You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
But with pytorch and ao nightlies with torch.export.export_for_inference it seems that I still require unwrap_tensor_subclass.
So what is the use case of this conditional unwrap?
Currently we need to do:
to make tensor subclass work with export/aot_compile, this should be added to default export path directly
The text was updated successfully, but these errors were encountered: