QUO VADIS - SCHIELE on NIGHTCAFE

“I believe that at the end of the century the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted.”
― Alan Turing - Computing machinery and intelligence

gi5yj1ecCsDPbejKNkAP--3--j7267_9.6154x.jpg

EGON'S SISTER PEEK-A-BOO

^{this one was initially created on DreamWombo and the result taken to Nightcafe as a seed image. I also used a different mode of Stable Diffusion (more about that below).}

I have been quite involved lately creating AI images on Nightcafe. Recently I joined a chat group of creators that belive in sharing prompts. Already prior to this I had cooperated with a creator/artist who had also not hidden his prompts, and the one I use for 'Schiele' came from him, which I adapted and added modifiers.
The amazing thing is that with the same prompt you could get many different results as well as styles (more about this below), not just by using different seed images, as shown in these two examples.
G9owDw8DsxGE4UhzEmfr--1--g2izt_11.3636x.jpg

SPACE PILOTS ON DOGON^{for this one I used one of my paintings as a seed image.}
Except at rare occasions I always use seed images, most times my own stuff, be that artwork or photography. I work most of the time in Stable Diffusion. Rarely in DALL-E2 where you cannot use seed images.

The 'Schiele' Prompt:

Text Prompt
"self aware machines by egon schiele austin osman spare, toyen, caravaggio jean benoit Post-Production, artstation, john salt detailed intricate, dystopian hypermaximalist, dreamlike ornate, hyperrealistic, ethereal 16k surrealistic, spectral, imperial colors, fantastic view, cinematic postprocessing, Surrealism, H.R. Giger, Giger Alien"
Negative Prompt (set to automatic)
"ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft"

I had used this prompt across platforms, sometimes tweeking it, and with additional moderations. Occasionally I mix it up between Deep Dream Text 2 Dream, DreamWombo and Nightcafe using respective results on those other platforms and sometimes switching back and forth using results as seed images. AI can create thousands of different iterations, so 'evolving' one result can give quite a few interesting additional results. For the most time I opt to create four images at medium resolution on Nightcafe, the odd time nine. The best ones I 'upscale' and also publish separately.

Here a few more examples created with Stable Diffusion v1.5 (the one I use most often):

▼ Below are a couple 'Schiele' creations where I used different seed images.
Sometimes you can make a silk purse out of a sows ear. This was the case with the next couple of images where I used results (unpublished) that didn't come up to snuff, but once 'evolved' turned out quite good.

▲ Before we forget that I do analog work - this is a drawing from some years ago (2008), called TRIPPING THROUGH BOGOMIL'S MIND
it is still floating around in Portugal someplace since it was exhibited there a few times. This was the seed image (hard to believe) for all of the above.

^{All these above and to the left of it were iterations and evolved from the same prompt - the next couple looking a bit different came about using the same prompt, but different input images.}

▲ EGON, EAT YOUR VEGETABLES!

◄ SIGNA TE SIGNA

^{same prompt, different seed images}

There are many more in my 'Schiele' Album THE SCHIELE COLLECTION ^{hopefully you can open that link}

Here a brief explanation how things work on Nightcafe

You have a choice of 5 Algorithm:
Stable Diffusion - DALL-E2 - COHERENT CLIP guided Diffusion - Artistic VOGAN+CLIP - Style Transfer
The first two are on the new advanced creation page, the others on the old style page. But with the VOGAN+CLIP on the old style page you can also create videos, but it costs quite a few point to do so.
It is a bit tricky to get to all the choices, but briefly you need to look for links to open those, like in Stable Diffusion for prompts you can open the Advanced Prompt Editor the simplest one are the Presets, then you have the Moderators and the one I go to all the time My Favorites (you can save any prompts from your, as well as any other creations that didn't hide them, in that file and reuse them.
On Settings look for All Settings to open more choices. This is important for a number of reasons: for one thing, you can now choose your output image size, the run time, the Aspect Ratio and many more, but one very important setting for Stable Diffusion would be the Model where you can choose between eight different models (I use mostly v1.5, but v2.1 gives you quite different looking images).
Most times under Prompts I leave the default Negative Prompt setting to On, but I had saved some negative prompts that had worked good for others in my Favorites.
You can also add second, third, and more different prompts and assign different weights to them.
Something else to tweak would be Clip Guidance and Sampling Method. Most times I leave them at default.

Final Words
I see great results from Midjourney, but also heard complaints of excessive censorship, plus, I am no fan of PAYING for anything, while on Nightcafe I create for free (yes, there is a Pro Plan), but I amass so many points (5 points to claim each day) plus 3 points up to 10 times a day for sharing creations plus achievement points (for following and being followed, as well as participating in chat groups), I have so many points on my account to burn, I never run out. Conversely, I am also on Deep Dream, a senior user with daily renewing 70 points to use, while DreamWombo is free for all (there is a pro feature as well, for a few extra perks).