tensorflow / tensorflow/probability

Best way to use/implement MixtureLogistic layer with tfp in VAE setting

Open
#1,051 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

layers
Dominant language
Jupyter Notebook
Stars
4.4k
Forks
1.1k
PR merge metrics
No merged PRs in 30d

Description

Hi,

I would be interested in using a discretized mixture of logistic layer for a Decoder in a VAE settings (similar to NVAE). Exploring the tfp API, it seems like the most appropriate layer would be the MixtureLogistic layer. Does it make sense? Would it work in tf.keras model built with keras functional API?
I will just try but wanted to ask whether you would suggest any other approach, or whether, for that purpose, I should go on and re-write the layer in TF (saw some implementations in pytorch around).

Thank you for any feedback!

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

The issue names the TFP MixtureLogistic layer, tf.keras Functional API, and the NVAE paper, but no repository files or tests. Start by reviewing the linked layer API and the NVAE decoder usage; a useful outcome would be a confirmed integration approach or a documented alternative.

Written by the indexing model from the issue text.

Assessment

Tech stack
tensorflow
Domain
machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
15/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.