Here's what the prompt was supposed to be, minus quotation marks: "very revealing male pinup photo, Harry Potter, 25 years old, gorgeous, detailed face, amazing, thighs!!! [out of frame, ugly, disfigured, poorly drawn hands, extra fingers, extra limbs, deformed, bad anatomy, missing limbs, signature, words, watermark, logo] under the whomping willow, extremely detailed digital painting, Pastel colors, intricate detailed and complex and coherent, soft lighting, midjourneyart style"

Seed 776665930, cfg 7, 45 steps  gave me the pic shown earlier. Needless to say, none of the Harrys that I got were that old -- half looked anywhere from 15-19 at the most -- and only 3 of the batch were shirtless, but otherwise dressed. Not much pinup, mostly portraits that were very nice but a couple also went awry. No big deal that the prompt was not spot on for all of them, this was just an experiment to see if the new model could do better. Naturally, I expected that when I cut the prompt down to "under the whomping willow, extremely detailed digital painting, Pastel colors, intricate detailed and complex and coherent, soft lighting, midjourneyart style" and used the exact same seed, cfg and number of steps, I'd get the exact same scenery as before, without that hexed Harry or any other figure. Instead, it gave me this:

...which is very nice and I can use it elsewhere but it's certainly not the same. Perhaps I should have completely rewritten the scene instead of just negative-prompting the damaged main figure out? Would that make a difference, I wonder?

Maybe the model is totally screwy. I also prompted a full body render of Snape, with care taken to avoid badly drawn hands, faces and other deformities -- and of course face correction was turned on. Instead it mangled poor Sevvie and gave me a surprise horse that looks to be wearing or carrying a tree. Completely off the rails!

Either that, or I've got Peeves and his merry band of poltergeists haunting my computer, pranking the heck out of it whenever I go to draw anything. Wouldn't surprise me in the least. ;-)

Ha ha that dragon in the pumkin is just perfect Bakapo!  My daughter sent me a photo of a carved cat pumkin she made that was cute.  For  next year she's getting an idea for a dragon carved pumkin, lol, thanks.  You may have started a new family tradition for halloween pumpkins...  ;)

Also have to have a black cat, so continued the adventure with a Dall-E 2 image via Night Cafe.  A better result than in Mid-J, even with --testp, so a thumbs up for more realistic stable diffusion animals in the Dall-E corner for now...  ;)

Give Piper a pet for me Shaun of the UK kingdom...  

the cat pumpkin is fabulous! I really like the eyes.  a dragon would look cool all lit up, too.  

Some Halloween backgrounds... use as you like, I kinda used the first one already lol

And some dragons...

Twiz has been a busy kid, happy halloween and a happy Draco'hain also, the dragon fall gathering of the clans.  To celebrate a good harvest of humans of course, how many knights in shiny armor have you devoured this season...  ;)

Hi rhol,

I gave wombo the prompt "cat piper" - and see what it produced!

Tricksy Treats!

Not sure how much use they are as stock since I haven't gotten round to upscaling yet but do as thou wilt and Happy Hallowed Weenies to all!

only because it's All Hallow's...

Wombo's idea of what a vampire bunny might look like...

I like the vampire bunny!

the Frankenstein monster by wombo 

Hi G, thanks for droping by, luv that vampire bunny! I'm going to try a few of those and maybe a few other vampire critters, lol.  Hope you post some more...  :)

Frankenstein's monster is cool, perfectly, properly scary. Good job! Vampire Bunny... totally adorable! Reminds me of a book my sister had -- not sure if school or library -- that I read when we were oh, so young. Can't remember if the title was the same as the character, but it was about some kids with a pet rabbit that got named Count Bunnicula because he was "odd".  He was, of course, a vampire. 

A Wombo vampire bunny with the Spectral effect...

And a vampire fox...  ;)

Wombo

i love me some foxeses.. heh...

yanno, Perry, what i found interesting was a while back you posted an image outta MJ to your gallery, and it wasn't that long after i had had a go in MJ myself to see what it was all about, blahblahblah.  anyway, the short of it is this:  one of the iterations i did on a fae must've contained a similar seed to your prompt somewhere along the line, as the outfit i got wasn't that incredibly dissimilar to yours.  when i saw your image, all i could think was 'things that make you go hm...'

prolly help if i showed ya, huh... keep in mind though this was cropped and then blended with other iterations, none of which i have the original output for any longer, but it hints at the outfit, in style and colour

and this is part of my first Wombo vampire bunny... just wasn't crazy about it, hence the second attempt i posted above... though, gotta say, i do love those frickin' teeth :)

Kitsune looking for Giana

@Giana - https://arc.tencent.com/en/ai-demos/faceRestoration

Use it to fix wonky faces... It's Len's find... and it's free.

