真的很好用，参照作者博客的一键
git clone https://github.com/wvthoog/proxmox-vgpu-installer.git && cd proxmox-vgpu-installer && bash proxmox-installer.sh
会在当前的目录下创建几个文件，确保有空间
[+] Nvidia driver version: NVIDIA-Linux-x86_64-535.161.05-vgpu-kvm.run
[+] Nvidia driver properly loaded, version matches 535.161.05
[+] Enable nvidia-vgpud.service
[+] Enable nvidia-vgpu-mgr.service
[+] In your VM download Nvidia guest driver for version: 535.161.05
[-] Linux: https://storage.googleapis.com/nvidia-drivers-us-public/GRID/vGPU16.4/NVIDIA-Linux-x86_64-535.161.07-grid.run
[-] Windows: https://storage.googleapis.com/nvidia-drivers-us-public/GRID/vGPU16.4/538.33_grid_win10_win11_server2019_server2022_dch_64bit_international.exe
Step 2 completed and installation process is now finished.
List all available mdevs by typing: mdevctl types and choose the one that fits your needs and VRAM capabilities
Login to your Proxmox server over http/https. Click the VM and go to Hardware.
Under Add choose PCI Device and assign the desired mdev type to your VM
Removing the config.txt file.
[?] Do you want to license the vGPU? (y/n): n
Exiting script.
Install the Docker container in a VM/LXC yourself.
By using this guide: https://git.collinwebdesigns.de/oscar.krause/fastapi-dls#docker
Select vGPU driver version:
1: 17.0 (550.54.10)
2: 16.4 (535.161.05)
3: 16.2 (535.129.03)
4: 16.1 (535.104.06)
5: 16.0 (535.54.06)
Enter your choice: 5
[-] Driver version: NVIDIA-Linux-x86_64-535.54.06-vgpu-kvm.run
[+] Downloading vGPU NVIDIA-Linux-x86_64-535.54.06-vgpu-kvm.run host driver using megadl
Downloaded NVIDIA-Linux-x86_64-535.54.06-vgpu-kvm.run
[+] MD5 checksum matched. Downloaded file is valid.
[+] Patching driver
[+] Installing patched driver
[+] Driver installed successfully.
[+] Nvidia driver version: NVIDIA-Linux-x86_64-535.54.06-vgpu-kvm.run
[+] Nvidia driver properly loaded
[+] Enable nvidia-vgpud.service
[+] Enable nvidia-vgpu-mgr.service
[+] In your VM download Nvidia guest driver for version: 535.54.06
[-] Linux: https://storage.googleapis.com/nvidia-drivers-us-public/GRID/vGPU16.0/NVIDIA-Linux-x86_64-535.54.03-grid.run
[-] Windows: https://storage.googleapis.com/nvidia-drivers-us-public/GRID/vGPU16.0/536.25_grid_win10_win11_server2019_server2022_dch_64bit_international.exe
Step 2 completed and installation process is now finished.
List all available mdevs by typing: mdevctl types and choose the one that fits your needs and VRAM capabilities
Login to your Proxmox server over http/https. Click the VM and go to Hardware.
Under Add choose PCI Device and assign the desired mdev type to your VM
Removing the config.txt file.
[?] Do you want to license the vGPU? (y/n): y
[+] Installing Docker-CE
[+] Docker pull FastAPI-DLS
[?] Enter the desired port number for FastAPI-DLS (default is 8443):
[!] Don't use port 80 or 443 since Proxmox is using those ports
[+] Generate Docker YAML compose file
[+] Running Docker Compose
[+] Generate FastAPI-DLS Windows/Linux executables
[+] license_windows.ps1 and license_linux.sh created and stored in: /root/proxmox-vgpu-installer/licenses
[-] Copy these files to your Windows or Linux VM's and execute
Exiting script.