@calebmkim collected results for several neural networks generated from TVM for his SRC paper. Let's add them to the repository and separate them into "large" and "small" with the latter ones being a part of the default benchmark suite.