Renderosity Forums / AI Generated Art
Welcome to the AI Generated Art Forum
Forum Moderators: wheatpenny, RedPhantom, Deenamic
AI Generated Art F.A.Q (Last Updated: 2024 Aug 27 11:07 am)
Dream, by Wombo https://dream.ai/create
Subject: Figament's AI Corner
There's just something about eyes no matter what color they are, some just pull me in...
Peekaboo
Base prompt for first: "Sherlock with John Watson, highly detailed skin, dramatic action pose, volumetric light, high quality, digital painting, artstation, smooth, sharp focus"
Modified to: "Sherlock with John Watson in the Tardis, highly detailed skin, dramatic action pose, volumetric light, high quality, digital painting, artstation, smooth, sharp focus 8k resolution holographic astral cosmic illustration mixed media by Pablo Amaringo" for the second
Negative prompt used on both: out of frame, off camera, ugly, deformed, poorly drawn hands, extra fingers, missing fingers, missing thumb, missing limbs, extra limbs, NSFW, nudity, doubles, text, signature, words, watermark, logo
And NightCafe's version of SD merrily ignored most of it. Can't figure what I'm doing wrong but I've never had much luck there. :-(
I used both the beta and the SDXL 0.9. I too used up my 10 free's on the 1.0 but the .09 and beta both gave my diversity images.@Steve -- those are excellent! Which SD version are you using on NightCafe and is it free or Pro? Because all I ever get from there is extremely ugly more often than not. Just ran a couple sets now and just 2 pics out of 32 were worth bothering with, even though not exactly as prompted -- just had one character not both. Here's what I mean by ugly:
If I had tried this in my local install, I'd probably have gotten lookalikes but they would at least be recognizable. I used the SD 1.5 model and my typical negative prompt to ward off disaster: "ugly, deformed, poorly drawn hands, extra fingers, missing fingers, missing thumb, missing limbs, extra limbs, twins, clones, doubles, text, signature, words, watermark, logo" but as you can easily observe, I may as well not have bothered.
Yes, I know they also have the SDXL model which gives better results with less fuss. But after 10 (I think, might be less) uses of that you have to pay. With real money, not credits, because they have decided it's "too expensive" and made it a Pro-users only feature.
One of the "wrong" but usable NightCafe renders:
Kenmac posted at 4:45 PM Sun, 30 July 2023 - #4471506Thanks very much. I appreciate it. :o)Those diversity images that have been posted made me go down the rabbit hole earlier this afternoon. The first one i did was done in Kandinsky using the pencil setting:
Nice soft graphite style...great looking creation...!!
I will eventually do some PantherMen but I don't care to see them wearing as much clothing (jewellery OK) as the ladies do in these videos so they may get age restricted which is why I have held off a bit.🤣
Biodigital Jazz, Man...V5.2
Daft Punk - Around the World / Harder Better Faster Stronger (Official Live Video 2007)
Nice sword work DG, reminds me a bit of Cloud's Buster Sword from Final Fantasy VII... 😎
Yeah your half way there Cloud's Buster Sword from Final Fantasy VII and a medieval scottish claymore....
Nice city light work and views...! 
This turned out to be a versatile prompt. I like how you made her swedish. Very nice.Mage ai
swedish woman, early 20s, long blond hair, freckles, blue eyes, wearing a paisley bandana, relaxing in a blossoming orchard, wildflower bouquet around her wrist, close up, colour photo
Here are some new semi-random prompts from the perchance website:
somber weather mage calling forth a gentle rain in a sun-scorched town
knowledgeable tinkerer crafting a mysterious potion in a secret alchemy den
greek goddess, draping white robes, silver laurel crown, in an ancient temple, cypress trees, setting sun
talented street artist adding color on a stark urban cityscape, character portrait
fae, iridescent wings, long wavy hair, in a fairy ring, forest glade, glow worms around
restless ghost wandering the rooms of a derelict mansion entwined with vines, character portrait
Have fun! :o)
Very nice Rhol. It's an adaptable prompt. BTW I like your dust bunnies as well. :o)A_Sunbeam - A few tweaks of your Swedish lass prompt...
swedish woman, early 20s, long blond hair, freckles, blue eyes, wearing a paisley bandana, relaxing in a blossoming orchard, wildflower bouquet around her wrist, close up, colour photo --ar 3:4 --s 899
Starting to get a backlog again...so time to thin the herd...couple of various sets... instead of sets of each...
Privacy Notice
This site uses cookies to deliver the best experience. Our own cookies make user accounts and other features possible. Third-party cookies are used to display relevant ads and to analyze how Renderosity is used. By using our site, you acknowledge that you have read and understood our Terms of Service, including our Cookie Policy and our Privacy Policy.
@Steve -- those are excellent! Which SD version are you using on NightCafe and is it free or Pro? Because all I ever get from there is extremely ugly more often than not. Just ran a couple sets now and just 2 pics out of 32 were worth bothering with, even though not exactly as prompted -- just had one character not both. Here's what I mean by ugly:
If I had tried this in my local install, I'd probably have gotten lookalikes but they would at least be recognizable. I used the SD 1.5 model and my typical negative prompt to ward off disaster: "ugly, deformed, poorly drawn hands, extra fingers, missing fingers, missing thumb, missing limbs, extra limbs, twins, clones, doubles, text, signature, words, watermark, logo" but as you can easily observe, I may as well not have bothered.
Yes, I know they also have the SDXL model which gives better results with less fuss. But after 10 (I think, might be less) uses of that you have to pay. With real money, not credits, because they have decided it's "too expensive" and made it a Pro-users only feature.
One of the "wrong" but usable NightCafe renders: