LPViT: Low-Power Semi-structured Pruning for Vision Transformers

TL;DR

This paper introduces LPViT, a block-structured pruning method for vision transformers that reduces resource consumption while maintaining accuracy, achieving significant speedup and power savings through hardware-aware optimization.

Contribution

It proposes a novel block pruning scheme with a hardware-aware learning objective and a lightweight post-training algorithm for efficient ViT model compression.

Findings

Achieves 3.93x speedup on hardware and GPUs for DeiT-B.

Reduces power consumption by 1.4x on GPUs.

Maintains competitive accuracy compared to other pruning methods.

Abstract

Vision transformers have emerged as a promising alternative to convolutional neural networks for various image analysis tasks, offering comparable or superior performance. However, one significant drawback of ViTs is their resource-intensive nature, leading to increased memory footprint, computation complexity, and power consumption. To democratize this high-performance technology and make it more environmentally friendly, it is essential to compress ViT models, reducing their resource requirements while maintaining high performance. In this paper, we introduce a new block-structured pruning to address the resource-intensive issue for ViTs, offering a balanced trade-off between accuracy and hardware acceleration. Unlike unstructured pruning or channel-wise structured pruning, block pruning leverages the block-wise structure of linear layers, resulting in more efficient matrix multiplications. To optimize this pruning scheme, our paper proposes a novel hardware-aware learning objective that simultaneously maximizes speedup and minimizes power consumption during inference, tailored to the block sparsity structure. This objective eliminates the need for empirical look-up tables and focuses solely on reducing parametrized layer connections. Moreover, our paper provides a lightweight algorithm to achieve post-training pruning for ViTs, utilizing second-order Taylor approximation and empirical optimization to solve the proposed hardware-aware objective. Extensive experiments on ImageNet are conducted across various ViT architectures, including DeiT-B and DeiT-S, demonstrating competitive performance with other pruning methods and achieving a remarkable balance between accuracy preservation and power savings. Especially, we achieve 3.93x speedup on dedicated hardware and GPUs respectively for DeiT-B, and a power reduction by 1.4x on GPUs. Code released to https://github.com/Akimoto-Cris/LPViT.