Llama transformer walkthrough

[bdzwillo]

Hello,

I made a llama2 transformer overview based on code examples available
here on github: https://github.com/bdzwillo/llama_walkthrough

This is based on a presentation I gave to my colleagues. This uses a
cloned version of the very small llama2.c https://github.com/karpathy/llama2.c
implementation from Andrej Karpathy with some added command line options.

It served very well for educational purposes, because the source code
is self contained and less complex than that of llama.cpp. So perhaps
this might be of interest to someone looking for concepts behind the code.