“交替训练两个数据集”

交替训练两个数据集可以通过以下示例代码实现：

import torch
import torch.utils.data as Data

# 定义数据集1和数据集2，分别为train_dataset1和train_dataset2
train_dataset1 = Data.TensorDataset(torch.randn(20, 10), torch.randint(0, 2, (20,)))
train_dataset2 = Data.TensorDataset(torch.randn(30, 10), torch.randint(0, 2, (30,)))

# 定义数据加载器1和数据加载器2，分别为train_loader1和train_loader2
train_loader1 = Data.DataLoader(train_dataset1, batch_size=4, shuffle=True)
train_loader2 = Data.DataLoader(train_dataset2, batch_size=6, shuffle=True)

# 定义模型和优化器
model = torch.nn.Linear(10, 1)
optimizer = torch.optim.SGD(model.parameters(), lr=0.01)

# 定义交替训练函数
def alternate_train(model, optimizer, loader1, loader2, num_epochs):
    for epoch in range(num_epochs):
        # 训练数据集1
        model.train()
        for step, (batch_x, batch_y) in enumerate(loader1):
            pred_y = model(batch_x)
            loss = torch.nn.functional.binary_cross_entropy_with_logits(pred_y, batch_y.float())
            optimizer.zero_grad()
            loss.backward()
            optimizer.step()
        
        # 训练数据集2
        model.train()
        for step, (batch_x, batch_y) in enumerate(loader2):
            pred_y = model(batch_x)
            loss = torch.nn.functional.binary_cross_entropy_with_logits(pred_y, batch_y.float())
            optimizer.zero_grad()
            loss.backward()
            optimizer.step()
    
    return model

# 进行交替训练
model = alternate_train(model, optimizer, train_loader1, train_loader2, num_epochs=10)