# Minimize Quantization Output Error with Bias Compensation

Source: [Minimize Quantization Output Error with Bias Compensation](<https://www.sota2.com/research/paper/68fa73e01e209131cea5e9db>)

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

Data updated: 2025-10-26T02:58:03.881Z

Authors: Cheng Gong, Haoshuai Zheng, Mengting Hu, Zheng Lin, Deng-Ping Fan, Yuzhi Zhang, Tao Li

Published: 2024-04-02T12:29:31.000Z

## Abstract

Quantization is a promising method that reduces memory usage and computational intensity of Deep Neural Networks (DNNs), but it often leads to significant output error that hinder model deployment. In this paper, we propose Bias Compensation (BC) to minimize the output error, thus realizing ultra-low-precision quantization without model fine-tuning. Instead of optimizing the non-convex quantization process as in most previous methods, the proposed BC bypasses the step to directly minimize the quantizing output error by identifying a bias vector for compensation. We have established that the minimization of output error through BC is a convex problem and provides an efficient strategy to procure optimal solutions associated with minimal output error,without the need for training or fine-tuning. We conduct extensive experiments on Vision Transformer models and Large Language Models, and the results show that our method notably reduces quantization output error, thereby permitting ultra-low-precision post-training quantization and enhancing the task performance of models. Especially, BC improves the accuracy of ViT-B with 4-bit PTQ4ViT by 36.89% on the ImageNet-1k task, and decreases the perplexity of OPT-350M with 3-bit GPTQ by 5.97 on WikiText2.The code is in https://github.com/GongCheng1919/bias-compensation.

## 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.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2404.01892>)

[PDF](<https://arxiv.org/pdf/2404.01892.pdf>)
