Search found 1490 matches

by torzdf
Wed Sep 15, 2021 9:01 am
Forum: Swap Showcase
Topic: First swap - Looking for advice (MKBHD x Michael Jordan)
Replies: 5
Views: 101

Re: First swap - Looking for advice (MKBHD x Michael Jordan)

Thanks for sharing your journey. I think/hope this will be helpful for other users. re: Phaze-A. Default settings should work on 4GB with Mixed Precision enabled, so I would guess that Windows is taking more VRAM than it should. If you can try on a Linux distro, then I would recommend this. A note a...
by torzdf
Thu Sep 09, 2021 12:05 am
Forum: Installation Support
Topic: [Guide] Using Faceswap on Nvidia RTX 30xx cards
Replies: 111
Views: 89595

Re: [Guide] Using Faceswap on Nvidia RTX 30xx cards

Thanks for the feedback, that is useful information :)

by torzdf
Tue Sep 07, 2021 9:41 pm
Forum: Extract Support
Topic: Manual Tool Can't Save - Index Out of Range
Replies: 1
Views: 1620

Re: Manual Tool Can't Save - Index Out of Range

I'd need recreatable examples to be able to look into this.

Either way, I'm not going to have time for a few weeks I'm afraid.

by torzdf
Tue Sep 07, 2021 9:35 pm
Forum: Training Support
Topic: Error on training
Replies: 1
Views: 1360

Re: Error on training

For the errors, you can safely ignore them. These have been introduced by Tensorflow 2.6 The custom_mask message is nothing to do with you and is most likely to do with a custom layer inside the faceswap application. It is a completely useless warning for the end user, but it has been decided by the...
by torzdf
Mon Sep 06, 2021 8:11 am
Forum: Installation Support
Topic: "WARNING Couldn't install nvidia-ml-py>=11.470.66 with pip. Please install this package manually"
Replies: 5
Views: 1168

Re: "WARNING Couldn't install nvidia-ml-py>=11.470.66 with pip. Please install this package manually"

Ok, I suspect what is happening is that pip cannot install nvidia-ml-py during an update, due the pre-existance of nvidia-ml-py3. In this case, you can ignore the warning. It is not important. I will look at steps to rectify in future. As long as there is a version of one of these packages on the sy...
by torzdf
Sun Sep 05, 2021 9:50 pm
Forum: Installation Support
Topic: "WARNING Couldn't install nvidia-ml-py>=11.470.66 with pip. Please install this package manually"
Replies: 5
Views: 1168

Re: "WARNING Couldn't install nvidia-ml-py>=11.470.66 with pip. Please install this package manually"

And are you installing or running an update?

It would make my life a lot easier if in future, you provided this kind of information rather than me having to probe for it. Also no need to @ me in every message.

Ultimately, that file exists, so...
https://pypi.org/project/nvidia-ml-py/

by torzdf
Sun Sep 05, 2021 4:59 pm
Forum: Installation Support
Topic: "WARNING Couldn't install nvidia-ml-py>=11.470.66 with pip. Please install this package manually"
Replies: 5
Views: 1168

Re: "WARNING Couldn't install nvidia-ml-py>=11.470.66 with pip. Please install this package manually"

Some context please. What's your OS, what's your CPU etc. etc. Just an output message like that helps me not at all.

by torzdf
Sun Sep 05, 2021 9:09 am
Forum: General Discussion
Topic: Warning on Ampere (30xx series) cards.
Replies: 4
Views: 86074

Re: Warning on Ampere (30xx series) cards.

Cuda 11.2/cuDNN 8.1 support has now been added to the default installer, so 30xx cards should just work.

See this thread for updating: viewtopic.php?f=4&t=1226

by torzdf
Sat Sep 04, 2021 9:08 am
Forum: Installation Support
Topic: [Tip] Installing Faceswap for AMD in a Docker File
Replies: 0
Views: 1376

[Tip] Installing Faceswap for AMD in a Docker File

Thanks to Discord user Cyborg, the following is a method that AMD users can use to install Faceswap with the amdgpu-pro driver inside a Docker Container. dockerfile: # sudo docker build -t faceswap/amdgpu . FROM ubuntu:bionic ENV DEBIAN_FRONTEND=noninteractive RUN \ amdgpu_version="amdgpu-pro-1...
by torzdf
Sat Sep 04, 2021 8:56 am
Forum: Installation Support
Topic: "Failed to extract font properties" when start Faceswap.
Replies: 4
Views: 3910

Re: "Failed to extract font properties" when start Faceswap.

