metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: input_ids
sequence: int32
splits:
- name: train
num_bytes: 91738479900
num_examples: 22375239
download_size: 13547146690
dataset_size: 91738479900