Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

revert to tensordot2 once tf.matmul is fixed #942

Open
mganahl opened this issue Sep 27, 2021 · 1 comment
Open

revert to tensordot2 once tf.matmul is fixed #942

mganahl opened this issue Sep 27, 2021 · 1 comment

Comments

@mganahl
Copy link
Collaborator

mganahl commented Sep 27, 2021 •
edited

The tf.matmul bug documented at tensorflow/tensorflow#52148 causes tensordot2 to fail under certain circumstances. Once tensorflow fixes the bug on their side we should revert to using tensordot2 in the TensorflowBackend.
@mganahl mganahl changed the title revert to tensordot2 once https://github.com/tensorflow/tensorflow/issues/52148 is fixed revert to tensordot2 once tf.matmul is fixed Sep 27, 2021
@refraction-ray
Copy link
Contributor

Even without tensordot2, bad interplay between tf.vectorized_map and complex valued ops (maybe still tf.matmul) can still emerge for gradient evaluations, see #940 (comment)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@mganahl @refraction-ray