/TARGCN

Primary LanguagePython

This is the code for the paper A Simple But Powerful Graph Encoder for Temporal Knowledge Graph Completion accepted to NeurIPS 2022 Temporal Graph Learning Workshop paper

Preprocessing

If you want you run the demo on ICEWS14, ICEWS05-15 or GDELT, go to 'Software/dataset/${DATASET}', and run: python preprocess.py Then you can train on corresponding datasets.

Training

To run the training demo, please run: python main.py --dataset ${DATASET}

Testing

To test a trained model, please run: python main.py --dataset ${DATASET} --test --resume --name ${CHECKPOINT_NAME}

Generalization to unseen timestamps

Please go to 'Software/dataset/icews14_unseen' and run: python preprocess_extrapolate.py Then go to the root directory and run: python main.py --dataset icews14_unseen

Generalization to irregular timestamped data

Please run: python main.py --dataset icews14_irr