Join the High Accuracy Club on ImageNet with A Binary Neural Network Ticket

TL;DR

This paper introduces BNext, a novel binary neural network architecture that, through improved optimization, knowledge distillation, and data augmentation, achieves over 80% accuracy on ImageNet, surpassing previous binary models.

Contribution

The paper presents BNext, a new binary neural network architecture with enhanced training techniques that achieve state-of-the-art accuracy on ImageNet.

Findings

BNext achieves 80.57% top-1 accuracy on ImageNet.

It significantly outperforms existing binary neural networks.

The proposed methods effectively close the accuracy gap with full-precision models.

Abstract

Binary neural networks are the extreme case of network quantization, which has long been thought of as a potential edge machine learning solution. However, the significant accuracy gap to the full-precision counterparts restricts their creative potential for mobile applications. In this work, we revisit the potential of binary neural networks and focus on a compelling but unanswered problem: how can a binary neural network achieve the crucial accuracy level (e.g., 80%) on ILSVRC-2012 ImageNet? We achieve this goal by enhancing the optimization process from three complementary perspectives: (1) We design a novel binary architecture BNext based on a comprehensive study of binary architectures and their optimization process. (2) We propose a novel knowledge-distillation technique to alleviate the counter-intuitive overfitting problem observed when attempting to train extremely accurate binary models. (3) We analyze the data augmentation pipeline for binary networks and modernize it with up-to-date techniques from full-precision models. The evaluation results on ImageNet show that BNext, for the first time, pushes the binary model accuracy boundary to 80.57% and significantly outperforms all the existing binary networks. Code and trained models are available at: https://github.com/hpi-xnor/BNext.git.