hltcoe/tdist-msmarco-scores

该数据集包含由MonoT5 reranker生成的MS MARCO训练查询-段落分数，这些分数由unicamp-dl/mt5-13b-mmarco-100k和castorini/monot5-3b-msmarco-10k两个模型生成。每个训练查询与ColBERTv2模型检索的前50个段落相关联。文件以gzip压缩格式存储，并遵循特定的命名规则。此外，还提供了MS MARCO训练查询的波斯语翻译，这些翻译未包含在neuMARCO或mMARCO中。

This dataset contains MS MARCO training query-passage scores generated by the MonoT5 reranker, specifically from the models unicamp-dl/mt5-13b-mmarco-100k and castorini/monot5-3b-msmarco-10k. Each training query is associated with the top-50 passages retrieved by the ColBERTv2 model. The files are stored in gzip compressed format and follow a specific naming scheme. Additionally, Persian translations of the MS MARCO training queries are provided, which were not included in either neuMARCO or mMARCO.

原始信息汇总

MS MARCO Distillation Scores for Translate-Distill

数据集描述

该数据集包含由MonoT5重排序器生成的MS MARCO训练查询-段落分数，具体使用的是以下两个模型：

• unicamp-dl/mt5-13b-mmarco-100k
• castorini/monot5-3b-msmarco-10k

每个训练查询关联了由ColBERTv2模型检索的前50个段落。

文件格式

文件采用gzip压缩，命名格式为{teacher}-monot5-{msmarco, mmarco}-{qlang}{plang}.jsonl.gz，其中qlang表示查询语言，plang表示段落语言。对于非英语语言，使用mmarco和neuMarco提供的翻译文本。

额外提供

提供了MS MARCO训练查询的波斯语翻译，文件位于msmarco.train.query.fas.tsv.gz。

引用信息

使用该数据集时，请引用以下论文：

bibtext @inproceedings{translate-distill, author = {Eugene Yang and Dawn Lawrie and James Mayfield and Douglas W. Oard and Scott Miller}, title = {Translate-Distill: Learning Cross-Language Dense Retrieval by Translation and Distillation}, booktitle = {Proceedings of the 46th European Conference on Information Retrieval (ECIR)}, year = {2024}, url = {https://arxiv.org/abs/2401.04810} }