Furkan Gözükara patreon

VisoMaster Automatic Installer - The Most Advanced 0-Shot Face Swap / Deep Fake APP - State of the Art - Windows and Massed Compute

🕑 Added 2025-09-23 22:37:00 +0000 UTC

Comments

Furkan Gözükara

i had no idea of this thank you so much. i will update installer to select whichever branch you want. which one is best and should be default?

Lou

Please check out the Visomaster Discord. There are 3 forks of Visomaster, each one implements the latest AI tech and all 3 developers sync their versions so support for each is available. The Visomaster original program is no where near as good as the 3 new versions, 2 of which offer portable versions. Simple install, no cloning repository, just one bat file and everything Visomaster is installed. It's so much better now with developer mods included. You can even edit and combine embeddings within the program to help clean up all saved json files. Very efficient, I rarely hit 12gig ram using the latest tech. I dont personally have anything to do with the projects, other than writing a few simple guides for better results, but still think Visomaster is the best local face swap utility available. SeCourses is still the best place for learning AI programs but I know he is very busy and can't follow up on every post.

Furkan Gözükara

TensorRT works but it takes like 10-15 minutes to compile everytime. i recommend cuda

Robbie Westerman

What isnt TENSOR Working?

Furkan Gözükara

sadly i dont know . this project abondoned also. but it works for me . did you select output folder?

cool1

When I try running VisoMaster, and I press the "start recoding" option, it processes all frames (it currently doesn't seem to replace any faces. Though sometimes it does detect a face). Then at the end (after processing all frames) it says: "Option to (record or transcode stop time) cannot be applied to input file D:/the folder and filename -- you are trying to apply an input option to an output file or vice versa. Move this option before the file it belongs to.". I tried removing spaces from the filename bu that didn't help. Also when I selected "show landmarks" it said in the cmd window: File "G:\SECourses-VisoMaster-face-replacement\VisoMaster\app\processors\workers\frame_worker.py", line 255, in paint_face_landmarks img[int(kpoint[1])+i][int(kpoint[0])+j][0] = kcolor[0] ~~~^^^^^^^^^^^^^^^^^^ IndexError: index 995 is out of bounds for dimension 0 with size 772 Is there a way to solve this, and to get it to easily replace the face in the video (that is mostly a side view throughout, of someone playing piano).

Furkan Gözükara

i am waiting swarmui implementation to make tutorials :D following all. and this is different than wan this is face swapping with more features

DanO..

It seems as though WAN 2.2 Animate does this quite well. Plus everything else it does. Simplify it and get it working with lower VRAM and that would be great. I guess you might be busy with WAN 2.5 though!

Furkan Gözükara

let me update installer for those

Mike

Will this work with cuda 12.9 and cudnn 9.12?

Furkan Gözükara

yep it is up to date

Tyler

Is this still working well?

Furkan Gözükara

you need to have python 3.11 installed for this app to work. dont make it default just install as in tutorial video

umut yilmaz

ImportError: Error importing numpy: you should not try to import numpy from its source directory; please exit the numpy source tree, and relaunch your python interpreter from there.

Furkan Gözükara

Hello you need Python 3.11 for this. for Rope V14 you need Python 3.10. do you have both of them in your PC? also did you follow requirements step by step ? https://youtu.be/DrhUHnYfwC0

Kent Lindström

I get a flash_attn_2_cuda error when trying both this and the rope v14. ImportError: DLL load failed while importing flash_attn_2_cuda: Det går inte att hitta den angivna proceduren. Dont know what im doing wrong here 🫤

Lou

Something to keep in mind.... VisoMaster is completely uncensored. FaceFusion is censored. I'm not advocating the making of anything "unethical" BUT if FaceFusion decides your content is "unethical" you won't get the results your looking for. To me this is, in a sense, a performance hit. VisoMaster will give you exactly what your expecting provided you know what your doing, no matter the content. FaceFusion will limit your creativity. AI won't always get it right so your experience can become frustrating. Work smart, not hard.

Furkan Gözükara

nice info

Lou

There is a lot of info available on their discord. Unfortunately I don't have a link. I should also add, if your looking for quality results, you will most likely change settings frequently. For example, if your video target face is similar to your Swap face half the work is done for you. If it's not then your in for a lot of tweaking to make it work. It really depends on how picky you are. There are tricks I learned on my own which produce excellent results. The discord channel is a great resource.

Furkan Gözükara

hi your python installation is totally wrong. first uninstall it and then install python 3.10 and 3.11 and 3.12 as shown in this video : https://youtu.be/DrhUHnYfwC0

Furkan Gözükara

hi your python installation is totally wrong. first uninstall it and then install python 3.10 and 3.11 and 3.12 as shown in this video : https://youtu.be/DrhUHnYfwC0

Furkan Gözükara

if this is for visomaster please report the error here : https://github.com/visomaster/VisoMaster/issues

Fco Muñoz

Clearing Cache Successfully Stopped Processing play_video: Starting video processing. Starting video processing. Starting frame_read_timer with an interval of 26 ms. Stopping video processing. Adding audio... [in#1 @ 0000023241e65540] -to value smaller than -ss; aborting. Error opening input file H:/Nueva carpeta\0382e033-ea0c-4f98-ade9-a5bf13407826_video.mp4. Error opening input files: Invalid argument (on local machine)

Pascal Meienberg

File "", line 88, in _run_code File "C:\Users\USER\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.12_qbz5n2kfra8p0\LocalCache\local-packages\Python312\Scripts\crumb.exe\__main__.py", line 4, in File "C:\Users\USER\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.12_qbz5n2kfra8p0\LocalCache\local-packages\Python312\site-packages\crumb\main.py", line 6, in from crumb import Crumb, NotInRepoError ImportError: cannot import name 'Crumb' from 'crumb' (C:\Users\USER\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.12_qbz5n2kfra8p0\LocalCache\local-packages\Python312\site-packages\crumb\__init__.py) 'crumb' is not recognized as an internal or external command, operable program or batch file.

Pascal Meienberg

s\Python312\site-packages\crumb\main.py", line 6, in from crumb import Crumb, NotInRepoError i get this error....any ideas what i fucked up?

Furkan Gözükara

Sadly i don't have such comparison but currently best 2 ones are : VisoMaster_v1 and FaceFusion_8 - they are most up to date and getting updated and developed

Mimic

VisoMaster_v1, Rope Alucard24_V13, Roop Unleashed_v6, FaceFusion_8 ... Where can I find a comparison of their performance and features? Or is VisoMaster just the best? Thanks a lot.

Furkan Gözükara

we have these 2 tutorials https://youtu.be/whDt36YwEKQ https://youtu.be/RdWKOUlenaY

BecauseReasons

Thanks for this, did you ever make a tutorial for it?

Furkan Gözükara

I wish i knew. the only way is trying and seeing sadly

Unrecognised

is there a way to find out what the setting options all mean? ive got it working but i have no idea what settings i should be trying to tweak things

Furkan Gözükara

yes you need requirements installed. great solution to have ubuntu 22 as well

363ls2gto

To clarify, this is not a 1 click installer and does not install python so I need to have python 3.10 already on my system. Much easier to use masscompute as they use Ubuntu 22.04 which runs Python 3.10 natively. It was fairly easy to dual install Python 3.10 on windows and edit the path per your video, but looks a little more complicated on Ubuntu 24 which runs on Python 3.12. Pyenv from GitHub looks promising. And your massedcompute install script would also need the python version clarification as you stated. Might be easiest to just install Ubuntu 22.04 on a spare disk to run all the massedcompute stuff easily at home. Thank you.

Furkan Gözükara

yes. but you need to have python 3.10 in your system to generate venv of that. just modify the script to generate python 3.10 venv like below this is windows example py -3.10 -m venv venv this is in ubuntu (didnt test) python3.10 -m venv venv

363ls2gto

Ubuntu 24 basically has to run on python 3.12. I thought the point of a Venv was to contain programs so that you could run different versions of i.e. python without affecting the entire system?

Furkan Gözükara

it is included and automatic

Furkan Gözükara

not yet. i am trying to purchase rtx 5090 but due to shameless nvidia they are being sold from 4200 USD. waiting to get it under like 3600 USD

pb1323

Do you have instructions for cuda 12.8 / blackwell support?

Furkan Gözükara

0 reason. you must have enviroment error. you need cuda 12.4 and python 3.10.11 and cudnn and git and ffmpeg and c++ tools installed

363ls2gto

Any reason the instructions for massed compute and installer would not work on my home version of Ubuntu 24?

Skuuurt

Do you have to manually install Tensor RT optimization by Glat0s or is it included in the install ?

Furkan Gözükara

For RunPod i recommend roop unleashed https://www.patreon.com/posts/119081500 or face fusion : https://www.patreon.com/posts/103765029 they are gradio based