[Feature request] Add PaddleOCR standalone executable for simpler usage

[timminator]

Hi there!
It's me again :-)

The last week while you (@niksedk) were working on implementing PaddleOCR into Subtitle Edit I worked on a new little site project of mine.
I was curious if I could create a standalone executable of paddleocr that could be run without setting up python and installing the required packages.
It took quite a bit of tinkering but I managed to get it working, my repository can be found here. I've created a CPU and GPU version.

It should work exactly like the command line version before.
I also fixed the bug in this release that prevented implementing the batch mode. :-)

So in theory Subtitle Edit could just prompt the user to download this version now.

And don't be confused when it takes a longer time to start when running it the first time. This is the normal behavior. Afterwards it runs normally.

We would only need to detect if an Nvidia GPU is present in the users system to decide which version should be used. I think we could do this by checking for example if the command "nvidia-smi" exists.

Edit: The GPU version falls back to using the CPU if no CUDA device is present so it also works in this case, just for your info.