Vervette, A.I. and 21st Century Art Education:

Cool results!
Ai can (or not) make up for imagination !
Problem : how to do for see all that is created! ?

Dreamtime is a local Ai app very heavy (many gigas and with a very long complex install ?
So absolutely not for me and my very old config!
I use only online app !
Fortunately for me there are hundred of them And free!

[Taron]

I just tried that install of Dreamtime and it's a really bad nudifier?! Seems useless, but maybe there is another with the same name?
The online Dreamtime is with "credits" and I really don't like that.
Vizcom is thus far the best when it comes to just using it. Is there another like that?
Just like I expected there are now truly hundreds of them and they all try to earn money, which is- in my eyes- total bottom-feeding, because it's open source stuff and they're all just hoping to make a quick buck.

For absolutely free online with any number of images and very Speedy!
Just have a Gmail !
You have Wombo! Of course very simple without tones of regulates!
https://dream.ai/create
Perfect for you nothing to learn ! (more free time for coding!
Numerous styles maybe 80...only one format 960 x 1568 jpg in your download folder
in fact today 143 - 33 "Premium" = 110 free styles!
For the moment my favorite style is Botany v3

So Prompt, click Style, click Create
When image is good for you : Button "Finalize" --> Button "Download" --> "Original" (With a little Logo) or "Background image" (without Logo) !
From this Download section : "Generate again" if you want an another image with same regulates or go to "Back to creation for new regulates!
All is free!

My little trick : open 10 Tabs of Wombo!!! Like this you will have an image at each time you go to a tab!
So a calculated image (to finalized of course) at each click!

Have fun...of course i have also many another ones but this one is my favorite and my first one after Gaugan who don't exist anymore!
Some ones i have used https://moiscript.weebly.com/stable-diffusion.html (+ Menu Divers / Free On Line / Ai

Following : prompt : woman in forest
style Botany v3
You can of course refine it in free Vizcom (unlimited), or Free Krea (40 / day)

Dream_Background (31).jpg (168.35 KiB) Viewed 5402 times

Refine Vizcom 100 % (...asian)


Krea 100 % Ressemblance

I just got back in from a long day. It looks like to me that Wombo really is the best free option considering the output and quality of Pilous image sets. The one I use is mage https://legacy.mage.space (legacy is ending soon they have a new version) The moment you are there you can create as many as you like but 1 at a time amd you can only use 2 AI models. Dont worry. They are both very good and can get quality results... To bring an image in simply click the image Icon at the end of the descriptor box and you can also make everything you do private by clicking the eye icon. I paid a month for it to really dig deep. and plan on using it free again at the end of the month. https://www.daz3d.com/ai-studio can be used for free and will generate 4 images at a time if wanted but the number of images is limited and the AI is limited as well. It is an experiment that focuses on Daz 3d products. Still in spite of its limitations it can produce some really nice results. Yup.. it appears to me that Pilou's beloved Wombo is the best out of the pack. I would use them all

Thanks for the comments on Dreamtime... I wasnt aware there was an AI named Dreamtime lol. I named the version I used Prince of the Dreamtime a long time ago. Here is a blog link for more insight into the workings of AI https://www.jonstokes.com/p/getting-sta ... diffusion#§diffusion-steps

I wonder if we could build an AI that would teach art basics? Almost seems simple but when people say simple.. It always becomes complicated

[Taron]

Well, it's highly conceivable to create an A.I. Art Teacher, but a pain in the butt to pull off. The trick is in the enormous training-set you'd have to assemble and you have to accept the revelation that stable diffusion is ultimately far from "intelligent". Like someone who read thousands of books and can recite any given page without knowing what it means whatsoever. ...so, yeah, I call A.I. actually H.L. (Hyper Librarian), hehehehe.

Yeah, Wombo was really one of the first that felt accessible and pleasant. Styles are a plenty and the process is pleasantly simple, but the restrictive image format and general sense of "dead-end" makes it not nearly as engaging as Vizcom. But Vizcom- as we have established already- makes asian girls. My goodness, if they were to share image-pools with Wombo...pffff.

I made an enhance with Stylar who has a new Portrait enhance system without deformation... Cosmetic...No deformation!
(the eyes and hairs) for the rest my old eyes don't see many change!

Free but credits (200) / month! https://www.stylar.ai/faq#link
an Enhance = 8 Credits!

@ Tartan
I don't know the DAZ AI!
I know Daz because it has 2 French products Bryce 3D By Eric Wenger, and Hexagon without Nurbs but cool polymodeler ( successor or Amapi the first to blend Nurbs and Poly)
Sometime free, sometime not...

Edit: seems no more available ?
Curious because
https://www.daz3d.com/bryce-7-pro
https://www.daz3d.com/hexagon-3d-modeling-free

Actually you can get both. But Bryce is supposed to be free also and Still manages to produce quality images that have an illustrative feel to them. Vue knocked it out of contention long ago. I have been involved with Zygote/Daz3d a long time and have a very small fingerprint in Daz3d s emergence from my days as one of the founders of Renderosity.com in the early 2000s. Taron and you should check it out from time as they improve their 3d AI based on their products.

Taron you should try mage then, its very accessible, and has a huge hyper library. https://legacy.mage.space and the new mage site https://www.mage.space. Post back if you like it.You are absolutely correct about asian girls lol. I consider Vizcom to have the best potential and like their creative approach. You are also correct in the incredible amount of effort it would take. But if we blink our eyes we will see it. https://www.producthunt.com/categories/ ... rative-art has some interesting approaches to AI.

Pilou I didnt know there were any image limitations with Wombo. If its a size limitation thats easy for me to get around with Topaz Gigapixel (an older version) I bought a few years back. When my current Mage runs out I will go and test it extensively. I like the experiments you just showed.

Taron if you feel like it post back here or to Pilous anything you do with Mage or Daz.. same to Pilou. Maybe we can brainstorm an approach to true education and development of the human side of things as we fumble our way into the future.