AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

"BK-SDM: A Lightweight, Fast, and Cheap Version of Stable Diffusion"

Bo-Kyeong Kim, Hyoung-Kyu Song, Thibault Castells, Shinkook Choi

ECCV 2024 · Image generation › Fast diffusion sampling · approach: Training-free diffusion acceleration · 65 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
remove blocks from U-Net and retrain with distillation (BK-SDM-Base)SDM-v1.4 · GPU latency, 512x512whole-pipeline latency: 1.41s -> 0.96s for BaseImproved
add distillation (output + feature KD) during retraining (ablation)BK-SDM-Base · MS-COCO 30KFID: 23.57 -> 15.76Improved
add output-level KD on top of denoising task loss (ablation)BK-SDM-v2-Small (teacher init) · MS-COCO 30KFID: 16.71 -> 14.27Improved
add feature-level KD on top of output KD (ablation)BK-SDM-v2-Small (teacher init) · MS-COCO 30KFID: 14.27 -> 16.61 FID; CLIP 0.2777 -> 0.2901Mixed
initialize from teacher weights instead of random (ablation)BK-SDM-v2-Small · MS-COCO 30KFID: 24.47 -> 16.61Improved
use magnitude pruning criterion instead of hand-designed block removal (ablation)BK-SDM-Small · MS-COCO 30KFID: >150Hurt
use Taylor pruning criterion instead of hand-designed removal (ablation)BK-SDM-Small · MS-COCO 30KU-Net MACs: 217.7G -> 305.4G MACs; FID 16.83 -> 16.73Hurt
use CLIP-score sensitivity criterion instead of hand-designed removal (ablation)BK-SDM-Small · MS-COCO 30KFID: 16.83 -> 21.86Hurt
increase batch size from 64 to 256 (ablation)BK-SDM-Base · MS-COCO 30KFID: FID 14.61 -> 15.76; IS 31.44 -> 33.79Mixed
remove the entire mid-stage of the U-Net without retraining (ablation)SDM-v1.4 · MS-COCO 256x256 30KFID: 13.05 -> 15.60; U-Net params -11.3%Mixed
remove no KD from personalization finetuning backbone (add KD retraining) (ablation)BK-SDM-Base DreamBooth · DreamBoothCLIP-I: 0.669 -> 0.708Improved
use BK-SDM backbone for DreamBooth finetuning instead of SDM-v1.4SDM-v1.4 · DreamBoothDINO: 0.728 -> 0.723 (Base)No clear effect

Later papers that compared with this method

5 later accepted papers report BK-SDM Base as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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