LukasStruppek/Rickrolling-the-Artist
[ICCV 2023] Source code for our paper "Rickrolling the Artist: Injecting Invisible Backdoors into Text-Guided Image Generation Models".
This project helps evaluate and secure text-guided image generation models like Stable Diffusion. It allows you to inject hidden commands into the model's text understanding component. The input is a pre-trained text encoder, and the output is a modified encoder that responds to secret triggers, enabling control over generated images or concept removal. This tool is for AI security researchers, model developers, and anyone concerned with the integrity of generative AI.
No commits in the last 6 months.
Use this if you need to test the security of text-to-image models against hidden tampering or want to remove undesired concepts from their generation capabilities.
Not ideal if you are looking for a simple, non-technical tool to generate images or modify existing images without understanding AI model architecture.
Stars
65
Forks
7
Language
Python
License
MIT
Category
Last pushed
Nov 20, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/LukasStruppek/Rickrolling-the-Artist"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.