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 supported59 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Doc] Add startup-cache-image agent skill
- [Bug]: Mamba align prefix cache keys a partial-hit state copy at a block boundary that no chunk wrote when an internal prefill checkpoint is used
- [Bug]: GLM-5.3 (DSA sparse MLA) is not batch invariant under `VLLM_BATCH_INVARIANT=1`; today it is refused at startup
- [Core] Priority scheduling: preempt lower-priority requests for KV cache space
- [MoE] Add a generic fused gate router and backend interface
- [Model] Add GRT (Geometric Rotation Transformer) model support
- [CI/Build] Make Python 3.14 the default for images and CI
- [Bugfix][Frontend] Put required properties first in strict tool grammars
- [Bugfix][Metrics] Count preemptions in the scheduler, not on resume
- [Core]: DRY BlockPool.get_new_blocks with caching + metrics on/off variable factored out of the loop
- Docs
- Python not yet supported