AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

CODA: Repurposing Continuous VAEs for Discrete Tokenization

Zeyu Liu, Zanlin Ni, Yeguo Hua, Xin Deng, Xiao Ma, Cheng Zhong, Gao Huang

ICCV 2025 · Image generation › Autoregressive and latent generation · approach: Hierarchical VQ-VAE generation · 6 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
add residual quantization (10 levels) to plain VQ on a continuous VAE (ablation)CODA · ImageNet 256x256, 16xrFID: 41.52 -> 14.23Improved
add attention-based sparse quantization (ablation)CODA · ImageNet 256x256, 16xrFID: 14.23 -> 7.06; utilization 51% -> 100%Improved
adapt the VAE with LoRA during discretization (ablation)CODA · ImageNet 256x256, 16xrFID: 7.06 -> 1.34Improved
tune LoRA on decoder versus encoder (ablation)CODA · ImageNet 256x256, 16xrFID: encoder 7.06 -> 5.75; decoder 7.06 -> 1.59Mixed
scale codebook size from 1024 to 65536 (ablation)CODA · ImageNet 256x256, 16xrFID: 2.07 -> 1.34Improved
normalize query and key in attention quantizer (ablation)CODA · ImageNet 256x256, 16xrFID: 1.87 -> 1.34Improved
use RMSNorm instead of LayerNorm (ablation)CODA · ImageNet 256x256, 16xrFID: 1.37 -> 1.34No clear effect
use CODA tokenizer instead of VQGAN trainingdiscrete tokenizer · ImageNet 256x256training iterations and utilization: 300K -> 50K iterations; 5.9% -> 100% utilizationImproved
use CODA tokenizer in MaskGIT instead of original tokenizerMaskGIT · ImageNet 256x256 generationFID: 4.92 -> 3.17Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

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