General Question

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
groverjones
Posts: 1
Joined: Mon Mar 07, 2022 9:38 pm

General Question

Post by groverjones »

I can not get good results with anything. I have followed many tutorials. ran millions of iterations for weeks. thousand of HQ pictures. 10 second source, I keep getting blurry exports at the end. I have given up at this point. However, I was curious about something...
How is it so bad with so much time and data but there are apps I can get that I can upload one photo of a face and have better results in about 5 seconds on my phone? How is that even possible and why isn't it that easy on PC?

User avatar
torzdf
Posts: 2651
Joined: Fri Jul 12, 2019 12:53 am
Answers: 159
Has thanked: 129 times
Been thanked: 622 times

Re: General Question

Post by torzdf »

That's next to an impossible question to answer due to the huge number of variables.

Generally 'single photo apps' use morphing models. They are quicker, but generally less robust and of lower quality.

A well trained 'deepfake' model will always be better. The reason for potential blurriness are numerous and covered extensively elsewhere in these forums. Nearly always it's data.

My word is final

Locked