Graph Neural Networks (GNN) Learning

This repository contains implementations of various Graph Neural Networks (GNNs) for node classification tasks.

Table of Contents

• Setup
• Usage
• Models
• Contributing

Setup

The project is implemented in Python and uses PyTorch and PyTorch Geometric. You can install the dependencies with pip:

1. clone the repo into your local machine

cd path/to/your/workspace
git clone https://github.com/Bingzw/DL_genAI_practicing.git

2. create python virtual environment

python3 -m venv venv
source venv/bin/activate

3. install the required packages

pip install -r requirements.txt

4. run model training

cd path/to/model/folder
python <train.py>

Models & Tasks

The repository contains models solving the following tasks:

• Node classification: predicting the label of a node in a graph
  • GCN Model
  • GAT Model
  • GraphSAGE
• Graph classification: predicting the label of an entire graph
  • GNN based on GraphConv (this can also be changed to GCN, GAT or GraphSAGE)
• Link prediction: predicting whether an edge exists between two nodes
  • Heterogeneous GraphSAGE

References

• PyTorch Geometric
• standford course: CS224W: Machine Learning with Graphs