I often use Convex.jl and Flux.jl. When I use them, it is really tough to combine them. Especially, I wanna train a network using Flux.jl when the loss function includes optimal values by solving optimisation problems using Convex.jl. For example, for d data points, Flux.jl usually provides a netw…

Best practice to combine Flux.jl and Convex.jl

iHany March 14, 2021, 2:18pm 3

OMG, it sounds like it would be impossible for me…
Bad news. Thanks.

Do you think it’s possible to differentiate Convex.jl problems by expanding some methods?

Best practice for Flux.jl: how to untrack gradient?

Topic		Replies	Views
Is it possible to solve "convex" neural network using `Convex.jl`? General Usage question , flux , optimization	5	598	December 23, 2020
Best practice for Flux.jl: how to untrack gradient? General Usage optimization , machine-learning	1	559	March 15, 2021
Can I extend other functions to make it compatible with Convex.jl (element-wise operation) Optimization (Mathematical) optimization	0	281	December 29, 2020
[ANN] FluxOptTools Package Announcements optim , flux , visualization	0	989	July 1, 2019
Convex.jl and Icnn Optimization (Mathematical)	4	438	May 12, 2022