Atompaint: datasets and trained model weights

This archive contains a variety of noteworthy artifacts created during the process of developing the Atompaint method for designing protein/non-protein binding interfaces. These artifacts include:The dataset used to train the models.Model weights for the neighbor location classifier, used to calculate Fréchet distance.Model weights for the diffusion model.Model weights for the amino acid classifier.Model weights for the end-to-end model (i.e. the diffusion model combined with the amino acid classifier).