Customising existing model

Hello everyone. I’m new to the world of object detection and currently have a university project which requires me to detect motorbikes through an object detection system. I am permitted to use an existing system but edit accordingly. I basically want to use an SSDmobile net with TensorFlow eventually onto a Raspberry Pi and use it on my own dataset including videos. Maybe convert this to TensorFlow Lite. I am confused about how I can edit an existing model. Could anyone please advise? Thank you in advance.

© Copyright 2013-2020 Analytics Vidhya