mlsanityData to 7 Oct 2026

Mobile-Former: Bridging MobileNet and Transformer

arXiv 2108.05895 ↗First posted 12 Aug 2021cs.CV cs.LGcited by 234

Not on the map. Why?

Abstract (arXiv)

We present Mobile-Former, a parallel design of MobileNet and transformer with a two-way bridge in between. This structure leverages the advantages of MobileNet at local processing and transformer at global interaction. And the bridge enables bidirectional fusion of local and global features. Different from recent works on vision transformer, the transformer in Mobile-Former contains very few tokens (e.g. 6 or fewer tokens) that are randomly initialized to learn global priors, resulting in low computational cost. Combining with the proposed light-weight cross attention to model the bridge, Mobile-Former is not only computationally efficient, but also has more representation power. It outperforms MobileNetV3 at low FLOP regime from 25M to 500M FLOPs on ImageNet classification. For instance, Mobile-Former achieves 77.9% top-1 accuracy at 294M FLOPs, gaining 1.3% over MobileNetV3 but saving 17% of computations. When transferring to object detection, Mobile-Former outperforms MobileNetV3 by 8.6 AP in RetinaNet framework. Furthermore, we build an efficient end-to-end detector by replacing backbone, encoder and decoder in DETR with Mobile-Former, which outperforms DETR by 1.1 AP but saves 52% of computational cost and 36% of parameters.

Key facts. Cited by 234 papers in the map (data through 7 Oct 2026). Builds on 47 of its 55 references that are in the map. First posted on arXiv 12 Aug 2021.

Builds on

47 of 55 references.

References · 47

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  4. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  5. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,082
  6. 2020End-to-End Object Detection with Transformerscited by 7,384
  7. 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,415
  8. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  9. 2016Feature Pyramid Networks for Object Detectioncited by 5,021
  10. 2018MobileNetV2: Inverted Residuals and Linear Bottleneckscited by 4,773
  11. 2017Squeeze-and-Excitation Networkscited by 4,763
  12. 2020Training data-efficient image transformers & distillation through attentioncited by 4,560
35 more
  1. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,827
  2. 2019Searching for MobileNetV3cited by 1,984
  3. 2021Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutionscited by 1,960
  4. 2021MLP-Mixer: An all-MLP Architecture for Visioncited by 1,441
  5. 2017Random Erasing Data Augmentationcited by 1,235
  6. 2017ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devicescited by 1,194
  7. 2018ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Designcited by 1,132
  8. 2021Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNetcited by 1,096
  9. 2019EfficientDet: Scalable and Efficient Object Detectioncited by 1,018
  10. 2021CvT: Introducing Convolutions to Vision Transformerscited by 946
  11. 2021CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classificationcited by 806
  12. 2021Going deeper with Image Transformerscited by 759
  13. 2017Focal Loss for Dense Object Detectioncited by 749
  14. 2021CSWin Transformer: A General Vision Transformer Backbone with Cross-Shaped Windowscited by 526
  15. 2021ConViT: Improving Vision Transformers with Soft Convolutional Inductive Biasescited by 484
  16. 2021LeViT: a Vision Transformer in ConvNet's Clothing for Faster Inferencecited by 475
  17. 2019GhostNet: More Features from Cheap Operationscited by 410
  18. 2021Early Convolutions Help Transformers See Bettercited by 395
  19. 2021Bottleneck Transformers for Visual Recognitioncited by 364
  20. 2019Dynamic Convolution: Attention over Convolution Kernelscited by 355
  21. 2021DeepViT: Towards Deeper Vision Transformercited by 288
  22. 2019CondConv: Conditionally Parameterized Convolutions for Efficient Inferencecited by 273
  23. 2021Co-Scale Conv-Attentional Image Transformerscited by 224
  24. 2021Scaling Local Self-Attention for Parameter Efficient Visual Backbonescited by 220
  25. 2021Conformer: Local Features Coupling Global Representations for Visual Recognitioncited by 164
  26. 2021How to represent part-whole hierarchies in a neural networkcited by 123
  27. 2019MixConv: Mixed Depthwise Convolutional Kernelscited by 108
  28. 2019AdderNet: Do We Really Need Multiplications in Deep Learning?cited by 82
  29. 2020Dynamic ReLUcited by 55
  30. 2020Rethinking Bottleneck Structure for Efficient Mobile Network Designcited by 54
  31. 2020WeightNet: Revisiting the Design Space of Weight Networkscited by 50
  32. 2020Model Rubik's Cube: Twisting Resolution, Depth and Width for TinyNetscited by 49
  33. 2021ConTNet: Why not use convolution and transformer at the same time?cited by 38
  34. 2021Refiner: Refining Self-attention for Vision Transformerscited by 38
  35. 2019Butterfly Transform: An Efficient FFT Based Neural Architecture Designcited by 16

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 8; 2022: 57; 2023: 64; 2024: 57; 2025: 36; 2026: 12.

234 papers.

