Comment by albertgoeswoof

What inference performance are you getting on this with llama?

How long would it take to recoup the cost if you made the model available for others to run inference at the same price as the big players?