This repository is the official PyTorch implementation of Neighborhood Enlargement in Graph Network which is submitted. It's based on the implementation of How Powerful are Graph Neural Networks?Code
Default parameters are not the best performing-hyper-parameters. Hyper-parameters need to be specified through the commandline arguments. Please refer to submitted paper for the details of how we set the hyper-parameters. The proposed method is set as --neighbor_pooling_type = sum1
Type
python main.py --help
to learn hyper-parameters to be specified.
Results of more benchmarks will be released soon.