metadata
dataset_info:
features:
- name: translation
struct:
- name: en
dtype: string
- name: mai
dtype: string
- name: new
dtype: string
splits:
- name: train
num_bytes: 700903
num_examples: 2494
download_size: 323847
dataset_size: 700903
configs:
- config_name: default
data_files:
- split: train
path: data/train-*