AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

DC-AE 1.5: Accelerating Diffusion Model Convergence with Structured Latent Space

Junyu Chen, Dongyun Zou, Wenkun He, Junsong Chen, Enze Xie, Song Han, Han Cai

ICCV 2025 · Image generation › Autoregressive and latent generation · approach: High-resolution diffusion synthesis · 19 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
use DC-AE 1.5 (structured latent + augmented diffusion training) instead of DC-AE at f32c128DiT-XL · ImageNet 256x256gFID: 26.44 -> 17.31Improved
use DC-AE 1.5 instead of DC-AE at f32c128UViT-H · ImageNet 256x256gFID: 17.38 -> 10.82Improved
use DC-AE 1.5 instead of DC-AE at f64c128USiT-H · ImageNet 512x512gFID: 6.88 -> 5.35Improved
add Structured Latent Space alone, without augmented diffusion training (ablation)DiT-XL with DC-AE-f32c128 · ImageNet 256x256gFID: 26.44 -> 26.75No clear effect
add Augmented Diffusion Training alone, without structured latent space (ablation)DiT-XL with DC-AE-f32c128 · ImageNet 256x256gFID: 26.44 -> 36.83Hurt
add Structured Latent Space alone (ablation)USiT-H with DC-AE-f32c128 · ImageNet 256x256gFID: 8.45 -> 6.72Improved
add Augmented Diffusion Training alone (ablation)USiT-H with DC-AE-f32c128 · ImageNet 256x256gFID: 8.45 -> 8.99No clear effect
combine Structured Latent Space with Augmented Diffusion Training (ablation)USiT-H with DC-AE-f32c128 · ImageNet 256x256gFID: 8.45 -> 6.10Improved
add Augmented Diffusion Training on structured latent channels (ablation)UViT-H · ImageNet 256x256convergence speed: 6x faster convergenceImproved
apply DC-AE 1.5 at small channel number c32 instead of DC-AEDiT-XL / USiT-H · ImageNet 256x256 (f32c32)gFID: 10.18 -> 10.50Hurt
impose channel-wise structure by random channel masking in autoencoder trainingDC-AE-f32c128 · ImageNet 256x256 reconstructionrFID: 0.26 -> 0.26No clear effect
use DC-AE-1.5-f64c128 instead of DC-AE-f32c32 with USiT-2BUSiT-2B · ImageNet 512x512gFID (no CFG): 2.90 -> 2.18Improved

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