metadata
dataset_info:
features:
- name: qid
dtype: string
- name: question
dtype: string
- name: choices
dtype: image
- name: images
sequence: image
- name: answer
dtype: string
- name: transformations
dtype: string
- name: difficulty_level
dtype: string
- name: question_info
dtype: string
- name: answer_info
dtype: string
splits:
- name: train
num_bytes: 2444658336
num_examples: 816
download_size: 1541978360
dataset_size: 2444658336
configs:
- config_name: default
data_files:
- split: train
path: data/train-*