metadata
dataset_info:
features:
- name: filename
dtype: string
- name: image
dtype: image
- name: question
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 1496767601
num_examples: 108782
download_size: 1483606219
dataset_size: 1496767601