用戶LoRA
miratsu \(miratsu169\), chibi, full body
My first creation of artstyle LoRAs
This LoRA isn’t perfectly trained — expect issues
Add 'detailed background' prompt for a more complex bg
Trained with Kohya-SS
Base Model: Illustrious-XL 2.0 (i should've choose 0.1 bc it was better :<)
53 images, 20 repeats, 2 epochs, 1,040 steps
