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 supported35 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Bug]: MXFP4 MoE Marlin fallback crashes with illegal memory access on Hopper (DeepSeek V4, TP+EP) — non-tile-aligned intermediate size
- [Bugfix] Add DSV4-Flash-DSpark gsm8k and pipe kv_quant_mode through SWA MLA unify
- [CPU] Add MoE DP/EP support
- [Bug]: DSV4 sparse MLA emits garbage (gsm8k ~0%) with the packed KV layout from #44577
- [Core] Refactor Flashinfer BatchPrefill* / BatchDecode* and trtllm_batch* wrappers' forward paths.
- [Bug]: Qwen3-Omni use_audio_in_video via --mm-processor-kwargs is broken in online serving: engine crash at startup, audio silently never fed, interleave placeholder/mRoPE errors
- [RFC]: Separate Flashinfer GQA into TRTLLM and flashinfer backends.
- [ROCm][DSV4][Perf] Step 1/N: Port ATOM attention into vLLM: port single-node single-token-prediction case
- [Core][Hardware][NVIDIA] Add custom all-reduce suspend hooks
- [WIP][Spec Decode] DSpark confidence-scheduled verification
- Docs
- Python not yet supported