metadata
dataset_info:
features:
- name: img
dtype: image
- name: en_cap
dtype: string
- name: hu_cap
dtype: string
splits:
- name: train
num_bytes: 4914324133
num_examples: 500000
download_size: 4649748345
dataset_size: 4914324133
configs:
- config_name: default
data_files:
- split: train
path: data/train-*