metadata
dataset_info:
features:
- name: image
dtype: image
- name: question
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 423714953411.192
num_examples: 110344
- name: test
num_bytes: 52200451417.162
num_examples: 13794
- name: val
num_bytes: 51651694517.614
num_examples: 13793
download_size: 520842108548
dataset_size: 527567099345.968
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: val
path: data/val-*