SEA-GWNN: Simple and Effective Generalized Adaptive Graph Wavelet Neural Network

This is a pytorch implementation of the paper titled: "SEA-GWNN: Simple and Effective Adaptive Graph Wavelet Neural Network".

Environment Settings

• pytorch 1.12.0
• numpy 1.23.4
• torch-geometric 1.7.2
• tqdm 4.64.1
• scipy 1.9.3
• seaborn 0.12.0
• scikit-learn 1.1.3
• ogb 1.3.5

All the requirements are given inside the "requirements.txt" file.

Datasets

We provide the Cora and Chameleon datasets in the folder 'data/' and 'new_data/' direactory. Additionally, you can download all the datasets utilized in this paper form the Pei et. al. 2018, "Geom-GCN: Geometric Graph Convolutional Networks".

Code Structure

The folder "homophilic graphs" is the code for the for standard citation networks (Cora, Citeseer, PubMed); and the folder "heterophilic graphs" is the code for the results in all the heterophilic datasets. Finally "large graphs" contain code for the large scale graphs.

Environment Setup

You will require pytroch gpu version 1.12.0 from pytorch (https://pytorch.org/get-started/previous-versions/).

conda install pytorch==1.12.0 torchvision==0.13.0 torchaudio==0.12.0 cudatoolkit=11.3 -c pytorch

Then run the requirements.txt file.

pip install -r requirements.txt