Guide Fake Video Face Swap (moderate complexity, Rope)

[ghibit]

Please, Log in or Register to view quotes

Installation guide is here:

Please, Log in or Register to see links and images

The performance is dependant on your pc + gpu but it is exactly the same if you install Rope independently or with VoC.

 

[dudeuber]

I'm getting an error when running the rope bat "modulenotfounderror: no module named cv2" any ideas?

 

[ghibit]

Please, Log in or Register to view quotes

If you followed the guide in the first post and got that error try reinstalling it using the other guide in the youtube video posted in this thread.

 

[dudeuber]

ok i take it back. following the video worked. all you have to do before this is download python and anaconda prompt then follow the video instructions.

 

[thebadbone]

Please, Log in or Register to view quotes

Running on a RTX 3070 with a R7 5700X and 32 GB DDR4

 

[Dvorak]

Please, Log in or Register to view quotes

Error when I try to use GFPGAN, without it it runs, but with low quality.

 

[bobjeff69]

Hi, im having an issue where my Source folder will not load any of the images I put in, I have tried both png and jpg. Any help?

 

[jj1888]

Please, Log in or Register to view quotes

When I've had this issue it has been because one of the files just would not work, try having one photo in your source folder and adding one at a time to find out which one is causing an issue, out of around 40 photos I had 5 which wouldn't work.

 

[JonasMan]

Any help with this error?

C:\Windows\System32\Rope>call venv\Scripts\activate.bat
Traceback (most recent call last):
File "C:\Windows\System32\Rope\Rope.py", line 3, in <module>
from rope import Coordinator
File "C:\Windows\System32\Rope\rope\Coordinator.py", line 4, in <module>
import torch
ModuleNotFoundError: No module named 'torch'
Press any key to continue . . .

Using pip show torch command, it shows that torch is installed.

Name: torch
Version: 2.0.1+cu118
Summary: Tensors and Dynamic neural networks in Python with strong GPU acceleration
Home-page:

Please, Log in or Register to see links and images

Author: PyTorch Team
Author-email: packages@pytorch.org
License: BSD-3
Location: c:\users\user\appdata\local\programs\python\python310\lib\site-packages
Requires: filelock, jinja2, networkx, sympy, typing-extensions
Required-by: basicsr, clip, facexlib, gfpgan, open-clip-torch, torchaudio, torchvision

 

[Axelhart69]

Please, Log in or Register to view quotes

Do you have rope installed in the system32 folder? Seems a weird location ngl :D

Did you activate the environment with the venv command?

 

[JonasMan]

Please, Log in or Register to view quotes

Yeah somehow it chose system32 folder as default. and yes i did venv command
Changed installation location and it works fine.

 

[leersfan34]

Please, Log in or Register to see links and images

I see at the start of the thread it says you do not need a Nvidia GPU. But this prompts me after I double-click the rope.bat file asking me for a Nvidia GPU. Is there another way I am missing as to how to do it with AMD GPU's? Or is it simply strictly Nvidia as of right now?

 

[tremordick]

Please help. Error in Rope opal.3 when using GPFGAN:
Traceback (most recent call last):
File "D:\laragon\bin\python\python-3.10\lib\tkinter\__init__.py", line 1921, in __call__
return self.func(*args)
File "D:\laragon\bin\python\python-3.10\lib\tkinter\__init__.py", line 839, in callit
func(*args)
File "D:\Work\Projects\STD\Rope\rope\Coordinator.py", line 55, in coordinator
vm.get_requested_video_frame(action[0][1], marker=False)
File "D:\Work\Projects\STD\Rope\rope\VideoManager.py", line 228, in get_requested_video_frame
temp = [self.swap_video(target_image, self.current_frame, marker), self.current_frame] # temp = RGB
File "D:\Work\Projects\STD\Rope\rope\VideoManager.py", line 574, in swap_video
img = self.func_w_test("swap_video", self.swap_core, img, fface[0], s_e, parameters, control)
File "D:\Work\Projects\STD\Rope\rope\VideoManager.py", line 622, in func_w_test
result = func(*args, **argsv)
File "D:\Work\Projects\STD\Rope\rope\VideoManager.py", line 749, in swap_core
swap = self.func_w_test('Restorer', self.apply_restorer, swap, parameters)
File "D:\Work\Projects\STD\Rope\rope\VideoManager.py", line 622, in func_w_test
result = func(*args, **argsv)
File "D:\Work\Projects\STD\Rope\rope\VideoManager.py", line 1105, in apply_restorer
self.models.run_GFPGAN(temp, outpred)
File "D:\Work\Projects\STD\Rope\rope\Models.py", line 152, in run_GFPGAN
io_binding.bind_output(name='output', device_type='cuda', device_id=0, element_type=np.float32, shape=(1,3,512,512), buffer_ptr=output.data_ptr())
File "D:\laragon\bin\python\python-3.10\lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py", line 582, in bind_output
self._iobinding.bind_output(
RuntimeError: Failed to find NodeArg with name: output in the def list

 

[coins]

sess = C.InferenceSession(session_options, self._model_path, True, self._read_config_from_model)
onnxruntime.capi.onnxruntime_pybind11_state.NoSuchFile: [ONNXRuntimeError] : 3 : NO_SUCHFILE : Load model from .\models\det_10g.onnx failed:Load model .\models\det_10g.onnx failed. File doesn't exist

I can't find source faces

I'm using Windows 10

 

[thomas1986]

I had the same problem (Failed to find NodeArg with name: output in the def list), I downloaded all the models files from

Please, Log in or Register to see links and images

and then it was OK.

 

[googly3]

Will this work on Intel GPU? Installed the CUDA and keep running into "no Nvidia driver" error on Anaconda prompt. Followed the video from a previous post. Everything was good until then

 

[stargateluffy]

hi does this software works with amd gpu or is it only for nvidia?

 

[latinaover]

which visual studio do i need for this? im installing via winget and it tells me that it can't find the native desktop package and nothing is selected so i dont know which to pick, thanks

 

[axic]

Please, Log in or Register to view quotes

download the c++ module via

Please, Log in or Register to see links and images

- the download at most will be 8-9GB
winget never works in grabbing the correct package needed for this stuff so I'd just recommend against it.