So I need to randomly append few false samples to my input but I already initiated the correct samples and customized my dataset by overriding the getitem function (and defining the false samples at the init is far too tedious at this point). Is there anyway for me to randomly append few samples after defining my batch without initiating these samples in my Dataset class?