High-Throughput Prediction of Gas Adsorption Performance of MOFs via Data-Efficient CGCNN Transfer Learning

This dataset contains the H₂ and CH₄ MOF (Metal-Organic Framework) datasets, together with the analysis features and topology enrichment factor results, used in the manuscript: "High-Throughput Prediction of Gas Adsorption Performance of MOFs via Data-Efficient CGCNN Transfer Learning" The repository contains the following compressed archives: h2_dataset.tar.gz — H₂ source domain dataset (65,271 MOF structures with CIF files, id_prop.csv , and atom_init.json ) ch4_dataset.tar.gz — CH₄ target domain dataset (5,896 MOF structures with CIF files, id_prop.csv , and atom_init.json ) analysis_data.tar.gz — Statistical analysis features and topology enrichment factor results