Sign In

Illustration Photo

46

249

159

15

Verified:

SafeTensor

Type

LoRA

Stats

44

0

Reviews

Published

Dec 21, 2025

Base Model

Qwen

Training

Steps: 3,000

Usage Tips

Clip Skip: 2
Strength: 1

Trigger Words

CinnaPhotostration

Hash

AutoV2
FB245EC0D1
Vampiric Essence
Daalis's Avatar

Daalis

Illustration Photos — LoRA

Illustration Photos is a whimsical, colorful LoRA created in collaboration with the original artist Cinnadust, designed to blend photorealistic environments with expressive 2D scratch-style illustration. Using the trigger keyword CinnaPhotostration, this LoRA transforms scenes into playful, storybook-like moments where hand-drawn characters and elements exist inside richly lit, realistic worlds. The result feels magical, cozy, and emotionally readable—like an illustrated dream layered directly onto reality.

At its heart, Illustration Photos celebrates contrast: soft, imperfect lines against crisp lighting, flat 2D subjects floating inside believable spaces, and charming whimsy balanced with cinematic realism. It excels at cute, curious, and imaginative scenarios, staying true to the joyful spirit found throughout Cinnadust’s work.


🔑 Trigger Keyword

CinnaPhotostration

Include this keyword in your prompt to activate the style. (with some of the models, while optional I still find that it makes it stronger, particularly with Z Image Turbo)


  • 0.5 – 0.8 → Light stylistic touch, more natural realism, subtle illustration influence

  • 1.0 (Recommended) → Strong, balanced adherence to the Illustration Photos look

  • 1.2 – 1.5 → More cartoony, expressive, and abstract results with exaggerated charm

Lower strengths allow the base model to dominate, while higher strengths push the image deeper into stylized, illustrated territory.


🎨 What It Works Best With

Illustration Photos performs well across a wide range of subjects:

  • Animals (especially cute or expressive poses)

  • People and characters

  • Places and environments

  • Objects and still-life scenes

  • Narrative or emotional moments

It shines brightest with whimsical, cozy, playful, and magical concepts. As prompts become more rigid, technical, or heavily constrained to a specific image type, LoRA adherence may naturally soften.


🧠 Usage Tips & Best Practices

Based on extensive testing, here are some helpful guidelines:

  • Explicitly hint at a 2D subject
    Use phrases like “2D hand-drawn,” “flat illustrated character,” “sketched with colored pencils or gel pens” to help the LoRA lock onto the intended contrast.

  • Describe the environment with realism
    Detailed lighting, reflections, atmosphere, and materials help reinforce the photorealistic backdrop the style relies on.

  • Expect subtle edge halos
    Soft outlines or slight glow around 2D subjects are a natural artifact of the style and part of its illustrated charm.

  • Keep the subject emotionally clear
    Simple actions and readable expressions produce the strongest results.

  • Avoid overloading the prompt
    Too many conflicting instructions can dilute the LoRA’s influence. Let the style breathe.


🚀 Model Compatibility

This LoRA has been tested and released with support for:

  • Z Image Turbo

  • Flux

  • Qwen

  • Illustrious

It works especially well with base Illustrious and Illustrious 2, showing the expected and healthy variance between trained and tuned checkpoints.


💖 Join the Fun

If you enjoy Illustration Photos, please consider giving it a thumbs up and posting your first generated image on the LoRA page. Seeing what others create helps spread the buzz, fuels inspiration, and keeps this kind of crazy growing.

Thank you for exploring this style, and for helping keep the magic alive.