FaceFusion V3.4 (Most Advanced 1-Click DeepFake) 1-Click Installers for Windows, RunPod and Massed Compute - Use locally and cloud

Furkan Gözükara

I don't know also I don't support NSFW.

Cawao Kaito

Furkan Gözükara

no you have to have python and git and ffmpeg c++ tools and msvc installed. so first follow this video - 1 time necessary. then simply downloading zip files will work with just 1 click : https://youtu.be/DrhUHnYfwC0

Cawao Kaito

Can I install it by simply downloading the FaceFusion folder and double clicking the files inside? Will it download all the necessary installers without having to install Python, git, or conda separately to support rtx5000 series?

Furkan Gözükara

you can delete zip files. it should work now since those are extracted into accurate folders.

David

Hi, I installed it and in the main root I have TensorRT_10.9.0.34.zip and cudnn_9.12_cuda_12.9.zip, what I have to do?

Furkan Gözükara

currently facefusion is most advanced up to date best one

Patrick Major

Is this better than the Rope that you have in different sections? I've tried Rope and the UX is quite beautiful (for something that is free). The ability to play and see "live-ish" representations of your changes really helps. Since you have experience with both, which one do you go to most?

Furkan Gözükara

dont run as administrator that is your problem

Rahul Gautam

Its installing in C drive although I had put the installer in D drive...how Can i change the install location?

Furkan Gözükara

that means that pod was broken. get a new pod.

dokoissho

RunPod PyTorch 2.8.0 container doesn't start, even after setting container disk to 50Gb and 200Gb network storage

Furkan Gözükara

just updated and fixed all. please use this template: RunPod Pytorch 2.8.0 runpod/pytorch:2.8.0-py3.11-cuda12.8.1-cudnn-devel-ubuntu22.04

Furkan Gözükara

i dont know you need to play with parameters and test, it supports too many

Myworld Hello

cool1

With FaceFusion, is there a way to edit a video with someone in but keep the same face but any place where their eyes look down, change it so their eyes look straight ahead instead (facing the camera)? I saw in the info about it that it has "Expression Adjustments". Could that allow that changing eye direction (to straight ahead always)? Or is there something else that could do that better?

Robert Waters

If you want to bypass the censorship you can replace the content_analyser.py file with: from typing import List # NSFW censorship disabled: all analysis functions now always allow content def pre_check() -> bool: # Skip any model checks return True def clear_inference_pool() -> None: # Stub for FaceFusion pipeline cleanup, no-op pass def analyse_stream(vision_frame, video_fps) -> bool: # Never flag any frame return False def analyse_frame(vision_frame) -> bool: # Never flag any frame return False def analyse_image(image_path: str) -> bool: # Never flag any image return False def analyse_video(video_path: str, trim_frame_start: int, trim_frame_end: int) -> bool: # Never flag any video return False

Furkan Gözükara

well my downloader downloads and sets into downloaded folder. so you have to use it otherwise really hard to make work

Furkan Gözükara

well this is not developed by me so your best shot is opening issue threads

koon grizzly

It doesn't want to open the browser by itself when i run the windows start bat file ? C:\audiocraft\audiocraft_plus\venv\lib\site-packages\gradio\components\dropdown.py:231: UserWarning: The value passed into gr.Dropdown() is not in the list of choices. Please update the list of choices to include: No models found or set allow_custom_value=True. warnings.warn( Running on local URL: http://127.0.0.1:7861 So i went to that url myself and it's working fine from there. Quickly tested, is this only photo to photo ? or can it do video like Visomaster? Also it has a lot of censorship, everything nude gets blurred results

decoRyder

OK, wonderful work, as always, and thank you very much. I have CUDA 12.8, and CUDNN 9.8, paths correctly set, and they work in for example ComfyUI. When I run facefusion, I get this: 2025-05-13 09:31:43.3240701 [E:onnxruntime:Default, provider_bridge_ort.cc:2036 onnxruntime::TryGetProviderInfo_CUDA] D:\a\_work\1\s\onnxruntime\core\session\provider_ bridge_ort.cc:1695 onnxruntime::ProviderLibrary::Get [ONNXRuntimeError] : 1 : FAIL : LoadLibrary failed with error 126 "" when trying to load "F:\SECourses\FaceFusion_ 14\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_providers_cuda.dll I saw that you set the capi path in the script, and no matter if I set the path in my ENV or not I still get the same error. LoadLibrary 126 I think refers to a dendency issue - does facefusion need a specific version of CUDA? Thanks!

