Saved in:
Bibliographic Details
Main Authors: Tsao, Hsi-Ai, Hsiung, Lei, Chen, Pin-Yu, Ho, Tsung-Yi
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2409.01821
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909304108351488
author Tsao, Hsi-Ai
Hsiung, Lei
Chen, Pin-Yu
Ho, Tsung-Yi
author_facet Tsao, Hsi-Ai
Hsiung, Lei
Chen, Pin-Yu
Ho, Tsung-Yi
contents Adapting pre-trained models to new tasks can exhibit varying effectiveness across datasets. Visual prompting, a state-of-the-art parameter-efficient transfer learning method, can significantly improve the performance of out-of-distribution tasks. On the other hand, linear probing, a standard transfer learning method, can sometimes become the best approach. We propose a log-likelihood ratio (LLR) approach to analyze the comparative benefits of visual prompting and linear probing. By employing the LLR score alongside resource-efficient visual prompts approximations, our cost-effective measure attains up to a 100-fold reduction in run time compared to full training, while achieving prediction accuracies up to 91%. The source code is available at https://github.com/IBM/VP-LLR.
format Preprint
id arxiv_https___arxiv_org_abs_2409_01821
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle When Does Visual Prompting Outperform Linear Probing for Vision-Language Models? A Likelihood Perspective
Tsao, Hsi-Ai
Hsiung, Lei
Chen, Pin-Yu
Ho, Tsung-Yi
Computer Vision and Pattern Recognition
Machine Learning
Adapting pre-trained models to new tasks can exhibit varying effectiveness across datasets. Visual prompting, a state-of-the-art parameter-efficient transfer learning method, can significantly improve the performance of out-of-distribution tasks. On the other hand, linear probing, a standard transfer learning method, can sometimes become the best approach. We propose a log-likelihood ratio (LLR) approach to analyze the comparative benefits of visual prompting and linear probing. By employing the LLR score alongside resource-efficient visual prompts approximations, our cost-effective measure attains up to a 100-fold reduction in run time compared to full training, while achieving prediction accuracies up to 91%. The source code is available at https://github.com/IBM/VP-LLR.
title When Does Visual Prompting Outperform Linear Probing for Vision-Language Models? A Likelihood Perspective
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2409.01821