LLaMA_MPS by jankais3r supports the community trained 7B alpaca model.

https://github.com/jankais3r/llama_mps


The 13B model also works with slight modification. We are getting somewhere nice!

https://github.com/jooray/alpaca-lora-13b/