AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

SwiftBrush v2: Make Your One-step Diffusion Model Better Than Its Teacher

Trung Tuan Dao, Thuan Hoang Nguyen, Thanh Van Le, Duc H Vu, Khoi Nguyen, Cuong Pham, Anh T Tran

ECCV 2024 · Image generation › Text-to-image diffusion · approach: Text-driven synthesis and semantic guidance · 14 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
initialize SwiftBrush student with SD Turbo pretrained weights (ablation)SwiftBrush · MS COCO-2014 30KFID: 15.46 -> 14.59Improved
add 2M LAION prompts to the training set (ablation)SwiftBrush (Turbo init.) · MS COCO-2014 30KFID: 14.59 -> 11.27Improved
add clamped CLIP loss in LoRA efficient training (ablation)SwiftBrush v2 · MS COCO-2014 30KFID: 13.21 -> 11.70Improved
add unclamped (naive) CLIP loss (ablation)SwiftBrush v2 · MS COCO-2014 30KFID: 14.59 -> 21.32Hurt
clamp CLIP loss with ReLU margin instead of naive CLIP (ablation)SwiftBrush v2 · MS COCO-2014 30KFID: 21.32 -> 13.19Improved
add weight scheduler for clamped CLIP loss (ablation)SwiftBrush v2 · MS COCO-2014 30KFID: 13.19 -> 11.70Improved
merge fully trained Model A and LoRA Model B by weight interpolation (ablation)SwiftBrush v2 · MS COCO-2014 30KFID: 11.27 / 11.02 -> 8.77Improved
add extra image regularization on minimal real data (ablation)SwiftBrush v2 merged model · MS COCO-2014 30KFID: 8.77 -> 8.14Improved
add LoRA and TinyVAE efficient training scheme (ablation)SwiftBrush v2 with CLIP loss · training costGPU days: 7.8 -> 4.3Improved
scale training prompts from 1.3M to 3.3M (ablation)SwiftBrush (Turbo init.) · MS COCO-2014 30KRecall: 0.55 -> 0.54Mixed
initialize from SwiftBrush instead of SD Turbo (ablation)SwiftBrush v2 · MS COCO-2014 30KFID: 8.77 -> 11.69Hurt
add HPSv2 loss to efficient training (ablation)Model B · MS COCO-2014 30K and HPSv2FID / HPSv2 Photo: FID 11.02 -> 11.15; HPSv2 Photo 26.71 -> 28.00Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

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