<h1 id="light-illustration-night-illustration">Light Illustration / Night Illustration</h1>An experimental LoRA that teaches two opposite concepts within a single model, allowing the image to morph depending on which trigger token is used. It does not simply change the time of day, brightness, or contrast. Instead, it attempts to morph the image itself into a different visual world based on the chosen concept.<h2 id="main-triggers">Main Triggers</h2><ul><li><code>nightillu</code></li><li><code>lightillu</code></li></ul>You can use either token individually, or experiment with both together (not fully tested on Krea2 yet).<hr /><h1 id="recommended-lora-weight">Recommended LoRA Weight</h1>0.5 – 1.5If the effect is too weak or barely visible, consider using a weight above 1.0<hr /><h1 id="how-to-use">How to Use</h1><h2 id="1.-no-trigger">1. No Trigger</h2>Attempts to produce a neutral effect.<hr /><h2 id="2.-nightillu">2. Nightillu</h2>Transforms the image into a dark ethereal world.On Anima,The trigger token alone is sufficient; descriptive prompt is optional.The weight and position of the trigger token can affect the result. Placing the token near the beginning of the prompt generally produces a stronger effect.On Krea2, the token alone may not be sufficient. Descriptive natural language prompts generally produce stronger and more consistent results.Examples:<pre><code>nightillu,
the image morphs into a dark ethereal world.
The atmosphere becomes mysterious and silent.</code></pre><pre><code>nightillu,
the image morphs into a dark ethereal world with symbolic arcane runes, stars, and moonlight.
The atmosphere becomes mysterious and silent.</code></pre><pre><code>nightillu,
the image morphs into a cursed dark ethereal world with swirling dark magical energy.
The atmosphere becomes intense.</code></pre><hr /><h2 id="3.-lightillu">3. Lightillu</h2>Transforms the image into a radiant luminous world.On Anima,The trigger token alone is sufficient; descriptive prompt is optional.The weight and position of the trigger token can affect the result. Placing the token near the beginning of the prompt generally produces a stronger effect.On Krea2, the token alone may not be sufficient. Descriptive natural language prompts generally produce stronger and more consistent results.Examples:<pre><code>lightillu,
the image morphs into a radiant luminous iridescent world.
The atmosphere becomes warm, sacred, and brilliant.</code></pre><pre><code>lightillu,
the image morphs into a radiant luminous iridescent world with holy symbols and divine angelic motifs.
The atmosphere becomes warm, sacred, and brilliant.</code></pre><pre><code>lightillu,
the image morphs into a radiant luminous iridescent world.
The atmosphere becomes luxurious, regal, and brilliant.</code></pre><hr /><h2 id="4.-mixing-both-triggers">4. Mixing Both Triggers</h2>Using both triggers together may produce interesting and unexpected results.Example:<pre><code>(nightillu:0.8)
(lightillu:1.1)</code></pre>Not fully tested on Krea2 yet!<h3 id="note">Note;</h3>(1)Recently, I found that the position of the trigger tokens can affect the results. It's probably some technical tokenizer stuff that I don't fully understand.<hr /><h2 id="5.-prompt-harmony">5. Prompt Harmony</h2>Consider the harmony of your entire prompt.Descriptions of lighting, color, atmosphere, mood, and environmental effects can either reinforce, compete with, or even override the intended transformation.<hr /><h1 id="feedback-welcome!">Feedback Welcome!</h1>I'd love to see what you create with this LoRA!Feel free to share:<ul><li>Your generated images</li><li>Your prompt discoveries</li><li>Your favorite settings</li><li>Any interesting behaviors or unexpected results</li></ul>Every piece of feedback helps improve future versions. ❤️

Files with SHA256: fac62ceb7b400bdb3943d269e34bc10b1851c3f2d1a058d5234652eb90ece290