<ul><li>Recent studies have introduced various approaches for prompt-tuning black-box vision-language models, referred to as black-box prompt-tuning (BBPT).</li><li>To address the issue of excessive queries in prompt-tuning, a new approach called Zeroth-order Intrinsic-dimensional Prompt-tuning (ZIP) is proposed.</li><li>ZIP reduces problem dimensionality and variance of zeroth-order gradient estimates for efficient and robust prompt optimization.</li><li>ZIP achieves state-of-the-art performance on multiple vision-language tasks, with improved few-shot accuracy and query efficiency.</li></ul>

ZIP: An Efficient Zeroth-order Prompt Tuning for Black-box Vision-Language Models

Discover more