DMNS MixedMedia Collage LoRA

Deconstruct. Rebuild. Collage the chaos.

A versatile LoRA trained on Higgsfield.ai's Mixed Media preset, designed to transform any subject into a handcrafted collage masterpiece. Whether you’re generating demonic horror characters, avant-garde fashion editorials, or surreal pop-art posters, this model wraps your creations in ripped paper layers, xerox overlays, tape seams, burn marks, and raw sketch elements — evoking a rebellious, zine-like DIY aesthetic with AI-level precision.

This LoRA doesn’t just apply a style — it tears your image apart and reconstructs it into an intricate, chaotic visual composition, fusing digital detail with analog imperfection.
It’s fully adaptive: characters, fashion, objects, abstract concepts — all get the same deconstructed, visceral mixed-media treatment.

Key Features:

• Ripped paper edges, layered fragments, and tape overlays

• Collage-style compositions with sketched symbols & doodles

• Glitch-inspired distortions & analog imperfections

• Burn marks, ink splatters, xerox textures

• Works across characters, fashion, posters, album covers, objects, and surreal concepts

Training Base:

• Trained on Juggernaut XL

• No trigger words necessary

• Since Juggernaut XL uses caption-based learning, this LoRA performs best with natural language prompts — don’t tag-spam, just describe the scene clearly.

Recommended Settings:

• LoRA Strength: 0.5 – 1

• Sampler: DPM++ 2M Karras / Euler A

• CFG Scale: 7.5 – 8

Creator:

virtualdmns (trained on images generated with higgsfield.ai)