metadata
dataset_info:
features:
- name: answer
dtype: string
- name: images
sequence: image
- name: problem
dtype: string
splits:
- name: train
num_bytes: 63067838
num_examples: 2000
- name: test
num_bytes: 51406891
num_examples: 400
download_size: 86303813
dataset_size: 114474729
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*