[Stable Diffusion] Prompt Sharing and Learning Thread

[Cerealkiller23]

I was just wondering if someone could explain something to me. I'm starting out with AI generation using stable diffusion and I'm not sure how I would generate a consistent character ( planning to make a game with it).

How exactly did the developer of this game

You must be registered to see the links

, create a consistent character with different clothes, poses etc. ?

 

[Mr-Fox]

I was just wondering if someone could explain something to me. I'm starting out with AI generation using stable diffusion and I'm not sure how I would generate a consistent character ( planning to make a game with it).

How exactly did the developer of this game

You must be registered to see the links

, create a consistent character with different clothes, poses etc. ?

controlnet and/or img2img

 

[me3]

In this case the IPAdapter is your tool. You show it what dress you want, then inject it's output into the sampler. It is a "more better" approach then using prompt weight's. In fact it becomes a micromodel that must improve consistency.

If you need a snippet of the workflow let me know.

I did some tests with IPadapter, but i've tested so many different methods that i'm mixing up what i tried with what. I don't think i did it specifically with just the clothing though, that might work better.

Here's an upscale of an image i got for something else. It's using IPadap, with a closeup of the person and "far off" shot of a town/settlement with with foggy/cloudy surrounding as "seed images". SD15 without any detailer fix.

You should check out this video here for an alternative method for video generation that is perhaps better at temporal coherence:

You must be registered to see the links

Unfortunately the controlnet seems to be made for animatediff which works rather poorly with low VRAM and it's very limited in frame count and rate. Really wish someone would figure out a "non batching" method for it. If batching makes it possible to keep this "alike", it has to somehow be referencing each other or the same image/noise/whatever. So in theory you should be able to do much the same with looping. We already have ways to use images or parts of images as reference points, in a sense we have 100% recreateability for a single image too, so there has to be some way to combine the various "parts"...anyway
I'll take a look at the controlnet model, it might have some impact even without animatediff, thanks for pointing it out.

 

[Cerealkiller23]

controlnet and/or img2img

how exactly should I use img2img?

Can I use an image of a character as a reference then add prompts and LORA's to change clothes, poses etc. ?

 

[Mr-Fox]

how exactly should I use img2img?

Can I use an image of a character as a reference then add prompts and LORA's to change clothes, poses etc. ?

I haven't experimented with this scenario so I can only give general tips. I suspect that you would either create a character in txt2img or use a ready made reference image. Then in img2img use controlnet with openpose for the poses and something like ip adapter for the clothes. The thing is you need to give more information in order for anyone to help you better because the workflow will depend what is your starting point and what is your goal in more specifics and details.
- Do you have a character already?
- Are you going to work in A1111 or ComfyUi?
Most importantly. Are you going to make something like the link you gave? Then count me out. I don't do lolly type stuff or anything other questionable content that looks even vaguely like CP

 

[Cerealkiller23]

I haven't experimented with this scenario so I can only give general tips. I suspect that you would either create a character in txt2img or use a ready made reference image. Then in img2img use controlnet with openpose for the poses and something like ip adapter for the clothes. The thing is you need to give more information in order for anyone to help you better because the workflow will depend what is your starting point and what is your goal in more specifics and details.
- Do you have a character already?
- Are you going to work in A1111 or ComfyUi?
Most importantly. Are you going to make something like the link you gave? Then count me out. I don't do lolly type stuff or anything other questionable content that looks even vaguely like CP

Thank you for replying, to answer your last question, not at all, I share your disgust for people who like that, it was just the first thing I saw on dlsite that was similar to what I want to make.

I'm working on A1111, as for the character, I currently have nothing particular in mind regarding the character, I just started researching AI art yesterday and have no concrete plan yet.

My plan is to make a ntr game in RPGmaker that I'd sell on the japanese market, I've been seeing alot of these ai art games sell pretty quickly making a nice profit so I thought I'd try it out as a side hustle.

I haven't heard of controlnet and openpose so I'll have to read up on that. Thank you for answering

 

[Sepheyer]

Thank you for replying, to answer your last question, not at all, I share your disgust for people who like that, it was just the first thing I saw on dlsite that was similar to what I want to make.

I'm working on A1111, as for the character, I currently have nothing particular in mind regarding the character, I just started researching AI art yesterday and have no concrete plan yet.

