Search found 141 matches

by bryanlyon
Mon Jan 20, 2020 8:54 pm
Forum: Extract
Topic: Creating masks for heavily obstructed faces
Replies: 2
Views: 8

Re: Creating masks for heavily obstructed faces

Right now, if no masks work for you there isn't much you can do about it. Torzdf does have an updated manual tool in the works that will allow you to manually mask the face, but right now it is not yet available.
by bryanlyon
Mon Jan 20, 2020 4:28 am
Forum: Extract
Topic: Error when masking with Unet post extraction
Replies: 3
Views: 10

Re: Error when masking with Unet post extraction

This is often masking a memory error. If you've already turned on Allow Growth in the Extract plugins, you may try turning down the BatchSize in the Settings/Extract Plugins/Mask/Unet DFL.
by bryanlyon
Mon Jan 20, 2020 4:23 am
Forum: Training
Topic: Speed/Quality and Batch size
Replies: 3
Views: 62

Re: Speed/Quality and Batch size

viewtopic.php?f=6&t=146
This is the best source of information on this. Honestly, much more than what this guide shows is overthinking things. Just go for the largest Batch Size you can without turning on any memory hacks (as those slow down training for Vram).
by bryanlyon
Sat Jan 18, 2020 7:08 pm
Forum: Training
Topic: Speed/Quality and Batch size
Replies: 3
Views: 62

Re: Speed/Quality and Batch size

Training speed is not just iterations. It's actually best to use the number of images in a time period. So Batch Size * Iterations is a better calculation So 50 * 240 = 12,000 images/hour 30 * 500 = 15,000 images/hour 6 * 900 = 5,400 images/hour (I however, believe you've likely mistaken the speeds ...
by bryanlyon
Sat Jan 18, 2020 5:27 pm
Forum: General Chat
Topic: Nvidia recently posted information saying that NVME could be used as
Replies: 1
Views: 13

Re: Nvidia recently posted information saying that NVME could be used as

Not really, no. Storage is much slower than the vtam on board the gpu and relying on it as vram replacement is not viable. Also, the images must be augmented so even streaming the raw images isn't viable. Further, consumer pcs don't have separate pci-e switches that allow the gpu to directly read fr...
by bryanlyon
Sat Jan 18, 2020 6:10 am
Forum: Training
Topic: Failed and return code: 3221225477
Replies: 3
Views: 25

Re: Failed and return code: 3221225477

This is a weird error that we've seen once before. See https://github.com/deepfakes/faceswap/issues/954 Unfortunately, none of the developers have been able to reproduce it. We are trying to look into it, but it may simply be a corrupted install. That error code is not coming from our application, b...
by bryanlyon
Fri Jan 17, 2020 5:44 pm
Forum: Training
Topic: Failed and return code: 3221225477
Replies: 3
Views: 25

Re: Failed and return code: 3221225477

Please post your crash log (if present). It'll be in the faceswap folder named "crash_report[date].[longnumber].log . If not, post your system info: GUI Users: Go to Help -> Output system information CLI Users: From inside your virtual environment, inside your faceswap folder, run: python -c "from l...
by bryanlyon
Fri Jan 17, 2020 5:52 am
Forum: Training
Topic: No flip for specific faces
Replies: 1
Views: 21

Re: No flip for specific faces

Any time there is a mark you want to preserve on one side of the face you should disable the flip option throughout the entire training process. If you do this please be aware that you'll want to include a significantly larger dataset to ensure the quality of the results.
by bryanlyon
Thu Jan 16, 2020 4:00 pm
Forum: Hardware
Topic: Hardware best practices
Replies: 17
Views: 2366

Re: Hardware best practices

It's not critical. You could certainly get away without upgrading the CPU, but you *WILL* see better speeds if you upgrade it later. Might be worth scheduling it sometime down the line.
by bryanlyon
Thu Jan 16, 2020 3:52 pm
Forum: Hardware
Topic: Hardware best practices
Replies: 17
Views: 2366

Re: Hardware best practices

Yes. a 2060 super would let you turn off a lot of the memory preservation you're currently using which is honestly the thing most slowing you down with your 960. I'm not going to promise double the speed, but you'll see some strong speed improvements. The only problem is that with a 2060 super, your...
by bryanlyon
Wed Jan 15, 2020 1:33 am
Forum: Hardware
Topic: free genesiscloud gpus any good?
Replies: 2
Views: 43

Re: free genesiscloud gpus any good?

I've never used it. They *look* fairly legit but you can never be sure. Maybe you'd rather try the free tier of Paperspace https://docs.paperspace.com/gradient/in ... -instances ?
by bryanlyon
Tue Jan 14, 2020 3:54 pm
Forum: Installation
Topic: vggface2_resnet50_v2.h5 install
Replies: 1
Views: 28

Re: vggface2_resnet50_v2.h5 install

From the faceswap folder, plugins/extract/recognition/.cache/vggface2_resnet50_v2.h5
by bryanlyon
Tue Jan 14, 2020 3:52 pm
Forum: Training
Topic: Improving quality
Replies: 3
Views: 128

Re: Improving quality

Villain is barely able to run on 8gb of vram. Try a BS of 1. If that doesn't work, try closing extra programs (Like browsers). Unfortunately MSG and other memory saving features come at a cost of time, that's just the tradeoff that you have to make to run larger models with limited memory.
by bryanlyon
Mon Jan 13, 2020 6:47 pm
Forum: Extract
Topic: What size should I extract faces at?
Replies: 2
Views: 59

Re: Image size

No, leave it at 256.
by bryanlyon
Sun Jan 12, 2020 5:27 am
Forum: Training
Topic: Crash When Resuming Training
Replies: 1
Views: 32

Re: Crash When Resuming Training

It sounds like you were marginally capable at that BS. Sometimes just by luck you run at a batch size once that wont work again later. Try a lower BS or close some applications that may be using vram (Browsers and Video Players are good examples).
by bryanlyon
Thu Jan 09, 2020 6:27 pm
Forum: Training
Topic: How to record a video for the sole purpose of training/face swapping?
Replies: 6
Views: 109

Re: How to record a video for the sole purpose of training/face swapping?

Flash, probably, but you should definitely use set lights. Lighting is probably the most important thing that people miss out on.
by bryanlyon
Thu Jan 09, 2020 6:23 pm
Forum: Training
Topic: How to record a video for the sole purpose of training/face swapping?
Replies: 6
Views: 109

Re: How to record a video for the sole purpose of training/face swapping?

It's important to remember that there is a difference between Training data videos and the swap video. Making a quality training set needs a huge variety, but the swap video you want to minimize complexity in any way you can. For training data, just run the camera, catch video when not actually film...
by bryanlyon
Thu Jan 09, 2020 3:51 pm
Forum: Training
Topic: Swapping Jeff Bezos and Elon Musk: blurry result
Replies: 3
Views: 137

Re: Swapping Jeff Bezos and Elon Musk: blurry result

You want the training faces to be at least that 128x or you're training to lower quality faces and you wont get full quality for the swap.

Villain is already a very heavy model. Without some exotic hardware with a lot of vram, it's pretty much as big as it can get.
by bryanlyon
Wed Jan 08, 2020 7:08 pm
Forum: Swap Showcase
Topic: Corridor Digital's Faceswap videos.
Replies: 8
Views: 2875

Re: Corridor Digital's Faceswap videos.

1. "with 100% coverage and extended mask" See https://forum.faceswap.dev/viewtopic.php?f=6&t=146#config 2. "pretrained the villain model for 50k iterations on random faces" why would he do this...or did he mean random faces of Keaunu ? There is some debate as to whether pretrain is useful or not. P...