StreamChat: Chatting with Streaming Video
This addresses the issue of inefficient real-time interaction with streaming video for users of multimodal AI systems, representing an incremental improvement over existing methods.
The paper tackles the problem of Large Multimodal Models (LMMs) experiencing delays when interacting with streaming video due to reliance on outdated visual context, and it introduces StreamChat, which updates visual context at each decoding step to achieve superior capabilities in streaming interaction scenarios compared to state-of-the-art video LMMs.
This paper presents StreamChat, a novel approach that enhances the interaction capabilities of Large Multimodal Models (LMMs) with streaming video content. In streaming interaction scenarios, existing methods rely solely on visual information available at the moment a question is posed, resulting in significant delays as the model remains unaware of subsequent changes in the streaming video. StreamChat addresses this limitation by innovatively updating the visual context at each decoding step, ensuring that the model utilizes up-to-date video content throughout the decoding process. Additionally, we introduce a flexible and efficient crossattention-based architecture to process dynamic streaming inputs while maintaining inference efficiency for streaming interactions. Furthermore, we construct a new dense instruction dataset to facilitate the training of streaming interaction models, complemented by a parallel 3D-RoPE mechanism that encodes the relative temporal information of visual and text tokens. Experimental results demonstrate that StreamChat achieves competitive performance on established image and video benchmarks and exhibits superior capabilities in streaming interaction scenarios compared to state-of-the-art video LMM.