The code has been tested under Pytorch 2.0. Datasets names are 'Digg', 'Memetracker', 'Android', 'Christianity', 'Twitter'.