This repo contains the implementation for the paper "Lorentzian Residual Neural Networks" to be presented at KDD 2025.
- For our implementation on homophilous graphs for LResNet in GNN, see the homophilous folder.
- For the implementation of LResNet in GNN on heterophilic graphs and LResNet in graph transformer, see the hetrophilic folder.
- For LResNet in hyperbolic ResNet for image classification, see the Classification folder in Images.
- For OOD robutness of LResNet, see the OOD-Detection folder in Images.
- In each of the above foler, we provide an example foler that include examples of running our code with sensible hyperparameters. Please see each folder for more details.