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
- [AMD] Pass USE_PDL explicitly to Triton kernel launches
- [Bug] AttributeError: 'list' object has no attribute 'tolist' in move_logprobs_to_cpu
- [Observability] Support custom OTLP service names
- [Bugfix] Fix TVM FFI tuple access in CUDA IPC JIT
- Qwen3.8-27B NVFP4 on RTX 5090: cookbook mem-fraction OOMs at decode-graph capture (~5GB) and torch.compile + decode graph is a 6x regression
- [NPU] Use fused Triton MRoPE for Qwen3.x
- Triton version mismatch
- [Bug] Qwen3.5 MTP weights silently dropped on compressed-tensors checkpoints (accept len pinned at 1.0)
- [Bug] Using hicache in the main branch of deepseekv4-flash will result in illegal memory access.
- [Bug] Grammar token sync initializes a singleton NCCL group under DP attention
- Docs
- Python not yet supported