>>29207
>>29201
Seems to not have a problem generating Adolf as a still image. The thing about the Aurora model behind Imagine makes me want my own autoregressive model. Sequential generations of an image. Diffusion models instead denoise from a field of static into an image. Aurora's outputs are realistic, which makes me want to snatch this fucking thing from their servers. Too many damn Content Moderated blurring denials for animating the greatest man who ever lived, as the post-render pixel pass is a cuck.