Post #2769490

I would pickup an inference GPU to run models locally. I can see a benefit in that especially if they're on the cheap