Papers › Pattern Attention Transformer with Doughnut Kernel

Pattern Attention Transformer with Doughnut Kernel

30 Nov 2022arXiv:2211.16961archive 2025-07-28

Wenyuan Sheng

We present in this paper a new architecture, the Pattern Attention Transformer (PAT), that is composed of the new doughnut kernel. Compared with tokens in the NLP field, Transformer in computer vision has the problem of handling the high resolution of pixels in images. In ViT, an image is cut into square-shaped patches. As the follow-up of ViT, Swin Transformer proposes an additional step of shifting to decrease the existence of fixed boundaries, which also incurs 'two connected Swin Transformer blocks' as the minimum unit of the model. Inheriting the patch/window idea, our doughnut kernel enhances the design of patches further. It replaces the line-cut boundaries with two types of areas: sensor and updating, which is based on the comprehension of self-attention (named QKVA grid). The doughnut kernel also brings a new topic about the shape of kernels beyond square. To verify its performance on image classification, PAT is designed with Transformer blocks of regular octagon shape doughnut kernels. Its architecture is lighter: the minimum pattern attention layer is only one for each stage. Under similar complexity of computation, its performances on ImageNet 1K reach higher throughput (+10%) and surpass Swin Transformer (+0.8 acc1).

Paper PDF

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Image Classification image-classification

Results from the paper archive 2025-07-28

Task	Dataset	Model	Metric	Value	Rank at snapshot	Leaderboard	Report
Image Classification	ImageNet	PAT-B	Top 1 Accuracy	83.6%	#409 of 1060	Archive leaderboard	report
Image Classification	ImageNet	PAT-S	Top 1 Accuracy	83.1%	#467 of 1060	Archive leaderboard	report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Absolute Position Encodings Adam Attention BPE Dense Connections Dropout Label Smoothing Layer Normalization Linear Layer Multi-Head Attention Position-Wise Feed-Forward Layer Residual Connection Softmax Stochastic Depth Swin Transformer Transformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections