metadata
dataset_info:
features:
- name: context
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 1700022862
num_examples: 4500000
download_size: 523999854
dataset_size: 1700022862
configs:
- config_name: default
data_files:
- split: train
path: data/train-*