Lora trained in recreate the Maeshima Ryo's style.
If you ask for the unknow model used in some images, is just a mix of Hassaku XL (Hentai) and MomoiroPony1.2 at 50-50
Version 1
Learning from my last lora. I pruned some tags to make it more like the style. But this time I used around 2000 images in the data set... maybe was excessive, It took around of 16 hours and I had to stop the process because it still remain around 50 hours...
Theres an issue that it is so unstable in the Pony base model (Pony v6) but works rlly well in merges or derivated models, so please if the model doesn't work at all, try with other model based in Pony v6 or increase the weight
Recommended weight: 1-1.2