D\'ej\`a Vu: Efficient Video-Language Query Engine with Learning-based Inter-Frame Computation Reuse

TL;DR

D'je9 Vu is a system that speeds up video-language models by reusing computations across frames, making large-scale video querying more practical without sacrificing much accuracy.

Contribution

It introduces ReuseViT, a modified Vision Transformer that learns to identify inter-frame reuse opportunities, combined with memory-compute techniques for real performance gains.

Findings

Up to 2.64x acceleration in embedding generation

Maintains within 2% accuracy of baseline models

Effective for large-scale video analytics

Abstract

Recently, Video-Language Models (VideoLMs) have demonstrated remarkable capabilities, offering significant potential for flexible and powerful video query systems. These models typically rely on Vision Transformers (ViTs), which process video frames individually to extract visual embeddings. However, generating embeddings for large-scale videos requires ViT inferencing across numerous frames, posing a major hurdle to real-world deployment and necessitating solutions for integration into scalable video data management systems. This paper introduces D\'ej\`a Vu, a video-language query engine that accelerates ViT-based VideoLMs by reusing computations across consecutive frames. At its core is ReuseViT, a modified ViT model specifically designed for VideoLM tasks, which learns to detect inter-frame reuse opportunities, striking an effective balance between accuracy and reuse. Although ReuseViT significantly reduces computation, these savings do not directly translate into performance gains on GPUs. To overcome this, D\'ej\`a Vu integrates memory-compute joint compaction techniques that convert the FLOP savings into tangible performance gains. Evaluations on three VideoLM tasks show that D\'ej\`a Vu accelerates embedding generation by up to a 2.64x within a 2% error bound, dramatically enhancing the practicality of VideoLMs for large-scale video analytics.