Any Success with Dlight 384 on Local Hardware

Want to understand the training process better? Got tips for which model to use and when? This is the place for you


Forum rules

Read the FAQs and search the forum before posting a new topic.

This forum is for discussing tips and understanding the process involved with Training a Faceswap model.

If you have found a bug are having issues with the Training process not working, then you should post in the Training Support forum.

Please mark any answers that fixed your problems so others can find the solutions.

Locked
User avatar
r4v1np1nk
Posts: 1
Joined: Sun Mar 15, 2020 7:12 pm

Any Success with Dlight 384 on Local Hardware

Post by r4v1np1nk »

I'm wondering if anyone has had success training Dlight with a 384px output on local (i.e. not cloud procured) hardware. If so, I'm wondering what your config was, specifically in terms of VRAM. I've got a GeForce card with 6GB right now and that's only enough to get up to 128px. Looking at a separate GPU card for this purpose and wondering if 12GB is going to be enough or if I should be looking at 24GB cards.

(I mean if money were no object, I'd buy a V100 and be done with this conversation but that's not happening for obvious reasons).

Top
User avatar
torzdf
Posts: 2682
Joined: Fri Jul 12, 2019 12:53 am
Answers: 159
Has thanked: 133 times
Been thanked: 626 times

Re: Any Success with Dlight 384 on Local Hardware

Post by torzdf »

12GB should be fine. More VRAM is always going to mean the ability to run larger models though, so really it's up to your budget/priorities.

My word is final

Top
Locked

Return to “Training Discussion”