GPA

This is a pytorch and pytorch-geometric based implementation of Graph Contrastive Learning with Personalized Augmentation.

Installation

The required packages can be installed by running pip install -r requirements.txt.

The datasets used in our paper can be automatically downlowad.

Train on the TUDatasets (PROTEINS, NCI1, DD, COLLAB, MUTAG, IMDB-BINARY, REDDIT-BINARY, REDDIT-MULTI-5K, github_stargazers):

python main_gpa.py --dataset "MUTAG" --mode unsuper
python main_gpa.py --dataset "MUTAG" --mode semi

Name		Name	Last commit message	Last commit date
Latest commit History 1 Commit
__MACOSX		__MACOSX
__pycache__		__pycache__
dataset		dataset
semi_dataset		semi_dataset
unsuper_dataset/MUTAG		unsuper_dataset/MUTAG
weights		weights
README.md		README.md
__init__.py		__init__.py
eval_graph.py		eval_graph.py
main_gpa.py		main_gpa.py
model.py		model.py
requirements.txt		requirements.txt