metadata
dataset_info:
features:
- name: custom_id
dtype: string
- name: author_info
struct:
- name: email_address
dtype: string
- name: question_categories
sequence: string
- name: image_attribution
dtype: string
- name: question
dtype: string
- name: question_images
sequence: image
- name: final_answer
dtype: string
- name: rationale_text
dtype: string
- name: rationale_images
sequence: image
splits:
- name: train
num_bytes: 33467517
num_examples: 87
- name: test
num_bytes: 78533251
num_examples: 203
download_size: 38575645117
dataset_size: 112000768
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*