b_main.py: The main program entry of the algorithm IHGNN b_dataset.py: Dataset related methods and classes b_evaluate.py: Evaluation related methods and classes data_generator.py: Data preprocess methods and classes HetG.py: Heterogeneous graph related methods and classes IHGNN: the algorithm floder utils/tools.py: Uitls methods and classes floder feature: Feature preprocess methods and classes
We implement all experiments on 1 Nvidia GTX 1080Ti GPUs. The experimental environment is set as requirements.txt.
The setting details of training and evaluation are in b_main.py.
For Training and evaluation the algorithm, you can just use "python b_main.py"