AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Taming Sampling Perturbations with Variance Expansion Loss for Latent Diffusion Models

Qifan Li, Xingyu Zhou, Jinhua Zhang, Weiyi You, Shuhang Gu

CVPR 2026 · Image generation › Autoregressive and latent generation · approach: Latent diffusion with learned tokenizers · 3 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 KL with VE loss (ablation)LDM tokenizer · ImageNet 256 (DiT-B)FID-10K: 31.93 -> 29.03Improved
replace KL with VE loss (ablation)LDM tokenizer · ImageNet 256 (LightningDiT-B)FID-10K: 22.25 -> 19.70Improved
replace KL with VE loss (ablation)LDM tokenizer · ImageNet 256 tokenizerrFID: 0.55 -> 0.60Hurt
replace KL with VE loss (ablation)VA-VAE tokenizer · ImageNet 256 (DiT-B)FID-10K: 22.27 -> 19.42Improved
replace KL with VE loss (ablation)VA-VAE tokenizer · ImageNet 256 (LightningDiT-B)FID-10K: 19.85 -> 15.50Improved
replace KL with VE loss (ablation)VA-VAE tokenizer · ImageNet 256 tokenizerPSNR: 27.43 -> 26.54Hurt
train VA-VAE+VE for 16 epochs versus VA-VAE for 50 epochs (ablation)VA-VAE tokenizer · ImageNet 256 (LightningDiT-B, 160 ep)FID-10K: 15.82 -> 12.89Improved
use VE loss instead of raising KL weight to beta=8 (ablation)VAE tokenizer · ImageNet 256 (no CFG)FID-10K: 27.54 -> 18.90Improved
sweep KL weight from 1e-6 to 1 (ablation)VAE tokenizer · ImageNet 256 (no CFG)FID-10K: 23.12 -> 23.18No clear effect
increase KL weight from 1e-6 to 8 (ablation)VAE tokenizer · ImageNet 256 tokenizerrFID: 0.39 -> 2.36Hurt
replace KL with VE loss (ablation)VAE tokenizer · ImageNet 256 (no CFG)FID-10K: 23.12 -> 18.90Improved
fine-tune VA-VAE with VE lossVA-VAE tokenizer · ImageNet 256 finetunedrFID: 0.28 -> 0.26Improved

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