>>16535
I2I might help if you can draw, at least at a basic level (I cannot), so you can put together a sketch and then turn it into a photo - if the sketch is clean enough it works reasonably well, so this you can pose the characters by drawing those poses.

>>16565
> I2I might help if you can draw, at least at a basic level (I cannot), so you can put together a sketch and then turn it into a photo - if the sketch is clean enough it works reasonably well, so this you can pose the characters by drawing those poses.

I could just trace over some JustSketchMe poses for that, but then, if I'm trying to match a particular character to that pose, I'm going to have to do some more LoRA character training. I've trained two characters so far, didn't get the level of consistency I'd want, and those LoRAs only work with Wan 2.2.

Qwen is the image editor I've used most, but (at least using Ostris AI Toolkit) there wasn't an specific training option for Qwen. I tried to train for SDXL (upon which Qwen is supposedly based) and it didn't woork.

I think Pony (which comes in several variants) looks promising, but I'm not sure how to train a LoRA for Pony yet.

Training on my own RTX 3090 setup is painfully slow, and running training on Runpod can get expensive if one multi-hour training session doesn't work and you have to try again (and again).

>>16424

Actualy that cake looks surprisingly good and acurate breast anatomy! LOL
but anyway prety good results

>>16318
love your works!
bump!

Love seeing those tits sliced open... Good luck! I'm eager to see more.

Fake is good. I like to kill and rape my anime waifu, then go on a romantic date with her the next day like nothing ever happened. If it were real, she’d just be gone and I’d end up in jail. Besides, photo-realistic guro is aesthetically inferior to "Fake" anime/2D/2.5D/3D styles in my opinion. You can get all the photo-realistic gore you want just by going to a butcher shop at the market anyway.

anything

What I really need are models that actually match real kids in terms of body shape and appearance. Most so-called child models out there are just clumsily adapted from adult ones, so they never look convincingly like children. Plus, once they do any complex movement, their proportions revert back to adult ones. I'm always having to fix it with all kinds of post-processing and repainting—it's such a pain. So I'm really hoping to find a model that can just handle this properly.

From my experience, regular non-porn models are good with child-like characters by default. It's the gore part where it stars breaking.


This is stock Z-Image Turbo.

... and this is what has been generated for original prompt:

> studio photography, girl, 6 years old, child, caucasian, full slim body
> girl stands straight, grin, doing a "V" sign.
> one-piece swimsuit, simple color, summer shoes
> elf, unnatural bright yellow hair color, pointy ears, fluffy cat tail


I've then painted over it just with two colors and had it refined with:

> studio photography, mauled body
> stump, cut-through body, gore, blood, torn white swimsuit, missing arm

>>17435
Try MiaoMiao RealSkin checkpoint with following prompt:

aged down,
undeveloped body, immature body, petite body, child,
loli,child body proportion,

Links here:
https://boards.guro.cx/ai/koko.php?res=13910

For the Horde!

>>17062
Fuck yeah... Loving this. Would love to see more tits getting sliced up nice and slow. Shame there's no blood, though.

>>17164

Big tit damage fetishist here. Been waiting a long time for AI to figure out how to do some boob damage. Seems I gotta wait longer. We've figure out pussy stabbing, decapitation, hanging, disembowelment.. fucking everything BUT beautiful big titties getting sliced and diced.

I tried doing it on my own, no luck. I suck at this

>>17062
it very nice! do more please

>>17178
It wasn't me, found it on 4chan. Someone asking Grok to do this. I wouldn't risk it, I don't need to be on a watchlist. lol.

Brains for dessert? The spoon is too big (or head too small) and they're goopy-er than I'd like, but it works.

https://files.catbox.moe/oopenr.mp4

Ground breast meat.

https://files.catbox.moe/z5iupb.mp4

I'm not usually a fan of cannibalism, but these are excellent works.

That fried womb was one of the hottest things I've ever seen. We definitely need more of that!

>>16611
Pretty bad, no inherent training on related concept.

Anima is the one that has the most potential. It understands most guro and nsfw concepts out of box, just need bit more training to add gore details. Have much better prompting control over IL/Pony/Noobs (tag+caption). You can migrate LoRA training dataset from SDXL to it very easily as well.

>>16613
Got it. Thanks!

It's not meant for guro. This is the best result I could get with Z-image. It doesn't understand wounds and most organs.

>>16708
> Sure it can do images photoreal.

I think you missed my point. There are some source images that Qwen won't do a good job of converting from non-realistic to realistic. But that doesn't mean you have to give up on using those sources. I found that a good way of getting around that problem is to use a two-step process of converting difficult sources into anime or line art style as an intermediate step.

> ...also don't publish realistic CP here, or this board will be busted.

So far it seems to be okay to post links to some iffy content so long as the images/videos are hosted elsewhere, but, thinking about that more, it has been guro with young subjects but nothing so overtly sexual, so I've deleted my original post and replaced it with a post that links to a redacted image instead.

>I think you missed my point. There are some source images that Qwen won't do a good job of converting from non-realistic to realistic. But that doesn't mean you have to give up on using those sources. I found that a good way of getting around that problem is to use a two-step process of converting difficult sources into anime or line art style as an intermediate step.

I agree with need of two step process. I've met this trouble too.
I was experimenting with enhancing of 3D renders, and found that qwen-image-edit badly photorealizes things which it considers already photoreal.
And we need to provide some "freedom" to the sampler, to allow it to convert to realistic style better.
I war trying to add noise after image-load, but finally come to "intermediate style".

Also I prefer to render the image as "colored pencil painting" on middle stage, not as "line art". This will transfer original colors into final stage, while could add more freedom to sampler. It also preserves face and body props better. Adding black liner could improve small details, like finger positioning.
And intermediate results could be nice too, and acceptable on some boards.

Image attached.
Or see https://boards.guro.cx/ai/koko.php?res=16160&page=0#q16481
(of course I've stopped on that stage for this board 😉)

Comment too long, view post No.16742 to see the full comment.

But how do you send a loli photo to Qwen without him saying he cant modified photo with cp on it?

>>16742
> Also I prefer to render the image as "colored pencil painting" on middle stage, not as "line art".

Hey, that works well. Thanks for the tip.

The colored pencil look seems so close the the kind of semi-realistic stuff that doesn't work (like Daz 3D output) that I'm kind of surprised, but pleasantly so.

I did find, however, that details like dripping cum tend to get lost going from color pencil to photorealistic, but if I dare add to my prompt a request to keep the cum or put it back, I end up getting no PR conversion at all.

>>16854
>But how do you send a loli photo to Qwen without him saying he cant modified photo with cp on it?

It's about locally deployed models, not online.
Because of that it just can't "say anything". Just obey as it can 😄

Also I won't recommend you to ask any online model to do such things.

>>16855
>I did find, however, that details like dripping cum tend to get lost going from color pencil to photorealistic, but if I dare add to my prompt a request to keep the cum or put it back, I end up getting no PR conversion at all.

Somehow, when you're trying to convert image to some drawn style (line, pencil, etc), it percepts things "as is", including body fluids, reproduces them perfectly.
But if there's NO additional requests like "make face more younger" - in this case it starts to rethink mentioned areas and remove some certain details, like body fluids.

Comment too long, view post No.16856 to see the full comment.

>>16167
What are you talking about! He said another video of this girl where we can see her face while hanging. It is a great video.

I haven't see the video of her face after she hangs

>>16171
That is why people are requesting it.

Can't you give it to Sora or Grok and ask to move camera bit up?

>>16166
why don't we have more of this video?