windows FaceFusion安装
显卡安装cuda
1. 安装检测
在安装之前先打开CMD窗口,粘贴如下命令,查看您的系统中是否已安装CUDA。
nvcc -V
如果您的电脑中已安装过CUDA,将会看到如下提示:
C:\Users\guoyg>nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2024 NVIDIA Corporation
Built on Fri_Jun_14_16:44:19_Pacific_Daylight_Time_2024
Cuda compilation tools, release 12.6, V12.6.20
Build cuda_12.6.r12.6/compiler.34431801_0
2. 适配下载
在安装CUDA之前,我们应该先知道自己的电脑应该安装哪个版本的CUDA,因为新旧版本兼容性不同。
想查看自己应该安装的CUDA版本,先在CMD命令行中执行如下命令:
nvidia-smi
在执行完上述命令之后,将会自动弹出如下内容,我们只需要查看第一行中的CUDA Version: 12.6这一部分,这代表需要下载CUDA版本为12.1的安装程序。
C:\Users\guoyg>nvidia-smi
Thu Aug 29 09:12:15 2024
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 560.76 Driver Version: 560.76 CUDA Version: 12.6 |
|-----------------------------------------+------------------------+----------------------+
| GPU Name Driver-Model | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA GeForce RTX 4060 Ti WDDM | 00000000:01:00.0 On | N/A |
| 0% 45C P8 10W / 165W | 791MiB / 16380MiB | 2% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| 0 N/A N/A 2260 C+G ...oogle\Chrome\Application\chrome.exe N/A |
| 0 N/A N/A 2464 C+G ....7122.0_x64__8wekyb3d8bbwe\Todo.exe N/A |
| 0 N/A N/A 2916 C+G ...siveControlPanel\SystemSettings.exe N/A |
| 0 N/A N/A 7644 C+G ...iversal\CtyunClouddeskUniversal.exe N/A |
| 0 N/A N/A 7844 C+G ...\bin-7.1.9\Nutstore.WindowsHook.exe N/A |
| 0 N/A N/A 8736 C+G D:\Typora\Typora.exe N/A |
| 0 N/A N/A 9364 C+G C:\Windows\explorer.exe N/A |
| 0 N/A N/A 9816 C+G ...nt.CBS_cw5n1h2txyewy\SearchHost.exe N/A |
| 0 N/A N/A 10360 C+G ...cal\Microsoft\OneDrive\OneDrive.exe N/A |
| 0 N/A N/A 11316 C+G ...2txyewy\StartMenuExperienceHost.exe N/A |
| 0 N/A N/A 12640 C+G ...les\AMD\CNext\CNext\AMDRSSrcExt.exe N/A |
| 0 N/A N/A 12800 C+G D:\PyCharm 2023.2.5\bin\pycharm64.exe N/A |
| 0 N/A N/A 13356 C+G ...ekyb3d8bbwe\PhoneExperienceHost.exe N/A |
| 0 N/A N/A 14868 C+G ...\AMD\CNext\CNext\RadeonSoftware.exe N/A |
| 0 N/A N/A 15148 C+G ...tstore\bin-7.1.9\NutstoreClient.exe N/A |
| 0 N/A N/A 16112 C+G ...5n1h2txyewy\ShellExperienceHost.exe N/A |
| 0 N/A N/A 18308 C+G ...__8wekyb3d8bbwe\WindowsTerminal.exe N/A |
+-----------------------------------------------------------------------------------------+
3. 下载地址
CUDA官网:CUDA Toolkit Archive | NVIDIA Developer
打开CUDA官网,在其中找到适合自己的版本,找到所需要的CUDA版本是12.6.0,所以直接点击CUDA Toolkit 12.6.0。
在点击适合自己的CUDA版本之后,将会看到如下界面,我们按照自身的系统情况选择下载即可:


4. 安装教程
双击运行下载的CUDA EXE安装包文件,下载后打开将会看到如下界面,然后根据图片上方的文字提示操作即可。
安装cuDNN
cuDNN(CUDA Deep Neural Network library)是由 NVIDIA 开发的一个加速深度神经网络的库,它为深度学习应用提供了 GPU 加速的功能。cuDNN 主要提供了深度神经网络的基本操作,例如卷积、池化、归一化等等,这些操作都可以在 GPU 上进行加速。在安装cuDNN之前,我们需要先安装Windows版本的CUDA
1. Nvidia|注册|登陆
值得注意的是在注册完善个人资料时要勾选加入NVIDIA开发人员计划,访问下载内容(如cuDNN)、操作视频等。选项,必须要勾选,否则注册后也无法下载cuDNN。
2. cuDNN|下载
请确保你已经完成了Nvidia账号的注册或登陆,否则无法下载下载。
点击此处前往cuDNN的下载页面,打开之后将会看到如下所示的内容,我们需要选择适合自己的版本。

