Graph neural network based AI model to predict timeseries data based on node information

I have 10 different Graph networks with time series information. Graph structure is same for all the scenarios except node information is different. Based on the specific node (node type) information, the time series node data is changing. For a given node type information, AI model should be able to predict the node properties. The following repo must be modified for this purpose.
https://github.com/sungyongs/dpgn