HARIVO: Harnessing Text-to-Image Models for Video Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kwon, Mingi, Oh, Seoung Wug, Zhou, Yang, Liu, Difan, Lee, Joon-Young, Cai, Haoran, Liu, Baqiao, Liu, Feng, Uh, Youngjung
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910643861323776
author Kwon, Mingi
Oh, Seoung Wug
Zhou, Yang
Liu, Difan
Lee, Joon-Young
Cai, Haoran
Liu, Baqiao
Liu, Feng
Uh, Youngjung
author_facet Kwon, Mingi
Oh, Seoung Wug
Zhou, Yang
Liu, Difan
Lee, Joon-Young
Cai, Haoran
Liu, Baqiao
Liu, Feng
Uh, Youngjung
contents We present a method to create diffusion-based video models from pretrained Text-to-Image (T2I) models. Recently, AnimateDiff proposed freezing the T2I model while only training temporal layers. We advance this method by proposing a unique architecture, incorporating a mapping network and frame-wise tokens, tailored for video generation while maintaining the diversity and creativity of the original T2I model. Key innovations include novel loss functions for temporal smoothness and a mitigating gradient sampling technique, ensuring realistic and temporally consistent video generation despite limited public video data. We have successfully integrated video-specific inductive biases into the architecture and loss functions. Our method, built on the frozen StableDiffusion model, simplifies training processes and allows for seamless integration with off-the-shelf models like ControlNet and DreamBooth. project page: https://kwonminki.github.io/HARIVO
format Preprint
id arxiv_https___arxiv_org_abs_2410_07763
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle HARIVO: Harnessing Text-to-Image Models for Video Generation
Kwon, Mingi
Oh, Seoung Wug
Zhou, Yang
Liu, Difan
Lee, Joon-Young
Cai, Haoran
Liu, Baqiao
Liu, Feng
Uh, Youngjung
Computer Vision and Pattern Recognition
Artificial Intelligence
We present a method to create diffusion-based video models from pretrained Text-to-Image (T2I) models. Recently, AnimateDiff proposed freezing the T2I model while only training temporal layers. We advance this method by proposing a unique architecture, incorporating a mapping network and frame-wise tokens, tailored for video generation while maintaining the diversity and creativity of the original T2I model. Key innovations include novel loss functions for temporal smoothness and a mitigating gradient sampling technique, ensuring realistic and temporally consistent video generation despite limited public video data. We have successfully integrated video-specific inductive biases into the architecture and loss functions. Our method, built on the frozen StableDiffusion model, simplifies training processes and allows for seamless integration with off-the-shelf models like ControlNet and DreamBooth. project page: https://kwonminki.github.io/HARIVO
title HARIVO: Harnessing Text-to-Image Models for Video Generation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2410.07763