This is an Embedding built for Stable Diffusion 2.0.
Trained on 9 Images of mech/cybersuits Training was done with the Automatic1111 WebUI
I have included all the model files from training but have 4 selected out.
ZiCyb: Highest stepping embedding - https://huggingface.co./Arron17/ZiCyb/resolve/main/ZiCyb.pt
ZiCybb: Slightly lower stepping, gives a slightly different armour pattern - https://huggingface.co./Arron17/ZiCyb/resolve/main/ZiCybB.pt
ZiCybL: Lower Stepping, this can be used to have less armour, or if your subject has a fairly low weight - https://huggingface.co./Arron17/ZiCyb/resolve/main/ZiCybL.pt
ZiCybP: This version tends to bring out Portrait images and also has a slight CGI effect if not negative prompted out - https://huggingface.co./Arron17/ZiCyb/resolve/main/ZiCybP.pt
No Negative prompts were used in the below generations, sometimes it can display some weirdness with eyes and hands, you can use the following negatives to mostly remove it "disfigured hand fingers, claws" Note: This will sometimes generate NSFW images. It is also generally bad at objects, if you try to create a car for example, it will usually just add the colours from the car to the armour.
Default Generations with the "by ZiCyb" phrase:
Prompt: A Cinematic Photograph of Keira Knightley by ZiCyb
