Add reset function for Diagonal QN operators, fix bug SpectralGradient #266
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Following JuliaSmoothOptimizers/NLPModelsModifiers.jl#94, I added
reset!
functions that restore the QN operators to the identity operator.At the same time, I spotted a bug with
SpectralGradient
where the push was not updatingd
in the closureprod = (res, v, α, β) -> mulSquareOpDiagonal!(res, d, v, α, β)
(becaused
was a Real).I fixed it by changing
d
to[σ]
so that modifyingd[1]
modifiesmulSquareOpDiagonal!(res, d, v, α, β)
accordingly.This modifies the way we update a
SpectralGradient
: we have to doop.d[1] = 2.0
instead ofop.d = 2.0
, but since we are supposed to update this operator withpush!
, this should not break code.Does the adding of the
reset!
function requires a new minor release2.6.0
, or can we stick to a patch2.5.1
?