Recommended Prompts
score_9,score_8_up,score_7_up
Recommended Negative Prompts
score_6,score_5,score_4
Recommended Parameters
samplers | Euler a | |
steps | 20 | |
cfg | 5-7 | |
clip skip | 2 | |
vae | xlVAEC_f1.safetensors | |
resolution | 1024×1344, 1024×1280 |
Recommended Hires (High-Resolution) Parameters
upscaler | 2x-UniScale_CartoonRestore-lite | |
upscale | 1.5 | |
denoising strength | 0.4 |
Tips
- Almost all models often require usage of ADetailer or Hires fix.
- Try mixing two Kakkoii LORAs for better results.
- GemuHentai model often requires adjustment of strength.
- Use ShonkaAnime, ShonkaSemi, or ShonkaReal to force specific styles in SHONKA MODEL.
Creator Sponsors
All sponsors are not affiliates of Diffus. Diffus provides an alternative online Stable Diffusion WebUI experience.
NEW DISCORD SERVER (work in progress) – https://discord.gg/bPRpQQRc4Y
Almost all models often requires useage of ADetailer or Hires fix!
If you like my work you can tip me on – patreon.com/Natakaro or buymeacoffee.com/natakaro 🙂
KakkoiiPony2 model is more trained version, little bit more style but sometimes need adjustments, you can also try mixing two Kakkoii LORAs.
GemuHentai model is harder to use – often requires adjustment of strength etc!!!
SHONKA MODEL integrates 3 styles in itself – you can use ShonkaAnime, ShonkaSemi, ShonkaReal to force style or use default similar to semi style.
All LORAs are trained in such a way that they can be mixed in reduced proportion, they do not require special parameters or way to generate.