Stable Diffusion AI is the latest trend that’s making rounds on social media, but what is it and how can you use it to generate images?
If your social media feed has been bombarded with anime or futuristic portraits, it’s because AI filters have been all the hype in recent weeks.
Fans have been playing around with filters readily available on platforms such as Meitu and TikTok, while MyHeritage is used for historical portraits through its “AI time machine.”
AI image fans are now looking to Stable Diffusion as the latest method to generate images rather than simply placing filters on top of existing photos. Read on to find out more about the new technology.
What is Stable Diffusion AI and how does it work?
Stable Diffusion is a machine-learning model that converts text into realistic, high-resolution images. It enables users to produce artwork quickly, thus allowing creatives to generate more ideas. It can also be used for inpainting and outpainting.
Created in collaboration between Stability AI, Runway, and LMU Munich’s CompVis group, Diffusion models are used to create data that resembles the data they were trained on, writes Medium.
Stable Diffusion works by adding noise to an image. The model then reverses the noising process and gradually improves the quality of the image until there is no noise, thus generating a realistic image to match the text prompt.
Popular models include Open AI’s Dalle-E 2, Midjourney, and Dream Studio.
@squteart A human without a soul?… these eyes, empty… scary. It moved me. #women #art #midjourney #photography #portait #religion #soul #beautifulgirls #fyp #fy #usa_tiktok #us #englishtiktokvideo #stablediffusion #photoshopediting
♬ Infinity – Jaymes Young
How to use the image generator
There are many free websites to try Stable Diffusion, including:
- Dream Studio
- Dream By Wombo
- Hugging Face
- Night Cafe
To create your next AI-generated artwork, create an account (if needed), type in any prompt you desire, and select the art style you want.
For example, if you type in “a dream of a distant galaxy, concept art, matte painting,” the model will create a futuristic landscape in space.
The best thing about stable diffusion is the options are endless so, even if you process the prompt again, it’s very unlikely to receive the same results.
Hugging Face also allows users to generate images from image inputs, so if you have poor artistic skills a simple drawing can be transformed into a more realistic version.
Simply upload your image and click submit for a new impressive makeover.
@amanabdulalim getting my feet wet with stable diffusion! #techtok #stablediffusion #dalle2 #aiart #codinglife #softwareengineer
♬ Your Day Off – Flying Lotus
