mlsanityData to 7 Oct 2026

WeightNet: Revisiting the Design Space of Weight Networks

arXiv 2007.11823 ↗First posted 23 Jul 2020cs.CVcited by 50

Not on the map. Why?

Abstract (arXiv)

We present a conceptually simple, flexible and effective framework for weight generating networks. Our approach is general that unifies two current distinct and extremely effective SENet and CondConv into the same framework on weight space. The method, called WeightNet, generalizes the two methods by simply adding one more grouped fully-connected layer to the attention activation layer. We use the WeightNet, composed entirely of (grouped) fully-connected layers, to directly output the convolutional weight. WeightNet is easy and memory-conserving to train, on the kernel space instead of the feature space. Because of the flexibility, our method outperforms existing approaches on both ImageNet and COCO detection tasks, achieving better Accuracy-FLOPs and Accuracy-Parameter trade-offs. The framework on the flexible weight space has the potential to further improve the performance. Code is available at https://github.com/megvii-model/WeightNet.

Key facts. Cited by 50 papers in the map (data through 7 Oct 2026). First posted on arXiv 23 Jul 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 13; 2022: 9; 2023: 10; 2024: 10; 2025: 4; 2026: 4.

50 papers.

Sort:

By area:

  1. 2021Dynamic Neural Networks: A Surveycited by 405
  2. 2021Mobile-Former: Bridging MobileNet and Transformercited by 234
  3. 2022Dynamic Spatial Propagation Network for Depth Completioncited by 70
  4. 2022Omni-Dimensional Dynamic Convolutioncited by 55
  5. 2021Decoupled Dynamic Filter Networkscited by 52
  6. 2023Adaptive Rotated Convolution for Rotated Object Detectioncited by 51
  7. 2021Revisiting Dynamic Convolution via Matrix Decompositioncited by 35
  8. 2022Reversible Column Networkscited by 32
  9. 2022PointConvFormer: Revenge of the Point-based Convolutioncited by 29
  10. 2021MicroNet: Improving Image Recognition with Extremely Low FLOPscited by 26
  11. 2024Residual Quantization with Implicit Neural Codebookscited by 24
  12. 2025Frequency Dynamic Convolution for Dense Image Predictioncited by 19
38 more
  1. 2021ELSA: Enhanced Local Self-Attention for Vision Transformercited by 18
  2. 2024DyFADet: Dynamic Feature Aggregation for Temporal Action Detectioncited by 18
  3. 2021Relational Self-Attention: What's Missing in Attention for Video Understandingcited by 15
  4. 2022AGConv: Adaptive Graph Convolution on 3D Point Cloudscited by 13
  5. 2022RRSR:Reciprocal Reference-based Image Super-Resolution with Progressive Feature Alignment and Selectioncited by 12
  6. 2023Decomposed Guided Dynamic Filters for Efficient RGB-Guided Depth Completioncited by 10
  7. 2021Dynamic Fusion Network for RGBT Trackingcited by 9
  8. 2023RefConv: Re-parameterized Refocusing Convolution for Powerful ConvNetscited by 9
  9. 2022TVConv: Efficient Translation Variant Convolution for Layout-aware Visual Processingcited by 7
  10. 2024KernelWarehouse: Rethinking the Design of Dynamic Convolutioncited by 6
  11. 2021Collaboration of Experts: Achieving 80% Top-1 Accuracy on ImageNet with 100M FLOPscited by 5
  12. 2023PSLT: A Light-weight Vision Transformer with Ladder Self-Attention and Progressive Shiftcited by 5
  13. 2023KernelWarehouse: Towards Parameter-Efficient Dynamic Convolutioncited by 4
  14. 2021DRPN: Making CNN Dynamically Handle Scale Variationcited by 3
  15. 2024Encoding Semantic Priors into the Weights of Implicit Neural Representationcited by 3
  16. 20233D Human Pose Lifting with Grid Convolutioncited by 2
  17. 2023Dynamic Mobile-Former: Strengthening Dynamic Convolution with Attention and Residual Connection in Kernel Spacecited by 2
  18. 2023Evolution: A Unified Formula for Feature Operators from a High-level Perspectivecited by 2
  19. 2023A Feature Reuse Framework with Texture-adaptive Aggregation for Reference-based Super-Resolutioncited by 2
  20. 2024NOAH: Learning Pairwise Object Category Attentions for Image Classificationcited by 2
  21. 2024Conv-INR: Convolutional Implicit Neural Representation for Multimodal Visual Signalscited by 2
  22. 2025Test-Time Training for Speech Enhancementcited by 2
  23. 2022Lightweight Structure-Aware Attention for Visual Understandingcited by 1
  24. 2023Dynamic Shuffle: An Efficient Channel Mixture Methodcited by 1
  25. 2025Adaptive Blind Super-Resolution Network for Spatial-Specific and Spatial-Agnostic Degradationscited by 1
  26. 2021Dynamic Normalizationcited by 0
  27. 2021Shift-and-Balance Attentioncited by 0
  28. 2021Compact Multi-level Sparse Neural Networks with Input Independent Dynamic Reroutingcited by 0
  29. 2022Dual Complementary Dynamic Convolution for Image Recognitioncited by 0
  30. 2024SGDM: Static-Guided Dynamic Module Make Stronger Visual Modelscited by 0
  31. 2024Task-Aware Dynamic Transformer for Efficient Arbitrary-Scale Image Super-Resolutioncited by 0
  32. 2024Puppet-CNN: Continuous Parameter Dynamics for Input-Adaptive Convolutional Networkscited by 0
  33. 2024Adaptive High-Pass Kernel Prediction for Efficient Video Deblurringcited by 0
  34. 2025Attentive Convolution: Unifying the Expressivity of Self-Attention with Convolutional Efficiencycited by 0
  35. 2026cMoLLM at Scale: Horizontal Scaling Laws for Mixture-of-LLMscited by 0
  36. 2026Core-KAN: Continuous Vision Kernels with Kolmogorov-Arnold Networkscited by 0
  37. 2026Overcoming Kernel Redundancy for Scaling Logic Gate Networkscited by 0
  38. 2026MCL: Meta Convolution Layercited by 0