resnet 在残差结构内增加注意力机制并且预训练-CSDN博客

本文链接：https://blog.csdn.net/weixin_45724232/article/details/116082250

该博客介绍了如何在PyTorch中加载预训练的ResNet34模型权重，并进行筛选和更新，以适应自定义网络结构。通过在BasicBlock模块中添加注意力机制并控制权重加载流程，实现模型的微调。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

所填添加的注意机制只是一个演示，怎么添加还需要自己琢磨

class BasicBlock(nn.Module):
    expansion = 1
    def __init__(self, inplanes, planes, stride=1, downsample=None, att=False):
        super(BasicBlock, self).__init__()
        self.conv1 = conv3x3(inplanes, planes, stride)
        self.bn1 = nn.BatchNorm2d(planes)
        self.relu = nn.ReLU(inplace=True)
        self.conv2 = conv3x3(planes, planes)
        self.bn2 = nn.BatchNorm2d(planes)
        self.downsample = downsample
        self.stride = stride
        self.is_att = att
        self.att = Attention(planes)
    def forward(self, x):
        residual = x
        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)
        out = self.conv2(out)
        out = self.bn2(out)
        if self.is_att:
            scale = self.att(out)
            out = out * scale
        if self.downsample is not None:
            residual = self.downsample(x)
        out += residual
        out = self.relu(out)
        return out

在加载与训练时，
1.先加载预训练权重
2.筛选所需的预训练权重
3.再将预训练权重更新到model_dict中
4.load权重

if pretrained:
        pretrained_dict=model_zoo.load_url(model_urls['resnet34'],model_dir='../pretrained')# Modify 'model_dir' according to your own path
        print('Petrain Model Have been loaded!')
        pretrained_dict = {k: v for k, v in pretrained_dict.items() if k in model_dict}
        model_dict.update(pretrained_dict)
        model.load_state_dict(model_dict)
return model