metadata
dataset_info:
features:
- name: prefix
dtype: string
- name: suffix
dtype: string
- name: reference
dtype: string
splits:
- name: train
num_bytes: 22352
num_examples: 50
download_size: 13700
dataset_size: 22352
configs:
- config_name: default
data_files:
- split: train
path: data/train-*