Bernini model generates consistent video+audio from reference images

A Reddit user showcases Bernini, a model that creates consistent video and audio from character reference images and scene backgrounds. The output maintains character consistency across frames.