[Stable Diffusion] Prompt Sharing and Learning Thread

Mr-Fox

Well-Known Member
Jan 24, 2020
1,401
3,794
I understand exactly what you are talking about. Very apt analogy also. Using the prompt on the fly to steer the boat in the direction you want. I think it's a good thing to have an idea in your head what it is you are going for. Nothing wrong with exploring though. Exploring and testing without a specific goal is a good way to learn new things. But with that being said, when it's time to create something it's best to have a clear "plotted course".
 
  • Like
Reactions: Sepheyer

fr34ky

Active Member
Oct 29, 2017
812
2,170
I understand exactly what you are talking about. Very apt analogy also. Using the prompt on the fly to steer the boat in the direction you want. I think it's a good thing to have an idea in your head what it is you are going for. Nothing wrong with exploring though. Exploring and testing without a specific goal is a good way to learn new things. But with that being said, when it's time to create something it's best to have a clear "plotted course".
Yes, I was talking specifically about the negative prompt, which many times prevents the picture from going in the direction we want, I normally have some standard negative prompts, but when I see the picture doesn't go in the direction I'm looking for, I start from zero with the negative prompts.
 

xmodo99

Newbie
Feb 20, 2020
19
17
First attempt at SD since 1.0.

It kinda went off the rails once urine jello happened.

It's surprisingly hard to get characters to eat anything other than dick. Anyone else have luck with translucent solids?

Also, anyone have success with making the character seem very sad/broken?

All these images went through a bunch of inpainting, but latest prompts should still be embedded in images.

00014-643552423.png

00014-1983795286.png

You don't have permission to view the spoiler content. Log in or register now.
 

xmodo99

Newbie
Feb 20, 2020
19
17
Listen... it just didn't cross my mind.

Though I do think you've just ruined fruit jelly for me.
 
  • Haha
Reactions: Mr-Fox

Mr-Fox

Well-Known Member
Jan 24, 2020
1,401
3,794
Edited, prompt below.

Can't shake a feeling that's what Playboy used to print in its glory days, circa ~90s. Am I off, and I have false memories?
Yup and 00's, but a bit less curvy. This is hot though, I don't like skinny. Curvy is feminine, skinny is not.
PLayboy used to be smoothed out skin and a bit of diffuse glow to make everything slightly shiny.
You nailed the look.(y)
 

Sepheyer

Well-Known Member
Dec 21, 2020
1,531
3,618
Hmm, she looks familiar somehow... A certain Monica? :unsure:
The chick from "Friends". Courteney Cox. The name is both in the prompt and in the file name:

"00006-1543782640-Courteney_Cox._chubby_plussize._wide_hips_cameltoe.__anfas.__stewarde...png"
 
  • Like
Reactions: Mr-Fox

Mr-Fox

Well-Known Member
Jan 24, 2020
1,401
3,794
The chick from "Friends". Courteney Cox. The name is both in the prompt and in the file name:

"00006-1543782640-Courteney_Cox._chubby_plussize._wide_hips_cameltoe.__anfas.__stewarde...png"
Thought so. Real milf candy right there.. :love:
 
  • Like
Reactions: Sepheyer

fr34ky

Active Member
Oct 29, 2017
812
2,170
Hey guys, you might be interested in this, the guy from this channel is on fire with the updates.



I made some testings and this is one of the most useful things, you can get the 'spice' from a model and put it into another model.

Here are some simple tests I made, I extracted the LORA from the model Anything 4.5 which is a 'anime' style model that has all that sexy hentai stuff inside, and applied the LORA to the model SamDoesSexy, which is a beautiful model that creates insane more 'western' sexy style art.

In the video the guy says you have to apply the LORA and trigger words, in my case, this model doesn't have trigger words, but I know that it triggers certain details I talked about before, so that's the idea to use this model.

The concept I was looking for was to get the underboob efect on the picture.

Both pictures are the exact same prompt and seed, in the second one I applied the LORA.

PS: I was so 'focused' that I didn't realize that in the second picture she looks less like the original character, but trust me the face is the easiest thing, and you can calibrate the lora prompt using 0.9, etc.

04466-3628252346-a photograph of Sophie Turner, beautiful face, wearing a tight tanktop, ((und...png 04467-3628252346-a photograph of Sophie Turner, beautiful face, wearing a tight tanktop, ((und...png
 

CoyotteStudio

Newbie
Game Developer
Feb 10, 2023
16
220
The evolution of these models is impressive.

I was looking through old posts of mine on a social network, and I found a post from August, with horrible results taken from the Dall-E Mini. And I was ecstatic about it, totally satisfied and amazed.

I can't even imagine what AI will bring us in the coming months.
 
  • Yay, new update!
  • Like
Reactions: Mr-Fox and Elefy