metadata
dataset_info:
features:
- name: tokens
sequence: int64
- name: tokens_reversed
sequence: int64
splits:
- name: train
num_bytes: 2959962112
num_examples: 9999872
download_size: 192990622
dataset_size: 2959962112
configs:
- config_name: default
data_files:
- split: train
path: data/train-*