Today I will explain how to the
artificial intelligence image generation program called "Stable Diffusion" to Krita. Stable diffusion is a free artificial
intelligence-based image generation program. You can generate high-resolution and unlimited images as long as your computer can handle it. First, if you don't have Krita on your computer, you should install it. For this video, I set up a clean Windows. If any issues arise, we will address them. I will download it from Krita.org. Version 5.2.0 has been released, and I was using it as a beta. I'm happy about that . I choose it because my operating system is Windows 11, but it also works on Linux and OS X. Krita is like an undiscovered treasure. I'm starting the installation I'm continuing by clicking the "Next" button. There's nothing special. I've sped up the installation video to save time. It usually takes 1-2 minutes. Krita has been installed and is working. I'm closing Krita and starting to install the plugin.
I will also add it to the description below I prepared the link for this beforehand. I will also add it to the description below. I'm copying and pasting it into the browser.
A new version of this is also available. I hope I hope it has improved. I should mention that it's quite successful. I also installed the Stable Diffusion program separately. Although it's good for fine-tuning, I must say it's not as easy as this. I will download this ZIP file it's only
606 kilobytes .When installed it will be around 15 GB. I will open this file in the folder where Krita plugins are located. This is important . I must open it in the correct folder. You should
find this folder according to your username. If the AppData folder is not visible check the "Show hidden
files" option in the view menu. The target folder is named "pykrita." I'm opening the zip file here. Let's Open Krita I select "Settings" from the
menu and then "Python Plugin Manager." It should appear here. I need
to activate it and restart the program I'm activating "Settings" - "Dockers"
and "AI image generation" It didn't work. I should create a blank file first Yes; it's open now. I press the config figure button. Here you can see where the installation will take place. It requires at least 10 gigs
of space. Since my graphics card is NVIDIA, CUDA should be selected. I'm pressing
the install button. The rest should be installed automatically. It takes a little
time. I will fast forward the video here and there The installation is complete but it gave
an error. I think I made a mistake here I had performed a clean Windows installation I didn't update the graphics card driver. I believe the error is due to this. I will uptate my graphics card and try again The graphics card driver is installed. Now I'm trying again yes it worked Two checkpoints have been loaded. One is for photo stic images and the other is for Imaginary drawings You can also load different checkpoints later Let me write something simple let's draw a cat It seems to be missing a leg These things usually happen Let me try something more complex While it's working let me take this Let me enlarge these layer
images so we can see them more clearly The image has arrived this is not very good I'm adding it to a layer by pressing
the "Apply" button. it was just a Preview moment ago. I will keep the prompt
the same and try changing the checkpoint. I might resort to a little trick. My ChatGPT account is not active on this computer right now. For now, I'll manage with Bing but ChatGPT is definitely more successful. it's struggling a lot I'm copying this and pasting it into Krita Let's see what happens. it looks a bit better let me try something else I ask it for a lural landscape
of a fantasy world this time By the way our previous Crystal drawing is complete I was expecting a drawing closer to digital art but it's not that beautiful therefore you should try downloading different checkpoints from CivitAI.com and experiment. Dream shaper 10 XL yields better results. While Dreamshaper 8 is only 2 GB, 10 XL is 6.4 GB in size. Generating images also takes a long time, but the results are more satisfying. So, I recommend experimenting with different models. I'm saving the file There's nothing more to do at a basic level. You can try it out for yourself. I've put in quite an effort to create these kinds of images. Thank you for watching in the rest of the video I'm leaving a
few examples of the images I created using this method