metadata
dataset_info:
features:
- name: images
list: image
- name: problem
dtype: string
- name: answer
dtype: string
- name: id
dtype: int64
splits:
- name: train
num_bytes: 2936388866.25
num_examples: 38870
download_size: 2896803156
dataset_size: 2936388866.25
configs:
- config_name: default
data_files:
- split: train
path: data/train-*