Furkan Gözükara

sadly i dont have you need to try and compare

PhoB J

Are somewhere any suggestion, what settings use, to squeze most of the framerate? I mean execution thread and queue counts, memory strategy and limit. Im on ryz 9900x, 16gb 5070ti gpu, 64gb ram.

Furkan Gözükara

it is your system related. something in your system blocking the graido app. first time seeing. but you can do that. my runpod and massed compute has enabling share feature. download this : https://huggingface.co/MonsterMMORPG/Generative-AI/resolve/main/face_fix_next.py rename it to default.py put it inside facefusion/facefusion/uis/layouts start and let me know if works

PhoB J

Hi, installed all, run Facefusion app bat and throwing: "ValueError: When localhost is not accessible, a shareable link must be created. Please set share=True or check your proxy settings to allow access to localhost. " (it happends the same error also when i install it simply in Pinokio). Can you help?

Furkan Gözükara

not yet sadly

Roy Davis

Is there a youtube tutorial for this?

Michał Kowalski

I think roop is nice for live cam but face fusion gave me better results in quality after swapping, sometimes have to use for example topaz video to make it upscaled, I noticed that facefusion have potential to do it in one move

Furkan Gözükara

i dont have atm for this but you can watch this and it is same to install and start : (starts at 22:03) : https://youtu.be/KW-MHmoNcqo?si=QN8X8Sjn13ZYu-EU&t=1323

rino damar jati

especially in runpod

rino damar jati

can you show me the tutorial to do this?

Furkan Gözükara

i see. i dont know atm either

bazzamore

I did look into it a bit further. I believe the training was done via DeepFaceLab then the training is exported to a ".dfm" file.

Furkan Gözükara

i don't know that. did you find any more info regarding that?

bazzamore

Hi Furkan, I noticed in FaceFusion they have option "deep_swapper" which allows users to select a premade face models. Is there a way to train our own models? These seem to produce somewhat better results in some cases.

Furkan Gözükara

yes please no NSFW sharing here

Error_404_unknown

"You have any idea how to remove nsfw censor or any kind of censorship in facefusion?" I found a way but i don't think i can say here.

Furkan Gözükara

makes sense

jw

facefusion has better eyes and mouth movement adaptation and better colors, but it is way slower than rope , i think due to the use of live_portrait

Lou

Yup, the dev start.bat and dev install.bat should be included in the main rope install folder, from secourses.

Furkan Gözükara

we have Alucard24 main branch installer and i see both are updated last 2 weeks ago

Lou

Alucard24 has a fork of rope with the option to run the development branch. It has features available not in the stable version. Run the development bat install/start in the rope directory to check it out. For me, it runs stable and has the best outputs I've seen so far. I haven't tried Facefusion yet.

Furkan Gözükara

awesome thanks for letting people here know as well

Long Phan

see discord for the debug info but basically installing cuda 12.4 fixed the issue (download links on the public tutorial page: https://www.patreon.com/posts/111553210 , thanks a lot !

Furkan Gözükara

you are welcome

Lou

Your right. curl.exe was blocked. Thanks.

Arcon Septim

Np figured it out

Arcon Septim

You have any idea how to remove nsfw censor or any kind of censorship in facefusion?

Furkan Gözükara

i think it is better than roop. but i cant say for sure better than rope pearl - next

Arcon Septim

so it's not good as roop?

Furkan Gözükara

can you show me screenshots of paths of user and system and also screenshot of capi folder and the error you can message from discord

Furkan Gözükara

it installs some libraries system wide true can break. best to run in a new machine

Arcon Septim

This facefusion completely broke my masedcompute, nothign works

Arcon Septim

Is this better than ROOP

Long Phan

hmm, the dlls are all there and the environment path has this: (if I execute echo %PATH% from the commande line) \FaceFusion_6\facefusion\venv\Lib\site-packages\onnxruntime\capi; But I still get the same issue.

