LoRA for industrial, technical, cyberpunk-style detailing in a graphic, digital-art illustrated style. Particularly for architecture but works for a range of sci-fi subjects. Not trained on human figures.

Trained on SDXL 1.0 base model using 1024x1024 input images.