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
- [Mamba] Add opt-in prefill-only radix caching (--radix-cache-skip-decode-insert)
- [Bugfix] Disable mixed chunk for PP speculative decoding
- [XPU] Dispatch the Inkling gate epilogue to sgl-kernel-xpu
- [Bug] NPU HiCache: MHATokenToKVPoolHost crashes when the device k_buffer is a single tensor (since #40326)
- [sgl-router] Serve native gRPC ChatComplete
- [XPU] Norm Inkling's q in place on the packed QKVR row via sgl-kernel-xpu
- [Fix] Honor --max-total-tokens for unified Mamba memory
- feat: explicit cache_id / load_cache_id for KV-cache save/load
- [sgl-router] Serve the remaining native gRPC inference RPCs
- [Docs] Update the ngram speculative decoding notes: devices, overlap scheduler, external corpus flags
- Docs
- Python not yet supported