카스크

  • 설명 :

QASC는 문장 구성에 중점을 둔 질의 응답 데이터 세트입니다. 그것은 초등학교 과학에 관한 9,980개의 8방향 객관식 질문(8,134 훈련, 926 개발, 920 시험)으로 구성되어 있으며 17M 문장의 말뭉치와 함께 제공됩니다.

나뉘다
'test' 920
'train' 8,134
'validation' 926
  • 기능 구조 :
FeaturesDict({
   
'answerKey': Text(shape=(), dtype=string),
   
'choices': Sequence({
       
'label': Text(shape=(), dtype=string),
       
'text': Text(shape=(), dtype=string),
   
}),
   
'combinedfact': Text(shape=(), dtype=string),
   
'fact1': Text(shape=(), dtype=string),
   
'fact2': Text(shape=(), dtype=string),
   
'formatted_question': Text(shape=(), dtype=string),
   
'id': Text(shape=(), dtype=string),
   
'question': Text(shape=(), dtype=string),
})
  • 기능 문서 :
특징 수업 모양 D타입 설명
풍모Dict
답변키 텍스트
선택 순서
선택/라벨 텍스트
선택/텍스트 텍스트
결합 사실 텍스트
사실1 텍스트
사실2 텍스트
formatted_question 텍스트
ID 텍스트
문제 텍스트
  • 인용 :
@article{allenai:qasc,
      author    
= {Tushar Khot and Peter Clark and Michal Guerquin and Peter Jansen and Ashish Sabharwal},
      title    
= {QASC: A Dataset for Question Answering via Sentence Composition},
      journal  
= {arXiv:1910.11473v2},
      year      
= {2020},
}