Concatenating tensors error: Expected object of scalar type Float but got scalar type Double

1

Hi,

I want to concatenate a tensor with a number that has multiple decimal places (0.990989 etc.) basically a float.

I extract features from a pretrained network like this:

class DenseNetConv(torch.nn.Module): 
    def __init__(self):
        super(DenseNetConv,self).__init__()
        original_model = models.densenet161(pretrained=True)
        self.features = torch.nn.Sequential(*list(original_model.children())[:-1])
        for param in self.parameters():
            param.requires_grad = False

    def forward(self, x):
        x = self.features(x)
        x = F.avg_pool2d(x, kernel_size=7).view(x.size(0), -1)
        return x

where I then define the fully connected layer like:

classifier = nn.Linear(2208 + 1, args.num_classes)

where “2208” is the number of features from the above class and the “1” is the additional score I want to concatenate:


score = score.to(device)
x10 = densenet(inputs10x)
x_all = torch.cat([x10, score], dim=1)

However, when I concatenate I get an error:

RuntimeError: Expected object of scalar type Float but got scalar type Double for sequence element 1 in sequence argument at position #1 'tensors'

Is the problem with the different dimensions of the score tuple and features tuple?

2

you can convert x10 to a double by doing

x10.double()
1 Like