LADD: latent generative features for adversarial diffusion distillation

What changed

Replaced ADD's pixel-space fixed visual discriminator features with features from pretrained latent diffusion models, simplifying high-resolution/multi-aspect distillation.

Scientific scope

Replaced ADD's pixel-space fixed visual discriminator features with features from pretrained latent diffusion models, simplifying high-resolution/multi-aspect distillation. Inherits ADD and latent diffusion. Delta is the latent-feature distillation mechanism, not first adversarial distillation or a separate invention for each SD3-Turbo checkpoint. Comparisons depend on original metrics and sampling budgets.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits ADD and latent diffusion. Delta is the latent-feature distillation mechanism, not first adversarial distillation or a separate invention for each SD3-Turbo checkpoint. Comparisons depend on original metrics and sampling budgets.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• Fast High-Resolution Image Synthesis with Latent Adversarial Diffusion Distillation

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction