片可能不需要将所有信息都注意到，例如图片的左上角和右下角很难出现在同一特征中。这样就会导致一些多余的计算，浪费了资源。
于是Swin Transformer提出了类似CNN卷积核的Shifted Windows（移动窗口），以减少patch之间做自注意力的次数，达到节省时间的目的。
接下来我将参照Swin Transformer模型整体结构图来讲解该模型，网络结构图如下所示：
1.1 输入
以输入图像大小为 224x224 为例，将图像tensor 224x224x3 从1处传入模型作为开始。
1.2 Patch Partition
到达第2步Patch Partiton，因为是以Transformer为基础，如同ViT一样需要将图片打成patch传入模型。这里将图像大小 224x224 通过卷积核大小为 4x4，步伐为4的卷积将图像转换为 56x56 的大小，通道数从3变为3x4x4=48维。
1.3 Linear Embedding
接下来数据将传入Swin Transformer最重要的第3大部分，首先在第一个stage会经过全连接层，这里的通道数在论文中是设置好的，即C=96。如上图所示，56x56x48 的tensor经过Linear Embedding维度变化为56x56x96。
1.4 Patch Merging
这里先说一下Patch Merging，因为只要第一个stage先经过Linear Embedding，后3个stage都是Patch Merging。这里的Patch Merging类似于CNN中的池化，为了使窗口获得多尺度的特征，需要在每一个block之前进行下采样，这样在窗口大小不变的情况下图像变小、维度增加，可以获得更多的特征信息。如下图所示：
论文中提到将图像下采样2倍，在Patch Merging中会隔一个点采样。细心的人这时会发现，我们图像的维度从 4x4x1 变为了 2x2x4，但是在模型中是从
变为
，即大小缩小一半，维度扩大2倍。但是在上图的操作中维度扩大了4倍，论文中为了使其与CNN中池化保持一致，在Patch Merging之后会在再通过一次卷积使维度缩小一半，以达到大小缩小一半，维度扩大2倍的效果。
到这一步Swin Transformer已经拥有了感知多尺度特征的手段。
1.5 Swin Transformer Block
接下来看看Swin Transformer中最重要的模块，这里是用到了Transformer中的编码器，但是做了一些改动，Transformer中是单纯的多头自注意力（MSA），而这里用到了W-MSA和SW-MSA。
- W-MSA
引入Windows Multi-head Self-Attention（窗口多头自注意力）就是为了解决开篇提到的计算量问题。W-MSA会将图像划分为 MxM （在论文中M默认为7）的不重叠窗口，然后多头自注意力只在同一个窗口内做，这样就从全局的自注意力变为了窗口内的自注意力，大大减小了计算量。
如上公式，若图像长宽为 224x224 、M=7，则从
降为
，还是肉眼可见的降低不少。
- SW-MSA
如果只进行W-MSA会发现一个问题，窗口之间是没用通信的，也就是窗口与窗口之间是独立计算的，就不能像卷积一样移动去感受不同像素范围。于是，作者引入了Shifted Windows Multi-Head Self-Attention（SW-MSA）模块让窗口移动起来。
例如上图所示，将图像分为了4个窗口，在W-MSA中4个窗口分别做MSA；然后在下一层SW-MSA左上角的窗口会向右和向下移动
个步伐。如右图所示将图像分割为9个窗口，会发现上一层不属于同一个窗口的patch，经过移动后融合为一个窗口，窗口与窗口之间成功进行了信息交流。
如果9个窗口分别进行SW-MSA，那么从上一层4个相同的窗口变为9个不完全相同的窗口进行计算，不但不利于并行计算，而且窗口数量也增大两倍多，通过W-MSA积攒的计算优势一去不复返。
那么该如何解决呢？有人想到padding补0，这样虽然可以保证窗口大小相同进行并行计算，但是数量是增多的。论文中，作者提出通过掩码的方式进行自注意力的计算。
A、B、C经过顺时针循环位移，将图像窗口补为大小相同的4个窗口。但是同一个窗口中，颜色不同的部分不能直接进行自注意力，需要在计算之后分别乘上对应的掩码。
我的理解如下：
掩码是加上一个很大的负值，在经过softmax之后该部分就会变为0。
1的部分全属于同一窗口可直接进行自注意力计算。
作者也对这部分进行详细的解释，另两个部分的掩码各不相同，但是都是按照上图逻辑计算，我就不一一展开，作者解释如下图所示：
在完成上述掩码计算之后，还需将循环移动的窗口还原。
图像在经过最后一个stage之后大小变为 7x7x768，也就是图像全局特征。该模型在该层之后没有再连接其他操作，因为作者是想把该模型打造成一个万能的模块，可以将这些全局特征应用到各种下游任务。例如，将 7x7x768 送入全连接层输出 1x768，再将 1x768 升维至 1x1000 便可用作ImageNet数据集的分类操作。
在Swin Transformer block之中W-MSA和SW-MSA一般是成对使用，但具体的参数变化和窗口移动按具体任务而定，上述文章主要介绍核心内容。
1.6 代码
代码采用Swin Transformer tiny模型，训练利用迁移学习，模型预训练来自ImageNet-1K数据集，模型参数为1k，窗口大小为 7x7，Swin Transformer block分别为2、2、6、2。
Swin Transformer网络模型如下：
""" Swin Transformer
A PyTorch impl of : `Swin Transformer: Hierarchical Vision Transformer using Shifted Windows`
- https://arxiv.org/pdf/2103.14030
Code/weights from https://github.com/microsoft/Swin-Transformer
"""
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.utils.checkpoint as checkpoint
import numpy as np
from typing import Optional
def drop_path_f(x, drop_prob: float = 0., training: bool = False):
"""Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).
This is the same as the DropConnect impl I created for EfficientNet, etc networks, however,
