SliderSpace uses PCA on CLIP embeddings of a diffusion model's own samples, then trains low-rank adapters for each principal component, turning them into composable image control sliders.
Ledits++: Limitless im- age editing using text-to-image models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
SliderSpace: Decomposing the Visual Capabilities of Diffusion Models
SliderSpace uses PCA on CLIP embeddings of a diffusion model's own samples, then trains low-rank adapters for each principal component, turning them into composable image control sliders.