App Crash Issue with TensorFlow Lite Model in MIT App Inventor's Personal Image Classifier Extension

I'm attempting to create an app in MIT App Inventor using the Personal Image Classifier (PIC) extension to classify traffic signs. For this purpose, I created a TensorFlow model and converted it to TensorFlow Lite format. However, when I upload the .tflite model to the PIC extension and run the app via a QR code, the app immediately crashes. It seems that the PIC extension may not support TensorFlow Lite models effectively, despite my efforts to use TensorFlow for better classification performance. Is it not possible to use TensorFlow Lite models with the Personal Image Classifier extension in App Inventor?