metadata
dataset_info:
features:
- name: sentence
dtype: string
- name: language
dtype: string
splits:
- name: train
num_bytes: 17402218
num_examples: 100000
download_size: 9326084
dataset_size: 17402218
configs:
- config_name: default
data_files:
- split: train
path: data/train-*