metadata
dataset_info:
features:
- name: image
dtype: image
- name: caption
dtype: string
splits:
- name: train
num_bytes: 4463593122.6
num_examples: 27000
- name: test
num_bytes: 487117631.4
num_examples: 3000
download_size: 4886531958
dataset_size: 4950710754
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*