기가워드

  • 설명 :

약 400만 개의 기사로 구성된 Gigaword의 기사 쌍 코퍼스에서 헤드라인 생성. https://github.com/harvardnlp/sent-summary 와 동일하지만 형식이 더 나은 https://github.com/microsoft/unilm/ 에서 제공하는 'org_data'를 사용합니다.

두 가지 기능이 있습니다. - 문서: 기사. - 요약: 제목.

나뉘다
'test' 1,951
'train' 3,803,957
'validation' 189,651
  • 기능 구조 :
FeaturesDict({
    'document': Text(shape=(), dtype=string),
    'summary': Text(shape=(), dtype=string),
})
  • 기능 문서 :
특징 수업 모양 D타입 설명
풍모Dict
문서 텍스트
요약 텍스트
  • 인용 :
@article{graff2003english,
  title={English gigaword},
  author={Graff, David and Kong, Junbo and Chen, Ke and Maeda, Kazuaki},
  journal={Linguistic Data Consortium, Philadelphia},
  volume={4},
  number={1},
  pages={34},
  year={2003}
}

@article{Rush_2015,
   title={A Neural Attention Model for Abstractive Sentence Summarization},
   url={http://dx.doi.org/10.18653/v1/D15-1044},
   DOI={10.18653/v1/d15-1044},
   journal={Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing},
   publisher={Association for Computational Linguistics},
   author={Rush, Alexander M. and Chopra, Sumit and Weston, Jason},
   year={2015}
}