File size: 600 Bytes
3e20849 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
from datasets import load_dataset
import json
import random
dataset = load_dataset("ethos", "binary")
id2label = dataset['train'].features['label'].names
rows = [{'text': row['text'], 'label': row['label'], 'label_text': id2label[row['label']].replace("_", " ")} for row in dataset['train']]
random.seed(42)
random.shuffle(rows)
num_test = 400
datasplits = {'train': rows[num_test:], 'test': rows[0:num_test]}
for split in ['train', 'test']:
with open(f'{split}.jsonl', 'w') as fOut:
for row in datasplits[split]:
fOut.write(json.dumps(row)+"\n")
|