<h2 id="yfg-crowdform-subjects-formed-from-thousands-of-tiny-human-figures">YFG Crowdform – Subjects Formed from Thousands of Tiny Human Figures</h2>Base Model Krea2 RAW (Primary inference/testing was done on Krea2 Turbo, 8 steps)<h3 id="style">Style</h3>Transforms subjects into forms that appear to be constructed from thousands of tiny human figures. The effect can range from subtle human-texture surface detail to full structural formation, where the entire object, creature, character, or scene looks as though it is made from miniature people arranged into a larger shape.This LoRA is especially strong for surreal concept art, symbolic imagery, creatures, vehicles, architecture, statues, and abstract forms.<h3 id="trigger">Trigger</h3>Official trigger word: <code>YFG-Crowdform style</code>That said, the trigger itself is less important than the descriptive phrases used in prompting. The model responds especially well to phrases such as:<ul><li>formed from thousands of tiny human figures</li><li>dark background</li><li>dramatic golden lighting</li><li>formed from thousands of tiny human figures walking on a white background, casting long shadows</li></ul>…and similar wording that clearly describes the intended crowd-made structure.<h3 id="strength">Strength</h3>Tested from 0.5 to 2.0<ul><li>0.5 – 0.8 → lighter influence, more subject fidelity</li><li>1.0 – 1.3 → strong Crowdform identity</li><li>1.5+ → very assertive, surreal, highly stylized formation effect</li><li>2.0 → maximum takeover territory, useful when you really want the crowd-built look to dominate</li></ul>Adjust based on how fully you want the subject to become “human-figure constructed.”<hr /><h3 id="key-characteristics">Key Characteristics</h3><ul><li>Crowd-built forms – subjects appear sculpted or assembled from countless miniature people.</li><li>Strong surrealism – excellent for conceptual and symbolic prompts.</li><li>Lighting-sensitive – dramatic lighting, especially golden rim / spotlight / low-key lighting, enhances the effect significantly.</li><li>Works on many subjects – vehicles, insects, portraits, abstract objects, environments, statues, creatures, and more.</li><li>Dark or white presentation modes – can work beautifully on dark cinematic backgrounds or on bright white layouts with visible shadow play.</li></ul><hr /><h3 id="inference-settings-used">Inference Settings Used</h3>All provided sample generations were run with:<ul><li>Inference model: Krea2 Turbo</li><li>Steps: 8</li><li>Sampler: <code>er_sde</code></li><li>Scheduler: <code>simple</code></li></ul>These settings appear to work very well with the LoRA.<hr /><h3 id="prompt-ideas">Prompt Ideas</h3><ul><li>“A giant cathedral formed from thousands of tiny human figures, dark background, dramatic golden lighting.”</li><li>“A futuristic beetle made of thousands of tiny walking people, cinematic shadows, surreal concept art.”</li><li>“A heroic portrait formed from thousands of tiny human figures walking on a white background, casting long shadows.”</li><li>“A massive spaceship constructed from miniature human silhouettes, moody atmosphere, golden highlights.”</li><li>“An abstract geometric sculpture built from thousands of tiny people, dramatic side lighting, black void backdrop.”</li></ul><hr /><h3 id="tips-and-tricks">Tips &amp; Tricks</h3><ol><li>Use descriptive formation language<ul><li>This matters more than the trigger. The clearer you describe the “made of tiny people” idea, the better the result.</li></ul></li><li>Lighting helps a lot<ul><li>“dramatic golden lighting,” “strong shadows,” “rim lighting,” and “spotlit darkness” all seem to push the style nicely.</li></ul></li><li>Try both dark and white backgrounds<ul><li>Dark backgrounds emphasize mass and mood.</li><li>White backgrounds with long shadows make the individual tiny figures easier to read.</li></ul></li><li>Best for symbolic / surreal subjects<ul><li>This LoRA really shines when used for creatures, monuments, masks, insects, ships, sculptures, and abstract concepts.</li></ul></li><li>Push strength when needed<ul><li>If the crowd effect is too weak, raising the LoRA weight can help force the transformation more strongly.</li></ul></li></ol><hr />YFG Crowdform is a concept-heavy surreal LoRA that turns subjects into living structures made from thousands of tiny people. For best results, rely less on the trigger alone and more on clear descriptive phrases that explain the formation effect, then pair it with dramatic lighting and either deep dark backgrounds or clean white shadow-heavy layouts.

Files with SHA256: 84fd162579c149809a629f00b032eee5f77e87bb0b3c1025b840cbe8e5720807