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] Accept zero-valued token IDs in completion prompts
- [Fix] Fall back from fused KV writes for RL native RoPE
- [Fix] Preserve LFM2 tool calls across streaming chunks
- [Fix] Preserve text around non-streaming LFM2 tool calls
- [Fix] Enforce API key auth for WebSocket routes
- [HiCache] Attribute asynchronous CUDA faults to the host-pool transfer that caused them
- [DSA] tilelang fp8 sparse decode: fix NoPE (d_tail=0) crash
- [Feature] Make SGLANG_DEBUG_MEMORY_POOL effective on the default page_size=1 allocator
- [DSV4.1][PD][1/N] Support prefill node use cp parallel strategy
- [Fix] Exclude output manifest when regenerating model checksums
- Docs
- Python not yet supported