- Mar 23, 2019
- 36
- 458
These are great. And they look more like OG Le Stag Tara than my "simplified" Tara I always use.Nothing sexual. Just Tara at the gym.
View attachment 4229264 View attachment 4229268 View attachment 4229270
Watersports was the furthest I've gone in pics I've made, and yeah I recognize that isn't even all that "extreme". I think dr.maker made some feminization stuff with MC which isn't for me but lots of folks like it. I'm only a prude on the stuff I produce myself. I'm glad to give anybody the tools to make whatever they're into with the AI.Not offended yet huh? ... Tbf, I am kinda surprised we didn't have someone opening up the agreement between Laura and the MC to include some dogs or a horse... I think my watersports request was the "hardest" we got so far no? xD
Honestly all I did was sign up for a free account at that klingai site in the watermark and feed it a Laura image.Tutorial coming soon?
I would like to experiment more, but the Buzz coins deplete fast, especially when the AI decides to make weird distorted images rather than the strict prompt. I am hoarding buzz for now.
Yeah there's really a lot of ways to answer this one. For Laura and Tara, we achieve a consistent look by describing their look to the AI. Eye color, hair color and style, "cute", stuff like that. And then apply style LoRAs which affect the whole image to make them "disney princess"-ish and anime-styled. That's what you can do in Civitai. But the txt2img AI has a short attention span so by the time you describe all the traits of both girls, they get mixed up sometimes. So maybe Tara ends up with a ponytail instead of Laura, etc.nickohoma, could you explain something, please? Say, you have a sample image of Laura and a sample image of Tara. How does one put them together into a new image, so that they are close to their samples and not some other athletic girl with long, pink, curly hair? Is it just about using the same resources and similar prompts?
Many thanks!Yeah there's really a lot of ways to answer this one. For Laura and Tara, we achieve a consistent look by describing their look to the AI. Eye color, hair color and style, "cute", stuff like that. And then apply style LoRAs which affect the whole image to make them "disney princess"-ish and anime-styled. That's what you can do in Civitai. But the txt2img AI has a short attention span so by the time you describe all the traits of both girls, they get mixed up sometimes. So maybe Tara ends up with a ponytail instead of Laura, etc.
This doesn't work for Tara and Laura, but if you have specific characters the AI recognizes, and you keep your prompt simple, you can sometimes get Civitai to build an image with multiple characters. "Tracer kissing Princess Peach" works.
View attachment 4229623
Everything from here down involves running a local Stable Diffusion instance like A1111 or ComfyUI.
My go-to is Regional Prompts in A1111. I tell the AI to put Laura in region 1 and Tara in region 2, and define the regions.
If you have a starter image you want to incorporate, you can do an img2img workflow. You can create or manipulate an image in Gimp or MSPaint or whatever and then feed it to the AI along with a prompt. So if you have the perfect profiles of Laura and Tara, you can put them in an image, then have the AI "build around" them.
Then there are IPAdapters which are basically sorcery. Only available in ComfyUI AFAIK, you could feed the AI images of characters and have it generate new images using those characters in completely different positions. At least in YouTube tutorials. I am not smart enough to get them to work.
Ouh, was the 0.35 of Moonripple something like that? Haven't played that yet, but guess I need to! :xWhen I was playing with the AI CG I did some watersport... But is a bit explicit
Also I could try later doing something like in the VN Moonripple Lake's Halloween Special: Laura with a Minotaur xD. Which when I think about it, the original Laura has not even done a creampie, yet here we are thinking about watersport, minotaurs, femboys, futas... Laura from this universe is insane and wild xD
Yes, dont worry. Im just too eager, like a kid with a new toy . I will stick to more basic images for now, until I feel more learned with AI CG.Honestly all I did was sign up for a free account at that klingai site in the watermark and feed it a Laura image.
It looks like there may be a ComfyUI module called CogVideo to do similar stuff. Which might be enough to get me back into using Comfy.
Your "basic" images are great. Are you posting your stuff on Civitai? You can get a little bit of buzz that way as people give you thumbs-up. And I've been known to tip newcomers occasionally .Yes, dont worry. Im just too eager, like a kid with a new toy . I will stick to more basic images for now, until I feel more learned with AI CG.
Very great pics, man! I could definitely see these as movie posters. I can't wait until AI can finally figure out how to render guns properly, though, hahaI'm a cinephile, apart from a very horny man xD, so I made a few homage images of the filmYou must be registered to see the links. Sadly the AI really struggled with the rifle model:
View attachment 4230798
View attachment 4230801
View attachment 4230802
View attachment 4230803
BTW, thanks a lot nickohoma for the buzz
AI truly mimics humanity... we just can't be not horny... xDThose are super impressive dude. Great stuff.
I also love that one of the shots is basically bj-pov. Minus any nudity/sex of course. Like the AI couldn't help itself and had to throw that in.
It's not fun for me to do these arts anymore because I can't find creative and different things. But whenever I enter this site, it is very nice to see that the previous posts are liked and taken as an example. nickohoma will help you to correct the flaws of both beginners and people who find flaws in their art after they start. Have fun again to all of you.First try based on ichi3000001's early stuff. And, ofc, special thanks to nickohoma's for their tutorial!
99 percent luck that she's actually holding the weights correctlyThese are great. And they look more like OG Le Stag Tara than my "simplified" Tara I always use.
Also impressed she's actually holding the weights. I've had trouble with that.