Supplementary data for "Generalized Machine Learning Potentials for Water Adsorption in Al-Based Metal–Organic Frameworks Considering Local Flexibility"

This zenodo includes three parts, i.e., three compressed folders below. Dataset: 1. original 420 Al-MOFs from public datasets2. optimized 420 Al-MOFs by the MACE-AlMOFs-v1 model3. DFT-labeled dataset for fine-tuning the MACE model, which including 3 xyz files, i.e, the training dataset, the test dataset and the additional dataset for energy offsets. You can use the property in xyz file, i.e., config_type and config_name in the xyz file to identify structures. Scripts: 1. CP2K calculation scripts2. MACE training scripts Fine-tuned MACE models: For both MACE-AlMOFs-68 model and MACE-AlMOFs-v1 model, we provide three models fine-tuned by the same dataset but with different random seeds.