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] scheduler: Publish KV events on every decode step
- [Bug] FlashInfer prefill is not split-invariant: radix-cache hit or chunked prefill flips greedy output on a Qwen3.5 hybrid GDN model
- [Fix] midashenglm: Remove full prompt logging
- [Fix] MegaMoE: stop double-dividing the DP prefill budget
- [Fix] deterministic: Run the MoE routing functions eagerly in batch-invariant mode
- [Fix] disagg: Size the decode reserve by what a request can still decode
- [Fix] [NPU] modelslim: Fix grouped W4A8 MoE loading and TP scale sharding
- [Fix] session: Order a waiting streaming-session turn under dfs-weight
- [Fix] diffusion: Keep each request's rollout trajectory and conditioning under dynamic batching
- [Fix] session: Keep a retracted streaming turn in flight
- Docs
- Python not yet supported