Developers looking to gain a better understanding of machine learning inference on local hardware can fire up a new llama ...