jacrev
and jacfwd
fail for torch.roll
when input is a scalar
#94925
Labels
module: functorch
Pertaining to torch.func or pytorch/functorch
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
馃悰 Describe the bug
jacrev
andjacfwd
fail fortorch.roll
when input is a scalarHowever, the
jacobian
function will succeed and returns the correct gradientVersions
cc @zou3519 @Chillee @samdow @soumith @kshitij12345 @janeyx99 @ezyang @albanD @gqchen @pearu @nikitaved @soulitzer @lezcano @Varal7
The text was updated successfully, but these errors were encountered: