metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: special_tokens_mask
sequence: int8
splits:
- name: train
num_bytes: 27157416224
num_examples: 5295908
download_size: 10137252952
dataset_size: 27157416224
configs:
- config_name: default
data_files:
- split: train
path: data/train-*