NVIDIA / NVIDIA/TransformerEngine

Enable Router custom call from JAX

Open
#2,710 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
3.5k
Forks
831
Avg merge
3d 11h
Merged PRs (30d)
65

Description

Current router kernels are present in common and callable from Pytorch side but not JAX. Need to support JAX router for either standalone use or later intergation to Maxtext moe layer.

These router kernels are CUDA kernels so using jax ffi lowering is sufficient.

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 locating the existing common router kernels and their PyTorch-side call path. Then investigate how JAX FFI lowering can expose the CUDA kernels for standalone router use or later MaxText MoE integration. Done means JAX can call the router kernels successfully.

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.