GraPHmax: Datasets: There is a folder graphdatasets that contains all the required datasets. How to run: Use the following command to run GraPHmax with default values of the parameters on MUTAG dataset. python main.py Requirements: Install python in your system. Requirements file has been given in the model folder. To install the same libraries run pip install -r requirements.txt These are python libraries that are required to run the code.
- Tensorflow (version: 1.14.0) with python 3.6.9.
- Networkx (version: 2.3)
- matplotlib
- pandas
- keras
- scipy
- pickle
- collections Hyperparameters Settings: There are some hyperparameters one can change during execution.
- plrt : Pooling ratio
- plly : Number of Pooling layers
- emb : Embedding dimenion
- lrp : Learning rate for periphery representation
- lrh : Learning rate for hierarchical representation
- lrl : Learning rate for classification task
- dropout : Dropout rate
- negpr : Number of negative samples for periphery discriminator
- neghr: Number of negative samples hierarchical discriminator
- dataset: Name of the dataset to perform classification on.
One can specify these hyperparameters during running the code. python main.py --dataset MUTAG