Zhou Xiaohuan - Zhuxian

The bottom model used for training is: v1-5-pruned.ckpt

The image above shows the weight opened by the Lora and the corresponding effect map

Parameter recommendations:

1. Lora weight recommendation: 0.6~1.0

2. clip: 2

3. VAE: vae-ft-mse-84000-ema

4. HD restoration and enlargement algorithm recommendation: R-ESRGAN 4x+ Anime6B; recommended redraw range: 0.2~0.4

5. Sampling method recommendation: DPM++2M Karras

6. Sampling steps: 20-30

7. When generating a full-body image, it's recommended to enable Adetailer plugin for repair and HD restoration to ensure normal facial features.

8. Welcome feedback!