AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

VQT: Vector Quantization Tuning for Efficient Fine-tuning and Compression of Pre-trained Vision Transformers

Yinglong Li, jiyuan Xia, Jingcheng Xie, Zhiwei Xiong

ECCV 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Efficient ViT adaptation · 0 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 LoRA with STE and codeword reassignment to codebook-only VQ (ablation)VQ fine-tuning · VTAB-1K ViT-B 1-bitaccuracy (%): 55.1 -> 57.6Improved
add LoRA with STE and codeword reassignment to codebook-only VQ (ablation)VQ fine-tuning · VTAB-1K ViT-B 1.5-bitaccuracy (%): 64.4 -> 69.0Improved
add stochastic sub-vector replacement to LoRA & STE (ablation)VQT · VTAB-1K ViT-B 1-bitaccuracy (%): 57.6 -> 63.9Improved
add stochastic sub-vector replacement to LoRA & STE (ablation)VQT · VTAB-1K ViT-B 2-bitaccuracy (%): 72.5 -> 72.3No clear effect
add progressive alignment on top of stochastic replacement (ablation)VQT · VTAB-1K ViT-B 1-bitaccuracy (%): 63.9 -> 67.5Improved
add progressive alignment on top of stochastic replacement (ablation)VQT · VTAB-1K ViT-B 1.5-bitaccuracy (%): 69.4 -> 70.8Improved
use both continuous-anchored and VQ-anchored modes instead of one (ablation)VQT · VTAB-1K ViT-B 1-bitaccuracy (%): 66.2 -> 67.5Improved
replace LoRA with DoRA inside VQT instead of direct VQ+DoRAVQT · VTAB-1K ViT-B 1-bit with DoRAaccuracy (%): 57.4 -> 67.9Improved
unified fine-tune-and-quantize VQT instead of post-training VQ4DiT on fine-tuned LoRAVQT · VTAB-1K ViT-B 1-bitaccuracy (%): 60.9 -> 67.5Improved
unified VQT instead of VQ4DiT with task loss at 2-bitVQT · VTAB-1K ViT-B 2-bitaccuracy (%): 72.9 -> 72.6No clear effect
allow codeword reassignment instead of fixed assignments (PQF)VQT · VTAB-1K ViT-B 1.5-bitaccuracy (%): 65.7 -> 70.8Improved
use VQT instead of SSVQ on ViT-L at 1 bitVQT · VTAB-1K ViT-Laccuracy (%): 59.2 -> 66.4Improved

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