metadata
dataset_info:
features:
- name: source_images
sequence: image
- name: source_captions
sequence: string
- name: prompt
dtype: string
- name: target_image
dtype: image
- name: target_caption
dtype: string
splits:
- name: train
num_bytes: 168173014864
num_examples: 360036
download_size: 90350985520
dataset_size: 168173014864
configs:
- config_name: default
data_files:
- split: train
path: data/train-*