cosmos_qa

  • 説明:

Cosmos QA は、常識に基づいた読解力を必要とする 35.6K の問題の大規模なデータセットであり、多肢選択問題として作成されています。人々の日常の物語の多様なコレクションの行間を読むことに焦点を当て、文脈内の正確なテキスト範囲を超えた推論を必要とする出来事の可能性のある原因または結果について質問します。

スプリット
'test' 6,963
'train' 25,262
'validation' 2,985
  • 機能構造:
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),
})
  • 機能のドキュメント:
特徴クラスDtype説明
特徴辞書
答え0文章ストリング
答え1文章ストリング
答え2文章ストリング
答え3文章ストリング
環境文章ストリング
ID文章ストリング
ラベルクラスラベルint64
質問文章ストリング
  • 引用
@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"
}