metadata
dataset_info:
features:
- name: image
dtype: image
- name: question
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 5962957671
num_examples: 6600
- name: validation
num_bytes: 1822918259
num_examples: 2200
- name: test
num_bytes: 1783703292
num_examples: 2200
download_size: 9264858771
dataset_size: 9569579222
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*