☀ Yan Paper Art _ Hunyuan

Model Introduction:The Lora model in the Yan paper art style can be considered a combination of Yan paper art and paper cutting art, featuring unique Chinese elements. This model is trained based on the Tencent Mix-DiT architecture and supports Chinese and English prompts.

Parameter Suggestions:

For online rendering (Web UI), please refer to:

VAE: automatic matching

Resolution: 768*1024, 768*1152

Weight: 0.6-0.8

CFG: 4.5-7

Sampler suggestion: DPM++ 2M Karras, steps: ≥30 steps, official recommendation is 50 steps

For online rendering (Comfy UI), you can refer to my workflow. Currently, some plugins may be missing in liblibai's Comfy UI.

Resolution: 768*1024, 768*1152

Weight: 0.8

Base film recommendation: HunyuanDiT-v1.2

For prompts, you can refer to my suggestions or use your own imagination.

For Chinese description triggers, use:Yan paper art, and for English description triggers, use: paper art