Furkan Gözükara

then manually verify if all files looking accurate. the capi folder has dlls and enviroment paths are there or not

Long Phan

I still have the exact same issue, even with the v6 installer. I'll try to reboot

Furkan Gözükara

hi there was an error please try v6 zip file : Windows_Fix_cuDNN_Errors_Run_As_Administrator.bat

Furkan Gözükara

ah thanks for reminder i am going to add tqdm package now

Furkan Gözükara

hi it must be your antivirus blocking downloading files. you need to fix it somehow . i would say also try warp vpn

Furkan Gözükara

hi you need cudnn fix. verify that dll files are inside D:\FaceFusion_5\facefusion\venv\lib\site-packages\onnxruntime\capi and also this path added to the top of your user path and system path in environment

Albert

I followed the instructions and the red error when generating. 2024-10-05 14:56:26.9995408 [E:onnxruntime:Default, provider_bridge_ort.cc:1992 onnxruntime::TryGetProviderInfo_CUDA] D:\a\_work\1\s\onnxruntime\core\session\provider_bridge_ort.cc:1637 onnxruntime::ProviderLibrary::Get [ONNXRuntimeError] : 1 : FAIL : LoadLibrary failed with error 126 "" when trying to load "D:\FaceFusion_5\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_providers_cuda.dll"

Lou

After install and the second step, followed by Start Windows.bat, it just hangs after selecting CUDA execution providers. Downloading:0 percent 0/8kb file=open_nsfw.hash cpu is working at something but nothing happens, just hangs. Did this with the last version too so I gave up. Thought this new version might work.

ApathyJones

Nice! The newest instructions as of Oct 5 fixed all the Cuda errors for me. The .bat file didn't work because it says it is missing the tqdm package, so I just did the steps manually in lieu of installing it, but regardless, it works now finally :)

Furkan Gözükara

just updated the installers and added info please follow new info and installers

Furkan Gözükara

just updated the installers and added info please follow new info and installers

Furkan Gözükara

just updated the installers and added info please follow new info and installers

Furkan Gözükara

just updated the installers and added info please follow new info and installers

Furkan Gözükara

just updated the installers and added info please follow new info and installers

Furkan Gözükara

hopefully making a new video today wait it please - windows 11 fresh install

Shepard4k

can't get CUDA to work, even with LTSC 17.8 no idea what i'm doing wrong :( [E:onnxruntime:Default, provider_bridge_ort.cc:1992 onnxruntime::TryGetProviderInfo_CUDA] D:\a\_work\1\s\onnxruntime\core\session\provider_bridge_ort.cc:1637 onnxruntime::ProviderLibrary::Get [ONNXRuntimeError] : 1 : FAIL : LoadLibrary failed with error 126 "" when trying to load "C:\Users\Shepard4k\FaceFusion_3_v1\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_providers_cuda.dll"

Furkan Gözükara

awesome i am glad you made it work

Error_404_unknown

I had the same error, It's taken me 2 days of install/delete reinstall. Finally I deleted everything py, cuda, VS, ONNX, CDnn and followed the tutorial and now it works. Everything else is now broken but I guess a reinstall of everything will fix that. I needed to clean the junk from my drive anyway this had just expidited the process.

Furkan Gözükara

hi it wont work anymore the developer of facefusion deleted older githubs - merged you can manually git checkout to specific commits but may not work if you upgrade gold member i can connect your pc and fix

Paul Pfab

hi, can u provide a link to the last version of your windows installer, that s working with cuda 11.8, please? (before facefusion_3_v1). thx

Furkan Gözükara

it is not sufficient our app downloads and extracts dll files. look at the installer source code and verify that those were put into places accurately . face fusion is so annoying to make it run if you also upgrade to gold membership i can connect your pc and try to fix

Paul Pfab

hi, i think i made cuda 12.4 default and i installed LTSC 17.8. getting this: " [E:onnxruntime:Default, provider_bridge_ort.cc:1992 onnxruntime::TryGetProviderInfo_CUDA] D:\a\_work\1\s\onnxruntime\core\session\provider_bridge_ort.cc:1637 onnxruntime::ProviderLibrary::Get [ONNXRuntimeError] : 1 : FAIL : LoadLibrary failed with error 126 "" when trying to load "C:\Users\paul\Facefusion\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_providers_cuda.dll" nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2024 NVIDIA Corporation Built on Tue_Feb_27_16:28:36_Pacific_Standard_Time_2024 Cuda compilation tools, release 12.4, V12.4.99 Build cuda_12.4.r12.4/compiler.33961263_0 any idea? regards paul

