AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

RePack then Refine: Efficient Diffusion Transformers with Vision Foundation Models

Guanfang Dong, Luke Schultz, Negar Hassanpour, Chao Gao

ICML 2026 · Image generation › Autoregressive and latent generation · approach: High-resolution diffusion synthesis

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 Latent-Guided Refiner (ablation)RePack-DiT-XL/1 · ImageNet 256x256 (CFG)gFID: 1.82 -> 1.65Improved
add Latent-Guided Refiner (ablation)RePack · ImageNet reconstructionrFID: 0.83 -> 0.69Improved
add Refiner (ablation)RePack-DiT-B/1 · ImageNet (64 epochs)gFID: 4.12 -> 3.99Improved
add Refiner (ablation)RePack-DiT-XL/2 · ImageNet (64 epochs)gFID: 3.66 -> 3.20Improved
add Refiner with fewer sampling steps (25 steps) (ablation)RePack-DiT-XL/1 · ImageNet (64 epochs)gFID: 2.09 -> 2.04Improved
add Refiner without CFG (ablation)RePack · ImageNet 256x256 (no CFG)gFID: 3.81 -> 3.79No clear effect
apply Refiner to VA-VAE or SD-VAE (ablation)VA-VAE, SD-VAE · ImageNetgFIDHurt
use bias-free linear projection instead of linear with bias (ablation)RePack · ImageNet (DiT-B/1, 64 epochs)gFID: 5.08 -> 4.12Improved
use learnable linear projection instead of PCA (ablation)RePack · ImageNet (DiT-B/1, 64 epochs)gFID: 7.56 -> 4.12Improved
reduce latent dimension from 32 to 16 (ablation)RePack · ImageNet (DiT-B/1, 64 epochs)gFID: 4.12 -> 3.61Improved
increase latent dimension from 32 to raw 768 (ablation)RePack · ImageNet (DiT-B/1, 64 epochs)gFID: 4.12 -> 66.43Hurt
switch VFM from DINOv3-B to DINOv2-B (ablation)RePack · ImageNet (DiT-B/1, 64 epochs)gFID: 4.12 -> 3.84Improved

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