Ruichen0424/AB-BNN

A&B BNN: Add&Bit-Operation-Only Hardware-Friendly Binary Neural Network

🚀 Introduction

This repository contains the pre-trained weights for the paper "A&B BNN: Add&Bit-Operation-Only Hardware-Friendly Binary Neural Network", published in CVPR 2024.

A&B BNN proposes to directly remove part of the multiplication operations in a traditional BNN and replace the rest with an equal number of bit operations. It introduces the mask layer and the quantized RPReLU structure based on the normalizer-free network architecture.

✨ Key Highlights

• Hardware-Friendly: Removes multiplication operations, replacing them with bit operations.
• Competitive Performance: Achieves 92.30%, 69.35%, and 66.89% on CIFAR-10, CIFAR-100, and ImageNet respectively.
• Innovative Structures: Introduces mask layer and quantized RPReLU.

🏆 Model Zoo & Results

We provide pre-trained models for CIFAR-10, CIFAR-100, and ImageNet. You can download the .h5 files directly from the Files and versions tab in this repository.

Dataset	Structure	# Params	Top-1 Acc
CIFAR10	ReActNet-18	11.18 M	91.94%
	ReActNet-A	28.32 M	89.44%
CIFAR100	ReActNet-18	11.23 M	69.35%
	ReActNet-A	28.41 M	63.23%
ImageNet	ReActNet-18	11.70 M	61.39%
	ReActNet-34	21.82 M	65.19%
	ReActNet-A	29.33 M	66.89%

💻 Usage

This repository hosts the model weights only.

For the training scripts, inference codes, and detailed usage instructions, please refer to our official GitHub repository.

📜 Citation

If you find our code useful for your research, please consider citing:

@inproceedings{ma2024b,
  title={A\&B BNN: Add\&Bit-Operation-Only Hardware-Friendly Binary Neural Network},
  author={Ma, Ruichen and Qiao, Guanchao and Liu, Yian and Meng, Liwei and Ning, Ning and Liu, Yang and Hu, Shaogang},
  booktitle={2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  pages={5704--5713},
  year={2024},
  organization={IEEE}
}