Dataset/Models for: Active learning accelerated exploration of the single atom local environments in multimetallic systems for oxygen electrocatalysis

This is the datasets and trained models for the work "Active learning accelerated exploration of the single atom local environments in multimetallic systems for oxygen electrocatalysis", by Hoje Chun, Jaclyn R. Lunger, Jeung Ku Kang, Rafael Gómez-Bombarelli, and Byungchan Han. Folder named Models contains the trained models of "m-PaiNN" and "per-site PaiNN". Folder named Dataset contains the torch dataset and Dataset_raw contains the raw Density Functional Theory (DFT) dataset parsed in format of pymatgen Structure. Some structures (868) in the search space are missing due to the lost track of the geometry optimization during the initial dataset curation.