YOLOv5 Improvement Series (23) - MobileViTv2 to replace the backbone network (an efficient separable self-attention mechanism for mobile vision Transformer) - Code World

YOLOv5 Improvement Series (23) - MobileViTv2 to replace the backbone network (an efficient separable self-attention mechanism for mobile vision Transformer)

Enterprise 2023-09-03 09:17:53 views: null

【YOLOv5 Improvement Series】Preliminary review:

YOLOv5 Improvement Series (0) - Evaluation and Analysis of Important Performance Indicators and Training Results

YOLOv5 improvement series (1) - adding SE attention mechanism

Guess you like

Origin blog.csdn.net/weixin_43334693/article/details/132428203

YOLOv5 Improvement Series (23) - MobileViTv2 to replace the backbone network (an efficient separable self-attention mechanism for mobile vision Transformer)

YOLOv5 improvement series (24) - MobileViTv3 that replaces the backbone network (a further upgrade of the mobile lightweight network)

YOLOv5 Improvement Series (21) - RepViT to replace the backbone network (Tsinghua ICCV 2023|The latest open source mobile terminal ViT)

YOLOv5 Improvement Series (19) - Swin TransformerV1 to replace the backbone network (ViT model with smaller parameters)

Yolov5/Yolov7 improvement --- attention mechanism: Polarized Self-Attention, the effect spikes CBAM, SE

CVPR2023 Plug and Play Series | An Efficient and Lightweight Self-Attention Mechanism Helps Image Restoration Network Win SOTA!

Self-attention mechanism and transformer

YOLOV5 improvement: CVPR2023: join the EfficientViT backbone: memory access efficient ViT with cascade group attention

Improving the YOLOv5 series: Combining CVPR2021: Multi-head attention Efficient Multi-Head Self-Attention

YOLOV5/YOLOV8 improvement: CVPR23: join the ConvNeXt V2 backbone, effectively increase points

YOLOv8/YOLOv7/YOLOv5/YOLOv4/Faster-rcnn series algorithm improvement [NO.71] attention mechanism Expectation-Maximization Attention (EMA module)

YOLOv5 Improvement Series (20) - Add BiFormer Attention Mechanism (CVPR2023|Small Goal Increase Point Artifact)

Revolutionizing Object Detection with YOLOv5 and MobileViTs Backbone Networks: The Future of Efficient and Accurate AI Vision

Vision Transformer (ViT): Analysis of image segmentation, image block embedding, category marking, QKV matrix and self-attention mechanism

YOLOv8/YOLOv7/YOLOv5/YOLOv4/Faster-rcnn series algorithm improvement [NO.67] Add attention mechanism MHSA

YOLOv8/YOLOv7/YOLOv5/YOLOv4/Faster-rcnn series algorithm improvement [NO.72] attention mechanism Linear Context Transform Block (LCT module)

[Lightweight] Practical improvement of YOLOv8 | Replacing the backbone network Backbone with GhostnetV2

YOLOv5 Improvement | 2023 Attention Chapter | iRMB Inverted Residual Block Attention Mechanism (Lightweight Attention Mechanism)

[Self-attention neural network] Swin Transformer network

[YOLOv8/YOLOv7/YOLOv5/YOLOv4/Faster-rcnn series algorithm improvement NO.65] CVPR 2023 | Tsinghua team's plug-and-play network architecture—Slide-Transformer

YOLOV5 adds attention mechanism, taking ECAttention as an example (series)

Yolov5/Yolov7 improvement --- attention mechanism: ShuffleAttention, ECA, EffectiveSE, SE

Yolov5/Yolov7 improvement---attention mechanism: ICASSP2023 EMA is based on efficient multi-scale attention of cross-space learning, the effect is better than ECA, CBAM, CA | small target rises significantly

Add feature pyramid and self-attention mechanism to the network

YOLOv8/YOLOv7/YOLOv5/YOLOv4/Faster-rcnn series algorithm improvements [NO.83] Change the backbone feature extraction network Backbone to RevCol

YOLOv5 improvement series (13) - replace the activation function of SiLU, ReLU, ELU, Hardswish, Mish, Softplus, AconC series, etc.

Self-Attention self-attention mechanism

Translation: Detailed illustration of Transformer's multi-head self-attention mechanism Attention Is All You Need

Trying to help you understand the essence of transformer attention mechanism (Self-Attention) in one article

Improve YOLOv8 | Backbone Network | YOLOv8 Replaces ShuffleNetv2 of the Backbone Network | "ShuffleNet v2: A Practical Guide to Efficient Convolutional Neural Network Architecture Design"

Recommended

Ranking

Reason Codes enhanced accounting documents

The first test case appium

Force command and dispatch emergency communication plan

Interview - What is concurrent programming

21 classic Python interview questions [recommended collection]

[Dahua Data Structure-Introduction to Data Structure ①]

Spring @ConfigurationProperties

Why do we want to broadcast live on WeChat public account? What are the advantages?

How to bring up the toolbar under the desktop of Apple laptop

Dialog: Manually enter information

Daily

More

2024-05-21(35)

2024-05-20(5)

2024-05-19(0)

2024-05-18(31)

2024-05-17(6)

2024-05-16(23)

2024-05-15(5)

2024-05-14(9)

2024-05-13(8)

2024-05-12(28)