EXTREME IMPORTANT NOTE!!!!:
Update 1 May 2024:
Update the optimized version of ROCmLibs
RX 6750XT included in this category -(gfx1031) GPU

Update 30 April 2024:
To install latest version, omit the git checkout part
If you already install it and want to update to latest version,
1. At the folder you install the stable diffusion, go to the address bar, type cmd and press enter
2. git checkout master
3. webui-user.bat

Update 20 April 2024:
If you have "ZLUDA device failed to pass basic operation test: index=None, device_name=AMD Radeon 'GPU' [ZLUDA]", try this:
At the folder you install the stable diffusion, go to the address bar, type cmd and press enter
1. venv\Scripts\activate
2. pip uninstall torch torchvision -y
3. webui-user.bat
---------------------------------------------------------------------------------------------------------------------------------------------------------------
List of AMD GPUs that work with ZLUDA
7900 XTX
7900 XT
7800 XT
7700 XT
7600
6950 XT
6900 XT
6800 XT
6750 XT
6700 XT
6700
6650 XT
6600 XT
6600

Music
Classic 1985 Movie Soundtrack by FreedomTrailStudio
[ Ссылка ]

Donate me at
Kofi: [ Ссылка ]
Paypal: tongtong83@gmail.com

Note:
Works from version 1.8.0-RC to 1.9.0

Installing Python 3.10.6
[ Ссылка ]

Installing Git
[ Ссылка ]

Download ZLUDA
[ Ссылка ]

Installing HIP SDK
[ Ссылка ]

Add System Environment Variables
%HIP_PATH%bin
D:\PATH\zluda

Download ROCmLibs Optimized For (credits to @BrknSoul ):
6700, 6700xt or 6750xt (gfx1031) GPU
[ Ссылка ]=

6600, 6600xt, or 6650xt (gfx1032) GPU
[ Ссылка ]=

Update Rocblas Library Folder
%HIP_PATH%bin\rocblas

Git Clone Code
git clone [ Ссылка ] && cd stable-diffusion-webui-directml && git submodule init && git submodule update

Git Checkout Code
git checkout 25a3b6c

All stable diffusion with ZLUDA installation out there needs to take 30 minutes or more for image generation for the first time and my tutorial is the only one in youtube that will run instantly!

1. Setup Python and Git:
First, you need to install two programs: Python and Git. Just follow the link provided and choose the right installer for your computer. Make sure to pick the 64-bit version. When installing, there will be some options to select; make sure to tick the box that says "Add to Path" for Python and then proceed with the installation.

2. Setup ZLUDA and HIP SDK:
Next, download ZLUDA and extract it somewhere on your computer. I put mine in a folder called "path" on my D drive. Then, download HIP SDK from the link provided in the video. Remember to agree to their terms before downloading. When you're installing the HIP SDK, make sure to choose the option for a full installation each time.

3. Update Environment Variables:
Now, we need to update some settings on your computer. Type "environment variable" in the search bar on your computer's taskbar. Then, click on "Edit the system environment variables." A window will pop up. Click on "Environment Variables" again. Look for "Path" under System variables and double-click it. Then, click "New" and add the paths for where you put ZLUDA and the HIP SDK.

4. Extra Step for Certain GPUs:
If your computer's GPU falls into this category, you'll need to do an extra step. Download a file called ROCmLibs from the link provided in the video. Type in a code in your computer's file explorer to find a folder, and then rename library folder to "libraryoriginal." After that, move library folder from the ROCmLibs.zip file you downloaded into the same place.

5. Setup Stable Diffusion:
Create a new folder where you want to set up your stable diffusion. Type "CMD" in the address bar and press enter. Then, paste git clone code from the video description. Once it's done, paste git checkout code to make sure you're installing the right version. Next, copy your stable diffusion model checkpoint into this specific folder. If you don't do this, stable diffusion will just download a default model for you. Finally, click on a file called webui-user.bat. You don't need to change anything in there. And that's it! You've got your stable diffusion running with ZLUDA. If you liked the video, subscribe to the channel.

Timecode
0:00 - Intro
0:22 - Supported GPUs
0:29 - Setup Python and Git
0:52 - Setup ZLUDA and HIP SDK
1:25 - Update Environment Variables
2:02 - Extra Step for Certain GPUs
2:35 - Setup Stable Diffusion

#AIForAMD #StableDiffusionForAMD #SDAMD #stablediffusionZLUDA

Related Video:
[ Ссылка ]

свернуть