compress_classifier.py: add an option to load a model in serialized mode

By default, when we create a model we  wrap it with DataParallel to benefit
from data-parallelism across GPUs (mainly for convolution layers).

But sometimes we don't want the sample application to do this: for
example when we receive a model that was trained serially.
This commit adds a new argument to the application to prevent
the use of DataParallel.