Grid: Omni Visual Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wan, Cong, Luo, Xiangyang, Luo, Hao, Cai, Zijian, Song, Yiren, Zhao, Yunlong, Bai, Yifan, Wang, Fan, He, Yuhang, Gong, Yihong
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908427915100160
author Wan, Cong
Luo, Xiangyang
Luo, Hao
Cai, Zijian
Song, Yiren
Zhao, Yunlong
Bai, Yifan
Wang, Fan
He, Yuhang
Gong, Yihong
author_facet Wan, Cong
Luo, Xiangyang
Luo, Hao
Cai, Zijian
Song, Yiren
Zhao, Yunlong
Bai, Yifan
Wang, Fan
He, Yuhang
Gong, Yihong
contents Visual generation has witnessed remarkable progress in single-image tasks, yet extending these capabilities to temporal sequences remains challenging. Current approaches either build specialized video models from scratch with enormous computational costs or add separate motion modules to image generators, both requiring learning temporal dynamics anew. We observe that modern image generation models possess underutilized potential in handling structured layouts with implicit temporal understanding. Building on this insight, we introduce GRID, which reformulates temporal sequences as grid layouts, enabling holistic processing of visual sequences while leveraging existing model capabilities. Through a parallel flow-matching training strategy with coarse-to-fine scheduling, our approach achieves up to 67 faster inference speeds while using <1/1000 of the computational resources compared to specialized models. Extensive experiments demonstrate that GRID not only excels in temporal tasks from Text-to-Video to 3D Editing but also preserves strong performance in image generation, establishing itself as an efficient and versatile omni-solution for visual generation.
format Preprint
id arxiv_https___arxiv_org_abs_2412_10718
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Grid: Omni Visual Generation
Wan, Cong
Luo, Xiangyang
Luo, Hao
Cai, Zijian
Song, Yiren
Zhao, Yunlong
Bai, Yifan
Wang, Fan
He, Yuhang
Gong, Yihong
Computer Vision and Pattern Recognition
Visual generation has witnessed remarkable progress in single-image tasks, yet extending these capabilities to temporal sequences remains challenging. Current approaches either build specialized video models from scratch with enormous computational costs or add separate motion modules to image generators, both requiring learning temporal dynamics anew. We observe that modern image generation models possess underutilized potential in handling structured layouts with implicit temporal understanding. Building on this insight, we introduce GRID, which reformulates temporal sequences as grid layouts, enabling holistic processing of visual sequences while leveraging existing model capabilities. Through a parallel flow-matching training strategy with coarse-to-fine scheduling, our approach achieves up to 67 faster inference speeds while using <1/1000 of the computational resources compared to specialized models. Extensive experiments demonstrate that GRID not only excels in temporal tasks from Text-to-Video to 3D Editing but also preserves strong performance in image generation, establishing itself as an efficient and versatile omni-solution for visual generation.
title Grid: Omni Visual Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.10718