效果立竿见影！YOLOv5引入 ScConv，mAP 提升 2.11（+7.9%）

分析如下：

文章目录

• 移植
• • 创建ultralytics/cfg/models/v5/yolov5-C2f-SCcConv.yaml
  • 修改ultralytics\nn\tasks.py
  • 修改ultralytics/nn/modules/__init__.py
  • 修改ultralytics\nn\extra_modules\block.py
  • 拉起训练

论文：SCConv: Spatial and Channel Reconstruction Convolution for Feature
Redundancy

移植

创建ultralytics/cfg/models/v5/yolov5-C2f-SCcConv.yaml

# Ultralytics YOLO 🚀,AGPL-3.0 license # YOLOv8 object detection model with P3-P5 outputs. For Usage examples seehttps://docs.ultralytics.com/tasks/detect # Parameters #nc:80 # number of classes