Skip to content
New issue

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

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

Already on GitHub? # to your account

support Zygote@0.7 #566

Merged
merged 4 commits into from
Apr 5, 2025
Merged

support Zygote@0.7 #566

merged 4 commits into from
Apr 5, 2025

Conversation

oscardssmith
Copy link
Member

This is needed to support ForwardDiff@1.0. Testing separately to see how CI looks.

This is needed to support ForwardDiff@1.0. Testing separately to see how CI looks.
@oscardssmith
Copy link
Member Author

unfortunately it looks like there are real inference failures caused by making everything a thunk in Zygote@0.7.

@ChrisRackauckas
Copy link
Member

These are real failures from this? I don't see how it would be related. Run master tests?

@oscardssmith
Copy link
Member Author

From talking to Avik, Zygote@0.7 creates more thunks than 0.6 which can regress type stability.

@ChrisRackauckas
Copy link
Member

But that should only effect tests of reverse mode, since the standard only uses forward mode?

@oscardssmith
Copy link
Member Author

But that should only effect tests of reverse mode, since the standard only uses forward mode?

The test that is failing is SimpleNonlinearSolve with AutoReverseDiff. #567 doesn't see the same failure.

@ChrisRackauckas ChrisRackauckas merged commit d56a975 into master Apr 5, 2025
166 of 216 checks passed
@ChrisRackauckas ChrisRackauckas deleted the os/Zygote-0.7 branch April 5, 2025 21:34
# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants