Home Artists Posts Import Register

Downloads

Content

This is the best chess I could play. #️⃣

See the tutorial further below for improving results⏬

This was one of my most difficult jobs, but I think I can still improve this LoRA.️‍

🖼️ See some examples 🖼️

LoRA prompt weight: 0.9~1
Model used for training: Uber Realistic Porn Merge 1.3.

Recomendations:

  • CFG ~9
  • Steps >=35
  • DPM family

ACTIVATION PROMPT:  jkfishnetnobra

Prompts that can improve your results: fishnet, nipples, pointy_nipples, large_breasts, huge_breasts, medium_breasts, sagging_breasts, large_fishnets, small_fishnets, pale_nipples, dark_nipples...

🌟Tutorial for improving results:

1. Do your prompt as normal, when you find a result almost like the one you want, but still imperfect, go to the next step. If the result is almost perfect go to step 3.

2. Recycle the seed and use the extra seed option. I recommend using strengths between 0.01 and 0.1. Find the generation that suits you best. Don't forget that for the next step you need to recycle the extra seed as well.

3. Use the hires fix to increase the resolution and enlarge the details, I recommend leaving the denoise between 0.5 and 0.6 and use the Latent (bicubic antialiased) upscaler.

4. To make it even better, you can still use inpaint in the nipples area, always retouching both at the same time and using between 0.3 - 0.5 denoise and DDIM sampler, with the same CFG you used for the original prompt. Also use the "nipples" prompt with 1.2~1.3 weight, if they don't come up. To speed up the process use "only masked" in inpaint area option and set a square resolution, such as 512x512 or 768x768.

Tendencies of this LoRA (and resolution):

  • Use "large_fishnets" to improve spaces on fishnets.
  • This Lora also fatten people: use "plump", "fat" on neg with ~1.3 weight.
  • To make nipples paler, use "ebony" prompt on neg.

Inside zip I'll put example images and tags used for training. You can try using it on your prompts.

Comments

No comments found for this post.