pytorch - RuntimeError：输入类型（torch.FloatTensor）和权重类型（torch.cuda.FloatTensor）即使在放置模型和输入到cuda之后也应该相同

Question

即使在我将输入和模型放入设备之后，我也遇到了“RuntimeError：输入类型 (torch.FloatTensor) 和权重类型 (torch.cuda.FloatTensor) 应该相同”。如果有人可以帮助我，我将不胜感激。

point_net.to(device)
for epoch in range(100):
    running_loss = 0.0
    correct = 0.0
    total = 0
    
    for i in range(0, points_tensor.shape[0], 2048):
        X = points_tensor[i:i + 2048]
        y = labels_tensor[i:i + 2048]
        
        X = X.to(device)
        y = y.to(device)
        optimizer.zero_grad()

        y_pred = point_net(X)
        loss = criterion(y_pred, y)
        loss.backward()
        optimizer.step()
        
        running_loss += loss.item()

标签： pytorch