Visual Arts & Animation Student and AI enthusiast Esdrico demonstrated an amazing use case for a recently-open-sourced text-to-image diffusion model Stable Diffusion. Using the AI, the author managed to recreate a dance scene from the 2016 musical La La Land and turn it into an anime-style animation. To create the animation, the artist utilized Euler Ancestral sampler and the "a film still of lalaland, artwork by studio ghibli, makoto shinkai, pixv" prompt. The result took around 30 minutes to generate and runs at a framerate of 8 FPS.
