JuliaDiff / JuliaDiff/ChainRulesCore.jl
Ability to identify rules that always return AbstractZero
Open
Nobody has claimed this yet.
design
rule definition helper
- Dominant language
- Julia
- Stars
- 267
- Forks
- 66
- PR merge metrics
- No merged PRs in 30d
Description
This related to @non_differentiable
I think for operator overloading based AD,
if a rule's propagator is always going to return a AbstractZero the correct thing to do quiet different.
One wants to accept the overloaded type, but return a non-overloaded type
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reading the @non_differentiable handling and operator-overloading AD rule machinery referenced in the issue. Clarify how to identify propagators that always return AbstractZero, including the expected overloaded input and non-overloaded result, then define tests for those cases.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- julia
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100