tumorailab/Protein2Text-QA
收藏官方服务：
资源简介：
Protein2Text-QA数据集是为了生成基于蛋白质序列的蛋白质功能的人类可读解释而设计的。它包含了从PubMed Central (PMC)文章中使用LLaMA3.1-8B-Instruct模型生成的约210,000个问答对。数据集分为预训练数据集、微调数据集、测试集、零样本集以及跨领域数据集。
The Protein2Text-QA dataset is designed to generate human-readable explanations of protein functions based on protein sequences. It consists of approximately 210,000 QA pairs generated from PubMed Central (PMC) articles using the LLaMA3.1-8B-Instruct model. The dataset is divided into pretraining, fine-tuning, test, zero-shot, and cross-domain subsets.
提供机构：
tumorailab
