Add Llama 8B model option

[jcabrero]

Adding a new model option which is Llama 7B. Consider whether adding a quantized version of Llama.cpp is a more efficient option for the current goal.