How do I retrain "Primary_Detector_Nano" for my images

Can somebody point me in the direction to learn how to retrain the resnet10.caffemodel for the “Primary_Detector_Nano” file and the build the engine for my images

This model is only for demo. We will not open network
You can check our TLT tool: https://developer.nvidia.com/transfer-learning-toolkit

What does " We will not open network" mean.
Will the TLT tool kit allow me to get the same performance as the Demos?

It’s an internal work in NVIDIA.
https://devtalk.nvidia.com/default/topic/1059008/deepstream-sdk/deepstream-sdk-4-0-deepstream-app-framework/post/5373748/#5373748