metadata
dataset_info:
features:
- name: context
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
- name: answer_start
dtype: int64
splits:
- name: train
num_bytes: 255316160
num_examples: 1366625
download_size: 76395903
dataset_size: 255316160
configs:
- config_name: default
data_files:
- split: train
path: data/train-*