Currently not supported. See: https://discuss.pytorch.org/t/how-do-use-logical-not-on-bytetensor/2074