Faceswap Forum

It might make more sense to think of it as a "forest for the trees" sort of thing. Low batch sizes will get a single tree looking better, while high batch sizes will get the whole forest looking better. Larger batch sizes do train faster, but they can't necessarily train better. I did a te...

A basic rule of thumb is that doubling the resolution multiplies by 4 the amount of memory required. if 256x256 is taking 20gb, then 512 would take 80gb and 1024 would take 320gb. So, no, 256gb of ram wouldn't be able to handle 1024x1024 training out of the box. However, another key point is that re...

Your model has probably corrupted. Probably due to running out of storage or something similar. Try using one of the backups.

Intel macs are very different from the Apple silicon and aren't supported the same way. Since you have an AMD card you should try the amd install. But you could also try using ROCM.

That could be a perfectly normal speed. It depends on which model you're using, larger models run slower and will take longer to process.

You need to clean your alignments file. See viewtopic.php?t=27#clean-the-alignments-file

Unfortunately I can't be sure what is happening here with this report. You may need to enable "trace" logging and send the log over to see what is actually causing the freeze.

NPUs can't generally be used to train. The Intel one included with that CPU is the same issue. Basically, Faceswap itself needs to train a model and while that NPU could help with tasks like extracting, those are a tiny part of what FaceSwap does, so we wont be supporting NPUs as the gains would be ...

So, it seems you've gotten going. A few pieces of advice: It looks like you'll need to clean up your data more. Side A is very lacking in variety, while Side B has a number of non-face objects you'll probably want to clean.

Welcome to the group and feel free to ask any questions.

This really isn't enough information to know what's going on. Please tell us what problems you're seeing and include any crash logs if present so we can troubleshoot.

No, we extract for use in the AI which is at a far lower resolution in almost all cases. If you're trying to extract for some other usage it'll be best to find a different face extractor (or write your own).

The only ways to know are to try it or find someone with experience to examine the dataset to give advice. I suggest trying it. Start with a small model until you can learn the processes.

If you installed Faceswap properly (using the installer or proper guide) then installing other Python instances shouldn't do anything to FaceSwap. However, since you've already got Conda installed with Faceswap I suggest that you install that other application in a proper conda environment. See the ...

Hello and Welcome. * if one has a video of oneself and wants to put their face onto an "original" where the person is wearing glasses, does the software recognize it? Or does it pixelate the object (the glasses), or even cause image distortions? As long as a decent chunk of the training da...

This actually is not a problem with FaceSwap but instead originates on your system. I'm assuming it's something due to drivers or the specific setup of Kali, WSL, Windows and your local system. Since there are so many variables there I wouldn't even know where to begin to find the cause. This is why...

"Violation of the segment ('core' generated)" This implies that something is failing to get a memory allocation it's seeking. It's also clearly outside of python and in the realm of compiled code. Probably high up in a library. Unfortunately without more information the best we can do is g...

The only 30xx card with the nvlink fingers is the 3090. All others cannot do any sort of "linking". That said, you can still use multiple GPUs in Faceswap. It is diminishing returns where each additional GPU slows down the collective more, and like torzdf says, only the batch size can be i...

The trouble with hardware recommendations since the 20xx series has been that you always have to add "if you can get it for a good price" to the end of ANY recommendation. Will a 4060 TI with 16gb of vram work? Definitely. Will it beat a $100 1080? Yes. Is it worthwhile to buy it for FaceS...

You don't have to hook into the passes at all to add your own architecture. You can add new models by following the plugin architecture here: https://github.com/deepfakes/faceswap/tree/master/plugins/train/model or add a new training paradigm following the plugins here: https://github.com/deepfakes/...

As the previous message said, this is a driver issue. Nothing to do with Faceswap and nothing we can do to prevent or work around it. As far as I've been able to find on it though, it's almost 100% an overclocking issue. Chances are the card you have is factory overclocked and it's hitting instabili...

Faceswap Forum

Search found 802 matches

Re: question about batch size

Re: Scaled Memory - Apple Silicone vs Nvidia

Re: Error: 0 state model found

Re: Intel Mac with AMD GPU running in CPU mode

Re: Low EG/s even after fresh install.

Re: How To Ensure The Interface Considers The Right Face To Swap In A Video With Two Faces?

Re: conversion problem

Re: Intel NPU / OpenVINO support

Re: Am I using it properly?

Re: HELP !!! After reinstall the OS and Faceswap, relocate all folder files, it does not work as it used to!

Re: Faceswap face extraction

Re: Question about convergence

Re: Will installing new Python version mess with app?

Re: A Short Hello with some questions

Re: GUI Glitching

Re: FaceSwap Linux running, but suddenly stops "Proces Exited" with no Errors

Re: Hardware best practices

Re: Hardware best practices

Re: Location of “forward pass” and/or Where to implement a Custom Training Loop

Re: Tensorflow error happens randomly while training