uncompatibility between google-ar/arcore-ios SDK and Firebase/MLModelInterpreter from Firebase MLKit custom models

[bzok2]

1. installing Arcore/augmented face and firebase/MLmodelInterpreter via cocoapods
2. go through the firebase mlkit customs model quickstart https://github.com/firebase/quickstart-ios/tree/master/mlmodelinterpreter/MLModelInterpreterExample
3. launch the app on a real device

The app work well whit firebase custom model tensor flow lite and google arcore ios separately. But not with both, when i add arcore/augmented face firebase mlkit crash at the model interpreter
with a
EXC_BAD_ACCESS. from lldb

navigator console output
#0 0x0000000100405bc0 in tflite::GetRegistrationFromOpCode(tflite::OperatorCode const*, tflite::OpResolver const&, tflite::ErrorReporter*, TfLiteRegistration const**) ()
#1 0x00000001003f3dcc in tflite::InterpreterBuilder::BuildLocalIndexToRegistrationMapping() ()
#2 0x00000001003f533c in tflite::InterpreterBuilder::operator()(std::__1::unique_ptr<tflite::Interpreter, std::__1::default_deletetflite::Interpreter >*, int) ()
#3 0x0000000101132614 in TFL_NewInterpreter ()
#4 0x000000010112fa70 in ::-[TFLInterpreter initWithModelPath:options:error:](NSString *, TFLInterpreterOptions *, NSError **) at /Volumes/cristoff disk/Cristoff/Cristoff_Freight_2020/Cristoff/Pods/TensorFlowLiteObjC/tensorflow/lite/experimental/objc/sources/TFLInterpreter.mm:105
#5 0x000000010112f6e0 in ::-[TFLInterpreter initWithModelPath:error:](NSString *, NSError **) disk/Cristoff/Cristoff_Freight_2020/Cristoff/Pods/TensorFlowLiteObjC/tensorflow/lite/experimental/objc/sources/TFLInterpreter.mm:59
#6 0x0000000100c14854 in __32-[FIRModelInterpreter loadModel]_block_invoke ()

and in the debugger console i have
[Firebase/MLKit][I-MLK015000] Input Firebase App can not be nil: statsLoggerForApp:clearcutLogger:loggerType:userDefaults:tfLiteVersion:processInfo:

Also the app work pretty well on simulator but crash on a real device.

i'm using MacOS catalina and Xcode 11.3 iOS 13

[rsfuller]

We are aware of this issue. Thank you for the detailed report!