More images in the dataset, continued training from v2.0.
A fine tune on almost 500 hand picked mega and giga giantess images.