A high-throughput and memory-efficient inference and serving engine for LLMs

Transformers-compatible library for applying various compression algorithms to LLMs for optimized deployment with vLLM

Community maintained hardware plugin for vLLM on Spyre

Community maintained hardware plugin for vLLM on Ascend

This repo hosts code for vLLM CI & Performance Benchmark infrastructure.

Evaluate and Enhance Your LLM Deployments for Real-World Inference Needs

vLLM’s reference system for K8S-native cluster-wide deployment with community-driven performance optimization

Cost-efficient and pluggable Infrastructure components for GenAI inference