CFG input for GNN

vision
1

I’m going to use pytorch for training GNN with CFGs as input
The question is: what is the standard form ( adjacency matrix ?? ) that i have to respect when i generete the CFGs