metadata
dataset_info:
features:
- name: original_image
dtype: image
- name: edit_prompt
dtype: string
- name: japanese_image
dtype: image
- name: mask
dtype: image
splits:
- name: train
num_bytes: 21278821
num_examples: 28
download_size: 21277851
dataset_size: 21278821
configs:
- config_name: default
data_files:
- split: train
path: data/train-*