This model is based on RouWei.
Note: Please read the RouWei page carefully. The usage is different from the conventional Illustrious model.
You can use RouWei's unique image tags, such as flat colors and low brightness.
※Fingers are not stable
※The recommended resolution is 1240*1240. Higher resolutions such as 1536*1536 are not supported.
content is as follows
Lora(Trained by 360 images AI-generated)
Recommended settings
Sampler: Euler a
CFG: 5（v-pred model is 4）
Steps: 24～30
Quality tags:
masterpiece, best quality
※The quality tag must be at the beginning
Negative prompt(e-pred):
bad quality, worst quality, 3d,realistic, backlighting,
Negative prompt(v-pred):
bad quality, worst quality, 3d,realistic, backlighting, bad hands, mutated hands
※3d, realistic, backlighting are essential.
使用しないと、画像がサンプルに近づきません。