APLA: A Simple Adaptation Method for Vision Transformers

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sorkhei, Moein, Konuk, Emir, Smith, Kevin, Matsoukas, Christos
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909549098696704
author Sorkhei, Moein
Konuk, Emir
Smith, Kevin
Matsoukas, Christos
author_facet Sorkhei, Moein
Konuk, Emir
Smith, Kevin
Matsoukas, Christos
contents Existing adaptation techniques typically require architectural modifications or added parameters, leading to high computational costs and complexity. We introduce Attention Projection Layer Adaptation (APLA), a simple approach to adapt vision transformers (ViTs) without altering the architecture or adding parameters. Through a systematic analysis, we find that the layer immediately after the attention mechanism is crucial for adaptation. By updating only this projection layer, or even just a random subset of this layer's weights, APLA achieves state-of-the-art performance while reducing GPU memory usage by up to 52.63% and training time by up to 43.0%, with no extra cost at inference. Across 46 datasets covering a variety of tasks including scene classification, medical imaging, satellite imaging, and fine-grained classification, APLA consistently outperforms 17 other leading adaptation methods, including full fine-tuning, on classification, segmentation, and detection tasks. The code is available at https://github.com/MoeinSorkhei/APLA.
format Preprint
id arxiv_https___arxiv_org_abs_2503_11335
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle APLA: A Simple Adaptation Method for Vision Transformers
Sorkhei, Moein
Konuk, Emir
Smith, Kevin
Matsoukas, Christos
Computer Vision and Pattern Recognition
Existing adaptation techniques typically require architectural modifications or added parameters, leading to high computational costs and complexity. We introduce Attention Projection Layer Adaptation (APLA), a simple approach to adapt vision transformers (ViTs) without altering the architecture or adding parameters. Through a systematic analysis, we find that the layer immediately after the attention mechanism is crucial for adaptation. By updating only this projection layer, or even just a random subset of this layer's weights, APLA achieves state-of-the-art performance while reducing GPU memory usage by up to 52.63% and training time by up to 43.0%, with no extra cost at inference. Across 46 datasets covering a variety of tasks including scene classification, medical imaging, satellite imaging, and fine-grained classification, APLA consistently outperforms 17 other leading adaptation methods, including full fine-tuning, on classification, segmentation, and detection tasks. The code is available at https://github.com/MoeinSorkhei/APLA.
title APLA: A Simple Adaptation Method for Vision Transformers
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.11335