FUNIT face swap instead of pet swap

https://github.com/shaoanlu/fewshot-face-translation-GAN uses modules from FUNIT and SPADE for face-swapping but the quality of swaps isnt as good as FUNIT, are there particular limits like expressions and rotations?

Sep 03 '19 16:09 ak9250

I just ported funit to pure keras and started training with 2 classes for face swap

Sep 14 '19 06:09 iperov

python_2019-09-14_13-08-11

Sep 14 '19 09:09 iperov

@iperov cool, is this port public or will it be part of deepfacelab?

Sep 14 '19 13:09 ak9250

part of dfl

Sep 14 '19 14:09 iperov

when approx it will appear in your dfl?

Sep 16 '19 19:09 eps696

@eps696 I dont know , currently testing... Do you want source code for keras ?

Sep 16 '19 19:09 iperov

would love to try it [dfl looks a bit overloaded by end-user convenience features, cleaner task-oriented code would b perfect]

Sep 16 '19 20:09 eps696

In the last page of the FUNIT arxiv paper (which will be presented in ICCV 2019), we do have the few-shot face translation experiments. We simply use CelebA for the experiments. I expect combining with SPADE and utilizing landmarks should lead to better performance thought.

Sep 16 '19 21:09 mingyuliutw

@mingyuliutw I am trying to train FUNIT on celebA but I am confused. What are the classes? If I take every person as a different class then there would be too much classes. Or can I take real as 1 and fake as 0, nothing more? Please Guide me.

Oct 20 '19 11:10 niley1nov

Each human identity is a class. CelebA has the person name for each photo. You can divide the training set into different classes using the name.

Oct 20 '19 11:10 mingyuliutw

I trained FUNIT on 256 VGGFace dataset clases for a week.

Currently, it seems that it cannot correctly convert unknown persons.

FUNIT_preview_TEST

Oct 20 '19 13:10 iperov

@mingyuliutw, are 256 persons with 69k total photos enough for the model?

should I expand to 1024, or 2048 persons ?

Oct 20 '19 14:10 iperov

@mingyuliutw can you give me advice for network config for 1024 person classes and 200k photos?

Oct 21 '19 07:10 iperov

@mingyuliutw is it good that one class has 60 photos and other class has 500 photos?

Oct 21 '19 07:10 iperov

@iperov would you like to share me your keras code? thank you.

Oct 26 '19 09:10 miaoYuanyuan

https://github.com/iperov/DeepFaceLab/blob/master/nnlib/FUNIT.py

Oct 26 '19 10:10 iperov

thank you ~~

Oct 26 '19 10:10 miaoYuanyuan

@iperov Hi I'm quite interested how you trained fuint for face swap Note N faces in the trainig data set, you just followed the origin fuint traning set and treat N ids as N classes right?