This example with a LeNet model trained on the fashionMnist dataset should work correctly with the Image Processing DSL It includes conversion to the Grayscale image preprocessor added in the #202
