Mark Ward

### What is the issue? When calling the chat completion for the following vision models, I get the following prompt eval count value. llava:7b-v1.6-mistral-q4_0 -> prompt eval count `1` llava-llama3:8b-v1.1-q4_0...