SENSETIME-W Unveils 'Vimi', First 'Controllable' Character Video Model

$SENSETIME-W(00020.HK)$ released Vimi, a model for generating one-minute, high-quality character videos without distortion. Vimi targets C-end users for activities like chatting, singing, and dancing. Based on the SenseNova foundation model, it creates human-like videos from any style photo.

SENSETIME-W is showcasing Vimi at WAIC 2024 in Shanghai.