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] phi4mm: Avoid dense chunk lookup masks
- [Bug] compressed-tensors: Untargeted layers fail with "Unable to find matching target"
- [Fix] compressed-tensors: Load layers that match no target as unquantized
- [Fix] spec: Keep routing captures aligned with accepted tokens
- [Fix] tokenizer_manager: Serialize disk weight updates while the engine is paused
- [Bug] dsv41: deepseek-recipe rendering differs from the checkpoint's encoding.py (tool namespaces, current-turn reasoning before tool results, action)
- [Perf] fa4: Use SplitKV for MTP target-verify and forward max_seqlen_k
- [Fix] mimo-v2: Defer fused-qkv FP8 scale_inv in MTP loader
- [Fix] scheduler: Reserve only the remaining decode when re-admitting a retracted request
- [Fix] disaggregation: Slice GDN conv state per head when TP exceeds key heads
- Docs
- Python not yet supported