the original name is misleading as 'Drop Connect' is a different form of dropout in a separate paper...
See discussion: https://github.com/tensorflow/tpu/issues/494#issuecomment-532968956 ... I've opted for
changing the layer and argument names to 'drop path' rather than mix DropConnect as a layer name and use
'survival rate' as the argument.
"""
if drop_prob == 0. or not training:
return x
keep_prob = 1 - drop_prob
shape = (x.shape[0],) + (1,) * (x.ndim - 1) # work with diff dim tensors, not just 2D ConvNets
random_tensor = keep_prob + torch.rand(shape, dtype=x.dtype, device=x.device)
random_tensor.floor_() # binarize
output = x.div(keep_prob) * random_tensor
return output
class DropPath(nn.Module):
"""Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).
"""
def __init__(self, drop_prob=None):
super(DropPath, self).__init__()
self.drop_prob = drop_prob
def forward(self, x):
return drop_path_f(x, self.drop_prob, self.training)
def window_partition(x, window_size: int):
"""
将feature map按照window_size划分成一个个没有重叠的window
Args:
x: (B, H, W, C)
window_size (int): window size(M)
Returns:
windows: (num_windows*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)
# permute: [B, H//Mh, Mh, W//Mw, Mw, C] -> [B, H//Mh, W//Mh, Mw, Mw, C]
# view: [B, H//Mh, W//Mw, Mh, Mw, C] -> [B*num_windows, Mh, Mw, C]
windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
return windows
def window_reverse(windows, window_size: int, H: int, W: int):
"""
将一个个window还原成一个feature map
Args:
windows: (num_windows*B, window_size, window_size, C)
window_size (int): Window size(M)
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))
# view: [B*num_windows, Mh, Mw, C] -> [B, H//Mh, W//Mw, Mh, Mw, C]
x = windows.view(B, H // window_size, W // window_size, window_size, window_size, -1)
# permute: [B, H//Mh, W//Mw, Mh, Mw, C] -> [B, H//Mh, Mh, W//Mw, Mw, C]
# view: [B, H//Mh, Mh, W//Mw, Mw, C] -> [B, H, W, C]
x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, H, W, -1)
return x
class PatchEmbed(nn.Module):
"""
2D Image to Patch Embedding
"""
def __init__(self, patch_size=4, in_c=3, embed_dim=96, norm_layer=None):
super().__init__()
patch_size = (patch_size, patch_size)
self.patch_size = patch_size
self.in_chans = in_c
self.embed_dim = embed_dim
self.proj = nn.Conv2d(in_c, embed_dim, kernel_size=patch_size, stride=patch_size)
self.norm = norm_layer(embed_dim) if norm_layer else nn.Identity()
def forward(self, x):
_, _, H, W = x.shape
# padding
# 如果输入图片的H，W不是patch_size的整数倍，需要进行padding
pad_input = (H % self.patch_size[0] != 0) or (W % self.patch_size[1] != 0)
if pad_input:
# to pad the last 3 dimensions,
# (W_left, W_right, H_top,H_bottom, C_front, C_back)
x = F.pad(x, (0, self.patch_size[1] - W % self.patch_size[1],
0, self.patch_size[0] - H % self.patch_size[0],
0, 0))
# 下采样patch_size倍
x = self.proj(x)
_, _, H, W = x.shape
# flatten: [B, C, H, W] -> [B, C, HW]
# transpose: [B, C, HW] -> [B, HW, C]
x = x.flatten(2).transpose(1, 2)
x = self.norm(x)
return x, H, W
class PatchMerging(nn.Module):
r""" Patch Merging Layer.
