Since 2022, we have been dealing with IAIA generating quality images, starting with Midjourney, Stable Diffusion or Dall.e. Now, if you were able to try such applicationsapplications in their early days, you may remember that we had a crazy amount of freedom then. It was easy to represent personalities such as Elon MuskElon Musk, Tom Hanks, Anne Hathaway and other celebrities. Only here: the cabinetslawyerslawyers stepped up to the plate and issued warnings to the leading companies in the field. Nowadays, there is no point asking Dall.e 3 or Leonardo to display personalities from the world of entertainment, business or politics side by side. They refuse, explaining that such a request is contrary to their regulations.
Let’s face it, it’s still a shame to have lost this ability. Let’s rest assured: it is still possible to achieve this by using products created by independent developers. We will do this here using a tool, DeeFooocus, which is a variant of an image generative AI called Fooocus.
Living or virtual celebrities
Here are some examples of visuals that we were able to create with DeeFoocus. You can see that it was possible to represent politicians as actors but also video game characters living together in the same shot like Lara Croft and Pikachu. We were even able to play BugsBugs Bunny at tennis with Mick Jagger.
Image generative AI: how to obtain extraordinary images with your prompts?
To be able to access DeeFooocus, we need to launch a relatively complex code from a programming interface. GoogleGoogle : Google Colab. Rest assured: you don’t need to understand the code involved here. Go to this address – just click on this link:
https://colab.research.google.com/github/ehristoforu/DeFooocus/blob/main/DeFooocus_colab.ipynb
Towards the top, you see a mention: Launch the DeFocus interface. Click on the arrow on the left. The arrow turns into a square surrounded by a rotating half-circle. You will see a series of obscure instructions appear on the screen. Don’t worry about it. You must wait several minutes for the processing to be completed. Ultimately you will see a line appear with this mention: Running on public URL. Click on the link to the right of this notice to access DeeFoocus.
Use the DeeFooocus
The basic use of DeeFooocus is simplicity itself: you type a prompt and it produces the requested image – you see it gradually build up. And there you can have fun. Note that we tried certain prompts in French and deduced that the application understands our language.
If you are not satisfied with the default result, click Advanced and select Quality in the Performance section – it will be slower but of a higher standard. A little further down, you can choose to generate more than one image for a single prompt (Image Number) which gives you more choices.
And There you go. The initial procedure for launching DeeFooocus may not be simple, but here you are dealing with a completely free and unrestricted generative AI. So, you might as well take advantage of it.
