Skip to content

WIP: bugfix AD #126

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

Draft
wants to merge 5 commits into
base: master
Choose a base branch
from
Draft

WIP: bugfix AD #126

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
57be075
support ScalMat in FiniteGP observation covariance
st-- Apr 1, 2022
693147d
WIP: ProjectTo type piracy AD fix
st-- Apr 1, 2022
4be201b
Merge branch 'master' of github.com:JuliaGaussianProcesses/Approximat…
st-- Apr 2, 2022
09044de
Update src/ApproximateGPs.jl
st-- Apr 2, 2022
a9f748f
Merge branch 'master' into st/bugfix-ad
st-- Apr 8, 2022
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: 6 additions & 0 deletions src/ApproximateGPs.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,4 +23,10 @@ include("deprecations.jl")

include("TestUtils.jl")

import ChainRulesCore: ProjectTo, Tangent
using PDMats: ScalMat
ProjectTo(x::T) where T <: ScalMat = ProjectTo{T}(; dim=x.dim, value=ProjectTo(x.value))
(pr::ProjectTo{<:ScalMat})(dx::ScalMat) = ScalMat(pr.dim, pr.value(dx.value))
(pr::ProjectTo{<:ScalMat})(dx::Tangent{<:ScalMat}) = ScalMat(pr.dim, pr.value(dx.value))

end
8 changes: 4 additions & 4 deletions src/LaplaceApproximationModule.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -179,16 +179,16 @@ function _check_laplace_inputs(
end

struct LaplaceCache{
Tm<:AbstractMatrix,Tv<:AbstractVector,Td<:Diagonal,Tf<:Real,Tc<:Cholesky
Tm<:AbstractMatrix,Tv1<:AbstractVector,Tv2<:AbstractVector,Tv3<:AbstractVector,Td<:Diagonal,Tf<:Real,Tc<:Cholesky
}
K::Tm # kernel matrix
f::Tv # mode of posterior p(f | y)
f::Tv1 # mode of posterior p(f | y)
W::Td # diagonal matrix of ∂²/∂fᵢ² loglik
Wsqrt::Td # sqrt(W)
loglik::Tf # ∑ᵢlog p(yᵢ|fᵢ)
d_loglik::Tv # ∂/∂fᵢloglik
d_loglik::Tv2 # ∂/∂fᵢloglik
B_ch::Tc # cholesky(I + Wsqrt * K * Wsqrt)
a::Tv # K⁻¹ f
a::Tv3 # K⁻¹ f
end

function _laplace_train_intermediates(dist_y_given_f, ys, K, f)
Expand Down
4 changes: 2 additions & 2 deletions src/SparseVariationalApproximationModule.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@ using LinearAlgebra
using Statistics
using StatsBase
using FillArrays: Fill
using PDMats: chol_lower
using PDMats: chol_lower, ScalMat

using AbstractGPs: AbstractGPs
using AbstractGPs:
Expand Down Expand Up @@ -306,7 +306,7 @@ Statistics. PMLR, 2015.
"""
function AbstractGPs.elbo(
sva::SparseVariationalApproximation,
fx::FiniteGP{<:AbstractGP,<:AbstractVector,<:Diagonal{<:Real,<:Fill}},
fx::FiniteGP{<:AbstractGP,<:AbstractVector,<:Union{Diagonal{<:Real,<:Fill},ScalMat}},
y::AbstractVector{<:Real};
num_data=length(y),
quadrature=DefaultExpectationMethod(),
Expand Down