metadata
dataset_info:
features:
- name: image
dtype: binary
- name: captions
dtype: string
splits:
- name: train
num_bytes: 857772
num_examples: 20
download_size: 857547
dataset_size: 857772
configs:
- config_name: default
data_files:
- split: train
path: data/train-*