Dreams and nightmares aesthetic lora for z-image-turbo. The lora will effectively change your regular prompts to a different style but prompting specifically for "dreamlike" or "nightmare" scenes will provide better results.

NOTE: works on BOTH the Base and Turbo models!

Trained on a dense 350 image dataset. Images include dream scenes, nightmare scenes, various creatures and items, comically large items, deformed or mutant people, some nfsw with aesthetic change and character detail implementation (such as luminescant veins under skin), dynamic lighting shots, AND MUCH MUCH MORE.

WARNING: LoRA can be too strong in some cases. i highly suggest starting at .2 and working your way up in strength until you find your preferred setting.

my recommended strength: 0.4 - 0.8

Also NSFW but not like you think. Youll see.