metadata
dataset_info:
features:
- name: tokens
sequence: int64
- name: mask
sequence: bool
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 165182051
num_examples: 51760
download_size: 35368437
dataset_size: 165182051
configs:
- config_name: default
data_files:
- split: train
path: data/train-*