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] [AMD] aiter: Run the DSpark MHA draft on non-causal unified_attention
- [Fix] stop_regex: Record the matched span, not the pattern source
- [Fix] observability: Assign correct dp_rank under attention-dp for KV metrics and Prometheus
- [Perf] [AMD] attention: Attend AITER varlen's live prefix as one segment
- [Feat] dsv41-engram: Share Engram tables across P/D using EGM
- [Fix] sglang-miles: Clear the multimodal embedding cache on flush_cache and weight updates
- [Performance] Reproducible cache hit-rate dip during L2/L3 capacity pressure with HiCache + Mooncake
- [Fix] disaggregation: Back up separate draft KV pools in PD CPU retraction
- [Perf] [AMD] mori: pack KV pages for heterogeneous TP
- [Perf] [AMD] mooncake: pack heterogeneous-TP SWA KV pages
- Docs
- Python not yet supported