metadata
dataset_info:
features:
- name: image
dtype: image
- name: question
dtype: string
- name: answer
dtype: string
- name: position
dtype: string
splits:
- name: train
num_bytes: 3968377621
num_examples: 199400
- name: validation
num_bytes: 2081646
num_examples: 100
- name: test
num_bytes: 9776592
num_examples: 500
download_size: 3416932388
dataset_size: 3980235859
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*