Exploring Visual Prompts for Adapting Large-Scale Models

Autor: Bahng, Hyojin, Jahanian, Ali, Sankaranarayanan, Swami, Isola, Phillip
Rok vydání: 2022
Předmět:
Druh dokumentu: Working Paper
Popis: We investigate the efficacy of visual prompting to adapt large-scale models in vision. Following the recent approach from prompt tuning and adversarial reprogramming, we learn a single image perturbation such that a frozen model prompted with this perturbation performs a new task. Through comprehensive experiments, we demonstrate that visual prompting is particularly effective for CLIP and robust to distribution shift, achieving performance competitive with standard linear probes. We further analyze properties of the downstream dataset, prompt design, and output transformation in regard to adaptation performance. The surprising effectiveness of visual prompting provides a new perspective on adapting pre-trained models in vision. Code is available at http://hjbahng.github.io/visual_prompting .
Comment: 16 pages, 10 figures
Databáze: arXiv