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.
rolnor
Posts: 78 Joined: Tue Feb 23, 2021 10:09 am
Has thanked: 2 times
Been thanked: 4 times
Post
by rolnor » Fri Jun 04, 2021 5:03 pm
I have the newest version of faceswap and there is no place to type in the alignements when running training as far as I see. This is a change from the previous version. Where does the software get the alignements from? Is there a way to add alignements manually??
rolnor
Posts: 78 Joined: Tue Feb 23, 2021 10:09 am
Has thanked: 2 times
Been thanked: 4 times
Post
by rolnor » Sun Jun 06, 2021 9:52 am
Thanks.
So I dont need alignements anymore when doing training?
torzdf
Posts: 2681 Joined: Fri Jul 12, 2019 12:53 am
Answers: 159
Has thanked: 133 times
Been thanked: 625 times
Post
by torzdf » Tue Jun 08, 2021 10:11 am
You don't need an alignments file. You do still need alignments, but they are stored within the face's PNG metadata when you extract faces.
rolnor
Posts: 78 Joined: Tue Feb 23, 2021 10:09 am
Has thanked: 2 times
Been thanked: 4 times
Post
by rolnor » Wed Jun 09, 2021 3:52 pm
OK, great! That simplifies.
rolnor
Posts: 78 Joined: Tue Feb 23, 2021 10:09 am
Has thanked: 2 times
Been thanked: 4 times
Post
by rolnor » Thu Jun 10, 2021 10:59 am
I am using two folders with pictures extracted in the earlier version of fs and running training in the new version and the result is partally good, partially bad. Is this very wrong? Does extracted fases from the earlier version also containing alignement info or not?
rolnor
Posts: 78 Joined: Tue Feb 23, 2021 10:09 am
Has thanked: 2 times
Been thanked: 4 times
Post
by rolnor » Thu Jun 10, 2021 1:55 pm
What I can see when I convert is that the faces have very different size so I will extract in the new version of fs and redoo the training.