English

MicroViTv2: Beyond the FLOPS for Edge Energy-Friendly Vision Transformers

Computer Vision and Pattern Recognition 2026-05-12 v1

Abstract

The Vision Transformer (ViT) achieves remarkable accuracy across visual tasks but remains computationally expensive for edge deployment. This paper presents MicroViTv2, a lightweight Vision Transformer optimized for real-device efficiency. Built upon the original MicroViT, the proposed model is designed based on reparameterized design, specifically Reparameterized Patch Embedding (RepEmbed) and Reparameterized Depth-Wise convolution mixer (RepDW) for faster inference, and introduces the Single Depth-Wise Transposed Attention (SDTA) to capture long-range dependencies with minimal redundancy. Despite slightly higher FLOPs, MicroViTv2 improves accuracy up to 0.5% compared to its predecessor and surpassing MobileViTv2, EdgeNeXt, and EfficientViT while maintaining fast inference and high energy efficiency on Jetson AGX Orin. Experiments on ImageNet-1K and COCO demonstrate that hardware-aware design and structural re-parameterization are key to achieving high accuracy and low energy consumption, validating the need to evaluate efficiency beyond FLOPs. Code is available at https://github.com/novendrastywn/MicroViT.

Keywords

Cite

@article{arxiv.2605.10148,
  title  = {MicroViTv2: Beyond the FLOPS for Edge Energy-Friendly Vision Transformers},
  author = {Novendra Setyawan and Chi-Chia Sun and Mao-Hsiu Hsu and Wen-Kai Kuo and Jun-Wei Hsieh},
  journal= {arXiv preprint arXiv:2605.10148},
  year   = {2026}
}
R2 v1 2026-07-22T07:03:35.892Z