cat_loaf-lumina2-lora

<style> .main-div { background-color: #9999ff; padding: 20px; border-radius: 10px; margin-bottom: 20px; } h1 { color: #24292e; margin-top: 0; font-size: 25px; } h2 { color: #24292e; margin-top: 0; font-size: 22px; } h3 { color: #24292e; margin-top: 0; font-size: 20px; } h4 { color: #24292e; margin-top: 0; font-size: 15px; } p { color:#24292e; } a { color: #0366d6; text-decoration: none; } .section-div { background-color: white; padding: 15px; border-radius: 8px; margin: 15px 0; box-shadow: 0 2px 4px rgba(0,0,0,0.1); } .subsection-div { background-color: #bbb; padding: 15px; border-radius: 8px; margin: 15px 0; box-shadow: 0 2px 4px rgba(0,0,0,0.1); } ul { margin-bottom: 0; color: #24292e; } ol { margin-bottom: 0; color: #24292e; } td { margin-bottom: 0; color: #24292e; } code { background-color: #333; padding: 3px 6px; border-radius: 4px; } table { color: #24292e; } </style> <div class="main-div"> <h1>Cat Loaf</h1> <div style="background-color: #ffff99; padding: 20px; border-radius: 10px; margin-bottom: 20px;"> <img src="docs/Images/2025-07-03-114007_NetaAniLumina_Alpha_full_roundnnnn_ep7_s160000_999816718778675.png"> <h3 >V1.1: Kohya Update Edition<br>For Lumina 2</h3> <a href="https://huggingface.co/mrpopsalot/cat_loaf-lumina2-lora"><img alt="Static Badge" src="https://img.shields.io/badge/HuggingFace-yellow?logo=huggingface&logoColor=white&style=for-the-badge"></a> <a href="https://tensor.art/models/881777742356418251"><img alt="Static Badge" src="https://img.shields.io/badge/TensorArt-teal?logoColor=white&style=for-the-badge"></a> <a href="https://www.seaart.ai/models/detail/d1jfdmte878c739qhrj0"><img alt="Static Badge" src="https://img.shields.io/badge/SeaArt-blue?logoColor=white&style=for-the-badge"></a> <img alt="GitLab License" src="https://img.shields.io/gitlab/license/mrpopsalot%2Faidl-omoaishi?gitlab_url=http%3A%2F%2Fgitgud.io&style=for-the-badge"> </div> <div class="section-div"> <h2>Overview</h2> <p>After a couple of painful weeks of trying to get Lumina 2 LoRA training using tools such as Kohya, Diffusion Pipe, and AI Tookit, I finally present you the first fully working concept LoRA for Lumina 2 (and the first LoRA for Neta Lumina).</p> <p>This LoRA teaches the cat loaf sitting techniques to cats. Without the LoRA, cats will struggle to loaf in most positions</p> <ul> <li><b>Trained Models:</b></li> <ul> <li><b>Lumina Image 2:</b> neta-art/lu2 a5nA EP64 S16520 @ 1024</li> <li><b>1.0:</b> neta-art/NetaAniLumina_Alpha Round NNNN EP7 S160000 @ 1024</li> </ul> <li><b>Developed by:</b> Mr Pops Alot</li> <li><b>Model type:</b> LoRA</li> <li><b>License:</b> Apache 2.0</li> </ul><br> </div> <div class="section-div"> <h2>Showcase</h2> <p><img src="docs/Images/EasyGrid_00036_.png"></p> <p><img src="docs/Images/2025-07-03-110059_NetaAniLumina_Alpha_full_roundnnnn_ep7_s160000_452023707135964.png"></p> </div> <div class="section-div"> <h2>Features</h2> <ul> <li>Turn cats into cute loafs </li> <li>Consistent results across different angles, aspect ratios, and art styles</li> </ul> </div> <div class="section-div"> <h2>How to Use</h2> <p>The key trigger phrase is: <code>cat loaf</code></p> </div> <div class="section-div"> <h2>Bias, Risks, and Limitations</h2> <p>This LoRA was made as a test lora to test out training on Lumina 2. While intended to be used as a final product, the LoRA's limited image count may cause issues on out of scope use cases.</p> </div> <div class="section-div"> <h2>Training Details</h2> <p>64 images have been used in the dataset, with another 60 images for regularization. Images used are <b>a mix of own sourced photorealisic pictures and drawn</b> art scraped from Danbooru.</p> <p><b>No synthetic data has been used</b> for the model. Tagging (Danbooru tag style) is <b>all manually done</b>.</p> <div class="subsection-div"> <h3>Training Procedure</h3> <h4>Lumina 2</h4> <ul> <li>Program: Kohya-ss</li> <li>Optimizer: CAME</li> <li>Scheduler: REX</li> <li>LoRA type: LoRA</li> <li>Steps: 140</li> <li>Precision: Full BF16</li> <li>Hardware: 1x RTX 3090 24GB</li> </ul> </div> </div> <div class="section-div"> <h2>FAQ</h2> <ul> <li><b>Will this model be ported to X?</b> This model will be trained on various models when I feel like training on them. </ul> </div> <div class="section-div"> <h2>License</h2> <p>This LoRA model is provided under the <a href="https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/apache-2.0.md">Apache License 2.0</a> license.</p> <!-- <p>Training configurations and dataset will also be released later. You can use these to further train AIDL.</p> --> </div> <div class="section-div"> <h2>Special Thanks</h2> <ul> <li>Ly, and others in the NoobAI discord to for helping me train a LoRA on Lumina and for emotional support.</li> </ul> </div> <div class="section-div"> <h2>Model Card Contact</h2> <ul> <p>I can be contacted on:</p> <li><b>Matrix:</b> @mr_pops_alot:cutefunny.art</li> <li><b>All The Fallen:</b> @MrDoesNothingAlot</li> </ul> </div> </div>