Torch.mm - pytorch 0.3.1 - unable to perform (t.sp.FloatTensor, t.FloatTensor),

autograd
1

I currently have two variables:
v1 = Variable[torch.sparse.FloatTensor]
v2 = Variable[torch.FloatTensor]

I am attempting to perform matrix multiplication on it through:
torch.mm(v1, v2)
but receiving the following error:

RuntimeError: Expected object of type Variable(torch.sparse.FloatTensor) but found type
Variable[torch.FloatTensor] for argument #1 ‘mat2’.

should this not be possible? I am currently using pytorch 0.3.1

2

Hi,

No I don’t think this is possible with such an old version of pytorch :confused:

3

I tried in pytorch 0.4.1 and it seems to work there. Just weird as I borrowed this code from a github repository based on a paper.