Researchers from UC Berkeley and Stanford tapped Test Time Training to generate one-minute videos

They added TTT layers to a pre-trained Transformer and fine-tuned it to generate cartoons with strong temporal consistency

Here's a single-shot example:pic.twitter.com/YCqNTHSvpu

— Rowan Cheung (@rowancheung) 8 avril 2025

Researchers from UC Berkeley and Stanford tapped Test Time Training to generate one-minute videos They added TTT layers to a pre-trained Transformer and fine-tuned it to generate cartoons with strong temporal consistency Here's a single-shot example:

→ View original post on X — @rowancheung