cosmos_qa

  • الوصف :

Cosmos QA عبارة عن مجموعة بيانات واسعة النطاق تتكون من 35.6 ألف مشكلة تتطلب فهمًا للقراءة قائمًا على المنطق ، وقد تمت صياغته على هيئة أسئلة متعددة الخيارات. يركز على القراءة بين السطور على مجموعة متنوعة من روايات الناس اليومية ، وطرح أسئلة تتعلق بالأسباب أو الآثار المحتملة للأحداث التي تتطلب التفكير خارج نطاق النص الدقيق في السياق.

انشق، مزق أمثلة
'test' 6963
'train' 25262
'validation' 2،985
  • هيكل الميزة :
FeaturesDict({
    'answer0': Text(shape=(), dtype=string),
    'answer1': Text(shape=(), dtype=string),
    'answer2': Text(shape=(), dtype=string),
    'answer3': Text(shape=(), dtype=string),
    'context': Text(shape=(), dtype=string),
    'id': Text(shape=(), dtype=string),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=4),
    'question': Text(shape=(), dtype=string),
})
  • وثائق الميزة :
ميزة فصل شكل نوع وصف
الميزات
الإجابة 0 نص سلسلة
الإجابة 1 نص سلسلة
الإجابة 2 نص سلسلة
الجواب 3 نص سلسلة
سياق الكلام نص سلسلة
هوية شخصية نص سلسلة
ضع الكلمة المناسبة ClassLabel int64
سؤال نص سلسلة
  • الاقتباس :
@inproceedings{huang-etal-2019-cosmos,
    title = "Cosmos {QA}: Machine Reading Comprehension with Contextual Commonsense Reasoning",
    author = "Huang, Lifu  and
      Le Bras, Ronan  and
      Bhagavatula, Chandra  and
      Choi, Yejin",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)",
    year = "2019",
    url = "https://www.aclweb.org/anthology/D19-1243"
}