关于cv2.dnn.readNetFromONNX(path)就报ERROR during processing node with 3 inputs and 1 outputs的解决过程【独家发布】

问题是什么：

如图，对vgg16使用opencv的dnn模块进行推理时出现错误。

 

错误的详细日志为：

[ERROR:[email protected]] global D:\a\opencv-python\opencv-python\opencv\modules\dnn\src\onnx\onnx_importer.cpp (1021) cv::dnn::dnn4_v20220524::ONNXImporter::handleNode DNN/ONNX: ERROR during processing node with 3 inputs and 1 outputs: [Conv]:(onnx_node!StatefulPartitionedCall/vgg16/conv_bn_two_relu_pool1/conv_bn_one_relu/conv2d/Conv2D) from domain='ai.onnx'
Traceback (most recent call last):
  File "D:/DLAI/qwenAILearn/StudyDNN/trainOnnx/test2.py", line 6, in <module>
    net = cv2.dnn.readNetFromONNX(path)  # 加载训练好的识别模型
cv2.error: OpenCV(4.6.0) D:\a\opencv-python\opencv-python\opencv\modules\dnn\src\onnx\onnx_i