metadata
dataset_info:
config_name: corpus
features:
- name: doc_id
dtype: string
- name: passage
dtype: string
- name: metadata
struct:
- name: last_modified_datetime
dtype: timestamp[us]
splits:
- name: train
num_bytes: 3472453
num_examples: 3523
download_size: 2002668
dataset_size: 3472453
configs:
- config_name: corpus
data_files:
- split: train
path: corpus/train-*