Ollama Modelfile Generator CLI options receive output file CANCELED receive model to be used (default: llama3) Initial Setup create tests Model Generation Prompt improve prompt to llama3