README.md

metadata

dataset_info:
  features:
    - name: query
      dtype: string
    - name: positive
      dtype: string
    - name: negative
      dtype: string
  splits:
    - name: train
      num_bytes: 34534407690
      num_examples: 39780811
  download_size: 13523306019
  dataset_size: 34534407690
task_categories:
  - question-answering
language:
  - es
size_categories:
  - 10M<n<100M
license: apache-2.0

Dataset Card for "ms-marco-es"

QA asymmetric Spanish dataset filtered from multilingual version of MS Marco

import datasets


ms_marco_es = datasets.load_dataset('unicamp-dl/mmarco', name='spanish', split='train')
ms_marco_es.push_to_hub("dariolopez/ms-marco-es", token=os.environ['hg_token'])