JuliaDiff / JuliaDiff/ChainRulesCore.jl

Support chunked frule

Open
#92 20 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

design forward-mode
Dominant language
Julia
Stars
267
Forks
66
PR merge metrics
No merged PRs in 30d

Description

from #90 it seems that @YingboMa wants frule to be able to be called on a Vector of sensitivies for the same primal value,
and get a of sensitivities vector back,
but without broadcasting ? (presumably because that would also recompute the forwad primal)

I don't understand properly.
So this thread is get @YingboMa @shashi or @ChrisRackauckas to explain that.

This might need a redesign of frrule again similar to solving #74.
Maybe we went too far there, since broadcasting the pushforward would presumably solve that case.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reading the context in #90 and the redesign discussion in #74, then review the current frule and frrule interfaces. Clarify whether a Vector of sensitivities should produce a sensitivities vector without broadcasting or recomputing the primal, and document the intended behavior before implementation.

Written by the indexing model from the issue text.

Assessment

Tech stack
julia
Domain
tooling
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.