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
- [Feature] Treat max_tokens as an upper bound when the prompt fits the context
- /v1/systemone requests are not aborted when the client disconnects (0.5.22)
- [CI] test_basic_sanity_dspark.py fails on main: decode vs prefill top-logprob mismatch exceeds 0.5 delta
- [Failure Tracker] PR Test (AMD)
- [CI] [AMD] disaggregation: test_disaggregation_pp.py fails with Mooncake v0.3.14 (aux data sends over same-node RDMA)
- [CI] [AMD] disaggregation: test_nixl_transfer_engine_e2e.py fails with NIXL_ERR_BACKEND (UCX destination unreachable)
- [Fix] moe: Quantize contiguous SiTU with the packed UE8M0 scale
- [Fix] docker: Relax frozen python-dateutil pin so runai-model-streamer[s3] resolves
- [Bug] PD decode: Qwen3.8-Flash-Next resumes with another request's QSA indexer state after retraction
- [Fix] qsa: Carry QSA indexer state through cpu_tensor retraction
- Docs
- Python not yet supported