metadata
dataset_info:
features:
- name: images
dtype: image
- name: questions
dtype: string
- name: answers
sequence: string
- name: source
dtype: string
splits:
- name: train
num_bytes: 163423360
num_examples: 1000
download_size: 161733967
dataset_size: 163423360
configs:
- config_name: default
data_files:
- split: train
path: data/train-*