Saved in:
Bibliographic Details
Main Authors: Shi, Bo-Yun, Lo, Yi-Cheng, An-Yeu, Wu, Tsai, Yi-Min
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2501.16738
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917920706134016
author Shi, Bo-Yun
Lo, Yi-Cheng
An-Yeu
Wu
Tsai, Yi-Min
author_facet Shi, Bo-Yun
Lo, Yi-Cheng
An-Yeu
Wu
Tsai, Yi-Min
contents The Mamba model, utilizing a structured state-space model (SSM), offers linear time complexity and demonstrates significant potential. Vision Mamba (ViM) extends this framework to vision tasks by incorporating a bidirectional SSM and patch embedding, surpassing Transformer-based models in performance. While model quantization is essential for efficient computing, existing works have focused solely on the original Mamba model and have not been applied to ViM. Additionally, they neglect quantizing the SSM layer, which is central to Mamba and can lead to substantial error propagation by naive quantization due to its inherent structure. In this paper, we focus on the post-training quantization (PTQ) of ViM. We address the issues with three core techniques: 1) a k-scaled token-wise quantization method for linear and convolutional layers, 2) a reparameterization technique to simplify hidden state quantization, and 3) a factor-determining method that reduces computational overhead by integrating operations. Through these methods, the error caused by PTQ can be mitigated. Experimental results on ImageNet-1k demonstrate only a 0.8-1.2\% accuracy degradation due to PTQ, highlighting the effectiveness of our approach.
format Preprint
id arxiv_https___arxiv_org_abs_2501_16738
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Post-Training Quantization for Vision Mamba with k-Scaled Quantization and Reparameterization
Shi, Bo-Yun
Lo, Yi-Cheng
An-Yeu
Wu
Tsai, Yi-Min
Image and Video Processing
The Mamba model, utilizing a structured state-space model (SSM), offers linear time complexity and demonstrates significant potential. Vision Mamba (ViM) extends this framework to vision tasks by incorporating a bidirectional SSM and patch embedding, surpassing Transformer-based models in performance. While model quantization is essential for efficient computing, existing works have focused solely on the original Mamba model and have not been applied to ViM. Additionally, they neglect quantizing the SSM layer, which is central to Mamba and can lead to substantial error propagation by naive quantization due to its inherent structure. In this paper, we focus on the post-training quantization (PTQ) of ViM. We address the issues with three core techniques: 1) a k-scaled token-wise quantization method for linear and convolutional layers, 2) a reparameterization technique to simplify hidden state quantization, and 3) a factor-determining method that reduces computational overhead by integrating operations. Through these methods, the error caused by PTQ can be mitigated. Experimental results on ImageNet-1k demonstrate only a 0.8-1.2\% accuracy degradation due to PTQ, highlighting the effectiveness of our approach.
title Post-Training Quantization for Vision Mamba with k-Scaled Quantization and Reparameterization
topic Image and Video Processing
url https://arxiv.org/abs/2501.16738