AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Unifying CNNs and ViTs for Learning-Efficient and Scalable Variational AutoEncoder

Zhiying Lu, Shang Chai, Chuanbin Liu, Litong Gong, Pandeng Li, Tiezheng Ge, Hongtao Xie

ECCV 2026 · Image generation › Autoregressive and latent generation · approach: Hierarchical VQ-VAE generation · 0 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 RoPE only for positional encoding instead of absolute position embedding (ablation)ViT-VAE · Extrapolation 256 to 512reconstruction qualityImproved
add a positional encoding (RoPE) to a ViT-VAE without any (ablation)ViT-VAE · Early training on ImageNet-1kvalidation reconstruction lossImproved
replace patch embedding and pixel-shuffle with sequential overlapping convolutions (SeqConv) (ablation)ViT w/ RoPE · Early training on ImageNet-1kvalidation reconstruction lossImproved
add a multi-stage design with ResBlock front-end (ablation)ViT w/ RoPE · Early training on ImageNet-1kvalidation reconstruction lossImproved
add depth-wise convolution FFN (ablation)ViT w/ RoPE w/ Multi-Stage · Early training on ImageNet-1kvalidation reconstruction lossImproved
add full convolution FFN (ablation)ViT w/ RoPE w/ Multi-Stage · Early training on ImageNet-1kvalidation reconstruction lossImproved
replace FullConv-FFN with MoE-FFN (ablation)TransVAE · Early training on ImageNet-1kvalidation reconstruction lossHurt
add RMSNorm before QKV projections (ablation)TransVAE · long training at scaletraining stabilityImproved
scale TransVAE from Tiny to LargeTransVAE · ImageNet-1k val 256x256 (f16d32)PSNR: 29.08 -> 29.78Improved
scale TransVAE from Tiny to LargeTransVAE · ImageNet-1k val 1024x1024 (f16d32)PSNR: 36.38 -> 38.25Improved
add VF alignment lossTransVAE-T · ImageNet-1k val 256x256 (f16d32)PSNR: 29.08 -> 28.76Hurt
add VF alignment lossTransVAE-T · ImageNet-1k val 1024x1024 (f16d32)rFID: 0.08 -> 0.24Hurt

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