README.md

Dataset Summary

This repository provides the mMARCO dataset. This dataset was created by translating the MS MARCO Passage Ranking dataset into 8 different languages.

Dataset Structure

You can load mMARCO dataset by choosing a specific language. For each language, there are four splits: the first one, called collection, which contains all the dataset passages, followed by three queries splits, regarding to the training, development and evaluation sets.

Load Dataset

An example to load the dataset:

dataset = load_dataset('unicamp-dl/mmarco', language='portuguese')

Citation Information

@misc{bonifacio2021mmarco,
      title={mMARCO: A Multilingual Version of the MS MARCO Passage Ranking Dataset}, 
      author={Luiz Henrique Bonifacio and Israel Campiotti and Vitor Jeronymo and Roberto Lotufo and Rodrigo Nogueira},
      year={2021},
      eprint={2108.13897},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}