Update 2025-01-03: Added contrast slider
I noticed that my previous Amnesia + Avilia LoRA always looks too warm, so I made some patch LoRAs to fix that.
No triggers needed. LoRA weights can be set from -2 to 2, depending on your use case and your base model, but please be careful that larger weights may change the result composition.
Color Temperature preview (色温):
positive weights make pictures colder
Saturation preview (饱和度):
positive weights make pictures less saturated
Brightness preview (亮度):
positive weights make pictures brighter
Contrast preview (对比度):
positive weights make pictures less contrast
Description
FAQ
Comments (11)
amazing lora's 10/10, shall test on v-pred later but so far they do exactly as advertised, just missing a contrast ver xD
nice work!
一个饱和度lora能有效缓解角色tag带来的过度鲜艳问题，多谢了！
这种lora是怎么训练的呢？
搜索关键词《差异法》，现在正经slider应该是要用这个：https://github.com/rohitgandikota/sliders，但是我一直没学，这几个我是用多个单图差异法融起来的
@Akiseki 学到了，谢谢！
@Akiseki 可以冒昧问一下用了哪些差异丹做融合啊？理论上不是一张图差异练一个lora就有很好的效果了吗？
@lse14 我的意思是10张图各自差异训练了10个色温，然后融合成一个，随机挑的图，融一下感觉容错率高一点（当然也有可能多此一举了）
@Akiseki 原来如此，受教了
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.