Furkan Gözükara

my installer downloads them auto working. if not working your solution is this post : https://www.patreon.com/posts/solving-cuda-and-111553210 follow that post and make cuda 12.4 default

jw

same problem for cuda.dll and tensorrt.dll, i think you have to fix the requirements.txt based on this: https://github.com/microsoft/onnxruntime/issues/22019

Furkan Gözükara

ye library error. set cuda 12.4 default and follow here : https://www.patreon.com/posts/solving-cuda-and-111553210 after that reinstall

Tobe2d

I trying using update file and it did not work so I deleted venv and run install again and now i have this error: Enter the execution provider (1 for CUDA - preffered and faster, 2 for CPU): 1 Using CUDA execution provider Traceback (most recent call last): File "F:\face_fusion_v3\facefusion\facefusion.py", line 7, in from facefusion import core File "F:\face_fusion_v3\facefusion\facefusion\core.py", line 8, in from facefusion import content_analyser, face_classifier, face_detector, face_landmarker, face_masker, face_recognizer, logger, process_manager, state_manager, voice_extractor, wording File "F:\face_fusion_v3\facefusion\facefusion\content_analyser.py", line 7, in from facefusion import inference_manager, state_manager, wording File "F:\face_fusion_v3\facefusion\facefusion\inference_manager.py", line 5, in import onnx File "F:\face_fusion_v3\facefusion\venv\lib\site-packages\onnx\__init__.py", line 77, in from onnx.onnx_cpp2py_export import ONNX_ML ImportError: DLL load failed while importing onnx_cpp2py_export: A dynamic link library (DLL) initialization routine failed. Press any key to continue . . .

Furkan Gözükara

it has some advantageous like interface is faster. both are pretty par at the moment

Wendel President

Better than Rope for you?

Furkan Gözükara

sadly impossible you need to use cloud

James Adams

Please add support for AMD! Thank you :-)

Furkan Gözükara

it works already in both cases use latest installer

Furkan Gözükara

sadly not possible

James Adams

If you could add support for AMD via DirectML, that would be amazing :-)

James Adams

Hi, FF Next branch now requires CUDA 12.4.1 or TensorRT 10.3.0. CUDA 11.8 is no longer supported. Can you please update your installer? Thanks!

Furkan Gözükara

sorry for late reply. can you make a fresh install and send me all install cmd screen logs? so i can see the install error reason

Furkan Gözükara

sorry for late reply. can you make a fresh install and send me all install cmd screen logs? so i can see the install error reason

eyal giron

it's now give https://pastebin.com/TUKNY5Nt

eyal giron

i get this when trying start with GPU https://pastebin.com/Bn3dyDir

Furkan Gözükara

i think rope is the best i don't like facefusion dev :D

Jason Dalton

what would you say is the overall best for deepfake quality this or ROPE?

Furkan Gözükara

it is mistake of facefusion also it is development branch you can expect it to be broken :D

Furkan Gözükara

bunların next diye bir branchı var development branch , 2. zip onu yüklüyor

Cemil Hacimahmutoglu

hocam 2 zip dosyası arasındaki fark ne acaba

Tobe2d

After restarting it is working fine now. I dont know what was wrong ;-)

Tobe2d

It is in new folder. the only button not working is AGE MODIFIER

Furkan Gözükara

i told you to make a fresh install into a new folder :) or you have to keep restarting until it deletes and redownloads all

Tobe2d

When I try to activate AGE MODIFIER I get this error: [FACEFUSION.DOWNLOAD] Validating hash for styleganex_age failed in FaceFusion NEXT

Furkan Gözükara

you are welcome thanks for support

Tobe2d

Thank you 🙏 This windows update is so important to keep up 😂

Furkan Gözükara

i am gonna add windows update button now to zip files

Furkan Gözükara

for next version please install into a new folder. for regular version you can do just git pull over the folder but there is no update atm

Tobe2d

What is the correct way to update the previews one we had on the 9th of August?

Furkan Gözükara

massed compute best. 31 cents per hour and much faster to install

Eyepatch_101

Ohh....okay...out of the two, which one do you recommend as I'm a occasional user (4-5 hrs usage per week) and I only follow your channel...should I go for massed compute or Runpod?

