metadata
language:
- en
dataset_info:
features:
- name: id
dtype: string
- name: title
dtype: string
- name: context
dtype: string
- name: question
dtype: string
- name: answers
sequence:
- name: text
dtype: string
- name: answer_start
dtype: int32
splits:
- name: train
num_bytes: 69824466.34546056
num_examples: 77087
- name: valid
num_bytes: 9521641.654539436
num_examples: 10512
- name: test
num_bytes: 10472984
num_examples: 10570
download_size: 52413878
dataset_size: 89819092
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*
- split: test
path: data/test-*