metadata
dataset_info:
features:
- name: subject
dtype: string
- name: image
dtype: image
- name: question
dtype: string
- name: ground_truth
dtype: string
splits:
- name: train
num_bytes: 6002202
num_examples: 225
download_size: 4725273
dataset_size: 6002202
configs:
- config_name: default
data_files:
- split: train
path: data/train-*