Running a PyTorch model on Raspberry pi and Nvidia Jetson Nano

I've trained an image segmentation algorithm based on pyTorch and I would like to make inference on the Raspberry pi and Nvidia Jetson Nano. I can give you trained model in standard format and can give you prediction script in python, but first i would like to run the same python Script in the Raspberry pi and Nvidia Jetson Nano. And later running in optimised way, might be in C++ or LibTorch etc. Please write me message for more details.