AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Adversarial Diffusion Distillation

Axel Sauer, Dominik Lorenz, Andreas Blattmann, Robin Rombach

ECCV 2024 · Image generation › Fast diffusion sampling · approach: Distillation for few-step generation · 633 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add image conditioning to the discriminator (in addition to or instead of text) (ablation)ADD-M discriminator · COCO2017 5k (FID, CLIP score)FID / CLIP scoreImproved
use text conditioning on the discriminator (ablation)ADD-M discriminator · COCO2017 5kFID / CLIP scoreImproved
use DINOv2 ViT as discriminator feature network (ablation)ADD-M · COCO2017 5kFID / CLIP scoreImproved
initialize the student from a pretrained diffusion model instead of from scratch (ablation)ADD-student · COCO2017 5kFID / CLIP scoreImproved
add adversarial loss to the distillation loss (ablation)ADD · COCO2017 5kFID / CLIP scoreImproved
use SDS or NFSD distillation weighting instead of exponential weighting (ablation)ADD distillation loss · COCO2017 5kFID / CLIP scoreMixed
use a bigger student and teacher (SDXL instead of SD) (ablation)ADD · COCO2017 5kFID / CLIP scoreNo clear effect
take several consecutive steps with the teacher (ablation)ADD distillation · COCO2017 5kFID / CLIP scoreNo clear effect
distill SDXL with ADD and sample in four steps instead of 50-step SDXLADD-XL · PartiPrompts user study (100 prompts)human preference / ELOImproved
single-step ADD-XL versus four-step LCM-XLADD-XL · PartiPrompts user studyhuman preference / ELOImproved
distill into student with adversarial lossADD-XL versus teacher SDXL · sample diversitydiversityHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy