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
- [Bug] gpt-oss on /v1/chat/completions drops replayed reasoning: the template reads `thinking`, SGLang only forwards `reasoning_content`
- [Fix] srt: cap finished_len at max_new_tokens on request abort with speculative decoding
- [Bug] FP8 block-quantized checkpoint on SM86 (FP8 Marlin) intermittently gives wrong tokens with the overlap scheduler; exact without it
- [Fix] multimodal: Convert non-RGB images to RGB in decode_media_snapshot
- [Bug] /v1/messages returns HTTP 500 for gpt-oss when the history contains a thinking block
- Exchange latent-width MoE rows under DP attention (Nemotron-H, Kimi K3)
- [Test][NPU] kimi_k3: aisbench cold-prefill perf trial (16k in / 1 out) on 4-node mix
- [Refactor] platforms: route device and stream state through current_platform
- [Perf] gemma3: Pool the projector input in channels-last
- [Fix] humming: Respect SGLANG_CACHE_DIR for cache and temporary files
- Docs
- Python not yet supported