G - I see what you mean, firefly fae in the forest.  Glad the twisting quantum ways of the universe let you find your pretty avatar, just don't let Twiz offer to do an eye makeover, lol...  ;)

When you get ideas like this you should post them right away, thats a smile any mother could luv eh...

rhol_figament posted at 11:24 PM Mon, 31 October 2022 - #4448108

G - I see what you mean, firefly fae in the forest.  Glad the twisting quantum ways of the universe let you find your pretty avatar, just don't let Twiz offer to do an eye makeover, lol...  ;)

1st one I had to borrow her eyes from one of my cuties, this one is from a stock photo lol

but i like wonky, Rick. look at the figure i've been predominantly mucking with this past year - trust me when i say Krystal is full of wonk, esp. in her mesh [she's all tris, can't accept any subdivision, and if you export her out as an OBJ and bring her back in to stick her own rig back into her, her mesh simply doesn't receive it well at all as there are some rotational and bending joint parametres that are bad in the original rig, at least in P10. but god do i love her wonk regardless... lol

the second one you did has shades of Christina Ricci in the face now... to my eyes at least

actually, like Christina and Mila Kunis had a love child.. snort

... and the first one a young Catherine Zeta Jones

Two fluffy tarantulas... they're only like 20 ft. tall

Kitsune carved in bronze...  ;)

AHH! Where's the LIKE button (except for the spiders!!)

Be rude not to :)

Yesterday was Samhain today is Dia de Los Muertos

Coupla things I found in the woods 

ZDNet has something on this topic regarding DALL-E:How to use DALL•E 2 to turn your wildest imaginations into tangible art

The article also point out the issue of art ownership.

Girl with a Pearl Earring

Love 'em all -- even the giant spiders. Here's a hint I got from facebook SD group I'm in last night: adding "adorable eyes" close to the start of your prompt gives very nice eyes and fewer facial train wrecks. Works with both humans and animals. Had great results in 5 out of every 6 renders I tested so you might wanna try it out.

Tried doing a vampire bunny of my own. I think SD must have been smoking something funny, 4 out of 6 were way more vamp than ebil wabbit.

And from my "the day after Halloween" prompt:

Those two were done using the midjourneyart style model for Stable Diffusion. Same prompt in regular SD 1.5 was rather meh, artistically speaking.

Do anyone know?  Just out of curiosity, was wondering why Deep Dream Generator has a spot of its own, while all other AI is lumped into just AI Generated?   Glad there is an AI category now.

Good question. I'd forgotten all about Deep Dream Generator and had to look it up. Turns out I had a free account there back maybe 5-6 years ago when it first came out. I recall not being very impressed with it at the time, since all it seemed to be was a filter that made a mess out of one's photos and Daz/Poser renders. It sure seems to have improved since then.

Quetzalcóatl

I guess some undead bunnies have been around for a while...

To get the old photo look you can add 'WW1 found photograph' at the end of your prompt... ;)

Or 'vintage found photograph' will also work.

A possible solution for Wombo users who have trouble with their figures getting their head and or feet clipped out of the image.  So far it works great for me...  ;)

You just need to upload an image of a figure in the pose you're looking for.  Full body, or just the upper torso for a portrait style, looking in any direction.  For the first one I used an image by Soffy, for a portrait image of a dark wraith.  Almost every reder had the final image in the same pose and direction.  A library of Victoria and Micheal poses in Poser or Daz Studio could be a fun reference.  Something to try at least... 

And for a full body figure I uploaded one of my own old daz renders...

Did a bunch of renders and the figure stayed in the middle 'almost' every time, lol.  

Thanks for the tip rhol_figament.   Will try it.   Boy, they are trashing us users of AI over on that other forum.  (even some that hang here...surprised me)

Which forum is that, deviant or the other D-place? 

love the WarBunny, Perry!! yumyumyum...

this was my third and last vampbunbun from last night, which i didn't like as much as number 2, which is why i posted number 2

i put together the prompt for the following back in early August when i was trying various AI sites for the first time. looking at it now, all i see are hints of Leonard.. heh

maybe Rick needs to give it some TwizEyes, too!! [chuckle]

Your vamp bunny is just fine!  Thanks for sharing....  ;)

He's not in a good mood, must be an anti ai ranter eh...  ;)

Digital Media

As for my own opinion on ai, it's just another tool to create images that I could never do in other media.  I've always been a poor drawer, even with lessons, my hands looked just like the wonky ai ones, lol.  AI is just digital media, all the great artist's in the world have not dissappeared, they are still drawing and painting and crafting.  Even if many of them have traded in their canvas board for a high tech wide screen tablet and pen eh.  People just need to chill, I guess nothing is supposed to be fun anymore...  ;(

Nah, I'm not anti or pro AI artwork... Why does my opinion matter so much, anyway? I don't if you remember rhol but I was the first one that told you about Wombo. We're here to have fun and enjoy each other's pics, not to point fingers or witch hunt. I want to be part of the solution, not part of the problem.

A lot of those AI threads got locked because of users going at each other, let's keep this one going and be excellent to each other ❤️

Be excellent to each other...

We're on a road to nowhere...

https://www.youtube.com/watch?v=LQiOA7euaYA