metadata
dataset_info:
features:
- name: input_ids
list:
list: int32
- name: labels
list:
list: int64
- name: num_mask_tokens
dtype: int64
- name: original
list:
list: int64
splits:
- name: train
num_bytes: 4842788940
num_examples: 1014020
- name: test
num_bytes: 48227796
num_examples: 10243
download_size: 861915692
dataset_size: 4891016736
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*