Hi,
I am trying a fix an issue. Some methods are broken. I wanted to know the source of where tensor is created in Pytorch. Like when we pass arguments like requires_grad, how are they taken care of in C++.
What does wrap and THPVariable_wrap exactly do ?
I’m looking into /csrc/autograd/generated/.
Hi,
Is it this issue: https://github.com/pytorch/pytorch/issues/36455 ?
I think this particular comment on the issue answers your question https://github.com/pytorch/pytorch/issues/36455#issuecomment-614097391?
Yes, thanks. Still not sure how to go about it. Waiting for someone more experienced to pitch in as suggested by ezyang.