Wanxiang_Multiple models (independent)

V7 model update content: remove European and American face shapes, add Asian faces, and fine-tune multiple scenes and elements at the same time to reduce background blurring in some scenes. (The model drawing effect can be viewed in the image area above) Can be used as a training base model and is very friendly to LORA compatibility Recommended amplification: Swinir_4x, 4x-UltraSharp Recommended size: any size Magnification ratio: maximum not more than 1.55 times (to avoid rendering distortion) To ensure image quality, it is recommended to open ADetailer repair Note: This model is publicly available and can be easily integrated. Please indicate the author. Thank you friends!

Files with SHA256: 87c91e2fb2f50c8fd71c91a733f998947ac1682469e8d61d674ff61592344219