Comment by fl0id

you can load the safetensors with ollama, you just have to provide a modelfile. or wait for someone to do it. It will in theory also quantize it for you, as I guess most ppl cannot load a 129 GB model...