I would like to apply GNNExplainer’s explain_node module to a model that predicts one value per graph that mimics a molecule, so how do I generate node masks about all nodes in a graph with referring to the same label?
Let me paraphrase your question to make sure I understand you correctly: you were trying to compute node-level explanations with GNNExplainer on only a subset of the nodes?
Thank you for your response. Perhaps that interpretation is correct.