Papers › MogaNet: Multi-order Gated Aggregation Network

MogaNet: Multi-order Gated Aggregation Network

7 Nov 2022arXiv:2211.03295archive 2025-07-28

Siyuan Li, Zedong Wang, Zicheng Liu, Cheng Tan, Haitao Lin, Di wu, ZhiYuan Chen, Jiangbin Zheng, Stan Z. Li

By contextualizing the kernel as global as possible, Modern ConvNets have shown great potential in computer vision tasks. However, recent progress on \textit{multi-order game-theoretic interaction} within deep neural networks (DNNs) reveals the representation bottleneck of modern ConvNets, where the expressive interactions have not been effectively encoded with the increased kernel size. To tackle this challenge, we propose a new family of modern ConvNets, dubbed MogaNet, for discriminative visual representation learning in pure ConvNet-based models with favorable complexity-performance trade-offs. MogaNet encapsulates conceptually simple yet effective convolutions and gated aggregation into a compact module, where discriminative features are efficiently gathered and contextualized adaptively. MogaNet exhibits great scalability, impressive efficiency of parameters, and competitive performance compared to state-of-the-art ViTs and ConvNets on ImageNet and various downstream vision benchmarks, including COCO object detection, ADE20K semantic segmentation, 2D\&3D human pose estimation, and video prediction. Notably, MogaNet hits 80.0\% and 87.8\% accuracy with 5.2M and 181M parameters on ImageNet-1K, outperforming ParC-Net and ConvNeXt-L, while saving 59\% FLOPs and 17M parameters, respectively. The source code is available at \url{https://github.com/Westlake-AI/MogaNet}.

PaperPDFCodeCode Syntology ran

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2211.03295")

Code

Syntology Ran 12 of 15 code samples harvested from 2 repositories linked to this paper; 3 have no recorded run. Of those that ran: 2 ran · our draft was wrong; 10 ran with no contract checked.

By repository: official repository: 15 samples from 2 repositories, 12 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Westlake-AI/MogaNet officialmentioned in papermentioned on GitHubpytorch report
Westlake-AI/openmixup officialmentioned in papermentioned on GitHubpytorch report
chengtan9907/OpenSTL officialmentioned in papermentioned on GitHubpytorch report
chengtan9907/simvpv2 mentioned in papermentioned on GitHubpytorchApache-2.0 report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

15 samples harvested; 12 ran; 0 honoured the contract we drafted; 3 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

2ran · our draft was wrong
10ran
3unverified

Licence: 0 of the 15 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 2 repositories linked to this paper, official or community; each sample names its own and says which. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

ChannelAggregationFFN Westlake-AI/MogaNet/models/moganet.py official repository ran · metamorphic tier: deterministic fingerprinted Apache-2.0 (permissive) · cc2b8b29fdd8ae54 · report
ChannelAggregationFFN chengtan9907/OpenSTL/openstl/modules/simvp_modules.py official repository ran fingerprinted Apache-2.0 (permissive) · cf05fdcf563e2f19 · report
ConvPatchEmbed Westlake-AI/MogaNet/models/moganet.py official repository ran · metamorphic tier: deterministic fingerprinted Apache-2.0 (permissive) · 415533ec94ba9379 · report
ElementScale Westlake-AI/MogaNet/models/moganet.py official repository ran · metamorphic tier: deterministic fingerprinted Apache-2.0 (permissive) · f347afec9d2fadaa · report
LayerNorm2d Westlake-AI/MogaNet/models/moganet.py official repository ran · metamorphic tier: deterministic Apache-2.0 (permissive) · 080b40f8bc571c80 · report
MultiOrderDWConv Westlake-AI/MogaNet/models/moganet.py official repository ran Apache-2.0 (permissive) · 34e96c682001e3a4 · report
MultiOrderDWConv chengtan9907/OpenSTL/openstl/modules/simvp_modules.py official repository ran Apache-2.0 (permissive) · 528b9b4cc6a8170a · report
MultiOrderGatedAggregation Westlake-AI/MogaNet/models/moganet.py official repository ran Apache-2.0 (permissive) · 629c8ffe83ae92cb · report
MultiOrderGatedAggregation chengtan9907/OpenSTL/openstl/modules/simvp_modules.py official repository ran Apache-2.0 (permissive) · c06bcd5ffc9492cb · report
StackConvPatchEmbed Westlake-AI/MogaNet/models/moganet.py official repository ran · metamorphic tier: deterministic fingerprinted Apache-2.0 (permissive) · 46949c7157f260ad · report
build_act_layer Westlake-AI/MogaNet/models/moganet.py official repository ran · our draft was wrong Apache-2.0 (permissive) · 63ec6cf8064edb07 · report
build_norm_layer Westlake-AI/MogaNet/models/moganet.py official repository ran · our draft was wrong Apache-2.0 (permissive) · 176cfedc20e4532c · report
MogaBlock Westlake-AI/MogaNet/models/moganet.py official repository unverified Apache-2.0 (permissive) · 1b34bbc3671c6f4a · report
MogaNet Westlake-AI/MogaNet/models/moganet.py official repository unverified Apache-2.0 (permissive) · 8e113d262a51881c · report
MogaSubBlock chengtan9907/OpenSTL/openstl/modules/simvp_modules.py official repository unverified Apache-2.0 (permissive) · e17304f230d6cf53 · report

