metadata
dataset_info:
features:
- name: image
dtype: image
- name: caption_basic
dtype: string
- name: caption_artsy
dtype: string
- name: caption_wt_parts
dtype: string
- name: conditioning_image
dtype: image
- name: mask_image
dtype: image
splits:
- name: train
num_bytes: 898016172.78
num_examples: 22885
download_size: 484949965
dataset_size: 898016172.78
configs:
- config_name: default
data_files:
- split: train
path: data/train-*