metadata
dataset_info:
features:
- name: answer
dtype: string
- name: image
dtype: image
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 1101443348.656
num_examples: 7748
- name: test
num_bytes: 89786903
num_examples: 567
download_size: 3546373791
dataset_size: 1191230251.656
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*