pyro-ppl / pyro-ppl/numpyro

Implement `RelaxedOneHotCategoricalStraightThrough`

Open
#559 3 comments 1 reaction 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement
Dominant language
Python
Stars
2.8k
Forks
315
Avg merge
3d 9h
Merged PRs (30d)
27

Description

Following #548 discussion, and while we wait for discrete latent variables, it would be nice to have a Gumbel-Softmax categorical approximation as featured in Pyro. Didn't realize this was the name given to Gumbel-Softmax in Pyro, but hopefully replication might be straight-forward?

numpyro (i.e. Jax) seems uniquely suited for problems involving large discrete structures (e.g. networks), so an ability to recover latent discrete variables (or their approximations) would be fantastic!

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reading the #548 discussion and the corresponding Pyro implementation of RelaxedOneHotCategoricalStraightThrough. Determine how the distribution should fit NumPyro's existing distribution APIs; done means the named approximation is implemented and its behavior matches the intended Pyro semantics.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.