CulturalRecipes

该数据集独特之处在于，它包含了自动配对的中文和英文食谱，并且还加入了由人工编写和筛选的测试集。该数据集在每种适应方向（英译中和中译英）上都包含了训练集和验证集，并附有一个小规模的人工筛选测试集。数据集的规模宏大，包含了来自RecipeNLG的超过200万份英文食谱，以及来自下厨房的150万份中文食谱。这项任务旨在研究中英语言背景下的食谱文化适应性。

This dataset is distinctive in that it encompasses automatically aligned Chinese and English recipes, alongside a manually written and curated test set. It provides training and validation sets for each translation adaptation direction (English to Chinese and Chinese to English), together with a small-scale manually filtered test set. Boasting a large corpus, the dataset contains over 2 million English recipes sourced from RecipeNLG and 1.5 million Chinese recipes from Xiachufang. This task aims to explore the cultural adaptability of recipes in the context of Chinese and English language backgrounds.