metadata
dataset_info:
features:
- name: question
dtype: string
- name: image
dtype: image
- name: answer
dtype: string
splits:
- name: train
num_bytes: 7126455711.4
num_examples: 20800
download_size: 1231022525
dataset_size: 7126455711.4
configs:
- config_name: default
data_files:
- split: train
path: data/train-*