A fork of a subfolder from Nvidia DeepLearningExamples

Adds valgrind option. To run the server with Valgrind profiling run the launch_server.sh script with --valgrind argument