metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 55751845968
num_examples: 5440266
- name: test
num_bytes: 6194659800
num_examples: 604475
download_size: 23930738929
dataset_size: 61946505768
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*