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
- [Scheduler] Add cache-hit-aware over-admission for bounded Mamba requests
- [Bug] Deterministic inference hangs when chunked prefill is smaller than the prefill alignment
- Enable and fuse the Inkling MoE gate epilogue on Intel XPU
- [Bug] [ROCm] DeepSeek-V4-Flash FP8 on MI300X (gfx942): output degrades with context length — coherent <100 tokens, garbage by ~300, NIAH 0/3
- Qwen3.8-Flash-Next QSA crashes with FP8 KV cache (BF16 q vs FP8 k dtype assertion)
- [Bug] DFlash2 state corruption with concurrent requests
- fix(attn): treat DECODER_BIDIRECTIONAL as non-causal on FlashInfer prefill
- [Bug] GLM-5.3-Flash thinking output degenerates into repeated '!' under multi-tool agentic prompts
- [Fix] mistral: only the first tool call of an array survives streaming
- [Bug] GLM-5.3-Flash (glm5_next) crashes on startup with --moe-runner-backend flashinfer_trtllm: IndexError index 288 out of bounds in logical_to_all_physical
- Docs
- Python not yet supported