metadata
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
list: string
- name: multi_document_question
dtype: string
- name: multi_document_answer
dtype: string
- name: extracted_ref_numbers
list: int64
- name: search_result
list: string
splits:
- name: train
num_bytes: 12764425
num_examples: 485
download_size: 7198131
dataset_size: 12764425
configs:
- config_name: default
data_files:
- split: train
path: data/train-*