In VLC 4.0, we'll add RAVU, NNEDI3 and a few other to upscale.
I'm not sure open source really applies to an ML model, I suppose you can obfusticate it to "close it". Regardless, prior art here is open source (http://avisynth.nl/index.php/Nnedi3).
Or is there a post-processing solution to upscale video for VLC?
I've seen some impressive neural net solutions for upscaing, but they haven't been open source.