metadata
dataset_info:
features:
- name: Sequence
dtype: string
splits:
- name: train
num_bytes: 80382214
num_examples: 147508
download_size: 79888144
dataset_size: 80382214
configs:
- config_name: default
data_files:
- split: train
path: data/train-*