metadata
dataset_info:
features:
- name: URL
dtype: string
- name: conditioning
dtype: image
- name: caption
dtype: string
- name: hed
dtype: image
- name: thinning_hed
dtype: image
splits:
- name: train
num_bytes: 2985432907.163
num_examples: 17593
download_size: 3038713172
dataset_size: 2985432907.163
configs:
- config_name: default
data_files:
- split: train
path: data/train-*