metadata
dataset_info:
features:
- name: conditioning_image
dtype: image
- name: mask_image
dtype: image
- name: caption_wt_parts
dtype: string
splits:
- name: train
num_bytes: 48487
num_examples: 2
download_size: 23456
dataset_size: 48487
configs:
- config_name: default
data_files:
- split: train
path: data/train-*