jarod2212/LumaBlend_Anima

LumaBlend Lora for Anima Checkpoint

Prompt

-

Prompt

-

Prompt

-

Prompt

-

Prompt

-

Prompt

-

Model description

LumaBlend Anima – LoRA Description

Trigger word: anylum Recommended weight: 0.6 – 1.0

LumaBlend Anima is a LoRA designed to enhance ANIMA with a soft, luminous, semi‑realistic aesthetic, while maintaining strong prompt adherence and avoiding face‑locking or morphological override. It performs exceptionally well on portraits, half‑body shots, and full‑body compositions, delivering clean rendering, smooth textures, and consistent color grading.

It emphasizes:

natural / soft cinematic lighting

smooth, realistic skin textures

uniform color grading

depth and atmospheric presence

stable, clean frame composition

Thanks to its balanced training, it does not overpower the base model and adapts easily to different subjects while preserving an elegant, professional visual identity.

⭐ Recommended Settings (ANIMA)

Sampler: er_sde

Scheduler: simple or normal

CFG: 4

Steps: 30

LoRA weight: 0.6 – 1.0

Resolution: 768–1024 on the long side

VAE: ANIMA’s native VAE

Trigger: place anylum at the beginning of the prompt

EXAMPLE PROMPTS FOR USE:

1. anylum, cinematic portrait of a young woman in soft daylight, natural skin texture, shallow depth of field, semi‑realistic style, detailed eyes, subtle smile, elegant composition

2. anylum, full‑body shot of a man in casual modern outfit, standing in warm sunset light, realistic proportions, soft shadows, semi‑realistic rendering, balanced color grading

3. anylum, close‑up portrait of a fantasy character with glowing eyes, intricate costume details, soft rim light, semi‑realistic texture, atmospheric background, expressive mood

4. anylum, artistic full‑body composition of a dancer mid‑motion, dynamic pose, flowing fabric, cinematic lighting, semi‑realistic anatomy, smooth color transitions, detailed highlights

Trigger words

You should use anylum to trigger the image generation.

Download model

Download them in the Files & versions tab.