Version 3 released

Base model changed to IllumiYume XL v2.0

It boasts a stable 1536*1536 resolution and character learning and prompt responsiveness that rivals or exceeds NoobAI-XL.

Version 3 is v-pred Model.

This is a merge model based on Noob e-pred.

illustration-style pattern and focuses on skin tones.

The merged content is as follows

1.IllumiYume XL v2.0(Version 3 Only)

2.NoobaHoshiv1.0（Noob e-pred1.1 + Nova Orange XL v3 + Raehoshi illust XL v3）

3.paruparu_illustrious_v4

4.My Lora(Trained by 360 images AI-generated)

(v1 is 200 images AI-generated)

Recommended settings

• Sampler: Euler a

• CFG: 5

• Steps: 24～30

Quality tags:

masterpiece, best quality

Negative prompt（ver1、ver2）:

bad quality, worst quality, watermark

※no problem without it

Negative prompt（ver3）:

bad quality,worst quality,worst detail,sketch,censor, simple background,transparent background, 3d, realistic, backlighting

3d, realistic, backlighting is your choice.