Tasks

3D Human Pose EstimationImage ClassificationInstance SegmentationObject DetectionPose EstimationRepresentation LearningSemantic SegmentationVideo Predictionobject-detection

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Classification ImageNet MogaNet-XL (384res) GFLOPs 102 #69 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-XL (384res) Number of params 181M #69 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-XL (384res) Top 1 Accuracy 87.8% #69 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-L GFLOPs 15.9 #299 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-L Number of params 83M #299 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-L Top 1 Accuracy 84.7% #299 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-B GFLOPs 9.9 #328 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-B Number of params 44M #328 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-B Top 1 Accuracy 84.3% #328 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-S GFLOPs 5 #429 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-S Number of params 25M #429 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-S Top 1 Accuracy 83.4% #429 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-T (256res) GFLOPs 1.44 #725 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-T (256res) Number of params 5.2M #725 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-T (256res) Top 1 Accuracy 80% #725 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-XT (256res) GFLOPs 1.04 #880 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-XT (256res) Number of params 3M #880 of 1060 Archive leaderboard report
Image Classification ImageNet MogaNet-XT (256res) Top 1 Accuracy 77.2% #880 of 1060 Archive leaderboard report
Instance Segmentation COCO test-dev MogaNet-XL (Cascade Mask R-CNN) mask AP 48.8 #27 of 112 Archive leaderboard report
Instance Segmentation COCO test-dev MogaNet-L (Cascade Mask R-CNN) mask AP 46.1 #39 of 112 Archive leaderboard report
Instance Segmentation COCO test-dev MogaNet-B (Cascade Mask R-CNN) mask AP 46 #40 of 112 Archive leaderboard report
Instance Segmentation COCO test-dev MogaNet-S (Cascade Mask R-CNN) mask AP 45.1 #43 of 112 Archive leaderboard report
Instance Segmentation COCO test-dev MogaNet-L (Mask R-CNN 1x) mask AP 44.1 #47 of 112 Archive leaderboard report
Instance Segmentation COCO test-dev MogaNet-B (Mask R-CNN 1x) mask AP 43.2 #50 of 112 Archive leaderboard report
Instance Segmentation COCO test-dev MogaNet-S (Mask R-CNN 1x) mask AP 42.2 #55 of 112 Archive leaderboard report
Instance Segmentation COCO test-dev MogaNet-T (Mask R-CNN 1x) mask AP 39.1 #82 of 112 Archive leaderboard report
Instance Segmentation COCO test-dev MogaNet-XT mask AP 37.6 #91 of 112 Archive leaderboard report
Instance Segmentation COCO test-dev MogaNet-T mask AP 35.8 #97 of 112 Archive leaderboard report
Instance Segmentation COCO val2017 MogaNet-S (256x192) AP50 90.7 #1 of 1 Archive leaderboard report
Instance Segmentation COCO val2017 MogaNet-S (256x192) AP75 82.8 #1 of 1 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-XL (Cascade Mask R-CNN) AP 56.2 #9 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-L (Cascade Mask R-CNN) AP 53.3 #10 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-B (Cascade Mask R-CNN) AP 52.6 #11 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-S (Cascade Mask R-CNN) AP 51.6 #14 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-L (Mask R-CNN 1x) AP 49.4 #18 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-L (RetinaNet 1x) AP 48.7 #21 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-B (Mask R-CNN 1x) AP 47.9 #22 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-B (RetinaNet 1x) AP 47.7 #23 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-S (Mask R-CNN 1x) AP 46.7 #25 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-S (RetinaNet 1x) AP 45.8 #26 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-T (Mask R-CNN 1x) AP 42.6 #28 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-T (RetinaNet 1x) AP 41.4 #29 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-XT (Mask R-CNN 1x) AP 40.7 #30 of 33 Archive leaderboard report
