metadata
dataset_info:
features:
- name: question
dtype: string
- name: cropped_image
dtype: image
- name: answer
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 408334081.2011177
num_examples: 26041
download_size: 408093962
dataset_size: 408334081.2011177
configs:
- config_name: default
data_files:
- split: train
path: data/train-*