coqa

  • Description:

CoQA: A Conversational Question Answering Challenge

Split Examples
'test' 500
'train' 7,199
  • Feature structure:
FeaturesDict({
    'answers': Sequence({
        'answer_end': tf.int32,
        'answer_start': tf.int32,
        'input_text': Text(shape=(), dtype=tf.string),
    }),
    'questions': Sequence(Text(shape=(), dtype=tf.string)),
    'source': Text(shape=(), dtype=tf.string),
    'story': Text(shape=(), dtype=tf.string),
})
  • Feature documentation:
Feature Class Shape Dtype Description
FeaturesDict
answers Sequence
answers/answer_end Tensor tf.int32
answers/answer_start Tensor tf.int32
answers/input_text Text tf.string
questions Sequence(Text) (None,) tf.string
source Text tf.string
story Text tf.string
  • Citation:
@misc{reddy2018coqa,
    title={CoQA: A Conversational Question Answering Challenge},
    author={Siva Reddy and Danqi Chen and Christopher D. Manning},
    year={2018},
    eprint={1808.07042},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}