huggingface / huggingface/datasets
OSCAR unshuffled_original_ko: NonMatchingSplitsSizesError
- Dominant language
- Python
- Stars
- 22k
- Forks
- 3.4k
- Avg merge
- 5d 7h
- Merged PRs (30d)
- 17
Description
## Describe the bug
Cannot download OSCAR `unshuffled_original_ko` due to `NonMatchingSplitsSizesError`.
## Steps to reproduce the bug
```python
>>> dataset = datasets.load_dataset('oscar', 'unshuffled_original_ko')
NonMatchingSplitsSizesError: [{'expected': SplitInfo(name='train', num_bytes=25292102197, num_examples=7345075, dataset_name='oscar'), 'recorded': SplitInfo(name='train', num_bytes=25284578514, num_examples=7344907, dataset_name='oscar')}]
```
## Expected results
Loading is successful.
## Actual results
Loading throws above error.
## Environment info
- `datasets` version: 1.12.1
- Platform: Linux-5.4.0-81-generic-x86_64-with-glibc2.29
- Python version: 3.8.10
- PyArrow version: 5.0.0
Contributor guide
Assessment
This issue has not been assessed yet.