Saved in:
Bibliographic Details
Main Authors: Chen, Kai, Gou, Yunhao, Huang, Runhui, Liu, Zhili, Tan, Daxin, Xu, Jing, Wang, Chunwei, Zhu, Yi, Zeng, Yihan, Yang, Kuo, Wang, Dingdong, Xiang, Kun, Li, Haoyuan, Bai, Haoli, Han, Jianhua, Li, Xiaohui, Jin, Weike, Xie, Nian, Zhang, Yu, Kwok, James T., Zhao, Hengshuang, Liang, Xiaodan, Yeung, Dit-Yan, Chen, Xiao, Li, Zhenguo, Zhang, Wei, Liu, Qun, Yao, Jun, Hong, Lanqing, Hou, Lu, Xu, Hang
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2409.18042
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908275730022400
author Chen, Kai
Gou, Yunhao
Huang, Runhui
Liu, Zhili
Tan, Daxin
Xu, Jing
Wang, Chunwei
Zhu, Yi
Zeng, Yihan
Yang, Kuo
Wang, Dingdong
Xiang, Kun
Li, Haoyuan
Bai, Haoli
Han, Jianhua
Li, Xiaohui
Jin, Weike
Xie, Nian
Zhang, Yu
Kwok, James T.
Zhao, Hengshuang
Liang, Xiaodan
Yeung, Dit-Yan
Chen, Xiao
Li, Zhenguo
Zhang, Wei
Liu, Qun
Yao, Jun
Hong, Lanqing
Hou, Lu
Xu, Hang
author_facet Chen, Kai
Gou, Yunhao
Huang, Runhui
Liu, Zhili
Tan, Daxin
Xu, Jing
Wang, Chunwei
Zhu, Yi
Zeng, Yihan
Yang, Kuo
Wang, Dingdong
Xiang, Kun
Li, Haoyuan
Bai, Haoli
Han, Jianhua
Li, Xiaohui
Jin, Weike
Xie, Nian
Zhang, Yu
Kwok, James T.
Zhao, Hengshuang
Liang, Xiaodan
Yeung, Dit-Yan
Chen, Xiao
Li, Zhenguo
Zhang, Wei
Liu, Qun
Yao, Jun
Hong, Lanqing
Hou, Lu
Xu, Hang
contents GPT-4o, an omni-modal model that enables vocal conversations with diverse emotions and tones, marks a milestone for omni-modal foundation models. However, empowering Large Language Models to perceive and generate images, texts, and speeches end-to-end with publicly available data remains challenging for the open-source community. Existing vision-language models rely on external tools for speech processing, while speech-language models still suffer from limited or totally without vision-understanding capabilities. To address this gap, we propose the EMOVA (EMotionally Omni-present Voice Assistant), to enable Large Language Models with end-to-end speech abilities while maintaining the leading vision-language performance. With a semantic-acoustic disentangled speech tokenizer, we surprisingly notice that omni-modal alignment can further enhance vision-language and speech abilities compared with the bi-modal aligned counterparts. Moreover, a lightweight style module is introduced for the flexible speech style controls including emotions and pitches. For the first time, EMOVA achieves state-of-the-art performance on both the vision-language and speech benchmarks, and meanwhile, supporting omni-modal spoken dialogue with vivid emotions.
format Preprint
id arxiv_https___arxiv_org_abs_2409_18042
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle EMOVA: Empowering Language Models to See, Hear and Speak with Vivid Emotions
Chen, Kai
Gou, Yunhao
Huang, Runhui
Liu, Zhili
Tan, Daxin
Xu, Jing
Wang, Chunwei
Zhu, Yi
Zeng, Yihan
Yang, Kuo
Wang, Dingdong
Xiang, Kun
Li, Haoyuan
Bai, Haoli
Han, Jianhua
Li, Xiaohui
Jin, Weike
Xie, Nian
Zhang, Yu
Kwok, James T.
Zhao, Hengshuang
Liang, Xiaodan
Yeung, Dit-Yan
Chen, Xiao
Li, Zhenguo
Zhang, Wei
Liu, Qun
Yao, Jun
Hong, Lanqing
Hou, Lu
Xu, Hang
Computer Vision and Pattern Recognition
Computation and Language
GPT-4o, an omni-modal model that enables vocal conversations with diverse emotions and tones, marks a milestone for omni-modal foundation models. However, empowering Large Language Models to perceive and generate images, texts, and speeches end-to-end with publicly available data remains challenging for the open-source community. Existing vision-language models rely on external tools for speech processing, while speech-language models still suffer from limited or totally without vision-understanding capabilities. To address this gap, we propose the EMOVA (EMotionally Omni-present Voice Assistant), to enable Large Language Models with end-to-end speech abilities while maintaining the leading vision-language performance. With a semantic-acoustic disentangled speech tokenizer, we surprisingly notice that omni-modal alignment can further enhance vision-language and speech abilities compared with the bi-modal aligned counterparts. Moreover, a lightweight style module is introduced for the flexible speech style controls including emotions and pitches. For the first time, EMOVA achieves state-of-the-art performance on both the vision-language and speech benchmarks, and meanwhile, supporting omni-modal spoken dialogue with vivid emotions.
title EMOVA: Empowering Language Models to See, Hear and Speak with Vivid Emotions
topic Computer Vision and Pattern Recognition
Computation and Language
url https://arxiv.org/abs/2409.18042