Data for Predicting the Thermal Behavior of Semiconductor Defects with Equivariant Neural Networks

figures.zip : Contains all data and plotting scripts used to generate the figures in the paper models.zip: Includes the trained machine-learning models used in the active-learning workflow data_train_mace.zip: Dataset used for training the MACE machine-learning force field. Contains structures in .extxyz format with force and energy label. data_train_deeph_round0.zip: Dataset used for training the initial DeepH Hamiltonian model. Contains the processed DFT Hamiltonians, graph representations, and structural data used in DeepH-E3. data_train_deeph_active.zip: Dataset used for training the active learning rounds. Contains the processed DFT Hamiltonians, graph representations, and structural data.