metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 234340261728
num_examples: 91253996
- name: test
num_bytes: 12333698256
num_examples: 4802842
download_size: 5208492851
dataset_size: 246673959984
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*