metadata
dataset_info:
features:
- name: passage
dtype: string
- name: index
dtype: int64
splits:
- name: train
num_bytes: 12681688023
num_examples: 27244679
download_size: 6925698199
dataset_size: 12681688023
configs:
- config_name: default
data_files:
- split: train
path: data/train-*