Object detection model can't be used in C++

[ly0303521]

Have I written custom code (as opposed to using a stock example script provided in MediaPipe)

None

OS Platform and Distribution

ubuntu20

Python Version

3.11

MediaPipe Model Maker version

0.2.1.3

Task name (e.g. Image classification, Gesture recognition etc.)

Object detection

Describe the actual behavior

model maker train is successful, but failed in C++ inference

Describe the expected behaviour

the trained model can inference in C++

Standalone code/steps you may have used to try to get what you need

In the begining, I use mediapipe follow the example "mediapipe/graphs/object_detection", its inferenced in C++, the model is "ssdlite_object_detection.tflite". 
Now, I tried retrain the object detection model, follow "https://developers.google.com/mediapipe/solutions/customization/object_detector", I found the models are 
MOBILENET_V2
MOBILENET_V2_I320
MOBILENET_MULTI_AVG
MOBILENET_MULTI_AVG_I384,
these are different from "ssdlite_object_detection.tflite".
I choosed "MOBILENET_V2" to try, training is successful, but I found MOBILENET_V2 model can't replace ssdlite_object_detection

Other info / Complete Logs

No response

[ly0303521]

#5308 more details in this issue I raised 1 month ago

[schmidt-sebastian]

Have you tried the model in MediaPipe Studio?

https://mediapipe-studio.webapps.google.com/studio/demo/object_detector