This AI Turns Your Rudimentary Doodles Into Realistic Photos
If you’re not a great artist but still want to create realistic images, then you need to check out this amazing AI tool.
The AI tool is called Diffuse The Rest which can be accessed for free here.
The dashboard consists of a blank canvas and a few painting controls, a text box for your prompt, a diffuse button, and an upload image button.
Transform Steps
The steps are easy. Draw your doodle art, write a little description of what the drawing is about, then click the “Diffuse the f rest” button.
Here’s a glorious drawing of a sunflower I did in under a minute.
The generation process takes only a few seconds and AI spits out four different variations. Here’s the result of my little sunflower.
Awesome, right?
The Hugging Face web app has a pretty basic painting menu so if you want more control, you can use other painting apps like Procreate, ClipStudio, or Gimp.
Upload An Existing Image
If you don’t want to bother with making a drawing by hand, you can upload any art you already have.
Just click “upload img,” describe the picture, and click “diffuse the f rest.” Here’s an example of a mug with a picture of a cute little puppy on it.
The final render looks incredibly realistic. One wouldn't immediately think this is AI-generated.
The Tech Behind
Stable Diffusion, an AI model that recently became open source, powers the web application.
Stable Diffusion is a text-to-image latent diffusion model created by the researchers and engineers from CompVis, Stability AI and LAION.
It’s trained on 512x512 images from a subset of the LAION-5B database. This model uses a frozen CLIP ViT-L/14 text encoder to condition the model on text prompts.
With its 860M UNet and 123M text encoder, the model is relatively lightweight and runs on a GPU with at least 10GB VRAM. See the model card for more information.
If you’re interested in running the AI text-to-image tool on your own hardware, follow this guide.
Can You Sell These Images?
Yes, you can sell or redistribute your AI-generated images under CreativeML Open RAIL-M license.
You are free to use them any way you want but they must not go against the provisions set in this license.
Final Thoughts
This tool is perfect for anyone who wants to be able to create images for their projects but doesn’t have the time or talent to do so. All you need is a basic drawing, and the AI will take care of the rest.
Also, if you’re bored and want to have some fun, this AI tool is a great way to entertain yourself.
I imagine this will be popular with children. Especially since kids have the wildest imagination, I am sure they’ll make some great art with this AI tool.
If you like to read more stuff like this in Medium, consider supporting me and thousands of other writers by signing up for a membership.