---
dataset_info:
  features:
  - name: title
    dtype: string
  - name: news_category
    dtype: string
  - name: source
    dtype: string
  - name: context
    dtype: string
  - name: question
    dtype: string
  - name: question_type
    dtype: int64
  - name: is_impossible
    dtype: bool
  - name: answer_text
    dtype: string
  - name: answer_start
    dtype: int64
  - name: negative_samples
    sequence: string
  - name: search_result
    sequence: string
  - name: answer
    dtype: string
  - name: extracted_ref_numbers
    sequence: int64
  splits:
  - name: train
    num_bytes: 4900918
    num_examples: 196
  download_size: 2824532
  dataset_size: 4900918
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---