metadata
dataset_info:
features:
- name: token_ids
list: int64
- name: token_ids_length
dtype: int64
splits:
- name: train
num_bytes: 11219883084
num_examples: 2940889
- name: valid
num_bytes: 383951252
num_examples: 99999
download_size: 11590380288
dataset_size: 11603834336
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*