shufflenet分类pytorch代码

下面是使用PyTorch实现ShuffleNet的分类代码：

import torch
import torch.nn as nn

class ShuffleNetBlock(nn.Module):
    def __init__(self, inp, oup, mid_channels, ksize, stride):
        super(ShuffleNetBlock, self).__init__()
        self.stride = stride
        self.mid_channels = mid_channels
        self.inp = inp
        self.oup = oup

        assert stride in [1, 2]

        if stride == 2:
            self.branch1 = nn.Sequential(
                nn.Conv2d(inp, inp, 3, 2, 1, groups=inp, bias=False),
                nn.BatchNorm2d(inp),
                nn.Conv2d(inp, mid_channels, 1, 1, 0, bias=False),
                nn.BatchNorm2d(mid_channels),
                nn.ReLU(inplace=True),
            )

            self.branch2 = nn.Sequential(
                nn.Conv2d(inp, mid_channels, 1, 1, 0, bias=False),
                nn.BatchNorm2d(mid_channels),
                nn.ReLU(inplace=True),
                nn.Conv2d(mid_channels, mid_channels, ksize, stride, ksize//2, groups=mid_channels, bias=False),
                nn.BatchNorm2d(mid_channels),
                nn.Conv2d(mid_channels, mid_channels, 1, 1, 0, bias=False),
                nn.BatchNorm2d(mid_channels),
                nn.ReLU(inplace=True),
            )

        else:
            assert inp == oup
            self.branch1 = nn.Sequential()

            self.branch2 = nn.Sequential(
                nn.Conv2d(mid_channels, mid_channels, ksize, stride, ksize//2, groups=mid_channels, bias=False),
                nn.BatchNorm2d(mid_channels),
                nn.Conv2d(mid_channels, oup, 1, 1, 0, bias=False),
                nn.BatchNorm2d(oup),
                nn.ReLU(inplace=True),
            )

    def forward(self, x):
        if self.stride == 1:
            x1, x2 = x.chunk(2, dim=1)
            out = torch.cat((x1, self.branch2(x2)), dim=1)
        else:
            out = torch.cat((self.branch1(x), self.branch2(x)), dim=1)

        return out

class ShuffleNet(nn.Module):
    def __init__(self, num_classes=1000):
        super(ShuffleNet, self).__init__()
        self.conv1 = nn.Conv2d(3, 24, 3, 2, 1, bias=False)
        self.bn1 = nn.BatchNorm2d(24)
        self.maxpool = nn.MaxPool2d(3, 2, 1)

        self.stage2 = self._make_stage(24, 144, 3, 2)
        self.stage3 = self._make_stage(144, 288, 7, 2)
        self.stage4 = self._make_stage(288, 576, 3, 2)
        self.conv5 = nn.Conv2d(576, 1024, 1, 1, 0, bias=False)
        self.bn5 = nn.BatchNorm2d(1024)
        self.fc = nn.Linear(1024, num_classes)

    def _make_stage(self, inp, oup, ksize, stride):
        layers = []
        layers.append(ShuffleNetBlock(inp, oup, int(oup/2), ksize, stride))

        for i in range(1, 4):
            layers.append(ShuffleNetBlock(oup, oup, int(oup/2), ksize, 1))

        return nn.Sequential(*layers)

    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = nn.ReLU(inplace=True)(x)
        x = self.maxpool(x)

        x = self.stage2(x)
        x = self.stage3(x)
        x = self.stage4(x)

        x = self.conv5(x)
        x = self.bn5(x)
        x = nn.ReLU(inplace=True)(x)

        x = x.mean([2, 3])
        x = self.fc(x)

        return x

其中，ShuffleNetBlock是ShuffleNet的基本模块，_make_stage是构建ShuffleNet每个阶段的函数，ShuffleNet是整个模型的定义。

可以使用以下代码来进行模型的实例化和训练：

import torch.optim as optim
import torchvision.datasets as datasets
import torchvision.transforms as transforms

# 实例化模型
model = ShuffleNet()

# 定义损失函数和优化器
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(model.parameters(), lr=0.1, momentum=0.9, weight_decay=0.0001)

# 加载数据集
train_dataset = datasets.ImageFolder(root='./train', transform=transforms.ToTensor())
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=32, shuffle=True, num_workers=4)

# 训练模型
for epoch in range(100):
    for i, (inputs, targets) in enumerate(train_loader):
        optimizer.zero_grad()
        outputs = model(inputs)
        loss = criterion(outputs, targets)
        loss.backward()
        optimizer.step()

        if i % 10 == 0:
            print('Epoch: %d, Batch: %d, Loss: %.3f' % (epoch+1, i, loss.item()))

其中，train文件夹下存放的是训练数据集，每个子文件夹代表一个类别。可以使用torchvision.transforms对数据进行预处理，例如将图像转换为tensor、对图像进行随机裁剪、随机翻转等。在训练过程中，使用优化器对模型的参数进行更新，同时计算损失函数，以监督模型的训练效果。