metadata
dataset_info:
features:
- name: caption
dtype: string
- name: url
dtype: string
- name: image
dtype: image
- name: image_bw
dtype: image
splits:
- name: train
num_bytes: 2036324403
num_examples: 10000
download_size: 425648735
dataset_size: 2036324403
configs:
- config_name: default
data_files:
- split: train
path: data/train-*