I really wish I could do it like that. But for what I'm trying to do it's necessary for me to build upon some specific NVIDIA CUDA images that have ubuntu 16.04 as the most recent version.
That Dockerfile that I linked to contains the code used to create one the images that can be pulled in the tutorial. While translating it from Debian stretch to Ubuntu 16, I noticed the error which persisted when I retraced my steps, running the test right after the "make" command in the original code.
That Dockerfile that I linked to contains the code used to create one the images that can be pulled in the tutorial. While translating it from Debian stretch to Ubuntu 16, I noticed the error which persisted when I retraced my steps, running the test right after the "make" command in the original code.