#using the llamavision model to realize the zero-shoot image tag

open the ollama first

Usage: python detect_tag.py /path/to/image_or_folder model_name

'config white list must list as single word.(use 'parking' instead of 'parking lot') so that the proggram can detect the items correctly'