Saved in:
Bibliographic Details
Main Authors: Wang, Xinlong, Zhang, Xiaosong, Luo, Zhengxiong, Sun, Quan, Cui, Yufeng, Wang, Jinsheng, Zhang, Fan, Wang, Yueze, Li, Zhen, Yu, Qiying, Zhao, Yingli, Ao, Yulong, Min, Xuebin, Li, Tao, Wu, Boya, Zhao, Bo, Zhang, Bowen, Wang, Liangdong, Liu, Guang, He, Zheqi, Yang, Xi, Liu, Jingjing, Lin, Yonghua, Huang, Tiejun, Wang, Zhongyuan
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2409.18869
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914959004270592
author Wang, Xinlong
Zhang, Xiaosong
Luo, Zhengxiong
Sun, Quan
Cui, Yufeng
Wang, Jinsheng
Zhang, Fan
Wang, Yueze
Li, Zhen
Yu, Qiying
Zhao, Yingli
Ao, Yulong
Min, Xuebin
Li, Tao
Wu, Boya
Zhao, Bo
Zhang, Bowen
Wang, Liangdong
Liu, Guang
He, Zheqi
Yang, Xi
Liu, Jingjing
Lin, Yonghua
Huang, Tiejun
Wang, Zhongyuan
author_facet Wang, Xinlong
Zhang, Xiaosong
Luo, Zhengxiong
Sun, Quan
Cui, Yufeng
Wang, Jinsheng
Zhang, Fan
Wang, Yueze
Li, Zhen
Yu, Qiying
Zhao, Yingli
Ao, Yulong
Min, Xuebin
Li, Tao
Wu, Boya
Zhao, Bo
Zhang, Bowen
Wang, Liangdong
Liu, Guang
He, Zheqi
Yang, Xi
Liu, Jingjing
Lin, Yonghua
Huang, Tiejun
Wang, Zhongyuan
contents While next-token prediction is considered a promising path towards artificial general intelligence, it has struggled to excel in multimodal tasks, which are still dominated by diffusion models (e.g., Stable Diffusion) and compositional approaches (e.g., CLIP combined with LLMs). In this paper, we introduce Emu3, a new suite of state-of-the-art multimodal models trained solely with next-token prediction. By tokenizing images, text, and videos into a discrete space, we train a single transformer from scratch on a mixture of multimodal sequences. Emu3 outperforms several well-established task-specific models in both generation and perception tasks, surpassing flagship models such as SDXL and LLaVA-1.6, while eliminating the need for diffusion or compositional architectures. Emu3 is also capable of generating high-fidelity video via predicting the next token in a video sequence. We simplify complex multimodal model designs by converging on a singular focus: tokens, unlocking great potential for scaling both during training and inference. Our results demonstrate that next-token prediction is a promising path towards building general multimodal intelligence beyond language. We open-source key techniques and models to support further research in this direction.
format Preprint
id arxiv_https___arxiv_org_abs_2409_18869
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Emu3: Next-Token Prediction is All You Need
Wang, Xinlong
Zhang, Xiaosong
Luo, Zhengxiong
Sun, Quan
Cui, Yufeng
Wang, Jinsheng
Zhang, Fan
Wang, Yueze
Li, Zhen
Yu, Qiying
Zhao, Yingli
Ao, Yulong
Min, Xuebin
Li, Tao
Wu, Boya
Zhao, Bo
Zhang, Bowen
Wang, Liangdong
Liu, Guang
He, Zheqi
Yang, Xi
Liu, Jingjing
Lin, Yonghua
Huang, Tiejun
Wang, Zhongyuan
Computer Vision and Pattern Recognition
While next-token prediction is considered a promising path towards artificial general intelligence, it has struggled to excel in multimodal tasks, which are still dominated by diffusion models (e.g., Stable Diffusion) and compositional approaches (e.g., CLIP combined with LLMs). In this paper, we introduce Emu3, a new suite of state-of-the-art multimodal models trained solely with next-token prediction. By tokenizing images, text, and videos into a discrete space, we train a single transformer from scratch on a mixture of multimodal sequences. Emu3 outperforms several well-established task-specific models in both generation and perception tasks, surpassing flagship models such as SDXL and LLaVA-1.6, while eliminating the need for diffusion or compositional architectures. Emu3 is also capable of generating high-fidelity video via predicting the next token in a video sequence. We simplify complex multimodal model designs by converging on a singular focus: tokens, unlocking great potential for scaling both during training and inference. Our results demonstrate that next-token prediction is a promising path towards building general multimodal intelligence beyond language. We open-source key techniques and models to support further research in this direction.
title Emu3: Next-Token Prediction is All You Need
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2409.18869