LampQ: Towards Accurate Layer-wise Mixed Precision Quantization for Vision Transformers

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Minjun, Lee, Jaeri, Kim, Jongjin, Yun, Jeongin, Kwon, Yongmo, Kang, U
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908650934632448
author Kim, Minjun
Lee, Jaeri
Kim, Jongjin
Yun, Jeongin
Kwon, Yongmo
Kang, U
author_facet Kim, Minjun
Lee, Jaeri
Kim, Jongjin
Yun, Jeongin
Kwon, Yongmo
Kang, U
contents How can we accurately quantize a pre-trained Vision Transformer model? Quantization algorithms compress Vision Transformers (ViTs) into low-bit formats, reducing memory and computation demands with minimal accuracy degradation. However, existing methods rely on uniform precision, ignoring the diverse sensitivity of ViT components to quantization. Metric-based Mixed Precision Quantization (MPQ) is a promising alternative, but previous MPQ methods for ViTs suffer from three major limitations: 1) coarse granularity, 2) mismatch in metric scale across component types, and 3) quantization-unaware bit allocation. In this paper, we propose LampQ (Layer-wise Mixed Precision Quantization for Vision Transformers), an accurate metric-based MPQ method for ViTs to overcome these limitations. LampQ performs layer-wise quantization to achieve both fine-grained control and efficient acceleration, incorporating a type-aware Fisher-based metric to measure sensitivity. Then, LampQ assigns bit-widths optimally through integer linear programming and further updates them iteratively. Extensive experiments show that LampQ provides the state-of-the-art performance in quantizing ViTs pre-trained on various tasks such as image classification, object detection, and zero-shot quantization.
format Preprint
id arxiv_https___arxiv_org_abs_2511_10004
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LampQ: Towards Accurate Layer-wise Mixed Precision Quantization for Vision Transformers
Kim, Minjun
Lee, Jaeri
Kim, Jongjin
Yun, Jeongin
Kwon, Yongmo
Kang, U
Computer Vision and Pattern Recognition
How can we accurately quantize a pre-trained Vision Transformer model? Quantization algorithms compress Vision Transformers (ViTs) into low-bit formats, reducing memory and computation demands with minimal accuracy degradation. However, existing methods rely on uniform precision, ignoring the diverse sensitivity of ViT components to quantization. Metric-based Mixed Precision Quantization (MPQ) is a promising alternative, but previous MPQ methods for ViTs suffer from three major limitations: 1) coarse granularity, 2) mismatch in metric scale across component types, and 3) quantization-unaware bit allocation. In this paper, we propose LampQ (Layer-wise Mixed Precision Quantization for Vision Transformers), an accurate metric-based MPQ method for ViTs to overcome these limitations. LampQ performs layer-wise quantization to achieve both fine-grained control and efficient acceleration, incorporating a type-aware Fisher-based metric to measure sensitivity. Then, LampQ assigns bit-widths optimally through integer linear programming and further updates them iteratively. Extensive experiments show that LampQ provides the state-of-the-art performance in quantizing ViTs pre-trained on various tasks such as image classification, object detection, and zero-shot quantization.
title LampQ: Towards Accurate Layer-wise Mixed Precision Quantization for Vision Transformers
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.10004