Model Information:

Base model: IllustriosXL 2.0

Sampling Type: Epsilon

Sampling Method:

1. Euler A

2. Euler Dy (Recomended)

3. Euler DY CFG++

Schedule Type:

• Auto

• Normal

• Exponential

• React Cosinusodinal Dynsf

Resolution

• 1440x2560 (2K)

• 1080x1920

• Other default Stable Diffusion Resolution:

  • 1024 x 1024 (1:1 Square)

  • 1152 x 896 (9:7) & 896 x 1152 (7:9)

  • 1216 x 832 (19:13) & 832 x 1216 (13:19)

  • 1344 x 768 (7:4 Horizontal) & 768 x 1344 (4:7 Vertical)

  • 1536 x 640 (12:5 Horizontal) & 640 x 1536 (5:12 Vertical)

Future Work:

• Change base model to NetaYume Lumina 3.5

Q & A

Q: Why stop training on Illustrious?

A: Limitation model.

Q: Can generate NSFW?
A: You're Committed a sin, if you REALLY wanna do it, just do it.

Q: Is there v-prediction?

A: NO

Q: Newest dataset?
A: no, ++knowledge -> hard to maintain resolution

Changelog:

0.3 (Same as 0.3-TEST before)

• This model will be deprecated. Thanks for using my model. don't be sad, there better than our model >:)

0.3-TEST

0.22

0.21

0.2

0.15

0.1

License:

NoobAI
IllustriousXL

If you'd like to support my work, you can gimme coffee

https://paypal.me/IRedDragonICY

https://ko-fi.com/ireddragonicy

https://saweria.co/IRedDragonICY

(Actually, I don't need money, just need happiness)