Object Detection COCO 2017 val MogaNet-XT (RetinaNet 1x) AP 39.7 #32 of 33 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-B (384x288) AP 77.3 #2 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-B (384x288) AP50 91.4 #2 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-B (384x288) AP75 84 #2 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-B (384x288) AR 82.2 #2 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-S (384x288) AP 76.4 #4 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-S (384x288) AP50 91 #4 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-S (384x288) AP75 83.3 #4 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-S (384x288) AR 81.4 #4 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-S (256x192) AP 74.9 #8 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-S (256x192) AR 80.1 #8 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-T (256x192) AP 73.2 #9 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-T (256x192) AP50 90.1 #9 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-T (256x192) AP75 81 #9 of 11 Archive leaderboard report
Pose Estimation COCO val2017 MogaNet-T (256x192) AR 78.8 #9 of 11 Archive leaderboard report
Semantic Segmentation ADE20K MogaNet-XL (UperNet) Validation mIoU 54 #71 of 235 Archive leaderboard report
Semantic Segmentation ADE20K MogaNet-L (UperNet) GFLOPs (512 x 512) 1176 #108 of 235 Archive leaderboard report
Semantic Segmentation ADE20K MogaNet-L (UperNet) Validation mIoU 50.9 #108 of 235 Archive leaderboard report
Semantic Segmentation ADE20K MogaNet-B (UperNet) GFLOPs (512 x 512) 1050 #120 of 235 Archive leaderboard report
Semantic Segmentation ADE20K MogaNet-B (UperNet) Validation mIoU 50.1 #120 of 235 Archive leaderboard report
Semantic Segmentation ADE20K MogaNet-S (UperNet) GFLOPs (512 x 512) 946 #136 of 235 Archive leaderboard report
Semantic Segmentation ADE20K MogaNet-S (UperNet) Validation mIoU 49.2 #136 of 235 Archive leaderboard report
Semantic Segmentation ADE20K MogaNet-S (Semantic FPN) GFLOPs (512 x 512) 189 #158 of 235 Archive leaderboard report
Semantic Segmentation ADE20K MogaNet-S (Semantic FPN) Validation mIoU 47.7 #158 of 235 Archive leaderboard report
Video Prediction Moving MNIST MogaNet (SimVP 10x) MAE 51.84 #4 of 31 Archive leaderboard report
Video Prediction Moving MNIST MogaNet (SimVP 10x) MSE 15.67 #4 of 31 Archive leaderboard report
Video Prediction Moving MNIST MogaNet (SimVP 10x) SSIM 0.9661 #4 of 31 Archive leaderboard report
Video Prediction Moving MNIST VAN (SimVP 10x) MAE 53.57 #5 of 31 Archive leaderboard report
Video Prediction Moving MNIST VAN (SimVP 10x) MSE 16.21 #5 of 31 Archive leaderboard report
Video Prediction Moving MNIST VAN (SimVP 10x) SSIM 0.9646 #5 of 31 Archive leaderboard report
Video Prediction Moving MNIST HorNet (SimVP 10x) MAE 55.7 #6 of 31 Archive leaderboard report
Video Prediction Moving MNIST HorNet (SimVP 10x) MSE 17.4 #6 of 31 Archive leaderboard report
Video Prediction Moving MNIST HorNet (SimVP 10x) SSIM 0.9624 #6 of 31 Archive leaderboard report
Video Prediction Moving MNIST ConvNeXt (SimVP 10x) MAE 55.76 #7 of 31 Archive leaderboard report
Video Prediction Moving MNIST ConvNeXt (SimVP 10x) MSE 17.58 #7 of 31 Archive leaderboard report
Video Prediction Moving MNIST ConvNeXt (SimVP 10x) SSIM 0.9617 #7 of 31 Archive leaderboard report
Video Prediction Moving MNIST Uniformer (SimVP 10x) MAE 57.52 #9 of 31 Archive leaderboard report
Video Prediction Moving MNIST Uniformer (SimVP 10x) MSE 18.01 #9 of 31 Archive leaderboard report
Video Prediction Moving MNIST MLP-Mixer (SimVP 10x) MAE 59.86 #10 of 31 Archive leaderboard report
Video Prediction Moving MNIST MLP-Mixer (SimVP 10x) MSE 18.85 #10 of 31 Archive leaderboard report
Video Prediction Moving MNIST Swin (SimVP 10x) MAE 59.84 #12 of 31 Archive leaderboard report
Video Prediction Moving MNIST Swin (SimVP 10x) MSE 19.11 #12 of 31 Archive leaderboard report
Video Prediction Moving MNIST ViT (SimVP 10x) MAE 61.65 #13 of 31 Archive leaderboard report
Video Prediction Moving MNIST ViT (SimVP 10x) MSE 19.74 #13 of 31 Archive leaderboard report
Video Prediction Moving MNIST ViT (SimVP 10x) SSIM 0.9539 #13 of 31 Archive leaderboard report
Video Prediction Moving MNIST Poolformer (SimVP 10x) MAE 64.31 #15 of 31 Archive leaderboard report
Video Prediction Moving MNIST Poolformer (SimVP 10x) MSE 20.96 #15 of 31 Archive leaderboard report
Video Prediction Moving MNIST ConvMixer (SimVP 10x) MAE 67.37 #16 of 31 Archive leaderboard report
Video Prediction Moving MNIST ConvMixer (SimVP 10x) MSE 22.3 #16 of 31 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

1x1 ConvolutionAttentionConvolutionDense ConnectionsGated ConvolutionGated Linear UnitResidual ConnectionVision 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