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
- [mem_cache][Fix] Retain undonated non-int8 Mamba checkpoint metadata
- [AMD][V4.1] Cheaper mHC reduce + sinkhorn row, bitwise unchanged
- [Feat] epd: Support and optimize EPD performance for DeepSeek V4.1 Flash
- [Bug] sglang-mm: token-layout offsets wrap at u32 for prompts expanding past 2^32 tokens, silently corrupting M-RoPE positions
- [mm] Reject token-layout media spans that wrap the u32 offsets pair
- [AMD][V4.1] Trim more per-cycle launches in DSpark decode
- [AMD] Skip the unused PP-spec compare on the GDN and KDA verify path
- [Bug] PD disagg: bootstrap_room normalization diverges between the Rust worker, the Python path, and the sgl-model-gateway router — n>1 chat requests to a Rust PD worker are always rejected
- [rust-server] expand PD bootstrap rooms across parallel choices
- [Bug] gemma-4-31B-it: audio request crashes the scheduler; /get_model_info misreports has_audio_understanding: true
- Docs
- Python not yet supported