Adapting Vision-Language Models to Open Classes via Test-Time Prompt Tuning

TL;DR

This paper introduces a test-time prompt tuning method for vision-language models that dynamically adapts prompts based on input, improving open-set recognition by leveraging concept matching scores.

Contribution

It proposes a novel test-time prompt tuning approach that combines learned and hand-crafted prompts using concept matching scores for better open-set adaptation.

Findings

Outperforms existing methods on 11 datasets

Improves recognition of both base and new classes

Demonstrates robustness in open-set scenarios

Abstract

Adapting pre-trained models to open classes is a challenging problem in machine learning. Vision-language models fully explore the knowledge of text modality, demonstrating strong zero-shot recognition performance, which is naturally suited for various open-set problems. More recently, some research focuses on fine-tuning such models to downstream tasks. Prompt tuning methods achieved huge improvements by learning context vectors on few-shot data. However, through the evaluation under open-set adaptation setting with the test data including new classes, we find that there exists a dilemma that learned prompts have worse generalization abilities than hand-crafted prompts. In this paper, we consider combining the advantages of both and come up with a test-time prompt tuning approach, which leverages the maximum concept matching (MCM) scores as dynamic weights to generate an input-conditioned prompt for each image during test. Through extensive experiments on 11 different datasets, we show that our proposed method outperforms all comparison methods on average considering both base and new classes. The code is available at https://github.com/gaozhengqing/TTPT