huggingface / huggingface/datasets
Something is wrong with WMT 19 kk-en dataset
Open
dataset bug
- Dominant language
- Python
- Stars
- 22k
- Forks
- 3.4k
- Avg merge
- 5d 7h
- Merged PRs (30d)
- 17
Description
The translation in the `train` set does not look right:
```
>>>import nlp
>>>from nlp import load_dataset
>>>dataset = load_dataset('wmt19', 'kk-en')
>>>dataset["train"]["translation"][0]
{'kk': 'Trumpian Uncertainty', 'en': 'Трамптық белгісіздік'}
>>>dataset["validation"]["translation"][0]
{'kk': 'Ақша-несие саясатының сценарийін қайта жазсақ', 'en': 'Rewriting the Monetary-Policy Script'}
```
Contributor guide
Assessment
This issue has not been assessed yet.