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
- [Perf] [NVIDIA] kimi-k3: Bound q6 preparation by server concurrency
- [Feat] [NPU] prefill: breakable prefill graph for DSV4 and ascend_ops MoE
- [Perf] flashinfer-autotune: Tune MoE with varied warm-up tokens
- [Feat] [NPU] dsa: Store the DSA MLA KV cache as INT8 on Ascend A2
- [Bug] rust-server: /generate text prompts get no BOS when the tokenizer relies on add_bos_token
- [Fix] sampling: Reject oversized n before it fans out
- [Feat] mem_cache: Evict Mamba states by GreedyDual-Size (--radix-eviction-policy gds)
- [Refactor] graph: Decouple eager_on_graph from CUDA capture (2/3)
- [Perf] hicache: Overlap PD-prefill write_back eviction backups with the in-flight forward
- [Perf] kimi-k3.1: Breakable prefill CUDA graphs under DCP, including prefix-hit extends
- Docs
- Python not yet supported