UBC-NLP/PEARL
收藏Hugging Face2025-05-29 更新2025-05-31 收录
下载链接：
https://hf-mirror.com/datasets/UBC-NLP/PEARL
下载链接
链接失效反馈官方服务：
资源简介：
PEARL基准数据集是从更大的PEARL数据集中精心挑选出的6867个高质量的问题/答案对。它专门设计用于评估视觉语言模型(VLMs)对阿拉伯文化内容的理解。该基准涵盖了十个重要的文化领域（例如，建筑、服装、美食）和十三种不同的问题类型，旨在对模型的感知、知识回忆和关于阿拉伯文化的推理能力进行稳健评估。
The PEARL Benchmark is a meticulously curated subset of 6,867 high-quality Question/Answer pairs derived from the larger PEARL dataset. It is specifically designed for evaluating Vision-Language Models (VLMs) understanding of Arab cultural content. The benchmark covers ten significant cultural domains (e.g., architecture, clothing, cuisine) and thirteen diverse question types, tailored for robust model assessment of perception, knowledge recall, and reasoning about Arab culture.
提供机构：
UBC-NLP
