paper-with-me

홈 › Papers

DiT: Efficient Vision Transformers with Dynamic Token Routing

2023-08-07 · Yuchen Ma, Zhengcong Fei, Junshi Huang

Recently, the tokens of images share the same static data flow in many dense networks. However, challenges arise from the variance among the objects in images, such as large variations in the spatial scale and difficulties of recognition for visual entities. In this paper, we propose a data-dependent token routing strategy to elaborate the routing paths of image tokens for Dynamic Vision Transformer, dubbed DiT. The proposed framework generates a data-dependent path per token, adapting to the object scales and visual discrimination of tokens. In feed-forward, the differentiable routing gates are designed to select the scaling paths and feature transformation paths for image tokens, leading to multi-path feature propagation. In this way, the impact of object scales and visual discrimination of image representation can be carefully tuned. Moreover, the computational cost can be further reduced by giving budget constraints to the routing gate and early-stopping of feature extraction. In experiments, our DiT achieves superior performance and favorable complexity/accuracy trade-offs than many SoTA methods on ImageNet classification, object detection, instance segmentation, and semantic segmentation. Particularly, the DiT-B5 obtains 84.8\% top-1 Acc on ImageNet with 10.3 GFLOPs, which is 1.0\% higher than that of the SoTA method with similar computational complexity. These extensive results demonstrate that DiT can serve as versatile backbones for various vision tasks.

📄 PDF Abstract BibTeX arXiv:2308.03409

Code (1)

maycbj/dit 공식 구현

Tasks

Instance SegmentationObjectobject-detectionObject DetectionSemantic Segmentation

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Adam 설명 없음
Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Residual Connection 설명 없음
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…

Similar Papers 제목 키워드 기반

DTRNet: Dynamic Token Routing Network to Reduce Quadratic Costs in Transformers

2025-08-31 · Aman Sharma, Saeed Najafi, Parsa Farinneya, Benyamin Jamialahmadi 외 arxiv

Transformers achieve state-of-the-art results across many tasks, but their uniform application of quadratic self-attention to every token at every layer makes them computationally expensive. We introduce DTRNet (Dynamic …

Routing Matters in MoE: Scaling Diffusion Transformers with Explicit Routing Guidance

2025-10-28 · Yujie Wei, Shiwei Zhang, Hangjie Yuan, Yujin Han 외 arxiv

Mixture-of-Experts (MoE) has emerged as a powerful paradigm for scaling model capacity while preserving computational efficiency. Despite its notable success in large language models (LLMs), existing attempts to apply Mo…

Computational Efficiency

BiFormer: Vision Transformer with Bi-Level Routing Attention

2023-03-15 · CVPR 2023 1 · Lei Zhu, Xinjiang Wang, Zhanghan Ke, Wayne Zhang 외

As the core building block of vision transformers, attention is a powerful tool to capture long-range dependency. However, such power comes at a cost: it incurs a huge computation burden and heavy memory footprint as pai…

Computational EfficiencyGPUimage-classificationImage Classification+3

Vision Transformers with Mixed-Resolution Tokenization

2023-04-01 · Tomer Ronen, Omer Levy, Avram Golbert

Vision Transformer models process input images by dividing them into a spatially regular grid of equal-size patches. Conversely, Transformers were originally introduced over natural language sequences, where each token r…

image-classificationImage Classification

CAT: Curvature-Adaptive Transformers for Geometry-Aware Learning

2025-10-02 · Ryan Y. Lin, Siddhartha Ojha, Nicholas Bai arxiv

Transformers achieve strong performance across diverse domains but implicitly assume Euclidean geometry in their attention mechanisms, limiting their effectiveness on data with non-Euclidean structure. While recent exten…

Knowledge Graph CompletionRelational Reasoning