JuliaDiff / JuliaDiff/ChainRulesTestUtils.jl

Errors from `PermutedDimsArray`

Open
#240 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Julia
Stars
50
Forks
18
PR merge metrics
No merged PRs in 30d

Description

Something is confused here, I'm pretty sure the rule gives the same result as the one for permutedims:

julia> test_rrule(PermutedDimsArray, rand(3, 4, 5), (3, 1, 2); atol=0.1)
test_rrule: PermutedDimsArray on Array{Float64, 3},Int64,Int64,Int64: Test Failed at /Users/me/.julia/packages/ChainRulesTestUtils/8dFTY/src/check_result.jl:24
  Expression: isapprox(actual, expected; kwargs...)
   Evaluated: isapprox([0.02 2.61 -7.71 0.42; -8.68 -8.99 -3.53 -4.96; 8.23 -2.05 7.62 -0.88;;; -6.06 0.12 3.47 4.54; -0.02 7.06 -3.1 0.65; 0.47 -5.54 2.34 0.55;;; -4.3 -4.81 -0.08 4.66; -1.29 1.05 2.44 2.48; 3.17 -3.19 1.14 -5.52;;; -6.65 -6.25 8.15 -3.25; 3.51 -4.9 5.53 -2.8; -8.0 0.13 -7.71 1.46;;; 0.42 8.74 8.1 2.85; 0.9 2.67 3.36 -8.24; 0.95 -6.47 -5.11 8.27], [0.019999999999746908 -6.6500000000002375 -0.02000000000025288 0.8999999999997429; -6.060000000000144 0.41999999999974325 -1.2900000000002418 8.229999999999814; -4.300000000000264 -8.680000000000021 3.5099999999997613 0.469999999999744;;; 3.169999999999719 2.609999999999724 -6.250000000000212 7.059999999999704; -8.00000000000028 0.11999999999974628 8.739999999999668 1.0499999999997618; 0.9499999999997412 -4.8100000000002225 -8.990000000000263 -4.900000000000136;;; 2.66999999999973 -3.1900000000002566 -7.710000000000218 8.149999999999768; -2.0500000000002374 0.12999999999974626 3.4699999999997257 8.099999999999698; -5.540000000000103 -6.470000000000212 -0.08000000000025227 -3.5300000000002676;;; -3.1000000000002617 3.3599999999997268 1.1399999999997399 0.4199999999997444; 2.439999999999729 7.619999999999814 -7.710000000000186 4.5399999999997185; 5.529999999999713 2.3399999999996295 -5.1100000000002215 4.659999999999719;;; -3.2500000000002327 0.6499999999997431 -8.240000000000203 -5.520000000000219; 2.849999999999756 2.4799999999997318 -0.8800000000002475 1.459999999999738; -4.960000000000223 -2.800000000000235 0.5499999999997423 8.269999999999694]; rtol = 1.0e-9, atol = 0.1)
Stacktrace:
  [1] test_approx(actual::Union{Number, AbstractArray{<:Number}}, expected::Union{Number, AbstractArray{<:Number}}, msg::Any; kwargs::Base.Pairs{Symbol, V, Tuple{Vararg{Symbol, N}}, NamedTuple{names, T}} where {V, N, names, T<:Tuple{Vararg{Any, N}}})
    @ ChainRulesTestUtils ~/.julia/packages/ChainRulesTestUtils/8dFTY/src/check_result.jl:24
  [2] _test_cotangent(accum_cotangent::Any, ad_cotangent::Any, fd_cotangent::Any; check_inferred::Any, kwargs::Base.Pairs{Symbol, V, Tuple{Vararg{Symbol, N}}, NamedTuple{names, T}} where {V, N, names, T<:Tuple{Vararg{Any, N}}})
    @ ChainRulesTestUtils ~/.julia/packages/ChainRulesTestUtils/8dFTY/src/testers.jl:301
  [3] (::ChainRulesTestUtils.var"#49#53"{Bool, NamedTuple{(:rtol, :atol), Tuple{Float64, Float64}}})(::Any, ::Vararg{Any})
    @ ChainRulesTestUtils ~/.julia/packages/ChainRulesTestUtils/8dFTY/src/testers.jl:228
  [4] #57
    @ ./tuple.jl:556 [inlined]
  [5] BottomRF
    @ ./reduce.jl:81 [inlined]
  [6] _foldl_impl
    @ ./reduce.jl:62 [inlined]
  [7] foldl_impl
    @ ./reduce.jl:48 [inlined]
  [8] mapfoldl_impl(f::typeof(identity), op::Base.var"#57#58"{ChainRulesTestUtils.var"#49#53"{Bool, NamedTuple{(:rtol, :atol), Tuple{Float64, Float64}}}}, nt::Nothing, itr::Base.Iterators.Zip{Tuple{Tuple{NoTangent, Array{Float64, 3}, NoTangent}, Tuple{NoTangent, Array{Float64, 3}, NoTangent}, Tuple{NoTangent, Array{Float64, 3}, NoTangent}}})
    @ Base ./reduce.jl:44
  [9] mapfoldl(f::Function, op::Function, itr::Base.Iterators.Zip{Tuple{Tuple{NoTangent, Array{Float64, 3}, NoTangent}, Tuple{NoTangent, Array{Float64, 3}, NoTangent}, Tuple{NoTangent, Array{Float64, 3}, NoTangent}}}; init::Nothing)
    @ Base ./reduce.jl:162
 [10] #foldl#259
    @ ./reduce.jl:185 [inlined]
 [11] foreach(::Function, ::Tuple{NoTangent, Array{Float64, 3}, NoTangent}, ::Tuple{NoTangent, Array{Float64, 3}, NoTangent}, ::Vararg{Tuple{NoTangent, Array{Float64, 3}, NoTangent}})
    @ Base ./tuple.jl:556
 [12] macro expansion
    @ ~/.julia/packages/ChainRulesTestUtils/8dFTY/src/testers.jl:227 [inlined]
 [13] macro expansion
    @ ~/.julia/dev/julia/usr/share/julia/stdlib/v1.8/Test/src/Test.jl:1356 [inlined]
 [14] test_rrule(::RuleConfig, ::Any, ::Any, ::Vararg{Any}; output_tangent::Any, check_thunked_output_tangent::Any, fdm::Any, rrule_f::Any, check_inferred::Bool, fkwargs::NamedTuple, rtol::Real, atol::Real, kwargs::Base.Pairs{Symbol, V, Tuple{Vararg{Symbol, N}}, NamedTuple{names, T}} where {V, N, names, T<:Tuple{Vararg{Any, N}}})
    @ ChainRulesTestUtils ~/.julia/packages/ChainRulesTestUtils/8dFTY/src/testers.jl:198
Test Summary:                                                        | Pass  Fail  Total  Time
test_rrule: PermutedDimsArray on Array{Float64, 3},Int64,Int64,Int64 |    7     1      8  0.0s
ERROR: Some tests did not pass: 7 passed, 1 failed, 0 errored, 0 broken.

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

Reproduce the failure with test_rrule(PermutedDimsArray, rand(3, 4, 5), (3, 1, 2); atol=0.1). Start by comparing the PermutedDimsArray rule with the permutedims rule and inspect the failing result reported through ChainRulesTestUtils. Done means the test passes with matching cotangents for this permutation.

Written by the indexing model from the issue text.

Assessment

Tech stack
julia
Domain
testing
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.