open-vdb/glove-100-angular

这是一个名为glove-100-angular的数据集，用于基准测试和研究目的。它包含了训练集、测试集和邻居集，每个集都包含了索引和嵌入向量。训练集大小为456.526 MB，包含1,183,514行数据；测试集大小为3.711 MB，包含10,000行数据；邻居集大小为98.829 MB，也包含10,000行数据。数据集的模式包括索引和嵌入向量列表。此外，邻居集还包含了邻居ID、邻居距离、度量方法、查询表达式、主键字段名、向量字段名和top_k等信息。数据集的使用受到原始数据源许可的限制，并且地面真实部分遵循Apache 2.0许可。

This is a dataset named glove-100-angular, used for benchmarking and research purposes. It includes training, test, and neighbors sets, each containing indices and embedding vectors. The training set is 456.526 MB in size with 1,183,514 rows of data; the test set is 3.711 MB with 10,000 rows of data; and the neighbors set is 98.829 MB with 10,000 rows of data as well. The schema of the dataset includes an index and a list of embedding vectors. Additionally, the neighbors set also contains neighbor IDs, neighbor distances, metric methods, query expressions, primary key field names, vector field names, and top_k information. The use of the dataset is subject to the restrictions of the original data source licenses, and the ground truth part is licensed under Apache 2.0.