My plan is to make a ntr game in RPGmaker that I'd sell on the japanese market, I've been seeing alot of these ai art games sell pretty quickly making a nice profit so I thought I'd try it out as a side hustle.

I haven't heard of controlnet and openpose so I'll have to read up on that. Thank you for answering

There is a hope to make consistent images where you first model them in DAZ / HS2 / Other and then run image-2-image with AI to add gloss as well as good old photoshop. I think the consensus so far is that no AI-only method can give you true consistency, although some combinations of methods can get you reasonably close.

Here is a start of one such discussion about "converting" DAZ girls into any style you want: https://f95zone.to/threads/stable-diffusion-prompt-sharing-and-learning-thread.146036/post-12343859

 

[Cerealkiller23]

There is a hope to make consistent images where you first model them in DAZ / HS2 / Other and then run image-2-image with AI to add gloss as well as good old photoshop. I think the consensus so far is that no AI-only method can give you true consistency, although some combinations of methods can get you reasonably close.

Here is a start of one such discussion about "converting" DAZ girls into any style you want: https://f95zone.to/threads/stable-diffusion-prompt-sharing-and-learning-thread.146036/post-12343859

I've never used those two so I'll try to find some alternative solution for now, thank you anyways

 

[me3]

Having been posting a lot of prompt/workflow-less images/clips lately i thought i'd do a cleaned up one to maybe give ppl ideas on how to do something. While this is from comfyui the image itself should make it possible to see what would be needed to set it up in a1111, i hope.
This is basically the same setup as i used to create the image in this post, and it can be used to layer or create "depth" in images. It obviously depends on what image you put to each layer and how they are handled/applied. I used a bunch of math and composite nodes to create the colored masking image, this can obviously be done other ways and in other shapes etc. IE if you put each "color/mask" next to each left to right, you could use it for panoramic views or multiple characters, a bit like regional prompts but with "image inputs".

This is not a very tidy/nice flow, i tried to keep things compact for the "screenshot" while still readable. Ipadapter nodes are colored by mask channel for visual aid. I've never installed nodes through anything else than the manager so any missing ones should be found there. I've included the mask image if anyone should want that.



Edit: Adding an image using the workflow, just dropping 3 image i had into it.

 

[me3]

Before stable diffusion - I always wanted my 3D Models to have a painted / artistic aesthetic to them. As if hand drawn.
I would of expected a solution to this be available through custom shaders but I was always disappointed with the results.

AI + Rotoscoping feels like the more likely technology to get there. Imagine extremely detailed hand draw art, animated and rendered at the speed of 3D - if that can be achieved it's almost the best of both worlds.

--

Before SD this was impossible.

Animation in old Disney movies always had extremely detailed backgrounds then simple flat shaded characters / animation because they had to draw them frame-by-frame. If a single frame takes an artist 3 weeks, you can't possibly achieve 24 frames per second and the likelihood of consistency falls dramatically as well.

This would be something AI could do (hopefully) that is essentially impossible today.

You mentioning hand drawn stuff reminded me of some images i got quite a while ago. I was running prompts of descriptions of fictional places, and for some reason this is the "style" and look the AI decided to give when using various descriptions of Rivendell. While maybe not hand drawn it reminds me of old book illustrations and that kind of thing.

 

[me3]

Considering the movement is mostly down to the AIs choice, you think there's one seed that has both halves moving?
View attachment SVD_00017.mp4

 

[Sepheyer]

You mentioning hand drawn stuff reminded me of some images i got quite a while ago. I was running prompts of descriptions of fictional places, and for some reason this is the "style" and look the AI decided to give when using various descriptions of Rivendell. While maybe not hand drawn it reminds me of old book illustrations and that kind of thing.

View attachment 3168298 View attachment 3168299

I love hand drawn stuff too.

These are Doomer Boomer checkpoint which is hot AF:

You must be registered to see the links

And I think here is what the checkpoint was actually meant for:

 

[Sepheyer]

Before stable diffusion - I always wanted my 3D Models to have a painted / artistic aesthetic to them. As if hand drawn.
I would of expected a solution to this be available through custom shaders but I was always disappointed with the results.

