Update Local Llama llama.cpp

[NigelNelson]

Context

The Local Llama tutorial uses the Llama.cpp repo, and previously breaking changes were introduced. In these changes, the model format was changed from GGML -> GGUF. These changes are now stable and there is enough GGUF support that the tutorial should be updated to stay relevant.

Description

Updates:

• Remove hack needed for ARM64 support (Updated llama.cpp includes this fix now)
• Correct all mentions and uses of GGML and use the GGUF equivalent.
• Update the ./server output snippet

[jjomier]

LGTM!