How do you troubleshoot Mendix giving the wrong ONNX random forest model output?
0
I have developed an onnx model via python/scikitlearn and have converted using code found on the onnx site and in the Mendix github site. Regardless, my test dataset is running through the onnx model and the incorrect label is being applied for about a third of the test data. It would be helpful if the probabilities were displayed, but they are blank. The dataset is the iris data set found in scikit learn.