milenagazdieva/LightUnbalancedOptimalTransport

PyTorch implementation of "Light Unbalanced Optimal Transport" (NeurIPS 2024)

30
/ 100
Emerging

This project helps machine learning researchers and practitioners perform image-to-image translation when the source and target image collections have different numbers of items or contain irrelevant examples. It takes two sets of images, such as 'adult faces' and 'young faces,' and efficiently finds the optimal way to transform images from one set to match the style or characteristics of the other, even if categories like gender are imbalanced. The output is a mapping that allows you to translate an image from the source to the target domain while handling discrepancies in the input data.

No commits in the last 6 months.

Use this if you need to translate images from one domain to another and your datasets are unbalanced (e.g., uneven class representation) or contain outliers that traditional methods struggle with.

Not ideal if your image datasets are perfectly balanced and free of outliers, as simpler optimal transport solvers might suffice.

image-to-image-translation unbalanced-data outlier-robustness generative-models computer-vision
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 6 / 25
Maturity 16 / 25
Community 8 / 25

How are scores calculated?

Stars

22

Forks

2

Language

Jupyter Notebook

License

MIT

Last pushed

Dec 23, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/milenagazdieva/LightUnbalancedOptimalTransport"

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