A Unified Pruning Framework for Vision Transformers
This work addresses resource constraints in deploying ViTs for computer vision applications, but it is incremental as it builds on existing pruning methods for a specific model type.
The paper tackles the problem of high computational costs and training data requirements in vision transformers (ViTs) by proposing a unified pruning framework (UP-ViTs) that compresses ViTs while maintaining structural consistency, achieving accuracy improvements such as 3.59% on ImageNet for UP-DeiT-T compared to vanilla DeiT-T.
Recently, vision transformer (ViT) and its variants have achieved promising performances in various computer vision tasks. Yet the high computational costs and training data requirements of ViTs limit their application in resource-constrained settings. Model compression is an effective method to speed up deep learning models, but the research of compressing ViTs has been less explored. Many previous works concentrate on reducing the number of tokens. However, this line of attack breaks down the spatial structure of ViTs and is hard to be generalized into downstream tasks. In this paper, we design a unified framework for structural pruning of both ViTs and its variants, namely UP-ViTs. Our method focuses on pruning all ViTs components while maintaining the consistency of the model structure. Abundant experimental results show that our method can achieve high accuracy on compressed ViTs and variants, e.g., UP-DeiT-T achieves 75.79% accuracy on ImageNet, which outperforms the vanilla DeiT-T by 3.59% with the same computational cost. UP-PVTv2-B0 improves the accuracy of PVTv2-B0 by 4.83% for ImageNet classification. Meanwhile, UP-ViTs maintains the consistency of the token representation and gains consistent improvements on object detection tasks.