8000 Use Mooncake to derive ForwardDiff and ReverseDiff rules · Issue #518 · chalk-lab/Mooncake.jl · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

Use Mooncake to derive ForwardDiff and ReverseDiff rules #518

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
Open
yebai opened this issue Mar 10, 2025 · 0 comments
Open

Use Mooncake to derive ForwardDiff and ReverseDiff rules #518

yebai opened this issue Mar 10, 2025 · 0 comments
Labels
enhancement New feature or request

Comments

@yebai
Copy link
Member
yebai commented Mar 10, 2025 •
edited
Loading

One can consider using Mooncake to write ForwardDiff and ReverseDiff rules. Here is a motivating example: TuringLang/JuliaBUGS.jl#280 (comment).

This helps reduce the maintenance burden of the Turing.jl ecosystem, for example, from DistributionsAD and Bijectors, which have many redundant rules for different autodiff backends.

Initial consultation suggests this might be feasible, so this issue is to continue the discussion.

Related: ForwardDiffChainRules.@ForwardDiff_frule and ReverseDiff.@grad / ReverseDiff.@grad_from_chainrules macro.

EDIT: We should also add Mooncake as a new backend for DI.DifferentiateWith

@willtebbutt willtebbutt added the enhancement New feature or request label Mar 10, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
enhancement New feature or request
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@yebai @willtebbutt
0