search

lanl-ansi / MathOptAI.jl

Embed trained machine learning predictors in JuMP
https://lanl-ansi.github.io/MathOptAI.jl/
Other
29 stars 1 forks source link

[breaking] return a formulation object #80

Closed odow closed 1 month ago

odow commented 2 months ago

Experiment with #67

I don't really like this yet. I need to look better at how other people do this.

odow commented 1 month ago

So. This is ready for review.

But I'm still not convinced what the best approach is.

odow commented 1 month ago

On réflection. I'm leaning towards merging this. It is a pretty disruptive PR to keep up to date with merge conflicts.

odow commented 1 month ago

I'm giving a talk about MathOptAI on the 24th, and I'd like this to be merged before then. The current implementation might not be everyone's ideal, but I think it's an okay start, and we can always change in future?

odow commented 1 month ago

I'm going to merge this, but leave #67 open, and we can continue to iterate.