输入类型（torch.cuda.DoubleTensor）和重量类型（torch.cuda.FloatTensor）应该相同

Question

我遇到了
错误 我应该更改哪一个来解决此错误？

  

输入类型（torch.cuda.DoubleTensor）和重量类型   （torch.cuda.FloatTensor）应该相同

for i in range(num_epoch):
    model.train()
    for j,[image,label] in enumerate(train_loader):
        x = Variable(image.float().cuda())
        y_= Variable(label.long().cuda())

        optimizer.zero_grad()
        output = model.forward(x)
        loss = loss_func(output,y_)
        loss.backward()
        optimizer.step()


    top_1_count = torch.FloatTensor([0])
    total = torch.FloatTensor([0])
    model.eval() 
    for image,label in test_loader:
        x = Variable(image,volatile=True).cuda()
        y_= Variable(label).cuda()

        output = model.forward(x)

        values,idx = output.max(dim=1)
        top_1_count += torch.sum(y_==idx).float().cpu().data

        total += label.size(0)

    print("Test Data Accuracy: {}%".format(100*(top_1_count/total).numpy()))
    if (top_1_count/total).numpy() > 0.98:
        break