GNNs-Baseline

A Pytorch based implementation of classical GNNs.
GNN基准测试模型.
Some components of this code are adapted from GPRGNN and H2GCN.

Requirement

This project should be able to run without any modification after following packages installed.

pytorch
torch_geometric
networkx
torch-sparse

H2GCN FAGCN APPNP GCN GAT ChebNet APPNP JKNet GPRGNN

DeepWalk Line NetMF

Download by yourself, dataset_utils.py will help you.

cora, citeseer, pubmed, computers, photo, chameleon, squirrel, film, texas, cornell, wisconsin

go to folder src

python train_model.py --RPMAX 1 \
        --net GCN \
        --train_rate 0.05 \
        --val_rate 0.05 \
        --dataset cora