1 day Ago By Mid-day
Despite being around for nearly 20 years not many know about Krita. It is an open-source, free graphics editor that is designed primarily for digital art and 2D animation. It’s available on several operating systems, including Windows, macOS and various Linux distros.
In itself, Krita is an exceptional drawing program, offering support for drawing tablets, various brush options, layers, and excellent built-in animation tools. However, we are most interested in Krita’s ability to use plugins. Krita AI DiffusionThis is possibly the most important plugin you can install in Krita right now.
It seamlessly integrates Stable Diffusion into Krita by installing and running ComfyUI in the background. It comes with preset options, and you can build your own workflows, too.This image was drawn by hand while the image on the right is what was generated using this image as the baseIt enables you to leverage the latest image-generating models to help bring your ideas to life.
You can use either a cloud-based, comfortable UI installation or a local installation. This one plugin will let you generate images, perform inpainting, refine existing images, create live paintings, upscale images, and even customise poses.How to installThe first step is to download Krita.
(https://krita.org/en/)After installing Krita on your OS, download the Krita AI Diffusion Plugin (https://github.com/Acly/krita-ai-diffusion/releases)The next step is to start Krita.
Then, in the menu, go to Tools>Scripts>Import Python Plugin from File. Select the Krita Diffusion zip file you just downloaded. Once installation is done, restart Krita.
After restarting the app. Create a new image, and use any size you want. It doesn’t matter.
Once the image appears, click on Settings>Dockers> AI Image Generation. This will add AI Image Generation at the bottom of the right toolbox. In the AI Image Generation section, click the gear icon.
This will open a new window titled ‘Configure Image Diffusion’. In that window select the first option on the left called ‘Connections’. For this example, we will use a Local Managed Server.
Select a path where you want the AI model files to stay, make sure it has 30-40GB free space at least. Then under Workloads, select Stable Diffusion XL, everything under Upscalers, everything under Stable Diffusion XL models, and everything under NoobAI XL models. Click on Launch and go away for a couple of hours.
Depending on your Internet connection, this can take a while. It will download all the files, auto-configure Comfy UI, and get everything ready. That’s it.
Now, every time you start up, AI diffusion will also start Comfy UI in the background. Be warned that the time it takes to generate an image will strongly depend on the graphics card you have or lack thereof. Nvidia cards work the best.
Also, start with a smaller canvas; 512x512 pixels is ideal. You can always upscale. What can you do?With the AI plugin installed, you can for starters generate images.
It’s straightforward: enter a prompt in the box labelled ‘Describe the content you want to see.’ Then, click on Generate. The drop-down menu allows you to select from the available models.
You can even create your own workflow using the gears option on the right. The brush icons with the sparkles icon left in the AI Image Generation tab lets you select the mode. There is Generate, Upscale, Live, Animation and Graph.
The images above were made in the Live option.You can also select bits of the image and retouch or repaint using AI. There are also tutorials online to use depth maps and a pose tool.
These are linked on the Krita AI Diffusion Github page. You can also take images and use them as a base to recreate the images in specific styles, such as the Ghibli images that are currently everywhere. Krita with Diffusion is a game changer.
It works so well that it may be time to bid farewell to the not-so-highly affordable subscription of Photoshop.Important links>> Krita: https://krita.org/en/>> Krita Diffusion: https://>> kritaaidiffusion.
com/>> Krita Tutorial: https://docs.krita.org/en/>> Krita Diffusion Tutorial: https://docs.
interstice.cloud/.
