   Llama.prg sample
   ----------------

      There are two executable files: llama_exsrv.exe and test_exllm.exe.

      The test_exllm.exe is a small Harbour + HwGUI program, compiled with 32-bit Borland C 5.5.
   It launches the llama_exsrv.exe as a background process and keeps connection with this process
   via special text files, using a technology from my GuiServer project.

      The llama_exsrv.exe is a Harbour program, compiled with 64-bit MSVC 2019. It uses a llama.lib -
   a library, built with Llama.prg project. This program works directly with AI local models. It may be
   launched from a command line with a path to model as a parameter - in this case it provides a simple
   console interface to communicate with a model. But the main mission is to serve as background process,
   a mediator between your Harbour program, GUI or console, with AI models.

      Models.ini must be edited before using the test_exllm.exe, fill it with full paths to your models.

   Source files for llama_exsrv.exe (llama.lib is needed also):
      llama_exsrv.prg, hbextsrv.prg, fconnect.prg

   Source files for test_exllm.exe:
      test_exllm.prg, hbextcli.prg,  fconnect.prg

   hbextsrv.prg, hbextcli.prg, fconnect.prg are a part of Guiserver project.

   test_exllm.hwprj, llama_exsrv.hwprj are HwBuilder project files to build test_exllm.exe and llama_exsrv.exe.

   Links:
      Guiserver: http://www.kresin.ru/en/guisrv.html
      Llama.prg: https://gitflic.ru/project/alkresin/llama_prg
