ChatBridge: Bridging Modalities with Large Language Model as a Language Catalyst
Zhao, Zijia, Guo, Longteng, Yue, Tongtian, Chen, Sihan, Shao, Shuai, Zhu, Xinxin, Yuan, Zehuan, Liu, Jing
–arXiv.org Artificial Intelligence
Building general-purpose models that can perceive diverse real-world modalities and solve various tasks is an appealing target in artificial intelligence. In this paper, we present ChatBridge, a novel multimodal language model that leverages the expressive capabilities of language as the catalyst to bridge the gap between various modalities. We show that only language-paired two-modality data is sufficient to connect all modalities. ChatBridge leverages recent large language models (LLM) and extends their zero-shot capabilities to incorporate diverse multimodal inputs. ChatBridge undergoes a two-stage training. The first stage aligns each modality with language, which brings emergent multimodal correlation and collaboration abilities. The second stage instruction-finetunes ChatBridge to align it with user intent with our newly proposed multimodal instruction tuning dataset, named MULTIS, which covers a wide range of 16 multimodal tasks of text, image, video, and audio modalities. We show strong quantitative and qualitative results on zero-shot multimodal tasks covering text, image, video, and audio modalities. All codes, data, and models of ChatBridge will be open-sourced.
arXiv.org Artificial Intelligence
May-25-2023
- Country:
- Asia (0.28)
- Genre:
- Personal > Interview (0.46)
- Research Report (0.82)
- Industry:
- Health & Medicine > Consumer Health (0.67)
- Leisure & Entertainment
- Materials > Chemicals
- Specialty Chemicals (0.61)
- Media > Film (1.00)
- Transportation
- Ground (0.93)
- Infrastructure & Services (0.93)
- Technology: