Faceswap Forum

You can do this with the effmpeg tool and the "mux-audio" option.

input - Your converted video
reference video - The original video (which has the audio you want to put on to your convert)

Thanks for this informative post. There is no question that Tensorflow compiled with ROCm support will perform far greater than plaidML with it's OpenCL support, but we didn't go that way for a couple of reasons.... As you alluded to, ROCm is only available on Linux, so for us and our desire to supp...

No, This is all handled for you.

In fact you SHOULDN'T install Cuda as it can lead to issues with the local install.

djandg wrote: ↑

Tue Apr 07, 2020 10:37 am

Takes a long time !

I'll second that. Took me over 10 days to train Villain to convergence

I have trained Villain at a very low batchsize (4 iirc) on a GTX 1080.

This was on Linux, so will be tougher with Windows, which holds back more of your GPU RAM.

You should be able to get it going with some of the VRAM saving options though.

To answer your questions in reverse order.....

The interface is updated at the same time.

I will review feature adds when I have a bit more time.

Try enabling the "Allow Growth" option. If that doesn't work, put your log in Pastebin and link here

This is a local lib issue.

https://www.google.com/search?client=fi ... +directory

This looks like a memory error.

Please go Help > Output System Information and paste here

A 2GB GPU will run the Lightweight model, and *maybe* the Original model with optimizer settings enabled.

If you have a GPU you want to use GPU mode.

Help > Output System Information. Copy + Paste here

If you go to your model folder, and copy the <model_name>_state.json file somewhere, then open the copy, it will tell you how many iterations it has done, so you should be able to tell if it is still incrementing...

Make sure you take a copy rather than open the original (whilst training)

The alignments tool is currently being re-written, so I won't be looking to bugfix.

However, the tool has no impact on output pngs (they aren't used for the tool) so I'm not sure how it would corrupt pngs?

plaidml is AMD only.


Check this answer:
app.php/faqpage#f0r3

Your frames location is wrong.

It should be "G:/faceswap test/POLARRANA/videos/low to up.mp4" (not with the word "done" in it.... assuming you haven't renamed your alignments file)

You should check out the guides, they cover a lot of this stuff:

app.php/tag/Guide

Please press the "Generate" button on your command and post the output