Loading…
OGCNN is a repository that implements an Orbital Graph Convolutional Neural Network for predicting crystalline material properties. It builds on Crystal Graph Convolutional Neural Networks and orbital-field ideas to improve predictive accuracy. The repo provides core code (main.py, model.py, predict.py), datasets, and documentation with prerequisites (PyTorch, scikit-learn, pymatgen, ASE) and references to key papers.
AI named OGCNN from December 2025 to February 2026.