Dear developers and contributors of DIG, I have a simple multi-label classification task based on GNN. In the last layer of the model, I select the value of the component of the output feature greater than a certain threshold as the classification result. I want to use DIG's subgraphx to explain the classification results, but there is only num_classes= in the parameter list. Does this mean that DIG can only handle multi-classification tasks?
Dear developers and contributors of DIG, I have a simple multi-label classification task based on GNN. In the last layer of the model, I select the value of the component of the output feature greater than a certain threshold as the classification result. I want to use DIG's subgraphx to explain the classification results, but there is only
num_classes=in the parameter list. Does this mean that DIG can only handle multi-classification tasks?