Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

More robust BroadcastStyle handling #43

Merged
merged 1 commit into from
Dec 30, 2023
Merged

More robust BroadcastStyle handling #43

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 2 additions & 4 deletions src/array.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -143,11 +143,9 @@ end
Adapt.adapt_structure(T, x::OneHotArray) = OneHotArray(adapt(T, _indices(x)), x.nlabels)

function Base.BroadcastStyle(::Type{<:OneHotArray{<:Any, <:Any, var"N+1", T}}) where {var"N+1", T <: AbstractGPUArray}
# We want CuArrayStyle{N+1}(). There's an AbstractGPUArrayStyle but it doesn't do what we need.
# We want CuArrayStyle{N+1}(). There's an AbstractGPUArrayStyle but it doesn't do what we need.
S = Base.BroadcastStyle(T)
# S has dim N not N+1. The following hack to fix it relies on the arraystyle having N as its first type parameter, which
# isn't guaranteed, but there are not so many GPU broadcasting styles in the wild. (Far fewer than there are array wrappers.)
(typeof(S).name.wrapper){var"N+1"}()
typeof(S)(Val{var"N+1"}())
end

Base.map(f, x::OneHotLike) = Base.broadcast(f, x)
Expand Down
Loading