DiffOpt.jl

DiffOpt.jl copied to clipboard

Best practice for DiffOpt.jl implementation with Flux (logsumexp)

20

Hi, developers! Thanks for this promising and potentially useful package. I'm studying differentiable convex optimisation and trying to implement it to the [PLSE](https://scholar.google.com/citations?view_op=view_citation&hl=ko&user=K7wrYmoAAAAJ&citation_for_view=K7wrYmoAAAAJ:RYcK_YlVTxYC), a neural network that I proposed. I...

JinraeKim

Currently, we only test giving QPs in standard form and use QPDiff or giving CPs in standard for and use ConicDiff. The user might not give these form and get...

blegat

Replace Ipopt by HiGHS

1

joaquimg

fix `@show` in the end of Differentiating trivial QP 1

2

joaquimg

Add missing get method for vector contraints

2

```julia function MOI.get(model::Optimizer, ::ForwardIn{ConstraintCoefficient}, vi::VI, ci::CI{F,S}) where {F,S} ``` only work for scalar

joaquimg

put tests into functions

2

joaquimg

Add missing tests in the end of: Differentiating conic with PSD and POS constraints

1

joaquimg

joaquimg

joaquimg

joaquimg