metadata
dataset_info:
features:
- name: id
dtype: string
- name: context
dtype: string
- name: question
dtype: string
- name: is_impossible
dtype: bool
- name: answers
struct:
- name: text
list: string
- name: answer_start
list: int64
- name: context_id
dtype: string
splits:
- name: train
num_bytes: 51471186
num_examples: 64439
- name: validation
num_bytes: 6398382
num_examples: 8012
- name: test
num_bytes: 6461559
num_examples: 8109
- name: gold_test
num_bytes: 153028
num_examples: 200
download_size: 12936397
dataset_size: 64484155
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
- split: gold_test
path: data/gold_test-*