metadata
dataset_info:
features:
- name: image
dtype: image
- name: image_id
dtype: string
- name: caption
dtype: string
- name: cui
sequence: string
splits:
- name: train
num_bytes: 188134485
num_examples: 850
download_size: 188061002
dataset_size: 188134485
configs:
- config_name: default
data_files:
- split: train
path: data/train-*