Thank you for sharing. I don't think I saw similar problems. May I ask - what's the version of the llama cpp which you are running? I wonder if there was some breaking change - I didn't pin specific library version in my dockerfile so it's just latest. Also I will share public repo with the docker image this week.