mit-han-lab/data-efficient-gans

[NeurIPS 2020] Differentiable Augmentation for Data-Efficient GAN Training

48
/ 100
Emerging

This project helps image creators or researchers generate realistic images even when they have very few training examples. You provide a small collection of images (e.g., just 100 pictures of pandas or specific landmarks), and it generates new, high-quality images that look like they belong to that collection. It's ideal for anyone in creative fields, scientific imaging, or research who needs to expand a limited visual dataset.

1,313 stars. No commits in the last 6 months.

Use this if you need to create diverse, high-fidelity images from a very small existing dataset, without needing extensive pre-training.

Not ideal if you already have a very large dataset for image generation or if your primary goal is not image synthesis from limited data.

generative-art synthetic-data image-generation computer-vision-research low-data-imaging
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 22 / 25

How are scores calculated?

Stars

1,313

Forks

176

Language

Python

License

BSD-2-Clause

Last pushed

Sep 24, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/mit-han-lab/data-efficient-gans"

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