Chinese-medical-instruction-tuning

基于医学知识图谱以及医学文献，结合ChatGPT API构建了中文医学指令微调数据集。医学知识库围绕疾病、药物、检查指标等构建，字段包括并发症，高危因素，组织学检查，临床症状，药物治疗，辅助治疗等，可以利用该数据集对ChatGLM或者LLaMA模型进行训练，提高模型在医疗领域的问答效果

A Chinese medical instruction fine-tuning dataset was developed using medical knowledge graphs, medical literature, and the ChatGPT API. The medical knowledge base integrated within this dataset is built around core medical concepts such as diseases, medications, and examination indicators, with its fields covering complications, high-risk factors, histological examinations, clinical symptoms, pharmacotherapy, adjuvant therapy, and more. This dataset can be employed to fine-tune models like ChatGLM or LLaMA, thereby improving their question-answering performance in the medical field.

搜集汇总

背景与挑战

背景概述

该数据集是一个中文医学指令微调数据集，基于医学知识图谱和文献构建，旨在训练大型语言模型以提升医疗领域的问答能力。它围绕疾病、药物和检查指标等医学概念，包含并发症、临床症状、药物治疗等多个字段，并通过GPT-3.5生成多样化的问答数据。数据集目前质量有限，正在持续迭代优化中。