Microsoft has introduced a new model dubbed Visual ChatGPT, which combines ChatGPT with visual platform models such as Transformers, ControlNet, and Stable Diffusion. Not only can you import images and create new ones, but you can also edit your images.
The purpose of Visual ChatGPT is very simple. You can create and modify images in chat format, creating a different kind of user experience for working with AI bioarts and images.
While we’ve used specific platforms and apps for images and art, this combines the concept of chat + visual stimulation, that hasn’t really been explored.
For now, you have to download Visual ChatGPT via the GitHub page , but I predict that we will see this in other UIs soon. Possibly as a new feature in ChatGPT and exposed through its API.
There’s also a “hugging face” option if you want to test it out.
Below is the system architecture diagram… 
Basically, Visual ChatGPT uses biological algorithms to analyze image data, then predict pixel values to create new images. It combines both methods of giving visual information in the form of text to ask questions and get answers.
Visual ChatGPT can be used to create custom images, including landscapes, objects, animals, people, and more. It can also be used to edit existing images, including adding or removing objects, modifying colors or changing lighting.