If we attempt to do this:
A = nn.Conv1d(100,50,3)
B = nn.Conv1d(50,100,3)
A.weight = B.weight.t()
We get an error:
TypeError: cannot assign 'torch.FloatTensor' as parameter 'weight' (torch.nn.Parameter or None expected)
The problem is that B.weight.t() returns a Tensor, even though B.weight is a parameter.
I’m running on master branch. Any thoughts?