TUKE-KEMT/reranking-skquad

该数据集用于训练和评估斯洛伐克信息检索的重排模型，包含用户查询、相关文档和无关文档，适用于评估交叉编码模型。数据集由queries.jsonl、corpus.jsonl和qrels.jsonl三个文件生成，每个实例包括一个查询和与之相关的正面和负面段落。

This dataset is for training and evaluating reranking models for Slovak information retrieval, containing user queries, relevant documents, and irrelevant documents, which is particularly useful for evaluating cross-encoder models. The dataset is generated from three files: queries.jsonl, corpus.jsonl, and qrels.jsonl, with each instance including a query and associated positive and negative passages.