metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: token_type_ids
sequence: int8
- name: attention_mask
sequence: int8
- name: word_ids
sequence: int64
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 42113819652
num_examples: 14830301
- name: validation
num_bytes: 4836576832
num_examples: 1703172
download_size: 21646576639
dataset_size: 46950396484
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*