metadata
dataset_info:
features:
- name: image_path
dtype: string
- name: answer
dtype: string
- name: question
dtype: string
- name: image
dtype: image
splits:
- name: train
num_bytes: 135755294
num_examples: 810
download_size: 71148130
dataset_size: 135755294
configs:
- config_name: default
data_files:
- split: train
path: data/train-*