VETIM: Expanding the Vocabulary of Text-to-Image Models only with Text This page is under construction. Project website