metadata
dataset_info:
features:
- name: title
dtype: string
- name: context
dtype: string
- name: id
dtype: string
- name: question
dtype: string
- name: is_impossible
dtype: bool
- name: answers
struct:
- name: answer_start
list: int64
- name: text
list: string
- name: is_bad
dtype: bool
- name: bad_reason
dtype: string
splits:
- name: train
num_bytes: 35707844
num_examples: 18567
- name: test
num_bytes: 2540540.916
num_examples: 999
download_size: 4644637
dataset_size: 38248384.916
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*