<h1 id="electric-vehicle-interior-lora-model-wituwt1jo">Electric Vehicle Interior LoRA Model</h1><h2 id="overview-8a5fqj0va">Overview</h2><p>"Electric Vehicle Interior" is a specialized LoRA model designed to enhance the generation of modern car interiors, with a particular focus on electric vehicles. This model addresses the limitations of existing models in producing contemporary automotive interior designs.</p><h2 id="key-features-7wzurl4vs">Key Features</h2><ul><li><p>Tailored for generating up-to-date car interior images</p></li><li><p>Specifically trained on electric vehicle interiors</p></li><li><p>Enhances detail and modernity compared to standard models</p></li></ul><h2 id="development-background-n0kdnx1wi">Development Background</h2><p>As an automotive designer, I noticed that the Flux 1 series models, while excellent at improving image details, tended to generate outdated car interiors. To address this, I curated a dataset of modern electric vehicle interiors and used it to train this LoRA model. The goal was to create a tool capable of producing more contemporary and innovative interior designs, especially for electric vehicles.</p><h2 id="usage-guide-klhc1vi12">Usage Guide</h2><ul><li><p>Compatible with ComfyUI</p></li><li><p>Installation: Place the model in <code>./Comfyui/models/loras/</code></p></li><li><p>Application: Use "Load Lora" as you would with SD1.5 or SDXL models</p></li><li><p>Use guidance 3.5</p></li><li><p>Steps &gt;=25</p></li><li><p>Recommended <span style="color:rgb(201, 209, 217)">resolution is </span><strong><span style="color:rgb(201, 209, 217)">1344x768</span></strong></p></li><li><p><span style="color:rgb(201, 209, 217)">When using, ensure the resolution is 1344x768. This is not an issue with Flux1, but rather a problem from my training. Other resolutions are yet to be tested, and you're welcome to conduct your own tests.</span></p></li></ul><h2 id="performance-notes-wi7x40tks">Performance Notes</h2><ul><li><p>No specific trigger words required</p></li><li><p>Best results achieved with more detailed prompts</p></li><li><p>Optimal performance may require longer prompts (possibly due to T51_1 XXL base or training parameters)</p></li></ul><h2 id="training-specifications-nmsr0c8ra">Training Specifications</h2><ul><li><p>Dataset: 383 images</p></li><li><p>Resolution: 1024</p></li><li><p>Maximum training epochs: 50</p></li><li><p>Network module: Lokr</p></li><li><p>Network dimensions: 4</p></li><li><p>Factor: 6</p></li><li><p>Optimizer: AdamW8bit</p></li><li><p>UNet learning rate: 2e-4</p></li><li><p>Training UNet only</p></li><li><p>Additional parameters:</p></li></ul><p>min_snr_gamma = 5</p><p>multires_noise_iterations = 10</p><p>multires_noise_discount = 0.3</p><h2 id="advanced-usage-g2yjj2bpr">Advanced Usage</h2><ul><li><p>Training script: <a target="_blank" rel="ugc" href="https://github.com/kohya-ss/sd-scripts">kohya-ss/sd-scripts</a></p></li><li><p>For broader compatibility, use the merge.py script from kohya-ss/sd-scripts to combine with <a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/FLUX.1-dev">Flux </a><a target="_blank" rel="ugc" href="http://1.dev">1.dev</a> than you can test ConntrolNet or Ip-adapter published by <a rel="ugc" href="https://huggingface.co/XLabs-AI">Xflux</a></p></li><li><p>Compatible with XLabs-AI/flux-controlnet-canny-v3 in ComfyUI</p></li><li><p>This LoRA model represents an experimental approach to improving electric vehicle interior visualization, offering designers and enthusiasts a tool to explore modern automotive design concepts.</p></li></ul><p></p><h3 id="note:-wjhw9axg0">Note:</h3><p>This model card was written with the assistance of <a target="_blank" rel="ugc" href="https://claude.ai/">Claude 3.5</a>, an AI language model developed by Anthropic.</p>

electric_vehicle_interior

<p>This is an experimental version aimed at exploring the capabilities of the FLUX1 series models. It serves as an initial release to test the concept and effectiveness of the approach. Future versions with improved performance and refinements may be released based on the feedback and results from this experimental model.</p><p>This LoRA model represents an experimental approach to improving electric vehicle interior visualization, offering designers and enthusiasts a tool to explore modern automotive design concepts.</p>

Files with SHA256: 3ce6e4de306c5bca5cbfb84610813d5827377e263fdb088eceecebcc5ac1ee0c