SnapGen-V: Generating a Five-Second Video within Five Seconds on a Mobile Device

Yushu Wu; Zhixing Zhang; Yanyu Li; Yanwu Xu; Anil Kag; Yang Sui; Huseyin Coskun; Ke Ma; Aleksei Lebedev; Ju Hu; Dimitris N. Metaxas; Yanzhi Wang; Sergey Tulyakov; Jian Ren

doi:10.1109/cvpr52734.2025.00237

CVPR2025顶会

SnapGen-V: Generating a Five-Second Video within Five Seconds on a Mobile Device

Yushu Wu, Zhixing Zhang, Yanyu Li, Yanwu Xu, Anil Kag, Yang Sui, Huseyin Coskun, Ke Ma, Aleksei Lebedev, Ju Hu, Dimitris N. Metaxas, Yanzhi Wang

arXiv DOI 出版方

2025年份

9顶会引用

摘要

A cat wearing sunglasses at a pool" "Spectacular continuous fireworks" "Massive tidal wave peaks in a historical hall" "Camera focus on a face of a fashion girl"

Figure 1. Example generation results from our mobile text-to-video model. Our model can generate high-quality and motion consistent 5-second videos on a mobile device (e.g., iPhone 16 Pro Max) within 5 seconds.