Year: 2025
Type: preprint
Abstract: The Transformer architecture has achieved significant success in natural language processing, motivating its adaptation to computer vision tasks. Unlike convolutional neural networks, vision transformers inherently capture long-range dependencies and enable parallel processing, yet lack inductive biases and efficiency benefits, facing significant computational and memory challenges that limit its real-world applicability. This paper surveys various online strategies for generating lightweight vision transformers for image recognition, focusing on three key areas: Efficient Component Design, Dynamic Network, and Knowledge Distillation. We evaluate the relevant exploration for each topic on the ImageNet-1K benchmark, analyzing trade-offs among precision, parameters, throughput, and more to highlight their respective advantages, disadvantages, and flexibility. Finally, we propose future research directions and potential challenges in the lightweighting of vision transformers with the aim of inspiring further exploration and providing practical guidance for the community. Project Page: https://github.com/ajxklo/Lightweight-VIT
Source: arXiv (Cornell University)
Authors: Zhang, Zherui, Xu, Rongtao, Zhou, Jie, Wang, Changwei, Pei, Xingtian, Xu, Wenhao, Zhang, Jiguang, Guo, Li, Gao, Longxiang, Xu, Wenbo, Xu, Shibiao
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6948079174