Hi, I am new to this whole deepfake stuff that you guys have here, and I need to get some vital information out quickly. You see, I am working on a summer internship and we were assigned the topic of deepfakes and how it could affect society and everything. I kind of got a jist of it, however, I want to know more about the programing behind it all. In other words, how exactly do deepfakes work?
Anyone able to answer this? That would be great!
You might want to check out our talk
Beyond that, we're happy to answer any specific questions you ask.
Ah, thanks, this video could be vital for my research. But in that case, a specific question I have is for training part. I have finished extracting my face and another face and I came out with two totals. My face having about 15000 images, and the other face, 3000 images. Does the number of faces or images found while extracting affect the training? Also, I am having trouble locating my alignment files, they don't appear in the folder I collected the images from.
Thanks in advance.
viewtopic.php?f=5&t=27#extract
In particular
Alignments: This option is in case you want to specify a different name/location for the alignments file. It's a lot easier to leave this blank, and let the process save it out to the default location (which will be in the same location as the source frames/video).
Alright, I have moved on to the merging step, however, I am a bit lost. It is asking for multiple alignment files to start the merging process... is it asking for the alignment files of two separate sets of faces or are there more alignment files I am not aware of?
Edit: hold on, I am going through the steps, and I noticed the key word "facesets" so I guess it is more than one sets of faces...
I have run into a problem and I do not have enough VRAM. It tells me what I can do, however, how much VRAM is even needed?
We would probably be more willing to help if it looked like you had done at least some basic research.
This is covered in the faqs, which is effectively the first thing you should have read.
My word is final
Understandable, I did jump onto this program without thinking twice, and did not check out any other information or requirements. However, I wouldn't say this was a waste of time for me, this experience so far would be good to speak on the accessibility of deepfake programs, and difficulty of creating one if one is not properly prepared or informed. With that being said, thank you for your time and information you were able to provide, and pardon me for you with questions I could have found the answers to myself.
No, it's fine. Enquiring minds are good. Unfortunately, time is limited, so I prefer to try to avoid retreading the same ground again and again.
We are of course happy to help with clarifying issues, but we have tried to put together a wide array of documentation to help ease the entry into what is a complicated process.
Our Discord server is also a good place for discussion with other users.
My word is final