Furkan Gözükara

you have to use paid services or your computer. free service as kaggle and colab won't allow faceswapping they ban the account :/ like use runpod or massed compute

Eyepatch_101

Ohh, Thats bad...guess I'll have to wait

Furkan Gözükara

I am going to add v3 branch installer as well working on it now

Furkan Gözükara

Kaggle immediately banned my very old account while i was trying to install :/ I still couldn't get it opened yet sadly

Eyepatch_101

Hi Do you have the Kaggle version ready for this?

Furkan Gözükara

yes this runs locally. yes hopefully flux training tutorial coming soon

Carlos L. Romero

Can this run local? also flux lora training coming?

Furkan Gözükara

yes rope is more optimized and faster. but face fusion is also working for v3 which may speed up

Steve

Is it me or is this very slow compared to the rope application you posted?

Furkan Gözükara

can you give me its repo link? i think this is the only app they have. they also have v3 branch but still in development. by the way we have rope web cam live version : https://www.patreon.com/posts/most-advanced-1-105123768 use zip file : Rope_Live_Stream_v4.zip

wei smith

can you make 1 click installer for facefusion webcam live?

Furkan Gözükara

I am using Cuda 11.8 in my system and facefusion running. I am using onnxruntime==1.17.1 onnxruntime-gpu==1.17.1

Blatta

Thank you so much for your fast support! I followed the tutorial very closely and was able to get your installs for A1111, Roop unleashed, CodeFormer and LivePortrait_v3 running. But with FaceFusion, when i hunt down the error messages in an internet search, i end up thinking the onnxruntime don´t fit the Cuda 11.8. The requirements.txt points to 11.17.3, the onnxruntime-gpu --upgrade from the bat file leads to a install of 1.18.1. According to https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html this would require Cuda 12.*. But i only understand like 1% of all of this, so i could be completely on a wrong track. You think i should run the steps from the tutorial all again and just give it a new try?

Furkan Gözükara

we automatically install specific onnxruntime. so cuda 11.8 is really necessary. easiest way is please follow this tutorial exactly : https://youtu.be/-NjNy7afOQ0

Blatta

The required onnxruntime 1.17.3 needs CUDA 12.x. From your tutorial i have installed CUDA 11.8. What would be the easiest fix for someone new to windows?

Furkan Gözükara

it is the best gpu. so my guess would be that you have installation error. please reinstall python c++ exactly as shown here into C drive and uninstall previous ones : https://youtu.be/-NjNy7afOQ0

Anshul Gupta

i have rtx 4090 paired with i7 14700k

Furkan Gözükara

What is your gpu? For Cuda you need Nvidia gpu

Furkan Gözükara

You can't solve it unless the app you are using have optimizations that you didn't enable yet

Furkan Gözükara

It depends on each model. You can check their command line arguments to give model folder paths. I show this in my tutorial videos

Furkan Gözükara

I will hopefully make a tutorial for the best ai voice cloning. It is not ready yet sadly

Anshul Gupta

One last help is needed from you i think Piper TTS is the best voice cloning and Text-to-speech generation AI tool. Can you help me in the cloning of voice as directed by the YouTube channel "Natlamir" in his video "Local Piper TTS Training (Using WSL)" Url link of this video is "https://www.youtube.com/watch?v=kGl0a652XtE&t=1s". Please help me to clone voices in English and hindi. i am ready to pay for it to you. Please Please

Anshul Gupta

Please help on how to reduce the redundancy of large AI models that is large, how to set the path in each model for each AI program and tool, the same models are used by stable diffusion automatic 1111 web UI and stable swarm UI. How to collect all these models in one place and then how to use them for all AI tools to set the path of folders of these models

Anshul Gupta

can you help me and guide me on how to solve the error "cuda out of memory", I think you should make a video on these types of error

Anshul Gupta

Hi, sir glad to talk with you, you are a genius, I installed face Fusion with your 1 click installer and I am getting the error "run.py: error: argument --execution-providers: invalid choice: 'cuda' (choose from 'CPU')". Please help me to solve it

Sahil Sangani

I used that before still same. but now i used pinokia installer it worked so thankfully finally. thanks :)

