metadata
dataset_info:
features:
- name: ids
dtype: string
- name: image_names
dtype: string
- name: images
dtype: image
- name: questions
dtype: string
- name: answers
dtype: string
splits:
- name: train
num_bytes: 44386448
num_examples: 635
- name: test
num_bytes: 10388111
num_examples: 159
download_size: 17127914
dataset_size: 54774559
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*