AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

REPA-E: Unlocking VAE for End-to-End Tuning of Latent Diffusion Transformers

Xingjian Leng, Jaskirat Singh, Yunzhong Hou, Zhenchang Xing, Saining Xie, Liang Zheng

ICCV 2025 · Image generation › Autoregressive and latent generation · approach: Latent diffusion with learned tokenizers

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
end-to-end tune VAE with REPA lossREPA with SiT-XL · ImageNet 256 (80 epochs, no CFG)gFID: 7.90 -> 4.07Improved
end-to-end tune VAE with REPA lossREPA with SiT-XL · ImageNet 256 (20 epochs, no CFG)gFID: 19.40 -> 12.83Improved
backpropagate diffusion loss to VAE (naive E2E) instead of using stop-gradient (ablation)REPA-E · ImageNet 256 (SiT-L, 100K)gFID: 16.3 -> 444.1Hurt
add batch-norm latent normalization (ablation)REPA-E · ImageNet 256 (SiT-L, 100K)gFID: 18.1 -> 16.3Improved
add GAN regularization loss on VAE (ablation)REPA-E · ImageNet 256 (SiT-L, 100K)gFID: 19.2 -> 16.3Improved
add REPA-E to SiT-BREPA · ImageNet 256 (SiT-B, no CFG)gFID: 49.5 -> 34.8Improved
add REPA-E to SiT-XLREPA · ImageNet 256 (SiT-XL, no CFG)gFID: 19.4 -> 12.8Improved
use CLIP-L as REPA target with REPA-EREPA · ImageNet 256 (SiT-L, CLIP-L encoder)gFID: 29.2 -> 23.4Improved
add REPA-E to VA-VAEREPA · ImageNet 256 (SiT-L, VA-VAE)gFID: 12.8 -> 11.1Improved
add REPA-E to IN-VAEREPA · ImageNet 256 (SiT-L, IN-VAE)gFID: 22.7 -> 12.7Improved
apply REPA alignment at 6th layer with REPA-EREPA · ImageNet 256 (SiT-L, alignment depth 6)gFID: 23.0 -> 16.4Improved
initialize VAE from pretrained SD-VAE instead of from scratch (ablation)REPA-E · ImageNet 256 (400K iters, no CFG)gFID: 4.34 -> 4.07Improved

Later papers that compared with this method

10 later accepted papers report REPA-E as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.

2 more not listed.

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