Most diffusion-based models use a fixed (model-tuned) guidance schedule. We show that picking the guidance value during inference, conditioned on the prompt/capability, significantly improves performance.
arxiv.org/abs/2509.16131
Aida Nematzadeh
@aidanematzadeh.bsky.social
181 followers 35 following 10 posts
Research scientist at Google DeepMind.🦎 She/her. www.aidanematzadeh.me
