nunchaku-ai/nunchaku

[ICLR2025 Spotlight] SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models

54
/ 100
Established

This project helps AI practitioners and researchers run powerful diffusion models, like those for generating or editing images, much more efficiently. It takes existing 4-bit neural network models and optimizes them to run faster and use less memory on various GPUs. The result is quicker image generation, editing, or other diffusion model tasks, making advanced AI capabilities more accessible.

3,724 stars.

Use this if you are working with large diffusion models and need to accelerate their performance or reduce their memory footprint for faster image generation and editing tasks.

Not ideal if you are working with AI models other than diffusion models, or if you do not require specialized optimizations for 4-bit neural networks.

AI image generation diffusion models AI model deployment image editing AI AI research
No Package No Dependents
Maintenance 10 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 18 / 25

How are scores calculated?

Stars

3,724

Forks

229

Language

Python

License

Apache-2.0

Last pushed

Mar 07, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/nunchaku-ai/nunchaku"

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