vllm
https://github.com/vllm-project/vllm
Python
A high-throughput and memory-efficient inference and serving engine for LLMs
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 supported62 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Feature][Observability] Attribute inference metrics and traces to API surface
- [Bug]: Stop sequences match inside reasoning, so responses come back empty (breaks Claude Code auto mode)
- [Bug]: Adaptive verification fails at startup with the default attention backend on FA2 GPUs
- [Bug]: `--load-format ipc_cache` leaves MoE routers on meta placeholders (illegal memory access, meta-tensor errors)
- [Attention][KV Connector] FlashInfer: optional separate K/V head slots for unquantized KV
- [Observability] Add tool-calling inference tracing
- [Bug]: /pause clears SHM multimodal cache, can cause fatal ValueError for queued requests
- [Bugfix][Frontend] Preserve HTTP error response headers
- [RFC]: Request-scoped auxiliary decision heads on a shared generative backbone
- [Bug]: Qwen3.6-35B-A3B NVFP4 on Jetson AGX Thor (sm_110): illegal memory access in GDN `_warmup_prefill_kernels` on cold-cache boots (3 of 4); warm caches 0 of 3
- Docs
- Python not yet supported