metadata
dataset_info:
features:
- name: data_id
dtype: int64
- name: image
dtype: image
- name: image_description
dtype: string
- name: language
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
- name: original_category
dtype: string
- name: source
dtype: string
- name: atomic_question
dtype: string
- name: atomic_fact
dtype: string
- name: vqa_category
dtype: string
splits:
- name: test
num_bytes: 831610643
num_examples: 1013
download_size: 831179869
dataset_size: 831610643
configs:
- config_name: default
data_files:
- split: test
path: data/test-*