metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: token_type_ids
sequence: int8
- name: attention_mask
sequence: int8
- name: special_tokens_mask
sequence: int8
splits:
- name: train
num_bytes: 13579352160
num_examples: 56580634
- name: validation
num_bytes: 145881600
num_examples: 607840
download_size: 5561650628
dataset_size: 13725233760
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*