Datasets:
lmqg
/

Languages:
Chinese
Multilinguality:
monolingual
Size Categories:
10K<n<100K
ArXiv:
License:
File size: 435 Bytes
87b3f07
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
import json

for i in ["test.jsonl", "train.jsonl", "validation.jsonl"]:
    with open(f"data/processed/{i}") as f:
        tmp = [json.loads(l) for l in f.read().split("\n") if len(l) > 0]
    print(len(tmp))
    for x in tmp:
        x["paragraph_question"] = f"question: {x['question']}, context: {x['paragraph']}"
    input()
    with open(f"data/processed/{i}", "w") as f:
        f.write("\n".join([json.dumps(x) for x in tmp]))