Datasets:
Tasks:
Text Generation
Modalities:
Text
Sub-tasks:
language-modeling
Languages:
Japanese
Size:
10K<n<100K
ArXiv:
License:
File size: 579 Bytes
a1681d2 82b8459 a1681d2 e1c267b 82b8459 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 |
import os
from glob import glob
from datasets import load_dataset
os.makedirs('./reference_files', exist_ok=True)
for split in ['validation', 'test']:
dataset = load_dataset('asahi417/qg_jaquad', split=split)
for data in ['question', 'answer', 'sentence', 'paragraph']:
with open('./reference_files/{}-{}.txt'.format(data, split), 'w') as f:
f.write('\n'.join([i.replace('\n', '') for i in dataset[data]]))
length = [len(open(i).read().split('\n')) for i in glob(f'reference_files/*{split}.txt')]
assert len(list(set(length))) == 1, length
|