Supplementary Data: Convolutional neural networks enable high-fidelity prediction of rotation-invariant properties of amorphous and crystalline materials

The training datasets, the code of computing SDMs, the scripts of training and applying CNN models as well as the well-trained CNN models are provided.If the data, code and/or CNN models are useful for your research, please cite: Z. Fan, Convolutional neural networks for learning rotation-invariant properties of both amorphous and crystalline materials. Phys. Rev. Materials 10, 053803 (2026). DOI: https://doi.org/10.1103/x3fw-pjcfPlease contact Zhao (zfan2016@gmail.com) if there is any question.