Training in Faceswap 2.0

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
thebrave246
Posts: 1
Joined: Wed Oct 28, 2020 1:43 am

Training in Faceswap 2.0

Post by thebrave246 »

Hey everyone I'm pretty new to Deepfaking and I've been training now over 200K iterations using DFL SAE at 256 input res pixels. The results aren't bad so far but I want to get the best results I can get, my question at the moment is can I switch model trainers during the process and see if it helps or would that reset everything? Also I am only using a batch size of 2 right now because the 256 input res is demanding of my gpu and I am using an RTX 2080 Super 8GB. Any help is appreciated, and here is some of my results at the moment.

Image
Image

Top
User avatar
torzdf
Posts: 2684
Joined: Fri Jul 12, 2019 12:53 am
Answers: 159
Has thanked: 134 times
Been thanked: 626 times

Re: Training in Faceswap 2.0

Post by torzdf »

You can't change model type. Each model is unique and independent from each other.

You may be able to run bigger batch sizes by using "Mixed Precision" (in the training settings menu), but you must start a new model to use this.

My word is final

Top
Locked

Return to “Training Discussion”