Sort:

By area:

  1. 2021MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformercited by 636
  2. 2022Multimodal Learning with Transformers: A Surveycited by 399
  3. 2022EfficientFormer: Vision Transformers at MobileNet Speedcited by 240
  4. 2022TinyViT: Fast Pretraining Distillation for Small Vision Transformerscited by 232
  5. 2022UniFormer: Unifying Convolution and Self-attention for Visual Recognitioncited by 206
  6. 2021PeCo: Perceptual Codebook for BERT Pre-training of Vision Transformerscited by 204
  7. 2023EfficientViT: Memory Efficient Vision Transformer with Cascaded Group Attentioncited by 172
  8. 2023FLatten Transformer: Vision Transformer using Focused Linear Attentioncited by 170
  9. 2021BEVT: BERT Pretraining of Video Transformerscited by 169
  10. 2022Separable Self-attention for Mobile Vision Transformerscited by 163
  11. 2023Run, Don't Walk: Chasing Higher FLOPS for Faster Neural Networkscited by 129
  12. 2023RepViT: Revisiting Mobile CNN From ViT Perspectivecited by 123
38 more
  1. 2022EdgeNeXt: Efficiently Amalgamated CNN-Transformer Architecture for Mobile Vision Applicationscited by 121
  2. 2022Rethinking Vision Transformers for MobileNet Size and Speedcited by 121
  3. 2022ViTAEv2: Vision Transformer Advanced by Exploring Inductive Bias for Image Recognition and Beyondcited by 117
  4. 2022MobileOne: An Improved One millisecond Mobile Backbonecited by 109
  5. 2024MobileNetV4 -- Universal Models for the Mobile Ecosystemcited by 98
  6. 2022Protecting Celebrities from DeepFake with Identity Consistency Transformercited by 95
  7. 2023MedViT: A Robust Vision Transformer for Generalized Medical Image Classificationcited by 89
  8. 2022TopFormer: Token Pyramid Transformer for Mobile Semantic Segmentationcited by 82
  9. 2022Conv2Former: A Simple Transformer-Style ConvNet for Visual Recognitioncited by 81
  10. 2023Full Stack Optimization of Transformer Inference: a Surveycited by 79
  11. 2022Fast Vision Transformers with HiLo Attentioncited by 78
  12. 2023SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applicationscited by 66
  13. 2022Next-ViT: Next Generation Vision Transformer for Efficient Deployment in Realistic Industrial Scenarioscited by 65
  14. 2021Lite Vision Transformer with Enhanced Self-Attentioncited by 64
  15. 2023Omni Aggregation Networks for Lightweight Image Super-Resolutioncited by 62
  16. 2023SparseViT: Revisiting Activation Sparsity for Efficient High-Resolution Vision Transformercited by 60
  17. 2022MogaNet: Multi-order Gated Aggregation Networkcited by 59
  18. 2023Spatially-Adaptive Feature Modulation for Efficient Image Super-Resolutioncited by 56
  19. 2022A Transformer-Based Feature Segmentation and Region Alignment Method For UAV-View Geo-Localizationcited by 55
  20. 2022Hydra Attention: Efficient Attention with Many Headscited by 55
  21. 2024SHViT: Single-Head Vision Transformer with Memory Efficient Macro Designcited by 55
  22. 2024Rewrite the Starscited by 55
  23. 2023Scale-Aware Modulation Meet Transformercited by 49
  24. 2022MOAT: Alternating Mobile Convolution and Attention Brings Strong Vision Modelscited by 47
  25. 2023Adaptive Frequency Filters As Efficient Global Token Mixerscited by 45
  26. 2022MobileViTv3: Mobile-Friendly Vision Transformer with Simple and Effective Fusion of Local, Global and Input Featurescited by 44
  27. 2023Image as Set of Pointscited by 43
  28. 2022EfficientViT: Multi-Scale Linear Attention for High-Resolution Dense Predictioncited by 42
  29. 2023EdgeSAM: Prompt-In-the-Loop Distillation for SAMcited by 40
  30. 2022MSP-Former: Multi-Scale Projection Transformer for Single Image Desnowingcited by 39
  31. 2024ViTamin: Designing Scalable Vision Models in the Vision-Language Eracited by 39
  32. 2023HRDFuse: Monocular 360degDepth Estimation by Collaboratively Learning Holistic-with-Regional Depth Distributionscited by 37
  33. 2021Pale Transformer: A General Vision Transformer Backbone with Pale-Shaped Attentioncited by 36
  34. 2022Reduce Information Loss in Transformers for Pluralistic Image Inpaintingcited by 36
  35. 2023Rethinking Mobile Block for Efficient Attention-based Modelscited by 36
  36. 2023A Survey of Techniques for Optimizing Transformer Inferencecited by 35
  37. 2023DilateFormer: Multi-Scale Dilated Transformer for Visual Recognitioncited by 34
  38. 2022Expediting Large-Scale Vision Transformer for Dense Prediction without Fine-tuningcited by 33

50 most cited of 234 All 234 (JSON)