Molecule Generation Datasets
收藏arXiv2025-09-30 收录
下载链接:
https://github.com/zhejiangzhuque/Diffusion-based-Graph-Generative-Methods
下载链接
链接失效反馈官方服务:
资源简介:
该数据集用于评估在各种分子生成任务中的指标,包括全新分子设计、构象设计、配体设计、配体对接以及蛋白质生成等。它包含了用于评估生成分子的一系列指标,如有效性、独特性、新颖性、质量、均方根偏差、类药物线性、合成可及性等。该数据集的任务是分子生成与评估。
This dataset is designed for evaluating metrics across various molecular generation tasks, including de novo molecular design, conformational design, ligand design, ligand docking, and protein generation, among others. It encompasses a suite of metrics for evaluating generated molecules, such as validity, uniqueness, novelty, quality, root-mean-square deviation (RMSD), drug-like linearity, synthetic accessibility, and more. The core tasks of this dataset are molecular generation and evaluation.



