AiMed: Artificial Intelligent Large Language Model for Medicine in China

With the continuous growth of medical information and the rapid development of artificial intelligence technology, processing medical knowledge efficiently and accurately becomes more and more important. In this study, we introduce a large language model called AiMed, which is specialized in processing Chinese medical information. AiMed is designed to solve multiple tasks such as medical knowledge QA, medical literature retrieval and reading, and it is expected to play an important role in medical research. The AiMed model is developed using a multi-stage process including pre-training, supervised fine-tuning, reinforcement learning based on human feedback, and direct preference optimization. These steps ensure high performance of the model in understanding medical terminology, analyzing complex contexts, and generating high-quality text. The codes and parameters of this study have been open-sourced and a series of experiments was conducted to verify the validity of AiMed.