rgeirhos/Stylized-ImageNet

Code to create Stylized-ImageNet, a stylized version of standard ImageNet (ICLR 2019 Oral)

55
/ 100
Established

This project helps machine learning researchers generate a unique version of the ImageNet dataset where images are 'stylized' like paintings while preserving their original object shapes. By feeding in original ImageNet images and a collection of style images, it outputs a new Stylized-ImageNet dataset. This modified dataset is used by researchers to train Convolutional Neural Networks (CNNs) to focus more on an object's shape rather than its texture.

524 stars.

Use this if you are a machine learning researcher studying computer vision models and want to create a specialized ImageNet dataset to investigate how CNNs learn object shapes versus textures.

Not ideal if you want to stylize a dataset other than ImageNet, as this tool is specifically configured for ImageNet.

computer-vision-research dataset-generation machine-learning-training neural-network-robustness image-stylization
No Package No Dependents
Maintenance 10 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 19 / 25

How are scores calculated?

Stars

524

Forks

65

Language

Python

License

MIT

Last pushed

Feb 16, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/rgeirhos/Stylized-ImageNet"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.