metadata
dataset_info:
features:
- name: data
struct:
- name: paragraphs
list:
- name: context
dtype: string
- name: document_id
dtype: int64
- name: qas
list:
- name: answers
list:
- name: answer_category
dtype: 'null'
- name: answer_end
dtype: int64
- name: answer_id
dtype: int64
- name: answer_start
dtype: int64
- name: document_id
dtype: int64
- name: question_id
dtype: int64
- name: text
dtype: string
- name: id
dtype: int64
- name: is_impossible
dtype: bool
- name: question
dtype: string
splits:
- name: train
num_bytes: 41377188
num_examples: 9835
- name: test
num_bytes: 9205752
num_examples: 2108
download_size: 25244681
dataset_size: 50582940
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*