Converting from Keras Model - Output Type

## ❓Question

- If this is a question about the Core ML Frame work or Xcode, please ask your question in the Apple Developer Forum: https://developer.apple.com/forums/
Hey,

Three years ago I have converted a classifier model built in keras (in tensorflow 1) with coremltools version 3.3.
I used the function ct.converters.keras.convert.
The output type specified from the model was of class VNClassificationObservation.

Now I have retrained the model and improved it using tensorflow 2.
Working with recent coremltools versions I can only use the coremltools.convert function. (since coremltools.converters.keras doesn't exist).
The output type of the model converted is multi array.

Is there a way to convert the new keras model so the output type will be 'labelProbability" of class VNClassificationObservation instead of multiarray?

Thanks

Start with the current coremltools.convert path for TensorFlow 2 and compare it with the older ct.converters.keras.convert behavior described in the issue. Determine whether the converted classifier can expose labelProbability for VNClassificationObservation rather than a multiarray; the work is done when a supported conversion path or limitation is clearly established.

Written by the indexing model from the issue text.