-
dataWithGCNSetting: Cora， Citeseer and PubMed. The dataset is split as training, validate, and test following the prior work (Kipf GCN).
-
dtaGenerator.py: to read dataset
-
Model.py: GCN model
-
centeralizedTrain.py: train GCN in centeralized scenario.
-
utils.py: includes some functions, such as the function to process adjacency matrix.
-
clients: clients end
-
sever.py: the server end to train GCN in Federated scenario.
-
experimentsOfFed.py & experimentsOfLEpochs: experiments codes