After reading yor code,I hava a question about the input data. Does the input data of the SADE network have to be labeled data?If so, isn't this still supervised learning?
No, you do not need labelled data for any training code. Wherever there is a placeholder for labels, feel free to fill it up with random labels. Labels are indeed only used for evaluation.
After reading yor code,I hava a question about the input data. Does the input data of the SADE network have to be labeled data?If so, isn't this still supervised learning?