torchsummary输出重复问题求助：附CNN模型复现代码

我正在复现一篇基于CNN的分类方案论文，写出了如下简化版代码，但运行torchsummary时出现输出重复的情况，查过其GitHub问答区没找到相关问题记录。

复现代码 #

import torch
import torch.nn as nn
from torchsummary import summary

class CNN_Pred2D(nn.Module):
    def __init__(self, n_filters=[8,8,8], debug=True):
        super().__init__()
        self.debug = debug
        
        self.model = nn.Sequential(
            nn.Conv2d(1, n_filters[0], kernel_size=(1,82)),
            nn.ReLU(),
            nn.Conv2d(n_filters[0], n_filters[0], kernel_size=(3,1)),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=(2,1)),
            
            nn.Conv2d(n_filters[0], n_filters[1], kernel_size=(3,1)),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=(2,1)),
            
            nn.Flatten(),
            nn.Linear(104,1),
            nn.Sigmoid()
        )

        
    def forward(self, X):
        out = self.model(X)
#         print(out.shape)
        return out

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = CNN_Pred2D().to(device)

summary(model, [(1, 60,82)])

异常输出截图 #

解决方法 #

• 升级/替换torchsummary：旧版torchsummary对嵌套nn.Sequential的模型结构识别存在bug，建议升级到最新稳定版，或改用torchinfo（torchsummary的官方维护分支），它对复杂模型结构的兼容性更好。
• 取消嵌套Sequential：将self.model中的层直接定义在模型类的__init__方法中，不通过嵌套的nn.Sequential封装，让torchsummary能正确遍历每一层结构。修改示例如下：

  class CNN_Pred2D(nn.Module):
      def __init__(self, n_filters=[8,8,8], debug=True):
          super().__init__()
          self.debug = debug
          
          # 直接定义每层，取消嵌套Sequential
          self.conv1 = nn.Conv2d(1, n_filters[0], kernel_size=(1,82))
          self.relu1 = nn.ReLU()
          self.conv2 = nn.Conv2d(n_filters[0], n_filters[0], kernel_size=(3,1))
          self.relu2 = nn.ReLU()
          self.pool1 = nn.MaxPool2d(kernel_size=(2,1))
          
          self.conv3 = nn.Conv2d(n_filters[0], n_filters[1], kernel_size=(3,1))
          self.relu3 = nn.ReLU()
          self.pool2 = nn.MaxPool2d(kernel_size=(2,1))
          
          self.flatten = nn.Flatten()
          self.fc = nn.Linear(104,1)
          self.sigmoid = nn.Sigmoid()
  
      def forward(self, X):
          x = self.relu1(self.conv1(X))
          x = self.relu2(self.conv2(x))
          x = self.pool1(x)
          x = self.relu3(self.conv3(x))
          x = self.pool2(x)
          x = self.flatten(x)
          x = self.sigmoid(self.fc(x))
          return x
  

• 验证模型结构：通过print(list(model.named_modules()))打印模型所有模块，确认是否存在重复定义的层，排除代码逻辑错误。
• 清理环境缓存：重启Python内核，或执行torch.cuda.empty_cache()清理缓存，避免因缓存导致的结构识别异常。

内容的提问来源于stack exchange，提问作者sourabh gupta