PyTorch从零开始实现ResNet

代码实现

本文实现 ResNet原论文 Deep Residual Learning for Image Recognition 中的50层，101层和152层残差连接。

代码中使用基础残差块这个概念，这里的基础残差块指的是上图中红色矩形圈出的内容：从上到下分别使用3, 4, 6, 3个基础残差块，每个基础残差块由三个卷积层组成，核大小分别为1x1, 3x3, 1x1 。

残差连接的结构

复现代码如下：

import torch
import torch.nn as nn

# 基础残差块，后面ResNet要多次重复使用该块
class block(nn.Module):
    def __init__(self, in_channels, out_channels, identity_downsample=None, stride=1):
        super(block, self).__init__()
        self.expansion = 4  
        self.conv1 = nn.Conv2d(in_channels, out_channels, kernel_size=1, stride=1, padding=0)
        self.bn1 = nn.BatchNorm2d(out_channels)
        self.conv2 = nn.Conv2d(out_channels, out_channels, kernel_size=3, stride=stride, padding=1)
        self.bn2 = nn.BatchNorm2d(out_channels)
        self.conv3 = nn.Conv2d(out_channels, out_channels*self.expansion, kernel_size=1, stride=1, padding=0)
        self.bn3 = nn.BatchNorm2d(out_channels*self.expansion)
        self.relu = nn.ReLU()
        self.identity_downsample = identity_downsample
    
    def forward(self, x):
        identity = x

        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.conv2(x)
        x = self.bn2(x)
        x = self.relu(x)
        x = self.conv3(x)
        x = self.bn3(x)

		# x 和 identity形状一致，才能相加
        if self.identity_downsample is not None:
            identity = self.identity_downsample(identity)

        x += identity
        x = self.relu(x)
        return x
    
class ResNet(nn.Module):
    def __init__(self, block, layers, image_channels, num_classes):
        super(ResNet, self).__init__()
        # 初始化的层
        self.in_channels = 64
        self.conv1 = nn.Conv2d(image_channels, 64, kernel_size=7, stride=2, padding=3)
        self.bn1 = nn.BatchNorm2d(64)
        self.relu = nn.ReLU()
        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)

        # ResNet layers
        self.layer1 = self._make_layer(block, layers[0], out_channels=64, stride=1)
        self.layer2 = self._make_layer(block, layers[1], out_channels=128, stride=2)
        self.layer3 = self._make_layer(block, layers[2], out_channels=256, stride=2)
        self.layer4 = self._make_layer(block, layers[3], out_channels=512, stride=2)

        self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
        self.fc = nn.Linear(512*4, num_classes)

    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.maxpool(x)

        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)
        x = self.layer4(x)

        x = self.avgpool(x)
        x = x.reshape(x.shape[0], -1)
        x = self.fc(x)
        return x

	# 核心函数：调用block基础残差块，构造ResNet的每一层
    def _make_layer(self, block, num_residual_blocks, out_channels, stride):
        identity_downsample = None
        layers = []
        
		# 修改形状，使得残差连接可以相加：x + identity
        if stride != 1 or self.in_channels != out_channels * 4:
            identity_downsample = nn.Sequential(nn.Conv2d(self.in_channels, out_channels*4, kernel_size=1,
                                                          stride=stride),                                               
                                                nn.BatchNorm2d(out_channels*4))
            
        layers.append(block(self.in_channels, out_channels, identity_downsample, stride))
        self.in_channels = out_channels * 4

        for i in range(num_residual_blocks - 1):
            layers.append(block(self.in_channels, out_channels)) # 256 -> 64, 64*4(256) again

        return nn.Sequential(*layers)
 
 # 构造ResNet50层：默认图像通道3，分类类别为1000
def resnet50(img_channels=3, num_classes=1000):
    return ResNet(block, [3, 4, 6, 3], img_channels, num_classes)

 # 构造ResNet101层  
def resnet101(img_channels=3, num_classes=1000):
    return ResNet(block, [3, 4, 23, 3], img_channels, num_classes)

 # 构造ResNet152层  
def resnet152(img_channels=3, num_classes=1000):
    return ResNet(block, [3, 8, 36, 3], img_channels, num_classes)

# 测试输出y的形状是否满足1000类
def test():
    net = resnet152()
    x = torch.randn(2, 3, 224, 224)
    y = net(x)
    print(y.shape) # [2, 1000]

test()

参考

[1] Deep Residual Learning for Image Recognition
[2] https://www.youtube.com/watch?v=DkNIBBBvcPs&list=PLhhyoLH6IjfxeoooqP9rhU3HJIAVAJ3Vz&index=19