metadata
dataset_info:
features:
- name: qid
dtype: string
- name: question
dtype: string
- name: images
sequence: image
- name: answer
dtype: string
- name: question_info
dtype: string
- name: type
dtype: string
- name: choices
sequence: string
splits:
- name: train
num_bytes: 18622762
num_examples: 376
download_size: 15587200
dataset_size: 18622762
configs:
- config_name: default
data_files:
- split: train
path: data/train-*