docs: update installation instructions for CAM with FTorch

[TomMelt]

Hi @jatkinson1000 , here's my take on the build instructions.

There's no major differences. I just tried to align it with the steps I actually used to build the code. Biggest one is probably that I removed running module load libtorch because it conflicts with the base ncarenv/23.06 module, and we don't actually need to load it. I did add an explanation though as to why it's been removed.

One thing I realised in the process though is the description for the xml vars gw_convect_dp_ml_net_path and gw_convect_dp_ml_norms possibly needs updating. Am I right in thinking that even if gw_convect_dp_ml is set to .false., if gw_convect_dp_ml_compare is .true. do we not run both models anyway. For this README I took the definitions of the variables from bld/namelist_files/namelist_definition.xml.

Here's the relevant section from bld/namelist_files/namelist_definition.xml.

<entry id="gw_convect_dp_ml_norms" type="char*132" input_pathname="abs" category="gw_drag"
       group="gw_drag_nl" valid_values="" >
  Absolute filepath to the deep convection gravity wave normalisation weights (NetCDF)
  used when `gw_convect_dp_ml` is set to `.true.`.
</entry>

<entry id="gw_convect_dp_ml_compare" type="logical" category="gw_drag"
       group="gw_drag_nl" valid_values="" >
  Whether or not to run a piggybacking comparison of the ML deep convection gravity·
  waves to the original scheme. Only one scheme will be used to advance the simulation
  as dictated by `gw_convect_deep_ml`.
Default: .false.
</entry>

Let me know what you think 👍

(I wouldn't normally tag all of you in the review, I just thought you might like to see the build instructions. Probs just @j-emberton and @jatkinson1000 that need to approve)

[j-emberton]

This looks good Tom. The changes make sense and add to readability and comprehension me.

[jatkinson1000]

Thanks @TomMelt
I have left a few suggestions/comments.

My overall impression after reading this is that the ### Obtaining FTorch section now feels quite empty. I wonder if there is some reshuffling that could make this easier to follow.

I'm happy for this to be my responsibility, as I realise this is a PR onto my PR, so the buck eventually stops with me.

I also note that on reading it is some time since I wrote these instructions (due to the delays in receiving a review (not from yourself!!)) and there are things I would do differently after working on the project for some time.
So, I'm keen to get this merged and then have a second pass through myself, so feel free to comment in #5 to make some of the points tasks for me to do.

[jatkinson1000]

I think this is at a point where we should merge, and make further adjustments in the host PR.

Thanks @TomMelt