Args:
dim (int): Number of input channels.
norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
"""
def __init__(self, dim, norm_layer=nn.LayerNorm):
super().__init__()
self.dim = dim
self.reduction = nn.Linear(4 * dim, 2 * dim, bias=False)
self.norm = norm_layer(4 * dim)
def forward(self, x, H, W):
"""
x: B, H*W, C
"""
B, L, C = x.shape
assert L == H * W, "input feature has wrong size"
x = x.view(B, H, W, C)
# padding
# 如果输入feature map的H，W不是2的整数倍，需要进行padding
pad_input = (H % 2 == 1) or (W % 2 == 1)
if pad_input:
# to pad the last 3 dimensions, starting from the last dimension and moving forward.
# (C_front, C_back, W_left, W_right, H_top, H_bottom)
# 注意这里的Tensor通道是[B, H, W, C]，所以会和官方文档有些不同
x = F.pad(x, (0, 0, 0, W % 2, 0, H % 2))
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]
x = torch.cat([x0, x1, x2, x3], -1) # [B, H/2, W/2, 4*C]
x = x.view(B, -1, 4 * C) # [B, H/2*W/2, 4*C]
x = self.norm(x)
x = self.reduction(x) # [B, H/2*W/2, 2*C]
return x
class Mlp(nn.Module):
""" MLP as used in Vision Transformer, MLP-Mixer and related networks
"""
def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
super().__init__()
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.drop1 = nn.Dropout(drop)
self.fc2 = nn.Linear(hidden_features, out_features)
self.drop2 = nn.Dropout(drop)
def forward(self, x):
x = self.fc1(x)
x = self.act(x)
x = self.drop1(x)
x = self.fc2(x)
x = self.drop2(x)
return x
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
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, attn_drop=0., proj_drop=0.):
super().__init__()
self.dim = dim
self.window_size = window_size # [Mh, Mw]
self.num_heads = num_heads
head_dim = dim // num_heads
self.scale = head_dim ** -0.5
# define a parameter table of relative position bias
self.relative_position_bias_table = nn.Parameter(
torch.zeros((2 * window_size[0] - 1) * (2 * window_size[1] - 1), num_heads)) # [2*Mh-1 * 2*Mw-1, nH]
# get pair-wise relative position index for each token inside the window
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], indexing="ij")) # [2, Mh, Mw]
coords_flatten = torch.flatten(coords, 1) # [2, Mh*Mw]
# [2, Mh*Mw, 1] - [2, 1, Mh*Mw]
relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :] # [2, Mh*Mw, Mh*Mw]
relative_coords = relative_coords.permute(1, 2, 0).contiguous() # [Mh*Mw, Mh*Mw, 2]
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 = relative_coords.sum(-1) # [Mh*Mw, Mh*Mw]
self.register_buffer("relative_position_index", relative_position_index)
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)
nn.init.trunc_normal_(self.relative_position_bias_table, std=.02)
self.softmax = nn.Softmax(dim=-1)
def forward(self, x, mask: Optional[torch.Tensor] = None):
"""
Args:
x: input features with shape of (num_windows*B, Mh*Mw, C)
mask: (0/-inf) mask with shape of (num_windows, Wh*Ww, Wh*Ww) or None
"""
# [batch_size*num_windows, Mh*Mw, total_embed_dim]
B_, N, C = x.shape
# qkv(): -> [batch_size*num_windows, Mh*Mw, 3 * total_embed_dim]
# reshape: -> [batch_size*num_windows, Mh*Mw, 3, num_heads, embed_dim_per_head]
# permute: -> [3, batch_size*num_windows, num_heads, Mh*Mw, embed_dim_per_head]
qkv = self.qkv(x).reshape(B_, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
