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] benchmark: Honor chat stream overrides when decoding responses
- [Perf] [NPU] multimodal: Add bounded cross-request Qwen3-VL image batching
- [Fix] metrics: Populate pending_prealloc_token_usage and add pre_allocated_token_usage
- [Fix] embedding: Do not reuse prefixes or split prefills for encoder models
- [Fix] attention: Do not default to FlashInfer for head dims FA2 prefill cannot run
- [Fix] trinity: Keep think tags inside tool-call arguments
- [Feat] deepseek-v4.1: Add experimental resident-weight recovery and warm standby
- [Fix] auth: Require admin API key for HiCache clear and remote weight export
- [Fix] zimage: Expand text conditioning for multiple outputs per prompt
- [PD] Bound the mooncake early-registration room lifecycle
- Docs
- Python not yet supported