RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (torch.FloatTensor) should be the

心已赠人 2023-01-06 13:59 137阅读 0赞

运行时出现这个错误：RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (torch.FloatTensor) should be the same。

主要是mode没有把mode没有调用cuda，而input却调用了，造成了不统一。这两个地方要保持一致。下面举个同时调用cuda的例子

autoencoderModel = autoencoderModel.cuda() 1,声明mode时，调用cuda
    criterion = nn.MSELoss()
    optimizer = torch.optim.Adam(autoencoderModel.parameters(), lr=LEARNING_RATE)
    #=======================================================================================================================
    #=======================================================================================================================
    # Model Training and Saving
    bestLoss = 1
    for epoch in range(EPOCHS):
        autoencoderModel.train()
        if epoch % 50 == 0 and epoch > 0:
            optimizer.param_groups[0]['lr'] = optimizer.param_groups[0]['lr'] * 0.1
        for i, autoencoderInput in enumerate(train_loader):
            autoencoderInput = autoencoderInput.cuda()2,input时，调用cuda
            autoencoderOutput = autoencoderModel(autoencoderInput)
            loss = criterion(autoencoderOutput, autoencoderInput)
            optimizer.zero_grad()
            loss.backward()
            optimizer.step()
            if i % PRINT_RREQ == 0:
                print('Epoch: [{0}][{1}/{2}]\t' 'Loss {loss:.4f}\t'.format(epoch, i, len(train_loader), loss=loss.item()))