sglang
https://github.com/sgl-project/sglang
Python
SGLang is a high-performance serving framework for large language models and multimodal models.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported1 Subscribers
Add a CodeTriage badge to sglang
Help out
- Issues
- [Fix] moe: Stabilize MXFP4 Marlin expert token ordering
- [Feat] platforms: Let platforms default unset profile request fields
- [Fix] disaggregation: Fail parked prefill requests when their decode goes away
- [Perf] [AMD] monokernels: Capture a routed step's DSA graph once, without the host read
- [Feat] [AMD] monokernels: Serve a draft spec's own steps, and skip the DSA replay metadata they never read
- [Fix] multimodal: Handle None data_token_pairs fallback in pad_mm_token_pairs
- [Perf] rust-server: Flatten egress logprob columns only for rows that carry them
- [Fix] gemma4: Refold the router scale when it is reloaded
- [Perf] diffusion: Wan-Animate-2 1.7x faster denoising via exact FA3 in-context attention, opt-in FP8 reference K/V
- [Fix] hicache: Disable the NIXL L3 cleaner in the NIXL storage tests
- Docs
- Python not yet supported