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
- [ROCm][RDNA][Perf] W4A16: fused scale/zero-point carrier for Triton prefill
- [RFC]: R3 delivery protocol: keys, sharing scope, write/read/release contract
- [RFC]: R3 as a KV cache layer: where routed-experts rows of prefix-cached blocks live
- [Feature][KV Connector] Add spec-owned transfer region export
- [Model] Muse Spark: support compressed-tensors MXFP4 experts on NVIDIA
- [KV Offload] SimpleCPUOffload: opt-in cap on GPU blocks pinned by pending stores
- [Quantization] Support CT channel FP8 requantization
- [Platform][TPU] Fail fast with installation guidance when tpu_inference is missing
- [ROCm][Qwen3.8-Flash-Next] gfx950 fast path for the mxfp4_qsa QSA reader (stacks on #64123)
- [Perf][MoE] OCP MX emulation: dequantize only the routed experts
- Docs
- Python not yet supported