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
- [RFC]: MoRIIO WRITE failure handling and safe KV block reclamation
- Fix suppress_stdout to target the real stdout fd
- [Bug] DeepSeek-V4.1-Flash cannot use TP > 8: Engram hash-head sharding assertion rejects tp_size=16 across 2 nodes
- [Bug] vendored flash_linear_attention chunk_o.py drops upstream mask before exp(), so boundary chunks can exponentiate unspecified lanes
- [KV Cache] Skip the zero fill in allocate_kv_cache when the platform does not need it
- [Startup UX]: Serial per-process Python imports across the API server → EngineCore → worker tree add significant startup time
- [Bugfix] Propagate SIGTERM during engine startup
- fix: prevent path traversal outside cache directory in FilesystemResolver (#59270)
- [HiSparse] Standardize terminology and add a nomenclature to the design doc
- [Performance]: FlashAttention backend rebuilds KV-cache views on every call (+~21 µs CPU/layer since #44455)
- Docs
- Python not yet supported