This will be a series of segmented creatures starting with humanoids, Segmented Sally and Segmented Sam.

It seems that Z-Image, and likely many other models, were trained primarily with images of woman, at least with regard to robotic augmentation, cyborgs, androids and the like, and the reason I make this assertion is because even using the term man/male I will often get female characteristics and very light on the sharpness of the male form. However, add just 0.2 of Segmented Sam and the character starts taking shape.

Segmented Sally and Sam were built on 100 images each of varying types of cybernetic parts on a female and male subject respectively.

With the variety of images provided for training you'll be able to assemble crisp, extraordinarily detailed, mechanical/biological subjects, or just put a complex prompt in and let it fly to see the changes that take effect.

While the source images weren't designed to provide material I did allow for the options, making sure to include references so that any augmentation of body parts is likely to align properly.

This LoRA is not intended to produce a real world photographic quality image and is deliberately guided towards producing a 3d effect.