ambiguity error in size(rec, dims...)

[CarloLucibello]

The following code works well

julia> f(x)=(p=size(x); p[1]*sum(x.^2))

julia> grad(f)(ones(3))
3-element Array{Float64,1}:
 6.0
 6.0
 6.0

and also size(x,1) doesn't give any problem, but here I have an ambiguity error

julia> f(x)=(p=size(x,(1,2)...); p[1]*sum(x.^2))

julia> grad(f)(ones(3,2,2,3))
ERROR: MethodError: size(::AutoGrad.Rec{Array{Float64,4}}, ::Int64, ::Int64, ::Int64) is ambiguous. Candidates:
  size{N}(x, d1::Integer, d2::Integer, dx::Vararg{Integer,N}) at abstractarray.jl:48
  size{##305}(x::AutoGrad.Rec{##305}, i...)
 in f(::AutoGrad.Rec{Array{Float64,4}}) at ./REPL[7]:1
 in forward_pass(::Function, ::Tuple{Array{Float64,4}}, ::Array{Any,1}, ::Int64) at /home/carlo/.julia/v0.5/AutoGrad/src/core.jl:88
 in (::AutoGrad.##gradfun#1#3{#f,Int64})(::Array{Any,1}, ::Function, ::Array{Float64,4}, ::Vararg{Array{Float64,4},N}) at /home/carlo/.julia/v0.5/AutoGrad/src/core.jl:39
 in (::AutoGrad.#gradfun#2)(::Array{Float64,4}, ::Vararg{Array{Float64,4},N}) at /home/carlo/.julia/v0.5/AutoGrad/src/core.jl:39
 in eval_user_input(::Any, ::Base.REPL.REPLBackend) at ./REPL.jl:64
 in macro expansion at ./REPL.jl:95 [inlined]
 in (::Base.REPL.##3#4{Base.REPL.REPLBackend})() at ./event.jl:68

This could be fixed easily defining

size(x::AutoGrad.Rec, d1::Integer, d2::Integer, dx::Vararg{Integer}) = size(getval(x), d1, d2, dx...)

as I already tested. If someone points me to the right location where this function should reside I can file a PR.

crossref denizyuret/Knet.jl#139

[denizyuret]

`size` is handled in AutoGrad/src/interfaces.jl L229-249.

…

On Fri, Jun 30, 2017 at 11:28 AM Carlo Lucibello ***@***.***> wrote: The following code works well julia> f(x)=(p=size(x); p[1]*sum(x.^2)) julia> grad(f)(ones(3))3-element Array{Float64,1}: 6.0 6.0 6.0 and also size(x,1) doesn't give any problem, but here I have an ambiguity error julia> f(x)=(p=size(x,(1,2)...); p[1]*sum(x.^2)) WARNING: Method definition f(Any) in module Main at REPL[5]:1 overwritten at REPL[7]:1. f (generic function with 1 method) julia> grad(f)(ones(3,2,2,3)) ERROR: MethodError: size(::AutoGrad.Rec{Array{Float64,4}}, ::Int64, ::Int64, ::Int64) is ambiguous. Candidates: size{N}(x, d1::Integer, d2::Integer, dx::Vararg{Integer,N}) at abstractarray.jl:48 size{##305}(x::AutoGrad.Rec{##305}, i...) in f(::AutoGrad.Rec{Array{Float64,4}}) at ./REPL[7]:1 in forward_pass(::Function, ::Tuple{Array{Float64,4}}, ::Array{Any,1}, ::Int64) at /home/carlo/.julia/v0.5/AutoGrad/src/core.jl:88 in (::AutoGrad.##gradfun#1#3{#f,Int64})(::Array{Any,1}, ::Function, ::Array{Float64,4}, ::Vararg{Array{Float64,4},N}) at /home/carlo/.julia/v0.5/AutoGrad/src/core.jl:39 in (::AutoGrad.#gradfun#2)(::Array{Float64,4}, ::Vararg{Array{Float64,4},N}) at /home/carlo/.julia/v0.5/AutoGrad/src/core.jl:39 in eval_user_input(::Any, ::Base.REPL.REPLBackend) at ./REPL.jl:64 in macro expansion at ./REPL.jl:95 [inlined] in (::Base.REPL.##3#4{Base.REPL.REPLBackend})() at ./event.jl:68 This could be fixed easily defining size(x::AutoGrad.Rec, d1::Integer, d2::Integer, dx::Vararg{Integer}) = size(getval(x), d1, d2, dx...) as I already tested. If someone points me to the right location where this function should reside I can file a PR — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#18>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABvNpjaWqvhx6mWjYW5ZDzx5NW-NHz9Tks5sJLGngaJpZM4OKSr6> .