metadata
dataset_info:
features:
- name: id
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
- name: image
dtype: image
splits:
- name: train
num_bytes: 2010676997.122
num_examples: 12486
- name: val
num_bytes: 497514571.052
num_examples: 3122
download_size: 2487935699
dataset_size: 2508191568.174
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*