JamieSJS/stanford-online-products

该数据集包含三个主要配置：qrels、corpus和query。qrels配置用于存储查询与语料库之间的相关性评分，包含查询ID、语料库ID和分数特征，测试分割包含840,927个示例。corpus配置用于存储语料库内容，包含ID、模态和图像特征，语料库分割包含120,053个示例。query配置用于存储查询内容，包含ID、模态和图像特征，测试分割包含120,053个示例。数据文件以parquet格式存储。

The dataset consists of three main configurations: qrels, corpus, and query. The qrels configuration is used to store relevance scores between queries and corpus, containing features such as query-id, corpus-id, and score, with the test split containing 840,927 examples. The corpus configuration is used to store corpus content, containing features such as id, modality, and image, with the corpus split containing 120,053 examples. The query configuration is used to store query content, containing features such as id, modality, and image, with the test split containing 120,053 examples. Data files are stored in parquet format.