somosnlp-hackathon-2025/memes_instagram_chilenos_es_small

memes_instagram_chilenos_es_small数据集是一个专门为训练和评估视觉语言模型而设计的，重点在于理解和解释智利文化背景下的幽默。该数据集包含了4776个例子，来自于1194个独特的智利Instagram页面上的memes，每个memes都附带了基于智利文化背景的解释。数据集的创建包括爬取、文本生成、手动标注和指令调整等步骤。

The memes_instagram_chilenos_es_small dataset is designed specifically for training and evaluating vision-language models, with a focus on understanding and interpreting humor within the Chilean cultural context. The dataset includes 4,776 examples from 1,194 unique memes found on popular Chilean Instagram pages, each accompanied by explanations based on the Chilean cultural background. The creation of the dataset involves steps such as scraping, text generation, manual annotation, and instruction tuning.