Enregistré dans:
Détails bibliographiques
Auteurs principaux: Fu, Fengyi, Huang, Mengqi, Wu, Shaojin, Jiang, Yunsheng, Huo, Yufei, Li, Hao, Song, Yinghang, Ding, Fei, Guo, Jianzhu, He, Qian, Fu, Zheren, Mao, Zhendong, Zhang, Yongdong
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2605.18678
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866910238629691392
author Fu, Fengyi
Huang, Mengqi
Wu, Shaojin
Jiang, Yunsheng
Huo, Yufei
Li, Hao
Song, Yinghang
Ding, Fei
Guo, Jianzhu
He, Qian
Fu, Zheren
Mao, Zhendong
Zhang, Yongdong
author_facet Fu, Fengyi
Huang, Mengqi
Wu, Shaojin
Jiang, Yunsheng
Huo, Yufei
Li, Hao
Song, Yinghang
Ding, Fei
Guo, Jianzhu
He, Qian
Fu, Zheren
Mao, Zhendong
Zhang, Yongdong
contents We present Lance, a lightweight native unified model supporting multimodal understanding, generation, and editing for both images and videos. Rather than relying on model capacity scaling or text-image-dominant designs, Lance explores a practical paradigm for unified multimodal modeling via collaborative multi-task training. It is grounded in two core principles: unified context modeling and decoupled capability pathways. Specifically, Lance is trained from scratch and employs a dual-stream mixture-of-experts architecture on shared interleaved multimodal sequences, enabling joint context learning while decoupling the pathways for understanding and generation. We further introduce modality-aware rotary positional encoding to mitigate interference among heterogeneous visual tokens and boost cross-task alignment. During training, Lance adopts a staged multi-task training paradigm with capability-oriented objectives and adaptive data scheduling to strengthen both semantic comprehension and visual generation performance. Experimental results demonstrate that Lance substantially outperforms existing open-source unified models in image and video generation, while retaining strong multimodal understanding capabilities. The homepage is available at https://lance-project.github.io.
format Preprint
id arxiv_https___arxiv_org_abs_2605_18678
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Lance: Unified Multimodal Modeling by Multi-Task Synergy
Fu, Fengyi
Huang, Mengqi
Wu, Shaojin
Jiang, Yunsheng
Huo, Yufei
Li, Hao
Song, Yinghang
Ding, Fei
Guo, Jianzhu
He, Qian
Fu, Zheren
Mao, Zhendong
Zhang, Yongdong
Computer Vision and Pattern Recognition
Artificial Intelligence
We present Lance, a lightweight native unified model supporting multimodal understanding, generation, and editing for both images and videos. Rather than relying on model capacity scaling or text-image-dominant designs, Lance explores a practical paradigm for unified multimodal modeling via collaborative multi-task training. It is grounded in two core principles: unified context modeling and decoupled capability pathways. Specifically, Lance is trained from scratch and employs a dual-stream mixture-of-experts architecture on shared interleaved multimodal sequences, enabling joint context learning while decoupling the pathways for understanding and generation. We further introduce modality-aware rotary positional encoding to mitigate interference among heterogeneous visual tokens and boost cross-task alignment. During training, Lance adopts a staged multi-task training paradigm with capability-oriented objectives and adaptive data scheduling to strengthen both semantic comprehension and visual generation performance. Experimental results demonstrate that Lance substantially outperforms existing open-source unified models in image and video generation, while retaining strong multimodal understanding capabilities. The homepage is available at https://lance-project.github.io.
title Lance: Unified Multimodal Modeling by Multi-Task Synergy
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2605.18678