metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 11586858926
num_examples: 52722772
- name: test
num_bytes: 4963957200
num_examples: 22595475
download_size: 7388440953
dataset_size: 16550816126
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*