# [batch_size*num_windows, num_heads, Mh*Mw, embed_dim_per_head]
q, k, v = qkv.unbind(0) # make torchscript happy (cannot use tensor as tuple)
# transpose: -> [batch_size*num_windows, num_heads, embed_dim_per_head, Mh*Mw]
# @: multiply -> [batch_size*num_windows, num_heads, Mh*Mw, Mh*Mw]
q = q * self.scale
attn = (q @ k.transpose(-2, -1))
# relative_position_bias_table.view: [Mh*Mw*Mh*Mw,nH] -> [Mh*Mw,Mh*Mw,nH]
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() # [nH, Mh*Mw, Mh*Mw]
attn = attn + relative_position_bias.unsqueeze(0)
if mask is not None:
# mask: [nW, Mh*Mw, Mh*Mw]
nW = mask.shape[0] # num_windows
# attn.view: [batch_size, num_windows, num_heads, Mh*Mw, Mh*Mw]
# mask.unsqueeze: [1, nW, 1, Mh*Mw, Mh*Mw]
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)
# @: multiply -> [batch_size*num_windows, num_heads, Mh*Mw, embed_dim_per_head]
# transpose: -> [batch_size*num_windows, Mh*Mw, num_heads, embed_dim_per_head]
# reshape: -> [batch_size*num_windows, Mh*Mw, total_embed_dim]
x = (attn @ v).transpose(1, 2).reshape(B_, N, C)
x = self.proj(x)
x = self.proj_drop(x)
return x
class SwinTransformerBlock(nn.Module):
r""" Swin Transformer Block.
Args:
dim (int): Number of input channels.
num_heads (int): Number of attention heads.
window_size (int): Window size.
shift_size (int): Shift size for SW-MSA.
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
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, num_heads, window_size=7, shift_size=0,
mlp_ratio=4., qkv_bias=True, drop=0., attn_drop=0., drop_path=0.,
act_layer=nn.GELU, norm_layer=nn.LayerNorm):
super().__init__()
self.dim = dim
self.num_heads = num_heads
self.window_size = window_size
self.shift_size = shift_size
self.mlp_ratio = mlp_ratio
assert 0 <= self.shift_size < self.window_size, "shift_size must in 0-window_size"
self.norm1 = norm_layer(dim)
self.attn = WindowAttention(
dim, window_size=(self.window_size, self.window_size), num_heads=num_heads, qkv_bias=qkv_bias,
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)
def forward(self, x, attn_mask):
H, W = self.H, self.W
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)
# pad feature maps to multiples of window size
# 把feature map给pad到window size的整数倍
pad_l = pad_t = 0
pad_r = (self.window_size - W % self.window_size) % self.window_size
pad_b = (self.window_size - H % self.window_size) % self.window_size
x = F.pad(x, (0, 0, pad_l, pad_r, pad_t, pad_b))
_, Hp, Wp, _ = x.shape
# 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
attn_mask = None
# partition windows
x_windows = window_partition(shifted_x, self.window_size) # [nW*B, Mh, Mw, C]
x_windows = x_windows.view(-1, self.window_size * self.window_size, C) # [nW*B, Mh*Mw, C]
# W-MSA/SW-MSA
attn_windows = self.attn(x_windows, mask=attn_mask) # [nW*B, Mh*Mw, C]
# merge windows
attn_windows = attn_windows.view(-1, self.window_size, self.window_size, C) # [nW*B, Mh, Mw, C]
shifted_x = window_reverse(attn_windows, self.window_size, Hp, Wp) # [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
if pad_r > 0 or pad_b > 0:
# 把前面pad的数据移除掉
x = x[:, :H, :W, :].contiguous()
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
class BasicLayer(nn.Module):
"""
A basic Swin Transformer layer for one stage.
