FluxML / Optimisers.jl

Optimisers.jl defines many standard optimisers and utilities for learning loops.
https://fluxml.ai/Optimisers.jl
MIT License
72 stars 20 forks source link

doc improvement: working with custom model types #84

Open CarloLucibello opened 2 years ago

CarloLucibello commented 2 years ago

We should give more prominence in the docs to the usage of Flux.@functor and Optimisers.trainable to define trainable parameters of custom types.

mcabbott commented 2 years ago

Yes. Where's the best place?

I don't want every docstring to become a novel, but perhaps the one for Optimisers.setup is one good place to show that this recurses into custom structs.

The main doc page uses a Flux model as the first example of applying a rule. It could instead first construct some simpler model, and show that a newly written optimisation rule is applied to all children etc?