Wanxiang_Multiple models (independent)

This model is different from the past, so if you use it, please read the following carefully:

This model compares sensitivity values of Lightning Large-scale Model (text emphasis/CFG Scale), the higher the value, the higher the saturation. Based on official recommendations and personal testing, the expression of CFG value "2" stands out the most realistically.

Sampling steps are recommended to be between 10-30, through multiple vocabulary tests the best performing value is "25"

This model has been repaired and integrated with VAE, maintaining the VAE state as (None) during operation and execution.

The zoom algorithm can use 4x-UltraSharp (High score iteration steps "15")

The recommended sampling method is DPM++2S a

This model is a multi-class large model that can generate various styles, so you can innovate boldly.