metadata
dataset_info:
features:
- name: text
dtype: string
- name: input_ids
sequence: int32
splits:
- name: train
num_bytes: 3388323834
num_examples: 25600000
download_size: 2430131280
dataset_size: 3388323834
configs:
- config_name: default
data_files:
- split: train
path: data/train-*