I'm wondering if it is possible to run the openvino model server in a docker container without the --privileged flag. I am using the original NCS(not the NCS2), but I'm not sure if that makes a difference. Currently, I'm unable to mount the device into a container as the device ID changes whenever a neural network model is loaded onto the stick. I saw that there was support for this with the NCSDK for the NCS1, and I wanted to know if there is a way to do this without it(the NCSDK).
Using the Intel Neural Compute Stick with a docker container requires the container to be ran in privileged mode as it gives the container access to the host's devices. However, there may be another way to configure the docker container to have access to specific devices (--device) but I have not tried this myself. I recommend posting your question on the Docker community forums.
Hope this helps!