Reproduce “Exploring the latent space of diffusion models directly through singular value decomposition” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Li Wang, Boyan Gao, Yanran Li, Zhao Wang, Xiaosong Yang, David A. Clifton, Jun Xiao
AVI edits diffusion-generated images by decomposing prompt-guided latent codes with singular value decomposition and recombining components from different denoising steps, using an MLP to predict weights while returning to the original prompt. For the Male attribute in Stable Diffusion 2.1, it reports FID 10.08, CLIP 25.4, and LPIPS 0.14, with manual prompt-pair selection and seed sensitivity.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.