metadata
dataset_info:
features:
- name: answer
dtype: string
- name: images
list: image
- name: problem
dtype: string
splits:
- name: train
num_bytes: 2014756365.25
num_examples: 19654
- name: validation
num_bytes: 602640992.591
num_examples: 6259
- name: test
num_bytes: 671378946.919
num_examples: 6719
download_size: 527002225
dataset_size: 3288776304.76
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*