This repository contains the code for our EMNLP 2024 paper: Can visual language models resolve textual ambiguity with visual cues? Let visual puns tell you!
Jiwan Chung, Seungwon Lim, Jaehyun Jeon, Seungbeen Lee, Youngjae Yu

Paper

Our recommendation is to access the corpus on HuggingFace:

from datasets import load_dataset

# load main data
dset = load_dataset("jiwan-chung/VisualPun_UNPIE", "heterographic")

dset = load_dataset("jiwan-chung/VisualPun_UNPIE", "homographic")

tbu