Quantization for Rapid Deployment of Deep Neural Networks

2018/10/12 by Jun Haeng Lee, Sangwon Ha, Lee, Jun Haeng +7 · 1 citation
Computer Science · #Advanced Image and Video Retrieval Techniques #Advanced Neural Network Applications #Adversarial Robustness in Machine Learning #FOS: Computer and information sciences #Neural and Evolutionary Computing (cs.NE)

paper · pdf · doi:10.48550/arxiv.1810.05488

openalex publication_date 2018/10/12 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

This paper aims at rapid deployment of the state-of-the-art deep neural networks (DNNs) to energy efficient accelerators without time-consuming fine tuning or the availability of the full datasets. Converting DNNs in full precision to limited precision is essential in taking advantage of the accelerators with reduced memory footprint and computation power. However, such a task is not trivial since it often requires the full training and validation datasets for profiling the network statistics and fine tuning the networks to recover the accuracy lost after quantization. To address these issues, we propose a simple method recognizing channel-level distribution to reduce the quantization-induced accuracy loss and minimize the required image samples for profiling. We evaluated our method on eleven networks trained on the ImageNet classification benchmark and a network trained on the Pascal VOC object detection benchmark. The results prove that the networks can be quantized into 8-bit integer precision without fine tuning.

Citations

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• BinaryConnect: Training Deep Neural Networks with binary weights during propagations
• Ristretto: Hardware-Oriented Approximation of Convolutional Neural\n Networks
• Ternary Weight Networks
• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• Towards Effective Low-bitwidth Convolutional Neural Networks
• Rethinking the Inception Architecture for Computer Vision
• Deep Residual Learning for Image Recognition
• YOLO9000: Better, Faster, Stronger
• Xception: Deep Learning with Depthwise Separable Convolutions

Cited by

• DMQ: Dissecting Outliers of Diffusion Models for Post-Training Quantization

Related

• Bit Efficient Quantization for Deep Neural Networks
• Exploring Quantization and Mapping Synergy in Hardware-Aware Deep Neural Network Accelerators
• Training High-Performance and Large-Scale Deep Neural Networks with Full 8-bit Integers
• A White Paper on Neural Network Quantization
• A Closer Look at Hardware-Friendly Weight Quantization