只训练了脸部，用来改善甘城画风的眼睛还原度。
需要配合原画风lora或者猫羽雫甘城的角色lora使用，
全身照2k分辨率以上眼睛效果才能较好，以下是推荐配合用的lora：
https://civarchive.com/models/21301/amashiro-natsuki-style
推荐分层：<lora:amashiroNatsukiV5:1.2:1:1:1:1:1:1:0:0:0:0:0:1:1:1:1:1:1>
https://civarchive.com/models/30791/lycoris-amaki-natsuki-style-for-anything-v5
这个是lycoris模型我没用过，就不试了
推荐权重参考：0.8以下
推荐lora分层参考：<lora:Nachoface:0.8:1,1,1,1,1,1,0,0,0,0,0.9,0,1,1,1,1,1>
【这样的推荐分层配合NatsukiV5可以跑涩图，我本地跑的涩图就不传了，反正应该也传不上来】
Description
FAQ
Comments (3)
English Translation via ChatGPT
"Only the facial part was trained to improve the eye restoration accuracy in the style of Amashiro Natsuki
It needs to be used in conjunction with the original art style, either Lora or the character Lora from Cat's Cradle.
For full-body images with a resolution of 2K or higher, the eye effect will be better. Here are the recommended options to use with Lora:
Lora model: Amashiro Natsuki Style (https://civitai.com/models/21301/amashiro-natsuki-style) Recommended configuration: lora:amashiroNatsukiV5:1.2:1:1:1:1:1:1:0:0:0:0:0:1:1:1:1:1:1
Lycoris Amaki Natsuki Style model (https://civitai.com/models/30791/lycoris-amaki-natsuki-style-for-anything-v5) I haven't personally used this Lycoris model, so I won't try it.
Recommended weight reference: Below 0.8
Recommended Lora configuration reference: lora:Nachoface:0.8:1,1,1,1,1,1,0,0,0,0,0.9,0,1,1,1,1,1
[With this recommended configuration combined with NatsukiV5, you can generate NSFW images. I won't share any examples I generated locally because I can't provide them here anyway.
Can you share the hassakumeina model you used to generate the images and Vae please, if possible.
想问一下用的大模型是什么w