metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: image
dtype: image
- name: conditioning_image
dtype: image
- name: text
dtype: string
- name: params
struct:
- name: downsample
dtype: int64
- name: grid_size
dtype: int64
- name: high_threshold
dtype: int64
- name: low_threshold
dtype: int64
- name: sigma
dtype: float64
splits:
- name: train
num_bytes: 205989049941
num_examples: 500000
download_size: 84179071741
dataset_size: 205989049941