通过pytorch中在unet网络中定义一个可以修改类别权重的代码

下面是一个示例代码，展示如何在PyTorch中在UNet网络中定义一个可以修改类别权重的代码：

import torch
import torch.nn as nn

class UNet(nn.Module):
    def __init__(self, n_channels, n_classes, bilinear=True, class_weights=None):
        super(UNet, self).__init__()
        self.n_channels = n_channels
        self.n_classes = n_classes
        self.bilinear = bilinear
        self.class_weights = class_weights
        
        self.inc = DoubleConv(n_channels, 64)
        self.down1 = Down(64, 128)
        self.down2 = Down(128, 256)
        self.down3 = Down(256, 512)
        factor = 2 if bilinear else 1
        self.down4 = Down(512, 1024 // factor)
        self.up1 = Up(1024, 512 // factor, bilinear)
        self.up2 = Up(512, 256 // factor, bilinear)
        self.up3 = Up(256, 128 // factor, bilinear)
        self.up4 = Up(128, 64, bilinear)
        self.outc = OutConv(64, n_classes)
        
    def forward(self, x):
        x1 = self.inc(x)
        x2 = self.down1(x1)
        x3 = self.down2(x2)
        x4 = self.down3(x3)
        x5 = self.down4(x4)
        x = self.up1(x5, x4)
        x = self.up2(x, x3)
        x = self.up3(x, x2)
        x = self.up4(x, x1)
        logits = self.outc(x)
        
        if self.class_weights is not None:
            # apply class weights
            weight_tensor = torch.tensor(self.class_weights, dtype=torch.float32)
            weight_tensor = weight_tensor.to(logits.device)
            weight_tensor = weight_tensor.unsqueeze(0).unsqueeze(-1).unsqueeze(-1)
            logits = logits * weight_tensor
            
        return logits

在这个示例代码中，我们使用了PyTorch中的nn.Module类来定义了一个UNet网络。我们还定义了一个可选的类别权重参数class_weights，它可以是一个列表，其中包含每个类别的权重。

在前向传递方法中，如果class_weights参数不为None，则我们会将权重应用于输出的逻辑张量。我们首先将权重列表转换为张量，然后将其移动到模型当前设备上。最后，我们将权重张量与逻辑张量相乘，以将权重应用于每个类别的预测。