metadata
dataset_info:
features:
- name: question
dtype: string
- name: answer
dtype: string
- name: question_type
dtype: string
- name: visual_indices
sequence: int64
- name: images
sequence: image
splits:
- name: train
num_bytes: 91356720
num_examples: 400
download_size: 77787520
dataset_size: 91356720
configs:
- config_name: default
data_files:
- split: train
path: data/train-*