本人是B站UP主：言萧的AI画室，
主页链接：言萧的AI画室的个人空间_哔哩哔哩_bilibili
此模型是老王炼制，B站UP：大咦把狼
本人仅为代发布模型，以下内容由老王编写：
基于XL1.0炼制。以前那个同名LORA的FOR XL
感叹一下，1.5只需要练1个小时的素材， 现在必须要干到7个小时。
像素的扩大，导致对显卡要求大幅增加。
参数设置也有一些区别了。
Description
使用压缩方法sv_fro，注意当前lora仅支持SDXL
FAQ
Comments (8)
Could we get an english translation? Are there any trigger words?
I'm sorry, there are no trigger words. You can use it directly. Also, due to my limited English proficiency, I'm worried that using technical terms might cause confusion. That's why I haven't translated it into English in the introduction.（by chatgpt translate）
translation by Bing.com
I am only a model for proxy release, and the following content is written by Lao Wang:
Based on XL1.0 refining. The previous same name LORA FOR XL
Exclaimed that 1.5 only needs to practice 1 hour of material, now it must be done for 7 hours.
The expansion of pixels has greatly increased the requirements for graphics cards.
There are also some differences in parameter settings."
Is there anything else I can help you with? 😊
Why is this under 1.5 if its SDXL?
my bad, forgot to change, >.<
This dont work with sdxl 1.0 model!
comfyui can be used
Doesn't work for me either. Error:
NansException: A tensor with all NaNs was produced in Unet. This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.