https://github.com/senya-ashukha/simple-equivariant-gnn

A short and easy PyTorch implementation of E(n) Equivariant Graph Neural Networks
https://github.com/senya-ashukha/simple-equivariant-gnn

deep-learning deep-neural-networks egnn equivariance gnn graph-neural-networks homo molecules neural-networks pytorch qm9

Last synced: about 23 hours ago
JSON representation

A short and easy PyTorch implementation of E(n) Equivariant Graph Neural Networks

• Host: GitHub
• URL: https://github.com/senya-ashukha/simple-equivariant-gnn
• Owner: senya-ashukha
• License: apache-2.0
• Created: 2022-01-13T10:38:16.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2022-01-14T12:25:53.000Z (almost 3 years ago)
• Last Synced: 2023-11-07T18:51:58.617Z (about 1 year ago)
• Topics: deep-learning, deep-neural-networks, egnn, equivariance, gnn, graph-neural-networks, homo, molecules, neural-networks, pytorch, qm9
• Language: Python
• Homepage:
• Size: 193 KB
• Stars: 112
• Watchers: 3
• Forks: 11
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Simple implementation of Equivariant GNN

[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/senya-ashukha/simple-equivariant-gnn/blob/main/simple-egnn.ipynb)

- A short implementation of [E(n) Equivariant Graph Neural Networks](https://arxiv.org/pdf/2102.09844.pdf) for [HOMO energy](https://en.wikipedia.org/wiki/HOMO_and_LUMO) prediction.

- Just 50 lines of code; 

- The implementation is based on pure PyTorch & Numpy, it has no external packages (like PyTorch-geometric).

- Closely matches the Mean Absolute Error reported in the paper.