metadata
dataset_info:
features:
- name: image
dtype: image
- name: question
dtype: string
- name: answers_text
dtype: string
- name: type
dtype: string
- name: subject
dtype: string
- name: relation
dtype: string
- name: image_path
dtype: string
splits:
- name: train
num_bytes: 2694878.3
num_examples: 60
- name: test
num_bytes: 51202687.7
num_examples: 1140
download_size: 45521286
dataset_size: 53897566
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*