Saved in:
Bibliographic Details
Main Authors: Liu, Kai, Li, Jungang, Sun, Yuchong, Wu, Shengqiong, Gao, Jianzhang, Zhang, Daoan, Zhang, Wei, Jin, Sheng, Yu, Sicheng, Zhan, Geng, Ji, Jiayi, Zhou, Fan, Zheng, Liang, Yan, Shuicheng, Fei, Hao, Chua, Tat-Seng
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.22905
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911349755346944
author Liu, Kai
Li, Jungang
Sun, Yuchong
Wu, Shengqiong
Gao, Jianzhang
Zhang, Daoan
Zhang, Wei
Jin, Sheng
Yu, Sicheng
Zhan, Geng
Ji, Jiayi
Zhou, Fan
Zheng, Liang
Yan, Shuicheng
Fei, Hao
Chua, Tat-Seng
author_facet Liu, Kai
Li, Jungang
Sun, Yuchong
Wu, Shengqiong
Gao, Jianzhang
Zhang, Daoan
Zhang, Wei
Jin, Sheng
Yu, Sicheng
Zhan, Geng
Ji, Jiayi
Zhou, Fan
Zheng, Liang
Yan, Shuicheng
Fei, Hao
Chua, Tat-Seng
contents This paper presents JavisGPT, the first unified multimodal large language model (MLLM) for joint audio-video (JAV) comprehension and generation. JavisGPT has a concise encoder-LLM-decoder architecture, which has a SyncFusion module for spatio-temporal audio-video fusion and synchrony-aware learnable queries to bridge a pretrained JAV-DiT generator. This design enables temporally coherent video-audio understanding and generation from multimodal instructions. We design an effective three-stage training pipeline consisting of multimodal pretraining, audio-video fine-tuning, and large-scale instruction-tuning, to progressively build multimodal comprehension and generation from existing vision-language models. For instruction tuning, we construct JavisInst-Omni, a high-quality instruction dataset with over 200K GPT-4o-curated audio-video-text dialogues that cover diverse and multi-level comprehension and generation scenarios. On JAV comprehension and generation benchmarks, our experiments show that JavisGPT outperforms existing MLLMs, particularly in complex and temporally synchronized settings.
format Preprint
id arxiv_https___arxiv_org_abs_2512_22905
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle JavisGPT: A Unified Multi-modal LLM for Sounding-Video Comprehension and Generation
Liu, Kai
Li, Jungang
Sun, Yuchong
Wu, Shengqiong
Gao, Jianzhang
Zhang, Daoan
Zhang, Wei
Jin, Sheng
Yu, Sicheng
Zhan, Geng
Ji, Jiayi
Zhou, Fan
Zheng, Liang
Yan, Shuicheng
Fei, Hao
Chua, Tat-Seng
Computer Vision and Pattern Recognition
This paper presents JavisGPT, the first unified multimodal large language model (MLLM) for joint audio-video (JAV) comprehension and generation. JavisGPT has a concise encoder-LLM-decoder architecture, which has a SyncFusion module for spatio-temporal audio-video fusion and synchrony-aware learnable queries to bridge a pretrained JAV-DiT generator. This design enables temporally coherent video-audio understanding and generation from multimodal instructions. We design an effective three-stage training pipeline consisting of multimodal pretraining, audio-video fine-tuning, and large-scale instruction-tuning, to progressively build multimodal comprehension and generation from existing vision-language models. For instruction tuning, we construct JavisInst-Omni, a high-quality instruction dataset with over 200K GPT-4o-curated audio-video-text dialogues that cover diverse and multi-level comprehension and generation scenarios. On JAV comprehension and generation benchmarks, our experiments show that JavisGPT outperforms existing MLLMs, particularly in complex and temporally synchronized settings.
title JavisGPT: A Unified Multi-modal LLM for Sounding-Video Comprehension and Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.22905