metadata
dataset_info:
features:
- name: image1
dtype: image
- name: image2
dtype: image
- name: caption1
dtype: string
- name: caption2
dtype: string
splits:
- name: train
num_bytes: 4512766259
num_examples: 9928
download_size: 4512816658
dataset_size: 4512766259
configs:
- config_name: default
data_files:
- split: train
path: data/train-*