AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

ViQ: Text-Aligned Visual Quantized Representations at Any Resolution

Xumin Yu, Zuyan Liu, Zhenyu Yang, Yuhao Dong, Shengsheng Qian, Jiwen Lu, Han Hu, Yongming Rao

ECCV 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Vision-language pretraining · 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
add BN + l2 bottleneck before SimVQ (gradual regularization) (ablation)ViQ · ViQ ablation (Qwen2.5-3B)average score: 60.9 -> 66.6Improved
use FSQ instead of SimVQ after BN + l2 (ablation)ViQ · ViQ ablation (Qwen2.5-3B)average score: 66.6 -> 67.9Improved
use l-infinity instead of l2 regularization (ablation)ViQ · ViQ ablation (Qwen2.5-3B)average score: 67.9 -> 68.7Improved
inject RoPE with attention for position information (ablation)ViQ · ViQ ablation (Qwen2.5-3B)average score: 65.3 -> 68.7Improved
use RoPE instead of learnable position embedding (ablation)ViQ · ViQ ablation (Qwen2.5-3B)average score: 65.7 -> 68.7Improved
add self-distillation loss to text loss (ablation)ViQ · ViQ ablation (Qwen2.5-3B)average score: 61.3 -> 66.8Improved
add VAE-latent reconstruction loss (ablation)ViQ · ViQ ablation (Qwen2.5-3B)average score: 66.8 -> 68.7Improved
use VAE latent loss instead of MSE + LPIPS (ablation)ViQ · ViQ ablation (Qwen2.5-3B)average score: 67.0 -> 68.7Improved
use larger FSQ codebook of 128,000 (ablation)ViQ · ViQ ablation (Qwen2.5-3B)average score: 68.7 -> 68.3No clear effect
widen bottleneck from 128 to 1536 (ablation)ViQ · ViQ ablation (Qwen2.5-3B)average score: 69.1 -> 69.3No clear effect
use ViQ discrete codes instead of SigLIP2-g features for VLM training (0.5B LLM, forward time)ViQ · VLM SFT efficiencyforward time speed-up: 70% (4k), 78% (16k)Improved
use ViQ instead of UniTok quantized encoderViQ · Multimodal understanding (Qwen2.5-1.5B)average score: 33.0 -> 57.2Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 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