---
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](https://huggingface.co/datasets/unicamp-dl/mmarco)

```python
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'])
```