ScaDyG: A New Paradigm for Large-scale Dynamic Graph Learning

This repository is the PyTorch implementation of ScaDyG.

Requirements

pip install -r requirements.txt

How to run

You can run the ScaDyG with the following commands:

python scalable_tgn_main_link_prediction.py --dataset mooc

We provide the mooc dataset as an example.

The raw data set can be downloaded from https://tgb.complexdatalab.com/ for node affinity prediction, and https://zenodo.org/records/7008205#.Yv_a_3bMJPZ for dynamic link prediction.