使用pytorch实现CNN和LSTM并列的文本分类代码并训练数据

使用 PyTorch 实现 CNN 和 LSTM 并列的文本分类的代码可以这样写：

import torch
import torch.nn as nn

class TextClassifier(nn.Module):
    def __init__(self, vocab_size, embedding_dim, num_class):
        super().__init__()
        self.embedding = nn.Embedding(vocab_size, embedding_dim)
        self.cnn = nn.Conv1d(embedding_dim, 128, 3, padding=1)
        self.lstm = nn.LSTM(embedding_dim, 128, batch_first=True)
        self.fc = nn.Linear(256, num_class)
        
    def forward(self, x):
        # x: (batch_size, seq_len)
        x = self.embedding(x) # (batch_size, seq_len, embedding_dim)
        x_cnn = self.cnn(x.transpose(1, 2)).max(dim=2)[0] # (batch_size, 128)
        x_lstm, _ = self.lstm(x) # (batch_size, seq_len, 128)
        x_lstm = x_lstm[:, -1, :] # (batch_size, 128)
        x = torch.cat((x_cnn, x_lstm), dim=1) # (batch_size, 256)
        x = self.fc(x) # (batch_size, num_class)
        return x

model = TextClassifier(vocab_size=len(vocab), embedding_dim=100, num_class=2)

# 训练数据
def train(model, iterator, optimizer, criterion):
    model.train()
    for batch in iterator:
        optimizer.zero_grad()
        logits = model(batch.text)
        loss = criterion(logits, batch.label)
        loss.backward()
        optimizer.step()

# 验证数据
def evaluate(model, iterator, criterion):
    model.eval()
    with torch.no_grad():
        total_loss = 0
        total_correct = 0
        total_num = 0
        for batch in iterator:
            logits = model(batch.text)
            loss = criterion(logits, batch.label)
            total_loss += loss.item()
            total_correct += (logits.argmax(dim=1) == batch.label).sum().item()
            total_num += len(batch)
    return total_loss / total_num, total_correct / total_num

其中，`vocab` 是词汇表，`batch.text` 和 `batch.label` 分别代表一个 mini-batch 的输入和标签。

然后你就可以