This model was trained on 1018 images/300 step, so in total for 305400 step.

All the images were created with ADetailer.

If you don't/do want cat ears just add it to the negative/positive prompt, nothing else needed!

THIS IS AN EXPERIMENTAL LORA

Images were gather with a custom script from tiktok. Further LoRa will be published with this method to polish training image collecting

Sources: linktr.ee/devdar_ia