Learned a lot about LoRA training so fixed the amount trained for, increased the amount of images, less epochs. Resulted in a much more usable LoRA that is safe to use at 1 weight.











If you'd like to support me and the LoRAs that I create consider throwing a tip my way @ https://ko-fi.com/kouyasimp
LoRA trained to replicate the character Kouya Aotsuki from the visual novel Morenatsu. Trained on 96 images with character and outfit pieces tagged. Compatible with other style and concept LoRAs. Be aware some example images are from in development versions.
Recommended weights are between 0.8-1. Recommended lower weight when combining with an art style LoRA.
Uses the trigger word Kouya for the character and the following trigger words for his outfit pieces: tank top, armbands, dog tags, pants, speedo
Bugs: The asymmetrical armbands are inconsistent on placement and both appearing.
This LoRA is in active development as I want the quality and consistency to increase.
Learned a lot about LoRA training so fixed the amount trained for, increased the amount of images, less epochs. Resulted in a much more usable LoRA that is safe to use at 1 weight.
1- النموذج المعاد النشر هو فقط لأغراض التعلم والتبادل والمشاركة، حقوق النشر والتفسير النهائي تعود للمبدع الأصلي.
2- إذا رغب المبدع الأصلي للنموذج في استلام نموذجه، يرجى التواصل مع موظفي SeaArt AI عبر القنوات الرسمية للتوثيق والاعتماد. نحن نلتزم بحماية حقوق كل مبدع. انقر للاستلام
