Creating free images with an image generation AI

in #stablediffusion2 years ago (edited)

2870380115_An_epic_painting__a_mix_of_Moria_and_Coruscant_by_John_Avon__3D__realistic_graphics__photorealism___cinematic__Unreal_Engine__V_Ray__path_tracing__dynamic_lighting__dynamic_shadows__dynamic_reflection__ray_tra.png

Hey steemit bloggers!

Do you know the situation? You want to write a blog and you are looking for an image that's suitable for your text. There are the free image galleries and they are fine, but sometimes you need a special image and you can't find it...

Yesterday, I accidently stumbled across an Image generation AI and it blew my mind. I just sat there creating image after image and forgot about time. At DreamStudio you can try it out for yourself!

The Stability AI uses the Stable Diffusion process and it creates images from text. Basically, you write a description of what you want to see, which style you want to use and maybe add some key words, you click on the Dream button and see the result. This is what "an epic portrait of Elrond the half elf, realistic, oil painting" can look like:

967022612_An_epic_portrait_of_Elrond_the_half_elf__realistic__oil_painting.png

There are various parameters that you can play with, like the image width and height. Some are more advanced and tie into the mechanics of creating the image, I didn't do a lot with them.
After the AI created the image, you can download it and use it for personal and commercial purposes. The license allows you to do that, the developers just ask you to be responsible for what you create, in addition you should use the #StableDiffusion hashtag. Images created through DreamStudio are fully open source, explicitly falling under the CC0 1.0 Universal Public Domain Dedication.

2937919043_An_epic_mushroom_forest_with_some_dwarves__papyrus_style__3D__realistic_graphics__photorealism___cinematic__ArtStation__Unreal_Engine__Autodesk_Arnold__Pixar_RenderMan__V_Ray__path_tracing__dynamic_lighting__d.png

It may take some time to generate an image that you really can use. Finding out which key words work best and where the limits of the AI are is a little time consuming. At the moment, the AI has sometimes problems to generate images, for example images with a rider or an archer. They looked surreal or unfinished. But looking on free image galeries for something to fit your needs can take time, too. It's up to you where you spend it ;-)

If you are interested in the project, here is some further information.