Using trained model as function in AIP logic

Ask the Community
1

I am trying to use a model I’ve deployed from Jupyter notebook as function and use it from the AIP logic flow.

Model was successfully deployed and I’ve created a function to run the model.

I already have the object which has matching properties for the model input.

Here’s the steps I took in AIP logic.

  1. Take the Input from the prepared Object
  2. Call function created from the model
  3. map all the matching properties for the model input

but when I tried to run the logic I would get this message:

ServiceException: INVALID_ARGUMENT (LogicExecutor:FunctionHasInvalidOutput): {unstructuredErrorContent=InvalidOutputError{details: InvalidOutputDetails{value: UntypedValueConversionFailedWrapper{value: UntypedValueConversionFailed{value: Optional[false]}}}}}

and argument:

{
“unstructuredErrorContent”: “{“details”:{“type”:“untypedValueConversionFailed”,“untypedValueConversionFailed”:{“value”:false}}}”
}

I think this might be a pretty simple issue to solve, but I don’t know how to. Is there anyone who could help me with this?

2

It seems like the issue could be your Function return type, what are you trying to return? Would you be able to share the code of your function maybe?

Do you get the same error if you run it through workshop?

1 Like