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 supported60 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Bugfix][ROCm] attention: Bound AITER FA varlen graph support
- [CI] Lock cmake and restore (CPU) Params, Env, Tokenizers, Parser job
- [RFC]: Load plan — an engine-neutral API for sharded weight transfer
- [Feature][Elastic EP] Coordinate native local-range resizing
- [Feature]: [MRV2][DBO] Report executed microbatch counts and decision reasons
- [Frontend] Serve all multi-port DP ranks from one Rust frontend
- [Bug][Spec Decode] Model Runner V2 + DSpark at n=8: target degenerates to one repeated token while acceptance reports 87.5%
- [Bug]: Gemma4 MoE cannot load PEFT expert LoRA adapters without --enable-mixed-moe-lora-format
- [Bugfix][LoRA] Use 3D MoE LoRA wrapper for Gemma4
- [Misc][MoE] Support Qwen4Exp in benchmark_moe.py
- Docs
- Python not yet supported