# Incremental Network Quantization: Towards Lossless CNNs with Low-Precision Weights

Source: [Incremental Network Quantization: Towards Lossless CNNs with Low-Precision Weights](<https://www.sota2.com/research/paper/68fa63a01e209131ce9ffd52>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T20:45:17.380Z

Authors: Aojun Zhou, Anbang Yao, Yiwen Guo, Lin Xu, Yurong Chen

Published: 2017-02-10T02:30:22.000Z

## Abstract

This paper presents incremental network quantization (INQ), a novel method, targeting to efficiently convert any pre-trained full-precision convolutional neural network (CNN) model into a low-precision version whose weights are constrained to be either powers of two or zero. Unlike existing methods which are struggled in noticeable accuracy loss, our INQ has the potential to resolve this issue, as benefiting from two innovations. On one hand, we introduce three interdependent operations, namely weight partition, group-wise quantization and re-training. A well-proven measure is employed to divide the weights in each layer of a pre-trained CNN model into two disjoint groups. The weights in the first group are responsible to form a low-precision base, thus they are quantized by a variable-length encoding method. The weights in the other group are responsible to compensate for the accuracy loss from the quantization, thus they are the ones to be re-trained. On the other hand, these three operations are repeated on the latest re-trained group in an iterative manner until all the weights are converted into low-precision ones, acting as an incremental network quantization and accuracy enhancement procedure. Extensive experiments on the ImageNet classification task using almost all known deep CNN architectures including AlexNet, VGG-16, GoogleNet and ResNets well testify the efficacy of the proposed method. Specifically, at 5-bit quantization, our models have improved accuracy than the 32-bit floating-point references. Taking ResNet-18 as an example, we further show that our quantized models with 4-bit, 3-bit and 2-bit ternary weights have improved or very similar accuracy against its 32-bit floating-point baseline. Besides, impressive results with the combination of network pruning and INQ are also reported. The code is available at https://github.com/Zhouaojun/Incremental-Network-Quantization.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Image Classification on ImageNet ILSVRC-2012 (val)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-ilsvrc-2012-val.md>): Image Classification; ImageNet ILSVRC-2012 (val); Top-1 Accuracy: 68.1.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1702.03044>)

[PDF](<https://arxiv.org/pdf/1702.03044.pdf>)
