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
- [Feat] disagg: Make Mooncake session failure blacklisting threshold cā¦
- [Fix] [AMD] quantization: make aiter optional for the quark schemes imported by every model
- [Fix] [AMD] layernorm: call vLLM fused_add_rms_norm with its in-place signature
- [Fix] [AMD] mla: Cap the absorb-BMM waves-per-EU budget so gfx950 can compile it
- [Perf] hicache: Wait once per write_back eviction walk, not once per victim
- [Fix] hicache: Restore sub-window SWA L3 hits under a mid-tree anchor
- [Feat] hicache: Store MXFP8 KV pages in L3 as payload + scales objects
- [Feat] PD: Decode KV offload for Mamba and SWA hybrid models
- [Feat] KVShard: Support KV Sharding with Mooncake storage
- support API key in multi-tokenizer mode
- Docs
- Python not yet supported