metadata
dataset_info:
features:
- name: text
dtype: string
- name: mask
sequence: int64
splits:
- name: train
num_bytes: 406123
num_examples: 5000
download_size: 152455
dataset_size: 406123
configs:
- config_name: default
data_files:
- split: train
path: data/train-*