Lightning-AI / Lightning-AI/lightning-thunder

Determine the configuration and setup end-to-end test and benchmark unmodified SGLang as baseline.

Open
#2,592 1 comment 0 reactions 1 assignee View on GitHub

@kshitij12345 is already working on this.

Since Oct 9, 2025.

Dominant language
Python
Stars
1.5k
Forks
121
PR merge metrics
No merged PRs in 30d

Description

For the models (Llama 4, GPT OSS, DeepSeek V3.1, and Qwen3-Next), we should determine the configuration of interest and write a script to test and demonstrate an end-to-end inference with default unmodified SGLang code.

NOTE:
* We are planning to integrate with sglang and sglang may have it's own implementation of this model. ([Eg](https://github.com/sgl-project/sglang/blob/main/python/sglang/srt/models/llama4.py))
* While setting up the CI job, we should remember that the end goal is to also support nvFP4. So we need figure out how to set this up in the CI with a machine that supports nvFP4 to track any regression.

Contributor guide

No contributing guide indexed for this repository

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.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.