metadata
dataset_info:
features:
- name: image
dtype: image
- name: question_categories
dtype: string
- name: question
dtype: string
- name: answer
sequence: string
splits:
- name: test
num_bytes: 22938510
num_examples: 500
download_size: 22401358
dataset_size: 22938510
configs:
- config_name: default
data_files:
- split: test
path: data/test-*