metadata
dataset_info:
features:
- name: translation
struct:
- name: en
dtype: string
- name: hi
dtype: string
- name: text
dtype: string
splits:
- name: train
num_bytes: 1018603574
num_examples: 1659083
download_size: 506831511
dataset_size: 1018603574
configs:
- config_name: default
data_files:
- split: train
path: data/train-*