AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

MacTok: Robust Continuous Tokenization for Image Generation

Hengyu Zeng, Xin Gao, Guanghao Li, Yuxiang Yan, Jiaoyang Ruan, Junpeng Ma, Haoyu Albert Wang, Jian Pu

CVPR 2026 · Image generation › Autoregressive and latent generation · approach: Hierarchical VQ-VAE generation · 2 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 local representation alignment to random masking (ablation)MacTok-128 · ImageNet 256x256 (SiT-B)gFID: 6.01 -> 3.53Improved
add DINO-guided semantic masking (ablation)MacTok-128 · ImageNet 256x256 (SiT-B)gFID: 3.53 -> 3.32Improved
add global representation alignment (ablation)MacTok-128 · ImageNet 256x256 (SiT-B)gFID: 3.32 -> 3.15Improved
use equal mix of random and semantic masking instead of random only (ablation)MacTok-128 · ImageNet 256x256 (SiT-B, w/o CFG)gFID: 14.59 -> 13.95Improved
use 100% semantic masking instead of random only (ablation)MacTok-128 · ImageNet 256x256 (SiT-B, w/o CFG)gFID: 14.59 -> 14.84No clear effect
raise max mask ratio from 0.4 to 0.7 (ablation)MacTok-128 · ImageNet 256x256 (SiT-B, w/o CFG)gFID: 15.49 -> 14.59Improved
raise max mask ratio from 0.7 to 0.8 (ablation)MacTok-128 · ImageNet 256x256 (SiT-B, w/o CFG)gFID: 14.59 -> 14.92Hurt
add decoder fine-tuning (ablation)MacTok-128 · ImageNet 256x256 (SiT-B, w/o CFG)rFID: 0.57 -> 0.43Improved
use MacTok tokenizer instead of SoftVQ-VAE at 64 tokensSiT-B · ImageNet 256x256 (optimal CFG)gFID: 4.09 -> 3.22Improved
increase MacTok tokens from 64 to 128MacTok+SiT-XL · ImageNet 512x512 (w/o CFG)gFID: 4.63 -> 5.12Hurt

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