CPT: Colorful Prompt Tuning for Pre-trained Vision-Language Models
This addresses the need for large labeled data in vision-language tasks, offering a novel tuning method that enhances few-shot and zero-shot capabilities, though it is incremental as it builds on existing prompt tuning paradigms.
The paper tackles the gap between pre-training and fine-tuning in vision-language models by introducing Cross-modal Prompt Tuning (CPT), which reformulates visual grounding as a fill-in-the-blank problem using color-based markers, resulting in significant improvements such as a 17.3% absolute accuracy gain and 73.8% relative standard deviation reduction in few-shot evaluations.
Pre-Trained Vision-Language Models (VL-PTMs) have shown promising capabilities in grounding natural language in image data, facilitating a broad variety of cross-modal tasks. However, we note that there exists a significant gap between the objective forms of model pre-training and fine-tuning, resulting in a need for large amounts of labeled data to stimulate the visual grounding capability of VL-PTMs for downstream tasks. To address the challenge, we present Cross-modal Prompt Tuning (CPT, alternatively, Colorful Prompt Tuning), a novel paradigm for tuning VL-PTMs, which reformulates visual grounding into a fill-in-the-blank problem with color-based co-referential markers in image and text, maximally mitigating the gap. In this way, CPT enables strong few-shot and even zero-shot visual grounding capabilities of VL-PTMs. Comprehensive experimental results show that the prompt-tuned VL-PTMs outperform their fine-tuned counterparts by a large margin (e.g., 17.3% absolute accuracy improvement, and 73.8% relative standard deviation reduction on average with one shot in RefCOCO evaluation). We make the data and code for this paper publicly available at https://github.com/thunlp/CPT.