aten.linear_backward meta registration needed for running resnet and bert with compiled autograd #114358
Labels
module: meta tensors
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
🐛 Describe the bug
When running resnet or bert with compiled autograd I'm getting following error:
torch._subclasses.fake_tensor.UnsupportedOperatorException: aten.linear_backward.default
Versions
.
cc @ezyang @eellison @bdhirsh @crcrpar @mcarilli
The text was updated successfully, but these errors were encountered: