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 supported49 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [ROCm] Cut 69 wasted contiguous copies per decode step from the skinny GEMM path
- [Feature][KV Connector] Add UMBP core and embedded DRAM offloading
- [ROCm][Perf][GLM-5.3-Flash] Improve ragged sparse prefill kernel
- [Security] Cap OpenPangu video sampling at the source frame count
- [Feature][Core] Add opt-in multi-worker CUDA process checkpoint sleep
- [Bug]: New watermarking feature (v0.30.0) allows per-request opt-out even when server enforces --watermark-config
- [Bugfix][KV Offload] Namespace KV offload cache by layout
- [KV Offload] Request-free CPU->GPU prefetch primitive
- [Bugfix] Don't assert on already-freed request in _update_from_kv_xfer_finished
- [Frontend] Add watermark_config.enforce to prevent per-request watermark opt-out
- Docs
- Python not yet supported