Skip to content

Conversation

@DhairyaLGandhi
Copy link
Member

UDEs and more generally, for AD getting gradients passed via Tangents is best practice. Due to the number of custom adjoints and rules, we end up with silently dropped gradients from various spots. This is a step in continuation of #701

Checklist

  • Appropriate tests were added
  • Any code changes were done in a way that does not break public API
  • All documentation related to code changes were updated
  • The new code follows the
    contributor guidelines, in particular the SciML Style Guide and
    COLPRAC.
  • Any new documentation only uses public API

Additional context

Add any other context about the problem here.

@DhairyaLGandhi DhairyaLGandhi marked this pull request as ready for review April 8, 2025 11:56
@DhairyaLGandhi
Copy link
Member Author

I've run formatting for the files in the PR, the one failing is test/downstream/initialization.jl. Lmk if I should add a format commit for that one as well.

@DhairyaLGandhi
Copy link
Member Author

pinging for a review @ChrisRackauckas

@DhairyaLGandhi
Copy link
Member Author

Unsure what is going on with the Core tests, it seems to fail to precompile Plots.jl, but that seems unrelated to the PR. cc @AayushSabharwal for a review as well

@ChrisRackauckas
Copy link
Member

That's on the Julia pre-release v1.12-beta2. Ignore those.

@ChrisRackauckas ChrisRackauckas merged commit 785f378 into SciML:master Apr 9, 2025
40 of 52 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants