A review/like/follow helps keep me motivated! I'd love to see what you create and hear your feedback! ೕ(•̤̀ᵕ•̤́๑)ᵒᵏᵎᵎᵎᵎMixtape Maximalism – a visual aesthetic rooted in the chaotic, unapologetically loud design sensibilities of '90s to early 2000s hip-hop mixtape and bootleg album covers, with a strong influence from No Limit Records' iconic album art. The goal of this LoRA is to reproduce that distinctive over-the-top style with questionable editing, where design restraint is nonexistent and everything competes for your attention. <h3 id="usage-tips-eclcmkol9">Usage Tips</h3><ul><li>Weight: 0.7-1</li><li>Trigger: <code>mixtape maximalism style</code></li><li>I generally start the prompt with <code>An album cover in a mixtape maximalism style.</code>, followed by details about the main subject(s), then the text, and finally the background/any other misc elements</li><li>Feel free to experiment, but the source image text elements were captioned in the format: <code>The artist name "NAME" is written [position] in a [font] with [style/effect(s)]</code>, and <code>The album title "TITLE" ...</code><ul><li>E.g., <code>The artist name "Oaky" is written across the top in a bold sans-serif font with a molten gold texture fill, a thin black stroke, a white glow, and lens flare effects.</code></li></ul></li><li>Text effect triggers: fill, stroke, drop shadow, glow, 3d, beveled, embossed</li><li>Misc triggers: <code>lens flare effect</code> (they'll probably show up anyway, but if you want more...)</li><li>Use <code>A parental advisory label is positioned [position]</code> if you'd like to include a parental advisory label<ul><li>E.g., <code>A parental advisory label is positioned in the bottom right corner</code></li></ul></li></ul><h3 id="notes-g7ukvm5hq">Notes</h3><ul><li>The LoRA captures the overall style I was aiming for, but text style/effect adherence isn't consistent at all. Looking to improve this in v2.</li><li>Showcases images have been upscaled by 1.5 with 0.15-0.2 denoising</li></ul><h3 id="training-info-p7g466wtb">Training Info</h3><ul><li>Images: 19</li><li>Steps: 3000</li><li>Network Rank/Alpha: 32</li><li>Learning Rate: 1e-4</li><li>Tool: <a target="_blank" rel="ugc" href="https://github.com/ostris/ai-toolkit">ai-toolkit</a></li><li>Captioning: Florence-2-large-PromptGen v1.5, Claude 3.7, manual editing</li></ul>

Files with SHA256: 239d5574dc0a40dc15e658e97ed3aad2654705c78ebe7590eace2c2ed39f01d5