metadata
dataset_info:
features:
- name: image
dtype: image
- name: label
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 582033134
num_examples: 550000
download_size: 75829966
dataset_size: 582033134
configs:
- config_name: default
data_files:
- split: train
path: data/train-*