Search found 5 matches
- Tue Mar 30, 2021 10:33 am
- Forum: Training Discussion
- Topic: Manual Masking
- Replies: 1
- Views: 2359
Manual Masking
Hi, is there any efficient way to demask objects or arms and fingers in front of faces? Is it even worth the time to do it for all source frames? Also you need to do manual masking in the source frames. So you cant filter the images you use for training after you sorted some of them out right? Becau...
- Sun Mar 28, 2021 2:29 pm
- Forum: Training Discussion
- Topic: Changing output size mid-train
- Replies: 2
- Views: 2609
Re: Changing output size mid-train
Thanks for your answer and the quick sumup.
- Sun Mar 28, 2021 12:05 pm
- Forum: Training Discussion
- Topic: Changing output size mid-train
- Replies: 2
- Views: 2609
Changing output size mid-train
Hi, is it possible to change the output size for one trainer in mid-train process or will it mess up the model? For example change dfaker output from 128 to 256? Can i expect better results in the endphase of training or is it not really helpful? Im aware of that a bigger output size needs more VRam...
- Sun Mar 28, 2021 11:59 am
- Forum: Extract Discussion
- Topic: Merge Videos --> Extract
- Replies: 2
- Views: 8127
Re: Merge Videos --> Extract
Thanks for your answer. I guess i had a false understanding of how the alignments are generated. There isn't one alignment for the whole video, instead it's one alignment generated for every face, right? In the past all those alignments were merged into one bigger alignment file. That's at least how...
- Sat Mar 27, 2021 8:03 pm
- Forum: Extract Discussion
- Topic: Merge Videos --> Extract
- Replies: 2
- Views: 8127
Merge Videos --> Extract
Hey, is it safe to merge Videos (Sources) first and then extract the faces, after the alignment update? Or is it better to extract every video and put the generated images into one folder afterwards?
Thanks for your help.