update

generate_reference_files.py

@@ -9,10 +9,7 @@ for split in ['validation', 'test']:
     dataset = load_dataset('asahi417/qg_jaquad', split=split, download_mode='force_redownload')
     for data in ['question', 'answer', 'sentence', 'paragraph']:
         with open('./reference_files/{}-{}.txt'.format(data, split), 'w') as f:
-            if data == 'paragraph':
-                tmp_data = dataset['paragraph_id']
-            else:
-                tmp_data = dataset[data]
+            tmp_data = dataset[data]
             f.write('\n'.join([i.replace('\n', '') for i in tmp_data]))
     length = [len(open(i).read().split('\n')) for i in glob(f'reference_files/*{split}.txt')]
     assert len(list(set(length))) == 1, length