Page 1 of 1

Same Models, pretrained model files

Posted: Sun Dec 01, 2019 12:52 am
by kaemen

I have a short maybe 10 secs video source, which I have succesfully swaped faces going thru the whole process and results are awesome..

Ive played around a little and found out something.

Another video of almost same source... (same person, equal face position, etc).

Like I said I played around, and set the former trained model folder and the previews fit perfectly.

But the alignment isnt correct. I wonder, if the preview is fitting... how can I adjust the alignement before convert? Why is the preview aligned but when I try to convert I get errors?

I found out that the training data contains the path to the original video file. So I copied the new video to the same directory and changed the name.

Process would continue then a little further but still end in another error.

Explain please


Re: Same Models, pretrained model files

Posted: Sun Dec 01, 2019 7:54 pm
by kaemen

No answer?

Let me more precise.

I am using the same Models as before.

The Video clip I am extracting Model A Faces is in the almost exact position. So I can use Model B Faces from before without problem.

New training process however takes equally long as before.

So I set the old trained folder with 40000 iterations. And the previews were perfect right at start, so I let it run and iterations.

But when trying convert this newly extracted faces. The Model B Face was discplaced and all blurry.

If the damn preview was correct how can I fix this with some sort command option?

Ive not used masks. Must be a solution ?


Re: Same Models, pretrained model files

Posted: Sun Dec 01, 2019 10:53 pm
by torzdf

Run extraction on your new video to generate an alignments file.


Re: Same Models, pretrained model files

Posted: Tue Dec 03, 2019 9:38 pm
by kaemen

Thats it?

Just recreate alignment for new video clip?`

Thanks I will try and feedback!


Re: Same Models, pretrained model files

Posted: Wed Dec 04, 2019 2:02 am
by kaemen

Your tip and new latest version works wonders!!

Thanks alot and big thank you for everyone putting time effort into the development of this amazing tool!

Thank you!