metadata
dataset_info:
features:
- name: image
struct:
- name: bytes
dtype: binary
- name: path
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 48811901
num_examples: 1400
download_size: 42366000
dataset_size: 48811901
configs:
- config_name: default
data_files:
- split: train
path: data/train-*