The official implementation of the paper "To Generate or Not? Safety-Driven Unlearned Diffusion Models Are Still Easy To Generate Unsafe Images ... For Now". This work introduces one fast and efficient attack methods to generate toxic content for safety-driven diffusion models.