Abstract LoRA, that actually doesn't need much of a coherent prompt – which is why I call it a 'Visual Haiku' LoRA. Optional trigger for all versions: w1z_v1zu4l_h41ku
Version 2 – Macrographic
Version 2 brings more texture and depth. Like a petri dish of nebulas and gooey slime.
Prompt tips: abstract macro photography, rubber, liquids, floral texture, fungi, bacteria
———————
Version 1 – Marbling
Great at textures - especially marbling textures, not so good with humans. Capable of objects.
Handles really high guidance, so don't be afraid to experiment.
Prompt tips: marbling, fiber texture, grainy, rock, vibrant colors, mesmerizing
Description
FAQ
Comments (2)
Some great looking examples- your LORA really seems to get the 'fluid' concept.
Disclaimer: Its maybe a bit too rigid, which means it won't easily translate the concept to other applications, even at lower levels or longer prompts, iit tends to keep the core until it completely breaks it.
I also made the "mistake" of including a single image of a face in the dataset – which means it is a bit overtrained of making people as living X-rays – see it as a feature