Args:
dim (int): Number of input channels.
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
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.
"""
def __init__(self, dim, depth, num_heads, window_size,
mlp_ratio=4., qkv_bias=True, drop=0., attn_drop=0.,
drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False):
super().__init__()
self.dim = dim
self.depth = depth
self.window_size = window_size
self.use_checkpoint = use_checkpoint
self.shift_size = window_size // 2
# build blocks
self.blocks = nn.ModuleList([
SwinTransformerBlock(
dim=dim,
num_heads=num_heads,
window_size=window_size,
shift_size=0 if (i % 2 == 0) else self.shift_size,
mlp_ratio=mlp_ratio,
qkv_bias=qkv_bias,
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
if downsample is not None:
self.downsample = downsample(dim=dim, norm_layer=norm_layer)
else:
self.downsample = None
def create_mask(self, x, H, W):
# calculate attention mask for SW-MSA
# 保证Hp和Wp是window_size的整数倍
Hp = int(np.ceil(H / self.window_size)) * self.window_size
Wp = int(np.ceil(W / self.window_size)) * self.window_size
# 拥有和feature map一样的通道排列顺序，方便后续window_partition
img_mask = torch.zeros((1, Hp, Wp, 1), device=x.device) # [1, Hp, Wp, 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
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, Mh, Mw, 1]
mask_windows = mask_windows.view(-1, self.window_size * self.window_size) # [nW, Mh*Mw]
attn_mask = mask_windows.unsqueeze(1) - mask_windows.unsqueeze(2) # [nW, 1, Mh*Mw] - [nW, Mh*Mw, 1]
# [nW, Mh*Mw, Mh*Mw]
attn_mask = attn_mask.masked_fill(attn_mask != 0, float(-100.0)).masked_fill(attn_mask == 0, float(0.0))
return attn_mask
def forward(self, x, H, W):
attn_mask = self.create_mask(x, H, W) # [nW, Mh*Mw, Mh*Mw]
for blk in self.blocks:
blk.H, blk.W = H, W
if not torch.jit.is_scripting() and self.use_checkpoint:
x = checkpoint.checkpoint(blk, x, attn_mask)
else:
x = blk(x, attn_mask)
if self.downsample is not None:
x = self.downsample(x, H, W)
H, W = (H + 1) // 2, (W + 1) // 2
return x, H, W
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:
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
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
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.
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, 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,
drop_rate=0., attn_drop_rate=0., drop_path_rate=0.1,
norm_layer=nn.LayerNorm, patch_norm=True,
use_checkpoint=False, **kwargs):
super().__init__()
self.num_classes = num_classes
self.num_layers = len(depths)
self.embed_dim = embed_dim
self.patch_norm = patch_norm
# stage4输出特征矩阵的channels
self.num_features = int(embed_dim * 2 ** (self.num_layers - 1))
self.mlp_ratio = mlp_ratio
# split image into non-overlapping patches
self.patch_embed = PatchEmbed(
patch_size=patch_size, in_c=in_chans, embed_dim=embed_dim,
norm_layer=norm_layer if self.patch_norm else None)
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
self.layers = nn.ModuleList()
for i_layer in range(self.num_layers):
# 注意这里构建的stage和论文图中有些差异
# 这里的stage不包含该stage的patch_merging层，包含的是下个stage的
layers = BasicLayer(dim=int(embed_dim * 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,
