paper-with-me

홈 › Papers

bViT: Investigating Single-Block Recurrence in Vision Transformers for Image Recognition

2026-05-11 · Michal Byra, Pawel Olszowiec, Grzegorz Stefanski, Grzegorz Gruszczynski, Alberto Presta arxiv

Vision Transformers (ViTs) are built by stacking independently parameterized blocks, but it remains unclear how much of this depth requires layer specific transformations and how much can be realized through recurrent computation. We study this question with bViT, a single-block recurrent ViT in which one transformer block is applied repeatedly to process an image. This architecture preserves the iterative structure of a deep ViT while removing layer specific block parameterization, providing a controlled setting for studying recurrence in vision. On ImageNet-1K, a 12-step bViT-B achieves accuracy comparable to standard ViT-B under the same training recipe and computational budget, while using an order of magnitude fewer parameters. We observe that recurrent performance improves with representation width, with wider bViTs recovering much more of the performance of standard ViTs than narrow variants. We interpret this behavior as implicit depth multiplexing, where a shared block expresses multiple step-dependent computations through the evolving hidden state. Beyond ImageNet classification, bViT transfers competitively to downstream tasks and enables parameter-efficient fine-tuning. Mechanistic analyses of activations, attention and step-specific pruning show that the shared block changes its effective behavior across recurrent steps rather than simply repeating the same computation. Our results suggest that a large fraction of ViT depth can be implemented through recurrent reuse, provided that the representation space is sufficiently wide.

📄 PDF Abstract BibTeX arXiv:2605.10661

Code (0)

등록된 구현이 없습니다.

Tasks

parameter-efficient fine-tuning

Similar Papers 제목 키워드 기반

Subtoken Vision Transformer for Fine-grained Recognition

2026-07-10 · Jie Zhu, Ivy Zhang, Minchul Kim, Xiaoming Liu arxiv

We present Subtoken Vision Transformer (SubViT), a selective image tokenization method for fine-grained visual recognition. Standard Vision Transformers compress each fixed-size patch into a single token, although fine-g…

Fine-Grained Visual Recognition

BViT: Broad Attention based Vision Transformer

2022-02-13 · Nannan Li, Yaran Chen, Weifan Li, Zixiang Ding 외

Recent works have demonstrated that transformer can achieve promising performance in computer vision, by exploiting the relationship among image patches with self-attention. While they only consider the attention in a si…

image-classificationImage ClassificationObject Recognition

MABViT -- Modified Attention Block Enhances Vision Transformers

2023-12-03 · Mahesh Ramesh, Aswinkumar Ramkumar

Recent studies have demonstrated the effectiveness of Gated Linear Units (GLU) in enhancing transformer models, particularly in Large Language Models (LLMs). Additionally, utilizing a parallel configuration within each T…

image-classificationImage Classification

How Much Is One Recurrence Worth? Iso-Depth Scaling Laws for Looped Language Models

2026-04-22 · Kristian Schwethelm, Daniel Rueckert, Georgios Kaissis arxiv

We measure how much one recurrence is worth to a looped (depth-recurrent) transformer, in equivalent unique parameters. From an iso-depth pretraining sweep across recurrence counts $r \in \{1, 2, 4, 8\}$ spanning ${\sim}…

Intra-Layer Recurrence in Transformers for Language Modeling

2025-05-03 · Anthony Nguyen, Wenjun Lin

Transformer models have established new benchmarks in natural language processing; however, their increasing depth results in substantial growth in parameter counts. While existing recurrent transformer methods address t…

Language ModelingLanguage Modelling