每一个下载后面都会有2个版本号,【Download cuDNN v8.9.2 (June 1st, 2023), for CUDA 12.x】,第一个是cuDNN的版本,其次是CUDA的版本。CUDA版本为12.1,那么就可以下载CUDA 12.x的版本,因为这其中的x是通用的意思。
所以,如果你的CUDA版本是11.x,那么就下载第2个即可,其它同理。
点击该按钮之后将会自动展开该选项,这将会包含所有相关系统的cuDNN下载地址,我们选择Local Installer for Windows (Zip)按钮即可自动开始下载。
3. cuDNN配置
需要知道自己的CUDA安装位置,如果你不知道自己的CUDA安装在什么位置,可以在CMD中执行如下内容。
where nvcc
执行完上面的命令之后,将会自动获取CUDA的安装目录。
C:\Users\guoyg>where nvcc
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\bin\nvcc.exe
只要到版本号这一目录即可,也就是C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6,后面的部分省略。
注意:每个人所获取到的版本可能不同,所以不要照抄,一定要自行获取。
此时将我们刚刚下载好的压缩包文件解压,然后将文件内的所有文件移动是我们获取到的目录内。(如果提示已存在,直接选择覆盖所有文件即可。)
4. 添加环境变量
我们完成上面的所有步骤之后,还需要将相关目录添加至系统环境变量内。
在安装CUDA时,应该在默认情况下已经添加了两个关于CUDA的环境变量,分为如下:
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\CUDA版本\bin
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\CUDA版本\libnvvp
我们还需要添加额外两个,我们刚刚已经获取到了CUDA的安装目录,另外两个文件夹也位于该目录内,分别是include文件夹和lib文件夹。
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\lib
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include
5. 验证安装
进入到C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\extras\demo_suite 目录中打开命令行窗口,然后执行如下命令:
bandwidthTest.exe
返回以下内容则代表配置成功:
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\extras\demo_suite>bandwidthTest.exe
[CUDA Bandwidth Test] - Starting...
Running on...
Device 0: NVIDIA GeForce RTX 4060 Ti
Quick Mode
Host to Device Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 6493.7
Device to Host Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 6433.9
Device to Device Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 218805.8
Result = PASS
NOTE: The CUDA Samples are not meant for performance measurements. Results may vary when GPU Boost is enabled.
在执行如下命令:
deviceQuery.exe
返回以下内容则代表配置成功:
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\extras\demo_suite>deviceQuery.exe
deviceQuery.exe Starting...
CUDA Device Query (Runtime API) version (CUDART static linking)
Detected 1 CUDA Capable device(s)
Device 0: "NVIDIA GeForce RTX 4060 Ti"
CUDA Driver Version / Runtime Version 12.6 / 12.6
CUDA Capability Major/Minor version number: 8.9
Total amount of global memory: 16379 MBytes (17175019520 bytes)
MapSMtoCores for SM 8.9 is undefined. Default to use 128 Cores/SM
MapSMtoCores for SM 8.9 is undefined. Default to use 128 Cores/SM
(34) Multiprocessors, (128) CUDA Cores/MP: 4352 CUDA Cores
GPU Max Clock rate: 2565 MHz (2.57 GHz)
Memory Clock rate: 9001 Mhz
Memory Bus Width: 128-bit
L2 Cache Size: 33554432 bytes
Maximum Texture Dimension Size (x,y,z) 1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
Maximum Layered 1D Texture Size, (num) layers 1D=(32768), 2048 layers
Maximum Layered 2D Texture Size, (num) layers 2D=(32768, 32768), 2048 layers
Total amount of constant memory: zu bytes
Total amount of shared memory per block: zu bytes
Total number of registers available per block: 65536
Warp size: 32
Maximum number of threads per multiprocessor: 1536
Maximum number of threads per block: 1024
Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
Max dimension size of a grid size (x,y,z): (2147483647, 65535, 65535)
Maximum memory pitch: zu bytes
Texture alignment: zu bytes
Concurrent copy and kernel execution: Yes with 1 copy engine(s)
Run time limit on kernels: Yes
Integrated GPU sharing Host Memory: No
Support host page-locked memory mapping: Yes
Alignment requirement for Surfaces: Yes
Device has ECC support: Disabled
CUDA Device Driver Mode (TCC or WDDM): WDDM (Windows Display Driver Model)
Device supports Unified Addressing (UVA): Yes
Device supports Compute Preemption: Yes
Supports Cooperative Kernel Launch: Yes
Supports MultiDevice Co-op Kernel Launch: No
Device PCI Domain ID / Bus ID / location ID: 0 / 1 / 0
Compute Mode:
< Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >
deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 12.6, CUDA Runtime Version = 12.6, NumDevs = 1, Device0 = NVIDIA GeForce RTX 4060 Ti
Result = PASS
安装FaceFusion
打开官网https://docs.facefusion.io/installation,Installation 选择Windows然后执行对应的命令即可

1. Prepare Your Platform
GIT
winget install -e --id Git.Git
Conda
winget install -e --id Anaconda.Miniconda3 --override "/AddToPath=1"
FFmpeg
winget install -e --id Gyan.FFmpeg
Codec
winget install -e --id CodecGuide.K-LiteCodecPack.Basic
2. Prepare Your Environment
Installation
conda init --all
conda create --name facefusion python=3.10
Usage
conda activate facefusion
3. Install Your Accelerator
CUDA
conda install conda-forge::cuda-runtime=12.4.1 cudnn=8.9.2.26 conda-forge::gputil=1.4.0
conda install conda-forge::zlib-wapi
OpenVINO ⁿᵉˣᵗ
conda install conda-forge::openvino=2024.2.0
4. Download Your Copy
Clone the repository:
git clone https://github.com/facefusion/facefusion
Ensure to enter the directory:
cd facefusion
5. Install The Application
python install.py
6. Done
Finally, run the command:
python run.py
注意:第一次启动会自动下载模型,耗时巨长,可以自行下载好模型后再启动
直接访问对应的url下载后将模型放到models文件夹下即可

可能出现的错误
启动后,上传照片后执行换脸时报错
[Could not locate zlibwapi.dll. Please make sure it is in your library path](https://stackoverflow.com/questions/72356588/could-not-locate-zlibwapi-dll-please-make-sure-it-is-in-your-library-path)
解决方案
进入目录C:\Program Files\NVIDIA Corporation\Nsight Systems 2024.4.2\host-windows-x64可以找到zlib.dll文件,将其复制为C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\bin\zlibwapi.dll即可