# MixFormer: Mixing Features across Windows and Dimensions

Source: [MixFormer: Mixing Features across Windows and Dimensions](<https://www.sota2.com/research/paper/68fa6c361e209131cea32d35>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-01-04T05:18:52.307Z

Authors: Qiang Chen, Qiman Wu, Jian Wang, Qinghao Hu, Tao Hu, Errui Ding, Jian Cheng, Jingdong Wang

Published: 2022-04-06T03:13:50.000Z

## Abstract

While local-window self-attention performs notably in vision tasks, it suffers from limited receptive field and weak modeling capability issues. This is mainly because it performs self-attention within non-overlapped windows and shares weights on the channel dimension. We propose MixFormer to find a solution. First, we combine local-window self-attention with depth-wise convolution in a parallel design, modeling cross-window connections to enlarge the receptive fields. Second, we propose bi-directional interactions across branches to provide complementary clues in the channel and spatial dimensions. These two designs are integrated to achieve efficient feature mixing among windows and dimensions. Our MixFormer provides competitive results on image classification with EfficientNet and shows better results than RegNet and Swin Transformer. Performance in downstream tasks outperforms its alternatives by significant margins with less computational costs in 5 dense prediction tasks on MS COCO, ADE20k, and LVIS. Code is available at \\url{https://github.com/PaddlePaddle/PaddleClas}.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Keypoint Detection on COCO (val)](<https://www.sota2.com/research/sota/keypoint-detection-on-coco-val.md>): Keypoint Detection; COCO (val); AP: 75.3.

- [Instance Segmentation on COCO (val)](<https://www.sota2.com/research/sota/instance-segmentation-on-coco-val.md>): Instance Segmentation; COCO (val); APmk: 44.9.

- [Instance Segmentation on COCO 2017](<https://www.sota2.com/research/sota/instance-segmentation-on-coco-2017.md>): Instance Segmentation; COCO 2017; APm: 41.2.

- [Object Detection on COCO (val)](<https://www.sota2.com/research/sota/object-detection-on-coco-val.md>): Object Detection; COCO (val); mAP: 47.6.

- [Classification on ImageNet-1K 1.0 (val)](<https://www.sota2.com/research/sota/classification-on-imagenet-1k-1-0-val.md>): Classification; ImageNet-1K 1.0 (val); Top-1 Accuracy (%): 83.8.

- [Image Classification on ImageNet-1K 1.0 (val)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-1k-1-0-val.md>): Image Classification; ImageNet-1K 1.0 (val); Top-1 Accuracy: 83.8.

- [Semantic Segmentation on ADE20K (val)](<https://www.sota2.com/research/sota/semantic-segmentation-on-ade20k-val.md>): Semantic segmentation; ADE20K (val); mIoU: not recorded.

- [Object Detection on COCO 2017](<https://www.sota2.com/research/sota/object-detection-on-coco-2017.md>): Object Detection; COCO 2017; AP (Box): not recorded.

- [Instance Segmentation on LVIS v1.0 (val)](<https://www.sota2.com/research/sota/instance-segmentation-on-lvis-v1-0-val.md>): Instance Segmentation; LVIS v1.0 (val); AP (Rare): not recorded.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2204.02557>)

[PDF](<https://arxiv.org/pdf/2204.02557.pdf>)

[Implementation](<https://github.com/PaddlePaddle/PaddleClas>)