---
license: apache-2.0
dataset_info:
  features:
  - name: image
    dtype: image
  splits:
  - name: train
    num_bytes: 2172348871.562
    num_examples: 162971
  download_size: 2166224426
  dataset_size: 2172348871.562
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---



## This repository contains Unofficial access for SDIP-horses dataset


[Official Repository](https://github.com/self-distilled-stylegan/self-distilled-internet-photos), [Project Page](https://self-distilled-stylegan.github.io/), [Paper](https://arxiv.org/abs/2202.12211)


Self-Distilled Internet Photos (SDIP) is a multi-domain image dataset. The dataset consists of Self-Distilled Flickr (SD-Flickr) and *Self-Distilled LSUN (SD-LSUN) that were crawled from [Flickr](https://www.flickr.com/) and [LSUN dataset](https://www.yf.io/p/lsun), respectively, and then curated using the method described in our Self-Distilled StyleGAN paper:

> **Self-Distilled StyleGAN: Towards Generation from Internet Photos**<br>
> Ron Mokady, Michal Yarom, Omer Tov, Oran Lang, Daniel Cohen-Or, Tali Dekel, Michal Irani, Inbar Mosseri


![](SDIP.png)

## Overview

[StyleGAN’s](https://github.com/NVlabs/stylegan2-ada-pytorch) fascinating generative and editing abilities are limited to structurally aligned and well-curated datasets. It does not work well on raw datasets downloaded from the Internet. The SDIP domains presented here, which are StyleGAN-friendly, were automatically curated by our [method](https://arxiv.org/abs/2202.12211) from raw images collected from the Internet. The raw uncurated images in *Self-Distilled Flicker (SD-Flickr)* were first crawled from [Flickr](https://www.flickr.com/) using a simple keyword (e.g. 'dog' or 'elephant').

The dataset in this page exhibits 4 domains: **SD-Dogs (126K images) 1024X1024**, **SD-Elephants (39K images) 512X512**, **SD-Bicycles (96K images) 256X256**, and **SD-Horses (162K images) 256X256**. Our curation process consists of a simple pre-processing step (off-the-shelf object detector to crop the main object and then rescale), followed by a sophisticated StyleGAN-friendly filtering step (which removes outlier images while maintaining dataset diversity). This results in a more coherent and clean dataset, which is suitable for training a StyleGAN2 generator (see more details in our [paper](https://arxiv.org/abs/2202.12211)).

The data itself is saved in a json format: for SD-Flickr we provide urls of the original images and bounding boxes used for cropping; for SD-LSUN we provide image identifiers with the bounding boxes. In addition to the SDIP dataset, we also provide weights of pre-trained StyleGAN2 models trained using each image domain presented in the paper.