Suggestion weight: 0~1.5
Tag:@xi410v3
This version adopts NaViT native-resolution training. To achieve the best results with this version, it is recommended to generate at high resolution, such as 1080x1920 or 1664x2342 (This is also the resolution of the LoRA training set)

Use https://github.com/WalkingMeatAxolotl/AnimaLoraStudio Training