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] dsv4: HIP flat RoPE kernel with one cos/sin load and int64 offsets, bitwise unchanged
- [Fix] marlin_moe: Load grouped zero points at group boundaries
- [Perf] DSv4.1: cache image preprocessing across requests
- [Chore] deps: Add nvidia-cuda-tileiras for the optional cuTile router
- [Chore] docker: Pin Dynamo and grpcio-tools in the Beam image
- [CI] [NPU] tests: Add GSM8K PR gate for DSV4-Flash W8A8 8P GPQA
- [Perf] hybrid-swa: Detect hybrid SWA models from layer_types
- [Fix] [NPU] Correct DSA Models Sparse offload index-cache budgeting
- [observability] Preserve tracing through the multi-tokenizer router
- [Perf] [AMD] dsa: Share KV pages across MTP verify rows in the gfx950 indexer logits kernel
- Docs
- Python not yet supported