We are considering CogVLM for our application and I have a few questions: * what scale are you running it at? Can this run in a production setting without blow

We are considering CogVLM for our application and I have a few questions: * what scale are you running it at? Can this run in a production setting without blowing up budget (GPU == $$$) * what usecases are you captioning for? any other insights/lessons learned you could share?