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(layers)
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)
def _init_weights(self, m):
if isinstance(m, nn.Linear):
nn.init.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)
def forward(self, x):
# x: [B, L, C]
x, H, W = self.patch_embed(x)
x = self.pos_drop(x)
for layer in self.layers:
x, H, W = layer(x, H, W)
x = self.norm(x) # [B, L, C]
x = self.avgpool(x.transpose(1, 2)) # [B, C, 1]
x = torch.flatten(x, 1)
x = self.head(x)
return x
def swin_tiny_patch4_window7_224(num_classes: int = 1000, **kwargs):
# trained ImageNet-1K
# https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_tiny_patch4_window7_224.pth
model = SwinTransformer(in_chans=3,
patch_size=4,
window_size=7,
embed_dim=96,
depths=(2, 2, 6, 2),
num_heads=(3, 6, 12, 24),
num_classes=num_classes,
**kwargs)
return model引入预训练模型之后，进行花类数据集（上篇博客提到）进行微调，代码如下：
import os
import argparse
import torch
import torch.optim as optim
from torch.utils.tensorboard import SummaryWriter
from torchvision import transforms
from my_dataset import MyDataSet
from model import swin_tiny_patch4_window7_224 as create_model
from utils import read_split_data, train_one_epoch, evaluate
def main(args):
device = torch.device(args.device if torch.cuda.is_available() else "cpu")
if os.path.exists("../weights") is False:
os.makedirs("../weights")
tb_writer = SummaryWriter()
train_images_path, train_images_label, val_images_path, val_images_label = read_split_data(args.data_path)
img_size = 224
data_transform = {
"train": transforms.Compose([transforms.RandomResizedCrop(img_size),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])]),
"val": transforms.Compose([transforms.Resize(int(img_size * 1.143)),
transforms.CenterCrop(img_size),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])])}
# 实例化训练数据集
train_dataset = MyDataSet(images_path=train_images_path,
images_class=train_images_label,
transform=data_transform["train"])
# 实例化验证数据集
val_dataset = MyDataSet(images_path=val_images_path,
images_class=val_images_label,
transform=data_transform["val"])
batch_size = args.batch_size
nw = min([os.cpu_count(), batch_size if batch_size > 1 else 0, 8]) # number of workers
print('Using {} dataloader workers every process'.format(nw))
train_loader = torch.utils.data.DataLoader(train_dataset,
batch_size=batch_size,
shuffle=True,
pin_memory=True,
num_workers=nw,
collate_fn=train_dataset.collate_fn)
val_loader = torch.utils.data.DataLoader(val_dataset,
batch_size=batch_size,
shuffle=False,
pin_memory=True,
num_workers=nw,
collate_fn=val_dataset.collate_fn)
model = create_model(num_classes=args.num_classes).to(device)
if args.weights != "":
assert os.path.exists(args.weights), "weights file: '{}' not exist.".format(args.weights)
weights_dict = torch.load(args.weights, map_location=device)["model"]
# 删除有关分类类别的权重
for k in list(weights_dict.keys()):
if "head" in k:
del weights_dict[k]
print(model.load_state_dict(weights_dict, strict=False))
if args.freeze_layers:
for name, para in model.named_parameters():
# 除head外，其他权重全部冻结
if "head" not in name:
para.requires_grad_(False)
else:
print("training {}".format(name))
pg = [p for p in model.parameters() if p.requires_grad]
