[Outdated][Guide] Converting a DFL (DeepFaceLab) Face Set for Faceswap

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
deephomage
Posts: 33
Joined: Fri Jul 12, 2019 6:09 pm
Answers: 1
Has thanked: 2 times
Been thanked: 8 times

[Outdated][Guide] Converting a DFL (DeepFaceLab) Face Set for Faceswap

Post by deephomage »

 ! Message from: torzdf

This method is now deprecated. Unfortunately it is not possible to use DFL generated facesets in Faceswap.

NB: Whilst this process should work for full face DFL face sets, it may fail for larger extractions.

DeepFaceLab stores the image alignments in each image header. To use a face set created in DFL for training a model in faceswap, you need to convert the DFL stored image alignments to an alignments.json file.

Faceswap normally creates the alignments.json file during the extraction process. But you can create an alignments.json from an existing DFL face set by opening a Windows Command prompt, changing to your faceswap root directory (usually C:\Users\[Your User Name]\faceswap) and running the command:

Code: Select all

python tools.py alignments -a dfl -j reformat -fc "<insert the full path to the DFL face set here>"

You can also convert the DFL alignments using the GUI:

dfl_convert.png
dfl_convert.png (135.43 KiB) Viewed 15154 times
Locked