AI image generator takes a big leap › jojo Reviews

We have been following the developments around the AI ​​image generator Stable Diffusion with great interest for several months now. Not only because we are fascinated by the graphically impressive results that appear out of nowhere after entering short text descriptions, but also because the first simple implementations are now finally available to everyone.

Stable distribution 2

Stable Diffusion 2.0: generates images from scratch

Simple use becomes possible

With Mac applications such as Diffusion Bee or CHARL-E and iPhone applications such as the freeware download Draw Things recently presented on jojo Reviews, you do not need a computer science degree to do your own experiments with Stable Diffusion, and you have also no basic trust in any of the many online platforms that allow the use of stable distribution in the network.

With the macOS and iOS downloads, image creation can be tried out locally on your own machines using intelligent algorithms. Account and free without leaving data traces. And it can only get better.

Stable Pigeon Fusion Qualitaet

Stable Distribution 2.0: Graphical results should be of higher quality

Stable distribution 2.0

Version 2.0 of the stable diffusion model was released today on the GitHub code platform. The new version offers many great improvements. New text-to-image models are available that can generate images with significant quality improvements in 512px x 512px and 768px x 768px resolutions.

Low Res Super Res

Super Resolution Upscaler upgrades resolutions

A new super resolution upscaler can also upscale small images to resolutions up to 2048px x 2048px. The also new depth-to-image model extends the so-called image-to-image function in such a way that image input can be combined with text and depth information.

StableDiffusionV2 Depth2Img

Stable diffusion 2.0: the depth-to-image model

This allows images to be completely redesigned in terms of content, but while preserving the existing depth information and thus also their composition. Image component replacement would have been greatly improved by the new inpainting module.


Depth-to-image model in Aktion

We’ll be in touch when Stable Diffusion 2.0 is back in easy-to-use Mac and iPhone apps.

What do you think?

Written by Kim Anderson

Leave a Reply

Your email address will not be published. Required fields are marked *

GIPHY App Key not set. Please check settings

    It runs in December 2022 › jojo Reviews

    This is the Black Friday deal on eBay ›jojo Reviews