Honestly, I don't know why this happens. However, it is just an info message generated by Matplotlib when scanning your font files, that it can't process a specific font. I don't believe it impacts the workings of Faceswap in any way. Whilst it is annoying to be nagged every time you launch faceswap...
by torzdf
Fri Sep 03, 2021 9:13 am
Forum: Training Support
Topic: Error During Training: CUDA_ERROR_ILLEGAL_ADDRESS: an illegal memory access was encountered
Replies: 3
Views: 1270

Re: Error During Training: CUDA_ERROR_ILLEGAL_ADDRESS: an illegal memory access was encountered

[/code] "OP, your problem sounds a lot like this guy's. Learn to use the search bar lol" Yes, nothing I've tried from there has helped and as far as I'm aware, this person never solved their problem either. The only thing I haven't been able to try yet is upgrading my PSU, which I plan to...
by torzdf
Wed Sep 01, 2021 10:07 am
Forum: Training Discussion
Topic: Question about Phaze-A model scaling
Replies: 2
Views: 1137

Re: Question about Phaze-A model scaling

Ha! Your link just showed me a bug in my -1 to 1 scaling (now fixed), so thanks for that. I just pulled this information from keras applications, for example this for efficientnet (https://www.tensorflow.org/api_docs/python/tf/keras/applications/efficientnet/EfficientNetB0): Note: each Keras Applica...
by torzdf
Mon Aug 30, 2021 9:47 am
Forum: Training Support
Topic: Question on training.
Replies: 1
Views: 1089

Re: Question on training.

You can't train Faceswap on a single image. The Neural Network needs lots of images at a high variety of data. Lots of identical images are not going to get it anywhere.

Related: viewtopic.php?f=27&t=831

by torzdf
Sun Aug 29, 2021 9:49 am
Forum: Training Support
Topic: "New" training method taking too much VRAM?
Replies: 7
Views: 3336

Re: "New" training method taking too much VRAM?

Ok, so I'm doing some tests. If I take the model that works with more iterations than the others and change just the name of the model dir, forcing it to generate a new one, the software runs into the same VRAM problem as the others. Are you using Conv Aware Init. This is known to take up more vram...
by torzdf
Sun Aug 29, 2021 9:46 am
Forum: Training Discussion
Topic: Strange Red Area/Artifact Creeping into Model
Replies: 26
Views: 8375

Re: Strange Red Area/Artifact Creeping into Model

Interested to know if the Adabelief optimizer can be used with this model (Phase-A StoJo 256). We've set EE to -16 but get complete model collapses after 10k iterations. We've also disabled mixed precision to see if we can fix the issue further into training as described above. Of course, we can't ...
by torzdf
Fri Aug 27, 2021 9:38 am
Forum: Training Discussion
Topic: Face Coverage
Replies: 2
Views: 1414

Re: Face Coverage

This is a 'bug'.... IE when I introduced face centering and set it as default, it looks like I did not up the default coverage. I will amend this at some point (hence me tagging this post as bug). 68.75 coverage is a hangover from legacy centering. In your particular example you're probably going to...
by torzdf
Fri Aug 27, 2021 9:34 am
Forum: Training Discussion
Topic: Please give me advise on my approach
Replies: 3
Views: 1758

Re: Please give me advise on my approach

I would add, in terms of workflow: Don't bother adding any maskers during extract time. It just slows down extraction. Do a "first pass" extraction, which is used purely to identify and extract faces from a video. Sort the faces (one source at a time) and clean up the face set. Clean the a...
by torzdf
Fri Aug 27, 2021 9:28 am
Forum: Extract Discussion
Topic: Questions on Multiple Sources
Replies: 1
Views: 2161

Re: Questions on Multiple Sources

Lets say I have 4 videos to extract FaceB from. So I extract each one, and end up with all of them in a single output directory, and I have 4 separate alignment files. So then I can run a Tools->Sort, and sort by faces, delete the faces I don't want, and then I can Remove Faces from each of the 4 a...