What is this ?
Well basically instead of merging lora (method tested here on my other lora page) to get an unique style. This time i tried to merge and mix directly from the dataset. Then trained it. And this is the result.
How to use ?
Well it's a lora style. So no trigger word. Just prompt for it and mess with the weigths. I recommend roughly between 0.5 and 0.8. Below it doesn't show enough effect and above it deep fry real quick.
Description
New dataset, various artwork of mostly stuff under tags such as goth fashion, goth, Use roughly 0.5 strenght and up.