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 supported24 Subscribers
Add a CodeTriage badge to vllm
Help out
- Issues
- skip non-FA attention layers in flash sliding-window scan
- [Feature]: Pre-ViT visual token pruning for VLMs (PixelPrune)
- [Distributed] Fix PCIe custom-allreduce eligibility on >2 GPU topologies
- [ROCm] Use unified decode fallback for sliding-window AITER FA
- Revert "fix(moe): fix RoutedExpertsCapturer assertion failure with DP>1 and MK path" (#37879)
- [Bugfix] Add bias to SKIP_TENSORS to fix online FP8 for models with biased linears
- [Bugfix] Fix Gemma 4 pipeline parallelism crash — IntermediateTensors .get() and None values
- [RFC]: Async parallel startup for EngineCore processes in DP/TP scenarios
- [Bug]: Qwen3.5 `thinking_token_budget` causes `reasoning_end_str` to leak into `content` field
- [CI/Build] Fix use_existing_torch.py to recursively match requirements subdirectories
- Docs
- Python not yet supported