AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Memory-Efficient Fine-Tuning Diffusion Transformers via Dynamic Patch Sampling and Block Skipping

Sunghyun Park, Jeongho Kim, Hyoungwoo Park, Debasmit Das, Sungrack Yun, Munawar Hayat, Jaegul Choo, Fatih Porikli, Seokeon Choi

CVPR 2026 · Image generation › Autoregressive and latent generation · approach: High-resolution diffusion synthesis · 5 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
replace fixed 256x256 resize with dynamic timestep-based patch sampling (ablation)LoRA on FLUX · DreamBooth datasetDINO: 0.7164 -> 0.7253Improved
dynamic patch sampling at 256x256 vs LoRA 512x512 (ablation)LoRA on FLUX · DreamBooth datasetDINO: 0.7324 -> 0.7253Hurt
add precomputed residual features when skipping blocks (30%) (ablation)DiT-BlockSkip on FLUX · DreamBooth datasetDINO: 0.4313 -> 0.7282Improved
add precomputed residual features when skipping blocks (50%) (ablation)DiT-BlockSkip on FLUX · DreamBooth datasetDINO: 0.4301 -> 0.7150Improved
skip selected blocks (cross-attention masking selection) instead of first 50% (ablation)DiT-BlockSkip on FLUX · DreamBooth datasetDINO: 0.6651 -> 0.7150Improved
skip selected blocks instead of last 50% (ablation)DiT-BlockSkip on FLUX · DreamBooth datasetDINO: 0.4808 -> 0.7150Improved
increase skip ratio from 30% to 50% with residual features (ablation)DiT-BlockSkip on FLUX · DreamBooth datasetDINO: 0.7282 -> 0.7150Hurt
use DiT-BlockSkip instead of HollowedNet block selectionDiT-BlockSkip 30% · DreamBooth dataset (FLUX)DINO: 0.4899 -> 0.7194Improved
use DiT-BlockSkip instead of LISA on SANADiT-BlockSkip 30% · DreamBooth dataset (SANA)DINO: 0.5226 -> 0.7388Improved
DiT-BlockSkip vs LoRA in user study (subject fidelity)DiT-BlockSkip · User studypreference (%): 40.0% LoRA vs 37.7% oursNo clear effect
DiT-BlockSkip vs LoRA in user study (text fidelity)DiT-BlockSkip · User studypreference (%): 29.4% -> 45.6%Improved

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