Need documentation for possible training data inputs

[gnperdue]

We need to provide documentation for the variables we are logging - unpacked names and a basic explanation of the values. This wiki page:

https://github.com/fermilab-accelerator-ai/meetings/wiki/training_inputs

is a good place to keep the information. @bschup is "in charge" of the effort but @jasonstjohn and Kiyomi should help as available. I plan on checking with @r-keller to see if any notes from the summer could seed the effort.