metadata
dataset_info:
features:
- name: tokens
sequence: int64
splits:
- name: train
num_bytes: 8048933352
num_examples: 1000000
download_size: 1960940460
dataset_size: 8048933352
configs:
- config_name: default
data_files:
- split: train
path: data/train-*