Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Implement RelaxedOneHotCategoricalStraightThrough #559

Open
rtbs-dev opened this issue Mar 13, 2020 · 3 comments
Open

Implement RelaxedOneHotCategoricalStraightThrough #559

rtbs-dev opened this issue Mar 13, 2020 · 3 comments
Labels
enhancement New feature or request

Comments

@rtbs-dev
Copy link

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!
@rtbs-dev
Copy link
Author

Pyro link for the original implementation.

@daydreamt
Copy link
Contributor

If no one else commits this until then (and no hard feelings if they do), I will give this a stab starting next weekend.

@neerajprad neerajprad added the enhancement New feature or request label Mar 23, 2020
@neerajprad
Copy link
Member

@daydreamt - Please go ahead, you will be assured a thorough and timely review. If you have any questions around the codebase, please let us know.

@daydreamt daydreamt mentioned this issue Apr 6, 2020
Open
17 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
enhancement New feature or request
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@daydreamt @neerajprad @rtbs-dev