The code has been tested on an environment with Python 3.9.7, PyTorch 1.9.1, Numpy 1.21.2, and DGL 0.7.1.
The datasets are taken from the previous work Self-Supervised Hypergraph Convolutional Networks for Session-based Recommendation (AAAI'21). You can download the datasets in this GitHub repo and extract them to the datasets folder.
To train FOCOL, run the following command:
python run.py expts/focol/focol.py --dataset-dir datasets/diginetica
which trains FOCOL on diginetica with default hyperparameters.
You can see the detailed usage with the following command:
python run.py expts/focol/focol.py -h