metadata
dataset_info:
features:
- name: candidate
dtype: image
- name: candidate_id
dtype: string
- name: caption
sequence: string
- name: category
dtype: string
- name: split
dtype: string
- name: target
dtype: image
- name: target_id
dtype: string
splits:
- name: val
num_bytes: 158557094
num_examples: 6016
download_size: 156940312
dataset_size: 158557094
configs:
- config_name: default
data_files:
- split: val
path: data/val-*