I was wondering if the mc is going to have any bad ends revolving around Ophelia. Sorry if I'm asking to much but are planning on updating the pictures for the mc being futa or keeping the way that it is, because I'm fine with either I'm just wondering that's all. Also, it's fine, take your time the more you polish the better it becomes.
Not around Ophelia, no, but I keep bouncing back and forth on ideas for the massage boi who's been corrupting her. I don't currently plan on changing images to accurately reflect a futa MC, but that may change if all this AI stuff works out as I hope it will.
I can only think of one off the top of my head, but it was just an example of another kind of girl whose anus is only there for sex rather than biological functions.
Well unless she can only use her power for mass destruction, and not at all for the kinds of things that all the other demons we know do, it would still be better to have it and not need it than to need it and not have it.
Oh, and those AI images are surprisingly good, better than what I usually see in AI images. I tried installing StableDiffusion myself a little while ago, but I could never get it to produce anything even slightly recognizable for some reason.
I dunno how good Abba would be at restraining herself if she were ever sicc'd on someone or something
Yeah image generation AI has come a long way. Luckily there's an all-in-one installer now that can self-update otherwise I'd never manage to wrap my head around it.
While I can't totally speak for slacker, I think the main reasons these images look so good are:
1. He's on more recent version of Stable Diffusion than the one you might be using.
2. He uses a model called Don Corneo's Seventh Heaven as his base, which is really high quality. Although I've found meinamix and meinahentai to be really good too for generating 2d porn.
3. He probably has a VAE installed, which helps a lot with color and contrast. I've never noticed a difference switching between them tho, so you could probably just grab one and forget about it.
4. He's using LORAs, which basically nudge the images in a specific direction. Loras that nudge the final image towards certain poses, towards consistent characters, and that nudge towards specific artstyles. There's a lot of knobs you can turn:
5. He also has ControlNet installed, tho idk which ones he's used it on. Basically you give controlnet an image and you can tell it to interpret the image's composition for generation. It's really powerful, especially for getting specific poses:
6. I believe he also has textual inversions, like EasyNegative, which reduce the chances of things like bad anatomy a bit.
7. Finally, you can specify you want a higher resolution picture with HiRes fix, which causes a massive jump in quality at the cost of about 4-5x the generation time. Faces especially benefit from this, but overall the detail jump is crazy:
While it's a lot of steps, once you know what you're looking for it's kinda just a lot of download times:
Grab the
You must be registered to see the links
, it's supposed to be all-in-one.
Grab a model from civitAI, here are
You must be registered to see the links
,
You must be registered to see the links
, and
You must be registered to see the links
Grab a vae, usually the model will specify the vae that was meant to be used but I just use
You must be registered to see the links
for everything because I'm lazy.
Grab textual inversions (I use
You must be registered to see the links
and
You must be registered to see the links
)
And then from there if you want more control:
Install the LORA plugin via the a111 web ui.
Grab LORAs from CivitAI or wherever, slacker's posted a ton of HU themed ones in the discord.
Install the ControlNet plugin via the a111 web ui.
Grab the ControlNet models
You must be registered to see the links
. I only really use openpose, anime lineart, and depth, but they all have their uses and you can see examples
You must be registered to see the links
.
And bam you should have everything you need, although it broke a few times for me and I had to do a lot of fixing and tweaking, but no tech every works for me first time. It is still limited: doing more than one person in a scene is really hard, decensoring images is for some reason near impossible for me, and some loras are really hard to work with. That's why I've spent 22 hours now on my Oreteki Lora alone so that it's as high a quality as possible.