metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: special_tokens_mask
sequence: int8
splits:
- name: train
num_bytes: 16995635700
num_examples: 1381759
download_size: 4366533537
dataset_size: 16995635700
configs:
- config_name: default
data_files:
- split: train
path: data/train-*