Data for the manuscript "Transfer Learning of Equivariant Interatomic Potentials across Numerical Fidelity Levels in Periodic DFT"

These data correspond to the manuscript “Transfer Learning of Equivariant Interatomic Potentials across Numerical Fidelity Levels in Periodic DFT.” They contain model-training configuration files, molecular dynamics trajectories, and the data used for model training. The model-training files include the config.yaml files, the indices of configurations in the datasets and scripts used to freeze model layers. The molecular dynamics trajectories are stored in XYZ format. The "frame_number" label in each XYZ file is retained as the frame index. The Allegro-TL MLMD trajectories and the DFT convergence tests are also stored in the data. The README file describes the directory structure and the contents of each directory.