metadata
dataset_info:
features:
- name: Content
dtype: string
splits:
- name: train
num_bytes: 669391
num_examples: 559
download_size: 239418
dataset_size: 669391
configs:
- config_name: default
data_files:
- split: train
path: data/train-*