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]: Scheduler-owned KV transfer obligations: release blocks only after both recv and send complete
- Opt-in request security middleware for the API server (rate limiting, IP lists, payload screening)
- [Perf][Sampler] Add opt-in FlashInfer Cake top-k/top-p sampling
- [Bugfix][Spec Decode] Include draft models in DeepGEMM warmup
- [ROCm][Perf][Kimi-K3] Specialized decode on gfx942 (MI325X)
- [Bugfix] Reject non-int values for integer SamplingParams fields
- [Bugfix][KV] Ignore stale Mooncake completion notifications
- [Bug]: GB10 (sm_121) engine hangs forever inside a cuDNN FP8 GEMM picked via FlashInferFP8ScaledMMLinearKernel; CutlassFP8ScaledMMLinearKernel avoids it
- [Model] Apply Transformers checkpoint renamings during weight loading
- [Bugfix][GPT-OSS] Scope engine-level structured outputs to the Harmony final channel
- Docs
- Python not yet supported