metadata
dataset_info:
features:
- name: input_ids
sequence: int64
splits:
- name: train
num_bytes: 96959162196
num_examples: 41821747
download_size: 3744535420
dataset_size: 96959162196
configs:
- config_name: default
data_files:
- split: train
path: data/train-*