AI + Rotoscoping feels like the more likely technology to get there. Imagine extremely detailed hand draw art, animated and rendered at the speed of 3D - if that can be achieved it's almost the best of both worlds.

--

Before SD this was impossible.

Animation in old Disney movies always had extremely detailed backgrounds then simple flat shaded characters / animation because they had to draw them frame-by-frame. If a single frame takes an artist 3 weeks, you can't possibly achieve 24 frames per second and the likelihood of consistency falls dramatically as well.

This would be something AI could do (hopefully) that is essentially impossible today.

I remember the negative shock I had when "they" continued Tank Police anime in a 3D-ish format instead of a traditional anime look. I went FML my life is over despite still being in pre-school.

Indeed, hoping for a rebirth of the esthetics since hand drawn look starts looking really easy to knock out:

 

[Sepheyer]

- Can you do a romantic pose?
- Sure.
- Is that... Luke Skywalker on Tatooine?
- That's right!



Control-net'ing scenes from movies is fun. Here the very same scene but with an elf:

 

[me3]

This took me about 40min to generate. With a setting of 16 frames, i wanted to see what happened if i increased that too 24...
View attachment SVD_00018.mp4
That took 33421 seconds...and tbh i can't really say it's worth the almost 9 hours extra...It took nearly 11gb vram (i still have only 6gb), overflowing has it's uses, but at a cost....

 

[Cerealkiller23]

Can someone explain prompts to me?



Why can't I get this to look how I want to?
I want there to be a hand grabbing her ass while she's looking at it embarrassed. I don't understand why she's half naked when I specified against it. Here are my prompts.

<lora:uraraka-10:1.15> <lora:grabbing_anothers_ass_v.2.3:1.15> grabbing another's ass, anime (style:1.2), (anime:1.3)

male hand grabbing woman's ass
the woman is wearing a (white_shirt:1.1), (pantyhose_over_panties:1.3), black_panties, dark blue skirt, brown pantythose, (fully_clothed:1.5)
the woman has brown eyes, brown hair, lips, medium breasts, fit ass, (clothes:1.2),
the woman looks surprised, embarrased, (looking_at_hand:1.2)
office background, desks, computers

Negatives: (badhandv4:1.2) (easynegative:1.2) , visible man, (nude:1.2), bad anatomy, naked_ass, naked, naked_butt

Can someone please explain this to me, Im going insane lmao

 

[Cerealkiller23]

Can someone explain prompts to me?

View attachment 3171260

Why can't I get this to look how I want to?
I want there to be a hand grabbing her ass while she's looking at it embarrassed. I don't understand why she's half naked when I specified against it. Here are my prompts.

<lora:uraraka-10:1.15> <lora:grabbing_anothers_ass_v.2.3:1.15> grabbing another's ass, anime (style:1.2), (anime:1.3)

male hand grabbing woman's ass
the woman is wearing a (white_shirt:1.1), (pantyhose_over_panties:1.3), black_panties, dark blue skirt, brown pantythose, (fully_clothed:1.5)
the woman has brown eyes, brown hair, lips, medium breasts, fit ass, (clothes:1.2),
the woman looks surprised, embarrased, (looking_at_hand:1.2)
office background, desks, computers

Negatives: (badhandv4:1.2) (easynegative:1.2) , visible man, (nude:1.2), bad anatomy, naked_ass, naked, naked_butt

Can someone please explain this to me, Im going insane lmao

Is it because the LORA's are clashing with each other?

 

[namhoang909]

This is my sd ui

this is ui of a guy on Youtube, his video was recorded 7 months ago, why don't I have those symbol under generate button? I really need that save prompt button

 

[Mr-Fox]

Is it because the LORA's are clashing with each other?

Stable diffusion can only do what the models have been trained to do. For something this specific, either you need to train your own lora or you need to find a lora someone else has trained. This is the point of sites like civitai or tensor.art . You can always experiment with controlnet and see what you can get from it. Maybe you can faint or fake it this way. Perhaps an approximation will be close enough and you can do the rest with dialoge.

 

[Synalon]

This is my sd ui
View attachment 3171593
this is ui of a guy on Youtube, his video was recorded 7 months ago, why don't I have those symbol under generate button? I really need that save prompt button
View attachment 3171600

Use this button here to save prompts/styles