Hi! Super super new at this (downloaded yesterday and have been poking around a bit here and there). I have a few things I want to perform swaps on, primarily smaller clips with a variety of faces. All of them I intend to swap with the same B face. I'm still trying to understand how the model works (I took a course in neural nets like ... a decade ago, I'm kinda rusty by now) but I'm wondering if I can use the same trained model for the clips in question. Is this a horrible idea, an okay idea, or am I completely 100% misunderstanding how this works from the ground up?
If it helps for context, I'm using torzdf's tutorials to start with, those are quite helpful, but much of the other stuff on the forums is going over my head so far because I'm new.