Looking at nvcr.io/nvidia/deepstream:4.0.2-19.12-devel, it is missing the /usr/src/tensorrt directory as well as the /usr/lib/python2.7/dist-packages/uff which makes it impossible to build the objectDetector_SSD example in the container.
Let me give it a try quickly and get back to you soon.
This issue can reproduce on our environment.
It looks like there is no python TensorRT package in the container.
We will feedback to our internal team and update more information with you later.
Temporally, you can generate the uff file on the other platform first.
And run the sample with the pre-generated model directly.
Thanks and sorry for the inconvenience.