metadata
dataset_info:
features:
- name: question_id
dtype: string
- name: image
dtype: image
- name: question
dtype: string
- name: answer_letter
dtype: string
splits:
- name: test
num_bytes: 189610594
num_examples: 3640
download_size: 175217102
dataset_size: 189610594
configs:
- config_name: default
data_files:
- split: test
path: data/test-*