QT server frontend

[tohurtv]

I just wrote over the past couple days a QT frontend to Llama.cpp's server! This is the first desktop GUI I have ever written and mainly wrote it for my uses but I figure others out there could use this too if you have alot of models you like to run with pure Llama.cpp as from my experience pure Llama.cpp is way faster then Koboldcpp as that runs Llama.cpp directly in python as far as I can tell and you can feel the python overhead and most certainly faster then Llama-cpp-python. I did write the GUI in pyqt5 but it uses the compiled server binary and it runs on its own process so not much overhead regarding python and it shows for me, I get about twice as many tokens per second compared to Koboldcpp and llama-cpp-python.

I do have few planned things I going to add over the new few days to weeks as I learn how to.

You can check it out at my github repo here llama.cpp-qt Any and all contributions are welcomed if anyone finds it useful.