metadata
dataset_info:
features:
- name: subset
dtype: string
- name: example_id
dtype: string
- name: images
list: string
- name: mc_question
list: string
- name: oe_question
list: string
- name: direct_answer
list: string
- name: choices
list:
list: string
- name: correct_choice_idx
list: int64
splits:
- name: train
num_bytes: 11610242
num_examples: 3358
- name: validation
num_bytes: 1786999
num_examples: 512
download_size: 5649527
dataset_size: 13397241
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*