metadata
dataset_info:
features:
- name: context
dtype: string
- name: question
dtype: string
- name: is_impossible
dtype: bool
- name: answers
struct:
- name: answer_start
list: int64
- name: text
list: string
- name: id
dtype: string
splits:
- name: train
num_bytes: 272331281
num_examples: 145455
- name: validation
num_bytes: 21260217
num_examples: 11370
download_size: 138874378
dataset_size: 293591498
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*