您的位置:

残差结构:从原理到应用

一、残差结构的原理

残差结构在深度学习中的应用越来越广泛,其核心原理是将输入特征和参考特征拼接在一起进行训练,以增强模型的学习能力和泛化能力。

具体地,残差结构引入了跨层连接,使得模型可以直接利用浅层的信息跟踪梯度,从而更好地学习到深层次特征表示。这种跨越层次的连接机制将输入和输出进行残差学习,通过残差的计算和积累,模型可以更快速、准确地逼近真实函数。

下面是一个简单的残差结构示例:

import torch.nn as nn
class ResidualBlock(nn.Module):
    def __init__(self, in_channels, out_channels, stride=1):
        super(ResidualBlock, self).__init__()
        self.conv1 = nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=stride, padding=1)
        self.bn1 = nn.BatchNorm2d(out_channels)
        self.relu1 = nn.ReLU(inplace=True)
        self.conv2 = nn.Conv2d(out_channels, out_channels, kernel_size=3, stride=1, padding=1)
        self.bn2 = nn.BatchNorm2d(out_channels)
        self.relu2 = nn.ReLU(inplace=True)
        self.shortcut = nn.Sequential()
        if stride != 1 or in_channels != out_channels:
            self.shortcut = nn.Sequential(
                nn.Conv2d(in_channels, out_channels, kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm2d(out_channels)
            )
    def forward(self, x):
        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu1(out)
        out = self.conv2(out)
        out = self.bn2(out)
        out += self.shortcut(x)
        out = self.relu2(out)
        return out

二、残差结构的改进

虽然残差结构在很多模型中表现出了优异的性能,但它也存在着一些问题,例如如果网络结构复杂度过高,很容易产生梯度消失问题。

因此,一些改进方法被提出,例如Res2Net、PVT、HRNet等。Res2Net与ResNeXt类似,但是它将多个并行分支的特征更多地交互,以增强网络表达能力。PVT(Pyramid Vision Transformer)则是基于Transformer的一种新型网络模型,它拥有多级金字塔特征融合机制,可以更好地应对不同大小的对象。HRNet(High-Resolution Network)则是一种具有多分支特征提取模块的网络结构,可以更好地保留高分辨率信息。

三、残差结构的应用

残差结构已经被广泛应用于图像分类、目标检测、语音识别等各种领域。以下是在图像分类上的一个例子:

import torch
import torchvision.transforms as transforms
import torchvision.datasets as datasets
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import DataLoader

train_transforms = transforms.Compose([transforms.Resize(256),
                                       transforms.RandomCrop(224),
                                       transforms.RandomHorizontalFlip(),
                                       transforms.ToTensor(),
                                       transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                            std=[0.229, 0.224, 0.225])])

test_transforms = transforms.Compose([transforms.Resize(256),
                                      transforms.CenterCrop(224),
                                      transforms.ToTensor(),
                                      transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                           std=[0.229, 0.224, 0.225])])

train_dataset = datasets.CIFAR10('data/', train=True, transform=train_transforms, download=True)
test_dataset = datasets.CIFAR10('data/', train=False, transform=test_transforms, download=True)

train_loader = DataLoader(train_dataset, batch_size=32, shuffle=True, num_workers=4)
test_loader = DataLoader(test_dataset, batch_size=32, shuffle=False, num_workers=4)


class ResNet(nn.Module):
    def __init__(self, block, num_blocks, num_classes=10):
        super(ResNet, self).__init__()
        self.in_channels = 64
        self.conv = nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1, bias=False)
        self.bn = nn.BatchNorm2d(64)
        self.relu = nn.ReLU(inplace=True)
        self.layer1 = self.make_layer(block, 64, num_blocks[0], stride=1)
        self.layer2 = self.make_layer(block, 128, num_blocks[1], stride=2)
        self.layer3 = self.make_layer(block, 256, num_blocks[2], stride=2)
        self.layer4 = self.make_layer(block, 512, num_blocks[3], stride=2)
        self.avg_pool = nn.AdaptiveAvgPool2d((1, 1))
        self.fc = nn.Linear(512 * block.expansion, num_classes)
        
    def make_layer(self, block, out_channels, num_blocks, stride):
        strides = [stride] + [1] * (num_blocks - 1)
        layers = []
        for stride in strides:
            layers.append(block(self.in_channels, out_channels, stride))
            self.in_channels = out_channels * block.expansion
        return nn.Sequential(*layers)
    
    def forward(self, x):
        out = self.conv(x)
        out = self.bn(out)
        out = self.relu(out)
        out = self.layer1(out)
        out = self.layer2(out)
        out = self.layer3(out)
        out = self.layer4(out)
        out = self.avg_pool(out)
        out = out.view(out.size(0), -1)
        out = self.fc(out)
        return out

def ResNet18():
    return ResNet(ResidualBlock, [2, 2, 2, 2])

model = ResNet18()

criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(model.parameters(), lr=0.1, momentum=0.9, weight_decay=1e-4)
lr_scheduler = optim.lr_scheduler.MultiStepLR(optimizer, milestones=[150, 225], gamma=0.1)

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
model.to(device)
criterion.to(device)

num_epochs = 300

for epoch in range(num_epochs):
    model.train()
    for i, (images, labels) in enumerate(train_loader):
        images = images.to(device)
        labels = labels.to(device)
        optimizer.zero_grad()
        outputs = model(images)
        loss = criterion(outputs, labels)
        loss.backward()
        optimizer.step()
    lr_scheduler.step()
    model.eval()
    with torch.no_grad():
        correct = 0
        total = 0
        for images, labels in test_loader:
            images = images.to(device)
            labels = labels.to(device)
            outputs = model(images)
            _, predicted = torch.max(outputs.data, 1)
            total += labels.size(0)
            correct += (predicted == labels).sum().item()
    accuracy = 100 * correct / total
    print('Epoch [{}/{}], Loss: {:.4f}, Accuracy: {:.2f}%'.format(epoch+1, num_epochs, loss.item(), accuracy))