EMOVA: Empowering Language Models to See, Hear and Speak with Vivid Emotions
This work addresses a gap for the open-source community in developing omni-modal models that integrate vision, language, and speech without external tools, though it builds incrementally on existing multimodal approaches.
The paper tackles the challenge of enabling Large Language Models to process images, text, and speech end-to-end with publicly available data, achieving state-of-the-art performance on both vision-language and speech benchmarks while supporting omni-modal spoken dialogue with vivid emotions.
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.