RuinedFooocus

The Future of AI Art is Here - Introducing RuinedFooocus

Forget everything you thought you knew about AI art generation - RuinedFooocus is here to completely reinvent the game!

This groundbreaking new image creator combines the best aspects of Stable Diffusion and Midjourney into one seamless, cutting-edge experience. The days of messy installations and manual tweaking are over. With RuinedFooocus, stunning images spring to life with just a few words - no technical skills required.

Built from the DNA of its predecessors but enhanced with next-level AI, RuinedFooocus generates jaw-dropping visuals in seconds. Just type a prompt and watch in awe as your wildest creative visions are instantly brought to life. The possibilities are truly endless.

But the magic doesn't stop there. RuinedFooocus puts the power in your hands like never before. You're in complete control of the experience from start to finish. And with its streamlined interface and intuitive controls, anyone can create like a pro.

Don't waste another minute stuck in the past. The future is calling, and it's time to answer with RuinedFooocus - the new standard in AI artistry. Install it today and unlock your creative potential like never before.

The future of art starts now!

Download

Windows

You can directly download Fooocus with:

>>> Click here to download <<<

After you download the file, please uncompress it, and then run the "run.bat".

In the first time you launch the software, it will automatically download models:

1. It will download sd_xl_base_1.0_0.9vae.safetensors from here as the file "Fooocus\models\checkpoints\sd_xl_base_1.0_0.9vae.safetensors".

If you already have these files, you can copy them to the above locations to speed up installation.

Note that if you see "MetadataIncompleteBuffer", then your model files are corrupted. Please download models again.

Linux

The command lines are

git clone https://github.com/runew0lf/RuinedFooocus.git
cd RuinedFooocus
conda env create -f environment.yaml
conda activate fooocus
pip install -r requirements_versions.txt

Then download the models: download sd_xl_base_1.0_0.9vae.safetensors from here as the file "Fooocus\models\checkpoints\sd_xl_base_1.0_0.9vae.safetensors", and download sd_xl_refiner_1.0_0.9vae.safetensors from here as the file "Fooocus\models\checkpoints\sd_xl_refiner_1.0_0.9vae.safetensors". Or let Fooocus automatically download the models using the launcher:

python launch.py

Or if you want to open a remote port, use

python launch.py --listen

Mac/Windows(AMD GPUs)

Coming soon ...

Ruined Edition Features

1. Supports custom styles in settings/styles.csv
2. Changed Resolutions and Styles to be in a dropdown instead of radio buttons
3. Apply multiple styles to one prompt and a send style to prompt button.
4. Ability to save full metadata for generated images embedded in PNG.
5. Ability to change default values of UI settings (loaded from settings/settings.json).
6. Generate a completely random prompt (taken from onebuttonprompt) with its own "special" tab (now officially supported by Arljen)
7. Made Resolutions mode readable
8. Added theme support in settings/settings.json you can see available themes HERE
9. Added settings/resolutions.json to allow users to add their own custom resolutions
10. Wildcards are now supported place see wildcards\colors.txt for an example. In your prompt make sure you type __<filename>__ to activate ie shiny new __colors__ Chevrolet pickup truck with big wheels
11. If the option --nobrowser is passed the web browser won't automatically launch
12. Added Custom paths in settings/paths.json to point to chekcpoints / loras and outputs director (Note: for windows paths either use / or \\ instead of \)
13. Added support for custom Performance - enables samplers/scheduler, steps, refiner steps, cfg & clip skip (Check advanced tab)
14. Displays time taken for each render in the console. If notification.mp3 exists in the root directory, this will play when the generations are complete.
15. Added Cancel button to stop generation. Thanks to Yownas and MoonRide
16. Pressing Ctrl-Enter is the same as pressing the generate button!
17. Adds the ability to add loras to just prompt ala auto111 style eg <lora:MyAwesomeLora-SL:0.90>
18. Drag and drop previous Ruined images to the main window, it will autofill the prompt with the generation metadata.
19. The ability to post json metadata into the prompt window so you dont have to go into ANY settings tabs when you click generate. (Note: setting a seed of -1 will generate a random seed)
20. Render different subjects so you can process a whole list of prompts. Seperate each prompt by placing --- on a new line
21. Supports subdirectories for models/loras so you can get all organised!
22. Add style:stylename to prompt to set the required style
23. Controlnet! Check out the PowerUp tab, once selected you'll be able to upload your controlnet base image into a new special area in the tab
24. Save your own custom perfomances and easily load them back in with the dropdown
25. Create your own default ControlNet settings, or just get into the hardcore details with the custom PowerUp mode
26. Image2Image mode is now active and available in the PowerUp tab
27. Added Support for SSD-1B Models
28. Automatically read triggerwords from <lora_filename>.txt just place a .txt file of the same name (minus the .safetensors) and it will read the triggerwords from the file
29. Upscaling using whatever upscaler you prefer 4xUltrasharp by default. Look for the new option in the powerup tab
30. Metadata Viewer now available in the Info tab
31. Now supports the SDXL LCM LORA

Thanks

This codebase is a fork from the original amazing work by lllyasviel The wonderful MoonRide who also maintains an amazing fork! The codebase starts from an odd mixture of Automatic1111 and ComfyUI. (And they both use GPL license.)

Discord Support Server

You can join our discord support server Here

Update Log

The log is here.