Furkan Gözükara

cuda 12.1 won't work. you have to have cuda 11.8. follow this video and set exactly as in there : https://youtu.be/-NjNy7afOQ0 also install into d:/face_ai folder nothing else. your folder path has space character

Sahil Sangani

No error while installation. just this error while using GPU mode. CPU works fine but too slow so i need GPU

Sahil Sangani

2024-06-15 04:25:47.8215414 [E:onnxruntime:Default, provider_bridge_ort.cc:1744 onnxruntime::TryGetProviderInfo_CUDA] C:\a\_work\1\s\onnxruntime\core\session\provider_bridge_ort.cc:1426 onnxruntime::ProviderLibrary::Get [ONNXRuntimeError] : 1 : FAIL : LoadLibrary failed with error 126 "" when trying to load "D:\1Git\face fusion\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_providers_cuda.dll" *************** EP Error *************** EP Error C:\a\_work\1\s\onnxruntime\python\onnxruntime_pybind_state.cc:866 onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasnt able to be loaded. Please install the correct version of CUDA andcuDNN as mentioned in the GPU requirements page (https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements), make sure they're in the PATH, and that your GPU is supported. when using [('CUDAExecutionProvider', {'device_id': '0', 'cudnn_conv_algo_search': 'DEFAULT'})] Falling back to ['CUDAExecutionProvider', 'CPUExecutionProvider'] and retrying. **************************************** 2024-06-15 04:25:47.8783911 [E:onnxruntime:Default, provider_bridge_ort.cc:1744 onnxruntime::TryGetProviderInfo_CUDA] C:\a\_work\1\s\onnxruntime\core\session\provider_bridge_ort.cc:1426 onnxruntime::ProviderLibrary::Get [ONNXRuntimeError] : 1 : FAIL : LoadLibrary failed with error 126 "" when trying to load "D:\1Git\face fusion\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_providers_cuda.dll" Traceback (most recent call last): File "D:\1Git\face fusion\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py", line 419, in __init__ self._create_inference_session(providers, provider_options, disabled_optimizers) File "D:\1Git\face fusion\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py", line 483, in _create_inference_session sess.initialize_session(providers, provider_options, disabled_optimizers) RuntimeError: C:\a\_work\1\s\onnxruntime\python\onnxruntime_pybind_state.cc:866 onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasnt able to be loaded. Please install the correct version of CUDA andcuDNN as mentioned in the GPU requirements page (https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements), make sure they're in the PATH, and that your GPU is supported. The above exception was the direct cause of the following exception: Traceback (most recent call last): File "D:\1Git\face fusion\facefusion\run.py", line 6, in core.cli() File "D:\1Git\face fusion\facefusion\facefusion\core.py", line 112, in cli run(program) File "D:\1Git\face fusion\facefusion\facefusion\core.py", line 234, in run ui.launch() File "D:\1Git\face fusion\facefusion\facefusion\uis\core.py", line 88, in launch ui_layout_module.render() File "D:\1Git\face fusion\facefusion\facefusion\uis\layouts\default.py", line 45, in render preview.render() File "D:\1Git\face fusion\facefusion\facefusion\uis\components\preview.py", line 42, in render conditional_append_reference_faces() File "D:\1Git\face fusion\facefusion\facefusion\core.py", line 280, in conditional_append_reference_faces reference_face = get_one_face(reference_frame, facefusion.globals.reference_face_position) File "D:\1Git\face fusion\facefusion\facefusion\face_analyser.py", line 447, in get_one_face many_faces = get_many_faces(vision_frame) File "D:\1Git\face fusion\facefusion\facefusion\face_analyser.py", line 504, in get_many_faces bounding_box_list_yoloface, face_landmark_5_list_yoloface, score_list_yoloface = detect_with_yoloface(vision_frame, facefusion.globals.face_detector_size) File "D:\1Git\face fusion\facefusion\facefusion\face_analyser.py", line 257, in detect_with_yoloface face_detector = get_face_analyser().get('face_detectors').get('yoloface') File "D:\1Git\face fusion\facefusion\facefusion\face_analyser.py", line 95, in get_face_analyser face_detectors['yoloface'] = onnxruntime.InferenceSession(MODELS.get('face_detector_yoloface').get('path'), providers = apply_execution_provider_options(facefusion.globals.execution_device_id, facefusion.globals.execution_providers)) File "D:\1Git\face fusion\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py", line 432, in __init__ raise fallback_error from e File "D:\1Git\face fusion\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py", line 427, in __init__ self._create_inference_session(self._fallback_providers, None) File "D:\1Git\face fusion\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py", line 483, in _create_inference_session sess.initialize_session(providers, provider_options, disabled_optimizers) RuntimeError: C:\a\_work\1\s\onnxruntime\python\onnxruntime_pybind_state.cc:866 onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasnt able to be loaded. Please install the correct version of CUDA andcuDNN as mentioned in the GPU requirements page (https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements), make sure they're in the PATH, and that your GPU is supported. i have PYTHON 3.10.11. C++ INSTALLED, PYTHON DEVELOPMENT INSTALLED. CUDA 12.1 installed. i added cuDNN to path after downloding it. i installed everything 3 times still same error please help

