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
- [CI] [AMD] pr-test: Move test_quark_mxfp4 from the MI35x 1-GPU PR stage to nightly
- [MiniMax-M3] GB300 cookbook recipe leaves ~11-15% median interactivity on the table: dev image predates in-tree trtllm_mha page-128 support, and the recipe never engages the faster decode backend
- [Perf] [AMD] Qwen3.5: fuse the MXFP4 input quantization of AITER ASM FP4 GEMMs into the preceding RMSNorm and SiLU-and-mul (bit-exact)
- [Perf] [AMD] gdn: Tune the ROCm decode Conv1D update and recurrent state streaming
- [Feat] observability: Report a TTFT phase breakdown in response meta_info
- [Perf] lora: Avoid O(tokens) per-token expansion in build_lm_head_pass_segments
- [Fix] benchmark: Share request logging across Python backends
- [RFC] Model Independent Parallelism Layer Boundaries and Weight Loading
- [Fix] mooncake: Await all transfer futures before reporting failures
- [Fix] disaggregation: Retain failed drained-ACK deliveries for retry
- Docs
- Python not yet supported