First and only LoRa attempt.

Trained and generated using Homochi-XL.

Version 1 images generated with josman_style, leaning towards a more flat, comic style.

Version 2 images generated with no specific tag (trained on josman_style) combining Bara Illustrious lora at 0.2, Josman LoRA at 0.6 for a more detailed style and shading.

Both versions generally ran around 25-50 steps at 5.0 CFG.