sentence-transformers/NanoTouche2020-bm25

这是一个名为NanoBEIR Touche2020 with BM25 rankings的数据集，是BEIR信息检索基准的一个子集Touche2020的更新版本。这个数据集设计得更加高效，以便于运行。它包括三个配置：语料库(corpus)、查询(queries)和相关性(relevance)。语料库包含文本数据，查询包含查询语句，相关性配置包含语料库段落与查询的相关性信息，包括BM25排名。该数据集用于评估Sentence Transformers中的交叉编码器模型，通过重新排序BM25的前*k*个结果。

This is a dataset called NanoBEIR Touche2020 with BM25 rankings, which is an updated version of a subset of the BEIR Information Retrieval Benchmark called Touche2020. This dataset is designed to be more efficient for operation. It includes three configurations: corpus, queries, and relevance. The corpus contains text data, queries contain query statements, and the relevance configuration contains information about the relevance of corpus passages to queries, including BM25 rankings. This dataset is used to evaluate CrossEncoder models in Sentence Transformers by reranking the top *k* results from BM25.