This repository is a PyTorch implementation of "Causal Invariant Learning for Out-of-Distribution Generalization in Dynamic Graphs"

Dependencies

• cuda = 11.3
• python >= 3.9
• pytorch >= 1.12.0
• numpy >= 1.24.3
• pyg >= 2.3.0
• scikit-learn >= 1.2.2
• scipy >= 1.16.0
• tqdm >= 4.65.0

Quick Start

• python main.py --dataset 'dataset'
• dataset for link prediction: collab, act, synthetic (0.4, 0.6, 0.8)
• dataset for node classification: Aminer, dymotif_data