Skip to content

[AutoDiff] Enable control flow AD by default. #25273

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.

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

Merged

Conversation

dan-zheng
Copy link
Contributor

Preemptively remove -Xllvm -differentiation-enable-control-flow flag.

Note that limited control flow AD support is not yet robust:

  • Memory leaks: TF-552. (work in progress)
  • Incorrect zero gradients involving buffers + recursion: TF-554.

Preemptively remove `-Xllvm -differentiation-enable-control-flow` flag.

Note that limited control flow AD support is not yet robust:
- Memory leaks: TF-552.
- Incorrect zero gradients involving buffers + recursion: TF-554.
@dan-zheng dan-zheng added the tensorflow This is for "tensorflow" branch PRs. label Jun 5, 2019
@dan-zheng dan-zheng requested a review from rxwei June 5, 2019 21:12
@dan-zheng
Copy link
Contributor Author

@swift-ci Please test tensorflow

@dan-zheng dan-zheng merged commit 2d9111d into swiftlang:tensorflow Jun 5, 2019
@dan-zheng dan-zheng deleted the autodiff-enable-control-flow branch June 5, 2019 22:46
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
tensorflow This is for "tensorflow" branch PRs.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants