Cleves-Jain

该数据集包含了1149个化合物，规模相对较小，涵盖了22个不同的药物靶点，并为模型训练提供了每个靶点2-3个模板活性化合物，同时还包括了850个诱饵化合物。该数据集旨在通过模板化合物，从诱饵化合物池中识别出活性化合物，为药物发现提供少样本学习任务。其规模被归类为小型，任务类型为药物发现领域的少样本学习。

This dataset consists of 1,149 compounds with a relatively small scale. It covers 22 distinct drug targets, and provides 2 to 3 template active compounds for each target to support model training. Additionally, it includes 850 decoy compounds. The purpose of this dataset is to identify active compounds from the decoy compound pool via the template compounds, creating a few-shot learning task for drug discovery. It is classified as a small-scale dataset, with the task belonging to few-shot learning in the drug discovery domain.