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
- [Fix] multimodal: Preserve strong identities across VLM caches
- [Perf] [AMD] gdn: Use strided prefill projection views on ROCm
- [Fix] function_call: Keep a split InternLM start marker for the next call
- [Fix] tokenizer: Use the model revision pin in post-load fixes
- [Fix] hicache: Enable DFlash L2/L3 caching with DCP
- [Feat] openai: Also return chat reasoning text as `reasoning`
- [Fix] benchmark: Pass req_to_token_pool to bench_one_batch dummy tree cache
- [Doc] minimax-m3: Route GB300 decode to trtllm_mha in the cookbook recipe (prefill stays fa4)
- [Fix] openai: Refuse tools + output constraint consistently instead of silently dropping the tool-call grammar
- [Perf] [NVIDIA] speculative-decoding: Add generic CUDA runtime fast paths
- Docs
- Python not yet supported