metadata
dataset_info:
features:
- name: image_path
dtype: string
- name: caption
dtype: string
- name: image
dtype: image
splits:
- name: train
num_bytes: 1478220981.428
num_examples: 28004
- name: validation
num_bytes: 201756083.828
num_examples: 4036
download_size: 1780153815
dataset_size: 1679977065.256
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*