convert to onnx and then convert to CoreML , but the prediction is wrong

[MarcSteven]

[Shuwrood-SSW]

any way to deal with it ? I have the same issue now.

[MarcSteven]

@Shuwrood-SSW You can check my blog about the issue and recheck the model. Prediction wrong to check You must ensure the same image processing on the model and then do it via CoreML. And Even though you did the image Processing on the PyTorch model, but you convert it to onnx and then it doesn't include image processing. You don't need to write code in your project directly, you can use CoreMLTools to verify your modelCoreMLTools And you have image dataSet, you can use Create ML to train and it will generate the model directly(the size of model is small)

[bhushan23]

@MarcSteven can you provide more details? Can you share ONNX model and input for which you observe incorrect prediction?

[MarcSteven]

onnx model @bhushan23 FYI

[MarcSteven]

Hope it is useful to modify