如何使用Pytorch实现将Alexnet中的附加输入数据与最后一个丢弃层的输出连接起来？

import torch
import torch.nn as nn
import torchvision.models as models

alexnet = models.alexnet(pretrained=True)
dropout_output_dim = alexnet.classifier[-1].in_features

class CustomModel(nn.Module):
    def __init__(self, dropout_output_dim):
        super(CustomModel, self).__init__()
        self.alexnet = models.alexnet(pretrained=True)
        self.dropout = nn.Dropout()
        self.fc = nn.Linear(dropout_output_dim + additional_input_dim, num_classes)

    def forward(self, x, additional_input):
        alexnet_output = self.alexnet(x)
        alexnet_output = self.dropout(alexnet_output)
        combined_output = torch.cat((alexnet_output, additional_input), dim=1)
        output = self.fc(combined_output)
        return output

model = CustomModel(dropout_output_dim)
criterion = nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(), lr=0.001, momentum=0.9)

# 在训练过程中，将附加输入数据和标签传递给模型
for images, additional_input, labels in dataloader:
    optimizer.zero_grad()
    outputs = model(images, additional_input)
    loss = criterion(outputs, labels)
    loss.backward()
    optimizer.step()