Obstructions- howto?

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
dafassi
Posts: 3
Joined: Wed Feb 15, 2023 11:24 am

Obstructions- howto?

Post by dafassi »

Hello there!
I am new to FaceSwap and after a lot of learning and trying I finally can get great results in my project.

Thank you to all developers!

However - one thing I cannot get working correctly is obstruction’s like microphones, hands etc.
This is always blurred on my results and the swap face laying over the obstruction.

I am using the latest FaceSwap Version running on a RTX3070ti.

Can some one explain me what I need to do, to get this working correctly?

  1. what extraction method should I choose?
  2. I am training with h128 - Is there any setting I am missing ? Or should I use an other trainer?
  3. is there a preferred conversion mode?
  4. do i Need to do anything else?

Thanks to all!

Top
User avatar
torzdf
Posts: 2672
Joined: Fri Jul 12, 2019 12:53 am
Answers: 159
Has thanked: 131 times
Been thanked: 625 times

Re: Obstructions- howto?

Post by torzdf »

Use the BiSeNet-FP mask (Faceswap weights) for both training and convert.

More info:
https://github.com/deepfakes/faceswap/pull/1210

My word is final

Top
Locked

Return to “Training Discussion”