yolov7-opencv-dnn-cpp icon indicating copy to clipboard operation
yolov7-opencv-dnn-cpp copied to clipboard

使用opencv部署yolov7

Results 5 yolov7-opencv-dnn-cpp issues
Sort by recently updated
recently updated
newest added

对yolov5s进行轻量化设计,骨干网络用shufflenetV2来进行代替,其他保持不变,代码中的种类已经修改 报错位置: ![image](https://github.com/UNeedCryDear/yolov7-opencv-dnn-cpp/assets/91108548/254e7dee-05ad-48f2-a575-7dd3ade76fb7) 修改后的网络输出: ![image](https://github.com/UNeedCryDear/yolov7-opencv-dnn-cpp/assets/91108548/b00f8619-ebf7-4c42-ab0f-f15e7cb03576) 原模型网络输出: ![image](https://github.com/UNeedCryDear/yolov7-opencv-dnn-cpp/assets/91108548/1f6a6b89-0cf4-4fcd-bfb5-3d6e20b5b5a3)

大佬您好,我是想了解一下,YOLOv7与YOLOv8获取模型输出后的后处理算法有什么区别吗,想要移植部署到华为昇腾的设备上,YOLOv8的模型推理输出为[1,8400,84],而YOLOv7的是[1,25200,85]。

![image](https://github.com/user-attachments/assets/05a73997-73b2-4424-ad46-90c4ad4b4e98) 大佬,为啥我直接拿yolov7.pt模型转onnx 后,代码在执行forward操作的时候速度很慢,请问一下可能是啥原因