NikosEfth/crafting-shifts

Official PyTorch implementation of the WACV 2025 Oral paper "Crafting Distribution Shifts for Validation and Training in Single Source Domain Generalization".

24
/ 100
Experimental

When developing AI models that work with images, it's common for models to perform well in a controlled environment but fail in the real world when image styles or conditions change. This project helps AI researchers and machine learning engineers evaluate how robust their image recognition models are to these real-world variations. It takes an existing dataset of images and helps you create new, 'shifted' versions of that data to better test and train your models, helping you build more reliable AI.

No commits in the last 6 months.

Use this if you need to thoroughly test and improve the generalization of your image classification models across various unseen visual conditions from a single source domain.

Not ideal if your AI model doesn't involve image data or if you are not working on problems where the visual style of inputs can drastically change.

AI model generalization image classification computer vision research machine learning robustness domain adaptation
Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 6 / 25
Maturity 16 / 25
Community 0 / 25

How are scores calculated?

Stars

23

Forks

Language

Python

License

Apache-2.0

Last pushed

Aug 31, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/NikosEfth/crafting-shifts"

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