Search found 2638 matches

by torzdf
Wed Mar 31, 2021 10:03 am
Forum: General Discussion
Topic: Auto alignment to "fill in the gaps" based on manually cleaned frames?
Replies: 2
Views: 10985

Re: Auto alignment to "fill in the gaps" based on manually cleaned frames?

If you are needing to manually align 1000s of frames for a single video, then something doesn't sound right. Do you have image examples you can share?

by torzdf
Wed Mar 31, 2021 10:02 am
Forum: Artificial Intelligence
Topic: Voice swap ?
Replies: 10
Views: 30263

Re: Voice swap ?

Voice swapping is a difficult problem to solve, as evidenced by the lack of decent solutions. Ultimately there is a lot more nuance in voice than there are in images.

by torzdf
Tue Mar 30, 2021 11:12 am
Forum: Training Discussion
Topic: Manual Masking
Replies: 1
Views: 2339

Re: Manual Masking

Unfortunately, at this moment in time, no there is no quick way to mask obstructions. It is a laborious face-by-face task....

It is most important for convert, although can help training too.

It is something on my road map to address.

by torzdf
Mon Mar 29, 2021 11:48 am
Forum: Installation Support
Topic: Setting environment variables for Tensorflow (on Linux)
Replies: 6
Views: 12875

Re: Setting environment variables for Tensorflow (on Linux)

Honestly? I'd need to investigate why that hard coded value was put in in the first place. My preference would just be to remove it altogether, as I am not particularly a fan of setting environment variables inside applications.

by torzdf
Mon Mar 29, 2021 11:41 am
Forum: Extract Support
Topic: Fan Aligner Crashing - CRITICAL An unexpected crash has occurred.
Replies: 1
Views: 10363

Re: Fan Aligner Crashing - CRITICAL An unexpected crash has occurred.

The FAN model file has corrupted. Most likely during download....
Delete the file:

faceswap/plugins/extract/align/.cache/face-alignment-network_2d4_keras_v2.h5

and try again

by torzdf
Sun Mar 28, 2021 11:50 am
Forum: Installation Support
Topic: Setting environment variables for Tensorflow (on Linux)
Replies: 6
Views: 12875

Re: Setting environment variables for Tensorflow (on Linux)

I would say that editing the code directly is rarely the best idea, as it can prevent Faceswap from updating.

by torzdf
Sun Mar 28, 2021 11:49 am
Forum: Installation Support
Topic: [Guide] Using Faceswap on Nvidia RTX 30xx cards
Replies: 124
Views: 201298

Re: Troubleshoot slow training speeds

Hi Everyone, I am trying to fake my friends face as a birthday present but am a total noob so any help would be really appreciated. I have got to the training stage and am getting extremely slow speeds. GPU: 3080 CPU ryzen 5 3600 batch size: 16 Villain: 1.3 EG/s people with comparable hardware seem...
by torzdf
Sun Mar 28, 2021 11:48 am
Forum: Convert Support
Topic: Crash Report
Replies: 3
Views: 11018

Re: Crash Report

Go into Training Settings and try reducing the Convert Batchsize

by torzdf
Sun Mar 28, 2021 11:46 am
Forum: Extract Discussion
Topic: Merge Videos --> Extract
Replies: 2
Views: 8023

Re: Merge Videos --> Extract

Honestly? It's entirely up to you for whichever works best for you workflow wise.

Personally I tend to stick to separate sources, but there is no reason not to compile all your sources together.

by torzdf
Sun Mar 28, 2021 11:43 am
Forum: General Discussion
Topic: Can faceswap be used to swap specific features such as ONLY eyes/nose/mouth?
Replies: 2
Views: 13824

Re: Can faceswap be used to swap specific features such as ONLY eyes/nose/mouth?

Sure it can, but it's going to require work on your part, and is not a feature we are likely to ever build in.

Ultimately, it is a masking problem. You have access to the 68 point face landmarks, so you can create "sub-masks" for any area of the face from that, and use those for conversion.

by torzdf
Fri Mar 26, 2021 9:35 am
Forum: Training Support
Topic: Session Stats no longer appearing after a few hours of training
Replies: 15
Views: 11576

Re: program crashes with this error message in the console

Yeah. It's a GUI bug. Doesn't impact training, just stops the graph updating for that session.

I'll fix it one day.

by torzdf
Fri Mar 26, 2021 9:34 am
Forum: Extract Discussion
Topic: Missing Alignment Merge Job
Replies: 20
Views: 28373

Re: Missing Alignment Merge Job

That has nothing to do with merging alignments.

Merging alignments file was an old training process. You're trying to convert.

by torzdf
Fri Mar 26, 2021 9:33 am
Forum: Installation Support
Topic: Setting environment variables for Tensorflow (on Linux)
Replies: 6
Views: 12875

Re: Setting environment variables for Tensorflow (on Linux)

In the first instance, run Faceswap in VERBOSE mode and you should get a message from Tensorflow telling you why the GPU is not being used.

You can set an env variable by doing:

VARIABLE=variable ; python faceswap.py ...

by torzdf
Thu Mar 25, 2021 1:41 pm
Forum: Extract Discussion
Topic: Missing Alignment Merge Job
Replies: 20
Views: 28373

Re: was the merge option removed?

14XXX88 wrote: Thu Mar 25, 2021 12:47 pm

I don't see it under tools/alignment anymore

Answered in this thread

by torzdf
Tue Mar 23, 2021 7:23 pm
Forum: Convert Support
Topic: Looks good during training but not after converting.
Replies: 10
Views: 16561

Re: Looks good during training but not after converting.

The training previews are generated in exactly the same way as the converted image.

by torzdf
Tue Mar 23, 2021 7:17 pm
Forum: General Discussion
Topic: Advice to maximize results for low VRAM cards
Replies: 4
Views: 12452

Re: Advice to maximize results for low VRAM cards

Do the fancier models provide better results with less-good data (obstructed, lower quality images, etc.), or is it more like, you still need as good source data as you'd need with lightweight, but you get a better/less blurry rendering in fewer iterations, because of the wider firehose? The latter...
by torzdf
Tue Mar 23, 2021 11:50 am
Forum: Convert Support
Topic: Looks good during training but not after converting.
Replies: 10
Views: 16561

Re: Looks good during training but not after converting.

Preview image is tiny, convert can be huge.

Save a copy of the preview image during training and zoom into it. It should give you a better idea of why you are seeing the differences you are.

by torzdf
Tue Mar 23, 2021 11:48 am
Forum: Training Discussion
Topic: training with profile faces
Replies: 5
Views: 5028

Re: training with profile faces

You can't delete the "right eye". All 68 Landmarks need to be present. Therefore for profile shots the hidden landmarks will tend to "stack" onto the side of the face.

As long as the face image looks aligned, this is fine.

by torzdf
Tue Mar 23, 2021 11:45 am
Forum: General Discussion
Topic: Advice to maximize results for low VRAM cards
Replies: 4
Views: 12452

Re: Advice to maximize results for low VRAM cards

Honestly, dude. I'm amazed that you can run any of the models on a 2GB card, so kudos to you :) What OS are you running? Lightweight and LowMem (original) are fairly similar. Lightweight should be the lightest model we have, although it is balanced slightly different to the LowMem original model (li...
by torzdf
Mon Mar 22, 2021 12:12 pm
Forum: Convert Discussion
Topic: Face Not Converting most of the frames
Replies: 2
Views: 11519

Re: Face Not Converting most of the frames

Most likely you are using an alignments file that had an "Extract Every N" number set to higher than 1.

Re-check the convert guide for how to do it right.