metadata
dataset_info:
features:
- name: image
dtype: image
- name: qa_pairs
dtype: string
- name: lang
dtype: string
- name: id
dtype: int64
splits:
- name: train
num_bytes: 3078386824.832
num_examples: 6678
- name: test
num_bytes: 1052448167.396
num_examples: 2116
download_size: 4239702402
dataset_size: 4130834992.2279997
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*