Lite-Me-LLaMA: Resource-Efficient Large Language Models for Medical Applications

To overcome the challenges of high costs, infrastructure demands, and data privacy concerns in medical applications of large language models (LLMs), we introduce Lite-Me-LLaMA, a suite of lightweight, open-source medical LLMs. By enhancing LLaMA3-8B with extensive medical pre-training and instruction fine-tuning, it is designed to be efficient, transparent, and readily deployable in clinical settings, requiring minimal computational resources for seamless integration into workflows while maintaining robust data privacy and security.