metadata
dataset_info:
features:
- name: task
sequence: string
- name: image_path
sequence: string
- name: input_image
sequence: image
- name: edited_image
sequence: image
- name: instruction
sequence: string
- name: caption
sequence: string
- name: mask
sequence: image
splits:
- name: train
num_bytes: 299799054
num_examples: 252
download_size: 298693980
dataset_size: 299799054
configs:
- config_name: default
data_files:
- split: train
path: data/train-*