3 ms·They are using Ollama which is based on llama.cpp; florence is not supported on that backend.by Eisenstein 2y agoThey are using Ollama which is based on llama.cpp; florence is not supported on that backend.