This LoRA is meticulously designed to generate Pyramid Head ( pyramidhd), a charater from Silent Hill.

Available for Krea2 and Anima.

Recommended LoRA Strength Krea2: 0.8 – 1.1

Recommended LoRA Strength Anima: 0.7 - 1.0

Trigger words: pyramidhd

Base Character: KREA2 pyramidhd, man with a large dark brown pyramid-shaped metal helmet with rivets and a jagged edge,

ANIMA: pyramidhd, 1boy, male focus, solo, muscular male, pyramid helmet, metal helmet, faceless male, rusty, perforated panel, rivets,

Default Outfit (Optional): KREA2: tattered dark gray apron-like garment,

Piece by piece: KREA2

• Torso: a bare muscular torso with pale skin scarred with red cuts

• Waist: a long tattered leather skirt in brown, gray-brown or olive-green, sometimes with cloth bands wrapped around the waist

• Weapon: a huge rusted great knife with a hooked handle and a wide straight blade

ANIMA:

• Torso: topless male, bare pectorals, abs, scars, red cuts, blood splatter,

• Waist: skirt, long skirt, brown skirt, leather skirt, tattered clothes, cloth bands,

• Weapon: holding weapon, great knife, huge weapon, rusty weapon, hooked handle,

Tips:

KREA2: write your prompt as natural sentences, starting with the trigger word, like "pyramidhd, A muscular man..." and then "He is wearing ...". Quality tags and negative prompts are not needed. Include "his entire head hidden beneath a large rusty metal pyramid helmet" to keep the helmet covering the whole head.

ANIMA: use Danbooru-style tags with the trigger word first. Quality tags: "masterpiece, best quality". The LoRA was trained with mixed captions (tags plus a natural-language sentence), so you can add one short sentence after the tags.

Please check the gallery images to see the exact prompts and settings I used to get an idea of how it behaves. I'd love to see your generations, so feel free to share your results in the comments!

Training Base Models

This LoRA was trained using the following base models:

• krea2_raw_bf16

• anima-base-v1.0

Tested in Forge Neo UI.