AirLLM

Overview

AirLLM provides a powerful inference engine for 70B parameter models, optimized for single GPU usage. It is designed for developers and researchers looking to leverage large language models efficiently on limited hardware.