`test_rrule`: also check output tangents with thunked fields
还没有人认领这个 Issue。
评估
- 难度
- 3/5
- 预计耗时
- 1-2 天
- 新手友好度
- 45/100
- Issue 类型
- 功能
- 描述清晰度
- 基本清楚
- 活跃度
- 活跃
- 领域
- testing
调研方向
Start with test_rrule and the existing check_thunked_output_tangent option, which currently calls the pullback with only the outer output tangent thunked. Reproduce the failure by wrapping each field of a Tangent output in a thunk and calling the pullback, as in the findnz example. Done means the new check runs for Tangent, Tuple and NamedTuple outputs, and that the unthunked and thunked results match. Decide with a maintainer whether to reuse the existing flag or add a new one.
由索引模型根据 Issue 内容生成。
描述
check_thunked_output_tangent=true only checks the pullback with the whole output tangent thunked, i.e. pullback(@thunk(ȳ)). But AD systems also pass structural tangents whose fields are thunks. For example, Zygote's tuple getindex adjoint keeps thunks, so for f(x)[i] the rule receives Tangent{Any}((ZeroTangent(), …, thunk, …)).
Rules that handle the outer thunk correctly can still fail on this. In JuliaDiff/ChainRules.jl#843, the findnz rule passed test_rrule, but with Zygote it failed:
using SparseArrays, Zygote
Zygote.gradient(A -> sum(findnz(A)[3]), sprand(5, 5, 0.4))
# MethodError: no method matching sparse(::Vector{Int64}, ::Vector{Int64}, ::InplaceableThunk{...}, ::Int64, ::Int64)
The pullback destructured _, _, V̄ = unthunk(Δ). Destructuring a Tangent does not unthunk its fields, unlike getindex/getproperty (see JuliaDiff/ChainRulesCore.jl#712).
Proposal: when the output tangent ȳ is a Tangent, or a Tuple/NamedTuple that is converted to one, test_rrule should additionally call the pullback with every field wrapped in a thunk, e.g. Tangent{P}(map(x -> @thunk(x), backing(ȳ))...), and compare the result with the unthunked cotangents. This could reuse the existing check_thunked_output_tangent flag or get a new one.
- 主要语言
- Julia
- 星标
- 50
- 派生
- 18
- 平均合并
- 1 小时 54 分钟
- 30 天内合并 PR
- 1
环境准备
这个项目没有提供开发容器、Dockerfile 或贡献指南,环境需要你自己搭建:先看它的 README,通用步骤见我们的新手贡献指南。
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
JuliaDiff/ChainRulesTestUtils.jl 的其他 Issue
-
难度 2/5 1-3 小时 新手友好度 65/100
JuliaDiff/ChainRulesTestUtils.jl#148 · 1 个 reaction ·
-
documentation good first issue
难度 2/5 1-3 小时 新手友好度 62/100
-
难度 2/5 1-3 小时 新手友好度 45/100
-
mutation
难度 3/5 1-2 天 新手友好度 45/100
-
enhancement
难度 5/5 一周以上 新手友好度 25/100
查看 JuliaDiff/ChainRulesTestUtils.jl 的全部 Issue
相似的 Issue
-
mul! with a Symmetric or Hermitian factor throws for a matrix view whose columns are not contiguous未关闭
难度 2/5 1-3 小时 新手友好度 72/100
JuliaLang/LinearAlgebra.jl#1829 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 78/100
MagneticResonanceImaging/MRIReco.jl#320 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 78/100
rfordatascience/tidytuesday#1098 ·
-
ci system:windows test
难度 1/5 1 小时以内 新手友好度 76/100
维护者通常 1 天内回复
-
Single-column TKE example: boundary-layer depth description doesn't match Han & Bretherton (2019)未关闭
难度 1/5 1-3 小时 新手友好度 82/100
NumericalEarth/Breeze.jl#1143 ·
维护者通常 1 天内回复