SD3.5 and Flux datasets
Datasets used in the paper "Training-Free Reconstruction-Based AI-Generated Image Detectors Are Inherently Vulnerable to Adversarial Examples" Contains 1000 images generated by SD3.5 and Flux-schnell models using the prompts from the paper "Synthbuster: Towards Detection of Diffusion Model Generated Images". The prompts can be downloaded using the official link: https://doi.org/10.5281/zenodo.10066460
Roman Demchenko
· Zenodo (CERN European Organi... · 0 citations