metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 27338511984
num_examples: 10645838
- name: validation
num_bytes: 2983605120
num_examples: 1161840
download_size: 7088975130
dataset_size: 30322117104