metadata
dataset_info:
features:
- name: image
dtype: image
- name: question
dtype: string
- name: answer
dtype: string
- name: id
dtype: int64
splits:
- name: train
num_bytes: 430646390
num_examples: 635
- name: test
num_bytes: 173859026
num_examples: 250
download_size: 589640106
dataset_size: 604505416
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*