Bencao - LLaMA fine-tuning model based on Chinese medical knowledge

! [GitHub - SCIR-HI/Huatuo-Llama-M Med China: Repo for BenTsao [original name: HuaTuo], Llama-7B tuned with Chinese medical knowledge. Bencao (formerly known as HuaTuo) model repository, fine tuned LLaMA model instructions based on Chinese medical knowledge]

Bencao [formerly known as HuaTuo]: LLaMA fine-tuning model based on Chinese medical knowledge

This project has open-source the LLaMA-7B model that has undergone Chinese medical instruction fine-tuning/instruction fine-tuning. We constructed a Chinese medical instruction dataset using a medical knowledge graph and GPT3.5 API, and based on this, fine tuned LLaMA's instructions to improve its question answering effectiveness in the medical field.

Based on the same data, we also trained a medical version of the ChatGLM model: ChatGLM-6B Med

In addition, we also attempted to incorporate the conclusions from medical literature as external information into multiple rounds of conversations using the GPT3.5 API, and fine tuned LLaMA instructions based on this. At present. We only open model parameters trained for a single disease of 'liver cancer'. In the future, we plan to release a medical dialogue dataset that incorporates literature conclusions and train models for 16 diseases related to liver, gallbladder, and pancreas.