Comment by ljouhet

Something like

    ollama run hf.co/ngxson/GLM-4.7-Flash-GGUF:Q4_K_M

It's really fast! But, for now it outputs garbage because there is no (good) template. So I'll wait for a model/template on ollama.com