metadata
dataset_info:
features:
- name: candidate_id
dtype: string
- name: caption
dtype: string
- name: group
sequence: string
- name: split
dtype: string
- name: target_id
dtype: string
- name: target
dtype: image
- name: candidate
dtype: image
splits:
- name: val
num_bytes: 902818814.275
num_examples: 4181
download_size: 608604226
dataset_size: 902818814.275
configs:
- config_name: default
data_files:
- split: val
path: data/val-*