aniruddha007/example-retrieval-reranking-dataset

该数据集包含两个配置：generate_reranking_pairs和generate_retrieval_pairs，每个配置都有特定的特征和结构。数据集用于生成与锚点句子相关的正负样本，适用于检索和重排序任务。数据集是通过distilabel工具生成的，并提供了用于复现生成过程的pipeline.yaml文件。

This dataset contains two configurations: generate_reranking_pairs and generate_retrieval_pairs. Each configuration includes training data with features such as filename, anchor, repo_name, positive, negative, distilabel_metadata, and model_name. The distilabel_metadata includes raw input and output information. The dataset is used for generating reranking and retrieval pairs, respectively for semantic similarity and query generation tasks. The dataset can be reproduced and explored using the distilabel CLI.