Saved in:
Bibliographic Details
Main Authors: Cui, Fangming, Yang, Xun, Wu, Chao, Xiao, Liang, Tian, Xinmei
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.19674
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910698783637504
author Cui, Fangming
Yang, Xun
Wu, Chao
Xiao, Liang
Tian, Xinmei
author_facet Cui, Fangming
Yang, Xun
Wu, Chao
Xiao, Liang
Tian, Xinmei
contents Prompt learning represents a promising method for adapting pre-trained vision-language models (VLMs) to various downstream tasks by learning a set of text embeddings. One challenge inherent to these methods is the poor generalization performance due to the invalidity of the learned text embeddings for unseen tasks. A straightforward approach to bridge this gap is to freeze the text embeddings in prompts, which results in a lack of capacity to adapt VLMs for downstream tasks. To address this dilemma, we propose a paradigm called EnPrompt with a novel External Layer (EnLa). Specifically, we propose a textual external layer and learnable visual embeddings for adapting VLMs to downstream tasks. The learnable external layer is built upon valid embeddings of pre-trained CLIP. This design considers the balance of learning capabilities between the two branches. To align the textual and visual features, we propose a novel two-pronged approach: i) we introduce the optimal transport as the discrepancy metric to align the vision and text modalities, and ii) we introduce a novel strengthening feature to enhance the interaction between these two modalities. Four representative experiments (i.e., base-to-novel generalization, few-shot learning, cross-dataset generalization, domain shifts generalization) across 15 datasets demonstrate that our method outperforms the existing prompt learning method.
format Preprint
id arxiv_https___arxiv_org_abs_2407_19674
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Advancing Prompt Learning through an External Layer
Cui, Fangming
Yang, Xun
Wu, Chao
Xiao, Liang
Tian, Xinmei
Computer Vision and Pattern Recognition
Prompt learning represents a promising method for adapting pre-trained vision-language models (VLMs) to various downstream tasks by learning a set of text embeddings. One challenge inherent to these methods is the poor generalization performance due to the invalidity of the learned text embeddings for unseen tasks. A straightforward approach to bridge this gap is to freeze the text embeddings in prompts, which results in a lack of capacity to adapt VLMs for downstream tasks. To address this dilemma, we propose a paradigm called EnPrompt with a novel External Layer (EnLa). Specifically, we propose a textual external layer and learnable visual embeddings for adapting VLMs to downstream tasks. The learnable external layer is built upon valid embeddings of pre-trained CLIP. This design considers the balance of learning capabilities between the two branches. To align the textual and visual features, we propose a novel two-pronged approach: i) we introduce the optimal transport as the discrepancy metric to align the vision and text modalities, and ii) we introduce a novel strengthening feature to enhance the interaction between these two modalities. Four representative experiments (i.e., base-to-novel generalization, few-shot learning, cross-dataset generalization, domain shifts generalization) across 15 datasets demonstrate that our method outperforms the existing prompt learning method.
title Advancing Prompt Learning through an External Layer
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2407.19674