AI Paper Atlas

Research map / Architectures and efficiency / Quantization

PTQ4ARVG: Post-Training Quantization for AutoRegressive Visual Generation Models

Xuewen Liu, Zhikai Li, Jing Zhang, Mengjuan Chen, Jianquan Li, Qingyi Gu

ICLR 2026 · Architectures and efficiency › Quantization · approach: Post-training quantization of ViTs · 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 GPS scaling (ablation)SmoothQuant baseline · RAR-B W6A6FID: 63.77 -> 36.51Improved
add STWQ on top of GPS (ablation)SmoothQuant+GPS · RAR-B W6A6FID: 36.51 -> 6.67Improved
add DGC calibration on top of GPS and STWQ (ablation)PTQ4ARVG · RAR-B W6A6FID: 6.67 -> 5.13Improved
add DGC on top of GPS and STWQ (ablation)PTQ4ARVG · VAR-d16 W6A6FID: 8.75 -> 8.34Improved
use GPS instead of SmoothQuant scaling (ablation)PTQ4ARVG · RAR-B W6A6FID: 10.26 -> 5.13Improved
use GPS instead of RepQ* scaling (ablation)PTQ4ARVG · RAR-B W6A6FID: 33.79 -> 5.13Improved
use GPS instead of OS+ scaling (ablation)PTQ4ARVG · RAR-B W6A6FID: 7.71 -> 5.13Improved
use GPS instead of SmoothQuant then RepQ* scaling (ablation)PTQ4ARVG · RAR-B W6A6FID: 7.43 -> 5.13Improved
use static instead of dynamic token-wise quantization (ablation)SmoothQuant · VAR-d16 W6A6FID: 30.14 -> 10.41Improved
add dynamic token-wise quantization (ablation)SmoothQuant · VAR-d16 W6A6FID: 18.54 -> 30.14Hurt
add static token-wise quantization (ablation)SmoothQuant · VAR-d16 W6A6FID: 18.54 -> 10.41Improved
use static instead of dynamic token-wise quantization (ablation)SmoothQuant 8-bit deployment · VAR-d16 batch 100speedup: 2.457x -> 2.922xImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 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