Tiny Video Networks

TL;DR

This paper introduces Tiny Video Networks, a novel approach to automatically design highly efficient video understanding models that achieve competitive performance with significantly reduced computational time on CPUs and GPUs.

Contribution

The paper presents a new method for automatically designing tiny, efficient video models that outperform existing solutions in speed while maintaining accuracy.

Findings

Models run in 37 ms per video on CPU

Models run in 10 ms per video on GPU

Achieve competitive accuracy with high efficiency

Abstract

Video understanding is a challenging problem with great impact on the abilities of autonomous agents working in the real-world. Yet, solutions so far have been computationally intensive, with the fastest algorithms running for more than half a second per video snippet on powerful GPUs. We propose a novel idea on video architecture learning - Tiny Video Networks - which automatically designs highly efficient models for video understanding. The tiny video models run with competitive performance for as low as 37 milliseconds per video on a CPU and 10 milliseconds on a standard GPU.