oitnews/googlenews_vectors

该数据集包含多个时间点的版本，每个版本的数据集包含以下特征：ids（字符串类型）、document（字符串类型）、metadatas（包含author、content、country、description、image、link、priority、published、source、summary、title等子字段）和embeddings（浮点数序列）。每个版本的数据集只有一个训练集，且每个训练集仅包含一个样本。数据集的大小和下载大小因版本不同而有所差异。

The dataset contains multiple versions at different time points, each with the following features: ids (string type), document (string type), metadatas (including subfields such as author, content, country, description, image, link, priority, published, source, summary, title), and embeddings (sequence of float64). Each version of the dataset has only one training set, and each training set contains only one sample. The size of the dataset and the download size vary depending on the version.