Furkan Gözükara

then install it as shown in this video : https://youtu.be/-NjNy7afOQ0

Adan Aguilar Cisneros

[FACEFUSION.CORE] FFMpeg is not installed i get this error when trying to start the app

Furkan Gözükara

if you become a Gold member i can connect your computer and debug the reason. sadly that message not telling a lot

Ivan Stoyneshki

I did, I run cuda 11.8 from the tutorial

Furkan Gözükara

please install python cuda c++ as in this video : https://youtu.be/-NjNy7afOQ0

Ivan Stoyneshki

how to mitigate this error - RuntimeError: D:\a\_work\1\s\onnxruntime\python\onnxruntime_pybind_state.cc:857 onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasnt able to be loaded. Please install the correct version of CUDA andcuDNN as mentioned in the GPU requirements page (https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements), make sure they're in the PATH, and that your GPU is supported.

Furkan Gözükara

great

Furkan Gözükara

this works great as well

droidgame maniac

very informative video, but just use pinokio for facefusion. just works without any problems

Mika Kujanpää

That's correct, and deepspeed dropped some other module version after installing it. I might have made some error because I had plenty of problems with onnxruntime. But I was able to solve them.

Furkan Gözükara

my installers makes their own virtual environments. so as long as you install them into different folders, they will not affect each other.

Mark Sutherland

Thanks for the reply. I solved it by just using the cpu. But to answer your question it's on a laptop with a GTX1050i. My PC has a A4000 (16gb) but I have that set up with stable diffusion and one of your one click installs of InstantID and everytime I install something it overwrites the versions etc so I use the laptop for stuff I'm not fussed about. But thanks again for the reply.

Furkan Gözükara

what is your python version and did you install c++ tools? what is your GPU? reinstall and send me all installation log so i can see if any errors : monstermmorpg@gmail.com i think you didnt follow this tutorial to install as shown there : https://youtu.be/-NjNy7afOQ0

Furkan Gözükara

i think you dont need should work

Mark Sutherland

I've installed the facefusion but get this error, any help appreciated cheers 2024-05-07 10:13:52.1596399 [E:onnxruntime:Default, provider_bridge_ort.cc:1548 onnxruntime::TryGetProviderInfo_CUDA] D:\a\_work\1\s\onnxruntime\core\session\provider_bridge_ort.cc:1209 onnxruntime::ProviderLibrary::Get [ONNXRuntimeError] : 1 : FAIL : LoadLibrary failed with error 126 "" when trying to load "C:\Users\Mark\Documents\facefusion\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_providers_cuda.dll" 2024-05-07 10:13:52.3393932 [E:onnxruntime:Default, provider_bridge_ort.cc:1548 onnxruntime::TryGetProviderInfo_CUDA] D:\a\_work\1\s\onnxruntime\core\session\provider_bridge_ort.cc:1209 onnxruntime::ProviderLibrary::Get [ONNXRuntimeError] : 1 : FAIL : LoadLibrary failed with error 126 "" when trying to load "C:\Users\Mark\Documents\facefusion\facefusion\venv\lib\site-packages\onnxruntime\capi\onnxruntime_providers_cuda.dll"

Mika Kujanpää

I'm installing this to Linux. Should I downgrade triton from 2.2.0 to 2.1.0? Also deepspeed from 0.14.2 to 0.11.2?