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
- [CI] tests: Rendezvous the tile-sharding gloo test through a file store
- [Fix] hicache: Withhold only protected bytes from page-cache credit
- [Perf] minimax-h3: Skip attention for the packing's alignment rows in the Ulysses pipeline
- [Fix] [AMD] quark: Widen the Quark MXFP4 W4A16 MoE gate to consumer RDNA and fail clearly without aiter
- feat: add NCCL EP as optional moe-a2a-backend (low-latency decode)
- [Test] [Ascend] npu: Add GLM-5.2 2P2D gpqa cases and unify evalscope retry loop
- [Perf] [AMD] attention: Forward decode graph variants through wrapper backends
- [Fix] scoring: Return token_logprobs for setwise CausalLM scoring
- [Perf] hicache: Space out write_back eviction waits with a bounded reserve
- [Fix] stop: Store the matched regex span, not the pattern
- Docs
- Python not yet supported