optimizer = optim.AdamW(pg, lr=args.lr, weight_decay=5E-2)
for epoch in range(args.epochs):
# train
train_loss, train_acc = train_one_epoch(model=model,
optimizer=optimizer,
data_loader=train_loader,
device=device,
epoch=epoch)
# validate
val_loss, val_acc = evaluate(model=model,
data_loader=val_loader,
device=device,
epoch=epoch)
tags = ["train_loss", "train_acc", "val_loss", "val_acc", "learning_rate"]
tb_writer.add_scalar(tags[0], train_loss, epoch)
tb_writer.add_scalar(tags[1], train_acc, epoch)
tb_writer.add_scalar(tags[2], val_loss, epoch)
tb_writer.add_scalar(tags[3], val_acc, epoch)
tb_writer.add_scalar(tags[4], optimizer.param_groups[0]["lr"], epoch)
torch.save(model.state_dict(), "../weights/model-{}.pth".format(epoch))
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--num_classes', type=int, default=5)
parser.add_argument('--epochs', type=int, default=10)
parser.add_argument('--batch-size', type=int, default=8)
parser.add_argument('--lr', type=float, default=0.0001)
# 数据集所在根目录
# https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz
parser.add_argument('--data-path', type=str,
default="../data/flower_photos")
# 预训练权重路径，如果不想载入就设置为空字符
parser.add_argument('--weights', type=str, default='../weights/swin_tiny_patch4_window7_224.pth',
help='initial weights path')
# 是否冻结权重
parser.add_argument('--freeze-layers', type=bool, default=False)
parser.add_argument('--device', default='cuda:0', help='device id (i.e. 0 or 0,1 or cpu)')
opt = parser.parse_args()
main(opt)训练结果如下图所示：
即使迁移学习的模型训练数据集较小，训练和测试结果准确率还是很高的。
模型预测代码如下：
import os
import json
import torch
from PIL import Image
from torchvision import transforms
import matplotlib.pyplot as plt
from model import swin_tiny_patch4_window7_224 as create_model
def main():
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
img_size = 224
data_transform = transforms.Compose(
[transforms.Resize(int(img_size * 1.14)),
transforms.CenterCrop(img_size),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])])
# load image
img_path = "../data/Image/flower.png"
assert os.path.exists(img_path), "file: '{}' dose not exist.".format(img_path)
img = Image.open(img_path)
plt.imshow(img)
plt.show()
img2 = img
# [N, C, H, W]
img = img.convert('RGB')
img = data_transform(img)
# expand batch dimension
img = torch.unsqueeze(img, dim=0)
# read class_indict
json_path = 'class_indices.json'
assert os.path.exists(json_path), "file: '{}' dose not exist.".format(json_path)
with open(json_path, "r") as f:
class_indict = json.load(f)
# create model
model = create_model(num_classes=5).to(device)
# load model weights
model_weight_path = "../weights/model-9.pth"
model.load_state_dict(torch.load(model_weight_path, map_location=device))
model.eval()
with torch.no_grad():
# predict class
output = torch.squeeze(model(img.to(device))).cpu()
predict = torch.softmax(output, dim=0)
predict_cla = torch.argmax(predict).numpy()
print_res = "class: {} prob: {:.3}".format(class_indict[str(predict_cla)], predict[predict_cla].numpy())
plt.title(print_res)
for i in range(len(predict)):
print("class: {:10} prob: {:.3}".format(class_indict[str(i)], predict[i].numpy()))
plt.imshow(img2)
plt.show()
if __name__ == '__main__':
main()代码预测结果如下图所示：
大约99.8%的可能性为玫瑰类别。
