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
- [Bug]: Cross-request context contamination via shared-prefix KV reuse in PD disaggregation with automatic prefix caching — clean with `--no-enable-prefix-caching`
- [Bug]: VLLM_MAX_AUDIO_CLIP_FILESIZE_MB=0 rejects non-empty audio
- [Bug]: MTP draft ignores target `--revision` (loads default branch) – regression vs 0.29, can crash with ZeroDivisionError
- [Bugfix][KV Connector] NIXL: reject a known engine_id seen at a different address
- [Bugfix][ROCm] Stop forcing legacy HSA IPC in ROCm images
- [Perf][Qwen3-VL] Resize video frames during decoding to reduce peak m…
- [Kernel] Specialize Triton MoE on fixed weight dimensions
- [Perf][LoRA] Avoid empty split-K programs in shrink
- [Bugfix][Distributed] Prevent silent split-brain in stateless DP on port collision (#61049)
- [ROCm][Perf] Add native per-token FP8 Triton linear backend and RDNA4 tuning
- Docs
- Python not yet supported