AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Scalable Image Tokenization with Index Backpropagation Quantization

Fengyuan Shi, Zhuoyan Luo, Yixiao Ge, Yujiu Yang, Ying Shan, Limin Wang

ICCV 2025 · Image generation › Autoregressive and latent generation · approach: Hierarchical VQ-VAE generation · 47 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 VQ with IBQ (index backpropagation to all codes) (ablation)VQGAN tokenizer (16,384 codes, dim 256) · ImageNet val reconstructionrFID: 3.98 -> 1.67 rFID; usage 5.3% -> 98%Improved
add double quantization loss (ablation)IBQ · ImageNet val reconstructionrFID: 1.67 -> 1.55Improved
use a deeper tokenizer (4 ResBlocks instead of 2) (ablation)IBQ + double quantization · ImageNet val reconstructionrFID: 1.55 -> 1.37Improved
increase codebook size from 1,024 to 262,144 (ablation)IBQ · ImageNet val reconstructionrFID: 2.24 -> 1.00Improved
increase codebook size to 262,144 (ablation)IBQ · ImageNet val reconstructioncodebook usage: 96% -> 84%Hurt
increase code dimension from 32 to 256 (ablation)IBQ · ImageNet val reconstructionrFID: 2.04 -> 1.37Improved
increase code dimension from 64 to 128 to 256 (ablation)IBQ · ImageNet val reconstructionrFID: 1.39 -> 1.38 -> 1.37No clear effect
increase ResBlocks in encoder/decoder from 1 to 4 (ablation)IBQ · ImageNet val reconstructionrFID: 1.80 -> 1.37Improved
increase code dimension from 8 to 256 in standard VQ (LlamaGen)LlamaGen tokenizer · ImageNet val reconstructionrFID: 2.19 -> 9.21; usage 97% -> 0.29%Hurt
use IBQ instead of LFQ tokenizer (ablation)vanilla AR transformer 342M, 16,384 codes · ImageNet 256x256 generationgFID: 3.40 -> 2.88Improved
scale AR model from 342M to 2.1B with IBQ tokenizerIBQ-B to IBQ-XXL · ImageNet 256x256 generationgFID: 2.88 -> 2.05Improved
replace hard-index backprop with soft quantization (weighted code average) in training (ablation)IBQ · ImageNet val reconstructionrFID: 16.17 vs 4.03 (same epochs); soft train/hard infer 233.17Hurt

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