Since the model is outputting the predicted adjacency matrix, all graphs in the dataset must have the same number of nodes. However, we can fix this by padding each target adjacency matrix by 0s to the max number of nodes in the dataset and set the output dimension of the model to max_num_nodes * max_num_nodes.
Since the model is outputting the predicted adjacency matrix, all graphs in the dataset must have the same number of nodes. However, we can fix this by padding each target adjacency matrix by 0s to the max number of nodes in the dataset and set the output dimension of the model to max_num_nodes * max_num_nodes.