Outlier-Aware Training for Low-Bit Quantization of Structural Re-Parameterized Networks

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Niu, Muqun, Ren, Yuan, Li, Boyu, Ding, Chenchen
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866909102078164992
author Niu, Muqun
Ren, Yuan
Li, Boyu
Ding, Chenchen
author_facet Niu, Muqun
Ren, Yuan
Li, Boyu
Ding, Chenchen
contents Lightweight design of Convolutional Neural Networks (CNNs) requires co-design efforts in the model architectures and compression techniques. As a novel design paradigm that separates training and inference, a structural re-parameterized (SR) network such as the representative RepVGG revitalizes the simple VGG-like network with a high accuracy comparable to advanced and often more complicated networks. However, the merging process in SR networks introduces outliers into weights, making their distribution distinct from conventional networks and thus heightening difficulties in quantization. To address this, we propose an operator-level improvement for training called Outlier Aware Batch Normalization (OABN). Additionally, to meet the demands of limited bitwidths while upkeeping the inference accuracy, we develop a clustering-based non-uniform quantization framework for Quantization-Aware Training (QAT) named ClusterQAT. Integrating OABN with ClusterQAT, the quantized performance of RepVGG is largely enhanced, particularly when the bitwidth falls below 8.
format Preprint
id arxiv_https___arxiv_org_abs_2402_07200
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Outlier-Aware Training for Low-Bit Quantization of Structural Re-Parameterized Networks
Niu, Muqun
Ren, Yuan
Li, Boyu
Ding, Chenchen
Computer Vision and Pattern Recognition
Machine Learning
Neural and Evolutionary Computing
Lightweight design of Convolutional Neural Networks (CNNs) requires co-design efforts in the model architectures and compression techniques. As a novel design paradigm that separates training and inference, a structural re-parameterized (SR) network such as the representative RepVGG revitalizes the simple VGG-like network with a high accuracy comparable to advanced and often more complicated networks. However, the merging process in SR networks introduces outliers into weights, making their distribution distinct from conventional networks and thus heightening difficulties in quantization. To address this, we propose an operator-level improvement for training called Outlier Aware Batch Normalization (OABN). Additionally, to meet the demands of limited bitwidths while upkeeping the inference accuracy, we develop a clustering-based non-uniform quantization framework for Quantization-Aware Training (QAT) named ClusterQAT. Integrating OABN with ClusterQAT, the quantized performance of RepVGG is largely enhanced, particularly when the bitwidth falls below 8.
title Outlier-Aware Training for Low-Bit Quantization of Structural Re-Parameterized Networks
topic Computer Vision and Pattern Recognition
Machine Learning
Neural and Evolutionary Computing
url https://arxiv.org/abs/2402.07200