二、MLP-Mixer
除了通过卷积和自注意力的方式处理图像，还有哪些方法可以实现呢？当然还有万能的全连接，MLP-Mixer采用全连接的方式实现图像分类操作。
全连接层处理图像难道不会导致参数过于庞大？这些问题在算力足够富裕的情况下都不是问题，使用全连接层会有更低的归纳偏置，而且在同准确率的情况下，处理速度会更快，如下图所示：
接下来我们就来看看MLP-Mixer是如何运转的吧！
2.1 网络模型整体结构
如上图所示，MLP-Mixer整体逻辑和ViT类似，但是该模型将ViT中的Transformer Encoder替换为N个Mixer Layer。
同样输入 224x224x3 的图像将其打成49个patch，即每个patch维度为 32x32x3 = 3072；再将 49x3072 的patch传入Per-patch Fully-connected进行降维操作，该层输出为 49x512；再将其传入N层Mixer Layer， Mixer Layer的输入输出维度是相同的；最后将Mixer Layer的输出做一次全局平均池化得到维度大小的向量之后，最后接入全连接层后输出分类概率。
2.2 Mixer Layer
Mixer Layer共进行两次MLP操作：
第一次为channel-mixing，即将每一维进行融合以混合每个位置的特征；
第二次是token-mixing，即将每一个token内部进行自融合，以混合空间信息。
- MLP1
首先输入张量为 49x512，先经过层归一化后，将其进行转置T变为 512x49，然后对每一个patch（49维）进行MLP操作后，仍得到 512x49 的张量，再转置T回 49x512。
在各层归一化之间都采用残差链接，即Skip-connections。
因为channel-mixing相当于所有patch在通道上连接后，做 1x1 的卷积获取同位置的空间信息，所以MLP-Mixer是CNN的特例。
- MLP2
经过MLP1输出的 49x512 张量先进行层归一化，然后分别对每个patch（512维）进行MLP2操作后，仍输出 49x512 的张量。
可能图片上的标记更好理解：
2.3 MLP
MLP是两个全连接层之间加一个GELU激活函数。
GELU激活函数：
![GULE(x)=0.5\times [1+tanh(\sqrt{\frac{2}{\pi }}(x+0.047715x^{3}))]](https://image.shishitao.com:8440/aHR0cHM6Ly9sYXRleC5jc2RuLm5ldC9lcT9HVUxFJTI4eCUyOSUzRDAuNSU1Q3RpbWVzJTIwJTVCMSZwbHVzO3RhbmglMjglNUNzcXJ0JTdCJTVDZnJhYyU3QjIlN0QlN0IlNUNwaSUyMCU3RCU3RCUyOHgmcGx1czswLjA0NzcxNXglNUUlN0IzJTdEJTI5JTI5JTVE.jpg?w=700)
因为MLP-Mixer采用全连接的方式，所以无需进行位置编码，因为token之间交换位置所对应的神经元权重不同，所以“语言”也会不同。
2.4 代码
MLP-Mixer网络搭建PyTorch如下所示：
class MlpBlock(nn.Module):
def __init__(self, in_mlp_dim=196, out_mlp_dim=256):
super(MlpBlock, self).__init__()
self.mlp_dim = out_mlp_dim
self.dense1 = nn.Linear(in_mlp_dim, out_mlp_dim)
self.gelu = nn.GELU()
self.dense2 = nn.Linear(out_mlp_dim, in_mlp_dim)
def forward(self, x):
y = self.dense1(x)
y = self.gelu(y)
y = self.dense2(y)
return y
class MixerBlock(nn.Module):
def __init__(self, tokens_mlp_dim=256, channels_mlp_dim=2048, batch_size=32):
super(MixerBlock, self).__init__()
self.batch_size = batch_size
self.norm1 = nn.LayerNorm(512) # 对512维的做归一化，默认给最后一个维度做归一化
self.token_Mixing = MlpBlock(out_mlp_dim=tokens_mlp_dim)
self.norm2 = nn.LayerNorm(512) # 对512维的做归一化
self.channel_mixing = MlpBlock(in_mlp_dim=512, out_mlp_dim=channels_mlp_dim)
def forward(self, x):
y = self.norm1(x)
y = y.permute(0, 2, 1)
y = self.token_Mixing(y)
y = y.permute(0, 2, 1)
x = x + y
y = self.norm2(x)
return x + self.channel_mixing(y)
class MlpMixer(nn.Module):
def __init__(self, patches, num_classes, num_blocks, hidden_dim, tokens_mlp_dim, channels_mlp_dim):
super(MlpMixer, self).__init__()
self.stem = nn.Conv2d(3, hidden_dim, kernel_size=patches, stride=patches)
self.mixer_block_1 = MixerBlock()
self.mixer_blocks = nn.ModuleList([MixerBlock(tokens_mlp_dim, channels_mlp_dim) for _ in range(num_blocks)])
self.pre_head_norm = nn.LayerNorm(hidden_dim)
self.head = nn.Linear(hidden_dim, num_classes) if num_classes > 0 else nn.Identity()
def forward(self, x):
x = self.stem(x)
b, c, h, w = x.shape
x = x.view(b, c, -1).permute(0, 2, 1)
for mixer_block in self.mixer_blocks:
x = mixer_block(x)
x = self.pre_head_norm(x)
x = x.mean(dim=1)
x = self.head(x)
return x
model = MlpMixer(16, 10, 6, 512, 256, 2048)因为算力不够我这里就没有拿数据去训练。
总结
本周的学习到此结束，目前图像领域仍是Transformer思想为主导的模型霸榜，所以下周将会继续有关Transformer for Vision的学习。
如有错误，请各位大佬指出，谢谢！