metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 65716109592
num_examples: 6412579
- name: test
num_bytes: 7301792232
num_examples: 712509
download_size: 28765116415
dataset_size: 73017901824
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*