This is a Pytorch implementation of ASTGNN. Now the corresponding paper is available online at https://ieeexplore.ieee.org/document/9346058.


We take the commands on PEMS04 for example.

Step 1: Process dataset:

• for ASTGNN

python prepareData.py --config configurations/PEMS04.conf

• for ASTGNN(p)

python prepareData.py --config configurations/PEMS04_rdw.conf

Step 2: train and test the model:

• for ASTGNN

nohup python -u train_ASTGNN.py --config configurations/PEMS04.conf --cuda=1 > pems04.out &

• for ASTGNN(p)

nohup python -u train_ASTGNN.py --config configurations/PEMS04_rdw.conf --cuda=1 > pems04_rdw.out &

The settings for each experiments are given in the "configurations" folder.

If you find the repository is useful, please cite our paper. Thank you.