Scaling the Long Video Understanding of Multimodal Large Language Models via Visual Memory Mechanism
This addresses the problem of limited input length in video-MLLMs for researchers and practitioners, though it is an incremental improvement on memory mechanisms.
The paper tackles the challenge of long video understanding in Multimodal Large Language Models (MLLMs) by proposing FlexMem, a training-free visual memory mechanism that mimics human video watching behavior. The result shows that FlexMem processes over 1,000 frames on a single 3090 GPU, achieving improvements over existing methods and comparable or better performance than SOTA models like GPT-4o and Gemini-1.5 Pro on some benchmarks.
Long video understanding is a key challenge that plagues the advancement of \emph{Multimodal Large language Models} (MLLMs). In this paper, we study this problem from the perspective of visual memory mechanism, and proposed a novel and training-free approach, termed \emph{Flexible Memory} (\textbf{FlexMem}). In principle, FlexMem aims to mimic human behavior of video watching, \emph{i.e.}, continually watching video content and recalling the most relevant memory fragments to answer the question. In this way, FlexMem can help MLLMs achieve video understanding of infinite lengths, unlike previous methods that process all video information at once and have input upper-limit. Concretely, FlexMem first consider the visual KV caches as the memory sources, and realize the effective memory transfer and writing via a dual-pathway compression design. Afterwards, FlexMem also explores different memory reading strategies for the diverse video understanding tasks, including the popular streaming one. To validate FlexMem, we apply it to two popular video-MLLMs, and conduct extensive experiments on five long video and one streaming video task. The experimental results show that on \textbf{a single 3090 GPU}, our FlexMem can achieve obvious improvements than existing efficient video understanding methods and process more than \textbf{1k frames}, which also helps the base MLLMs achieve comparable or even better performance than SOTA MLLMs on some benchmarks, \emph{e.g.} , GPT-4o and Gemini-1.5 Pro.