metadata
dataset_info:
features:
- name: tokens
sequence: int64
splits:
- name: train
num_bytes: 9676052584
num_examples: 20000000
- name: validation
num_bytes: 1836463376
num_examples: 3796010
download_size: 1026466555
dataset_size: 11512515960
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*